Paternal MTHFR deficiency leads to hypomethylation of young retrotransposons and reproductive decline across two successive generations

Abstract

5,10-Methylenetetrahydrofolate reductase (MTHFR) is a crucial enzyme in the folate metabolic pathway with a key role in generating methyl groups. As MTHFR deficiency impacts male fertility and sperm DNA methylation, there is the potential for epimutations to be passed to the next generation. Here, we assessed whether the impact of MTHFR deficiency on testis morphology and sperm DNA methylation is exacerbated across generations in mouse. Although MTHFR deficiency in F1 fathers has only minor effects on sperm counts and testis weights and histology, F2 generation sons show further deterioration in reproductive parameters. Extensive loss of DNA methylation is observed in both F1 and F2 sperm, with >80% of sites shared between generations, suggestive of regions consistently susceptible to MTHFR deficiency. These regions are generally methylated during late embryonic germ cell development and are enriched in young retrotransposons. As retrotransposons are resistant to reprogramming of DNA methylation in embryonic germ cells, their hypomethylated state in the sperm of F1 males could contribute to the worsening reproductive phenotype observed in F2 MTHFR-deficient males, compatible with the intergenerational passage of epimutations.

Keywords: DNA methylation; Intergenerational epigenetic inheritance; MTHFR; Male germ cell development; Mouse; Young retrotransposons.

Fig. 1.

MTHFR deficiency impacts epigenetic reprogramming in male germ cells and results in reproductive decline across two generations. (A) Experimental design for the production of F1 and F2 MTHFR-deficient males. Based on high levels of expression of MTHFR in prospermatogonia (PSG), MTHFR deficiency is expected to affect F1 generation (1st hit) primordial germ cells (PGCs) when DNA methylation patterns are established. Epimutations in the sperm of MTHFR-deficient fathers will either be corrected postfertilization during reprogramming in F2 pre-implantation embryos or, if they escape reprogramming, be passed on to the germ cells of the F2 post-implantation MTHFR-deficient embryos (sons). A second phase of reprogramming takes place in the PGCs of the F2 MTHFR-deficient sons, preceding re-acquisition of DNA methylation in PSG, which are deficient in MTHFR (2nd hit). If PGC reprogramming is not complete, F2 generation PSG may carry epimutations from their fathers in addition to being impacted by MTHFR at the time of de novo methylation. (B) Testis weight, sperm count and proportion of abnormal testicular tubules in Mthfr^+/+ (WT) and Mthfr^−/− F1 males (WT, n=3; F1 Mthfr^−/−, n=4). (C) Representative histological cross-sections of testes. (D) Testis weight, sperm count and proportion of abnormal testicular tubules in F2 generation males (WT, n=3; F2 Mthfr^−/−, n=4). (E) Representative testicular histological cross-sections from F2 generation males. (F) Testis weight, sperm count and proportion of abnormal testicular tubules in the maternal deficient (Mat. Def.) group of WT (n=4) and Mthfr^−/− (n=6) males. Data are mean+s.e.m. *P<0.05; **P<0.01; ****P<0.0001. Scale bars: 100 µm.

Fig. 2.

Genome-wide loss of sperm DNA methylation in MTHFR-deficient F1 and F2 generation males. (A) Number of 100 bp tiles that significantly lost (hypomethylated) or gained (hypermethylated) methylation in the sperm of F1 Mthfr^−/− compared with WT males. (B) Distribution of DMTs into genomic elements is shown for all sequenced F1 tiles as well as F1 generation DMTs. (C) GO enrichment analysis of genic DMTs in F1 generation males. The dotted line indicates the P<0.05 threshold for significance for FDR. The dotted bars indicate common enriched pathways between the F1 and F2 generations. (D) Number of 100 bp tiles that were significantly hypomethylated or hypermethylated in the sperm of F2 Mthfr^−/− compared with WT males. (E) Distribution of DMTs into genomic elements is shown for all sequenced F2 tiles as well as F2 generation DMTs. (F) GO enrichment analysis of genic DMTs in F2 generation males. The dotted line indicates the P<0.05 threshold for significance for FDR. The shaded bars indicate common enriched pathways between the F1 and F2 generations. (G) Euler diagrams of common hypo- and hypermethylated tiles between sperm of F1 and F2 generation males. Hypermethylated tiles are shown proportional (in size) to the hypomethylated tiles on the top right, with the magnified version shown below.

Fig. 3.

MTHFR-associated sperm hypomethylation extends beyond isolated CpGs to encompass larger regions in F1 and F2 generation males. (A,B) Comparison of sperm hypomethylated DMRs in MTHFR-deficient males of the F1 (A) and F2 (B) generations. Left: The number of all differentially methylated CpG sites (DMCs), single isolated CpGs (not merged into regions) and merged regions acquired by adjoining DMCs within 100 bp from each other. Right: The distribution of merged DMRs. (C) Comparison of sizes of hypomethylated merged regions (regions equal to or smaller than 100 bp and larger than 100 bp) between F1 and F2 generations. ****P<0.0001. (D) Euler diagram of the common sperm hypomethylated merged DMRs between the F1 and F2 generations. (E) A large sperm hypomethylated DMR in MTHFR-deficient males showing an example of an F2 region within an F1 region (a L1MdA repeat region) is shown on chromosome 5, with CpG sites in the region indicated as red boxes. In the graph, filled shapes indicate significant DMCs and unfilled shapes indicate non-significant DMCs.

Fig. 4.

MTHFR-sensitive sites are enriched for intergenic sequences and young retrotransposons. (A) Distribution of F1 merged hypomethylated DMRs into genomic elements in comparison with all sequenced F1 regions. ***P<0.001; ****P<0.0001. (B) Location of F1 merged hypomethylated DMRs with respect to CpG islands/shores/shelves and open sea regions in comparison with all sequenced F1 regions. (C) The proportion of overlaps between merged hypomethylated DMRs with regions identified as repeats using the RepeatMasker program (LINE, long interspersed nuclear element; LTR, long terminal repeat; SINE, short interspersed nuclear element). (D) The proportion of overlaps between merged hypomethylated DMRs with young LINEs (L1Md family of retrotransposons, Fig. S5) compared with the remainder of the regions.

Fig. 5.

MTHFR-sensitive sites are subject to late de novo methylation and marked by H3K4me3. (A) Scatterplots showing percentage DNA methylation (DNAme) in 4803 F1 hypomethylated DMRs (red dots) compared with whole-genome 1 kb bins (gray dots) in E13.5 male PGCs versus E16.5 PSG (left), E16.5 PSGs versus P0 PSG (middle) and P0 PSG versus sperm (right). For genome 1 kb bins, 50,000 randomly selected data points are plotted. (B) Violin plots showing the distribution of the percentage DNAme levels of whole-genome 1 kb bins, F1 hypo DMRs, top 200 F1 hypo DMRs (by size) and CpG islands (CGIs) during spermatogenesis, including spermatocytes (Spcyte). (C) Scatterplots of H3K4me3 levels in the top 200 F1 hypomethylated DMRs (red dots) compared with whole-genome 1 kb bins (gray dots) in E10.5 versus E13.5 PGCs (left), E13.5 PGCs versus E16.5 PSGs (middle) and E16.5 versus P0 PSGs (right). (D) Violin plots showing the distribution of H3K4me levels for whole-genome 1 kb bins, F1 hypo DMRs, the top 200 F1 hypo DMRs (by size) and CGIs during spermatogenesis, including spermatocytes and sperm. (E) Scatterplots showing the percentage of DNAme versus H3K4me3 levels at F1 hypomethylated DMRs (red dots) compared with whole-genome 1 kb bins (gray dots) for E16.5 PSG (left) and P0-P1 PSG (right). RPKM, reads per kilobase million.

Fig. 6.

MTHFR-sensitive sites are targets for DNMT3l and DNMT3C. (A) Scatterplots showing the percentage DNA methylation in 4803 F1 hypomethylated DMRs (red dots) compared with 50,000 whole-genome 1 kb bins (gray) in Nsd1 heterozygotes (HET) versus knockout (KO) P0 PSG (left) (Shirane et al., 2020), Dnmt3l wild-type (WT) versus KO P10 spermatogonia (middle) and Dnmt3c WT versus KO P10 spermatogonia (right) (Barau et al., 2016). (B) The proportion of DMRs (±1 kb) overlapping with DNMT3C-sensitive regions that showed more than 5-fold increased expression in Dnmt3c KO compared with Dnmt3c^+/− (DNMT3C sensitive) (Barau et al., 2016) in P20 testes, within young LINEs compared with the remainder of the regions. (C) Proposed model showing the effect of MTHFR deficiency in male germ cells during development. De novo DNA methylation patterns are established normally in male germ cells when MTHFR is expressed at normal high levels in prospermatogonia (blue line). MTHFR is normally expressed at high levels during the phase of late de novo DNA methylation and thus deficiency of MTHFR is predicted to limit methyl donor levels and preferentially impact H3K4me3-marked sequences that would normally be methylated by DNMT3L and DNMT3C after E16.5 days. The worsening of reproductive parameters in MTHFR-deficient sons versus their fathers suggests that epigenetic defects can accumulate across generations. The preferential loss of DNA methylation at young retrotransposons, sequences that are normally kept highly methylated through both PGC and pre-implantation reprogramming phases, could contribute to this effect. Loss of DNA methylation at these specific regions could potentially result in an increased expression of young retrotransposons and lead to germ cell death and subfertility. tub., tubules.