Therapeutic Opportunities in the Vaginal Microbiome

ABSTRACT

The reproductive tract of females lies at the core of humanity. The immensely complex process that leads to successful reproduction is miraculous yet invariably successful. Microorganisms have always been a cause for concern for their ability to infect this region, yet it is other, nonpathogenic microbial constituents now uncovered by sequencing technologies that offer hope for improving health. The universality of Lactobacillus species being associated with health is the basis for therapeutic opportunities, including through engineered strains. The manipulation of these and other beneficial constituents of the microbiota and their functionality, as well as their metabolites, forms the basis for new diagnostics and interventions. Within 20 years, we should see significant improvements in how cervicovaginal health is restored and maintained, thus providing relief to the countless women who suffer from microbiota-associated disorders.

SCOPE OF THIS REVIEW

In order to design and apply a therapeutic, there must be a vaginal disease or disorder in need of treatment or a condition influenced by the vaginal microbiota. There must also be a way for a new therapeutic agent to function through the microbiome. Those conditions include bacterial vaginosis (BV), aerobic vaginitis (AV), urinary tract infection, urethritis, cervicitis, vulvovaginal candidiasis, vulvodynia, endometriosis, and chorioamnitis. In addition, sexually transmitted infections are included, as arguably some may be prevented by vaginal microbes. Cancer is included, not only because of the association with viral infection, but also because microbial dysbiosis has been associated with cancers at other sites.

The size of this list demonstrates the enormity of not only the burden of disease among females but also the challenge of reviewing each satisfactorily and suggesting opportunities for novel approaches to prevention and treatment. Rather than provide exhaustive details of these conditions, I will refer the reader to appropriate articles for further insight and try to focus on the elements that offer an opportunity for intervention through microbial manipulation. Much of this will, by necessity, be conjecture, but hopefully reasoned and realistic.

THE CERVICOVAGINAL MICROBIOME AND ITS INFLUENCE ON HEALTH

Studies using various molecular methods over the past 15 years have provided insight into the array of microbes that can be detected in the vagina, vulva/labia, cervix, and uterus (1–18). The list is extensive, exceeding 250 bacterial species as well as yeast, Chlamydia, Archaea, viruses, and protozoa. Using statistical methodologies, attempts have been made to create groupings of bacterial species or genera into which the status of a given female might fall. In the first use of Illumina next-generation sequencing (NGS), we reported two major clusters (community state types [CSTs]) associated with a normal vaginal microbiota, one dominated by Lactobacillus iners and another by Lactobacillus crispatus (7). There were four CSTs strongly associated with BV, and these were dominated by Prevotella bivia, Gardnerella vaginalis, Lachnospiraceae, or a mixture of different species. This study of 132 subjects was more insightful than the one involving four phylum groups proposed from only 8 subjects (6), but there was still consistency in the findings of dominant lactobacilli or a mixed microbiota with depleted lactobacilli. Another study suggested one Lactobacillus grouping and three related to dysbiosis with the order Lachnospiraceae and genera Sneathia and Prevotella being dominant (19). Further studies in pregnant and nonpregnant HIV-negative women (8, 11) proposed five CST groupings: (i) Lactobacillus crispatus dominant; (ii) Lactobacillus gasseri dominant; (iii) L. iners dominant; (iv) Peptoniphilus, Prevotella, and Anaerococcus species and elevated Gardnerella or Ureaplasma abundances; and (v) Lactobacillus jensenii dominant. Despite issues with sampling, storage, processing, DNA extraction kits’ error rates from bias in samples, technical variation with PCR amplification, and the use of different primers and statistical methods (20–22), the creation of subgroups of microbiota is intended to help design better treatment strategies. But is it realistic?

The abject failure of industry to develop novel treatments for vaginal health in 40 or so years surely suggests that creating five to eight new ones is not likely to happen soon. In addition, simply identifying which organisms are present is inadequate information to devise new treatment strategies. Understanding how they got there and what they are doing is critically important. We performed the only transcriptomic study to date and showed that L. iners has the ability to adapt to the very different BV environment from that of a healthy vagina (23). Ideally, such a study should examine changes that occur to move the pendulum towards dysbiosis (Fig. 1). Ravel and others have performed such regular sampling over 16 weeks (9), but not using transcriptomics. Nevertheless, their findings and those of others (24–28) have shown that menstruation, sexual activity, spermicides, douching, and antibiotics can dramatically alter the species abundance. Thus, it is impossible to define a single composition associated with health. So, for all the impressive NGS studies that have been undertaken, the basic principle is the same as it was based on culture results 40 years ago, with a high presence of lactobacilli associated with health (29, 30). It remains to be seen whether women with one of the four species (L. crispatus, L. iners, Lactobacillus jensenii, Lactobacillus gasseri) are more protected against dysbiosis than women who do not harbor such species. To prove this, one would have to provide in vivo evidence that certain factors confer protective functions.

FIGURE 1a.

Different mechanisms by which beneficial microbes might influence vaginal health.

TARGETS FOR THERAPY: MECHANISMS WHEREBY LACTOBACILLI PROTECT THE HOST

One of the features of a microscopic analysis of a vaginal swab from a healthy woman is the number of epithelial cells with few adherent bacteria, mostly presumptive lactobacilli. If so few beneficial bacteria are “protecting” the host, how are they doing it? The theory that hydrogen peroxide (H₂O₂) produced by lactobacilli is the key defense mechanism (31) makes some sense, as this compound could inhibit the growth of pathogens. However, it has not been verified by a study measuring concentrations of H₂O₂ in the vagina. Therefore, the concentration required per milliliter of vaginal fluid or per surface area of epithelium is not known. Also, if it is protective, why do strains producing it, such as L. crispatus, become so apparently easy to displace when BV arises? Arguably, a more plausible protective mechanism, but also one that has not been proven in humans, is that biosurfactants produced by some strains of lactobacilli can cover cell surfaces and interfere with pathogen binding (32, 33) (Fig. 1). This process represents a thermodynamic effect, which has a good theoretical basis (34). Not only do the surface tensions of bacteria and host cells influence bacterial adhesion but so does also the suspending fluid. Thus, biosurfactants can alter the latter and make it less receptive to bacteria. Lactobacillus strains vary in the extent of biosurfactant production (35), but no studies have been done to determine if common vaginal isolates produce more than others. Efforts to purify biosurfactants and define their components and associated gene pathways have been limited, as have efforts to enhance their production. Industrialization of yeast-derived biosurfactants is being attempted (36), but not with vaginal applications in mind.

It is possible that bacteriocins or bacteriocin-like compounds are produced and help protect the host against pathogens, thereby leaving a relatively clear epithelial cell (37). But these compounds tend to have quite restricted target organisms, and it seems difficult to imagine that they control the wide range of BV and AV causative agents. Having said that, a recent study identified bacteriocin-like substances produced by vaginal lactobacilli that inhibited the growth of aerobic bacteria Klebsiella, Staphylococcus aureus, Escherichia coli, and Enterococcus faecalis and the yeast Candida parapsilosis as well as some Lactobacillus species (38). A strain of a somewhat rarer vaginal species, Lactobacillus pentosus, isolated from a healthy Nigerian woman, was found to contain a putative cluster of genes for biosynthesis of a cyclic bacteriocin precursor (39). It is not clear if such a strain would be more effective in African women than the four Lactobacillus species of the CSTs. No such association has been tested to date. While bacteriocins have been known for many years, few if any, and none for the vagina, have been developed as therapeutic agents. If this approach is to be effective, it may require a more targeted bacteriocin, such as directly against one pathogen. Group B streptococci would be one such target, given their prevalence, their danger to newborns, and the need to administer antibiotics in women at the time of delivery (40). An oral probiotic, Streptococcus salivarius, may be one option, given its bacteriocin activity against group B streptococci (41).

The possibility of controlling pathogenesis by coaggregating lactobacilli with pathogens and thereby tying up the latter’s ability to spread across surfaces, or by interfering with virulence expression such as production of toxins, has been considered (37) (Fig. 1). The latter concept suggests that a pathogen may not need to direct its energy to producing a toxin, if it can survive and colonize a surface. From a therapeutic point of view, the existence of a pathogen able to infect the host is not ideal, especially one such as S. aureus that can produce toxic shock toxin. But if the cocci are bound to lactobacilli, this potentially makes them less able to propagate and infect. In the case of Lactobacillus reuteri RC-14, cyclic dipeptides were identified to counter S. aureus toxins (42). Compounds such as these and quorum-sensing molecules, including coumarin, a natural plant phenolic compound, are being considered for antivirulence activity against a spectrum of pathogens (43).

The action of lactobacilli on host cell surfaces has been considered for a number of reasons. The most primitive is to simply block the access of pathogens to the surface. Such a competitive-exclusion model would require high inoculum volumes to cover the vaginal epithelium and repeated application, which seem impractical. Rather than blocking access to the host epithelium, this might provide a surface to which pathogens adhere (coaggregate), and if their subsequent growth is not reduced, they might persist. Nevertheless, some groups continue to pursue this competitive-exclusion concept, with genes that aid adherence and inhibit G. vaginalis proposed as a means for candidate probiotic L. crispatus to improve vaginal health (44). The ability of lactobacilli to improve epithelial integrity by upregulating tight-junction proteins has been shown (45). The question is why this would be advantageous to the lactobacilli. In fact, it may not be important to lactobacilli but rather may be a by-product of the metabolites that they produce. An interesting study on intestinal cells showed that Lactobacillus rhamnosus GG could increase tight-junction proteins Zonula occludens-1, claudin-1, and occludin gene expression by utilizing polyamines (46). Polyamines are found in the vagina and can be used by Trichomonas to adhere and infect (47), so the potential exists for lactobacilli to compete for these compounds and improve epithelial integrity.

The nature of the vagina in reproductive-age women is such that different levels of inflammatory processes occur during the menstrual cycle and pregnancy. The ability to maintain immune homeostasis requires more than microbial modulation, but certainly microbes can influence the process. Alternatively, the process can influence the microbes. For example, during menstruation, the bacterial diversity appears to increase, albeit with variations among individuals (48). The ability to influence mucosal immunity using vaccines, antibodies, and antigens has been successful, but in the urogenital tract, modulation affecting bacterial dysbiosis has not been attempted. One exception has been efforts to create a multisubunit vaccine that elicits antibody against uropathogenic E. coli (49). Studies in mice are encouraging, and the concept has great appeal, as this process would stimulate IgG and not eradicate the commensal lactobacilli as antibiotics do. Other vaccine approaches include using flagellin proteins from Pseudomonas aeruginosa to passively immunize against pyelonephritis (50). Immune modulation to prevent BV has not been attempted, and one study of patients receiving a quadrivalent human papillomavirus (HPV) vaccine (HPV–6, –11, –16, and –18) surprisingly reported an increase in the management rate of Gardnerella/BV in women aged 50 years and older (51). If indeed vaccines increase the risk of BV, this requires closer investigation. A Hungarian vaccine comprising five inactivated strains of lactobacilli administered prophylactically to women was reported to prevent BV, but this approach seems more like a way to circumvent drug regulations requiring to deliver living bacteria than a true vaccine (52). Why would dead or inactivated lactobacilli be better at inducing a protective immune response than live lactobacilli? While one study has shown that heat-treated Lactobacillus paracasei NCC 2461 stabilizes interleukin-10 (IL–10) mRNA (53), IL-10 has not been shown to mediate vaginal protection against pathogens.

Aerobic vaginitis is more of an inflammatory disease than BV (54), so it would be interesting to test if vaccines against E. coli, streptococci, and staphylococci make an impact on this condition.

Overall, there is no clear path at present to enhance the vaginal mucosal immune parameters over the menstrual cycle, to reduce the recurrence of BV. Stimulation of anti-BV host defenses to the point of being effective without increasing inflammation and discharge will not be simple, but one case study using intravaginally administered L. rhamnosus GR-1 (1 × 10⁹ cells in 1 ml once daily for 5 days) suggested that it may potentially work (45).

Can Targets Be Developed from Metabolomics Data?

While the cell surfaces per se of bacteria mediate activity in the host and the response can influence outcomes, the metabolic by-products are also important. The compounds that bacteria produce are clearly influenced by the organisms’ genomic capabilities, the nutrients available, the microbial milieu in which they live, and environmental factors such as pH, mucins, antimicrobials, hormones, host cells, and immune factors.

Studies of the metabolome of vaginal bacteria are relatively recent, using liquid and gas chromatography with mass spectrometry, nuclear magnetic resonance spectroscopy, and databases that can interpret the resultant spectra (55–57). These studies have identified lactic acid, acetic acid, glycerol, and other metabolites consistent with the known capacities of lactobacilli and other vaginal organisms. Our study (57) was the first to identify and verify metabolites associated with high diversity and clinical BV, specifically 2-hydroxyisovalerate and γ-hydroxybutyrate (GHB), but not succinate as others had previously reported (58). This provides a means to develop a diagnostic system to detect GHB in vaginal swabs. If combined with elevated pH, it would likely be an effective tool to identify patients requiring therapy.

For AV, no such studies have been undertaken, but given the nature of the aerobic organisms and infection, it should be feasible to identify levels of lipopolysaccharide or other inflammatory compounds in vaginal fluid.

Engineered Strains to Deliver Vaccines

The encouraging results from probiotic applications to various areas of the host have stimulated the search for ways to further enhance benefits. One way would be to engineer bacteria either to deliver specific factors that promote health or prevent disease or to increase the production of compounds known to provide benefits.

The first efforts to engineer lactic acid bacteria for this purpose came from the insertion of murine IL-2 into lactococci (59). This was a precursor for the creation of strains engineered with biological containment (60) designed to deliver IL-10 to treat inflammatory bowel disease (61). However, these efforts have so far failed to make a clinical impact.

For vaginal applications, Mercenier et al. (62) were the first to consider using lactobacilli to increase mucosal immunity against pathogens such as Chlamydia trachomatis, but these efforts fell by the wayside. With the intent of preventing HIV infection in women, Lee and others (63) engineered an L. jensenii strain that secreted 2D CD4, which recognized a conformation-dependent anti-CD4 antibody and bound HIV type 1 gp120. They have continued their research to express single-chain and single-domain antibodies in L. jensenii 1153-1128 to passively transfer these antibodies to the mucosa and through colonization of the lactobacilli provide protection at the vaginal site of HIV transmission (64). Using a different approach, we genetically modified L. reuteri RC-14 to produce HIV entry or fusion inhibitors fused to the native expression and secretion signals of BspA, Mlp, or Sep proteins capable of blocking the three main steps of HIV entry into human peripheral blood mononuclear cells (65). The expression cassettes were stably inserted into the chromosome, but funding for the project was not forthcoming, so progress was halted. It remains to be seen if these approaches will be sufficiently effective and affordable to warrant large-scale applications, especially to countries where sexual transmission of the virus is most prevalent.

A final recombination strategy is being considered by Israeli researchers following studies that showed that lactobacilli are important for sperm motility and conception (66). The idea is to inhibit this Lactobacillus function and presumably then make the sperm less mobile and more apt to not reach the egg (http://nocamels.com/2013/02/the-new-contraceptive-suppository-that-disables-sperm/). The challenges with this approach are many, including overcoming the sheer number and mobility of the sperm, covering the whole area that semen reaches, and taking into account the vaginal microbiota, which if colonized by Prevotella or Pseudomonas (67) may inhibit fertilization but if colonized by competing lactobacilli may not.

COMMUNITY PROBIOTICS

Given the success of fecal microbiota transplantation in curing Clostridium difficile infections (68), the concept of using a full microbiome from one person to reset the aberrant microbiome of another is being considered for niches other than the gut. The use of antibiotics to prepare the gut for entrance of a new microbiota has been important so far, and potentially it could also eradicate many vaginal organisms. On the other hand, if it led to recalcitrant biofilms, persister strains, and invasion of epithelial cells, as in the bladder (69, 70), this could perhaps reduce the ability of the implanted microbiota to fully take over.

Much research is needed to determine the ideal composition of a donor vaginal microbiota. While the proportion of species may be identifiable, it should be appreciated that these organisms were presumably arranged by early life factors, diet, hormones, and metabolic linkages. Some of these will inevitably differ in a recipient. Nevertheless, this approach will no doubt be tested. This may require a sponge to collect the microbiota from the donor or use a synthetically prepared microbial collection, similar to that designed by Allen-Vercoe for the gut (71). The latter has the advantage of being devoid of virulence factors and phage, but the potential disadvantage of not coming from the vagina per se, where metabolic interactions had been actively occurring just prior to transplant. Presumably, this real-time functionality is at the core of fecal microbiota transplant success.

It would be particularly interesting to assess whether microbiota transplants could reverse the risk of reproductive tract cancers. One mechanism might be to reduce viral shedding and enhance immune defenses in women with early abnormalities, for example, those associated with HPV infection. Another might be to reduce recurrence of BV and AV, conditions that may be associated with a higher risk of cancer. Further research is required to link dysbiosis with cancer and to assess how lactobacilli influence carcinogenesis of the reproductive tract.

SUMMARY

It does not seem that long ago that our work on beneficial bacteria in the female urogenital tract was met with disdain or disregard. Today, the level of interest in this area is astounding, crossing all parts of life from soil and plants to fish and animals and every part of the human body. A multibillion-dollar industry continues to expand, and consumers the world over, with the exception of Africa, are using probiotics to restore and maintain health.

Likewise, interest in the reproductive tract microbiome has also seen a resurgence, which is long overdue considering the relatively primitive diagnostic and treatment options that have been available to women. Dysbiosis in the tract is frequent and can be debilitating for the woman and influential in the fetus. Our level of understanding of these microbiotas has increased significantly with NGS technology, but it must continue, particularly using functional investigations, if we are to truly improve lifelong reproductive tract health. Therapeutic options will emerge from such research for diagnostics and for treatment and maintenance. The ability to manipulate the microbiome through probiotics, prebiotics, and bacterial metabolites would reduce our reliance on unnatural chemical drugs and empower women if they wish to self-manage their vaginal health. Currently, only the use of urine dipsticks for urinary tract infection and antifungals for vulvovaginal candidiasis are accessible over the counter. This is changing with access to probiotics, but further clinical studies are needed to align which composition is best for which subject. Such personalized care will be feasible, especially since only relatively few CSTs are present, and each woman will presumably fit one of them.

Markers of success will be the extent to which innovations are funded sufficiently to continue through human testing and the extent to which clinicians and regulators embrace this new microbiome paradigm.

FIGURE 1b.

Different mechanisms by which beneficial microbes might influence vaginal health.