Differential genomic imprinting regulates paracrine and autocrine roles of IGF2 in mouse adult neurogenesis.

Genomic imprinting is implicated in the control of gene dosage in neurogenic niches. Here we address the importance of Igf2 imprinting for murine adult neurogenesis in the subventricular zone (SVZ) and in the subgranular zone (SGZ) of the hippocampus in vivo. In the SVZ, paracrine IGF2 is a cerebrospinal fluid and endothelial-derived neurogenic factor requiring biallelic expression, with mutants having reduced activation of the stem cell pool and impaired olfactory bulb neurogenesis. In contrast, Igf2 is imprinted in the hippocampus acting as an autocrine factor expressed in neural stem cells (NSCs) solely from the paternal allele. Conditional mutagenesis of Igf2 in blood vessels confirms that endothelial-derived IGF2 contributes to NSC maintenance in SVZ but not in the SGZ, and that this is regulated by the biallelic expression of IGF2 in the vascular compartment. Our findings indicate that a regulatory decision to imprint or not is a functionally important mechanism of transcriptional dosage control in adult neurogenesis.

Different epigenetic states define syncytiotrophoblast and cytotrophoblast nuclei in the trophoblast of the human placenta.

INTRODUCTION: The syncytiotrophoblast (STB) epithelial covering of the villous tree in the human placenta is a multi-nucleated syncytium that is sustained by continuous incorporation of differentiating cytotrophoblast (CTB) cells. STB nuclei display a variety of morphologies, but are generally more condensed in comparison to CTB nuclei. Here, we consider whether this condensation is a feature of epigenetic regulation of chromatin structure. METHODS: Semi-quantitative immunohistochemical investigations of a panel of histone modifications were performed to determine the relative proportions in CTB and STB nuclear populations. We also investigated the patterns of DNA methylation and distribution of DNA methyltransferases enzymes in these populations. RESULTS: Unexpectedly DNA methylation, and H3K9me3 and H3K27me3, which are modifications associated with heterochromatin, are present at lower levels in STB nuclei compared to CTB, despite the intensive condensation in the former nuclear population and the progenitor state of the latter. By contrast, STB nuclei are enriched for H4K20me3, which is also associated with repressive states. 5'hydroxymethylcytosine immunoreactivity is higher in STB, with intense staining observed in the highly condensed nuclei within syncytial knots. DISCUSSION: Cell-type specific epigenetic states exist within the trophoblast populations potentially regulating their different functions and developmental properties and suggesting non-canonical epigenetic states associated with the properties of these cells.

Perturbations to the IGF1 growth pathway and adult energy homeostasis following disruption of mouse chromosome 12 imprinting.

AIM: Disruption to insulin-like growth factor (IGF) signalling pathways during early life causes growth retardation and defects of developing metabolic organs that can alter set points of energy homeostasis for a lifetime. Inheritance of two maternal copies of human chromosome 14q32.2 (Temple syndrome) causes severe foetal growth retardation and post-natal failure to thrive. Disruption of imprinted gene dosage in the orthologous region on mouse chromosome 12 also affects growth. Here, we investigated whether altering chromosome 12-imprinted gene dosage can affect IGF signalling. METHODS: We investigated mice with a transgene insertion at the imprinted domain of chromosome 12. This lesion causes misexpression of neighbouring genes such that the expression of non-coding RNAs is elevated, and levels of delta-like homologue 1 (Dlk1), retrotransposon-like 1 (Rtl1) and deiodinase 3 (Dio3) transcripts are reduced. RESULTS: We observed three key phenotypes in these mice: (i) embryonic growth retardation associated with altered expression of IGF1 binding proteins, (ii) peri-natal failure to thrive accompanied by hypothyroidism and low serum IGF1. Unexpectedly this phenotype was growth hormone independent. (iii) Adult animals had reduced glucose tolerance as a result of endocrine pancreatic insufficiency. CONCLUSIONS: We propose that all of these phenotypes are attributable to impaired IGF action and show for the first time that the chromosome 12 cluster in the mouse is an imprinted locus that modulates the IGF signalling pathway. We propose that growth retardation observed in human Temple syndrome might have a similar cause.

Jdp2 downregulates Trp53 transcription to promote leukaemogenesis in the context of Trp53 heterozygosity.

We performed a genetic screen in mice to identify candidate genes that are associated with leukaemogenesis in the context of Trp53 heterozygosity. To do this we generated Trp53 heterozygous mice carrying the T2/Onc transposon and SB11 transposase alleles to allow transposon-mediated insertional mutagenesis to occur. From the resulting leukaemias/lymphomas that developed in these mice, we identified nine loci that are potentially associated with tumour formation in the context of Trp53 heterozygosity, including AB041803 and the Jun dimerization protein 2 (Jdp2). We show that Jdp2 transcriptionally regulates the Trp53 promoter, via an atypical AP-1 site, and that Jdp2 expression negatively regulates Trp53 expression levels. This study is the first to identify a genetic mechanism for tumour formation in the context of Trp53 heterozygosity.

A quantitative analysis of transcriptionally active syncytiotrophoblast nuclei across human gestation.

The syncytiotrophoblast (STB) epithelial covering of the human placenta is a unique terminally differentiated, multi-nucleated syncytium. No mitotic bodies are observed in the STB, which is sustained by continuous fusion of underlying cytotrophoblast cells (CTB). As a result, STB nuclei are of different ages. Morphologically, they display varying degrees of chromatin compaction, suggesting progressive maturational changes. Until recently, it was thought that STB nuclei were transcriptionally inactive, with all the mRNAs required by the syncytium being incorporated upon fusion of CTB. However, recent research has shown the presence of the active form of RNA polymerase II (RNA Pol II) in some STB nuclei. In this study, we confirm the presence of transcriptional activity in STB nuclei by demonstrating immunoreactivity for a transcription factor and an RNA polymerase I (RNA Pol I) co-factor, phospho-cAMP response element-binding protein and phospho-upstream binding factor, respectively. We also show, through immunoco-localisation studies, that a proportion of STB nuclei are both RNA Pol I and II transcriptionally active. Finally, we quantify the numerical densities of nuclei immunopositive and immunonegative for RNA Pol II in the STB of normal placentas of 11-39 weeks gestational age using an unbiased stereological counting tool, the physical disector. These data were combined with estimates of the volume of trophoblast to calculate total numbers of both types of nuclei at each gestational age. We found no correlation between gestational age and the numerical density of RNA Pol II-positive nuclei in the villous trophoblast (r = 0.39, P > 0.05). As the number of STB nuclei increases exponentially during gestation, we conclude that the number of transcriptionally active nuclei increases in proportion to trophoblast volume. The ratio of active to inactive nuclei remains constant at 3.9:1. These findings confirm that the majority of STB nuclei have intrinsic transcriptional activity, and that the STB is not dependent on CTB fusion for the provision of transcripts.

Genomic imprinting effects in a compromised in utero environment: implications for a healthy pregnancy.

Genomic imprinting in gametogenesis marks a subset of mammalian genes for parent-of-origin-dependent monoallelic expression in the offspring. In mice, the identification and manipulation of individual imprinted genes has shown that the diverse products of these genes are largely devoted to controlling pre- and postnatal growth. Human syndromes with parental origin effects have been characterized both at the phenotypic and genotypic levels, allowing further elucidation of the function and regulation of imprinted genes. Evidence suggests that a compromised in utero environment influences fetal growth through the modulation of epigenetic states. However it is not known whether imprinted genes, by their nature, might be more or less susceptible to such environmental influences. Here we review the progress made in addressing the influence of a compromised in utero environment on the behavior of imprinted genes. We also examine whether these environmental influences may have an impact on the later development of human disease.

Evidence for transcriptional activity in the syncytiotrophoblast of the human placenta.

The aim was to test for evidence of transcriptional activity within the nuclei of the syncytiotrophoblast of the human placenta. The syncytiotrophoblast forms the epithelial covering of the villous tree, and is a multinucleated, terminally-differentiated syncytium generated through fusion of the underlying progenitor cytotrophoblast cells. Its nuclei are heterogeneous with respect to chromatin condensation, and previous functional studies of 3H-uridine uptake in vitro have indicated that they are transcriptionally inactive. This observation is surprising given the key roles this tissue plays in active transport, hormone synthesis and metabolic regulation, and has widespread implications for trophoblast physiology and pathophysiology. We used three different approaches to look for evidence of transcriptional activity. First, immunofluorescence staining was performed on paraffin-embedded early pregnancy and term placental villi, using an antibody directed specifically against the actively transcribing form of RNA polymerase II. Second, a nucleoside incorporation assay was applied to placental villi maintained in short-term culture, with and without the transcription blocker alpha-amanitin. Third, histone modifications associated with active chromatin were identified by immunohistochemistry and immunofluorescence. Each of these methods showed transcription to be occurring in a proportion of syncytiotrophoblast nuclei, with qualitative evidence for transcription being more abundant in the first trimester than at term. These findings correlated with electron microscopical observations of prominent nucleoli within the nuclei, particularly during early pregnancy, signifying transcription of ribosomal RNA. Contrary to previous findings, these results confirm that a proportion of syncytiotrophoblast nuclei actively produce mRNA transcripts.

Disproportional effects of Igf2 knockout on placental morphology and diffusional exchange characteristics in the mouse.

Both complete knockout of the Igf2 gene (Igf2null(+/-)) and knockout of its placental specific transcript alone (Igf2P0(+/-)) lead to fetal growth restriction in mice. However, in the Igf2null(+/-) this growth restriction occurs concurrently in gestation with placental growth restriction, whereas, placental growth restriction precedes fetal growth restriction in the Igf2P0(+/-) mouse. Previous studies have shown that the Igf2P0(+/-) placenta has proportionate reductions in its cellular compartments and its diffusional exchange characteristics. Yet, nothing is known about the structural development or diffusional exchange characteristics of the Igf2null(+/-) mouse. Hence, this study compares the structural properties (using stereology) and diffusional exchange characteristics (using measurement of permeability-surface area product, P.S, of three inert hydrophilic tracers) of the Igf2null(+/-) and the Igf2P0(+/-) placenta to identify the role of Igf2 in the development of the labyrinthine exchange membrane and its functional consequences. Our data show disproportionate effects of complete Igf2 ablation on the compartments of the placenta, not seen when the placental-specific transcript alone is deleted. Furthermore, although the theoretical diffusing capacity (calculated from the stereological data) of the Igf2null(+/-) placenta was reduced relative to control, there was no effect of the complete knockout on permeability surface area available for small hydrophilic tracers. This is in contrast to the Igf2P0(+/-) placenta, where theoretical diffusion capacity and P.S values were reduced similarly. Total ablation of the Igf2 gene from the fetoplacental unit in the mouse therefore results in a disproportionate growth of placental compartments whereas, deleting the placental specific transcript of Igf2 alone results in proportional placental growth restriction. Thus, placental phenotype depends on the degree of Igf2 gene ablation and the interplay between placental and fetal Igf2 in the mouse.

Oxidative stress and the induction of cyclooxygenase enzymes and apoptosis in the murine placenta.

Placental oxidative stress has been implicated in many complications of human pregnancy, including preterm delivery and preeclampsia. It is now appreciated that reactive oxygen species can induce a spectrum of changes, ranging from homeostatic induction of enzymes to apoptotic cell death. Little is known regarding the occurrence of placental oxidative stress in other species. We investigated markers of oxidative stress in the labyrinthine (LZ) and junctional (JZ) zones of the murine placenta across gestational age, and correlated these with expression of the cyclooxygenase enzymes COX-1 and COX-2, and apoptosis. We tested a causal link between the two by subjecting placental explants to hypoxia-reoxygenation (H/R) in vitro, a known stimulus for generation of oxidative stress. Western blotting demonstrated significant increases in the concentrations of hydroxynonenal (HNE), COX-1 and COX-2 with gestational age. Dual-labelling demonstrated co-localisation of HNE, and COX-1 and COX-2 within the trophoblast of the LZ, and glycogen cells of the JZ. An apoptotic index based on TUNEL-positivity demonstrated an increase with gestational age, and dual-labelling showed co-localisation of TUNEL labelling with HNE and active caspase-3 within the trophoblast of the LZ. H/R significantly increased oxidative stress, induction of COX-1 and COX-2, and the apoptotic index. Co-localisation demonstrated the increases in COX to be within the trophoblast of the LZ, and in particular the glycogen cells of the JZ. Apoptosis was restricted to the LZ. We speculate that the induction of COX enzymes is a physiological response to oxidative stress, and may play a role in initiating or augmenting parturition. Generation of oxidative stress may also play a role in influencing the growth trajectory of the placenta, and its component cell types. The mouse may provide an experimental genetic model in which to investigate these phenomena.

Origin and characteristics of glycogen cells in the developing murine placenta.

The junctional zone (Jz) of the mouse placenta consists of two main trophoblast populations, spongiotrophoblasts and glycogen cells (GCs), but the development and function of both cell types are unknown. We conducted a quantitative analysis of GC size, number, and invasion of cells into the decidua across gestation. Furthermore, we identified markers of GC function to investigate their possible roles in the placenta. While the spongiotrophoblast cell volume doubles, and cell number increases steadily from E12.5 to E16.5, there is a remarkable 80-fold increase in GC numbers. This finding is followed by a notable decrease by E18.5. Surprisingly, the accumulation of GCs in the decidua did not fully account for the decrease in GC number in the Jz, suggesting loss of GCs from the placenta. Glucagons were detected on GCs, suggesting a steady glucose release throughout gestation. Connexin31 staining was shown to be specific for GCs. GC migration and invasion may be facilitated by temporally regulated expression of matrix metalloproteinase 9 and the imprinted gene product, Decorin. Expression of the clearance receptor for type II insulin-like growth factor (IGF-II), IGF2R, in a short developmental window before E16.5 may be associated with regulating the growth effects of IGF-II from glycogen cells and/or labyrinthine trophoblast on the expansion of the Jz. Thus stereology and immunohistochemistry have provided useful insights into Jz development and function of the glycogen cells.

Analysis of mouse conceptuses with uniparental duplication/deficiency for distal chromosome 12: comparison with chromosome 12 uniparental disomy and implications for genomic imprinting.

Distal mouse chromosome 12 is imprinted. Phenotypic analysis of mouse embryos with maternal or paternal uniparental disomy for the whole of chromosome 12 has characterized the developmental defects associated with the altered dosage of imprinted genes on this chromosome. Here we conduct a characterization of maternal and paternal Dp(dist12) mice using the reciprocal translocation T(4;12)47H. This limits the region analysed to the chromosomal domain distal to the T47H breakpoint in B3 on mouse chromosome 12. Both MatDp(dist12)T47H and PatDp(dist12)T47H conceptuses are non-viable and the frequency of recovery of Dp(dist12) conceptuses by 10.5 days post coitum (dpc) was lower than expected after normal adjacent-1 disjunction. A subset of MatDp(dist12) embryos can survive up to one day post partum. In contrast to paternal uniparental disomy 12 embryos, no live PatDp (dist12) embryos were recovered after 16.5 days of gestation. Other phenotypes observed in maternal and paternal chromosome 12 uniparental disomy mice are recapitulated in the Dp(dist12) mice and include placental, muscle and skeletal defects. Additional defects were also noted in the skin of both MatDp(dist12) and maternal uniparental disomy 12 embryos. This study shows that the developmental abnormalities associated with the altered parent of origin for mouse chromosome 12 can be attributed to the genomic region distal to the T47H breakpoint.

Epigenetics and imprinting of the trophoblast -- a workshop report.

Genomic imprinting is a remarkable process that causes genes to be expressed or repressed depending on their parental-origin. Imprinted genes play important roles in prenatal growth and organ development. Postnatally, imprinted genes can contribute to the regulation of metabolic pathways and behaviour associated with the control of resources. One of the most important sites of imprinted gene action is the placenta. During this workshop at the 11th meeting of the International Federation of Placenta Associations/European Placenta Group held in Glasgow, a series of short talks were presented providing an overview of the evolution, function and mechanisms of imprinting in mammals with particular reference to the placenta. In addition, epigenetic control of trophoblast development and function were considered. This report summarises the contributions to the workshop.

Ultrastructural changes in the interhaemal membrane and junctional zone of the murine chorioallantoic placenta across gestation.

The mouse is an extremely useful experimental model for the study of human disease owing to the ease of genetic and physiological manipulation. A more detailed knowledge of murine placental development will, we hope, increase our understanding of the pathogenesis of placentally related complications of human pregnancy. The murine placenta consists of two main fetally derived compartments: the labyrinthine zone and the junctional zone. Exchange in the labyrinthine zone takes place across an interhaemal membrane comprising an outer layer of cytotrophoblast cells and two inner layers of syncytial trophoblast. The cytotrophoblast layer thins as gestation advances, and in addition becomes highly perforated after embryonic day (E)12.5. Furthermore, as gestation advances cytotrophoblast nuclear volume and DNA content increase, suggesting the formation of labyrinthine trophoblast giant cells. The syncytial layers become increasingly microvillous, enlarging the surface area for exchange. Separate basement membranes support the syncytium and the fetal capillary endothelium throughout gestation, although these appear to fuse where the capillaries are closely approximated to the trophoblast. The junctional zone consists of two principal trophoblast cell types, spongiotrophoblasts and invasive glycogen cells, yet the functions of each remain elusive. Spongiotrophoblasts vary in their appearance even when not fully differentiated, but a striking feature is the extensive endoplasmic reticulum of the more mature cells. Early glycogen cells are distinguished by the presence of electron-dense glycogen granules, and large amounts of surrounding extracellular matrix. Later the accumulations of glycogen granules occupy almost all the cytoplasm and there are few organelles. This is the first study to use both scanning and transmission electron microscopy in an ultrastructural description of murine placental development and is complementary to contemporary genetic investigations.

Imprinted genes in the placenta--a review.

Imprinted genes are expressed monoallelically depending on their parental origin. High expression of the majority of imprinted genes tested to date has been demonstrated in extraembryonic tissues; placenta and yolk sac. Several mouse models where specific imprinted genes have been disrupted demonstrate that fetal and placental growth may be regulated by imprinted genes, in which paternally expressed genes enhance, and maternally expressed genes restrain, growth. We review the current information on, and suggest possible functional roles for, imprinted genes in placental development.

Epigenetic alteration at the DLK1-GTL2 imprinted domain in human neoplasia: analysis of neuroblastoma, phaeochromocytoma and Wilms' tumour.

Epigenetic alterations in the 11p15.5 imprinted gene cluster are frequent in human cancers and are associated with disordered imprinting of insulin-like growth factor (IGF)2 and H19. Recently, an imprinted gene cluster at 14q32 has been defined and includes two closely linked but reciprocally imprinted genes, DLK1 and GTL2, that have similarities to IGF2 and H19, respectively. Both GTL2 and H19 are maternally expressed RNAs with no protein product and display paternal allele promoter region methylation, and DLK1 and IGF2 are both paternally expressed. To determine whether methylation alterations within the 14q32 imprinted domain occur in human tumorigenesis, we investigated the status of the GTL2 promoter differentially methylated region (DMR) in 20 neuroblastoma tumours, 20 phaeochromocytomas and, 40 Wilms' tumours. Hypermethylation of the GTL2 promoter DMR was detected in 25% of neuroblastomas, 10% of phaeochromocytoma and 2.5% of Wilms' tumours. Tumours with GTL2 promoter DMR hypermethylation also demonstrated hypermethylation at an upstream intergenic DMR thought to represent a germline imprinting control element. Analysis of neuroblastoma cell lines revealed that GTL2 DMR hypermethylation was associated with transcriptional repression of GTL2. These epigenetic findings are similar to those reported in Wilms' tumours in which H19 repression and DMR hypermethylation is associated with loss of imprinting (LOI, biallelic expression) of IGF2. However, a neuroblastoma cell line with hypermethylation of the GTL2 promoter and intergenic DMR did not show LOI of DLK1 and although treatment with a demethylating agent restored GTL2 expression and reduced DLK1 expression. As described for IGF2/H19, epigenetic changes at DLK1/GTL2 occur in human cancers. However, these changes are not associated with DLK1 LOI highlighting differences in the imprinting control mechanisms operating in the IGF2-H19 and DLK1-GTL2 domains. GTL2 promoter and intergenic DMR hypermethylation is associated with the loss of GTL2 expression and this may contribute to tumorigenesis in a subset of human cancers.

Placental-specific insulin-like growth factor 2 (Igf2) regulates the diffusional exchange characteristics of the mouse placenta.

Restricted fetal growth is associated with postnatal mortality and morbidity and may be directly related to alterations in the capacity of the placenta to supply nutrients. We proposed previously that imprinted genes can regulate nutrient supply by the placenta. Here, we tested the hypothesis that the insulin-like growth factor 2 gene (Igf2) transcribed from the placental-specific promoter (P0) regulates the development of the diffusional permeability properties of the mouse placenta. Using mice in which placental-specific Igf2 had been deleted (P0), we measured the transfer in vivo of three inert hydrophilic solutes of increasing size (14C-mannitol, 51CrEDTA, and 14C-inulin). At embryonic day 19, placental and fetal weights in P0 conceptuses were reduced to 66% and 76%, respectively, of wild type. In P0 mutants, the permeability.surface area product for the tracers at this stage of development was 68% of that of controls; this effect was independent of tracer size. Stereological analysis of histological sections revealed the surface area of the exchange barrier in the labyrinth of the mouse placenta to be reduced and thickness increased in P0 fetuses compared to wild type. As a result, the average theoretical diffusing capacity in P0 knockout placentas was dramatically reduced to 40% of that of wild-type placentas. These data show that placental Igf2 regulates the development of the diffusional exchange characteristics of the mouse placenta. This provides a mechanism for the role of imprinted genes in controlling placental nutrient supply and fetal growth. Altered placental Igf2 could be a cause of idiopathic intrauterine growth restriction in the human.

Comparative developmental anatomy of the murine and human definitive placentae.

The placenta of eutherian mammals is a remarkable biological structure. It is composed of both zygote-derived and maternal cells, and mediates the complex interactions between the mother and the fetus that are necessary for fetal growth and survival. While the genetic basis of human placental development and function is largely unknown, its understanding is of immense clinical importance because placentopathies of unknown genetic aetiology are thought to be the cause of many types of pregnancy complications including unexplained miscarriage and intrauterine growth retardation. The mouse is the best-studied mammalian experimental genetic model system and research is not restricted by the inherent ethical and practical limitations associated with the human. As a result, knowledge about the genetic control of mouse placental development has expanded greatly in recent years. In order for this to be of benefit to medical practice, extrapolations from murine to human placentation have to be made. However, comprehensive comparisons of the placentae of these two species are rare. This review therefore compares the developmental anatomy of the placenta between humans and mice with emphasis on structures and cell types that might be analogous between the two species. This could be of particular benefit to mouse developmental geneticists who study placental development and have an interest in the possible clinical implications of their work.

Comparative sequence analysis of the imprinted Dlk1-Gtl2 locus in three mammalian species reveals highly conserved genomic elements and refines comparison with the Igf2-H19 region.

The Dlk1-Gtl2 domain on mouse chromosome 12 contains reciprocally imprinted genes with the potential to contribute to our understanding of common features involved in imprinting control. We have sequenced this conserved region in the mouse and sheep and included the human sequence in a three species comparison. This analysis resulted in a precise conservation map and identification of highly conserved sequence elements, some of which we have shown previously to be differentially methylated in the mouse. Additionally, this analysis facilitated identification of a CpG-rich tandem repeat array located approximately 13-15 kb upstream of Gtl2. Furthermore, we have identified a third imprinted transcript that overlaps with the last Dlk1 exon in the mouse. This transcript lacks a conserved open reading frame and is probably generated by cleavage of extended Dlk1 transcripts. Because Dlk1 and Gtl2 share many of the imprinting properties of the well-characterized Igf2-H19 domain, it has been proposed that the two regions may be regulated in the same way. Comparative genomic examination of the two domains indicates that although there are similarities, other features are very different, including the location of conserved CTCF-binding sites, and the level of conservation at regulatory regions.

Dynamic temporal and spatial regulation of the cdk inhibitor p57(kip2) during embryo morphogenesis.

The complete developmental expression pattern of the cyclin dependent kinase inhibitor (CDKI) p57(kip2) has not been reported, here we report a detailed study of the localization of p57(kip2) protein during mouse organogenesis. We show that p57(kip2) is coincident with key stages of differentiation of several organs, some but not all of which are affected in Beckwith-Weidermann syndrome, a human congenital syndrome characterized by foetal overgrowth and childhood tumours.

DNA methylation in genomic imprinting, development, and disease.

Changes in DNA methylation profiles are common features of development and in a number of human diseases, such as cancer and imprinting disorders like Beckwith-Wiedemann and Prader-Willi/Angelman syndromes. This suggests that DNA methylation is required for proper gene regulation during development and in differentiated tissues and has clinical relevance. DNA methylation is also involved in X-chromosome inactivation and the allele-specific silencing of imprinted genes. This review describes possible mechanisms by which DNA methylation can regulate gene expression, using imprinted genes as examples. The molecular basis of methylation-mediated gene regulation is related to changes in chromatin structure and appears to be similar for both imprinted and biallelically expressed genes.