CDKN2A homozygous deletions and TSC2 somatic mutations in metastatic pancreatic neuroendocrine tumors

Abstract

Despite improvements in the molecular profiling of pancreatic neuroendocrine tumors (PanNETs), predicting their clinical behavior and response to specific therapies remains challenging. We sought to elucidate the molecular basis underlying the broad phenotypic variations in these neoplasms through a genetic characterization of primary and metastatic PanNETs. Our findings revealed an enrichment of CDKN2A homozygous deletions and TSC2 somatic mutations in metastatic PanNETs when compared to non-metastatic lesions. Tumor evolution analysis further revealed the acquisition of such genetic alterations as late events in the progression of these neoplasms, conferring poor survival outcomes to the affected patients. Biallelic loss of DNA damage repair genes, ATRX and/or DAXX, was associated with a high fraction of the genome altered in PanNETs, with pathogenic alterations affecting those genes also being associated with a homologous recombination deficiency signature. These findings highlight molecular mechanisms driving PanNET progression and underscore the need for further molecular characterization and tumor evolution studies to evaluate targeted therapies for such a challenging disease.

Fig. 1. Genomic features of primary neoplams and metastases from pancreatic neuroendocrine tumors.

Comparisons between primary neoplams and metastases from pancreatic neuroendocrine tumors (PanNETs) for a, recurrent somatic alterations, b, mutual exclusivity analysis between DAXX and ATRX, and c, copy number gains and losses. Comparisons between liver metastases from PanNETs stratified according to the number of distant metastases for d, recurrent somatic alterations. Alteration types and number of metastases are color-coded according to the legend. Statistical significance was evaluated using Fisher’s exact test. CN copy number, PanNETs pancreatic neuroendocrine tumors.

Fig. 2. Genomic features of primary neoplams and metastases from pancreatic neuroendocrine tumors with low and high fraction of the genome altered.

Comparisons between primary neoplasms and metastases from pancreatic neuroendocrine tumors (PanNETs) with low and high fraction of the genome altered for a and b, recurrent somatic alterations, and c, tumor mutational burden. Comparisons between PanNETs harboring single-hit alterations affecting DAXX and ATRX, and double genetic alterations affecting the same genes for d, fraction of the genome altered. Alteration types are color-coded according to the legend. Statistical significance was evaluated in (a) and (b) using Fisher’s exact test, and in (c) and (d) using the Mann-Whitney U test. FGA, fraction of the genome altered; PanNETs, pancreatic neuroendocrine tumors.

Fig. 3. Survival analysis of metastatic pancreatic neuroendocrine tumors harboring genetic alterations affecting ATRX, DAXX, MEN1, TSC2, CDKN2A and KRAS.

Kaplan-Meier overall survival curves for 175 pancreatic neuroendocrine tumors stratified by altered gene status, namely a, ATRX, b, DAXX, c, MEN1, d, TSC2, e, CDKN2A (homozygous deletions), f, KRAS (G12D/V/R). Statistical significance was evaluated in (a), (b), (c), (d), (e) and (f) using the log-rank test. PanNETs, pancreatic neuroendocrine tumors.

Fig. 4. Clonal composition and phylogenetic analysis of primary pancreatic neuroendocrine tumors and matched metastases.

Clonal composition and phylogenetic analysis of three cases (a–c). Heatmaps depicting the frequency of genetic alterations and cancer cell fractions of each somatic mutation in the primary pancreatic neuroendocrine tumors (PanNETs) and matched metastasis of a given case are shown ((a) and (b), top left; (c), top). The alteration types and cancer cell fraction are color-coded according to the legend. Copy number plots depicting segmented Log2 ratios (y-axis) according to genomic position (x-axis) of primary PanNETs and matched metastases are depicted ((a) and (b), top right; (c), bottom). Phylogenetic trees of primary PanNETs and matched metastases are shown ((a) and (b), bottom left; (c), middle). Trunk and branches are colored according to clusters, and the number of somatic mutations that result in the divergence of a clone/subclone from its ancestor is shown. Pathogenic mutations that define a given clone are depicted. Phylogenetic trees based on copy number alterations are shown ((a) and (b), bottom right; (c), bottom left). The numbers alongside the branches represent the number of copy number alterations. Gains and losses are shown in parentheses. Alkyl alkylating agent, dROS damage by reactive oxygen species, M metastasis, MSI microsatellite instability, NA not applicable, P primary, UV ultra-violet light, T trunk.

Fig. 5. Genomic features of primary pancreatic neuroendocrine tumors with and without homologous recombination deficiency features.

Comparisons between primary pancreatic neuroendocrine tumors stratified according to the presence/absence of homologous recombination deficiency features assessed by mutational signatures (a) for b, recurrent somatic mutations and c, copy number alterations (gains, losses, and amplifications). Alteration types are color-coded according to the legend. Statistical significance was evaluated using Fisher’s exact test. ALT, alternative lengthening of telemores; BER base-exchange repair, CN copy number, HRD homologous recombination deficiency, MSI microsatellite instability, PanNETs pancreatic neuroendocrine tumors.

Fig. 6. Signaling pathways most affected by genetic alterations present in pancreatic neuroendocrine tumors.

Frequency of activating (red) or loss-of-function (blue) somatic genetic alterations affecting genes in pancreatic neuroendocrine tumors (PanNETs) and present in the canonical a PI3K/AKT/mTOR, and b p53/DNA damage repair/cell cycle signaling pathways. The frequency of primary PanNETs with no evidence of metastases, primary tumors with evidence of metastases, and metastatic samples is depicted under the gene name. Statistical significance was evaluated using using Mann-Whitney U test.