Epigenetic determinants of reproductive potential augment the predictive ability of the semen analysis.

Full semen parameters and male demographic information can be found in Supplemental Table 1 (available online). Of the 1344 men analyzed for this study, 12.0% had a sperm concentration of less than 15 million/mL, 14.3% had a total motile count (TMC) less than 20 million, and 65.5% had morphology results greater than or equal to 4.0%. An overview of the female partner demographics can be found in Supplemental Table 2 (available online). In an additional analysis of IUI-treated fertility outcomes, 21.1% of men were removed from the study because their partners had known female infertility factors. We performed a gene promoter methylation variability analysis on sperm samples from 1344 men seeking infertility care to quantify how many genes had dysregulated promoters. Thresholds for these signatures of irregular methylation were set on the basis of fertile controls. Distribution of dysregulated promoters among the infertile men displayed an average of 12.7 dysregulated promoters and a median of 9.0 dysregulated promoters ( Fig. 1A ). We performed regression analyses between the number of dysregulated promoters and several factors such as male BMI, age, TMC, concentration, and morphology of sperm, and found no meaningful relationships to the number of dysregulated promoters ( Fig. 1B and C , Supplemental Figure 3 , available online). Previous data in other disease types had shown that increased promoter dysregulation is associated with pathologic phenotypes ( 18 ). We sought to understand the relationship between the number of dysregulated promoters and clinical outcomes for different fertility treatments while controlling for female infertility factors. To do this, the top and bottom 10 th percentile of dysregulated promoters were identified to the nearest integer. The top 10th percentile included men with ≥22 dysregulated promoters (n = 140) and was designated as the “Poor” sperm quality group. The bottom 10th percentile of dysregulated promoters included men with ≤3 dysregulated promoters (n = 114) and was designated as the “Excellent” sperm quality group. All remaining men with >3 and <22 dysregulated promoters (n = 1090) were designated as the “Normal” sperm quality group. Supplemental Table 1 contains the semen parameters and demographics associated with each group. When creating these 3 distinct groups, we identified a statistically significant enrichment of men with low TMC in the Poor group compared with the Excellent group. Analysis of the percentage of live births and pregnancies of couples undergoing IUI (n = 544) showed a statistically significant difference between the Excellent and Poor sperm quality groups, as well as between the Average and Poor sperm quality groups ( Fig. 2A ). Similar pregnancy and live birth results were seen for couples whose female partners had no female infertility factors (n = 344) ( Fig. 2B ), indicating a relationship between sperm DNA methylation promoter dysregulation and fertility potential. A permutation analysis was completed to determine if the differences seen in live birth rates could be due to random chance. We found the real difference in live birth rates to be in the 99.5 percentile of permutations, indicating a very low probability that these results are due to chance ( Supplemental Figure 2 ). When completing the same analysis for men undergoing IVF (primarily with ICSI), we saw no statistical difference between any of the sperm quality groups (Excellent, Average, or Poor), with or without controlling for female factors ( Figs. 2C and D ). These data together show that the analysis of accumulated dysregulated promoters, termed Epigenetic SpermQT, appears to identify men with lower fertility potential for IUI procedures. Additionally, IVF appears to overcome sperm quality issues identified with SpermQT. Even with enrichment of low TMC in the Poor sperm quality group ( Supplemental Table 1 ), 77.8% of men with a Poor SpermQT result had a TMC ≥20 M and 75.6% had both a TMC ≥20 M and a sperm concentration ≥15 M/mL, suggesting that SpermQT identifies a new subset of men with low fertility potential that would have been missed by semen analysis alone. Interestingly, in these data, TMC alone is not statistically predictive of live birth rates in individuals undergoing IUI ( Fig. 3A ). However, when combined with SpermQT, the integration of both assessments provides a statistically significant prediction of live births from IUI ( Fig. 3C ). Similar to SpermQT alone, the combination of TMC with SpermQT is not predictive of outcomes of IVF (primarily with ICSI) ( Fig. 3D ). With both tests identifying a unique subset of subfertile men, the combination of SpermQT with TMC could provide a more comprehensive assessment of male fertility, identifying previously undiagnosed men as subfertile and helping physicians guide their treatment recommendations more accurately. Analysis of the dysregulated promoters in 1233 target gene promoters across all samples with a Poor SpermQT score revealed a broad distribution of irregular methylation ( Fig. 4 ). This reflects the biological complexity and heterogeneity between male infertility patients. Ten gene promoters were epigenetically dysregulated in more than 20% of samples ( Fig. 4 inset) with 3 dysregulated genes (ACTR5, ASGR1, and HSD17B7) present in more than 30% of samples (36.2%, 33.3%, and 31.2%, respectively). ACTR5 is an Actin Repair Protein known for UV-damage repair and double-strand break repair and has been previously identified to be highly expressed in the testis ( 22 ). ASGR1 is a protein subunit of the asialoglycoprotein receptor largely known for glycoprotein homeostasis in the liver. However, ASGR1 has also been identified as enriched in early and late-stage spermatids because of glycoproteins’ essential role in sperm development and function ( 21 ). HSD17B7 is an enzyme involved in estrogen and androgen metabolism as well as cholesterol biosynthesis. Deletion of HSD17B7 has been shown to cause reduced testosterone production and early fetal death in mice ( 23 , 24 ). Additional analysis of the distribution of dysregulated promoters can be found in Supplemental Figure 4 (available online). We also performed an ontology analysis ( 21 ) on these 1233 promoter regions to gain more insight into the possible pathways and mechanisms contributing to this phenotype of subfertility. Analysis showed 4 terms were enriched in the GO Molecular Function ontology: peptidase regulator activity, peptidase inhibitor activity, endopeptidase regulator activity, and endopeptidase inhibitor activity ( Supplemental Figure 1 ). Interestingly, it has been hypothesized for years that these enzymes could play a role in the physiology of sperm ( 25 ), with more recent work in mice and humans supporting this hypothesis ( 26 – 30 ).

Sperm DNA methylation data (Infinium MethylationEPIC Array) from fertile sperm donors were obtained from Miller et al. ( 18 ). Additionally, sperm DNA methylation data from a clinical multi-site National Institutes of Health study of men experiencing infertility were used from previously published data from Jenkins et al. ( 17 ). The trial was approved by the Institutional Review Boards at all study centers and the data coordinating center. Written informed consent was obtained from all participants, and a Data and Safety Monitoring Board provided external oversight. Analysis was completed on 1344 de-identified patient sperm DNA methylation data and clinical outcomes previously published in Jenkins et al. ( 17 ). Outcomes included both live birth and pregnancy data, where pregnancy was determined by either ultrasound or biochemical (human chorionic gonadotropin) assessment. The clinical information of patients was described in detail in the previous publication. The population of couples undergoing intrauterine insemination (IUI) completed a cumulative average of 2–3 cycles during the study. For couples undergoing in vitro fertilization (IVF), a cumulative average of 1–2 embryos were transferred per couple and 76% of fertilization occurred via IVF with intracytoplasmic sperm injection (ICSI) ( 19 ). Additionally, when controlling for female factors, women <35 years old with no prior diagnosis of polycystic ovary syndrome, endometriosis, fibroids, blocked tubes, or diminished ovarian reserves were included. The sperm DNA methylation data were preprocessed as described in Miller et al. ( 18 ) with minor modifications as detailed in Supplemental File 1 (available online). We also removed any sperm samples from analysis that did not have a mean methylation value less than 0.24 of all the cytosine-phosphate guanine dinucleotide beta values in the differentially methylated region of DLK1 as described by Jenkins et al. ( 20 ) (chr14:101,191,893–101,192,913, GRCh37). Jenkins et al. ( 20 ) showed the methylation states of the probes in this region are a good discriminator between sperm and somatic cells. This procedure ensured analyses were performed only on samples containing sperm DNA methylation and not contaminating somatic cell DNA methylation. Gene promoters with the least variable methylation values (n = 1233) in sperm from fertile sperm donors (n = 43) and the corresponding gene promoter variability cutoffs were selected as described in Miller et al. ( 18 ). These promoters and corresponding cutoffs were then used to perform sperm methylation variability analyses on the sperm methylation data from men experiencing infertility (n = 1344). We observed the promoter methylation variability within selected promoters and identified the number of promoters falling outside the prescribed gene methylation promoter cutoffs (i.e., “dysregulated promoters”). These analyses were performed as outlined in Miller et al. ( 18 ) with a minor modification noted in Supplemental File 1 . In addition, an ontology analysis was performed on these 1233 promoter regions using the web application implementation of the GREAT algorithm ( https://great.stanford.edu ) ( 21 ) as shown in Supplemental Figure 1 (available online). We then established thresholds for the number of dysregulated promoters for samples with “Excellent” (≤ 3 dysregulated promoters), “Average” (between 4 and 21 dysregulated promoters), and “Poor” sperm quality (≥22 dysregulated promoters). Two-sided t-tests were subsequently performed on the pregnancy and live birth outcomes of these couples, categorized by sperm quality and the type of infertility treatment received. A permutation analysis (n = 10,000) was performed by shuffling the live birth results of couples receiving IUI treatment, and the live birth rates of couples in the Excellent sperm quality category was compared with those in the Poor sperm quality category ( Supplemental Figure 2 , available online).

Our retrospective analysis supports 2 key findings: First, infertile men in our sample population displayed greater numbers of dysregulated promoters than healthy sperm donors. Second, analysis of these dysregulated promoters can assess IUI treatment outcomes. The lack of a similar correlation for IVF treatment outcomes is encouraging because it suggests that there is no loss of fitness because of dysregulated promoters with the assistance of IVF (primarily with ICSI). Moreover, when paired with the traditional semen analysis, SpermQT could direct clinicians and couples toward the most effective treatment plan. Building on this research will not only improve clinical outcomes, but provide longoverdue guidance for those seeking infertility care.

This study was grounded in the concept that there are multiple biological pathways that may lead to decreased fertility potential, and that these biological pathways likely differ among infertile men. Here, we have defined a threshold for epigenetic stability that serves as an indicator of “healthy” sperm. Once a man’s sperm crosses this threshold, there emerges a phenotype of lower fertility potential. We have shown the development of a DNA methylation assessment in sperm that has a statistically significant association with pregnancy and live birth percentages of couples undergoing IUI treatment. SpermQT’s ability to identify a subset of men that are largely missed by the current standard of care could allow for a more comprehensive assessment of male fertility by healthcare providers. For example, men with Poor sperm quality results may be advised to forgo IUI treatment in favor of IVF with ICSI, saving the patient time and expenses. Male infertility is a complex disorder that will require a combination of assessments for physicians to understand it more effectively. The primarily visual and superficial aspects of the current standard of care (basic semen analysis) are useful but fall short of a comprehensive diagnostic for male infertility. When combined with the initial semen analysis (particularly TMC), SpermQT metrics may provide more guidance and help set expectations for couples seeking infertility care. To ensure these results are conserved across population types, we call for additional independent, prospective studies to further validate the application of this potential diagnostic test. Because this study was conducted with retrospective data, future studies on prospective data are needed for continued validation of this new sperm assessment. One prospective clinical trial using this biomarker is already in progress, furthering this important research ( https://clinicaltrials.gov/study/NCT05966883 ). Additionally, future analysis with sperm RNA and protein analysis is needed to validate differences in gene expression between men with Excellent (or Normal) and Poor SpermQT results. Future work into the relationship of peptidase/endopeptidase regulator activity as well as peptidase/endopeptidase inhibitor activity and male infertility could also prove very fruitful as indicated by the ontology analysis of these 1233 promoters. We believe future studies coupling epigenetic and gene expression patterns as well as chromatin structure will be important to give more insights into the epigenetic underpinnings and overall mechanism of this biomarker. Future knowledge of the mechanism(s) identified by this potential biomarker could also be invaluable in developing future infertility therapeutics. We anticipate that the clinical analysis of sperm DNA methylation will become increasingly relevant, as it has been shown that sperm DNA methylation can change in response to environmental exposures, diet, lifestyle, and medications ( 13 , 31 – 33 ). Future research is in progress on the effects of these types of changes in decreasing the number of dysregulated promoters and improving fertility outcomes.