Trifluoperazine, a novel autophagy inhibitor, increases radiosensitivity in glioblastoma by impairing homologous recombination

Human glioma cell lines U87 and U251 were purchased from the cell bank of the Chinese Academy of Sciences and were cultured in Dulbecco’s modified Eagle’s medium (ThermoFisher Scientific; Waltham, MA, USA) containing 10% fetal bovine serum (ThermoFisher Scientific), glutamine (4 mM), penicillin (10 U/mL), and streptomycin (100 mg/mL). Normal human astrocytes (NHA) were purchased from Lonza (Walkersville, MD, USA) and were cultured in Astrocyte Medium BulletKit (Lonza) according to the manufacturer’s instructions. P3 is a primary GBM cell line isolated from a patient biopsy. The P3 tumor has the following molecular characteristics (+ [Chr 7, Chr19, 20q], −[1q42-q43, Chr9, Chr10, 20p] --[PIK3R1, CDKN2A/B]. P3 cells were cultured in Neurobasal Medium (ThermoFisher Scientific) containing penicillin (10 U/mL), streptomycin (100 mg/mL), B27 supplement (20 μL/mL), FGF (20 ng/mL), EGF (20 ng/mL) and heparin (32 IE/mL).

The cytotoxic effect of TFP (Sigma; St. Louis, MO, USA) on the GBM cell lines, P3 and NHA cells was determined using the CCK-8 assay (Dojindo; Kumamoto, Japan). Cells were suspended in DMEM with 10% fetal bovine serum (FBS) or Neurobasal Medium (for P3) and seeded into 96-well, flat-bottomed plates (5 × 10³ cells/well). After incubation overnight at 37 °C, cells were pretreated with PBS or TFP (0–30 μM). After 24 h or 48 h of culture, cells were incubated for an additional 2 h at 37 °C with 100 μL of serum-free DMEM or Neurobasal Medium (for P3 cells) containing 10 μL of CCK-8, and absorbance was measured at a wavelength of 450 nm using a microplate reader (BioRad; Hercules, CA, USA).

The tumor cells (2.5 × 10⁴ cells/well) were seeded into 24-well, flat-bottomed plates. After 24 h, cells were treated with PBS, 5 and 10 μM of TFP for an additional 48 h in DMEM with 10% serum, and subsequently stained with EdU using the Apollo Detection Kit (Ribobio; Guangzhou, China) according to the manufacturer’s instructions. EdU positive cells were counted from at least 10 random fields by fluorescence microscopy (Leica DMi8; Leica Microsystems, Wetzlar, Germany).

After treatment of different doses of TFP, 100 nM bafilomycin A1, 2.5 μM rapamycin, 4 Gy radiation at a dose rate of 1.8 Gy/min using a linear accelerator (Primus Hi; Siemens Medical Instruments; Erlangen, Germany) or 5 μM TFP for 24 h before receiving one dose of 4 Gy, whole-cell protein extracts (20-50 mg) were prepared using a radioimmunoprecipitation assay buffer (RIPA; Thermo Fisher Scientific) supplemented with protease inhibitor cocktail (Cell Signaling Technology; Beverly, MA, USA). Proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane. Membranes were blocked with 5% skimmed milk in Tris-buffered saline containing 0.1% Tween-20, and subsequently incubated with primary and indicated secondary antibodies (Thermo Fisher Scientific). Proteins on western blots where visualized using the Chemiluminescent Reagents Kit (Millipore, Billerica, MA, USA). Chemiluminescent signals were detected with the ChemiDoc XRS+ (Bio-Rad, Hercules, CA, USA) and quantified using Image Lab 3.0 software (Bio-Rad). The following primary antibodies were used for western blotting: beta-tubulin, LC3BI/II, phospho-histone H2A.X (Ser139; also known as γ-H2A.X), P62 and survivin (Cell Signaling Technology; Beverly, MA, USA); GAPDH, BRCA1, BRCA2 and Rad51 were purchased from Santa Cruz (Dallas, TX, USA).

Cells were transiently transfected with the pSELECT-GFP-LC3 plasmid (Genepharma; Shanghai, China) using Lipofectamine 2000 reagent (ThermoFisher Scientific) according to the manufacturer’s instructions. After being treated with PBS or 10 μM TFP for 24 h, cells were observed using a Leica TCS SP5 Confocal Laser Scanning Microscope (Leica Microsystems) and GFP-LC3 puncta per cell was counted. Ten random fields were obtained per treatment group.

Cells were fixed in 3% glutaraldehyde in PBS for 2 h, washed five times with 0.1 M cacodylate buffer, and post-fixed with 1% OsO₄ in 0.1 M cacodylate buffer containing 0.1% CaCl₂ for 1.5 h at 4 °C. Cells were dehydrated in graded alcohol series and embedded in epoxy resin. Ultrathin sections were cut and stained with uranyl acetate and lead citrate. Images were obtained using a JEM-1200EX II electron microscope (JEOL, Tokyo, Japan).

Following treatment with PBS (control), 100 nM bafilomycin A1 and 10 μM TFP for 48 h, U251 and U87 cells were rinsed 3 times with fresh medium, and Lyso-Tracker Red (diluted in DMEM with 10% FBS; ThermoFisher Scientific) was added to a final concentration of 66 nM. Cells were incubated for 30 min at 37 °C and rinsed with phosphate-buffered saline (PBS). Nuclei were stained with 5 μg/ml Hoechst 33,342 (ThermoFisher Scientific), and live cells were observed using a Leica DMi8 fluorescence microscope.

Comet assays were performed according to a previously described protocol [17]. Briefly, after treatment with PBS, TFP, radiation or combination treatment, cells were thoroughly mixed with low melting point agarose solution. Radiation treatment was carried out with a single dose of 4 Gy at a dose rate of 1.8 Gy/min using a linear accelerator (Primus Hi). The cell suspension was spread on a Comet Slide (CometAssay® Kit, Trevigen; Gaithersburg, MD, USA) covered with 1.5% normal melting agarose. Slides were immersed in prepared lysis solution, treated with Tris-EDTA buffer (10 mM TrisCl, pH 7.5, 1 mM EDTA), and then placed horizontally on an electrophoresis tray filled with alkaline solution (300 mM NaOH, 1 mM EDTA). Electrophoresis was conducted at room temperature with an electrical field of 25 V and a current of 300 mA for 20 min. After electrophoresis, the slides were stained with GelRed (Biotium; Fremont, CA, USA). Slides were examined under fluorescence microscopy. Cells were analyzed using the Comet Assay Software Project (CASP). Olive tail moment (OTM) was used to quantify the extent of DNA damage.

Immunofluorescence detection of γ-H2A.X foci was performed to monitor formation of DNA double strand breaks (DSBs). Cells cultured on coverslips were treated with PBS or 5 μM of TFP before receiving one dose of 4 Gy at a dose rate of 1.8 Gy/min using a linear accelerator (Primus Hi). At indicated time points (2, 6, 12 and 24 h), the cells were rinsed with PBS and then fixed in 4% paraformaldehyde before permeabilisation with 0.3% Triton X-100. After blocking with 5% BSA (Sigma), cells were incubated with diluted primary antibody for γ-H2A.X overnight at 4 °C, followed by staining with Fluorescein (FITC)-conjugated goat anti-rabbit IgG (ThermoFisher Scientific). Finally, the samples were mounted in mounting medium containing DAPI (ThermoFisher Scientific). Three random fields were examined at a magnification of ×63 by a Leica TCS SP5 Confocal Laser Scanning Microscope.

U251 and U87 cells (3 × 10³ cells/well) were plated in six-well plates. The adherent cells were then treated with PBS or 5 μM TFP for 24 h before receiving one dose of 4 Gy at a dose rate of 1.8 Gy/min whereupon they were incubated for 14 days. Then colonies were washed with PBS, fixed in 4% paraformaldehyde, stained with 0.1% crystal violet and counted. Colonies consisting of more than 50 cells were counted as surviving colonies.

Apoptosis was measured by quantifying cleaved-caspase 3 and 7 activity using Cell Event Caspase 3/7 Green Detection Reagent (Life Technologies, Carlsbad, CA, USA). Cells were plated in 96-well plates with a density of 3000 cells per well and allowed to adhere. Thereafter, the cells were treated with PBS or 5 μM TFP for 24 h before receiving one dose of 4 Gy at a dose rate of 1.8 Gy/min in a linear accelerator (Primus Hi). Then the cells were inspected using an IncuCyte Zoom live cell imaging system (Essen BioScience, Ann Arbor, MI, USA).

An HR assay (Norgen Biotek, Thorold, ON, Canada) was performed on U251 and U87 cells according to manufacturer’s instructions. Briefly, at day 3 after TFP treatment, cells were transfected with a positive control plasmid or two HR dl plasmids (dl-1 and dl-2). After 24 h of transfection, DNA was isolated using the Wizard genomic DNA purification kit (Promega, Madison, WI, USA). qPCR was performed with the supplied primers using a Roche LightCycler 480 II (Roche Applied Science, Indianapolis, IN, USA).

Activity of cathepsin B and L in U251 and U87 cells was tested using a Fluorometric Assay Kit (Abcam, Cambridge, UK) according to the manufacturer instructions. Briefly, after treatment with PBS (control) or 5 μM TFP for 24 h, the cells were lysed and supernatants were incubated with cathepsin-B (Ac-RR-AFC) or L (AC-FR-AFC) substrates at 37 °C for 1.5 h. Then samples were measured on a fluorescent microplate reader at excitation/emission wavelength = 400/505 nm. After subtracting the background control (buffer) from sample readings, activity of cathepsin B and L was determined by comparing results from TFP treated cells with the level from controls.

U251 and U87 cells were transfected with siRNA twice at a 24-h interval with lipofectamine 2000. The final concentration of siRNAs was 50 nM. Sequences for the siRNAs used were the following: cathepsin L, 5′-GATGCACAACAGATTATACTT-3′; nontargeting siRNA controls, 5′-UUCUCCGAACGUGUCACGUTT-3′ (Genepharma). Western blot analysis was used to assess the downregulation of cathepsin L.

All animal protocols were approved by the ethics committee at the Shandong University (Jinan, China) and conducted according to the national regulations in China. For implantations, nude mice were anesthetized with 4% chloral hydrate (300 mg/kg) and placed in a stereotactic frame. Using aseptic surgical procedures, an incision was made in the parietal scalp, and a small burr hole was drilled 2.5 mm lateral to the bregma. U251 and P3 cells (1 × 10⁶ cells/mouse) were implanted 2.0 mm into the right striatum using a Hamilton syringe (Hamilton Co., Reno, NV, USA). Two weeks later, mice were randomly divided into four groups (6 mice/group). Groups 1 and 2 were injected intraperitoneally (IP) with PBS or TFP (1 mg/kg, 5 days/week). Group 3 was given three doses of localized irradiation (5 Gy) at days 15, 20, and 25 after implantation following IP injection of PBS. Group 4 was irradiated three times following IP injection of TFP. Mice were sacrificed when central nervous system symptoms (such as poor ambulation, lethargy, hunched posture) or weight loss > 20% body mass developed. The mice were anesthetized with chlorohydrate and perfused transcardially with 4% paraformaldehyde in PBS. Whole brains were removed, post-fixed overnight in 4% paraformaldehyde in PBS, coronally sectioned into 5 slices, and paraffin embedded. Tissue sections were cut (10 μm) and incubated with primary antibodies as indicated. The following primary antibodies were used for immunohistochemistry: Ki67 (Abcam); γ-H2A.X (Cell Signaling Technology), and Rad51 (Santa Cruz).

The cytotoxic effects of TFP on tumor cells in vitro were determined using the CCK-8 assay (Fig. 1a, b). The IC50 values of TFP for U251, U87 and P3 cells were 16 μM, 15 μM, 15.5 μM, respectively. GBM cells were significantly more sensitive to TFP compared to normal human astrocytes (IC50 22.5 μM, P < 0.05). These results were further substantiated by the use of a proliferation assay where EdU was incorporated in U251 and U87 cells. The number of EdU positive cells decreased in a dose dependent manner in both cell populations (Fig. 1c, d). These results were further confirmed by a reduced clonogenic ability and increased caspase activity following TFP treatment (Fig. 5c, d).

Previous studies have suggested that TFP induces autophagy [18]. LC3B, a membrane component of autophagosomes, is used as a marker for the induction of autophagy. Analyses of LC3B by western blot showed increased levels after upon TFP treatment (Fig. 2a). To visualize the development of autophagosomes following TFP treatment, we transiently transfected GBM cells with a GFP-LC3 expression construct. Confocal microscopy showed an increase in fluorescence puncta in TFP treated cells after 24 h (Fig. 2b). Transmission electron microscopy (TEM) represents the gold standard for detecting autophagic vacuoles (AVs). More AVs were detected in TFP treated cells compared to the control group (Black arrows, Fig. 2c).

Autophagy is a dynamic multistage process where the assembly of autophagosomes is an early event. The assembly of autophagosomes is therefore not directly connected to autophagosome degradation. Thus, an accumulation of AVs in cells can occur either as a consequence of an increased autophagosome formation or decreased degradation [19]. The levels of SQSTM/p62, one of the most important long-lived proteins critical for autophagy, accumulated in a dose-dependent manner in response to TFP (Fig. 2d). This indicates that TFP blocks autophagy.

Rapamycin, which induces autophagy by inhibition of mTOR [20], and Bafilomycin A1 (BAF), which blocks autophagy by selectively inhibiting H⁺-ATPase [21], were used as positive controls. Our results show that p62 levels decreased in rapamycin treated cells whereas it was clearly increased in cells treated with TFP or BAF (Fig. 2e).

BAF blocks autophagy flux by reducing the acidification of lysosomes. Altered lysosome function prevents fusion of autophagosomes with lysosomes. LysoTracker Red, which labels highly acidic lysosomal vacuoles and thus detects activity of vacuolar H⁺-ATPase (v-ATPase) [22], was used to determine the status of lysosomes after TFP or BAF treatment. Under fluorescence microscopy, lysosomes could not be detected with LysoTracker Red in TFP and BAF treated cells (Fig. 3a). These results indicate that TFP blocks autophagy by inhibiting the acidification of lysosomes.

A number of studies have demonstrated that inhibitors of autophagy enhance the sensitivity of cells to radiation therapy [6, 8, 23]. To investigate whether TFP affected radiosensitivity, we used the comet assay to measure DNA integrity in U251 and U87 cells. TFP or radiation treatment alone slightly increased the levels of DNA tails above controls, whereas combined treatment caused a significant increase in the appearance of DNA tails (Fig. 3b, c). Protein levels of γ-H2AX, a gold standard to detect the presence of DSBs [24], were correspondingly increased following combination treatment (Fig. 3d).

We next quantified γ-H2AX foci at different time points after radiation treatment or TFP combined with radiation using immunofluorescence. Radiation induced more γ-H2AX foci after 2 h of treatment. This was reduced considerably from 6 h and onwards. However, TFP treatment resulted in a significant prolongation of the γ-H2AX signal at least 24 h post irradiation in U251 and U87 cells (27% and 21.6%, respectively) compared with radiation alone (10%, P < 0.01; 2.3%, P < 0.05. Figure 4a, b). Radiation induces apoptosis mainly by generating DSBs. DSBs can be repaired by either non-homologous end joining (NHEJ) or homologous recombination (HR). In NHEJ, ligation occurs regardless of whether the ends come from the same chromosome. Thus loss of genetic information and translocations might occur [25]. HR uses the information that is contained in genetically identical, or almost identical, DNA molecules (usually the sister chromatid) to repair damaged DNA, and therefore, has a higher accuracy of maintaining DNA integrity [26]. The DNA repair protein Rad51 polymerizes onto resected DNA ends to form a nucleoprotein filament and promotes strand exchange between homologous DNA duplexes. As such, Rad51 plays a central role in HR and is crucial for the stability of the genome and the normal cell cycle [27]. After treatment of TFP, we found that the expression of Rad51 and the associated DNA repair proteins BRCA1 and BRCA2, decreased, while γ-H2AX increased in a dose-dependent manner (Fig. 5a). In addition, HR efficiency was calculated using a PCR-based HR assay kit. We found that after TFP treatment, the HR efficiency decreased significantly (P < 0.05) compared to the control group (Fig. 5b). TFP might therefore lead to an increase in DSBs and radiosensitivity by inhibiting the expression of central DNA repair proteins.

Lysosomes play an important role during autophagy and among all the lysosomal proteases, cathepsins have multiple roles, not only in the degradation of nonfunctional organelles, but also in the process of tumorigenesis [28]. Earlier studies have also shown that proteasomes as well as lysosomes can contribute to cellular tolerance to various proteotoxic stressors, and confer resistance to chemo-, radio- and immunotherapy [29]. We therefore wanted to assess a putative role of lysosomes as a link between autophagy and DNA damage repair following TFP treatment. Moreover, impaired acidification of lysosomes might affect the activity of cathepsins, which is optimized for low pH [30]. We first examined the enzymatic activities of two main cathepsins (B and L), in response to TFP. With a fluorometric-based kit, we found that after TFP treatment, the activity of cathepsin B and L decreased significantly (Fig. 6a). By western blot analysis, we also found that cathepsin B heavy chain (the mature form of cathepsin), procathepsin B and cathepsin L decreased in a dose-dependent manner (Fig. 6b). Previous studies have shown that cathepsin L is not only involved in autophagy [31‐33], but also in radioresistance [34, 35]. After knocking down cathepsin L, we found that the expression of P62 and γ-H2AX increased, while Rad51 decreased (Fig. 6c), which was in consistent with the results observed after TFP treatment. Finally, HR efficiency also decreased significantly after knocking down cathepsin L (Fig. 6d). In conclusion, these data suggest that down-regulation of cathepsin L might explain the radiosensitivity effect of TFP.

To evaluate the antitumor effect of TFP in combination with radiation in vivo, we established orthotopic xenograft models with U251 and P3 cells (Fig. 7a). The survival curves demonstrated no survival benefit from radiation treatment alone (median survival 24.8 and 29.7 days, −control vs radiation respectively, P > 0.05), while TFP significantly prolonged survival (median survival 24.8 days vs 33.8 days, control vs TFP respectively). Combined treatment with TFP and radiation significantly increased the overall median survival of animals and produced some long-term survivors (median survival 46 days vs 29.7 days, radiation vs combined treatment, P < 0.01; median survival 46 days vs 33.8 days, TFP vs combined treatment, P < 0.05. Fig. 7b). Analyses of Ki67 positive cells by immunohistochemistry from TFP treated sections and the combination group showed a dramatic decrease in tumor cell proliferation compared to the control and radiation group (Fig. 7c). Sections stained for γ-H2AX and Rad51 showed that in the TFP treatment group, the expression of Rad51 decreased compared with controls while in the radiation treatment group it was increased. In the combined treatment group, a significantly reduced fraction of cells was positive for Rad51 compared to controls. Animals treated with TFP and radiation displayed a significant increase in γ-H2AX staining compared with either treatment alone (Fig. 7c). In summary, our data demonstrates that TFP sufficiently penetrates the BBB and increases radiosensitivity in vivo.