Targeting Heat Shock Protein 27 and Fatty Acid Oxidation Augments Cisplatin Treatment in Cisplatin-Resistant Ovarian Cancer Cell Lines

Abstract

Most ovarian cancer patients develop recurrent cancers which are often resistant to commonly employed chemotherapy agents, such as cisplatin. We have previously shown that the inhibition of heat shock protein 27 (HSP27) or fatty acid oxidation (FAO) sensitizes cisplatin-resistant ovarian cancer cell lines to cisplatin and dual inhibition of both HSP27 and FAO induces substantial cell death in vitro. However, it is unclear how HSP27 and FAO promote cisplatin resistance, and if dual inhibition of both HSP27 and FAO would augment cisplatin treatment in vivo. Here we showed that HSP27 knockdown in two cisplatin-resistant ovarian cancer cell lines (A2780CIS and PEO4) resulted in more ROS production upon cisplatin treatment. HSP27-knockdown cancer cells exhibited decreased levels of reduced glutathione (GSH) and glucose6phosphate dehydrogenase (G6PD), a crucial pentose phosphate pathway enzyme. ROS depletion with the compound N-acetyl cysteine (NAC) attenuated cisplatin-induced upregulation of HSP27, FAO, and markers of apoptosis and ferroptosis in cisplatin-resistant ovarian cancer cell lines. Finally, inhibition of HSP27 and FAO with ivermectin and perhexiline enhanced the cytotoxic effect of cisplatin in A2780CIS xenograft tumors in vivo. Our results suggest that two different cisplatin-resistant ovarian cancer cell lines upregulate HSP27 and FAO to deplete cisplatin-induced ROS to attenuate cisplatin's cytotoxic effect.

Keywords: FAO; HSP27; ROS; cisplatin resistance; ovarian cancer.

Figure 1

HSP27 knockdown increases cisplatin-induced ROS in cisplatin-resistant ovarian cancer cells. (A,B) Flow cytometry of A2780CIS and PEO4 cells treated with vehicle (Veh) or 16 μM cisplatin (CIS) and then stained with MitoSOX dye. Quantification of MitoSOX Mean Fluorescence Intensity (MFI) was shown to right. (C) Western blot of GPX4 (normalized to GAPDH) in A2780CIS cells treated with vehicle (Veh) or cisplatin (16 or 32 μM). (D,E) Flow cytometry of A2780CIS (D; N = 4) or PEO4 (E) cells. Scrambled control or HSP27 knockdown cells were treated with vehicle (Veh) or 16 μM CIS and stained with MitoSOX. Quantification of relative MitoSOX MFI on graphs to right. Unpaired two-tailed two-sample t-test (for A,B) or one-way ANOVA (For C–E). Data are presented as mean ± SD. N = 3 unless otherwise indicated. *, p < 0.05. Uncropped Western blot images are displayed in Figure S12.

Figure 2

ROS depletion by NAC attenuates cisplatin-induced HSP27 expression and cell death (A,B) Western blot of total HSP27, GPX4, or cleaved caspase-3 (CC3) in A2780CIS (A) or scrambled control (scrm) and HSP27-knockdown A2780CIS cells (B) treated with DMSO (Veh), 16 µM cisplatin (CIS), 2.5 mM NAC, or combination (CIS + NAC). Quantification of band intensities (normalized to GAPDH) on the right. One-way ANOVA. Mean ± SD. *, p value < 0.05. N = 3. NS, not significant. Uncropped Western blot images are displayed in Figure S14.

Figure 3

HSP27-knockdown cancer cells exhibit decreased levels of reduced thiols and G6PD. (A,B) Flow cytometry of scramble control (Scrm) or HSP27-knockdown A2780CIS (A) and PEO4 (B) cells treated with vehicle (Veh) or 16 μM cisplatin (CIS) and stained with ThiolTracker. Quantification of relative ThiolTracker MFI on the right. (C) Flow cytometry of A2780CIS cells treated with vehicle (Veh) + CIS or 1.5 μM ivermectin (IVM) + CIS and assessed with ThiolTracker. Quantification of ThiolTracker MFI on the right. (D,E) Western blot of G6PD in A2780CIS-scrm or A2780CIS-shHSP27 cells treated with 16 µM CIS. Quantification of G6PD (normalized to GAPDH) on the right. One-way ANOVA. Mean ± SD. *, p value < 0.05. N = 3. Uncropped Western blot images are displayed in Figure S15.

Figure 4

HSP27 knockdown increases lipid peroxidation in cisplatin-treated cells, and NAC attenuates cisplatin induction of FAO markers. (A) Flow cytometry of scramble control or HSP27 knockdown A2780CIS cells treated with vehicle (Veh) or 16 μM Cisplatin (CIS) and then assessed with Image-iT Lipid peroxidation dye. Graph shows quantification of 510 nM to 590 nM ratio emitted from lipid peroxidation dye. (B,C) Western blot of CD36 in A2780CIS (B) cells and CPT1A in A2780CIS (B) and PEO4 (C) cells treated with DMSO (Veh), 16 µM or 32 µM cisplatin (CIS), 2.5 mM NAC, or combination (CIS + NAC). Quantification of CD36 and CPT1A (normalized to GAPDH) on the right. (D) Western blot of CPT1A and HSP27 in scramble control and HSP27-knockdown A2780CIS cells treated with 16 µM CIS or CIS + NAC. Quantification of CPT1A (normalized to GAPDH) on the right. One-way ANOVA. Mean ± SD. *, p value < 0.05. NS, not significant. Uncropped Western blot images are displayed in Figure S16.

Figure 5

Dual inhibition of HSP27 and FAO augments the antitumor effect of cisplatin in vivo. (A) A schematic depiction of A2780CIS xenograft experiment. Once subcutaneous tumors are formed 12–18 days after tumor cell injection, mice were distributed to five groups and treated with: vehicle (Veh), 3 mg/mL cisplatin (CIS), 3 mg/mL cisplatin and 3 mg/mL ivermectin (CIS + IVM), 3 mg/mL cisplatin and 8 mg/mL perhexiline (CIS + PER), or all three drugs (CIS + IVM + PER). (B) Relative tumor volume (measured in cm³, normalized to starting tumor volume at day 1 of the treatment) over the course of treatment. N = number of tumors. One-way ANOVA. (C) Representative images of A2780CIS xenograft tumors from each treatment group. (D) Wet weights of excised subcutaneous tumors (in grams (g)) after 6 treatment cycles. Each dot indicates an individual tumor. (E) Western blot of GPX4 and cleaved caspase-3 (CC3) in A2780CIS xenograft tumors. Quantification of markers normalized to GAPDH is shown below. One-way ANOVA. Mean ± SD. *, p value < 0.05. ns, not significant. Uncropped Western blot images are displayed in Figure S17.

Figure 6

Differential expression of HSP27 and CPT1A in A2780CIS xenograft tumors. (A) Western blot of CTP1A, pHSP27 (S78), and total HSP27 in A2780CIS xenograft tumors. Quantification of markers normalized to GAPDH on the right. One-way ANOVA. Mean ± SD. *, p value < 0.05. ns, not significant (B) Representative immunofluorescence images of DAPI, CPT1A, and HSP27 in vehicle or cisplatin-treated A2780CIS xenograft tumors. Scale bars = 100 μM (C) Proposed model of how dual inhibition of HSP27 and FAO promotes the antitumor effect of cisplatin. Images created with Biorender. Uncropped Western blot images are displayed in Figure S17.