Repositioning of antiarrhythmics for prostate cancer treatment: a novel strategy to reprogram cancer-associated fibroblasts towards a tumor-suppressive phenotype

Abstract

Background: Cancer-associated fibroblasts (CAFs) play a significant role in fueling prostate cancer (PCa) progression by interacting with tumor cells. A previous gene expression analysis revealed that CAFs up-regulate genes coding for voltage-gated cation channels, as compared to normal prostate fibroblasts (NPFs). In this study, we explored the impact of antiarrhythmic drugs, known cation channel inhibitors, on the activated state of CAFs and their interaction with PCa cells.

Methods: The effect of antiarrhythmic treatment on CAF activated phenotype was assessed in terms of cell morphology and fibroblast activation markers. CAF contractility and migration were evaluated by 3D gel collagen contraction and scratch assays, respectively. The ability of antiarrhythmics to impair CAF-PCa cell interplay was investigated in CAF-PCa cell co-cultures by assessing tumor cell growth and expression of epithelial-to-mesenchymal transition (EMT) markers. The effect on in vivo tumor growth was assessed by subcutaneously injecting PCa cells in SCID mice and intratumorally administering the medium of antiarrhythmic-treated CAFs or in co-injection experiments, where antiarrhythmic-treated CAFs were co-injected with PCa cells.

Results: Activated fibroblasts show increased membrane conductance for potassium, sodium and calcium, consistently with the mRNA and protein content analysis. Antiarrhythmics modulate the expression of fibroblast activation markers. Although to a variable extent, these drugs also reduce CAF motility and hinder their ability to remodel the extracellular matrix, for example by reducing MMP-2 release. Furthermore, conditioned medium and co-culture experiments showed that antiarrhythmics can, at least in part, reverse the protumor effects exerted by CAFs on PCa cell growth and plasticity, both in androgen-sensitive and castration-resistant cell lines. Consistently, the transcriptome of antiarrhythmic-treated CAFs resembles that of tumor-suppressive NPFs. In vivo experiments confirmed that the conditioned medium or the direct coinjection of antiarrhythmic-treated CAFs reduced the tumor growth rate of PCa xenografts.

Conclusions: Collectively, such data suggest a new therapeutic strategy for PCa based on the repositioning of antiarrhythmic drugs with the aim of normalizing CAF phenotype and creating a less permissive tumor microenvironment.

Keywords: Antiarrhythmics; Cancer-associated fibroblasts; Drug-repositioning; Prostate cancer.

Fig. 1

Cation channels are up-modulated in CAFs and involved in fibroblast activation. a Heatmap (bottom) reporting normalized enrichment scores (NES) for gene sets related to voltage gated channels and bar plot (top) reporting mean + sd, as calculated by GSEA on three independent datasets of prostate cancer patient-derived CAFs vs. matched NPFs. b qRT-PCR showing CACNA1H, CACNB1, CACNB3, SCN2A, SCN1B, and KCNS3 expression levels in an independent setting of three CAF and matched NPF cultures. Data were reported as relative expression compared to NPF and were representative of three independent experiments. c Western blotting analysis showing selected ion channel protein levels in a pair of matched CAFs and NPFs. β-actin was used as endogenous control. d qRT-PCR indicating relative expression levels of α-SMA, FAP and COL1A1 in NPF#1, NPF#2 and WPMY-1 fibroblasts exposed to CM of DU145 cells with respect to control fibroblasts. e Western blotting showing α-SMA, FAP and COL1A1 expression levels in WPMY-1 fibroblasts exposed or not to CM of DU145 cells. β-tubulin was used as endogenous control. f Immunofluorescence microphotographs showing α-SMA (green) and Col1a1 (red) expression in NPF and NPF exposed to CM of DU145 cells. Nuclei counterstained with DAPI (blue). Scale bar, 50 µm. g Cation channel mRNA expression levels in NPF#1, NPF#2 and WPMY-1 fibroblasts exposed to CM of DU145 cells with respect to untreated fibroblasts. h Western blotting displaying cation channel protein levels in NPF#1 and WPMY-1 fibroblasts exposed or not to CM of DU145 cells. β-tubulin was used as endogenous control. Results reported in the figure represent the mean (+ SD) of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.005, Student’s t-test

Fig. 2

Activated fibroblasts show increased membrane conductance for potassium, sodium and calcium. a Representative Voltage-clamp recordings of inward and outward currents from activated and control WPMY-1 cells fibroblasts. A schematic of the applied protocol is shown in the insert. Upon activation, an inward current as well as a fast-inactivating outward current can be detected in fibroblasts. b The inward current shows significant sensitivity to both flecainide (2.5 μM) and nifedipine (2.5 μM) when applied to the bath solution. c Pie chart summarizing the percentage of cells expressing the main currents detected in A upon depolarization in both the control and the activated group. d The IV plot for the K + steady shows a significant increase in the current density in the activated group compared to the control condition. The I-V plot for the inward current is also displayed

Fig. 3

Antiarrhythmics counteract the activated state of prostate CAFs. a Western blotting and relative quantification showing protein levels of fibroblast activation markers (α-SMA and COL1A1) in CAFs treated for 48 h with sub-toxic doses of antiarrhythmics. β-tubulin was used as endogenous control. b Bar plots showing the wound-healing rate assessed by scratch assay on CAFs exposed to antiarrhythmics. Data are reported as wound healing ratio at 24 h compared to 0 h point. c Representative immunofluorescence microphotographs (upper panel) showing the organization of β-actin cytoskeleton (green) and p-FAK (red) in CAFs treated with verapamil as compared to untreated. Scale bar, 50 μm. Western blotting analysis (lower panel) showing p-FAK, FAK protein levels in CAFs treated with sub-toxic doses of antiarrhythmics. β-tubulin was used as endogenous control. d Representative images (upper panel) showing 3D-collagen gel remodeling of CAFs exposed to sub-toxic doses of antiarrhythmics. NPFs were used as negative control. The dotted lines define gel areas. Bar plots (lower panel) showing ECM remodeling ratio of treated CAFs assessed by 3D-collagen gel assay. Data are reported as ECM remodeling ratio at 48 h compared to 0 h point. e Western blotting showing levels of pro-MMP2 and active-MMP2 in CM from CAFs exposed or not to sub-toxic doses of antiarrhythmics and, pro-MMP2 intracellular levels in treated cells. Gapdh was used as endogenous control for cell lysate. Results reported in the figure represent the mean (+ SD) of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.005, Student’s t-test

Fig. 4

Antiarrhythmics affect PCa cell growth by impairing CAF function. a Schematic representation of CM experiment work-flow (Created with Biorender.com). b Graph reporting the growth of DU145 cells cultured with CM from CAFs treated or not with antiarrhythmics, or CM from NPFs at different time points (24, 48, 72 h). c Graph reporting the growth of LNCaP cells cultured with CM from CAFs treated or not with antiarrhythmics, or CM from NPFs at different time points (24, 48, 72 h). d Cell cycle phase distribution of DU145 cells cultured with CM from CAFs treated or not with antiarrhythmics, or CM from NPFs at 72 h until treatment. Results reported in the figure represent the mean (+ SD or ± SD) of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.005, Student’s t-test

Fig. 5

Antiarrhythmics affect PCa cell plasticity by impairing CAF function. a-d Western blotting analysis showing E-cadherin, β-catenin, Vimentin and Snail protein amount in DU145 cells (a-b) and LNCaP cells (c-d) exposed to CM from NPFs or CM from CAFs treated or not to antiarrhythmics. β-tubulin was used as endogenous control. e Representative bright-field microphotographs (left panel) showing migration rate of DU145 cells exposed to CM from NPFs or CM from CAFs treated or not with nifedipine or flecainide. Scale bar, 100 μm. The dotted lines define the areas lacking cells. Bar plots (right panel) showing the wound-healing rate of DU145 cells upon the indicated treatments, as from the scratch assay. Data are reported as wound healing ratio at 24 h compared to 0 h. f Cytokine and chemokine protein array blots (left panel) of CM from CAFs treated or not with nifedipine or flecainide, and CM from NPFs. Bar plot (right panel) depicts the pixel density of each cytokine or chemokine (mean). The signal intensity of each cytokine or chemokine was expressed relative to the mean of the intensity of the corresponding spots from vehicle control sample. g Western blotting and relative quantification showing the expression of stemness markers (CD133 and CD44) in DU145 cells exposed to CM from CAFs treated or not with antiarrhythmics, or CM from NPFs, with respect to untreated cells. β-tubulin was used as endogenous control. Results reported in the figure represent the mean (+ SD) of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.005, Student’s t-test., when calculated against untreated cells

Fig. 6

Antiarrhythmics normalize the transcriptome of CAFs. a Scatter plot of t-SNE components showing similarity of transcriptomes of CAFs, NPFs and antiarrhythmics-treated CAFs. b Venn diagram showing overlap between Reactome gene sets enriched (GSEA, NES < 0, FDR p-val < 0.05) in genes down-regulated in CAFs upon nifedipine and flecainide treatments. c Bar plot showing NES of representative gene sets down-regulated in nifedipine- and flecainide-treated CAFs

Fig. 7

Antiarrhythmics impair the capability of CAFs to sustain PCa cell growth in vivo. a DU145 cells were subcutaneously injected into the right flanks of SCID mice. When tumors reached the volume of ~ 100 mm.³, mice were randomized into four groups and were intra-tumorally treated for 5 days per 2 weeks with CM from CAFs treated or not with nifedipine or flecainide, or CM from NPFs (see insert on the right for a schematic representation of the experiment). The graph reports tumor volumes along the experiment. Black rows indicate when treatment was administered. Schematic representation of the experimental workflow (created with Biorender.com) (b) Western blotting showing E-cadherin levels in PCa tumors excised at the end of the treatment with CM from CAFs treated or not with antiarrhythmics, or CM from NPFs. β-tubulin was used as endogenous control. c Representative bright-field microphotographs (upper panel) showing Ki-67 staining in PCa tumors excised at the end of the treatment with CM from CAFs treated or not with antiarrhythmics, or CM from NPFs. Bar plot (lower panel) showing Ki-67 positive cells in PCa tumors upon the relative treatment. Data were reported as percentage of Ki-67 positive cells with respect to total number of cells. Eight fields were evaluated for each condition. d Representative bright-field microphotographs (upper panel) showing CD31 staining in PCa tumors excised at the end of the treatment with CM from CAFs treated or not with antiarrhythmics, or CM from NPFs. Bar plot (lower panel) showing CD31 positive cells in PCa tumors upon the relative treatment. Data were reported as percentage of CD31 positive cells with respect to total number of cells. Eight fields were evaluated for each condition. e DU145 cells were subcutaneously co-injected with CAFs, CAFs pretreated with nifedipine or flecainide, or with NPFs, into the right flanks of SCID mice. The graph report tumor volumes along the experiment. f Timeline indicating tumor take (n. of tumors/n. of co-injected mice) in the different experimental groups, as from the experiment described in panel e. (Created with Biorender.com)