Vaginal Dysbiosis from an Evolutionary Perspective

Evolutionary approaches are powerful tools for understanding human disorders. The composition of vaginal microbiome is important for reproductive success and has not yet been characterized in the contexts of social structure and vaginal pathology in non-human primates (NHPs). We investigated vaginal size, vulvovaginal pathology and the presence of the main human subtypes of Lactobacillus spp./ BV-related species in the vaginal microflora of baboons (Papio spp.). We performed morphometric measurements of external and internal genitalia (group I, n = 47), analyzed pathology records of animals from 1999–2015 (group II, n = 64 from a total of 12,776), and evaluated vaginal swabs using polymerase chain reaction (PCR) (group III, n = 14). A total of 68 lesions were identified in 64 baboons. Lactobacillus iners, Gardnerella vaginalis, Atopobium vaginae, Megasphaera I, and Megasphaera II were not detected. L. jensenii, L. crispatus, and L. gasseri were detected in 2/14 (14.2%), 1/14 (7.1%), and 1/14 (7.1%) samples, respectively. BVAB2 was detected in 5/14 (35.7%) samples. The differences in the vaginal milieu between NHP and humans might be the factor associated with human-specific pattern of placental development and should be taken in consideration in NHP models of human pharmacology and microbiology.


Discussion
Host-microbiome interactions are critical for host development. The reproductive evolution of the host is accompanied by microbial evolution and vice versa 16 . Numerous examples of this microbial evolution have recently been reported for baboons and include Papio-unique Brucella sub-species 17,18 and papilloma and HPV2 19 . The definition of "normal" vaginal microbial communities differs among species. A healthy human vaginal environment is characterized by the dominance of lactobacilli 20,21 . These lactobacilli transform glycogen into lactic acid, generating an acidic environment 22 and forming protective biofilms 23 that prevent the colonization and proliferation of potentially pathogenic organisms.
NHPs may rely on different defense mechanisms for protection against sexually transmitted diseases. The differences between humans and NHPs include the vaginal pH (acidic in humans (pH < 4.5) 22 and acidic-alkaline in baboons (pH = 5.5-6.5 24 ), the anatomy of the utero-cervical junction (sharp anterflexio in women compared to "scarcely noticeable" ventroflexio in baboons) 25 , and increased diversity of microbial communities in baboons compared to humans 24 . Interestingly, microbial diversity in primates is determined by the size of the vagina (or baculum length) 8 . The length of the vagina is 10-12 cm in humans 26 and approximately 7 cm in baboons in our study. The discrepancies between published observations (decreased microbial diversity despite increased vaginal size in humans) could be explained, among others things, by the great ability for the vagina to stretch 25 and increase vaginal size due to sexual swelling 27 in baboons. Additionally, social structure and copulative behavior of baboons and humans also differ 28 . Baboons live in harem communities (one male and typically 10-15 females), and males require several vaginal introductions before ejaculation. In general, the specific social structure and higher promiscuity might have been important for promoting species development 29,30 . A comparison of the general distribution of parasites between NHP and humans revealed a relative abundance of fungi and bacteria (22% and 38%, respectively) in humans compared to NHPs (3% and 10%, respectively) 31 . These differences in the overall microbial landscape may be responsible for the development of specific local, including vaginal, protective mechanisms. Interestingly, the differences in vaginal lactobacilli between baboons and humans are not accompanied by differences in vaginal fungal composition 32 . The histological and cytological changes of the vagina during the menstrual cycle are similar in humans and baboons 33 , including an increased level of glycogen-enriched cells during ovulation 33,34 . Differences in the structural morphology of the vagina include epithelial maturation (which occurs in the early proliferative phase in baboons but the ovulation phase in humans), the absence of erythrocytes in the vaginal smear around ovulation 35 and the presence of cornification of the vaginal epithelium in 10% of baboon specimens 36 ; in humans, hyperkeratosis represents a metaplastic change 37 . In Papio spp. the microbial milieu does not change upon the administration of exogenous progestins and is independent of menstrual cycle phases 9,24 , whereas levonorgestrel therapy and menstrual cycle phases are associated with changes in microbial communities in humans 38,39 . Evolutionary pressure may have resulted in the formation of hormone-sensitive microbial communities.
The frequencies of vaginal and vulvar pathologies among all pathological diagnoses in baboons are 0.6% and 0.04% (respectively) 40 . In our study, the most common pathology was vaginal stricture (45%), presumably associated with HPV2 (Simian agent 8) 12 . The disease, which is the most common STD in captive baboons, has devastating consequences in Papio spp., preventing intercoitus 14 . However, recent publications have suggested that these lesions may also be associated with Treponema infection 41,42 . The course of infection with herpesvirus simplex is not as devastating in humans 43 , possibly due to the protective role of L. crispatus during viral infection. Conversely, the clinical course of infection with Treponema pallidum in baboons 41,44,45 is mild compared to that in humans. Baboons have not been reported to have STDs caused by Ureaplasma, Gardnerella vaginalis, Atopobium vaginae, or Megasphaera I. In agreement with this observation, we did not detect these four species in our sample set. Interestingly, in contrast to humans, baboons do not exhibit increased numbers of infection-related stillbirths and preterm births 25,46 .
The abundance of lactobacilli in our study (21.5%) is in agreement with a previous report 9 in which lactobacilli were detected in 16% of wild-caught baboons but lower than the rate reported by Skangalis et al. (47.4%) 47 . L. crispatus is one of the most frequently detected phylotypes in the human vaginal microbiome (85%) 11 , but is among the lactobacilli with the lowest abundances in baboons 8 . In agreement with this observation, L. crispatus was detected in only one animal in our study (7.1%), a young female in a harem cage of 11 females. Yildirim et al. detected L. crispatus in olive but not yellow baboons 8 . The species in our study are hybrids of yellow, olive, and hamadryas baboons; therefore, it is difficult to draw conclusions regarding the specificity of subspecies. In Rhesus macaques (another Old World NHP), the abundance of L. crispatus is much lower (0.65%) 48 , and L. johnsonii 49 and L. reuteri 48 are predominant. In humans, L. crispatus protects against G. vaginalis 50 , which has not been detected in the baboon vaginal microbiome. Remarkably, the genome of G. vaginalis includes the tetracycline resistance gene (tet(M)). This gene is also detected in N. gonorrhoeae and U. urealyticum, vaginal microbial species that are present in humans but absent in baboons 51 . However, the tetM gene was the most abundant gene in vaginal swabs of wild and captive baboons 52 . The source of this gene remains to be elucidated. L. crispatus protects against viral infection 50,53 . Viral infection of cytotrophoblasts decrease their invasive capacity 54 , leading to shallow trophoblast invasion. Trophoblast invasion in baboons is shallow in contrast to deep invasion in humans 55 . In humans, the abundance of L. crispatus may decrease the viral load and thus promote trophoblast invasion (Fig. 1).
but L. gasseri was present in 2/304 samples, and the most common was L. johnsonii (85/304 48 ), which is related to L. iners and L. gasseri. L. iners has the shortest genome 57 and is dominant in Caucasian/Asian women (34.1%) 58 , whereas L. gasseri is present at a much lower abundance (6.3%) 58 . Considering the evolution of macaques, baboons and hominids 59,60 , the absence of L. iners might be the result of intra-species evolution. In humans, bacterial vaginosis is associated with an abundance of Megasphaera type I, BVAB2, Gardnerella vaginalis and Atopobium vaginae 61 . Megasphaera type I, BVAB2, and G. vaginalis are rare or absent in sexually unexposed women. In our study, we did not detect G. vaginalis, Atopobium vaginae, and Megasphaera type I in baboons. In agreement with observations in humans, the majority of BVAB2-positive animals (four out of five) were multiparous 14-to 15-year-old animals, an age comparable to perimenopause in humans 62 . Only one nulliparous young animal was BVAB2-positive, which was attributed to the housing of this baboon in the harem cage with the other BVAB2-positive animals. The diagnosis of BV is non-existent in NHPs. Interestingly, the majority of the vaginal anaerobic flora in baboons is represented by the common species of BV in humans (Sneathia from the phylum Fusobacteria 24 ). These microbes produce short chain fatty acids (SCFAs) 63 , volatile substances, which stimulate the mating behavior of NHPs 64 . Lactobacilli and an acidic environment in the vagina may be predisposing factors for the acquisition of BV in baboons.
In conclusion: our study confirmed the low abundance of human-specific Lactobacillus spp. in baboons. The absence of L. iners, Gardnerella vaginalis, Atopobium vaginae, and Megasphaera I in the vaginal microflora of Papio spp. is a novel finding. The presence of lactobacilli might indicate a predisposition to BV in NHPs.

Materials and Methods
Animal characteristics, housing and handling. Overall study design. This study included three groups of baboon, hybrids of yellow, olive, and hamadryas baboons (Papio spp.). In group I, morphometric measurements of external and internal external genitalia were obtained during bi-annual health checks (n = 16) or necropsy (n = 31). In group II, animals with available pathology records on pathological vulvar and vaginal changes were retrospectively analyzed (n = 64). In group III, vaginal swabs from baboons obtained during health exams were analyzed by polymerase chain reaction (PCR) (n = 14).
Group composition and animal housing. Group I. Baboons were housed in two open-top 6-acre metal and concrete corrals with dirt floors and gang cages with concrete floors at the SNPRC (Southwest National Primate Research Center, Texas Biomedical Research Institute) as previously described) 65 Group II. Pathology records Morphometry of external genitalia. Animals were sedated via intramuscular injection of ketamine (10 mg/kg) as described previously 65 . The ano-genital distance was measured with a measuring tape from the middle of the anus to the middle of the introitus. The diameter of the introitus was measured from the upper to the lower pole ( Fig. 2A). During necropsy, the length of the vagina was measured using a ruler from the introitus to the cervix (introitus to cervix distance) and to the left fornix (introitus to fornix distance) (Fig.2B).
Collection of vaginal specimens. Vaginal specimens were collected using sterile cotton swabs after the perineal skin was cleaned with Betadine solution and rinsed several times with sterile saline solution. Specimens were stored at − 80 °C until further evaluation (8-9 years).

Polymerase chain reaction.
A real-time PCR (qPCR) assay was used to detect and determine the relative concentrations of the vaginal flora as described previously 66,67 . The qPCR assays identified vaginal Lactobacillus spp., including L. crispatus, L. gasseri, L. iners, and L. jensenii. The assays also detected facultative anaerobic bacteria (Gardnerella vaginalis, Atopobium vaginae (AV), bacterial vaginosis-associated bacteria (BVAB2), and Megasphaera I and II). qPCR analysis of gene transcripts was performed using a Bio-Rad iCycler RealTime PCR machine and 2× Taqman Master Mix. RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA). Primer probe sets were designed in-house using the software packages Primer ExpressTM v2.0 (Applied Biosystems) and Beacon Designer v2.0 (PREMIER Biosoft International). Additionally PCR, detecting tuf gene, encoding elongation factor Tu, from 33 strains representing 17 Lactobacillus gene target was performed 68 .