Critical role of ROR-γt in a new thymic pathway leading to IL-17-producing invariant NKT cell differentiation

α-GalCer Stimulation Induces Thymic IL-17 Production.

We recently characterized a distinct subset of iNKT cells that produce high levels of IL-17 and are therefore termed iNKT17 cells (8). To address the question of whether this population is already present in the thymus, we stimulated iNKT cell-enriched thymic CD8^neg cells with α-GalCer and found that they did effectively produce substantial amounts of IL-17 in addition to IL-4 and IFN-γ (Fig. 1A).

CD44^posNK1.1^negCD4^neg Thymic iNKT Cells Are the Major Source of IL-17.

The understanding of thymic iNKT cell development has considerably progressed in recent years. It is generally acknowledged that precursors of this lineage diverge from mainstream thymocyte development at the CD4^posCD8^pos double-positive (DP) stage to undergo a series of expansion and differentiation steps giving rise to peripheral IL-4- and IFN-γ-producing iNKT cells (18). During their progression to the mature CD24^low phenotype, 3 intermediate differentiation stages have been distinguished ranging from CD44^lowNK1.1^neg to CD44^highNK1.1^neg, and CD44^highNK1.1^pos cells (18–20). Based on this definition, we sorted thymic CD8^neg CD1d-tetramer^pos iNKT cells into 4 subsets, namely CD44^lowNK1.1^neg (stage 1), CD44^highNK1.1^negCD4^pos (stage 2, CD4^pos), CD44^highNK1.1^negCD4^neg (stage 2, CD4^neg), and CD44^highNK1.1^pos (stage 3) (Fig. 1B) to assess their functional characteristics. IL-17 production was very low in stages 1 and 3 whereas most of the activity was concentrated in the CD44^posNK1.1^negCD4^neg fraction (stage 2, CD4^neg) (Fig. 1C). We used intracellular cytokine staining to assess IL-17 on the single cell level and confirmed that producer cells are actually predominant in the CD44^posNK1.1^negCD4^neg iNKT subset because >50% of cells was positive (Fig. 2A).

IL-17-Producing CD44^posNK1.1^negCD4^neg iNKT Thymocytes Represent a Mature Stage of iNKT Cell Development.

Little is known about CD44^posNK1.1^negCD4^neg iNKT cells and their functions. It has been reported that they derive from HSA^lowNK1.1^negCD4^pos iNKT cell precursors, similarly to their CD4^pos counterpart (21–23), but it has not been established whether they can give rise to progeny. To address this issue, we seeded fetal thymic organ cultures (FTOC) from iNKT cell-deficient Jα18^−/− mice with the 4 thymic iNKT cell subsets represented in Fig. 1B and analyzed their differentiation potentials under these conditions. Our results are consistent with the conclusion that the CD44^posNK1.1^negCD4^neg (stage 2, CD4^neg) iNKT cells represent a mature differentiation stage because they failed to develop in FTOC, similarly to mature CD44^posNK1.1^pos (stage 3) iNKT cells, whereas CD44^negNK1.1^neg (stage 1) and CD44^posNK1.1^negCD4^pos (stage 2, CD4^pos) precursors gave rise to mature NK1.1^pos progeny (Fig. 2B).

IL-17-Producing iNKT Cells Express ROR-γt.

We performed qPCR array profiling to determine the expression of genes typically associated with iNKT cell differentiation and to identify the cytokine profile generated by the 4 thymic iNKT cell subsets described in Fig. 1B. These subsets were sorted and immediately analyzed without further manipulation (Fig. 3). IL-4 and GATA-3 mRNAs were expressed at different levels whereas both IFN-γ and T-bet transcripts were highly up-regulated in mature NK1.1^pos cells (stage 3). IL-17 mRNAs were detected only at the stage 2 whether these cells were CD4^neg or CD4^pos whereas IL-23R mRNA expression was basically observed only among stage2 CD4^neg cells. Analysis of mRNA encoding ROR-γt, which is required for the differentiation of conventional Th17 cells (24), revealed high expression in CD44^posNK1.1^negCD4^neg iNKT cells (stage 2, CD4^neg), but also among stage 2, CD4^pos, prompting us to verify whether ROR-γt and IL-17 production was also associated in thymic iNKT cells.

It has been reported before that most thymic double-positive precursors express ROR-γt before engaging into the differentiation process during which receptor expression is gradually lost (21, 22). We confirmed this notion in experiments performed with Rorc(γt)-Gfp^TG mice because the majority of immature CD8^posCD1d tetramer^pos or CD24^posCD1d tetramer^pos iNKT cells expressed ROR-γt (up to 90%) in contrast with their more mature CD8^negCD1d tetramer^pos (Fig. 4A) or CD24^negCD1d tetramer^pos (Fig. 4B) iNKT cell counterparts. Nevertheless, a small fraction of CD8^negCD1d tetramer^pos iNKT cells conserved their ROR-γt expression, and most of them were CD44^posNK1.1^negCD4^neg (Fig. 4C), similar to IL-17-producing thymic iNKT cells (Fig. 2 A and B). It can therefore be concluded that nearly all ROR-γt^pos iNKT cells and IL-17-producing iNKT cells share the same phenotype. It is noteworthy that stage 1 CD44^negNK1.1^neg cells were present among gated ROR-γt^pos CD8^neg iNKT cells, suggesting that a fraction of iNKT cell precursors retains ROR-γt expression to give rive rise ultimately to more mature ROR-γt^pos iNKT cells.

It remained to be determined whether ROR-γt expression and IL-17 production were effectively associated in iNKT cells. Intracellular staining of sorted populations showed that this was actually the case because nearly all IL-17-producing iNKT cells were both ROR-γt^pos and CD44^posNK1.1^negCD4^neg (Fig. 5A). Furthermore, sorted CD44^posNK1.1^negCD4^neg ROR-γt^pos iNKT cells produced high levels of IL-17 upon stimulation whereas their ROR-γt^neg counterpart did not (Fig. 5B). CD44^posNK1.1^negCD4^pos ROR-γt^pos iNKT can also produce some IL-17, even though they are less efficient than the corresponding CD4^neg subset (Fig. 5C).

We have reported that iNKT17 cells are present in the periphery, which raises the question of whether they conserve their ROR-γt expression. We found that 2.5% iNKT cells in the spleen were ROR-γt^pos (Fig. 6A) and among these, 91% lacked the NK1.1 marker. This observation is in keeping with our previous data (8), recently confirmed by other groups (25, 26), that iNKT17 cells are indeed NK1.1^neg. Similar to their thymic counterpart, sorted splenic ROR-γt^pos but not ROR-γt^neg iNKT cells produce high levels of IL-17 after α-GalCer stimulation (Fig. 6B), designating ROR-γt as a useful marker to identify peripheral IL-17-producing iNKT cell subsets.

IL-17-Producing ROR-γt^pos iNKT Cells Evolve from an Alternative Pathway of Development of Thymic iNKT Cells.

We addressed the question whether thymic ROR-γt^neg iNKT cells could reacquire the expression of ROR-γt lost after the double-positive differentiation stage or inversely, whether ROR-γt^pos iNKT cells could eventually become negative. Because of the low number of thymic ROR-γt^pos iNKT cells, we developed a culture system that, similar to FTOC, could sustain iNKT cell differentiation. We took advantage of the fact that IL-7 induces preferential iNKT cell expansion (27) to set up cocultures between total thymocytes from Jα18^−/− mice, used as feeder cells, and CD44^negNK1.1^neg (stage 1), CD44^posNK1.1^negCD4^neg (stage 2, CD4^neg), or CD44^posNK1.1^negCD4^pos (stage 2, CD4^pos) cells purified from sorted thymic CD1d tetramer^pos ROR-γt^neg iNKT cells. This system performed as well as FTOC, because differentiation from both stage 1 (Fig. 7B) and 2 (Fig. 7 C and D) was readily obtained. Even though ROR-γt^neg iNKT populations generated more mature progeny, they failed to acquire ROR-γt and to produce IL-17 (Fig. 7 B–D). It can therefore be assumed that the developmental signals requisite for the maturation of these cells are inadequate in terms of ROR-γt expression or IL-17 production. By contrast, sorted ROR-γt^pos iNKT cells maintain their massive IL-17 production (Fig. 7 E and F), despite the down-regulation that occurs in some cells after culture and stimulation (Fig. 7F). These data suggest that once the cells have been instructed to produce IL-17 through a ROR-γt-dependent process that is probably initiated in vivo after the double-positive (CD4^posCD8^pos) or CD24^pos stage, a continuous high expression of ROR-γt is no longer necessary to produce IL-17, at least in these IL-7-dependent in vitro conditions.

Taken together, our findings show that ROR-γt is essential for the alternative developmental program that generates highly effective iNKT17 cells. Our findings are consistent with the notion that some iNKT cells are predetermined to become IL-17 producers, although the intra and/or extracellular signals that induce or sustain ROR-γt expression during the maturation process remain unknown. In this context, we assessed whether the generation of IL-17-producing iNKT cells was affected by the absence of factors shown to be required for conventional Th17 cell development, such as IL-6. Using IL-6^−/−Rorc(γt)-Gfp^TG mice we observed no modification in IL-17 production by thymic iNKT ROR-γt^pos cells (data not shown), proving that IL-6 is not required for the generation of iNKT17 cells.

It is possible that some specific endogenous antigens direct iNKT cell precursor cells preferentially toward iNKT17 cell differentiation during their positive selection at the double-positive CD4^posCD8^pos stage. However, until now, we were unable to detect any significant difference in terms of Vβ repertoire, whether or not iNKT cells produced IL-17 (8). Another possibility would be that some unspecified costimulatory molecules might intervene at some important branching point, to set off the program that favors iNKT17 cell development, eventually by maintaining the expression of ROR-γt. Hence, our data suggest that ROR-γt plays an important role in the program that determines iNKT17 differentiation in the thymus, in addition to its implication in the development of peripheral Th17 cells.

The specific physiologic or pathologic functions of iNKT17 cells remains to be elucidated even though they have already been implicated in certain models of lung inflammation (8, 28). Whatever the answer to these issues, our present study has uncovered an alternative thymic pathway from which IL-17-producing iNKT cells with the mature CD44^posNK1.1^negCD4^negROR-γt^pos phenotype will emerge (Fig. 8).

Animals.

Jα18^−/− (29), Rorc(γt)-Gfp^TG (30), IL-6^−/−Rorc(γt)-Gfp^TG (30), and C57BL/6 male mice from 4–8-weeks-of-age were used. Animal experiments were performed according to the French Institutional Committee.

iNKT Cell Enrichement and Sorting.

Enriched iNKT cells were obtained by depleting the thymus of CD8 cells and splenocytes of CD8, CD19, and CD62L cells labeled with the corresponding mAb (BD PharMingen) and with anti-rat Ig-coated magnetic beads (Dynabeads and Invitrogen), according to the manufacturers' protocols. Cells were then stained and iNKT cell subsets were sorted by using a FACSAria cell sorter (Becton Dickinson).

Immunofluorescence.

After iNKT cell enrichment, cells were incubated with CD1d-tetramer-APC (provided by NHI tetramer facilities), anti-NK1.1-PerCP-Cy5.5, anti-CD4-Pacific Blue, anti-CD8-PE or FITC (when cells from Rorc(γt)-Gfp^TG mice were not used), anti-CD24-PE, and anti-CD44-FITC or -PECy7 (BD PharMingen). For intracellular staining, cells were fixed in 4% PFA, washed, and permeabilized with 0.5% saponin (Sigma-Aldrich), then incubated with anti-IL-17-PE or isotype controls (Biosciences). In some experiments, anti-GFP Ab and anti-rabbit IgG-Alexa Fluor 488 were used. The cells were washed and analyzed in a FACSCanto II (Becton Dickinson) by using FlowJo software.

Cell Culture.

Thymic CD8^neg or sorted iNKT cells were cultured at a final concentration of 5 × 10⁶ or 5 × 10⁵ cells per ml, respectively, with or without peritoneal macrophages from Jα18^−/− mice as APCs, at a ratio of 1:2, and stimulated with 100 ng/ml α-GalCer (Alexis). All culture supernatants were harvested and stored at −80°C until cytokine detection by ELISA, as described (8). In some experiments, thymic iNKT cells were stimulated for 4 h with 10⁻⁸ M phorbol 12-myristate 13-acetate (PMA) (Sigma–Aldrich), 5 × 10⁻⁶ M ionomycin, and 10 μg/ml brefeldin A. Cells were then stained for intracellular cytokine detection.

Expansion of iNKT Cell Precursors.

In some experiments, fetal thymus organ cultures (FTOCs) were carried out. In brief, thymic lobes were collected at day 17 of gestation from Jα18^−/− mice. Hanging drops were prepared in Terasaki plates by adding to each well 30,000 sorted cells per thymic lobe. The plates were immediately inverted to form hanging drops and incubated for 48 h in a humidified incubator (5% CO₂, 37°C). After incubation, the lobes were removed from the hanging drops, washed, and placed on a nuclepore filter (Nuclepore, Costar) resting on a grid in 2 ml of complete media, for 14 days. In another set of experiments, 2 × 10⁴ sorted iNKT cells per ml were cocultured with total thymocytes from Jα18^−/− mice in 24-well culture plates (20 × 10⁶ cells per ml) and stimulated with IL-7 (20 ng/ml) The plates were incubated for 4 days. Dead cells were removed by FicollPaque centrifugation.

Gene Expression Analysis.

Total RNA for gene expression analysis by real time RT-PCR was extracted from 500–5,000 cells electronically sorted into vials, by using an RNeasy Micro Kit (Qiagen). High quality RNA (250–500 pg) was subjected to 1 linear mRNA amplification cycle by using the MessageBooster Kit for qRT-PCR (Epicentre Biotechnologies). Amplified mRNA (50–100 ng) was then converted to cDNA by using SuperScript III (Invitrogen) according to the manufacturer's protocol. The expression of 84 different genes was measured by using the Th17 RT2 Profiler PCR Array for mouse (SuperArray Bioscience Corporation) and confirmed for individual genes by using specific primer pairs (SuperArray Bioscience Corporation). Real-time PCR was performed in a PTC-200 thermocycler equipped with a Chromo4 detector (Bio-Rad Laboratories). Data were analyzed by using Opticon Monitor software (Bio-Rad Laboratories).