Screening And Identification Of Candidate Esophageal Cancer Associated Antigens And Evaluation Of Their Diagnostic Values For Esophageal Squamous Cell Carcinoma

Objective Identify candidate of tumor associated antigens(TAAs)which were associated with esophageal squamous cell carcinoma(ESCC)by using serological proteome analysis,searching literature related to genomics and ESCC,systematic review of the literature on autoantibodies in EC diagnosis,and summarizing our previous research results about autoantibodies in ESCC diagnosis.After designing and optimizing a mini-array of multiple TAAs which was useful for ESCC diagnosis,we tried to construct the corresponding prediction model of ESCC,evaluate the possibility and feasibility of the model for early diagnosis of ESCC.The diagnostic value of prediction model for ESCC was verified by using a set of independent population data.The ultimate goal is to provide evidence and information for the establishment of a novel non-invasive diagnostic approach for the immunodiagnossis of early stage ESCC.Methods 1.Screening and identification of candidate antigens associated with ESCC by proteomic analysis.1)Screening candidate ESCC associated antigens and anti-TAAs antibodies positive serum by using Western blot.Sera from 102 patients with ESCC and 58 normal controls were hybridized with proteins from four types of ESCC celllines,TE-2,TE-8,EC1 and EC9706.The proteins which had corresponding antibody positive rate more than 10% in sera with ESCC and had higher antibody positive rate in sera with ESCC than normal control(p <0.05)were observed and defined as candidate ESCC-associated antigens.In the process of screening,the sera of patients with ESCC which were positive response to the candidate ESCC associated antigens were selected,too.2)The anti-TAAs antibodies positive sera were validated and the cell locations of the candidate TAAs were observed by immunofluorescence.3)Identification of the candidate ESCC associated antigens by serological proteome analysis(SERPA).After extracting the total cell proteins of TE-8 cells,three IPG strips were used to run isoelectric focusing electrophoresisusing the same parameter.After SDS-PAGE gel electrophoresis,one of the gels was stained by coomassie brilliant blue R250,the other two gels were transferred to nitrocellulose membrane,one of the nitrocellulose membranes was hybridized with the anti-TAAs antibodies positive sera and the other one was hybridized with normal sera,respectively.The protein spots which have positive reaction with the anti-TAAs antibodies positive sera were analyzed and identified by mass spectrometry.2.Expression and purification of candidate TAAs.The candidate TAAs which were identified by SERPA,reported in literatures or identified based on the previous research results were expressed and purificated by prokaryotic expression system.1)Construction of recombinant expression plasmids of candidate TAAs.Total RNA of EC9706 cells was extracted and reversed to c DNA,target genes were amplified by PCR and ligated into cloning vector.After identification by enzyme digestion and DNA sequencing,the correct target genes were inserted into expression plasmid of p ET-28 a.2)Expression and purification of candidate TAAs.The recombinant expression plasmids were transformed into expression host strains,after optimized of IPTG concentration,the solubility of target proteins were determined.Purification soluble proteins under native conditions by using different concentrations of imidazole,and purification insoluble proteins under denaturing conditions by using different p H of 8M urea.3)Identification of the immunological activity of purified proteins.The immunological activities of purified proteins were identified by Western blot hybridization with specific antibodies.3.Evaluation the diagnostic value of the candidate TAAs autoantibodies for ESCC and establishment of logistic regression model for prediction of ESCC.1)The diagnostic value of single TAA.Serum autoantibodies were detected by indirect enzyme-linked immunosorbent assay(ELISA).Rank-sum test or independent sample t-test was carried out to compare the expression levels of autoantibodies between ESCC cases and healthy controls.Logistic regression models were constructed based on the expression of every anti-TAAs antibody in training cohort which included 324 ESCC patients and 324 age-and gender-matched healthy controls to estimate the risk of being diagnosed with ESCC.In order to predict the sensitivity and specificity of the single anti-TAA autoantibody in the diagnosis of ESCC,the predictive probability of 0.5 was used as a cut-off value.Receiver operating characteristic(ROC)curve analysis was used to evaluate the diagnostic performance for every kind of TAA in ESCC.ESCC patients were divided into early stage ESCC(TNM stage 0,I and II)and late stage ESCC(TNM stage III and IV).We compared the positive rates of anti-TAA autoantibodies in different stages of ESCC to determine whether there were specific indexs which were early stage ESCC associated antigens.2)Establishment the predictive TAAs panel.A backward step(condition)logistic regression model was constructed based on the expression levels of autoantibodies in training cohort to estimate the risk of individuals being diagnosed with ESCC.The predictive probability of 0.5 was used as cut-off value to determine the sensitivity and specificity of the prediction model in the diagnosis of ESCC.And the diagnostic performances for the TAA panels in ESCC were evaluated by ROC curve analysis.In order to optimize the combination of anti-TAAs autoantibodies,excluding the autoantibody which had the minimum diagnostic value(including sensitivity,specificity,AUC,and diagnostic accuracy,etc.)for the predictive model one by one,constructing new prediction models and determining the value of the predictive models for ESCC diagnosis.3)Validation of the TAAs panel.Using an independent cohort which included 186 ESCC patients and 186 age-and gender-matched healthy controls to validate the predictive models.Determine the diagnostic accuracy of prediction according to the prediction probability of diagnosis as ESCC,evaluate the diagnostic value of the models by ROC curve analysis.4)Prediction model for different stages of ESCC diagnosis.Evaluation and comparison the diagnostic performance for early stage of ESCC and late stage of ESCC by the prediction models to analyze whether the prediction models are specificity for early stage ESCC.To further determine the diagnostic performance of prediction models for early stage ESCC,the prediction models were used to compare the diagnostic value of all patients with early stage ESCC and normal controls to determine the ability of the prediction model to distinguish the early ESCC patients from normal population.Results 1.Identification of four candidate ESCC associated antigens by proteomics approach.Results from Western blot showed that there were three protein bands with molecular weight near 40 k D,50 k D and 65 k D had higher positive rates in ESCC patients compared with normal controls.Sera hybridized with the proteins from TE-8 cell line celllines by SERPA,results from mass spectrometry analysis showed that the four candidate TAAs were hn RNP1A2B1 isform A2(hn RNPA2B1),calreticulin(CALR),enolase 1(ENO1)and heat shock protein 60(HSPD1).2.The prokaryotic expression systems of hnRNPA2B1,CALR,HSPD1 and NICD were successfully constructed,and the four proteins were successfully expressed and purified.All of the purified protein had immunological activity.3.15 kinds of anti-TAAs autoantibodies were detected and a logistic regression model for the probability of diagnosed patients with ESCC was constructed based on the expression levels of ten anti-TAAs autoantibodies(p53,p62,HCCR,c-myc,MDM2,hn RNPA2B1,NICD,ENO1,Koc and IMP1)in training cohort.The sensitivity of this TAAs panel was 71.6% and the specificity was 85.2%,the AUC was 0.871(95% CI: 0.843-0.896),the diagnostic accuracy was 78.4%.Further optimization of the combination of anti-TAAs autoantibodies,a prediction model based on 9 anti-TAAs autoantibodies(p53,p62,HCCR,c-myc,MDM2,hn RNPA2B1,NICD,ENO1 and Koc)had a highest AUC(AUC = 0.872,95% CI: 0.844 – 0.897);A prediction model based on 7 anti-TAAs autoantibodies(p53,p62,HCCR,c-myc,MDM2,hn RNPA2B1 and NICD)had a highest diagnostic accuracy(79.2%)for diagnosis of ESCC in training cohort.The AUC of the prediction model based on 7 anti-TAAs autoantibodies was 0.864 for diagnosis of ESCC in training cohort,the sensitivity was 71.9% and the specificity was 86.4%.In validation cohort,the AUC,sensitivity,and specificity of this panel were 0.864,65.6% and 90.3%,respectively.This 7 anti-TAAs autoantibodies model had relatively higher diagnostic performance and cost effectiveness compared with other combinations.The 7 TAAs panel could differentiate early-stage ESCC patients from normal controls,with AUC of 0.864,sensitivity of 71.9% and specificity of 87.8%,and the detection accuracy rate was 83.7%.Conclusion 1.CALR,hn RNP1A2B1,ENO1 and HSPD1 may be candidate esophageal squamous cell carcinoma associated antiges.2.A prediction model based on 7 anti-TAAs autoantibodies(p53,p62,HCCR,c-myc,MDM2,hn RNPA2B1 and NICD)could provide a high diagnostic performance for ESCC,especially for early-stage ESCC patients.