Influence of the Vaginal Microbiota on Toxic Shock Syndrome Toxin 1 Production by Staphylococcus aureus

Abstract

Menstrual toxic shock syndrome (TSS) is a serious illness that afflicts women of premenopausal age worldwide and arises from vaginal infection by Staphylococcus aureus and concurrent production of toxic shock syndrome toxin-1 (TSST-1). Studies have illustrated the capacity of lactobacilli to reduce S. aureus virulence, including the capacity to suppress TSST-1. We hypothesized that an aberrant microbiota characteristic of pathogenic bacteria would induce the increased production of TSST-1 and that this might represent a risk factor for the development of TSS. A S. aureus TSST-1 reporter strain was grown in the presence of vaginal swab contents collected from women with a clinically healthy vaginal status, women with an intermediate status, and those diagnosed with bacterial vaginosis (BV). Bacterial supernatant challenge assays were also performed to test the effects of aerobic vaginitis (AV)-associated pathogens toward TSST-1 production. While clinical samples from healthy and BV women suppressed toxin production, in vitro studies demonstrated that Streptococcus agalactiae and Enterococcus spp. significantly induced TSST-1 production, while some Lactobacillus spp. suppressed it. The findings suggest that women colonized by S. aureus and with AV, but not BV, may be more susceptible to menstrual TSS and would most benefit from prophylactic treatment.

INTRODUCTION

Toxic shock syndrome (TSS) is a systemic illness characterized by extensive T-cell proliferation throughout the body. This leads to systemic inflammation and concurrent health problems, including rash formation, multiple organ failure, and potentially death. Menstrual TSS is a form of the disease prominent in premenopausal women, afflicting approximately 1 in every 100,000 women in the Western world (1, 2). This condition arises from vaginal infection by Staphylococcus aureus and subsequent production of toxic shock syndrome toxin-1 (TSST-1), a superantigen thought to be unique with the capability of crossing the vaginal epithelial layer (3, 4). It is estimated that 70% of women in North America use tampons (5), and although the rates of menstrual TSS have dropped significantly since the early 1980s, it remains crucial to better understand the pathogenesis of this condition to prevent its re-emergence.

Production of TSST-1, encoded by the tst gene, is regulated in large part by the agr system of S. aureus (6), a well-characterized quorum-sensing system (7) that regulates many secreted and surface-expressed virulence factors (8). Activation of this system is regulated by several factors, including the SarA protein (9, 10), RAP/RIP proteins (11), and alternative sigma factor B (σ^B) (12). Much research has been focused on investigating the nature of TSST-1 production in response to environmental cues that may be present in the vagina. For instance, Yarwood and Schlievert (13) revealed that elevated oxygen levels in the presence of carbon dioxide are necessary for TSST-1 production. It is also well established that a neutral pH of 6.5 to 7.0, which is found under aberrant vaginal conditions, is optimal for production of this toxin, in contrast to a healthy vaginal pH of 4.5 (14, 15). It is therefore readily apparent that the vaginal environment, characteristic of fluctuating gas levels and pH, could influence the degree to which colonizing S. aureus produces TSST-1. However, little research has been conducted to evaluate the putative role of the residing vaginal microbiota, which arguably shapes this environment as a whole.

The vaginal microbiota consists largely of Lactobacillus species, which produce lactic acid and other antimicrobial metabolites that prevent the overgrowth of opportunistic pathogens (16). These bacteria have been shown to inhibit the growth of several urogenital pathogens, including S. aureus (17, 18). Furthermore, the vaginal probiotic strain Lactobacillus reuteri RC-14 has been shown to produce cyclic dipeptides capable of suppressing TSST-1 production in S. aureus (19), suggesting that indigenous vaginal lactobacilli may protect the host from menstrual TSS progression.

Changes in the vaginal environment render the vaginal microbiota susceptible to constant fluctuations in its members. This occasionally leads to depletion in the lactobacillus population and a subsequent increase in pathogenic bacteria, resulting in bacterial vaginosis (BV) or aerobic vaginitis (AV). BV is characterized by the overgrowth of strict anaerobic bacteria, including Gardnerella vaginalis, Atopobium vaginae, Prevotella bivia, and Leptotrichia amnionii (20, 21), and has been associated with increased susceptibility to sexually transmitted infections, preterm labor (22), and pelvic inflammatory disease (23, 24). In contrast, AV is characterized by the overgrowth of enteric aerobic and facultatively anaerobic commensals, such as Escherichia coli, Streptococcus agalactiae, Enterococcus faecalis, and S. aureus (25), and is an inflammatory condition that can often lead to pregnancy complications (26–29).

The aim of the present study was to investigate a potential relationship between the vaginal microbiota and TSST-1 production by S. aureus and, thus, susceptibility to menstrual TSS. We hypothesize that an aberrant microbiota that alters the vaginal pH and environment leads to increased production of TSST-1 and thus might represent a risk factor for the development of TSS. In contrast, we believe that TSST-1 production is suppressed in a healthy vaginal environment dominated by Lactobacillus species. If this link between the vaginal microbiota and TSST-1 production is established, it may be feasible to control toxin production of S. aureus by altering the vaginal microbiota through the application of exogenous probiotic bacteria.

MATERIALS AND METHODS

Clinical study.

Details of the clinical study were reviewed and approved by the Health Sciences Research Ethics Board at The University of Western Ontario. Participants were provided with a package detailing all relevant information about the study, including an in-depth explanation of the clinical procedure, and each person signed a consent form prior to sample collection.

Recruitment of premenopausal women between the ages of 18 and 40 years with BV as well as healthy controls took place in London, Ontario, Canada, and was based on selective criteria designed to ensure that the vaginal samples reflected a representative microbiota of the general female premenopausal population. Recruitment posters were placed in various locations in London, Ontario, Canada, including The University of Western Ontario and local hospitals and medical clinics. The poster emphasized a target audience of premenopausal women with a suspected history of BV. Participants were excluded if they had reached menopause; had a urogenital infection other than BV in the past 6 months; were pregnant; had a history of gonorrhea, chlamydia, estrogen-dependent neoplasia, abnormal renal function, or pyelonephritis; were taking prednisone, immune-suppressive agents, or antimicrobial medication; or had undiagnosed abnormal vaginal bleeding. Participants were asked to refrain from oral or vaginal intercourse and consuming probiotic supplements or foods for 48 h prior to the clinical visit. No participants were menstruating at the time of the clinical visit. At the conclusion of the study, a total of 34 participants were recruited: 11 healthy participants, 10 with an intermediate status, and 13 with BV. The mean age was 24.6 years. None of the subjects were diagnosed with AV or had a microbiome indicative of this condition.

Weekly clinics were held at the Victoria Family Medical Centre (London, Ontario, Canada) over a 6-month period. A total of 7 samples were collected from each participant at a single time point: one pHem-alert applicator (Gynex) was used to detect the vaginal pH; three CultureSwab polyester-tipped swabs (BD Biosciences) were used for bacterial isolation (for diagnosis of BV by the Nugent score, for bacterial DNA collection); two ESwabs (Copan) were processed for culturing the organisms; and one cytobrush (Cooper Surgical) was processed for vaginal epithelial cell RNA. Swabs were inserted approximately 5 cm into the vaginal canal, pressed against the wall, and rotated 4 times.

The microbial status was assessed through the Nugent scoring method (30). A CultureSwab polyester-tipped swab (BD Biosciences) was pressed against the vaginal wall and smeared onto a glass microscope slide, heat fixed, and Gram stained. Four fields of view were chosen at random at a magnification of ×1,000, and bacterial numbers from each were recorded. A participant was diagnosed as having BV if the Nugent score was 7 to 10 and if she had a vaginal pH greater than 4.5. All participants with a healthy Nugent score of 0 to 3 had a pH that was equal to or less than 4.5.

Sequencing by Illumina system.

Bacteria were collected on a CultureSwab polyester-tipped swab (BD Biosciences), which was immediately placed in 700 μl RNAprotect reagent (Qiagen). The tube was vortexed for 5 s, the swab was discarded, and the solution was stored at −20°C. DNA isolation was performed using an InstaGene matrix (Bio-Rad) DNA preparation kit according to the manufacturer's instructions. The collected DNA was stored at −20°C and was used as the DNA template for subsequent PCRs.

PCR was carried out in 50-μl reaction mixtures consisting of 40 μl master mix and 10 μl template (or 10 μl Milli-Q H₂O [Millipore] for the negative control). The master mix consisted of 1× PCR buffer (Invitrogen), 1.7 mM MgCl₂, 210 μM deoxynucleoside triphosphates (dNTPs), 640 nM forward and reverse primers, 0.05 U/μl Platinum Taq polymerase (Invitrogen), and Milli-Q H₂O (Millipore). Amplification was done using eubacterial primers flanking the V6 region of the 16S rRNA gene: V6-F (5′-CAACGCGARGAACCTTACC-3′) and V6-R (5′-ACAACACGAGCTGACGAC-3′). The PCRs were carried out in a Mastercycler apparatus (Eppendorf) under the following annealing temperature touchdown program: 94°C for 2 min, followed by 10 cycles of 94°C for 45 s, 61°C to 51°C for 45 s (with the temperature in each cycle dropping by 1°C), and 72°C for 45 s and then 15 cycles of 94°C for 45 s, 51°C for 45 s, and 72°C for 45 s and ending with a final elongation step of 72°C for 2 min. PCR products (5 μl) were mixed with 1 μl 6× loading dye, and the mixture was loaded onto a 1.5% agarose gel. The gel was then viewed under UV light in an AlphaImager system (Alpha Innotech Corporation). The ImageJ program was used to quantify the brightness of each DNA band so that equal amounts of each sample were added to the subsequent Illumina reaction. DNA samples were then sent for Illumina sequencing at The Next-Generation Sequencing Facility in The Centre for Applied Genomics at the Hospital for Sick Children in Toronto, Ontario, Canada. Data analysis of the sequencing reads was performed according to our colleagues' protocol (31).

Luminescence assay.

Vaginal swab samples were collected from participants using the ESwab (Conan) swab and transport system containing 1 ml liquid Amies medium. Following sample collection, the ESwab was submerged into the Amies medium for 2 h at room temperature, followed by a 10-s vortex at high speed. The swab was removed and discarded. The ESwab solution was aliquoted into two portions of 350 μl each and centrifuged at 10,000 rpm for 5 min. The supernatants were transferred to a fresh tube and stored at −80°C. These were later used for the in vitro luminescence assay.

A culture of a reporter strain, S. aureus MN8, containing the cloning vector pAmilux with an incorporated promoter for tst, was grown in brain heart infusion (BHI) medium with 10 μg/μl chloramphenicol overnight at 37°C with a shaking speed of 250 rpm (19, 32). The culture was diluted to an optical density at 600 nm (OD₆₀₀) of 0.02 with (i) supernatant from the clinical sample, (ii) Amies transport medium, or (iii) BHI, of which the last two acted as controls. Next, 200 μl of the subculture was grown for 16 h in a Fluoroskan Ascent FL luminometer (Thermo Scientific), where luminescence production was detected from the activated tst promoter every 30 min. Another 200 μl of the same subculture was grown in a Bioscreen C reader (MTX Lab Systems) at 37°C with continuous shaking for 24 h to monitor growth of the bacteria.

Bacterial cultures and conditions for supernatant challenge assay.

All bacterial strains used in the supernatant challenge assay, along with the respective growth conditions, are listed in Table 1. These bacteria were obtained from either ATCC or previous clinical isolations. The strains were divided into 3 groups for analysis: bacteria from healthy women and AV-associated and BV-associated strains. The S. aureus MN8 strain was selected as a prototype of menstrual TSS strains, as it was isolated from a premenopausal woman suffering from the condition (33). All aerobic bacteria were cultured under shaking conditions for 16 h at 37°C and then subcultured to an OD₆₀₀ of 0.02 using their growth medium. The subcultures were grown for 12 h. The lactobacilli were grown under anaerobic conditions using a GasPak EZ system (Becton, Dickinson) for 24 h at 37°C and subcultured for an additional 24 h, starting at an OD₆₀₀ of 0.02. The strict anaerobes Atopobium vaginae, Gardnerella vaginalis, and Prevotella bivia were cultured and subcultured in a Forma anaerobic chamber (model 1025; Thermo Scientific) for 48 h at 37°C to allow growth into stationary phase. Subcultures underwent centrifugation at 3,500 × g for 10 min at 4°C, and the supernatants were transferred to a fresh tube. The pH of the supernatants was determined and adjusted to that of the S. aureus MN8 supernatant (pH 5.95), sterilized via filtration using 0.2-μm-pore-size filters, transferred to a clean tube, and frozen at −20°C for no longer than 1 week. Next, S. aureus MN8 cultures were grown for 16 h and subcultured starting at an initial OD₆₀₀ of 0.02 in half BHI and half supernatant of the challenger bacteria. The subcultures were grown for 16 h at 37°C under shaking conditions and were then centrifuged at 3,500 × g for 10 min at 4°C. Subcultures were also grown in a Multiskan Ascent plate reader (Thermo Scientific) to detect OD₆₀₀ values to ensure similar S. aureus growth across the various conditions. The supernatants were subsequently used for real-time PCR.

Table 1.

Bacterial strains used for supernatant challenge assay

Bacteria	Source	Growth medium/conditiona
Staphylococcus aureus MN8	Clinical isolate	BHI/aerobic
Escherichia coli J96	Clinical isolate	BHI/aerobic
Streptococcus agalactiae ATCC 13813	ATCC	BHI/aerobic
Enterococcus faecalis ATCC 33186	ATCC	BHI/aerobic
Enterococcus faecium ATCC 19434	ATCC	BHI/aerobic
Atopobium vaginae BAA-55	Reid Culture Collection	VDMP/anaerobic
Prevotella bivia 29303	Reid Culture Collection	VDMP/anaerobic
Gardnerella vaginalis	Clinical isolate	VDMP/anaerobic
Lactobacillus crispatus ATCC 33820	ATCC	MRS/anaerobic
Lactobacillus jensenii ATCC 25258	ATCC	MRS/anaerobic
Lactobacillus johnsonii DSM20553	DSMZ	MRS/anaerobic
Lactobacillus gasseri ATCC 33323	ATCC	MRS/anaerobic
Lactobacillus reuteri RC-14	Clinical isolate	MRS/anaerobic
Lactobacillus rhamnosus GR-1	Clinical isolate	MRS/anaerobic
Lactobacillus iners AB-1	Clinical isolate	MRS/anaerobic

^aVDMP, vaginally defined medium plus 0.5% proteose peptone; MRS, de Man, Rogosa, and Sharpe.

Quantitative real-time PCR.

A culture of S. aureus MN8 challenged with supernatants from the other bacteria (as described above) was grown from an initial OD₆₀₀ of 0.02 for 16 h. Ten milliliters of RNAprotect (Qiagen) was added to 5 ml culture, and the solution was vortexed and incubated for 10 min at room temperature and then centrifuged at 6,000 × g for 20 min at 4°C. The supernatant was then discarded, and the pellet was stored at −80°C. To isolate bacterial RNA, the pellet was suspended in 5 ml lysis solution (10 mM Tris-HCl, 1 mM EDTA [pH 8.0], 50 μg/ml lysostaphin) and vortexed. The solution was incubated for 10 min at 37°C and then centrifuged at 6,000 × g for 20 min. Isolation of RNA from the pellet was performed using TRIzol reagent (Invitrogen) in accordance with the manufacturer's instructions.

Contaminants were removed from the RNA samples using an RNeasy minikit (Qiagen). RNA products were viewed via electrophoresis in a 1% agarose gel using 1× TBE (Tris-borate-EDTA), stained with ethidium bromide, and viewed under UV light in an AlphaImager (Alpha Innotech Corporation). The RNA concentration and quality in the samples were then determined through a biophotometer (Eppendorf) (RNA quality cutoffs, ≥1.8 at 260/280 nm and ≥1.6 at 260/230 nm). Five hundred nanograms of RNA from each sample was used as a template for reverse transcription-PCR. Conversion to cDNA was done using Multiscribe reverse transcriptase mix (Invitrogen). The reactions were carried out in a Mastercycler apparatus (Eppendorf) with the following program: initial step of 25°C for 10 min, followed by 37°C for 120 min and a final step of 85°C for 5 min.

The cDNA samples were used as the templates for real-time PCRs. Primers included tst-f (5′-CTGATGCTGCCATCTGTGTT-3′) and tst-r (5′-GTAAGCCCTTTGTTGCTTGC-3′) for expression of the tst gene and rpoB-f (5′-TCCTGTTGAACGCGCATGTAA-3′) and rpoB-r (5′-GCTGGTATGGCTCGTGATGGTA-3′) for expression of the rpoB housekeeping gene. The iQ SYBR green Supermix (Bio-Rad) was used to carry out reactions in 20-μl reaction mixtures, which were composed of the following: 10 μl 2× iQ SYBR green Supermix, 1 μl of each primer at 10 μM, 7 μl RNase-free H₂O, and 1 μl cDNA template. Reactions were run in a Rotor-Gene 6000 thermocycler (Corbett) under the following program: an initial melting ramp from 72°C to 95°C, followed by 40 cycles of 95°C for 10 s, 60°C for 15 s, and 72°C for 20 s, ending with a hold at 95°C for 10 min. Data were analyzed using the Rotor-Gene 6000 series software (version 1.7; Corbett). Expression of tst by S. aureus under each condition was compared to tst expression when S. aureus was grown in medium alone. For the purpose of comparison, the control expression was set at 100%. Overall, the supernatant challenge experiment was performed three times, with each run having three technical replicates starting from the reverse transcription stage.

A series of standards was created to quantify gene expression during data analysis. A subculture of S. aureus MN8 was grown for 12 h, and total DNA was isolated using InstaGene matrix (Bio-Rad). PCR was carried out in 50-μl reaction mixtures consisting of the following components: 10× PCR buffer, 1.7 mM MgCl₂, 210 mM dNTP mix, 640 nM each primer (the tst- and rpoB-specific primer pairs used for real-time PCR), 5 U Platinum Taq DNA polymerase (Invitrogen), 5 μl template, and Milli-Q H₂O (Millipore) to reach 50 μl. Amplification was performed in a Mastercycler (Eppendorf) using the following program: 94°C for 1 min, followed by 30 cycles of 94°C for 45 s, 60°C for 45 s, and 72°C for 45 s and a final elongation step of 72°C for 2 min. Products were viewed via electrophoresis in a 1% agarose gel using 1× TBE, stained with ethidium bromide, and viewed under UV light in an AlphaImager (Alpha Innotech Corporation). The concentration of DNA in each sample was determined through spectrophotometry, and a series of standards with concentrations ranging from 5,000 ng/μl to 5 × 10⁻⁹ ng/μl was made. These standards were used as the templates in the real-time PCRs to construct standard curves with known concentrations of DNA.

Statistical analysis.

Statistical analysis was performed on the real-time PCR data using the GraphPad Prism (version 4) program. Gene expression under the various conditions was compared to that for the control using one-way analysis of variance with Dunnett's multiple-comparison test. Changes in gene expression were considered significant for P values of <0.05. The standard error of the mean (SEM) for each condition tested is displayed on the bar plots.

RESULTS

Vaginal microbiota profile of premenopausal women.

The vaginal microbiota of participants was determined through 16S rRNA sequencing by Illumina, and bacterial DNA from 21 participants was successfully isolated and amplified for sequencing (9 healthy participants, 5 participants of intermediate status, and 7 participants with BV). The vaginal microbiotas of healthy versus BV participants were very distinct. The predominant species associated with healthy and intermediate samples was Lactobacillus crispatus (50.7% and 41.6%, respectively), followed by Lactobacillus iners (27.4% and 20.5%, respectively), with an average species abundance of 3.6 species per sample (Fig. 1). Other bacterial species, including those associated with BV, were present in relatively low abundance in these subjects. Three outliers were present, having diversities of 11, 12, and 17 species per sample. One intermediate sample had no L. iners or L. crispatus isolates (at ≥1% abundance) but instead consisted of Lactobacillus jensenii, Lactobacillus gasseri/L. johnsonii, and G. vaginalis of nearly equal abundances. Participants assigned to the BV group, however, had a predominant Gardnerella vaginalis population (average abundance, 49.7%), along with Leptotrichia amnionii (3.9%) and Atopobium vaginae (3.7%). The BV samples had a higher bacterial species diversity than the healthy and intermediate samples, with an average of 10.5 species per sample.

Fig 1.

Vaginal microbiota of 21 subjects recruited from the clinical study, as determined through 16S rRNA sequencing by Illumina. Vaginal state was determined via Nugent scoring and measurement of pH: a score of 1 to 3 is healthy, one of 4 to 6 is intermediate, and one of 7 to 10 is BV. Each column represents a participant (marked by their corresponding identification [ID] number), with each colored bar representing a type of bacterium.

Expression of tst in S. aureus in women with and without bacterial vaginosis.

In order to investigate the ability of the different vaginal environments to influence the production of TSST-1, vaginal samples were collected from nonpregnant, premenopausal subjects in London, Ontario, Canada, and were tested against cultures of the reporter strain S. aureus MN8/pAmilux-Ptst to monitor the activity of the tst promoter (reflecting tst expression). Samples from all three vaginal states (healthy, intermediate, and BV) were tested in this luminescence assay, and following 16 h of growth, expression of tst in S. aureus was found to be completely suppressed from all 3 sample categories, as illustrated in Fig. 2A. However, samples from two healthy participants were found to have no effect (Fig. 2B and C, respectively). Total protein levels in these two samples, as well as in three samples demonstrating strong inhibition, were determined via a bicinchoninic acid (BCA) assay (data not shown). Results showed similar protein levels in all the samples, regardless of anti-tst activity, suggesting that the lack of suppression observed in these samples was likely not due to the absence of host proteins.

Fig 2.

Luminescence activity of the S. aureus MN8 tst gene in response to vaginal swab contents from women with various types of vaginal health (healthy, intermediate, and BV). (A) Vaginal swab samples from all 3 health groups were capable of suppressing tst expression, as shown by three representative samples. (B, C) Two clinical samples from the healthy group failed to suppress tst expression. S, Staphylococcus aureus MN8; numbers, participant identification number; TM, transport medium, the medium in which the swabs were preserved.

TSST-1 production in response to bacteria associated with BV, AV, and a healthy microbiota.

In order to investigate whether vaginal bacterial species have an influence on TSST-1 production, cultures of S. aureus MN8 were challenged with supernatants of AV-associated, BV-associated, and healthy microbiota-associated bacteria, and tst expression was monitored by real-time PCR. The probiotic L. reuteri RC-14 was also included in the experiment, as this strain has been shown to suppress TSST-1 production (19). The bacteria grew to similar OD₆₀₀ values (data not shown). Bacteria used to challenge S. aureus were separated into 3 groups for data analysis: lactobacilli, AV-associated bacteria, and BV-associated bacteria. Several Lactobacillus species were found to significantly decrease tst expression (Fig. 3A): L. johnsonii, L. jensenii, and L. gasseri decreased tst expression by 54%, 59%, and 49%, respectively (P < 0.05). The probiotic strains L. rhamnosus GR-1 and L. reuteri RC-14 decreased expression by 57% and 72%, respectively (P < 0.05). Interestingly, L. iners and L. crispatus (the most common vaginal residents) did not significantly decrease tst expression.

Fig 3.

Fold change in gene expression of tst in S. aureus MN8 in response to supernatants from lactobacilli (A), AV-associated bacteria (B), and BV-associated bacteria (C), detected by real-time PCR. *, P < 0.05.

In general, the AV-associated bacteria increased tst expression. Products secreted from S. agalactiae significantly increased tst expression 3.7-fold relative to that for S. aureus grown in BHI (P < 0.05) (Fig. 3B). E. faecalis and Enterococcus faecium increased tst expression 1.7- and 1.9-fold, respectively, while E. coli did not affect expression. However, the latter changes in gene expression were not statistically significant compared to the gene expression for the S. aureus control grown in BHI. Interestingly, we acquired preliminary data showing that S. aureus challenged with a combination of S. agalactiae and L. reuteri RC-14 supernatants, as well as with S. agalactiae and L. jensenii supernatants, expressed basal levels of the tst gene, suggesting a dampening effect from the lactobacilli (data not shown).

Finally, the BV-associated organisms slightly decreased tst expression. In particular, G. vaginalis, A. vaginae, and P. bivia decreased expression by 15%, 16%, and 27%, respectively (the last of these was deemed statistically significant) (Fig. 3C), compared to the tst expression of S. aureus grown in half BHI and half vaginally defined medium plus 0.5% proteose peptone.

DISCUSSION

To our knowledge, this is the first report of indigenous vaginal bacteria altering TSST-1 production in S. aureus. In particular, several species of Lactobacillus significantly suppressed tst expression, including L. gasseri and L. jensenii, which are common vaginal residents, and L. johnsonii, which is a gut commensal occasionally found in the vagina. Surprisingly, however, L. iners and L. crispatus did not demonstrate tst suppression. We confirmed anti-TSST-1 activity from the vaginal probiotic strain L. reuteri RC-14 as well as a distal urethral isolate, the probiotic L. rhamnosus GR-1.

An interesting discovery was that aerobic bacteria induced tst expression. During menses, increased oxygen levels and a neutral pH increase the chances of TSST-1 production (13–15), and the presence of neighboring aerobic bacteria may be enough for S. aureus to secrete a threshold level of the toxin, leading to disease progression. The finding that S. agalactiae (group B Streptococcus [GBS]) significantly induced expression by more than 3-fold was particularly intriguing, considering the seemingly symbiotic relationship between this species and S. aureus. Multiple studies have found that S. agalactiae inhibits Lactobacillus spp. and G. vaginalis without inhibiting S. aureus (34–36) and that S. aureus colonization is significantly associated with colonization by S. agalactiae in a population of pregnant women (37). Our findings thus illustrate another way in which GBS could promote the virulence of S. aureus. In the context of menstrual TSS, women with AV who are colonized predominantly with these two species might be at a higher risk of developing the disease; elevated oxygen levels in the vagina would presumably induce TSST-1 production in S. aureus (13), and the presence of GBS may further perpetuate toxin production. If this relationship were to hold true in suitable animal studies, the application to women of certain Lactobacillus species intravaginally may be worth considering, since they can displace S. aureus and S. agalactiae from vaginal epithelial cells (38).

A number of Lactobacillus species tested (L. jensenii, L. gasseri, L. johnsonii) suppressed tst expression in the supernatant challenge assay, as did the vaginal probiotic strain L. reuteri RC-14, which is known to confer this effect (19). Screening these supernatants for the cyclic dipeptides linked to L. reuteri RC-14 is under way to determine whether this or another substance is responsible for the tst suppression. Surprisingly, L. rhamnosus GR-1, another urogenital probiotic strain, also suppressed the toxin. A study by Laughton et al. (2006) found that this strain, in contrast to L. reuteri RC-14, was not able to suppress the RNAIII-initiating P3 promoter in S. aureus Newman (39). This suggests that the inhibitory effect on tst expressed by L. rhamnosus GR-1 was not due to inhibition of agr but perhaps arose from altered SarA activity, as this regulator has been shown to bind and activate the tst promoter directly (10). Lastly, L. iners and L. crispatus, the predominant vaginal residents of healthy women, did not demonstrate anti-tst activity. Considering the rare nature of menstrual TSS, this suggests that other factors, such as vaginal metabolites from the host, are at play in preventing S. aureus in the vagina from secreting this toxin. Nevertheless, this finding will aid in identifying a putative tst-suppressing factor(s) secreted by the other Lactobacillus members by comparing their secretory products under the experimental conditions.

The complete contents of vaginal swabs from women of healthy, BV, and intermediate status were introduced into cultures of the reporter strain S. aureus MN8/pAmilux-Ptst. Monitoring the luminescence produced from this strain allowed us to determine tst promoter activity in real time during the growth of S. aureus. It was discovered that samples from participants with all three vaginal states were able to significantly suppress tst expression. However, the microbiota sequencing data showed that these samples had high abundances of L. iners, L. crispatus, and G. vaginalis, all of which failed to suppress tst in the challenge assay of the present study. This suggests that either there is a collective effect of the microbiota or premenopausal women have host-derived factors that prevent the development of menstrual TSS. Identifying host-derived vaginal metabolites shared among both healthy and BV women will be important for shedding light on putative candidate compounds to test. It is important to note that two samples from the healthy group failed to show any suppression, which illustrates that there may be a subpopulation of women who would benefit from TSST-1-suppressive agents more so than the general population. Of note, these were the same samples from healthy women to show increased bacterial diversity, and although no unique bacteria were identified within the samples, this suggests that an altered vaginal environment may contribute to increased menstrual TSS susceptibility. The suppression of tst expression in response to BV samples was surprising, as these samples had pHs greater than 4.5, with some having pHs as high as 6.0, which is more conducive to TSST-1 production (14). However, the effect of pH would likely be minimized, as the transport medium used to preserve the contents of the swabs contained various buffering agents. Levels of oxygen and carbon dioxide would not likely contribute to suppression, as all samples were exposed to atmospheric conditions during collection and preparation.

The vaginal microbiota profiles of the women recruited here were similar to those seen in other studies, in that L. iners, L. crispatus, and G. vaginalis made up the core members of the microbiota (21, 40). However, the microbiota of the subjects varied from person to person, regardless of health status. For instance, 4 of the 5 intermediate samples consisted primarily of L. crispatus and L. iners, while the other sample was composed of L. jensenii, L. gasseri/L. johnsonii, and G. vaginalis. As well, although vaginal swab samples from healthy women typically contain less-diverse populations than BV samples, very complex populations were seen in two healthy women (17 and 12 species per sample) and one woman of intermediate status (11 species per sample) compared to the other healthy women (average, 3.6 species per sample). Members of these populations included Streptococcus and Enterococcus spp. Therefore, healthy women whom S. aureus colonizes may already have a diverse population of bacteria, some of which could release compounds that induce TSST-1 production.

Overall, this study has shown that the vaginal environment of women with BV and with a healthy background is able to suppress TSST-1, consistent with the rare occurrence of menstrual TSS. However, samples from a few subjects failed to suppress the toxin, illustrating that a subpopulation of women may be more susceptible to developing this condition. Characterizing the factor(s) responsible for the observed suppression will help identify a target population of premenopausal women who would benefit the most from prophylactic options, such as the application of exogenous lactobacilli. Perhaps the major finding of this study was that the aerobic bacteria S. agalactiae, E. faecalis, and E. faecium had a TSST-1-inducing effect on S. aureus. This suggests that women with AV or a complex aerobic microbiota, which is occasionally seen during menses, may be at an increased risk of menstrual TSS and would benefit the most from a vaginally applied probiotic.