1. Introduction
Bacterial vaginosis (BV) is the most prevalent vaginal condition among women of reproductive age, affecting approximately 30% of this population globally [1,2]. BV is particularly prevalent among women of childbearing age in sub-Saharan Africa, with rates of 42.1% in those aged 15–24 and 41.2% in those aged 25–49 [3]. Clinically, BV manifests through symptoms such as increased vaginal discharge with a distinctive fish-like odor, vaginal discomfort, and/or urinary symptoms. BV is characterized by a reduction in beneficial Lactobacillus species and by colonization with a diverse spectrum of primarily anaerobic bacteria [4]. BV is associated with a range of negative sexual and reproductive outcomes, such as a higher risk of preterm birth, miscarriage, increased genital inflammation, sexually transmitted infections (STIs), and higher susceptibility to HIV [5,6]. Treatment failure and relapse in up to 60% of patients within 6–12 months [7,8], leading to recurrent (≥3 BV episodes per year) or refractory (no response upon antibiotic treatment) BV, creates a high burden of disease and impairs the quality of life of affected women [9].
Vaginal microbiome communities are complex and can be classified into five community state types (CSTs) [10], although intermediate states and transitions within the menstrual cycle have been described [11,12]. CSTs are characterized by the abundance of specific Lactobacillus species, except for CST IV, which has high diversity [10]. CST IV is commonly associated with BV, but also the L. iners-abundant CST III is considered imbalanced and often shifts into the BV state [13,14,15]. The predominance of non-iners lactobacilli is less commonly observed among African women compared to Caucasian women, highlighting geographical and ethnic differences in vaginal microbiome composition [16,17]. During the development of BV, the vaginal microbiome shifts towards increased bacterial diversity, particularly of anaerobic microbes such as Gardnerella, Prevotella, Fannyhessea (formerly Atopobium), Mobiluncus, and BV-associated bacteria 1 (BVAB1) (Candidatus Lachnocurva vaginae [18]) and BVAB2 and 3 [19]. The formation of a Gardnerella-dominated biofilm was described as characteristic of BV [5] and provides favorable conditions for other anaerobic bacteria to proliferate [19,20]. A symbiotic relationship between Gardnerella and Prevotella strains further establishes the biofilm and degrades the vaginal mucous layer, which in turn increases the adherence of other BVAB [19].
Although it is not always classified as an infection, clinically apparent BV is usually treated with oral or topical metronidazole (MTZ) or clindamycin (CLI) as the standard of care, with emerging experimental alternatives including antiseptic treatments, Lactobacillus-based live biotherapeutics, antimicrobial agents such as endolysins, pro- and prebiotics, and vaginal microbiome transplantation [21]. Several studies have shown a modest reduction in the bacterial abundances of BV-associated microorganisms after antibiotic treatment. While this was inversely associated with increased relative abundances of Lactobacillus species [13,22,23], often this increase is largely represented by L. iners, which has been described as destabilizing itself [1,13,24,25]. The high rate of BV recurrence of 60% may be caused by the inability of antibiotics to eliminate the biofilm [20,26] but also by antibiotic resistance [2,27]. In vitro passaging experiments with sub-minimum inhibitory concentration (MIC) concentrations of MTZ showed that resistance developed after 5–10 passages in six out of six tested Gardnerella strains [28]. Previously, we reported that 12/20 (60%) Gardnerella strains isolated from women with BV were resistant to MTZ, whereas all were susceptible to CLI when grown in the planktonic form [29]. However, when grown as biofilm in vitro, CLI was ineffective against seven out of nine Gardnerella strains [28]. In a larger study by Petrina et al. (2017), which analyzed 605 BV-related bacterial strains, all 110 Gardnerella isolates demonstrated sensitivity to CLI, indicating its potential effectiveness in treating infections caused by this genus [30]. However, 27% of the isolates exhibited resistance to MTZ. Furthermore, this study reported that all 25 Fannyhessea isolates were also sensitive to CLI, yet 18% showed resistance to MTZ [31], suggesting that MTZ may not be a reliable treatment option for a significant portion of these bacteria.
Previous research by Mtshali et al. (2021) demonstrated that treatment with MTZ was associated with a gradual shift of vaginal microbiome profiles away from predominance of strict anaerobes towards an increased abundance of L. iners in the CAPRISA 083 (CAP083) clinical trial [13]. Participants were diagnosed for BV by both Nugent Score (NS) and assessment of clinical symptoms, and were treated according to the South African Management Guidelines for sexually transmitted infections [32]. All enrolled 56 participants completed two follow-up visits at 6 and 12 weeks. The limited efficacy of MTZ in fully eradicating BV-related pathogens raises important concerns regarding treatment success and the potential for recurrence. This highlights the necessity of examining the bacterial resistance profile in the vaginal microbiome of the CAP083 participants, as this may influence treatment outcomes and the persistence of BV.
In the CAP083 study [13,33], intermediate BV was defined as NS = 4–6, BV as NS ≥ 7, and NS ≤ 3 was considered BV negative. At the baseline visit, all participants received a single oral dose of 2 g MTZ. At the 6-week visit, participants with a NS ≥ 4 (39/56) received further treatment with 400 mg MTZ orally daily for 5 days. At the second follow-up visit, 12 weeks after treatment, a microbiological cure rate (NS ≤ 3) of 23% was reported. At all three visits, vaginal swabs were taken prior to treatment to investigate genital inflammation markers and changes in the vaginal microbiota in response to MTZ treatment by 16S rRNA sequencing.
Pharmacokinetic (PK) studies of MTZ conducted by other groups reported a maximum plasma concentration (cmax) of ~40 µg/mL after oral application of 2 g MTZ, dropping to 4–5 µg/mL after 24 h [31,34]. In another study, oral application of 400 mg MTZ twice daily for 7 days resulted in a steady-state concentration of 6.9 µg/mL in plasma [35]. We used these values to interpret MIC values measured for bacterial isolates from CAP083.
In the present study, we isolated bacterial strains from vaginal swabs from CAP083, that is from women with laboratory-diagnosed STIs and/or BV (NS ≥ 4), both prior to and following MTZ exposure. We characterized the resistance levels of the vaginal microbiota and compared the prevalence of MTZ-resistant strains in women in whom treatment failed (NS ≥ 4) with those who were microbiologically cured (NS ≤ 3). Through this investigation, our research sought to enhance understanding of MTZ resistance patterns and their implications for treatment effectiveness, ultimately contributing to the development of improved strategies for managing BV.
3. Discussion
BV is a prevalent condition characterized by dysbiosis of the vaginal microbiome. Despite MTZ being a first-line treatment, many women experience treatment failure and recurrent episodes of BV [13,22,23]. This study utilized culturomics along with previously published 16S rRNA gene sequencing data to assess abundance and resistance levels of multiple BV-associated bacteria before and after MTZ treatment. Our data showed that Gardnerella was the most frequently cultured isolate, while L. iners was underrepresented compared to its relative abundance in 16S rRNA sequencing.
Notably, 3%, 9%, and 14% of isolates at the baseline, 6-week, and 12-week visits, respectively, were identified as L. iners, an organism that is often missed in culture-based studies [38]. This organism has specific nutritional (such as L-cysteine) and environmental requirements that are not adequately replicated in culture media, resulting in its underrepresentation or complete absence in cultured samples [39,40]. Moreover, 16S rRNA sequencing can detect viable but non-culturable states, thus capturing a broad diversity of L. iners that may exist in their natural habitats yet remain unobservable through traditional culturing methods. More studies are needed to culture and investigate the role of L. iners in the vaginal microbiota as a potential predictive biomarker or risk factor for BV, along with its impact on the mucosal immune microenvironment. The increasing relative abundance of L. iners from baseline to follow-up visits is in line with one of the main conclusions of Mtshali and colleagues, that treatment with MTZ strongly increases the overall abundance of this species [13]. Additionally, BVAB1, an uncultured bacterial species [18,41], could not be isolated despite demonstrating a high relative abundance indicated by 16S rRNA gene sequencing [13].
It has been suggested that resistance in Gardnerella spp., along with its role as the primary producer of biofilms that hinder antimicrobial penetration, is associated with reduced efficacy of MTZ, contributing to the persistence and recurrence of BV [1,26,42,43,44,45]. Consistent with these findings, this study observed that a large share of Gardnerella isolates were resistant to MTZ at baseline, and the resistance level increased throughout this study. Additionally, the stable median MIC from baseline to the 6-week visit, followed by a notable increase among participants receiving multiple doses of MTZ by the 12-week visit, suggests the potential development of resistance as a direct consequence of the treatment. We did not investigate the underlying mechanism of acquired resistance to MTZ, as it was described as complex and multifactorial (reviewed in [46]).
In a recent study, Petrina and colleagues evaluated MTZ susceptibility of vaginal isolates from women in the US [30]. They reported a median MIC of MTZ of 8 µg/mL for 110 Gardnerella isolates, with 27% of strains being classified as MTZ-resistant according to CLSI. This contrasts with our study, in which we found a median MIC of 64 µg/mL and 64% resistant Gardnerella isolates at the baseline. The differences between the two studies may come from the geographic locations but also from the study populations, which in the case of Petrina et al. also included women without BV.
The majority of isolated strains at baseline (52%, 76/145) exhibited MICs higher than the maximum plasma concentration of MTZ reported to be reached after oral dosing of 2 g MTZ (~40 µg/mL) in PK studies conducted by other groups [31,34]. At the 6-week visit, 91% (84/92) of determined MIC values exceeded the reported steady-state plasma concentration after dosing twice daily with 400 mg for 7 days (~7 µg/mL) [35]). The plasma concentrations reached in CAP083 participants may have been even lower, because the participants were only dosed once a day for 5 days. The reported levels of MTZ in plasma after similar dose regimens as in CAP083, along with the high median MICs for MTZ among bacterial isolates, may explain the low change in the relative abundance of Gardnerella observed in CAP083 after treatment. This is also reflected in the low cure rate of only 23% at the 12-week follow-up. Instead, the 5-day treatment after the 6-week visit significantly increased the median MIC of Gardnerella and L. iners isolates. Consistently, previous studies have reported recurrence rates of BV of up to 60% within 12 months after oral MTZ treatment, with MTZ resistance being a contributing factor [2,8,9,28]. Our analysis indicates a certain association between pre-existing Gardnerella resistance and poor outcomes of MTZ treatment. However, the high share of participants with at least one MTZ-resistant Gardnerella strain, which was 70% and 46% before the first and second treatment, respectively, may have confounded a clear outcome.
Furthermore, we observed a relatively stable average abundance of Gardnerella in participants throughout this study (re-analyzed data from [13]) across NS states. The low reduction in Gardnerella and the low cure rate of 23% in CAP083 is consistent with (but obviously does not prove) a keystone role of Gardnerella in BV [19]. Other studies report 4-week cure rates of above 60% after oral MTZ therapy [47], but high recurrence rates within 12 months [13].
A spread of MIC values from sensitive to highly resistant Gardnerella strains was observed within individual swabs. This spread was also found at the 12-week visit, indicating that i) strains with different MTZ sensitivity coexist in the vaginal microbiome and that ii) sensitive strains can persist after treatment when resistant strains are present. This might indicate that MTZ-resistant strains protect sensitive ones, at least at the dose regimen applied in CAP083. A mechanistic model that would explain these observations could be that the outer layers of the biofilm reduce the MTZ exposure of the inner layers, meaning that resistant bacteria might be most abundant in the outer layers of the biofilm. Recent studies reported that all currently known Gardnerella species can coexist in single microbiomes [48], that Gardnerella species diversity is high in BV [49], and that different Gardnerella species may contribute differently to BV [50], adding another layer of complexity. Consistently, the re-analysis of CAP083 NS and 16S rRNA sequencing data [13] indicated a lack of significant changes in the average Gardnerella abundance across all participants before and after treatment.
This study encountered several limitations that may influence the interpretation of its findings. Firstly, the reliance on existing 16S rRNA sequencing data may restrict our ability to fully capture the complexity of the vaginal microbiome, particularly for underrepresented species like L. iners. Furthermore, the isolation of only eight isolates per participant and timepoint may not adequately reflect the full diversity of strains present in the vaginal microbiome. Additionally, we were unable to differentiate between Gardnerella species, which may have distinct roles in BV and could contribute variably to treatment outcomes. Lastly, the study’s participants may have had a history of BV treatment prior to enrollment, which could skew results by leading to higher rates of Gardnerella detection. This prior treatment history may influence the observed resistance patterns within the cohort, thereby affecting the overall conclusions regarding antibiotic resistance and treatment efficacy.
To better understand the pathogenesis of BV, future studies are needed to investigate the coexistence of sensitive and resistant bacteria. Also, knowledge of the most important bacterial genera that need to be targeted to allow the vaginal microbiome to shift away from CSTs III and IV needs to be deepened to develop an improved therapy. Our study highlights the need for more effective treatment options due to widespread resistance to the current standard of care in bacteria associated with disease relapse and recurrence.
Source link
Timo Schwebs www.mdpi.com
