The frequency of KRAS and BRAF mutations in intrahepatic cholangiocarcinomas and their correlation with clinical outcome

Abstract

The incidence of intrahepatic cholangiocarcinoma (ICC) is increasing worldwide. The prognosis of ICC is poor and a better understanding of ICC tumor biology is needed to more accurately predict clinical outcome and to suggest potential targets for more effective therapies. v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) and BRAF are frequently mutated oncogenes that promote carcinogenesis in a variety of tumor types. In this study, we analyze a large set of ICC tumors (N = 54) for mutations in these genes and compare the clinical outcomes of wild type versus KRAS and BRAF mutant cases. Out of 54 cases, 7.4% were mutant for KRAS, 7.4% were mutant for BRAF and these were mutually exclusive. These mutant cases were associated with a higher tumor stage at time of resection and a greater likelihood of lymph node involvement. These cases were also associated with a worse long-term overall survival. Therefore, testing for KRAS and BRAF mutations could be a valuable adjunct in improving both prognosis and outcome stratification among patients with ICC.

Introduction

Cholangiocarcinoma is the second most common type of liver cancer after hepatocellular carcinoma [1]. In the United States, there are 5000 cases annually, constituting nearly 3% of all gastrointestinal cancers [2]. Approximately two-thirds of cases arise from the extrahepatic biliary tree while the remainder arise within the liver. Although intrahepatic (ICC) and extrahepatic cholangiocarcinomas (ECC) arise from similar epithelium, each tumor type has distinct risk factors, clinical presentations, genetic changes and management.

Alarmingly, the incidence of ICC is rising while the incidence of ECC has remained stable [3]. The factors driving the increased incidence of ICC are not well-understood, but increases in hepatitis C infection and nonalcoholic fatty liver disease may be partially responsible [4]. Other risk factors for ICC include hepatitis B infection, primary sclerosing cholangitis, advanced age, hepatolithiasis, chemical exposure and liver fluke infection [2].

ICC has a dismal prognosis and surgical resection is the only therapy with the possibility of long-term cure. Unfortunately, most patients present with advanced lesions and are not suitable candidates for resection [1]. While chemotherapy provides improved quality of life and a modest increase in survival, it is not curative [5]. At these later stages, the median survival for patients with unresectable ICC ranges from five and eight months [6]. Even with surgical resection, five-year overall survival is still poor and ranges from only 14% to 40% [1]. Certainly, a better understanding of cholangiocarcinoma tumor biology is needed to advance therapeutic strategies to improve survival.

The molecular alterations that drive tumorigenesis in cholangiocarcinomas are beginning to be identified. Single-gene studies have identified a variety of genetic derangements, most of which involve well-known tumor genes including KRAS, BRAF, TP53 and epidermal growth factor receptor (EGFR) [7–12]. More recently, several studies have used comprehensive genomic and proteomic analyses to identify additional molecular determinants as well as to suggest multi-gene signatures to stratify cholangiocarcinomas into different functional and prognostic groups. In many of these studies, mutations in KRAS and BRAF are reported, however, the frequency of these mutations varies considerably between studies. With respect to KRAS, the reported frequency ranges from as low as 8% to as high as 53% [9,13]. Similarly, the frequency of BRAF mutations has been reported over a wide incidence ranging from 0% to 22% [7,8].

The large variance across studies is likely the result of small samples sizes. Furthermore, some studies report aggregated data on cases consisting of both ICC and ECC, which have very different frequencies of KRAS mutation as suggested by two recent reports [14,15]. Similarly, a subset of liver neoplasms also have both neoplastic ductal and hepatocellular epithelial elements. These tumors almost certainly have different underlying pathogenesis, possibly arising from hepatic stem cells rather than more differentiated cholangiocytes[16]. To more accurately define the mutational frequency, a more carefully defined study with a large and carefully defined set of intrahepatic tumors is needed.

An accurate report of KRAS and BRAF mutations has important clinical implications since epidermal growth factor receptor (EGF-R) inhibitor therapy is largely ineffective in the presence of KRAS and BRAF mutations [17–19]. On the other hand if KRAS or BRAF mutations are frequent, these should be specifically targeted.

Moreover, the prognostic and predictive implications of KRAS and BRAF mutations are not defined. There are no studies that report clinical follow-up data in conjunction with KRAS and BRAF mutational status. One study examined a mixed population of sixty-nine ICC and ECC cases and mutations in KRAS and BRAF were not associated with differences in prognosis [8]. Given the limited data on KRAS and BRAF in ICC, additional studies linked to clinical outcome are clearly indicated. As such, in the present study we sought to examine a carefully defined set of patients with ICC tumors to define the frequency of KRAS and BRAF mutations and compare the mutational status with patient demographics, tumor pathological characteristics and clinical outcome.

Materials and Methods

Patients and tissue samples

The use of human tissues in this study was approved by the Johns Hopkins Institution Review Board. All human liver samples are from patients undergoing surgical resection at the Johns Hopkins Hospital, Baltimore, MD. Fifty-four cases of ICC derived from patients treated with curative intent surgical resection at The Johns Hopkins Hospital from 1990 through 2011 were identified. From these surgical cases, formalin-fixed paraffin-embedded tumor tissue was retrieved. Data on patient and treatment-related variables were also obtained. Specifically, information on tumor number, size, as well as margin and nodal status were ascertained. Tumor staging was assigned according to the AJCC cancer staging manual, 7^th edition [20]. Date of last follow-up, vital status and date of recurrence were collected on all patients.

Pathology Review

The pathology of all cases was reviewed by an experienced hepatobiliary pathologist to verify the tumor type and evaluate other pathologic parameters including degree of differentiation, vascular invasion, perineural invasion, biliary invasion and satellite lesions. The pathology was independently reviewed by two additional pathologists. All tumors were composed of glandular differentiated cells. No tumors contained areas with hepatocellular differentiation (biphenotypic tumors).

DNA Extraction and Pyrosequencing

Paraffin embedded blocks with tumor tissue were selected and tumor DNA was macro-dissected and isolated using the Pinpoint DNA Isolation System (Zymo Research, Irvine, CA). Extracted DNA was then purified using the QIAamp DNA Mini Kit (Qiagen Valencia, CA), according to the manufacturer’s “Blood and Body Fluid Spin” protocol. Pyrosequencing for detecting codon 12 and 13 mutations of the KRAS gene was performed as described previously [20]. Pyrosequencing for detecting codon 600 mutations of the BRAF genes was conducted using forward primer 5′-GAAGACCTCACAGTAAAAATAG-3′ and biotinylated reverse primer 5′-ATAGCCTCAATTCTTACCATCC-3′ for PCR and primer 5′ GACCTCACAGTAAAAATAGGTGATTTTG-3′ for sequencing. PCR conditions were: 95°C for 15 min, 42 cycles of 95°C for 20 sec, 53°C for 30 sec and 72°C for 20 sec, and 72°C for 5 min. The nucleotide dispensation order for pyrosequencing was: 5′-GTAGCTAGCTATCAGCATCGACTCTCGATGAGTG-3′. The limit of mutant allele detection by pyrosequencing was 5%, based on sequencing colon margin of resection negative control samples.

Statistical analysis

Summary statistics were obtained using established methods and presented as percentages of median values. Categorical and continuous variables were compared using r exact and Kruskal-Wallis tests, respectively. Median overall and recurrence free survival times were estimated using the Kaplan-Meier method and differences were compared using the log-rank test, where p< 0.05 was considered statistically significant. All statistical tests were two-tailed. Statistical analyses were performed using SAS 9.3 (SAS Institute, Cary, NC).

Results

Frequency of KRAS and BRAF mutations

A total of fifty-four cases of ICC were evaluated. The demographic, disease and treatment characteristics are summarized in Table 1. Median patient age was 64 years and the cohort had a predominance of white females. The tumor from each case was sequenced for KRAS (codon 12 and 13) and BRAF (codon 600) mutations by pyrosequencing, with a mutant allele detection limit of 5% [20]. Representative pyrograms showing mutant KRAS and BRAF sequences are shown in Figure 1. We found that the tumor from four cases (7.4%) had mutations in KRAS and an additional, but distinct four cases (7.4%) had mutations in BRAF. All KRAS mutations were G12D substitutions and all BRAF mutations were V600E substitutions. When cases with either a KRAS or BRAF mutation were grouped together, we found that mutant cases were more likely to have a higher TNM stage at the time of surgical resection. Two of four (50%) KRAS mutant cases and two of four (50%) BRAF mutant cases presented as T4 lesions compared with only seven of 46 (16%) wild type (WT) lesions (p=0.03). Similarly, two of four (50%) KRAS mutant cases and two of four (50%) BRAF mutant cases presented with lymph node involvement at the time of surgical resection compared with only three of forty-six (7%) without either mutation (p=0.01). Evaluation of other pathologic parameters including tumor size, tumor multifocality, tumor differentiation, vascular invasion, perineural invasion, biliary invasion and satellite lesions showed no correlation with mutational status (all p>0.05).

Table 1.

Demographic, disease, and treatment characteristics of patients

	No Mutation (n=46)	KRAS (n=4)	BRAF (n=4)	P-value
Age, median (range)	64 (37–81)	67.5 (55–78)	59 (45–78)	0.12
Male gender	14 (30.4)	1 (25.0)	2 (50.0)	0.82
Race
White	39 (84.8)	3 (75.0)	4 (100)	0.75
Other	7 (15.2)	1 (25.0)	0
Body mass index, median	26.0 (16.0–44.1)	24.0 (20.9–28.1)	36.2 (26.6–45.9)
HBV infection	3(6.5)	0	0	1.00
Cirrhosis	4(8.7)	0	0	1.00
CA 19–9, median (range)	31 (11–56)	81 (–)	503 (1–1005)	0.08
Size, median (range)	5.5(2.5–19.0)	5.2 (3.7–5.7)	4.6(3.5–5.6)	0.22
Bilobar involvement	11(23.9)	1(25.0)	0	0.81
Nodes
N0	43 (93.5)	2(50.0)	2(50.0)	0.01*
N1	3 (6.5)	2 (50.0)	2(50.0)
AJCC stage
0	1 (2.2)	0	0	0.03*
1	28(60.9)	0	2(50.0)
2	10(21.7)	1(25.0)	0
3	0	1(25.0)	0
4	7(15.2)	2(50.0)	2(50.0)
Multiple tumors	9(19.6)	2(50.0)	1(25.0)	0.38
Liver resection
< Hemihepatectomy	11(23.9)	0	3(75.0)	0.18
Hemihepatectomy	24(52.2)	3(75.0)	1(25.0)
Extended Hepatectomy	11(23.9)	1(25.0)	0
Margins
R0	38 (82.6)	2 (50.0)	3(75.0)	0.20
R1	8(17.4)	2(50.0)	1(25.0)
Tumor Differentiation
Well	8(17.3)	2(50.0)	0	0.32
Moderate	20(43.5)	2(50.0)	2(50.0)
Poor	18(39.1)	0	2(50.0)
Vascular Invasion	11(23.9)	2(50.0)	1 (25.0)	0.52
Perineural invasion	7(15.2)	2(50.0)	0	0.24
Biliary invasion	6(13.0)	2(50.0)	0	0.19
Satellite lesions	7(15.2)	2(50.0)	1(25.0)	0.16
Intrahepatic metastases	2 (4.4)	0	0	1.00
Adjuvant therapy	20 (43.5)	3 (75.0)	1 (25.0)	0.39

^*p< 0.05

Figure 1.

Representative pyrograms of KRAS (A) and BRAF (B) mutations detected in ICC by pyrosequencing.

Clinical Outcome

Overall survival (OS) and recurrence free survival (RFS) were analyzed for each group and are summarized in Table 2. Median overall survival was 13.5 months for KRAS mutant cases and 24.4 months for BRAF mutant cases, compared to 37.3 months for wild type cases. Combining KRAS mutant cases with the BRAF mutant cases resulted in a median OS of 23.2 months (p = 0.05, compared with the wild type group). The Kaplan-Meier survival curves for these two groups are shown in Figure 2.

Table 2.

Description of patients with a mutation of KRAS or BRAF (n=4 each)

	Age	Sex	Race	Tumor Size	Nodal status	AJCC Stage	Overall Survival (m)	Median OS (m)	P-value OS (WT* vs. KRAS + BRAF)
KRAS	55	F	Asian	5.4	N0	4a	13.5	13.5	0.05
	66	M	White	5	N0	2	6.8
	78	F	White	3.7	N1	3c	8.4
	69	F	White	5.7	N1	4a	13.0
BRAF	45	F	White	4.5	N1	4a	85.4	24.4
	78	M	White	3.5	N0	1	25.6
	63	M	White	4.7	N0	1	23.2
	55	F	White	5.6	N1	4a	2.7

^*OS for WT = 37.3 months

Figure 2.

Kaplan-Meier plot demonstrating overall survival for patients with either KRAS or BRAF mutated tumors compared to wild type (p = 0.05).

Discussion

In this study, we examined the frequency of KRAS and BRAF mutations in a set of surgically resected intrahepatic cholangiocarcinomas. We identified KRAS mutations in 7.4% of cases and BRAF mutations in an additional 7.4% of cases. When correlated with pathological parameters and survival data, we found that either mutation is associated with higher TNM stage at presentation and decreased overall survival. These findings suggest that BRAF and KRAS mutational analysis could serve as a useful adjunctive tool for evaluating patients with ICC. Furthermore, BRAF and KRAS mutational analysis is commonplace in the workup of other tumor types and the infrastructure needed for this analysis is already widely available.

The identification of several BRAF mutated tumors in our data, in agreement with several previous studies, also suggests a potential application for targeted therapy in these cases. Vemurafenib, for example, has demonstrated dramatic antitumor activity in malignant melanomas with BRAF codon V600 mutations[22]. Perhaps, this agent could also be of use in the context of appropriately selected biliary malignancies. On the other hand, the 15% of patients with either KRAS or BRAF mutations would likely not be good candidates for chemotherapy trials involving EGF-R inhibitors. Given the low frequency of these mutations, this information would only benefit a subset of patients. However, any improvement in survival would be a significant step forward considering the current poor prognosis of this disease.

It is important to note that almost all studies investigating gene mutations in cholangiocarcinomas, including the current study, have used tissue obtained at the time of surgical resection. Since only a subset of patients presenting with ICC have disease limited enough to allow for surgery, these cases only represent a subset of total ICC, one enriched for less aggressive disease. Therefore, estimations about the frequency of gene mutations may underestimate the true incidence of KRAS and BRAF mutation if these genes are indeed associated with more aggressive and advanced disease. Further work is needed to determine if the characteristics of resectable tumors are representative of the larger set of all cholangiocarcinomas regardless of the amenability to resection.

Lastly, while our data suggest that mutations in KRAS and BRAF portend a poor prognosis, this is only applicable to a minority of cases, a mere 15% in our study. In a report by Andersen et al., the authors examined global gene expression in cholangiocarcinomas and reported that by integrating the KRAS mutational status with a 238-gene molecular signature, grouped all patients with mutated KRAS/BRAF into a poor prognostic group [14]. However, a KRAS mutation was not an independent prognostic factor in this circumstance. Interestingly, these data are similar to the outcomes presented in the current study. Collectively, these data suggest that stratification of patients into prognostic groups is feasible based on certain genetic profiles. Other means of evaluating ICC to identify additional potential gene candidates for better patient stratification into prognostic categories is still needed to address the majority of cases. Recent studies employing whole genome expression profiling have revealed fascinating insights into the tumor biology of cholangiocarcinomas and suggest that individual cases can be placed into prognostically significant categories based on an array of gene expression changes [13–15,23].