Characterization of Staphylococcus aureus strains in arabbit model of osseointegrated pin infections

Dustin Williams,1 Roy Bloebaum,2,3 Cathy A. Petti1,41Associated Regional and University Pathologists (ARUP) Laboratories, ARUP Institute forClinical and Experimental Pathology, Salt Lake City, Utah2Department of Veterans Affairs, Salt Lake City, Utah3Department of Orthopaedics, University of Utah School of Medicine, Salt Lake City, Utah4Departments of Medicine and Pathology, University of Utah School of Medicine, Salt Lake City, Utah

Received 10 November 2006; revised 5 March 2007; accepted 3 May 2007Published online 9 August 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.31485

Abstract: Staphylococcus aureus is a common infectingagent of many surgical sites. As a commensal organism tohumans and rabbits, the infection process may occur dueto native or exogenous S. aureus. We applied exogenousS. aureus ATCC 49230 once weekly to the surgical site ofan osseointegrated pin in 20 New Zealand white rabbits.Clinical signs of infection resulted in euthanasia and at ne-cropsy samples were collected from putatively infectedsites. The predominant organism cultured was S. aureus.We observed various b-hemolysis patterns of S. aureus onculture media and used pulsed field gel electrophoresis(PFGE) to determine whether there were distinct strains ofS. aureus collected from various sites of the rabbits. On the

basis of PFGE results, we found that the exogenous S. aur-eus ATCC 49230 was not the S. aureus cultured during ne-cropsy, but that S. aureus native to the rabbits was in factthe infecting agent. We conclude that this rabbit model forS. aureus infection, which has not been described previ-ously, may contribute to understanding the pathogenesisof S. aureus infections in future studies with simulatedosseointegrated pin infections secondary to S. aureus.� 2007 Wiley Periodicals, Inc. J Biomed Mater Res 85A:366–370, 2008

Key words: infection; osseointegration; Staphylococcus aur-eus; PFGE; rabbit model

INTRODUCTION

Biomaterial and device-related infections are com-monly attributed to biofilm-producing Staphylococcusaureus,1 a pathogen often commensal with humanand rabbit normal flora.2–5 Exit-sites of percutaneousdevices in rabbit models have been used as surro-gate markers for evaluating and identifying rates ofS. aureus infection with the rate of natural infectioncaused by S. aureus occurring in approximately 10%of rabbits in a pin tract infection model.6 Infectionrates have been shown to increase with the inocula-tion of exogenous S. aureus at a surgical site, butthese previous reports were limited by investigatorsnot distinguishing between native and inoculatedtype-strains of S. aureus that were responsible forsurgical site infection.7,8

Aware that surgical site infections can be monomi-crobial or polymicrobial, and distinguishing truepathogens from commensal skin flora would be im-portant, we evaluated the colony morphologies ofS. aureus cultured from infected rabbit skin and softtissue. To distinguish whether the infecting microor-ganism originated from the normal rabbit flora orexogenous S. aureus, we used pulsed field gel elec-trophoresis (PFGE) to determine strain relatedness.PFGE uses pulsing electrical currents, as opposed toconstant current, through agarose gel to separateDNA fragments cut by nuclease enzymes. The DNAfingerprint then gives an organism an identity bybanding pattern, making it possible to distinguishbacterial samples (i.e. S. aureus) of the same origin orof separate identity. We compared PFGE patterns ofexogenous S. aureus (type strain inoculated at surgi-cal site) with S. aureus strains cultured from sites ofinfection and noninfected sites (e.g. intact skin,throat) to define the source of S. aureus implantinfection.

The goal of this study was to use a rabbit osseoin-tegrated implantation model to determine whetherS. aureus applied to a surgical site increased the risk

Correspondence to: D. Williams; e-mail: dustin.williams@aruplab.comContract grant sponsor: U.S. Army Medical Research

and Materiel Command; contract grant number: 05342001

' 2007 Wiley Periodicals, Inc.

of infection and to determine whether the exoge-nously applied S. aureus would in fact be the infect-ing agent. We hypothesized that adding an exoge-nous strain of S. aureus would increase the infectionrate in osseointegrated implants.

MATERIALS AND METHODS

Pin treatment

Eighteen New Zealand white rabbits of either sex, age>4.5 months, and weight >4 kg were implanted with ahexagonal, 2.7 mm maximum diameter 3 36 mm length ti-tanium alloy (Ti-6AI-4V) pin free of surface finish, and 19others were implanted with porous tantalum metal pins.Pins were custom manufactured by the Zimmer Corpora-tion with sterilization by gamma radiation. Implantationinvolved transcortical insertion of the pin into the lateralaspect of the upper tibia in a unilateral, percutaneous fash-ion using aseptic technique. Nine of the 18 rabbits with ti-tanium alloy pins served as controls for both test groupsafter 8 of the 19 rabbits with tantalum metal pins thatwere to serve as controls were removed from the studydue to skin displacement and enclosure of the pin site.

At 14 days postsurgery and once weekly thereafter, a1 3 108 CFU/200 lL concentration of S. aureus AmericanType Culture Collection (ATCC) 49230 (ATCC strain frompatient with osteomyelitis) was applied to the pin siteusing a sterile transfer pipet of 9 rabbits with titaniumpins and the remaining 11 with porous tantalum metalpins. The 1 3 108 S. aureus ATCC 49230 concentration wasobtained by adjusting the bacterial suspension in 1 mLphosphate buffered saline (PBS)/25% glycerol to a 1.0McFarland standard equivalent to *3 3 108 S. aureus/mL.Samples were centrifuged and resuspended in 600 lLPBS/25% glycerol solution then separated into three cryo-vial tubes in 200 lL aliquots giving *1 3 108 S. aureus/200 lL.

This study was approved by the Institutional AnimalCare and Use Committee (IACUC) at the Veterans AffairsMedical Center, Salt Lake City, Utah and conducted in ac-cordance with National Institutes of Health (NIH) guide-lines for the care and use of laboratory animals.

Sample collection

Immediately prior to surgery, random swab samplesfrom 27 of 29 rabbits’ skin, 9 of 29 rabbits’ nares, and 8 of29 rabbits’ throats were obtained to determine whethernormal flora bacteria were present before contact withequipment or humans. Infection data was collected over a6 month period. Rabbits that displayed grade II clinicalsigns of infection as per Checketts et al.,9 wherein a quali-tative observation of swelling, redness, purulent discharge,lowered limb usage, and/or limping caused by infectionor other observation suggesting pain or discomfort, wereeuthanized by injection to the rabbit ear prior to the 6month completion. The euthanasia solution contained phe-

nytoin sodium and pentobarbital sodium at a dose of1 mL/4.5 kg.

Immediately prior to euthanasia, blood culture samplesfrom the rabbit ear were collected aseptically with beta-dine and isopropyl alcohol application. Following euthana-sia, swabs of putatively noninfected, intact rabbit skin,throat, and posterior nares were taken as controls of nor-mal microbial flora. Swabs were also taken of the infectedpin site including purulent material. After blood and swabcollection, deeper, sub-dermal tissue was obtained withaseptic technique by a punch biopsy 5 mm distal to theinfected pin site. With a drill bit, bone samples were col-lected through the medullary canal to the pin site by ream-ing action. All swab, tissue, and blood samples were pro-cessed and cultured by a standard laboratory protocol.After growth, the presence or absence of b-hemolysis forS. aureus isolates were noted.

Infection was defined as isolation of S. aureus from tis-sue from a surgical site clinically suspicious for infectionor from a sterile body site. Isolation of S. aureus only fromthe skin/pin tract interface was not considered an infectionsince S. aureus can be present as normal skin and oropha-ryngeal flora, which could colonize the site, without signsof clinical infection, from rabbit licking. Samples culture-positive for S. aureus were saved for further analysis.

Pulsed field gel electrophoresis

To determine the relatedness among S. aureus isolatesincluding the exogenously applied S. aureus strain, we per-formed PFGE. S. aureus isolates recovered from rabbit sam-ples with S. aureus ATCC 49230 as a control were pro-cessed and PFGE was performed by a standard laboratoryprotocol.10 DNA banding patterns of samples collectedfrom different sites were compared and analyzed. Thenumber of DNA band differences determined whetherthe strains were the same or distantly related based on thework of Tenover et al.10 Strains with seven or more banddifferences were determined to be of separate origin.Strains with three or less band differences were defined asindistinguishable or closely related.

RESULTS

Native versus exogenous S. aureus infection

Of sites sampled to assess native microbial flora,S. aureus was recovered from random skin samplesin 4 (15%) of 27 rabbits, but no S. aureus wasdetected in the nares or throat of any rabbitssampled (Table I). S. aureus was recovered from theskin/pin interface from six of the nine control rab-bits, but only two were classified as infected withS. aureus according to our case definition (Table II)—one being euthanized prior to the 6 month comple-tion and 1 at the 6 month completion. Three of theseven noninfected control rabbits (euthanized at 6months) displayed grade I signs—irritation or

CHARACTERIZATION OF STAPHYLOCOCCUS AUREUS STRAINS 367

Journal of Biomedical Materials Research Part A

inflammation (e.g., cage rubbing)—as per Checkettset al.9 at necropsy.

For the 20 rabbits treated weekly with S. aureus,all were euthanized prior to a 6 month completionperiod. Of these, S. aureus was recovered from 8 of 9(89%, p ¼ 0.0045) with titanium pins and 10 of 11(91%, p ¼ 0.0152) with tantalum pins. Combining thedata, all 18 out of 20 were found to have S. aureus ata site other than the skin/pin interface, confirmingclinical infection (p ¼ 0.0007). One rabbit from eachof these groups displayed grade II clinical signs ofinfection, but had negative culture results. The pre-dominant organism cultured from each collectionsite—pin/skin interface, muscle, tissue, bone, andblood—was S. aureus (Table III).

Bone ingrowth was observed in seven representa-tive tantalum metal pins by 6 weeks, suggesting asuitable level of osseointegration occurred (data notshown). A noticeable gap at the skin/pin interfacewas observed by light microscopy around each tan-talum metal pin, suggesting failure of adequate skiningrowth (data not shown).

Microbiologic findings

Following rabbit sample collection and growth ofS. aureus on sheep blood agar, various patterns of b-hemolysis were observed. Variations were present inS. aureus from each site of the rabbits, infected andnoninfected alike. Some strains were slightly b-he-molytic, wherein the hemolysis was seen only whenthe bacterial colony was removed from the plate andthe agar underneath the colony had visible hemoly-

sis. Other strains were highly b-hemolytic, whereinhemolysis was seen surrounding bacterial colonies.

PFGE

PFGE was performed on 36 S. aureus isolates withS. aureus ATCC 49230 as a control. The 36 sampleswere representative of 6 of 9 rabbits with titaniumpins and 5 of 11 with tantalum pins. With PFGE,three different banding patterns were found (Fig. 1).Two of the banding patterns had only one band dif-ference and were determined to be the same strain.The third banding pattern had a difference of at leastseven bands from each of the other two patternswhich classified it as a distinct S. aureus strain (Fig.1). Banding patterns 1 and 2 were consistently dis-played in all (36/36) S. aureus samples from infected

TABLE IRandom Sampling of Normal Flora

Nares Throat Skin (Distalto Pin Site)

Recovery of S. aureus 0 0 4a

Recovery of other bacteria 8 8 1Samples with no growth 1 0 22Total number sampled 9 8 27

aAll S. aureus strains were classified as native strain bypulsed field gel electrophoresis.

TABLE IIControl Rabbits (n = 9)

Samples Collected at Necropsy

Skin/PinInterface

Tissue Bone Blood

Recovery of S. aureus 6 1a 1b 1b

Recovery of other bacteria 1 0 0 0Samples with no growth 2 8 8 8Total number sampled 9 9 9 9

aRabbit 7341.bRabbit 7345.

TABLE IIIRabbits with Exogenously Applied S. aureus (n = 20)

Samples Collected at Necropsy

Skin/PinInterface

Tissue Bone Blood

Recovery of S. aureus 19 16 1 1Recovery of other bacteria 1 2 1 9Samples with no growth 0 2 18 10Total number sampled 20 20 20 20

Figure 1. DNA banding patterns of native (1, 2) S. aureusstrains and the S. aureus ATCC 49230 (3) control strain.

368 WILLIAMS, BLOEBAUM, AND PETTI

Journal of Biomedical Materials Research Part A

and noninfected (‘‘native’’ colonized) sites (Fig. 2).Banding pattern 3 was unique to the control S. aur-eus ATCC 49230 strain (i.e. inoculated strain) andwas never cultured from any sample sites.

DISCUSSION

We found distinct colony morphologies of S. aur-eus on culture media suggesting a different infectingstrain, a unique observation that has not beendescribed previously in osseointegrated pin animalmodels. We also found that a weekly application ofS. aureus increases the infection rate at the surgicalsite of osseointegrated pins in rabbits when com-pared to untreated controls. Reports of animal mod-els have shown that an exogenously inoculated typestrain of bacteria at surgical sites was identified asthe infecting strain11,12; however, in our uniquemodel the exogenous bacterial strain (S. aureusATCC 49230) was not identified as the infectingagent. Instead, the infecting S. aureus strain appearedto be from the rabbit’s normal flora.

A potential limitation of this study was the ab-sence of a tantalum control group because of prema-ture removal from the study. Nevertheless, collec-tively and individually, the treated titanium andtantalum metal groups displayed a statistically sig-nificant increase in the rate of infection. Thoughbone ingrowth occurred into porous tantalum metal,the lack of skin ingrowth perhaps deterred its abilityto prevent infection. The increased infection rate of

the tantalum metal group suggests that tantalummetal alone does not reduce infection rates in thisanimal model. Additionally, while none of the naso-or oropharyngeal swabs and only 15% of the ran-dom skin swabs were positive for the native S. aur-eus strain, we believe that sampling error and bacte-rial overgrowth may have been responsible for thepoor recovery of organisms. Although data are notshown, we have continued to randomly sample rab-bits to determine colonization rates for this native S.aureus strain, and 3 (30%) of 10 were colonized withthis native strain. In an attempt to determine the ori-gin of the infecting S. aureus strain, we sampled sev-eral items such as clippers used for rabbit hair re-moval, totes used for rabbit transfer, cages, andother surfaces. Only a tote was positive for thenative S. aureus strain; however, rabbits that hadnever been exposed to that tote still displayed thesame strain of S. aureus. We did not test caretakers,but rabbits from different locations with differentcaretakers displayed the same native S. aureus strain.After eliminating these factors as potential sources ofS. aureus, we suspect that the infecting S. aureusstrain originated from the initial environment of therabbits before shipment then becoming part of therabbits’ normal flora, an observation which has beencorroborated by a previous study.2

The observation that the control rabbits did nothave a significant rate of infection underscores theunique interaction between the exogenously appliedand native S. aureus strains. Though not yet under-stood, the infection process may have similarities tooral microbial ecology.13 For instance, the matrix putin place by Streptococcus mutans on teeth allows attach-ment of other bacteria, followed by metabolic acidbyproduct causing dental carries in the mouth. Aswith any environment, certain bacteria have a geneticpredisposition, or ‘‘preference,’’ for attachment sites.Normal flora S. aureus may have had firm attachmentto skin extracellular matrix or other peptides andwhen large amounts of exogenous S. aureus wereapplied, a microbial environment was created that fos-tered native S. aureus to infect the surgical site. Thisrelationship between native and exogenous S. aureusstrains requires further study that may include therole of virulence factors. Overall, we conclude thatthis model demonstrates the complexities of microbialcommunities and their role in the pathogenesis of S.aureus infections, and serves as an excellent model tofurther study the prevention of osseointegrated pros-thetic limb and other percutaneous infections.

Opinions, interpretations, conclusions, and recommen-dations are those of the author and are not necessarilyendorsed by the U.S. Army. This material is the result ofwork supported with resources and the use of facilities atthe George E. Wahlen Department of Veterans Affairs, Salt

Figure 2. DNA banding patterns consistently found in allS. aureus samples (native and infected sites). Far left lane isDNA ladder, far right lane is S. aureus ATCC 49230.

CHARACTERIZATION OF STAPHYLOCOCCUS AUREUS STRAINS 369

Journal of Biomedical Materials Research Part A

Lake City, Utah Health Care Systems. The authorsacknowledge the Department of Orthopaedics and theUniversity of Utah School of Medicine, Salt Lake City,Utah as well as John Hibbs, M.D. for his support andinsightful comments, Teri Rosenbaum for her data analysisand Amit Phansalkar for his statistical analysis.

References

1. Maderazo EG, Judson S, Pasternak H. Late infections of totaljoint prostheses. A review and recommendations for preven-tion. Clin Orthop Rel Res 1988;229:131–142.

2. Ajuwape TJ, Aregbesola EA. The bacterial flora of the upperrespiratory tract of normal rabbits. Israel J Vet Med 2002;52:121–123.

3. Devriese LA, Godard C, Okerman L, Renault L. Characteris-tics of Staphylococcus aureus strains from rabbits. Ann Vet Res1981;12:327–332.

4. Kloos WE, Musselwhite MS. Distribution and persistence ofStaphylococcus and Micrococcus species and other aerobic bac-teria on human skin. App Micro 1975;30:381–395.

5. Renquist D, Soave O. Staphylococcal pneumonia in a labora-tory rabbit: An epidemiologic follow-up study. J Am VetMed Assoc 1969;155:1221–1223.

6. Gerritsen M, Lutterman JA, Jansen JA. The influence ofimpaired wound healing on the tissue reaction to percutane-ous devices using titanium fibre mesh anchorage. J BiomedMat Res 2000;52:135–141.

7. Mghir AS, Cremieux AC, Bleton R, Ismael F, Manteau M,Dautrey S, Massias L, Garry L, Sales N, Maziere B, Carbon C.Efficacy of teicoplanin and autoradiographic diffusion patternof [14C]teicoplanin in experimental Staphylococcus aureus infec-tion of joint prostheses. Antimic Agents Chem 1998;42:2830–2835.

8. Sarda L, Mghir AS, Peker C, Meulemans A, Cremieux AC,Guludec DL. Evaluation of 99mTc-ciprofloxacin scintigraphyin a rabbit model of Staphylococcus aureus prosthetic jointinfection. J Nuc Med 2002;43:239–245.

9. Checketts RG, MacEachern AG, Otterburn M. Pin tract infec-tion and the principles of pin site care. In: De Bastiani G,Apley AG, Goldberg A, editors. Orthofix External Fixationin Trauma and Orthopaedics. London: Springer; 2000. pp 97–103.

10. Tenover FC, Arbeit RD, Goering RV, Mickelsen PA, MurrayBE, Persing DH, Swaminathan B. Interpreting chromosomalDNA restriction patterns produced by pulsed-field gel elec-trophoresis: Criteria for bacterial strain typing. J Clin Micro-biol 1995;33:2233–2239.

11. Kalicke T, Schierholz J, Schlegel U, Frangen TM, Koller M,Printzen G, Seybold D, Klockner S, Muhr G, Arens S. Effecton infection resistance of a local antiseptic and antibioticcoating on osteosynthesis implants: An in vitro and in vivostudy. J Orthop Res 2006;24:1622–1640.

12. Wenke JC, Owens BD, Svoboda SJ, Brooks DE. Effectivenessof commercially-available antibiotic-impregnated implants.J Bone Joint Surg Br 2006;88:1102–1104.

13. Gibbons RJ, Houte JV. Bacterial adherence in oral microbialecology. Ann Rev Micro 1975;29:19–44.

370 WILLIAMS, BLOEBAUM, AND PETTI

Journal of Biomedical Materials Research Part A