Association of Claudin-1 overexpression with immune exclusion in nasopharyngeal carcinoma and its potential as a novel therapeutic target.

Abstract

151 Background: Claudin-1 (CLDN1) is a tight junction protein overexpressed in several cancers, where it is involved in the tumor microenvironment remodeling and stiffening of the tumor-stroma interface. This potentially prevents penetration of tumor infiltrating immune cells, in turn hindering the efficacy of immune checkpoint inhibitors (ICI). Given recent evidence supporting ICI efficacy in nasopharyngeal carcinoma (NPC), we investigated whether CLDN1 is overexpressed in NPC and its association with the tumor immune microenvironment (TME). Methods: 131 NPC patients treated at a single institution were prospectively recruited. TME profiling was performed using deconvolutional methods (CIBERSORT, quanTIseq) on tumor transcriptomes to estimate proportions of immune cell subsets, and expression of immune-related genes and pathways. Transcriptome profiling was performed on tumors with >70% cellularity and normal controls using Illumina Truseq (CA). We stratified tumors by their median CLDN1 expression and compared the immune profile of CLDN1high versus CLDN1low tumors. Cox regression model was used to analyze disease-free survival (DFS) as the primary clinical endpoint. Results: We observed CLDN1 overexpression in NPC versus normal tissues ( P=0.0046) that was independent of known prognostic clinical parameters e.g., TNM-stage and EBV DNA levels. Consistent with our hypothesis, CLDN1 expression was negatively correlated with total immune cell fraction in the tumor ( P<0.001). We observed reduced abundance of most immune subsets in CLDN1high tumors, notably CD8 T cells ( P<0.001), activated CD4 memory T cells ( P<0.001), and M1 and M2 macrophages ( P=0.048 and P<0.001). There was an overall downregulation of both stimulatory and suppressive GO immune-related gene-sets and genes in CLDN1high versus CLDN1low tumors, thereby corroborating our preceding observation of reduced immune infiltration in the TME in CLDN1 overexpressing NPC. Our findings were validated in a published dataset of 113 NPC tumors (Zhang et al., 2017). Of clinical relevance, reduced immune infiltration in the TME was prognostic for DFS; HR of CLDN1high, CD8 T celllow (stratified by median) NPC was 2.1 (95% CI:0.9-5.0) when compared with CLDN1low, CD8 T cellhigh NPC. Conclusions: Here, we report the overexpression of CLDN1 in NPC that is ubiquitous across disease stages and associated with an immune-low TME. This influence on the TME is associated with an inferior outcome of NPC patients. Given the role of the tumor TME in ICI response, our findings provide a rationale for future trials investigating combination of CLDN1 targeting with ICI to reverse this immune exclusion effect, potentially leading to improved outcomes.