Ezrin Inhibition Overcomes Acquired Resistance to Vemurafenib in BRAFV600E-Mutated Colon Cancer and Melanoma Cells In Vitro

Abstract

Despite the advancements in targeted therapy for BRAFV600E-mutated metastatic colorectal cancer (mCRC), the development of resistance to BRAFV600E inhibition limits the response rate and durability of the treatment. Better understanding of the resistance mechanisms to BRAF inhibitors will facilitate the design of novel pharmacological strategies for BRAF-mutated mCRC. The aim of this study was to identify novel protein candidates involved in acquired resistance to BRAFV600E inhibitor vemurafenib in BRAFV600E-mutated colon cancer cells using an integrated proteomics approach. Bioinformatic analysis of obtained proteomics data indicated actin-cytoskeleton linker protein ezrin as a highly ranked protein significantly associated with vemurafenib resistance whose overexpression in the resistant cells was additionally confirmed at the gene and protein level. Ezrin inhibition by NSC305787 increased anti-proliferative and pro-apoptotic effects of vemurafenib in the resistant cells in an additive manner, which was accompanied by downregulation of CD44 expression and inhibition of AKT/c-Myc activities. We also detected an increased ezrin expression in vemurafenib-resistant melanoma cells harbouring the BRAFV600E mutation. Importantly, ezrin inhibition potentiated anti-proliferative and pro-apoptotic effects of vemurafenib in the resistant melanoma cells in a synergistic manner. Altogether, our study suggests a role of ezrin in acquired resistance to vemurafenib in colon cancer and melanoma cells carrying the BRAFV600E mutation and supports further pre-clinical and clinical studies to explore the benefits of combined BRAF inhibitors and actin-targeting drugs as a potential therapeutic approach for BRAFV600E-mutated cancers.

Keywords: BRAF inhibitor; BRAFV600E; NSC305787; actin cytoskeleton; colon cancer; ezrin; ezrin inhibitor; melanoma; vemurafenib.

Figure 1

Bioinformatics analysis of combined proteomics datasets obtained by 2-DE/MALDI-TOF/TOF MS and LC–MS/MS to select new protein targets and associated cellular processes pertinent to the development of resistance to vemurafenib in BRAFV600E-mutated RKO colon cancer cells. (A) Network of interactions between the proteins differentially expressed in RKO colon cancer cells resistant to BRAFV600E targeting by vemurafenib (STRING). The PPI analysis included a combined list of proteins identified as overexpressed in the resistant cell line by 2-DE/MALDI-TOF/TOF MS and/or LC–MS/MS analysis. Nodes correspond to the dataset proteins and the edges to the combined evidence of the protein interactions (>0.7 evidence score). KMeans (10 clusters) clustering of the nodes was performed to refine visualisation of the connectivity patterns. The dashed lines depict edges connecting the clusters. Associations of the proteins with selected functional categories from GO Biological Functions ontologies are color-coded as follows: GO:0006903, vesicle targeting (p = 0.0018)—brown; GO:0042866, pyruvate biosynthetic process (p = 0.0459)—dark blue; GO:0042254, ribosome biogenesis (p = 2.39 × 10⁻⁵)—dark yellow; GO:0030036, actin cytoskeleton organisation (p = 1.50 × 10⁻⁶)—green; GO:0042594, response to starvation (p = 0.0083)—yellow; GO:0045055, regulated exocytosis (p = 1.39 × 10⁻⁸)—blue; GO:0009165, nucleotide biosynthetic process (p = 0.00059)—dark purple; GO:0040017, positive regulation of locomotion (7.12 × 10⁻⁵)—light green; GO:1901987, regulation of cell cycle phase transition (p = 0.0146)—purple; HSA-5673001 RAF/MAP kinase cascade (p = 0.0299)—red. (B) Cytoscape analysis. Topological connectivity characteristics of the selected proteins/potential targets. In bold—proteins with network connections and in the top 25 differential (MS) log₂ fold change (p < 0.05); ADAM10 and RAF1—proteins with strong network connectivity and with low but positive differential (MS) log₂ fold change; EZR and ALDOA—top differentially expressed proteins in MALDI data; average path length—average number of steps along the shortest paths for all possible pairs of network; betweenness centrality—the number of times a node acts as a bridge along the shortest path between two other nodes; closeness centrality—node’s average farness (inverse distance) to all other nodes; clustering coefficient—the degree to which nodes in a graph tend to cluster together; eccentricity—the degree of a node’s combined paths’ variation from being circular. (C) KEGG pathway enrichment analysis. (D) The network reconstructed in STRING shows the interconnection of the EZR (ezrin), CAV1, RAF1, ADAM10 and CD44 and the key functions involved in colorectal cancer. Network settings: 0.7 confidence, confidence view. Color-coded associations of the proteins with selected functional categories from GO Biological Functions, cellular components and KEGG pathways: Red—GO:0072584, caveolin-mediated endocytosis (0.00053); blue—GO:2000641, regulation of early endosome to late endosome transport (0.0029); purple—GO:0070161, anchoring junction (1.12 × 10⁻⁵); cyan—GO:0005925, focal adhesion (1.12 × 10⁻⁵); dark green—GO:0098805, whole membrane (0.00053); magenta—GO:0016323, basolateral plasma membrane(0.00091); orange—GO:0045121, membrane raft (0.0015); lime green—GO:0005901, caveola (0.0354); yellow—GO:0005902, microvillus (0.0399); maroon—hsa05218, melanoma (8.75 × 10⁻⁹); grey—hsa05210, colorectal cancer (8.75 × 10⁻⁹).

Figure 1

Bioinformatics analysis of combined proteomics datasets obtained by 2-DE/MALDI-TOF/TOF MS and LC–MS/MS to select new protein targets and associated cellular processes pertinent to the development of resistance to vemurafenib in BRAFV600E-mutated RKO colon cancer cells. (A) Network of interactions between the proteins differentially expressed in RKO colon cancer cells resistant to BRAFV600E targeting by vemurafenib (STRING). The PPI analysis included a combined list of proteins identified as overexpressed in the resistant cell line by 2-DE/MALDI-TOF/TOF MS and/or LC–MS/MS analysis. Nodes correspond to the dataset proteins and the edges to the combined evidence of the protein interactions (>0.7 evidence score). KMeans (10 clusters) clustering of the nodes was performed to refine visualisation of the connectivity patterns. The dashed lines depict edges connecting the clusters. Associations of the proteins with selected functional categories from GO Biological Functions ontologies are color-coded as follows: GO:0006903, vesicle targeting (p = 0.0018)—brown; GO:0042866, pyruvate biosynthetic process (p = 0.0459)—dark blue; GO:0042254, ribosome biogenesis (p = 2.39 × 10⁻⁵)—dark yellow; GO:0030036, actin cytoskeleton organisation (p = 1.50 × 10⁻⁶)—green; GO:0042594, response to starvation (p = 0.0083)—yellow; GO:0045055, regulated exocytosis (p = 1.39 × 10⁻⁸)—blue; GO:0009165, nucleotide biosynthetic process (p = 0.00059)—dark purple; GO:0040017, positive regulation of locomotion (7.12 × 10⁻⁵)—light green; GO:1901987, regulation of cell cycle phase transition (p = 0.0146)—purple; HSA-5673001 RAF/MAP kinase cascade (p = 0.0299)—red. (B) Cytoscape analysis. Topological connectivity characteristics of the selected proteins/potential targets. In bold—proteins with network connections and in the top 25 differential (MS) log₂ fold change (p < 0.05); ADAM10 and RAF1—proteins with strong network connectivity and with low but positive differential (MS) log₂ fold change; EZR and ALDOA—top differentially expressed proteins in MALDI data; average path length—average number of steps along the shortest paths for all possible pairs of network; betweenness centrality—the number of times a node acts as a bridge along the shortest path between two other nodes; closeness centrality—node’s average farness (inverse distance) to all other nodes; clustering coefficient—the degree to which nodes in a graph tend to cluster together; eccentricity—the degree of a node’s combined paths’ variation from being circular. (C) KEGG pathway enrichment analysis. (D) The network reconstructed in STRING shows the interconnection of the EZR (ezrin), CAV1, RAF1, ADAM10 and CD44 and the key functions involved in colorectal cancer. Network settings: 0.7 confidence, confidence view. Color-coded associations of the proteins with selected functional categories from GO Biological Functions, cellular components and KEGG pathways: Red—GO:0072584, caveolin-mediated endocytosis (0.00053); blue—GO:2000641, regulation of early endosome to late endosome transport (0.0029); purple—GO:0070161, anchoring junction (1.12 × 10⁻⁵); cyan—GO:0005925, focal adhesion (1.12 × 10⁻⁵); dark green—GO:0098805, whole membrane (0.00053); magenta—GO:0016323, basolateral plasma membrane(0.00091); orange—GO:0045121, membrane raft (0.0015); lime green—GO:0005901, caveola (0.0354); yellow—GO:0005902, microvillus (0.0399); maroon—hsa05218, melanoma (8.75 × 10⁻⁹); grey—hsa05210, colorectal cancer (8.75 × 10⁻⁹).

Figure 2

Western blot analysis of total ezrin (A) and phosphorylated ezrin (B) protein expression in vemurafenib-sensitive (parental, RKO) and -resistant (RKOr) colon cancer cells cultured in the absence (CTRL) and presence of vemurafenib (3 µM) for indicated time periods (24 h, 48 h, 72 h). The expression levels of ezrin and phospho-ezrin phosphorylated at threonine 567 were measured by densitometry analysis using the Quantity One 1-D Analysis Software version 4.6.6. The level of ezrin was normalised to the total amount of proteins loaded in the lane (total protein normalisation), whereas the level of phospho-ezrin was normalised to ezrin. The data represent the mean and standard deviation obtained from three independent biological experiments. Statistical significance (p < 0.05) is denoted with an asterisk.

Figure 3

Anti-proliferative effect of combination treatment with vemurafenib and ezrin inhibitor NSC305787 in vemurafenib-resistant RKO colon cancer cells after a 72 h treatment period. (A) Concentration-dependent effect of combination treatment on cell viability measured by MTT assay. IC50 values were calculated using linear regression analysis and the results were statistically analysed by ANOVA with Tukey’s post hoc test, and statistically significant differences (p < 0.05) are indicated by an asterisk. Data were obtained from three independent biological experiments. (B) The combination index (CI) was calculated for each drug combination from growth inhibition curves using CompuSyn software (version 1.0), where CI = 1, <1 and >1 indicate additive effect, synergism and antagonism, respectively. Data were obtained from three independent biological experiments. The treatment combination exerting an additive anti-proliferative effect is highlighted in red and was selected for further functional studies.

Figure 3

Anti-proliferative effect of combination treatment with vemurafenib and ezrin inhibitor NSC305787 in vemurafenib-resistant RKO colon cancer cells after a 72 h treatment period. (A) Concentration-dependent effect of combination treatment on cell viability measured by MTT assay. IC50 values were calculated using linear regression analysis and the results were statistically analysed by ANOVA with Tukey’s post hoc test, and statistically significant differences (p < 0.05) are indicated by an asterisk. Data were obtained from three independent biological experiments. (B) The combination index (CI) was calculated for each drug combination from growth inhibition curves using CompuSyn software (version 1.0), where CI = 1, <1 and >1 indicate additive effect, synergism and antagonism, respectively. Data were obtained from three independent biological experiments. The treatment combination exerting an additive anti-proliferative effect is highlighted in red and was selected for further functional studies.

Figure 4

Pro-apoptotic effect of combination treatment with vemurafenib and ezrin inhibitor NSC305787 in vemurafenib-resistant RKO colon cancer cell line after a 48 h treatment period. (A–D) Representative images of Annexin V assay performed to detect induction of apoptosis when cells were treated with either single agents or their combination at indicated concentrations for 48 h. Confocal microscopy images were taken at 20× magnification: (A) untreated cells; (B) cells treated with vemurafenib (VEM); (C) cells treated with EZR inhibitor (EZRi); (D) cells treated with dual therapy (VEM + EZRi); (E) quantification of the percentage of unaffected cells (Ann−/PI−), early apoptotic cells (Ann+/PI−) and late apoptotic/primary necrotic cells (Ann+/PI+). Data are representative of two independent biological experiments performed in two replicates. Ann—Annexin V; PI—propidium iodide. Values represent the mean ± SD. The number of visual fields was 8 for all the treatments at both replicates. Two-way ANOVA with Tukey’s post hoc test was used for statistical analysis where only the most significant differences between the groups were marked by three asterisks (p ≤ 0.001). ns = non-significant.

Figure 4

Pro-apoptotic effect of combination treatment with vemurafenib and ezrin inhibitor NSC305787 in vemurafenib-resistant RKO colon cancer cell line after a 48 h treatment period. (A–D) Representative images of Annexin V assay performed to detect induction of apoptosis when cells were treated with either single agents or their combination at indicated concentrations for 48 h. Confocal microscopy images were taken at 20× magnification: (A) untreated cells; (B) cells treated with vemurafenib (VEM); (C) cells treated with EZR inhibitor (EZRi); (D) cells treated with dual therapy (VEM + EZRi); (E) quantification of the percentage of unaffected cells (Ann−/PI−), early apoptotic cells (Ann+/PI−) and late apoptotic/primary necrotic cells (Ann+/PI+). Data are representative of two independent biological experiments performed in two replicates. Ann—Annexin V; PI—propidium iodide. Values represent the mean ± SD. The number of visual fields was 8 for all the treatments at both replicates. Two-way ANOVA with Tukey’s post hoc test was used for statistical analysis where only the most significant differences between the groups were marked by three asterisks (p ≤ 0.001). ns = non-significant.

Figure 5

Western blot analysis of protein targets potentially associated with the chemosensibilisation effect of combination treatment with vemurafenib and ezrin inhibitor NSC305787 in vemurafenib-resistant RKO colon cancer cells after a 72 h treatment period. Relative protein expression levels of selected proteins were measured by densitometry analysis using Quantity One software and normalised to the total amount of protein loaded in the lane (total protein normalisation). The data represent the mean and standard deviation obtained from three independent biological experiments. Statistical significance (p < 0.05) is denoted with an asterisk, and p ≤ 0.001 with three asterisks. CTRL—untreated/control cells; VEM—vemurafenib/27.91 µM; EZRi—ezrin inhibitor/1.61 µM; VEM + EZRi—27.91 µM vemurafenib + 1.61 µM ezrin inhibitor.

Figure 6

Western blot analysis of total ezrin (A) and phosphorylated ezrin (B) protein expression in vemurafenib-sensitive (parental, A375) and vemurafenib-resistant (A375-r) melanoma cells carrying the BRAFV600E mutation in the absence (CTRL) and in the presence of 0.8 µM vemurafenib for 72 h. The expression levels of ezrin and phospho-ezrin phosphorylated at threonine 567 were measured by densitometry analysis using the Quantity One software. The level of ezrin was normalised to the total amount of proteins loaded in the lane (total protein normalisation), whereas the level of phospho-ezrin was normalised to ezrin. The data represent the mean and standard deviation obtained from three independent biological experiments.

Figure 7

Anti-proliferative effect of the combination treatment with vemurafenib and ezrin inhibitor NSC305787 in vemurafenib-resistant A375 melanoma cells carrying the BRAFV600E mutation after a 72 h treatment period. (A) Concentration-dependent effect of combination treatment on cell viability measured by MTT assay. IC50 values were calculated using linear regression analysis and the results were statistically analysed by ANOVA with Tukey’s post hoc test, and statistically significant differences (p < 0.05) are indicated by an asterisk. Data were obtained from three independent biological experiments. (B) The combination index (CI) was calculated for each drug combination from growth inhibition curves using CompuSyn software, where CI = 1, <1 and >1 indicate additive effect, synergism and antagonism, respectively. Data were obtained from three independent biological experiments. Combinations of concentrations exerting the most potent synergistic effect are highlighted in red.

Figure 8

Pro-apoptotic effect of the combination treatment with vemurafenib and ezrin inhibitor NSC305787 in vemurafenib-resistant A375 melanoma cell line after a 48 h treatment period. (A–D) Representative images of Annexin V assay performed to detect induction of apoptosis when cells were treated with either single agents or their combination at indicated concentrations for 48 h. Confocal microscopy images were taken at 20x magnification: (A) untreated cells; (B) cells treated with vemurafenib (VEM); (C) cells treated with EZR inhibitor (EZRi); (D) cells treated with dual therapy (VEM + EZRi); (E) quantification of the percentage of unaffected cells (Ann−/PI−), early apoptotic cells (Ann+/PI−) and late apoptotic/primary necrotic cells (Ann+/PI+). Data are representative of two independent biological experiments performed in two replicates. Ann—Annexin V; PI—propidium iodide. Values represent the mean ± SD. The number of visual fields was 8 for all the treatments at both replicates. Two-way ANOVA with Tukey’s post hoc test was used for statistical analysis where only the most significant differences between the groups were marked by three asterisks (p ≤ 0.001).