Well-differentiated G1 and G2 pancreatic neuroendocrine tumors: a meta-analysis of published expanded DNA sequencing data

Abstract

Introduction: Well-differentiated pancreatic neuroendocrine tumors (PNETs) can be non-functional or functional, e.g. insulinoma and glucagonoma. The majority of PNETs are sporadic, but PNETs also occur in hereditary syndromes, primarily multiple endocrine neoplasia type 1 (MEN1). The Knudson hypothesis stated a second, somatic hit in MEN1 as the cause of PNETs of MEN1 syndrome. In the recent years, reports on genetic somatic events in both sporadic and hereditary PNETs have emerged, providing a basis for a more detailed molecular understanding of the pathophysiology. In this systematic review and meta-analysis, we made a collation and statistical analysis of aggregated frequent genetic alterations and potential driver events in human grade G1/G2 PNETs.

Methods: A systematic search was performed in concordance with the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) reporting guidelines of 2020. A search in Pubmed for published studies using whole exome, whole genome, or targeted gene panel (+400 genes) sequencing of human G1/G2 PNETs was conducted at the 25^th of September 2023. Fourteen datasets from published studies were included with data on 221 patients and 225 G1/G2 PNETs, which were divided into sporadic tumors, and hereditary tumors with pre-disposing germline variants, and tumors with unknown germline status. Further, non-functioning and functioning PNETs were distinguished into two groups for pathway evaluation. The collated genetical analyses were conducted using the 'maftools' R-package.

Results: Sporadic PNETs accounted 72.0% (162/225), hereditary PNETs 13.3% (30/225), unknown germline status 14.7% (33/225). The most frequently altered gene was MEN1, with somatic variants and copy number variations in overall 42% (95/225); hereditary PNETs (germline variations in MEN1, VHL, CHEK2, BRCA2, PTEN, CDKN1B, and/or MUTYH) 57% (16/30); sporadic PNETs 36% (58/162); unknown germline status 64% (21/33). The MEN1 point mutations/indels were distributed throughout MEN1. Overall, DAXX (16%, 37/225) and ATRX-variants (12%, 27/225) were also abundant with missense mutations clustered in mutational hotspots associated with histone binding, and translocase activity, respectively. DAXX mutations occurred more frequently in PNETs with MEN1 mutations, p<0.05. While functioning PNETs shared few variated genes, non-functioning PNETs had more recurrent variations in genes associated with the Phosphoinositide 3-kinase, Wnt, NOTCH, and Receptor Tyrosine Kinase-Ras signaling onco-pathways.

Discussion: The somatic genetic alterations in G1/G2 PNETs are diverse, but with distinct differences between sporadic vs. hereditary, and functional vs. non-functional PNETs. Increased understanding of the genetic alterations may lead to identification of more drivers and driver hotspots in the tumorigenesis in well-differentiated PNETs, potentially giving a basis for the identification of new drug targets. (Funded by Novo Nordisk Foundation, grant number NNF19OC0057915).

Keywords: Knudson’s two-hit hypothesis; MEN1; genetics; germline; meta-analysis; pancreatic neuroendocrine tumors; somatic; systematic review.

Figure 1

PRISMA diagram of the literature search. A total of 14/300 datasets were included. PNET, Pancreatic neuroendocrine tumor; G, histological grade. .

Figure 2

Genes frequently altered in pancreatic neuroendocrine tumors (PNETs). Oncoplot of genes altered in more than 2% of all samples are presented in the datasets (n=225). The sum of mutations annotated in each PNET is shown in the upper panel. Each colored bar indicates a somatic variant colored based on the mutation type and copy number variation as depicted. Colored filled boxes represents mutations, partly filled boxes represents copy number variations. The tumor types of each of the PNETs are shown in the lower bar. Panel on the right sums up the number of mutations (upper) and copy number variations (CNVs) (lower) identified in each specific gene. Further, the percentage of PNETs with mutations (- CNV) and mutations plus CNVs (+CNV) are presented at the right. TMB, tumor mutation burden; cnLOH, copy-neutral loss of heterozygosity; CNV, copy number variations.

Figure 3

MEN1 mutational distribution in MEN1 patients and sporadic PNETs, where MEN1 is altered in 30% on mutational level. (Upper panel) MEN1 mutations found in germline from 11 (73%) of MEN1 patients. One MEN1 patient harbored germline copy number variations in MEN1, and one MEN1 patients had a splice site variant (not shown). Lastly, one PNET had an unspecified MEN1-germline variant. (Lower panel) somatic MEN1 mutations from two patient with MEN1 (p.Leu175Pro and p.G469Afs*35) and 65 sporadic PNETs. Nine somatic variants were further identified in the MEN1 splice sites (not shown). Twelve (80%) MEN1 patients had somatic copy number variations in the MEN1 locus. Transcript: NM_130799 (isoform 2), protein identifier: NP_570711, menin length: 610 amino acids.

Figure 4

DAXX and ATRX somatic alterations in pancreatic neuroendocrine tumors (PNETs). Somatic variations uncovered in the frequently altered genes (A) DAXX (15%) and (B) ATRX (8.5%). Each protein is highlighted in gray, using the transcripts NM_001141970 and NM_000489, respectively. Domains are represented as colored boxes. The pins correspond to single somatic mutations identified in PNETs of the cohort and the color of the pin indicates the mutation type, and the height depicts the number of the variant type in the locus. Annotated domains are adapted from Wang et al. (40). Domains in DAXX: SIM, Sumo-interaction motif; DHB, DAXX helical bundle; HBD, histone binding domain; A, Acidic segment rich in Glu/Asp residues; SPE, segment rich in Ser/Pro/Glu residues; SPT, segment rich in Ser/Pro/Thr residues. Domains in ATRX: ADD, ATRX-DNMT1-DNMT1L domain; HP1, HP1-binding motif; DBM, DAXX binding motif; ATPase, ATPase domain.

Figure 5

Co-occurrence plot showing mutually exclusive (brown) or co-occurring (cyan) set of altered genes (mutations) identified in PNETs. The plot shows the top of mutated genes (altered in more than 2% of PNETs) and p-values are indicated as asterics (p-value < 0.05) or dots (p-value < 0.1) determined by pair-wise Fisher’s exact test. * P-value < 0.05 and the dot represents p-value < 0.1

Figure 6

Frequently altered genes in hereditary and sporadic pancreatic neuroendocrine tumors (PNETs). The gene list is ordered after the frequency of somatic mutations in specific genes. (A) Data from PNETs from patients with germline mutation in MEN1, MUTYH, CHEK2, BRCA, CDKN1B, CHEK2/MUTYH, MEN1/MUTYH or VHL (n=30) and (B) PNETs from patients with sporadic PNETs (n=162).The tumor syndrome of each of the PNETs is shown in the lower bar. Each colored bar in the oncoplot indicates a somatic variant colored based on the mutation type or copy number variation as depicted. Filled colored boxes represent gene-specific somatic mutations, and partly colored boxes represent copy number variations in the specific gene region or chromosome. Panel on the right sums up the number of PNETs with mutations (upper), and mutations plus copy number variations (CNVs) (lower) identified in each specific gene. Likewise, the percentage is represented for PNETs with mutations (-CNV) and for PNETs with mutations and CNVs (+CNV). PNETs with more than one mutation in a gene were represented as a multi-hit (black), and genes with mutation and copy number variation were depicted as a complex event (grey). The sum of variations annotated in each PNET is shown in the upper panel. CNV, copy number variation; cnLOH, copy-neutral loss of heterozygosity; TMB, tumor mutation burden.

Figure 7

Oncoplots of somatic mutations and copy number variations identified in pancreatic neuroendocrine tumors (PNETs) divided into tumor types. Frequently altered genes identified in (A) non-functioning PNETs and (B) functioning PNETs are presented (found in more than 3%). Variant type is depicted in the barplot and the sum of mutations (upper) and the sum of mutations plus copy number variations (lower) in each gene is represented in the right bar. The percentage represents the percentage of PNETs with mutations (-CNV) and mutations plus copy number variations (+CNV) in each gene. The upper bar shows the total number of mutations in each specific PNET sample. (C) Groupwise comparison of gene variants based on tumor type. Copy number variations are not included in the analysis. The corresponding bars are colored by tumor type and indicates genes significantly altered (p-value < 0.05) between tumor types. The y-axis represents the odds ratio. TMB, Tumor mutation burden; CNV, copy number variation; cnLOH, copy-neutral loss of heterozygosity.

Figure 8

Top oncogenic pathway enriched in (A) non-functioning (n=30) and (B) functioning (n=5) pancreatic neuroendocrine tumors (PNETs). The oncogenic pathways uncovered in PNETs are based on somatic mutations. The oncogenic signaling pathways are based on pathways from The Cancer Genome Atlas (TCGA) cohorts. Genes highlighted in red are tumor suppressor genes, and genes highlighted in blue are proto-oncogenes. Each column indicates a PNET and a somatic mutation is represented by a red box in the row of the altered gene. PI3K, Phosphoinositide 3-kinase; RTK-RAS, Receptor Tyrosine Kinase-Ras.