Regulation of GST-MDA-7 toxicity in human glioblastoma cells by ERBB1, ERK1/2, PI3K, and JNK1-3 pathway signaling

Glioblastoma multiforme (GBM) is diagnosed in ∼20,000 patients per annum in the United States, with anaplastic astrocytoma and GBM accounting for the majority of astrocytic tumors (1, 2). Despite surgical removal of the tumor and intensive use of radiation and chemotherapy, the mean survival of this disease is only in general extended by these interventions from 2 to 3 months to 1 year (1–5).

The gene for mda-7 [interleukin (IL)-24], an IL-10 family member (6–12), was isolated from human melanoma cells induced to undergo terminal differentiation by treatment with IFN-β and mezerein (6). In agreement with MDA-7/IL-24 acting as a tumor suppressor, MDA-7/IL-24 protein expression is decreased in advanced melanomas with nearly undetectable levels in metastatic disease (6, 13, 14). Enforced expression of MDA-7/IL-24 inhibits the growth of transformed cells and ultimately causes cell death, without apparently exerting deleterious effects in normal human epithelial cells, fibroblast cells, melanocytes, and astrocytes (6, 14–26). The mechanisms by which Ad.mda-7 induces tumor cell death are not fully understood; however, several studies suggest the involvement of proteins important for the onset of growth inhibition and apoptosis, including BCL-2 family members (16–26). For example, in melanoma, malignant glioma, and prostate and breast cancer cell lines but not in normal melanocytes, astrocytes, or prostate or breast epithelial cells, respectively, infected with Ad.mda-7, the ratio of proapoptotic to antiapoptotic proteins in cancer cells changes in favor of cell death, thereby facilitating induction of apoptosis (12, 16–20, 25, 26). In prostate cancer lines, overexpression of either BCL-2 or BCL-x_L in a cell type–dependent fashion can protect cells from Ad.mda-7 toxicity (24, 26). Thus, MDA-7/IL-24 lethality can occur by different pathways in different cell types. More recently, MDA-7 toxicity has been linked to alterations in homeostatic endoplasmic reticulum stress signaling as judged by its binding to BiP/GRP78 (27).

The ability of MDA-7/IL-24 to modulate cell signaling processes in transformed cells has been investigated by several groups, with the majority of studies presenting data from adenoviral transduction and expression of the MDA-7/IL-24 transgene in the infected cell population. Our laboratories have presented evidence that Ad.mda-7 kills established glioma and melanoma cells in part by promoting p38 mitogen-activated protein kinase (MAPK)–dependent expression of GADD153 (25, 28). Other groups have argued that inhibition of phosphatidylinositol-3-kinase (PI3K) signaling, but not extracellular signal-regulated kinase 1/2 (ERK1/2) signaling, modestly promotes Ad.mda-7 lethality in breast and lung cancer cells (29, 30). Prior studies using purified MDA-7/IL-24 proteins have tended to treat tumor cells with high nanomolar to micromolar concentrations of the cytokine in studies to modulate tumor cell biology, which are unlikely to be therapeutically achievable in vivo. Prior work by our groups has shown, using bacterially synthesized GST-MDA-7 protein, that in the 0.25 to 2.0 nmol/L concentration range GST-MDA-7 primarily causes growth arrest with little cell killing, whereas at ∼20-fold greater concentrations the cytokine causes profound growth arrest and a delayed induction of tumor cell death, 48 to 96 h after exposure (31–33). The toxicity of low nanomolar GST-MDA-7 concentrations were elevated by multiple agents that generate reactive oxygen species, which correlated with prolonged activation of the c-Jun NH₂-terminal kinase (JNK) 1/2 pathway that was linked to tumor cell killing. However, the dose-dependent effects of GST-MDA-7 on tumor cell signaling processes have not been studied in any cell line in a detailed manner, and the relative abilities of low (∼1 nmol/L) and higher (∼30 nmol/L) GST-MDA-7 concentrations to alter signaling pathway activities in tumor cells, and the relevance of any changes in cell signaling processes to cell survival, are presently unknown.

Using primary human GBM isolates, maintained in vivo and cultured in vitro, and established human glioma cell lines, we examined the effect of GST-MDA-7 on cell viability and cell signaling and linked changes in cell signaling to some of the molecular mechanisms by which GST-MDA-7 enhances glioma cell death.

Antibody reagents, kinase inhibitors, caspase inhibitors cell culture reagents, primary human GBM cells, and noncommercial recombinant adenoviruses, which were kindly provided by Dr. K. Valerie (Virginia Commonwealth University) and Dr. J. Moltken (University of Cincinnati), have been previously described in refs. 18–20, 24–28, 31–35.

Generation of Ad.mda-7 and Synthesis of GST-MDA-7. Recombinant type V adenovirus to express MDA-7 (Ad.mda-7), control (CMV vector), or control (β-galactosidase) were generated using recombination in HEK293 cells as described in refs. 18–20, 24–28, 32.

Culture and In vitro Exposure of Cells to GST-MDA-7 and Drugs. All established cell lines (U251 and U118) were cultured at 37°C [5% (v/v) CO₂] in vitro using RPMI supplemented with 5% (v/v) FCS and 10% (v/v) nonessential amino acids. Primary human glioma cells were initially subcultured following isolation from the animal in 2% (v/v) FCS to prevent growth of contaminating rodent fibroblasts for 1 week before in vitro analyses, after which cells were cultured in 5% (v/v) FCS. For short-term cell killing assays and immunoblotting, cells were plated at a density of 3 × 10³ per cm² and 36 h after plating treated with GST-MDA-7 and/or various drugs as indicated. Cells were not cultured in reduced/no serum medium during any study in this research.

Cell Treatments, SDS-PAGE, and Western Blot Analysis. Cells were treated with various GST-MDA-7 concentrations as indicated in the figure legend. For SDS-PAGE and immunoblotting, cells were lysed in a nondenaturing lysis buffer and prepared for immunoprecipitation as described in refs. 24–33.

Recombinant Adenoviral Vectors: Infection In vitro. We generated and purchased previously noted recombinant adenoviruses to express constitutively activated and dominant-negative AKT (dnAKT) and MEK1 (dnMEK1) proteins, dominant-negative caspase-9 (dncasp.9), XIAP, c-FLIP-s, CRM A, and BCL-x_L (Vector Biolabs). Cells were infected with these adenoviruses at an approximate multiplicities of infection (MOI) of 50. As noted above, cells were further incubated for 24 h to ensure adequate expression of transduced gene products before drug exposures.

Detection of Cell Death by Trypan Blue, Hoechst, Terminal Deoxynucleotidyl Transferase–Mediated dUTP Nick End Labeling, and Flow Cytometric Assays. Cells were harvested by trypsinization with trypsin/EDTA for ∼10 min at 37°C. As some apoptotic cells detached from the culture substratum into the medium, these cells were also collected by centrifugation of the medium at 1,500 rpm for 5 min. The pooled cell pellets were resuspended and mixed with trypan blue dye. Trypan blue stain, in which blue dye incorporating cells were scored as being dead, was done by counting of cells using a light microscope and a hemacytometer. Five hundred cells from randomly chosen fields were counted and the number of dead cells was counted and expressed as a percentage of the total number of cells counted. For confirmatory purposes, the extent of apoptosis was evaluated by assessing Hoechst-stained and terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling–stained cytospin slides under fluorescent light microscopy and scoring the number of cells exhibiting the “classic” morphologic features of apoptosis and necrosis. For each condition, 10 randomly selected fields per slide were evaluated, encompassing at least 1,500 cells. Alternatively, the Annexin V/propidium iodide assay was carried to determine cell viability out as per the manufacturer’s instructions (BD PharMingen) using a Becton-Dickinson FACScan flow cytometer (18, 32, 34).

Cell Survival Analyses. Cells were assayed for the effect(s) of GST-MDA-7 cell survival. Cells were plated (250-2,000 per 60-mm dish) and 12 h after plating treated with GST or GST-MDA-7. Ninety-six hours after GST-MDA-7 treatment, cells were washed with fresh medium and permitted to continue growing for 20 to 28 days in the absence of GST or GST-MDA-7 (32).

Preparation of S-100 Fractions and Assessment of Cytochrome c Release. Cells were harvested after GST-MDA-7 treatment as described previously by centrifugation at 600 rpm for 10 min at 4°C and washed in PBS. Cells (∼1 × 10⁶) were lysed by incubation for 3 min in 100 μL lysis buffer containing 75 mmol/L NaCl, 8 mmol/L Na₂HPO₄, 1 mmol/L NaH₂PO₄, 1 mmol/L EDTA, and 350 μg/mL digitonin. The lysates were centrifuged at 12,000 rpm for 5 min, and the supernatant was collected and added to an equal volume of 2× Laemmli buffer. The protein samples were quantified and separated by 15% SDS-PAGE.

Data Analysis. Comparison of the effects of various treatments was done using one-way ANOVA and a two-tailed Student’s t test. Differences with P < 0.05 were considered statistically significant. Experiments shown are the individual experimental mean ± SE (of multiple individual data points) from multiple separate experiments.

The effect of low (∼1 nmol/L), intermediate (∼5 nmol/L), and high (∼30 nmol/L) GST-MDA-7 concentrations on cell signaling and viability in primary human glioma cells expressing activated ERBB1vIII (GBM6) and mutant active wild-type ERBB1 (GBM12) was determined (Fig. 1A-C). Low concentrations of GST-MDA-7 had a modest effect on cell viability 0 to 96 h after exposure as judged by caspase-3 cleavage, poly(ADP-ribose) polymerase cleavage, trypan blue staining, terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling staining, and Hoechst staining; however, low concentrations of GST-MDA-7 significantly suppressed cell growth as judged by cell counting and by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays (Fig. 1B and C; data not shown). Higher concentrations of GST-MDA-7 promoted significant higher amounts of cell killing that were manifested 48 to 96 h after treatment, which correlated with cleavage of caspase-3 and poly(ADP-ribose) polymerase, and a significantly greater suppression of cell numbers than were observed with lower cytokine doses (Fig. 1A-C). Near identical data to that in GBM6 and GBM12 cells were obtained in GBM5 cells that express a constitutively active PI3K protein (Supplementary Fig. S1).⁷

Based on these observations, we next explored the effect 1 to 30 nmol/L GST-MDA-7 had on the activities of multiple signal transduction pathways in primary human GBM cells over a 72-h time course as assessed by the phosphorylation of individual protein kinases. Treatment of GBM6 and GBM12 cells with 1 nmol/L GST-MDA-7 in a cell type–dependent fashion modestly suppressed or enhanced ERK1/2 and AKT activity and modestly enhanced p38 MAPK and JNK1-3 signaling (Fig. 2A and B; cf. data in ref. 32). Treatment of these cells with 30 nmol/L GST-MDA-7 abolished ERK1/2 (GBM6) or AKT (GBM12) activity and more strongly promoted phosphorylation of JNK1-3 and p38 MAPK. JNK1-3 activity, as measured using immunoprecipitation of the enzyme and a 96-well colorimetric plate assay, was also noted to be increased in GBM6 cells in a dose-dependent fashion (data not shown). GBM5 cells responded to GST-MDA-7 exposure in a similar manner to GBM6 cells (Supplementary Fig. S1, inset).⁷ An intermediate GST-MDA-7 concentration of 5 nmol/L, which caused a similar level of toxicity to that observed in cells treated with 1 nmol/L GST-MDA-7, enhanced JNK1-3 and p38 MAPK activity to the same extent as that observed in cells treated with 30 nmol/L GST-MDA-7, but in contrast to a 30 nmol/L GST-MDA-7 exposure, treatment with 5 nmol/L GST-MDA-7 did not profoundly suppress ERK1/2 or AKT signaling (Fig. 2; cf. data in Fig. 1).

Additional studies then attempted to further delineate the pathway by which JNK1-3 became activated. Expression of dominant-negative MEKK1 abolished GST-MDA-7-induced activation of JNK1-3 in GBM6 and GBM12 cells (Fig. 2C). Expression of dominant-negative RhoA N19 also abolished JNK1-3 activation by GST-MDA-7, whereas expression of dominant-negative Rac1 N17/Cdc42 N17/Ras N17 had no effect on JNK1-3 activation (Fig. 2C). Expression of dominant-negative MKK4 or MKK7 weakly suppressed JNK1-3 activation, whereas combined expression of dominant-negative MKK4 and MKK7 abolished GST-MDA-7-induced activation of JNK1-3 in GBM6 cells (Fig. 2C).

Further analyses then determined using molecular approaches whether any of the alterations in GBM cell viability were causally linked to altered signaling pathway function(s) in cells treated with 30 nmol/L GST-MDA-7. In GBM6 and GBM12 cells, combined inhibition of MEK1 and AKT function promoted the lethality of GST-MDA-7 (Fig. 3A and B). In GBM6 and GBM12 cells, activation of AKT or MEK1 variably and modestly suppressed GST-MDA-7 lethality in the primary GBM cells, whereas activation of both AKT and MEK1 was required to promote a potent antiapoptotic effect regardless of the cell type under investigation (Fig. 3). In all GBM cells, inhibition of JNK1-3 signaling suppressed GST-MDA-7 toxicity; surprisingly, based on prior work in established tumor cell lines, wherein p38 MAPK signaling was toxic, expression of dominant-negative p38α MAPK (dnp38α) enhanced basal levels of GBM cell morbidity as well as GST-MDA-7 lethality: the enhanced level of killing observed in cells expressing dnp38α was abolished by inhibition of JNK1-3 function. In agreement with the hypothesis that the majority of cell killing was being mediated via the intrinsic apoptosis pathway, expression of dncasp.9 and overexpression of BCL-x_L suppressed GST-MDA-7 lethality (Fig. 3A-C). Identical data to that observed in GBM6 and GBM12 cells were obtained in GBM5 cells (Supplementary Fig. S2).⁷

Based on the above findings, we did studies to determine how or whether signaling pathway cross-talk occurred in the control of cell viability in cells treated with GST-MDA-7. In GBM6 and GBM12 cells, treatment with 30 nmol/L GST-MDA-7 suppressed ERK1/2 and AKT activity, respectively (Fig. 4A; cf. data in Fig. 2). Although inhibition of JNK1-3 signaling protected GBM6 and GBM12 cells against the toxic effects of GST-MDA-7, it did not restore ERK1/2 and AKT activity to basal levels. These findings are in contrast to our prior work in which we combined 1 nmol/L GST-MDA-7 and ionizing radiation exposure, where inactivation of ERK1/2 was JNK1-3 dependent and caspase-3 dependent (30). Expression of dominant-negative p38 MAPK enhanced basal levels of cell death as well as 30 nmol/L GST-MDA-7 toxicity, which in the case of basal viability levels correlated with modest inactivation of AKT, and in the case of GST-MDA-7 toxicity correlated with a weak (GBM12) and stronger (GBM6) promotion of JNK1-3 activation, which was causal in cell killing (Fig. 4A; cf. data in Fig. 3). Activation of JNK1-3 has been linked by many groups to the induction of mitochondrial dysfunction, and we examined the regulation of BAX activity in GST-MDA-7-treated glioma cells. GST-MDA-7 (30 nmol/L) activated BAX in a JNK1-3-dependent fashion and promoted cell killing: constitutive activation of MEK1 and AKT suppressed MDA-7/IL-24-induced activation of JNK1-3 and of BAX and suppressed cell killing (Fig. 4B). To confirm a role for JNK1-3 signaling in long-term cell survival, we did colony formation assays; a 96-h treatment of GBM12 or U251 glioma cells with GST-MDA-7 (30 nmol/L) suppressed colony formation 14 to 21 days after exposure, an effect that was blocked by inhibition of JNK1-3 signaling and by inhibition of caspase-9 function (Fig. 4C).

In prior studies, Ad.mda-7 and GST-MDA-7 have been shown to promote cell killing via activation of the p38 MAPK pathway and increased expression of GADD transcription factors (25, 28, 33). Hence, we determined whether the role of p38 MAPK signaling in GST-MDA-7 toxicity in GBM6 and GBM12 cells was dose dependent and different from established tumor cell types examined previously. Treatment of GBM6 cells with GST-MDA-7 (1-300 nmol/L) enhanced in a dose-dependent fashion phosphorylation of p38 MAPK and its substrate heat shock protein (HSP) 27 (Fig. 4D, inset; cf. data in Fig. 2A). In contrast to GBM6 cells, treatment of GBM12 cells with GST-MDA-7 (1-300 nmol/L) resulted in a peak p38 MAPK activation, which occurred at ∼30 nmol/L GST-MDA-7 and that was not significantly further elevated by exposure to 300 nmol/L GST-MDA-7 (Fig. 4D, inset; cf. data in Fig. 2B). In agreement with our data expressing dominant-negative p38 MAPK, treatment of GBM6 and GBM12 cells with a low specific concentration of the p38α/β MAPK inhibitor SB203580 (1 μmol/L), but not the inactive analogue of this drug SB202474, enhanced basal levels of cell morbidity and modestly enhanced GST-MDA-7 (30 nmol/L) lethality within 48 h (Fig. 4D, bottom). Treatment of GBM6 cells with GST-MDA-7 (300 nmol/L) caused ∼4-fold more cell killing than treatment with GST-MDA-7 (30 nmol/L), and inhibition of p38 MAPK suppressed the lethality of GST-MDA-7 (300 nmol/L). In GBM12 cells, however, inhibition of p38 MAPK signaling enhanced GST-MDA-7 lethality regardless of GST-MDA-7 concentration (30-300 nmol/L).

In established glioma cells, GADD transcription factors have been linked to the toxic effects of MDA-7/IL-24. We noted that GADD153 expression was enhanced in a p38 MAPK-dependent fashion in GBM6 cells treated with GST-MDA-7 (30 and 300 nmol/L), but in contrast to the dose-dependent increases in p38 MAPK phosphorylation that correlated with changes in cell viability, there was no marked dose-dependent increase in GADD153 levels with increasing GST-MDA-7 doses (Fig. 4D, inset, left). Similar findings for GADD153 expression were noted in GBM12 cells. GST-MDA-7 very weakly modulated GADD34 and GADD45 protein levels in GBM6 and GBM12 cells. Collectively, these findings argue that the role of p38 MAPK signaling and GADD transcription factors in modulating GST-MDA-7 toxicity in primary human glioma cells is a more complicated process than has been observed in other established tumor cell types. Collectively, the findings in Figs. 2 to 4 show that GST-MDA-7-induced activation of the JNK pathway coupled to inactivation of the MEK1/2 pathway and/or the AKT pathway plays a central role in the lethality of GST-MDA-7 in primary human glioma cells. Our data also argue that, in a dose-dependent manner, activation of p38 MAPK by GST-MDA-7 has the potential to either promote cell survival or cell killing.

The probability that physicians will be able to achieve systemic prolonged high toxic concentrations of MDA-7/IL-24 (>30 nmol/L) in patients is low. Hence, further analyses determined whether the toxicity of lower GST-MDA-7 concentrations (∼1 nmol/L) could be enhanced by the manipulation of AKT and MEK1 signaling. In GBM6 and GBM12 cells, the promotion of 1 nmol/L GST-MDA-7 toxicity to a level approximating the lethality of a 30 nmol/L GST-MDA-7 treatment required inhibition of both the PI3K-AKT pathways and the MEK-ERK1/2 pathways (Fig. 5A and B). Similar data were obtained in GBM5 cells (Supplementary Fig. S3).⁷ In GBM6 and GBM12 cells, expression of dnMEK1 enhanced GST-MDA-7-induced activation of JNK1-3 that was further enhanced by inhibition of AKT (Fig. 5A and B, inset). In GBM6 and GBM12 cells, suppression of JNK1-3 signaling abolished the promotion of GST-MDA-7 toxicity caused by inhibition of MEK1 and AKT (Fig. 5C, inset, left and bottom; data not shown). Inhibition of MEK1 and AKT enhanced the activation of BAX by low 1 nmol/L concentrations of GST-MDA-7; the activation of BAX was also JNK1-3 dependent (Fig. 5C, inset, right). Collectively, our data show that simultaneous inhibition of both PI3K-AKT and MEK-ERK1/2 pathways enhances the toxicity of low 1 nmol/L GST-MDA-7 concentrations via a JNK1-3-dependent and BAX-dependent mechanism.

ERBB1 is a growth factor receptor whose actions have been linked via the ERK1/2 and PI3K-AKT pathways in many gliomas to an aggressive phenotype and to rapid patient morbidity (3, 5, 32). As ERBB1 signaling can modulate ERK1/2 and AKT activity, and inhibition of these pathways enhances GST-MDA-7 toxicity (Figs. 3–5), we determined whether inhibition of ERBB1 function enhanced GST-MDA-7 toxicity and could replicate our observations for inhibition of ERK1/2 and AKT. In GBM12 cells that express a full-length mutated active ERBB1 protein, expression of dominant-negative ERBB1 (COOH-terminal deletion of 533 amino acids, CD533) or treatment with an ERBB1 inhibitor AG1478 caused a similar amount of cell death to that induced by GST-MDA-7 (30 nmol/L) exposure. Combined treatment of GBM12 cells with GST-MDA-7 and inhibition of ERBB1 resulted in essentially a null effect on cell killing (Fig. 6A). To further explore whether ERBB1 signaling alters the survival of glioma cells exposed to GST-MDA-7, we treated established U118 human glioma cells transfected to express wild-type ERBB1 or ERBB1vIII with GST-MDA-7. Expression of either wild-type ERBB1 or ERBB1vIII suppressed the toxicity of GST-MDA-7 in U118 cells compared with cells transfected with vector control plasmid (Fig. 6B). Collectively, these findings argue in human glioma cells that downstream inhibition of PI3K and MEK1/2 may represent a better approach to enhance GST-MDA-7 lethality that upstream inhibition of ERBB1.

Geldanamycins are therapeutic agents that act by inhibiting the function of HSP90 and lower expression, and reduce the activity, of the PI3K-AKT and the MEK-ERK1/2 pathways (35, 36). This suggests that geldanamycins could be a useful family of clinically relevant compounds that may interact with MDA-7. In GBM6 and GBM12 cells, 17AAG modestly suppressed the activity of ERK1/2 and AKT, and combined with GST-MDA-7 to further suppress ERK1/2 and AKT activity, as well as activating JNK1-3 (Fig. 6C). As 17AAG suppressed both AKT and ERK1/2 activity and activated JNK1-3, we determined whether constitutive activation AKT and MEK1 prevented the toxic interaction between GST-MDA-7 and 17AAG. In GBM6 cells, the toxicity of 17AAG was suppressed by combined, but not individual, activation of AKT and MEK1 (Fig. 6D). Individual activation of either AKT or MEK1 suppressed the toxic interaction between 17AAG and GST-MDA-7. Tumor cell viability in cells treated with 17AAG and GST-MDA-7 was maintained by inhibition of JNK1-3 signaling.

Previously, we have noted that treatment of cells with GST-MDA-7 reduced proliferation and caused tumor-specific cell killing and radiosensitization (25, 31, 32). The studies in this research were designed to determine how primary human glioma cells respond to GST-MDA-7 exposure over a range of cytokine concentrations with respect to alterations in cell signaling and how, mechanistically, alterations in signaling affect viability.

Low concentrations of GST-MDA-7 (∼1 nmol/L) suppressed glioma cell growth with minimal toxicity, which correlated with transient weak enhancement of p38 MAPK and JNK1-3 activity and with a modest reduction in ERK1/2 and AKT signaling. An intermediate 5-fold higher concentration of GST-MDA-7 also suppressed glioma cell growth with modest toxicity, which correlated with a stronger transient activation of p38 MAPK and JNK1-3 but with only a modest reduction in ERK1/2 and AKT signaling. A 30-fold higher GST-MDA-7 concentration caused profound levels of toxicity that correlated with stronger activation of JNK1-3 and with a stronger transient activation of p38 MAPK and that in a cell type–dependent fashion nearly abolished either ERK1/2 and/or AKT signaling.

Multiple studies using a variety of cytokine and toxic stimuli have shown that JNK1-3 activation in astrocytes, neurons, and transformed versions of these cells can cause cell death (37, 38). The balance between the outputs of ERK1/2 and AKT signaling and that of JNK1-3 signaling has been argued to represent a key homeostatic mechanism, which regulates cell survival versus cell death processes (39, 40). GST-MDA-7-induced JNK1-3 signaling was causal in BAX activation that will act to promote caspase-9-dependent cell killing. Maintenance of ERK1/2 and/or AKT signaling in cells exposed to high toxic concentrations of GST-MDA-7 suppressed JNK1-3 phosphorylation as well as abolishing subsequent BAX activation and cell killing. Inhibition of AKT and ERK1/2 promoted the toxicity of GST-MDA-7, which was associated with increased JNK1-3 signaling, and that was causal in BAX activation and tumor cell death. Thus, GST-MDA-7, at low apparently nontoxic concentrations, activates the JNK1-3 pathway in glioma cells but does not suppress ERK1/2 and/or AKT signaling, and for GST-MDA-7 to become a toxic agent, even in glioma cells exhibiting relatively high levels of JNK1-3 activity requires the suppression of ERK1/2 and/or AKT function. Thus, the loss of ERK1/2 and AKT signaling appears to be a primary response associated with cell killing, and the gain of JNK1-3 signaling a secondary response, in GST-MDA-7-induced lethality in primary human glioma cells.

In prior studies, we and others have argued that Ad.mda-7 and GST-MDA-7-induced activation of the p38 MAPK pathway is causal in the lethal effects of MDA-7 expression (27, 28). The present studies in primary human glioma cells, using molecular and small-molecule inhibitor approaches, argue that basal levels of p38 MAPK activity and GST-MDA-7-induced p38 MAPK activation in the 1 to 30 nmol/L concentration range are protective signals. We noted that inhibition of p38 MAPK function could suppress basal levels of AKT phosphorylation in GBM6 cells, in general agreement with data reported by Park et al. (41). Sainz-Perez et al. argued that loss of MDA-7 expression in leukemia cells reduced p38 MAPK activity, which was causal in reduced viability (42). Inhibition of p38 MAPK function also modestly enhanced GST-MDA-7-induced JNK1-3 phosphorylation, and enhanced JNK1-3 signaling was causal in BAX activation and elevated levels of cell death.

The reasons why virus-mediated intracellular overexpression of MDA-7 or high dose ∼400 to 600 nmol/L GST-MDA-7 treatment may cause cell killing via p38 MAPK signaling whereas we observed lower dose 1 to 30 nmol/L MDA-7 acting in a protective manner in glioma cells were unclear. In other works, we and others have shown that the biological effect of transient intense or prolonged low-level activation of the ERK1/2 or JNK1/2 pathways can promote cell growth and cell survival, whereas prolonged intense activation of these pathways inhibits growth and causes cell death (43). Thus, it is possible that based on the dosage of MDA-7 the relative amplitude and duration of p38 MAPK activation could be a marker for whether the “readout” signals from this pathway are protective or toxic. In agreement with this hypothesis, we noted in GBM6 cells that an order of magnitude higher GST-MDA-7 concentration (300 nmol/L) promoted intense activation of p38 MAPK compared with a 30 nmol/L GST-MDA-7 exposure; 300 nmol/L GST-MDA-7-induced p38 MAPK activation was a cytotoxic signal. However, in GBM12 cells, regardless of the GST-MDA-7 concentration, p38 MAPK signaling appeared to act in a protective manner. Hence, our findings suggest that p38 MAPK signaling may promote or diminish GST-MDA-7 toxicity, which is dependent on the cell type and the relative cytokine concentration.

ERBB1 expression in glioblastoma, either as a full-length receptor or as a truncated active mutant receptor (variant III), is known to be an indicator for a more malignant cellular phenotype and for poorer patient survival (3, 5). Recently, inhibition of ERBB1 function was shown to enhance the lethality of Ad.mda-7 in lung cancer cells, in part by derepressing inhibitory signaling from the ERBB1 receptor, which acts to suppress MDA-7/IL-24 protein synthesis (44). To our surprise, we found that inhibition of ERBB1 using small-molecule or molecular approaches did not significantly alter GST-MDA-7 toxicity in GBM12 cells that express a full-length mutant active ERBB1 protein. In contrast to these findings, expression of ERBB1 or ERBB1vIII suppressed GST-MDA-7 lethality in stably transfected U118 established human glioma cells. Collectively, these findings in glioma cells argue against a simplistic interpretation that ERBB1 inhibition obligatorily promoting GST-MDA-7 toxicity and suggest that combined inhibition of PI3K-AKT signaling and MEK1/2-ERK1/2 signaling may represent a more ubiquitous approach to promoting cytokine toxicity.

Treatment with GST-MDA-7 in a dose-dependent fashion caused a JNK-dependent activation of BAX. BAX is a BH3 domain protein that promotes pore formation in the mitochondrial outer membrane leading to loss of mitochondrial integrity, including the release into the cytosol of cytochrome c (16, 17, 45, 46). In agreement with data showing BAX activation, expression of dncasp.9 or of BCL-x_L suppressed GST-MDA-7 lethality, showing that activation of the intrinsic pathway was a central response to GST-MDA-7 exposure. Prior studies, in which low concentrations of GST-MDA-7 were shown to radiosensitize primary human GBM cells by activation of JNK1-3 and the intrinsic pathway, together with the present findings using higher individual concentrations of GST-MDA-7, suggest that in human glioma cells agents that promote activation of other apoptosis pathways could have potential for combinatorial therapeutic approaches.

Based on our finding that the toxicity of GST-MDA-7 correlated with reduced ERK1/2 and AKT activity, we hypothesized that geldanamycin-induced inhibition of AKT and ERK1/2, by suppressing HSP90 function, would enhance cytokine toxicity. The toxic interaction between the clinically relevant geldanamycin 17AAG and GST-MDA-7 was suppressed by constitutive activation of either AKT or MEK1, whereas the toxicity of 17AAG as a single agent was only suppressed by expression of both activated AKT and MEK1. A recent contemporaneous study by Pataer et al. noted that viability of established lung cancer cells was reduced in a greater than additive manner in cells infected with Ad.mda-7 and treated with HSP90 antagonist geldanamycins without suppression of HSP90 expression (47). In this study, the relative role of AKT versus MEK1 signaling was not investigated. Several groups have shown that geldanamycins, alone or in combination with other agents, can alter the viability of brain cancer cells in vitro. Additional studies will be required to determine whether the lethality of MDA-7 towards glioma tumors in vivo can be enhanced by agents that simultaneously suppress both AKT and ERK1/2 activity.