Since previous studies have shown that iNKT17 cells can secrete IL-17 through TCR engagement 20, we investigated whether CD1d was HSP assay required for IL-17A mRNA
expression by iNKT17 cells in the pancreas (Fig. 3E). To address this question, we used Vα14 NOD mice expressing CD1d solely in the thymus (CD1dpLck Vα14 NOD mice) 31. RORγt, IL-23R and IFN-γmRNA expression was similar in pancreatic iNKT cells from both types of mice. However, IL-17A mRNA expression was significantly decreased (3-fold) in iNKT cells from mice lacking peripheral CD1d expression. Altogether, our data suggest that iNKT17 cells are activated locally in the pancreas in a CD1d-dependent manner. To evaluate the role of iNKT17 cells in type 1 diabetes, we reconstituted immunodeficient NOD mice with different iNKT cell subsets and analyzed the induction of diabetes after transfer of anti-islet BDC2.5 T cells 32. Since there is no specific antibody available to purify iNKT17 cells, we first determined the frequency of iNKT17 cells in different iNKT cell subpopulations divided according to CD4 and NK1.1
expression of donor cells. As shown in Fig. 3A and Supporting Information Fig. 2, iNKT17 cells are mainly present in the CD4− iNKT cell population and at a higher frequency among NK1.1− CD4− iNKT cells. Therefore, we enriched iNKT17 cells based on their lack of CD4 expression and they were found to represent around 23% of the injected CD4− iNKT cell population (Fig. 3B). Recipient NOD mice were reconstituted buy SAR245409 with CD4− or CD4+ iNKT cells, which were detected in pancreas before BDC2.5 T-cell transfer (Fig. 3B). In order to detect an eventual pathogenic role of iNKT17 cells, all recipient mice were injected with a low number of BDC2.5 T cells, which induces around 30% of diabetes in control mice devoid of iNKT cells (Fig. 3C). Interestingly, in the group of mice reconstituted with CD4− iNKT cells, the incidence of diabetes was significantly (p=0.036) increased Quinapyramine and reached 70%. In contrast, reconstitution with CD4+ iNKT
cells significantly (p=0.033) prevented the development of diabetes. Moreover, when CD4− iNKT cells were further divided according to NK1.1 expression, only NK1.1− CD4− iNKT cells containing the higher frequency of iNKT17 cells exacerbated diabetes (Fig. 3D). Since diabetes induced by diabetogenic BDC2.5 T cells is associated with their production of IFN-γ 13, we have analyzed whether the presence of iNKT cell subsets have influenced their production of IFN-γ and IL-17. As previously described 13, in diabetic control mice devoid of iNKT cells, BDC2.5 T cells produced large amount of IFN-γ in both PLNs and pancreas (Fig. 4A). In diabetic mice reconstituted with CD4− iNKT cells, production of IFN-γ by BDC2.5 T cells was similar as in diabetic control mice and production of IL-17 remained low, less than 1%. While cytokine production by BDC2.5 T cells was similar in both groups of mice, the frequency of BDC2.