Bacterial vaginosis (BV) is a dysbiosis of the vaginal flora characterized by a shift from a Lactobacillus-dominant environment to a polymicrobial mixture including Actinobacteria and Gram-negative bacilli. BV is a common vaginal condition in women and is associated with increased risk of sexually transmitted infection and adverse pregnancy outcomes such as preterm birth. Gardnerella vaginalis is one of the most frequently isolated bacterial species in BV. However, there has been much debate in the literature concerning the contribution of G. vaginalis to the etiology of BV, since it is also present in a significant proportion of healthy women. Here we present a new murine vaginal infection model with a clinical isolate of G. vaginalis. Our data demonstrate that this model displays key features used clinically to diagnose BV, including the presence of sialidase activity and exfoliated epithelial cells with adherent bacteria (reminiscent of clue cells). G. vaginalis was capable of ascending uterine infection, which correlated with the degree of vaginal infection and level of vaginal sialidase activity. The host response to G. vaginalis infection was characterized by robust vaginal epithelial cell exfoliation in the absence of histological inflammation. Our analyses of clinical specimens from women with BV revealed a measureable epithelial exfoliation response compared to women with normal flora, a phenotype that, to our knowledge, is measured here for the first time. The results of this study demonstrate that G. vaginalis is sufficient to cause BV phenotypes and suggest that this organism may contribute to BV etiology and associated complications. This is the first time vaginal infection by a BV associated bacterium in an animal has been shown to parallel the human disease with regard to clinical diagnostic features. Future studies with this model should facilitate investigation of important questions regarding BV etiology, pathogenesis and associated complications.
Citation: Gilbert NM, Lewis WG, Lewis AL (2013) Clinical Features of Bacterial Vaginosis in a Murine Model of Vaginal Infection with Gardnerella vaginalis. PLoS ONE 8(3): e59539. doi:10.1371/journal.pone.0059539
Editor: Adam J. Ratner, Columbia University, United States of America
Received: January 14, 2013; Accepted: February 15, 2013; Published: March 19, 2013
Copyright: © 2013 Gilbert et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported in part by a Basil O’ Connor award from the March of Dimes (to ALL), a Translational Research Pilot Grant funded by the Barnes-Jewish Hospital Foundation (to ALL and WGL), and a postdoctoral fellowship from the American Heart Association (to NMG). The study was also supported in part by NIH grant P50DK064540-11 (PI: Hultgren). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
One in three women in the U.S. have bacterial vaginosis (BV) , a microbial imbalance of the vaginal flora characterized by the absence of normally dominant lactobacilli and overgrowth of complex communities dominated by Gram-negative bacteria and Actinobacteria , . BV can be asymptomatic, maybe even part of a spectrum of ‘normal’ from the patient perspective, but often displays characteristic clinical features, including “thinning” of vaginal fluid secretions, increased pH (>4.5), a fishy odor upon potassium hydroxide treatment, and the presence of clue cells (epithelial cells studded with bacteria) in wet mounts. An additional defining feature of BV is the presence of vaginal sialidase –, an enzyme that cleaves terminal sialic acid residues from complex glycans, which are abundant on host cell surfaces and secreted mucus proteins –. Women with BV are at increased risk of pelvic inflammatory disease, infections following surgery or other routine gynecologic procedures, sexually transmitted infections including HIV, and serious pregnancy complications such as intrauterine infection and preterm birth –.
Unlike most common infectious diseases, BV appears to be polymicrobial in nature. Recent genomic studies have illustrated the complexity and heterogeneity of BV, which can vary in bacterial composition from day to day and from one individual to another , , –. Although more than a dozen bacterial species have been associated with BV, the potential causal contributions of each to the biochemical, cellular, and clinical features of BV remain elusive. Gardnerella vaginalis was the first bacterium implicated in the pathogenesis of BV and continues to be associated with the disease . However, there has been much debate in the literature concerning the contribution of G. vaginalis to the development and pathogenesis of BV. G. vaginalis can be isolated/detected from asymptomatic women that do not meet the criteria for BV diagnosis at the time of detection , , –, causing some to question its potential role in BV. However, consistent with the notion of G. vaginalis as a potential pathogen, strains identified as G. vaginalis have been isolated from invasive perinatal infections –. Moreover, several investigations have described the pathogenic potential of some G. vaginalis isolates in cell adhesion and entry, cytolytic toxin production, biofilm formation, and other phenotypes that may reflect virulence , –.
One important diagnostic feature of BV is the presence of clue cells, which are thought to be exfoliated epithelial cells coated with bacteria. G. vaginalis, among other BV-associated bacteria, has been shown to interact with vaginal epithelial cells in culture , , and clinical studies have shown that vaginal specimens from women with BV have adhered bacteria on their surfaces , –. However, experimental investigation of the potential role of Gardnerella vaginalis in generating clue cells in vivo requires an animal model with features of human BV.
Upon infectious challenge, epithelial cells within the urinary and genital tracts can undergo a process termed exfoliation, in which superficial cells appear to be actively shed from the epithelial surface. In some cases, this is beneficial to the host, flushing potential pathogens from the mucosa , , while in other cases, potential pathogens can take advantage of their access to underlying mucosal tissue , . Despite much speculation in the medical literature regarding the potential causes of epithelial exfoliation in BV, including enzymes such as sialidase and prolidase or the combination of amines and organic acids produced by anaerobes , –, the degree of vaginal epithelial shedding among women with BV or normal flora does not appear to have been directly measured or reported in the clinical literature. Distinguishing whether epithelial exfoliation is actively induced in BV is necessary for establishing the pathophysiology of the disease and may be important for understanding why women with BV are at increased risk of secondary urogenital and intrauterine infections.
Defining a role for G. vaginalis (or other associated bacteria) in BV pathogenesis has been hampered by the absence of robust small animal models that displays phenotypes seen in human BV. Here we describe the development of a new murine model of G. vaginalis vaginal infection. This model displays several key features of BV, including presence of vaginal sialidase activity, and the presence of epithelial cells with attached bacteria (reminiscent of clue cells). Additionally, we provide the first quantitative evaluation of vaginal epithelial exfoliation in BV, demonstrating increased shedding of epithelial cells in both G. vaginalis infected mice and in clinical specimens from women with BV compared to mock-infected mice or women with normal flora respectively. This epithelial response is contrasted by an absence of inflammatory cell infiltrate, consistent with the lack of vaginal inflammation found in human BV throughout the literature –. The results from this murine model suggest that G. vaginalis alone is sufficient to yield BV phenotypes and provide further justification for considering G. vaginalis as a contributor to the causes and complications associated with BV.
Murine Vaginal and Ascending Uterine Infection by G. vaginalis Occurs in the Absence of an Overt Inflammatory Infiltrate
To establish a G. vaginalis murine vaginal infection model, we used a sialidase-positive G. vaginalis strain (JCP8151B), a clinical isolate from a woman with BV. From our initial experiments we found that the majority of β-estradiol-treated C57/Bl6 mice contained vaginal flora producing large, mucoid colonies on Gardnerella semi-selective media that occluded the smaller G. vaginalis colonies. These ‘contaminating’ vaginal flora colonies were resistant to the addition of a combination of sulfamethoxazole, trimethoprim sulfate, gentamicin and perfloxacin to selection plates. Previous vaginal models of G. vaginalis did not acknowledge the presence of endogenous flora or describe methods by which G. vaginalis was distinguished and measured , . We performed sequencing of genes encoding 16S rRNA from isolated colonies of the murine vaginal flora and identified several isolates to be species of Enterococcus (faecalis and gallinarum). To circumvent this issue and enable accurate determination of G. vaginalis titers in the murine reproductive tract, we isolated a spontaneous streptomycin resistant (SmR) mutant of JCP8151B. We confirmed that the SmR isolate contained the canonical mutation in the Rpsl gene (K43N) and displayed logarithmic growth rate and sialidase activity indistinguishable from the parent strain (data not shown). Using this new JCP8151B-SmR strain, we inoculated G. vaginalis vaginally into C57/Bl6 mice and determined CFU levels in vaginal washes and vaginal homogenates at 24 and 72 hours post infection (hpi) (Fig. 1A–B). Colonies presumed to be G. vaginalis due to growth on streptomycin-containing plates were also confirmed by PCR using primers reported to be specific for G. vaginalis  (data not shown).
G. vaginalis JCP8151B-SmR4 titers were determined by enumerating colony forming units (CFU) in vaginal washes (A) and tissue homogenates (B) at 24 and 72 hpi. For samples containing no colonies, the limit of detection for each was determined and is displayed instead of a value of 0. Results are meta data from 2 independent experiments, each with 10 mice/infection group. (C) Histological inflammation was assessed by hematoxylin-eosin (H&E) staining of formalin-fixed, paraffin-embedded vaginal tissue sections.
BV has been characterized clinically as a microbial condition that most often lacks obvious signs of inflammation in vaginal tissue . To determine the extent of histological inflammation (edema, neutrophil infiltrate) in our murine model, we performed hematoxylin and eosin (H&E) staining on vaginal tissues collected at 24 and 72 hpi. Similar to clinical observations of BV, G. vaginalis infection did not result in marked tissue inflammation, edema, or polymorphonuclear (PMN) cell infiltrate (Fig. 1C). The mean CFU levels in both vaginal washes and homogenates decreased significantly from 24 hpi to 72 hpi (Fig. 1A–B), suggesting clearance of the bacteria is occurring by PMN-independent mechanisms.
There was a strong positive correlation between CFU determined from vaginal washes and vaginal homogenates (Fig. 2A). Additional experiments determined that G. vaginalis could persist in the murine vagina of ∼50% of mice for as long as 8 days post infection (data not shown). G. vaginalis was also found at low levels (mean 69.5 CFU/g) in the uterine horns of 55% and 45% of mice at 24 and 72 hpi, respectively. As might be expected, animals with higher titers of vaginal bacteria were more likely to exhibit ascending infections of the uterine horns (Fig. 2B–C).
Vaginal wash and tissue titers determined as in Fig. 1 were plotted and analyzed using GraphPad® Prism 5.0. ***P<0.001; *P<0.05. Results are meta data from 2 independent experiments, each with 10 mice/infection group.
G. vaginalis Leads to Vaginal Sialidase Activity in vivo
A hallmark feature of bacterial vaginosis is the presence of high levels of sialidase activity in vaginal fluid compared to specimens from women with normal flora. Our JCP8151B-SmR4 isolate produces robust sialidase activity in culture. To determine whether JCP8151B-SmR4 expresses sialidase activity during vaginal infection, we performed sialidase activity assays on vaginal washes collected above. Sialidase activity was present in washes from 67% of G. vaginalis infected mice at 24 hpi, while the majority (86%) of mock-infected mice contained no detectable sialidase activity (Fig. 3A). Isolation of vaginal bacteria from the few mock-infected animals with vaginal sialidase activity demonstrated that these mice were colonized with sialidase-positive Eubacteria consortium or Enterococcus spp. (data not shown). The level of sialidase activity present in washes of infected animals correlated positively with G. vaginalis CFU levels in vaginal washes and homogenized vaginal tissue, strongly suggesting that the observed sialidase activity in infected animals is in fact produced by G. vaginalis (Fig. 3B and C). Together these results strongly suggest that G. vaginalis expresses sialidase in the murine vagina and for the first time establish a prominent biochemical feature of BV in a murine infection model.
(A) Sialidase enzymatic activity was determined in vaginal washes from mock and G. vaginalis infected mice at 24 and 72 hpi and statistical significance determined by the Mann-Whitney test. The dotted line represents the cutoff value used to distinguish sialidase (+) from sialidase (−) vaginal washes for performing Fisher’s exact test, shown in the box above. Results are meta data from 4 independent experiments (24 hpi-mock n = 44, G. vaginalis n = 70; 72 hp-mock n = 29, G. vaginalis n = 40). The sialidase activity value from each G. vaginalis-infected mouse was plotted against G. vaginalis CFU in the corresponding vaginal wash (B) or tissue homogenate (C) and Spearman correlations were determined. ***P<0.0001; **P<0.005. (D) Fisher’s exact test to determine whether ascending uterine horn infection by G. vaginalis was contingent upon the presence of vaginal sialidase activity. Results shown represent combined data from both 24 and 72 hpi, though each time point was significant when analyzed individually.
A greater percentage of mice displaying vaginal sialidase activity also had G. vaginalis in their uterine horns than those mice that were sialidase negative (Fig. 3D). These data are consistent with the observation that higher vaginal titers increase the propensity for ascending infection.
G. vaginalis Interacts with Murine Vaginal Epithelial Cells in vitro and in vivo
Another feature of BV, and a component of the Amsel criteria for BV diagnosis, is the presence of epithelial cells with adherent bacteria, termed clue cells. G. vaginalis has been shown to interact with cultured human vaginal epithelial cells and experiments using vaginal biopsy  and vaginal fluid  samples suggests that bacteria may form an adherent biofilm on epithelial cells in the human vagina. Vaginal washes from G. vaginalis infected mice often contained clumps of epithelial cells with apparent attached bacteria (Fig. 4A, panel b, arrowheads).
(A) Phase contrast light microscopy of vaginal wash wet mounts (panels a and b) and H&E staining of vaginal tissue sections (panels c and d) from mock (a and c) and G. vaginalis (b and d) infected mice. Arrows in panel b and dark purple puncta in panel d illustrate potentially adherent bacteria in G. vaginalis-infected samples. (B) Fluorescent confocal microscopy images (with corresponding bright-field images) of vaginal epithelial cells either mock (top) or RBITC labeled G. vaginalis (bottom) infected in vitro for 3 h at 37°C. (C) Fluorescent confocal microscopy images of vaginal washes from mock or RBITC labeled G. vaginalis infected mice, collected at 4 hpi.
Additionally, abundant hematoxylin-rich puncta, indicative of adherent bacteria, were also apparent in some samples upon H&E staining of vaginal sections (Fig. 4A, panel d). These results suggested that G. vaginalis may interact with murine vaginal epithelial cells; however, we could not definitively distinguish G. vaginalis from murine vaginal flora using histology. Previously used antibodies for immunofluorescence  were unavailable. Therefore, to provide further confidence that G. vaginalis could interact with murine vaginal epithelial cells we performed infection assays with fluorescently labeled G. vaginalis. First we assessed adherence ex vivo using epithelial cells present in vaginal washes from uninfected, β-estradiol treated mice. Epithelial cells were washed extensively to remove endogenous flora, then infected with fluorescently labeled G. vaginalis and visualized by fluorescent confocal microscopy. G. vaginalis adhered to mouse vaginal epithelial cells, appearing as distinct fluorescent puncta or biofilm-like collections decorating the epithelial cell surface (Fig. 4B). Similar puncta were not observed on epithelial cells inoculated with a “mock” preparation (containing the fluorescent label but no bacteria), demonstrating that this pattern of fluorescence was specific to G. vaginalis and not an artifact of the introduction of the RBITC label. Next we determined whether we could observe such interactions during in vivo infection. Following vaginal inoculation with fluorescently labeled G. vaginalis, similar fluorescent puncta were detected on epithelial cells in vaginal washes from infected mice at 4 hpi. Together, these results demonstrate that G. vaginalis adheres to murine vaginal epithelial cells, similar to what has been seen for human-derived cultured epithelial cells , .
G. vaginalis Infection Results in Robust Epithelial Exfoliation
While G. vaginalis did not elicit a robust inflammatory response, we observed that the epithelial surfaces of G. vaginalis infected mice displayed evidence of epithelial cell exfoliation (see Fig. 1C). To gain a semi-quantitative perspective of this phenotype, we scored the degree of histological exfoliation in slides that were blinded to the observer, with 0 being none and 3 being very robust (see Fig. 5B for representative images). Vaginal sections from G. vaginalis-infected mice had significantly higher exfoliation scores compared to mock-infected controls (Fig. 5A). Additionally, G. vaginalis infection resulted in increased thickness of the transitional epithelium (Fig. 5C), which correlated positively with exfoliation score (Fig. 5D), suggesting that there may be increased epithelial proliferation in response to the surface exfoliation. Inoculation of heat-killed G. vaginalis did not result in a significant increase in either epithelial exfoliation score or thickness (Fig. 5A&C), demonstrating that live bacteria are required to induce this response.
(A) Histological exfoliation scoring. H&E stained, formalin-fixed, paraffin-embedded vaginal tissue sections from the 24 hpi time point were assessed for evidence of epithelial exfoliation (eosin-rich layers of epithelial cells superficial to the transitional epithelium) and assigned a numerical value from 0–3, with 0 = none and 3 = very robust. The Kruskal-Wallis test was used to for statistical evaluation of differences between groups (P = 0.0161). For pairwise comparisons, post-hoc testing was performed using the Mann-Whitney U-test, with significance indicated in the figure. *P<0.05. Upon conservative correction for multiple comparisons with Bonferroni-Holm, the non-parametric tests remain significant. (B) Representative images of epithelial scoring, with white, capped lines representing the measured thickness of the transitional epithelial layer shown in (C). (C) Using the same samples from (A), average thickness of the transitional epithelium was determined from five measurements per sample using StreamStart® software. Results in each graph are meta-data from 4 independent experiments (mock n = 36, G. vaginalis n = 25, heat-killed (HK) G. vaginalis n = 10). The data passed the D’Agostino & Perason omnibus normality test. Therefore, a one-way ANOVA was used detected significant differences between groups (P = 0.0004), followed by post-hoc pairwise comparisons using the unpaired t-test. **P<0.01; ***P<0.001. Again, differences remained significant with the conservative Bonferroni correction. (D) Transitional epithelium thickness values from (C) were plotted against exfoliation scores from (A) and Spearman correlation was determined. ***P<0.001.
As an additional assessment of this apparent exfoliation response, we examined vaginal washes by wet mount light microscopy. Consistent with our observations of H&E stained vaginal sections, vaginal washes from both mock and G. vaginalis infected mice contained predominantly epithelial cells with very limited, if any, leukocytes (representative images shown in Fig. 6A). We enumerated epithelial cells, again in a blinded manner, and found that G. vaginalis infected mice had significantly higher numbers of epithelial cells in vaginal washes compared to mock-infected animals or those exposed to heat-killed G. vaginalis (Fig. 6B). Finally, the degree of epithelial exfoliation in mice infected with G. vaginalis correlated positively with both vaginal wash CFU and sialidase activity (Fig. 6C–D), consistent with this response being relative to infectious burden. Together these results show that G. vaginalis JCP8151B induces a robust vaginal epithelial exfoliation response in a murine vaginal infection model.
(A) Representative images of phase contrast light microscopy fields of vaginal wash wet mounts from mock, G. vaginalis and HK G. vaginalis infection groups, as indicated, used in (B) to enumerate epithelial cells. (B) The average numbers of epithelial cells per field of view were calculated from 5 images per sample. Results are meta-data from 4 independent experiments (mock n = 51, G. vaginalis n = 74, HK G. vaginalis n = 19). The Kruskal-Wallis test was used to for statistical evaluation of differences between groups (P = 0.0065). The Mann-Whitney U-test was used for post-hoc pairwise comparisons, with significance indicated in the figure. **P<0.01; *P<0.05. Differences remained significant with the conservative Bonferroni-Holm correction for multiple comparisons. (C) and (D) Epithelial counts were plotted against their corresponding vaginal wash CFU (C) or sialidase activities (D) and Pearson correlations were determined.
Epithelial Exfoliation as a Clinical Feature of Bacterial Vaginosis
Pathogen induction and blockade of host epithelial exfoliation responses has been described in other murine urogenital infection models. For example, uropathogenic E. coli (UPEC) induces exfoliation of superficial umbrella cells lining the bladder, which is thought to be consistent with the progression of acute UTI in humans –. However, N. gonorrhea has been shown to block vaginal epithelial exfoliation through interactions with human specific receptors (carcinoembryonic antigen-related cell adhesion molecules, CEACAMs) , . Although the presence of epithelial clue cells is a well-established clinical feature of BV, we found no examples in the literature measuring whether the relative number of epithelial cells in clinical specimens is increased in women with BV compared to those with normal flora. To determine whether the increased number of epithelial cells seen in our murine model is a verifiable feature of human BV, we performed microscopic enumeration of vaginal epithelial cells on slides prepared from human vaginal swabs from women with (+) and without (−) BV as defined by Nugent score (7–10 and 0–3 respectively) (Fig. 7). Consistent with the murine model, significantly higher levels of epithelial cells were observed on slides prepared from specimens of women with BV compared to women with normal flora (Fig. 7). These results strongly suggest that an epithelial exfoliation host response occurs in the clinical setting of BV.
Heat-fixed, Gram stained slides were prepared from fresh human vaginal swabs and classified as BV (+) or no BV (−) by Nugent scoring (scores 7–10 and 0–3 respectively). The average number of epithelial cells per field of view was calculated from 3 images per sample. BV (−) n = 28, BV (+) n = 27. The Mann-Whitney U-test was used for statistical evaluation of differences between groups. ***P<0.0001.
Bacterial vaginosis (BV) is a common vaginal condition in women ,  and is associated with increased risk of sexually transmitted infection and adverse pregnancy outcomes, including preterm birth –. Despite its prevalence, the etiolog(ies) of BV symptoms and complications are poorly understood. There is an obvious and dramatic shift in the overall vaginal flora from a Lactobacillus-dominant state to one overrun by high titers of Gram-negative anaerobes and Actinobacteria. A handful of bacteria have come to be known as BV-associated bacteria, including G. vaginalis. However the contribution of these BV-associated bacteria to the overall disease state is largely undefined. A significant contributing factor to this dearth of understanding is the absence of relevant small animal models.
We found a few recent reports in the literature describing murine vaginal inoculation with G. vaginalis. Two of these papers investigated the effects of Lactobacillus probiotic strains on G. vaginalis colonization in an outbred mouse strain , . However, the methods of isolation and enumeration of G. vaginalis vaginal titers in these papers were only loosely described. For example, G. vaginalis recovered from infected mice were reported as a percentage of the no probiotic control group rather than an absolute enumeration of recovered colony forming units, hampering assessment of the overall bacterial load. Although possibly explained by a difference in mouse strain used or conditions of the housing or breeding facility, there was also no reference to the incidence of contaminating vaginal flora, which we found invariably present in our mouse model. In addition to these two studies, there is one report of vaginal G. vaginalis infection in gnotobiotic mice , which naturally circumvents the issue of contaminating flora.
Here we describe a murine vaginal infection model with G. vaginalis that is, to our knowledge, the first to recapitulate key BV phenotypes. Technically speaking, this model is rather straightforward and very similar to infection models utilized for other vaginal pathogens. However, the model described here is distinguishable by the fact that previous studies of G. vaginalis infection in mice did not investigate BV-related phenotypes in vivo. Furthermore, we took multiple measures to ensure accuracy of our G. vaginalis titer enumeration, including 1) generating a streptomycin-resistant (SmR) isolate, and 2) confirming by PCR that bacteria re-isolated from mouse vaginal washes were G. vaginalis. Vaginal inoculation with G. vaginalis was sufficient to yield 1) vaginal sialidase activity; 2) epithelial cells with adherent bacteria (reminiscent of clue cells); 3) epithelial exfoliation in the absence of an inflammatory infiltrate. Together, results from this model demonstrate that G. vaginalis likely plays a role in generating these clinical features observed in humans and each is discussed in further detail below.
Sialidase activity is rarely detected in women with normal flora , , . In pregnant women, sialidase activity has also been independently correlated with increased risk of chorioamnionitis and PTB , . Production of sialidase by isolated BV-associated bacteria grown in culture strongly suggests that BV-associated sialidases are bacterial in origin , . Bacterial sialidases have been characterized as virulence factors in bacterial infections of various mucosal sites . Our previous in vitro studies have highlighted the potential role of G. vaginalis sialidase in the deglycosylation and degradation of host glycoproteins . In our murine model, the level of sialidase activity correlates with vaginal G. vaginalis titers and sialidase positive mice are more likely to develop uterine horn infection. These data suggest that higher vaginal titers are more likely to result in ascending infection. Although further studies are required, these results may also suggest potential role for sialidase in the mechanisms facilitating ascending infection. Consistent with this idea, sialidase activity levels in pregnant women correlated directly with increased risk of chorioamnionitis and preterm birth , .
G. vaginalis and the Formation of Clue Cells
It has long been established that BV bacteria interact with epithelial cells. In fact, the presence of exfoliated clue cells is a qualitative diagnostic feature of the disease , . It has previously been suggested that G. vaginalis may be responsible for clue cell formation, since it was detected on the surface of exfoliated vaginal epithelial cells more frequently and at higher levels than the BV-associated anaerobes Mobiluncus, Bacteroides, and Fusobacterium . A more recent high-resolution phylogenetic study examining correlations between different species of BV bacteria and clinical features revealed that G. vaginalis is positively associated with the presence of clue cells . G. vaginalis has also been shown to interact with a vaginal epithelial cell line  and epithelial cells present in human clinical samples , . We demonstrated, using fluorescently labeled bacteria, that G. vaginalis interacts with murine vaginal epithelial cells, forming clue-like cells in vitro. Labeled G. vaginalis was also detected on the surface of epithelial cells recovered from vaginal washes of mice following infection with G. vaginalis. Together these data provide evidence that our murine model of G. vaginalis infection displays this important feature of BV and provides an avenue for investigating clue cell formation in vivo.
G. vaginalis Induces Epithelial Exfoliation in the Absence of Overt Inflammation
BV is characterized by a heavy overgrowth of Actinobacteria and Gram-negative anaerobes but a surprising absence of the type of inflammatory infiltrate seen in other urogenital infections such as gonococcal infection ,  or urinary tract infection , , . Epithelial shedding, or exfoliation, appears to be a common mechanism of protection employed by mucosal surfaces , . G. vaginalis vaginal infection in mice produced a robust exfoliation response that correlated directly with vaginal titers and vaginal sialidase activity.
Consistent with the lack of overt inflammation in women with BV , our histological and wet mount microscopy analyses of vaginal specimens from mice infected with G. vaginalis displayed no signs of polymorphonuclear leukocyte (PMN) recruitment. Even though estradiol treatment suppresses the influx of PMNs that naturally occurs in mice after ovulation , literature evidence has shown that PMN recruitment to the vagina can still occur upon infectious challenge in C57/Bl6 mice . Although no quantitative data was given, Teixeira et al. reported the presence of ‘inflammatory lesions’ in the vaginas of gnotobiotic mice infected with G. vaginalis . It is reasonable to suspect that indigenous microflora in the vagina may contribute to host innate immune responses, as has been shown in the gut , . This may influence inflammatory responses to G. vaginalis in mice lacking endogenous flora. Further studies are required to provide a better understanding of the role of G. vaginalis in the apparent suppression of inflammatory responses that occurs during BV. Recent biochemical and genomic investigations have revealed that G. vaginalis isolates can be remarkably diverse –, perhaps allowing different host responses to various strains of G. vaginalis.
Epithelial Exfoliation as a Measurable Clinical Feature: Implications for Understanding BV
Epithelial exfoliation has long been discussed in the BV literature, most often with regard to clue cells. Although there are countless references to “increased epithelial exfoliation” in BV, we found no examples of quantitative analysis in a defined experimental or clinical setting. We present evidence that an increased vaginal epithelial exfoliation response, a robust feature of the G. vaginalis animal model described here, is also apparent in Gram-stained images of vaginal fluids from women with BV (Nugent 7–10) compared to normal controls (Nugent 0–3). Our analysis of clinical samples provides the first concrete evidence to classify an exfoliation response as a BV phenotype.
The modest but significant increase in the number of exfoliated epithelial cells in women with BV may be a beneficial response if it removes adherent potential pathogens and when surface epithelial layers can be replenished. However, excessive exfoliation may promote access to underlying tissue, which may facilitate the establishment of BV-associated bacteria and increase the risk of secondary infection. In fact, BV is known to be associated with increased risk of certain sexually-transmitted infections (STI) , ,  and some vaginal pathogens can take advantage of the exfoliation process to facilitate access to underlying tissue. For example, T. vaginalis causes contact-dependent cytotoxicity upon adherence to vaginal epithelial cells, thereby leading to exfoliation and erosion of the epithelium. It has been suggested that this may allow trichomonads into extracellular matrix and basement membrane sites within the vaginal tissue . It is possible that removal of the outer epithelial cell layer by G. vaginalis provides a niche for formation of an adherent biofilm, which has been observed in vaginal biopsies from women diagnosed with BV . Interestingly, BV correlates with T. vaginalis infection, therefore it is conceivable that exfoliation induced by G. vaginalis and T. vaginalis could be mutually beneficial and may also impact other vaginal organisms. Interestingly, some vaginal microbicides have been shown to have paradoxical effects, actually increasing susceptibility to HIV and other sexually-transmitted pathogens. This increased susceptibility was found to be coincident with rapid exfoliation and re-growth of epithelial cell layers . These previous findings are consistent with the idea that epithelial exfoliation in BV may contribute to increased STI risk. Future studies should examine whether epithelial exfoliation may contribute to the overall risk of secondary infections associated with BV.
We found a single reference in the literature that performed quantitative analysis of vaginal epithelial shedding (by counting epithelial cells present in vaginal lavage samples). Interestingly, this study uncovered a link between vaginal epithelial exfoliation and smoking , a behavior that has been shown to be associated with BV . If exfoliation promotes vaginal colonization by BV bacteria, it is tempting to speculate that the link between smoking and BV could be explained, at least in part, by the initiation of epithelial exfoliation that occurs in smokers. Ultimately, the downstream ramifications of epithelial exfoliation for the overall pathophysiology of BV remain to be explored and the new murine model presented here provides a valuable tool for these investigations.
G. vaginalis as a Pathogen
Since it was first described, there has been vigorous debate in the literature regarding the role of G. vaginalis in BV. G. vaginalis is one of the most frequently isolated bacterial species from women with BV. Consistent with the notion of G. vaginalis as a potential pathogen, strains identified as G. vaginalis have been isolated from placenta, amniotic fluid, and blood –. G. vaginalis has also been implicated in uterine infections and development of endometritis . Results from a comparison of epithelial adhesion, cytotoxicity and biofilm formation between several BV associated bacteria suggested that G. vaginalis may be more virulent than other species associated with the disease . The observation that G. vaginalis produces BV phenotypes in our murine model, in the absence of other BV-associated bacteria, emphasizes its likely role in BV etiology. However, the main controversy appears to lie in the fact that G. vaginalis can also be detected from women with normal flora , , –. We argue that the presence of G. vaginalis in healthy individuals does not constitute a basis for disregarding this bacterium in the causes and complications of BV. Just as “healthy” people can be asymptomatic carriers of such pathogens as Streptococcus pneumoniae , Group A Streptococcus , , Haemophilus influenzae  or Clostridium difficile , carrier states may also exist for G. vaginalis. Indeed, recent genomic and phenotypic studies support the hypothesis that variations in bacterial strain virulence, titers, and/or windows of host susceptibility may bring a colonization state to a state of pathogenesis , , , , .
In summary, our results demonstrate for the first time that G. vaginalis is sufficient to yield key BV phenotypes in an animal model. The quantitative experimental methods described here show that infection with G. vaginalis leads to vaginal sialidase activity, bacterial adherence to vaginal epithelial cells, and a robust epithelial exfoliation response that we demonstrate as a relevant clinical feature of BV in quantitative, controlled experiments. These data provide strong evidence that G. vaginalis can play an active role in generating clinical features associated with BV. The phenotypic parallels to human BV displayed in the murine system provide a new experimental tool with great potential to expand our understanding of G. vaginalis-host interactions in the vagina.
Materials and Methods
Vaginal swabs were collected as part of the Contraceptive CHOICE project  according to protocols approved by the Washington University Institutional Review Board (IRB ID# 201108155). Mouse experiments were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals. The protocol was approved by the Animal Studies Committee of Washington University School of Medicine (Protocol Number: 20110149).
Bacterial Strains and Growth Conditions
Gardnerella vaginalis clinical isolate JCP8151B (GenBank JX860320) was obtained from a vaginal swab from a woman with BV (based on Nugent score) obtained in accordance with IRB-approved protocols in collaboration with the Washington University Contraceptive CHOICE Project (IRB ID# 201108155). The vaginal swab was transported from the clinic to the lab using Port-A-Cul™ pre-reduced anaerobic transport media tubes (BD). Tubes were brought into a Vinyl anaerobic airlock chamber (Coy Products) under an atmosphere maintained at approximately 1% hydrogen and 0 ppm oxygen. Within 24 hours and the swab was used to inoculate “Gardnerella semi-selective media,” (agar plates with 5% defibrinated sheep blood, 10 mg/L colistin, 10 mg/L nalidixic acid, and 4 mg/L amphotericin B). Plates were pre-incubated in the chamber for at least 16 hours for equilibration to anaerobic conditions and were incubated anaerobically post-inoculation at 37°C for 24–48 hours. Translucent pinpoint colonies were isolated, and candidate G. vaginalis strains were tested by diagnostic PCR using primers previously reported to be specific for G. vaginalis (forward primer GGGCGGGCTAGAGTGCA and reverse primer GAACCCGTGGAATGGGCC) . Additional validation of the G. vaginalis identity was obtained by sequencing 16S rDNA. Full details of this strain are being reported in a separate manuscript.
A spontaneous streptomycin resistant mutant, JCP8151B-SmR #4, was isolated by plating JCP8151B (concentrated from a 2 day, 25 mL NYC-III culture) on Gardnerella semi-selective media +1 mg/mL streptomycin and selecting resistant colonies after incubating anaerobically at 37°C for 72 hours. We confirmed the streptomycin resistance in this isolate was due to point mutation of the Rpsl gene (as that reported for other bacteria) by amplifying G. vaginalis Rpsl with primers rpsL F2 (CATGGTTTAAGGTGTGCTG) and rpsL R (GTTAATCAACTGAGCCACG) and sequencing using rpsL F2. To confirm that the point mutation did not result in any apparent growth defects or changes in sialidase activity, JCP8151B and JCP8151B-SmR #4 were grown anaerobically overnight in 5 ml NYC-III medium at 37°C, then diluted to OD600 of 0.1 in 5 mL NYC-III medium. A 25 µL aliquot of each bacterial suspension was analyzed for sialidase activity as described below. The remaining bacterial suspension was incubated anaerobically at 37°C for 24 hr to monitor growth. Aliquots were removed at 0, 1, 3, 6 and 24 h, serial diluted and plated on Gardnerella semi-selective media to enumerate colonies. Results for sialidase activity and growth curves were indistinguishable between JCP8151B and JCP8151B-SmR. The JCP8151B-SmR isolate was used for murine infection model experiments described below. Control experiments (as dictated in figures and figure legends) treated a parallel group of mice infected with JCP8151B-SmR that was heat-killed by incubation of the bacterial inoculum at 80°C for 10 min.
Murine Vaginal Infection Model
Female C57/Bl6 mice (6–8 weeks) were injected intraperitoneally with 0.5 mg β-estradiol in 100 µL filter-sterilized sesame oil three days prior to and on the day of inoculation. Mice were anaesthetized with isofluorane and inoculated vaginally with ∼5×107 CFU G. vaginalis in 20 µL sterile PBS (OD600 = 5.0). Vaginal washes were collected by flushing vaginas with 50 µL sterile PBS using a P200 pipet (GeneMate), pipetting up and down 10x, followed by rinsing into an additional 10 µL PBS in a sterile 1.5 mL Eppendorf tube. G. vaginalis titers were determined from washes by preparing 10-fold serial dilutions in PBS (in the anaerobic chamber) and spotting 5 µL of each dilution in quadruplicate onto 1 mg/mL streptomycin selection plates (either Gardnerella semi-selective media or NYC-III agar). Colonies were then enumerated and reported as recovered colony forming units (CFU) per mL of vaginal fluid. Vaginal washes were also analyzed for sialidase activity and epithelial exfoliation as described below.
Mice were sacrificed at 24 hpi or 72 hpi to harvest vaginas and uterine horns. One uterine horn and half of the vagina (bisected longitudinally) from each mouse was homogenized followed by serial dilution and plating as for vaginal washes. Colonies were enumerated and reported as CFU per gram of tissue. The remaining vaginal tissue and uterine horn from each mouse were fixed in 10% buffered formalin phosphate at room temperature followed by paraffin embedding. Histological slide preparation and H&E staining were performed by the Washington University School of Medicine Histology Core.
Sialidase Activity Assays
Vaginal wash samples collected as described above (25 µL) were diluted 1∶2 with 100 mM sodium acetate pH 5.5 containing 300 µM 4-methylumbelliferyl-Neu5Ac (50 µL). Substrate hydrolysis was monitored using a Tecan M200 plate reader.
G. vaginalis Interaction with Murine Vaginal Epithelial Cells in vitro and in vivo
G. vaginalis JCP8151B was fluorescently labeled with Rhodamine B isothiocyanate (RBITC; Aldrich 283924). RBITC was prepared fresh at 0.2 mg/mL in 20 mM HCl and 5 µL of this stock was added to 500 µL G. vaginalis JCP8151B in sterile PBS (prepared as described for mouse experiments above). The bacteria were then incubated anaerobically at 37°C for 30 min, centrifuged, resuspended in 500 µL NYC-III and allowed to recover with an additional 30 min, 37°C anaerobic incubation. Finally, the labeled bacteria were washed twice with PBS and resuspended in 500 µL PBS for inoculation of epithelial cells, as described below. For each experiment a “mock” labeled sample, lacking bacteria, was prepared in parallel.
While G. vaginalis was being labeled, vaginal washes were collected from β-estradiol-treated mice, as described above. Washes were pooled and spun at 300 g for 5 min to collect epithelial cells. Epithelial cells were then washed 3 times with sterile PBS to remove the majority of endogenous flora and then distributed in 50 µL aliquots into 1.5 mL eppendorf tubes. RBITC-labeled G. vaginalis or the “mock” labeled sample (5 µL) were added to the epithelial cells and the samples were rotated at 37°C for 3 hours. Finally, epithelial cells were washed twice to remove unassociated bacteria and then visualized on an Olympus BX61 confocal fluorescent microscope using SlideBook 5.0 software.
For in vivo analyses, G. vaginalis was labeled with RBITC as described for in vitro experiments and then inoculated vaginally into β-estradiol-treated mice as described above. An additional group of mice was inoculated with a “mock”-label preparation containing no bacteria. At 4 hpi, mouse vaginas were washed with 50 µL PBS and epithelial cells were visualized by confocal fluorescent microscopy as described above.
Analysis of Murine Epithelial Cell Exfoliation
H&E stained mouse vaginal histology sections collected above were visualized on an Olympus BX61 microscope to assess the degree of inflammation (24 hpi and 72 hpi time points) and epithelial exfoliation (24 hpi time point). Images were captured and epithelial thickness was measured using StreamStart® software, with averages calculated from 5 measurements per vaginal section.
For assessment of epithelial exfoliation in mouse vaginal washes, wet mounts were prepared with 5 µL vaginal wash (from the 24 hpi time point) and visualized by phase contrast microscopy using an Olympus BX61 microscope. Five representative images were captured from each specimen (1 per mouse) and epithelial cells were counted from each image to determine an average.
Clinical Specimen Handling and Analysis of Epithelial Exfoliation
Vaginal swabs (Starplex) were collected as part of the Contraceptive CHOICE project [104)] according to protocols approved by the Washington University Institutional Review Board (IRB ID# 201108155) and underwent Nugent scoring using published methods as previously described , . Gram stained slides (the same used for Nugent scoring) were analyzed for epithelial exfoliation by enumerating epithelial cells in three representative images as described for murine samples above.
GraphPad Prism 5.0 software was used for all statistical analyses presented. The statistical tests used to analyze each set of data are indicated in the figure legends.
We thank Jeffrey Peipert, Jenifer Allsworth, Gina Secura, and Jennifer Bick for contributions of Contraceptive CHOICE project clinical specimens and confirmatory data. We recognize and thank Justin Perry and Lloyd Robinson for technical assistance. We are thankful for help from Fredrick Kraus with analyzing H&E stained histology slides and are grateful to Jennifer Lodge for use of her confocal microscope.
Conceived and designed the experiments: NMG WGL ALL. Performed the experiments: NMG WGL. Analyzed the data: NMG WGL ALL. Wrote the paper: NMG WGL ALL.
- 1. Allsworth JE, Peipert JF (2007) Prevalence of bacterial vaginosis: 2001–2004 National Health and Nutrition Examination Survey data. Obstetrics and gynecology 109(1): 114–120. doi: 10.1097/01.aog.0000247627.84791.91
- 2. Srinivasan S, Liu C, Mitchell CM, Fiedler TL, Thomas KK, et al. (2010) Temporal variability of human vaginal bacteria and relationship with bacterial vaginosis. PLoS One 5(4): e10197. doi: 10.1371/journal.pone.0010197
- 3. Oakley BB, Fiedler TL, Marrazzo JM, Fredricks DN (2008) Diversity of human vaginal bacterial communities and associations with clinically defined bacterial vaginosis. Appl Environ Microbiol 74(15): 4898–4909. doi: 10.1128/aem.02884-07
- 4. Smayevsky J, Canigia LF, Lanza A, Bianchini H (2001) Vaginal microflora associated with bacterial vaginosis in nonpregnant women: reliability of sialidase detection. Infect Dis Obstet Gynecol 9(1): 17–22. doi: 10.1155/s1064744901000047
- 5. McGregor JA, French JI, Jones W, Milligan K, McKinney PJ, et al.. (1994) Bacterial vaginosis is associated with prematurity and vaginal fluid mucinase and sialidase: results of a controlled trial of topical clindamycin cream. Am J Obstet Gynecol 170(4): 1048–1059; discussion 1059–1060.
- 6. Cauci S, Guaschino S, Driussi S, De Santo D, Lanzafame P, et al. (2002) Correlation of local interleukin-8 with immunoglobulin A against Gardnerella vaginalis hemolysin and with prolidase and sialidase levels in women with bacterial vaginosis. J Infect Dis 185(11): 1614–1620. doi: 10.1086/340417
- 7. Cauci S, McGregor J, Thorsen P, Grove J, Guaschino S (2005) Combination of vaginal pH with vaginal sialidase and prolidase activities for prediction of low birth weight and preterm birth. Am J Obstet Gynecol 192(2): 489–496. doi: 10.1016/j.ajog.2004.07.023
- 8. Howe L, Wiggins R, Soothill PW, Millar MR, Horner PJ, et al. (1999) Mucinase and sialidase activity of the vaginal microflora: implications for the pathogenesis of preterm labour. Int J STD AIDS 10(7): 442–447. doi: 10.1258/0956462991914438
- 9. Wiggins R, Crowley T, Horner PJ, Soothill PW, Millar MR, et al. (2000) Use of 5-bromo-4-chloro-3-indolyl-alpha-D-N-acetylneuraminic acid in a novel spot test To identify sialidase activity in vaginal swabs from women with bacterial vaginosis. J Clin Microbiol 38(8): 3096–3097.
- 10. Varki NM, Varki A (2007) Diversity in cell surface sialic acid presentations: implications for biology and disease. Lab Invest 87(9): 851–857. doi: 10.1038/labinvest.3700656
- 11. Varki A (2007) Glycan-based interactions involving vertebrate sialic-acid-recognizing proteins. Nature 446(7139): 1023–1029. doi: 10.1038/nature05816
- 12. Lewis AL, Lewis WG (2012) Host sialoglycans and bacterial sialidases: a mucosal perspective. Cell Microbiol 14(8): 1174–1182. doi: 10.1111/j.1462-5822.2012.01807.x
- 13. Marrazzo JM, Martin DH, Watts DH, Schulte J, Sobel JD, et al. (2010) Bacterial vaginosis: identifying research gaps proceedings of a workshop sponsored by DHHS/NIH/NIAID. Sex Transm Dis 37(12): 732–744. doi: 10.1097/olq.0b013e3181fbbc95
- 14. Hillier SL, Nugent RP, Eschenbach DA, Krohn MA, Gibbs RS, et al. (1995) Association between bacterial vaginosis and preterm delivery of a low-birth-weight infant. The Vaginal Infections and Prematurity Study Group. The New England journal of medicine 333(26): 1737–1742. doi: 10.1056/nejm199512283332604
- 15. Watts DH, Krohn MA, Hillier SL, Eschenbach DA (1990) Bacterial vaginosis as a risk factor for post-cesarean endometritis. Obstetrics and gynecology 75(1): 52–58.
- 16. Hitti J, Hillier SL, Agnew KJ, Krohn MA, Reisner DP, et al. (2001) Vaginal indicators of amniotic fluid infection in preterm labor. Obstetrics and gynecology 97(2): 211–219. doi: 10.1016/s0029-7844(00)01146-7
- 17. Wiesenfeld HC, Hillier SL, Krohn MA, Amortegui AJ, Heine RP, et al. (2002) Lower genital tract infection and endometritis: insight into subclinical pelvic inflammatory disease. Obstetrics and gynecology 100(3): 456–463. doi: 10.1016/s0029-7844(02)02118-x
- 18. Wiesenfeld HC, Hillier SL, Krohn MA, Landers DV, Sweet RL (2003) Bacterial vaginosis is a strong predictor of Neisseria gonorrhoeae and Chlamydia trachomatis infection. Clin Infect Dis 36(5): 663–668. doi: 10.1086/367658
- 19. Zhang X, Xu X, Li J, Li N, Yan T, et al. (2002) [Relationship between vaginal sialidase bacteria vaginosis and chorioammionitis]. (Translated from chi). Zhonghua Fu Chan Ke Za Zhi 37(10): 588–590 (in chi)..
- 20. Svare JA, Schmidt H, Hansen BB, Lose G (2006) Bacterial vaginosis in a cohort of Danish pregnant women: prevalence and relationship with preterm delivery, low birthweight and perinatal infections. BJOG 113(12): 1419–1425. doi: 10.1111/j.1471-0528.2006.01087.x
- 21. Rezeberga D, Lazdane G, Kroica J, Sokolova L, Donders GG (2008) Placental histological inflammation and reproductive tract infections in a low risk pregnant population in Latvia. Acta Obstet Gynecol Scand 87(3): 360–365. doi: 10.1080/00016340801936487
- 22. Donders GG, Van Calsteren K, Bellen G, Reybrouck R, Van den Bosch T, et al. (2009) Predictive value for preterm birth of abnormal vaginal flora, bacterial vaginosis and aerobic vaginitis during the first trimester of pregnancy. BJOG 116(10): 1315–1324. doi: 10.1111/j.1471-0528.2009.02237.x
- 23. Gajer P, Brotman RM, Bai G, Sakamoto J, Schütte UM, et al. (2012) Temporal dynamics of the human vaginal microbiota. Sci Transl Med 4(132): 132ra152. doi: 10.1126/scitranslmed.3003605
- 24. Ravel J, Gajer P, Abdo Z, Schneider GM, Koenig SS, et al. (2011) Vaginal microbiome of reproductive-age women. Proc Natl Acad Sci U S A 108 Suppl 14680–4687. doi: 10.1073/pnas.1002611107
- 25. Ling Z, Kong J, Liu F, Zhu H, Chen X, et al. (2010) Molecular analysis of the diversity of vaginal microbiota associated with bacterial vaginosis. BMC Genomics 11: 488. doi: 10.1186/1471-2164-11-488
- 26. Zozaya-Hinchliffe M, Lillis R, Martin DH, Ferris MJ (2010) Quantitative PCR assessments of bacterial species in women with and without bacterial vaginosis. J Clin Microbiol 48(5): 1812–1819. doi: 10.1128/jcm.00851-09
- 27. Santiago GL, Tency I, Verstraelen H, Verhelst R, Trog M, et al. (2012) Longitudinal qPCR study of the dynamics of L. crispatus, L. iners, A. vaginae, (sialidase positive) G. vaginalis, and P. bivia in the vagina. PLoS One 7(9): e45281. doi: 10.1371/journal.pone.0045281
- 28. Gardner HL, Dukes CD (1955) Haemophilus vaginalis vaginitis: a newly defined specific infection previously classified non-specific vaginitis. Am J Obstet Gynecol 69(5): 962–976.
- 29. Schellenberg JJ, Links MG, Hill JE, Dumonceaux TJ, Kimani J, et al. (2011) Molecular definition of vaginal microbiota in East African commercial sex workers. Appl Environ Microbiol 77(12): 4066–4074. doi: 10.1128/aem.02943-10
- 30. Hill JE, Goh SH, Money DM, Doyle M, Li A, et al. (2005) Characterization of vaginal microflora of healthy, nonpregnant women by chaperonin-60 sequence-based methods. Am J Obstet Gynecol 193(3 Pt 1): 682–692. doi: 10.1016/j.ajog.2005.02.094
- 31. Tosun I, Alpay Karaoğlu S, Ciftçi H, Buruk CK, Aydin F, et al. (2007) [Biotypes and antibiotic resistance patterns of Gardnerella vaginalis strains isolated from healthy women and women with bacterial vaginosis]. (Translated from tur). Mikrobiyol Bul 41(1): 21–27 (in tur)..
- 32. Berardi-Grassias L, Roy O, Berardi JC, Furioli J (1988) Neonatal meningitis due to Gardnerella vaginalis. European journal of clinical microbiology infectious diseases : official publication of the European Society of Clinical Microbiology 7(3): 406–407. doi: 10.1007/bf01962348
- 33. Silver HM, Sperling RS, St Clair PJ, Gibbs RS (1989) Evidence relating bacterial vaginosis to intraamniotic infection. Am J Obstet Gynecol 161(3): 808–812. doi: 10.1016/0002-9378(89)90406-7
- 34. Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK, et al. (1988) A case-control study of chorioamnionic infection and histologic chorioamnionitis in prematurity. The New England journal of medicine 319(15): 972–978. doi: 10.1056/nejm198810133191503
- 35. Gelber SE, Aguilar JL, Lewis KL, Ratner AJ (2008) Functional and phylogenetic characterization of Vaginolysin, the human-specific cytolysin from Gardnerella vaginalis. J Bacteriol 190(11): 3896–3903. doi: 10.1128/jb.01965-07
- 36. Marrs CN, Knobel SM, Zhu WQ, Sweet SD, Chaudhry AR, et al. (2012) Evidence for Gardnerella vaginalis uptake and internalization by squamous vaginal epithelial cells: implications for the pathogenesis of bacterial vaginosis. Microbes Infect 14(6): 500–508. doi: 10.1016/j.micinf.2011.12.009
- 37. Patterson JL, Stull-Lane A, Girerd PH, Jefferson KK (2010) Analysis of adherence, biofilm formation and cytotoxicity suggests a greater virulence potential of Gardnerella vaginalis relative to other bacterial-vaginosis-associated anaerobes. Microbiology 156(Pt 2): 392–399. doi: 10.1099/mic.0.034280-0
- 38. Harwich MD Jr, Alves JM, Buck GA, Strauss JF 3rd, Patterson JL, et al (2010) Drawing the line between commensal and pathogenic Gardnerella vaginalis through genome analysis and virulence studies. BMC Genomics 11: 375. doi: 10.1186/1471-2164-11-375
- 39. Cook RL, Reid G, Pond DG, Schmitt CA, Sobel JD (1989) Clue cells in bacterial vaginosis: immunofluorescent identification of the adherent gram-negative bacteria as Gardnerella vaginalis. J Infect Dis 160(3): 490–496. doi: 10.1093/infdis/160.3.490
- 40. Villegas-castrejon H, Villanueva-diaz C, Arredondo-garcia JL, Narcio-reyes L, Karchmer-krivitzky S (1992) Bacterial vaginosis associated with G vaginallis/Mobiluncus sp: ultrastructural parameters. Arch AIDS Res 6(3): 177–182.
- 41. De Boer JM, Plantema FH (1988) Ultrastructure of the in situ adherence of Mobiluncus to vaginal epithelial cells. Can J Microbiol 34(6): 757–766. doi: 10.1139/m88-129
- 42. Swidsinski A, Mendling W, Loening-Baucke V, Ladhoff A, Swidsinski S, et al. (2005) Adherent biofilms in bacterial vaginosis. Obstetrics and gynecology 106(5 Pt 1): 1013–1023. doi: 10.1097/01.aog.0000183594.45524.d2
- 43. Bien J, Sokolova O, Bozko P (2012) Role of Uropathogenic Escherichia coli Virulence Factors in Development of Urinary Tract Infection and Kidney Damage. Int J Nephrol 2012: 681473. doi: 10.1155/2012/681473
- 44. Fukushi Y, Orikasa S, Kagayama M (1979) An electron microscopic study of the interaction between vesical epitherlium and E. Coli. Invest Urol 17(1): 61–68.
- 45. Crouch ML, Alderete JF (1999) Trichomonas vaginalis interactions with fibronectin and laminin. Microbiology 145 (Pt 10): 2835–2843.
- 46. Mulvey MA, Schilling JD, Martinez JJ, Hultgren SJ (2000) Bad bugs and beleaguered bladders: interplay between uropathogenic Escherichia coli and innate host defenses. Proc Natl Acad Sci U S A 97(16): 8829–8835. doi: 10.1073/pnas.97.16.8829
- 47. Sobel JD (2000) Bacterial vaginosis. Annu Rev Med 51: 349–356. doi: 10.1146/annurev.med.51.1.349
- 48. Thurman AR, Doncel GF (2011) Innate immunity and inflammatory response to Trichomonas vaginalis and bacterial vaginosis: relationship to HIV acquisition. Am J Reprod Immunol 65(2): 89–98. doi: 10.1111/j.1600-0897.2010.00902.x
- 49. Gonzalez Pedraza Aviles A, Ortiz Zaragoza MC, Irigoyen Coria A (1999) Bacterial vaginosis a “broad overview”. Rev Latinoam Microbiol 41(1): 25–34.
- 50. Cauci S, Culhane JF, Di Santolo M, McCollum K (2008) Among pregnant women with bacterial vaginosis, the hydrolytic enzymes sialidase and prolidase are positively associated with interleukin-1beta. Am J Obstet Gynecol 198(1): 132 e131–137.
- 51. Weissenbacher T, Walter C, Mylonas I, Scholz C, Gingelmaier A, et al. (2010) Interleukin-6, interleukin-10 and interleukin-12 in vaginal fluid from women with bacterial vaginosis. Arch Gynecol Obstet 281(1): 77–80. doi: 10.1007/s00404-009-1072-6
- 52. Cauci S, Guaschino S, De Aloysio D, Driussi S, De Santo D, et al. (2003) Interrelationships of interleukin-8 with interleukin-1beta and neutrophils in vaginal fluid of healthy and bacterial vaginosis positive women. Mol Hum Reprod 9(1): 53–58. doi: 10.1093/molehr/gag003
- 53. Giraldo PC, de Carvalho JB, do Amaral RL, da Silveira Gonçalves AK, Eleutério J Jr, et al. (2012) Identification of immune cells by flow cytometry in vaginal lavages from women with vulvovaginitis and normal microflora. Am J Reprod Immunol 67(3): 198–205. doi: 10.1111/j.1600-0897.2011.01093.x
- 54. Trinh HT, Lee IA, Hyun YJ, Kim DH (2011) Artemisia princeps Pamp. Essential oil and its constituents eucalyptol and alpha-terpineol ameliorate bacterial vaginosis and vulvovaginal candidiasis in mice by inhibiting bacterial growth and NF-kappaB activation. Planta Med 77(18): 1996–2002. doi: 10.1055/s-0031-1280094
- 55. Joo HM, Hyun YJ, Myoung KS, Ahn YT, Lee JH, et al. (2011) Lactobacillus johnsonii HY7042 ameliorates Gardnerella vaginalis-induced vaginosis by killing Gardnerella vaginalis and inhibiting NF-kappaB activation. Int Immunopharmacol 11(11): 1758–1765. doi: 10.1016/j.intimp.2011.07.002
- 56. Fredricks DN, Fiedler TL, Thomas KK, Oakley BB, Marrazzo JM (2007) Targeted PCR for detection of vaginal bacteria associated with bacterial vaginosis. J Clin Microbiol 45(10): 3270–3276. doi: 10.1128/jcm.01272-07
- 57. Danielsson D, Teigen PK, Moi H (2011) The genital econiche: focus on microbiota and bacterial vaginosis. Ann N Y Acad Sci 1230: 48–58. doi: 10.1111/j.1749-6632.2011.06041.x
- 58. Fredricks DN, Fiedler TL, Marrazzo JM (2005) Molecular identification of bacteria associated with bacterial vaginosis. The New England journal of medicine 353(18): 1899–1911. doi: 10.1056/nejmoa043802
- 59. Justice SS, Hunstad DA (2012) UPEC hemolysin: more than just for making holes. Cell Host Microbe 11(1): 4–5. doi: 10.1016/j.chom.2012.01.001
- 60. Dhakal BK, Mulvey MA (2012) The UPEC pore-forming toxin alpha-hemolysin triggers proteolysis of host proteins to disrupt cell adhesion, inflammatory, and survival pathways. Cell Host Microbe 11(1): 58–69. doi: 10.1016/j.chom.2011.12.003
- 61. Mulvey MA, Schilling JD, Hultgren SJ (2001) Establishment of a persistent Escherichia coli reservoir during the acute phase of a bladder infection. Infect Immun 69(7): 4572–4579. doi: 10.1128/iai.69.7.4572-4579.2001
- 62. Mulvey MA, Lopez-Boado YS, Wilson CL, Roth R, Parks WC, et al. (1998) Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science 282(5393): 1494–1497. doi: 10.1126/science.282.5393.1494
- 63. Muenzner P, Rohde M, Kneitz S, Hauck CR (2005) CEACAM engagement by human pathogens enhances cell adhesion and counteracts bacteria-induced detachment of epithelial cells. J Cell Biol 170(5): 825–836. doi: 10.1083/jcb.200412151
- 64. Muenzner P, Bachmann V, Zimmermann W, Hentschel J, Hauck CR (2010) Human-restricted bacterial pathogens block shedding of epithelial cells by stimulating integrin activation. Science 329(5996): 1197–1201. doi: 10.1126/science.1190892
- 65. Teixeira GS, Carvalho FP, Arantes RM, Nunes AC, Moreira JL, et al. (2012) Characteristics of Lactobacillus and Gardnerella vaginalis from women with or without bacterial vaginosis and their relationships in gnotobiotic mice. J Med Microbiol 61(Pt 8): 1074–1081. doi: 10.1099/jmm.0.041962-0
- 66. Bradshaw CS, Morton AN, Garland SM, Horvath LB, Kuzevska I, et al. (2005) Evaluation of a point-of-care test, BVBlue, and clinical and laboratory criteria for diagnosis of bacterial vaginosis. J Clin Microbiol 43(3): 1304–1308. doi: 10.1128/jcm.43.3.1304-1308.2005
- 67. Myziuk L, Romanowski B, Johnson SC (2003) BVBlue test for diagnosis of bacterial vaginosis. J Clin Microbiol 41(5): 1925–1928. doi: 10.1128/jcm.41.5.1925-1928.2003
- 68. Cauci S, Culhane JF (2011) High sialidase levels increase preterm birth risk among women who are bacterial vaginosis-positive in early gestation. Am J Obstet Gynecol 204(2): 142 e141–149.
- 69. Briselden AM, Moncla BJ, Stevens CE, Hillier SL (1992) Sialidases (neuraminidases) in bacterial vaginosis and bacterial vaginosis-associated microflora. J Clin Microbiol 30(3): 663–666.
- 70. Moncla BJ, Braham P, Hillier SL (1990) Sialidase (neuraminidase) activity among gram-negative anaerobic and capnophilic bacteria. J Clin Microbiol 28(3): 422–425.
- 71. Lewis AL, Lewis WG (2012) Host sialoglycans and bacterial sialidases: a mucosal perspective. Cellular microbiology.
- 72. Lewis WG, Robinson LS, Perry J, Bick JL, Peipert JF, et al. (2012) Hydrolysis of secreted sialoglycoprotein immunoglobulin A (IgA) in ex vivo and biochemical models of bacterial vaginosis. The Journal of biological chemistry 287(3): 2079–2089. doi: 10.1074/jbc.m111.278135
- 73. Amsel R, Totten PA, Spiegel CA, Chen KC, Eschenbach D, et al. (1983) Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations. Am J Med 74(1): 14–22. doi: 10.1016/0002-9343(83)91112-9
- 74. Blackwell A, Barlow D (1982) Clinic diagnosis of anaerobic vaginosis (non-specific vaginitis). A practical guide. Br J Vener Dis 58(6): 387–393. doi: 10.1136/sti.58.6.387
- 75. Srinivasan S, Hoffman NG, Morgan MT, Matsen FA, Fiedler TL, et al. (2012) Bacterial communities in women with bacterial vaginosis: high resolution phylogenetic analyses reveal relationships of microbiota to clinical criteria. PLoS One 7(6): e37818. doi: 10.1371/journal.pone.0037818
- 76. Villegas H, Arias F, Flores E, Casanova G, Karchmer S (1997) Ultrastructural characteristics of Gardnerella vaginalis infection in the heterosexual couple. Arch Androl 39(2): 147–153. doi: 10.3109/01485019708987913
- 77. Sarantis H, Gray-Owen SD (2012) Defining the roles of human carcinoembryonic antigen-related cellular adhesion molecules during neutrophil responses to Neisseria gonorrhoeae. Infect Immun 80(1): 345–358. doi: 10.1128/iai.05702-11
- 78. Liu Y, Feinen B, Russell MW (2011) New concepts in immunity to Neisseria gonorrhoeae: innate responses and suppression of adaptive immunity favor the pathogen, not the host. Front Microbiol 2: 52. doi: 10.3389/fmicb.2011.00052
- 79. Hang L, Haraoka M, Agace WW, Leffler H, Burdick M, et al. (1999) Macrophage inflammatory protein-2 is required for neutrophil passage across the epithelial barrier of the infected urinary tract. J Immunol 162(5): 3037–3044.
- 80. Schilling JD, Martin SM, Hung CS, Lorenz RG, Hultgren SJ (2003) Toll-like receptor 4 on stromal and hematopoietic cells mediates innate resistance to uropathogenic Escherichia coli. Proc Natl Acad Sci U S A 100(7): 4203–4208. doi: 10.1073/pnas.0736473100
- 81. Cliffe LJ, Humphreys NE, Lane TE, Potten CS, Booth C, et al. (2005) Accelerated intestinal epithelial cell turnover: a new mechanism of parasite expulsion. Science 308(5727): 1463–1465. doi: 10.1126/science.1108661
- 82. Monahan JJ, Hall RH (1975) Reduced effect of 50% formamide upon Tm of DNA absorbed on hydroxylapatite columns. Anal Biochem 63(2): 620–623. doi: 10.1016/0003-2697(75)90390-5
- 83. Cauci S (2004) Vaginal Immunity in Bacterial Vaginosis. Curr Infect Dis Rep 6(6): 450–456. doi: 10.1007/s11908-004-0064-8
- 84. Jerse AE, Wu H, Packiam M, Vonck RA, Begum AA, Garvin LE (2011) Estradiol-Treated Female Mice as Surrogate Hosts for Neisseria gonorrhoeae Genital Tract Infections. Front Microbiol 61(8): 1074–81. doi: 10.3389/fmicb.2011.00107
- 85. Yano J, Kolls JK, Happel KI, Wormley F, Wozniak KL, Fidel PL Jr (2012) The acute neutrophil response mediated by S100 alarmins during vaginal Candida infections is independent of the Th17-pathway. PLoS One 7(9): e46311. doi: 10.1371/journal.pone.0046311
- 86. Fagundes CT, Souza DG, Nicoli JR, Teixeira MM (2011) Control of host inflammatory responsiveness by indigenous microbiota reveals an adaptive component of the innate immune system. Microbes Infect 13(14–15): 1121–1132. doi: 10.1016/j.micinf.2011.07.012
- 87. Reading NC, Kasper DL (2011) The starting lineup: key microbial players in intestinal immunity and homeostasis. Front Microbiol 2: 148. doi: 10.3389/fmicb.2011.00148
- 88. Paramel Jayaprakash T, Schellenberg JJ, Hill JE (2012) Resolution and characterization of distinct cpn60-based subgroups of Gardnerella vaginalis in the vaginal microbiota. PLoS One 7(8): e43009. doi: 10.1371/journal.pone.0043009
- 89. Pleckaityte M, Janulaitiene M, Lasickiene R, Zvirbliene A (2012) Genetic and biochemical diversity of Gardnerella vaginalis strains isolated from women with bacterial vaginosis. FEMS Immunol Med Microbiol 65(1): 69–77. doi: 10.1111/j.1574-695x.2012.00940.x
- 90. Udayalaxmi J, Bhat GK, Kotigadde S (2011) Biotypes and virulence factors of Gardnerella vaginalis isolated from cases of bacterial vaginosis. Indian J Med Microbiol 29(2): 165–168. doi: 10.4103/0255-0857.81798
- 91. Santiago GL, Deschaght P, El Aila N, Kiama TN, Verstraelen H, et al.. (2011) Gardnerella vaginalis comprises three distinct genotypes of which only two produce sialidase. Am J Obstet Gynecol 204(5): 450 e451–457.
- 92. Allsworth JE, Peipert JF (2011) Severity of bacterial vaginosis and the risk of sexually transmitted infection. Am J Obstet Gynecol 205(2): 113 e111–116.
- 93. Kaul R, Nagelkerke NJ, Kimani J, Ngugi E, Bwayo JJ, et al. (2007) Prevalent herpes simplex virus type 2 infection is associated with altered vaginal flora and an increased susceptibility to multiple sexually transmitted infections. J Infect Dis 196(11): 1692–1697. doi: 10.1086/522006
- 94. Cone RA, Hoen T, Wong X, Abusuwwa R, Anderson DJ, et al. (2006) Vaginal microbicides: detecting toxicities in vivo that paradoxically increase pathogen transmission. BMC Infect Dis 6: 90.
- 95. Basu J, Mikhail MS, Palan PR, Payraudeau PH, Romney SL (1992) Factors influencing the exfoliation of cervicovaginal epithelial cells. Am J Obstet Gynecol 167(6): 1904–1909. doi: 10.1016/0002-9378(92)91795-c
- 96. Alnaif B, Drutz HP (2001) The association of smoking with vaginal flora, urinary tract infection, pelvic floor prolapse, and post-void residual volumes. J Low Genit Tract Dis 5(1): 7–11. doi: 10.1097/00128360-200101000-00002
- 97. Kristiansen FV, et al. (1987) Isolation of Gardnerella vaginalis in pure culture from the uterine cavity of patients with irregular bleedings. Br J Obstet Gynaecol 94(10): 979–984. doi: 10.1111/j.1471-0528.1987.tb02273.x
- 98. Dahlblom V, Soderstrom M (2012) Bacterial interactions in the nasopharynx - effects of host factors in children attending day-care centers. J Infect Public Health 5(2): 133–139. doi: 10.1016/j.jiph.2011.11.007
- 99. Mahendrappa KB (2010) Rajendra (2010) Upper limit of normal antistreptolysin-O titer in healthy school children. Indian Pediatr 47(7): 629.
- 100. Pires R, Ardanuy C, Rolo D, Morais A, Brito-Avô A, et al. (2010) Emergence of ciprofloxacin-nonsusceptible Streptococcus pyogenes isolates from healthy children and pediatric patients in Portugal. Antimicrob Agents Chemother 54(6): 2677–2680. doi: 10.1128/aac.01536-09
- 101. Ozaki E, Kato H, Kita H, Karasawa T, Maegawa T, et al. (2004) Clostridium difficile colonization in healthy adults: transient colonization and correlation with enterococcal colonization. J Med Microbiol 53(Pt 2): 167–172. doi: 10.1099/jmm.0.05376-0
- 102. Piot P, Van Dyck E, Peeters M, Hale J, Totten PA, et al. (1984) Biotypes of Gardnerella vaginalis. J Clin Microbiol 20(4): 677–679.
- 103. Benito R, Vazquez JA, Berron S, Fenoll A, Saez-Neito JA (1986) A modified scheme for biotyping Gardnerella vaginalis. J Med Microbiol 21(4): 357–359. doi: 10.1099/00222615-21-4-357
- 104. Secura GM, Allsworth JE, Madden T, Mullersman JL, Peipert JF (2010) The Contraceptive CHOICE Project: reducing barriers to long-acting reversible contraception. Am J Obstet Gynecol 203(2): 115 e111–117.
- 105. Nugent RP, Krohn MA, Hillier SL (1991) Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 29(2): 297–301.