1 Establishment Of Early Diagnostic Model For Esophageal Squamous Cell Carcinoma 2 MIR503HG Promotes Cell Proliferation,Invasion And Migration Via Hsa-miR-503 Pathway In Esophageal Squamous Cell Carcinoma

Establishment of early diagnostic model for esophageal squamous cell carcinomaEsophageal cancer is a malignant tumor of digestive system that seriously threatens human health.The early symptoms of esophageal cancer are not obvious,and the mortality is high.As a class of important non-coding RNAs,lncRNAs play an important role in tumor progression.Many lncRNAs have been shown to be potential biomarkers and targets for cancer diagnosis and treatment.Previously,whole transcriptome data of 153 pair esophageal squamous cell carcinoma(ESCC)and adjacent tissues were generated in our laboratory.We randomly selected the data of 93 pairs of ESCC and adjacent tissues as the training dataset,and the data of the remaining 60 pairs as the test dataset.In the training dataset,2468 lncRNAs differentially expressed in cancer and adjacent tissues were identified,among which 1326 lncRNAs were up-regulated and 1 142 were down-regulated.Unsupervised hierarchical clustering of these differential lncRNAs could significantly distinguish between tumor and normal tissue.Then,random forest algorithm and 10-fold crossvalidation strategy were adopted to perform recursive feature elimination,and 6 lncRNAs were finally screened out as a group of diagnostic markers.The expressions of these 6 lncRNAs were verified in the test dataset and four external datasets(GSE130078,GSE53622,GSE53624,TCGA+GTEX),and the results were consistent with the training dataset.In addition,the 6 lncRNAs were used for unsupervised hierarchical clustering in the training dataset,test dataset and four external datasets.The results showed that the 6 lncRNAs could effectively distinguish between tumors and normal tissues.These results indicate that the 6 lncRNAs have strong diagnostic potential and can be used to construct a diagnostic model of ESCC.According to the expression data of the 6 lncRNAs screened,we constructed the diagnostic model and the TD score of each patient was obtained.The value of TD score is between 0 and 1,representing the probability of ESCC.The cutoff value is 0.5.Test the effectiveness of the diagnostic model we built in the training dataset,test dataset and the results showed that the specificity,sensitivity,accuracy and AUC of the model are all 1.Then,the model was validated in four external datasets,and the results showed that the specificity,sensitivity,accuracy and AUC of the model in the GSE130078 dataset were 0.6957,0.9565,0.8261 and 0.968 respectively.In GSE53622,the sensitivity,accuracy,specificity and AUC of the model were 0.9,0.95,and 1.In GSE53624,the specificity,sensitivity,accuracy and AUC of the model were 0.9916,0.8908,0.9412 and 0.997.In the unpaired TCGA+GTEX data,the specificity,sensitivity,accuracy and AUC of the model were 0.8413,0.8642,0.8466 and 0.951,respectively.These results indicate that the diagnostic model established by us has a strong diagnostic efficiency.In addition,we found that the diagnostic model was independent of gender,smoking,and alcohol consumption.In order to further verify the robustness of the model in early diagnosis of ESCC,normal and stage I ESCC tissues were selected,and the cutoff value was calculated as 0.562 by using Youden index,and the cutoff value was used to predict stage I ESCC and normal tissues.The results showed that the sensitivity,specificity,accuracy and AUC of the model were I in both the training set and the test set.In the GSE130078 data set,the specificity and accuracy of the model were 0.9565,0.963,and the sensitivity and AUC were 1.In GSE53622 and GSE53624,the sensitivity.accuracy,specificity and AUC of the model were all 1.In TCGA+GTEX data,the specificity,sensitivity,accuracy and AUC of the model were 0.93727,0.57143,0.9281 and 0.901,respectively.Then,we further added the data of stage Ⅱ ESCC tissues and calculated the cutoff value of 0.551,which was used to predict the stage Ⅰ/Ⅱ ESCC and normal tissue.Similar to the previous results,the model showed good performance,especially in TCGA+GTEX data,where the sensitivity of the model increased from 0.57143 to 0.8333.These results indicate that the diagnostic model has a strong robustness in the early diagnosis of ESCC.Finally,we validated the model with a low-throughput qRT-PCR assay in 30 paired ESCC and normal tissues.The results showed that the sensitivity,specificity,accuracy and AUC of the diagnostic model were 0.9474,0.8947,0.9211 and 0.983,respectively.When predicting stage Ⅰ/Ⅱ ESCC and normal tissue,the sensitivity,specificity,accuracy,and AUC of the model were 1,0.8947,0.92,and 0.982,respectively.These results indicate that the diagnostic model also has strong diagnostic performance in low-throughput data.In summary,based on the whole transcriptome data of 153 pair ESCC and adjacent tissues,we identified 6 lncRNAs for the construction of diagnostic models of ESCC.The diagnostic model has strong efficacy and robustness in the diagnosis of ESCC,even in the early diagnosis.MIR503HG promotes cell proliferation,invasion and migration via hsa-miR-503 pathway in esophageal squamous cell carcinomaCurrently,a combination of surgery,radiotherapy and chemotherapy is predominant in the clinical treatment of esophageal cancer.Although the target therapies have been studied for many years,the overall benefits for patients remains limited,and 5-year survival rate for patients with esophageal cancer is still very low.Therefore,it is of great significance to continue to search for effective therapeutic targets for the treatment of esophageal squamous cell carcinoma.Previously,in the screening of early diagnosis models for esophageal squamous cell carcinoma,we identified an lncRNA MIR503HG with abnormally high expression in esophageal squamous cell carcinoma(ESCC)tissues.lncRNA is involved in a variety of biological processes and may play an important role in the tumorigenesis of ESCC.Therefore,this study aims to investigate the function and mechanism of MIR503HG in ESCC.First,we analyzed the expression of MIR503HG in three GEO databases and TCGA combined GTEX data,which indicated that MIR503HG was highly expressed in ESCC and was a potential oncogene.Next,we construct KYSE30 and KYSE510 cells with stable knockdown of MIR503HG using lentiviral system to study the effects of MIR503HG on the proliferation.invasion and migration in ESCC cells.CCK-8 and colony formation experiments showed that the cell proliferation ability of stable knockdown MIR503HG was significantly inhibited.Transwell assay showed that the cell invasion and migration abilities were significantly reduced when MIR503HG was stably knocked down.Consistently,the subcutaneous xenograft experiments showed that the tumor size and weight of the stable knockdown of MIR503HG group were significantly lower than that of the control group.In addition,the effect of stable knockdown of MIR503HG on the cell cycle was examined by flow cytometry.The results showed that stable knockdown of MIR503HG resulted in Gl-S arrest of the cell cycle.Next.we explored the mechanism of MIR503HG.As the host gene of hsa-miR503,MIR503HG may exert its function by regulating hsa-miR-503.We used qRT-PCR to detect the change of expression of hsa-miR-503-3p and hsa-miR-503-5p after stable knockdown of MIR503HG,and found that the expression of hsa-miR-503-3p and hsamiR-503-5p were all down-regulated.This suggests that MIR503HG has a regulatory effect on hsa-miR-503,which may be the underlying mechanism of its function.Furthermore,we performed rescue experiment.The results showed that the overexpression of hsa-miR-503-3p and hsa-miR-503-5p partially restored the proliferation,invasion and migration ability of stable knockdown MIR503HG cells.In summary,this study reveals the function of MIR503HG in ESCC and preliminarily revealed the mechanism by which MIR503HG promotes the proliferation,invasion and migration of ESCC cells by regulating the hsa-miR-503 signaling pathway.MIR503HG may be a new target for target therapy of ESCC.