The profiles of vaginal microbiota in fertile and infertile Thai women.

This cross-sectional study was conducted from August 2019 to February 2020. We recruited 30 fertile women at the Family Planning Clinic and 62 infertile women at the Infertility Clinic, Department of Obstetrics and Gynecology, King Chulalongkorn Memorial Hospital, Bangkok, Thailand. All 92 participants were Thai, born in Thailand, aged 18–45 years, and without significant underlying diseases. They were instructed to avoid sexual activity for at least seven days before sample collection. We excluded individuals who had pelvic inflammatory disease (PID), current BV, or sexually transmitted diseases. BV was excluded based on clinical history and gynecological examination at the time of participant recruitment. Women who had used immunosuppressive drugs, vaginal douching, or oral antibiotics within the past month were also excluded from the study. Additionally, fertile women were require to have regular menstrual cycles, have a live birth within past one year, and not currently use intrauterine device and hormonal contraception. Infertility was defined by the failure to achieve a pregnancy after 12 months or more of regular unprotected sexual intercourse. Among the infertile group, we excluded all tubal and male infertility. The minimum sample size of 90 participants (30 fertile and 60 infertile women) was determined according to the study by Gluseppina Campisciano, et al. [ 20 ]. Questionnaire interviews and physical examinations were conducted among all eligible participants by well-trained research assistants in a private and comfortable environment. Personal characteristics including age (years), date of last menstrual period, menstruation interval (days), number of live births, frequency of sexual intercourse (time/week), causes for infertility, and history of antibiotics use in past 6 months (yes vs. no) were recorded. Next, body weight (kg) and height (cm) were measured for each woman; and the body mass index (BMI) was calculated and recorded. Finally, each woman lay in the lithotomy position with an emptied bladder. A vaginal specimen was collected using a cotton swab from the mid-vaginal wall using a cotton swab and immediately placed into a 2 mL of transport medium. The medium consisted of Hank’s balanced salt solution (Gibco, ThermoFisher Scientific, USA), 1% (w/v) of bovine serum albumin (Gibco, ThermoFisher Scientific, USA), 15 µg/mL of amphotericin B (Supelco, Merck, Germany), 50 µg/mL of streptomycin (Calbiochem, Merck, Germany) and 100U/mL of penicillin G (Calbiochem, Merck, Germany) in a 15 ml collection tube. All collected specimens were sent to the laboratory within 4 h after collection. All vaginal samples were collected during the predicted periovulation period according to last menstrual period and interval of menstruation. Among the 62 infertile participants, 42 women were included in the infertile (pre-ovulation induction) group during natural cycles. Twenty women underwent ovarian induction and were included in the infertile (post-ovulation induction) group, of whom 19 received clomiphene citrate and 1 received letrozole. One sample failed sequencing quality control and was excluded from the final microbiome analyses. The infertile (pre-ovulation induction) and infertile (post-ovulation induction) groups consisted of different participants rather than repeated measurements collected from the same individuals. On the predicted date, urine sample was collected for testing the luteinizing hormone (LH) by LH ovulation test (Check Tru ® ). If women received positive result from LH test, we provided them pelvic examination and collected vaginal sample. In case with LH negative, participants were re-evaluated during the subsequent menstrual cycle. The participants underwent a transvaginal ultrasound on day 12 of their menstrual cycle to measure the size of dominant follicle. Once at least one folliclereached a mean diameter of 18 mm, vaginal samples were collected. If the dominant follicle did not meet the size requirement, we followed the participant until the dominant follicle met the size requirement. Participants received ovarian induction with 50–100 mg of clomiphene citrate daily (Ovinum ® ; BIOLAB, Thailand) or 5 mg letrozole (Letov ® ; Zydus Cadila, India) on days 3–5 of the menstrual cycle. On day 12 of the menstrual cycle, participants underwent transvaginal ultrasound to assess the size of dominant follicle. When at least one follicle, but no more than three follicles, reached a mean diameter of 18 mm, final oocyte maturation was triggered by subcutaneous injection of 0.25 mg of recombinant hCG (Ovidrel ® ; Merck Serono, Germany). Vaginal samples were collected prior to intrauterine insemination (IUI), which was performed 36–40 h later. The remaining procedures followed the same protocol as in the pre-ovulation induction group. Two milliliters of vaginal suspension were centrifuged at 2,000 rpm at room temperature for 5 min to spin down the debris and microbial cells. The supernatant was removed, and then DNA was extracted by GenUP™ gDNA extraction kit (Biotechrabbit, Germany) following the recommended protocol. Finally, the extracted DNA was reconstituted with 30 µl and stored at -20 °C until analysis. The 16S rDNA is the universal barcoding region for identifying bacteria. At this step, amplification of bacterial 16S rDNA (V4 region) was performed using Polymerase Chain Reaction (PCR). The phasing primer set contained phasing sequences, TruSeq adaptor, and targeted primer sequences: 515F:5’-GTGCCAGCMGCCGCGGTAA-3’ and 806R: 5’- GGACTACHVGGGTWTCTAAT-3’. Each of the 20 µl PCR reactions consisted of 20 ng of DNA template, 0.2 µM of each primer, 0.2 mM of dNTPs, 1x Phusion green HF Buffer, and 0.4 U of Phusion DNA Polymerase (2U/µl) (Thermo Scientific, USA). The amplification program started at 98 °C for 30 s to denature the DNA, followed by 25 cycles of 98 °C for 10 s, 53 °C for 25 s, 72 °C for 25 s and ended with one cycle of extension at 72 °C for 10 min. The genes from the amplified PCR product (size approximately 380 bp) were separated by electrophoresis using 2% agarose gel. The amplified products from the first round of PCR have been amplified again using another primer set containing Illumina sequencing primers, multiplexing indexes, and Illumina adaptors. After amplification, the PCR products from this step were incorporated with Illumina sequencing adaptors and various indexes. Then, the genes from the PCR products were separated by electrophoresis using 2% agarose gel. The desired gene/band was cut out of the gel and purified by QIAquick Gel Extraction Kit (QIAGEN, Germany). The DNA was quantified. Finally, DNA concentrations of more than 20 ng/µl were measured again by qPCR using KAPA library quantification kits for Illumina platforms (Kapa Biosystems, USA) and pooled with an equal amount to make the final concentration of the library 2 nM. The sequencing was performed by using the MiSeq v2 reagent kit on the MiSeq platform (Illumina, USA). Loading concentration of the library was 6 pM with 20% of spike-in PhiX. The library was sequenced for pair-end 2 × 250 cycles. Demultiplexing sequencing data (FASTQ files) were generated by MiSeq Reporter software (version 2.6.2.3). Raw data were processed and analyzed by the QIIME2 pipeline (version 2018.8). Briefly, paired-end reads were joined and trimmed based on the quality score. After that, merged reads were deduplicated and clustered with 97% similarity by VSEARCH. Chimeric sequences were also filtered out by the UCHIME algorithm. The filtered reads were identified based on 99% OTUs clustered 16 S Greengene database (2013.8) using the VSEARCH algorithm. The relative abundance (RA) of taxa was normalized with the total reads of each sample. Data analysis and comparison, including calculation of alpha diversity, beta diversity, rarefaction curve, and statistical analysis, were evaluated using the plugins available in the QIIME2 software. After the microbiota was processed, the relative abundance of bacterial taxa was obtained, and compared between groups as the outcomes. Moreover, taxa differentially represented between groups were identified using linear discriminant analysis effect size (LEfSe) algorithms with an LDA score threshold of > 2.0. LEfSe analyses were performed without additional multiple testing correction. Variables with normal distribution were summarized as mean ± standard deviation (SD), whereas non-normally distributed variables were presented as median (interquartile range, IQR). Comparisons between two groups were performed using Student’s t-test for parametric data and Mann–Whitney U test for non-parametric data. Kruskal–Wallis test was used for multiple-group comparisons. Relative abundance (%) and bacterial read counts were summarized as appropriate. Alpha rarefaction curves were generated to confirm adequate sequencing depth. Differential abundance was analyzed using LEfSe. Alpha and beta diversity metrics were compared across groups. A two-sided P  < 0.05 was considered statistically significant. All analyses were performed using SPSS version 22 (IBM Corporation ® ).

The mean age was 36.4 years in the infertile group and 34.4 years in the fertile group (Table  1 ). Mean duration of infertility was 27.2 months. Approximately 72.6% (45/62) of the infertile women were classsified as having primary infertility and 27.4% (17/62) with secondary infertility. More than half of infertility cases remained unexplained, accounting for 53.2%. Table 1 Baseline characteristics ( N  = 92) Infertile ( N  = 62) Control ( N  = 30) Age (year) 36.4 ± 3.6 34.4 ± 5.7 BMI (kg/m 2 ) 22.1 ± 3.6 25.2 ± 5.8 Menstrual cycle interval (days) 30.2 ± 5.2 29.4 ± 3.2 Number of live births 0 (0–0) 2 (1–2) Frequency of SI (time/week) 3 (2–3) 2 (2–3) Contraception  - No 7 (23.3)  - Condom 8 (26.7)  - Tubal sterilization 15 (50.0) Type of infertility  - Primary 45 (72.6)  - Secondary 17 (27.4) Causes of infertility  - Anovulation 13 (21.0)  - Endometriosis 16 (25.8)  - Unexplained 33 (53.2) History of ATB use in 6 months 20 (32.3) 14 (46.7) Data presented as mean ± SD or median (interquartile range) or number (%). BMI body mass index, SI sexual intercourse, ATB antibiotic Baseline characteristics ( N  = 92) Data presented as mean ± SD or median (interquartile range) or number (%). BMI body mass index, SI sexual intercourse, ATB antibiotic The 16 S rRNA gene sequencing was successfully performed and sequenced to ascertain the composition of the vaginal microbiota among 91 women (30 fertile and 61 infertile women). Alpha rarefaction curves suggested slightly higher microbial richness in fertile women (Fig.  1 ). Relative abundance of vaginal microbiota at the phylum and genus levels across study groups is shown in Fig.  2 . The most abundant bacterial phyla of all groups were Firmicutes, Actinobacteria, Bacteroidetes and Fusobacteria (Fig.  2 A). Genus-level analysis demonstrated that Lactobacillus was the predominant genus, comprising more than 60% of the vaginal microbiota across all groups (Fig.  2 B). Detailed view of the microbial diversity was shown in Fig.  3 . Fig. 1 Rarefaction curves of vaginal sample in three groups including fertile, infertile (pre-ovulation induction) and infertile (post-ovulation induction) groups Rarefaction curves of vaginal sample in three groups including fertile, infertile (pre-ovulation induction) and infertile (post-ovulation induction) groups Fig. 2 Comparison of the microbiome in each group. Taxonomic classification at the phylum ( A ) and taxa ( B ) levels were based on the relative abundance (RA). RA < 1% of the microbiota were labeled together as ‘others’ Comparison of the microbiome in each group. Taxonomic classification at the phylum ( A ) and taxa ( B ) levels were based on the relative abundance (RA). RA < 1% of the microbiota were labeled together as ‘others’ Fig. 3 Relative abundance of vaginal microbiota across study groups. A Relative abundance of dominant bacterial phyla. B Relative abundance of dominant bacterial genera. Samples are grouped into fertile women, infertile women before ovulation induction, and infertile women after ovulation induction. Each vertical bar represents an individual participant sample Relative abundance of vaginal microbiota across study groups. A Relative abundance of dominant bacterial phyla. B Relative abundance of dominant bacterial genera. Samples are grouped into fertile women, infertile women before ovulation induction, and infertile women after ovulation induction. Each vertical bar represents an individual participant sample Differential abundance analysis using LEfSe is presented in the taxonomic cladograms shown in Fig.  4 . Several taxa appeared to be relatively enriched in the fertile group compared with the infertile (pre-ovulation induction) group (Fig.  4 A). Figure  4 B presents the taxonomic cladogram comparing infertile women before and after ovulation induction. Genera including Clostridium , Shuttleworthia , Dialister and Ureaplasma appeared to be relatively enriched in the infertile (pre-ovulation induction) group compared with the infertile (post-ovulation induction) group. Fig. 4 Taxonomic cladograms generated from LEfSe analysis showing differentially enriched bacterial taxa between study groups. A Comparison between fertile women and infertile women before ovulation induction. B Comparison between infertile women before and after ovulation induction. Colored nodes represent taxa with differential abundance identified by LEfSe analysis, whereas yellow nodes indicate taxa without significant differences between groups. Node size reflects relative abundance of taxa Taxonomic cladograms generated from LEfSe analysis showing differentially enriched bacterial taxa between study groups. A Comparison between fertile women and infertile women before ovulation induction. B Comparison between infertile women before and after ovulation induction. Colored nodes represent taxa with differential abundance identified by LEfSe analysis, whereas yellow nodes indicate taxa without significant differences between groups. Node size reflects relative abundance of taxa Microbial diversity analysis was performed using the Chao1, Shannon, and Simpson indices (Fig.  5 ). No significant differences were observed between fertile women and infertile women (pre-ovulation induction) (Fig.  5 A–C), or between infertile women before and after ovulation induction (Fig.  5 D–F). Similarly, beta diversity analysis demonstrated no significant clustering differences between groups (Fig.  6 ). Fig. 5 The alpha diversity graph shows the richness and evenness of the vaginal microbiota using Chao1, Shannon and Simpson diversity index, respectively. Alpha diversity of vaginal microbiome in a fertile group compares to infertile group (pre-ovulation induction) indicating A-C . At the same time, figures D-F show the alpha diversity of the infertile group (pre-ovulation induction) versus the infertile group (post-ovulation induction). Abbreviation “ns” represents not significantly difference between groups The alpha diversity graph shows the richness and evenness of the vaginal microbiota using Chao1, Shannon and Simpson diversity index, respectively. Alpha diversity of vaginal microbiome in a fertile group compares to infertile group (pre-ovulation induction) indicating A-C . At the same time, figures D-F show the alpha diversity of the infertile group (pre-ovulation induction) versus the infertile group (post-ovulation induction). Abbreviation “ns” represents not significantly difference between groups Fig. 6 Beta diversity analysis of vaginal microbiota across study groups using principal coordinate analysis (PCoA). Each point represents an individual participant sample. A Comparison between fertile women and infertile women (pre-ovulation induction). B Comparison between infertile women (pre-ovulation induction) and infertile women (post-ovulation induction). No significant clustering differences were observed between groups Beta diversity analysis of vaginal microbiota across study groups using principal coordinate analysis (PCoA). Each point represents an individual participant sample. A Comparison between fertile women and infertile women (pre-ovulation induction). B Comparison between infertile women (pre-ovulation induction) and infertile women (post-ovulation induction). No significant clustering differences were observed between groups

In conclusion, this study provides preliminary insights into vaginal microbiota profiles among reproductive-aged Thai women. The overall vaginal microbiota composition was largely similar across fertile and infertile women, including women with different ovarian induction status. Although no significant differences were observed in overall microbial diversity, exploratory differential abundance analyses identified several taxa that may warrant further investigation. These findings contribute to the growing understanding of vaginal microbiota in reproductive health among Thai women. Future larger-scale prospective longitudinal studies are warranted to better understand the potential relationship between vaginal microbiota, vaginal microbiota variation, and reproductive health outcome, including microbiota profiles according to specific infertility etiologies.

Our results provide preliminary insights into the vaginal microbiota composition of reproductive-aged Thai women. Firmicutes, Actinobacteria, Bacteroidetes, and Fusobacteria were the dominant bacterial phyla across all groups, while Lactobacillus was the predominant bacterial genus. Alpha and beta diversity analyses did not demonstrate significant differences between fertile, infertile (pre-ovulation induction), and infertile (post-ovulation induction) groups. However, differential abundance analysis identified several taxa enriched in specific groups. In addition, marginally higher microbial richness was observed among fertile women compared with infertile women. The role of the microbiota in reproductive health has attracted considerable interest among clinical and scientific researchers in investigating the physiological and pathological roles of the microbiota in the reproductive systems [ 9 ]. A general opinion is that Lactobacillus is the dominant bacterial genus in the healthy vaginal environment among reproductive-aged women, playing a potential protective role by creating the acidic environment [ 21 ]. Consistent with previous studies, Lactobacillus iners was the most common species among Thai reproductive-aged women [ 22 ]. Previous studies have reported associations between Lactobacillus -dominant vaginal microbiota and favorable reproductive outcomes [ 23 – 25 ]. In our study, L. iners appeared more abundant in fertile women than in infertile women; however, these differences were not statistically significant (Supplementary Figure). Nevertheless, a similar trend has been reported in a previous study, in which a higher relative abundance of L. iners was observed among Indian women who had full-term deliveries compared with infertile women [ 26 ]. We also explored vaginal microbiota profiles according to ovarian induction status in pre-ovulation induction and post-ovulation induction women. In our study, exploratory taxa-level differences were identified between the pre-ovulation induction and post-ovulation induction groups. Similarly, a Chinese study also reported taxa-level differences identified by LEfSe analysis between women before and after ovarian stimulation [ 27 ]. The vaginal microbiota was positively correlated with the level of hormones, especially estrogen [ 15 , 28 ]. Women with anovulatory cycles had low sex hormones such as estrogen, progesterone, and luteinizing hormone during ovulation. The level of luteinizing hormone was low even at the peak of the ovulation period [ 29 ]. During ovarian induction cycles, estrogen levels increase, which may promote the relative abundance of Lactobacillus . Almost all the participants used clomiphene citrate to induce ovulation for the infertile group (post-ovulation induction). Clomiphene citrate is a selective estrogen receptor modulator that can act as an estrogen agonist and antagonist depending on the target tissue [ 30 ]. Clomiphene citrate has a negative effect on the cervical mucus and can increase vaginal dryness [ 29 , 31 , 32 ]. This effect may alter the vaginal environment and subsequently influence the baseline vaginal ecosystem. Alpha and beta diversity analysis showed no significant differences across groups, suggesting relatively similar microbial diversity profiles across groups. In contrast, the difference in diversity of the vaginal microbiome was observed in Indian and Chinese cohorts [ 26 , 27 ]. Many factors have been related to richness and diversity of vaginal microbiota, such as age, ethnicity, lifestyle, menstruation, vaginal hygiene practices, sexual activity, antibiotic use, hormonal status, pregnancy, and contraceptive methods [ 15 , 16 , 28 , 33 , 34 ]. Differences between Thai women and Indian or Chinese women may result from the difference in genetics, diet, cultural habits, and environmental factors. Differences in study design may also have contributed to the differences between our findings with previous reports, including sample selection, data collection methods, and criteria for recruiting participants. Although alterations in vaginal microbiota have been proposed to influence reproductive outcomes, existing clinical evidence remains inconsistent across different reproductive and obstetric populations [ 35 , 36 ]. Therefore, further prospective studies integrating microbiota findings with detailed clinical phenotypes and reproductive outcomes are required to better clarify the clinical significance of vaginal microbiota in women’s reproductive health. In addition to reproductive outcomes, vaginal microbiota imbalance may also have broader implications for women’s quality of life and psychosocial well-being. Vaginal dysbiosis has been associated with symptoms such as abnormal discharge, discomfort, and recurrent genital infections, and psychosexual burden, which may negatively affect sexual well-being and psychological stress [ 37 – 39 ]. Among infertile women, concerns related to reproductive health conditions may further contribute to emotional distress and reduced quality of life. Therefore, future studies integrating microbiome, clinical, and psychosocial outcomes may provide a more comprehensive understanding of women’s reproductive health. This study has several strengths. First, to our knowledge, this is among the few studies investigating vaginal microbiota profiles in Thai reproductive-aged women, providing population-specific data from an underrepresented Southeast Asian population. Second, vaginal samples were collected during the periovulatory period to reduce variability related to menstrual cycle fluctuations. Third, the study explored vaginal microbiota profiles not only between fertile and infertile women but also according to ovarian induction status, which remains relatively limited in the current literature. Finally, standardized 16 S rRNA sequencing and established microbiome analytical pipelines were applied to characterize vaginal microbial composition across study groups. Several limitations should be acknowledged in the present study. First, the cross-sectional design and the use of different participants in the pre- and post-ovulation induction groups limited causal and temporal interpretation of microbiota changes following ovarian induction. Second, defining the fertile group based on recent live birth history may not have fully reflected current fertility status and may also have introduced biological heterogeneity related to postpartum physiological changes and reproductive history. Then, the relatively small sample size, particularly in the post-ovulation induction subgroup, together with the heterogeneous ovarian induction regimens, limited statistical power and interpretation of subgroup findings. Therefore, subgroup findings should be considered exploratory. Forth, LEfSe findings should also be interpreted cautiously because of the potential for false discovery and the absence of additional multiple testing correction. In addition, OTU-based clustering and the Greengenes (2013.8) reference database are no longer considered the most up-to-date microbiota analytical approaches. Fifth, important confounding factors influencing vaginal microbiota composition, including age, hormonal variations, sexual behavior, hygiene practices, prior antibiotic exposure, contraceptive history, diet, and reproductive history, were not fully controlled or adjusted for. Sixth, hormone levels, including estrogen, progesterone, and LH, were also not measured, limiting interpretation regarding endocrine influences on vaginal microbiota composition. In addition, the infertile group included heterogeneous infertility etiologies, including unexplained infertility, ovulatory disorders, and endometriosis, which may independently influence vaginal microbiota composition. Eighth, Vaginal pH was not measured, bacterial vaginosis was not assessed using standardized diagnostic criteria, and sample collection timing may not have been fully standardized across groups. In addition, multivariable analyses were not performed because of the relatively limited sample size, which may have further limited control of potential confounding factors. Finally, the present study did not include longitudinal reproductive outcomes, such as pregnancy or in vitro fertilization outcomes, limiting the clinical interpretation of the microbiota findings.

Human microbiota inhabit various anatomical sites within and on the surface of the human body and encompass viruses, yeasts, fungi, protozoa, and bacteria [ 1 ]. Microbial composition and function vary greatly across anatomical regions and among individuals, influenced by both genetic and environmental factors, such as ethnicity, early microbial exposure, lifestyle factors, and diet [ 2 , 3 ]. The microbiome, comprising microbiota and their genomes, plays an important role in human health and disease [ 4 ]. Next-generation sequencing-based approaches have simplified the identification of human microbiota at various body sites, substantially improving the understanding of human health and disease [ 5 ]. The metagenomic studies in the Human Microbiome Project revealed that approximately 9% of the total human microbiota reside in the urogenital tract [ 6 ]. Particularly, approximately 10 10 -10 11 bacteria densely populate the female reproductive tract microbiome [ 7 ]. These bacteria play critical roles in various host reproductive processes by regulating physiological functions and influencing the onset and progression of health conditions, including gametogenesis, fertilization, pregnancy establishment and maintenance, and the microbial colonization of the fetus and/or newborn [ 4 , 8 – 10 ]. Infertility, as defined by the World Health Organization, is “a disease of the male or female reproductive system defined by the failure to achieve a pregnancy after 12 months or more of regular unprotected sexual intercourse”, affecting roughly 1 in 6 adults worldwide [ 11 , 12 ]. Increasing evidence suggests that vaginal microbiota composition may play a role in female reproductive health and infertility. Bacterial vaginosis (BV) has been associated with tubal factor infertility. Alterations in the lower genital tract microbiota are more prevalent in infertile patients [ 13 ]. Lactobacillus is the predominant bacterial genus in the healthy vaginal environment [ 14 ]. Their abundance is positively correlated with estrogen levels and changes during pregnancy and up to one year postpartum [ 15 , 16 ]. Lactobacillus helps maintain an acidic vaginal environment through glycogen metabolism, thereby promoting protection against pathogenic infections [ 17 , 18 ]. Moreover, a Lactobacillus -rich vaginal environment could modulate immune responses against sperm [ 19 ]. Therefore, understanding the potential role of the vaginal microbiota in reproductive health and pregnancy is important. However, the difference in vaginal microbiota profiles between Thai fertile and infertile women remains unclear. This study aimed to compare vaginal microbiota profiles according to fertility and ovarian induction status among Thai women. The findings from this exploratory study may contribute to future research on the potential role of vaginal microbiota in reproductive health and infertility among Thai women.

Supplementary Material 1: Supplementary Figure 1. Comparison of the relative abundance of Lactobacillus iners across fertile, infertile (pre-ovulation induction), and infertile (post-ovulation induction) groups. No statistically significant differences were observed between groups (all P > 0.05). Supplementary Material 1: Supplementary Figure 1. Comparison of the relative abundance of Lactobacillus iners across fertile, infertile (pre-ovulation induction), and infertile (post-ovulation induction) groups. No statistically significant differences were observed between groups (all P > 0.05).