Genes and Immunity (2012) 13, 388–398 & 2012 Macmillan Publishers Limited All rights reserved 1466-4879/12 www.nature.com/gene ORIGINAL ARTICLE Genome-scale profiling reveals a subset of genes regulated by DNA methylation that program somatic T-cell phenotypes in humans D Martino1,2, J Maksimovic1, J-HE Joo1, SL Prescott2,3 and R Saffery1,3 The aim of this study was to investigate the dynamics and relationship between DNA methylation and gene expression during early T-cell development. Mononuclear cells were collected at birth and at 12 months from 60 infants and were either activated with anti-CD3 for 24 h or cultured in media alone, and the CD4 þ T-cell subset purified. DNA and RNA were co-harvested and DNA methylation was measured in 450 000 CpG sites in parallel with expression measurements taken from 25 000 genes. In unstimulated cells, we found that a subset of 1188 differentially methylated loci were associated with a change in expression in 599 genes (adjusted P valueo0.01, b-fold 40.1). These genes were enriched in reprogramming regions of the genome known to control pluripotency. In contrast, over 630 genes were induced following low-level T-cell activation, but this was not associated with any significant change in DNA methylation. We conclude that DNA methylation is dynamic during early T-cell development, and has a role in the consolidation of T-cell-specific gene expression. During the early phase of clonal expansion, DNA methylation is stable and therefore appears to be of limited importance in short-term T-cell responsiveness. Genes and Immunity (2012) 13, 388–398; doi:10.1038/gene.2012.7; published online 12 April 2012 Keywords: T-cell epigenetics; immune epigenetics; DNA methylation; gene expression; T-cell development; reprogramming differentially methylated region INTRODUCTION cytokines, thus remaining ‘poised’ for commitment.5,14 Although Shortly after birth, there are rapid phenotypic and functional these observations have contributed to our understanding of the changes in both innate and adaptive immunity. This critical period mechanisms that govern T-cell plasticity and lineage commitment, of early immune programming is not only important for they have largely been observed under highly polarizing 15–18 establishing normal patterns of immunity, but also represents a experimental conditions, and studied for a restricted period of heightened susceptibility to various immune disorders. number of important cytokine gene loci. Therefore, a more In the adaptive immune compartment, there is a developmental complete picture of the epigenetic processes that govern normal transition as ‘less mature’ T-cells emerging from the thymus T-cell development is warranted, which has only recently been undergo maturation in the periphery.1 This transition becomes possible with the advent of genome-wide technologies. apparent over the first year of life;2 however, the molecular In the current study, we investigated the role of DNA processes that drive this are poorly understood. Understanding methylation and its association with patterns of gene expression these processes is critical, as disruption in these pathways may under two scenarios: (1) during the steady-state development of alter the normal course of T-cell development, and potentially naive CD4 þ T-cells shortly after birth; (2) following the activation program susceptibility may lead to a range of allergic and of T-cells during entry into the cell cycle. Our data provide insights autoimmune diseases.3,4 into the molecular pathways of early T-cell programming, and Epigenetic modifications are likely to mediate early develop- characterize developmental pathways potentially susceptible to mental changes in T-cells, because these modifications are known disruption through early-life environmental exposures. to have a well-defined role in determining both the diversity and plasticity of T-helper cell phenotypes.5–7 Variability in DNA methylation levels and histone modification profiles establishes RESULTS 8–11 active or repressive states of transcription at key cytokine loci. Combined genome-wide DNA methylation and gene expression In differentiated CD4 þ T-cells, these mechanisms are responsible analysis reveals a dynamic genomic program during steady-state for the somatic heritability of differentiated T-cell states, and T-cell development are described in close association with the acquisition of Neonatal CD4 þ T-cells are phenotypically and functionally effector phenotypes, and specialized patterns of cytokine unique compared with later ages. To investigate the epigenetic 11–13 gene expression. In undifferentiated (naive) T-cells, DNA differences between these cell types, we compared DNA methylation marks maintain the plasticity of CD4 þ T-cells, methylation and gene expression in neonatal CD4 þ cells with because these cells express low levels of a broad range of their 12-month counterparts under two conditions. The 1Cancer, Disease and Developmental Epigenetics, Murdoch Children’s Research Institute, Royal Melbourne Hospital, Parkville, Victoria, Australia and 2School of Paediatrics and Child Health, University of Western Australia, Perth, Western Australia, Australia. Correspondence: Dr R Saffery, Cancer, Disease and Developmental Epigenetics, Murdoch Children’s Research Institute, Royal Melbourne Hospital, Flemington Road, Parkville, Victoria 3052, Australia. E-mail: [email protected] 3Equal senior authors. Received 24 January 2012; revised 28 February 2012; accepted 29 February 2012; published online 12 April 2012 Epigenetic programming of CD4 þ T-cell phenotypes during early life D Martino et al 389 Figure 1. Experimental methodology and validation of in vitro protocol. (a) Experimental approach used in this study. (b) Experimental design for methylation comparisons between fresh and 24-h cultured CD4 þ cells. (c) Matrix scatterplot of fresh versus 24-h cord blood and adult blood samples. The figure shows scatterplot comparisons between all samples. MDS, multidimensional scaling plot. experimental strategy is outlined in Figure 1a. Briefly, mono- 1471 probes (31.9%) had no associated gene annotation and were nuclear cells collected from the same infants at birth and located either in intergenic regions or in regions occupied by two 12months were cultured with and without anti-CD3 and IL-2 for or more refseq transcripts. A total of 3224 CpGs (70%) showed 24 h, in a 371 incubator maintained at 5% CO2. After this time, increased methylation between neonatal and 12-month CD4 þ media supernatants were reserved for cytokine analysis and cells, and 1383 CpGs (30%) showed reduced methylation, CD4 þ cells were purified by magnetic bead sorting. DNA and indicative of dynamic changes in DNA methylation during early total RNA were co-harvested for microarray analysis. Studies of development. normal development were conducted in resting (unstimulated) Unsupervised sample clustering based on the 4607 differentially cells, and studies of T-cell activation were conducted in anti-CD3- methylated probes correctly discriminated neonatal from treated cultures. 12-month samples (Supplementary Figure 1). We performed To validate the experimental approach, we first needed to ontology enrichment analysis on the gene-associated probes determine whether T-cells rested in culture for 24 h undergo any and identified terms associated with gene expression, RNA non-physiological changes in DNA methylation that may compli- polymerase II activity and transcription. Alongside these, we also cate data interpretation. To address this, we performed a small observed a host of developmental terms, including cell and tissue pilot experiment detailed in Materials and methods (Figure 1b). morphogenesis, mesenchymal differentiation, skeletal muscle, and We compared freshly thawed T cells with cultured Tcells genome- olfactory and neuronal development (Supplementary Table 1). wide, using exploratory techniques and probe-wise tests for Examples of genes associated with these terms include the differential methylation (adjusted P-valueo0.05 and b-fold myosins (MYO1D, MYOIC), myosin light-chain kinases (MYLK), change40.1). We found no changes in DNA methylation profiles olfactory receptor family members (ORS1E2, OR4D2) and neuronal between fresh and cultured CD4 þ cells, for neonatal or adult peptides (NRP2, NRTN). Epigenetic changes at these loci are likely samples (Figure 1c). This demonstrated that short-term cell culture to reflect the developmental control of gene expression during does not distort the physiological patterns of genomic lineage commitment.19,20 Several immunological terms were also methylation. enriched in the list of differentially methylated genes, and these To gain a broad picture of the extent to which neonatal CD4 þ included antigen processing and presentation, immune response, cells are developmentally different from their 12-month counter- leukocyte activation, protein kinase signaling, TGFb signaling and parts, we compared unstimulated cultures. Probe-wise compar- MAPK signaling (Table 1). isons of DNA methylation between neonatal and 12-month CD4 þ In the gene expression data set, we observed 986 probes that cells identified a total of 4607 differentially methylated CpG sites varied significantly between unstimulated neonatal and 12-month (adjusted P valueo0.01 and b-fold change40.10), of which 3136 T cells (adjusted P-valueo0.01, b-fold change42). This consti- sites mapped to 1826 unique genes (Figure 2a). The remaining tuted 287 (29%) upregulated probes and 699 (71%) & 2012