Miniaturized Oligonucleotide Arrays : A new tool for discriminating colonization from infection due to Staphylococcus aureus in diabetic foot ulcers

Between 1 January 2004 and 30 April 2006, we prospectively enrolled all diabetic patients admitted to our department with any type of foot ulcer. Patients were included if they had never been treated by antibiotics (first episode) or if systemic antibiotic treatment was stopped at least 6 months before the time of onset of the current episode (recurrent wound). Using the IDSA clinical criteria (5) recently validated by Lavery et al. (13), wounds were considered either colonized (grade 1) or infected (grade ≥2). After wound debridement, samples for bacterial culture were obtained by swabbing the wound base, by needle aspiration, or by tissue biopsies and immediately sent to the bacteriology department. To minimize bias, laboratory technicians were kept blind to the clinical data. Only patients with monomicrobial culture positive for S. aureus were retained for inclusion in the study. Patients with grade 1 ulcers did not receive antibiotic treatment but were closely followed up to definitively establish the wound status (infection/colonization) at 1 month after inclusion and until 6 months thereafter. If any worsening was noticed by the patient, they were instructed to come immediately in our outpatient department. If this evolution appeared during the first 3 weeks, the wounds were considered to have a rapid worsening evolution. If the symptoms were observed after 3 weeks, the wounds were considered in a slow worsening evolution. In case of ulcer healing (complete re-epithelialization), a microbiological specimen was obtained 1 month later. Otherwise, a sample was obtained during the worsening phase of the ulcer, and ulcer grade was updated. Ulcer healing was definitively assessed at 6 months after inclusion.

Genus, species, and antibiotic susceptibilities were determined using the Vitek 2 card (BioMérieux, Marcy-l'Etoile, France). Strains were classified as antibiotic sensitive, intermediately resistant, or resistant, according to the recommendations of the Antibiotic Committee of the French Society for Microbiology (14). Sensibility to methicillin was screened by agar diffusion test using cefoxitin (30 μg) disks (BioRad, Marnes La Coquette, France) (14).

Each S. aureus strain collected during the study was analyzed in the INSERM laboratory. Genomic DNAs were prepared as previously described (12). A primer elongation reaction was used before hybridization with the DNA array. A 20-μl aliquot of this labeled sample was then transferred into the ArrayTube (ClonDiag, Jena, Germany) and incubated for 60 min at 50°C. The staining reaction was performed according to the manufacturer's instructions. The ArrayTube was then placed into the ATR 04 reading device (Ivagen, Bernis, France). The IconoClust software was used according to the manufacturer for recording and analyzing the arrays. Signal intensity and local background were measured for each probe position. Intensity of local background was subtracted from those of the automatically recognized spots, and averages for all results of a given probe were calculated. Resulting values <0.1 were considered as negative, >0.3 as positive, and between 0.1 and 0.3 as equivocal. The studied virulence genes encoded for the main toxins produced by S. aureus during cutaneous infections included: leukocidins (lukF, lukS, lukS-PV, and lukF-PV), enterotoxins (sea, seb, sec, sek, and seq), exfoliatins (etA and etB), and toxic shock syndrome toxin (tsst). Validation was performed using negative (Staphylococcus epidermidis strains) and positive (S. aureus, where PCRs were developed for every target included in the present study) controls. Validation of mecA genes and virulence genes was obtained by PCR using standard in-house PCR (mecA) and previously published methods (12,16). Finally, arrays results and clinical data were compared and analyzed by two independent experts.

For each qualitative variable, comparisons between grade 1 and 2 ulcers and between grade 1 and 2 and grade 3 and 4 ulcers were assessed using the χ² test and Fisher's exact test. A P value ≤0.05 was considered statistically significant. Analysis was performed with SAS software, version 9.1 (Cary, NC).

From 338 selected patients, 72 were included; in 38 patients the wound was the first episode of ulceration, whereas in 34 it was recurrent. The distribution and characteristics of the included patients are shown in Fig. 1 and Table 1. Twenty-two wounds (30.6%) were classified as grade 1 and followed up for 6 months, 21 ulcers as grade 2 (29.2%), 20 as grade 3 (27.8%), and 9 as grade 4 (12.5%). Thirty-five (48.6%) of the 72 S. aureus we isolated were methicillin resistant (MRSA), and 37 (51.4%) were methicillin sensitive. All recurrent ulcers were positive for MRSA. Only one ulcer appearing as the first episode was positive for MRSA.

During the follow-up period, 8 grade 1 ulcers healed (36.4%), whereas 14 worsened. Two of 14 had a rapid worsening evolution (≤10 days) and 12 of 14 a slow worsening evolution (≥27 days). Two healing ulcers had samples that remained positive for S. aureus compared with 11 of the 14 nonhealed ulcers. During the whole study, 85 S. aureus were isolated (72 on the initial cultures and 13 during the follow-up period): 48 were methicillin-sensitive S. aureus and 37 MRSA (43.5%).

The prevalence of resistance and virulence determinants are summarized in Table 2. Results of the resistant genes completely agreed with conventional susceptibility data (data not shown). All MRSA isolates were detected by cefoxitin test and were positive for mecA by both PCR and arrays. ermC and ermA macrolide resistance genes were found in 32.9% and 27.1% of the isolates, respectively. The most prevalent aminoglycoside resistance gene was aadD (23.5% of the isolates). The tet, mupR, and dfrA genes were uncommon (<7% of the isolates). No van genes were detected, in agreement with the in vitro susceptibility data. No significant trend in distribution was observed along the different grades, except for the aadD gene, which was more frequently associated with severe grades (2–4) (P = 0.04). A trend was also observed for a higher prevalence of mecA with increasing severity of infection; nevertheless, the difference was statistically significant only for grade 1 compared with grade 4 (P < 0.05).

Concerning virulence genes, enterotoxins were detected in 43.5% of the isolates: sea (43.5%) and sec (10.6%) genes were the most prevalent. lukF and lukS leukocidin genes were found in 68.2% and 63.5% of the isolates, respectively. Interestingly, genes encoding Panton-Valentine leukocidin, toxic shock syndrome toxin, and exfoliatin A toxins were found in 1.2, 1.2, and 4.7% of the isolates, respectively. lukF, lukS, sea, and sek were significantly more frequent in grade 2 ulcers compared with grade 1 (P < 0.001).

On admission, S. aureus cultured from 20 of 22 grade 1 ulcers showed no virulence genes, whereas these genes were present in all but one S. aureus isolated from grade 2–4 ulcers. The enterotoxin A gene (sea) was significantly associated with severe grades (3 and 4) (P < 0.001).

At follow-up, S. aureus was isolated from 13 grade 1 ulcers (Table 3). S. aureus from the two healing ulcers were of a different genotype profile compared with the initial sample, only due to difference in resistance genes. However, in both cases, no virulence genes were detected in initial or in follow-up samples. Two of the 11 worsening ulcers showed a rapid deterioration. Both genotype profiles of isolated S. aureus were unchanged. Unlike strains from other grade 1 ulcers, they carried virulence genes. S. aureus from slowly worsening recalcitrant ulcers showed genotype profiles completely different from those at inclusion, with appearance of many virulence genes in every case (Fig. 1).

Our study suggests that DNA arrays may be a useful tool for rapidly determining S. aureus genotype profiles and differentiating infection from colonization. Over the last few years, different DNA array technologies have been developed, but most systems are very expensive, time consuming, technically demanding, and difficult to adapt to the needs of clinical screening, restricting their use to research laboratories (12,16). In contrast, the miniaturized oligonucleotide array described in this study is relatively easy to perform; the use of tube-integrated arrays (Lab-on Chip system) with nonfluorescent labeling and a rapid hybridization protocol is time saving, easy to use and interpret, and allows large numbers of samples to be analyzed (12). An isolate, once cultured, can be processed within 1 day.

This study confirmed the correlation between results for arrays and PCR as previously demonstrated (12). Nearly one-half of S. aureus we isolated were methicillin resistant. This high-prevalence rate is likely attributable to the high proportion of patients with recurrent ulcers: all were previously infected and had been already treated with antibiotics during a preceding hospitalization, a well-established risk factor for selecting antibiotic-resistant organisms, especially in a diabetic foot clinic (2,4,5,8). However, in grade 1 ulcers, the prevalence of MRSA was limited (29.2%), ruling out the possibility for a clonal transmission of MRSA to explain our results. Finally, an additional interest of the array technology was its ability to identify Panton-Valentine genes; these genes coding for a cytotoxin are thought to be a major threat in severe tissue necrosis (17–21). Interestingly, these genes were present in a strain from a deep ulcer (grade 4).

The main result of our observational study is to show that the microarray technology enables clinicians to distinguish grade 1 from grade 2 ulcers, as the former generally displayed a very low level of virulence genes. Moreover, this tool allows us to predict wound outcome. Indeed, the two grade 1 ulcers with strains carrying virulence determinants had a rapidly worsening course, whereas the strains isolated in the initial and the follow-up samples showed exactly the same genotype profile, suggesting that the isolates were the same and that the wounds were actually infected—not only colonized. According to the genotype profile of S. aureus isolated for grade 1 ulcers, an adapted management of diabetic foot ulcers might be proposed. Furthermore, the array results provide rapid data on S. aureus in vitro susceptibility, notably the presence of the mecA gene, an important therapeutic concern.

One of the main limitations in this study is the small number of grade 1 ulcers we followed-up. Due to the low number of patients consulting for grade 1 ulcer with S. aureus as the sole pathogen and the low level of grade 1 wounds with rapid worsening, a large recruitment would be difficult to obtain. However, a large-scale study must be developed to definitively validate this tool. Moreover, the proposed array could be easily expanded to additional target genes and also adapted for testing other main pathogens isolated in diabetic foot ulceration (Enterobacteriaceae, Pseudomonas spp., or streptococci). Thus, this clinical platform appears suitable for use under routine conditions in a microbiology laboratory.

In conclusion, the increasing prevalence of staphylococci resistance and the small number of new antimicrobial drugs must stimulate the discovery of new solutions in the near future. The miniaturized oligonucleotide arrays may be an interesting tool in management of diabetic foot ulceration, allowing an earlier discrimination of infection from colonization of wounds and a more adequate usage of antibiotic treatment.

This work was supported by the Coloplast Foundation, the French Speaking Association for Diabetes and Metabolic Diseases (ALFEDIAM-Aventis grant), the Fondation pour la Recherche Médicale Languedoc-Roussillon-Rouergue, and l'Université Montpellier 1 (Bonus Qualité Recherche).

We thank all of the participants of the Department of Diabetology for help in recruiting patients.

This study was presented at the 26th Interdisciplinary Symposium on Anti-Infectious Chemotherapy (RICAI), Paris, France, December 2006.