Caveolin‐1 down‐regulation is required for Wnt5a‐Frizzled 2 signalling in Ha‐RasV12‐induced cell transformation

Abstract Caveolin‐1 (Cav1) is down‐regulated during MK4 (MDCK cells harbouring inducible Ha‐Ras V12 gene) transformation by Ha‐RasV12. Cav1 overexpression abrogates the Ha‐RasV12‐driven transformation of MK4 cells; however, the targeted down‐regulation of Cav1 is not sufficient to mimic this transformation. Cav1‐silenced cells, including MK4/shCav1 cells and MDCK/shCav1 cells, showed an increased cell area and discontinuous junction‐related proteins staining. Cellular and mechanical transformations were completed when MDCK/shCav1 cells were treated with medium conditioned by MK4 cells treated with IPTG (MK4+I‐CM) but not with medium conditioned by MK4 cells. Nanoparticle tracking analysis showed that Ha‐RasV12‐inducing MK4 cells increased exosome‐like microvesicles release compared with their normal counterparts. The cellular and mechanical transformation activities of MK4+I‐CM were abolished after heat treatment and exosome depletion and were copied by exosomes derived from MK4+I‐CM (MK4+I‐EXs). Wnt5a, a downstream product of Ha‐RasV12, was markedly secreted by MK4+I‐CM and MK4+I‐EXs. Suppression of Wnt5a expression and secretion using the porcupine inhibitor C59 or Wnt5a siRNA inhibited the Ha‐RasV12‐ and MK4+I‐CM‐induced transformation of MK4 cells and MDCK/shCav1 cells, respectively. Cav1 down‐regulation, either by Ha‐RasV12 or targeted shRNA, increased frizzled‐2 (Fzd2) protein levels without affecting its mRNA levels, suggesting a novel role of Cav1 in negatively regulating Fzd2 expression. Additionally, silencing Cav1 facilitated the internalization of MK4+I‐EXs in MDCK cells. These data suggest that Cav1‐dependent repression of Fzd2 and exosome uptake is potentially relevant to its antitransformation activity, which hinders the activation of Ha‐RasV12‐Wnt5a‐Stat3 pathway. Altogether, these results suggest that both decreasing Cav1 and increasing exosomal Wnt5a must be implemented during Ha‐RasV12‐driven cell transformation.


| INTRODUCTION
Caveolin-1 (Cav1), a major component of caveolae, interacts with many signalling molecules via its scaffolding domain and plays an important role in signal transduction, membrane trafficking and cholesterol transport. 1 Accumulating evidence has shown that Cav1 is reduced in tumour-derived cells or oncogene-transformed fibroblasts. [2][3][4][5][6] In addition to its role as a tumour suppressor, Cav1 is also associated with the regulation of focal adhesions and integrinmediated actin remodelling; both mechanisms have been widely studied with respect to mechanotransduction. 7,8 Recently, we showed that cancer cells or Ha-Ras V12 -overexpressing cells exhibit a different mechanical phenotype, showing cell softening and loss of stiffness sensing. 9 Cav1 expression is down-regulated as a consequence of Ha-Ras V12 -mediated oncogenic stimulus employed using an IPTG-inducible expression system. In NIH3T3 fibroblasts, Cav1 increases RhoA activity and Y397 FAK phosphorylation, which directed actin cap formation and contributes to cell elasticity and stiffness sensing. Therefore, the Ha-Ras V12 -induced fibroblast-transformed phenotype can be reversed by Cav1 re-expression and mimicked by Cav1 silencing. 9 Approximately 90% of human cancers occur in epithelial tissues. In the early stages of cancer, cell junctions are often disrupted. 10 Instead of stress fibres or actin caps, circumferential actin rings are prominent in epithelial cells. These actin filaments are associated with adherens junctions and tight junctions that generate actomyosin tension, 11 which plays a role in mechanotransduction and regulates cell stiffness. 12,13 Importantly, Cav1 recruits the E-cadherin/b-catenin complex to the membrane, which stabilizes the cell-cell adhesion of normal epithelia. 14,15 Nevertheless, whether and how Cav1 down-regulation is responsible for epithelial transformation remains unclear.
In this study, we showed that Cav1 was down-regulated after Ha-Ras V12 induction in MK4 cells. As expected, Cav1 overexpression averted the Ha-Ras V12 -driven cellular and mechanical transformation of MK4 cells. However, Cav1 silencing did not elicit the cellular and mechanical transformation of MK4 or Madin-Darby canine kidney (MDCK) cells, suggesting that multiple changes in gene expression collaboratively contribute to Ha-Ras V12 transformation. A growing body of evidence suggests that exosomes transfer proteins and functional RNA, contributing to the propagation of a transformed cell phenotype. [16][17][18][19] Using proteomics analysis, Simpson and colleagues demonstrated that several factors carried by exosomes contributed to the Ha-Ras V12 -induced epithelial-mesenchymal transition (EMT) in MDCK cells. 20 Thus, the impact of Ha-Ras V12 -activated exosomal factors on the transformation of Cav1-silencing MDCK cells was evaluated.

| RT-PCR
Total RNA was extracted from cells using TRIzol reagent (Invitrogen-Molecular Probes, Carlsbad, CA, USA) according to the manufacturer's instructions. The RNA quality was verified and reverse-transcribed using Moloney murine leukaemia virus reverse transcriptase (Promega, Madison, WI, USA). The cDNA was subsequently used as a template for PCR using primers specific for the following genes:  PCR was performed at 94°C for 5 minutes, followed by 25   cycles at 94°C for 30 seconds, 60°C for 30 seconds and 72°C for 30 seconds, with a final step at 72°C for 10 minutes. The PCR products were resolved on a 1.2% agarose gel containing ethidium bromide and visualized under a UV transilluminator.

| Measurements of cell stiffness by atomic force microscopy
The JPK NanoWizard â II AFM with BioCell (JPK Instruments, Berlin, Germany) was used as previously described. 24 The measurements of cell stiffness were performed as previously described. 9,24 2.6 | Immunofluorescence staining and confocal microscopy Immunofluorescence staining was performed as previously described. 22 The following primary antibodies were used: Cav1, bcatenin and E-cadherin (BD Biosciences Pharmingen; San Jose, CA, USA), and claudin-1 and ZO-1 (Invitrogen). After extensively washing with PBS, the cells were incubated with secondary antimouse or rabbit IgG conjugated with Alexa 488 (Invitrogen-Molecular Probes) and/or phalloidin-TRITC (Sigma-Aldrich) and 10 lg/mL Hoechst 33258 (Sigma-Aldrich) for 1 hour. The imaging was performed from sequential z-series scans using the FluoView TM FV1000 confocal microscope (Olympus, Tokyo, Japan) with a 609 water immersion lens, NA 1.35 (Uplsapo).

| Western blot analyses
Western blot analysis was performed as previously described. 22 Primary antibodies directed against the following proteins were used:

| Transwell migration assay
Migration was evaluated using a 24-well transwell assay (8 lm pore size polycarbonate membrane, Corning, MA, USA) as previously described. 9 Briefly, 5 9 10 4 cells from each clone were suspended in 300 lL of serum-free DMEM and subsequently seeded onto the upper chamber, whereas 600 lL of DMEM containing 10% FBS and 10 lg/mL of collagen I was added to the outer side of the chamber.
After culturing at 37°C and 5% CO 2 in a humidified incubator for 6 hour, the cells on the upper surface of the membrane were removed using a cotton-tipped swab, and the penetrated cells on the lower membrane surface were fixed using 4% paraformaldehyde and subsequently stained with crystal violet. Cell migration values were determined by counting all penetrated cells of each clone under a phase-contrast microscope (2009 magnification) and subsequently normalized to the control.

| Evaluation of cell proliferation with Click-iT â EdU
Cell proliferation was evaluated using a Click-iT EdU Alexa Fluor 488 Imaging Kit (Invitrogen-Molecular Probes) as previously described. 9 2.10 | Preparation of conditioned medium (CM) and

isolation of exosomes
The cells were cultured under normal culture conditions. At approximately 90% confluence, the cells were rinsed twice with sterile PBS and changed to fresh culture medium to initiate conditioning. After 48 h of incubation, CM was collected and centrifuged at 2000 g to remove cells and debris. The supernatants were filtered through a 0.22-lm filter, aliquoted and subsequently stored at À80°C until further use. Exosomes were isolated using total exosome isolation reagent (Invitrogen) according to the manufacturer's instructions.
Briefly, the cell-free CM was mixed well with 0.5 volumes of total exosome isolation reagent and incubated at 4°C overnight. Subsequently, the mixtures were centrifuged at 10 000 g for 1 hour at 4°C. Finally, the resulting exosomes pellets were resuspended in PBS and stored at 4°C for up to 1 week or at À20°C for long-term storage.

| Exosome labelling and uptake analysis
The exosomes were labelled with 1,1 0 -Dioctadecyl-3,3,3 0 ,3 0 -tetramethylindocarbocyanine perchlorate (DiI, Sigma). Briefly, the purified exosomes (1 mg/ml in PBS) were incubated with DiI (5 lg/mL) at 4°C for 20 minutes in the dark with gentle agitation. DiI-labelled exosomes were washed twice with PBS by centrifugation at 10 000 g for 1 hour at 4°C. The pelleted exosomes were finally resuspended in PBS and stored at 4°C. For exosome uptake studies, subconfluent cells were incubated with DiI-labelled exosomes (100 lg/mL) for 24 hour at 37°C. Surface-bound exosomes were removed after extensive washing with serum-free medium. Finally, cells were fixed with 2% paraformaldehyde at room temperature and observed using confocal microscopy. Exosomes uptake was calculated in representative cells using ImageJ software, and the results are represented as (the intensity of red pixels divided by cell area)*1000.

| Nanoparticle tracking analysis (NTA)
The data presented in this study were generated using NanoSight were set as follows: temperature, 21°C; viscosity, 1 cP; frames per second, 30; measurement time, 90 seconds; and detection threshold, 5. The data are presented as the average and standard deviation of the three video recordings.

| Statistical analyses
All data are expressed as the mean AE SEM of at least two independent experiments. The results were analysed via ANOVA and t-tests Notably, MK4+I-EXs, but not MK4-EXs, stimulated cell scattering ( Figures S4E, and 2H) and cell softening ( Figure 2I)    A previous study showed that Fzd2-Stat3 signalling plays a critical role in Wnt5a-mediated EMT and cell migration. 28 To assess whether Fzd2-Stat3 signalling is involved in Ha-Ras V12 -Wnt5a-induced transformation, we evaluated the Fzd2 expression and Stat3 phosphorylation in MK4 cells treated with or without IPTG. MK4 cells expressed low levels of Wnt5a and Fzd2, which were markedly increased after Ha-Ras V12 induction ( Figures 3D and F, 6A and B). In addition, the phosphorylation of Stat3 on Tyr 705 was elevated, revealing that the Wnt5a-Fzd2-Stat3 non-canonical pathway might be activated in Ha-Ras V12 -induced transformation ( Figure 6C and D).

However, Fzd2 mRNA levels remained unchanged in response to
Ha-Ras V12 ( Figure 6E and F), suggesting that Fzd2 expression is controlled through a post-transcriptional mechanism. Silencing Cav1 increased Fzd2 protein levels without affecting its mRNA levels in MDCK/shCav1 cells ( Figure 6G and H). Consequently, treatment of  Aberrant Wnt signalling, in response to overproduction in Wntsecreting cells or mutations in Wnt-receiving cells, has been implicated in many cancers. 30 Recent studies have pointed to a critical role of exosomes for Wnt secretion and extracellular travelling. 31,32 Exosome-bound Wnts and their signalling activities were functionally implicated during embryonic development and cancer progression. 31,33,34 Here, we showed that the overexpression of Ha-Ras V12 increased the synthesis and release of exosomal Wnt5a ( Figure 3D-F), which subsequently induced cellular and mechanical transformation in MK4 cells (Figures 3G and 4A-D). Noteworthily, the level of Cav1 determines the cellular responsive to Ha-Ras V12 -activated EXs, CM, or Wnt5a (Figures 5 and 6). Currently, a number of Wnt5a receptors were reported, including Fzd2, Fzd3, Fzd4, Fzd5, Fzd6, Fzd7, Fzd8, RYK, ROR2 and CD146. 35 Results from several independent studies indicated that Fzd2 expression might drive EMT through the non-canonical Wnt pathway in different cancer cells. 28,[36][37][38][39] Gujral et al 28 reported that Wnt5a and its ligand Fzd2 are overexpressed in several metastatic cancer cell lines and tumours. These authors further identified a non-canonical Fzd2-Fyn-Stat3 pathway that mediates Wnt5a-induced EMT and cell migration. RT-PCR results showed that Fzd2, but not Fzd5 and Fzd8, were highly expressed in the clones derived from MDCK cells ( Figure S7).
Because Ha-Ras V12 induced EMT-like morphological change in MK4 cells, we focused our research on Wnt5a-Fzd2-Stat3 pathway. We observed that Cav1-reduced cells, either due to Ha-Ras V12 overexpression or Cav1-targeted shRNA, showed augmented Fzd2 protein expression (Figure 6A  The importance of tumour-derived exosomes in tumour progression cannot be overemphasized. Cancer cell-derived exosomes promote the transformation of cells through an autocrine mechanism or through uptake by normal cells surrounding the tumour, which might confer the transformed characteristics of cancer cells upon normal recipient cells. 47,48 Bissell and Hines proposed that the microenvironment surrounding the tumour provides tumour-suppressive signals as long as the architecture of the tissue homeostasis is essentially controlled. 49 Although the initiation of tumours resulting from a potent oncogene is unavoidable, their progression to malignancy can and should be controllable. Overcoming the protective roles of the physiological microenvironment requires "promotion" agents, which are typically associated with aberrant repair and fibrosis. Indeed, wound healing and TGF b1 are considered highly effective promoting stimuli. 50 Williams et al 51 showed that Cav-1 expression in both epithelial and stromal cells provides a protective effect against mammary hyperplasia and mammary tumorigenesis. Cav1 negatively regulates the exosome internalization in glioblastoma cells. 52 A similar phenomenon was observed in Cav1-expressing or silenced MDCK cells treated with Ha-Ras V12 -activated exosomes ( Figure 5). Exosomes carrying Wnt proteins on their surfaces were reported to activate Wnt signalling in target cells. 31 Here, we showed that Cav1 negatively regulates the Wnt receptor Fzd2 F I G U R E 8 The inverse relation between protein levels of Cav1 and Fzd2 was observed in several cancer cell lines. Normal (underlined) and cancer cell lines were cultured on tissue culture dishes overnight and then harvested for Western blot analysis. GAPDH served as a loading control. and thereby confers a protective effect against Ha-Ras V12 /exosomal Wnt5a-induced transformation in MDCK cells. Thus, the levels of Cav1 in the normal cells surrounding tumour are critical for providing tumour-suppressive signals to constrain tumour progression. Cav1 is expressed at high levels in terminally differentiated cells and is often deregulated in cancer and fibrotic diseases. 53 The deregulation of Cav1 via TGFb, a potent fibrogenic cytokine, might disrupt tumoursuppressive signals, thereby promoting tumour progression.
In conclusion, based on the in vitro evidence, we suggest that the presence of Cav1 in recipient cells blocks exosome uptake and its downstream signalling. Cav1 might play an important physiological role in the defence against tumour-derived exosomes via the degradation of Fzd2, thereby suppressing Wnt5a-driven malignant transformation or inhibition of tumour-derived exosomes internalization through an unidentified mechanism. Although the loss of Cav1 is not sufficient to causally drive cell transformation, it is a critical step in the acquisition of the oncogene-induced transformed phenotypes in both tumour cells and normal cells surrounding the tumour.
These findings significantly advance the general understanding of exosome-mediated tumour progression and offer potential strategies for how this pathway may be targeted through the modulation of Cav1 expression.