Pleiotropic tumor suppressor functions of WWOX antagonize metastasis

Abstract

Tumor progression and metastasis are the major causes of death among cancer associated mortality. Metastatic cells acquire features of migration and invasion and usually undergo epithelia-mesenchymal transition (EMT). Acquirement of these various hallmarks rely on different cellular pathways, including TGF-β and Wnt signaling. Recently, we reported that WW domain-containing oxidoreductase (WWOX) acts as a tumor suppressor and has anti-metastatic activities involving regulation of several key microRNAs (miRNAs) in triple-negative breast cancer (TNBC). Here, we report that WWOX restoration in highly metastatic MDA-MB435S cancer cells alters mRNA expression profiles; further, WWOX interacts with various proteins to exert its tumor suppressor function. Careful alignment and analysis of gene and miRNA expression in these cells revealed profound changes in cellular pathways mediating adhesion, invasion and motility. We further demonstrate that WWOX, through regulation of miR-146a levels, regulates SMAD3, which is a member of the TGF-β signaling pathway. Moreover, proteomic analysis of WWOX partners revealed regulation of the Wnt-signaling activation through physical interaction with Disheveled. Altogether, these findings underscore a significant role for WWOX in antagonizing metastasis, further highlighting its role and therapeutic potential in suppressing tumor progression.

Fig. 1

mRNAs profiling using Affymetrix analysis. a Hierarchical unsupervised clustering of gene expression in MDA-MB435S cells revealed the presence of four major clusters. Of these, two clusters included genes that were strongly upregulated or downregulated in WWOX (blue) cells compared to EV (red) or WFPA (green) cells. b PCA results show the clustering of the replicates from each clone, emphasizing the clustering of the EV (RED) and WFPA (blue) in one dimension that is separate from that of WWOX (green). c Volcano plot analysis showing upregulated and downregulated genes comparing WWOX-expressing cells to EV and WFPA-expressing cells

Fig. 2

miRNAs profiling in WWOX-expressing MDA-MB435S metastatic cells and workflow of the intersection analysis with mRNA profiling. a Heatmap with candidate miRNAs. “W” samples are wild type, “E” samples are empty vector and M samples are mutated WWOX ones. b Workflow of the miRNA targeting prediction and function prediction analyses

Fig. 3

Outcome of mRNA-miRNA intersection analysis. a Pathway enrichment analysis on the validated miRNA targets. Heat map showing pathway enrichment analysis for the miRNA targets that were changed upon WWOX expression (both up- and downregulated). Scale-bar shows the −log(P-value); 3.6 is the lowest P-value (=0.000251). b Diseases and biological function analysis on the validated targets of miRNAs. Heat map showing diseases and biological function enrichment analysis for the targets of miRNAs that were changed upon introduction of WWOX expression (both up- and downregulated). Scale-bar shows the -log(P-value); 1.3 = 0.05 P-value

Fig. 4

WWOX modulates TGF-β signaling via miR-146a targeting of SMAD3. a Immunoblot of SMAD3 in MDA-MB435 cells; Ev-empty vector, WFPA-expressing mutated WWOX, WWOX-expressing wild-type WWOX. b Luciferase assay of SMAD3 3′UTR (wild-wild type, mut-mutated 3′UTR), with or without miR-146a. Bars indicates SD. c Immunoblot showing SMAD3 and SMAD2 levels in MDA-MB435S cells upon expression of miR-146a or miR-363. d Immunoblot showing Smad3 in MDA-MB435S cells; Ev-empty vector, and WWOX-expressing wild-type WWOX, and assays were performed after expression of anti-miR-146a or anti-Scr for 24 and 48 h. GAPDH was used as an endogenous control in (a), (c) and (d). Densitometry analysis for SMAD3 levels is shown below each blot

Fig. 5

WWOX physically interacts with DVL2 and functionally results in reduced Wnt pathway activation. a WWOX-interacting proteins in MDA-MB435S cells. b Immunoblot showing the co-IP of WWOX and DVL2 from HEK293T cells using Flag tagged DVL2 and Myc-tagged WWOX; mouse IgG was used as a negative control. Control- without Wnt ligands, Wnt- with Wnt ligands. Heavy chain IgG is indicated for each lane to demonstrate equal amounts of precipitate. c qRT-PCR of AXIN2 in MDA-MB231 cells treated with or without Wnt ligands; UBC was used as an endogenous control. The number of times an experiment was repeated (N) is stated in the legend. d Immunoblot showing Fibronectin and WWOX in SUM149 cells; EV- empty vector, WWOX- wild-type WWOX, and WFPA- mutated WWOX. e qRT-PCR of AXIN2 in SUM149 cells treated with or without Wnt ligands; UBC was used as an endogenous control. GAPDH was used as an endogenous control in (a) and (c). **P-value < 0.01. Statistical analyses (including error bars and p values) for (c) and (e) represent three independent experiments