EBV-Infected NKTCL Upregulated MHC-II Imprinted T Cell Exhaustion By Interacting with LAG3

Background: Extranodal natural killer/T-cell lymphoma (NKTCL) is a highly aggressive malignancy pathologically associated with Epstein‒Barr virus (EBV) infection. Previous studies have found that EBV is the main contributor to shaping the immunosuppressive microenvironment of NKTCL. However, the mechanisms and functional consequences of EBV-infected tumor cells in modulating T cell immune responses remain unclear. Methods: In this study, we collected an extensive dataset comprising scRNAseq data from 10 treatment-naïve NKTCL patients (GSE203663) and bulk RNAseq data from 278 patients including 128 from OEP000498, 84 from GSE160119 and 66 from GSE90597. Seurat was employed to identify T cells and their subtypes within the tumor microenvironment. Additionally, we explored the ligand-receptor pairs between T cell subtypes and NK cells by using CellChat. Additionally, we employed the corrplot R package to evaluate the correlation of candidate genes expression. Results: We first delineate the transcriptional profiles of both normal and malignant NK cells in NKTCL patients, which were categorized into EBV-high and EBV-low based on the transcription level of EBV-encoded genes in tumors. We found that MHC-II-related molecules, including HLA-DRA, HLA-DRB, HLA-DQA, HLA-DQB, HLA-DPA, HLA-DPB, etc., were remarkably expressed in EBV-high malignant NK cells, significantly differing from normal and EBV-low malignant NK cells (log2FC>1, FDR<0.01). Interestingly, MHC-I-related molecules, such as HLA-A, HLA-B, HLA-C, and HLA-E, were substantially overexpressed in EBV-high normal NK cells. Considering the pivotal roles of EBV and MHC molecules in shaping the immune microenvironment, particularly T cell immunity, we next analyzed CD8+ T cells within the tumor microenvironment. A total of 9087 CD8+ T cells were grouped into eight clusters: C1_NAIVE (naïve T cells), C2_CM (central memory T cells), C3_MAIT, C4_EM (effector memory T cells), C5_EFF (effector T cells), C7_XCL(exhausted T cells with high XCL-1), C8_EX (exhausted T cells), and C9_PROLIFE. We observed an enrichment of exhausted CD8+ T cells (25.03%), including 15.22% in C8_EX and 9.81% in C7_XCL, in EBV-high tumors versus EBV-low tumors (13.54%). We further evaluated the expression of inhibitory molecules in tumor-infiltrating T cells, and the results revealed that TOX, HAVCR2, PDCD1, CTLA4, TIGIT and LAG3 were upregulated in CD8+ T cells from EBV-high tumors. Among these, LAG3 was the most markedly upregulated and expressed at a higher proportion, with expression levels reaching 54.02% in CD8+ T cells from EBV-high tumors compared to 42.16% in EBV-low tumors. Previous studies reported that MHC-II is the primary ligand that interacted with LAG3 on T cells. Thus, we conducted ligand-receptor interactions analysis between NK cells and CD8+ T cells in NKTCL. We observed the strongest interactions between LAG3 on CD8+ T cells and MHC-II molecules on EBV-high malignant NK cells, including HLA-DPA1, HLA-DQA1 and HLA-DRA, compared with normal NK cells and EBV-low malignant NK cells. Besides, the interactions between LAG3 and MHC-II molecules were markedly stronger than that between LAG3 and LGALS3. We further confirmed the interactions in the bulk RNAseq dataset by assessing the correlation between LAG3 and MHC-II molecules, as well as EBV transcription levels. We observed significant positive correlations between LAG3 and HLA-DPA1/DQB1(LAG3- HLA-DPA1: R=0.32, P=0.001; LAG3-HLA-DQB1: R=0.63, P<0.001), as well as between LAG3 and EBV (R=0.44, P<0.001). The above results suggested that NKTCL tumor cell-specific upregulated MHC-II molecules strongly interacted with LAG3 on CD8+ T cells, contributing to sustaining T cell exhaustion. Conclusions: The study revealed that MHC-II molecules were significantly upregulated in EBV-high NKTCL tumor cells, strongly interacting with LAG3, which was overexpressed on tumor-infiltrating CD8+ T cells. The interactions between LAG3 and MHC-II molecules contributed to T cell exhaustion in NKTCL. Consequently, LAG3 could serve as a promising therapeutic target for NKTCL.