SHORT REPORT JBMR

Evidence of Staphylococcus aureus Deformation,

Proliferation, and Migration in Canaliculi of Live Cortical
Bone in Murine Models of Osteomyelitis
Karen L de Mesy Bentley,1,2,3 Ryan Trombetta,1,4 Kohei Nishitani,1 Sheila N Bello-Irizarry,1 Mark Ninomiya,1
Longze Zhang,1 Hung Li Chung,4 James L McGrath,4 John L Daiss,1 Hani A Awad,1,4 Stephen L Kates,1,5
and Edward M Schwarz1,2,3,4
1
University of Rochester School of Medicine and Dentistry, Rochester, NY, USA
2
Department of Pathology & Laboratory Medicine, University of Rochester School of Medicine and Dentistry, Rochester, NY, USA
3
Department of Orthopaedics, University of Rochester School of Medicine and Dentistry, Rochester, NY, USA
4
Department of Biomedical Engineering, University of Rochester School of Medicine and Dentistry, Rochester, NY, USA
5
Department of Orthopaedic Surgery, Virginia Commonwealth University Medical Center, Richmond, VA, USA

ABSTRACT
Although Staphylococcus aureus osteomyelitis is considered to be incurable, the major bacterial reservoir in live cortical bone has
remained unknown. In addition to biofilm bacteria on necrotic tissue and implants, studies have implicated intracellular infection of
osteoblasts and osteocytes as a mechanism of chronic osteomyelitis. Thus, we performed the first systematic transmission electron
microscopy (TEM) studies to formally define major reservoirs of S. aureus in chronically infected mouse (Balb/c J) long bone tissue.
Although rare, evidence of colonized osteoblasts was found. In contrast, we readily observed S. aureus within canaliculi of live cortical
bone, which existed as chains of individual cocci and submicron rod-shaped bacteria leading to biofilm formation in osteocyte lacunae.
As these observations do not conform to the expectations of S. aureus as non-motile cocci 1.0 to 1.5 mm in diameter, we also performed
immunoelectron microscopy (IEM) following in vivo BrdU labeling to assess the role of bacterial proliferation in canalicular invasion.
The results suggest that the deformed bacteria: (1) enter canaliculi via asymmetric binary fission; and (2) migrate toward osteocyte
lacunae via proliferation at the leading edge. Additional in vitro studies confirmed S. aureus migration through a 0.5-mm porous
membrane. Collectively, these findings define a novel mechanism of bone infection, and provide possible new insight as to why
S. aureus implant-related infections of bone tissue are so challenging to treat. © 2016 American Society for Bone and Mineral Research.

KEY WORDS: STAPHYLOCOCCUS AUREUS; OSTEOMYELITIS; CANALICULAR SYSTEM; CORTICAL BONE; ELECTRON MICROSCOPY; HAPTOTAXIS;
DUROTAXIS

Introduction However, to date there have been no systematic transmission

electron microscopy (TEM) studies to formally define the major
reservoirs of S. aureus in chronically infected live bone. To
S taphylococcus aureus accounts for the majority of chronic
osteomyelitis cases, and these bone infections remain a major
challenge in orthopedics.(1) Moreover, this condition is considered
address this we utilized murine models of tibial(4) and femoral(5)
osteomyelitis, and performed TEM studies on the behavior of
to be incurable because of biofilm-dwelling bacteria persisting prototypic methicillin-sensitive S. aureus (MSSA; UAMS-1),(10)
deep within cortical bone(2) and this condition costs billions of and the most common clinical strain of methicillin-resistant
dollars each year.(3) Animal models of implant-associated S. aureus S. aureus (MRSA; USA300 LAC).(11)
infections utilizing scanning electron microscopy (SEM) have
provided important insights on biofilm formation, maturation, and
accessory gene regulator-dependent bacterial emigration to Materials and Methods
perpetuate chronic infections.(4,5) Moreover, results from in vitro
S. aureus strains and mice
experiments,(6) animal model studies,(7,8) and clinical biopsies(9)
have demonstrated S. aureus infection of osteoblasts and Four S. aureus strains were used: UAMS-1,(10) UAMS-1 expressing
osteocytes, implicating intracellular persistence within bone cells the green fluorescent protein (GFPþ UAMS-1; provided by
as a mechanism of immune evasion and chronic infection. Dr. Steven Gill, University of Rochester, Rochester, NY), a

Received in original form August 19, 2016; revised form December 2, 2016; accepted December 5, 2016. Accepted manuscript online December 8, 2016.
Address correspondence to: Edward M Schwarz, PHD, University of Rochester Medical Center, 601 Elmwood Avenue, Box 665, Rochester, NY 14642, USA.
E-mail: Edward_Schwarz@URMC.Rochester.edu and Karen Bentley, MS, Pathology, Electron Microscope Resource, 601 Elmwood, Box 626, Rochester, NY 14642,
USA. E-mail: Karen_Bentley@URMC.Rochester.edu
Journal of Bone and Mineral Research, Vol. 32, No. 5, May 2017, pp 985–990
DOI: 10.1002/jbmr.3055
© 2016 American Society for Bone and Mineral Research

985
protein A-deficient mutant of UAMS-1 (DSpa; provided by which were placed in the four outermost screw holes of the
Dr. Paul Dunman, University of Rochester, Rochester, NY), and plate. A 0.7-mm transverse osteotomy was cut through the
USA300LAC.(11) These strains were used as described in the three femoral mid-diaphysis under the plate using a 0.67-mm wire
separate in vivo experiments, and an in vitro study, as listed in Gigli saw and a cutting guide (RISystem, Davos, Switzerland). A
the sections below. All animal studies were performed in 1-mm radius semicircle of fibrillar collagen sheet (Kensey Nash,
accordance with protocols approved by the University of Exton, PA, USA) was soaked for at least 2 hours in an overnight
Rochester’s Committee on Animal Resources. To be consistent culture of USA300 prepared as described in Experiment 1
with our previous studies,(4,5,12) normal and healthy female (contaminating dose 5  104 CFU, group 1, test), or was
Balb/c J mice (Jackson Research Labs, Bar Harbor, ME, USA) at 8 untreated (group 2, sterile control). A collagen sheet was
to 10 weeks of age were acclimated for 1 week prior to surgery placed into the femoral defect (n ¼ 5 per group; 10 total),
and randomly assigned to an infection group. All mice were fed which was secured with 5-0 nylon monofilament sutures
Purina Chow 5008 (St. Louis, MO), and were anesthetized just wrapped around the muscle and skin during wound closure to
prior to each type of surgery with xylazine (12 mg/kg) and ensure retention of the collagen sheet in the osteotomy defect
ketamine (130 mg/kg) injected intraperitoneally. The weight of for the duration of the in vivo experiment. The mice were
the mice at the time of surgery ranged from 21 to 28 g. They housed for 14 days prior to euthanasia and femur harvest for
typically lost 2 to 3 g 1 day postoperation, and recovered this TEM analysis.
weight loss by day 14. There were no significant weight
differences between the groups at any time point. Experiment 3
To assess the role of proliferation in S. aureus infection of bone
Experiment 1 following direct contamination, mice (n ¼ 5; 10 total) received
To assess S. aureus infection during the establishment of the femoral osteotomy surgery with a sterile collagen sheet
implant-associated osteomyelitis, we utilized a transtibial pin (group 1, negative control), or a collagen sheet soaked with
implant model.(4) Briefly, an overnight culture was established UAMS-1DSpa (group 2). Following surgery, the mice were fed
by inoculating 1  107 colony forming units (CFU) of UAMS-1 in 0.5 mg/mL bromodeoxyuridine (BrdU) (Sigma, St. Louis, MO,
10 mL of tryptic soy broth (TSB), which was incubated for USA) in their drinking water, which was available ad libitum
16 hours at 37°C. The next day, flat stainless steel wires were 24 hours/day until euthanasia on day 14, and the femurs were
incubated in this overnight culture for 30 min prior to surgical harvested for IEM.
implantation into the mice, which contaminated them with
5  105 CFU (group 1, test). Untreated wires were used as TEM
group 2 sterile pin controls (n ¼ 5 per group; 10 total). After Tibias were excised and immersion fixed in 0.1M sodium
anesthesia, a 5-mm to 7-mm longitudinal skin incision was made cacodylate buffered 2.5% glutaraldehyde/4% paraformalde-
on the medial side of the proximal right lower limb. The center of hyde overnight. Femurs were perfusion fixed via the
the incision was about 5 mm posteromedial to the tibial descending aorta using the same fixative. Afterward, the
tuberosity. Then the soft tissue was removed, and the medial long bones underwent decalcification for 7 to 10 days with
aspect of the proximal tibia was exposed. A medial hole was 14% EDTA. Following decalcification, all implants (pins,
created by manually inserting a 23G needle at the level of the plates, screws, and sponges) were removed prior to trimming
tibial tuberosity, and the center of the anteroposterior axis. A and processing for TEM. The bones were postfixed in 2.0%
30G needle was then used to enlarge the hole of the medial osmium tetroxide, dehydrated through a graded series of
surface, via insertion parallel to the knee joint line and posterior ethanol to 100%, transitioned into propylene oxide (PO), PO/
aspect of the tibia. Then, a flat stainless steel wire (cross-section EPON/Araldite epoxy resin, two changes of 100% EPON/
0.2 mm  0.5 mm; MicroDyne Technologies, Plainville, CT, USA) Araldite epoxy resin before embedment into large inverted
was cut to a 4 mm length, and bent into an L-shaped implant (flat cap side) BEEM size 00 capsule molds for 48 hours of
(long side 3 mm, short side 1 mm), which was inserted along a polymerization at 60°C. A variation of our large block
medial-to-lateral path through the medial and lateral holes, sectioning technique(13) was utilized to produce 1-mm
leaving the 1-mm dimension protruding on the medial aspect of sections of the entire proximal tibia (>3 mm), and the
the tibia, but tightly apposed to its periosteal surface. Afterward, femoral diaphysis flanking the osteotomy (two pieces 3 mm
the muscle and skin were closed with sutures. The mice were proximal and distal to the defect) were embedded as intact
housed for 14 days prior to euthanasia and tibia harvest for TEM bone samples. One-micron (1-mm) sections cut using a
analysis. diamond Histoknife, and stained with Toluidine blue, were
used to identify S. aureus infiltrating into the canalicular
Experiment 2 system of cortical bone. Serial step (every 300 nm) thin
To assess S. aureus infection of bone following direct sections were cut at 70 nm using a diamond knife, placed
contamination, we utilized a femoral osteotomy model.(5) onto formvar/carbon slot grids, stained with aqueous uranyl
Briefly, after anesthesia the right femur was exposed by a acetate and lead citrate then imaged using a Hitachi 7650
direct lateral approach blunt dissection, and a six-hole TEM with an attached Gatan Erlangshen 11 megapixel digital
radiolucent polyether ether ketone (PEEK) femoral plate camera.
(11 mm in length  1.75 mm in width  1 mm thick) with a
IEM
40-nm titanium coating (RIS.401.130 MouseFix plate 6 hole,
rigid; RISystem, Davos, Switzerland), was implanted 2 to 3 mm UAMS-1DSpa was used for IEM to eliminate background
from the distal end of the femur on the anterolateral surface of antibody binding to protein A. The femurs were perfusion
the bone using just four titanium screws (RIS.401.100 fixed, excised, and postfixed an additional 8 hours with 4.0%
MouseFix screw, L ¼ 2.00 mm; RISystem, Davos, Switzerland), paraformaldehyde in 0.1M sodium cacodylate buffer and then

986 DE MESY BENTLEY ET AL. Journal of Bone and Mineral Research

decalcified with 14% EDTA for 10 days before embedding into using the same Hitachi 7650 TEM and the same digital camera
LR White resin. The blocks were sectioned at 1-mm for system.
Toluidine blue staining to identify by light microscopy cortical
In vitro experiment
bone infection sites (Fig. 2A). Thin sections at 80 nm were
placed onto carbon/formvar 150 mesh grids and incubated in To validate S. aureus migration through submicron pores,
5N HCl for 15 min to denature double stranded DNA into we utilized an in vitro transwell system (Fig. 2E). Briefly,
single strands. The grids were rinsed in distilled water, PBS, 0.4-mm-thick silicon nitride membrane chips containing pores
incubated in blocking solution containing 2% normal donkey 0.5 mm in diameter were fabricated by SiMPore Inc. (West
serum in PBS, then overnight at 4°C in sheep anti-BrdU Henrietta, NY, USA) as described.(14) Individual chips were housed
(Fitzgerald) primary antibody at 1/100, rinsed in PBS and within a circular culture well system made of a firm silicon gel. A
incubated at room temperature for 1.5 hours in donkey anti- silicon gasket was used to form a tight seal maintaining 150 mL of
sheep 12-nm gold-tagged secondary (Jackson ImmunoRe- GFPþ UAMS-1 in TSB exclusively on the topside of the chip
search). For a negative control, the primary antibody was (covered with a sterile coverslip). Real-time fluorescent imaging
substituted with 2% normal serum in PBS. The grids were of the bacteria at 37°C was performed using an Olympus FV 1000
stained with uranyl acetate and lead citrate and examined microscope, in which z-stack images, starting at the topside of

Fig. 1. TEM evidence of submicron-elongated S. aureus in the osteocytic lacunar-canaliculi network of infected live bone tissue. Long bones from mice
(n ¼ 5) infected with a UAMS-1–contaminated tibial pin (A–F, H, I), or a USA300-infected femoral osteotomy (G), were harvested on day 14 postinfection
for TEM. (A) Low magnification TEM image (4000) of UAMS-1 invasion of live bone tissue (note osteocyte OC) in a canaliculus (green arrow)
communicating with the marrow cavity (MC). Also note the proximal neutrophils (yellow arrow) within the marrow. (B) Low magnification TEM image
(4000) of UAMS-1 invasion of an osteocytic lacunar-canaliculus adjacent to a channel infected with S. aureus (arrows) containing necrotic cells ( ).
Higher magnification TEM images (C: 8000; D: 10,000) of UAMS-1 colonization of osteocytic lacunae. (E) Low magnification TEM image of three
parallel canaliculi in various states of colonization (1: severely infected; 2: moderately infected; and 3: uninfected) by the invading UAMS-1 within the live
cortical bone (3500). (F) Higher magnification TEM image measuring submicron-elongated UAMS-1 (15,000). (G) Similar bacterial invasion of
canaliculi adjacent to the osteotomy (red arrow), and neutrophils in the marrow cavity ( ) were observed in USA300-infected femurs (4000), but not in
long bones that received sterile implants (data not shown). (H) Low magnification TEM image (4000) documenting cortical bone damage adjacent to
the infected tibia pin (red arrows), and a cavity filled with UAMS-1 (yellow bracket) that leads to a canaliculus (black arrow). (I) High magnification TEM of
the infected cavity in H demonstrating mitotic S. aureus in the live cortical bone (25,000). Note that only the bacterium entering the canaliculus has an
asymmetric septal plane (red arrows), which is aligned perpendicularly with the canaliculus orifice, perhaps to anchor and propel the emerging daughter
cell into the submicron channel in the cortical bone during binary fission.

Journal of Bone and Mineral Research S. AUREUS IN CANALICULI OF CORTICAL BONE IN MURINE MODELS OF OSTEOMYELITIS 987
Fig. 2. In vivo and in vitro evidence of proliferating S. aureus through canaliculi and submicron pores. Mice (n ¼ 5) received a femoral osteotomy that was
infected with UAMS-1 DSpA, and were fed BrdU in their drinking water continuously postoperation until euthanasia on day 14. (A) Light microscopy
image (4) of the toluidine blue–stained cortical bone section containing a defect (yellow bracket) leading to a canaliculus (black bracket) colonized with
S. aureus, which was interrogated by immunoelectron microscopy (IEM). (B) Low (15,000 boxed region in A), (C) high (40,000 boxed region in B), and
(D) ultrahigh (boxed region in C enlarged from 150,000 original) magnification of IEM images of BrdU-positive S. aureus. Note the rod-shaped bacterium
with a diameter of 0.36-mm at the leading edge of bacterial infiltration (C), and the immunogold-labeled chromosome (12 nm black dots in D) confirming
BrdU incorporation. IEM on S. aureus–infected femurs from mice that were not fed BrdU were all negative for immunogold labeling (data not shown).
(E) Our in vitro transwell culture system loads via an open chamber on the topside (white arrow), through which GFPþ UAMS-1 in TSB medium is placed
onto a 0.4-mm-thick silicon nitride submicron porous (0.5 mm in diameter) membrane. Note left and right access channels (black arrows) for loading TSB
medium into the underside reservoir, which is a sealed chamber that is physically separated from the topside of the submicron porous membrane.
(F) SEM (15,000) of the static biofilm that forms on the top surface of the membrane at 3 hours. Validation of GFPþ UAMS-1 migration through the
submicron pores is shown by confocal fluorescent microscopy of the bacteria (G top view) and a 3D reconstructed image of five confocal slices (H).
S. aureus migration through the submicron pores via proliferation is evidenced by low (I: 2000) and high (J: 12,000) magnification SEM imaging of the
underside of the membrane demonstrating extrusion of the bacteria at 6.5 hours. (K) The increase in GFPþ S. aureus occupancy of the submicron pores
was quantified via real-time fluorescent confocal microcopy from the topside of the membrane, and the data are presented as the mean of three
independent experiments  SEM in which the plateau (saturation time) is estimated at 35.6 min (95% CI, 25.8 to 45.4 min;  p < 0.05 versus 5 minutes).
(L–P) Representative fluorescent images (60) from the topside obtained at 15-min intervals from 0 to 60 min are shown to illustrate the increase in pore
occupation over the first 30 min, and subsequent saturation thereafter.

the membrane, and collected through to the underside of the Results

membrane, were acquired at 15-min intervals to quantify pore
occupancy and migration of GFPþ bacteria through the pores. Consistent with prior radiology, histology, and SEM studies of
SEM was also performed as described(5) to document the static uninfected controls with the tibia implant and femoral
biofilm growth on the topside, and visualize bacteria emerging (at osteotomy models,(4,5) TEM analyses did not identify bacteria
6.5 hours) through the submicron pores from the underside. in murine long bones (n ¼ 5 per experiment; n ¼ 15 total) that

988 DE MESY BENTLEY ET AL. Journal of Bone and Mineral Research

were not challenged with S. aureus (data not shown). of this bacterial species, which has been historically defined as
Additionally, consistent with some reports,(15–17) TEM analysis non-motile cocci, based on the absence of any discernible
of five infected tibias and 10 infected femurs rarely demon- motility structures (ie, flagella, cilia, pseudopods), and its
strated colonized bone cells (osteoblasts/bone lining cells, morphology during both planktonic and biofilm growth.(20)
osteocytes, and osteoclasts) that are reported in some This definition is also consistent with the most recent
publications.(6–9) In contrast, we readily observed S. aureus pathogenesis studies, which demonstrated that S. aureus
within canaliculi of live cortical bone, which existed as chains of utilizes an array of virulence genes to establish immune
individual cocci and submicron rod-shaped bacteria leading to privileged staphylococcal abscess communities (SAC) during
biofilm bacteria in large canaliculi and osteocyte lacunae (Fig. 1). chronic infection of soft tissues.(21) However, a complete
Further scrutiny of these TEM images revealed several remark- explanation of chronic osteomyelitis, and why extensive
able features of S. aureus invasion and colonization of the live surgical debridement beyond the margins often fails to
cortical bone tissue, as evidenced by bacteria filled canaliculi eradicate S. aureus bone infections (ie, reinfection rates
(10-mm away from osteocytes) and lacunae (Fig. 1A–D). The following revision surgery for MRSA-infected total joint
first is differential colonization based on location, as evidenced replacements have been reported to be up to 40%(22))
by the severely infected, moderately infected, and uninfected remains elusive. Thus, identifying the nature and location of
parallel osteocytic lacunar-canaliculi (Fig. 1E). Another remark- the S. aureus reservoir in chronic osteomyelitis is of critical
able finding was the extent of S. aureus deformation in the importance.
narrowest canaliculi, which limited the bacterial diameter to Here we utilized TEM to analyze two murine models of bone
20% of its native 1.0 to 1.5-mm size (Fig. 1F). Of note is that the infection, which demonstrated bacteria reside within canal-
features of UAMS-1–infected bone could also be found in tibias iculi of the infected bone. Thus, these data provide strong
and femurs colonized with USA300LAC, and no obvious evidence that S. aureus colonization of the osteocytic lacunar-
differences between strains were observed, suggesting similar canaliculi network within live bone could be an important
colonization mechanisms between MSSA and MRSA strains (Fig. reason these infections are incurable. Moreover, we observed
1G). Last, we observed two unprecedented apparent features of what appears to be a novel motility mechanism that allows
S. aureus at the entrance of a canaliculus (Fig. 1H, I). The first is S. aureus to: (1) identify canaliculi in cortical bone; (2) deform
metamorphosis from conventional rounded cocci to rod-shaped from spherical cocci into submicron rod-shaped bacteria; and
bacteria oriented to enter the canaliculus. The second is (3) propagate throughout these channels, potentially via
asymmetric binary fission in which the septal plane is aligned haptotaxis (cell migration defined by the 3D structure of the
perpendicular to the canaliculus, perhaps to anchor the extracellular matrix), durotaxis (cell migration defined by the
bacterium and propel the developing daughter cell into the rigidity of the extracellular matrix), and asymmetric binary
submicron bone channel. Collectively, these static data may fission.
suggest that S. aureus is a “motile” organism in vivo, which uses There are several noteworthy limitations to this study. The first
this hitherto unreported behavior to invade and colonize live is that the results from prototypic MSSA and MRSA cannot be
cortical bone in a mammalian host. generalized to all S. aureus strains. Similarly, we utilized a
To further define the mechanism of S. aureus motility in prototypic Th2 host (Balb/c), and thus subsequent studies in a
bone, we performed IEM following in vivo BrdU labeling of Th1 host (ie, C57B/6) are warranted.(23,24) Additionally, we used
infected mice (Fig. 2A–D). The results confirmed that deformed 8-week-old to 10-week-old growing mice with cortical bone
S. aureus at the leading edge of submicron canaliculi undergo containing primary Haversian systems to represent a condition
proliferation in vivo, as determined by BrdU incorporation into that most often occurs in the adult human skeleton in which
the bacterial chromosome. This finding is consistent with an many cycles of remodeling have occurred. Thus, additional
extrusion motility mechanism in which S. aureus: (1) enters experiments to confirm similar S. aureus invasion of aged cortical
submicron bone channels via asymmetric binary fission; and bone tissue are needed. Finally, genetic screens utilizing our in
(2) migrates along canaliculi to invade osteocyte lacunae via vitro real-time imaging of S. aureus migration through a
proliferation at the leading edge. Because in vivo dynamic submicron membrane filter are warranted to identify the genes
imaging of this process is not possible, we generated an in involved and develop potential novel antibiotics to prevent and
vitro transwell model to document S. aureus migration treat osteomyelitis.
through submicron pores with real time fluorescent imaging
of GFPþ UAMS-1 (Fig. 2E–P). We also performed SEM of the Disclosures
underside of the membrane, which confirmed extrusion of
actively divided bacteria transmigrating through the 0.5-mm
All authors state that they have no conflicts of interest.
pores.

Acknowledgments
Discussion
This work was supported by grants from AOTrauma (Davos,
Due to its ability to cause an assortment of serious infectious Switzerland); National Institutes of Health, Grant numbers: P30
diseases, S. aureus has been intensely studied for centuries. It AR069655, P50AR054041, S10RR026542, and T32AR53459. We
has been over 200 years since Sir Benjamin Brodie described thank Gayle Schneider of the URMC Electron Microscope
musculoskeletal infections likely caused by this microbe in his Shared Resource Laboratory for her technical assistance of
1813 treatise Pathological Researches Respecting the Diseases of serial thin sectioning for TEM; Drs. Linda Callahan and Paivi
Joints,(18) and the lesion that is now commonly referred to as a Jordan of the Confocal and Conventional Microscopy Shared
“Brodie’s abscess” from his first description of subacute Resource; and Dr. Christopher A. Beck for biostatistical
osteomyelitis.(19) As a result, we have a general understanding support.

Journal of Bone and Mineral Research S. AUREUS IN CANALICULI OF CORTICAL BONE IN MURINE MODELS OF OSTEOMYELITIS 989
References protects mice from implant-associated osteomyelitis by mediating
opsonophagocytosis of Staphylococcus aureus megaclusters.
1. Darouiche RO. Treatment of infections associated with surgical J Orthop Res. 2014;32(10):1389–96.
implants. N Engl J Med. 2004;350(14):1422–9. 13. de Mesy Jensen KL, di Sant’Agnese PA. Large block embedding and
2. Brady RA, Leid JG, Calhoun JH, Costerton JW, Shirtliff ME. “pop-off” technique for immunoelectron microscopy. Ultrastruct
Osteomyelitis and the role of biofilms in chronic infection. FEMS Pathol. 1992;16(1-2):51–9.
Immunol Med Microbiol. 2008;52(1):13–22. 14. Agrawal AA, Nehilla BJ, Reisig KV, et al. Porous nanocrystalline silicon
3. Hernandez-Vaquero D, Fernandez-Fairen M, Torres A, et al. Treat- membranes as highly permeable and molecularly thin substrates for
ment of periprosthetic infections: an economic analysis. Scienti- cell culture. Biomaterials. 2010;31(20):5408–17.
ficWorldJournal. 2013;2013:821650. 15. Gristina AG, Costerton JW. Bacterial adherence to biomaterials and
4. Nishitani K, Sutipornpalangkul W, de Mesy Bentley KL, et al. tissue. The significance of its role in clinical sepsis. J Bone Joint Surg
Quantifying the natural history of biofilm formation in vivo during Am. 1985;67(2):264–73.
the establishment of chronic implant-associated Staphylococcus 16. Horst SA, Hoerr V, Beineke A, et al. A novel mouse model of
aureus osteomyelitis in mice to identify critical pathogen and host Staphylococcus aureus chronic osteomyelitis that closely mimics the
factors. J Orthop Res. 2015;33(9):1311–9. human infection: an integrated view of disease pathogenesis. Am J
5. Inzana JA, Schwarz EM, Kates SL, Awad HA. A novel murine model of Pathol. 2012;181(4):1206–14.
established Staphylococcal bone infection in the presence of a 17. Power ME, Olson ME, Domingue PA, Costerton JW. A rat model of
fracture fixation plate to study therapies utilizing antibiotic-laden Staphylococcus aureus chronic osteomyelitis that provides a suitable
spacers after revision surgery. Bone. 2015;72:128–36. system for studying the human infection. J Med Microbiol.
6. Ellington JK, Harris M, Hudson MC, Vishin S, Webb LX, Sherertz R. 1990;33(3):189–98.
Intracellular Staphylococcus aureus and antibiotic resistance: im- 18. Brodie BC. Pathological researches respecting the diseases of joints.
plications for treatment of staphylococcal osteomyelitis. J Orthop Med Chir Trans. 1813;4:210–80.
Res. 2006;24(1):87–93. 19. Brodie BC. An account of some cases of chronic abscess of the tibia.
7. Reilly SS, Hudson MC, Kellam JF, Ramp WK. In vivo internalization of Med Chir Trans. 1832;17:239–49.
Staphylococcus aureus by embryonic chick osteoblasts. Bone. 20. Baird-Parker AC. Classification and identification of staphylococci
2000;26(1):63–70. and their resistance to physical agents. In: Cohen JO, editor. The
8. Hamza T, Li B. Differential responses of osteoblasts and macrophages staphylococci. New York: Wiley; 1972. Chapter 1, p. 1–20.
upon Staphylococcus aureus infection. BMC Microbiol. 2014;14:207. 21. Cheng AG, Kim HK, Burts ML, Krausz T, Schneewind O, Missiakas DM.
9. Bosse MJ, Gruber HE, Ramp WK. Internalization of bacteria by Genetic requirements for Staphylococcus aureus abscess formation
osteoblasts in a patient with recurrent, long-term osteomyelitis. A and persistence in host tissues. FASEB J. 2009;23(10):3393–404.
case report. J Bone Joint Surg Am. 2005;87(6):1343–7. 22. Azzam K, McHale K, Austin M, Purtill JJ, Parvizi J. Outcome of a
10. Gillaspy AF, Hickmon SG, Skinner RA, Thomas JR, Nelson CL, Smeltzer second two-stage reimplantation for periprosthetic knee infection.
MS. Role of the accessory gene regulator (agr) in pathogenesis of Clin Orthop Relat Res. 2009 Jul;467(7):1706–14.
staphylococcal osteomyelitis. Infect Immun. 1995;63(9):3373–80. 23. Prabhakara R, Harro JM, Leid JG, Harris M, Shirtliff ME. Murine
11. Kourbatova EV, Halvosa JS, King MD, Ray SM, White N, Blumberg immune response to a chronic Staphylococcus aureus biofilm
HM. Emergence of community-associated methicillin-resistant infection. Infect Immun. 2011;79(4):1789–96.
Staphylococcus aureus USA 300 clone as a cause of health care- 24. Prabhakara R, Harro JM, Leid JG, Keegan AD, Prior ML, Shirtliff ME.
associated infections among patients with prosthetic joint Suppression of the inflammatory immune response prevents the
infections. Am J Infect Control. 2005;33(7):385–91. development of chronic biofilm infection due to methicillin-
12. Varrone JJ, de Mesy Bentley KL, Bello-Irizarry SN, et al. Passive resistant Staphylococcus aureus. Infect Immun. 2011;79(12):
immunization with anti-glucosaminidase monoclonal antibodies 5010–8.

990 DE MESY BENTLEY ET AL. Journal of Bone and Mineral Research