Identification of functional pathways and molecular signatures in neuroendocrine neoplasms by multi-omics analysis

Abstract

Background: Neuroendocrine neoplasms (NENs) represent a heterogeneous class of rare tumors with increasing incidence. They are characterized by the ability to secrete peptide hormones and biogenic amines but other reliable biomarkers are lacking, making diagnosis and identification of the primary site very challenging. While in some NENs, such as the pancreatic ones, next generation sequencing technologies allowed the identification of new molecular hallmarks, our knowledge of the molecular profile of NENs from other anatomical sites is still poor.

Methods: Starting from the concept that NENs from different organs may be clinically and genetically correlated, we applied a multi-omics approach by combining multigene panel testing, CGH-array, transcriptome and miRNome profiling and computational analyses, with the aim to highlight common molecular and functional signatures of gastroenteropancreatic (GEP)-NENs and medullary thyroid carcinomas (MTCs) that could aid diagnosis, prognosis and therapy.

Results: By comparing genomic and transcriptional profiles, ATM-dependent signaling emerged among the most significant pathways at multiple levels, involving gene variations and miRNA-mediated regulation, thus representing a novel putative druggable pathway in these cancer types. Moreover, a set of circulating miRNAs was also selected as possible diagnostic/prognostic biomarkers useful for clinical management of NENs.

Conclusions: These findings depict a complex molecular and functional landscape of NENs, shedding light on novel therapeutic targets and disease biomarkers to be exploited.

Keywords: ATM signaling; Circulating biomarkers; Molecular signatures; Neuroendocrine neoplasms.

Fig. 1

H&E and neuroendocrine markers expression in NEN tumors categories. A NETG1 H&E (×20); B positive CgA expression in NETG1 (20×); C positive Syn expression in NETG1; D NETG2 H&E; E positive CgA expression in NETG2(×20); F positive Syn expression in NETG2 (×20); G NECG3 H&E (×20); H positive CgA expression in NECG3 (×20); I positive Syn expression in NECG3(×20); L H&E in MTC (×20); M positive CgA expression in MTC (×20); N positive Calcitonin expression in MTC. Scale bar 100 µm

Fig. 2

Mutational landscape of 44 NENs. A Pie chart showing proportions and genomic localizations of the identified variants after filtering and their B exonic function. C Oncoplot representation of the TOP 20 mutated genes in the analyzed samples. Only exonic and splicing mutations are represented. Each column represents individual patients with proper numeric code listed lower the graph and mutated genes are listed on the y-axis. The box colors indicate the type of mutation. The upper bar plots represent the total number of exonic/splicing variants identified for each sample within the represented gene set, while the right bar plots indicate the percentage of mutated samples for each gene. Colored bars in the bottom figure depict the pathological features of the patients; the upper is referred to cancer histotype: MTC, Gut, Pancreas and Others (lung-NEN and metastases), the middle indicates the WHO grade for GEP-NENs and the lower is referred to the presence of familiar syndrome. D Ingenuity Pathway Analysis (IPA) of the TOP 20 mutated genes. The red line indicates p-value threshold. E Boxplots showing Tumor Mutational Burden (TMB), as number of mutations per megabases, distribution comparing either GEP-NENs and MTCs (left panel) or NETs and NECs (right panel). T-test p-values are shown on the top

Fig. 3

Genomic rearrangements of NENs. A Circos plot showing aCGH-derived genomic rearrangements (middle ring) and fusion transcripts (inner ring) retrieved among the analyzed samples. External colored ring indicates chromosomes. Red bars represent deletions, while green ones show amplifications. The grid behind the bars shows the number of samples presenting rearrangements. The thin lines in the bars represent the presence of only a part of the aberration in some samples. The curved lines in the inner ring indicate the fusion genes in their exact location. Gene names are shown on the outer ring. B Heatmap representing aCGH-derived genomic rearrangements in the 30 samples considered for analysis. Each raw represents a chromosome while each column represents a patient whose numeric code is listed in the lower side. Shades of red and green indicate the rearrangement type as shown in the side legend. The bar with the different color blue shades represents tumor grades. C Boxplot representing gene counts between samples showing (green) or lacking (grey) amplification of chromosome 19. D Boxplot representing gene counts between samples having (red) or not (grey) deletion of chromosome 22

Fig. 4

miRNA profiling in NENs tissue and serum. A Venn Diagram comparing MTC and GEP-NEN expressed miRNAs. B Dot plot of the most significant IPA canonical pathways involving commonly expressed miRNAs. Dot color ranges from red to purple depending on the –log of the adjusted p-value. Dot sizes depend on the Gene Ratio as described in the figure. C GO-Plot showing some of the most important pathways involving the predicted target genes of miRNAs selected for validation. Different pathways are drawn with different colors. The pathway-involved target genes are reported in the outer ring. D Boxplots of the relative expression (2^-DDCT) of each serum validated miRNA in NEN (blue) samples versus healthy controls (red). Asterisks indicate statistical significance (p-value < 0.05). E Boxplots of the relative expression (2^-DDCT) of each serum validated miRNA in different samples groups mentioned in the figure. The asterisks indicate statistical significance (p-value < 0.05)

Fig. 5

Functional interaction networks. A Network of amplified genes found through multigene panel sequencing. B Functional network involving only mutated genes. C HIF1a signaling network involving miRNA targets. D Multi-level network considering both miRNA targets and gene variations. Circles are labeled with gene names, rectangles represent pathways, while diamonds the miRNAs. Different colors of the circles indicate if the genes are amplified (green), mutated (blue), targeted by miRNA (pink) or targeted and/or amplified (pink and green). Dashed lines indicate correlation between miRNAs and their targets, while full lines indicate gene connections to specific signalings