Investigation into the immune microenvironment of gastroenteropancreatic high-grade neuroendocrine carcinoma.

Abstract

e16244 Background: Gastroenteropancreatic high-grade neuroendocrine carcinomas (GEP HG-NEC) are aggressive cancers with poor survival and limited therapies. Using high-throughput gene expression analysis, we report the tumor-immune system interactions within the tumor microenvironment (TME), differentially expressed genes based on age, tumor location, survival, and objective response rate (ORR) to systemic therapy with the aim to drive future effective therapy research. Methods: We evaluated patients diagnosed with GEP HG-NEC from 2010-2021 for a pilot study and included those with adequate tumor samples. RNA was analyzed using the NanoString nCounter PanCancer IO 360 panel. Logistic regression modeling was used for hypothesis testing associated with best overall response. Cox model was used for testing overall survival (OS). One-sided P values were calculated per the hypothesized positive association between immune signatures and improved clinical outcomes. Results: 21 patients were identified. Mean age 64 years (yrs) with a 2:1 gender ratio (M=14; F=7). Most common primary locations - pancreas (PNEC = 8) and esophagus (3). ORR = 78.6%. TME trends were dampened anti-cancer immune response in patients >60 yrs old and <6-month OS on treatment. TME gene signatures of age >60 yrs (vs. <60 yrs), were indicative of fewer neutrophils, myeloid cells, CD8+ T-cells and CD45+ cells. Downregulated genes in the >60yr old population associated with dampened immune response - ICAM1, TNFAIP6, TWIST1, and PTGER4. A smaller natural killer (NK) cell gene signature was the only notable association with OS <6 months (vs. >6 month). Downregulated genes in >6-month population were HK2 and REN while upregulated genes included CX3CR1, ZAP70, KLRK1, XCL1/2, and MMP7 (immunostimulatory). Increased immune infiltration was associated with prolonged OS, with elevated gene signature of T-cells, lymphoid cells, mast cells, and cytotoxic cells. Downregulated genes related to poor OS were HNF1A and ESR1 (immunosuppressive). A greater potential for a tumor adaptive immune resistance/response were noted; age <60 yrs and prolonged OS having elevated TME PD-L1 and PD-1/TIGIT gene signatures respectively. PNEC was associated with elevated gene signatures of TME MHCII vs. non-PNEC. Upregulated genes in non-PNEC include CXCL2/3 while downregulated genes include TMEM173, IFI16 and RELN (diminished adaptive immune response). Conclusions: This pilot study identified trends in gene expression associated with clinical outcomes. While statistical significance was precluded by limited sample size, our results suggest trends toward increased TME immune cell infiltration in age <60yrs and prolonged OS, increased potential for adaptive immune responses in PNEC, and elevated NK cell infiltration in patients with >6 month OS. Rational drug development for affected patients should target these specific abnormalities.