Research Highlights:Highlights from the latest articles in epigenomics

Stem Cell Dna Methylation: A Consequence of Intrauterine Growth Restriction

Evaluation of: Einstein F, Thompson RF, Bhagat TD et al.: Cytosine methylation dysregulation in neonates following intrauterine growth restriction. PLoS One 5(1), E8887 (2010).

Summary of results

Intrauterine growth restriction (IUGR) predisposes afflicted infants towards a myriad of postnatal diseases, including Type 2 diabetes. Data from multiple animal models suggest that predisposition toward Type 2 diabetes involves epigenetic modifications to chromatin. Einstein et al. hypothesized that the in utero environment that causes IUGR would affect the DNA methylation status of CD34⁺ hematopoietic stem and progenitor cells obtained from cord blood [1]. To test this, cord blood was obtained from IUGR and control infants (n = 5 for each), with IUGR defined as a birth weight and ponderal index of less than the tenth percentile. Controls were matched to IUGR by gestational age, ethnicity and gender. CD34⁺ cells were isolated. High resolution HpaII tiny fragment enrichment by ligation-mediated PCR (HELP) assays were used to initially identify differences in DNA methylation, and the results were verified by bisulfate mass-array validation. The most important observation of the study is that IUGR subjects demonstrated a number of consistent differences in methylation near genes in processes involved in stem-cell function. Limitations of the study include the small sample size and the applicability to other tissues.

Translational relevance

A considerable translational strength of this study is the focus upon a population of stem cells. Hematopoietic stem cells are multipotent cells characterized by multiple lineages. Stem cells, while containing a unique epigenetic signature, propagate their epigenetic ‘message‘ to future differentiated cells. A subpopulation of these cells, identified by the presence of the cell surface marker CD34⁺, are thought to be involved in injury repair and give rise to components of the immune system. CD34⁺ hematopoietic stems cells are readily accessible from cord blood and, thus, can be used to indicate potential correlations between pregnancy outcomes and epigenetic potential.

Scientific relevance

This study moves the field forward by demonstrating the power of the HELP assay. The HELP assay provides a means to look for IUGR-induced changes in DNA methylation that occur in novel loci [1]. The method utilizes methylation-sensitive restriction digestion followed by ligation-mediated PCR. HpaII and Msp1 digested DNA is amplified in the presence of two different fluorophores and labeled products are cohybridized on a custom genomic microarray. The technique is capable of using different methylation-sensitive restriction enzymes to increase the number of sites that can be examined and hybridization against any customized microarray. The group looked at DNA methylation in multipotent hematopoietic (CD34⁺) stem cells of IUGR and control-matched infants in an attempt to identify epigenetic modifications induced at an early age that may contribute to increased susceptibility to age-related diseases. It would be of great interest to compare HELP assay results from stem cells versus differentiated cells.

A number of papers have defined stem-cell epigenetic modifications that mark sites for future modifications that will be required for gene expression in differentiated tissues [2–4]. Interference with some of these early epigenetic events can block progression to gene expression. Cytosine methylation as a means of controlling gene expression, first proposed by Riggs, has now been linked to histone covalent modifications [5,6]. By screening for DNA methylation in samples that are limited, it may also be possible to obtain evidence of histone changes. In addition to screening stem cell populations for CG methylation, recent evidence indicates asymmetrical cytosine methylation is also important.

Asymmetrical cytosine methylation, once thought only to be important in plants, has now been shown to be relevant in mammals. Lister et al. have recently shown that embryonic and differentiated cells contain 60 and 45 million methylated cytosines, respectively, in the human genome [4]. In differentiated cells, 99% of the methyl cytosines are in symmetrical CG sites, however in stem cells up to 25% of methyl cytosines are in asymmetrical CHG and CHH sites (where C represents cytosine, H represents adenosine, cytosine or thymine and G represents guanine). Around transcription start sites, methylated CHH and methylated CHG densities are decreased similar to methylated CG, while at enhancer sites, embryonic and differentiated cells utilize symmetrical and asymmetrical methylation differently.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

Intrauterine Growth Restriction, Neural Microrna Expression and Hypertension in the Mouse

Evaluation of: Goyal R, Goyal D, Leitzke A, Gheorghe CP, Longo LD: Brain renin–angiotensin system: fetal epigenetic programming by maternal protein restriction during pregnancy. Reprod. Sci. 17(3), 227–238 (2010).

Summary of study

The basis of this study is three key observations. First, the brain expresses all components of the renin–angiotensin system (RAS). Second, the RAS functions as both an endocrine and paracrine system. Third, epidemiologic studies of low birth weight humans and experimental studies of low birth weight animals reveal an association between altered RAS and hypertension [1]. Subsequently, these authors set out to test the hypothesis that ‘antenatal maternal protein deprivation leads to epigenetic changes and alterations in gene expression of brain RAS of the mouse fetus‘. To test this hypothesis, FVB/NF pregnant mice experienced protein deprivation, and a subsequent increase in carbohydrate intake, during the second half of pregnancy. Promoter methylation analysis and microRNA (miR) studies were used to define the epigenetic consequences of diet, as well as measures of mRNA and protein expression. The most important observations of these studies include hypomethylation of the angiotensin converting enzyme -1 (ACE-1) promoter and upregulation of miRs, which are relevant to ACE-1 expression. Limitations of the study include uncertainty of whether the consequences are owing to changes in maternal protein or carbohydrate diet content, use of presumed ‘whole‘ fetal brain, focus upon promoters only and exclusion of postnatal juvenile or adult data.

Translational relevance

Recent years have seen an increase in the understanding of the adult onset consequences of IUGR; these morbidities include the development of hypertension, obesity and Type 2 diabetes [2–4]. Emphasis in the field has turned to the mechanism behind the long term effects of IUGR, with epigenetics taking center stage [5]. Of the many long-term complications of IUGR, renal insufficiency and adult onset hypertension present a significant clinical problem. Much of the understanding of the molecular effects of IUGR on long-term outcomes such as hypertension, have come from rodent and sheep models of IUGR. In these animal models, IUGR is induced at various times during gestation and by various means. A unifying characteristic of the long-term effects of IUGR in the kidney are alterations in the expression of renal transcription factors driving development [6], altered glucocorticoid signaling molecules [7] and increased apoptotic processes [8].

These studies, noting the impact of maternal dietary changes upon expression of RAS components, suggest that the RAS is vulnerable to perinatal events and deprivation in multiple organ systems. An important question is how generalizable these differences are from tissue to tissue. Indeed, one of our goals as a field needs to be defining this tissue specificity. This is an important goal because of the temptation to impose systemic interventions, such as histone deacetylase inhibitors or a pharmacological dietary manipulation, upon a targeted epigenetic ‘mutation‘ in a specific tissue. What may modify morbidity in a single tissue, may exacerbate morbidity in another.

Scientific relevance

A finding that moves the field forward in this study is the affect of maternal protein deprivation upon miR expression [1]. The regulation of genes involved in blood pressure has been shown to involve miR. The downregulation of human angiotensin type II receptor (AT2) by miR-155 can be blocked by a polymorphism at the miR target site within the AT2 3´-UTR and this single base polymorphism has been linked to hypertension [9,10]. miR-mediated inhibition can occur at the transcriptional level by degrading the mRNA, or in the case of miR-155, at the translational level [11]. Similarly, the mineralocorticoid receptor gene, which is also linked to blood pressure homeostasis, is regulated at the translational level by brain-specific miR-124 and miR-135a, which is expressed in the kidney and brain [12,13]. Goyal et al. found protein restriction upregulated miR-27a and miR-27b expression and this was correlated with reduced Ace-1 protein levels. However, direct inhibition of Ace-1 mRNA translation by miR-27a and/or miR-27b remains to be confirmed. Whether or not miR-27a or miR-27b is found to block ACE-1 mRNA translation, perhaps the most significant finding is that the nutritional environment can alter expression of miR.

However, predicting the outcome from altered miR expression may not be straight forward. A gene can have multiple miR target sequences, while a single miR can have multiple targets and the extent of miR suppression could depend on the level of miR expression. For example, miR-124 can also target the glucocorticoid receptor in the brain while Sober et al. observed the greatest repression from miR with minimal or undetectable levels of expression [14,15].

Placental Variability in Methylation Reflects Maternal and Fetal Pathogenesis

Evaluation of: Bourque DK, Avila L, Penaherrera M, von Dadelszen P, Robinson WP: Decreased placental methylation at the H19/IGF2 imprinting control region is associated with normotensive intrauterine growth restriction but not preeclampsia. Placenta 31(3), 197–202 (2010).

Summary of study

This study focuses on the concept that parent-of-origin imprinted genes play a role in pathological fetal growth states such as IUGR, as well as the regulation of ‘normal‘ growth [1]. Investigators have identified 50 parent-of-origin imprinted genes. Two clusters of these genes appear to be particularly relevant to placental function and fetal growth, and the expression of genes within these clusters is regulated by two epigenetic control regions, imprinting control region (ICR)-1 and ICR2. The investigators hypothesized that placental DNA methylation of these control regions will be ‘aberrant‘ in human pregnancies complicated by preeclampsia and/or IUGR. To test this hypothesis, placentas were collected from control, IUGR, preeclampsia and preeclampsia with IUGR pregnancies. DNA methylation was determined for ICR1 and ICR2 using bisulfite modification with methylation-sensitive single nucleotide primer extension and pyrosequencing, respectively. The most important observations of this study are the considerable intraplacental variations in DNA methylation for ICR1; decreased methylation for ICR1 in IUGR placentas; and no significant differences in DNA methylation in ICR2 between all four groups. Limitations of this study include the lack of controls for maternal diet or race and questions about whether the study was powered enough to detect the affect of multiple variables, including gestational age. Insufficient detail is included to make the latter determination.

Translational relevance

This study is particularly relevant and an important first step. Placental integrity is essential for fetal health and placental tissue is amenable to molecular studies as it is readily obtainable from a variety of pregnancies. Failure of the trophoblast to invade the uterine wall contributes to both preeclampsia and IUGR [2]. It has been suggested that IUGR is associated with more profound changes in placental structure than preeclamsia but this appears to be dependent upon the timing of preeclampsia. Poor trophoblast invasion has been associated with dysregulation of IGF2, a paternally imprinted growth factor. IGF2 levels are regulated by the long noncoding RNA, H19, a maternally imprinted gene. The ICR1 associated with H19 is differentially methylated at its CpGs according to parental inheritance. Methylation of the H19 promoter is negatively correlated with H19 expression. At the same time as H19 expression decreases, the expression of IGF2 increases.

The focus on parent-of-origin imprinted genes such as IGF2 and H19 is important, particularly considering the key role that these genes play in early embryonic development and placental function throughout the pregnancy. It is noteworthy that the parent-of-origin imprinted genes represent a specific subset of epigenetically regulated genes. Questions remain as to how vulnerable this subset of genes is to environmental perturbations (e.g., maternal dietary deprivation), and whether vulnerable time periods are similar between the different categories of epigenetically regulated genes. Considering recent implications of epigenetics in syndromes initially believed to be primarily genetic, this information will be crucial.

Scientific relevance

Syndromes associated with IUGR can be due to multiple factors, including epigenetic events [3]. One example involving DNA imprinting is Silver–Russell syndrome. Hypomethylation of the differentially methylated region in the ICR results in expression of H19 and suppression of IGF2[4]. The loss of IGF2 causes severely restricted placental and fetal growth. Differential methylation within the placenta has been demonstrated by analyzing free fetal DNA isolated from maternal blood [5]. In order to access the role of aberrant DNA methylation associated with imprinted genes, particularly involving ICR1 and ICR2 of the H19/IGF2 locus, Bourque et al. looked at CpG methylation in human placentas from pregnancies associated with preeclampsia and/or IUGR [5].

In spite of the evidence of defects in imprinting being associated with IUGR, the authors found only small statistical differences. This could be explained by the observation that, in mice, imprinting profiles at some sites in the placenta are maintained by histone modifications rather than DNA methylation. The imprinting center 2 on mouse distal chromosome 7 is flanked by several paternally repressed genes, with more distal ones inprinted exclusively in the placenta [6]. Repression involved H3K9me2 and K27me3, which was lost when ICR2 was deleted, whereas genetic ablation of DNA methylation within ICR2 had no effect. Something similar may be occurring at the H19/IGF2 locus. Fauque et al. found H19 and IGF2 expression was altered by conditions associated with assisted reproductive technologies [7]. H19 mRNA expression in placenta was significantly increased in spite of the fact that methylation of the ICR was no different to controls.

Singh et al. have recently characterized the histone modifications associated with the H19/IGF2 ICR in mouse embryonic fibroblasts [8]. They found that when paternal H19 is off, H3K79me3 is localized to the H19 promoter and H19/IGF2 ICR while H3K79me1–2 and H4K91ac accumulate at the expressed IGF2 promoter. However it is not known whether these histone marks exist within the placental locations.

Intrauterine Growth Restriction Disturbs Mediators of Placental Protein Turnover

Evaluation of: Gascoin-Lachambre G, Buffat C, Rebourcet R et al.: Cullins in human intra-uterine growth restriction: expressional and epigenetic alterations. Placenta 31(2), 151–157 (2010).

Summary of study

This study is based upon a previous placental transcriptome analysis of a rat model of IUGR [1]. This analysis identified members of the Cullin gene family as placental moderators of fetal growth, which is further supported by a cell culture model of preeclampsia and involvement of a Cullin family member in dolichospondylic dysplasia (also know as 3M syndrome). Cullin genes produce proteins that direct proteolysis via the ubiquitin-proteasome system. Transgenic studies demonstrate that Cullins are necessary for early embryonic development. Based upon this background, the authors hypothesized that aberrant Cullin expression and promoter DNA methylation occurs in IUGR (particularly CUL4B and CUL7). To test this hypothesis, placentas were collected from pregnancies complicated by preeclampsia, preeclampsia associated with IUGR, idiopathic IUGR, vascular IUGR and controls. The most important findings from these investigations include increased Cullin expression, particularly CUL7, in pathological placentas; CUL4A and CUL7 mRNA levels are increased by preeclampsia; and changes in CUL7 placental promoter methylation, while appearing to be associated with disease status, did not appear to directly correlate with gene expression. A unique strength of this article is the measurement of mRNA levels of the Cullin family from multiple exons. Limitations of this study include the paucity of clinical information, as well as the lack of analysis of factors such as fetal gender and gestational age. In addition, in the human studies, it is unclear whether Cullins are a downstream marker or a primary cause. Regardless, the observation that Cullin expression is aberrant in IUGR moves the field forward.

Translational relevance

Animal models, as well human epidemiological studies, consistently observe that placental protein homeostasis plays a role in the pathophysiology of IUGR. Tight regulation of protein degradation is a vital component of cell cycle control, cellular differentiation and protein metabolism. A vast amount of cellular protein degradation is accomplished by the ubiquitin–proteosome system. Proteolysis is accomplished by the 26S proteosome and protein targets are ubiquitinated by a series of ubiquitin activating, conjugating and ligating enzymes. Target specificity is accomplished at the level of the ubiquitin ligase. An important group of ubiquitin ligases are the Cullin family and mRNA levels of these enzymes are known to be dysregulated in rat IUGR placenta.

The observation that Cullins are increased in placentas associated with disease is indicative of alterations in protein metabolism. Although it would be beyond the scope of this article, it would be valuable to know whether the changes in Cullin expression correlated with the degradation of individual targets. We would speculate that the increase in Cullin expression reflects an attempt by the placenta to maximize amino acid availability. Subsequently, this study provides an important initial insight into physiological mechanism through which the placenta adapts to IUGR.

Scientific relevance

Gascoin-Lachambre et al. investigated the role of Cullin proteins in placental diseases including preeclampcia, preeclampcia associated with IUGR, idiopathic IUGR and vascular IUGR [1]. Cullin proteins participate in a multiple array of cellular processes, including DNA replication and repair, transcription, cell cycle transition and signal transduction [2]. Their particular function is to serve as a scaffold in assembling E3 ubiquitin ligase complexes that target proteins for ubiquitination and subsequent degradation by the 26S proteosome [3]. CUL1 is a negative regulator of the cell cycle [4]. CUL2 is required for normal vasculogenesis that involves, in part, regulation of hypoxia-inducible factor 1-mediated transcription [5]. CUL4, which has alternative transcription start sites encoding different N-termini, is essential for histone H3K4, 9 and 27 methylation [6]. Finally, mutations that result in loss of function of the CUL7 gene have been linked to 3M syndrome, which is characterized by severe IUGR but not placental insufficiency [7].