Identification Of Prognostic Molecular Subgroups Using Immune-related Genes And A Novel Biomarker TNNT1 In Pancreatic Cancer

Objective:Pancreatic cancer is one kind of malignant tumors in the digestive tract with high malignancy,low early diagnosis rate,and high mortality,the morbidity and mortality of pancreatic cancer are gradually increasing in recent years.According to the global cancer data releases by the International Agency for Research on Cancer,there were 495773 new cases of pancreatic cancer worldwide in 2020,accounting for 2.6%of the total new tumors,and 466003 deaths caused by pancreatic cancer accounting for4.7%of total new deaths.Due to the low sensitivity to chemotherapy and radiotherapy of pancreatic cancer cells,the surgical resection is still considered as the main treatment for pancreatic cancer,but the postoperative recurrence rate is relatively high,so the overall prognosis of pancreatic cancer is still not improved recently.The traditional TNM staging only relied on the basic clinical and pathology information,or combined with the clinical manifestations of pancreatic cancer patients,this method could only roughly estimate the survival time of patients which couldn’t meet the requirements of personalized and accurate treatment.The sensitivity and accuracy of classical tumor biomarkers such as CA19-9,CA125,and CEA in the aspects of early diagnosis and prognosis prediction of pancreatic cancer are also insufficient to meet the demands of precision medicine.Therefore,improving the early diagnosis rate and establishing new risk prediction models will be helpful to guide the clinical decisions and improve the overall prognosis of pancreatic cancer patients.Several previous studies based on different omics data had confirmed that pancreatic cancer has significantly different molecular features.In addition,previous studies had shown that the interaction between tumor cells and immune microenvironment plays a vital role in regulating the malignant behaviors of pancreatic cancer.Further investigating the mechanisms behind the interaction may help to understand the immune regulation in the progression of pancreatic cancer,contributing to the development of immunotherapy.Therefore,we try to use immune related genes to identify novel molecular subgroups for pancreatic cancer,providing a new perspective to investigate the mechanisms behind the occurrence and development of pancreatic cancer.In addition,the newly constructed prognostic prediction model can assist to evaluate the prognosis of pancreatic cancer patients and help clinicians to make clinical decisions.TNNT,also named Troponin T,a 30-35 k Da protein containing 220-300 amino acids.It plays an important role in the contraction and relaxation of striated muscle.In vertebrates,there are three isoforms of TNNT including TNNT1,TNNT2,and TNNT3.Several previous studies had shown that TNNT1 plays an important regulatory role in multiple human diseases.For example,overexpression of TNNT1 in the retinal pigment epithelial cells can promote its migration ability.In the ketamine-induced neuropsychiatric disorders mice model,the TNNT1 expression level of the brain tissue was up-regulated.Clinical studies had shown that mutations of TNNT1 are significantly correlated with HDL-C expression level,coronary heart disease and linear myopathy.In the field of cancer research,TNNT1 was significantly overexpressed in breast cancer tissues,and significant correlations with clinical stage,T stage,and N stage were also identified in breast cancer.Overexpression of TNNT1 could significantly improve the proliferation of breast cancer cells.Studies had shown that TNNT1 is significantly overexpressed in colon cancer tissues and patients with high expression of TNNT1 have a poor prognosis.Overexpression of TNNT1 could promote the proliferation,migration,and invasion of colon cancer cells by regulating the EMT process.The expression of TNNT1 was significantly correlated with the prognosis of patients with malignant pleural mesothelioma.Above researches showed that TNNT1 has a promoting effect in various malignant diseases,but the biological effect of TNNT1 in pancreatic cancer has not been reported.Therefore,this study tries to further explore the role of TNNT1 in the occurrence and development in pancreatic cancer and its related molecular signaling pathways.Methods:1.The multi-omics data of pancreatic cancer were obtained from the TCGA database and preprocessed properly.The prognostic genes of pancreatic cancer were identified and selected by the Log-rank and COX survival analysis.We obtained the immune-related genes from the Imm Port database.The intersection of immune-related genes and prognostic genes were reserved for further consensus cluster analysis.The specific molecular subgroups were identified based on the consensus cluster analysis and prognostic immune-related genes.The GO,KEGG,and GSEA methods were utilized to identify the signaling pathways associated with the molecular subgroups.Immune infiltration and mutation spectrum analyses were also used to reveal the molecular characteristics of the molecular subgroups.The WGCNA method was used to identify the geneset that is significantly associated with the molecular subgroups.The Cytoscape software and Metascape database were used to identify the hub genes of the above geneset.Finally,we constructed a prognostic prediction model based on the immune-related genes via the LASSO regression model.The ROC curve was utilized to test the predictive efficiency of the model in the internal test set and internal/external validation sets.The univariate and multivariate COX models were used to determine whether the prognostic prediction model could be recognized as an independent prognostic factor.2.The expression level of TNNT1 in 37 pairs of pancreatic cancer and adjacent normal tissues were detected and comparatively analyzed.Besides,the expression level of TNNT1 in four human pancreatic cancer cell lines(As PC-1,Capan-2,MIA Pa Ca-2,and PANC-1)and human normal pancreatic cell line(HPDE6-C7)were detected.The As PC-1 and MIA Pa Ca-2 cell lines with stably knockdown of TNNT1 were constructed by using the lentivirus transfection method.Subsequently,CCK8,clonal formation,wound-healing and transwell assays were used to detect the changes of proliferation,migration,and invasion ability of the pancreatic cancer cells after stably knockdown of TNNT1.The subcutaneous tumor model of nude mice was utilized to investigate the effect of stable knockdown of TNNT1 on the progression of pancreatic cancer in vivo.3.The relative bioinformatic analyses were conducted to predict the TNNT1 gene related signaling pathways in pancreatic cancer based on the data obtained from the TCGA database.The changes of the key molecules in the EMT and MEK/ERK/GSK-3βsignaling pathways in the As PC-1 and MIA Pa Ca-2 cell lines with TNNT1 stable knockdown were detected by the Western blot experiment.In order to investigate the relationship between TNNT1 and MEK/ERK/GSK-3βsignaling pathways,we performed the rescue experiments with three groups,including negative control,TNNT1overexpression,and TNNT1 overexpression+U0126 inhibitor groups.Results:1.The prognostic molecular subgroups based on the immune-related genes were identified in this study.Firstly,the 1811 immune-related genes were extracted from the Imm Port database.Then,the gene expression data and clinical follow-up information of pancreatic cancer patients were integrated based on the TCGA database.The Log-Rank and COX survival analyses were performed to select the prognostic genes,and genes with P<0.01 in the both tests were considered as genes with prognostic significance.Finally,a total of 67 immune-related genes with significantly prognostic significance were obtained from the intersection of the immune-related genes and prognostic genes.The total samples of pancreatic cancer were clustered into two subgroups using the consensus cluster analysis,named C1 and C2.The survival analysis between C1 and C2subgroups showed significant differences.Then,the differentially expressed genes between the two molecular subgroups were calculated,and a total of 2698 differentially expressed genes were obtained including 827 up-regulated and 1871 down-regulated genes.Subsequently,the GO,KEGG,and GSEA analyses were performed based on the above differentially expressed genes,and multiple signaling pathways related with the occurrence and development of tumor were identified.The immune infiltration analysis between the two subgroups showed that the infiltration levels of 6 kinds of immune cells in the C2 group are higher than that in the C1 subgroup,and significant differences presented in the macrophages,myeloid dendritic cells,T cells CD4+,and T cells CD8+.The mutation spectrum analysis suggested that the mutational frequency of KRAS,TP53,RNF43,FLG,PCDH15,and ADAMTS16 in the C1 subgroup are higher than that in the C2 subgroup.Above results suggested that the poor prognosis of the C1 subgroup may associate with the immunosuppressed status and high mutational rate of cancer-related genes.The WGCNA analysis was utilized to identify the gene modules related with the clinical traits using the above 2698 differentially expressed genes.After analyzing,the brown module was significantly correlated with C1 subgroup,T stage,and clinical stage,suggesting that the brown module may be associated with the progression of pancreatic cancer.Then,the Cytoscape software and Metascape database were used to further identify the hub genes of the brown module.The expression levels of TNNT1,KCNN4,SH2D3A,and PHLDA2 genes were significantly overexpressed in pancreatic cancer compared with normal pancreatic tissue based on the results from the GEPIA database.In addition,the survival analysis showed that the expression levels of above four genes were significantly correlated with overall survival and relapse-free survival in pancreatic cancer,and higher expression levels tended to shorter overall survival and relapse-free survival.The LASSO regression model was used to construct prognostic prediction model based on the above 67 immune-related genes with prognostic significance,and a model consisting of 19 genes was established.The risk score formula of the prognostic prediction model was provided as follows:Risk score=~1 ~9=1Coefficient*Expression of gene.Each sample of different datasets could obtain a risk score according to the above formula.According to the comparison with the median value,those with a risk score greater than the median value were defined as the high-risk group,while those with a risk score lower than the median value were defined as the low-risk group.The survival analysis between the high-risk and low-risk groups in the internal test set and internal/external validation sets showed that the high-risk group had a worse prognosis than the low-risk group.The ROC model was further used to calculate the AUC area of the prognostic model in the different datasets.It was found that the prognostic prediction model could serve as a satisfactory tool in the different datasets.Then,the univariate COX regression analysis was performed on the prognostic model and common clinicopathological information.The results showed that the prognostic model,age,tumor stage,T stage,and N stage were significant prognostic factors in pancreatic cancer(P<0.05).Further multivariate COX regression analysis of the above significant prognostic factors showed that the prognostic model and N stage could serve as independent prognostic factors in pancreatic cancer(P<0.05).2.The western blot and immunohistochemistry assays were used to detect the protein expression level of TNNT1 in the 37 pairs of pancreatic cancer and corresponding adjacent normal tissues.The results showed that the expression level of TNNT1 in the pancreatic cancer tissues were significantly higher than that in the adjacent normal tissues.Based on the correlation analysis between TNNT1 gene expression data and corresponding clinicopathological information of 177 pancreatic cancer samples from TCGA database,the result showed that the samples with high expression of TNNT1tended to be the advanced stage of tumor.Western blot experiment showed that the TNNT1 expression levels of four human pancreatic cancer cell lines(As PC-1,Capan-2,MIA Pa Ca-2,and PANC-1)were higher than that of normal pancreatic cell line(HPDE6-C7).Besides,the TNNT1 expression level of As PC-1 and MIA Pa Ca-2 cell lines were relative higher among these cell lines.Therefore,As PC-1 and MIA Pa Ca-2were selected to construct stably knockdown TNNT1 cell lines via lentivirus transfection method for further cellular functional experiments.The results of CCK8 and clonal formation experiments confirmed that stable knockdown of TNNT1 could inhibit the proliferation and clonal formation abilities of pancreatic cancer cells.The results of wound-healing and transwell assays also showed that stable knockdown of TNNT1 could inhibit the migration and invasion abilities of pancreatic cancer cells.The subcutaneous tumor model with nude mice also confirmed that knockdown of TNNT1 expression could significantly inhibit the progression of pancreatic cancer in vivo.Above results suggested that TNNT1 might be a potential therapeutic target for pancreatic cancer.3.In this study,we tried to predict the potential molecular mechanisms of TNNT1 based on the data from the TCGA database and GO/KEGG/GSEA analyses.After analysis,we found that TNNT1 is significantly correlated with the epithelial mesenchymal transition,apical junction,and MAPK signaling pathways.After stably knockdown of TNNT1 in pancreatic cancer cells,the expression level of E-cadherin was significantly up-regulated,while the expression levels of N-cadherin,Vimentin,MMP9,and Twist were significantly down-regulated.These results suggested that TNNT1 knockdown could significantly inhibit the epithelial mesenchymal transformation in pancreatic cancer.In addition,the expression levels of P-MEK1/2,P-ERK1/2,and P-GSK-3βwere significantly inhibited after stably knockdown of TNNT1 in As PC-1 and MIA Pa Ca-2cell lines,while the expression levels of MEK1/2,ERK1/2,and GSK-3βwere not significantly changed.The rescue experiment was designed in this study to further confirm the regulatory relationship between TNNT1 and related signaling pathways in pancreatic cancer.The rescue experiment contained three groups including negative control,TNNT1 overexpression,and TNNT1 overexpression+U0126 inhibitor groups.The results of CCK8,clonal formation,wound healing,and transwell assays showed that overexpression of TNNT1 could significantly enhance the proliferation,migration,and invasion of pancreatic cancer cells,while adding the U0126 inhibitor could partially eliminate the promoting effect of TNNT1 overexpression.The western blot assay was used to detect the expression of TNNT1 among the three groups.The results showed that the expression level of TNNT1 in the TNNT1 overexpression group of As PC-1 and MIA Pa Ca-2 were significantly up-regulated,while adding U0126 inhibitor did not affect the TNNT1 expression.The expression levels of EMT related proteins in the above two cell lines were further detected.After overexpression of TNNT1 in As PC-1 and MIA Pa Ca-2cells,the expression levels of N-cadherin and Vimentin were significantly up-regulated,while the expression level of E-cadherin was significantly down-regulated.The addition of U0126 inhibitor could partially reverse the changes caused by TNNT1 overexpression.In addition,the changes of related proteins in the ERK/GSK-3βsignaling pathways were further detected in this study.The results showed that overexpression of TNNT1 could significantly increase the expression levels of P-ERK1/2 and P-GSK-3β,while the expression levels of ERK1/2 and GSK-3βmanifested no obvious change.Similarly,adding U0126 inhibitor could inhibit the changes caused by TNNT1 overexpression partially.These results proved that TNNT1 could regulate the malignant biological behaviors of pancreatic cancer via regulating EMT and MEK/ERK/GSK-3βsignaling pathways.Conclusions:1.In this study,we successfully identified and constructed specific prognosis-related molecular subgroups by the consensus cluster analysis using immune-related genes in pancreatic cancer.The GO,KEGG,and GSEA enrichment analysis of the differentially expressed genes between the two molecular subgroups provided a new perspective for further understanding the pathogenesis of pancreatic cancer.The comparative analysis of immune infiltration,gene mutation profiles,and clinical features between molecular subgroups provided theoretical support for the difference in prognosis.Four novel biomarkers for pancreatic cancer were identified from the differentially expressed genes between the molecular subgroups using the WGCNA analysis,including TNNT1,KCNN4,SH2D3A,and PHLDA2.The prognostic prediction model based on the 19 immune-related genes was established by the LASSO regression analysis for pancreatic cancer which could be used to evaluate the prognosis of pancreatic cancer patients and aid clinicians in making management-related decisions.2.The expression level of TNNT1 in the pancreatic cancer tissue was significantly higher than that in the corresponding paracancerous tissue.The comparative analysis of the clinicopathological information suggested that pancreatic cancer patients with high expression of TNNT1 gene tend to be in the advanced stage.The higher expression level of TNNT1 indicated poorer prognosis.Knockdown of TNNT1 could significantly inhibit the malignant biological behaviors of pancreatic cancer cells based on the in vitro and in vivo experiments,which also indicate that TNNT1 could be a potential therapeutic target for pancreatic cancer.3.It was further confirmed that TNNT1 regulates the malignant biological behaviors of pancreatic cancer cells via regulating the epithelial-mesenchymal transformation and activating the MEK/ERK/GSK-3βsignaling pathways based on the rescue experiments.