SEARCH

SEARCH BY CITATION

Keywords:

  • Colonization;
  • diabetic foot ulcer;
  • EDIN ;
  • oligonucleotide array;
  • severe infection;
  • Staphylococcus aureus

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements and Funding
  8. Transparency Declaration
  9. References

Staphylococcus aureus is both a common colonizer of human skin and the most frequently isolated pathogen in diabetes foot infections (DFIs). The spread of DFI to soft tissue and bony structures is a major causal factor for lower-limb amputation. It is therefore of great importance to differentiate colonizing from infecting strains of S. aureus. Epidermal cell differentiation inhibitors known as EDIN and EDIN-like factors, a group of toxins targeting RhoA master regulator of the actin cytoskeleton, may confer virulence properties on S. aureus. In this study, for the first time, analysis of S. aureus strains, recovered in DFIs at an initial stage and during the follow-up, showed that 71.4% of edin-positive strains were associated with moderate-to-severe infections (grades 3 and 4 of the IDSA/IWGDF classification) compared with 28.6% of edin-positive strains associated with low-grade infections. Most of these strains were edin-B positive (86.7%) and belonged to CC25/28-MSSA (n = 10). One edin-B-positive ST152-MSSA strain was negative for the two highly prevalent predictive markers of infecting strains (lukDE and hlgv). Collectively, this points towards the edin-B encoding gene as a bonafide subsidiary predictive risk marker of DFI.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements and Funding
  8. Transparency Declaration
  9. References

Foot ulcers are common in diabetic patients, with a prevalence rate as high as 25% [1]. These ulcers frequently become infected, and spread of infection to soft tissue and bony structures is a major causal factor for lower-limb amputation [2]. Staphylococcus aureus is the most frequently isolated pathogen in diabetic foot infections (DFIs). Both the occurrence and progression of staphylococcal infection result from the action of a variety of virulence factors combined with the immunological status of the host [3, 4]. Bacterial factors are extremely heterogeneous in structure and mode of action, giving the bacteria specific metabolic and adhesive properties, as well as offering them a protection against the large body of innate immune effectors [3]. Defining the contributing role of virulence factors by combining cellular microbiological techniques and epidemiological approaches would be of interest in predicting S. aureus-associated risk of pejorative infections [5].

The epidermal differentiation factor (EDIN) and EDIN-like factors of S. aureus belong to such factors of well-determined mode of action, although their implication in staphylococcal infection remains ill defined. EDIN belongs to a family of bacterial exotoxins targeting the small host protein RhoA, which they inhibit [6, 7]. This small GTPase is a major regulator of the host cell actin cytoskeleton [8]. Several cell biology studies have revealed that inhibition of RhoA has a detrimental impact on the cohesion of epithelium and endothelium barriers, which is likely to favour bacterial dissemination [9, 10]. Inhibition of RhoA also impairs complement-mediated phagocytosis [11]. As a whole, a large number of studies aimed at investigating the impact of RhoA inhibition, point towards a contributing role of EDIN factors in bacterial colonization and host tissue invasion [5, 12]. Thus defining the distribution of edin-encoding genes in clinical isolates of S. aureus might help in predicting the risk factors for severe and spreading infection.

A previous multicentre study conducted on clinical isolates of S. aureus had suggested that two clonal complexes (CC) and an array of virulence factors can be used to discriminate uninfected from infected diabetic foot ulcers (DFUs) [13]. Indeed, we determined that colonizing S. aureus strains (grade 1) with a favourable outcome are linked to the clonal complexes CC5/CC8. In parallel, we defined a group of virulence factors associated with worse outcomes of DFU, notably sea, sei, lukDE and cap8 [14]. Refined analysis of virulence factors associated with infecting strains, using DNA array technology, identified two highly prevalent factors, lukDE and hlgv, as suitable genetic markers to predict risk of infection by S. aureus (p <0.005). Nevertheless, it would be interesting to determine additional markers, which might be used to discriminate a higher number of colonizing from infecting strains of S. aureus in DFU. We therefore analysed the distribution of edin genes in clinical isolates of S. aureus responsible for diabetic foot infection (DFI) in a French national collection of 195 isolates.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements and Funding
  8. Transparency Declaration
  9. References

Study design

We prospectively enrolled a sample of outpatients attending one of 12 participating French foot clinics between 1 April 2008 and 30 June 2010 for any type of DFU, after informed consent was obtained (Fig. 1). This study was approved by the local ethics committee (South Mediterranean III) and carried out in accordance with the Declaration of Helsinki as revised in 2000. Patients were included if they had not received any antibiotic agents in the previous week. Every patient was examined by trained physicians to grade infection severity. According to the IDSA-IWGDF criteria [15, 16], wounds were considered either uninfected (grade 1) or infected (grade ≥2). After wound debridement, samples for bacterial culture were obtained by swabbing the wound base, needle aspiration or tissue biopsy and immediately sent to the bacteriology department. Only patients with monomicrobial culture for S. aureus were included in the study. Antibiotic treatment was not prescribed in patients with uninfected ulcers. Thereafter, patients were closely monitored for 30 days to definitively assess the wound status (infected/uninfected). If the wound condition was worsening before the follow-up visit, patients were instructed to return to the outpatient department for early review and a further sample was taken for bacterial culture. At the follow-up visit, if the wound was healing but not completely re-epithelialized, a sample for bacterial culture was obtained, and the outcome was considered as favourable. For completely re-epithelialized wounds, no microbiological specimens were sampled and the wound was considered healed.

image

Figure 1. Flowchart of the study.

Download figure to PowerPoint

Microbiological study

Genus, species and antibiotic susceptibilities were determined using the Vitek 2 card (BioMérieux, Marcy-l'Etoile, France) and interpreted according to the recommendations of the French Society for Microbiology [17]. Susceptibility to methicillin was screened by agar diffusion using cefoxitin disks (BioRad, Marnes-La-Coquette, France) [17].

Oligonucleotide DNA arrays to determine EDIN-isoforms

Each S. aureus strain collected during the study was analysed at the INSERM laboratory in Nîmes, France. The Alere StaphyType DNA microarray was used according to protocols and procedures previously detailed [18, 19]. The test was able to screen numerous markers simultaneously in 5 hours. The DNA microarray covers 334 target sequences including two edin-isofoms: edinA and edinB. Primer and probe sequences have been published previously [18, 19]. DNA was extracted from each S. aureus strain and after amplification and hybridization, markers were identified. This technology determines the clonal complex (CC) of strains. A CC may be best defined as a cluster of strains (clones) that are close enough together to be claimed to share a common origin. Thus this group of bacteria is genetically identical to a single ancestral clone. This affiliation of isolates was assessed by an automated comparison of hybridization profiles with a collection of reference strains previously characterized [18, 19].

Statistical analysis

The presence of each gene in S. aureus strains was compared according to the ulcer grades and the outcome of ulcers using Fisher's exact test. The performances of the edin-isoforms for diagnosing grade 4 DFU were assessed as usual. Statistical analysis was performed using the s-plus 2000 software package (Insightful Corporation, Seattle, WA, USA) and results were considered significant for p <0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements and Funding
  8. Transparency Declaration
  9. References

Clinical and bacterial data

During the study period, 195 patients were enrolled. At baseline, seventy-five wounds (38.5%) were classified as uninfected and 120 as infected (Fig. 1). During the follow-up period, 18 patients (24%) returned to the outpatient department because of a rapidly worsening wound: four developed a grade 2 ulcer, 12 a grade 3 ulcer and two a grade 4 ulcer. At the follow-up visit, 33 patients (44%) had their wound not fully re-epithelialized: a sample for bacterial culture was obtained, and the outcome was considered as favourable. In 24 patients (32%), the wound was considered healed.

At baseline, S. aureus isolate was found to be positive for edin (A and B) genes in 14 patients (7.2%): six isolates were recovered from the patients with clinically uninfected DFUs (8%) and eight from patients with infected DFUs (7%). Six of those 14 patients (43%) had a grade 4 DFI compared with six of the 187 patients (3%) in whom S. aureus was edin-negative (p <0.0003, Table 1). Other clinical characteristics were identical for both groups (Table 1). Moreover, no significant differences in infection outcomes were found in the 75 patients of the uninfected group, whether the strains were positive or negative for edin.

Table 1. Demographic and clinical characteristics of the study population
CharacteristicsPatients with edin(+) S. aureus n = 14Patients with edin(−) S. aureus n = 181Total n = 195p edin vs. total
  1. Values are median and interquartile ranges (25th to 75th percentile) or numbers and percentages. NS, non-significant.

Age (range), y69 (58–91)66 (33–101)66.5 (33–101)NS
Male/female, n (%)8 (57.1)/6 (46.9)131 (72.4)/50 (27.6)139 (71.3)/56 (28.7)NS
Type 1/Type 2 diabetes mellitus3/1126/15529/166NS
HbA1c (%)8.4 ± 2.98.3 ± 2.48.3 ± 2.2NS
Cardiovascular disease
 Absence5 (35.7)49 (27.1)54 (27.7)NS
 Coronary heart disease6 (42.9)57 (31.5)63 (32.3)NS
 Peripheral arterial disease8 (57.5)114 (63.0)122 (62.6)NS
 Arterial hypotension12 (85.7)142 (78.5)154 (79.0)NS
 Stroke2 (14.4)14 (7.7)16 (8.2)NS
Nephropathy
 Absence5 (35.7)64 (35.4)69 (35.4)NS
 Microalbuminuria3 (21.4)47 (26.0)50 (25.6)NS
 Proteinuria3 (21.4)48 (26.5)51 (26.2)NS
 Renal failure5 (35.7)57 (31.5)62 (31.8)NS
Neuropathy
 Peripheral10 (71.4)177 (97.8)187 (95.9)NS
 Autonomic0 (0)40 (22.1)40 (20.5)NS
Retinopathy
 Absence7 (50)72 (39.8)79 (40.5)NS
 Non-proliferative diabetic retinopathy6 (42.8)85 (47.0)91 (46.7)NS
 Proliferative diabetic retinopathy1 (7.1)30 (16.6)31 (15.9)NS
Lifestyle factors
 Obesity4 (28.6)73 (40.3)77 (39.5)NS
 Smoking3 (21.4)51 (28.2)54 (27.7)NS
 Alcoholism3 (21.4)25 (13.8)28 (14.4)NS
 Sedentary8 (57.1)139 (76.8)147 (75.4) 
First wound/recurrence3 (21.4)/11 (78.6)48 (26.5)/133 (73.5)51 (26.2)/144 (73.8)NS
IDSA grade
 16 (42.8)69 (38.1)75 (38.5)NS
 20 (0)38 (21.0)38 (19.5)NS
 32 (14.4)68 (37.6)70 (35.9)NS
 46 (42.8)6 (3.3)12 (6.1)0.0003
Outcome of Grade 1 ulcers (1 month after enrollment)
 Complete healing0 (0)24 (13.3)24 (32)NS
 Improved4 (66.7)29 (16.0)33 (44)NS
 Worsening2 (33.3)16 (8.8)18 (24)NS

During the follow-up period, 24 of the initially uninfected wounds completely healed, whereas 33 improved without being closed and 18 worsened. No S. aureus strain from the healed wounds was initially positive for edin genes. From the six uninfected DFUs with edin-positive S. aureus, four had a favourable outcome (66.7%), whereas two (33.3%) worsened rapidly (≤30 days) with occurrence of a grade 4 DFI. Of note, an S. aureus strain was isolated from the second bacterial sampling carried out in those uninfected ulcers that were not completely healed at the follow-up visit, whether the outcome was favourable or not.

Collectively, here we determined that edin-positive strains of S. aureus were more frequently recovered from severe grade 3/4 ulcers.

Distribution of edin-A, B encoding genes

The prevalence of edin-positive isolates was 7.2% (14/195). DNA array analysis revealed that among the 14 edin-positive S. aureus strains, only one (0.5%) was edin-A positive whereas 13 (6.7%) were edin-B positive. The edin-A-positive strain of S. aureus was recovered from a colonizing sample. The edin-B-positive strains were recovered from colonizing samples in five patients and in eight patients suffering from grade 3/4 infected ulcers.

In the second bacterial sampling carried out in those uninfected ulcers that were not completely healed, edin-B-positive S. aureus was recovered only from the two rapidly and severely worsening DFUs. In the four DFUs with favourable outcome, the strains isolated at follow-up no longer harboured any edin-isoforms; moreover, these strains were not the same as initially assessed by difference in their CCs.

If we used the edin-isoforms to differentiate localized infections (grade 2–3) from systemic infections (grade 4), sensitivity was 0.5, specificity 0.96, and positive and negative predictive values 0.43 and 0.97, respectively. Moreover, if we used the edin-isoforms to predict outcome of initially uninfected DFUs, sensitivity was very poor (11.1%) but specificity high (93.0%); positive and negative predictive values were 33.3% and 76.8%, respectively.

Main characteristics of edin+ S. aureus isolates

The univariate analysis found that eight genes were significantly associated with edin-positive isolates: lukDE, hlgv, lukS-PV lukF-PV, etD, cap5, fib, splA and splB (p <0.05). In contrast, four genes (cap8, sea, cna and the regulator agrII) were significantly associated with edin-negative strains (p <0.001).

In 12 strains, edin-B and etD genes were found to be associated. They were all positive for the sarA gene regulon, which can control edin-B expression [20]. Ten of these strains belonged to the agrI cluster. Five of these ten patients had a grade 4 DFU. Regarding the virulence profiles of these ten isolates, all the strains carried the following genes: the haemolysin-encoding genes (hlgv, hla and hlb), the egc cluster of enterotoxins, the genes encoding leukocidin lukDE, the genes encoding intracellular adhesion proteins (icaA, icaC and icaD), cap5, and five genes encoding MSCRAMM (clfA, clfB, fib, ebpS and fnbA). Of note, cna and fnbB genes were absent whereas tet and fosB encoding resistance genes were always present.

As shown in Table 2, the edin-positive S. aureus isolates displayed a low clonal diversity compared with the other strains (40 CCs). None of the edin-positive isolates belonged to CC5/CC8. A total of three known CCs and one unknown singleton were identified. CC25/28 methicillin-sensitive S. aureus (MSSA) (n = 10, 76.9%) was the most common CC in edin-positive strains. This CC and the CC80-MRSA (European Community MRSA Clone, n = 3) were significantly associated with edin-positive isolates (p <0.01). The edin-B-positive singleton isolated from a grade 4 ulcer was found to be negative for etD. It was identified as ST152-MSSA negative for lukDE and hlgv. This reinforced the interest of using edin-B, together with the highly prevalent markers lukDE and hlgv, to decipher strains of S. aureus with a high potential for infection in DFI.

Table 2. Main virulence profiles of S. aureus harbouring edin genes isolated from diabetic foot ulcers
 edinA n = 1edinB n = 13Other strains n = 181p edin vs. other
  1. a

    20 other CCs determined in this group. NS, non-significant

Resistance markers
 mecA 02 (15.4)37 (20.4)NS
 tet 113 (100)177 (97.8)NS
 fosB 110 (76.9)114 (63.0)NS
Enterotoxins
 sea 11 (7.7)78 (43.1)0.04
 seb 03 (23.1)9 (5.0)NS
 sei 010 (76.9)77 (42.5)0.05
 egc cluster 010 (76.9)80 (44.2)NS
Toxins
 lukDE 112 (92.3)114 (63.0)0.04
 hlgv 112 (92.3)95 (52.5)0.003
 hlb 16 (46.2)100 (55.2)NS
 lukS-PV lukF-PV 03 (23.1)1 (0.6)0.001
 etA 00 (0)0 (0)NS
 etB 00 (0)0 (0)NS
 etD 012 (92.3)0 (0)<0.0001
Capsules
 cap5 112 (92.3)98 (54.1)0.004
 cap8 01 (7.7)83 (45.9)0.004
MSCRAMM
 fib 112 (92.3)87 (48.1)0.001
 fnbA 113 (100)162 (89.5)NS
 fnbB 10 (0)48 (26.5)NS
 bbp 111 (84.6)147 (81.2)NS
 ebpS 113 (100)181 (100)NS
 clfA 113 (100)181 (100)NS
 clfB 113 (100)181 (100)NS
 sarA 113 (100)181 (100)NS
Other virulence markers
 isdA 113 (100)181 (100)NS
 splA 112 (92.3)112 (61.9)0.02
 splB 112 (92.3)109 (60.2)0.01
 cna 01 (7.7)75 (41.4)0.01
 chp 010 (76.9)102 (56.4)NS
agr
 agrI 111 (84.6)114 (63.0)NS
 agrII 00 (0)42 (23.2)0.04
 agrIII 02 (15.4)24 (13.3)NS
 agrIV 00 (0)1 (0.6)NS
Clonal complexes
 Unknown CC 10 (0)18 (9.9)NS
 CC25/28-MSSA 010 (76.9)0 (0)<0.0001
 ST152-MSSA 01 (7.7)0 (0)NS
 CC80-MRSA 02 (15.4)0 (0)0.005
 CC5-MSSA 00 (0)14 (7.7)NS
 CC8-MSSA 00 (0)36 (19.9)NS
 Others 00 (0)113 (62.5)a<0.001

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements and Funding
  8. Transparency Declaration
  9. References

DFUs are prone to infections that may spread systematically and lead to amputation of lower extremities and possibly to death. From a clinical point of view, differentiating colonization from infection may be problematic in DFUs due to the masking impact of the neuropathy and/or ischemia [15]. Monomicrobial DFI thus represents a unique model in humans to compare the distribution of bacterial factors between infecting and colonizing strains of S. aureus. Although the virulence potential of S. aureus relies on the combination of multiple factors, those associated with colonizing strains of grade 1 DFUs have probably a poor contributing role in virulence and notably in invasion. In this multicentre study, we found that among the 14 edin-positive S. aureus isolates, ten were recovered from deep grade 3 or 4 infections (at baseline or during follow-up). Moreover, in the four other edin-positive cases in which outcomes were favourable, the strains isolated at follow-up were no longer positive for edin and were from different CCs. These findings suggest that these edin-positive strains were successfully removed during debridement, allowing colonization of wounds by other strains. At least, these findings clearly indicate that edin genes have a higher prevalence in S. aureus from high-grade ulcer isolates.

Three alleles of edin have been characterized [21-23]. They all share similar biochemical activities toward Rho proteins but are encoded within different genetic backgrounds. The gene encoding edin-A is located on a plasmid, as for edin-C and the closely associated exfoliative toxin-B encoding gene [22]. edin-B is chromosomally encoded within a highly variable pathogenicity island in close proximity to exfoliative toxin-D encoding gene [20, 23]. Although the DNA array we used did not efficiently detect the edin-C gene (unpublished data), the absence of detection of the exfoliative type-D strongly suggests that edin-C-positive strains were not present in this type of infection. Thus, edin-B is the most prevalent edin gene associated with DFIs.

The clonal complex analysis indicates that edin-positive strains belonged to four major groups: a singleton near CC8 (edin-A, n = 1), a singleton belonging to ST152-MSSA (edin-B, n = 1), CC80-MRSA (edin-B, n = 2) and CC25/28-MSSA (edin-B, n = 10). Here we identified an edin-B-positive/et-D-negative strain that belongs to ST152-MSSA. Importantly, this strain appeared negative for the haemolysin (hlgv) and the leukocidin (lukDE) genes, which have been defined as two highly prevalent virulence markers of infectious strains in DFU. Given that the edinB-positive ST152 strain of S. aureus was isolated from a Grade 4 ulcer, this result further emphasizes the interest in screening edin-B as a surrogate maker to discriminate colonizing from infecting strains of S. aureus in DFU. It is worth noting that none of the edin-positive S. aureus belonged to CC5/CC8; two CCs claimed to be predictive of a favourable outcome [13].

Our analysis of the distribution of edin in DFU supports the idea that edin might work together with the arsenal of S. aureus virulence factors to give the bacteria a higher potential for infection [5, 12]. In keeping with this concept, previous studies indicated that EDIN factors might favour induction of deep-seated infectious foci [24, 25]. Edin encoding genes thus represent additional markers of interest to differentiate infecting from colonizing Staphylococcus aureus strains in DFU and to predict the wound outcome. Our results highlight the potential usefulness of the molecular biology in patients in whom the clinical diagnosis of infection is made difficult by peripheral arterial disease, neuropathy or impaired leukocyte functions [26]. The DNA microarray might provide important help for clinicians and might allow adequate management of uninfected ulcers carrying S. aureus, according to its genotype profile. The detection of some CCs and virulent genes in grade 1 ulcers raises the question about antibiotic treatment in apparently uninfected DFU and challenges the dogma about abstention of antibiotic therapy in clinically uninfected ulcers. Indeed, the absence of CC5/CC8 corresponding to colonizing strains and the presence of CC25/28 and/or of lukDE, hlgv and edin genes suggest the need for antibiotic treatment as a therapeutic strategy. However, to date the techniques we used are largely in the field of research.

Acknowledgements and Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements and Funding
  8. Transparency Declaration
  9. References

This work was supported by grants from the Hospital Project of Clinical Research 2007 (2007-A00678-45), the French Society of Diabetes (ALFEDIAM grant 2008), a Pfizer Grant, the Languedoc-Roussillon Region (Chercheur d'avenir Grant 2009) and the National Institute of Health and Medical Research (INSERM). NM is a PhD student from the InfectioPôle Sud Foundation. We thank all the participants of the foot clinics for help in recruiting patients. We thank Laure Vidal-Navarro, INSERM U1047, for her technical assistance and Anne Blanc-Potard, Julien Chamard, Céline Groul-Viaud and the Research Clinical Technicians of the study, Nîmes University Hospital, for their help in data processing.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements and Funding
  8. Transparency Declaration
  9. References