Prenatal transportation stress alters genome-wide DNA methylation in suckling Brahman bull calves

Abstract

The objective of this experiment was to identify genome-wide differential methylation of DNA in young prenatally stressed (PNS) bull calves. Mature Brahman cows (n = 48) were transported for 2-h periods at 60 ± 5, 80 ± 5, 100 ± 5, 120 ± 5, and 140 ± 5 d of gestation or maintained as nontransported Controls (n = 48). Methylation of DNA from white blood cells from a subset of 28-d-old intact male offspring (n = 7 PNS; n = 7 Control) was assessed via reduced representation bisulfite sequencing. Samples from PNS bulls contained 16,128 CG, 226 CHG, and 391 CHH (C = cytosine; G = guanine; H = either adenine, thymine, or cytosine) sites that were differentially methylated compared to samples from Controls. Of the CG sites, 7,407 were hypermethylated (at least 10% more methylated than Controls; P ≤ 0.05) and 8,721 were hypomethylated (at least 10% less methylated than Controls; P ≤ 0.05). Increased DNA methylation in gene promoter regions typically results in decreased transcriptional activity of the region. Therefore, differentially methylated CG sites located within promoter regions (n = 1,205) were used to predict (using Ingenuity Pathway Analysis software) alterations to canonical pathways in PNS compared with Control bull calves. In PNS bull calves, 113 pathways were altered (P ≤ 0.05) compared to Controls. Among these were pathways related to behavior, stress response, metabolism, immune function, and cell signaling. Genome-wide differential DNA methylation and predicted alterations to pathways in PNS compared with Control bull calves suggest epigenetic programming of biological systems in utero.

INTRODUCTION

Environmental stimuli or stressors incurred by a gestating dam can alter postnatal phenotypes of the offspring (Clarke et al., 1996; Lay et al., 1997; Mueller and Bale, 2008). In response to a stressor experienced by a gestating female, the placental barrier enzyme 11-βHSD2 might be suppressed. Because 11-βHSD2 functions to convert cortisol to an inactive form before crossing the placenta to fetal circulation, decreased 11-βHSD2 results in increased exposure of the fetus to corticosteroids (Benediktsson et al., 1997; Stirrat et al., 2018). Altered prenatal environment affects postnatal outcomes, in part, by epigenetic modifications, such as DNA methylation (Szyf, 2012). DNA methylation is a covalent modification in which a methyl group is added to the carbon 5 position of a cytosine nucleotide (Wyatt, 1950; Razin and Riggs et al., 1980). This chemical modification of DNA plays a crucial role in regulating unique functions between different cell types, despite each cell type having an identical genome (Razin and Riggs, 1980; Suelves et al., 2016). It is not well understood how a prenatal or life experience might alter DNA methylation differently between cell types, though there is evidence that immune cells and brain cells can be similarly influenced (Provençal et al., 2012; Massart et al., 2016a). Increased methylation of DNA in promoter regions of a gene has been reported to suppress transcription and gene expression (Levine et al., 1991; Tate and Bird, 1993). Alternatively, increased methylation of DNA in gene body regions (i.e., introns and exons) may increase gene expression (Hellman and Chess, 2007). Therefore, characterization of methylation status of DNA in promoter and gene body regions may provide insight regarding the influence of prenatal environment on postnatal phenotype. Beef cattle health, animal welfare/well-being, behavior, and production trait phenotypes are important to producers and consumers. This study investigated how environmental stressors incurred by gestating dams could affect relevant traits in beef cattle, by evaluating the influence of prenatal transportation stress on genome-wide DNA methylation in their offspring. We hypothesized that prenatal transportation stress would alter genome-wide methylation of DNA from white blood cells (WBC) of 28-d-old Brahman bull calves.

MATERIALS AND METHODS

All experimental procedures were in compliance with the Guide for the Care and Use of Agricultural Animals in Research and Teaching (FASS, 2010) and were approved by the Texas A&M University Animal Care and Use Committee.

Animal Procedures

Multiparous Brahman cows were assigned to 1 of 2 treatment groups (Transported: n = 48 and Control: n = 48) according to age, parity, and temperament (Littlejohn et al., 2016). Transported cows were hauled for a 2-h duration at 60 ± 5, 80 ± 5, 100 ± 5, 120 ± 5, and 140 ± 5 d of gestation (Price et al., 2015). Control cows were maintained in the same manner as Transported cows with the exception of being transported. The 2 groups were housed in the same pasture (at the Texas A&M AgriLife Research and Extension Center at Overton, TX) and fed the same diet (Littlejohn et al., 2016). From these cows, 26 male and 18 female calves (Control group) were born to Control dams, while 20 male and 21 female calves [prenatally stressed (PNS) group] were born to Transported dams. Male calves were maintained as bulls throughout the study. Each calf was restrained manually at 28 d of age for less than 5 min for blood sample collection. One 10-mL vacuum tube (BD, Franklin Lakes, NJ) containing EDTA was used for blood sample collection from each calf via jugular venipuncture with a sterile 18-gauge needle. The WBC were isolated and stored at −80 °C until DNA was extracted. For methylation analysis, samples from bull calves were stratified into groups based on the following criteria: 1) adequate availability of DNA, 2) calf treatment group, and 3) calf sire. From those stratified groups, samples from 7 PNS and 7 Control bull calves were randomly selected for this study. Our lab chose to focus on bull calves rather than heifer calves because of the reported transgenerational impact of PNS attributed to the male germline (Rodgers et al., 2013; Rodgers and Bale, 2015).

Sample Analysis

Processing of blood samples

Blood samples were centrifuged at 2,671 × g for 30 min at 6 °C. The WBC layer was isolated and transferred into 2-mL nuclease-free microcentrifuge tubes. Samples were repeatedly washed in red blood cell lysis buffer solution until a clean cell pellet was produced. Specifically, one wash included the following: red blood cell lysis buffer was added to the tube, mixed by vortex for 20 s, shaken for 3 min, centrifuged at 1600 × g for 5 min, and the supernatant was discarded. Clean WBC pellets were stored at −80 °C until they were thawed for DNA extraction.

DNA extraction

Phenol-chloroform extraction procedures were used to isolate DNA from each WBC pellet for methylation analysis. The following description summarizes extraction procedures: Samples were removed from the −80 °C freezer and placed on wet ice for thawing and kept on ice between procedural steps. White blood cell pellets were homogenized in extraction buffer (100 mM NaCl, 10 mM Tris, 1 mM EDTA, pH 7.5), to which 10 mg/mL proteinase K and 20% SDS was added for proteinase K digestion and incubated at 55 °C for 2 h. Samples were extracted twice with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1) and twice with an equal volume of 1-bromo-3-chloropropane (substituted for chloroform). After each extraction, samples underwent centrifugation for 5 min at 8,000 × g. DNA was precipitated by addition of 10% 3 M sodium acetate (pH 5.2) and 1 volume isopropanol to the solution, followed by centrifugation for 5 min at 13,000 × g. Pelleted DNA was rinsed with 70% ethanol, centrifuged at 13,000 × g for 5 min, air-dried, rinsed with 95–100% ethanol, centrifuged, air-dried, and suspended in 150–200 µL TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0). Purified DNA was stored at −80 °C until selected samples were shipped to Zymo Research Corp (Irvine, CA) for DNA methylation analysis.

Library construction for DNA methylation analysis

A method of reduced representation bisulfite sequencing (RRBS), Methyl-MiniSeq (Zymo Research; Irvine, CA), was used to assess differential DNA methylation in samples from PNS compared with Control bull calves. Libraries were prepared from 200–500 ng of genomic DNA digested with 60 units of TaqαI and 30 units of MspI (NEB) sequentially and then extracted with Zymo Research (ZR) DNA Clean & Concentrator-5 kit (Cat#: D4003). Fragments were ligated to preannealed adapters containing 5′-methyl-cytosine instead of cytosine according to Illumina’s specified guidelines (www.illumina.com). Adaptor-ligated fragments of 150–250 bp and 250–350 bp in size were recovered from a 2.5% NuSieve 1:1 agarose gel (Zymoclean Gel DNA Recovery Kit, ZR Cat#: D4001). The fragments were then bisulfite-treated using the EZ DNA Methylation-Lightning Kit (ZR, Cat#: D5020). Preparative-scale PCR was performed and the resulting products were purified (DNA Clean & Concentrator - ZR, Cat#D4005) and sequenced on an Illumina HiSeq.

Methyl-MiniSeq sequence alignments and data analysis

Sequence reads from bisulfite-treated EpiQuest libraries were identified with standard Illumina base-calling software and analyzed via a Zymo Research proprietary analysis pipeline, which is written in Python and used Bismark (http://www.bioinformatics.babraham.ac.uk/projects/bismark/) to perform the alignment. Bismark is used to map bisulfite converted sequence reads to determine methylation status of cytosine nucleotides. Index files were constructed using the bismark_genome_preparation command (creates a subdirectory in the fasta file directory) and the entire reference genome. The --non_directional parameter was applied while running Bismark. All other parameters were set to default. Filled-in nucleotides were trimmed off for methylation calling. Methylation ratio was defined as the measured number of cytosines (number of reads reporting a C) divided by the total number of cytosines (total number of reads reporting a C or T) covered at that site. Each methylation difference was calculated by subtracting the average Control methylation ratio at a site from the average PNS methylation ratio at a site. Fisher’s exact test or t-test was performed for each CpG site which has at least 5 reads coverage, and promoter, gene body and CpG island annotations were added for each CpG included in the comparison.

Prediction of pathways and functions altered by prenatal stress

An enrichment analysis was executed (analyzed on April 3, 2018) using Ingenuity Pathway Analysis software (IPA; Redwood City, CA) to assess alterations to signaling pathways and biological functions in PNS compared to Control bull calves by overlaying genes with differentially methylated (i.e., at least 10% more or less methylated than Controls) CpG sites onto networks generated by IPA. All P values (P values of overlap) for those pathways and functions were calculated using a right-tailed Fisher exact test. These measured the probability of association of the significant genes with a pathway or function group by random chance alone. Only differentially methylated CpG sites (P ≤ 0.05) that had a methylation difference of ≥ 10% and were located within promoter regions were utilized in the enrichment analysis. The promoter region was selected, because DNA methylation in the promoter region regulates gene activity. Specifically, increased DNA methylation within gene promoter regions has been reported to cause suppressed transcription and gene expression (Levine et al., 1991; Tate and Bird, 1993). Therefore, for the purpose of predicting alterations to signaling pathways, hypomethylation within gene promoters was assumed to result in increased gene activity, while hypermethylation within gene promoters was assumed to result in decreased gene activity. It is important to note that genes can be inhibitors or activators of specific cellular processes. Therefore, regulation of gene expression by increased or decreased DNA methylation can result in activation or inhibition of cellular processes, depending on the function of that specific gene. For example, gene expression that is down-regulated by increased DNA methylation in an inhibitory gene could result in activation of a cellular process.

RESULTS AND DISCUSSION

Genome-wide DNA Methylation in PNS and Control Bull Calves

Coverage summary

Samples from Control and PNS calves had an average total read number of 38,534,348 and 34,894,063 read pairs, respectively. Both Control and PNS samples had an average mapping efficiency of 33% and bisulfite conversion rate of 99%. Sequence depths of unique CpG sites were 6,392,121 (9.0 times) for samples from Control and 6,358,081 (8.2 times) for samples from PNS bull calves on average. Sequence depths of unique CHG sites were 11,046,912 (7.3 times) for samples from Control and 10,985,061 (6.6 times) for samples from PNS bull calves on average. Sequence depths of unique CHH sites were 27,388,851 (6.9 times) for samples from Control and 27,218,778 (6.3 times) for samples from PNS bull calves on average.

CpG sites

A summary of genome-wide distribution of differential DNA methylation in CpG sites (i.e., in the CpG context; defined as a cytosine followed by a guanine) across promoters, introns, exons, and intergenic regions in PNS compared with Control bull calves is located in Table 1. Briefly, 16,128 CpG sites (52.76% of which were located within CpG islands) were differentially methylated in PNS compared to Control bull calves (P ≤ 0.05). The majority of differentially methylated CpG sites were found in intergenic regions (65.5%), followed by introns (19.2%), exons (7.8%), and promoters (7.5%). Of those affected sites, 45.93% were hypermethylated (at least 10% more methylated than Controls; P ≤ 0.05) and 54.07% were hypomethylated (at least 10% less methylated than Controls; P ≤ 0.05). Methylation of DNA in mammals primarily occurs within CpG contexts (Ehrlich et al., 1982), although reports of substantial DNA methylation within non-CpG contexts (i.e., in the CHG or CHH context; in which C = cytosine; G = guanine; H = either adenine, thymine, or cytosine) have been established in embryonic stem cells, gametes, and brain cells (Ramsahoye et al., 2000; Shirane et al., 2013; Varley et al., 2013). Previously, methylation of cytosine nucleotides in non-CPG contexts was thought to vanish upon cellular differentiation (Lister et al., 2009). However, methylation of DNA in non-CpG contexts of differentiated mammalian cell types has gained attention in various tissues in recent years (Ziller et al., 2011; Barua et al., 2014; Zhou et al., 2016).

Table 1.

Summary of genome-wide distribution of differential DNA methylation (HYPER= Hypermethylation, HYPO=Hypomethylation) across CpG, CHG, and CHH sites in prenatally stressed (PNS) compared with Control bull calves^1,2,3

Genomic region	Affected regions (N)	Percent of total regions affected	HYPER regions (N)	4Percent of affected regions with HYPER	HYPO regions (N)	Percent of affected regions with HYPO	HYPER regions within CpG Islands (N)	Percent of HYPER regions within CpG Islands	HYPO regions within CpG Islands (N)	Percent of HYPO regions within CpG Islands
CpG Sites
Promoter	1,205	7.5%	543	3.37%	662	4.10%	307	4.14%	364	4.17%
Intron	3,103	19.2%	1,386	8.59%	1,717	10.65%	303	4.09%	354	4.06%
Exon	1,260	7.8%	602	3.73%	658	4.08%	364	4.91%	383	4.39%
Intergenic	10,560	65.5%	4,876	30.23%	5,684	35.24%	1,025	13.84%	1,146	13.14%
Total	16,128	100.0%	7,407	45.93%	8,721	54.07%	1,999	26.99%	2,247	25.77%
CHG Sites
Promoter	10	4.4%	3	1.33%	7	3.10%	1	0.93%	2	1.69%
Intron	55	24.3%	28	12.39%	27	11.95%	1	0.93%	2	1.69%
Exon	15	6.6%	9	3.98%	6	2.65%	3	2.78%	2	1.69%
Intergenic	146	64.6%	68	30.09%	78	34.51%	5	4.63%	7	5.93%
Total	226	100.0%	108	47.79%	118	52.21%	10	9.26%	13	11.02%
CHH Sites
Promoter	14	3.6%	6	1.53%	8	2.05%	2	0.96%	1	0.55%
Intron	121	30.9%	65	16.62%	56	14.32%	4	1.91%	5	2.75%
Exon	12	3.1%	6	1.53%	6	1.53%	2	0.96%	1	0.55%
Intergenic	244	62.4%	132	33.76%	112	28.64%	10	4.78%	5	2.75%
Total	391	100.00%	209	53.45%	182	46.55%	18	8.61%	12	6.59%

¹Affected regions were considered P ≤ 0.05.

²In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

³Differentially methylated sites were at least 10% more or less methylated than Controls.

⁴Percent of hypermethylated regions and percent of hypomethylated regions add up to 100%.

Non-CpG sites

A summary of genome-wide distribution of differential DNA methylation in CHG sites across promoters, introns, exons, and intergenic regions in PNS compared with Control bull calves is located in Table 1. Briefly, 226 CHG sites (20.28% of which were located within CpG islands) were differentially methylated in PNS compared to Control bull calves (P ≤ 0.05). The majority of differentially methylated CHG sites were found in intergenic regions (64.6%), followed by introns (24.3%), exons (6.6%), and promoters (4.4%). Of those affected sites, 47.79% were hypermethylated and 52.21% were hypomethylated.

A summary of genome-wide distribution of differential DNA methylation in CHH sites across promoters, introns, exons, and intergenic regions in PNS compared with Control bull calves is located in Table 1. Briefly, 391 CHH sites (15.20% of which were located within CpG islands) were differentially methylated in PNS compared to Control bull calves (P ≤ 0.05). The majority of differentially methylated CHH sites were found in intergenic regions (62.4%), followed by introns (30.9%), promoters (3.6%), and exons (3.1%). Of those affected sites, 53.45% were hypermethylated and 46.55% were hypomethylated.

Zhou et al. (2016) assessed DNA methylation of 10 bovine cell types using RRBS and found that of the CpG-enriched regions, 33.5% were in the CpG context, 1.1% were in the CHG context, and 1.5% were in the CHH context. Of the cytosines that were differentially methylated compared to one or more of the other 10 cell types, 94.34% were in the CpG context and 5.66% were in the non-CpG context (Zhou et al., 2016). Furthermore, Barua et al. (2014) reported differential methylation of CHG and CHH contexts due to prenatal programming in rodents. Thus, increasing evidence suggests an influence of DNA methylation in non-CpG contexts in differentiated mammalian cells.

Differential DNA Methylation in CpG Sites

Promoter regions

A total of 1,205 differentially methylated CpG sites were identified within promoter regions in PNS compared with Control bull calves (Table 1). A greater percentage of these CpG sites were hypomethylated (54.07%) compared to hypermethylated (45.93%; Table 1). Richetto et al. (2017) reported a similar occurrence in mice that were exposed to a prenatal viral challenge on gestational day 9 or 17, with 64% and 61% of differentially methylated regions being hypomethylated and 36% and 39% being hypermethylated, respectively. Differential methylation within promoter regions in PNS compared with Control bulls was distributed throughout the genome, with all chromosomes having both hypomethylated and hypermethylated CpG sites (Fig. 1). The number of hypermethylated compared with hypomethylated CpG sites within promoter regions in PNS compared with Control bulls is represented in Fig. 1. Increased DNA methylation within gene promoter regions has been reported to cause suppressed transcription and gene expression (Levine et al., 1991; Tate and Bird, 1993). This suggests that hypermethylated CpG sites within gene promoter regions might have downregulated gene expression and hypomethylated CpG sites within gene promoter regions might have upregulated gene expression in PNS calves compared with Controls.

Figure 1.

Comparison of hypermethylated and hypomethylated CpG sites within promoter regions in prenatally stressed (PNS) compared with Control bull calves.

Strongly differentially methylated CpG sites (“strongly” being defined as a degree of methylation in PNS calves that was at least 33% different from Controls) were specifically examined to highlight those sites that exhibited the greatest changes in degree of methylation due to PNS treatment. In agreement with this study’s overall findings, a greater percentage of strongly hypomethylated (methylation difference ≤ -0.33 ratio) compared to strongly hypermethylated (methylation difference ≥ 0.33 ratio) CpG sites were identified within promoter regions (P ≤ 0.05) in PNS bull calves compared with Controls. Six strongly hypermethylated CpG sites were located within promoter regions in PNS compared with Control calves (Table 2). Among these were 2 CpG sites that were located within the promoter region of the NQO2 gene. This hypermethylated gene was also involved in the upregulation of the “NRF2-mediated Oxidative Stress Response” pathway in PNS calves (Supplementary Table S1). Furthermore, a base pair deletion in the promoter region of this gene has been associated with schizophrenia in humans (Harada et al., 2003), suggesting the potential for physiological and behavioral alterations due to differences in neurotransmitter pathways.

Table 2.

Strongly hypermethylated (methylation difference ≥ 0.33 ratio) CpG sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr7†	CCDC105	9178140	9178141	−	7.86	8.71	0.36	0.003097
chr11†*	MRPL41	105596122	105596123	−	8.57	7.00	0.35	0.003315
chr12†	PCDH17	5200175	5200176	−	11.14	7.29	0.34	0.0004446
chr21*	MIR655	67587446	67587447	+	8.57	9.43	0.45	0.008982
chr23	NQO2	50497771	50497772	−	6.71	5.86	0.39	0.002682
chr23†	NQO2	50497904	50497905	−	6.71	5.86	0.42	0.02622

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region.

The top (lowest P values) 30 of 543 significantly (P ≤ 0.05) hypermethylated CpG sites located within promoter regions in PNS compared with Control calves are listed in Table 3. Among these was a CpG site within the promoter region of the Caudal Type Homeobox 2 (CDX2) gene. CDX2 is involved in early embryo pluripotency, differentiation, and development (Xie et al., 2013). Therefore, it is consistent that this gene was also involved in the alterations to the pathways, “Role of NANOG in Mammalian Embryonic Stem Cell Pluripotency, Transcriptional Regulatory Network in Embryonic Stem Cells, Role of Oct4 in Mammalian Embryonic Stem Cell Pluripotency” in PNS calves (Supplementary Table S1). In vitro hyperosmolar stress resulted in reduced CDX2 and inhibited potency in the early embryo of the mouse (Xie et al., 2013). Bovine embryos exposed to heat stress exhibited reduced CDX2 gene expression in total blastocyst RNA (Silva et al., 2013). These studies agree with the current study, which is predictive of downregulated expression of CDX2 due to increased DNA methylation in the promoter region of CDX2 in PNS bulls compared with Controls.

Table 3.

Top (lowest P values) 30 hypermethylated CpG sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1†	CLSTN2	130108179	130108180	−	11.71	13.00	0.11	0.000009504
chr1	CLDN8	5036161	5036162	+	16.57	14.29	0.13	0.0003032
chr2†	TMEM200B	125008770	125008771	+	12.43	9.71	0.13	0.00028
chr3†	SHE	16143944	16143945	+	7.14	12.29	0.18	0.0009009
chr7†	HNRNPM	18289571	18289572	−	9.43	8.71	0.21	0.0002804
chr7†	HNRNPM	18289557	18289558	−	9.43	8.71	0.22	0.0006671
chr7	SIL1	52143957	52143958	+	13.00	10.71	0.1	0.001113
chr7†	HNRNPM	18289508	18289509	−	9.43	8.57	0.21	0.001569
chr7†	HNRNPM	18289493	18289494	−	9.29	8.71	0.22	0.001877
chr7†	HNRNPM	18289497	18289498	−	9.43	8.71	0.22	0.001877
chr7†	HNRNPM	18289505	18289506	−	9.43	8.71	0.22	0.001877
chr10*	EIF3J	103983752	103983753	+	9.43	8.86	0.13	6.067E−07
chr10†*	SIX6	72952940	72952941	−	13.00	10.14	0.19	0.0001318
chr10†	SHC4	61503850	61503851	+	8.86	9.14	0.17	0.001072
chr10*	EIF3J	103983746	103983747	+	9.29	8.86	0.14	0.001454
chr10	SHC4	61504208	61504209	−	9.86	7.29	0.15	0.001506
chr10†	LHFPL2	9527046	9527047	+	13.00	8.86	0.11	0.001618
chr11	BRE, RBKS	71827429	71827430	−	12.14	10.86	0.3	0.0007929
chr12*	CDX2	32316737	32316738	−	18.86	13.00	0.25	0.0002012
chr12†	PCDH17	5200175	5200176	−	11.14	7.29	0.34	0.0004446
chr13†	BMP7	59425545	59425546	−	8.00	8.00	0.11	0.001394
chr15	MMP7	6390211	6390212	+	8.57	6.29	0.19	0.000205
chr15†	CHRM4	77253713	77253714	+	10.71	6.14	0.13	0.000309
chr16†*	MMP23B	52263865	52263866	+	10.71	8.86	0.12	0.001214
chr16†*	SLC45A1	45908696	45908697	−	11.57	9.00	0.11	0.001281
chr19	ITGA3	37230331	37230332	+	12.29	11.00	0.12	0.0000896
chr19*	CORO6	21435132	21435133	+	21.00	18.00	0.17	0.001546
chr25†	METRN	613852	613853	−	8.71	6.00	0.12	0.001253
chr26	ZWINT	2851210	2851211	+	10.00	11.71	0.15	0.0001167
chr28†	TRIM67	3519622	3519623	+	18.14	19.14	0.11	0.001037

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region.

There were 22 strongly hypomethylated CpG sites located within promoter regions in PNS compared with Control calves (Table 4). Among these was a CpG site within the promoter region of the Guanine Nucleotide-Binding Protein G(S) Subunit Alpha (GNAS) gene. In cattle and other mammals, GNAS is an imprinted gene that has been reported to play a major role in development, growth, and metabolism (Plagge et al., 2004; Sikora et al., 2011). DNA methylation of a paternally expressed transcript of this gene, GNASXL, has been reported to be positively associated with prenatal maternal stress (as quantified by increased depression, anxiety, and cortisol) in humans (Vangeel et al., 2015). This is in contrast with the hypomethylated CpG site within the promoter region of the GNAS gene in PNS bull calves in this study.

Table 4.

Strongly hypomethylated (methylation difference ≤ −0.33 ratio) CpG sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	MethylDiff	P value
chr1*	EPCAM	157207666	157207667	−	7.4	8.9	−0.34	0.0495
chr1†*	GNAS	154101009	154101010	−	10.6	5.4	−0.35	0.0147
chr7*	TNFRSF10D	18053077	18053078	−	10.9	16.0	−0.38	0.01531
chr8	KHNYN, CBLN3	104517901	104517902	−	11.9	9.0	−0.35	0.03365
chr9	PLA2G4B	103674611	103674612	+	14.7	6.9	−0.44	0.04478
chr11	ATF6B	103182975	103182976	+	13.7	9.4	−0.33	0.002325
chr11†	SNX19	100835712	100835713	−	8.0	11.4	−0.35	0.006669
chr13†	LOXL4	58167618	58167619	−	12.7	12.4	−0.34	0.01073
chr13†	MGRN1	58167617	58167618	+	11.3	23.0	−0.33	0.004263
chr13†	YDJC	54817882	54817883	−	6.7	10.9	−0.39	0.007604
chr17†	YDJC	73172496	73172497	−	7.3	12.0	−0.45	0.009104
chr19	ATF6B	22423681	22423682	−	14.3	14.0	−0.38	0.01513
chr20*	CSNK1D	69935936	69935937	−	9.0	9.4	−0.39	0.01086
chr21†*	GNAS	30886861	30886862	−	11.3	9.6	−0.33	0.01602
chr23†	MGRN1	51619275	51619276	−	11.0	8.6	−0.33	0.004263
chr23†	SNX19	17940352	17940353	−	9.4	11.1	−0.35	0.02976
chr23†	TMOD1	51239299	51239300	+	11.6	9.0	−0.35	0.0389
chr23*	UBXN8	48211861	48211862	−	10.9	6.3	−0.37	0.001421
chr24†	MGRN1	3104521	3104522	−	10.9	7.0	−0.33	0.004263
chr25*	CARD9	1721752	1721753	−	9.7	9.7	−0.34	0.002587
chr25	MIR99B, MIRLET7E	4471948	4471949	+	8.4	6.3	−0.38	0.005474
chr25*	TNFRSF10D	2238087	2238088	+	12.3	9.1	−0.37	0.02807

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region.

The top (lowest P values) 30 of 662 significantly (P ≤ 0.05) hypomethylated CpG sites located within promoter regions in PNS compared with Control calves are listed in Table 5. Among these was a CpG site within the promoter region of the Dopamine Receptor D1 (DRD1) gene. This gene has been associated with behavioral disorders such as psychosis and schizophrenia (Andreou et al., 2016). The DRD1 gene was involved in the following pathways altered in PNS calves in this study: “cAMP-mediated signaling, G-Protein Coupled Receptor Signaling, Gαs Signaling, Dopamine-DARPP32 Feedback in cAMP Signaling, CDK5 Signaling, Gap Junction Signaling, and Dopamine Receptor Signaling” (Supplementary Table S1). Rat pups that were separated from their mothers for 6-h periods each day during the first 2 wk of life had downregulated DRD1 gene expression in the nucleus accumbens (Zhu et al., 2010). Rat pups that were separated from their mothers for 3-h periods had decreased mid-brain tyrosine hydroxylase-immunoreactive dopaminergic neurons as juveniles (15 d of age) but increased numbers as adolescents (35 d of age) and adults (70 d of age; Chocyk et al., 2011). These varying results suggest that alterations to the DRD1 gene due to prenatal or early life stress are specific to life stage.

Table 5.

Top (lowest P values) 30 hypomethylated CpG sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr3†	ECHDC2	94016695	94016696	+	9.14	12.29	−0.28	0.0006247
chr4†	OSBPL3	71296952	71296953	+	17.29	15.57	−0.13	0.001308
chr7	FDX1L	16072432	16072433	+	9.43	9.57	−0.18	0.0004641
chr7†	NMRK2	21294100	21294101	+	10.00	7.43	−0.16	0.000824
chr8	MIR2887-2, MIR2887-1, MIR2904-1, MIR2904-3, MIR2904-2	59158057	59158058	+	9.71	13.57	−0.23	0.0008405
chr8†*	XKR6	8255834	8255835	−	14.57	7.43	−0.11	0.001542
chr9*	SEC63	42650623	42650624	+	14.14	9.29	−0.17	0.0002254
chr9	OPRM1	92152196	92152197	+	8.14	9.43	−0.13	0.0002767
chr9	OPRM1	92152194	92152195	+	8.14	9.43	−0.2	0.001245
chr10*	DRD1	5653417	5653418	+	9.14	7.86	−0.19	0.0003017
chr10	LTB4R	20689223	20689224	−	8.43	9.86	−0.16	0.0007834
chr11†	SIX2	27263419	27263420	+	21.00	17.00	−0.11	0.0005395
chr11†*	SURF4	104335032	104335033	−	11.29	11.14	−0.14	0.001013
chr15†	APLNR	81738036	81738037	+	25.57	24.57	−0.23	0.0004384
chr18*	EGLN2	50330819	50330820	−	8.00	6.57	−0.15	0.0007258
chr18†*	NAT14	62474955	62474956	−	11.57	10.86	−0.25	0.001256
chr18	FUZ	56625931	56625932	−	10.57	8.43	−0.3	0.001281
chr19†	PMP22	33382476	33382477	−	11.43	12.57	−0.12	0.00007101
chr19	ERBB2	40722503	40722504	+	13.14	9.57	−0.1	0.0004212
chr19*	MPDU1	27924874	27924875	−	8.14	6.57	−0.15	0.001674
chr21†*	NKX2-8	47199004	47199005	+	12.86	14.86	−0.24	0.0003029
chr21*	CHRNA7	30181352	30181353	−	11.00	7.29	−0.29	0.0005365
chr21†	CSPG4	33574762	33574763	−	7.14	6.86	−0.17	0.00072
chr22†	ITGA9	10946832	10946833	−	17.00	17.43	−0.26	0.001275
chr25†	PGP	1746496	1746497	+	18.57	17.57	−0.25	0.001152
chr26	MIR2397	49743379	49743380	+	12.43	9.57	−0.14	0.00004049
chr26	PRLHR	39220962	39220963	+	11.43	9.29	−0.22	0.00008747
chr27*	UBXN8	26008284	26008285	−	10.00	8.57	−0.37	0.001421
chr29*	TALDO1	50864154	50864155	−	8.57	8.14	−0.27	0.0001148
chr29*	AP2A2	50445180	50445181	+	10.71	7.71	−0.2	0.0004282

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region.

Gene body regions

Differential methylation of 4,363 CpG sites were identified within gene body (i.e., introns and exons) regions in PNS compared with Control bull calves (Table 1). A greater percentage of these CpG sites were hypomethylated (Table 1). Differential methylation within gene body regions in PNS compared with Control bulls was distributed throughout the genome, with both hypomethylation and hypermethylation observed in CpG sites across all chromosomes (Fig. 2). The number of hypermethylated compared with hypomethylated CpG sites within gene body regions in PNS compared with Control bulls are represented in Fig. 2. Increased DNA methylation in gene bodies may result in activated gene expression (Hellman and Chess, 2007), suggesting that hypermethylated gene body regions might have upregulated gene expression and hypomethylated gene body regions might have downregulated gene expression.

Figure 2.

Comparison of hypermethylated and hypomethylated CpG sites within gene body regions in prenatally stressed (PNS) compared with Control bull calves.

The top (lowest P values) 30 of 1,988 significantly (P ≤ 0.05) hypermethylated CpG sites located within gene body regions in PNS compared with Control calves are listed in Table 6. Among these was a CpG site within the gene body region of the Iodothyronine Deiodinase 3 (DIO3) gene, suggesting a potential upregulation in gene expression. The DIO3 gene is an imprinted gene in mice, cattle, and other species that encodes a protein, which inactivates thyroid hormones (Tsai et al., 2002; Yang et al., 2017). Thyroid hormones influence central nervous system development (Bernal, 2005) and behavior (Stohn et al., 2018). Male and female mice deficient in DIO3 exhibited increased aggression in response to an intruder (Stohn et al., 2018). Furthermore, rats whose dams consumed ethanol from day 8 to 21 of gestation had increased placental DIO3 mRNA compared to Controls (Shukla et al., 2011). These studies suggest that prenatal environment can shape nervous system development through alterations to the DIO3 gene.

Table 6.

Top (lowest P values) 30 hypermethylated CpG sites located within gene body regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1§	CLDN8	5036161	5036162	+	16.57	14.29	0.13	0.0003032
chr1†‡	CLSTN2	130108179	130108180	−	11.71	13.00	0.11	0.000009504
chr1*‡	EPHB1	135353859	135353860	−	12.43	12.29	0.16	0.00003959
chr2†‡	TMEM200B	125008770	125008771	+	12.43	9.71	0.13	0.00028
chr3†*§	DMRTA2	96456243	96456244	−	19.29	15.29	0.15	0.0003634
chr3*‡	C3H1orf52	59027290	59027291	−	6.29	8.71	0.15	0.00001461
chr4*‡	NUDCD3	77653827	77653828	−	5.86	6.29	0.4	0.0004187
chr5*‡	ZNF384	104130929	104130930	+	7.71	8.14	0.15	0.0003626
chr7*§	F12	40258637	40258638	−	7.71	5.71	0.13	0.0004367
chr7†§	HNRNPM	18289571	18289572	−	9.43	8.71	0.21	0.0002804
chr7*‡	CRTC1	4435796	4435797	+	11.29	10.14	0.14	0.0004044
chr11*§	SLC8A1	22824023	22824024	+	9.57	8.86	0.29	0.0001272
chr11*‡	EHMT1	105461042	105461043	+	7.29	6.57	0.63	0.0003017
chr11*‡	RALGDS	103166859	103166860	−	7.14	6.29	0.55	0.00003935
chr11*‡	RALGDS	103166862	103166863	−	7.14	6.29	0.48	0.0001696
chr11*‡	RALGDS	103166855	103166856	−	7.14	6.29	0.55	0.0003
chr12†*‡	CARS2	89257092	89257093	−	11.71	9.00	0.2	0.0001168
chr12*‡	DOCK9	79833906	79833907	+	10.00	6.00	0.2	0.0002803
chr12†*‡	GRTP1	90628533	90628534	−	6.57	6.43	0.11	0.0004244
chr15†§	CHRM4	77253713	77253714	+	10.71	6.14	0.13	0.000309
chr15§	MMP7	6390211	6390212	+	8.57	6.29	0.19	0.000205
chr15†*‡	PHOX2A	52625075	52625076	−	6.29	6.14	0.17	0.0002289
chr19‡	ITGA3	37230331	37230332	+	12.29	11.00	0.12	0.0000896
chr21†*§	DIO3	68152290	68152291	+	14.86	14.14	0.18	0.0001115
chr22*‡	IL17RD	44515500	44515501	−	18.57	18.14	0.27	0.00007418
chr23*‡	MDGA1	11586464	11586465	−	21.14	15.43	0.17	0.0004023
chr23*‡	PHACTR1	43380907	43380908	+	10.14	9.29	0.11	0.0002705
chr26‡	ZWINT	2851210	2851211	+	10.00	11.71	0.15	0.0001167
chr28*§	C28H10orf35	26104046	26104047	+	14.00	8.57	0.15	0.0003049
chr29†*§	NRXN2	43477994	43477995	+	10.57	6.71	0.28	0.0001266

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CpG island.

*DNA methylation was exclusively contained in the gene body region of the gene.

‡DNA methylation was located within an intron region.

§DNA methylation was located within an exon region.

The top (lowest P values) 30 of 2,375 significantly (P ≤ 0.05) hypomethylated CpG sites located within gene body regions in PNS compared with Control calves are listed in Table 7. Among these was a CpG site within the gene body region of the Adenylate Cyclase Activating Polypeptide 1 (ADCYAP1) gene, suggesting a potential downregulation in gene expression. The ADCYAP1 gene encodes for the PACAP peptide and is associated with stress disorders and posttraumatic stress disorder (Ressler et al., 2011). Mice that were deficient in PACAP exhibited behavioral abnormalities such as hyperactivity, jumping, and depression (Ishihama et al., 2010). Furthermore, mice whose dams were exposed to a posttraumatic stress disorder model (restrained for 2-h period and then forced to swim for a 20-min period with 6 other rats) had downregulated ADCYAP1 gene expression (Zhang et al., 2016). These studies support the potential influence of prenatal stress on the ADCYAP1 gene in this study and suggest altered behavioral effects due to prenatal stress. Behavioral alterations were previously observed in PNS calves (Littlejohn et al., 2016).

Table 7.

Top (lowest P values) 30 hypomethylated CpG sites located within gene body regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr2*§	CRYBA2	107655452	107655453	+	17.86	12.57	−0.28	0.0001131
chr2†*§	TMEM200B	125010636	125010637	+	9.00	7.14	−0.12	6.12E−08
chr2*‡	DPP10	68059420	68059421	+	13.43	10.43	−0.29	0.00007136
chr3*§	LRRC42	92815815	92815816	−	15.57	11.71	−0.12	0.0003827
chr3*‡	GRIK3	109641283	109641284	−	13.86	9.43	−0.15	0.00006294
chr5*§	FMNL3	30363597	30363598	+	10.14	13.00	−0.16	0.00001178
chr6†*‡	NSG1	106743364	106743365	−	11.14	8.71	−0.18	0.0001679
chr7‡	FDX1L	16072432	16072433	+	9.43	9.57	−0.18	0.0004641
chr9§	OPRM1	92152196	92152197	+	8.14	9.43	−0.13	0.0002767
chr11†*‡	GBGT1	103182975	103182976	+	9.43	9.43	−0.6	0.0001948
chr11†*‡	NR5A1	95535022	95535023	−	12.57	11.57	−0.12	0.00006977
chr13*‡	CUBN	31602752	31602753	−	24.86	31.14	−0.1	0.0002177
chr13*‡	TGIF2	66340155	66340156	+	8.86	10.14	−0.11	0.0003646
chr14†*‡	AGO2	4114895	4114896	−	7.00	6.14	−0.43	0.000509
chr14*‡	CPQ	69484191	69484192	+	9.86	9.71	−0.1	0.000005467
chr15†§	APLNR	81738036	81738037	+	25.57	24.57	−0.23	0.0004384
chr16*‡	ACOT7	47839343	47839344	+	6.43	7.00	−0.11	0.0001594
chr16*‡	RGS7	36591986	36591987	−	23.86	20.29	−0.13	0.000143
chr17*‡	SELM	72065070	72065071	−	16.14	18.14	−0.27	0.0003458
chr18*‡	CBLC	52956154	52956155	+	8.86	7.86	−0.2	0.0002327
chr19*§	ABR	22423681	22423682	−	13.71	14.00	−0.64	0.00003286
chr19‡	ERBB2	40722503	40722504	+	13.14	9.57	−0.1	0.0004212
chr19†‡	PMP22	33382476	33382477	−	11.43	12.57	−0.12	0.00007101
chr22*‡	ALDH1L1	61235727	61235728	+	17.00	12.57	−0.15	0.0000172
chr22*‡	LTF	53523469	53523470	−	11.14	8.14	−0.24	0.0003724
chr24†*‡	ADCYAP1	36118731	36118732	+	13.29	10.43	−0.16	0.0004943
chr26§	PRLHR	39220962	39220963	+	11.43	9.29	−0.22	0.00008747
chr26‡	GLRX3	49743379	49743380	+	12.43	9.57	−0.14	0.00004049
chr27*‡	MTUS1	18657599	18657600	−	10.57	11.71	−0.21	0.00008578
chr29†*‡	MOB2	50986147	50986148	−	13.86	11.29	−0.26	0.0003578

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CpG island.

*DNA methylation was exclusively contained in the gene body region of the gene.

‡DNA methylation was located within an intron region.

§DNA methylation was located within an exon region.

Differential DNA Methylation in CHG Sites

Promoter regions

There were 10 differentially methylated CHG sites within promoter regions in PNS compared with Control bull calves (Table 1). A greater percentage of these CHG sites were hypomethylated compared to hypermethylated (Table 1).

The 3 significantly (P ≤ 0.05) hypermethylated CHG sites located within promoter regions in PNS compared with Control calves are listed in Table 8. Among these was a CHG site within the promoter region of the Crystallin Beta B3 (CRYBB3) gene. The CRYBB3 gene is a member of the crystallin gene family, dysregulations of which have been associated with neural disorders, such as Alzheimer’s disease and schizophrenia (Shinohara et al., 1993; Middleton et al., 2002). Crystallin family heat shock proteins likely have a neuroprotective function (Masilamoni et al., 2006; Ousman et al., 2007). Specifically, CRYBB3 was upregulated in 3 models of mice whose dams underwent immune activation during gestation (dams were administered human influenza virus, poly (I:C), or recombinant IL-6 during gestation; Garbett et al., 2012). Previous studies suggest differential methylation of the CRYBB3 gene might result in alterations in neural and behavioral characteristics of PNS calves.

Table 8.

Hypermethylated CHG sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1	COLQ	154285093	154285094	+	9.92	7.71	0.27	0.02224
chr12†	FAM155A	87425041	87425042	+	8.00	5.71	0.16	0.000134
chr17*	CRYBB3	67543284	67543285	+	12.43	12.86	0.17	0.02543

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region.

The 7 significant (P ≤ 0.05) hypomethylated CHG sites located within promoter regions in PNS compared with Control calves are listed in Table 9. Among these was a CHG site within the promoter region of the Coagulation Factor II Receptor-Like 3 (F2RL3) gene. The F2RL3 gene is hypomethylated in adult smokers and has been associated with mortality risk due to cardiovascular diseases (Breitling et al., 2012).

Table 9.

Hypomethylated CHG sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr4†	AKR1B1	99033535	99033536	+	5.50	9.00	−0.1	0.00876
chr7*	F2RL3	6100112	6100113	−	9.13	13.14	−0.13	0.04596
chr17	HIC2	74182640	74182641	+	7.15	8.00	−0.11	0.01282
chr18	CYP2S1	50656104	50656105	+	9.63	7.43	−0.12	0.04467
chr21	FURIN	22213152	22213153	−	13.28	14.71	−0.11	0.04053
chr22	QARS	51484825	51484826	−	9.30	8.86	−0.13	0.01024
chr25†*	SBDS	28636530	28636531	+	9.39	10.00	−0.11	0.01421

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region

Gene body regions

There were 70 differentially methylated CHG sites within gene body regions in PNS compared with Control bull calves (Table 1). A slightly greater percentage of these CHG sites were hypermethylated compared to hypomethylated (Table 1).

The top (lowest P values) 30 of 37 significantly (P ≤ 0.05) hypermethylated CHG sites located within gene body regions in PNS compared with Control calves are listed in Table 10. Among these was a CHG site within the gene body region of the Protein Kinase C Alpha (PRKCA) gene. The PRKCA gene encodes for Protein Kinase C Alpha, a member of the serine- and threonine-specific protein kinases, which phosphorylates peptide substrates and is involved in various cell-signaling processes (Lee et al., 2017). The PRKCA gene has been associated with PTSD and memory in humans (De Quervain et al., 2012). Furthermore, rats whose dams were stressed by placing them on an elevated platform made of Plexiglass twice each day for a 10-min period between 12 and 16 d of gestation resulted in genome-wide alterations in gene expression, including expression of PRKCA (Mychasiuk et al., 2011). Due to its influence on various cell-signaling pathways, differential methylation of PRKCA has the potential to influence many biological processes in PNS calves.

Table 10.

Top (lowest P values) 30 hypermethylated CHG sites located within gene body regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1‡	COLQ	154285093	154285094	+	11.29	7.71	0.27	0.02224
chr2*‡	CAPZB	133916622	133916623	+	10.57	8.57	0.1	0.03793
chr4*§	ADCY1	76872462	76872463	−	9.43	6.43	0.11	0.01105
chr4*‡	LRGUK	98721906	98721907	+	29.29	29.43	0.16	0.007548
chr4*‡	LRGUK	98750277	98750278	+	11.57	14.57	0.17	0.03429
chr4*‡	SSPO	113439057	113439058	+	8.29	5.00	0.11	0.04102
chr5*‡	CNTN1	40172096	40172097	−	5.43	5.57	0.11	0.01133
chr5*‡	BTBD11	71077326	71077327	+	11.14	13.86	0.13	0.01252
chr5*‡	BTBD11	71077307	71077308	+	12.71	15.29	0.12	0.01786
chr5*‡	BTBD11	71077336	71077337	+	9.71	11.86	0.14	0.0197
chr5*‡	BTBD11	71077331	71077332	+	10.57	13.43	0.14	0.0197
chr5*‡	GTSE1	117731307	117731308	+	14.14	13.71	0.12	0.02663
chr10*‡	SLC8A3	82134282	82134283	+	8.14	6.57	0.11	0.008943
chr11*‡	ASS1	100828396	100828397	−	8.14	9.14	0.37	0.01106
chr11†*§	GBGT1	103182822	103182823	+	7.14	6.71	0.45	0.003044
chr12†§	FAM155A	87425041	87425042	+	9.14	5.71	0.16	0.000134
chr13*‡	PLCB1	1664698	1664699	+	9.14	7.71	0.39	0.02851
chr14*‡	TRAPPC9	4363947	4363948	+	9.86	7.57	0.26	0.03318
chr15*‡	GDPD5	55422374	55422375	+	9.29	7.00	0.2	0.002641
chr16*‡	RGS7	36590606	36590607	+	16.86	22.71	0.11	0.03004
chr16*‡	RGS7	36590615	36590616	+	10.43	11.86	0.1	0.03718
chr18*§	MYADM	62023403	62023404	−	11.00	6.14	0.18	0.04267
chr19*‡	PRKCA	63490629	63490630	+	9.57	10.71	0.1	0.02086
chr19*‡	TRPV2	33822975	33822976	−	8.57	5.86	0.12	0.01706
chr22*§	RNF123	51074315	51074316	−	9.14	6.00	0.15	0.01264
chr22*§	RNF123	51074297	51074298	−	9.14	6.00	0.12	0.03985
chr23*‡	EFHC1	24617468	24617469	+	9.71	7.71	0.11	0.01582
chr23*‡	GMDS	51389933	51389934	+	10.14	12.29	0.14	0.0292
chr25*§	INTS1	41995171	41995172	+	8.86	6.00	0.12	0.01351
chr27†*‡	CSGALNACT1	37970135	37970136	+	10.86	8.29	0.42	0.01612

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CpG island.

^*DNA methylation was exclusively contained in the gene body region of the gene.

^‡DNA methylation was located within an intron region.

^§DNA methylation was located within an exon region.

The top (lowest P values) 30 of 33 significantly (P ≤ 0.05) hypomethylated CHG sites located within gene body regions in PNS compared with Control calves are listed in Table 11. Among these was a CHG site within the gene body region of the Diacylglycerol Acyltransferase 1 (DGAT1) gene. The DGAT1 gene encodes for a protein enzyme involved in the conversion of diacylglycerol and fatty acyl CoA to triacylglycerol diacylglycerol and has been associated with metabolic diseases (Yen et al., 2008) and with milk production traits in dairy cattle (Mohammed et al., 2015). Mice whose dams underwent daily restraint stress for a 3-h period from 8 d of gestation to birth had increased accumulation of lipids in the liver and increased DGAT1 gene expression (Maeyama et al., 2015). Differential methylation of DGAT1 suggests an influence of prenatal stress on postnatal metabolic processes.

Table 11.

Top (lowest P values) 30 hypomethylated CHG sites located within gene body regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1*‡	OXNAD1	155073206	155073207	−	13.71	14.29	−0.34	0.01431
chr3*‡	TRAF3IP1	118294302	118294303	−	7.57	6.57	−0.35	0.02739
chr4†§	AKR1B1	99033535	99033536	+	6.29	9.00	−0.1	0.00876
chr6*§	PDGFRA	71409734	71409735	+	12.57	8.29	−0.1	0.00002454
chr7*‡	NMRK2	21292863	21292864	−	9.14	7.71	−0.35	0.001535
chr11*§	IFITM5	107203536	107203537	+	10.00	8.57	−0.17	0.03473
chr11*‡	FAM102A	98633234	98633235	+	12.29	9.57	−0.22	0.01901
chr11*‡	GTF3C5	103098035	103098036	−	8.43	7.29	−0.11	0.03774
chr11*‡	POMT1	101661140	101661141	+	8.71	7.71	−0.45	0.005054
chr13*‡	CTNNBL1	67382273	67382274	+	11.43	9.14	−0.11	0.01673
chr14*‡	DGAT1	1799370	1799371	+	5.86	6.57	−0.11	0.01688
chr14*‡	DPYS	62412347	62412348	+	10.57	11.43	−0.12	0.01354
chr16†*§	CAMK1G	75614071	75614072	+	7.86	7.43	−0.39	0.006132
chr17*‡	ISCU, LOC533308	66689455	66689456	+	7.57	8.14	−0.11	0.01066
chr17‡	PRODH	74182640	74182641	+	8.14	8.00	−0.11	0.01282
chr18*§	GAN	8018228	8018229	+	7.43	8.14	−0.1	0.008568
chr18†*‡	DDX19A	1774349	1774350	+	9.29	7.14	−0.35	0.03018
chr18*‡	VAT1L	5066203	5066204	−	9.29	7.43	−0.27	0.03996
chr19*‡	BAIAP2	52216891	52216892	−	6.14	6.00	−0.1	0.01496
chr19*‡	EXOC7	56206159	56206160	+	7.29	5.71	−0.11	0.02192
chr19*‡	SAMD14	37166191	37166192	+	6.43	5.86	−0.1	0.01496
chr21‡	FURIN	22213152	22213153	−	15.14	14.71	−0.11	0.04053
chr21*‡	LOC524810	71546924	71546925	−	33.14	18.57	−0.12	0.02952
chr21*‡	OTUD7A	30644587	30644588	+	12.71	10.00	−0.43	0.02269
chr22*‡	PTPRG	39397678	39397679	−	10.00	9.00	−0.31	0.0106
chr22‡	QARS	51484825	51484826	−	10.57	8.86	−0.13	0.01024
chr27†*‡	MTMR7	18983825	18983826	−	7.29	9.86	−0.3	0.03781
chr27*‡	RNF122	28663566	28663567	−	15.86	18.00	−0.24	0.007464
chr29*§	TSSC4	49837555	49837556	+	11.14	10.14	−0.15	0.003493
chr29*‡	FAT3	2565099	2565100	+	16.29	11.86	−0.41	0.01684

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CpG island.

*DNA methylation was exclusively contained in the gene body region of the gene.

‡DNA methylation was located within an intron region.

§DNA methylation was located within an exon region.

Differential DNA Methylation in CHH Sites

Promoter regions

There were 14 differentially methylated CHH sites within promoter regions in PNS compared with Control bull calves (Table 1). A greater percentage of these CHH sites were hypomethylated compared to hypermethylated (Table 1).

The 6 significantly (P ≤ 0.05) hypermethylated CHH sites located within promoter regions in PNS compared with Control calves are listed in Table 12. Among these was a CHH site within the promoter region of the Immediate Early Response 2 (IER2) gene. The IER2 gene encodes the Immediate Early Response 2 gene and is involved in neural development. Prenatal exposure to arsenic has been associated with altered IER2 gene expression, with upregulation of IER2 gene expression potentially serving as a biomarker of prenatal arsenic exposure (Fry et al., 2007).

Table 12.

Hypermethylated CHH sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1*	C1H21orf91	18719042	18719043	−	7.00	7.43	0.13	0.04541
chr1*	C1H21orf91	18719029	18719030	−	7.00	7.43	0.1	0.01395
chr4†*	ZYX	107597956	107597957	−	5.13	6.57	0.19	0.01628
chr5	LOC511240	72050111	72050112	+	11.00	9.86	0.22	0.03802
chr5*	POLR3B	70062294	70062295	−	15.50	14.14	0.11	0.03267
chr7†*	IER2	13545518	13545519	−	7.50	6.57	0.1	0.02712

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region.

The 8 significantly (P ≤ 0.05) hypomethylated CHH sites located within promoter regions in PNS compared with Control calves are listed in Table 13. Among these was a CHH site within the promoter region of the Interferon Induced Transmembrane Protein 1/Interferon Induced Transmembrane Protein 2 (IFITM1/IFITM2) genes. Patients with schizophrenia had increased IFITM1 and IFITM2 gene expression, which was likely indicative of an early environmental insult (Arion et al., 2007; Hwang et al., 2013).

Table 13.

Hypomethylated CHH sites located within promoter regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr4	CPA5	94878884	94878885	+	8.54	11.29	−0.14	0.03917
chr5†*	SRGAP1	50119329	50119330	−	9.28	11.29	−0.12	0.0108
chr11	IFITM2, IFITM1	107192312	107192313	+	12.05	10.00	−0.22	0.006774
chr11	IFITM2, IFITM1	107192386	107192387	+	12.93	11.00	−0.19	0.01444
chr13*	ZSWIM1	75359296	75359297	−	7.73	8.57	−0.2	0.003527
chr15*	MAML2	14155574	14155575	+	6.77	7.14	−0.15	0.01866
chr15	C15H11orf74	67844693	67844694	+	10.75	13.43	−0.1	0.0426
chr18	POP4	40371128	40371129	−	10.53	14.29	−0.11	0.007739

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CPG island.

^*DNA methylation was exclusively located within the promoter region.

Gene body regions

There were 133 differentially methylated CHH sites within gene body regions in PNS compared with Control bull calves (Table 1). A slightly greater percentage of these CHH sites were hypermethylated compared to hypomethylated (Table 1).

The top (lowest P values) 30 of 71 significantly (P ≤ 0.05) hypermethylated CHH sites located within gene body regions in PNS compared with Control calves are listed in Table 14. Among these was a CHH site within the gene body region of the Peroxisome Proliferator Activated Receptor Delta (PPARD) gene. The PPARD gene plays a key role in glucose and lipid metabolism (Brunmair et al., 2006). Male and female rats whose dams were administered dexamethasone between 13 d of gestation and birth exhibited hyperinsulinemia, altered glucose and fatty acid metabolism, and females (only) had increased PPARD gene expression in skeletal muscle (Wyrwoll et al., 2008). Holstein cows fed a moderate-energy (1.47 Mcal/kg) diet compared to Controls (1.24 Mcal/kg) during late gestation had lower PPARD expression compared to Controls after parturition (Osorio et al., 2013).

Table 14.

Top (lowest P values) 30 hypermethylated CHH sites located within gene body regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1*‡	COLQ	154250800	154250801	−	14.00	10.71	0.31	0.02325
chr1*‡	SH3BP5	154100975	154100976	−	6.43	5.43	0.55	0.006741
chr3*‡	ATG16L1	113620590	113620591	−	12.71	12.00	0.14	0.001883
chr5*‡	BTBD11	71077321	71077322	+	12.00	14.14	0.14	0.006485
chr5*‡	BTBD11	71076725	71076726	−	26.00	17.14	0.11	0.009977
chr5*‡	BTBD11	71077305	71077306	+	12.71	15.29	0.14	0.01081
chr5*‡	BTBD11	71077300	71077301	+	11.57	13.57	0.13	0.01306
chr5*‡	BTBD11	71077320	71077321	+	13.00	15.29	0.11	0.01355
chr5*‡	BTBD11	71077302	71077303	+	12.29	15.00	0.12	0.01688
chr5*‡	BTBD11	71077322	71077323	+	12.00	14.29	0.13	0.01713
chr5*‡	BTBD11	71077304	71077305	+	12.86	15.14	0.12	0.01713
chr5*‡	BTBD11	71077313	71077314	+	12.86	15.14	0.12	0.01786
chr6*‡	PDE5A	7031186	7031187	−	13.57	9.71	0.25	0.00821
chr8*‡	MOB3B	16811946	16811947	+	9.57	11.14	0.35	0.01953
chr13*‡	PKIG	73718134	73718135	+	6.57	7.57	0.13	0.002584
chr13*‡	PKIG	73718143	73718144	+	6.57	7.57	0.13	0.003265
chr13*‡	PKIG	73718119	73718120	+	6.43	7.43	0.11	0.01428
chr19*§	SEZ6	20882874	20882875	+	11.00	6.86	0.41	0.02207
chr19*‡	NXN	22530169	22530170	−	10.29	9.86	0.39	0.005695
chr20*‡	DAP	62637623	62637624	+	15.71	12.14	0.14	0.02114
chr20*‡	SKIV2L2	23806050	23806051	−	14.00	15.00	0.14	0.00286
chr21*‡	FAN1	27938912	27938913	−	10.71	9.29	0.29	0.01662
chr23*‡	PHACTR1	43337982	43337983	−	10.86	11.00	0.23	0.02088
chr23*‡	PPARD	9321494	9321495	+	7.86	8.57	0.11	0.01898
chr23*‡	PRIM2	2729219	2729220	−	5.86	9.71	0.15	0.02263
chr25*§	INTS1	41995205	41995206	+	8.86	6.00	0.12	0.01351
chr25†*‡	LFNG	41308254	41308255	−	13.29	12.00	0.17	0.01471
chr26*‡	RBM20	31655067	31655068	−	11.71	8.43	0.12	0.008981
chr27*‡	RNF122	28624366	28624367	+	29.00	14.86	0.11	0.01688
chr29*‡	RCOR2	42870766	42870767	+	11.00	9.57	0.1	0.01591

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CpG island.

*DNA methylation was exclusively contained in the gene body region of the gene.

‡DNA methylation was located within an intron region.

§DNA methylation was located within an exon region.

The top (lowest P values) 30 of 62 significantly (P ≤ 0.05) hypomethylated CHH sites located within gene body regions in PNS compared with Control calves are listed in Table 15. Among these was a CHH site within the gene body region of the Dihydropyrimidinase-like 2 (DPYSL2) gene. The DPYSL2 gene is a collapsin response mediator protein that is involved in neurodevelopment, neurotransmission, and neurodegenerative diseases (Charrier et al., 2003). Rats whose dams underwent gestational stressors (i.e., restraint stress, food deprivation, forced swimming, reversed light-dark cycles, and overcrowding stress during dark cycles) between 14 d of gestation and birth exhibited decreased DPYSL2 expression and potentially increased susceptibility to schizophrenic characteristics (Lee et al., 2015). Furthermore, rats whose dams were stressed by placing them on an elevated platform made of Plexiglass twice each day for a 10-min period between 12 and 16 d of gestation resulted in genome-wide alterations in gene expression, including expression of DPYSL2 (Mychasiuk et al., 2011).

Table 15.

Top (lowest P values) 30 hypomethylated CHH sites located within gene body regions of genes in prenatally stressed (PNS) compared with Control calves¹

Chrom	Gene	Start	End	Strand	Average Total CPG (Control)	Average Total CPG (PNS)	Methyl Diff	P value
chr1*‡	TBC1D5	156103383	156103384	+	12.71	11.43	−0.27	0.02232
chr2*‡	UBR4	134135665	134135666	+	9.29	7.86	−0.37	0.02694
chr3*‡	BARHL2	52695313	52695314	−	15.43	12.57	−0.15	0.01694
chr7*‡	MAP2K2	21141408	21141409	−	23.00	16.86	−0.29	0.02658
chr7*‡	PLVAP	5687715	5687716	−	12.43	8.00	−0.12	0.0146
chr7*‡	SLC12A2	27042120	27042121	+	9.00	10.57	−0.11	0.005449
chr8*‡	DPYSL2	75136546	75136547	+	7.14	8.43	−0.32	0.03065
chr8*‡	GSN	112604634	112604635	+	7.86	5.86	−0.11	0.02649
chr8*‡	TNFRSF10D	71053599	71053600	+	72.00	44.71	−0.21	0.03105
chr10*‡	LRRC16B	20910389	20910390	+	14.00	11.71	−0.12	0.01559
chr11*‡	FUBP3	100947908	100947909	+	11.43	13.71	−0.56	0.001667
chr11‡	IFITM2, IFITM1	107192312	107192313	+	13.71	10.00	−0.22	0.006774
chr11‡	IFITM2, IFITM1	107192386	107192387	+	14.71	11.00	−0.19	0.01444
chr11*‡	NACC2	103605839	103605840	−	20.71	12.29	−0.49	0.006619
chr11*‡	NEK6	95337824	95337825	+	12.86	9.57	−0.13	0.01493
chr12*‡	FARP1	79248800	79248801	−	10.29	11.43	−0.28	0.02768
chr13*‡	BCAS1	82210449	82210450	+	7.00	7.29	−0.24	0.03362
chr14*‡	ASAP1	11452197	11452198	+	12.86	9.71	−0.13	0.02803
chr14*‡	DPYS	62412342	62412343	+	10.43	11.43	−0.12	0.01296
chr14*‡	DPYS	62412355	62412356	+	10.43	11.29	−0.12	0.01296
chr14*‡	DPYS	62412345	62412346	+	10.57	11.43	−0.12	0.01354
chr15*‡	LOC509058	76786259	76786260	−	10.86	8.57	−0.43	0.01323
chr18‡§	POP4	40371128	40371129	−	12.00	14.29	−0.11	0.007739
chr18*‡	MTHFSD	12404197	12404198	−	10.14	9.29	−0.12	0.03374
chr18*‡	PEPD	44033725	44033726	−	6.86	7.29	−0.5	0.0007633
chr19*‡	RAB37	57344842	57344843	−	17.43	13.71	−0.11	0.01538
chr21*‡	AKAP6	43500188	43500189	+	14.86	12.00	−0.24	0.02095
chr21*‡	LOC524810	71577837	71577838	−	19.86	16.29	−0.17	0.02587
chr21*‡	PSTPIP1	32652323	32652324	−	6.86	6.86	−0.24	0.01042
chr27*‡	KCNU1	31979794	31979795	+	8.71	8.57	−0.27	0.03064

¹In limited cases, multiple genes were represented within one recorded region. In such cases, the record was considered one site.

^†DNA methylation was located within a CpG island.

*DNA methylation was exclusively contained in the gene body region of the gene.

‡DNA methylation was located within an intron region.

§DNA methylation was located within an exon region.

Canonical Pathways Altered in PNS Compared with Control Bull Calves

There were 113 canonical pathways altered (P ≤ 0.05) in PNS compared with Control bull calves. Those signaling pathways and the differentially methylated genes in each pathway are represented in Supplementary Table S1. Among the pathways altered in PNS bull calves were pathways related to behavior, stress response, immune function, metabolism, and cell signaling.

Pathways related to behavior, stress response, and neural function

Many genes and multiple canonical pathways related to behavior, stress response, and neural function were significantly altered in PNS compared with Control bull calves. Several of these pathways involved the hypothalamic-pituitary-adrenal (HPA) axis, neurotransmitter signaling, and opioid signaling. The “Corticotropin Releasing Hormone Signaling” pathway was activated in PNS bull calves (Supplementary Table S1). Other studies have reported differences in methylation of genes involved in HPA axis regulation, especially at the level of the glucocorticoid receptor gene, NR3C1 (Perroud et al., 2014). Although this study did not show differences in NR3C1 methylation, it did show differences at other levels of the HPA axis, including POMC methylation. “Dopamine-DARPP32 Feedback in cAMP Signaling and Dopamine Receptor Signaling” pathways were activated in PNS bull calves (Supplementary Table S1). Alterations in methylation of genes and canonical pathways related to behavior, stress response, and neural function agree with increased HPA axis activity and more excitable temperaments observed in calves in the larger population from which bull calves in this study were derived (Littlejohn et al., 2016). Previous reports suggest an influence of prenatal stress on dopamine regulation, especially at the level of COMT (Thompson et al., 2012) and dopamine receptors (Berger et al., 2002). The COMT gene and 2 dopamine receptor gene subtypes (DRD1 and DRD5) were differentially methylated in PNS bull calves. The “GABA Receptor Signaling” pathway was altered in PNS bull calves (Supplementary Table S1). Other studies have reported differences in genes involved in GABA regulation, development of GABAergic cells, and associated anxious behavior (Berger et al., 2002; Lussier and Stevens, 2016). The “Serotonin Receptor Signaling” pathway, with an emphasis on serotonin receptor subtypes, was altered in PNS bull calves (Supplementary Table S1). Prenatal stress has been associated with differences in serotonin receptor binding, serotonin synthesis, and associated behavioral alterations (Peters, 1986; Van den Hove et al., 2006). Richetto et al. (2017) reported “Neuronal Differentiation” to be the most enriched gene ontology term associated with cell differentiation in mice that were exposed to a prenatal viral challenge on gestational day 9 or 17. Significant subterms of “Neuronal Differentiation” included: “Gamma-Aminobutyric Acidergic Differentiation, Central Nervous System Differentiation, Noradrenergic System Differentiation, and Dopamine Differentiation.” Alterations to neurotransmitter pathways, such as dopamine, GABA, and serotonin have been associated with psychiatric disorders such as depression, anxiety, psychosis, and schizophrenia (Markham and Koenig, 2011). Furthermore, SNPs within the genes POMC, DRD2, DRD3, HTR2A, and SLC18A2 have been associated with temperament in cattle (Garza-Brenner et al., 2017). These genes are part of the HPA, dopamine, and serotonin signaling pathways, each of which were predicted to be altered due to prenatal stress in this study. Previous studies are consistent with alterations in predicted pathways related to behavior, stress response, and neural function in PNS bull calves. Immune cells and brain cells may exhibit similarities in differential methylation induced by prenatal or life experiences (Provençal et al., 2012; Tylee et al., 2013; Massart et al., 2016a; Seifuddin et al., 2017). For example, rhesus monkeys that were reared by an inanimate surrogate and age-matched peers rather than their biological dam exhibited significant overlap of differential methylation of DNA in T cells and cells from the prefrontal cortex (Provençal et al., 2012). Furthermore, 9 mo after a peripheral nerve injury (i.e., a chronic pain model) was induced in young rats, 72% of the promoters that were differentially methylated in T-cells were also differentially methylated in cells from the prefrontal cortex (Massart et al., 2016a). WBC have been reported to serve as an acceptable surrogate to reflect differential methylation in brain cells; however, differential methylation in WBC may only represent a portion of differential methylation in cells from neural tissues of interest.

Pathways related to immune function

Many genes and multiple canonical pathways related to immune function were significantly altered in PNS compared with Control bull calves. This might have been expected considering DNA methylation was assessed in WBC. Several of these key pathways included, “Leukocyte Extravasation Signaling, IL-15 Production, and IL-8 Signaling, Phagosome Formation, and B Cell Activating Factor Signaling” (Supplementary Table S1). Alterations in methylation of immune function related genes and canonical pathways in PNS bull calves were related to differences in cytokine concentrations and leukocyte counts in response to an endotoxin challenge (Littlejohn et al., 2018). Specifically, PNS bulls had a larger increase in IFN-γ from basal concentrations and a larger decrease in circulating monocyte counts suggesting increased extravasation of monocytes. These observations can be directly related to some of the altered pathways in this study. Alterations in immune function due to prenatal stress have been evidenced by hematology and cytokine alterations in primates, rodents, and swine (Reyes and Coe, 1997; Vanbesien-Mailliot et al., 2007; Couret et al., 2009). Richetto et al. (2017) reported the “Leukocyte Differentiation” process to be significantly altered in mice that were prenatally exposed to a viral challenge on gestational day 9 or 17. Cao-Lei et al. (2014) reported that children whose mothers were in the 1998 ice storm in Quebec during gestation had altered genome-wide DNA methylation in T cells at 13 yr of age. Six of the top 10 functions that were reported to be altered in those PNS children were related to immune function. Specifically, each of those functions was directly involved in T lymphocyte function (Cao-Lei et al., 2014). Previous studies as well as methylomic and phenotypic differences related to immune function in PNS bull calves suggest a potential influence of prenatal stress on overall health and immune response in bovine.

Pathways related to metabolism

Multiple genes and canonical pathways related to metabolic function were significantly altered in PNS compared with Control bull calves. Several of these pathways included, “Leptin Signaling in Obesity, Adipogenesis pathway, and Glycine Cleavage Complex” (Supplementary Table S1). Alterations in methylation of genes and canonical pathways that were related to metabolic function were linked to differences in metabolic function observed in a subset of the larger population of bulls from which bulls in this study were derived (d’Orey Branco et al., 2016). These alterations included a differential insulin response to a glucose challenge. Specifically, PNS bulls took less time to reach peak insulin response to glucose administration, decreased time to return to baseline, and a smaller area under the insulin response curve compared to Controls. This suggested PNS bulls to have an increased sensitivity to insulin (d’Orey Branco et al., 2016). Cao-Lei et al. (2014) reported the influence of the prenatal Quebec ice storm stressor on “Type 1 Diabetes Mellitus Signaling,” as predicted by differential DNA methylation of T cells at 13 yr of age. Furthermore, male mice whose dams were exposed to repeated exposure to an aggressive lactating female during late gestation had increased circulating triglyceride concentrations, decreased hepatic 5α-reductase, decreased Pparα mRNA expression, and decreased subcutaneous fat PEPCK mRNA expression (Brunton et al., 2013). Alterations to metabolism related pathways have implications for metabolic diseases, such as diabetes. However, alterations to metabolic processes in cattle might translate to differences in growth, gain, or feed efficiency, which could result in profitability differences in PNS cattle.

Pathways related to cell pluripotency and signaling

Many genes and canonical pathways related to cell pluripotency and cell signaling were significantly altered in PNS compared with Control bull calves.

Two pathways related to cell pluripotency that were predicted by IPA to be altered in PNS bulls were also predicted to be altered in rhesus monkeys that experienced early life maternal separation stress and a lack of maternal-rearing, including: “Human Embryonic Stem Cell Pluripotency and Role of NANOG in Mammalian Embryonic Stem Cell Pluripotency (Massart et al., 2016b).” Because DNA methylation is a primary regulator of cell-specific functions (Razin and Riggs, 1980; Suelves et al., 2016), predicted alterations to pluripotency of cells is logical and suggests developmental programming of various cell types.

Among altered cell signaling pathways was, “cAMP-mediated signaling, G-Protein Coupled Receptor Signaling, Gαs Signaling, Phospholipase C Signaling, Tec Kinase Signaling, Gαi Signaling, and TGF-β Signaling” (Supplementary Table S1). Cao-Lei et al. (2014) also reported an influence of the prenatal Quebec ice storm stressor on “Phospholipase C Signaling,” as predicted by differential DNA methylation of T cells at 13 yr of age. Massart et al. (2016b) reported that rhesus monkeys undergoing early life maternal separation stress and a lack of maternal-rearing had altered genome-wide DNA methylation in CD3+ T cells from day 14 to 2 y of age. From that study, 5 of the top 6 canonical pathways (assessed by IPA) altered in PNS monkeys were also significantly differentially methylated in PNS bull calves in this study. Two of those mutually altered pathways were related to cell signaling: “G-protein coupled receptor signaling and cAMP-mediated cell signaling.” Alterations to cell signaling pathways in previous studies and in PNS bull calves, suggests a potential influence of prenatal stress on many biological processes through cell signaling processes employed across many cell types.

Physiological Functions Altered in PNS Compared with Control Bull Calves

The top (most enriched) 10 “Physiological System Development and Function” terms generated by IPA software (Fig. 3) suggest a broad influence of prenatal stress on physiological systems. Six of the top 10 terms were related to developmental processes, which could occur during prenatal programming of physiological systems in utero. It is consistent with this laboratory’s previous phenotypic findings (Littlejohn et al., 2016) that “Behavior” and “Nervous System Development and Function” were within the top 10 most enriched terms (Fig. 3). These data suggest a broad influence of prenatal stress on prenatal development of physiological systems.

Figure 3.

Top (most enriched) 10 “Physiological System Development and Function” terms associated with alterations to methylation of DNA in prenatally stressed bull calves (generated by IPA software).

Genome-wide Overview of Differentially Methylated Regions

Overall, genome-wide distribution of differential DNA methylation (hypermethylation and hypomethylation) in PNS compared with Control bull calves were similar to previous reports in humans and nonhuman primates (Cao-Lei et al., 2014; Massart et al., 2016b). Heat maps in Figs. 4 and 5 compare the most significant (lowest P values) 100 methylation ratios (specific to each individual animal) that were hypermethylated and hypomethylated, respectively, in PNS compared with Control bull calves. Hierarchical cluster analysis was performed with regard to prenatal treatment. The dendrograms above and to the left of the heat maps represent this clustering.

Figure 4.

Comparison of the top 100 methylation ratios that were hypermethylated in prenatally stressed (PNS) compared with Control bull calves (1.0=Greatest degree of methylation; 0.0=Least degree of methylation).

Figure 5.

Comparison of the top 100 methylation ratios that were hypomethylated in prenatally stressed (PNS) compared with Control bull calves (1.0=Greatest degree of methylation; 0.0=Least degree of methylation).

Genome-wide chromosome distribution of differentially methylated CpG sites with regard to significance, -log₁₀(p-value), is represented in the form of a Manhattan plot in Fig. 6. Fourteen CpG sites surpassed the -log₁₀(p-value) threshold of 5. Of these sites, 1 was located exclusively within a promoter region (EIF3J), 1 was located within a promoter and gene body region (CLSTN2), 2 were located exclusively within a gene body region (TMEM200B and CPQ), and 10 were located within an intergenic region. These data reveal a diverse distribution of differentially methylated CpG sites in PNS calves, suggesting a substantial influence of prenatal environment on gene function. Genome-wide chromosomal distributions of DNA methylation ratios are represented in Fig. 7; these ratios were calculated from CpG sites within all region types with a minimum sequence read depth of 5 times (P ≤ 1.0). Although there was a greater percentage of significantly hypomethylated than hypermethylated CpG sites in PNS compared with Control bull calves (Table 1), the genome-wide illustration of all analyzed CpG sites in PNS and Control calves suggests the opposite relationship (Fig. 7). Overall, these data show genome-wide distribution of differential methylation across each chromosome in PNS compared with Control bull calves.

Figure 6.

Manhattan plot of –log₁₀(p-values) for all differentially methylated CpG sites across the genome (all region types).

Figure 7.

Comparison of genome-wide prenatally stressed (PNS; red line) and Control (blue line) CpG site methylation ratios (all region types).

Methylation of DNA acts to control gene activity and is a primary regulator of cell-specific functions (Razin and Riggs, 1980). Methylation status of DNA can be influenced by external stimuli, resulting in altered gene expression and phenotype (Feinberg, 2010; Szyf, 2012). The epigenome is most sensitive to change during embryogenesis and perinatal development (Reik, 2007). Early life development affects many biological mechanisms, which shape phenotype in beef cattle (Alford et al., 2007; Brickell et al., 2009). Therefore, the objective of this study was evaluation of the influence of a prenatal transportation stressor on the postnatal epigenome, and how PNS-induced alterations to the epigenome might affect biological systems, and thereby economically relevant traits in cattle. In the current study, methylation was assessed in DNA of WBC from PNS and Control bull calves. Prenatal and early life stressors have been reported to alter DNA methylation in various leukocyte types (Provençal et al., 2012; Cao-Lei et al., 2014). It is important to acknowledge that differential methylation has been reported among leukocyte types (Adalsteinsson et al., 2012), which raises concern with assessing a fundamentally heterogeneous population of WBC. However, Heiss and Brenner (2017) reported relatively low variation in DNA methylation among leukocyte types in humans, suggesting WBC could be an acceptable population of cells for DNA methylation analysis. Changes in methylation of DNA in WBC induced by prenatal or life experiences have been reported to be correlated with changes in methylation of DNA in various brain tissues (Provençal et al., 2012, Tylee et al., 2013; Massart et al., 2016a; Seifuddin et al., 2017); therefore, methylation of DNA from WBC can provide insight to methylation in other tissues such as neural tissue. Because differential methylation in WBC may only represent a portion of differential methylation in cells from neural tissues of interest, greater insight might be achieved by assessing tissue-specific methylation. It is also important to understand the influence of changes in DNA methylation on the transcriptome and resultant phenotype. Therefore, future studies should assess tissue-specific differences in methylation of DNA and its association with differences in transcription and phenotype due to prenatal stress.

CONCLUSIONS

To our knowledge, these data are the first reports of a genome-wide assessment of DNA methylation in PNS calves. Overall, these data exhibited similarities with data from models of prenatal stress in humans and nonhuman primates (Cao-Lei et al., 2014; Massart et al., 2016b). Prenatal transportation stress in cattle altered genome-wide DNA methylation profiles, which were predicted by IPA to alter canonical pathways related to behavior, stress response, neural function, immune function, metabolism, cell signaling, and other biological processes. Alterations in behavior, stress response, metabolism, and immune function are related with phenotypic differences observed in the population of calves from which bull calves in this study were derived (d’Orey Branco, et al., 2016; Littlejohn et al., 2016; 2018). The data presented herein demonstrate alterations of the methylome in PNS calves. Future evaluation of the inter-relationships of the methylome and transcriptome in specific tissues with phenotype of PNS calves will increase the understanding of the impact of prenatal stress on economically and biologically relevant phenotypic traits in a bovine model.

SUPPLEMENTARY DATA

Supplementary data are available at Journal of Animal Science online.