Vemurafenib Inhibits Active PTK6 in PTEN-null Prostate Tumor Cells

Abstract

Protein tyrosine kinase 6 (PTK6, also called BRK) is overexpressed and activated in human prostate cancer. Loss of the tumor suppressor PTEN, a frequent event in prostate cancer, leads to PTK6 activation at the plasma membrane and its oncogenic signaling. The small molecule inhibitor vemurafenib, also known as PLX4032, and its tool analog PLX4720 were designed to inhibit constitutively active BRAF V600E, yet they also have potent effects against PTK6. Vemurafenib is used in the treatment of metastatic melanoma, but its efficacy in prostate cancer has not been assessed. When activated at the plasma membrane, PTK6 promotes signaling through FAK, EGFR, and ERK1/2, and we show this can be blocked by vemurafenib. In addition, PTK6-mediated cell growth, migration, and invasion are inhibited upon vemurafenib administration. Using a flank xenograft model, vemurafenib treatment reduced tumor burden. Using saturation transfer difference NMR and molecular docking, we demonstrate that vemurafenib binds in the active site of PTK6, inhibiting its activation. These structural studies provide insight into the PTK6-vemurafenib complex, which can be utilized for further refinement chemistry, whereas functional studies demonstrate that active PTK6 is a viable drug target in prostate cancer.

Figure 1.. Vemurafenib inhibits activation and signaling downstream of PTK6.

(A) PC3 cells stably expressing Palm-PTK6-YF were treated with vemurafenib or DMSO vehicle and incubated for the times described. Changes in phosphorylation of PTK6, FAK, and BCAR1 were monitored by immunoblotting. Activation of PTK6 is monitored by phosphorylation of Y342. (B) PC3 Vector and PC3 Palm-PTK6-YF cells were treated with vemurafenib and total cell lysates were prepared. Changes in protein levels and phosphorylation were monitored by immunoblotting. Each lane represents one plate of cells (independent experimental replicate). Quantitation of immunoblot data is presented for DMSO and vemurafenib treated Palm-PTK6-YF (Palm-YF) expressing cells in the lower panel. Relative changes in phosphorylation of PTK6 targets were normalized to total protein levels. Residual signal following vemurafenib treatment is displayed ± SEM. (C) siRNA mediated knockdown of BRAF does not inhibit PTK6 activity. Cells were transiently transfected with scrambled or BRAF targeting siRNA (SCRsi and BRAFsi) and harvested at 72 hours. Activating phosphorylation of PTK6 (PY342) and its direct substrates FAK and EGFR are not reduced following knockdown of wild type BRAF. (D) PC3 cells subjected to stable PTK6 knockdown by two shRNA targeting vectors (sh49 and sh52) or scrambled shRNA control (shSCR) were incubated with 1 μM vemurafenib or DMSO vehicle. Similar reductions in downstream signaling are detected in PC3 cells by knockdown of PTK6 (DMSO treated cells) or vemurafenib treatment (shSCR lane). Changes in protein levels and phosphorylation were monitored by immunoblotting.

Figure 2.. Vemurafenib inhibits PTK6 activation in multiple cell types.

(A) PC3 Vector cells, PC3 cells stably expressing Palm-PTK6 YF (PTK6-YF), MDA-MB-468, T47D, MDA-MB-231 and MIA PaCa-2 cells were treated with DMSO or 1 μM Vemurafenib for 24 hours. Cells were stained for active PTK6 PY342 (Alexa Fluor 488; green). Cells were counterstained with DAPI (blue). Treatment with vemurafenib inhibits active PTK6 in cells of different origin. Scale bar, 50 μm. (B) (B) Quantification of the PY342 Alexa Fluor 488 immunofluorescence signal is presented for the DMSO and vemurafenib treated cells. Signal intensities in vemurafenib treated cells were normalized to the average signal observed in DMSO treated cells and expressed as the average ± SEM. Significance was assessed using the Student’s two-tailed t-test (n=15 cells/cell type/treatment, **** P < 0.0001).

Figure 3.. Oncogenic functions of PTK6 are inhibited upon vemurafenib treatment.

(A) PC3 cells expressing empty vector (PC3-VEC) or Palm-PTK6-YF (PC3-Palm-YF) were used in a proliferation assay with the addition of vemurafenib (1μM) or DMSO vehicle. Cell number was plotted ± SEM. Significance was assessed at Day 6 using the Student’s two-tailed t-test (n=3 independent trials run in triplicate, ** P < 0.01). (B) PC3 cells expressing Palm-PTK6-YF were used in a scratch migration assay with the addition of vemurafenib (1μM) or DMSO vehicle. Wound gap distance was measured after the time described ± SEM. Significance was assessed at each time point using the Student’s two-tailed t-test (n=3 independent trials run in triplicate, ** P < 0.01). (C) PC3 cells expressing empty vector or Palm-PTK6-YF were used in a transwell invasion assay with the addition of vemurafenib (1μM) or DMSO vehicle and invaded cells were quantified after 24 hours and plotted ± SEM. Significance was assessed using the Student’s two-tailed t-test (n=3 independent trials run in triplicate, *** P < 0.001).

Figure 4.. Vemurafenib reduces PTK6-induced tumor burden in vivo.

PC3 cells expressing empty vector or Palm-PTK6-YF (PC3-Palm-YF) were injected into the flanks of nude mice. After tumors were palpable (~100 mm³), the mice were fed control chow or chow containing 417 mg/kg PLX4720 (tool analog of vemurafenib). (A) Tumor volume was measured on a regular basis (n=6/group). Significance was assessed at Day 41 post-injection using the Student’s two-tailed t-test. (B) At the completion of the study (41 days post-injection), the masses of the tumors were recorded (n=6). Significance was assessed at Day 41 post-injection using the Student’s two-tailed t-test. (C) At the completion of the study, the masses of the mice were recorded (n=3). Significance was assessed at Day 41 post-injection using the Student’s two-tailed t-test. (D). At the completion of the study, pictures of the mice and tumors were recorded. Scale bar, 2 cm. (E) Tumors from mice bearing PC3-VEC and PC3-Palm-YF expressing xenografts were sectioned and immunostained for Ki67. (F) Tumors from mice bearing PC3-VEC and PC3-Palm-YF expressing xenografts were sectioned and immunostained for cleaved caspase-3 (CC3+). For E and F, antibody stained xenograft tumor sections were examined for hot spots, areas of higher staining than the surrounding tissue, and positive cells were counted in at least four different hot spots per tumor sample. Results are expressed as an average ± SEM, and one-way ANOVA followed by Tukey’s multiple comparison test (GraphPad Prism 7) were used to determine P-values (bottom panels; n= 5, * P < 0.05, *** P < 0.001, **** P < 0.0001).

Figure 5.. Vemurafenib directly binds the catalytic domain of PTK6.

(A) Full-length (FL), SH3, SH2, SH3-SH2 and kinase domains of PTK6 fused to GST or GST alone were expressed in bacteria and affinity-purified using glutathione agarose. Purified proteins were analyzed by SDS-PAGE. (B) 1D ¹H NMR spectrum of 1 mM vemurafenib and STD NMR spectra of vemurafenib (1 mM) alone, GST-FL-PTK6 (10 μM) alone or 1 mM vemurafenib plus 10 μM GST-FL-PTK6. Inset shows signals from 1D ¹H NMR spectra of vemurafenib (1 mM) in the aromatic region. (C) STD NMR spectra of GST or PTK6 GST-fusion proteins (10 μM) plus 1 mM vemurafenib. (D) Molecular docking of vemurafenib (green) into the crystal structure of PTK6 (grey) (PMID: 27993680). Distinct regions of the kinase domain are color coded as follows: Pi loop (yellow), α-C helix (magenta), Hinge (blue), HRD loop (orange), DFG loop (cyan), Catalytic triad (red). Magnified images show how the sulfonamide and propyl groups of vemurafenib are oriented towards the PTK6 catalytic triad in our model of the drug-kinase complex.