Genomic Characterization of Chinese Lung Adenocarcinoma and Comprehensive Analyses to Identify Clinical Association

TY - BOOK

T1 - Genomic Characterization of Chinese Lung Adenocarcinoma and Comprehensive Analyses to Identify Clinical Association

AU - Wu, Kui

PY - 2018

Y1 - 2018

N2 - Lung cancer is the most leading cause of cancerous deaths worldwide1 with two major types:non-small-cell lung cancer (NSCLC) and small cell lung cancer (SCLC), which account forapproximately 85% and 15% of all diagnosed lung cancers, respectively2. Lung adenocarcinoma isthe most common histological type of non-small-cell lung cancer, resulting in more than 500,000deaths globally every year3, and is also the most diagnosed malignancy in China. Though cigarettesmoking is thought to be the major causative factor in lung carcinogenesis, an increasing proportionof diagnosed NSCLC is observed in never smokers, of which lung adenocarcinoma is the predominantsubtype4. Prognosis of lung adenocarcinoma remains poor and contributes to high mortality, thereasons mainly are: 1) Late diagnosis when locally advanced malignancy develops or even spreadsto regional and distant tissues2; 2) Though molecular profiling has been developed to improvepathologic diagnosis and subtype classification through several well-known biomarkers such asEGFR, KRAS, ALK , ERBB2, and BRAF, still around half of diagnosed lung adenocarcinoma harborsno known clinically relevant oncogenic drivers5,6. 3) Inter- and intratumor heterogeneities of lungadenocarcinoma contribute to difficult diagnosis and poor therapy response. 4) Poor therapeuticefficiency on metastasis due to lack of clear mechanisms governing metastasis formation andpotential drugable targets. Previous studies through genomic characterization, either by array-based copy number variationprofiling7, target region sequencing of candidate cancer related genes8 or complete whole genomesequencing3,9 have identified many potential driver genes associated with lung adenocarcinomatumorigenesis, including activated oncogenes like EGFR, KRAS, ERBB2, BRAF, EPHA3, andPIK3CA, and translocations or fusions involving ALK, ROS1, RET and KIF5B, most of which haveapproved target drugs or under clinical investigation. Loss of function alteration or deletion haveaffected several tumor suppressor genes in lung adenocarcinoma such as TP53, APC, STK11, RB1,NF1, LRP1B, CDKN2A, SMARCA4, and KEAP1, which play important role in tumorigenesis butnone of them can be targeted by specific drugs. By sequencing large cohort of patients, these studieshave also identified frequently mutated genes other than known driver genes, like U2AF1, ARID1A,RBM103 and DACH1, CFTR, RELN, ABCB5, and HGF9. Though most of the identified oncogenic“driver” events evidently contribute to cancerogenesis, these studies mainly focused on lungadenocarcinomas at early stages, and it was hard to decipher tumor metastasis responsible for poorprognosis or even death by these findings. This is in part due to difficulty in obtaining sufficientspecimens as most advanced patients are not subjected to surgery and are instead offered conservativetreatment. However, investigation on additional altered genes in advanced lung adenocarcinomas,especially those harbored in the corresponding metastases is biological worth-while with practicalclinical implications, which could not only expand the number of potential driver mutations for lungcancerogenesis, but also improve our knowledge on metastasis formation and further guide diagnosisand therapies on metastatic lung adenocarcinomas. A recent study demonstrated high concordance ofrecurrent somatic alterations between primary tumor and matched metastases from patients with non–small-cell lung cancer10. This initial survey was limited by targeted sequencing of only 189 cancerrelated genes and was short of deep exploration of molecular mechanism associated with metastasisdevelopment. As cancer is considered to be a disease with high complexity, genetic research through one ortwo technological platforms would not provide a comprehensive understanding at the molecular level.Multi-omics solutions, integrating genomic mutations, gene differential expression and fusedtranscripts as well as methylation and microRNA-mediated regulatory profiling, are needed toprovide novel and more complete insight into the etiology of cancer. One of the advances of integratedanalyses using multi-omics data relates to the possible identification of potentially novel mechanismof carcinogenesis. For instance, traditionally mutations of the DNA sequence, especially those incancer driver genes, have been heavily studied to understand the genetic basis of carcinogenesis.However, RNA editing, a post-transcriptional modification event that can flexibly and dynamicallychange RNA transcripts, could lead to effects similar to genomic mutations and potentially contributeto tumorigenesis as well. In mammals, RNA editing is mainly accomplished by the adenosinedeaminase acting on RNA (ADAR) family of enzymes that convert adenosine to inosine (A -> I)11,12.RNA editing is enriched in Alu elements13 present in many introns and untranslated regions(UTRs)14, and may modulate RNA structure, RNA splicing and transcript expression15. RNA editinghas previously been associated with a few cancer types16-20, such as hepatocellular carcinoma andprostate cancer. Pan-cancer characterization of editing events, including lung adenocarcinoma, hasalso been carried out in two recent studies21,22 which found that RNA editing levels were significantlyhigher in lung adenocarcinomas than in paired normal tissues. Han et al. also found that ADAR1expression levels were significantly higher in lung adenocarcinomas than in normal samples, whileADAR2 expression levels were in the opposite direction. Moreover, ADAR1 was reported to be anoncogene undergoing gene amplification-associated activation that affects downstream RNA editingpatterns and patient prognosis in non-small-cell lung cancer cells23. Although most RNA editing incancer is likely to be passenger events, a small portion may act as drivers and serve as potentialmarkers for personalized diagnosis and therapy. For example, elevated editing of AZIN1S367G inhepatocellular carcinoma has been reported as oncogenic activity and may thus be a potential driverin pathogenesis16. Furthermore, the profiling of the genetic landscape could also promote investigation of relateddiseases that may be associated with lung cancer. For instance, Chronic Obstruction PulmonaryDisease (COPD) and lung cancer are common lung diseases that frequently coexist. COPD ischaracterized by progressive airflow obstruction and chronic inflammation in the airways24.WHOestimates that COPD will become the third leading cause of death worldwide by 2030. A number ofepidemiological studies has demonstrated that the presence of COPD increased the risk ofdevelopment of lung cancer25-27. Cigarette smoking is considered as a common cause of COPD andlung cancer, and the smokers with airflow obstruction are up to 6-fold more likely to develop lungcancer than those with normal lung function28. Despite the fact that cigarette smoking is a principalcause of both COPD and lung cancer, several population-based studies show that COPD confers riskfor lung cancer regardless of patients’ smoking history25,29,30. The high prevalence of lung cancer inCOPD subjects suggests that there may be certain mechanisms linking COPD to lung cancer. In fact,several mechanisms including oxidative stress, genetic predisposition, epigenetic modifications andchanges in inflammatory milieu and immune defense have been proposed to link the pathogenesis ofCOPD and development of lung cancer25,31,32. Genome-wide association studies (GWAS) haveidentified several single nucleotide polymorphisms (SNPs) which predispose to increasedsusceptibility to COPD and lung cancer in the genes encoding SERPIN2, HHIP, FAM13A, IREB2,CHRNA3 and CHRNA525,33. Other studies also showed that inflammatory response factors, aberrantNF-煯B activation, cytokine release, and high levels of CD8+T cells mediate the link between COPDand lung cancer34,35. There is no doubt that understanding the mechanistic link between COPD andlung cancer would provide therapeutic and preventive benefit for patients with COPD. However, themolecular mechanisms linking COPD with lung cancer development are far from clear, and theheterogeneous nature of lung cancer and COPD makes it difficult to identify the mechanisms linkingCOPD to lung cancer. It is also unknown whether LUAD patients with COPD harbor distinct geneticcharacteristics compared to those without COPD

AB - Lung cancer is the most leading cause of cancerous deaths worldwide1 with two major types:non-small-cell lung cancer (NSCLC) and small cell lung cancer (SCLC), which account forapproximately 85% and 15% of all diagnosed lung cancers, respectively2. Lung adenocarcinoma isthe most common histological type of non-small-cell lung cancer, resulting in more than 500,000deaths globally every year3, and is also the most diagnosed malignancy in China. Though cigarettesmoking is thought to be the major causative factor in lung carcinogenesis, an increasing proportionof diagnosed NSCLC is observed in never smokers, of which lung adenocarcinoma is the predominantsubtype4. Prognosis of lung adenocarcinoma remains poor and contributes to high mortality, thereasons mainly are: 1) Late diagnosis when locally advanced malignancy develops or even spreadsto regional and distant tissues2; 2) Though molecular profiling has been developed to improvepathologic diagnosis and subtype classification through several well-known biomarkers such asEGFR, KRAS, ALK , ERBB2, and BRAF, still around half of diagnosed lung adenocarcinoma harborsno known clinically relevant oncogenic drivers5,6. 3) Inter- and intratumor heterogeneities of lungadenocarcinoma contribute to difficult diagnosis and poor therapy response. 4) Poor therapeuticefficiency on metastasis due to lack of clear mechanisms governing metastasis formation andpotential drugable targets. Previous studies through genomic characterization, either by array-based copy number variationprofiling7, target region sequencing of candidate cancer related genes8 or complete whole genomesequencing3,9 have identified many potential driver genes associated with lung adenocarcinomatumorigenesis, including activated oncogenes like EGFR, KRAS, ERBB2, BRAF, EPHA3, andPIK3CA, and translocations or fusions involving ALK, ROS1, RET and KIF5B, most of which haveapproved target drugs or under clinical investigation. Loss of function alteration or deletion haveaffected several tumor suppressor genes in lung adenocarcinoma such as TP53, APC, STK11, RB1,NF1, LRP1B, CDKN2A, SMARCA4, and KEAP1, which play important role in tumorigenesis butnone of them can be targeted by specific drugs. By sequencing large cohort of patients, these studieshave also identified frequently mutated genes other than known driver genes, like U2AF1, ARID1A,RBM103 and DACH1, CFTR, RELN, ABCB5, and HGF9. Though most of the identified oncogenic“driver” events evidently contribute to cancerogenesis, these studies mainly focused on lungadenocarcinomas at early stages, and it was hard to decipher tumor metastasis responsible for poorprognosis or even death by these findings. This is in part due to difficulty in obtaining sufficientspecimens as most advanced patients are not subjected to surgery and are instead offered conservativetreatment. However, investigation on additional altered genes in advanced lung adenocarcinomas,especially those harbored in the corresponding metastases is biological worth-while with practicalclinical implications, which could not only expand the number of potential driver mutations for lungcancerogenesis, but also improve our knowledge on metastasis formation and further guide diagnosisand therapies on metastatic lung adenocarcinomas. A recent study demonstrated high concordance ofrecurrent somatic alterations between primary tumor and matched metastases from patients with non–small-cell lung cancer10. This initial survey was limited by targeted sequencing of only 189 cancerrelated genes and was short of deep exploration of molecular mechanism associated with metastasisdevelopment. As cancer is considered to be a disease with high complexity, genetic research through one ortwo technological platforms would not provide a comprehensive understanding at the molecular level.Multi-omics solutions, integrating genomic mutations, gene differential expression and fusedtranscripts as well as methylation and microRNA-mediated regulatory profiling, are needed toprovide novel and more complete insight into the etiology of cancer. One of the advances of integratedanalyses using multi-omics data relates to the possible identification of potentially novel mechanismof carcinogenesis. For instance, traditionally mutations of the DNA sequence, especially those incancer driver genes, have been heavily studied to understand the genetic basis of carcinogenesis.However, RNA editing, a post-transcriptional modification event that can flexibly and dynamicallychange RNA transcripts, could lead to effects similar to genomic mutations and potentially contributeto tumorigenesis as well. In mammals, RNA editing is mainly accomplished by the adenosinedeaminase acting on RNA (ADAR) family of enzymes that convert adenosine to inosine (A -> I)11,12.RNA editing is enriched in Alu elements13 present in many introns and untranslated regions(UTRs)14, and may modulate RNA structure, RNA splicing and transcript expression15. RNA editinghas previously been associated with a few cancer types16-20, such as hepatocellular carcinoma andprostate cancer. Pan-cancer characterization of editing events, including lung adenocarcinoma, hasalso been carried out in two recent studies21,22 which found that RNA editing levels were significantlyhigher in lung adenocarcinomas than in paired normal tissues. Han et al. also found that ADAR1expression levels were significantly higher in lung adenocarcinomas than in normal samples, whileADAR2 expression levels were in the opposite direction. Moreover, ADAR1 was reported to be anoncogene undergoing gene amplification-associated activation that affects downstream RNA editingpatterns and patient prognosis in non-small-cell lung cancer cells23. Although most RNA editing incancer is likely to be passenger events, a small portion may act as drivers and serve as potentialmarkers for personalized diagnosis and therapy. For example, elevated editing of AZIN1S367G inhepatocellular carcinoma has been reported as oncogenic activity and may thus be a potential driverin pathogenesis16. Furthermore, the profiling of the genetic landscape could also promote investigation of relateddiseases that may be associated with lung cancer. For instance, Chronic Obstruction PulmonaryDisease (COPD) and lung cancer are common lung diseases that frequently coexist. COPD ischaracterized by progressive airflow obstruction and chronic inflammation in the airways24.WHOestimates that COPD will become the third leading cause of death worldwide by 2030. A number ofepidemiological studies has demonstrated that the presence of COPD increased the risk ofdevelopment of lung cancer25-27. Cigarette smoking is considered as a common cause of COPD andlung cancer, and the smokers with airflow obstruction are up to 6-fold more likely to develop lungcancer than those with normal lung function28. Despite the fact that cigarette smoking is a principalcause of both COPD and lung cancer, several population-based studies show that COPD confers riskfor lung cancer regardless of patients’ smoking history25,29,30. The high prevalence of lung cancer inCOPD subjects suggests that there may be certain mechanisms linking COPD to lung cancer. In fact,several mechanisms including oxidative stress, genetic predisposition, epigenetic modifications andchanges in inflammatory milieu and immune defense have been proposed to link the pathogenesis ofCOPD and development of lung cancer25,31,32. Genome-wide association studies (GWAS) haveidentified several single nucleotide polymorphisms (SNPs) which predispose to increasedsusceptibility to COPD and lung cancer in the genes encoding SERPIN2, HHIP, FAM13A, IREB2,CHRNA3 and CHRNA525,33. Other studies also showed that inflammatory response factors, aberrantNF-煯B activation, cytokine release, and high levels of CD8+T cells mediate the link between COPDand lung cancer34,35. There is no doubt that understanding the mechanistic link between COPD andlung cancer would provide therapeutic and preventive benefit for patients with COPD. However, themolecular mechanisms linking COPD with lung cancer development are far from clear, and theheterogeneous nature of lung cancer and COPD makes it difficult to identify the mechanisms linkingCOPD to lung cancer. It is also unknown whether LUAD patients with COPD harbor distinct geneticcharacteristics compared to those without COPD

UR - https://rex.kb.dk/permalink/f/h35n6k/KGL01012061463

M3 - Ph.D. thesis

BT - Genomic Characterization of Chinese Lung Adenocarcinoma and Comprehensive Analyses to Identify Clinical Association

PB - Department of Biology, Faculty of Science, University of Copenhagen

ER -