The primate vaginal microbiome: Comparative context and implications for human health and disease



The primate body hosts trillions of microbes. Interactions between primate hosts and these microbes profoundly affect primate physiology, reproduction, health, survival, and ultimately, evolution. It is increasingly clear that primate health cannot be understood fully without knowledge of host-microbial interactions. Our goals here are to review what is known about microbiomes of the female reproductive tract and to explore several factors that influence variation within individuals, as well as within and between primate species. Much of our knowledge of microbial variation derives from studies of humans, and from microbes located in nonreproductive regions (e.g., the gut). We review work suggesting that the vaginal microbiota affects female health, fecundity, and pregnancy outcomes, demonstrating the selective potential for these agents. We explore the factors that correlate with microbial variation within species. Initial colonization by microbes depends on the manner of birth; most microbial variation is structured by estrogen levels that change with age (i.e., at puberty and menopause) and through the menstrual cycle. Microbial communities vary by location within the vagina and can depend on the sampling methods used (e.g., swab, lavage, or pap smear). Interindividual differences also exist, and while this variation is not completely understood, evidence points more to differences in estrogen levels, rather than differences in external physical environment. When comparing across species, reproductive-age humans show distinct microbial communities, generally dominated by Lactobacillus, unlike other primates. We develop evolutionary hypotheses to explain the marked differences in microbial communities. While much remains to be done to test these hypotheses, we argue that the ample variation in primate mating and reproductive behavior offers excellent opportunities to evaluate host-microbe coevolution and adaptation. Am J Phys Anthropol 57:119–134, 2013. © 2013 Wiley Periodicals, Inc.

Over the past century, anthropologists have made enormous strides toward understanding the evolutionary relationships between primates and their macroecologies (e.g., Chapman et al., 2000; Campbell et al., 2011; Dixson, 2012; Fleagle, 2013). However, by focusing on the relationships between primates and their physical and social environments, we have inadvertently overlooked a key component of the primate body: the microbiome. The primate body hosts trillions of microbes. Humans, for example, have 10 times as many microbial cells as human cells (Turnbaugh et al., 2007; Yang et al., 2009; Fig. 1). Just as microbes have shaped the evolution of life on earth (Woese, 2002), they have influenced the direction of multicellular evolution (King, 2004; Ley et al., 2008; Rokas, 2008) and may even be responsible for host speciation (Miller et al., 2010).

Figure 1.

Humans harbor more than 10 times as many microbial cells as human cells. These microbes contain many more genes than are found in the human genome.

The majority of microbes inhabiting primate hosts are bacteria (Mitreva, 2012). A smaller percentage consists of Archaea (e.g., methanogens), which are single-cell microbes that resemble bacteria in their lack of cell nucleus or organelles, but with a very different cell wall structure and biochemical composition (see Table 1 for a list of terms and concepts). Eukaryotic microbes (e.g., fungi and amoebas) make up the remainder. These diverse microbial communities reside on all external surfaces of primate bodies as well as most internal tracts and passages that connect to the exterior, such as the skin, mouth, nose, gastrointestinal tract, and vagina (Table 2), as well as regions once thought to be sterile [e.g., brain, amniotic sac (DiGiulio et al., 2008; Branton et al., 2013)].

Table 1. Some important terms and concepts
Metabolic functionsProcesses needed to maintain a living organism through breakdown of organic material to obtain energy, or the synthesis of new components using energy.
MicrobiotaThe microscopic living organisms of a region.
MicrobiomeThe combined genetic material of the microorganisms in a particular environment.
BacteriaMicroscopic, single-celled organisms belonging to Kingdom Monera with non-compartmentalized cells and their DNA (usually circular) can be found throughout the cytoplasm rather than within a membrane-bound nucleus.
ArchaeaSingle-cell microbes that resemble bacteria in their lack of cell nucleus or organelles, but with a very different cell wall structure and biochemical composition. They make up their own Domain, separate from eukaryotes and bacteria.
16S rRNAA component of the 30S small subunit of prokaryotic ribosomes, the sequence of which can be used as a genetic marker, as it exists in all bacteria and has a species-specific sequence that serves as an indicator of evolutionary relationships among taxa.
Clone librariesA collection of DNA fragments that is stored and propagated in a population of micro- organisms through molecular cloning.
MetagenomicsThe study of genetic material recovered directly from environmental samples.
MetabolomicsThe study of the unique chemical fingerprints produced by specific cellular processes.
MetatranscriptomicsThe study of the sets of RNA molecules (including mRNA, rRNA, tRNA, and other non-coding RNA) produced in one or many cells in a group of interacting organisms.
MetaproteomicsThe study of all protein samples recovered directly from an environmental sample.
Radial treeA method of displaying a phylogenetic tree structure whereby the root is located in the left and nodes are displayed in concentric rings around the left. The number of branches serves to identify major groupings of taxa, while the size of the terminal “fans” illustrates the diversity of closely related species.
Table 2. Common microbial populations found in different regions of the human body
RegionTypical bacteria
Respiratory tractStaphylococcus
Gastro-intestinal tractLactobacilli
Urogenital tractCorynebacteria

Microbial communities are niche specific, differing considerably from one body site to another, both in quantity and in species composition (Faust et al., 2012; Zhou et al., 2013). For example, mouth microbiota are abundant and diverse. In contrast, stomach microbiota are typically sparse as a result of the acidic environment. In the human small intestine, microbial constituents are relatively few, and are predominantly found adhering to the intestinal lining, where they process the rapidly moving fluids. In contrast, in the slower moving, less acidic lumen of the human colon, microbes are plentiful. Even within a particular anatomical region, microniches exist, harboring distinct microbial communities (Kim et al., 2009). On the other hand, microbial communities inhabiting the same body site show similarities across individuals, indicating that microbial populations are highly adapted to the different physical and chemical properties of these sites (Zhou et al., 2013), and that site properties are more important than other interindividual factors (e.g., age, sex; Costello et al., 2009).

The sheer scale of host-microbe interactions points to major microbial influences on primate evolution. Interactions between hosts and microbes profoundly affect host physiology, reproduction, health, survival, and ultimately, evolution (Nunn and Altizer, 2006; Leitich, 2005; Steer, 2005; Heath and Schuchat, 2007; Leitich and Kiss, 2007). For example, gastrointestinal microbes can alter liver function by modulating gene-expression (Björkholm et al., 2009) and early microbial colonization can affect brain development (Heijtz et al., 2011), modulate serotonin levels (Desbonnet et al., 2008), and affect host behavior (Maes et al., 2008; Bercik et al., 2011; Forsyth and Kunze, 2013).

Microbes affect hosts through relationships that range from highly beneficial and codependent to severely pathological (Table 3). Most colonizing microbes are either commensal or mutualistic with their hosts and are a vital part of host health and development (Xu and Gordon, 2003; Nyholm and McFall-Ngai, 2004; Verhelst et al., 2004; Verstraelen et al., 2004; Zhou et al., 2004; Hyman et al., 2005). In mutualistic relationships, the host benefits the resident microbial communities by providing nutrients and optimal environmental conditions to support bacterial growth. Microbes reciprocate by providing numerous benefits to their host. For example, some microbes break down indigestible plant polymers, thus increasing host ability to extract energy from their diet (Gill et al., 2005; Turnbaugh et al., 2006; Tasse et al., 2010), while other microbial constituents provide essential vitamins and nutrients for the host (Moran et al., 2008). Indigenous microbial communities play a protective role by increasing resistance and thus preventing colonization of pathogenic microbes through direct competition (Brownlie and Johnson, 2009). Other benefits include metabolic functions (e.g., disease resistance) that hosts could not have evolved independently (cf. Bäckhed et al., 2005, 2007). Still others play a key role in development of adaptive immunity against respiratory infections (Ichinohe et al., 2011) or mediate postnatal maturation of immune function (Dobber et al., 1992; McFall-Ngai, 2007; Mazmanian et al., 2005, 2008; Gaboriau-Routhiau et al., 2009). In contrast, pathogenic microbes (e.g., Wolbachia and Spiroplasma) negatively manipulate host physiology (Moran et al., 2008) and have caused selective sweeps of host populations [e.g., plague, smallpox, and cholera (Wolf et al., 2001)]. Thus, relationships between hosts and their microbiomes are of critical importance for host survival and fitness (Dethlefsen et al., 2006). However, to date, primate host-microbial relationships have largely been overlooked.

Table 3. Types of host-microbe relationships
MutualisticBoth the host and the microbe benefit as a consequence of the interaction
CommensalInduces no significant damage after primary infection
PathogenicCapable of causing disease to the host

The goal of this review is to introduce microbes and the role of microbes in a comparative context, with a focus on what is currently known about microbiomes of the primate vaginal tract (in terms of the number and relative abundance of microbial species present). We focus on vaginal tract microbes because they have a clear evolutionary effect. More specifically, mating drives the evolution of behaviors and morphologies (Darwin, 1871). Because reproductive fitness depends on the immunological acceptance of certain foreign cells (i.e., sperm), females are particularly susceptible to microbes, including pathogens via their reproductive tracts. The close interaction between sex, reproduction, and evolution presupposes that selective pressures on vaginal ecology must be considerable and that coevolutionary relationships between females and microbiota are expected (Sobel et al., 1999; Stumpf et al., 2011). Evidence for the influences of reproductive tract microbiota on health, fecundity and pregancy outcomes is widespread (Hill, 1993; Koumans and Kendrick, 2001; Johansson and Lycke, 2003; Leitich, 2005; Schwebke and Desmond, 2005; Schwebke, 2005; Steer, 2005; Heath and Schuchat, 2007; Leitich and Kiss, 2007; White et al., 2011). For example, microbes play causal roles in disease processes [e.g., sexually transmitted diseases (STDs)], and select against hosts or alternatively, play a protective role in preventing colonization of microbes responsible for STDs (Wren, 2000; Gupta and Maiden, 2001). Such functions may be critical to variation in mating systems, either by increasing or decreasing the importance of microbes in conferring STD resistance. These relations may hold great importance in the evolution of primate sexuality, given the fundamental role of the reproductive tract in fitness.

We are now at the initial stages of understanding microbial variation within primates, with research seeking to address a broad set of questions, ranging from what drives that variation to whether or not that variation is adaptive and what the implications of that variation are for the health and fitness of any individual host or host species. Primates are the ideal group to investigate questions of microbial diversity as this Order is particularly variable in their reproductive systems, anatomy, morphology and behavior (Dixson, 2012). Though behavioral and anatomical aspects of primate reproduction are relatively well studied (Dixson, 2012; Fleagle, 2013), we still lack a clear understanding of the microbial dimensions of these systems and how the corresponding microbiomes have coevolved with differences in primate sexual and reproductive behavior. This diversity enables comparative tests for microbial adaptation, coadaptation, and coevolution with their hosts, as well as valuable insights into human microbiomes.

New Methods Available

Although the importance of host-microbe interactions in health and disease was recognized, technological limitations hindered scientific understanding of these relationships until recently. Previously, the predominant method of analysis, growth of cell cultures on laboratory media, was problematic because the vast majority of microbial species (>99%) were resistant to cultivation by standard laboratory techniques (Bakken, 1985). However, these culture-based methods have been superseded by culture-independent genomic technologies that now permit accurate and less biased identification of microbial taxa from diverse samples (Venter et al., 2004; Tringe et al., 2005; Fig. 2). The first of the new technologies involved use of polymerase chain reaction (PCR) to amplify 16S rRNA genes that differ markedly among bacterial or archaeal species (Table 1), followed by direct sequencing and analysis of 16S rDNA clone libraries from isolated vaginal samples to determine the microbial phylogenies that are present.

Figure 2.

Flowchart of culture-independent genomic analyses of microbes. DNA is extracted and PCR is used to amplify 16S rRNA genes that code for different bacteria or archaeal species, which are then sequenced.

More recently, state-of the art sequence technologies (next generation and recently third generation) have transformed studies of microbial ecology by enabling direct metagenomic analysis of samples without the need to generate DNA clone libraries (Mardis, 2008). Employing these technologies, genomic DNA from the bacterial population is isolated and cut into small fragments, which are directly sequenced. By sequencing portions of all genes in the bacterial population, metagenomic analyses looks at the host as a metaorganism of host and microbial chromosomes and provides an overview of the microbial functional gene content. This type of analysis can provide insight about the metabolic potential (i.e., the range of possible biochemical processes encoded by the genes present) of the microbial population that can occur within a particular host niche.

Microarrays are another recently developed technique used for understanding microbial levels of gene expression (Miller and Tang, 2009). Using microarrays, DNA segments representing thousands of genes are placed on a slide-sized surface (a chip). Fluorescently labeled DNA copies of RNA segments (cDNA) from the sample population bind to the DNA segments on the microarray chip and binding events are monitored to detect which genes were expressed. Newer next-generation gene-expression approaches such as RNAseq, also called whole transcriptome shotgun sequencing, use high-throughput sequencing to directly sequence the cDNA and gain information about the RNA content of the samples (Marguerat et al., 2008).

These new, cost-effective, high-throughput sequencing-based approaches have enabled researchers to analyze, define, and compare both the composition and dynamics of human and nonhuman primate (NHP) microbial communities (Verhelst et al., 2004; Verstraelen et al., 2004; Zhou et al., 2004; Fredricks et al., 2005; Hyman et al., 2005; Kim et al., 2009; Thies et al., 2007; White et al., 2011). Thus, these new technologies add strong analytical and conceptual tools to studies of primate physiology, evolution and behavior, opening the door to novel questions about how microbes covary with primate health as well as phylogenetic and behavioral diversity.

Factors Affecting Vaginal Microbial Ecologies

The vaginal microbiome shows remarkable inter and intraindividual variability, which is influenced by many endogenous (genetic, hormonal), as well as exogenous (e.g., behavioral) factors. While most of what is known about microbes comes from studies of human hosts, the factors affecting the vaginal microbial community composition are incompletely understood. Here, we review current evidence of some of the known and presumed factors affecting vaginal microbial community composition in humans and other primate species and highlight areas for further research.

Intraindividual variation

Factors such as age, life history phase, and stage in the menstrual cycle affect vaginal environments, including the microbiota and vaginal pH (Weinstein et al., 1936). In neonates, the vaginal environment is influenced by high maternal estrogens that keep the vaginal mucosa rich in glycogen. Within the first day after birth, the neonatal vagina is colonized by Lactobacillus spp., as well as Corynebacteria, Staphylococci, Streptococci, and E. coli species (Jarvis, 1996; Dominguez-Bello et al., 2010). Lactobacillus is a mutualistic microbe (Martin et al., 2007; Witkin et al., 2007) that metabolizes glycogen into lactic acid (Wylie and Henderson, 1969; Singh et al., 1972; Boskey et al., 1999, 2001), lowering vaginal pH. However, microbial colonization varies in part on the type of birth (i.e., vaginal or Ceasarean-section delivery). Vaginally delivered babies are predominantly colonized by maternal vaginal microbes, whereas babies born by C-section are initially colonized by microbes typical of adult skin (Dominguez-Bello et al., 2010). As the maternal estrogen effects diminish, Lactobacillus abundance declines and vaginal pH rises to neutral or alkaline (Hammerschlag et al., 1978; Thoma et al., 2011). The juvenile phase is characterized by low abundance of Lactobacillus. Similarly, during premenarche, there are few lactobacilli or Gardnerella vaginalis and Mobiluncus spp., and pH levels average around 7.0 (Hammerschlag et al., 1978; Thoma et al., 2011). At puberty, the vaginal environment shifts in response to rising circulating estrogen levels (Thoma et al., 2011), thickening the vaginal epithelium and increasing levels of glycogen in the vaginal walls (Fienberg and Cohen, 1968; Gregoire et al., 1971; Paavonen, 1982). The increasing glycogen near sexual maturity enables proliferation of Lactobacillus and consequent lowering of vaginal pH (3.5–4.4) typical for women of reproductive age (Paavonen, 1982; Boskey et al., 1999, 2001; Danielsson et al., 2011; but see Linhares et al., 2011).

The acidic vaginal environment facilitated by Lactobacillus spp. is believed to play a protective role in limiting growth of pathogenic bacteria, preventing bacterial vaginosis ((BV), a common vaginal disorder in humans, characterized by disruption of the equilibrium of the normal vaginal microbiota, decreases in Lactobacillus content, increases in other bacterial species such as Gardnerella vaginalis, and increased vaginal pH), yeast infections, STDs such as HIV, and Neisseria gonorrhoeae infections (Sewankambo et al., 1997; Gupta et al., 1998; Taha et al., 1998; Martin et al., 1999; Pybus and Onderdonk, 1999; Sobel, 1999; Donders et al., 2000; Cherpes et al., 2003; Watts et al., 2005; Oakley et al., 2008; Mirmonsef et al., 2012). Lactobaccilli competitively exclude pathogens such as Prevotella, Atopobium, Gardnerella, Sneathia, Megasphaera, and Mobiluncus, from binding sites on vaginal epithelial cells, and also produce antimicrobial compounds and can enhance the host immune response to pathogens (Klebanoff et al., 1991; Boskey et al., 2001; Kaewsrichan et al., 2006; Voravuthikunchai et al., 2006; Ma et al., 2012). This protective environment is very advantageous during the reproductive period, providing a clear benefit to host fitness and making this relationship a good candidate for host-microbe coevolution.

After menopause, the abundance of lactobacilli sharply declines (Burton and Reid, 2002; Cauci et al., 2002; Pabich et al., 2003; Heinemann and Reid, 2005; Galhardo et al., 2006). The postmenopause drop in lactobacilli prevalence and abundance likely reflects the concurrent drop in circulating estrogens and glycogen production. Reduction in lactobacilli after the drop in estrogen during menopause is accompanied by an increase in vaginal pH to 6.5–7.0 and a potential decrease in the defenses against pathogenic bacteria (Marrazzo et al., 2010; Srinivasan et al., 2010). These microbiome changes are associated with increased risks for contracting STDs, and endocervical and pelvic inflammatory disease (Marrazzo et al., 2010; White et al., 2011). Such age-related changes in vaginal environments suggest that maintaining a glycogen-rich ecosystem dominated by lactobacilli is costly, but that the benefits (e.g., greater fitness) outweigh these costs.

Menstrual cycle phase

Vaginal microbial ecologies are also highly affected by the menstrual (estrous) cycle phase (Onderdonk et al., 1986; Keane et al., 1997; Narushima et al., 1997; Gajer et al., 2012). In humans, high midcycle estrogen levels accompanied by increased glycogen production, lead to Lactobacillus proliferation (Wylie and Henderson, 1969; Singh et al., 1972; Boskey et al., 1999, 2001). Increased mucosal secretions also potentially enhance the growth of other microbial taxa, including Candida (Schwebke and Weiss, 2001). However, high levels of estrogen or estrogen and progesterone at midcycle are associated with higher stability of microbial communities (Gajer et al., 2012), and epithelial cell receptivity to lactobacilli adhesion increases at peak midcycle estrogen levels (Chan et al., 1984), emphasizing the protective role for this taxon during this important reproductive phase.

Menses are accompanied by considerable changes in microbial diversity. DNA sequence analyses of baboon samples revealed lower prevalence, intensity and diversity of microbes during menstruation compared to other phases of the estrous cycle (Fig. 3). Moreover, a longitudinal study of 32 healthy reproductive-age women sampled twice weekly over a 16-week period showed that menstruation was responsible for the greatest alteration in vaginal bacterial community composition (Gajer et al., 2012; Fig. 4). These findings have potential implications for the debate concerning the evolution of menstruation. Specifically, Profet (1993) argued that menstruation evolved to rid the vaginal tract of pathogens. This idea has been largely discounted in favor of the notion that sloughing endometrial tissue may require less energy than constant uterine readiness, and evidence for elevated pathogens in the uterus before menstruation is lacking (Strassmann, 1996). However, until recently, testing such hypotheses has been hampered by technical limitations. This research (Rivera et al., unpublished data; Gajer et al., 2012) points to major shifts in microbial community composition with menses, including the reduction of potentially harmful pathogens, suggesting of Profet's (1993) hypothesis merits some reconsideration.

Figure 3.

A menstruating baboon (right) shows extremely low microbial diversity compared to a nonmenstruating baboon (left).

Figure 4.

Interpolated bar plot shows the temporal (and interindividual) dynamics of vaginal bacterial communities in three women sampled twice weekly over 16 weeks. Colors show relative abundance of phylotypes. Key for each phylotype is shown at the bottom. Adapted from Gajer et al. (2012).


Recent studies indicate that microbial composition (content and relative abundance) also differs by location within the vagina (Kim et al., 2009; Yeoman et al. 2012). 16S rRNA gene clone libraries of samples obtained from different locations (cervix, fornix, outer vaginal canal) and by different methods (swab, scrape, lavage) from the human vaginal tracts revealed that, although there are considerable co-occurrences (i.e., lactobacilli are found at each site), the vaginal microbiota are not homogenous (Kim et al., 2009; Yeoman et al., 2013; Faust et al., 2012). Instead, microbial composition differs significantly within an individual with regard to anatomical site and sampling method used (Kim et al., 2009). The proximal third of the vaginal tract (e.g., fornix) exhibited more microbial diversity than other regions (Yeoman et al., 2013). The bacterial diversity detected by scrape samples exceeded that detected using swab and lavage samples from multiple locations within an individual (Kim et al., 2009; Yeoman et al., 2013). These studies are biomedically relevant, highlighting microniches and the importance of the sampling site and sampling method in determining the composition of vaginal microbial communities as comprehensive sampling is more predictive of vaginal health (Kim et al., 2009).

Interindividual Variation

Along with intraindividual variation, microbial ecologies also are influenced by interindividual variation. Significant person-to-person variation in the vaginal microbial populations occurs among humans (Zhou et al., 2004, 2010b; Kim et al., 2009; Ravel et al., 2011). For example, lactobacilli are generally predominant in human vaginal ecosystems worldwide (Pavlova et al., 2002; Anukam et al., 2006a; Stoyancheva et al., 2006). However, healthy women of reproductive age have markedly different bacterial community compositions (Verhelst et al., 2004; Verstraelen et al., 2004; Zhou et al., 2004; Fredricks et al., 2005; Hyman et al., 2005; Kim et al., 2009). These communities either are dominated by one of four common Lactobacillus species (L. crispatus, L. iners, L. gasseri, and L. jensenii) or, for the 20–40% of women who lack a Lactobacillus prevalence, a diverse anaerobe-dominant microbial community predominates, composed of varying amounts of Atopobium, Corynebacterium, Anaerococcus, Peptoniphilus, Prevotella, Gardnerella, Sneathia, Eggerthella, Mobiluncus, or Finegoldia (Ravel et al., 2011; Yeoman et al., 2013).

Among humans, vaginal microbial communities differ by self-reported race (Zhou et al., 2010b; Ravel et al., 2011). Studies find higher prevalence of vaginal communities that are not dominated by Lactobacillus spp. in Hispanic and African-American women and higher median pH levels [Hispanic (pH 5.0 ± 0.59); black (pH 4.7 ± 1.04)] compared with Asian (pH 4.4 ± 0.59) and white (pH 4.2 ± 0.3) women (Ravel et al., 2011). Such differences in Lactobacillus might correlate to differences in estrogen levels (cf. Pinheiro et al., 2005), but other host factors play a role (Ma et al., 2012).

Interindividual variation in vaginal microbial ecology is also evident among baboons (Rivera et al., 2010a). Using denaturing gel electrophoresis, a method for analyzing PCR-amplified regions of 16S rRNA genes and screening large numbers of samples, the study found that even baboons living together in a highly controlled captive environment show considerable interindividual vaginal microbial variation (Rivera et al., 2010a). This suggests that individuals have different genetic, physiological, or behavioral factors that influence the resident vaginal microbiota and that the vaginal tract accommodates a considerable range of microbial populations.

External Environment

One factor that does not readily explain variation in microbial diversity is external environment. While it is reasonable to assume that physical contact between the host and its associated soil or other environmental microbes might influence vaginal microbial diversity, several lines of evidence point to individual or host species-specific physiology as a more dominant factor. Documenting microbial diversity in (environmental) macrohabitats is notoriously difficult (Hughes et al., 2001), and microbiologists have yet to demonstrate microbial richness differences among terrestrial macroenvironments. Also, while soils should be expected to harbor the majority of environmental microbes given the major food source of organic detritus, worldwide soil surveys show no major differences among large-scale habitats (rainforest, prairie, and desert) for microbial species richness (Fierer and Jackson, 2006; Fierer et al., 2007). Soil pH affects microbial diversity and abundance, with reduced microbial diversity in low pH rainforest soils possibly offsetting effects of abundant forest floor detritus, high year-round temperatures, and humidity (Feirer and Jackson, 2006). Similar levels of microbial diversity characterize primate habitats (tropical forests and grasslands), with low diversity compared to more temperate areas with less acidic soils (Fierer and Jackson, 2006).

Microbial communities found on the primate body are quite different in composition from the microbial populations found in soil and water. Similarly, the diversity observed for vaginal microbiomes of captive baboons, which demonstrated differences in microbial communities in spite of a shared and controlled environment (Rivera et al., 2010a), suggests that it is internal, host-specific differences rather than external environmental factors that are the major determinants of vaginal microbial diversity (see also Stumpf et al., 2010; Yildirim et al., 2011). Comparable results have been obtained for primate gut microbes where Ochman et al. (2010) found that the relationships among microbial communities were congruous with primate host phylogeny.

This is not to suggest that inoculation of microbes from external sources does not occur. Rather, both pathogenic and mutualistic microbes can be transferred from close contact with other species in a shared environment (Woolhouse et al., 2001; Davies and Pedersen, 2008; Nunn et al., 2011; Degnan et al., 2012; Moeller et al., 2012; Faith et al., 2013; Goldberg et al., 2008). Interestingly, physical proximity rather than evolutionary relationships of the host species can matter more particularly for disease transmission: despite the more distant evolutionary relationships, domesticated animals pose greater risk of transmitting pathogenic microbes to humans than NHPs (Pedersen et al., 2007; Goldberg et al., 2008). Recent evidence indicates that sympatric African ape hosts share 53% more bacterial phylotypes than allopatric hosts (Moeller et al., 2012). Generally, however, microbial phylogenies are distinguishable by host species regardless of environmental similarities (cf. Ochman et al., 2010; Yildirim et al., 2010, 2011; Moeller et al., 2012).

Microbial Variation among NHPs

Most microbe studies to date have focused on human host-microbe relationships. However, recent culture-independent studies of primate vaginal microbiomes indicate that there is considerable host species-specific microbial diversity (Rivera et al., 2010a, 2010b; Stumpf et al., 2010; Yildirim et al., 2011; Spear et al., 2010; Fig. 5). Understanding both variation in primate vaginal microbial ecologies and the factors (e.g., socioecological, sexual, morphological, or genetic) that influence microbial variation is crucial to understanding host-microbe relationships.

Figure 5.

Pie charts indicate main bacterial phyla in vaginal samples of humans (A) (data from Hyman et al., 2005), captive baboons (B), and wild baboons (C). Wild and captive baboon microbial taxa more closely resemble one another than humans, with some more fine grained differences in the most numerically abundant group, Firmicutes (Figure from Rivera et al., 2010b).

Numerous factors potentially influencing vaginal microbial community compositions include host phylogeny, body size, mating system, and substrate use. For example, variation in mating systems may depend in part on how primates, including humans, conduct routine associations with vast numbers of microbes (Turnbaugh et al., 2007). Promiscuity is thought to be a driving factor leading to systemic differences in the primate immune system (Nunn et al., 2000) and the rate of molecular evolution of genes impacting immune function (Wlasiuk and Nachman, 2010). This raises expectations that promiscuity may also influence the composition of the vaginal microbiome both directly and indirectly, and which in turn might contribute to the evolution of mating behavior (Immerman, 1986; Kokko et al., 2002; Loehle, 1995; Sharon et al., 2010; Thrall et al., 1997, 2000). For example, primate sexual swellings are associated with multimale mating systems and promiscuity (Stumpf et al., 2011). Sexual swellings signal female receptiveness and are obvious attractants to males. Significantly, they occur when the vaginal tract is most hospitable to foreign cells. Perineal swellings increase the length and volume of the vagina, facilitate mucosal contact and transmission of microbes among many mating partners, and their proximity to the anus may increase exposure to gastrointestinal microbes. These factors may influence patterns of microbiota in species with pronounced sexual swellings such as chimpanzees and baboons. Paralleling established correlates of multimale mating, such as testes size and number of mates (Short, 1979; Harcourt et al., 1981; Kappeler and van Schaik, 1999; Nunn et al., 2000, 2003; Nunn et al., 2004), more promiscuous host species may show greater microbial diversity than less promiscuous species, consistent with recent analyses of mice (MacManes, 2011).

Along with anatomical, behavioral, and physiological characteristics, multimale mating probably evolved in tandem with mechanisms to reduce infections and STDs (Harvey et al., 1991; Kappeler and van Schaik, 1999). STDs significantly affect host fitness (Lockhart et al., 1996; Nunn et al., 2004), and STDs are known to “co-opt” hosts (Nunn and Altizer, 2006). Species with more promiscuous mating systems may be especially vulnerable to STDs (Brown and Brown, 1986; Moller et al., 1993), with behaviors affecting the distribution and spread of STDs (e.g., May and Anderson, 1979; Anderson and May, 1979, 1992; Nunn et al., 2000; Thrall et al., 2000; Arneberg, 2002; Roberts et al., 2002). Consequently, greater degrees of sociality and promiscuity among host species are expected to influence the prevalence and diversity of microbes, and lead to increased selection for evolved immune defenses against infection (Freeland, 1976; Loehle, 1995). This example offers a testable hypothesis for the ways by which hosts and microbes could coevolve in polygynandrous species.

Other mating systems may have different factors that influence selection on hosts and microbial communities. For example, vaginal environments clearly influence sperm survival, and variation in vaginal microbial communities may play a role in cryptic female choice. Recent research indicates that microbes influence mating preferences in fruit flies (Sharon et al., 2010). Focusing on the interactions of hosts and microbes may help to guide future research into the evolution of mating systems and behavior.

Human Uniqueness

Recent, culture-independent studies of NHPs highlight the distinctiveness of the human vaginal microbiome (e.g., Rivera et al., 2010a, 2010b; Yildirim et al., 2011). In particular, the human vaginal microbiome is dominated by Lactobacillus spp. and is characterized by low pH (≤4.5), and low Shannon's Diversity Index (SDI) (0.81 ± 0.54). In contrast, Lactobacillus is not prevalent among NHPs (Rivera et al., 2010a, 2010b; Spear et al., 2010; Yildirim et al., 2011), constituting 5% or less of the NHP vaginal microbial communities (Spear et al., 2010). The paucity of Lactobacillus correlates with significantly lower glycogen and lactic acid levels in NHPs compared to women with healthy vaginal microbiota (Mirmonsef et al., 2012; Table 4). Data from vervets, baboons, and mangabeys (Street et al., 1983; Johnson et al., 1984) as well as rabbits, mice and rats (Linhares et al., 2011) also indicate that the vaginal environment of these species is characterized by a higher, near neutral pH. Consequently, the low Lactobacillus loads, low lactic acid and high pH characteristic of NHPs resemble human vaginal dysbiosis (microbial imbalances), BV, or postmenopausal changes (e.g., Sneathia, Anaerococcus, Prevotella, Gardnerella, Gemella, Facklamia, Peptoniphilus) (Spear et al., 2010; Hummelen et al., 2011; Yildirim et al., 2011).

Table 4. Interspecies difference in median levels of glycogen and lactate (Mirmonsef et al., 2012)
 μg glycogen/μg proteinmol lactate/μg protein
Rhesus macaque0.0041.2
Pigtailed macaque00.7
Healthy human0.24.2

Several hypotheses could explain the marked differences in NHP and human vaginal microbiomes. First, we propose a “reproductive phase hypothesis”. Because we know that the Lactobacillus abundance in humans is closely linked to estrogen levels in reproductive-aged women (Ravel et al., 2011), the most likely explanation for the interspecies differences in vaginal microbiomes lies in differences in reproductive cycling. For example, estrus in most NHPs is relatively brief (Dixson, 2012). At an extreme, the mating season of several lemur species is only 2–4 days over an entire year. Throughout the majority of the year, the vagina is sealed and thus is isolated from the environment (Ankel-Simons, 2007). Most anthropoid taxa, while not as extreme as strepsirrhines in this regard, have more limited periods of sexually activity than do humans (Dixson, 2012). Consequently, NHP vaginal microbial profiles may more closely resemble the microbial patterns of nonreproductive humans (juvenile or postmenopausal), with lower Lactobacillus, lower glycogen, and neutral pH levels, yet may be more similar to humans during their relatively brief periovulatory periods. At such times, we predict that greater Lactobacillus abundance would confer protection from pathogens and microbial disruptions when such protection is most important. Although no significant differences in glycogen or lactate levels have been observed in longitudinally collected samples from cycling pigtail macaques (Mirmonsef et al., 2012; see also Spear et al., 2010), NHP species have not been well sampled across their ovulatory cycle to evaluate this hypothesis and more data are clearly necessary.

Alternatively, a second explanation, the “disease risk” hypothesis, proposes that the distinctiveness of the human vaginal microbiome corresponds to humans' unique sexuality. Specifically, continuous sexual receptivity throughout the cycle and even throughout pregnancy, prolonged intromission, and the larger vaginal size in humans compared to NHPs all increase microbial exposure. This hypothesis posits that these factors would expose women to greater risk of STDs, other pathogens, and microbial disruptions, subjecting humans to selection for commensal microbial communities that minimize risks of infection, whereas NHPs may not have been subject to this selection to such an extent. Thus, the considerable selective pressure resulting from disruption to the vaginal ecosystem may explain the uniqueness of the human vaginal microbiome. The protective presence of Lactobacillus in humans would confer advantages necessary for the more continuous sexual receptivity that is characteristic of our species, but which is not common in other NHPs.

This model presumes that the vaginal microbiomes have been selected by STDs. This disease-risk hypothesis is supported by the fact that STDs are relatively common among humans and that they are well known to affect mortality and morbidity across the lifespan. For example, gonorrhea, when present in the mother during parturition, can be transmitted to the newborn and cause blindness in newborns (Schulz et al., 1987; Whitcher et al., 2001). Similarly, untreated syphilitic infections have caused substantial mortality (Schulz et al., 1987; McDermott et al., 1993). HIV is another relatively recent STD causing high mortality in humans (e.g., Rosenberg, 1995). Lower Lactobacillus dominance, and consequently lower lactic acid concentrations and higher pH levels, are associated with increased susceptibility to HIV infection (Sewankambo et al., 1997; Gupta et al., 1998; Taha et al., 1998; Sobel, 1999; Mirmonsef et al., 2012). Few NHP species have been tested for STDs, but it is interesting to note that chimpanzees, in spite of their considerable promiscuity, show little evidence of microbial STDs (Rushmore et al., 2013; Rushmore, 2013).

Third the “obstetric protection hypothesis” suggests that, humans are also unique among primates in that gestation is longer than most other species and women face enormous constraints during birth. Unlike other primates, the dimensions of the pelvic outlet are smaller than the infant head. Delivery of the infant requires passage through this restricted canal and birth constraints greatly elevate risks of infection to both the fetus and mother (Rosenberg and Trevathan, 2002). Pregnant women are particularly vulnerable to microbial community disruptions (McGregor and French, 2000). Numerous studies of pregnant women suggest that Lactobacillus plays an important protective role: depletion of lactobacilli, shifts to a polymicrobial system, with increased anaerobes, facultative anaerobes (Donders, 2007), and overgrowth of BV-associated microbiota are linked to chorioamnionitis, premature membrane rupture, and preterm labor (Gantert et al., 2010). Disruption of the mutualism between human hosts and resident vaginal lactobacilli tremendously elevates the risk of preterm birth over sevenfold with severe fitness consequences (Hillier et al., 1995; McGregor and French, 2000; Lamont and Sawant, 2005; Steer, 2005; Verstraelen and Senok, 2005; Heath and Schuchat, 2007; Leitich and Kiss, 2007; Pettersson, 2007; Thies et al., 2007; Schwebke, 2009). The selective potential of the microbial environment is clearly quite high.

The “common function hypothesis” is a fourth potential explanation for the unique human microbiome. This hypothesis supposes that the protective role of Lactobacillus in humans may be fulfilled by other microbial species in NHPs. Bacteria are characterized by tremendous variability. Even phylogenetically similar bacterial species may exhibit different ecological and virulence properties (Jaspers and Overmann, 2004). New direct DNA analyses show functional diversity in Lactobacillus and many other taxa (Verhelst et al., 2004; Verstraelen et al., 2004; Zhou et al., 2004; Hyman et al., 2005; Kim et al., 2009; Yeoman et al., 2010, 2013). In addition, bacterial communities composed of different species may still occupy similar ecological niches and exhibit similar metabolic functions (Langenheder et al., 2005). Thus, it is possible that different microbial communities may accomplish protective roles without elevated lactate levels or lower pH. For example, secreted proteins (e.g., S-layer bacteriocins) and perhaps small antimicrobial peptides might provide this protective role (cf. Kalyoussef et al., 2012). The vaginal microbiomes of NHPs are characterized by a number of novel, unclassifiable taxa, particularly in those primate species more evolutionarily distant from humans (Yildirim et al., 2011; Rivera et al., 2010b; Fig. 6). Support for this idea can be found in the fact that microbial species such as Atopobium, Streptococcus, Staphylococcus, Megasphaera, and Leptotrichia, are capable of homolactic or heterolactic acid fermentations (Jovita et al., 1999; Zhou et al., 2004; Ma et al., 2012). Thus, the highly diversified, NHP microbial community may function to resist microbial disruptions and maintain a healthy vaginal environment despite low Lactobacillus content and a higher pH.

Figure 6.

Differences between human and baboon microbiota are highlighted in this radial tree with human sequences (from Hyman et al., 2005) shown in blue and baboon shown in red. Even within the same phylogenetic group of bacteria, the baboon sequences clustered separately from the human strains (from Rivera et al., 2010b).


Microbes are an important component of our internal and external environments, affecting many aspects of primate health and fitness. The application of culture-independent, direct next-generation high-throughput sequencing technologies have enabled researchers to generate both 16S rRNA gene-based and metagenomic DNA-based microbial phylogenies that identify the microbes in primate vaginal (and other) ecosystems. These technologies allow researchers to define, analyze, and compare the composition and dynamics of these microbial communities and to better understand primate host-microbe relationships and their influence on health, disease and evolution.

We are just beginning to discern and test for some of the causal factors that underlie patterns of variation within and among vaginal microbiomes. Individual differences in genetics and physiology seem to be important factors in discriminating among microbiomes both intra- and interindividually. Whereas the external physical environment seems to play a minor role, contact with other hosts' microbiomes seems important. Sexual partners (male and female as well as female-female) share related microbial compositions and strains (Jaspers and Overmann, 2004; Marrazzo et al., 2009; Eren et al., 2011; Danielsson et al., 2011). Sexual activity is a strong risk factor for BV (Fethers et al., 2008). BV is rare in women who have never had vaginal intercourse (Fethers et al., 2009) and is most common in women who have new or multiple sexual partners (Eren et al., 2011; Yeoman et al., 2013). The ways in which primates have evolved (and coevolved) strategies to defend against pathogenic invasion and infection, to contend with exposure from conspecifics (e.g., through intercourse), and to respond to the dynamics of vaginal bacterial communities throughout life and during the menstrual cycle offer much fodder for future study. We anticipate that some of the interspecific variation will be structured by differences in mating system and mating behavior, and our group is currently working within this framework.

Studies of the vaginal microbiome offer clear illustrations of how microbes affect primates in ways that range from highly beneficial and codependent to severely pathological in nature. The health and fitness consequences of these interactions are enormous. For example, strong evidence suggests that vaginal colonization with group B Streptococcus during pregnancy and/or delivery can lead to life-threatening infections in neonates (Verani and Schrag, 2010). Given different microbial colonization patterns, the neonatal health implications of Caesarian sections remain to be discerned. Some C-section babies may be more likely to get skin infections from methicillin-resistant Staphylococcus aureus because they lack the protective microbes from their mother's birth canal (Zimmer, 2010).

Vaginal microbes play important roles in other conditions with high fitness consequences, including debilitating health problems such as BV and STDs, adverse pregnancy, and preterm birth. These findings emphasize the need for evolution-based perspectives on the selection pressures affecting host microbiomes and the consequences on host reproduction, health, and disease. The close association with BV and preterm birth and the discovery of microbes in amniotic fluid (DiGiulio et al., 2008; Fig. 7) (which was thought to be sterile) underscores the need for understanding the host-microbe relationships for predicting and preventing STDs, BV, and adverse pregnancy outcomes. In light of the link between sexual activity with novel partners and BV (Eren et al., 2011; Yeoman et al., 2013), educating pregnant women about the risk factors for BV is an important step toward the prevention of preterm birth.

Figure 7.

Bacterial rDNA concentrations in human amniotic fluid during preterm labor correlate negatively with gestational age (adapted from DiGiulio et al., 2008).

Comparative research done to date illustrates the complexity of the vaginal microbiome and the considerable differences that exist among host species. Cercopithecines, particularly macaques and baboons with simian immunodeficiency virus, are used as models to study HIV sexual transmission and human female reproductive disorders (Khan-Dawood, 1998). However, the uniqueness of the human vaginal microbiome has important implications for NHP models used to study STDs. The paucity of the key human vaginal microbe, lactobacilli, the low lactic acid and resultant high pH of NHPs, coupled with the considerable variation in microbial communities across species (Rivera et al., 2010b) raises questions of the appropriateness of the NHP model for humans.

The finding that microbial communities can vary by sampling methods and by microenvironments within the vagina (Kim et al., 2009) may hold key implications for etiology and diagnosis of other diseases, such as ovarian and endometrial cancers. These cancers are notably difficult to detect, rarely caught early, and have a very poor prognosis (Stirling et al., 2005). However, recent findings indicate that ovarian and endometrial cancerous cells can be found in the vagina (Kinde et al., 2013). While pap smears (i.e., brushing the cervix to remove cells) are the standard clinical practice used to detect cervical cancer, findings from microbial studies of the vagina (e.g., Kim et al., 2009) indicate that localized sampling may be inefficient for detecting microbial populations (and presumably rare cancerous cells) dispersed through the vagina. These findings highlight the necessity to use techniques to obtain a broader sampling to provide greater diagnostic capability and improve patient outcomes.

Understanding the function of different microbes across species will also be crucial for understanding the dynamics of host-microbe interactions, health and disease, and facilitating the development of new strategies for disease diagnosis and prevention and for personalized treatments to improve the quality of women's lives. Functional differences cannot be assumed simply by species identification due to species-level or strain-level genomic heterogeneity (Yeoman et al., 2010). Metagenomics can help to identify the functional properties of the microbiomes and the extent of fuctional redundancy across hosts. Metabolomics can be used to characterize the products of ecosystem-level physiological processes and metabolic output (Ma et al., 2012).

Metatranscriptomics and metaproteomics can assess shed light on the functional interaction between the vaginal microbiota and the host. While some studies are moving in this direction for humans (Yeoman et al., 2013), none has been applied to NHPs.

With the exception of one human study (Ravel et al., 2011) and a limited study in macaques (Spear et al., 2010), the temporal dynamics of vaginal communities are currently poorly understood. Most studies focus largely on cross-sectional data with one sample per individual. However, the vaginal environment is dynamic and the microbiome may change (in composition or function) in ways that increase susceptibility to infection. For example, microbiota may vary from preconception to gestation and throughout pregnancy, with important implications for neonatal health and morbidity. Does pregnancy predict hormonal changes associated with anovulation, and a resultant decrease in Lactobacillus? If so, are women more susceptible to diseases such as BV due to the absence of the protective role of Lactobacillus? More longitudinal metagenomic studies are needed, in which individuals are re-sampled and variation in microbial community composition and metabolic funtion is assessed over time, coupled with extensive demographic data on subjects' behaviors, health history, and reproductive status. The overall patterns of change in microbial ecosystems may better predict infection risk than the presence or absence of a particular microbial species. Longitudinal studies would also help to clarify the roles of estrogen and Lactobacillus on vaginal pH and to identify a NHP equivalent. Studies of the temporal dynamics of the vaginal ecosystem will also contribute to our understanding of what constitutes a healthy ecosystem.

The breadth of sampling of indigenous human populations has been quite limited. There have only been a handful of studies of extant (De Filippo et al., 2010; Yatsunenko et al., 2012) and extinct (Tito et al., 2008, 2012; Adler et al., 2013) native peoples. Inclusion of more data from a broader sample of human diversity may facilitate better understanding of vaginal microbial dynamics and aid in improving health (Lewis et al., 2012).

Interdisciplinary study of microbiomes also extends to conservation applications. Considering the strong evidence implicating the critical importance of microbiomes to human health, development, and reproduction (Pedersen et al., 2007; Zhou et al., 2010a; Kau et al., 2011; White et al., 2011), NHP host-microbe interactions are arguably of similar importance, particularly in the context of individual and population fitness. A final point of emphasis: we are in the beginning phases of assessing primate microbial diversity. There clearly remains many fruitful lines of research to pursue, but it is the comparative study of host-microbe relationships that will allow us to better understand the health consequences and evolution of our own microecologies, and those of our primate relatives.


The authors gratefully acknowledge Trudy Turner for inviting this review and for her patience.