Kinesin Facilitates Phenotypic Targeting of Therapeutic Resistance in Advanced Prostate Cancer

Abstract

Understanding the mechanisms underlying resistance is critical to improving therapeutic outcomes in patients with metastatic castration-resistant prostate cancer. Previous work showed that dynamic interconversions between epithelial-mesenchymal transition to mesenchymal-epithelial transition defines the phenotypic landscape of prostate tumors, as a potential driver of the emergence of therapeutic resistance. In this study, we use in vitro and in vivo preclinical MDA PCa patient-derived xenograft models of resistant human prostate cancer to determine molecular mechanisms of cross-resistance between antiandrogen therapy and taxane chemotherapy, underlying the therapeutically resistant phenotype. Transcriptomic profiling revealed that resistant and sensitive prostate cancer C4-2B cells have a unique differential gene signature response to cabazitaxel. Gene pathway analysis showed that sensitive cells exhibit an increase in DNA damage, while resistant cells express genes associated with protein regulation in response to cabazitaxel. The patient-derived xenograft model specimens are from patients who have metastatic lethal castration-resistant prostate cancer, treated with androgen deprivation therapy, antiandrogens, and chemotherapy including second-line taxane chemotherapy, cabazitaxel. Immunohistochemistry revealed high expression of E-cadherin and low expression of vimentin resulting in redifferentiation toward an epithelial phenotype. Furthermore, the mitotic kinesin-related protein involved in microtubule binding and the SLCO1B3 transporter (implicated in cabazitaxel intracellular transport) are associated with resistance in these prostate tumors. Combinational targeting of kinesins (ispinesib) with cabazitaxel was more effective than single monotherapies in inducing cell death in resistant prostate tumors. Implications: Our findings are of translational significance in identifying kinesin as a novel target of cross-resistance toward enhancing therapeutic vulnerability and improved clinical outcomes in patients with advanced prostate cancer.

Figure 1.. Differential response to anti-androgens and taxane chemotherapy in prostate cancer resistant cell lines.

Panels A and B, show the results of cell viability of C4-2B, C4-2BER, C4-2BAR and C4-2BDR in response to enzalutamide (10μM), 10μM abiraterone or docetaxel (50nM) at 24 (panel A) and 48hrs (panel B) measured by the MTT assay. Panel C, Western blot analysis of E-cadherin, IGFBP-3, HSET, BCL-2 and cofilin protein levels in response to enzalutamide (10μM), abiraterone (10μM) or docetaxel (50nM) (for 24hrs), GAPDH was used as a loading control. Panel D shows representative images of matrigel invasion assay, cells were counted in 4 fields of view from three independent experiments at 24 (panel E) and 48 (panel F) hours. Following wound scratch assay, images were taken in 3 fields of view and quantified at 24 (panel G) and 48 (panel H) hrs. Data represents mean of three independent experiments in triplicate ±SEM, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 as determined by two-way ANOVA.

Figure 2.. Phenotypic profiling of cross-resistance in docetaxel resistant prostate cancer cells to cabazitaxel.

Panels A-D, reveal the dose response of C4-2B and C4-2BDR cells to increasing concentrations (0–50nM) of cabazitaxel at 24 (panel A), 48 (panel B), 72 (panel C) and 96hrs (panel D). The results represent the mean of three independent MTT assay experiments in triplicate as percentage of untreated controls. Panel E, Western blot analysis of E-cadherin, IGFBP-3, HSET, BCL-2 and cofilin protein levels in response to 20nM cabazitaxel in C4-2B and C4-2BDR cells (12 to 48hrs). GAPDH was used as a loading control; the data is representative of three experiments. Panel F, Bcl-2 co-immunoprecipitation of C4-2B and C4-2BDR cells treated with cabazitaxel (12 and 24hrs). Panels G-J, C4-2B and C4-2BDR cells were treated with cabazitaxel (20nM) as indicated and mRNA was analyzed by RT-PCR for transcriptional regulators of EMT were measured by qRT-PCR; panel G-SNAIL, panel H-SLUG, panel I-TWIST, panel J-ZEB1. Data presented as average relative expression to untreated C4-2B cells from three independent experiments performed in triplicate (mean±SEM, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 by ANOVA).

Figure 3.. RNAseq analysis of differential response of cabazitaxel sensitive and resistant prostate cancer cells.

Panel A, PCA analysis between C4-2B and C4-2BDR in response to cabazitaxel (20nM). Panels B-C, heat-map of the unique gene signatures after cabazitaxel treatment of C4-2B and C4-2BDR, respectively. Volcano plot between C4-2B and C4-2BDR at 0hrs (panel D) and 12hrs (panel E) cabazitaxel treatment. Top enriched gene pathways in C4-2B and C4-2BDR cells in response to cabazitaxel (panel F).

Figure 4.. Loss of SLCO1B3 as a mediator of cabazitaxel resistance.

Panels A-B, RNAseq expression signatures of SLC family transporter genes in C4-2B (sensitive) and C4-2BDR (resistant) cells after 6 and 12hrs cabazitaxel treatment. Panel C, Dose response of parental PC3 and cabazitaxel resistant PC3-CR cells to cabazitaxel for 96hrs. Panel D, RNAseq analysis of PC3 and PC3-CR cells. Panels E-F, expression of SLCO1B3 in PC3 and PC3-CR cells via RT-PCR normalized to 18s (panel E), and western blot (panel F, GAPDH as loading control). Panel G, Cell viability analysis of PC3-shSLCO1B3 stable knockdown in dose response to cabazitaxel. Panel H, Kaplan-Meier plot depicts progression free survival of patients with alterations in the SLCO1B3 gene (red) and no alterations (blue) as determined by TCGA database. Statistical significance indicated by *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, students t-test.

Figure 5.. Profiling of kinesin and survival protein in lethal prostate cancer MDA PCa PDX models.

Panel A, PDX prostate cancer specimens were subjected to HSET and Bcl-2 immunohistochemistry, shown by representative images at 20x and 40x magnification. Panel B, Quantification analysis of HSET immunohistochemistry represented as the mean percentage of positive cells (from 6 fields of view), analyzed by two independent observers. Statistical significance indicated by *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, students t-test. Panel C, Kaplan-Meier plot depicts progression free survival of patients with alterations in the HSET gene (red) and no alterations (blue) as determined by TCGA database.

Figure 6.. Combinational targeting of kinesins to overcome therapeutic resistance in prostate cancer cells.

Dose response of cell viability of C4-2B and C4-2BDR prostate cancer cells to kinesin inhibitor ispinesib at 24, 48, 72 and 96hrs, respectively shown on panels A, B, C, and D;. (0–50nM). C4–2b and C4-2BDR cells were treated with kinesin inhibitors CW069 (50μM) or ispinesib (5nM) alone or in combination with cabazitaxel (10nM) for 24 and 48hrs. Viability was determined by MTT assay after 24 (panel E), and 48hrs (panel F). Data presented as mean ± SEM of three independent experiments in triplicate, expressed as percentage of untreated controls. Western blot analysis of EMT, HSET, and apoptotic proteins in C4-2B and C4-2BDR cells following treatment with combinations of CW069 (panel G) or Ispinesib (panel H) and cabazitaxel. GAPDH was used as a loading control. Statistical significance indicated by *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 as determined by two-way ANOVA. Panel I, Schematic illustration revealing that in taxane resistant prostate cancer cells there is upregulation of HSET and Bcl-2 proteins, loss of SLCO1B3 transporter, and increased gene expression of ELF5, CBLN2 and ELOVL6. These key players contribute to the interconversion dynamic between epithelial and mesenchymal phenotypes in response to cabazitaxel. Inhibition of mitotic kinesins (HSET) and survival (Bcl-2) proteins, along with microtubule targeting cabazitaxel provides a potentially powerful platform to overcome therapeutic resistance in CRPC.

Figure 7.. Transcriptional regulation of EMT by kinesins and androgens.

Human prostate cancer cells LNCaP, were treated with 50μM CW069 (HSET inhibitor) in the presence or absence of DHT (1nM) for 6, 12 and 24hrs in CSS. RT-PCR was subsequently performed to evaluate the mRNA levels PSA (panel A), KIFC1 (panel B), E-cadherin (panel C), and SNAIL (panel D) Data presented as average relative expression to untreated controls from three independent experiments performed in triplicate (mean±SEM, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001).