Inhibition of inducible heat shock protein-70 (hsp72) enhances bortezomib-induced cell death in human bladder cancer cells

Abstract

The proteasome inhibitor bortezomib (Velcade) is a promising new agent for bladder cancer therapy, but inducible cytoprotective mechanisms may limit its potential efficacy. We used whole genome mRNA expression profiling to study the effects of bortezomib on stress-induced gene expression in a panel of human bladder cancer cell lines. Bortezomib induced strong upregulation of the inducible HSP70 isoforms HSPA1A and HSPA1B isoforms of Hsp72 in 253J B-V and SW780 (HSPA1A(high)) cells, but only induced the HSPA1B isoform in UM-UC10 and UM-UC13 (HSPA1A(low)) cells. Bortezomib stimulated the binding of heat shock factor-1 (HSF1) to the HSPA1A promoter in 253JB-V but not in UM-UC13 cells. Methylation-specific PCR revealed that the HSPA1A promoter was methylated in the HSPA1A(low) cell lines (UM-UC10 and UM-UC13), and exposure to the chromatin demethylating agent 5-aza-2'-deoxycytidine restored HSPA1A expression. Overexpression of Hsp72 promoted bortezomib resistance in the UM-UC10 and UM-UC13 cells, whereas transient knockdown of HSPA1B further sensitized these cells to bortezomib, and exposure to the chemical HSF1 inhibitor KNK-437 promoted bortezomib sensitivity in the 253J B-V cells. Finally, shRNA-mediated stable knockdown of Hsp72 in 253J B-V promoted sensitivity to bortezomib in vitro and in tumor xenografts in vivo. Together, our results provide proof-of-concept for using Hsp72 inhibitors to promote bortezomib sensitivity in bladder cancers and suggest that selective targeting of HSPA1B could produce synthetic lethality in tumors that display HSPA1A promoter methylation.

Figure 1. Effects of bortezomib on cell death and Hsp72 mRNA and protein expression in a subset of bladder cancer cells.

A. Effects of bortezomib on cell death. Bladder cancer cell lines (253JB-V, SW780, UM-UC10, UM-UC13) were incubated with or without 100nM bortezomib for 24 hours and PI/FACS was used to quantify cell death. Mean ± SEM, n = 3. B. Effects of bortezomib on the mRNA expression of Hsp72 isoform HSPA1A. Cells were exposed to 30 nM bortezomib for 6 h and HSPA1A was measured by quantitative RT-PCR. RQ = relative quantity (to GAPDH). Values represent mean ± SEM (N≥3) C. Effects of bortezomib on Hsp72 protein levels. Cells were incubated for 16–18 h with 30 nM of bortezomib (nM), and Hsp72 levels were measured in whole cell lysates by immunoblotting. Blots are representative of two independent experiments.

Figure 2. Increased HSPA1B expression compensates for loss of HSPA1A expression in UM-UC10 and UM-UC13 cells.

A. Basal expression of HSPA1A and HSPA1B isoforms across the four cell lines. B. Bortezomib-induced expression of HSPA1A and HSPA1B across the four cell lines. Specific primers for each isoform, as well as a pan-primer that recognized both isoforms, were used to measure expression by quantitative RT-PCR. Values represent mean±SE (n = 2). RQ = relative quantity (to GAPDH).

Figure 3. Selective methylation of the HSPA1A promoter suppresses gene expression in UM-UC10 and UM-UC13 cells.

A. Expression of HSF1. 253J B-V and UM-UC13 bladder cancer cell lines were exposed to bortezomib for 12 h and HSF1 mRNA (left panel) and protein levels (right panel) were measured by quantitative RT-PCR and immunoblotting, respectively. Values represent mean±SE (n = 2); blots are representative of at least two independent experiments. B. Chromatin immunoprecipitation (ChIP) analysis of HSF1 binding to the HSPA1A promoter. Note that basal and bortezomib-induced HSF1 binding was greatly reduced in the bortezomib-sensitive UM-UC13 cells. Mean ± SEM, n = 3. *P<0.05 C. Selective methylation of the HSPA1A promoter in bortezomib-sensitive human bladder cancer cells. Methylation-specific PCR was used to assess chromatin methylation in drug-sensitive (UM-UC10, UM-UC13) and drug-resistant (253J B-V, SW780) cell lines as described in Materials and Methods. m, methylated; u, unmethylated. m/u ratios were calculated using densitometry. Data are representative of at least two independent experiments. D. The DNA methyltransferase inhibitor 5-aza-2′-deoxycytidine restores HSPA1A expression in bortezomib-sensitive cells. Cells were incubated with 5 µM 5-AzdC for 5 days and then incubated with or without 30nM bortezomib for 6 h, and HSPA1A expression was measured by quantitative real-time PCR. Values represent mean ± SEM, n = 3. RQ = relative quantity (to GAPDH).

Figure 4. Genetic modulation of HSPA1A and HSPA1B in HSPA1A-low UM-UC10 and UM-UC13 cells.

A. Effects of enforced HSPA1A overexpression on HSPA1A mRNA and Hsp72 total protein levels in UM-UC10 and UM-UC13 cells. Cells were stably transduced with a lentiviral HSPA1A expression construct, and HSPA1A levels were measured by quantitative RT-PCR. Mean ± SEM, n = 3. RQ = relative quantity (to GAPDH). Protein levels were measured via immunoblotting. Blots are representative of two independent experiments. B. Effects of HSPA1A overexpression on bortezomib-induced cell death. Cells transduced with empty vector (NT) or with the HSPA1A expression construct were exposed to 30 nM bortezomib for 24 h and plasma membrane integrity was measured by trypan blue uptake. (The presence of RFP in the expression construct prevented our use of the PI/FACS cell death assay.) Mean ± SEM, n = 3. *, P<0.02. C. Knockdown of HSPA1B in UM-UC10 cells. Left panel, knockdown efficiencies of siRNA against HSPA1A, HSPA1B, or both isoforms as measured by quantitative RT-PCR following exposure to 30 nM bortezomib for 6 h. RQ = relative quantity (to GAPDH). Right panel, corresponding knockdown of Hsp72 protein levels by the difference siRNA sequences following exposure to 30 nM BZ for 14 h. Data are representative of two independent experiments. D. Effect of HSPA1B knockdown on bortezomib-induced cell death in UM-UC10 cells. Following 72 h knockdown, cells were exposed to 30 nM bortezomib for 24 h and cell death measured by PI/FACS analysis. Values represent mean±SE (n = 3). *P<0.01.

Figure 5. Inhibition of Hsp72 induction sensitizes resistant cells to bortezomib.

A. Effects of stable Hsp72 knockdown on basal and bortezomib-induced Hsp72expression. 253J B-V cells were transduced with a non-targeting (NT) or HSPA1A/B (KD5, KD9) lentiviral shRNA constructs as described in Materials and Methods. Top panel, cells were incubated with or without indicated concentrations of bortezomib for 6 h and HSPA1A expression was measured by quantitative RT-PCR. Mean ± SEM, n = 3. RQ = relative quantity (to GAPDH). Bottom panel, effects of stable Hsp72 knockdown on Hsp72 protein levels. Cells were incubated for 12 h with indicated concentrations of bortezomib (nM), and Hsp72 levels were measured in whole cell lysates by immunoblotting. B. Effects of stable Hsp72 knockdown on bortezomib-induced cell death. Cells were incubated with the indicated concentrations of bortezomib for 48 h and PI/FACS analyses were used to quantify cell death. Mean ± SEM, n = 3. *, P<0.01 compared to corresponding NT values. C. Effects of stable Hsp72 knockdown on lysosomal membrane integrity. Cells were exposed to the indicated concentrations of bortezomib for 18 h prior to staining with Lysotracker Red, and loss of red fluorescence was measured by FACS. Left panel: representative FACS histograms. Right panel: results were quantified (mean±SEM; n = 3). *P<0.03. D. Effects of pharmacologic inhibition by KNK-437 on bortezomib-induced HSPA1A levels. Bortezomib-resistant 253J B-V cells were exposed to 25 µM KNK-437 with or without 100 nM bortezomib for 12 h, and HSPA1A levels were measured by quantitative RT-PCR. Mean ± SEM, n = 3. RQ = relative quantity (to GAPDH). E. Effects of KNK-437 on cell death. 253J B-V cells were exposed to 25 or 50 µM KNK-437 in combination with 30 or 100 nM bortezomib for 48 h, and loss of plasma membrane integrity was quantified by PI/FACS. Mean ± SEM, n = 3. *, P<0.05.

Figure 6. Knockdown of HSPA1A promotes bortezomib-induced growth inhibition in vivo.

A. Effects of bortezomib on HSPA1A induction. Animals were injected twice with 1 mg/kg bortezomib (3 days apart), tumor RNA was harvested, and HSPA1A expression was measured by quantitative RT-PCR. Mean ± SEM, n = 3. RQ = relative quantity (to GAPDH). B. Effects of bortezomib on tumor growth. Athymic nude mice were inoculated s.c. with 253JB-V.KD9 or 253JB-V.NT cancer cells. When tumors became palpable (5–7 days), the mice were treated i.v. biweekly with bortezomib at 1 mg/kg/dose or with saline control. Tumor volumes were measured twice a week after the start of treatment. Values represent mean ± SE (N = 5).