Causal associations between systemic inflammation and polycystic ovary syndrome: a Mendelian randomisation study emphasising the role of CXCL11
•,,,
...
Abstract
Purpose Systemic inflammation has been increasingly implicated in the pathogenesis of polycystic ovary syndrome (PCOS), but the causal nature and direction of this relationship remain uncertain. This study aimed to evaluate the potential causal associations between circulating inflammatory cytokines and the risk of PCOS using a Mendelian randomisation (MR) approach.
Methods We conducted a two-sample MR analysis using summary-level data from large-scale genome-wide association studies involving 91 systemic inflammatory markers (n=14 824) and PCOS (10 074 cases and 103 164 controls) among individuals of European ancestry. Genetic variants associated with cytokines at genome-wide significance (p<5×10−8) were selected as instrumental variables. The inverse-variance weighted method was used as the primary analytical strategy, supplemented by sensitivity analyses and correction for multiple testing.
Results Genetically predicted higher circulating levels of C-X-C motif chemokine ligand 11 (CXCL11) were significantly associated with a reduced risk of PCOS (OR=0.740, 95% CI 0.625 to 0.871, p<0.001), and this association remained statistically significant after multiple testing correction (adjusted p=0.030). Nominal associations with decreased PCOS risk were also observed for interleukin-13 (IL-13), IL-10 and adenosine deaminase (ADA), but these did not withstand correction for multiple comparisons. No evidence of horizontal pleiotropy was detected, and leave-one-out sensitivity analyses supported the robustness of the findings.
Conclusion These results support a potential causal role of systemic inflammation in the development of PCOS, with CXCL11 emerging as a promising inflammatory marker and potential therapeutic target. Further studies are needed to validate these findings and explore their clinical relevance in PCOS management.
What is already known on this topic
Systemic inflammation has been implicated in polycystic ovary syndrome (PCOS), but the causal nature of this relationship remains unclear.
What this study adds
Mendelian randomisation analysis shows that genetically predicted higher circulating CXCL11 levels are causally associated with a reduced risk of PCOS.
How this study might affect research, practice or policy
CXCL11 may serve as a promising biomarker and therapeutic target in PCOS management.
Introduction
Polycystic ovary syndrome (PCOS) is a common and complex endocrine disorder affecting 11%–13% of women of reproductive age worldwide, with prevalence varying across ethnicities and diagnostic criteria.1 2 Characterised by hyperandrogenism, ovulatory dysfunction and polycystic ovarian morphology, PCOS poses substantial challenges in both diagnosis and long-term management. In addition to reproductive implications, PCOS is associated with a wide range of metabolic disturbances—including insulin resistance, dyslipidaemia and obesity—as well as increased risks for type 2 diabetes, cardiovascular disease and psychological disorders.3 4
Recent studies have emphasised the critical role of chronic low-grade inflammation in the pathogenesis of PCOS. Elevated levels of proinflammatory markers such as interleukin-6 (IL-6), tumour necrosis factor-alpha, C-reactive protein and vascular endothelial growth factor have been consistently reported in women with PCOS.3 5–7 Furthermore, inflammatory pathways are tightly interconnected with endocrine and metabolic dysregulation—particularly hyperandrogenism and insulin resistance—creating a vicious cycle that exacerbates ovarian dysfunction.3 8 These insights suggest that immune-inflammatory mechanisms may represent a critical yet underexplored axis in the development and progression of PCOS.
However, most existing evidence is derived from cross-sectional or observational studies, which are inherently limited by confounding and reverse causation.9 Mendelian randomisation (MR) offers a robust alternative by using genetic variants as instrumental variables to infer potential causal relationships.10 Previous MR analyses have suggested associations between PCOS and certain inflammatory markers, yet the directionality and specificity of these associations remain poorly defined.5–7
Therefore, this study aims to investigate the causal relationships between systemic inflammatory cytokines and PCOS using a two-sample MR approach. By leveraging large-scale genome-wide association study (GWAS) data, we seek to clarify the causal roles of key inflammatory mediators in the pathophysiology of PCOS and, conversely, to determine whether PCOS itself influences systemic inflammatory profiles. This work may offer novel insights into disease mechanisms and identify potential targets for therapeutic intervention.
Materials and methods
Study overview
This MR study aimed to investigate the causal relationship between inflammatory cytokines and PCOS. The study adhered to the Strengthening the Reporting of Mendelian Randomization Studies guidelines and used publicly available datasets with prior ethical approval obtained for all original studies. A study diagram is provided in figure 1.
Study flow diagram. This study performed a Mendelian randomisation analysis to determine if systemic inflammatory cytokines were causally associated with PCOS. PCOS, polycystic ovary syndrome; SNP: single nucleotide polymorphism.
Data sources
Summary statistics for inflammatory cytokines were obtained from a GWAS that examined 91 inflammatory cytokines in 14 824 participants of European ancestry using the Olink Target Inflammation panel with proximity extension assay technology.8 For PCOS, GWAS data were sourced from a study involving 10 074 cases and 103 164 controls of European ancestry.11 PCOS diagnoses were based on either the National Institutes of Health criteria (requiring hyperandrogenism and oligo/anovulation) or the Rotterdam criteria (requiring at least two of the following three features: hyperandrogenism, oligo/anovulation and polycystic ovarian morphology).12
Selection of genetic instruments
Genetic instruments were selected based on stringent criteria to ensure validity. Single-nucleotide polymorphisms (SNPs) that were strongly and independently (P＜5×10-8, linkage disequilibrium coefficient r2 <0.001, region width 10 000 kb) associated with inflammatory cytokines were selected as genetic instruments. 13The selected SNPs and their characteristics, including effect sizes, P-values, and other related data, are presented in online supplemental Table S1). To maximise the genetic variance explained by these variants, we used a threshold of 1×10-5 for selecting genetic variants. The F value was calculated from the formula
N represents the sample size of the dataset, beta is the effect size of the SNPs on the exposure variable, SE denotes the SE of beta and EAF refers to the effect allele frequency. We selected genetic variants with an F value greater than 10, indicating a strong correlation with the exposure.14 15 To ensure that the effects of the SNPs on both the exposure and the outcome were aligned with the identical allele, we harmonised the SNPs between exposures and outcomes. The final remaining SNPs constituted the instrumental variables for the exposures.
Mendelian randomisation analysis
Causal effects were estimated using the inverse-variance weighted (IVW) method as the primary analysis, as it provides the most statistically efficient estimates when all instruments are valid. To address potential violations of MR assumptions, complementary methods such as MR-Egger, weighted median and weighted mode analyses were also performed. These methods provide robust estimates under different patterns of pleiotropy and invalid instruments, though at the cost of reduced statistical power.16–18
To detect and account for horizontal pleiotropy, the MR-Egger intercept test was used. Heterogeneity among SNPs was assessed using Cochran’s Q statistic. In the presence of significant heterogeneity (p＜0.05), the IVW random-effects model was applied. Leave-one-out analyses and scatter plots were conducted to assess the robustness of the results and identify influential SNPs.
Statistical analysis
All analyses were conducted using the TwoSampleMR (V.0.5.6) and MR-PRESSO (V.1.0) packages in R software (V.4.4.0) (https://github.com/MRCIEU/TwoSampleMR). Effect estimates are presented as ORs with 95% CIs. The Benjamini-Hochberg method was used to control for multiple testing, with an adjusted p<0.05 considered statistically significant. Supplementary analyses, including sensitivity testing and visualisation, ensured the robustness and reliability of the findings.
Results
Main causal effects of inflammatory cytokines on polycystic ovary syndrome (PCOS)
Among the 91 inflammatory cytokines evaluated, four showed evidence of a causal association with PCOS (table 1).
Table 1
•
Inflammatory cytokines significantly associated with PCOS
Genetically predicted levels of C-X-C motif chemokine ligand 11 (CXCL11) were significantly associated with a decreased risk of PCOS (OR=0.740, 95% CI 0.625 to 0.871, P＜0.001), and this association remained statistically significant after Benjamini–Hochberg correction for multiple testing (adjusted p=0.030). In addition, genetically predicted levels of IL-13 (OR=0.730, 95% CI 0.591 to 0.896, p=0.003, adjusted p=0.123), IL-10 (OR=0.777, 95% CI 0.616 to 0.979, p=0.033, adjusted p=0.895) and ADA (OR=0.881, 95% CI 0.780 to 0.994, p=0.039, adjusted p=0.895) were nominally associated with reduced PCOS risk, but did not remain significant after multiple testing correction (online supplemental Table S2).
Sensitivity analyses and pleiotropy assessment
The MR-Egger regression analysis showed no evidence of directional pleiotropy for any of the four significant cytokines (figure 2). The intercept tests yielded non-significant results for C-X-C motif chemokine 11 (intercept=−0.0028, p=0.893), IL-13 (intercept=−0.0054, p=0.826), IL-10 (intercept=−0.0069, p=0.779) and adenosine deaminase (intercept=0.0089, p=0.794). Results of the horizontal pleiotropy analysis are presented in online supplemental Table S3.
Circular plot of Mendelian randomisation study on inflammatory cytokines and PCOS by MR methods. Colours indicate beta values (red, high; blue, low). MR, Mendelian randomisation; PCOS, polycystic ovary syndrome.
Cochran’s Q statistics indicated no significant heterogeneity for C-X-C motif chemokine 11 (Q=21.1,p=0.969), IL-13 (Q=26.1, p=0.459) and adenosine deaminase (Q=1.4, p=0.839). IL-10 showed marginal evidence of heterogeneity (Q=42.9, p=0.060), but this did not substantially affect the main findings (online supplemental Table S4).
Individual single nucleotide polymorphism (SNP) effects and leave-one-out analysis
Leave-one-out sensitivity analyses demonstrated that no single SNP substantially influenced the overall causal estimates for any of the four cytokines (figure 3). Detailed SNP-specific results and their individual contributions to the causal estimates are presented in online supplemental Table S5. For C-X-C motif chemokine 11, all 36 SNPs showed consistent direction of effect, with rs148828177 showing the strongest individual effect.
The causal relationships between systemic inflammatory cytokines and PCOS using different MR methods. Each panel represents the causal estimates for PCOS on a specific inflammatory cytokine: (A) C-X-C motif chemokine 11, (B) interleukin-13, (C) interleukin-10 (D) adenosine deaminase levels. The slope of each line corresponds to the causal estimates for each method. Individual SNP effects on the outcome (represented by points and vertical lines) against their effects on the exposure (represented by points and horizontal lines) are illustrated in the background. MR, Mendelian randomisation; PCOS, polycystic ovary syndrome; SNP, single nucleotide polymorphism.
Complementary analysis methods
The consistency of effect estimates across different MR methods (IVW, weighted median and weighted mode) further supported the robustness of our findings (figure 4). All four analytical approaches (IVW, MR-Egger, weighted median and weighted mode) showed consistent direction and magnitude of effects for the significant cytokines, particularly for C-X-C motif chemokine 11. Scatter plots of SNP effects on exposure vs outcome showed consistent patterns of association, particularly for C-X-C motif chemokine 11, which demonstrated the most stable and significant relationship with PCOS risk (figure 2).
The leave-one-out plot of the causal relationships between four systemic inflammatory cytokines and PCOS. Each panel corresponds to MR estimates for a systemic inflammatory cytokine on PCOS: (A) C-X-C motif chemokine 11, (B) interleukin-13, (C) interleukin-10 and (D) adenosine deaminase levels. MR, Mendelian randomisation; PCOS, polycystic ovary syndrome.
Discussion
In this two-sample MR study, we explored the bidirectional causal relationship between systemic inflammatory regulators and the risk of PCOS. Our findings indicate that elevated levels of certain inflammatory markers—particularly CXCL11 and IL-13—may exert protective effects against PCOS, whereas PCOS itself appears to causally increase the levels of IL-10 and ADA. These insights shed new light on the complex interplay between inflammation and PCOS pathogenesis, underscoring the potential therapeutic utility of targeting inflammatory pathways in the management of PCOS.
CXCL11, a CXC chemokine that signals through the CXCR3 receptor, plays a central role in immune cell recruitment, particularly activated T cells and NK cells, to sites of inflammation or disease. Its dual functionality is evident in its pro-inflammatory role in autoimmune disorders and transplant rejection, as well as its immune-activating effects in cancer by promoting cytotoxic lymphocyte infiltration.19–21 These observations highlight the microenvironment-dependent regulation of adaptive immunity by CXCL11.
Our MR analysis revealed that genetically predicted higher circulating levels of CXCL11 were associated with a reduced risk of PCOS, suggesting a potential protective role of this chemokine in ovarian function. However, recent evidence highlights the context-dependent nature of CXCL11 in reproductive disorders. For instance, a study reported significantly elevated CXCL11 levels in the follicular fluid of patients with autoimmune thyroiditis (AIT), where it synergised with interferon-γ to recruit CXCR3+T lymphocytes, exacerbating ovarian inflammation and impairing folliculogenesis.22 This apparent contradiction—where CXCL11 appears protective in PCOS but pro-inflammatory in AIT—may reflect tissue-specific or disease-context-dependent pleiotropic effects. Additionally, a recent clinical study reported a strong correlation between serum CXCL11 levels and both prolactin and 17-OH progesterone levels in PCOS patients.23 These observations suggest that CXCL11 may influence key features of PCOS such as hyperandrogenism and dysregulation of the hypothalamic–pituitary–ovarian axis. Together, these findings support the emerging view of CXCL11 as a multifaceted player in PCOS pathophysiology. Therapeutic strategies targeting CXCL11 may thus require precision approaches, potentially involving tissue-specific delivery, to avoid paradoxical effects across different reproductive disorders.
Similarly, IL-13 demonstrated a nominally protective effect in our analysis, although this association did not remain statistically significant after correction for multiple testing. IL-13 is a well-characterised anti-inflammatory cytokine that modulates immune responses by promoting T-helper 2 cell differentiation and enhancing immune tolerance.24 In PCOS, dysregulated immune function contributes to chronic low-grade inflammation, which is thought to exacerbate the disorder’s metabolic and reproductive features. Several studies have shown that IL-13 may play a role in modulating key aspects of PCOS, including immune responses and inflammatory pathways. One study observed elevated IL-13 levels in the follicular fluid of PCOS patients, suggesting a potential role in modulating local immune responses.25 Another study reported that IL-13 may influence the balance between pro- and anti-inflammatory signals in PCOS, with effects that may vary according to body mass index.26 In addition, IL-13 has been shown to participate in immune regulation in obesity-related inflammation, a common feature of PCOS.27 While its association with PCOS risk was not statistically significant after multiple testing adjustment in our study, the directionality and consistency of the findings suggest that IL-13 may remain a relevant modulator of PCOS-associated immune dysfunction and merits further investigation.
In our study, genetically predicted levels of IL-10 and ADA were also nominally associated with reduced PCOS risk, although these associations did not meet statistical significance after correction. IL-10 is an anti-inflammatory cytokine known for suppressing pro-inflammatory responses and promoting immune tolerance.28 29 The nominal association observed here aligns with its immunoregulatory role and suggests a potential protective effect in PCOS. However, several observational studies have reported reduced circulating IL-10 levels in PCOS patients compared with healthy controls,30 31 suggesting a deficiency in anti-inflammatory signalling as a contributor to PCOS-associated inflammation. The discrepancy between our MR findings and previous observational studies may reflect differences in study design: MR relies on germline genetic variants as proxies for lifelong exposure, minimising confounding and reverse causation, whereas cross-sectional studies may capture disease-related secondary changes in cytokine levels.
ADA is a key enzyme in purine metabolism and is also involved in immune activation and T-cell function.32 33 Previous studies have reported increased ADA activity in women with PCOS, particularly in those with obesity or insulin resistance.34 35 In contrast, our MR analysis found a nominal inverse association between genetically predicted ADA levels and PCOS risk. This apparent inconsistency may reflect differences between genetically determined basal ADA expression and acquired elevations secondary to metabolic or inflammatory stress. It is also possible that different ADA isoforms, or tissue-specific expression patterns, exert divergent biological effects that are not adequately captured by current GWAS-based summary statistics. Future studies employing bidirectional MR, tissue-specific expression analyses and longitudinal inflammatory profiling are needed to further clarify the causal and temporal dynamics of ADA in PCOS pathogenesis.
The clinical implications of our findings are potentially important, as they suggest that therapeutic strategies targeting CXCL11—a cytokine with a robust inverse association with PCOS—may offer a novel avenue for disease management. Given the multifactorial nature of PCOS, interventions that modulate inflammatory signalling pathways involved in insulin resistance, hyperandrogenism and ovarian dysfunction may help alleviate both metabolic and reproductive manifestations of the disorder. Although IL-13, IL-10 and ADA showed nominal associations with reduced PCOS risk, these findings did not remain significant after correction for multiple testing and should be interpreted with caution. Further investigation is warranted to explore their potential roles in PCOS pathophysiology and therapeutic development.
Several limitations of our study should be noted. While MR provides a robust framework to infer causality, it relies on key assumptions, including the validity of genetic instruments, absence of horizontal pleiotropy and minimal confounding by population stratification. Although we performed comprehensive sensitivity analyses to address these issues, the possibility of residual bias cannot be excluded. In addition, our use of summary statistics from GWAS limits the assessment of population-specific effects, particularly in non-European populations with distinct genetic backgrounds. Finally, MR captures lifetime genetic predisposition and may not reflect dynamic or tissue-specific cytokine regulation.
In conclusion, our study provides genetic evidence supporting the role of systemic inflammation in PCOS, with CXCL11 emerging as a promising therapeutic target. Further studies in diverse populations and experimental models are needed to validate these findings.
Contributors: XX carried out the experiments and drafted the manuscript. WM performed the statistical analysis. DW participated in data acquisition, analysis and interpretation. QL contributed to the study design and critically revised the manuscript. All authors have read and approved the final version of the manuscript. QL is the guarantor of this work and accepts full responsibility for the integrity of the data and the accuracy of the analysis.
Funding: This study was supported by the Multidisciplinary Clinical Research Innovation Team Project of Beijing Chao-Yang Hospital (grant no. CYDXK202203).
Competing interests: QL has served as an editorial member of GOCM. There are no competing interests.
Patient and public involvement: Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
Provenance and peer review: Not commissioned; externally peer-reviewed.
Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Data availability statement
The data are included within the article and its supplementary information. The data are included within the article and supplementary information.
Ethics statements
Patient consent for publication:
Not applicable.
Ethics approval:
Not applicable.
Acknowledgements
We gratefully acknowledge the investigators and institutions that made the GWAS datasets used in this study publicly available.
Stener-Victorin E, Teede H, Norman RJ, et al. Polycystic ovary syndrome. Nat Rev Dis Primers2024; 10:27.
Teede HJ, Tay CT, Laven J, et al. Recommendations from the 2023 International evidence-based guideline for the assessment and management of polycystic ovary syndrome. Fertil Steril2023; 120:767–93.
Jin L, Yu J, Chen Y, et al. Polycystic ovary syndrome and risk of five common psychiatric disorders among European women: A two-sample Mendelian randomization study. Front Genet2021; 12.
Jia M-J, Zhou L, Liu X-N, et al. Genetically predicted serum metabolites mediate the association between inflammatory proteins and polycystic ovary syndrome: a Mendelian randomization study. Front Med2024; 11:1433612.
Chen H, Zhang Y, Li S, et al. The association between genetically predicted systemic inflammatory regulators and polycystic ovary syndrome: A Mendelian randomization study. Front Endocrinol2021; 12:731569.
Zhao JH, Stacey D, Eriksson N, et al. Genetics of circulating inflammatory proteins identifies drivers of immune-mediated disease risk and therapeutic targets. Nat Immunol2023; 24:1540–51.
Smith GD, Ebrahim S. “Mendelian randomization”: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol2003; 32:1–22.
Tang B, Lin N, Liang J, et al. Leveraging pleiotropic clustering to address high proportion correlated horizontal pleiotropy in Mendelian randomization studies. Nat Commun2025; 16.
Day F, Karaderi T, Jones MR, et al. Large-scale genome-wide meta-analysis of polycystic ovary syndrome suggests shared genetic architecture for different diagnosis criteria. PLoS Genet2018; 14.
Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril2004; 81:19–25.
Pierce BL, Ahsan H, Vanderweele TJ, et al. Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol2011; 40:740–52.
Smith GD, Timpson N, Ebrahim S, et al. Strengthening causal inference in cardiovascular epidemiology through Mendelian randomization. Ann Med2008; 40:524–41.
Burgess S, Small DS, Thompson SG, et al. A review of instrumental variable estimators for Mendelian randomization. Stat Methods Med Res2017; 26:2333–55.
Bowden J, Davey Smith G, Haycock PC, et al. Consistent estimation in Mendelian randomization with some invalid instruments using a weighted Median estimator. Genet Epidemiol2016; 40:304–14.
Hartwig FP, Davey Smith G, Bowden J, et al. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol2017; 46:1985–98.
Mitsuhashi N, Wu GD, Zhu H, et al. Rat chemokine CXCL11: structure, tissue distribution, function and expression in cardiac transplantation models. Mol Cell Biochem2007; 296:1–9.
Li Y, Han S, Wu B, et al. CXCL11 correlates with immune infiltration and impacts patient immunotherapy efficacy: A Pan-cancer analysis. Front Immunol2022; 13:951247.
Hatziagelaki E, Pergialiotis V, Kannenberg JM, et al. Association between biomarkers of low-grade inflammation and sex hormones in women with polycystic ovary syndrome. Exp Clin Endocrinol Diabetes2020; 128:723–30.
Gallinelli A, Ciaccio I, Giannella L, et al. Correlations between concentrations of interleukin-12 and interleukin-13 and lymphocyte subsets in the follicular fluid of women with and without polycystic ovary syndrome. Fertil Steril2003; 79:1365–72.
Vasyukova E, Zaikova E, Kalinina O, et al. Inflammatory and anti-inflammatory parameters in PCOS patients depending on body mass index: A case-control study. Biomedicines2023; 11.
Mishra S, Ashish A, Rai S, et al. The impact of inflammatory cytokines on recurrent pregnancy loss: A preliminary investigation. Reprod Sci2025; 32:804–14.
Wu Y-R, Hsing C-H, Chiu C-J, et al. Roles of IL-1 and IL-10 family cytokines in the progression of systemic lupus erythematosus: Friends or foes? IUBMB Life2022; 74:143–56.
Abraham Gnanadass S, Divakar Prabhu Y, Valsala Gopalakrishnan A, et al. Association of metabolic and inflammatory markers with polycystic ovarian syndrome (PCOS): an update. Arch Gynecol Obstet2021; 303:631–43.
Nikseresht M, Shahrebabaki AM, Mohammad-Sadeghipour M, et al. Comparison of serum levels of IL-10 and IL-11 and mRNA expression of IL-10, IL-11, COX-2, BCL6, and ZEB Family in peripheral blood mononuclear cells (PBMC) of women with polycystic ovary syndrome and healthy individuals. J Reprod Immunol2024; 164:104281.
Passos DF, Bernardes VM, da Silva JLG, et al. Adenosine signaling and adenosine deaminase regulation of immune responses: impact on the immunopathogenesis of HIV infection. Purinergic Signal2018; 14:309–20.
Kaljas Y, Liu C, Skaldin M, et al. Human adenosine deaminases ADA1 and ADA2 bind to different subsets of immune cells. Cell Mol Life Sci2017; 74:555–70.
Kahraman S, Eroglu Altinova A, Elgun S, et al. Serum activities of dipeptidyl peptidase-4 and adenosine deaminase in polycystic ovary syndrome: association with obesity. Gynecol Endocrinol2019; 35:714–8.
