Starvation-induced suppression of DAZAP1 by miR-10b integrates splicing control into TSC2-regulated oncogenic autophagy in esophageal squamous cell carcinoma

Abstract

Esophageal squamous cell carcinoma (ESCC) accounts for about 90% of all incident esophageal cancers, with a 5-year survival rate of < 20%. Autophagy is of particular importance in cancers; however, the detailed regulatory mechanisms of oncogenic autophagy in ESCC have not been fully elucidated. In the present study, we address how splicing control of TSC2 is involved in mTOR-regulated oncogenic autophagy. Methods: Alternative splicing events controlled by DAZAP1 in ESCC cells were identified via RNAseq. Differential phosphorylation of short or long TSC2 splicing variants by AKT and their impacts on mTOR signaling were also examined. Results: We found that starvation-induced miR-10b could enhance autophagy via silencing DAZAP1, a key regulator of pre-mRNA alternative splicing. Intriguingly, we observed a large number of significantly changed alternative splicing events, especially exon skipping, upon RNAi of DAZAP1. TSC2 was verified as one of the crucial target genes of DAZAP1. Silencing of DAZAP1 led to the exclusion of TSC2 exon 26 (from Leu947 to Arg988), producing a short TSC2 isoform. The short TSC2 isoform cannot be phosphorylated at Ser981 by AKT, which resulted in continuous activation of TSC2 in ESCC. The active TSC2 inhibited mTOR via RHEB, leading to continually stimulated oncogenic autophagy of ESCC cells. Conclusions: Our data revealed an important physiological function of tumor suppressor DAZAP1 in autophagy regulation and highlighted the potential of controlling mRNA alternative splicing as an effective therapeutic application for cancers.

Keywords: DAZAP1; TSC2; autophagy; esophageal squamous cell carcinoma; miR-10b; selective splicing.

Figure 1

Nutrient deprivation-induced miR-10b promotes autophagy and functions as an oncogene in ESCC. (A) Nutrient deprivation significantly induced the miR-10b expression in ESCC KYSE450 and KYSE510 cells. (B) Over-expressed miR-10b increased starvation-induced conversion of LC3-I to LC3-II as well as starvation-induced SQSTM1 degradation in KYSE450 and KYSE510 cells. (C, D) miR-10b enhanced accumulation of autophagosomes which were visualized via the LC3B-II immunofluorescence in KYSE450 and KYSE510 cells. Scale bars, 10 μm. The number of LC3 puncta in cells of each group was calculated from 3 random fields, and at least 30 cells were chosen. (E, F) miR-10b significantly promotes proliferation of KYSE450 and KYSE510 cells. (G) There was elevated miR-10b expression in ESCC tissues compared to normal tissues of cases from Shandong set or the Jiangsu set. (H) Increased miR-10b expression was significantly associated with poor survival of ESCC patients. All results of the mean of triplicate assays with standard deviation are presented. The difference between two groups was calculated using Student's t test (assuming Gaussian distributions) or Wilcoxon Signed Rank Test (not assuming Gaussian distributions). ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.

Figure 2

MiR-10b promotes autophagy by silencing DAZAP1 expression in ESCC. (A) Venn diagram of potential candidate target genes of miR-10b by integrating the results of the algorithms TargetScan, PICTAR, Micro-RNA and MiRDB. (B) qRT-PCR validation of the nine potential target genes of miR-10b in KYSE450 and KYSE510 cells transfected with either miR-10b mimics or NC RNA. (C) miR-10b could significantly inhibit DAZAP1 protein and mRNA expression in ESCC celllines. (D) Schematic constructions of pGL3-DAZAP1 and pGL3-Mut10b. (E) pGL3-DAZAP1 and pGL3-Mut10b were co-transfected into KYSE450 and KYSE510 cells with miR-10b mimics or NC RNA. Luciferase activity was detected at 48h after transfection and normalized relative to the Renilla luciferase expression. Inhibition effects of miR-10b mimics on pGL3-DAZAP1 or pGL3-Mut10b were showed. (F, G) Immunoblot results of extracts from non-starved or starved KYSE450 and KYSE510 cells. Silencing DAZAP1 expression (siDAZ1-1 and siDAZ1-2) increased starvation-induced conversion of LC3B-I to LC3B-II and accelerated rapamycin-induced SQSTM1 degradation in both ESCC celllines. Over-expressed DAZAP1 suppressed conversion of LC3B-I to LC3B-II and down-regulation of SQSTM1 in KYSE450 and KYSE510 cells. (H) DAZAP1 inhibited starvation-induced GFP-LC3 LC3B⁺ autophagosomes formation in both ESCC celllines. The number of LC3 punctae in cells of each group was calculated from 3 random fields, and at least 30 cells were chosen. Autophagy was assessed under non-starved (STV-) or starved (STV+) conditions. The difference between two groups was calculated using Student's t test (assuming Gaussian distributions) or Wilcoxon Signed Rank Test (not assuming Gaussian distributions). All results of the mean of triplicate assays with standard deviation are presented. ^*P < 0.05; ^**P < 0.01; ^***P < 0.001.

Figure 3

DAZAP1 inhibits cell proliferation of ESCC cells. (A) Relative DAZAP1 expression in 86 pairs of ESCC tissues and normal esophageal tissues (Jiangsu set and Shandong set). Significantly down-regulated DAZAP1 expression was observed in ESCC tissues compared with normal esophageal samples in both patient sets. (B) ESCC patients with high DAZAP1 expression exhibited significantly prolonged survival. (C, D) Silencing DAZAP1 expression (siDAZ1-1 and siDAZ1-2) promoted cell proliferation. However, ectopic DAZAP1 expression inhibited KYSE450 and KYSE510 cell growth. Cell numbers were counted at 24h, 48h and 72h after transfection. (E, F) Colony formation assays. On the 14th day after transfection of DAZAP1 siRNAs or expression constructs, colony number in each well was counted. The difference between two groups was calculated using Student's t test (assuming Gaussian distributions) or Wilcoxon Signed Rank Test (not assuming Gaussian distributions). All results of the mean of triplicate assays with standard deviation are presented. ^**P < 0.01; ^***P < 0.001.

Figure 4

MiR-10b increases migration and invasion abilities of ESCC cells via targeting DAZAP1. (A,B) miR-10b and DAZAP1 RNAi (siDAZ1-1 and siDAZ1-2) promoted wound-healing of KYSE450 and KYSE510 cells. Wound-healing area in both celllines is presented as a histogram. (C, D) Over-expressed DAZAP1 evidently inhibited wound-healing. (E, F) miR-10b and silencing DAZAP1 promoted invasion ability of KYSE450 and KYSE510 cells. However, ectopic DAZAP1 expression suppressed invasion ability of both ESCC celllines. Cells on the lower surface of the chamber were stained by crystal violet at 48 h after small RNA transfection. Cell counts data are presented as a histogram. The difference between two groups was calculated using Student's t test (assuming Gaussian distributions) or Wilcoxon Signed Rank Test (not assuming Gaussian distributions). All results of the mean of triplicate assays with standard deviation are presented. ^*P < 0.05; ^**P < 0.01; ^***P < 0.001. (siDAZ1-1 and siDAZ1-2)

Figure 5

DAZAP1 regulates alternative splicing of TSC2 mRNA. (A) RNAseq of KYSE510 cells transfected with siRNAs (siDAZ1-1 and siDAZ1-2) or NC RNA was performed to identify endogenous splicing events controlled by DAZAP1. MATS analyses indicate that dysregulated DAZAP1 led to six hundred and thirteen alternative splicing events, including exon skipping (SE, n = 352), intron retention (RI, n = 22), alternative 5' splice site (A5SS, n = 34), alternative 3' splice site (A3SS, n = 36), and mutually exclusive exon (MXE, n = 102). (B) Venn diagram of the overlapped genes between the alternative splicing (AS) genes and autophagy genes. (C) Decreased inclusion of TSC2 exon26 in mature mRNA of KYSE510 cells transfected with siRNAs (siDAZ1-1 and siDAZ1-2). (D) A 2706bp DNA fragment including human TSC2 exons 25, 26 and 27 as well as introns 25 and 26 was cloned into the pcDNA3.1 vector (pcDNA3.1-TSC2-minigene). (E) Different alternative splicing patterns of TSC2 pre-mRNA were detected using RT-PCR in KYSE450 and KYSE510 cells after silencing DAZAP1 or over-expressed DAZAP1.

Figure 6

DAZAP1 inhibits oncogenic autophagy via the TSC2/RHEB/mTOR signal pathway. (A) KYSE450 and KYSE510 cells were transfected with NC RNA, miR-10b mimics, or DAZAP1 siRNAs (siDAZ1-1 and siDAZ1-2). After 48h, cell lysates were analyzed by western blotting. GAPDH levels were measured as loading controls. (B) KYSE450 and KYSE510 cells were transfected with pcDNA3.1 or pcDNA-DAZAP1. After 48h, cell lysates were analyzed by western blotting. GAPDH levels were measured as loading controls. (C) Model for regulation of alternative splicing of TSC2 mRNA and oncogenic autophagy by DAZAP1 under conditions of nutrient sufficiency.