167 Exploration of different immunogenomic TMEs in gastric cancer based on HER2 and PD-L1 expression

Background Recent results from the KEYNOTE-811 study showed that first-line pembrolizumab was superior to placebo with respect to progression-free survival (PFS) in combination with trastuzumab and chemotherapy in patients with HER2 positive metastatic gastric cancer (GC). Interestingly, in the subgroup analysis, the PFS benefit with pembrolizumab was significant in patients with tumors with PD-L1 CPS ≥1 (positive), whereas there was no benefit in the small subgroup of patients with tumors with CPS < 1 (negative). Therefore, this study is to investigate immune-genomic differences in the tumor microenvironment according to the composition of surrounding cells and the expression status of HER2 and PD-L1. Methods Utilizing single-cell RNA sequencing and TCR sequencing, 14 HER2-positive and 18-negative primary GC samples were examined, PD-L1 expression stained with pharmDx IHC 22C3 and analyzed by CPS. Among the HER2-positive group, there were 4 samples in the PD-L1 negative and 10 samples in the PD-L1 positive. In the HER2 negative group, there were 6 samples in the PD-L1 negative and 12 samples in the positive group. Results First, when analyzing the distribution of immune cells, the distribution of CD4 Treg (p=0.084) and CD8 exhausted T (Tex) (p=0.036) cells was higher in the HER2-positive group compared to the negative. When regrouped according to PD-L1, only the HER2 positive group showed a tendency to increase the observation of CD4 Treg and CD8 Tex cells compared to positive in PD-L1 negative. Furthermore, TCR analysis revealed that Tex TCR in the HER2-positive and PD-L1 negative group was more hyperexpanded than in the HER2-negative group. In HER2-positive patients, hyperexpanded Tex cells exhibited significantly higher exhaustion and cytotoxic scores in the PD-L1 negative group, indicating a higher prevalence of terminally exhausted T cells in this group. In addition, this group had higher expression of immune checkpoint proteins, such as PDCD1, LAG3, BTLA, VISTA and TIGIT. On the other hand, in the HER2-negative group, the expression pattern of immune checkpoint molecules was similar regardless of PD-L1, but some samples with high expression of these proteins were observed in the PD-L1-negative group. Conclusions The high proportion of CD4 Tregs, CD8 Tex and the elevated exhausted score in the subgroup with PD-L1 negative in HER2 positive GCs suggesting immune exclusive environment compared to HER2 and PD-L1 coexpressing group. Considering increased expression of immune checkpoint proteins other than PD-1/PD-L1, for patients with HER2-positive and PD-L1 negative, combination therapy with other immune checkpoint inhibitors rather than anti-PD-1/PD-L1 may be suggested.