Unraveling the causal association between leukocyte telomere length and infertility: A two-sample Mendelian randomization study

Telomeres, the protective caps at the ends of eukaryotic linear chromosomes [ 1 ], play a crucial role in maintaining genomic stability and cellular health [ 2 ]. Shortened telomeres have been associated with various age-related diseases and conditions, including cancer, cardiovascular disease, and neurodegenerative disorders [ 3 ]. In recent years, there has been growing interest in exploring the potential link between telomere length and infertility, a condition that affects a significant proportion of the population. With the rapid modernization of society, infertility has emerged as a formidable hurdle for individuals and couples worldwide, with a prevelance ranging from 10% to 15% [ 4 ]. It encompasses a range of conditions, including male and female infertility, sperm abnormalities, and endometriosis. Understanding the underlying mechanisms and potential biomarkers for infertility is crucial for developing effective diagnostic tools and therapeutic interventions. Observed studies have found the relationship between leukocyte telomere length (LTL) and infertility. Shorter telomere length was associated with female infertility factors, such as polycystic ovary syndrome (PCOS), ovarian insufficiency and tubal factor in oocyte granulosa cells, endometrial tissue and leukocytes [ 5 ]. And shorter LTL was also associated with greater odds of endometriosis [ 6 ]. As for male infertility, the mean of LTL and sperm telomere length (STL) were significantly shorter in infertile men compared with fertile individuals [ 7 ]. However, conflicting results regarding the association between leukocyte telomere length and infertility have been observed in numerous RCT studies [ 5 ], and the causal relationship and underlying mechanisms are still to be elucidated. Mendelian randomization (MR), a method that utilizes genetic variants as instrumental variables, offers a unique opportunity to investigate causal relationships between exposures and outcomes [ 8 ]. Unlike traditional observational studies, which are prone to confounding and reverse causality, mendelian randomization leverages genetic variants that are randomly assigned at conception and are not influenced by confounders or disease status. This approach allows researchers to overcome limitations of observational studies and provide more robust evidence for causal associations [ 9 ]. By leveraging genetic variants that are associated with telomere length, Mendelian randomization can provide insights into the causal association between telomere length and infertility. In this study, we aim to utilize two-sample Mendelian randomization to explore the causal relationship between leukocyte telomere length and major causes of infertility. Specifically, we will investigate the potential causal association between telomere length and male and female infertility, sperm abnormalities, and endometriosis. By utilizing large-scale genetic data and statistical approaches, we aim to provide robust evidence regarding the role of telomere length in infertility.

We conducted a comprehensive analysis using the male infertility and sperm abnormalities datasets from FinnGen to evaluate the causal association between leukocyte telomere length (LTL) and the occurrence of male infertility and sperm abnormalities. The findings from Table 1 indicate that no causal relationships were observed between LTL and male infertility (OR 1.269, 95%CI (0.838, 1.921), p = 0.261). Similarly, there were no causal associations found between LTL and sperm abnormalities (OR 0.918, 95%CI (0.640, 1.317), p = 0.643). The analysis did not uncover any heterogeneity or pleiotropy. MR, Mendelian randomization; LTL, leukocyte telomere length; OR, odds ratio; CI, confidence interval; p, p value; MVMR, multivariable Mendelian randomization. To investigate the potential causal relationship between leukocyte telomere length (LTL) and female infertility, we conducted a rigorous analysis utilizing the comprehensive collection of five datasets obtained from FinnGen. The analysis revealed no causal relationships between LTL and female infertility (OR 1.108, 95%CI (0.962, 1.277), p = 0.155). Consistently, LTL did not have any causal effects on female infertility caused by anovulation (OR 0.932, 95% CI (0.667, 1.303), p = 0.682), endometriosis (OR 1.174, 95%CI (0.891, 1.547), p = 0.255), tubal factors (OR 1.154, 95%CI (0.760, 1.753), p = 0.502), or other factors related to cervigal, vaginal, and other organs (OR 1.086, 95%CI (0.937, 1.259), p = 0.273) ( Table 2 ). The analysis did not uncover any heterogeneity or pleiotropy. MR, Mendelian randomization; LTL, leukocyte telomere length; OR, odds ratio; CI, confidence interval; p, p value; MVMR, multivariable Mendelian randomization. We utilized three datasets from FinnGen and UKbiobank to assess the causal relationships between LTL and endometriosis. In the FinnGen dataset, heterogeneity (P heterogeneity = 0.023) and pleiotropy (P MR-PRESSO = 0.019) were observed in this dataset. After removing outliers identified by MR-PRESSO, the MR analyses indicated a causal relationship between LTL and endometriosis. Specifically, the risk of endometriosis increased by 0.304 times with a one standard deviation decrease in genetically predicted LTL, as determined by the random effects IVW method (OR = 1.304, 95% CI = 1.122–1.517, p = 0.001) ( Table 3 ). In the UKbiobank dataset, no clear evidence of heterogeneity or pleiotropy was found, leading us to employ the IVW method with a fixed effects model for causal estimation. We also identified causal relationships between LTL and a higher risk of self-reported endometriosis (OR 1.004, 95%CI (1.002, 1.005), p = 2.65E-5) and ICD10 diagnosed endometriosis (OR 1.002, 95%CI (1.000,1.003), p = 0.049). The effects of the SNPs on LTL and endometriosis are depicted in Fig 2 through scatter plots. After excluding the effects of confounding factors such as BMI and smoking, the results of MVMR showed that the causal effect of LTL on both FinnGen dataset (P MVMR = 0.023) and self-reported endometriosis (P MVMR = 0.002) remained significant. (A) Endometriosis as the outcome. (B) Endometriosis and infertility occurring together as the outcome. (C) ICD10 endometriosis of uterus as the outcome. (D) Self-reported endometriosis as the outcome. Trend lines derived from five different MR methods are also included in each scatter plot to indicate cause and effect. LTL, leukocyte telomere length; MR, Mendelian Randomization; ICD, International classification of Diagnose. MR, Mendelian randomization; LTL, leukocyte telomere length; OR, odds ratio; CI, confidence interval; p, p value; MVMR, multivariable Mendelian randomization. Furthermore, we investigated the causal effects of LTL on endometriosis in different organs. The forest plot ( Fig 3 ) revealed that the causal effects of LTL were particularly pronounced in endometriosis of the intestine (OR 3.58, 95%CI (1.68, 8.02), p = 0.002) and ovary (OR 1.37, 95%CI (1.10, 1.70), p = 0.004). Additional details regarding the results of other statistical methods, as well as the results of pleiotropy and heterogeneity testing, can be found in S4 – S7 Tables. The leave-one-out analysis demonstrated consistent results with the MR studies, indicating that no single SNP significantly influenced the findings. This suggests the robustness and reliability of the MR studies ( S1 Fig ). LTL, leukocyte telomere length; OR, odds ratio; IVW, inverse-variance weighted; CI, confidence interval.

Our study utilized MR analysis with data summaries from a large sample GWAS analysis to investigate the causal association between LTL and the risk of male and female infertility, as well as abnormal spermatozoa in the European population. Our findings indicate that there is no significant causal relationship between LTL and these infertility factors. However, we did observe a strong causal relationship between LTL and endometriosis, which has important implications for the diagnosis and prognosis of this condition.

In our research, we employed a two-sample mendelian randomization approach to meticulously examine the causal impacts of LTL on various aspects of infertility, encompassing male and female infertility, sperm abnormalities, and endometriosis. The MR study based on three fundamental assumptions: (1) the instrumental variables (single nucleotide polymorphisms or SNPs) exhibit robust associations with the exposures under scrutiny; (2) the instrumental variables are not linked to any confounding factors; (3) the instrumental variables exclusively influence the outcome through the exposure, without exerting any direct effect on the outcome itself. The schematic representation of our study design is visually depicted in Fig 1 . LTL leukocyte telomere length, snps single nucleotide polymorphisms. The genetic association data for LTL were obtained from a genome-wide association study (GWAS) involving 472,174 European participants [ 10 ], adjusting for age, sex, and the first ten principal components (PCs). These participants were aged 40–69 years and had an equal distribution of males (45.8%) and females (54.2%). LTL measurements were determined using a quantitative polymerase chain reaction (qPCR) assay. To mitigate any potential biases stemming from body mass index (BMI) and smoking, we procured comprehensive BMI GWAS summary data from the UK Biobank and smoking GWAS sunmmary data from the GWAS & Sequencing Consortium of Alcohol and Nicotine use (GSCAN) [ 11 ]. Leveraging these datasets, we conducted multivariable Mendelian Randomization (MVMR) to ensure a robust analysis. In this study, we investigated the outcomes of male infertility, female infertility, sperm abnormalities, and endometriosis. We utilized the Ieu Open GWAS Project database ( https://gwas.mrcieu.ac.uk/ ) to identify the most suitable GWAS summary data that align with our research objectives. Finally, we obtained summary data on infertility from the FinnGen Consortium R5 release and the UK Biobank. S1 Table provides detailed information about the specific outcomes examined in our analysis. The diagnostic criteria for infertility-related diseases were based on the International Classification of Diseases (ICD) codes, including ICD-9 and ICD-10. Female infertility was further categorized into different subtypes, including Female infertility, Female infertility associated with anovulation, Female infertility of cervigal, vaginal, other or unspecified origin, and Female infertility of tubal origin. In addition, we investigated endometriosis in various organs, such as the fallopian tube, intestine, ovary, pelvic peritoneum, rectovaginal septum and vagina, uterus, and unspecified locations. Instrument variables associated with LTL, BMI and smoking in GWAS datasets were selected using a stringent set of inclusion criteria. Initially, SNPs with genome-wide significance (p ≤ 5 × 10 −8 ) were chosen. Subsequently, we performed clumping of SNPs by excluding variants in linkage disequilibrium (LD, R 2 > 0.001 and within 10,000 kb). To ensure consistency, all selected SNPs were harmonized to correspond to the same allele for effect estimation. The strength of the instrumental variables (IVs) was assessed using the F statistic, with a threshold of less than 10 defining weak instruments that were subsequently excluded [ 12 ]. Furthermore, we eliminated palindromic SNPs that could introduce uncertainty in determining the effect allele in the exposure GWASs. After this rigorous screening process, the remaining SNPs were deemed eligible instrumental variables. S2 Table provides comprehensive information on all instrumental variables used in this study, and S3 Table gives summary information of these instrument variables. We utilized various statistical techniques, such as inverse variance weighting (IVW), the weighted median (WM), MR-Egger, the weighted model, and the simple model, to assess the causal association between exposure (LTL) and outcome (infertility). The IVW method was chosen as the primary statistical analysis approach, with the random effects model used in the presence of heterogeneity [ 13 ]. Various sensitivity analyses were conducted to assess the robustness of the findings. These included tests for heterogeneity, pleiotropy, leave-one-out analyses, and the Mendelian randomization pleiotropy residual sum and outlier (MR-PRESSO) method [ 14 ]. Heterogeneity was evaluated using Cochran’s Q statistic, with a significance level of p < 0.05 indicating the presence of heterogeneity [ 15 ]. Directional pleiotropy was assessed using the MR-Egger intercept analysis, with a p-value greater than 0.05 suggesting the absence of pleiotropy [ 16 ]. The MR-PRESSO method was employed to identify and remove outliers that may have influenced the results. Additionally, the leave-one-out test involved systematically removing single nucleotide polymorphisms (SNPs) one by one and assessing the stability of the results. Stable and reliable causal relationships were indicated if the remaining results did not exhibit significant changes. To address potential confounding effects stemming from BMI) and smoking, we utilized the MVMR to investigate the influence of BMI, leukocyte telomere length (LTL), and smoking on the occurrence of infertility, sperm abnormalities, and endometriosis. The results were reported as odds ratios (ORs) with corresponding 95% confidence intervals. Two-sided p-values were used, and statistical significance was determined at p < 0.05. All statistical analyses were performed using the "Two-Sample MR" packages in R software (version 4.3.1).

(PDF) (XLSX) (XLSX) (XLSX) (XLSX) (XLSX) (XLSX) (XLSX)