Background Pancreatic ductal adenocarcinoma is one of the most aggressive cancers, with a 5-year survival rate of less than 8%. The complicated tumor microenvironment, particularly TGF-β, provides possible convenience for the progression of PC cells. TGF-β regulates critical cellular processes, including autophagy. However, the mechanism and effects of TGF-β-mediated autophagy are still poorly understood. Methods Bioinformatics analysis, western blot, transmission electron microscopy and confocal microscopy were used to identify that TFEB is the key factors in TGF-β-induced autophagy. The biological effects of TFEB-driven autophagy were investigated in vitro using transwell and wound healing assays and in vivo using liver metastasis and LSL-KrasG12D/Pdx1-Cre mice models. Luciferase assays and motif analysis were used to assess regulation of RAB5A gene promoter activity by TGF-β-induced TFEB. TFEB levels were measured by real-time PCR, western blot and immunohistochemical staining in clinical pancreatic ductal adenocarcinoma tissues. Results We demonstrated that TGF-β induces TFEB expression via the canonical smad pathway in Smad4-positive PC cells and facilitates TFEB-mediated autophagic activation. TFEB-driven autophagy caused by TGF-β regulates RAB5A-dependent endocytosis of Itgα5 and promotes progression of PC cells. We further showed that enhanced TFEB expression and its direct target RAB5A both predict poor prognosis in PC patients. Conclusions Our findings reveal TFEB-driven autophagy is required for TGF-β induced migration and metastasis of PC cells by promoting endocytosis of Itgα5β1 and focal adhesion disassembly through the TGF-β-TFEB-RAB5A axis. Our results highlight the potential utility of suppressing TFEB-driven autophagy to block PC metastasis. Electronic supplementary material The online version of this article (10.1186/s13046-019-1343-4) contains supplementary material, which is available to authorized users.

No abstract

Novel biomarkers for pancreatic adenocarcinoma are urgently needed because of its poor prognosis. Here, by using The Cancer Genome Atlas (TCGA) RNA-seq data, we evaluated the prognostic values of the differentially expressed miRNAs and constructed a five-miRNA signature that could effectively predict patient overall survival (OS). The Kaplan-Meier overall survival curves of two groups based on the five miRNAs were notably different, showing overall survival in 10.2% and 47.8% at five years for patients in high-risk and low-risk groups, respectively. The ROC curve analysis achieved AUC of 0.775, showing good sensitivity and specificity of the five-miRNA signature model in predicting pancreatic adenocarcinoma patient survival risk. The functional enrichment analysis suggested that the target genes of the miRNA signature may be involved in various pathways related to cancer, including PI3K-Akt, TGF-β, and pluripotent stem cell signaling pathways. Finally, we analyzed expression of the five specific miRNAs in the miRNA signature, and validated the reliability of the results in 20 newly diagnosed pancreatic adenocarcinoma patients using qRT-PCR. The expression results of qRT-PCR were consistent with the TCGA results. Taken together, these findings suggested that the five-miRNA signature (hsa-miR-203, hsa-miR-424, hsa-miR-1266 hsa-miR-1293, and hsa-miR-4772) could be used as a prognostic marker for pancreatic adenocarcinoma.