Fetal growth restriction alters transcription factor binding and epigenetic mechanisms of renal 11β-hydroxysteroid dehydrogenase type 2 in a sex-specific manner

Abstract

Intrauterine growth restriction (IUGR) increases the risk of serious adult morbidities such as hypertension. In an IUGR rat model of hypertension, we reported a persistent decrease in kidney 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) mRNA and protein levels from birth through postnatal (P) day 21. This enzyme deficiency can lead to hypertension by limiting renal glucocorticoid deactivation. In the present study, we hypothesized that IUGR affects renal 11β-HSD2 epigenetic determinants of chromatin structure and alters key transcription factor binding to the 11β-HSD2 promoter in association with persistent downregulation of its mRNA expression. To test this hypothesis, we performed bilateral uterine artery ligation on embryonic day 19.5 pregnant rats and harvested kidneys at day 0 (P0) and P21. Key transcription factors that can affect 11β-HSD2 expression include transcriptional enhancers specificity protein 1 (SP1) and NF-κB p65 and transcriptional repressors early growth response factor (Egr-1) and NF-κB p50. Our most important findings were as follows: 1) IUGR significantly decreased SP1 and NF-κB (p65) binding to the 11β-HSD2 promoter in males, while it increased Egr-1 binding in females and NF-κB (p50) binding in males; 2) IUGR increased CpG methylation status, as well as modified the pattern of methylation in several CpG sites of 11β-HSD2 promoter at P0 also in a sex-specific manner; and 3) IUGR decreased trimethylation of H3K36 in exon 5 of 11β-HSD2 at P0 and P21 in both genders. We conclude that IUGR is associated with altered transcriptional repressor/activator binding in connection with increased methylation in the 11β-HSD2 promoter region in a sex-specific manner, possibly leading to decreased transcriptional activity. Furthermore, IUGR decreased trimethylation of H3K36 of the 11β-HSD2 gene in both genders, which is associated with decreased transcriptional elongation. We speculate that alterations in transcription factor binding and chromatin structure play a role in in utero reprogramming.

uteroplacental insufficiency (UPI) is associated with many complications of pregnancy such as hypertensive disorders and preeclampsia and affects 3–10% of pregnancies in Western society (58). UPI causes an abnormal intrauterine environment, exposing the fetus to stressors such as hypoglycemia, hypoinsulinemia, acidosis, hypoxia, and increased circulating glucocorticoids (GCs) and results in intrauterine growth restriction (IUGR) (16). This particular environment in the human fetus has been associated with increased risk of developing serious long-term morbidities including hypertension (3, 8, 14).

Deficiency of the enzyme 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) constitutes a very important mechanism that can lead to hypertension. This enzyme regulates renal steroid sensitivity by metabolizing GCs to an inactive form in aldosterone target tissues such as the kidney (21, 48). 11β-HSD2-reduced activity causes hypertension in both humans and animal models through overactivation of the mineralocorticoid receptor (MR) by corticosterone, leading to renal sodium retention and a salt-sensitive increase in blood pressure (20, 29, 36).

Importantly, in a well-characterized animal model of IUGR and adult onset hypertension, we have previously reported persistently decreased kidney 11β-HSD2 mRNA and protein levels through day 21 of life (juvenile rat) (7). This occurs in association with increased circulating corticosterone levels both at birth as well as at day 21 of life (4, 6). However, the mechanisms that regulate kidney 11β-HSD2 expression and how they can be affected by IUGR are largely unknown and constitute the target of the present investigation.

11β-HSD2 activity can be compromised by at least two different known mechanisms. Mutations in the 11β-HSD2 gene cause a rare severe form of inherited hypertension (syndrome of apparent mineralocorticoid excess) in which cortisol activates the MR resulting in hypertension (57). Alternatively, 11β-HSD2 abundance and activity can be affected by corticosterone hormones, growth factors, shear stress, inflammatory cytokines, and hypoxia, all environmental factors that can be present in the altered intrauterine milieu following UPI (13, 15, 20, 26).

A significant characteristic of the IUGR kidney is the fact that the decrease in 11β-HSD2 mRNA and protein levels is persistent through day 21 of life in our animal model (7). Therefore, we now investigate whether a third mechanism, epigenetic phenomena, is associated with alterations in transcriptional regulation observed in the 11β-HSD2 gene. The role of epigenetics as a potential mechanism in the developmental programming of hypertension constitutes a novel area of investigation.

Epigenetics constitutes an important mechanism capable of regulating gene transcription over time, linking an early life event to adult morbidity. It entails heritable changes in chromatin that alter gene expression without altering the DNA sequence (28, 44, 46, 59). Prime examples of these changes include DNA methylation and covalent modifications of histones, such as acetylation (17, 43). In general, increased DNA methylation of normally unmethylated CpG islands correlates with transcriptional repression (29).

Furthermore, a mechanism by which increased promoter methylation affects gene expression is by modulating the binding of transcription factor complexes (9). In the case of 11β-HSD2, this gene promoter contains a GC-rich region in very close proximity to the transcription start site that is enriched in specificity protein 1 (SP1) binding sites (2, 60). Other key transcription factors that can interact in this GC-rich region and affect 11β-HSD2 expression include early growth response factor (Egr-1) and NF-κB (p50/p65), as previously shown by Kostadinova et al. (Fig. 1) (35). For instance, decreased binding activity of SP1 has been associated with decreased 11β-HSD2 expression (45). Similarly, hypoxia has been shown to downregulate the expression of 11β-HSD2 by induction of the early growth response gene Egr-1 in renal cell culture as well as in vivo experiments (26). Finally, the NF-κB inducible factor can act as a transcriptional activator or repressor upon stimulation and translocation to the cell nucleus. For example, the NF-κB subunit p65 (RelA) has been shown to bind to the κB1 site of the 11β-HSD2 promoter functioning as a transcriptional activator, whereas binding of subunit p50 has been associated with repression of 11β-HSD2 gene expression (35).

Fig. 1.

Schematic representation of the 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) gene depicting CpG sites along the promoter and exon 1, represented as vertical bars. Consensus sites for transcriptional factors are represented in dark grey [specificity protein 1 (Sp1)/early growth response factor (Egr-1)] and light grey (NF-κB) lines; transcription factors are represented in oval shapes. Numbers refer to the position relative to the transcription start in the kidney transcript (+1). Arrows represent sequence of primers designed for chromatin immunoprecipitation (ChIP) studies for forward 1 (F1), reverse 1 (R1), and forward 2 (F2), reverse 2 (R2) in the promoter region and forward 3 (F3), reverse 3 (R3) in exon 1 region [based on original work by Kostadinova et al. (35) investigating 11β-HSD2 genomic footprinting].

On the other hand, the amino terminal tail of the histone core proteins around which DNA wraps can be subject to posttranslational modifications including acetylation, methylation, and phosphorylation (31). The combination of histone amino-terminal modifications thus reveals a “histone code” that provides another level through which chromatin structure can regulate transcription factor contact points with DNA (55).

A large number of epidemiological and animal studies (4, 25) point to the fact that gender differences play a very important role in the development of cardiovascular and renal disease following UPI. Moreover, our group has previously observed gender-specific responses to IUGR in regards to epigenetic marks along the liver insulin growth factor 1 gene (IGF-1) that are affected even before the onset of insulin resistance and diabetes in our animal model (24).

Based on this background, we hypothesized that in the kidney, IUGR affects transcription factor binding, as well as modifies epigenetic determinants of 11β-HSD2 chromatin structure, in association with its persistent mRNA downregulation through day 21 of life. We further hypothesized that these alterations are gender specific.

To test these hypotheses, we induced IUGR using a well-established model of UPI in which bilateral uterine artery ligation is performed on day 19.5 of gestation in Sprague-Dawley rats (term of 21.5 days) (5, 39, 40, 42). UPI in this animal model results in offspring with low birth weight and asymmetrical IUGR (53), with a 25% reduction in glomeruli number and adult onset hypertension (7, 50).

We first evaluated the hypothesis that IUGR affects transcription factors SP1, Egr-1, and NF-κB (subunits p50 and p65) binding to the promoter region of 11β-HSD2 using chromatin immunoprecipitation (ChIP) in both control and IUGR rats. Second, to test the hypothesis that IUGR alters kidney 11β-HSD2 epigenetic characteristics, we assessed 1) DNA methylation of CpG sites in the promoter and exon 1 of 11β-HSD2, using Na-bisulfite sequencing; and 2) 11β-HSD2 DNA-histone modifications using ChIP. The markers we chose to analyze were those that we had previously shown to be affected in the IUGR rat (24).

We tested our hypotheses at day 0 (P0) and at day 21 of postnatal life (P21) in both genders to determine the consequences of IUGR on 11β-HSD2 epigenetic characteristics during the perinatal and postnatal period. This time frame includes the completion of nephrogenesis (P8) through juvenile stages. Moreover, in this particular animal model, we have observed that by P21 most gender-specific responses have become evident for several processes, while potentially confounding factors, such as overt hypertension, are not yet present.

METHODS

Animals.

All procedures were approved by the University of Utah Animal Care Committee and are in accordance with the American Physiological Society's Guiding Principles (1). These surgical methods have been previously described (50, 53). In brief, on day 19.5 of gestation, the maternal rats (Sprague-Dawley) were anesthetized with intraperitoneal xylazine (8 mg/kg) and ketamine (40 mg/kg), and both inferior uterine arteries were ligated (IUGR; n = 12 litters). Control animals received anesthesia (control; n = 12 litters). Rats recovered within a few hours and had ad libitum access to food and water. At term (21.5 days gestation), P0 pups were delivered by caesarian section, weighed, and decapitated (n = 6 litters IUGR and control, respectively). To minimize litter to litter variation, one male and one female pup from each litter was used for all P0 studies. In this rat model of assymetrical growth restriction, IUGR pups are 20–25% lighter than the control animals, and birth weights are normally distributed within and among litters. Litter size does not differ between IUGR and control groups (38). To study P21 rats, the remaining maternal rats were allowed to deliver spontaneously at term (n = 6 litters IUGR and control, respectively), and litters were randomly culled to 6 pups. Based on our previous studies where we observed that operated dams are able to provide normal lactation, we did not cross-foster to nonoperated female rats. Furthermore, our group has recently reported the breast milk content of operated dams compared with control nonoperated dams measured after dams were separated from P21 pups. Breast milk from dams that underwent the IUGR surgery did not significantly differ from control breast milk in terms of caloric, fat, protein, zinc, and sodium content (34).

At P21, one male and one female pup from each litter was randomly selected and separated from their dams, anesthetized, and decapitated. For P0 and P21, both male and female rats were included in the study in equal numbers (n = 6 male and 6 female animals for IUGR and control groups, respectively). Pup gender was determined by dissection and visualization. For all dates, kidneys were quickly harvested and frozen in liquid nitrogen.

Na-bisulfite sequencing.

Bisulfite modification was performed as described earlier (22) . We elected to analyze a total 17 CpG sites that were upstream 11β-HSD2 transcription initiation site in the promoter in 2 CpG-rich regions (Fig. 1). The 13 CpG sites in the first region (forward 1/reverse 1) were at −223, −219, −213, −210, −203, −197, −166,−157,−151, −134, −129, and −110. Bisulfite-treated genomic DNA was amplified with the following primers: CpG sites between −258 and −81: set 1 forward 5′-GTAGGGYGTAGGGTTGGGTAGATTG and reverse 5′- CTTTCCTCCACTTCTATCTCAACAC. The four CpG sites in the second region (forward 2/reverse 2) were at −65, −59, −54, and −43 on the genomic DNA. The primer sequences for CpG sites −105 to −12: set 2 forward 5′-GTGTTGAGATAGAAGTGGAGGAAAG and reverse 5′-CTCRAACCCAACTTATAAAACRCCCT. In exon 1, we analyzed 19 CpG sites (forward 3/reverse 3) including 235, 251, 257, 273, 286, 289, 299, 326, 328, 336, 343, 345, 354, 361, 372, 374, 376, 389, and 405 on the genomic DNA. The primer sequences for CpG sites +210 to within intron 1 (+427): set 3 forward 5′-GTCTGGGTYGTTYGTTGTTGGYGG and reverse 5′-CCTCTAAATATACRATCCCACCCTC. For each group, five animals were analyzed by bisulfite sequencing using CpGenome DNA modification kit (Chemicon International, Temecula, CA) following the manufacturer's instructions (11). In brief, genomic DNA was treated with sodium bisulfite to covert unmethylated cytosines to uracil, which then converts to thymine, leaving 5-methylcytosines unchanged. PCR conditions for the primers were 95°C for 10 min, followed by 94°C for 30 s, annealing at 57°C (sets 1 and 2) or 56°C (set 3) for 30 s, 72°C for 30 s, 35 cycles. The PCR products from bisulfite-treated genomic DNA were cloned into the vector pSC-A (Stratagene, Cedar Creek, TX). Eight colonies from each PCR cloning were inoculated into SeqPrep 96 plates (Edge BioSystems, Gaithersburg, MD). The plasmid DNA was prepared by using SeqPrep 96 plasmid prep kit (Edge BioSystems) and sequenced according to the manufacturer's instructions for double-stranded plasmid DNA using the BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems) with M13 reverse primer.

ChIP assay and real-time PCR.

A quantity of 100 mg kidney tissue was ground in liquid nitrogen and fixed with formaldehyde, with a final concentration of 1%, for 10 min. The chromatin was sonicated (Sonic Dismembrator, model 100; Fisher Scientific, Pittsburgh, PA) 10 times for 10 s on ice at the highest level to generate chromatin fragments of 500–2,000 bp. Sonicated chromatin was quantified on the basis of DNA content at A₂₆₀. The chromatin equivalent of 40 μg DNA based on the absorption at A₂₆₀ was used in each immunoprecipitation. The protocol described by Oberley and Farnham (47) was used with the following modifications. To measure transcription binding in the 11β-HSD2 promoter region, ChIP with SP-1 (Active Motif, North America, Carlsbad, CA), NF-kB (p50), NF-kB (p65), and Egr-1 (Santa Cruz Biotechnology, Santa Cruz, CA) was performed. For SP-1 (Active Motif, North America, Carlsbad, CA), antibody concentrations of 2.5 ug were used. For NF-kB (p50), NF-kB (p65), and Egr-1 (Santa Cruz Biotechnology), antibody concentrations of 2 ug were used. To assess histone covalent modifications within the 11β-HSD2 promoter, exon 2, and exon 5, ChIP with anti-acK9H3 (Cell Signaling Technologies, Beverly, MA), anti-acK14H3, anti-me2K4H3, anti-me3K4H3, anti-me3K9H3 (Millipore Upstate, Charlottesville, VA), or anti-me3K36H3 (Abcam, Cambridge, MA) was performed. For anti-acK14H3, 20 μg of antibody were used. For all of the others, the volume of antibody was equal to the volume of formaldehyde cross-linked chromatin. The DNeasy tissue kit (Qiagen, Valencia, CA) was used to purify the DNA from the total amount of DNA extracted from 40 μg of chromatin. DNA was quantitated by measuring A₂₆₀/A₂₈₀. DNA fragments containing 11β-HSD2 site-specific sequences, including promoter, exon 2, exon 5, and an intergenic region, were quantified by real-time PCR. Primer and probe sequences designed using Primer Express software (Applied Biosystems) are listed in Table 1. For the transcription factor ChIP experiments, total DNA input was used as a positive control, whereas the intergenic region at site 263.8 kb (accession no. BH351084) upstream of the IGF-1 gene was used as a negative control. For the six histone modifications studied, the same intergenic region was used for control as previously described by our group (24). Relative quantification of PCR products was based on value differences between the target and the intergenic control using the comparative C_t method (TaqMan Gold RT-PCR manual; PE Biosystems).

Table 1.

ChIP/real-time PCR primer and probe sequences

Transcript	Sequence
Promoter
Forward	5′-TGCGCAGGTGAGAGTGCA
Reverse	5′-CCTCACTCAGTGCCATGGC
Probe	5′-6FAM-ATTTGGGCGTGTGTACCGAGGACCT
Exon 2
Forward	5′-CTGGATGCCATGGGCTTC
Reverse	5′-AGCACCAGGGCCATTCAA
Probe	5′-6FAM-CGGTGCTGGCCACTGTGTTGGA
Exon 5
Forward	5′-AGCCTTTGAGAGGATGCTACAGA
Reverse	5′-GGCCCCACATGTCTAATCATTC
Probe	5′-6FAM-ACTTCTGTGCTGACTTT
Intergenic region
Forward	5′-AAGTGGCAACTCCATGACTCAA
Reverse	5′-GCCTGGTGTCACACCCAAA
Probe	5′-6FAM-ATGCCTTCCAGAGAGGTTTGGTACTGCC

ChIP, chromatin immunoprecipitation.

Statistics.

All data presented are expressed as means ± SE percentage of control. Real-time PCR was analyzed using ANOVA (Fisher's protected least significance difference) and Student's unpaired t-test as applicable. Student's 2-tailed t-test was used for analysis of DNA methylation. A value of P < 0.05 was considered statistically significant.

RESULTS

IUGR alters transcription factor binding to 11β-HSD2 at P0.

The highly GC-rich 11β-HSD2 promoter contains several binding sites for SP1, Egr-1, and NF-κB transcription factors. We evaluated the presence of these transcription factors in the promoter after UPI in vivo using ChIP at P0 (Fig. 2A) and P21 (Fig. 2B). IUGR was associated with significantly decreased SP1 binding to the 11β-HSD2 promoter as measured by ChIP at P0 in IUGR males (35 ± 8% of control; P < 0.05), with no difference in IUGR females (45 ± 19% of control; P < 0.08).

Fig. 2.

Uteroplacental insufficiency affects transcription binding to the promoter of kidney 11β-HSD2. ChIP analysis of transcription factors SP1, Egr-1, and theNF-κB subunits p65 and p50 binding to renal 11β-HSD2 promoter in intrauterine growth restriction (IUGR) and control rats at day 0 (P0; A) and postnatal (P) day 21 (B). Results are expressed as means ± SE percentage relative to controls (n = 6 litters; * P < 0.05; **P < 0.01; ***P < 0.001).

In contrast, IUGR was associated with increased Egr-1 binding to the 11β-HSD2 promoter in IUGR females (394 ± 19% of control; P < 0.001), with no significant difference observed in IUGR males (262 ± 25% of control; P < 0.08).

Finally, IUGR induced a decrease in NF-κB subunit p65 localization in the κB site together with a significant increase in p50 NF-κB subunit in the11β-HSD2 promoter in IUGR males at P0 (25.5 ± 8 and 233 ± 17% of control; P < 0.05, respectively). There was only a trend towards decreased p65 binding in the IUGR females at this stage (41 ± 15% of control; P < 0.07), with no change in NF-κB p50 subunit presence.

UPI alters transcription factor binding to 11β-HSD2 at P21.

At P21, IUGR was associated with persistently decreased SP1 binding to the 11β-HSD2 promoter as measured by ChIP in IUGR males (23 ± 9% of control; P < 0.01), with no difference observed in IUGR females (52 ± 14% of control; P < 0.09). In contrast, we found no alteration in Egr-1 binding to the 11β-HSD2 promoter following UPI at P21 (males: 110 ± 19% of control; females: 121 ± 15% of control). Similar to P0, UPI induced a significant increase in binding of NF-κB p50 subunit to the 11β-HSD2 promoter in IUGR males (154 ± 5%; P < 0.01). In contrast to P0, there was no difference in NF-κB p65 binding to the κB site of 11β-HSD2 in IUGR males and females at this age (Fig. 2B).

UPI induced increased renal 11β-HSD2 DNA methylation at P0.

We analyzed a total of 17 CpG sites within the promoter region (−12 to −258) and 19 CpG sites in exon 1 (+235 to +405) of the rat 11β-HSD2 gene (Figs. 1 and 3). These GC-rich regions contain several confirmed binding sites for SP1, Egr-1, and NF-κB transcription factors. To define methylation status of these sites, we performed Na-bisulfite sequencing.

Fig. 3.

Uteroplacental insufficiency alters CpG methylation status in the 11β-HSD2 promoter. Rat kidney 11β-HSD2 promoter CpG methylation analysis between −258 and −12 in IUGR and control rats. A: P0 males. B: P0 females. Graphs represent means ± SE percentage methylation; white bars are control values, and black bars are IUGR values. Consensus sites for transcription factor are represented above lines. Right: methylation patterns; each horizontal row of beads represent 1 of the 17 CpG sites, and each vertical line of beads represents a clone. ○, unmethylated CpG sites; ●, methylated CpG sites (n = 5; *P < 0.05; **P < 0.01).

In the present study, IUGR significantly increased the methylation status of the 11β-HSD2 promoter in males and females at P0, without affecting the methylation status of exon 1 (data not shown).

Importantly, the pattern of methylation was different between genders. At P0 in males, IUGR increased CpG methylation at the −65, −157, −210, and −223 CpG sites (Fig. 3A), whereas in females, IUGR increased CpG methylation at the −65, −110, −157, −203, and −213 CpG sites relative to controls (Fig. 3B). At P21, there was no significance difference in methylation status of the renal 11β-HSD2 promoter and exon 1 when comparing IUGR and control rats (data not shown).

UPI affects renal 11β-HSD2 histone code.

We analyzed three different sites along the 11β-HSD2 gene for six histone H3 covalent modifications in IUGR and control kidneys. These included acetylation of H3 at lysine 9 and 14 (acK9 and acK14), dimethylation of H3 at lysine 4 (me2K4), and trimethylation of H3 at lysines 4 (me3K4), 9 (me3K9), and 36 (me3K36).

The extent of each modification was quantified using ChIP/real-time PCR and expressed as IUGR percentage of controls (controls considered 100%) after normalization with the 263.8-kb intergenic region for both groups (Fig. 4).

Fig. 4.

IUGR affects histone modifications along the 11β-HSD2 gene. Six histone modifications at 3 sites (promoter, exon 2, and exon 5) were analyzed by ChIP and real-time PCR at P0 and P21 in IUGR and control rats. A: P0 males. B: P21 males. C: P0 females. D: P21 females. Graphs represent means ± SE percentage of control value (n = 6; *P < 0.05; **P < 0.01).

As with the DNA methylation studies, IUGR affected renal 11β-HSD2 patterns of H3 acetylation and methylation in a gender- and age-specific manner compared with controls. At P0, IUGR was associated with increased acK9 in the 11β-HSD2 promoter and exon 2 regions in IUGR males, whereas in females IUGR increased me3K4 in these same regions. In addition, in exon 2, IUGR significantly decreased me2K4 in males but was associated with increased me2K4 in females. Finally, both male and female IUGR animals exhibited decreased me3K36 following UPI in the exon 5 region at P0 (Fig. 4, A and C).

At P21, the pattern of the covalent modifications was intriguing in that we observed no significant changes in the IUGR male promoter and exon 2 regions following UPI. The only persistent change observed in males at P21 was decreased me3K36 in exon 5.

In contrast, in IUGR females, UPI significantly increased acK9 in the promoter and me3K4 in exon 2 regions, while it decreased acK14 in exon 2. Similar to P0, me3K36 remained lower in exon 5 at P21 in IUGR females (Fig. 4, B–D).

DISCUSSION

The most important and novel findings in the present study are that UPI affects transcription factor binding and modifies the epigenetic determinants of the chromatin structure of kidney 11β-HSD2 in IUGR rats. This occurs in the context of decreased kidney 11β-HSD2 mRNA levels at birth (P0) through day 21 of life in this animal model of IUGR offspring predisposed to developing hypertension (7). A secondary yet important finding in the present study is that the effect of UPI on 11β-HSD2 epigenetic transcriptional modulation is gender specific.

Regulation of 11β-HSD2 is important to the understanding of the molecular mechanisms underlying hypertension in the IUGR offspring. This is supported by several studies (20, 36) that suggest that renal 11β-HSD2 has a pivotal relevance for blood pressure control. Much of the research involving regulation of 11β-HSD2 in relation to programming of hypertension has been carried out in placental, renal, colonic, and breast carcinoma cell lines (26, 33, 41, 54). Studies of human and other primate placentas or derived trophoblast cells have shown that placental 11β-HSD2 activity is reduced among other factors by nitric oxide, progesterone, estrogen, prostaglandins, and inflammatory cytokines (54). For instance, among inflammatory cytokines, TNF-α has been shown to decrease 11β-HSD2 expression in renal cells by a MAPK-dependent mechanism (27). Alternatively, and as mentioned earlier in this manuscript, other researchers (13, 15, 20, 26) have shown that 11β-HSD2 abundance and activity can be affected by corticosterone hormones as well as growth factors, shear stress, and hypoxia. Notably, several of the above regulating factors can be present in the altered intrauterine milieu following UPI and could contribute to the downregulation of kidney 11β-HSD2 at birth. However, we have observed that kidney 11β-HSD2 mRNA levels are persistently downregulated beyond the perinatal period in IUGR rats, suggesting that in utero reprogramming could be taking place following UPI.

Epigenetics disregulation is an important molecular mechanism through which a prenatal insult such as IUGR can persistently affect postnatal phenotype. Mechanisms such as DNA CpG methylation and histone covalent modifications affect gene expression by controlling the three-dimensional structure of chromatin and thereby regulate transcription machinery and transcription factor access to DNA. Data from our laboratory demonstrate that epigenetic characteristics of chromatin structure are vulnerable to change in response to the intrauterine milieu associated with UPI and IUGR. For example, our laboratory (23, 42) has demonstrated that IUGR induces site-specific changes in histone H3 acetylation and affects DNA-histone H3 positioning in the rat liver. Furthermore, the changes in epigenetic characteristics are closely associated with alterations in gene expression in several tissues, including the kidney (23, 34, 42, 50). Among those genes, our laboratory (50) has observed that UPI is associated with altered CpG methylation within the p53 promoter in the IUGR kidney, in association with increased levels of p53 and apoptosis. The 11β-HSD2 promoter and exon 1 also contain highly GC-rich regions (>80% of the sequence) that have been shown to be epigenetically regulated, both in vitro and in vivo (2). Therefore, in the present study, we propose that epigenetic regulation is a good candidate mechanism for kidney 11β-HSD2 expression.

Our present findings of a relatively modest increase of 11β-HSD2 DNA methylation in IUGR kidneys are not surprising. Traditionally, it has been proposed that DNA methylation in the promoter region regulates gene expression in an “on-off” manner (32). Nonetheless, our group as well as other investigators has shown that modest increases in methylation could dampen gene expression at a tissue level, most likely secondary to alterations in nucleosome positioning and decreased access to key transcription factors (18, 19, 24). Indeed, as shown in the present study, UPI induced a significant decrease in the binding of the transcriptionally active SP1 and NF-κB (p65) to the GC-rich region of the 11β-HSD2 promoter in IUGR males. In contrast, UPI was associated with increased binding activity of the repressor Egr-1 in IUGR females and the inactive NF-κB (p50) dimer in IUGR males to this same region. However, the interplay of transcription factors is complex and dynamic, particularly, following pathologic conditions such as those encountered following UPI.

Interestingly, as previously demonstrated by genomic footprinting, the binding sites for SP1, Egr-1, and NF-κB (κB1) in the 11β-HSD2 promoter are in very close vicinity, allowing for direct protein-protein interaction between all three transcription factors (35). Several studies have described these interactions. For instance, Perkins et al. (49) showed that Sp1 and NF-κB work synergistically in the activation of human immunodeficiency virus enhancer. In contrast to SP1, Egr-1 was shown to repress NF-κB (p65 subunit) transcriptional activity through its zinc finger domain (10). During T-cell activation, however, Egr-1 has been shown to act synergistically with NF-κB (p65 subunit), mediating T-cell proliferation (12). In many promoters, including 11β-HSD2, SP1, and Egr-1 binding sites, overlap and stimulation of Egr-1 can prevent Sp1 from interacting with its recognition element (30, 35). In most instances, the switch in the transcription binding activity is triggered by changes in the local environment such as hypoxia, exposure to cytokines, and growth factors. Therefore, this transcriptional regulatory mechanism may have an important role in downregulation of kidney 11β-HSD2 following the altered UPI milieu, especially in the perinatal period.

Little is known about how sex influences kidney and subsequent gene expression in response to the IUGR insult. Therefore, the present findings of differential DNA methylation and transcription factor binding of 11β-HSD2 in males vs. females demonstrates that while the general IUGR phenotype is similar between sexes, many of the underlying molecular mechanisms differ. Importantly, we have previously observed comparable gender-specific responses to UPI in other tissues such as the IUGR brain and liver (24, 34)

Finally, following UPI, IUGR males and females also responded differently in terms of most histone modifications analyzed. For instance, UPI significantly decreased me2K4 in males but was associated with increased me2K4 in females in exon 2. Similarly, UPI increased acK9 in the 11β-HSD2 promoter and exon 2 regions in IUGR males, whereas in females it increased me3K4 in these same regions at birth. However, we did observe one persistent modification that was common to both genders: the decrease in me3K36 in exon 5 following UPI at birth through day 21 of life.

Traditional epigenetic concepts describe histone covalent modifications including acetylation and methylation. For instance, gene activation is often associated with histone H3 acetylation at K9 and K14, and methylation at K4 and K36. Trimethylation of H3K9 has traditionally been implicated in gene silencing, whereas dimethylation of H3K4 and trimethylation of H3K36 have been considered to be responsible for gene elongation (37, 52, 56). Nonetheless, it has become evident in recent years that these histone modifications do not affect transcription in an isolated fashion, but that modifications at one site can affect neighboring histone codes (31, 51). As a result, the interpretation of how each modification can affect expression is influenced by the entire set of markers and not by a single histone modification. For instance, in the present study, the unexpected increase in acK9 and me3K4 of H3 in the context of decrease in 11β-HSD2 mRNA expression in the IUGR rat could indicate that the initial stages of gene activation in terms of histone acetylation are preserved or even enhanced in IUGR rats. The significance of these findings remains speculative, but it would appear that when exposed to IUGR, the elongation stages of 11β-HSD2 transcription could be affected, as evidenced by persistent decrease in me3K36 in exon 5 in both genders. Interestingly, our group (24) has recently reported similar results when analyzing the effects of IUGR upon the hepatic IGF-1 histone code.

Our observations of altered transcription binding and epigenetic marks along the 11β-HSD2 gene at a time when these animals have completed nephrogenesis but still do not exhibit overt hypertension have profound importance for a mechanistic understanding of fetal programming. Furthermore, this study sets the stage for future investigation that can identify interventions at an early stage, before the onset of hypertension. For instance, simple cost-effective dietary interventions that can affect epigenetic marks of the 11β-HSD2 gene may be protective and temper the consequences of IUGR.

However, we are aware that the present study does not provide data at later stages of life in these animals, especially at the time of exhibiting hypertension, which constitutes a limitation of this investigation.

Our present findings suggest that decreased levels of substrate (hypoglycemia, decreased folate, and zinc) in utero, as seen in UPI, can induce significant changes in chromatin structure and therefore alter the functional genome's response to subsequent environmental stimuli, laying the groundwork for future postnatal morbidities such as those described in the IUGR kidney.

Perspectives and Significance

The present work has identified in the perinatal and postnatal IUGR rat kidney altered binding of key transcription factors as well as changes in the epigenetic characteristics of 11β-HSD2 gene. Importantly, these alterations are gender specific and occur at a time when there is no overt hypertension in either sex.

We speculate that the present findings could have a role in the perinatal programming of hypertension observed in the IUGR offspring, through decreased 11β-HSD2 mRNA levels. Importantly, the present results set the stage for future investigation that can address interventions, such as dietary modifications, that can modulate the epigenetic response to the IUGR insult.

While epigenetic mechanisms are key for the persistent and heritable changes in gene expression that occur without alteration of DNA sequence, its contribution to the observed gender differences in developmental programming are still uncertain. Therefore, studies of altered epigenetic marking that incorporate sexual dimorphism will be of great importance for understanding the developmental origins of health.

Ongoing research in our laboratory is further investigating whether altered epigenetic marks and changes in transcription factor binding of the 11β-HSD2 gene persist through the development of hypertension at later age in this IUGR animal model.

GRANTS

This research was supported by National Institute of Child Health and Human Development Grant R03-HD-058782-01 (to M. Baserga) and the University of Utah Primary Children's Medical Center Innovative Grant Award (to M. Baserga).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).