Chan et al. Clinical Epigenetics (2023) 15:5 Clinical Epigenetics https://doi.org/10.1186/s13148-022-01417-1 RESEARCH Open Access Sperm DNA methylome abnormalities occur both pre‑ and post‑treatment in men with Hodgkin disease and testicular cancer Donovan Chan1, Kathleen Oros Klein2, Antoni Riera‑Escamilla3, Csilla Krausz3,4, Cristian O’Flaherty1,5,6, Peter Chan1,7, Bernard Robaire6,8 and Jacquetta M. Trasler1,6,9* Abstract Background Combination chemotherapy has contributed to increased survival from Hodgkin disease (HD) and testicular cancer (TC). However, questions concerning the quality of spermatozoa after treatment have arisen. While studies have shown evidence of DNA damage and aneuploidy in spermatozoa years following anticancer treatment, the sperm epigenome has received little attention. Our objectives here were to determine the impact of HD and TC, as well as their treatments, on sperm DNA methylation. Semen samples were collected from community controls (CC) and from men undergoing treatment for HD or TC, both before initiation of chemotherapy and at multiple times post- treatment. Sperm DNA methylation was assessed using genome-wide and locus-specific approaches. Results Imprinted gene methylation was not affected in the sperm of HD or TC men, before or after treatment. Prior to treatment, using Illumina HumanMethylation450 BeadChip (450 K) arrays, a subset of 500 probes was able to distinguish sperm samples from TC, HD and CC subjects; differences between groups persisted post-treatment. Comparing altered sperm methylation between HD or TC patients versus CC men, twice as many sites were affected in TC versus HD men; for both groups, the most affected CpGs were hypomethylated. For TC patients, the promoter region of GDF2 contained the largest region of differential methylation. To assess alterations in DNA methylation over time/post-chemotherapy, serial samples from individual patients were compared. With restriction landmark genome scanning and 450 K array analyses, some patients who underwent chemotherapy showed increased alterations in DNA methylation, up to 2 to 3 years post-treatment, when compared to the CC cohort. Similarly, a higher-resolution human sperm-specific assay that includes assessment of environmentally sensitive regions, or “dynamic sites,” also demonstrated persistently altered sperm DNA methylation in cancer patients post-treatment and suggested prefer‑ ential susceptibility of “dynamic” CpG sites. Conclusions Distinct sperm DNA methylation signatures were present pre-treatment in men with HD and TC and may help explain increases in birth defects reported in recent clinical studies. Epigenetic defects in spermatozoa of some cancer survivors were evident even up to 2 years post-treatment. Abnormalities in the sperm epigenome both pre- and post-chemotherapy may contribute to detrimental effects on future reproductive health. Keywords Sperm, DNA methylation, Chemotherapy, Testicular cancer, Hodgkin disease *Correspondence: Jacquetta M. Trasler jacquetta.trasler@mcgill.ca Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Chan et al. Clinical Epigenetics (2023) 15:5 Page 2 of 19 Background DNA methylation, the most well-studied epigenetic Therapeutic advances in the field of oncology have mark, involves the covalent addition of a methyl group allowed for an increase in the number of cancer survi- onto the cytosine ring. This modification occurs at 80% of vors. However, the treatments used for effectively cur- CpG dinucleotides in the human genome and is involved ing patients of their disease do not target cancer cells in a number of cellular processes including transcrip- alone, but also adversely affect normal cells. Male germ tional repression, X-inactivation and chromatin stability. cells are particularly susceptible to anticancer drugs due Proper establishment of genomic imprinting, the mono- to the ongoing division of spermatogonial stem cells and allelic expression of genes in a parent-of-origin man- continued production of mature spermatozoa following ner, also involves DNA methylation, with key events of puberty; any possible negative impacts to the sperm cell demethylation and remethylation occurring in develop- may have implications for the next generation. Of par- ing germ cells. Imprinted genes have important roles in ticular interest are the survivors of Hodgkin disease (HD) embryonic growth and development, placental function and testicular cancer (TC), since they are often diagnosed and postnatal behavior [18–21]. at a young age. With a greater than 95% 5-year survival The importance of DNA methylation is underscored rate of adolescents and young adults for these two dis- by the role this epigenetic modification plays in different eases [1, 2], concerns about fertility and the possibility of disease states. Abnormal expression of imprinted genes parenthood are particularly relevant [3, 4]. has been implicated in several developmental disorders There are a number of studies showing genetic defects including the Beckwith–Wiedemann, Silver Russell and in spermatozoa, as well as more recent evidence of epige- Angelman syndromes [22, 23]. Aberrant DNA meth- netic abnormalities, following chemotherapy treatments ylation has also been implicated in human infertility but [5, 6]; with this come questions regarding the possible remains controversial; altered methylation of imprinted impact of exposed germ cells to future generations. In a loci has been reported in some studies examining human previous study, we reported increases in aneuploidy in spermatozoa from men presenting with infertility and the spermatozoa of HD and TC patients [7]. In the same decreased sperm counts [24, 25], while other found no cohort, increased sperm DNA damage, altered chroma- such associations [26]. Furthermore, genome-wide hypo- tin quality and decreased DNA compaction were found methylation, as well as abnormal methylation of onco- [8–10]. In these studies, abnormalities were found prior genes and/or tumor suppressor genes, are all hallmarks to the onset of treatment and remained up to 24 months of cancer progression [27, 28]. after the termination of chemotherapy. Interestingly, With roles in cancer, as well as evidence of DNA meth- studies have demonstrated that the risk of fathering off- ylation alterations associated with infertility and inter- spring with a major congenital abnormality is increased generational effects [29, 30], our aim was to assess the in men with a history of cancer, even if the children were sperm DNA methylation patterns in a cohort of HD conceived prior to diagnosis and treatment, and this and TC patients as well as a cohort of community con- increase does not seem to diminish even years after treat- trol subjects (CC). The present study was designed to ment was given [11–13]. Thus, it is possible that both the investigate the impact of HD and TC, as well as their cancer and its treatment may adversely affect male germ treatments, on sperm DNA methylation patterns. Using cells. several approaches to study genome-wide and locus-spe- Animal models can help separate out effects of antican- cific DNA methylation, our goals were to determine 1) cer drug treatment from those of the underlying cancers whether abnormal germ cell DNA methylation patterns being treated in humans. In one study, in which male rats were present in patients prior to initiation of chemother- were treated using similar drug regimens to those used in apy and 2) if, within the same patient, patterns of sperm human TC patients, survival of offspring sired from these DNA methylation were altered as a result of their cancer rats was found to be significantly reduced [14]. As well, treatment, as well as over time post-treatment. there were increased DNA damage [15] and altered gene expression [16] in germ cells post-treatment. In the same Results model, altered sperm DNA methylation was detected Following data processing and quality control, DNA immediately after the termination of combination chem- methylation data generated using the Illumina Human- otherapy treatments [17]. These studies indicate that Methylation 450 K BeadChip (450 K) were available for along with persistent genetic and/or DNA damage in 7 CC, 7 HD and 6 TC patients. Table 1 provides patient sperm, anticancer drugs used in human TC can result in characteristics of these subjects at baseline, prior to sperm epigenetic alterations, in a rodent model, where initiation of chemotherapy. No significant differences there were also detrimental effects of the paternal treat- between the three cohorts were observed in any of the ment on the offspring. quantitative measures. Chan et al. Clinical Epigenetics (2023) 15:5 Page 3 of 19 Table 1 Characteristics at baseline, prior to chemotherapy initiation, of the Montreal patient cohort Measure CC HD TC n 7 7 6 Age (years) 27.00 ± 2.29 25.00 ± 1.72 28.83 ± 1.14 Smokers 0 1 1 Sperm Conc. (millions/mL) 31.29 ± 5.16 91.29 ± 28.79 44.50 ± 10.46 Fast + Slow % Motility 48.29 ± 10.89 44.43 ± 1-.24 49.50 ± 8.99 Treatment N/A ABVD Non-seminoma—BEP Mean ± SEM CC community control, HD Hodgkin disease, TC testicular cancer, N/A not applicable, ABVD doxorubicin, bleomycin, vinblastine, dacarbazine, BEP bleomycin, etoposide, cisplatin Imprinted gene DNA methylation is not affected by cancer 450 K array, several probes are located within the ger- or chemotherapy treatments mline imprinting control regions (ICRs) of numerous As key events of erasure and reestablishment of DNA imprinted genes. We therefore extracted and averaged methylation of imprinted genes occur in male germ cells the beta-values/methylation of the probes within each during their development, we first examined this cat- ICR of individual imprinted genes from all the partici- egory of genes as a preliminary indication of the impact pants prior to and following treatment (Fig. 1 and Addi- of HD and TC and their treatment on sperm. With the tional file 1: Table S1). We examined methylation of the Fig. 1 Sperm DNA methylation of imprinted genes is not affected by time (CC men), disease (HD or TC men) or treatments (HD or TC men). Methylation (beta-value) at the imprinting control regions of A H19, B MEST, C KCNQ1OT1 and D SNRPN, examining samples at baseline (prior to initiation of treatment; solid bars) and post-time or post-treatment (hatched bars) times points. E Clustering of the average beta-values across the ICR of the different imprinted genes analyzed using the 450 K array Chan et al. Clinical Epigenetics (2023) 15:5 Page 4 of 19 ICRs of H19, MEST, KCNQ1OT1 and SNRPN, genes Sperm DNA methylation differs between control previously reported to be altered in male infertility and/ and cancer groups prior to treatment or regions known to have roles in imprinting disorders. Sperm samples were examined prior to chemotherapy For all four genes, there were no differences between the to determine if the underlying cancer had an impact on baseline samples for the CC, HD and TC groups (Fig. 1A- the sperm DNA methylome. The baseline samples were D, solid bars). Thus, for all pre-treatment sperm sam- first studied using the 450 K arrays. Of the over 450,000 ples, high levels of methylation (> 85%) were found at the probes found on the array, we removed probes with mul- paternally methylated ICR of H19, while low levels (< 5%) tiple mapping to locations, as well as probes contain- were observed for the ICRs of the maternally methyl- ing single nucleotide polymorphisms at a minor allele ated genes MEST, KCNQ1OT1 and SNRPN. Similarly, frequency of greater than 1%. Hierarchical clustering of at all times post-chemotherapy treatment, normal lev- the beta-values/methylation for the remaining 385,553 els of methylation/beta-value (Fig. 1A–D, hatched bars) probes was not able to differentiate the cancer cohorts were found. Sperm DNA methylation levels of other from that of CC (Fig. 2A). Analysis of variance (ANOVA) imprinted genes found on the 450 K arrays also appeared was then applied to analyze the differences among the normal and clustering based solely on imprinted gene group means in the three different cohorts; this resulted methylation from all the samples was not able to distin- in 11,525 probes with p values < 0.05 (Additional file 1: guish the CC, HD and TC groups (Fig. 1E). In conclu- Table S2). Hierarchical clustering from these sites was sion, imprinted gene methylation was not affected in the able to distinguish TC patients and, to a lesser extent, sperm of HD or TC patients either before or after anti- the HD patients from CC (Fig. 2B). With clustering of cancer drug therapy. the top 500 most significant probes, we were able to dif- ferentiate perfectly the three cohorts from one another (Fig. 2C). To ensure that these results did not happen Fig. 2 Sperm DNA methylation is capable distinguishing patients according to cancer diagnosis. Hierarchical cluster of 450 K probes at baseline, prior to chemotherapy treatment, for A all probes, B significant probes discovered by three-group ANOVA and C top 500 most significant probes. D Significant probes/sites discovered were merged to discover differentially methylated regions (DMRs), with E the top 10 DMRs containing the greatest number of merged probes/sites shown Chan et al. Clinical Epigenetics (2023) 15:5 Page 5 of 19 by chance, we randomly selected the same number of used to look at gene ontology (GO) focusing on biologi- probes found to be significant following the three-group cal processes using the DMRs; the top 20 most significant ANOVA and observed no similar grouping of samples terms are presented in Additional file 2: Fig. S2A, and all (Additional file 2: Fig. S1). These results demonstrate that significant terms can be found in Additional file 3. Many a sperm DNA methylation signature, consisting of a sub- terms pertaining to nervous system/synaptic signals as set of 450 K CpG sites, was able to distinguish HD and well as gene expression were discovered. TC patients and CC subjects. The DMRs were then sorted according to the number While probes represent individual CpG sites, we were of ANOVA significant sites, and the top 10 are listed interested to ascertain whether larger regions of altered (Fig. 2E). The large region found within the promoter methylation existed. To do so, we merged the signifi- and intronic elements of the transcript variant 3, of the cantly altered sites found within 1 kb of each other in spondin 2, SPON2, gene and the long non-coding RNA order to obtain differentially methylated regions (DMRs). LOC100130872 (Fig. 3A) contains 20 sites (Fig. 3B, C, Of the 11,525 total sites discovered by ANOVA, a large shaded box). Differential methylation can be observed majority of the sites remained isolated: 9248 single sites, between the three cohorts, with HD and TC patients representing 80.2% of the total sites (Fig. 2D). Despite possessing the highest and lowest averaged beta-value/ this large proportion of isolated CpGs, DMRs were dis- methylation across the entire region, respectively. These covered, resulting in 973 merged regions. TopGO was results suggest that, beyond individual CpG sites, there Fig. 3 Schematic of the largest differentially methylated region (DMR) discovered following three-group ANOVA and post hoc Tukey’s test for HD. A UCSC genome browser view of the region encompassing the long non-coding RNA, LOC100130872, and the SPON2 (Spondin 2) gene, with tracks showing location of 450 K probes (450 K CpGs) and all CpGs (CpG sites) within the region. Graphs of the average beta-value from each patient group are depicted, plotting the sites according to their B relative position within the genome and C separated to indicate SEM and significant CpG sites as discovered following post hoc testing. Highlighted areas in B and C indicate the DMR discovered following three-group ANOVA, while * indicates significant differences in post hoc Tukey’s testing. Mean ± SEM, *p < 0.05 Chan et al. Clinical Epigenetics (2023) 15:5 Page 6 of 19 are regions in the genome that are able to differentiate Hodgkin disease‑ and testicular cancer‑specific sperm DNA CC, HD and TC subjects using their sperm DNA meth- methylation profiles ylation patterns. Next, we used the 450 K data to determine whether there DNA methylation of the same baseline sperm sam- were sites of differential methylation between specific ples was examined with a second approach, restriction cancer types (HD or TC) when compared with the CC landmark genomic scanning (RLGS), a technique that group. Using post hoc Tukey’s testing, following our ini- examines approximately 3000 loci in the human genome. tial ANOVA analysis, a total of 3265 significantly altered RLGS produces large two-dimensional gels where each probes were found between the HD and CC cohorts locus is represented by a spot (Additional file 2: Fig. (Fig. 4A and Additional file 1: Table S3); more than twice S3A); by overlapping gels, we are able to observe differ- the number of probes, 7760, were discovered for the TC ences in methylation by the appearance (hypomethyla- cohort (Fig. 4C and Additional file 1: Table S4). For both tion), disappearance (completely hypermethylation) or cancer types, the majority of the differentially methylated changes in spot intensity. Using a single CC patient as a sites demonstrated hypomethylation; 2277 sites (69.7%) reference (CC7), we compared all other sperm baseline showed lower methylation in the HD cohort, while 4904 samples from all other patients and found that many sites (63.2%) were found in the TC cohort as compared to loci were altered in methylation (Additional file 2: Fig. the CC. S3B); however, no cancer-specific locus was observed Similar to the analyses above, cancer-specific DMRs when comparing all the HD and/or TC samples to the were also identified. Once again, the majority of CpGs reference (i.e., locus specifically in cancer cohorts com- were single sites; 3112 and 6661 single sites (represent- pared to CC). As RLGS analysis only allows loci differ- ing 93.3% and 84.2% of total significant sites) for the HD ing in methylation between samples by more than 25% and TC groups, respectively, when compared to the CC to be detected, the results suggest that large differences group (Fig. 4A, C). When compared to the 68 DMRs in methylation do not distinguish the 3 groups from one for the HD cohort, greater than sevenfold more (493 another using this technique. DMRs) were found to be different between the TC and CC subjects. When looking at all DMRs together or with hyper- and hypomethylated DMRs separately, many bio- logical processes related to immune system/response Fig. 4 Discovery of differentially methylated regions following post hoc Tukey’s testing. Significant probes/sites discovered following post hoc Tukey’s testing between A HD vs. CC and C TC vs. CC patients were merged to discover differentially methylated regions (DMRs). A list of the top 10 DMRs found from B HD and D TC containing the greatest number of merged probes/sites Chan et al. Clinical Epigenetics (2023) 15:5 Page 7 of 19 and development were found to be significantly enriched some subjects in which their post-treatments sample was in the HD cohort (Additional file 2: Fig. S2B and Addi- found to be further away from their baseline sample (dis- tional file 3). For the TC cohort, many GO terms related cussed further below). to development were found to be enriched (Additional Three different genome-wide DNA methylation assays file 2: Fig S2C and Additional file 3); interestingly, biolog- were used to assess the impact of anticancer drug treat- ical processes related to sexual/male reproduction were ment on human sperm. Samples collected from individ- also discovered. ual patients were compared directly to their own baseline A list of the top 10 regions with the most CpG sites is (pre-treatment) sample. For 450 K array data, a cutoff of shown for HD and TC (Fig. 4B and D, respectively). For greater than 0.2 beta-value change (~ 20% methylation) HD DMRs, the largest region, containing the most sites was chosen, similar to the 25% cutoff used for the RLGS (9 sites), was located in the SPON2 gene, in the same approach. Examination of the different time points dem- region found through three-group ANOVA (Fig. 3). One onstrated both hyper- and hypomethylation; the total large region, common to both the HD and TC cohorts, number of sites showing altered methylation for each was found overlapping the non-coding RNA for the individual subject using RLGS or the 450 K array is listed pseudogene brain-abundant, membrane-attached sig- in Additional file 2: Fig. S5B-C, respectively. By RLGS, nal protein 1 pseudogene 1, BASP1P1 (Additional file 2: CC individuals showed few loci with altered methylation Fig. S4A). All 450 K array probes within the vicinity of over time (Fig. 6A, left panel); an average of 12 spots were the pseudogene were found to be affected in at least one changed when looking at samples collected at 6, 12 and of the cancer cohorts, with lower levels of sperm DNA 18 months compared to baseline. For the cancer survi- methylation observed in both compared with CC (Addi- vors, most had an above average number of changes post- tional file 2: Figs. S4B and S4C). chemotherapy (Fig. 6A, right panel); the greatest number For the TC group, the largest region that demonstrated of changes, with 50 (~ fourfold increase), was observed differential methylation was found within the promoter– for a HD patient, 6 months following treatment (HD3-1). transcription start site (TSS) of the growth differentiation By 450 K array, CC individuals demonstrated on average factor 2, GDF2 (Fig. 5A). For 450 K array probes in GDF2, 145 sites with altered methylation at 6, 12 and 18 months 11 out of 12 CpG sites covered were found to differenti- (Fig. 6B, left panel). Compared to this cohort, half of the ate our cohorts through ANOVA (Fig. 5B and C, shaded men undergoing treatment (HD2-1, HD3-1 and TC2-1) box); of those sites, 6 were determined to be altered demonstrated 20-fold higher alterations in DNA meth- between TC and CC cohorts following post hoc Tukey’s ylation at 6 months post-chemotherapy (Fig. 6B, right testing (Fig. 5C). The altered DNA methylation was vali- panel). While most patients showed similar levels to CC dated through several bisulfite pyrosequencing assays, patients at later time points, one HD patient still dem- which examined a total of 21 CpGs (Fig. 5D). All 21 CpGs onstrated altered sperm DNA methylation at 12 and examined showed decreased methylation (~ 20%) in TC 24 months (HD2-2 and 4). Samples where an increased patients compared to CC (Fig. 5E). Taken together, these number of changes were observed are the same sam- results suggest that cancer-specific CpG sites and regions ples that were found to be further away in the PCA from exist for HD and TC patients as compared to CC. While their baseline timepoint (Additional file 2: Fig. S5A). the 450 K array examines only a limited number of CpG Taken together, our initial results indicate that in some sites within a given genomic feature, larger areas of dif- cancer survivors higher than pre-treatment DNA meth- ferential methylation may exist that distinguish the differ- ylation defects in sperm are present, even up to 2 years ent cohorts. post-treatment. We recently developed a human sperm-specific Increased DNA methylation alterations in a subset 5-methylcytosine capture sequencing (MCC-Seq) of cancer patients following chemotherapy treatment panel that interrogates 3.18 million CpG sites across Next, we examined the effect of chemotherapy on sperm the genome [31]. The panel targets all sites on Illumi- DNA methylation. All CC subjects had at least one na’s MethylationEPIC array (850 K), but also examines time point available following their baseline collection. CpGs of intermediate levels of methylation (20–80% Conversely, not every cancer patient had samples avail- methylated) found to possess greater variability in able for analysis with the 450 K array due to insufficient human sperm DNA and be susceptible to environmen- spermatozoa/DNA post-treatments (HD: 4 patients; TC: tal effects; we refer to the latter as dynamic CpG sites. 3 patients). Principal component analysis (PCA) using We hypothesized that the human sperm-specific MCC- the data from all probes demonstrated that samples Seq panel might be more sensitive than the 450 K array from each individual patient had a tendency to group at detecting altered DNA methylation at later time together (Additional file 2: Fig. S5A); however, there are points post-treatment. The capture panel was used Chan et al. Clinical Epigenetics (2023) 15:5 Page 8 of 19 Fig. 5 Altered methylation within the GDF2 promoter is observed exclusively with TC patients. A UCSC genome browser view of the GDF2 gene (grow differentiation factor 2), with tracks showing location of 450 K probes (450 K CpGs) and all CpGs (CpG sites) within the region. Graphs of the average beta-value from each patient groups are depicted, plotting the sites B according to their relative position with the genome and C separated to indicate SEM and significant CpG sites as discovered following post hoc testing. Highlighted areas in B and C indicate the DMR discovered following three-group ANOVA, while * indicates significant differences in post hoc Tukey’s testing. D A schematic view of CpGs found within promoter region of GDF2 and pyrosequencing PCR and sequencing primers designed to validate altered methylation. CpG sites analyzed using the 450 K array are indicated. E Pyrosequencing results of all CpG sites analyzed between CC and TC patients. Mean ± SEM; p < 0.05 * Chan et al. Clinical Epigenetics (2023) 15:5 Page 9 of 19 Fig. 6 Altered sperm DNA methylation detected months following chemotherapy treatments. A RLGS and B 450 K array analysis of DNA methylation over time/post-chemotherapy treatment of CC patients (left) and from HD and TC patients (right). The dotted lines drawn within each graph indicate the average number of sites altered in CC patients for each technique used. C MCC-Seq (5-methylcytosine capture sequencing) analysis showing the number of CpG sites that demonstrated > 20% methylation difference within individual patients, comparing the 12 and 18 months post-treatment, with the number of dynamic CpG sites indicated. D The total number of sites altered is expressed as a percentage of all the sites (1,711,875) analyzed. ***p < 0.0001, chi-squared test with Yate’s correction to examine two CC and one each of the HD and TC had an elevated number of altered sites, with 53,011 patients. While no baseline time points were available and 48,882 CpGs, respectively (Fig. 6C). In all patients, for further analyses, we analyzed samples collected 12 the number of hypo- and hypermethylation events and 18 months post-chemotherapy (time points 2 and was similar. Interestingly, a vast majority of the sites 3) and determined the number of differences between found to be different between 12 and 18 months were these two times, for each subject. With a minimum found in the dynamic sites (hatched bars); on aver- tenfold coverage, a total of 1,711,875 CpG sites were age 82.9% of the changes occurring were at dynamic commonly sequenced between all the samples. Using CpGs. The total number of sites altered from each sub- a similar cutoff as the techniques above, we found ject was also expressed as a percentage of all the CpGs that the CC patients had a total of 39,656 and 38,383 analyzed (Fig. 6D); a significantly larger proportion CpG sites which demonstrated a minimum 20% dif- of CpGs was altered in cancer patients as compared ference in methylation; the HD and TC patients both to the CC subjects (p < 0.0001, chi-squared test with Chan et al. Clinical Epigenetics (2023) 15:5 Page 10 of 19 Yate’s correction). The use of the MCC-Seq approach resulted from the underlying cancer and would likely provides further evidence that DNA methylation disappear post-treatment. To determine if any of the defects may persist at later time points post-treatment pre-treatment differences between the groups persisted (12–18 months). post-treatment, we examined a single sample from each cancer survivor, where the number of altered probes/ Time‑course analysis of post‑treatment samples sites from the 450 K array was similar to that of the CC demonstrates altered sperm DNA methylation post-chemotherapy, in an attempt to minimize the effects at 6 months, with some evidence of persistence at later seen by chemotherapy treatment. In examining the times and common sites affected 11,525 significant probes discovered with our baseline We next wanted to examine how alterations in sperm (pre-treatment) three-group ANOVA, hierarchical clus- DNA methylation behave over time post-treatment, tering still segregated the TC survivor patients from the within individual cancer patients, using the 450 k array HD and CC cohorts post-chemotherapy (Fig. 8A). Using data. Venn diagrams were created to depict the altera- the top 500 significant probes, there was an even better tions in the sperm DNA methylome for cancer patients distinction of the three cohorts (Fig. 8B). We next took for whom several time points post-treatment were availa- the 450 K array data from all subjects/patients, regard- ble (Fig. 7A). A preponderance of changes in DNA meth- less of the degree of altered methylation post-treatment, ylation at the initial 6-month follow-up were specific to and examined the methylation of DMRs that had shown this time point. Examining patterns in sperm DNA meth- altered methylation pre-treatment between the different ylation for these sites across all collection periods, we see cancer groups. We examined the individual probes dis- that, compared to baseline, hypo- and hypermethylation covered through ANOVA, highlighting those specifically are observed at 6 months following chemotherapy and altered following post hoc Tukey’s testing, and examined that these patterns return to near normal levels at subse- the averaged methylation throughout the region. No sig- quent time points (Fig. 7B). nificant difference is observed between CC subjects at For patients HD2 and HD3, where a large number of baseline and at their time points afterwards for the three sites were affected at 6 months, we asked whether com- genes analyzed (Fig. 8C-E and Additional file 2: Fig. S6). mon sites were affected between patients. A total of 1234 In contrast, the BASP1P1 locus demonstrated signifi- CpGs were altered in both patients, representing ~ 25% cantly lower methylation in all cancer patients at all time of DMCs from each patient; these individual sites were points (Fig. 8C and Additional file 2: Fig. S6A). For the merged into 225 DMRs (Fig. 7C, right). TopGO analysis region spanning SPON2/LOC100130872, HD patients at of these regions revealed a number of significant biologi- baseline and post-treatment showed significantly higher cal processes, many relating to brain and overall develop- levels of methylation (Fig. 8D). Looking at individual ment as well as neurobiology (Fig. 7C, left and Additional CpGs, all sites, including those not significant following file 3). Next, while most changes following treatment post hoc Tukey’s testing, showed increased sperm DNA appeared to be resolved after the 6-month time point, methylation in HD subjects at all time points (Addi- we asked whether some altered sites persisted over time. tional file 2: Fig. S6B). For the SPON2/LOC100130872 For patient HD2, a total of 256 sites were found in com- region, TC patients showed a nonsignificant decrease mon at all the collection times post-chemotherapy (6, 12 at baseline (p = 0.23), which became significant post- and 24 months, Fig. 7A). GO analysis of these sites, found treatment. Similarly, within the promoter-TSS region that many biological processes involved in development of GDF2, TC subjects demonstrated decreased average as well as terms related to the nervous system were sig- methylation post-treatment (Fig. 8E) and all sites across nificantly enriched (Fig. 7D and Additional file 3). Taken the region possessed lower methylation, as well (Addi- together, while the majority of the defects in sperm DNA tional file 2: Fig. S6C); this latter result is in line with our methylation patterns immediately following chemother- pyrosequencing validation which revealed the decreased apy appear to be resolved, common changes due to treat- methylation at many CpG sites across the locus. Taken ment for HD were identified. These changes and those together, these results indicate that sperm samples col- found to persist up to 2 years post-treatment are related lected post-chemotherapy treatment in survivors show to processes important for development and the nervous persistent differential methylation. system. Sperm DNA methylation within the GDF2 promoter Pre‑treatment differential methylation of CpG sites demonstrates increased variability in TC patients in sperm persists post‑chemotherapy treatment compared to CC in a validation cohort We postulated that the pre-treatment differences in Having demonstrated that, within our study, decreased sperm methylation between the three groups might have methylation was observed in TC patients throughout the Chan et al. Clinical Epigenetics (2023) 15:5 Page 11 of 19 A) B) HD1 HD2 HD3 TC1 TC2 1.0 1.0 1.0 1.0 1.0 0.8 0.8 0.8 0.8 0.8 Beta-value Hypo 0.6 0.6 0.6 0.6 0.6 0.4 0.4 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.2 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 0.8 0.8 0.8 0.8 0.8 Beta-value Hyper 0.6 0.6 0.6 0.6 0.6 0.4 0.4 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.2 0.0 0.0 0.0 0.0 0.0 m ne m hs m hs s m ne m hs m hs s m e 18 on s 42 on s m ths s e s s e s s th th m th m th th th th th th 6 elin lin lin 12 ont 24 ont 24 nt 30 ont li li on on 12 on on on on on on se se se se o s m m m m Ba Ba Ba Ba Ba 6 6 6 18 6 36 Time Points C) Top 20 terms ordered by weighted Fisher p-value D) Top 20 terms ordered by weighted Fisher p-value insulin secretion involved in cellular response ... trachea cartilage development midbrain-hindbrain boundary development negative regulation of mesenchymal cell ... anterograde axonal protein transport positive regulation of neuron projection... positive regulation of hepatocyte proliferation protein localization to basolateral plas... MAPK cascade asymmetric cell division Go Biological processes Go Biological processes negative regulation of cardiac muscle tissue dev. mitral valve morphogenesis homophilic cell adhesion via plasma... neural tube closure positive regulation of epithelial cell apoptosis ... multicellular organism growth negative regulation of voltage-gated cal... positive regulation of GTPase activity cell-cell junction organization negative regulation of synaptic plastici... axon guidance negative regulation of oxidative stress-... regulation of postsynaptic neurotransmit... glandular epithelial cell development response to oxidative stress non-canonical Wnt signaling pathway via ... regulation of canonical Wnt signaling pa... regulation of prostatic bud formation protein localization to cell junction face development beta-catenin destruction complex disasse... regulation of protein localization to ce... free ubiquitin chain polymerization lens morphogenesis in camera-type eye receptor-mediated endocytosis of virus ... pallium development plasma membrane raft assembly epithelium migration mesenchymal-epithelial cell signaling Fig. 7 Altered sperm DNA methylation in cancer patients demonstrates mainly acute changes, with some patients showing common alterations and persistent changes over time. A Venn diagrams of altered sperm DNA methylation within patients across different collection time points. B Acute hypo- and hypermethylation observed at 6 months following chemotherapy treatments. C Intersection of DMCs from two HD patients (right) with gene ontology (GO) analysis performed using common DMRs (left). D GO analysis of DMCs found to persist throughout all time points from patient HD2. Top 20 most significant GO biological processes, by weighted Fisher p values, are plotted with the number in bars representing the number of significant genes annotated to DMRs/DMCs. Enrichment is calculated as a ratio of genes affected and the number of genes found within each GO biological process (see Additional file 1 for all significant GO terms). Dashed vertical line, p = 0.05 Chan et al. Clinical Epigenetics (2023) 15:5 Page 12 of 19 Fig. 8 Sperm DNA methylation of samples post-chemotherapy treatment can still distinguish subjects from each other. Hierarchical clustering of cancer patients post-chemotherapy and baseline CC using the previously discovered A 11,525 significant ANOVA probes and B the top 500 most significant probes. The averaged beta-values of the probes found to differentiate cancer cohorts were used to examine the regions of C SPON2/ LOC100130872, D BASP1P1 and E) GDF2 in CC, HD and TC at baseline (solid bars) and post-chemotherapy treatments (hatched bars). Mean ± SEM; t test *p < 0.05, **p < 0.01 promoter region of GDF2 (Fig. 4E) and that this differ- Discussion ence is still found in samples post-treatment (Fig. 7E), Hodgkin disease and testicular cancer are two forms of we wanted to determine whether the same tendencies cancers that are commonly diagnosed in men of young held true when examining a different cohort (Italian) reproductive age. While treatments result in high sur- of subjects. The patient characteristics of these sam- vival rates, questions about fertility and the possibility ples are shown in Additional file 1: Table S5. Sperm of fathering children post-treatment are of particular samples from 6 CC subjects and 12 TC patients, 6 sam- concern. Are there negative impacts due to the disease ples each collected at 2 and 3 years post-chemotherapy, state itself and/or caused by the chemotherapy treat- were obtained, and bisulfite pyrosequencing was used ments that can affect the sperm? Can these have implica- to assay the same 21 CpG sites within GDF2. The aver- tions for the next generation? This study adds to previous age methylation across the locus was similar for the two work, including reports from our group, demonstrat- groups (55.71% and 54.17% for CC and TC, respectively). ing an increased incidence of aneuploidy, DNA damage However, there was more variability in GDF2 methyla- and altered chromatin quality in human sperm up to 2 tion among the TC patients, particularly at 2 years post- years following combination chemotherapy treatments treatment, where 2 of the 6 patients showed low levels for HD and TC [7–10]. Our findings here reveal dis- of methylation across the locus (Fig. 9A). Variability in tinct DNA methylation patterning in sperm prior to the GDF2 methylation decreased but remained significant at onset of treatment that are able to distinguish HD, TC 3 years post-treatment (Fig. 9B). and CC subject groups from one another; interestingly, these profiles between the different groups remained Chan et al. Clinical Epigenetics (2023) 15:5 Page 13 of 19 Fig. 9 Promoter region of GDF2 of an external validation cohort demonstrates a large variation in sperm DNA methylation. A The averaged methylation of all samples from an external validation cohort is plotted along with the methylation from individual CC and TC patients across the GFD2 promoter locus; B Plot of the average methylation of all samples, from both Canadian and Italian cohorts, in which pyrosequencing results are available. Mean ± SEM; *p < 0.05, Mann–Whitney test post-treatment. As well, by comparing baseline and post- to chemotherapy for osteosarcoma, no imprinting gene chemotherapy samples, while most initial defects appear methylation abnormalities were reported [6]. to be resolved at later time points, altered sperm DNA The study presented here is novel, in that we were able methylation was detected in patients up to 18–24 months to examine the human sperm DNA methylome in sam- post-treatment, indicating that treatment-induced epi- ples acquired prior to the initiation of treatment and genetic abnormalities, like genetic damage, persist long follow the same patients at several time points post- after patients become disease free. chemotherapy. Our results with the 450 K arrays demon- Imprinted genes acquire sperm-specific DNA meth- strated that, prior to treatment, sperm DNA methylation ylation patterns during male germ cell development and signatures are present in men and are able to distinguish play crucial roles in normal development; disturbances HD and, to a greater extent, TC patients from CC sub- in genomic imprinting can lead to a number of disorders jects. Interestingly, these group-specific DNA methyla- [32]. We analyzed the methylation of several imprinted tion patterns persisted in sperm following chemotherapy, genes represented on the 450 K arrays and found nor- up to the 3.5-year the period studied. Pre-treatment mal profiles in the sperm of men diagnosed with HD or signatures were not evident using the lower resolution TC, prior to, and following, combination chemotherapy RLGS method. RLGS examines DNA methylation at a treatments. Our pre-treatment results are in keeping relatively small number of loci compared to the 450 K with those of a recent study examining the DNA meth- arrays and can only reproducibly detect larger changes ylation status of imprinted genes in cryopreserved sperm in DNA methylation, greater than 25% in magnitude. samples from normospermic before treatment semi- Similar to our findings, sperm aneuploidy has also been noma patients and control subjects, where no altera- detected in pre-treatment samples from testicular can- tions in imprinted gene methylation were found [33]. The cer and Hodgkin disease patients [7, 34]. Explanations absence of imprinted gene DNA methylation abnormali- for the presence of pre-treatment signatures include the ties in testicular cancer patients prior to treatment in two impact of a cancer environment on germline stem cells, different studies is reassuring. We are not aware of stud- underlying genetic susceptibility, and an altered somatic ies that have compared imprinted gene methylation in environment. Recent studies examining the incidence patients before and after treatment. However, we previ- of birth defects in the children of men with cancer, sup- ously examined effects of human testicular cancer treat- port the potential clinical relevance of the presence of ment regimens (BEP) on sperm DNA methylation in a rat pre-treatment genetic and epigenetic alterations in the model [17]. Similar to our human studies, imprinted gene spermatozoa of cancer patients [13]. While studying methylation was not affected by the BEP treatment. In a sperm aneuploidy is labor-intensive and requires special- post-treatment genome-wide study of sperm DNA meth- ized expertise, genome-wide DNA methylation assays are ylation in adult patients 10 years after pubertal exposure more widely accessible and may facilitate the screening of pre-treatment spermatozoa for prospective studies. Chan et al. Clinical Epigenetics (2023) 15:5 Page 14 of 19 Our study suggests that pre-treatment DNA methyla- region found in HD patients was within the SPON2 tion differences between groups can persist post-treat- locus, showing increased methylation in HD versus CC ment. In support of the finding in our prospective cohort, individuals. Altered methylation within the promoter of validation of a locus found to be altered in TC patients this extracellular matrix protein has been observed both also showed greater variation in an external cohort, in prostate cancer and in cancer-associated myofibro- where samples were collected 2–3 years following chem- blast cell lines from gastric and esophageal adenocarci- otherapy. The most likely explanation would be persis- nomas compared to normal tissues [46, 47]. Decreased tence of abnormal germline stem cells or spermatogonial methylation within the promoter led to a corresponding stem cells (SSCs) in the testis post-treatment. SSCs that increase in SPON2 expression within the cell lines, which had been quiescent may not have been eliminated by the was also detected in patients with the poorer prognosis chemotherapy. That SSCs with abnormal DNA meth- and survival of gastric/colorectal carcinoma patients [48, ylation patterns can persist in the testis for many years 49]. Specific to testicular cancer, decreased methylation is supported by a study of osteosarcoma patients treated in sperm was observed within the promoter of GDF2 as adolescents and examined as adults [6]. Although pre- in our patient cohort, a result also seen in a subset of treatment samples were not available for the latter study, TC patients 2–3 years post-treatment in the validation altered DNA methylation patterns were found in sperma- cohort. This locus encodes the GDF2 protein, also known tozoa 10 years after pubertal exposure to chemotherapy. as bone morphogenic protein 9 (BMP9), a member of The pre-treatment aberrant epigenetic marks found the TGFβ superfamily, which has roles in cancer pro- in the sperm of HD and TC patients may identify a pre- gression [50]. Specifically, GDF2 activation of SMAD1/5 existing signature that could be used to screen patients. signaling has been found to enhance cell death in ovar- Indeed, both diseases have etiologies associated with ian and breast cancer cell lines; treatment of these cells altered epigenetic patterning. The distinctive cells used with 5-azacytidine, a potent demethylating agent, dem- for the diagnosis of HD, the Hodgkin/Reed–Sternberg onstrated decreased methylation of the GDF2 locus, cells, have been shown to have lost or decreased many increased signaling and higher sensitivity to cell death of the B cell-specific gene expression profiles [35, 36]. [51]. This same study also observed higher GDF2 pro- By altering DNA methylation and/or histone acetyla- moter methylation in tumors of ovarian cancer patients tion patterning, studies in cell lines have been able to compared to normal individuals. Similarly, loss of GDF2/ alter the expression patterns of B cells to resemble more BMP9, in a mouse model of breast cancer, was associated closely those characteristics of HD, and vice versa [37– with increased tumor growth and metastasis, indicating 39]. Testicular cancer, hypospadias, cryptorchidism and a protective role of this gene [52]. From the cancer stud- poor semen quality, are included in the clinical entity, ies, the increased methylation of SPON2 in the sperm of testicular dysgenesis syndrome, a disorder postulated to HD patients would result in decreased expression, while be due to environmental effects that disrupt embryonic the decreased methylation of GDF2 in the sperm of the and fetal germ cell development [40, 41]. Indeed, the pre- TC patients would result in increased expression. We cursor of testicular germ cell tumors, intratubular germ hypothesize that such altered expression patterns, per- cell neoplasia (ITGCN) cells, has been theorized to origi- haps involving other genes identified in the HD and TC nate from fetal gonocytes, through morphological and cohorts, may provide a growth or differentiation advan- gene expression profiling [42, 43]. Epigenetic marks of tage for germ cells. the ITGCN have also been demonstrated to be similar to Immediately following chemotherapy treatments, those of fetal germ cells [44, 45], allowing for increased many patients in the HD and TC cohorts showed an proliferative activity. Interestingly, biological processes elevated number of sites with altered sperm DNA meth- relating to the immune system/immune response and ylation compared to the CC groups. Indeed, a significant sexual/male reproduction were observed in the differen- increase in DNA damage, as measured by tail extent tially methylated regions from the baseline analysis of the moment (Comet assay), was detected in a similar group HD and TC cohorts, respectively, alluding to the origins of both HD and TC patients 6 months after treatments of these diseases. While the altered epigenetic signatures [9]. We also observed that many of the initial defects from our patient cohorts were discovered in sperm, it observed were resolved and the number of altered sites remains to be determined whether more accessible bio- diminished with time, similar to the treatment-induced logical samples, such as blood or saliva, would also carry damage to sperm chromatin structure being differen- perturbed epigenetic patterning. tially repaired over time in cancer survivors. However, From cancer studies, methylation of loci identified in one patient demonstrated increased alterations in the both HD and TC patient sperm are known to influence sperm methylome in the following 2 years; similar signifi- gene expression. The largest differentially methylated cant sperm DNA damage and low DNA compaction have Chan et al. Clinical Epigenetics (2023) 15:5 Page 15 of 19 been observed up to 24 months post-treatment [10]. It is environmental exposures, as well as the process of aging, also intriguing that developmental and nervous system has demonstrated that differential methylation is mainly pathways were enriched in examining our post-treatment found in these novel dynamic CpG sites targeted [31, 59]. effects for GO biological process terms. Taken together, Future studies using our human sperm targeted panel, or alterations in sperm DNA methylation, altered sperm with whole-genome bisulfite sequencing, may be useful chromatin structure and DNA damage can be observed to examine in greater detail the effect of cancer as well as in HD and TC patients prior to and up to several years their treatments on the sperm methylome. following chemotherapy regimens. This damage to the genetic/epigenetic cargo of spermatozoa may help to Conclusion explain the observed increased risk of congenital abnor- This is the first study that demonstrates distinct sperm malities in children fathered by men with a history of DNA methylation signatures in cancer patient cohorts: cancer, whether the offspring were conceived prior to or differences in the sperm DNA methylome between CC, years following treatments [11, 12]. HD and TC patients were observed prior to the initiation One limitation of our study is the small number of of treatment and these signatures persisted post-chemo- patients available in each cohort, as well as the reduced therapy, up to the 3.5-year time period studied. Follow- numbers of samples available following treatments. ing treatments, we observed higher levels of methylation While these are common forms of cancers in men of defects, which decreased over time. However, similar to reproductive age, the incidence is still low, with 2.8 and observed DNA damage, in some patients, these defects 6.8 per 100,000 in the United States for HD and TC, persisted up to 2 years following treatment. Using a tar- respectively [53, 54]. As well, the impact of treatments on geted human sperm capture sequencing technique, dif- spermatogenesis is evident with decreased sperm counts ferences in sperm DNA methylation were observed in immediately following chemotherapy, which only return novel dynamic CpG sites, which warrants further inves- to pre-treatment levels after 24 months [55]. Despite tigation. Our findings emphasize the importance of col- the low patient numbers in our different cohorts, we lecting pre-treatment samples in future studies and may were able to detect differential methylation signatures help explain increases in birth defects reported in cancer in sperm. As mentioned previously, a validation in other patients both before and after their treatment. cohorts of HD and TC patients of these signatures would be beneficial. Another limitation is that we were only Methods able to examine a subset of the ~ 32.8 million possible Semen sample collection from cancer patients CpG methylation sites found in the human genome [56]. and community controls The majority of the results presented here are from the Semen samples were obtained from a cohort of subjects Illumina 450 K array, examining approximately 450,000 with non-seminoma testicular cancer (TC, stages II–III, CpG sites, with these mainly covering CpG islands and n = 7) or Hodgkin lymphoma (HD, stage II–IV, n = 7) over gene regions. Our early study examining sperm who underwent combination chemotherapy and no DNA methylation using the 450 K array, showed that radiation therapy treatments; TC patients were treated the majority of CpGs interrogated (86% of probes) were with 2–4 cycles of BEP chemotherapy (bleomycin, etopo- hypo- or hypermethylated (< 20% or > 80%, respectively) side, cisplatin) and HD patients underwent 4–8 cycles with methylation levels conserved between individu- of ABVD treatments (doxorubicin, bleomycin, vinblas- als; only a small proportion of probes (0.3%) were vari- tine, dacarbazine). Collection time points from these able between individuals and demonstrated intermediate individuals were at baseline (t = 0), prior to initiation of (20–80%) methylation [57]. We reported similar finding chemotherapy, and at 6 months intervals (t = 1, 2, 3, etc.), (i.e., a minority of variable CpG sites) in a more recent up to 3.5 years, after termination of treatments. Due study using the 450 K array and examining the effects to dropout and/or azoospermia, samples could not be of low dose folic acid supplementation on sperm DNA obtained throughout the post-chemotherapy period for methylation [58]. Our use of the human sperm-specific all patients. A cohort of healthy untreated community MCC-Seq that examines 3.18 million sites, did demon- control (CC, n = 8) volunteers was also recruited. The strate a greater number of alterations in patients who CC subjects had no significant past medical history or underwent chemotherapy treatments; the vast majority history of infertility. All CC, HD and TC subjects were of the altered DNA methylation was found in dynamic non-smokers and were part of a larger cohort of sub- CpGs, sites of intermediate methylation and not well cov- jects, where studies examining sperm quality and damage ered on the different Illumina arrays. Indeed, the use of over time and following chemotherapy treatments were our sperm-specific panel for the analysis of human DNA previously conducted [8, 9]. The baseline samples from methylation differences due to a genetic polymorphism, the CC cohort were included as a post mandatory folic Chan et al. Clinical Epigenetics (2023) 15:5 Page 16 of 19 acid fortification group in a previous study of ours [58]. at over 480,000 CpGs dinucleotides of CC (n = 8), HD Sperm from men in all groups were collected during the (n = 7) and TC (n = 7) subjects. The assay was conducted same time period; the men would have been exposed to by Genome Quebec at the McGill University Innova- folic acid fortification in their diets, due to the mandatory tion Centre using a protocol previously described [62]. addition of this vitamin to white flour and enriched grain Briefly, ~ 500 ng of DNA was bisulfite converted using an products in 1998; no additional folic acid supplementa- EZ DNA methylation Kit (Zymo Research), followed by tion was given to any of the cohorts. Semen samples were whole-genome amplification, enzymatic fragmentation collected by masturbation, with a minimum of 3 days and hybridization to the BeadChip. Arrays were imaged abstinence and allowed to liquefy at 37 °C for 30 min. on the Illumina iScan, resulting in raw data files in IDAT Samples were then aliquoted and stored at −80 °C in format. The annotation of the 450 K array to determine the absence of cryo-protectant. Samples from an Italian the genomic location of probes was determined with Illu- validation cohort, were also collected in a similar fash- minaHumanMethylation450kanno.ilmn12.hg19 (version ion. Here, TC patient survivors who had undergone 1–4 0.6.0) [63]. Preprocessing of the raw data files (IDAT), cycles of BEP chemotherapy were recruited and samples quality control (QC) assessment and normalization were collected 2 or 3 years after termination of treatments performed using the R package minfi [64], with normali- (n = 6–3 seminoma and 3 non-seminoma, per time zation based on the study by Fortin et al. [65]. The QC point); six control samples from this cohort were also demonstrated that a single sample from a CC patient examined. The study was approved by the different insti- (CC7-2) clustered separately from other samples and tutional ethics review boards and informed consent was failed due to lower median intensities (Additional file 2: obtained from all subjects (see Declarations). Fig. S7A, red arrow). PCA of raw and normalized data was also carried out using the R package shiny methyl Sperm DNA extraction (version 1.22.0) [66]. This analysis indicated that the same Semen samples were thawed on ice and centrifuged at sample failing QC was an outlier of PC2 (Additional 4000 g for 5 min in order to obtain sperm pellets. Pellets file 2: Fig. S7B, red arrow); as well, three other samples were resuspended in lysis buffer (10 mM Tris•HCl, pH 8; (one CC and two TC; CC8-0, TC7-0 and TC7-1; blue 150 mM EDTA, pH 8; 100 mM DTT) and high molecular arrows) appeared to be outliers within PC1. We exam- weight DNA isolated by Proteinase K digestion and phe- ined whether these samples were possibly contaminated, nol–chloroform extraction, as described previously [60]. by examining the imprinted gene methylation. The three Precipitated DNA was allowed to dissolve in TE buffer samples were identified as outliers through PCA analysis, (10 mM Tris•HCl, pH 8; 1 mM EDTA) for 1 week before which demonstrated deviation in methylation across all proceeding with DNA methylation analyses. germline ICRs analyzed to a similar extent; this would indicate somatic cell contamination (Additional file 1: Global assessment of methylation by RLGS Table S1). Hierarchical clustering of the averaged meth- Several techniques were used for the assessment of DNA ylation from all ICRs of imprinted genes demonstrated methylation. Initial analysis relied on restriction land- that the failed QC sample as well as the outliers clustered mark genomic scanning, RLGS [60]. Briefly, DNA sam- clearly apart (Additional file 2: Fig. S7C); therefore, all ples (CC, n = 8; HD, n = 7; TC, n = 6) were digested with these (along with any other post-treatment samples) were the methylation sensitive restriction enzyme, NotI. Fol- removed from further analysis, leaving an n = 7 for the lowing this, radioactive end-labeling of overhangs was CC and HD cohorts and an n = 6 for the TC cohort. performed, proceeded by further digestion of samples with EcoRV. DNA fragments were electrophoresed on a Customized 5‑Methyl‑C capture sequencing (MCC‑Seq) first dimension agarose gel and in situ digestion with a of DNA with human sperm‑specific panel third restriction enzyme, Hinf1, was carried out. Samples Our recently developed human sperm-specific capture were finally run on a second dimension acrylamide gel, panel [31] was used to assess over 3 million CpG sites resulting in the final RLGS profile. Profiles were imaged within the genome of 8 samples: two CC and one each on phosphorimaging plates and film. Films were visually HD and TC at times 2 and 3 following chemotherapy assessed and a scoring system previously described was treatment. Using 1 µg of DNA from each sample, whole- used for analysis [61]. genome bisulfite sequencing (WGBS) library prepara- tion and targeting bisulfite sequencing were performed as Methylation profiling with HumanMethylation450 described previously [67, 68]. Briefly, DNA was sonicated BeadChip to obtain 300–400 bp fragments, proceeded by DNA- Illumina’s Infinium HumanMethylation 450 BeadChip end repair, 3’-end adenylation, adaptor ligation and clean (450 K) array was used in order to assess the methylation up, according to the KAPA High Throughput Library Chan et al. Clinical Epigenetics (2023) 15:5 Page 17 of 19 Preparation kit protocol (Roche/KAPA Biosystems). Sam- DNA methylation of each CpG analyzed by bisulfite ples were then bisulfite converted using the EpiTect Fast pyrosequencing, as well as the averaged normalized DNA bisulfite kit (Qiagen) followed by 9–12 cycles of beta-value across the different DMRs. GO term enrich- PCR amplification. The final WGBS libraries were purified ment was performed using the R package topGO (ver- using Agencourt AMPure Beads (Beckman Coulter) and sion. 2.42.0). Chi-squared with Yates’ correction was were then enriched using our human sperm-specific panel, applied to compare the frequency of CpG sites having using the SeqCap Epi Enrichment System protocol devel- > 20% methylation changed, as measured by MCC-Seq, oped and optimized by Roche NimbleGen. Equal amounts in cancer patients compared to the individual CC sub- of the multiplexed WGBS libraries (84 ng each) were com- jects. Unless otherwise specified, statistical significance bined and hybridized to the human sperm-specific cap- was set at p < 0.05 for the different analyses. ture panel at 47 °C for 72 h. The captured libraries were washed, recovered and PCR-amplified before final puri- Abbreviations fication was conducted. The final capture libraries were TC Testicular cancer sequenced in a single lane of the Illumina HiSeq4000 sys- HD Hodgkin disease tem using 100-bp paired end sequencing. Raw reads from CC Community control 450 k array Illumina HumanMethylation 450 K BeadChip the sequencers were processed using pipelines developed PCA Principal component analysis at the McGill University and Génome Québec Innovation ICR Imprinting control region Centre (MUGQIC), as part of the GenAP project (https://​ RLGS Restriction landmark genome scanning DMR Differentially methylated region bitbu​cket.​org/​mugqic/​genpi​pes). Common CpG sites from TSS Transcriptional start site all samples, with > 10 × coverage, were retained. MCC-seq 5-Methylcytosine capture sequencing BEP treatments Bleomycin, etoposide, cisplatin treatments ABVD treatments Doxorubicin, bleomycin, vinblastine, dacarbazine Locus‑specific analysis of DNA methylation by bisulfite treatments pyrosequencing Bisulfite pyrosequencing was used to validate differen- Supplementary Information tial methylation at the promoter region of GDF2. Here, The online version contains supplementary material available at https://​doi.​ baseline sperm DNA samples prior to any treatments org/​10.​1186/​s13148-​022-​01417-1. were analyzed from CC (n = 8) and TC (n = 5) cohorts Additional file 1. Supplementary Tables S1–S6. of our main study, as well the Italian validation cohort. Additional file 2. Supplementaion Figures S1–S7. For pyrosequencing assays, sperm DNA (200–500 ng) underwent bisulfite conversion with the EpiTect Bisulfite Additional file 3. Complete list of significant GO terms. Kit (Qiagen) according to the manufacturer’s protocol. Primers were designed (PyroMark Assay Design 2.0, Qia- Acknowledgements We thank the team at the Génome Québec Innovation Centre for performing gen) to overlap the areas where differential methylation the Illumina HumanMethylation 450K BeadChip experiments. Research was was observed following 450 K array analysis; two sets enabled in part by support provided by Calcul Québec and Compute Canada. of PCR primers, using a total of five sequencing prim- Author contributions ers, were designed in order to analyze a total of 21 CpG D.C., B.R. and J.T. conceived the study. D.C. and J.T. designed the study. A.R., sites (Additional file 1: Table S6). Bisulfite PCR was con- C.K., C.O. and P.C. provided samples analyzed in this study. D.C. performed the ducted, and pyrosequencing was performed as previously experiments and acquired the data. D.C., K.K. analyzed the data. D.C. and J.T. wrote the manuscript. All authors read and approved the final manuscript. described [69] using the PyroMark Q24 kit (Qiagen) as per the manufacturer’s protocol. Funding This work was supposed by a grant from the Canadian Institutes of Health Research (CIHR) to JT (FDN-148425) and to BR (HGG-62294). JT and BR are Statistical analyses James McGill Professors. Fundación Merck Salud (Ayudas Merck de Inves‑ The R environment and language (version 3.5.2) were tigación 2017 en el área de Investigación Clínica en Fertilidad) provided used to perform statistical computing and graphing of an award to C.K. the of 450 K array data. Specifically, graphical heatmaps Availability of data and materials and hierarchical clustering, as well as ANOVA and Illumina HumanMethylation 450 K BeadChip data (IDAT files) and MCC-Seq post hoc Tukey’s tests, were conducted using RStudio data have been submitted to Gene Expression Omnibus (GEO) under the accession number GSE211122. (version 1.2.1335). Significant probes within 1000 bp were merged using the merge function of BEDtools Declarations (version 2.29.0) to determine DMRs. Other graphs were generated, and statistical analyses were per- Ethics approval and consent to participate formed using GraphPad Prism Version 6.01. Unpaired The study was approved by the different institutional ethics review boards, and informed consent was obtained from all subjects. Montreal cohort: The t tests were performed to compare the mean sperm Chan et al. Clinical Epigenetics (2023) 15:5 Page 18 of 19 McGill Research Ethics Office (Institutional Review Board): McGill IRB (A04- 13. Al-Jebari Y, Rylander L, Stahl O, Giwercman A. Risk of congenital malfor‑ M32-03A); Italian cohort: ethics committee from the Fundacio Puigvert (Ref FP: mations in children born before paternal cancer. JNCI Cancer Spectr. 2014/07c) and the regional ethics committee for clinical experimentations in 2018;2(2):pky027. Tuscany (Ref: 27-11 and 2019-481). 14. Bieber AM, Marcon L, Hales BF, Robaire B. Effects of chemotherapeutic agents for testicular cancer on the male rat reproductive system, sperma‑ Consent for publication tozoa, and fertility. J Androl. 2006;27(2):189–200. Non-applicable—no individual person’s data in any form are included in this 15. Delbes G, Hales BF, Robaire B. Effects of the chemotherapy cocktail manuscript. used to treat testicular cancer on sperm chromatin integrity. J Androl. 2007;28(2):241–9; discussion 50–1. Competing interests 16. Delbes G, Chan D, Pakarinen P, Trasler JM, Hales BF, Robaire B. Impact of The authors have no competing interests to declare. the chemotherapy cocktail used to treat testicular cancer on the gene expression profile of germ cells from male Brown-Norway rats. Biol Author details Reprod. 2009;80(2):320–7. 1 Research Institute of the McGill University Health Centre, 1001 Décarie Boul. 17. Chan D, Delbes G, Landry M, Robaire B, Trasler JM. Epigenetic alterations Block E, Montréal, QC, Canada. 2 Lady Davis Institute for Medical Research, Jew‑ in sperm DNA associated with testicular cancer treatment. Toxicol Sci. ish General Hospital, Montréal, QC, Canada. 3 Andrology Department, Fundació 2011. Puigvert, Universitat Autònoma de Barcelona, Instituto de Investigaciones 18. Fowden AL, Sibley C, Reik W, Constancia M. Imprinted genes, placental Biomédicas Sant Pau (IIB-Sant Pau), Barcelona, Catalonia, Spain. 4 Department development and fetal growth. Horm Res. 2006;65(Suppl 3):50–8. of Biomedical, Experimental and Clinical Sciences Mario Serio, University 19. Isles AR, Holland AJ. Imprinted genes and mother-offspring interactions. of Florence, Florence, Italy. 5 Department of Surgery, McGill University, Mon‑ Early Hum Dev. 2005;81(1):73–7. tréal, QC, Canada. 6 Department of Pharmacology and Therapeutics, McGill 20. Tucci V, Isles AR, Kelsey G, Ferguson-Smith AC, Erice IG. Genomic imprint‑ University, Montréal, QC, Canada. 7 Department of Urology, McGill University, ing and physiological processes in mammals. Cell. 2019;176(5):952–65. Montréal, QC, Canada. 8 Department of Obstetrics and Gynecology, McGill 21. Robles-Matos N, Artis T, Simmons RA, Bartolomei MS. Environmental University, Montréal, QC, Canada. 9 Departments of Pediatrics and Human exposure to endocrine disrupting chemicals influences genomic imprint‑ Genetics, McGill University, Montréal, QC, Canada. ing, growth, and metabolism. Genes (Basel). 2021;12(8). 22. Monk D, Mackay DJG, Eggermann T, Maher ER, Riccio A. Genomic Received: 17 May 2022 Accepted: 21 December 2022 imprinting disorders: lessons on how genome, epigenome and environ‑ ment interact. Nat Rev Genet. 2019;20(4):235–48. 23. Butler MG. Genomic imprinting disorders in humans: a mini-review. J Assist Reprod Genet. 2009;26(9–10):477–86. 24. Santi D, De Vincentis S, Magnani E, Spaggiari G. Impairment of sperm References DNA methylation in male infertility: a meta-analytic study. Andrology. 1. Daugaard G, Gundgaard MG, Mortensen MS, Agerbaek M, Holm NV, 2017;5(4):695–703. Rorth M, et al. Surveillance for stage I nonseminoma testicular cancer: 25. Asenius F, Danson AF, Marzi SJ. DNA methylation in human sperm: a outcomes and long-term follow-up in a population-based cohort. J Clin systematic review. Hum Reprod Update. 2020;26(6):841–73. Oncol. 2014;32(34):3817–23. 26. Leitao E, Di Persio S, Laurentino S, Woste M, Dugas M, Kliesch S, et al. The 2. Engert A. ABVD or BEACOPP for advanced Hodgkin Lymphoma. J Clin sperm epigenome does not display recurrent epimutations in patients Oncol. 2016;34(11):1167–9. with severely impaired spermatogenesis. Clin Epigenetics. 2020;12(1):61. 3. Lee SH, Shin CH. Reduced male fertility in childhood cancer survivors. 27. Feinberg AP, Tycko B. The history of cancer epigenetics. Nat Rev Cancer. Ann Pediatr Endocrinol Metab. 2013;18(4):168–72. 2004;4(2):143–53. 4. Delessard M, Saulnier J, Rives A, Dumont L, Rondanino C, Rives N. 28. Kanwal R, Gupta K, Gupta S. Cancer epigenetics: an introduction. Meth‑ Exposure to chemotherapy during childhood or adulthood and conse‑ ods Mol Biol. 2015;1238:3–25. quences on spermatogenesis and male fertility. Int J Mol Sci. 2020;21(4). 29. Kobayashi H, Hiura H, John RM, Sato A, Otsu E, Kobayashi N, et al. DNA 5. Wyrobek AJ, Schmid TE, Marchetti F. Relative susceptibilities of male germ methylation errors at imprinted loci after assisted conception originate in cells to genetic defects induced by cancer chemotherapies. J Natl Cancer the parental sperm. Eur J Hum Genet. 2009;17(12):1582–91. Inst Monogr. 2005;34:31–5. 30. Kovalchuk A, Ilnytskyy Y, Woycicki R, Rodriguez-Juarez R, Metz GAS, 6. Shnorhavorian M, Schwartz SM, Stansfeld B, Sadler-Riggleman I, Beck D, Kovalchuk O. Adverse effects of paternal chemotherapy exposure Skinner MK. Differential DNA methylation regions in adult human sperm on the progeny brain: intergenerational chemobrain. Oncotarget. following adolescent chemotherapy: potential for epigenetic inheritance. 2018;9(11):10069–82. PLoS ONE. 2017;12(2):e0170085. 31. Chan D, Shao X, Dumargne MC, Aarabi M, Simon MM, Kwan T, et al. Cus‑ 7. Tempest HG, Ko E, Chan P, Robaire B, Rademaker A, Martin RH. Sperm tomized MethylC-capture sequencing to evaluate variation in the human aneuploidy frequencies analysed before and after chemotherapy in sperm DNA methylome representative of altered folate metabolism. testicular cancer and Hodgkin’s lymphoma patients. Hum Reprod. Environ Health Perspect. 2019;127(8):87002. 2008;23(2):251–8. 32. Carli D, Riberi E, Ferrero GB, Mussa A. Syndromic disorders caused by dis‑ 8. O’Flaherty C, Vaisheva F, Hales BF, Chan P, Robaire B. Characterization of turbed human imprinting. J Clin Res Pediatr Endocrinol. 2020;12(1):1–16. sperm chromatin quality in testicular cancer and Hodgkin’s lymphoma 33. Bruno C, Blagoskonov O, Barberet J, Guilleman M, Daniel S, Tournier B, patients prior to chemotherapy. Hum Reprod. 2008;23(5):1044–52. et al. Sperm imprinting integrity in seminoma patients? Clin Epigenetics. 9. O’Flaherty C, Hales BF, Chan P, Robaire B. Impact of chemotherapeutics 2018;10(1):125. and advanced testicular cancer or Hodgkin lymphoma on sperm deoxyri‑ 34. Burrello N, Vicari E, La Vignera S, Romeo G, Campagna C, Magro E, bonucleic acid integrity. Fertil Steril. 2010;94(4):1374–9. et al. Effects of anti-neoplastic treatment on sperm aneuploidy rate in 10. O’Flaherty CM, Chan PT, Hales BF, Robaire B. Sperm chromatin structure patients with testicular tumor: a longitudinal study. J Endocrinol Invest. components are differentially repaired in cancer survivors. J Androl. 2011;34(6):e121–5. 2011;33(4):629–36. 35. Kuppers R, Klein U, Schwering I, Distler V, Brauninger A, Cattoretti G, et al. 11. Stahl O, Boyd HA, Giwercman A, Lindholm M, Jensen A, Kjaer SK, et al. Risk Identification of Hodgkin and Reed-Sternberg cell-specific genes by gene of birth abnormalities in the offspring of men with a history of cancer: a expression profiling. J Clin Invest. 2003;111(4):529–37. cohort study using Danish and Swedish national registries. J Natl Cancer 36. Schwering I, Brauninger A, Klein U, Jungnickel B, Tinguely M, Diehl V, et al. Inst. 2011;103(5):398–406. Loss of the B-lineage-specific gene expression program in Hodgkin and 12. Al-Jebari Y, Glimelius I, Berglund Nord C, Cohn-Cedermark G, Stahl O, Reed-Sternberg cells of Hodgkin lymphoma. Blood. 2003;101(4):1505–12. Tandstad T, et al. Cancer therapy and risk of congenital malformations 37. Ehlers A, Oker E, Bentink S, Lenze D, Stein H, Hummel M. Histone acetyla‑ in children fathered by men treated for testicular germ-cell cancer: A tion and DNA demethylation of B cells result in a Hodgkin-like pheno‑ nationwide register study. PLoS Med. 2019;16(6):e1002816. type. Leukemia. 2008;22(4):835–41. Chan et al. Clinical Epigenetics (2023) 15:5 Page 19 of 19 38. Doerr JR, Malone CS, Fike FM, Gordon MS, Soghomonian SV, Thomas RK, 60. Okazaki Y, Okuizumi H, Sasaki N, Ohsumi T, Kuromitsu J, Hirota N, et al. An et al. Patterned CpG methylation of silenced B cell gene promoters in expanded system of restriction landmark genomic scanning (RLGS Ver. classical Hodgkin lymphoma-derived and primary effusion lymphoma 1.8). Electrophoresis. 1995;16(2):197–202. cell lines. J Mol Biol. 2005;350(4):631–40. 61. Oakes CC, La Salle S, Smiraglia DJ, Robaire B, Trasler JM. A unique configu‑ 39. Ushmorov A, Leithauser F, Sakk O, Weinhausel A, Popov SW, Moller P, et al. ration of genome-wide DNA methylation patterns in the testis. Proc Natl Epigenetic processes play a major role in B-cell-specific gene silencing in Acad Sci USA. 2007;104(1):228–33. classical Hodgkin lymphoma. Blood. 2006;107(6):2493–500. 62. Bibikova M, Fan JB. GoldenGate assay for DNA methylation profiling. 40. Skakkebaek NE, Rajpert-De Meyts E, Buck Louis GM, Toppari J, Anders‑ Methods Mol Biol. 2009;507:149–63. son AM, Eisenberg ML, et al. Male reproductive disorders and fertility 63. Hansen KD. IlluminaHumanMethylation450kanno.ilmn12.hg19: Annota‑ trends: influences of environment and genetic susceptibility. Physiol Rev. tion for Illumina’s 450k methylation arrays. . R package version 060. 2016. 2016;96(1):55–97. 64. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen 41. Skakkebaek NE, Rajpert-De Meyts E, Main KM. Testicular dysgenesis KD, et al. Minfi: a flexible and comprehensive Bioconductor package for syndrome: an increasingly common developmental disorder with envi‑ the analysis of Infinium DNA methylation microarrays. Bioinformatics. ronmental aspects. Hum Reprod. 2001;16(5):972–8. 2014;30(10):1363–9. 42. Skakkebaek NE, Berthelsen JG, Giwercman A, Muller J. Carcinoma- 65. Fortin JP, Labbe A, Lemire M, Zanke BW, Hudson TJ, Fertig EJ, et al. Func‑ in-situ of the testis: possible origin from gonocytes and precursor of tional normalization of 450k methylation array data improves replication all types of germ cell tumours except spermatocytoma. Int J Androl. in large cancer studies. Genome Biol. 2014;15(12):503. 1987;10(1):19–28. 66. Fortin JP, Fertig E, Hansen K. shinyMethyl: interactive quality control of 43. Sonne SB, Almstrup K, Dalgaard M, Juncker AS, Edsgard D, Ruban L, et al. Illumina 450k DNA methylation arrays in R. F1000Res. 2014;3:175. Analysis of gene expression profiles of microdissected cell populations 67. Allum F, Shao X, Guenard F, Simon MM, Busche S, Caron M, et al. indicates that testicular carcinoma in situ is an arrested gonocyte. Cancer Characterization of functional methylomes by next-generation capture Res. 2009;69(12):5241–50. sequencing identifies novel disease-associated variants. Nat Commun. 44. Almstrup K, Nielsen JE, Mlynarska O, Jansen MT, Jorgensen A, Skak‑ 2015;6:7211. kebaek NE, et al. Carcinoma in situ testis displays permissive chroma‑ 68. Cheung WA, Shao X, Morin A, Siroux V, Kwan T, Ge B, et al. Functional tin modifications similar to immature foetal germ cells. Br J Cancer. variation in allelic methylomes underscores a strong genetic contribu‑ 2010;103(8):1269–76. tion and reveals novel epigenetic alterations in the human epigenome. 45. Kristensen DG, Skakkebaek NE, Rajpert-De Meyts E, Almstrup K. Epige‑ Genome Biol. 2017;18(1):50. netic features of testicular germ cell tumours in relation to epigenetic 69. Dejeux E, El abdalaoui H, Gut IG, Tost J. Identification and quantification characteristics of foetal germ cells. Int J Dev Biol. 2013;57(2–4):309–17. of differentially methylated loci by the pyrosequencing technology. 46. Kim JW, Kim ST, Turner AR, Young T, Smith S, Liu W, et al. Identification Methods Mol Biol. 2009;507:189–205. of new differentially methylated genes that have potential functional consequences in prostate cancer. PLoS ONE. 2012;7(10):e48455. 47. Najgebauer H, Liloglou T, Jithesh PV, Giger OT, Varro A, Sanderson Publisher’s Note CM. Integrated omics profiling reveals novel patterns of epigenetic Springer Nature remains neutral with regard to jurisdictional claims in pub‑ programming in cancer-associated myofibroblasts. Carcinogenesis. lished maps and institutional affiliations. 2019;40(4):500–12. 48. Jin C, Lin JR, Ma L, Song Y, Shi YX, Jiang P, et al. Elevated spondin-2 expression correlates with progression and prognosis in gastric cancer. Oncotarget. 2017;8(6):10416–24. 49. Zhang Q, Wang XQ, Wang J, Cui SJ, Lou XM, Yan B, et al. Upregulation of spondin-2 predicts poor survival of colorectal carcinoma patients. Onco‑ target. 2015;6(17):15095–110. 50. Elliott RL, Blobe GC. Role of transforming growth factor Beta in human cancer. J Clin Oncol. 2005;23(9):2078–93. 51. Varadaraj A, Patel P, Serrao A, Bandyopadhay T, Lee NY, Jazaeri AA, et al. Epigenetic regulation of GDF2 suppresses Anoikis in Ovarian and Breast Epithelia. Neoplasia. 2015;17(11):826–38. 52. Ouarne M, Bouvard C, Boneva G, Mallet C, Ribeiro J, Desroches-Castan A, et al. BMP9, but not BMP10, acts as a quiescence factor on tumor growth, vessel normalization and metastasis in a mouse model of breast cancer. J Exp Clin Cancer Res. 2018;37(1):209. 53. Townsend W, Linch D. Hodgkin’s lymphoma in adults. Lancet. 2012;380(9844):836–47. 54. Smith ZL, Werntz RP, Eggener SE. Testicular cancer: epidemiology, diagno‑ sis, and management. Med Clin North Am. 2018;102(2):251–64. 55. Bujan L, Walschaerts M, Moinard N, Hennebicq S, Saias J, Brugnon F, et al. Impact of chemotherapy and radiotherapy for testicular germ cell tumors Ready to submit your research ? Choose BMC and benefit from: on spermatogenesis and sperm DNA: a multicenter prospective study from the CECOS network. Fertil Steril. 2013;100(3):673–80. • fast, convenient online submission 56. Gershman A, Sauria MEG, Guitart X, Vollger MR, Hook PW, Hoyt SJ, et al. Epigenetic patterns in a complete human genome. Science. • thorough peer review by experienced researchers in your field 2022;376(6588):eabj5089. • rapid publication on acceptance 57. Krausz C, Sandoval J, Sayols S, Chianese C, Giachini C, Heyn H, et al. Novel • support for research data, including large and complex data types insights into DNA methylation features in spermatozoa: stability and peculiarities. PLoS ONE. 2012;7(10):e44479. • gold Open Access which fosters wider collaboration and increased citations 58. Chan D, McGraw S, Klein K, Wallock LM, Konermann C, Plass C, et al. • maximum visibility for your research: over 100M website views per year Stability of the human sperm DNA methylome to folic acid fortification and short-term supplementation. Hum Reprod. 2017;32(2):272–83. At BMC, research is always in progress. 59. Cao M, Shao X, Chan P, Cheung W, Kwan T, Pastinen T, et al. High-resolu‑ tion analyses of human sperm dynamic methylome reveal thousands of Learn more biomedcentral.com/submissions novel age-related epigenetic alterations. Clin Epigenetics. 2020;12(1):192.