The sequential activity of Gata3 and Thpok is required for the differentiation of CD1d-restricted CD4⁺ NKT cells

Introduction

While the vast majority of CD4⁺ T cells are MHC II-restricted, small but functionally important contingents of CD4⁺ cells are restricted by other MHC or MHC-like molecules. The most abundant of these cells recognize lipid-bound CD1d molecules, express markers typical of NK cells, including the surface antigen NK 1.1, and can be identified in vitro through their binding to tetramers of α-galactosylceramide (Gal-ceramide)-bound CD1d 1. NKT cells express an invariant TCR-α chain and a limited repertoire of TCR-β chains (hence their name, invariant NKT cells, iNKT), and NKT-cell development is largely dependent on the transcription factor PLZF 2, 3. Although they are selected by MHC I-like CD1d molecules, iNKT cells typically do not express CD8, and appear as CD4⁺CD8⁻ (CD4 single positive (SP)) or CD4⁻CD8⁻ (double negative (DN)). This is unlike subsets of MHC I-restricted cells that acquire effector function in the thymus and adopt a CD8-SP phenotype 4, 5. Currently, it is unclear why, unlike conventional MHC I-restricted thymocytes, iNKT precursors differentiate into CD4⁺ rather than CD8⁺ cells.

The transcription factors Gata3 and Thpok are both required for the generation of CD4⁺ T cells from CD4⁺CD8⁺ (double positive, DP) thymocytes 6. Gata3 is expressed throughout T-cell differentiation and is required at multiple branch points during T-cell development or effector differentiation 7. Gata3 expression is transiently up-regulated during the differentiation of CD4-lineage thymocytes 8; it is required at an early step of CD4⁺ T-cell development, and has been proposed to promote the expression of genes specifically expressed in CD4⁺ T cells, or required for their differentiation 9. Thpok is not expressed in DP thymocytes and its up-regulation during CD4-lineage differentiation requires Gata3 9-11. Thpok is required for the commitment of MHC II-restricted thymocytes to the CD4-lineage, notably by repressing the expression of the transcription factor Runx3 and of cytotoxic genes, and by promoting CD8 silencing 12-15.

In the current study, we found that Thpok is expressed at a high level in CD1d-restricted NKT cells in the thymus and peripheral lymphoid organs. Thpok is required for the differentiation of CD4⁺ iNKT cells and for their repression of CD8, a function reminiscent of the one it serves in MHC II-restricted T cells. We further document that Thpok expression in iNKT cells is highly dependent on Gata3. We conclude from these findings that a similar cascade, sequentially involving Gata3 and Thpok, promotes the differentiation of MHC II- and CD1d-restricted thymocytes into CD4⁺ T cells.

Results and discussion

We previously described a bacterial artificial chromosome reporter transgene in which the first Thpok coding exon (exon 2) has been replaced by a GFP cDNA 9, 15 (Supporting Information Fig. 1, thereafter referred to as Thpok^GFP transgene). In two independently derived mouse lines, this transgene expressed GFP in essentially all CD4⁺ T cells and mature CD4-SP thymocytes but not in CD8-SP thymocytes, or in naïve or memory spleen or LN CD8⁺ T cells 9. However, we found low-level GFP expression in activated Thpok^GFP CD8⁺ cells (Fig. 1A), in agreement with a recent report 16. Immunoblot analyses demonstrated expression of endogenous Thpok protein (Fig. 1A), although at levels well below those in CD4⁺ cells.

This suggested that Thpok expression in the T-cell lineage was not strictly limited to MHC II-restricted cells. To further explore this possibility, we introduced the Thpok^GFP reporter into MHC II-deficient mice. In such mice, a small subset of thymocytes expressed GFP, of which most were TCR^hi CD24^lo (Fig. 1B). These cells shared two additional characteristics: they did not express CD8 and were CD44^hi (Fig. 1B, C). Invariant NKT (iNKT) cells, which recognize CD1d-bound lipids 17, form the main subset of CD4⁺ T cells in MHC-II-deficient mice and are TCR^hi CD24^lo CD44^hi, prompting us to assess whether they express Thpok. Supporting this idea, most GFP⁺ thymocytes in MHC-II-deficient Thpok^GFP mice expressed the NKT cell marker NK1.1 (Fig. 1C). Indeed, most iNKT thymocytes, identified using α-Gal-ceramide-loaded CD1d tetramers, expressed the Thpok reporter (Fig. 1D and Supporting Information Fig. 2), and so did peripheral iNKT cells (Fig. 1E).

iNKT cells derive from DP thymocytes 18, cease CD8 expression early during their differentiation, and therefore become CD4-SP cells. Subsequently, a subset of them terminates CD4 expression and becomes CD4⁻CD8⁻ DN thymocytes. These DN iNKT cells also expressed GFP, although slightly less than their CD4-expressing counterparts (Fig. 1D). This differs from MHC II-restricted cells, which all express both Thpok and CD4. Future studies will determine the mechanistic basis for this difference. Unlike CD4-differentiating thymocytes and naïve CD4⁺ T cells 13, 15, iNKT cells co-express Runx3 and Thpok, with similar Runx3 levels in CD4⁺ and DN subsets (Y. X. and R. B., unpublished results); whether or not Runx3 affects CD4 expression in iNKT cells remains to be determined, although we and others previously found that Thpok antagonizes the Cd4-silencing function of Runx3 14, 15, 19.

To examine whether Thpok is required for iNKT-cell development, we evaluated iNKT populations in Thpok-deficient mice 9. While there was no reproducible difference in the frequency of CD1d-specific cells between Thpok-deficient and Thpok-sufficient thymi, Thpok-deficient iNKT cells failed to express CD4, and a subset of them expressed low levels of CD8 (Fig. 2A). iNKT cells were also found in normal frequency in the spleen and liver of Thpok-deficient mice, where they also displayed a DN or CD8-expressing phenotype (Fig. 2B and data not shown). Thus, unlike the related zinc finger transcription factor PLZF 2, 3, Thpok is not needed for the development of iNKT cells, but it is required for their continued CD4 expression. Of note, Thpok-deficient iNKT cells expressed less CD8 than conventional CD8⁺ cells, and a subset of them expressed CD8α but not CD8β (Fig. 2B, right and Supporting Information Fig. 3A). This suggests that so far unidentified transcription factors, expressed in conventional CD8⁺ cells but not in iNKT cells, promote CD8 expression. In addition, Thpok-deficient iNKT cells had impaired expression of markers of effector function, including NK1.1 and Granzyme B (Fig. 2C and D), although not of CD44 (Supporting Information Fig. 3B), indicating that Thpok controls multiple aspects of iNKT-cell differentiation. The reduced Granzyme B expression by Thpok-deficient iNKT cells (Fig. 2D) was especially noticeable since Thpok represses Granzyme B expression in conventional T cells 12, 15. Thus, Thpok affects multiple aspects of iNKT cells development in addition to their coreceptor expression. A recent report similarly found that Thpok promotes IFN-γ expression in iNKT cells 20, whereas it has the opposite effect in conventional CD4⁺ cells 15.

Similar to their Thpok-deficient counterparts, Gata3-deficient iNKT cells have reduced CD4 expression 21. We had previously proposed a sequential model for Gata3 and Thpok function during the development of MHC II-restricted CD4⁺ T cells, whereby Gata3 is required for Thpok expression and Thpok promotes CD4-commitment and CD8 silencing 9. Consequently, we wondered whether Gata3 was required for Thpok expression in iNKT cells. To assess this, we evaluated the expression of Thpok mRNA by RT-PCR in iNKT cells purified from Gata3-sufficient mice and from mice in which floxed Gata3 alleles were deleted in DP cells by a Cd4-Cre transgene (referred to as Gata3^ΔDP mice)22. These analyses showed a 90% reduction in Thpok expression in Gata3-deficient relative to wild-type iNKT cells (Fig. 3A). To determine whether this resulted from a homogenous reduction of Thpok expression, we introduced the Thpok^GFP transgene into Gata3^ΔDP mice. Consistent with our previous finding that Gata3 is required for Thpok expression, there was little or no GFP fluorescence in conventional thymocytes (CD44^lo NK1.1⁻) in Thpok^GFP Gata3^ΔDP mice (Supporting Information Fig. 4A). In iNKT thymocytes, Gata3 disruption made GFP expression barely detectable, compared with their Gata3-sufficient counterparts (Fig. 3B, columns 2 and 4, and Supporting Information Fig. 4B). We conclude from these experiments that Gata3 promotes Thpok expression in iNKT cells.

As previously reported 21, Gata3 disruption resulted in the disappearance of the CD4⁺ iNKT subset. We observed a minor re-expression of CD8 on Gata3^ΔDP iNKT thymocytes (data not shown and Fig. 3B), with substantial biological variability among mice. This re-expression was not nearly as pronounced as on Thpok-deficient iNKT cells, and it is possible that this is due to the small residual Thpok expression in Gata3-deficient iNKT thymocytes (Supporting Information Fig. 4B). Of note, Gata3 disruption results in drastically reduced peripheral iNKT-cell numbers 21, whereas Thpok disruption had no such effect, indicating that Gata3 serves other functions in addition to promoting Thpok expression. This is supported by the altered development of Gata3-deficient iNKT thymocytes, illustrated by the premature expression of NK1.1 by Gata3- but not Thpok-deficient iNKT thymocytes (Ref. 21 and Supporting Information Fig. 3B).

The expression of Thpok by CD1d-selected iNKT cells indicates that Thpok can be expressed in thymocytes in the absence of MHC-II co-engagement of TCR and CD4. The selection of CD1d-restricted iNKT cells differs from that of conventional T cells in two critical aspects 1, 23. First, selecting CD1d molecules are expressed by DP thymocytes but not by the thymic epithelium 24. Second, the selection of iNKT cells requires homotypic interactions between SLAM family receptors and signaling through the adaptor SAP, none of which is involved in selection of conventional T cells 25. Future work will determine whether either factor promotes Thpok expression by iNKT cells. Because CD1d does not bind CD8, the expression of Thpok by iNKT cells is consistent with the ‘kinetic signaling’ model of lineage choice, which proposes that CD8-independent TCR signaling in thymocytes promotes CD4 differentiation, and therefore results in Thpok expression 26.

The expression of Thpok in iNKT cells depends on Gata3, similar to what we previously reported in MHC II-restricted thymocytes 9. However, while Gata3-deficient MHC II-restricted thymocytes fail to differentiate into mature T cells, iNKT cells develop despite Gata3 disruption, allowing us to dissect effects of Gata3 on cell development from those on Thpok expression. As a result, we could identify a Gata3 requirement for Thpok expression in iNKT cells, distinct from that for selection. Such a demonstration had not been possible in MHC II-restricted thymocytes, which in absence of Gata3 are arrested before the stage at which they would normally express Thpok 9, 27. We had previously shown that Gata3 was recruited to the Thpok locus in MHC II-restricted thymocytes 9. Recent ‘Chipseq’ large scale analyses have found that Gata3 binds the same site, close to the 3′ extremity of Thpok intron 1 in iNKT cells (within DNase I hypersensitivity site A in Ref. 9) (J.Z., William E. Paul and Keji Zhao, unpublished data). In MHC II-restricted thymocytes, this activity of Gata3 is part of a broader network of positive and negative transcriptional regulators that determine Thpok expression 6, 28, and it is likely that this is the case in iNKT cells as well.

Previous studies have led to the idea that Thpok is up-regulated as a result of MHC II signaling in thymocytes 10, 11, 13, 14. We now add an important amendment to this paradigm, namely that Thpok is also expressed in iNKT cells. As in conventional MHC II-restricted thymocytes T cells, Thpok is required for proper repression of CD8 expression by CD1d-restricted iNKT thymocytes and for their development into CD4⁺ cells, and its expression is Gata3-dependent. We propose that the sequential activation of Gata3 and Thpok is a characteristic of CD4⁺ T cells, regardless of their antigenic specificity.

Materials and methods