Invariant natural killer T-cell subsets have diverse graft-versus-host-disease-preventing and antitumor effects

Abstract

Invariant natural killer T (iNKT) cells are a T-cell subset with potent immunomodulatory properties. Experimental evidence in mice and observational studies in humans indicate that iNKT cells have antitumor potential as well as the ability to suppress acute and chronic graft-versus-host-disease (GVHD). Murine iNKT cells differentiate during thymic development into iNKT1, iNKT2, and iNKT17 sublineages, which differ transcriptomically and epigenomically and have subset-specific developmental requirements. Whether distinct iNKT sublineages also differ in their antitumor effect and their ability to suppress GVHD is currently unknown. In this work, we generated highly purified murine iNKT sublineages, characterized their transcriptomic and epigenomic landscape, and assessed specific functions. We show that iNKT2 and iNKT17, but not iNKT1, cells efficiently suppress T-cell activation in vitro and mitigate murine acute GVHD in vivo. Conversely, we show that iNKT1 cells display the highest antitumor activity against murine B-cell lymphoma cells both in vitro and in vivo. Thus, we report for the first time that iNKT sublineages have distinct and different functions, with iNKT1 cells having the highest antitumor activity and iNKT2 and iNKT17 cells having immune-regulatory properties. These results have important implications for the translation of iNKT cell therapies to the clinic for cancer immunotherapy as well as for the prevention and treatment of GVHD.

Graphical abstract

Figure 1.

Identification of surface molecules for sorting of iNKT sublineages using scRNA-seq. (A) Uniform manifold approximation and projection (UMAP) plot of scRNA-seq data showing distinct clusters of iNKT cell subsets: iNKT1 (blue), iNKT2 (red), and iNKT17 (green) cells. (B) Dot-plot showing the proportion of cells (dot size) and the scaled (z score) gene expression of genes encoding for the iNKT sublineage-defining transcription factors T-Bet (Tbx21), PLZF (Zbtb16), and RORγT (Rorc). (C) Single‐cell heatmap representing the 10 most highly differentially expressed genes in thymic iNKT cell subsets. Expression for each gene is scaled (z scored) across single cells. (D) Normalized counts of Icos, Pdcd1, and Cd4 RNA expression. (E) FACS-sorting strategy for isolation of iNKT sublineages based on surface molecules starting from CD19^–, CD8a^–, CD62L^–, TCRγδ^–, GR-1^–, Ter119^–, and CD24^– cells. Cell purity of FACS-sorted iNKT sublineages assessed by intranuclear staining for the transcription factors PLZF and RORγT. (F) Heatmap representing the 10 most highly differentially expressed genes identified in scRNA-seq analysis in bulk RNA-seq analysis performed on sorted populations. Expression for each gene is scaled (z scored) across single rows.

Figure 2.

Analysis of the chromatin accessibility landscape in iNKT sublineages. (A) Volcano plots showing significance and log₂ fold change of ATAC-seq peaks in pairwise comparisons between iNKT sublineages. Peaks colored in gray on volcano plots indicate a log₂ fold change of −2 to +2 and/or an adjusted P value >.05. Peaks with a log₂ fold change of less than −2 or >2 and an adjusted P value <.05 are colored according to the iNKT sublineage represented (iNKT1, blue; iNKT2, red; and iNKT17, green). Vertical dotted lines on volcano plots indicate a log₂ fold change of 2; horizontal dotted line indicates an adjusted P value of .05. (B) Heatmap showing clusters for the top 2500 varying ATAC-seq peaks. Colors indicate z score of reads in each peak scaled to the mean across all iNKT cell subsets and replicates.

Figure 3.

iNKT1, but not iNKT2 and iNKT17, cells display cytotoxic potential. (A) Genome tracks showing a comparison of ATAC-seq and RNA-seq profiles for genes encoding cytotoxic molecules in iNKT1 (blue), iNKT2 (red), and iNKT17 (green) cells. Data are merged from 2 biological replicates. Black boxes highlight differentially accessible ATAC-seq peaks. (B) In vitro cytotoxicity by iNKT sublineages against α-galactosylceramide (αGalCer)-loaded CD1d-transduced A20 lymphoma (A20-CD1d) cells after 24 hours of coculture (effector:target ratio of 4:1). Represented data are pooled from 3 independent experiments performed in triplicate. (C) Survival after A20-CD1d (2 × 10e4) cells intravenously injected into sublethally irradiated (4.4 Gy) wild-type BALB/c mice treated with iNKT cell subsets (5 × 10e4 cells; iNKT1, blue; iNKT2, red; iNKT17, green) or untreated (black). Results are pooled from 2 independent experiments with a total of 10 mice per group. Survival curves were plotted by using the Kaplan-Meier method and compared by using a log-rank test. P values are indicated when significant.

Figure 4.

iNKT sublineages display different in vitro suppressive potential. (A) Genome tracks showing a comparison of ATAC-seq and RNA-seq profiles for genes encoding cytokines in iNKT1 (blue), iNKT2 (red), and iNKT17 (green) cells. Data are merged from 2 biological replicates. Black boxes highlight differentially accessible ATAC-seq peaks. In vitro suppression assay of indicated iNKT sublineages. (B-C) Representative profiles (B) and summary (C) of violet tracer dilution as well as CD25 and ICOS expression of CD4 T cells stimulated with anti-CD3/anti-CD28 activation beads in the presence or absence of FACS-sorted iNKT1, iNKT2, and iNKT17 cells. Data are representative of 2 independent experiments. MFI, mean fluorescence intensity.

Figure 5.

iNKT2 and iNKT17 cells protect from GVHD, whereas iNKT1 cells have no suppressive potential. (A) BALB/c recipient mice were irradiated with 2 × 4.4 Gy, followed by transplantation of 4 × 10⁶ TCD-BM cells and 1.0 × 10⁶ Tcon from luc⁺ FVB/N donor mice. In addition, 5 × 10⁴ FACS-sorted iNKT1, iNKT2, or iNKT17 cells from FVB/N donors were transferred together with the graft. Shown are the survival curve (left graph), weight (middle graph), and GVHD score (right graph). Data are pooled from 3 independent experiments with a total of 15 mice per group (except irradiation control, n = 10). Error bars indicate standard error of the mean. (B) BALB/c recipient mice were transplanted as aforementioned, and 1 × 10⁴ FACS-sorted iNKT1, iNKT2, or iNKT17 cells from FVB/N donors were transferred together with the graft. The survival graph is depicted on the left; the data are pooled from 3 independent experiments with a total of 20 mice per group (except total body irradiation [TBI, n = 10] and BM [n = 15]). For weight (middle graph) and GVHD score (right graph), data are pooled from 2 independent experiments with 10 mice per group. Error bars indicate standard error of the mean. (C) Serum Th1/Th2 cytokine levels (IFN-γ, TNF-α, IL-2, IL-6, IL-18, IL-5, and IL-13) at day 7 from transplanted mice are shown (n = 6-8 in each group). P values are indicated when significant. *P < .05; **P < .01; ***P < .001. BMT, BM transplantation.

Figure 6.

Adoptive transfer of distinct iNKT subsets differentially affect CD4 and CD8 Tcon transcriptome during murine acute GVHD. RNA-seq analysis of CD4 (A,C) and CD8 (B,D) Tcon recovered at day 7 post-HCT from animals treated with the indicated iNKT sublineages (untreated, gray; iNKT1, blue; iNKT2, red; iNKT17, green). (A-B) Principal component analysis of transcriptome based on the top 500 differentially expressed genes across all samples. (C-D) Heatmap and hierarchical clustering based on the 500 most highly differentially expressed genes across all samples. Immune-related genes are highlighted. Expression for each gene is scaled (z scored) across single rows. Analysis was performed on 3 biological replicates from 2 independent experiments.

Figure 7.

iNKT2 and iNKT17 cells inhibit IL-18Rα and CD49d expression and prevent tissue damage during murine acute GVHD. IL-18Rα (A) and CD49d (B) expression at the surface of CD4 (upper panels) and CD8 (lower panels) Tcon recovered at day 7 after HCT from spleen of mice treated with different iNKT sublineages (untreated, gray; iNKT1, blue; iNKT2, red; iNKT17, green). Representative FACS histograms (left panels) and summary bar plots (right panels) are shown. Data are pooled from 2 independent experiments with 3 to 5 mice per group. (C) Representative photomicrographs of hematoxylin and eosin–stained sections of small intestine and colon (×200) collected 7 days after transplantation from indicated groups. GVHD tissue damage manifests as inflammation (open arrow), crypt apoptosis (solid arrow), and crypt loss (solid asterisk). Histopathologic GVHD score (consisting of scores for crypt apoptosis and inflammation) for small intestine and colon for the indicated groups. Data are pooled from 2 independent experiments with a total of 6 mice per group. P values are indicated when significant.