Integrated Analysis Of Genomic And Transcriptomic Data Of Adenocarcinoma At Gastroesophageal Junction

Background and objective:Adenocarcinoma at the gastroesophageal junction(ACGEJ)refers to a malignant tumor that occurs at the esophagogastric junction.In recent decades,the incidence of ACGEJ has been rising throughout the world.While China is one of the high incidence areas of ACGEJ,the mechanism of development and progression of ACGEJ is yet unclear.Surgical resection is still the mainstay of treatment for ACGEJ and currently few specific effective therapies are available except ERBB2 inhibitors,which contributes to the dismal clinical outcomes of this malignancy.In recent years,a large number of multi-omics data of tumor samples have been generated,making it possible to interpret the molecular characteristics of tumors in a multi-dimensional and comprehensive way.The research results could provide the theoretical foundation basis for the diagnosis,treatment and prognosis prediction of cancer.However,the lack of omics studies focusing on ACGEJ has greatly hindered the deep understanding of the molecular characteristics of this malignancy.Therefore,the present study has assembled the largest Chinese ACGEJ cohort with whole genome sequencing and transcriptome sequencing data to profile the genomic and transcriptomic characteristics of ACGEJ and their roles in the ACGEJ development and progression.We also aim to identify alterations vulnerable to drugs and prognostic predictive biomarkers for ACGEJ,which could provide a theoretical basis for its early diagnosis,precise treatment and prognostic prediction.Methods:In this study,we sequenced the whole genome of 124 ACGEJ tumor tissues and paired peripheral blood samples,of which 123 tumor tissues and paired adjacent normal tissues were submitted to RNA-sequencing.To identify genomic alterations,we performed the analysis of whole-genome mutation profiles,mutational signatures,significantly mutated genes,copy number variations and structural variations.Transcriptome data were analyzed to interpret the possible functional consequences of genomic alterations.We further screened our ACGEJ samples for genomic alterations predicting vulnerabilities to therapeutic agents approved by clinical guidelines,currently in clinical trials,or supported by evidence from pre-clinical assays or case reports,comparing with that of Caucasian ACGEJ samples included in The Cancer Genome Atlas(TCGA).Last but not least,we analyzed the relationships between key genomic and transcriptomic changes of ACGEJ and survival time in patients to identify ACGEJ prognostic prediction biomarkers.We then validated these potential biomarkers in TCGA dataset and Gene Expression Omnibus(GEO)dataset.Results:Firstly,we found that TP53 was the only gene significantly mutated in ACGEJ samples.On the contrary,there were a large number of copy number variations(CNVs)in genomes of ACGEJ.Protein-coding genes subject to CNVs were 5.7 times more than those altered by non-silent SNVs(single nucleotide variations)or indels(insertion and deletion).In particular,we detected 67 significantly amplified peaks and 55 significantly deleted peaks and identified 25 CNV driver genes located in these peaks.Among these CNV drivers,13 had consistent expression changes.Besides,we proved that ACGEJ has a chromosomal unstable phenotype at the transcriptomic and genomic levels and chromosomal instability(CIN)was significantly related to catastrophic genomic events including chromothripsis and kataegis,as well as the number of CNVs and structure variations(SVs).TP53 mutations and loss of CIN suppressors could be the inducement of CIN.Secondly,we found Signature 17-featured SNVs in ACGEJ genomes,which covered most of the significant mutations in CTCF binding sites.There was a significant correlation between the activity of Signature 17 and CIN-associated genomic features.We also clarified for the first time the possible transcriptomic consequences of Signature 17,including increased proliferation abilities,metabolic reprogramming and the immunosuppressive microenvironment.Thirdly,we comprehensively described the response rates of ACGEJ patients to various types of chemotherapeutic drugs and targeted therapeutic drugs,as well as the frequencies of alterations targeted by these drugs.Among them,chemotherapeutic drugs with>70%predicted response rates included Anthracyclines,Gemcitabine and Mitomycin C.The response rates of the three drugs were substantially higher in our patients than in Caucasian patients.For targeted therapeutic drugs,besides ERBB2 inhibitors,WEE1,CDK4/6,PARP and CDK2 inhibitors also had high response rates(>20%)in ACGEJ patients.Among them,the response rate of CDK2 inhibitors in our patients was higher than Caucasian ACGEJ.The target alteration of CDK2 inhibitors was CIN-related CCNE1 amplification.Finally,we found that low tumor mutation load(TMB),more gene-level CNVs and high Signature 17 activity were associated with poor prognosis in patients with ACGEJ.We then constructed a multivariate Cox model to predict prognosis using these three risk factors.In addition,the correlation between Signature 17 activity and prognosis also existed in TCGA gastric cancer data set.As for transcriptomic alterations,we found that the low activity of interferon ?(IFN-?)response pathway,which was negatively correlated with Signature 17,was related to the poor prognosis of ACGEJ patients.Low expression levels of IFI44 and IFI30,two genes involved in the IFN-? response pathway,were also associated with a shorter survival time of ACGEJ patients.The correlation between IFI44 expression level and prognosis for gastric cancer was validated in GEO dataset.The correlation between IFI30 expression level and prognosis for gastric cancer was validated both in TCGA dataset and in GEO dataset.In conclusion,these two genes could be used as a clinical biomarker for predicting the prognosis of ACGEJ.Conclusion:We show that the major changes in the ACGEJ genomes of Chinese patients are CIN promoted tumorigenic focal copy-number variations and Signature 17-featured SNVs.Signature 17 in ACGEJ may be one of the causes of its CIN phenotype and is associated with several malignant phenotypes of tumor.ACGEJ patients are predicted to be sensitive to several chemotherapeutic and targeted therapeutic drugs.Low TMB,more gene-level CNVs and high Signature 17 activity are related to the poor prognosis of ACGEJ.The low activity of the IFN-a response pathway and the low expression level of IFI44 and IFI30 are also related to shorter survival time of ACGEJ.The two genes could serve as prognosis prediction biomarkers clinically.These findings further our understanding of ACGEJ and help develop effective therapies to precisely fight this malignancy.