T-bet Orchestrates CD8αα IEL Differentiation

Abstract

Very little is known about the transcription factor network that regulates the development of intestinal intraepithelial lymphocytes (IELs). In this issue of Immunity, Klose et al. (2014b) and Reis et al. (2014) demonstrate an essential role for T-bet in regulating the CD8αα IEL differentiation program.

Intraepithelial lymphocytes (IELs) are a heterogeneous population of antigen-experienced T cells strategically distributed in the epithelium of the small and large intestine, where they play critical roles in providing the first line of defense against infections while preserving the integrity of the mucosal barrier (Cheroutre et al., 2011). Natural CD8αα⁺ IELs are selected in the thymus from a small population of immature CD4⁺CD8αβ⁺CD8αα⁺ (triple positive) thymocytes, which after agonist selection develop into double-negative (CD4⁻CD8α⁻) T cell receptor (TCR)ab⁺ or TCRγδ⁺ cells, the precursors of TCRαβ⁺ and TCRγδ⁺ IELs, respectively (Gangadharan et al., 2006). At this stage, IEL precursors upregulate the gut-homing receptors CCR9 and integrin α4β7 and migrate to the intestinal epithelium to complete their differentiation into IELs, characterized by the expression of CD8αα, CD103, and various natural killer (NK) cell-associated markers including CD122 (interleukin-15Rβ [IL-15Rβ] subunit), Ly49A, Ly49E, Ly49G, and NK1.1 (Andrew et al., 1996; Gangadharan et al., 2006; Wurbel et al., 2001). CD8αα⁺ IELs can also be generated via the “developmental diversion” of strongly self-reactive T cells that have not received a CD28-B7 signal in the thymus and thereby escape negative selection (Pobezinsky et al., 2012). In contrast, induced CD8αα⁺ IELs arise from conventional TCRαβ⁺ CD4⁺ and CD8αβ⁺ T cells in the periphery after undergoing a conversion process characterized by the loss of classical CD4⁺ T helper functions and acquisition of the IEL-associated markers CD103 and CD8αα (Reis et al., 2013).

IELs have been well-characterized in their developmental pathways and localization, their reactivity to self and non-self-antigens, and their innate and adaptive effector functions, yet the transcription factor networks responsible for programming IELs have not been fully investigated. In this issue of Immunity, Klose et al. and Reis et al. report that the transcription factor T-bet (T-box expressed in T cells) is critically important for promoting the CD8αα IEL differentiation program. Both groups demonstrate that T-bet is highly expressed in almost all CD8αα⁺ IELs and that T-bet-deficient mice lack CD8αα⁺ IELs. While Klose et al. focused on identifying the molecular mechanism(s) by which T-bet mediates the development of natural CD8αα⁺ IELs, Reis et al. investigated the role of T-bet in the reprogramming of conventional, MHCII-restricted TCRαβ⁺CD4⁺ T cells into induced CD8αα⁺ IELs.

Originally described as the transcription factor regulating CD4⁺ Th1 lineage commitment, T-bet is now recognized to have a fundamental role in various adaptive and innate immune cells. T-bet promotes the differentiation of innate lymphoid cells in the gut, specifically the NKp46⁺ IL-7Rα⁺ innate lymphoid cell 1 (ILC1) and CCR6^–RORγt⁺ ILC3 subsets (Klose et al., 2014a; Klose et al., 2013). In combination with the transcription factor Eomesodermin, T-bet also regulates the development of NK cells, while the survival of natural killer T (NKT) cells is dependent on T-bet-mediated up-regulation of the Il2rb gene (encoding IL-15Rβ, CD122). Because CD8αα⁺ IELs share phenotypic and functional similarities with ILCs and NK and NKT cells, Klose et al. proceeded to investigate whether T-bet and Eomesodermin are required for the development of CD8αα⁺ IELs. The results from their current study now establish that T-bet, but not Eomesodermin, regulates CD8αα⁺ IEL generation. Interestingly, the number of IEL precursors (IELPs) in the thymi of T-bet-deficient (Tbx21^–/–) or wild-type mice was not significantly different, suggesting that T-bet is not required for IEL commitment but rather for their survival or differentiation. Indeed, the authors found that T-bet expression is induced in IELPs in the thymus following strong agonist stimulation and that it is further augmented by IL-15 signaling. Moreover, Tbx21^–/– IELPs were able to successfully home to the intestine when adoptively transferred into immunodeficient Rag2^–/–Il2rg^–/– mice, but they failed to expand in vivo (Klose et al., 2014b). Since it has been reported that IEL expansion is vitally dependent on intestinal epithelial-cell-derived IL-15, the authors examined whether T-bet affects the responsiveness of IEL precursors to IL-15 in the gut and discovered that Tbx21^–/– IELPs did not upregulate CD122 (IL-15Rβ) expression, and thus failed to respond to IL-15. On the other hand, microbial stimuli and food-derived aryl hydrocarbon receptor (AhR) ligands, which are known to significantly influence the composition of the gut-associated immune system, had no effect on T-bet expression. In summary, Klose et al. concluded that T-bet is primarily required for the upregulation of CD122, and hence for the IL-15-dependent activation, differentiation, and expansion of thymus-derived IEL precursors in the periphery (Figure 1).

Figure 1. T-bet Regulates the Differentiation of Natural and Induced CD8αα⁺ Intestinal Intraepithelial Lymphocytes.

IEL populations comprise natural TCRαβ⁺CD8αα⁺ or TCRγδ⁺CD8αα⁺ cells derived from thymic precursors and induced TCRαβ⁺CD8αβ⁺CD8αα⁺ or TCRαβ⁺CD4⁺CD8αα⁺ cells derived from conventional peripheral T cells. (Bottom left) Thymic IEL precursors (IELp) express a low amount of IL-15R via a T-bet-dependent mechanism. Following the migration of IELp to the gut, intestinal epithelium-derived IL-15 triggers increased T-bet (encoded by Tbx21) expression, which in turn upregulates expression of CD122 (IL-15Rβ subunit), creating a positive feedback loop to promote CD8αα⁺ IEL maturation, survival, and proliferation. (Bottom right) Peripheral TCRαβ⁺ T cells upregulate T-bet expression upon exposure to TGF-β, retinoic acid (RA), and either IFN-γ or IL-27 in the intestinal milieu. T-bet activates the transcription factor Runx3, which induces the expression of CD103, CD8αα⁺, and other aspects of the CD8αα⁺ IEL differentiation program. T-bet also suppresses the T-helper lineage-specific factor Thpok and other conventional T-helper effector functions, thus promoting the conversion of conventional CD4⁺ T cells into induced CD8αα⁺ IELs.

The work of Reis et al. offers a complementary perspective by examining the role of T-bet in the development of peripherally induced CD8αα⁺ IELs. Reis et al. have had a long-standing interest in mechanisms that prevent the aberrant activation of TCRαβ⁺CD4⁺ T cells in the gut. Their previous study elegantly showed that one such mechanism is the conversion of TCRαβ⁺CD4⁺ T cells into CD8αα⁺ IELs, a process that is dependent on the transcription factor Runx3, encompassing the acquisition of CD103 and CD8αα expression, as well as loss of the T-helper lineage-specific transcription factor ThPOK (Reis et al., 2013). In the current report, the authors continue their work on the gut-specific reprogramming of TCRαβ⁺CD4⁺ T cells, providing unequivocal evidence that T-bet is required for the upregulation of Runx3 and is therefore essential for induction of the CD8αα IEL differentiation program. Additionally, the authors demonstrate that T-bet further promotes CD8αα IEL differentiation by inhibiting CD4⁺ T helper functions in a Runx3-independent manner (Figure 1).

Using both in vitro and in vivo systems, Reis et al. first demonstrated that exposure of TCRαβ⁺CD4⁺ T cells to TGF-β, retinoic acid, and either IFN-γ or IL-27 resulted in rapid acquisition of a CD8αα IEL-specific gene-expression profile, including the upregulation of Tbx21 and Runx3 followed by Thpok downregulation (Reis et al., 2014). T-bet-deficient TCRαβ⁺ CD4⁺ T cells failed to adopt the features of CD8αα⁺ IELs, indicating that T-bet plays an indispensable role in regulating the CD8αα⁺ differentiation program. Interestingly, IFN-γ and IL-27, both known to be strong inducers of T-bet, were found to have distinct roles in the differentiation of specific CD8αα⁺ IEL populations: the differentiation of TCRαβ⁺CD4⁺ CD8αα⁺ IELs depended on IFN-γ, while IL-27 was needed for the development of TCRγδ⁺ CD8αα⁺ and TCRαβ⁺CD8αβ⁺ CD8αα⁺ IELs (Reis et al., 2014). One could speculate that distinct CD8αα⁺ IEL populations differ in their expression of cytokine receptors or their preferential localization in the vicinity of IFN-γ-producing NK cells or IL-27-expressing dendritic cells; however, the authors did not elaborate further on these differences in cytokine responsiveness. Because a deficiency in either T-bet or Runx3 results in a drastic reduction of all CD8αα⁺ IELs, Reis et al. focused on dissecting the hierarchy of these transcription factors in the CD8αα⁺ IEL differentiation program. Runx3 was ectopically expressed in Tbx21^–/– cells to determine whether CD8αα⁺ IEL differentiation can be rescued in the absence of T-bet, and vice versa. The results from these experiments, in combination with ChIP analyses, suggested the following model for the generation of induced CD8αα⁺ IELs: factors within the intestinal milieu can induce T-bet expression, which in turn regulates Runx3 by directly binding to regulatory elements in the Runx3 locus. Once Runx3 is induced, the subsequent upregulation of CD8αα⁺ IEL-associated genes is primarily driven by Runx3. The authors also found that T-bet promotes the loss of ThPOK, which could be facilitated directly through the binding of T-bet to Thpok regulatory elements or indirectly through the modulation of Runx3 expression. In either case, this T-bet-mediated downregulation of Thpok is dependent on the presence of Runx3, demonstrating the importance of T-bet-Runx3 protein-protein interactions in at least some aspects of the CD8αα⁺ IEL differentiation program.

These two ground-breaking reports provide exciting clues regarding the transcriptional circuitry that controls the terminal maturation of CD8αα⁺ IELs, and uncover a role of T-bet in this process. Together, their findings raise some questions about the relationships between natural, induced, and developmentally diverted CD8αα⁺ IELs. For example, Reis et al. demonstrated that the differentiation of induced CD8αα⁺ IELs revolves around T-bet-mediated upregulation of Runx3, whereas Klose et al. reported that IL-15 can induce Runx3 expression in thymic IEL precursors independently of T-bet. It is possible that these contrasting observations on the regulation of Runx3 expression reflect epigenetic differences that have yet to be elucidated, between thymic IELPs and mature conventional TCRαβ⁺ T cells in the periphery prior to initiation of CD8αα⁺ IEL differentiation. Nevertheless, the fact that normal levels of Runx3 expression in Tbx21^–/– thymic IELPs were not sufficient for the development of natural Tbx21^–/– CD8αα⁺ IELs in the gut indicates that natural CD8αα⁺ IELs, similar to induced CD8αα⁺ IELs, depend on the concerted expression and cooperative action of T-bet and Runx3 for their development. Furthermore, while IL-15 signaling mediates the survival of CD8αα⁺ IELs by inducing expression of the antiapoptotic protein Bcl2, overexpression of Bcl2 in Tbx21^–/– mice did not restore the CD8αα⁺ IEL compartment, suggesting that the requirement for T-bet in CD8αα⁺ IELs is not solely limited to the regulation of CD122 (Klose et al., 2014b). It remains to be seen whether additional target genes, directly or indirectly downstream of T-bet, play a similarly critical role in the differentiation, maintenance, or function of either natural or induced CD8αα⁺ IELs. With that in mind, Klose et al. and Reis et al. have provided an intriguing framework from which future studies can be developed in order to unravel the complexities of these (and other) intestinal lymphocyte populations that are so critical to healthy gut homeostasis.