Font Size: a A A

Genomic Characterization Of Colorectal Carcinogenesis And Metastasis In Chinese Population

Posted on:2020-01-03Degree:DoctorType:Dissertation
Country:ChinaCandidate:F LiaoFull Text:PDF
GTID:1524306551991769Subject:Cell biology
Abstract/Summary:PDF Full Text Request
Background and purpose:Colorectal cancer(CRC)is one of the most malignant tumors in digestive system,with higher mobidity and mortality.most patients have a late onset age and only 10%of the CRC patients diagnosed younger than 50 years old,while 75%of them were diagnosed between 40 and 49 years old.From the clinical prognosis,tumor metastasis is the leading cause of death.The most common metastatic organs of CRC are liver and lung.Classical metastatic model basing on anatomical analysis suggest that tumor cells can transfer to the distant organs through lymphatic vessels and blood vessels.In the blood circulation,tumor cells fall off the primary tumors and entered the blood vessels,transferred and colonized in the liver through the portal system firstly,and then transferred to the lung,peritoneum,or other organs through the systemic circulation,which is a sequential progression pattern from the liver to the lung,but there is still no molecular biology evidence to support such transfer model.With the development of sequencing and bioinformatics,some germline mutations associated with familial young CRCs has been identified(such as Lynch syndrome raised by germline mutation in MLH1),studies on tumor heterogeneity and phylogeny from the genomic levels have also demonstrated the independent origin of CRC distant metastases and lymph node metastases and also demonstrated the polyclonal seeding of lymph node metastases.But,it is nor clear whether the molecular variations driving CRC in tumor cells will vary according to the age of onset,and the small tumor lesion types and specimens also restrict the deeply explore of tumor metastasis as well phylogeny patterns.So,we urgently need to enroll large-scale young patients as well patients with diverse lesion types,and investigated their tumor development and metastasis characteristics from molecular level.As to find out the key gene mutations that play a driving role in tumorigenesis and development,recall the evolution history of CRC evolution patterns and to provide potential biomarkers for clinical individualized prevention and treatment of CRC patients.Method:the whole study containing two part,part 1 focus on young CRC pathiens and conducted whole-exome sequencing(WES)as well RNA sequencing(RNA-seq).part 2 focus on metastatic patients and also conducted WES.We utilized bioinformatics as well biostatistics to systmaticly analysis the inherit,tumor genomicts and metastatic characteristics respectively.What’s more,"REVOLVER"machine learning method was used to track the genomic evolution trajectory of CRC metastasis.Finally,the newly identified candidate CRC metastatic driver gene were screened by Sanger sequencing in independent CRC patient cohortsResults:young CRC patients genomicts study enrolled 111 pathient younger than 40,including 25 colon cancers and 86 rectal cancers,67 males and 44 females.Basing on the total somatic mutation counts,patients were divided into hypermutated(N=24)and non-hypermutated group(N=87).Besides,16 patients from the hypermutated group exhibited high level of tumor microsatellite instability(MSI),and the remaining 8 hypermutated patients exhibited tumor microsatellite stability(MSS).Significantly mutated gene(SMG)analysis showed that TP53(61%)and SMAD4(22%)mutations were enriched in non-hypermutated group,while APC(83%)and KRAS(71%)were more likely to be mutated in hypermutated group.Interestingly,significant enrichment of mutations in PTEN and PIK3CA were observed in patients with hypermutated and MSS,while frameshift mutations of BAX were restrictedly detected in patients with hypermutated and MSI.In addition to the well reported CRC related SMGs,we also identified a candidate SMG encoded the nerve growth factor(NGF),and survival analysis suggest the positive relationship between NGF lower expression with patient survival.Cell biology experiments found that cell proliferation as well migration ability increased when increase the expression level of NGF.In addition,germline mutations of mismatch repair(MMR)genes were identified in 10 patients,and those patients were thus diagnosed as Lynch syndrome.6 of them blongs to hypermutaed/MSI,1 patients belong to hypermuated/MSS,while 3 patients belong to non-hymermutaed group.Genome wide association study(GWAS)conducted basing on the East Asians from The Exome Aggergation Consortium(ExAC)database CRC patiens(N=3933)found single nucleotide polymorphisms(SNP)rs144494034(P=1.06×10-12,OR=5.0,95%CI:3.1-8.1)assoticated with disease susceptibility.Further examining rs144494034 in larger patient cohort with all ages through Sanger sequencing found the mutant allele were significantly enriched in patients younger than 50 years old(P=5.88×10-6,OR=4.5,95%Cl:2.3-8.8).Comparing our data as well patients in International Cancer Genome Consortium(ICGC)cohort with CRC patients in TCGA cohort,significant higher mutation rate of APC was observed in Caucasians(81.2%)than that in Chinese CRC patients(45.8%)(P=4.75×10-29).Besides,mutation frequency of APC also significantly enriched in old patients(P=0.03)and exhibited positive relationship with patient survival(P=0.0029),which could be used as a prediction marker for CRC treatment.RNA sequencing(RNA-seq)identified PTPRK-RSP03 fusion gene in 3 patients.Further,reverse transcription polymerase chain reaction(RT-PCR)found a new PTPRK-RSPO3 fusion pattern,of which exon 13 of PTPRK fusioned to the exon 2 of RSP03.Specially,patients carried fusion genes are considered to have specific fusion circular RNA(f-circRNA)in some cancers,we thus determined the presence of fcircRNA for PTPRK-RSP03 and multiple fusion patterns have been identified but with no specific junction site shared among PTPRK-RSP03 carriers.Study on metastatic analysis enrolled six paients with accessiblie primary tumors,liver metastases and lung metastases samples at the same time.The average age of the patients was 65 years old,5 male patients,1 female patient,3 patients with colon cancer,2 patients with rectal cancer,1 patient with colon and rectal tumors,3 patients with lymph node metastasis,and all patients presented with tumor microsatellite stability(MSS).Among them,patient C06 received neoadjuvant therapy and five patients received adjuvant therapy between surgeries except patient C01.WES was performed on 80 lesions of the six CRC patients,and the sequencing data of 74 lesions were included in the tumor phylogenetic analysis.patient C01 was found to have tumor lesions in both the colon and the rectum.WES analysis showed that the two primarys lesion hold different drive gene mutaions.The major driver gene mutations of colon tumors were:APC,CSDE1,CYLD,SETD2 and TP53,while the maj or driver gene mutations of rectal tumors were APC,FBXW7,TP53 and ZFP36L2.Phylogenetic analysis showed that all metastatic tumors of patient C01 were derived from the recrum primary tumors,while the subsequent lung metastases were derived from the early liver metastases and obtained JAK2 mutaion during the followed progression in partial subclones.Patient C02 was found to have primary tumor in the transverse colon,and WES analysis showed that the major driver gene mutations of the tumor progenitor cells were APC,KRAS,TP53 and ZFP36L2.In addition,tumor phylogenetic analysis showed that lymphatic metastasis and distant metastas were originated from different subclone cells of the primary tumor,and had different driver gene mutations,POLE and RUNX1 respectivily.In addition,tumor metastases in the late progression stage occurred different driver gene mutations,such as ATM,PPP2R1A,PPTRD,BRCA2 and DHX9 mutations in the liver metastases,FAT1 mutations in the lung metastases,and EP300,KMT2B,IRF2,EGFR,TET2 mutations in the lymphatic metastases.Patient C03 had the primary tumor in the rectum,and WES analysis showed that the main driver gene mutations of the tumor progenitor cells were APC,CACNA1A,FBXW7,H3F3A,KRAS,TCF7L2 and ZFP36L2,and the ZFP36L2 gene mutation even occurred simultaneously on both alleles.In addition,tumor phylogenetic analysis suggests that the lung metastases originated from the earlier primary tumor subclones,and some subclones aquired CHD3 gene mutations during the later progression courses.On the other hand,the liver metastases originated from the later subclones of the primary tumor different from that lung metastases.Patient C04 had primary tumor in the ascending colon,and WES analysis showed that the major driver gene mutations of the tumor progenitor cells were AMER1,APC,KRAS,PIK3CA,MAP2K4 and SOX9.In addition,tumor phylogenetic analysis suggested that liver metastases were derived from the earlier primary subclones,and no driver gene mutations occurred during subsequent progression,while the advanced lung metastases were independently derived from the different primary subclones that obtained secondary APC gene mutations,and mutations in the EPHA3,ERBB4 and KMT2B driver genes occurred during the later tumor progression.Patient C05 had primary tumor in the colon,and WES analysis showed that the major driver gene mutations of the progenitor tumor cells were APC,NRAS,TCF7L2 and TP53.In addition,tumor phylogenetic analysis suggests that the liver metastases and lung metastases occured simultaneously and directly from different subclones of the primary tumor.The tumor subclones occured EPHA3 gene mutation or KMT2B gene mutation during the course of liver or lung metastasis and colonization respectively.The same tumor metastasis pattern was observed in patient C06 with primary tumor found in the rectum.WES analysis showed that the major driver gene mutations of the tumor progenitor cells were APC,SCAF4 and TP53.Tumor phylogenetic analysis suggested that the liver metastasis and lung metastasis occurred simultaneously,and were directly transferred from the different subclones of the primary tumor,and some tumor cells colonized in the liver occured NOTCH1 gene mutation in the later stage,while some tumor cells had mutation in FLT3 during the lung metastasis and colonization.In addition,according to the results of WES analysis,we divided the tumor nonsynonymous mutations into four major mutation catigories:Trunk,Branch,Shared and Private.Basing on the 299 driver genes listed in TCGA database,we further dividing the Trunk and Branch mutations into driver gene mutations and non-driver gene mutaitons.Totally,the number of driver gene mutations ranged from 5 to 8 in each patietn,and the driver gene mutation proportion in the Trunk category(8.1%)was significantly higher than the Branch category(2.4%)(two-side paired t-test,P=0.04).Among all the driver gene mutations,the APC,TP53 and KRAS genes had the highest mutation frequency,and the APC mutation was detected in all patients,while TP53 mutaiton and KRAS mutaion was detected in all the lesions of 4 patients and all lesions of 3 patients respectivly.Using the "REVOLVER" machine learning method to track genomic evolution trajectories during CRC metastasis,and a genome evolutionary trajectory APC/TP53→ZFP36L2→FBXW7 that is consistent with the whole patient population was identified.In addition,we also discovered a new candidate driven gene mutation,ZFP36L2,which is mutated in patients C01(p.C206Y),C02(p.R160fs),and C03(p.G107fs and p.K102X).Patient C03 even showed biallelic mutation.Structurally,the point mutation(p.C206Y)of patient C03 in the conserved region can result in loss of interaction of this domain with zinc ions,further disrupting the interaction of ZFP36L2 zinc domain with the 3’UTR of the mRNA,which is regulated by ZFP36L2.Immunohistochemical(IHC)analysis showed that ZFP36L2 mutations in tumor tissues of patients C01,C02 and C03 resulted in the loss of protein expression,whereas the absence of ZFP36L2 expression in the tumor tissue of patient C06 may be caused by epigenetic silencing.Further,we screened the ZFP36L2 somatic mutations in two independent CRC cohorts(MET and PRM),and ZFP36L2 mutation frequencies in the MET cohort(N=146)and PRM cohort(N=210)were 10.3%(15/146)and 7.6%(16/210)respectively.After taking tumor metastasis into consideration and excluding microsatellite instability(MSI)tumors,the frequency of ZFP36L2 mutations in the MET cohort(12/122,9.84%)was significantly higher than that in the TCGA database(15/368,4.08%)(Pearson chi-square test,P=0.01),also the mutations frequency in the mPRM cohort(6/44,13.64%)was higher than that in the nPRM cohort(3/135,2.22%)(Continuous correction chi-square test,P=0.009),suggests that ZFP36L2 may play an important role in the CRC metastasis.Conclusions:Young CRC patients in Chinese populations displayed specific genomic characteristics,including well reported CRC related SMGs mutation frequency and newly identified candidate SMGs,such as NGF,of which were suggested to participate in CRC tumor metastasis through cell biology experiments.Specially,significant higher mutation rate of APC was observed in Caucasians than that in Chinese CRC patients,and its mutations were significantly enriched in hypermutated group as well old patients and exhibited positive correlation with patients’ survival.In addition,RNA-seq analysis found frequently and recurrently PTPRK-RSPO3 fusion gene,which could form different f-circRNA and may be the potential biomarker for CRC patients in liquid biopsy.There are two progression models existed in CRC distant metastasis,the Sequential model and Branch-off model.In the sequential metastasis model,the primary tumor cells can be orderly transferred to the liver and the lung via the blood circulation.In the Branch-off progression model,different tumor subclones can be transferred to the liver or lung simultaneous or metachronous.In addition,during CRC metastasis,the genomic evolutionary trajectory of APC/TP53→ZFP36L2→FBXW7 was identified,and the high frequency ZFP36L2 gene mutation was mainly enriched in metastatic colorectal cancer,suggesting that ZFP36L2 may be involved in CRC metastasis and play an important role.
Keywords/Search Tags:Colorectal cancer, whole-exome sequencing, RNA-sequencing, bioinformatics, young patients, tumor metastasis, genomic characteristic, driver gene mutation
PDF Full Text Request
Related items