In Utero Oxidative Stress Epigenetically Programs Antioxidant Defense Capacity and Adulthood Diseases

Preeclampsia and oxidative stress

Preeclampsia, a pregnancy-associated disease characterized by hypertension and proteinuria, is considered one of the leading causes of maternal mortality, and its rates continue to rise (108). Numerous studies have shown that preeclampsia is accompanied by increased inflammation and oxidative stress as well as a decrease in AOX capacity (99), maternal plasma biomarkers of which have included increased leptin (8), MCP-1 (59), interleukin (IL)-6 (91), IL-8 (59), malondialdehyde (MDA); thiobarbituric acid reactive substances (TBARS) (75), H₂O₂ (124), and tumor necrosis factor-α (14). Preeclampsia has also been associated with placental pathologies, and a concurrent increase in placental oxidative stress (7, 75, 84, 105, 140), which is thought to decrease placental efficiency and thereby impair fetal growth capacity (52). Despite the extensive data connecting preeclampsia and placental oxidative stress, there is still no definitive cause and effect established between the disease, oxidative balance, and fetal outcomes. It is likely that numerous factors tip the balance toward increased placental oxidative stress, including general maternal inflammation and excessive ROS production, and these factors likely contribute to the development of preeclampsia, which, in turn, feeds the cycle of increased ROS production.

Intrauterine growth restriction and oxidative stress

In addition to preeclampsia, intrauterine growth restriction (IUGR) is now widely accepted as being associated with poor birth outcomes and an increased risk for many adult-onset diseases. Additionally, many (but not all) cases of IUGR are thought to occur due to placental insufficiency (61) [some due to preeclampsia (89)], which is also accompanied by placental oxidative stress (114). As just bdiscussed, the placenta is subjected to high volumes of oxygenated blood to support fetal development (23), and is, therefore, at risk for increased ROS production. A study comparing women who gave birth to either normal-weight or IUGR babies showed that the levels of MDA and xanthine oxidase were higher in maternal plasma, umbilical cord plasma, and the placenta when compared with normal mothers, and this corresponded to a decrease in the AOX potential within these samples (15). In cord blood from IUGR neonates, AOXs as well as the activity of AOX enzymes, including SOD, catalase, and glutathione peroxidase, were significantly decreased when compared with normal-weight babies, and this was accompanied by increased lipid peroxidation (50). Another study comparing small, average, and large-for-gestational-age infants showed that while oxidative stress was higher and AOX status was lower in small and large infants when compared with those of an average size, the AOX status of mothers who gave birth to small but not large babies was lower when compared with mothers of average-sized infants (100), suggesting that IUGR pregnancies are accompanied by whole-system oxidative stress that is potentially passed to the infant from maternal circulation.

Diabetes and oxidative stress

While patients with diabetes often exhibit signs of whole-system oxidative stress (78), several recent studies have suggested that in utero exposure to hyperglycemia, either in pregnancies complicated by gestational diabetes mellitus (GDM) or in women with uncontrolled Type 1 or Type 2 diabetes, results in oxidative stress in the embryo or fetus. This likely occurs, because the metabolism of excess glucose results in the increased production of ROS. Cell culture studies using rat embryos showed that incubation with glucose led to an increase in the activity of SOD, the mRNA expression of Cu/ZnSOD, MnSOD, and GPx (40), as well as a decrease in the activity of the Glutathione (GSH) AOX enzyme (123). In humans, placentas of women with GDM had increased isoprostane release, SOD activity, and protein carboyl activity (30), and an analysis in placental explants after exposure to oxidative stress demonstrated that placentas of women with gestational diabetes were less responsive in terms of isoprostane release and NF-kappaB DNA-binding than women without GDM, suggesting a decreased capacity to respond to oxidative stress in these tissues (29). These molecular alterations in response to oxidative stress could be potentially linked to embryonic malformations (90), and due to this, elevated gestational glucose has been classified by some as a teratogen (134).

Maternal obesity and oxidative stress

Obesity is a state of increased inflammation and oxidative stress, and there is evidence that gestational obesity not only affects maternal oxidative status, but can also impact the offspring. It has been suggested that the reason for this may be the fact that preeclampsia in obese women has been shown to occur two- to fourfold more frequently than in normal-weight women (64, 102), and as discussed earlier, preeclampsia is associated with increased oxidative stress, in both the mother and offspring. However, several studies in animals suggest that the oxidative stress during obese pregnancy may modulate oxidative stress independently of preeclampsia. In a mouse model of diet-induced obesity, oocytes and zygotes of obese dams had significantly increased rates of ROS generation as well as a depletion in glutathione (51), and in offspring of diet-induced obese rats, oxygen radical absorbance capacity (ORAC) levels and catalase (CAT) activity were decreased from birth through adulthood, and SOD activity was increased by the time offspring reached adulthood (17). Additional studies in animals and humans will be needed to tease out the effects of obesity-induced preeclampsia and maternal obesity alone on fetal outcomes and oxidative balance.

Xenobiotics, heavy metals, and oxidative stress

In rats, a gestational exposure to lead and cadmium led to a gender-specific activation of hepatic phase 1 and 2 xenobiotic-metabolizing enzymes in adult offspring, and cadmium had the strongest effect on depressing the activities of AOX enzymes, including Cu/Zn-SOD, Mn-SOD, CAT, GPx, and Glutathione Reductase (GR) (95). In mice, an injection with the endotoxin lipopolysaccharide (LPS) during gestation resulted in fetal malformations, and these were associated with an increase in lipid peroxidation and decrease in glutathione content in maternal liver, embryo, and the placenta. Additionally, a free-radical trapping agent reversed these outcomes, including reducing the number of external malformations (136), suggesting that the teratogenic effects of LPS are, at least in part, mediated by oxidative stress. Exposure to polycyclic aromatic hydrocarbons (PAHs), from either environmental pollutants or maternal tobacco use, during pregnancy has been shown to influence both physical (96) and cognitive development (92) in the offspring. PAH exposure is associated with increased oxidative stress, and a study comparing PAH exposure, pregnancy oxidative stress, and fruit/vegetable intake found that an association between markers of PAH exposure and oxidative stress, and interestingly, the increased intake of fruits and vegetables during pregnancy eliminated this association (60), suggesting not only that PAH exposure in pregnancy is directly associated with maternal oxidative stress, but also that this outcome can be modulated by diet.

Bisphenol A and oxidative stress

Bisphenol A (BPA) exposure has been shown to be toxic to numerous organ systems, in both humans and animals, presumably due to its activity as a pseudo-estrogen (45). Few studies are available to clearly establish the relationship between BPA and oxidative stress, except that it has been proposed that estrogens cause ROS-induced DNA damage, and BPA may behave in a similar manner to induce cellular damage (27). In mice, a 5-day injection of BPA induced SOD and decreased catalase activity in the liver, and led to an increase in reduced glutathione and glutathione disulfide in the brain, kidney, liver, and testes, and these outcomes were hypothesized to be due to the BPA-mediated increase in hydrogen peroxide production in mouse organs (58). Due to its teratogenic effects, the in utero exposure to BPA has received considerable attention (42), and recent animal studies have attempted to provide the molecular basis for the deleterious effects of BPA exposure on fetal outcomes. BPA exposure limited to the periods of gestation and lactation in a mouse model resulted in a decrease in the weight of the brain, kidneys, and testes in 4-week old offspring, with a concurrent increase of TBARS in these tissues and an increase in the activity of catalase in the liver and glutathione peroxidase in the kidneys (57). While confirming the toxic nature of BPA has become of critical importance, primarily due to its ubiquitous nature, other environmental endocrine disruptors should also be closely monitored and studied to determine their effects on oxidative stress, especially if these also play a role in modulating fetal outcomes and health in adulthood.

Maternal cigarette or nicotine exposure and oxidative stress

The exposure to tobacco smoke during pregnancy has been linked to negative pregnancy outcomes (101), and numerous studies have suggested that these outcomes may be directly associated with the capacity of cigarette smoke or nicotine to induce oxidative stress, both in the mother and the offspring. A recent transcriptome analysis of pregnant smokers and non-smokers showed that smoking was associated with altered expression of numerous oxidative stress-related genes in peripheral blood, placenta, and cord blood, suggesting that smoking significantly affects oxidative balance in maternal and fetal tissues (128). In fetal explant cultures, treatment with cigarette smoke extract resulted in increased markers of oxidative stress and a decrease in anti-apoptotic markers (81). While poor pregnancy outcomes have been historically associated with cigarette smoke itself, studies suggest that nicotine, independent of smoking, has dire consequences during pregnancy, potentially by inducing oxidative stress. In rats, nicotine administration increased arterial contractions and hypertension in the offspring, and this was associated with decreased SOD activity, and increased superoxide, MDA, and nitrotyrosine protein levels in the vascular walls of adult offspring (133). Additionally, offspring of rats that were prenatally exposed to nicotine had pancreatic islet oxidative stress, mitochondrial abnormalities, glucose intolerance, and reduced glucose-stimulated insulin secretion, indicating that prenatal nicotine exposure induced a diabetic phenotype in offspring (20). Specifically, the markers of pancreatic oxidative stress in offspring in response to nicotine exposure included increased Gpx-1, MnSOD, islet ROS, and protein carbonyl production (19). These studies unquestionably support the notion that maternal smoking induces oxidative damage, in both the mother and the fetus, and future studies are needed to determine whether addressing the oxidative effects of smoking and nicotine can ameliorate the damage caused by maternal smoking.

Maternal HF diet and oxidative stress

A HF diet is thought to alter oxidative balance by increasing lipid peroxidation and the production of ROS, and our group recently showed that a gestational HF in rats repressed the expression of AOX genes in livers of adult offspring. Despite being on a standard control diet for 12 weeks after birth, offspring exposed to a gestational HF diet had increased plasma triglycerides, and their livers had increased TBARS, as well as decreased transcription of Gpx-1, Sod1, Pon1, Pon2, and Pon3 (135). Our data demonstrates that the maternal environment is a critical programmer of the AOX capacity, and that molecular events programmed in utero can last into adulthood. Unpublished observations from our group have also found that a gestational HF diet resulted in the gender-specific increased mRNA expression of AOX enzymes in livers of neonatal rats. These results suggest that while the fetus initially responds to a maternal HF diet by increasing its AOX capacity, this response may be short lived. Under a chronic stress, the AOX system no longer functions with time, which eventually leads to oxidative stress-induced tissue damage (Fig. 2). Additionally, what remains to be clarified is whether the dysregulation in oxidative balance in the offspring is due to the increased in utero supply of maternal lipids itself, or the innate increase in the production of lipids or ROS due to the offspring's programmed altered lipid handling and metabolism.

OPEN IN VIEWER

FIG. 2.

Maternal HF diet and the offspring's AOX defense system. Studies have suggested that a maternal HF contributes to numerous diseases in the offspring, including nonalcoholic liver disease, and this may be accompanied by a dysregulation of the AOX defense system in the offspring. The exact mechanism behind these observations remains undefined and is likely complicated and multi-faceted. Potentially, a maternal HF diet dysregulates maternal lipid metabolism, which increases maternal oxidative stress. The increased ROS produced during maternal lipid peroxidation are sensed by the placenta, which either increases its own ROS production, or transports maternal peroxidation products to the offspring. This is also accompanied by an increased transport of free fatty acids from the mother to the fetus. The excess fatty acids may lead to endogenous ROS production, and these, along with those from the mother, initially activate the AOX defense system. However, it is possible that with time, and under a chronic ROS burden, the AOX system no longer functions at capacity, which leads to oxidative damage and eventual tissue damage. AOX, antioxidant; HF, high fat; NEFA, non-esterified fatty acids; TG, triglycerides.

Maternal protein restriction and oxidative stress

Our group has shown that gestational protein restriction in rats has severe consequences for the growth capacity of offspring (109), and Ozanne et al. have repeatedly demonstrated that protein restriction during gestation or lactation influences oxidative balance in the offspring. A gestational low-protein (LP) diet, when compared with post-natal protein restriction, was shown to increase renal albuminuria in young mice, and this was accompanied by altered expression of genes that function to protect the kidney against oxidative stress (26). Additionally, a gestational LP diet resulted in an increase in markers of fibrosis in pancreatic islets of adult offspring, and this was accompanied by increased islet expression of xanthine oxidase, and decreased MnSOD, Cu/Zn-SOD, and heme oxygenase-1 (117), demonstrating increased oxidative stress and diminished AOX capacity in pancreatic cells of offspring exposed to a gestational LP diet. Using the same model, aortas of offspring exposed to a gestational LP diet had lower MnSOD protein content than the postnatally LP-fed group, and there was a significant decrease of MnSOD between 3 and 12 months of age in the gestational LP group, suggesting a decrease in AOX capacity in these animals. These changes were also associated with increased urinary 8-oxo-dG, a marker of oxidative stress, in the gestational LP group by 12 months of age when compared with both the control and the postnatal LP groups (118). How gestational protein restriction modulates the AOX defense system in the offspring remains unclear. Potentially, protein malnutrition can be sensed by several molecular signaling pathways as a cellular stressor, and this activates the AOX response, which diminishes with age. Alternatively (or concurrently), inadequate protein supply during pregnancy may increase the metabolism of other substrates, such as lipids and carbohydrates, by both the mother and fetus, therefore increasing free fatty acids, ketones, and glucose, all of which have been shown to be capable of dysregulating oxidative balance.

Gestational dietary interventions that decrease oxidative stress

Certain maternal dietary interventions have been shown to positively influence oxidative status. When a HF diet supplemented with dietary AOXs (Vitamins A, C, and E, as well as selenium) was fed to pregnant rat dams, the offspring of these animals had significantly improved outcomes when compared with offspring of dams fed the HF diet without AOX supplementation. These beneficial changes included decreased adiposity at 2 weeks and 2 months of age, decreased TBARS in the fetus and neonate, and decreased ROS and increased GSH in the preimplantation embryos (103). A high fiber diet fed to pregnant mice that were also consuming a HF diet improved the hydroxyl radical scavenging capacity in the placenta as well as maternal plasma when compared with the HF diet alone. Additionally, the livers of offspring whose mothers were supplemented with fiber, in conjunction with a HF diet, had increased total SOD, Cu/ZnSOD, and MnSOD protein, as well as increased mRNA expression of these genes, Hif-1α, Trx1, Trx2, and Gpx1. A similar increase of Hif-1α, Trx1, Cu/ZnSOD, and MnSOD mRNA was also observed in fetal hearts when compared with offspring of dams consuming a HF diet alone without fiber supplementation (73). In a rat model of gestational high-saturated fat feeding, fetuses of HF-fed dams had lower total mineralized tissues, crown-rump length, and total bone volume than controls, and quercetin supplementation improved these outcomes. The authors proposed that the HF diet increased oxidative stress, and quercetin, which has been shown to have strong AOX capacity, decreased it, thus improving fetal outcomes (67).

While soy has been shown to have some beneficial effects, the intake of soy during pregnancy has been somewhat controversial, as some animal studies have shown that the intake of phytoestrogenic genistein, the major isoflavone in soy, during either pregnancy or lactation has the ability to interfere with fertilization, oocyte maturation, and general development (10, 25, 28, 98). However, there is also evidence that soy isoflavones may positively affect endothelial tissues by mediating the oxidative response. A study of aged male rats that were fed a soy protein-rich diet during gestation and adult life showed that animals fed the soy diet had improved endothelial function, reduced blood pressure, and these improvements were associated with decreased oxidative stress (76). At the molecular level, these differences in oxidative status between animals fed soy-rich and soy-deficient diets were likely due to the observed increase in mitochondrial glutathione, as well as the increased mRNA expression of endothelial nitric oxide synthase (eNOS), MnSod, and cytochrome c oxidase. Additionally, when the soy-deficient group was re-fed the soy-rich diet for 6 months, they showed improvements in blood pressure, endothelial function, and the expression of AOX genes in vascular tissues. As previously discussed, estrogens have been shown to induce ROS production, and, interestingly, further studies with isoflavone-rich diets in animals have suggested that dietary estrogens improve vascular function by inducing nitric oxide and, therefore, the AOX defense system, potentially through the activation of Nrf2, thus improving endothelial function and defense capacity (106). While these studies suggest that soy intake may be cardioprotective, it is critical to reiterate that ROS production and overproduction exist in a fine balance, and further studies with dietary phytoestrogens will be needed to set intake safety levels, especially during pregnancy, when it is critical to maintain oxygen and ROS levels within a narrow range.

Classification and functions of epigenetic modifications

Both DNA methylation and the covalent modification of histone tails are epigenetic modifications that have been shown to be powerful regulators of gene expression (Fig. 3). DNA methylation most often occurs at “CG”-rich regions, so called “CpG Islands,” and is thought to be relatively permanent (110). Under normal conditions, the genome undergoes several regulated bouts of methylation and demethylation during the earliest stages of development (69); however, in adulthood, any considerable change in the methylation of DNA is primarily associated with abnormal pathologies, such as carcinogenesis (79). Clinical, animal, and cell studies have demonstrated the role of DNA methylation in the development and progression of numerous cancers, including colon, breast, hepatic, ovarian, prostate, and others. While DNA methylation is relatively permanent and results in the silencing of genes, the modification of histone tails has been thought to be less static and more responsive to diet and the environment. Additionally, the complicated interplay or crosstalk between the various histone modifications has made it more difficult to classify these as absolute activators or inhibitors of transcription. Histone modifications act by modifying the physical configuration of the DNA, either loosening or tightening the chromatin, and these arrangements either activate or inhibit gene transcription by regulating the binding of transcription factors and the transcription-initiation machinery (62). Therefore, while certain histone modifications, such as acetylation, are thought to be activators, and others, such as sumoylation, have been shown to be inhibitors of transcription, these are general observations and have many exceptions. Adding to the complexity of epigenetic modifications is the proposed importance of the interaction between them. In addition to the complicated interplay between the many histone modifications, it has also been shown that DNA methylation and histone modifications can regulate each other and/or be regulated by the same factors in response to various stimuli (85). Therefore, while epigenetics has become a new and exciting frontier for the study of gene transcription, the context within which these modifications occur should be considered when making any concrete conclusions about their regulation of specific genes.

OPEN IN VIEWER

FIG. 3.

Epigenetic modifications. The two most common epigenetic modifications include DNA methylation and histone modifications. Methylation of the DNA strand and CG dinucleotides is considered an inhibitor of transcription, while histone modifications can either activate or inhibit gene transcription. Covalent modifications to the tails of histones regulate transcription by either loosening or tightening the DNA helix, and include acetylation (Ac), methylation (Me), sumoylation (S), phosphorylation (P), and ubiquitination (Ub).

The role of epigenetics in the programming of gene expression in utero

Maternal physiology and diet are potent regulators of fetal outcomes, and there is mounting evidence suggesting that these changes occur in response to the in utero programming of gene expression. Epigenetics was suggested to contribute to these outcomes, because early epidemiological studies related to the “Fetal Origins Of Disease Hypothesis” showed that the prevalence of certain diseases was associated with environmental events rather than evolutionary transformations in gene expression (11). Since then, numerous studies have shown that maternal diet results in alterations in fetal birth outcomes and may predict adult disease development (38). Additionally, in animals, studies that manipulate gestational diet have confirmed that these physiological outcomes in the offspring are accompanied by molecular alterations to vital pathways that regulate almost all chronic diseases (2, 13, 22, 49, 55, 82). In an attempt to discover the molecular mechanisms behind the in utero programming of disease, recent studies have focused on connecting the physiological response to maternal diet and any epigenetically induced changes in gene transcription. Table 3 presents an overview of outcomes from studies discussed next as well as from several others.

The hallmark example of the effect of maternal diet on the fetal epigenome came from studies in Avy mice, which were shown to have epigenetic modifications to the Agouti gene that altered coat color, as well as increased their susceptibility toward obesity and tumor development (132). The DNA methylation that led to the altered expression of the Agouti gene was shown to be modified by the addition of methyl donors to the maternal diet, showing a direct link between genotype, phenotype, and maternal intake. In later studies, Jirtle et al. showed that the supplementation of the maternal diet with genistein (a plant isoflavone from soy) increased the methylation of the Agouti gene, which not only changed the coat color in Avy mice, but also protected them from developing obesity, and these effects persisted into adulthood (35). Consequently, further studies of these “Agouti” mice [those carrying the viable yellow agouti A(vy) epiallele] showed that supplementation with other dietary methyl donors, which are known to influence DNA methylation, induced the pseudoagouti phenotype (131). These initial discoveries paved the way for the study of the developmental origins of disease from the epigenetic perspective, specifically focusing on the interaction between maternal nutrition and the fetal epigenome in programming adult-onset diseases.

While research in the field is currently somewhat limited, there are several other animal models of maternal nutritional manipulation that directly connect gestational dietary intake and alterations within the fetal epigenome. Lillycrop et al. have demonstrated that maternal protein restriction results in considerable DNA methylation in the offspring, and affects the transcription of genes critical for hepatic metabolic processes, such as Pppar-α (71) and glucocorticoid receptor (72). Using a similar model of gestational protein restriction, our group also showed that a gestational low-protein diet resulted in a significant increase of hepatic Igf2 transcription, and this was correlated with the methylation of the gene (43). A gestational HF diet in mice was shown to lead to the hypomethylation of dopamine and opioid-related genes in brains of offspring that also had a preference for sucrose and fat (129). Either supplementation or deficiency of folic acid and other methyl donors during pregnancy has also been implicated in regulating methylation events in the offspring (129), as the process of methylation requires a constant supply of one-carbon molecules. As discussed earlier, histone modifications, rather than DNA methylation, may represent a more immediate response of the fetal epigenome to the maternal nutrient supply. A maternal low-protein diet resulted in gender-specific histone modifications within the Cebp/β gene in skeletal muscle of adult offspring that corresponded to the transcription of this critical metabolic regulator (138). The same treatment also resulted in histone modifications and the repression of p16/Cdkn2a, a key cell-cycle control gene, in mammary glands of adult female offspring, an event that has been linked to breast carcinogenesis (137). Gestational choline supplementation led to an increase in H3K9Me2 and H3K27Me2, while choline-deficiency increased the levels of H3K4Me2 in offspring rats (33). In another study in rats, gestational food restriction led to the dimethylation of histone H3 at lysine residue 4 (H3K4Me2) within the Igf1 promoter, which affected the mRNA expression of the gene (122), and a gestational low-protein diet in mice resulted in a decrease of H3K4Me3 and H4K20Me3, and an increase of H3K9Me3 and H3K27Me3 within Igf2, which was accompanied by a decrease in the mRNA expression of the gene (104). Although it is clear that a maternal diet that is high in fat is deleterious to fetal development, the consequences of a gestational HF diet on the histone code have not been thoroughly elucidated. We recently reported that a maternal HF diet, independent of maternal obesity development, results in a decreased association of acetylated histone H3 (H3Ac), H3K4Me2, H3K9Me3, and H3K27Me3 within the promoter of the hepatic gluconeogenic Pck1 gene. Additionally, our analysis of the Pck1 coding region showed that a maternal HF diet led to a decrease in H3K9Me3, and an increase in acetylated histone H4 (H4Ac) and H3K4Me2, and that these modifications were associated with an increased transcription and transcriptional rate of the Pck1 gene (109). In primates, the chronic consumption of a maternal HF diet led to an increase in H3K14Ac, a trend of increase in H3K9Ac, H3K18Ac, H3K9Me2, H3K9Me3, and H3K27Me3, and these fetal hepatic tissues had an increase in triglycerides and non-alcoholic fatty liver disease (1). The studies just mentioned suggest that a maternal diet in which either macro- or micro-nutrients are manipulated alters fetal chromatin structure in rodents as well as in primates via covalent histone modifications, and this contributes to changes to gene expression as well as to the phenotype of the offspring.

Epigenetic modifications of the AOX defense system

Due to the AOX defense system's ability to quickly and efficiently respond to numerous environmental events, it is not surprising that it has been shown to be regulated at the epigenetic level (Table 4). There is a strong support from human, animal, and cell studies showing that epigenetic events are critical for the regulation of cancer development and progression, and since cancer patients often exhibit increased oxidative stress, several studies have focused on the role of epigenetics in oxidative balance during carcinogenesis. While ROS themselves are potent epigenetic modifiers of critical carcinogenesis-related genes (139), genes involved in oxidative balance have themselves also been shown to be epigenetically modified. In breast cancer cells, which often have decreased MnSod activity, the gene that encodes the enzyme (SOD2) was hypermethylated, suggesting that its expression was epigenetically regulated (48). Additionally, the decrease in SOD2 transcription was accompanied by a decrease in dimethylation of histone H3 at K4 (a transcriptional activator) and hypoacetylation of histone H3 at K9, and reversing the hypoacetylation increased SOD2 expression (47), suggesting that the gene is at least in part regulated by histone modifications that influence the condensation state of the chromatin. In prostate cancer, the epigenetic regulation of glutathione-s-transferase P1 (GSTP1) has also been shown to be responsible for this AOX's decreased transcription (127) and an increase in oxidative damage (125). The GSTP1 promoter is hypermethylated during prostate cancer (83), thus decreasing its transcription and AOX activity, which has been linked to the development of high-grade intraepithelial neoplasia (87).

Atherosclerosis is marked by increased oxidative stress and a decrease in AOX capacity, and epigenetic events appear to regulate genes in atherosclerotic patients. In the atherosclerotic aorta, inducible nitric oxide synthase (iNOS) has been shown to be upregulated (93), which could potentially be due to its hypomethylation (24). Additionally, both extracellular SOD (66) and 15-Lipoxygenase (46) are hypomethylated in atherosclerotic lesions, which likely regulates their activation in response to oxidative stress. In an animal model of spontaneous pulmonary artery hypertension (PAH), SOD2 was significantly decreased in the pulmonary artery smooth muscle cells [which has also been shown in humans with PAH (16)], and this decrease was accompanied by a significant increase of SOD2 promoter methylation (6), suggesting that the decreased transcription and activity of this AOX during pulmonary hypertension is a result of its epigenetic status.

In utero epigenetic programming of the AOX system: The need for further research

While little data are available regarding the epigenetic modifications occurring within the AOX system in response to maternal diet, the response of AOX genes to maternal nutrition is likely similar to that of other vital pathways, in that they are regulated in utero, which likely occurs at the epigenetic level. A few studies are available relating to nutrition and epigenetics in humans, but a study of mucosal swabs showed that MnSOD was differentially expressed in vegetarians versus meat-eaters, and that this was associated with differential global DNA methylation as well as methylation of CpG islands within the MnSod gene (119). Unpublished observations from our lab have shown that a maternal HF diet impacts the expression of the Pon genes in liver, and these changes are gender-specific and may occur through the epigenetic modifications of these genes. We showed that Pon1, Pon2, and Pon3 were increased after a maternal HF diet in livers of male but not female neonates. Histone modifications within the Pon1 gene promoter corresponded to the mRNA expression of the gene, with increased H4Ac, H3K4Me2 and a decrease in H3K9Me3. These modifications were associated with increased oxidative stress in the liver, suggesting that the fetal AOX system responds to maternal nutrition, which occurs at the epigenetic level (Fig. 4).

OPEN IN VIEWER

FIG. 4.

Epigenetic regulation of hepatic Pon1 by maternal HF diet. Although the direct mechanism behind the regulation of the AOX system remains unknown, unpublished data from our lab suggests that a maternal HF diet can induce epigenetic changes in the fetal epigenome and affect AOX genes, including hepatic Pon1. We observed that fetal livers of offspring whose mothers consumed an HF diet during gestation but did not develop obesity (image represents OR-CD [Obese Resistant] rats from Charles River Laboratories) had increased TG and TBARS, which was accompanied by the increased mRNA expression of the Pon1 gene, a hepatic antioxidant. Chromatin immunoprecipitation, a technique utilized to test proteins associated with the chromatin, showed that the Pon1 promoter was enriched with H3K4Me2 as well as H4Ac, both of which are associated with active transcription and act by loosening the DNA to allow for the binding of various transcription factors. H3K4Me2, dimethylated histone H3 at Lysine residue 4; H4Ac, acetylation of histone H4.