Whole exome and targeted sequencing were utilized to better characterize the genetic alterations in IPMNs. The most commonly mutated gene in IPMNs is KRAS with 50–80% of IPMNs harboring a KRAS mutation [6,7]. The most common location of the KRAS mutation is in codon 12, resulting in a G12D mutation. This KRAS mutation encodes a constitutively active GTP-binding protein that regulates various signaling cascades including cell growth and proliferation. Furthermore, KRAS mutations are found in >90% of PDACs and, hence, are thought to be essential and an early event for tumorigenesis [8]. Mutation in the GNAS gene is another common alteration in IPMNs and was identified in 40–70% of lesions [9]. GNAS encodes the stimulatory alpha subunit of the guanine nucleotide-binding protein which activates cyclic adenosine monophosphate (cAMP), leading to the activation of multiple effectors including protein kinase A and EPAC (Exchange Protein directly activated by cAMP). Although the relevance of these pathways in pancreatic cancer was not fully established, these pathways were implicated in other cancer types [10,11]. The GNAS mutation is predominantly located at codon 201, commonly leading to R201H or R201C alterations [9]. This is thought to be an activating mutation that results in the constitutive activation of GNAS. In contrast to KRAS, GNAS mutations are not commonly observed in non-IPMN PDACs but were found in 25–61% of IPMN- derived PDACs [7,12,13].

KRAS and/or GNAS mutations are found in >90% of IPMNs, both in those with advanced neoplasia (high-grade dysplasia or adenocarcinoma) and low-grade dysplasia [7]. The prevalence of their alteration suggests that they play an important role in IPMN development. These mutations were found to coexist, suggesting that they are not mutually exclusive. Interestingly, transgenic mouse models of IPMNs required synergistic mutation between GNAS and KRAS to generate cystic lesions that are histologically similar to human IPMNs [9,14]. Although there are no significant differences in KRAS and GNAS alterations in IPMNs with and without advanced neoplasia, a difference in mutation frequency were observed when analyzing the histological subtypes [7].

IPMNs can be divided into histological subtypes including gastric, intestinal, and pancreatobiliary. The oncocytic subtype is now considered a separate entity, i.e., “Intraductal Oncocytic Papillary Neoplasm”, by WHO’s classification of tumors, due to its different molecular and clinical features. These subtypes can give rise to different types of invasive carcinoma [15]. In general, intestinal IPMN gives rise to colloid carcinoma while tubular carcinoma can arise from gastric or pancreatobiliary subtype. Interestingly, there is a higher frequency of GNAS mutation in IPMN-associated colloid carcinoma and KRAS mutations in IPMN-associated tubular carcinoma [7]. This difference in mutational frequency may be clinically relevant, given that tubular adenocarcinoma is associated with worst clinical outcomes, and thus, it may offer prognostic value.

Inactivating mutation in the RNF43 gene is another common genetic alteration found in IPMNs, occurring in 10–75% of cystic lesions [7]. RNF43 encodes a transmembrane E3 ubiquitin ligase that negatively regulates the Wnt pathway. Thus, the loss of function of RNF43 confers Wnt pathway activation, which was shown to play a potential role in mediating PDAC progression [16,17]. Other loss of function mutations found in IPMNs include TP53, CDKN2A, and SMAD4. These genetic alterations occur at a lower rate in IPMNs with low-grade dysplasia compared to those with advanced neoplasia, suggesting that these genes may play a role in mediating progression to malignancy [7]. Furthermore, these genetic alterations are also commonly detected in invasive pancreatic cancer.

More recently, Noe et al. performed whole exome sequencing (WES) of IPMNs and MCNs and their associated invasive carcinomas to better understand the genetic alteration driving tumorigenesis in these cystic precursors [18]. WES analysis revealed the high prevalence of previously identified drivers of pancreatic tumorigenesis including KRAS, GNAS, RNF43, CDKN2A, TP53, and SMAD4. In addition, the use of WES revealed novel mutations in IPMN tumorigenesis. Somatic mutations in the ATM gene were found in 17% of lesions. ATM encodes a serine/threonine kinase that is involved in DNA double-strand break repair as well as other cellular processes including metabolism and chromatin remodeling. Moreover, individuals with germline pathogenic ATM variants appear to have an increased lifetime risk of pancreatic cancer [19,20]. In addition, loss of ATM expression in a mouse model of PDAC resulted in a greater number of proliferative precursor lesions [21]. Somatic mutation in GLI3 gene was also observed in 8% of samples. GLI3 gene encodes a transcription factor that is a member of the Hedgehog signaling pathway and is involved in normal cellular processes including tissue and immune cell development. Furthermore, studies implicated GLI3 in mediating pro-tumorigenesis pathways in various cancers, including pancreatic cancer [22,23,24].