INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Toxic shock syndrome (TSS) is a life-threatening multisystem disorder caused by strains of Staphylococcus aureus. It is characterized by rapid onset of fever, arterial hypotension, scarlatiniform rash, and multiorgan failure (4). Originally described for children by Todd and Fishaut in 1978 (32), TSS has been extensively studied over the past 18 years, since the occurrence of major outbreaks associated with menstruation and the use of a newly introduced superabsorbant brand of tampon (3, 4). These tampons were withdrawn from the market, and most cases of TSS now occur in settings other than menstruation and among individuals of both sexes and all ages. Nonmenstrual TSS is usually secondary to S. aureus infection or to skin or mucosal trauma with S. aureus colonization (4).

S. aureus TSS toxin 1 (TSST-1) was the first toxin shown to be involved in TSS, in both menstrual and nonmenstrual cases (2, 30). Staphylococcal enterotoxins A to D and H (SEA to SED and SEH) also appear to have caused some cases of nonmenstrual TSS (8, 14, 16, 21, 28, 29). TSST-1 and SEA to SED have been linked to other staphylococcal syndromes such as staphylococcal scarlet fever (SSF) and recalcitrant erythematous desquamating disorder (REDD), both of which were suggested to be variants of TSS on the basis of toxin production and certain clinical similarities (6, 18). Another staphylococcal enterotoxin (SEE) was isolated from chicken and food specimens but has not been associated with TSS (7). All of these toxins exhibit superantigen activity, stimulating polyclonal T-cell proliferation through coligation between major histocompatibility complex class II molecules on antigen-presenting cells and the variable portion of the T-cell antigen receptor  chain (22). This superantigen activity can be detected with mitogenic assays (with mouse, rabbit, or human leukocytes) or in a CD69-specific flow cytometric assay of T-cell activation (15, 17).

In a French epidemiological survey of S. aureus isolates from patients with TSS, SSF, and REDD, several strains produced none of the known superantigenic toxins (TSST-1, SEA to SEE, and SEH), but superantigen activity was detected in culture supernatants in a CD69-specific flow cytometric assay, pointing to unknown superantigenic toxins (16). Recently, two new staphylococcal enterotoxins, SEG and SEI, have been isolated from S. aureus strains collected from human nares but have not been linked to TSS (24). In this study we used PCR to detect the SEG and SEI genes (seg and sei, respectively) in CD69-positive, SEA- to SEE-, SEH-, and TSST-1-negative strains isolated from patients with a diagnosis of TSS or SSF. seg and sei were detected in all of these strains, suggesting a clinical importance for these new toxins.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Patients. The 12 patients included in this study corresponded to nine cases of TSS (including two menstrual case) and three cases of SSF. They were all colonized or infected by S. aureus strains that did not produce the usual superantigenic toxins associated with TSS. They were selected from among 170 cases of TSS, 5 cases of REDD, and 105 cases of SSF reported to the Centre National de Référence des Toxémies à Staphylocoques (Lyon, France) between 1 January 1985 and 31 January 1999 from hospitals throughout France. Cases were first identified by chart review if the patient was from the Lyon area or otherwise from accompanying notes sent to the staphylococcal reference center. Cases met the definition of TSS, REDD, or SSF (4, 6, 16). The patients were epidemiologically unrelated.

Strains. S. aureus strains from patients with TSS or SSF were cultured from sites including the genital tract, blood, skin, throat, and soft tissues. Strains were identified as S. aureus by their ability to coagulate citrated rabbit plasma (bioMérieux, Marcy-l'Etoile, France) and to produce a clumping factor (Staphyslide test; bioMérieux). Isolates were typed by using phage and serotyping techniques (33).

Toxins. Superantigen activity in culture supernatants was detected by measuring T-cell activation in a CD3- and CD69-specific flow cytometry assay (15). Since fewer than 1% of unstimulated CD3⁺ lymphocytes spontaneously expressed CD69, T cells were considered to be activated when more than 2% expressed CD69 (15).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The 12 S. aureus strains induced CD69 expression by over 2% of CD3⁺ lymphocytes (Table 2), in a manner similar to that observed with supernatants from control strains producing known superantigenic toxins (Table 3). In contrast, the toxin-negative control strain (RN450) induced CD69 expression in only 0.4% ± 0.2% of CD3⁺ lymphocytes, as observed with strains that do not produce superantigenic toxins (15).

The 12 strains were examined for the presence of sea to see, seg to sei, eta, and etb by PCR amplification, and were all positive for seg and sei only (Table 2), while RN-450 was negative for all of these genes. Several of the control strains harboring sea to see, seh, tst, eta, or etb were also positive for both seg and sei (Table 3). seg and sei amplicons from 3 of the 12 clinical strains were sequenced, and the sequences were 100% identical to the published sequences (GenBank accession no. AF064773 and AF064774, respectively). Since a change in only a few bases could cause false-negative results in seg or sei PCRs, DNAs from three PCR-positive strains and five PCR-negative strains were analyzed by Southern blotting with seg and sei probes. Only strains which were positive for seg and sei by PCR hybridized to both DNA probes (Fig. 1), thus confirming the PCR results. The sizes of the hybridizing fragments were identical for both probes but differed between strains, from 2.9 kb (strains A900322 and TC7) to >7.1 kb (strain A980483) (Fig. 1).

View larger version (75K): [in this window] [in a new window]   FIG. 1.   Southern blot hybridization of DNAs of S. aureus strains with sei and seg probes. Total DNA was digested with HindIII, separated by agarose gel electrophoresis, transferred to positively charged nylon membranes, and probed with the indicated DIG-labelled probes. Lanes contain DNAs of three S. aureus strains found to be PCR positive for seg and sei (lanes 1, A980483; lanes 2, TC-7; and lanes 3, A900322) and five S. aureus strains found to be PCR negative for seg and sei (lanes 4, FRI-1169; lanes 5, FRI-569; lanes 6, RN-450; lanes 7, A970237; and lanes 8, A990204).

Since all 12 strains were positive for both genes by PCR and the sizes of the hybridizing fragments were identical for both probes, the possibility that the seg and sei loci were adjacent to each other was investigated by attempting to coamplify the two genes with all combinations of seg- and sei-specific primers. Only the combination of primer SEI-1 with primer SEG-2 produced an amplicon, of about 3.2 kb, while the other primer combinations gave negative results. Partial sequence analysis showed that the 3.2-kb amplicon contained portions of both seg and sei, in tandem orientation with a 1.9-kb intergenic segment.

Eight of the clinical strains harboring seg and sei were analyzed by phage typing and serotyping. They were not clonal, as they had distinct phage types. Four strains belonged to phage group III, one belonged to group V, and three were untypeable (Table 2). Serotyping confirmed the absence of clonality.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Among the staphylococcal superantigenic toxins, only TSST-1, SEA, SEB, SEC, and SED have been linked to TSS or SSF (8, 18, 21). This study shows that two additional staphylococcal enterotoxins, SEG and SEI, are likely associated with human TSS and SSF. We selected strains of S. aureus isolated from patients with TSS or SSF and which did not produce TSST-1, SEA to SEE, SEH, ETA, or ETB. Using PCR amplification with primer sets designed to be specific for seg or sei, we detected both genes in all 12 strains and also in several control strains (Tables 2 and 3). The specificity of seg and sei amplification was confirmed by sequence analysis of PCR products and Southern blotting (Fig. 1). SEG and SEI produced by these strains were probably responsible for the T-cell activation detected in a CD69-specific flow cytometry assay. It is also conceivable that the SEG and SEI produced by these strains caused the clinical manifestations of TSS or SSF.

As seg and sei were initially cloned from two different strains (FRI-572 and FRI-445, respectively) (24), we were surprised that both seg and sei were detected in all 12 clinical strains and also in several reference strains (Table 3). The positive PCR amplification obtained with the SEI-1 and SEG-2 primers in our study indicates that sei and seg are in tandem orientation and are separated by 1.9 kb of intergenic DNA. Sequencing of this intergenic region is under way to determine whether it contains additional open reading frames, as suggested by the reported observation that strain FRI-445 contains part of an enterotoxin-like gene upstream of sei (24). The link between two staphylococcal superantigenic toxins such as seg and sei is not uncommon; it has been described for plasmid pIB485, which contains both sed and sej (34), and for pathogenicity islands containing both tst and an enterotoxin-like gene (19). Phage typing and serotyping ruled out a clonal origin of our S. aureus clinical strains harboring seg and sei and responsible for TSS, contrasting with the clonal origin of strains isolated from patients with menstrual TSS that produced TSST-1 (25).

Three of the strains that produced SEG and SEI were associated with two cases of menstrual TSS and a case of puerperal SSF, respectively. Previous findings suggested that the vast majority of cases of menstrual TSS were due to TSST-1-producing strains (29). However, some vaginal isolates from women with menstrual TSS did not produce TSST-1, suggesting that other staphylococcal toxins might be responsible for the clinical manifestations (4, 9, 10, 21); indeed, cases of menstrual TSS related to strains producing only SEA to SED have been described (8, 23). Our data suggest that S. aureus strains producing SEG and SEI but not TSST-1 can also cause menstrual TSS, although this needs to be confirmed by experimental and epidemiological studies.

In conclusion, S. aureus strains that produce both SEG and SEI may be associated with SSF and TSS (including menstrual cases), and the SEG and SEI determinants are close to each other on the S. aureus chromosome. The PCR amplification method used in this study is an efficient way of identifying strains harboring seg and sei.