e14078 Background: Immune checkpoint inhibitors have brought great breakthroughs in cancer therapy. Activated immune response is known to be the prerequisite for exerting immunotherapy efficacy. Epstein-Barr virus (EBV) infection is associated with longer survival in gastric cancer (GC) patients due to enhanced anti-tumor immune response, and therefore it was reportedly played an important role in modulating immune checkpoint blockade therapy efficacy. However, molecular dimensions underlying the good response to immune checkpoint inhibitors in presence of EBV infection are still unclear. The aim of this study is to identify a gene signature related to EBV induced anti-tumor immune response, and select a tag gene from this signature to predict which patients are most likely to benefit from immune checkpoint blockade therapy. Methods: Two large transcriptome datasets from Gene Expression Omnibus(GEO) database (GSE51575 and GSE62254) were used to screen gene signature for EBV infected gastric cancer tissues. We further selected genes that showed a trend towards differential co-expression independent of EBV infection status. The tag gene of this differential co-expression signature was finally identified by bioinformatics analysis. To make an external validation, we performed RNA sequencing in 20 colorectal caner (CRC) tissues and 20 GC tissues, respectively. Meanwhile, tissue microarrays of CRC cohort (36 paired tumor and normal tissues) and GC cohort (75 paired tumor and normal tissues) were used to analyze the association of SLAMF8 with CD8 protein expression by immunohistochemistry (IHC). Results: Analysis of GEO datasets indicated 788 genes as feature gene cluster for EBV-positive gastric cancer, from which 290 genes were selected to be characterized by differential co-expression in either EBV-positive or EBV-negative gastric cancers. SLAMF8 was identified as the tag gene for this differential co-expression signature. This signature, tagged by SLAMF8, was successfully validated by our RNA sequencing data in presence of its good performance in dividing CRC and GC patients into two subsets. Moreover, we observed a significant association between SLAMF8 and CD8 expression in our CRC and GC tissue samples, in terms of either mRNA or protein level. Conclusions: SLAMF8, a potential indicator for T cell‐mediated immune response induced by EBV infection, may be served as a biomarker for individualized immune checkpoint blockade therapy in gastrointestinal cancer. Further SLAMF8 guided drug sensitivity tests are warranted to validate our results.
Targeting the myeloid-derived suppressor cell compartment for inducing responsiveness to immune-checkpoint blockade is best limited to specific subtypes of gastric cancer
