Interpretation
Since PCOS patients are characterized by irregular menstrual cycles and
sex hormone disorders, while many studies have indicated that the
diversity of vaginal bacterial species increases along with the decrease
in the content of Lactobacillus 15, 16,
24. We also found a significant decrease in the content ofLactobacillus and was significantly enriched in potential
pathogens, such as U. parvum, G. vaginalis, P. buccalis ,P. timonensis, and A. baumanni, in PCOS patients.
After lactobacilli spp. dominance was destroyed, G.vaginalis become the dominant bacterium in the vaginal
biome15, 16, 24. G. vaginalis is closely
associated with the pathogenesis of bacterial
vaginosis26 and in combination with U.parvum further increases the risk of spontaneous preterm
birth27.
Further, the abundance of P. buccalis and P.timonesis were inversely associated with Lactobacillusabundance, which maintains vaginal homeostasis28,
29. P . timonensis abundance was associated with mucosal
inflammation30 and cervical intraepithelial
neoplasia31. Additionally, A. baumanniican cause preterm delivery and chorioamnionitis during
pregnancy32, as well as infection in vaginally
delivered infants33. In conclusion, the reduction inLactobacillus abundance in PCOS patients leads to changes in
vaginal pH and destruction of the immune barrier34,
making it easier for potentially vaginal commensal anaerobic bacteria,
such as G. vaginalis, U. parvum, Prevotella spp., A.
baumannii to colonize and
reproduce.
These potential vaginal pathogens mentioned above cause disease at a
greater frequency and severity and can potentially result in BV,
miscarriage, preterm birth, pre-cancerous, and cancerous cervical
lesions8, 35-38. In addition, B. breve was a
subdominant group of the vaginal microbiota in the VMB of healthy
indivuals39, 40 but its abundance was depleted in the
PCOS group. Giordani et al.39 found that the oral
administration of B. breve as a probiotic bacteria can prevent
urogenital infections. Since PCOS is a common cause of infertility and
increases the risk of abortion, fetal arrest, and preterm
birth15, we suspect that there is a correlation
between altered compositions of the VMB and adverse reproductive
outcomes in PCOS patients. Therefore, the adverse pregnancy prognosis of
PCOS patients may not be related only to the disease itself but also to
their vaginal microbiological status, which should not be neglected
either. Modifying the composition of vaginal microorganisms may be a
novel method of clinical treatment for clinicians to improve the
clinical pregnancy outcome in patients with PCOS. Nevertheless, we
cannot ignore the role of endocrine disorders in PCOS, as they may play
a role in impacting the diversity and composition of the vaginal
microbiota. There may be a mutual causal relationship between PCOS and
vaginal microbiota.
Intriguingly, the within-group variability was significantly higher in
the PCOS group than in the control group. PCOS phenotype A patients
showed slightly lower vaginal microbiota diversity compared with PCOS
phenotype D patients. Serum levels of testosterone, LH, and AMH in PCOS
may contribute to this discrepancy in findings. First, a high level of
testosterone is associated with an elevated abundance ofLactobacillus 24. Some researchers have also
demonstrated that testosterone can alter the gut microbiota and reduce
alpha-diversity7, 8, 41. Second, this study indicates
that AMH is the biochemical marker that is most closely with the vaginal
microbiota of PCOS women. Third, serum levels of T, LH and AMH were
positively correlated with U. parvum, A. baumanniiabundances, but negatively correlated with the abundance of P.
buccalis , suggesting that the abundance of vaginal microflora was
affected. Therefore, a greater degree of variation was observed in PCOS
samples but the exact mechanism involved needs to be further
investigated.
We performed the validation of a PCOS prediction model (AUC, 0.8). We
conducted in-group verification and obtained satisfactory results. In
future, we will need to use independent samples to verify our model. Our
model was able to accurately distinguish PCOS patients from healthy
controls. The dysbiotic vaginal microbiome captured by the classifier
offered further evidence for the identification of PCOS-associated
microbial composition and indicated that those bacteria were highly
correlated with PCOS.
Furthermore, we explored microbiome community interactions between the
two groups. In the VMB network,
module 1 and module 2 showed thatL.
crispatus and L. iners were clustered together, and indicated a
negative correlation between L. crispatus and L. iners.The results were consistent with the results of many previous
studies42, 43. Module 3 clustered potential pathogens,
including G. vaginalis , P. bivia, and A. vaginae .A. vaginae and P. bivia can be incorporated as they
influence gene expression in G. vaginalisbiofilms44. Gardnerella spp . can produce
metabolite amino acids that are used by Prevotella spp. as
nutrients to produce ammonia, which in turn is used byGardnerella spp45 . Indeed, bacterial
genera clustered in modules had similar requirements different vaginal
micro environments. Next, the network analyses revealed that the unique
connectors of the PCOS group wereL.iners , A.christensenii ,A. vaginae, P.timonensis, A. omnicolens, and Sneath sanguinegens , which
have been reported as potential pathogen involved in BV, miscarriage,
preterm birth, pre-cancerous, and cancerous cervical
lesions8, 35-38. We further found thatL.crispus , P. timonensis, and P. buccalis can be identified
as key bacteria that drive changes in the vaginal microbial interaction
network in PCOS patients. These key bacteria are potential threats for
the development of PCOS and we can facilitate the conversion of the
vaginal microbial interactions network back to normal by restoring
normal levels of these bacteria, by using techniques such as vaginal
microflora transplantation (VMT) 46and
Prebiotics40, which provide a natural protection by
restoring vaginal homeostasis.
Finally, we predicted the function of high-abundance vaginal bacteria,
while the metabolic pathways involved and microbiome influence to the
host were investigated in our study. In addition, a higher concentration
of activated peptide/nickel transport system permease proteins that are
associated with quorum sensing (QS), which is a regulatory system that
allows bacteria to share information on cell density and can adjust gene
expression accordingly47, was found in the control
group. Mixed species of bacteria communities can influence and be
influenced by the activities of QS or other neighboring
species48. PCOS showed a low expression of QS or a
lack of intercommunity regulation. These differential functional
pathways further suggest a possible pathogenic mechanism exerted by
vaginal flora in PCOS patients.