Background: Pancreatic cancer is one of the most devastating diseases without early detection, effective screening biomarkers and therapeutic treatments. In the past decades, genetic studies have indicated various genes related to this malignancy. Summary: Genetic alterations have been involved in the initiation, progression and invasion of pancreatic cancer, which might indicate promising targets for early screening, diagnosis and future intervention. Here we will review genetic changes in pancreatic cancer and analyze their correlations with several common precursors and familial syndromes. Key Message: Genetic analysis for pancreatic cancer or its precursors might help us to characterize patients into subtype individuals in the future and have significant implications for individualized treatments. Practical Implications: At present, pancreatic cancer is regarded as a disease with a wide range of genetic alterations, including germline and somatic mutations. Some genetic alterations such as KRAS, p16CDKN2A, TP53 and SMAD4 were specifically correlated with different types of histological precursors of pancreatic cancer and some familial syndromes highly related to pancreatic cancer. Moreover, genetic changes also predict drug sensitivity and implicate novel therapeutic targets.
Pancreatic cancer is the fourth leading cause of cancer mortality among males and females in the United States and seventh in China [1,2]. Despite enormous breakthroughs in cancer therapy in the past decades, the mortality rate for pancreatic cancer is still increasing . Surgery is believed to be the only curative tool for pancreatic cancer patients . However, the 5-year survival rate for patients with localized disease after surgical resection is 24% and only 2% for those with metastatic disease . Moreover, due to lack of specific symptoms at an early stage and the absence of effective diagnostic biomarkers, up to 80% of patients are diagnosed at an advanced stage and not eligible for surgery . For most patients, adjuvant and neoadjuvant chemotherapy based on gemcitabine might be the most effective treatment . However, the median survival for locally advanced or metastatic pancreatic cancer patients treated with gemcitabine is only 7.9 months .
Although the overall survival rate for pancreatic cancer patients has remained dismal in the past years, various approaches have been investigated to improve this result. Prevention and early detection appear to be the first effective ways. Prevention includes controlling risk factors such as cigarette smoking, blood glucose level, obesity, etc. [7,8]. Recent surveillance methods include the detection of serum CEA and CA19-9 level . Imaging has been greatly improved with the development of computed tomography scans, magnetic resonance imaging and endoscopic ultrasonography (EUS) [10,11]. However, none of these methods can be used alone for diagnosis and the optimal strategy has not yet been determined. Although EUS is effective in detecting small lesions, abnormalities on EUS are not specific for early detection of pancreatic cancer because of lacking specific biomarkers .
In the past years, researchers have studied the molecular basis of cancer initiation, progression and metastasis in various tumor entities. Substantial progress has been made in understanding the genomic initiation of pancreatic cancer. Moreover, it has been recognized that pancreatic cancer has close associations with three types of histological precursors, namely pancreatic intraepithelial neoplasia (PanIN), intraductal papillary mucinous neoplasm (IPMN) and mucinous cystic neoplasm (MCN). Interestingly, all of these precursor lesions are related to specific genetic alterations . In particular, several key gene mutations were found in a majority of familial pancreatic cancer (FPC) cases. Additionally, novel genetic studies indicated that mutated KRAS and p53 are involved in inherited or acquired drug resistance, including gemcitabine treatment [14,15,16,17].
In this review, we will summarize the related genetic changes in pancreatic cancer and their correlations with several important precursors of pancreatic cancer and specific familial syndromes, which might provide help for early screening, pathological classification, diagnostic, predictive and prognostic implications.
Genetics of Pancreatic Cancer
Pancreatic ductal adenocarcinoma (PDAC) accounts for >90% of all pancreatic neoplasms . These tumors are derived from pancreatic ductal cells. Until now, PDAC has been regarded as a disease with a wide range of genetic alterations, including germline and somatic mutations. In the majority of invasive PDAC, the genetic abnormalities with high frequency are mutational activation of the KRAS oncogene, inactivation of tumor suppressor genes including p16CDKN2A, TP53, SMAD4/DPC4 and BRCA2, widespread chromosomal losses, gene amplifications and telomere shortening . In a few (<20%) pancreatic cancers, genes mutated with low frequency include oncogenes such as BRAF, MYB, AKT2 and EGFR, and tumor suppressor genes such as MKK4, MAP2K4, STK11, TGFBR1, TGFBR2, ACVR1B, ACVR2A, FBXW7 and EP300 [19,20].
KRAS. The KRAS gene is located on chromosome arm 12p and encodes a membrane-bound guanosine triphosphate-binding protein. It mediates various cellular functions, including proliferation, cellular survival, motility and cytoskeletal remodeling . Activation of KRAS mutations appears to be the initial genetic change that is detected in the progression series, occurring occasionally in histologically normal pancreas and in about 30% of lesions that show the earliest stages of histological disturbance. This has been found in >95% of pancreatic cancers and the gene is mainly activated by point mutations, a single amino acid substitution from G to D at codon 12 [19,22]. In pancreatic cancer, KRAS has been involved in three major pathways, including RAF/MEK/ERK, PI3K/Pdk1/AKT and the Ral guanine nucleotide exchange factor pathway [23,24,25]. Loss of KRAS expression results in massive cell death and arrested proliferation, leading to rapid tumor regression . Recently, tumors with KRAS mutations have been characterized by constitutively high levels of autophagy . KRAS participated in this process by altering the expression of enzymes involved in glucose utilization .
GNAS. GNAS mutations are reported as a frequently observed early genetic aberration in IPMNs. Studies found that nearly 50% of IPMNs harbored GNAS mutations, and 51% harbored both GNAS and KRAS mutations . GNAS is located on chromosome 20q13.32 and encoded G-protein alpha subunit . GNAS mutations were observed both in low- and high-grade tumors as well as in invasive tumors . They might play an important role in the initiation of IPMN.
Tumor Suppressor Genes
p16CDKN2A. The p16CDKN2A protein is a cyclin-dependent kinase inhibitor. It is related to several different mechanisms such as homozygous deletion or promoter hypermethylation . The p16CDKN2A gene has been found to be inactivated in 95% of PDACs. It locates at 9q21 and encodes two tumor suppressors, INK4A and ARF . In most pancreatic cancers, homozygous deletion of 9p21 causes the loss of both transcripts, thereby disrupting the related RB and p53 tumor suppression pathways. INK4A inhibits CDK4/CDK6-mediated phosphorylation of RB, thereby blocking entry into the S phase of the cell cycle. ARF stabilizes p53 by inhibiting its MDM2-dependent proteolysis. Compared to ARF, INK4A seems to be a dominant suppressor gene at this locus, as it is more frequent in germline and sporadic mutations .
TP53. The mutation of tumor protein 53 (TP53) is observed in up to 75% of pancreatic cancers . It is regarded as a late event during PanIN progressing to PDAC and seems to be more frequent in IPMNs . TP53 encodes the p53 protein. p53 plays an important role in cellular stress responses, particularly by activating DNA repair, inducing growth arrest and triggering apoptosis. It mediates G1 block  and has a close relationship with G2/M block  by upregulating p53 up-regulated modulator of apoptosis (PUMA) and binding to BCL2 . Loss of p53 function through mutation of the TP53 gene therefore promotes pancreatic neoplasm through the loss of critical cell functions . In mice containing oncogenic KRAS and lacking p53, loss of autophagy no longer blocks tumor progression, but actually accelerates carcinogenesis . Unlike most tumor suppressor genes, inactivation of the TP53 gene typically occurs through missense mutations of one allele, accompanied by loss of the other allele. The majority of these missense mutations are clustered in hot spot residues, mainly within the DNA binding domains .
SMAD4/DPC4. Another frequent genetic alteration in PDAC is the loss of SMAD4/DPC4. This gene maps to chromosome 18q21, a region that sustains deletion in approximately 55% of pancreatic cancers, mainly by homozygous deletion . SMAD4 encodes a transcriptional regulator that is a keystone component in the transforming growth factor-β (TGF-β) family signaling cascade. Loss of SMAD4 function abolishes the SMAD4-dependent TGF-β pathway and gives rise to unregulated cellular proliferation . Patients with tumors that harbor a wild-type SMAD4 gene have a lower propensity for widespread metastasis than those with loss of SMAD4, and loss of DPC4 expression was closely related to a low survival rate . Loss of nuclear SMAD4 frequently occurs in pancreatic adenocarcinomas, but not in extrapancreatic malignancies .
BRCA2. Inherited BRCA2 mutations are typically associated with hereditary breast-ovarian cancer (HBOC) syndrome, but also carry a significant risk for the development of pancreatic cancer. BRCA2, located on 13q12.3, approximately accounts for 24% of sporadic pancreatic cancers . In normal cells, BRCA2 maintains genomic stability through regulating the homologous recombination-based DNA repair process. Tumor cells with defective BRCA2 lack the ability to repair DNA, rendering the cells more susceptible to cell death . Thus, BRCA2 deficiency results in the accumulation of lethal chromosomal aberrations . Loss of wild-type BRCA2 seems to be a late event in those individuals who inherit germline heterozygous mutations of BRCA2, which predisposes to severely dysplastic PanINs and adenocarcinomas .
Other Less Frequent Mutations
There are many less frequently mutated genes in pancreatic cancer. AKT2 gene mutation was found in 20% of pancreatic cancers . Amplifications of some oncogenes such as C-MYC, KRAS and GATA6 were less frequent [19,49]. Homozygous deletions of the suppressor gene MKK4 were found in about 2% of pancreatic cancer cases. Interestingly, MKK4 functions as a downstream effector of DPC4, p16CDKN2A, TP53 and BRCA2 . Recent reports indicated that genes mutated in a few (<20%) pancreatic cancers, including oncogenes such as BRAF, MYB, AKT2 and EGFR, and tumor suppressor genes such as MAP2K4, STK11, TGFBR1, TGFBR2, ACVR1B, ACVR2A and FBXW7 . Moreover, structural analysis of mutated genes indicated PIK3CG, DGKA, STK33, TTK and PRKCG as low-frequency driver mutations .
Clinical Implications of Genetic Alterations in Pancreatic Cancer
Screening of Hereditary and FPC Syndromes
Approximately 10% of pancreatic cancers have a familial basis [52,53]. Familial aggregations are related to several syndromes, such as Peutz-Jeghers syndrome (PJS), familial atypical multiple mole melanoma (FAMMM) syndrome, HBOC syndrome, hereditary nonpolyposis colorectal carcinoma (HNPCC) syndrome, familial adenomatous polyposis (FAP) syndrome, Li-Fraumeni syndrome, ataxia telangiectasia, hereditary pancreatitis, cystic fibrosis and FPC syndrome [54,55]. Of note, various germline gene alterations have been found to be involved in these syndromes.
PJS is an autosomal dominant hamartomatous polyposis syndrome. Patients with PJS have an increased incidence of pancreatic cancer . Interestingly, >80% of PJS cases are caused by inherited mutations in the tumor suppressor gene STK11/LKB1 (chromosome 19p13) . PJS patients have a high predisposition to other gastrointestinal cancers, including colorectal cancer (39%), gastric cancer (29%) and small bowel adenocarcinomas (13%) [58,59]. Females with PJS are even at high risk for breast cancer (54%) and gynecologic malignancies, including ovarian tumors (21%), endometrial cancer (9%) and adenoma malignum of the cervix (10%) . Individuals with PJS harbor a 132-fold increased risk of pancreatic cancer (table 1). Moreover, patients with PJS are at increased risk for IPMNs and their precursors . All this is a profound implication for the screening and management of pancreatic cysts in patients with PJS .
FAMMM syndrome is an autosomal dominantly inherited syndrome with incomplete penetrance. The germline mutation p16CDKN2A gene is strongly associated with FAMMM syndrome, although the prevalence was inconsistent [62,63,64]. Except for pancreatic cancer, individuals with FAMMM syndrome are also at increased risk of developing sarcomas, lung and breast cancers [65,66]. The risk of pancreatic cancer in kindreds with FAMMM syndrome is 13- to 22-fold higher than in the average population [65,67], and individuals with a germline p16CDKN2A mutation have a 38-fold higher risk of pancreatic cancer than the general population (table 1) .
HBOC syndrome is an autosomal, dominantly inherited syndrome associated with germline mutations in BRCA1, BRCA2 and other genes in the Fanconi anemia pathway (FANCN/PALB2) [11,69]. Among them, germline mutations of BRCA2 have the highest prevalence (5-17%) in inherited PDAC [11,70,71]. In patients with pancreatic cancer and two or more first-degree relatives with pancreatic cancer, the prevalence of BRCA2 mutations has been estimated to range from 17 to 19% [69,71]. In a study of 173 families with germline BRCA2 mutations and breast/ovarian cancers, representing a cohort of 3,728 individuals, the relative risk of pancreatic cancer was 3.51 (table 1) . The incidence of pancreatic cancer is especially high in families of Ashkenazi Jewish descent, where a germline BRCA26174delT mutation has been reported .
HNPCC syndrome, also named Lynch syndrome, is a dominant autosomal disease . It is associated with germline mutations of mismatch repair genes, including MLH1, MSH2, PMS1, PMS2 and MSH6/GTBP . Except for pancreatic cancer, patients with HNPCC syndrome also have an increased risk of other cancers such as endometrial, gastric, small intestinal and ureteral cancers . A recent study showed that 147 families containing mismatch repair gene mutation in a mismatch gene had an 8.6-fold (95% CI 4.7-15.7) increased risk of pancreatic cancer as compared to the general population (table 1) . This corresponds to a 3.68% (95% CI 1.45-5.88%) lifetime risk of pancreatic cancer (by age 70) . When patients with HNPCC syndrome develop pancreatic cancer, they usually have a characteristic medullary phenotype. Some often harbor microsatellite instability and a distinctly poor differentiated medullary histopathology .
FAP syndrome is an autosomal, dominantly inherited disease [78,79]. Germline mutation of APC is linked with FAP syndrome. It is estimated that this alteration increases 4-fold the risk for pancreatic cancer, while it also has an elevated incidence of early colonic polyps and adenocarcinoma (table 1) . APC mutation causes derangement of Wnt signaling, which leads to pancreatic ampullary cancer or duodenal cancer.
Hereditary pancreatitis is a rare inherited form of pancreatitis . Mutations in the cationic trypsinogen gene (PRSS1) on chromosome 7q35 cause an autosomal dominant form of hereditary pancreatitis, whereas mutations in the serine protease inhibitor gene (SPINK1) on chromosome 5q32 cause an autosomal recessive form of hereditary pancreatitis [82,83,84]. Patients with hereditary pancreatitis have a remarkable 58-fold (95% CI 23-105) increased risk of pancreatic cancer and a lifetime risk of pancreatic cancer of 30-40% (by age 70) (table 1) [85,86].
Pancreatic Cancer Families Screening
Individuals with at least a 5- to 10-fold increased risk of pancreatic cancer, such as members of FPC families with at least two affected first-degree relatives, might be candidates for long-term surveillance and screening [11,87]. Genetic alterations might be effective for screening . Pancreatic surgery can be performed when a suspicious lesion is identified.
With the increasing sensitivity of novel imaging modalities, occasional pancreatic cysts become more discernible. Most of these do not require surgical intervention because of the low risk of malignancy. However, some might be high-risk lesions, such as IPMN and MCN. Studies on IPMN indicate that approximately 70% of main pancreatic duct IPMNs may progress into invasive cancer . Thus, it is important to identify some biomarkers to differentiate low- from high-risk lesions, consequently appropriately stratifying patients for either surveillance or surgical intervention.
KRAS might be an essential biomarker (table 2). During the progression from PanIN-1 to PanIN-2 and PanIN-3 lesions, KRAS mutation is an important early event and the rate in PanIN patients becomes higher [22,89]. It is reported that approximately 45% of PanIN-1 lesions carried a KRAS mutation . Moreover, through loss of heterozygosity analyses, p16CDKN2A, TP53 and SMAD4 have also been related to the progression of PanIN . Inactivating mutations of p16CDKN2A through promoter hypermethylation are found in PanIN-2 lesions . It has been demonstrated that functional abrogation of p16CDKN2A is one of the early events in pancreatic ductal carcinogenesis . Inactivation of TP53 and SMAD4 are generally associated with PanIN-3 . These findings might provide a genetic basis for the progression of PanINs to PDAC and the staging of PanINs.
The analysis of gene mutation might also shed light on the differential diagnosis for pancreatic premalignant mucinous lesions. Activating point mutations of KRAS occur in approximately 50% of IPMNs with low-grade dysplasia, and the prevalence of KRAS mutations increases the grading of dysplasia . It has been reported that EUS fine needle aspiration plus KRAS mutation analysis might significantly improve accuracy in the diagnosis of pancreatic cancer .
Moreover, inactivated mutations of p16CDKN2A and TP53 are found in IPMNs with high-grade dysplasia  (table 2). GNAS mutations are reported as a frequently observed early genetic aberration in IPMNs. Studies of MCNs have indicated the prevalence of KRAS mutations and aberrant nuclear p53 accumulation with increasing grades of dysplasia . Although SMAD4 mutation and loss of nuclear expression are not observed in most noninvasive MCNs, its loss of expression was often found in the progression from MCNs to infiltrating cancers. Inactivation of SMAD4 is regarded as the late event of neoplastic progression from MCNs . Furthermore, tumor microsatellite instability testing in patients with a family history appears to be an effective way for the surveillance of Lynch syndrome or the presence of young-onset medullary cancer of the pancreas.
During the past decade, numerous chemotherapeutic and molecular targeted agents have been evaluated alone or in combination in various clinical trials. Chemotherapeutic agents including DNA-damaging agents, such as fluorouracil, oxaliplatin and gemcitabine, are widely used in pancreatic cancer. Recently, many novel therapies were tested, including molecular targeted therapies. Among them, only erlotinib, which targets the epidermal growth factor receptor, has been shown to improve survival in combination with gemcitabine as compared to gemcitabine alone . Still, innate and acquired drug resistance remains unknown . KRAS and BCL2 have been involved in drug resistance in pancreatic cancer . Mutation of KRAS results in constitutive activation of the RAS/RAF/MEK/ERK pathway, with loss of epidermal growth factor receptor signaling control, rendering inhibitors of EGFR ineffective . KRAS also induces PEAK1 amplification and desensitizing cancer cells to trastuzumab and gemcitabine . Gemcitabine stabilized mutant p53 protein in the nuclei and induced chemoresistance . KRAS knockdown abolished the insulin-like growth factor-1-induced ERK pathway in KRAS-mutated cancer cells and enhanced the therapeutic efficacy of everolimus . The overexpression of c-Met, a mesenchymal-epithelial transition factor gene, promotes anti-apoptotic effects via PI3K-AKT activation, consequently reducing the sensitivity towards gemcitabine [101,102]. SIRT1 was overexpressed in pancreatic cancer and downregulation of SIRT1 could enhance chemosensitivity .
It has been reported that p53 is strongly associated with tumor differentiation and presence of locoregional recurrence . Loss of p16CDKN2A was associated with lymphatic invasion and postoperative widespread metastasis . A significant correlation was found among SMAD4/DPC4 and tumor size, lymphatic invasion and lymph node metastasis. Loss of SMAD4/DPC4 was significantly associated with shorter overall survival. Multivariate analysis revealed that loss of SMAD4/DPC4 was an independent and significant poor prognostic factor for overall and disease-free survival .
With better understanding of chemoresistance in pancreatic cancer, new approaches have emerged. For example, to overcome the resistance mechanism related to the RAS/RAF/MEK/ERK pathway, several novel inhibitors including MEK inhibitors, BRAF inhibitors, HSP90 inhibitors, KRAS-directed immunotherapy, mTOR inhibitors and several combinations thereof have been tested . An interplay between the N-terminus domain of secreted protein and rich in cysteine (SPARC), BCL2 and CASPASE-8 helps to augment apoptosis and consequently increase the response to treatment . Combination treatment of gemcitabine and p53-reactivating molecules (CP-31398 and RITA) inhibits the proliferation of pancreatic cancer cells and induced apoptosis . The inhibition of poly(adenosine diphosphate-ribose) polymerase (PARP) might be a potential synthetic lethal therapeutic strategy for the treatment of cancers with specific DNA repair defects, including carriers with BRCA1 or BRCA2 mutation .
Although pancreatic cancer remains as a dismal disease, genetic studies indicate that genetic alterations might be useful for diagnosis, treatment and prognosis (table 2). Mutations including KRAS, p16CDKN2A, p53 and SMAD4 might provide a molecular basis for the progression from PanINs to pancreatic cancer (fig. 1). With more genes being identified, diagnosis and staging of pancreatic cancer and precursor lesions according to gene mutations might be more discernible and advisable for further treatment. Moreover, it might help to characterize and stratify cancer patients into subgroups for individualized treatments.
This work was supported by the Natural Science Foundation of China (81472214, 81272678) and by the Science and Technology Innovation Team of Zhejiang Province in China (2013TD13).
The authors declare that there are no conflicts of interest.