Advanced paternal age (APA) has been shown to be a significant risk factor in the offspring for neurodevelopmental psychiatric disorders, such as schizophrenia and autism spectrum disorders. During aging, de novo mutations accumulate in the male germline and are frequently transmitted to the offspring with deleterious effects. In addition, DNA methylation during spermatogenesis is an active process, which is susceptible to errors that can be propagated to subsequent generations. Here we test the hypothesis that the integrity of germline DNA methylation is compromised during the aging process. A genome-wide DNA methylation screen comparing sperm from young and old mice revealed a significant loss of methylation in the older mice in regions associated with transcriptional regulation. The offspring of older fathers had reduced exploratory and startle behaviors and exhibited similar brain DNA methylation abnormalities as observed in the paternal sperm. Offspring from old fathers also had transcriptional dysregulation of developmental genes implicated in autism and schizophrenia. Our findings demonstrate that DNA methylation abnormalities arising in the sperm of old fathers are a plausible mechanism to explain some of the risks that APA poses to resulting offspring.
Advanced paternal age (APA) is a risk factor for a wide range of health conditions in the offspring. Previously, APA was associated with sporadic cases of rare single gene disorders,1 such as achondroplastic dwarfism, Apert syndrome and Pfeiffer syndrome (reviewed in ref. 2) More recently, numerous studies show that APA is a risk factor for relatively common, polygenic disorders, including neurodevelopmental psychiatric disorders, such as schizophrenia (SCZ),3 autism spectrum disorders (ASD),4,5 bipolar disease6 and attention deficit/hyperactivity disorder.7 However, the mechanism by which APA contributes to these complex disorders remains unresolved.
Male fertility requires ongoing cell division of spermatogonia in the testes throughout life. With each cell division, both genetic and epigenetic information must be preserved in the newly synthesized DNA to maintain stability of the germline. Failure to maintain the fidelity of information contained within the genome of spermatogonia can have a substantial impact on offspring health and has been proposed as a general mechanism of the APA effect.8,9 Recent evidence indicate that de novo single nucleotide variations10, 11, 12, 13 and copy number variations,14, 15 contribute to psychiatric disease and that APA is associated with an increased rate of these types of mutations.16,17
Likewise, aging is also associated with altered DNA methylation in both mammalian somatic and germ cells,18 and has been hypothesized to underlie the paternal age effect.8 Furthermore, DNA methylation abnormalities have been associated with SCZ, ASD and bipolar disease.19,20 Hence, we reasoned that aging effects on sperm DNA methylation might be a factor influencing offspring susceptibility to relatively common, complex neurodevelopmental disorders associated with APA.
Subjects and methods
Paternal age breeding
Old and young father offsprings (OFO and YFO) were generated by breeding 12–14-month and 3-month-old male 129SvEv/Tac mice from Taconic (Hudson, NY, USA), respectively, with 3-month-old female 129SvEv/Tac mice. To control for maternal and litter effects on the offspring, each male was placed with two females to generate litters from different mothers. After 2 weeks, the male mouse was removed to prevent any paternal contact with the offspring and the females were separated into individual cages. Newborn litters were monitored for pup number and survival (Supplementary Table S1). The offspring were left in the cage with the mother until weaning at 21 days. Throughout the breeding all mice were individually marked to determine the paternal and maternal origin of the offspring. Mice were kept on a 7:00 to 19:00 h light cycle.
At 12 weeks of age all male offspring were weighed (Supplementary Table S1) and underwent behavioral testing. Mice received 1 week of rest between each behavioral task. All testing was performed during the light cycle and during the same time of day for each behavior.
Open field (OF)
The mice were placed in the center of a large square (17 × 17 × 12 inches) plexiglas chamber under bright ambient light conditions (800 lux). Activity was monitored in 5 min bins for 30 min using infrared beams. The measures recorded included total ambulatory distance and time, vertical activity and center activity.
Mice were placed in acoustically isolated startle chambers (MED Associates, St Albans, VA, USA). The test started with a 5 min acclimation followed by three consecutive sessions of trials. The background noise was 70 dB throughout the acclimation and trial periods. Sessions 1 and 3 included 10 trials of startle stimuli (120 dB; 40 ms). Session 2 consisted of 56 trials in which startle response magnitude, peak latency and onset latency to each stimulus were recorded for trials, in which the startle stimulus was presented alone or preceded by 100 ms with a 15 ms prepulse. The prepulse amplitude was 2, 4 or 8 dB above background. Startle responses were measured by the downward force (N) applied to a forcemeter between 30 and 70 ms after the onset of the startle stimulus. Prepulse inhibition (PPI) was calculated as 1−ppr[x]/sr, where ppr[x] is the average startle response across trials presenting a prepulse of amplitude x, and sr is the average startle response across trials in which the startle stimulus was presented alone.
Statistical analysis of behavioral data
All behavioral data was analyzed using a two-tailed student t-tests and one- or two-way analysis of variance (ANOVA).
Genomic DNA was isolated from one brain hemisphere of four OFO and four young father offspring (YFO) mice. This preparation included midbrain and forebrain structures, but not the hindbrain or cerebellum. Sperm DNA was isolated from epididymal sperm of the old and young fathers of the above offspring. Methylation Mapping Analysis by Paired-end Sequencing (Methyl-MAPS) libraries were constructed, sequenced and then mapped onto the mouse genome as previously described.21,22 Individual mate-pair libraries were created for each offspring, while sperm DNA was pooled to generate old and young father libraries. Briefly, 7 μg DNA was digested with methylation-dependent (McrBC) and methylation-sensitive (restriction enzyme) enzymes in parallel. McrBC endonuclease generates the unmethylated compartment and is able to interrogate the methylation state of >80% of the CG sites in the genome. The methylated compartment is generated by digestion with a panel of all known methylation-sensitive tetranucleotide restriction enzymes, collectively termed RE (HpaII, HhaI, AciI, BstUI and HpyCH4IV), each of which cuts at a specific 4 bp sequence only if the CG in the recognition site is unmethylated. Libraries of fragments >800 bp in size were constructed utilizing EcoP15’s unique digestion properties. Deep sequencing of digested sequence fragments was performed on the SOLiD sequencing platform from Life Technologies by Beckman Coulter Genomics (Danvers, MA, USA). Mapping of paired-end sequenced fragments were performed with the SOLiD software analysis package (Life Technologies, Grand Island, NY, USA).
Model gene analysis
Methyl-MAPS data analysis pipeline along with estimation of CpG methylation frequencies were previously described.23 Pair-wise comparison between YFO and OFO individual samples did not show significant differences, thus we pooled the YFO and OFO samples to increase coverage as well as to be consistent with the young father sperm (YFS) and old father sperm libraries (OFS), which were pooled DNA from four mice per group. Reads were summed up for YFO and OFO and the combined counts were processed with the Methyl-MAPS data analysis pipeline. For the gene model analyses, genomic feature annotations were downloaded from the UCSC Bioinformatics website (http://genome.ucsc.edu/), including CpG islands (CGIs) and RefSeq gene annotations. All annotations and methylation data were indexed by individual CpG sites and stored in a MySQL database for use in subsequent analyses. CpG dinucleotides were overlapped with the following mutually exclusive genomic features: promoters, first exons, first introns, internal exons, internal introns and last exon. RefSeq gene annotations were based on the Mouse Genome NCBI Build 37. Only genes with complete start and end coding sequences were used in our analysis. The promoter regions were defined as 1 kb upstream of the transcription start site (TSS). We used the definitions of Maunakea et al.24 for CGI promoters and intragenic CGIs.
The gene model was built based on the following gene feature criteria that included: 4000 bp upstream of TSS; 300 bp of the first exon; 10 000 bp of the first intron; 200 bp of the internal exon; 4000 bp of the internal intron; 1000 bp of the last exon; and 2000 bp of the region after poly(A) signal. A total number of 20 496 RefSeq genes that satisfy (1) annotation of complete coding region start and end sites and (2)⩾3 exons were included in the gene model generation. The 5′- and 3′-ends of exons were anchors of the model, where each RefSeq gene was aligned. The methylation level of each position was the average of methylation of corresponding CpGs from Refseq genes. To avoid stochastic variation, we applied a 50-bp non-overlapping smoothing window.
Determination of DNA methylation differences
For each gene and associated genomic features, methylation levels were averaged within each group. These average values were used to compute methylation differences (mDiffs), which were then summed across all genes and corresponding gene features in the gene model. As noted above, 100-bp sliding windows were used to determine mDiffs within genomic regions for each gene feature. We used Wilcox rank sum test to identify regions with statistically significant DNA mDiff between groups and adjusted for multiple testing with the Benjamini–Hochberg method.
Both library construction and sequencing were performed by the Columbia Genome Center. Briefly, poly-A containing messenger RNA was purified using poly-T oligo-attached magnetic beads (200 ng–1 μg per sample, RNA Integrity Number>8 required) and libraries were prepared using the Illumina TruSeq RNA prep kit (Illumina, San Diego, CA, USA). Libraries were then sequenced using an Illumina HiSeq2000 (Illumina) at the Columbia Genome Center. We multiplexed samples in each lane, which yields targeted number of single-end 100 bp reads for each sample, as a fraction of 180 million reads for the whole lane.
Off-Line Basecaller (OLB-1.9.4) was used for base calling and the pass filter reads were mapped to the mouse genome (NCBI37/mm9) using Tophat (version 2.0.4; ref. 25) with 2 mismatches (read-mismatches=2) and 10 maximum multiple hits (max-multihits=10). To tackle the mapping issue of reads that are from exon–exon junctions, Tophat infers novel exon–exon junctions ab initio, and combine them with junctions from known messenger RNA sequences (refgenes) as the reference annotation. We estimated the relative abundance (aka expression level) of genes and splice isoforms using cufflinks (version 2.0.2; ref. 26) with default settings. We tested for differentially expressed genes under various conditions using cuffdiff, a program included in the cufflinks package. The variance of the expression level was estimated based on a negative binomial distribution that models the number reads from high-throughput sequencing experiments.
The Ingenuity Pathways Knowledge Base (IPA) was used to identify enriched functional gene networks and canonical pathways among the differentially expressed genes between OFO vs YFO. The full 17-gene list that resulted from the differential gene expression analysis was used for the analysis. The P-values were calculated by IPA by using a right-tailed Fisher's exact test. A cutoff of P<0.05 was used for significance, as suggested by the software.
To determine the effect of aging on sperm DNA methylation, we performed genome-wide methylation profiling of epididymal sperm DNA pooled from young father sperm (YFS; 3-month-old; n=4) or old father sperm (OFS; 12–14 mo; n=4) using the Methyl-MAPS approach21,22 to query 18 779 892 (88%) CpG sites ≥8 × coverage. We organized the resulting OFS and YFS DNA methylation data by aligning CpGs by their genomic position in relation to the TSS, first, internal and last exons across 20 496 RefSeq genes. We compared DNA methylation changes and we calculated the mDiff by subtracting YFS methylation values from OFS values at each qualifying CpG site (Figure 1a). In the proximal promoter region (±1 kb of TSS), DNA methylation levels were similar between the two groups, with only a small region of significant hypomethylation in OFO 300–550 bp upstream of the TSS. However, further up and downstream of the TSS, we observed a profound loss of CpG methylation in OFS compared with YFS. The difference in methylation between OFS and YFS (Figure 1b) was significant at 300–550 bp, 850–3250 bp upstream of the TSS and 750–3000 bp, 3100–3200 bp downstream of the 3′ position of the first exon (P<0.05–0.01). These regions correspond to CGI shore regions that are known to contain variability in DNA methylation that is associated with changes in gene expression.27
In addition to the methylation changes near the TSS, we also observed a significant loss of methylation in OFS at intron/exon splice junctions. Specifically, there was a significant mDiff 50–800 bp upstream and downstream of internal exons, 450–700 bp upstream and 450–1500 bp, 1700–1750 bp and 1850–1950 bp downstream of the last exon (P<0.05–0.01). Despite these local changes we found no evidence of global changes28 in methylation or hydroxymethylation of mouse sperm DNA (Supplementary Figure S1).
We next sought to determine whether the alterations in sperm DNA methylation observed in OFS would affect brain DNA methylation patterns and behavioral performance of the offspring. Thus, before harvesting the sperm from the old and young fathers, they were mated with 3-month-old females to obtain OFO and YFO. As described in the Methods, each male was mated with two females to determine potential maternal and litter effects. There were no differences between the two groups in litter size and viability at birth, and only a small difference in body weight at the time of behavioral testing (Supplementary Table S1).
OFO and YFO were tested for exploratory and acoustic startle behaviors. OFO displayed significantly decreased exploratory activity (P<0.05; YFO=100, OFO n=87) in the OF compared with YFO (Figure 2a). The effect of APA on exploratory activity is robust and has been observed in several independent cohorts (Supplementary Figure S2). Although the OFO had a slight weight difference compared with YFO, a one-way ANOVA using weight as a covariate found no significant effect of it on exploratory activity. This effect did not appear to be the result of increased anxiety as there was no significant difference between groups in the time spent in the center of the OF (Supplementary Figure S3) or on other anxiety measures (Supplementary Figure S4).
Abnormal prepulse inhibition is observed in both SCZ (reviewed in ref. 29) and ASD.30 We found that the acoustic startle response was significantly reduced in OFO (n=87) compared with YFO (Figure 2b, n=100; P<0.05) and that 72, 76 or 78 db prepulse stimuli (P<0.001) were less effective in reducing startle in OFO (Figure 2c). No significant difference was observed between OFO and YFO in their latency to startle, and weight was not a contributing factor to the effect.
To test for maternal effects contributing to the behavioral deficits in OFO, we performed a two-way ANOVA using maternal identity as an independent variable but found no effect on exploratory activity and prepulse inhibition.
In different cohorts, we expanded the behavioral battery to assess additional endophenotypes of SCZ and ASD, including social interaction, anxiety and learning and memory, but found no effect of APA on these domains (Supplementary Figures S4d–f).
Based on our behavioral findings in OFO and YFO mice, we reasoned that if transmitted patterns of sperm DNA methylation were responsible for these behavioral differences, evidence of altered DNA methylation should be detectable in the brains of these two groups. To test this hypothesis, we performed Methyl-MAPS on DNA from individual brains of OFO and YFO and mapped genic methylation patterns in the same way as performed for the sperm DNA methylation analysis (Figure 3a). Similar to the sperm data, for the brain DNA methylation data we analyzed only high coverage CpG sites (⩾8 × coverage), across all samples comprising 6 283 961 CpG sites equivalent to ~29% of all CpGs in the genome. Statistical analysis of the mDiff between YFO and OFO confirmed that, similar to OFS, OFO have significantly reduced methylation 900–950 bp, 1050–1200 bp upstream of the TSS and 800–1600 bp downstream of the first exon (P<0.05–0.01, Figure 3b). A previous study found global hypermethylation of cerebellar DNA from offspring of aged fathers,31 however, we found no evidence of this in the brains of OFO (Supplementary Figure S1c).
The main effect of APA on both paternal sperm and offspring brain DNA methylation is a loss of methylation at the regions flanking the TSS. These regions, which are about 1 kb upstream and downstream of the TSS, overlap with CGI shores where alterations in DNA methylation is known to be associated with changes in gene expression.27 To determine if the observed changes in DNA methylation in OFS and OFO are specific to CGI promoters and corresponds to CGI shores, we compared the mDiff of both sperm and offspring brain samples at CGI and non-CGI promoters. As shown in Figures 4a and b, there is a significant loss of methylation at the regions flanking both CGI and non-CGI promoters in OFS compared with YFS. However, the mDiff is greater at CGI promoters and the regions corresponding to CGI shores. In the offspring, there is a significant loss of methylation in OFO only in the regions flanking CGI promoters and not in non-CGI promoters. A similar pattern was observed at intragenic CGI shores, with OFS showing significant loss of methylation, however, this was not significantly different in their offspring (Supplementary Figure S5).
Because DNA methylation at CGI shores is associated with alterations in gene expression,27 we performed RNA-seq on the remaining hemisphere of the same offspring, which was used to determine whole-genome methylation patterns. We identified 17 genes as differentially expressed between the two groups (Supplementary Table S2). These genes have mainly neuronal functions and an IPA revealed that 9 out of 17 genes form a network involved in ‘cellular development’, ‘nervous system development and function’, and ‘embryonic development’ (Supplementary Figure S6). The ‘physiological system development and function’ analysis similarly revealed a significant association with ‘nervous system development and function’, which is driven by a significant number of genes which have been previously associated with cerebellar granule cell innervation and morphology (Cbln1, Cbln3, En2 and NeuroD1). The ‘disease and disorder’ analysis revealed a significant association with ‘neurological disease’ (seven genes) and ‘developmental disorders’ (four genes), which appears to be driven by disorders affecting the cerebellum, such as cerebellar ataxia (CA8, Cbln1, Cbln3 and NeuroD1), as well as autism (En2) and mental retardation (CA8). These genes have an important role in the functioning of the nervous system, although to date most associated with cerebellar function and dysfunction. The differentially expressed genes in OFO brains (En2, Cbln1 and NeuroD1) are also associated with regulation of cortical, hippocampal and striatal function, during development.
To determine the relationship between DNA methylation and gene expression in OFO compared with YFO, we performed a correlation analysis of the mDiff values from the Methyl-MAPS data set, with the differential gene expression log2 scores (Supplementary Figure 7). We specifically focused on the mDiff of CpG sites in the distal and proximal promotor region (−1 to −4000 kb from the TSS). Pearson correlation co-efficient revealed a negative correlation (r=−0.01098) that was not quite significant (P=0.2059). This suggests that within our data set, overall, DNA methylation follows the traditional observation of having an inhibitory effect on gene expression, however, there are other influencing factors as well. These are likely to include developmentally regulated expression changes and microRNA expression changes.
This study demonstrates that aging is associated with a loss of sperm DNA methylation at CGI shores, that these abnormalities are also present in the ensuing offspring with effects on both behavior and gene expression. The main effect of APA on offspring behavior, is reduced exploratory activity and PPI, the latter of which is commonly affected in patients with SCZ or ASD (reviewed in refs. 32, 33).
The effects of APA on behavior in mice, has been reported by other groups in past (reviewed by Foldi et al.34), however, the results have been quite variable between different labs. While one group reports a similar decrease in exploratory activity as we observe,35 others report an increase36 or no effect.37 This variability is not simply due to strain differences, but is likely to result from slight variations in the paternal age at breeding, handling of the mice and procedural differences. In our hands, the effect of APA on locomotor activity is relatively small, but consistent between cohorts. Maternal influences on offspring behavior are difficult to exclude. To minimize sire influences on maternal behavior some studies have used either cross-fostering or in vitro fertilization experiments. However, both strategies have been shown to induce epigenetic changes in the offspring38,39 and would likely produce their own confounds.
Paternal transmission of DNA methylation patterns has been reported as a consequence of dietary conditions,40 exposure to environmental toxins41 and odor memory formation.42 However, it is widely held that DNA methylation differences occurring during spermatogenesis are not transmitted to offspring due to the reprogramming of methylation patterns that occurs in the embryo after fertilization. Generally, reprogramming in the preimplantation embryo results in demethylation of the paternal genome43 and subsequent remethylation according to the maternal genome. Nevertheless, some genomic regions, including intracisternal A-particle elements44 and imprinted genes,45 are known to escape reprogramming and the paternal methylation pattern is inherited by the offspring. Interestingly, in mice, evidence suggests that DNA methylation of imprinted genes is altered in the offspring brains of older fathers, however, the methylation state of these genes in the paternal sperm was not explored.31 Our results indicate that some of the DNA methylation abnormalities in OFS are transferred to the offspring, and that reprogramming resets some but not all of these alterations.
It is striking that CGI shores in sperm DNA are particularly sensitive to demethylation with advancing age. At the end of mammalian spermatogenesis, chromatin is extensively remodeled and all but a small percent of nucleosomes are replaced by protamines, allowing the DNA to be packaged more densely. These residual nucleosomes are retained specifically at unmethylated CpG-dense regions, such as CGI promoters,46,47 suggesting a mechanism by which CGI shores are more susceptible to methylation abnormalities in the sperm. Intriguingly, nucleosome retention in sperm is associated with the establishment of DNA methylation-free regions in the early embryo46 and has been proposed as a mechanism of transgenerational inheritance.46,47 Whether this mechanism is involved in the process of how paternal age-associated DNA methylation abnormalities are transferred to the offspring remains to be explored.
Our finding that both CGI shores and splice junctions are abnormally methylated in OFO, suggests that APA can have broad effects on gene expression and cell signaling in the offspring. The RNA-seq results from offspring brain RNA, suggest that most of the genes that are abnormally expressed in the OFO are involved in early stages of neurodevelopment, including brain patterning and synaptogenesis. These are processes that are implicated in the pathology of both SCZ and autism. Some of the genes, such as En2 and CA8, have previously been associated with ASD and mental retardation and the majority of the genes act in signaling networks associated with neurodevelopmental psychiatric disorders, suggesting that these could be novel candidate genes for further exploration.
The exact relationship between APA-associated DNA methylation alterations and previously reported de novo mutations16,17 remains unclear. APA-associated alterations in DNA methylation may be: (1) independent of single nucleotide variations,48 (2) a consequence of mutational events,49 or (3) may influence the mutation rate of certain regions.50 While our current experiments do not address these possibilities, our findings nonetheless demonstrate that age-related DNA methylation abnormalities of paternal origin are a viable candidate mechanism underlying de novo cases of neurodevelopmental disorders such as autism and SCZ.
Risch N, Reich EW, Wishnick MM, McCarthy JG Spontaneous mutation and parental age in humans. Am J Hum Genet 1987; 41: 218–248.
Goriely A, Wilkie AO Paternal age effect mutations and selfish spermatogonial selection: causes and consequences for human disease. Am J Hum Genet 2012; 90: 175–200.
Malaspina D, Harlap S, Fennig S, Heiman D, Nahon D, Feldman D et al. Advancing paternal age and the risk of schizophrenia. Arch Gen Psychiatry 2001; 58: 361–367.
Reichenberg A, Gross R, Weiser M, Bresnahan M, Silverman J, Harlap S et al. Advancing paternal age and autism. Arch Gen Psychiatry 2006; 63: 1026–1032.
Hultman CM, Sandin S, Levine SZ, Lichtenstein P, Reichenberg A . Advancing paternal age and risk of autism: new evidence from a population-based study and a meta-analysis of epidemiological studies. Mol Psychiatry 2011; 16: 1203–1212.
Frans EM, Sandin S, Reichenberg A, Lichtenstein P, Långström N, Hultman CM . Advancing paternal age and bipolar disorder. Arch Gen Psychiatry 2008; 65: 1034–1040.
D'Onofrio BM, Rickert ME, Frans E, Kuja-Halkola R, Almqvist C, Sjölander A et al. Paternal age at childbearing and offspring psychiatric and academic morbidity. JAMA Psychiatry 2014; 71: 432–438.
Malaspina D . Paternal factors and schizophrenia risk: de novo mutations and imprinting. Schizophr Bull 2001; 27: 379–393.
Perrin MC, Brown AS, Malaspina D . Aberrant epigenetic regulation could explain the relationship of paternal age to schizophrenia. Schizophr Bull 2007; 33: 1270–1273.
Sanders SJ, Murtha MT, Gupta AR, Murdoch JD, Raubeson MJ, Willsey AJ et al. De novo mutations revealed by whole-exome sequencing are strongly associated with autism. Nature 2012; 485: 237–241.
Neale BM, Kou Y, Liu L, Ma'ayan A, Samocha KE, Sabo A et al. Patterns and rates of exonic de novo mutations in autism spectrum disorders. Nature 2012; 485: 242–245.
Iossifov I, Ronemus M, Levy D, Wang Z, Hakker I, Rosenbaum J et al. De novo gene disruptions in children on the autistic spectrum. Neuron 2012; 74: 285–299.
O'Roak BJ, Vives L, Girirajan S, Karakoc E, Krumm N, Coe BP et al. Sporadic autism exomes reveal a highly interconnected protein network of de novo mutations. Nature 2012; 485: 246–250.
Malhotra D, McCarthy S, Michaelson JJ, Vacic V, Burdick KE, Yoon S et al. High frequencies of de novo CNVs in bipolar disorder and schizophrenia. Neuron 2011; 72: 951–963.
Sebat J, Lakshmi B, Malhotra D, Troge J, Lese-Martin C, Walsh T et al. Strong association of de novo copy number mutations with autism. Science 2007; 316: 445–449.
Kong A, Frigge ML, Masson G, Besenbacher S, Sulem P, Magnusson G et al. Rate of de novo mutations and the importance of father's age to disease risk. Nature 2012; 488: 471–475.
Flatscher-Bader T, Foldi CJ, Chong S, Whitelaw E, Moser RJ, Burne TH et al. Increased de novo copy number variants in the offspring of older males. Transl Psychiatry 2011; 1: e34.
Richardson B . Impact of aging on DNA methylation. Ageing Res Rev 2003; 2: 245–261.
Dempster EL, Pidsley R, Schalkwyk LC, Owens S, Georgiades A, Kane F et al. Disease-associated epigenetic changes in monozygotic twins discordant for schizophrenia and bipolar disorder. Hum Mol Genet 2011; 20: 4786–4796.
Nguyen A, Rauch TA, Pfeifer GP, Hu VW . Global methylation profiling of lymphoblastoid cell lines reveals epigenetic contributions to autism spectrum disorders and a novel autism candidate gene, RORA, whose protein product is reduced in autistic brain. FASEB J 2010; 24: 3036–3051.
Edwards JR, O'Donnell AH, Rollins RA, Peckham HE, Lee C, Milekic MH et al. Chromatin and sequence features that define the fine and gross structure of genomic methylation patterns. Genome Res 2010; 20: 972–980.
Xin Y, O'Donnell AH, Ge Y, Chanrion B, Milekic M, Rosoklija G et al. Role of CpG context and content in evolutionary signatures of brain DNA methylation. Epigenetics 2011; 6: 1308–1318.
Xin Y, Ge Y, Haghighi F . Methyl-Analyzer—Whole Genome DNA Methylation Profiling. Bioinformatics 2011; 27: 2296–2297.
Maunakea AK, Nagarajan RP, Bilenky M, Ballinger TJ, D'Souza C, Fouse SD et al. Conserved role of intragenic DNA methylation in regulating alternative promoters. Nature 2010; 466: 253–257.
Trapnell C, Pachter L, Salzberg SL . TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 2009; 25: 1105–1111.
Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol 2010; 28: 511–515.
Doi A, Park IH, Wen B, Murakami P, Aryee MJ, Irizarry R et al. Differential methylation of tissue- and cancer-specific CpG island shores distinguishes human induced pluripotent stem cells, embryonic stem cells and fibroblasts. Nat Genet 2009; 41: 1350–1353.
Jenkins TG, Aston KI, Cairns BR, Carrell DT . Paternal aging and associated intraindividual alterations of global sperm 5-methylcytosine and 5-hydroxymethylcytosine levels. Fertil Steril 2013; 100: 945–951.
Swerdlow NR, Weber M, Qu Y, Light GA, Braff DL . Realistic expectations of prepulse inhibition in translational models for schizophrenia research. Psychopharmacology (Berl) 2008; 199: 331–388.
Perry W, Minassian A, Lopez B, Maron L, Lincoln A . Sensorimotor gating deficits in adults with autism. Biol Psychiatry 2007; 61: 482–486.
Smith R.G, Reichenberg A, Kember RL, Buxbaum JD, Schalkwyk LC, Fernandes C et al. Advanced paternal age is associated with altered DNA methylation at brain-expressed imprinted loci in inbred mice: implications for neuropsychiatric disease. Mol Psychiatry 2013; 18: 635–636.
Kohl S, Heekeren K, Klosterkötter J, Kuhn J . Prepulse inhibition in psychiatric disorders—apart from schizophrenia. J Psychiatr Res 2013; 47: 445–452.
Braff DL, Geyer MA, Swerdlow NR . Human studies of prepulse inhibition of startle: normal subjects, patient groups, and pharmacological studies. Psychopharmacology (Berl) 2001; 156: 234–258.
Foldi CJ, Eyles DW, Flatscher-Bader T, McGrath JJ, Burne TH . New perspectives on rodent models of advanced paternal age: relevance to autism. Front Behav Neurosci 2011; 5: 32.
Garcia-Palomares S, Pertusa JF, Miñarro J, García-Pérez MA, Hermenegildo C, Rausell F et al. Long-term effects of delayed fatherhood in mice on postnatal development and behavioral traits of offspring. Biol Reprod 2009; 80: 337–342.
Foldi CJ, Eyles DW, McGrath JJ, Burne TH . Advanced paternal age is associated with alterations in discrete behavioural domains and cortical neuroanatomy of C57BL/6J mice. Eur J Neurosci 2010; 31: 556–564.
Smith RG, Kember RL, Mill J, Fernandes C, Schalkwyk LC, Buxbaum JD et al. Advancing paternal age is associated with deficits in social and exploratory behaviors in the offspring: a mouse model. PLoS One 2009; 4: e8456.
Li L, Wang L, Xu X, Lou H, Le F, Li L et al. Genome-wide DNA methylation patterns in IVF-conceived mice and their progeny: a putative model for ART-conceived humans. Reprod Toxicol 2011; 32: 98–105.
Hager R, Cheverud JM, Wolf JB . Change in maternal environment induced by cross-fostering alters genetic and epigenetic effects on complex traits in mice. Proc Biol Sci 2009; 276: 2949–2954.
Carone BR, Fauquier L, Habib N, Shea JM, Hart CE, Li R et al. Paternally induced transgenerational environmental reprogramming of metabolic gene expression in mammals. Cell 2010; 143: 1084–1096.
Anway MD, Cupp AS, Uzumcu M, Skinner MK . Epigenetic transgenerational actions of endocrine disruptors and male fertility. Science 2005; 308: 1466–1469.
Dias BG, Ressler KJ . Parental olfactory experience influences behavior and neural structure in subsequent generations. Nat Neurosci 2014; 17: 89–96.
Reik W, Dean W, Walter J . Epigenetic reprogramming in mammalian development. Science 2001; 293: 1089–1093.
Lane N, Dean W, Erhardt S, Hajkova P, Surani A, Walter J et al. Resistance of IAPs to methylation reprogramming may provide a mechanism for epigenetic inheritance in the mouse. Genesis 2003; 35: 88–93.
Branco MR, Oda M, Reik W . Safeguarding parental identity: Dnmt1 maintains imprints during epigenetic reprogramming in early embryogenesis. Genes Dev 2008; 22: 1567–1571.
Vavouri T, Lehner B . Chromatin organization in sperm may be the major functional consequence of base composition variation in the human genome. PLoS Genet 2011; 7: e1002036.
Erkek S, Hisano M, Liang CY, Gill M, Murr R, Dieker J et al. Molecular determinants of nucleosome retention at CpG-rich sequences in mouse spermatozoa. Nat Struct Mol Biol 2013; 20: 868–875.
Hitchins MP, Ward RL . Erasure of MLH1 methylation in spermatozoa-implications for epigenetic inheritance. Nat Genet 2007; 39: 1289.
Ligtenberg MJ, Kuiper RP, Chan TL, Goossens M, Hebeda KM, Voorendt M et al. Heritable somatic methylation and inactivation of MSH2 in families with Lynch syndrome due to deletion of the 3' exons of TACSTD1. Nat Genet 2009; 41: 112–117.
Xia J, Han L, Zhao Z . Investigating the relationship of DNA methylation with mutation rate and allele frequency in the human genome. BMC Genomics 2012; 13 (Suppl 8): S7.
We thank Dr Timothy Bestor for invaluable help with experimental design and helpful discussions; Jackie Tinsley, Prashant Donthamsetti, Heather El-Amamy and Matthew Gingrich for experimental help; Caitlin McOmish for helpful discussions. This research was supported by grants from the Simons Foundation and the National Institute of Mental Health (5R21MH073794) to JAG, the G Harold & Leila Y Mathers Foundation to DM and JAG; a NARSAD Young Investigator Award and Sackler Award to MHM.
The authors declare no conflict of interest.
Supplementary Information accompanies the paper on the Molecular Psychiatry website
About this article
Cite this article
Milekic, M., Xin, Y., O’Donnell, A. et al. Age-related sperm DNA methylation changes are transmitted to offspring and associated with abnormal behavior and dysregulated gene expression. Mol Psychiatry 20, 995–1001 (2015). https://doi.org/10.1038/mp.2014.84
Clinical Epigenetics (2021)
A multifactorial model for the etiology of neuropsychiatric disorders: the role of advanced paternal age
Pediatric Research (2021)
Sperm DNA methylation mediates the association of male age on reproductive outcomes among couples undergoing infertility treatment
Scientific Reports (2021)
Journal of Assisted Reproduction and Genetics (2021)
Repeated sampling facilitates within- and between-subject modeling of the human sperm transcriptome to identify dynamic and stress-responsive sncRNAs
Scientific Reports (2020)