Heat Shock Factor 1 Depletion Sensitizes A172 Glioblastoma Cells to Temozolomide via Suppression of Cancer Stem Cell-Like Properties

We first confirmed the effect of BIS depletion on HSF1. A172 glioblastoma cells were transfected with 100 nM BIS small interfering RNA (siRNA) for 48 h and cultured under SP-forming conditions, which we established in a previous study [15]. As shown in Figure 1A,B, BIS depletion was verified via Western blotting compared with control siRNA (si-CTL) in both the standard monolayer (ML) and SP-forming culture conditions. Interestingly, protein expression of HSF1 was not altered in cells of the standard ML but was significantly downregulated in SPs following BIS depletion. To examine if the downregulated HSF1 protein level was due to mRNA level changes, we performed quantitative real-time polymerase chain reaction (qRT-PCR). As shown in Figure 1C, HSF1 mRNA was slightly increased in SP-forming conditions and BIS depletion did not alter HSF1 mRNA. We observed similar results in U87-MG glioblastoma cells (Figure S1A,B).

Next, we examined whether downregulation of BIS affected the subcellular localization of HSF1 in SPs with nuclear and cytosolic fractionation. Each fraction was verified with the nuclear and cytosolic markers Lamin B1 and β-actin, respectively. Consistent with our previous data, nuclear SOX2 was decreased with siBIS treatment. When BIS was knocked down with siRNA in A172 cells, nuclear translocation of HSF1 was significantly suppressed compared with control SPs (Figure 1D,E). Similarly, we observed decreased nuclear localization of HSF1 in BIS-depleted SPs via confocal microscopy (Figure 1F).

Concomitant with increased HSF1 expression, we observed that BIS expression was also increased under SP-forming conditions [16]. Based on previous reports in which BIS was reported as a target gene of HSF1 [8,9,10], as well the data presented in Figure 1, we inquired whether the interaction between HSF1 and BIS may be involved in the regulation of glioblastoma stemness. We first investigated whether HSF1 was altered under SP-forming culture conditions, in which reactive oxygen species (ROS) as well as mRNA and protein expression of SOX2 are increased [15]. We compared HSF1 protein levels between the standard ML and SP-forming culture conditions for 72 h in A172 glioblastoma cells. Consistent with a previous report [15], SOX2 protein was increased in SP-forming conditions. Protein expression of HSF1, as well as its target gene HSPA8, was also increased in the SP-forming condition (Figure 2A). Following siRNA knockdown of HSF1 for 48 h, cells were transferred to ML or SP-forming culture conditions. We verified that HSF1 was depleted; qRT-PCR showed that SOX2 mRNA was decreased in HSF1-depleted SPs (Figure 2B,C). Furthermore, sphere forming ability was suppressed by HSF1 depletion (Figure 2D,E). Since cancer stem cells were reported to display a high potential for epithelial–mesenchymal transition (EMT), which is subsequently linked to invasive phenotype, we checked matrix metalloprotease (MMP) activity as a critical player for EMT [17,18,19]. Zymography using supernatant from sphere incubation indicated HSF1 knockdown led to reductions in MMP2 activity, suggesting that HSF1 may play a role in the EMT which is might be associated with GSC-like phenotype (Figure 2D,F).

Next, we investigated whether representative HSF1 target genes, such as HSP, were altered by siRNA depletion of HSF1 or BIS. qRT-PCR results indicated that HSPs, such as HSPA8 and HSP27, were significantly upregulated with siHSF1 in SP-forming conditions (Figure 3A–D). Consistent with this, BIS depletion did not decrease transcription of HSP genes (Figure 3E–G). These results are similar to data obtained with the glioblastoma cell line U87 under the same conditions (Figure S1D–F).

TMZ is used for glioblastoma treatment, although resistance often develops. Hence, we investigated TMZ resistance of SPs expressing increased SOX2 compared with ML cells, which did not express SOX2. ML or SP cells were treated with increasing concentrations of TMZ (0, 167, 334, and 668 μM) for 48 h and chemosensitivity was measured using a viability assay kit to compare the cellular response to TMZ. As shown in Figure 4, A172 cells under SP condition were less sensitive to TMZ compared with the ML cells.

The results presented in Figure 2 and Figure 4 suggested that high levels of HSF1 may be important in TMZ resistance, raising the possibility that downregulation of HSF1 may sensitize the response to TMZ. To investigate this, we pretreated cells with 50 nM siHSF1 for 48 h, trypsinized, and transferred the cells to SP-forming culture conditions. The SPs were then exposed to different concentrations of TMZ (0, 167, 334 and 668 μM). After 24 to 48 h of treatment, SPs were imaged with an inverted microscope and analyzed with a viability assay, SP-formation assay, and Western blotting. HSF1 depletion suppressed SP formation and viability upon TMZ treatment (Figure 5A–C). SOX2 expression was decreased by TMZ in a concentration dependent manner, which was further decreased by HSF1 depletion (Figure 5D). HSF1 silencing also enhanced the cleavage of poly (ADP-ribose) polymerase (PARP), a marker of apoptosis, compared with TMZ treatment only (Figure 5D,E). Next, supernatant from the SP media was analyzed with zymography. We found that HSF1 depletion inhibited MMP2 activity; however, the inhibition was not increased when HSF1 depletion was combined with TMZ treatment (Figure 5F). Consistent with this, treatment with BIS siRNA and TMZ also significantly induced apoptosis concomitantly with suppression of CSC-like properties (Figure S2). Therefore, these results indicate that BIS and HSF1 depletion enhances the sensitivity to TMZ treatment in GSCs.

A172 and U87 human glioblastoma cells, A549 human lung cancer cells and HEP2 human laryngeal cancer cells were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA) and maintained in DMEM (Hyclone, Logan, UT, USA) contained with 10% heat-inactivated fetal bovine serum (FBS), 100 units/mL penicillin and 100 mg/mL streptomycin at 37 °C in 5% CO₂ atmosphere. For the sphere-forming assay, a single-cell suspension following trypsinization was cultured in B27-supplemented DMEM/F12 (Cellgro, Manassas, VA, USA) with epidermal growth factor and basic fibroblast growth factor (10 ng/mL each: R&D Systems, Minneapolis, MN, USA) without serum on ultralow attachment plates (Corning, Tewksbury, MA, USA) at a density of 1 × 10⁵ cells/mL as described previously [17]. Knockdown was performed by transfection of specific 100 nM siRNA G-fectin (Genolution Pharmaceuticals, Seoul, Korea) according to manufacturer’s instruction. siRNAs for control (si-CTL: 5′-CCUACGCCACCAAUUUCGU-3′), and BIS (si-BIS: 5′-AAGGUUCAGACCAUCUUGGAA-3′) were purchased from Bioneer (Daejeon, Korea). Specific siRNA targeted for HSF1 (si-HSF1: 5′-CUGAAGAGUGAAGACAUAAAGA-3′) was purchased from Genolution Pharmaceuticals. Temozolomide was purchased from Sigma-Aldrich (St. Louis, MO, USA) and dissolved in dimethyl sulfoxide (Sigma-Aldrich).

Spheres after 3–5 days were attached to standard culture plates in media containing 5% FBS stained with crystal violet solution (Sigma-Aldrich, St. Louis, MO, USA). For morphological examination, pictures were taken under the inverted microscope.

Cells were lysed with lysis buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris–HCl pH 8.0) with protease inhibitor (Roche Diagnostics, Mannheim, Germany) on ice for 30 min. Equal amounts of protein were separated on 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes (GE Healthcare Life Sciences, Buckinghamshire, UK). The membranes were incubated for 1 h with 5% dry skim milk in TBST (20 mM Tris, 137 mM NaCl, 0.1% Tween 20) buffer and then incubated with antibodies against BIS [6], HSF1, HSPA8 (Enzo Life Sciences, Farmingdale, NY, USA), SOX-2 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), cleaved PARP (Abcam, Cambridge, UK) and β-actin (Sigma-Aldrich). After incubation with horseradish peroxidase-conjugated anti-mouse, anti-rabbit or anti-goat IgG (1:5000; Santa Cruz Biotechnology), the immunoreactive bands were visualized by an enhanced chemiluminescence substrate (Thermo Fisher Scientific, Waltham, MA, USA). Quantification for the intensities of each band was carried out on Image J software (NIH, Bethesda, MD, USA).

Cells were incubated in buffer A (10 mM HEPES, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol (DTT) and 0.05% NP40, pH 7.9) for 10 min on ice. After centrifugation at 3000 rpm for 10 min, supernatant was utilized as a cytosolic fraction. Pellet was suspended in buffer B (5 mM HEPES, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT, 300 mM NaCl and 26% glycerol, pH 7.9), homogenized 20 times with a plastic pestle homogenizer, and incubated on ice for 30 min. After centrifugation at 13,200 rpm for 30 min, supernatant was used as a nuclear fraction. Lamin B1 (Santa Cruz Biotechnology) was utilized as a nuclear marker protein.

Spheres attached on poly-l-ornithine-treated plates were fixed for 10 min with 3.7% formaldehyde in phosphate buffered saline (PBS) and permeabilized for 10 min at room temperature with 0.2% Triton X-100. The spheres were blocked for 1 h with 1% bovine serum albumin in PBS at room temperature. Specific primary antibody for HSF1 was added to the cells, and incubated overnight at 4 °C. The cells were washed in PBS and incubated for 1 h at 4 °C with fluorescent Texas Red—conjugated anti rabbit IgG (Santa Cruz Biotechnology). After washing, the cells were stained with 0.4 mg/mL 40, 6-diamidino-2-phenylindole (DAPI, Sigma) for 10 min at room temperature. Fluorescent images were acquired via confocal microscopy (Zeiss, Oberkochen, Germany).

Total RNA was isolated using AccuZol (Bioneer), and first strand of cDNA was synthesized by reverse transcribed to cDNA using a ReverTra Ace qPCR kit (Toyobo, Osaka, Japan). A quantitative real-time PCR was performed to validate the expression level using SYBR^® Premix Ex TaqTM (TaKaRa Bio, Shiga, Japan) with specific primers (Table 1) on Applied Biosystems 7300 machine (Applied Biosystems, Carlsbad, CA, USA). The relative values for specific mRNA were calculated after normalization to the Ct value from β-actin in the same sample using the DDCt method.

Gelatin zymography was performed using supernatants from sphere forming culture media as previously described [38].

Cell proliferation was assessed as a function of metabolic activity using an EZ-Cytox Cell Viability Assay kit (ItsBio, Seoul, Korea) as mentioned in the previous study [17]. EZ-Cytox Cell Viability assay is based on reduction of tetrazolium chloride to the water-soluble formazan by succinate-tetrazolium reductase, which forms part of the mitochondrial respiratory chain.

The Student t-test was used to compare the differences between two groups. Each experiment was repeated at least three times and p-values of <0.05 were considered as statistically significant.