Histone/protein deacetylase 3 dictates critical aspects of regulatory T cell development and function

Regulatory T cells (Tregs) comprise a subset of CD4⁺ T lymphocytes that holds promise as immunotherapy for inflammatory disease states including autoimmune disorders, solid organ allograft rejection, graft-versus-host disease, and acute inflammatory conditions¹. However, limited availability of Treg numbers restricts their usefulness in cell transfer applications, particularly as patients may need multiple infusions for chronic diseases. Pharmacologic manipulation of the Treg lineage – either as part of an ex vivo expansion strategy or via drugs administered to a patient – may augment the Treg suppressive effect, stabilize a regulatory phenotype, and ultimately promote immune homeostasis. Conversely, adjunctive drug therapy that diminishes Treg function could benefit malignant diseases in which anti-tumor immunity prevents metastasis and promotes tumor clearance. Epigenome-modifying drugs represent an attractive pharmacotherapeutic strategy because of their profound effects on Treg phenotype and function. For example, pan-histone/protein deacetylase inhibitors (pan-HDACi) augment Treg numbers and suppressive function in part by promoting acetylation of FOXP3 – the master Treg transcription factor – and enhancing DNA binding². In a recent issue of the Journal of Clinical Investigation, Wang and colleagues published unexpected results determining that histone/protein deacetylase 3 (HDAC3) is critical for Treg development and function³. Their results carry important implications for translating Treg epigenetic mechanisms into the clinical arena and raise interesting questions about biologic control of Treg function.

Because pan-HDACi display narrow therapeutic indices when administered systemically⁴, investigators have developed intense interest in which HDAC isoforms could be targeted to achieve a desired effect on Treg phenotype and function. Of the 11 classical zinc-dependent HDAC metalloenzymes, most available data describe the effects of class IIa HDACs (HDAC4, HDAC5, HDAC7, and HDAC9) on Treg biology. For example, knockdown of HDAC7 and HDAC9 can augment Treg suppressive function⁴. However, class IIa HDACs have weak histone/protein deacetylase activity owing to a histidine rather than a tyrosine residue in their catalytic domain⁵. Rather than acting via an enzymatic mechanism, class IIa HDACs may exert their effects by functioning as bromodomains (protein domains that recognize acetylated lysine residues) in complex with class I HDACs such as HDAC3^6,7.

The FOXP3 transcription factor suppresses interleukin-2 (IL-2) production at the transcriptional level; inability to produce IL-2 is a defining characteristic of the Treg lineage. As many gene repression complexes contain HDAC3, Wang and colleagues first performed immunoprecipitation assays determining that FOXP3 and HDAC3 physically interact. The FOXP3-HDAC3 interaction had functional implications, as HDAC3 bound to the Treg Il2 promoter in a FOXP3-dependent manner and inhibited its transcription. While increasing FOXP3 levels in intact Tregs augments their suppressive function⁸, HDAC3-deficient FOXP3⁺ Tregs produced significant amounts of IL-2 despite normal levels of Foxp3 mRNA and elevated or normal levels of other Treg-related gene products (Ctla4, Tgfb, Il10, Ebi3, and Gitr). Microarray analysis revealed that HDAC3-deficient Tregs possessed greatly dysregulated gene expression with activation of pro-inflammatory pathways including NF-κB. Figure 1 outlines the effects of Treg-restricted HDAC3 deficiency.

Figure 1.

Schematic representation of the effect HDAC3 deletion has on regulatory T cell (Treg) development and function. HDAC3 binds to FOXP3 and facilitates gene repression at the Il2 promoter and other pro-inflammatory gene loci. In the absence of HDAC3, Tregs produce IL-2 and fail to properly populate lymphoid organs. iTreg generation is likewise impaired. HDAC3-deficient Tregs display limited suppressive function that leads to systemic autoimmunity and early mortality. IL-2, interleukin 2; nTreg, natural Treg; iTreg, induced Treg; FOXP3, forkhead box protein 3; HDAC3, histone/protein deacetylase 3.

The authors created a Treg-specific HDAC3-deficient mouse (Hdac3^fl/flFoxp3^Cre) to explore the role HDAC3 plays in Treg development and function, as global Hdac3 knockout mice do not survive past embryonic stage E9.5. In contrast, Treg-specific HDAC3-deficient mice were viable but developed fatal autoimmunity with profound activation of conventional T and B cells in the lungs, liver, kidneys, and bone marrow. Male mice experienced more intense lung and liver inflammation compared to female littermates suggesting that sex differences may interact with the Treg epigenetic state to modulate the inflammatory phenotype. A mild proliferative glomerulonephritis underpinned the renal inflammation and stemmed from development of cryoglobulins. Cryoglobulin formation in the Hdac3^fl/flFoxp3^Cre mouse raises interesting questions about how Tregs control B cell responses – an understudied area in lymphocyte biology.

The Treg-specific HDAC-deficient mice died by six weeks of age; rescue adoptive transfer of wild-type (WT) Tregs to these mice delayed onset of the fatal autoimmune phenotype. Adoptive transfer prolonged median survival by about three weeks, but all adoptive transfer recipients still died before 12 weeks of age. The authors did not explore why adoptive transfer of WT Tregs, which are usually a long-lived cell type, failed to rescue Hdac3^fl/flFoxp3^Cre mice for a longer duration. It is possible that the pro-inflammatory state of the Hdac3^fl/flFoxp3^Cre mouse was hostile to WT Treg survival, proliferation, and function. Future studies could determine mechanisms underlying why WT Tregs failed to control inflammation in the Hdac3^fl/flFoxp3^Cre mouse on a long-term basis.

Hdac3^fl/flFoxp3^Cre mice demonstrated lymph node and spleen enlargement but atrophic thymuses. Interestingly, these mice contained markedly decreased numbers of splenic Tregs but increased numbers of lymph node and thymic Tregs with unimpaired thymic Treg production. The differential balance of Tregs in lymphoid organs points to an effect of HDAC3 deficiency on Treg homing and trapping, and the authors provide data that CCR7 and sphingosine-1-phosphate receptor expression may be responsible for the observed differences in Treg numbers. Importantly, HDAC3-deficient Tregs displayed similar proliferation and apoptosis when compared to WT Tregs, suggesting that differences in Treg numbers were not due to differential expansion or death.

Because both thymus-derived natural Tregs (nTregs) and peripherally induced Tregs (iTregs) comprise the extra-thymic FOXP3⁺ Treg pool, Wang and colleagues examined if iTreg induction was impaired in the absence of HDAC3. Peripheral conversion of HDAC3-deficient CD4⁺CD25⁻ cells (non-Tregs) into iTregs was compromised under iTreg-polarizing conditions. The authors found that HDAC3-deficent T cells had limited accessibility of the Foxp3 promoter and CNS2 sites, which are critical regions at the Foxp3 locus determining Treg development, phenotype, function, and stability. HDAC3-deficent T cells also displayed unrestrained IL-2 and IL-6 expression that blocked iTreg conversion. These data support the hypothesis that HDAC3 is critical for the normal development of both main Treg subsets: nTregs and iTregs.

The investigators also found that HDAC3 determines Treg suppressive function. Both spleen- and lymph node-derived Tregs from Hdac3^fl/flFoxp3^Cre mice demonstrated impaired function in an in vitro lymphocyte suppression assay; suppressive function improved with retroviral transduction of HDAC3 into HDAC3-deficient Tregs. Transduction of HDAC3 did not restore function completely to WT function, which suggests that developmental imprinting in the absence of HDAC3 may program Tregs in a manner that is irreversible by later correction of the HDAC3 deficiency. The investigators also tested how HDAC3 determines Treg function in three in vivo systems. First, they found that HDAC3-deficient Tregs exerted limited suppression of homeostatic conventional T cell proliferation when co-injected along with conventional T cells into lymphocyte-deficient Rag-1^−/− mice. Second, Wang and colleagues induced cardiac allograft rejection by grafting BALB/c hearts into a strain-mismatched C57BL/6 Rag-1^−/− mouse and subsequently injecting C57BL/6 conventional T cells. Administration of WT Tregs mitigated long-term rejection (more than 120 days) in the allograft rejection model; however, administration of HDAC3-deficient Tregs led to graft failure by 33 days post-transplant. Finally, HDAC3-deficient Tregs failed to control colitis induced by adoptive transfer of conventional T cells to Rag-1^−/− mice. These three models convincingly demonstrate the critical role HDAC3 plays in determining the ability of Tregs to suppress effector T cell function. However, the impact of Treg HDAC3 on controlling innate immunity – for example, Treg interactions with monocyte/macrophage populations – remains unknown. Future studies could seek to understand the effect of Treg epigenetic manipulation on their interaction with other innate and adaptive immune system components.

Wang and colleagues' results underscore the importance of testing the effect of epigenetic enzyme isoforms in a specific cell type of interest. Their work contrasts with a report of Hdac3 deletion driven by CD4-Cre that caused Hdac3 deletion in all T cells⁹. Conventional T cell development was normal in CD4-Cre HDAC3 conditional knockout mice, but these mice failed to develop invariant NKT cells. Wang and colleagues' data also provide important information for investigators interested in designing isoform- or class-specific HDACi. Drug design should carefully consider how to rationally approach Treg-based epigenetic therapies, as manipulation of specific HDAC isoforms – rather than inhibiting all HDACs with a pan-HDACi – can lead to counterintuitive and possibly harmful effects on Treg development and function. Going forward, rational drug design might target certain HDACs over others to achieve a desired consequence. For example, an HDAC3-specific inhibitor may be beneficial for malignant conditions in which Treg suppressive function is deleterious.

Collectively, Wang and colleagues generated unexpected data showing the critical role of HDAC3 in dictating Treg development and function. Their paper has translational implications for rational drug design, which if fruitful could benefit myriad conditions ranging from inflammatory diseases to malignancies. Future studies should work to understand the function of other HDAC isoforms specifically within Tregs, as more data will be necessary and useful for endeavors to create effective Treg-based epigenetic pharmacotherapies.