In this study, we investigated differential DNA methylation status in well-differentiated small intestinal and appendiceal NETs with different molecular subtypes. We found 901 genes with differentially methylated promoter regions in the study’s NETs having 18LOH, and these genes were enriched in tumor-related pathways, including "GPCR downstream signaling," "neuroactive ligand-receptor interaction," “beta defensins,” and “hemostasis”. Most of the genes were hypomethylated, broadly suggesting transcriptional activation of these genes. The lncRNA, which made up 22.8% of the differentially methylated genes, have been shown to regulate cancer genes and may represent additional disease pathways 34–37.
The GPCR pathway represents a large family of cell-surface molecules regulating the signal transmission for multiple cellular functions. GPCRs are also the most common class of therapeutic targets, with approximately 700 FDA-approved drugs targeting 128 GPCRs38,39. Pertinent to the findings of this study, GPCRs play multiple roles in cancer development and have been the focus of past studies to define differences in neuroendocrine tumors, primarily those originating from the small intestine versus pancreas and lung 40 41. Our significantly enriched GPCR pathwayssuggests that expression of GPCR pathways are controlled through methylation events which in turn drivestumor progression a subset of GINETs 41,42. Being neuroendocrine tumors, it makes sense that the ‘neuroactive ligand-receptor interaction’ pathway is involved, and accordingly, is part of the GPCR downstream signaling pathway. We describe hypermethylation of one such neuroactive ligand, the SST gene encoding somatostatin, in 18LOH tumors relative to NoCNV and MultiCNV. Somatostatin is the ligand for SSTRs, G-protein coupled somatostatin receptors that are overexpressed in a subset of GINETs and the target of somatostatin analog therapies. Examining the clinical response to somatostatin analog therapies relative to SST methylation state may provide guidance on more targeted treatment strategies.
Our survival analysis with the three molecular subgroups based on chromosomal changes showed there was no significant difference between subgroups (p = 0.41; supplementary Fig. S3) However, the ability to detect such differences is limited by the sample size. We did identify 24 genes with methylation levels associated with survival outcomes specifically in 18LOH tumors, potentially explaining why there is disagreement in the literature on survival benefits for 18LOH tumors. The most significant association was with TRHR. Hypomethylation of the TRHR promoter may be associated with 18LOH tumor progression and survival outcomes. The mechanism behind this survival difference is not obvious but may represent a constellation of methylation events that are clinically relevant, potentially as a pharmacologic target for small intestinal NET patients.
Selection of tumors from the cancer registry included ICDO codes for the small intestine as well as the appendix because we believe there is cross-over in the etiology of these tumors. Three of the 47 tumors examined for methylation were from the appendix. All three were classified as NoCNV subtype, and hierarchical clustering of their methylation profile was fully integrated with the other NoCNV samples from the small intestine, suggesting a similar etiology. Additionally, we have observed neuroendocrine tumors of the appendix in families with multiple NETs of the small intestine, which suggests there is overlap in the genetic etiology.
Limitations of the study included small sample size, methylation arrays were not run on matched normal tissues and a lack of treatment data for the NET patients. Also, the methylation changes were not validated with RNA expression profiling, but rather relied on published results in different tissues. Strengths are that the samples tested were representative of all tumors reported in the state for that period of time, and that the frequency of molecular subtypes were similar to previous reports 9,11−13.
In conclusion, by separating small intestine and appendiceal NETs into different molecular subtypes based on chromosomal changes, we were able to define major pathways that are differentially methylated. Two relevant pathways were the GPCR and the overlapping neuroactive ligand-receptor interaction. One gene that is differentially hypermethylated in 18LOH tumors is SST, encoding somatostatin. Somatostatin analogs are a primary treatment for GINETs, targeting overexpressed somatostatin receptors on the tumors. Somatostatins are also used for GINET visualization, radiotherapy, and repressing carcinoid-syndrome side effects. GPCR may be important for survival in small intestinal NETs and are targetable by drugs in some cases. In future work, methylation status in these genes can be explored as a prognostic and/or predictive biomarker to predict responses to SSA’s and for careful selection of patients for appropriate treatments. Future studies that incorporate detailed histologic characterization, GPCR expression, and treatments will be important. The molecular characterization of GINETs can lead to novel prognostic and predictive biomarkers to better inform treatments and may improve survival of patients with GINETs.