First Evidence of Sternal Wound Biofilm following Cardiac Surgery

Management of deep sternal wound infection (SWI), a serious complication after cardiac surgery with high morbidity and mortality incidence, requires invasive procedures such as, debridement with primary closure or myocutaneous flap reconstruction along with use of broad spectrum antibiotics. The purpose of this clinical series is to investigate the presence of biofilm in patients with deep SWI. A biofilm is a complex microbial community in which bacteria attach to a biological or non-biological surface and are embedded in a self-produced extracellular polymeric substance. Biofilm related infections represent a major clinical challenge due to their resistance to both host immune defenses and standard antimicrobial therapies. Candidates for this clinical series were patients scheduled for a debridement procedure of an infected sternal wound after a cardiac surgery. Six patients with SWI were recruited in the study. All cases had marked dehiscence of all layers of the wound down to the sternum with no signs of healing after receiving broad spectrum antibiotics post-surgery. After consenting patients, tissue and/or extracted stainless steel wires were collected during the debridement procedure. Debrided tissues examined by Gram stain showed large aggregations of Gram positive cocci. Immuno-fluorescent staining of the debrided tissues using a specific antibody against staphylococci demonstrated the presence of thick clumps of staphylococci colonizing the wound bed. Evaluation of tissue samples with scanning electron microscope (SEM) imaging showed three-dimensional aggregates of these cocci attached to the wound surface. More interestingly, SEM imaging of the extracted wires showed attachment of cocci aggregations to the wire metal surface. These observations along with the clinical presentation of the patients provide the first evidence that supports the presence of biofilm in such cases. Clinical introduction of the biofilm infection concept in deep SWI may advance the current management strategies from standard antimicrobial therapy to anti-biofilm strategy.


Introduction
Median sternotomy is the most common approach used in cardiac surgery procedures to access the heart. The incidence of human sternotomy wound site infection ranges from 1-8% [1,2,3]. This incidence is significant, both economically as well as with respect to health care impact, considering the annual volume of patients undergoing cardiac surgery procedures (more than 600,000 cardiac surgeries per year) and associated health complications [4]. Sternal wound infection (SWI) is a serious complication even after successful surgery with mortality rates reaching 40% [2,5,6]. Additionally, patients with SWI require prolonged antibiotic courses, repeated surgical interventions, longer hospital stay, and eventually higher health care cost [2]. One of the challenges in the management of these patients is overt clinical signs of infection despite ambiguous or negative culture results.
One of the most common pathogens isolated from SWI are Gram positive bacteria, with Staphylococci species being the most frequently reported [7,8,9]. Staphylococci strains (such as S. aureus & S. epidermidis) are known for their ability to grow as biofilms [10]. Once a component of the biofilm structure, the bacterial cells become recalcitrant to the host immune responses as well as antimicrobial therapies, resulting in persistent infection [11]. Microorganisms attach to the surfaces such as wound tissues, biomaterials and surgical implants (e.g. sutures and hard warestainless steel wires) to form biofilms [12,13,14]. Colonization of surgical wounds with biofilm makes them resistant to both antimicrobial as well as other interventions such as surgical debridement aimed at treating wound infection [14,15]. Given the poor prognosis of cardiac surgery wound infection complications, we sought to look for the presence of biofilm at the sternal wound site in patients undergoing cardiac surgery. This work provides the first direct evidence demonstrating presence of biofilm infection in sternal wound site cardiac surgery patients. The introduction of the concept of biofilm infection in deep SWI will help revisit wound management strategies.

Results
Stainless steel wires used for approximation of the sternum after cardiac surgery were tested in vitro for bacterial adhesion, biofilm formation, and recalcitrance to antimicrobial tobramycin. In the SWI cultures from patients, both Methicillin-resistant Staphylococcus aureus (MRSA) and Methicillin-sensitive Staphylococcus aureus (MSSA) were identified (Table 1). Methicillin resistance is independently associated with increased mortality and hospital charges among patients with S. aureus surgical site infections (SSI), therefore, we chose MRSA for in vitro studies [17]. Wires were twisted in a manner similar to that done during closing of sternotomy in the operating room and then incubated with MRSA PFGE strain type USA300 (source, Los Angeles correctional facility), for 24 h. Other wires from the same stock were used as un-inoculated controls. Examination of the wires under scanning electron microscope (SEM) showed attachment and accumulation of MRSA isolates on the wires within extracellular amorphous material forming three-dimensional structures (Fig. 1B). SEM imaging of the control wires showed no microorganisms attached to the metal surface (Fig. 1A).
Biofilms associated with biomedical implant infections are known for their resistance to antibiotics [18]. To determine whether wire-associated bacteria show characteristics of classical biofilm bacteria described in the literature, the wires were inoculated with MRSA and challenged with tobramycin. The resistance to tobramycin was evaluated in wire-associated bacteria  Table 1. Demographic characteristics of patients (n = 9) and SWI status.  versus planktonic bacteria. After 2-h of challenge, tobramycin failed to kill wire-associated bacteria but reduced the planktonic load greater than five-log (Fig. 2). For the clinical study, nine patients were recruited. Three of the nine patients (control non SWI) had a cardiac surgery procedure previously and were scheduled for a second surgical procedure in which they underwent re-sternotomy. The sternotomy wound sites in the three patients were intact with an old scar and no signs of infection were noted. In the test arm, remaining six patients had deep sternal wound infection (SWI) which complicated their cardiac surgery and were therefore scheduled for a sternal debridement procedure (SWI group). These wounds were initially classified as infected by the physician providing care using standard clinical criteria including systemic leukocytosis/fever and localized signs of infection including erythema, necrosis, discharge, and failure of healing. The infection involved the skin, subcutaneous tissue, and extended to the sternum. The sternotomy wound site displayed signs of active infection with localized erythema, exudates, friable wound edges and sternal instability (Fig. 3A). The average interval between the cardiac surgery procedure and the debridement procedure was 2 to 12 weeks in which different classes of antibiotics were used to manage infection (Table 1). Wound cultures showed colonization with MSSA, MRSA in two and other four showed negative culture data.
As an initial screening method, the debrided tissues taken from infected sternal wounds were stained using Gram staining. The staining showed patchy pattern of colonization with numerous Gram positive cocci. Some areas of the tissues showed extensive colonization with large aggregates, while other areas were infected with small micro-colonies or single cocci (Fig. 3B).
To visualize bacterial aggregates in the wound tissue, the surfaces of debrided tissue samples were analyzed using SEM imaging. The samples showed scattered aggregates of cocci attached to the wound surface and appeared as three-dimensional structures (Fig. 4). The bacteria were partially covered with extracellular fibers connecting the cocci together within the threedimensional structure.
According to previous clinical reports, staphylococcus is one of the most common organisms isolated from patients with SWI [7,8,9]. The culture reports from two out of six SWI patients showed MRSA colonization at wound site (Table 1). Morphological analysis of Gram stain and SEM images also suggested staphylococci infection. Identification of staphylococci in the debrided tissues was confirmed by immuno-fluorescence staining using anti-staphylococci antiserum. Indeed, tissue samples taken from infected sternal wounds were showed discrete intense granular green stain indicative of micro colonies of staphylococci (Fig. 5, lower panels,). No evidence of staphylococci was found in tissues taken from non-infected re-sternotomy wound of patients in the control arm (Fig. 5, upper panels).
The architecture of staphylococci micro colonies within the debrided tissues was further studied using confocal laser scanning microscope (CLSM) (Fig. 6). Three-dimensional images were developed to visualize the depth of staphylococci biofilms throughout thick tissue sections (20 mm). Most of the staphylococci were organized in three-dimensional clumps that were scattered across tissue sections (Fig. 6). Together, these clumps constituted a thick staphylococci biomass that traversed through over 70% of the whole tissue section (Fig. 6).
Stainless steel wires extracted from infected or non-infected sternal wound were examined under SEM. The metal surface of the wires was completely coated by mix of extracellular tissue matrix, fibers, and red blood cells. Interestingly, in test patients, we observed three-dimensional clusters of cocci attached to the hardware extracted from infected sternal wound (Fig. 7, lower panels). Such clusters of cocci were not found in wires from noninfected sternal wound of control patients (Fig. 7, upper panels). Additionally, we note that staphylococci were never isolated from non-infected sternal wound hardware using standard culture methods (data not shown).

Discussion
Post-sternotomy wound infections are classified into superficial and deep. Superficial sternal wound infections (SWI) are confined to the skin and/or subcutaneous tissue with overall good response to antimicrobial therapies and local wound care. On the other hand, deep SWI includes, besides the superficial wound infection, sternal osteomyelitis with or without infection of the retrosternal structure [2]. Despite being uncommon, deep SWI is a life threatening complication after cardiac surgery with associated high mortality (10 -40%) [2,5,6]. Antimicrobial therapies alone usually fail to successfully treat deep SWI, which necessitates adding physical therapies such as surgical debridement, vacuumassisted closure, rigid sternal fixation, and flap reconstruction [19]. These interventions are highly limited in productivity as the incidence of mortality in these case remains high [20,21]. In particular, non-responsiveness of post-sternotomy deep wound infections to broad spectrum antibiotics remains a major clinical challenge [19,22].
In the majority of SWI clinical reports, microbiological analysis revealed wound colonization with staphylococcal strains (methicillin-sensitive Staphylococcus aureus, methicillin-resistant Staphylococcus aureus, Staphylococcus epidermidis) [7,8,9,23,24]. Staphylococcus strains are known for their capability to form robust biofilms on exposed tissues or biomaterials surfaces [10,25]. Staphylococcus species are the most important pathogen responsible for biofilmassociated medical devices infection [18,26,27]. Staphylococci biofilms have been identified on intravascular catheters, endocardial pacemaker lead, vascular grafts, mechanical heart valves, orthopedic implants, and ventriculo-peritoneal shunts [28,29,30,31,32,33]. We noted that MRSA clinical isolates were able to accumulate on the wires and grow as three dimensional aggregates of cocci encased in an amorphous extracellular material. These MRSA aggregates on the wires displayed resistance to tobramycin compared to planktonic isolates. These data are consistent with previous report where Olsson et al. compared the adherence ability of staphylococcal clinical isolates to sternal fixation stainless steel wires in vitro [34]. They reported no difference in adherence and attachment between coagulase negative staphylococci isolated from deep SWI and contaminants of non-infected re-sternotomy wounds. However, accumulation as biofilms on the wires were more frequently observed in deep SWI isolates than in contaminants [34].
Following clinical diagnostic criteria of biofilm associated infections as proposed by Parsek and Singh were evaluated in this study: (a) Infecting bacteria were adherent to some substratum or are surface associated; (b) direct examination of infected tissue showed bacteria living in cell clusters, or micro colonies, encased in extracellular matrix; (c) the infection confined to a particular location although secondary dissemination is possible and (d) antibiotic resistance despite the fact that the responsible organisms are susceptible to killing in the planktonic state [35]. Scanning electron microscopy detected three-dimensional aggregates of cocci attached to the wound tissues and stainless steel wires. Confocal laser scanning microscopy helped visualize thick layers of three-dimensional staphylococci aggregates distributed throughout the debrided tissue sections. The infection was localized to the sternotomy wound site with negative blood cultures and no signs of systemic involvement except in one patient. All patients were scheduled for surgical debridement after unsuccessful extended rounds of different classes of broad spectrum antibiotics to control the infection. Matching these clinical findings and the in vitro work with the clinical diagnostic criteria mentioned above, we conclude that deep SWI in the six patients of the test arm is a biofilmassociated infection.
Biofilm infected-wounds represent a real clinical challenge because of the complex nature, in which different bacterial species act synergistically to survive together [36,37,38]. Not only are these populations resistant to the host immune response and antibiotics, but also many bacteria inside biofilms are difficult to be identified by standard culture methods [39,40]. This may explain why only 2 out of 6 subjects had positive cultures report using standard clinical diagnostics, while staphylococci biofilm was detected by histological analysis in all of these patients. This observation indicates a reason for the failure of standard antimicrobial therapies to control the infection suggesting a comprehensive biofilm directed strategy is required to manage such infections. Although debridement decreases the bacterial burden, removes devitalized tissues (nutrient sources for bacteria), and enhances the immune response (by improving microcirculation), it fails to completely eradicate all pocket-containing bacteria within the wound [41,42]. Concerns have been raised regarding the efficacy of debridement alone against wound-associated biofilm infection [43]. Debridement also increases the wound size attenuating previous re-epithelialization and wound closure progression. Bacterial colonies, missed by debridement, can invade the wound bed again, reattach, and establish biofilm communities within 48 hours post-debridement [44]. Current medical treatment of SWI includes empiric antibiotic therapy with broad coverage that will be changed to specific therapy according to cultures [19]. There is almost a need of surgical interference, along with the medical therapy, to overcome the infection. According to our findings in the current study, the resistant nature of SWI to medical therapy is attributed to biofilm-associated infection. We suggest that combined therapies are needed to improve the outcomes. Debridement to be applied early to reduce the microbial load, removes necrotic tissues and foreign bodies, and physically disperses complex bacterial communities within the wound. This will force bacterial cells to enter a growing phase and actively search for a new site for reattachment. During which, a combination of anti-biofilm agents and newly developed antibiotics such as Daptomycin or Linezolid should be administrated [22,45,46]. Molecular analysis of the wound microbiology is needed to target other uncultivable pathogens might be involved in deep SWI. Additionally, coating of sternal fixation wires with anti-biofilm agents should be considered for further research. Introduction of anti-biofilm agents as part of wound care was shown to enhance the healing of chronic wounds in a recent clinical study [47]. One of the anti-biofilm agents, ribonucleicacid-III-inhibiting peptide (RIP), has received significant attention regarding staphylococcus infections [48]. RIP is a regulator molecule of quorum sensing in staphylococcus that was shown to suppress genes responsible for toxins production and biofilm formation in experimental infection models [48].
In summary, this work provided the first direct evidence demonstrating biofilm in deep SWI. Introduction of biofilm concept in deep SWI will help revisit management strategies. New classes of drugs have been developed to target different steps of biofilm formation: attachment, extracellular matrix, cell-to-cell communication (quorum sensing), and biofilm dispersion [49,50]. Consideration of such anti-biofilm strategies to manage deep surgical wound infection in cardiac surgery patients is warranted.

Materials and Methods
In vitro Biofilm Culture on Surgical Steel Wires USA300 MRSA cultures [16] were grown in tryptic soy broth (TSB) for 18 h to prepare an inoculum. The 18 h culture was diluted in fresh TSB to an optical density (OD 600 ) equal to 0.1 and inserted into replicate wells of a 12-well plate. Sterile wires commonly used for sternal fixation after cardiac surgery were twisted and cut into 15 mm sections and placed into each of the wells containing the USA300 inoculum. The plates were incubated at 37uC and 5% CO 2 on an incubator for 24 h. Following the growth period, planktonic cells were pelleted at 20,000xg and washed using PBS while wires were dipped into fresh PBS for washing. Planktonic cells and wire-associated cells were treated with 10 mg/ml tobramycin for 2 h. Following treatment, cells were sonicated for 10 minutes in a bath sonicator. The suspensions were serially diluted and plated. Tobramycin untreated planktonic or wire-associated cells were used as controls and each case the untreated colony forming units (CFU) were approximately 10 6 CFU/ml. Percent survival was calculated using log (treated CFU)/log (untreated CFU)x100. Human Subjects and Sample Collection Subjects (n = 6) participating in the study were patients scheduled for debridement procedure of an infected sternal wound after cardiac surgery or a re-sternotomy procedure for a second cardiac surgery with healed non-infected sternal wound (control, n = 3). All procedures were done at the Wexner Medical Center of The Ohio State University. All approvals were obtained as required. Our study protocol was approved by the Biomedical Sciences Institutional Review Board at The Ohio State University and all participants provided informed written consent. Demographic characteristics of patients and wound related information are listed in Table 1. Protocol was approved by the Ohio State University's Institutional Review Board. Samples were obtained from individual subjects during the debridement or re-sternotomy procedure. The upper part of the debrided tissue, representing the surface of the wound bed, was kept in a fixative solution for imaging by SEM. The rest of the tissue sample was kept in 4% formalin or immediately embedded in optimum cutting temperature (OCT) compound and stored frozen in liquid N 2 for histological analyses. Stainless steel wires, commonly used for closing of sternum, were collected in a fixative solution for imaging by SEM.

Wound Culture Clinical Procedures
Wound cultures were done in Ohio State University Wexner Medical Center Clinical microbiology laboratory using standard culture procedures. In brief, wound specimens were plated on selective media for culture and examinations. The identification and susceptibility testing on following was performed: i) gram negative rod; ii) s. aureus; iii) iii) enterococcus.

Imaging
Fluorescence stained sections were imaged using a Zeiss Axiovert 200 inverted fluorescent microscope supported by an AxioCam digital camera, a motorized stage and guided by Axiovision software (Zeiss, Thornwood, NY). Mosaic images of Gram/Twort or immunofluorescence stained debrided tissues were collected.

Confocal Scanning Laser Microscope Imaging
The biomass of staphylococci clumps colonizing the debrided tissues was visualized using an Olympus Fluoview FV1000 spectral confocal microscope (Olympus, Pittsburgh, PA) under 600X magnification using an argon laser. Z-stack images were created by merging serial scans of thick tissue section (20 mm). Threedimensional orthogonal projections of the z-stack images were generated by the Fluoview FV1000 software.

Scanning Electron Microscope (SEM) Imaging
Debrided tissue samples and stainless steel wires were fixed in a 2.5% glutaraldehyde solution in 0.2 M phosphate buffer for 3 days. On the 4th day, samples were washed with 0.1 M phosphate buffer and then dehydrated using graded concentrations of ethanol. Samples were washed with hexamethyldisilazane (HMDS, Ted Pella Inc., Redding, CA) and left to dry overnight. Before scanning, samples were mounted and coated with gold. A FEI TM NOVA nano SEM (FEI TM , Hillsboro, OR) equipped with a field-emission gun electron source was used for imaging.

Acknowledgments
We thank the laser capture Molecular Analysis (LCM) facility and Brian Kemmenoe (Campus Microscopy & Imaging Facility, OSU) for assistance with CLSM and SEM imaging.