Genomic characterization of malignant progression in neoplastic pancreatic cysts

Abstract

Intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs) are non-invasive neoplasms that are often observed in association with invasive pancreatic cancers, but their origins and evolutionary relationships are poorly understood. In this study, we analyze 148 samples from IPMNs, MCNs, and small associated invasive carcinomas from 18 patients using whole exome or targeted sequencing. Using evolutionary analyses, we establish that both IPMNs and MCNs are direct precursors to pancreatic cancer. Mutations in SMAD4 and TGFBR2 are frequently restricted to invasive carcinoma, while RNF43 alterations are largely in non-invasive lesions. Genomic analyses suggest an average window of over three years between the development of high-grade dysplasia and pancreatic cancer. Taken together, these data establish non-invasive IPMNs and MCNs as origins of invasive pancreatic cancer, identifying potential drivers of invasion, highlighting the complex clonal dynamics prior to malignant transformation, and providing opportunities for early detection and intervention.

Intraductal Pancreatic Mucinous Neoplasms: A Tumor-Biology Based Approach for Risk Stratification

“…SMAD4 mutations are rarely observed in IPMNs while enriched in invasive carcinomas, thereby suggesting that they are acquired during late stages of tumorigenesis; the hypothesis that alterations involving SMAD4 may drive malignant transformation and invasion in a subset of lesions is supported by a recent work by Noë et al [ 66 , 67 , 72 , 88 ]. In order to be translocated into the nucleus, Smad4 has to be bound to phosphorylated Smad3, which can be phosphorylated as follows: COOH-terminally phosphorylated Smad3 (pSmad3C), frequently detected in normal pancreatic tissues, linker-phosphorylated Smad3 (pSmad3L), rarely found in normal tissues, and Smad3 phosphorylated at both sites (pSmad3L/C).…”

“…The role of KRAS in PDAC and other human cancers is well established [ 69 ]; mutations in this gene are an early event during IPMN development and are present in up to 80% of the cases [ 5 , 70 , 71 , 72 ]. A recent study confirmed the synergistic action of KRAS and tumor suppressor gene mutations for development of IPMN in animal model and highlighted the role of Wnt/β-catenin pathway in KRAS -associated lesions [ 73 ].…”

“…Of interest, in vivo studies highlighted an increased sensitivity of PDAC harboring RNF43 mutations to Wnt-pathway pharmacological inhibitors, which may represent a new option for Wnt-addicted subset of PDAC [ 86 , 87 ]. Noë et al recently showed that distinct RFN43 alterations are present in multiple precancerous clones, but are lacking in invasive cancer, suggesting that some mutations characterizing early noninvasive lesions are not selected for invasive carcinoma [ 72 ].…”

“…Other recently discovered novel driver genes may contribute to IPMN tumorigenesis. Both germline and somatic ATM mutations have been found in up to 17% cases; other newly proposed driver genes are GLI3 and SF3B1 [ 72 ].…”

“…In summary, further studies could provide insights into transition of IPMN to invasive carcinoma [ 72 , 81 , 82 ], with mutations in CDKN2A , SMAD4 , and TP53 genes probably not fundamental in the early evolution of IPMNs. Recently, Bayesian hierarchical models to study the evolutionary timeline of high-grade IPMN to PDAC estimated that at least three years are needed for acquisition of invasive properties: this significant time window provides us with an opportunity for surveillance and early detection of invasive cancer; it will therefore be of uttermost importance to find the correct strategy to pursue this goal [ 72 ].…”