Frequent Alterations in Cytoskeleton Remodelling Genes in Primary and Metastatic Lung Adenocarcinomas
Total Page:16
File Type:pdf, Size:1020Kb
ARTICLE Received 25 Jun 2015 | Accepted 6 Nov 2015 | Published 9 Dec 2015 DOI: 10.1038/ncomms10131 OPEN Frequent alterations in cytoskeleton remodelling genes in primary and metastatic lung adenocarcinomas Kui Wu1,2,3,4,5,*, Xin Zhang2,3,4,*, Fuqiang Li1,*, Dakai Xiao2,3,6,*, Yong Hou1,5,*, Shida Zhu1,5,*, Dongbing Liu1, Xiaofei Ye1,7, Mingzhi Ye1,8, Jie Yang1, Libin Shao1, Hui Pan2,3,6,NaLu1, Yuan Yu1, Liping Liu2,3,6, Jin Li2,3,6, Liyan Huang2,3, Hailing Tang2,3, Qiuhua Deng2,3,6, Yue Zheng1, Lihua Peng1, Geng Liu1, Xia Gu9, Ping He9, Yingying Gu3,9, Weixuan Lin6, Huiming He6, Guoyun Xie1, Han Liang1,NaAn1, Hui Wang1, Manuel Teixeira10, Joana Vieira10, Wenhua Liang2,3,4, Xin Zhao1, Zhiyu Peng1,8, Feng Mu1,11, Xiuqing Zhang1,8, Xun Xu1, Huanming Yang1,12, Karsten Kristiansen1,2, Jian Wang1,12, Nanshan Zhong3,4, Jun Wang1,5,**, Qiang Pan-Hammarstro¨m1,7,** & Jianxing He2,3,4,** The landscape of genetic alterations in lung adenocarcinoma derived from Asian patients is largely uncharacterized. Here we present an integrated genomic and transcriptomic analysis of 335 primary lung adenocarcinomas and 35 corresponding lymph node metastases from Chinese patients. Altogether 13 significantly mutated genes are identified, including the most commonly mutated gene TP53 and novel mutation targets such as RHPN2, GLI3 and MRC2. TP53 mutations are furthermore significantly enriched in tumours from patients harbouring metastases. Genes regulating cytoskeleton remodelling processes are also frequently altered, especially in metastatic samples, of which the high expression level of IQGAP3 is identified as a marker for poor prognosis. Our study represents the first large-scale sequencing effort on lung adenocarcinoma in Asian patients and provides a comprehensive mutational landscape for both primary and metastatic tumours. This may thus form a basis for personalized medical care and shed light on the molecular pathogenesis of metastatic lung adenocarcinoma. 1 BGI-Shenzhen, Beishan Industrial Zone, Yantian District, Shenzhen 518083, China. 2 Department of Thoracic Surgery, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou 510120, China. 3 Guangzhou Institute of Respiratory Disease & State Key Laboratory of Respiratory Disease, Guangzhou 510120, China. 4 National Clinical Research Center for Respiratory Disease, Guangzhou 510120, China. 5 Department of Biology, University of Copenhagen, DK-2200 Copenhagen N, Denmark. 6 Research Center for Translational Medicine, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou 510120, China. 7 Department of Laboratory of Medicine, Karolinska Institutet, Stockholm 14186, Sweden. 8 Guangzhou Key Laboratory of Cancer Trans-Omics Research, BGI-Guangzhou, Guangzhou 510006, China. 9 Department of Pathology, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou 510120, China. 10 Genetics Department and Research Center, Portuguese Oncology Institute, Porto 4200-072, Portugal. 11 BGI-Wuhan, Wuhan 430075, China. 12 James D. Watson Institute of Genome Sciences, Zhejiang University, Hangzhou 310058, China. * These authors contributed equally to this work. ** These authors jointly supervised this work. Correspondence and requests for materials should be addressed to J.H. (email: [email protected]) or to Q.P.-H. (email: [email protected]) or to J.W. (email: [email protected]). NATURE COMMUNICATIONS | 6:10131 | DOI: 10.1038/ncomms10131 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10131 ung cancer is the leading cause of cancerous deaths 1,2 Table 1 | Clinical feature summary of 335 sequenced lung worldwide with two major types: non-small-cell lung adenocarcinomas. Lcancer (NSCLC) and small cell lung cancer (SCLC), accounting for B85% and 15% of all diagnosed lung cancers, 3 Discovery Validation Total P value respectively . Lung adenocarcinoma is the most common cohort cohort histological type of NSCLC, resulting in 4500,000 deaths w 4 Age at surgery, 0.4918 globally every year . Despite advances in surgery, molecular years subtyping and targeted therapy, prognosis of lung Median 59.2 58.6 58.8 adenocarcinoma remains poor and the reasons for this could be Range 25.2–81.6 32.4–84.9 25.2–84.9 due to: (1) Diagnosis was often made already at a late stage when Gender 0.1303z localized malignant tumours spread to regional and distant Male 62 (61.4%) 121 (51.7%) 183 (54.6%) tissues3; (2) Lack of known targetable driver genes in Female 39 (38.6%) 113 (48.3%) 152 (45.4%) approximately half of the diagnosed patients5,6; (3) Complexity Smoking status 0.3373z of inter- and intra-tumour heterogeneity7,8; and (4) Poor Smoker 37 (36.6%) 68 (29.1%) 105 (31.3%) understanding on the mechanism of metastasis development, as Non-smoker 58 (57.4%) 141 (60.3%) 199 (59.4%) NA 6 (5.9%) 25 (10.7%) 31 (9.3%) well as lack of corresponding treatment. Follow-up, 0.1042w Previous studies have characterized the genomic landscape of months lung adenocarcinomas and identified many potential cancer Median 22 37 36 4,9–11 driver genes , of which targeting therapies have been Range 4–80 1–77 1–80 developed for several activated oncogenes such as EGFR, ERBB2 Tumour stage 0.1447z and BRAF6,12–14 and translocations or fusions involving ALK, I 19 (18.8%) 63 (26.9%) 82 (24.5%) ROS1 and RET15–17. Most of these studies, however, mainly II 18 (17.8%) 51 (21.8%) 69 (20.6%) focused on tumour samples obtained from European or North III 56 (55.4%) 98 (41.9%) 154 (46%) American patients, and the majority of specimens were collected IV 8 (7.9%) 21 (9%) 29 (8.7%) at early disease stages. Undoubtedly, it is important to have a NA 0 (0%) 1 (0.4%) 1 (0.3%) Metastasis 0.1087z comprehensive genomic analysis on lung adenocarcinomas from Negative 25 (24.8%) 80 (34.2%) 105 (31.3%) Eastern Asian population when considering its rapidly increasing Positive 76 (75.2%) 153 (65.4%) 229 (68.4%) incidence rate and potential genetic heterogeneity between NA 0 (0%) 1 (0.4%) 1 (0.3%) different ethnic populations. Moreover, genetic characterization Survival status 0.3612z of advanced lung adenocarcinomas especially those harbouring Alive 67 (66.3%) 160 (68.4%) 227 (67.8%) corresponding metastases will not only expand the spectrum of Dead 33 (32.7%) 60 (25.6%) 93 (27.8%) potential cancer driver genes involved in lung carcinogenesis, NA 1 (1%) 14 (6%) 15 (4.5%) but also improve our knowledge on metastasis formation NA, not applicable. and further guide diagnosis and therapies for metastatic wWilcoxon rank sum test. 2 lung adenocarcinomas. A recent study demonstrated high zPearson’s w -test. concordance of recurrent somatic alterations between primary tumours and matched metastases in NSCLCs18. This initial survey was limited though by focusing on targeted sequencing of of 24 Chinese lung adenocarcinoma patients. Genomes were 189 cancer-related genes. sequenced to a mean depth of 49.6 Â (range: 42.0–57.8 Â ) for Here we performed a comprehensive genetic analysis of 101 primary tumours and 51.2 Â (44.4–66.3 Â ) for metastatic Chinese lung adenocarcinomas, as well as 35 corresponding specimens, while it was 31.9 Â (range: 23.6–34.9 Â ) for adjacent lymph node metastases through multiplatform sequencing. Two normal tissues. On average 93M clean reads (73–110M) were hundred and thirty-four primary tumours were furthermore generated from whole-transcriptome sequencing. Whole-exome included into this study as an independent validation cohort. In sequencing (WES) was performed on both primary and matched an addition to a number of previously reported lung cancer driver lymph node metastases from an additional 11 cases, and from genes, we have identified several novel, potentially oncogenic primary tumour samples from a further group of 66 patients. genes that are significantly mutated in our cohort. Integrative RNA-seq analysis was furthermore performed in 32 cases from analyses through genomic data furthermore highlight pathways the latter group. Exome sequencing reached median fold coverage that may play a critical role in driving tumour metastasis. These of 81.4 Â (37.6–293.7 Â ) and 37.3M reads (20.0–53.4M) were results provide new insights on the pathogenesis of lung generated from RNA-seq (Supplementary Data 2 and 3). These adenocarcinomas and also form a basis for further improve- 101 cases constituted the discovery cohort. ment of clinical management of our patients in the precision For verification, an independent hybrid-recapture and medicine era. ultra-deep DNA sequencing were performed on a set of 51 selected genes with a mean fold coverage of 140 Â (48.2–455.1 Â ) on custom target regions in 98 cases of the Results discovery cohort described above and additional 234 primary Sample description and sequencing statistics. Paired tumours lung adenocarcinomas (the validation cohort) (Supplementary and normal adjacent tissues were obtained from 335 patients that Data 4). The cohort description with different data sets was provided written informed consent to carry out genomic studies summarized in Supplementary Data 5. in accordance with local Institutional Review Boards. All tumour specimens were reviewed by independent pathologists to determine the histological subtype, TNM stage and tumour Somatic DNA alterations and verification. Somatic point cellularity (Supplementary Fig. 1). Detailed clinical features were variations and small insertions/deletions (indels; o50 bp) were summarized in Table 1 and Supplementary Data 1. detected using MuTect19 and Platypus20, respectively. A mean of Whole-genome, transcriptome sequencing data were obtained 9.7 somatic mutations per Mb (1.7–64.4) were identified from the from primary tissues and corresponding lymph node metastases first 101 primary lung adenocarcinomas, while the mean somatic 2 NATURE COMMUNICATIONS | 6:10131 | DOI: 10.1038/ncomms10131 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10131 ARTICLE mutation rate in 35 lymph node metastases was 6.4 mutations per 6/24).