The vaginal mycobiome: A contemporary perspective on fungi in women's health and diseases

ABSTRACT

Most of what is known about fungi in the human vagina has come from culture-based studies and phenotypic characterization of single organisms. Though valuable, these approaches have masked the complexity of fungal communities within the vagina. The vaginal mycobiome has become an emerging field of study as genomics tools are increasingly employed and we begin to appreciate the role these fungal communities play in human health and disease. Though vastly outnumbered by its bacterial counterparts, fungi are important constituents of the vaginal ecosystem in many healthy women. Candida albicans, an opportunistic fungal pathogen, colonizes 20% of women without causing any overt symptoms, yet it is one of the leading causes of infectious vaginitis. Understanding its mechanisms of commensalism and patho-genesis are both essential to developing more effective therapies. Describing the interactions between Candida, bacteria (such as Lactobacillus spp.) and other fungi in the vagina is funda-mental to our characterization of the vaginal mycobiome.

Introduction

The community of fungal organisms residing within the lower female reproductive tract is referred to as the vaginal mycobiota. Among those is one of the leading causes of vaginal infection, Candida albicans, a well-studied fungal pathogen. Because vulvovaginal candidiasis (VVC) is the second most commonly reported form of infectious vaginitis, a great deal of effort has been invested in studying the mechanisms of C. albicans pathogenesis. However, the totality of fungal organisms present within the vagina has been grossly underappreciated. While we have extensive knowledge of the types of bacteria present in the vaginal milieu, very little is known about their fungal counterparts. Recent studies that explored the composition of fungi within the vagina have pointed to an exciting new frontier of research. Exploring the types, functional and compositional dynamics of fungal species in the context of the vaginal environment is an important objective, with potential implications for treating and preventing VVC, improving obstetric outcomes, and reproductive health in general. Recent advances in next-generation sequencing tools have enabled the high-throughput identification of fungi and provide burgeoning insights into fungal ecology.

Mycology of the vagina

The sum of the genomes and genes carried by fungal species that exist within a particular environmental or biological niche is termed the “mycobiome.”¹ In humans, the mycobiome is poorly understood compared to the microbiome, which was extensively described by the Human Microbiome Project — the largest and most comprehensive survey of bacterial taxa in healthy adults.^2-5 In the midst of an explosive new era of genomics, rapidly developing sequencing technologies and cutting-edge bioinformatics tools, the mycobiome has developed into its own “sub-specialty” within the field of microbial genomics, but one that lags far behind its bacterial counterpart.⁶

The earliest studies of vaginal microbiology underestimated the complexity of this ecosystem, in part, due to the limitations of the culture-dependent techniques used.^7,8 While bacteria have long been known to dominate the vaginal milieu, leading to a number of studies on the bacterial community, early investigators of vaginal mycology have attempted to draw attention to the importance of fungal members of this community.⁹

Using classical, culture-dependent methods, investigators have measured point prevalence of vaginal fungi in healthy volunteers and diabetics, as well as adolescents and pregnant women.^10-12 Fungi were recovered by culture in 20–60% of the samples. Without exception, the predominant member of the fungal community is identified as C. albicans (often >70%), though the rank prevalence for non-albicans species (including C. krusei, C. parapsilosis, C. tropicalis, C. glabrata, C. guilliermondii, C. pseudotropicalis, C. stellatoidea, and others) varies by population studied, geography, and cultivation methods. More recent epidemiological studies of Candida carriage report even greater representation by C. albicans, making up 85–95% of isolates recovered.^13-15

A study comparing Candida isolates from the United States and the United Kingdom concluded that physiological biotyping could not significantly distinguish the isolates examined.¹⁶ According to the authors, the evidence does not support strain tropism — a selection for “vaginotropic” vs. “vaginopathic” strains — since strains isolated during infection are statistically identical to those collected in asymptomatic colonization.^16,17 Using these same biotyping methods Odds et al^17,18 reported consistency in isolates from different sites (and tissue types) in the same patient and isolates collected months apart. Using different methods, a separate group also reported that no significant differences exist among Candida strains isolated from different body sites; but added that a single host can be colonized by multiple species, and even multiple genotypes of the same species.¹⁹

Using a cultivation-independent 18S rRNA gene clone library, Guo et al²⁰ identified 3 fungal phyla in vaginal samples: Ascomycota (22/28), of which Candida was the predominant genera, Basidiomycota (5/28), and Oomycota (1/28). They reported larger proportions of C. albicans and lower proportions of S. cerevisiae and uncultured (unidentified) fungi in women with allergic rhinitis and recurrent vaginal candidiasis, as compared to healthy controls. Fungal diversity was also increased in these patient populations compared to controls. The authors suggested that fungal dysbiosis might be correlated with the pathophysiology of recurrent vaginal candidiasis, highlighting a role for fungal community in disease.

Though next-generation sequencing technology has profoundly expanded our appreciation of the bacterial microbiota within the human vagina²¹, these sophisticated methods have been applied to the study of the vaginal mycobiota to a far lesser degree. In fact, the very first next-generation sequencing-based survey of fungal communities within the vagina was published in 2013 by a group whose aim was to describe the bacterial and fungal communities of women in Estonia.²² This cross-sectional study sequenced the Internal Transcribed Spacer 1 (ITS1) in 251 vaginal samples from 294 healthy women to characterize fungal taxonomic composition. Fifty-eight percent of sequences belonged to the Ascomycota Division, though Basidiomycota was also represented in a small number of sequences (3%). Within Ascomycota, hits to Saccharomycetales (dominated by the genus Candida), Capnodiales, Eurotiales, Pleosporales, and Helotiales were observed. The resulting mean taxonomic richness for each sample was about 8 fungal Operational Taxonomic Units (OTUs). Candida OTUs could be detected in 70% of samples. Within the group of sequences mapping to Candida, C. albicans was not surprisingly predominant (68%). Strikingly, 161 unique hits at the species-level were obtained from this data set, yet 38% of OTUs were unspecified — no taxonomic assignment lower than kingdom was available. These results speak to a very serious problem in molecular-based fungal taxonomy. The quality and relatively low representation of fungal species in reference databases has direct consequences on quality and accuracy of taxonomic assignments returned. Current databases not only lack the rich volume of sequences that exist for bacterial 16S rRNA gene, but taxonomic synonyms and misclassifications are widespread.^23-25 To protect the integrity of these investigations, careful sequence annotation and database curation is absolutely essential. The results published by Drell et al,²² nonetheless, reveal the underappreciated diversity of fungal members within the vaginal ecosystem and warrant follow-up on these findings. In light of our limited understanding of vaginal mycology, most of what we know derives from studies focused on the key player C. albicans.

Fungal commensalism in the vagina

Witkin and Ledger²⁶ described the characteristics and qualities of the human vaginal environment that make it both unique and complex. In contrast to oft-used animal models, the vagina of reproductive-age women is distinctly acidic (pH of 4.5 or less), owing to the presence of lactic-acid producing bacteria that thrive in this anaerobic niche²⁷ not otherwise described in the Class Mammalia.²⁸ Further intricacies of the vaginal environment in humans include cycles of growth, production of glycogen and its breakdown products by human α-amylases²⁹, and shedding of the epithelium in response to reproductive hormones, and primarily innate (as opposed to adaptive) mucosal immune protection.^30-32 These physiologic features are certainly key determinants of microbial colonization of the vaginal mucosa.^7,33

In 1976, Goplerud et al isolated Lactobacillus spp and C. albicans at consistently increased rates over all 3 trimesters in healthy pregnant women.³⁴ These early observations provided preliminary evidence for a positive correlation between estrogen levels and vaginal colonization of microbes. The data were further substantiated by Larsen and Galask³⁵ who demonstrated obligate estrogenization to achieve yeast colonization in a rat model of vaginal Candidiasis. At present, all laboratory animal models of vaginal candidiasis require estrogen-treatment to establish colonization and subsequent infection.^{36, 37} Some Candida species possess a cytosolic estrogen receptor that could mediate direct transcriptional responses to host hormones.³⁸ Furthermore, estrogen has been shown to disrupt neutrophil chemotaxis to the vaginal epithelium³⁹ and inhibit Th17 cell differentiation,⁴⁰ resulting in heightened host vulnerability to pathogens, such as Candida. Clinical and anecdotal reports frequently link symptomatic Candida vaginitis to the luteal phase, just prior to menses, which is marked by both a high estrogen state and increased vaginal pH.⁷ Researchers have documented enhanced C. albicans adherence to vaginal epithelial cells through estrogen signaling.⁴¹ Further, while Candida is known for its broad pH-range tolerability, adherence to vaginal epithelial cells is significantly enhanced at pH 6 compared to pH 3–4.⁷

The hormone-dependent production and accumulation of glycogen (and its breakdown products²⁹) by human vaginal epithelial cells should not be understated in its contribution to fungal colonization,⁴² however, Candida can utilize other nutrients (including lactate), making Candida highly adaptable to shifts in the nutritional microenvironment in the vagina.⁷ The food source used by Candida in a particular niche is not a trivial detail, as studies have clearly indicated the effects of environmental cues, such as nutrient availability, on cell wall architecture^43-45 which impacts interactions between Candida and immune cells. Recent in vitro work has shown that in the presence of lactic acid (as the sole carbon source) C. albicans is taken up by macrophages less efficiently and can alter immune cell cytokine profiles, specifically by increasing IL-10 and decreasing IL-17 production.^{46, 47} Interestingly, cells grown in a mixed lactate-glucose media behave more like lactate-grown cells. This has particular relevance in the vaginal context because of the abundance of both glycogen (and its breakdown sugar products) and lactic acid, which could effectively promote anti-inflammatory responses to Candida. Other studies have corroborated these findings and suggest that Candida may have evolved to curb immune responses to promote its own persistence and commensalism.⁴⁸ It has also been shown that Lactobacillus indoleamine 2,3-dioxygenase 1 (IDO1) in the gut leads to the production of tryptophan catabolites that act on regulatory T cells, resulting in increased local expression of IL-22⁴⁹ and immunoprotection to VVC.⁵⁰ This suggests that the bacterial microbiome could be mediating tolerance to C. albicans on the mucosa.

Attachment to the mucosal epithelium is mediated by binding to specific host receptors, of which the ALS (agglutinin-like sequence) adhesion family is best studied.^51-53 Hyphal formation is an important attachment factor as well.^54,55 In response to quorum sensing mechanisms, yeast growth is favored by high cell densities (>10 ⁷ cells mL⁻¹) whereas hyphal formation is stimulated by lower cell densities (<10 ⁷ cells mL⁻¹).⁵³ C. albicans adheres to vaginal and oral epithelial cells with greater levels than other species⁵⁶ — an important, but most likely partial, explanation for the predominance of C. albicans-associated infections. Interestingly, Sobel et al⁵⁷ reported marked person-to-person variability in Candida adherence to exfoliated vaginal epithelial cells, but enhanced attachment of C. albicans to epithelial cells from women with recurrent VVC.⁵⁸

Pathogenesis of Candida and vulvovaginal candidiasis

Despite its prominence as the second most common vaginal infection among US women of reproductive-age,^13,59 epidemiologic data on the incidence of VVC remains incomplete — primarily because this is not a reportable infection by public health authority standards.⁶⁰ And while exceptionally common—3 in 4 women will be affected at least once over their lifetime⁶¹—asymptomatic carriage rates for C. albicans in healthy women are estimated at around 20%.^12,62 Mechanisms of immunoprotection are still debated, and the factors that trigger the transition from commensal to pathogenic yeast are still obscure. However, it is generally accepted that predispositions for Candida growth/invasion are niche specific^63,64, though immune defects, breaches in epithelial integrity and microbial dysbiosis are common themes.⁵⁵ While many similarities exist between the mucosal environments of the mouth and vagina, the immunology of Candida vaginitis is undoubtedly distinct.^65,66 Excellent reviews have been written on the immunology of VVC, thus will not be discussed here.^67-69

A statistically significant increase was noted in number of C. albicans colonies cultured from swabs of women with VVC, representing an increase in fungal concentration, as compared to controls (healthy, no VVC), though no difference was measured with non-albicans Candida species.⁷⁰ Peeters et al⁷⁰ noted a positive correlation between the number of C. albicans colonies grown and amount/severity of vaginal discharge, as well as reported pruritus (itching). These findings support the theory of a fungal burden threshold above which inflammatory cells are recruited, resulting in the vaginal symptoms often reported, including itching, irritation, burning, and discharge.⁶⁸

Candida are polymorphic fungi, whose morphogenic transitions are essential mechanisms of pathogenesis in the human host.^71,72 The yeast form (blastoconidia) is typically associated with asymptomatic colonization, transmission or spread (particularly in the bloodstream)^60,73, while the hyphal (mycelial) form contributes mostly to adherence and mucosal invasion, characteristic of symptomatic disease.^57,60,74,75 Peters et al⁷⁶ recently noted that the genes which control C. albicans morphogenesis are required for the immunopathology associated with VVC. A variety of environmental stimuli affect a cell's morphology, including nutrient availability, pH, and temperature.⁷² Using sophisticated quorum sensing mechanisms, C. albicans regulates morphogenesis in response to these external cues.⁷⁷

In vitro proteolytic activity of C. albicans isolates from women with symptomatic VVC was greater than isolates from asymptomatic carriers.⁷⁸ Proteolytic enzymes, namely the secreted aspartyl protease (SAP) family, are well-studied virulence factors employed by C. albicans, in particular, to invade the mucosal layer during VVC and induce immunopathology.^79-83 In 1940, it was proposed that the immunopathology of VVC is caused by a Candida toxin.⁸⁴ Decades later, researchers negated this hypothesis by suggesting Candida cell wall glycoproteins resemble bacterial endotoxins in their structural location, pyrogenicity and immunogenicity, though much less potent.⁸⁵ Earlier this year, however, Moyes et al⁸⁶ identified a secreted cytolytic peptide toxin produced by C. albicans that is essential for mucosal pathogenesis in a murine model of oropharyngeal candidiasis. The C. albicans extent of cell elongation 1 (ECE1) gene, which encodes the toxin they named Candidalysin,⁸⁶ is also among the most highly expressed genes during murine VVC.⁸⁷

Like many pathogenic bacteria, Candida species in general and C. albicans in particular, are efficient at biofilm formation within the human host.^88,89 C. albicans biofilms have been identified on dentures, catheters, as well as mucosal epithelia.⁹⁰ Contact sensing (contact with abiotic or host substrates) has been described as an important trigger for C. albicans biofilm formation.⁹¹ Furthermore, yeast cells are stimulated to form hyphae upon contact with a surface,⁹¹ which in some cases may facilitate active penetration of host tissues^92,93 and in others may lead to mature biofilm formation.⁹⁴ Transcriptional regulation of biofilm formation has been mainly attributed to Bcr1, Tec1, and Efg1,⁹⁰ however, recent studies revealed novel transcription factors associated with this process: Ndt80, Rob1, Brg1.⁹⁵ Harriott et al⁹⁶ were the first to show using in vivo and ex vivo models of murine Candida vaginitis that C. albicans does, indeed, form Bcr1- and Efg1-dependent biofilms. Similar to other body sites, biofilms in the vagina are of major concern, as they have been implicated in immunopathology of VVC, anti-fungal treatment failures and recurrent infections.^97,98 Genomic microvariations in C. albicans, which include rearrangements, loss of heterozygosity, polymorphisms, and copy number variations, have also been associated with fungal persistence in a host and antifungal resistance.^99,100

Candida-Bacteria interactions in the vagina

As early as 1930, Doderlein's bacillus,¹⁰¹ which is now widely known as Lactobacillus, was positively attributed to protection of the vaginal mucosa and a “healthy” microenvironment.¹⁰² Because the lack of Lactobacillus spp. within the vaginal microbiota has been associated with susceptibility to urogenital infections, such as bacterial vaginosis,^103,104 HIV,¹⁰⁵ and urinary tract infection,¹⁰⁶ it stands to reason that Lactobacillus spp may, similarly, protect women from vulvovaginal candidiasis (VVC). Among others, Odds proposed in 1979 a model where symbiosis between some fungi and bacteria is established,^107,108 though these relationships have not been clearly elucidated within the vagina. Early bacterial-yeast co-culture experiments led to the hypothesis that the role of vaginal bacteria is not likely to prevent colonization of Candida but rather prevent their uninhibited proliferation.³⁵ In support of this hypothesis, short chain fatty acids and lactate produced by Lactobacillus spp. and other lactic-acid producing bacteria were shown to inhibit the yeast-to-hyphae switch in C. albicans.¹⁰⁹ Further, investigators have reported lower numbers of Lactobacillus spp in vaginal cultures from women with symptomatic VVC.^70,102 Supporting evidence for this hypothesis includes increased susceptibility to Candida vaginitis following antibiotic therapy, a well-documented risk factor for VVC.^110-112 But not all studies have substantiated the link between VVC and antibiotic usage,¹¹³ and even anecdotal reports are inconsistent. This is consistent with the different types of vaginal microbiota which could be differentially affected by antibiotics.^114,115 Not all women who take antibiotics develop VVC and most women who report VVC have not recently taken antibacterial therapy. Colonization with Candida appears to be a prerequisite risk factor for developing VVC following antibiotic therapy.¹¹⁶ Recently, however, one group has posited that their findings are more consistent with Lactobacillus being associated with greater risk for vulvovaginal candidiasis.¹¹⁷ By elucidating the mechanism by which lactic acid suppresses immune responses to C. albicans, Ene et al⁴⁶ have supplied evidence for this claim.

Candida-bacteria interactions within the vagina likely take place within the context of a polymicrobial biofilm on the epithelial surface.^118,119 An in vitro model of various C. albicans-bacterial biofilms concluded that bacteria negatively impact C. albicans biofilm formation by inhibiting fungal growth and suppressing genes responsible for hyphae formation.¹²⁰ Peleg et al¹²¹ described 5 types of bacterial-fungal interactions and many speculate these also take place within the vagina: physical interactions,^122,123 chemical exchanges,¹²⁴ use of metabolic by-products,^109,118 changes in the environment,¹²⁴ and alteration of the host immune response.¹²⁵ Further studies are required to better understand this important relationship between vaginal bacteria, Candida spp. and other fungi in health and diseases.

Importance of the fungal mycobiome

It is becoming increasingly clear that fungal communities play a more significant role in human health and disease than once assumed.^126,127 Concerted effort, such as that given to surveying bacterial composition and abundance, is required to carry this field into the translational and clinical arenas. Characterization of the human mycobiome has the potential to produce widespread clinical advances in diagnosis, treatment and prevention of fungal infections^23,128 and vulvovaginal candidiasis, in particular, but also potentially bacterial and viral infections. Made possible by next-generation, culture-independent sequencing technologies, new developments in fungal contribution to human health and disease have proven to be very promising. Though far less abundant than bacterial members of the environment, fungi (but primarily Candida albicans) have a pronounced effect on vaginal health, and thus require more in depth studies of the interaction between the mycobiome and the microbiome. While pathogenic mechanisms attributed to single fungal species have consumed much of mycology, it is believed that mycobiome studies will establish correlation between composition and function of the entire fungal community, and cross-kingdom interactions to disease processes. Notably, fungi-fungi interactions have been implicated in the pathogenic process; C. glabrata binds to the hyphae of C. albicans to establish oropharyngeal candidiasis,¹²⁹ thus supporting the need for mycobiome studies that consider the full context in which infection takes place. Culture-based isolation and characterization of pathogens remains of great necessity, however, development of novel in vitro (and even in vivo) models of polymicrobial communities would be ideal to test hypotheses into the role of the vaginal mycobiome in health and disease. And while we invest scientific capital to understand pathogenic mechanisms, we must not neglect to appreciate mechanisms of commensalism, as this will likely lead to preventative strategies that prohibit the commensal-to-pathogen transition.¹³⁰

Figure 1.

Interactions between Candida and microbiota at the mucosal interface have profound effects on the vaginal ecosystem.^109,120-125 Metabolites and small molecules made by the microbiota affect the metabolism and morphology of Candida species. Changes in microbiota relative abundance also impact the abundance of Candida and its ability to access the mucosal surface, where invasion occurs. In healthy states, when microbiota-derived lactic acid is produced, Candida can alter host cytokine production and promote anti-inflammatory signaling. The contribution of bacterial-fungal interactions to the ecology of the vaginal microbiota remains to be described.

Funding

This review was supported by the National Institute for Allergy and Infectious Diseases and the National Institute of Nursing Research of the National Institutes of Health under awards number U19AI084044, R01AI116799 and R01NR015495. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.