By continuing to browse this site you agree to us using cookies as described in About Cookies
Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishing platforms. The new Wiley Online Library will be migrated over the weekend of February 24 & 25 and will be live on February 26, 2018. For more information, please visit our migration page: http://www.wileyactual.com/WOLMigration/
Members of the University of Hong Kong Lung Cancer Study Group include Lik-Cheung Cheng, Daniel Tsin-Tien Chua, Lap-Ping Chung, James Chung-Man Ho, Elaine Lai-Han Leung, Alan Dart-Loon Sihoe, Vicky Pui-Chi Tin, and Maria Pik Wong.
The echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase (EML4-ALK) fusion gene resulting from the chromosome inversion inv(2)(p21;p23) recently was identified in nonsmall cell lung cancer (NSCLC). The authors of this study investigated the frequency, genetic and clinicopathologic profiles of EML4-ALK in Chinese patients with NSCLC.
EML4-ALK was investigated in 266 resected primary NSCLC, including adenocarcinomas (AD), lymphoepithelioma-like carcinomas, squamous cell carcinomas, mucoepidermoid carcinomas, and adenosquamous carcinomas, by reverse transcriptase-polymerase chain reaction and was verified by sequencing. EML4-ALK protein expression was studied by immunohistochemistry.
Thirteen tumors (4.9%) had EML4-ALK comprising 4 fusion transcript variants with fusion of the variable segments from 5′ EML4 to 3′ ALK and with preservation of the ALK kinase domain. The most common variant consisted of 8 tumors with variant 3 that involved EML4 exon 6. The others included 2 tumors with variant 1 (exon 13), 2 tumors with variant 2 (exon 20), and 1 tumor with the novel variant 5 (exon 18). There were 11 ADs and 2 unusual carcinomas with mixed squamous and glandular components. Immunohistochemistry demonstrated diffuse ALK fusion proteins in the tumor cell cytoplasm. EML4-ALK was associated with nonsmokers (P = .009). Tumors with the fusion gene had the wild-type epidermal growth factor receptor (EGFR) (P = .001) and v-Ki-ras2/Kirsten rat sarcoma viral oncogene homolog (KRAS) genes. Patients who had EML4-ALK-positive AD had a younger median age (P = .018) compared with patients who did not have the fusion gene.
Lung cancer is the leading cause of cancer deaths in the world.1 The main tumor types include adenocarcinoma (AD), small cell carcinoma, squamous cell carcinoma (SCC), and large cell carcinoma. AD is the most common tumor type among both nonsmokers and smokers. Mutational activation of the epidermal growth factor receptor (EGFR) pathway is the major carcinogenic mechanism in nontobacco-induced lung cancers, whereas different pathways, such as mutations in the v-Ki-ras2/Kirsten rat sarcoma viral oncogene homolog (KRAS), are involved in tobacco-mediated lung carcinogenesis. Several uncommon cancer types may develop in the lung, including Epstein-Barr virus (EBV)-induced lymphoepithelioma-like carcinomas (LELCs), but the molecular pathways of other tumors, such as mucoepidermoid carcinomas (MECs) and adenosquamous carcinomas (ADSQs), are not clear.2
Chromosomal translocation has long been linked to cancer causation3 and is related to approximately 20% of all cancers.4 Fusion genes are common in lymphoid, hematopoietic, and connective tissue cancers; however, apart from prostatic carcinoma, it is believed that fusion genes are rare in solid tumors.5, 6EML4-ALK, which consists of a fusion of the echinoderm microtubule-associated protein-like 4 (EML4) and anaplastic lymphoma kinase (ALK) genes resulting from the chromosome inversion inv(2)(p21;p23), recently has been identified in Japan in the AD from a man who was a smoker by screening for transforming activities in its complementary DNA expression library.7 Several studies have reported EML4-ALK in a low percentage of lung cancers, but the detailed clinicopathologic profiles is unclear because of the small number of cases identified. To investigate the relation of the EML4-ALK fusion gene and lung cancer profiles (especially smoking history) as well as EGFR and KRAS mutations, we studied 266 nonsmall cell lung cancers (NSCLCs) of common and uncommon types.
MATERIALS AND METHODS
Primary lung cancers from Chinese patients who did not receive neoadjuvant therapy and who underwent resection at Grantham Hospital or Queen Elizabeth Hospital, Hong Kong, and the corresponding non-neoplastic lung tissues were collected and snap-frozen in liquid nitrogen. Informed consent and ethics approval were obtained for research purposes. The samples included 209 ADs, 34 SCCs, 11 LELCs, and 12 other histologic types, including MEC and ADSQ. Tumor tissues were examined histologically before nucleic acid extraction to ensure that tissues with minimal necrosis and at least 75% tumor by area were used. Nonsmokers were patients who had smoked <100 cigarettes in their lifetime or who were exposed to environmental tobacco at home or in the workplace. Ever-smokers were current smokers or individuals who had stopped smoking for ≤1 year before surgery. The details of sample collection, tumor typing, and grading have been described previously.8 Twenty-four lung cancer cell lines that were purchased from the American Type Culture Collection (Manassas, Va) or kindly provided by Dr. J. Minna (Hamon Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas, Tex), including 18 NSCLC cell lines (A549, H23, H358, H441, H1299, H1437, H1648, H1650, H1819, H1975, H1993, H2122, H2347, HCC78, HCC366, HCC827, HCC1195, and HCC1833), 1 small cell cancer line (H526), and 5 NSCLC cell lines raised from local patients (HKULC1,9 HKULC2,9 HKULC3,9 HKULC4,9 and FA31), were analyzed.
RNA Extraction and Complementary DNA Synthesis
Total RNA was extracted from frozen cancer and normal lung tissues using Trizol (Invitrogen, Carlsbad, Calif) according to the manufacturer's instructions. RNA quality and the absence of genomic DNA contamination were verified by gel electrophoresis. Synthesis of complementary DNA was performed with oligo-dT primed reverse transcription from 5 μg total RNA using the SuperScript First-Strand Synthesis System (Invitrogen) according to the manufacturer's instructions.
Detection of EML4-ALK Fusion Transcripts by Reverse Transcriptase-polymerase Chain Reaction and Direct Sequencing
The primers that we used to identify the EML4-ALK fusion transcript were chosen to enable the detection of all possible in-frame fusions of EML4 to exon 20 of ALK in which the kinase domain of ALK would be preserved. The primers Fusion-RT-S and Fusion-RT-AS7 were used to detect fusions that involved EML4 exons 13, 18, and 20, as reported previously. The primers E6A20-S (5′-TTCGAGCATCACCTTCTCC-3′) and E6A20-AS (5′-GGACACCTGGCCTT CATAC-3′) were used to detect fusion involving EML4 exon 6. Polymerase chain reaction (PCR) was performed with initial denaturation at 95°C for 4 minutes, followed by 40 cycles of amplification (at 95°C for 30 seconds, 55°C for 30 seconds, and 72°C for 2 minutes 30 seconds), and final extension at 72°C for 10 minutes. Reverse transcriptase (RT)-PCR for β-actin was used as control for RNA quality. Resulting PCR products were verified by sequencing using a 3130XL Genetic Analyzer (Applied Biosystems, Foster City, Calif) with the BigDye Terminator version 3.1 cycle sequencing kit (Applied Biosystems) according to the manufacturer's instructions.
Identification of EML4-ALK Genomic Breakpoints
Genomic DNA was extracted from frozen tissues by using standard phenol-chloroform methods. The genomic breakpoints of EML4-ALK variants 1 and 2 were detected as reported previously.7 The genomic breakpoint of EML4-ALK variant 3 was identified by PCR with the primers Fusion3-G-E6 (5′-GCATAAAGATGTCATCATCAACCAAG-3′) and Fusion-genomic-AS,7 and the breakpoint of variant 5 was identified by using the PCR primers Fusion5-G-E18 (5′-CTGGATGCAGAAACCAGAGATCTAGT-3′) and Fusion-genomic-AS,7 followed by direct sequencing of the amplification products.
Detection of EML4 and ALK Messenger RNA Expression
EML4 messenger RNA (mRNA) was detected by RT-PCR using the primers EML4-F (5′-CAGCCAT-GTCACCAATGTC-3′) and EML4-R (5′-CACTTGG-CTCCACAGTTTGT-3′), which spanned the 3′ end of EML4. PCR was performed at 95°C for 4 minutes, 40 cycles of amplification (at 95°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds), and 10 minutes at 72°C for final extension. ALK mRNA expression was studied using the primers ALK-V-F and ALK-V-R, which flanked the extracellular region, as reported previously.10 Fifty samples each of normal lung and lung cancers, including 23 ADs, 9 LELCs, 8 SCCs, 8 ADSQs, and 2 MECs, were analyzed along with the 24 cancer cell lines.
Detection of EGFR and KRAS Mutations
Mutations of EGFR (exons 18-21) and KRAS (codons 12, 13, and 61) were detected as described previously. The results from 209 cases had been reported previously.8
Paraffin sections (5 μm thick) were subjected to antigen retrieval using microwave heating at 95°C for 30 minutes in 10 mM citrate buffer, pH 6.0. Endogenous peroxidase was quenched with 3% hydrogen peroxide for 10 minutes. Nonspecific binding was blocked with a biotin blocking system (DAKO, Glostrup, Denmark) and 10% normal goat serum for 10 minutes. Sections were incubated at 4°C overnight with 1:1000 polyclonal anti-ALK antibody (Invitrogen), which could detect both the full-length and fusion ALK products. The secondary antibody consisted of biotinylated goat-antirabbit immunoglobulin (DAKO). Color signals were developed using the streptavidin-biotinylated horseradish peroxidase complex (DAKO), 3,3′-diaminobenzidine, and hydrogen peroxide system. The primary antibody was omitted in the control reaction.
The chi-square test or Fisher exact tests (SPSS version 15.0; SPSS Inc., Chicago, Ill) were used to compare EML4-ALK fusion gene status with clinicopathologic characteristics, including sex, smoking status, tumor type, differentiation, stage, and EGFR and KRAS mutation status. The Mann-Whitney U test was used to compare tumor size and patient age with fusion gene status. The 2-sided statistical significance was defined as P < .05.
EML4-ALK Fusion Transcript Was Detected in 4.9% of Nonsmall Cell Lung Cancers
EML4-ALK transcripts were detected in 13 of 266 tumors (4.9%) that comprised 4 variants of EML4 fusion with exon 20 of ALK (Fig. 1). Eight tumors involved EML4 exon 6, which included 2 splice forms that differed by 33 nucleotides from intron 6 of EML4 (variants 3a and 3b).11 There were 2 cases each of variants 1 and 2 with fusion points at EML4 exons 13 and 20, respectively.7 One tumor yielded a novel fusion transcript that was sequenced, characterized as fusion of EML4 exon 18, and designated variant 5. No fusion product involving exon 15 of EML4 (variant 4) was detected.12 None of the 24 cancer cell lines that we examined harbored the fusion transcript.
Variation in Genomic Breakpoints of EML4-ALK
The genomic structure of EML4-ALK fusion was identified in 1 case each of variants 1, 2, 3 and 5. The breakpoints were located at 447 base pairs (bp) of EML4 intron 13 and 770 bp of ALK intron 19 (variant 1), at 182 bp of EML4 intron 20 and 1865 bp of ALK intron 19 (variant 2), at 805 bp of EML4 intron 6 and 1817 bp of ALK intron 19 (variant 3), and at 654 bp of EML4 intron 18 and 1760 bp of ALK intron 19 (variant 5) (Fig. 2). There was gain of 2 adenines and 1 thymidine at the fusion points of variants 3 and 5, respectively.
Clinicopathologic Characteristics of Patients With the EML4-ALK Fusion Gene
Table 1 summarizes the clinical and pathologic profile of all patients, and Table 2 lists the patients who had the fusion gene identified, which included 11 of 209 ADs (5.3%) from 1 chronic smoker, 2 passive smokers, and 8 never-smokers. Histologically, 9 of 11 ADs had the mixed tumor subtype, which displayed papillary, micropapillary, and cribriform growth (Fig. 3A). Two tumors (Patients 124 and 206) had a bronchioloalveolar carcinoma (BAC) pattern with lepidic tumor spread at the periphery (Fig. 3B). No significant association was observed between the fusion variants and the AD tumor subtypes. Two fusion gene tumors from nonsmokers had unusual histologic features that included both squamous and glandular components. One tumor (Patient 224) had prominent endobronchial growth and consisted of low-grade tumor cells arranged in papillary squamous areas interspersed with mucin-containing glandular spaces or individual mucinous cells (Fig. 3C). The diagnosis rendered by 4 experienced anatomic pathologists who specialized in lung tumor diagnosis agreed to designate this tumor as a low-grade mucoepidermoid carcinoma in the broad sense. The other tumor (Patient 249) consisted of distinct and well defined squamous and glandular components merged with poorly differentiated, indeterminate areas with extensive necrosis and moderate-to-marked nuclear pleomorphism. A consensus diagnosis could not be reached in this case, and the possibilities of ADSQ or mucoepidermoid carcinoma were considered (Fig. 3D). None of the 34 SCCs, 11 LELCs, and or 8 ADSQs contained the fusion gene. Overall, the fusion gene was identified in 12 of 141 nonsmokers (8.5%) and in 1 of 125 ever-smokers (0.8%) with a statistically significant association observed for nonsmokers (P = .009; chi-square test). However, when only ADs were considered, this association became insignificant (P = .053; Fisher exact test). There was a trend toward younger age among patients who had the fusion gene patients when all patients were included in the analysis (P = .065; Mann-Whitney U test). The difference was significant when only ADs were considered (median age, 59 years [interquartile range, 51-61 years] vs 64 years [interquartile range, 55-71 years]; P = .018; Mann-Whitney U test). No significant association was observed with patient sex, histologic tumor type, tumor differentiation, tumor size, or pathologic stage.
Table 1. Relations Between EML4-ALK Gene Fusion and Clinicopathologic Profile in Patients with Nonsmall Cell Lung Carcinoma
Detection of EML4 and ALK Messenger RNA Expression
EML4 mRNA expression was detected readily in all samples of normal lung and lung cancers, including the fusion gene cases (Fig. 4A,B). mRNA expression of the ALK extracellular region was detected in 6 lung cancer cell lines and in 7 tumors without EML4-ALK (Fig. 4C), but not in normal lung tissues or the in tumors that harbored the fusion gene.
ALK Protein Was Expressed in Tumors With EML4-ALK
Paraffin sections were available for 12 of the fusion gene cancers. All showed fine granular staining of various intensities in the tumor cell cytoplasm without membrane or nuclear accentuation (Fig. 3) (Table 2). Because full-length ALK mRNA was not expressed in these samples, the staining implicated expression of the fusion proteins. No expression was observed in normal lung bronchial epithelium, alveolar pneumocytes, alveolar macrophages, vascular tissues, or other connective tissues.
Detection of EGFR and KRAS Mutations
There were 125 patients with EGFR mutations and 22 patients with KRAS mutations, as summarized in Table 1. All 13 patients who had the EML4-ALK fusion gene had wild-type sequences for exons 18 through 21 of EGFR and for codons 12, 13, and 60 of KRAS. Statistical analysis demonstrated a significant association of the EML4-ALK fusion gene with wild-type EGFR (P = .001; chi-square test) but not KRAS (P = .610; Fisher exact test). Cases with EML4-ALK, EGFR, or KRAS mutations were mutually exclusive. The relation between the genomic status of the 3 genes, sex, and smoking history in patients with AD is shown in Table 3.
Table 3. EML4-ALK, EGFR, and KRAS Mutations in Patients with Adenocarcinoma
Fusion genes were not described previously in major primary lung cancers until the EML4-ALK fusion gene was identified. Currently, limited numbers of lung cancers that harbor EML4-ALK have been reported. Because lung cancers that develop in different backgrounds of tobacco exposure and/or ethnicity may involve distinct carcinogenetic pathways with different implications on treatment potentials, we have investigated the frequencies and clinicopathologic profiles of all potential oncogenic variants of EML4-ALK in a cohort of 266 primary NSCLCs from ethnic Chinese patients. The EML4-ALK fusion gene was identified in 13 tumors (4.9%). This proportion was comparable to that reported in another series that investigated 103 Chinese patients (3 of 103 tumors had the EML4-ALK fusion gene; 2.9%).13 Lung cancers from Japan also had similar proportions of EML4-ALK, as illustrated in the respective studies by Fukuyoshi et al. (1 of 104 tumors; 0.1%),14 Inamura et al. (5 of 221 tumors; 2.3%),15 Shinmura et al. (2 of 77 tumors; 2.6%),16 and Soda et al. (5 of 75 tumors; 6.7%).7 Recently, EML4 and/or ALK rearrangements were demonstrated by break-apart fluorescent in situ hybridization assay in 16 of 603 tumors (2.7%) from Caucasian patients.17 In another study, EML4-ALK also was identified in lung cancers from Caucasian patients (2 of 138 tumors; 1.5%) and Korean patients (6 of 167 tumors; 3.8%).12
Variants 1 and 2 have been the main EML4-ALK variants reported7, 14-17; however, in our population, variant 3 was the most common (8 of 13 tumors; 61.5%). Although variant 4 was not detected in our samples, another novel variant exhibited the fusion of EML4 exon 18 to ALK exon 20 and was designated variant 5. Thus, although the genomic breakpoints were variable, as demonstrated in our patients compared with those reported in the literature,7, 11, 14, 16 the basic structures of the EML4-ALK fusion gene in all described cases were similar, with in-frame fusion of 5′ fragments of EML4 to ALK exon 20 resulting in preservation of the coiled-coil domain of EML4 and the juxtamembrane intracellular region, including the tyrosine kinase domain, of ALK. Soda et al. demonstrated that the EML4 basic region that contains the coiled-coil domain is essential for the transforming ability of the fusion gene, most likely by mediating dimerization and ALK activation.7 Thus, the inclusion of the coiled-coil domain in all variants, including the novel variant 5, suggests that they are competent in mediating transformation.
Our study has revealed interesting data on the clinicopathologic profiles of cancers with the EML4-ALK fusion gene. The fusion gene most frequently involved AD; however, unlike Inamura et al.,15 we did not observe any association with AD subtypes. This could be related to interobserver variability in histologic subtyping of AD and the relatively small number of cases in the comparison. We have observed that most of the fusion gene ADs were the mixed tumor subtype, which also is the most common pattern of all lung ADs that have undergone thoroughly histologic examination. It is noteworthy that only 2 ADs had the lepidic growth pattern characteristic of BAC. Because it is believed that BAC indicates differentiation toward the terminal bronchiolar/alveolar epithelium and often is associated with EGFR mutations, the predominance of other AD subtypes suggests that EML4-ALK may play a role in the carcinogenesis of more proximal airways. It also is worth noting that 2 SCCs, which characteristically develop in proximal airways, also contained EML4-ALK,7, 13 and we also identified the fusion gene in 2 unusual tumors that displayed distinct squamous and glandular components, including 1 MEC. MECs usually are discovered in salivary glands18, 19 but they also develop in the lungs from bronchial glands of proximal airways and rarely in other unusual sites, such as the thyroid gland.20 The chromosomal translocation t(11;19) that results in the oncogenic CREB-regulated transcription coactivator 1-mastermind-like 2 fusion gene CRTC1-MAML2 has been identified in MECs of the salivary glands and lungs.20-22 Overall, our findings and the reported data together suggest that EML4-ALK may be involved in NSCLC subtypes derived from large or proximal airway cells.
Most of the patients who had EML4-ALK were primarily nonsmokers (10 never smokers; 2 passive smokers), and the association was statistically significant when all patients were included in the analysis. Data from other series also indicated more frequent involvement in nonsmokers.12, 13, 15 Because AD is the most common tumor type in both smokers and nonsmokers, we analyzed the AD subgroup, but the association became nonsignificant (P = .053).
In this study, a younger median age was observed among patients who had tumors with the EML4-ALK fusion gene. In particular, 1 patient aged 33 years with BAC-type carcinoma had no known risk factors, such as tobacco exposure or a family history of lung cancer. Inamura et al. also reported a similar trend toward younger age among patients who had tumors with EML4-ALK.15 It is noteworthy that patients with anaplastic large cell lymphoma and inflammatory myofibroblastic tumors that harbor ALK fusion genes also tend to be younger.23-25 Data from a larger number of patients will be needed to determine whether chromosomal translocations that involve ALK are more common in younger patients.
In mice, the expression of both EML4 and ALK is regulated developmentally and is restricted to the nervous system after birth.26, 27 Transfection of EML4 into mammalian cells has demonstrated that expression of the protein is colocalized with cytoplasmic microtubules in interphase cells and, specifically, with the mitotic spindle in cycling cells.26, 28 Our data indicate that full-length ALK mRNA was not expressed in normal human lung, and EML4 transcripts were detectable readily in lung cancers and in cancer cell lines. Unexpectedly, EML4 transcripts also were detected in normal human lung tissues. The findings indicate that constitutive EML4 promoter activity and protein dimerization may mediate activation of the ALK tyrosine kinase domain in the fusion gene, in turn leading to anomalous transmission of downstream oncogenic cellular signals.29 Consistent with this hypothesis, immunohistochemistry indicated that there was diffuse expression of the ALK fusion protein in the EML4-ALK tumors, but ALK proteins were not detected in normal lung tissues.
All of the EML4-ALK-positive tumors in our study were positive for wild-type EGFR and KRAS. In EGFR and KRAS wild-type tumors, 13 of 119 tumors (11%) tumors or 11 of 66 ADs (17%) harbored EML4-ALK. Inamura et al. also observed the mutually exclusive involvement of EGFR, KRAS, and EML4-ALK.15 Conversely, Koivunen et al. reported the coexistence of EGFR mutation and EML4-ALK in a patient with AD.12 Thus, further investigation is needed to determine the relative oncogenic role of the 2 genomic changes.
EML4-ALK appears to be a new oncogene involved in NSCLC, particularly in nonsmokers. Twenty-nine percent of lung ADs from nonsmoking women in our population that had the wild types of EGFR and KRAS involved EML4-ALK. Although EGFR mutations are confined almost exclusively to lung AD, EML4-ALK also is involved in other lung cancer types, including those that arise in the more proximal airways. These patients potentially could benefit from ALK-inhibitor targeted therapy.11, 12, 30
We thank Dr. John Chan (Department of Pathology, Queen Elizabeth Hospital), Dr. Elaine Wang (Department of Pathology, Grantham Hospital), and Dr. K. H. Fu (Tseung Kwan O Hospital) for review of the tumor histology and helpful discussions. We also gratefully acknowledge all surgeons of the Cardiothoracic Units at Grantham Hospital and Queen Elizabeth Hospital for providing cases for this study. This work represents the joint effort of clinicians and scientists of Grantham Hospital, Queen Elizabeth Hospital, and Queen Mary Hospital, Hong Kong.
Conflict of Interest Disclosures
Supported by a grant from the General Research Fund (HKU/778708) awarded by the Research Grants Council (Hong Kong Government) and by a grant from the Small Project Fund (10207989) awarded by the Committee on Research and Conference Grants, University of Hong Kong.