Lung cancer is the leading cause of death among all the malignant tumors worldwide.1 According to recent data from the WHO (Globocan 2012), in China lung cancer results in approximately 600,000 deaths every year. The prognosis of patients with advanced non-small cell lung cancer (NSCLC) is disappointing.2

Nowadays, the understanding of lung cancer has shifted from that of a simple tumor disease to neoplasms caused by a set of molecular driver mutations. Novel targeted agents against a specific mutant gene are designed to assist the standard platinum-based doublet chemotherapy of advanced NSCLC.2,3 The spectrum of a patient’s mutant genes is as important as the histological features of the tumor sample.

Treatment of patients with NSCLC harboring EGFR mutations best illustrates the therapeutic breakthrough in lung cancer, and has changed the way doctors manage NSCLC. After taking EGFR tyrosine kinase inhibitors (TKIs), these patients had a higher response rate and longer progression-free survival compared with single platinum-based doublet chemotherapy.4,5 The ALK TKIs have also proven to be more efficient in patients with NSCLC harboring ALK rearrangement.6,7

The BRAF gene was first discovered and identified by a group of Japanese scientists in 1988.8 It is a serine/threonine (Ser/Thr) kinase that is regulated by Ras GTPases to downstream control the protein of the mitogen-activated protein kinases (MAPK) family. The activation of Raf is initiated by the Ras-GTP complex with the Ras binding domain. Raf then changes conformation and recruits to the cell membrane. This promotes the Raf phosphorylation and stimulation of its Ser/Thr kinase activity, which triggers sequential phosphorylation and activation of MEK and ERK. The RAS-RAF-MEK-ERK-MAP signal pathway mainly functions to regulate cell proliferation, differentiation, and survival.9–11

Mutations in BRAF will cause the abnormal signal transduction in its pathway, making the cell more easier to proliferate, infiltrate, and metastasize.9,10 Somatic BRAF mutations have been reported in a variety of tumors. In melanomas, approximately 60 % of patients have BRAF mutations, and BRAF V600E mutations account for almost 80 % in these mutant patients.11 In vitro experiments have proven that the mutant BRAF V600E protein no longer requires upstream proteins for initiation or activation, and can activate its downstream signal protein constitutively. In a mouse model of lung cancer, the BRAF V600E mutation in lung epithelial cells helps the maintenance of the tumor.12

In NSCLC, according to published data, the BRAF mutation rate ranges from 0.7 to 10.0 %, while the BRAF V600E mutation accounts for <50 % among all the BRAF mutations.13–16 To date, BRAF mutations reported in lung cancer, including G466, G469, and L597, are nearly all located in the activation domain or the kinase domain. The substitution of amino acid leads to either elevated or reduced kinase activity of the mutant protein.17,18

Many of the BRAF inhibitors are already in use or under development, including vemurafenib, sorafenib and dabrafenib. Vemurafenib has been proven, with remarkable effect, in patients with melanoma harboring BRAF V600E mutations.19 Sorafenib, as an antiangiogenesis agent for treating NSCLC, can also work on BRAF mutations. Recently, several case reports demonstrated activity of vemurafenib and dabrafenib in BRAF-mutant lung cancer patients; such agents would have more clinical benefits than traditional chemotherapy.20,21

According to the latest reports,13,22 BRAF mutations are relatively frequent in Western populations (8–10 %); however, no large-scale data have reported on the situation in Eastern populations. Whether or not BRAF mutations are associated with poor prognosis is still unknown. We explore BRAF mutations in Chinese patients with lung adenocarcinoma, and the association between mutations and clinicopathological characteristics of these patients, in preparation for further study and target agent therapy.

Materials and Methods

Patients and Samples

All patients with newly diagnosed primary lung cancer were retrospectively selected from the Department of Thoracic Surgery, Fudan University Shanghai Cancer Center, between October 2007 and February 2013. All patients were pathologically confirmed as having lung adenocarcinoma, according to the new WHO classification of lung tumors, by two independent pathologists (YL and XS). No patients received neoadjuvant therapy, and all underwent complete surgical excision. Tumor tissues and normal tissues were sampled just after the surgical resection, and were immediately stored in liquid nitrogen. Written informed consent was obtained from the patients and the study was approved by the Ethics of Human Research Committee of Fudan University Shanghai Cancer Center.

RNA Extraction and Mutation Analysis

RNA was extracted from the frozen tumor and normal tissue according to the standard protocols (RNeasy Mini Kit; Qiagen, Hilden, Germany). Total RNA samples were reverse transcribed into complementary DNA using a Revert Aid First Strand cDNA Synthesis Kit (Fermentas, St Leon-Rot, Germany). EGFR (exons 18–22), KRAS (exons 2–3), HER2 (exons 18–21) and BRAF (exons 11–15) were amplified by polymerase chain reaction (PCR) using KOD-plus DNA polymerase and cDNAs. ALK rearrangement was detected using quantitative real-time PCR (qRT-PCR), with validation using fluorescent in situ hybridization (FISH), as previously reported.23

Clinicopathological Characteristics

Clinical characteristics, including sex, age at diagnosis, pathologic TNM stage according to the 7th edition of the lung cancer staging classification system, tumor differentiation, smoking status, and histologic subtypes of adenocarcinoma according to the new IASLC/ATS/ERS multidisciplinary classification of lung adenocarcinoma, were collected. Survival and disease relapse information were collected every 3 months after surgery in the clinic or by telephone.

Statistical Analysis

Pearson’s Chi squared test or Fisher’s exact test (when the count in any cell of a contingency table was less than required) was performed to detect the association between BRAF mutation and clinicopathological characteristics. The Kaplan–Meier method was used to estimate the survival curve, and the log-rank test was used to compare the survival data. Cox regression was used for multivariate analysis to assess the effect of covariates on relapse-free survival (RFS) and overall survival (OS). SPSS for Windows version 19.0 (IBM Corporation, Armonk, NY, USA) was used to process data; 0.05 was chosen as the type I error, therefore a p value <0.05 indicated statistical significance.

Results

Clinicopathological Characteristics

A total of 1358 patients were enrolled in our study cohort; 607 (44.7 %) were male and 751 (55.3 %) were female. The average age was 59.4 years, ranging from 22.4 to 84.3 years. There were 926 (68.2 %) never smokers, 74 (5.4 %) former smokers, and 358 (26.4 %) current smokers. The number of patients in pathological TNM stage I–IV was 738 (54.3 %), 161 (11.9 %), 401 (29.5 %), and 58 (4.3 %), respectively. Overall, tumor samples of 172 (12.7 %) patients were well-differentiated, 725 (53.4 %) were moderately differentiated, and 461 (33.9 %) were poorly differentiated. In histological subtypes, acinar predominant (44.9 %) was the most frequent, followed by solid predominant (16.9 %), papillary predominant (13.6 %), and lepidic predominant (8.8 %); the remaining subtypes were invasive mucinous adenocarcinoma (IMA) (6.0 %), minimally invasive adenocarcinoma (MIA) (2.4 %), adenocarcinoma in situ (AIS) (2.3 %), micropapillary predominant (1.6 %), and ‘other’ (3.5 %). More detailed clinicopathological characteristics are summarized in Table 1.

Table 1 Clinicopathological characteristics of the 1358 patients with lung adenocarcinoma
Full size table

Gene Mutation Spectrum

A total of 857 patients harbored the EGFR mutation, with frequency being as high as 63.1 %, followed by the KRAS mutation in 108 (8.0 %) patients, ALK rearrangement in 70 (5.2 %) patients, and HER2 mutation in 39 (2.9 %) patients. Only 20 patients were harboring the BRAF mutation, accounting for 1.5 % of all patients. (Electronic Supplementary Fig. 1). No patients with BRAF mutations had a concomitant mutation in EGFR, KRAS, HER2, or ALK.

BRAF Mutation Genotypes

A total of 11 different types of BRAF mutations were identified; five V600E mutations (25 %), four G469A/E/V mutations (20 %), three K601E mutations (15 %), three L597R mutations (15 %), and two G466A/V mutations (10 %), with the rest being N581I, G593S, and D594G mutations. Two BRAF mutations—N581I and G593S—were never reported (see Table 2 for more details).

Table 2 Individual patient characteristics harboring BRAF mutations
Full size table

Associations Between BRAF Mutations and Clinicopathological Variables

Among 1358 patients, statistical significance in smoking status was found between patients with or without BRAF mutations [odds ratio (OR) 3.28; 95 % CI 1.33–8.08; p = 0.008]. When subdividing patients into groups of less than or greater than 60 years of age, we found statistical significance remained in the less than 60 years age group (OR 5.29; 95 % CI 1.36–20.67; p = 0.012) but disappeared in the greater than 60 years age group (OR 2.10; 95 % CI 0.60–7.34; p = 0.195). When diagnosed before the age of 60 years, patients harboring BRAF mutations tended to have poor differentiation in tumor samples (70.0 vs. 35.1 %; p = 0.014) and were more likely to relapse (70 vs. 28 %; p = 0.008). Male patients had a higher probability of harboring BRAF mutations (OR 0.30; 95 % CI 0.08–1.18; p = 0.067), although there was less statistical significance (Table 3).

Table 3 Characteristics of patients less than 60 years of age with lung adenocarcinoma
Full size table

Among 20 BRAF-mutated patients, no statistical significance was found in sex, age, smoking status, TNM stage, tumor differentiation, or subtypes between patients with BRAF V600E or non-V600E mutations. Nevertheless, patients harboring V600E mutations had a higher frequency in IMA subtype (40 vs. 0 %; p = 0.053).

Clinical Outcomes

Overall, 875 qualified patients were involved in survival analysis, including 20 patients with BRAF mutations and 605 patients with EGFR, KRAS, ALK, or HER2 mutations. Follow-up data were collected from October 2007 to December 2014. During the follow-up, 9 of 20 patients with BRAF mutations had recurrence; in these nine patients, seven were less than 60 years of age, and only two were greater than 60 years of age (OR 0.11; 95 % CI 0.01–0.84; p = 0.035). The median RFS of patients with BRAF mutations was 21.5 months compared with 47.8 months in the entire series. Relapse-free time was significantly shorter for patients harboring BRAF mutations compared with patients harboring EGFR mutations (p = 0.011) and patients with ALK mutations (p = 0.026). Cox regression in RFS analysis also confirmed smoking history, advanced stage, poor differentiation, and BRAF mutation as risk factors for recurrence. Two patients with BRAF mutations died during follow-up. No statistical significance was found in OS between patients with BRAF mutations and patients with other mutations; however, patients with EGFR mutations lived longer than patients without any listed mutations (p = 0.011) (Fig. 1; Table 4).

Fig. 1
figure 1

Kaplan–Meier curve for relapse-free survival and overall survival in 875 patients with lung adenocarcinoma. In relapse-free survival, curve difference between BRAF and ALK, and between BRAF and EGFR, is statistically significant. In overall survival, curve difference between EGFR versus WT (wide type) and Other is statistically significant

Full size image
Table 4 Univariate and multivariate survival analysis of 875 patients with lung adenocarcinoma
Full size table

Discussion

In the present study, we reported BRAF mutations in 1.5 % of patients with lung adenocarcinoma. The prevalence of BRAF mutations ranged from 0.7 to 10.0 % in different studies.13,14,16 Japanese scientists reported that they detected only 0.8–1.3 % of BRAF mutations in NSCLC, and scientists in Taiwan reported only 0.7 % of BRAF mutations in lung adenocarcinoma. Some scientists attributed this low frequency to ethnic difference and high frequency of EGFR mutations in Asian people.14,24 Our mutation rate of 1.5 % was consistent with these previous Asian reports.

In 20 of the BRAF mutations, five were V600E mutations (25 %); the proportion was much smaller than that in melanomas (>90 %) but was still predominant in NSCLC, which was in agreement with other studies (31–81 %).11,24,25 In V600E mutations, two in five were females, and in non-V600E mutations, six in 15 were females (40 %), compared with the whole cohort (743 in 1358; 55 %). There seemed to be no association between sex and BRAF V600E/non-V600E mutations. Interestingly, Marchetti et al. found V600E mutations were more prevalent in females,15 while in contrast, others reported equal distribution of sex with V600E mutations.

We found a strong relationship between smoke and BRAF mutations (60 % of patients were smokers), especially in patients less than 60 years of age (70 % of patients were smokers). A recent study with a larger series of BRAF mutations also revealed that the majority of patients with BRAF mutations were smokers (92 %);26 therefore, we have reason to believe that smoke is an important influencing factor in BRAF mutations.

In our study, patients less than 60 years of age with BRAF mutations also tended to have poor differentiated tumors (70 vs. 35 %; p = 0.031), while patients greater than 60 years of age did not follow this pattern (32 vs. 20 %; p = 0.67). Remarkably, these BRAF-mutant patients in our study had a statistically significant shorter relapse-free time compared with patients with EGFR and ALK mutations, which showed BRAF mutations may indicate a poor prognosis. Taking the poor differentiation into account, this may partially explain the high recurrence rate in young patients (70 vs. 20 %), and indicated that when patients get older, the tumor tends to mildly differentiate and have a lower chance of relapse. No significance was found in OS for patients either with or without BRAF mutations or with different kinds of BRAF mutations. However, Litvak et al. concluded that patients with advanced lung adenocarcinomas harboring V600 mutations had an improved OS compared with those with non-V600 mutations. Such a surprising result may partially be due to half of those V600 mutation patients receiving a BRAF inhibitor agent during treatment.26

In a recent systematic review and meta-analysis of patients with NSCLC harboring BRAF mutations, the researcher collected 170 mutant patients in a total of 5599 patients; the mutation rate was 3 %. It was quite certain that the BRAF V600E mutation was more frequent in woman and was related to never smokers. However, no significant difference was found in sex, smoking status or stage between BRAF mutant patients and BRAF wide-type patients.27

Since our study only focused on BRAF mutations in lung adenocarcinoma, the prevalence of BRAF mutations in squamous cell carcinoma (SCC) for Chinese people was still unclear. According to mainstream research, a rare BRAF mutation has been detected in SCC. Kinno et al.24 and Marchetti et al.15 only reported one BRAF mutation in SCC. Surprisingly, in September 2012, Hammerman et al. reported that a comprehensive genomic detection in SCC revealed a 4 % higher frequency of BRAF mutations, with all non-V600E mutations.28 To verify this suspicious data, more research on BRAF mutations in SCC is required. BRAF copy number analysis was also not included in our study. Remarkably, Japanese scientists found 10.3 % of patients with lung adenocarcinoma had increased BRAF copy number. These variations were all presented in BRAF V600E cases and might serve as a marker of the more aggressive behavior of this type of mutation. If this was the case, further study is needed.29,30

In our 20 BRAF mutations, N581I and G593S were newly reported. After reviewing the published data, more than 50 BRAF mutations in lung cancer have been discovered to date. Mutations such as V600, G466, G469, D594, and L597 were frequently reported. BRAF mutations were believed to be found mainly in exon 11 and exon 15.11 Strikingly, most of the mutations are located in the P loop (AA: 464–471) or the activation segment within or near the Asp-Phe-Gly (DFG) motif (AA: 594–600). In vitro experiments showed an enhanced ability of kinase in BRAF to directly phosphorylate its downstream protein MEK in some of the mutations (G469A, V600E), which may promote the proliferation of cancer cells. On the other hand, some mutations (G466V, G596R) inactivated the BRAF kinase, while the rest did not have an impact on the BRAF kinase.17,18 Those mutations that induce decreased or intermediate BRAF kinase activity would not lead to a better outcome because they still stimulated the downstream proteins through other distinct mechanisms based on further studies. Activated mutations can directly phosphorylate and activate MEK, as previously mentioned; however, the inactivated mutations could stimulate CRAF activity, which could then ultimately activate MEK, and finally complete the signal transduction in cancer cells.31

For patients with BRAF V600E mutations, second-generation BRAF inhibitors, including vemurafenib and dabrafenib, would have selective efficiency on the mutant BRAF kinase. Dabrafenib has shown its efficacy in melanoma, anaplastic thyroid cancer, and other solid tumors with the BRAF V600E mutation.32 An interim analysis of a phase II study of dabrafenib showed an improved disease control rate of 60 % for patients with stage IV NSCLC harboring BRAF V600E mutations, and a partial response of 40 % for these patients. This result was encouraging, and the following study is in progress (NCT01336634).33 Additionally, in 2012 and 2014, Gautschi et al.20 and Robinson et al.34 reported a BRAF V600E-mutated NSCLC patient responding to vemurafenib. Trametinib (GSK1120212), an MEK inhibitor that had good results for metastatic melanoma carrying the BRAF V600E mutation, was included in a phase III clinical trial (NCT01362296).35

As for some non-V600E BRAF mutations, the BRAF inhibitor seems inefficient theoretically. Indeed, in vitro studies did not provide an encouraging result for vemurafenib against low-activity G466V mutation.36 Therefore, inhibitors of MEK, which located downstream of BRAF proteins, were developed. However, an in vitro experiment showed that non-V600E BRAF-mutant cell lines were selectively sensitive to MEK inhibition. Drug combination was required in some circumstances, and some preclinical data showed its efficiency.37 By inhibiting the CRAF pathway, a pan-RAF inhibitor—sorafenib—may play an important role in treating non-V600E BRAF-mutated patients. Some preclinical data have proved the efficiency of sorafenib.31

Conclusions

In our study, we identified BRAF mutations in 1.5 % (20/1358) of patients with lung adenocarcinoma. Two point mutations in our study—N581I and G593S—were newly reported. BRAF mutations were more frequent in current smokers (OR 3.28; 95 % CI 1.33–8.08; p = 0.008), and young patients harboring BRAF mutations showed poor differentiation in tumor tissues. BRAF-mutant patients had a higher rate of recurrence and worse RFS compared with other patients.