Abstract

Abstract

This study primarily employs Mendelian randomization analysis to establish causal evidence linking gut microbiota to polycystic ovary syndrome (PCOS). It will provide better insights for exploring how targeted modulation of specific gut microbes for the treatment of PCOS.

Open in a separate window

1. INTRODUCTION

Polycystic ovary syndrome (PCOS) is a complex and heterogeneous disorder characterized by the presence of hyperandrogenism, ovulatory dysfunction, and polycystic ovarian morphologic features. ¹ With a global prevalence ranging from 6% to 15% among women of reproductive age, ² the disease is widely acknowledged to be strongly associated with various diseases, such as infertility, insulin resistance (IR), type 2 diabetes mellitus, and cardiovascular disease, ^{3 , 4} thus posing a substantial economic burden worldwide. Fortunately, the focus on how to comprehensively understand the impact of PCOS on women's health has been escalating continuously. So far, however, the precise cause and pathogenesis of PCOS remain incompletely understood due to the intricate etiology and heterogeneous nature of the disease. Although accumulating evidence pointed towards PCOS as a complex polygenic disorder with prominent influences by epigenetic modifications and environmental factors, for instance, dietary choices and lifestyle, ⁵ yet none of these factors can entirely elucidate the wide range of clinical manifestations associated with PCOS despite their contributions.

Over the past decade, part evidence has underscored the significant involvement of the gut microbiome in the pathogenesis of PCOS. ⁶ Previous studies have highlighted alterations in the relative abundance of specific bacterial taxa, particularly for those belonging to the Bacteroidetes and Firmicutes phyla, ⁷ that may be relevant to PCOS. For example, Ji‐Hee et al. observed a significant increase in Ruminococcus levels in neonatally androgenized rats, positively correlating with serum testosterone levels. ⁸ Another study revealed a pronounced increase in beta diversity of Bacteroides vulgatus in PCOS patients compared to healthy controls. ⁹ Furthermore, investigations into the role of probiotics and prebiotics have shown potential in the prevention and treatment of PCOS. ^{10 , 11} Nonetheless, despite these observed links between the gut microbiota and PCOS, inconsistencies in findings persist, and the current understanding of the precise community structure and functional aspects of the gut microbiome in women with PCOS remains limited. A typical example is that, unlike previous studies, Lüll K et al. found no significant differences in gut microbial profiles between PCOS and non‐PCOS women of similar BMI in the late fertile age group. ¹²

Although results from observational studies have provided initial evidence of potential links between gut microbiota and PCOS, it is imperative to acknowledge the inherent limitations of these studies. Most previous studies were designed as case–control studies, making it challenging to establish the timing of exposure and outcome definitively. In addition, observational studies are susceptible to confounding factors such as age, environmental influences, phenotypic patterns, and lifestyle, which are difficult to effectively control for. Hence, residual confounding and reverse causation further limited the understanding of the intricate array of factors contributing to PCOS development. Thus, the causal relationship between the observed gut microbiota and the risk of PCOS remains unconfirmed. Additionally, there is a scarcity of studies that comprehensively explore the interaction between the intact gut microbiota and PCOS, as well as studies that establish explicitly causal relationships.

Mendelian randomization (MR) analysis is a powerful methodology that employs genetic variation as instrumental variables (IVs) to ascertain the causal effects of risk factors on outcomes. ¹³ By controlling for confounding factors, reverse causation, and various biases, MR analysis enhances the validity of causal inference in the exposure‐outcome association. Through the random allocation of exposure‐related genetic variant alleles during reproduction, independent of environmental influences and disease progression, MR analysis minimizes the likelihood of confounding and reverse causality. ¹⁴ Therefore, in the present study, using the genome‐wide association study (GWAS) summary statistics from the MiBioGen consortiums and Integrative Epidemiology Unit (IEU) Open GWAS datasets, a bidirectional MR analysis was conducted to evaluate and establish a causal association between gut microbiota and the risk of PCOS, further providing a clear understanding of the relationship between the two from a genetic perspective.

2. MATERIAL AND METHODS

2.1. Study design

The present study utilized a two‐sample MR analysis to investigate the causal association between gut microbiota and PCOS. Additionally, a reverse MR analysis was conducted to evaluate whether PCOS is associated with distinct taxa of gut microbiotas. To fulfill the assumptions of the two‐sample MR analysis, three requirements were ensured: (1) the genetic IVs had a direct relationship with the exposure (gut microbiota), (2) the IVs were independent of confounding factors, and (3) the IVs influenced the outcome variables (PCOS) solely through the exposure (gut microbiota). An overview of the study design is presented in Figure 1.

Open in a separate window

FIGURE 1

Overview of the present study design and the assumptions of Mendelian randomization studies.

3. RESULTS

3.2. Causal association between gut microbiota and PCOS

As shown in Figure 2, nine types of bacteria, specifically, Alphaproteobacteria, Bacilli, Bilophila, Blautia, Burkholderiales, Candidatus Soleaferrea, Cyanobacteria, Holdemania, and Lachnospiraceae and a total of 123 SNPs were found to be associated with PCOS in the above MR methods included at least IVW method (Table S1). The IVW estimate suggested that Bacilli (odds ratio [OR]=1.76, 95% confidence interval [CI]: 1.07–2.91, p=0.03) and Burkholderiales (OR=2.37, 95% CI: 1.22–4.62, p=0.01) had a positive effect on PCOS, and the promoting effect was still significant after considering the associated pleiotropy (cML‐MA‐BIC: OR=1.74, 95% CI: 1.07–2.84, p=0.03; OR=2.14, 95% CI: 1.14–4.02, p=0.02, respectively). The IVW estimate of Lachnospiraceae also showed a promoting effect on PCOS (OR=1.86, 95% CI: 1.04–3.35, p=0.03), while this supportive effect was no longer significant in the cML‐MA‐BIC method. The IVW estimate of Blautia (OR=0.16, 95% CI: 0.03–0.79, p=0.02) and Cyanobacteria (OR=0.51, 95% CI: 0.31–0.83, p=0.01) showed its considerable protective effect against PCOS, and cML‐MA‐BIC method also supported the protective effect of them against PCOS (OR=0.16, 95% CI: 0.03–0.87, p=0.03; OR=0.50, 95% CI: 0.31–0.79, p<0.01, respectively). Similarly, the IVW estimates of Alphaproteobacteria (OR=0.55, 95% CI: 0.30–0.99, p=0.04) and Candidatus Soleaferrea (OR=0.65, 95% CI: 0.43–0.98, p=0.04) also presented a suggestive association with PCOS. In addition, the IVW estimate of Bilophila (OR=0.42, 95% CI: 0.23–0.77, p<0.01) and Holdemania (OR=0.53, 95% CI: 0.35–0.81, p<0.01) also showed its distinctively protective effect against PCOS, while weighted median (OR=0.35, 95% CI: 0.16–0.78, p=0.01; OR=0.45, 95% CI: 0.25–0.84, p=0.01, respectively) and cML‐MA‐BIC estimate (OR=0.50, 95% CI: 0.30–0.83, p=0.01; OR=0.58, 95% CI: 0.39–0.86, p=0.01) also showed a negative effect on the risk of PCOS (Figure 2).

Open in a separate window

FIGURE 2

Forest plots of the association of nine significant gut microbiotas with PCOS. PCOS, polycystic ovary syndrome.

3.3. Sensitive analysis

For the heterogeneity test, the results of Cochrane Q statistics demonstrated no significant heterogeneity of these IVs (Table 1). And the results of MR‐Egger intercept test (Table 1) also showed no evidence of horizontal pleiotropy was observed in the identified significant phenotypes. Funnel plot (Figure S1) indicated there was no presence of bias or systematic heterogeneity. Leave‐one‐out plots (Figure S2) showed the stability of MR results and improves the reliability of the study. The scatter plot was also in line with other results (Figure 3). There were potential outliers of the IVs that were presented on visual inspection in scatter plots. However, further MR‐Egger intercept test did not find any significant outliers (p>0.05, Table 1). Therefore, there was insufficient evidence for horizontal pleiotropy in the association between these categories of gut microbiota and PCOS. Among these nine causal associations, the F‐statistics of the IVs ranged from 15.07 to 136.62, eliminating the bias of weak IVs. This also proves the stability of our results.

TABLE 1

Results of nine gut microbiotas from Cochrane Q test and MR‐Egger intercept test.

Exposure	Cochrane Q test			MR‐Egger intercept test			F‐statistic
	Q	Q_df	p‐value	Intercept	SE	p‐value
Class Alphaproteobacteria	1.463	6	0.962	−3.03E‐02	1.03E‐01	7.80E‐01	34.44
Class Bacilli	13.781	17	0.683	9.97E‐02	5.13E‐02	6.98E‐02	15.07
Family Lachnospiraceae	11.503	15	0.716	2.90E‐02	6.22E‐02	6.49E‐01	15.89
Genus Bilophila	13.996	12	0.301	−1.93E‐02	1.13E‐01	8.68E‐01	26.66
Genus Blautia	0.053	1	0.819	NA	NA	NA	136.62
Genus Candidatus Soleaferrea	10.643	10	0.386	7.73E‐02	1.10E‐01	5.00E‐01	22.72
Genus Holdemania	11.832	13	0.541	9.50E‐03	6.19E‐02	8.81E‐01	25.51
Order Burkholderiales	6.952	9	0.642	4.49E‐02	7.51E‐02	5.66E‐01	30.97
Phylum Cyanobacteria	4.96	7	0.665	−1.06E‐01	1.03E‐01	3.44E‐01	41.44

Open in a separate window

Open in a separate window

FIGURE 3

Scatter plots for the causal association between nine significant gut microbiotas and PCOS. PCOS, polycystic ovary syndrome.

3.4. Reverse MR analysis

According to the results of reverse MR analysis, a total of 69 taxa existed that were significant for at least the IVW method, of which 7 taxa were unknown and therefore not included; of these, 25 had a positive correlation, and 37 had a negative correlation (Table S4). It is worth mentioning that among the nine bacteria of significance in the MR analysis, by using reverse MR analysis, the IVW estimate showed that PCOS could significantly reduce the abundance of Alphaproteobacteria and Lachnospiraceae (OR=0.95, 95% CI: 0.91–0.98, P<0.01; OR=0.97, 95% CI: 0.94–0.99, p=0.02, respectively), and the attenuating effect was still significant after considering the associated pleiotropy (cML‐MA‐BIC: OR=0.95, 95% CI: 0.91–0.98, p<0.01; OR=0.96, 95% CI: 0.94–0.99, p=0.02, respectively) (Figure 4). Details about the results of reverse MR analysis are shown in Table S2 and Tables S4–S6.

Open in a separate window

FIGURE 4

Forest plots with reverse causality between nine significant gut microbiotas and PCOS. PCOS, polycystic ovary syndrome.

4. DISCUSSION

In the present study, using the summary statistics of gut microbiota from the largest GWAS meta‐analysis data for gut microbiota composition conducted by the MiBioGen consortium and the summary statistics of PCOS from the IEU Open GWAS, we performed a bidirectional MR analysis to explore the causal association between gut microbiota and PCOS. Previous studies have primarily focused on establishing correlations between gut microbiota and PCOS. However, due to the various confounding factors and individual differences in observational studies, the conclusions were likely to be controversial. Our study went beyond correlation and established causal relationships between specific taxa of bacteria and PCOS. Specifically, MR analysis showed a clear causal link between nine taxa of microbiotas and PCOS. Furthermore, through reverse MR analysis, we identified causal relationships between PCOS and 69 taxa of gut microbiota. Remarkably, bidirectional causality was observed for Alphaproteobacteria and Lachnospiraceae among these taxa, what's more, heterogeneity and horizontal pleiotropy tests further supported the robustness of our results.

Abundant of observational studies have reported the association between gut microbiota and PCOS. Proteobacteria playing a potential protective role in various metabolic disorders, including obesity, IR, and PCOS ²² have been extensively reported. It has been stated that an increased abundance of Lachnospiraceae may be associated with improved metabolic disorders such as IR, obesity, and type 2 diabetes mellitus, ²³ therefore mitigating the risk of PCOS. Enrichment of specific Blautia by remodeling the gut microbial structure of PCOS mice significantly ameliorated PCOS and alleviates inflammatory state. ²⁴ In addition, part of specific Bacillus was also shown effective in ameliorating the PCOS symptoms. ²⁵ As well, some studies have reported that compounds contained in certain types of Cyanobacteria may be able to be used to treat PCOS. ²⁶ Totally, our findings were consistent with some previous studies, in which we demonstrated nine gut microbiotas causally associated with PCOS by MR analysis. Specifically, Lachnospiraceae, Alphaproteobacteria, Blautia, Bacilli, and Cyanobacteria were previously identified as possibly associated with POCS. Moreover, we found that Holdemanella, Burkholderia, Candidatus Soleaferrea, and Bilophila were also associated with PCOS, which showed limited evidence indicating the direct link between them and PCOS, whereas most of them were reported to be associated with related metabolic disorders such as IR and impaired glucose tolerance. ²⁷ However, due to potential confounding factors in different observational studies, such as disparate study centers, inclusion criteria, and experimental conditions. The conclusions drawn for specific taxa are variable and inconsistent, while our findings, based on MR analysis, could be able to avoid the bias of observational studies and offer a more comprehensive and reliable perspective from a genetic standpoint.

The underlying mechanisms regarding how the gut microbiota interacts with PCOS are still not fully understood and currently are generally thought to be involved in the following aspects. Firstly, numerous studies have shown that gut microbiotas were linked to IR, chronic inflammatory state, and hyperandrogenism. ²⁸ Lipopolysaccharides produced by gram‐negative bacteria could trigger inflammatory responses, cause IR and hyperinsulinemia, ^{6 , 29} ultimately promoting the development of PCOS. Bile acids metabolism is another crucial mechanism ³⁰ through which gut microbiota influences PCOS. Manipulating bile acid metabolism could be a promising therapy for improving and treating PCOS. ⁹ Additionally, the gut microbiome played a crucial role in the decomposition of dietary fiber, and the production of short‐chain fatty acids (SCFAs), ³¹ which have been implicated in IR and PCOS. Moreover, the gut‐brain axis, as a novel mechanism, was thought to be involved in the pathology of PCOS. Patients with PCOS exhibited decreased levels of gut‐brain peptides such as ghrelin, cholecystokinin, and PYY. ³² And gut microbiota could be involved in the development of IR and hyperandrogenism by affecting the secretion of gut‐brain peptides. ³³ The microbiota also played a crucial role in the reproductive endocrine through interactions with estrogens, androgens, insulin, and other hormones. ³⁴ Certain bacteria such as Lactobacillus have been found to lower testosterone levels and alleviate PCOS. ³⁵ Overall, current studies have shown that gut microbiota and PCOS can interact through various mechanisms, including the regulations of inflammatory responses and IR, bile acid metabolism, SCFAs production, the gut‐brain axis, and sex hormones, ultimately influencing the development of PCOS.

Furthermore, of interest is that the results of reverse MR analysis revealed a causal relationship about PCOS on 69 taxa of gut microbiota. Particularly, Lachnospiraceae and Alphaproteobacteria were found to be bidirectionally causally related to PCOS. Several previous studies have suggested the possibility of the reverse relationship. Using linear discriminant analysis, Wang et al. found that the relative abundance of Lachnospiraceae (belonging to Firmicutes phylum) and Burkholderiaceae in the PCOS patients was significantly decreased. ²² Indeed, studies demonstrated that PCOS is associated with decreased alpha diversity and changes in specific Firmicutes. ^{36 , 37} Another observational study found those with PCOS presented a lower abundance of Blautia compared to healthy individuals. ³⁸ In addition, compared to those without nutritional intervention, PCOS patients undergoing nutritional intervention were observed distinct changes in the phylum Firmicutes and Bacteroidetes. ³⁹ Not only that, but there are also studies finding that in PCOS patients with visceral obesity vs those without visceral obesity, a significant alteration in the β‐diversity of gut microbiota emerged. ⁴⁰ All in all, these studies suggested that PCOS may affect the distribution and abundance of different taxa of gut microbiota, and that some of these bacteria may in turn act to influence the risk of PCOS. And our results further reinforced the causal relationship between PCOS and gut microbiota.

Overall, investigations on the correlation between different taxa of gut microbiota and PCOS remain unclear based on previous studies. There are also multiple potential mechanisms of interaction between gut microbiota and PCOS. Our results suggested a potential bidirectional interaction between PCOS and gut microbiota and also provided exact genetic evidence for the relationship of several taxa of gut microbiota and PCOS. Using MR analysis and reverse MR analysis, the present study redefined the correlation between gut microbiota and PCOS, further presenting an explicit understanding of the causality underlying this complex and intricate correlation. Furthermore, these findings have significant implications for future targeted therapies that focus on modifying the gut microbiota as a potential treatment strategy for PCOS. In the future, interventions targeting these nine specific gut microbiota taxa could involve the administration of exogenous supplementation or inhibitors to modulate their abundance, thereby reducing the risk of PCOS development in susceptible individuals.

There are several strengths in our present study. First and foremost, as mentioned before, the MR design enables the demonstration or verification of a causal relationship between exposure and outcomes in the absence of randomized clinical trials. Through conducting a bidirectional MR analysis, as well as subsequent sensitivity analysis, we were able to establish certain causality between gut microbiota and PCOS, while avoiding the confounders typically encountered in observational studies. Secondly, the present study also systematically and comprehensively analyzed 211 taxa of gut microbiota and had a clear overview of their relationship with PCOS, which served as a valuable reference for clinicians and researchers in this field and provided a foundation for subsequent studies exploring the potential of targeting the gut microbiota as a therapeutic strategy for PCOS. Thirdly, our study utilized genetic variants of both the exposure and outcome obtained from large, well‐powered GWAS studies, thereby ensuring that the instruments used in the MR analysis were robust. Furthermore, we employed several methods, such as Cochran's Q statistic, MR‐Egger intercept tests, and leave‐one‐out analysis, to detect and exclude pleiotropy from our study. This ensured that the potential causal associations identified were robust and reliable.

However, despite the strengths of our study, several limitations must be acknowledged. Firstly, to mitigate the potential bias arising from ethnic heterogeneity, our study was limited to individuals of European ancestry. As such, further research is required to determine whether our conclusions can be generalized to other racial groups. Secondly, as our analysis was based on summary statistics rather than raw data, we were unable to conduct subgroup analyses based on factors such as age and specific PCOS types. And the PCOS dataset which we used was not detailed for medication use. Thus, further research using more detailed data may be necessary to explore the potential impact of such factors on our findings. Thirdly, due to limited information, it was challenging to investigate the extent of overlap between the selected gut microbiota and PCOS populations. This potential overlap may have introduced weak instrument bias during the analysis. Therefore, future research with more extensive data collection may be required to further explore this issue.

5. CONCLUSION

The bidirectional MR analysis provided a robust understanding of identifying the causality underlying the intrinsic and complicated correlation between gut microbiota and PCOS based on genetic evidence. Additionally, this study provided a few taxa of gut microbiota that are positively or negatively causally associated with PCOS, which also provided better insights for exploring how targeted modulation of specific gut microbes for the treatment of PCOS. Future studies are required to elucidate the specific mechanisms by which the microbiome impacts the pathogenesis, progression, and phenotypes of PCOS, as well as how PCOS interacts with gut microbiota.

AUTHOR CONTRIBUTIONS

Ruo‐Lin Mao and Li‐Xia Zhu: study design and interpretation, data acquisition and analysis, statistical analysis and verification of the underlying data, initial draft of the manuscript, and critical revision of the manuscript for important intellectual content. Xiang‐Fei Wang: data acquisition and analysis, statistical analysis and verification of the underlying data, initial draft of the manuscript, and critical revision of the manuscript for important intellectual content. Meng Wang, Rui Long, and Lei Jin: initial draft of the manuscript and critical revision of the manuscript for important intellectual content. Jue‐Pu Zhou: statistical analysis and verification of the underlying data and critical revision of the manuscript for important intellectual content. The final manuscript was read and approved by all authors.

FUNDING INFORMATION

This work was supported by the National Key Research and Development Program of China under Grant 2021YFC2700603, and the Experimental technology research project from Huazhong University of Science and Technology (2024M108), and Hubei Key Laboratory Open Fund of Environmental and Health Hazards of Persistent Toxic Pollutants (PTS2022‐03).

CONFLICT OF INTEREST STATEMENT

The authors declare that they do not have any conflicts of interest.

ETHICS STATEMENT

This research was conducted using published studies and consortia providing publicly available summary statistics. All original studies have been approved by the corresponding ethical review board, and the participants have provided informed consent. In addition, no individual‐level data was used in this study. Therefore, no new ethical review board approval was required.

Supporting information

Notes

Mao R‐L, Wang X‐F, Zhou J‐P, et al. Causal relationships between gut microbiota and polycystic ovarian syndrome: A bidirectional Mendelian randomization study. Acta Obstet Gynecol Scand. 2024;103:2232‐2241. 10.1111/aogs.14957 [Europe PMC free article] [Abstract] [CrossRef] [Google Scholar]

REFERENCES