Introduction
Ubiquitination is a post-translational modification (PTM) based on the addition of the evolutionary conserved protein ubiquitin (Ub) to the lysine residue(s) of the modified protein (Hershko and Ciechanover, 1998). Ubiquitination controls the turnover, activation state, cellular localization, and interactions of target proteins. Undoubtedly, it is a process that has a direct impact on various aspects of cell biology (Rape, 2018). Ubiquitination requires sequential activation of ubiquitin, its transfer to the carrier protein, and subsequent linkage reaction with the substrate lysine residues. This last step, mediated by the E3 ubiquitin protein ligases (E3s), determines target specificity.
Ring Finger protein 43 (RNF43) is a E3 ubiquitin ligase with a single transmembrane domain from the PA-TM-RING family. RNF43 and its close homolog Zinc and Ring Finger 3 (ZNRF3) act as negative regulators of the Wnt/β-catenin signaling pathway (Koo et al., 2012; Hao et al., 2012). Wnt/β-catenin signaling is an evolutionary conserved pathway and a crucial regulator of embryonal development and tissue homeostasis. RNF43 and ZNRF3 control via regulation of Wnt/β-catenin multiple processes including liver zonation (Planas-Paz et al., 2016), limb specification (Szenker-Ravi et al., 2018), and mammalian sex determination (Harris et al., 2018). Mechanistically, RNF43 and ZNRF3 ubiquitinate plasma membrane Wnt receptors called Frizzleds (FZDs) and a co-receptor low-density lipoprotein receptor-related protein 6 (LRP6), which results in their internalization and degradation (Hao et al., 2012; Koo et al., 2012). Therefore, cells become less sensitive or insensitive to Wnt ligands. Activity of RNF43/ZNRF43 is regulated by secreted proteins from R-spondin (RSPO) family (Kazanskaya et al., 2004; Kim et al., 2008; Kim et al., 2006; Kim et al., 2005; Nam et al., 2007; Nam et al., 2006; Peng et al., 2013; Xie et al., 2013) that trigger internationalization of RNF43/ZNRF3 and function as physiologically relevant activators of Wnt/β-catenin pathway (Binnerts et al., 2007; Carmon et al., 2011; de Lau et al., 2011; Hao et al., 2016; Hao et al., 2012; Jiang et al., 2015; Koo et al., 2012; Zebisch et al., 2013; Zebisch and Jones, 2015).
Because deregulation of Wnt/β-catenin pathway promotes tumor formation (Lim and Nusse, 2013; van Kappel and Maurice, 2017; Wiese and Nusse, 2018), RNF43/ZNRF3 can act as tumor suppressors. Indeed, mutation or inactivation of
Some members of the Wnt family – such as WNT5A and WNT11 – preferentially activate downstream signaling that is distinct from Wnt/β-catenin pathway and is referred to as β-catenin-independent or noncanonical Wnt pathway (Pandur et al., 2002; Humphries and Mlodzik, 2018; VanderVorst et al., 2019; Andre et al., 2015). Noncanonical Wnt pathway shares some features with the Wnt/β-catenin pathway – such as the requirement for FZD receptors, dishevelled (DVL) phosphoprotein and casein kinase 1 (CK1) – but clearly differs in others. In the mammalian noncanonical pathway, receptor tyrosine kinase-like orphan receptor 1 (ROR1) and ROR2 act as primary (co-)receptors (in contrast to LRP5/6 that have this role in the Wnt/β-catenin pathway) and four-transmembrane Vang-like protein 1 (VANGL1) and VANGL2 participate in the signal transduction (Asem et al., 2016; VanderVorst et al., 2019). This signaling axis is also referred to as planar cell polarity pathway (PCP), and its activation leads to changes in actin cytoskeleton dynamics, facilitating, that is, polarized cell migration (Andre et al., 2015; Janovská and Bryja, 2017; Kaucká et al., 2015; Weeraratna et al., 2002).
FZD receptors, the best-defined targets of RNF43/ZNRF3, are shared among all Wnt pathways and their endocytosis and/or degradation have the potential, at least in theory, to prevent signaling by any Wnt ligands. So far, there is only one study that suggests the role of RNF43/ZNRF3 in noncanonical Wnt signaling in mammals (Tsukiyama et al., 2015). In addition, secreted inhibitor of RNF43/ZNRF3 called r-spondin 3 (RSPO3) potentiated noncanonical PCP pathway in
In this study, we have directly addressed the role of RNF43 in the WNT5A-induced signaling. We demonstrate that RNF43 controls the noncanonical Wnt pathway similarly to Wnt/β-catenin pathway. Further, we show that RNF43 is a relevant inhibitor of pro-metastatic WNT5A signaling in melanoma where it prevents both WNT5A-induced invasive behavior and WNT5A-assisted development of resistance to B-RAF and MEK inhibitors.
Results
RNF43 inhibits WNT5A-driven noncanonical Wnt signaling pathway
In order to test whether or not RNF43/ZNRF3 controls noncanonical Wnt signaling, we have decided to study T-REx 293 cells. T-REx 293 cells secrete endogenous WNT5A that constitutively activates the noncanonical Wnt pathway – this can be demonstrated by the CRISPR/Cas9-mediated knockout of
Figure 1.
RNF43 interactome is enriched with the Wnt planar cell polarity pathway components.
(A) Western blot analysis of T-REx 293 wild type (WT) and
RNF43 physically interacts with key proteins from the noncanonical Wnt pathway
To address the molecular mechanism of RNF43 action in the noncanonical Wnt pathway, we decided to describe RNF43 interactome by the proximity-dependent biotin identification (BioID; Roux et al., 2012), which was already successfully applied in the challenging identification of E3s substrates (Coyaud et al., 2015; Deshar et al., 2016). We have exploited our recently published dataset (Spit et al., 2020) based on T-REx 293 TetON cells that inducibly expressed RNF43 fused C-terminally (intracellularly) with BirA* biotin ligase. Several core proteins of the noncanonical Wnt signaling pathway – namely, ROR1, ROR2, VANGL1, VANGL2, SEC24B, and all three isoforms of DVL – were strongly and specifically biotinylated by RNF43-BirA* (Figure 1D and D′, Figure 1—source data 1). Furthermore, the noncanonical Wnt pathway was significantly enriched also in the gene ontology (GO) terms (Figure 1—source data 2). Altogether, it suggests that RNF43 can at least transiently interact with multiple proteins involved in the Wnt/PCP pathway, including essential receptor complex components from the ROR, DVL, and VANGL protein families.
To validate the protein-protein interactions identified by BioID, we performed a series of co-immunoprecipitation (co-IP) and co-localization experiments (Figure 2, Figure 2—figure supplement 1). We have focused on the interactions of RNF43 with ROR1/ROR2 and with VANGL1/VANGL2 mainly because these interactions are novel and at the same time highly relevant for the noncanonical Wnt pathway. RNF43 co-immunoprecipitated with both VANGL2 (Figure 2A) and VANGL1 (Figure 2—figure supplement 1A). More detailed analysis of VANGL2 showed co-localization of VANGL2 and RNF43 in the cell membrane (Figure 2B and B′). RNF43 also efficiently pulled down ROR1 (Figure 2C) and ROR2 (Figure 2—figure supplement 1B). Deletion of the cysteine-rich domain (CRD) (ROR2, Figure 2—figure supplement 1B) had no impact on the amount of co-immunoprecipitated RNF43, which suggests that RNF43 primarily interacts with RORs intracellularly. Both ROR1/ROR2 co-localized with RNF43 at the level of the plasma membrane (Figure 2D and D′ and Figure 2—figure supplement 1C and C′). It was described that RORs and VANGLs also bind DVL (Gao et al., 2011; Mentink et al., 2018; Seo et al., 2017; Witte et al., 2010; Yang et al., 2017) and at the same time DVL proteins mediate ubiquitination of FZD receptors by RNF43 in the Wnt/β-catenin pathway (Jiang et al., 2015). To address whether DVL also acts as a physical link between RNF43 and the analyzed PCP proteins, we performed the co-IP experiments with VANGL2 and ROR1 in the T-REx 293 cells lacking all free DVL isoforms (DVL triple knockout cells) (Paclíková et al., 2017). As shown in Figure 2—figure supplement 1D and E, RNF43 was able to bind both VANGL2 and ROR1 as efficiently as in the wild-type cells (compare with Figure 2A and D). In summary, our results indicate that RNF43 interacts, in a DVL-independent way, with PCP proteins from VANGL and ROR families.
Figure 2.
RNF43 interacts with Wnt/planar cell polarity (PCP) components.
(A) RNF43 interacts with VANGL2, but not with its mutants lacking N- or C-termini. VANGL2-EGFP and its variants (schematized) were overexpressed with RNF43-HA in Hek293 T-REx cells, immunoprecipitated by anti-HA and anti-GFP antibodies and analyzed by western blotting. Representative experiment from N = 3. Scheme illustrates secondary structure of the wild-type VANGL2 protein and its shortened variants used in this study. (B, B′) RNF43 (anti-HA, red) co-localized with transiently expressed VANGL2 (GFP, green). Co-localization was analyzed utilizing histograms of red, green, and blue channels signals along selection (yellow line) (B′). TO-PRO-3 Iodide was used to stain nuclei (blue). Scale bar: 25 μm. (C) RNF43 binds to the ROR1 and deletion of the intracellular part of ROR1 disrupts this interaction. RNF43-HA was detected in the ROR1 pull down prepared from lysates of Hek293 T-REx cells overexpressing RNF43-HA and ROR1-V5, N = 3. ROR1 wild-type and truncated mutants are represented in the scheme. (D, D′) RNF43 (anti-HA, red) co-localized with transiently expressed ROR1-V5 (anti-V5, green). Signals along selection (yellow line) were analyzed (D′). TO-PRO-3 was employed nuclei staining (blue). Scale bar: 25 μm. RNF43 interactions with VANGL1 and ROR2 are studied in Figure 2—figure supplement 1.
Figure 2—figure supplement 1.
RNF43 interacts with Wnt/planar cell polarity (PCP) components.
A. RNF43 interacts with VANGL1. VANGL1-Myc was co-immunoprecipitated in the HA pull-down, prepared from lysate of Hek293 T-REx cells transiently overexpressing RNF43-HA and VANGL1-Myc, but not from the lysate containing only VANGL1-Myc overexpressed transgene, N=3. B. RNF43 interacts with the ROR2 in the CRD domain dispensable manner. Wild type ROR2 and ΔCRD-ROR2 mutant were detected in HA and FLAG pull downs, prepared from lysates of the Hek293 T-REx cells transiently overexpressing RNF43-HA and ROR2-FLAG or ΔCRD-ROR2, N=3. C., C’ Exogenous ROR2 (anti-FLAG, green) colocalizes with the RNF43-BirA*-HA (anti-HA, red) in the TetON Hek293 T-Rex cells. DNA was visualized by TO-PRO-3 Iodide. Scale bar represents 25 μm. Co-localization of ROR2 and RNF43 was analyzed utilizing histograms (C’) of red, green and blue channels along selection (yellow line). Data is present in the Figure 2—figure supplement 1—source data 1. D. RNF43 interacts with the VANGL2 in the absence of all three Disheveled isoforms. RNF43 binding to VANGL2 in the DVL1/2/3 deficient cells was confirmed in the two-directional co-IP assay, N=3. E. Interaction between ROR1-V5 and RNF43-HA is preserved in the DVL1-3 null cells. ROR1 was detected in the HA pull-down and RNF43 in the V5 immunoprecipitation. T-REx DVL1/2/3 tKO cells were transfected with highlighted plasmids, N=3.
RNF43 ubiquitinates VANGL2 and triggers its degradation
Since RNF43 is an E3 ubiquitin ligase, we next tested whether it can ubiquitinate its binding partners from the noncanonical Wnt pathway. Enzymatically inactive RNF43 Mut1 variant (Koo et al., 2012) served here as a negative control. Using His-ubiquitin pulldown assay, we were able to show that VANGL2 (Figure 3A), as well as DVL1 and DVL2 (Figure 3—figure supplement 1A), was ubiquitinated when co-expressed with RNF43 but not with RNF43 Mut1. However, we were unable to detect RNF43-induced ubiquitination of ROR1 or ROR2 (negative data, not shown).
Figure 3.
Mechanism of Wnt/planar cell polarity (PCP) inhibition by RNF43.
(A) Hek293 T-REx cells were transfected with plasmid encoding His-tagged ubiquitin, VANGL2-GFP and HA-tagged wild-type or Mut1 RNF43 constructs. Ubiquitinated proteins were enriched by His pull down and analyzed by western blotting. VANGL2 is ubiquitinylated by the E3 ubiquitin ligase RNF43, but not by its enzymatically inactive variant (RNF43Mut1). Representative experiment from N = 3. RNF43-mediated ubiquitination of DVL1 and DVL2 together in Figure 3—figure supplement 1. (B) Tetracycline-induced overexpression of the wt RNF43 (HA), but not enzymatically inactive RNF43Mut1 (HA), decreased VANGL2 protein level and suppressed phosphorylation of ROR1 and DVL3 (empty arrowhead; full arrowhead indicates unphosphorylated protein). CRISPR/Cas9-derived
Figure 3—figure supplement 1.
Mechanism of Wnt/planar cell polarity (PCP) inhibition by RNF43.
(A) DVL1 and DVL2 are ubiquitinylated by the E3 ubiquitin ligase RNF43, but not by its enzymatically inactive mutant (RNF43Mut1). Hek293 T-REx cells were transfected with plasmid encoding His-tagged ubiquitin, DVL1-FLAG or DVL2-FLAG, and wild-type or Mut1 RNF43 constructs and subjected to His-tag pull down and subsequent western blotting. N = 3. (B) Quantification of western blots from Figure 3B. Unpaired two-tailed t-test, *p<0.05, **p<0.01, ***p<0.001, N = 3. (C) Western blotting showing the lack of ROR1 protein in the T-REx 293
Figure 3—figure supplement 2.
Mechanism of Wnt/planar cell polarity (PCP) inhibition by RNF43.
(A) Tetracycline-induced RNF43 (anti-HA, green) co-localized with transiently expressed lysosomal marker (Lamp1-mCherry, red). TO-PRO-3 Iodide was used to stain nuclei (blue). Scale bar: 25 μm. (B) Transiently overexpressed VANGL2 (GFP, green) and lysosomal marker (Lamp1-mCherry, red) in the presence and absence of tetracycline-induced RNF43 (anti-HA, magenta). Hoechst 33342 was used to stain nuclei (blue). Scale bar: 10 μm. (C) Tetracycline-forced expression of RNF43 (anti-HA, magenta) caused retention of VANGL2 (GFP, green) in Golgi complex (anti-endogenous Golgin-97, red). DAPI was used to stain nuclei (blue). Scale bar: 10 μm.
Figure 3—figure supplement 3.
Mechanism of Wnt/planar cell polarity (PCP) inhibition by RNF43.
(A, B) Confocal imaging of the inducible T-REx
Further analysis showed that overexpression of RNF43, but not its E3 ligase dead variant, decreased VANGL2 protein level (Figure 3B, quantified in Figure 3—figure supplement 1B). Decrease in VANGL2 caused by RNF43 was accompanied by impeded phosphorylation of ROR1 (Figure 3B) and DVL3 (Figure 3B, Figure 3—figure supplement 1B). On the other side, two independent clones of cells deficient in both RNF43 and ZNRF3 (
RNF43 induces ROR1 endocytosis by a clathrin-dependent pathway
ROR1 and ROR2 are the key receptors for WNT5A that we found to interact with RNF43 (Figures 1 and 2). We thus speculated that RNF43 can regulate ROR1/ROR2 surface levels. T-REx cells express dominantly ROR1, and indeed flow cytometric analysis demonstrated that cell lacking endogenous RNF43 and ZNRF3 have more ROR1 receptor on the surface than parental T-REx cells (Figure 3D). The staining is specific as demonstrated by the validation of the ROR1-APC antibody in
In our analysis of RNF43 interactors (Figure 1D), we identified also multiple proteins involved in endosomal transport. It included proteins involved in the clathrin endocytic pathway – STAM1, HRS, ZFYVE16, PICALM, NUMB, RAB11-FIP2, and subunits of the associated adaptor protein complexes AP-3 and AP-4 (Supplementary file 1; Bache et al., 2003; Cullis et al., 2002; Hirst et al., 2013; Raiborg et al., 2001; Santolini et al., 2000; Seet and Hong, 2005; Tebar et al., 1999). Based on the BioID results analysis, we speculated that RNF43 may promote clathrin-mediated endocytosis of ROR1. Thus, we applied dansylcadaverine to block this pathway (Blitzer and Nusse, 2006). In agreement with our hypothesis, treatment with this inhibitor prevented RNF43-mediated effect on the ROR1 surface expression in T-REx 293 RNF43/ZNRF3 dKO RNF43 TetON cells (Figure 3F).
To get a better insight into the mechanism of RNF43-induced internalization of ROR1, we analyzed the co-localization of ROR1 and RAB5 (marker of early endosomes) and RAB11 (marker of recycling endosomes) in T-REx 293 R/Z dKO RNF43 TetON (Figure 3G and Figure 3—figure supplement 3A) and T-REx 293 RNF43 TetON cells (Figure 3—figure supplement 1E and Figure 3—figure supplement 3B). Hyperactivation of Rab5 by overexpression of wild-type Rab5 leads to the formation of giant early endosomes (Bucci et al., 1992) where we observed ROR1/RAB5 co-localization after 3 hr of tetracycline treatment. The co-localization decreased after overnight exposition to tetracycline. RAB11+ endosomes were recruited to the ROR1 as well as after RNF43 induction and RAB11 co-localized strongly with ROR1 even after ON treatment. We conclude that surface ROR1 is controlled by RNF43 via interference with RAB5- and RAB11-mediated endocytosis and RNF43-mediated effect is not persistent due to the activity of recycling RAB11 recycling pathway (Ullrich et al., 1996).
RNF43 expression is decreased in human melanoma
Our data shown in Figures 1—3 demonstrate that RNF43 can inhibit WNT5A-induced noncanonical signaling via downregulation of the receptor complexes. But is RNF43 capable of blocking WNT5A-induced biological processes? WNT5A signaling plays a crucial role in melanoma, one of the most malignant tumor types. High expression of
Interestingly, the in silico analysis of gene expression (Talantov et al., 2005; Xu et al., 2008) showed that
Figure 4.
RNF43 in melanoma.
(A, B)
Figure 4—figure supplement 1.
RNF43 in melanoma.
(A)
Figure 4—figure supplement 2.
RNF43 in melanoma.
(A–D) RT-qPCR results – expression of the
Figure 4—figure supplement 3.
RNF43 in melanoma.
(A-A’’, B-B’’) Effects of the inducible RNF43 overexpression in RAS-mutant MelJuso (A) and BRAF V600E A375 (B) cells. Exogenous RNF43 expression blocked response to the 40 and 80 ng/ml 3 hr-long rhWNT5A treatments. Porcupine inhibitor LGK-974 was used to block endogenous Wnt ligands secretion and RNF43 was probed by HA antibody. β-actin is shown as a loading control. Phosphorylation status and total protein level of DVL2 (A’ B’) and DVL3 (A’’, B’’) were quantified, unpaired t-test *p<0.05, **p<0.01, ***p<0.001, N = 3. Results were normalized to 1, for the ‘DOX-’ condition. (C) Western blot analysis of control to A375 RNF43 TetON cell line. Cells were treated with doxycycline and Porcupine inhibitor LGK-974 for cells validation. Figure 4—figure supplement 3—source data 1 file presents raw data. Empty arrowhead: phosphorylated; full arrowhead: unphosphorylated protein.
RNF43 inhibits invasive properties of melanoma cells in vitro
A375 and A2058 are human melanoma cell lines carrying BRAF V600E mutations that are broadly used to study WNT5A role in melanoma (Anastas et al., 2014; Connacher et al., 2017; Da Forno et al., 2008; Ekström et al., 2014; Linnskog et al., 2016; Liu et al., 2018). For the purpose of our study, we chose A375 wild-type (A375) cells and their derivate with the increased metastatic potential referred to as A375 IV (Kucerova et al., 2014). Both A375 variants and A2058 express
To study the RNF43 function, we generated A375 cells lacking
WNT5A signaling has been related to numerous biological features that support the invasive properties of melanoma (Arozarena and Wellbrock, 2017b; O’Connell and Weeraratna, 2009; Prasad et al., 2015; Weeraratna et al., 2002). To address if RNF43 affects any of these WNT5A-controlled properties, we have compared parental and RNF43-derivative melanoma cells in a panel of functional assays that included (1) wound healing assay, (2) collagen I hydrogel 3D chemotaxis assay, (3) Matrigel invasion assay, (4) invadopodia formation assay, and (5) gelatin degradation assay. Firstly, all A375, A375 IV, and A2078 cells overexpressing RNF43 showed suppressed 2D collective migration in the wound healing assay (Figure 5A–E). Next, we analyzed the impact of RNF43 expression on directional invasion through the collagen I hydrogel in response to chemokines. This assay mimics the taxis of melanoma cells in body: CCL21 drives lymph nodes metastasis and CXCL12 promotes lung invasion (Figure 5F; Jacquelot et al., 2018; McArdle et al., 2016). A375 IV and A2058 cell lines showed significant response to these treatments and RNF43 blocked completely these invasion events (Figure 5G and H). Importantly, the response of A375 cells in this assay was negligible (Figure 5—figure supplement 1B), confirming their decreased metastatic capacity compared to the A375 IV. In line, invasion of individual cells through the extracellular matrix (ECM) mimicking Matrigel was higher in A375 IV in comparison to A375 but in both cases reduced by RNF43 OE (Figure 5—figure supplement 1D). The same is true for the number of invadopodia-specialized structures mediating adhesion and remodeling of the surrounding ECM (Eddy et al., 2017; Masi et al., 2020). A375 IV cells formed more invadopodia than A375 parental cells (Figure 5I) and cells overexpressing RNF43 formed less of them (Figure 5I). In agreement, we also observed reduced gelatin degradation activity in A375 and A375 IV cells overexpressing RNF43 (Figure 5J). Furthermore, treatment with WNT5A enhanced the gelatin degradation capacity of A375 cells, but not their RNF43-overexpressing derivate (Figure 5J). Representative images from the conducted assays are shown in Figure 5—figure supplement 1, Figure 5—figure supplement 2, Figure 5—figure supplement 3, Figure 5—figure supplement 4. All these assays strongly support the conclusion that RNF43 acts as a strong molecular inhibitor of WNT5A-triggered proinvasive features of melanoma. Interestingly,
Figure 5.
RNF43 inhibits WNT5A-dependent invasive properties of human melanoma.
(A–E) RNF43 reduced migration of A375 (B), A375 RNF43 TetON (C), A375 IV (D), and A2058 RNF43 TetON (E) in the wound healing assay. Wound was photographed 48 hr after scratch and presented as % of cell-free surface at the end of the experiment. Cells proliferation was suppressed by serum starvation, unpaired two-tailed t-test: *p<0.05, **p<0.01, ***p<0.001, N = 3 (B, D) or 4 (C, E). Representative photos at the end of the experiment are shown in (A) and in Figure 5—figure supplement 1A. (F–H) RNF43 blocked collagen I hydrogel 3D invasion in response to CCL21 (100 ng/ml) and CXCL12 (100 ng/ml) of A375 IV (G) and A2058 (H) cell lines. Cells were serum starved, collagen I (1.5 mg/ml) was overlaid and polymerized. Doxycycline was applied for RNF43 induction during starvation. After 24 hr, cells were fixed and stained for DNA (Hoechst 33342, blue) and F-actin (phalloidin, red) and imaged by confocal microscopy. Invasion index was calculated as the ratio of invaded cells at specified height to the number of noninvasive cells at the glass level, unpaired two-tailed t-test: **p<0.01, ****p<0.0001, N = 3 (G) or N = 4 (H). A375 cells did not invaded collagen I hydrogel (Figure 5—figure supplement 1B). Representative photos are presented in Figure 5—figure supplement 1C and C′. (I) RNF43 overexpression in A375 and A375 IV decreased number of invadopodia. Quantification of the invadopodia formed by melanoma cells, based on the analysis of confocal images. Number of cortactin/F-actin double-positive puncta in the individual cells was calculated in the ImageJ software, unpaired two-tailed t-test: ***p<0.001, ****p<0.0001. Examples of confocal imaging are shown: green, phalloidin; red, cortactin; blue, DNA. See Figure 5—figure supplement 2 for images from all experimental conditions. (J) Gelatin degradation assay; both A375 and A375 IV RNF43-overexpressing cell lines showed decreased capacity to locally degrade the extracellular matrix. rhWNT5A treatment induced gelatin degradation by A375 cells. Serum-starved cells were plated onto gelatin-Oregon Green-coated coverslips and incubated for 24 hr. Images obtained by Leica SP8 confocal microscope were analyzed for the presence of gelatin degradation by individual cells using ImageJ software, unpaired two-tailed t-test: *p<0.05, **p<0.01, N = 3. Example of gelatin degradation is shown; more pictures are presented in Figure 5—figure supplement 3 and Figure 5—figure supplement 4. Numerical data are given in Figure 5—source data 1 file.
Figure 5—figure supplement 1.
RNF43 inhibits Wnt5a-dependent invasive properties of human melanoma.
(A) Wound healing experiment – representative photos at the experimental end point (data Figure 5B–E). (B) Collagen I hydrogel chemotaxis assay A375 cell line results. After 24 hr, cells did not significantly respond to the treatments, unpaired two-tailed t-test: p>0.05, N = 3. (C, C′) Representative confocal photos of collagen I hydrogel chemotaxis assay for A375 IV (C) and A2058 (C′) cells (data presented in Figure 5G and H). (D, D′) Matrigel invasion assay – stable RNF43 overexpression-inhibited invasive properties of the A375 and A375 IV. Serum-starved cells were plated onto Matrigel-coated porous membrane. Medium containing 20% of serum was used as chemoattractant. After 18 hr of incubation, cells were fixed in methanol, noninvaded ones were removed from the upper part of transwell insert by cotton swab. Results were normalized to 1 for the number of invaded A375 cells, unpaired two-tailed t-test: *p<0.05, N = 4. (D′) Representative photos assay after crystal violet staining. Figure 5—figure supplement 1—source data 1 contains raw data.
Figure 5—figure supplement 2.
RNF43 inhibits Wnt5a-dependent invasive properties of human melanoma.
Confocal imaging of A375, A375 IV, and RNF43 overexpressing and
Figure 5—figure supplement 3.
RNF43 inhibits Wnt5a-dependent invasive properties of human melanoma.
Confocal imaging of gelatin degradation assay without (Figure 5—figure supplement 3) and after (Figure 5—figure supplement 4) rhWNT5A treatment. Serum-starved cells were plated onto gelatin-Oregon Green (green)-coated coverslips and incubated for 24 hr. Fixed cells were stained with phalloidin-Alexa 594 for F-actin visualization (red) and TO-PRO-3 Iodide for nuclei (blue). Foci showing gelatin degradation are marked. Scale bars represent 50 μm. Experiment was repeated three times. Results are presented in Figure 5J.
Figure 5—figure supplement 4.
RNF43 inhibits Wnt5a-dependent invasive properties of human melanoma.
Confocal imaging of gelatin degradation assay after rhWNT5A treatment. Serum-starved cells were plated onto gelatin-Oregon Green (green)-coated coverslips and incubated for 24 hr. Fixed cells were stained with phalloidin-Alexa 594 for F-actin visualization (red) and TO-PRO-3 Iodide for nuclei (blue). Foci showing gelatin degradation are marked. Scale bars represent 50 μm. Experiment was repeated three times. Results are presented in Figure 5J.
RNF43 prevents acquisition of resistance to BRAF V600E targeted therapy
The mitogen-activated protein kinase (MAPK) pathway is hyperactivated in melanoma (Davies et al., 2002) as a result of UV-induced mutations triggering constitutive activation of this signaling axis. The most common genetic aberration –
The process of vemurafenib resistance acquisition can be modeled in vitro. We applied an experimental scheme optimized for A375 (Anastas et al., 2014). This model (Figure 6A) allows to study both acute responses to vemurafenib (24 hr treatment) as well as the gradual adaptation of long-term cell culture to increasing vemurafenib doses. Vemurafenib-resistant (VR) cells can be obtained after approximately 2 months. As shown in Figure 6B, treatment with vemurafenib resulted in rapid and complete inhibition of ERK1/2 phosphorylation, the readout of MAPK activation (compare lanes 1 and 2). In contrast, A375 VR cells showed constitutive ERK1/2 phosphorylation in the presence of 2 μM vemurafenib (compare lanes 2 and 3). Interestingly, transient exposition to vemurafenib resulted in the impeded phosphorylation of ROR1, DVL2, and DVL3 and increased ROR2 protein level without change in its mRNA (Figure 6B–D, Figure 6—figure supplement 1A and B). On the other side, VR cells displayed elevated ROR1 levels (both protein and mRNA) and increased phosphorylation of DVL2 and DVL3 (Figure 6B–D, Figure 6—figure supplement 1A and B). Strikingly, the expression of
Figure 6.
RNF43-overexpressing melanoma cells do not develop resistance to BRAF V600E targeted therapies.
(A) Scheme showing the experimental model used for the analysis of vemurafenib resistance (VR) acquisition. Melanoma cells are exposed to the increasing doses of the
Figure 6—figure supplement 1.
RNF43-overexpressing melanoma cells do not develop resistance to BRAF V600E targeted therapies.
(A, B) Protein levels of ROR1 (A) and ROR2 (B) (western blot shown in Figure 6B) were quantified using ImageJ software, two-tailed t-test: *p<0.05, **p<0.01; N = 3. Expression ROR1 (A′) and ROR2 (B′) genes in parental, cells transiently treated with BRAFi and vermurafenib-resistant A375 cells and presented as 2−ΔΔCt ± SD. Relative expression level was normalized to the HSPCB and RPS13. (C, C′) Western blot analysis of 24 hr and chronic response to vemurafenib, cobimetinib, and their combination in A2058 cells, continuation of Figure 6H. Empty arrowhead represents phosphorylation-dependent shift; full arrowhead indicates unphosphorylated status. Scheme of the experiment is presented in (C′). (D–F) Expression of
To strengthen our results, we exposed A375 and A2058 parental and RNF43-expressing cells temporarily and chronically to vemurafenib, MEK inhibitor – cobimetinib and to the FDA-approved combination of these drugs (Ascierto et al., 2016; Signorelli and Shah Gandhi, 2017; experimental scheme in Figure 6—figure supplement 1C’). We failed to obtain A375-resistant line, but A2058 acquired resistance to cobimetinib alone and its combination with vemurafenib (Figure 6H and L). Treatments and resistance were validated by pERK1/2 levels (Figure 6H). A2058 enriched with ectopically expressed RNF43 did not survive cobimetinib monotherapy and cobimetinib with vemurafenib combination, whereas vemurafenib-only treatment did not fully eliminate those cells (Figure 6L). Thus, we aimed to detect the differences in mechanisms behind targeted therapies resistance acquisition to characterize better the function of RNF43 in melanoma. Firstly, we noticed the shift of VANGL1 in all delivered resistance models, while VANGL2 was shifted only in cells being in the long-term culture with cobimetinib and with cobimetinib and vemurafenib (Figure 6H). This corresponds with decreased
Altogether, these data confirm earlier findings on the importance of WNT5A signaling in the acquisition of resistance to targeted therapies and demonstrate that RNF43 can block this process. Moreover, we show that RNF43 could be a negative regulator of melanoma phenotype plasticity by targeting MITF-low/WNT5A-high melanoma cells (Hoek et al., 2006; Kim et al., 2017; Sensi et al., 2011; Tirosh et al., 2016).
RNF43 as onco-suppressor in vivo: impact on tumors and resistance to vemurafenib
Next, we aimed to confirm our results also in the in vivo model. We decided to use cell lines with the same origin but varying in the
Figure 7.
RNF43 inhibits melanoma proliferation and response to vemurafenib in vivo
(A) RT-qPCR results – expression of the
Figure 7—figure supplement 1.
RNF43 inhibits melanoma proliferation and response to vemurafenib in vivo.
(A) Vemurafenib formulations test for the in vivo application. Vemurafenib 2.5 mg/ml was prepared in aqueous solutions of 1% carboxymethyl cellulose (CMC), 10, 20, and 25% Kolliphor. DMSO was used as reference. Vemurafenib was added to solutions from DMSO-based saturated stock. Formulation in 25% Kolliphor was selected for further tests. (A′) A375 cells were treated overnight with 1 μM vemurafenib in DMSO and in 25% Kolliphor formulation, lysed and phosphorylated ERK1/2 signal was used as readout for vemurafenib inhibitory properties. (B) Tumors sizes formed by A375 RNF43 TetON cells (RNF43 mid/high after following doxycycline supplementation) in cohorts II–V at the beginning of treatments. (C) Tumor growth curve – caliper measurements of tumors sizes within experimental groups I–III. (D) Western blot analysis of primary tumors at the experimental end points. Tissues were homogenized, lysed in the 1% SDS lysis buffer, sonicated and clarified by centrifugation, followed by addition of Laemmli buffer and thermal denaturation. Each sample contains 25 µg of total protein. RNF43 was probed by HA tag-specific antibody. Empty arrowhead represents phosphorylation-dependent shift; black arrowhead indicates unphosphorylated status. (E) Quantifications of tumor samples western blots. Tumors formed by RNF43 low cells had scientifically higher ROR1 (E) and VANGL2 levels (E′) and higher phosphorylation of DVL2 (E′′) than RNF43 mid-derived tumors. RNF43 high vemurafenib-resistant cells had higher ROR1 protein level than in the RNF43 mid cohort (E). Vemurafenib treatment led to the increased WNT5A protein level in tumors at the end of the experiment (E′′′). RNF43 low N = 5; RNF43 mid N = 3; RNF43 high N = 4; RNF43 mid + vemurafenib N = 5; RNF43 high + vemurafenib N = 4, unpaired two-tailed t-test *p<0.05, **p<0.01. All used data are presented in Figure 7—figure supplement 1—source data 1.
Further, we also investigated the synergistic effect of RNF43 and vemurafenib in vivo. RNF43 mid and RNF43 high cohorts received vemurafenib, and this treatment significantly prolonged the survival (Figure 7F), validating our formulation and dosing efficacy. Importantly, doxycycline supplementation (
Figure 8.
RNF43 inhibits WNT5A-driven signaling and suppresses melanoma invasion and resistance to the targeted therapy.
Graphical summary. RNF43 is an inhibitor of the noncanonical WNT5A-induced pathway. RNF43 interacts with receptor complexes of the Wnt/PCP signaling and its enzymatic activity results in the reduced cells sensitivity to WNT5A (A). In melanoma, WNT5A promotes invasion and metastasis (B) as well as resistance to targeted therapies, including treatments with vemurafenib – inhibitor of commonly mutated BRAF kinase and cobimetinib-targeting activity of the MEK enzyme (C). RNF43 blocks melanoma-invasive properties via interference with the nonconical Wnt pathway, leading to the increased sensitivity to treatment (D).
Discussion
Our study identified RNF43 as an inhibitor of noncanonical WNT5A-induced signaling. RNF43 physically interacted with multiple receptor components of the Wnt/PCP pathway such as ROR1/2, VANGL1/2, or DVL1/2/3 and triggered degradation of VANGL2 and membrane clearance of ROR1, ultimately resulting in the reduced cell sensitivity to WNT5A. The newly discovered RNF43 action in WNT5A-mediated signaling seems to be mechanistically different than the well-known function in the Wnt/β-catenin pathway. For example, we observed ROR1 and VANGL2 interaction with RNF43 in the absence of DVL. In contrast, DVL seems to be essential for the activity of RNF43 in the Wnt/β-catenin pathway (Jiang et al., 2015). Further, the inhibitory action of RNF43 in WNT5A signaling could not be blocked by inhibition of lysosomal pathway, in contrast to the earlier observations in WNT/β-catenin pathway (Koo et al., 2012). On the other side, WNT5A signaling can be, similarly to Wnt/β-catenin, promoted by RNF43 inhibitors from R-SPO family. Also, in line with the earlier findings that RNF43 leads to the packing of ubiquitinated FZD to the RAB5+ endosomes (Koo et al., 2012), ROR1 is as well internalized via a clathrin-dependent mechanism into RAB5+ and RAB11+ endosomes. Interestingly, ROR1 internalization is transient. We can speculate that it represents the first step in the silencing of the noncanonical Wnt pathway, which further translates into stable inhibition, for example, via VANGL2 sequestration. It remains to be studied how RNF43 in a coordinated manner controls both WNT/β-catenin and noncanonical WNT pathways.
We demonstrate that the newly characterized RNF43-WNT5A regulatory module controls WNT5A signaling and biology in melanoma. WNT5A-induced signaling plays a crucial role in this cancer type. Up to date, 5-year survival of metastatic melanoma patients rates between 5 and 19%, depending on the location and the number of metastases (Sandru et al., 2014). Elevated expression of
Previous studies showed that melanoma displays remarkable phenotypic plasticity upon targeted therapy (Hoek et al., 2006; Hoek and Goding, 2010; Kemper et al., 2014; Rambow et al., 2018). This is also well demonstrated in our A2058 model where vemurafenib-resistant cells become melanotic (MITFhigh), whereas cobimetinib- and especially vemurafenib/cobimetinib-double-resistant cells have MITFlow/WNT5Ahigh phenotype that is characteristic for highly invasive melanoma with the dedifferentiated phenotype (Ahn et al., 2017; Anastas et al., 2014; Arozarena and Wellbrock, 2019; Massi et al., 2020; Webster et al., 2015). Inhibition of the WNT5A pathway by RNF43 could block one of the trajectories of resistance acquisition by Darwinian selection of preexisting subpopulation (Chisholm et al., 2015), thereby promoting less metastatic phenotype (Bai et al., 2019). This is supported by our observation that vemurafenib-treated A2058 cells can partially overcome RNF43 OE by phenotypic switch to MITFhigh phenotype, which is not the case in cobimetinib-treated cells. From the other angle, the low
The relevance of our findings is likely not limited to melanoma. Signaling cascade RSPO–LGR4/5–RNRF43/ZNRF3 has been shown to regulate a variety of biological processes. In light of our results, it is tempting to speculate that WNT5A-RNF43 axis regulates other developmental, physiological, and pathophysiological conditions. For example,
Further exciting avenues relate to the importance of RSPO-RNF43/ZNRF3 module in the regulation of multiple developmental processes dependent on WNT5A. There are literature hints that suggest that indeed WNT5A signaling is fine-tuned by RNF43/ZNRF3 during convergent extension movements. The regulation of Rspo3 has been proven in
Materials and methods
1. Cell lines and treatments
T-REx-293 (R71007, Thermo Fisher Scientific), GFP labeled human melanoma A375 wild-type (A375) and its metastatic derivates A375 IV (Kucerova et al., 2014), A2058 (ECACC 91100402), and MelJuso (Štětková et al., 2020) (kindly gifted by Stjepan Uldrijan) cell lines were propagated in the Dulbecco’s modified Eagle’s medium (DMEM, 41966-029, Gibco, Life Technologies) supplemented with the 10% fetal bovine serum (FBS, 10270-106, Gibco, Life Technologies), 2 mM L-glutamine (25030024, Life Technologies), 1% penicillin-streptomycin (XC-A4122/100, Biosera) under 5% (vol/vol) CO2 controlled atmosphere at 37°C. Routine checks for mycoplasma contamination were performed. For inhibition of endogenous Wnt ligands, cells were treated with the 0.5 μM Porcupine inhibitors C-59 (ab142216, Abcam) or 1 μM LGK-974 (1241454, PeproTech). The time points and doses have been chosen based on the purpose of the experiment. Changes in the phosphorylation of the WNT5A-downstream proteins in the noncanonical Wnt pathway transduction have been analyzed after 3 hr with the recombinant human WNT5A (645-WN, R&D Systems) in doses 40–100 ng/ml as given in the figure legends. Longer stimulation (overnight and longer) has been used in the functional experiments. For canonical Wnt signaling activation, the recombinant human WNT3A (5036-WN, R&D Systems) was used overnight in 40 ng/ml, 60 ng/ml, or 80 ng/ml concentrations. Co-treatment with the recombinant human R-Sponidin-1 (120-38, PeproTech) in 50 ng/ml dose was applied where indicated. Dansylcadaverine (D4008, Sigma-Aldrich) 50 µM treatment along with 3 hr tetracycline was applied to block clathrin-dependent endocytosis pathway (Blitzer and Nusse, 2006).
For preparation of stable cell lines, antibiotic selection after plasmid DNA transfection was performed using 5 μg/ml blasticidin S (3513-03-9, Santa Cruz Biotechnology) or 200 μg/ml of hygromycin B (31282-04-9, Santa Cruz Biotechnology) for T-REx-293 cells and accordingly 400 μg/ml and 5 μg/ml in case of A375 melanoma cell line. As a result, tetracycline-inducible T-REx-293 RNF43 and RNF43 Mut1 TetON, T-REx-293
T-REx-293
2. Plasmids/cloning
Backbone of the plasmid pcDNA4-TO-RNF43-2xHA-2xFLAG (kindly gifted by Bon-Kyoung Koo together with pcDNA4-TO-RNF43Mut1-2xHA-2xFLAG; Koo et al., 2012) was used for further cloning. Briefly, for generation of the BioID-inducible pcDNA4-TO-RNF43-BirA*-HA plasmid, cDNA encoding RNF43 without stop codon was amplified by the PCR and cloned into the pcDNA3.1 MCS-BirA(R118G)-HA (Addgene plasmid #36047; RRID:Addgene_36047) using HpaI (ER1031, Thermo Fisher Scientific) and EcoRI (ER0271, Thermo Fisher Scientific) restriction enzymes to fuse it in frame with the BirA*-HA sequence. Then, RNF43-BirA*-HA cDNA was amplified and cloned by the In-Fusion cloning method (639690, Takara Bio) into linearized by HindIII (ER0501, Thermo Fisher Scientific) and XbaI (ER0681, Thermo Fisher Scientific) pcDNA4-TO plasmid. To eliminate BirA* enzyme-mediated potential false-positive results, pcDNA3-RNF43-HA was prepared by subcloning RNF43 PCR product containing HA encoding sequence in the reverse primer to the pcDNA3 backbone (Invitrogen). pCW57-RNF43 (blast) plasmid was obtained by RNF43-HA-FLAG cDNA cloning into the EcoRI site of the pCW57-MCS1-P2A-MCS2 (Blast) backbone (Addgene #80921; RRID:Addgene_80921) by the In-Fusion cloning method. All obtained plasmids were verified by the Sanger sequencing method.
Other plasmids used were described previously and included myc-Vangl1, GFP-Vangl2, GFP-Vangl2ΔN, GFP-Vangl2ΔC, GFP-Vangl2ΔNΔC (Belotti et al., 2012), pLAMP1-mCherry (Addgene #45147), pEGFP-C1-Rab5a (Chen et al., 2009), GFP-rab11 WT (Addgene #12674), His-ubiquitin (Tauriello et al., 2010), pcDNA3-Flag-mDvl1 (Tauriello et al., 2010), pCMV5-3xFlag Dvl2 (Addgene #24802), pCDNA3.1-Flag-hDvl3 (Angers et al., 2006), pcDNA3.1-hROR1-V5-His (gifted by Kateřina Tmějová), pcDNA3-Ror2-Flag and pcDNA3-Ror2-dCRD-FLAG (Sammar et al., 2004), pRRL2_ROR1ΔCYTO and pRRL2_ROR1ΔTail (Gentile et al., 2011), hCas9 (Addgene #41815), gRNA_GFP-T1 (Addgene #41819), and PiggyBack-Hygro and Transposase coding plasmids (gifted by Bon-Kyoung Koo). Sequences of primers used for cloning are presented in Table 1.
Table 1.
Cloning and mutagenesis primers.
| Primer | Sequence | Purpose |
|---|---|---|
| RNF43 BirA*F |
| RNF43 cDNA cloning into pcDNA3.1 MCS-BirA(R118G)-HA |
| RNF43 BirA*R |
| |
| RNF43 InFusion F |
| RNF43-BirA(R118G)-HA into pcDNA4 |
| RNF43 InFusion R |
| |
| RNF43-HA F |
| RNF43-HA cloning into pcDNA3 |
| RNF43-HA R |
| |
| pCW57-RNF43 InFusion F |
| pCW57-RNF43 generation |
| pCW57-RNF43 InFusion R |
|
3. CRISPR/Cas9
For targeting
For targeting
4. RNF43 BioID analysis
Data are available via ProteomeXchange (Deutsch et al., 2020) with identifier PXD020478 in the PRIDE database (Perez-Riverol et al., 2019). The analysis of the mass spectrometric RAW data files was carried out using the MaxQuant software (version 1.6.2.10) using default settings unless otherwise noted. MS/MS ion searches were done against modified cRAP database (based on http://www.thegpm.org/crap) containing protein contaminants like keratin, trypsin, etc., and UniProtKB protein database for
5. Transfection
T-REx-293 cells were transected using 1 μg/ml, pH 7.4 PEI, and plasmid DNA in a 4:1 ratio (Paclíková et al., 2017). Plasmid DNA were in amount of 3 µg for 6 cm culture dish (ubiquitination assay) and 6 µg for 10 cm dish (co-immunoprecipitation or stable cell lines preparation). Approximately 1 × 106 of A375 and A375 IV cells were electroporated with 6 μg of plasmid DNA utilizing Neon Transfection System (Thermo Fisher Scientific) 1200 V, 40 ms, 1 pulse. Culture media were changed 6 hr post-transfection.
6. His-ubiquitin pulldown assay
Cells were transfected with the plasmid encoding polyhistidine-tagged ubiquitin, RNF43-HA, or enzymatically inactive RNF43, protein of interest, and cultured overnight. Next, cells were treated with 0.2 µM epoxomicin (E3652, Sigma) for 4 hr and lysed in the buffer containing 6 M guanidine hydrochloride (G3272, Sigma), 0.1 M NaxHxPO4 pH 8.0, and 10 mM imidazole (I5513, Sigma), sonicated, and boiled. Insoluble fraction was removed by the centrifugation (16,000 g, room temperature [RT], 10 min). For the pull down of tagged proteins, 10 µl of equilibrated in lysis buffer His Mag Sepharose beads Ni (GE28-9799-17, GE Healthcare) was added to each sample and kept on a roller overnight. Then, the beads were washed three times in the buffer containing 8 M urea (U5378, Sigma), 0.1 M NaxHxPO4 pH 6.3, 0.01 M Tris, and 15 mM imidazole, resuspended in 100 μl of western blot sample buffer, boiled for 5 min, and loaded onto SDS-PAGE gel. Approximately 10% of cellular lysate was used as a transfection control after ethanol precipitation and resuspension in the western blot sample buffer.
7. Western blotting and antibodies
Western blot analysis was performed as described before using samples with the same protein amount, measured by the DC Protein Assay (5000111, Bio-Rad), or lysed directly in the sample buffer (2% SDS, 10% glycerol, 5% β-mercaptoethanol, 0.002% bromophenol blue, and 0.06 M Tris HCl, pH 6.8) and protease inhibitor cocktail (11836145001, Roche) after PBS wash (Mentink et al., 2018). Protein extraction from mouse tissues was done by homogenizing in the 1% SDS, 100 mM NaCl, 100 mM Tris, pH 7.4 buffer, sonication, clarification by centrifugation (16,000 g, 4°C, 15 min), and protein concentration measurement. Next, volumes of samples containing the same protein amounts were mixed with western blot sampling buffer and loaded onto SDS-PAGE gels. Briefly, after electrophoretic separation, proteins were transferred onto Immobilon-P PVDF Membrane (IPVH00010, Millipore) and detected using primary and corresponding HRP-conjugated secondary antibodies on Fusion SL imaging system (Vibler) using Immobilon Western Chemiluminescent HRP Substrate (Merck, WBKLS0500). Molecular size of the bands is marked in each panel (kDa). A list of used antibodies is presented in Appendix 1—key resources table. Shifts of ROR1, ROR2, DVL2, DVL3, VANGL1, and VANGL2 are marked by arrowheads. Empty arrowhead marks phosphorylation-dependent shift. Densitometric analysis of western blot signals was performed using ImageJ software. Activation level of DVL2 and DVL3 (Bryja et al., 2007a) is presented as the ratio of intensities of upper band; representing active – phosphorylated protein fraction and lower band (black arrowhead; unphosphorylated). Total DVL2 and DVL3 levels were quantified as the sum of two bands intensities.
8. Immunofluorescence and confocal microscopy
Cells growing on the glass were fixed in 4% paraformaldehyde (PFA) in PBS. Fixed cells were permeabilized by 0.1% Triton X-100 in PBS and blocked in 1% solution of bovine serum albumin (BSA) in PBS. Then, samples were incubated overnight at 4°C with primary antibodies diluted in 1% BSA in PBS and washed. Corresponding Alexa Fluor secondary antibodies (Invitrogen) were incubated with samples for 1 hr at RT, along with 1 µg/ml Hoechst 33342 (H1399, Thermo Fisher Scientific) for nuclei staining. After PBS washes, the samples were mounted in the DAKO mounting medium (S3023, DAKO). Images were taken on the confocal laser scanning microscopy platform Leica TCS SP8 (Leica). For co-localization analysis, histograms for each channel were prepared in LAS X Life Science (Leica) software and plotted in GraphPad Prism 8. Co-localization is marked by arrowheads.
9. Immunoprecipitation
T-REx-293 cells were transfected with the proper plasmid DNA and cultured for 24 hr. Then, cells were washed two times with PBS and lysed for 15 min in the buffer containing 50 mM Tris pH7.6, 200 mM NaCl, 1 mM EDTA, 0.5% NP40, fresh 0.1 mM DTT (E3876, Sigma) and protease inhibitor cocktail (04693159001, Roche). Insoluble fraction was removed by centrifugation (16,000 g, RT, 15 min), 10% of total cell lysate was kept as western blot control. Lysates were incubated with 1 μg of antibody for 16 hr at 4°C on the head-over-tail rotator. Next, 20 μl of protein G-Sepharose beads (17-0618-05; GE Healthcare) equilibrated in complete lysis buffer were added to each sample and incubated for 4 hr at 4°C, following six washes using lysis buffer and resuspension in 100 μl of western blot sample buffer. Immunoprecipitation experiments were analyzed by the western blot.
10. Flow cytometric determination of ROR1 surface expression
Determination of the ROR1 surface expression of T-REx-293 and its derivates was done using anti-ROR1-APC (#130-119-860, Miltenyi Biotec) and Accuri C6 (BD Biosciences) (
11. Quantitative polymerase chain reaction (qPCR)
Messenger RNA was isolated using the RNeasy Mini Kit (74106; Qiagen) according to the manufacturer’s instructions. 1 μg of mRNA was transcribed to cDNA by the RevertAid Reverse Transcriptase (EP0442, Thermo Fisher Scientific) and analyzed by use of LightCycler 480 SYBR Green I Master (04887352001, Roche) and LightCycler LC480 (Roche). Results are presented as 2−ΔΔ
12. Databases
RNF43, VANGL1, and DVL3 gene expression in different melanoma stages was analyzed through Oncomine (RRID:SCR_007834; Rhodes et al., 2004) database in the different datasets (Talantov et al., 2005, Xu et al., 2008, Haqq et al., 2005). OncoLnc (Anaya, 2016) database was employed to elucidate whether the expression of the
13. Wound healing assay, Matrigel invasion assay, fluorescent gelatin degradation assay, invadopodia formation assay, and collagen I hydrogel 3D invasion assay
For the determination of cellular motility and invasive properties in vitro wound healing (O’Connell et al., 2008), Matrigel invasion towards 20% FBS as chemoattractant followed by crystal violet staining of invaded cells, fluorescent gelatin degradation in the presence of 5% FBS after overnight starvation, and invadopodia formation assays were prepared according to the established protocols (Makowiecka et al., 2016). The wound gap was photographed using the Olympus ix51 inverted fluorescence microscope after 48 hr from scratch. Percentage of the cell-free surface was measured by ImageJ software. For the fluorescent gelatin degradation assay purpose, 80 ng/ml of rhWNT5A was used during 16 hr of cells’ incubation on the coverslips coated with gelatin-Oregon Green conjugate (G13186, Thermo Fisher Scientific). Alexa Fluor 594 phalloidin (A12381, Thermo Fisher Scientific) and TO-PRO-3 Iodide (642/661) were employed for the cells’ visualization on confocal microscopy platform Leica TCS SP8. For the invadopodia formation assay, an immunofluorescence imaging protocol employing phalloidin and anti-cortactin antibody was performed. Invadopodia – as structures double positive for F-actin and cortactin staining – was quantified for tested cell lines and conditions and presented as the number of invadopodia per one cell. Two independent repetitions were performed.
Collagen I hydrogel 3D invasion assay is a modification of the inverted vertical invasion assay (McArdle et al., 2016). Cells were plated on the µ-Slide 8 Well glass bottom coverslips (80827, Ibidi). At 80% confluence, the full medium was replaced with one containing 0.5% FBS for proliferation suppression. Doxycycline for RNF43 induction was applied at this step when needed. Next day, a solution of rat tail collagen type I in final concentration 1.5 mg/ml prepared accordingly to the manufacturer’s protocol was overlaid over the cells and left for polymerization for 30 min at 37°C, 5% CO2. Then medium with final FBS concentration 10% and 100 ng/ml CXCL12 (350-NS, R&D Systems) or 100 ng/ml CCL21 (366-6C, R&D Systems) was added to the wells. After 24 hr, cells were PFA fixed, permeabilized, and stained with Hoechst 33342 (H1399, Thermo Fisher Scientific) and Alexa Fluor 594 phalloidin. Corresponding photos at the coverslip level and at 50 μm (A375 and A2058) or 70 μm (A375 IV) were taken using a confocal laser scanning microscopy platform Leica TCS SP8 (Leica). Invasion index was calculated as the ratio of invaded cells at a specified height to the number of noninvasive ones.
14. Colony formation assay
To assess the ability of colony formation in the presence of 0.3 μM vemurafenib, 300 of the melanoma cells were plated onto 24-well plate and were subsequently cultured for 7 days. After that time, the medium was removed and colonies were washed in PBS, fixed in the ice-cold methanol for 30 min, and stained with 0.5% crystal violet in 25% methanol. After washing and drying, bound crystal violate was eluted with 10% acetic acid and absorbance at 590 nm was measured on Tecan Sunrise plate reader. Results were normalized to the nontreated A375 wild-type results.
15. Animal studies
Animal experiments were approved by the Academy of Sciences of the Czech Republic (AVCR 85/2018), supervised by the local ethical committee, and performed by certified individuals (Karel Souček, Markéta Pícková, Ráchel Víchová).
A375 ctrl and A375 RNF43 TetON cells were implanted as a suspension of 50,000 cells in 50 μl saline intradermally into 9-week-old males of NOD-Rag1null IL2rgnull strain, obtained from Jackson Laboratory. Animals were checked daily, and weight and tumors sizes were measured weekly along perpendicular axes using an external caliper. Tumor volumes were calculated using the equation volume = ½ (length × width2). Upon tumor establishment, animals were divided into five cohorts based on administered cells and treatment: (I) A375 ctrl + vemurafenib (N = 5); (II) A375 RNF43 TetON Dox - (N = 3); (III) A375 RNF43 TetON Dox + (N = 4); (IV) A375 RNF43 TetON VEMURAFENIB Dox - (N = 5); and (V) A375 RNF43 TetON VEMURAFENIB Dox + (N = 4). Doxycycline supplemented for RNF43 expression induction was administrated in drinking water, 0.2 mg/ml in 0.1% weight:volume sucrose. Control cohort obtained drinking water with 0.1% sucrose. Vemurafenib was supplemented daily by oral gavage in concentration of 25 mg/kg/day as freshly prepared formulation in 25% Kolliphor ELP (61791-12-6, Sigma-Aldrich) with 2.5% DMSO. Animals not treated with vemurafenib received the same formulation without inhibitor. Mice were sacrificed when tumors reached approximately 1000 mm3. Tumor samples were analyzed by western blot, and the times to reach the experimental end point was compared.
16. Software and statistics
Statistical significance was confirmed by two-tailed paired or unpaired Student’s t-tests. Survival was analyzed by Mantel–Cox test. Correlation between RNF43(HA) protein level and its targets in the in vivo experiments was tested by one-tailed Spearman correlation test. Statistical significance levels were defined as *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. All statistical details including the number of biological or technical replicates can be found in each figure legend. Statistical analysis and data visualization were performed in GraphPad Prism 8.0 software. Graphs are presented with error bars as ± SD if not stated differently in the figure legends.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2021, Radaszkiewicz et al. This work is published under https://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
RNF43 is an E3 ubiquitin ligase and known negative regulator of WNT/β-catenin signaling. We demonstrate that RNF43 is also a regulator of noncanonical WNT5A-induced signaling in human cells. Analysis of the RNF43 interactome using BioID and immunoprecipitation showed that RNF43 can interact with the core receptor complex components dedicated to the noncanonical Wnt pathway such as ROR1, ROR2, VANGL1, and VANGL2. RNF43 triggers VANGL2 ubiquitination and proteasomal degradation and clathrin-dependent internalization of ROR1 receptor and inhibits ROR2 activation. These activities of RNF43 are physiologically relevant and block pro-metastatic WNT5A signaling in melanoma. RNF43 inhibits responses to WNT5A, which results in the suppression of invasive properties of melanoma cells. Furthermore, RNF43 prevented WNT5A-assisted development of resistance to BRAF V600E and MEK inhibitors. Next, RNF43 acted as melanoma suppressor and improved response to targeted therapies in vivo. In line with these findings,
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer