Background
The H19/IGF2 imprinted loci have attracted recent attention because of their role in cellular differentiation and proliferation, heritable gene regulation, and in utero or early postnatal growth and development. Expression from the imprinted H19/IGF2 locus involves a complex interplay of 3 means of epigenetic regulation: proper establishment of DNA methylation, promoter occupancy of CTCF, and expression of microRNA-675. We have demonstrated previously in a multigenerational rat model of intrauterine growth restriction the epigenetic heritability of adult metabolic syndrome in a F2 generation. We have further demonstrated abrogation of the F2 adult metabolic syndrome phenotype with essential nutrient supplementation of intermediates along the 1-carbon pathway and shown that alterations in the metabolome precede the adult onset of metabolic syndrome. The upstream molecular and epigenomic mediators underlying these observations, however, have yet to be elucidated fully.
Objective
In the current study, we sought to characterize the impact of the intrauterine growth−restricted lineage and essential nutrient supplementation on both levels and molecular mediators of H19 and IGF2 gene expression in the F2 generation.
Study Design
F2 intrauterine growth−restricted and sham lineages were obtained by exposing P1 (grandmaternal) pregnant dams to bilateral uterine artery ligation or sham surgery at gestational day 19.5. F1 pups were allocated to the essential nutrient supplemented or control diet at postnatal day 21, and bred at 6−7 weeks of age. Hepatic tissues from the resultant F2 offspring at birth and at weaning (day 21) were obtained. Bisulfite modification and sequencing was employed for methylation analysis. H19 and IGF2 expression was measured by quantitative polymerase chain reaction. Promoter occupancy was quantified by the use of chromatin immunoprecipitation, or ChIP, against CTCF insulator proteins.
Results
Growth-restricted F2 on control diet demonstrated significant down-regulation in H19 expression compared with sham lineage (0.7831 vs 1.287; P < .05); however, essential nutrient supplementation diet abrogates this difference (4.995 vs 5.100; P > .05). Conversely, Igf2 was up-regulated by essential nutrient supplemented diet on the sham lineage (2.0 fold, P = .01), an effect that was not observed in the growth restricted offspring. A significant differential methylation was observed in the promoter region of region H19 among the intrauterine growth−restricted lineage (18% vs 25%; P < .05) on a control diet, whereas the essential nutrient supplemented diet was alternately associated with hypermethylation in both lineages (sham: 50%; intrauterine growth restriction: 84%, P < .05). Consistent with essential nutrient supplementation impacting the epigenome, a decrease of CTCF promoter occupancy was observed in CTCF4 of the growth restricted lineage (2.45% vs 0.56%; P < .05) on the control diet, an effect that was repressed with essential nutrient supplementation.
Conclusion
Heritable growth restriction is associated with changes in H19 gene expression; these changes are reversible with diet supplementation to favorably impact adult metabolic syndrome.
The in utero environment is known to play a major role in the long-term health of the offspring. According to the hypothesis from the Developmental Origins of Health and Disease, an adverse in utero environment is associated with fetal programming, making the individual susceptible later in life to the onset of metabolic syndrome (MetS). This fetal programming is associated with epigenomic alterations and has been demonstrated to occur across multiple generations. Therefore, the identification of effective interventions during gestation to stop the cycle of the adult onset of disease is essential to ameliorate not only the health of the individual but that of future generations.
Intrauterine growth restriction (IUGR) resulting from uteroplacental insufficiency is an example of such an adverse in utero environment, where the fetus is subjected to hypoxia, acidosis, and substrate deprivation. We and others have shown that these individuals are at an increased risk of MetS in adulthood. Using our established model of uteroplacental insufficiency-induced fetal growth restriction, we have shown that the growth-restricted phenotype is multigenerational. In this model, late-gestation bilateral uterine artery ligation (or a control “sham” surgery) is performed on grandparental (P1) pregnant Sprague−Dawley rats at embryonic day 19 (e19), and the F1 pups are delivered at e21. The F1 generation exposed to the uterine artery ligation is born growth restricted compared with the offspring from the sham surgery. The F1 generation was allocated onto either a control diet, or an essential nutrient−supplemented (ENS) diet at weaning (postnatal day 21, D21). The ENS diet is enriched with components of the 1-carbon metabolic pathway. These F1 pups bred spontaneously to yield the F2 generation. Of note, the IUGR lineage-F2 generation, born to mothers on the control diet, was growth restricted, even though no surgical intervention was performed on the F1 animals.
In this model, we have found previously that only at postnatal day 160 (D160) a sex-specific MetS phenotype was apparent, with the males exhibiting obesity, increased central fat mass accumulation, glucose intolerance, insulin resistance, and increased triglyceride, very-low-density lipoprotein, and fatty acids. No sex-specific differences were observed early in the F2 offspring early in life, at either birth (D0) or D21. This phenotype was only observed in the IUGR lineage animals with no ENS diet intervention. We also have found distinct serum metabolomes between the F2 D160 males exposed to either a control or ENS diet in utero.
Such a multigenerational phenotype of IUGR begs the question: Are there epigenetic changes involved in the propagation of this phenotype? We and others have shown that IUGR is associated with epigenetic alterations in many tissues, including liver. Methylation levels are sensitive to the availability of the nutrients in the 1-carbon metabolism pathway, including methionine, folic acid, and choline. We therefore hypothesize that supplementation of the maternal diet with components of the 1-carbon metabolism pathway during gestation would be associated with changes in gene-specific DNA methylation levels in the offspring.
Expression from the imprinted H19/insulin-like growth factor 2 ( IGF2 ) locus involves a complex interplay of 3 means of epigenetic regulation: proper establishment of DNA methylation, promoter occupancy of CTCF, and expression of microRNA-675 ( miR675 ). DNA methylation is necessary for the establishment of genomic imprinting at this locus, an epigenetic mechanism leading to parent-of-origin monoallelic expression. Of note, the expression of imprinted genes is dictated by the parent of origin and not the sex of the offspring. Specific DNA methylation patterns throughout the imprinting control region (ICR) of these imprinted genes are necessary for proper transcriptional regulation. Both H19 and IGF2 are examples of imprinted genes integral to fetal growth and development. The IGF2 gene is expressed from the paternal allele throughout development, promoting fetal and placental growth. Alterations in Igf2 also have been implicated in postnatal growth control and the susceptibility to obesity. H19 is a long noncoding RNA expressed in fetal life from the maternal allele and thereafter repressed in early neonatal life. Within the first exon of H19 lies miR-675 , which is expressed in the placenta and is involved in regulating placental growth. Within the H19 promoter lies a differentially methylated region whose deletion in a murine model has been shown to disrupt completely H19 and IGF2 expression from this locus. This promoter region also contains multiple binding elements for the CTCF transcription factor. CTCF is a highly conserved transcription factor that can act as either a transcriptional activator or repressor ( Figure 1 ). The function of CTCF varies by cell type and is regulated through an epigenetic mechanism.
We hypothesized that an adverse in utero environment would be associated with changes in transcription from the H19/IGF2 locus and that these changes could be ameliorated with supplementation of the maternal diet with components of the 1-carbon metabolic pathway. We used our transgenerational model of IUGR to determine the effects of growth restriction and ENS supplementation on this imprinted locus and to specifically investigate the 3 epigenetic regulators at this locus: DNA methylation, CTCF occupancy, and miR675 expression.
Methods
Heritable multigenerational model of IUGR
Our multigenerational model of fetal growth restriction in Sprague−Dawley rats has been described previously. To summarize, pregnant rats were randomly allocated on e19 to receive either a sham surgery or bilateral uterine artery ligation with resultant F1 IUGR and control offspring born by cesarean delivery on e21. Pups remained with their mothers after birth and throughout lactation. On D21, weaned F1 were allocated to control (Harlan Teklad 8640; Harlan Laboratories, Indianapolis, IN) or to ENS (Harlan Teklad8640+folic acid, choline, B12, betaine, L-methionine, L-arginine, zinc) diet. Mating pairs were established with F1 generation on D80. The resultant offspring (F2) were maintained on the allocated diet of their parent. All experiments were approved by The University of Utah and Baylor College of Medicine Animal Care Committee(s).
cDNA analysis
The cDNA expression level of H19 and IGF2 in rat hepatic tissue was analyzed. Total RNA was extracted from D0, D21, and D160 of the F2 generation. Approximately 30 mg of hepatic tissue was lysed with a chaotropic buffer, and RNA was extracted with the Machery-Nagel NucleoSpin kit (Bethlehem, PA) and reverse-transcribed by use of the Applied Biosystems High Capacity cDNA Reverse Transcription Kit (Foster City, CA), as described previously. Reverse-transcription polymerase chain reaction (RT-PCR) analyses were then performed using 10 μL of cDNA and 2 μM final concentration of forward and reverse primers in a fixed 30 cycle PCR. The primers used were as follows: H19 , 5′-GAACATTTCCAGGGGAGTCA-3′ (forward) and 5′-CAGACATGAGCTGGGTAGCA-3′ (reverse); and IGF2 , 5′-GGAAGTCGATGTTGGTGCTT-3′ (forward) and 5′-TTCACTGATGGTTGCTGGAC-3′ (reverse). Products were analyzed by agarose gel electrophoresis.
Real-time RT-PCR
RNA was extracted from rat liver with a TRIzol protocol (Invitrogen, Carlsbad, CA). H19 , IGF2 , and miR-675 mRNA abundance was analyzed via real-time RT-PCR via an Icycler Thermal cycle (Bio-Rad, Hercules, CA). Real-time quantitative RT-PCR analyses were performed as described previously with the use of 2 μL of cDNA and 2 μM final concentration of forward and reverse primers and TaqMan probes in a total reaction volume of 5 μL. We used the iQ5 Real-Time PCR Detection System from Bio-Rad. Relative quantification of each gene was calculated after normalization to GAPDH by use of the comparative threshold cycle method.
Bisulfite sequencing and analysis
DNA was isolated from hepatic tissue with the DNeasy Blood and Tissue kit from QIAGEN (Valencia, CA). The purified DNA was subjected to bisulfite modification with the EZ DNA methylation kit (Zymo Research, Orange, CA). Bisulfite-treated DNA was PCR amplified with the use of primers specific for 3 distinct regions of the H19 proximal ICR (primers used are listed in Supplemental Table 1 ). The amplified PCR products subsequently were cloned with the Stratagenes Strataclone system (San Diego, CA), and a minimum of 11 samples from each reaction were sequenced by Lone Star Labs (Houston, TX) and aligned by the use of CLUSTAL W (Kyoto University, Kyoto, Japan). A representation of each individual CpG site, was created using ABI Methyl Prime software ( Figure 4 , A).
Chromatin immunoprecipitation (ChIP)
ChIP assay was performed as described previously with the following modifications. Approximately 100 mg of frozen D21 rat hepatic tissue from each animal was fixed at room temperature in 1% formaldehyde for 15 minutes. The reaction was quenched with 125 mM glycine. Cells were lysed with a Dounce homogenizer (Thermo Fisher Scientific, Waltham, MA). Chromatin was sonicated with the Bioruptor sonicator at 4°C into 200−500 base pair fragments.
Approximately 100 μg of chromatin was used for each reaction. Ten percent of the reaction was set aside before the addition of antibody. This served as the input. A mock immunoprecipitation (IP) was performed for each sample with the use of immunoglobulin G as a control. Five microliters of CTCF (Millipore, Billerica, MA) antibody were used per IP and incubated overnight with rotation at 4°C. Chromatin was purified by addition of Dynal Magnetic Beads conjugated to Protein A (Illumina, San Diego, CA). Beads were pulled down with a magnet, washed repeatedly, and resuspended in elution buffer. Crosslinks were reversed by incubation at 65°C for 5 hours. DNA was purified using QIAquick columns (QIAGEN).
Enrichment of promoter occupancy by virtue of histone modifications was determined by the use of quantitative PCR (primer sequences for the nonspecifc site and the 4 CTCF binding sites can be found in Supplemental Table 1 ). PCR was performed with SYBR Green (Applied Biosystems) in the iCycler (Bio-Rad). To determine percent IP, samples were first normalized to the input for each sample (ΔCt). Percent IP was calculated as 2 -ΔCt .
Statistical analysis
Univariate comparisons were performed with the χ 2 or Fisher exact test for discrete variables and the Wilcoxon test for the continuous variables. Real-Time PCR was analyzed with the 2 -ΔΔCT method as described previously. Results are displayed as fold change of growth-restricted lineage on ENS vs control diet compared with sham lineage. An independent-sample t test was performed for each gene. An independent-sample 1-tailed t test was performed to analyze the differential methylation between groups. An independent-sample 2-tailed t test was performed for the ChIP analysis in which we compared ENS with control diet. Prism software (GraphPad, v6.0.1) was used for analyses and a nominal P -value <.05 was considered significant.
Results
H19 expression is restricted to fetal and early neonatal life
To analyze the developmental patterns of H19 expression, full-length transcripts were analyzed by RT-PCR. At birth (D0), H19 was expressed ubiquitously in both sham and IUGR lineage F2 offspring regardless of gestational diet (demonstrated by equal band intensity across all groups; Figure 2 , A). At the time of weaning (D21), however, we observed an unanticipated silencing of H19 gene expression in IUGR lineage rats on control diet (absence of bands in far right lanes; Figure 2 , B). Of interest, H19 expression persisted in IUGR lineage F2 animals born to dams on ENS diet. By adulthood (D160), H19 was ubiquitously silenced, seen as absence of bands in all groups ( Figure 2 , C). Thus, fetal growth restriction is associated with premature loss of H19 expression among IUGR lineage F2 offspring on control diet whereas expression appropriately persists in offspring exposed to an ENS diet. These same ENS-exposed IUGR lineages will not develop adult MetS later in life.
Such differential developmental silencing is uniquely restricted to H19 , because IGF2 expression persists into adult life regardless of growth restricted lineage or diet, as evidenced by persistence of bands in D21 and D160 offspring ( Figure 2 , B). As a loading control, GAPDH was equivalently amplified in all samples ( Figure 2 , C).
Effect of ENS diet on quantitative H19 and IGF2 expression on F2 IUGR offspring
Hepatic expression of H19 and IGF2 were quantified via the use of real-time RT-PCR. Although the expression of H19 and IGF2 did not show a significant alteration in the IUGR lineage in the D0 offspring, expression of H19 was decreased significantly by virtue of the ENS diet in the sham lineage (0.8-fold, P < .002, Figure 3 , A). No effect was observed on the expression of IGF2 by virtue of lineage or diet at D0 ( Figure 3 , B). At D21, H19 expression was significantly decreased in the IUGR lineage on the control diet (0.4-fold, P < .05, Figure 3 , C) and significantly up-regulated in both the Sham (4.0-fold, P < .001) and IUGR lineage (6.6-fold, P < .001) animals on the ENS diet ( Figure 3 , C). At D21, IGF2 expression was up-regulated in the sham lineage on the ENS diet (2.0 fold, P = .01, Figure 3 , D), an effect that was not observed in the growth-restricted offspring. Therefore, maintaining the F1 generation on a diet rich in methyl donors is associated with relief of IUGR lineage H19 silencing in D21 F2 offspring and an up-regulation of IGF2 on sham lineage.
Hypermethylation of the ICR
Given the changes in H19 gene expression, we sought to investigate the mechanism underlying this differential regulation by characterizing the methylation pattern of the ICR. We analyzed 3 regions (regions I-III) that contain 4 distinct CTCF binding elements ( Figure 4 , A). In total, 43 CpG dinucleotide sites were analyzed within the designated regions: region I (−3001 to −3396), region II (−4278 to −4427), and region III (−4622 to −4894) ( Figure 4 , A).
Overall percent methylation was calculated for each group ( Supplemental Table 2 ). There was no significant difference in methylation by virtue of lineage or diet in region I of the ICR ( Figure 4 , B). Similarly, in region II, no significant difference in CpG site-specific methylation was observed ( Figure 4 , C). In contrast, we observed significant differential methylation in region III of the H19 ICR region by virtue of ENS supplementation. When compared with sham lineage on control diet, we observed a significant hypermethylation (18% vs 25% meCpG, P = .019). F2 offspring given an ENS diet, irrespective of lineage, also demonstrated significant hypermethylation in this region compared with sham lineage on control diet (18% vs 50% meCpG for sham ENS P = .002, 18% vs 84% meCpG IUGR ENS P < .001; Figure 4 , D).
Ideally, studies that aim to examine parent of origin effects will use allele-specific methylation marks to ascribe imprinting as either maternal or paternal; however, this is not possible among inbred rat strains such as those used in our current study. As an alternative, we elected to examine for parent of origin effects by delineating the IUGR phenotype early in life as arising from either the maternal or paternal lineage. Weight was measured at day 21 as a phenotypic measure of inherited changes after an IUGR lineage derived from either the maternal or paternal parent of origin ( Supplemental Figure ). For control-fed male rats of IUGR linage, there was a significantly lower weight for those with inherited IUGR through the paternal line ( P < .05) ( Supplemental Figure ), whereas for female rats the effect at day 21 was more pronounced, with significantly decreased expression from either maternal or paternal inheritance ( P < .05 and P < .001, respectively) ( Supplemental Figure ). The ENS-fed rats, however, appeared to be lower than non−ENS-fed rats at day 21, regardless of treatment. The epigenetic inheritance of the IUGR phenotype, although either maternal or paternal lineage in male and female rats in our model, echoes our previous findings and is consistent with extensive metabolomic changes discovered in F2 generation IUGR lineage rats in serum at day 21, regardless of F2 sex.
CTCF binding within the H19 ICR
Given the aforementioned findings and the known mechanisms of epigenomic inheritance at this locus, we sought to characterize other means of epigenomic regulation in the ICR, namely through CTCF occupancy within the H19 promoter. CTCF occupancy was quantified by the use of ChIP with an antibody against the CTCF protein. Quantitative PCR was used to determine CTCF enrichment within the 4 known binding sites within the H19 ICR. At CTCF binding site 2, occupancy was increased between the control diet IUGR animals and the ENS sham lineage ( Figure 5 , B). We observed a 77% decreased occupancy of CTCF at site 4 in growth-restricted offspring on control diet compared with sham (2.4 ± 0.6 vs 0.6 ± 0.3, P < .01, Figure 5 , D) as was expected because of the hypermethylation observed in region III ( Figure 4 , D) whereas the supplemented diet increased the CTCF binding, 3.43-fold ( Figure 5 , D), on the growth-restricted lineage compared with the control diet. No significant changes were observed in CTCF binding sites 1 and 3 ( Figures 5 , A and C).
