SEARCH

SEARCH BY CITATION

Keywords:

  • CD4/helper T-cells (Th cells Th0, Th1, Th2, Th17);
  • regulatory T (Treg) cells;
  • enhancers;
  • master regulator transcription factors;
  • environmental response factors;
  • GATA3

Summary

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

Mature naive CD4 T-cells possess the potential for an array of highly specialized functions, from inflammatory to potently suppressive. This potential is encoded in regulatory DNA elements and is fulfilled through modification of chromatin and selective activation by the collaborative function of diverse transcription factors in response to environmental cues. The mechanisms and strategies employed by transcription factors for the programming of CD4 T-cell subsets will be discussed. In particular, the focus will be on co-operative activity of environmental response factors in the initial activation of regulatory DNA elements and chromatin alteration, and the subsequent role of ‘master regulator’ transcription factors in defining the fidelity and environmental responsiveness of different CD4 T-cell subsets.


Introduction: CD4 T-cells, in context

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

Mature naive CD4 T-cells, when poised for effector differentiation, are near their final destination following a long developmental journey. Mesoderm-derived haemangioblasts – the multipotent progenitors of both endothelial cells and haematopoietic cells – develop into the embryonic haemogenic endothelial cells of the dorsal aorta. Definitive haematopoietic stem cells derived from this diminutive tissue go on to seed the fetal liver and eventually the adult bone marrow. These self-renewing haematopoietic stem cells differentiate into the common myeloid and common lymphoid progenitor cells that form the basis for the plethora of devoted immune cell lineages, including CD4 T-cells. Along this broad spectrum of differentiation – from germ layers to T-cell subsets – a number of mechanistic strategies are employed to access new developmental potential while restricting alternative fates.

Conrad Waddington (1905–1975) considerably progressed thinking on cellular differentiation by proposing that genes (and mutations) can affect differentiation potential. He visualized this concept as a marble rolling through an ‘epigenetic landscape’, shaped by the action of genes, with ridges and valleys representing irreversible developmental commitment and future potential (Fig. 2, reviewed in ref. [1]). Spatial and temporal control of gene expression creates this ‘epigenetic landscape’ and instructs diverse cellular differentiation from a single common genome. Mechanisms controlling varied gene expression can include instructive morphogen gradients, asymmetric cell division, and natural distributions or stochastic action of signalling, nuclear, or chromatin-associated factors (gene expression noise[2]) together with feedback and ‘feedforward’ transcriptional networks. Notably, both Caenorhabditis elegans (~ 1000 cells) and humans (> 1012 cells) encode 20 000–25 000 protein-coding genes. As such, the non-coding regulatory component of the genome (~ 9·7 × 107 base pairs in C. elegans, and 3 × 109 in humans) is an appealing environment for integrating signals into spatio-temporal and cell-type-specific gene expression patterns to confer diverse cellular function.[3] Chromatin accessibility at non-coding DNA—namely, proximal promoter sequences—was described first by Carl Wu[4] in 1980 and was suggested to facilitate recruitment of factors that regulate gene activity. Contemporary understanding of mammalian regulatory DNA elements places the majority at intronic or intergenic regions. However, unlike promoter studies, a major challenge of approaching the possibility of regulatory function in such distal DNA elements was determining where to look. Based on the observation that transcription only occurs at rearranged immunoglobulin heavy chain (Igh) genes, and never at non-rearranged genes, Susumu Tonegawa, Walter Schaffner and colleagues hypothesized that rearrangement brought downstream regulatory DNA into proximity with the promoter sequence to enhance transcription. Indeed, in 1983, they described a downstream endogenous enhancer element in the Igh gene that was active in a tissue-specific manner – in B cells, not in HeLa cells or fibroblasts.[5, 6] Recent advances in high-throughput sequencing technologies have improved our capacity to study and appreciate the role of the regulatory genome in controlling differentiation and cellular diversity. For example, mapping of chromatin accessibility and transcription factor binding sites demonstrates that ~ 1–2% of the genome is accessed as regulatory DNA in a given cell type. The cell-type-specific and largely non-overlapping nature of the regulatory DNA suggests that a substantial amount of intergenic sequence could encode regulatory information.[7] New genomic experimental approaches allow for incisive study of the role of this extensive regulatory DNA landscape in cellular differentiation.

Differentiation of T helper (Th) and regulatory T (Treg) cells from mature CD4 T-cells represents relatively late-stage differentiation. Although these cells can be considered close relatives, their faithful differentiation and phenotypic stability are critical, as their dysregulation can result in a broad spectrum of diseases, from autoimmunity to immunodeficiency. Th and Treg cell states are defined by expression of master regulator transcription factors [GATA binding protein 3 (GATA3), T-box 21 (TBET), RAR-related orphan receptor γ(RORγt) and Forkhead box P3 (FOXP3)] and associated phenotypic characteristics such as participation in particular types of inflammatory responses or the suppression of immune cell activation. Appropriate lineage stability or plasticity is encoded in the mechanisms instructing and maintaining the Th/Treg lineage-specific transcriptional programmes. Here, I will discuss several recent studies describing the role of key transcription factors in the de novo activation of regulatory DNA elements affecting lineage-defining transcriptional programmes. These studies underscore, quantitatively, the dominance and importance of signal-activated transcription factors downstream of T-cell receptor (TCR) signalling and cytokine receptor signalling in initiation of T-cell polarization. Further, they reflect how co-operative binding of transcription factors to combinatorial motifs across the genome is a common strategy for the activation of lineage-specific enhancers.

Master regulators and pioneer factors

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

Treatment of fibroblasts with the DNA methyltransferase inhibitor 5-azacytodine results in de-repression of a number of genes and their conversion to myoblasts. Davis, Weintraub and Lassar discovered myogenic differentiation 1 (MYOD) to be highly induced under these conditions and went on to show its sufficiency for myogenesis in a number of cell types.[8] Since this discovery, a number of ‘master regulator’ transcription factors have been described, with the notable characteristic that their expression in immediate precursor cells (and sometimes alternative lineages, in so-called ‘transdifferentiation’) is necessary and ‘sufficient’ for differentiation and acquisition of distinctive cell-type-specific characteristics. Genomic approaches allow for the study of the global activity of such transcription factors. For example, MYOD functions in the global de novo activation of enhancers involved in muscle growth and differentiation; MYOD is required for acquisition of chromatin characteristics associated with active enhancers: monomethylation of histone 3, lysine 4 (H3K4me1), recruitment of PolII and the histone acetyltransferase, p300, and histone acetylation (characteristically of H3K27).[9]

The ability of ‘master regulator’ transcription factors to “open” and activate latent lineage-specific regulatory DNA is intuitive and appealing in its simplicity – it represents a single-step mechanism for the extraction of information from dispersed regulatory DNA and its use in the control of cell-type-specific transcription. Enhancer activation typically progresses from transcription factor binding at specific DNA motifs to recruitment of ‘co-activators’ – histone and chromatin modifying factors such as the SWItch/Sucrose Non-Fermentable chromatin remodelling complex and histone-modifying enzymes, like p300 – and the recruitment of general transcription factors and PolII, often with physical interaction with the associated gene promoter.[10, 11] Several studies suggest that complex and incremental control of regulatory elements and their chromatin states by sequentially and co-operatively acting transcription factors underlies the progressive alteration of enhancer states through differentiation.[3, 12-15] However, some factors—definitive ‘pioneer factors’—have the capacity to bind to nucleosomal DNA or higher-order chromatin and establish enhancer accessibility and responsiveness to subsequent binding of other factors. FOXA factors are examples of this class of transcription factor, with essential roles in early embryonic development and organogenesis. The FOXA1 DNA-binding domain structurally mimics the linker histone, H1, and stably binds to nucleosomal DNA, probably through interactions with the core histones, H3 and H4. These characteristics are associated with slow nuclear diffusion, abundant non-specific nucleosomal interactions, and stable binding at some Forkhead recognition motifs followed by nucleosome displacement and accessibility of surrounding regulatory DNA to other transcription factors.[16, 17]

Although the critical functions of Th cell master regulator transcription factors TBET and GATA3 have been well established for over a decade,[18-20] mechanistic insights and global, genomic characterization have been recent. How do Th cell master regulator transcription factors function and how extensive is their transcriptional and regulatory footprint? What are their roles in de novo enhancer activation and gene expression? Through what mechanisms do they modulate the activity of the regulatory elements that they bind – as bona fide pioneer factors displacing nucleosomes, through co-operative binding with other factors, or through binding to previously accessible, poised elements? Early studies demonstrated the sufficiency of over-expressed TBET and GATA3 to induce DNase I accessibility and transcription at the interferon-γ (Ifng) and Th2 cytokine loci, respectively, and suggested their role in regulation of chromatin. In some cases this activity was shown to be independent of signals from cytokine receptors and downstream signal transducer and activator of transcription (STAT) factors or despite alternative lineage cytokine stimulation.[18, 19, 21-23] Loss of function studies established a requirement for these factors in Th differentiation in vivo.[20, 24] Importantly, these studies focused exclusively on small sets of signature Th1 and Th2 genes, usually the respective cytokine gene loci, and clearly established the important role of TBET and GATA3 in their regulation. Subsequently, master regulators were described for Treg (FOXP3) and Th17 (RORγt) cells and shown to be critical for differentiation and acquisition of their respective T-cell lineage transcriptional programmes and phenotypes.[25-29] Their defining roles in CD4 T-cell subset differentiation and requirement for signature gene expression, analogous to classical master regulator transcription factor function, implied that Th master regulator transcription factors act as pioneer factors in the nucleation of de novo enhancer accessibility and activation. Recent studies suggest a model (Figs 1 and 2) that contrasts with this view, in which master regulators have limited footprints and act through collaboration with signal-activated environmental response factors.[12-14, 30, 31]

image

Figure 1. CD4 T-cell polarization: environmental response factors (ERFs) control initial enhancer and gene expression activation, which is then modulated and stabilized by master regulator factors (MRFs). (a) Sensing of the extracellular environment through T-cell receptor (TCR) and cytokine receptor signalling activates ERFs that act co-operatively in the de novo activation of enhancers and gene expression (step 1). For example, activating protein 1 (AP-1) and interferon regulatory factor (IRF) bind co-operatively to AP-1/IRF composite elements (AICE), often together with signal transducer and activator of transcription (STAT) factors, recruit co-activators, and activate regulatory elements and associated gene expression. A small subset of ERF-regulated genes appear to be repressed (~ 10%). The transcriptional programme initiated in step 1 is broadly inclusive (Th0). Subsequently, in step 2, MRFs, which are induced by ERFs, act mostly on the previously conditioned regulatory elements to appropriately tailor their respective T-cell subset transcriptional programmes by stabilizing and amplifying the expression of lineage-specific genes and repressing alternative lineage genes. In T helper type 1 (Th1), Th2 and regulatory T (Treg) cells, lineage stability is controlled in part through MRF engagement of autoregulatory positive feedback loops. (b) Chromatin at regulatory elements is highly modified by ERF binding and co-activator recruitment. Inactive enhancers may either be poised with pioneer factor binding associated with moderate accessibility for subsequent ERF binding and low level H3K4me1, or inaccessible with nucleosomal transcription factor binding sites. Inaccessible regulatory elements probably require additional co-factors for co-operative ERF binding and activation. ERFs, namely AP-1, IRFs, and STATs (and additional co-factors, such as nuclear factor of activated T-cell, NFAT) recruit co-activators such as chromatin remodellers, histone acetyltransferases, like p300, and enzymes responsible for H3K4me1 (lysine methyltransferases, KMT). These ERF-recruited chromatin-modifying factors act to increase accessibility and nucleosomal depletion of regulatory elements, and to post-translationally modify flanking histones (acetylation of the H3 and H4 N-terminal tails, H3K4me1, and others). These features enable transcription of associated genes and subsequent MRF binding.

Download figure to PowerPoint

image

Figure 2. Waddington's Epigenetic Landscape for CD4 T-cells: Pioneer factors, environmental response factors (ERFs), and master regulator factors (MRFs) regulate chromatin state and control T-cell subset developmental restriction, potential, and stability. Early acting, classic MRFs and pioneer factors, such as FOXA, GATA1, and MYOD, instruct early cell lineage identity through chromatin alterations and the regulation of target gene transcription. Later, following T-cell receptor (TCR) and cytokine receptor signalling, ERFs similarly alter the existing chromatin state and transcription at target sites. In contrast, CD4 T-cell MRFs—GATA3, FOXP3, TBET and RORγt—predominantly associate with regulatory elements within regions of previously activated chromatin, where ERF co-factors are pre-bound or bind co-operatively. MRFs may define CD4 T-cell subset plasticity or stability, identity, and heritable phenotypes through participation in networks with stabilizing positive feedback loops and through the maintenance of chromatin states established by ERFs at lineage-specific regulatory elements.

Download figure to PowerPoint

A genomics era reinterpretation of T-cell master regulators

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

GATA3, TBET, FOXP3 and RORγt all have essential roles in regulating the phenotype of their respective CD4 T-cell subtypes. Three recent studies (described in detail below) characterized the relative contribution of these four transcription factors in the activation and function of lineage-specific regulatory DNA, or enhancers.[12-14] Surprisingly, despite differing approaches, all three studies demonstrated a quantitatively minor role for these four MRFs in the de novo activation of lineage-specific enhancers. In the two general models for T-cell lineage enhancer activation tested by these studies, the first step is the same: the ‘right’ combinations of environmentally activated or induced transcription factors – environmental response factors (ERFs) such as STATs, interferon regulatory factors (IRFs), activated protein 1 (AP-1), nuclear factor of activated T-cell (NFAT) and nuclear factor κB (NF-κB) – bind to, and initiate expression of, master regulator factors (MRF) – Tbx21, Gata3, Rorc, Foxp3. Simultaneously these ERFs activate a set of general activation response (Th0) regulatory DNA elements, and a subset of lineage-specific (for example Th1- or Th2-specific) regulatory elements. In the second step, the MRFs either co-ordinate de novo activation of remaining lineage-specific regulatory DNA allowing binding of ERFs (perhaps acting in a second wave), or alternatively, they mainly bind to enhancers previously activated by ERFs. The critical distinction between these models is whether MRFs pioneer the activation of lineage-specific regulatory elements, or bind to regulatory elements pre-activated by ERFs. Based on recent studies, it appears that most lineage-specific enhancers are initially activated by ERFs or other nuclear factors expressed and functioning before the induced expression of MRFs. In particular, STATs, IRFs and AP-1 factors acting co-operatively have a prominent role in the activation of T-cell subset enhancers.

Environmental response factors and enhancer activation

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

To determine the relative contributions of STATs and MRFs, O'Shea and colleagues extensively characterized the enhancers of in vitro differentiated Th1 and Th2 cells with and without the respective STATs and MRFs.[13] One exciting observation from this study was the uniqueness of the Th1-activated and Th2-activated enhancer landscapes. Just over half of all active enhancers in Th1 and Th2 cells, characterized by both H3K4me1 and p300 binding, were shared between the two lineages Considering how closely related Th1 and Th2 cells are in the context of expansive cellular diversity (and considering these particular cells derived from a homogeneous population of naive CD4 T-cells before TCR and cytokine driven in vitro differentiation), this extent of dissimilarity in their enhancer landscapes is interesting and suggests broad functional divergence and responsiveness. The likely explanation for this discrete enhancer repertoire is that differential activation of ERFs between the two lineages plays an extensive role in the activation of enhancers. Indeed, lineage-specific enhancer motif analysis and chromatin immunoprecipitation sequencing experiments for key transcription factors involved in Th1 and Th2 differentiation revealed a dominant role for STATs compared with MRFs in this lineage-specific enhancer activation. Analysis of the repertoire and characteristics of Th1 enhancers in the absence of STAT1 or STAT4 revealed these interleukin-12 (IL-12) and interferon-γ cytokine receptor-activated ERFs to be required for almost 60% of Th1 enhancer activation. Notably, while TBET regulated the expression of a number of Th1 genes, the levels of p300 at associated enhancers were largely independent of TBET. However, 17% of Th1 enhancer activation (p300 recruitment) was dependent on TBET. These data raise interesting questions about TBET's mechanism of action at target regulatory DNA. Elegant studies from Weinmann and colleagues have demonstrated the potential for TBET to act through at least two separable mechanisms mapped to distinct protein domains – recruitment of an H3K4me2 methyltransferase and direct transactivation.[32] Therefore, it will be interesting to determine if those few Th1 enhancers that require TBET for activation rely primarily on the chromatin-modifying potential of TBET, whereas the genes whose expression is augmented by TBET, independent of extensive modification of enhancer characteristics, rely more heavily on the transactivation domain and increased recruitment of the general transcription machinery.

As in Th1 cells, it appears that Th2 cell enhancer activation is heavily reliant on ERFs, namely STAT6 downstream of IL-4R signalling. STAT6 was required for the activation of 77% of all Th2-specific enhancers.[13] Although, like TBET, GATA3 plays a minor role in enhancer activation, when over-expressed, it is sufficient for enhancer activation at about half of STAT6-dependent enhancers. In this context, it is interesting to consider potential GATA3 dosage effects in chromatin regulation and target gene expression, and the possibility for GATA3 to function as a ‘pioneer’-like factor in some settings. In fact, during early T-cell development, GATA3 and PU.1 binding can precede full enhancer activation and gene expression in developing thymocytes.[33] However, during the initial events of Th cell polarization, GATA3 and TBET play a less substantial role in nucleating chromatin alterations, activating enhancers, and influencing gene expression compared with STATs. Although representing a minority, it will be interesting to better understand the enhancers and genes dependent on MRFs for activation, both in terms of their potentially distinct chromatin characteristics and functional roles.

Considering the relative function of ERFs and MRFs in Th cell differentiation, a study from Littman and colleagues thoroughly explored the transcriptional programme of Th17 cells as defined by five key transcription factors: basic leucine zipper transcription factor (BATF), IRF4, STAT3, cellular musculoaponeurotic fibrosarcoma oncogene homolog (cMAF) and RORγt.[12] Surprisingly, the authors find that Th17 master regulator, RORγt, binds almost exclusively to sites co-bound by all four other transcription factors, minimally effects p300 binding and chromatin accessibility of binding sites, and while it regulates many genes positively, it does so modestly, effecting only a few key Th17 genes more than twofold (when compared to RORγt-deficient cells; Fig. 1). In contrast, BATF and IRF4, binding co-operatively, as well as STAT3, were found to have pioneer-like function. Indeed, these factors were primarily responsible for Th17 cell enhancer activation as measured by p300 recruitment and increases in accessibility. Another study from Regev and colleagues provides additional details of Th17 cell transcriptional kinetics.[34] Th17 cell differentiation proceeds in three distinguishable stages, termed induction (within 4 hr), onset of phenotype and amplification (4–20 hr), and stabilization and IL-23 signalling (20–72 hr). Several factors act throughout these stages, including BATF, IRF4 and STAT3, but others are restricted in their activity to either the early induction stage (including several STAT and IRF factors) or the late, stabilization stage (for example, RORγt). Consistent with early activity of ERFs in establishing chromatin states and initializing transcriptional programmes, and late stabilizing activity of MRFs, STAT1 and IRF1 target gene binding and activity predominate early (along with the core factors BATF, IRF4 and STAT3), whereas RORγt binding and regulatory activity occur during stabilization and at sites previously occupied by other core factors.[34] Therefore, as in Th1 and Th2 cell differentiation, ERFs – notably, STAT1, STAT3, IRF4, AP-1 – play dominant roles in Th17 cell enhancer activation with the MRF, RORγt, subsequently binding to augment and stabilize gene expression.

Like Th cells, Treg cells can differentiate from mature naive T-cells with distinct environmental cues converging to induce the expression of sets of genes and the MRF, FOXP3, for instruction of the Treg cell phenotype and function.[29, 35] While FOXP3 has been shown to be necessary and sufficient for Treg cell differentiation and function, questions remain about its mechanism of action in regulating the Treg cell transcriptional programme. To address this, Rudensky and colleagues used combinations of DNase I hypersensitive site sequencing (DNase-seq) and transcription factor chromatin immunoprecipitation sequencing to ask if FOXP3 bound to inaccessible chromatin as a pioneer-like factor, initiating remodelling and regulatory element activation, or if it bound to previously accessible regulatory elements to modulate their activity. The vast majority of FOXP3 binding sites were found to be accessible in Treg cell precursors, before FOXP3 expression and binding, and were pre-bound by other transcription factors, notably E26 avian leukemia oncogene (ETS1), runt-related transcription factor 1 (RUNX1) and Forkhead box O1 (FOXO1), which acts as a FOX-family placeholder for FOXP3 and also functions independently to augment the Treg cell phenotype.[14, 36] A small set of seemingly FOXP3-activated, Treg-cell-specific enhancers existed, but even these were recapitulated in FOXP3-negative cells upon activation and were enriched for motifs of the TCR activated transcription factors, AP-1 and NFAT.[14] Therefore, as with GATA3, TBET and RORγt, FOXP3 has a minimal role in the de novo activation of enhancers during differentiation, and instead functions subsequently, binding to previously active regulatory elements to augment or tune activity.

The study by Rudensky and colleagues also reveals an extensive collection of regulatory DNA elements in ex vivo isolated, mature, unstimulated CD4 T-cells. Almost 6000 uniquely accessible chromatin sites were present in mature naive CD4 T-cells, compared with B cells. This array of DNase I hypersensitive sites probably represents poised or active regulatory elements and may reflect the differentiation potential of these cells (almost all of these were shared with Treg cell DNase I hypersensitive sites).[14] Certainly, in the context of T-cell activation, AP-1, NFAT, IRF4 and other TCR-activated or induced transcription factors have essential roles in de novo accessibility and activation of regulatory elements. However, while these recent studies expose the activity of several transcription factors in the activation of Th-cell-specific enhancers (previously inactive or poised in naive CD4 T-cells), the factors responsible for poising the enhancer landscape that exists in naive CD4 T-cells during thymocyte differentiation are largely unknown. Although a number of transcription factors are critical for thymocyte development (PU.1, NOTCH, GATA3, E2A, TCF-1, LEF-1, RUNX1, etc),[33] those responsible for the de novo accessibility and heritable maintenance of poised or active enhancer states are not well understood. Such factors could function analogously to PU.1 and C/EBP in myeloid cells and PU.1, EBF and E2A in early B-cell differentiation – binding co-operatively to lineage-specific enhancers to mediate de novo chromatin remodelling and acquisition of H3K4me1 on enhancer-flanking nucleosomes.[37-39] Notably these studies found AP-1 motif enrichment at a portion of lineage-specific enhancers, and AP-1 and NFAT motifs were also enriched among enhancers activated during Th cell polarization without Th1 or Th2 bias.[13] Furthermore, activation of a subset of MYOD enhancers appears to be dependent on AP-1; knockdown of c-Jun resulted in reduced H3K4me1 and H3K27ac at AP-1 and MYOD co-bound enhancers.[9] It is intriguing to consider then that both MRFs (MYOD) and ERFs (IRFs and STATs) could engage AP-1 as a common factor involved in de novo enhancer activation.

Given its broad expression, what determines the activity of AP-1 in a given cell type? Several recent studies have characterized co-operative binding of AP-1 with IRF4 and IRF8.[12, 30, 31, 40, 41] Together with the capacity for IRFs to bind with Ets-family members (such as PU.1) at ETS/IRF composite elements (EICE), description of AP-1/IRF composite elements (AICE) reveals how these factors function together to bind distinct elements co-operatively, and may explain some of their distinct functions in T-cell subsets, dendritic cells and B cells. At AICE, combinatorial integration is possible through both varied AP-1 dimer composition and choice of IRF family co-factors. For example, IRF8 co-operates with BATF3/JUN to instruct homeostatic classical dendritic cell (cDC) differentiation, and with BATF/JUN during inflammatory cDC differentiation.[40] BATF/JUN and IRF4 co-operative binding at AICE motifs is required for instruction of Th17 differentiation and B-cell class switch recombination.[12, 30, 31, 40] Further, it is likely that co-operation of these AP-1/IRF complexes with different STAT family members can confer additional integration of environmental cues for interpretation of combinatorial motifs in regulatory DNA elements.

Transcriptional programming of environmental responsiveness

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

Transcriptional programmes that integrate environmental signals with cell intrinsic features instruct cellular phenotypes, including plasticity. In this context, it is interesting to compare and contrast the transcriptional strategies of FOXP3 and RORγt in control of Treg and Th17 cell identity, respectively. Recent mechanistic insights into the transcriptional regulation of Foxp3 and Rorc and their targets explain some of the characteristics of the Treg and Th17 cellular phenotype. For example, both FOXP3 and RORγt have in common an activity that largely reinforces, stabilizes and maintains a chromatin and gene activation landscape initiated by ERFs. More specifically, these factors augment the expression of critical lineage-specific genes such as il2ra, ctla4, il10, il10ra, cd5, icos and notably, Foxp3 itself, in the case of FOXP3, and il17a, il17f, il1r1 and il23r for RORγt (Fig. 1). This target gene selection reflects the distinct behaviour and biology of Th17 and Treg cells. RORγt augments il23r expression in a positive feedback loop, as STAT3 signalling downstream of IL-23R activates Rorc expression. However, this feedback loop, and maintained expression of Rorc and Th17 lineage fidelity, is dependent on the persistence of environmental IL-23 and transforming growth factor-β (TGF-β), and altered environmental signals, especially IL-12 and interferon-γ, can subvert Rorc expression and the Th17 transcriptional programme, converting cells to the Th1 lineage (Fig. 2).[42-44] In contrast, FOXP3 regulates its own expression upon engagement of a positive feedback loop following activation and CpG demethylation at a Foxp3-intronic enhancer (CNS2), a heritable feature of mature Treg cells, effectively buffering mature Treg cells from changes in environmental signals.[45] These differences may reflect important phenotypic features of these distinct cell types. For example, once innocuous antigen has been encountered and Treg-cell-mediated tolerance has been established, breaking this tolerance, through loss of Treg cell stability, even in the context of inflammation, is likely to be detrimental to the organism. Therefore, the molecular mechanisms described above may have been selected because they achieve Treg cell lineage stability and prevent off-target, innocuous antigen-specific responses during inflammation.[46] In contrast, Th17 cells represent a potent inflammatory Th cell subset endowed with the ability to augment adaptive responses, tissue inflammation, and neutrophil recruitment, and are therefore often juxtaposed with Treg cells as frequent culprits of autoimmune disease.[25] Indeed, studies from both Rudensky and colleagues and Littman and colleagues validated the functional importance of Treg or Th17 cell regulatory elements through comparison with genome-wide association study data. For example, both sites of Treg-specific chromatin accessibility, and binding sites for the core Th17 cell transcription factors overlapped with different mutations linked to ulcerative colitis and rheumatoid arthritis, diseases in which Th17 cells and Treg cells have opposing roles and where dysregulation of either cell type can result in disease.[12, 14] Intuitively then, when not dysregulated by genetic lesions or environmental toxins, Th17 cell environmental responsiveness and lineage plasticity can allow for the harnessing of their potent inflammatory potential to fight infection and resolve tissue damage while assuring their appropriate restraint and reprogramming under homeostatic conditions.

Similarly, Th1 and Th2 cells have encoded appropriate environmental responsiveness and stability into their transcriptional programmes, enabling the maintenance of type-specific memory responses with some capacity for adaptation. Both TBET and GATA3 reinforce their own expression directly, through transcriptional positive feedback loops, and indirectly, through enhancement of cytokine receptor expression and autocrine signals upstream of MRF activation.[47] The TBET target HLX, and perhaps TBET itself can activate TBET gene expression.[23, 48] For both TBET and GATA3, retroviral expression can induce transcription of the endogenous genes.[23, 49] As with FOXP3 autoregulation, these cell intrinsic positive feedback loops confer a degree of environmental buffering and thereby bolster lineage fidelity. Indeed, Th1 or Th2 cells that have undergone several rounds of division, demethylated CpG motifs at key lineage genes, and established transcriptional autoregulatory loops, become highly committed.[50, 51] In contrast, newly differentiated Th1 and Th2 cells are highly responsive to reprogramming following exposure to alternative lineage-instructing cytokines.[52] While the relative contributions of specific mechanisms – transcription factor feedback loops and autocrine signalling – are not well understood for Th1 and Th2 cell plasticity in vivo, these features are reminiscent of early developmental plasticity and later acquisition of the remarkable stability seen in FOXP3-expressing cells.[53] In vivo, newly generated peripherally induced Treg cells (within their first week) retain some plasticity (~ 50% maintain FOXP3 expression) whereas mature peripherally induced Treg cells achieve remarkable stability (~ 99%),[54] through mechanisms also involving CpG demethylation and autoregulation.[45] Hence, the plasticity and stability phenotypes of distinct CD4 T-cell subsets are varied and developmentally regulated, and are controlled by transcriptional and epigenetic mechanisms.

Are master regulators living up to their name?

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

Several recent studies described here detail the relative roles and co-operative function of transcription factors in the initiation of T-cell subset differentiation and provide consensus on a primary role for ERFs in the early activation of enhancers and associated gene transcription. Indeed, with MRFs dispensable for much of the early Th cell transcriptional programme, and their relatively small regulatory footprint, some may see fit to question their ‘master’ status. However, while the in vitro studies are detailed and incisive in their control over comparative conditions, it is crucial to consider what we have learned from in vivo loss-of-function studies, and to appreciate the function of MRFs in heritable maintenance of cellular phenotype, environmental responsiveness and plasticity (see above), as well as the complexity of Th cell phenotypic delineation in the organism. The role of FOXP3 in Treg cell biology illustrates this distinction in perspective well. Stimulation of naive CD4 T-cells through the TCR, together with environmental sensing of TGF-β and IL-2 can recapitulate a significant fraction of the Treg cell transcriptional signature, independent of Foxp3 expression.[35, 55] Perhaps this is analogous to the minor role for TBET, GATA3 and RORγt in initializing Th1, Th2 and Th17 enhancer activation and transcriptional signatures. However, in vivo, FOXP3 is critical for Treg cell identity and loss of Foxp3 in mature Treg cells results in their dedifferentiation, acquisition of alternative T-cell subset phenotype, extensive immunopathologies and death.[29, 56] Although we can appreciate the major role of ERFs in the initial differentiation process and the mechanistic insights gained from these studies, we can also acknowledge that the transcriptional programmes they induce are insufficient for complete in vivo, faithful, CD4 T-cell subset commitment and maintenance. As quantitatively inferior as their roles may seem in the initialization of enhancers and transcriptional programmes, minute features such as modulation of a key set of genes or establishment of stabilizing positive feedback loops, establish MRFs as central and defining factors in CD4 T-cell subsets. Studies of mechanisms employed by MRFs to orchestrate these cellular phenotypes are important for a general understanding of cellular differentiation and identity. For example, MRFs engage in positive feedback loops for core lineage transcription factor expression and augment expression of co-operatively regulated genes, thereby adapting and stabilizing the transcriptional programme. MRFs simultaneously inhibit the expression of genes instructing alternative lineages such as other MRFs or cytokines instructing opposing lineages (Fig. 1). Furthermore, considering the heritable maintenance of most T-cell subsets, MRFs can potentially propagate chromatin and gene states, perhaps even in the absence of the original signals and ERF activation. In these ways, even with a seemingly small initial regulatory footprint, once induced, MRFs can dominantly influence cellular phenotype.

Summary and future issues

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

Recent genomic studies provide important examples of complex transcriptional control of cellular differentiation, and underscore the co-operative and networked action of several transcription factors in immune cell differentiation.[12-14, 30, 31, 39, 40] Whereas we can appreciate the simple significance of cellular instruction through over-expression of factors like MYOD, FOXP3, and in iPS cells, OCT4, SOX2, KLF4 and c-MYC, studies such as those discussed here reveal that such activity occurs through extensive collaboration with supporting factors. Indeed, experiments establishing the sufficiency of many MRFs for lineage instruction relied on stimulation-dependent over-expression and the coincident activation of crucial ERF co-factors.

The integration of information in the form of co-ordinated binding of environmental response factors and nuclear master regulator transcription factors to regulatory sites across the genome (Fig. 1) represents an elegant strategy for initial instruction and subsequent stabilization of immune cell phenotype in response to environmental cues. ERFs play a dominant and immediate role in altering chromatin state and initiating transcription, followed by induced MRF expression resulting in positive feedback transcription loops that stabilize the cellular phenotype. These consecutive steps during CD4 T-cell subset differentiation can be projected onto Waddington's epigenetic landscape, to indicate the contribution of key transcription factors to the restriction and instruction of developmental potential (Fig. 2).

Given the function of MRFs to stabilize cellular phenotype it will be interesting to assess the mechanisms conferring this activity. If not prominent in chromatin remodelling during initial differentiation, are MRFs involved in the maintenance of chromatin accessibility and gene state, in the absence of ERFs, either in quiescent or proliferating cells? Additionally, while early studies of MRFs and STATs focused on signature Th genes with established function, like ifng and the Th2 cytokines, we have little understanding of most of the hundreds to thousands of enhancers that are activated predominantly by ERFs. Certainly, not all enhancers carry equivalent weight in defining T-cell function, and the small set of signature lineage genes potently regulated by MRFs are indisputably crucial for cellular phenotype and function. Further, does the in vitro context of Th cell polarization recapitulate the potential variation of ERF activation downstream of TCR signalling in vivo? For example, increased TCR signal strength can affect mature T-cell polarization (biasing towards Treg and Th17 cell lineages), and one possibility is that signal strength differences result in dosage effects of TCR-associated transcription factors, such as AP-1, IRF4 and NFAT, with intended effects on target gene expression. Furthermore, it will be important to better understand the differences in chromatin states and transcription factor function in initial polarization compared with long-term maintenance of T-cell subsets. Whereas description of enhancer characteristics is extensive – chromatin accessibility, H3K4me1, H3K27ac, p300 recruitment, physical interaction with promoters – it will be exciting to learn more about the precise mechanisms of enhancer-mediated activation of transcription. Finally, we have much to learn about the graded, sequential progression of regulatory chromatin ‘maturation’, from condensed, to poised, to fully active, with augmentation of associated gene transcription, and the specific roles of DNA- and chromatin-binding factors in this process.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References

I appreciate ongoing support and mentorship from C. David Allis. I thank A.Y. Rudensky and members of the Allis and Rudensky laboratories for helpful discussions, and M. Sellars, A. Arvey, C. Li and R. Niec for insightful comments and input on the manuscript. S.Z.J. is supported by the National Institutes of Health under Ruth L. Kirschstein National Research Service Award (GM100616).

References

  1. Top of page
  2. Summary
  3. Introduction: CD4 T-cells, in context
  4. Master regulators and pioneer factors
  5. A genomics era reinterpretation of T-cell master regulators
  6. Environmental response factors and enhancer activation
  7. Transcriptional programming of environmental responsiveness
  8. Are master regulators living up to their name?
  9. Summary and future issues
  10. Acknowledgements
  11. Disclosures
  12. References