{"paper_id":"543b8bca-83d3-40bc-a976-c9b34c727098","body_text":"Sperm DNA methylation data (Infinium MethylationEPIC Array) from fertile\nsperm donors were obtained from Miller et al. ( 18 ). Additionally, sperm DNA methylation data from a clinical\nmulti-site National Institutes of Health study of men experiencing infertility\nwere used from previously published data from Jenkins et al. ( 17 ). The trial was approved by the Institutional\nReview Boards at all study centers and the data coordinating center. Written\ninformed consent was obtained from all participants, and a Data and Safety\nMonitoring Board provided external oversight.\nAnalysis was completed on 1344 de-identified patient sperm DNA\nmethylation data and clinical outcomes previously published in Jenkins et al.\n( 17 ). Outcomes included both live\nbirth and pregnancy data, where pregnancy was determined by either ultrasound or\nbiochemical (human chorionic gonadotropin) assessment. The clinical information\nof patients was described in detail in the previous publication. The population\nof couples undergoing intrauterine insemination (IUI) completed a cumulative\naverage of 2–3 cycles during the study. For couples undergoing in vitro\nfertilization (IVF), a cumulative average of 1–2 embryos were transferred\nper couple and 76% of fertilization occurred via IVF with intracytoplasmic sperm\ninjection (ICSI) ( 19 ). Additionally, when\ncontrolling for female factors, women <35 years old with no prior\ndiagnosis of polycystic ovary syndrome, endometriosis, fibroids, blocked tubes,\nor diminished ovarian reserves were included.\nThe sperm DNA methylation data were preprocessed as described in Miller\net al. ( 18 ) with minor modifications as\ndetailed in  Supplemental File\n1  (available online). We also removed any sperm samples from analysis\nthat did not have a mean methylation value less than 0.24 of all the\ncytosine-phosphate guanine dinucleotide beta values in the differentially\nmethylated region of DLK1 as described by Jenkins et al. ( 20 ) (chr14:101,191,893–101,192,913, GRCh37).\nJenkins et al. ( 20 ) showed the\nmethylation states of the probes in this region are a good discriminator between\nsperm and somatic cells. This procedure ensured analyses were performed only on\nsamples containing sperm DNA methylation and not contaminating somatic cell DNA\nmethylation.\nGene promoters with the least variable methylation values (n = 1233) in\nsperm from fertile sperm donors (n = 43) and the corresponding gene promoter\nvariability cutoffs were selected as described in Miller et al. ( 18 ). These promoters and corresponding cutoffs were\nthen used to perform sperm methylation variability analyses on the sperm\nmethylation data from men experiencing infertility (n = 1344). We observed the\npromoter methylation variability within selected promoters and identified the\nnumber of promoters falling outside the prescribed gene methylation promoter\ncutoffs (i.e., “dysregulated promoters”). These analyses were\nperformed as outlined in Miller et al. ( 18 ) with a minor modification noted in  Supplemental File 1 . In addition,\nan ontology analysis was performed on these 1233 promoter regions using the web\napplication implementation of the GREAT algorithm ( https://great.stanford.edu ) ( 21 ) as shown in  Supplemental Figure 1  (available online).\nWe then established thresholds for the number of dysregulated promoters\nfor samples with “Excellent” (≤ 3 dysregulated promoters),\n“Average” (between 4 and 21 dysregulated promoters), and\n“Poor” sperm quality (≥22 dysregulated promoters). Two-sided\nt-tests were subsequently performed on the pregnancy and live birth outcomes of\nthese couples, categorized by sperm quality and the type of infertility\ntreatment received.\nA permutation analysis (n = 10,000) was performed by shuffling the live\nbirth results of couples receiving IUI treatment, and the live birth rates of\ncouples in the Excellent sperm quality category was compared with those in the\nPoor sperm quality category ( Supplemental Figure 2 , available online).\n\nFull semen parameters and male demographic information can be found in\n Supplemental Table\n1  (available online). Of the 1344 men analyzed for this study, 12.0%\nhad a sperm concentration of less than 15 million/mL, 14.3% had a total motile\ncount (TMC) less than 20 million, and 65.5% had morphology results greater than\nor equal to 4.0%. An overview of the female partner demographics can be found in\n Supplemental Table\n2  (available online). In an additional analysis of IUI-treated\nfertility outcomes, 21.1% of men were removed from the study because their\npartners had known female infertility factors.\nWe performed a gene promoter methylation variability analysis on sperm\nsamples from 1344 men seeking infertility care to quantify how many genes had\ndysregulated promoters. Thresholds for these signatures of irregular methylation\nwere set on the basis of fertile controls. Distribution of dysregulated\npromoters among the infertile men displayed an average of 12.7 dysregulated\npromoters and a median of 9.0 dysregulated promoters ( Fig. 1A ). We performed regression analyses between the\nnumber of dysregulated promoters and several factors such as male BMI, age, TMC,\nconcentration, and morphology of sperm, and found no meaningful relationships to\nthe number of dysregulated promoters ( Fig.\n1B  and  C ,  Supplemental Figure 3 , available\nonline).\nPrevious data in other disease types had shown that increased promoter\ndysregulation is associated with pathologic phenotypes ( 18 ). We sought to understand the relationship between\nthe number of dysregulated promoters and clinical outcomes for different\nfertility treatments while controlling for female infertility factors. To do\nthis, the top and bottom 10 th  percentile of dysregulated promoters\nwere identified to the nearest integer. The top 10th percentile included men\nwith ≥22 dysregulated promoters (n = 140) and was designated as the\n“Poor” sperm quality group. The bottom 10th percentile of\ndysregulated promoters included men with ≤3 dysregulated promoters (n =\n114) and was designated as the “Excellent” sperm quality group.\nAll remaining men with >3 and <22 dysregulated promoters (n =\n1090) were designated as the “Normal” sperm quality group.  Supplemental Table 1 \ncontains the semen parameters and demographics associated with each group. When\ncreating these 3 distinct groups, we identified a statistically significant\nenrichment of men with low TMC in the Poor group compared with the Excellent\ngroup.\nAnalysis of the percentage of live births and pregnancies of couples\nundergoing IUI (n = 544) showed a statistically significant difference between\nthe Excellent and Poor sperm quality groups, as well as between the Average and\nPoor sperm quality groups ( Fig. 2A ).\nSimilar pregnancy and live birth results were seen for couples whose female\npartners had no female infertility factors (n = 344) ( Fig. 2B ), indicating a relationship between sperm DNA\nmethylation promoter dysregulation and fertility potential. A permutation\nanalysis was completed to determine if the differences seen in live birth rates\ncould be due to random chance. We found the real difference in live birth rates\nto be in the 99.5 percentile of permutations, indicating a very low probability\nthat these results are due to chance ( Supplemental Figure 2 ).\nWhen completing the same analysis for men undergoing IVF (primarily with\nICSI), we saw no statistical difference between any of the sperm quality groups\n(Excellent, Average, or Poor), with or without controlling for female factors\n( Figs. 2C  and  D ). These data together show that the analysis of\naccumulated dysregulated promoters, termed Epigenetic SpermQT, appears to\nidentify men with lower fertility potential for IUI procedures. Additionally,\nIVF appears to overcome sperm quality issues identified with SpermQT.\nEven with enrichment of low TMC in the Poor sperm quality group ( Supplemental Table 1 ),\n77.8% of men with a Poor SpermQT result had a TMC ≥20 M and 75.6% had\nboth a TMC ≥20 M and a sperm concentration ≥15 M/mL, suggesting\nthat SpermQT identifies a new subset of men with low fertility potential that\nwould have been missed by semen analysis alone. Interestingly, in these data,\nTMC alone is not statistically predictive of live birth rates in individuals\nundergoing IUI ( Fig. 3A ). However, when\ncombined with SpermQT, the integration of both assessments provides a\nstatistically significant prediction of live births from IUI ( Fig. 3C ). Similar to SpermQT alone, the combination of\nTMC with SpermQT is not predictive of outcomes of IVF (primarily with ICSI)\n( Fig. 3D ). With both tests identifying\na unique subset of subfertile men, the combination of SpermQT with TMC could\nprovide a more comprehensive assessment of male fertility, identifying\npreviously undiagnosed men as subfertile and helping physicians guide their\ntreatment recommendations more accurately.\nAnalysis of the dysregulated promoters in 1233 target gene promoters\nacross all samples with a Poor SpermQT score revealed a broad distribution of\nirregular methylation ( Fig. 4 ). This\nreflects the biological complexity and heterogeneity between male infertility\npatients. Ten gene promoters were epigenetically dysregulated in more than 20%\nof samples ( Fig. 4  inset) with 3 dysregulated genes (ACTR5, ASGR1, and HSD17B7)\npresent in more than 30% of samples (36.2%, 33.3%, and 31.2%, respectively).\nACTR5 is an Actin Repair Protein known for UV-damage repair and double-strand\nbreak repair and has been previously identified to be highly expressed in the\ntestis ( 22 ). ASGR1 is a protein subunit\nof the asialoglycoprotein receptor largely known for glycoprotein homeostasis in\nthe liver. However, ASGR1 has also been identified as enriched in early and\nlate-stage spermatids because of glycoproteins’ essential role in sperm\ndevelopment and function ( 21 ). HSD17B7 is\nan enzyme involved in estrogen and androgen metabolism as well as cholesterol\nbiosynthesis. Deletion of HSD17B7 has been shown to cause reduced testosterone\nproduction and early fetal death in mice ( 23 ,  24 ). Additional analysis\nof the distribution of dysregulated promoters can be found in  Supplemental Figure 4  (available\nonline). We also performed an ontology analysis ( 21 ) on these 1233 promoter regions to gain more insight into the\npossible pathways and mechanisms contributing to this phenotype of subfertility.\nAnalysis showed 4 terms were enriched in the GO Molecular Function ontology:\npeptidase regulator activity, peptidase inhibitor activity, endopeptidase\nregulator activity, and endopeptidase inhibitor activity ( Supplemental Figure 1 ).\nInterestingly, it has been hypothesized for years that these enzymes could play\na role in the physiology of sperm ( 25 ),\nwith more recent work in mice and humans supporting this hypothesis ( 26 – 30 ).\n\nThis study was grounded in the concept that there are multiple biological\npathways that may lead to decreased fertility potential, and that these biological\npathways likely differ among infertile men. Here, we have defined a threshold for\nepigenetic stability that serves as an indicator of “healthy” sperm.\nOnce a man’s sperm crosses this threshold, there emerges a phenotype of lower\nfertility potential.\nWe have shown the development of a DNA methylation assessment in sperm that\nhas a statistically significant association with pregnancy and live birth\npercentages of couples undergoing IUI treatment. SpermQT’s ability to\nidentify a subset of men that are largely missed by the current standard of care\ncould allow for a more comprehensive assessment of male fertility by healthcare\nproviders. For example, men with Poor sperm quality results may be advised to forgo\nIUI treatment in favor of IVF with ICSI, saving the patient time and expenses.\nMale infertility is a complex disorder that will require a combination of\nassessments for physicians to understand it more effectively. The primarily visual\nand superficial aspects of the current standard of care (basic semen analysis) are\nuseful but fall short of a comprehensive diagnostic for male infertility. When\ncombined with the initial semen analysis (particularly TMC), SpermQT metrics may\nprovide more guidance and help set expectations for couples seeking infertility\ncare. To ensure these results are conserved across population types, we call for\nadditional independent, prospective studies to further validate the application of\nthis potential diagnostic test. Because this study was conducted with retrospective\ndata, future studies on prospective data are needed for continued validation of this\nnew sperm assessment. One prospective clinical trial using this biomarker is already\nin progress, furthering this important research ( https://clinicaltrials.gov/study/NCT05966883 ). Additionally, future\nanalysis with sperm RNA and protein analysis is needed to validate differences in\ngene expression between men with Excellent (or Normal) and Poor SpermQT results.\nFuture work into the relationship of peptidase/endopeptidase regulator activity as\nwell as peptidase/endopeptidase inhibitor activity and male infertility could also\nprove very fruitful as indicated by the ontology analysis of these 1233 promoters.\nWe believe future studies coupling epigenetic and gene expression patterns as well\nas chromatin structure will be important to give more insights into the epigenetic\nunderpinnings and overall mechanism of this biomarker. Future knowledge of the\nmechanism(s) identified by this potential biomarker could also be invaluable in\ndeveloping future infertility therapeutics.\nWe anticipate that the clinical analysis of sperm DNA methylation will\nbecome increasingly relevant, as it has been shown that sperm DNA methylation can\nchange in response to environmental exposures, diet, lifestyle, and medications\n( 13 ,  31 – 33 ). Future research is\nin progress on the effects of these types of changes in decreasing the number of\ndysregulated promoters and improving fertility outcomes.\n\nOur retrospective analysis supports 2 key findings: First, infertile men in\nour sample population displayed greater numbers of dysregulated promoters than\nhealthy sperm donors. Second, analysis of these dysregulated promoters can assess\nIUI treatment outcomes. The lack of a similar correlation for IVF treatment outcomes\nis encouraging because it suggests that there is no loss of fitness because of\ndysregulated promoters with the assistance of IVF (primarily with ICSI). Moreover,\nwhen paired with the traditional semen analysis, SpermQT could direct clinicians and\ncouples toward the most effective treatment plan. Building on this research will not\nonly improve clinical outcomes, but provide longoverdue guidance for those seeking\ninfertility care.","source_license":"CC-BY-4.0","license_restricted":false}