KRAS and GNAS mutations and p53 overexpression in biliary intraepithelial neoplasia and intrahepatic cholangiocarcinomas




Similar to the pancreatic intraepithelial neoplasia (PanIN)-pancreatic carcinoma sequence model, intrahepatic cholangiocarcinoma (ICC) also reportedly follows a stepwise carcinogenesis process through the a precursor lesion: biliary intraepithelial neoplasia (BilIN). For this study, the authors investigated the status of v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) and GNAS complex locus (GNAS) mutations and tumor protein 53 (p53) overexpression in the stepwise process of cholangiocarcinogenesis.


Thirty patients with hepatolithiasis were surveyed, and their lesions were categorized as follows: non-neoplastic large bile duct (LBD) (n = 12), peribiliary gland (PBG) (n = 9), BilIN-1 (low-grade dysplasia; n = 12), BilIN-2 (high-grade dysplasia; n = 16), and BilIN-3 (noninvasive or in situ carcinoma; n = 10). KRAS mutation at codons 12 and 13 and GNAS mutations at codons 601 and 602 were analyzed using genomic DNA extracted from isolated lesions by laser capture microdissection. Immunohistochemical expression of p53 also was evaluated in BilIN lesions, ICCs, and extrahepatic cholangiocarcinomas (ExCCs).


A prevalence of KRAS mutations was identified in patients with ICC (31.5%), BilIN-3 (30%), and BilIN-2 (43.8%) compared with BilIN-1 (25%). Furthermore, KRAS mutations were detected in LBD lesions (41.7%) and PBG lesions (44.4%), and these mutations were observed with greater frequency in patients who had BilIN with KRAS mutations. GNAS mutations were not identified in any of the ICCs or other lesions examined. The overexpression of p53 was not identified in BilIN lesions and was less frequent in ICCs (18.2%) compared with ExCCs (38.1%) and gallbladder carcinomas (61.5%).


KRAS mutations, which were present in approximately 33% of BilIN lesions, occurred as an early molecular event during the progression of BilIN to ICC, whereas p53 overexpression was identified as a late molecular event. Furthermore, the current results indicted that BilIN also may arise from LBD and PBG lesions in patients who have hepatolithiasis with KRAS mutations. Cancer 2013. © 2013 American Cancer Society.


Cholangiocarcinomas (CCs) are highly malignant tumors with a dismal prognosis, and they account for 10% to 20% of deaths from hepatobiliary maliganacies.1 Recently, it was reported that intrahepatic CCs (ICCs) in particular have increased incidence and mortality rates worldwide, with an increase in incidence up to 165% in the United States over the last few years.2-4 Although ICCs usually arise in normal liver, chronic cholangitis, which may occur in the setting of hepatolithiasis or primary sclerosing cholangitis, occasionally precedes the development of ICC.5 CC arising in chronic cholangitis progresses through 2 types of preneoplastic or early neoplastic biliary lesions: flat and papillary types.6 The former type is characterized by biliary intraepithelial neoplasia (BilIN), previously called biliary epithelial dysplasia.7, 8 This type is frequently identified in hepatolithiasis5 and also has been identified in patients with primary sclerosing cholangitis, alcoholic cirrhosis, and hepatitis C with and without ICC.9, 10 According to the criteria for BilIN, BilIN-1 corresponds to low-grade dysplasia, BilIN-2 corresponds to high-grade dysplasia, and BilIN-3 corresponds to noninvasive or in situ carcinoma11; and a stepwise progression through BilIN-1, BilIN-2, and BilIN-3 to invasive ICC has been proposed in hepatolithiasis.7, 8, 11, 12

Recent progress suggests that the biliary tract and pancreas have similar pathologic features; for example, there are many similarities between BilIN and PanIN, intraductal papillary neoplasm of the bile duct (IPNB), and intraductal papillary mucinous neoplasm of the pancreas (IPMN).13 Several studies indicate that the biliary tree contains stem cell compartments for the liver, pancreas, and bile duct that persist into adulthood.14 The biliary tract exerts the potential for pancreatic differentiation; and the ventral pancreatic duct, which eventually becomes the main pancreatic duct, originates from the common bile duct.13 The common origin may explain the similarity of diseases that occurs in both the biliary tract and pancreas. Within the pancreatic model, v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation, tumor protein 53 (p53) mutation, and the sonic hedgehog signaling pathway all have been implicated as key players in pancreatic carcinogenesis.15-18 KRAS mutation is an early event within PanIN and occurs in up to 90% of early PanIN and in 95% of pancreatic adenocarcinomas (PDACs).18, 19 In contrast, p53 mutation is a late event in PanIN and is present in the majority of PDACs.17, 19 A recent report demonstrated that greater than 99% of the earliest stage, lowest grade, PanIN-1 lesions contained mutations in KRAS, cyclin-dependent kinase inhibitor 2A (p16/CDKN2A), GNAS complex locus (GNAS), or v-raf murine sarcoma viral oncogene homolog B1 (BRAF).19 GNAS encodes the α-subunit of the stimulatory G-protein (Gαs), which mediates the regulation of adenylate cyclase activity through G-protein–coupled receptors. Activating mutations of GNAS reportedly are prevalent in pituitary adenomas and in intraductal papillary mucinous neoplasms of the pancreas,20 whereas they are absent or rare in most other tumor types. There have been few studies on GNAS mutation in CC21; and, to our knowledge, there have been no studies on BilIN.

It has been hypothesized that pancreatic duct glands (PDGs) give rise to PanIN.22 We also have reported that florid, reactive, and atypical peribiliary glands (PBGs) often are observed in hepatolithiasis preceding the development of CC.5, 23 Given the similar morphology to PDGs, the PBGs also may take part in the development of biliary metaplasia and neoplasia. The role and genetic alterations of PBGs as possible precursors to CC have yet to be investigated. Because of the similarities between the precursor and background lesions involved in PDAC and CC, in this study, investigated the role of key PanIN-PDAC–associated molecular pathways (KRAS, GNAS, and p53) in BilIN, PBGs, and ICCs. We hypothesize that similar mutations present in the PanIN-pancreatic carcinoma sequence also may play a role in cholangiocarcinogesis. Although currently there is greater recognition and detection of the presence of BilIN in chronic biliary diseases, few studies have investigated the mutation status of KRAS and GNAS in BilIN lesions during the stepwise progression toward ICC. In this study, we examined the mutation status of KRAS and GNAS and the expression of p53 in BilIN and PBGs in patients with hepatolithiasis and ICC.


Patient Selection


Patients with hepatolithiasis who attended Kanazawa University Hospital and affiliated institutions from 1980 to 2009 were surveyed, and their lesions were categorized into the following types: non-neoplastic large bile duct (LBD), PBG, BilIN-1, BilIN-2, BilIN-3, and ICC. The Ethics Committee of Kanazawa University approved this study.

Nonhepatolithiasis intrahepatic cholangiocarcinoma and other pancreatic-biliary tract carcinomas

Patients who were diagnosed between 1990 and 2010 with ICC (n = 56), extrahepatic CC (ExCC) (n = 42), pancreatic carcinoma (n = 34), and gallbladder carcinoma (n = 13) were reviewed and selected from the files of Kanazawa University Hospital and used for comparison.

Control group

Normal bile ducts and peribiliary glands from 10 nonprimary hepatic or biliary diseased livers, from either resection specimens or autopsy specimens, also were used as controls. These specimens were fixed in 10% buffered formalin and embedded in paraffin, and 10 serial 3-μm thin sections and 5 serial 10-μm thin sections on membrane-coated slides for laser microdissection were cut from each paraffin block.

Microdissection and KRAS and GNAS Mutation Analysis

Bile duct lesions and PBGs were microdissected using the Leica AS LMD Laser microdissection system (Leica Microsystems, Wetzlar, Germany) according to the manufactures' instructions from air-dried, hematoxylin and eosin-stained, 10-μm sections on membrane-coated slides. Genomic DNA was then extracted from the microdissected tissues using the QIAamp DNA Micro Kit (Qiagen, Hilden, Germany). For PDACs and ICCs, tumor tissues were observed and scrapped off of 2 or 3 whole serial sections (3 μm), and DNA was isolated using the QIAMP DNA KIT (Qiagen). Isolated DNA was then subjected to polymerase chain reaction (PCR) amplification of the region of the KRAS2 gene that contained codons 12 and 13. The forward and reverse primers were 5′-AGGCCTGCTGAAAATGACTG-3′ and 5′-ATCAAAGAATGGTCCTGCAC-3′, respectively. Amplifications were done by initial denaturation at 94°C for 3 minutes followed by 35 cycles of denaturation at 94°C for 1 minute, annealing at 58°C for 1 minute, and extension at 72°C for 1 minute; followed by a 10-minute final extension at 72°C using TaqDNA polymerase (Takara EX Taq; Takara Bio, Kyoto, Japan). Because of the low DNA concentrations in most specimens, we conducted a second-round, 20-cycle PCR at the same conditions described above using a nested PCR primer pair with the forward and reverse primers 5′-GCCTGCTGAAAATGACTGAA-3′ and 5′-GAATGGTCCTGCACCAGTAA-3′, respectively. In addition, PCR amplification of the region of the GNAS gene coding codon 201 was performed using the forward and reverse primers 5′-ACTGTTTCGGTTGGCTTTGGTGA-3′ and 5′-AGGGACTGGGGTGAATGTCAAGA-3′, respectively. These PCR products were then purified using the Qiagen PCR purification kit and were sequenced using the Big Dye cyclic sequencing kit and the ABI 310 sequencer (Applied Biosystems, Forster City, Calif).


Immunohistochemical staining on the selected cases described above was performed using the M7001 antibody for p53 (Dako, Carpinteria, Calif; 1:100 dilution). In brief, after pretreatment using a microwave with citrate buffer, pH 6.0, at 95°C for 20 minutes and blocking endogenous peroxidase, the sections were incubated with the primary antibody at 4°C overnight. The Envision+ solution for mouse and rabbit (Dako) was then applied for 30 minutes at room temperature. The reaction products were observed using 3-3′-diaminobenizidine tetrahydrochloride (Sigma Chemical Company, St. Louis, Mo) and H2O2. The sections were then lightly counterstained with hematoxylin. Similar dilutions of control mouse or rabbit immunoglobulin G (Dako) were applied instead of the primary antibody as a negative control. Positive and negative controls were routinely included. Expression of p53 was scored as positive according to the presence of nuclear staining in >25% of the lesion.

Statistical Analysis

The Wilcoxon rank-sum test was used for statistical analysis of the differences in human studies. P values < .05 were considered statistically significant.


KRAS Mutation

Among the patients who had hepatolithiasis, 30 had available paraffin blocks. Twenty-seven blocks revealed hepatolithiasis with BilIN, and the remaining 3 blocks revealed hepatolithiasis without BilIN. Of these patients, 12 patients with LBD lesions, 9 with PBG lesions, 12 with BilIN-1 lesions, 16 with BilIN-2 lesions, and 10 with BilIN-3 lesions underwent successful microdissection, DNA extraction, PCR amplification, and DNA sequencing to evaluate KRAS mutation status. KRAS mutations were detected in 13 of 27 patients (48%) who had hepatolithiasis with BilIN but were not detected in 3 patients who had hepatolithiasis without BilIN. The numbers of KRAS mutations (all GGT to GAT in codon 12) in each type of lesion were as follows: 5 of 12 LBD lesions, 4 of 9 PBG lesions, 3 of 12 BilIN-1 lesions, 7 of 16 BilIN-2 lesions, and 3 of 10 BilIN-3 lesions (Fig. 1). Four of 6 patients who had KRAS mutations with LBD and PBG lesions also had KRAS mutations with BilIN lesions (Fig. 1D). Thirty-eight patients who had ICC underwent whole-section DNA extraction, PCR amplification, and DNA sequencing. Of these 38 patients, 12 had KRAS mutations (GGT to GAT in codon 12).

Figure 1.

v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation status is illustrated in samples of biliary intraepithelial neoplasia (BilIN) from patients with hepatolithiasis and intrahepatic cholangiocarcinoma (ICC). (A) These are examples of hematoxylin and eosin (H&E)-stained slides of BilIN (Left) before and (Right) after laser capture microdissection (H&E staining; original magnification, ×100) (B) These are images of (Left) representative histology and (Right) a sequencing trace of BilIN-1 (low-grade dysplasia) without KRAS mutation and BilIN-2 with KRAS mutation. An arrow indicates a missense mutation; WT, wild type (H&E staining; original magnification, ×100). (C) The prevalence of KRAS mutation is illustrated in large bile duct (LBD), peribiliary gland (PBG), BilIN-1, BilIN-2 (high-grade dysplasia), and BilIN-3 (noninvasive or in situ carcinoma) lesions from patients with hepatolithiasis and in ICC. (D) These are KRAS mutation profiles of patients who had hepatolithiasis with LBD, PBG, BilIN-1, BilIN-2, and BilIN-3 lesions. Blue boxes indicate wild type; yellow boxes, KRAS mutation; gray boxes, a focus that was not available or was unable to be amplified for sequencing.

GNAS Mutation

GNAS mutation was not detected in any of the LBD, PBG, or BilIN lesions examined. Furthermore, GNAS mutation was not detected in any of the ICCs examined (Fig. 2).

Figure 2.

GNAS complex locus (GNAS) mutation in biliary intraepithelial neoplasia (BilIN) is illustrated. Images show (Left) representative histology and (Right) a sequencing trace of a BilIN-2 lesion (high-grade dysplasia) without GNAS mutation. WT indicates wild type.

p53 Immunohistochemistry

Immunohistochemical overexpression of p53 was evaluated in 31 patients who had BilIN-1, BilIN-2, and BilIN-3 lesions and also in 33 ICCs, 42 ExCCs, and 13 gallbladder carcinomas. Overexpression of p53 was not identified in any BilIN lesions (Fig. 3). The overexpression of p53 was less frequent in ICCs (18.2%) than in ExCCs (38.1%) (Fig. 3) and gallbladder carcinomas (61.5%; P < .05). Of the 6 ICCs that had p53 overexpression, 2 also had a KRAS mutation.

Figure 3.

Tumor protein 53 (p53) expression is observed in samples of biliary intraepithelial neoplasia (BilIN) from patients who had hepatolithiasis and intrahepatic and extrahepatic cholangiocarcinoma (ExCC). (Left) A BilIN-2 lesion (high-grade dysplasia) is negative for p53 expression. (Right) An Ex-CC has uniform and strong p53 expression (immunostaining for p53 and H&E; original magnification, ×200).


In this study we observed that KRAS mutations take part in the stepwise progression of BilIN to ICC, similar to the PanIN-PDAC progression model. Also, for the first time, we observed that a large proportion of BilIN lesions (33%) had KRAS mutations. The prevalence of KRAS mutations was highest in BilIN-2 lesions compared with BIlIN-1 and BilIN-3 lesions in this study. The prevalence of KRAS mutations was 31.5% in ICCs, which was lower than that observed in PDACs, in agreement with previous reports.21, 24, 25 Although it is unclear why the prevalence of KRAS mutations differs between PDACs and ICCs, it is plausible that at least some ICCs may be associated with a KRAS mutation as an early event in carcinogenesis.

Furthermore, to our knowledge, our study is the first to report the finding of KRAS mutations in approximately 40% of LBD and PBG lesions. It is noteworthy that the majority of the LBD and PBG lesions that had KRAS mutations occurred in patients who also had BilIN lesions with KRAS mutations. Although the number of patients was limited, KRAS mutations were not detected in any patients who had hepatolithiasis without BilIN. These findings suggest that BilIN lesions with KRAS mutations may arise from LBD and PBG lesions that have KRAS mutations in patients with hepatolithiasis. To our knowledge, there have been no other studies demonstrating KRAS mutations in the background of bile duct and PBG lesions in patients with BilIN and ICC. It has been hypothesized that, in the pancreas, PDGs give rise to PanIN.22 A recent study suggested that PanIN may arise in the centroacinar-acinar region, possibly through a process of acinar-ductal metaplasia in mouse models with KRAS G12D mutations and in pancreatic tissue from patients with familial PDAC.26 Similar to the pancreatic model, BilIN also may arise from LBD and PBG lesions with KRAS mutations. Because PBG lesions also are present in the extrahepatic bile duct and have been suggested as precursors of ExCC, KRAS mutations may take part in the cholangiocarcinogenesis of ExCC. This important point remains to be elucidated in the near future.

Although our current study confirmed the presence of a KRAS mutation-dependent pathway in the BilIN-ICC model in approximately 33% of lesions, the other key molecular events that occur in the remaining 66% of lesions remain unknown. We examined GNAS mutation as a candidate gene; however, we did not detect any GNAS mutations in ICCs or in BilIN lesions. GNAS mutations are identified frequently in IPMNs, whereas they are absent to rare in PDACs.20 In a recent study, GNAS mutations were detected in 9.3% of liver fluke-associated CCs.21 Given the histologic similarities of IPNBs and IPMNs, it is possible that IPNBs also may harbor GNAS mutations; however, we did not examine IPNBs or CCs associated with IPNB. Our study suggests that GNAS mutations do not appear to participate in cholangiocarcinogenesis through flat precursor lesions (BilIN). Another candidate gene that may be involved in cholangiocarcinogenesis is myeloid/lymphoid or mixed-lineage leukemia 3 (MLL3), which encodes a histone-lysine N-methyltransferase and has a high prevalence of mutation in liver fluke-associated CC.21 Further study is warranted to elucidate a possible role for MLL3. We did not detect p53 overexpression in LBD, PBG, or BilIN lesions, although it was detected in 18.2% of ICCs. This finding suggests that p53 mutation may be a late event in cholangiocarcinogenesis, as reported previously.25, 27 We observed no correlation between p53 overexpression and KRAS mutation in this study.

In conclusion, KRAS mutation occurred in approximately 33% of our patients who had BilIN and was identified as an early molecular event during the progression of BilIN to ICC in patients with hepatolithiasis, whereas p53 overexpression was identified as a late molecular event. Furthermore, we observed that a subset of BilIN lesions may arise from precursor LBD and PBG lesions, in which we also observed KRAS mutations.


This study was supported in part by a Grant-in Aid for Scientific Research (B) from the Ministry of Education, Culture, Sports, and Science and Technology of Japan (22390067).


The authors made no disclosures.