The bull sperm epigenome: role in male fertility and transmission to the offspring

Abstract

Extensive DNA methylation and small non-coding RNA modifications occur during male germline differentiation, which are essential for normal spermatogenesis and sperm maturation. Recent evidence has underscored the role of these key regulators in male fertility. However, due to significant environmental and technical variations across the studies, the identification of epigenetic biomarkers relevant to fertility remains challenging. Nevertheless, recent studies using large cohorts and integrative approaches have highlighted sperm epigenetic signatures that improve fertility prediction. In addition, sperm-borne epigenetic marks have been proposed to influence early embryo development and offspring phenotype, supporting the concept of non-genetic inheritance across generations. In breeding sectors and particularly in cattle, while artificial insemination enables the dissemination of paternal genetics at the populational level, some of the broadly used reproductive biotechnologies may impact the sperm epigenetic profile. Therefore, deciphering sperm epigenetic mechanisms has become an important research topic for enhancing breeding efficiency. This review provides an overview of current knowledge on the bovine epigenetic dynamics in male germ cells, their association with fertility and their potential role in embryo development and offspring phenotype, with a focus on DNA methylation and small non-coding RNAs.

Introduction

Epigenetics refers to the molecular mechanisms, such as DNA methylation, small non-coding RNAs (sncRNAs) and histone post-translational modifications, which alter gene regulation in a DNA sequence-independent fashion and are transmitted to the daughter cells through cell divisions. DNA methylation plays an important role in the structure and stability of the genome when associated with heterochromatin and repetitive elements such as transposable elements (TEs; mobile elements in the genome of viral origin that have accumulated during evolution) and satellites (sequences located near or at the centromeric regions of chromosomes). DNA methylation also regulates transcription and is dynamic as a function of cell type, developmental stage, the animal’s physiology or the environment. Finally, DNA methylation is involved in X inactivation and the genomic imprinting of genes expressed in a mono-allelic manner depending on the parental origin of the allele, a phenomenon that is essential for harmonious growth of the fetus [1]. Genomic DNA is wrapped around octamers of histones to form the nucleosome. Histones contain N-terminal tails targeted by different types of modifications, such as acetylation and methylation. These modifications affect different amino acids, producing dozens of post-translational variants with different functional roles. The addition and removal of these modifications are highly flexible processes that directly affect the accessibility of the transcription machinery to genomic DNA, and hence the activation or repression of gene expression. The combinatorial nature of the different histone marks, together with DNA methylation and the presence of certain transcription factors or RNA polymerase, define specific chromatin states associated with specific transcriptional states. These chromatin states are transmitted to daughter cells, thus ensuring the continuity of cell identity through mitosis [2]. Finally, sncRNAs play an important role in post-transcriptional regulations and their expression is tightly regulated in a cell type-specific manner. The sncRNA subfamilies include sequences shorter than 200 nucleotides, such as the extensively studied microRNAs (miRNAs) [3], small interfering RNAs (siRNAs), PIWI-interacting RNAs (piRNAs) and fragments of transfer RNAs and ribosomal RNAs, referred to as tsRNAs and rsRNAs or tRFs and rRFs, respectively.

Sperm cells are transcriptionally inactive and represent the ultimate form of male germ cell differentiation. Their fate is to survive outside the organism and contribute to a new individual after fertilization of an oocyte. To support these functions, the epigenome of spermatozoa, acquired during the differentiation of male germ cells (MGCs) into mature cells, is unique [4, 5]. The differentiation of MGCs begins early during fetal life and is a finely regulated process orchestrated by extensive epigenetic remodeling and characterized by successive waves of transcription, with distinct groups of genes activated at different stages. This specific transcriptional program is silenced at the end of spermatogenesis, when most histones are replaced by protamines, resulting in tight compaction of the sperm nucleus and protection of the paternal genetic material during the transit through the female genital tract.

Reproductive management in livestock is one of the major angular stones of breeding programs. Initially developed to renew the population, reproductive biotechnologies, such as artificial insemination (AI), today remain the main drivers of genetic progress dissemination. In cattle and especially dairy breeds, where bull semen is widely used for AI, the use of less fertile ejaculates or bulls reduces the global efficiency of breeding programs, involving more resources to reach the expected results. Fertility controls, through semen quality assessments, have particularly become a challenge for breeding companies. These sperm functional measurements need to integrate molecular level information since several selection and breeding practices, such as hormonal treatments of the mothers, embryo technologies or the hastened growth and puberty of male calves [6], may interfere with proper establishment of the sperm epigenome. Additionally, the selected AI bulls are usually obtained from high breeding value sires and high-producing dairy cows. These cows are more likely to experience a negative energy balance in the event of concurrent lactation and gestation, which may lead to an unfavorable in utero environment for the developing fetus [7].

Because bull semen has a widespread diffusion potential, with hundreds of AIs being performed per batch and potentially hundreds of offsprings being generated, it is important to understand the role of epigenetic mechanisms in germ line differentiation and sperm maturation, male fertility and offspring phenotype construction. The goal of this review is therefore to provide an overview of the epigenome of MGCs in mammals, its variation according to bull fertility and its potential transmission to next generation, with a focus on DNA methylation and sncRNA expression.

Epigenetic remodeling during male germ cell differentiation and maturation

Given the crucial role of DNA methylation and sncRNAs in the regulation of gene expression, their involvement in MGC differentiation has been the focus of extensive investigation in model species and humans, but little information is available for cattle. DNA methylation is erased throughout the genome when primordial germ cells colonize the gonad (which occurs at 30 to 50 days post fertilization in bovine). DNA methylation erasure involves mechanisms that are both passive (through cell division and the absence of maintenance activity) and active (through the generation of 5 hydroxymethylcytosine (5hmC)). This erasure is not total: some genomic regions retain methylation to a degree that differs according to species [8]. In parallel with DNA demethylation, levels of repressive histone marks rise, which prevents the initiation of massive transcriptional activity following genome hypomethylation [8]. The loss of 5-methylcytosine (5meC) and then of repressive histone marks at specific loci, as well as the gain in 5hmC, are all essential for the expression of germline differentiation genes [9].

Following this first phase of massive demethylation, a re-methylation of genomic DNA occurs in prospermatogonia. This de novo DNA re-methylation is necessary for MGC differentiation. Indeed, mice lacking de novo methylation enzymes Dnmt3a [10, 11], Dnmt3l [12] and the rodent-specific Dnmt3c [13] are sterile due to meiotic arrest or lack of commitment into spermatogenesis. In humans as well, polymorphisms in DNMT genes have been associated with abnormal sperm parameters and male infertility [14]. The piRNAs contribute to de novo DNA methylation through their role in TE silencing, which is crucial to avoid invasion of the germline genome and propagation to the next generation. To achieve this silencing, piRNAs displaying partial sequence homology with TEs guide the recruitment of de novo DNA methylation and chromatin remodeling machineries towards nascent TE transcripts [15]. Although most 5meC in MGCs is acquired during life in utero, it appears that modifications still occur after birth [16]. Recent progress in whole genome DNA methylation analyses has revealed that DNA methylation remains highly dynamic at all stages of MGC differentiation in rats [17] and during human spermatogenesis [18]. In humans [19] and mice [20], a transient demethylation event is observed just before meiosis, which especially targets recombination hotspots and is required for meiotic recombination [19]. DNA methylation changes in sperm have also been reported during the transit through epididymis, including in the bull [21]; however, these changes may be related to extracellular DNA present in the lumen of caput epididymis and contaminated sperm DNA preparations [22, 23].

As largely supported by data in mice, the genome-wide erasure and re-apposition of DNA methylation is an important window of epigenetic plasticity that can be altered by deleterious environmental conditions. The living conditions of the mother may affect the reprogramming of MGCs and the sperm methylome in adulthood, with possible physiological effects on reproductive outcomes [24] and on the metabolism of the next generation [25]. In sheep, nutritional stress during pregnancy alters the DNA methylation landscape and the functional parameters of offspring spermatozoa [26]. The dynamics of de novo DNA methylation are not yet understood in male cattle; however, studies on other species suggest that environmental control during gestation (and particularly the diet of high-producing dams) may also prove crucial to ensure the proper differentiation of MGCs, optimal fertility traits and an adequate sperm methylome throughout adulthood. After birth in cattle, DNA methylation is still sensitive to environmental factors since the sperm methylome retains a memory of the nutrition offered during the first months of life [27, 28], as reported for other species [29].

Small non-coding RNAs also participate in epigenetic regulations during MGC differentiation and maturation. Postnatal inactivation of DICER, key miRNAs such as miR-34b/c or miR-449, or proteins from the piRNA biogenesis (Miwi, Mili, Tdrd1, Tdrd9, Mvh, Mov10L1, Maelstrom, Gasz, Pld6, Spocd1) led to smaller testis size in adulthood as well as severe defects in meiosis, sperm morphology including abnormalities in elongation, chromatin organization and compaction, defective flagellum formation, ultimately leading to sterility [30,31,32,33,34,35,36,37,38].

Bovine ejaculated sperm convey several sncRNA families including piRNAs, miRNAs, tRFs and rRFs [39,40,41,42]. Since post-spermiogenesis spermatozoa are transcriptionally inactive, their sncRNAs content was long thought to be stable and exclusively inherited from spermatogenesis. Interestingly, while a part of these sequences indeed remain from spermatogenesis, a significant part of them seem to be acquired during epididymal transit or in contact with seminal plasma. In bulls, the characterization of sperm content collected from distinct regions of the male reproductive tract, including the testis, caput, corpus, and cauda epididymis, as well as ejaculated spermatozoa with or without seminal plasma, has provided valuable insights into the dynamics of sncRNAs along the genital tract. [43]. The most pronounced modification in sncRNA content occurs between testicular and epididymal spermatozoa (Fig. 1). After spermiogenesis, in fully formed spermatozoa located within the testicular parenchyma, piRNAs account for more than 80% of total sncRNAs. Their abundance drastically decreases upon entry into the epididymis, representing half of the total sncRNA expression in the caput and less than 20% in the cauda epididymis.

Fig. 1.

Dynamic abundance of piRNAs, miRNAs and tRFs in spermatozoa from different segments of reproductive tract.

Several hypotheses have been proposed to explain this massive loss of piRNAs, including their degradation in testicular spermatozoa via the ubiquitin–AP/C system, their indirect involvement in the degradation of spermatogenic mRNAs, or their elimination through the cytoplasmic droplet. This dynamic regulation of piRNA expression is not specific to bovine sperm, as similar patterns have been reported in mice at comparable magnitudes [44, 45]. In parallel, post-testicular expression of miRNAs, tRFs and rRFs increases during the sperm journey in male tract. These post-testis acquired-sncRNAs may play specific roles in final sperm maturation, fertilizing ability, or early embryonic development.

Like the sperm DNA methylome, sperm-derived sncRNAs are sensitive to environmental influences occurring during MGC differentiation (Fig. 2).

Fig. 2.

Sensitive periods for epigenetic modifications in the bull germline. MGC, male germ cells; SSC, spermatogonial stem cells.

Variations of the sperm methylome and sncRNA content with bull field fertility

For many decades semen quality evaluation was based on parameters such as sperm concentration, morphology, progressive and total motility and viable sperm count, exploring the sperm quality as a proxy of sperm fertility. However, this classical approach failed to efficiently predict male fertility. The discovery of epigenetic regulations involved in spermatogenesis has prompted increasing scientific interest in elucidating how these mechanisms contribute to fertility and could represent a valuable source of fertility biomarkers.

Several studies have investigated the association between inter-individual variations of the semen methylome and fertility traits in bulls [46,47,48,49,50, 51]. Overall, the results reported were not concordant in terms of the genes or genomic regions targeted by differential methylation and the magnitude of DNA methylation changes. These inconsistencies may be related to both technical issues (e.g., different approaches and parameters used to generate DNA methylation data and detect differential methylation) and biological issues (different scores used to assess bull fertility, different breeds, high inter-individual variability and small numbers of samples involved in each study). Inadequate sample size has been proposed as a major obstacle to replicating fertility-related DNA methylation signatures in human sperm [52]. Large cohorts are thus required to address the predictive capacity of the sperm methylome in relation to male fertility. In humans, this question was investigated based on the methylation profiles of 127 infertile men undergoing in vitro fertilization and 54 fertile normozoospermic men [53]. Although the study did not include an independent test set, promising results were obtained on this large sample size, with 82% sensitivity of the model (ability to identify infertile men). In cattle, a large Montbéliarde bull cohort (n = 100) was harnessed to implement predictive models using a machine learning approach, resulting in 72% correct fertility predictions from the sperm methylome [54]. In this cohort the confounding effect of genetic polymorphism was limited by filtering putative variants. Interestingly, these results could be replicated in a completely independent cohort of 20 Montbéliarde bulls, suggesting robust fertility prediction with DNA methylation information. Although promising, these results have to be validated on a larger dataset and on different breeds. In addition, understanding the functional involvement of the genic regions containing the identified biomarkers would be of great interest.

Many studies have also focused on the statistical associations between sncRNAs and sperm quality in humans, pigs, rams and cattle [40, 55,56,57]. In the bovine species, several miRNAs have been associated with sperm motility, field fertility with conventional or sexed semen and morphological abnormalities, such as miR-2284, miR-2483, miR-6526, miR-34c, miR-93, miR-106b and miR-100 [40, 58, 59]. The study by Donnellan et al. (2022) identified 13 differentially expressed miRNAs between 10 high- and low-fertility Holstein bulls [60]. Among them, previously reported miRNAs such as miR-146a and miR-2285 were found. Interestingly, the gene pathways targeted by these 13 miRNAs were related to embryonic development and gene expression regulation. In small ruminants, Bodu et al. (2025) highlighted 227 differentially expressed sperm-borne miRNAs between high- and low-fertility rams. Among them, several key miRNAs, such as oar-miR-200 and oar-miR-370-3p, were found to be upregulated in the high-fertility group [61]. Fewer studies have focused on other sncRNA subfamilies, most of them in human assisted reproductive technologies [62, 63]. Interestingly, these studies have highlighted the role of specific tRFs, mainly those derived from the Gly and Glu isoacceptors. In both studies, an increased expression of these tRFs was observed in oligozoospermic patients or in individuals with repeated ICSI failures, suggesting a role of these seminal tRFs in male infertility. Overall, although numerous associations between sncRNAs and sperm quality/fertility have been reported, the lack of consistency in identified biomarkers suggests that their effects are likely context-dependent. This variability may reflect differences in experimental design and environmental conditions. Therefore, sncRNAs should be considered as part of a broader regulatory network rather than as standalone fertility markers.

Approaches based on a unique type of biomarkers may not be sufficiently effective to predict male fertility and multi-omics integration approach may be necessary. This idea has been approached in cattle using 98 Montbéliarde bulls with contrasting fertility values to integrate genetic and epigenetic data (sperm DNA methylation and sncRNAs). Interestingly, several statistical approaches including Logistic Lasso, Random Forest, Gradient Boosting, and Neural Networks, have selected different features from all “omics” layers, illustrating the relevance of the integration approach [64]. This suggests that sperm fertility may not rely on single epigenetic markers but rather on combinatorial patterns.

Impact of the sperm epigenome on embryo development and offspring phenotype

Major epigenetic changes occur after fertilization, supporting the transition from highly specialized gametes to totipotency. Overall 5meC levels fall sharply in embryos during preimplantation development [65,66,67]. DNA methylation erasure is more rapid and important in the bovine paternal genome (which is initially more methylated than the maternal genome) and requires active demethylation mechanisms. At the level of the genomic sequence, regions subject to genomic imprinting are not affected by this wave of DNA demethylation [68, 69]. These dramatic epigenetic changes participate in triggering embryonic genome activation (EGA). Indeed, initially, the genome of the newly fertilized embryo is transcriptionally inactive. Embryo development then depends strictly on the stock of RNA and proteins accumulated in the oocyte. EGA occurs at the 8-cell stage in cattle, when chromatin accessibility reaches a maximum [70], which is coincident with the minimal level of DNA methylation (15%) [68].

The importance of post-testis acquired sncRNAs for embryo development has particularly been highlighted in animal models. In mouse zygotes, inhibition of tRF-Val-CAC, a highly abundant sperm-borne sncRNA, modulates the expression of 505 genes at the 2-cell stage, particularly those involved in chromatin organization, cell-cycle progression, mRNA processing and alternative splicing. The subsequent embryos showed delayed first cleavage, reduced 2-cell progression, and impaired blastocyst formation in a dose-dependent manner [71]. Conine et al. (2018) investigated the role of the sncRNAs contained within epididymal EVs and demonstrated that embryos produced with caput sperm showed alterations in the expression of genes related to RNA-binding proteins as well as chromatin regulators, compared to embryos produced with cauda sperm, leading to developmental failure shortly after implantation [72]. In addition, injection of miRNAs isolated from the cauda into caput-derived zygotes restored normal gene expression. Following up to these findings, Trigg & Conine (2024) showed that mice sperm lacking epididymal miRNAs (particularly the miR-34/449 family and X-linked miR-880 cluster) are unable to fertilize eggs naturally or via in vitro fertilization despite normal production and motility [73]. Again, microinjection of epididymal miRNAs isolated from cauda epididymosomes rescued these defects. These findings provide direct evidence that epididymal-acquired miRNAs are essential for the regulation of early embryonic gene expression.

The sperm epigenetic features conveyed to the embryo are postulated to mediate the intergenerational transmission of non-genetic information that may impact the long-term phenotype of offspring in response to environmental changes affecting the father [74, 75]. Overall, studies on post-fertilization reprogramming in cattle [68] and other species have suggested a limited inheritance of DNA methylation patterns in the paternal genome. Recent work using an epigenome-edited mouse model has shown that a severe imprinting defect at the H19 differentially methylated region in sperm is partially corrected after fertilization, in an H3K9me3-dependent manner [76]. Although centered on one specific imprinted locus, this study suggests that H3K9me3 may help restrict the intergenerational inheritance of sperm-derived DNA methylation defects. On the other hand, individual loci such as imprinted loci, specific subfamilies of TEs, as well as a few genes, are specifically targeted by the DNA methylation maintenance machinery and are faithfully maintained throughout post-fertilization reprogramming in mice [77]. It has recently been demonstrated that engineered DNA methylation signatures and associated phenotypes are stably maintained across several generations in a mouse model, suggesting intergenerational transmission at specific loci [78]. It has also been proposed that the binding of a transcription factor to a CpG-containing site during periods of epigenetic reprogramming may protect that CpG from de novo methylation, thereby ensuring the transmission of a hypomethylated state across generations [79]. DNA methylation patterns related to paternal diet in F0 have been maintained across the F1 and F2 generations in the sheep [80], suggesting that transgenerational inheritance can also be observed in livestock. In outbred species, however, this phenomenon can easily be confounded with genetically-controlled epigenetic effects. Indeed, DNA methylation in sperm is partially controlled by genetic mechanisms, as demonstrated in cattle [81].

Beyond epigenetic inheritance, the sperm epigenome may also modulate the embryo development and offspring phenotype through other mechanisms. Indeed, growing evidence indicates that semen epigenetic alterations resulting from direct or indirect stress effects on the germline during spermatogenesis, epididymal maturation, or storage can affect the phenotype of the resulting offspring, a concept referred to as paternal programming [82, 83]. It has recently been demonstrated in mice that the DNA methylation status in sperm impacts the establishment of chromatin on the paternal genome in early embryos [84]. Aberrant epigenetic patterns in sperm resulting from suboptimal environment or physiology during the reprogramming of MGCs or during spermatogenesis may therefore interfere with embryo chromatin establishment and post-fertilization reprogramming. The resulting embryo may carry subtle molecular, morphological or metabolic defects that could drive long-term effects on the phenotype [24, 25]. This is exemplified by the epigenetic alterations that have been observed in semen collected at a peripubertal age. Embryos produced using such peripubertal semen display subtle modifications to the DNA methylome and transcriptome that particularly affect the genes involved in metabolic functions and protein synthesis [85]. Whether the epigenetic changes observed during the preimplantation development are maintained throughout lifespan and are susceptible to impact the long-term phenotype remains to be investigated.

Sperm-derived sncRNAs are also sensitive to environmental influences. Changes in sperm sncRNA content have been widely reported in response to various stressors and can have consequences on the offspring phenotype. In cattle, sperm exposed to heat shock is associated with a delay in embryonic developmental kinetics and alteration of sperm-borne miRNAs [86]. tRFs also are heavily affected by stress factors and mediate paternal effects. For instance, injecting specific sperm tRFs into fertilized embryos recapitulated the paternal high-fat diet effects on offspring glucose tolerance, while other sncRNA classes did not [87]. The mechanisms through which the content in sperm sncRNAs influence embryo development are yet to be explored. Sperm-borne sncRNAs can directly target maternal mRNAs accumulated in the oocyte, promoting their degradation or modulating their translation during the early stages of development, prior to zygotic genome activation. This notion has been exemplified by Rodgers et al., with the action of 9 sperm miRNAs (miR-29c, miR-30a, miR-30c, miR-32, miR-193-5p, miR-204, miR-375, miR532-3p, and miR-698) on a subset of maternal mRNA and several biological pathways, such as extracellular matrix, collagen organization or cellular to cellular recognition [88]. In addition, these sncRNAs may indirectly influence gene expression by interacting with other epigenetic mechanisms once the embryonic genome is activated. Together, these mechanisms may alter the expression of gene networks involved in key developmental processes such as cell proliferation, lineage specification, and differentiation (Fig. 3). Thus, even the smallest changes in RNA distribution at early embryonic stages could influence the trajectory of development.

Fig. 3.

Potential mechanisms through which sperm sncRNAs and DNA methylation influence embryo development. The sperm sncRNAs can regulate the pool of maternal mRNAs accumulated in the oocyte and/or modulate gene expression in the early embryo following zygotic genome activation. Key developmental pathways involved in cell proliferation, lineage specification, and differentiation may be targeted by these sperm-derived epigenetic marks.

Future directions

In cattle, several selection and breeding practices may interfere with proper establishment of the sperm epigenome of bulls used for AI, leading to epigenetic alterations that can potentially disseminate to many herds and have long-term consequences. Although data about epigenome remodelling during male germ cell differentiation is not yet available in cattle, reports on other species suggest that windows of epigenome sensitivity to environmental factors exist (Fig. 2), during which particular attention should be paid to the bull and its environment in order to maximize the epigenetic potential of sperm. Optimal in utero conditions should first be put in place by monitoring the nutrition, health and welfare of the dam in order to ensure the proper epigenetic reprogramming and development of fetal germ cells. Optimal conditions during the postnatal period, with particular focus on the transition phases, growth and puberty, as well as during the semen production period, should contribute to early and efficient spermatogenesis, the stable production of high-quality semen and the overall fertility of the bull.

In breeding companies, routine semen quality controls are carried out, allowing the identification of most bulls with severe infertility and observable effects on semen functional parameters. Consequently, the difference between marketed fertile and subfertile bulls is subtle, making the accurate prediction of fertility challenging. In the future, large cohorts of bulls that are well-characterized in terms of their genotypes, epigenetic features and fertility will certainly help reduce inter-individual variability and develop accurate models for fertility prediction.

In conclusion, while significant progress has been made in characterizing the bovine sperm epigenome, particularly DNA methylation and sncRNAs, its functional implications for fertility, offspring development and next generation phenotype remain largely unknown. The design of affordable epigenotyping tools, such as the RUMIGEN EpiChip that could be used to profile a large number of animals [89], together with the development of integrated approaches that combine both genetic and epigenetic information, may drive future progress. The discovery of efficient epigenetic biomarkers could transform reproductive management, improving both fertility prediction and long-term herd performance in cattle.

Conflict of interests

The authors have nothing to declare.

Funding Statement

We thank all INRAE and Eliance members of our team, our collaborators, our funders Horizon Europe, APIS-GENE, ANR and the PHASE division, as well as the breeding companies who support our work. The authors gratefully acknowledge Dr. Pritha Dey for reviewing the manuscript and for her valuable contribution to improving its grammatical quality. The publication fee of this article was covered by JSPS KAKENHI Grant Number 22HP2009.