The four major pathways in gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) including chromatin remodeling, DNA damage repair, activation of mTOR signaling, and telomere maintenance were mediated by some critical molecules and constituted critical processes of regulation in cancer-causing processes. However, the interplay and potential role of these pathway-related molecules in the tumor microenvironment of the primary and metastatic site remained unknown.
We systematically evaluated the mRNA expression of 34 molecules associated with the four pathways in 227 GEP−NEN samples from 5 datasets. We assigned the samples into two expression patterns of pathway-related molecules by an unsupervised clustering method. Subsequently, we explored the specific cell-related molecules, especially immune and stromal cells using the WGCNA method, based on differentially expressed genes (DEGs) responsible for the different patterns of pathway-related molecules, which provided a new method to qualify the pathway-related subtypes of individual tumors, then the PC_Score and PI_Score scoring systems were also constructed using obtained specific cell-related molecules. Furthermore, we performed the association of pathway-related subtypes with characteristics of immune landscape in primary and metastatic GEP-NENs.
We demonstrated that the specific pathway-related molecules (SMARCA4, MLH1, TSC1, ATRX, and ATR) were associated with cytolytic activity. Then we identified the two distinct patterns of pathway-related molecules, which were characteristic with a significantly distinct immune landscape. Using WGCNA, we also identified the fibroblast-related molecules, including ASPN, COL10A1, COL3A1, EDNRA, MYL9, PRELP, RAB31, SPARC, and THBS2, and immune-related molecules including CASP1, CCL5, CTSS, CYBRD1, PMP22, and TFEC. Based on these specific markers, we identified four distinct pathway-related subtypes, characterized by immune and fibrotic enriched (I/FE), immune enriched (IE), fibrotic enriched (FE), and immune and fibrotic desert (I/FD), of which I/FE was characteristic with the highest PC_Score and PI_Score whereas I/FD presents the opposite trend. I/FE was positively correlated with immune landscape of T-cell activation and immunosuppression. Furthermore, the I/FE marked GEP-NENs with increased immune activation scores (T-cell costimulation, MHC I presentation, and APC costimulation). Importantly, the four distinct pathway-related subtypes were not conserved in different tumor sites, because I/FE was lacking in the liver metastatic site even though IE, FE, and I/FD also could be observed in the metastatic site.
This study was the first to perform a comprehensive analysis of the four major pathways in GEP-NENs. We demonstrated the potential function of these pathway-related molecules in immune landscapes. Our findings indicated that the primary and metastatic GEP-NENs had distinct antitumor phenotypes. This work highlighted the interplay and potential clinical utility of these pathway-related molecules in GEP-NENs.