Introduction

Staphylococcus aureus is a major human pathogen that causes both suppurative and toxin-mediated nosocomial and community-acquired infections worldwide [1]. Numerous toxins have been identified in S. aureus, including some of them that have been potentially associated with specific infections or syndromes [2].

Among these virulence factors, Panton–Valentine leukocidin (PVL) is a pore-forming toxin generally associated with complicated skin and skin structure infections, extensive cellulitis, necrotizing pneumonia, and acute osteomyelitis in children [3–5]. In particular, the leukotoxic action of PVL is responsible for the high mortality rate (up to 75%) reported in necrotizing pneumonia [4].

Several toxins have superantigenic properties, such as toxic shock syndrome toxin-1 (TSST-1) and staphylococcal enterotoxins B (SEB) and C (SEC). These superantigens directly activate T-cells through bypassing normal antigen presentation. They bind to the specific variable region of the T-cell antigen receptor-β-chain with strong affinity [2, 6]. Uncontrolled release of Th1 pro-inflammatory cytokines mediated by massive T-cell proliferation may be responsible for the symptomatology of staphylococcal toxic shock syndrome (STSS) [2]. Clinically, STSS combines rapid onset of fever, diffuse macular rash, hypotension, and multiple organ failure, with a mortality rate ranging from 4 to 22% [6, 7].

Exfoliative toxins A (ETA) and B (ETB) are epidermolytic proteases mainly responsible for staphylococcal scalded skin syndrome (SSSS), which typically occurs in neonates and infants, but can also affect predisposed adults. The syndrome is characterized by a fever and a rapid generalized desquamation without mucosal damage. The mortality rate is low in infants but can reach 67% in adults [8]. The strains producing these toxins can also cause localized forms (namely, bullous impetigo), where skin lesions are restricted to the site of infection [8].

Even if these toxins have been related to severe infections, the clinical significance of their presence in S. aureus isolates has been poorly investigated, and only among collections of non-consecutive clinical isolates [9–13]. The aim of this study was, then, to: (i) screen consecutive S. aureus clinical isolates (resistant or not to methicillin) for the presence of these six major toxins and (ii) assess the clinical significance of these toxin-producing strains. The proportion of virulence genes was compared between the isolates groups associated with uncomplicated infection and severe infection.

Materials and methods

Bacterial isolates

From May to June 2009, all non-duplicate S. aureus clinical isolates recovered from patients hospitalized at the university hospital of Caen (ca. 2,000-bed) were studied. They were identified by using conventional phenotypic methods, such as Gram staining, catalase activity, and latex agglutination test (Pastorex Staph-Plus®; Bio-Rad, Marnes-la-Coquette, France). In doubtful cases, an identification by using the ID-GP card on the Vitek 2 system (bioMérieux, Marcy-l’Etoile, France) was performed. Antimicrobial susceptibility was determined by disc diffusion on Mueller–Hinton agar according to the Antibiogram Committee of the French Society for Microbiology guidelines (http://www.sfm.asso.fr).

Patient characteristics and definitions

Demographic and clinical data were retrospectively recorded for each patient regarding age, sex, underlying conditions (including heart disease, hemodialysis, diabetes mellitus, hematological disorder, HIV infection, current malignancy, immunosuppressive therapy), site of isolation, clinical presentation, hospital ward, treatment, and clinical outcome.

Sepsis, severe sepsis, septic shock, and multiple organ dysfunction syndrome were defined according to the criteria of the American College of Chest Physicians and Society of Critical Care Medicine Conference [14], while STSS was defined as described elsewhere [6]. Hospital-acquired infections were defined according to the CDC definitions [15] (http://www.cdc.gov/ncidod/dhqp/pdf/nnis/NosInfDefinitions.pdf). All other cases were considered as community-acquired infections.

Multiplex real-time PCR

Bacterial genomic DNA was extracted by using the InstaGene Matrix® Kit (Bio-Rad) according to the manufacturer’s recommendations. We developed two triplex real-time polymerase chain reaction (PCR) assays for the detection of pvl, tst, and eta genes (PCR A) and etb, seb, and sec genes (PCR B). Primers and TaqMan™ LNA probes were designed and synthesized by Sigma-Aldrich France (Table 1). PCR A and PCR B experiments were performed by using the QuantiFast® Multiplex PCR Kit (Qiagen, Courtaboeuf, France) in a 25-μl reaction mixture with final concentrations of 0.5 μM for each primer and 0.2 μM for each LNA probe. PCR amplifications were performed using a Rotor-Gene Q instrument (Qiagen) as follows: (i) initial denaturation step of 5 min at 95°C and (ii) 40 cycles of PCR, with one cycle consisting of 45 s at 95°C and 45 s at 60°C. The presence of amplified DNA was detected by measuring the fluorescence emitted at 470 nm (for pvl and etb), 530 nm (for tst and seb), and 585 nm (for eta and sec).

Table 1 Primers and LNA probes used in the study
Full size table

PFGE

Fingerprinting by pulsed-field gel electrophoresis (PFGE) was performed on toxin gene-positive strains as previously described [16]. SmaI-macrorestricted patterns were obtained by using a contour-clamped homogeneous electric field DR-III apparatus (Bio-Rad) as follows: ramped pulse times of 5 s and 35 s at 6 V/cm for 20 h at 14°C. Fingerprint® II software (Bio-Rad) was used to analyze PFGE patterns. The calculation of similarity matrices and dendrograms were obtained with the unweighted pair group method with arithmetic averages (UPGMA). Strains were considered to be clonally-related if the Dice coefficients were greater than 85%.

Comparative analysis

In order to evaluate the role of toxins in the severity of infections, patients were classified into two groups: (i) uncomplicated infections (UI) defined as sepsis and (ii) severe infections (SI) defined as severe sepsis, septic shock, multiple organ dysfunction syndrome, or STSS. These two groups were compared regarding demographic and clinical characteristics, as well as toxin gene prevalence.

Statistical analysis

Statistical analysis was performed using the GraphPad Software (http://www.graphpad.com/quickcalcs). The χ2 test or Fisher’s exact test was used, as appropriate, to compare categorical variables, whereas Student’s t-test was used to compare continuous variables. Differences between groups were considered to be statistically significant for p-values <0.05.

Results

Clinical data

During the two-month period, 186 strains were recovered from 186 patients, including 75 isolates obtained from patients with superficial infections (68 skin and soft tissue infections, six ear, nose, throat [ENT] infections, and one genital infection) and 111 from patients with invasive infections (33 respiratory tract infections, 30 bacteremia, 21 urinary tract infections, 17 bone or joint infections, and 10 deep abscesses). Among the 186 patients enrolled in the study, 151 suffered from uncomplicated infections and 35 suffered from severe infections (18 severe sepsis, 15 septic shock, two multiple organ dysfunction syndromes). No patients met the diagnostic criteria for STSS. Most infections were hospital-acquired (99/186; 53.2%), while the prevalence of methicillin-resistant Staphylococcus aureus (MRSA) infection was 25.8%. Forty surgical site infections (21.5%) were reported.

The demographic and clinical characteristics of the patients are shown in Table 2. There were no significant differences between UI and SI groups regarding age, sex, underlying diseases, prior hospitalization, prior surgery, nosocomial infection, MRSA infection, and surgical therapy (Table 2). By contrast, there was a significant difference for hospitalization ward, type of infection (superficial or invasive), and antimicrobial therapy (Table 2). Unsurprisingly, the 28-day crude mortality was significantly higher among SI patients than among UI patients (34.1% vs. 6.6%; p = 0.0002).

Table 2 Demographic and clinical characteristics of 186 patients with Staphylococcus aureus infections
Full size table

Prevalence of toxin genes

Of the 186 isolates, 55 (29.6%) harbored at least one toxin gene (Table 3), comprising 13 of 35 strains (37.1%) recovered from SI patients and 42 of 151 strains (27.8%) recovered from UI patients. The most prevalent gene was sec (15.1%), followed by seb (10.2%). Five (2.7%) strains were positive for both tst and sec genes, while seb and sec genes were detected in a single isolate. Finally, the eta gene was detected in one strain, whereas no strain was positive for the etb gene. Noteworthy, one or more toxin genes were detected in 46 (33.3%) of 138 methicillin-susceptible S. aureus (MSSA) and in 9 (18.7%) of 48 MRSA.

Table 3 Polymerase chain reaction (PCR) detection of S. aureus toxin genes and comparison of toxin profiles between uncomplicated and severe infections
Full size table

By using a cut-off value of 85% for the Dice coefficient, 16 and 10 different PFGE clusters were identified in the group of pvl-, tst- and sec-, seb-, seb- and sec-, and eta-positive strains (n = 27) and in the group of sec-positive strains (n = 28), respectively (Fig. 1). Noteworthy, three of the five tst-positive isolates were MRSA, while the single pvl-positive isolate was also MRSA (Fig. 1a). Concerning the 28 sec-positive strains, 13 (46.4%) clustered in a major group (Fig. 1b). Although they exhibited identical antimicrobial resistance profiles, they were recovered at different dates from patients hospitalized in different hospital wards (Fig. 1b).

Fig. 1
figure 1

Unweighted pair group method with arithmetic averages (UPGMA) dendrogram of pulsed-field gel electrophoresis (PFGE) results based on the Dice distance matrix of pvl-, tst and sec-, seb-, seb and sec-, and eta-positive (a) and sec-positive (b) Staphylococcus aureus clinical isolates. The 85% cut-off value of the Dice coefficient is shown by a vertical dotted line. Date of isolation, hospital ward, and antimicrobial resistance and toxin gene profiles are also indicated. OXA, oxacillin; KM, kanamycin; TM, tobramycin; LVX, levofloxacin; ERY, erythromycin; LIN, lincomycin; PT, pristinamycin; FA, fusidic acid; FOS, fosfomycin; RA, rifampicin

Full size image

Clinical significance of toxin-producing strains

The overall prevalence of toxin-producing strains among UI patients (27.8%) and SI patients (37.1%) was not statistically different (p = 0.3044) (Table 3). In addition, none of the detected genes was more prevalent in one of the two groups. Nonetheless, SI tended to be associated with a higher prevalence of the sec gene (25.7% vs. 12.5%; p = 0.0655).

Discussion

To the best of our knowledge, this is the first study that has evaluated the prevalence of several major toxins on a collection of consecutive (methicillin-resistant or -susceptible) S. aureus clinical isolates. To do that, we optimized a multiplex real-time PCR assay that appeared to be a useful tool for the detection of S. aureus toxin genes during epidemiological surveys, as previously reported [17, 18]. Overall, 29.6% of all tested isolates were positive for one or more toxin genes, this prevalence being similar to those reported in previous studies [10, 11, 13, 19]. More specifically, only a small proportion of strains were positive for pvl (0.5%) and eta (0.5%) genes and none for the etb gene. Such low prevalence has also been reported in other studies [11, 13, 20]. The single pvl-positive strain harbored a typical antimicrobial resistance pattern (resistance to methicillin, kanamycin, and fusidic acid, and susceptibility to tobramycin, gentamicin, and fluoroquinolones), suggesting that it could belong to the MRSA sequence type 80 (ST80) European clone that has emerged and spread in Europe since the late 1990s [21]. We found a lower proportion of tst gene (2.8%) in comparison with other studies (8.7 to 30% of infections) [10, 11, 13, 19], but in Peacock et al.’s study, this prevalence decreased after adjustment for effects of clonality [13]. Among the five tst-positive strains (also sec-positive), three isolates had a particular antimicrobial resistance phenotype (phenotype 1: resistance to methicillin, kanamycin, tobramycin, and fusidic acid, and susceptibility to gentamicin and fluoroquinolones; phenotype 2: resistance to methicillin and fusidic acid, and susceptibility to kanamycin, tobramycin, gentamicin, and fluoroquinolones) assumed to belong to the MRSA ST5 Geraldine clone, which was recently reported in France [22]. The analysis of PFGE patterns showed the polyclonality of circulating strains in our hospital during the period of the study. Only one major cluster of sec-positive isolates (46.4%) was identified, suggesting that these strains were epidemiologically related. However, no outbreak was identified during the two-month study period.

In this study, we demonstrated no statistically significant association between the severity of infection and the presence of the six detected toxin genes, even if the sec gene tended to be more common in SI patients than in UI patients (25.7% vs. 12.5%; p = 0.065). This tendency has not yet been reported [10, 11, 13, 19] and remains to be confirmed.

Our study presented several limitations: (i) this study was a single-center investigation, (ii) clinical data have been retrospectively collected, and (iii) the SI patients group (n = 35) was relatively small compared to the UI patients group. Furthermore, the molecular analysis of strains does not ascertain the in vivo production of toxins.

Other toxins have been shown to play a role in severity. For instance, Ferry et al. demonstrated in S. aureus bacteremia a link between the presence of the sea gene and septic shock, whereas the presence of the egc genes was negatively correlated with the severity of infection [10]. In addition, several virulence factors could act in combination for enhancing the clinical severity of infections [13]. Even if only one pvl-positive strain was identified in this study, the role of PVL as a major virulence determinant in severe infections remains controversial [9, 23–28]. Finally, the interaction between the host immune system (including genetic predispositions) and the pathogen may play a major part in S. aureus toxin-mediated infections’ pathogenesis. Not all colonized or infected patients with a toxin-producing strain of S. aureus go on to develop STSS. A lack of detectable antibodies to STSS-associated superantigens seems to be a major risk for the development of STSS [2, 6]. These individuals may be predisposed to repeated episodes of STSS. In addition, toll-like receptor type 2 (TLR-2) polymorphism may place individuals at a higher risk of Gram-positive septic shock, as well as certain HLA haplotypes can also impact on the clinical susceptibility to toxic effects of superantigens [29].

In conclusion, we did not observe a statistically significant relation between the severity of the infection and the presence of toxin gene(s). However, it would be interesting to complete these first results by a multicenter investigation with a larger population. So far, the detection of toxin genes in S. aureus has been considered to be of interest for the diagnosis of toxin-mediated diseases leading to an early antitoxinic therapy, such as clindamycin, linezolid, and/or intravenous immunoglobulins. Although these molecules have been shown to be effective for reducing exotoxin production in vitro [30–33], the clinical benefits should be more thoroughly evaluated.