The CXCL16-CXCR6 axis in glioblastoma modulates T-cell activity in a spatiotemporal context

Abstract

Introduction: Glioblastoma multiforme (GBM) pathobiology is characterized by its significant induction of immunosuppression within the tumor microenvironment, predominantly mediated by immunosuppressive tumor-associated myeloid cells (TAMCs). Myeloid cells play a pivotal role in shaping the GBM microenvironment and influencing immune responses, with direct interactions with effector immune cells critically impacting these processes.

Methods: Our study investigates the role of the CXCR6/CXCL16 axis in T-cell myeloid interactions within GBM tissues. We examined the surface expression of CXCL16, revealing its limitation to TAMCs, while microglia release CXCL16 as a cytokine. The study explores how these distinct expression patterns affect T-cell engagement, focusing on the consequences for T-cell function within the tumor environment. Additionally, we assessed the significance of CXCR6 expression in T-cell activation and the initial migration to tumor tissues.

Results: Our data demonstrates that CXCL16 surface expression on TAMCs results in predominant T-cell engagement with these cells, leading to impaired T-cell function within the tumor environment. Conversely, our findings highlight the essential role of CXCR6 expression in facilitating T-cell activation and initial migration to tumor tissues. The CXCL16-CXCR6 axis exhibits dualistic characteristics, facilitating the early stages of the T-cell immune response and promoting T-cell infiltration into tumors. However, once inside the tumor, this axis contributes to immunosuppression.

Discussion: The dual nature of the CXCL16-CXCR6 axis underscores its potential as a therapeutic target in GBM. However, our results emphasize the importance of carefully considering the timing and context of intervention. While targeting this axis holds promise in combating GBM, the complex interplay between TAMCs, microglia, and T cells suggests that intervention strategies need to be tailored to optimize the balance between promoting antitumor immunity and preventing immunosuppression within the dynamic tumor microenvironment.

Keywords: CXCL16; CXCR6; glioblastoma; immunotherapy; myeloid cells.

Figure 1

The CXCL16/CXCR6 axis is partitioned into myeloid/T-cell populations in GBM based on scRNAseq analyses. In (A–C), the expression of CXCL16 and CXCR6 was examined in the Neftel et al. human scRNA-seq GBM datasets. In (B – Left panel), the DEGs between CXCR6⁺ and CXCR6^- within the T-cell cluster (defined by CD3ϵ expression) in GBM are shown. In (B, right panel), the differential expression of genes in the CXCL16⁺ vs. CXCL16^- clusters within the myeloid cluster (defined by CD14 expression) in GBM. In (C), the area under the curve (AUC) for gene sets associated with T-cell activation and exhaustion was compared between CXCR6⁺ and CXCR6^- T cells in human GBM. In (D, E), multiplex immunohistochemistry was performed on 22 newly diagnosed and recurrent GBM samples. The correlation of T-cell infiltration is based on the expression of CXCL16⁺ cells in (D) and immunosuppressive/microglia in (E). In (F–H), scRNA analysis of CT-2A tumor-bearing mice is shown. In (F), UMAP projection and single R annotation of cellular infiltrates in CT2A tumors are shown. In (G), the expression of CXCL16 and CXCR6 was among the genes in these clusters. In (H), the DEGs between CXCR6+ and CXCR6- T cells in CT-2A tumors are shown. In (D, E), for 22 patients, both newly diagnosed and recurrent tumors were analyzed via simple linear regression. P values for all comparisons are directly stated in the figures.

Figure 2

Characterization of the CXCL16/CXCR6 axis in tumor-infiltrating immune cells. In (A, C), we analyzed immune cells isolated from mouse tumors by flow cytometry 14 days after tumor implantation. (B, D) are from n=5, showing the surface and intracellular levels of CXCL16 per cell type in tumors and spleens. Further analysis of PD-L1 and CXCL16 coexpression in tumor-associated macrophages and monocytes in the spleen and tumor tissue are shown in (E) and were analyzed by flow cytometry. In (F), we treated TAMCs under different conditions and analyzed CXCL16 expression by flow cytometry. The significance of differences in (B, D–F) was analyzed by one-way ANOVA; ****p<0.0001. ns = p>0.05, *p<=0.05, ***p<=0.001.

Figure 3

CXCR6 affects CD8⁺ T-cell migration and functions in tumors. (A, B) CD4⁺ and CD8⁺ T cells and TAMCs from tumors and draining lymph nodes (n=5) were isolated and analyzed by flow cytometry 14 days after tumor injection. In (C), the ratio of infiltrating lymphocytes to myeloid cells was analyzed by flow cytometry. The absolute numbers of infiltrating CD4+ and CD8+ T cells and the total number of TAMCs are shown in (D) and were analyzed via flow cytometry. In (E–H), naïve CD8+ T cells from control and CXCR6 KO mice were adoptively transferred and injected at a 1:1 ratio i.v. 5 days later. Tissues were harvested for flow cytometry analysis. Before and after the transfer, the percentages of CD45.1+ versus CD45.2+ CD8⁺ T cells were analyzed via flow cytometry (E, bottom panel). The data shown in (F) are from n=5, and the percentages of transferred CD8⁺ T cells in the organs were analyzed by flow cytometry. The expression of exhaustion markers (PD-1, Lag3, and Tim3) in CXCR6-deficient CD8⁺ T cells shown in (G) was analyzed via flow cytometry after the adoptive transfer experiment. In (H), flow cytometry was used to measure the inflammatory response of GzmB/IFN-γ/TNF-α-expressing tumor-isolated WT and CXCR6 KO CD8⁺ T cells 14 days after injection. Significance in (A, B) was analyzed by one-way ANOVA; (C–H) were analyzed by Student’s t test, *p<=0.05, **p<=0.01, ***p<=0.001, ****p<=0.0001.

Figure 4

CXCR6 deficiency prevents CD8⁺ T-cell–TAMC interactions and increases CD8⁺ T-cell activity and survival. (A) CD8⁺ T cells from control or CXCR6 KO mice were isolated, activated in vitro, labeled with a fluorescent-tracker dye, and then cocultured at a 1:1 ratio with tumor-associated myeloid cells (TAMCs). After 90 minutes, the slides were washed and analyzed via epifluorescence microscopy. The data shown in (A, B) are from 4 replicates, and two images per slide were analyzed. A paired Student’s t test was performed, and p<0.001 = **. (C) Tumors were isolated on the 19th day after the adoptive transfer experiment, and the data in (D) show the CD45.1 and CD45.2 ratios in tumors without and after transfer, as analyzed by flow cytometry. In (G–I), flow cytometry analysis of the coexpression of CD44 with CD62L, PD1, GzmB, INF-γ, and TNF-α in CD45.1⁺ or CD45.2⁺ CD8⁺ T cells is shown. The CXCR6 knockout group treated with PD-1 blockade achieved 50% survival. (J). Median survival: control = 24 days, control + PD-1 blockade = 36.5 days, CXCR6 KO = 30 days, CXCR6 KO + PD-1 blockade = 65 days; p= 0.002 = ** Log-rank test). (K) Surviving mice were rechallenged via tumor implantation. Tumor rejection in both groups represented the generation of immunologic memory against tumor antigens. Significance in (D–I) was analyzed by paired Student’s t test; *p<=0.05, **p<=0.01, ***p<=0.001.