T-cell identity and epigenetic memory

Abstract

T-cell development endows cells with a flexible range of effector differentiation options, superimposed on a stable core of lineage-specific gene expression that is maintained while access to alternative hematopoietic lineages is permanently renounced. This combination of features could be explained by environmentally responsive transcription factor mobilization overlaying an epigenetically stabilized base gene expression state. For example, "poising" of promoters could offer preferential access to T-cell genes, while repressive histone modifications and DNA methylation of non-T regulatory genes could be responsible for keeping non-T developmental options closed. Here, we critically review the evidence for the actual deployment of epigenetic marking to support the stable aspects of T-cell identity. Much of epigenetic marking is dynamically maintained or subject to rapid modification by local action of transcription factors. Repressive histone marks are used in gene-specific ways that do not fit a simple, developmental lineage-exclusion hierarchy. We argue that epigenetic analysis may achieve its greatest impact for illuminating regulatory biology when it is used to locate cis-regulatory elements by catching them in the act of mediating regulatory change.

Figure 1

Examples of histone marking patterns in neighborhoods of strongly-expressed T-cell genes. (A) ZAP70, in human CD4⁺ T cells (data from (Barski et al. 2007)). (B) Gata3, in four subsets of murine CD4⁺ T cells: 1, naïve T cells; 2, Th1 T cells; 3, Th2 T cells, and 4, “natural” regulatory T cells (nTregs)(data from (Wei et al. 2009)). Plots show enrichments of chromatin immune precipitation with antibodies against the indicated modified histones and RNA polymerase II, displayed as histograms against the human (hg18, panel A) and murine (mm9, panel B) genomes on the UCSC genome browser (http://genome.ucsc.edu). WIG files for (A) were downloaded from http://dir.nhlbi.nih.gov/papers/lmi/epigenomes/hgtcell.html. WIG files for (B) were constructed from raw data deposited in a public repository by Wei et al., and converted from mm8 to mm9 coordinates before plotting. H2A.Z, H3K4me1, H3K4me2, and H3K4me3 are all markers for accessibility and/or activation. H3K27me3, H3K9me2, H3K9me3, and H4K20me3 are all implicated in repression. H3K79me3 is associated with actively elongating polymerase. The repression mark H3K27me3 and the activation mark H3K4me3 are particularly useful, as shown in (B). Magenta arrows indicate the genes of interest and their directions of transcription. In (B), blue arrow shows site of upstream regulatory element for Gata3 that is selectively demethylated in Th2 cells.

Figure 2

Distinct modes of repression of lineage-inappropriate genes in T cells. (A) GATA2 in human CD4⁺ T cells: an example of “broad” H3K27me3-mediated repression of a gene that was expressed in the earliest precursor stages. (B) SPI1 in human CD4⁺ T cells: an example of repression without any repressive marks. Note the lack of any of the marks, H3K27me3, H3K9me2, H3K9me3, and H4K20me3, which are thought to mediate repression. (C) Pax5 in four subsets of murine CD4⁺ T cells: an example of “broad” H3K27me3-mediated repression of a gene which may never have been expressed in T-cell precursors. (D) Bcl11a in four subsets of murine CD4⁺ T cells: an example of promoter-associated H3K27me3-mediated repression of a gene that is expressed in T cells until DN3 stage. In (C, D): 1, naïve T cells; 2, Th1 T cells; 3, Th2 T cells, and 4, nTregs. Magenta arrows indicate the genes of interest and their directions of transcription. Data for (A, B) were from Barski et al. 2007, as in Fig. 1A. Data for (C, D) were from Wei et al. 2009, as in Fig. 1B.