2020 Book Biomaterials-AndMicrofluidics

Advances in Experimental Medicine and Biology 1230
J. Miguel Oliveira
Rui L. Reis Editors
Biomaterials- and
Microfluidics-Based
Tissue Engineered
3D Models
Advances in Experimental Medicine
and Biology
Volume 1230
Series Editors
Wim E. Crusio, CNRS and University of Bordeaux UMR 5287, Institut de
Neurosciences Cognitives et Intégratives d’Aquitaine, Pessac Cedex, France
John D. Lambris, University of Pennsylvania, Philadelphia, PA, USA
Heinfried H. Radeke, Clinic of the Goethe University Frankfurt Main,
Institute of Pharmacology & Toxicology, Frankfurt am Main, Germany
Nima Rezaei, Research Center for Immunodeficiencies, Children’s Medical
Center, Tehran University of Medical Sciences, Tehran, Iran
More information about this series at http://www.springer.com/series/5584
J. Miguel Oliveira • Rui L. Reis
Editors
Biomaterials-
and Microfluidics-Based
Tissue Engineered 3D
Models
Editors
J. Miguel Oliveira Rui L. Reis
3B’s Research Group 3B’s Research Group
University of Minho University of Minho
Guimarães, Portugal Guimarães, Portugal
ISSN 0065-2598 ISSN 2214-8019 (electronic)

Advances in Experimental Medicine and Biology
ISBN 978-3-030-36587-5 ISBN 978-3-030-36588-2 (eBook)
https://doi.org/10.1007/978-3-030-36588-2
© Springer Nature Switzerland AG 2020

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher nor
the authors or the editors give a warranty, expressed or implied, with respect to the material
contained herein or for any errors or omissions that may have been made. The publisher remains
neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
In recent years, we have assisted a new impetus in tissue engineering and

personalized regenerative medicine research due to adoption of emerging
technologies for fabrication of functional microtissues, which best emulate
human condition and allow minimizing animal experimentation. In this
respect, advanced biomaterials and microfluidics are truly game changers.
This book on biomaterials- and microfluidics-based tissue-engineered 3D
models provides different chapters concerning the fundamental aspects, chal-
lenges, and progress on the biomaterials and microfluidics applications in the
development of 3D tissue-engineered in vitro models. The type and role of
biomaterials and microfluidics systems considered in the designing of in vitro
models are overviewed. The microfluidics platforms used in processing of
biomaterials are succinctly addressed, while current challenges relating tis-
sue-, organ-, and body-on-a-chip models are explored in depth and relevant
examples given. Other hot topics discussed relate to the current applications
of the 3D models in the central nervous system, angiogenesis, and drug dis-
covery and development. The future trends in the field are concisely pro-
vided. This book will serve as a core reference for multidisciplinary students
and a wide range of established researchers and professionals working in
tissue engineering and precision medicine fields.
Guimarães, Portugal J. M. Oliveira
v
Contents
1 Microfluidic Devices and Three Dimensional-Printing

Strategies for in vitro Models of Bone�� 1
F. Raquel Maia, Rui L. Reis, Vitor M. Correlo,
and Joaquim M. Oliveira
2 Microfluidics for Processing of Biomaterials�� 15
Luca Gasperini, Alexandra P. Marques, and Rui L. Reis
3 Organs-on-a-Chip�� 27
Lucie A. Low, Margaret Sutherland, Nadya Lumelsky, Seila
Selimovic, Martha S. Lundberg, and Danilo A. Tagle
4 Engineering Patient-on-a-Chip Models for Personalized
Cancer Medicine�� 43
David Caballero, Rui L. Reis, and Subhas C. Kundu
5 Biomaterials and Microfluidics for Liver Models�� 65
Alain da Silva Morais, Joaquim M. Oliveira, and Rui L. Reis
6 Microfluidic Systems in CNS Studies�� 87
Anna Andrzejewska and Miroslaw Janowski
7 Microfluidics for Angiogenesis Research �� 97
Lígia Costa, Rui Luís Reis, Joana Silva-Correia,
and Joaquim Miguel Oliveira
8 Biomaterials and Microfluidics for Drug Discovery
and Development�� 121
Mariana R. Carvalho, Roman Truckenmuller, Rui Luís Reis,
9 Dynamic Culture Systems and 3D Interfaces Models
for Cancer Drugs Testing�� 137
Diogo C. Fernandes, Raphaël F. Canadas, Rui L. Reis,
10 Nanoparticles and Microfluidic Devices
in Cancer Research�� 161
F. Raquel Maia, Rui L. Reis, and Joaquim M. Oliveira
Index�� 173
vii
Microfluidic Devices and Three
Dimensional-Printing Strategies 1
for in vitro Models of Bone
F. Raquel Maia, Rui L. Reis, Vitor M. Correlo,

Abstract bone angiogenesis. More recently, 3D printing

Bone is a complex and highly dynamic tissue, has been pursued to produce more intricate
which has been worldwide studied, from fun- microfluidic devices and engineered tissues in
damental biology to tissue engineering fields. a single step. The ability to print spatially con-
Even so, current in vitro models do not truly trolled structures composed of different bio-
replicate the native bone tissue environment. materials, growth factors and cells caught the
For so, new and improved in vitro tissue mod- attention of scientists for the development of
els are necessary to obtain more reliable data, more efficient in vitro models. Additionally, it
not only in a development point of view, but allows obtaining microfluidic devices and/or
also to fasten the translation of new drugs into engineered tissues with the desired architec-
the clinics. In this reasoning, tissue- ture within a small amount of time and with
engineering strategies were applied to develop reduced costs. Recently, the use of high-reso-
mimetic and three-dimensional (3D) lution scanning boosted the production of
microenvironments, which were associated patient-specific implants. Despite the difficul-
with microfluidic devices for the development ties associated with 3D printed structures that
of more complex and realistic systems. Such still need to be overcome, it has been proven
devices mimic blood vessels that are present to be a valuable tool to accomplish a new gen-
in the native tissue, thus enabling the study of eration of 3D bioprinted bone-on-a-chip
complex biological mechanism as such as platforms.
F. R. Maia (*) · R. L. Reis · V. M. Correlo Keywords

J. M. Oliveira Bone tissue engineering · Microfluidic
3B’s Research Group, I3Bs – Research Institute on devices · Three-dimensional models ·
Biomaterials, Biodegradables and Biomimetics, Bioprinting · In vitro models
University of Minho, Headquarters of the European
Institute of Excellence on Tissue Engineering and
Regenerative Medicine, Guimarães, Portugal
ICVS/3B’s PT Government Associate Lab,
Braga, Guimarães, Portugal
The Discoveries Centre for Regenerative and
Precision Medicine, Headquarters at University
of Minho, Guimarães, Portugal
e-mail: raquel.maia@i3bs.uminho.pt
© Springer Nature Switzerland AG 2020 1

J. M. Oliveira, R. L. Reis (eds.), Biomaterials- and Microfluidics-Based Tissue Engineered 3D
Models, Advances in Experimental Medicine and Biology 1230,
https://doi.org/10.1007/978-3-030-36588-2_1
2 F. R. Maia et al.
Highlights have been pursued to understand the behavior of

osteoclast to new biomaterials (Maia et al. 2018).
• Microfluidic devices and 3D printing have Within this setting, where 3D and several types of
great potential for the development of highly cells are considered it is possible to obtain more
realistic bone tissue in vitro models. reliable results. Even so, most 3D in vitro models
• The most promising approaches to obtain were cultivated under static conditions, which
bone tissue in vitro models that better mimic resulted in poor nutrient and waste exchange. To
the native tissue are reviewed. overcome these issues, bioreactors were devel-
• Relevant examples of the approaches reviewed oped allowing not only to improve nutrients per-
are highlighted. fusion and waste elimination but also comprise
mechanical stimuli (Bouet et al. 2015).
Despite the incredible progress achieved, most
1.1 Introduction models only partly mimic the in vivo scenario. In
fact, it lacks the presence of vasculature
Bone is a very complex and highly dynamic tis- structures, and other features, as such as specific
sue continually renewing and controlled by sev- architectures and tissue-tissue interaction. For so,
eral transcription regulatory cascades. Bone innovative technological platforms and tech-
tissue is composed mainly of four different types niques have become available, as the microfluidic
of cells, osteoblasts, osteoclasts, osteocytes and devices (Aziz et al. 2017) and 3D printing strate-
bone lining cells (Florencio-Silva et al. 2015). gies (Bhattacharjee et al. 2016; Jariwala et al.
Osteoblasts and osteoclasts are involved in the 2015). Within microfluidic devices it is possible
bone remolding process, where osteoblasts are to obtain intricate multi-compartment co-
responsible for bone tissue neoformation, while cultures, connected by micro-channels that
osteoclasts are responsible for bone tissue resorp- mimic the vasculature (Carvalho et al. 2018;
tion. Osteocytes act as mechanosensors and are Jusoh et al. 2015). All this, using low reagents’
involved in the initiation of the bone remodeling amounts and cell numbers. In the case of 3D
process (Florencio-Silva et al. 2015). In the case printing, it has allowed pushing forward the
of bone lining cells, there is evidence that these development of in vitro models of bone, enabling
cells are a major source of osteoblasts in adult- the production of 3D printed microfluidic devices
hood (Matic et al. 2016). This intricate scenario and/or tissues with the desired complexity in a
has been studied worldwide to try to develop new single step.
therapeutic approaches to bone diseases. This chapter aims to overview the recent
Currently, it is common knowledge that two- development on 3D bone tissue engineered in
dimensional (2D) culture in vitro models do not vitro models and highlight the most relevant
truly replicate bone tissues (Duval et al. 2017). examples of the progress achieved. For so, four
This led scientist to develop new in vitro models different topics will be explored: (i) microfluidic-
using Tissue Engineering approaches. For exam- based in vitro models of bone; (ii) techniques
ple, Lima et al. (Lima et al. 2014b) developed a used to obtain 3D printed microfluidic devices;
scaffold with micro organized and architected (iii) 3D bioprinting-based bone tissue models;
structures that support cell adhesion and prolif- and (iv) 3D bioprinted microfluidic in vitro
eration. In fact, tissue-engineering can involve models.
the combination of cells and scaffolds resulting
in a three-dimensional (3D) environment that
better mimic the reality of bone tissue (Lima 1.2 Microfluidic-Based in vitro
et al. 2014a; Bouet et al. 2015). Amongst the dif- Models of Bone
ferent 3D strategies followed to develop new bio-
materials to guide bone tissue formation, Microfluidic devices have been used to mimic the
osteoblasts are usually used, while other cells physiological environment of tissues and organs.
like osteoclast are ignored. For so, new studies In fact, the ability to mimic blood vessels has step
1 Microfluidic Devices and Three Dimensional-Printing Strategies for in vitro Models of Bone 3
forward the development of efficient 3D in vitro sprouted, invading the mineralized matrix
models. Moreover, such devices have the (Fig. 1.1d).
capability to spatial and temporally control fluids As demonstrated, such platform offers new
and physical features. Additionally, they allow relevant models for different areas, namely,
screening several conditions through screening of new drugs, new biomaterials and
parallelization, using reduce amounts of reagents/ study biological mechanisms of disease (e.g. can-
cells. Altogether, these features boost its use in cer). For example, new in vitro models of bone
combination with in vitro models’ strategies for were developed to study the interaction of bone
drug discovery as well as in a development point cells with newly developed biomaterials aiming
of view. In this reasoning, microfluidic devices to speed up the development of novel orthopedic
have been showing to be powerful tools that implants (Lee et al. 2012). In a different study,
allow conducting experiments with high Hao et al. (2018) showed the development of a
sensitivity and within small amounts of time. microfluidic-based in vitro model of breast cancer
One of the most appealing characteristics of bone metastasis composed of a simultaneous-
microfluidic devices is the possibility to study growth-and-dialysis mechanism. The developed
complex biological mechanisms in vitro, such as mechanism allowed the exchange of nutrients
bone angiogenesis, as demonstrated by Jusoh and and waste continuously through the dialysis
colleagues (Jusoh et al. 2015). They developed a membrane. Moreover, they were able to obtain a
vascular network within a matrix composed of long-term culture of highly mineralized
fibrin and hydroxyapatite to obtain a microfluidic- osteoblastic tissue. Additionally, this microfluidic
based in vitro model of vascularized bone tissue tool allowed the co-culture of metastatic human
(Fig. 1.1). The designed microfluidic presented breast cancer cells and osteoblastic tissue,
channels separated by gaps using columns observing similar results as in vivo studies. As
(Fig. 1.1c), through where, blood vessels example, researchers observed that metastasis-
Fig. 1.1 Microfluidic based vascularized bone model. (a) umns from fibrin matrix with HA to mimic bone tissue;
Scheme of bone angiogenesis; (b) Fibrin matrix with (d) image showing blood vessels sprouting towards mimic
hydroxyapatite (HA) to mimic the entrapment of blood bone tissue. (Adapted with permission from (Jusoh et al.
vessel; (c) scheme of developed microfluidic device show- 2015). Copyright © 2015, Royal Society of Chemistry)
ing channels with endothelial cells (EC) separated by col-
4 F. R. Maia et al.
suppressed breast cancer cells proliferated at sectional slices; and finally, (iv) printed layer-by-
slower rates in opposite to 2D cultured cells, layer (Fig. 1.2) (Ho et al. 2015; Waheed et al.
where it was observed higher proliferation rates. 2016). Furthermore, 3D printing allows the
Moreover, those cells did not invade the matrix of standardization of procedures which results in
osteoblastic tissue. In the case of normal more reproducible devices as compared with soft
metastatic breast cancer cells, after 2 weeks they lithography. Even so, the cost, resolution and
observed that cells were capable to invade the materials necessary to print the microfluidic
osteoblastic tissue matrix and degraded it, devices need to be evaluated to guarantee optimum
reaching distant locations. Moreover, Veerla et al. results.
studied a drug delivery system within a
microfluidic channel for osteoblastic
differentiation (Veerla et al. 2018). For that, 1.3 echniques to 3D - Printing
T
researchers developed graphene oxide/alginate Microfluidic Devices
beads loaded with risedronate. The channel with
the beads showed controlled delivery of Nowadays, the fabrication of microfluidic devices
risedronate inducing osteoblasts differentiation, is predominantly obtained by casting-based 3D
without the need for additional supplements. molding approaches (soft lithography) of
Yet, although the research associated costs were PDMS. Nevertheless, this technique presents
very much reduced, the production of such device some constraints, such as difficulty to fully
using the traditional soft lithography technique can automatize the molding process, namely curing,
be very time consuming and expensive. In fact, the assembly and bonding. In this sense, a process that
development process involves several steps was supposed to be low cost is, in fact, expensive
including the device’s design using an engineering and time-consuming. In this reasoning, a new
drawing software, the fabrication of a master mold technique is being used to overcome the difficulties
according to the design and its replication using along the manufacturing process. 3D printing
polydimethylsiloxane (PDMS). After the curing allows obtaining microfluidic devices, printing
time, the PDMS is removed and cut into the desired layer by layer from a CAD file, mainly through
shape. Finally, oxygen plasma is used to improve three different techniques: fused deposition
PDMS-glass bound (Ho et al. 2015). Furthermore, modelling, inkjet modelling and stereolithography
the structural complexity achieved during its (Fig. 1.3), which will be discussed briefly below
processing it is reduced compared with the tissue (Macdonald et al. 2017; Bhattacharjee et al. 2016).
complexity observed in vivo. Altogether, these In a study comparing these three techniques
drawbacks encouraged a new line of thinking, (Macdonald et al. 2017), it was demonstrated that
namely the use of 3D-printing to produce fused deposition modelling was the most suitable
microfluidic devices (Bhattacharjee et al. 2016). for the development of microfluidic devices that
3D-printing is an additive technology with which it worked as micromixers, showing a rapid mix of
is possible to print complex and highly realistic different solutions. In the case of inkjet modelling,
structures within a small amount of time and with it was demonstrated that this technique is more
reduced costs (Bhattacharjee et al. 2016). In fact, suitable for cell culture since it allowed printing
3D printing enables to produce a microfluidic more intricate systems. Finally, stereolithography
device with the desired architecture in a single step, showed to be the technique that allowed higher
i.e. by using just one equipment digital data is control over fluid flow.
converted into the desired structure. The process
can be divided into different steps: (i) design of the
device using a computer-aided drawing (CAD); (ii) 1.3.1 Fused Deposition Modelling
convert this into standard triangulation language
(STL) file; (iii) move the STL file into the printer, Fused deposition modeling is the technique
where it is divided into 2D sequential cross- more extensively used. In fact, this technique
Fig. 1.2 3D printing of a microfluidic device. The differ- file into individual layers; (4) printing of microfluidic
ent steps of 3D printing: (1) production of computer-aided device in a layer-by-layer manner. (Reprinted with per-
drawing file; (2) conversion of CAD file into a standard mission from (Waheed et al. 2016) - Published by The
triangulation language (STL) file; (3) conversion of STL Royal Society of Chemistry)
relies on the extrusion, layer-by-layer, of a high 1.3.2 Inkjet Modelling

variety of thermoplastic materials through a
high temperature nozzle (Fig. 1.3a) (Waheed Inkjet modelling can produce microfluidic devices
et al. 2016; Macdonald et al. 2017). During through continuous or dropwise deposition
extrusion, the thickness of polymer thread (Fig. 1.3b) (Waheed et al. 2016; Macdonald et al.
depends on the distance from the nozzle to the 2017). During the continuous deposition, the
substrate, and on the ratio between flow rate and printing is performed using lower viscosity inks at
print speed, ranging between 0.1 and 0.3 mm. higher velocity as compared with drop deposition.
Most of the materials compatible with this Nevertheless, through drop deposition it is possible
technique are biocompatible, such as to achieve higher placement accuracy, resulting in
polycaprolactone (PCL), polylactic acid (PLA) reproducible microfluidic devices. To obtain the
and polycarbonate (PC) (Salentijn et al. 2017). printed device, jetting heads deposits the material
Additionally, fused deposition modeling support into trays, while a UV light that is coupled to the
the printing of composite materials, with jetting head, polymerize the material as it is being
different physicochemical properties, and printed (Rafał and Krzysztof 2015). To print the
extrude different materials at the same time, next layer, the tray moves and a new process
which makes it very appealing for different begins until getting the entire structure (Waheed
tissue engineering applications. Despite the et al. 2016; Macdonald et al. 2017). Additionally,
possibility to extrude an extensive variety of it allows the fast printing of intricate microfluidic
bioinks, the more viscous ones guarantee the devices with several materials, providing
integrity of the printed structures, preventing its physicochemical features accordingly with the
collapse (Yi et al. 2017). Moreover, this final approach. Nevertheless, this technique
technique allows the placement of other presents some limitations, including the removal
materials inside of the printing structure in the of the support material and the amount of waste
middle of the run (Salentijn et al. 2017). produced during the printing process. Also,
Nevertheless, fused deposition-modeling amongst the low number of materials used to print
present some restrictions related with the rough using this technique (≈ 100), it is still necessary to
surface obtained, and with the accuracy of develop materials with biocompatible features for
dimensions. For so, the resulting microfluidic biomedical devices applications.
devices present fusion lines and blank spaces
between the printed layers, which affects the
resolution obtained (Waheed et al. 2016; 1.3.3 Stereolithography
Macdonald et al. 2017). Finally, the velocity of
printing is slower than inkjet modelling and Stereolithography is a technique that allows to print
stereolithography. a curable material layer-by-layer using two
6 F. R. Maia et al.
Fig. 1.3 3D-printing strategies and images of microflu- the material is polymerized; (c) Stereolithography, where
idic devices printed. (a) Fused deposition modelling, a photopolymerizable promoter is projected onto a plat-
where thermoplastics are extruded through a heated noz- form immersed in a photopolymer. Scale bars = 1 cm.
zle, in an XY manner until complete the layer; (b) Inkjet (Adapted with permission from (Macdonald et al. 2017).
modelling, where the material is extruded through nozzles Copyright © 2017 American Chemical Society)
in a continuous or dropwise manner. Between each layer,
different approaches, free surface approach and process in both approaches just differs in the
constrained surface approach (Fig. 1.3c) (Waheed configuration of printing, being the first

et al. 2016; Macdonald et al. 2017). The printing configuration printed on top of a platform, while
the second configuration printed hanging from a Additionally, 3D printing allows the easy pro-
platform. The materials used are photopolymerized duction of constructs with different sizes, geom-
resins, as epoxides, which after the first layer being etries and porosities. With this in mind, Cox et al.
completed printed needs to be photopolymerized (2015) investigated the production of a 3D
by laser or UV light. Then, until the structure is printed scaffold composed of hydroxyapatite and
complete, the platform moves, a new layer is poly(vinyl)alcohol (PVA) through fused deposi-
printed and traced by the laser to be polymerized. tion modeling for bone tissue engineering appli-
This technique only needs a small amount of cations. In a first stage, researchers evaluated the
material which allow reducing the associated costs. precursor flowability and observed that this is a
Since there are two main variables along the very important feature during printing, since it
process, photopolymerizable promoter and resin influences the mechanical strength, microstruc-
features, the resolution achieved also varies with ture and porosity. The obtained scaffold exhibited
them (Waheed et al. 2016; Macdonald et al. 2017). a porosity around 55%, with pore channel size of
One of the main challenges resides in the clearance around 1460–1750 μm and micropores size
of uncured resin and the fact that it is only possible around 10–60 μm and rough topography, which
to print one material at a time. are very promising features for bone tissue engi-
neering applications. In a different approach, He
et al. (2015) produced a scaffold through
1.4 D – Bioprinting Based Bone
3 extrusion- based rapid prototyping technique
Tissue Models using gels, which is similar to fused deposition
modeling. For that, they used 15% PVA/calcined
3D bioprinting strategies have been pursued in goat spongy bone mixed with a biphasic ceramic
last few years as one of the most promising tech- powder. In this experiment, the obtained scaf-
nology to be used in tissue engineering approach folds presented a uniform pore size and porosity
(Jariwala et al. 2015). In fact, it enables the devel- around 68%. Moreover, they have shown that,
opment of highly reproducible and spatially con- when rabbit bone marrow stromal cells were cul-
trolled constructs with different materials, growth tured on top of developed scaffolds, these cells
factors and/or other molecules of interest as were able to spread and proliferate. Finally, they
reported by Faramarzi et al. (Faramarzi et al. injected in vivo via the marginal ear vein the
2018). They studied the development of a new leachate of bone scaffold and observed no acute
bioink composed of alginate and platelet-rich toxicity, pyrogenic reaction or stimulation. Using
plasma, which contains several growth factors a different strategy, Boga et al. (2018) develop a
that can improve tissue regeneration. Also, the scaffold that emulated the native hollow bone
developed bioink could be crosslinked upon through fused deposition modeling, using a mix-
reaction with native calcium and allowed the ture of tri-calcium phosphate, to mimic the inor-
controlled release of growth factors. Interestingly, ganic phase, and alginic acid, to mimic the
this bioink allowed the use of patient specific organic phase. Finally, they modified the mixture
platelet-rich plasma which ultimately would with graphene oxide to enhance the mechanical
minimize the host immune response. In a different and biological properties. Then, human
study, Cui et al. (2016) printed a vascularized osteoblasts were cultured on the designed
bone construct with immobilized bone scaffolds and the construct was evaluated in vitro.
morphogenetic protein 2 (BMP2) and vascular The results showed calcium deposition and an
endothelial growth factor (VEGF) peptides. increase of alkaline phosphatase activity after
Additionally, the printed construct presented a 21 days of culture, demonstrating its potential to
hard structure surrounded by a soft matrix to improve bone regeneration. In a different strategy
mimic native bone. The results have shown that, to mimic bone vasculature, Byambaa et al. (2017)
upon perfusion, it was possible to obtain an printed a bone-like structure with a perfusable
osteogenic construct with a vascular network. vascular lumen. After processing, the structure
8 F. R. Maia et al.
was colonized with human umbilical vein vided more anchorage points, while the calcium
endothelial cells and bone marrowderived human used for the crosslinking of the bioink provided
mesenchymal stem cells with the intent to the right mechanical properties to enable cell
promote different environments capable to migration. Although natural hydrogels are
support vasculogenesis and osteogenesis. The favored due to its similarities with extracellular
culture was maintained during 21 days showing matrix, among the most used there are still some
that the printed structure was capable to support drawbacks to overcome (Adepu et al. 2017). In
cell proliferation and viability. fact, there is a limited choice concerning the
Yet, a new 3D - printing approach emerged to number of hydrogels for bioprinting of cells and
answer the demand for effective engineered tis- compatible mechanisms for hydrogels’
sues, the bioprinting of 3D cell-laden tissue con- crosslinking. To tackle these issues Costa et al.
structs (Adepu et al. 2017; Wu et al. 2016). This (2017) studied silk fibroin bioink crosslinked
technique allows to print (e.g. through jetting and through an enzymatic reaction for 3D bioprint-
extrusion techniques), not only materials, but ing. The silk fibroin used in this study is produced
also, of cells embedded within chemically or by silk-worm Bombyx mori and has gained atten-
thermally crosslinkable hydrogels (bioink) (Ho tion due to its features, such as mechanical prop-
et al. 2015). In fact, it allows controlling cellular erties, biocompatibility, degradability,
distribution within the developed matrices. In the water-based processing, and accessibility of
work of Wang et al. (2016), it is reported the chemical groups for functional modifications.
development of a 3D hybrid construct composed The results showed that human adiposederived
of alginate, gelatin and adiposederived stem stem cells cultured on top of the developed matri-
cells. Upon the bioprinting process, researchers ces were viable and were able to proliferate.
have shown that the encapsulated cells remained Nevertheless, it is difficult to guarantee a specific
viable and proliferative. Additionally, osteogenic cellular distribution the developed matrices.
differentiation assays were also conducted in In the case of bone, it is easy to understand
vitro, showing cell differentiation along the the advantages of using bioprinting technique.
osteogenic lineage, and subcutaneously in vivo, In fact, constructs for bone tissue approaches
showing ectopic bone formation. In brief, this needs to be compliant enough to allow, e.g. cell
study demonstrated the efficacy of using 3D bio- proliferation, migration and formation of blood
printing constructs in bone tissue engineering. vessels, but also strong enough to support the
Even so, despite the excitement about the pos- mechanical stress that bone tissue is subjected
sibility to print complex structures with different to. Additionally, as described previously,
cell types and mechanical properties (Fedorovich several types of cells should be taken into
et al. 2011; Arrigoni et al. 2017; Wang et al. consideration. Bioprinting emerged as a
2014), there is a clear necessity to optimize some technique that allowed including several of the
conditions as the development of adequate desired features in bone constructs production
hydrogel type (bioink) and needle diameter to in a simple and controlled way. As an example,
maintain cell viability (Fedorovich et al. 2008). Cunniffe et al. (2017) proposed 3D bioprinted
Bendtsen and Wei (2017) studied the optimization bone tissue for the development of personalized
of a bioink composed of alginate- polyvinyl gene activated implant to treat complex bone
alcohol-hydroxyapatite with collagen to promote defects. They developed a bioink composed of
cell proliferation for bone tissue engineering alginate with RGD and nano-hydroxyapatite
applications. For that they printed cells with plasmid DNA to obtain an engineered
encapsulated within different bioink formula- non-viral gene activated matrix. Bone marrow-
tions. The results showed that the amount of col- derived mesenchymal stem cells were added to
lagen and level of crosslinking influenced cell this bioink and printed simultaneously with a
proliferation. In this reasoning, the 3D printed polycaprolactone supporting mesh. The authors
constructs with higher amount of collagen pro- were able to detect a sustained expression of
Fig. 1.4 3D bioprinted mandible bone. (a) CAD model bone; (e) Alizarin Red S staining showing mineral
produced from CT image data; (b) visualization of deposits characteristic of osteogenic differentiation.
dispensing paths along with the 3D structure; (c) picture (Reprinted with permission from (Kang et al. 2016)
of 3D printing process; (d) picture of 3D printed mandible Copyright © 2016, Springer Nature)
bone morphogenic protein and transforming strategy. In a different example, Kang et al. (2016)
growth factor genes during 14 days, resulting in printed cell-laden hydrogels together with
matrix deposition and mineralization. Upon biodegradable polymers combine with tricalcium
subcutaneous implantation, they observed phosphate and sacrificial hydrogels to emulate the
higher vascularization and mineralization as shape of a human mandible bone (Fig. 1.4). For
compared with constructs without cells. this, they used computed tomography (CT) images
Overall, this gene-activated bioink can be the of the bone and transformed it into a computer
first step to a totally new generation of bioinks. model (Fig. 1.4 a, b). Finally, this created model
Recently, bioprinting caught the attention of was printed, dispensing cells and materials into de
scientists to the possibility to produce patient- desired shape (Fig. 1.4 c, d). Additionally, they
specific implants emulating the complex tissues’ included microchannels along the construct to
shape and gradient composition (Costa et al. 2018; facilitate nutrients diffusion. After 24 hours of
Datta et al. 2017). One example is the work of Li culture, they observed around 90% of viability and
et al (2017). To study a new approach for treating after 28 days they were able to detect mineral
bone and cartilage defects, researchers explored deposits (Fig. 1.4 e). Although they did not observe
the use of 3D scanning followed by 3D printing. the efficacy of this construct once implanted, this
For that, they created different defects in animals was a step forward to the development of patient-
(rabbit and mini pig), produced digital models of specific implants. Despite the advances achieved
the defects by high-resolution scanning and in engineering bone tissues, most of the strategies
accurately printed using photopolymerized do not comprise vascularization, which has
hydrogels, demonstrating the feasibility of this hampered its successful maturation and
10 F. R. Maia et al.
implantation. Yet, the produced tissues can be used In a different approach, it was created a
as more reliable in vitro 3D models. method to produce a segment of a nephron sur-
rounded by an artificial matrix, as shown in
Fig. 1.5 (Homan et al. 2016). For this, they
1.5 3D – Bioprinted Microfluidic printed a bioink with gel-to-fluid transition fea-
in vitro Models tures, composed of Pluronic F127 mixed with
trombone, inside of a matrix composed of gelatin
Bioprinted tissues have revolutionized tissue- and fibrinogen (Fig. 1.5 b and c). After cooling
engineering field but it still has some limitations, the printed construct at 4 °C, the Pluronic F127
as the production of vascularized complex mixed with thrombin liquefied, opening a
biomimetic structures. For so, new strategies convoluted tubular channel. Then, human
have been pursued as the direct bioprinting of proximal tubular cells were perfused to
3D artificial tissues within microfluidic devices, circumscribe the convoluted tubules creating a
producing a new generation of 3D - bioprinted columnar renal tubular epithelium (Fig. 1.5 d, e
organ-on-a-chip platforms (Yi et al. 2017). These and f). With this technique, the authors were able
sophisticated platforms containing tissues with to maintain a perfusable open lumen in culture
complex microarchitectures capable of for long periods of time (≈ 2 months). This
humanlike functions, allowed proper interesting model showed to have morphological
investigation of physicochemical stimulus. features and markers similar to native cells.
Although these engineered tissues were obtained Additionally, the authors also showed that the
by bioprinting, the devices were obtained developed model could be used to study drugs’
through the traditional techniques, namely soft effect on epithelium’s permeability.
lithography. Even so, this was a huge step, since In a more complex study, Hyungseok Lee and
it allowed to obtain bioprinted tissues adapted to Dong-Woo Cho (2016) were able to print not
the specific designs of microfluidic devices. only cells but also the microfluidic device in a
Although no examples of bioprinted bone tissue- single procedure. In fact, this interesting work
on-a-chip were reported, there are some works describes the bioprinting of different cell types
worth to mention. One example is the work of and materials in a controlled manner to obtain a
Bhise et al. (2016), where the authors printed a bioprinted liver-on-a-chip. For that, they printed
construct composed of hepatic spheroids inside of an empty cavity, human hepatocellular
encapsulated inside of a hydrogel inside of a carcinoma cells and human umbilical vein
microfluidic device for continuous perfusion. endothelial cells entrapped within a matrix of
Using this model, they were able to culture collagen type 1 and gelatin. Then, they produced
spheroids for 1 month and study the effect of the walls of the channel in a layer-by-layer
acute exposure to acetaminophen, an analgesic fashion, the top wall and sealed it to avoid
and antipyretic drug. In fact, the model allowed leakage. The last step was the production of a
the prediction of drug toxicity emulating the in connection tube. Overall, the developed liver-on-
vivo environment. a-chip platform showed an improved liver
Fig. 1.5 (continued) by an artificial matrix; (c) images of the procedure for obtaining a segment of a nephron surrounded
by an artificial matrix; (d) image of developed convoluted proximal tubule (scale bar = 500 μm); (e) higher magnification
of area denoted by the white rectangle in (d) (scale bar = 200 μm); (f) image of cells circumscribing developed convoluted
proximal tubule after perfusion (tubulin staining showing primary cilia, scale bar = 50 μm); (g) picture of a columnar
epithelium; (h) human proximal tubular cells after 6 weeks of culture showing a basement membrane composed of
laminin and collagen IV (scale bar = 10 μm); (i, l) brightfield images of developed convoluted proximal tubule perfused
with different concentrations of Cyclosporine A (Cys A) (scale bars = 200 μm); (j, m) Images of actin and nuclei staining
of developed convoluted proximal tubule perfused with different concentrations of Cys A (scale bars = 200 μm); (k, n)
high magnification images of (j, m) respectively (scale bars = 200 μm). (Adapted with permission from (Homan et al.
2016) Copyright © 2016, Springer Nature)
Fig. 1.5 Development of a segment of a nephron sur- and convoluted proximal tubule; (b) scheme of the
rounded by an artificial matrix. (a) Picture of a nephron procedure for obtaining a segment of a nephron surrounded
function, translated by the secretion of albumin control of cells’ differentiation. But further
and urea, as compared with the static culture of optimization will be needed.
hepatocytes and endothelial cells. This Regardless of the progress made so far, 3D
observation was due to the presence of a flow of bioprinting by itself still struggles to produce
medium, which mimicked the native environment. tissues with blood vessels structures. For so, a
Noteworthy, the authors also compared the new generation of 3D - bioprinted organ-on-a-
culture of hepatocytes under static conditions chip platforms have emerged as a promising
with the co-culture of hepatocytes with approach. These complex structures are capable
endothelial cells, and showed that the presence of of human like functions, allowing a proper
endothelial cells promoted liver functions. investigation of physicochemical stimulus. In
fact, it allowed to print not only the tissue but also
the microfluidic device in one single step. In this
1.6 Conclusions reasoning the result organ-on-a-chip promise to
emulate the native tissue as well as the
Bone tissue engineering field has evolved greatly vascularization normally present.
along the years. Even so the development of an in Overall, despite the difficulties associated
vitro model that better mimic the native tissue is with 3D printed structures that still need to be
still a challenge. Nevertheless, it is clear that it overcome, it has been proven to be a valuable
would be an important step to the translation of tool to accomplish a new generation of 3D bio-
new drugs into the clinics. The main problem is printed bone-on-a-chip technology platform.
the complexity presented along with this tissue.
Not only due to the cells that compose it, but also Acknowledgments The authors thank the funds obtained
due to the different mechanical properties along through the FROnTHERA (NORTE-01-0145-
FEDER-0000232) project supported by Norte Portugal
the bone tissue and the presence of an intricate Regional Operational Program (NORTE 2020), under the
vasculature network. Given these challenges, PORTUGAL 2020 Partnership Agreement, through the
new techniques have been pursued. In one hand, European Regional Development Fund (ERDF). FRM
the combination of 3D tissue engineering acknowledges Portuguese Foundation for Science and
Technology (FCT) for her post-doc grant (SFRH/
approaches with microfluidic devices allowed to BPD/117492/2016), JMO thanks FCT for the distinction
mimic the vasculature naturally present in native attributed under the Investigator FCT program
tissues. On the other hand, the use of 3D printing (IF/01285/2015) and VMC acknowledges Investigator
has allowed developing spatially controlled FCT program (IF/01214/2014).
structures composed of different materials and/or
growth factors and/or cells. In fact, 3D printing
allows the fabrication of structures in a more References
reproducible manner and within small amounts
Adepu S, Dhiman N, Laha A, Sharma CS, Ramakrishna S,
of time and lower costs. Furthermore, it is Khandelwal M (2017) Three-dimensional bioprinting
important to mention that 3D printing has also for bone tissue regeneration. Curr Opin Biomed Eng
promoted a new generation of patient-specific 2:22–28. https://doi.org/10.1016/j.cobme.2017.03.005
implants, boosting personalized medicine. Still, Arrigoni C, Gilardi M, Bersini S, Candrian C, Moretti
M (2017) Bioprinting and organ-on-Chip applica-
one of the greatest challenges is the development tions towards personalized medicine for bone dis-
of a bioink for the printing of such structures with eases. Stem Cell Rev 13(3):407–417. https://doi.
properties similar to native bone tissue, org/10.1007/s12015-017-9741-5
maintaining cell viability and promoting cell Aziz AUR, Geng C, Fu M, Yu X, Qin K, Liu B (2017)
The role of microfluidics for organ on Chip simula-
proliferation and differentiation. Signs of pro- tions. Bioengineering 4(2):39. https://doi.org/10.3390/
gresses have been made to achieve a proper bio- bioengineering4020039
ink, such as the reinforcement of mechanical Bendtsen ST, Wei M (2017) In vitro evaluation of 3D bio-
properties of the bioink, the crosslinking used or printed tri-polymer network scaffolds for bone tissue
even the modification with plasmid DNA for
regeneration. J Biomed Mater Res A 105(12):3262– vated bioinks for bone tissue engineering. Tissue Eng
3272. https://doi.org/10.1002/jbm.a.36184 Part A 23(17–18):891–900. https://doi.org/10.1089/
Bhattacharjee N, Urrios A, Kang S, Folch A (2016) The ten.tea.2016.0498
upcoming 3D-printing revolution in microfluidics. Datta P, Ozbolat V, Ayan B, Dhawan A, Ozbolat IT (2017)
Lab Chip 16(10):1720–1742. https://doi.org/10.1039/ Bone tissue bioprinting for craniofacial reconstruc-
c6lc00163g tion. Biotechnol Bioeng 114(11):2424–2431. https://
Bhise NS, Manoharan V, Massa S, Tamayol A, Ghaderi doi.org/10.1002/bit.26349
M, Miscuglio M, Lang Q, Shrike Zhang Y, Shin Duval K, Grover H, Han L-H, Mou Y, Pegoraro AF,
SR, Calzone G, Annabi N, Shupe TD, Bishop CE, Fredberg J, Chen Z (2017) Modeling physiological
Atala A, Dokmeci MR, Khademhosseini A (2016) events in 2D vs. 3D cell culture. Physiology 32(4):266–
A liver-on-a-chip platform with bioprinted hepatic 277. https://doi.org/10.1152/physiol.00036.2016
spheroids. Biofabrication 8(1):014101. https://doi. Faramarzi N, Yazdi IK, Nabavinia M, Gemma A, Fanelli
org/10.1088/1758-5090/8/1/014101 A, Caizzone A, Ptaszek LM, Sinha I, Khademhosseini
Boga JC, Miguel SP, de Melo-Diogo D, Mendonca AG, A, Ruskin JN, Tamayol A (2018) Patient-specific bio-
Louro RO, Correia IJ (2018) In vitro characterization inks for 3D bioprinting of tissue engineering Scaffolds.
of 3D printed scaffolds aimed at bone tissue regen- e1701347. https://doi.org/10.1002/adhm.201701347
eration. Colloids Surf B Biointerfaces 165:207–218. Fedorovich NE, De Wijn JR, Verbout AJ, Alblas J, Dhert
https://doi.org/10.1016/j.colsurfb.2018.02.038 WJ (2008) Three-dimensional fiber deposition of
Bouet G, Marchat D, Cruel M, Malaval L, Vico L (2015) In cell-laden, viable, patterned constructs for bone tis-
vitro three-dimensional bone tissue models: from cells sue printing. Tissue Eng A 14(1):127–133. https://doi.
to controlled and dynamic environment. Tissue Eng org/10.1089/ten.a.2007.0158
Part B Rev 21(1):133–156. https://doi.org/10.1089/ Fedorovich NE, Wijnberg HM, Dhert WJ, Alblas J (2011)
ten.TEB.2013.0682 Distinct tissue formation by heterogeneous printing
Byambaa B, Annabi N, Yue K, Trujillo-de Santiago G, of osteo- and endothelial progenitor cells. Tissue Eng
Alvarez MM, Jia W, Kazemzadeh-Narbat M, Shin A 17(15–16):2113–2121. https://doi.org/10.1089/ten.
SR, Tamayol A, Khademhosseini A (2017) Bioprinted TEA.2011.0019
osteogenic and vasculogenic patterns for engineering Florencio-Silva R, GRDS S, Sasso-Cerri E, Simoes MJ,
3D bone tissue. Adv Healthcare Mat 6(16):1700015. Cerri PS (2015) Biology of bone tissue: structure,
https://doi.org/10.1002/adhm.201700015 function, and factors that influence bone cells. Biomed
Carvalho MR, Reis RL, Oliveira JM (2018) Mimicking the Res Int 17:1–17. https://doi.org/10.1155/2015/421746
3D biology of osteochondral tissue with microfluidic- Hao S, Ha L, Cheng G, Wan Y, Xia Y, Sosnoski DM,
based solutions: breakthroughs towards boosting drug Mastro AM, Zheng SY (2018) A spontaneous 3D
testing and discovery. Drug Discov Today 23:711– bone-on-a-chip for bone metastasis study of breast
718. https://doi.org/10.1016/j.drudis.2018.01.008 cancer cells. Small 14(12):1702787. https://doi.
Costa JB, Silva-Correia J, Oliveira JM, Reis RL (2017) org/10.1002/smll.201702787
Fast setting silk fibroin bioink for bioprinting of He HY, Zhang JY, Mi X, Hu Y, Gu XY (2015) Rapid pro-
patient-specific memory-shape implants. Adv totyping for tissue-engineered bone scaffold by 3D
Healthcare Mat 6(22):n/a-n/a. https://doi.org/10.1002/ printing and biocompatibility study. Int J Clin Exp
adhm.201701021 Med 8(7):11777–11785
Costa JB, Silva-Correia J, Ribeiro VP, da Silva Morais Ho CM, Ng SH, Li KH, Yoon YJ (2015) 3D printed
A, Oliveira JM, Reis RL (2018) Engineering patient- microfluidics for biological applications. Lab
specific bioprinted constructs for treatment of degen- Chip 15(18):3627–3637. https://doi.org/10.1039/
erated intervertebral disc. Mat Today Commun c5lc00685f
doi:https://doi.org/10.1016/j.mtcomm.2018.01.011, Homan KA, Kolesky DB, Skylar-Scott MA, Herrmann J,
506, 512, 19 Obuobi H, Moisan A, Lewis JA (2016) Bioprinting of
Cox SC, Thornby JA, Gibbons GJ, Williams MA, Mallick 3D convoluted renal proximal tubules on Perfusable
KK (2015) 3D printing of porous hydroxyapatite scaf- chips. Sci Rep 6:34845. https://doi.org/10.1038/
folds intended for use in bone tissue engineering appli- srep34845
cations. Mater Sci Eng C Mater Biol Appl 47:237–247. Jariwala SH, Lewis GS, Bushman ZJ, Adair JH, Donahue
https://doi.org/10.1016/j.msec.2014.11.024 HJ (2015) 3D printing of personalized artificial bone
Cui H, Zhu W, Nowicki M, Zhou X, Khademhosseini scaffolds. 3d Print Additive Manuf 2(2):56–64. https://
A, Zhang LG (2016) Hierarchical fabrication of doi.org/10.1089/3dp.2015.0001
engineered vascularized bone biphasic constructs Jusoh N, Oh S, Kim S, Kim J, Jeon NL (2015)
via dual 3D bioprinting: integrating regional bioac- Microfluidic vascularized bone tissue model with
tive factors into architectural design. Adv Healthc hydroxyapatite-incorporated extracellular matrix.
Mater 5(17):2174–2181. https://doi.org/10.1002/ Lab Chip 15(20):3984–3988. https://doi.org/10.1039/
adhm.201600505 C5LC00698H
Cunniffe GM, Gonzalez-Fernandez T, Daly A, Sathy BN, Kang H-W, Lee SJ, Ko IK, Kengla C, Yoo JJ, Atala A
Jeon O, Alsberg E, Kelly DJ (2017) Three-dimensional (2016) A 3D bioprinting system to produce human-
bioprinting of Polycaprolactone reinforced gene acti- scale tissue constructs with structural integrity. Nat
Biotechnol 34:312–319. https://doi.org/10.1038/ osteoblasts during adulthood. Stem Cells (Dayton,

nbt.3413. https://www.nature.com/articles/ Ohio) 34(12):2930–2942. https://doi.org/10.1002/
nbt.3413#supplementary-information stem.2474
Lee H, Cho D-W (2016) One-step fabrication of an organ- Rafał W, Krzysztof A (2015) Inkjet 3D printing of micro-
on-a-chip with spatial heterogeneity using a 3D bio- fluidic structures—on the selection of the printer
printing technology. Lab Chip 16(14):2618–2625. towards printing your own microfluidic chips. J
https://doi.org/10.1039/C6LC00450D Micromech Microeng 25(8):085013
Lee JH, Gu Y, Wang H, Lee WY (2012) Microfluidic Salentijn GIJ, Oomen PE, Grajewski M, Verpoorte E
3D bone tissue model for high-throughput evalua- (2017) Fused deposition modeling 3D printing for
tion of wound-healing and infection-preventing bio- (bio)analytical device fabrication: procedures, materi-
materials. Biomaterials 33(4):999–1006. https://doi. als, and applications. Anal Chem 89(13):7053–7061.
org/10.1016/j.biomaterials.2011.10.036 https://doi.org/10.1021/acs.analchem.7b00828
Li L, Yu F, Shi J, Shen S, Teng H, Yang J, Wang X, Jiang Veerla SC, Kim DR, Yang SY (2018) Fabrication
Q (2017) In situ repair of bone and cartilage defects of a microfluidic device for studying the in situ
using 3D scanning and 3D printing. Sci Rep 7(1):9416. drug-loading/release behavior of graphene oxide-
https://doi.org/10.1038/s41598-017-10060-3 encapsulated hydrogel beads. Biomat Res 22(1):7.
Lima MJ, Correlo VM, Reis RL (2014a) Micro/nano rep- https://doi.org/10.1186/s40824-018-0119-9
lication and 3D assembling techniques for scaffold Waheed S, Cabot JM, Macdonald NP, Lewis T, Guijt
fabrication. Mater Sci Eng C 42:615–621. https://doi. RM, Paull B, Breadmore MC (2016) 3D printed
org/10.1016/j.msec.2014.05.064 microfluidic devices: enablers and barriers. Lab
Lima MJ, Pirraco RP, Sousa RA, Neves NM, Marques Chip 16(11):1993–2013. https://doi.org/10.1039/
AP, Bhattacharya M, Correlo VM, Reis RL (2014b) C6LC00284F
Bottom-up approach to construct microfabricated Wang X, Tolba E, Schröder HC, Neufurth M, Feng Q,
multi-layer scaffolds for bone tissue engineering. Diehl-Seifert B, Müller WEG (2014) Effect of bio-
Biomed Microdevices 16(1):69–78. https://doi. glass on growth and biomineralization of SaOS-2
org/10.1007/s10544-013-9806-4 cells in hydrogel after 3D cell bioprinting. PLoS
Macdonald NP, Cabot JM, Smejkal P, Guijt RM, Paull One 9(11):e112497. https://doi.org/10.1371/journal.
B, Breadmore MC (2017) Comparing microfluidic pone.0112497
performance of three-dimensional (3D) printing Wang XF, Song Y, Liu YS, Sun YC, Wang YG, Wang Y,
platforms. Anal Chem 89(7):3858–3866. https://doi. Lyu PJ (2016) Osteogenic differentiation of three-
org/10.1021/acs.analchem.7b00136 dimensional bioprinted constructs consisting of
Maia FR, Musson DS, Naot D, da Silva LP, Bastos AR, human adipose-derived stem cells in vitro and in vivo.
Costa JB, Oliveira JM, Correlo VM, Reis RL, Cornish PLoS One 11(6):e0157214. https://doi.org/10.1371/
J (2018) Differentiation of osteoclast precursors on journal.pone.0157214
gellan gum-based spongy-like hydrogels for bone Wu Z, Su X, Xu Y, Kong B, Sun W, Mi S (2016) Bioprinting
tissue engineering. Biomed Mat (Bristol, England) three-dimensional cell-laden tissue constructs with
13(3):035012. https://doi.org/10.1088/1748-605X/ controllable degradation. Sci Rep 6:24474. https://doi.
aaaf29 org/10.1038/srep24474
Matic I, Matthews BG, Wang X, Dyment NA, Worthley Yi H-G, Lee H, Cho D-W (2017) 3D printing of organs-
DL, Rowe DW, Grcevic D, Kalajzic I (2016) on-chips. Bioengineering 4(1):10. https://doi.
Quiescent bone lining cells are a major source of org/10.3390/bioengineering4010010
Microfluidics for Processing
of Biomaterials 2
Luca Gasperini, Alexandra P. Marques,
and Rui L. Reis
Abstract silk fibroin are emphasized deu to their rele-

Microfluidics techniques can be used to pro- vance in the field. The critical characteristics
cess a wide range of biomaterials, from syn- of these materials are discussed, giving par-
thetic to natural origin ones. This chapter ticular consideration to those that directly
describes microfluidic processing of biomate- impact its processability using microfluidics.
rials, mainly polymeric materials of natural Furthermore, some microfluidic-based pro-
origin, focusing on water-soluble polymers cessing techniques are presented, describing
that form non-flowing phases after crosslink- their suitability to process materials with dif-
ing. Some polysaccharides and proteins, ferent sol-gel transition mechanisms.
including agarose, alginate, chitosan, gellan
gum, hyaluronic acid, collagen, gelatin, and Keywords
Microfluidics · Microfabrication · Hydrogels
L. Gasperini (*)
3B’s Research Group, I3Bs – Research Institute on
Biomaterials, Biodegradables and Biomimetics,
University of Minho, Headquarters of the European Highlights
Regenerative, Braga, Portugal • Microfluidics techniques allow a high level of
ICVS/3B’s – PT Government Associate Laboratory, control on fluid flow and can be employed to
Braga/Guimarães, Portugal process materials for microfabrication.
e-mail: luca.gasperini@i3bs.uminho.pt • Based on the nature of flowing the material,
A. P. Marques · R. L. Reis different microfluidics set-up of a wide range
3B’s Research Group, I3Bs – Research Institute on of complexity can be used.
Regenerative, Braga, Portugal
ICVS/3B’s – PT Government Associate Laboratory,
Braga/Guimarães, Portugal
Precision Medicine, Headquarters at University of
Minho, Guimarães, Portugal

https://doi.org/10.1007/978-3-030-36588-2_2
16 L. Gasperini et al.
2.1 Biomaterials ing, the formation of the gel is reversible, but in

and Crosslinking Strategies the case of agarose it presents a marked thermal
hysteresis (Indovina et al. 1979) meaning that
2.1.1 Thermally Cross-Linked after the formation of the gel the liquefaction
Biomaterials temperature is much higher than the UCST. The
thermal hysteresis allows decreasing the temper-
Some polymeric solutions form a gel upon tem- ature below UCST to form gels that remain stable
perature variations. In these cases, the solvent if reheated above the same temperature. For
and solute are miscible within a specific tempera- example, some formulations of agarose can be
ture range and up to a critical temperature at used to encapsulate cells because they form gels
which they form a gel. Based on the nature of the between 25 and 35 °C and the gels remain stables
polymer and its sol-gel mechanism, the gel is when heated up to the physiological temperature.
formed either by increasing or decreasing the Gelatin is another common natural-origin bioma-
temperature (Jeong et al. 2002). Some polymeric terial that forms gel upon cooling. Gelatin is a
solutions present a critical temperature below protein that results from breaking the triple helix
which the polymer and the solution are mixable, of collagen into single strain molecules after acid
lower critical solution temperature (LCST); the treatment or hydrolysis (Kuijpers et al. 1999).
gel is formed by increasing the temperature Gelatin solutions present a critical temperature
above LCST. On the other hand, the upper critical that depends on its structure and concentration.
solution temperature (UCST) indicates the tem- In some commonly used formulation this critical
perature above which polymer and water are mis- temperature is very close to physiological tem-
cible, and when the polymeric solution is cooled perature. While gelatin, forms gel upon cooling,
below the UCST the polymer becomes insoluble collagen solutions typically form gels upon heat-
and more hydrophobic, forming a gel. This is an ing. Commercial formulations of collagen are
enthalpy-driven transition where the polymer- commonly found dissolved or soluble in diluted
water interactions become unfavorable due to acid (e.g. 0.1% acetic acid). To form a gel, the
temperature variations. By changing the tempera- solution is neutralized with a strong base (usually
ture, the polymer rearranges its conformation to NaOH) in a concentrated buffer. When heated to
minimize the surface in-contact with water. Many physiological temperature collagen undergoes
polymeric solutions of natural-origin materials fibrillogenesis through an entropy-driven self-
follow this last sol-gel mechanism (Klouda and assembly process, forming collagen fibers indis-
Mikos 2008; Ward and Georgiou 2011). Agarose, tinguishable from the natural ones (Helseth and
for example, is a polysaccharide derived from red Veis 1981; Xu et al. 2009).
algae formed by alternating units of β-D-
galactopyranose and 3,6-anhydro-α-
L-galactopyranose that above UCST are 2.1.2 Polyelectrolytes and Ionically
organized in a random coil conformation. Below Cross-Linked Biomaterials
UCST the polymer chains assume a double helix
conformation where the helices aggregate and Some polymers carry a net charge along their
are stabilized by hydrogen bonds (Lahaye and backbone that makes them, in the presence of
Rochas 1991). The critical temperature depends ions or polyelectrolytes of opposite charge, to
on the molecular weight and the structure of the form insoluble complexes through those groups
polymer and can vary typically within a range of responsible for the solubility of the polymer in
20–50 °C. While its chemical composition can water(Bekturov and Bimedina 1981). The formed
change with the source, some commercial formu- gels can be dissolved in the presence of chelating
lations of agarose are modified to reduce the agents that break down the complex. Some com-
number of hydrogen bonds, decreasing its critical monly used polymers are negatively charged due
temperature. As other thermally driven crosslink- to the presence of carboxyl or sulfate groups
2 Microfluidics for Processing of Biomaterials 17
while positively charged polymers carry amino suspended in the polymeric solution and kept
groups that become protonated in water at low for extended periods of time, as for example
pH (mainly chitosan and polylysine) (Mano et al. prior bioprinting or microfluidics, leading to
2007). osmotic shock (Carvalho et al. 2017).
Alginate is a widely used, negatively charged, While Alginate and gellan gum are negatively
polysaccharide derived from brown seaweed. It charged, chitosan is one of the few natural posi-
was one of the first and still one of the most used tively charged polysaccharides that is being used
polymers for cell encapsulation (de Vos et al. as a biomaterial. Chitosan derives from the de-
2006; Goh et al. 2012). Alginate forms gels in the acetylation of chitin, a structural component of
presence of multivalent cations that bind adjacent the exoskeleton of some species such as crabs
chains creating a 3D hydrophilic network. and shrimps (Marguerite Rinaudo 2006).
Alginate is a linear block copolymer made of Chitosan contains an amine group that, upon pro-
mannuronic (M) and glucuronic (G) residues tonation in acidic solutions, renders it soluble and
being the latter the preferable binding sites for positively charged (Rinaudo et al. 1999). At
the cations. While the sol-gel transition is quick, physiological pH and in the presence of phos-
the strength of the gel that is formed depends on phate salts, chitosan is soluble and thermorespon-
the structure of the polymer, the relative ratio of sive, forming a gel when heated up to 37 °C
G and M blocks and the cations used (Drury et al. (Couto et al. 2009).
2004; Mancini et al. 1999).
Gellan gum, similarly to alginate, is also a
negatively charged polysaccharide that can form 2.1.3 Chemically Cross-Linked
hydrogels in the presence of cations but through Biomaterials
a different sol-gel mechanism due to the poly-
mer thermos-sensitive character. The polymeric The family of chemically cross-linked biomateri-
chains of gellan gum are in random coil confor- als is composed of those materials that form
mation in solution at high temperature, and hydrogels thanks to covalent bonds formed
upon cooling they rearrange to double-helices between adjacent polymeric chains after a chemi-
that, by further decreasing the temperature, self- cal reaction. Because of this, chemical crosslink-
assemble in tight structures (Grasdalen and ing is irreversible, in opposition to the ionic and
Smidsrød 1987; Oliveira et al. 2010). These thermal one that involve reversible mechanisms.
structures, necessary for the formation of stable Materials that form hydrogels when irradiated by
gels, are further stabilized in the presence of light are particularly interesting in the biomedical
positively charged ions that shield the carboxyl field. These crosslinking strategies require the
side groups of the polymer chain. For this rea- presence of a photoinitiator that forms radicals
son, the sol-gel transition can be obtained both when irradiated by light. The chemical nature of
by decreasing the temperature of a solution and/ the photoinitiator dictates the wavelength of the
or in the presence of ions. Solutions of gellan light necessary for the formation of those radi-
gum at relatively low concentration (1, 1.5%) do cals. The radicals then react with the functional
not form strong hydrogels at room temperature, groups of the polymeric backbone forming cova-
but they present characteristics of microstruc- lent intermolecular bonds. The attractiveness of
tured liquids (thixotropic). Differently than algi- crosslink materials just by the action of light has
nate where only multivalent cations trigger the been leading the research on polymer modifica-
transition to gel, in gellan gum any positively tion to introduce suitable functional groups, such
charged cation effectively shields the helices as acrylates, to allow its photocrosslinking
increasing the viscosity of the solution or form- (Ferreira et al. 2009; Smeds et al. 2001). One
ing a gel. This limits the use of buffers or any example is the modification of hyaluronic acid, a
solution containing salts to dissolve the poly- biodegradable linear anionic polysaccharide
mer. This is particularly relevant if cells are represent in the extracellular matrix of mammals,
that is soluble in water and forms highly viscous 2.2 Processability

solutions at relatively low concentrations but can-
not form stable hydrogels (Balazs 2004). The The processability of polymeric solutions for
chemical reaction can be optimized to obtain dif- microfluidics is characterized by their resistance
ferent degrees of substitution and as such to tune to flow. In microfluidics, fluids are forced to flow
the final properties of the gel in terms of porosity into microchannels and capillaries by applying
and mechanical properties. This type of reaction pressure to the reservoir where the microchannel
has been also applied to other biopolymers such is connected. Viscosity is the property of the fluid
as gelatin (Van Den Bulcke et al. 2000) and gel- that represents its resistance to flow and is related
lan gum (Coutinho et al. 2010), mentioned above. to how the molecules in the liquid phase interact.
In these cases, the intrinsic features of the poly- Generically, a highly viscous solution is defined
mers are likely to be changed after the modifica- as thick; highly viscous solutions have high resis-
tion of the polymer and thus, the original tance to flow and are characterized by high cohe-
crosslinking mechanisms might no longer apply. sive forces. On the opposite hand, low viscosity
Another chemical crosslinking method that solutions are defined as thin and possess a low
has been of particular interest is the chemical resistance to flow.
crosslinking mediated by enzymes. This method Like other properties of materials, viscosity is
is inspired by nature and uses enzymes as cata- a function of different parameters. For Newtonian
lyzers of the chemical reaction (Moreira Teixeira fluids, viscosity varies with the temperature and
et al. 2012). By controlling the activity of the viscosity tends to decrease as temperature
enzymes, gels with different properties are increases. Polymeric solutions such as the ones
formed. One of the most straightforward exam- that have been discussed in this chapter are gen-
ples is the combination of fibrinogen and throm- erally non-Newtonian. For non-Newtonian flu-
bin, naturally present in mammals and involved ids, the viscosity not only depends on temperature
in vital biological processes. Thrombin catalyzes but also on the shear rate or the shear rate history,
the cleavage of the fibrinopeptides in the middle being the shear rate the rate of deformation due to
of fibrinogen, a water-soluble glycoprotein, to shear stresses. In particular, polymeric solutions
form fibrin monomers. These monomers polym- tend to show a shear-thinning non-Newtonian
erize into half-staggered oligomers that lengthen behavior meaning that their viscosity decreases
into protofibrils, which then aggregate laterally to at higher shear rates. When polymers are forced
form fibers. The 3D structure/gel that is formed through a channel they preferentially rearrange in
by branching during polymerization is cross- the direction of the flow, minimizing the interac-
linked by calcium and the transglutaminase fac- tion between the chains and as such decreasing
tor XIIIa becoming mechanically stable. the viscosity of the solution. Thus, the viscosity
While the above gelling mechanisms are the profile for these materials correlates the shear
most common, others of different natures exist. rate to the value of viscosity. For the majority of
For example a polymer may rearrange forming the shear-thinning solutions a log-log plot of vis-
an insoluble three-dimensional structure as con- cosity vs shear rate allows identifying a linear
sequence of pH or ionic strength variations of the region corresponding to the power-law
coagulating bath (Agnello et al. 2017). Some distribution that connects two regions where the
polymers, such as silk fibroin, present complex viscosity is constant, called Newtonian plateaus.
sol-gel mechanisms influenced by many param- The plateau at low shear rates is called zero shear
eters. The water-soluble silk is metastable and viscosity region while the one at high shear rates
tends to stabilize forming insoluble β-sheets. The is called infinite shear viscosity region. Based on
sol-gel transition is influenced by mechanical the range of shear rates tested the polymeric solu-
shear stresses, concentration of the protein and tion may present all these regions or only some.
salts in solution, and pH and temperature (Kim Viscosity values are typically measured with a
et al. 2004; Matsumoto et al. 2006). rheometer by placing the polymeric solution
between plates, one fixed and the other rotating. measured viscosity depends on the deformation
A deformation of gradually increasing speed is history of the solution (Barnes and Barnes 1997).
applied, and the stress necessary to obtain that Those microstructures are usually made of weak
rotation is recorded by the machine. forces that are destroyed during deformation.
The behavior of a non-Newtonian shear thin- While the viscous forces can rebuild in time, the
ning solutions can be model using different math- viscosity will be lower if measured shortly after
ematical models, which leads to their the deformation. These materials are more chal-
identification with the name of the model lenging to process because their properties,
(Fig. 2.1). The Carreu and Cross equations can be namely viscosity, change while they are pro-
used to model fluids that present both the power- cessed. This means that the processing parame-
law regions and the Newtonian plateaus (H. A. ters should be tuned in function of processing
Barnes and Walters 1985; Rao 2007). The pres- time and adapted to the change of viscosity to
sure used in a microfluidic system and the flow- obtain consistent results. Furthermore, these
rate can be correlated using other equations such materials present an abrupt change of viscosity
as the Hagen-Poiseuille equation, typically pre- from the material at rest (off the shelf) and while
sented in the following form (Sutera and Skalak being processed. Thixotropic materials tend to
1993): present apparent yield stress, which is the mini-
8 LQ mum stress necessary to initiate the flow (Møller
P et al. 2006).
R4
There are also other properties than viscosity
Where ∆P is the difference of pressure between that influence the processability of biomaterials.
the reservoir and end of the channel, μ is the vis- For example, when working with thermorespon-
cosity, L is the length of the channel, Q is the flow sive materials is essential to define and character-
rate, and R is the radius of the channel. This ize the temperature range where the sol-gel
equation is only valid for Newtonian fluids. transition happens. Dynamic tests are performed
Other, sometimes called corrections, were devel- by applying a sinusoidal shear deformation and
oped for non-Newtonian fluids, and contain the recording the stress. This kind of measurement is
parameters obtained after the non-linear regres- performed to obtain information on the visco-
sion performed to fit the model to the viscosity elastic behavior of the sample. For an ideal elas-
data. Examples are the power-law index and the tic sample, the stress is proportional and in phase
consistency for the Rabinowitsch correction, with the deformation. For purely viscous samples
applicable to power-law fluids. the stress is proportional to the rate of deforma-
Other materials that in solution tend to form tion and stress and strain are out of phase.
microstructures, are thixotropic meaning that the Viscoelastic materials have both contributions,
Fig. 2.1 Viscosity
profiles of Gellan gum
1% and Alginate 1.5%.
The gellan gum profile
was fitted with a
power-law equation
(R2 = 0.9875) while the
alginate profile was
fitted with a Cross
equation (R2 = 0.9979)
and their behavior is described by the storage and this is one crucial feature commonly
modulus G’ and loss modulus G”, which charac- exploited to process biomaterials. Mixing two
terize respectively the elastic solid-like and vis- different fluids in microfluidic systems can be
cous liquid-like contributions. Different challenging. This is proved by the amount of lit-
experiments can be performed based on the erature that can be found related to this subject
nature of the sample. For example the sol-gel (Lee et al. 2011; Nguyen and Wu 2005). When
transition temperature of thermoresponsive poly- mixing is needed, a chip that contains a passive
meric solutions can be measured by a tempera- mixer is a simple solution to design. In this con-
ture sweep. A sinusoidal deformation in the linear figuration, the microchannels are designed to
viscoelastic region is applied at constant fre- increase the contact area between two or more
quency and by gradually changing temperature. fluids and thus favoring mixing. Differently,
G’ an G” are recorded, and a gradual increase of active mixers are miniaturized devices that apply
G’ with temperature is expected close to the gel- external forces to the fluids.
point, the point where G’ and G” cross.
2.3.1 Droplet-Based
2.3 Processing Approaches
Droplet-based microfluidics is a high throughput
In microfluidic systems used to process poly- method capable of producing highly monodis-
mers, the flow is characterized by low Reynolds persed and separated droplets of liquids in an
number, or in other terms the flow is laminar. immiscible phase. Droplet-based systems are
Reynolds number is the ratio between the inertial typically emulsion system, and as such it allows
and the viscous forces of the flow and, for the producing oil droplets in water or water-based
flow in a circular pipe, is defined as: droplets in oil. The material used for the channels
vD must be selected accordingly, preferable hydro-
Re philic materials for oil in water and hydrophobic

ones for water in oil emulsions since the continu-
Where ρ, v and μ are respectively the density, ous phase has to wet the walls of the channels and
velocity, and viscosity of the fluid, while D is the to keep the droplets separated avoiding their
diameter of the channel. The flow is considered coalescence after being formed (Jankowski et al.
laminar if the Reynolds number is under 2000. To 2013; Saeki et al. 2010). Surfactants (e.g. Tween
understand why in microfluidic systems the flow 20 and Span 80) are commonly added to lower
is always laminar it’s important to provide simple the interfacial energy of the phases facilitating
examples. Considering the case of water flowing the formation of new interfaces and stabilizing
in a channel of 500 μm of diameter with a flow the formed droplet.
rate of 20 mL/min the Reynolds number is less Droplets are formed due to a junction, typi-
than 1000. The flow rate used in this particular cally between two or more microfluidic channels
example is way above the flow rates typically contained in a chip. The geometry of a junction
used in microfluidic systems which are typically can vary, but the T and the flow-focusing are
in the microliters range, Considering that most most commonly used for the formation of drop-
polymeric solutions are more viscous than water, lets (Fig. 2.2). The T-junction is formed by join-
more realistic parameters such as lower flow rate ing two channels at a 90-degree angle. In this
and higher viscosity would result in an even type of junctions the dispersed phase coming
lower Reynolds number. Overall these consider- from a channel enters the continuous phase and is
ations ensure that in most (if not all) microfluid- deformed up to a break by the flow, forming a
ics set-up used to process polymers, the flow is droplet. Similarly, in a flow-focusing junction,
laminar. When two fluids in laminar flow get in the continuous phase flows from side channels
contact they tend to stay separate without mixing that encounter the droplet-forming solution (dis-
Fig. 2.2 Scheme
showing two
microfluidic junctions
used for the formation of
droplets. (a) T-junction
(b) Flow focus.
(Gasperini et al. 2014)
perse phase) at the junction forcing it into an ori- Importantly, abrupt expansions of the fluid dur-
fice present at the beginning of the exit channel ing this phase should be avoided because they
that focuses the flow and favors the formation of can lead to the formation of local vortexes that
droplets. In this junction the shearing forces that destroy the still liquid droplet. For photo cross-
form the droplet are symmetric. In some set-ups linkable materials the section of the microfluidic
two inlets can be used to carry two different poly- apparatus where the crosslinking takes place
meric solutions that intersect at the junction should be made of material that is transparent to
forming a droplet with two separate sections, the the particular light wavelength used. For exam-
Janus beads (Zhao et al. 2009). ple, polytetrafluoroethylene is a common choice
In both the T and flow-focusing junctions, the in many microfluidics setups because the fluori-
size of the droplet is determined by the nature of nated polymer has excellent resistance to a wide
the solutions (e.g., viscosity), the size of the range of chemicals and good rigidity. It can also
channels and the relative flow rate of the two easily be found from many suppliers in different
phases. Furthermore, there are three distinct sizes, although these pieces are typically semi-
mechanisms of formation of droplets: squeezing, transparent or opaque to light. Thus, for photo-
dripping and jetting (De Menech et al. 2008; Nie crosslinking, other fluorinated polymers, such as
et al. 2008). Depending on the relative flow rates fluorinated ethylene propylene, should be cho-
and the nature of the solutions one of these mech- sen. While maintaining excellent chemical resis-
anisms might be favored. In particular, in the tance this polymer has much better optical
dripping mode the formation of the droplet due to properties in the visible lights and close UV
the breakup of the fluid happens near the junc- range (Šiljegović et al. 2011).
tion. On the other hand, in the jetting mode the Ionically cross-linkable materials can be pro-
flow of the droplet-forming phase is squeezed cessed in the chip or out of the chip. One of the
creating a continuous thread that extends into the methods that have been followed to form a gel is
outlet channel and breaks far away from the junc- based on carriers that can be triggered in a con-
tion. This last regime leads to droplets with a trolled manner to release ions. For example, cal-
broader size range and is favored at high flow- cium carbonate can release calcium under acidic
rates (Seemann et al. 2011). conditions; calcium carbonate nanoparticles-
Depending on the nature of the material, the containing drops were crosslinked when sub-
droplets formed by microfluidics can then be jected to a continuous phase at low pH coming
crosslinked using different methods. Droplets from an inlet placed upstream the junction due to
made of photo cross-linkable and thermally the release of calcium (Tan and Takeuchi 2007).
cross-linkable materials are hardened by the Other authors have produced droplets by emulsi-
mechanisms discussed in the first section of this fying an aqueous solution of alginate in a phase
chapter and inside the channels composing the containing a dissolved cross-linking agent (Choi
microfluidic apparatus. For thermally cross- et al. 2007; Zhang et al. 2006). In this case the
linkable materials, parts of the microfluidic sys- continuous phase containing cross-linking ions is
tem can be placed in a water bath of controlled used directly in the junction during the formation
temperature guaranteeing the crosslinking before of the drop. External crosslinking can also be
the droplets reaching its end and being collected. achieved either by using crosslinking baths con-
Fig. 2.3 (a) Illustrates the low tendency of fluids to mix (green, either MA-HA:ALG or MA-CS:ALG) and the
in a microfluidic device. (b) Polyelectrolyte complexation positive counterpart (red, CHT). (Reprinted (adapted)
at the interface of the two fluids occurs through electro- with permission from (Raquel Costa-Almeida et al. 2016).
static interactions between the negatively charged solution Copyright 2016 American Chemical Society)
taining salts (Chen et al. 2013) or by collecting As mentioned above, fluids in laminar flow
the droplets in a solid substrate rich in calcium regimes have a low tendency to mix so that two
that allows the diffusion of the cations into the fluid flowing in parallel inside one channel will
droplets (Hirama et al. 2013). By changing the stay separate (Fig. 2.3). When two polyelectro-
viscosity of the crosslinking bath, during harden- lytes are used, the interaction at the interface can
ing the droplets can gain different shapes, from form a solid phase that aligns in the direction of
microgels with a dimple to spherical beads (Hu the flow forming an ordered microstructure
et al. 2012). (Costa-Almeida et al. 2015). Other geometries of
the chip allow obtaining different types of fibers.
If it is coaxial, the inner component of the fiber
2.3.2 Continuous-Flow will stay separate from the outer shell. Upon
crosslinking the separation is maintained and a
While droplet microfluidics deals with the fabri- core-shell fiber can be fabricated. Some complex
cation and manipulation of discrete droplets of geometries also exist that combine in one single
liquids, continuous microfluidics is a technique chip some of the geometries previously discussed
based on the manipulation of continuous flows forming multicomponent fibers with accurate and
(Jun et al. 2014). More straightforward or similar complex structures (Cheng et al. 2014).
microchip designs to the ones used for the forma-
tion of droplets can be used. The simpler design
comprising one single channel is used to spin a 2.4 Conclusions
single component fiber. It is a set up commonly
employed for wet-spinning fibers into a coagulat- In this chapter, different applications of microflu-
ing bath (Agnello et al. 2016). In this case the idics for processing different polymers were pre-
most commonly used materials are the ones that sented. The versatility of microfluidics techniques
crosslink quickly in the presence of the crosslink- can be exploited using different designs of the
ing agent present in the coagulating bath (e.g. chips and different set-ups to target specific
alginate spun in calcium chloride). This allows applications. Each design has its advantages, and
the formation of the gel immediately after exiting no single device or solution is universal and able
from the microchip guaranteeing the fabrication to process all materials. Overall there is a wide
of a round fiber. variety of solutions that can be adopted for pro-
Other chips may contain junctions similar to cessing polymers at the microscale.
the ones of droplet microfluidics to fabricate mul-
ticomponent fibers. In this case, instead of using Acknowledgments APM acknowledges the European
emulsions, the inherent low tendency of fluids to Research Council for Consolidator Grant Project “ECM_
INK” ERC-2016-COG-726061.
mix is commonly exploited.
References Couto DS, Hong Z, Mano JF (2009) Development of bio-

active and biodegradable chitosan-based injectable
systems containing bioactive glass nanoparticles. Acta
Agnello S, Gasperini L, Reis RL, Mano JF, Pitarresi G,
Biomater 5(1):115–123. https://doi.org/10.1016/j.
Palumbo FS, Giammona G (2016) Microfluidic pro-
actbio.2008.08.006
duction of hyaluronic acid derivative microfibers to
De Menech M, Garstecki P, Jousse F, Stone H a (2008)
control drug release. Mater Lett 182:309–313
Transition from squeezing to dripping in a microflu-
Agnello S, Gasperini L, Mano JF, Pitarresi G, Palumbo
idic T-shaped junction. J Fluid Mech 595:141–161.
FS, Reis RL, Giammona G (2017) Synthesis, mechan-
https://doi.org/10.1017/S002211200700910X
ical and thermal rheological properties of new gellan
de Vos P, Faas MM, Strand B, Calafiore R (2006)
gum derivatives. Int J Biol Macromol 98:646–653.
Alginate-based microcapsules for immunoisolation
https://doi.org/10.1016/j.ijbiomac.2017.02.029
of pancreatic islets. Biomaterials 27(32):5603–5617.
Balazs EA (2004) Viscoelastic properties of hyaluronan
https://doi.org/10.1016/j.biomaterials.2006.07.010
and its therapeutic use. In: Chemistry and biology of
Drury JL, Dennis RG, Mooney DJ (2004) The ten-
hyaluronan, vol 415. Elsevier, Amsterdam
sile properties of alginate hydrogels. Biomaterials
Barnes HHA, Barnes A (1997) Thixotropy – a review.
25(16):3187–3199. https://doi.org/10.1016/j.
Journal of Non-Newtonian Fluid Mechanics 70(97):1–
biomaterials.2003.10.002
33. https://doi.org/10.1016/S0377-0257(97)00004-9
Ferreira P, Coelho, JFJ, Almeida JF, Gil MH (2009)
Barnes HA, Walters K (1985) The yield stress myth?
Photocrosslinkable Polymers for Biomedical
Rheol Acta 24(4):323–326. https://doi.org/10.1007/
Applications. In R. Fazel (Ed.), Biomedical
BF01333960
Engineering - Frontiers and Challenges. InTech,
Bekturov EA, Bimedina LA (1981) Interpolymer com-
pp. 55–74
plexes. Adv Polymer Sci 41:99–147
Gasperini L, Mano JF, Reis RL (2014) Natural poly-
Carvalho AF, Gasperini L, Ribeiro RS, Marques AP, Reis
mers for the microencapsulation of cells. J R Soc
R l (2017) Control of osmotic pressure to improve cell
Interface 11(100):20140817–20140817. https://doi.
viability in cell-laden tissue engineering constructs. J
org/10.1098/rsif.2014.0817
Tissue Eng Regen Med 12(2):e1063–e1067. https://
Goh CH, Heng PWS, Chan LW (2012) Alginates as a use-
doi.org/10.1002/term.2432
ful natural polymer for microencapsulation and thera-
Chen W, Kim J-H, Zhang D, Lee K-H, Cangelosi GA,
peutic applications. Carbohydr Polym 88(1):1–12.
Soelberg SD, Furlong CE, Chung J-H, Shen AQ
https://doi.org/10.1016/j.carbpol.2011.11.012
(2013) Microfluidic one-step synthesis of alginate
Grasdalen H, Smidsrød O (1987) Gelation of gellan
microspheres immobilized with antibodies. J R Soc,
gum. Carbohydr Polym 7(5):371–393. Retrieved
Interface R Soc 10:20130566. https://doi.org/10.1098/
from http://www.sciencedirect.com/science/article/
rsif.2013.0566
pii/014486178790004X
Cheng Y, Zheng F, Lu J, Shang L, Xie Z, Zhao Y, Chen Y,
Helseth DL, Veis A (1981) Collagen self-assembly
Gu Z (2014) Bioinspired multicompartmental micro-
in vitro differentiating specific telopeptide-dependent
fibers from microfluidics. Adv Mater 26:5184–5190.
interactions using selective enzyme modification and
https://doi.org/10.1002/adma.201400798
the addition of free amino telopeptide. J Biol Chem
Choi C-H, Jung J-H, Rhee YW, Kim D-P, Shim S-E,
256(14):7118–7128. Retrieved from https://pdfs.
Lee C-S (2007) Generation of monodisperse alginate
semanticscholar.org/6d43/701ea790f272f10fdc437bc
microbeads and in situ encapsulation of cell in micro-
c8abfc79a38a5.pdf
fluidic device. Biomed Microdevices 9(6):855–862.
Hirama H, Kambe T, Aketagawa K, Ota T, Moriguchi
https://doi.org/10.1007/s10544-007-9098-7
H, Torii T (2013) Hyper alginate gel microbead for-
Costa-Almeida R, Gasperini L, Rodrigues MT, Babo PS,
mation by molecular diffusion at the hydrogel/drop-
Mano JF, Reis RL, Gomes ME (2015) Development
let interface. Langmuir: ACS J Surfaces Colloids
of biomimetic microengineered hydrogel fibers for
29(2):519–524. https://doi.org/10.1021/la303827u
tendon regeneration In 2015 4th TERMIS World
Hu Y, Wang Q, Wang J, Zhu J, Wang H, Yang Y (2012)
Congress
Shape controllable microgel particles prepared by
Costa-Almeida R, Gasperini L, Borges J, Babo PS,
microfluidic combining external ionic crosslinking.
Rodrigues MT, Mano JF, Reis RL, Gomes ME (2016)
Biomicrofluidics 6(2):26502–265029. https://doi.
Microengineered multicomponent hydrogel fibers:
org/10.1063/1.4720396
combining polyelectrolyte complexation and micro-
Indovina PL, Tettamanti E, Micciancio-Giammarinaro
fluidics. ACS Biomat Sci Eng 3(7):1322–1331. https://
MS, Palma MU (1979) Thermal hysteresis and
doi.org/10.1021/acsbiomaterials.6b00331
reversibility of gel-sol transition in agarose-water
Coutinho DF, Sant SV, Shin H, Oliveira JT, Gomes
systems. J Chem Phys 70(6):2841–2847. https://doi.
ME, Neves NM, Khademhosseini A, Reis RL
org/10.1063/1.437817
(2010) Modified Gellan gum hydrogels with tunable
Jankowski P, Ogończyk D, Derzsi L, Lisowski W,
physical and mechanical properties. Biomaterials
Garstecki P (2013) Hydrophilic polycarbonate chips
31(29):7494–7502. https://doi.org/10.1016/j.
for generation of oil-in-water (O/W) and water-
biomaterials.2010.06.035
in-
oil-in-water (W/O/W) emulsions. Microfluid
Nanofluid 14(5):767–774. https://doi.org/10.1007/ Nguyen N-T, Wu Z (2005) Micromixers—a review. J

s10404-012-1090-8 Micromech Microeng 15(2):R1–R16. https://doi.
Jeong B, Kim SW, Bae YH (2002) Thermosensitive org/10.1088/0960-1317/15/2/R01
sol-gel reversible hydrogels. Adv Drug Deliv Rev Nie, Z., Seo, M., Xu, S., Lewis, P. C., Mok, M., Kumacheva,
54(1):37–51. Retrieved from http://www.ncbi.nlm. E., … Stone, H. a. (2008). Emulsification in a micro-
nih.gov/pubmed/11755705 fluidic flow-focusing device: effect of the viscosities
Jun Y, Kang E, Chae S, Lee S-H (2014) Microfluidic of the liquids. Microfluid Nanofluid, 5, 585–594.
spinning of micro- and nano-scale fibers for tissue doi:https://doi.org/10.1007/s10404-008-0271-y
engineering. Lab Chip 14(13):2145–2160. https://doi. Oliveira JT, Martins L, Picciochi R, Malafaya PB, Sousa
org/10.1039/c3lc51414e RA, Neves NM, Mano JF, Reis RL (2010) Gellan
Kim U-J, Park J, Li C, Jin H-J, Valluzzi R, Kaplan DL gum: a new biomaterial for cartilage tissue engineer-
(2004) Structure and properties of silk hydrogels. ing applications. J Biomed Mater Res A 93(3):852–
Biomacromolecules 5(3):786–792. https://doi. 863. https://doi.org/10.1002/jbm.a.32574
org/10.1021/bm0345460 Rao MA (2007) Flow and functional models for rheological
Klouda L, Mikos AG (2008) Thermoresponsive hydrogels properties of fluid foods. Springer, Boston, pp 27–58.
in biomedical applications. Eur J Pharm Biopharm: https://doi.org/10.1007/978-0-387-70930-7_2
Off J Arbeitsgemeinschaft Für Pharmazeutische Rinaudo M (2006) Chitin and chitosan: properties and
Verfahrenstechnik eV 68(1):34–45. https://doi. applications. Prog Polym Sci 31(7):603–632. https://
org/10.1016/j.ejpb.2007.02.025 doi.org/10.1016/j.progpolymsci.2006.06.001
Kuijpers AJ, Engbers GHM, Feijen J, De Smedt SC, Rinaudo M, Pavlov G, Desbrie J (1999) Influence of ace-
Meyvis TKL, Demeester J, Krijgsveld J, Zaat SAJ, tic acid concentration on the solubilization of chitosan.
Dankert J (1999) Characterization of the network Polymer 40:7029–7032
structure of Carbodiimide cross-linked gelatin gels. Saeki D, Sugiura S, Kanamori T, Sato S, Ichikawa S
Macromolecules 32(10):3325–3333. https://doi. (2010) Microfluidic preparation of water-in-oil-in-
org/10.1021/ma981929v water emulsions with an ultra-thin oil phase layer.
Lahaye M, Rochas C (1991) Chemical structure and Lab Chip 10(3):357–362. https://doi.org/10.1039/
physico-chemical properties of agar. In: Juanes JA, B916318B
Santelices B, McLachlan JL (eds) International work- Seemann R, Brinkmann M, Pfohl T, Herminghaus S (2011)
shop on Gelidium. Springer, Dordrecht, pp 137–148. Droplet based microfluidics. Rep Prog Phys 75:016601.
https://doi.org/10.1007/978-94-011-3610-5 https://doi.org/10.1088/0034-4885/75/1/016601
Lee C, Chang C, Wang Y, Fu L (2011) Microfluidic mix- Šiljegović M, Kačarević-Popović ZM, Bibić N,
ing: a review. Int J Mol Sci 12:3263–3287. https://doi. Jovanović ZM, Maletić S, Stchakovsky M, Krklješ
org/10.3390/ijms12053263 AN (2011) Optical and dielectric properties of fluo-
Mancini M, Moresi M, Rancini R (1999) Mechanical rinated ethylene propylene and tetrafluoroethyl-
properties of alginate gels: empirical characteri- ene–perfluoro(alkoxy vinyl ether) copolymer films
sation. J Food Eng 39:369–378. Retrieved from modified by low energy N4+ and C4+ ion beams.
http://www.sciencedirect.com/science/article/pii/ Radiat Phys Chem 80(12):1378–1385. https://doi.
S0260877499000229 org/10.1016/J.RADPHYSCHEM.2011.08.012
Mano JF, Silva GA, Azevedo HS, Malafaya PB, Sousa Smeds KA, Grinstaff MW, Pfister-serres A, Miki
RA, Silva SS, Boesel LF, Oliveira JM, Santos TC, D, Dastgheib K, Inoue M, Hatchell DL (2001)
Marques AP, Neves NM, Reis RL (2007) Natural ori- Photocrosslinkable polysaccharides for in situ hydro-
gin biodegradable systems in tissue engineering and gel formation. J Biomed Mater Res 54(1):115–121
regenerative medicine: present status and some mov- Sutera SP, Skalak R (1993) The history of Poiseuille’s
ing trends. J R Soc Interface 4(17):999–1030. https:// law. Annu Rev Fluid Mech 25(1):1–20. https://doi.
doi.org/10.1098/rsif.2007.0220 org/10.1146/annurev.fl.25.010193.000245
Matsumoto A, Chen J, Collette AL, Kim U, Altman GH, Tan W-H, Takeuchi S (2007) Monodisperse alginate
Cebe P, Kaplan DL (2006) Mechanisms of silk fibroin hydrogel microbeads for cell encapsulation. Adv
sol-gel transitions. J Phys Chem B 110(43):21630– Mater 19(18):2696–2701. https://doi.org/10.1002/
21638. https://doi.org/10.1021/jp056350v adma.200700433
Møller PCF, Mewis J, Bonn D (2006) Yield stress and Van Den Bulcke AI, Bogdanov B, De Rooze N, Schacht
thixotropy: on the difficulty of measuring yield EH, Cornelissen M, Berghmans H (2000) Structural
stresses in practice. Soft Matter 2(4):274. https://doi. and rheological properties of Methacrylamide modi-
org/10.1039/b517840a fied gelatin hydrogels. Biomacromolecules 1(1):31–
Moreira Teixeira LS, Feijen J, van Blitterswijk CA, 38. https://doi.org/10.1021/bm990017d
Dijkstra PJ, Karperien M (2012) Enzyme-catalyzed Ward M, Georgiou TK (2011) Thermoresponsive poly-
crosslinkable hydrogels: emerging strategies for tissue mers for biomedical applications. Polymers 3(4):1215–
engineering. Biomaterials 33(5):1281–1290. https:// 1242. https://doi.org/10.3390/polym3031215
doi.org/10.1016/j.biomaterials.2011.10.067 Xu T, Molnar P, Gregory C, Das M, Boland T, Hickman
JJ (2009) Electrophysiological characterization of
embryonic hippocampal neurons cultured in a 3D
collagen hydrogel. Biomaterials 30(26):4377–4383. Zhao LB, Pan L, Zhang K, Guo SS, Liu W, Wang Y, Chen
https://doi.org/10.1016/j.biomaterials.2009.04.047 Y, Zhao XZ, Chan HLW (2009) Generation of Janus
Zhang H, Tumarkin E, Peerani R, Nie Z, Sullan RMA, alginate hydrogel particles with magnetic anisotropy
Walker GC, Kumacheva E (2006) Microfluidic pro- for cell encapsulation. Lab Chip 9(20):2981–2986.
duction of biopolymer microcapsules with controlled https://doi.org/10.1039/b907478c
morphology. J Am Chem Soc 128(37):12205–12210.
https://doi.org/10.1021/ja0635682
Organs-on-a-Chip
3
Lucie A. Low, Margaret Sutherland,
Nadya Lumelsky, Seila Selimovic,
Martha S. Lundberg, and Danilo A. Tagle
Abstract peutic efficacy in vitro; and (3) disease model-

Organs-on-chips, also known as “tissue chips” ing of human tissues to recapitulate
or microphysiological systems (MPS), are pathophysiology within specific subpopula-
bioengineered microsystems capable of recre- tions and even individuals, thereby advancing
ating aspects of human organ physiology and precision medicine efforts. This chapter will
function and are in vitro tools with multiple discuss the development and evolution of
applications in drug discovery and develop- three-dimensional organ models over the past
ment. The ability to recapitulate human and decade, and some of the opportunities offered
animal tissues in physiologically relevant by MPS technology that are not available
three-dimensional, multi-cellular environ- through current standard two-dimensional cell
ments allows applications in the drug develop- cultures, or three-dimensional organoid sys-
ment field, including; (1) use in assessing the tems. This chapter will outline future avenues
safety and toxicity testing of potential thera- of research in the MPS field, how cutting-edge
peutics during early-stage preclinical drug biotechnology advances are expanding the
development; (2) confirmation of drug/thera- applications for these systems, and discuss the
L. A. Low (*)
National Center for Advancing Translational Sciences S. Selimovic
(NCATS), National Institutes of Health, National Institute of Biomedical Imaging and
Bethesda, MD, USA Bioengineering (NIBIB), National Institutes of
e-mail: lucie.low@nih.gov Health, Bethesda, MD, USA
e-mail: Seila.selimovic@nih.gov
M. Sutherland
National Institute for Neurological Disorder and M. S. Lundberg
Stroke (NINDS), National Institutes of Health, National Heart, Lung, and Blood Institute (NHLBI),
Bethesda, MD, USA National Institutes of Health, Bethesda, MD, USA
e-mail: sutherlandm@ninds.nih.gov e-mail: lundberm@nhlbi.nih.gov
N. Lumelsky D. A. Tagle
National Institute of Dental and Craniofacial National Center for Advancing Translational Sciences
Research (NIDCR), National Institutes of Health, (NCATS), National Institutes of Health,
Bethesda, MD, USA Bethesda, MD, USA
e-mail: nadyal@nidcr.nih.gov e-mail: Danilo.tagle@nih.gov
© This is a U.S. government work and not under copyright protection in the U.S.; foreign copyright 27
protection may apply 2020
https://doi.org/10.1007/978-3-030-36588-2_3
28 L. A. Low et al.
current and future potential and challenges drug development ends up having direct effects
remaining for the field to address. on patient quality of life and disease prognoses.
Therefore, alternative approaches are desperately
Keywords needed to speed up the processes, reduce costs
and produce safer and more effective drugs.
Microphysiological systems · Worldwide efforts are underway to develop, vali-
Bioengineering · Microfluidics · Drug date and apply alternative methods and para-
development · Disease modeling digms to supplement existing platforms, and to
introduce novel assays and methodologies to help
speed, streamline and redevelop drug develop-
Highlights ment processes (see Sect. 3.2) (Balls 1995; Casey
et al. 2015).
• Organ-on-chips are three-dimensional tissue Over the last decade, significant advance-
platforms that could have utility in a wide ments have been realized in the development and
range of drug development paradigms utilization of human in vitro microphysiological
• Platforms could be used for disease modeling, systems (MPS), often termed “organs-on-chip”
drug screening and to identify patient sub- or “tissue chips”, for toxicity analysis, efficacy
groups, or those who are most likely to benefit testing and disease modeling. These microfluidic
from clinical treatments devices combine materials science and bioengi-
• Biological and technical challenges remain neering approaches with human stem cell, pri-
but validation and commercialization efforts mary and immortalized cell line-based biology to
are underway replicate human tissue responses and diseases.
3D cell and tissue models are more complex than
standard 2D cell cultures, and often incorporate
3.1 Introduction multiple heterogeneous cell types that result in
relatively more faithful recapitulation of human
It is well-known that the current processes for in vivo cellular environments by allowing cellular
discovering and developing new therapeutic communications in more appropriate contexts
drugs and treatments are hugely costly and com- (see Sect. 3.2). Additionally, the development of
plex, take many years, and have extremely high human bioengineered systems at preclinical
attrition rates, with the simplistic “chevron” phases reduces the need for animal testing, which
depiction of drug development recently being helps eliminate the limitations of testing thera-
updated to a more realistic and dynamic “map” peutics in organisms with different physiologies
(Wagner et al. 2017a, b). The failure rate of drugs (e.g. different liver metabolic enzymes (Fashe
entering clinical trials is as high as 90%, and et al. 2015, Sasahara et al. 2015)) as well as being
recent reports put the timeline and cost of a new highly responsive to the 3R principles of animal
drug entering the market as up to 15 years and testing (refinement, reduction, replacement).
over $2.6 billion from start to finish (Morgan Growth within this technology sector has been
et al. 2011; Profile Toolkit 2018). During preclin- driven by a need to increase success in new drug
ical stages, potential compounds mainly fail due development pipelines and reduce the escalating
to unacceptable toxicity levels in 2-dimensional R&D costs by offering technical alternatives
cellular and animal models (Roberts et al. 2014; (Martin et al. 2017), and is projected to experi-
Waring et al. 2015). At later clinical stages, fail- ence a 38% growth rate in commercial areas over
ures are predominantly a result of a lack of drug the next few years (Fig. 3.1).
efficacy (Arrowsmith 2011a, b). It is becoming widely acknowledged that
Because many potential therapeutics are lost MPS could serve as tools with applications across
during development or require years of clinical a broad spectrum of drug development, from pro-
trials data before approval for broader clinical viding more physiologically-relevant platforms
use, the length of time needed plus high costs of for early target validation of promising com-
3 Organs-on-a-Chip 29
Fig. 3.1 – Market analysis of projected growth of and gigantic promise”, Organs-On-Chips 2017 report,
Organs-on-chips market from 2017–2022. CAGR = com- April 2017 (http://www.yole.fr/OrgansOnChips_Market.
pound annual growth rate. Courtesy of Yole aspx#.W2B8pxpKiV4)
Développement, from “Organs-on-chips: small market
pounds; to disease modeling of both common tissues signifies recognition that they promise a
and rare diseases in order to shed light on patho- superior capacity over traditional 2D systems to
physiology and drug responses; to stratification mimic extracellular matrix (ECM) organization,
of patient subpopulations or even creating “you- tissue architecture, specific cell-cell/cell-ECM
on-a-chip” for truly personalized medicine interactions and sophisticated environmental
(Fig. 3.2). This chapter will describe the evolu- cues, including biomechanical and biophysical
tion of MPS over the last decade and compare the forces (Jackson and Lu 2016). However, 2D and
pluses and minuses of 2-dimensional (2D) and 3D systems each have their own advantages and
3-dimensional (3D) cellular model systems. It disadvantages, which should be considered
will highlight recent key innovations in MPS within the context of specific applications (Liu
platform development and finally discuss impor- et al. 2018) (Table 3.1). 2D cultures, where cells
tant opportunities and challenges that remain for are grown as monolayers on flat surfaces, have
the field. been employed for many years as ubiquitous
tools for drug screening, toxicity testing and cel-
lular and molecular analyses of disease mecha-
3.2 Comparing 2-Dimensional nisms (Guillouzo 1998). Because of their
and 3-Dimensional Tissue simplicity, ease of technical manipulation, scal-
Modeling Systems ability, and high-throughput capabilities, they
remain a widespread model. Indeed, much useful
The recent explosion of activity in the develop- information has been obtained with 2D cultures.
ment of 3D in vitro tissue systems such as organ- For example, over the years they have been
oids and organs-on-chips for a variety of human widely used for drug toxicity studies (Niu and
30 L. A. Low et al.
Fig. 3.2 A simplified schematic of the drug development dosage guidelines and toxicity profiles for ‘first-in-
process, excluding post-approval monitoring by regula- humans’ studies. Animal usage is also reduced/refined. At
tory bodies. Organs-on-chips could be useful at each stage preclinical and clinical stages, organs-on-chips can be cre-
of the drug development process. At preclinical stages, ated from healthy and diseased populations using primary
they allow screening of hit compounds in complex organ tissues or induced pluripotent stem cell (iPSC)
3-dimensional multi-cellular systems, allowing earlier sources and used to test existing and repurposed therapeu-
elimination of highly toxic compounds and validation and tics in a risk-free manner for patients. Patient populations,
optimization of lead targets more quickly. At later pre- or even individuals, can be modeled on a chip, and person-
clinical stages, they can model human diseases and con- alized therapeutics or treatment regimens tested before
firm compound efficacy. Before moving into Phase I human administration. ADME – absorption, distribution,
clinical trials, use of multiple integrated organs-on-chips metabolism, excretion. PK-PB modeling – physiologically-
allows ADME and PK-PB modeling and refinement of based pharmacokinetic modeling
Wang 2015), and for studies of the mechanisms tions with other cell types, within a highly-
and pathophysiology of many monogenic and organized ECM. Because of these limitations, the
polygenic diseases, such as inborn errors of differentiated 2D cultures of iPSCs and of pri-
hepatic metabolism (Pournasr and Duncan 2017), mary cells typically exhibit highly immature
Parkinson’s disease (Lopes et al. 2017; Singh transcriptional and metabolic profiles of fetal
Dolt et al. 2017), diabetes (Wallet et al. 2017) and rather than postnatal cells (Jiang et al. 2018).
others. Gene-editing technologies have been
employed for derivation of isogenic induced plu-
ripotent stem cell (iPSC) lines using 2D culture 3.2.2 A
dding a Third Dimension
systems, where derived cells for multiple tissues to Tissue Culture
originate from a single donor (Singh Dolt et al.
2017). Emergence of a new generation of biomaterials
with tunable properties as well as the use of
micro-fabrication, bioprinting and microfluidic
3.2.1 Limitations of 2-Dimensional technologies opens wide-ranging possibilities for
Systems adding “a third dimension” to traditional 2D cul-
tures. In this regard, organs-on-chips and organ-
Despite many useful features, the limitations of oids as well as organoid/chip hybrids are now at
2D systems in mimicking human physiology are the center of much of the activity in the field (Liu
obvious. These limitations are primarily related et al. 2018) (Table 3.1). Organoids are 3D tissue
to the loss of normal tissue architecture and bio- cultures, often derived from stem cells, which
physical forces, and the heterotypic 3D environ- self-organize into simplistic organs and tissues.
ment or niche in which cells normally reside in In MPS (AKA organs-on-chips) however, differ-
vivo that enable dynamic and reciprocal interac- ent cell types and ECMs are placed into pre-
Table 3.1 – Comparisons between 2D and 3D tissue platforms in terms of biological complexity, manufacturing and
outputs. Engineered tissues = cells + scaffolds. Adapted from Liu et al. (15)
Conventional 2D 3D systems
systems Engineered tissues Organoid Organ-on-chip
Production method Differentiated, grown on Fabricated with Embedded in Seeded in
and timing rigid flat surfaces as scaffold and casting Matrigel to engineered
monolayer; fast mold; slow self-organize; slow chambers with
perfusion; fast
Maturation Immature Improved but still Improved but still Improved but still
lacking lacking lacking
Cell morphology Unnatural, usually Size and shape Size and shape Depends on
and type monotype similar to in vivo similar to in vivo platform design
ECM Limited composition Similar to in vivo Similar to in vivo Depends on
and contact with cells platform design
Tissue architecture Absent Simple Complex, similar to Complexity depends
organ on platform design
developmental
stages
Diffusion of signal Short distances (through Concentration Interior cells may Precisely controlled
factors and cell membranes); gradients may die or lack maturity temporal and spatial
nutrients usually receiving exist, may be due to ineffective gradients
supraphysiological affected by ECM transport to interior
doses properties
Vascularization or No No No Yes
perfusion?
High throughput Present; high; absent Possibly; low; may Absent; very low; Present; very high;
feasibility; develop to may develop to microscale
controllability; macroscale macroscale
scalability
Variability; Low; high; easy High; low; difficult High; low; easy High; sometimes
reproducibility; low; difficult
difficulty of use
Possibility for Straightforward Difficult Difficult Straightforward
genome editing
Characterization Limited, but easy cell Tissue function Tissue function Real-time tissue/
and analyses retrieval analyses possible analyses possible organ function
but cell retrieval but cell retrieval analyses possible,
and phenotypic and phenotypic with easy cell
analyses can be analyses can be retrieval
hard. hard.
fabricated chambers and cultured under different with a range of biosensors, can be amenable to
conditions in complex bioreactors, with micro- high-throughput genetic sequencing, provide
engineered microfluidic channels for fluid flow context for high-content readouts, and allow real-
across the tissue to allow recapitulation of tissue time monitoring and sample retrieval. Moreover,
complexity. This design makes it possible to exert specific tissue-tissue and cell-cell interactions
fine control over tissue composition and architec- and biophysical forces such as shear stress, ten-
ture, as well as to pre-program precise tissue oxy- sion, mechanical loading, electrical forces, peri-
genation and nutrient diffusion profiles. These stalsis and breathing movements can be applied
features create opportunities for significant pro- to MPS in a highly controlled manner to promote
longation of the chips’ viability and function up cell differentiation, tissue maturation and acqui-
to several weeks or even months (Tanataweethum sition of tissue function readouts which are more
et al. 2018). Because of the sophisticated design useful than those obtained with 2D systems.
features in some systems, they can be equipped
32 L. A. Low et al.
3.2.3 O
rganoids vs. Organs-on- drives organoid formation, while taking advan-
Chips and Synergistic tage of the precise bioengineering design princi-
Engineering ples that characterize production of tissue chips.
In this regard, capitalizing on the new opportuni-
Unlike the creation of organs-on-chips driven by ties offered by sophisticated bioprinting technol-
specific features of their design and fabrication, ogies might be promising (Li et al. 2016). It is
formation of 3D organoids is primarily controlled envisioned that synergistic engineering could be
by stochastic self-organizing principles of cell- useful for designing multi-tissue composite mod-
cell and cell-ECM interactions where the cells els, such as vascularized organ buds that can be
coalesce to form masses mimicking certain fea- effectively perfused with culture medium or
tures of normal tissue organization and function blood substitutes (Takebe et al. 2015; Zhang et al.
(Lou and Leung 2018). Many investigators have 2016), or innervated tissues able to receive neural
used iPSCs to derive organoids for a variety of inputs (Workman et al. 2016).
human tissues, including gut, lung, kidney, heart
and brain (Dutta et al. 2017). During organoid
formation, as differentiation of the iPSCs pro- 3.3 The Evolution
ceeds in vitro in response to specific differentia- of Microphysiological
tion cues, the arising cell types self-assemble to Systems Funding
form 3D structures containing different cell types and Investment
in spatial configurations similar to those found in
organs during normal in vivo development. While Building on early versions of organs-on-chips,
organoids possess complex architecture reminis- MPS technology development in the U.S. has
cent of native tissues, their formation may be been accelerated through an eight-year invest-
much slower compared to that of tissue chips. ment of over $140 M from the National Institutes
Additionally, it is difficult to exert precise control of Health (NIH), the Food and Drug
over organoid formation, and they are not easily Administration (FDA), and the Defense
amenable to multiplexing and real-time monitor- Advanced Research Projects Agency (DARPA)
ing; further, sample retrieval to obtain analyzable (Sutherland et al. 2013). In Europe, a ban on the
readouts from these systems can be challenging use of experimental animals in 2012 under the
or impossible. Moreover, lack of vascularization/ Organisation for Economic Co-operation and
perfusion in these constructs can lead to ineffi- Development (OECD) guidelines stimulated the
cient oxygen and nutrient transport into interior European Union to initiate a €1.4 million,
cells resulting in a necrotic core; an important 36-month project to support alternative platform
limitation that reduces viability and shortens the development for drug testing. The core members
life of these cultures. Finally, while gene-editing of the project were InSphero AG, the
technologies can be easily combined with tissue Bioengineering Laboratory, ETH Zurich,
chip design and derivation, it is difficult to mod- AstraZeneca AB R&D, and F. Hoffmann-La
ify genes when employing organoid-based tech- Roche AG (Kanamori et al. 2018). Japanese
nology as there is little control over the pharmaceutical companies have also recognized
differentiation microenvironment. the potential of MPS technology. In July 2017,
Given that tissue chips and organoids are the Japan Agency for Medical Research and
endowed with their own set of advantages and Development (AMED) initiated a five-year
limitations, investigators in the field are now research program with a budget of approximately
attempting to improve them through so-called ¥500 million (~$4.5 M) per year with the goal of
synergistic engineering to develop new 3D in developing a chip-based MPS that will be used in
vitro models at the intersection of organoid and drug discovery processes (Kanamori et al. 2018).
tissue chip technologies (Takebe et al. 2017). A To broaden the access to this technology and
general design principle for such models is to support the use for both academic and industry
rely on a certain degree of stochasticity, which investigators, further development is required to
optimize reproducibility, robustness, and device ney (Jang et al. 2013) and blood-brain barrier
manufacturing. To enable increased adoption of (Herland et al. 2016).
these devices, additional demonstration of utility Since the publications describing the lung
and cost per chip will need to be considered. chip and demonstration of its utility, advance-
Based on advances in academic R&D, several ments in other organ-on-chip systems has
start-up companies have been entering the market expanded to include liver (Yoon No et al. 2015;
in an effort to transform organ-on-chip technol- Vernetti et al. 2016; Lee-Montiel et al. 2017;
ogy to industrially relevant applications (Zhang Soto-Gutierrez et al. 2017), vasculature (Moya
and Radisic 2017). To date, broad investment in et al. 2013; Vunjak-Novakovic et al. 2013;
organ-on-chip technology shows the magnitude Fernandez et al. 2016; Sobrino et al. 2016), mus-
of expectations for research related to the use of cle (cardiac (Agarwal et al. 2013, Wang et al.
these devices. 2014, Mathur et al. 2015) and skeletal (Madden
et al. 2015; Cheng et al. 2016)), kidney (Weber
et al. 2016), reproductive tissues including ovary,
3.3.1 Expansion and Integration uterus, cervix and fallopian tube (Laronda et al.
of Organ-on-Chip Technology 2013; Arslan et al. 2015; Zhu et al. 2016; Xiao
et al. 2017), testes, blood-brain barrier (Brown
There are many examples of the rapid develop- et al. 2015; Brown et al. 2016; Wang et al. 2017),
ment of organ-on-chip technology, but one exam- skin (Abaci et al. 2016; Schimek et al. 2018), gut
ple of the integration of chip technology with (Foulke-Abel 2016; Shah et al. 2016) and bone
mechanical engineering is epitomized by the (de Peppo et al. 2013), amongst many others (see
lung-on-chip developed by researchers at the organ-on-chip reviews (An et al. 2015, Esch et al.
Wyss Institute, which integrated relevant biome- 2015, Low and Tagle 2017a, b).
chanical forces to cultured tissues to illustrate the The power of MPS can be seen when the chips
key roles these forces play in faithfully modeling are linked together to form integrated systems.
human tissues (Huh et al. 2010). This lung chip Indeed, shared goals of the U.S. government
mimics the alveolar-capillary border by using a agencies’ MPS programs focused on organ-on-
dual culture system of human alveolar epithelial chip devices with viabilities of at least a twenty-
cells in the upper chamber and pulmonary micro- eight- day period, and the integration of these
vascular endothelial cells in the lower chamber devices into linked platforms to improve model-
separated by a flexible, microporous membrane. ing inter-organ interaction. One of the first inte-
After a sixteen-day incubation period, cyclic grated organ-on-chip devices from 2004 used
changes in pressure are applied to two air cham- microfabrication technology to create a three-
bers running parallel to the cell chambers, thereby chamber (lung, liver, “other”) microscale mam-
stretching and relaxing the membrane to mimic malian cell culture platform on a one- inch square
the mechanical forces of breathing. This chip has silicon chip (Sin et al. 2004). The device had a
been used to model pulmonary inflammatory flow rate of 1.76 μL/min and was designed to
responses to the proinflammatory mediator tumor model an approximated physiological liquid-to-
necrosis factor alpha (TNF-α); show the phago- cell ratio plus hydrodynamic shear stress, while
cytic response of neutrophils to foreign E.Coli maintaining the organ-specific liquid residence
bacteria introduced into the device; escalate the time parameters determined through physiologi-
toxic and inflammatory responses of the lung to cally based pharmacokinetic modeling (PBPK).
silica nanoparticles through mechanical stress; A dissolved oxygen sensor was integrated into
and demonstrate the impact of clinically-relevant the system, making it possible to integrate real-
concentrations of interleukin−2 (IL-2) on micro- time sensors into such devices (see Sect. 3.4).
vascular integrity (Huh et al. 2012). The basic The integration of multiple organs on a com-
design has since been used to model a number of mon platform provides a wealth of additional
other tissues including gut (Kim et al. 2012), kid- research possibilities, including analyses of
34 L. A. Low et al.
Fig. 3.3 – A TissUse

(Germany) four-organ
chip, incorporating
intestine (1), liver (2),
skin (3) and kidney (4)
tissues. A surrogate
blood medium is shown
in pink. An excretory
flow circuit is shown in
yellow. Each
microphysiological fluid
flow circuit is operated
by a separate peristaltic
micropump. (Image
reproduced under a
Creative Commons 3.0
license from
Maschmeyer et al.)
organ–organ interactions, pharmacokinetics/ cells (Takahashi and Yamanaka 2006; Yu et al.
pharmacodynamics (PK/PD), hormone/immune/ 2007) has led to the prospect of using iPSCs to
cytokine responses to drug exposure, and physi- seed organ-on-chip devices, both as a means to
ological reactions to drug metabolites. To address develop integrated multi-organ systems from the
the pharmacokinetics of drug absorption, metab- same individual and as a possible tool for person-
olism and secretion, Maschmeyer et al. alized medicine approaches to numerous dis-
(Maschmeyer et al. 2015) developed an inte- eases. The ability to differentiate iPSCs into
grated microphysiological platform that main- multiple cell types eliminates the need for isola-
tains the functionality of liver, kidney, intestine tion of primary tissues from donors or patients,
and skin over 28 days in co-culture (Fig. 3.3). A which is often a difficult and invasive procedure
peristaltic on-chip micropump ensures pulsatile (and is not possible for organ systems like the
media flow interconnecting the four tissue cul- brain), or may not produce high enough cell
ture compartments through microfluidic chan- yields to be of use. iPSCs also provide a renew-
nels. A second microfluidic circuit ensures able cell source for device seeding, enabling
drainage of the fluid excreted through the kidney wider reproducibility and therefore utility of the
epithelial cell layer. This dual microfluidic sys- platforms for the research communities. iPSC
tem enables ADME (absorption/distribution/ generation has led to rapid development of, as an
metabolism/excretion) profiling and repeated example, brain and blood-brain barrier (BBB)
dose toxicity testing of drug candidates. organ-on-chip devices, including a vascularized
motor neuron model, which demonstrates
enhanced neuronal activity, induction of vascu-
3.3.2 A Marriage of Technologies lar-neural interaction genes, and developmental
gene expression profiles that are indicative of
The revolution in stem cell technology driven by more in vivo-like signatures when compared to
the discovery that human induced pluripotent conventional 2D culture systems (Sances et al.
stem cells (iPSCs) can be generated from adult 2018).
3.4 ow Engineering Advances

H to environmental cues, e.g. the presence of a
Have, and Will Continue to, pathogen, drug, or toxicant. Researchers have
Push the Field Forward begun to incorporate existing cell analysis tech-
niques with microfluidics, such as enzyme-linked
Current approaches to developing organ-on-a- immunosorbent assays (ELISA) (Riahi et al.
chip platforms are rooted in plate-based cell cul- 2016; Luan et al. 2017), trans-epithelial electrical
ture, microfluidics, and 3-dimensional printing. resistance (TEER) (van der Helm et al. 2017;
Multi-well cell culture plates gave rise to bioen- Henry et al. 2017), and multi-electrode arrays
gineered tissues by allowing for the 3D culture of (MEA) (Koutsouras et al. 2017; Maoz et al.
cell suspensions and later organoids, e.g. liver 2017). Recently, several of these techniques have
embryoid bodies (Au et al. 2014) and cerebral been simultaneously included into a single organ-
organoids (Lancaster et al. 2013). Microfluidics on-a-chip platform (Junaid et al. 2017; Zhang
then contributed the capability to reduce the cul- et al. 2017). This platform features physical (e.g.
ture well size and control with great precision the temperature, pH detection), chemical (e.g. immu-
number of cells, and therefore the organoid size, nobiosensors to monitor excreted biomarkers or
as well as the quantities of reagents (Du et al. to detect oxygen levels), and optical (e.g. sensors
2016). It also enabled researchers to design chips to observe neuro-muscular contractility) sensors,
for larger biomimetic tissue constructs (Tan and which enable continuous measurements of the
Desai 2004) and introduce structural complexity, cellular environment and behavior. Future work
e.g. permeable membranes for lung (Kniazeva should focus on long-term (e.g. months) stability
et al. 2011) and intestinal tissue (Kim et al. 2012) of the embedded sensors without loss of accuracy
engineering. More recently, 3D bioprinting of or precision; multiplexing sensors for concurrent
biopolymers and cell suspensions has provided monitoring of key biomarkers from several engi-
the flexibility to pattern tissue components on a neered tissues; methods for continuous process-
larger scale (up to 1 mm in height), and with high ing of the large sensor and optical data; and
precision (usually on the order of 50 μm). Most feedback loops from the sensors to the environ-
importantly, bioprinting enables easy integration mental controls, including perfusate contents and
of fluidic channels for nutrient supply and cellu- flow rate. Together, these technical advances
lar waste removal – effectively vascular channels promise to increase the utility of MPS in a variety
(Kolesky et al. 2014; He et al. 2016). These of contexts and create cutting-edge technologies
developments in engineering, coupled with syn- that can give information on cell function in ways
ergistic engineering to combine organoid and that are currently not possible using any other
organ-on-chip technology, have moved cell cul- platform.
ture studies over the last two decades from two
dimensions into the 3D and 4D (temporal con-
trol) realms when looking at responses over time. 3.5 Present Considerations
These complex tissues and multi-tissue platforms and Future Challenges
that are viable and can function for several weeks
(Materne et al. 2015; Skardal et al. 2017; 3.5.1 Cell Sources
Ronaldson-Bouchard et al. 2018) have significant
translational potential for drug screening and rare For organ-on-chip development, cell sources cur-
disease research. rently come anywhere from primary tissue
One powerful advantage to organ-on-chip biopsies, to commercially available cell lines, to
technology is the possibility to monitor real-time stem cell-derived cell populations. Primary cells
outcomes of a variety of cell- and tissue-specific derived from some types of tissue biopsies may
responses in non-invasive ways. This, however, be grown and maintained in culture for long peri-
requires continuous monitoring of the engineered ods of time and therefore allow for longer term
tissue to ascertain its functionality and response studies than 2D cell culture systems. Comparison
36 L. A. Low et al.
of on-chip experiments to fresh tissue collection organ physiology which will require the efforts
is important because these collections often best of researchers in multiple fields of expertise to
mimic in vivo physiology, but they are not widely overcome. Some of these may be related to organ
available from human patients. Other cell sources scaling, which is the issue of how to represent the
are induced pluripotent stem cells (iPSCs) from size and/or cell number of linked organ systems
reprogrammed adult somatic cells, and are often in a physiologically relevant manner. For exam-
used in organ-on-chip platforms. iPSCs may be ple, the liver is considerably larger than the kid-
derived from patients and represent a potentially ney, so a joint system must scale the tissue
unlimited cell source, but this technology is still constructs accordingly. Additionally, cell metab-
evolving. For instance, although the maturation olism may be different in different organ systems,
signals recorded in the vascularized motor and with variable cell maturity and time in cul-
neuron-on-chip device are encouraging (Sances ture, the basis for normal flow rate for each tissue
et al. 2018), most systems still exhibit embryonic compartment is not completely understood
gene expression profiles and hence significant (Wikswo et al. 2013). Furthermore, to keep the
development is still needed to drive iPSC-derived multiple different cell types healthy and func-
cell types toward an adult phenotype. This is an tional, multi-organ-on-a-chip platforms require
essential step for application of organ-on-chip development of a perfusate that maintains physi-
devices in modeling of adult-onset diseases, but ologically relevant circulation for all included tis-
the recent development of genetic editing tech- sues – a fluid that functions either as a blood
niques offers exciting potential for the creation of surrogate or as interstitial fluid (Abaci and Shuler
multiple tissue types from the same donor (iso- 2015). Such a fluid would contain all nutrients,
genic cell lines), or introduction and deletion of growth factors, and chemokines required for all
specific genetic sequences to model or correct tissue types within the system. If attained, this
monogenetic diseases. design would greatly simplify the geometry of
microfluidic or 3D printed organ-on-a-chip plat-
forms as only one perfusion channel would be
3.5.2 Organ-on-Chip Integration: needed for all tissues.
Biological and Technical
Challenges • Technical challenges:
As previously stated, integration of multiple tis- These are challenges related to the physical
sue models is of great interest because cells and linkage and integration of multiple tissue plat-
organ systems communicate with each other by forms. They include basic challenges such as
secreting exosomes and soluble factors that are presence of air bubbles (see Sect. 3.5.3); mainte-
important for peripheral crosstalk. Microfluidic nance of sterility when linking separately cul-
advances have allowed individual organs-on- tured platforms (and preventing leakage); and
chips to be connected to one another to mimic the selection of materials for platform fabrication
in vivo role of the circulation thus enabling the (see Sect. 3.5.3). For physical coupling of micro-
recapitulation of certain aspects of homeostasis. fluidic systems, methods for fluidic integration
Comprehensive body-on-chip integration aims to are categorized as: (1) static, (2) single-pass per-
more fully mimic the physiology of drug delivery fusion, and (3) recirculation (of both common or
and metabolism in the body. However, several tissue-specific media). Static microfluidic con-
factors need to be considered when designing nections between organ chips depend on physical
these systems: proximity rather than convective flow, and organ
crosstalk is facilitated by culturing all cells and
• Biological challenges: organs in the same well (Li et al. 2012). However,
one limitation of such systems is a potential for
These are challenges which require ongoing exponential increase of cell toxicity, which can
research in order to understand basic cellular and complicate the interpretation of results. Single-
pass perfusion creates culture systems that Selecting the right material for microfluidic
arrange organ chips in parallel, in series, or both. device fabrication is another challenge for an
Here, the drawback is that unidirectionality of the organ-on-chip platform developer. The chip
perfusion only enables crosstalk to organs down- material should be biocompatible with the cho-
stream and does not allow for upstream feedback, sen tissue and should support a physiologically-
which occurs in vivo. Recirculating microfluidic relevant environment. Polydimethylsiloxane
flow more closely mimics in vivo circulation and (PDMS) is commonly used in many academic
allows for both downstream and upstream cross- laboratories because of its low cost, ease of fabri-
talk (Oleaga et al. 2016). Zhang et al. (2017) cre- cation, optical transparency, lack of cellular tox-
ated a recirculating microfluidic multi-organ icity and elastomeric properties that simplify
organoid platform which incorporates multiple chip bonding (Shirure and George 2017).
sensors, thereby allowing real-time information However, PDMS is hydrophobic and therefore is
capture that could be employed to create feed- difficult to surface-functionalize; because of
back algorithms for evaluating drug responses in these properties, PDMS absorbs hormones, small
near real-time. Other methods of fluidic flow may molecules and drugs thereby altering their con-
be driven by gravity and passively-controlled centrations, and affecting reproducibility of the
hydraulic resistances (Esch et al. 2016). results (Chen et al. 2016). PDMS is also perme-
Additional technical challenges include incorpo- able to gases (Merkel et al. 2000), which leads to
ration into platforms of organ-specific biomi- spontaneous formation of gas bubbles that impact
metic stimuli (mechanical, electrical and perfusion flow inside microfluidic devices – the
chemical) to drive the functionality of the organ dreaded “gas bubble problem” (Wang et al.
chips and maintain physiologically-relevant 3D 2012). No other currently used materials – glass,
tissues (Huh et al. 2010; Nunes et al. 2013; Sun thermoplastics, or hydrogels – fully satisfy the
and Nunes 2016; Marturano-Kruik et al. 2018). needs, hence the search continues for an optimal
material – one which is surface “functionaliz-
able”, yet non-sticky to cells; gas permeable only
3.5.3 Manufacturing, Materials under controlled, reversible conditions; and cost-
and Reproducibility effective and simple to handle.
Finally, although the predictive value and
Development of organ-on-chip technology has physiological relevance of various organ-on-chip
been hampered by several factors, which include systems have been demonstrated, their potential
inadequate community standards for manufactur- for medium or high-throughput screenings (HTS)
ing, poor performance of materials, and repro- is still to be realized. Adaptations such as massive
ducibility issues. These factors have limited the parallelization and automation are among the
high throughput capabilities for organ-on-chip biggest challenges for meeting industrial bench-
systems. While these difficulties remain, prog- marks of HTS. Further, low-throughput charac-
ress is continually being made, especially with teristics of perfusion, sampling, and cell injection,
technical advances in 3D printing. One of the as well as limitations of fabrication, scale up, and
oldest and most powerful 3D bioprinting tech- integration of online analytics/sensors must be
niques is laser-assisted stereolithography (SLA) addressed before chip platforms could be
(Groll et al. 2016). To date, there are many grow- employed for HTS (Probst et al. 2018).
ing companies utilizing SLA technologies,
including Organovo, N3D, Bio3D, and
LaserPrint. Although a 3D-printed organ-on-a- 3.5.4 Standardization
chip platform is yet to be fully commercialized, it and Validation Considerations
is expected that adoption of these devices will be
realized with continued advances in automation As organ-on-chip models functionally evolve, a
and additive manufacturing. need for standardized, good manufacturing prac-
tice (GMP) and good laboratory practice (GLP)
38 L. A. Low et al.
procedures are becoming more important for New Drug (IND) applications in the United
comparing data across laboratories. Validation of States. Additionally, if organ-on-chip and inte-
tissue health and functionality assays must be grated “body-on-chip” platforms aim to reduce
robust, reliable, and reproducible. However, the use or replace animal models, and be quali-
because there is no single ‘validation’ method or fied through the US FDA validation strategy of
assay that fits all tissue types, standardization of Drug Development Tools (DDTs), the validation
validation methods is problematic. For example, process of physiologically relevant in vitro sys-
if a system is designed to model and quantify tems should follow similar, if not the same crite-
transport across barrier tissues, it becomes impor- ria as outlined in the DDT guidance document
tant to verify the integrity of tight junctions that (Rebelo et al. 2016). In the US, the current FDA
comprise the barrier. One approach could be an DDT program is voluntary and uses a “fit-for-
assessment of tight junction viability through a purpose” qualification. Once an animal model is
quantification of tight-junction protein expres- qualified for a specific context of use as a DDT,
sion (Shah et al. 2016). However, these measures the pharmaceutical industry can use the tool for
would be less useful in non-barrier tissues, which the qualified purpose during product develop-
exemplifies the need to validate tissue functions ment, and FDA reviewers can be confident in
applicable to the experimental model for its spe- applying the DDT without additional supporting
cific context of use. To help address this chal- data. Therefore, meeting FDA DDT requirements
lenge, in 2017 the National Center for Advancing by the multi-organ or body-on-a-chip platforms
Translational Sciences (NCATS), part of the NIH should provide a reliable approach for earning
in the United States, funded Tissue Chip Testing industry acceptance of this technology for drug
Centers to independently assess the robustness discovery and development. The NIH-funded
and reliability of various MPS platforms. In col- Tissue Chip Testing Centers and MPS Database
laboration with the FDA and pharmaceutical in the US and the Organ-on-Chip Development
industry, these Testing Centers validate the out- (ORCHID) initiative in Europe are supporting
comes and assays provided by the platform these efforts.
designers, with continuously evolving validation
criteria being developed and published (Low and
Tagle 2017a, b; Sakolish et al. 2018). To house 3.6 Conclusions and Future
this data, and create a standardized, centralized Directions
database of MPS data from a variety of users,
NCATS also funded the establishment of an MPS Although existing challenges need to be
Database (MPS-Db) to provide developers and addressed to accommodate the use of organ-on-
end-users access to a publicly accessible reposi- chip technology in a routine lab workflow, the
tory of MPS data. enormous potential of these devices to expand
our understanding of human cellular biology,
improve toxicity and efficacy testing in drug
3.5.5 Commercialization development and model human disease in the
and Regulatory/Industry context of precision medicine will continue to
Buy-in of Organ-on-Chip accelerate the development of this technology. To
Systems give an example of a future avenue of potentially
impactful research: Today, chronic inflammatory
As noted above, manufacturing advancements processes are at the core of disease pathology in
are still needed to reach GMP and GLP levels, seven of the top ten leading causes of mortality in
both for commercialization of organ-on-chip sys- the developed world. Extensive evidence sup-
tems and for encouraging buy-in from regulatory ports the role of inflammation in heart disease
bodies that will receive data from these platforms (Awan and Genest 2014), cancer (Hanahan and
being included in, for example, Investigational Weinberg 2011), chronic pulmonary disease
(Gan et al. 2004), stroke (Kawabori and Yenari applications of the technology as it continues to
2015), Alzheimer’s disease (Amor et al. 2010), evolve. While challenges remain, the promise of
diabetes (Biondi-Zoccai et al. 2003), nephritis these tools remains transformative, inspiring
(Zheng and Zheng 2016) and Parkinson’s disease researchers and clinicians and providing hope for
(Stephenson et al. 2018), as well as holding a cru- patients.
cial role in general injury and healing. A better
understanding of the processes that initiate, drive,
and terminate inflammation, both acute and References
chronic, could help to uncover new treatments
with impact in wound healing, allergies, and the Abaci HE, Shuler ML (2015) Human-on-a-chip design
strategies and principles for physiologically based
health of the aging population. Utilization of pharmacokinetics/pharmacodynamics modeling.
organ-on-chip technology will enable researchers Integr Biol 7(4):383–391
to study chronic inflammation in multiple organs Abaci HE et al (2016) Human skin constructs with
including lung, gut, brain, and skin, as well as spatially controlled vasculature using primary and
iPSC-derived endothelial cells. Adv Healthc Mater
investigate the new roles of immune cells during 5(14):1800–1807
inflammation associated with cancer. Agarwal A et al (2013) Microfluidic heart on a chip for
When considering applications for disease higher throughput pharmacological studies. Lab Chip
modeling, organ-on-chip technology has also 13(18):3599–3608
Amor S et al (2010) Inflammation in neurodegenerative
provided new tools for modeling metastatic can- diseases. Immunology 129(2):154–169
cer mechanisms, as these models can recreate An F et al (2015) Organ-on-a-Chip: new platform for bio-
cell-cell or cell-ECM interactions, spatiotempo- logical analysis. Anal Chem Insights 10:39–45
ral chemical/physical gradients or the (hydro) Arrowsmith J (2011a) Trial watch: phase II failures:
2008–2010. Nat Rev Drug Discov 10(5):328–329
dynamic properties of the cellular microenviron- Arrowsmith J (2011b) Trial watch: phase III and sub-
ment. Over the coming years, the integration of mission failures: 2007–2010. Nat Rev Drug Discov
advanced quantitative optical imaging methods, 10(2):87–87
screening and biosensing techniques into on-chip Arslan SY et al (2015) Novel three dimensional human
Endocervix cultures respond to 28-day hormone treat-
systems together with Big Data analyses is ment. Endocrinology 156(4):1602–1609
expected to provide clinicians and pharmaceuti- Au SH et al (2014) Hepatic organoids for microfluidic
cal companies with a new generation of highly drug screening. Lab Chip 14(17):3290–3299
predictive metastasis-on-chip models (Caballero Awan Z, Genest J (2014) Inflammation modulation and
cardiovascular disease prevention. Eur J Prev Cardiol
et al. 2017). Regarding future applications of 22(6):719–733
MPS technology, clinical trial readiness studies Balls M (1995) Defining the role of ECVAM in the devel-
are beginning to collect peripheral blood mono- opment, validation and acceptance of alternative tests
nuclear cells (PBMCs) from blood samples for and testing strategies. Toxicol In Vitro 9(6):863–869
Biondi-Zoccai GGL et al (2003) Atherothrombosis,
iPSC generation of many tissue types, thus inflammation, and diabetes. J Am Coll Cardiol
enabling precision medicine efforts around 41(7):1071–1077
organ-on-chip technology and generation of Brown JA et al (2015) Recreating blood-brain barrier
“you-on-a-chip” to test individualized treat- physiology and structure on chip: a novel neuro-
vascular microfluidic bioreactor. Biomicrofluidics
ments. This effort has particularly impactful 9(5):054124
potential for rare diseases such as Amyotrophic Brown JA et al (2016) Metabolic consequences of inflam-
Lateral Sclerosis (ALS) and Frontotemporal matory disruption of the blood-brain barrier in an
Degeneration (FTD), where the use of tissue chip organ-on-chip model of the human neurovascular unit.
J Neuroinflammation 13(1):306
technology to screen therapeutics may lead to the Caballero D et al (2017) Organ-on-chip models of can-
identification of the right drug, for the right cer metastasis for future personalized medicine: from
patient at the right time. chip to the patient. Biomaterials 149(Supplement
The organ-on-chip field is fast-developing C):98–115
Casey W et al (2015) A new path forward: the
worldwide, as stakeholders from various sectors Interagency Coordinating Committee on the
are beginning to appreciate the broad potential
40 L. A. Low et al.
Validation of Alternative Methods (ICCVAM) and Huh D et al (2012) A human disease model of drug toxicity–
National Toxicology Program’s Interagency Center induced pulmonary Edema in a lung-on-a-Chip micro-
for the Evaluation of Alternative Toxicological device. Sci Trans Med 4(159):159ra147–159ra147
Methods (NICEATM). J Am Assoc Lab Anim Sci Jackson EL, Lu H (2016) Three-dimensional models for
54(2):170–173 studying development and disease: moving on from
Chen T et al (2016) A drug-compatible and temperature- organisms to organs-on-a-chip and organoids. Integr
controlled microfluidic device for live-cell imaging. Biol Quant Biosci Nano Macro 8(6):672–683
Open Biol 6(8):160156 Jang K-J et al (2013) Human kidney proximal tubule-on-
Cheng CS et al (2016) Cell density and joint microRNA- a-chip for drug transport and nephrotoxicity assess-
133a and microRNA-696 inhibition enhance differen- ment. Integr Biol 5(9):1119–1129
tiation and contractile function of engineered human Jiang, Y., et al. (2018). Maturation of cardiomyocytes
skeletal muscle tissues. Tissue Eng A 22(7–8):573–583 derived from human pluripotent stem cells: current
de Peppo GM et al (2013) Engineering bone tissue substi- strategies and limitations. Mol Cells 41(7):613–621
tutes from human induced pluripotent stem cells. Proc Junaid A et al (2017) An end-user perspective on organ-
Natl Acad Sci 110(21):8680–8685 on-a-chip: assays and usability aspects. Curr Opin
Du G et al (2016) Microfluidics for cell-based high Biomed Eng 1:15–22
throughput screening platforms—a review. Anal Chim Kanamori T et al (2018) Technical aspects of micro-
Acta 903(Supplement C):36–50 physiological systems (MPS) as a promising wet
Dutta D et al (2017) Disease Modeling in stem cell-derived human-in-vivo simulator. Drug Metab Pharmacokinet
3D organoid systems. Trends Mol Med 23(5):393–410 33(1):40–42
Esch EW et al (2015) Organs-on-chips at the frontiers of Kawabori M, Yenari MA (2015) Inflammatory responses
drug discovery. Nat Rev Drug Discov 14:248–260 in brain ischemia. Curr Med Chem 22(10):1258–1277
Esch MB et al (2016) Modular, pumpless body-on-a-chip Kim HJ et al (2012) Human gut-on-a-chip inhabited by
platform for the co-culture of GI tract epithelium and microbial flora that experiences intestinal peristalsis-
3D primary liver tissue. Lab Chip 16(14):2719–2729 like motions and flow. Lab Chip 12(12):2165–2174
Fashe MM et al (2015) Species-specific differences in Kniazeva T et al (2011) A microfluidic respiratory assist
the in vitro metabolism of Lasiocarpine. Chem Res device with high gas permeance for artificial lung
Toxicol 28(10):2034–2044 applications. Biomed Microdevices 13(2):315–323
Fernandez CE et al (2016) Human vascular microphysi- Kolesky DB et al (2014) 3D bioprinting of vascularized,
ological system for in vitro drug screening. Sci Rep heterogeneous cell-laden tissue constructs. Adv Mater
6:21579 26(19):3124–3130
Foulke-Abel, J., et al. (2016). Human enteroids as a model Koutsouras DA et al (2017) Simultaneous monitoring of
of upper small intestinal ion transport physiology and single cell and of micro-organ activity by PEDOT:PSS
pathophysiology. Gastroenterology 150(3): 638–649. covered multi-electrode arrays. Mater Sci Eng C
e638 81:84–89
Gan W et al (2004) Association between chronic obstruc- Lancaster MA et al (2013) Cerebral organoids model
tive pulmonary disease and systemic inflammation: human brain development and microcephaly. Nature
a systematic review and a meta-analysis. Thorax 501:373–379
59(7):574–580 Laronda MM et al (2013) Recreating the female repro-
Groll J et al (2016) Biofabrication: reappraising the defi- ductive tract in vitro using iPSC technology in a
nition of an evolving field. Biofabrication 8(1):013001 linked microfluidics environment. Stem Cell Res Ther
Guillouzo A (1998) Liver cell models in in vitro toxicol- 4(Suppl 1):S13–S13
ogy. Environ Health Perspect 106(Suppl 2):511–532 Lee-Montiel FT et al (2017) Control of oxygen ten-
Hanahan D, Weinberg RA (2011) Hallmarks of Cancer: sion recapitulates zone-specific functions in human
the next generation. Cell 144(5):646–674 liver microphysiology systems. Exp Biol Med
He Y et al (2016) Developments of 3D printing micro- 242(16):1617–1632
fluidics and applications in chemistry and biology: a Li AP et al (2012) Definition of metabolism-dependent
review. Electroanalysis 28(8):1658–1678 xenobiotic toxicity with co-cultures of human hepa-
Henry OYF et al (2017) Organs-on-chips with integrated tocytes and mouse 3T3 fibroblasts in the novel inte-
electrodes for trans-epithelial electrical resistance grated discrete multiple organ co-culture (IdMOC)
(TEER) measurements of human epithelial barrier experimental system: results with model toxicants
function. Lab Chip 17(13):2264–2271 aflatoxin B1, cyclophosphamide and tamoxifen. Chem
Herland A et al (2016) Distinct contributions of astrocytes Biol Interact 199(1):1–8
and Pericytes to Neuroinflammation identified in a Li J et al (2016) Recent advances in bioprinting tech-
3D human blood-brain barrier on a Chip. PLoS One niques: approaches, applications and future prospects.
11(3):e0150360 J Trans Med 14(1):271
Huh D et al (2010) Reconstituting organ-level lung func- Liu C et al (2018) Modeling human diseases with induced
tions on a Chip. Science 328(5986):1662–1668 pluripotent stem cells: from 2D to 3D and beyond.
Development 145(5):dev156166
Lopes FM et al (2017) Mimicking Parkinson’s disease in Probst C et al (2018) High-throughput organ-on-a-chip
a dish: merits and pitfalls of the most commonly used systems: current status and remaining challenges. Curr
dopaminergic in vitro models. NeuroMolecular Med Opin Biomed Eng 6:33–41
19(2):241–255 Profile Toolkit (2018) The Dynamic U.S. Research and
Lou Y-R, Leung AW (2018) Next generation organoids Development Ecosystem. from https://http://www.
for biomedical research and applications. Biotechnol phrma.org/industryprofile/2018/
Adv 36(1):132–149 Rebelo SP et al (2016) Validation of bioreactor and
Low LA, Tagle DA (2017a) Organs-on-chips: Progress, human-on-a-Chip devices for chemical safety assess-
challenges, and future directions. Exp Biol Med ment. Validation of alternative methods for toxicity
(Maywood, NJ) 242(16):1573–1578 testing. C. Eskes and M. Whelan. Springer, Cham,
Low LA, Tagle DA (2017b) Tissue chips - innovative pp 299–316
tools for drug development and disease modeling. Lab Riahi R et al (2016) Automated microfluidic platform
Chip 17(18):3026–3036 of bead-based electrochemical immunosensor inte-
Luan Q et al (2017) A microfluidic in-line ELISA for grated with bioreactor for continual monitoring of cell
measuring secreted protein under perfusion. Biomed secreted biomarkers. Sci Rep 6:24598
Microdevices 19(4):101 Roberts RA et al (2014) Reducing attrition in drug devel-
Madden L et al (2015) Bioengineered human myobundles opment: smart loading preclinical safety assessment.
mimic clinical responses of skeletal muscle to drugs. Drug Discov Today 19(3):341–347
elife 4:e04885 Ronaldson-Bouchard K et al (2018) Advanced maturation
Maoz BM et al (2017) Organs-on-chips with combined of human cardiac tissue grown from pluripotent stem
multi-electrode array and transepithelial electri- cells. Nature 556(7700):239–243
cal resistance measurement capabilities. Lab Chip Sakolish C et al (2018) Technology transfer of the micro-
17(13):2294–2302 physiological systems: a case study of the human prox-
Martin L et al (2017) How much do clinical trials cost? imal tubule tissue Chip. Sci Rep 8(1):14882–14882
Nat Rev Drug Discov 16:381–382 Sances S et al (2018) Human iPSC-derived endothelial
Marturano-Kruik A et al (2018) Biomechanical regulation cells and microengineered organ-Chip enhance neuro-
of drug sensitivity in an engineered model of human nal development. Stem Cell Reports 10(4):1222–1236
tumor. Biomaterials 150:150–161 Sasahara K et al (2015) Pharmacokinetics and metabolism
Maschmeyer I et al (2015) A four-organ-chip for inter- of Delamanid, a novel anti-tuberculosis drug, in ani-
connected long-term co-culture of human intes- mals and humans: importance of albumin metabolism
tine, liver, skin and kidney equivalents. Lab Chip in vivo. Drug Metab Dispos 43(8):1267–1276
15(12):2688–2699 Schimek K et al (2018) Bioengineering of a full-thickness
Materne E-M et al (2015) The multi-organ Chip - a micro- skin equivalent in a 96-well insert format for substance
fluidic platform for long-term multi-tissue Coculture. permeation studies and organ-on-a-chip applications.
J Vis Exp: JoVE 98:52526 Bioengineering 5(2)
Mathur A et al (2015) Human iPSC-based cardiac micro- Shah P et al (2016) A microfluidics-based in vitro model
physiological system for drug screening applications. of the gastrointestinal human–microbe interface. Nat
Sci Rep 5:8883 Commun 7:11535
Merkel TC et al (2000) Gas sorption, diffusion, and per- Shirure VS, George SC (2017) Design consider-
meation in poly(dimethylsiloxane). J Polym Sci B ations to minimize the impact of drug absorption in
Polym Phys 38(3):415–434 polymer-based organ-on-a-chip platforms. Lab Chip
Morgan S et al (2011) The cost of drug development: a 17:681–690
systematic review. Health Policy 100(1):4–17 Sin A et al (2004) The design and fabrication of three-
Moya ML et al (2013) In vitro perfused human capillary chamber microscale cell culture analog devices with
networks. Tissue Eng Part C Methods 19(9):730–737 integrated dissolved oxygen sensors. Biotechnol Prog
Niu N, Wang L (2015) In vitro human cell line mod- 20(1):338–345
els to predict clinical response to anticancer drugs. Singh Dolt K et al (2017) Modeling Parkinson’s dis-
Pharmacogenomics 16(3):273–285 ease with induced pluripotent stem cells harboring
Nunes SS et al (2013) Biowire: a platform for maturation α-synuclein mutations. Brain Pathol 27(4):545–551
of human pluripotent stem cell–derived cardiomyo- Skardal A et al (2017) Multi-tissue interactions in an inte-
cytes. Nat Methods 10:781–787 grated three-tissue organ-on-a-chip platform. Sci Rep
Oleaga C et al (2016) Multi-organ toxicity demonstration 7(1):8837
in a functional human in vitro system composed of Sobrino A et al (2016) 3D microtumors in vitro supported
four organs. Sci Rep 6:20030 by perfused vascular networks. Sci Rep 6:31589
Pournasr B, Duncan SA (2017) Modeling inborn Soto-Gutierrez A et al (2017) Pre-clinical and clinical
errors of hepatic metabolism using induced plu- investigations of metabolic zonation in liver diseases:
ripotent stem cells. Arterioscler Thromb Vasc Biol the potential of microphysiology systems. Exp Biol
37(11):1994–1999 Med 242(16):1605–1616
Stephenson J et al (2018) Inflammation in CNS neurode-
generative diseases. Immunology 154(2):204–219
42 L. A. Low et al.
Sun X, Nunes SS (2016) Biowire platform for maturation Wang G et al (2014) Modeling the mitochondrial cardio-
of human pluripotent stem cell-derived cardiomyo- myopathy of Barth syndrome with induced pluripotent
cytes. Methods 101:21–26 stem cell and heart-on-chip technologies. Nat Med
Sutherland ML et al (2013) The national institutes of 20(6):616–623
health microphysiological systems program focuses Wang YI et al (2017) Microfluidic blood–brain bar-
on a critical challenge in the drug discovery pipeline. rier model provides in vivo-like barrier properties
Stem Cell Res Ther 4(1):1–5 for drug permeability screening. Biotechnol Bioeng
Takahashi K, Yamanaka S (2006) Induction of pluripotent 114(1):184–194
stem cells from mouse embryonic and adult fibroblast Waring MJ et al (2015) An analysis of the attrition of drug
cultures by defined factors. Cell 126(4):663–676 candidates from four major pharmaceutical compa-
Takebe T et al (2015) Vascularized and complex organ nies. Nat Rev Drug Discov 14:475–486
buds from diverse tissues via mesenchymal cell-driven Weber EJ et al (2016) Development of a microphysiologi-
condensation. Cell Stem Cell 16(5):556–565 cal model of human kidney proximal tubule function.
Takebe T et al (2017) Synergistic engineering: organoids Kidney Int 90(3):627–637
meet organs-on-a-chip. Cell Stem Cell 21(3):297–300 Wikswo JP et al (2013) Engineering challenges for instru-
Tan W, Desai TA (2004) Layer-by-layer microfluidics for menting and controlling integrated organ-on-Chip
biomimetic three-dimensional structures. Biomaterials Systems. IEEE Trans Biomed Eng 60(3):682–690
25(7–8):1355–1364 Workman MJ et al (2016) Engineered human pluripotent-
Tanataweethum N et al (2018) Establishment and char- stem-cell-derived intestinal tissues with a functional
acterization of a primary murine adipose tissue-chip. enteric nervous system. Nat Med 23:49
Biotechnol Bioeng 0(0) Xiao S et al (2017) A microfluidic culture model of the
van der Helm MW et al (2017) Fabrication and validation human reproductive tract and 28-day menstrual cycle.
of an organ-on-chip system with integrated electrodes Nat Commun 8:14584
to directly quantify transendothelial electrical resis- Yoon No D et al (2015) 3D liver models on a microplat-
tance. JoVE 127:e56334 form: well-defined culture, engineering of liver tissue
Vernetti LA et al (2016) A human liver microphysiology and liver-on-a-chip. Lab Chip 15(19):3822–3837
platform for investigating physiology, drug safety, and Yu J et al (2007) Induced pluripotent stem cell lines
disease models. Exp Biol Med 241(1):101–114 derived from human somatic cells. Science
Vunjak-Novakovic G et al (2013) HeLiVa platform: 318(5858):1917–1920
integrated heart-liver-vascular systems for drug test- Zhang B, Radisic M (2017) Organ-on-a-chip devices
ing in human health and disease. Stem Cell Res Ther advance to market. Lab Chip 17(14):2395–2420
4(1):1–6 Zhang YS et al (2016) Bioprinting 3D microfibrous scaf-
Wagner J et al (2017a) A dynamic map for learning, com- folds for engineering endothelialized myocardium and
municating, navigating and improving therapeutic heart-on-a-chip. Biomaterials 110:45–59
development. Nat Rev Drug Discov 17:150 Zhang YS et al (2017) Multisensor-integrated organs-on-
Wagner JA et al (2017b) Application of a dynamic chips platform for automated and continual in situ
map for learning, communicating, navigating, and monitoring of organoid behaviors. Proc Natl Acad Sci
improving therapeutic development. Clin Transl Sci 114(12):E2293–E2302
11(2):166–174 Zheng Z, Zheng F (2016) Immune cells and inflammation
Wallet MA et al (2017) Isogenic cellular systems model in diabetic nephropathy. J Diabetes Res 2016:10
the impact of genetic risk variants in the pathogenesis Zhu J et al (2016) Human fallopian tube epithelium co-
of type 1 diabetes. Front Endocrinol 8(276) culture with murine ovarian follicles reveals cross-
Wang Y et al (2012) Systematic prevention of bubble talk in the reproductive cycle. Mol Hum Reprod
formation and accumulation for long-term culture of 22:756–767
pancreatic islet cells in microfluidic device. Biomed
Microdevices 14(2):419–426
Engineering Patient-on-a-Chip
Models for Personalized Cancer 4
Medicine
David Caballero, Rui L. Reis, and Subhas C. Kundu
Abstract of a new generation of highly predictable pre-

Traditional in vitro and in vivo models typi- clinical models capable to reproduce in vitro
cally used in cancer research have demon- the biological complexity of the human body.
strated a low predictive power for human Recent advances in nanotechnology and tissue
response. This leads to high attrition rates of engineering have enabled the development of
new drugs in clinical trials, which threaten predictive organs-on-a-chip models of cancer
cancer patient prognosis. Tremendous efforts with advanced capabilities. These models can
have been directed towards the development reproduce in vitro the complex three-dimen-
sional physiology and interactions that occur
between organs and tissues in vivo, offering
multiple advantages when compared to tradi-
tional models. Importantly, these models can
D. Caballero (*) · S. C. Kundu
3B’s Research Group, I3Bs – Research Institute on be tailored to the biological complexity of
Biomaterials, Biodegradables and Biomimetics, individual cancer patients resulting into bio-
University of Minho, Headquarters of the European mimetic and personalized cancer patient-on-
Institute of Excellence on Tissue Engineering and a-chip platforms. The individualized models
Regenerative Medicine, Barco, Guimarães, Portugal
provide a more accurate and physiological
ICVS 3Bs PT Government Associate Lab, environment to predict tumor progression on
e-mail: dcaballero@i3bs.uminho.pt; patients and their response to drugs. In this
kundu@i3bs.uminho.pt chapter, we describe the latest advances in the
R. L. Reis field of cancer patient-on-a-chip, and discuss
3B’s Research Group, I3Bs – Research Institute on about their main applications and current chal-
Biomaterials, Biodegradables and Biomimetics, lenges. Overall, we anticipate that this new
paradigm in cancer in vitro models may open
Regenerative Medicine, Guimarães, Portugal up new avenues in the field of personalized –
cancer – medicine, which may allow pharma-
Braga, Guimarães, Portugal ceutical companies to develop more efficient
drugs, and clinicians to apply patient-specific
Precision Medicine, Headquarters at University therapies.
e-mail: rgreis@i3bs.uminho.pt
© This is a U.S. government work and not under copyright protection in the U.S.; foreign copyright 43
protection may apply 2020
https://doi.org/10.1007/978-3-030-36588-2_4
44 D. Caballero et al.
Keywords tumor models and therapies. In this regard, recent

Cancer · In vitro model · Organ-on-a-chip · advances in nanotechnology and tissue engineer-
Patient-on-a-chip · Personalized medicine ing have enabled the development of microflu-
idic-based organs-on-a-chip (OoC) models of
human organs and tissues with high predictive
Highlights power (Bhatia and Ingber 2014; Esch et al. 2015).
OoC have been widely employed for the study of
• Cancer patient-on-a-chip (CPoC) are highly human physio-pathological processes, including
physiological patient-matched microfluidic the numerous events occurring during cancer
models for personalized cancer medicine. progression (Caballero et al. 2017b).
• CPoC allow the study of the mechanistic Nevertheless, the current level of development of
determinants of tumor growth and progression OoC have restricted their use mainly for funda-
as well as for drug ADME investigation. mental research applications (Junaid et al. 2017),
• An active collaboration between all the whereas their clinical and industrial applications
involved stakeholders is crucial to boost the still need to be proven. Major challenges that
use of CPoC-based models in the clinics. have limited their acceptance by the clinical and
industrial market include their technical com-
plexity, lack of standardization, low throughput,
4.1 Introduction and the need for specialized personnel, among
others. Indeed, clinical settings for cancer
Cancer remains a major burden on global health, research demand for user-friendly platforms
being the second leading cause of death world- capable to reproduce in vitro the functional units
wide after ischemic heart diseases (W.H.O 2018). and interactions of organs, and their tumors, from
Despite the significant advances made in under- patients, and predict their response to treatments.
standing and fighting cancer, it is yet not well This new generation of tumor models must,
understood why a large number of patients fail to therefore, be highly personalized requiring the
respond to the treatment, or why some of them integration of cells from cancer patients; this
respond differently to the same therapy. Current results into individualized cancer patient-on-a-
in vitro models, such as two- and three-dimen- chip models. This may be critical for cancer ther-
sional (2D and 3D, respectively) platforms typi- apeutics, where the efficiency (and toxicity) of
cally used for drug discovery, development, and specific chemo-, immune-, and even radio-thera-
screening, are obsolete (Caballero et al. 2017b). pies can be first tested in vitro in a microfluidic
They do not recapitulate the complexity of the model of a cancer patient prior to moving to its
human body and lack fundamental characteristics human counterpart; this may eventually help
of the tumor microenvironment (TME). Similarly, physicians in taking decisions. Overall, the devel-
in vivo models, despite being more physiological, opment of such a breakthrough in clinical diag-
are not predictive to humans. This lack of accu- nosis may have a tremendous impact not only in
racy has caused that many anti-cancerous drugs, the clinics by improving cancer patient progno-
which have progressed through (pre)clinical tri- sis, but also in the biotechnological and pharma-
als, being withdrawn from the market due to their ceutical industry where novel products may be
unanticipated toxicity in humans. Overall, the evaluated in a more efficient and predictable
one-size-fits-all paradigm to treat cancer is not manners. Finally, this may reduce the associated
valid anymore. Therefore, more personalized, costs of patient treatment in the healthcare
physiologically-relevant, and predictable tumor system.
models are needed. This new generation of tumor In this chapter, we show the latest advances in
models must reproduce the biological complexity the field of patient-on-a-chip models for person-
(molecular, structural, and dynamical) of each alized cancer medicine. We discuss on their
individual patient, resulting into patient-specific scientific, clinical, and industrial applications as
4 Engineering Patient-on-a-Chip Models for Personalized Cancer Medicine 45
well as their main advantages compared to the resources in the development of such new gener-
traditional models currently used in drug discov- ation of pre-clinical models.
ery, screening, and cancer research. We finally Overall, drug development would signifi-
point out on the current limitations and the chal- cantly benefit from more advanced and physio-
lenges. This drives us to advance towards the new logical pre-clinical models capable to predict
generation of patient-specific clinical models. patient outcomes and identify toxicity problems
at an early stage of the drug development pro-
cess. This will univocally improve the efficiency
4.2 hy Current Drug Testing
W of drug discovery, increase the clinical success
Fails? rates, and significantly reduce the associated
costs to the healthcare system.
A recent study has estimated the cost associated
to drug discovery and its translation to the market
to be around 2.6 billion dollars, with a 145% 4.3 Traditional Platforms
increase compared to the former decade (DiMasi for Tumor Modeling
et al. 2016). Currently, more than 70–80% of can- and Drug Discovery:
cer research is still based on simplistic in vitro An Overview
models, which oversimplify the in vivo context.
Indeed, in vitro models fail to recapitulate the Tumor models are as diverse as the number and
multicellular content, 3D architecture, vascular- type of cancers. The success or failure of an anti-
ization, or fluid exchange of the human body, cancerous compound is directly correlated to the
which is critical to assess correctly human selected model and its inherent strengths and
organ(s) response, drug diffusion kinetics, and weaknesses. This selection depends on a large
dose efficiency. Similarly, in vivo models are variety of parameters, including simplicity, cost,
incapable to predict satisfactorily the outcome of tumor type or the target under evaluation, among
drugs on patients due to the absence of the human others. The perfect tumor model must accom-
immune system that is of upmost importance plish a set of characteristics, such as (i) predic-
during cancer dissemination and therapeutic tive, (ii) reliable, (iii) reproducible, (iv) fast, (v)
response. Altogether, this lack of accuracy causes affordable, and (vi) simple to use. We next briefly
the failure of about 90% of the drugs (95% for describe about the traditional models typically
cancer-related drugs) that enter to clinical trials, used for modelling cancer and testing drugs,
being the primary causes of failure their efficacy which include both experimental (in vitro and in
and clinical safety (Hay et al. 2014; Mullard vivo) and computational (in silico) approaches,
2016; Seruga et al. 2015). In addition, multiple analyzing their advantages and limitations. For a
approved drugs after evaluation using traditional detailed description, the readers may refer to the
in vitro and animal models have been pulled from recent reviews on the topic (Caballero et al.
the market due to their unexpected toxic effects 2017b; Carvalho et al. 2017; Katt et al. 2016;
in organs not directly targeted by the drug Stock et al. 2016).
(Siramshetty et al. 2016). A recent work compil-
ing and analyzing the data on the attrition of drug
candidates developed by big pharmaceutical 4.3.1 In Vitro Tumor Models
companies have concluded that novel methodolo-
gies are indeed needed for an efficient develop- 4.3.1.1 One-Dimensional Models
ment of new drug (Waring et al. 2015). In this One-dimensional (1D) tumor models are widely
regard, the pharmaceutical and biotechnological unknown despite their ability to recapitulate in a
industries are currently investing huge amount of simple assay the fibrous architecture of the native
extracellular matrix (ECM). Indeed, solid
evidences have shown that the cells seeded in 1D cer progression. However, they display serious
environments display morphodynamic features inherent weaknesses, including a limitation in the
reminiscent of the native scenario. This category number and interaction with other cells, and
include 1D fibers (synthetic or natural-derived), importantly, the 3D architecture of the TME,
micro-patterned ECM lines, or topographic among others.
micro-structures. One-dimensional models have
mainly been employed for the study of direc- 4.3.1.2 Two-Dimensional Models
tional cell motion, a critical phenomenon Two-dimensional in vitro tumor models have
involved in cancer invasion. As an example, an been vastly employed in unravelling the mecha-
array of uniaxial lines coated with ECM proteins nistic determinants of tumor dissemination as
of 1.5 μm in width was employed to assess the well as for assessing the efficiency of anti-can-
dynamic properties of migrating cancer cells cerous drugs. This is mainly a consequence to
(Doyle et al. 2009). Interestingly, it was found their simplicity, easy handling, superior accessi-
that the cells displayed high velocity regardless bility, high throughput, and in general, reduced
of ECM density, similar to cells seeded in 3D cost (Katt et al. 2016). In this type of assay, cells
native-like matrices and in contrast with 2D sur- grow on flat surfaces typically made of plastic,
faces. Additionally, the cells seeded on 1D micro- such as in cell culture dishes, flasks, or multi-well
structures displayed phenotypes and cytoskeleton plates, which are homogenously coated, or pat-
organization similar to those in 3D environments. terned, with proteins of the ECM. Besides drug
This suggests that 1D micro-engineered systems screening applications, and similar to 1D case,
can provide a representative model where to 2D tumor models have widely been employed for
investigate signaling pathways, cellular adhe- the study of cancer cell migration and cell-cell/
sion, and the cytoskeleton organization involved ECM interactions (Doxzen et al. 2013). Recently,
in cancer invasion. this type of strategy was used to separate two iso-
One-dimensional tumor models have also genic populations of human ovarian cancer cells
been employed for the study of cell-cell interac- on micro/nano-patterned surfaces using their dif-
tions, an event that has been demonstrated to be ferent migration dynamics (Shu Fan et al. 2017).
fundamental in cancer dissemination. As an Surprisingly, on flat (homogenous) condition, the
example, micro-patterned fibronectin (FN) or cells displayed similar migration characteristics,
collagen I lines mimicking the native ECM fibers which suggested that on patterned environments
were used to explore the interaction between cancer cells motility varied in a metastasis-
tumor cells and tumor-associated macrophages dependent manner.
observed in vivo (Sharma et al. 2014). Time-lapse Two-dimensional tumor models display a high
microscopy experiments revealed that highly degree of customization that make them ideal
metastatic MTLn3 mammary tumor cells candidates to assess the influence of substrate
migrated with increased motility and lamellipo- interfaces in cancer dissemination. In this regard,
dia activity upon interacting with bone marrow- FN-coated surfaces and hyaluronic acid hydrogel
derived macrophages. Interestingly, this were employed to investigate the dissemination
interaction was perturbed using inhibitors of the of human glioblastoma cells along topological
involved paracrine signaling resulting in a interfaces (Rape and Kumar 2014). The cells
reduced cell migration dynamics. The obtained were seeded in-between both environments,
results validated the hypothesis that macrophages mimicking the vascular basement membrane of
were mediated by a paracrine loop during cancer the human brain and the brain parenchyma,
cell invasion and metastasis, providing key respectively. Results showed that the stiffness of
insights about the important role of macrophages- the FN surface, together with actomyosin activa-
tumor cells crosstalk in cancer expansion. tion, influenced cell migration speed and focal
Overall, 1D models, albeit simple, may be adhesion assembly. This suggested a mechano-
very useful to assess crucial information of can- chemical feedback mechanism governing cell
morphodynamics during tumor dissemination. mechanical properties and cell content (co-cul-
This work provided a simple model to mimic cell ture). They mimic the complexity and permeabil-
migration along vascular interfaces, which could ity of the in vivo scenario, preserving the cell
easily be extended to other types of tumors. phenotype and, in general, gene expression.
It is worth mentioning that 2D models are still Importantly, they also reproduce with higher
the preferred methodology used by pharmaceuti- fidelity the stromal reorganization during tumor
cal companies for drug discovery and screening progression, as recently reported (Brancato et al.
due to their intrinsic advantages. However, they 2017). In this work, biodegradable gelatin-based
also suffer from serious limitations. In particular, micro-carriers were employed to investigate the
2D tumor models fail to recapitulate the bio- desmoplastic reaction activated by the stroma-
chemical heterogeneity and 3D architecture of pancreatic cancer crosstalk. Results showed that
the native TME. Further, the cells grown on these pancreatic cancer cells co-cultured with normal
conditions display extreme phenotypes and gene fibroblasts within the micro-carriers formed
perturbation (Cukierman et al. 2001). They also native-like tumor micro-tissues and were able to
fail to reproduce organ-organ and tissue-tissue trigger the cancer-associated fibroblast pheno-
interactions, and dynamic drug/metabolic pro- type, displaying TME markers.
cesses. Consequently, the effect of drugs tested Finally, and despite all the strengths of 3D
on this type of models cannot be extrapolated to tumor models, they also display certain limita-
the human scenario. tions, which impacts on their predictive power.
Among them, their low throughput, and lack of
4.3.1.3 Three-Dimensional Models immune system, perfusion, vasculature, shear
Three-dimensional tumor models overcome sev- stress forces, organ-organ (and tissue-tissue)
eral of the limitations of their 1D and 2D counter- interactions, and standardization, which threaten
parts by mimicking in a better way the structural, the results relevancy and their clinical and indus-
biochemical, and cellular properties of native tis- trial translations.
sues. Cells seeded in this type of environment
display phenotypes and gene expression reminis-
cent of the in vivo scenario, making 3D tumor 4.3.2 In Vivo Tumor Models
models a physiologically-relevant platform
where to study (cancer) cell-cell/ECM interac- The lack of physiologically-relevant models
tions and the actual effect of drugs. Several works capable to reproduce in vitro the complexity of
have compared the performance of 3D and 2D tumors is one of the main limitations in preclini-
tumor models for drug discovery, showing that cal drug development. To overcome this, in vivo
cancer cells seeded in the former respond to tumor models are typically used before clinical
drugs in a similar way they do in vivo (Breslin trials on patients (Breyer et al. 2015). They are
and O’Driscoll 2016; Riedl et al. 2017; Stock essential to assess the efficiency and toxicity of
et al. 2016). This makes 3D tumor models unique anti-cancerous drugs (and other types of com-
candidates to enable the identification of predic- pounds) as well as to explore the mechanistic
tive biomarkers associated with tumor origin or determinants of the disease. Significant progress
with the resistance to a specific therapy, which is in the development of drugs has been obtained by
more predictive of their outcome in patients. using these models, mainly because they provide
The types of 3D tumor models available are as the biological complexity of a living system and
diverse as the etiology of the disease. Typical recapitulate the dynamic interplay between
models include standard transwell plates, syn- organs and tissues. In vivo tumor models range
thetic or naturally-derived hydrogels, scaffolds, from simple model systems typically used for
and matrices, micro-carriers, or spheroids. Most mechanistic investigations, such as Drosophila,
of these systems allow a high control on the C. elegans, or Zebrafish, to more complex animal
models, such as rodents or primates, which dis- experimentation in a rapid and simple way. They
play higher physiological and genetic similarity also avoid hard-to-measure parameters or com-
to humans. pensates non-measurable factors, and do not
The screening and validation of potential new require expensive and laborious materials and
anti-cancerous drugs is typically performed using experimentation.
(immuno-compromised) animals either with the In silico models have been employed to simu-
tumor implanted sub-cutaneous or in the same late different pathological events ranging from
organ of origin with cancer cells from patients subcellular mechanisms (e.g. gene modulation)
(patient-derived orthoxenograft model). The lat- to macroscopic interaction between cancer cells
ter is the most relevant cancer model system, as it and the TME (e.g. tumor invasion). They have
is presumed to faithfully represent the genomic also been used for rapidly modelling the pharma-
features of primary tumors, as well as the struc- cokinetic profile of anti-cancerous compounds,
ture, signaling, and metastasis of the human or for predicting the response of a specific ther-
tumor. However, animal models are highly apy to a patient as a function of a large set of
expensive, time-consuming, and allow reduced parameters, such as the genetic profile of a
manipulation. Further, they are in general, very patient.
limited in their ability to mimic human carcino- Overall, in silico tumor models can provide
genesis, physiology, and progression; they dis- important insights into the mechanistic determi-
play serious differences with human tumors, in nants of the disease and in predicting the effi-
particular regarding their metabolism and immu- ciency/toxicity of drugs, thus supporting
nology, which impact on their actual effect on physicians in clinical-decision making. However,
humans. they are still based on oversimplified algorithms;
Overall, despite the biological complexity that their precision depend on a limited number of
animal models provide, the above-mentioned experimental parameters, which threatens the
limitations represent a major drawback when try- possibility to patient-personalized prediction.
ing to elucidate new compounds behavior or to Implementing all the (big) data necessary to
understand molecular mechanisms underlying develop a personalized computational model and
drug response and/or acquired resistance. This is predict patient-specific treatment outcomes
indeed supported by the low percentage of suc- would be univocally associated with the need of
cessful translation of drugs from animal models high computing power. Therefore, a compromise
to clinical trials (Mak et al. 2014). between oversimplification and complexity must
be achieved to bring in silico modelling to the
clinical practice.
4.3.3 In Silico Tumor Models
The large amount and wide range of experimen- 4.4 Organ-on-a-Chip Tumor
tal data produced in the context of cancer have Models: Tumor-on-a-Chip
made in silico models of tumors indispensable.
Mathematical modelling and computer simula- Recent advances in micro- and nanofabrication
tions construct simplified, but representative, technologies and tissue engineering have led to
models of various factors and events involved in the development of miniaturized organ-on-a-chip
the progression of the disease (Vavourakis et al. (OoC) models of human physio-pathological
2017). In silico modelling span a broad spectrum processes. OoC are microfluidic devices contain-
of models, including statistical, multiscale, con- ing human cells, which precisely replicate key
tinuum, discrete, and agent-based (Edelman et al. features and the 3D architecture of human organs
2010). Typically, they are used to complement and tissues. OoC offers multiple advantages
experimental work since they are capable to compared to their in vitro, in vivo, and in silico
extrapolate, or to refine, in vivo and in vitro counterparts. They are 3D, highly physiological,
compatible with mass production, and cheap. topic and references therein (Caballero et al.
They also permit the co-culture of different cell 2017b; Katt et al. 2016).
types and high control over the physical (geome-
try, size), hydro/hemodynamic (flow, shear stress,
gradient), mechanical (stress, rigidity), and bio- 4.4.1 Cancer Invasion
chemical (functionalization) properties of the and Intravasation Models
model. Their design can be customized, or alter-
natively, standardized – commercial – chips can The first event in cancer dissemination is the
be used. Finally, OoC also allow the integration growth of the tumor and the invasion of cancer
of other tumor models, such as tumor organoids, cells. Several cancer invasion ToC models have
and are compatible with traditional and advanced been described in the literature (Chen et al.
microscopy technologies (Caballero et al. 2017b). 2016b; Lei et al. 2016; Zhang et al. 2012). These
Different OoC models of human organs have models differ in the architecture of the chip, the
been described during the last years, including type of tumor, and in particular, the composition
the lung (Huh et al. 2010), spleen (Rigat- and structure of the TME. The latter is a very
Brugarolas et al. 2014), kidney (Kim et al. 2016), dynamic system exposed to a large variety of
brain (Kilic et al. 2016), liver (Lee et al. 2013), mechanical cues, which influence the progres-
gut (Kim et al. 2012), and heart (Grosberg et al. sion of the tumor. This was recently addressed
2011), among many others (Bhatia and Ingber using an orthotopic human lung ToC model,
2014). They all differ in their design and cell con- which recapitulated the growth of lung cancer
tent but display a 3D-like arrangement of cells cells, cytokine secretion profiles, and the clinical
recreating the human scenario. OoC models have responses to therapy previously observed in
widely been employed to investigate pathological patients (Hassell et al. 2017). The model con-
processes, and in particular, tumor growth and tained two channels mimicking the vascular
progression. These OoC models of tumors are endothelium and the pulmonary epithelium,
typically denoted as cancer- or tumor-on-a-chip respectively, which were interconnected through
(ToC). ToC models can easily recreate within a a porous thin membrane. Next, two side cham-
micro-engineered chip the (cancer) cell-cell/ bers were used to apply cyclic stretching mimick-
ECM interactions, the spatiotemporal chemical/ ing physiological breathing within the chip.
physical gradients, or the (hydro)dynamic char- Interestingly, the results showed that the induced
acteristics of the TME. They may be used to bet- cyclic motion inhibited tumor growth and inva-
ter elucidate the efficiency and toxicity of (new) sion, which was mediated by changes in the epi-
drugs in the human body (Caballero et al. 2017a), dermal growth factor receptor and the MET
as well as to explore the mechanistic determi- protein kinase. This work demonstrated the high
nants of the disease in a highly physiological, capabilities of ToC models in recapitulating
low-cost, and high-throughput manner, solving human physiology and in unravelling the mecha-
the limitations of traditional models (Mosig nism at work governing pathological processes.
2017). Different ToC models have been reported During the last few years, intravasation-on-a-
reproducing several of the events occurring dur- chip models have been employed to investigate
ing cancer progression, including tumor growth the invasion of tumor cells into the vascular and
and invasion (Hassell et al. 2017), intra/extrava- lymphatic systems (see Fig. 4.1) (Lee et al. 2014;
sation (Chen et al. 2017; Zervantonakis et al. Zervantonakis et al. 2012). Typically, these mod-
2012), or organ specificity (Xu et al. 2016), and els are based on simple micro-engineered blood
others (see Fig. 4.1). Herein, we briefly discuss vessel-like channels and chambers mimicking
about some recent examples of ToC platforms the TME. These models differ on their complex-
highlighting their main applications, strengths, ity, ranging from simple hollow polydimethylsi-
and limitations. For a detailed description, the loxane (PDMS) channels coated with endothelial
readers are referred to the recent reviews on the cells, to more complex systems where the native
Fig. 4.1 Tumor-on-a-chip models. (a) Scheme showing lymphangiogenesis, (e) intravasation, (f) extravasation,
the different events occurring during cancer metastasis. and (g) organ specificity. Reproduced with permission
ToC models of (b) tumor invasion, (c) angiogenesis, (d) from (Caballero et al. 2017b). Copyright © Elsevier
TME is mimicked employing hydrogel-based endothelial permeability, and intravasation rate

ECMs encapsulating cancerous cells. Recently, a (Zervantonakis et al. 2012).
3D ToC device made of concentric three-layer Overall, even though ToC models of cancer
cell-laden hydrogels reproducing the tumor cells, invasion and intravasation can provide key
the stroma, and the vascular network, was used to insights about their mechanism at work, they
assess the breast cancer cell intravasation need some refinement in terms of throughput and
(Nagaraju et al. 2018). This model showed that compatibility with standard analytical and char-
the breast cancer cells displayed different inva- acterization technologies to boost their clinical
sion capabilities depending on their metastatic and industrial applicability.
capability and on the presence of the vascular
layer. In particular, MDA-MB-231 breast cancer
cells displayed enhanced invasion and intravasa- 4.4.2 Angiogenesis Models
tion in the presence of HUVEC cells, in contrast
to MCF7 cancer cells. Similarly, a 3-channel One of the main events in the cascade of tumor
microfluidic chip containing endothelial and can- dissemination is angiogenesis, which is defined
cer cells interconnected via a 3D ECM hydrogel as the formation of a new vascular network from
was employed to investigate the mechanism of pre-existing blood vessels. Angiogenesis is regu-
tumor cells-endothelium interactions. Results lated by an extensive variety of factors including
showed that the secretion of TNF-α by macro- mechanical and biochemical cues, which trigger
phages increased cancer cells invasion capability, the invasion of cancer cells into the microcircula-
tion system. Many ToC models of tumor angio-
genesis have been reported during the last few work of cancer cell extravasation (see Fig. 4.1)
years. They have provided valuable information (Chen et al. 2016a, 2017; Cui et al. 2017; Ma
about the regulators of cancer cell invasion, et al. 2018). Most of these models are based on
including the important role of fluid flow and chips containing three microfluidic channels sep-
shear stress, the interaction of cells with the arated by micro-pores or slits. Typically, one of
TME, or the influence of pro-angiogenic factors the channels is coated with endothelial cells
in the invasion of endothelial cells (see Fig. 4.1) mimicking the blood vessel; a second channel is
(Bischel et al. 2013; Theberge et al. 2015). As an used to recreate the TME, and typically, a third
example, a ToC model recreating a single – par- channel is employed to flow drugs or to create
ent – human blood vessel was developed to moni- gradients of biochemical compounds. As an
tor vascular sprouting (Pauty et al. 2018). In this example, this strategy was employed to study the
work, vascular endothelial growth factor-A was extravasation of human metastatic breast cancer
added to stimulate the growth of sprouts on the cells towards specific secondary bone-mimicking
vessel, indicators of the nascent formation of microenvironment (Jeon et al. 2015). The side
capillaries and mimicking in vivo sprouting channels were used to flow biochemical factors
angiogenesis. The model, albeit simple, provided and breast cancer cells. The middle channel con-
a good platform to examine the effect of anti- tained a fibrin gel mimicking the bone microenvi-
cancerous drugs. In particular, two Food and ronment and embedding human umbilical vein
Drug Administration (FDA)-approved anti- endothelial cells, human bone marrow-derived
angiogenic drugs (sorafenib and sunitinib) were mesenchymal stem cells, and osteoblast-differen-
used. Interestingly, these drugs inhibited the tiated cells. Importantly, a functional native-like
growth of new vessels even upon vascular endo- microvascular network was generated allowing
thelial growth factor-A stimulation, thus validat- the lateral flow of media across the gel and the
ing the ToC model for drug screening applications. arrest, and extravasation, of MDA-MB-231
Similarly, a cylindrical collagen-based microflu- breast cancer cells. To demonstrate the specificity
idic channel was used to seed endothelial cells of breast cancer cells towards bone, two addi-
that self-organized forming an artificial vessel. In tional microenvironments were generated,
vivo-like angiogenesis was mimicked by inject- including muscle and acellular collagen. Finally,
ing a gradient of different pro-angiogenic factors. the role of physiological fluid shear stress was
Results showed that a combination of growth fac- addressed finding that the extravasation rate was
tors and not an individual one was needed for larger in static condition, most likely a conse-
new vessel formation. Altogether, the above- quence of the observed decrease in microvessel
described approaches show promising clinical permeability, and highlighting the utility of this
applications even though they still need to be approach. Altogether, this work highlights the
proven. potential of ToC models to investigate the mecha-
nistic determinants of extravasation. This may be
of particular interest for the pharmaceutical
4.4.3 Extravasation Models industry to develop and evaluate the efficiency of
new drugs targeting this pathological event.
Extravasation is the event where cancer cells, To summarize, all the above-mentioned works
after invading and circulating along the vascular demonstrate that ToC models can be an extremely
and lymphatic systems, escape out of the vessels useful tool to reproduce the complex events
and invade the neighboring tissue to establish a occurring during tumor dissemination, providing
secondary tumor site. Similar to intravasation, key insights about the mechanism at work of the
this transendothelial invasion is a rare event, disease. This is of high interest for clinical appli-
which is difficult to capture and analyze in vivo. cations, where new simple and reliable tumor
Several extravasation-on-a-chip devices have models that overcome the limitations of current
been reported to investigate the mechanism at in vitro and in vivo models, are needed. However,
ToC models still lack fundamental elements of platforms follow a static configuration, where the
the human physiology, such as organ-organ and individual organ and tissue models are intercon-
tissue-tissue interconnection, which are critical nected following a pre-defined order. Typically, a
for cancer metastasis studies as well as for drug single-route microfluidic channel and a com-
discovery, screening, and toxicology mon – universal – medium are used. In this case,
applications. the flow is unidirectional where only upstream
organs interact with the downstream ones. In
contrast, semi-static systems interconnect several
4.5 Multi-Organ-on-a-Chip organ modules using a microfluidic channel plat-
Models form and Boyden-based chambers containing the
individual organ models (Edington et al. 2018).
Organs and tissues in the human body are not Typically, semi-static systems allow the recircu-
remained in isolation, but are integrated into a lation of culture media mimicking systemic
highly dynamic and interconnected microenvi- blood flow. Finally, flexible multi-OoC models
ronment with a continuous exchange of biochem- interconnect several – individual – OoC devices
ical and mechanical signals. Indeed, an action through microfluidic tubing allowing multiple
targeting one specific organ or tissue, such as the and modular configurations. In this case, the
addition of a drug, may affect others. In this number, type, and interconnection order can be
regard, standard OoC models mimicking indi- freely chosen.
vidual organs and their functional units are inca- Recent works have reported the use of multi-
pable to investigate secondary organ/tissue OoC devices for disease modelling and drug dis-
effects. To address this limitation, highly inte- covery/screening applications (Hwan and Hwan
grated OoC devices combining different organs 2018; Wang et al. 2018) (see Table 4.1); these
and tissues on a single chip have been developed. models differ in their designs and complexities.
These devices have been denoted as multi-organ As an example, a multi-OoC platform was
(or tissue) on-a-chip (multi-OoC) and provide a recently developed comprising the liver, lung,
highly physiological and dynamic microenviron- and heart, and used to study inter-organ response
ment where to recapitulate human organ-organ to drug administration (Skardal et al. 2017). The
and tissue-tissue interactions. Multi-OoC have, model interconnected three independent OoC
therefore, an enormous potential for disease models by means of a central fluid routing bread-
modelling as well as for drug discovery and board in a plug & play configuration, and used a
screening applications. This allows a comprehen- standard peristaltic pump to perfuse the drug-
sive testing in the drug absorption, distribution, containing media. Interestingly, drug response
metabolism, and excretion (ADME) process. was found to depend on inter-organ and tissue
Similarly, they offer large possibilities for the interactions, highlighting the importance of this
prediction of drug efficiency and toxicity to sec- approach.
ondary organs by assessing the passage of a drug Multi-OoC have also been applied to cancer
throughout the human body. research to assess organ specificity in cancer
Multi-OoC can be classified as static, semi- metastasis (Kong et al. 2016; Skardal et al. 2016a;
static, and flexible depending on their configura- Xu et al. 2016). This is critically important
tion (Rogal et al. 2017). It may be noted that because the dissemination of tumor cells from
alternative categories can also be attributed the primary tumor to secondary organs is respon-
depending on their hydrodynamic characteristics, sible for more than 90% of cancer-related deaths
namely static (diffusion-driven), single-pass (Valastyan and Weinberg 2011). Recently, a bone
(open loop), pump-driven (closed-loop), and perivascular niche-on-a-chip device was
pump-less (gravity-driven) recirculating plat- employed to investigate the metastatic coloniza-
forms (Wang et al. 2018). Traditional multi-OoC tion of breast cancer cells to bone (Marturano-
Table 4.1 Recent works on multi-organ and patient-on-a-chip devices

Organs Main
modelled applications Advantages Limitations References
Liver Toxicity Multiple real-time sensors Low throughput Zhang et al. (2017)
Heart integrated Reduced number of
Platform compatible with organs
existing OoC models Large size
Mini-microscope integrated Scaling between organs
not addressed
Liver Toxicity Recirculation of media Reduced number of Kam(ei et al. 2017)
Heart Parallelization organs
Integrated micro-valves Scaling between organs
Common media not addressed
Gut Metastasis Simple No control on tumor Skardal et al.
Liver Drug Customizable structure or growth (2016a)
screening Vascular network mimicking Lack of both endothelial
and stromal cells
No alternative metastatic
sites
Low throughput
Reduced number of
organs
Gut Inflammation Long-term cultures Reduced number of Chen Wen et al.
Liver Pharmaco- Recirculation organs (2017),
kinetics Scaling of organs Low throughput Tsamandouras et al.
(2017)
Lung Drug Programmable micro-pump Reduced number of Coppeta et al. (2017)
Liver discovery integrated organs
Precise fluid flow control Scaling between organs
Reconfigurable not addressed
Non-sorption materials Low throughput
Liver Drug Independent or interconnected Low throughput Skardal et al. (2017)
Heart screening response No pharmaco-kinetic
Lung Toxicology Integrated biosensors studies
Use of 3D organoids Reduced number of
Modular configuration organs
Intestine ADME of Long period co-culture stability Low throughput. Maschmeyer et al.
Liver drugs and tissue functionality Lack of both endothelial (2015)
Skin Reproducible 4-tissue and stromal cells
Kidney homeostasis
Dynamic perfusion
Cardiac Toxicology Multiple organs and drugs tested Important organs for Oleaga et al. (2016)
Muscle Drug Measurement of electrical, ADME applications
Neuronal discovery mechanical, and metabolic missing
Liver response Scaling between organs
Modular, reconfigurable and not addressed
pumpless system For some organs, cell
Common serum-free medium lines were used
Lung Metastasis Multiple metastatic sites No assessment of shear Xu et al. (2016)
Brain Multiple features of the in vivo stress and vascular
Bone lung TME, μ-vascular targeting for drug
Liver endothelium, immune cells, delivery
fibroblasts and distant organs Low throughput
(continued)
Table 4.1 (continued)
Organs Main
modelled applications Advantages Limitations References
GI tract ADME of Multiple organs Complex Miller and Shuler
Lung drugs. Pumpless (gravity-driven) Use of cell lines (2016)
Liver Scalable Low term culture
Bone Scaling arguments considered
marrow
Kidney
Intestine ADME of Multiple organs Complex Vernetti et al. (2017)
Liver drugs. In vivo-like response Low throughput
Kidney Drug Dynamic perfusion Scaling between organs
proximal organ- not addressed
tubule specificity. No vascularization
Blood- Toxicology
brain
barrier
Skeletal
muscle
Kruik et al. 2018). The chip incorporated a native It may be noted that the above-mentioned
decellularised 3D bone matrix embedding human works albeit informative, mainly address the
endothelial and bone marrow-derived mesenchy- metastasis of cancer cells into a specific second-
mal stem cells. This in vivo-like composition ary tissues, limiting the understanding of the key
allowed the formation of a stable microvascula- players involved in organs specificity during
ture in controlled flow conditions. MDA-MB-231 metastasis. To address this issue, multi-OoC
breast cancer cells were introduced into the devices displaying multiple metastatic sites have
device both under static and interstitial flow con- also been reported. As an example, a multi-organ
ditions. Surprisingly, results showed a higher lung-on-a-chip device was recently employed to
four-fold proliferation rate of cancer cells in replicate the specificity of metastasis from the
static conditions, suggesting that physiologically- lung to the brain, bone, and liver (Xu et al. 2016).
relevant interstitial flow activated a slow-prolifer- The model contained the upstream native-like
ative state in cancer cells colonizing the niche. lung cancer microenvironment, the microvascu-
Next, in a similar work, the specificity of human lar endothelium, immune cells, fibroblasts, and
breast cancer cells to bone was studied using a the three downstream organs. Results showed
3-channel multi-OoC device (Bersini et al. 2014). that the cells from the respective metastatic sites
In this work, MDA-MB-231 human breast cancer overexpressed specific proteins characteristic of
cells, human umbilical vascular endothelial cells, metastasis, and importantly, similar to the in vivo
and human osteo-differentiated bone marrow- scenario.
derived mesenchymal stem cells, were employed Overall, the advanced capabilities of multi-
to replicate the invasive malignant tumor cells, OoC platforms provide them with unique capa-
the vascular network, and the metastatic site of bilities for the development of highly predictive
bones, respectively. Results showed that highly cancer models, where inter-organs and tissues
metastatic breast cancer cells introduced in the interactions can be easily assessed. Therefore,
microvascular network extravasated in larger multi-OoC may find solid applications in drug
amounts towards the pre-conditioned bone discovery, screening, and toxicology, with an
microenvironment compared to control enormous potential in the field of personalized
conditions. cancer medicine (Esch et al. 2014). However,
improvements must also be addressed in using the model. Typically, for drug delivery and
primary cells (or iPS-derived cells) and in inte- screening studies, the model may include the gut,
grating personalized biomimetic 3D microenvi- liver, kidney, blood vessel, and the targeted organ
ronments if these models are intended to be of interest. Additional organs may also be
adopted by the clinical market. included to assess off-target side effect of drugs,
such as the heart, to investigate cardiotoxicity, or
the brain to examine blood-brain barrier (BBB)
4.6 Multi-Organ-on-a-Chip permeability.
Models of Cancer Patients: A special category of patient-on-a-chip mod-
Cancer Patient-on-a-Chip els is cancer patient-on-a-chip (CPoC), where
cells from cancer patients are used. These models
The combination of multi-OoC technology and display a tremendous potential in personalized
tissue engineering tools with primary cells from medicine by reproducing in a multi-OoC the
patients, has led to the development of clinical- tumor of an individual patient. This may permit
oriented patient-on-a-chip models (Abaci and to (i) unravel the specific mechanism at work of
Shuler 2015; Hampton 2017; Low and Tagle patient tumor dissemination and organ-specific-
2018; Skardal et al. 2016b; Williamson et al. ity during metastasis, (ii) reproduce organ-organ
2013) (see Fig. 4.2). In these models, the differ- and tissue-tissue interactions, (iii) provide real
ent organs and tissues are interconnected follow- time data about treatment efficiency and side
ing a physiologically-relevant arrangement to effects, (iv) assess the ADME and toxicity of
mimic the systemic organization and diversity of chemo- and immune-therapeutic drugs, or radio-
events occurring inside the human body (see therapeutics treatments, (v) investigate drug
Fig. 4.2). The number of organs required to pharmacokinetics and pharmacodynamics, and
develop a patient-on-a-chip are not fixed, which (vi) test, and predict, personalized therapies in a
may depend on the biological characteristics of high-throughput and controlled microenviron-
each individual patient and the clinical goal of ment (Zhang et al. 2017). Importantly, since most
Fig. 4.2 The patient-on-a-chip concept. Schemes and excretion. In this case, drug administration can be per-
showing (a) several organs in the human body, and (b) the formed through the lung (inhaled) or through the gut
patient-on-a-chip model representing some of them, (orally administrated) to study their metabolism in the
namely the lung, gut, liver, kidney, bone, and heart. The liver, their transport and clearance in the kidney, their
different organs are interconnected following a immune response in e.g., the bone, and their toxicity in
physiologically-
relevant configuration to account for e.g., the heart. Panel (b) is reproduced with permission
organ/tissue-organ/tissue interactions capable to recapitu- from (Huh et al. 2011). Copyright © Elsevier
late systemic drug absorption, distribution, metabolism,
cancer patient fail to respond equally to the same interconnecting functional heart, muscle, brain,
treatment, this precision medicine approach may and liver models was used to assess the systemic
also help physicians in taking decisions on ther- toxicity of five drugs with known side effects
apy selection and identify treatment specific during long-term time periods and continuous
regimes or unexpected risk factors. Finally, CPoC recirculation flows (see Fig. 4.3b and Table 4.1)
may also be employed to model genetic variation (Oleaga et al. 2016). Interestingly, the model
or to identify population sub-groups that display integrated metabolic, electrical, and mechanical
a higher risk from adverse side effects. However, readouts to analyze organ functionality in real
and despite the enormous potential of CPoC, the time. In general, the obtained results were consis-
amount of truly CPoC models reported in the lit- tent with those reported in the literature, namely
erature are scarce. This is due to the technical and a decrease in cell functionality and viability upon
biological complexity required, which makes that drug treatment. In particular, doxorubicin, a well-
most of the reported works are based on tradi- known chemotherapeutic drug, showed acute
tional multi-OoC platforms (see Table 4.1). cardio- and hepatotoxicity. Doxorubicin cardio-
One of the main advantages of CPoC models toxicity was also reported in a different work by
is their ability to investigate fundamental aspects using a simple CPoC model, which intercon-
of the human – cancer – physiology by linking nected a heart- and a cancer-liver-on-a-chip (see
multiple organs and tissues together, such as the Fig. 4.3c and Table 4.1) (Kamei et al. 2017).
sequential metabolism of individual (or combina- Results showed that cardiotoxicity was a result of
tions of) drug administration in the human body the metabolic production of doxorubicinol by
(Ronaldson-Bouchard and Vunjak-Novakovic HepG2 cells. Similarly, the interconnection of
2018). This may provide valuable information heart, liver, and lung was used to assess the toxic-
about their toxicities and pharmacokinetics, ity and organ specificity of bleomycin, a drug
which is critical during the process of drug devel- typically used for treating some cancers (see
opment, thus increasing the efficiency of the ther- Fig. 4.3d and Table 4.1) (Skardal et al. 2017). The
apy and reducing the risk of drug toxicity. addition of the drug exhibited an unanticipated
Recently, a CPoC model was developed by cardiotoxicity, where cardiac cells ceased beat-
sequentially interconnecting several organ mod- ing. Surprisingly, the addition of the drug using
els, namely the intestine, liver, kidney proximal an individual heart-on-a-chip model did not show
tubule, BBB, and skeletal muscle, following an in any effect; cells did not stop beating. This sug-
vivo-like configuration (see Fig. 4.3a and gested that bleomycin might be metabolized in
Table 4.1) (Vernetti et al. 2017). The model was the liver and produce secondary toxic factors to
employed to assess organ-specific response upon cardiac cells. Unfortunately, the device did not
compound administration, where the intestine, integrate sensing capabilities to identify and
liver, and kidney represented the organs involved quantify such toxic compounds. This was
in drug absorption, metabolism, and excretion, addressed in a similar work using a fully-autom-
respectively. The BBB and the skeletal muscle atized CPoC platform integrating physical, elec-
were added due to their importance in neurovas- trochemical, and optical biosensors into a human
cular drug administration and as a potential off- heart-and-liver-on-a-chip device (see Fig. 4.3e
target organ to examine metabolism-based and Table 4.1) (Zhang et al. 2017). Again, the
toxicity. The effect on specific organs of three CPoC model was used to assess the cardiotoxic-
known compounds, namely terfenadine, trimeth- ity of common chemotherapeutic drugs towards
ylamine-N-oxide, and vitamin D3, was investi- hepatocellular carcinoma cells. It was found that
gated. Results showed that the compounds were liver cancer organoids shrunk upon drug treat-
successfully absorbed, metabolized, and trans- ment as demonstrated by the measured reduction
ported, and importantly, in agreement with in the amount of secreted albumin and in the
reported clinical data. Similarly, a CPoC approach increase of characteristic hepatotoxic biomark-
Fig. 4.3 Cancer patient-on-a-chip models. (a) Five- actual chip containing three sets of co-culture systems
organ CPoC (intestine, liver, kidney proximal tubule, (Kamei et al. 2017) © Royal Society of Chemistry. (d)
BBB, and skeletal muscle) used to assess organ-specific Three-organ CPoC model (heart, liver, and lung) used to
response upon compound administration. The scheme assess the toxicity and organ specificity of bleomycin. The
shows four of the organs systems used for functional cou- top and bottom images show a scheme and an actual pho-
pling, A) intestine, B) liver, C) kidney, and D) BBB tograph of the entire system set up, respectively (Skardal
(Vernetti et al. 2017) © Nature Publishing Group. (b) et al. 2017) © Nature Publishing Group. (e) Two-organ
Four-organ CPoC (heart, muscle, brain, and liver) model CPoC model (heart and liver) used to assess the cardiotox-
used to assess the systemic toxicity of five drugs with icity of common chemotherapeutic drugs. The scheme
known side effects. The scheme shows the different cell shows the entire system set up, and the inset, an actual
compartments (Oleaga et al. 2016) © Nature Publishing picture of the two-organs model (Zhang et al. 2017) ©
Group. (c) Two-organ CPoC model (heart and liver) used National Academy of Sciences. Figures are reproduced
to assess doxorubicin cardiotoxicity. The image shows the with the permission from the references
ers. Similarly, the addition of doxorubicin Overall, CPoC represent a new generation of
induced pronounced cardiotoxicity as demon- patient-matched models capable to evaluate and
strated by the high levels of CK-MB, a cardiac predict the outcome of individualized therapies.
biomarker characteristic of heart damage. This is a consequence of their unique ability to
replicate into a microfluidic platform the biologi-
cal multi-organ complexity of individual patients (CTCs), DNA/RNA, exosomes or antibodies

and their tumors. Similarly, CPoC models offer a emerge as potential biomarkers of cancer metas-
great potential for drug discovery, screening, and tasis with high clinical value. This liquid biopsy
toxicology applications, where biotechnological approach has led to the development of liquid
and pharmaceutical companies can develop drugs biopsy-on-a-chip devices for the individualized
in a more efficient and fast manner. This will uni- diagnosis, prognosis, and optimized treatments
vocally reduce the current high attrition rates of on cancer patients (Tadimety et al. 2017). Indeed,
drugs. several CPoC-based systems have already been
reported for the detection of CTCs from different
tumor origins, which highlights the enormous
4.7 Clinical and Industrial potential of these platforms for industrial and
Applications: Current clinical applications (Alix-Panabieres and Pantel
Challenges and New 2014; Ferreira et al. 2016; Warkiani et al. 2016).
Directions Unfortunately, despite the latest progress in
the development of ToC, (multi)-OoC, and CPoC
Besides fundamental research, where ToC, models, their clinical use is still very limited. The
(multi)-OoC, and CPoC have been widely future generation of these type of models may
employed in cancer research applications, the use need to display advanced properties and be com-
of these models by the industry is just beginning. patible with current and modern imaging, screen-
Pharmaceutical companies, such as Pfizer™, ing, and biosensing technologies. Besides the
Roche™, Merck™, AstraZeneca™, and Johnson technical and biological challenges related to the
& Johnson™, have started to employ this tech- interconnection of different organs and tissues,
nology to accelerate the screening and develop- the development of CPoC-based devices is also
ment of novel therapeutic agents. In addition, accompanied with considerable number of eco-
several OoC-based companies have started to nomic barriers, which restrict their incorporation
receive public funds to explore the clinical and into the pharmaceutical, biotechnological, and
industrial capabilities of their technology. As an clinical market. A recent set of articles reported
example, 45 M$ have recently been invested in on the challenges and requirements of employing
Emulate™ (US) to boost the application of OoC this type of models in drug development and
technology for drug discovery. The chips have mechanistic studies, including the important
been employed by the FDA to assess the toxicol- visions of researchers, clinicians, industry, and
ogy of food supplements and additives policy makers (Watson et al. 2017). The topics
(Fitzpatrick 2017). Interestingly, collaborations discussed included the optimal volume of sam-
with hospitals have also been initiated to estab- ples to be employed, scaling arguments, model
lish Patient-on-a-Chip clinical programs aiming cost, and market needs, among others, highlight-
at developing personalized CPoC-based models ing the challenges to be considered when devel-
of patient diseases using reprogrammed stem oping predictable CPoC-based models to obtain a
cells from patients. systemic-like response. In the following, we
Many OoC companies have emerged during briefly discuss about some of these challenges.
the last few years offering their chips for a diverse For a detailed information, the readers are
varieties of biomedical applications, including referred to specialized reviews on the topic.
pharmacological and toxicology studies, drug
discovery, and others. Other industrial and clini-
cal applications of CPoC include the screening of 4.7.1 Technological Challenges
cancer biomarkers to assess the properties of
metastatic tumor. In this regard, the identification Technological challenges can be categorized as
in peripheral blood of circulating tumor cells basic and complex. The former include standard
bubble formation or sterility maintenance. The 4.7.2 Biological Challenges

need for new non-sorptive materials for chip fab-
rication instead of standard PDMS, or the lack of Biological challenges include the integration of a
an integrated perfusion system capable to provide TME with controlled mechanochemical and bio-
physiological fluid flows for the different tissues logical properties, where the cells are capable to
(speed and shear stress) to account for the recapitulate the phenotype observed in vivo over
mechanical stimuli/response of cells, are also long culture periods. In this regard, the culture of
included in this category. Next, new biomaterials different cells types, such as tumor cells, endo-
mimicking better the native environment will thelial, macrophages, or fibroblasts, is still a
need to be created. In this regard, the use of cells major limitation in most models. In addition,
obtained from patients capable to build the TME keeping the cells and tissues alive for long-term
are anticipated to be a major revolution. Complex periods (>1 month) in physiological-like condi-
challenges include the lack of standardization, tions, together with a systemic liquid-to-cell vol-
which limits the direct comparison of the data ume ratio, is a major challenge. Similarly,
between laboratories. To address this, several gathering a sufficient number of mature cells
generic microfluidic devices have been devel- from patients to seed inside the chips, as well as
oped and are commercially available (Junaid the time required to build a functional CPoC, are
et al. 2017). Alternatively, the use of modular, major bottlenecks. The use of iPSCs may provide
plug & play, pre-fabricated, and universal unlimited number of cells and the creation of spe-
Lego™-like microfluidic units may standardize cific patient and disease-specific cells (Singh
the fabrication of CPoC-based systems facilitat- et al. 2015). Next, organ/tissue functionality must
ing the comparison of results between labs be ensured by supplying the different organs and
(Bhargava et al. 2014; Loskill et al. 2015). Each tissues with specific culture media, or alterna-
individual unit may mimic a specific body func- tively, a universal one. Further, medium recircu-
tion, and integrate technical, sensing, and pump- lation and functional vascularization is crucial
ing functions. Next, CPoC-based models are for linking the different organs and tissues to the
more complex compared to static models and, in vasculature as well as for mimicking systemic
general, cells are more difficult to collect for drug administration (Osaki et al. 2018).
molecular analysis. Other technical challenges Importantly, scaling arguments must be consid-
include the development of models compatible ered. This accounts for designing the different
with traditional or advanced imaging (e.g., super- tissues and organs, which are physiologically
resolution microscopy), screening (e.g., multi- scaled relatively to each other as well as in terms
OMICS), and biosensing technologies (e.g., of tissue-media ratios. Next, the role of dormant
single-molecule biosensing). The latter is of cancer cells or cancer stem cells in resuming
upmost importance for monitoring physiological tumor proliferation must be assessed to analyze
parameters (e.g. O2, H2O2, pO2, pH, and others) chemotherapeutic resistance in cancer patients.
in real-time. Similarly, CPoC-based systems In this regard, primary cells from cancer patients
must also be compatible with automatized (robot- (or even reprogrammed iPSCs from adult somatic
ized) platforms for an efficient and fast handling. cells) may be used instead of standard cell lines,
However, all these technological improvements which behavior may differ between labs as well
may be associated with an increase in complex- as their response to drugs compared to in vivo
ity, which may limit the reproducibility of results scenario. Even more complex, having (and dem-
and manipulation speed/simplicity. Therefore, onstrating) that the tissue(s)/organ(s) model dis-
future CPoC may need to find a balance between play a gene expression pattern characteristic of
technological simplicity, biological relevance, the in vivo scenario is challenging, but funda-
and adaptability. mental for clinical applications. Similarly, the
integration of several and personalized organs alternative physico-chemical properties, such as

and tissues on chip is critical to assess for meta- synthetic or natural-derived hydrogels, and fabri-
static-site influence. This is of upmost impor- cation methods, such as 3D (bio)printing, offer
tance for the development of personalized an enormous potential for the low-cost fabrica-
treatments since not all cancer patients respond tion of organ-on-a-chip devices (Caballero et al.
equally to a specific treatment. Next, the integra- 2017a), thus offering better market opportunities.
tion of human microvasculature, containing both Finally, important market limitations include the
the blood and lymphatic vasculature is crucial to lack of specialized industry and personnel capa-
elucidate the mechanistic determinants of tumors ble to manipulate the chips and interpret the
dissemination, to investigate how drugs transit results, which makes their adoption by the clini-
along the vasculature and are administrated to the cal market in the near future unrealistic.
TME, and to assess how the tumor respond to Overall, the current level of development of
drugs injected into the “human” vasculature. CPoC-based devices, and the technological, bio-
Finally, the impact of toxic compounds exchange logical and economic barriers that they still chal-
between organs must be explored (Frey et al. lenge, limit their current use into the clinical
2014). This is highly important for investigating practice. The collaboration between different
the efficiency, toxicity, and side effects of drugs communities, namely academia, clinics, and
in secondary organs. This means that the number industry, as well as the involvement of policy
of organs and tissue models shall be as high as makers, research organizations, societies,
possible. However, an increase in the number of patients, and other stakeholders, is fundamental
integrated organ/tissue models is associated with to boost their development and use in the clinics.
a significant increase in the complexity related to However, the future looks promising and several
chip fabrication, biological content, handling, European and International-funded organizations
and cost, among others. and initiatives have been created to consolidate
personalized and patient-centred technologies in
healthcare treatment.
4.7.3 Economical and Market
Challenges
4.8 Conclusions
Economical and market barriers include the high
manufacturing cost of typical CPoC systems. A Organ-on-a-chip models of cancer patients, or
decrease in the fabrication cost must be accom- cancer patient-on-a-chip, provide a large plethora
plished to facilitate their adoption by the pharma- of advantages compared to their in vitro, in vivo,
ceutical, biotechnological, and clinical market. and in silico counterparts. This new generation of
To this aim, the chips must be simple, miniatur- highly personalized cancer models displays bio-
ized, mass produced, and integrate bioelectronic mimetic compositions and 3D architectures capa-
components for sample monitoring and screen- ble to reproduce in a single microfluidic chip the
ing. Next, clinical-oriented CPoC models will biological, structural, and dynamical complexity
need to be developed under GMP (good manu- of human organs and tissues, and their interac-
facturing practice) conditions using biomaterials tions, of individual cancer patients. Indeed, can-
(and cells) displaying rigid requirements in terms cer-patient-on-a-chip (CPoC) models are a viable
of cost, biocompatibility, and durability. In this physiologically-relevant platform for personal-
regard, PDMS, typically used for the fabrication ized medicine, drug discovery and screening, as
of chips due to its intrinsic properties (e.g. optical well as to explore the mechanistic determinants
transparency, rigidity and biocompatibility), is of cancer progression. By using CPoC models,
not compatible with mass industrial and commer- patient-matched therapies can be tested by physi-
cial production for its elevated cost and fabrica- cians to select the most adequate treatment for
tion methodology. Other materials displaying each individual patient. Importantly, by combin-
ing different organs and tissues models, the – Bischel LL, Young EWK, Mader BR, Beebe DJ (2013)
Tubeless microfluidic angiogenesis assay with
metabolic – interaction between different organs three-dimensional endothelial-lined microvessels.
can be modelled and investigated. This is of par- Biomaterials 34:1471–1477. https://doi.org/10.1016/j.
ticular importance for metastasis studies and to biomaterials.2012.11.005
assess the side effect of drugs into secondary Brancato V et al (2017) Bioengineered tumoral micro-
tissues recapitulate desmoplastic reaction of pancre-
organs. Altogether, CPoC are expected to replace atic cancer. Acta Biomater 49:152–166. https://doi.
standard in vitro static assays, bridging the gap org/10.1016/j.actbio.2016.11.072
between pre-clinical (animal) models and Breslin S, O’Driscoll L (2016) The relevance of using 3D
humans. To become this a reality, an active col- cell cultures, in addition to 2D monolayer cultures,
when evaluating breast cancer drug sensitivity and
laboration between the different stakeholders is resistance. Oncotarget 7:45745–45756. https://doi.
fundamental to boost their translation finally org/10.18632/oncotarget.9935
towards the biotechnological and clinical Breyer MD, Look AT, Cifra A (2015) From bench
market. to patient: model systems in drug discovery. Dis
Model Mech 8:1171–1174. https://doi.org/10.1242/
dmm.023036
Acknowledgements D.C. acknowledges the financial Caballero D, Blackburn SM, de Pablo M, Samitier J,
support from the Portuguese Foundation for Science and Albertazzi L (2017a) Tumour-vessel-on-a-chip mod-
Technology (FCT) under the program CEEC Individual els for drug delivery. Lab Chip 17:3760–3771. https://
2017 (CEECIND/00352/2017). All the authors acknowl- doi.org/10.1039/c7lc00574a
edge the financial support from the European Union Caballero D, Kaushik S, Correlo VM, Oliveira JM, Reis
Framework Programme for Research and Innovation RL, Kundu SC (2017b) Organ-on-chip models of can-
Horizon 2020 on Forefront Research in 3D Disease cer metastasis for future personalized medicine: from
Cancer Models as in vitro Screening Technologies chip to the patient. Biomaterials 149:98–115. https://
(FoReCaST) under grant agreement no. 668983. D.C. and doi.org/10.1016/j.biomaterials.2017.10.005
S.C.K also acknowledge the support from the FCT under Carvalho MR, Lima D, Reis RL, Oliveira JM, Correlo VM
the scope of the project Modelling Cancer Metastasis into (2017) Anti-cancer drug validation: the contribution of
the Human Microcirculation System using a Multiorgan- tissue engineered models. Stem Cell Rev Rep 13:347–
on-a-Chip Approach (2MATCH) (02/SAICT/2017 – n° 363. https://doi.org/10.1007/s12015-017-9720-x
028070) funded by the Programa Operacional Regional Chen Wen LK et al (2017) Integrated gut/liver microphys-
do Norte supported by FEDER. Conflicts of interest: iological systems elucidates inflammatory inter-tissue
none. crosstalk. Biotechnol Bioeng 114:2648–2659. https://
doi.org/10.1002/bit.26370
Chen MB, Lamar JM, Li R, Hynes RO, Kamm RD (2016a)
Elucidation of the roles of tumor integrin β1 in the
References extravasation stage of the metastasis cascade. Cancer
Res 76:2513–2524. https://doi.org/10.1158/0008-
Abaci HE, Shuler ML (2015) Human-on-a-chip design 5472.can-15-1325
strategies and principles for physiologically based Chen Z-z et al (2016b) Establishment of a gastric cancer
pharmocokinetics/pharmacodynamics modeling. subline with high metastatic potential using a novel
Integr Biol 7:383–391. https://doi.org/10.1039/ microfluidic system. Sci Rep 6:38376. https://doi.
c4ib00292j org/10.1038/srep38376
Alix-Panabieres C, Pantel K (2014) Technologies Chen M, Whisler J, Fröse J, Yu C, Shin Y, Kamm R (2017)
for detection of circulating tumor cells: facts and On-chip human microvasculature assay for visualiza-
vision. Lab Chip 14:57–62. https://doi.org/10.1039/ tion and quantification of tumor cell extravasation
c3lc50644d dynamics. Nat Protoc 12:865–880
Bersini S et al (2014) A microfluidic 3D in vitro model Coppeta JR et al (2017) A portable and reconfigurable
for specificity of breast cancer metastasis to bone. multi-organ platform for drug development with
Biomaterials 35:2454–2461. https://doi.org/10.1016/j. onboard microfluidic flow control. Lab Chip 17:134–
biomaterials.2013.11.050 144. https://doi.org/10.1039/c6lc01236a
Bhargava KC, Thompson B, Malmstadt N (2014) Discrete Cui X, Guo W, Sun Y, Sun B, Hu S, Sun D, Lam RHW
elements for 3D microfluidics. Proc Natl Acad Sci (2017) A microfluidic device for isolation and char-
U S A 111:15013–15018. https://doi.org/10.1073/ acterization of transendothelial migrating can-
pnas.1414764111 cer cells. Biomicrofluidics 11:014105. https://doi.
Bhatia SN, Ingber DE (2014) Microfluidic organs-on- org/10.1063/1.4974012
chips. Nat Biotechnol 32:760. https://doi.org/10.1038/ Cukierman E, Pankov R, Stevens DR, Yamada KM (2001)
nbt.2989 Taking cell-matrix adhesions to the third dimension.
Science 294:1708–1712. https://doi.org/10.1126/ Huh D, Hamilton GA, Ingber DE (2011) From 3D cell
science.1064829 culture to organs-on-chips. Trends Cell Biol 21:745–
DiMasi JA, Grabowski HG, Hansen RW (2016) 754. https://doi.org/10.1016/j.tcb.2011.09.005
Innovation in the pharmaceutical industry: new esti- Hwan LS, Hwan SJ (2018) Organ-on-a-chip technol-
mates of R&D costs. J Health Econ 47:20–33. https:// ogy for reproducing multiorgan physiology. Adv
doi.org/10.1016/j.jhealeco.2016.01.012 Healthc Mater 7:1700419. https://doi.org/10.1002/
Doxzen K et al (2013) Guidance of collective cell migra- adhm.201700419
tion by substrate geometry. Integr Biol 5:1026–1035. Jeon JS, Bersini S, Gilardi M, Dubini G, Charest JL,
https://doi.org/10.1039/c3ib40054a Moretti M, Kamm RD (2015) Human 3D vascularized
Doyle AD, Wang FW, Matsumoto K, Yamada KM (2009) organotypic microfluidic assays to study breast cancer
One-dimensional topography underlies three-dimen- cell extravasation. Proc Natl Acad Sci U S A 112:214–
sional fibrillar cell migration. J Cell Biol 184:481– 219. https://doi.org/10.1073/pnas.1417115112
490. https://doi.org/10.1083/jcb.200810041 Junaid A, Mashaghi A, Hankemeier T, Vulto P (2017)
Edelman LB, Eddy JA, Price ND (2010) In silico models An end-user perspective on organ-on-a-chip: assays
of cancer Wiley interdisciplinary. Rev Syst Biol Med and usability aspects Curr Op. Biomed Eng 1:15–22.
2:438–459. https://doi.org/10.1002/wsbm.75 https://doi.org/10.1016/j.cobme.2017.02.002
Edington CD et al (2018) Interconnected microphysiolog- Kamei K-i et al (2017) Integrated heart/cancer on a chip to
ical systems for quantitative biology and pharmacol- reproduce the side effects of anti-cancer drugs in vitro.
ogy studies. Sci Rep 8:4530. https://doi.org/10.1038/ RSC Adv 7:36777–36786. https://doi.org/10.1039/
s41598-018-22749-0 c7ra07716e
Esch MB, Smith AS, Prot JM, Oleaga C, Hickman JJ, Katt ME, Placone AL, Wong AD, Xu ZS, Searson PC
Shuler ML (2014) How multi-organ micro-devices (2016) In vitro tumor models: advantages, disad-
can help foster drug development. Adv Drug Deliv vantages, variables, and selecting the right platform.
Rev 69–70:158–169 Front Bioeng Biotechnol 4. https://doi.org/10.3389/
Esch EW, Bahinski A, Huh D (2015) Organs-on-chips at fbioe.2016.00012
the frontiers of drug discovery. Nat Rev Drug Discov Kilic O et al (2016) Brain-on-a-chip model enables analy-
14:248. https://doi.org/10.1038/nrd4539 sis of human neuronal differentiation and chemotaxis.
Ferreira MM, Ramani VC, Jeffrey SS (2016) Circulating Lab Chip 16:4152–4162. https://doi.org/10.1039/
tumor cell technologies. Mol Oncol 10:374–394. c6lc00946h
https://doi.org/10.1016/j.molonc.2016.01.007 Kim HJ, Huh D, Hamilton G, Ingber DE (2012) Human
Fitzpatrick S (2017) ‘Organs-on-chips’ technology: FDA gut-on-a-chip inhabited by microbial flora that expe-
testing groundbreaking science. U.S. Food & Drug riences intestinal peristalsis-like motions and flow.
Administration, Silver Spring Lab Chip 12:2165–2174. https://doi.org/10.1039/
Frey O, Misun PM, Fluri DA, Hengstler JG, Hierlemann c2lc40074j
A (2014) Reconfigurable microfluidic hanging Kim S, LesherPerez S, Kim B, Yamanishi C, Labuz J,
drop network for multitissue interaction and analy- Leung B, Takayama S (2016) Pharmacokinetic pro-
sis. Nat Commun 30:4250. https://doi.org/10.1038/ file that reduces nephrotoxicity of gentamicin in a
ncomms5250 perfused kidney-on-a-chip. Biofabrication 8:015021.
Grosberg A, Alford PW, McCain ML, Parker KK (2011) https://doi.org/10.1088/1758-5090/8/1/015021
Ensembles of engineered cardiac tissues for physi- Kong J et al (2016) A novel microfluidic model can mimic
ological and pharmacological study: heart on a chip. organ-specific metastasis of circulating tumor cells.
Lab Chip 11:4165–4173. https://doi.org/10.1039/ Oncotarget 7:78421–78432. https://doi.org/10.18632/
c1lc20557a oncotarget.9382.
Hampton T (2017) Body-on-a-chip reveals multitissue Lee S-A, No DY, Kang E, Ju J, Kim D-S, Lee S-H (2013)
interactions during drug exposure. JAMA 318:2069– Spheroid-based three-dimensional liver-on-a-chip to
2070. https://doi.org/10.1001/jama.2017.17869 investigate hepatocyte-hepatic stellate cell interactions
Hassell BA, Goyal G, Lee E, Sontheimer-Phelps A, Levy and flow effects. Lab Chip 13:3529–3537. https://doi.
O, Chen CS, Ingber DE (2017) Human organ chip org/10.1039/c3lc50197c
models recapitulate orthotopic lung cancer growth, Lee H, Park W, Ryu H, Jeon NL (2014) A microfluidic
therapeutic responses, and tumor dormancy in vitro. platform for quantitative analysis of cancer angiogen-
Cell Rep 21:508–516. https://doi.org/10.1016/j. esis and intravasation. Biomicrofluidics 8:054102.
celrep.2017.09.043 https://doi.org/10.1063/1.4894595
Hay M, Thomas DW, Craighead JL, Economides C, Lei KF, Tseng H-P, Lee C-Y, Tsang N-M (2016)
Rosenthal J (2014) Clinical development success rates Quantitative study of cell invasion process under
for investigational drugs. Nat Biotechnol 32:40–51. extracellular stimulation of cytokine in a microflu-
https://doi.org/10.1038/nbt.2786 idic device. Sci Rep 6:25557. https://doi.org/10.1038/
Huh D, Matthews BD, Mammoto A, Montoya-Zavala M, srep25557
Hsin HY, Ingber DE (2010) Reconstituting organ-level Loskill P, Marcus SG, Mathur A, Reese WM, Healy KE
lung functions on a chip. Science 328:1662–1668. (2015) μOrgano: a lego®-like plug & play system for
https://doi.org/10.1126/science.1188302
modular multi-organ-chips. PLoS One 10:e0139587. Lab Chip 14:1715–1724. https://doi.org/10.1039/

https://doi.org/10.1371/journal.pone.0139587 c3lc51449h
Low LA, Tagle DA (2018) ‘You-on-a-chip’ for precision Rogal J, Probst C, Loskill P (2017) Integration concepts
medicine. Expert Rev Precis Med Drug Dev 3:137– for multi-organ chips: how to maintain flexibility?!
146. https://doi.org/10.1080/23808993.2018.1456333 Future Sci OA 3:FSO180. https://doi.org/10.4155/
Ma Y-HV, Middleton K, You L, Sun Y (2018) A review fsoa-2016-0092
of microfluidic approaches for investigating cancer Ronaldson-Bouchard K, Vunjak-Novakovic G (2018)
extravasation during metastasis. Microsys Nanoeng Organs-on-a-chip: a fast track for engineered human
4:17104. https://doi.org/10.1038/micronano.2017.104 tissues in drug development. Cell Stem Cell 22:310–
Mak IWY, Evaniew N, Ghert M (2014) Lost in transla- 324. https://doi.org/10.1016/j.stem.2018.02.011
tion: animal models and clinical trials in cancer treat- Seruga B, Ocana A, Amir E, Tannock IF (2015) Failures
ment. Am J Transl Res 6:114–118 in phase III: causes and consequences. Clin Cancer
Marturano-Kruik A et al (2018) Human bone perivascular Res 21:4552–4560. https://doi.org/10.1158/1078-
niche-on-a-chip for studying metastatic colonization. 0432.ccr-15-0124
Proc Natl Acad Sci U S A. https://doi.org/10.1073/ Sharma VP, Beaty BT, Cox D, Condeelis JS, Eddy RJ
pnas.1714282115 (2014) An in vitro one-dimensional assay to study
Maschmeyer I et al (2015) A four-organ-chip for intercon- growth factor-regulated tumor cell–macrophage inter-
nected long-term co-culture of human intestine, liver, action. Methods Mol Biol Clifton, NJ 1172:115–123.
skin and kidney equivalents. Lab Chip 15:2688–2699. https://doi.org/10.1007/978-1-4939-0928-5_10
https://doi.org/10.1039/c5lc00392j Shu Fan Z, Singaram G, Yuan Hao X, Sally KYT, Alice
Miller PG, Shuler ML (2016) Design and demonstration STW, Stella WP, Yun Wah L (2017) Substrates with
of a pumpless 14 compartment microphysiological patterned topography reveal metastasis of human
system. Biotechnol Bioeng 113:2213–2227. https:// cancer cells. Biomed Mater 12:055001. https://doi.
doi.org/10.1002/bit.25989 org/10.1088/1748-605X/aa785d
Mosig AS (2017) Organ-on-chip models: new opportuni- Singh VK, Kalsan M, Kumar N, Saini A, Chandra R
ties for biomedical research. Future Sci OA 3:FSO130. (2015) Induced pluripotent stem cells: applications
https://doi.org/10.4155/fsoa-2016-0038 in regenerative medicine, disease modeling, and
Mullard A (2016) Parsing clinical success rates. Nat drug discovery. Front Cell Dev Biol 3:2. https://doi.
Rev Drug Discov 15:447. https://doi.org/10.1038/ org/10.3389/fcell.2015.00002
nrd.2016.136 Siramshetty VB, Nickel J, Omieczynski C, Gohlke B-O,
Nagaraju S, Truong D, Mouneimne G, Nikkhah M Drwal MN, Preissner R (2016) WITHDRAWN – a
(2018) Microfluidic tumor–vascular model to study resource for withdrawn and discontinued drugs.
breast cancer cell invasion and intravasation. Adv Nucleic Acids Res 44:D1080–D1086
Healthc Mater 0:1701257. https://doi.org/10.1002/ Skardal A, Devarasetty M, Forsythe S, Atala A, Soker S
adhm.201701257 (2016a) A reductionist metastasis-on-a-chip platform
Oleaga C et al (2016) Multi-organ toxicity demonstration for in vitro tumor progression modeling and drug
in a functional human in vitro system composed of screening. Biotechnol Bioeng 113:2020–2032. https://
four organs. Sci Rep 6:20030. https://doi.org/10.1038/ doi.org/10.1002/bit.25950
srep20030 Skardal A, Shupe T, Atala A (2016b) Organoid-on-a-chip
Osaki T, Sivathanu V, Kamm RD (2018) Vascularized and body-on-a-chip systems for drug screening and
microfluidic organ-chips for drug screening, dis- disease modeling. Drug Discov Today 21:1399–1411.
ease models and tissue engineering. Curr Opin https://doi.org/10.1016/j.drudis.2016.07.003
Biotechnol 52:116–123. https://doi.org/10.1016/j. Skardal A et al (2017) Multi-tissue interactions in an inte-
copbio.2018.03.011 grated three-tissue organ-on-a-chip platform. Sci Rep
Pauty J et al (2018) A vascular endothelial growth factor- 7:8837. https://doi.org/10.1038/s41598-017-08879-x
dependent sprouting angiogenesis assay based on an Stock K et al (2016) Capturing tumor complexity in vitro:
in vitro human blood vessel model for the study of comparative analysis of 2D and 3D tumor mod-
anti-angiogenic drugs. EBioMedicine 27:225–236. els for drug discovery. Sci Rep 6:28951. https://doi.
https://doi.org/10.1016/j.ebiom.2017.12.014 org/10.1038/srep28951
Rape AD, Kumar S (2014) A composite hydrogel plat- Tadimety A, Syed A, Nie Y, Long CR, Kready KM, Zhang
form for the dissection of tumor cell migration at tis- JXJ (2017) Liquid biopsy on chip: a paradigm shift
sue interfaces. Biomaterials 35:8846–8853. https:// towards the understanding of cancer metastasis. Integr
doi.org/10.1016/j.biomaterials.2014.07.003 Biol 9:22–49. https://doi.org/10.1039/c6ib00202a
Riedl A et al (2017) Comparison of cancer cells in 2D vs Theberge AB, Yu J, Young EWK, Ricke WA, Bushman
3D culture reveals differences in AKT–mTOR–S6K W, Beebe DJ (2015) Microfluidic multiculture assay
signaling and drug responses. J Cell Sci 130:203–218. to analyze biomolecular signaling in angiogenesis.
https://doi.org/10.1242/jcs.188102 Anal Chem 87:3239–3246. https://doi.org/10.1021/
Rigat-Brugarolas LG et al (2014) A functional micro- ac503700f
engineered model of the human splenon-on-a-chip. Tsamandouras N, Chen WLK, Edington CD, Stokes CL,
Griffith LG, Cirit M (2017) Integrated gut and liver
microphysiological systems for quantitative in vitro microfluidics. Nat Protoc 11:134–148. https://doi.
pharmacokinetic studies. AAPS J 19:1499–1512. org/10.1038/nprot.2016.003
https://doi.org/10.1208/s12248-017-0122-4 Watson DE, Hunziker R, Wikswo JP (2017) Fitting tissue
Valastyan S, Weinberg RA (2011) Tumor metasta- chips and microphysiological systems into the grand
sis: molecular insights and evolving paradigms. scheme of medicine, biology, pharmacology, and
Cell 147:275–292. https://doi.org/10.1016/j. toxicology. Exp Biol Med 242:1559–1572. https://doi.
cell.2011.09.024 org/10.1177/1535370217732765
Vavourakis V, Wijeratne PA, Shipley R, Loizidou M, Williamson A, Singh S, Fernekorn U, Schober A (2013)
Stylianopoulos T, Hawkes DJ (2017) A validated The future of the patient-specific body-on-a-chip.
multiscale in-silico model for mechano-sensitive Lab Chip 13:3471–3480. https://doi.org/10.1039/
tumour angiogenesis and growth. PLoS Comput c3lc50237f
Biol 13:e1005259. https://doi.org/10.1371/journal. Xu Z et al (2016) Design and construction of a multi-organ
pcbi.1005259 microfluidic chip mimicking the in vivo microenvi-
Vernetti L et al (2017) Functional coupling of human ronment of lung cancer metastasis. ACS Appl Mater
microphysiology systems: intestine, liver, kidney Interfaces 8:25840–25847. https://doi.org/10.1021/
proximal tubule, blood-brain barrier and skeletal acsami.6b08746
muscle. Sci Rep 7:42296. https://doi.org/10.1038/ Zervantonakis IK, Hughes-Alford SK, Charest JL,
srep42296 Condeelis JS, Gertler FB, Kamm RD (2012) Three-
W.H.O (2018) Cancer. http://www.who.int/cancer/en/ dimensional microfluidic model for tumor cell intrava-
Wang YI, Carmona C, Hickman JJ, Shuler ML (2018) sation and endothelial barrier function. Proc Natl Acad
Multiorgan microphysiological systems for drug Sci U S A 109:13515–13520. https://doi.org/10.1073/
development: strategies, advances, and chal- pnas.1210182109
lenges. Adv Healthc Mater 7:1701000. https://doi. Zhang Q, Liu T, Qin J (2012) A microfluidic-based device
org/10.1002/adhm.201701000 for study of transendothelial invasion of tumor aggre-
Waring MJ et al (2015) An analysis of the attrition of gates in realtime. Lab Chip 12:2837–2842. https://doi.
drug candidates from four major pharmaceutical org/10.1039/c2lc00030j
companies. Nat Rev Drug Discov 14:475. https://doi. Zhang YS et al (2017) Multisensor-integrated organs-on-
org/10.1038/nrd4609 chips platform for automated and continual in situ
Warkiani ME, Khoo BL, Wu L, Tay AKP, Bhagat AAS, monitoring of organoid behaviors. Proc Natl Acad
Han J, Lim CT (2016) Ultra-fast, label-free isolation Sci U S A 114:E2293–E2302. https://doi.org/10.1073/
of circulating tumor cells from blood using spiral pnas.1612906114
Biomaterials and Microfluidics
for Liver Models 5
Alain da Silva Morais, Joaquim M. Oliveira,
and Rui L. Reis
Abstract several tissue engineering challenges are associ-

Over the past years, important progresses have ated with applying this scaffold technology to
been made in the field of tissue engineering. one vital organ construct: liver. The development
Many of the early trials to improve the develop- of microscale tissue (“micro-tissue”) constructs
ment of an engineered tissue construct were cen- to mimic partially the complex structure–func-
tered on the concept of seeding cells onto tion interactions of liver parenchyma have been
biomaterial scaffold. By means of innovative obtained through the engineering of sophisti-
manufacturing machineries, the conception of a cated biomaterial scaffolds, liver-cell sources,
preformed scaffold became possible. Nowadays, and in vitro culture techniques. For in vitro
applications, micro-tissue constructs are being
upgraded into cell-based assays for testing acute,
chronic and idiosyncratic toxicities of drugs or
A. da Silva Morais (*) pathogens. The present chapter will focus on the
3B’s Research Group, I3Bs – Research Institute on biomaterials currently used for the development
Biomaterials, Biodegradables and Biomimetics, of in vitro liver constructs as well as the descrip-
tion of the microfluidic-based models that show
Regenerative Medicine, AvePark, Parque de Ciência e great promise for liver regenerative medicine
Tecnologia, Zona Industrial da Gandra, Barco, approaches.
Guimarães, Portugal
ICVS/3B’s–PT Government Associate Laboratory, Keywords
Braga, Guimarães, Portugal Biomaterials · Decellularized matrices · Liver
e-mail: alain.morais@i3bs.uminho.pt
models · Microfluidics · Natural polymers ·
J. M. Oliveira · R. L. Reis Hydrogels
ICVS/3B’s PT Government Associate Lab, Highlights
The Discoveries Centre for Regenerative and • Micro-technology approaches, combining
Precision Medicine, Headquarters at University specific biomaterials and microfluidic

https://doi.org/10.1007/978-3-030-36588-2_5
66 A. da Silva Morais et al.
latforms, present a high potential in the cre-

p Nowadays, orthotopic liver transplantation
ation of liver models. (OLT) is the only clinical accepted therapeutic
• Most promising approaches, some already option for patients with acute or advanced liver
applied in vivo, are highlighted. disease (Burra and Freeman 2012; Dutkowski
et al. 2015). However this surgical procedure is
limited due to low organ availability and the
5.1 Introduction requirements for donor livers (Adam et al. 2012),
resulting in a continuing need for alternative
Over the past years, important progresses have strategies for liver transplantation. Among these,
been made in the field of tissue engineering. hepatocytes transplantation (Hughes et al. 2012),
Many of the early trials to improve the develop- extracorporeal liver support (Stange 2011;
ment of an engineered tissue construct were cen- Stutchfield et al. 2011) and stem cell therapy
tered on the concept of seeding cells onto (Drosos and Kolios 2013) are being used in order
biomaterial scaffold. By means of innovative to extend the life of patient or to regenerate the
manufacturing machineries, the conception of a hepatic tissue until availability of a donor liver
preformed scaffold became possible. (see Fig. 5.1).
The liver, the largest solid organ of the body, Tissue engineering (TE) is based on the
plays a central role in the maintenance of meta- involvement of cells, signals (physically from a
bolic homeostasis and the biotransformation and bioreactor or chemically from growth factors),
elimination of xenobiotic substances from the and biocompatible scaffolds, providing an ideal
body, associated with endocrine and exocrine environment for cell growth and differentiation
activities. Despite his highly regenerative poten- with the goal to develop constructs relevant both
tial, damages to hepatocytes by drug and viral in vitro and in vivo (Bajaj et al. 2014). Liver tis-
infections or toxins, resulted in a significant sue engineering aimed to develop extracorporeal
reduction of functionality and regeneration capa- liver-support systems or hepatic regeneration
bility (Protzer et al. 2012). Liver failure can be based on cell transplantation or implantable
divided into acute liver failure (ALF), in the devices (Ananthanarayanan et al. 2011). These
absence of pre-existing liver disease, chronic strategies planned to integrate these structures of
decompensation of an end-stage liver disease, or liver function and mimic hepatic structural design
Acute-on-chronic liver failure (ACLF) of known array from keeping liver cells in biocompatible
or unknown underlying chronic liver disease and biodegradable scaffolds to a more bottom-up
(Kim and Kim 2013). To understand the burden method that involves the elaboration of the
of liver disease in Europe, the incidence and microscopic subunit of the whole tissue. The use
prevalence of two end-stage of liver pathology of biocompatible and biodegradable materials is
have been studied, cirrhosis and primary liver relevant for TE applications. Those biomaterials
cancer. Liver diseases have been found to con- play an important role by providing a defined and
tribute significantly to the global burden of mor- tunable 3D environment that is required for com-
tality and morbidity (Collaborators 2015) causing plex cellular organization and organoid growth in
approximately 2% of all deaths. World Health vitro or in vivo. For liver tissue engineering, the
Organization (WHO) estimate that hepatocellular development of new artificial matrices needed to
carcinoma (HCC) accounts 47,000 deaths in satisfy general requirements of biomaterials,
Europe with a growing incidence over the next including cell and tissue compatibility. Moreover,
10–15 years, and represents more than 80% of the chemical and physical properties of the mate-
primary liver cancers (Siegel et al. 2012). The rials are often customized to encourage cell func-
most common causes of HCC and cirrhosis in tions, including liver cells (Liu and Chen 2005).
Europe are alcohol, viral hepatitis B and C and For example, to mimic the hepatic native extra-
metabolic syndromes related to overweight and cellular matrix (ECM), the stiffness of a devel-
obesity (Singal and Anand 2013; Poh et al. 2015). oped biomaterial-based matrix is required to be
5 Biomaterials and Microfluidics for Liver Models 67
Fig. 5.1 Rationale Scheme. Once the conventional ther- approaches resulting in the development of liver tissue
apeutic methods are unsuccessful to treat decompensated engineering strategies
liver diseases, there is a continuing need for advanced
soft (Discher et al. 2005). Also, the matrix topol- process currently used for the development of
ogy and chemistry are engineered to guide hepa- liver models.
tocytes to form spheroids and differentiate
correctly (Jiang et al. 2010). For liver TE, a vari-
ety of natural and synthetic biomaterials have 5.2 Biomaterials – Based In Vitro
been assessed (Willerth and Sakiyama-Elbert Liver Models
2008; Ananthanarayanan et al. 2011).
In combination with biomaterial, the improve- Different types of polymers have been used for
ment of the microenvironment will control the the development of artificial matrices for liver tis-
cells for liver regeneration. The co-culture of sue engineering. These are classified in main cat-
cells allows the communication through para- egories of natural and synthetic polymers. In
crine factors and helps the formation of more addition to these, native extracellular matrices in
functional cellular constructs (Su et al. 2017). the form of a decellularized biomatrix have been
Additionally, the culture of cells in 3D environ- also applied. The biomaterials and their applica-
ments allows them to reorganize in a way similar tions in liver tissue engineering are summarized
as in native tissue (Knight and Przyborski 2015). herein. The advantages and drawbacks of some
Mimic in vitro the in vivo condition remains pos- of these biomaterials in liver tissue engineering
sible through the use of bioreactors allowing to applications were described in Table 5.1.
finely control the supply and circulation of oxy-
gen and depleted nutrients (Ebrahimkhani et al.
2014; Schuerlein et al. 2017).
Important advances have been made in micro- 5.2.1 Naturally-Based Biomaterials
fabrication and nanotechnology fields. These
aimed to precisely control the spatial distribution Liver extracellular matrix (ECM) is mainly com-
of biomolecules and substrate topography at posed by protein collagen, mostly type I and III,
micro- to nanometer resolution. Then, they can and glycoprotein fibronectin (Martinez-
be used to engineer extracellular micro- Hernandez and Amenta 1993). Thus, natural
environments, cell shape and intercellular tissue polymers such as proteins (e.g. collagen and
structures with sophisticated topographies. fibrin) and polysaccharides (e.g. chitosan, algi-
Microfluidic platforms for cell culture are rap- nate, hyaluronic acid) may allow the artificial
idly gaining importance in biological research matrix to mimic the composition of the natural
and drug development applications, such as stem liver ECM.
cell differentiation studies (Khademhosseini Collagen, the most abundant component of
et al. 2006) and drug metabolism or toxicity stud- ECM (Aycock and Seyer 1989), is composed by
ies (Sivaraman et al. 2005; Kane et al. 2006). a triple-helical structure to form a 3D architec-
Microfluidic channels are particularly interesting ture, highly organized, to interrelate with other
since they can be simply multiplexed with inte- components then playing an essential role to keep
grated fluid handling procedures for efficient and the structural and biological integrity of ECM
high throughput cellular analysis (Dittrich and (Cen et al. 2008). Therefore, collagen has been
Manz 2006), in situ monitoring of cellular events used in several TE approaches including between
(Tourovskaia et al. 2005; Lee et al. 2006) and can others cartilage (Deponti et al. 2013), bone
mimic a physiological cellular microenvironment (Zhang et al. 2018), skin (Pensalfini et al. 2017)
with controlled shear stresses and distribution of and vessels (Chan et al. 2016). Despite the use of
biochemical molecules at cellular level (Walker monolayer and sandwich forms of collagen in
et al. 2004; Kim et al. 2006). liver TE (Zeigerer et al. 2017), collagen alone did
Herein, we overview and discuss the advances not present enough stiffness to support cells
on the biomaterials as well as on the microfluidic (Vasanthan et al. 2012). Moreover, the degrada-
tion of collagen is much faster than synthetic
Table 5.1 Biomaterials applied for fabrication of in vitro liver models

Some liver models
Biomaterials Advantages Drawbacks (Refs.)
Natural Collagen Native component of the ECM liver Lacks mechanical Aycock and Seyer
matrix, biodegradable, properties, costly (1989), Cen et al.
biocompatible, presented several (2008), Lyu and
cell binding motifs Untereker (2009),
Vasanthan et al.
(2012), Chan et al.
(2016) and
Zeigerer et al.
(2017)
Fibrin Promotes vascularization, Fast degradation, poor Bruns et al.
hydrolytically biodegradable mechanical strength, able (2005), Ahmed
to induce immune et al. (2008),
responses Banihashemi et al.
(2015) and
Stevens et al.
(2017)
Chitosan Glycosoaminoglycan-like Low mechanical strength, Rinaudo (2006),
polysaccharide, supports spheroid able to induce immune Seo et al. (2006),
formation, biocompatible, promotes responses Wang et al. (2016)
high cellular levels of liver and Fan and Yang
specificity (2017)
Alginate Hydrophilic, allows cell seeding Reduced adherence, able Dvir-Ginzberg
and microencapsulation, supports to induce immune et al. (2003), Lee
spheroid formation, easy handled responses and Mooney
(2012), Alan et al.
(2015) and Kim
et al. (2017)
Gelatin Derived from natural polymer Dissolved at temperatures Mariod and Fadul
collagen, excellent at or above 37 °C, made (2013) and Mari
biocompatibility, non-immunogenic gels at lower et al. (2017)
temperatures around
room temperature
Hyaluronic acid Natural constituent of the liver Highly viscous, reduced Collins and
biomatrix, biocompatible, allows mechanical strength, Birkinshaw (2013)
coculture of hepatocytes quickly degraded and Hemshekhar
et al. (2016)
Silk Low cost purification, Possibility of minimal Naresh and Utpal
biocompatible, ease of chemical inflammatory responses (2012) and Kundu
modification et al. (2013)
Elastin Accurate control over scaffold Need functionalization to Kanta (2016) and
composition, biodegradable, improve cell adhesion Miranda-Nieves
biologically relevant and Chaikof
functionalization, stimuli (2017)
responsive
Laminin Promote hepatogenic differentiation Expensive Khalaj et al.
(2016), Watanabe
et al. (2016) and
Garg (2017)
(continued)
Table 5.1 (continued)
Some liver models
Biomaterials Advantages Drawbacks (Refs.)
Synthetic PLA Allowed hepatocyte aggregation, Acidic degradation Shah Mohammadi
biocompatible product, initiate peptide et al. (2014) and
degradation and Perez et al. (2017)
PLLA Preserved hepatic function for up to inflammation, lack cell Török et al. (2010)
4 weeks, promote hepatogenic recognition signals and Wang et al.
differentiation (2013)
PGA Good mechanical properties, Kazemnejad
degradation rate (2009) and Knight
and Payne (2013)
PLGA Biocompatible, biodegradable, can Hammond et al.
be modulated by changes in ratio of (2006), Makadia
PLLA:PLG, promote stem cell and Siegel (2011),
differentiation Mironov Anton
et al. (2016) and
Caddeo et al.
(2017)
PCL Inert, biocompatible, biodegradable Highly hydrophobic, Semnani et al.
slow degradation rate (2016) and
Malikmammadov
et al. (2018)
PDMS Oxygen permissible membrane, Highly hydrophobic, Powers et al.
allow high cell seeding density permeable to water (2002) and Sia and
vapours Whitesides (2003)
PEG Hydrophilic, resistant to protein Highly hydrophobic, Stevens et al.
adsorption, responsive to chemical slow degradation rate (2015)
modification, can be polymerized in
presence of cells
PVA Inert hydrophilic matrix, preserved Deficient of cell binding Bidault et al.
functions of cryopreserved motifs (2013), Jain et al.
hepatocytes (2015) and Kumar
and Han (2017)
Polyacrylamide Promotes development of spheroid Non-biodegradable Cozzolino et al.
with controlled size and high yield (2016)
Other Decellularized Whole structural and functional Transplanted graft Lee and Cho
matrix constituents of the native liver survival ≤8 days, starts (2012), Skardal
matrix stimulates effective cell clotting in vivo due to et al. (2012)
functions in vitro exposed collagen
PLA Polylactic acid, PGA Polyglycolic acid, PLLA Poly (L-lactic acid), PLGA Poly(lactic-co-glycolic acid), PCL Poly
(ε-caprolactone), PDMS Poly (dimethylsiloxane), PEG Polyethylene glycol, PVA Polyvinyl alcohol. Source: Adapted
with permission from (Jain et al. 2014) Copyright © 2013, Asian Pacific Association for the Study of the Liver
polymers (Lyu and Untereker 2009). These fold (Banihashemi et al. 2015), the main
weaknesses limited the use of collagen alone in disadvantages of fibrin are the reduced mechani-
liver TE. cal properties and fast enzymatic degradation (de
Fibrin is a soluble protein made and secreted la Puente and Ludena 2014). Current approaches
by the liver. Since it can be easily obtained from for liver TE are based on the use of fibrin gel as
blood plasma of the own patient, reducing then matrix (Bruns et al. 2005; Lee et al. 2010a).
the risk of immunological reaction or infections, Recently Stevens et al. (Stevens et al. 2017) used
it has been widely used TE applications (Ahmed fibrin hydrogel to encapsulate endothelial cell
et al. 2008). Despite the ability of the cells to pro- cords and hepatic cellular aggregates in order to
liferate and produce ECM within the fibrin scaf- create “liver tissue seeds”. These liver seed grafts,
appropriate for ectopic implantation, were able to 2015) thanks to its source abundance, low price
significantly increase cell expansion in response and biocompatibility. In vitro studies have
of tissue injury (Stevens et al. 2017). assessed the encapsulation of different cell types
Chitosan is a unique biopolymer obtained into alginate scaffolds (e.g. hydrogel, micro-
from N-deacetylation of the chitin (Rinaudo spheres, gel and bioink). Human induced plurip-
2006; Elieh-Ali-Komi and Hamblin 2016) which otent stem cells (hiPSCs) and human embryonic
showed significant analgesic and antibacterial stem cells (hESCs) were printed into alginate
proprieties as well as mucoadhesive capacity hydrogels and, after hepatocyte differentiation,
(Croisier and Jérôme 2013). Up to now, different demonstrated good viability as well as liver mor-
chitosan-based architectures were developed phology and functions (Alan et al. 2015). The
including gels, sponges, fibers and films (Croisier same results have been observed when encapsu-
and Jérôme 2013) and their potential for tissue lation of mouse primary hepatocytes or hepato-
engineering application has been assessed (Kim cytes spheroids within alginate scaffolds
et al. 2008b; Sultana et al. 2015; Agrawal and (Miranda et al. 2010; Kim et al. 2017). The
Pramanik 2016; Dash et al. 2017; Silva et al. impact of the functionalization of alginate has
2017). In liver tissue engineering approaches, also been assessed. Capsules using galactosyl-
due to its low bioactivity toward tissue cells and ated alginate have been developed for hepato-
the absence of cell bonding domains, chitosan is cytes entrapment (Yang et al. 2002; Lou et al.
frequently modified with functional molecules or 2017). The authors observed that the encapsu-
combined with other materials (Seo et al. 2006; lated hepatocytes exhibited higher cell viability
Cheng et al. 2016; Fan and Yang 2017). As exam- and maintained liver-specific functions.
ple, the modification of chitosan nanofibrous Gelatin and its derivatives is a natural bio-
scaffold with surface-galactose ligands can both polymer derived from animal collagen and is
reduce the time of degradation of the scaffold and divided in two types: type A obtained from par-
improves the bioactivity of primary hepatocytes tial acid hydrolysis of animal collagen and type B
in vitro (Feng et al. 2009). Also, functionalization obtained from alkaline hydrolysis (Mariod and
of chitosan with lactose moieties showed to Fadul 2013). However this polymer presented a
improve cell adhesion and that this effect is huge limitation since it is soluble in aqueous
dependent of the increase of the substitution solutions. A recent review focused on the devel-
degree (Wang et al. 2016). However the authors opment of gelatin-based composites in order to
observed that moderate substitution degree of surpass this limitations and modulate the proper-
chitosan with lactose moieties is related with bet- ties of the formulations for tissue engineering
ter mechanical stability and biocompatibility of applications (Mari et al. 2017). For example,
the complex, suggesting it as a promising mate- Choi et al. (2014) developed a human gelatin
rial for liver tissue engineering (Wang et al. tissue-adhesive hydrogel with haemostatic poten-
2016). Moreover, Teotia et al. (2015) modified tial assessed using a haemorrhaging liver rat
chitosan with porous polysulfone (Psf) and model. Recently, functionalization of gelatin into
polysulfone-tocopheryl polyethylene glycol suc- a photo-crosslinkable gelatin methacryloyl
cinate (Psf-TPGS). The composite membrane (GelMA) has been used to create protein-based
developed improved the adhesion and expansion 3D scaffolds with uniform pore interconnectivity,
of HepG2 cells on the external surface, maintain- structural stability and tailorable degradation
ing the internal surface with a decent blood com- properties (Lee et al. 2017). The scaffolds devel-
patibility. Then the authors suggested the use of oped have been shown to improve in vitro hepa-
this hollow fiber to vascularize the bioartificial tocytes attachment and intercellular interaction
liver (Teotia et al. 2015). compared to 2D culture system.
Alginate is a natural polysaccharide widely Hyaluronic acid (HA) is the principal compo-
used in tissue engineering (Dvir-Ginzberg et al. nent of perisinusoidal space in liver (Nallagangula
2003; Lee and Mooney 2012; Venkatesan et al. et al. 2018). Then this non-sulfated glycosamino-
glycan represented an excellent choice for applica- nance of primary human hepatocytes (Watanabe
tion in various tissue engineering strategies (Collins et al. 2016), hepatogenic differentiation of human
and Birkinshaw 2013; Hemshekhar et al. 2016) bone marrow mesenchymal stem cells (Khalaj
including its use as hydrogel for liver application. et al. 2016) and hepatic and hepatic specification
Catapano et al. used non-woven fabrics of HA of human pluripotent stem cells (Kanninen et al.
esters to culture liver cells (Catapano et al. 2001). 2016).
These non-wave fabrics were later enriched with A single biomaterial can barely meet all the
components of ECM by Zavan et al. (2005). The requirements of hepatic regeneration, thus blends
authors demonstrated an increase of cell survival of materials appeared to take advantages of each
the hepatocytes up to 14 days in vitro and up to component. Several studies have been based in
35 days after implantation of hepatocytes aggre- the use of these composite materials for liver tis-
gates (Zavan et al. 2005). Since transdifferentiation sue engineering applications. For examples: (i)
of hepatic stellate cells (Ito cells) into myofibro- Lee et al. developed a highly bioactive ink con-
blasts is a key event in hepatic fibrogenesis (Higashi sisting in a mixture of collagen/extracellular
et al. 2017) and the stiffness of the environment has matrix (ECM) and alginate (Lee Hyeong et al.
been shown to influence the progression of the 2015), (ii) Shang et al. prepared a galactosylated
fibrosis (Olsen et al. 2011), the methacrylation of chitosan (GCs)/HA porous hybrid sponge (Shang
HA (MeHA) hydrogel has also been assessed. et al. 2014), (iii) Tripathi and Melo created a
Studies have demonstrated that MeHA “stiff” sponge-like biocomposite agarose–chitosan scaf-
hydrogels affected the morphology and differentia- fold (Tripathi and Melo 2015), (iv) Deng et al.
tion potential of HSC into myofibroblasts used an alginate/xyloglucan scaffold (Deng et al.
(Guvendiren et al. 2014; Caliari et al. 2016). 2015), (v) Nakamura et al. prepared a heparin-
Other natural-based biomaterials have also conjugated gelatin scaffold able to hold vascular
been investigated in order to develop new models endothelial growth factor (VEGF) (Nakamura
for liver tissue engineering applications. Silk et al. 2013), (vi) Arisaka et al. modified a thermo-
fibroins, which can be obtained from several responsive surface with heparin-binding epider-
sources such as silkworms, spiders and flies, have mal growth factor-like growth factor (HB-EGF)
demonstrated huge potential for tissue regenera- (Arisaka et al. 2016), (vii) Yang et al. character-
tion (Kundu et al. 2013) including the develop- ized in vitro and in vivo silk fibroin (SF)/gelatin
ment of a biomimetic artificial liver extracellular composite scaffolds (Zhao et al. 2011), (viii)
matrix (Naresh and Utpal 2012). Elastin, one of Kasoju and Bora fabricated SF and GCs nanofi-
the most stable protein found in the ECM of elas- brous mesh (Naresh and Utpal 2012), (ix) Gong
tic organs like arteries, lung and liver. Liver fibro- et al. developed chitosan/gelatin scaffolds with
sis development has been associated to a fine three-dimensional channels (Gong et al.
dysregulation of elastin synthesis and metabo- 2014), and (x) Xia et al. developed a bioreactor-
lism (Kanta 2016). However, although the pres- based bioartificial liver (BAL) with stacked sand-
ence of elastin in hepatic tissues, the use of wich culture plates containing hepatocytes
elastin-based biomaterials in liver tissue engi- together with porous and non-porous collagen-
neering is limited and mostly focused on improv- coated PET films (Xia et al. 2012). Some exam-
ing hepatocytes 2D cultures and spheroids ples of naturally-based in vitro models were
formation (Miranda-Nieves and Chaikof 2017). presented in Fig. 5.2.
Laminin, a structural protein of the ECM of
skeletal muscle fibers with a crucial role in tissue
development, maintenance and function regula- 5.2.2 Synthetic Biomaterials
tion, was also involved in the development of
scaffolds for tissue engineering applications While naturally-based biomaterials presented
(Garg 2017). Recent studies have focused on the several advantages for liver tissue engineering
use of laminin-based matrices for the mainte- applications, they still present many deficiencies
Fig. 5.2 Naturally-based in vitro models for liver tissue printing of alginate scaffold (Reprinted with permission
engineering applications. (I) Development of cell polarity (Kim et al. 2017) © 2017, the Korean Surgical Society.
in collagen sandwich hepatocytes (Reprinted with permis- https://creativecommons.org/licenses/by-nc/4.0/). (IV)
sion (Zeigerer et al. 2017) © 2016 The Authors. Published A. Gelatin methacryloyl-based (GelMA) inverted colloi-
by Elsevier Inc.). (II) SEM images of rat primary hepato- dal crystal (ICC) hydrogels. B. Evaluation of liver-specific
cytes and hepatocyte aggregates cultured on chitosan and functions of Huh7.5 cell constructs in 3D GelMA ICC
galactosylated chitosan (GC) films and nanofibers after scaffolds and on 2D GelMA substrates. (Reprinted with
7 days (Reprinted with permission (Feng et al. 2009) © permission (Watanabe et al. 2016) © 2018 Springer
2009 Elsevier Ltd. All rights reserved). (III) Isolation of Nature Limited. All rights reserved. https://creativecom-
mouse primary hepatocytes and 3-dimensional (3D) bio- mons.org/licenses/by/4.0/)
as for example reduced mechanical properties, et al. 2017). Several studies have applied PLA-
limited availability, and variable degradation based scaffolds for different tissue regeneration
rate. Thus, as alternative, synthetic biomaterials applications. For example PLA scaffolds, pro-
with tunable degradability and mechanical prop- duced with different architecture, have been
erties have been developed and assessed. To widely used for tissue engineering (Shah
improve cell adhesion, the synthetic biomaterials Mohammadi et al. 2014; Magiera et al. 2017; Xu
were functionalized with polysaccharides, pro- et al. 2017). Regarding liver applications,
teins or polypeptides (Vasanthan et al. 2012). The Bierwolf et al. demonstrated that PLA nanofi-
key synthetic biomaterials used for liver tissue bers, created by electrospinning technique,
engineering applications are aliphatic polyesters improved proliferation and expression of liver
(e.g. PLA, PGA, PCL…), polyethers (e.g. PEG) markers by primary rat hepatocytes (Bierwolf
and polyamides (e.g. PVA). et al. 2010). The biodegradable poly(L-lactic
Aliphatic polyesters are synthetic materials acid) (PLLA) aliphatic polymer has demon-
which possess ester bonds vulnerable to hydroly- strated potential to improve hepatocyte expan-
sis and which can offer a microenvironment to sion and function in vitro (Wang et al. 2013) as
orient tissue regeneration through improvement well as it represented an optimized method for
of material degradability and mechanical sup- primary hepatocyte transplantation in vivo (Török
port. Polylactic acid (PLA) is a biocompatible et al. 2010). Polyglycolic acid (PGA) is a poly(α-
and biodegradable aliphatic polyester (Perez hydroxyl) ester which can be tuned to improve
degradation rate and mechanical properties. tosylated PCL (GPCL) membranes which main-
Studies have been focused on the use of PGA for tained the physiological functions of hepatocytes
different tissue engineering applications (Cao in vitro. Semnani et al. (2016) created nanofi-
and Kuboyama 2010; Knight and Payne 2013; brous scaffolds through co-electrospinning of
Toosi et al. 2016). For liver regeneration, some PCL with chitosan (CS) aiming to use this com-
encouraging results have been presented posite complex as matrix to enhance liver regen-
(Kazemnejad 2009). In order to improve the eration. Bishi et al. (2013) synthetized poly
mechanical property and the degradation rate, (L-lactic acid)-co-poly(ε-caprolactone)
PLA and PGA are frequently blended into the (PLACL) and developed PLACL/collagen nano-
co-polymer poly(lactic-co-glycolic acid) fibrous scaffolds able to support the proliferation
(PLGA). Compared to the biocompatibility of and differentiation potential of human MSCs into
alginate and chitosan, PLGA has been associated hepatocyte-like cells. Shim et al. (2013) fabri-
with significant inflammatory response (Lee cated PCL/PLGA/collagen scaffolds. The authors
et al. 2012). Despite that, PLGA-based scaffolds observed that encapsulated rat primary hepato-
have been frequently investigated and studies cytes within the composite scaffolds are viable
have been focused on the use of these PLGA and able to secrete albumin to up to 10 days.
scaffolds for the development of models for drug Moreover, PLC has also been used to develop
delivery and tissue engineering applications delivery systems of specific proteins. For exam-
(Makadia and Siegel 2011; Pan and Ding 2012; ple, Yu et al. (2014) developed a biodegradable
Gentile et al. 2014; Mironov Anton et al. 2016). PCL nanofiber matrix able to release the vascular
PLGA has also been assessed from a while in endothelial growth factor (VEGF) which
liver studies (Hammond et al. 2006; Caddeo et al. improved, through a concentration-dependent
2017). As example, Liu et al. (2016) developed a effect, liver regeneration in rats submitted to 70%
PLGA scaffold allowing the co-culture of hepa- partial hepatectomy.
tocytes and mesenchymal stem cells (MSCs). A possible drawback for the use of aliphatic
They observed that the differentiation of MSCs polyesters is based on their degradation products
and the survival of hepatocytes within the scaf- which can alter the microenvironment of the
fold is dependent of the ratio hepatocytes/MSCs, cells as well as their natural hydrophobic proper-
with the ratio 5:1 respectively showed to be the ties, limiting their application in tissue engineer-
most efficient. Moreover, they showed that the ing (Hammond et al. 2006). Therefore others
direct transplantation of the construct could synthetic biomaterials have been considered
improve the rejection response and the restora- such as polyethers and polyamides. Polyethylene
tion of liver functions in mice model. As for other glycol (PEG) is a synthetic polyether which sol-
biomaterials, composites with different PLGA- ubility in aqueous solutions and organic solvents
based compositions and forms have been created contributed to its biocompatibility and process-
such as PLGA combined with collagen, gelatin ability (Ivanova et al. 2014). Therefore PEG-
or plasma treated (Hasirci et al. 2001), and based hydrogels were widely used in tissue
PLGA-based porous scaffolds (Hammond et al. engineering applications (Burdick and Stevens
2011) or microspheres incorporating growth fac- 2005). For liver tissue engineering approaches, a
tors (Zhu Xin et al. 2007). Poly (ε-caprolactone) degradable hydrogel has been developed
(PCL), and the PLC-based biomaterials, are ali- (Stevens et al. 2015). This model was based on
phatic polyester used for several biomedical PEG- diacrylamide (PEGDAAm) grafted with
applications based on the good processability and matrix- metalloproteinase sensitive peptides
biocompatibility of PCL (Malikmammadov et al. which improved 3D cell migration while the
2018). For liver regeneration purpose applica- hydrogel is degraded by cell action. Stevens
tions, several studies have been realized. Zhang et al. showed that, after ectopic implantation in
et al. (2014) functionalized PCL with galactose nude mice, the hepatic PEGDAAm-based tissue
to improve PCL’ wettability and prepared galac- was still present and preserve functionality for
Fig. 5.3 Synthetic-based in vitro models for liver tissue mission (Semnani et al. 2016) © 2017 Taylor & Francis).
engineering applications. (I) A. Nanofibrous scaffolds (III) Co-culture of rat primary hepatocytes with liver
seeded with rat primary hepatocytes. B. Immunofluorescent endothelial cells in PEG-diacrylamide hydrogel improves
staining of hepatocyte-specific factors (Reprinted with engineered tissue longevity (Reprinted with permission
permission (Bierwolf et al. 2010) © 2010, John Wiley and (Stevens et al. 2015) Copyright © 2015 The Authors.
Sons). (II) Scanning electron microscope (SEM) results of Journal of Biomedical Materials Research Part A
Hepa 1–6 epithelial liver mouse cells cultured on PCL/ Published by Wiley Periodicals, Inc.)
chitosan nanofibers after 4 and 168 h (Reprinted with per-
more than 3 weeks. Polyvinyl alcohol (PVA), a hydrogels’ stiffness. Some examples of synthetic
water-soluble polymer, has been widely used for in vitro models were presented in Fig. 5.3.
tissue engineering applications (Kumar and Han
2017). For liver regeneration models, different
approaches have been investigated since the 5.2.3 Decellularized Matrices
development of self-supported fibrin-PVA inter-
penetrating polymer networks (IPN) (Bidault Decellularization protocol is based on the
et al. 2013) to macroporous hydrogel consisting removal of all cells of the tissue resulting in an
of acrylonitrile and N-vinyl-2-pyrrolidone copo- extracellular matrix (ECM) scaffold of the native
lymer (PANNVP) and IPN together with chito- tissue (Lee and Cho 2012). Decellularized matri-
san as hepatocyte carriers (Jain et al. 2015). ces, with defined chemical composition and
Moreover, application of polyacrylamide has architecture mimicking the targeting tissue but
also been studied in liver regeneration. Cozzolino deprived of immune responses, are widely
et al. (2016) optimized the in vitro conditions for applied in tissue engineering strategies. For
hepatic stem cells and differentiated hepatocytes example, Skardal et al. showed that extracts from
through the modulation of polyacrylamide whole fresh liver tissue and from acellular liver
ECM can be both mixed with collagen, HA or biomolecules (Nikkhah et al. 2012). So, this
heparin-conjugated HA in order to create hydro- technology represented a sophisticated tool for
gels able to maintain primary human hepatocytes the engineering of intercellular tissue structures
functions in vitro (Skardal et al. 2012). The and extracellular micro-environments. The devel-
authors proposed that this 3D model could be opment of microfluidic technologies has engen-
applied for cell therapy and drug and toxicology dered huge interest for a large variety of domains
screening purposes. A similar approach has been due to the capacity to create highly structured
applied to develop a liver ECM (L-ECM) scaf- systems to evaluate the physical importance of
fold (Nakamura and Ijima 2013). fluid mechanisms and transport. Then, there has
Functionalization of the scaffold with hepatocyte been a huge increase in microfluidics develop-
growth factors (HGF) was showed to promote- ment in terms of improvement of fabrication pro-
liver specific functions of hepatocytes. cess and strategies for surface alterations (Kim
Even if L-ECM scaffolds are interesting, the et al. 2008a).
lack of a vascular network which is essential for Microfluidic or lab-on-a-chip platforms have
nutrient and metabolites distribution, represented quickly gain significance in biological research
a huge inconvenient. Then, different emerging and drug development applications, once they
strategies to surpass this problem have been propose a valuable technology able to be use for
developed. One is based on the development of a the development of new biomaterials (Barata
vascular network through endothelialization of et al. 2016). Indeed, this platform has been fre-
the decellularized liver scaffold (Shirakigawa quently applied for drug metabolism and toxicity
et al. 2013), another is centered in the functional- assays (Kimura et al. 2018) as well as for stem
ization of the L-ECM scaffold, presenting a per- cell differentiation studies (Zhang et al. 2017).
severed native vasculature, with heparin to reduce Microfluidic channels are of particular interest
the thrombogenicity without affecting the re- once, in order to promote high throughput and
cellularization of the scaffold neither the hepatic useful cellular assay, they can be multiplexed
functions of the cells in vitro (Bruinsma et al. with integrated fluid handling procedures (Lin
2015). Promising results have also been obtained et al. 2016). Moreover, this procedures have also
by repopulation of L-ECM scaffolds with hepa- been applied for the evaluation of cellular distri-
tocytes and endothelial progenitor cells (Zhou bution of biochemical molecules (Banerjee et al.
et al. 2015a) (Fig. 5.4). 2017), to reproduce a physiological cellular
Despite the decellularized scaffolds presented microenvironment with controlled shear stresses
huge advantages in their composition and struc- (Bissoyi et al. 2016), and the monitoring of cel-
ture compared to other biomaterial-based scaf- lular events in situ (Kwapiszewska et al. 2014).
folds, some challenges persist to be overcome Manufacturing of microfluidics is based on
including the development of optimized decellu- different approaches. Bhattacharjee et al. (2016)
larization protocols in order to keep the native have reviewed the following strategies: injection
ECM structure and allowing the complete molding, soft lithography, 3D printing and paper
removal of the native cells as well as the improve- microfluidics. After comparing them to each
ment of the re-cellularization strategies. other, the authors concluded that the 3D printing
strategy, due to its ability to quickly offer a rapid
prototype of a physical model, is the most rele-
5.3 Microfluidics-Based Liver vant and then represented a more appropriate
Models technology for specific microfluidic models
(Macdonald et al. 2017). Besides the conven-
Microfabrication, and more specifically nano- tional microfluidics models, centrifugal microflu-
technology domain, have recognized important idic or lab-on-a-disc platforms have emerged
advances aiming to determine precisely, up to the since the past few years (Tang et al. 2016). These
nanometer resolution, the spatial dispersal of the platforms presented several advantages com-
Fig. 5.4 Decellularization and recellularization of rat liv- tion method. (Reprinted with permission (Zhou et al.
ers with hepatocytes and endothelial progenitor cells. (I) 2015a) © 2015 International Center for Artificial Organs
Sequential whole-organ decellularization progress (a–f). and Transplantation and Wiley Periodicals, Inc.)
(II) Recellularization process through parenchymal injec-
pared to the traditional-one including a reduced liver-on-a-chip constructs were usually

amount of instrumentation, the proficient removal developed to improve the hepatic functions in
of residual volumes or unwanted bubbles, as well comparison to the conventional culture models
as a density-based sample separation and trans- (Harink et al. 2013). Stimulation of primary rat
portation (Strohmeier et al. 2015). Microfluidic hepatocytes with hepatocyte growth factor
chip-based organs-on-chips technology has been (HGF), diffusing through microfluidic channels,
highlighted due to its capacity to mimic and rec- improved cell phenotype compared to unstimu-
reate the complex functions and structure of lated cells (Patel et al. 2015). Moreover, the
human organs through the cells inside micro- development of a microfluidic device with liver
fabricated channels (Sun et al. 2016). The use of microsomes encapsulated within a 3D PEG-
these microfluidic devices for cell-cell interac- based hydrogel showed that this model was able
tions studies and as platforms for theranostic to screen the impact of liver metabolism on dif-
applications has been recently reviewed (Choi ferent compounds (Lee et al. 2013).
et al. 2017; Rothbauer et al. 2018). Another tech- Microfluidic technology has also been applied
nology recently described and mainly applied for to develop liver microtissues or scaffolds using
the manipulation of biological units, such as par- different biomaterials (Ahadian et al. 2017). In
ticle separation in microfluidic models, design- the beginning of the twenty-first century, Powers
ing, cell electroporation, and isolation of dead et al. developed a microfabricated array bioreac-
versus live cells, is dielectrophoresis (DEP) tor based on a silicon sheet scaffold (Powers et al.
(Chandrasekaran et al. 2016). This technique has 2002). Poly(dimethylsiloxane) (PDMS), a syn-
been applied in the development of an artificial thetic polymer widely used in the field of
microchip composed by liver sinusoid-like struc- biotechnology (Sia and Whitesides 2003), was
tures (Schütte et al. 2011) and of a liver-cell pat- used as frame to create for example a microfabri-
terning lab chip system (Ho et al. 2013). cated bioreactor (Ostrovidov et al. 2004), a
The liver, one of the most complex organ, chitosan-based bioartificial liver (Lee et al. 2010b)
plays a central role in the development of new as well as an extended bile canalicular structure
drugs through the fabrication of 3D bioprinted (Goral et al. 2010). These models showed
liver devices. Several studies have described the improvement of liver specific functions by the
development of appropriate drugs and gene car- cells used for the assays. Alginate, a natural bio-
riers associated with the rise of different micro- material previously described, was also used in
fluidic devices allowing to assess the impact of combination with PDMS to develop an “organ in
drugs under conditions mimicking the biological a droplet” device (Chen et al. 2016) through the
system (Damiati et al. 2018). As depicted in use of a droplet-based microfluidic method. That
Fig. 5.5, in the past decade, different hepatic technique permitted the co-culture of fibroblasts
models were developed from microscale primary and hepatocytes cell lines in a spatially defined
rat and human hepatocyte culture chips (Lee organization resulting in a 3D liver construct.
et al. 2007) (Fig. 5.5-II) to a multi-walled liver Nevertheless, the evaluation of the influence
microfluidic device (Khetani and Bhatia 2007) of hepatic disorders on hepatocytes functions
(Fig. 5.5-I). More recently, Yu et al. (2017) cre- remained essential. Thus, microfluidic devices
ated a perfusion-incubator-liver-chip (PIC) able have been recently used for this purpose.
to maintain hepatocyte viability and functions Alcoholic liver disease (ALD), one of the major
for long-term 3D culture. This model could be cause of chronic liver disease worldwide, is char-
used for chronic hepatotoxicity testing (Fig. 5.5- acterized between others by hepatocytes apopto-
II). Therefore, liver-on-a-chip devices were fre- sis, activation of the hepatic stellate cells (HSCs)
quently developed to test the acute drug toxicity and alterations in the innate immunity, leading
impact through assessing pharmacokinetic the liver to fibrosis and in most of the cases to
parameters such as bioavailability, cell permea- cirrhosis (Gao and Bataller 2011). Zhou et al.
bility, and hepatic clearance. Nowadays, the developed, by the microfluidic co-culture
Fig. 5.5 Examples of microfluidics liver models. (I) Soft like barrier for primary hepatocyte culture. A. SEM
lithographic process for the fabrication of microscale liver micrograph illustrating the microfluidic sinusoid unit.
hepatocytes cultures (multi-well format) A. Schematic of B. Viability of rat primary hepatocytes when loaded at
the process flow aside photomicrographs taken at each high density after 7 days (Reprinted with permission (Lee
step. B. Photograph of a 24-well device with repeating et al. 2007) © 2007 Wiley Periodicals, Inc.). (III)
hepatic microstructures. C. Phase-contrast micrographs of Schematic of the perfusion-incubator-liver-chip (PIC).
micropatterned co-cultures (Reprinted with permission (Reprinted with permission (Yu et al. 2017) © The
(Khetani and Bhatia 2007) © 2007, Springer Nature). (II) Author(s) 2017 http://creativecommons.org/licenses/
An artificial liver sinusoid with a microfluidic endothelial- by/4.0/)
approach, a liver injury-on-a-chip device (Zhou can lead from simple steatosis to hepatocellular
et al. 2015b). The authors demonstrated that alco- carcinoma. Gori et al. fabricated the first in vitro
hol induced secretion of transforming growth on-chip model of NAFLD established within a
factor beta (TGF-β) by the injured hepatocytes microfluidic device in a sinusoid-like custom
resulting in a paracrine activation of the HSCs. (Gori et al. 2016) (Fig. 5.6). This model will
Non-alcoholic Fatty Liver Disease (NAFLD) was allow to understand the cellular, molecular and
also assessed using the microfluidic approach. epigenetic mechanisms that coordinate NAFLD
NAFLD, a chronic liver disease characterized by development.
the accumulation of lipids in the hepatocytes and Blood vessels are vital for organ survival since
related to insulin resistance (Leclercq et al. 2007), they manage the delivery of nutrients and oxygen
Fig. 5.6 Liver-on-a-Chip microfluidic model to investi- cytotoxicity after treatment of HepG2 cells with free fatty
gate NAFLD. (I) Microarchitecture and geometric con- acids (FFAs). (Reprinted with permission (Gori et al.
figuration of the NAFLD-on-a-chip device. (II) Evaluation 2016) © 2016 Gori et al. http://creativecommons.org/
of cell viability/cytotoxicity (A) and cell morphology (B) licenses/by/4.0/)
over a week in culture. (III) Evaluation of cell viability/
through the body (Ding et al. 2010). So, studies

have been focused in developing organ-on-chip 5.4 Conclusions
devices integrating the vasculature for further
application in liver tissue engineering (Zhang Liver, the largest organ of the body, has a central
et al. 2016). Moreover, recent investigations have role in the control of metabolic homeostasis and
focused on the development of vasculature-on-a- the xenobiotic metabolism from the body, associ-
chip devices for drug delivery applications in ated with endocrine and exocrine activities.
vivo (Goudie et al. 2016; Uguz et al. 2017). However, despite the high regenerative potential,
However, until now no implantable microfluidic liver injuries resulted in significant impact on
models have been established for liver regenera- functionality and regeneration capability leading
tion application. to liver failure. Nowadays, the only clinically
accepted therapeutic option for patients with References

advanced liver disease remains the orthotopic
liver transplantation, which has the disadvantage Adam R, Karam V et al (2012) Evolution of indications
and results of liver transplantation in Europe. A report
to be limited due to low organ availability. from the European Liver Transplant Registry (ELTR).
Therefore, currently, there is a continuous need J Hepatol 57(3):675–688
for alternative strategies. Agrawal P, Pramanik K (2016) Chitosan-poly(vinyl alco-
Over the last decade, several approaches have hol) nanofibers by free surface electrospinning for tis-
sue engineering applications. Tissue Eng Regen Med
been developed to improve or replace, even only 13(5):485–497
temporarily, the crucial liver functions in order to Ahadian S, Civitarese R et al (2017) Organ-on-a-Chip
increase patient’s life or to regenerate the hepatic platforms: A convergence of advanced materials,
tissue until find a compatible organ. These cells, and microscale technologies. Adv Healthc Mater
7(2):1700506
approaches include hepatocytes transplantation, Ahmed TAE, Dare EV et al (2008) Fibrin: a versatile scaf-
extracorporeal liver support and stem cells ther- fold for tissue engineering applications. Tissue Eng
apy. Besides those, liver tissue engineering and Part B Rev 14(2):199–215
regenerative medicine (TERM) strategies attempt Alan F-J, Catherine F et al (2015) Bioprinting of human
pluripotent stem cells and their directed differentiation
to forward prominently. Even if the choice of an into hepatocyte-like cells for the generation of mini-
adequate source of cells represents an important livers in 3D. Biofabrication 7(4):044102
step for liver TERM, the use of appropriate sup- Ananthanarayanan A, Narmada BC et al (2011) Purpose-
ports and the development of adequate culture driven biomaterials research in liver-tissue engineer-
ing. Trends Biotechnol 29(3):110–118
conditions were essential for future clinical Arisaka Y, Kobayashi J et al (2016) A heparin-modified
applications. Now, the use of natural and thermoresponsive surface with heparin-binding epi-
synthetic-based biomaterials were broadly dermal growth factor-like growth factor for main-
applied for the development of liver TERM strat- taining hepatic functions in vitro and harvesting
hepatocyte sheets. Regen Ther 3:97–106
egies. But, despite the potential and advantages Aycock RS, Seyer JM (1989) Collagens of normal and cir-
presented by both types, new biomaterials have rhotic human liver. Connect Tissue Res 23(1):19–31
been investigated. Then, new approaches have Bajaj P, Schweller RM et al (2014) 3D biofabrication
been exploiting the use of decellularized liver strategies for tissue engineering and regenerative med-
icine. Annu Rev Biomed Eng 16:247–276
matrices and nanomaterials showing both huge Banerjee A, Weaver I et al (2017) Bioelectronic measure-
potential to widely support the development of ment and feedback control of molecules in living cells.
advanced devices for clinical therapeutic applica- Sci Rep 7(1):12511
tions and TERM. Banihashemi M, Mohkam M et al (2015) Optimization of
three dimensional culturing of the HepG2 cell line in
Micro-technology approaches propose impor- fibrin scaffold. Hepat Mon 15(3):e22731
tant advantages in the creation of liver models, Barata D, van Blitterswijk C et al (2016) High-throughput
essentially in terms of increment of hepatic screening approaches and combinatorial development
microscale systems. Microfluidic platforms are of biomaterials using microfluidics. Acta Biomater
34:1–20
highly attractive and efficient technologies to Bhattacharjee N, Urrios A et al (2016) The upcoming
mimic liver microenvironment and microstruc- 3D-printing revolution in microfluidics. Lab Chip
ture. Therefore, several companies have as key 16(10):1720–1742
goal the development and commercialization of Bidault L, Deneufchatel M et al (2013) Self-supported
fibrin-polyvinyl alcohol interpenetrating polymer net-
chip devices, including organ-on-chip models, works: an easily handled and rehydratable biomaterial.
mainly for drug screening application. However, Biomacromolecules 14(11):3870–3879
until now, only few models have been designed Bierwolf J, Lutgehetmann M et al (2010) Primary rat
for in vivo liver regeneration assays. In the future, hepatocyte culture on 3D nanofibrous polymer scaf-
folds for toxicology and pharmaceutical research.
it is projected that construction and transplanta- Biotechnol Bioeng 108(1):141–150
tion of such kind of micro-structures may sustain Bishi DK, Mathapati S et al (2013) Trans-differentiation
or surpass many liver functions in order to be con- of human mesenchymal stem cells generates func-
sidered as a liver regenerative medicine approach. tional hepatospheres on poly(l-lactic acid)-co-poly(ε-
caprolactone)/collagen nanofibrous scaffolds. J Mater Collaborators, G. M. a. C. o. D (2015) Global, regional,

Chem B 1(32):3972 and national age–sex specific all-cause and cause-
Bissoyi A, Bit A et al (2016) Enhanced cryopreservation specific mortality for 240 causes of death, 1990–2013:
of MSCs in microfluidic bioreactor by regulated shear a systematic analysis for the global burden of disease
flow. Sci Rep 6:35416 study 2013. Lancet 385(9963):117–171
Bruinsma BG, Kim Y et al (2015) Layer-by-layer hepa- Collins MN, Birkinshaw C (2013) Hyaluronic acid based
rinization of decellularized liver matrices to reduce scaffolds for tissue engineering—a review. Carbohydr
thrombogenicity of tissue engineered grafts. J Clin Polym 92(2):1262–1279
Transl Res 1(1):04 Cozzolino AM, Noce V et al (2016) Modulating the sub-
Bruns H, Kneser U et al (2005) Injectable liver: a novel strate stiffness to manipulate differentiation of resident
approach using fibrin gel as a matrix for culture and liver stem cells and to improve the differentiation state
intrahepatic transplantation of hepatocytes. Tissue of hepatocytes. Stem Cells Int 2016:5481493
Eng 11(11–12):1718–1726 Croisier F, Jérôme C (2013) Chitosan-based biomaterials
Burdick JA, Stevens MM (2005) Chapter 11: Biomedical for tissue engineering. Eur Polym J 49(4):780–792
hydrogels. In: Hench LL, Jones JR (eds) Biomaterials, Damiati S, Kompella U et al (2018) Microfluidic devices
artificial organs and tissue engineering. Woodhead for drug delivery systems and drug screening. Genes
Publishing, Cambridge, pp 107–115 9(2):pii: E103
Burra P, Freeman R (2012) Trends in liver transplantation Dash M, Samal SK et al (2017) Enzymatically biominer-
2011. J Hepatol 56:S101–S111 alized chitosan scaffolds for tissue-engineering appli-
Caddeo S, Boffito M et al (2017) Tissue engineering cations. J Tissue Eng Regen Med 11(5):1500–1513
approaches in the design of healthy and pathological de la Puente P, Ludena D (2014) Cell culture in autolo-
in vitro tissue models. Front Bioeng Biotechnol 5:40 gous fibrin scaffolds for applications in tissue engi-
Caliari SR, Perepelyuk M et al (2016) Stiffening hydro- neering. Exp Cell Res 322(1):1–11
gels for investigating the dynamics of hepatic stellate Deng X, Cao Y et al (2015) Enhanced liver functions
cell mechanotransduction during myofibroblast acti- of HepG2 cells in the alginate/xyloglucan scaffold.
vation. Sci Rep 6:21387 Biotechnol Lett 37(1):235–240
Cao H, Kuboyama N (2010) A biodegradable porous Deponti D, Giancamillo AD et al (2013) Collagen scaf-
composite scaffold of PGA/β-TCP for bone tissue fold for cartilage tissue engineering: the benefit of
engineering. Bone 46(2):386–395 fibrin glue and the proper culture time in an infant car-
Catapano G, De Bartolo L et al (2001) Morphology and tilage model. Tissue Eng A 20(5–6):1113–1126
metabolism of hepatocytes cultured in Petri dishes Ding B-S, Nolan DJ et al (2010) Inductive angiocrine sig-
on films and in non-woven fabrics of hyaluronic acid nals from sinusoidal endothelium are required for liver
esters. Biomaterials 22(7):659–665 regeneration. Nature 468:310
Cen L, Liu W et al (2008) Collagen tissue engineering: Discher DE, Janmey P et al (2005) Tissue cells feel and
development of novel biomaterials and applications. respond to the stiffness of their substrate. Science
Pediatr Res 63(5):492–496 310(5751):1139
Chan Choi Y, Choi JS et al (2014) Human gelatin tissue- Dittrich PS, Manz A (2006) Lab-on-a-chip: microfluidics
adhesive hydrogels prepared by enzyme-mediated in drug discovery. Nat Rev Drug Discov 5:210
biosynthesis of DOPA and Fe3+ ion crosslinking. J Drosos I, Kolios G (2013) Stem cells in liver regeneration
Mater Chem B 2(2):201–209 and their potential clinical applications. Stem Cell Rev
Chan EC, Kuo S-M et al (2016) Three dimensional 9(5):668–684
collagen scaffold promotes intrinsic vascularisa- Dutkowski P, Linecker M et al (2015) Challenges to liver
tion for tissue engineering applications. PLoS One transplantation and strategies to improve outcomes.
11(2):e0149799 Gastroenterology 148(2):307–323
Chandrasekaran A, Abduljawad M et al (2016) Have Dvir-Ginzberg M, Gamlieli-Bonshtein I et al (2003) Liver
microfluidics delivered for drug discovery? Expert tissue engineering within alginate scaffolds: effects of
Opin Drug Discovery 11(8):745–748 cell-seeding density on hepatocyte viability, morphol-
Chen Q, Utech S et al (2016) Controlled assembly of het- ogy, and function. Tissue Eng 9(4):757–766
erotypic cells in a core-shell scaffold: organ in a drop- Ebrahimkhani MR, Neiman JAS et al (2014) Bioreactor
let. Lab Chip 16(8):1346–1349 technologies to support liver function in vitro. Adv
Cheng Z, Runzhou Z et al (2016) Hepatocyte growth Drug Deliv Rev 0:132–157
factor-loaded collagen-chitosan scaffold containing Elieh-Ali-Komi D, Hamblin MR (2016) Chitin and chito-
differentiated bone marrow-derived mesenchymal san: production and application of versatile biomedi-
stem cells as a model for hepatic tissue engineering. J cal nanomaterials. Int J Adv Res 4(3):411–427
Biomater Tissue Eng 6(8):621–628 Fan J, Yang J (2017) Preparation and characterization of a
Choi YJ, Yi HG et al (2017) 3D cell printed tissue ana- chitosan/galactosylated hyaluronic acid/heparin scaf-
logues: a new platform for Theranostics. Theranostics fold for hepatic tissue engineering. J Biomater Sci
7(12):3118–3137 Polym Ed 28(6):569–581
Feng Z-Q, Chu X et al (2009) The effect of nanofibrous Ho C-T, Lin R-Z et al (2013) Liver-cell patterning Lab
galactosylated chitosan scaffolds on the formation of Chip: mimicking the morphology of liver lobule tis-
rat primary hepatocyte aggregates and the maintenance sue. Lab Chip 13(18):3578–3587
of liver function. Biomaterials 30(14):2753–2763 Hughes RD, Mitry RR et al (2012) Current status of hepa-
Gao B, Bataller R (2011) Alcoholic liver disease: patho- tocyte transplantation. Transplantation 93(4):342–347
genesis and new therapeutic targets. Gastroenterology Ivanova EP, Bazaka K et al (2014) Chapter 3: Advanced
141(5):1572–1585 synthetic polymer biomaterials derived from organic
Garg K (2017) Laminin enriched scaffolds for tissue sources. In: Ivanova EP, Bazaka K, Crawford RJ (eds)
engineering applications. Adv Tissue Eng Regen Med New functional biomaterials for medicine and health-
Open Access 2(3):194–200 care. Woodhead Publishing, Oxford, pp 71–99
Gentile P, Chiono V et al (2014) An overview of Jain E, Damania A et al (2014) Biomaterials for liver tis-
poly(lactic-co-glycolic) acid (PLGA)-based bio- sue engineering. Hepatol Int 8(2):185–197
materials for bone tissue engineering. Int J Mol Sci Jain E, Damania A et al (2015) Fabrication of macropo-
15(3):3640–3659 rous cryogels as potential hepatocyte carriers for bio-
Gong H, Agustin J et al (2014) Biomimetic design and artificial liver support. Colloids Surf B: Biointerfaces
fabrication of porous chitosan–gelatin liver scaffolds 136:761–771
with hierarchical channel network. J Mater Sci Mater Jiang H-L, Kim Y-K et al (2010) Roles of spheroid forma-
Med 25(1):113–120 tion of hepatocytes in liver tissue engineering. Int J
Goral VN, Hsieh Y-C et al (2010) Perfusion-based micro- Stem Cells 3(2):69–73
fluidic device for three-dimensional dynamic primary Kane BJ, Zinner MJ et al (2006) Liver-specific functional
human hepatocyte cell culture in the absence of bio- studies in a microfluidic Array of primary mammalian
logical or synthetic matrices or coagulants. Lab Chip hepatocytes. Anal Chem 78(13):4291–4298
10(24):3380–3386 Kanninen LK, Harjumäki R et al (2016) Laminin-511 and
Gori M, Simonelli MC et al (2016) Investigating nonalco- laminin-521-based matrices for efficient hepatic spec-
holic fatty liver disease in a liver-on-a-Chip microflu- ification of human pluripotent stem cells. Biomaterials
idic device. PLoS One 11(7):e0159729 103:86–100
Goudie MJ, Ghuman AP et al (2016) Investigation of dif- Kanta J (2016) Elastin in the liver. Front Physiol 7:491.
fusion characteristics through microfluidic channels https://doi.org/10.3389/fphys.2016.00491
for passive drug delivery applications. J Drug Deliv Kazemnejad S (2009) Hepatic tissue engineering using
2016:7913616 scaffolds: state of the art. Avicenna J Med Biotechnol
Guvendiren M, Perepelyuk M et al (2014) Hydrogels 1(3):135–145
with differential and patterned mechanics to study Khademhosseini A, Langer R et al (2006) Microscale
stiffness-mediated myofibroblastic differentiation of technologies for tissue engineering and biology. Proc
hepatic stellate cells. J Mech Behav Biomed Mater Natl Acad Sci U S A 103(8):2480–2487
38:198–208 Khalaj Z, Lotfi AS et al (2016) Laminin matrix promotes
Hammond JS, Beckingham IJ et al (2006) Scaffolds for hepatogenic terminal differentiation of human bone
liver tissue engineering. Expert Rev Med Devices marrow mesenchymal stem cells. Iran J Basic Med Sci
3(1):21–27 19(1):34–42
Hammond JS, Gilbert TW et al (2011) Scaffolds contain- Khetani SR, Bhatia SN (2007) Microscale culture
ing growth factors and extracellular matrix induce of human liver cells for drug development. Nat
hepatocyte proliferation and cell migration in normal Biotechnol 26:120
and regenerating rat liver. J Hepatol 54(2):279–287 Kim TY, Kim DJ (2013) Acute-on-chronic liver failure.
Harink B, Le Gac S et al (2013) Regeneration-on-a-chip? Clin Mol Hepatol 19(4):349–359
The perspectives on use of microfluidics in regenera- Kim L, Vahey MD et al (2006) Microfluidic arrays for
tive medicine. Lab Chip 13(18):3512–3528 logarithmically perfused embryonic stem cell culture.
Hasirci V, Berthiaume F et al (2001) Expression of liver- Lab Chip 6(3):394–406
specific functions by rat hepatocytes seeded in treated Kim SM, Lee SH et al (2008a) Cell research with physi-
poly(lactic-co-glycolic) acid biodegradable foams. cally modified microfluidic channels: a review. Lab
Tissue Eng 7(4):385–394 Chip 8(7):1015–1023
Hemshekhar M, Thushara RM et al (2016) Emerging roles Kim I-Y, Seo S-J et al (2008b) Chitosan and its deriva-
of hyaluronic acid bioscaffolds in tissue engineer- tives for tissue engineering applications. Biotechnol
ing and regenerative medicine. Int J Biol Macromol Adv 26(1):1–21
86:917–928 Kim Y, Kang K et al (2017) Three-dimensional (3D)
Higashi T, Friedman SL et al (2017) Hepatic stellate cells printing of mouse primary hepatocytes to generate 3D
as key target in liver fibrosis. Adv Drug Deliv Rev hepatic structure. Ann Surg Treat Res 92(2):67–72
121:27–42 Kimura H, Sakai Y et al (2018) Organ/body-on-a-chip
based on microfluidic technology for drug discovery.
Drug Metab Pharmacokinet 33(1):43–48
Knight TA, Payne RG (2013) Characterization of a PGA- Liu WF, Chen CS (2005) Engineering biomaterials to
based scaffold for use in a tissue-engineered neo- control cell function. Mater Today 8(12):28–35
urinary conduit. In: Basu J, Ludlow JW (eds) Organ Liu M, Yang J et al (2016) Superior performance of co-
regeneration: methods and protocols. Humana Press, cultured mesenchymal stem cells and hepatocytes in
Totowa, pp 179–188 poly(lactic acid-glycolic acid) scaffolds for the treat-
Knight E, Przyborski S (2015) Advances in 3D cell cul- ment of acute liver failure. Biomed Mater (Bristol,
ture technologies enabling tissue-like structures to be UK) 11(1):015008
created in vitro. J Anat 227(6):746–756 Lou R, Yu W et al (2017) Fabrication of stable galacto-
Kumar A, Han SS (2017) PVA-based hydrogels for tis- sylated alginate microcapsules via covalent coupling
sue engineering: a review. Int J Polym Mater Polym onto hydroxyl groups for hepatocytes applications.
Biomater 66(4):159–182 Carbohydr Polym 155:456–465
Kundu B, Rajkhowa R et al (2013) Silk fibroin bioma- Lyu S, Untereker D (2009) Degradability of polymers
terials for tissue regenerations. Adv Drug Deliv Rev for implantable biomedical devices. Int J Mol Sci
65(4):457–470 10(9):4033–4065
Kwapiszewska K, Michalczuk A et al (2014) A Macdonald NP, Cabot JM et al (2017) Comparing micro-
microfluidic- based platform for tumour spheroid fluidic performance of three-dimensional (3D) print-
culture, monitoring and drug screening. Lab Chip ing platforms. Anal Chem 89(7):3858–3866
14(12):2096–2104 Magiera A, Markowski J et al (2017) PLA-based hybrid
Leclercq IA, Da Silva Morais A et al (2007) Insulin resis- and composite electrospun fibrous scaffolds as poten-
tance in hepatocytes and sinusoidal liver cells: mecha- tial materials for tissue engineering. J Nanomater
nisms and consequences. J Hepatol 47(1):142–156 2017:11
Lee JS, Cho S-W (2012) Liver tissue engineering: recent Makadia HK, Siegel SJ (2011) Poly lactic-co-glycolic
advances in the development of a bio-artificial liver. acid (PLGA) as biodegradable controlled drug deliv-
Biotechnol Bioprocess Eng 17(3):427–438 ery carrier. Polymers 3(3):1377–1397
Lee Hyeong J, Kim Yong B et al (2015) A new approach Malikmammadov E, Tanir TE et al (2018) PCL and PCL-
for fabricating collagen/ECM-based bioinks using based materials in biomedical applications. J Biomater
Preosteoblasts and human adipose stem cells. Adv Sci Polym Ed 29(7–9):863–893
Healthc Mater 4(9):1359–1368 Mari CE, Laura Saenz del B et al (2017) Gelatin as bio-
Lee KY, Mooney DJ (2012) Alginate: properties and biomaterial for tissue engineering. Curr Pharm Des
medical applications. Prog Polym Sci 37(1):106–126 23(24):3567–3584
Lee PJ, Hung PJ et al (2006) Nanoliter scale microbio- Mariod AA, Fadul H (2013) Review: gelatin, source,
reactor array for quantitative cell biology. Biotechnol extraction and industrial applications. Acta Sci Pol
Bioeng 94(1):5–14 Technol Aliment 12(2):135–147
Lee PJ, Hung PJ et al (2007) An artificial liver sinu- Martinez-Hernandez A, Amenta PS (1993) The hepatic
soid with a microfluidic endothelial-like barrier extracellular matrix. Virchows Arch A 423(2):77–84
for primary hepatocyte culture. Biotechnol Bioeng Miranda JP, Rodrigues A et al (2010) Extending hepa-
97(5):1340–1346 tocyte functionality for drug-testing applications
Lee Y-B, Polio S et al (2010a) Bio-printing of collagen using high-viscosity alginate–encapsulated three-
and VEGF-releasing fibrin gel scaffolds for neural dimensional cultures in bioreactors. Tissue Eng Part C
stem cell culture. Exp Neurol 223(2):645–652 Methods 16(6):1223–1232
Lee KH, Shin SJ et al (2010b) Microfluidic synthesis of Miranda-Nieves D, Chaikof EL (2017) Collagen and elas-
pure chitosan microfibers for bio-artificial liver chip. tin biomaterials for the fabrication of engineered liv-
Lab Chip 10(10):1328–1334 ing tissues. ACS Biomater Sci Eng 3(5):694–711
Lee Y, Kwon J et al (2012) Reduction of inflammatory Mironov Anton V, Grigoryev Aleksey M et al (2016) 3D
responses and enhancement of extracellular matrix printing of PLGA scaffolds for tissue engineering. J
formation by vanillin-incorporated poly(lactic- Biomed Mater Res A 105(1):104–109
co-glycolic acid) scaffolds. Tissue Eng Part A Nakamura S, Ijima H (2013) Solubilized matrix
18(19–20):1967–1978 derived from decellularized liver as a growth factor-
Lee J, Kim SH et al (2013) Fabrication and characteriza- immobilizable scaffold for hepatocyte culture. J
tion of microfluidic liver-on-a-chip using microsomal Biosci Bioeng 116(6):746–753
enzymes. Enzym Microb Technol 53(3):159–164 Nakamura S, Kubo T et al (2013) Heparin-conjugated
Lee BH, Shirahama H et al (2017) Colloidal templating gelatin as a growth factor immobilization scaffold. J
of highly ordered gelatin methacryloyl-based hydro- Biosci Bioeng 115(5):562–567
gel platforms for three-dimensional tissue analogues. Nallagangula KS, Nagaraj SK et al (2018) Liver fibrosis:
NPG Asia Mater 9:e412 a compilation on the biomarkers status and their sig-
Lin CH, Chang HC, Hsu CHA (2016) Microfluidic nificance during disease progression. Future Sci OA
platform for high-throughput single-cell isola- 4(1):FSO250
tion and culture. J Vis Exp 112:e54105. https://doi.
org/10.3791/54105
Naresh K, Utpal B (2012) Silk fibroin based biomimetic Shang Y, Tamai M et al (2014) Hybrid sponge comprised
artificial extracellular matrix for hepatic tissue engi- of galactosylated chitosan and hyaluronic acid medi-
neering applications. Biomed Mater 7(4):045004 ates the co-culture of hepatocytes and endothelial
Nikkhah M, Edalat F et al (2012) Engineering microscale cells. J Biosci Bioeng 117(1):99–106
topographies to control the cell–substrate interface. Shim J-H, Kim AJ et al (2013) Effect of solid freeform
Biomaterials 33(21):5230–5246 fabrication-based polycaprolactone/poly (lactic-co-
Olsen AL, Bloomer SA et al (2011) Hepatic stellate cells glycolic acid)/collagen scaffolds on cellular activities
require a stiff environment for myofibroblastic dif- of human adipose-derived stem cells and rat primary
ferentiation. Am J Physiol Gastrointest Liver Physiol hepatocytes. J Mater Sci Mater Med 24(4):1053–1065
301(1):G110–G118 Shirakigawa N, Takei T et al (2013) Base structure con-
Ostrovidov S, Jiang J et al (2004) Membrane-based PDMS sisting of an endothelialized vascular-tree network
microbioreactor for perfused 3D primary rat hepato- and hepatocytes for whole liver engineering. J Biosci
cyte cultures. Biomed Microdevices 6(4):279–287 Bioeng 116(6):740–745
Pan Z, Ding J (2012) Poly(lactide-co-glycolide) porous Sia SK, Whitesides GM (2003) Microfluidic devices fab-
scaffolds for tissue engineering and regenerative med- ricated in poly(dimethylsiloxane) for biological stud-
icine. Interface Focus 2(3):366–377 ies. Electrophoresis 24(21):3563–3576
Patel D, Haque A et al (2015) Using reconfigurable micro- Siegel R, Naishadham D et al (2012) Cancer statistics,
fluidics to study the role of HGF in autocrine and para- 2012. CA Cancer J Clin 62(1):10–29
crine signaling of hepatocytes. Integr Biol (Camb) Silva JM, Rodrigues LC et al (2017) Engineered tubular
7(7):815–824 structures based on chitosan for tissue engineering
Pensalfini M, Ehret AE et al (2017) Factors affecting the applications. J Biomater Appl 32(7):841–852
mechanical behavior of collagen hydrogels for skin Singal AK, Anand BS (2013) Recent trends in the epi-
tissue engineering. J Mech Behav Biomed Mater demiology of alcoholic liver disease. Clin Liver Dis
69:85–97 2(2):53–56
Perez RA, Jung CR et al (2017) Biomaterials and culture Sivaraman A, Leach JK et al (2005) A microscale in vitro
technologies for regenerative therapy of liver tissue. physiological model of the liver: predictive screens for
Adv Healthc Mater 6(2):1600791 drug metabolism and enzyme induction. Curr Drug
Poh Z, Goh BB et al (2015) Rates of cirrhosis and hepato- Metab 6(6):569–591
cellular carcinoma in chronic hepatitis B and the role Skardal A, Smith L et al (2012) Tissue specific synthetic
of surveillance: a 10-year follow-up of 673 patients. ECM hydrogels for 3-D in vitro maintenance of hepa-
Eur J Gastroenterol Hepatol 27(6):638–643 tocyte function. Biomaterials 33(18):4565–4575
Powers MJ, Domansky K et al (2002) A microfabri- Stange J (2011) Extracorporeal liver support.
cated array bioreactor for perfused 3D liver culture. Organogenesis 7(1):64–73
Biotechnol Bioeng 78(3):257–269 Stevens KR, Miller JS et al (2015) Degradable hydrogels
Protzer U, Maini MK et al (2012) Living in the liver: derived from PEG-diacrylamide for hepatic tissue engi-
hepatic infections. Nat Rev Immunol 12(3):201–213 neering. J Biomed Mater Res A 103(10):3331–3338
Rinaudo M (2006) Chitin and chitosan: properties and Stevens KR, Scull MA et al (2017) In situ expansion of
applications. Prog Polym Sci 31(7):603–632 engineered human liver tissue in a mouse model of
Rothbauer M, Zirath H et al (2018) Recent advances in chronic liver disease. Sci Transl Med 9(399):eaah5505
microfluidic technologies for cell-to-cell interaction Strohmeier O, Keller M et al (2015) Centrifugal microflu-
studies. Lab Chip 18(2):249–270 idic platforms: advanced unit operations and applica-
Schuerlein S, Schwarz T et al (2017) A versatile modular tions. Chem Soc Rev 44(17):6187–6229
bioreactor platform for tissue engineering. Biotechnol Stutchfield BM, Simpson K et al (2011) Systematic review
J 12(2):1600326 and meta-analysis of survival following extracorporeal
Schütte J, Hagmeyer B et al (2011) “Artificial micro liver support. Br J Surg 98(5):623–631
organs”—a microfluidic device for dielectrophoretic Su N, Gao P-L et al (2017) Fibrous scaffolds potentiate the
assembly of liver sinusoids. Biomed Microdevices paracrine function of mesenchymal stem cells: a new
13(3):493–501 dimension in cell-material interaction. Biomaterials
Semnani D, Naghashzargar E et al (2016) Evaluation of 141(Supplement C):74–85
PCL/chitosan electrospun nanofibers for liver tissue Sultana N, Mokhtar M et al (2015) Chitosan-based nano-
engineering. Int J Polym Mater 66(3):149–157 composite scaffolds for tissue engineering applica-
Seo S-J, Choi Y-J et al (2006) Alginate/Galactosylated tions. Mater Manuf Process 30(3):273–278
chitosan/heparin scaffold as a new synthetic extracel- Sun W, Chen Y-Q et al (2016) Organs-on-chips and its
lular matrix for hepatocytes. Tissue Eng 12(1):33–44 applications. Chin J Anal Chem 44(4):533–541
Shah Mohammadi M, Bureau MN et al (2014) Chapter Tang M, Wang G et al (2016) A review of biomedical
11: Polylactic acid (PLA) biomedical foams for tissue centrifugal microfluidic platforms. Micromachines
engineering. In: Netti PA (ed) Biomedical foams for 7(2):pii: E26
tissue engineering applications. Woodhead Publishing,
Cambridge, pp 313–334
Teotia RS, Kalita D et al (2015) Bifunctional Polysulfone- Xu T, Yang H et al (2017) Polylactic acid nanofiber scaf-
chitosan composite hollow fiber membrane for bioarti- fold decorated with chitosan Islandlike topography for
ficial liver. ACS Biomater Sci Eng 1(6):372–381 bone tissue engineering. ACS Appl Mater Interfaces
Toosi S, Naderi-Meshkin H et al (2016) PGA-incorporated 9(25):21094–21104
collagen: toward a biodegradable composite scaffold Yang J, Goto M et al (2002) Galactosylated alginate as
for bone-tissue engineering. J Biomed Mater Res A a scaffold for hepatocytes entrapment. Biomaterials
104(8):2020–2028 23(2):471–479
Török E, Lutgehetmann M et al (2010) Primary human Yu Y-Q, Jiang X-S et al (2014) Local delivery of vascu-
hepatocytes on biodegradable poly(l-lactic acid) lar endothelial growth factor via nanofiber matrix
matrices: a promising model for improving transplan- improves liver regeneration after extensive hepatec-
tation efficiency with tissue engineering. Liver Transpl tomy in rats. J Biomed Nanotechnol 10(11):3407–3415
17(2):104–114 Yu F, Deng R et al (2017) A perfusion incubator liver chip
Tourovskaia A, Figueroa-Masot X et al (2005) for 3D cell culture with application on chronic hepato-
Differentiation-on-a-chip: a microfluidic platform for toxicity testing. Sci Rep 7(1):14528
long-term cell culture studies. Lab Chip 5(1):14–19 Zavan B, Brun P et al (2005) Extracellular matrix-
Tripathi A, Melo JS (2015) Preparation of a sponge-like enriched polymeric scaffolds as a substrate for hepato-
biocomposite agarose-chitosan scaffold with primary cyte cultures: in vitro and in vivo studies. Biomaterials
hepatocytes for establishing an in vitro 3D liver tissue 26(34):7038–7045
model. RSC Adv 5(39):30701–30710 Zeigerer A, Wuttke A et al (2017) Functional properties
Uguz I, Proctor CM et al (2017) A microfluidic ion pump of hepatocytes in vitro are correlated with cell polarity
for in vivo drug delivery. Adv Mater 29(27):1701217 maintenance. Exp Cell Res 350(1):242–252
Vasanthan KS, Subramanian A et al (2012) Role of bio- Zhang Y, Zhang Y et al (2014) Galactosylated poly(epsilon-
materials, therapeutic molecules and cells for hepatic caprolactone) membrane promoted l iver-specific func-
tissue engineering. Biotechnol Adv 30(3):742–752 tions of HepG2 cells in vitro. Mater Sci Eng C Mater
Venkatesan J, Bhatnagar I et al (2015) Alginate compos- Biol Appl 41:52–58
ites for bone tissue engineering: a review. Int J Biol Zhang B, Montgomery M et al (2016) Biodegradable
Macromol 72:269–281 scaffold with built-in vasculature for organ-on-a-
Walker GM, Zeringue HC et al (2004) Microenvironment chip engineering and direct surgical anastomosis. Nat
design considerations for cellular scale studies. Lab Mater 15(6):669–678
Chip 4(2):91–97 Zhang J, Wei X et al (2017) Stem cell culture and dif-
Wang T, Feng Z-Q et al (2013) Nanoporous fibers of type- ferentiation in microfluidic devices toward organ-on-
I collagen coated poly(l-lactic acid) for enhancing pri- a-chip. Future Sci OA 3(2):FSO187
mary hepatocyte growth and function. J Mater Chem Zhang D, Wu X et al (2018) The development of colla-
B 1(3):339–346 gen based composite scaffolds for bone regeneration.
Wang B, Hu Q et al (2016) Porous lactose-modified chi- Bioact Mater 3(1):129–138
tosan scaffold for liver tissue engineering: influence of Zhao Y, Sha XUL et al (2011) In vitro and in vivo char-
galactose moieties on cell attachment and mechanical acterization of silk fibroin/gelatin composite scaffolds
stability. Int J Polym Sci 2016:8 for liver tissue engineering. J Dig Dis 13(3):168–178
Watanabe M, Zemack H et al (2016) Maintenance of Zhou P, Huang Y et al (2015a) Decellularization and recel-
hepatic functions in primary human hepatocytes lularization of rat livers with hepatocytes and endothe-
cultured on Xeno-free and chemical defined human lial progenitor cells. Artif Organs 40(3):E25–E38
recombinant laminins. PLoS One 11(9):e0161383 Zhou Q, Patel D et al (2015b) Liver injury-on-a-chip:
Willerth SM, Sakiyama-Elbert SE (2008) Combining microfluidic co-cultures with integrated biosensors for
stem cells and biomaterial scaffolds for constructing monitoring liver cell signaling during injury. Lab Chip
tissues and cell delivery. StemBook, Cambridge, MA 15(23):4467–4478
Xia L, Arooz T et al (2012) Hepatocyte function within Zhu Xin H, Gan Seng K et al (2007) Proteins combi-
a stacked double sandwich culture plate cylindrical nation on PHBV microsphere scaffold to regulate
bioreactor for bioartificial liver system. Biomaterials Hep3B cells activity and functionality: a model of
33(32):7925–7932 liver tissue engineering system. J Biomed Mater Res
A 83A(3):606–616
Microfluidic Systems in CNS
Studies 6
Anna Andrzejewska and Miroslaw Janowski
Abstract Keywords
Current technological progress facilitates the Microfluidic chambers · Microfluidic
introduction of micro-devices into biotechnology channels · PDMS · CNS
research including studies on central nervous sys-
tem. Wide range of micro-chambers with diver-
sity of channel systems and multiple compartments
enable users to create models which closely
mimic nervous tissue structure which nowadays is Highlights
often called as brain-on-a- chip technology.
Heretofore experiments showing the influence of • Microfluidic devices consist of channel sys-
substance gradients, cell interactions, spatial con- tems which enable to work with a great accu-
ditions, neuroinflammation, stem cells migration, racy simultaneously reducing time and cost of
drug delivery, mechanisms controlling progres- experiments performance
sion of diseases like Alzheimer, Parkinson, multi- • Microfluidic systems better mimic natural
ple sclerosis or nerve injury were performed in brain tissue homeostasis than conventional
microfluidic devices. Moreover, the integration of cell culture due to assuring of constant flow of
bio-sensors and development of dedicated soft- culture medium and discharge of unnecessary
ware for microfluidic studies can enable perform- products, providing natural cells spatial orga-
ing high throughput and good quality automated nization and inducing shear stress
experiments investigating regeneration and • Nowadays microfluidic systems are
degeneration processes in models well emulating employed in research investigating
central nervous system structures. pathomechanisms of Alzheimer disease,
Parkinson disease, multiple sclerosis, nerve
A. Andrzejewska degeneration, neuro-inflammation and brain
NeuroRepair Department, Mossakowski Medical metastasis as well as in attempts of develop-
Research Centre, Polish Academy of Sciences, ment of new therapeutic strategies
Warsaw, Poland
e-mail: aandrzejewska@imdik.pan.pl
M. Janowski (*)
Center for Advanced Imaging Research, Department
of Diagnostic Radiology and Nuclear Medicine,
University of Maryland, Baltimore, MD, USA
e-mail: miroslaw.janowski@som.maryland.edu

https://doi.org/10.1007/978-3-030-36588-2_6
88 A. Andrzejewska and M. Janowski
6.1 Introduction technology allowed creating very complex

micro-channel systems including one and multi
Microfluidic devices (micro-devices) consist of compartment chambers connected with each
channel systems with a cross-section in range of other by channels. These systems may be hori-
a few micrometres, and historically have been zontally or vertically oriented, and possess singu-
developed as a tool for chemical analysis hence lar or many inputs and outputs. All these
they enable to work with a great accuracy using combinations were developed in an effort to cre-
even very small volumes of reagents (Harrison ate models that best reflects the in vivo situation
et al. 1993). Micro-devices save time required to what nowadays is often called as “organ-on-a-
run experiments, as well as significantly reduce chip” technology (Prabhakarpandian et al. 2013;
their costs mainly due to low consumption of Brown et al. 2015; Wang et al. 2017a). This task
often very expensive reagents. The same advan- is even more demanding considering the chal-
tages have led to adaptation of microfluidic sys- lenge of recreation of the most complicated and
tems in a wide variety of disciplines including inscrutable system of human body – central ner-
biotechnology. Technological progress provided vous system (CNS).
a new set of materials for micro-device produc-
tion which fit biological applications. The most
commonly used micro-chambers are fabricated 6.2 Microfludics in CNS
by photolithography method from polydimethyl-
siloxane (PDMS). Photolithography is the pro- The microfluidic systems possess a few advan-
cess used to obtain a variety of channel shape tages over standard two- or even three dimen-
projection in micro-chamber. Briefly, the UV sional culture system what make them more
light exposure causes the etching of PDMS, appropriate to mimic CNS-like microenviron-
therefore in order to obtain the desired micro- ment. Constant flow of culture medium contain-
pattern, the PDMS is placed on the photo-stable ing nutritional compounds together with
surface (like glass or silicone), and then irradi- simultaneous discharge of unnecessary products
ated by UV light through the layer of photoresis- of cell metabolism allow obtaining proper condi-
tant mask of particular shape. This leads to the tions for nervous tissue homeostasis which is not
mapping of the mask shape in the PDMS (Das provided in conventional cell culture methods.
and Srivastava 2016) (Fig. 6.1). Low cost and The constant flow provides also stability of solu-
highly reproducible fabrication of PDMS cham- ble factors gradient in studies aiming to investi-
bers fulfils mass production requirements. PDMS gate their influence on cellular response. Taking
is a synthetic polymer with low auto-fluorescence advantage of this property of microfluidic sys-
and good transparency facilitating its use in fluo- tem, the influence of retinoic acid and sonic
rescent microscope-based imaging. Moreover, hedgehog (SHH) gradient on differentiation of
PDMS is gas-permeable and biocompatible neural stem cells was studied in analogy to phe-
therefore it is well suited for cell culture experi- nomenon of neural tube formation during foetal
ments, however due to hydrophobicity it is not development (Uzel et al. 2016). Similar analysis
favourable for cell adhesion. Heretofore the spec- was made for concentration gradient of SHH,
trum of surface modification techniques has been BMP4 (bone morphogenetic protein 4) and FGF8
employed in order to enable cell growth inside (fibroblast growth factor 8) regarding to their role
PDMS chambers including plasma treatment or in neural progenitor differentiation into neurons
coating with functional groups and/or extracellu- and generation of neural networks (Park et al.
lar matrix. Alternatively in some studies we can 2009). The ability of factors like netrin-1 and
found micro-devices produced from natural sub- slit-2 in axonal guidance was investigated ana-
strates like collagen, fibrin or agarose which bet- logically (Kothapalli et al. 2011) (Fig. 6.2).
ter mimic the natural conditions (Zhou et al. Another interesting improvement in analysis
2012; Berthier et al. 2012). Currently available of CNS cells physiology through utilization of
6 Microfluidic Systems in CNS Studies 89
Fig. 6.1 The schematic overview of photolithography technique
Fig. 6.2 The example of systems of channels providing cells exposition on gradient of factors in microfluidic
chamber
micro-devices is the possibility to isolate in dis- mation study, myelination process analysis and
tinct compartments soma from axons and den- investigation on transmission of axonal signals
drites of singular neurons by micro-grooves and (Neto et al. 2016). Interactions between different
trace the signalling between them. This system types of CNS cells were observed in microfluidic
was used to assess the gene expression changes systems, too. The complex net of mutual physical
in neuronal nuclei after BDNF-treatment of their and functional relations existing among neurons,
dendrites (Cohen et al. 2011). The separation of astrocytes, oligodendrocytes, microglia cells,
axons from soma has also showed that neurothro- pericytes and endothelial cells have led to cre-
phin-3 and BDNF provide strong stimulation sig- ation of neurovascular unit (NVU) concept. It is
nal for axon elongation independently from the basic entity providing proper function of CNS
exposition of cell body (Taylor et al. 2005). and damage of even one of its element induces
Similar strategies have been used for synapse for- dysfunction of the whole system. Thus all studies
investigating separated parts of NVU are biased The presence of blood-brain-barrier (BBB) is one
by the lack of crucial factors provided by each of the most unique features of CNS. In physio-
cells of this entity (Muoio et al. 2014). The logical state its existence is truly beneficial since
majority of NVU models are based on cell co- it protects the CNS from harmful substances
culture method, however this type of systems diluted in blood, however at the same time during
does not provide natural spatial organization of pathological process it creates an obstacle in drug
cells. Much closer mapping of natural CNS cell delivery following systemic administration.
configuration was gained by the use of vertically Many theoretically effective substances failed in
situated two-compartments microfluidic chamber treatment of CNS diseases due to their inability
system created by Bown and co-workers, where to cross BBB. This makes the challenge of cre-
endothelial cells adhered to the walls of lower ation of proper BBB model in vitro to become the
chamber separated by semi-permeable mem- high priority issue in order to perform high-
brane from upper compartment containing layer throughput screening of potential CNS penetrat-
of pericytes and astrocytes together with neurons ing drugs. Heretofore transwell system was
embedded in three-dimensional collagen gel mainly adopted to obtain this goal however the
(Brown et al. 2015). Noteworthy work has also layer of brain endothelial cells placed inside of
been made by Kim et al., who created a model of transwell insert do not create intracellular con-
hippocampal neural network with architecture of nections analogical to this observed in BBB. It
parallel axon projection between CA3 and CA1 might be related lack of shear stress which is
regions maintained due to use of microfluidic naturally induced by the blood flow through ves-
chamber with properly aligned channel systems sels. It was shown that endothelial cells growing
in monolithic gel (Kim et al. 2017). The spatial in microfluidic chamber under influence of con-
and bio-mechanical cues delivered to the cells stant perfusion inducing shear stress reveal fea-
were shown to play an important role in neural tures reflecting the BBB structure much more
stem cells (NSCs) fate determination. Wang and closely including stronger expression of tight
co-authors created micro-device containing 16 junction proteins like ZO-1 and claudin-1
separated micro-wells reflecting cellular niches (Prabhakarpandian et al. 2013) and higher values
which allowed them to simultaneously investi- of trans-endothelial electrical resistance (TEER)
gate different culture conditions for neural stem (Takeshita et al. 2014). The functionality of rec-
cells differentiation including variables like two reated in microfluid device BBB endothelial
and three dimensional culture or static and perfu- layer was also confirmed by its responsiveness to
sion state. This study revealed that three dimen- the presence of mannitol, urea and histamine
sional culture in the presence of extracellular solutions causing drop of TEER value (Takeshita
matrix proteins stimulates self-renewal of NSCs, et al. 2014; Wang et al. 2016). Microfluidic BBB
whereas two-dimensional and spheroid cultures model has already been proven to be useful for
induce their differentiation under continuous per- screening BBB permeability of drugs such as caf-
fusion conditions (Wang et al. 2017a). Among feine, cimetidine or doxorubicin (Wang et al.
other of the crucial mechanical factors which are 2017b). Alternatively, the same model of BBB
better mimicked by the microfluidic system than can be used to study neuro-inflammation by addi-
by the conventional culture are more natural spa- tion of pro-inflammatory cytokine to the chamber
tial ratio between nutrition availability and tissue to induce activation of endothelial cells. The flow
area as well as presence of shear stress. The of leukocytes through microfluidic channel cov-
exchange of oxygen and nutrients delivered by ered by activated endothelial cells perfectly
fluid in micro-channels take place by diffusion mimics the process of leukocyte extravasation
mechanism analogical to in vivo process what is (Cho et al. 2015) (Fig. 6.3). Next possible appli-
in opposite in other culture systems where cells cation of BBB microfluidic model is to emulate
are immersed in medium (Harink et al. 2013). the process of metastasis (Xu et al. 2016).
Fig. 6.3 Movement patterns of Mesenchymal Stem Cells charts were automatically generated by dedicated soft-
flowing through microfluidic chamber in analogy to steps ware and show examples of velocity variation typical for
of leucocytes diapedesis process observed in a VCAM-1 unstopped (a), rolling (b), arrested (c), and crawling (d)
protein covered channel of microfluidic chamber. Line cells. (Reprinted from Andrzejewska et al. 2019)
Moreover this system can be utilized to predict et al. 2013). Similar technology was used in com-
the efficacy of delivery and homing of stem cells bination with culture of neurospheroids, which
like mesenchymal stem cells or glia progenitors showed that amyloid β can be more toxic for cells
to CNS structures after their systemic administra- causing their death and interrupting neural cir-
tion as a therapeutic approach in neurological cuits, if it is delivered via interstitial perfusion in
disorders (Gorelik et al. 2012; Jablonska et al. comparison to static conditions (Park et al. 2015).
2018; Andrzejewska et al. 2019). Very important This experiment highlights the huge additional
area of CNS research where microfluid device cognitive value of research performed in environ-
can be employed is investigation of pathophysi- ment better mimicking the natural brain tissue
ology of neurodegenerative diseases. Beyond the homeostasis like microfluidic systems than con-
blood flow imitation, the stream of fluid in micro- ventional culture. Through displaying neurons
chamber can reflect a much slower brain intersti- culture in microfluid device containing separated
tial fluid perfusion. This system was resourcefully compartments for soma and axons the neuron-
used in study by Choi and co-workers to analyse neuron and axon to soma transport of amyloid β
the neurotoxicity of amyloid beta oligomeric was identified (Song et al. 2014). Movement of
assemblies gradient on neurons as an equivalent Parkinson disease pathognomonic protein –
of Alzheimer disease progression model (Choi α-synuclein between neurons and along axons
was also shown using microfluidic device through this area (Dollé et al. 2013) (Fig. 6.4).
(Freundt et al. 2012). Except of amyloid plaques Alternatively bottom chamber may be replaced
formation another feature of Alzheimer disease is by smaller sectors placed inside of the main
the impairment of microglia function. The accu- chamber, which deformation induced by pressure
mulation and migration of these cells toward increase causes crushing of adjacent axons
soluble or bound amyloid β fraction was also (Hosmane et al. 2011). By manipulation of
assessed in microfluidic device and revealed the applied pressure intensity and its duration the
difference in mechanisms responsible for microg- different level of severity of nerve injury can be
lia chemotaxis induced by higher and lower con- obtained in very reproducible manner (Dollé
centration of particular forms of this protein. et al. 2013). These techniques allowed researches
Among others the researchers found that migra- to reveal the existence of difference in sensitivity
tion of microglia can be inhibited by high amount on compression-injury between particular types
of bound form of amyloid β, which can have a of neurons such as hippocampal neurons and dor-
significant impact on development of therapies of sal root ganglion neurons (Magdesian et al.
Alzheimer disease (Cho et al. 2013a). 2012). Another method of axonal damage relies
Furthermore, microfluidic device was used as a on administration of chemical substances to axo-
platform for investigation of mechanism of dam- nal compartment in microfluidic chamber.
age caused by CD8+ lymphocytes to interferon Exposure to particular toxins can mimic different
gamma activated cortical neuronal axons in anal- in vivo situation. The infusion of excess of neu-
ogy to process occurring in multiple sclerosis. rotransmitters like glutamate was used as a model
During development of this neurodegenerative of excitotoxicity which is a cause of secondary
disease axonal injury is induced primarily only damage in stroke (Samson et al. 2016), in addi-
with subsequent cell body disruption. The appli- tion to blood emulated haemorrhage-related axo-
cation of multi-chamber microfluidic device with nal injury (Charier et al. 2015), whereas the
separated soma and axon compartments allowed treatment with detergent like Triton X-100 was a
researchers to perform study providing more model for chemically-induced axotomy
accurate system for proper event sequence, where (Deleglise et al. 2013). Axon dissection can be
axons could be attacked first being isolated from also very efficiently obtained inside of microflu-
cells’ body (Sauer et al. 2013). Taking advantage idic chamber by laser based injury which is an
of possibility to culture even singular neuron in immensely precise and high throughput method
chambers with soma and axon division this sys- however it demands the availability of more
tem was widely adapted to study nerve injury and sophisticated equipment (Yanik et al. 2004). All
regeneration process. Currently a few methods of these platforms gave us a unique opportunity to
induction of nerve degeneration in microfluidic study the degeneration and regeneration process
devices are available which reflect the differences of axons which were exposed to toxins or treat-
in injury mechanisms in vivo. The physical injury ment independently from cellular body.
of axons like those created by stretching or strain- Moreover, cell culture in micro-devices can be
ing is mirrored by application of pneumatic- easily combined with assessment of fractions
stress or direct compression applied within (like somas or axons) isolated from particular
micro-chambers. These types of systems can be chamber compartments by analytical techniques.
obtained by addition of bottom chamber passing The material acquired from microfluidic devices
under axons projections. This chamber is covered after cell culture was proven to be suitable for
by flexible PDMS membrane which upward mRNA analysis, microarrays, biochemical assays
deformation is induced by increase of internal and morphology assessment by fluorescent and
pressure in bottom chamber. This elongate the Atomic Force Microscopy (Park et al. 2006;
interval distance between upper chamber com- Taylor et al. 2009; Magdesian et al. 2012).
partments and cause the tension of axons passing
Fig. 6.4 The mechanism of induction of model of nerve injury caused by stretching or straining created in microfluidic
chamber
The fusion of technologies is critical for fur- spatial as well as functional aspects. In turn, this
ther progress in immersing microfluidic devices approach requires very close collaboration of
within CNS studies. Implementation of sensors specialists from different disciplines such physi-
like electrodes inside the microfluidic chambers cists, material scientists, biotechnologists and
can enable users to extend the range of gained software engineers.
information with neurons functionality experi-
ments as these cells exert electrical and electro-
chemical activity (Cho et al. 2013b; Kafi et al. 6.3 Conclusions
2016) and to provide the possibility of charac-
terisation of compounds on which they react on Implementation of new technologies in CNS
(MacKerron et al. 2017). Another possible appli- studies is indispensable since the broadening of
cation of additional biosensors is connected with our knowledge introduces more and more vari-
very detailed environment control including ables which are shown to have significant impact
parameters like temperature, metabolic activity, on biological processes. Recent discoveries indi-
oxygen level (Harink et al. 2013), or sensation of cate on spatial arrangement, fluid pressure and
TEER value which was already mentioned above chemical – physical surface properties as factors
(Wang et al. 2017b). All these methods require determining the fate of cells and course of the
the computer control of the experimental condi- regeneration process. Microfluidic systems allow
tions. Therefore the development of high quality us to control most of these parameters thus meet-
dedicated software systems is necessary to ing current challenges in recreation of CNS
obtain quantitative results from high throughput structures and functionality.
studies, which currently are not broadly avail-
able on market (Andrzejewska et al. 2019). The Acknowledgments
This work was supported by
future directions of microfluidic device develop- NCR&D STRATEGMED1/235773/19/NCBR/2016
“EXPLORE ME” project and NIH R01NS091100.
ment should be linked with multi-electrode
assays and three- dimensional printing tech-
niques in order to obtain optimal platforms to
mimic such complicated systems like CNS in
References Gorelik M, Orukari I, Wang J et al (2012) Use of MR cell

tracking to evaluate targeting of glial precursor cells
to inflammatory tissue by exploiting the very late
Andrzejewska A, Nowakowski A, Grygorowicz T, et al
antigen-4 docking receptor. Radiology 265:175–185.
(2019) Single-cell, high-throughput analysis of cell dock-
https://doi.org/10.1148/radiol.12112212
ing to vessel wall. J Cereb Blood Flow Metab 39:2308–
Harink B, Le Gac S, Truckenmüller R et al (2013)
2320. https://doi.org/10.1177/0271678X18805238
Regeneration-on-a-chip? The perspectives on use
Berthier E, Young EWK, Beebe D (2012) Engineers are
of microfluidics in regenerative medicine. Lab Chip
from PDMS-land, biologists are from Polystyrenia.
13:3512–3528. https://doi.org/10.1039/c3lc50293g
Lab Chip 12:1224–1237. https://doi.org/10.1039/
Harrison DJ, Glavina PG, Manz A (1993) Towards
C2LC20982A
miniaturized electrophoresis and chemical analy-
Brown JA, Pensabene V, Markov DA et al (2015)
sis systems on silicon: an alternative to chemical
Recreating blood-brain barrier physiology and struc-
sensors. Sens Actuators B 10:107–116. https://doi.
ture on chip: a novel neurovascular microfluidic
org/10.1016/0925-4005(93)80033-8
bioreactor. Biomicrofluidics 9:054124. https://doi.
Hosmane S, Fournier A, Wright R et al (2011) Valve-
org/10.1063/1.4934713
based microfluidic compression platform: single axon
Charier D, Beauchet O, Bell M et al (2015) Memantine
injury and regrowth. Lab Chip 11:3888–3895. https://
plus vitamin D prevents axonal degeneration caused
doi.org/10.1039/c1lc20549h
by lysed blood. ACS Chem Neurosci 6:393–397.
Jablonska A, Shea DJ, Cao S et al (2018) Overexpression
https://doi.org/10.1021/cn500303k
of VLA-4 in glial-restricted precursors enhances their
Cho H, Hashimoto T, Wong E et al (2013a) Microfluidic
endothelial docking and induces diapedesis in a mouse
chemotaxis platform for differentiating the roles of
stroke model. J Cereb Blood Flow Metab 38:835–846.
soluble and bound amyloid-β on microglial accu-
https://doi.org/10.1177/0271678X17703888
mulation. Sci Rep 3:1823. https://doi.org/10.1038/
Kafi MA, Cho H-Y, Choi J-W (2016) Engineered peptide-
srep01823
based nanobiomaterials for electrochemical cell
Cho H-Y, Eun-Bi-Ko, Kim T-H, Choi J-W (2013b)
chip. Nano Converg 3:17. https://doi.org/10.1186/
Fabrication of carbon nanotubes/RGD peptide com-
s40580-016-0077-7
posites to enhance electrochemical performance of
Kim SH, Im S-K, Oh S-J et al (2017) Anisotropically
cell chip. J Biomed Nanotechnol 9:1398–1402
organized three-dimensional culture platform for
Cho H, Seo JH, Wong KHK et al (2015) Three-
reconstruction of a hippocampal neural network.
dimensional blood-brain barrier model for in vitro
Nat Commun 8:14346. https://doi.org/10.1038/
studies of neurovascular pathology. Sci Rep 5:15222.
ncomms14346
https://doi.org/10.1038/srep15222
Kothapalli CR, van Veen E, de Valence S et al (2011) A
Choi YJ, Chae S, Kim JH et al (2013) Neurotoxic amyloid
high-throughput microfluidic assay to study neurite
beta oligomeric assemblies recreated in microfluidic
response to growth factor gradients. Lab Chip 11:497–
platform with interstitial level of slow flow. Sci Rep
507. https://doi.org/10.1039/c0lc00240b
3:1921. https://doi.org/10.1038/srep01921
MacKerron C, Robertson G, Zagnoni M, Bushell TJ
Cohen MS, Bas Orth C, Kim HJ et al (2011) Neurotrophin-
(2017) A microfluidic platform for the characterisation
mediated dendrite-to-nucleus signaling revealed by
of CNS active compounds. Sci Rep 7:15692. https://
microfluidic compartmentalization of dendrites. Proc
doi.org/10.1038/s41598-017-15950-0
Natl Acad Sci U S A 108:11246–11251. https://doi.
Magdesian MH, Sanchez FS, Lopez M et al (2012) Atomic
org/10.1073/pnas.1012401108
force microscopy reveals important differences in
Das S, Srivastava VC (2016) Microfluidic-based photo-
axonal resistance to injury. Biophys J 103:405–414.
catalytic microreactor for environmental application:
https://doi.org/10.1016/j.bpj.2012.07.003
a review of fabrication substrates and techniques,
Muoio V, Persson PB, Sendeski MM (2014) The neuro-
and operating parameters. Photochem Photobiol Sci
vascular unit – concept review. Acta Physiol Oxf Engl
15:714–730. https://doi.org/10.1039/c5pp00469a
210:790–798. https://doi.org/10.1111/apha.12250
Deleglise B, Lassus B, Soubeyre V et al (2013) Synapto-
Neto E, Leitão L, Sousa DM et al (2016)
protective drugs evaluation in reconstructed neuronal
Compartmentalized microfluidic platforms: the unri-
network. PLoS One 8:e71103. https://doi.org/10.1371/
valed breakthrough of in vitro tools for neurobiologi-
journal.pone.0071103
cal research. J Neurosci 36:11573–11584. https://doi.
Dollé J-P, Morrison B, Schloss RS, Yarmush ML (2013)
org/10.1523/JNEUROSCI.1748-16.2016
An organotypic uniaxial strain model using microflu-
Park JW, Vahidi B, Taylor AM et al (2006) Microfluidic
idics. Lab Chip 13:432–442. https://doi.org/10.1039/
culture platform for neuroscience research. Nat Protoc
c2lc41063j
1:2128–2136. https://doi.org/10.1038/nprot.2006.316
Freundt EC, Maynard N, Clancy EK et al (2012) Neuron-
Park JY, Kim S-K, Woo D-H et al (2009) Differentiation of
to-neuron transmission of α-synuclein fibrils through
neural progenitor cells in a microfluidic c hip-generated
axonal transport. Ann Neurol 72:517–524. https://doi.
cytokine gradient. Stem Cells 27:2646–2654. https://
org/10.1002/ana.23747
doi.org/10.1002/stem.202
Park J, Lee BK, Jeong GS et al (2015) Three-dimensional mammalian axons. J Neurosci 29:4697–4707. https://
brain-on-a-chip with an interstitial level of flow and doi.org/10.1523/JNEUROSCI.6130-08.2009
its application as an in vitro model of Alzheimer’s dis- Uzel SGM, Amadi OC, Pearl TM et al (2016) Simultaneous
ease. Lab Chip 15:141–150. https://doi.org/10.1039/ or sequential orthogonal gradient formation in a 3D
c4lc00962b cell culture microfluidic platform. Small 12:612–622.
Prabhakarpandian B, Shen M-C, Nichols JB et al (2013) https://doi.org/10.1002/smll.201501905
SyM-BBB: a microfluidic blood brain barrier model. Wang JD, Khafagy E-S, Khanafer K et al (2016)
Lab Chip 13:1093–1101. https://doi.org/10.1039/ Organization of endothelial cells, pericytes, and astro-
c2lc41208j cytes into a 3D microfluidic in vitro model of the
Samson AJ, Robertson G, Zagnoni M, Connolly CN blood-brain barrier. Mol Pharm 13:895–906. https://
(2016) Neuronal networks provide rapid neuropro- doi.org/10.1021/acs.molpharmaceut.5b00805
tection against spreading toxicity. Sci Rep 6:33746. Wang Y, Ma J, Li N et al (2017a) Microfluidic engi-
https://doi.org/10.1038/srep33746 neering of neural stem cell niches for fate deter-
Sauer BM, Schmalstieg WF, Howe CL (2013) Axons are mination. Biomicrofluidics 11:014106. https://doi.
injured by antigen-specific CD8(+) T cells through a org/10.1063/1.4974902
MHC class I- and granzyme B-dependent mechanism. Wang YI, Abaci HE, Shuler ML (2017b) Microfluidic
Neurobiol Dis 59:194–205. https://doi.org/10.1016/j. blood-brain barrier model provides in vivo-like barrier
nbd.2013.07.010 properties for drug permeability screening. Biotechnol
Song H-L, Shim S, Kim D-H et al (2014) β-Amyloid Bioeng 114:184–194. https://doi.org/10.1002/
is transmitted via neuronal connections along axo- bit.26045
nal membranes. Ann Neurol 75:88–97. https://doi. Xu H, Li Z, Yu Y et al (2016) A dynamic in vivo-like
org/10.1002/ana.24029 organotypic blood-brain barrier model to probe met-
Takeshita Y, Obermeier B, Cotleur A et al (2014) An astatic brain tumors. Sci Rep 6:36670. https://doi.
in vitro blood-brain barrier model combining shear org/10.1038/srep36670
stress and endothelial cell/astrocyte co-culture. Yanik MF, Cinar H, Cinar HN et al (2004) Neurosurgery:
J Neurosci Methods 232:165–172. https://doi. functional regeneration after laser axotomy. Nature
org/10.1016/j.jneumeth.2014.05.013 432:822. https://doi.org/10.1038/432822a
Taylor AM, Blurton-Jones M, Rhee SW et al (2005) A Zhou J, Khodakov DA, Ellis AV, Voelcker NH (2012)
microfluidic culture platform for CNS axonal injury, Surface modification for PDMS-based microflu-
regeneration and transport. Nat Methods 2:599–605. idic devices. Electrophoresis 33:89–104. https://doi.
https://doi.org/10.1038/nmeth777 org/10.1002/elps.201100482
Taylor AM, Berchtold NC, Perreau VM et al (2009)
Axonal mRNA in uninjured and regenerating cortical
Microfluidics for Angiogenesis
Research 7
Lígia Costa, Rui Luís Reis, Joana Silva-Correia,
Abstract esis has been a subject of major interest among

Angiogenesis is a natural and vital phenome- the scientific community, being transverse to
non of neovascularization that occurs from different areas and founding particular atten-
pre-existing vasculature, being present in many tion in tissue engineering and cancer research
physiological processes, namely in develop- fields. Microfluidics has emerged as a power-
ment, reproduction and regeneration. Being a ful tool for modelling this phenomenon,
highly dynamic and tightly regulated process, thereby surpassing the limitations associated
its abnormal expression can be on the basis of to conventional angiogenic models. Holding a
several pathologies. For that reason, angiogen- tremendous flexibility in terms of experimental
design towards a specific goal, microfluidic
systems can offer an unlimited number of
opportunities for investigating angiogenesis in
many relevant scenarios, namely from its fun-
damental comprehension in normal physiolog-
L. Costa · J. Silva-Correia (*) ical processes to the identification and testing
3B’s Research Group, I3Bs – Research Institute on of new therapeutic targets involved on patho-
University of Minho, Headquarters of the European logical angiogenesis. Additionally, microvas-
Institute of Excellence on Tissue Engineering and cular 3D in vitro models are now opening up
Regenerative Medicine, Barco, Guimarães, Portugal new prospects in different fields, being used
ICVS/3B’s – PT Government Associate Laboratory, for investigating and establishing guidelines
Braga, Guimarães, Portugal for the development of next generation of 3D
e-mail: joana.correia@i3bs.uminho.pt functional vascularized grafts. The promising
R. L. Reis · J. M. Oliveira applications of this emerging technology in
3B’s Research Group, I3Bs – Research Institute on angiogenesis studies are herein overviewed,
University of Minho, Headquarters of the European encompassing fundamental and applied
Institute of Excellence on Tissue Engineering and research.
ICVS/3B’s PT Government Associate Lab, Keywords
Braga, Guimarães, Portugal Angiogenesis · Sprouting · Endothelization ·
The Discoveries Centre for Regenerative and Microfluidics · Microvascular models ·
Precision Medicine, Headquarters at University Tumour angiogenesis

https://doi.org/10.1007/978-3-030-36588-2_7
98 L. Costa et al.
Highlights scaffolds before implantation, thus can help

reducing the in vivo period of hypoxia and simul-
• Microfluidic technology for modelling the taneously favouring the angiogenesis by cell
angiogenesis process can unveil key players of releasing of angiogenic factors (Laschke and
angiogenesis regulation in physiological and Menger 2016). The functionalization of biomate-
pathological conditions. rials with biological cues aiming to stimulate
• Complex microvascular systems on a chip can angiogenesis, such as the vascular endothelial
possibly investigating the effect of hemody- growth factors (VEGFs) or the adhesive extracel-
namic forces, cellular interactions, biochemi- lular matrix (ECM)-derived peptides, especially
cal and biophysical stimuli involved on those that specifically binds to endothelial cells
angiogenesis. (ECs), has been one of the most attempted strate-
• Microfluidic platforms have gained promi- gies to direct in vivo angiogenesis on engineered
nence in angiogenesis models but the great tissue substitutes (Rouwkema and
majority of these platforms are confined to Khademhosseini 2016; Akintewe et al. 2017).
proof-of-concept models. However, it has been seen that the simple avail-
ability of angiogenic factors on engineered con-
structs is not sufficient to drive the formation of a
7.1 Introduction mature vascular network (Rouwkema and
Khademhosseini 2016). Thus, the creation of dif-
Vascularization has been one of the major barri- ferent angiogenic factors’ patterns onto scaffolds
ers on the clinical translation of scalable substi- and the control over their spatiotemporal releas-
tutes fabricated using tissue engineering (TE) ing is nowadays seen as a potential strategy to
strategies (Novosel et al. 2011). Although great direct/control neovascularization (Kant and
efforts have been made over the past decade to Coulombe 2018). As a complex coordinated pro-
find a gold standard approach for engineering 3D cess, physical cues have also a critical role on
vascularized tissues, with a significant contribu- angiogenesis modulation, being some strategies
tion of materials science and microfabrication focused on the smart design of TE constructs,
tools, this goal still remains a virtual reality (Kim called scaffold-based approaches. In fact, an ade-
et al. 2016a). Different vascularization approaches quate microarchitecture of the TE substitutes,
have been introduced in order to overcome this comprising a suitable porosity/interconnectivity
hurdle, which can be generically divided in two for being easily perfusable by the host vascular
main groups: (i) cell-based, and (ii) scaffold- system, is essential to promote the organization
based strategies (Novosel et al. 2011). The first of the ECs into a new vascular network (Datta
assents on the intrinsic potential of vessel- et al. 2017). Advanced high precision technolo-
forming cells for neovascularization, which gies, such as 3D bioprinting and photolithogra-
underlies on two basic phenomena: vasculogen- phy, have arisen as potential tools to create highly
esis and angiogenesis. The first, meaning the defined microvascular network’ templates on TE
assembly of undifferentiated endothelial cells substitutes to guide the in vivo formation/organi-
(ECs) into a primordial microvascular system, zation of new vasculature (Griffith et al. 2005;
whereas the second corresponds to the sprouting Chan et al. 2012; Laschke and Menger 2016;
of new blood vessels from pre-existing ones Sudo et al. 2016).
(Novosel et al. 2011). Cell-based strategies con- Vascularization control has a pivotal role not
sist on the incorporation of ECs on TE constructs, only in the context of TE, but also in pathological
being the concept of in vitro pre-vascularization conditions, such as cancer, in which hindering
introduced as initial approach that can possibly vascular network formation is a potential treat-
generating a microvasculature network on the ment strategy (Weis and Cheresh 2011).
7 Microfluidics for Angiogenesis Research 99
Integrated approaches for understanding the bio- potential drugs targeting pathological angiogen-
logical and physical mechanisms underlying esis are also briefly reviewed.
angiogenesis constitute a fundamental step to the
progress on these different fields.
The progress that have been made on micro- 7.2 Microfluidic Platforms
fabrication techniques, such as soft lithography, for Angiogenesis Research
3D printing, and micropatterning techniques have
turned microfluidics a potential tool to the in vitro Angiogenesis is a complex orchestrated phenom-
miniaturization of the complex biochemical and enon of new blood vessels’ formation involving
biophysical phenomena involved on microvascu- the sprouting or splitting of pre-existing ones,
lature morphogenesis (Khademhosseini et al. being a vital and natural process during develop-
2006). By means of applying microfluidic tech- ment and wound repair (Otrock et al. 2007). For
nologies to the design and fabrication of complex this reason, angiogenesis research has been one of
microvascular systems, with a tight control of flu- the main focus on several scientific areas, from
ids perfusion, a plenty of opportunities can arise fundamental studies, for the comprehension of the
for the study of hemodynamic forces, cellular biological and biophysical mechanisms, to
interactions, biochemical and biophysical stimuli applied areas, such as, TE and cancer research,
involved on angiogenesis (Weis and Cheresh where its modulation is imperative. Some signal-
2011; Kim et al. 2017). ling pathways involved on angiogenesis induction
In vitro microvascular models have evolved have been clarified in the past decades (Carmeliet
from simple 2D microvascular systems, widely and Jain 2011). The signalling pathways associ-
used to conduct simple assays such as the influ- ated with recognized angiogenesis’ molecular
ence of different fluid shear stress on cellular inducers, such as the vascular endothelial growth
viability and the chemotactic effect of growth factor (VEGF) (Koch and Claesson-Welsh 2012),
factors (GFs) on cell behaviour (Lee et al. 2018), angiopoietins (ANG) (Pafumi et al. 2015), fibro-
to more physiologically relevant 3D microvascu- blast growth factors (FGF) and VEGF crosstalk
lar models (Bersini and Moretti 2015). (Song and Finley 2018), platelet derived growth
In fact, it is known the vital role of the ECM in factor (Dubrac et al. 2018), have been elucidated.
the context of TE, so the merge of microfluidics Intercellular communication represents also a key
and hydrogels, particularly extracellular matrix factor during vascular morphogenesis. It has been
derived-hydrogels, have naturally arisen to better stated that under pro-angiogenic conditions, the
approach the 3D in vivo niches of vascular for- phenotypic specialization of the ECs into tip or
mation (Lewis and Gerecht 2016). Through a stalk cells is regulated via Notch-signalling
precise control of cell patterning, biochemical (Kofler et al. 2011). The role of several adhesion-
gradients and dynamic physical stimuli, it has membrane proteins (e.g., αvβ3-integrin, ephrin-
been possible to recreate the angiogenesis pro- 2B and VE-cadherin which are involved in
cess in a more realistic and physiological relevant cell-cell and cell-ECM communication) in ECs
conditions (Bersini and Moretti 2015; Lewis and migration, sprouting and new blood vessels’ sta-
Gerecht 2016). bilization, has also been described (Carmeliet and
Herein, an overview of the latest advances on Jain 2011). The ECM remodelling comprises a
microfluidics platforms contributing for the com- natural step of vascular morphogenesis (Mongiat
prehension of the angiogenesis process is pro- et al. 2016). The role of some proteolytic systems,
vided. The microvascular models unveiling key such as plasmin or metalloproteinases (MMPs),
factors involved on this complex process in nor- on angiogenesis regulation, as well as, its influ-
mal physiological and pathological scenarios are ence by delivering ECM-degradation derived GFs
highlighted. Additionally, the development of and cytokines have also been reported (Mongiat
microvascular platforms for the screening of et al. 2016). Besides all these regulatory’ systems,
100 L. Costa et al.
hypoxia has been identified as a major player on Photolithography, a preferentially used tech-
angiogenesis, by the synergistic effect of different nique to create patterns in photosensitive materi-
molecular pathways involved on angiogenesis als, which function as templates for replica
activation, via HIF-1α complex (Krock et al. moulding of biologically inert and optically
2011). Also, the shear stress resulting from hemo- transparent materials (e.g., polydimethylsilox-
dynamic forces is a regulator factor of angiogen- ane, PDMS), has been a recurrent microfabrica-
esis, by means of altering the gene expression of tion method to create microvascular models
ECs through mechanotransduction signals, via (Mannino et al. 2018b). Novel microfabrication
CD31(PECAM) cell-cell junction complex, and techniques have been developed to emulate the
further activation of downstream pathways 3D microenvironment of vascular niches and
(Wragg et al. 2014). Although several pathways overcome some constraints associated with
coordinating angiogenesis have been uncovered lithography techniques, such as the requirement
in the past decades, angiogenesis is a highly com- of specialized training with labour intensive prac-
plex and dynamic regulated process, being its tice and costs. The 3D printing techniques involv-
underlying mechanisms in some pathological ing the use of sacrificial materials to generate
conditions and even in normal circumstances not microchannels directly inside of an hydrogel/
fully understood and therefore unpredictable. scaffold, together with microfiber fabrication,
Microfluidics have emerged as a potential tool comprises some of these non-traditional micro-
to bridge the gap between the poor representative fabrication methods (Mannino et al. 2018b).
conventional 2D culture models and the in vivo Although several microfabrication methods can
models with an intricate and uncontrollable vari- be employed depending on the experimental
ability (Bhatia and Ingber 2014; Bogorad et al. design and envisioned objectives, the endotheli-
2017). By providing micrometer-sized platforms zation of these microchannels, consisting of cul-
that can be designed in different microchannels’ turing/patterning them with ECs, is an essential
geometries with distinct compartmentalization step towards recreating the in vivo vascular mor-
and using an unlimited combination of cell types/ phogenesis (Mannino et al. 2018a).
biomaterials continuously perfused with a Some authors have categorized the in vitro
selected fluid, it have been possible to recreate microvascular models based on the approaches
target microenvironments, such as the vascular used to produce them, namely: (1) templating,
morphogenesis in different contexts, broaden and (2) self-organizing methods (Hasan et al.
application from basic to applied research (van 2014). The templating approach refers to the
der Meer et al. 2009; Bogorad et al. 2017; methods in which hollow lumen templates are
Mannino et al. 2018b). generated inside of ECM-like matrices through
In fact, the flexible design of these microflu- subtractive moulding techniques (Fig. 7.1a, b)
idic platforms can allow to set up different chem- (Sadr et al. 2011; Yoshida et al. 2013; Buchanan
ical and physical cues with control of their et al. 2014), sacrificial materials (Wu et al. 2011;
spatiotemporal deliver, have unveiled the poten- Miller et al. 2012) or bioprinting (Norotte et al.
tial of this technology, not only to the fundamen- 2009). The second approach is based on the
tal study of the subjacent mechanisms of vascular innate morphogenic potential of ECs to sponta-
formation and regulation, but also for the con- neously generate microvascular networks, under
struction of high-throughput models for drugs GFs gradients or co-culture of supporting cells,
screening targeting angiogenesis (Young 2013). mainly supported by photolithography
Moreover, since this type of technology com- techniques. These self-assembly approach can

monly makes use of optically transparent materi- offer a closer proximity to the in vivo physiologi-
als, it is compatible with several imaging cal conditions (Lee et al. 2018).
microscopy techniques, enabling real time moni- Despite this approach-based categorization,
toring and a full set of more reliable functional microvascular models can go from the 2D endo-
readouts (Young 2013). thelialized microvascular models, comprehending
those in which ECs are simple lined on the micro- poly(ethylene glycol) (PEG) hydrogel containing
channels without a supportive cellular matrix cell-adhesive and MMP-degradable peptides in a
(Fig. 7.1c, d) (Theberge et al. 2015), usually used tri-channel microfluidic system (Fig. 7.1h–j). In
to conduct simple shear stress studies on ECs that study, the ability of this tuned matrix towards
response (Raasch et al. 2015) or endothelium per- generating in vivo-like lumenized vascular net-
meability studies (Sato et al. 2015), to more com- works was demonstrated, thus evidencing the
plex and realistic 3D models. Using a 2D utility of combining supportive matrices with
approach, Theberge et al. (Theberge et al. 2015) standard microfluidic platforms for the in vitro
have investigated the chemical crosstalk between study and recapitulation of vascular morphogen-
ECs and macrophages during angiogenesis on a esis in a more realistic and 3D context (Zanotelli
co-culture microfluidics platform (Fig. 7.1c, d) et al. 2016).
(Theberge et al. 2015). The role of macrophage The design of such types of microvascular
derived-MMP12 as an anti-angiogenic factor was models has been adapted according to the final
confirmed by observing tubule formation impair- objective of the study. Since these devices have
ment in co-culture conditions (ECs cultured with been applied to the systematic study of different
fibroblasts under macrophages stimulation), factors affecting angiogenesis, such as the role of
which can be reversed by applying an MMP12 GFs (Chung et al. 2009; Shin et al. 2011; Jeong
inhibitor (Theberge et al. 2015). et al. 2011a; Dai et al. 2011), flow conditions
However, the importance of having the ECs in (Song and Munn 2011; Buchanan et al. 2014),
their natural in vivo organization, i.e. in lumen- intercellular communication (Kim et al. 2015b),
ized structures on their phenotype and cell behav- tumour angiogenesis (Buchanan et al. 2014;
iour (Bischel et al. 2014), have prompted the use Chung et al. 2017; Miller et al. 2018), and drugs
of 3D microvascular models. Supportive cellular screening targeting angiogenesis (Kim et al.
matrices have been used for the development of 2015a; Theberge et al. 2015; Amann et al. 2017),
more physiologically relevant 3D microvascular a panoply of increasingly complex microfluidic’
models. Interestingly, the 3D lumenized struc- designs have been generated. Hence, instead of
tures involving microfluidic devices with ECM- focusing on the different complexity of these
matrices (Bischel et al. 2013) can allow studying microvascular models, further attention will be
the vascular morphogenesis phenomenon, given to their final aim. In this context, some
through vasculogenesis and angiogenesis studies using different microvascular models for
(Fig. 7.1e–g) (Chung et al. 2009; Jeong et al. assessing the effect of several factors on angio-
2011a; Dai et al. 2011; Yeon et al. 2012; Kim genesis will be discussed in the next sub-sections.
et al. 2013a; Jeon et al. 2014; Cochrane et al. Moreover, the application of these microfluidic
2018; Mannino et al. 2018b). These are typically platforms for modelling angiogenesis-related
composed by interleaved controlled and condi- pathologic conditions and its potential use for
tioned flow channels (where different stimuli can discovery of new therapeutic targets is also
be applied) connected to one or more compart- reviewed.
ments, typically micropatterned with different
3D matrices, such as fibrin, collagen or Matrigel,
to emulate the ECM, where ECs are attached and 7.2.1 Regulatory Factors
allowed to sprout, with the final objective of Controlling Angiogenesis
investigate microvascular network formation in
direction of an applied stimuli (Cochrane et al. Although several molecular-signalling pathways
2018). For example, Zanotelli MR et al. (Zanotelli involved in angiogenesis induction have been
et al. 2016) have developed a 3D morphogenic- uncovered in past years, their intricate crosstalk
based microvascular model by using induced plu- towards new vasculature formation, has not been
ripotent stem cell-derived endothelial cells fully understood. By its turn, the 3D microfluidic
(iPSC-ECs) encapsulated in a functionalized models of angiogenesis have both enabled the
Fig. 7.1 Approach-based categorization of the microvas- 3D μVM tumour model to study the effect of wall shear
cular models (μVM) into templating [a-b] and self- stress (WSS) on tumour-endothelium paracrine signalling
organizing methods, from the simpler 2D μVM [c-d] to of angiogenesis. [b] Lumen formation after 72h of pre-
the more realistic 3D μVM [e-g; h-j]. [a] Scheme of the condition flow rate in co-culture system of breast cancer
precise application of a set of chemical and phys- tions, ANG-1 was identified as a major player on
ical stimulus, and their reliable and quantitative its regulation (Shin et al. 2011). By it turn,
monitoring in terms of the synchronized cellular Nguyen et al. (Nguyen et al. 2013) have demon-
responses involved on vascular morphogenesis. strated the effect of different combined angio-
The VEGF and ANG-1 crosstalk guiding the genic factors on the multicellular responses
coordinated cellular responses during angiogen- involved on neovascularization and the opposite
esis was assessed by Shin et al. (Shin et al. 2011). effect of angiogenic inhibitors. This biomimetic
The authors have investigated EC sprout into a model of angiogenesis was based on two cylin-
3D collagen-type I matrix, encased on a micro- drical channels (one lined with ECs and perfused
fluidic device and linked to three channels, i.e. with medium and the other used to apply the
one serving as control and the others perfused angiogenic stimuli) encased and linked by a 3D
with culture media supplemented with the bio- collagen matrix. Several proangiogenic factors
chemical factors, being the ECs cultured on the (i.e. basic fibroblast growth factor (bFGF), hepa-
ANG-1 supplemented channel (Fig. 7.2a). The tocyte growth factor (HGF), VEGF, monocyte
role of ANG-1 on the collective and coordinated chemotactic protein-1 (MCP-1), sphingosine-1-
migration of tip and stalk ECs during vascular phosphate (S1P) and phorbol 12-myristate
morphogenesis was confirmed on this study. 13-acetate (PMA)), were tested individually and
Differences on the morphogenetic behaviour of in different combinations. Results showed that
ECs under VEGF alone and VEGF combined only S1P and PMA gradients have resulted in
with ANG-1 were clearly observed, thus evidenc- significant EC migration, when tested indepen-
ing the role of ANG-1 on enhancing and stabiliz- dently. Although, PMA promoted a more cohe-
ing the tip-stalk EC connections (Fig. 7.2b, c). sive and coordinated EC migration when
Since the morphogenesis and life cycle of stalk compared to the EC response under S1P. Two
ECs are regulated by direct tip-stalk cell connec- distinct cocktails were tested and the combina-
Fig. 7.1 (continued) cells (MDA-MB-231 cells) and and angiogenesis [f.5-8]. [g] Immunofluorescence images
endothelial cells expressing a red fluorescent protein of 3D endothelial networks stained for CD31 (red),
(RFP) (scale bar: 200 μm). Reprinted with permission F-actin (green) and nuclei (blue) formed by vasculogene-
from Buchanan et al. (Buchanan et al. 2014). Copyright © sis [g.1] and angiogenesis [g.2-3] (scale bar: 50 μm).
2014 Taylor & Francis Group, LLC. [c] 2D self-organiz- F-actin-rich filopodia extension of angiogenic sprouts
ing μVM configuration for the study of different cell [g.4] (scale bar: 20 μm) and hollow lumen formation [g.5]
derived-soluble factor signalling on endothelial tubule (scale bar: 10 μm). [g.6] Immunofluorescence images of
formation. [d] Fluorescence micrographs of the right the microvascular networks stained for junction proteins
channel containing the co-culture system (ECs and fibro- (scale bar: 50 μm) and for vessel maturation markers,
blasts) immunocytochemistry stained for CD31 (green) ICAM-1 [g.7] (scale bar: 10 μm), laminin [g.8] and col-
and nuclei (DAPI-blue) in different conditions (co- lagen IV [g.9] (scale bar: 20 μm). Reprinted with permis-
culture/tri-culture systems and in the presence of MMP12 sion from Kim et al. (Kim et al. 2013a). Copyright © 2013
inhibitor). Reprinted with permission from Theberge et al. Royal Society of Chemistry. [h] Scheme of the 3D μVM
(Theberge et al. 2015). Copyright © 2015 American comprising a tri-channel interconnected system: one cen-
Chemical Society. [e] Schematic representation of a 3D tral channel and two fluidic channels. Confocal images of
self-morphogenic μVM for studying vasculogenesis and microvascular network formation from iPSC-ECs, encap-
angiogenesis. The microfluidic platform comprises the sulated in different MMP degradable peptide-crosslinked
outermost channels (LO-left outside and RO-right out- PEG matrices (polymerized into the central channel) and
side) for human normal lung fibroblasts (LFs) seeding, cultured in basal medium [i.1, 3, 5] or in VEGF-
flanked with two fluidic channels and one central channel. supplemented medium [i.2, 4,6] (scale bar: 250 μm).
For vasculogenesis experiment [e.3-4] HUVECs-loaded Capillary network formation in 50% crosslinked PEG
fibrin matrix were patterned in the central channel, while hydrogel stained for CD31 (green), VE-Cadherin (red)
for angiogenesis study [e.5-6], HUVECs were attached on and nuclei (blue-DAPI) at different magnifications [j.1, 2,
the side of central channel patterned with acellular 3D 3]. Scale bars: 250 μm [j.1] and 100 μm [j.2]. Reprinted
fibrin matrix, while the fibroblasts were only cultured in with permission from Zanotelli et al. (Zanotelli et al.
the same matrix on the RO channel. [f] Micrographs of 2016). Copyright © 2016 Acta Materialia Inc. Published
endothelial network formation by vasculogenesis [f.1-4] by Elsevier Ltd.
Fig. 7.2 Microvascular platforms for the study of the erative angiogenic sprout response. [b] Angiogenic
biochemical cues driving angiogenesis. [a.i-ii] sprouts formation into the collagen scaffold, under the
Microfluidic system and its schematic representation to effect of horizontal and vertical VEGF gradients [b.i] and
investigate the VEGF-ANG-1 crosstalk on the 3D coop- under the same VEGF applied gradients supplemented
tion containing MCP-1, VEGF, PMA, and S1P Recently, van Duinen et al. (2019) have devel-
has shown a superior effect on multicellular oped a robust and standardized angiogenesis
sprouting response, resembling better the in vivo model on a microfluidic platform for high-
vascular morphogenesis. The effects of a VEGF throughput assessment of gradients driving this
receptor inhibitor on the angiogenic sprouting of process. The scalability of this microvascular
the developed systems were also evaluated. The platform was achieved by combining an array of
results unveiled a VEGF-induced filopodia for- 40 microfluidic devices, integrated underneath a
mation and sprout elongation in a context- 384-well plate. Maintaining the platform under
dependent manner (Nguyen et al. 2013). The passive levelling perfusion on a rocker platform,
proximity of this in vitro model to the in vivo vas- reproductible gradients were generated and dif-
cular morphogenesis process, which recapitu- ferent experiments were allowed. Each device
lated all the coordinated steps involved on was composed by three PhaseGuides-
angiogenesis, such as sprouting, elongation and interconnected channels: one central channel pat-
lumen formation/stabilization, could be poten- terned with collagen type I, serving as sprouting
tially applied to the study of the molecular mech- channel, and two adjacent perfusion channels
anisms driving neovascularization. (one for culturing ECs and the other to generate
In a similar study, Farahat et al. (Farahat et al. different angiogenic gradients after the establish-
2012) have developed a 3D high-throughput ment of a confluent endothelial monolayer)
microfluidic platform for the study of two proan- (Fig. 7.2d). Results have proven the utility of the
giogenic factors combination, VEGF and S1P, on platform to conduct in vitro relevant studies of
the EC response. The authors have demonstrated factors driving induction and regression of micro-
the synergistic effect of both factors on angiogen- vasculature, due to its physiological resemblance
esis sprouting induction, particularly on influenc- with the in vivo angiogenesis EC multicellular
ing the growth and direction of the sprouts responses. Moreover, an optimal cocktail for the
(Farahat et al. 2012). Inspired on this microflu- generation of a robust angiogenic sprouting
idic architecture, Del Amo et al. (2016), have response, was identified, being composed by
investigated the effect of some GFs, involved on VEGF-165, PMA and S1P, with the last having a
regeneration processes, namely VEGF, critical role on the directional guidance and
PDGF-BB, TGFβ and BMP-2 on the initial stage maintenance of angiogenesis sprouting
of endothelial sprout. The results suggested that (Fig. 7.2e). Under prolonged exposure to gradi-
VEGF is among the main angiogenesis regula- ents, anastomosis have occurred, forming perfus-
tors, being its concentration and spatiotemporal able stable lumens with proper barrier function
location decisive for sprouting evolution (Del and sprouts regression, emulating the in vivo
Amo et al. 2016). neovessel stabilization and maturation (van
Duinen et al. 2019). As it can be noticed from
Fig. 7.2 (continued) with ANG-1 [b.ii]. The white arrow- [d.iv.1]. Tip cell filopodia formation observed, at day 1,
heads point to tip-stalk cell sprouting structures while with lumen formation visible at day 2 [d.iv.2,3], after GFs
empty arrowheads indicate isolated tip cells (scale bar: gradient application. Angiogenic sprouts formation under
100 mm). [c] Quantification of the number of tip cells [c.i] different angiogenic factor’ combinations (VEGF+PMA;
and the number of tip cells attached to stalk cells sprout- VEGF+S1P and VEGF+PMA + S1P), after 4 days of
ing structures [c.ii] under VEGF gradients or ANG-1 stimulation [E.i] and the respective sprouts, after 6 days of
supplemented VEGF gradients (∗P < 0.05; ∗∗P < 0.01; stimulation, stained for F-actin (red) and nucleus (blue)
∗∗∗
P < 0.001). Reprinted with permission from Shin et al. [e.ii,iv,v]. Close up view of angiogenic sprouting under
(Shin et al. 2011). Copyright © 2011 Royal Society of VEGF+PMA + S1P, stained for F-actin (red), nucleus
Chemistry. [d] Schematic representation of the standard- (blue) and VE-cadherin (green) (Scale bars: 100 μm) [e.
ized microfluidic platform for the high-throughput screen- iii]. Tip-stalk cell structures with filopodia and lumen for-
ing’ factors governing angiogenesis [d.i-ii]. Microvessel mation [e.iii], traversing the entire gel channel [e.ii] under
culturing ‘steps [d.iii] on the collagen’ adjacent perfusion the combination VEGF+PMA + S1P. Reprinted with per-
channel, for the investigation of ECs sprouting response mission from van Duinen et al. (2019). Copyright © 2018,
under gradients imposed on the other perfusion channel Springer Nature
106 L. Costa et al.
Fig. 7.2, microfluidic angiogenic models have in a NO-signalling dependent manner. By its
been accepted and broadly used to study the turn, interstitial flow and VEGF interplay have
effect of several biochemical-angiogenesis mod- shown a synergistic effect, promoting EC inva-
ulators and their crosstalk on ECs behaviour dur- sion and sprouting with tip cell filopodia forma-
ing the different vascular morphogenic steps. tion against the transverse applied flow and
Mechanical forces, imposed by both blood directed to VEGF increasing gradients (Song and
flow and interstitial plasma are also key regula- Munn 2011).
tors of vascular morphogenesis and remodelling, Using a microfluidic device Galie et al. (2014)
having a profound impact on EC gene expression have reported a common shear stress threshold
and behaviour. It has been reported that varia- value (∼10 dyn/cm2) for both luminal and trans-
tions of the normal shear stress, experienced by a mural flow, in which angiogenesis sprouting is
quiescent endothelium, can induce intracellular induced through up-regulation of the downstream
signalling pathways mediated by effectors, such as the MMP-1, critical on 3D cell
EC-mechanosensing molecules, such as the migration and sprouting (Galie et al. 2014).
CD31 platelet endothelial cell adhesion complex In another study, the individual and combined
(Wragg et al. 2014; Lertkiatmongkol et al. 2016). effect of interstitial flow (IF) and proangiogenic
Nitric oxide (NO) has also been identified as factors, such as VEGF an S1P, were investigated
mediator of shear stress-induced angiogenesis (Kim et al. 2016b). It was reported a microfluidic
(Wragg et al. 2014). Although it is known that platform comprehending six parallel microchan-
angiogenesis is a tightly and dynamically regu- nels interconnected by arrays of microposts. The
lated process by biochemical and mechanical sig- two outermost channels were cultured with stro-
nalling, the interplay between these factors have mal fibroblasts (known to secret pro-angiogenic
been poorly explored. Song and Munn (2011) factors), the immediately adjacent two intermedi-
have investigated the crosstalk between mechani- ate microchannels served as fluidic channels, and
cal forces, imposed by tangential shear stress and the two central microchannels were patterned/
transverse interstitial flow, and VEGF, in the vas- cultured using two different configurations. One
cular morphogenesis regulation. That study was was cultured with ECs embedding on a fibrin
conducted on a microfluidic device model for matrix (for EC assembly on a vascular network
angiogenesis sprouting. The microfluidic plat- through vasculogenesis) and the other patterned
form comprised two EC lined microchannels, with the acellular fibrin matrix (serving as a
where controllable fluid forces were imposed, as sprouting channel) (Fig. 7.3a). Different direc-
well as, established VEGF gradients, intercon- tional IF pressures were applied, as well as, spa-
nected by a collagen compartment for EC sprout- tially defined pro-angiogenic gradients were
ing. The study demonstrated that physiological established by varying the fibroblast densities’
shear stress reduces VEGF-induced angiogenesis ratio between the outermost channels. The results
Fig. 7.3 (continued) (scale bar: 100 μm). Representative namely density, on endothelial sprouting response, under
fluorescence images of the angiogenic sprouting response appropriate established VEGF gradients. Representative
(stained for nuclei-blue, F-actin-green and CD31-red) to phase contrast images of endothelial sprout morphology
different directional applied flow pressures (indicated by at different matrix densities [d]. Favoured endothelial
white arrows) [b.i] (scale bar: 200 μm). Higher magnifica- multicellular response with the formation of stable elon-
tion images of the angiogenic sprouts depicted on the dot- gated sprouts at intermediate collagen densities [b.iii-iv].
ted boxes [b.ii], with an increase of actin-rich filopodia Sprouting features of ECs-coated microbeads at different
formation (yellow arrows) in response to opposite sprout- collagen matrix densities [e], depicting the total number
directional applied IF (scale bar: 40 μm). Quantification of sprouts per bead [e.i], the average sprout thickness [e.
of total sprout area [b.iii] and total microvascular network ii], the average aspect ratio (sprout length to width) [e.iii]
area [b.iv]. Reprinted with permission from Kim et al. and the number of new sprouts at different time points
(2016b). Copyright © 2016 Royal Society of Chemistry. normalized to the total number of observed sprouts [e.iv]
[c] Schematic configuration of the microfluidic system to (∗p < 0.05; ∗∗p < 0.01). Reprinted with permission from
investigate the effect of biomechanical matrix properties, Shamloo et al. (Shamloo and Heilshorn 2010). Copyright
© 2010 Royal Society of Chemistry
Fig. 7.3 Microvascular platforms for the study of the factors-fibroblast stimulation, with its schematic configu-
biophysical cues driving angiogenesis. [a] Microfluidic ration [a.i-ii]. Sequential assembly of a microvascular net-
device to investigate the influence of interstitial flow (IF) work, through vasculogenesis, on channel N and
on microvascular morphogenesis, under pro-angiogenic angiogenic sprouting into channel S, after 4 days [a.iii]
108 L. Costa et al.
have shown a synergistic effect of IF and the pro- Given the high adaptability of microfluidic
angiogenic factors synthetized by fibroblasts, on platforms to be designed in such a way that
the directional guidance and stimulation of endo- enables the in vitro recreation of a target 3D
thelial sprouting (Kim et al. 2016b) (Fig. 7.3b). microenvironment, it is not surprising that micro-
The ECM role on angiogenesis regulation has fluidics constitute an usefulness platform for the
also been described. It is known that ECM pro- investigation of the guidelines for the construc-
vides adhesive motifs enabling EC migration tion of advanced vascularized TE substitutes
during vascular morphogenesis and functions as (Smith and Gerecht 2014). Jeong et al. (2011b)
a pool of ECM-remodelling derived biomole- have reported the ability of an ECM-like hydro-
cules having a critical and intricate role on angio- gel system composed by a semi-interpenetrating
genesis regulation (Mongiat et al. 2016). network (SIPNs) of collagen and hyaluronic acid
Although the importance of the ECM as a physi- (HA) to modulate EC migration in the presence
cal support of the living tissues and its crucial of a VEGF gradient, on a study conducted on a
role on cell behaviour is consensual, the biome- three channel microfluidic device. In other study,
chanical properties of ECM on angiogenesis reg- Baker et al. (2013) have investigated the influ-
ulation have been neglected. Shamloo and ence of different molecular inducers’ diffusion
Heilshorn (Shamloo and Heilshorn 2010) have patterns on vascular morphogenesis, namely on
investigated the influence of ECM properties, the location and morphology of EC sprouting.
such as matrix density/stiffness and VEGF cross- Using different lithography-defined sacrificial
talk on ECs sprouting and lumen stabilization. templates, different microchannel outlines were
The researchers have used a microfluidic device drawn within a collagen matrix, enabling the
for culturing adult human dermal microvascular generation of differential diffusion patterns of
endothelial cells (HDMVEC)-coated microbeads several GFs. Results have demonstrated that, EC
on a 3D collagen/fibronectin matrix compartment sprouting is defined on the highest gradient loca-
connected by microcapillaries to two channels, tions, being the concerted EC migration/sprout-
where the VEGF gradients were established ing with the typical formation of stalk-like
(Fig. 7.3c). Different matrix stiffnesses were structures, favoured by intermediate GF gradi-
tested, by means of varying the collagen concen- ents (Baker et al. 2013).
tration on the matrix blend. Several VEGF (e.g.,
VEGF-A and VEGF-165) concentrations were
also tested. Results suggested that matrix density 7.2.2 Cellular Interactions
mediates VEGF-induced sprouting and lumen
formation by having critical role on the balance Cellular communication and cellular-ECM inter-
between endothelial tip cell migration and stalk action are key players on homeostasis mainte-
cell proliferation. While higher matrices’ densi- nance in any tissue, being this intercommunication
ties promoted the formation of thicker sprouts fundamental to generate integrated cellular
with better ability for lumen formation, lower responses when homeostasis is perturbed, such as
stiffnesses impaired lumen stabilization on regeneration/wound healing processes or on
(Fig. 7.3d–e) (Shamloo and Heilshorn 2010). pathological conditions (e.g. cancer) (Sakthivel
That study stated the importance of ECM biome- et al. 2019). All these conditions are accompa-
chanical properties on angiogenesis regulation, nied by vascular remodelling through
evidencing the utility of this microfluidic plat- angiogenesis-based processes. In such scenario,
form, not only for the investigation of mechani- the organized and interconnected quiescent endo-
cally suitable biomaterials for regenerative thelial cells monolayer (via junction molecules-
applications, but also serving as a platform for VE-cadherin and claudins) protected by pericytes,
the discovery of new potential biochemical thera- initiates a cascade of cellular responses. Briefly,
pies targeting ECM-related pathological in response to an angiogenic signal, pericytes
angiogenesis. detach from the basement membrane of blood
vessels by proteolytic degradation mediated by showing morphological alterations of the ECs

MMPs. Subsequently, ECs lose their junctions and pericytes in response to this biochemical
and the new forming vessel dilates. The permea- stimulus. Together, the results demonstrated the
bility of the growing endothelium increases (via utility of this platform as a dynamic model of
VEGF signalling) leading to a parallel increase pericytes and EC communication during angio-
of plasma proteins extravasation creating a tem- genesis, to study pathological situations, as well
porary ECM. ECs migrate in direction of this as, for drug screening of potential therapeutics
provisional ECM, through integrin signalling and targeting abnormal ECs-pericytes cell signalling
the simultaneous remodelling of ECM release observed, for instance, in tumour-induced angio-
pro-angiogenic factors providing an angiogenic- genesis (Kim et al. 2015b).
propitious microenvironment. The new vessel is Angiogenesis is a dynamically regulated pro-
then formed by the concerted action of the cess dependent on the activation/suppression of
selected tip EC (DLL4 and JAGGED1), which integrated molecular and cellular signalling path-
migrates, through integrins mediated-adhesion, ways. Comprehending the cell-interaction and
in the direction of signalling molecules (sema- cell-ECM involved on vascular morphogenesis is
phorins and ephrins), and the stalk EC, which of major interest for several research areas, in
divides and elongates the new capillary. The sta- which, the control of this process is crucial.
bilization/maturation of the newly formed blood Microfluidics have been applied to the study of
vessel is crucial for its functionality, consisting in cellular interactions involved on angiogenesis
the covering of the ECs layer by pericytes, the and their abnormal communication in pathologi-
deposition of a basement membrane and ECs- cal conditions, by providing suitable co-culture
junctional re-establishment (Fig. 7.4) (Carmeliet systems for the in vitro recreation of vascular
and Jain 2011). morphogenesis under targeted microenviron-
The role of pericytes during the angiogenesis ments (Khademhosseini et al. 2006; Sakthivel
process was studied on a microfluidic platform et al. 2019).
by Kim et al. (2015b). In that study, the authors The importance of EC source for the in vitro
recreated in vitro, the sequential process of angio- reconstruction of organ-specific microvascular
genesis: sprouting, elongation and maturation, by models was asserted by H. Uwamori et al. (2019),
means of using a microfluidic device. This device who conducted a study in a trichannel microflu-
was composed of a central acellular channel, pat- idic device using different ECs sources: brain
terned with a 3D fibrin matrix, connected to two microvascular endothelial cells (BMECs) and
adjacent media channels, and an outermost HUVEC in co-culture with mesenchymal cells,
fibroblast cultured channel. A co-culture of endo- for modelling the blood brain barrier (BBB). The
thelial cells (HUVECS) and pericytes were different co-culture systems (MSC with HUVEC
attached to one side of the fibrin central channel and MSC with BMECs) were attached to one
and the sprouting was directionally observed in side of the fibrin gel prepatterned on the central
response of the biochemical factors generated by compartment, where microvascular networks
dermal and lung-derived fibroblasts. In vivo-like were allowed to grow. Results from permeability
sprouting, elongation and stabilization of the tests of the developed microvasculatures and
newly formed blood vessels, with pericytes cov- junction proteins’ expression profiles showed
ering, was successfully recapitulated in the pres- significant differences in terms of ECs morphol-
ence of the biochemical factors secreted by ogy and barrier function. It was demonstrated
dermal fibroblasts. Thus, the results evidenced that BMECs formed more robust endothelial bar-
that pericytes play an important role on both ves- rier, being less permeable than the microvascular
sel diameter and branching regulation. Moreover, networks formed with HUVECS, due to higher
the effects of VEGF and inflammatory cytokines expression of tight junction proteins. The barrier
on pericytes-ECs interaction were assessed. The function of the BMEC-microvasculature, under
results were in agreement with previous studies, induced inflammatory conditions by addition of
110 L. Costa et al.
Fig. 7.4 Molecular signalling involved on neovascular- genic stimulators. Tip EC migrates in the direction of sig-
ization trough angiogenesis sprouting. Under a pro- nalling molecules accompanied by stalk cells proliferation
angiogenic microenvironment, the quiescent endothelium allowing sprout elongation and lumen formation. After
dilates, and the tip EC is selected to initiate vessel branch- sprouts fusion (anastomosis), stalk cells recruit pericytes
ing. The process is accompanied by degradation of base- for neovessel stabilization. EC junctions are re-
ment membrane, pericyte detachment and EC junctions established, the basement membrane deposited, and the
loosening. The permeability of the branching vessel vessel maturation promoted by maintenance signals.
increases and the plasm proteins-extravasion provides a Reprinted with permission from Carmeliet and Jain
temporary ECM, while its degradation liberates angio- (2011). Copyright © 2011 Springer Nature
thrombin in cell culture media, has shown an ments, in physiological and pathological condi-
increased permeability with leaky characteristics. tions, and predict ECs response under drug effect,
This organ specific-endothelium model mimick- have been demonstrated by Abaci et al. (Abaci
ing BBB functions can be useful to conduct stud- et al. 2015). The authors have developed a micro-
ies of BBB-associated pathologies (Uwamori vascular model to study EC phenotype changes
et al. 2019). in pathological settings, such as on ischemic con-
ditions associated to atherosclerosis. The physi-
ological relevance of the microvascular model
7.2.3 Modelling Pathological was assessed by treating the healthy microvascu-
Conditions lar system with a commonly used vasotoxic anti-
cancer drug, 5-Fluorouracil (5-FU), in
Beyond the critical importance of angiogenesis combination with a vasoprotective drug
in normal physiological processes, such as, in the (Resveratrol), being the outcomes compared with
context of tissue regeneration, its aberrant regula- an in vivo mouse model. Similarly to what have
tion is in the origin of many pathological condi- been reported on in vivo studies (Beck et al. 2000;
tions (Chung and Ferrara 2011), such as in Hayashi et al. 2003; Skuli et al. 2012), ischemic
inflammatory scenarios (Costa et al. 2007; Kim conditions trigger the transition of a quiescent
et al. 2013b; Varricchi et al. 2015), arthritis endothelium to an activated angiogenic state,
(Marrelli et al. 2011; Elshabrawy et al. 2015), with higher proliferation rates, up-regulation of
autoimmune diseases (Liakouli et al. 2011; Liu angiogenic genes and compromised endothelial
et al. 2015), and cancer progression and metasta- barrier function. In agreement with previous data,
sis (Gavalas et al. 2013; Multhoff et al. 2014; the treatment with 5-FU revealed that the drug
Bielenberg and Zetter 2015), among others. does not promote EC apoptosis. Nevertheless, it
Molecular and cellular mechanisms involved on impairs EC barrier function by means of interfer-
pathological angiogenesis in different circum- ing with adhesive cell junctions, which can be
stances, have been investigated on in vitro micro- attenuated by co-administration with Resveratrol.
vascular platforms (Abaci et al. 2013). Choroidal This microvascular model has, thus, proven its
pathological-angiogenesis associated to macular potential for being used as a physiologically real-
degeneration was investigated by Chen et al. istic platform to investigate EC responses/pheno-
(2017). In that study, the ocular fundus was recre- types in healthy and diseased conditions and to
ated through the set-up of a co-culture system of screen potential drug therapies (Abaci et al.
human retinal pigmental epithelial (ARPE-19) 2015).
cells and HUVECs, separated by a porous mem- The great majority of these microvascular
brane in a microfluidic device. The abnormal platforms have been devoted to recapitulate the
secretion of VEGFs, appointed as the trigger of complex tumour microenvironment (TME)
pathological choroidal vascularization, was (Michna et al. 2018), for studying the interplay
imposed on the co-culture microfluidic system between its components during tumour develop-
through a combination of low glucose and ment and progression. The intricate interaction
hypoxic conditions. Results demonstrated that between endothelial cells, tumour (Lin et al.
the changes on the co-culture microenvironment 2017) and stromal (Chung et al. 2017) cells has
can increase the VEGF levels secreted by ARPE- been investigated using microfluidic platforms.
19 cells, leading to an induced directional growth Chung et al. (Chung et al. 2017) have studied
of HUVECs, which seems to cause ARPE-19 tumour cell-stroma interaction in a paracrine
monolayer detachment (Chen et al. 2017). fashion and angiogenesis driven by tumour-
The potential of these microvascular systems stroma complex on a microfluidic platform. In
for modelling complex vascular microenviron- that study, it was evidenced the differential influ-
112 L. Costa et al.
ence of stroma on morphological changes of dif- static cases, the understanding of the complex
ferent tumour cell types. Moreover, proangiogenic phenomenon behind tumour angiogenesis devel-
factors resulting from different tumour cells- opment is imperative for the identification of new
fibroblasts coculture systems (simulating tumour targets for therapeutic treatment.
stroma), have stimulated the angiogenic sprout- Microfluidic technologies hold a tremendous
ing. Although, differences between the ability of potential to contribute to the progress on this
different tumour cell types on EC sprouting field, by providing co-culture cell platforms,
induction, phenotype and on the permeability of where specific tumour microenvironments can be
new formed blood vessels were noticed (Chung modelled, and the intricate tumour-EC interac-
et al. 2017). tions studied in order to possibly identifying key
By it turn, it has been proposed an integrated molecular and cellular factors involved in tumour
microfluidic platform which allows the simulta- angiogenesis, while new therapeutic drugs can be
neous investigation of tumour-EC interactions tested (Caballero et al. 2017; Sontheimer-Phelps
(through a co-culture system of HUVECs and et al. 2019). Several cancer models have been
cervical carcinoma cells (CaSki), identification developed on microfluidic platforms (Sontheimer-
of cell secreted angiogenic proteins, and evalua- Phelps et al. 2019), aiming to elucidate the com-
tion of paclitaxel drug resistance (Lin et al. 2017). plex multistep involved on cancer metastasis,
It was noticed that HUVECs can induce being the existing model mainly focused on can-
paclitaxel-drug resistance in co-culture condi- cer intravasion (Zervantonakis et al. 2012) and
tions. The potential of this integrated platform for extravasion processes (Chen et al. 2013; Jeon
studying/monitoring tumour-EC interaction, for et al. 2013).
identification of tumour biomarkers and, simulta- Tumour angiogenesis, a putative target for
neously, for functioning as an anti-cancer drug cancer treatment (Zhao and Adjei 2015), have
screening platform was evidenced on that study. been, yet, underexplored using microfluidic
The promising use of such technology for approaches, considering the potential of this
engineering vascularized tumour models that can technology, not only, to the fundamental compre-
be useful for investigating the complex molecular hension of this process, but also to the high-
and cellular mechanisms involved on cancer pro- throughput screening of potential drugs targeting
gression and metastasis (Wang et al. 2018), tumour angiogenesis.
namely, by modelling tumour angiogenesis will Lee et al. (2014) have developed a microflu-
be highlighted in the following sections. idic platform modelling tumour angiogenesis and
metastasis. The reported microfluidic chip was
7.2.3.1 Tumour Angiogenesis Study organized in seven microchannels, intercon-
Tumour angiogenesis is a natural process of nected by a set of microposts, spatially organized
tumour growth and development, by establishing to culture different cells and to allow the analysis
a vascular blood supply to support tumour cells’ of different processes. Briefly, microvessel was
viability and proliferation. Tumour microenvi- generated by vasculogenesis of HUEVCs, which
ronment comprises a pool of signalling mole- were able to self-assemble on a perfusable
cules and pathways driving the uncontrolled microvessel under lung fibroblasts-angiogenic
activation of the quiescent endothelium by creat- factors’ stimuli. The immediately adjacent micro-
ing pro-angiogenic conditions (Weis and Cheresh channel was divided into two smaller microchan-
2011). The formation of tortuous and dysregu- nels: the upper and bridge channels, serving for
lated tumour vascular networks, also character- cancer culture, in different configurations,
ized by the lack of proper barrier functions (due according to the objective of study (cancer angio-
to a deficient pericyte coverage), has also been genesis or metastasis). For cancer angiogenesis
closely associated to tumour metastasis study, human glioblastoma multiform cells
(Bielenberg and Zetter 2015). Considering that a (U87MG), embedded in a fibrinogen solution,
vast majority of cancer deaths results from meta- were patterned into the upper channel while acel-
lular fibrinogen was injected into the adjacent idic system composed by one perfusable micro-
bridge channel. The angiogenic stimulation of channel, lined with ECs and concentrically
the surrounding microvasculature by cancer co-cultured with breast cancer cells, encased onto
cells-secreted factors was successfully modelled a collagen type I matrix, different flow shear
on this in vitro platform. A significant formation stress rates were experimented. Results have
of sprouting structures compared with the control shown that flow shear stress can modulate the
group, with no angiogenic-cancer stimulation, angiogenic phenotype of tumour cells in a para-
was clearly noticed. Moreover, under bevaci- crine fashion, since higher values of shear stress
zumab treatment, a common anti-VEGF drug, have promoted a decrease on tumour-expression
the angiogenic sprouting was blocked (Lee et al. of angiogenic factors (Buchanan et al. 2014).
2014). Although such type of in vitro model was Miller et al. (2018) has reported the first 3D
able to recapitulate the tumour angiogenesis and human renal carcinoma model on a microfluidic
intravasion of cancer cells, more realistic plat- chip. In that study, primary human renal carci-
forms which can better resemble the tumour noma cells from patient donors were used. The
microenvironment and EC-tumour communica- primary-donor derived tumour cell clusters were
tion during these processes are in great need. embedded in a collagen matrix and patterned
In a different study, Zheng et al. (2016) have around an engineered human microvessel sub-
studied leukaemia-induced bone marrow angio- jected to continuous flow. Results showed that
genesis on a microfluidic device comprising three the human renal carcinoma cells stimulate ECs
parallel and interconnected microchannels. The sprouting, being their phenotype, namely the
microfluidic platform was composed by two cell expression of angiogenic factors, donor-
culture channels, one for endothelial cell culture dependent. The endothelial sprouting was then
(HUVECs) and the other for leukemic cell mono- successfully blocked by means of adding a vas-
cultures (e.g., U937, HL60, and K562 leukemic cular endothelial growth factor receptor 2
tested cell lines) or co-cultures with bone marrow (VEGFR2) blocking chimera. This patient-
stromal cell line (e.g., HS5). These channels were specific vascularized tumour model has, thus,
interfaced by a collagen-patterned compartment, proven its potential to the discovery of new tar-
and both were perfused with endothelial growth gets for blocking tumour angiogenesis in a per-
medium (EGM). Results have shown the differ- sonalized therapeutic approach (Miller et al.
ent vascular morphogenic ability, assigned by 2018).
different leukemic cell lines. Moreover, in co-
culture systems, it was evident a more discon-
certed EC response, with formation of endothelial 7.2.4 Drugs Screening
sprouts without the typical multicellular stalk
structures and a higher number of isolated tip EC, A deep understanding of the biochemical, bio-
when compared to the EC response obtained physical and cellular interplay during this pro-
under stromal cell monoculture conditions. cess on specific conditions would opening new
Although the results were in agreement with leu- opportunities for the discover of new potential
kaemia patient-derived bone marrow samples, targets for therapeutic treatment (Yoo and Kwon
the proposed microfluidic model of leukemic- 2013).
induced bone marrow-angiogenesis has to be Surpassing the natural constrains associated to
improved in order to better mimic the complex the conventional cell culture systems, microfluid-
bone marrow microenvironment (Zheng et al. ics has arisen as powerful technology for in vitro
2016). modelling specific conditions, in a more relevant
By it turn, Buchanan et al. (2014) have inves- physiological context, allowing the study of
tigated how shear stress influence the paracrine angiogenesis in different scenarios. Although the
signalling involved in angiogenesis activation, on potential and versatility of microfluidic platforms
a microvascular tumour model. Using a microflu- for fundamental research and also their promis-
114 L. Costa et al.
ing application as preclinical models for drug ously, suppressing endothelial sprouting.
screening and testing is recognized, their routine Moreover, it was observed that the drug increases
application in biological research and acceptance the permeability of the existing microvasculature
in pharmaceutical industry is, yet, limited (Lee (Zeinali et al. 2018). Although the microvascular
et al. 2018). Therefore, efforts are now directed to model has demonstrated its potential for the
turn these platforms as primary experimental study of cell behaviour under therapeutic drugs, a
models for clinical research trials, focusing more more realistic model of the pathological cellular
in their end-applicability design instead of mod- niche would prompt this type of microvascular
el’s development and validation. models to patient-specific models, where new
For instance, Kim et al. (2015a) have devel- therapeutics drugs can be screened in a personal-
oped a high-throughput microfluidic platform for ized approach.
the quantitative assessment effect of an anti- Carvalho et al. (2019) have developed a
angiogenic drug. This microfluidic device com- colorectal tumour-on-a-chip model comprising a
prised three main components: i) the ECM-like microvascular part to access the effect and effi-
chamber patterned with collagen and linked to a cacy of anti-cancer drugs delivered by nanopar-
tumour-mimic channel, where the VEGF gradi- ticles. The chip was composed by a core chamber,
ents were generated, and ii) an endothelial cell where the tumoral niche was created, by cultur-
culture channel, allowing the simultaneous appli- ing colorectal cancer cells (HCT-116 cells) into a
cation of different concentrations of the anti- 3D Matrigel supplemented with VEGF, laterally
angiogenic drug bortezomib. This quantitative connected to two microfluidic channels, cultured
platform has demonstrated its potential by means with human colonic microvascular endothelial
of enabling the determination of a specific con- cells (HCoMECs) and perfused with anticancer
centration of bortezomib inhibiting VEGF- drug-loaded nanoparticles. Although the main
induced angiogenesis, with minimal cytotoxic objective of that study was to recreate a physio-
effect on ECs (Kim et al. 2015a). logically relevant model of colorectal cancer to
Targeting a different pathological-function as a nanomedicines screening platform,
angiogenesis condition, such as the vascular the natural ability of the tumour niche, in the
remodelling associated with idiopathic pulmo- presence of VEGF, to induce endothelial sprout-
nary fibrosis, Zeinali et al. (2018) have created a ing of the HCoMECs, suggested the ability of
human microvascular model to study the effects this platform for further investigation of key tar-
of nintedanib, a potent inhibitor of tyrosine- gets involved on tumour-angiogenesis induction
kinases-angiogenic receptors, on vascular remod- and the simultaneous screening of drugs target-
elling. The microfluidic chip was composed by ing this process (Carvalho et al. 2019).
one central chamber, for co-culture of healthy- By it turn, Phan et al. (2017) have developed
patient derived normal lung fibroblasts (NL-FBs) a promising custom-fit 96 well plate array of 12
and HUVECs in a fibrin matrix, interconnected standardized and highly reproductible vascular-
with two perfusable microchannels, which were ized micro-organs, for high-throughput screen-
connected to two outermost channels, cultured ing of potential drugs. As a proof-of-concept,
with NL-FBs in the same matrix. After the assem- vascularized microtumours (VMT), consisting
bly of a perfusable microvascular network, the of colorectal cancer cell line HCT-116 embed-
system was treated with three different concen- ded in ECs and stromal cells, were established
trations of nintedanib. The results have shown in the platform and used to blindly test a set of
that nintedanib strongly decreases microvascular FDA- approved anti-cancer drugs. Anti-
density by inducing the weakening of the ECs angiogenic and anticancer drugs were success-
intercellular junctions, by inhibiting the support- fully identified, unveiling the potential of this
ive and stabilizing role of fibroblasts on the platform for screening drug effect on multiple
existing microvasculature and by, simultane- tissues (Phan et al. 2017).

7.3 Concluding Remarks driving angiogenesis overcoming the major

hurdle that TE faces, the lack of vascularization
In the past decade, the advance of microfabrica- of developed grafts.
tion techniques has enabled a great progress in Although these microfluidic platforms have
microfluidic technologies, by providing new gained prominence in disease models and for
tools for the designing of more complex and pharmaceutical research, the great majority of
sophisticated platforms. In fact, microfluidics has these platforms are confined to proof-of-concept
an unparallel potential for in vitro mimicking 3D models. Therefore, now the main challenge is to
complex cellular microenvironments in a closer move forward this technology from its theoretical
proximity to physiological conditions. For this potential to its practical application on funda-
reason, microfluidics holds the potential to fill the mental and applied research. The process would
gap between the poor representative conventional pass for the simplification and standardization of
culture systems and the complex in vivo models these platforms, turning this technology more
with their intrinsic variability and unpredictable accessible by the end-users, thus, encouraging its
associated’ factors. acceptance as integral part of a pharmaceutic
By means of offering an unlimited number of research line.
channel geometry designs, cellular/biomaterials
combination and a controlled spatiotemporal Acknowledgments The authors would like to acknowl-
delivery of biochemical and biophysical cues, by edge the financial support provided by the Portuguese
Foundation for Science and Technology (FCT) through
a continuous and controllable fluid perfusion sys- the project B-FABULUS (PTDC/BBB-ECT/2690/2014)
tems, this technology have been providing singu- and Fun4TE (PTDC/EMD-EMD/31367/2017). The FCT
lar assessment of cellular behaviour, under distinctions attributed to J. Silva-Correia (IF/00115/2015)
targeted conditions. All these advantages have and J. Miguel Oliveira (IF/01285/2015) under the
Investigator FCT program are also greatly
positioned microfluidic platforms as promising acknowledged.
primary experimental in vitro models.
Although several angiogenesis regulatory
molecules and signalling pathways have been References
identified/clarified in the past years, the under-
standing of its regulation in different contexts, Abaci HE, Drazer G, Gerecht S (2013) Recapitulating the
i.e. signalling mechanisms in a specific surround- vascular microenviroment in microfluidic platforms.
ing microenvironment, such as, on pathological Nano Life 03:1340001. https://doi.org/10.1142/
S1793984413400011
scenarios, is imperative. Abaci HE, Shen Y, Tan S, Gerecht S (2015) Recapitulating
By incorporating vasculature on these microphysiological and pathological shear stress and oxy-
fluidic platforms, it has been possible to increase gen to model vasculature in health and disease. Sci
the complexity and physiological relevance of Rep 4:4951. https://doi.org/10.1038/srep04951
Akintewe OO, Roberts EG, Rim N-G et al (2017) Design
organ-on-a-chip models and, by the other side, to approaches to myocardial and vascular tissue engi-
study angiogenesis in a variety of normal physi- neering. Annu Rev Biomed Eng 19:389–414. https://
ological and pathological contexts. Beyond the doi.org/10.1146/annurev-bioeng-071516-044641
tremendous potential of this microvascular mod- Amann A, Zwierzina M, Koeck S et al (2017)
Development of a 3D angiogenesis model to study
els to elucidate angiogenesis mechanisms in tumour – endothelial cell interactions and the effects
pathological scenarios, they hold promising of anti-angiogenic drugs. Sci Rep 7:2963. https://doi.
application as direct drug screening platforms for org/10.1038/s41598-017-03010-6
simultaneous identification and testing of new Baker BM, Trappmann B, Stapleton SC et al (2013)
Microfluidics embedded within extracellular matrix
potential targets/drugs for pathological angiogen- to define vascular architectures and pattern diffusive
esis. Additionally, microfluidic platforms incor- gradients. Lab Chip 13:3246. https://doi.org/10.1039/
porating vasculature founds also potential c3lc50493j
application for the screening of biomaterials
116 L. Costa et al.
Beck H, Acker T, Wiessner C et al (2000) Expression Chung AS, Ferrara N (2011) Developmental and
of Angiopoietin-1, Angiopoietin-2, and tie receptors pathological angiogenesis. Annu Rev Cell
after middle cerebral artery occlusion in the rat. Am Dev Biol 27:563–584. https://doi.org/10.1146/
J Pathol 157:1473–1483. https://doi.org/10.1016/ annurev-cellbio-092910-154002
S0002-9440(10)64786-4 Chung S, Sudo R, Zervantonakis IK et al (2009)
Bersini S, Moretti M (2015) 3D functional and perfusable Surface-treatment-induced three-dimensional cap-
microvascular networks for organotypic microfluidic illary morphogenesis in a microfluidic platform.
models. J Mater Sci Mater Med 26:180. https://doi. Adv Mater 21:4863–4867. https://doi.org/10.1002/
org/10.1007/s10856-015-5520-5 adma.200901727
Bhatia SN, Ingber DE (2014) Microfluidic organs- Chung M, Ahn J, Son K et al (2017) Biomimetic model
on-chips. Nat Biotechnol 32:760–772. https://doi. of tumor microenvironment on microfluidic plat-
org/10.1038/nbt.2989 form. Adv Healthc Mater 6:1700196. https://doi.
Bielenberg DR, Zetter BR (2015) The contribu- org/10.1002/adhm.201700196
tion of angiogenesis to the process of metasta- Cochrane A, Albers HJ, Passier R et al (2018) Advanced
sis. Cancer J 21:267–273. https://doi.org/10.1097/ in vitro models of vascular biology: human induced
PPO.0000000000000138 pluripotent stem cells and organ-on-chip technol-
Bischel LL, Young EWK, Mader BR, Beebe DJ (2013) ogy. Adv Drug Deliv Rev. https://doi.org/10.1016/j.
Tubeless microfluidic angiogenesis assay with addr.2018.06.007
three-dimensional endothelial-lined microvessels. Costa C, Incio J, Soares R (2007) Angiogenesis and
Biomaterials 34:1471–1477. https://doi.org/10.1016/j. chronic inflammation: cause or consequence?
biomaterials.2012.11.005 Angiogenesis 10:149–166. https://doi.org/10.1007/
Bischel LL, Sung KE, Jiménez-Torres JA et al (2014) s10456-007-9074-0
The importance of being a lumen. FASEB J 28:4583– Dai X, Cai S, Ye Q et al (2011) A novel in vitro angiogene-
4590. https://doi.org/10.1096/fj.13-243733 sis model based on a microfluidic device. Chin Sci Bull
Bogorad MI, DeStefano J, Wong AD, Searson PC (2017) 56:3301. https://doi.org/10.1007/s11434-011-4717-3
Tissue-engineered 3D microvessel and capillary net- Datta P, Ayan B, Ozbolat IT (2017) Bioprinting for
work models for the study of vascular phenomena. vascular and vascularized tissue biofabrication.
Microcirculation 24:e12360. https://doi.org/10.1111/ Acta Biomater 51:1–20. https://doi.org/10.1016/j.
micc.12360 actbio.2017.01.035
Buchanan CF, Verbridge SS, Vlachos PP, Rylander MN Del Amo C, Borau C, Gutiérrez R et al (2016)
(2014) Flow shear stress regulates endothelial barrier Quantification of angiogenic sprouting under dif-
function and expression of angiogenic factors in a 3D ferent growth factors in a microfluidic platform.
microfluidic tumor vascular model. Cell Adhes Migr J Biomech 49:1340–1346. https://doi.org/10.1016/j.
8:517–524. https://doi.org/10.4161/19336918.2014.9 jbiomech.2015.10.026
70001 Dubrac A, Künzel SE, Künzel SH et al (2018) NCK-
Caballero D, Blackburn SM, de Pablo M et al (2017) dependent pericyte migration promotes patho-
Tumour-vessel-on-a-chip models for drug delivery. logical neovascularization in ischemic retinopathy.
Lab Chip 17:3760–3771. https://doi.org/10.1039/ Nat Commun 9:3463. https://doi.org/10.1038/
C7LC00574A s41467-018-05926-7
Carmeliet P, Jain RK (2011) Molecular mechanisms and Elshabrawy HA, Chen Z, Volin MV et al (2015) The
clinical applications of angiogenesis. Nature 473:298– pathogenic role of angiogenesis in rheumatoid arthri-
307. https://doi.org/10.1038/nature10144 tis. Angiogenesis 18:433–448. https://doi.org/10.1007/
Carvalho MR, Barata D, Teixeira LM et al (2019) s10456-015-9477-2
Colorectal tumor-on-a-chip system: A 3D tool for Farahat WA, Wood LB, Zervantonakis IK et al (2012)
precision onco-nanomedicine. Sci Adv 5:eaaw1317. Ensemble analysis of Angiogenic growth in three-
https://doi.org/10.1126/sciadv.aaw1317 dimensional microfluidic cell cultures. PLoS
Chan JM, Zervantonakis IK, Rimchala T et al (2012) One 7:e37333. https://doi.org/10.1371/journal.
Engineering of in vitro 3D capillary beds by self- pone.0037333
directed angiogenic sprouting. PLoS One 7:e50582. Galie PA, Nguyen D-HT, Choi CK et al (2014) Fluid
https://doi.org/10.1371/journal.pone.0050582 shear stress threshold regulates angiogenic sprout-
Chen MB, Whisler JA, Jeon JS, Kamm RD (2013) ing. Proc Natl Acad Sci 111:7968–7973. https://doi.
Mechanisms of tumor cell extravasation in an in vitro org/10.1073/pnas.1310842111
microvascular network platform. Integr Biol 5:1262. Gavalas N, Liontos M, Trachana S-P et al (2013)
https://doi.org/10.1039/c3ib40149a Angiogenesis-related pathways in the pathogenesis
Chen L, Ito S, Kai H et al (2017) Microfluidic co-cultures of ovarian Cancer. Int J Mol Sci 14:15885–15909.
of retinal pigment epithelial cells and vascular endothe- https://doi.org/10.3390/ijms140815885
lial cells to investigate choroidal angiogenesis. Sci Rep Griffith CK, Miller C, Sainson RCA et al (2005) Diffusion
7:3538. https://doi.org/10.1038/s41598-017-03788-5 limits of an in vitro thick prevascularized tissue.
Tissue Eng 11:257–266. https://doi.org/10.1089/ Kim S, Chung M, Ahn J et al (2016b) Interstitial flow
ten.2005.11.257 regulates the angiogenic response and phenotype of
Hasan A, Paul A, Vrana NE et al (2014) Microfluidic endothelial cells in a 3D culture model. Lab Chip
techniques for development of 3D vascularized tissue. 16:4189–4199. https://doi.org/10.1039/C6LC00910G
Biomaterials 35:7308–7325. https://doi.org/10.1016/j. Kim S, Kim W, Lim S, Jeon J (2017) Vasculature-on-a-
biomaterials.2014.04.091 chip for in vitro disease models. Bioengineering 4:8.
Hayashi T, Noshita N, Sugawara T, Chan PH (2003) https://doi.org/10.3390/bioengineering4010008
Temporal profile of angiogenesis and expression of Koch S, Claesson-Welsh L (2012) Signal transduction
related genes in the brain after ischemia. J Cereb Blood by vascular endothelial growth factor receptors. Cold
Flow Metab 23:166–180. https://doi.org/10.1097/01. Spring Harb Perspect Med 2:a006502–a006502.
WCB.0000041283.53351.CB https://doi.org/10.1101/cshperspect.a006502
Jeon JS, Zervantonakis IK, Chung S et al (2013) In vitro Kofler NM, Shawber CJ, Kangsamaksin T et al (2011)
model of tumor cell extravasation. PLoS One 8:e56910. Notch signaling in developmental and tumor angio-
https://doi.org/10.1371/journal.pone.0056910 genesis. Genes Cancer 2:1106–1116. https://doi.
Jeon JS, Bersini S, Whisler JA et al (2014) Generation org/10.1177/1947601911423030
of 3D functional microvascular networks with Krock BL, Skuli N, Simon MC (2011) Hypoxia-induced
human mesenchymal stem cells in microfluidic sys- angiogenesis: good and evil. Genes Cancer 2:1117–
tems. Integr Biol 6:555–563. https://doi.org/10.1039/ 1133. https://doi.org/10.1177/1947601911423654
C3IB40267C Laschke MW, Menger MD (2016) Prevascularization
Jeong GS, Han S, Shin Y et al (2011a) Sprouting angio- in tissue engineering: current concepts and future
genesis under a chemical gradient regulated by directions. Biotechnol Adv 34:112–121. https://doi.
interactions with an endothelial monolayer in a micro- org/10.1016/j.biotechadv.2015.12.004
fluidic platform. Anal Chem 83:8454–8459. https:// Lee H, Park W, Ryu H, Jeon NL (2014) A microfluidic
doi.org/10.1021/ac202170e platform for quantitative analysis of cancer angiogen-
Jeong GS, Kwon GH, Kang AR et al (2011b) Microfluidic esis and intravasation. Biomicrofluidics 8:054102.
assay of endothelial cell migration in 3D interpene- https://doi.org/10.1063/1.4894595
trating polymer semi-network HA-Collagen hydro- Lee S, Ko J, Park D et al (2018) Microfluidic-based
gel. Biomed Microdevices 13:717–723. https://doi. vascularized microphysiological systems. Lab Chip
org/10.1007/s10544-011-9541-7 18:2686–2709. https://doi.org/10.1039/C8LC00285A
Kant RJ, Coulombe KLK (2018) Integrated approaches Lertkiatmongkol P, Liao D, Mei H et al (2016) Endothelial
to spatiotemporally directing angiogenesis in host and functions of platelet/endothelial cell adhesion mol-
engineered tissues. Acta Biomater 69:42–62. https:// ecule-1 (CD31). Curr Opin Hematol 23:253–259.
doi.org/10.1016/j.actbio.2018.01.017 https://doi.org/10.1097/MOH.0000000000000239
Khademhosseini A, Langer R, Borenstein J, Vacanti JP Lewis DM, Gerecht S (2016) Microfluidics and bio-
(2006) Microscale technologies for tissue engineer- materials to study angiogenesis. Curr Opin
ing and biology. Proc Natl Acad Sci 103:2480–2487. Chem Eng 11:114–122. https://doi.org/10.1016/j.
https://doi.org/10.1073/pnas.0507681102 coche.2016.02.005
Kim S, Lee H, Chung M, Jeon NL (2013a) Engineering Liakouli V, Cipriani P, Marrelli A et al (2011) Angiogenic
of functional, perfusable 3D microvascular networks cytokines and growth factors in systemic scle-
on a chip. Lab Chip 13:1489. https://doi.org/10.1039/ rosis. Autoimmun Rev 10:590–594. https://doi.
c3lc41320a org/10.1016/j.autrev.2011.04.019
Kim Y-W, West XZ, Byzova TV (2013b) Inflammation Lin L, Lin X, Lin L et al (2017) Integrated microfluidic
and oxidative stress in angiogenesis and vascular dis- platform with multiple functions to probe tumor–
ease. J Mol Med 91:323–328. https://doi.org/10.1007/ endothelial cell interaction. Anal Chem 89:10037–
s00109-013-1007-3 10044. https://doi.org/10.1021/acs.analchem.7b02593
Kim C, Kasuya J, Jeon J et al (2015a) A quantitative Liu J, Wang X, Yang X et al (2015) Investigating
microfluidic angiogenesis screen for studying anti- the role of angiogenesis in systemic lupus ery-
angiogenic therapeutic drugs. Lab Chip 15:301–310. thematosus. Lupus 24:621–627. https://doi.
https://doi.org/10.1039/C4LC00866A org/10.1177/0961203314556293
Kim J, Chung M, Kim S et al (2015b) Engineering of a Mannino RG, Pandian NKR, Jain A, Lam WA (2018a)
biomimetic Pericyte-covered 3D microvascular net- Engineering “endothelialized” microfluidics for
work. PLoS One 10:1–15. https://doi.org/10.1371/ investigating vascular and hematologic processes
journal.pone.0133880 using non-traditional fabrication techniques. Curr
Kim JJ, Hou L, Huang NF (2016a) Vascularization of Opin Biomed Eng 5:13–20. https://doi.org/10.1016/j.
three-dimensional engineered tissues for regenera- cobme.2017.11.006
tive medicine applications. Acta Biomater 41:17–26. Mannino RG, Qiu Y, Lam WA (2018b) Endothelial cell
https://doi.org/10.1016/j.actbio.2016.06.001 culture in microfluidic devices for investigating
118 L. Costa et al.
microvascular processes. Biomicrofluidics 12:042203. ing static networks. Trends Biotechnol 34:733–745.
https://doi.org/10.1063/1.5024901 https://doi.org/10.1016/j.tibtech.2016.03.002
Marrelli A, Cipriani P, Liakouli V et al (2011) Sadr N, Zhu M, Osaki T et al (2011) SAM-based cell trans-
Angiogenesis in rheumatoid arthritis: a disease spe- fer to Photopatterned hydrogels for microengineering
cific process or a common response to chronic inflam- vascular-like structures. Biomaterials 32:7479–7490.
mation? Autoimmun Rev 10:595–598. https://doi. https://doi.org/10.1016/j.biomaterials.2011.06.034
org/10.1016/j.autrev.2011.04.020 Sakthivel K, O’Brien A, Kim K, Hoorfar M (2019)
Michna R, Gadde M, Ozkan A et al (2018) Vascularized Microfluidic analysis of heterotypic cellular inter-
microfluidic platforms to mimic the tumor microenvi- actions: a review of techniques and applications.
ronment. Biotechnol Bioeng 115:2793–2806. https:// TrAC Trends Anal Chem. https://doi.org/10.1016/j.
doi.org/10.1002/bit.26778 trac.2019.03.026
Miller JS, Stevens KR, Yang MT et al (2012) Rapid Sato M, Sasaki N, Ato M et al (2015) Microcirculation-
casting of patterned vascular networks for perfus- on-a-chip: A microfluidic platform for assaying
able engineered three-dimensional tissues. Nat Mater blood- and lymphatic-vessel permeability. PLoS
11:768–774. https://doi.org/10.1038/nmat3357 One 10:e0137301. https://doi.org/10.1371/journal.
Miller CP, Tsuchida C, Zheng Y et al (2018) A 3D human pone.0137301
renal cell carcinoma-on-a-chip for the study of tumor Shamloo A, Heilshorn SC (2010) Matrix density medi-
angiogenesis. Neoplasia 20:610–620. https://doi. ates polarization and lumen formation of endothelial
org/10.1016/j.neo.2018.02.011 sprouts in VEGF gradients. Lab Chip 10:3061. https://
Mongiat M, Andreuzzi E, Tarticchio G, Paulitti A (2016) doi.org/10.1039/c005069e
Extracellular matrix, a hard player in angiogen- Shin Y, Jeon JS, Han S et al (2011) In vitro 3D collec-
esis. Int J Mol Sci 17:1822. https://doi.org/10.3390/ tive sprouting angiogenesis under orchestrated ANG-1
ijms17111822 and VEGF gradients. Lab Chip 11:2175. https://doi.
Multhoff G, Radons J, Vaupel P (2014) Critical role org/10.1039/c1lc20039a
of aberrant angiogenesis in the development of Skuli N, Majmundar AJ, Krock BL et al (2012) Endothelial
tumor hypoxia and associated radioresistance. HIF-2α regulates murine pathological angiogen-
Cancers (Basel) 6:813–828. https://doi.org/10.3390/ esis and revascularization processes. J Clin Invest
cancers6020813 122:1427–1443. https://doi.org/10.1172/JCI57322
Nguyen D-HT, Stapleton SC, Yang MT et al (2013) Smith Q, Gerecht S (2014) Going with the flow: micro-
Biomimetic model to reconstitute angiogenic sprouting fluidic platforms in vascular tissue engineering. Curr
morphogenesis in vitro. Proc Natl Acad Sci 110:6712– Opin Chem Eng 3:42–50. https://doi.org/10.1016/j.
6717. https://doi.org/10.1073/pnas.1221526110 coche.2013.11.001
Norotte C, Marga FS, Niklason LE, Forgacs G (2009) Song M, Finley SD (2018) Mechanistic insight into acti-
Scaffold-free vascular tissue engineering using bio- vation of MAPK signaling by pro-angiogenic fac-
printing. Biomaterials 30:5910–5917. https://doi. tors. BMC Syst Biol 12:145. https://doi.org/10.1186/
org/10.1016/j.biomaterials.2009.06.034 s12918-018-0668-5
Novosel EC, Kleinhans C, Kluger PJ (2011) Song JW, Munn LL (2011) Fluid forces control endothe-
Vascularization is the key challenge in tissue engi- lial sprouting. Proc Natl Acad Sci 108:15342–15347.
neering. Adv Drug Deliv Rev 63:300–311. https://doi. https://doi.org/10.1073/pnas.1105316108
org/10.1016/j.addr.2011.03.004 Sontheimer-Phelps A, Hassell BA, Ingber DE (2019)
Otrock Z, Mahfouz R, Makarem J, Shamseddine A (2007) Modelling cancer in microfluidic human organs-
Understanding the biology of angiogenesis: review on-chips. Nat Rev Cancer 19:65–81. https://doi.
of the most important molecular mechanisms. Blood org/10.1038/s41568-018-0104-6
Cells Mol Dis 39:212–220. https://doi.org/10.1016/j. Sudo R, Chung S, Shin Y, Tanishita K (2016) Integrated
bcmd.2007.04.001 vascular engineering: vascularization of reconstructed
Pafumi I, Favia A, Gambara G et al (2015) Regulation tissue. In: Tanishita K, Yamamoto K (eds) Vascular
of Angiogenic functions by angiopoietins through engineering. Springer Japan, Tokyo, pp 297–332
calcium-dependent signaling pathways. Biomed Res Theberge AB, Yu J, Young EWK et al (2015) Microfluidic
Int 2015:1–14. https://doi.org/10.1155/2015/965271 multiculture assay to analyze biomolecular signaling
Phan DTT, Wang X, Craver BM et al (2017) A vascu- in angiogenesis. Anal Chem 87:3239–3246. https://
larized and perfused organ-on-a-chip platform for doi.org/10.1021/ac503700f
large-scale drug screening applications. Lab Chip Uwamori H, Ono Y, Yamashita T et al (2019) Comparison
17:511–520. https://doi.org/10.1039/C6LC01422D of organ-specific endothelial cells in terms of micro-
Raasch M, Rennert K, Jahn T et al (2015) vascular formation and endothelial barrier functions.
Microfluidically supported biochip design for cul- Microvasc Res 122:60–70. https://doi.org/10.1016/j.
ture of endothelial cell layers with improved perfu- mvr.2018.11.007
sion conditions. Biofabrication 7:015013. https://doi. van der Meer AD, Poot AA, Duits MHG et al (2009)
org/10.1088/1758-5090/7/1/015013 Microfluidic technology in vascular research.
Rouwkema J, Khademhosseini A (2016) Vascularization J Biomed Biotechnol 2009:1–10. https://doi.
and angiogenesis in tissue engineering: beyond creat- org/10.1155/2009/823148
van Duinen V, Zhu D, Ramakers C et al (2019) Perfused Yoshida H, Matsusaki M, Akashi M (2013) Multilayered
3D angiogenic sprouting in a high-throughput in vitro blood capillary analogs in biodegradable hydro-
platform. Angiogenesis 22:157–165. https://doi. gels for in vitro drug permeability assays. Adv
org/10.1007/s10456-018-9647-0 Funct Mater 23:1736–1742. https://doi.org/10.1002/
Varricchi G, Granata F, Loffredo S et al (2015) adfm.201201905
Angiogenesis and lymphangiogenesis in inflamma- Young EWK (2013) Advances in microfluidic cell
tory skin disorders. J Am Acad Dermatol 73:144–153. culture Systems for Studying Angiogenesis.
https://doi.org/10.1016/j.jaad.2015.03.041 J Lab Autom 18:427–436. https://doi.
Wang X, Sun Q, Pei J (2018) Microfluidic-based 3D engi- org/10.1177/2211068213495206
neered microvascular networks and their applications Zanotelli MR, Ardalani H, Zhang J et al (2016) Stable
in vascularized microtumor models. Micromachines engineered vascular networks from human induced
9:493. https://doi.org/10.3390/mi9100493 pluripotent stem cell-derived endothelial cells cul-
Weis SM, Cheresh DA (2011) Tumor angiogenesis: tured in synthetic hydrogels. Acta Biomater 35:32–41.
molecular pathways and therapeutic targets. Nat Med https://doi.org/10.1016/j.actbio.2016.03.001
17:1359–1370. https://doi.org/10.1038/nm.2537 Zeinali S, Bichsel CA, Hobi N et al (2018) Human
Wragg JW, Durant S, McGettrick HM et al (2014) Shear microvasculature- on-a chip: anti-neovasculogenic
stress regulated gene expression and angiogenesis in effect of nintedanib in vitro. Angiogenesis 21:861–
vascular endothelium. Microcirculation 21:290–300. 871. https://doi.org/10.1007/s10456-018-9631-8
https://doi.org/10.1111/micc.12119 Zervantonakis IK, Hughes-Alford SK, Charest JL et al
Wu W, DeConinck A, Lewis JA (2011) Omnidirectional (2012) Three-dimensional microfluidic model for
printing of 3D microvascular networks. Adv tumor cell intravasation and endothelial barrier func-
Mater 23:H178–H183. https://doi.org/10.1002/ tion. Proc Natl Acad Sci 109:13515–13520. https://
adma.201004625 doi.org/10.1073/pnas.1210182109
Yeon JH, Ryu HR, Chung M et al (2012) In vitro for- Zhao Y, Adjei AA (2015) Targeting angiogenesis in
mation and characterization of a perfusable three- Cancer therapy: moving beyond vascular endothelial
dimensional tubular capillary network in microfluidic growth factor. Oncologist 20:660–673. https://doi.
devices. Lab Chip 12:2815. https://doi.org/10.1039/ org/10.1634/theoncologist.2014-0465
c2lc40131b Zheng Y, Sun Y, Yu X et al (2016) Angiogenesis in liq-
Yoo SY, Kwon SM (2013) Angiogenesis and its therapeu- uid tumors: an in vitro assay for leukemic-cell-
tic opportunities. Mediat Inflamm 2013:1–11. https:// induced bone marrow angiogenesis. Adv Healthc
doi.org/10.1155/2013/127170 Mater 5:1014–1024. https://doi.org/10.1002/
adhm.201501007
Biomaterials and Microfluidics
for Drug Discovery 8
and Development
Mariana R. Carvalho, Roman Truckenmuller,
Rui Luís Reis, and Joaquim Miguel Oliveira
Abstract nificant interest in the drug discovery and

Microfluidic devices are now one of the most development domain. Herein, the most recent
promising tools to mimic in vivo like condi- advances in the field of microfluidics for drug
tions, either in normal or disease scenarios, discovery are overviewed, and the role of bio-
such as tumorigenesis or pathogenesis. materials in 3D in vitro models and the contri-
Together with the potential of biomaterials, its bution of organ-on-a-chip technologies
combination with microfluidics represents the highlighted.
ability to more closely mimic cells’ natural
microenvironment concerning its three- Keywords
dimensional (3D) nature and continuous per- 3D models · Biomaterials · Drug discovery ·
fusion with nutrients and cells’ crosstalk. Due Microfluidics · Organ-on-a-chip · Cancer
to miniaturization and increased experimental
throughput, microfluidics have generated sig-
Highlights
M. R. Carvalho · R. L. Reis · J. M. Oliveira (*) • Microfluidic devices are now one of the most
Biomaterials, Biodegradables and Biomimetics, promising tools to mimic in vivo like condi-
University of Minho, Headquarters of the European tions, either in normal or disease scenarios,
Institute of Excellence on Tissue Engineering and such as tumorigenesis or pathogenesis.
Regenerative Medicine, Guimarães, Portugal • In the case of cancer drugs, studies report the
ICVS/3B’s PT Government Associate Lab, rate of successful translation from animal
Braga/Guimarães, Portugal models to clinical cancer trials is less than 8%,
The Discoveries Centre for Regenerative and revealing the urgent need to develop better
Precision Medicine, Headquarters at University of models.
Minho, Guimarães, Portugal
e-mail: miguel.oliveira@dep.uminho.pt • The advances in the fields of biomaterials
engineering combined with microfluidics
R. Truckenmuller
Department of Complex Tissue Regeneration, have advanced to impressive levels of
MERLN Institute for Technology-Inspired
Regenerative Medicine, Maastricht University,
Maastricht, The Netherlands

https://doi.org/10.1007/978-3-030-36588-2_8
122 M. R. Carvalho et al.
sophistication, helping in the drug discovery drug development (Zheng et al. 2013). Estimates
and development. predict HTS products only contribute 19–33%
(or less) to successful marketed drugs (Khanna
2012).
8.1 Introduction Regarding the animal experimentation,
although this step is indispensable for pre-clinical
Many diseases, particularly acute disorders, are screening in the drug discovery process, various
now treatable or manageable very effectively. issues such as ethical considerations and species
The discovery of new medications for health differences remain. In the case of cancer drugs,
problems that affect a great part of the world’s studies report that the average rate of successful
population, such as cancer, cardiovascular dis- translation from animal models to clinical cancer
eases, metabolic disorders, arthritis, depression, trials is less than 8% (Mak et al. 2014; Szymański
anxiety, gastrointestinal disorders, pain, infec- et al. 2012).
tious diseases and many others, have led to Although three-dimensional (3D) culture
improvement in overall health (Mullard 2011). techniques were introduced to the HTS tech-
As a result, the increased quality of life may lead niques, to construct a more natural extracellular
to increased life expectancy (Mullard 2011). microenvironment, most of these conventional
The drug discovery development process tra- screening platforms are still limited in applicabil-
ditionally starts in the laboratory where cells cul- ity due to their reliance on static culture mainte-
tured on flat bottom flasks are observed in their nance. To tackle these issues, 3D cell-based in
response to a drug, and finally moves through the vitro assays have been actively pursued, most of
established chain of drug development: animal the times using hydrogels. However, it remains
and human clinical trials. Only then the product difficult to accurately predict drug efficacy, toxic-
can be approved and commercialized (Hughes ity, and organs interactions, because cultivated
et al. 2011). However, it is well described in the cells often do not retain their original organ func-
literature that the vast majority of “hits” fail tions and morphologies in conventional 2D in
between animal testing and clinical trials. The vitro cell culture systems (Chandrasekaran et al.
very one successful drug, from being a promising 2016b).
“hit” to be launched as a finished product takes As a promising solution, microfluidic tech-
between 12 and 15 years and can exceed $1 bil- nologies emerged, in this case those also provid-
lion in costs (Paul et al. 2010; Mak et al. 2014). ing 3D microenvironment with microvascular-like
Among the primary causes of failure, nonclini- perfusion and diffusion between mimicked
cal/clinical safety (>50%) and efficacy (>10%) microvessels and 3D cell culture, which is closer
stand out from all other factors (e.g., strategic, to what cells encounter in real tissues or organs.
commercial or operational) (Hay et al. 2014). Briefly, microfluidics is the field of fluids
More recently, scientific development lead us manipulation in channels with dimensions of tens
to cell-based high-throughput screening (HTS) of micrometers (Whitesides 2006). Microfluidic
techniques, which have enabled researchers to chips are small platforms comprising channel
screen more than a million compounds in a cou- systems connected to liquid reservoirs by, for
ple of months (Hughes et al. 2011). However, example, tubing systems in turn linked to
still only about 1 out of 5000 promising drugs syringes. The size of the channels is in the range
successfully makes it to the market. Even though of a few micrometers, which greatly facilitates
HTS improves the screening speed and is now handling of volumes much smaller than a micro-
considered a standard platform for drug discov- liter (Bettinger and Borenstein 2010).
ery, it leaves several unresolved issues, being the In combination with appropriate biological
monetary costs and efficiency on top. Current assays and high-sensitivity detection techniques,
pharmaceutical industries employ HTS primarily such systems allow the identification and isola-
for chemical optimization at the early stage of the tion of individual cells or molecules (Dittrich and
8 Biomaterials and Microfluidics for Drug Discovery and Development 123
Manz 2006). A great variety of microchip fabri- extensively reported (Radhakrishnan et al. 2017).
cation techniques and materials are available for The use of a variety of hydrogels based on syn-
producing highly sophisticated 2½D and 3D thetic polymers (e.g. poly(ethylene glycol, PEG),
microstructures with integrated modules. polyvinyl alcohol, poly(N,Ndimethylacrylamide),
Sensors, detectors and optical components can be and methoxy poly(ethylene glycol)-poly(e-
integrated on chip (Tsao 2016). caprolactone) have also been widely reported in
With rapid growing on emphasis on transla- literature (Lai et al. 2013; Cui et al. 2012; Bichara
tional research and more matured bio-et al. 2014; de Girolamo et al. 2015; Inagaki et al.
microfabrication technologies, currently in the 2014).
USA alone there are over three companies using Cross-linked hydrogel networks can be engi-
microchips for in vivo drug delivery, at least five neered on a molecular basis to tune for example
microchip-based cancer cell/biomarker diagnos- physical parameters such as mesh size and with it
tic companies and about 10 companies working mechanical properties such as stiffness according
on microfabricated cell chips for drug screening to the desired characteristics (Zeng et al. 2014).
and discovery (FluidicMEMS 2018). Hydrogel networks, both natural and synthetic,
This book chapter focuses on current and are therefore able to match the physical, chemi-
future applications of biomaterials in microfluid- cal, and mechanical properties of native
ics, microfluidics systems in drug discovery and Extracellular Matrix (ECM) molecules, justify-
highlights the potential of ‘organ-on-a-chip’ ing the widespread use of hydrogels as ECM
technology for drug discovery. mimics in microfluidic models (Lutolf and
Hubbell 2005; Barata et al. 2016). Various types
of hydrogels have been used in microfluidic sys-
8.2 Biomaterials Applied tems (Wan 2012) with collagen (Baker et al.
in Microfluidic Systems 2013; Toh et al. 2009), poly(ethylene glycol)
(PEG) (Dinh et al. 2013; Kobel and Lutolf 2011),
Biomaterials are gradually being developed as in agarose (Si et al. 2012), fibrin (Bita et al. 2010;
vitro microenvironments mimicking in vivo cell Toh et al. 2009), and Matrigel (Anguiano et al.
niches, as well as widely used in tissue engineer- 2017; Chaw et al. 2007) being some of the most
ing for regeneration and replication of diverse commonly used ones.
types of both normal and diseased tissues Toh et al. (2009) used collagen to develop a
(Shantanu et al. 2016). The development of 3D microfluidic 3D hepatocyte chip (‘3D HepaTox
culture models with tunable biomaterials has Chip’) for in vitro drug toxicity testing. The 3D
facilitated the investigation of biological phe- HepaTox Chip is based on multiplexed microflu-
nomena with a higher level of complexity and idic channels where a 3D microenvironment is
physiological relevance than the traditional 2D engineered in each of the eight channels channels
models (Carvalho et al. 2015). to maintain the hepatocytes’ synthetic and meta-
The combination of microfluidic technology bolic functions (Toh et al. 2009). Positively
and biomaterials now promises to enable the charged methylated collagen and negatively
recapitulation of important aspects of the spatial charged HEMA-MMA-MAA terpolymer were
and temporal characteristics of complex native chosen to form a localized 3D matrix because
niches, therefore facilitating the identification of they have been shown to support hepatocyte
new mechanisms of cell regulation and pathogen- functions.
esis, profoundly impacting drug discovery (Kobel Dinh et al. (2013) combined PEG with the
and Lutolf 2011). benefits of microfluidics for precision single-cell
Briefly, hydrogels can be derived from natural handling. The result was a model comprising
polymers, including chitosan, alginate, hyal- compartmentalized neuron arraying (CNA)
uronic acid (HA), gellan gum (GG), agarose, col- microfluidic circuits that allow high-throughput
lagen, gelatin, and fibroin, which have been experimentation (Fig. 8.1)(Dinh et al. 2013).
Fig. 8.1 – Biomaterials applied in microfluidic systems. HepG2 cells were encapsulated inside cylindrical hydro-
(1) In situ neuron patterning. Differentiated human gel microstructures (78 μm Ø × 50 μm), while A549 cells
SH-SY5Y neurons cultured for 5 days within the CNA were encapsulated inside rectangular hydrogel micro-
devices on a PL/PLL-g-PEG pattern (a) and on a continu- structures (330 μm × 217 μm × 50 μm). (b) Corresponding
ous PL coating (b). Biomaterial patterning increases neu- fluorescence image for the cell viability expressed by live/
rite guidance into the outgrowth microchannels. dead assay (live cells green, dead cells red). (d) Enlarged
(Reprinted with permission from (Dinh et al. 2013). optical and fluorescence images of one array element.
Copyright © 2018 Royal Society of Chemistry). 2 – (C) (Reprinted with permission from (Gao et al. 2010).
Optical image of hydrogel microstructure arrays contain- Copyright © 2010 Elsevier B.V)
ing two phenotypes of cells: HepG2 and A549 cells.
Similarly, Gao et al. (2010) employed photoli- To address the global challenge of vascular-
thography to geometrically encapsulate human ization, Zhang et al. (2013) employed additive
hepatoma HepG2 and human lung epithelial manufacturing to build a blood vessel-like micro-
A549 cells in a PEG hydrogel inside a microflu- fluidic structure, which was then embedded
idic device. The cells in the device were exposed inside the bulk material. To achieve this, alginate
to anticancer drugs and apoptosis was studied and chitosan hydrogels were used for direct print-
(Fig. 8.2). Two anticancer drugs (Actinomycin D ing of the channels with a wall thickness below
and methotrexate) exhibited distinct effects on 200 μm. The printed microfluidic network was
the levels of intracellular glutathione and reactive able to sustain media perfusion, both without and
oxygen species, indicating the selectivity of these with being embedded in bulk hydrogels to sup-
drugs on the disturbance of redox balance within port cell viability, representing an improvement
cells (Gao et al. 2010). towards vascularization models (Zhang et al.
2013).
Fig. 8.2 – Methods of microfluidics in drug discovery. arrays. Initially, a PDMS microfluidic mold in the form of
(1) On-chip reaction. (a) Schematic of the process of on- an array of microchannels was aligned on an array of
chip reaction. (b) Frames of a video that correspond to the microwells. Each cell type was flown through an indepen-
steps of A. A concentrated solution of crystal violet (a pH dent channel. The cells docked onto the corresponding
indicator) in 0.1 N sodium hydroxide solution is loaded micro-wells, which resulted in a patterned array of cells.
into the microchamber with the three-state valve half- To deliver multiple solutions, the PDMS microfluidic
open; loading an equal amount of 1 N HCl into the cham- mold was removed and replaced with another mold which
ber turns the solution from blue to yellow. Reprinted with was placed orthogonally to create multiphenotype cell
permission from (Wu et al. 2004). Copyright © 2004 The arrays inside each microchannel. (Reprinted with permis-
National Academy of Sciences; (2) Schematic diagram of sion from (Khademhosseini et al. 2005). Copyright ©
reversible sealing of microfluidic arrays onto microwell- 2018 Copyright Clearance Center, Inc)
patterned substrates to fabricate multiphenotype cell
Regarding fibrin hydrogels, Carrion et al. tions in several fields from drug discovery to
(Bita et al. 2010) described a novel 3D microflu- stem cell research.
idic device as a model system to study the molec- ECM derived biomaterials approaches are
ular regulation of perivascular stem cell niches. now being studied for their potential to preserve
Endothelial cells (ECs) suspended within 3D the tissue-specific biochemical composition as
fibrin hydrogels patterned in the device adjacent well as the ultrastructure of the ECM (Hussey
to stromal cells (either fibroblasts or human bone et al. 2018). The development of biomaterials
marrow-derived mesenchymal stromal cells, that promote the formation of functional tissues
hBM-MSCs) executed a morphogenetic process in clinical applications is therefore of the utmost
similar to vasculogenesis, maturing into a robust importance in tissue engineering. So far, many
capillary network with well-defined hollow studies have focused on the decellularization of
lumens. Both MSCs and fibroblasts formed peri- tissues and organs, including small intestinal sub-
cytic associations with the ECs but promoted mucosa, heart valve, blood vessel, skin, nerve,
capillary morphogenesis with distinct kinetics tendon, ligament, bladder, amniotic membrane,
(Bita et al. 2010). heart, liver and lung (Uygun et al. 2010; Choi
The fields of biomaterials engineering and et al. 2010; Lu et al. 2011). For example, Choi
microfluidics have advanced to impressive levels et al. (2010) decellularized an ECM scaffold
of sophistication. The pioneering efforts reported derived from porcine cartilage and proved it was
above to further integrate biomaterials technol- more efficient in maintain chondrogenesis of rab-
ogy into microfluidic chips tackle pertinent ques- bit mesenchymal stem cells (rMSCs) in vitro and
in the nude mouse model in vivo than the PGA can be used to address this issue. These devices
scaffold used as a control (Choi et al. 2010). are integrated to manipulate, label, lyse, separate
Uygun et al. (2010) generated a transplantable and quantify the protein contents of single cells
liver graft using perfusion decellularization tech- using single-molecule fluorescence counting
nique, and further introduced perfusion-seeding (Wu et al. 2004) (Fig. 8.2). One of the many suc-
and culture techniques for the preparation of cessful examples consists of measuring the num-
recellularized liver matrix for transplantation. ber of epitope-tagged human β2-adrenergic
The recellularized graft supports liver-specific receptors (β2ARs), an important pharmacologic
function including albumin secretion, urea syn- target in a number of airway and cardiovascular
thesis and cytochrome P450 expression at com- diseases. Another interesting application of
parable levels to normal liver in vitro (Uygun microfluidic devices in drug discovery is for
et al. 2010). ligand binding studies, minimizing interaction
Inorganic compounds such as hydroxyapatite times, improving sensitivity and increasing
(HA) have also been integrated in microfluidic throughput. In this sense, a microfluidic platform
models for mimicking bone tissues. For example, has been used to characterize DNA binding
Jusoh et al. (2015) developed a microfluidic plat- energy using four eukaryotic transcription factors
form that integrates fibrin ECM with the syn- to predict their in vivo function (Maerkl and
thetic bone mineral HA nanoparticles to provide Quake 2007).
in vivo-like microenvironments for bone vessel
sprouting. The observed that fibrin with 0.2% HA (b) For hit identification
exhibited a higher number of sprouts, and higher b1) Compound generation: Another way is
length of sprouts (Jusoh et al. 2015). through the generation of compounds in
microfluidic devices, either at a 2D or 3D
device-architectural level. Developments
8.3 Microfluidics in Drug in combinatorial chemistry have greatly
Discovery enhanced our ability to generate drug can-
didates (Kang et al. 2008). Merging chan-
Microfluidic technologies represent an improve- nel geometries, various reagents can be
ment to existing technologies in key areas of drug mixed and biochemical reactions can be
discovery. Microfluidics involves materials and induced. Reduction of reaction chamber’s
techniques for controlling the movement of min- size and consequently amount of reagents
ute quantities of fluids, and its ability to miniatur- and time needed for reaction empower
ize assays and increase experimental throughput this strategy. Mitchell et al. (2001) devel-
have generated significant interest of the drug oped a miniaturized-synthesis/total analy-
discovery and development domain (Kang et al. sis system that incorporated a chemical
2008). microprocessor with time-of-flight mass
Microfluidics can be applied to drug discovery spectrometry (TOF-MS) which has been
in a variety of ways. For instance, it can be used used in the generation of compound
in target identification, through platforms for cell libraries (Mitchell et al. 2001). This sys-
culture in which cell’s response to a drug in a tem enabled real-time processing of many
more in vivo like environment is assessed (Barata reactions due to the integration of
et al. 2017). continuous-flow synthesis and provision
for on-line analysis within a microfabri-
(a) To select specific targets cated structure. In another example, Zhou
et al. (2004) developed a microfluidic pico
In drug discovery, the signal transduction array device for the synthesis and purifi-
pathways and protein-protein interactions within cation of oligonucleotides. This device
cells must be understood. Microfluidic devices consisted of fluidic channels and 3698
individual pico-reaction chambers for Concentration gradients also have a great role
parallel synthesis of oligonucleotides in drug screening and cell-based studies. A
(Zhou et al. 2004). Microfluidic systems gradient-generating microfluidic device can gen-
can be also used to understand and assess erate a certain range of doses and thereby control
natural drug candidates. Natural products different cell behaviors at the same time. By
are undoubtedly one of the major sources using gradients, cells can be stimulated with con-
of new chemical entities, with an increase trolled temporal and spatial resolution to study
in world market growth rate (Shen 2015). the effects of drug concentration on chemotaxis.
Natural molecules are “evolutionarily Occhetta et al. (2015) introduced a branched
optimized” as drug-like molecules and channel to generate a gradient and culture 3D
remain one of the best sources of drugs micromasses of adult hBM-MSCs under continu-
leads (Shen 2015). To identify the physi- ous and controlled laminar flow perfusion
ological action of a natural compound out (Occhetta et al. 2015). Moreover, the platform
of a complex array of constituent mole- allows for conditioning of the cellular microag-
cules in the source material, microfluidic gregates with different combinations of growth
devices can also be used. Gramowski factors, morphogens or drug molecules, in a
et al. (2006) applied multielectrode micro- high-throughput fashion.
chips for the screening of herbal medi-
cines. The screening of complex mixtures
of neuroactive substances was assessed in 8.4 Organ on Chip and Drug
the chip, where changes in the spontane- Discovery
ous activity of cultured networks of pri-
mary cortical neurons were quantified to Recent scientific developments in areas such as
evaluate the action of the drugs stem cell research, regenerative medicine, bioma-
(Gramowski et al. 2006). terials, tissue engineering and microfluidics have
b2) HTS is considered one of the most impor- been integrated into new 3D in vitro models in
tant methods for identifying “hit” com- order to better mimic human tissues and organs.
pounds (Chandrasekaran et al. 2016a). These “organs-on-chip” devices can provide not
Over the last decades, pharmaceutical only the biological relevance but also the requi-
companies established HTS as a major site high-throughput applications in drug discov-
tool to screen the properties of new chem- ery. In the following, we provide an overview of
ical entities. A significant number of the developed technologies of organ-on-chip
microfluidic technologies can now be devices and their role on drug testing and
used to enable HTS studies, including discovery.
multiplexed systems, microwell arrays,
plug-based methods and gradient-
generating devices (Du et al. 2016). 8.4.1 Lung on a Chip
Khademhosseini et al. (2005) studied the Huh et al. (2012) developed a biomimetic micro-
response of multiple cell types (e.g. hepatocytes, device that is able to reconstitute organ-level lung
fibroblasts, and embryonic stem cells) to differ- functions to create a human disease: pulmonary
ent compounds by seeding cells within microw- edema (as well as a model that mimics normal
ells that were integrated within an array of lung function) (Fig. 8.3). In this model, the
reversibly sealed microfluidic channels (Fig. 8.2). authors cultured two types of human lung cells
The ability to position many cell types on a single (epithelial and endothelial primary lung cells) in
chip could be useful for studying the effects of a stacked parallel microchannels separated by a
series of compounds on different cell types. (porous) membrane. Alike the human lung, the
upper alveolar channel was filled with air,
Fig. 8.3 – (1) Wxamples of organs on chips: A microen- of the hepatocyte microenvironment in liver tissue. (c)
gineered model of human pulmonary edema. (a) IL-2 The sinusoid space is bordered by a sheet of highly fenes-
therapy is associated with vascular leakage. (b) IL-2– trated endothelial cells. (d) The single microfluidic sinu-
induced pulmonary edema is modeled in a microengi- soid had three fluid terminals: a flow inlet, a flow outlet,
neered lung-on-a-chip. The top or “air” portion is the and a cell inlet. Scale bar represents 20 μm. Reprinted
alveolar channel; the bottom or “liquid” portion is the with permission from (Lee et al. 2007). (Copyright ©
vascular channel. Scale bar represents 200 μm. (c) 2007 Wiley Periodicals, Inc. (3). Higher-throughput heart
Endothelial exposure to IL-2 (1000 U/ml) causes liquid on a chip and fluidic microdevice). (a) Schematic of fabri-
from the lower, microvascular channel to leak into the cation process. (b) Image of the engraving laser process-
upper, alveolar chamber. Scale bars represent 200 μm. (d) ing of Muscular Thin Films (MTFs) in the 18 mm diameter
During IL-2 treatment, prothrombin (100 μg/ml) and fluo- chip. (c) Exploded view of the conception and assembly
rescently labeled fibrinogen (2 mg/ml) introduced into the of the fluidic device which fits the chip. (d) Image of an
microvascular channel form fluorescent fibrin clots actual device in action. (Reprinted with permission from
(white) over the course of 4 days. Scale bar represents Agarwal et al. 2013). Copyright © 2018 Royal Society of
200 μm. (e) A fluorescence confocal microscopic image Chemistry. (4) In vivo bone marrow engineering. (a)
shows that the fibrin deposits (red) in (d) are found on the Workflow to generate a bone marrow-on-a-chip system
upper surface of the alveolar epithelium (green). Scale bar (b) PDMS device containing bone-inducing materials in
represents 50 μm. (f) The clots in (d) and (e) are highly its central cylindrical chamber before implantation (top).
fibrous networks. Scale bar represents 5 μm. Reprinted (c) Low- (left) and high-magnification views (right) of
with permission from (Huh et al. 2012). (Copyright © histological hematoxylin-eosin stained sections of the
2012, American Association for the Advancement of bone marrow formed in the PDMS device with two open-
Science). (2) Microfluidic liver sinusoid. (a) The liver ings (top) or one lower opening (center) at 8 weeks.
receives blood flow from the hepatic artery and the portal (Reprinted with permission from (Torisawa et al. 2014).
vein, which carries xenobiotics such as ingested drugs Copyright © 2014, Springer Nature)
from the small intestine to be metabolized. (b) Schematic
whereas the lower microvascular channel was assessment of new drugs (Uygun et al. 2010). To
filled with liquid. Stretch was cyclically applied develop such models, hepatocytes with mem-
to the membrane by additional vacuum actuator brane polarity and functional bile canaliculi are
channels to mimic the breathing movements of essential (Nakao et al. 2011). Lee et al. (2007)
the lung. After adding interleukin-2 (IL-2) to the created a biologically inspired artificial liver
microvascular channel, the fluid started to leak sinusoid with a microfluidic endothelial-like bar-
into the air compartment, mimicking pulmonary rier having mass transport properties similar to
edema. The authors tested their pulmonary dis- the liver acinus (Fig. 8.3). This unit consisted of a
ease model against the pharmacological agent cord of hepatocytes (50 μm × 30 μm × 500 μm)
GSK2193874. GSK2193874 blocks ion channels fed by diffusion of nutrients across the microflu-
activated by mechanical strain, and inhibited idic endothelial-like barrier from a convective
leakage of fluid to the air compartment, suggest- transport vessel. Combining photolithography-
ing it would be a viable option for patients with based microfluidic technologies and primary cell
pulmonary edema. The next step will be to hook culture, the team was able to assess liver toxicity
this lung up to other chip-based organs with the of diclofenac, since it has long been implicated in
goal of being able to rapidly screen many drugs drug-related liver damage (Lee et al. 2007).
across many organs (Huh et al. 2012). Following Lee’s work, Nakao et al. (Nakao
Jain et al. (2018) developed another lung et al. 2011) presented a microfluidic device com-
model on a chip with primary lung alveolar and prising an hepatic cord-like structure model with
endothelial cells. This new chip design enables primary rat hepatocytes, which was proven func-
whole human blood to be perfused through the tional: computer simulation results show that the
vascular channel without producing thrombus flow velocity in the medium flow channel in the
formation, while allowing high-resolution, real- model is 1 mm∕s, which is close to the respective
time analysis of interactions between human blood flow velocity in vivo. Moreover, when
blood cells and endothelial cells in an in vivo-like 5-(and-6)-carboxy-2′,7′-dichloro-fluorescein
context. Using this technology in combination diacetate (CDFDA) was injected into the device,
with novel analytical tools for quantitation of it was readily absorbed by the cells and metabo-
dynamic platelet-endothelial interactions and lized into CDF by esterase. The metabolites were
clot formation, the authors demonstrated a key actively excreted into bile canaliculi by the MRP2
role of the epithelium in inflammation-driven protein (Nakao et al. 2011).
vascular thrombosis during lipopolysaccharide
endotoxin (LPS)-induced acute lung injury. The
study showed that the model can be used to eval- 8.4.3 Kidney on a Chip
uate cytoprotective effects of Protease-activated
Receptor-1 (PAR-1), a potential therapeutic with Jang et al. (Jang and Suh 2010) developed a sim-
anti-thrombotic and anti-inflammatory activities, ple multi-layer microfluidic device by integrating
in vitro (2018). a polydimethylsiloxane (PDMS) microfluidic
channel (using photolithography technology) and
a porous membrane substrate (from polyester;
8.4.2 Liver on Chip 0.4 μm pore size;, 10 μm thickness) to culture
primary rat inner medullary collecting duct
The liver is one of the most relevant organs (IMCD) cells (Jang and Suh 2010). In order to
involved in maintaining physiological homeosta- evaluate molecular transport in renal tubule cells,
sis. Models that could recapitulate functional the team measured water and sodium uptake after
liver tissue will have an enormous impact in hormonal stimulations of the IMCD cells using
pharmaceutical fields, allowing for toxicity vasopressin and aldosterone. The cells were sub-
jected to a fluidic shear stress of 1 dyn/cm2 for the 8.4.6 Intestine on a Chip
time period of 5 h on the porous membrane,
which was found to be sufficient for the IMCD When discussing drug development, the small
cells to enhance cell polarization and rearrange intestine is crucial, given that orally administered
cytoskeleton and cell junctions (Jang and Suh drugs are mainly absorbed in the small intestine
2010). (Fasinu et al. 2011). There they diffuse across a
mucous layer covering an epithelial cell layer lin-
ing the intestinal wall. Therefore, it is logical that
8.4.4 Heart on a Chip drugs and chemicals are tested on intestinal cells
and tissues to assess pharmacological parameters
Agarwal et al. (Agarwal et al. 2013) developed such as absorption, distribution, metabolism,
anisotropic cardiac microtissues recapitulating elimination and toxicity, before any other.
the laminar architecture of the heart ventricle, Having this in mind, several groups started
which were engineered on cantilevers. The sub- developing models destined for absorption and
millimeter-sized thin film cantilevers of soft elas- metabolism studies. Realistic models should
tomers were fabricated by a laser-based process include certain key components such as cellular
(Fig. 8.3) (Agarwal et al. 2013). They used the (i.e. enterocytes, goblet and vascular endothelial
microdevice seeded with cardiac myocytes from cells), structural (i.e. villi and mucus) and
mice to test the positive inotropic effect of iso- dynamic (i.e. peristalsis) features. Some of these
proterenol on cardiac contractility at dosages issues were addressed by Kimura et al. (2008).
ranging from 1 nM to 100 μM. Tissues were The authors engineered an intestinal model with
exposed to different drug concentrations with the a membrane and vascular flow simulating the
advantage of complete washout between dos- epithelial barrier and the epithelial-endothelial
ages, leading to a fast collection of data from just interface. The microfluidic model comprised
one chip (Agarwal et al. 2013). optical fiber inserts for on-line and real-time
pharmacokinetic measurements to detect the
amount of fluorescence of rhodamine-123 in cul-
8.4.5 B one Marrow on a Chip ture, working as a tracer substrate to study diffu-
sion (Kimura et al. 2008).
Torisawa et al. (2014) developed a tissue-
engineered bone marrow model that is able to
retain hematopoietic stem and progenitor cells in 8.4.7 Tumor on a Chip
in vivo-like conditions for 1 week. In this exam-
ple, the bone marrow is first generated in mice, Cancer remains one of the main causes of death,
and then explanted as a whole and maintained in despite enormous efforts to cure the disease
vitro within a microfluidic device fabricated (Stroock and Fischbach 2010; Buchanan and
using standard photolithography (Fig. 8.3). Rylander 2013). It is no longer arguable that we
Results show this model mimics complex tissue- need to have a better understanding of the tumour
level responses to Ƴ-radiation toxicity normally microenvironment, as well as a more effective
observed only in vivo, as well as to a therapeutic means of screening anti-cancer drug leads
countermeasure agent (G-CSF) that is known to (Carvalho et al. 2017). The emerging tumour-on-
accelerate recovery from radiation-induced tox- a-chip technology can be of extreme usefulness
icity in cancer patients (Torisawa et al. 2014). in these processes.
Moreover, this research sheds light on the fact In an approach to better understand breast
that it is crucial to keep the whole hematopoietic cancer metastasis to bone, Jeon et al. (2015)
niche to obtain an in vivo like response, rather developed a microfluidic model of a human vas-
than culturing particular cell types (Torisawa cularized organ-specific microenvironment,
et al. 2014). which was used to explore and tune the extrava-
sation process of metastatic tumor cells (Jeon a pharmacokinetic-based manner (Sung et al.
et al. 2015). The experiment comprises an organ- 2010). Three cell lines representing the liver, a
specific model that enables the study of human tumor and bone marrow were cultured in the
metastatic breast cancer cell extravasation within three-chamber microCCA to test the toxicity of
a perfusable human microvascularized bone- an anticancer drug, 5-fluorouracil (5-FU). The
mimicking microenvironment. The microfluidic result was analyzed with the PK-PD model of the
model was characterized by anastomoses with device, and compared with the result of the static
the lateral media channels. MDA-MB-231 breast conditions (cell culture flasks). Each cell type
cancer cells were introduced and extravasation exhibited differential responses to 5-FU, and the
events monitored. Once established, the platform responses in the microfluidic environment were
was used to understand the anti-metastatic role of different from those in the static environment.
adenosine in the human breast cancer metastasis The authors showed a relatively new approach
to bone (Jeon et al. 2015). based on the combination of mathematical mod-
In its turn, Bischel et al. (2014) integrated a eling and an in vitro experimentation, with
microfluidic co-culture platform with a multi- improved predictability for testing drug toxicity.
photon imaging-based technique to determine More recently, Wong et al. (2017) exploited a
phenotypic cell behavior and flavin adenine dinu- 2.4 × 2.4 cm PDMS-based microfluidic chip to
cleotide (FAD) fluorescence intensity and fluo- generate droplets and perform drug screens
rescence lifetime simultaneously in the same cell. against suspended and adherent cancer cell lines,
This platform combines two independent assays as well as cells dissociated from primary tumors
normally performed with two different cell popu- of human patients. Bortezomib and Vorinostat
lations into a single device, allowing researchers were chosen as target drugs for leukemia in con-
to simultaneously assess both phenotypic cell junction with Jurkat cells, whereas Cisplatin and
behavior and enzyme activity (Bischel et al. Epirubicin were chosen as target drugs for breast
2014). The metastasis towards bone was mim- cancer in conjunction with MDA-MB-231 cells.
icked by differentiating MC3T3-E1 cells (osteo- Cell viability was measured by ethidium homodi-
blast precursor cell line) towards osteoblast mer-1 staining, enabling the capture of single cell
lineage and seeded on the sidewall of a micro- drug response, without compromising population
channel. After a collagen I hydrogel coating was analysis. This method provides a versatile and
applied as extracellular matrix, either Lymph rapid drug screening method using cells in differ-
Node Carcinome of the Prostate cells (LNCaP) ent states (Wong et al. 2017).
or (Bone metastatic LNCaP-derivative C4-2B)
C4-2B cells were seeded into the channel. By
engineering this cancer-bone microenvironment, 8.5 Conclusions
it was demonstrated that the cross-talk between a
bone metastatic prostate cancer cell line, C4-2B, New micro-engineering-based strategies have
and bone stromal cells, MC3T3-E1, increases the significantly improved the drug development
invasive behavior of the C4-2Bs, possibly by an process. Microfluidics makes it possible to pre-
increase in reactive oxygen species (ROS)- dict therapeutic activity by enabling high-
producing N1-acetyl polyamine oxidase (APAO) throughput functional readouts; to conduct
activity (Bischel et al. 2014). high-throughput screens for drug discovery in
Sung et al. (2010) presented a microscale cell realistic engineered 3D culture microenviron-
culture analog (mCCA), which is basically a ments; to develop organ-on-a-chip platforms to
microfluidic device based on pharmacokinetics- screen those drugs, which would probably fail
pharmacodynamics (PK-PD) mathematical mod- during the expensive animal testing and human
els. In this model, multiple cell culture chambers clinical trials later in the drug discovery process
are connected with fluidic channels to mimic pipeline. Critical challenges remain in making
multi-organ interactions and test drug toxicity in these technological advances practically feasible
for real-world pharmaceutical discovery. Thus, References

the optimal solution for drug discovery may rely
on sufficiently complex and realistic technology, Agarwal A, Goss JA, Cho A, McCain ML, Parker KK
(2013) Microfluidic heart on a chip for higher through-
while being simple enough to allow for assay put pharmacological studies. Lab Chip 13(18):3599–
readouts, reliability and operational simplicity. In 3608. https://doi.org/10.1039/c3lc50350j
addition, mathematical and computational Anguiano M, Castilla C, Maska M, Ederra C, Pelaez R,
modeling will guide the design of the next gen- Morales X, Munoz-Arrieta G, Mujika M, Kozubek M,
Munoz-Barrutia A, Rouzaut A, Arana S, Garcia-Aznar
eration of organ-on-chip devices, and nanosens- JM, Ortiz-de-Solorzano C (2017) Characterization
ing will enable the real-time and in situ monitoring of three-dimensional cancer cell migration in mixed
of the cellular metabolic responses. collagen-Matrigel scaffolds using microfluidics and
Regarding biomaterials, in vitro platforms are image analysis. PLoS One 12(2):e0171417. https://
doi.org/10.1371/journal.pone.0171417
just a scientific attempt at mimicking the natural Baker BM, Trappmann B, Stapleton SC, Toro E, Chen CS
system. It is therefore important to make the right (2013) Microfluidics embedded within extracellular
choice of the function to be mimicked, as this matrix to define vascular architectures and pattern dif-
choice will eventually determine whether a sys- fusive gradients. Lab Chip 13(16):3246–3252. https://
doi.org/10.1039/c3lc50493j
tem is considered valuable or not. Because of Barata D, van Blitterswijk C, Habibovic P (2016) High-
this, there is now a huge demand for new and tun- throughput screening approaches and combinatorial
able biomaterials to interface with microfluidic development of biomaterials using microfluidics.
chips for cell biological studies. Acta Biomater 34:1–20. https://doi.org/10.1016/j.
actbio.2015.09.009
However, the reality is that despite rapid Barata D, Spennati G, Correia C, Ribeiro N, Harink B,
development of expertise in designing, fabricat- van Blitterswijk C, Habibovic P, van Rijt S (2017)
ing and operating microfluidic systems for nearly Development of a shear stress-free microfluidic gra-
20 years now, no commercially available drugs dient generator capable of quantitatively analyz-
ing single-cell morphology. Biomed Microdevices
have been discovered as a unique result of micro- 19(4):81. https://doi.org/10.1007/s10544-017-0222-z
fluidic technologies, to the best of our Bettinger CJ, Borenstein JT (2010) Biomaterials-based
knowledge. microfluidics for engineered tissue constructs. Soft
Griffith, a world-class scientist in organ-on-a- Matter 6(20):4999–5015. https://doi.org/10.1039/
C0SM00247J
chip research said that more questions would be Bichara DA, Bodugoz-Sentruk H, Ling D, Malchau E,
answered if Pharma collaborated more with engi- Bragdon CR, Muratoglu OK (2014) Osteochondral
neers, as Pharma knows what problems it needs defect repair using a polyvinyl alcohol- polyacrylic
to solve, but engineers know how to build chips. acid (PVA-PAAc) hydrogel. Biomed Mater
(Bristol, England) 9(4):045012. https://doi.
Therefore, success in this endeavor must start org/10.1088/1748-6041/9/4/045012
with alliances between the fields of pharmaceuti- Bischel LL, Casavant BP, Young PA, Eliceiri KW, Basu
cal industry, engineering, medicine, biology and HS, Beebe DJ (2014) A microfluidic coculture and
computational biology. Maybe that is our future multiphoton FAD analysis assay provides insight into
the influence of the bone microenvironment on pros-
answer. tate cancer cells. Integr Biol 6(6):627–635. https://doi.
org/10.1039/C3IB40240A
Acknowledgments This chapter is a result of the project Bita C, Huang CP, Ghajar CM, Kachgal S, Kniazeva E,
FROnTHERA (NORTE-01-0145-FEDER-000023), sup- Jeon NL, Putnam AJ (2010) Recreating the perivas-
ported by Norte Portugal Regional Operational cular niche ex vivo using a microfluidic approach.
Programme (NORTE 2020), under the PORTUGAL 2020 Biotechnol Bioeng 107(6):1020–1028. https://doi.
Partnership Agreement, through the European Regional org/10.1002/bit.22891
Development Fund (ERDF). Mariana Carvalho acknowl- Buchanan C, Rylander MN (2013) Microfluidic culture
edges her PhD scholarship NORTE-08-5369-FSE-000044, models to study the hydrodynamics of tumor progres-
funded by Programa Operacional Regional do Norte, sion and therapeutic response. Biotechnol Bioeng
Fundo Social Europeu, Norte2020 TERM&SC and 110(8):2063–2072. https://doi.org/10.1002/bit.24944
EMBO Short-Term Fellowship 7232 J. M. Oliveira thanks Carvalho MR, Lima D, Reis RL, Correlo VM, Oliveira
FCT for his distinction attributed under the FCT JM (2015) Significance of tissue-engineered 3D in-
Investigator program (IF/01285/2015). vitro models for cancer research and drug screening.
Trends Biotechnol
Carvalho MR, Lima D, Reis RL, Oliveira JM, Correlo VM Chim Acta 665(1):7–14. https://doi.org/10.1016/j.
(2017) Anti-Cancer drug validation: the contribution aca.2010.03.015
of tissue engineered models. Stem Cell Rev 13:347– Gramowski A, Jugelt K, Stuwe S, Schulze R,
363. https://doi.org/10.1007/s12015-017-9720-x McGregor GP, Wartenberg-Demand A, Loock J,
Chandrasekaran A, Abduljawad M, Moraes C (2016a) Schroder O, Weiss DG (2006) Functional screen-
Have microfluidics delivered for drug discovery? ing of traditional antidepressants with primary cor-
Expert Opin Drug Discov 11(8):745–748. https://doi. tical neuronal networks grown on multielectrode
org/10.1080/17460441.2016.1193485 neurochips. Eur J Neurosci 24(2):455–465. https://doi.
Chandrasekaran A, Abduljawad M, Moraes C (2016b) org/10.1111/j.1460-9568.2006.04892.x
Have microfluidics delivered for drug discovery? Hay M, Thomas DW, Craighead JL, Economides C,
Expert Opin Drug Discovery 11(8):745–748. https:// Rosenthal J (2014) Clinical development success
doi.org/10.1080/17460441.2016.1193485 rates for investigational drugs. Nat Biotechnol 32:40.
Chaw KC, Manimaran M, Tay FE, Swaminathan S https://doi.org/10.1038/nbt.2786. https://www.nature.
(2007) Matrigel coated polydimethylsiloxane based com/articles/nbt.2786#supplementary-information
microfluidic devices for studying metastatic and non- Hughes JP, Rees S, Kalindjian SB, Philpott KL
metastatic cancer cell invasion and migration. Biomed (2011) Principles of early drug discovery. Br
Microdevices 9(4):597–602. https://doi.org/10.1007/ J Pharmacol 162(6):1239–1249. https://doi.
s10544-007-9071-5 org/10.1111/j.1476-5381.2010.01127.x
Choi K-H, Choi BH, Park SR, Kim BJ, Min B-H (2010) Huh D, Leslie DC, Matthews BD, Fraser JP, Jurek S,
The chondrogenic differentiation of mesenchy- Hamilton GA, Thorneloe KS, McAlexander MA,
mal stem cells on an extracellular matrix scaffold Ingber DE (2012) A human disease model of drug
derived from porcine chondrocytes. Biomaterials toxicity-
induced pulmonary edema in a lung-on-a-
31(20):5355–5365. https://doi.org/10.1016/j. chip microdevice. Sci Transl Med 4(159):159ra147.
biomaterials.2010.03.053 https://doi.org/10.1126/scitranslmed.3004249
Cui X, Breitenkamp K, Finn MG, Lotz M, D'Lima DD Hussey GS, Dziki JL, Badylak SF (2018) Extracellular
(2012) Direct human cartilage repair using three- matrix-based materials for regenerative medicine.
dimensional bioprinting technology. Tissue Eng Part Nat Rev Mat 3:159–173. https://doi.org/10.1038/
A 18(11–12):1304–1312. https://doi.org/10.1089/ten. s41578-018-0023-x
TEA.2011.0543 Inagaki Y, Kitamura N, Kurokawa T, Tanaka Y,
de Girolamo L, Niada S, Arrigoni E, Di Giancamillo A, Gong JP, Yasuda K, Tohyama H (2014) Effects
Domeneghini C, Dadsetan M, Yaszemski MJ, Gastaldi of culture on PAMPS/PDMAAm double-network
D, Vena P, Taffetani M, Zerbi A, Sansone V, Peretti gel on chondrogenic differentiation of mouse
GM, Brini AT (2015) Repair of osteochondral defects C3H10T1/2 cells: in vitro experimental study.
in the minipig model by OPF hydrogel loaded with BMC Musculoskelet Disord 15:320. https://doi.org/
adipose-derived mesenchymal stem cells. Regen Med 10.1186/1471-2474-15-320
10(2):135–151. https://doi.org/10.2217/rme.14.77 Jain A, Barrile R, van der Meer AD, Mammoto A,
Dinh ND, Chiang YY, Hardelauf H, Baumann J, Jackson Mammoto T, De Ceunynck K, Aisiku O, Otieno MA,
E, Waide S, Sisnaiske J, Frimat JP, van Thriel C, Louden CS, Hamilton GA, Flaumenhaft R, Ingber
Janasek D, Peyrin JM, West J (2013) Microfluidic DE (2018) Primary human lung Alveolus-on-a-chip
construction of minimalistic neuronal co-cultures. model of intravascular thrombosis for assessment of
Lab Chip 13(7):1402–1412. https://doi.org/10.1039/ therapeutics. Clin Pharmacol Ther 103(2):332–340.
c3lc41224e https://doi.org/10.1002/cpt.742
Dittrich PS, Manz A (2006) Lab-on-a-chip: microfluidics Jang KJ, Suh KY (2010) A multi-layer microfluidic device
in drug discovery. Nat Rev Drug Discov 5:210–218. for efficient culture and analysis of renal tubular
https://doi.org/10.1038/nrd1985 cells. Lab Chip 10(1):36–42. https://doi.org/10.1039/
Du G, Fang Q, den Toonder JMJ (2016) Microfluidics b907515a
for cell-based high throughput screening platforms— Jeon JS, Bersini S, Gilardi M, Dubini G, Charest JL,
A review. Anal Chim Acta 903:36–50. https://doi. Moretti M, Kamm RD (2015) Human 3D vascularized
org/10.1016/j.aca.2015.11.023 organotypic microfluidic assays to study breast cancer
Fasinu P, Pillay V, Ndesendo VM, du Toit LC, Choonara cell extravasation. Proc Natl Acad Sci 112(1):214–
YE (2011) Diverse approaches for the enhancement 219. https://doi.org/10.1073/pnas.1417115112
of oral drug bioavailability. Biopharm Drug Dispos Jusoh N, Oh S, Kim S, Kim J, Jeon NL (2015)
32(4):185–209. https://doi.org/10.1002/bdd.750 Microfluidic vascularized bone tissue model with
FluidicMEMS (2018) List of microfluidics companies. hydroxyapatite-incorporated extracellular matrix.
http://fluidicmems.com/list-of-microfluidics-lab-on-a- Lab Chip 15(20):3984–3988. https://doi.org/10.1039/
chip-and-biomems-companies/ c5lc00698h
Gao D, Liu J, Wei HB, Li HF, Guo GS, Lin JM (2010) Kang L, Chung BG, Langer R, Khademhosseini A (2008)
A microfluidic approach for anticancer drug analy- Microfluidics for drug discovery and development:
sis based on hydrogel encapsulated tumor cells. Anal from target selection to product lifecycle manage-
ment. Drug Discov Today 13(1–2):1–13. https://doi. cells, towards engineering developmental pro-
org/10.1016/j.drudis.2007.10.003 cesses. Sci Rep 5:10288. https://doi.org/10.1038/
Khademhosseini A, Yeh J, Eng G, Karp J, Kaji H, srep10288. https://www.nature.com/articles/
Borenstein J, Farokhzad OC, Langer R (2005) Cell srep10288#supplementary-information
docking inside microwells within reversibly sealed Paul SM, Mytelka DS, Dunwiddie CT, Persinger CC,
microfluidic channels for fabricating multiphenotype Munos BH, Lindborg SR, Schacht AL (2010) How to
cell arrays. Lab Chip 5(12):1380–1386. https://doi. improve R&D productivity: the pharmaceutical indus-
org/10.1039/b508096g try's grand challenge. Nat Rev Drug Discov 9(3):203–
Khanna I (2012) Drug discovery in pharmaceutical indus- 214. https://doi.org/10.1038/nrd3078
try: productivity challenges and trends. Drug Discov Radhakrishnan J, Subramanian A, Krishnan UM,
Today 17(19):1088–1102. https://doi.org/10.1016/j. Sethuraman S (2017) Injectable and 3D bioprinted
drudis.2012.05.007 polysaccharide hydrogels: from cartilage to osteochon-
Kimura H, Yamamoto T, Sakai H, Sakai Y, Fujii T (2008) dral tissue engineering. Biomacromolecules 18(1):1–
An integrated microfluidic system for long-term per- 26. https://doi.org/10.1021/acs.biomac.6b01619
fusion culture and on-line monitoring of intestinal Shantanu P, Iman H, CJ M, LE A (2016) Polymeric bioma-
tissue models. Lab Chip 8(5):741–746. https://doi. terials for in vitro Cancer tissue engineering and drug
org/10.1039/b717091b testing applications. Tissue Eng Part B Rev 22(6):470–
Kobel S, Lutolf MP (2011) Biomaterials meet microfluid- 484. https://doi.org/10.1089/ten.teb.2015.0567
ics: building the next generation of artificial niches. Shen B (2015) A new Golden age of natural products
Curr Opin Biotechnol 22(5):690–697. https://doi. drug discovery. Cell 163(6):1297–1300. https://doi.
org/10.1016/j.copbio.2011.07.001 org/10.1016/j.cell.2015.11.031
Lai JH, Kajiyama G, Smith RL, Maloney W, Yang F (2013) Si G, Yang W, Bi S, Luo C, Ouyang Q (2012) A parallel
Stem cells catalyze cartilage formation by neonatal diffusion-based microfluidic device for bacterial che-
articular chondrocytes in 3D biomimetic hydrogels. motaxis analysis. Lab Chip 12(7):1389–1394. https://
Sci Rep 3:3553. https://doi.org/10.1038/srep03553. doi.org/10.1039/C2LC21219F
http://dharmasastra.live.cf.private.springer.com/ Stroock AD, Fischbach C (2010) Microfluidic cul-
articles/srep03553#supplementary-information ture models of tumor angiogenesis. Tissue Eng Part
Lee PJ, Hung PJ, Lee LP (2007) An artificial liver sinusoid A 16(7):2143–2146. https://doi.org/10.1089/ten.
with a microfluidic endothelial-like barrier for primary TEA.2009.0689
hepatocyte culture. Biotechnol Bioeng 97(5):1340– Sung JH, Kam C, Shuler ML (2010) A microfluidic device
1346. https://doi.org/10.1002/bit.21360 for a pharmacokinetic-pharmacodynamic (PK-PD)
Lu H, Hoshiba T, Kawazoe N, Chen G (2011) Autologous model on a chip. Lab Chip 10(4):446–455. https://doi.
extracellular matrix scaffolds for tissue engineer- org/10.1039/b917763a
ing. Biomaterials 32(10):2489–2499. https://doi. Szymański P, Markowicz M, Mikiciuk-Olasik E (2012)
org/10.1016/j.biomaterials.2010.12.016 Adaptation of high-throughput screening in drug
Lutolf MP, Hubbell JA (2005) Synthetic biomaterials as discovery—toxicological screening tests. Int J
instructive extracellular microenvironments for mor- Mol Sci 13(1):427–452. https://doi.org/10.3390/
phogenesis in tissue engineering. Nat Biotechnol ijms13010427
23(1):47–55. https://doi.org/10.1038/nbt1055 Toh YC, Lim TC, Tai D, Xiao G, van Noort D, Yu H
Maerkl SJ, Quake SR (2007) A systems approach to (2009) A microfluidic 3D hepatocyte chip for drug
measuring the binding energy landscapes of transcrip- toxicity testing. Lab Chip 9(14):2026–2035. https://
tion factors. Science 315(5809):233–237. https://doi. doi.org/10.1039/b900912d
org/10.1126/science.1131007 Torisawa Y-S, Spina CS, Mammoto T, Mammoto A,
Mak IW, Evaniew N, Ghert M (2014) Lost in translation: Weaver JC, Tat T, Collins JJ, Ingber DE (2014) Bone
animal models and clinical trials in cancer treatment. marrow–on–a–chip replicates hematopoietic niche
Am J Transl Res 6(2):114–118 physiology in vitro. Nat Methods 11:663. https://doi.
Mitchell MC, Spikmans V, Mello AJD (2001) Microchip- org/10.1038/nmeth.2938. https://www.nature.com/
based synthesis and analysis: control of multicom- articles/nmeth.2938#supplementary-information
ponent reaction products and intermediates. Analyst Tsao C-W (2016) Polymer microfluidics: simple, low-cost
126(1):24–27. https://doi.org/10.1039/B007397K fabrication process bridging academic lab research to
Mullard A (2011) 2010 FDA drug approvals. Nat Rev commercialized production. Micromachines 7(12).
Drug Discov 10(2):82–85. https://doi.org/10.1038/ https://doi.org/10.3390/mi7120225
nrd3370 Uygun BE, Soto-Gutierrez A, Yagi H, Izamis ML,
Nakao Y, Kimura H, Sakai Y, Fujii T (2011) Bile cana- Guzzardi MA, Shulman C, Milwid J, Kobayashi N,
liculi formation by aligning rat primary hepatocytes Tilles A, Berthiaume F, Hertl M, Nahmias Y, Yarmush
in a microfluidic device. Biomicrofluidics 5(2):22212. ML, Uygun K (2010) Organ reengineering through
https://doi.org/10.1063/1.3580753 development of a transplantable recellularized liver
Occhetta P, Centola M, Tonnarelli B, Redaelli A, Martin graft using decellularized liver matrix. Nat Med
I, Rasponi M (2015) High-throughput microflu- 16(7):814–820. https://doi.org/10.1038/nm.2170
idic platform for 3D cultures of mesenchymal stem
Wan J (2012) Microfluidic-based synthesis of hydrogel sizes on chondrocyte culture for cartilage tissue engi-
particles for cell microencapsulation and cell-based neering. Mater Sci Eng C 34:168–175. https://doi.
drug delivery. Polymers (Basel) 4(2):1084–1108. org/10.1016/j.msec.2013.09.003
https://doi.org/10.3390/polym4021084 Zhang Y, Yu Y, Ozbolat IT (2013) Direct bioprint-
Whitesides GM (2006) The origins and the future of ing of vessel-like tubular microfluidic channels.
microfluidics. Nature 442:368–373. https://doi. J Nanotechnol Eng Med 4(2):0210011–0210017.
org/10.1038/nature05058 https://doi.org/10.1115/1.4024398
Wong AH-H, Li H, Jia Y, Mak P-I, Martins RPS, Liu Y, Zheng XT, Yu L, Li P, Dong H, Wang Y, Liu Y, Li CM
Vong CM, Wong HC, Wong PK, Wang H, Sun H, (2013) On-chip investigation of cell-drug interactions.
Deng C-X (2017) Drug screening of cancer cell lines Adv Drug Deliv Rev 65(11–12):1556–1574. https://
and human primary tumors using droplet microflu- doi.org/10.1016/j.addr.2013.02.001
idics. Sci Rep 7(1):9109. https://doi.org/10.1038/ Zhou X, Cai S, Hong A, You Q, Yu P, Sheng N, Srivannavit
s41598-017-08831-z O, Muranjan S, Rouillard JM, Xia Y, Zhang X, Xiang Q,
Wu H, Wheeler A, Zare RN (2004) Chemical cytometry Ganesh R, Zhu Q, Matejko A, Gulari E, Gao X (2004)
on a picoliter-scale integrated microfluidic chip. Proc Microfluidic PicoArray synthesis of oligodeoxynucle-
Natl Acad Sci U S A 101(35):12809–12813. https:// otides and simultaneous assembling of multiple DNA
doi.org/10.1073/pnas.0405299101 sequences. Nucleic Acids Res 32(18):5409–5417.
Zeng L, Yao Y, Wang D-A, Chen X (2014) Effect of https://doi.org/10.1093/nar/gkh879
microcavitary alginate hydrogel with different pore
Dynamic Culture Systems and 3D
Interfaces Models for Cancer 9
Drugs Testing
Diogo C. Fernandes, Raphaël F. Canadas,
Rui L. Reis, and Joaquim M. Oliveira
Abstract used for the development of pharmaceutical

The mass use of biological agents for pharma- agents from this point on. In the last decades,
ceutical purposes started with the develop- the use of bioreactors and microfluidic sys-
ment and distribution of vaccines, followed by tems has been expanded in different fields.
the industrial production of antibiotics. The The emergence of the tissue engineering led to
use of dynamic systems, such as bioreactors, the development of in vitro models cultured in
had been already applied in the food industry dynamic systems. This is particularly relevant
in fermentation processes and started being considering the urgent reduction of the total
dependence on animal disease models that is
undermining the development of novel drugs,
using alternatively human-based models to
make the drug discovery process more reli-
able. The failure out coming from animal
D. C. Fernandes · R. F. Canadas (*) models has been more prevalent in certain
3B’s Research Group, I3Bs – Research Institute on types of cancer, such as glioblastoma multi-
Biomaterials, Biodegradables and Biomimetics, form and in high-grade metastatic cancers like
Institute of Excellence on Tissue Engineering and bone metastasis of breast or prostatic cancer.
Regenerative Medicine, AvePark, Parque de Ciência e The difficulty in obtaining novel drugs for
Tecnologia, Barco, Guimarães, Portugal these purposes is mostly linked to the barriers
ICVS/3B’s–PT Government Associate Laboratory, around the tumors, which these bioactive mol-
Braga, Guimarães, Portugal ecules have to overcome to become effective.
e-mail: raphael.canadas@i3bs.uminho.pt
For that reason, the individualized study of
R. L. Reis · J. M. Oliveira each interface is paramount and is only realis-
tic once applying human-based samples (e.g.
University of Minho, Headquarters of the European cells or tissues) in three-dimensions for in
Institute of Excellence on Tissue Engineering and vitro modeling under dynamic conditions. In
Regenerative Medicine, Guimarães, Portugal this chapter, the most recent approaches to
ICVS/3B’s PT Government Associate Lab, model these interfaces in 3D systems will be
Braga, Guimarães, Portugal explored, highlighting their major contribu-
The Discoveries Centre for Regenerative and tions to the field. In this section, these sys-
Precision Medicine, Headquarters at University tems’ impact on increased knowledge in

https://doi.org/10.1007/978-3-030-36588-2_9
138 D. C. Fernandes et al.
relevant aspects of cancer aggressiveness as The lack of innovative drugs is mostly related
invasive or motile cellular capacity, or even to the failure during the preclinical stage due to:
resistance to chemotherapeutic agents will poor mimicry of human tissue specific barriers
have particular focus. The last section of this such as the blood-brain barrier (BBB) (Lim et al.
chapter will focus on the integration of the 2018); and the translation of false-positive results
tumor interfaces in dynamic systems, particu- from pre-clinical to clinical stage (Mak et al.
larly its application on high-throughput drug 2014). This last one is particularly bound to the
screening. The industrial translation of such fact that 92% of the molecules selected from ani-
platforms will be discussed, as well as the mal models fail in efficacy when applied to
main upcoming challenges and future humans (Mak et al. 2014).
perspectives. Although the most promising alternative to
the use of animal models is the use of in vitro
Keywords models (Fig. 9.1) achieving clinically relevant
3D interfaces models · In vitro tumor models complexity together with reproducibility is the
· Drug screening · Bioreactors major challenge of this approach. Despite tumor
models have increasingly mimicked human phys-
iology over the last decade (Katt et al. 2016), the
recreation of general and tissue-specific hall-
Highlights marks and surrounding barriers is still an unre-
solved quest in the field.
• The impairment in novel drugs development is There are several sources of variability that
mainly due to the failure of results in animal influence tumor progression and, eventually, the
models to translate to humans, particularly in patient’s outcome. The surroundings of a tumor
certain cancer types, and the best alternative is site turn into a series of interfaces to which can-
the use of in vitro models; cer cells contact and interact, undergoing com-
• The three-dimensional (3D) interfaces con- plex feedback loops that have proven relevant in
tacting cancer cells work as barriers to thera- chemo resistance (Castells et al. 2012; Rice et al.
pies effectiveness, becoming essential features 2017) and tumor recurrence (Walens et al. 2019)
in in vitro models; (Fig. 9.2). The most recent strategies to mimic
• 3D tumor interfaces can be cellular, molecular the several tumor interfaces will be reviewed and
or physical and have been modeled in recent analyzed herein. In addition, the use of physio-
years, existing already some examples of logically relevant dynamic conditions involving
industrial application for drug testing bioreactors and microfluidic systems will be dis-
purposes. cussed and an author’s analysis on the future out-
look of the field will be presented.
9.1 Introduction
9.2 D Interface Models Critical
3
During the past few decades, a decrease in the Features – Recreation
number of new pharmaceutical drugs has become of the Tumor’s Interfaces
evident, with the newly found drugs becoming
also increasingly costly (DiMasi et al. 2016). 9.2.1 Structural
Although recent therapies, such as immunother-
apy, have looked promising for different tumors, Interactions between cells and the surrounding
they have failed in treating certain primary cancer ECM are paramount in cancer development and
types, such as glioblastoma multiform (Lim et al. progression. The molecular assembly of ECM
2018) or castration-resistant prostate metastatic can promote aberrant behavior, even in non-
cancer (Twardowski et al. 2019). cancerous cells, revealing the importance of the
9 Dynamic Culture Systems and 3D Interfaces Models for Cancer Drugs Testing 139
Fig. 9.1 Advantages of in vitro models as compared to factors, biomimetic extracellular matrix (ECM) and the
animal models. Controlled flow rate at physiological shear use of human cells are the main reasons why in vitro mod-
stress, administration of defined concentrations of soluble els may outperform animal models as disease models for
drug testing
ECM remodeling in the development of malig- Strikingly, benign tumor cells and regular epithe-
nant phenotype (Paszek et al. 2005). In the fol- lial cells, when contacting the membrane, had
lowing sub-sections, the latest biomimetic controlled proliferation and the integrity of the
strategies to generate and study the feedback membrane was kept. However, when contacting
loops between cancer cells and structural cues of malignant cancer cells, the basement membrane
the surrounding tissue will be addressed. got disrupted and the invasiveness of these cells
got exacerbated (Guzman et al. 2017).
9.2.1.1 Tumor-Basement Membrane Nonetheless, the interplay between a tumor
The basement membrane is an adjacent structure and the surrounding basement membrane is
to the epithelium, endothelium and mesothelium dependent on the cancer grade and sensitive to
that supply architectural support to the tissue. other cell types surronding the tumor site. This
These membranes are steep and discrete inter- was observed in patient-derived organoids of
faces that tumors have to overcome during tissue pancreatic cancer surrounded by basement mem-
invasion. The interface between the connective brane invasive behavior and consequent mem-
tissue and the cancer and cancer-associated cells brane disruption induced by the presence of
is associated with tumor progression and malig- pancreatic stellate cells. Pancreatic adenocarci-
nancy (Engbring and Kleinman 2003). noma cells cultured in spheroids encapsulated in
Due to its relevance, when designing a 3D Matrigel® developed their own basement mem-
tumor in vitro model, such interface should be brane maintaining homeostasis. Nonetheless,
structurally reproduced. Using this strategy, it once pancreatic stellate cells were present, the
was reported that this interface can promote, but cancer cells disrupted the basement membrane
not induce, malignant phenotype on cancer cells and invaded the surrounding matrix (Koikawa
in a model where spheroids were culture contact- et al. 2018). This interference on the interplay
ing an in vitro healthy basement membrane. between cancer cells and the basement mem-
Fig. 9.2 Tumor microenvironment’s interfaces. Every cell type, physical and biochemical features surrounding cancer
cells establish interplays that affect and condition the tumor’s outcome. (Figure adapted from Joyce 2005)
brane differs by other cell types uncovers a level in the epithelial-to-mesenchymal transition
of complexity that urges for increased under- (EMT), namely how the rearrangement of the
standing and following studies. ECM leads to the development of fibrotic tumor
These results highlighted the importance of microenvironment (Guzman et al. 2017; Sharma
the tumor-basement membrane interface in an in et al. 2017; Puls et al. 2017; Shi et al. 2019;
vitro model. In high-throughput systems, consid- Tanaka et al. 2019). These models, however, are
ering it is a discrete 2D homogeneous interface, quite diverse in complexity, uncovering different
the addition of this interface would increase bio- dynamics according to its purpose.
mimetic capacity to the existing models without One-dimensional (1D) and two-dimensional
presenting a major fabrication challenge. (2D) models, simplistic by design, deliver
straightforward quantifications that can be useful
9.2.1.2 Tumor-ECM Interface: Fibrotic to account for the contribution of each feature in
vs. Healthy a more complex model, particularly in 3D.
The structural remodeling of the ECM of the The 1D breast cancer fibrotic models have
tumor’s surrounding tissue is a hallmark in sev- been used for example to quantify the correlation
eral cases, such as pancreatic (Puls et al. 2017) or between fiber’s diameter and cell linear motility.
breast cancers (Vermeulen et al. 2017). Since the Measuring the average linear velocity of cultured
most relevant component of the ECM of these tis- cancer cells onto fibers with diameters ranging
sues is collagen, its reassembly remodels the from 20 to 0.7 μm, it was found that fibers with
microenvironment and cell behavior. Thus, most smaller diameter, closer to the ones observed in
models are based on the fibrotic assemble of this clinical assessments, led to increased migration
protein. velocity. Moreover, motility-related nuclear
Recently, models mimicking the changes in deformation, usually associated with the squeez-
tissue fibrosis have been used to uncover its role ing process in 3D motion, was found to be related
to the forces balance necessary for linear cellular 9.2.2 Mechanical Properties
motion (Sharma et al. 2017).
Thus, for the observation of specific phenom- 9.2.2.1 Tumor-Solid Interface
ena, simpler models enable a clear classification Over the last years it was established that cancer
and quantification of particular aspects of the dis- cells undergo EMT upon variations of the matrix
ease, therefore, aiding the assembly of better and stiffness (Liou et al. 2014; Ko et al. 2016; Bauer
more representative 3D models. However, it was et al. 2018) and for more than a decade it is
already demonstrated that only 3D cancer models known to drive carcinogenesis (Paszek et al.
can realistically mimic the dynamics of the native 2005). Nonetheless, only by using novel 3D
tumor tissue, which is critical either for drugs models, the role of stiffness in chemoresistance
testing or for the optimization of medical devices was uncovered (Rice et al. 2017).
(Ridky et al. 2010). Polyacrylamide gels with tunable rigidity, but
To do so, in vitro disease models using clini- no bioactive role in cell adhesion, induced an
cally observed features enable information that increase in mesenchymal markers on pancreatic
could not be achieved in vivo without advanced cancer by three-fold when compressive stiffness
microscopy strategies. For instance, such models increased from 1 kPa to 25 kPa. Furthermore, it
allowed the clarification that pancreatic stellate was concluded that the increasing stiffness led to
cells, not cancer-associated fibroblasts, are the an increased resistance to paclitaxel, a chemo-
cells responsible for the fibrotic reassembly of therapeutic agent that stabilizes the microtubules
collagen and fibronectin that drives the malignant cytoskeleton, but not to gemcitabine, a nucleotide
phenotype in cancer cells, using clinically analogue that competes with other nucleotides
observed tissue thickness (Tanaka et al. 2019). during replication of DNA (Rice et al. 2017).
However, none of the mentioned models Thus, it was shown that stiffness itself is a suffi-
accounted for the effect of stiffness increase cient factor to induce resistance to chemothera-
related to the increased fibrotic tissue fraction. peutic agents.
With that purpose, in a 3D pancreatic cancer Strikingly, the compressive modulus of the
model of invasion, three different pancreatic duc- interstitial matrix does not affect uniquely the
tal adenocarcinoma cell lines were encapsulated cancerous cells but also all the tumor microenvi-
in engineered fibrotic tissue of variable stiffness. ronment. Increased stiffness was found to arrest
Matrigel® was used to counter the increase of the differentiation of pre-adipocytes cultured in
stiffness caused by having a thicker fibrotic tissue between tumor spheroids (Yue et al. 2018).
fraction, maintaining the stiffness constant. It Therefore, stiffness-related malignancy affects
was then found that an increase in fibrotic tissue, not only cancer cells, modulating their cytoskel-
with a constant stiffness, led by itself to the over- eton, but also other cells in the tumor microenvi-
expression of mesenchymal markers (e.g., ronment and even healthy stromal cells, changing
vimentin), and a decrease of epithelial markers their behavior towards a disease one.
(Puls et al. 2017). Importantly, these results dem-
onstrate the distinct roles of stiffness and tissue 9.2.2.2 Tumor-Liquid Interface
structure in driving tumor malignancy. The circulation of fluid from the vascular system
Considering the presented models, the impor- through the surrounding tissue and collected in
tance of the ECM architecture in tissue homeo- the lymphatic one produces a shear stress effect
stasis is undeniable. However, it is paramount to that is inversely proportional to the distance from
consider the variation of biophysical properties the blood vessel. When the tumor promotes
with the architectural and chemical changes of angiogenesis or undergoes surrounding tissue
the tissue, which can have drastic effects over the invasion, approaching blood vessels, the shear
tumor’s fate, as it will be addressed in the follow- stress increases due to the proximity to the blood
ing sub-section. flow (Sefidgar and Soltani 2014).
The shear stress felt by cancer cells influences faces in conditions of high humidity to recreate
their physiology and is relevant for tumor out- the native conditions of the tissue. For that pur-
come (Mak 1986; Hillsley and Frangos 1994; pose, an ex vivo 3D model of colorectal cancer
Mow et al. 1999; Triantafillu et al. 2017). Once was developed by culturing primary tissue in
cancer cells undergo endothelial-to-mesenchymal inner wells, exposed to a humidified atmosphere,
transition (EMT) and start intravasation into sur- while being narrowly submerse by the medium
rounding blood vessels, the blood flow and the dispensed in the outer well. The liquid moves
fluid permeating the tissue exert increased influ- into the inner well by capillarity and keeping the
ence over cancer cells. To investigate how impor- cells nearly submerged, mimicking the circula-
tant the tumor-blood interface can be on the tissue tion of fluid observed in the native tissue. The
and blood vessels invasion by cancer cells, in model kept the physiology of the biopsied tissue
vitro models have recently proven that fluid shear for a minimum of 14 days, either healthy or can-
stress can promote malignant phenotype in cancerous, showing to be a robust system for ex vivo
cer cells. For instance, in ovarian (Hyler et al. tissue culture (Usui et al. 2018).
2018) and breast cancer (Bratton et al. 2019) Skin tissue, on the other hand, is the only tis-
models, low magnitude fluid shear stress drove sue that contacts directly with air that has not
genomic instability and mesenchymal pheno- been firstly humidified. Therefore, a non-
type, leading to metastasis. In the case of the vascularized melanoma model should include
ovarian cancer model, malignant and non- this unique feature of the skin. To mimic that
malignant cells were subjected to fluid shear tissue-air interface of skin cancer, melanoma
stress, being cultured while attached and in sus- spheroids were inserted in the dermis of the
pension, it was concluded that suspension culture model, consisting of fibroblasts and collagen
severely promoted malignant (Hyler et al. 2018). type I, below a layer of keratinocytes, mimicking
There are still numerous limitations to the cre- the epidermis. This structure was developed into
ations of these models. The heterogeneity of the an insert well slightly submerged in medium.
density of capillary vessels within tissues, or the Histology results, when compared to explants,
pressure variation with the distance from the showed that the model resembled the native tis-
heart makes the creation of models for personal- sue, either in the skin construct, either in the het-
ized medicine particularly. Nonetheless, these erogeneity of the tumor, having a proliferative
models shall account for the near in vivo com- front and necrotic core, which is characteristic of
plexity with the application of advanced fabrica- non-vascularized melanomas (Houghton 1987).
tion techniques, like bioprinting, that will allow The generation of this model was performed in
the creation of hierarchical 3D model. multiple simultaneous batches, allowing the pro-
duction of a high number of air-tissue organo-
9.2.2.3 Tumor-Air Interface tropic replicas within a 17 days period. This way,
The air-tissue interface of specific organs is this model showed a method for a reproducible,
known to be a critical factor in carcinogenesis mass production of organotypic melanoma mod-
due to the nefarious effects of certain external els (Müller and Kulms 2018).
factors, like smoking (Van Duuren and This interface, however, is still quite unex-
Goldschmidt 1976) or pollution (Demetriou and plored, particularly in lung cancer where the
Vineis 2015), or sun radiation (Kripke 1981). exogenous stimuli are the main cause of cancer,
Moreover, these interfaces are critical for the according to the United Kingdom National
physiological behavior of the correspondent tis- Health Institute (NHS) website for causes of lung
sues, being responsible for the tissue’s main cancer. Models of skin or the respiratory or gas-
function, such as lung, skin or intestines. tric tract should include, in following years, this
In the case of lung, intestines and stomach interface and validate the disease with organ
models, it is necessary to create tissue-air inter- function impairment.
9.2.3 Biochemical known and well-established tumor’s hallmarks is

the oxygen gradient within the tumor mass, that
The biochemical interactions of interest in a can- is known to drive a pH gradient known as the
cer model are diverse. These molecular cues Warburg effect (Warburg 1956; Lu et al. 2002).
either happen in the form of gradient or as a uni- Recently, 2D breast cancer models, using soluble
form signal across the tumor mass, solubilized or oxygen gradients, allowed a deeper understand-
deposited in the surrounding ECM. Furthermore, ing about the mechanisms of hypoxia-related
they can be composed of small molecules or pro- migration (Enokida et al. 2017; Tsuruno et al.
teins, thus making the characterization and emu- 2018). Inducing a gradient of 0–3%, generated
lation of the tumor’s biochemical interactions cell migration towards the oxygenated region.
particularly challenging. However, over the last This migration significantly decreases with the
few years, a number of promising works have cell density. Thus, oxygen-dependent migration
been suggesting innovative designs incorporating is, actually, a pH-dependent migration, from
these cues. acidic to more neutral locations (Enokida et al.
2017; Tsuruno et al. 2018).
9.2.3.1 Soluble Factors To further understand the role of hypoxia-
The interplay between cells within the tumor related genes on chemo resistance, spheroids of
microenvironment is often done by the intermedi- colorectal cancer cell line were categorized cor-
acy of soluble factors. These are the most diverse relating size with oxygen levels. Overall, the
signals responsible for cell-cell communication spheroids that were in hypoxia and developed
and the characterization of the activated pathways necrosis presented gene expression levels closer
by several of these molecules has been extensively to in vivo ones, demonstrating an increasing
reviewed (Kessenbrock et al. 2010; Jung et al. resistance to 5FU, a chemotherapeutical agent
2015). However, to identify and characterize the clinically used for colorectal cancer treatment
role most relevant soluble factors for the develop- (Däster et al. 2017).
ment of malignant phenotype to be particularly However, some behavioral consequences are
targeted by new drugs, there is the need to create still unclear, namely the role of hypoxia and
tumor models where the intercellular relations can hypoxia-related genes on chemo resistance still
be isolated and individually assessed. requires further investigation, despite the clear
Recently, the role of neural Wiskott-Aldrich relevance of hypoxia on the adaptive capacity of
syndrome protein (N-WASP), an actin cancer cells to chemotherapy.
cytoskeleton-related protein, in the motile and
invasive behavior of pancreatic ductal adenocar- 9.2.3.2 ECM-Bound Factors
cinoma, was studied using a 2D model. Even The delivery of biochemical stimuli is often done
though the promotion of motile behavior on can- by fixing molecular factors known to be present
cer cells by autocrine stimulation with N-WASP surrounding the tumor site to the ECM, therefore
was already known, this model revealed that producing an enhanced effect with lower concen-
impairing N-WASP led to unresponsive cancer trations than when present in solution (Hollier
cells to chemo attractant such as lysophospha- et al. 2005). Such factors have shown to be altered
tidic acid (Juin et al. 2018). The assessment of in diseased tissues and to drive aggressive behav-
these feedback loops that can now be targeted ior in cancer cells (Huleihel et al. 2016).
with novel therapies, would not be possible with- To replicate this effect two main approaches
out an in vitro simplified construct. can be applied: top-down, through decellulariza-
Small molecules and secreted proteins are tion, and bottom-up, by the regular assembly of
soluble factors for intercellular communication components in vitro (Lee et al. 2011). Despite the
but they can also be metabolites that indicate an most common one is the bottom-up due to the con-
abundance or shortage of nutrients or over- trol achieved in every component of the model, the
population. For example, one of the most well- top-down approach has been the one with larger
success in recreating this interaction due to its biomimetic modeling approach for aggressive
complexity. glioblastoma (Koh et al. 2018).
Decellularization, as a top-down approach, However, this dependence on patient-derived
consists in extracting a part of the tissue, with- tissues limits the applications of decellularized
draw the cellular part of the excised tissue and matrices, since the availability is short and vari-
use it to create a full in vitro model that has the able from donor to donor. Further research into
native ECM’s biochemical signature. the composition, particularly the spatial distribu-
In vitro models using decellularized tissue tion of factors within these matrices is para-
evaluate only the effect of ECM-bound motifs, mount, which can potentially be recapitulated by
when compared to matrices with the same physi- synthetic/engineered approaches (Yang et al.
cal properties. For such properties, a model for 2016; Zhu et al. 2018b).
colorectal cancer was recently proposed using
decellularized matrix allowed to conclude that,
when compared with a healthy control, tumor tis- 9.2.4 Cellular
sue showed a decreased angiogenic potential in
chorioallantoic membrane assays and led to the The tumor’s cell mass is surrounded by stromal,
expression of autocrine and paracrine factors epithelial, endothelial and immune cells that have
characteristic of native tumors by cancer cells been proven relevant in carcinogenesis, EMT and
after recellularization with a colorectal cancer metastasis settling (Magdeldin et al. 2017; Chung
cell line (Piccoli et al. 2018). et al. 2017; Angeloni et al. 2017; Heinrich et al.
The manifestation of cancer hallmarks by the 2019). However, intercellular crosstalk is complex
in vitro model validates it as a biomimetic effort. and the feedback mechanisms are not fully charac-
Thus, to verify if a patient-derived decellular- terized. In animal models is impossible to assure
ized matrix of a glioma triggered native-like control on the interplay among the cellular niches.
behavior on glioblastoma in vitro, glioblastoma For that reason, in vitro models represent a
spheroids were encapsulated in such matrix and flexible tool to address interfaces and assess the
their behavior was assessed. The cellular and influence of cellular composition over cancer
collective dynamics were quantified using tumor outcome, as it will be addressed in the following
spheroids cultured in decellularized sub-sections.
ECM. Taking into account the heterogeneity of
the glioblastoma multiform tumor, the shape 9.2.4.1 Stromal
and expression levels of ECM remodeling pro- There have been some models focusing on the
teins, such as matrix metalloproteinases, were study of individual or group interactions among
assessed and compared to the results obtained in stromal cells and cancer-associated cells over
collagen matrices. Using tumor spheroids, glio- last years. The development of some techniques
blastoma’s migration patterns were isotropic, allowed, in the last few years, the generation of
meaning that cells did not exhibit a particular or interfaces with controlled features and dimen-
preferable migration direction. Moreover, it was sions, increasing reproducibility and monitoring
verified that cells in the tumor invasive front capacity.
were faster, showed an elongated morphology, Following this reasoning, metastatic colorec-
while the core consisted in round, less motile tal, ovarian and stomach tumors’ microenviron-
cells, similarly to what is known from in vivo ment models were developed by culturing
studies (Sottoriva et al. 2013). Tumor cells cancer cell lines with primary fibroblasts
maintained in the decellularized tissue predomi- (Chung et al. 2017).
nantly presented higher gene expression of One of the most relevant interactions to be
ECM remodeling factors when compared to the explored is the paracrine interactions in cancer
control group cultured in collagen, exhibiting an cells’ viability and stromal angiogenic potential.
invasive profile. These features validate this as a Using in vitro models, it was made possible to
perceive how the paracrine interactions affect sue specific hierarchy in comparison to explants
cancer cells’ viability and stromal angiogenic to understand how the spatial distribution affects
potential. By co-culturing of a cancer cell spher- the interplay between the tumor and the sur-
oid with a 2D monolayer of fibroblasts, cancer rounding stromal cells.
cells cytoplasmic area and number of nuclei were
significantly increased when compared to a 9.2.4.2 Immune Cells
monoculture of cancer cells (Chung et al. 2017). Recently, the interest on the triple interaction
Using a microfluidic platform designed for the between the immune, stromal and cancer cells
study of this cancer model paracrine effect on the has risen, particularly considering the panoply of
recruitment of endothelial cells, researchers dem- models lately proposed for the study of these
onstrated that the in vitro cultured ovarian and interfaces (Herter et al. 2017; da Costa et al.
metastatic colorectal cancers contributed to 2018; Suresh et al. 2018; Pereira et al. 2019).
reduce the angiogenic potential of the fibroblast This interest is mainly based on the applications
culture, while the stomach cancer model led to an of immunotherapy for cancer treatment
increase of angiogenesis (Chung et al. 2017). (Gopalakrishnan et al. 2018).
Thus, this model showed that stroma-tumor inter- Strikingly, several evidences of cooperation
action occurs in a site-dependent manner. among these cells have also been reported both in
However, though several models were created carcinogenesis and in tumor progression. To
along the years, the comparison of the results understand this cooperation, a 3D bioprinted co-
obtained with the in vitro construct with ex-vivo culture model of glioblastoma cells and macro-
models cultured in the same conditions is usually phages showed that this interplay led, firstly, to a
absent. polarization of the macrophages towards a
To perform such validation, the migration and cancer-associated macrophage phenotype, which
invasion patterns of glioblastoma cellular spher- is followed by an increased aggressiveness and
oids cultured onto several layers of astrocytes invasiveness of the glioblastoma cells. Moreover,
were compared to patient-derived samples. The when this interplay is disrupted the tumor’s
model exhibited glioblastoma hallmarks also growth is diminished and its sensitivity to che-
verified in the brain cancer slices, such as indi- motherapy increased.
vidually migrating cells, collective strands Considering this contribution of immune cells
extending along blood vessels, and multicellular for cancer progression, the interplay between stro-
networks of interconnected glioma cells infiltrat- mal and immune cells is also of particular curios-
ing the neuropil. These observations were not ity since it is an interface that has shown related to
present in a control group of spheroids without cancer progression (Heinrich et al. 2019). For
the surrounding astrocytes. Essentially, this study instance, to study the influence of cancer-
showed the relevance of the presence of stromal associated fibroblasts and mast cells on carcino-
cells around the tumor to mimic the invasion and genesis and first cancer progression, a 3D model
intravasation is revealed by this study (Gritsenko using patient-derived cancer-associated and
et al. 2017). healthy fibroblasts, mast cells and healthy prostate
Stromal cells are essential to create represen- epithelial cells was developed sequentially to
tative models for drug testing. However, for that understand the effects of an established niche over
purpose, validation of the results found in vitro cancer cells. Thus, in an initial stage, fibroblasts,
with in vivo or patient-derived samples is crucial healthy and cancer-associated, were cultured in a
to understand the relevance of these models. 3D-printed scaffold to produce native tissue.
Moreover, the observation and characterization Afterwards, healthy prostate epithelial cells were
of these cancer-stromal cells interplay also has cultured onto the scaffold and morphogenesis was
become increasingly complete, by using more observed and measured. Finally, mast cells were
advanced materials, fabrication and visualization introduced into the model and the morphological
techniques. The following years must include tis- changes were assessed. When compared to healthy
fibroblasts, cancer-associated cells led to signifi- cancer targeting. However, the manipulation of
cant morphogenesis towards a cancerous pheno- the immune cells still requires a profound charac-
type, which was exacerbated with the presence of terization of the effects of the biochemical loops
mast cells (Pereira et al. 2019). between cancer cells and cancer-associated
Furthermore, this synergistic interplay has immune cells. For this purpose, proteomic and
been observed in some other in vitro models (da secretomic approaches must be assessed in the
Costa et al. 2018; Suresh et al. 2018), showing future when applying in vitro models, ideally
the potential to be a hallmark and a therapeutic with real-time monitoring technology.
target for cancer progression. For example, het-
erocellular spheroids, containing pancreatic can- 9.2.4.3 Epithelial
cer, fibroblast and macrophage cell lines, were During the metastatic process of extravasation,
proposed as a pancreatic cancer microenviron- depending on the site, cancer cells can face a
ment model. The authors verified that cancer layer of epithelial cells or a bulk of ECM and
cells underwent severe morphological changes stromal cells. The importance of the interplay
and significantly increased the expression of between cancer cells and the epithelial layer and
stem cell factor when compared to homotypic the relevance of the target cells on metastasis for-
spheroids (Suresh et al. 2018). mation is still unclear and it is challenging to
Alternatively, using a bone metastasis prostate measure when using animal models. In vitro
cancer model, da Costa et al. 2018 showed that tumor-epithelium models have been helping in
tumor-associated macrophages and cancer- the characterization of these interfaces and
associated fibroblasts were able to modulate the enabling quantification of their interactions for
viability, cell proliferation and response to growth metastasis formation.
factors of patient-derived cancer cells (da Costa For instance, to study tumor and target tissue
et al. 2018). So, the presence of immune cells has interaction, two models have recently revealed
been shown to directly influence the progression the importance of epithelial cells in metastatic
of varied cancer types. cancer cells adhesion (Angeloni et al. 2017;
Despite the evident role of the cells from the Marques-Torrejon et al. 2018). A simplified in
immune system in cancer aggressiveness high- vitro model of metastatic breast cancer metastasis
lighted during the present section, immune cells in bone was assembled. Adipose stem cells
also have therapeutic potential, which can also be (ASC)-derived osteoblasts onto polyurethane and
reproduced and tested using in vitro models. For co-culturing breast cancer in a layer-by-layer
the colorectal cancer, to model the immunother- procedure onto the osteoblasts. Interestingly,
apy fibroblast and cancer cell lines were encapsu- breast cancer cells were not able to adhere and
lated and co-culture were treated with a spread onto the polyurethane surface as they did
fibroblast-targeted T-cell biospecific antibody. when cultured onto osteoblasts (Angeloni et al.
Primary peripheral mononuclear blood cells were 2017), demonstrating that the tissue’s epithelial
added to the model and their effects over the layer is essential for metastasis formation.
treatment were assessed by means of measuring Similar results were achieved after the injection
viability of cancer cells over time. For that pur- of glioblastoma stem cells into ex vivo cultured
pose, the mononuclear cells were cultured in sus- mouse adult brain slices. Cancer cells particularly
pension and stimulated using cytokines to engrafted and responded to endothelial stimuli
promote infiltration into the spheroid. Strikingly, when inserted in a biomimetic neuronal epithe-
the immune cells led to a significant enhance- lium (Marques-Torrejon et al. 2018).
ment of the treatment by acting as both perme- However, the interplay between metastatic
ation and killing agents (Herter et al. 2017). cancer cells and the target tissue’s epithelium
Overall, different recent studies have been is more complex than the management of the
demonstrating that immune cells can drive cancer adhesion of circulating tumor cells, affecting
progression or be molecularly reprogrammed for permanently the biochemical profile of the tis-
sue’s epithelial cells. To study these changes, ilarly to clinically observed metastatic lesions
the influence of metastatic lung cancer cells (Xu et al. 2016b).
over brain, liver and bone was studied using a Overall, the interface in between the tumor
body-on-a-chip device with astrocytes, hepato- and the epithelial layer has demonstrated to be
cytes or osteocytes cultured in interconnected extremely relevant to be reproduced in vitro,
chambers, respectively (Fig. 9.3). With no particularly in the study of the mechanisms

endothelial barriers, metastasis can be formed underlying metastasis settling. However, the cur-
over the three-epithelium layers. However, the rent state-of-the-art still lacks further investiga-
newly formed metastasis did not simply tion on quantifying the contribution of cancer
adhered and proliferated, having also induced cells secreted extracellular vesicles previous to
aberrant gene expression in the target tissues intravasation and cell-cell contact upon extrava-
epithelium, leading to severe cell damage, sim- sation that will require specific and advanced
Fig. 9.3 Body-on-a-chip for the affinity assessment of tasis formation sites is related to the interplay of cancer
lung metastasis for bone, liver and brain epithelia settling. cells and tissue’s specific epithelium. (Adapted with the
With interconnected epithelium chambers and metastatic permission of (Xu et al. 2016b). Copyright 2016 American
lung cancer cells in circulation, the differential settling of Chemical Society)
cancer cells is an indication that the preference of metas-
analytical methods integrated on epithelial-can- able enough for an efficient drug delivery. For
cer interface model devices. that reason, in vitro drug tests for glioblastoma
have to take into account the impermeability of
9.2.4.4 Vascular the surrounding vascularization of the tumor,
Tumor vascularization is a key parameter in requiring a BBB model based on human cells and
tumor evaluation and it is linked to late diagnosis interfacing the in vitro tumor model. This way,
and consequent poor prognosis. Nonetheless, the the models can recreate the biggest obstacle to
vascular system is a central interface in each the drug delivery.
grade of cancer. Angiogenesis is exacerbated in Using a tumor-blood interface, glioblastoma
primary tumor sites, metastatic cancer cells targeting drug delivery systems were effective in
undergo intravasation into the blood stream overcoming BBB with a value of transendothelial
through these capillaries and undergo extravasa- electrical resistance of 250 Ω/cm2 in vitro
tion at specific tissue sites, without having a (Heggannavar et al. 2018). Despite being promis-
known motif for it. To study each of the particu- ing, these results were obtained using a BBB pre-
larities of this interface, in vitro vascular models senting a transendothelial electrical resistance
with an established cancer grade were developed one order of magnitude inferior to the one of
and contributed over last years to understand the native BBB measured in mice (Butt et al. 1990),
mechanisms underlying angiogenesis in primary revealing the urge to create tumor and BBB mod-
tumors and extavasation, in particular, in meta- els more similar to the native tissues.
static cancer as it will be explored herein. These interactions, however, are not particular
of brain cancer. There have been other proposed
Primary Tumor vascular cancer models of different origins
Increased vascularization is a generalized tumor (Magdeldin et al. 2017; Klimkiewicz et al. 2017;
hallmark, but it is a particular characteristic in Sánchez-Rodríguez et al. 2017; Racordon et al.
some cancer types. The tumor-vascular interface 2017; Jung et al. 2017; Roudsari et al. 2018;
is characterized by a gradient of nutrients, fluid Boussommier-Calleja et al. 2019; Jiménez-Torres
shear stress and oxygen but also by an interaction et al. 2019). Such studies have been providing
between cancer and endothelial cells that modu- insights about type-specific and general biomark-
late tumor’s progression. Although it is difficult ers and hallmarks that are particularly relevant
to distinguish the contribution of each effect, in for diagnosis and therapeutic development.
vitro models revealed that a simple cellular inter- The 3D models of tumor vasculature is crucial
play reduced the expression of endothelial mark- to properly mimic the cancer environment for
ers by endothelial cells and increased proliferation new drug development, since the 3D constructs
on cancer cells (Heggannavar et al. 2018). within scaffolds or hydrogels enable the spatial
The capacity of the glioblastoma to manipu- formation of vascular-like segments, improving
late the surrounding endothelia, disrupting its the mimicry of the cancer-vasculature interface.
characteristic tight junctions by secreting vascu- For example, using a co-culture of a colorectal
lar endothelial growth factor (VEGF) is a hall- cancer cell line, human fibroblasts and human
mark of this tumor type (Herter et al. 2017), umbilical vein endothelial cells (HUVECs), the
which is a particular case of a primary tumor- interplay between stromal cells and HUVECs
blood interaction. The BBB is the brain endothe- was highly overdriven by the presence of the
lium covered by monolayers of pericytes and colorectal cancer cells, a classic cancer hallmark,
astrocytes, respectively, which create an imper- when cultured in 2D. However, once the tumor
meable membrane to most compounds based on started growing in 3D, cancer cells decreased
the tight junctions that glioblastoma cells are their expression of CK20, a primary colorectal
known to disrupt. cancer molecular biomarker, while increasing
Despite this tumor-blood interaction, the their invasive phenotype, disrupting the newly
cancer-destabilized vessels are still not perme- formed vessels (Magdeldin et al. 2017).
As a different 3D approach, spheroids are par- cancer and endothelial cells regulates the behav-
ticularly useful 3D models to create biochemical ior of both cell types when in co-culture.
gradients. Thus, when combining spheroids with However, the extent of these alterations and how
endothelial cells, it enables the study of angio- permanent they are is still unclear.
genesis in a tumor mass, allowing the study of the To better understand the consequences of
loops that regulate such interaction, with special cancer-endothelium interface over the behavior
interest for drug discovery. To further do so, lung of endothelial cells, patient-derived renal tumor
(Roudsari et al. 2018), liver (Jung et al. 2017) and tissue and healthy endothelia were cultured in a
skin (Klimkiewicz et al. 2017) cancers were microfluidic device. The tumor-derived endothe-
modulated using heterotypic spheroids of angio- lial cells generated an increased number of
genesis to compare primary cancer types of dif- sprouts from the cultured vessels when compared
ferent grades. to healthy tissue derived ones, exhibiting an
From the lung cancer model it was possible to increased angiogenic potential. This pro-
verify that lung cancer spheroids with more met- tumorigenic behavior of recruited endothelial
astatic phenotype recruited twice the endothelial cells is something that can play a role in cancer
cells that benign tumor spheroids did, showing reappearance, and should be targeted in next gen-
that with a grade increase, the pro-angiogenic eration therapies (Jiménez-Torres et al. 2019).
potential also increases (Roudsari et al. 2018). Despite the fact that tumor cells alter the phys-
One of the most important roles of angiogen- iology of tumor surrounding endothelial cells,
esis is the control of oxygen levels within a tissue the aberrant response of immune cells may also
that is particularly relevant in cancer, since it is be responsible for a significant part of tumor’s
correlated to tumor aggressiveness. Since physiological angiogenesis (Shi et al. 2019). In a
hypoxia is key aspect on cancer aggressiveness, a validated vascular model for monocyte differen-
hypoxic tumor model containing a surrounding tiation and maturation, where melanoma and
endothelium was designed to understand how the breast cancer cells were introduced, it was veri-
balance between hypoxia and normoxia is con- fied that while monocytes stopped cell progres-
trolled by the tumor. sion and, particularly, extravasation,
Specifically, a melanoma model was devel- monocyte-derived macrophages did not produce
oped to study the effects of hypoxia over cancer any significant effect over cancer cell viability
cells recruitment of endothelial cells. It was (Boussommier-Calleja et al. 2019). This model
found that the combined effect of the co-culture showed immune cell type dependent modulation
and hypoxia conditions led to the recruitment and of the tumor-blood interface, highlighting the
posterior stabilization of the population of endo- urge of deeper understanding of the immune sys-
thelial and epithelial cancer stem cells, showing a tem role on tumor angiogenesis processes.
tissue-like behavior similar to in vivo observa- The creation of vascular models to study pri-
tions (Klimkiewicz et al. 2017). Nonetheless, this mary cancer behavior is important to understand
led to the hypothesis of a cancer-mediated oxy- the balance of oxygen and nutrients supply within
gen balance by endothelial cells recruitment. the tumor microenvironment. This balance
To assess the angiogenesis-mediated oxygen between what is necessary to allow cancer cell
equilibrium in tumors, liver cancer spheroids viability and what is required to induce malig-
were co-cultured with endothelial cells at 2% of nant phenotype may then be targeted in future
the total number of cells. HUVECs not only therapies and significantly increase treatments
reduced the necrotic core of the spheroid but also efficiency.
induced EMT transition-related genes and
angiogenesis-related genes in cancer cells, lead- Metastasis
ing to a gene expression profile closer to the one Metastasis are of tumors formed in secondary
verified in patient samples (Jung et al. 2017). sites, within a different tissue from the primary
These models showed that the interplay between site, after dissemination of metastatic cancer cells
through the blood stream and extravasation at a extravasation, which are relevant for the develop-
specific site and tissue. The cellular consequences ment of novel therapeutics (Xu et al. 2016a).
that the extravasation process induces and the Metastatic models like the one described in
selectivity behind it are still unclear and particu- this sub-section will be used in following years
lar challenging to observe in vivo due to the for further understanding the molecular mecha-
spatio-temporal unpredictability of the process. nisms behind tissue specific recognition, even
The existence of a capillary network is a key though these developments are dependent on
feature to observe biochemical alterations pro- advanced imaging techniques coupled with trans-
moted by the extravasation process. However, the parent devices.
creation of these capillary networks are rather
challenging and, consequently, uncommon in in 9.2.4.5 Microbiological
vitro models. This issue was overcome in a breast Several tissues present a symbiotic relationship
cancer metastatic in bone model was developed with microbes, making them a crucial part not
where the bone ECM and vascularization were only of tissues function, but also of systemic
mimicked. To do so, bone mesenchymal stem homeostasis (Gopalakrishnan et al. 2018; Ma
cells (BMSCs) were co-cultured with endothelial et al. 2018; Mrázek et al. 2019), and consequently
cells at an optimized flow rate and oxygen con- a relevant variable to be considered for in vitro
centration to achieve a perivascular bone model. models. For instance, the intestinal mucosa and
After so, cancer cells were introduced to the the skin are tissues where the microbiome signifi-
model and extravasation occurred, as predicted, cantly contributes for healthy physiology (Chng
validating it as a metastasis model and grew in et al. 2016; Daniel et al. 2017).
low-proliferative state when compared to static Deregulations in the profile of these microbi-
culture. However, once exposed to sunitinib, a omes have been implied in several diseases and
chemotherapeutic agent, the settled metastasis their presence in healthy tissue in vitro models
cultured in dynamic conditions, post-has been increasing over last years. This way,
extravasation, showed resistance to the drug, several probiotics have been proposed for exam-
exhibiting a decreased impairment of its growth ple to fight intestinal infections, substituting, at
rate. This way, this platform not only showed to least in early stages, the use of antibiotics (Hao
be a valid model for breast cancer metastatic in et al. 2015). Since some microbes, namely bacte-
bone as it showed to reveal mechanisms behind ria, were enriched in tumor sites, the hypothesis
chemo resistance of these tumors. of having a probiotic disrupting this cooperative
Some metastasis models have been developed situation has been proposed for possible future
over recent years. An interesting work stood out cancer therapy.
by the increased potential for allow studying of To investigate this possibility, Lactobacillus
the mechanisms of cancer-tissue selective rhamnosus, a common probiote, was co-cultured
extravasation. with Fusobacterium nucleatum, bacteria shown
Using a blood-brain barrier model built in a to be particularly present in colorectal cancer,
microfluidic device, malignant cancer cell lines and together with spheroids of a colorectal can-
originated from lung, breast, skin and liver cancer cell line.
cer were circulated in contact with the vascular Interestingly, Lactobacillus rhamnosus con-
barrier. Once it contacted the model, the invasion siderably impaired the attachment of
speed was measured. Lung cancer was the fastest Fusobacterium nucleatum when cultured
to invade, followed by breast cancer and mela- simultaneously and in the same cell density,
noma, respectively, while liver cancer was not mainly due to its binding capacity. In protective
able to perform extravasation across the and competitive models, the probiotic was able to
BBB. Thus, the in vitro model led to significant reduce in 80% the binding capacity of
differences among cancer cell types in BBB Fusobacterium nucleatum, where Lactobacillus
rhamnosus was cultured previous and simultane-
ously to Fusobacterium nucleatum, respectively. the idea of classical stirring using propellers
Nonetheless, in displacement assays, mimicking (Zhao et al. 2016).
an eventual treatment of previously infected Recently, dynamic systems have been minia-
tumor tissue with Fusobacterium nucleatum, the turized and adapted to tissue engineering applica-
removal only achieved 20% (Mokhtar et al. tions. Particularly in the field of in vitro models,
2019). microfluidic devices, a miniaturized bioreactor
Still, tumor models incorporating microbi- concept, has become increasingly popular (Zhao
omes are yet underdeveloped (Villenave et al. et al. 2016). These devices allow a controlled cul-
2017; Jalili-Firoozinezhad et al. 2018; Réquilé ture, real-time visualization and the multiplex
et al. 2018; Zhu et al. 2018a; Shin et al. 2019). In measurements with small sample volume.
following years, the incorporation of the microbi- Nonetheless, microfluidic platforms are pro-
ome in 3D models will be done by means of duced using different architectures, dependent on
establishing the most representative bacteria in the target mimicked tissue and the balance
native microbiome and their influence over can- between complexity and processability. Besides
cer cells viability and proliferation, identifying the discussion about this dichotomy, this section
what can be the largest contributors for microbe- will also address some innovative bioreactors
induced cancer development. designed for 3D tumor interfaces that are exam-
ples of the balance between the achievement of a
representative model and the reproducibility nec-
9.3 D Models Integration
3 essary for the clinical translation.
in Dynamic Culture Systems
The presence of fluid flow influences the behav- 9.3.1 Dynamic vs. Static
ior of tissue culture in vitro. Nonetheless, it is
only relevant to model if it is used to recapitulate The theoretical effects of physiological fluid
in vivo dynamics. There are interfaces, namely shear stress over tumor progression have been
the tumor-blood, where the fluid shear stress can previously discussed, particularly the role of 3D
be highly important for the progression of the models in the characterization of this interface.
disease. However, under a biomimetic perspective, the
The importance of the establishment of a practical use of dynamic conditions is particu-
dynamic culture of 3D tumor interfaces will be larly relevant because the human body is a con-
addressed in the first part of this section, with a struct of organs perfused by blood indirectly
special focus on the relevance of its application in through vessels. This perfusion occurs by a leak-
anti-cancer drugs standardized screenings. age from the vessels, with the fluid being col-
The second part of this section will be directed lected by the lymphatic system.
mostly at volume scales and how scale-up strate- The differential of pressure between these two
gies have been designed in the development of systems, vascular and lymphatic, drives the fluid
engineered tumor interfaces. over cells with a flux that decreases proportion-
Standard bioreactors are usually classified ally to the distance from the blood vessel. The
according to its geometry, stirring, and feeding flux that flows through the tissue varies quite sig-
mode. In classical bioreactors classification, nificantly according to the activity of the organ-
there were two main types of systems – tank and ism. All the mentioned features can be potentially
tubular (Cooney 1983). Regarding the stirring, mimicked using diverse dynamic systems,
the knowledge gathered about the effect of shear namely by applying advanced microfabrication
stress over the phenotypic homeostasis of sensi- techniques.
tive cell types, namely mammalian cells and, par- Advanced designs allow the establishment of
ticularly, the development of in vitro models constructs with well-defined interfaces under
using biomimetic strategies have almost retired real-time visualization of the interactions along
the interface. Cao et al. (2019) developed a perfu- This device allowed the verification of a cell
sion bioreactor that mimicked the near tumor viability decrease as the distance from the mim-
blood and the lymphatic systems, siding the icked blood vessel increases, as was predictable
tumor mass encapsulated in photo-crosslinked from a mass-transfer perspective. The gene
methacrylated gelatin. The vessels were built expression levels of tumor progression-related
using 3D printing of photo-crosslinkable biomarkers like metalloproteinase and prolifera-
poly(ethylene glycol) gels, instantly gelified tion markers also verified the effects of the drug,
using ultraviolet light. This design increased the showing the robustness of this model. Although
fluid flux through the tumor mass creating easily further details about time of fabrication and pro-
tunable dynamics, enabling high-throughput fab- cessing would be required to classify this as a
rication and near in vivo flow dynamics relevant high-throughput system, the robustness of the
for drug testing over different tumor models (Cao device makes it a candidate for industrial use.
et al. 2019). The use of these channel-based microfluidic
However, the commitment with the simplifica- devices require a total change in the protocols of
tion of processes for reproducibility is a major the pharmaceutical industry. Most industrial pro-
challenge for widespread dissemination. Li et al. tocols are still optimized for well plates-based
(2017) developed a platform for ovarian cancer drug screening.
behavior assessment when contacting the perito- Also, the translation to the widespread use by
neal membrane, as it often happens in this cancer the pharmaceutical industry requires more than
type. However, this device differs from others due the mere achievement of the design parameters.
to the generation of physiological interstitial fluid Within the parameters mentioned as the key ones
shear stress. Strikingly, the design and fabrication for platform design, uniformity is the main goal
process are particularly reproducible, since the to achieve. As a relevant example for dynamic
visualization is optimized by the use of the mini- systems, flow pumping can cause differential
malistic design arresting cells just by cell adhe- pressure along the device, triggering fluctuations
sion to fibronectin or to the mesenthelial cells and non-conformity.
using a layer-by-layer approach. Thus, this device To face this situation, Chen et al. (2019) cre-
shows enormous potential for future standardized ated a high-throughput pumpless system for drug
high-throughput drug screening (Li et al. 2017). testing on a tumor-stromal interface that allows a
The translation of these devices to the phar- uniform shear stress over the entire tissue, using
maceutical market is hugely dependent on the a swinging well-plate design. Using arrays of 4
capacity of up scaled standardized production, as wells per condition, the device is a simple sys-
well as on the robustness of disease models pro- tem, which avoids the use of external pumps,
duction. For that purpose, these systems must therefore more translational and user-friendly.
ally advanced microfabrication techniques with Regarding the tissue model, it consists of a
real-time assessment capacity. Carvalho et al. 2-layer mode, one of cancer other of stromal
2019 developed a platform for standardized cells. Though there are interactions in 3D, the
colorectal targeting drug testing through a tumor- system do not recapitulate 3D heterogeneities or
blood interface. Using an advanced design, this gradients and cells do not form an architecture
platform allowed the creation of a centered tumor that characterizes in vivo tumors (Chen et al.
mass encapsulated in Matrigel® surrounded by a 2019).
circular HUVEC blood vessel by a set of chan- There is still, however, a large market need for
nels and baffles. As a proof-of-concept of the use tumor interfaces, with simplified protocols, auto-
of this platform for drug screening, the effect of a mation, and up scalable features.
drug-loaded advanced delivery system, through In the following sub-section, the most recently
the tumor-blood interface, over the tumor viabil- industrially translated devices that were used for
ity was assessed and quantified using image pro- 3D interfacial tumor models will be reviewed.
cessing techniques (Carvalho et al. 2019). This analysis will allow better understanding of
the features that characterize the industrial sys- since the system was developed for a cultured
tems and what makes them different from most tumor mass. It has only a chamber with no spe-
of the academic outcomes. cific design interface models.
The development of interface models often
requires the existence of two inlets or a valves
9.3.2 Industrial Transition system. For the creation of gradients, that can
drastically alter cancer cell dynamics as was veri-
Several devices containing 3D tumor models are fied in the biochemical interfaces sub-section, the
often considered promising in a field, but not use of multiple channels is the most common and
suitable for industrial drug screening. The main robust strategy. For that reason, Mimetas© devel-
reason behind it relates to the equilibrium among oped the OrganoPlate®, a system for the purpose
mimicry, reproducibility, productivity, and indus- of having perfusable tumor interface models
trial validation, which are necessary for industrial (Fig. 9.4). This bioreactor is a pumpless device,
production and general use on the pharmaceuti- with contacting channels that can provide 364
cal industry. simultaneous replicas. The number and position
Disregarding the fluid dynamics, models as of channels can be tuned according to desired
those presented along this chapter often differ model and it has been applied in healthy (van
from each other in one or several parameters (e.g. Duinen et al. 2019) and in tumor tissue (Lanz
cell origin or type; biomimetic stimuli, such as et al. 2017; Ogmundsdottir et al. 2018; Queiroz
growth factors or metabolic cues; mechanical et al. 2018) modeling, validating its usefulness in
properties of the supporting material or shape of standardized drug testing.
the 3D model, design of the bioreactor platform However, none of these systems allow the
when existing). However, while the increasing establishment of a more complex microenviron-
diversity of approaches is paramount for the evo- ment that can be insufficient for the replacement
lution of the research field, it also represents a of drug testing in animal models. For that pur-
delay on standardization for industrial uptake. pose, there is the need to create dynamic models
Devices for tumor-targeting drug high- with the mimicking of the tissue’s hierarchical
throughput screenings shall have designs that structure that vary enormously among tissues.
consider productivity, increasing conformity and Therefore, such system must have a specific
standardization simultaneously to biological vali- design interface constructs that allow the use of
dation. To exhibit the robustness for widespread optimized material methods, such as the use scaf-
drug testing using tumor models, the system must folds or hydrogels, and the real-time assessment
proof to work regardless of the cancer type or of all the components of the interfacial model. A
kind of interface. dual-chamber 6 well-plate sized bioreactor
With that intent, Kiyatec© validated the per- (Canadas et al. 2017) showed to be a robust strat-
fusable system 3DKUBE® for the culture of egy for interfacial models. This innovative sys-
patient-derived xenografts for immuno-oncology, tem can allow the creation of gradients and
glioblastoma (Ashley et al. 2018), breast (Guo interfaces, either solid or soluble in dynamic con-
et al.), and ovarian cancers (Desrochers et al. ditions (Canadas et al. 2018).
2015; Shuford et al. 2019). Moreover, in each Nonetheless, the currently increasing industri-
model, cancer cells were isolated and cultured ally translated devices constitute robust upgrades
with exogenous stromal, endothelial and immune to 2D and static screenings, opening-up ways for
cells, usually present at the tumor site, and the even more realistic systems. There is still, how-
models were validated by molecular character- ever, the market need to combine efforts to build
ization and chemo resistance to known models of tumor interfaces with defined features,
compounds. simplified protocols allowing automation and
However, the recreated interfaces are not actu- upscale of the ones currently existent in the
ally discrete nor the dimensions were controlled academia.
Fig. 9.4 Example of OrganoPlate® for interface forma- central channel while perfusing two different medias in
tion. (a) Design and placement of 3 lane OrganoPlate®, the outside lanes. (A and B were adapted from van Duinen
where (b) two different gradients can be formed within a et al. 2019)
9.3.3 Concluding Remarks and expensive, which creates important barriers

and Future Outlook for industrial application.
Strikingly, when specifically mimicking meta-
The combination of 3D cues of the tumor’s inter- static cancer models, either based on organism
faces in a dynamic platform contributes for the models, like body-on-a-chip, or just focusing on
recapitulation of shear stress, which is extremely one vascular model of barrier to monitor intrava-
relevant for drug effectiveness. Finally, the trans- sation/extravasation, the current approaches are
lation of the academic platforms to the industry often too simplistic. Frequently, the state-of-the-
was discussed. For that purpose, the reviewed art still neglects models able to replicate systemic
models clearly show the progress margin and the effects, a requirement to replace animal experi-
milestones to be achieved in the following years. mentation in the future. The models that address
Importantly, bioreactors and microfluidic sys- systematic consequences are often built on the
tems are now tending to avoid pumping systems basis of proof-of-concept approaches, without
due the industry’s demand for user-friendly, sim- using features with physiological and pathologi-
pler setups. cal values. The validation with commercially
Specifically from the tissue modeling perspec- successful and non-successful drug candidates is
tive, the most recent interface tumor models are, also uncommon, being also a main requirement
in general, already developed as 3D constructs. to prove the model’s robustness.
The increased complexity when compared to 2D Despite that, the recent advances represent an
contributes to a better representation of native tis- increased effort to replace or minimize animal
sue. To do so, advanced microfabrication tech- models and improve the predictability of false/
niques are applied, which makes the developed positive results before the clinical trials.
models more biomimetic when compared to 2D Furthermore, the regulatory pressure to use in
and homogeneous 3D, since it represents a larger vitro systems started with the cosmetics industry,
number of interactions among cancer cells and and will expand across the pharma industry in the
surrounding ECM. On the other hand, these pro- following years. So, the continuous evolution of
cesses are often increasingly complex, diversified material sciences, microfabrication techniques
and measurement tools, altogether with the intro- rats: a developmental study. J Physiol 429:47–62.
https://doi.org/10.1113/jphysiol.1990.sp018243
duction of artificial intelligence-based data pro- Canadas RF, de Oliveira JMA, Marques AMP, dos Reis
cessing, allow the anticipation that the in vitro RLG (2017) Multi-chambers bioreactor, methods and
models will outperform the animal ones for drug uses. PCT/IB2015/057210
testing, personalized medicine, and diagnostics, Canadas RF, Ren T, Tocchio A et al (2018) Tunable aniso-
tropic networks for 3-D oriented neural tissue models.
including systemic effects, in the upcoming Biomaterials 181:402–414. https://doi.org/10.1016/j.
decades. biomaterials.2018.07.055
Cao X, Ashfaq R, Cheng F et al (2019) A tumor-on-a-
Acknowledgments The authors would like to acknowl- chip system with bioprinted blood and lymphatic ves-
edge the financial support provided by the Portuguese sel pair. Adv Funct Mater 29:1807173. https://doi.
Foundation for Science and Technology (FCT) through org/10.1002/adfm.201807173
the projects B-FABULUS (PTDC/BBB-ECT/2690/2014) Carvalho MR, Barata D, Teixeira LM et al (2019)
and 3BioMeD (FCT/4773/4/5/2017/S). This work has Colorectal tumor-on-a-chip system: A 3D tool for
been funded under the FCT doctoral programme for precision onco-nanomedicine. Sci Adv 5:eaaw1317.
Advanced Therapies for Health (PATH) and supported by https://doi.org/10.1126/sciadv.aaw1317
the Frontiers of technology for theranostics of cancer, Castells M, Thibault B, Delord J-P, Couderc B
metabolic and neurodegenerative diseases (FROnTHERA) (2012) Implication of tumor microenvironment in
Structured Project NORTE-01-0145-FEDER-000023, Chemoresistance: tumor-associated stromal cells
supported by Norte of Portugal Regional Operational protect tumor cells from cell death. Int J Mol Sci
Programme (NORTE 2020), under the PORTUGAL 2020 13:9545–9571. https://doi.org/10.3390/ijms13089545
Partnership Agreement, through the European Social Chen Z, He S, Zilberberg J, Lee W (2019) Pumpless
Fund. Raphaël F. Canadas thanks to FCT for the Postdoc platform for high-throughput dynamic multicellular
fellowship in the scope of the Project PTDC/EMD- culture and chemosensitivity evaluation. Lab Chip
EMD/29139/2017 funded by FEDER program. The FCT 19:254–261. https://doi.org/10.1039/C8LC00872H
distinction attributed to J. Miguel Oliveira (IF/01285/2015) Chng KR, Tay ASL, Li C et al (2016) Whole metage-
under the Investigator FCT program is also greatly nome profiling reveals skin microbiome-depen-
acknowledged. dent susceptibility to atopic dermatitis flare.
Nat Microbiol 1:16106. https://doi.org/10.1038/
nmicrobiol.2016.106
Chung M, Ahn J, Son K et al (2017) Biomimetic model
References of tumor microenvironment on microfluidic plat-
form. Adv Healthc Mater 6:1700196. https://doi.
Angeloni V, Contessi N, De Marco C et al (2017) org/10.1002/adhm.201700196
Polyurethane foam scaffold as in vitro model for Cooney CL (1983) Bioreactors: design and operation.
breast cancer bone metastasis. Acta Biomater 63:306– Science (80-) 219:728–733. https://doi.org/10.1126/
316. https://doi.org/10.1016/j.actbio.2017.09.017 science.219.4585.728
Ashley MS, Lauren O, Lillia H et al (2018) Profiling da Costa FHB, Navone NM, Farach-Carson MC, Carson
patient-specific glioblastoma drug response in vitro DD (2018) Abstract 3799: Heparan sulfate modifiers
using complex 3D microtumors. Cancer Res 78:1121 and signaling between tumor-associated macrophages
Bauer J, Staudacher JJ, Mancinelli G et al (2018) and cancer-associated fibroblasts in a biomimetic
Increased stiffness of the tumor microenvironment in hydrogel model of bone-metastatic prostate cancer.
colon cancer leads to an increase in Activin and meta- In: Immunology. American Association for Cancer
static potential. Cancer Res 78:177 LP – 177 Research. Chicago, pp 3799–3799
Boussommier-Calleja A, Atiyas Y, Haase K et al (2019) Daniel SG, Ball CL, Besselsen DG et al (2017) Functional
The effects of monocytes on tumor cell extrava- changes in the gut microbiome contribute to trans-
sation in a 3D vascularized microfluidic model. forming growth factor β-deficient colon cancer. mSys-
Biomaterials 198:180–193. https://doi.org/10.1016/j. tems 2. https://doi.org/10.1128/mSystems.00065-17
biomaterials.2018.03.005 Däster S, Amatruda N, Calabrese D et al (2017) Induction
Bratton M, Byrne CE, Rahman SM et al (2019) Using an of hypoxia and necrosis in multicellular tumor
in vitro fluidics approach to model the evolution of spheroids is associated with resistance to chemother-
metastatic breast cancer reveals shear stress as a possi- apy treatment. Oncotarget 8. https://doi.org/10.18632/
ble driver of genomic instability and somatic mutation. oncotarget.13857
FASEB J 33:lb175–lb175. https://doi.org/10.1096/ Demetriou CA, Vineis P (2015) Carcinogenicity of ambi-
fasebj.2019.33.1_supplement.lb175 ent air pollution: use of biomarkers, lessons learnt and
Butt AM, Jones HC, Abbott NJ (1990) Electrical resis- future directions. J Thorac Dis 7:67–95. https://doi.
tance across the blood-brain barrier in anaesthetized org/10.3978/j.issn.2072-1439.2014.12.31
Desrochers T, Shuford S, Mattingly C et al (2015) Abstract Bioeng 43:573–581. https://doi.org/10.1002/

LB-282: ex vivo 3d drug response profiling of primary bit.260430706
human ovarian cancer differentiates treatment-naive Hollier B, Harkin D, Leavesley D, Upton Z (2005)
and relapsed patients and molecular subtypes. In: Responses of keratinocytes to substrate-bound
Molecular and cellular biology. American Association vitronectin:growth factor complexes. Exp Cell
for Cancer Research. Philadelphia, p LB-282-LB-282 Res 305:221–232. https://doi.org/10.1016/j.
DiMasi JA, Grabowski HG, Hansen RW (2016) yexcr.2005.01.004
Innovation in the pharmaceutical industry: new esti- Houghton AN (1987) Phenotypic heterogeneity of mel-
mates of R&D costs. J Health Econ 47:20–33. https:// anoma. Relation to the differentiation program of
doi.org/10.1016/j.jhealeco.2016.01.012 melanoma cells. J Exp Med 165:812–829. https://doi.
Engbring JA, Kleinman HK (2003) The basement mem- org/10.1084/jem.165.3.812
brane matrix in malignancy. J Pathol 200:465–470. Huleihel L, Hussey GS, Naranjo JD et al (2016) Matrix-
https://doi.org/10.1002/path.1396 bound nanovesicles within ECM bioscaffolds. Sci Adv
Enokida Y, Tsuruno Y, Okubo K, Yamaoka Y, Takahashi E 2:e1600502. https://doi.org/10.1126/sciadv.1600502
(2017) Directional migration of MDA-MB-231 cells Hyler AR, Baudoin NC, Brown MS et al (2018) Fluid
under O2/pH gradients. In: Halpern H, LaManna J, shear stress impacts ovarian cancer cell viability, sub-
Harrison D, Epel B (eds) Oxygen transport to tissue cellular organization, and promotes genomic instabil-
XXXIX. Advances in experimental medicine and biol- ity. PLoS One 13:e0194170. https://doi.org/10.1371/
ogy, vol 977. Springer, Cham journal.pone.0194170
Gopalakrishnan V, Spencer CN, Nezi L et al (2018) Jalili-Firoozinezhad S, Prantil-Baun R, Jiang A et al
Gut microbiome modulates response to anti–PD-1 (2018) Modeling radiation injury-induced cell death
immunotherapy in melanoma patients. Science and countermeasure drug responses in a human
(80-) 359:97–103. https://doi.org/10.1126/science. Gut-on-a-Chip. Cell Death Dis 9:223. https://doi.
aan4236 org/10.1038/s41419-018-0304-8
Gritsenko P, Leenders W, Friedl P (2017) Recapitulating Jiménez-Torres JA, Virumbrales-Muñoz M, Sung KE
in vivo-like plasticity of glioma cell invasion along et al (2019) Patient-specific organotypic blood
blood vessels and in astrocyte-rich stroma. Histochem vessels as an in vitro model for anti-angiogenic
Cell Biol 148:395–406. https://doi.org/10.1007/ drug response testing in renal cell carcinoma.
s00418-017-1604-2 EBioMedicine 42:408–419. https://doi.org/10.1016/j.
Guo Q, Millard M, Lotstein A, et al Complex, patient- ebiom.2019.03.026
derived, multi-cell type, 3d models of breast cancer for Joyce JA (2005) Therapeutic targeting of the tumor micro-
personalized prediction of therapeutic response. http:// environment. Cancer Cell 7:513–520. https://doi.
kiyatec.com/wp-content/uploads/2019/03/kiyatec_ org/10.1016/j.ccr.2005.05.024
poster_Breast-Model-v2.pdf Juin A, Spence HJ, Martin KJ et al (2018) Chemotaxis in
Guzman A, Sánchez Alemany V, Nguyen Y et al (2017) pacreatic ductal adenocarcinoma metastasis: an unex-
A novel 3D in vitro metastasis model elucidates dif- pected role of NWASP in maintaining self-generated
ferential invasive strategies during and after breaching gradients and LPA receptor recycling. SSRN Electron
basement membrane. Biomaterials 115:19–29. https:// J. https://doi.org/10.2139/ssrn.3275292
doi.org/10.1016/j.biomaterials.2016.11.014 Jung H-Y, Fattet L, Yang J (2015) Molecular pathways:
Hao Q, Dong BR, Wu T (2015) Probiotics for prevent- linking tumor microenvironment to epithelial-
ing acute upper respiratory tract infections. Cochrane mesenchymal transition in metastasis. Clin Cancer
Database Syst Rev. https://doi.org/10.1002/14651858. Res 21:962–968. https://doi.org/10.1158/1078-0432.
CD006895.pub3 CCR-13-3173
Heggannavar GB, Hiremath CG, Achari DD et al (2018) Jung H-R, Kang HM, Ryu J-W et al (2017) Cell spheroids
Development of doxorubicin-loaded magnetic silica– with enhanced aggressiveness to mimic human liver
Pluronic F-127 Nanocarriers conjugated with transfer- Cancer in vitro and in vivo. Sci Rep 7:10499. https://
rin for treating glioblastoma across the blood–brain doi.org/10.1038/s41598-017-10828-7
barrier using an in vitro model. ACS Omega 3:8017– Katt ME, Placone AL, Wong AD et al (2016) In vitro tumor
8026. https://doi.org/10.1021/acsomega.8b00152 models: advantages, disadvantages, variables, and
Heinrich MA, Bansal R, Lammers T et al (2019) selecting the right platform. Front Bioeng Biotechnol
3D-bioprinted mini-brain: a glioblastoma model to 4. https://doi.org/10.3389/fbioe.2016.00012
study cellular interactions and therapeutics. Adv Mater Kessenbrock K, Plaks V, Werb Z (2010) Matrix metal-
31:1806590. https://doi.org/10.1002/adma.201806590 loproteinases: regulators of the tumor microenvi-
Herter S, Morra L, Schlenker R et al (2017) A novel ronment. Cell 141:52–67. https://doi.org/10.1016/j.
three-dimensional heterotypic spheroid model for the cell.2010.03.015
assessment of the activity of cancer immunotherapy Klimkiewicz K, Weglarczyk K, Collet G et al (2017) A
agents. Cancer Immunol Immunother 66:129–140. 3D model of tumour angiogenic microenvironment to
https://doi.org/10.1007/s00262-016-1927-1 monitor hypoxia effects on cell interactions and can-
Hillsley MV, Frangos JA (1994) Review: bone tissue engi- cer stem cell selection. Cancer Lett 396:10–20. https://
neering: the role of interstitial fluid flow. Biotechnol doi.org/10.1016/j.canlet.2017.03.006
Ko P, Kim D, You E et al (2016) Extracellular matrix Model Mech 11:dmm031435. https://doi.org/10.1242/
rigidity-dependent Sphingosine-1-phosphate secretion dmm.031435
regulates metastatic cancer cell invasion and adhesion. Mokhtar N, Wong K, Raja A, Ali R (2019) IDDF2019-
Sci Rep 6:21564. https://doi.org/10.1038/srep21564 ABS- 0227 manipulation of Gut microbiota
Koh I, Cha J, Park J et al (2018) The mode and dynam- in vitro model of colorectal cancer: strong
ics of glioblastoma cell invasion into a decellularized adherence ability of Lactobacillus Rhamnosus.
tissue-derived dimensional tumor model. Sci Rep Gut 68:A23 LP-A23. https://doi.org/10.1136/
8:1–12. https://doi.org/10.1038/s41598-018-22681-3 gutjnl-2019-IDDFAbstracts.43
Koikawa K, Ohuchida K, Ando Y et al (2018) Basement Mow VC, Wang CC, Hung CT (1999) The extracellular
membrane destruction by pancreatic stellate cells leads matrix, interstitial fluid and ions as a mechanical sig-
to local invasion in pancreatic ductal adenocarcinoma. nal transducer in articular cartilage. Osteoarthr Cartil
Cancer Lett 425:65–77. https://doi.org/10.1016/j. 7:41–58. https://doi.org/10.1053/joca.1998.0161
canlet.2018.03.031 Mrázek J, Mekadim C, Kučerová P et al (2019)
Kripke ML (1981) Immunologic mechanisms in UV Melanoma-related changes in skin microbiome.
radiation carcinogenesis. Adv Cancer Res 34:69–106. Folia Microbiol (Praha) 64:435–442. https://doi.
Elsevier, Frederick, Maryland org/10.1007/s12223-018-00670-3
Lanz HL, Saleh A, Kramer B et al (2017) Therapy response Müller I, Kulms D (2018) A 3D Organotypic mela-
testing of breast cancer in a 3D high-throughput per- noma spheroid skin model. J Vis Exp. https://doi.
fused microfluidic platform. BMC Cancer 17:709. org/10.3791/57500
https://doi.org/10.1186/s12885-017-3709-3 Ogmundsdottir H, Kowal N, Cherek P, Olafsdottir E (2018)
Lee K, Silva EA, Mooney DJ (2011) Growth factor PO-420 mechanisms of action of anti- proliferative
delivery-based tissue engineering: general approaches lichen compound protolichesterinic acid. In: PO-416
and a review of recent developments. J R Soc Interface Cytotoxicity of Ru (II) and Ru (III) salen complexes
8:153–170. https://doi.org/10.1098/rsif.2010.0223 against breast and colorectal cancer cell lines. BMJ
Li S-S, Ip CKM, Tang MYH et al (2017) Modeling Publishing Group Ltd, pp A187.3–A1A189
ovarian Cancer multicellular spheroid behavior in a Paszek MJ, Zahir N, Johnson KR et al (2005) Tensional
dynamic 3D peritoneal microdevice. J Vis Exp. https:// homeostasis and the malignant phenotype.
doi.org/10.3791/55337 Cancer Cell 8:241–254. https://doi.org/10.1016/j.
Lim M, Xia Y, Bettegowda C, Weller M (2018) Current ccr.2005.08.010
state of immunotherapy for glioblastoma. Nat Rev Pereira BA, Lister NL, Hashimoto K et al (2019) Tissue
Clin Oncol 15:422–442. https://doi.org/10.1038/ engineered human prostate microtissues reveal key
s41571-018-0003-5 role of mast cell-derived tryptase in potentiating
Liou Y-R, Torng W, Kao Y-C et al (2014) Substrate stiff- cancer-associated fibroblast (CAF)-induced morpho-
ness regulates filopodial activities in lung cancer cells. metric transition in vitro. Biomaterials 197:72–85.
PLoS One 9:e89767. https://doi.org/10.1371/journal. https://doi.org/10.1016/j.biomaterials.2018.12.030
pone.0089767 Piccoli M, D’Angelo E, Crotti S et al (2018) Decellularized
Lu H, Forbes RA, Verma A (2002) Hypoxia-inducible colorectal cancer matrix as bioactive microenviron-
factor 1 activation by aerobic glycolysis implicates ment for in vitro 3D cancer research. J Cell Physiol
the Warburg effect in carcinogenesis. J Biol Chem 233:5937–5948. https://doi.org/10.1002/jcp.26403
277:23111–23115. https://doi.org/10.1074/jbc. Puls TJ, Tan X, Whittington CF, Voytik-Harbin SL (2017)
M202487200 3D collagen fibrillar microstructure guides pancreatic
Ma C, Han M, Heinrich B, et al (2018) Abstract A02: cancer cell phenotype and serves as a critical design
Gut microbiome controls growth of liver tumors. In: parameter for phenotypic models of EMT. PLoS
Cancer and the microbiome. American Association for One 12:e0188870. https://doi.org/10.1371/journal.
Cancer Research. Boston, pp A02–A02 pone.0188870
Magdeldin T, López-Dávila V, Pape J et al (2017) Queiroz K, Habani Y, Lanz H, et al (2018) PO-449 3d
Engineering a vascularised 3D in vitro model of cancer glioma-on-a-chip models for personalised medi-
progression. Sci Rep 7:44045. https://doi.org/10.1038/ cine in organoplates®. In: Poster presentation: trans-
srep44045 lational research. BMJ Publishing Group Ltd, p
Mak AF (1986) The apparent viscoelastic behavior A408.1–A4A408
of articular cartilage—the contributions from the Racordon D, Valdivia A, Mingo G et al (2017) Structural
intrinsic matrix viscoelasticity and interstitial fluid and functional identification of vasculogenic mim-
flows. J Biomech Eng 108:123–130. https://doi. icry in vitro. Sci Rep 7:6985. https://doi.org/10.1038/
org/10.1115/1.3138591 s41598-017-07622-w
Mak IW, Evaniew N, Ghert M (2014) Lost in translation: Réquilé M, Gonzàlez Alvarez DO, Delanaud S et al (2018)
animal models and clinical trials in cancer treatment. Use of a combination of in vitro models to investigate
Am J Transl Res 6:114–118 the impact of chlorpyrifos and inulin on the intesti-
Marques-Torrejon MA, Gangoso E, Pollard SM (2018) nal microbiota and the permeability of the intestinal
Modelling glioblastoma tumour-host cell interactions mucosa. Environ Sci Pollut Res 25:22529–22540.
using adult brain organotypic slice co-culture. Dis https://doi.org/10.1007/s11356-018-2332-4
Rice AJ, Cortes E, Lachowski D et al (2017) Matrix stiff- Triantafillu UL, Park S, Klaassen NL et al (2017) Fluid
ness induces epithelial–mesenchymal transition and shear stress induces cancer stem cell-like phenotype
promotes chemoresistance in pancreatic cancer cells. in MCF7 breast cancer cell line without inducing epi-
Oncogene 6:e352–e352. https://doi.org/10.1038/ thelial to mesenchymal transition. Int J Oncol 50:993–
oncsis.2017.54 1001. https://doi.org/10.3892/ijo.2017.3865
Ridky TW, Chow JM, Wong DJ, Khavari PA (2010) Tsuruno Y, Okubo K, Fujiwara T et al (2018) An
Invasive three-dimensional organotypic neoplasia in vitro model for determining tumor cell migra-
from multiple normal human epithelia. Nat Med tion under metabolic gradients. Adv Exp Med Biol
16:1450–1455. https://doi.org/10.1038/nm.2265 1072:201–205
Roudsari LC, Jeffs SE, West JL (2018) Lung adenocar- Twardowski P, Wong JYC, Pal SK et al (2019) Randomized
cinoma cell responses in a 3D in vitro tumor angio- phase II trial of sipuleucel-T immunotherapy preceded
genesis model correlate with metastatic capacity. ACS by sensitizing radiation therapy and sipuleucel-T alone
Biomater Sci Eng 4:368–377. https://doi.org/10.1021/ in patients with metastatic castrate resistant prostate
acsbiomaterials.7b00011 cancer. Cancer Treat Res Commun 19:100116. https://
Sánchez-Rodríguez C, Cruces KP, Riestra Ayora J et al doi.org/10.1016/j.ctarc.2018.100116
(2017) BCG immune activation reduces growth and Usui T, Sakurai M, Umata K et al (2018) Preparation of
angiogenesis in an in vitro model of head and neck human primary colon tissue-derived organoid using
squamous cell carcinoma. Vaccine 35:6395–6403. air liquid interface culture. In: Current protocols in
https://doi.org/10.1016/j.vaccine.2017.10.008 toxicology. Wiley, Hoboken, pp 22.6.1–22.6.7
Sefidgar M, Soltani M (2014) Interstitial flow in can- van Duinen V, Zhu D, Ramakers C et al (2019) Perfused
cerous tissue: effect of considering remodeled cap- 3D angiogenic sprouting in a high-throughput in vitro
illary network. J Tissue Sci Eng 01. https://doi. platform. Angiogenesis 22:157–165. https://doi.
org/10.4172/2157-7552.S4-003 org/10.1007/s10456-018-9647-0
Sharma VP, Williams J, Leung E, et al (2017) Abstract Van Duuren BL, Goldschmidt BM (1976) Cocarcinogenic
1883: Fibrillar 1D tumor microenvironment is the and tumor-promoting agents in tobacco
key driver associating high-speed tumor cell motil- Carcinogenesis. J Natl Cancer Inst 56:1237–1242.
ity with nuclear shape in breast cancer. In: Tumor https://doi.org/10.1093/jnci/56.6.1237
biology. American Association for Cancer Research, Vermeulen MA, Slaets L, Cardoso F et al (2017)
pp 1883–1883 Pathological characterisation of male breast can-
Shi Z, Fultz RS, Engevik MA et al (2019) Distinct roles of cer: results of the EORTC 10085/TBCRC/BIG/
histamine H1- and H2-receptor signaling pathways in NABCG International Male Breast Cancer Program.
inflammation-associated colonic tumorigenesis. Am Eur J Cancer 82:219–227. https://doi.org/10.1016/j.
J Physiol Liver Physiol 316:G205–G216. https://doi. ejca.2017.01.034
org/10.1152/ajpgi.00212.2018 Villenave R, Wales SQ, Hamkins-Indik T et al (2017)
Shin W, Hinojosa CD, Ingber DE, Kim HJ (2019) Human Human gut-on-A-Chip supports polarized infection of
intestinal morphogenesis controlled by transepithelial Coxsackie B1 virus in vitro. PLoS One 12:e0169412.
morphogen gradient and flow-dependent physical cues https://doi.org/10.1371/journal.pone.0169412
in a microengineered Gut-on-a-Chip. iScience 15:391– Walens A, DiMarco AV, Lupo R et al (2019) CCL5 pro-
406. https://doi.org/10.1016/j.isci.2019.04.037 motes breast cancer recurrence through macrophage
Shuford S, Wilhelm C, Rayner M et al (2019) Prospective recruitment in residual tumors. elife 8. https://doi.
validation of an ex vivo, patient-derived 3D spher- org/10.7554/eLife.43653
oid model for response predictions in newly diag- Warburg O (1956) On the origin of cancer cells. Science
nosed ovarian Cancer. Sci Rep 9:11153. https://doi. 123:309–314
org/10.1038/s41598-019-47578-7 Xu H, Li Z, Yu Y et al (2016a) A dynamic in vivo-like
Sottoriva A, Spiteri I, Piccirillo SGM et al (2013) organotypic blood-brain barrier model to probe met-
Intratumor heterogeneity in human glioblastoma astatic brain tumors. Sci Rep 6:36670. https://doi.
reflects cancer evolutionary dynamics. Proc Natl org/10.1038/srep36670
Acad Sci 110:4009–4014. https://doi.org/10.1073/ Xu Z, Li E, Guo Z et al (2016b) Design and construc-
pnas.1219747110 tion of a multi-organ microfluidic chip mimicking the
Suresh M, Mattheolabakis G, Singh A, Amiji M (2018) in vivo microenvironment of lung cancer metastasis.
In vitro model of inflammatory, hypoxia, and cancer ACS Appl Mater Interfaces 8:25840–25847. https://
stem cell signaling in pancreatic cancer using hetero- doi.org/10.1021/acsami.6b08746
cellular 3-dimensional spheroids. bioRxiv 454397. Yang M, Nelson R, Ros A (2016) Toward analysis of pro-
https://doi.org/10.1101/454397 teins in single cells: a quantitative approach employing
Tanaka HY, Kitahara K, Sasaki N et al (2019) Pancreatic isobaric tags with MALDI mass spectrometry realized
stellate cells derived from human pancreatic cancer with a microfluidic platform. Anal Chem 88:6672–
demonstrate aberrant SPARC-dependent ECM remod- 6679. https://doi.org/10.1021/acs.analchem.5b03419
eling in 3D engineered fibrotic tissue of clinically rel- Yue X, Nguyen TD, Zellmer V et al (2018) Stromal
evant thickness. Biomaterials 192:355–367. https:// cell-laden 3D hydrogel microwell arrays as tumor
doi.org/10.1016/j.biomaterials.2018.11.023 microenvironment model for studying stiff-

ness dependent stromal cell-cancer interactions. model linking microbiome alteration and immune dys-
Biomaterials 170:37–48. https://doi.org/10.1016/j. regulation. Br J Dermatol 179:570–581. https://doi.
biomaterials.2018.04.001 org/10.1111/bjd.16734
Zhao J, Griffin M, Cai J et al (2016) Bioreactors for tissue Zhu Y, Clair G, Chrisler WB et al (2018b) Proteomic anal-
engineering: an update. Biochem Eng J 109:268–281. ysis of single mammalian cells enabled by microflu-
https://doi.org/10.1016/j.bej.2016.01.018 idic Nanodroplet sample preparation and ultrasensitive
Zhu TH, Zhu TR, Tran KA et al (2018a) Epithelial bar- NanoLC-MS. Angew Chem Int Ed 57:12370–12374.
rier dysfunctions in atopic dermatitis: a skin-gut-lung https://doi.org/10.1002/anie.201802843
Nanoparticles and Microfluidic
Devices in Cancer Research 10
F. Raquel Maia, Rui L. Reis,
Abstract can hamper the proper delivery of the nanopar-

Cancer is considered the disease of the cen- ticles into the desired site. In this reasoning, to
tury, which can be easily understood consider- get a step closer to the in vivo reality, it has
ing its increasing incidence worldwide. Over been proposed of the use of microfluidic
the last years, nanotechnology has been pre- devices. In fact, microfluidic devices can be
senting promising theranostic approaches to designed on-demand to exhibit complex struc-
tackle cancer, as the development of tures that mimic tissue/organ-level physiolog-
nanoparticle-based therapies. But, regardless ical architectures. Even so, despite
of the promising outcomes within in vitro set- microfluidic-based in vitro models do not
tings, its translation into the clinics has been compare with the reality and complexity of
delayed. One of the main reasons is the lack of the human body, the most complex systems
an appropriate in vitro model, capable to created up to now have been showing similar
mimic the true environment of the human results to in vivo animal models. Microfluidic
body, to test the designed nanoparticles. In devices have been proven to be a valuable tool
fact, most of in vitro models used for the vali- to accomplish more realistic tumour’s envi-
dation of nanoparticle-based therapies do not ronment. The recent advances in this field, and
address adequately the complex barriers that in particular, the ones enabling the rapid test
naturally occur in a tumor scenario, as such as of new therapies, and show great promise to
blood vessels, the interstitial fluid pressure or be translated to the clinics will be overviewed
the interactions with surrounding cells that herein.
F. R. Maia (*) · R. L. Reis · J. M. Oliveira Keywords

3B’s Research Group, I3Bs – Research Institute on Nanoparticles · Microfluidic in vitro models ·
Biomaterials, Biodegradables and Biomimetics, Cancer therapies · Drug delivery
Precision Medicine, Headquarters at University
e-mail: raquel.maia@i3bs.uminho.pt

https://doi.org/10.1007/978-3-030-36588-2_10
Highlights tools for the development of new and more effi-

cient cancer pharmacotherapies. Additionally,
• New nanoparticle-based therapies have been nanoparticles present some therapeutic features
failing before reaching the clinic due to the that were not possible to obtain with the present
lack of a proper in vitro model capable of therapeutic approaches. In fact, nanoparticles can
mimic human body reality and efficiently vali- transport different therapeutic agents simultane-
date those therapies. ously, even the ones that present low water solu-
• The development of more and more complex bility, and can be designed to control the
microfluidic systems, capable to emulate therapeutic agents’ release over time. But, regard-
tumours’ environment have been showing less of all these advantages, the translation of
results similar with in vivo animal models. nanoparticles into the clinics have been delayed.
• Microfluidic-based 2D and 3D in vitro models One of the main reasons is the lack of an appro-
may help boost decrease the cost of develop- priate in vitro model, capable to mimic the true
ment and testing of nanomedicines. environment of the human body, to test the
designed nanoparticles. In one hand, most of the
models widely used nowadays (e.g., two dimen-
10.1 Introduction sional standard cultures or cell spheroid cultures
in static conditions) do not address adequately
Cancer is already designated the disease of the the complex barriers that naturally occur in a
century, which can be easily understood consid- tumor scenario, as such as blood vessels, the
ering its increasing incidence worldwide (Bray interstitial fluid pressure or the interactions with
et al. 2018). In fact, cancer cells have the capacity surrounding cells that can hamper the proper
to proliferate continuously resisting death, induce deliver of the nanoparticles into the desired site
angiogenesis, suffer constant mutations and (Ozcelikkale et al. 2017). On the other hand,
metastasize forming a second tumor (Fouad and since these same models do not replicate the true
Aanei 2017). Altogether, these features have been in vivo environment, the obtained results are not
hampering the efficacy of current theranostic accurate enough nor reliable. This has urged the
strategies and cancer has become a global scientists to develop new in vitro models to more
burden. accurately test the nanoparticles-based therapies,
Over the last years, nanotechnology has been as photodynamic therapy (Chudy et al. 2018;
presenting promising theranostic approaches to Wang et al. 2018).
tackle cancer, as the development of nanoparti- In this reasoning, to bridge the gap between
cles for the detection of circulating cancer cells the typically two-dimensional standard cultures
(Song et al. 2019) and for drug delivery (Khanna and the in vivo reality, it has been proposed of
et al. 2019). Nanoparticles present very interest- the use of microfluidic devices. Microfluidic
ingly characteristics. One of the most appealing devices can be designed on-demand to exhibit
is the possibility to be designed with different microchannels and/or microchambers that
sizes (in the range of 1–100 nm) and shapes, mimic tissue/organ-level physiological architec-
allowing to take advantage of the enhanced per- tures (Zhang et al. 2018; Xu et al. 2014).
meation and retention (EPR) effect for cancer Moreover, these devices require only a few
treatment approaches (Yuancheng et al. 2018). In microliters of volume, resulting in lower
this sense, nanoparticles can cross blood vessels, amounts of reagents and consequently lower
transporting therapeutic agents until the disease associated costs, which makes them very attrac-
site and diminishing systemic side effects. These tive models (Yun et al. 2013). In the special case
features associated with the possibility to modify of nanoparticle-based therapeutics, it allows
the nanoparticles’ surface with molecules of evaluating the nanoparticle targeted accumula-
interest, allowing a targeted therapy, easily makes tion, cytotoxicity, transport and uptake, result-
these nanostructures one of the most promising ing in more relevant data.
10 Nanoparticles and Microfluidic Devices in Cancer Research 163
Herein, we aim to discuss in depth the evalua- of designed nanoparticles (Tokarova et al. 2013;
tion of new nanoparticle-based therapies for can- Sarfati et al. 2011). For example, Sarfati et al.
cer research using microfluidic devices, (2011) used microfluidic devices to culture lung
highlighting the most relevant and recent exam- cancer cells in monolayer and test the cellular
ples. Two main models are overviewed, i.e. uptake of nanoparticles modified with specific
microfluidic-based two-dimensional (2D) in vitro peptides to target these cells, under a constant
models, where cells are cultured in a monolayer flow (Fig. 10.1). As control, the authors used
inside the device, and microfluidic-based three- nanoparticles modified with scramble peptides.
dimensional (3D) in vitro models, where cells are Within this setting, the authors observed that the
cultured as tumour spheroids or within matrices. flow within the microchannels prevented the
nanoparticle precipitation, in opposite to what
usually happened in 2D standard cultures.
10.2 Microfluidic-Based 2D in vitro Moreover, they detected that in the absence of
Models flow, cells internalized both nanoparticles,
nanoparticles with target peptide and scramble
Nanoparticle-based therapies have been widely peptide, as depicted in Fig. 10.1 E. But when they
studied to tackle cancer. In fact, the idea of deliver the nanoparticles under a flow rate, cells
enabling the delivery of high amounts of thera- internalized higher amounts of the nanoparticles
peutic agents directly into a cancer environment, modified with the target peptide (Figs. 10.1 C, D
without damaging any healthy tissue/cells has and E).
been very attractive to scientists (Han et al. 2016). Besides fluid flow influence, blood vessels’
Nevertheless, many drug delivery therapies failed wall can as well influence the effectiveness of
during clinical trials and only a few ends up new therapies, namely the endothelial cells can
approved by Food and Drug Administration hinder its success. In this sense, microfluidic
(FDA) and translated into the clinics. One of the devices have been used to study nanoparticles’
main reasons is the lack of a suitable model to interactions with endothelial cells. In fact, thera-
accurately predict nanoparticle-based therapies’ peutic nanoparticles are supposed to show no
performance once within the human body. In effect on healthy endothelial cells and only inter-
fact, despite the incredible value of the static 2D act with cancer cells (Kang et al. 2016; Khor
standard cultures during the first steps of thera- et al. 2018). In the work of Kang et al. (2016),
pies’ study, it cannot emulate the in vivo environ- they cultured endothelial cells inside a micro-
ment and the barriers normally present. For channel and observed that the shear stress pro-
example, the blood vessels’ wall, the blood flow moted by the fluid flow influenced the alignment
within the vascular environment that induces of actin stress fibers inside endothelial cells,
fluid shear stress, and the interstitial flow within which hampered the internalization of nanopar-
cancer’s environment. Altogether, these barriers ticles. Interestingly, in the work of Khor et al.
have been preventing the accumulation of the (2018) they also took advantage of the environ-
necessary amounts of therapeutic agents in situ. ment promoted by the microchannel that mimics
Within this reasoning, microfluidic devices have the blood vessels, to optimize nanoparticles
been proposed to develop a more physiological design. In this sense, they were able to tune the
relevant model and this way overcome the limita- surface of nanoparticles to control its interaction
tions observed within the 2D standard cultures. with endothelial cells.
In the special case of drug delivery systems, body In a different example, Bagley et al. (2015)
fluids are expected to interfere with the interac- also used the similarity of microchannels with
tion of drug-delivery nanoparticles and cells, and blood vessels to test the use of plasmonic
ultimately with the proper release of drugs. For nanoparticles as an improved drug delivery strat-
so, many scientists have been assessing the influ- egy. In fact, these nanoparticle’s can convert
ence of fluid flow into the cellular internalization near-infrared light into localized heat, perturbing
Fig. 10.1 Lung cancer cells uptake of targeted nanopar- target peptide (green); D) Cancer cells with DiL stain
ticles within a microfluidic device. (a) Scheme of a micro- (red) and internalized nanoparticles modified with scram-
fluidic device, composed of (1) 6 syringes, which are ble peptides (green); and E) Internalization of nanoparti-
connected by tubes to (2) microchannel and finally to the cles modified with target peptide (white column) or
(3) reservoir; (b) Monolayer of cells cultured inside the scramble peptide (black column) at different flow rates. (∗
microchannel; (c) Cancer cells with 1,1′-Dioctadecyl- represents p < 0.001). (Reprinted with permission from
3,3,3′,3’-Tetramethylindocarbocyanine Perchlorate (DiL) (Sarfati et al. 2011) Copyright © 2010 Elsevier Ltd. All
stain (red) and internalized nanoparticles modified with rights reserved)
the vasculature within the cancer environment, 2017; Zhai et al. 2017). For example, in the work
and allowing the successful penetration of the of Carvalho et al. (2017) the authors developed
therapeutic agents. In this study, the authors fluorescent-labeled nanoparticles to easily moni-
produced endothelialized channels inside the
tor different cancer cells’ fate. For this study, the
microfluidic device and subjected cells to several authors compared microfluidic devices with 2D
heating cycles to unravel the contribution of the standard settings. When tested the efficiency of
endothelium to the vascular thermotolerance in a internalization, they observed that different cells
cancer setting (Bagley et al. 2015). At the first presented different internalization rates.
cycle of high temperatures, the endothelial mono- Nevertheless, the internalization rate was higher
layer showed some disruption, allowing an in microfluidic devices than in 2D standard set-
increase of permeability as expected. But, upon a tings. Moreover, they observed that cells pre-
second cycle of treatment, the disruption sented higher sensitivity to nanoparticles’
decreased and less permeability was observed. internalization in microfluidic devices, resulting
With this study, it was clear that endothelial cells in higher amounts of dead cells. While in 2D con-
were able to adapt to the high temperatures, ditions less dead cells were observed. Since no
which can hamper the success of plasmonic therapeutic agents were added to these nanopar-
nanoparticles for therapies delivery. ticles, the authors hypothesized that the fluores-
Furthermore, microfluidic devices can be used cent label could be cytotoxic at high
to visualize nanoparticles’ internalization and its concentrations, once cell death was higher in
effect on cellular behavior, as cell death, making microfluidic devices where more nanoparticles
it a promising tool for the validation and optimi- were internalized. In fact, the number of nanopar-
zation of new nanoparticle-based therapies as ticles’ internalized is one important parameter
demonstrated with several works (Carvalho et al. since it can influence new strategies efficiency. It
was with this in mind that Zhai et al. (2017) asso- ment through the mimicry of the physiological
ciated surface-enhanced Raman spectroscopy conditions, it is possible to obtain more reliable
(SERS) to microfluidic devices to validate newly data concerning the validation of new
designed nanoparticles. In this sense, they were nanoparticle-based therapies. In this reasoning,
able to observe the exact amount of nanoparticles microfluidic devices have been playing important
that cells internalized and consequently associate roles in the mimicry of the in vivo conditions.
with its therapeutic effect on cancer cells. Besides the features mentioned in the previous
Ultimately, microfluidic devices present point, microfluidic devices allow the 3D culture
another very appealing feature that is the possi- of cancer and healthy cells, in the form of spher-
bility to perform simultaneously the evaluation of oids or within a matrix (Yang et al. 2015;
multiple therapeutic nanoparticles as done by Zuchowska et al. 2017; Kalinowska et al. 2019).
Mytxelena-Iribarren et al. (2019). They devel- For example, Yang et al. (2015) co-cultured
oped a microfluidic device with multiple micro- breast cancer cells with adipose-derived stem
channels, where they culture osteosarcoma cells. cells within a microfluidic for the assessment of
In each microchannel, they submitted cells to the efficiency of photodynamic therapy (PDT)
nanoparticles loaded with different therapeutic associated with therapeutic agents. PDT is a very
agents, expediting the selection of more efficient attractive therapy since it presents fewer side
therapies. By substituting the cell line tested by effects than the typical therapies, radiotherapy
patients’ cells it is easy to realize that it would be and chemotherapy. In fact, it uses photosensitiz-
a powerful tool to develop personalized ing agents that upon the exposition to visible
therapies. light radiation produces reactive oxygen species,
Nevertheless, despite the importance of these which kills only the adjacent cells. To test this
microfluidic-based 2D in vitro models for the therapy, the authors developed a 3D model within
validation of new therapies, it lacks the three a microfluidic device. For that, cells were allowed
dimensionality found in vivo. That limitation has to proliferate and produce their extracellular
urged the development of new and improved matrix, creating a natural 3D environment inside
three-dimensional (3D) models on a microfluidic the microfluidic device. Then, gold nanoparticles
device. were perfused along the microchannels and its
effects associated with PDT treatment were
observed. The results showed that cancer cells
10.3 Microfluidic-Based 3D in vitro cultured in the developed 3D environment within
Models the microfluidic device were more resistant than
2D standard cultures. While in the 3D environ-
Currently, much attention has been paid to 3D, ment the treatment was efficient in only 50% of
especially in the development of new in vitro cells, in 2D environment, the treatment was 100%
models for the validation of new cancer thera- efficient. In a different study, the efficacy of PDT
pies. In the work of Carvalho et al. (2019b) it was was assessed on lung cancer cells’ spheroids cul-
clear the importance of three dimensional (3D) tured within a microfluidic device, as depicted in
features for the assessment of new therapies to Fig. 10.2 (Zuchowska et al. 2017). Upon 5 days
fight cancer. They studied the effect of new tar- of treatment, cancer cells showed higher resis-
geted nanoparticles in 2D and 3D culture set- tance to PDT than the observed in the previous
tings. Besides the decrease of nanoparticles’ example. This difference can be due to the fact
internalization in the 3D cultures, as compared to that in the first example cells were cultured in a
2D standard cultures, they also observed that the 3D environment less dense than the environment
therapeutic effect decreased by 50% in 3D cul- presented by the spheroids. In fact, in the case of
tures after 72 h in culture. spheroids it is possible that the photosensitizing
Additionally, it is widely accepted that by agents did not reach to all cells, as well as the
generating a more specific cancer microenviron- oxygen, essential for the success of this therapy.
Fig. 10.2 Microfluidic device for the assessment of PDT a microwell of a 384-well plate. (d) Microfluidic device
efficacy on lung spheroids. (a) Picture of the microfluidic incorporated in holder with the same dimensions as a 384-
device. (b) Schematic representation of spheroid forma- well plate. (Adapted with permission from (Zuchowska
tion inside the microfluidic device along the time. (c) et al. 2017) © 2017 Elsevier B.V. All rights reserved)
Picture of microwells dimensions, which are the same as
Noteworthy, this assay was performed on a to test. Additionally, the microfluidic device
microfluidic device incorporated in a holder with allows for a more realistic setting, since it allows
the same dimensions as a 384-well plate, which to mimic the distribution of gold nanoparticles
allowed to easily monitor the culture using a along the blood vessels until the cancer tissue. To
spectrofluorimetric microplate reader (Fig. 10.2c test this therapy, the authors cultured cancer
and d). Altogether, the developed model seems to spheroids or healthy spheroids inside the device
be a promising tool for new nanoparticle-based and gold nanoparticles were perfused, mimicking
therapies assessment and optimization. the in vivo reality. Upon irradiation, the authors
More recently, spheroids cultured inside a successfully observed cells death and reduction
microfluidic device were used to test other thera- of the spheroid, demonstrated that PTT may be
pies as photothermal therapy (PTT), an alterna- more powerful for tumour treatment than PDT.
tive to PDT (Kalinowska et al. 2019). In fact, Despite the relevance of previously described
PTT allows for a deeper tissue penetration in models, it is widely accepted that for the success
opposite to PDT, due to the use of near-infrared of new nanoparticle-based therapies, it is essen-
radiation, which makes spheroids a good model tial to fully understand cancer environment and
mimic all the building blocks. In this sense, the infiltration of nanoparticles within the tumour tis-
system should not only mimic the tumour or the sue, drugs were able to penetrate further, as dem-
blood vessels, but should allow their co-culture, onstrated by the increase on the cytotoxicity
envisioning the screening of new drugs in a more along the throughout the tumour.
realistic setting (DeWitt and Rylander 2018; Despite the complexity obtained by introduc-
Carvalho et al. 2019a). Dewitt and Rylander pre- ing new building blocks of the tumour environ-
pared a protocol where they demonstrated how ment, as such as blood vessels, tumour and
they prepared an optically clear model to enable healthy tissues, altogether within a microfluidic
the visualization of nanoparticles transport along device, it is also possible to control the features
an endothelialized channel and along with the of such building blocks, as done by
tumour tissue (DeWitt and Rylander 2018). Prabhakarpandian et al. and later on by Vu et al.
Additionally, this model also allowed for the (Prabhakarpandian et al. 2015; Vu et al. 2019). In
post-culture analysis of the uptake of nanoparti- the first example they cultured an endothelial cell
cles by tumour and endothelial cells, which is a lined leaky capillary vessel with a 3D spheroid
plus for the validation of new nanoparticles- tumour (Prabhakarpandian et al. 2015). This co-
based therapeutic strategies. Carvalho et al. culture emulated the typical cervical cancer envi-
(2019a) as well developed a model that allowed ronment normally composed by leaky vessels
the visualization of the effect of new nanoparti- within a solid tumour. This setting showed simi-
cles-based treatments and the posterior cell lar results as compared with in vivo data, validat-
recovery for gene expression evaluation, envi- ing the capability of using this device to predict
sioning its use for personalized treatment devel- the real efficiency of new drugs in vivo. In the
opment. With the intent to better mimic the second example they developed a model com-
tumour environment, Chen et al. (2018) devel- posed of a tumour tissue within a leaky microvas-
oped a breast tumour model in a microfluidic cular network (Vu et al. 2019). For that they
device where they incorporated the mimicry of activated a regulator of the endothelial junction’s
blood vessels wall, the matrix that surrounds the formation, which promoted the reorganization of
tumour and the tumour itself. The device was the cytoskeleton and disrupted the junctions
incorporated within a well-plate for the real-time resulting in a leak endothelium. In this sense they
monitoring of nanoparticles-based therapies effi- were able to study the delivery of new
ciency. In this sense, the authors were able to nanoparticle-based therapies with different sizes
observe with high spatiotemporal resolution the and surface chemistries throughout the blood
transport of the drug delivery system across the vessels and its accumulation within the tumour
microvascular wall, then the matrix, and finally tissue in a more realistic manner. In fact, they
its infiltration within the spheroid. Interesting, conclude that it is not only the size of the parti-
with this system, the authors observed different cles that influences the success of the delivery
infiltration rates into spheroids composed of dif- process but also the surface chemistry. Depending
ferent cancer cells’ types. In a different example, on the surface chemistry, the nanoparticles inter-
Jarvis et al. (2017) used, as well, a microchannel acted with the endothelial cells hindering its
colonized with endothelial cells, but instead of interaction with the tumour cells. Interestingly,
studying just tumour tissues, they also studied Tang et al. (2017) showed that by close emulate
healthy tissues in a 3D arrangement, to better the in vivo tumour scenario, i.e. culture a vascular
mimic the tumour environment and more reliably network under shear stress and in contact with
test new nanoparticle-based therapeutic systems. highly metastatic tumours is enough to simulate
Additionally, they introduced the nanoparticles the enhanced permeability and retention (EPR)
under physiologically relevant shear stress. With effect. For that, they created the vascular network
this setting, they were able to test different by using an image database of in vivo networks
nanoparticles for the delivery of super hydropho- and comprised a tumour tissue (Fig. 10.3).
bic drugs and realized that despite the superficial Additionally, they included shear stress in the
Fig. 10.3 Vascular
network comprising a
tumour tissue within a
microfluidic device. (a)
Low levels of
VE-cadherin (white
arrows) were observed
under low shear stress;
(b) VE-cadherin is
inhibited in the presence
of highly metastatic
tumour cells; and (c) No
differences in the
VE-cadherin levels were
observed in the presence
of less metastatic cells.
(Adapted with
permission from (Tang
et al. 2017) under a
Creative Commons
Attribution 4.0
International License)
model, which showed to be essential for the for- to naturally include in a model, as such as differ-
mation of endothelial cellular junctions, shown ent cut-off pore size, interstitial fluid pressures
by the low levels of vascular endothelial cadherin and tissue microstructures as done by Kwak et al.
(VE-cadherin, the molecule responsible for the (2014). For that study, the authors developed a
adhesion of endothelial cells) under low shear model composed of microchannels that mimic
stress (Fig. 10.3 A). With this realistic model, not only, the blood vessels, where the nanoparti-
they observed that the extravasation of nanopar- cles were introduced at physiological relevant
ticles along the endothelial wall to tumour tissue velocity and pressure, but also, the lymphatic
increased in the presence of metastatic tumour vessels. In fact, despite the importance of blood
cells. Additionally, the weakening of the endo- vessels for the delivery of therapies, lymphatic
thelial cell barrier function showed similar per- vessels are also key factors for the success of
meability values as the ones found in vivo for those therapies since lymphatic vessels should
similar tumour animal models. Noteworthy, present reduced drainage to limit the clearance of
when studying less metastatic tumour cells, no such delivery therapies from the tumour. In this
permeability was observed. reasoning, the authors designed a model where
Finally, microfluidic devices can as well be blood and lymphatic vessels were surrounding a
designed with different features that are difficult tumour tissue composed of cancer cells within a
collagen matrix. In this sense, they were able to microenvironment, which influences the effi-
mimic the pore size of tumour blood vessels, the ciency of new therapies, as such as vascular net-
high interstitial fluid pressure and the dense works or permeable blood vessels. These same
extracellular matrix of the tumour and the high building blocks are not only important for the
cell density of tumour tissues. At the end they validation of new therapies, but also playing
were able to obtain a model where it was possible important role in the optimization of the novel
to change all the previously mentioned parame- nanoparticle-based therapies. In fact, the delivery
ters but maintain the interactions between the of nanoparticles-based therapies is a process that
vessels and the tumour, revealing the effect of the comprises several steps. First nanoparticles are
physiological conditions of the therapeutic injected intravenously, then they should cross
nanoparticles’ transport. through the blood vessels endothelium and pen-
etrate into the tumour tissue, finally, they should
be internalized by the tumour cells without dam-
10.4 Conclusions and Future aging the healthy cells in the way. But all these
Perspectives steps comprise several problems, as such as rapid
clearance by the lymphatic vessels, reduced pen-
The study of new therapies to tackle tumour have etration, and unexpected organ uptake and accu-
been the main focus of researchers worldwide. mulation. Each one of those steps has different
One of the most studied approaches is the use of requisites and need nanoparticle-based therapies
nanoparticle-base therapies. Nevertheless, most with features that enable to fulfil those requisites
of the developed therapies fail along the process for optimal transport. Additionally, tumour tis-
of validation. For so, only a few nanoparticle- sues differ from patient to patient and within the
based therapies reach clinics. Part of the reason same tissue it is possible to find a heterogeneous
lays behind the lack of models capable to predict population of tumour cells, which as well impairs
truthfully the effect of those therapies once inside the successful effect of new therapies. Even so,
the human body. the most complex systems created by now within
Currently, microfluidic devices have emerged the microfluidic devices have been showing simi-
as privileged devices since they can be designed lar results to in vivo animal models, which is a
on demand. Microfluidic devices can include step forward the better in vitro model for
appealing features as mimicking of blood ves- nanoparticle- based therapies validation.
sels, by the endothelialisation of the microchan- However, they do not yet compare with the real-
nels and addition of shear stress, or by the ity of the human body because of the most com-
possibility to mimic the tumour itself, by culture plex system created by now do not replicate the
of tumour cells within the microchambers. interactions of tumour tissues and distant organs.
Nevertheless, many assays, performed within this In fact, it is difficult to emulate the signals path-
setting, are still two dimensional which has been ways between tumours and distant organs, and
shown to underestimate the effect of nanoparti- even harder to emulate the signals between
cles. To overcome this limitation, several efforts tumours and the endocrine and immune system.
have been made to increase the complexity and For so, it looks like that the connection of tumour
include three-dimensionality within the model. and distant organs within one device or several
Many times by the inclusion of spheroids tumour devices may be the next step in the development
tissues or by the culture of tumour cells within a of more physiologically relevant models in vitro.
matrix, which can be in contact with endothelial Overall, microfluidic devices seem to be the set-
cells. ting for the production of more and more realistic
Despite the superiority demonstrated as com- tumour’s environment, enabling the rapid test of
pared with 2D settings, it was still missing many new therapies, increasing the probability to be
of the building blocks that compose the tumour translated into the clinical scenario.
Acknowledgments The authors thank the funds obtained protocols. Springer, New York, pp 159–178. https://
through the FROnTHERA (NORTE-01-0145- doi.org/10.1007/978-1-4939-8661-3_12
FEDER-0000232) project supported by Norte Portugal Fouad YA, Aanei C (2017) Revisiting the hallmarks of
Regional Operational Program (NORTE 2020), under the cancer. Am J Cancer Res 7(5):1016–1036
PORTUGAL 2020 Partnership Agreement, through the Han B, Qu C, Park K, Konieczny SF, Korc M (2016)
European Regional Development Fund (ERDF). FRM Recapitulation of complex transport and action of
acknowledges Portuguese Foundation for Science and drugs at the tumor microenvironment using tumor-
Technology (FCT) for her contract under the Transitional microenvironment-on-chip. Cancer Lett 380(1):319–
Rule DL 57/2016 (CTTI-57/18-I3BS(5)), JMO thanks 329. https://doi.org/10.1016/j.canlet.2015.12.003
FCT for the distinction attributed under the Investigator Jarvis M, Arnold M, Ott J, Pant K, Prabhakarpandian B,
FCT program (IF/01285/2015). Mitragotri S (2017) Microfluidic co-culture devices
to assess penetration of nanoparticles into cancer cell
mass. Bioeng Trans Med 2(3):268–277. https://doi.
org/10.1002/btm2.10079
References Kalinowska D, Grabowska-Jadach I, Liwinska M, Drozd
M, Pietrzak M, Dybko A, Brzozka Z (2019) Studies
Bagley AF, Scherz-Shouval R, Galie PA, Zhang AQ, on effectiveness of PTT on 3D tumor model under
Wyckoff J, Whitesell L, Chen CS, Lindquist S, microfluidic conditions using aptamer-modified
Bhatia SN (2015) Endothelial Thermotolerance nanoshells. Biosens Bioelectron 126:214–221. https://
impairs nanoparticle transport in tumors. Cancer Res doi.org/10.1016/j.bios.2018.10.069
75(16):3255–3267. https://doi.org/10.1158/0008- Kang T, Park C, Lee B-J (2016) Investigation of biomi-
5472.can-15-0325 metic shear stress on cellular uptake and mechanism
Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre of polystyrene nanoparticles in various cancer cell
LA, Jemal A (2018) Global cancer statistics 2018: lines. Arch Pharm Res 39(12):1663–1670. https://doi.
GLOBOCAN estimates of incidence and mortal- org/10.1007/s12272-016-0847-0
ity worldwide for 36 cancers in 185 countries. CA Khanna V, Kalscheuer S, Kirtane A, Zhang W, Panyam
Cancer J Clin 68(6):394–424. https://doi.org/10.3322/ J (2019) Perlecan-targeted nanoparticles for drug
caac.21492 delivery to triple-negative breast cancer. Fut Drug
Carvalho MR, Maia FR, Silva-Correia J, Costa BM, Reis Discovery 1(1):FDD8. https://doi.org/10.4155/
RL, Oliveira JM (2017) A semiautomated microflu- fdd-2019-0005
idic platform for real-time investigation of nanopar- Khor SY, Vu MN, Pilkington EH, Johnston APR,
ticles' cellular uptake and cancer cells' tracking. Whittaker MR, Quinn JF, Truong NP, Davis TP
Nanomedicine (Lond) 12(6):581–596. https://doi. (2018) Elucidating the influences of size, surface
org/10.2217/nnm-2016-0344 chemistry, and dynamic flow on cellular association
Carvalho MR, Barata D, Teixeira LM, Giselbrecht S, of nanoparticles made by polymerization-induced
Reis RL, Oliveira JM, Truckenmüller R, Habibovic self-assembly. Small (Weinheim an der Bergstrasse,
P (2019a) Colorectal tumor-on-a-chip system: A Germany) 14(34):1801702. https://doi.org/10.1002/
3D tool for precision onco-nanomedicine. Sci Adv smll.201801702
5(5):eaaw1317. https://doi.org/10.1126/sciadv. Kwak B, Ozcelikkale A, Shin CS, Park K, Han B (2014)
aaw1317 Simulation of complex transport of nanoparticles
Carvalho MR, Carvalho CR, Maia FR, Caballero D, around a tumor using tumor-microenvironment-on-
Kundu SC, Reis RL, Oliveira JM (2019b) Peptide- chip. J Control Release 194:157–167. https://doi.
modified dendrimer nanoparticles for targeted therapy org/10.1016/j.jconrel.2014.08.027
of colorectal Cancer. Adv Therapeutics 0(0):1900132. Mitxelena-Iribarren O, Zabalo J, Arana S, Mujika M
https://doi.org/10.1002/adtp.201900132 (2019) Improved microfluidic platform for simulta-
Chen Y, Gao D, Wang Y, Lin S, Jiang Y (2018) A neous multiple drug screening towards personalized
novel 3D breast-cancer-on-chip platform for thera- treatment. Biosens Bioelectron 123:237–243. https://
peutic evaluation of drug delivery systems. Anal doi.org/10.1016/j.bios.2018.09.001
Chim Acta 1036:97–106. https://doi.org/10.1016/j. Ozcelikkale A, Moon HR, Linnes M, Han B (2017) In
aca.2018.06.038 vitro microfluidic models of tumor microenvironment
Chudy M, Tokarska K, Jastrzębska E, Bułka M, Drozdek to screen transport of drugs and nanoparticles. Wiley
S, Lamch Ł, Wilk KA, Brzózka Z (2018) Lab-on-a- Interdiscip Rev Nanomed Nanobiotechnol 9(5):e1460.
chip systems for photodynamic therapy investiga- https://doi.org/10.1002/wnan.1460
tions. Biosens Bioelectron 101:37–51. https://doi. Prabhakarpandian B, Shen MC, Nichols JB, Garson
org/10.1016/j.bios.2017.10.013 CJ, Mills IR, Matar MM, Fewell JG, Pant K (2015)
DeWitt MR, Rylander MN (2018) Tunable collagen Synthetic tumor networks for screening drug delivery
microfluidic platform to study nanoparticle transport systems. J Control Release 201:49–55. https://doi.
in the tumor microenvironment. In: Sirianni RW, org/10.1016/j.jconrel.2015.01.018
Behkam B (eds) Targeted drug delivery: methods and Sarfati G, Dvir T, Elkabets M, Apte RN, Cohen S (2011)
Targeting of polymeric nanoparticles to lung metas-
tases by surface-attachment of YIGSR peptide from evaluation. Biotechnol Adv 32(7):1256–1268. https://
laminin. Biomaterials 32(1):152–161. https://doi. doi.org/10.1016/j.biotechadv.2014.07.009
org/10.1016/j.biomaterials.2010.09.014 Yang Y, Yang X, Zou J, Jia C, Hu Y, Du H, Wang H (2015)
Song Y, Shi Y, Huang M, Wang W, Wang Y, Cheng J, Lei Evaluation of photodynamic therapy efficiency using
Z, Zhu Z, Yang C (2019) Bioinspired engineering of an in vitro three-dimensional microfluidic breast can-
a multivalent aptamer-functionalized Nanointerface to cer tissue model. Lab Chip 15(3):735–744. https://doi.
enhance the capture and release of circulating tumor org/10.1039/c4lc01065e
cells. Angew Chem Int Ed 58(8):2236–2240. https:// Yuancheng L, Kai Z, Xiao L, Yu Shrike Z (2018) Blood-
doi.org/10.1002/anie.201809337 vessel-on-a-Chip platforms for evaluating nanopar-
TangY, Soroush F, Sheffield JB, Wang B, Prabhakarpandian ticle drug delivery systems. Curr Drug Metab
B, Kiani MF (2017) A biomimetic microfluidic 19(2):100–109. https://doi.org/10.2174/13892002186
tumor microenvironment platform mimicking the 66170925114636
EPR effect for rapid screening of drug delivery sys- Yun H, Kim K, Lee WG (2013) Cell manipulation in
tems. Sci Rep 7(1):9359. https://doi.org/10.1038/ microfluidics. Biofabrication 5(2):022001. https://doi.
s41598-017-09815-9 org/10.1088/1758-5082/5/2/022001
Tokarova V, Pittermannova A, Kral V, Rezacova P, Zhai Z, Zhang F, Chen X, Zhong J, Liu G, Tian Y, Huang
Stepanek F (2013) Feasibility and constraints of Q (2017) Uptake of silver nanoparticles by DHA-
particle targeting using the antigen-antibody inter- treated cancer cells examined by surface-enhanced
action. Nanoscale 5(23):11490–11498. https://doi. Raman spectroscopy in a microfluidic chip. Lab
org/10.1039/c3nr04340a Chip 17(7):1306–1313. https://doi.org/10.1039/
Vu MN, Rajasekhar P, Poole DP, Khor SY, Truong NP, C7LC00053G
Nowell CJ, Quinn JF, Whittaker M, Veldhuis NA, Zhang H, Zhu Y, Shen Y (2018) Microfluidics for
Davis TP (2019) Rapid assessment of nanoparticle cancer nanomedicine: from fabrication to evalu-
extravasation in a microfluidic tumor model. ACS Appl ation. Small (Weinheim an der Bergstrasse,
Nano Mat 2(4):1844–1856. https://doi.org/10.1021/ Germany) 14(28):1800360. https://doi.org/10.1002/
acsanm.8b02056 smll.201800360
Wang Y, Cuzzucoli F, Escobar A, Lu S, Liang L, Wang Zuchowska A, Jastrzebska E, Chudy M, Dybko A,
S (2018) Tumor-on-a-chip platforms for assessing Brzozka Z (2017) 3D lung spheroid cultures for
nanoparticle-based cancer therapy. Nanotechnology evaluation of photodynamic therapy (PDT) proce-
29(33):332001. https://doi.org/10.1088/1361-6528/ dures in microfluidic lab-on-a-Chip system. Anal
aac7a4 Chim Acta 990:110–120. https://doi.org/10.1016/j.
Xu X, Farach-Carson MC, Jia X (2014) Three-dimensional aca.2017.07.009
in vitro tumor models for cancer research and drug
Index
A Cen, L., 69
Agarwal, A., 130 Central nervous system (CNS), v, 87–93
Ahmed, T.A.E., 69 Chaikof, E.L., 69
Alan, F.-J., 69 Chan, E.C., 69
Andrzejewska, A., 87–93 Chen, L., 111
Angiogenesis, v, 3, 50–51, 97–115, 141, 145, 148, Chen Wen, L.K., 53
149, 162 Chen, Y., 167
Aycock, R.S., 69 Chen, Z., 152
Cho, D.-W., 10
Choi, Y.J., 71
B Cho, S.-W., 70
Baker, B.M., 108 Collins, M.N., 69
Banihashemi, M., 69 Coppeta, J.R., 53
Bhattacharjee, N., 76 Correia, J. S.-, 97–115
Bhise, N.S., 10 Correlo, V.M., 1–12
Bidault, L., 70 Costa, J.B., 8
Bioengineering, 28, 32 Costa, L., 97–115
Biomaterials, v, 2, 3, 15–22, 30, 59, 60, 65–81, 98, 100, Cox, S.C., 7
108, 115, 121–132 Cozzolino, A.M., 70, 75
Bioprinting, 2, 7–10, 12, 17, 30, 32, 35, 37, 60, Cui, H., 7
98, 100, 142 Cunniffe, G.M., 8
Bioreactors, 2, 31, 66, 68, 72, 78, 138, 151–154
Birkinshaw, C., 69
Bischel, L.L., 131 D
Bishi, D.K., 74 Da Silva Morais, A., 65–81
Boga, J.C., 7 Decellularized matrices, 70, 75–76, 81, 144
Bone tissue engineering, 2, 7, 8, 12 Del Amo, C., 105
Bruns, H., 69 Dewitt, M.R., 167
Buchanan, C.F., 113 Disease modelling, 52
Byambaa, B., 7 Dvir-Ginzberg, M., 69
Drug delivery, 4, 36, 53, 55, 74, 80, 90, 123, 148, 162,
163, 167
C Drug development, 28, 30, 38, 47, 56, 58, 68, 76, 122,
Caballero, D., 43–61 130, 131, 148
Caddeo, S., 70 Drug discovery, 3, 32, 38, 44–48, 52–54, 58, 60,
Canadas, R.F., 137–155 121–132, 149, v
Cancer, 3, 38, 39, 43–61, 66, 98, 99, 102, 108, 111–114, Drug screening, 29, 35, 46, 51, 53, 81, 109, 112, 114,
122, 123, 130, 137–155, 161–169 115, 123, 127, 131, 152, 153
Cancer therapies, 150, 165
Cao, X., 152
Carvalho, M.R., 114, 121–132, 152, 164, E
165, 167 Endothelization, 100

https://doi.org/10.1007/978-3-030-36588-2
174 Index
F Kundu, S.C., 43–61

Fadul, H., 69 Kwak, B., 168
Fan, J., 69
Faramarzi, N., 7
Fernandes, D.C., 137–155 L
Lee, H., 112
Lee, J.S., 70
G Lee, K.Y., 69
Galie, P.A., 106 Li, I., 9
Gao, D., 124 Li, S.-S., 152
Garg, K., 69 Liver models, 56, 57, 65–84
Gasperini, L., 15–22 Low, L.A., 27–39
Gong, H., 72 Lumelsky, N., 27–39
Gramowski, A., 127 Lundberg, M.S., 27–39
Lyu, S., 69
H
Hammond, J.S., 70 M
Han, S.S., 70 Maia, F.R., 1–12, 161–169
Hao, S., 3 Makadia, H.K., 70
He, H.Y., 7 Malikmammadov, E., 70
Hemshekhar, M., 69 Mari, C.E., 69
Huh, D., 127 Mariod, A.A., 69
Hydrogels, 8–10, 17, 18, 37, 46, 47, 50, 60, 70–76, 78, Marques, A.P., 15–22
99–101, 103, 108, 122–125, 131, 148, 153 Maschmeyer, I., 53
Microfabrication, 33, 68, 76, 98–100, 115, 123, 151,
152, 154
I Microfluidic chambers, 89–93
In vitro models, v, 1–12, 32, 44, 45, 72, 73, 75, 105, 113, Microfluidic channels, 4, 20, 31, 34, 51. 52, 68, 76, 78,
115, 127, 138–140, 144, 146, 148, 150, 151, 155, 90, 114, 123, 127, 129
162, 163, 169 Microfluidic devices, 1–12, 22, 28, 37, 48, 59, 78, 79, 88,
In vitro tumor models, 45–47, 148 92, 93, 101, 103, 105–109, 111, 114, 124–127,
129–131, 149–152, 161–169
Microfluidic in vitro models, 2, 10–12
J Microfluidics, v, 15–22, 35, 65–81, 97–115,
Jain, A., 129 121–132
Jain, E., 70 Microphysiological systems, 32–34
Jang, K.J., 129 Microvascular models, 99–102, 109, 111, 114, 115
Janowski, M., 87–93 Miller, J.S., 113
Jarvis, M., 167 Miller, P.G., 54
Jeon, J.S., 130 Miranda-Nieves, D., 69
Jusoh, N., 126 Mironov Anton, V., 70
Mitchell, M.C., 126
Mooney, D.J., 69
K Munn, L.L., 106
Kamei, K.-i., 53 Mytxelena-Iribarren, O., 165
Kang, H.-W., 9
Kanta, J., 69
Kazemnejad, S., 70 N
Khademhosseini, A., 127 Nakao, Y., 129
Khalaj, Z., 69 Nanoparticles, 21, 33, 114, 161–169
Kim, C., 114 Naresh, K., 69
Kim, J., 109 Natural polymers, 68, 69, 123
Kimura, H., 130
Kim, Y., 69
Knight, T.A., 70 O
Kumar, A., 70 Occhetta, P., 127
Kundu, B., 69 Oleaga, C., 53
Index 175
Oliveira, J.M., 1–12, 65–81, 97–115, 121–132, 137–155, Three-dimensional (3D) models, 47, v, 10, 47, 76, 101,
161–169 141, 142, 145, 148, 149, 151, 153, 165
Organ-on-a-chip (OoC), v, 10, 12, 27–39, 44, 48–53, 58, Torisawa, Y.-S., 130
60, 78, 80, 81, 85, 88, 115, 123, 127–132 Török, E., 70
Truckenmuller, R., 121–132
Tsamandouras, N., 53
P Tumour angiogenesis, 101, 112–114
Patient-on-a-chip, 43–61
Payne, R.G., 70
Perez, R.A., 70 U
Personalized medicine, 12, 29, 34, 43–61, 142, 155 Untereker, D., 69
Phan, D.T.T., 114 Utpal, B., 69
Polydimethylsiloxane (PDMS), 4, 37, 49, 59, 60, 70, 78, Uygun, B.E., 126
92, 100, 125, 128, 129, 131
Powers, M.J., 70
V
Van Duinen, V., 105
R Vasanthan, K.S., 69
Reis, L.R., 1–12, 15–22, 43–61, 65–81, 97–115, Vernetti, L., 54
121–132, 137–155, 161–169
Rinaudo, M., 69
Rylander, M.N., 167 W
Wang, B., 70
Wang, X.F., 8
S Watanabe, M., 69
Selimovic, S., 27–39 Whitesides, G.M., 70
Semnani, D., 70, 74 Wong, A.H.-H., 131
Seo, S.-J., 69
Seyer, J.M., 69
Shah Mohammadi, M., 70 X
Shim, J.-H., 74 Xia, L., 72
Shin, Y., 103 Xu, Z., 53
Shuler, M.L., 54
Sia, S.K., 70
Siegel, S.J., 70 Y
Skardal, A., 53, 70 Yang, J., 69, 72
Song, J.W., 106 Yu, F., 78
Sprouting, 3, 51, 98, 99, 105–110, 112–114, 126 Yu, Y.-Q., 74
Stevens, K.R., 69, 70
Sung, J.H., 131
Sutherland, M., 27–39 Z
Zanotelli, M.R., 101
Zeigerer, A., 69
T Zeinali, S., 114
Tagle, D.A., 27–39 Zhai, Z., 165
Tang, Y., 167 Zhang, Y., 74, 124
Teotia, R.S., 71 Zhang, Y.S., 37, 53
Theberge, A.B., 101 Zheng, Y., 113
3D interfaces models, 137–155 Zhou, X., 126

2020 Book Biomaterials-AndMicrofluidics

Uploaded by

Copyright:

Available Formats

2020 Book Biomaterials-AndMicrofluidics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2020 Book Biomaterials-AndMicrofluidics

Uploaded by

Copyright:

Available Formats

Advances in Experimental Medicine and Biology 1230

ISSN 0065-2598 ISSN 2214-8019 (electronic)

© Springer Nature Switzerland AG 2020

In recent years, we have assisted a new impetus in tissue engineering and

Guimarães, Portugal J. M. Oliveira

1 Microfluidic Devices and Three Dimensional-Printing

F. Raquel Maia, Rui L. Reis, Vitor M. Correlo,

Abstract bone angiogenesis. More recently, 3D printing

F. R. Maia (*) · R. L. Reis · V. M. Correlo Keywords

© Springer Nature Switzerland AG 2020 1

Highlights have been pursued to understand the behavior of

relies on the extrusion, layer-by-layer, of a high 1.3.2 Inkjet Modelling

Biotechnol 34:312–319. https://doi.org/10.1038/ osteoblasts during adulthood. Stem Cells (Dayton,

Abstract silk fibroin are emphasized deu to their rele-

© Springer Nature Switzerland AG 2020 15

2.1 Biomaterials ing, the formation of the gel is reversible, but in

that is soluble in water and forms highly viscous 2.2 Processability

References Couto DS, Hong Z, Mano JF (2009) Development of bio-

Nanofluid 14(5):767–774. https://doi.org/10.1007/ Nguyen N-T, Wu Z (2005) Micromixers—a review. J

Abstract peutic efficacy in vitro; and (3) disease model-

Fig. 3.3 – A TissUse

3.4  ow Engineering Advances

David Caballero, Rui L. Reis, and Subhas C. Kundu

Abstract of a new generation of highly predictable pre-

Keywords tumor models and therapies. In this regard, recent

TME is mimicked employing hydrogel-based endothelial permeability, and intravasation rate

Table 4.1 Recent works on multi-organ and patient-on-a-chip devices

cal multi-organ complexity of individual patients (CTCs), DNA/RNA, exosomes or antibodies

bubble formation or sterility maintenance. The 4.7.2 Biological Challenges

integration of several and personalized organs alternative physico-chemical properties, such as

modular multi-organ-chips. PLoS One 10:e0139587. Lab Chip 14:1715–1724. https://doi.org/10.1039/

Abstract several tissue engineering challenges are associ-

© Springer Nature Switzerland AG 2020 65

­ latforms, present a high potential in the cre-

Table 5.1 Biomaterials applied for fabrication of in vitro liver models

pared to the traditional-one including a reduced liver-on-a-chip constructs were usually

through the body (Ding et al. 2010). So, studies

accepted therapeutic option for patients with References

caprolactone)/collagen nanofibrous scaffolds. J Mater Collaborators, G. M. a. C. o. D (2015) Global, regional,

© Springer Nature Switzerland AG 2020 87

6.1 Introduction technology allowed creating very complex

Fig. 6.1 The schematic overview of photolithography technique

References Gorelik M, Orukari I, Wang J et al (2012) Use of MR cell

Abstract esis has been a subject of major interest among

© Springer Nature Switzerland AG 2020 97

Highlights scaffolds before implantation, thus can help

vessels by proteolytic degradation mediated by showing morphological alterations of the ECs

7.3 Concluding Remarks driving angiogenesis overcoming the major

Abstract nificant interest in the drug discovery and

© Springer Nature Switzerland AG 2020 121

for real-world pharmaceutical discovery. Thus, References

Abstract used for the development of pharmaceutical

© Springer Nature Switzerland AG 2020 137

9.2.3 Biochemical known and well-established tumor’s hallmarks is

9.3.3 Concluding Remarks and expensive, which creates important barriers

Desrochers T, Shuford S, Mattingly C et al (2015) Abstract Bioeng 43:573–581. https://doi.org/10.1002/

Abstract can hamper the proper delivery of the nanopar-

F. R. Maia (*) · R. L. Reis · J. M. Oliveira Keywords

© Springer Nature Switzerland AG 2020 161

Guimarães, Portugal J. M. Oliveira

1 Microfluidic Devices and Three Dimensional-Printing

relies on the extrusion, layer-by-layer, of a high 1.3.2 Inkjet Modelling

2.1 Biomaterials ing, the formation of the gel is reversible, but in

that is soluble in water and forms highly viscous 2.2 Processability

3.4 ow Engineering Advances

bubble formation or sterility maintenance. The 4.7.2 Biological Challenges

latforms, present a high potential in the cre-

6.1 Introduction technology allowed creating very complex

7.3 Concluding Remarks driving angiogenesis overcoming the major

9.2.3 Biochemical known and well-established tumor’s hallmarks is

9.3.3 Concluding Remarks and expensive, which creates important barriers