MHC class II expressing thymocytes suppress invariant NKT cell development

Abstract

NKT cells are positively selected on cortical thymocytes expressing the non-classical MHC class I CD1d molecules. However, it is less clear how NKT cells are negatively selected in the thymus. In this study, we investigated the role of MHC class II expression in NKT cell development. Transgenic mice expressing MHC class II on thymocytes and peripheral T cells had a marked reduction in invariant NKT (iNKT) cells. Reduced numbers of iNKT cells correlated with the absence of in vivo production of cytokines in response to the iNKT cell agonist α–galactosylceramide. Using mixed bone marrow chimeras, we found that MHC class II expressing thymocytes suppressed the development of iNKT cells in trans in a CD4-dependent manner. Our observations have significant implications for human iNKT cell development since human thymocytes express MHC class II, which can lead to inefficient selection of iNKT cells.

INTRODUCTION

NKT cells are a specialized subset of T cells that co-express NK cell markers, such as NK1.1. They originate in the thymus and are positively selected by CD1d on cortical DP thymocytes^{1,2, 3}. There are at least two sub-populations of NKT cells based on their TCR diversity¹. The major subset is called invariant NKT (iNKT) cells because their TCR repertoire is very limited with the majority of them carrying the invariant Vα14-Jα18 αchain paired mostly with one of the three Vβ chains (Vβ8.2, Vβ7, and Vβ2) in mice. The other subset has a diverse TCR repertoire and represents the non-invariant NKT (non-iNKT) cells.

iNKT cells can be negatively selected when iNKT TCR receives too strong signal^2–4. The addition of iNKT cell agonist α-galactosylceramide (α-GalCer) to fetal thymic organ culture leads to iNKT cell deletion^{5, 6} and the transgenic expression of CD8α and CD8β chains in T cells also results in a reduction of iNKT cells⁷. Transgenic mice overexpressing the NK activating receptor LY49D have reduced iNKT cell development⁸. Furthermore, over-expression of CD1d in dendritic cells decreased the number of iNKT cells^{6, 9}.

Unlike murine NKT cells, the number of Vα24⁺ iNKT cells in humans is highly variable among individuals and they represent only a small portion of total NKT cells, whereas non-iNKT cells are the major population^{1, 10–14}. Human thymocytes can express not only CD1 but also MHC class II¹⁵. However, it is not known whether MHC class II on thymocytes plays a role during NKT cell development because murine thymocytes do not express MHC class II. In this study, we showed that MHC class II-expressing thymocytes suppress the development of iNKT cells partly due to the interaction of MHC class II with CD4.

RESULTS AND DISCUSSION

To study the effect of MHC class II expression on NKT cell development, we utilized mice that express MHC class II in thymocytes and peripheral T cells by introducing the MHC class II transactivator (CIITA) gene as a transgene (CIITA^Tg)¹⁶. CIITA^Tg mice had reduced thymic NK1.1⁺TCRβ⁺ NKT cells but not in the periphery (Figure 1a). We next examined iNKT cells using CD1d tetramers loaded with α–GalCer (hereafter referred to as tetramers) that specifically identify iNKT cells. As shown in Figure 1b and 1c, CIITA^Tg mice exhibited a marked reduction in iNKT cells both in the thymus and the peripheral tissue, indicating CIITA expression in T cells led to a selective failure of iNKT development and a corresponding expansion of non-iNKT cells. The reduction of iNKT cell number was MHC haplotype independent (Figure 1d). However, unlike CD1d^−/− mice that had no iNKT cells, residual iNKT cells were present in CIITA^Tg mice (Figure 1b). These residual iNKT cells in the CIITA^Tg thymus seemed to have undergone proper developmental stages and become mature iNKT cells as defined by CD44 and NK1.1 expression, albeit a slight increase and decrease in percentages of immature and the most mature iNKT cells, respectively (Fig. 1e). The decrease in iNKT cell number in the periphery resulted in very little circulating IL-4 or IFN-γ upon in vivo activation (Supplementary Figure 1a). The CD1d levels on CIITA^Tg DP thymocytes (Supplementary Figure 1b) were comparable to WT cells.

Figure 1.

Reduction of iNKT cells in CIITA^Tg mice. Expression of TCRβ and NK1.1 markers (a) or TCRβ and tetramer (b) on indicated cell types from WT (CIITA^Tg littermates), CIITA^Tg, and CD1d^−/− mice. The number in the dot plots represents the percentage of NK1.1⁺TCRβ⁺ NKT cells (a) or iNKT cells (b). Mice used were 10–11 weeks old and similar results were obtained from at least four mice per group. (c) Absolute numbers of iNKT cells. Values are mean ± SD of iNKT cells present in thymi (n=10), spleens (n=6), and livers (n=4). Mice were 2–3 months old. (d) Representative iNKT cell staining of 3 pairs of WT (CIITA^Tg littermates) and CIITA^Tg on the H-2^d background. (e) Analysis of NK1.1 and CD44 expression on thymic iNKT cells. Tetramer⁺HSA^lo cells were gated and percentages of cells in each quadrant are shown. Results are representative of four pairs of WT and CIITA^Tg littermates. Mice were 10 weeks old.

Because CIITA can turn on the expression of multiple genes involved in MHC class II antigen presentation, such as the invariant chain (Ii)¹⁷, we wanted to determine whether the expression of MHC class II or a gene product regulated by CIITA is responsible for the observed inefficient iNKT cell development. To do this, we generated CIITA^Tg mice lacking MHC class II by crossing them with MHC class II deficient Aβ ^−/− mice (CIITA^Tg /Aβ^−/−). In these mice, developing thymocytes express the CIITA transgene but not MHC class II. The absence of MHC class II in CIITA^Tg/Abb^−/− mice significantly improved the generation of iNKT cells (Figure 2a and 2b), suggesting that MHC class II participates in development of iNKT cells. The incomplete restoration implies the possible involvement of other CIITA-regulated components in iNKT cell development. One putative candidate is Ii that is known to associate with CD1d in endosomes and enhances CD1d antigen presentation of dendritic cells and B cells to iNKT cells¹⁸. Further study is required to determine whether Ii plays a suppressive role in iNKT cell development.

Figure 2.

MHC class II inhibits iNKT cell development. (a) MHC class II expression in thymocytes is necessary for the inhibition of iNKT cell generation. Total thymocytes from the indicated mice were analyzed for TCRβ and tetramer reactivity. WT and CIITA^Tg mice on the Abb^+/− and Abb^−/− background were littermates. Results are representative of at least 4 experiments. (b) Relative percentages of thymic iNKT cells in at least 8 mice of Abb^+/−, CIITA^Tg/Abb^+/−, CIITA^Tg /Abb^−/−, or Abb^−/−. The value of the percentage of iNKT cells in WT mice was set at 100%. Mice were 2–3 months old. (c) Suppression of iNKT cell development by MHC class II-expressing thymoyctes in trans. BM cells from WT (CD45.2) and WT (CIITA^Tg littermates; CD45.1/CD45.2) or WT (CD45.2) together with CIITA^Tg (CD45.1/CD45.2) were co-injected into B6.SJL (CD45.1) recipients. Shown at the top is the composition of total thymocytes in reconstituted mice. TCRβ versus tetramer profiles of thymocytes are shown at the bottom. Chimeras were analyzed 10–12 weeks post BM transplantation. (d) MHC class II and CD4 expression is required for the disappearance of iNKT cells in mixed BM chimeric mice as indicated. BM were prepared from WT (CD45.1/2), Abb^−/− (CD45.2), CIITA^Tg/Abb^−/− (CD45.2), CIITA^Tg/CD1d^−/− (CD45.2), CIITA^Tg (CD45.1/2), and CD4^−/− (CD45.2) mice and transferred to B6.SJL mice. The number in the dot plots represents the percentage of iNKT cells developed from WT BM cells except the bottom group that shows CD4^−/− iNKT cells. Chimeras were analyzed 2–3 months post BM transplantation.

To test whether MHC class II-expressing thymocytes fail to develop into iNKT cells in a cell autonomous manner, we constructed mixed bone marrow (BM) chimeric mice. Different CD45 markers were used to distinguish cells derived from the two different BM sources and the host. In WT+WT→B6.SJL mice, each WT BM population generated a sizable iNKT cell population (Figure 2c, left group; Supplementary Figure 2). However, in the presence of CIITA^Tg BM, WT BM cells failed to generate a comparable proportion of iNKT cells, implying that MHC class II expressing thymocytes suppress neighboring thymocytes to progress to iNKT cells in trans (Figure 2c, right group; Supplementary Figure 2).

To further substantiate the role of MHC class II in iNKT cell development, we co-transferred WT BM cells together with CIITA^Tg/Abb^−/− BM to B6.SJL recipients. If MHC class II expression itself is responsible for deleting thymocytes, iNKT cell development should be restored in WT+CIITA^Tg/Abb^−/− →B6.SJL mice. Indeed, iNKT cells originated from WT BM were present at a normal level (compare the 2^nd row in Figure 2d to the right group in Figure 2c; Supplementary Figure 2). In the control WT+Abb^−/−→B6.SJL chimeras, iNKT cells were also generated (Figure 2d, top row; Supplementary Figure 2). Previously, over-expressing CD1d on DC that also express MHC class II lead to negative selection of iNKT cells^{6, 9}. Therefore, we asked whether co-expression of CD1d together with MHC class II on the same cells is necessary to delete iNKT cells. iNKT cells derived from WT BM were dramatically decreased in the presence of CIITA^Tg/CD1d^−/− thymocytes indicating un unlikely role of CD1d in deletion of iNKT cells (Figure 2d, 3^rd row; Supplementary Figure 2). Lastly, MHC class II on the thymocyte surface might function as a co-receptor for CD4 molecules, enhancing the avidity of TCR on NKT progenitors that are being selected. If so, the strength of signaling could reach beyond the threshold for survival of cells. Indeed, in the CD4^−/−+CIITA^Tg→B6.SJL chimeras, CD4^−/− thymocytes were developed to iNKT cells normally in the presence of MHC class II expressing thymocytes (Figure 2d, bottom row; Supplementary Figure 2), suggesting that the interaction of MHC class II with CD4 is at least partly responsible for the disappearance of iNKT cells. Thus, in addition to its role in iNKT cell activation^{19, 20}, CD4 seems to participate in the regulation of iNKT cell development.

Taken all together, we have demonstrated, for the first time, that MHC class II plays an important role during iNKT cell development. Similar to CIITA^Tg mice, humans have much fewer iNKT cells than non-iNKT cells compared to mice. Perhaps, this is related to the fact that human fetal and neonatal thymocytes express MHC class II, which likely plays an inhibitory role in iNKT cell generation. The final tuning of the strength of the signal delivered by CD1d and/or MHC class II may be critical for survival or death of developing thymocytes to progress toward iNKT cells.

METHODS

Mice

Mice carrying the human type III CIITA transgene (CIITA^Tg), CIITA^Tg mice on the MHC class II deficient Abb^−/− background (CIITA^Tg/Abb^−/−), and CD1d^−/− on the C57BL/6 background were described previously^{15, 16, 21}. CIITA^Tg and CD1d^−/− mice were backcrossed with C57BL/6 (B6) mice at least 6 to 8 times, respectively. CIITA^Tg mice on the Balb/c background were backcrossed 6 times. CD4^−/− mice and B6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and C57BL/6.SJL (B6.SJL) mice from Taconic (Germantown, NY). Non-CIITA^Tg littermates from heterozygous CIITA^Tg breeding were used as wild type (WT) controls. CIITA^Tg and WT mice were also bred to carry CD45.1 and CD45.2 congenic markers. All mice were housed in the animal facility at the Indiana University School of Medicine (IUSM) or the University of Michigan Medical School under SPF conditions and used at 6–12 weeks of age. All animal experiments were performed under protocols approved by the institutions.

Mixed bone marrow chimeras

Preparation of mature lymphocyte-depleted BM cells and BM transfers were done as described previously¹⁵. To generate mixed BM chimeras, BM cells from two different types of donor mice were mixed at a ratio of 1:1 and each lethally-irradiated recipient mouse (950 rad) was infused with 2–5×10⁵ cells via tail vein injection.

Flow cytometry

All antibodies used for flow cytometry were purchased from BD Biosciences (San Diego, CA). Cells were pre-incubated with the anti-FcγR mAb 2.4G2 to block non-specific antibody binding before stained with the following FITC-, PE-, PerCP-, CyChrome-, APC-or biotin-conjugated antibodies: TCRβ (H57), CD4 (L3T4), B220 (RA3-6B2), NK1.1 (PK136), CD1d (1B1), CD45.1 (A20), CD45.2 (104). Flourochome-conjugated streptavidin was used to visualize staining by biotinylated primary antibodies. APC-conjugated CD1d-α– GalCer tetramers were generously provided by Dr. Albert Bendalac from University of Chicago and the NIH Tetramer Facility. Events were acquired on a FACSCalibur or FACSCanto (BD Biosciences) flow cytometer and the data were analyzed using the CELLQuest Pro or FlowJo software.

In vivo Cytokine Production Assay

Mice were injected via i.v with 2 μg α– GalCer²² or vehicle in 100 μl of PBS. After 2 hours, mice were sacrificed and serum IL-4 and IFN-γ were detected using ELISA kits (BD Biosciences).

Statistic analysis

The nonparametric Wilcoxon Rank Sum Test was performed to calculate statistical significance. P values less than 0.05 were considered statistically significant.