TMZ-induced PrPc/par-4 interaction promotes the survival of human glioma cells

Authors


Abstract

Malignant gliomas recur even after extensive surgery and chemo-radiotherapy. Although a relatively novel chemotherapeutic agent, temozolomide (TMZ), has demonstrated promising activity against gliomas, the effects last only a few months and drug resistance develops thereafter in many cases. It has been acknowledged that glioma cells respond to TMZ treatment by undergoing G2/M arrest, but not apoptosis. Here we demonstrate a phase-specific chemotherapy resistance due to cellular prion protein (PrPc) in human glioma cells upon TMZ treatment. TMZ-induced G2/M-arrested cultures show an upregulation of PrPc expression and are more resistant, whereas G1/S-phase cells that show decreased levels of PrPc are more sensitive to apoptosis. Furthermore, an investigation into the biological significance of PrPc association with par-4 provided the first evidence of a relationship between the endogenous levels of PrPc and the resistance of glioma cells to the apoptotic effects of TMZ. Upon TMZ treatment, PrPc exerts its antiapoptotic activity by inhibiting PKA-mediated par-4 phosphorylation that are important for par-4 activation, nuclear entry and initiation of apoptosis. In context with cell cycle-dependent responses to chemotherapy, the data from this study suggest the possibility of exploiting the PrPc-dependent pathway to improve the efficacy of TMZ-based regimen for patients with gliomas.

Malignant gliomas are the most aggressive brain tumor in adults, accounting for more than half of the primary malignant brain tumors diagnosed each year and having a combined incidence of 5–8/100,000 population. Treatment of malignant gliomas has changed substantially over the last few years. The standard of care, until recently, has involved maximum possible resection, followed by adjuvant irradiation and chemotherapy. However, even with aggressive treatment, the median reported survival is less than 1 year.1 An oral alkylating agent, temozolomide (TMZ), has greatly extended our armamentarium in the treatment of malignant glioma.2 It undergoes rapid chemical conversion in the systemic circulation at physiological pH to the active compound, MTIC (monomethyl triazeno-imidazole carboxamide), inducing methylation in multiple locations on DNA. The methylation of the O6 position of GMP in DNA, although relatively infrequent (6–7%), is usually regarded as the lethality-causing lesion.3, 4

Despite the fact that TMZ contributes the significant therapeutic benefits in glioma patients, its opposite effects that may enhance the malignancy of the treated cancers have also drawn increasing attention. The addition of TMZ to radiotherapy only resulted in a slightly longer median survival time in the newly diagnosed GBM patients, 14.6 versus 12.1 months and the best 5-year survival of 9.8%.5 TMZ has been shown to induce prolonged G2/M arrest in p53-intact glioma cells and transient G2/M arrest in p53-inactivated glioma cells.6 On the one hand, these mechanisms are responsible for TMZ-induced cytotoxicity in vitro. On the other hand, G2/M delay seems to provide more time to repair the potentially lethal DNA lesions and thereby reduce the genotoxic effects of antitumor drugs. In fact, postchemotherapy residual mass frequently lead to treatment failure (i.e., cancer recurrence) in malignant glioma patients. Thus, understanding the cellular mechanisms involved in acquisition of apoptotic-resistant phenotype after TMZ treatment will allow the development of rational molecular-based tumor therapies in gliomas.

Several intriguing lines of evidence have emerged recently suggesting that PrPc, a glycosylphosphatidylinositol-anchored protein expressed predominantly in the brain, may exert a cytoprotective activity to rescue neural cells from pro-apoptotic stimuli and anticancer drug treatments. For example, PrPc interaction with stress-inducible protein 1 (SIT1) induce neuroprotective signals that rescue cells from apoptosis.7 PrPc has also been found to rescue cultured cerebellar granule neurons8 and N2a neuroblastoma cells9 from apoptosis induced by Dpl. In our study, we demonstrate the differential sensitivity of the G1/S- and G2/M-synchronized glioma cells to apoptosis, and suggest a critical role for PrPc in imparting resistance in the G2/M population. In particular, upon TMZ treatment, PrPc interacts with par-4 via SAC domain, the cancer-specific proapoptotic effector domain,10 to inhibit the proapoptotic activities of par-4. These studies suggest PrPc as an effective target for abolishing the opposite effects of TMZ-based regime during clinical treatment against gliomas.

Material and Methods

Cell culture

The human glioma cell lines U251 and U87 (American Type Culture Collection) were used in our study. Each cell line was grown in RPMI medium 1640 (Life Technologies) containing 5% FBS and maintained at 37°C in an atmosphere of 5% CO2 and 95% room air.

Immunoblotting and coimmunoprecipitation

Total protein was extracted from cells using RIPA lysis buffer. For immunoblotting, protein extract (50 μg/lane) were electrophoresed, transferred to PVDF membranes, and incubated overnight with the antibodies against PrPc (Sigma), par-4 (Santa Cruz, CA), p-par-4 (Santa Cruz), cytochrome C (Santa Cruz), bcl-2 (Santa Cruz) and bax (Santa Cruz), respectively. Membranes were treated with the appropriate secondary antibodies (Invitrogen) and analyzed using the Enhanced Chemiluminescence System (Amersham). For coimmunoprecipitation, 1 mg of protein lysates from each sample was incubated with 2 μg of anti-PrPc, or par-4 overnight at 4°C. Then immunoblotting detections were performed as described earlier.

Flow cytometry assay

Flow cytometry was used to quantitatively detect the apoptotic rate as described previously.11

Caspase activity assay

The caspase activities were determined using the ApoAlert Caspase Assay Kit (clontech) as described previously.11

Inhibition of PrPc, par-4 expression by RNA interference

Cells (2 × 105) were seeded in six-well plates and after an overnight incubation, the cells were transfected with various concentrations of siRNA in the serum-free Opti-MEM medium using the HiPerfect Reagent (Qiagen) as suggested by the manufacturer's instructions.

Immunofluorescence microscopy analysis

Cells were grown on glass slides and treated as indicated. The slides were quickly washed with PBS followed by fixing in 100% methanol at −20°C for 10 min. The samples were subjected to probing with the antibodies against PrPc (Sigma) or par-4 (Santa cruz) and the appropriate secondary antibodies. The fluorescence was visualized under the immunofluorescence microscopy (Leica).

Plasmids construction

The plasmids encoding human par-4-GFP, SAC-GFP, HA-par-4 and Flag-PrPc are generous gifts from Dr. Chao Lu (Nanjing University, Nanjing, Jiangsu Province, PRC). ΔSAC-GFP mutant was made by site-directed mutagenesis (QuikChange kit; Stratagene) to delet the SAC domain (amino acids 147–206) of par-4, followed by ligation into the pcDNA3.1/CT-GFP-TOPO vector (Invitrogen).

ShRNA-expressing lentivirus production

Oligonucleotides corresponding to the shRNA sequence of PrPc11 were annealed and cloned into the lentivirus vector, pll3.7 (Addgene). Virus stocks were prepared by cotransfecting with three packaging plasmids (pMDLg/pRRE, CMV-VSVG and RSV-Rev) into 293T cells. The viral supernatants were harvested 48 h later, filtered and centrifuged (90 min at 25,000g). Viral titers were determined by fluorescence-activated cell sorting analysis.

Luciferase reporter assay

Cells were cotransfected with the constant amount of HA-par-4 or ΔSAC-GFP mutant, the luciferase reporter construct carrying bcl-2 promoter together without or with the increasing amounts of Flag-PrPc. Forty-eight hours after transfection, cells were lysed and the luciferase activities were measured by the dual-luciferase reporter assay system (Promega).

GST-pull-down assay

PrPc-Ig Fc protein was incubated with protein A-sepharose beads for 30 min, and subsequently washed and incubated with cell extracts (10 mg) for 18 h at 4°C. Protein eluted from the beads was resolved by SDS-PAGE, and subjected to Coomassie blue staining and subsequent mass spectrometry.

Animal studies

For in vivo treatment studies, 6-week-old NOD/SCID female mice were injected with 1.5 × 106 proliferating U87 cells subcutaneously into the right flank under isoflurane anaesthesia. Approximately 12 days after inoculation, mice developed palpable tumors and were treated with virus containing the shRNA against PrPc or/and TMZ (50 mg/kg for five consecutive). Tumor volume was calculated using the following formula: V = (DL × Dmath image) × π/6. Where DL is the largest diameter and DS is the smallest diameter. All experiments received prior approval from the Huashan Hospital's Research Institute Institutional Animal Use and Care Committee.

Immunohistochemistry

Three-micrometer sections were cut from paraffin blocks onto the silanized slides, and the sections were immunostained using the antibody against PrPc. Scoring of PrPc was graded from − to +++, based on an assessment of the intensity of the reaction product, and the percentage of positive cells: score -, no reactivity or membranous reactivity in <10% of tumor cells; +, faint/barely perceptible reactivity is detected in >10% of tumor cells; ++, weak to moderate complete reactivity is seen in >10% of tumor cells; +++, strong complete reactivity is seen in >10% of tumor cells.

Statistical analysis

Statistics were calculated by SPSS software. The results are presented as mean ± standard error (SEM). ANOVA, Student's t-test analysis and Dunnett's multiple comparison tests were used to compare mean values. A p-value of less than 0.05 was defined as statistically significance.

Results

TMZ-induced G2/M arrest in glioma cells correlates with resistance to drug-mediated apoptosis

After U87 cells were treated with TMZ (100 μM for 6 hr) and incubated for another 48 hr in the absence of TMZ, we found that the cells accumulated at the G2/M boundary which was associated with the gradual loss of cells with 2N DNA content (Fig. 1a). However, the sub-G1 population, which represents the apoptotic cells, was small and did not significantly increase after TMZ treatment (Fig. 1b). We hypothesized that the low rate of apoptosis in the TMZ-treated cells may be attributed to drug-induced G2/M arrest, which plays a key role in the antiapoptotic process.

Figure 1.

Glioma cells arrested in G1/S and G2/M phases display differential sensitivity to the TMZ-induced apoptosis. After U87 cells were treated with TMZ (100 μM for 6 h), cell cycle distribution (a) and the apoptosis rate (b) of the TMZ-treated cells were assessed by flow cytometry. The nonsynchronized cells and cells synchronized at G1/S or G2/M phase were treated with TMZ for the indicated times, after which the apoptosis rates were determined by FCM (c) and caspase activity assay (d), respectively. **p < 0.05 versus corresponding the controls cells (Student's t-test).

To test this hypothesis, U87 cells were synchronized with aphidicolin in the G1/S phase or with nocodazole in the G2/M phase. It was observed that with 6hs of TMZ treatment, the G1/S-arrested cells exhibited 46.82% apoptosis that increased to 87.53% by 12 h in comparison with the unsynchronized controls, which had 2.82% apoptosis at 6 h and 3.04% at 12 h. The G2/M population treated with TMZ showed no significant apoptosis, with an apoptotic population of 3.46% at 6 h and 2.31% at 12 h of treatment (Fig. 1c). In accordance, the TMZ-induced caspase activity was unaltered in the nonsynchronized cells and cells at G2/M phase, but significantly higher in the cells at G1/S phase (Fig. 1d). These results indicated that the G2/M cells are more resistant to TMZ-mediated apoptosis as compared to the G1/S phase cells.

PrPc is expressed in a cell cycle-dependent manner in glioma cells

Since a significant difference in response to TMZ between the G1/S and G2/M cells was observed, we performed transcriptome profiling of the two cell populations which revealed 83 differentially expressed genes. PrPc was further investigated because it was the most significantly changed gene in the G2/M-arrested cells (Fig. 2a). As shown in Figure 2b, in comparison with the unsynchronized cells, G2/M-arrested cells exihibited a significant increase in PrPc expression, whereas PrPc was negligibly expressed in the G1/S-arrested populations. Cell cycle-dependent expression of PrPc was further confirmed in cells by dual staining with antibody specific for PrPc and propidium iodide for DNA. The expression of PrPc could only be detected in the cells at G2/M but not G1/S phase, corroborating the results of immunoblot analysis (Fig. 2c). As TMZ has been proved to induce prolonged G2/M arrest, we found that TMZ induced PrPc expression in a dose-dependent manner in both U87 and U251 cells (Fig. 2d). To establish the clinical relevance of PrPc expression in TMZ-mediated effects, we assessed PrPc in 5 patients who had paired initial and recurrent glioma specimens available from before and after adjuvant TMZ combined with radiotherapy, respectively. As compared to the initial tumors, all recurrent lesions were characterized by increased expression of PrPc. Similar results were not detected, however, in the recurrent patients who just underwent radiotherapy after primary surgery (Fig. 2e and Table 1).

Figure 2.

PrPc is expressed in a cell cycle-dependent manner. (a). Array hybridization analysis of gene expression extracted from U87 cells at G1/S and G2/M phase, respectively. The density of the hybridization signals reflects the relative expression level of particular genes. Expression of PrPc in the unsynchronized cells and cells synchronized at various phases were analyzed by immunoblotting (b) and immunofluorescence (c), Scale bars: 20 μm. (d). Dose effect of TMZ-induced PrPc expression in U87 and U251 cells was evaluated by immunoblotting (e). PrPc-stained sections of gliomas taken from the patients who had paired initial and recurrent glioma specimens available from before and after radiotherapy combined without or with TMZ treatment, respectively. Scale bars: 50 μm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Table 1. TMZ-based therapy enhances PrPc expression in gliomas
inline image

Identification of a novel interaction between PrPc and par-4

To identify the proteins that could conceivably link PrPc signaling to the antiapoptotic phenotype of glioma in vivo, a soluble form of PrPc fused to Ig Fc was immobilized on protein A-sepharose beads and used in a pull-down assay with cell lysates from U87 cells treated with or without TMZ, respectively. After extensive washing, the bound proteins were resolved on SDS-PAGE gels and stained with Coomassie blue. A unique band of about 38 kDa in the lysates of the TMZ-treated cells, which was absent from the lysates of control cells, was pulled down by the PrPc-Fc fusion protein (Fig. 3a). The band was excised, and sequencing identified the protein as par-4.

Figure 3.

Interaction of PrPc with the SAC domain of par-4 protects glioma cells from the TMZ-induced apoptosis. (a). Soluble PrPc-Fc bound to protein-A Sepharose was incubated with the cell lysates from U87 cells treated without or with TMZ, and proteins that bound to PrPc-Fc were resolved on SDS-PAGE and stained with Coomassie blue. A 38 kD cellular protein that interacts with PrPc-Fc is present in the TMZ-treated cell lysates. After U87 cells were treated with TMZ (b) or synchronized at different phases (c), reciprocal coimmunoprecipitation of the PrPc/par-4 complexes from the equal amounts of cell lysates were then used for immunoblotting with the indicated antibodies. (d). U87 cells were transfected with the GFP, SAC-GFP, or ΔSAC-GFP expression constructs in the presence of TMZ, and the cell lysates were subjected to immunoprecipitation with the par-4 or GFP antibody. The immunoprecipitated complexes were then subjected to immunoblotting for PrPc or GFP. Whole-cell extracts from the transfectants were used as input. (e). Upon TMZ treatment, the apoptosis rates in the PrPc- or/and par-4-depleted or/and G2/M-arrested U87 cells were determined by FCM.

Using the TMZ-treated U87 cells as an in vitro model, we performed reciprocal coimmunoprecipitation assays. Figure 3b showed that the PrPc/par-4 complex existed in the untreated cells. TMZ treatment significantly increased the association of PrPc with par-4. Neither of these proteins was detected in controls that had been treated with IgG alone. To further determine whether the increased formation of PrPc/par-4 complex is attributed to TMZ-induced G2/M arrest or direct effects of TMZ treatment, we evaluated the interaction between PrPc and par-4 in cells at different phases. As compared to the unsynchronized cells, the physical association of PrPc with par-4 was significantly augmented in the G2/M-arrested cells, whereas relatively less PrPc/par-4 complex was noted in the cells at G1/S phase (Fig. 3c). These observations indicated that the association between PrPc and par-4 was relatively stable and had been established when the cells were arrested at G2/M phase in response to TMZ treatment.

Previous studies have identified the core domain, designated SAC (selective for apoptosis in cancer cells), as the effector domain of Par-4.10 Although the above data indicate PrPc binds with par-4, it was unclear if the SAC domain and PrPc interact in the same manner. To test the potential interaction between PrPc and the SAC domain, we transfected U87 cells with the GFP, SAC-GFP or ΔSAC-GFP (SAC domain-deleted mutant of par-4) expression constructs in the presence of TMZ, and subjected the cell lysates to immunoprecipitation with either the GFP or par-4 antibody. As shown in Figure 3d, PrPc coimmunoprecipitated with SAC-GFP when we used the par-4 antibody with lysates from the SAC-GFP transfectants. In coimmunoprecipitations using par-4 antibody and lysates from the cells that were transfected with GFP or ΔSAC-GFP, we coimmunoprecipitated PrPc with endogenous par-4. Likewise, when we used the GFP antibody with lysates from the SAC-GFP transfectants, PrPc coimmunoprecipitated with SAC-GFP, whereas coimmunoprecipitations using the GFP antibody with lysates from either the GFP-transfected cells or the ΔSAC-GFP-transfected cells did not yield PrPc (Fig. 3d). Accordingly, the SAC domain is requisite and sufficient for interaction with PrPc.

The PrPc/par-4 complex protects glioma cells from the apoptotic effects of TMZ

We then attempted to determine whether the PrPc/par-4 interaction directly participates in the commitment of cells to the antiapoptotic effects upon TMZ treatment. An immunoprecipitation and immunoblot analysis indicated that PrPc siRNA had reduced the basal levels of PrPc protein by about 80% in the subconfluent cultures of TMZ-treated cells but had no effect on par-4 expression, and vice versa (Supporting Information Fig. S1A). Upon TMZ treatment, PrPc depletion greatly enhanced the apoptosis rate as well as the caspase activities in either the unsynchronized cells or the G2/M-arrested cells. However, these effects were completely blocked in the cultures pretreated with par-4 siRNA plus PrPc siRNA (Fig. 3e and Supporting Information Fig. S1B), indicating that PrPc protects cells from the proapoptotic effects of TMZ mediated via par-4.

PrPc blocks the transcriptional and pro-apoptotic activities of par-4 in the TMZ-treated cells

Par-4 protein, which is identified in cancer cells undergoing apoptosis, functions as a transcriptional repressor in the nucleus. We further investigated the mechanism by which PrPc suppresses the proapoptotic functions of par-4. On the one hand, to address whether PrPc could affect the par-4-mediated transcriptional activation, we performed luciferase reporter assays. Both U87 and U251 cells were cotransfected with the constant amount of the expression plasmids for HA-par-4 or ΔSAC-GFP and the luciferase reporter construct carrying par-4-responsive element derived from bcl-2 in the presence or absence of the increasing amounts of the Flag-PrPc expression plasmid. As seen in Figure 4a, Flag-PrPc significantly abolished the par-4-mediated reduced luciferase activities driven by bcl-2 promoter, whereas Flag-PrPc alone or/with ΔSAC-GFP had a marginal effect on the luciferase activities.

Figure 4.

TMZ triggers apoptosis via the PKA-par-4-bax-dependent pathway which can be neutralized by PrPc overexpression in glioma cells. (a). U87 and U251 cells were cotransfected with the constant amounts of HA-par-4 or ΔSAC-GFP mutant, the luciferase reporter construct carrying bcl-2 promoter together without or with the increasing amounts of Flag-PrPc. Forty-eight h after transfection, the cells were lysed and their luciferase activities were measured. After the PrPc-depleted U87 cells were treated with TMZ, the par-4 activity, the PrPc/par-4 interaction (b) and the localization of par-4 (c) were assessed by coimmunoprecipitation and immunofluorescence, respectively. (d). Upon TMZ treatment, the PKA activity in the PrPc-depleted U87 cells was evaluated. (e). After pretreated without or with the PKA inhibitor, H89, the par-4 activity in the PrPc-depleted U87 cells upon TMZ treatment was determined. (f). Expressions of cytochrome C, bcl-2, bax were assessed by immunoblotting in the PrPc- or/and par-4-depleted U87 cells after TMZ treatment. (g). The apoptosis rates and caspase activity in the PrPc-depleted cells treated with TMZ or/and the bax inhibitor peptide, V5, were assessed. **p < 0.05 versus corresponding controls cells (Student's t-test). Scale bars: 20 μm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

On the other hand, phosphorylation of human par-4 at T163 (within SAC domain) is required for activation of the apoptotic potential in cancer cells. Thus, changes in par-4 activity were examined using antibody that specifically recognizes phosphorylation of the T163 residue of par-4. Figure 4a indicated that active par-4 was negligibly detected in the untreated cells. However, after the PrPc-depleted or G1/S-arrested cells were treated with TMZ, increased amounts of activated par-4 were immunoprecipitated from the cell lysates. When we immunoprecipitated these cell lysates with the par-4 antibody, no detectable PrPc protein was detected in the par-4 immunoprecipitates (Fig. 4b, lane4 and Supporting Information Fig. S2, lane2). In contrast, both the TMZ-treated and the G2/M-arrested cells protected by PrPc overexpression demonstrated a block in this response in that no or very little active par-4 was detected in these immunoprecipitates. Accordingly, an immunoblot analysis of par-4 immunoprecipitates demonstrated that, under these conditions, par-4 remained associated with PrPc in these cells (Fig. 4b, lane2 and Supporting Information Fig. S2, lane4).

After stimulation, par-4 is phosphorylated and translocates from cytoplasm to nucleus, which is essential for initiation of apoptosis. Using an antibody against par-4, we were able to visualize the translocation of par-4 to the nucleus in the PrPc-depleted and G1/S-arrested cells after TMZ treatment. On the contrary, when the cells were treated with TMZ or synchronized at G2/M phase, par-4 was not detected in the nucleus but within the cytoplasm (Fig. 4c). These data provided the first evidence to support the hypothesis that upon TMZ treatment, PrPc prevents par-4 from executing its proapoptotic functions via binding with the SAC domain.

PKA-mediated par-4 phosphorylation drives the TMZ-induced apoptosis in glioma cells

We next explored the potential molecular basis of sensitivity to apoptosis by par-4 phosphorylation. PKA activity was significantly elevated after TMZ treatment in the PrPc-depleted cells (Fig. 4d). However, no similar effects were detected on other known upstream kinases of par-4, such as ZIPK and Akt (data not shown). Furthermore, after PrPc depletion, the TMZ-induced par-4 activation was largely abolished by cotreatment with the PKA inhibitor, H89, suggesting that par-4 served as a downstream target of PKA (Fig. 4e).

Studies have indicated roles for par-4 in activation of the Fas-FADD-caspase-8 pathway as well as inhibition of the NF-κB pro-survival pathway. To investigate the downstream signaling events that contribute to the TMZ-induced par-4-mediated cell death, we transfected U87 cells with PrPc siRNA and found that TMZ raised the expression of bax while reduced bcl-2, and that more cytochrome C released into cytoplasm in the PrPc-depleted cells, which were significantly attenuated by pretreatment with par-4 siRNA (Fig. 4f). Meanwhile, either Fas, FasL expressions or NF-κB activity remained unchanged under the same conditions (data not shown). A Bax inhibitor peptides V5, which inhibits the Bax-mediated apoptosis in vitro, also reversed the effects of PrPc RNAi as indicated by FCM assays as well as caspase activity assays (Fig. 4g). These results demonstrated the involvement of the PKA-par-4-Bax pathway in the TMZ-induced apoptosis upon PrPc depletion in glioma cells.

Combined PrPc shRNA and TMZ treatment inhibit the tumorigenicity of glioma cells in in vivo models

Since PrPc depletion seems to potentiate the TMZ-induced apoptosis in vitro, we evaluated the effects of PrPc targeted therapy on U87 glioma xenografts growth in vivo. A significant growth inhibitory effect was observed in the treated groups as compared to the control groups. Indeed, after 30 days of treatment with TMZ in the U87 and U87/PrPc-shRNA tumors, the mean tumor volume was reduced by 31.3% (p > 0.05) and 78.2% (p < 0.05; Fig. 5a), respectively, whereas the mean tumor weight was decreased by 35.9% (p > 0.05) and 79.6% (p < 0.05; Fig. 5b), respectively. Animal survival was represented by a Kaplan-Meier curve. Figure 5c shows that the median survival time of the untreated mice (PBS) was 43 days. In contrast, the median survival time of the TMZ+PrPc shRNA–treated mice was significantly longer (116 days) than that of the mice treated with TMZ or PrPc shRNA alone (79 and 44 days, respectively). The difference in survival between the groups was significant as demonstrated by log rank test (p < 0.05).

Figure 5.

The antiglioma effects of TMZ is significantly enhanced by PrPc shRNA. When the tumors of U87 xenograft reached an average volume of 80 mm3, mice (15/group) were treated with the PrPc-specific shRNAs or/and TMZ (50 mg/kg for five consecutive), as described in the Methods section. Tumor volume (a) and weight (b) were measured at 40 days after the inoculation. (c). Survival analysis is represented by a Kaplan-Meier curve (n = 10). **p < 0.05 versus the control cells.

Discussion

Treatment of malignant gliomas remains one of the greatest challenges facing the adult and pediatric oncologists today. Although initial treatment of glioma with surgery, radiotherapy and chemotherapy often produces some palliation of the symptoms, these tumors almost universally recur with an unrelenting progression to death. TMZ is an alkylating agent which is widely used for treating primary and recurrent high-grade gliomas.4, 12 In accordance with the previous reports, we found that at a clinically achievable dose, TMZ induces G2/M arrest which is associated with senescence, but not apoptosis in malignant glioma cells. Given enough time, the survival cells may reentry into the cell cycle and result in the creation of cells that, by virtue of their ability to bypass G2/M arrest, may divide and undergo further alterations that give rise to the cells with more aggressive phenotypes.6 Furthermore, we observed the TMZ-induced apoptosis in glioma cells synchronized at G1/S phase, whereas a dramatic decreased apoptosis rate was detected in the cells at G2/M phase. These results implied that the most likely factor that determines the susceptibility of glioma cells to TMZ, and thus the different fates of cells, lies in the cell cycle stages.

PrPc, abundantly expressed in neurons, was found to be the most differentially expressed gene between the glioma cells at G1/S and G2/M phase. PrPc is expressed in a cell cycle-dependent manner and particularly overexpressed when the cells transit through the G2/M phase upon TMZ treatment. Also, PrPc is significantly upregulated in the resected specimen from recurrent patients underwent TMZ treatment, suggesting that the altered PrPc expression should be ascribed to the TMZ-induced G2/M arrest. As PrPc plays a central role as an effective antiapoptotic protein in the neuronal cells, we hypothesized that PrPc expression upon TMZ treatment contributes to the antiapoptotic activity of glioma cells.

Most PrPc molecules are normally localized on the cell surface, where they are attached to the lipid bilayer via a C-terminal, glycosyl-phosphatidylinositol (GPI) anchor.13 Several intriguing lines of evidence have emerged recently suggesting that PrPc may exert a cytoprotective activity via interacting with downstream effectors.14–16 Functional proteomic analysis and reciprocal pull-down assays were performed to provide the first evidence of an association between PrPc and the anti-apoptotic protein, par-4, in the TMZ-treated glioma cells, as well as in the cells synchronized at G2/M phase. Par-4, a 340-amino-acid protein, is a cancer cell-selective proapoptotic protein that functions in the cytoplasmic and nuclear compartments as a tumor suppressor.17–19 Ectopic par-4 and its SAC domain are well-known to exert their apoptotic effects depending on the cellular context and post-translational modifications. For instance, par-4 phosphorylation by Akt prevents its nuclear translocation thereby promoting cell survival.20 In contrast, phoshorylation of par-4 at T163 by PKA appears to positively regulate its proapoptotic activity.21 Our study identifies a novel pathway to regulate the proapoptotic function of par-4. Upon TMZ treatment, PrPc overexpressed in the G2/M-arrested cells binds with par-4 via its SAC domain. This physical interaction prevents par-4 from the PKA-mediated activation and subsequent trafficking to nucleus, which is required for either the TMZ-induced mitochondrial pathway-dependent apoptosis or the direct suppression of bcl-2 promoter. So far, distinct partner proteins of intracellular par-4 have been confirmed. For example, par-4 interactions with ζPKC or TOP1 in the cytoplasm or nucleus, respectively, impede NF-kB activity,22, 23 and par-4 interactions with WT1 inhibit bcl-2 promoter.24 In this sense, considering the negative regulation of par-4 by a membranous protein, PrPc, these interactions reveal distinct, cell compartment specific roles for the par-4 partners in growth regulation.

We further reasoned that if the PrPc/par-4 complex is responsible for protecting the G2/M cells against apoptosis by TMZ, then disruption of the PrPc/par-4 interaction should effectively induce cell death. PrPc depletion greatly enhanced the TMZ-induced apoptotic rate of glioma cells. This effect was primarily reversed by disrupting par-4 expression, indicating that the PrPc/par-4 complex is responsible for eluding the cytotoxic attacks via its antiapoptotic activity. Two mechanisms of resistance to TMZ have been elucidated. The first mechanism of resistance involves the DNA repair protein, O6-alklyguanine-DNA alkyltransferase. AGT acts by directly removing the methyl adduct on the O6 guanine position.25 Independent of AGT levels, resistance to temozolomide has also been shown in tumor cells deficient in the DNA mismatch repair system. This deficiency leads to the tolerance of O6 methylguanine and the ability of the cell to survive despite the presence of persistent DNA damage.26, 27 Our studies present a novel model in which par-4 may act as one of the ligand of PrPc receptor and interaction between the ligand/receptor accounts for antiapoptotic activity of glioma upon TMZ treatment. As PrPc depletion has greatly improved the efficacy of TMZ-based therapy both in vitro and in vivo, a better understanding of the molecular mechanisms of action of PrPc in our study may contribute to the development of further strategies for rationally and selectively manipulating the sensitivity of glioma cells to the TMZ-based chemotherapy.

Ancillary