Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 120 (2015) 26 30

EUROSENSORS 2015

Advanced electrochemical in vitro detection of superoxide radicals

with fully integrated microsensor system
H. Flamm*, J. Kieninger, A. Weltin, G.A. Urban
Department of Microsystems Engineering IMTEK, University of Freiburg, 79110 Freiburg, Germany

Abstract
An electrochemical sensing system for in vitro measurement of superoxide radical is presented. The approach of the system is
based on a fully integrated sensor setup in conventional tissue culture containment, utilizing novel direct oxidation of ROS on
polymer covered gold microelectrode. The concept was shown to provide stable and sensitive measurement procedure without
the need for enzymatic recognition elements. Sensor performance was investigated from artificial enzymatic superoxide
production, with a sensitivity of 2247 AM-1m2 and outstanding high selectivity for main interfering substances. Online
monitoring of superoxide release from human breast carcinoma cell line T-47D was successfully demonstrated.
2015
2015The
TheAuthors.
Authors.
Published
by Elsevier
Ltd.

Published
by Elsevier
Ltd. This
is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the organizing committee of EUROSENSORS 2015.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of EUROSENSORS 2015
Keywords: Superoxide; electrochemical sensor; in vitro; direct oxidation; enzymeless; cancer cell culture

1. Introduction
The in vitro detection of short lived reactive oxygen species superoxide is mainly based on spectrophotometric or
electron spin resonance techniques [1], which include extensive chemical and physical cell culture treatment.
Further the monitoring of dynamic radical changes in cellular microenvironment is a challenge using cumulative
techniques. Electrochemical sensors could provide reliable methods for the detection of superoxide in cell culture.
State of the art electrochemical superoxide sensors are mainly based on the coupling of biological recognition
elements cytochrome c (Cyt c) [2, 3] or superoxide dismutase (SOD) [4, 5] to gold electrodes. Sensor sensitivity is
often not sufficient for the detection of superoxide radical release from cell culture. Further, the selectivity and long

* Corresponding author. Tel.: +49-761-203-7266; fax: +49-761-203-7262.

E-mail address: hubert.flamm@imtek.uni-freiburg.de

1877-7058 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of EUROSENSORS 2015

doi:10.1016/j.proeng.2015.08.558

H. Flamm et al. / Procedia Engineering 120 (2015) 26 30

term stability of these sensor approaches is not guaranteed, due to possible degradation of the functional
biocomponents. Thus, electrode fouling by unspecific adsorption hinders their usage for extended cell culture
monitoring. In this work we present a measurement concept, which is based on a direct oxidation of the analyte
superoxide on a gold microelectrode, as illustrated in figure 1. A polymeric protection cover was introduced to avoid
unspecific binding of molecules from complex cell culture media and to enhance sensor selectivity towards the main
interferent hydrogen peroxide by its weight cut-off ability. A full integration of the sensor chip into conventional
tissue culture containment allows convenient cell cultivation on chip over days and subsequent online monitoring of
superoxide release upon drug stimulation without the need for a cell transfer.

Fig. 1. Illustration of the electrochemical sensor setup for direct superoxide oxidation and transferred electron upon electrode polarization.

2. Measurements and Results

2.1. Chip fabrication
Sensor chips were fabricated as fully integrated 3-electrode setup on chip, comprising gold working electrodes,
platinum counter and Ag/AgCl reference electrode as illustrated in figure 2 left. Chip fabrication was done as
standard clean room process on 100 mm diameter Pyrex borosilicate glass wafer using platinum metal structuring
by physical vapor deposition, lift-off processing and chip isolation with silicon nitride/oxide layer. Integration of
silver/silver chloride reference electrode was achieved on wafer-level by cathodic electrodeposition of silver from
Arguna S (Umicore) solution and subsequent partial conversion of the silver layer to silver chloride. The reference
was covered with poly-HEMA based long term stable hydrogel cover by automated dispensing robot. Gold working
electrodes were electroplated on the platinum electrodes as full-wafer process from Puramet (Doduco)
galvanization bath. For measurement in cell culture application, the chip was mounted in standard 25 cm2 tissue
culture flask (BD Biosciences) by CNC machining of openings and permanent fixation of the chip with UV-curable
adhesive Loctite 3201 (Henkel). Prior to cell culture measurement, the flask was sterilized by UV irradiation and
the polymeric semipermeable membrane was applied by incubation of the sensor setup in linear branched
polyethylenimine (PEI) solution for 45 min. For detailed information about sensor fabrication steps, please refer to
authors previous work [6].

Fig. 2. Left: On-chip 3-electrode setup. Inset marks the used gold working electrode (golden circle), platinum counter electrode (white square)
and the on-chip Ag/AgCl reference (dark circle); Right: Integrated sensor chip in standard cell culture flask for long term cell cultivation in
system and superoxide monitoring.

27

28

H. Flamm et al. / Procedia Engineering 120 (2015) 26 30

2.2. Superoxide sensor calibration

Sensor calibration was done from artificial superoxide production by enzymatic conversion of hypoxanthine with
xanthine oxidase (XOD), as described by Ge and Lisdat [7]. Briefly, the enzyme XOD converts hypoxanthine to uric
acid and superoxide and xanthine as intermediates with overall reaction rate k1. After formation, superoxide radicals
decompose spontaneously to hydrogen peroxide and molecular oxygen with rate k2. Thus, a short term steady state
concentration of superoxide can be calculated from the reaction rates k1, k2 and the enzyme activity of XOD,
according to Tammeveski et al. [8] with

k1
>XOD@
k2

>O @


2 SS

Reaction rate k1 was determined using cytochrome c method with an UV/vis spectrophotometer (Lambda 40,
Perkin Elmer), as described by Ge and Lisdat [7]. For the used xanthine oxidase, rate k1 was derived with
k1 = 2.55 s-1. Rate k2 is given by Behar et al. [9] with 2.3x10-5 M-1s-1. From these values and the used XOD activity,
the maximum peak concentration of produced superoxide was calculated. Figure 3 left shows amperometric
measurements of superoxide at 120 mV vs. on-chip Ag/AgCl reference electrode for different XOD activities in
100 M hypoxanthine, prepared in PBS solution, pH 7.4 which was supplemented with 100 M EDTA as chelating
agent to avoid metal ion catalyzed side reactions. It can be observed, that directly after mixing of hypoxanthine with
the enzyme XOD (Time = 0 min) superoxide is produced depending on enzyme activity. Signals decreased within
short time with different slopes before they drop back to baseline. The signal shape is in agreement with the
expected reaction kinetics, since the substrate hypoxanthine is consumed by reaction, which explains the signal
decrease and drop to baseline after total substrate consumption. Reaction times for inserted XOD activities in figure
3 left correlate with spectrophotometric measurement of uric acid formation (data not shown). From artificial
superoxide production, a calibration plot was derived in the range of 300 nM to 3.6 M and a linear sensor response
above 700 nM radical concentration with a sensitivity of 2247 AM-1m-2, as shown in figure 3 right.

700

Superoxide concentration:
300 nM
700 nM
850 nM

60
50

Current density in nAcm-2

Current density in nAcm-2

70

40
30
20
10
0

600
500
400
slope: 2247 AM-1m-2

300
200
100

-10

-20
0

10

15

20

Time in min

25

30

35

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Superoxide concentration in M

Fig. 3. Left: Amperometric sensor signals, measured at 120 mV vs. on-chip reference polarization on a PEI covered gold electrode, upon artificial
superoxide production from various XOD activities, resulting in short term steady state radical concentrations of 300 nM, 700 nM and 850 nM;
Right: Calibration plot for PEI covered gold electrode. R2 = 0.994, Standard error = 7.14 AM-1m-2, n = 3.

H. Flamm et al. / Procedia Engineering 120 (2015) 26 30

Main interfering substances during sensor calibration are xanthine, uric acid and hydrogen peroxide which are
formed always as coproducts from superoxide production and decomposition. The selectivity of direct superoxide
oxidation at 120 mV vs. on-chip Ag/AgCl was tested on PEI covered gold microelectrodes. Xanthine and uric acid
showed no signal above the background up to tested concentrations of 100 M and 1 mM, respectively. Hydrogen
peroxide is reduced on blank gold electrodes with a sensitivity of -3.9 AM-1m-2 which is dramatically decreased to
-0.6 AM-1m-2 after PEI membrane deposition. Further, the addition of hydrogen scavenger catalase during sensor
calibration did not change the signal height. Thus, with polymeric protection of the electrodes, the influence of
hydrogen peroxide can be neglected.
The stability of superoxide electrodes was tested by calibration of the sensor before and after 48 h incubation at
37 C in CO2-Independent Medium (Gibco), supplemented with 10% fetal bovine serum (FBS), 1% L-glutamine,
1% pen/strep and 0.2 units/ml insulin at which the sensor retained more than 80% of its initial sensitivity.
2.3. Cell culture measurement

Current density in nAcm-2

Superoxide release in activated cell culture was determined from human breast carcinoma cell line T-47D (ATCC
umber HTB-133) [10]. The cells were passaged from RPMI-1640 culture medium (Gibco) into CO2-Independent
Medium (Gibco) for measurement in ambient air condition. Both media were supplemented with 10% fetal bovine
serum (FBS), 1% L-glutamine, 1% pen/strep and 0.2 units/ml insulin. Prior to the measurement, the cells were
passaged into sensor equipped culture flask and cultivated for 48 h at 37 C in humidified atmosphere containing
5% CO2. Cell response on drug stimulation was measured at 37 C under normoxic conditions in a temperature
controlled heat cabinet. During all states of proliferation, the cells could be observed through the transparent glass
chip by inverted light microscopy. The measurement of superoxide radical release was conducted on sub-confluent
cell layer by polarization of the PEI covered working electrode at 120 mV vs. on-chip Ag/AgCl reference. After
reaching a baseline signal, cells were stimulated by addition of 5 M phorbol 12-myristate 13-acetate (PMA) from
1 mM stock solution prepared in DMSO. The amperometric sensor response is shown in figure 4. Sensor signal
shows a little response directly after PMA addition, which is also observed in control experiments without cells. The
peak might arise from DMSO, which is used for PMA stock solution preparation, as suggested by other authors
[11]. After 5 min of incubation the sensor signal increased from 25 nAcm-2 to 125 nAcm-2 within 13 minutes and
decreased afterwards, as shown in figure 4. To verify sensor selectivity, the enzyme superoxide dismutase (SOD,
750 units) was added into the cell culture flask. Superoxide dismutase specifically enhances superoxide
decomposition reaction, which could be observed by a signal drop back to baseline within a few seconds.

150
125

SOD 750 units

100
75

PMA 1M

50
25
0
50

60

70

80

90

Time in min
Fig. 4. Amperometric signal of superoxide sensor, upon the stimulation of T-47D cell culture with phorbol 12-myristate 13-acetate (PMA) and
superoxide scavenging by addition of superoxide dismutase (SOD).

29

30

H. Flamm et al. / Procedia Engineering 120 (2015) 26 30

3. Conclusion
In this study we demonstrate the successful application of fully integrated electrochemical sensor setup for the
measurement of superoxide release from stimulated cancer cell culture. The sensing principle with direct oxidation
of the analyte at polymer covered gold microelectrodes was verified and outstanding sensor performance with
enhanced sensitivity could be demonstrated. Selectivity of the superoxide sensor was achieved by proper choice of
operating potential region and the introduction of polymeric molecular weight cut-off membrane. Thus the long term
stability of the sensor could be enhanced by avoiding unspecific adsorption from complex measurement medium.
The enzymeless amperometric measurement principle allows cell cultivation directly on chip for time period of at
least 48 h. By system design, pericellular superoxide measurement can be conducted in situ directly at the place of
formation, which is in our opinion an inherent requirement for the reliable quantification of short lived radicals.
Thin-film based sensor chip fabrication using standard clean room processing provides high fabrication yield and
good reproducibility of the sensor batches. The use of transparent glass substrate for the superoxide sensor chips
allows cell inspection during all stages of experiment with inverted light microscopy and the integration of sensor
chips in conventional tissue culture containment provides a tool for convenient cell culture monitoring in the system
without the need for a cell transfer, which could reduce cellular response on extrinsic chemical and mechanical
stress.
Acknowledgements
This study was funded by the European Union Integrated Project of METOXIA (HEALTH-F2-2009-222741).
The authors are grateful to Erik Pettersen and Joe Sandvik, University of Oslo, for providing human breast cancer
cells.
References
[1] S.I. Dikalov, A.E. Dikalova, R.P. Mason, Noninvasive diagnostic tool for inflammation-induced oxidative stress using electron spin
resonance spectroscopy and an extracellular cyclic hydroxylamine, Arch. Biochem. Biophys. 402 (2002) 218226.
[2] J. M. Cooper, K. R. Greenough, C. J. McNeil, Direct electron transfer reactions between immobilized cytochrome c and modified gold
electrodes, J. Electroanal. Chem. 347/12 (1993) 267275.
[3] W. Scheller, W. Jin, E. Ehrentreich-Frster, B. Ge, F. Lisdat, R. Bttemeier, U. Wollenberger, F. W. Scheller, Cytochrome C Based
Superoxide Sensor for In Vivo Application, Electroanalysis 11/1011 (1999) 703706.
[4] L. Campanella, G. Favero, M. Tomassetti, Superoxide Dismutase Biosensors for Superoxide Radical Analysis, Anal. Lett. 32/13 (1999)
25592581.
[5] M. I. Song, F. F. Bier, F. W. Scheller, A method to detect superoxide radicals using Teflon membrane and superoxide dismutase,
Bioelectrochemistry Bioenerg. 38/2 (1995) 419422.
[6] H. Flamm, J. Kieninger, A. Weltin, G.A. Urban, Superoxide microsensor integrated into a Sensing Cell Culture Flask microsystem using
direct oxidation for cell culture application, Biosens. Bioelectron. 65 (2015) 354-359.
[7] B. Ge, F. Lisdat, Superoxide sensor based on cytochrome c immobilized on mixed-thiol SAM with a new calibration method, Anal. Chim.
Acta 454 (2002) 5364.
[8] K. Tammeveski, T.T. Tenno, A.A. Mashirin, E.W. Hillhouse, P. Manning, C.J. McNeil, Superoxide electrode based on covalently
immobilized cytochrome c: modelling studies, Free radical biology & medicine 25 (1998) 973978.
[9] D. Behar, G. Czapski, J. Rabani, L.M. Dorfman, H.A. Schwarz, The acid dissociation constant and decay kinetics of the perhydroxyl radical,
J. Phys. Chem 74 (1970) 32093213.
[10] I. Keydar, L. Chen, S. Karby, F.R. Weiss, J. Delarea, M. Radu, S. Chaitcik, H.J. Brenner, Establishment and characterization of a cell line of
human breast carcinoma origin, Europ J. Cancer 15 (1979) 659670.
[11] S.-C. Chang, N. Pereira-Rodrigues, J. R. Henderson, A. Cole, F. Bedioui, C. J. McNeil, An electrochemical sensor array system for the
direct, simultaneous in vitro monitoring of nitric oxide and superoxide production by cultured cells, Biosens. Bioelectron. 21/6 (2005) 917
922.