Identification Of LOX And ACSL5 As Potential Prognostic Predictors For Pancreatic Cancer Patients

Objective:Pancreatic cancer is the 14th most common among malignant tumors in the world,with a mortality rate of 9th.At present,surgery is the main method for the treatment of pancreatic cancer.However,as there are no obvious specific symptoms in its early stages,patients usually go to the hospital at late stage.At the same time,pancreatic cancer has the characteristics of high metastasis and invasiveness,and prone to chemotherapy tolerance,which seriously affects the treatment effect.Despite the continuous development of multidrug combinations and neoadjuvant therapy,the 5-year survival rate of patients with pancreatic cancer is between 2% and 9% for many years,and has not improved significantly.Therefore,what indicators to take as a reference and adopt more precise treatment strategies to balance the quality of life and overall survival of patients,has become the hot direction of current researches.In recent years,high-throughput analysis has played an important role in the screening of biomarkers such as tumor diagnostic indicators,drug targets,and prognosis prediction.Although the researches on pancreatic cancer biomarkers have received widespread attention,further screening of biomarkers with the ability to predict the prognosis of pancreatic cancer and optimization the related prognostic prediction models are still needed.Methods:99 patients with pancreatic cancer diagnosed and treated at the Tianjin Medical University Cancer Hospital from 2008 to 2015 were included in this study.Patients were grouped based on progression-free survival(PFS).The patients with PFS > 18 months was obtained as the good prognosis group,while those with PFS <12 months were obtained as the poor prognosis group.Three cases were selected from each of the above two groups,and the differential expression genes(DEGs)between the two groups were screened using a genome-wide expression chip.The genomic RNA samples of 68 patients were collected,and qRT-PCR was used to evaluate the correlation between the expressions of the candidate genes in the tumor tissues and the prognosis of patients.The candidate genes that significantly associated with the prognosis of patients were preferred,and immunology histochemistry staining were used to evaluate their protein expressions.Subsequently,the relationship betweencandidate gene expressions as well as the clinical pathological parameters,such as lymphatic invasion and tumor staging,and the prognosis of patients were analyzed by multivariate analysis.Finally,the TCGA database and STRING online tools were used to investigate the potential mechanisms of the obtained potential biomarkers affecting the progression of pancreatic cancer.Results:(1)178 DEGs were screened from pancreatic cancer patients from three sampleswith good prognosis and poor prognosis respectely by using microarray analysis of gene expression profiles.Among them,110 genes were up-regulated and 68 genes were down-regulated.(2)Based on the differential expression levels of the DEGs andreferences,24 candidate genes were selected for qRT-PCR detection,and the correlation between the mRNA levels of these genes and the patient’s PFS and OS was analyzed.Finally,4 candidate biomarkers were identified.The high expression of LOX suggested a poor prognosis,whereas the high expression of ACSL5,TOX3,and SLC44A4 suggested a good prognosis.(3)Immunohistochemistry(IHC)stainingand bidirectional scoring system were used to quantitatively analyze the correlation between protein expression of candidate biomarkers and clinical indicators.The differential expression of ACSL5 was associatedd to tumor T stage of patients.The level of TOX3 low-expression was significantly associated to clinical stage and vascular invasion of the patients.Meanwhile,patients with high LOX expression and low ACSL5 and TOX3 expression had significantly poor prognosis.(4)The results of multivariate analysis using Pearson correlation coefficient further showed that LOX and ACSL5 were independent prognostic factors of pancreatic cancer and had potential for application of prognosis prediction.(5)The possible molecularmechanism of LOX affecting the progression of pancreatic cancer were analyzed with TCGA database,and the results suggest that LOX had significant correlation with HIF-1a and epithelial mesenchymal transition(EMT)indicators.The protein-protein interaction network was constructed using STRING online tools,and the results showed that ACSL5 may affect the lipid metabolism-related pathways.Conclusion In this study,we analyzed the DEGs between pancreatic cancer tissues of patients with good prognosis and poor prognosis,and identified LOX and ACSL5 as thepotential biomarkers for predicting the prognosis.The high expression of LOX indicated a poor prognosis,while the high expression of ACSL5 indicated a good prognosis.Combining the clinical and pathological characteristics of patients with molecular interaction network analysis results,we speculated that LOX might promote tumor metastasis by regulating the EMT process,while ASCL5 might be participate in tumor progression by affecting lipid metabolism.