C57BL/6 mice were purchased from CLEA Japan, Inc. (Tokyo, Japan). C57BL/6 Ly5.1 congenic mice were obtained from RIKEN BRC (Tsukuba, Japan). IL-21 receptor (IL-21R)-deficient (IL21R^-/-) mice (15) were backcrossed over 8 generations onto C57BL/6 mice. TCRδ^-/- mice were obtained from the Jackson Laboratory (Bar Harbor, ME). All the mice were housed in microisolator cages under specific pathogen-free conditions. The Chiba University Animal Care and Use Committee approved animal procedures used in this study.

RPMI 1640 medium containing 10% fetal calf serum, antibiotics, 1 x GlutaMAX (Gibco), 10 mM HEPES (Gibco), 1 mM sodium pyruvate (Gibco), 54 μM 2-mercaptoethanol, and 1 x non-essential amino acids (Gibco) was used for cell culture (complete γδT medium). For the purification of nγδT17 cells, B220^-CD11b^-TER119^-CD4^-CD8^- cells were purified from splenocytes and LN cells of naïve WT mice or IL21R^-/- mice using biotin-conjugated antibodies and MojoSort streptavidin nanobeads according to the manufacturer’s instructions (BioLegend, San Diego, CA), and then CD27^-TCRγ/δ⁺ cells, previously defined as γδT17 cells in the periphery (16), were sorted using an SH800 cell sorter (Sony, Tokyo, Japan). 4 x10³ cells were stimulated with plate-bound anti-TCRγ/δ antibody (1 μg/ml) in the presence of anti-IFN-γ antibody (10 μg/ml) and indicated cytokines (IL-1β: 5 ng/ml, IL-23: 5 ng/ml, IL-21: 20 ng/ml) in a 96 well round-bottom plate for three days. Cells were then washed and cultured in a new well without anti-TCRγ/δ antibody in the presence of indicated cytokines for another three days.

γδT17 cell expansion culture from whole splenocytes was performed as described elsewhere (17) with minor modifications. In brief, 2.5 x10⁵ cells from the spleen and LNs were stimulated with plate-bound anti-TCRγ/δ antibody (1 μg/ml) in the presence of IL-1β (5 ng/ml) and IL-23 (5 ng/ml) in 96 well flat-bottom plates for three days. Cells were washed and cultured in a new well without anti-TCRγ/δ antibody in the presence of IL-1β and IL-23 for another three days. After washing, cells were cultured in the presence of IL-7 (20 ng/ml) for another three days.

Antibodies used in this study are listed in Supplementary Table 1 . Recombinant murine IL-1β and IL-23 were purchased from R&D Systems (Minneapolis, MN), and recombinant murine IL-7 and IL-21 were purchased from BioLegend. STAT3 inhibitor (S3I-201) was purchased from Selleck Chemicals (Houston, TX).

Vγ6 staining was performed as described elsewhere (18) with minor modifications. In brief, before staining of cell surface markers, cells were stained with anti-TCRγ/δ antibody (clone: GL3) (BioLegend), followed by 17D1 hybridoma supernatants that recognize both Vγ5 and Vγ6 TCR. FITC-conjugated mouse anti-rat IgM monoclonal antibody (MRM-47, BioLegend) was used to capture cells that were positively stained with 17D1 hybridoma supernatants. Anti-Vγ1 and Vγ5 TCR staining and other cell surface staining were performed along with FITC-conjugated mouse anti-rat IgM antibody.

Apoptotic cells were stained with Annexin V according to the manufacturer’s instruction (BD bioscience).

To detect intracellular cytokines, cells were stimulated with PMA and ionomycin for 4 hours in the presence of BD GolgiPlug (BD Biosciences, Franklin Lakes, NJ). Intracellular cytokine and transcription factor staining were performed using an eBioscience Foxp3 transcription factor fixation/permeabilization kit (Thermo Fisher, Waltham, MA) as described previously (19). Cells were analyzed by FACS Canto II, FACS LSR Fortessa (BD Bioscience), or Novocyte Penteon (Agilent Technologies, Santa Clara, CA). FACS profiles were analyzed using the FlowJo software ver. 10.8 (BD Biosciences).

Bone marrow cells were isolated from CD45.1⁺ WT mice or CD45.2⁺ IL21R^-/- mice, and then CD90⁺ cells were depleted to remove mature αβ and γδ T cells using a biotin-conjugated anti-CD90.2 antibody (30-H12, BioLegend) and Mojosort streptavidin nanobeads (BioLegend). CD90^- bone marrow cells (total 1x10⁷ cells) were injected intravenously to lethally irradiated (950 rad) TCRδ^-/- mice. For generating mixed BMC mice, a mixture of CD90^- bone marrow cells (total 1x10⁷ cells) of CD45.1⁺ WT mice and CD45.2⁺ IL21R^-/- mice at a 1:1 ratio was injected intravenously to lethally irradiated TCRδ^-/- mice.

Mice were injected subcutaneously with complete Freund’s adjuvant (CFA), which is an emulsion of PBS, incomplete Freund’s adjuvant (IFA) (Chondrex), and M. tuberculosis extract H37 Ra (Difco) (PBS 100 μl, IFA 100 μl, and M. tuberculosis extract 0.4 mg/mouse) into the tail base, and then intraperitoneally injected with pertussis toxin diluted in PBS (200 ng/mouse) 6 hours and 48 hours after CFA administration. For the induction of EAE, MOG 35-55 peptide (MEVGWYRSPFSRVVHLYRNGK) was emulsified together with CFA (100 μg/mouse). The induction of EAE and scoring are described elsewhere (19, 20).

Statistical analyses were performed using GraphPad Prism ver.9 (GraphPad Software), and the results are described in each Figure legend. Data are summarized as mean ± SEM.

We first investigated the role of IL-21 in developing nγδT17 cells in the fetal thymus. IL-17A⁺RORγt⁺CCR9⁺γδT cells, which represent nγδT17 cells in the fetal thymus (6), and Vγ4⁺nγδT17 cells were comparable between IL21R^-/- mice and littermate WT mice ( Figure 1A ). As previously reported (10), nγδT17 cells were lower in the spleen in adult IL21R^-/- mice than in WT mice ( Figure 1B ). Notably, Vγ4⁺nγδT17 cells but not Vγ4^-nγδT17 cells were significantly decreased in IL21R^-/- mice ( Figure 1B ). Inconsistent with a previous report showing increased lung γδT17 cells in IL21R^-/- mice (12), we found that the number of nγδT17 cells in the lung was decreased in IL21R^-/- mice as compared to WT mice ( Figure 1C ). Again, Vγ4⁺nγδT17 cells but not Vγ4^-nγδT17 cells were decreased in the lungs of IL21R^-/- mice ( Figure 1C ). This discrepancy may be due to environmental factors because the lungs are rich in Vγ6⁺γδT17 cells, whose pool is affected by commensal bacteria and increases with age (14, 18). Meanwhile, the frequencies of Vγ4⁺nγδT17 cells and Vγ4^-nγδT17 cells in the skin and large intestine were not significantly different between WT and IL21R^-/- mice ( Supplementary Figure 1 ).

Given that IL-21 is constitutively produced by approximately 5% of splenic CD4⁺ T cells in naïve mice (21), we next analyzed proliferation and apoptosis markers of Vγ4⁺nγδT17 cells and Vγ4^-nγδT17 cells in the spleen of naïve mice. The active proliferation marker Ki-67 positive cells were significantly decreased in Vγ4⁺nγδT17 cells but not in Vγ4^–nγδT17 cells in naive IL21R^-/- mice compared with those in naïve WT mice ( Figure 1D ). Meanwhile, the proportion of annexin V⁺ DAPI^– apoptotic cells in each subset of IL21R^-/- mice did not differ from that of WT mice ( Figure 1E ). These findings suggest that IL-21 is dispensable for developing nγδT17 cells in the fetal thymus but is indispensable for the maintenance and/or proliferation of Vγ4⁺nγδT17 cells in the periphery, especially in the spleen and the lung.

We next investigated the effect of IL-21 on the proliferation of nγδT17 cell subsets. For this purpose, CD27^- γδT cells, which represent γδT17 cells in the periphery (16), were purified from the spleen and lymph nodes of WT and IL21R^-/- mice ( Figure 2A ). At the start of the culture, the number of Vγ4⁺nγδT17 cells among CD27^- γδT cells was slightly decreased in IL21R^-/- mice ( Supplementary Figure 2 ). Consistent with previous reports (7–9, 22), IL-1+IL-23 stimulation together with anti-TCRγ/δ induced approximately 10-fold proliferation of Vγ4⁺nγδT17 cells in WT mice and IL21R^-/- mice. Notably, IL-21 in combination with IL-1, but not IL-21 alone, induced the proliferation of Vγ4⁺nγδT17 cells as effectively as IL-1+IL-23 in WT mice but not in IL21R^-/- mice ( Figure 2A ). Importantly, Vγ6⁺nγδT17 cells, the major subset of Vγ4^–γδT17 cells, did not significantly proliferate in response to IL-21+IL-1, but they responded well to IL-1+IL23 ( Figure 2A ), suggesting that the proliferative effect of IL-21+IL-1 is somewhat unique to Vγ4⁺nγδT17 cells.

We next examined cytokine receptor expression to determine how IL-21+IL-1 synergistically promotes Vγ4⁺γδT17 cell proliferation. Consistent with a previous report (22), IL-23 significantly upregulated the expression of IL-1R in both Vγ4⁺nγδT17 cells and Vγ6⁺nγδT17 cells ( Figure 2B , left panel). These results may explain the synergistic effect of IL-1+IL23 on the proliferation of Vγ4⁺nγδT17 cells and Vγ6⁺nγδT17 cells. However, IL-21 did not significantly change the expression of IL-21R, IL-1R, and IL-23R in Vγ4⁺nγδT17 cells and Vγ6⁺nγδT17 cells ( Figure 2B ), suggesting that the alteration of cytokine receptor expression might not be the mechanism of different responses of Vγ4⁺nγδT17 cells and Vγ6⁺nγδT17 cells on IL-21+IL-1 stimulation. Moreover, the effect of IL-21+IL-1 on Vγ4⁺nγδT17 cell proliferation was not mediated through the production of IL-23, as the addition of an anti-IL-23 p19 neutralizing antibody did not have a significant effect ( Supplementary Figure 3 ).

Given that IL-21 exerts its effect mainly through STAT3, PI3K, and MAPK pathway downstream of IL-21R (23) and that STAT3 is essential for IL-17 production from IL-23-stimulated Vγ4⁺ γδT17 cells (22), we next examined the role of STAT3 pathways. As shown in Figure 2C , a STAT3 inhibitor reduced the proliferation of IL-1+IL-23-stimulated Vγ4⁺nγδT17 cells and Vγ6⁺nγδT17 cells, as well as IL-21+IL-1-stimulated Vγ4⁺nγδT17 cells. Taken together, IL-21 induces the proliferation of IL-1-stimulated Vγ4⁺nγδT17 cells but not Vγ6⁺nγδT17 cells via the STAT3 pathway.

Next, we aimed to determine the role of IL-21 in the development of γδT17 cells that do not originate from the fetal thymus. Previous studies have shown that in lethally irradiated TCRδ^-/- mice or Rag1^-/- mice that received bone marrow (BM) cell transfer from WT mice, γδT17 cells originated from the fetal thymus are absent, but newly induced γδT17 cells can develop in the draining LNs (dLNs) following immunization (7–9). We employed these newly induced γδT17 cell systems to avoid the contamination of nγδT17 cells ( Figure 3A ).

First, we analyzed the development of γδT17 cells in the thymus of BMC mice before immunization. As expected, IL-17A⁺RORγt⁺CCR9^-CXCR6⁺ nγδT17 cells were absent in the thymus of BMC mice but were present in the thymus of control naïve CD45.1 WT mice ( Figure 3A ). Importantly, IL-17A^-RORγt⁺CCR9⁺CXCR6^- immature γδT17 cells could develop in the adult thymus even in TCRδ^-/- mice if these mice were received CD90⁺ mature T cells-depleted WT bone marrow cell transfer ( Figure 3A ). These results are consistent with previous reports showing that IL-17-producing γδT cells are absent in the thymus of BMC mice if IL-17 is used as a marker of γδT17 cells (7, 8). A recent study employing single-cell RNA sequencing has shown that immature γδT17 cells in the adult thymus are presumed to be a precursor of inducible γδT17 cells, which requires antigen stimulation to produce IL-17 (6). Therefore, both inducible γδT17 and newly induced γδT17 may originate from immature γδT17 cells, which can develop even in the adult thymus.

To determine the role of IL-21 in the development of immature γδT17 cells and newly induced γδT17 cells, we employed mixed BMC (mBMC) mice in which a mixture of CD45.1⁺WT and CD45.2⁺IL21R^-/- BM cells was injected intravenously into lethally irradiated adult TCRδ^-/- mice ( Figure 3B ). After the engraftment of donor cells, the frequencies of CD45.1⁺ T cells and CD45.2⁺ T cells in the spleen and lymph nodes were comparable ( Supplementary Figure 4 ). The frequencies of γδT cells and IL-17A^-RORγt⁺ immature γδT17 cells in the thymus were comparable between IL21R^-/- and WT cells, suggesting that the engraftment of γδT cells is equivalent and that IL-21 is indispensable for the development of immature γδT17 cells in the thymus ( Figure 3C ). However, upon stimulation with complete Freund’s adjuvant (CFA) and pertussis toxin (Ptx), the frequencies of newly induced γδT17 cells and Vγ4⁺ newly induced γδT17 cells in the dLNs were significantly lower in IL21R^-/- cells than in WT cells in mBMC mice ( Figure 3D ). Upon the induction of EAE, the frequencies of newly induced γδT17 cells and Vγ4⁺ newly induced γδT17 cells in the brain and spinal cord were significantly lower in IL21R^-/- cells in mBMC mice ( Figures 3E, F ). These findings indicate that IL-21 is dispensable for the development of immature γδT17 cells in the thymus but is required for the expansion of newly induced γδT17 cells and Vγ4⁺ newly induced γδT17 cells under inflammatory conditions.

We finally examined the pathophysiological roles of IL-21R expressed on γδT17 cells in EAE. In this experiment, in vitro-expanded WT- or IL21R^-/-CD27^-γδT17 cells were prepared ( Figure 4A ) and intravenously injected into TCRδ^-/- mice, and EAE was induced in the recipient mice ( Figure 4B ). As shown in Figure 4A , after γδT17 cell expansion with IL-1+IL-23 and, subsequently, with IL-7 (17), most of the CD27^-γδT17 cells were IL-17A⁺RORγt⁺ in WT and IL21R^-/- cells. The ratio of Vγ4⁺CD27^-γδT17 cells to Vγ4^-CD27^-γδT17 cells and the capacity of IL-17 production in Vγ4⁺CD27^-γδT17 cells and Vγ4^-CD27^-γδT17 cells were comparable between WT and IL21R^-/- cells ( Figure 4A ). After the adoptive transfer of γδT17 cells and the induction of EAE, IL21R^-/- γδT17 cell-transferred TCRδ^-/- mice showed lower EAE disease scores than WT γδT17 cell-transferred TCRδ^-/- mice ( Figure 4C ). TCRδ^-/- mice without γδT17 cell transfer (PBS) exhibited mild symptoms of EAE with a delayed onset, as previously reported (24).

Seven days after EAE induction, Vγ4⁺γδT17 cells but not total γδT17 cells in the spleen were significantly lower in IL21R^-/- γδT17 cell-transferred TCRδ^-/- mice than in WT γδT17 cell-transferred TCRδ^-/- mice ( Figure 4D ). Twenty-one days after EAE induction, although γδT17 cells were undetectable in the brain and spinal cord in all groups, Th17 cells in the spinal cord were significantly lower in IL21R^-/- γδT17 cell-transferred TCRδ^-/- mice than in WT γδT17 cell-transferred TCRδ^-/- mice ( Figure 4E ), consistent with a previous study showing that γδT17 cells amplify Th17 cell responses (25). These findings suggest that IL-21R on γδT17 cells maintains Vγ4⁺γδT17 cells, amplifies Th17 responses, and exacerbates EAE.