Development of a Novel and Rapid Antibody-Based Diagnostic for Chronic Staphylococcus aureus Infections Based on Biofilm Antigens

Prosthetic joint infections are difficult to diagnose and treat due to biofilm formation by the causative pathogens. Pathogen identification relies on microbial culture that requires days to weeks, and in the case of chronic biofilm infections, lacks sensitivity. Diagnosis of infection is often delayed past the point of effective treatment such that only the removal of the implant is curative. Early diagnosis of an infection based on antibody detection might lead to less invasive, early interventions.

ABSTRACT

Prosthetic joint infections are difficult to diagnose and treat due to biofilm formation by the causative pathogens. Pathogen identification relies on microbial culture that requires days to weeks, and in the case of chronic biofilm infections, lacks sensitivity. Diagnosis of infection is often delayed past the point of effective treatment such that only the removal of the implant is curative. Early diagnosis of an infection based on antibody detection might lead to less invasive, early interventions. Our study examined antibody-based assays against the Staphylococcus aureus biofilm-upregulated antigens SAOCOL0486 (a lipoprotein), glucosaminidase (a domain of SACOL1062), and SACOL0688 (the manganese transporter MntC) for detection of chronic S. aureus infection. We evaluated these antigens by enzyme-linked immunosorbent assay (ELISA) using sera from naive rabbits and rabbits with S. aureus-mediated osteomyelitis, and then we validated a proof of concept for the lateral flow assay (LFA). The SACOL0688 LFA demonstrated 100% specificity and 100% sensitivity. We demonstrated the clinical diagnostic utility of the SACOL0688 antigen using synovial fluid (SF) from humans with orthopedic implant infections. Elevated antibody levels to SACOL0688 in clinical SF specimens correlated with 91% sensitivity and 100% specificity for the diagnosis of S. aureus infection by ELISA. We found measuring antibodies levels to SACOL0688 in SF using ELISA or LFA provides a tool for the sensitive and specific diagnosis of S. aureus prosthetic joint infection. Development of the LFA diagnostic modality is a desirable, cost-effective option, potentially providing rapid readout in minutes for chronic biofilm infections.

INTRODUCTION

Surgical site infections (SSIs), such as prosthetic joint infections (PJIs), are a major cause of prolonged hospitalization, increased health care costs, morbidities associated with additional clinical procedures, and mortality (1–5). Staphylococcus aureus and coagulase-negative staphylococci are responsible for the majority of SSIs (6). Due to the rapid development and acquisition of multiple antibiotic resistance markers, as well as the propensity to change from an acute to a chronic and recurrent infection, S. aureus has reemerged as an important human pathogen. An inherent mechanism by which S. aureus persists in the host is through biofilm formation. A biofilm is a sessile community of microorganisms attached to a substratum, interface, or each other and embedded in a microbe-derived matrix of extracellular polymeric substances (EPSs). Here, they exhibit an altered phenotype with respect to growth, gene expression, and protein production (7). Delays in the diagnosis of PJIs can be significant due to the lag between colonization, biofilm formation, presentation of signs and symptoms of an inflammatory response to biofilm infection, and, ultimately, appropriate diagnosis. As this time interval increases, the difficulty of treatment rises, and these deep infections commonly cannot be managed without surgery (8).

The ability of current diagnostic tests to detect biofilms before clinical symptoms develop is inadequate. General host response mechanisms, such as an elevated white blood cell count, are indicators of infection but are not specific enough to render a diagnosis nor target a treatment to a surgical site or implant. Synovial biomarkers of inflammation used to diagnose suspected PJI include elevated C-reactive protein (CRP), leukocyte esterase, alpha-defensin, human beta-defensin-2 (HBD-2), HBD-3, and cathelicidin LL-37 (9–11). However, these biomarkers fail to identify the cause of inflammation (infectious or noninfectious) and, if infectious, the microbe(s) responsible for the inflammation (12, 13). In addition, traditional microbial techniques of culturing intraoperative purulence or wound swabs on agar are unreliable, untimely, or ineffectual for cultivating biofilms (14, 15). An improved method for identifying a microbe is through biopsy and culture, which could potentially miss <0.1-mm³ biofilm aggregates (16). Imaging technology, such as X ray, computed tomography (CT) scans, and magnetic resonance imaging (MRI), can provide the exact location of infection, but lacks the ability to identify the causative agent of infection (17–22).

The advent of molecular techniques based upon PCR or protein-based mass spectrometry (MS) platforms have increased sensitivity in the species-level identification of pathogenic microbes, including in cases of PJI, and were heralded as major advancements (23–25). However, the sensitivity of PCR diagnostics has created problems, mainly false positives due to poor quality control and exogenous, contaminating DNA (26–28). A positive PCR result may lack clinical significance, as samples from sterile body sites that lack clinical signs of pathology can be positive by PCR (29). Additionally, 10% to 40% of the global population is colonized by S. aureus, providing the potential for misdiagnosis when insufficient skin-disinfecting procedures are employed. Therefore, there is a continued and imperative need for innovative culture-independent tests that will accurately and efficiently identify biofilm infections before clinical symptoms present and persist. To that end, the objective of the study was to develop rapid serology-based assays that detect and quantify antibodies produced against previously identified S. aureus genes with upregulated expression in a biofilm mode of growth (30).

MATERIALS AND METHODS

Rabbit model of osteomyelitis.

Osteomyelitis was initiated in New Zealand White female rabbits (Charles River Laboratories, Wilmington, MA) by injecting M2, a sequence type 30 (ST30), spa type T019, agr III methicillin-resistant S. aureus (MRSA) strain, which was isolated from an osteomyelitis patient into the intramedullary space of the tibiae as described previously (30–36). Although this model attempts to replicate conditions within host tissue, it does not contain a foreign device or simulate medical conditions or tissue damage that cause devascularization and, thereby, inhibit proper infiltration of host immune cells. To compensate for these factors, an infectious dose higher than the physiologic level of bacteria that initiate human infection is required. Separate infection studies were completed to examine (1) the humoral response at chronic stage osteomyelitis versus naive and (2) the kinetics of IgG production throughout infection versus after resolution. In the preliminary study, three rabbits were infected with MRSA and the infection was allowed to progress to day 42. Serum samples were collected on day 0 (naive) and day 42 postinfection (chronic stage), and bone cultures on day 42 confirmed chronic infections. In the second study, five rabbits were infected with MRSA, and the infection was allowed to progress for 14 days. Radiographs confirmed osteomyelitis. Vancomycin (80 mg/kg of body weight) was administered twice a day (BID) for 2 weeks, and the rabbits were subsequently housed for another 7 weeks. On day 78, rabbits were euthanized, and bone cultures confirmed S. aureus clearance. Serum samples were collected prior to infection and then at 10, 29, 43, and 78 days postinfection. Animal studies were performed as approved by the Institutional Animal Care and Use Committee and under the supervision of the Veterinary Resource Staff at the University of Maryland, Baltimore, MD.

Purification of recombinant biofilm-specific proteins.

We previously characterized and described MRSA biofilm-specific proteins that were immunogenic in the rabbit model of osteomyelitis and upregulated in the biofilm mode of growth in vitro (30). To develop a biofilm-specific assay, recombinant proteins for select S. aureus in vivo-expressed proteins (SACOL0486, glucosaminidase, or SACOL0688) were produced and purified as described previously (31, 37).

Enzyme-linked immunosorbent assay.

ELISAs assessed the antibody titers elicited against SACOL0486, glucosaminidase, and SACOL0688 during S. aureus-mediated osteomyelitis in rabbits. Microtiter plates were coated with 0.3 μg of protein (SACOL0486, glucosaminidase, or SACOL0688) per well in a coating buffer (32 mM Na₂CO₃ and 68 mM NaHCO₃) and incubated overnight at 4°C. The coating buffer was discarded, and then the wells were washed 3 times with wash buffer (phosphate-buffered saline [PBS] with 0.4% Tween 20) and blocked with 0.2 ml/well of blocking buffer (PBS with 0.1% bovine serum albumin [BSA] and 0.02% Tween 20) for 1 h at room temperature (RT). The blocking buffer was discarded, serum samples were added, and 2-fold serial dilutions in blocking buffer were performed for each serum sample from 1/10 to 1/1,280 (in duplicate). The plates were incubated for 1 h at RT and then washed three times with wash buffer. In each well, 0.05 ml of a 1/1,000 dilution of anti-rabbit-horseradish peroxidase (HRP) antibody (Pierce, Rockford, IL) was added, and the plates were incubated for 1 h at RT. The wells were rinsed 3 times with wash buffer. Finally, 0.05 ml of the chromogenic substrate (10 mg 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic-acid [ABTS] dissolved in 10 ml citrate/phosphate buffer) and 0.1 ml H₂O₂ was added to each well and incubated for 10 min at RT. Absorbance values were read at 450 nm using an Opsys MR microplate reader (Dynex, Chantilly, VA). A two-sample paired t test was performed for each set of serum dilutions to determine if there was a significant difference (P < 0.05) between absorbance values for naive serum and infected serum. SACOL0688 ELISAs demonstrated sensitivity in detecting S. aureus infection compared with SACOL0486 and glucosaminidase and were subsequently used to evaluate the synovial fluid (SF) specimens. ELISAs were performed as previously described for the serum samples from rabbits with the following alterations. Microtiter plates were coated with 0.5 μg/well SACOL0688. After the wash and blocking steps, 10-fold serial dilutions were performed for each synovial sample (in duplicate) from 1/10 through 1/10,000. Incubation with the secondary antibody was performed using a 1/1,000 dilution of anti-human-HRP antibody (Pierce). Patients were designated infected with S. aureus if the absorbance values were found to be three standard deviations (SDs) above the mean (38).

Lateral flow assay.

The lateral flow assays (LFAs) were made using 0.5-cm by 5-cm strips of nitrocellulose impregnated with sequential lines of protein A-gold conjugates, recombinant S. aureus antigen (test line), and anti-protein A antibody (control line). A 1/200 dilution of protein A conjugated to colloidal gold (courtesy of Shang Li) was applied near the end of the sheet; then 0.1 to 1.0 mg/ml of SACOL0486, glucosaminidase, or SACOL0688 was applied 3 cm from the end; and a 1/5 dilution of anti-protein A antibody (Biomeda, Foster City, CA) was applied a further 1 cm from the test line. Serum was diluted (1/100) in 0.2 ml of running buffer (50 mM HEPES, 0.35% BSA, and 0.1% polyethylene glycol [PEG] [pH 7.4]) and applied to the nitrocellulose strip. Naive (day 0) and infected (day 42) sera from three rabbits were run in triplicate. Sera from healthy human subjects served as negative controls for exposure to the antigens. Blood was collected from two consenting human donors in accordance with the Human Subjects Institutional Review Board (IRB) at Northern Arizona University. Each assay was allowed to run for 10 min, and results were recorded as positive if two lines were detected or negative if only the control line appeared. For an overview of the lateral flow assay procedure, see Fig. 1.

FIG 1.

LFA overview. Serum is applied onto an absorbent pad and then transferred using running buffer onto a conjugation pad. Sera or synovial IgGs conjugate to protein A-gold, and then labeled IgGs migrate to the test line (T) that contains a species-specific in vivo-expressed antigen. If the patient is infected with that microbial species, labeled IgG will adhere to the test line, producing a red stripe, thereby indicating infection. Unbound labeled IgG then migrates with the fluid front to the control line (C) containing an anti-protein A antibody, where remaining protein A-gold conjugates bind and appear as a second red stripe, confirming proper function.

Patient samples.

Patients undergoing primary or revision knee or hip arthroplasty from 2009 to 2014 at the Rothman Orthopaedic Institute were eligible for recruitment into this study. Intraoperative SF specimens were collected from patients during revision total joint arthroplasties. Revision patients had undergone routine institutional standardized infection workups. There is not a single “gold standard” for diagnosis of PJI. The Rothman Orthopaedic Institute uses the multiple criteria for the diagnosis of PJI, defined by the International Consensus Meeting (39, 40). Major criteria include two positive cultures of the same microbe or a sinus tract associated with joint or prosthesis, while minor criteria include elevated host immune cell counts and enzyme levels in the serum and/or synovial fluid. Preoperative diagnosis of infection is based upon screening serum (elevated CRP or erythrocyte sedimentation rate [ESR]) and/or synovial samples (elevated white blood cell [WBC], polymorphonuclear neutrophil (PMN), CRP, or positive for alpha-defensin) and scoring each criterion. The cumulative score defines the patient as infected, possibly infected, or not infected. In cases of inconclusive preoperative evaluation, diagnosis of infection is delineated by histology, culture, and purulence findings from intraoperative specimens. Microbiological cultures were cultivated for 48 to 72 hours. The Rothman Orthopaedic Institute routinely records culture results for PJI specimens after 48 to 72 hours and typically only maintains culture plates through day 14 for shoulder and elbow cases under suspicion of infection. Standard practices at other institutions routinely retain cultures for 7 to 14 days to detect fastidious pathogens; therefore, the short culture may have missed fastidious organisms. SF samples were further analyzed by Ibis T5000 technology (select samples listed in Table S1).

Another set of 10 SF samples isolated during 2009 to 2014 at the Rothman Orthopaedic Institute was utilized in this study. These specimens were only evaluated by microbiological culture and were not analyzed by Ibis T5000 technology (see Table S2 in the supplemental material). These samples were selected based on S. aureus-positive culture and sent to the University of Maryland, Baltimore, MD, without corresponding culture data. Culture data were provided upon completion of diagnostic testing.

Molecular analyses with Ibis T5000.

DNA was extracted from 1-ml aliquots of SF using the DNeasy blood and tissue kit (Qiagen, Valencia, CA) as instructed by the manufacturer. Bac (bacteria, antibiotic resistance, and Candida sp.) detection plates (Abbott Molecular, Des Plaines, IL) were used to amplify 16S/18S rRNA amplicons from 0.01-ml DNA aliquots for each specimen, and PCR products were then analyzed using the Ibis T5000 system as described previously (41–44). The Ibis T5000 analysis demonstrated good reliability in staphylococci diagnosis, as this technology showed complete concordance with routine microbiological testing for staphylococci in 87 patient samples (25). The present study used 30 of those samples to test the specificity of the SACOL0688 diagnostic (Table 1; see Table S1 in the supplemental material).

TABLE 1.

Description of the 30 patients with infectious etiologies defined by Ibis T5000 technology^a

Patient	Procedureb	Reason for procedure	ESR (mm/h)	CRP (mg/dl)	WBC (103 cell/μl)	PMN%	Culture result
1	Total knee revision	Loosening	21	n/a	6.8	61.2	Negative
2	Total knee revision	Loosening	40	n/a	7.5	70.8	Negative
3	Total knee revision	Loosening	44	5.2	7.0	68.0	Negative
4	Total knee revision	Loosening	16	0.9	7.3	64.6	Negative
5	Total hip revision	Infection	34	2.5	8.8	58.4	Negative
6	Total knee revision	Stiffness	5	<0.5	8.5	66.5	Negative
7	Total knee revision	Instability	n/a	n/a	9.5	n/a	n/a
8	Total knee revision	Extensor mech	n/a	n/a	7.9	n/a	Negative
9	Total knee revision	Loosening	24	1.0	10.5	73.0	Negative
10	Total knee revision	Loosening	10	0.5	n/a	n/a	Negative
11	Total knee revision	Loosening	10	<0.5	7.2	62.5	Negative
12	Total knee revision	Loosening	10	<0.5	8.3	69.6	Negative
13	Total hip revision	Loosening	37	0.6	9.3	60.7	Methylobacterium species
14	Total knee revision	Aseptic loosening	14	<0.5	5.5	60.3	Negative
15	Total knee revision	Loosening	59	0.9	10.3	75.8	Negative
16	Total knee revision	Patella maltracking	73	3.1	11.4	69.1	Negative
17	Total hip revision	Loosening	21	0.49	6.3	63.7	Negative
18	Total knee revision	Loosening	70	1.0	5.0	68.4	Negative
19	Knee spacer insertion	Infection	74	7.3	3.7	57.3	Staphylococcus aureus X2
20	Total hip revision	Polyethylene wear	47	1.3	5.5	n/a	Negative
21	Total knee revision	Polyethylene wear	14	<0.5	7.3	56.3	Negative
22	Reimplantation	Reimplantation	25	0.6	6.2	60.3	Negative
23	Total knee (primary)		n/a	0.7	8.8	n/a	Negative
24	Total hip revision	Loosening	10	<0.5	6.3	52.8	Staphylococcus sp., coagulase negative
25	Total knee revision	Infection	45	<0.5	6.9	64.4	Staphylococcus lugdunensis
26	Knee spacer insertion	Infection	33	<0.5	6.8	52.9	Negative
27	Total hip revision	Pain	21	0.8	5.5	n/a	Negative
28	Total hip revision	Pain	18	<0.5	3.7	63.9	Negative
29	Total knee revision	Pain, stiffness	30	<0.5	7.8	55.8	Negative
30	Total hip revision	Pain, femoral fibrous ingrowth	20	<0.5	7.0	51.8	Negative

^aESR, erythrocyte sedimentation rate; CRP, C-reactive protein; WBC, white blood cell; PMN%, polymorphonuclear neutrophil percentage; n/a, not available.

^bAll procedures categorized as revision.

Statistical methods.

Calculations were made for means and standard deviations for any continuous variables and frequency distributions for categorical variables. The t test was used to compare means for continuous variables, and chi-square analysis was used for categorical variables. A P value of 0.05 was considered significant.

RESULTS

Diagnostic development.

To identify a strong candidate for detection of Staphylococcus aureus chronic infection, the antibody response to biofilm antigens (SACOL0486, glucosaminidase, and SACOL0688) in sera from naive and infected rabbits were analyzed by ELISA. Significant increases in the absorbance values postinfection were identified by a two-sample paired t test (Fig. 2; Table 2). ELISAs for each antigen demonstrated a significant increase in absorbance at the 1:10 dilution (P < 0.05), but a significant difference in SACOL0486 and glucosaminidase titers disappeared after the 1:40 and 1:10 dilutions, respectively. However, we found a >2.5-fold titer increase for all dilutions in the ELISAs with SACOL0688 (Fig. 2). The threshold for detection was calculated using the optical density at 450 nm (OD₄₅₀) SD of naive samples; three SDs above the naive mean defined a value of >0.260 at the 1:640 dilution in SACOL0688 ELISAs as “positive.”

FIG 2.

ELISAs for SACOL0486, glucosaminidase, and SACOL0688 comparing absorbance from day 0 (naive) to day 42 (postinfection). SACOL0486 and glucosaminidase display only slight differences, while absorbance of SACOL0688 exhibits >2.5× increase after infection.

TABLE 2.

Statistics of ELISAs evaluating differences between naive and infected rabbit serum samples for SACOL0486, glucosaminidase, and SACOL0688

Biofilm antigen	Two-sample paired t-test results on 1:10 dilution		Dilution factor of last significant difference
	t	P value
SACOL0486	5.81	0.0011	40
Glucosaminidase	3.32	0.0106	10
SACOL0688	21.4	<0.0005	1,280+

Infection kinetics.

Antibody titers for SACOL0688 were measured and corrected against naive IgG levels at multiple points postinfection and after vancomycin treatment in rabbits (Fig. 3). IgG titers rapidly increased during the 2 weeks postinfection until administration of an antibiotic. Antibody levels peaked at completion of vancomycin treatment (4 weeks postinfection), after which the infection resolved. Once the infection resolved, anti-SACOL0688 IgG levels began to decrease due to cessation of staphylococcal antigen presentation.

FIG 3.

Total anti-SACOL0688 IgG changes over time following S. aureus biofilm infection and subsequent resolution in the rabbit model, as measured by ELISA. Rabbits (n = 5) were infected with S. aureus and then treated with vancomycin (80 mg/kg) starting 14 days postinfection for 2 weeks. Serum was obtained prior to infection and then on 10, 29, 43, and 78 days postinfection. Serum IgG levels are expressed as absorbance minus background (naive serum).

LFA functionality.

An ELISA is a robust test providing quantifiable data, but the multistep process requires several hours, dedicated technician time, and many reagents. Therefore, we developed a simple diagnostic test based upon the LFA format (Fig. 1) using SACOL0486, glucosaminidase, and SACOL0688. A proof of concept for the LFA was evaluated using rabbit serum. Sensitivity and specificity percentages were calculated for each assay (Table 3). The SACOL0486 and glucosaminidase LFAs demonstrated sensitivity of 89% and specificity of 56%. In contrast, the SACOL0688 LFA demonstrated sensitivity and specificity of 100%. Of note, noninfected human sera reacted with SACOL0486 and glucosaminidase but did not cross-react with SACOL0688. Representative LFAs for each antigen are shown in Fig. 4.

TABLE 3.

Statistical measures for SACOL0486, glucosaminidase, and SACOL0688 LFAs^a

Biofilm antigen	No. of negative results		No. of positive results		Sensitivity	Specificity
	True negative	False negative	True positive	False positive	(%)	(%)
SACOL0486	5	1	8	4	89	56
Glucosaminidase	5	1	8	4	89	56
SACOL0688	9	0	9	0	100	100

^aLFAs performed in triplicate with each rabbit sample.

FIG 4.

LFAs with SACOL0486, glucosaminidase, or SACOL0688 at the test line and anti-protein A antibody at the control lines. Naive rabbit sera (0) were expected to be negative and sera collected during S. aureus infection (42) were expected to be positive, whereas sera from healthy human subjects were expected to be negative for S. aureus biofilm-specific antibodies.

Clinical validation.

To validate the effectiveness and to test the specificity of SACOL0688 as a diagnostic, ELISAs were performed by researchers who were blind to infection status with an initial set of SF samples (n = 30) (Fig. 5) from patients who underwent revision arthroplasty surgery. Microbiological culture and Ibis T5000 technology had already established the microbiological diagnosis. One sample (patient number 19) had an anti-SACOL0688 IgG titer greater than the positive threshold or three SDs above the mean sample absorbance at the 1:500 dilution (0.252 ± 0.43) (38). This analysis successfully identified the only patient with Ibis and culture-confirmed S. aureus among patients with PJIs mediated by other species (n = 25) or negative (n = 4) (Table S1). PJI cases caused by coagulase-negative staphylococci (n = 6) were clearly under the cutoff, except patient 25, who was infected with Staphylococcus lugdunensis that contains a close homolog to SACOL0688 and had an IgG titer near the positive threshold. Overall, the validation analysis demonstrated 100% specificity for PJI caused by S. aureus.

FIG 5.

ELISA quantification of anti-SACOL0688 IgG levels in synovial fluid of 30 patients with suspected PJI. Sample(s) are considered positive when the absorbance reading(s) measured above threshold (red line) of three SDs above the mean (0.252 ± 0.43) (38).

Although the data showed high specificity, in order to assess sensitivity, a second set of SF samples (n = 10) were analyzed in a blind manner by ELISA. These patient specimens were not affiliated with the aforementioned Ibis study. Microbiological culture identified S. aureus in all specimens (Table S2). Nine of the specimens had IgG titers above the positive threshold (data not shown), demonstrating a sensitivity of 90.9%.

DISCUSSION

Successful treatment of a biofilm infection is significantly affected by the timeliness of diagnosis. However, diagnosis can be hindered by the days-to-weeks delay and low sensitivity of microbial culture, as well as the inability of many current molecular techniques used to identify the causative agent of these infections. (8). These drawbacks are especially problematic when the infection involves a surgical implant, such as an infected joint replacement (e.g., PJI) (22, 27, 45). As a consequence, higher failure rates for treatment of PJI are reported following debridement with prosthesis retention compared with a two-stage surgical procedure because the former treatment often does not adequately remove a biofilm-based prosthesis infection (43). Therefore, any method that contributes to the early diagnosis of PJI and that identifies a biofilm-mediated infection has great potential to direct treatment in these difficult cases.

Prosthetic joint infections are classified by time to infection into early (within 1 to 3 months), delayed onset (after 3 months and within 12 to 24 months), or late onset (after 12 to 24 months). Acute prosthetic joint infections typically present with the presence of a draining sinus as definitive evidence of infection, but multiple pathogens can initiate early onset infections, including S. aureus, aerobic Gram-negative bacilli, and coagulase-negative staphylococci (CoNS). S. aureus can also mediate delayed onset PJI and is a primary agent of late onset PJI via hematogenous seeding from another nidus of infection. While S. aureus is a leading cause of PJI, differential diagnosis is necessary to identity the causative agent, and rapid diagnosis can direct appropriate antibiotic regimens until antibiotic susceptibility is determined over a few days. Our strategy aims to target a major causative agent of PJI and is not meant to be a comprehensive test to identify infected patients. Although homologs to the S. aureus antigen SACOL0688 are present in CoNS, the antibody response elicited against CoNS appears to have limited cross-reactivity with recombinant SACOL0688. Presumably, a negative result with our diagnostic modalities would not rule out infection mediated by other Gram-negative or Gram-positive bacteria; therefore, microbiological culture or other diagnostic techniques must be employed.

We developed rapid serology-based assays that detect and quantify antibodies against S. aureus, which avoid the need for pathogen isolation and provide rapid turnaround times. Furthermore, pathogen-specific diagnostics have greater utility over assays measuring general biomarkers of inflammation, where diagnosis of infection can be confounded by postsurgical inflammation or underlying inflammatory conditions (e.g., acute gout flare-up) and often fail to identify the causative agent. Host antibody levels against biofilm-specific proteins were measured in a rabbit model of osteomyelitis and in patients with PJIs using ELISA. Antibodies against SACOL0688 from rabbit sera demonstrated 100% sensitivity and 100% specificity, and antibodies from patient SF showed 91% sensitivity and 100% specificity. While the use of ELISA in the clinical diagnostic laboratory has become fairly standard, the use of lateral flow assays is a desirable option in the clinical setting because it avoids the need for pathogen isolation and provides rapid turnaround times. The LFA provides the utility of an ELISA with less time required (<10 minutes) and acts as a cost-effective ($10), specific (100%), and sensitive (>90%) diagnostic test. Based upon these results, we have identified a S. aureus protein with great potential as a diagnostic tool using multiple modalities. One limitation of the current diagnostic is our inability to differentiate between methicillin-resistant and methicillin-sensitive S. aureus strains, unlike PCR assays (29). Another limitation of this diagnostic is the inability to identify polymicrobial infections.

One concern with the diagnostic is the ability to detect acute and chronic infections in the host. In general, the host immune system recognizes the invading pathogen and generates an activated B cell response within 2 to 4 weeks of infection. Increased IgG titers against S. aureus antigens have been observed during acute and chronic infection. Verkaik et al. demonstrated variability in time to reach peak IgG response for multiple staphylococcal microbial surface components recognizing adhesive matrix molecules (MSCRAMMS) in patients following onset of bacteremia. Measurement of the IgG titers found the median time to peak titer level ranged from 15 to 28 days depending on the MSCRAMM (46). In these patients, the origin of bacteremia was catheter related or linked to preexisting osteomyelitis, which suggests the blood infection began after S. aureus dispersal from a biofilm on the catheter or bone matrix. The MSCRAMMS would have been downregulated during the biofilm model of growth on the catheter but induced during planktonic growth in blood. Importantly, den Reijer et al. reported that mean IgG levels for SACOL0688 are significantly increased in both bacteremia and osteomyelitis patients compared with healthy individuals (47).

Another concern is the potential for false positives due to circulating antibodies in S. aureus carriers. Dryla et al. evaluated the circulating IgG titer against 19 staphylococcal cell surface and secreted proteins in patients with acute infections compared with healthy adults. Overall, significantly higher IgG titers against SACOL0688 were observed in infected patients than in healthy adults, and similarly low levels of circulating antibodies were observed in healthy patients regardless of carrier status. Furthermore, longitudinal analysis of IgG titers against SACOL0688 found the circulating antibody levels remained stable over a period of 6 months in 8 healthy adults, but these levels were extremely low in 7 of the individuals (48). Presumably, the low anti-SACOL0688 IgG titers represent a baseline titer following resolution of a previous S. aureus infection. In other studies, IgG levels against the majority of staphylococcal proteins were similar in persistent, intermittent, and noncarriers, although high IgG levels against toxic shock syndrome toxin 1 (TSST-1) were found in individuals with persistent carriage (49, 50). Based on these studies, it appears that SACOL0688 expression and recognition by the host immune response, which is necessary to maintain a high level of circulating antibody, is only achieved during infection.

Our kinetic analysis of the humoral response during osteomyelitis found anti-SACOL0688 titers rapidly increase during the first 2 weeks of infection (Fig. 3). While S. aureus pathology persists over an extended period, the clinical signs of PJI can be delayed weeks to months (5, 51). Detection of anti-S. aureus antibodies within 2 weeks of initial colonization would allow for rapid diagnosis, possibly prior to clinical presentation. Early detection and therapeutic intervention with antimicrobial therapy alone can potentially resolve prosthetic implant infections without surgical removal of the infected joint (8, 52, 53). However, the feasibility of diagnosis before clinical signs needs to be confirmed in a clinical study that is under way. In contrast, anti-SACOL0688 titers declined following resolution since cessation of antigen presentation prevents further antibody production. We expect antibody titers to decline over time until a baseline level is reached. Therefore, we hypothesize that monitoring anti-SACOL0688 titers may allow clinicians to follow treatment outcome and infection resolution, thereby guiding surgical revision of prosthetic devices following infection, and must be further evaluated in patients treated for S. aureus-mediated infections. Future studies will further define the clinical efficacy of this assay; including its use as a noninvasive sentinel to track implant-associated infections commonly associated with orthopedic devices. This ability to diagnosis or monitor would be made dramatically easier using the LFA that was shown to be efficacious in this study (Fig. 1 and 4).

Serological assays provide the most utility when diagnosing biofilm pathogens because they are rapid and do not require invasive biopsy. Although the diagnostic modalities in this study target S. aureus, our discovery platform can expand the assay to diagnose additional microbial pathogens found in biofilm-mediated infections. The antigens presented by S. aureus in a biofilm mode of growth are different from those exhibited by its planktonic counterpart (54). Therefore, researchers can exploit the unique protein profile of biofilm microbes for antigen discovery. Future studies should include testing the diagnostic modalities against other S. aureus infections. In addition, further studies will determine the timing of a positive ELISA or LFA result following surgical placement of the prosthetic joint in clinical cases of PJI and when these measures become negative following successful therapy.

As awareness of the importance of biofilms in hospital and community-acquired illnesses increases in the health care community, a biofilm-specific diagnostic test for the most common pathogens, such as the one we have developed for S. aureus, will be the first step to direct appropriate treatment. As a result, the rapid diagnosis of biofilm infections could potentially diminish health care costs as well as the duration of patient hospital-stay and will holistically improve patient outcome.