33.1 Introduction

A particularly important, yet largely underappreciated, aspect of transgender health relates to the creation of novel microbiome niches with gender-confirmation surgery. In transwomen specifically, gender-confirming surgery may involve the creation of a vault through vaginoplasty [Reference Hadj-Moussa, Ohl and Kuzon1], open to microbial colonization. In ciswomen, it is rather well established that the vaginal microbiome tends to be dominated by niche-specific Lactobacillus taxa, an ecology understood as central to human reproductive health [Reference van de Wijgert, Borgdorff and Verhelst2,Reference Charbonneau, Blanton and DiGiulio3]. Perturbations of this community ecology in turn are known to predispose to an array of adverse health conditions, including an overall increased acquisition risk of sexually transmitted infections (STIs). Apart from surgical interventions, transwomen as well as transmen will most often receive gender-affirming hormonal treatment, which may also affect the structure and function of the vaginal as well as of the neovaginal microbiome. In this chapter, we will review what is known of the neovaginal microbiome, as well as the effects of estrogen and androgen exposure on the vaginal and neovaginal microbiome, respectively.

33.2 The Vaginal Microbiome in Ciswomen and Its Relation to Infectious Disease

As known from culture-based studies [Reference Charbonneau, Blanton and DiGiulio3], microbiome data obtained through molecular approaches do confirm that in a majority of women of reproductive age, the vaginal microbiome tends to be dominated by niche-specific Lactobacillus taxa [Reference van de Wijgert, Borgdorff and Verhelst2]. Such a remarkable microbial community assembly, with a single bacterial genus largely dominating an ecological niche [Reference Anahtar, Byrne and Doherty4], has been suggested to represent a co-evolutionary adaptation to preserve reproduction in response to a unique human lifestyle [Reference Charbonneau, Blanton and DiGiulio3–Reference Mendes-Soares, Suzuki, Hickey and Forney5]. Only a few Lactobacillus taxa appear to have co-evolved with the human vaginal niche, specifically, L. crispatus, L. iners, L. gasseri, and L. jensenii, with only one of these absolutely dominating the vaginal microbiota in a given woman at a given point of time.

A primary biological function of the vaginal microbiome in this respect, is considered to relate to the colonization resistance conferred by this open ecosystem, thereby protecting to a relative extent against invading microorganisms from adjacent niches as well as against microorganisms that are introduced through sexual contact. The vaginal microbiome is further thought to take part in human reproduction, and vaginal microbiome ecology has been associated with various aspects thereof, notably fertility [Reference Reid, Brigidi and Burton6,Reference Koedooder, Mackens and Budding7], pregnancy course and timing of delivery [Reference Charbonneau, Blanton and DiGiulio3], and possibly fetal development [Reference Jašarević and Bale8]. Of further note is, that, albeit contentious, the vaginal microbiome has been proposed as the seeding bank to the uterine microbiome [Reference Verstraelen, Vilchez-Vargas and Desimpel9], and therefore potentially also of importance to uterine transplantation [Reference Jones, Saso and L’Heveder10].

Perturbations of vaginal community ecology, broadly referred to as vaginal dysbiosis, in turn predispose a host of adverse health outcomes. This relates at least in part to lower genital tract inflammation and decreased colonization resistance in the absence or relative absence of indigenous lactobacilli [Reference Anahtar, Byrne and Doherty4,Reference Borgdorff, Gautam and Armstrong11]. Both phenomena likely explain for instance the overall increased risk of STI acquisition observed with vaginal dysbiosis, as most extensively documented for HIV (see also Chapter 37) and HPV (see Chapter 36). The most common manifestation of vaginal dysbiosis is known as bacterial vaginosis (BV) and involves the establishment of a Gardnerella vaginalis-dominated biofilm adherent to the cervicovaginal epithelium [Reference Verstraelen and Swidsinski12]. Bacterial vaginosis remains most often asymptomatic, yet in those women reporting symptoms, it typically persists and recurs as a frustrating condition presenting as foul-smelling vaginal discharge. In ciswomen, it is considered the most common microorganism-associated vulvovaginal condition, followed by Candida vulvovaginitis, which in turn is most often caused by C. albicans. Often present as part of the normal mycobiome or fungal microbiome in the gastrointestinal and genital tracts, C. albicans may assume an infectious, invasive lifestyle [Reference Swidsinski, Guschin and Tang13] in an inflammation-driven condition, known as Candida vulvovaginitis. Candida vulvovaginitis is indifferently associated with genital itch against a variable background of inflammatory symptoms, such as edema and skin fissures. Development and recurrence of this common condition does however not seem to be related in a consistent manner with the vaginal (bacterial) microbiome, albeit Lactobacillus dominance may enhance the presence of Candida as such [Reference Willems, Ahmed, Liu, Xu and Peters14].

33.3 The Vaginal Microbiome and the Role of Hormones in Ciswomen and Transmen

The primary driver of Lactobacillus dominance in the vaginal niche is generally assumed to be circulating estradiol, although mechanisms involved have not been fully elucidated. The availability of glycogen [Reference Mirmonsef, Hotton and Gilbert15], which accumulates in the cervicovaginal environment in women of reproductive age, in an estrogen-dependent and insulin-independent manner [Reference Mirmonsef, Hotton and Gilbert16,Reference Verstraelen, Vervaet and Remon17] is, arguably, a key mechanism here. However, for reasons we ignore, the relationship between serum estradiol, cervicovaginal fluid glycogen, and Lactobacillus dominance tends be less straightforward than anticipated [Reference Mirmonsef, Hotton and Gilbert15,Reference Mirmonsef, Hotton and Gilbert16]. If anything, major hormonal shifts that occur with puberty and menopause mark reproductive life transitions, respectively to and from Lactobacillus dominance. In a similar vein, hormonal replacement therapy involving systemic or local estrogen applied intravaginally is an effective means of counteracting postmenopausal atrophy and associated symptoms, collectively referred to as genitourinary syndrome of menopause (GSM). While invariably accompanied by (re-)colonization with lactobacilli, arguably, the estrogen effect primarily relates to the (re-)establishment of a suitable niche for lactobacilli, the cervicovaginal epithelium stacked with glycogen under these conditions, as well as providing a platform for a number of other mucosal–microbial interactions, including adhesion. This may also explain why oral contraceptives seem to provide a relative protective effect against vaginal dysbiosis and BV in particular, while increasing the risk of Candida vulvovaginitis [Reference van de Wijgert, Verwijs, Turner and Morrison18,Reference Vodstrcil, Hocking and Law19]. Along the same lines, the widely used progestogen injectable, depot medroxyprogesterone acetate (DMPA), which has been of concern owing to its association with increased HIV acquisition in high-risk settings, presumably compromises Lactobacillus recruitment through its hypo-estrogenic effect [Reference Wessels, Felker, Dupont and Kaushic20], while allowing more diverse communities to establish in the cervicovaginal niche. Overall, albeit much remains to be determined, and although different estrogens and progestins may differentially impact the vaginal microbiota, further dependent on administration route [Reference Wessels, Felker, Dupont and Kaushic20], current data suggest a protective effect on the vaginal microbiome exerted by estrogens, while progestins may have more detrimental effects, at least, under some conditions involving relative hypo-estrogenism.

Considerably less is known on the effect of androgens on the vaginal microbiome, and hence there is also a dearth of data on the vaginal microbiota in transmen. In one controlled study, involving long-term intramuscular testosterone administration in transmen, a loss of normal architecture of the vaginal epithelium was revealed, intermediate and superficial layers were completely lost, and glycogen content was depleted, which was different from comparable samples obtained from menopausal and even from postmenopausal women [Reference Baldassarre, Giannone and Foschini21]. As may be expected from the above, such conditions are invariably accompanied by loss of Lactobacillus dominance in ciswomen. Indeed, a single study recently compared the vaginal microbiota through sequencing of the V3–V4 regions of the bacterial 16S ribosomal RNA (rRNA) gene between eight ciswomen and 28 transmen. Transmen received mostly systemically administered testosterone for at least 1 year, with a few of them (n = 4) also being treated with local estrogen [Reference Winston McPherson, Long and Salipante22]. It was found that, relative to ciswomen, transmen were significantly less likely to have Lactobacillus-dominated microbiota, while a more diverse microbiota was generally in place (Shannon diversity index of 3.14 (interquartile range 1.93–3.56) compared to 0.75 (interquartile range 0.59–1.29) in ciswomen). Although key taxa of a BV (G. vaginalis and Atopobium vaginae) were present in some transmen, this more diverse microbiota were not those typically seen in women with a non-Lactobacillus-dominated microbiota. Interestingly, although numbers were very small, it was also observed that even under testosterone treatment, topical administration of estrogen again enhanced L. iners dominance in two (2/4) men.

33.4 The Neovaginal Microbiome

As indicated above, gender-confirming surgery may involve the creation of a vault through vaginoplasty by making use of a penile skin flap, scrotal skin grafts, or sigmoid grafts [Reference Hadj-Moussa, Ohl and Kuzon1] as discussed elsewhere. From a microbial ecology perspective, the neovaginal vault may be considered a so-called vacant niche [Reference Litvak and Bäumler23], open to colonization from adjacent microbiome communities. While no mechanistic studies have been performed in this setting, theoretically, colonization of the neovaginal skin and/or mucosa can be expected to originate from the gastro-intestinal tract and the skin, as is thought to occur in ciswomen. Of note here is that the urinary and genital tract microbiomes were recently also suggested to be interconnected [Reference Thomas-White, Forster and Kumar24]. Even when bacteria from adjacent niches [Reference Bauer, Kainz, Carmona-Gutierrez and Madeo25] compete for colonization of the neovaginal niche, arguably, other factors are expected to determine microbiome composition, tissue type in particular, depending on vaginoplasty technique applied [Reference Hadj-Moussa, Ohl and Kuzon1]. A further unknown is how the relatively anaerobic environment, presumably with a fluctuating low pO2, might affect colonization of skin flaps or sigmoid grafts that line the neovaginal vault. While we do largely lack such insights at present, understanding the dynamics of the neovaginal microbiome is of defined importance, as to understand not only how this bacterial community is established, but also as to how it can be preserved, or possibly modified, e.g. in response to vulvovaginal complaints or infectious disease risk.

Microbiome composition of the neovagina has been addressed in only a handful of studies over the past three decades, including one culture-based study [Reference Toolenaar, Freundt and Wagenvoort26] and one bacterial 16S rRNA gene-based study of the sigmoid neovagina [Reference van der Sluis, Bouman and Mullender27] and following vaginoplasty involving penile inversion [Reference Weyers, Verstraelen and Gerris28], and one metaproteomics study in a mixed population, in whom vaginoplasty was performed through miscellaneous techniques [Reference Birse, Kratzer and Zuend29].

In what is presumably the very first study on the issue, Toolenaar and colleagues reported on the microbiome of the sigmoid neovagina in 15 patients, of which 10 were transwomen, and five patients had aplasia as a part of the Mayer-Rokitansky-Küster-Hauser syndrome [Reference Toolenaar, Freundt and Wagenvoort26]. The study was published before the so-called microbiome revolution and was entirely culture-based. Overall, in this pioneering study, the authors found a neovaginal microbiome that strongly resembled the sigmoid colon microbiome with high abundances of some signature taxa such as Escherichia coli, Bacteroides spp., and Prevotella spp., as well as Lactobacillus spp., however with an overall lower abundance as compared to the colon, with reported total counts of 10³ to 10¹¹ colony-forming units (cfu)/ml neovaginal fluid at an average pH of 8 (range 7 to 9). Through conventional culture techniques, a total of 85 different species representing 17 different genera were isolated with an average of six species per patient (range 1 to 9). Lactobacilli in this study were not further identified at the species level, so it remains inconclusive if these were representative for a vaginal and/or gut microbiomes. Of note, the average neovaginal pH in this study was 8, not resembling the typical low pH of a Lactobacillus-dominated vaginal microbiome. The authors further report that 11 patients (73.3%) complained of sticky neovaginal discharge, but no attempts were made to correlate clinical and microbiological data. Use of estrogen replacement therapy was not mentioned in this study.

In what remains the single largest and most comprehensive study on the subject, Weyers and colleagues enrolled 50 transwomen in whom gender-confirmation surgery involved vaginoplasty through an inverted penile skin flap technique [Reference Weyers, Verstraelen and Gerris28]. Microbiome analysis in this patient series encompassed analysis of Gram-stained vaginal smears, aerobic and anaerobic culture followed by identification of isolated colonies through transfer RNA intergenic spacer length polymorphism analysis followed by automated fluorescent capillary electrophoresis. Of note is that the vast majority of patients in this study (47/50 or 94.0%) received estrogen replacement therapy while under study. Gram-stained smears in this patient group revealed highly diverse communities in 44/50 smears (the remainder had few bacteria), and albeit the authors did not report Nugent scores, they do mention that smears showed similarity to a BV microbiota. In fact, none of the Gram-smears indeed showed the presence of lactobacilli as most often present in the vaginal smears of ciswomen, but rather Gram-variable coccobacilli than Gram-negative rods and no clue cells. A selection of images obtained from Gram-stained smears from eight different neovaginal smears (kindly provided by the authors), not published previously, is shown in Figure 33.1.

Figure 33.1 Microscopic images (10x 100x) representing the microbial diversity observed in Gram-stained neovaginal smears in the study from Weyers and colleagues [Reference Weyers, Verstraelen and Gerris28]. Various morphotypes are observed, such as polymorphous Gram-negative rods (A), Gram-positive cocci (B), Gram-negative coccobacilli (C), and fusiform Gram-negatives (D).

Further analysis through culture in a subset of 30 women led to an average of 8.6 species per woman (range 4–14). Among the more common species were Staphylococcus epidermidis (n = 19), Streptococcus anginosus group spp. (n = 16), Enterococcus faecalis (n = 13), Corynebacterium sp. (n = 12), Bacteroides ureolyticus (n = 10 women), and Mobiluncus curtisii (n = 10). Only a single isolate of Lactobacillus species was found, i.e. L. casei, typically not found as part of the vaginal microbiota in ciswomen. Only three isolates were identified as species considered as key markers of BV, specifically one G. vaginalis and two A. vaginae isolates. The microbiological findings were consistent with the elevated pH of the neovagina (mean 5.8; range 5.0–7.0). Patients also addressed a number of questions on subjective well-being through a standardized questionnaire and thereby reported regular episodes of vulvar irritation (22.0%), dysuria (9.0%), and foul-smelling vaginal discharge (23.5%). The authors did not report on any ecological community analysis, yet they did not identify any relation between such vulvovaginal complaints, and lifestyle habits, hygiene, and sexual activity on one hand, and bacterial species identified on the other hand.

Petricevic and colleagues investigated the neovaginal microbiomes of 63 transwomen who underwent sex reassignment surgery using an inverted penile skin flap technique at least 1 year before enrolment [Reference Petricevic, Kaufmann and Domig30]. Neovaginal smears were analyzed by means of PCR-denaturing gradient gel electrophoresis (PCR-DGGE), using lactobacilli species-specific primers. Lactobacillus amplicons were found in 47/63 samples (75%). The 279 Lactobacillus signals belonged to 13 different species, the most abundant being L. gasseri (n=38), L. crispatus (n = 36), L. johnsonii (n = 35), and L. iners (n = 35). More than 90% of women harbored a combination of two or more neovaginal Lactobacillus species. In a related study in (largely) the same transwomen, Petricevic and colleagues further investigated rectal smears for the presence of lactobacilli [Reference Petricevic, Kaufmann and Domig31]. A total of 43 of the 61 transwomen (70.5%) simultaneously harbored the same lactobacilli in both the neovagina and rectum, suggesting that the rectum acts as reservoir for neovaginal colonization of lactobacilli. The findings of Petricevic and colleagues seem to be in contrast with other studies that document a neovaginal microbiome different from a typical Lactobacillus-dominated vaginal microbiome found in women of reproductive age. In both studies however, Petricevic and colleagues used a molecular method (PCR) to specifically amplify Lactobacillus DNA only, which might explain these contradictory findings. Likely, Lactobacillus DNA detected results from minority strains present at low abundance and not from a single Lactobacillus species dominating the econiche, as seen in a typical Lactobacillus-dominated vaginal microbiome found in women of reproductive age. The fact these authors found that over 90% of investigated transwomen harbored two or more different Lactobacillus species also points towards a microbiome not typically seen in ciswomen, where in the majority of cases, only a single Lactobacillus species dominates the econiche.

van der Sluis and colleagues recently reported on a series of 28 patients following sigmoid colon vaginoplasty, which was part of gender-confirmation surgery in the vast majority of women (26/28 or 93.0%) described here [Reference van der Sluis, Bouman and Mullender27]. Sampling of the neovagina was partially longitudinal, albeit the authors refrained from any longitudinal analysis. Quite unique, and of defined interest here, however, is that in all patients, paired neovaginal and rectal samples were obtained. The authors used a 16S-23S ribosomal RNA intergenic spacer PCR-based profiling approach for bacterial identification, in part of the patients further complemented by analysis of the skin microbiota sampled in the hypogastric area. Using a principal coordinate analysis, it was found that neovaginal microbiota showed defined similarity to the rectum, however the neovaginal microbiota also turned out to be distinctively different. According to the authors this was mostly apparent at the phylum level for Bacteroidetes, showing lower abundance and diversity relative to the rectum, while no such differences were observed for other phyla, notably Firmicutes, Actinobacteria, Fusobacteria, and Verrucobacteria. The authors reported the most discriminative species between neovaginal and rectal microbiome but did not specify the microbial composition of the neovaginal microbiome. Hence, it is unknown in this study to what extent the neovaginal microbiome resembled the vaginal microbiome. The authors did not specifically mention estrogen replacement yet stated that none of the patients described in this study were treated with topical (or systemic) steroids or other agents before sample collection.

A fifth study, encompassing five transwomen, employed metaproteomics as an unconventional approach for microbiome analysis [Reference Birse, Kratzer and Zuend29]. Vaginoplasty in this small series involved penile inversion with a scrotal cutaneous graft in four transwomen, and the latter technique complemented with colon sigmoid extension in a fifth transwoman. As in the study by van der Sluis and colleagues, the authors also attempted to obtain paired neovaginal and rectal samples, which is of defined interest to the topic under study. Not all samples obtained however had sufficient amounts of proteins to allow for proteomics, leaving out four neovaginal samples (4/9) and two rectal samples (2/9). Estrogen replacement therapy was documented for three transwomen (3/5) in the analysis and unknown for one patient (1/5). The authors further included data on 30 ciswomen, from whom a vaginal sample was analysed with the very same approach. For the neovaginal samples, an average of eight species were identified (range 2–17), as compared to an average of six species per rectal sample (range 1–18), and of five species per cis vaginal sample (range 1–14). The most abundant protein profiles over all neovaginal samples were from Porphyromonas (30.2%), Peptostreptococcus (9.2%), Prevotella (9.0%), Mobiluncus (8.0%), and Jonquetella (7.3%). Not unexpectedly, the most abundant profiles in ciswomen were Lactobacillus (64.8%), and Gardnerella (18.2%) in transwomen, respectively. In an ecological cluster analysis, the authors showed distinct clusters of cisvaginal Lactobacillus-dominated, cisvaginal Gardnerella-dominated, and rectal microbiomes. One exception to this was the presence of a cisvaginal microbiome dominated by Prevotella, the profile that also dominated the rectal microbiomes. A fourth cluster consisted of microbiomes from rectal, cisvaginal, and neovaginal samples, and these were largely characterized by a rather high diversity of species. Lactobacillus was present in only one neovaginal microbiome, with a relative abundance of about 25%.