Molecular Subtypes of Extra-pulmonary Neuroendocrine Carcinomas Identified by the Expression of Neuroendocrine Lineage-Specific Transcription Factors

Abstract

Extra-pulmonary neuroendocrine carcinomas (EPNEC) represent a group of rare and heterogenous neoplasms with adverse clinical outcome. Their molecular profile is largely unexplored. Our aim was to investigate if the major transcriptional drivers recently described in high-grade pulmonary neuroendocrine carcinomas characterize distinct molecular and clinical subgroups of EPNEC. Gene expression of ASCL1, NEUROD1, DLL3, NOTCH1, INSM1, MYCL1, POU2F3, and YAP1 was investigated in a series of 54 EPNEC (including 10 cases with mixed components analyzed separately) and in a group of 48 pulmonary large cell neuroendocrine carcinomas (P-LCNEC). Unsupervised hierarchical cluster analysis classified the whole series into four major clusters. P-LCNEC were classified into two major clusters, the first ASCL1/DLL3/INSM1-high and the second (including four EPNEC) ASCL1/DLL3-low but INSM1-high. The remaining EPNEC cases were sub-classified into two other clusters. The first showed INSM1-high and alternative ASCL1/DLL3 or NEUROD1 high expression. The second was characterized mainly by MYCL1 and YAP1 overexpression. In the ten cases with mixed histology, ASCL1, DLL3, INSM1, and NEUROD1 genes were significantly upregulated in the neuroendocrine component. Higher gene-expression levels of NOTCH1 and INSM1 were associated with lower pT stage and negative nodal status. Low INSM1 gene expression was associated with shorter overall survival in the entire case series (p = 0.0017) and with a trend towards significance in EPNEC, only (p = 0.06). In conclusion, our results show that EPNEC possess distinct neuroendocrine-lineage-specific transcriptional profiles; moreover, low INSM1 gene expression represents a novel potential unfavorable prognostic marker in high-grade NECs including those in extra-pulmonary location.

Keywords: Classification; Extra-pulmonary neuroendocrine carcinoma; INSM1; Prognosis; Transcriptional profile.

Fig. 1

Representative EPNEC cases from the series investigated. a–c EPNEC of large cell type, unknown primary, brain metastasis, positive for INSM1 protein (b) and chromogranin A (c), with focal dot-like pattern. d, e EPNEC of large cell type, unknown primary, lymph node metastasis, positive for synaptophysin (e) and with high Ki-67 index (f). g–i EPNEC, large cell type primary of the colon, positive for chromogranin A (i). l, m Mixed neuroendocrine/non-neuroendocrine neoplasm of the duodenum with adenocarcinoma component with mucin production (l and m) and small cell carcinoma component (n) (all original magnifications 200 × , except g 100 × and l 40 ×)

Fig. 2

Representation of the interaction between investigated molecules (a) and their most significant partners (b) visualized by STRING software. Black line type and saturation of the edges represent the confidence score of the functional association

Fig. 3

Unsupervised cluster analysis of the entire cohort and in P-LCNEC and EPNEC subgroups based on gene expression patterns. GEP, gastro-entero-pancreatic system; GU, genito-urinary tract; UK, unknown; P, pulmonary, L, large; S, small; M, male; F, female

Fig. 4

Gene expression patterns in neuroendocrine and non-neuroendocrine microdissected tumor cell populations of ten cases. NE, neuroendocrine; GU, genito-urinary tract; GI, gastrointestinal tract

Fig. 5

Overall survival curves in the entire cohort according to T stage and nodal status

Fig. 6

Overall survival curves in the entire cohort and in EPNEC, only, according to INSM1 expression (low and high as determined by median gene expression values, see text)