Senescent Tumoral HLA-E Reshapes Microenvironment through Impairing NK Cell-Dendritic Cell-T Cell Network in Malignant Pleural Effusion from Lung Cancer

Abstract

Background: Malignant pleural effusion (MPE) is ominous in lung cancer patients. However, comprehensive studies of both innate and adaptive immune responses within the pleural tumor microenvironment remain limited. Methods: We collected samples from patients with heart failure and lung cancer-MPE. By single-cell RNA sequencing, we analyzed alternations in cancer cells, NK cells, DCs, and T cells. Key cytokines involving in cell-cell interactions were quantified using Luminex or ELISA, while HLA-E and aging markers were assessed via immunohistochemistry. Results: Our findings revealed that CD56⁺CD16⁺ and CD56⁻CD16⁻ NK cells exhibited reduced cytotoxicity, mainly through HLA-E-expressing senescent cancer cells interacting with NK cells inhibitory receptor, leading to NK cell dysfunction and reduced XCL2 expression, which might impair cDC1 recruitment. Consequently, aDC2 cells evolved into exhausted phenotype, resulting in inadequate T cell activation. In CD8 T cells, transcription factors such as FOXO1 contributed to diminished cytotoxicity. Despite presence of GZMA CD4 T cells, their cytotoxicity was suppressed in MPE. Th1-like and Th2-like regulatory T cells further inhibited CD4 T cell responses. Key molecules, CXCL16, BAG6, and IL-7, bridging innate and adaptive immunity conferred poor prognosis. Conclusions: Our study demonstrates that senescent cancer cells promote immunoevasion through HLA-E, suppressing NK cell cytotoxicity, impairing DC function, and disrupting T cell activation. Cell-cell interaction and imbalanced Th1/Th2 contribute to microenvironmental remodeling, driving disease progression. These findings provide insights into the immunological landscape and therapeutic targets for intervention.

Keywords: HLA-E; dendritic cell; nature killer cell; pleural metastasis; senescent cancer cell.

Figure 1

Landscape of NK cell subsets in MPE. (A) t-distributed stochastic neighbor embedding (tSNE) projection showing three clusters of NK cells in HP (n = 4) and LCP (n = 5). (B) Gene expression of NK cell surface markers: CD56 (NCAM1) and CD16 (FCGR3A). (C) Comparison of ratio in total cells between HP and LCP. (D) Heatmap of activating and inhibitory receptor expressions in NK subsets. (E) Cytotoxic and inflammatory profiles across three subsets of NK cells. (F) Tumor antigen-receptor interactions between cancer cells and two major subsets of NK cells in MPE. (G) Violin plot of HLA-E expression in cancer cells. (H) Quantification of XCL2 expression in pleural fluid was assessed by ELISA, comparing HP with LCP samples. * p < 0.05. HP: pleural effusion of heart failure; LCP: pleural effusion of lung cancer.

Figure 2

Deconvolution of HLA-E positive cancer cells in LCP. (A) t-SNE plot presenting cancer cells from MPE of lung cancer patients. (B) Differential expression of HLA-E in pleural metastatic cancer cells. (C) Heatmap comparing distinct gene expression profiles between HLA-E positive and HLA-E negative cancer cells. (D) Cell cycle arrest and senescence-weighted scores in the HLA-E positive and HLA-E negative cancer cells. (E) Heatmap illustrating senescence-associated secretory phenotype gene expression in HLA-E positive and HLA-E negative cancer cells. (F) Enrichment plots reveal the senescence phenotype associated with the HLA-E positive cluster. (G) The expression of senescence-related genes positively correlated with HLA-E within the HLA-E positive subset. (H) Immunohistochemical analysis of p21 and HLA-E expression in pleural metastatic cancer lesions. Representative images indicate the localization and relative expression levels of p21 and HLA-E in tumor cells within the pleura.

Figure 3

Distinct phenotypes of aDC2 within MPE. (A) tSNE map demonstrating 7 subsets of DCs. (B) Proportional distribution of 7 DC clusters in HP and LCP. (C) Violin plots illustrated functional differences (phagocytosis, antigen processing and presentation) in each DC subtype. (D) Partition-based graph abstraction (PAGA) analysis illustrating the developmental transitions among DC subsets. Connectivity strength between clusters reflects transcriptional similarity and suggests possible lineage relationships or maturation pathways across DC states. (E) PAGA paths including upregulated genes were involved in the aDC2 transition. (F) Heatmap showing DEGs in aDC2 subset from pleural effusion of HP and LCP. (G) Violin plots depicting upregulated genes in CD274 negative and CD274 positive aDC2. (H) Quantification of CCL22 levels in pleural effusion samples from HP and lung cancer patients. CCL22 levels in pleural effusion were measured by ELISA. * p < 0.05.

Figure 4

Profile of dysfunctional CD8 T cells. (A) tSNE plot illustrating the distribution of 9 transcriptionally distinct CD8 T cell subsets identified from pleural effusion of HP and lung cancer patients. (B) Different subsets of CD8 T cells in the pleural effusion of HP and LCP. (C) The proportion of various subsets of CD8 T cells in pleural effusion of HP and LCP. (D) Violin plots showing the cytotoxic scores among CD8 T cell subsets. Scores were calculated based on the expression of cytotoxic genes (GZMA, GZMB, GZMH, GZMK PRF1, GNLY etc). (E) Pseudotime trajectory analysis of three effector memory CD8 T cell (CD8 TEM) subsets. (F) key genes contributing to the pseudotemporal transition of CD8⁺ TEM subsets. (G) Genes of step-up trend during the transition from CXCR3⁺ to CCL4⁺ CD8 TEM.

Figure 5

Functional impairment of cytotoxic CD4 T cells. (A) Overview of 4 CD4 T cell clusters presented as a tSNE plot. (B) Localization of four CD4 T cell subsets in the pleural effusion of HP and LCP. (C) Violin plots showing the expression of various genes regulating the cytotoxicity ability of different subsets of CD4 T cells. (D) Bar plots comparing the percentage of CD4 T subsets of HP and LCP. (E) GSCA analysis predicting the overall survival rate of lung cancer patients based on the top 100 genes of GZMA CD4 T cells. (F) Representative flow cytometry plots show the gating strategy and the presence of CD45⁺CD8⁻CD4⁺GZMA⁺ T cells in MPE from lung cancer patients. These cells were identified based on surface markers CD45, CD4, and CD8, along with intracellular expression of granzyme A (GZMA). (G) Violin plots representing cytotoxicity scores and GZMA expression in Cycling GZMA CD4 and GZMA CD4 T cells in the pleural effusion of HP and LCP. (H) Heatmap depicting upregulated transcription factors in Cycling GZMA CD4 and GZMA CD4 T cells in the MPE of LCP.

Figure 6

Tregs were associated with poor survival in LCP patients. (A) Box plot showing the percentage of Treg (CD4⁺CD25⁺FOXP3⁺) in the pleural effusion of HP and LCP by flow cytometry. (B) Heatmap illustrating the expression of helper T-specific transcription factors in Tregs. (C) Violin plots depicting CTLA4 and TIGIT expression levels in Treg in the pleural effusion of HP and LCP. (D) BATF expression in Tregs in the pleural effusion of HP and LCP patients. (E) Differences in Th1 cytokines (IL-12p70, IFN-γ, and TNF-α) between the pleural effusion of HP and LCP. (F) Levels of Th2 cytokines (IL-10, IL-4, and IL-5) in the pleural effusion of HP and LCP. (G) Levels of inflammatory cytokines (IL-6 and IL-8) in the pleural effusion of HP and LCP. (H) Th1/Th2 cytokine ratio comparing TNF-α to IL-10, IL-4, and IL-5 in the pleural effusion of HP and LCP. (I) Kaplan-Meier survival analysis of high and low TNF-α/IL-10 ratios. HR: hazard ratio; ns: not significant; * p < 0.05; ** p < 0.01; **** p < 0.001.

Figure 7

Cell-cell interactions predict poor survival biomarkers in MPE. (A) Number and weight scores of interactions between DCs, LA-TAMs, NK cells, T cells, mesothelial and cancer cells. (B) Ligand-receptor interactions between HLA-E positive cancer cells and NK cells. (C) Interactions between aDC2 and various subsets of CD4 and CD8 T cells. (D) Heatmap showing ligand-receptor interactions from LA-TAMs to various subsets of CD8 T cells. (E) Relationship between Tregs and CD4/CD8 clusters as well as cancer cells. Expression levels of CXCL16 (F), BAG6 (G), and IL-7 (H) in the pleural effusion of HP and LCP. Survival analysis of BAG6 (total: right; EGFR mutation: left) (I) and IL-7 (J) in lung cancer patients with MPE. HR: hazard ratio; ns: not significant; *, p <0.05.

Figure 8

Mechanisms for lung cancer pleural metastasis. DC, dendritic cell; HLA-E, human leukocyte antigen; NK, nature killer cell; PD1, programmed cell death protein 1; Th, helper T cell;