OSU-03012 stimulates PERK-dependent increases in HSP70 expression, attenuating its lethal actions in transformed cells

We have further defined mechanism(s) by which OSU-03012 (OSU), a derivative of the COX2 inhibitor Celecoxib but lacking COX2 inhibitory activity, kills transformed cells. In cells lacking expression of PKR-like endoplasmic reticulum kinase (PERK -/-) the lethality of OSU was attenuated. OSU enhanced the expression of Beclin 1 and ATG5 and cleavage of pro-caspase 4 in a PERK-dependent fashion and promoted the Beclin 1and ATG5-dependent formation of vacuoles containing LC3, followed by a subsequent caspase 4-dependent cleavage of cathepsin B and a cathepsin B-dependent formation of low pH intracellular vesicles; cathepsin B was activated and released into the cytosol and genetic suppression of caspase 4, cathepsin B or AIF function significantly suppressed cell killing. In parallel, OSU caused PERK-dependent increases in HSP70 expression and decreases in HSP90 and Grp78/BiP expression. Changes in HSP70 expression were post-transcriptional. Knock down or small molecule inhibition of HSP70 expression enhanced OSU toxicity and over-expression of HSP70 suppressed OSU-induced low pH vesicle formation and lethality. Our data demonstrate that OSU-03012 causes cell killing that is dependent on PERK-induced activation of multiple toxic proteases. OSU-03012 also increased expression of HSP70 in a PERK-dependent fashion, arguing that OSU-03012-induced PERK signaling promotes both cell survival and cell death processes. This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 4 Introduction. Inhibitors of cyclooxygenase 2 (COX2) were originally developed to inhibit inflammatory immune responses, with a primary intention to use such agents clinically in the treatment of chronic diseases e.g. rheumatoid arthritis (Hawkey et al., 2005; , Kiefer et al., 2004). It was also noted that COX2 was over-expressed in many tumor cells and that agents which inhibited COX2 e.g. Celecoxib (Celebrex) could suppress tumor cell growth in vitro and when grown as xenografts in animals (Klenke et al., 2006; Koehne et al., 2004; Cui et al, 2005; Kang et al., 2006). Studies in patients demonstrated that individuals with prolonged exposure to COX2 inhibitors as part of an anti-inflammatory therapeutic regimen also had a lower incidence of developing cancer, suggestive that COX2 inhibitors were cancer preventative (Kashfi and Rigas, 2005; Narayanan et al., 2006). However, as the sensitivity of tumor cells to COX2 inhibitors was investigated in greater detail, it became apparent that expression of COX2 did not per se correlate with tumor cell sensitivity to COX2 inhibitor treatment (Patel et al, 2005; Kulp et al., 2004). The agent OSU-03012 was developed as an anti-cancer agent, with Celecoxib as the chemical backbone (Zhu et al., 2004). In vitro OSU-03012 has an order of magnitude greater anti-tumor activity than Celecoxib, but lacks COX2 inhibitory activity. Based on these preliminary observations, OSU-03012 was approved for development by the National Cancer Institute RAID program, with likely initiation of a Phase I drug trial in 2007. Studies by our laboratory recently argued that OSU-03012 caused cell death through mechanisms which involved a form of endoplasmic reticulum (ER) stress signaling and mitochondrial dysfunction but that were a caspase-independent form of cell death, as initially judged by a lack of effect of the caspase inhibitor zVAD and expression of dominant negative caspase 9. Instead, our findings argued in HCT116 cells that knock down of apoptosis inducing factor (AIF) expression significantly attenuated OSU-03012 lethality (Yacoub et al., 2006). In the last 5-10 years, multiple growth factor receptors and downstream signal transduction pathways have been linked to the advantage tumor cells have over non-transformed cells in terms of increased rates of proliferation and cell survival following exposure to toxic stresses (reviewed in Dent et al., 2003; Valerie et al., 2007). The This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 5 toxicity of OSU-03012 in tumor cells was initially argued to be due to inhibition of the enzyme PDK-1, part of the PI3 kinase pathway, in as much as OSU-03012 can suppress AKT phosphorylation and showed measurable inhibition of PDK-1 activity in the 5-50 μM range in vitro (Dent et al., 2003). OSU-03012 has also been shown to interact in a synergistic fashion with BCR-ABL inhibitors and with the ERBB2 inhibitor Herceptin to suppress tumor cell viability and to kill in a manner that is in many cell types at least partially caspase independent (Johnson et al., 2005; Tseng et al., 2005; Tseng et al., 2006; To et al., 2007). In our previous studies, we also noted that inhibition of either MEK1/2 or PI3K enhanced the toxicity of OSU-03012 in glioma, colon cancer and transformed rodent fibroblast cell types (Yacoub et al, 2006). However, while OSU-03012 has been noted to suppress PDK-1 function and AKT activity, other data has also strongly argued that OSU-03012 toxicity, and its radiosensitizing effects, could not attributed to suppression of AKT signaling (Yacoub et al, 2006, Caron et al., 2004). In the present studies, we have again utilized established and primary human glioma and established colon cancer cell lines, and transformed fibroblasts lacking expression of pro-apoptotic proteins, and examined the impact of OSU-03012 on cell viability, and further defined the molecular mechanisms by which OSU-03012 enhances tumor cell death (Yacoub et al, 2006). This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 6 Materials and Methods. Materials. Phospho-/total(ERK1/2; JNK1/2; p38 MAPK) antibodies, phospho-/total-AKT (T308; S473) and the total and cleaved caspase 3 antibodies were purchased from Cell Signaling Technologies (Worcester, MA). Anti-PERK, anti-BID, anti-caspase 2, anti-caspase 4, anti-cathepsin B, pan-anti-HSP70, anti-HSP90, anti-eIF2α and anti-eIF2α S51 antibodies were purchased from Cell Signaling Technologies (Worcester, MA). All the secondary antibodies (anti-rabbit-HRP, anti-mouse-HRP, and anti-goat-HRP) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Enhanced chemi-luminescence (ECL) and TUNEL kits were purchased from NEN Life Science Products (NEN Life Science Products, Boston, MA) and Boehringer Mannheim (Manheim, Germany), respectively. Trypsin-EDTA, DMEM, RPMI, penicillin-streptomycin were purchased from GIBCOBRL (GIBCOBRL Life Technologies, Grand Island, NY). HCT116 and U251 cells were purchased from the ATCC. BID -/fibroblasts were kindly provided by Dr. S. Korsmeyer (Harvard University, Boston, MA). Transformed PKR like endoplasmic reticulum (PERK) -/cells were a kind gift from the Ron Laboratory, Skirball Institute, NYU School of Medicine. Immortalized cathepsin B -/fibroblasts and matched wild type fibroblasts were kindly supplied by Christoph Peters, Thomas Reinheckel (Medizinische Universitaetsklinik Freiburg, Freiburg, Germany) and Paul Saftig (Christian-Albrechts-Universitaet Kiel, Kiel, Germany). Short hairpin RNA plasmid constructs targeting ATG5 (pLVTHM/Atg5) was a generous gift from Dr. Yousefi, Department of Pharmacology, University of Bern, Bern Switzerland and targeting Beclin-1 (pSRP-Beclin 1), kindly provided by Dr. Yuan, Department of Cell Biology, Harvard Medical School, Boston, MA (Yang et al., 2005; Levine et al., 2005; Yousefi et al., 2006; Shibata et al., 2008). The level of knock-down of these autophagy related proteins in the tumor cells was determined by western blotting with an anti-ATG5 and an anti-Beclin-1 antibody (both from Santa Cruz Biotechnology). Commercially available validated short hairpin RNA molecules to knock down RNA / protein levels were from Qiagen (Valencia, CA): ATG5 (SI02655310); Beclin 1 (SI00055573, SI00055587); AIF (SI02662114, SI02662653); caspase 4 (SI02225692); HSP70A/B (SI00442967, SI03128370, SI00442988, SI00442995). Primary human GBM cells and information on the genetic background of such cells were very kindly supplied for our use by Dr. C. David James, University of This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 7 California San Francisco, CA. The plasmids to express green fluorescent protein (GFP)tagged human LC3; (his)6 tagged hamster Grp78/BiP; dominant negative PERK (Myc-tagged PERK∆C); human HSP70 promoter linked to luciferase were kindly provided by Dr. S. Spiegel, V.C.U., Dr. A.S. Lee, USC/Norris Cancer Center, Los Angeles, CA, Dr. J.A. Diehl University of Pennsylvania, Philadelphia, PA, and Dr. S. Calderwood respectively. Other reagents and techniques were as described in (Yacoub et al., 2006; Caron et al., 2005, Yacoub et al., 2004; McKinstry et al., 2002; Fehrenbacher et al., 2004). Methods. Culture and in vitro exposure of cells to drugs. All established cell lines were cultured at 37 C (5% (v/v CO2) in vitro using RPMI supplemented with 5% (v/v) fetal calf serum and 10% (v/v) Non-essential amino acids. Primary human glioma cells were cultured in 2% (v/v) fetal calf serum to prevent growth of contaminating rodent fibroblasts during in vitro analyses. For short term cell killing assays, immunoblotting and AIF/cathepsin release studies, cells were plated at a density of 3 x 10 per cm (~2 x 10 cells per well of a 12 well plate) and 48h after plating treated with various drugs, as indicated. In vitro OSU-03012 treatment was from a 100 mM stock solution of each drug and the maximal concentration of Vehicle (DMSO) in media was 0.02% (v/v). Cells were not cultured in reduced serum media during any study in this manuscript. In vitro cell treatments, microscopy, SDS-PAGE and Western blot analysis. For in vitro analyses of short-term cell death effects, cells were treated with Vehicle or OSU-03012 for the indicated times in the Figure legends. For apoptosis assays where indicated, cells were pre-treated with vehicle (VEH, DMSO), zVAD (50 μM), calpain inhibitor (Acetyl-Calpastatin (aa184-210)) (5 μM) or Cathepsin B inhibitor ([L-3-trans(Propylcarbamoyl)oxirane-2-carbonyl]-L-isoleucyl-L-proline Methyl ester) (1 μM); cells were isolated at the indicated times, and either subjected to trypan blue cell viability assay by counting in a light microscope or fixed to slides, and stained using a commercially available Diff Quick (Geimsa) assay kit (Yacoub et al., 2006; Caron et al., 2005, Yacoub et al., 2004). OSU-03012 lethality, as judged in trypan blue exclusion assays or by This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 8 Geimsa assays, was first evident ~12h after drug exposure (data not shown). 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 (Mansfield, MA). Lysotracker visualization of lysosomes / acidic vacuoles. The Lysotracker red dye was added at 50 nM (or 75 nM depending on cell type) and incubated ~20 min. Cells were fixed in 3.4% paraformaldehyde and visualized on an Axiovert 200 fluorescent microscope under the 40x objective. For 3-methyladenine inhibition of vacuole formation, 5 mM of 3MA was added to the cells for 30 minutes before OSU-03012 was added. Cells were stained with lysotracker red as explained above. For SDS PAGE and immunoblotting, cells were plated at 5 x 10 cells / cm and treated with drugs at the indicated concentrations and after the indicated time of treatment, lysed in whole-cell lysis buffer (0.5 M Tris-HCl, pH 6.8, 2%SDS, 10% glycerol, 1% β-mercaptoethanol, 0.02% bromophenol blue), and the samples were boiled for 30 min. The boiled samples were loaded onto 10-14% SDS-PAGE and electrophoresis was run overnight. Proteins were electrophoretically transferred onto 0.22 μm nitrocellulose, and immunoblotted with various primary antibodies against different proteins. All immunoblots were visualized by ECL. For presentation, immunoblots were digitally scanned at 600 dpi using Adobe PhotoShop 7.0, and their color removed and Figures generated in MicroSoft PowerPoint. Infection of cells with recombinant adenoviruses. Cells were plated at 3 x 10 per cm in each well of a 12 well, 6 well or 60 mm plate. After plating (24h), cells were infected (at a multiplicity of infection of 50) with a control empty vector virus (CMV) and adenovirus to express HSP70, dominant negative AKT or dominant negative MEK1 (Vector Biolabs, Philadelphia, PA). Twenty four hours after infection cells were treated with the indicated concentrations of OSU-03012 and/or drugs, and cell survival or changes in expression / protein phosphorylation determined 0-48h after drug treatment by trypan blue assay and immunoblotting, respectively. This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 9 Transfection of cells with siRNA or with plasmids. For Plasmids: Cells were plated as described above and 24h after plating, transfected. For mouse embryonic fibroblasts (2-5μg) or other cell types (0.5μg) plasmids expressing a specific mRNA (or siRNA) or appropriate vector control plasmid DNA was diluted in 50μl serum-free and antibiotic-free medium (1 portion for each sample). Concurrently, 2μl Lipofectamine 2000 (Invitrogen), was diluted into 50μl of serum-free and antibioticfree medium (1 portion for each sample). Diluted DNA was added to the diluted Lipofectamine 2000 for each sample and incubated at room temperature for 30 min. This mixture was added to each well / dish of cells containing 200μl serum-free and antibiotic-free medium for a total volume of 300 μl, and the cells were incubated for 4 h at 37 °C. An equal volume of 2x medium was then added to each well. Cells were incubated for 48h, then treated with OSU-03012. For analysis of cells transfected with GFP-LC3 constructs, the GFPLC3-positive vacuolated cells were examined under the 40x objective of a Zeiss Axiovert fluorescent microscope. Forty LC3-GFP-positive cells were analyzed per condition. Each vacuole was counted and the average number of vacuoles per cell for each, including cells that did not exhibit vacuolization, was calculated. For HSP70 promoter-luciferase assays: Cells, MEFs or U251, were plated as described above and 24h after plating, transfected with either a control luciferase plasmid (1μg) + β-galactosidase (β-Gal) plasmid (15ng), or Luciferase plasmid (1μg) + β-Gal plasmid (15ng) were incubated for 5 min in serum-free medium, then added to Gene Juice (EMD Biosciences, 2μl / condition), and incubated 15 min together at room temperature (Yacoub et al., 2004). This mixture was added to cells and incubated at 37 C for 24 h after which cells were treated with OSU-03012 for 0-24h, then washed 2x with PBS, and harvested in cell lysis buffer (25mM Tris phosphate pH 7.8, 2mM DTT, 2mM CDTA (trans-1,2-diaminocyclohexane-N,N,N’,N’-tetra-acetic acid), 10% glycerol and 1% (v/v) Triton X-100). The lysate was centrifuged for 5 min at 13,000 x g at 4 C to pellet debris. The luciferase assay was performed according to the manufacturer’s instructions (Promega, Madison, WI). Briefly, luciferase substrate was brought to room temperature, then added to 20μl lysate and measured immediately on a Perkin Elmer luminometer. The luciferase measurement was normalized to β-Galactosidase measurement to This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 10 control for transfection efficiency; 50μl 2x β-Galactosidase reagent (200 mM Na2HPO4/NaH2PO4, pH7.4, 2mM MgCl2, 200mM β-mercaptoethanol, 1.34 mg/mL O-nitrophenylβ-D-Galactopyranoside) was added to 50μl cell lysate and incubated at 37 C for 10 min. The product of the assay was measured at OD405. Transfection with siRNA: Cells were plated in 60 mm dishes from a fresh culture growing in log phase as described above, and 24h after plating transfected. Prior to transfection, the medium was aspirated and 1 ml serum-free medium was added to each plate. For transfection, 10 nM of the annealed siRNA targeting AIF, ATG5, Beclin 1, caspase 4 or HSP70, the positive sense control doubled stranded siRNA targeting GAPDH or the negative control (a “scrambled” sequence with no significant homology to any known gene sequences from mouse, rat or human cell lines) were used (predominantly Qiagen, Valencia, CA; occasional alternate siRNA molecules were purchased from Ambion, Inc., Austin, Texas). Ten nM siRNA (scrambled or experimental) was diluted in serum-free media. Four μl Hiperfect (Qiagen) was added to this mixture and the solution was mixed by pipetting up and down several times. This solution was incubated at room temp for 10 min, then added dropwise to each dish. The medium in each dish was swirled gently to mix, then incubated at 37 C for 2h. One ml of 10% (v/v) serum-containing medium was added to each plate, and cells were incubated at 37 C for 48h before re-plating (50 x 10 cells each) onto 12-well plates. Cells were allowed to attach overnight, then treated with OSU-03012 (0-48h). Trypan blue exclusion assays and SDS PAGE / immunoblotting analyses were then performed at the indicated time points (Yacoub et al., 2004, Caron et al., 2004). Manipulation of drug treated cells to isolate a crude cytosolic fraction. A crude membrane fraction was prepared from treated cells as described in (Yacoub et al., 2006). Briefly, cells were washed twice in ice cold isotonic HEPES buffer (10 mM HEPES pH 7.5, 200 mM mannitol, 70 mM sucrose, 1 μM EGTA, 10 μM protease inhibitor cocktail (Sigma, St. Louis, MO). Cells on ice were scraped into isotonic HEPES buffer and lysed by passing 20 times through a 25 gauge needle. Large membrane pieces, organelles and unlysed cells were removed from the suspension by centrifugation for 5 min at 120 x g. The crude granular fraction and This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 11 cytosolic fraction was obtained from by centrifugation for 30 min at 10,000 x g, leaving the cytosol as supernatant. Data analysis. Comparison of the effects of various treatments was performed following ANOVA using the Student’s t test. Differences with a p-value of < 0.05 were considered statistically significant. Experiments shown are the means of multiple individual points (± SEM). This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 12 Results. Treatment of HCT116 and U251 human tumor cells with OSU-03012 caused a dose-dependent induction of cell death that was suppressed by use of multiple transient or stable siRNA molecules to knock down AIF expression; and inhibition of cathepsin function but was not altered by incubation with the “pan”-caspase inhibitor zVAD (Figures 1A and 1B) (Yacoub et al, 2006, Hart et al., 2007). Previously we demonstrated that OSU-03012 lethality correlated with increased release of cathepsin B into the cytosol as well as its cleavage into active forms, as judged using a small molecule cathepsin inhibitor (Yacoub et al, 2006). One potential mechanism by which OSU-03012 could cause AIF release into the cytosol via cathepsin B activation was by promoting the cleavage of BID and previously we noted that the lethality of OSU-03012 was suppressed in transformed BID -/transformed fibroblasts (Yacoub et al, 2006). In the present studies we demonstrate that OSU-03012 lethality was also suppressed in cathepsin B -/fibroblasts as judged in trypan blue and Annexin-PI staining assays that correlated with reduced OSU-03012 –stimulated cleavage of BID and reduced release of AIF into the cytosol (Figures 1C and 1D). Over-expression of either BCL-XL or BCL-2 can protect mitochondria, as well as the endoplasmic reticulum (ER), from toxic stresses and the parental compound of OSU-03012, Celecoxib, has been proposed to utilize ER stress to kill malignant cells (Fehrenbacher et al., 2004). Previously, we noted that over-expression of BCL-XL modestly suppressed the toxicity of OSU-03012 in human glioma cells whilst others noted over-expression of BCL-2 in malignant hematologic cells did not suppress OSU-03012 toxicity (Yacoub et al, 2006; Tseng et al., 2005; Tseng et al., 2006; To et al., 2007). We also found in PKR-like endoplasmic reticulum kinase null (PERK -/-) fibroblasts that the ability of OSU-03012 to cause cell death was significantly reduced, whereas the toxicity of thapsigargin was enhanced in PERK -/cells (Figure 2A, data not shown (Yacoub et al, 2006)). OSU-03012 –induced release of AIF into the cytosol and cleavage of cathepsin B were PERK-dependent (Figure 2A, upper immunoblotting panels). Identical cell survival data were obtained when a truncated dominant negative PERK protein was stably expressed in K562 leukemic cells treated with OSU-03012 (data not shown). In contrast to This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 13 loss of PERK function, expression of a dominant negative eIF2α protein (eIF2α S51A) did not significantly alter the toxicity of OSU-03012 in 3T3 fibroblasts (data not shown (Yacoub et al, 2006)). Pro-apoptotic ER stress signaling downstream of PERK – eIF2α CHOP and IRE – CHOP has been linked in a variety of cell types to changes in the activities of pro-caspase 12, pro-caspase 2 and pro-caspase 4, enzymes that can also directly promote BID cleavage / mitochondrial dysfunction and that can thus also directly activate the intrinsic apoptosis pathway (reviewed in Yeung et al., 2006, Hitomi et al., 2004). Caspases 2 and 4 have been noted to be relatively refractory to the protective actions of the “pan”-caspase inhibitor zVAD (Ekert et al., 1999). OSU-03012 promoted pro-caspase 4 degradation which was suppressed in PERK -/MEFs, and caused pro-caspase 2 degradation that was largely PERK-independent (Figure 2B, upper inset panel (i)). Transient knock down of caspase 4 expression using a short hairpin RNA in U251 cells suppressed OSU-03012 toxicity; of note, following OSU-03012 treatment, loss of caspase 4 expression suppressed cathepsin B cleavage and inhibition of cathepsin B activity did not alter OSU-03012 –induced cleavage of pro-caspase 4; indeed, inhibition of cathepsin B appeared to promote caspase 4 activation (Figures 2B and 2C, lower inset panel (ii)). Similar data were obtained following OSU-03012 treatment in cathepsin B -/fibroblasts and in HCT116 cells, and in U251 and U937 cells with knock down of pro-caspase 4 expression using a plasmid expressed siRNA molecule (data not shown; Rahmani et al., 2007). Cathepsin B is localized in endosomes in resting glioma cells and this enzyme plays a role in cell migration, angiogenesis as well as cell death processes (Lakka et al., 2005). As OSU-03012 caused cathepsin B release into the cytosol and proteolytic cleavage of cathepsin B, and endosomal dysfunction has been linked to cell death processes we determined whether endosome function was altered by OSU-03012 treatment. OSU-03012 caused vacuolization of acidic endosomes in transformed MEFs within 6h of exposure, as judged by intense lysotracker red staining that was almost an all or nothing effect (Figure 3A). OSU-03012 did not cause vacuolization of acidic endosomes in PERK -/MEFs or in wild type MEFs treated with a non-specific This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 14 autophagy inhibitor, 3 methyl adenine (3MA) (Figure 3A, data not shown). Similar data were also obtained in U251, GBM6 and GBM12 human cancer cells (Figure 3B, data not shown). Vacuolization of the protein LC3 is one recognized marker for autophagy, and transfection of U251 glioma cells with a construct to express a green fluorescent protein (GFP) tagged LC3 protein demonstrated that OSU03012 treatment induced vacuolization of GFP tagged LC3 within 3h (Figure 3C); the appearance of LC3-GFP positive vacuoles preceded lysotracker red staining acidic endosomes by ~3h (Figure 3C) (20-23, 32-34). Expression of dominant negative PERK in glioma cells or treatment of these cells with 3MA significantly suppressed OSU-03012 –induced vacuolization of the LC3-GFP protein as well as the induction of acidic endosomes (Figure 3D, Table 1, data not shown). The ATG12-ATG5 and the ATG8 (LC3)-PE conjugation systems are interdependent and a disruption in one system has a direct negative effect on the autophagic process (Yang et al., 2005; Levine et al., 2005; Yousefi et a., 2006; Shibata et al., 2006). Beclin-1 is a functional component of the lipase signaling complex which is essential for the induction of autophagy (Yang et al., 2005; Levine et al., 2005; Yousefi et a., 2006; Shibata et al., 2006). Therefore, perturbation of the levels of ATG5 or Beclin-1 should result in reduced autophagy and the attenuation of the biological effects of OSU. To test this, differing short hairpin and plasmid expressed RNA interference approaches were used to specifically suppress ATG5 and Beclin-1 protein levels in tumor cells. Cells were transiently transfected with short hairpin RNA constructs targeting ATG5 or Beclin-1. OSU03012 treatment rapidly increased the expression of ATG5, and caused rapid complete cleavage of endogenous LC3 protein in U251 and in wild type MEFs; the OSU-03012-stimulated elevation of ATG5 expression and LC3 cleavage were not present in PERK -/MEFs (Figures 3E and 3F). Knock down of ATG5 or Beclin 1 expression in U251 cells significantly suppressed the appearance of LC3 positive vacuoles in glioma cells following OSU-03012 treatment (Figure 3E, Table 2, data not shown). Knock down of Beclin 1 or ATG5 expression suppressed the toxicity of OSU-03012 in glioma cells (Figure 3G, data not shown). This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 15 We next attempted to place the activation of cathepsin B within the context of OSU-03012 –induced vacuolization. Loss of cathepsin B expression abolished OSU-03012 –induced acidic endosome vacuolization, but did not alter the ability of OSU-03012 to cause LC3-GFP vacuolization. This data argues that the apparent secondary lysotracker red staining / acidic endosome vacuolization after OSU-03012 exposure was a cathepsin B dependent process (Figure 4). Collectively, our data demonstrate that OSU-03012 causes a PERK-dependent induction of ATG5 and Beclin 1 expression that are causal in the formation of vacuoles that contain LC3; this data suggests OSU-03012 causes an autophagic response in glioma cells and in rodent fibroblasts. Based on the observation that OSU-03012 caused cell death, in part, via endosomal dysfunction, as well as via AIF release into the cytosol (Yacoub et al., 2006), we examined whether any of these effects correlated with any parallel compensatory alterations in the expression of a protein whose functions could ameliorate the proapoptotic actions of such events i.e. HSP70 (Nylandsted et al., 2004; Ravagnan et al., 2001 Demidenko et al., 2006; Mosser et al., 2000; Gurbuxani et al., 2003; Mambula and Calderwood, 2006). Treatment of transformed MEFs with OSU-03012 rapidly increased HSP70 expression and decreased HSP90 expression, effects that were PERK-dependent (Figure 5A, sections (i) and (ii), and Figure 5B). After ~6-9h of OSU-03102 exposure, expression of BiP/Grp78 surprisingly began to decline, contrary to a “classical” ER stress response, whereas expression of Grp94 and CHOP variably changed from experiment to experiment (Figure 5A, data not shown) (Rutkowski and Kaufman, 2004; Ron, 2002). The increase in HSP70 protein expression in cells was not due to altered rates of transcription; in fibroblasts and U251 cells OSU-03012 treatment caused a modest albeit significant reduction in the activity of the HSP70 promoter (Figure 5C). Based on data in Figure 5A, we determined whether changes in HSP70 function altered the toxicity of OSU03012 using an established small molecule inhibitor of HSP70 function, NZ28. Treatment of U251 glioma cells and transformed rodent fibroblasts with NZ28 enhanced the toxicity of OSU-03012 suggesting that inhibition of HSP70 function could promote OSU-03012 toxicity (Figure 5D). As HSP70 inhibits AIF function, and in order to determine one potential protective site of HSP70 action, we made further use of U251 cells lacking AIF This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 16 expression (Ravagnan et al., 2001; Gurbuxani et al., 2003; Hart et al., 2007). In a dose-dependent fashion, inhibition of HSP70 with NZ28 enhanced OSU-03012 toxicity in vector control siRNA expressing U251 cells but did not enhance OSU-03012 toxicity in U251 cells lacking AIF expression (Figure 5E (see also Figure 1B)). This suggests at least one site of HSP70 action in protecting cells from OSU-03012 toxicity is by blocking the pro-apoptotic actions of AIF. To extend our findings, we performed further studies using molecular approaches, and determined that transient knock down of HSP70 expression increased, and over-expression of HSP70 decreased, the toxicity of OSU03012 in HCT116 and U251 tumor cells (Figures 6A-6D, data not shown). Similar data were obtained using an additional siRNA molecule to knock down HSP70 expression (data not shown). We next determined whether over-expression of HSP70 altered OSU-03012 –induced acidic endosome vacuolization (Nylandsted et al., 2004; Mambula and Calderwood, 2006). Over-expression of HSP70 suppressed the low pH endosome vacuolization response of OSU-03012 treated HCT116 and U251 cells as judged microscopically (Figure 6C, inset panel). Over-expression of Grp78/BiP, a PERK binding protein, whose expression declined after OSU03012 treatment, was noted to be protective against OSU-03012 toxicity (data not shown) (Lee, 2005). Overexpression of HSP70 suppressed the GFP-LC3 vacuolization response of OSU-03012 treated U251 cells as judged microscopically (Figure 6D, inset panel). Thus OSU-03012 causes cell death in a PERK-dependent fashion as well as increases expression of a protective protein, HSP70, in a PERK-dependent fashion, arguing that OSU-03012-induced PERK signaling promotes both cell survival and cell death processes. In prior studies by Yacoub et al. we noted that inhibition of both the ERK1/2 and PI3K pathways enhanced OSU-03012 toxicity (Yacoub et al., 2006). A drug that could potentially mediate simultaneous inhibition of ERK1/2 and PI3K signaling is the geldanamycin 17AAG; however, geldanamycins also increase expression of HSP70 which could act to protect cells from OSU-03012 toxicity (Nylandsted et al., 2004; Ravagnan et al., 2001; Demidenko et al., 2006). Simultaneous exposure of transformed wild type MEFs to 17AAG and OSUThis article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 17 03012 resulted in a greater than additive increase in cell killing (Figure 7A). This effect was abolished in PERK -/cells. In wild type MEFs, 17AAG modestly increased the expression of HSP70 whereas in PERK -/cells, 17AAG-induced HSP70 levels were very much greater (Figure 7A, inset panel). Identical data were obtained in U251 glioma cells when dominant negative PERK was expressed (Figure 7B). These findings argue that PERK signaling acts to suppress the induction of HSP70 expression after geldanamycin exposure and further emphasize the protective role of HSP70 in maintaining viability. This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 18 Discussion Previous studies have demonstrated that the novel Celecoxib derivative OSU-03012 at concentrations an order of magnitude below those stably achievable in mouse plasma, and that are without observable normal tissue toxicity, killed hematopoietic, glioblastoma, lung and colon cancer cells in vitro. In these studies, OSU-03012 toxicity was variously linked to inhibition of PDK-1 and to the induction of a form of ER stress signaling with activation of a cell death / cathepsin B – AIF pathway (Yacoub et al, 2006; Tseng et al., 2005; Tseng et al., 2006; To et al., 2007). The present studies were initiated to further elucidate the molecular mechanisms by which OSU-03012 kills transformed cells in vitro. OSU-03012 promoted a dose-dependent induction of transformed cell killing that was significantly reduced in U251 cells in which AIF expression was stably suppressed. Genetic deletion of Cathepsin B suppressed OSU03012 –induced cleavage of BID, AIF release into the cytosol and fibroblast cell killing. Loss of PERK function suppressed OSU-03012 –induced activation of cathepsin B. OSU-03012 also promoted a PERK-dependent processing of pro-caspase 4 and knock down of caspase 4 expression protected cells against OSU-03012 toxicity; cathepsin B activation was dependent upon caspase 4 in glioma cells. Collectively, these findings provide additional support to those previously reported by our laboratory and argue that OSU-03012 causes activation of multiple pro-apoptotic proteases downstream of PERK, but independently of eIF2α and CHOP, to cause cell death (Yacoub et al., 2006). OSU-03012 treatment caused a rapid ~3h-6h PERK-dependent induction of intracellular vesicles in human cancer cells and in rodent fibroblasts. The vacuolization effects included the appearance of low pH vesicles that stained for lysotracker red and also vesicles that were associated with a transfected GFP-tagged LC3 protein. OSU-03012 increased the expression of ATG5 and Beclin 1, and promoted cleavage of endogenous LC3 protein, in a PERK-dependent fashion. Knock down of either ATG5 or Beclin 1 expression significantly reduced the PERK-dependent induction of vesicles that were associated with a transfected GFP-tagged LC3 This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 19 protein. Collectively, these findings together with our prior work, strongly argue that OSU-03012 exposure causes an early autophagic response in transformed cell types that precedes the apparent AIF release into the cytosol, or morphological i.e. trypan blue positive, manifestation of cell death (Yacoub et al., 2006). Our prior findings with OSU-03012 were consistent with the hypothesis that OSU-03012 promotes lysosomal dysfunction and cathepsin protease translocation into the cytosol which catalyzes BID cleavage that in turn promotes AIF release into the cytosol (Yacoub et al., 2006). In the present studies we found that loss of PERK function blocked the vacuolization of low pH vesicles as well as those containing LC3, and the release of cathepsin B and AIF into the cytosol. Loss of cathepsin B function also suppressed low pH vesicle vacuolization arguing that the physical manifestation of the low pH vesicles was a secondary response after cathepsin B activation. In contrast, loss of cathepsin B function did not alter the induction of LC3 containing vacuoles, arguing that OSU-03012 induced true autophagic vesicles prior to causing either cathepsin B activation or to causing the induction of low pH vesicle vacuolization. It has been observed by others that during an autophagic response LC3 concentration into vacuoles increases at an earlier time point than the low pH vacuolization effects observed using lysotracker red dye. The first step in autophagy has been shown to be the envelopment of an organelle by the isolating membrane (Yang et al., 2005; Levine and Yuan, 2005; Yousefi et al., 2006; Shibata et al., 2006). This compartment is called an autophagosome, which is marked by the appearance of LC3 (Yang et al., 2005; Levine and Yuan, 2005; Yousefi et al., 2006; Shibata et al., 2006). The autophagosomes are then enveloped by the lysosomal compartment, forming autoysosomes, which are, in the case of our present studies, likely detected by the acidic vacuole stain lysotracker red. Thus, collectively, this data demonstrates that OSU-03012 causes a form of ER stress response that leads to activation of a PERK and lysosomal vacuolization -dependent cathepsin B / BID / AIF –dependent cell death pathway. Celecoxib-induced apoptosis has been argued to be ER stress dependent, with loss of GADD153 (CHOP) function preventing cell killing (Tsutsumi et al., 2006; Tsutsumi et al., 2004). In this model, simplistically, Celecoxib –induced apoptosis would be impeded in cells lacking PERK expression, which is the opposite of our This article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 20 findings with OSU-03012 treatment wherein transformed fibroblasts were more resistant to drug toxicity when PERK was not expressed. Indeed, parallel studies in wild type and PERK -/cells using OSU-03012 and thapsigargin demonstrate diametrically opposite effects in drug sensitivity. Knock down of CHOP expression modestly reduced OSU-03012 toxicity in transformed MEFs but had no effect on survival in HCT116 or U251 cells (Park and Dent, Unpublished observations). OSU-03012 decreased Grp78/BiP expression and had no affect on CHOP expression whereas in a “classical” ER stress response the expression of these proteins would be expected to rise (Rutkowski and Kaufman, 2004; Ron, 2002). Recently, we noted that the novel drug Sorafenib, which has multiple intracellular targets, caused an ER stress response that also correlated with a reduction in the expression of Grp78/BiP, suggesting that the response pattern we observed with OSU-03012 may be a characteristic of drugs which inhibit both “signaling pathways” and that also cause ER stress responses (Rahmani et al., 2007). Of additional note with respect to the actions of OSU-03012, some novel agents e.g. the proteasome inhibitor Velcade (Bortezomib) have been shown to kill tumor cells via increased expression of ER stress markers such as CHOP whilst inhibiting PERK activity and the phosphorylation of eiF2α (Nawrocki et al., 2005). It has been noted in many studies that tumor cells, when exposed to moderately toxic concentrations of therapeutic agents, exhibit compensatory survival responses by activating survival signaling pathways or by increasing the expression of certain proteins that maintain cell viability (reviewed in Grant and Dent, 2004). “Classical” ER stress signaling has also been noted to either promote a toxic response e.g. elevated CHOP expression, or promote a protective response e.g. enhanced BiP/Grp78 expression, based on the duration or intensity of the stress signal (Rahmani et al., 2007). OSU-03012 treatment of transformed human and rodent cells initially increased expression of HSP70 in parallel to the formation of autophagic vacuoles, which are both putatively protective effects, in a PERK-dependent fashion. However, OSU-03012 treatment of cells then subsequently caused a decrease in HSP90 and BiP/Grp78 expression in parallel to the release of cathepsin B and AIF into the cytosol, all of which are putatively toxic effects, and these effects also occurred in a PERKThis article has not been copyedited and formatted. The final version may differ from this version. Molecular Pharmacology Fast Forward. Published on January 8, 2008 as DOI: 10.1124/mol.107.042697 at A PE T Jornals on O cber 9, 2017 m oharm .aspeurnals.org D ow nladed from MOL 2007/42697 21 dependent fashion. In HCT116 and U251 cells over-expression of HSP70 suppressed, and knock down of HSP70 levels enhanced, OSU-03012 toxicity. Over-expression of HSP70 blocked acidic endosome vacuolization, in general agreement with the findings of others (e.g. Nylandsted et al., 2004). Collectively these findings support the hypothesis that OSU-03012 causes a form of ER stress which induces a protective PERKdependent response involving elevated expression of HSP70 and autophagic vesicle formation. This ultimately degenerates however, presumably due to prolonged PERK signaling, into a cytotoxic response with the formation of low pH acidic endosomes, and the release of cathepsin B and AIF into the cytosol, which causes a non-apoptotic form of cell death. Increased expression of HSP70 has been shown by several groups to stabilize endosomes, to suppress the apoptotic activity of AIF, and collectively to promote cell survival (Nylandsted et al., 2004; Ravagnan et al., 2001 Demidenko et al., 2006; Mosser et al., 2000; Gurbuxani et al., 2003; Mambula and Calderwood, 2006). In our analyses, we demonstrated that over-expression of HSP70 blocked the formation of GFP-LC3 vacuoles and low pH acidic endosomes following OSU-03012 exposure demonstrating that one site at which HSP70 acted to promote survival following OSU-03012 treatment was at the earliest stages of ER stress signaling. However, our data in U251 cells, in which we modified HSP70 function using the pharmacologic agent NZ28, also suggested that HSP70 could act downstream in the cell death pathway at the level of AIF. As OSU-03012 was competent to overcome the protective effects of compensatory increases in HSP70 expression it is tempting to speculate whether this compound may act as a general suppressive agent of “protective” HSP70 function at