Identification and evaluation of the microbiome in the female and male reproductive tracts

Abstract

BACKGROUND

The existence of an extensive microbiome in and on the human body has increasingly dominated the scientific literature during the last decade. A shift from culture-dependent to culture-independent identification of microbes has occurred since the emergence of next-generation sequencing (NGS) techniques, whole genome shotgun and metagenomic sequencing. These sequencing analyses have revealed the presence of a rich diversity of microbes in most exposed surfaces of the human body, such as throughout the reproductive tract. The results of microbiota analyses are influenced by the technical specifications of the applied methods of analyses. Therefore, it is difficult to correctly compare and interpret the results of different studies of the same anatomical niche.

OBJECTIVES AND RATIONALE

The aim of this narrative review is to provide an overview of the currently used techniques and the reported microbiota compositions in the different anatomical parts of the female and male reproductive tracts since the introduction of NGS in 2005. This is crucial to understand and determine the interactions and roles of the different microbes necessary for successful reproduction.

SEARCH METHODS

A search in Embase, Medline Ovid, Web of science, Cochrane and Google scholar was conducted. The search was limited to English language and studies published between January 2005 and April 2018. Included articles needed to be original microbiome research related to the reproductive tracts.

OUTCOMES

The review provides an extensive up-to-date overview of current microbiome research in the field of human reproductive medicine. The possibility of drawing general conclusions is limited due to diversity in the execution of analytical steps in microbiome research, such as local protocols, sampling methods, primers used, sequencing techniques and bioinformatic pipelines, making it difficult to compare and interpret results of the available studies. Although some microbiota are associated with reproductive success and a good pregnancy outcome, it is still unknown whether a causal link exists. More research is needed to further explore the possible clinical implications and therapeutic interventions.

WIDER IMPLICATIONS

For the field of reproductive medicine, determination of what is a favourable reproductive tract microbiome will provide insight into the mechanisms of both unsuccessful and successful human reproduction. To increase pregnancy chances with live birth and to reduce reproduction-related health costs, future research could focus on postponing treatment or conception in case of the presence of unfavourable microbiota and on the development of therapeutic interventions, such as microbial therapeutics and lifestyle adaptations.

female/male, reproduction, microbiota, vagina, cervix, uterus, endometrium, semen, coronal sulcus

Introduction

Microbiome

The number of microorganisms coexisting in and on the human body is estimated to be in the same order of magnitude as the total number of cells that make up the body (Sender et al., 2016). The microorganisms colonising our body are referred to as our microbiota and besides bacteria, include viruses, fungi, yeasts, archaea and protozoa. The collective genetic content of these microorganisms and the surrounding environmental conditions are known as the microbiome (Marchesi and Ravel, 2015). It has been suggested that the combined collection of microbial and human cells should be seen as one single ecological and biological unit, the so-called human holobiont. The host–microbiome relationship can be categorised as commensalism (one species benefits, while the other remains unaffected), mutualism (beneficial to both species) or parasitism (one species benefits at the expense of the other) (Bordenstein and Theis, 2015).

The diversity of microorganisms within a given body site is defined as the number of different microorganisms. The most abundant microorganism is called the dominant one, and each region within the human body has its own characteristic composition. A distinction in diversity is made within (i.e. alpha) and between (i.e. beta) samples. Alpha-diversity refers to the mean species diversity at the region of interest, while the beta diversity reflects the diversity between different regions of interest (Human Microbiome Project C, 2012; Huttenhower et al., 2012), although all microbiota influence and actively interact with each other. However, the focus of this review is on bacteria only, since these microbiota are currently the most extensively investigated. Bacteria are taxonomically classified into different taxa, which are subsequently classified into different ranks, i.e. domains, kingdoms, phyla, classes, orders, families, genera and, finally, species (Fig. 1) (Yarza et al., 2014; Pirih and Kunej, 2018). The evolutionary relationship between bacteria is called phylogeny, and the evolutionary relationships among taxa of bacteria is typically depicted in a phylogenetic tree (Ehrlich, 1965). Microscopy-dependent techniques have been used to identify bacteria based on phenotypic or morphological characterisation and specific cell staining characteristics, whereas current technologies use sequencing of taxonomy-associated markers genes, such as the 16 s rRNA gene, or whole genome sequences (Morgan et al., 2013) to identify bacteria.

Figure 1

Taxonomy versus phylogenetic tree of Lactobacillus crispatus from domain to species.

Preparation of samples for microbiome analyses

For microbiome research, it is important to sample what you want to sample, identify what you want to identify and avoid bacterial contamination to prevent misinterpretation of the results. Therefore, the different necessary steps in microbiome analyses and possible pitfalls are discussed.

Microbiome sampling

Microbiome samples can be collected with many different (commercially available) sets consisting of a swab, tube and, often, an optional buffer solution. Collection can be performed by either a medical professional or by the individual via self-collection. The sampling should be standardised, which implicates the need to understand and strictly adhere to the protocol, and should prevent contamination by surroundings (e.g. skin, hands, dust, clothes) to preserve the original sample.

Sample storage

After sample collection, most samples are stored before further processing. The optimum storage condition generally involves immediate placement on ice and storage at −80°C until further processing (Aagaard et al., 2013; Kim et al., 2017a, 2017b; Chu et al., 2018). One should be aware that storage buffers can be contaminated (Salter et al., 2014) and that the time to place the sample on ice and to store impacts on the final analyses (Schellenberg et al., 2017).

DNA extraction and isolation

Upon sample processing, the first steps are DNA extraction and isolation, which nowadays is mostly done by commercially available DNA extraction kits. Sample contamination can occur by reagents in the kit (Salter et al., 2014; Kim et al., 2017a, 2017b), by the use of laboratory instruments and by the local environment (e.g. gloves and air).

In general, especially when handling low-biomass samples (e.g. semen or meconium) which have low abundance of microbiota, special care is needed to avoid contamination and hence false results (Chafee et al., 2015). It is strongly advised to collect samples in a standardised manner, to use only sterile equipment, to minimise accidental exposure to the environment and to always include positive and negative controls in the analysis (Salter et al., 2014) to prevent erroneous conclusions downstream in the microbiome analysis.

Early microbiome sample analyses

Microscopy

Light microscopy has for decades been the corner stone of microbial (viruses excepted) identification. Bacteria can be identified and taxonomically divided based on characteristics like Gram staining and cell morphology (shape, arrangement, form and size). Microscopy of clinical specimens is still used for the diagnosis of certain infectious diseases. In the field of human reproduction, microscopy still has an important role in the detection of clue cells or determination of the Nugent Score in case of suspicion of bacterial vaginosis (see below) (Nugent et al., 1991). However, the role of microscopy in daily clinical practice of reproductive medicine has diminished due to the relative time-consuming aspect, the disappearance of clinical microscopy training, experience and exposure, and the introduction of new techniques, such as polymerase chain reaction (PCR)-techniques (Bennett et al., 2014). A recent study showed that the sensitivity of a clinician’s diagnosis for bacterial vaginosis is nowadays significantly lower than that by a PCR/fluorogenic probe-based investigational test (P < 0.0001) (Schwebke et al., 2018).

Culture-dependent methods

Culture-dependent methods were the first techniques, after light microscopy, to study microbes and make use of defined growth media for the culture and proliferation of an organism (Sandle, 2011). After a certain time period of culturing, different genera can be assigned by judgement of cell staining characteristics, morphology and/or capacity for biochemical reactions. Although these methods are not expensive, they are time-consuming and elaborate and only microbiota whose metabolic and physiological needs are provided by the specific culture medium will proliferate (Nadkarni et al., 2009), with high-abundant and fast-growing species suppressing others (Hiergeist et al., 2015). Therefore, the biggest limitation of this technique is that the resulting microbial composition is incomplete and not representative (Ward et al., 1990).

Microbiome sequencing

Sequencing techniques based on 16 S ribosomal DNA (rDNA) genes overcome the problems of culture. The region of interest (amplicon or target sequence) in microbiome research is a specific gene, which is highly conserved during evolution and unique for all bacterial species, namely the 16 S rRNA gene (Fig. 2) (Lane et al., 1985). The 16 S rRNA gene is ~1500 base pairs (bp) long, contains nine hypervariable regions (V1–V9) interspersed between highly conserved regions, and codes for a component of the 30 S small subunit of the prokaryotic ribosome (Gray et al., 1984). Universal primers complementary to the highly conserved sequences between the hypervariable regions of the 16 S rRNA gene, can reliably identify different taxa (Větrovský and Baldrian, 2013). The hypervariable regions (V1–9) classify the microorganisms into taxonomic units and are used to perform phylogenetic analyses (Woese et al., 1990). The more sequences of 16S rRNA genes of different bacteria match, the more likely the microbes are related at a higher taxonomic rank, e.g. the threshold sequence identity is 94.5% for genera and 86.5% for families (Yarza et al., 2014, 2014).

Figure 2

A representation of the circular chromosome of bacteria. The 16S and 23S ribosomal RNA genes are highlighted together with the intergenic space (IS) region. V1–V9 marks the variable regions, with the conserved regions between them.

In 1975, the first sequencing technique (Sanger sequencing) was introduced (Sanger and Coulson, 1975), offering more insight and detail into the diversity of microbiota than the former classification relying on the combination of Gram staining and microscopy. Pyrosequencing was the first variant on the classical Sanger sequencing (Nyrén et al., 1993) and was based on detection of luminescence. In 2005, the development of next-generation (also called high-throughput or second generation 16S rRNA) sequencing (NGS) technologies allowed massive parallelisation of sequencing of bacterial DNA (Metzker, 2005) and eliminated the need for isolating cultures prior to analysis. NGS techniques were quickly commercialised by Roche Life Sciences’ 454 platform (Margulies et al., 2005) and Illumina’s MiSeq and HiSeq platforms (Caporaso et al., 2012; Loman et al., 2012). The biggest disadvantage of NGS is the generation of chimeric sequences and sequencing errors (Haas et al., 2011). Furthermore, the different 16 S rRNA primers are of major importance for the accuracy of the analysis, since the specific targeted region influences the types of bacteria included in the analysis (Baker et al., 2003; Mizrahi-Man et al., 2013; Tremblay et al., 2015; D’Amore et al., 2016).

Bioinformatic pipelines and operational taxonomic unit clustering

In order to analyse the large datasets of NGS analyses, bioinformatical pipelines (prescribed sets of processing steps converting raw data to interpretable material) are available. Frequently used examples of such bioinformatical pipelines in the field of microbiota research are mothur (Schloss et al., 2009) and Quantitative Insights Into Microbial Ecology (QIIME) (Caporaso et al., 2010). The purpose of a pipeline is to assign the resulting DNA sequences to taxonomic levels (from phylum to species) to determine the microbial composition and richness of the sample. The resulting DNA sequences are clustered to one another according to their similarity. Typically, above 97% sequence similarity, sequences are combined into operational taxonomic units (OTU). Nowadays, reassessment of the threshold of 97% has been proposed due to the availability of large numbers of sequences. It was suggested that the optimal identity thresholds is increased to around or higher than 99% (Edgar, 2018). These OTUs can then be further classified taxonomically. OTUs can be constructed by several methods (Westcott and Schloss, 2015) including the closed-reference method, the de novo method and the open method.

In the closed-reference method, resulting sequences are clustered against an external (often open source) reference database, such as Greengenes (McDonald et al., 2012), SILVA (Pruesse et al., 2007) or Ribosomal Database Project (Cole et al., 2008). The process is parallelizable, quick and suited for large datasets. However, sequences which are not present in the database will be processed as non-recognisable reads and will be discarded in further analysis.

In the de novo method, resulting sequences are clustered against one another without the use of an external reference database. All sequences will be clustered, however not in parallel, which can be time consuming and will often be unsuitable for large datasets.

The open method combines both of the above-mentioned methods. Sequences are initially clustered against an external reference database, and non-recognised ones will subsequently be clustered de novo.

Since the genetic differences of species clustered in the same OTU are neglected, clustering leads to loss of actual diversity (Chen et al., 2013). Importantly, the method used for OTU clustering will directly influence the final results (Nguyen et al., 2016).

Open source bioinformatics software packages, such as mothur (Schloss et al., 2009) and QIIME (Caporaso et al., 2010) combine all the sequence processing steps, without applying a standardised procedure and changes in analytic parameters which can lead to over- or under-interpretation in the microbial composition.

Whole genome shotgun sequencing

Whereas 16 S rRNA-based analyses are limited to the examination of the bacterial diversity, whole genome sequencing (WGS) techniques allow sequencing of both the whole genome presented as well as all the genomes present (Ranjan et al., 2016). WGS allows information about the function of genes and identification of novel genes, encoded metabolic pathways, the structure and organisation of genomes and the community structure (Roumpeka et al., 2017).

IS-pro technique

The intergenic spaces (IS)-pro technique is another microbial profiling technique, based on amplification of the IS regions, whose lengths are specific for each bacterial species (Fig. 2) (Budding et al., 2010).

Quantitative (real-time) PCR

Among the sequencing-related techniques currently available, quantitative PCR (qPCR) has proven itself as a sensitive method for the specific detection of individual species or bacterial groups (Huijsdens et al., 2002; Ott et al., 2004). Quantitative real-time PCR with species-specific probes is especially able to monitor quantitative changes, since this method is based on the continuous monitoring of changes in fluorescence during PCR (Malinen et al., 2003).

The microbiome in health and disease

In 2008, the NIH Common Fund Human Microbiome Project (HMP) in the USA was established (Peterson et al., 2009) to characterise the human microbiome at five different body sites (nasal passages, oral cavity, skin, gastrointestinal and urogenital tract) in healthy individuals and analyse how our microbiota contribute to normal physiology and predisposition to disease (Turnbaugh et al., 2007; Methé et al., 2012). Simultaneously, in Europe, the international microbiome consortium, the Metagenomics of the Human Intestinal Tract (MetaHIT) was initiated in 2008 (Qin et al., 2010) with the objective to establish associations between the genes of the human intestinal microbes and health and disease (Li et al., 2014).

The HMP showed that interpersonal variation (beta diversity) was significantly larger than intrapersonal variation (alpha-diversity). In addition, the vagina contained the lowest alpha-diversity, with relatively low beta diversity at the genus level but very high diversity among OTUs due to the presence of distinct Lactobacillus spp. (Human Microbiome Project C 2012).

A symbiotic relationship between host and the residing microorganisms is necessary to maintain health and avoid disease (Martin and Schwab, 2012) and an imbalance in this relationship can result in a dysbiotic state (Knight et al., 2017). These dysbiotic states have been shown to be associated with diseases such as dental caries and bacterial vaginosis. Dental caries is associated with increased phylogenetic diversity and overabundance of Prevotella taxa (Yang et al., 2012; Peterson et al., 2013). Bacterial vaginosis is characterised by a shift from a ‘healthy’ state with a low-pH, Lactobacillus-dominated community to a higher-pH, more diverse microbial community (Fredricks et al., 2005). Importantly, non-communicable diseases (NCDs) such as obesity, cardiovascular disease and inflammatory bowel disease, but also malnutrition have been linked to dysbiotic states (Knight et al., 2017).

Nowadays, it is known that the gut microbiome assists in digestion of food and produces vitamins and other compounds that influence human health by affecting host metabolism and immune responses (Flint et al., 2012; Walker and Lawley, 2013; Maranduba et al., 2015; Li et al., 2017). In addition, environmental factors such as use of antibiotics, diet and geography, as well as ethnicity, can strongly influence the composition of the gut microbiome (Gill et al., 2006; Cresci and Bawden, 2015; Doré and Blottière, 2015).

However, the shifts between symbiosis and dysbiosis and vice versa are unpredictable processes and their causes are not yet understood.

The microbiome in reproductive health and disease

The different parts of the male and female reproductive tract constitute a system subdivided by anatomical or physiological barriers. The male reproductive tract consists of an external part, the penis and the scrotum, and the internal part, the testes, epididymis, accessory glands, vas deferens and urethra. The female reproductive tract constitutes the vagina, the cervix and the uterine cavity which extends into the fallopian tubes, localising the fimbriae near the ovaries.

Whether microbes in the reproductive tract have a beneficial role for female health and reproductive success comparable to the behaviour of symbiotic microbiota in the local gut is unknown. However, more and more evidence is accumulating indicative of comparable roles.

Recently, Chen et al. (2017) showed the presence of a specific microbial composition which changes along the course of the female reproductive tract, confirming other findings that the reproductive tract is indeed not sterile. Mitchell et al. (2015), e.g. found the presence of endometrial bacteria at a significantly lower quantity as compared to the vaginal quantity, suggesting that the cervix serves as a partial filter or barrier for ascending microbiota. The vaginal microbiome of the non-pregnant, healthy women appears to be dynamic and influenced by ethnicity, sexual activity, the menstrual cycle and the local microbiota (Zhou et al., 2007; Ravel et al., 2011; Gajer et al., 2012), and is mostly dominated by four Lactobacillus spp., i.e. L. crispatus, L. iners, L. jensenii or L. gasseri (Ravel et al., 2011).

In the assisted-reproductive technology (ART) setting, the presence of a diverse vaginal microbiome seems to influence pregnancy outcome negatively (Hillier et al., 1995; Moore et al., 2000; Hyman et al., 2012, 2014), while the opposite has been observed for a Lactobacillus dominant vaginal microbiome (Moore et al., 2000). Vaginal microbiota composed solely of Lactobacillus (L. crispatus, L. iners, L. jensenii, L. gasseri or other Lactobacillus species) at the cycle before embryo transfer have been associated with successful outcome of the IVF-embryo transfer procedure (Hyman et al., 2012). More specifically, Hyman et al. investigated the correlation between the vaginal microbiome composition during infertility therapy with the subsequent clinical outcomes in 30 patients. They concluded that the number of bacterial genera on swabs taken at the time of embryo transfer was significantly different (P = 0.028) between patients who had a live birth and patients who did not have a live birth.

Colonisation of the follicular fluid by microorganisms has been suggested as a potential cause of adverse pregnancy outcomes in IVF since colonisation of the follicular fluid at the time of oocyte retrieval was associated with higher embryo discard rates (P < 0.0001), lower rates of embryo transfer (P = 0.0001) and lower pregnancy rates (P < 0.05) in both fertile and infertile women (Pelzer et al., 2013). These results were based on 263 women who had two types of specimens collected: follicular fluid samples from the left and right ovary (n = 463) and vaginal swabs (n = 263) (Pelzer et al., 2013). Lower embryo transfer rates were associated with the presence of Propionibacterium spp. (P < 0.05) and Streptococcus spp. (P < 0.01) in right follicles, whereas higher embryo transfer rates were associated with the presence of Lactobacillus spp. (P < 0.05) in both the right and left follicles. Negative pregnancy outcomes were found with the presence of Actinomyces spp., Bifidobacterium spp., Propionibacterium spp. and Streptococcus spp. (P < 0.01) within the left ovarian follicular fluid and Actinomyces spp., Bifidobacterium spp. and Streptococcus intermedius (P < 0.01) within the right ovary. Positive pregnancy outcomes were found with the presence of Lactobacillus spp. (P < 0.001) in the left ovary (Pelzer et al., 2013). Additionally, Moore et al. (2000) demonstrated that the presence of Streptococcus viridans on the embryo transfer catheter tip is associated with adverse ART outcomes.

During pregnancy, a decrease in richness and diversity of the vaginal microbiome occurs (Aagaard et al., 2012) with a transition towards a Lactobacillus-dominated community (Aagaard et al., 2012; Romero et al., 2014; MacIntyre et al., 2015). Aagaard et al. (2012) compared vaginal samplings of 24 pregnant women with 60 non-pregnant controls and found diversity and richness to be reduced during pregnancy, whereas the microbiome of women closer to term returned to the non-pregnant microbiome state. Two independent studies, MacIntyre et al. (2015) characterising 42 pregnant women and Romero et al. (2014) characterising 22 pregnant and 32 non-pregnant women, confirmed these findings of a vaginal microbiome during pregnancy dominated by Lactobacillus spp. with low alpha-diversity. MacIntyre et al. also reported opposite finding during the postpartum period with a vaginal microbiome less dominated by Lactobacillus spp. albeit with increased alpha-diversity. Women without Lactobacilli-dominance accompanied by elevated Gardnerella or Ureaplasma abundances and pregnant women with an increased vaginal microbiome instability have a higher risk of preterm birth (DiGiulio et al., 2015; Stout et al., 2017). During pregnancy, the shift in microbial composition is possibly a response to the increased oestrogen levels (MacIntyre et al., 2015), whereby the shift to Lactobacillus-dominated communities serves as a protection against bacterial vaginosis (Ravel et al., 2011) with an associated reduced risk of preterm birth (Hyman et al., 2014). These finding indicate that the vaginal microbiome in general is not static with shifts in microbial composition occuring regularly and influenced by many factors (Forney et al., 2006; Zhou et al., 2007; Kim et al., 2009; Gajer et al., 2012).

The future goal of research in the field of the reproductive microbiome should be determination of a ‘healthy’ microbiome (Gevers et al., 2012; Lloyd-Price et al., 2016) to provide more insight in how and when an altered microbiome leads to disease and in how to use the microbiome as a biomarker of reproductive health. However, the increasing variety of analytical and bioinformatic tools and methods used for processing of the sequencing data creates a challenge if one tries to compare and interpret the different studies as a whole. An extensive summary of the currently published data of the microbiome in the healthy female and male reproductive tract is provided within this article. The findings will be discussed in light of clinical implications and future perspectives of reproductive health and ART.

Methods

Search strategy

We performed a search in Embase, Medline Ovid, Web of Science, Cochrane and Google Scholar. The search strategy included keywords related to scientific literature concerning the microbiome of the female and male reproductive tract, such as (alone or in combination): microflora, microbiota, genital tract, reproduction, semen, vagina, uterus, cervical, placenta, conception, assisted reproduction and urogenital microbiome. A protocol for this review has been registered in PROSPERO International prospective register of systematic reviews (2016: CRD42016042506).

Inclusion and exclusion criteria

Articles published from 2005 (after the introduction of the NGS technique) until April 2018, written in English and available online, were eligible for inclusion. Eligible studies, had to be related to original microbiome research in the female or male reproductive tract and could include culture-dependent or culture-independent techniques. Reviews, publications with no new results, and articles of which only the abstract was available were excluded, as were as comments on published articles and reports on animal research. Studies describing the microbiome of body parts other than the reproductive tract were not withheld. The populations of interest were males, women of reproductive age before conception and pregnant women until mid-gestation. We focused on publications which included at least one group of a healthy population and excluded adolescents (under 18 years of age), postmenopausal women and patients with diseases such as Human Immunodeficiency Virus (HIV), (chronic) prostatitis or vaginitis.

Selection procedure

Titles and/or abstracts of studies retrieved using the above-mentioned search strategy were screened independently by two reviewers (R.K. and S.S) to identify studies eligible for full-text screening. The full texts of these eligible articles were retrieved and independently assessed by the same two reviewers. Any disagreements concerning the eligibility of particular studies were resolved through discussion with a third reviewer (J.L.).

Study selection

A flowchart of the search strategy and study selection process of the articles is shown in Fig. 3. The search yielded 5201 results, with 2944 unique articles. After screening of titles and abstracts, 2792 articles were excluded. After reading the full-text articles, 51 articles were eligible for inclusion. A quality assessment was not performed, due to the variety of collection/extraction protocols and analytical/bioinformatical methods. Tables 1 and 2 provide an overview of the definitive selected articles and summarise the applied methods, techniques and results. A subdivision was made between articles that used culture-dependent techniques, culture-independent techniques or a combination of both. In addition, a distinction has been made between articles that studied a predetermined selection of microorganisms (selective) and articles that analysed the whole spectrum of detected microbiota (non-selective). Some articles describe multiple sample sites (e.g. vagina and cervix) and therefore, the total number of articles used in Tables 1 and 2 adds up to 89 from the 51 original articles.

Figure 3

PRISMA flow methodology for the selection of relevant manuscripts.

Table I

Included studies for the female reproductive tract. Overview of study characteristics and reported taxonomic assignments.

No.	(Author, year) Country	Sample	Population (N)	Selective/non-selective	Technique	16 S rRNA region	Database	Taxonomic assignment (reported)
Anatomical region: Vagina
1	Pascual et al. (2006) Argentina	Posterior fornix	Reproductive-age women (N = 100)	S	Culture			L. acidophilus, L. fermentum, L. gasseri, L. brevsi, L. jensenii, L. casei subsp. casei, L. delbrueckii subsp. delbrueckii, Peptostreptococci, Streptococci, Bifidobacteria, Propionibacteria
2	Aleshkin et al. (2006) Russia	Vaginal wall	Pregnant and non-pregnant women, healthy pregnant women (first trimester) (N = 200)	NS	Culture			Lactobacillus spp., Gardnerella vaginalis, Bifidobacterium spp., Clostridium spp., Propionibacterium spp., Mobiluncus spp., Peptostreptococcus spp., Peptococcus spp., Bacteroides spp., Prevotella spp., Porphyromonas spp., Fusobacterium spp., Veillonella spp., Corynebacterium spp., Staphylococcus spp., Streptococcus spp., Streptococcus group B, Streptococcus group D, Neisseria spp., Enterobacteriaceae, Candida spp.
3	Anukam et al. (2006) Nigeria	Vaginal	Healthy premenopausal women (N = 241)	S	PCR	V2–V3	GenBank DNA databases, BLAST algorithm	L. iners, L. gasseri, L. plantarum, L. suntoryeus, L. crispatus, L. rhamnosu, L. vaginalis, Lactobacillus spp., L. fermentum, L. helveticus, L. johnsonii, L. salivarius
4	Jakobsson and Forsum (2007) Sweden	Upper third vagina	IVF patients (N = 22)	S	Culture, NGS			L. iners, L. gasseri, L. jensenii, Mobiluncus
5	Garg et al. (2009) India	High vaginal wall	Healthy reproductive-age women (N = 80)	S	Culture, PCR		BLAST	L. reuteri, L. fermentum, L. salivarius, L. plantarum, L. crispatus, L. jensenii), L. gasseri, L. acidophilus, L. casei, L. paracasei, L. rhamnosus, L. delbruckii
6	Pelzer et al. (2011) Australia	Vaginal	IVF patients (N = 71)	S	Culture, PCR		Basic Local Alignment Search Tool (BLAST, NCBI)	A. meyeri, Bacteroides spp., Bifidobacterium spp., Bifidobacterium spp., Candida albicans, C. glabrata, Clostridium butyricum, C. ramosum, Corynebacterium spp., Escherichia coli, Enterococcus faecalis, Egghertella lenta, Gemella spp., L. crispatus, L. gasseri, L. jensenii, Propionibacterium acnes, S. epidermidis, S. lugdunensis, Sterptococcus spp., S. agalactiae, S. viridans
7	Hyman et al. (2012) USA	Posterior fornix	IVF patients (N = 30)	NS	Sanger Sequencing		Ribosomal Database Project (RDP)	Lactobacillus
8	Ekanem et al. (2012) Nigeria	Posterior fornix	Non-pregnant reproductive-age women (N = 220)	NS	Culture			Lactobacillus sp., Diphtheroids, Staphylococcus epidermidis, Staphylococcus aureus, Staphylococcus faecalis, Bacteroides sp., Peptostreptococcus sp., Eschericia coli, Candida albicans, Gardnerella vaginalis, Streptococcus agalactiae, Peptococcus sp., Clostridium sp., Proteus sp.
9	Gajer et al. (2012) USA	Mid-vaginal	Reproductive-age women (N = 32)	NS	NGS	V1–V2	RDP Naïve Bayesian Classifier, Lactobacillus: speciateIT	L. iners, Atopobium,L. jensenii, Prevotella, Aerococcus, Shigella, Megasphaera, Eggerthella, Gemella, Peptoniphilus, L. gasseri, Finegoldia, Other Phyloptypes
10	Mangot-Bertrand et al. (2013) France	Vaginal	IVF patients (N = 307)	S	qPCR			Lactobacillus spp., G. vaginalis, A. vaginae, Mycoplasma hominis
11	Pendharkar et al. (2013) South Africa	Vaginal	Premenopausal black women with or without BV (N = 30)	S	Culture, PCR	Complete 16 S rRNA gene	BLASTN, Genbank accession number	L. crispatus, L. iners, L. gasseri, L. jensenii, L. vaginalis, L. ruminis, L. mucosae, L. paracasei, L. coleohominis
12	Brotman et al. (2014) USA	Mid-vaginal	Premenopausal women (30)	NS	NGS	V1–V2	RDP Classifier, Lactobacillus: speciateIT	L. crispatus L. iners, L. gasseri, L. jensenni, Atopobium, Megasphaera, Prevotella, Sneathia, Streptococcus, Ruminococcaceae, Lachnospiraceae, Aerococcus, Lachnospiraceae, Anaerococcus, Diaphorobacter, Peptinophilus, Lachnospiraceae, Parvimonas, L.otu2, Proteobacteria, Proteobacteria, Dialister, Veillonella, Ruminococcaceae, Finegoldia
13	Liu, et al. (2013) China	Vaginal fornix and lower third of vagina	Healthy women and women with BV and/or VVC (N = 95)	NS	NGS	V6	Global Alignment for Sequence Taxonomy (GAST)	Lactobacillus, Gardnerella, Streptococcus, Prevotella, Granulicatella, Bifidobacterium, Dialister, Sneathia, Alloscardovia, Parvimonas, Escherichia, Peptostreptococcus, Anaerococcus, Haemophilus, Peptinophilus, Bacillus, Aquabacterium, Mobiluncus, Sphingomonas, Ralstonia
14	Bahaabadi et al. (2014) Iran	Vaginal	Infertile women (N = 100)	S	PCR		NCBI gene bank	M. hominis
15	Albert et al. (2015) Canada	Vaginal	Healthy reproductive-age women (N = 310)	NS	NGS, cpn60 PCR	V3	Bowtie 2, mPUMA, cpn60 reference database	L. crispatus, L. jensenii, Atopobium vaginae, Streptococcus devriesei, L. acidophilus, L. iners, Weissella viridescens, Desulfotalea psychorophila, Peptoniphilus harei, Clostridium innocuum Streptococcus parasanguinis, Gardnerella vaginalis subgroup A, Gardnerella vaginalis subgroup C, Prevotella tannerae, Faecalibacterium prausnitzii, L. gasseri, Sphingobium yanoikuyae, Gardnerella vaginals subgroup B, Massilia timonae, Acidaminococcus fermentans, Megasphaera sp. genomsp. type 1, Prevotella timonensis
16	Gautam et al. (2015) Kenya, Rwanda, South Afrcia, Tanzania	Cervicovaginal	Pregnant and non-pregnant women (N = 430)	NS	Microarray		Ribosomal Database Project, Genbank	L. crispatus, L. iners, Gardnerella vaginalis, Atopobium vaginae, Prevotella spp., G. vaginalis, A. vaginae, Prevotella spp., Dialister, Megasphaera spp., Mobiluncus spp., lowest abundance L. iners, Prevotella spp., Megasphaera spp.
17	Jespers et al. (2015) Africa	Vaginal	Pregnant and non-pregnant women (N430)	S	Culture, qPCR			Lactobacillus genus, Lactobacillus crispatus, Lactobacillus iners, Lactobacillus jensenii, Lactobacillus gasseri, Lactobacillus vaginalis, Gardnerella vaginalis, Atopobium vaginae, Prevotella bivia, Escherichia coli, Candida albicans
18	Mitchell et al. (2015) USA	Vaginal	Women undergoing hysterectomy for benign disease (N = 58)	S	Culture, qPCR			Prevotella spp., L. Iners, L. Crispatus, G vaginalis, A vaginae, L. jensenii
19	Moreno et al. (2016) Spain	Posterior fornix	Fertile women (N = 13)	NS	NGS	V3–V5	QIIME, UCLUST algorithm	Lactobacillus spp., Atopobium, Clostridium, Gardnerella, Megasphaera, Parvimonas, Prevotella, Sphingomonas, Sneathia genera, Gardnerella, Clostridium, Sneathia, Prevotella spp., Atopobium, Gardnerella, Prevotella, or Sneathia
20	Haahr et al. (2016) Denmark	Posterior fornix	IVF patients (N = 130)	S	Culture, qPCR			Atopobium vaginae, Gardnerella vaginalis, L. Iners, L. Crispatus, L. Jensenii, L. Gasseri
21	de Vieira Santos-Greatti et al. (2016) Brazil	Vaginal	Non-pregnant reproductive-age women (N = 783)	S	qPCR			G. vaginalis
22	Zozaya et al. (2016) USA	Vaginal	Women with or without BV (N = 96)	NS	Pyrosequencing		Ribosomal Database Project	Megasphaera, BVAB1, P. bivia, Prevotella, Gardnerella, Aerococcus, L. iners, Porphyromonas, Sneathia, Leptotrichia, Atopobium, Actinomyces, Megasphaera1, Eggerthella, Anaerococcus, Dialister, BVAB2, M. hominis, Peptoniphilus, Lactobacillus sp., Barnesiella, Gemella, Peptostreptococcus, Parvimonas, P. disiens
23	Babu et al. (2017) India	Posterior fornix	Healthy women and women with infertility problems (N = 200)	NS	Culture			Healthy: Lactobacillus, Micrococcus, Enterococcus, Coagulase-negative Staphylococcus spp.
24	Freitas and Hill (2017) Canada	Vaginal	Healthy reproductive-age women (N = 492)	S	cpn60 PCR, qPCR	V3	cpnDB reference database	Bifidobacterium breve, B. longum, B. dentium, Alloscardovia omnicolens
25	Kim et al. (2017) Korea	Posterior fornix	Pregnant women (N = 168)	S	qPCR			L. crispatus, L. iners, L. jensenii, L. gasseri, L., vaginalis, G. vaginalis and A. vaginae
26	Nasioudis et al. (2017) USA	Posterior vaginal wall	First trimester pregnant women (N = 154)	NS	NGS	V1–V3		Lactobacillus crispatus, L. iners, L. gasseri, Gardnerella, L. jensenii, Streptococcus, Bifidobacterium, L. helveticus, L. acidophilus, L. johnsonii
27	Campisciano et al. (2017) Italy	Cervical-vaginal	Infertile and fertile women (N = 96)	NS	NGS	V1–V3	Vaginal 16 S rDNA Reference Database	Idiopathic bacilli, Actinobacteria, Gammaproteobacteria, Tenericutes, Clostridia, Bacteroidia
28	Wee et al. (2017) Australia	Posterior fornix	Infertile women and fertile controls (N = 31)	NS	NGS, PCR			Bacilli, Actinobacteria, Gammaproteobacteria, Tenericutes
29	Son et al. (2018) Korea	Posterior fornix	Pregnant women (1) first trimester (N = 221), (2) second trimester (N = 138)	NS	Culture			E. coli, Enterobacteriae, Klebsiella pneumonia, Pseudomonas aeruginosa, Staphylococcus aureus, Streptococcus agalactiae, Enterococcus faecalis, Candida, Gram-negative bacteria, Gram-positive cocci, Other gram-positive cocci E. coli, Enterobacteriae, Klebsiella pneumonia, Staphylococcus aureus, Streptococcus agalactiae, Candida, Gram-negative bacteria, Gram-positive cocci, Other gram-positive cocci
Anatomical region: Cervix
30	Fotouh and Al-Inany (2008) Egypt	Cervical mucus samples, catheter tip	IVF/ICSI patients (N = 25)	NS	Culture			Staphylococcus aureus, Coagulase-negative staphylococcus Streptococci, Diphteroids, Lactobacilli, Gram-negative bacteria, Klebsilla spp., Pseudomonous spp., Proteus, Non-lactose fermenters, E. coli
31	Simhan and Krohn (2009) USA	Cervical	Pregnant women first trimester (N = 218)	S	Culture or PCR			Neisseria gonorrhoeae, Chlamydia trachomatis, Trichomonas vaginalis
32	Prabha, Aanam, and Kaur (2011) India	Cervical area	Women with unexplained infertility (N = 27)	NS	Culture			Staphylococci, Micrococci, Streptococci, Bacillus, E. coli, Pseudomonas
33	Costoya et al. (2012) Chile	Intracervical	Patients with tubo-peritoneal infertility and normal fertile patients (N = 60)	S	PCR		SYBR Safe DNA gel stain	Mycoplasmas
34	Cicinelli et al. (2012) Italy	Cervical	Women referred for diagnostic hysteroscopy (N = 404)	S	Culture			Streptococci, E. coli, E. Faecalis, Ureaplasma, Gardnerella vaginalis
35	Ekanem et al. (2012) Nigeria	Cervical canal	Non-pregnant reproductive-age women (N = 225)	NS	Culture			Lactobacillus sp., Diphtheroids, Staphylococcus epidermidis, Staphylococcus aureus, Staphylococcus faecalis, Bacteroides sp., Peptostreptococcus. sp., E. coli, Candida albican, Gardnerella vaginalis, Streptococcus agalactiae, Peptococcus sp., Clostridium sp., Proteus sp.
36	Smith et al. (2012) Costa Rica	Exfoliated cervical cells	Women (N = 10)	NS	Sanger sequencing, NGS	V6, V6–V9	usearch, RDP Classifier, pplacer	Lactobacillus, Gardnerella, Prevotella, Megasphaera, BVAB1/Clostridiales, Howardella
37	Kasprzykowska et al. (2014) Poland	Cervical	Women with no symptoms of genital tract infection (N = 40)	S	PCR			Mycoplasma spp., U. Parvum, U. Urealyticum
38	Anahtar et al. (2015) South African	Cervical	HIV-negative women (N = 94)	NS	NGS, WGS	V4		Fusobacterium, Aerococcus, Sneathia, Gemella, Mobiluncus, Prevotella, Shuttleworthia, Clostridiales, Mycoplasma, Lactobacillus iners, Leptotrichiaceae
39	Gautam et al. (2015) Kenya, Rwanda, South Africa, Tanzania	Cervicovaginal	Pregnant and non-pregnant women (N = 430)	S	PCR			Neisseria gonorrhoeae, Chlamydia trachomatis
40	de Vieira Santos-Greatti et al. (2016) Brazil	Endocervical	Non-pregnant reproductive-age women (N = 783)	S	PCR			C. trachomatis, N. gonorrhoeae
41	Seo et al. (2016) South Korea	Cervical	Women with CIN and control women (N = 137)	NS	NGS	V1–V3	EzTaxon-e, BLASTN, Mothur
42	Panda et al. (2016) India	Cervical	Unexplained infertile women (N = 296)	NS	culture			Micrococcus spp., diptheroids, non-enterococcal group D Streptococcus, Staphylococcus aureus, coagulase negative Staphylococcus, Enterococcus spp., Bacillus spp., E. coli, Klebsiella spp., Acinetobacter spp., Candida spp.
43	Campisciano et al. (2017) Italy	Cervical-vaginal	Idiopathic (1), Infertile (2) and fertile (3) women (N = 96)	NS	NGS	V1–V3		(3) Bacilli, Actinobacteria, Gammaproteobacteri, Tenericutes
44	Wee et al. (2017) Australia	Endocervical canal	Infertile women and fertile controls (N = 31)	NS	NGS, PCR			No information
45	Campos et al. (2018) Brazil	Endocervix	Women with (1) and without (2) endometriosis (N = 104)	S	PCR			(2) Mycoplasma genitalium, Mycoplasma hominis, Ureaplasma urealyticum, Ureaplasma parvum
46	Di et al. (2018) Italy	Endocervical	Women (N = 35)	S	NGS	V3–V4	SILVA rRNA reference database	C. trachomatis, Neisseria gonorrhoeae, Trichomonas vaginalis, Mycoplasma, Candida, Firmicutes, Actinobacteria, Fusobacteria, Proteobacteria, Tenericutes, Bacteroidetes, Lactobacillus, Atopobium, Bifidobacterium, L. crispatus, L. gasseri, L. inesr, Leptotrichia amnionii, Gardnerella vaginalis, Prevotella spp. Actinobacteria, L. crispatus, L. gasseri, Leptotrichia amnionii, G. vaginalis, Prevotella spp.
47	Graspeuntner et al. (2018) Germany	Cervix	Women with infectious (1) and non-infectious infertility (2), female sex workers (3) and healthy controls (4) (N = 190)	NS	culture, PCR, NGS	V3/V4		(3) Lactobacillus, Gardnerella, Prevotella, Sneathia, Clostridiales, N. gonorrhoeae, C. trachomatis (4) Lactobacillus, Gardnerella, Prevotella, Sneathia, C. trachomatis
48	Taylor et al. (2018) USA	Cervical	Women (N = 250)	S	PCR			C. trachomatis, N. gonorrhoeae, M. genitalium, G. vaginalis, Sneathia spp., U. urealyticum, A. vaginae, BVAB1
Anatomical region: Endometrium
49	Fotouh and Al-Inany (2008) Egypt	Cervical mucus samples, catheter tip	IVF/ICSI patients (N = 25)	NS	Culture			Staphylococcus aureus, Coagulase-negative staphylococcus, Streptococci, Diphteroids, Lactobacilli, Klebsilla spp., Pseudomonous spp., Proteus, Non-lactose fermenters, E. coli
50	Cicinelli et al. (2012) Italy	Endometrial	Women referred for diagnostic hysteroscopy (N = 404)	S	Culture			Streptococcus Agalactiae, Enterococcus faecalis, E. coli, U. urealyticum, Mycoplasma, Staphylococci, Gardnerella vaginalis
51	Moreno et al. (2016) Spain	Endometrial fluid	Fertile women and IVF patients (N = 70)	S	NGS	V3–V5	Ribosomal database project classifier method v2.2	Lactobacillus, Gardnerella, Bifidobacterium, Streptococcus, Prevotella
52	Verstraelen et al. (2016) The Netherlands	Endometrial tissue and mucus	Women with various reproductive conditions (N = 19)	NS	NGS	V1–2	Ribosomal Database Project, NCBI database	Bacteroides xylanisolven, Bacteroides thetaiotaomicron, Bacteroides fragilis, Bacteroides vulgatus, Bacteroides ovatus,Pelomonas, Betaproteobacteria, Escherichia/Shigella, Chitinophagaceae. Lactobacillus iners, Prevotella amnii, Lactobacillus crispatus, Gardnerella vaginalis, Atopobium vaginae
53	Franasiak et al. (2016) USA	Endometrial (transfer catheter)	Patients undergoing embryo transfer (N = 33)	NS	NGS	V2–4–8, V3–6, V7–9	RDP classifier (Naïve Bayesian classification), Greengenes database	Flavobacterium, Lactobacillus, Limnohabitans, Polynucleobacter, Bdellovibrio, Chryseobacterium, Spirochaeta, Clostridium, Blvii28, Pseudomonas, Fluviicola, Paludibacter, Curvibacter, Methylotenera, Pelosinus, Acidovorax, Delftia, Janthinobacterium, Streptococcus, Candidatus Aquiluna, Pedobacter, Caloramator, Sulfuricurvum, Shuttleworthia, Salinibacterium, Sulfurospirillum, Paucibacter, Acinetobacter, Microbacterium, Cellvibrio
54	Tao et al. (2017) USA	Endometrial (transfer catheter)	IVF patients (N = 70)	NS	NGS	V4	RDP 2.2 in QIIME using the Greengenes database	Lactobacillus spp., Corynebacterium spp., Bifidobacterium spp., Staphylococcus spp., Streptococcus
55	Wee et al. (2017) Australia	Endometrial biopsy	Infertile women and controls (N = 31)	NS	qPCR			No information
56	Moreno et al. (2018) Italy	Endometrial biopsy	Patients assessed for chronic endometritis (N = 113)	S	Culture, PCR, NGS	V2–4–8, V3–6, V7–9	Greengenes database	Chlamydia trachomatis, Enterococcus, E. coli, Gardnerella vaginalis, Klebsiella pneumoniae, Mycoplasma hominis, Neisseria gonorrhoeae, Staphylococcus, Streptococcus
57	Taylor et al. (2018) USA	Endometrial	Women (N = 250)	S	PCR			E. faecalis, E. coli, Staphylococcus epidermidis, Staphylococcus aureus, Streptococcus agalactiae, Streptococcus mitis and yeasts
Anatomical region: Upper Genital Tract
58	Costoya et al. (2012) Chile	Fallopian tubal flushings	Patients with tubo-peritoneal infertility and normal fertile patients (N = 60)	S	PCR		SYBR Safe DNA gel stain	Mycoplasmas
59	Kasprzykowska et al. (2014) Poland	Fluid from the pouch of Douglas	Women with no symptoms of genital tract infection (N = 40)	S	qPCR			Mycoplasma spp., U. Parvum, U. Urealyticum
60	Pelzer et al. (2011) Australia	Follicular fluid	ART patients (N = 71)	NS	Culture, PCR		Basic Local Alignment Search Tool (BLAST, NCBI)	Actinomyces species, A. israelii, A. naeslundii, C. parapsilosis, C. auromucosum, Fusobacterium spp., Lactobacillus spp., L. iners, P. avidum, P. granulosum, P. propionicus, Prevotella disiens, P. melanogenicus, Peptinophilus asaccharolyticus, Peptostreptococcus spp., Staphylococcus spp., Propionibacterium, Prevotella, Staphylococcus spp.
61	Pelzer et al. (2012) Australia	Follicular fluid	IVF patients (N = 36)	NS	Culture, qPCR		Basic Local Alignment Search Tool (BLAST, NCBI)	Lactobacillus gasseri, L. Crispatus, L. Jensenii, CoNS, Propionibacterium spp., Peptostreptococcus spp., B. Longum, S. Agalactiae, S. Anginosus, Micrococcus spp., Salmonella enterica, E. coli, Lactobacillus species, Propionibacterium spp., Peptostreptococcus spp., Salmonella enterica
62	Pelzer et al. (2013) Australia	Follicular fluid	IVF couples (N = 263)	S	Culture		Basic Local Alignment Search Tool (BLAST, NCBI)	Lactobacillus spp., Bifidobacterium spp., Staphylococcus spp.
63	Mitchell et al. (2015) USA	Upper genital tract: endocervix + endometrial fluid	Women undergoing hysterectomy for benign disease (N = 58)	S	Culture, qPCR			Lactobacillus iners, L crispatus, L jensenii, Gardnerella vaginalis, Atopobium vaginae, Megasphaera spp., Prevotella spp., Leptotrichia/Sneathia, BVAB1, BVAB2, BVAB3
64	Campos et al. (2018) Brazil	Peritoneal fluid	Women with (1) and without (2) endometriosis (N = 104)	S	PCR			(2) Mycoplasma genitalium, Mycoplasma hominis
65	Campos et al. (2018) Brazil	Biopsied tissue samples	Women with (1) and without (2) endometriosis (N = 104)	S	PCR			(2) Mycoplasma genitalium

Table II

Included studies for the male reproductive tract. Overview of study characteristics and reported taxonomic assignments.

No.	Year (Author) Country	Sample	Population (N)	Selective/non-selective	Technique	16 S rRNA region	Database	Taxonomic assignment (reported)
Sample: Semen
66	Virecoulon et al. (2005) France	Semen	Patients from infertile couples (N = 543)	NS	Culture, PCR			Coagulase-negative Staphylococci, Streptococcus agalactiae, S. anginosus, S. constellatus, S. mitis, S. oralis, non-hemolytic Streptococci, Enterococcus spp., Escherichia coli, Proteus mirabilis, Gardnerella vaginalis, Corynebacterium spp., Lactobacillus spp., Ureaplasma urealyticum
67	Gdoura et al. (2007) Tunisia	Semen	Infertile men (N = 120)	S	PCR			U. parvum, M. hominis, M. genitalium
68	Kiessling et al. (2008) USA	Semen	Men undergoing fertility evaluation or vasectomy (N = 34)	NS	Taq polymerase		GenBank BLASTn search	Peptoniphilis, Anaerococcus, Finegoldia, Peptostreptococcus spp., Corynebacterium spp., Staphylococcus, Lactobacillus, Streptococcus, Pseudomonas spp., Haemophilus, Acinetobacter spp.
69	Zinzendorf et al. (2008) Africa	Semen	Asymptomatic men undergoing fertility evaluation (N = 927)	NS	Culture			U. urealyticum, M. hominis
70	Ivanov, Kuzmin, and Gritsenko (2009) Russia	Seminal fluid	Healthy (1) men and men with chronic prostatis syndrome (2) (N = 108)	NS	Culture			(1) S. haemolyticus, S. saprophyticus, S. capitis, S. hominis, S. aureus, Corynebacterium genitalium, C. pseudogenitalium, Lactobacillus spp., Streptococcus spp., Micrococcus spp.
71	De Francesco et al. (2011) Italy	Semen	Men investigated for subfertility and healthy normo-zoospermic controls (N = 732)	NS	Culture			Gardnerella vaginalis, Escherichia coli, Enterococcus spp. Ureaplasma urealyticus, Staphylococcus spp., Streptococcus agalactiae, Enterococcus sp., Streptococcus viridans, Escherichia coli, Morganella morganii, Proteus mirabilis, Citrobacter koseri, Enterobacter aerogenes, Klebsiella pneumoniae, Serratia fonticola
72	Domes et al. (2012) Canada	Semen	Non-azoospermic subfertile men (N = 4935)	NS	Culture			Enterococcus fecalis, E. coli, group B Streptococcus, Staphylococcus aureus, Klebsiella pneumoniae Proteus mirabilis, Citrobacter koseri, Morganella morganii
73	Hou et al. (2013) China	Semen	Sperm donors and infertility patients (N = 77)	NS	Culture, pyrosequencing	V1–V2	SILVA bacterial sequence database with the use of Mothur	Streptococcus, Corynebacterium, Finegoldia, Veillonella. Lactobacillus, Prevotella, Staphylococcus, Anaerococcus, Peptoniphilus, Incertae sedis, Porphyromonas, Clostridiales, Corynebacterium, Finegoldia, Anaerococcus, Ralstonia, Streptococcus, Pelomonas, Acidovorax, Atopobium, Veillonella, Prevotella, Aerococcus, Gemella
74	Bahaabadi et al. (2014) Iran	Semen	Infertile men (N = 100)	S	qPCR		NCBI gene bank	Mycoplasma, M. hominis
75	Weng et al. (2014) Taiwan	Semen	Men (N = 96)	NS	NGS			Lactobacillus iners, Prevotella sp., Gardnerella sp., Lactobacillus sp., Pseudomonas sp., Prevotella bivia. Genera; Lactobacillus Pseudomonas, Prevotella, Gardnerella, Rhodanobacter, Streptococcus, Finegoldia, Haemophilus
76	Filipiak et al. (2015) Poland	Semen	Infertile men (N = 72)	S	Culture			E. faecalis, E. coli, S. aureus, Ureaplasma sp., Ch. Trachomatis, Klebsiella oxytoca, Morganella morganii, Proteus mirabilis, M. hominis, Chlamydia
77	Palini et al. (2016) Italy	Semen	Patients admitted to semen analysis (N = 20)	NS/S	PCR, culture			Staphylococcus spp., viridans streptococci, Gram-negative bacilli (not identified), Proteus mirabilis, Escherichia coli, Enterococci
78	Godovalov and Karpunina (2016) Russia	Seminal plasma	Men of infertile couples (N = 71)	S	Culture			Streptococci, Enterococci, Staphylococci, Candida fungi, Enterobacteria, anaerobes
79	Ahmadi et al. (2017) Iran	Seminal fluid	(1) Infertile men having abnormal semen parameters and (2) healthy fertile men (N = 330)	S	qPCR, culture			(2) M. hominis
80	Mändar et al. (2017) Estonia	Semen	Men with (1) and without (2) prostatitis (N = 67)	NS	NGS	V6		(2) Lactobacillus iners, Lactobacillus crispatus, Gardnerella vaginalis, Corynebacterium seminale, Peptoniphilus asaccharolyticus, Atopobium vaginae, Enterobacter cowanii, Pseudomonas veronii, Campylobacter rectus, Bacteroides ureolyticus, Anaerococcus hydrogenalis, Streptococcus infantis, Acinetobacter johnsonii, Varibaculum cambriense, Peptostreptococcus anaerobius, Janthinobacterium lividum
81	Chen et al. (2018) China	Seminal plasma	(1) Healthy men, (2) patients with obstructive and non-obstructive azoospermia (N = 17)	NS	NGS		RDP classifier	(1) Lactobacillus, Prevotella, Proteus, Pseudomonas, Veillonella, Corynebacterium, Rhodococcus, Staphylococcus and Bacillus
82	Italy	Urine, semen	(1) Infertile patients and (2) healthy volunteers (N = 660)	NS	Culture			(2) Enterococcus faecalis, E. coli, Staphylococcus haemolyticus, Streptococcus agalactiae, Proteus mirabilis, Klebsiella pneumoniae
83	Monteiro et al. (2018) Portugal	Semen	(1) Infertility-related cases and (2) controls (N = 118)	NS	NGS	V3–V6	Greengenes database	(2) Enterococcus, Staphylococcus, Anaerococcus, Peptoniphilus, Caulobacteraceae, Pasteurellaceae Aggregatibacter, Pasteurellaceae Haemophilus, Enterobacteriaceae Klebsiella, Enterobacteriaceae Morganella, Actinobacteria Actinomycetaceae, Actinobacteria Corynebacterium, Actinobacteria Propionibacterium, Bacteriodetes Flavobacteriaceae
Anatomical region: Coronal Sulcus
84	Price et al. (2010) Uganda	Coronal sulcus	HIV-negative men before (1) and after (2) circumcision (N = 12)	NS	Pyrosequencing	V3–V4	Ribosomal Database Project (RDP) Naı¨ve Bayesian Classifier	(1) Pseudomonadaceae, Oxalobacteraceae, Corynebacteriaceae, Clostridiales Family XI, Staphylococcaceae, Prevotellaceae, Moraxellaceae, Comamonadaceae, Bifidobacteriaceae. Xanthomonadaceae, Enterobacteriaceae, Fusobacteriaceae, Aeromonadaceae, Veillonellaceae, Sphingomonadaceae, Aerococcaceae, Peptostreptococcaceae, Carnobacteriaceae, Streptococcaceae, Micrococcaceae, Flavobacteriaceae, Burkholderiales Family V, Porphyromonadaceae, Caulobacteraceae, Enterococcaceae, Lachnospiraceae, Burkholderiaceae, Campylobacteraceae, Coriobacteriaceae, Rhodocyclaceae, Actinomycetaceae, Intrasporangiaceae, Planctomycetaceae, Halomonadaceae, Brevibacteriaceae, Bradyrhizobiaceae, Mycoplasmataceae, Pseudomonadales Family VI (2) Pseudomonadaceae, Oxalobacteraceae, Corynebacteriaceae, Clostridiales Family XI, Staphylococcaceae, Prevotellaceae, Moraxellaceae, Comamonadaceae, Bifidobacteriaceae. Xanthomonadaceae, Enterobacteriaceae, Fusobacteriaceae, Aeromonadaceae, Veillonellaceae, Sphingomonadaceae, Aerococcaceae, Peptostreptococcaceae, Carnobacteriaceae, Streptococcaceae, Micrococcaceae, Flavobacteriaceae, Burkholderiales Family V, Bacillaceae, Caulobacteraceae, Enterococcaceae, Burkholderiaceae, Rhodocyclaceae, Actinomycetaceae, Intrasporangiaceae, Planctomycetaceae, Halomonadaceae, Brevibacteriaceae, Neisseriaceae, Bradyrhizobiaceae, Dermabacteraceae, Rhodobacteraceae, Pseudomonadales Family VI
85	Nelson et al. (2012) America	Coronal sulcus	Adolescent men (N = 18)	NS	Sanger, PCR, pyrosequencing	V1–V3, V3–V5, V6–V9	NCBI using BLASTN, subset: SILVA database	Corynebacteria, Staphylococcus, Anaerococcus. Peptoniphilus, Prevotella, Finegoldia, Porphyromonas, Propionibacterium, Delftia. Corynebacterium, Staphylococcus, Anaerococcus, Unclassified, Prevotella, Peptoniphilus, Finegoldia, Porphyromonas, Propionibacterium, Delftia
86	Liu et al. (2013) Uganda	Coronal sulcus	Circumcised (1) and uncircumcised (2) men (N = 156)	NS	qPCR, pyrosequencing	V3–V6	Ribosomal Database Project Naïve Bayesian Classifier	(1) Peptoniphilus spp., Anaerococcus spp., Unclassified Clostridiales, Prevotella spp., Finegoldia spp., Murdochiella spp., Porphyromonas spp., Corynebacterium spp., Dialister spp., Negativicoccus spp., Peptostreptococcus sp., Mobiluncus spp., Gardnerella spp., Lactobacillus spp., Staphylococcus spp., Saccharofermentans spp., Streptococcus spp., Actinomyces spp., Veillonella spp., Peptococcus spp., Olsenella spp., Arcanobacterium spp., Howardella spp., Parvimonas spp., Atopobium spp., Sneathia spp., Sutterella spp., Moryella spp., Peptostreptococcaceae family, Treponema spp., Fusobacterium spp., Pyramidobacter spp., Facklamia spp., Anaerosphaera spp., Kocuria spp., Megasphaera spp., Micrococcus spp., Gemella spp., Ralstonia spp.
								(2) Peptoniphilus spp., Anaerococcus spp., Prevotella spp., Finegoldia spp., Murdochiella spp., Porphyromonas spp., Corynebacterium spp., Dialister spp., Negativicoccus spp., Peptostreptococcus spp., Mobiluncus spp., Gardnerella spp., Lactobacillus spp., Staphylococcus spp., Saccharofermentans spp., Streptococcus spp., Actinomyces spp., Veillonella spp., Peptococcus spp., Olsenella spp., Arcanobacterium spp., Howardella spp., Parvimonas spp., Atopobium spp., Sneathia spp., Sutterella spp., Moryella spp., Peptostreptococcaceae family, Treponema spp., Fusobacterium spp., Pyramidobacter spp., Facklamia spp., Anaerosphaera spp., Kocuria spp., Megasphaera spp., Micrococcus spp., Gemella spp., Ralstonia spp.
87	Zozaya et al. (2016) USA	Urethral and penile skin	Male partners of women with (1) and without (2) BV (N = 130)	NS	Pyrosequencing			Megasphaera, BVAB1, P. bivia, Prevotella, Gardnerella, Aerococcus, L. iners, Porphyromonas, Sneathia, Leptotrichia, Atopobium, Actinomyces, Megasphaera1, Eggerthella, Anaerococcus, Dialister, BVAB2, M. hominis, Peptoniphilus, Lactobacillus sp., Barnesiella, Gemella, Peptostreptococcus, Parvimonas, P. disiens
Sample: Urine
88	Virecoulon et al. (2005) France	Ffirst void urine	Patients from infertile couples (N = 543)	S	PCR			Chlamydia
89	Nelson et al. (2012) USA	Urine	Adolescent men (N = 18)	NS	Sanger, PCR, pyrosequencing	V1–V3, V3–V5, V6–V9	NCBI using BLASTN, subset: SILVA database	Corynebacteria, Staphylococcus, Anaerococcus, Peptoniphilus), Prevotella, Finegoldia, Porphyromonas, Propionibacterium, Delftia, Streptocccus, Lactobacillus, Staphylococcus, Gardnerella, Unclassified, Corynebacterium, Veillonella, Anaerococcus, Prevotella, Escherichia/Shigella

The reported microbiome composition of the female reproductive tract is discussed and subdivided in four anatomical regions: vaginal, cervix, endometrium and upper reproductive tract (including the fallopian tubes, ovaries and peritoneal fluid). The reported microbiome composition of the male reproductive tract is discussed and subdivided in two categories: semen and coronal sulcus. Figure 4 gives an overview of all the reported microbiota, at the family level for each part of the reproductive tracts. Overview of detected microbiota reported in a minimum of two independent scientific studies, per individual part of the tracts, are shown in Figs 5 and 6. We opted to mainly focus on results of non-selective studies, including both culture-dependent and culture-independent, to minimise bias and to compare the results as fairly as possible.

Figure 4

Representation of all the detected and reported microbiota of the 51 included articles. Bacteria are shown at the level of family, for each part of the reproductive tracts. The legend on the right explains the different bacteria by colour.

Figure 5

Representation of described bacteria described at the level of family per anatomical region or sample. Female: vagina, cervix, endometrium, upper genital tract. Male: semen and coronal sulcus. On the Y-axis are the names of the reported family of bacteria. The X-axis mentions the number of independent publications which detected and reported the microbiota.

Figure 6

Heatmaps indicating the presence or absence of the bacteria at the level of family classified in the publications per anatomical region or sample. Female: vagina, cervix, endometrium and upper genital tract. Male: semen and coronal sulcus. The numbers at the top of each heatmap refers to the selected article in Tables 1 and 2. The Y-axis represents the dendrogram of the reported bacteria.

Results and Discussion

Overview of reported microbiota in literature

Tables 1 and 2 summarise the results of the reported microbiota compositions of the different anatomical regions of the female and male reproductive tract. A subdivision has been made between articles that solely made use of culture-dependent techniques, culture-independent techniques or a combination of a culture-dependent technique followed by culture-independent techniques. In addition, a distinction can be made between articles that made use of a predetermined selection of microorganisms (selective studies) and articles that analysed the whole spectrum of detected microbiota (non-selective studies). To establish an visual overview of the composition of the microbial communities in the different anatomical regions of healthy individuals, we summarised all microbiota on family level when they were reported in at least two independent studies (Figs 4, 5 and 6). To minimise bias and distorted composition, non-selective studies were chosen to compare the microbiome results, since these cover and detect the most extensive overview of present bacteria.

The most common present microbial family present throughout the female reproductive tract as a whole, were Lactobacillaceae. However, each anatomical region seems to carry its own unique distinct composition of microbes (Figs 4, 5 and 6). Within the lowest part of the female reproductive tract, Bifidobacteriaceae, Prevotellaceae and Veillonellaceae, besides Lactobacillaceae, are the most commonly detected bacterial families in the vagina. In addition, only Lactobacillaceae were found regardless of the used technique. Although, the cervical channel is in direct contact with the environment of the vagina, the detected microbiota are different. The cervical microbiota are predominantly composed of Clostridiaceae, Enterobacteriaceae, Staphylococcaceae and Streptococcaceae, along with Lactobacillaceae. When the results are separated by the two different techniques, the compositions of the vaginal and cervical microbiome resemble each other. The culture-dependent technique results for both regions show Enterobacteriaceae, Staphylococcaceae and Streptococcaceae, whereas with the culture-independent technique Lactobacillaceae and Prevotellaceae are detected. In other words, the comparison of the results of the different studies which used the same techniques seems to be more similar.

The reported uterine cavity and endometrial microbiome mainly consist of Streptococcaceae, besides the above-mentioned Lactobacillaceae. As both families are also reported in the vagina (Lactobacillaceae) and cervix (Streptococcaceae), possible contamination during collection cannot be excluded, despite preventive measures. In order to correct for possible contamination by vaginal microbes, endometrial fluid paired with cervical or vaginal samples were compared in several studies (Fotouh and Al-Inany, 2008; Cicinelli et al., 2014a, b; Wee et al., 2017; Taylor et al., 2018). Contamination was assumed if microorganisms were detected in both samples, but the conclusions of these studies vary. Wee et al. (2017) reported that the dominant microbial community members are consistent in the vagina, cervix and endometrium, although the relative proportions varies, while Cicinelli et al. (2012) concluded that the cervical samples have a low concordance (33%) with the endometrial samples.

In our literature overview, no family was reported more than once in uterine or endometrial samples in the culture-dependent techniques articles. We conclude that it is not possible to determine the standard microbiome of the most commonly found bacterial families in the uterine cavity with this technique. However, in the culture-independent studies, Lactobacillaceae, Bifidobacteriaceae, Comamonadaceae and Streptococcaceae were reported more than once in the uterine cavity, suggesting that the culture-independent techniques are more valid. It seems that we can conclude that the microbiome in the uterus indeed has its own composition, since with the exception of Lactobacillaceae, the bacterial families reported were not previously found as the most common bacterial families in the vaginal and cervical microbiomes.

In the upper female reproductive tract, besides Lactobacillaceae, Peptostreptococcaceae and Propionibacteriaceae were detected, both of which have not been reported for any other anatomical part of the reproductive tract.

Microbial collections of the different internal anatomical regions of the male reproductive tract are more difficult to obtain and pose a great risk of contamination. It still needs to be determined whether each of these different parts carry a unique microbiome. Until then, the seminal microbiome is regarded as the collective end result to which all different parts have contributed. The microbial composition of the semen seems to contain Staphylococcaceae, Streptococcaceae, Enterobacteriaceae and Enterococcaceae as well as Lactobacillaceae. The reported presence of microbes in semen are in agreement with our own results (internal communication). The microbiome of the easy accessible coronal sulcus showed different bacterial families as compared to the semen, namely Porphyromonadaceae and Prevotellaceae. The results of the male reproductive tract also indicate the presence of a unique microbiome of the different accessible parts.

Clinical implications for reproductive health

The vagina is the gateway between the external environment and the reproductive tract higher up and can be affected by changes in both exogenous and endogenous sources. One of the possible roles of the microbiome is to protect the reproductive tract against various infections. The above-mentioned results show that the majority of women have a Lactobacillus spp. dominated reproductive tract microbiome, which seems to be essential for preventing entry or overgrowth of pathogens. An important characteristic of Lactobacillus is the production of lactic acid (Gajer et al., 2012), necessary to provide an acidic environment, which interferes with proliferation of other bacteria (Alakomi et al., 2000; O’Hanlon et al., 2011; O’Hanlon et al., 2013). The difference in vaginal microbial composition between women is reflected in their vaginal pH-levels (see below) (Gajer et al., 2012). Another property of the Lactobacilli that interferes with growth of other bacteria is the production of connections, called bacteriocines (Mendes-Soares et al., 2014; Ojala et al., 2014). Furthermore, Lactobacillus produces both d- and l-isomers of lactic acid, whereas the human body itself is only capable of producing the l-isomer (Mendes-Soares et al., 2014). The main advantage of d-lactic acid is that it down-regulates matrix metalloproteinase (MMP)−8, enabling the cervical plug to maintain integrity and thereby limit vertical transmission of vaginal bacteria into the uterus (Witkin et al., 2013). At the same time, Lactobacilli act as a mechanical barrier by binding to the surface of epithelial cells, which prevents the binding of other bacteria (Mendes-Soares et al., 2014; Ojala et al., 2014).

The vaginal microbiome can be divided into five specific major community state types (CST) I–V by using targeted sequencing of the 16 S rRNA gene (Ravel et al., 2011), of which four are dominated by Lactobacillus. Group I is dominated by the specie Lactobacillus crispatus (26.2%), group II by L. gasseri (6.3%), group III by L. iners (34.1%) and group V by L. jensenii (5.3%). Group IV is not dominated by Lactobacillus, but contains a variety of more strict anaerobes (Ravel et al., 2011). CST IV-A is characterised by some Lactobacillus spp. and a variety of strictly anaerobic bacteria, whereas CST IV-B is characterised by a mix of the genus Atopium, Prevotella, Sneathia and Gardnerella among others (Gajer et al., 2012). In addition, Albert et al. (2015) expanded the range of CSTs by using a slightly different technique and observed a CST dominated by Gardnerella subgroups (CST IVC and IVD). Although the impact of hormonal variations in a natural cycle or during ART (Jakobsson and Forsum, 2007; Hyman et al., 2012) on the human vaginal microbiome needs further research, in animal studies hormone therapy has been shown to alter the composition of the vaginal flora due to Lactobacillus being dependent on oestrogens (Bezirtzoglou et al., 2008).

The vaginal microbiome has been found to be dynamic, since women experience transitions between CSTs over time (Gajer et al., 2012). However, not all transitions between CSTs are equally common. CST IV-B often changes into III but rarely into I. CST-I often changes to III or IV-A. CSTIII changes twice as often to IV-B, as compared to IV-A. CST-II rarely changes and no change from CST-I to CST-II has been observed. In addition, CST-II is relatively stable compared to CST IV-A over a 16-week period (Gajer et al., 2012). These findings suggest that point estimates of community composition could be misleading in case participants belong to a community state type that shows considerable changes over time. Most of the transitions to other state types were, however, transient in nature with 35% of all alternative state types persisting for less than a week. Transition between CSTs seems mainly affected by the timing in the menstrual cycle e.g. menstruating or not, the community class itself and by sexual activity (Gajer et al., 2012).

The difference in microbial composition is also reflected in vaginal pH-levels, as mentioned above. CST-I seems to have the lowest median pH (4.0 ± 0.3), whereas CST-IV shows the highest median pH (5.3 ± 0.6). The difference in pH between the different CSTs is most likely explained by the specific dominance of Lactobacilli and ability per Lactobacillus to produce lactic acid (Gajer et al., 2012).

Ethnicity also seems to influence CST. Anahtar et al. (2015) reported a lower percentage (37%) of Lactobacillus dominance and 45% Gardnerella dominant communities in cervicovaginal microbiota of South African women when compared to the Lactobacillus percentage in white (90%) and black (62%) women from earlier publications (Ravel et al., 2011; Zhou et al., 2007). Of the investigated women with Lactobacillus dominance, 77% had L. iners (Anahtar et al., 2015). Despite a difference in CST, Anukam et al. (2006) and Pendharkar et al. (2013) demonstrated that African and Caucasian women are colonised by the same Lactobacillus species.

Vaginitis

Vaginal symptoms such as discharge, odour, itching or burning can be caused by vaginitis. The most common causes are bacterial vaginosis (BV), vulvovaginal candidiasis and trichomoniasis (Sobel, 1997), respectively, associated with overgrowth of the following species: Bacteroides and Mobiluncus, Candida spp. and Trichomonas vaginalis. The pathogens of vaginitis can be determined by use of combination of clinical symptoms and microscopy. Clinical methods to diagnose bacterial vaginosis are based on either Amsel et al. (1983) or Nugent et al. (1991). BV is the most common vaginal microbial disorder, described as a polybacterial dybiosis (van de Wijgert et al., 2014), affecting 30% of the women during reproductive age (Workowski and Bolan, 2015). Anaerobes like Gardnerella, Atopobium, Mobiluncus, Mycoplasma, Dialister, Sneathia and Prevotella are examples of possible agents in BV (Onderdonk et al., 2016; Liu et al. 2013a, b). With BV, the bacterial composition changes to a higher diversity (Gottschick et al., 2017), leading to an increase in vaginal pH (Brooks et al., 2017). Importantly, BV is associated with adverse reproductive outcomes, such as infertility, miscarriage (Donders et al., 2009), recurrent pregnancy loss (three or more successive miscarriages) (Işik et al., 2016) and preterm birth (Onderdonk et al., 2016). Besides microscopy, determination of the vaginal microbiome composition could have great clinical potential for assessing, predicting and treating BV.

Detection of budding yeast or pseudohyphae on wet mount by light microscopy or positive culture are used to diagnose vulvovaginal candidiasis (Workowski and Bolan, 2015).

Trichomoniasis is diagnosed through microscopic observation on wet mount, by culturing or by biochemical detection through antigen-, nucleic acid hybridisation-, or nucleic acid amplification-based assays (Prevention, 2015; Workowski and Bolan, 2015). Although these microbes are distinguishable by light microscopy, we expect the NGS-technique could play a bigger role in diagnosing vaginitis in the near future due to the lack of experienced clinical microscopists.

Infertility and ART

The microbiome of the reproductive tract has been associated with the chance of conception for natural conceptions as well as in ART cycles. Fertility problems could be caused by changes in the microbiome of the female genital tract by ascending pathogens from the vagina to parts of the upper genital tract, local microbial distortion due to haematogenous spread of infective microbes, retrograde spread from the peritoneal cavity (Schoenmakers et al., 2018) or hormonal influences ultimately leading to a dysbalanced and dysbiotic uterine environment (Haahr et al., 2016). Besides lactic acid, Lactobacilli produce bacteriocins (as mentioned above) and hydrogen peroxide (Petrova et al., 2015), which aid in inhibiting pathogens and promote a supportive environment for embryonic implantation and survival.

Women with infertility problems show a reduced number of cervical Lactobacillus (Graspeuntner et al., 2018), a lower abundance of vaginal L. iners and a higher abundance of Candida, a higher prevalence of asymptomatic bacterial vaginosis, the presence of certain specific bacteria such as Atopobium vaginae, Ureaplasma vaginae, U. parvum, U. urealyticum and Gardnerella and a lower frequency of Mycoplasmateceae species, as compared to healthy women (Costoya et al., 2012; Urszula et al., 2014; Panda et al., 2016; Babu et al., 2017; Campisciano et al., 2017; Wee et al., 2017). In infertility due to infection (Graspeuntner et al., 2018), a decrease in Lactobacillus and a higher cervical microbial diversity was detected, with a significant higher read count of Gardnerella, Prevotella, Leptotrichia amnionii and Sneathia as compared to fertile controls (Di et al., 2018; Graspeuntner et al., 2018). In line with the protective and supportive characteristics of Lactobacilli, are the recent findings that a Lactobacillus-dominated (>90% Lactobacillus spp.) endometrial microbiome profile correlates with reproductive success (Moreno et al., 2016) and that the percentages of Lactobacilli spp. in both the vagina and the endometrium of IVF patients versus non-IVF patients and healthy volunteers were significantly lower (Kyono et al., 2018). Moreover, the presence of non-Lactobacillus-dominated microbiota, especially with detection of the genera Gardnerella (family Bifidobacteriaceae) and Streptococcus (family Streptococcaceae), in the endometrial fluid seems to be associated with significant decreases in implantation, ongoing pregnancy and live birth rates (Moreno et al., 2016). Women with a live birth showed a lower species diversity index of the vaginal microbiome as compared to women with no live birth, with a vaginal composition consisting solely of Lactobacillus yielding the highest chance of success (Hyman et al., 2014). In contrast, Fotouh and Al-Inany (2008) and Franasiak et al. (2016) demonstrated that the role of the microbial composition of, respectively, the cervical canal and the endometrium during embryo transfer is limited and does not significantly affect pregnancy rates. In addition, specific endometrial microbiome profiles could be related to chronic endometritis (CE; Cicinelli et al., 2014a, b), which seems to cause a predisposition to infertility caused by endometriosis (Khan et al., 2014), repeated implantation failure (Cicinelli et al., 2014a, b) and recurrent miscarriage (Cicinelli et al., 2014a, b). In IVF patients, a negative correlation was found between abnormal vaginal microbiota or BV and clinical pregnancy rate (Mangot-Bertrand et al., 2013; Haahr et al., 2016). However, it has been shown that an increase of opportunistic pathogens in the female genital tract always correlates with decreased frequency of Lactobacillus species (Aleshkin et al., 2006) and has been associated with lower success rates of ART. However, a Cochrane review investigating the prophylactic use of antibiotics in relation to the subsequent clinical pregnancy rate found no significant influence (Kroon et al., 2012). This conclusion is based on a study (Brook et al., 2006) with 350 ART patients in whom no difference was found in clinical pregnancy rate between those receiving antibiotics prior to embryo transfer and those not (OR 1.02, 95% CI 0.66–1.58), although genital tract colonisation was significantly more likely in women who did not receive antibiotics prior to ET compared to those who did (OR 0.59, 95% CI 0.37–0.95).

Obstetrical complications

After achieving an ongoing pregnancy, the microbiome of the female reproductive tract continues to play a role (Schoenmakers et al., 2018). Due to the pregnancy, only the vagina is accessible, limiting microbiome research of the reproductive tract during pregnancy. Embryonic development and growth are largely dependent on placental function, which suggests that the recently determined placental microbiome could influence foetal and pregnancy outcome (Aagaard et al., 2012). A state of dysbiosis in the vagina, endometrium or placenta could ultimately lead to an adverse implantation and pregnancy outcome. Although the timing of preterm birth (between 24 and <37 weeks of gestation) is outside of the scope of this review, it seems to find its origin just before mid-gestation (Stout et al., 2017).

At the end of the first trimester of pregnancy, the vaginal microbiome is mostly composed of L. crispatus, L. iners, L. gasseri or L. jensenii (Kim et al., 2017b). Son et al. (2018) concludes that abnormal vaginal colonisation, with Klebsiella pneumonia as the most significant microbe, in the second trimester is associated with a significant increase in preterm delivery before 28 weeks of gestation, whereas S. agalactiae colonisation in the second trimester demonstrated a higher late miscarriage rate. In addition, abnormal vaginal colonisation detected in the second trimester was associated with a lower rate of live births compared with the group without bacterial colonisation. Dominant presence of L. iners at 16 weeks of gestation is significantly associated with early preterm birth before 34 weeks, while high abundance of L. crispatus seems to predict term birth (Kindinger et al., 2017). Women who deliver at term seem to have a stable diversity and richness in vaginal microbes, whereas in women who deliver preterm, microbial richness and diversity is decreased significantly between the first and second trimester (Stout et al., 2017). Additionally, placentas of preterm birth show a different microbiome as compared to term placentas (McElrath et al., 2008).

Semen microbiome and reproductive health

Another route of introduction of microbiota into the female reproductive tract is via semen. Recent analyses have shown the presence of a seminal microbiome (Table 2), which most likely is the combined result of each part of the male reproductive tract. NGS techniques showed that seminal bacterial communities can be clustered into three groups, dominated by either Lactobacillus, Pseudomonas or Prevotella. Importantly, 80% of the normal semen samples belonged to the Lactobacillus-dominated group (Weng et al., 2014). Not only does the semen harbour its own microbiome, the microbiota also seems to be able to attach to spermatozoa prior to ejaculation (Toth et al., 1982; Svenstrup et al., 2003; Keith and Berger, 1985) and are able to hitchhike into the female reproductive tract.

The presence of Mycoplasma spp. has been associated with low sperm concentration and abnormal sperm morphology (Gdoura et al., 2007; Zinzendorf et al., 2008). According to the data of Ahmadi et al. (2017), the frequency of Mycoplasma hominis is significantly higher in infertile compared to fertile men and moreover, antibiotic therapy improved the semen quality of infertile men. In line with the female reproductive tract, less Lactobacilli and a higher species diversity was seen in male reproductive disease (Mändar et al., 2017). An increase of Neisseria, Klebsiella and Pseudomonas and a reduction in Lactobacillus has recently been linked to seminal hyperviscosity and oligoasthenoteratozoospermia (Monteiro et al., 2018), indicating that sexual transmitted diseases (STDs) hamper not only female fertility.

Interaction between male and female reproductive tract microbiomes

Besides discussing the microbiome of the female and male reproductive tract separately, we want to stress the fact that the male and female microbiome are influenced by each other and seem to interact.

When Mandar et al. (2015) compared the seminal and vaginal microbiomes of couples, they found a high number of shared DNA sequences or phylotypes (85%). Among the shared phylotypes, the most abundant genera were Lactobacillus, Veillonella, Streptococcus, Porphyromonas and Atopobium. Although, seminal communities were more diverse, in line with above-mentioned finding, semen had lower total bacterial concentrations than vaginal communities.

The semen microbiome significantly, although temporarily, affects the vaginal microbiome (Borovkova et al., 2011). Intercourse results in vaginal alkalisation, reflected in an increase in Nugent scores accompanied by shifts in local microbiota, with Staphylococci and Streptococci being the most frequent cultured species. Importantly, in case of lower Nugent scores, these shifts occurred less, indicating a protective role of the acid pH produced by Lactobacillus dominant microbiota. However, Eschenbach et al. (2001) showed no effect on vaginal Lactobacilli and pH measurement 8–12 h after intercourse, yet significantly more E. coli were found in the vagina. A physiological post-coital, but temporary, condition has been proposed by Leppaluoto (2011), which involves the replacement of vaginal Lactobacillus by Gardnerella vaginalis through the neutralising power of the ejaculate, resulting in pH changes.

Extensive research in rodents, has shown that exposure to seminal fluid leads to cytokine signalling within the female reproductive tract altering endometrium receptivity and dynamics of preimplantation embryo development. If in humans (repeated) exposure to seminal fluid will have the same effect also need to be determined (Robertson and Sharkey, 2016).

How the interaction between the female and male reproductive microbiome influences each other is still unknown. Also, whether a temporary combined female–male microbiome occurs during the post-coital period, perhaps even persisting into the preimplantation period and aiding in a successful conception (Fig. 7), should be clarified in future research.

Figure 7

The seminal microbiome and uterine microbiome of human reproduction before and during conception.

Future perspectives in microbiome research

Comparability

This review shows that the variety of studied cohorts and the different techniques used to assess the microbiota make it difficult to compare studies. Moreover, it has been shown that bacterial communities vary among different populations and ethnicities. Last but not least, it seems that there are discernible differences between microbiota from healthy controls and women with reproductive diseases. Lack of standardisation hampers reliable comparison between different study outcomes. These differences within and between women and men, either healthy (Gajer et al., 2012) or diseased (Mehta et al., 2015), or subjected to different interventions (Liu et al. 2013a, b), make it difficult to extend these results to a wider population of infertile couples. In this review, all microbial compositions of the included studies were analysed, which has resulted in a refined impression of the microbiota present per anatomical region of the reproductive tract in healthy women and men. Still, the need to define differences between individuals is essential to link bacterial community composition to states of health and treatment outcomes. In addition to the interpersonal variability, there might also be intrapersonal variability since most studies only sampled participants once. However, studies that sampled several times over a longer period of time do report microbial changes over time. Switches between CSTs has been observed in vaginal microbiota within the same women over time (Gajer et al., 2012; Mehta et al., 2015). In future studies, the Jensen–Shannon divergence index (Abou-Moustafa, 2014) should be used in order to measure the variability, to provides a quantitative measure of community stability. Moreover, precise phenotyping of patients and controls should be done in order to reduce differences between populations.

Dynamics

After the initial microbiome studies allowed definition of the microbial composition, nowadays research is focussing on factors that may impact on the composition of the microbiome, such as the effect of circumcision (Liu et al. 2013a, b) and the menstrual cycle (Johnson et al., 1985). Longitudinal studies have demonstrated that the vaginal microbiota can be dynamic. Dramatic shifts in bacterial composition and concentration have been observed in response to numerous endogenous and exogenous factors. It seems that the reproductive microbiome is responsive to sexual intercourse, a prior history of BV, a greater number of sexual partners and a greater number of recent episodes of receptive oral sex (Schwebke et al., 1999).

The fact that the microbiome in general, including that of the vagina, is dynamic, provides new opportunities to adjust a disbalanced composition or simply await a return to a balanced composition in order to improve reproductive outcomes.

Function

Understanding which microbial species or community types represent potentially dysbiotic states, and whether manipulation to reduce risk is feasible or effective, first requires a better understanding of the composition and their function. The ultimate goal is determining the function of the female and male reproductive tract microbiota, and which factors can change it for the better. Nonetheless, fluctuations in community composition through time or endogenous and exogenous factors does not necessarily lead to a change in community performance. Fluctuations could occur while maintaining community performance due to perseverance of the metabolic state, when there is functional redundancy among community members and when shifts in the relative abundances of species occur due to changes in environmental conditions that favour one population over another. Future research into the composition of the metabolome will possibly reveal this complex interaction between species and host-microbe interactions.

Applicability

Research into the reproductive tract microbiome has focused on identification of specific bacterial species or microbiome compositions that might impact on clinical outcomes. In this review, we described studies that investigated the role of the microbiome in female health, fertility, infertility, ART and obstetrical outcomes.

Standardisation remains a critical issue hampering implementation of microbiome analysis in clinical practice. Each research group should thoroughly describe the whole process of their microbiome assessment, including used protocols, platforms used, sequenced regions and databases used, including data analysis pipelines, which will probably accelerate the implementation of results into daily clinical practice.

Given the reported associations between the microbiome of the female and male reproductive tract and clinical outcomes in the field of reproductive health, clinical analysis of the microbiome could become a tool for possible risk assessment and therapeutics in the future. Routine screening of the vaginal microbiome, as a proxy for the reproductive outcome, in all women undergoing fertility treatments, might become an option.

Acknowledgements

We would like to thank D.A.J. Hilster for creating the figures.

Authors’ roles

R.K., S.M., A.B., J.L. and S.S. were involved in the design, execution and analysis of this review. All authors contributed to the drafting of the manuscript. R.K., S.M., D.F., S.S. and J.L. were responsible for the final editing, A.B., C.B., J.L. and S.S. reviewed and edited the manuscript. All authors approved the final version of the article.

Funding

S.M was supported by the Research Fund of Flanders (Fonds Wetenschappelijk Onderzoek (FWO), Flanders, Belgium, 11M9415N). Otherwise, no specific funding was sought for the study, and departmental funds were used to support the authors throughout the study period and manuscript preparation: Department of Obstetrics and Gynaecology, Erasmus University Medical Centre, Rotterdam, The Netherlands; Department of Medical Microbiology and Infection Control, Vrije Universiteit (VU) University Medical Centre, Amsterdam, The Netherlands; and Centre for Reproductive Medicine, Universitair Ziekenhuis Brussel, Vrije Universiteit Brussel, Brussels, Belgium.

Conflict of interest

A.B. reports funding from IS-Diagnostics Ltd, outside the submitted work, and has a patent 392EPP0 pending. C.B. reports honoraria and/or research grants from MSD, Ferring, Merck, Abbott and Besins. J.L. reports grants from Dutch Heart Foundation, Ferring, Metagenics Inc. J.L. has also received personal consultancy fees from ARTPred B.V., Danone, Euroscreen and Roche, during the conduct of the study. In addition, J.L. is a co-applicant on a Erasmus MC patent, that predicts IVF outcome based on the urinary microbiome. This particular patent is licensed to ARTPred B.V. The other authors declare that they have no conflict of interest.