| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Division of Neonatology, Department of Pediatrics, University of Utah School of Medicine, Salt Lake City, Utah
2 David Geffen School of Medicine and Department of Pediatrics, Mattel Children's Hospital, UCLA, Los Angeles, California
3 Department of Pediatrics, University of Arkansas Medical Sciences, Little Rock, Arkansas
4 Department of Nutrition, University of California, Davis, California
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Barker's fetal origins of adult disease hypothesis; zinc; DNA methylation; histone deacetylase; histone acetylation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Uteroplacental insufficiency caused by bilateral uterine artery ligation in the pregnant rat subjects the rat fetus to a intrauterine environment that is similar to the human condition (77–79, 96, 97). The IUGR rat recovers quickly from the initial insult during the perinatal period and appears to be metabolically normal until it develops insulin resistance in young adulthood (88, 94). Because of presumed protective mechanisms such as the "diving reflex," the brain of the IUGR rat does not experience the same magnitude of insult as other tissues; however, the IUGR brain is affected by the altered intrauterine environment, as evidenced by altered perinatal mRNA levels of key apoptotic proteins as well as by persistent changes in mRNA levels of genes related to energy metabolism. These changes last beyond the period of recovery and occur before the onset of insulin resistance (57, 58).
The molecular mechanisms underlying these changes in mRNA levels are unknown, but the aforementioned findings suggest an alteration in transcriptional regulation that is relatively persistent. Epigenetic modifications of chromatin structure cause persistent alterations in transcriptional regulation and involve processes such as DNA methylation and histone acetylation.
Methylation of CpG dinucleotides is an important epigenetic mechanism that alters chromatin structure and thereby influences processes such as DNA replication and DNA transcription (12, 71, 76). DNA methylation inversely correlates with histone acetylation, which alters histone-DNA contact and affinity (40, 63, 76). Considering the ubiquitous nature of DNA methylation and histone acetylation in genome control pathways, these phenomena are possible mechanisms through which uteroplacental insufficiency could initiate a "metabolic imprint." By permanently altering cerebral chromatin structure and subsequently affecting patterns of gene expression, IUGR could contribute to the pathogenesis of postnatal complications such as neurodevelopmental delay and neuroendocrine reprogramming (98).
We therefore hypothesized that uteroplacental insufficiency in the rat would alter cerebral DNA methylation and histone acetylation. To prove this hypothesis, we measured DNA methylation, as well as histone acetylation, in both control and IUGR rat brains at day of life 0 (d0) and day of life 21 (d21). Because of the differences we found at d0, we used two approaches to identify cellular mechanisms through which these differences occur. First, we measured and localized expression of proteins involved in DNA methylation [DNA methyltransferase 1 (DNMT1); methyl-CpG binding protein 2 (MeCP2)] and histone acetylation [histone deacetylase (HDAC)1] in control and IUGR d0 rat brains; d0 cerebral HDAC activity was also measured.
Second, we measured cerebral levels of trace minerals, with a particular interest in zinc, and one-carbon metabolites in control and IUGR d0 rat brains. Cerebral zinc levels are relevant because zinc deficiency leads to IUGR, and zinc is a cofactor for many of the enzymes involved in determining chromatin infrastructure, including DNMT1 and HDAC1 (26, 69, 74, 86). We investigated the effect of uteroplacental insufficiency on whole cerebral one-carbon metabolism because it potentially contributes to the regulation of genome-wide DNA methylation via S-adenosylmethionine (SAM) and S-adenosylhomocystine (SAH) levels.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
To study 21-day-old (d21) rats, dams were allowed to deliver spontaneously and litters were culled to six pups, as previously described (47, 48, 52–55). At d21, Con and IUGR rats were separated from their dams for 4 h, anesthetized, and killed (n = 6 litters for Con and IUGR, respectively). Pups were studied at this age because they have not yet developed overt insulin resistance or dyslipidemia that may confound our studies, and because they are normally separated from their dams as weanlings at this age (52, 88, 94). The pups were fasted because of the potential affects of heterogeneous nutritional states on epigenetics (63). Analysis on the basis of sex was performed because previous studies demonstrated that both rodents and humans respond to a deprived intrauterine milieu in a sex-specific manner (28, 37, 52, 54, 91). d21 whole brains were collected as described for d0 above. In addition, two d21 rats per litter underwent transcardiac perfusion-fixation with 0.9% NaCl, followed by 10% formalin. Brains were removed and then were put in 10% formalin at 4°C overnight for embedding (90).
Cytosine extension assay.
The cytosine extension assay was used to measure global and CpG island DNA methylation in IUGR and Con brains with the restriction enzymes HpaII and BssHII, which assess global and CpG island methylation, respectively (82). In brief, 2 µg of genomic DNA was digested overnight with 50 U of BssHII, HpaII, or MspI endonuclease according to the manufacturer's protocol. A second DNA aliquot was incubated without restriction enzyme addition to serve as a background control. The single nucleotide extension reaction was performed in a 25-µl reaction mixture containing 0.25 µg of DNA, 1x PCR buffer II, 1.0 mM MgCl2, 0.25 U of AmpliTaq DNA polymerase, and [3H]dCTP (57.4 Ci/mmol), incubated at 56°C for 1 h, and then placed on ice. Six 10-µl aliquots from each reaction were applied into Whatman DE-81 ion-exchange filters and washed three times with 0.5 Na-phosphate buffer (pH 7.0) at room temperature. The filters were dried and processed for scintillation counting. Background radiolabel incorporation was subtracted from enzyme-treated samples, and the results were expressed as relative [3H]dCTP incorporation/0.5 µg DNA or as percent change from Con samples.
Histone isolation and Western blotting.
Western blotting was used to measure overall acetylation for histones H3 and H4 at d0, as well as histone 3 lysine 9 (H3/K9) and histone 3 lysine 14 (H3/K14) site-specific acetylation at d0 and d21. In brief, histones were isolated from d0 and d21 whole brain tissue by acid extraction according to Galasinski et al. (33). Histone concentrations were determined with a micro-bicinchoninic acid (BCA) protein assay kit (Pierce Biotechnology). Ten to twenty micrograms of histones were separated on 15% SDS-PAGE gels and transferred by electroblotting to polyvinylidene difluoride membranes (Millipore, Billerica, MA). Blocking was carried out with freshly prepared PBS plus 3% nonfat milk. After washing, the membrane was incubated overnight with primary antibodies diluted in PBS-milk. Primary antibodies included anti-acetyl-H3 at 1:1,000 (Upstate Cell Signaling, Lake Placid, NY), anti-acetyl-H4 at 1:1,000 (Upstate Cell Signaling), anti-acetyl-H3/K9 at 1:400 (Cell Signaling, Beverly MA), anti-acetyl-H3/K14 at 1:5,000, anti-acetyl histone 3 lysine 18 at 1:500, and anti-histone H3 at 1:2,000 (Upstate Cell Signaling). Separate membranes were used for each primary antibody (histone and nonhistone). After a result was obtained from acetylated H3, the same membrane was stripped with stripping buffer (62.5 mM Tris·HCl pH 6.7, 2% SDS, 100 mM 2-mercaptoethanol) at 55°C for 30 min and then was reprobed with total H3. Secondary antibodies conjugated with horseradish peroxidase (HRP) were incubated for 1 h at room temperature. Signal was detected with enhanced chemiluminescence (ECL) according to the manufacturer's instructions (Amersham, Little Chalfont, UK). The amount of site-specific acetylated H3 was quantified relative to the amount of total H3 in the sample.
Immunohistochemistry analysis.
Immunohistochemistry was used to localize the differences in H3 acetylation occurring between IUGR and Con groups, as well as DNMT1, MeCP2, and HDAC1 expression. In brief, coronal sections of embedded brain tissue were deparaffinized and rehydrated in a graded series of ethanol and distilled H2O (dH2O) with a final wash in PBS. Sections were then incubated in a 3% H2O2 solution for 30 min at room temperature to quench endogenous peroxidase activity. After a brief PBS wash, slides were subjected to an antigen retrieval procedure (Biogenex Laboratories, San Ramon, CA), in which the slides were put in a slide tray with 10 mM citrate solution, pH 6.0, placed in a microwave oven, and heated on high power for 165 s and on low power for 8 min, after which they were cooled to room temperature. Slides were then rinsed with tap water, washed in PBS for 10 min, and incubated in a blocking buffer (2% normal goat serum, 2% bovine serum albumin, 0.8% Triton X-100, 0.2% nonfat dry milk in PBS) at room temperature for 1 h. They were then probed overnight at 4°C in a humidified chamber with DNMT1, MeCP2, or HDAC1, acetyl-H3/K14, or acetyl-H3/K9 antibodies at 1:100 to 1:1,000 dilution in blocking buffer. The next day, sections were washed in PBS containing 0.2% Tween 20 three times and exposed to biotinylated goat anti-rabbit antibody for 1 h. After exposure to a Vectastain avidin-biotin complex mixture for 1 h, slides were washed in PBS for 15 min, stained with diaminobenzidine (Sigma), counterstained with hematoxylin, dehydrated, and coverslipped with Cytoseal 60 (Stephens Scientific, Kalamazoo, MI).
Transmission electron microscopy.
Transmission electron microscopy was performed to visualize chromatin ultrastructure. The analysis focused on the hippocampus because this region is vulnerable to IUGR insult and contains a high density of hormone receptors (42, 59). Brains of d0 rat pups were fixed for transmission electron microscopy. Briefly, the brains were placed in 2.5% glutaraldehyde-1% paraformaldehyde in buffer (pH 7.4, 310 mosmol/kgH2O, 4°C for 24 h), after which the hippocampus was isolated. The principles of systematic, uniform, and random sampling were used to collect tissue blocks of hippocampus (1 mm3; 4–6 per hippocampal region per brain) (7). The tissue blocks were postfixed in 1% osmium tetroxide, dehydrated in a graded acetone series, and infiltrated and embedded in epoxy resin. Thin sections (80-nm thickness) were cut with the aid of a diamond knife, counterstained with uranyl acetate and lead citrate, and analyzed with a Hitachi H-7100 transmission electron microscope. Neurons in the hippocampal region were photographed at the same magnification in the upper left corner of each grid square for an entire thin section per tissue block. Thin sections from four tissue blocks were photographed per brain. Image analysis for nuclear dimensions was performed on digitized images. Nuclear average, perimeter, and maximum and minimum diameters (measured perpendicularly) were determined with Bioquant Image Analysis software (Bioquant, Nashville, TN). Ten nuclei per thin section were measured.
RNA isolation and real-time RT-PCR.
Brain mRNA levels of DNMT1, HDAC1, and MeCP2 were measured by real-time RT-PCR, as previously described (54, 81). In brief, total RNA was extracted from d0 brains with an RNeasy Minute Kit (Qiagen, Valencia, CA), treated with DNase I (Ambion, Austin, TX), and quantified by ultraviolet absorbance (16). Sample integrity was confirmed by gel electrophoresis.
The probe and primers were designed with Primer Express (PE Applied Biosystems, Foster City, CA) with a reporter dye FAM and a TAMRA quencher dye (Table 1). cDNA was synthesized from 2 µg of DNase-treated total RNA. cDNA- and gene-specific probe and primers were added to Taqman universal PCR master mix (PE Applied Biosystems), and samples were run on an ABI Prism 7900. Real-time RT-PCR quantification was then performed with the Taqman GAPDH as an internal control. Before the use of GAPDH as a control, serial dilutions of cDNA were quantified to prove the validity of using GAPDH as an internal control. Relative quantification of PCR products was based on value differences between the target and GAPDH control by the comparative threshold cycle method (Taqman Gold RT-PCR manual, PE Applied Biosystems). Cycle parameters were 50°C for 2 min, 95°C for 10 min, and then 40 cycles at 95°C for 15 s and 60°C for 60 s. For each set of reactions, samples were run in triplicate.
|
Nuclear protein extracts preparation.
Nuclear protein was isolated to provide substrate for HDAC nuclear activity assays. Whole brains were ground under liquid nitrogen. Tissue that was used to prepare nuclear extracts for the HDAC activity assay was first washed with ice-cold PBS to lyse red blood cells. The ground tissue was resuspended in 5 ml of ice-cold PBS and then centrifuged (2,000 g) for 5 min. The washing step was repeated until all color was removed. The ground tissue was resuspended in 5 volumes of buffer A (in mM: 1 DTT, 0.5 PMSF, 10 KCl, 10 HEPES, pH 7.9, 1.5 MgCl2) plus protease inhibitors, incubated for 15 min on ice, and then centrifuged (2,000 g) for 5 min. The supernatant was removed by aspiration, and the tissue pellet was resuspended in 2 volumes of buffer A. The tissue was disrupted with a dounce homogenizer, using the tight pestle. Lysis was checked for every 10 strokes with Trypan blue (39). The lysate was centrifuged (1,000 g) for 10 min. The nuclear pellet was resuspended in 0.5 ml of buffer C [in mM: 1.5 MgCl2, 1 DTT, 420 NaCl, 0.5 PMSF, 0.2 EDTA, 20 HEPES, pH 7.9, with glycerol 25% (vol/vol)] plus protease inhibitors, incubated at 4°C for 30 min, and then centrifuged at 17,000 g for 15 min. The supernatant was collected and stored at –80°C.
HDAC enzyme activity assay.
An HDAC activity assay kit (Abcam, Cambridge, MA) was used to measure HDAC activity in the d0 brain nuclear extracts as described by the manufacturer. Fifty micrograms of nuclear protein was added to each well of a 96-well plate. Two microliters of the HDAC inhibitor Trichostatin A at a final concentration of 1 mg/ml was added to the negative control wells.
Mineral analysis.
Whole brain zinc, iron, and copper levels were measured as previously described (43). Briefly, brain tissue was minced and digested with concentrated nitric acid and wet-ashed with a modification of the method of Clegg et al. (14). Whole brain zinc, iron, and copper levels were analyzed by flame atomic absorption spectroscopy (model Smith-Heifjie 4000, Thermo Jarrell Ash, Franklin, MA).
One-carbon metabolite measurement.
Intracellular SAM, SAH, homocysteine, methionine, cysteine, and adenosine were measured by HPLC with coulometric electrochemical detection as previously described (72, 73). Briefly, samples of frozen brain (10–15 mg wet wt) were homogenized in 200 µl of PBS. To reduce sulfhydral bonds, 50 µl of freshly prepared 1.43 M sodium borohydride solution containing 1.5 mM EDTA, 66 mM NaOH, and 10 µl isoamyl alcohol was added. To precipitate proteins, 250 µl of ice-cold 10% meta-phosphoric acid was added and mixed well, and the sample was incubated for 30 min on ice. After centrifugation (18,000 g) for 15 min at 4°C, the supernatant was filtered through a 0.2-µm nylon membrane filter (PGC Scientific, Frederick, MD). Metabolite elution was performed by HPLC with a Shimadzu solvent delivery system (ESA model 580) and a reverse-phase C18 column (5 µm; 4.6 x 150 mm; MCM, Tokyo, Japan) obtained from ESA (Chelmsford, MA). An isocratic elution at 0.9 ml/min was used. A 20-µl aliquot of plasma extract was directly injected onto the column with a Beckman Autosampler (model 507E). All plasma metabolites were quantified with a model 5200A Coulochem II and CoulArray electrochemical detection systems (ESA) equipped with a dual analytical cell (model 5010), a four-channel analytical cell (model 6210), and a guard cell (model 5020). The concentrations of plasma metabolites were calculated from peak areas and standard calibration curves with ESA HPLC software provided by the manufacturer.
Collection of breast milk and characterization of content.
Dams were separated from d21 pups for 4 h and subsequently anesthetized by intraperitoneal injection with 60 mg/kg ketamine and 12 mg/kg xylazine. Each dam was injected intraperitoneally with 100 U of oxytocin. Gentle pressure was applied around each gland to collect milk drops by capillary action into a sterile glass Pasteur pipette. The dams were killed after milk collection. Caloric, fat, and protein content were determined as previously described (10, 64, 92). Mineral analysis was performed by atomic absorption spectrometry as described above.
Statistics.
All data are expressed as mean ± SE percentages of control. ANOVA (Fisher's protected least significant difference) and the Student's unpaired t-test determined statistical significance. P < 0.05 was accepted as indicating statistical difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
d0 and d21 cerebral histone acetylation.
In the d0 IUGR rat brain, total histone H3 acetylation was increased to 157 ± 16.1% of control values (P < 0.05), whereas no difference was noted in total H4 histone acetylation. Site-specific changes were also evident; uteroplacental insufficiency significantly increased whole brain histone acetylation at H3/K14 to 188 ± 24 of control values (P < 0.05), without affecting H3/K9 acetylation (129 ± 27% of control values) (Fig. 2A). Interestingly, immunohistochemistry revealed that that H3/K14 acetylation (Fig. 3A) and H3/K9 acetylation (Fig. 4A) were increased in d0 white matter and hippocampus, respectively, in the IUGR brains relative to control. No sex-specific differences were noted at d0 in terms of histone acetylation.
|
|
|
In contrast, in d21 male rats, total brain H3/K14 acetylation was not significantly different between IUGR and Con animals, whereas H3/K9 acetylation was significantly reduced in brain from IUGR male rats to 68.3 ± 1.9 of control values (P < 0.001; Fig. 2C). Moreover, decreased acetylated H3/K14 and H3/K9 immunostaining were evident in the internal capsule and hippocampus of the male IUGR rats, respectively (Figs. 3C and 4C).
d0 Transmission electron microscopy.
The ultrastructural appearance of nuclei of neurons in the hippocampus was dependent on the rat group. Nuclei appeared smaller with more condensed chromatin around the nuclear envelope in the brains from IUGR pups compared with the Con group, whereas the larger nuclei in the Con group had an open-faced vesicular profile (Fig. 5). Image analysis supported the observed difference in size and showed that neuronal nuclei were significantly smaller in the hippocampus from the IUGR pups (Table 2).
|
|
|
d0 Cerebral mineral concentrations and one-carbon metabolites.
Uteroplacental insufficiency also significantly increased whole brain levels of zinc, while decreasing whole brain levels of iron (Table 3). Copper levels were not significantly affected by uteroplacental insufficiency. Uteroplacental insufficiency significantly increased cerebral levels of adenosine, cysteinylglycine, and cysteine in the IUGR pups relative to the sham-operated controls (Table 3). No significant differences were noted in homocysteine, methionine, SAM, and SAH between the two groups.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
6% of all pregnancies, and impacts human health because of morbidities such as neurodevelopmental delay and neuroendocrine reprogramming (13, 27, 32, 102, 103). These long-term morbidities suggest epigenetics as a responsible mechanism. Epigenetics involves changing determinants of chromatin structure and allows adaptation to the deprived environment through gene expression via an initial investment of energy. The latter is true because DNA methylation and histone acetylation can be maintained through cell division and therefore encode heritable information (45, 95). However, little has been presented identifying specific molecular epigenetic changes in the brain of animals suffering a prenatal insult. We therefore focused this study on the IUGR brain. The novel findings from this work include 1) IUGR affects DNA methylation and histone acetylation; 2) epigenetic alterations occur through d21 and show sex specificity; 3) the effects of IUGR on cerebral epigenetic determinants are region specific; 4) a possible trigger for these changes includes cerebral zinc levels; and 5) at d21, caloric, protein, and fat breast milk content do not differ between Con and IUGR dams, although iron content is significantly increased in IUGR breast milk. Our findings of H3 hyperacetylation and DNA hypomethylation in IUGR rat brain at d0 demonstrate that IUGR impacts epigenetic determinants of cerebral chromatin structure, which is visible in the hippocampus by transmission electron microscopy. To understand the process though which these changes occur, we initially targeted DNMT1, as opposed to other DNA methyltransferases, because DNMT1 represents the majority of methyltransferase activity in embryo lysates (62, 107). The association between decreased cerebral DNMT1 mRNA and protein expression suggests that DNMT1 plays a role in IUGR hypomethylation.
Similarly, global cerebral DNA hypomethylation characterizes transgenic animals in which DNMT1 is knocked out in neuroblasts (25). In these transgenic animals, cerebral cells affected by DNA hypomethylation were lost within the first 3 wk of life, and global DNA methylation normalized relative to the wild-type animals. Our postnatal findings that global methylation was similar between d21 Con and IUGR pups also suggest a loss of postnatal hypomethylated cells. A notable difference between the transgenic study and the present study is that we assessed CpG island methylation and found that CpG hypomethylation persisted in the female IUGR rats at d21. These latter data support the concept that separate mechanisms regulate global DNA methylation and CpG methylation (45). Interestingly, adult mice with decreased levels of DNMT1 expression resist mild to moderate cerebral ischemia (22, 23). These findings raise the intriguing possibility that the decrease in fetal DNMT1 expression in the IUGR brain may be a protective response to minimize the impact of uteroplacental insufficiency on the central nervous system.
The decreased levels of MeCP2 within the IUGR brain may also contribute to cerebral DNA hypomethylation. MeCP2 is required to maintain CpG status of genomic DNA (44, 61). As a result, the appearance of decreased MeCP2 in the hippocampus of the IUGR brain is consistent with the notion that DNA from this region is relatively hypomethylated.
This notion is further supported by the finding of decreased HDAC1 protein in the IUGR fetal hippocampus because DNA hypomethylation associates with histone hyperacetylation, which occurs if HDAC activity is reduced. We focused on HDAC1 because it complexes with DNMT1 and in vitro studies demonstrate that hypoxia decreases HDAC1 protein levels (30, 38). Other investigators have found that multiple stimuli affect hippocampal neuronal histone acetylation. Levenson et al. (60) found that activation of N-methyl-D-aspartate receptors increased hippocampal H3 acetylation, although their studies did not identify the lysines that were specifically affected. Crosio et al. (15) demonstrated that kainic acid (a glutamate receptor agonist) increased H3/K14 acetylation. We found increased acetylation of both H3/K9 and H3/K14 in the d0 IUGR and d21 female hippocampus and white matter, respectively.
Several studies provide insight into the meaning of these findings. In yeast, different cellular stresses induced specific histone acetylation patterns that led to expression of genes from similar functional classes (46). This referenced study found that acetylation of H3/K9 associated with intergenic regions of expressed genes and acetylation of H3/K9 and H3/K14 associated with open reading frames of expressed genes (46). Similarly, a large-scale study of histone modification patterns in human and mouse cells for chromosomes 21 and 22 found that 58% of the acetylated H3/K9 and H3/14 sites coincided within 1 kb of the transcription start of a known gene (5). These studies identify histone acetylation as a mechanism through which uteroplacental insufficiency affects chromatin structure near transcription start sites.
Our findings of altered H3/K14 and H3/K9 acetylation in white matter and hippocampus, respectively, are consistent with these regions being vulnerable to perinatal insults. In rats, uteroplacental insufficiency decreases hippocampal weight and neurogenesis as well as decreasing gliogenesis in the white matter of the cerebellum (11). In humans, placental abnormalities such as infarction and intrauterine hypoxia have been associated with white matter injury (8, 31). Furthermore, voxel-based morphometry has been used to demonstrate that perinatal asphyxia is also associated with hippocampal atrophy in "healthy adolescents" (67).
The vulnerability of the hippocampus to injury is particularly interesting in light of its function and the adult morbidities associated with IUGR. The function of the hippocampus is to "sense" soluble molecules in the blood to perform feedback control, and evidence exists that the hippocampus modulates body physiology, including the hypothalamus-pituitary-adrenal (HPA) axis (26, 39, 59). A major morbidity of IUGR is the neuroreprogramming of the HPA axis. Fernald and Grantham-McGregor (27) found that IUGR in school-aged Jamaican children altered their stress response to psychological and physiological stimuli. Similarly, Cianfarani et al. (13) demonstrated that the neuroendocrine response is permanently altered is some IUGR children, which may affect their catch-up growth. Of particular interest in the latter study, IUGR children who did not catch up were characterized by increased serum levels of cortisol.
Little is known about how sex influences the cerebral epigenetics and subsequent gene expression response to the IUGR insult, although animal studies focusing on other tissues provide some insight. Studies utilizing rats rendered IUGR through uteroplacental insufficiency demonstrate sex-specific differences in growth, serum triglycerides, and gene expression (37, 52, 54). In liver, IUGR rats of both sexes showed decreased liver expression of CPTI at d21, but only the male IUGR rats showed decreased expression of CPTI at day 120 of postnatal life relative to the controls (52). Furthermore, at d21, the livers of male IUGR rats are characterized by a relative H3 hyperacetylation, whereas the livers of female IUGR rats are characterized by a relative H3 hypoacetylation. We speculate that these differences between tissues explain, at least in part, why the general IUGR phenotype is grossly similar between the sexes but many specific characteristics appear to differ.
To identify a possible cellular trigger of the IUGR cerebral epigenetic response, we measured cerebral zinc levels. Zinc deficiency is associated with IUGR, and zinc is a cofactor for DNMT1 and HDAC1 (4, 26, 41, 68, 69, 74, 99, 100). Furthermore, zinc deficiency markedly alters histone solubility compared with other ion deficiencies (24, 70). In our model, IUGR significantly increased cerebral zinc levels, which is likely caused by the rise in metallothionein levels that other investigators have observed in rodent brains stressed by ischemia (93, 101, 106).
We also measured cerebral levels of the important trace elements iron and copper. Considering the association between IUGR and neurodevelopmental delay, our findings of decreased cerebral iron in the IUGR brain are relevant because of the link between decreased hippocampal iron and function in infants of diabetic mothers (75, 80, 87). Unfortunately, little is known about the effects of IUGR on cerebral iron concentrations. Georgieff et al. (34) suggest that live born infants who suffer from restricted maternal-fetal blood flow may be at significant risk for postnatal iron deficiency in multiple tissues, including brain. Moreover, a study from this group found that fetal iron deficiency increased the vulnerability of the rat hippocampus to hypoxic ischemic insult (84).
To investigate another possible cellular mechanism through which uteroplacental insufficiency alters epigenetic determinants of chromatin structure, we measured cerebral levels of metabolites involved in one-carbon metabolism. The most relevant of these metabolites are SAH and SAM, which play a role in the regulation of genome-wide DNA methylation (86). In liver, IUGR significantly increases SAH, a response previously demonstrated to be associated with DNA hypomethylation (9, 66). The subsequent failure of IUGR to affect cerebral SAH levels in association with DNA hypomethylation highlights the tissue-specific nature of the IUGR response. This tissue-specific response is not surprising in light of previous work that demonstrated that folate/methyl donor deficiency causes DNA hypomethylation and decreased DNMT activity in liver but not in other tissues (83).
Interestingly, although cerebral SAH levels were not affected, uteroplacental insufficiency did increase cerebral cysteine, cysteinylglycine, and adenosine levels, all by-products of one-carbon metabolism. Cysteine is the rate-limiting precursor for glutathione synthesis (6). Astrocytes provide cysteine to neurons by releasing cysteinylglycine (105). Adenosine is released from cells under oxidative stress and acts as a neuroprotectant in the rodent hippocampus by attenuating the cellular consequences of reactive oxygen species (1, 85). As a result, the increased levels of all three metabolites (cysteine, cysteinylglycine, adenosine) may reflect an attempt by the IUGR fetus to protect the brain from oxidative stress.
Early postnatal nutrition is important, and the d21 cerebral epigenetic response may also reflect differences in postnatal nutrition, which is why we characterized breast milk from Con and IUGR dams after 21 days of supporting their respective pups. No significant differences in calories, protein, fat, or zinc were noted between the two groups, although differences may have existed earlier. Furthermore, other components of the breast milk not assessed in this study may have contributed to the differences at d21 between the epigenetic characteristics of the Con and IUGR rat brains. The differences in iron content between Con and IUGR dam breast milk are intriguing, because iron content is independent of maternal mineral status and decreased iron levels characterize the IUGR brain (17). We chose not to cross-foster, because this rarely occurs in the human situation and is unlikely to be recommended in the future because of infection issues.
Despite our best efforts, this latter point emphasizes that caution is always necessary when attempting to apply data from a rat model to human pathophysiology. The fetal and juvenile rat is physiologically immature relative to the human, and the insult imposed on the fetal rat in this model of uteroplacental insufficiency is severe. In contrast, the impact of uteroplacental insufficiency experienced by humans ranges across a continuum.
In summary, uteroplacental insufficiency and subsequent IUGR affect epigenetic determinants of chromatin structure in regions of the fetal brain that are known to be vulnerable to perinatal insults such as hypoxia or ischemia. The effects are both sex- and tissue specific. We speculate that changes in epigenetic determinants of chromatin structure alter gene expression in the IUGR rat hippocampus and subsequently trigger the neuroendocrine reprogramming that complicates the IUGR phenotype.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: R. H. Lane, Univ. of Utah School of Medicine, Dept. of Pediatrics, Div. of Neonatology, PO Box 58129, Salt Lake City, UT 84158 (e-mail: robert.lane{at}hsc.utah.edu).
* X. Ke and Q. Lei contributed equally to this manuscript. 
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
coactivator-1 gene expression in a fiber specific manner. Pediatr Res 53: 994–1000, 2003.[CrossRef][ISI][Medline] coactivator-1 gene expression in a rat model of intrauterine growth retardation and subsequent insulin resistance. Endocrinology 143: 2486–2490, 2002.
P < 0.001. D: representative Western blots for d21 female acetyl-H3/K14 and d21 male acetyl-H3/K9. Top bands are acetylated H3, and bottom bands are total H3 internal control; n = 6 litters.
