ABSTRACT
BACKGROUND/OBJECTIVES: Median survival of pancreatic ductal adenocarcinoma (PDAC) is around eight months and new prognostic tools are needed. Circular RNAs (circRNAs) have gained interest in different types of cancer. However, only a few studies have evaluated their potential in PDAC. We aimed to identify the most differentially expressed circRNAs in PDAC compared to controls and to explore their potential as prognostic markers. METHODS: Using frozen specimens with PDAC and controls, we performed RNA sequencing and identified 20,440 unique circRNAs. A custom code set of capture- and reporter probes for NanoString nCounter analysis was designed to target 152 circRNAs, based on abundancy, differential expression and a literature study. Expression of these 152 circRNAs was examined in 108 formalin-fixed and paraffin-embedded surgical PDAC specimens and controls. The spatial expression of one of the most promising candidates, ciRS-7 (hsa_circ_0001946), was evaluated by chromogenic in situ hybridization (CISH) using multi-punch tissue microarrays (TMAs) and digital imaging analysis. RESULTS: Based on circRNA expression profiles, we identified different PDAC subclusters. The 30 most differentially expressed circRNAs showed log2 fold changes from -3.43 to 0.94, where circNRIP1 (hsa_circ_0004771), circMBOAT2 (hsa_circ_0007334) and circRUNX1 (hsa_circ_0002360) held significant prognostic value in multivariate analysis. CiRS-7 was absent in PDAC cells but highly expressed in the tumor microenvironment. CONCLUSIONS: We identified several new circRNAs with biomarker potential in surgically treated PDAC, three of which showed an independent prognostic value. We also found that ciRS-7 is absent in cancer cells but abundant in tumor microenvironment and may hold potential as marker of activated stroma.
ABSTRACT
Pancreatic ductal adenocarcinoma (PDAC) is an aggressive type of cancer with an overall 5-year survival of around 10 %. New prognostic tools to stratify patients are needed. Our main aim was to evaluate the prognostic value of overall copy number variation (CNV) burden in surgically treated PDAC. DNA extracted from 108 surgical PDAC specimens was examined to collect data on the genome-wide DNA methylation status of >850,000 CpG sites in promoter, gene body, and enhancer regions (Illumina Infinium Methylation EPIC BeadChip Kit). CNV profiles were obtained and all PDACs were stratified into one of three groups: Low, moderate, or high overall CNV burden. Tumors histologically showing a dominant conventional and/or tubulopapillary pattern in 60 %-100 % and 0-59 % were categorized as Group A and Group B as per Kalimuthu. We also performed targeted next-generation sequencing (NGS) and immunohistochemistry. High overall CNV burden held independent negative prognostic value with poor survival (HR 4.01 (95%CI 1.96-8.19), p = 0.00014) and was more frequent in Group B (p = 0.0003). Most frequent chromosomal arm-level aberrations were gains of 8q (29 %) and 1q (19 %) and losses of 17p (55 %), 18q (43 %), 6q (37 %), 9p (36 %), 6p (26 %), 19p (26 %), and 8p (25 %). Most frequent mutations found were in KRAS (95 %), TP53 (62 %), CDKN2A (24 %), SMAD4 (23 %), ATM (9 %), ARID1A (7 %), RNF43 (7 %), GNAS (6 %), and KDM6A (6 %). Group A PDACs showed more frequently KRAS variants other than Gly12Val and Gly12Asp (p = 0.012). Our data indicate that overall CNV burden using genome-wide methylation profiling may be a useful prognostic tool in surgically treated PDAC. Importantly, our approach, using data from genome-wide methylation profiling for analysis of overall CNV burden, can be performed on formalin-fixed and paraffin embedded PDAC tissues. Future studies should examine the prognostic value of overall CNV burden in unresectable PDAC.
