Abundant PD-L1 expression in Epstein-Barr Virus-infected gastric cancers.

Charles S Rabkin,Tania Fleitas,M Constanza Camargo,Scott J Rodig,Sarah Derks,Xiaoyun Liao,Enrico Solcia,Anna M Chiaravalli,Gordon J Freeman,Fausto Sessa,Adam J Bass,Xinsen Xu

doi:10.18632/oncotarget.9076

Abstract

Gastric cancer (GC) is a deadly disease with limited treatment options. Recent studies with PD-1 inhibition have shown promising results in GC, but key questions remain regarding which GC subclass may respond best. In other cancers, expression of the PD-1 ligand PD-L1 has been shown to identify cancers with greater likelihood of response to PD-1 blockade. We here show with immunohistochemistry that Epstein-Barr Virus (EBV)+ GCs (n = 32) have robust PD-L1 expression not seen in other GCs. In EBV+ GC, we observed PD-L1 staining in tumor cells in 50% (16/32) and immune cells in 94% (30/32) of cases. Among EBV-negative GCs, PD-L1 expression within tumors cells was observed only in cases with microsatellite instability (MSI), although 35% of EBV-/MSS GCs possessed PD-L1 expression of inflammatory cells. Moreover, distinct classes of GC showed different patterns of PD-L1+ immune cell infiltrations. In both EBV+ and MSI tumors, PD-L1+ inflammatory cells were observed to infiltrate the tumor. By contrast, such cells remained at the tumor border of EBV-/MSS GCs. Consistent with these findings, we utilized gene expression profiling of GCs from The Cancer Genome Atlas study to demonstrate that an interferon-γ driven gene signature, an additional proposed marker of sensitivity to PD-1 therapy, were enriched in EBV+ and MSI GC. These data suggest that patients with EBV+ and MSI GC may have greater likelihood of response to PD-1 blockade and that EBV and MSI status should be evaluated as variables in clinical trials of these emerging inhibitors.

Highlights

Gastric cancer (GC) is a deadly disease that, despite its molecular heterogeneity, has been largely treated through uniform approaches
In order to determine whether PD-L1 expression is restricted to 9p24.1 amplified EpsteinBarr Virus (EBV)+ GCs we collected a sample series of 12 EBV+ and 10 EBV negative GCs (2 microsatellite instable (MSI), 3 genomic instable (GS), 5 chromosomal instable (CIN)) from the The Cancer Genome Atlas (TCGA) study, using cases in which we were able to retrieve tissue slides [4]
When we broke out the EBV negative GC by MSI status we found that MSI GCs had PD-L1+ tumor and immune cells in 33% (5/15) and 46% (7/15) of cases respectively, which was higher than EBV negative microsatellite stable (MSS) GCs that had no PD-L1 positive tumor cells (0% (0/34), P < 0.001) and PD-L1 positive immune cells in only 35% of cases (12/34, P < 0.001)

Summary

Introduction

Gastric cancer (GC) is a deadly disease that, despite its molecular heterogeneity, has been largely treated through uniform approaches. As new profiling approaches and molecular characterization efforts enable better subclassification of GC, one hope is that these subclasses may help guide optimal selection of therapy, both cytotoxic and biologic therapies. There is great enthusiasm regarding use of novel immune checkpoint inhibitors in cancer therapy, most notably those targeting the programmed cell death protein 1 (PD-1) pathway. As only a portion of patients respond to PD-1 inhibition, key questions remain regarding both possible biomarkers that could guide the use of PD-1 inhibitors as well as the potential that distinct biologic classes of GC may differ in sensitivity to these checkpoint inhibitors

Objectives

Methods

Results

Conclusion