Apoptosis Induced by Proteasome Inhibition in Cancer Cells: Predominant Role of the P53/PUMA Pathway

Oncogene (2007) 26, 1681–1692 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc ORIGINAL ARTICLE Apoptosis induced by proteasome inhibition in cancer cells: predominant role of the p53/PUMA pathway

CG Concannon1, BF Koehler1,2, Claus Reimertz2, BM Murphy1, C Bonner1, N Thurow2, MW Ward1, AVillunger 3, AStrasser 4,DKo¨ gel2,5 and JHM Prehn1,5

1Department of Physiology and Medical Physics, Royal College of Surgeons in Ireland, Dublin, Ireland; 2Experimental Neurosurgery, Centre for Neurology and Neurosurgery, Johann Wolfgang Goethe University Clinics, Theodor-Stern-Kai 7, Frankfurt/Main, Germany; 3Division of Experimental Pathophysiology and Immunology, Biocenter, Innsbruck Medical University, Innsbruck, Austria and 4The Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia

The proteasome has emerged as a novel target for Introduction antineoplastic treatment of hematological malignancies and solid tumors, including those of the central nervous The correct functioning of the ubiquitin-proteasome system. To identify cell death pathways activated in pathway is essential for the degradation of the majority response to inhibition of the proteasome system in cancer of intracellular proteins. Several key regulatory proteins cells, we treated human SH-SY5Y neuroblastoma cells involved in cell proliferation and differentiation are with the selective proteasome inhibitor (PI) epoxomicin regulated by proteasome-mediated proteolysis resulting (Epoxo). Prolonged exposure to Epoxo was associated in the activation or inhibition of specific cell signaling with increased levels of poly-ubiquitinylated proteins and pathways (Adams, 2004a). The proteasome is also p53, release of cytochrome c from the mitochondria, and central to the regulation of cell death and apoptosis. activation of caspases. Analysis of global gene expression The proapoptotic Bcl-2 family proteins Bim, Bik and using high-density oligonucleotide microarrays revealed Bid, which regulate the release of proapoptotic factors that Epoxo triggered transcriptional activation of the two such as cytochrome c and Smac from mitochondria, are Bcl-2-homology domain-3-only (BH3-only) genes p53 known substrates for targeted ubiquitination and upregulated modulator of apoptosis (PUMA) and Bim. degradation (Breitschopf et al., 2000; Marshansky Subsequent studies in PUMA- and Bim-deficient cells et al., 2001; Ley et al., 2003). Indeed, caspases, the key indicated that Epoxo-induced caspase activation and proteases activated during apoptosis, are also regulated apoptosis was predominantly PUMA-dependent. Further by the proteasome. An endogenous caspase inhibitor characterization of the transcriptional response to Epoxo family of proteins, the inhibitor of apoptosis (IAPs), not in HCT116 human colon cancer cells demonstrated that only inhibit active caspases but also target them for PUMA induction was p53-dependent; with deficiency in destruction by acting as ubiquitin ligases and promoting either p53 or PUMA significantly protected HCT116 polyubiquitination of activate caspases (Huang et al., cells against Epoxo-induced apoptosis. Our data suggest 2000; Suzuki et al., 2001). that p53 activation and the transcriptional induction of In recent years modulation of proteasomal function its target gene PUMA play an important role in with specific inhibitors has evolved as a novel target for the sensitivity of cancer cells to apoptosis induced by the treatment of cancers such as multiple myeloma proteasome inhibition, and imply that antineoplastic (Chauhan et al., 2005) as well as lung, colon and therapies with PIs might be especially useful in cancers prostate cancer (Aghajanian et al., 2002; Papandreou with functional p53. et al., 2004; Chauhan et al., 2005). Proteasome Oncogene (2007) 26, 1681–1692. doi:10.1038/sj.onc.1209974; inhibitors (PIs) have been demonstrated to overcome published online 18 September 2006 chemoresistance of tumor cells by enhancing chemosensitivity and even acting in synergy with other agents Keywords: proteasome; apoptosis; epoxomicin; Bcl-2 to induce apoptotic cell death of tumor cells (Adams, family; BH3-only protein; p53 2002; Park and Lenz, 2004). Indeed, clinical efficiency of Bortezomib (PS-341/Velcade), a reversible PI has recently been observed in multiple myeloma patients (Richardson et al., 2003). Inhibitors with a broader Correspondence: Professor JHM Prehn, Department of Physiology specificity and irreversible binding have likewise been and Medical Physics, Royal College of Surgeons in Ireland, 123 identified and are currently being tested in preclinical St Stephen’s Green, Dublin 2, Ireland. trials (Melino, 2005). One such inhibitor is epoxomicin E-mail: [email protected] 0 0 5These authors share equal senior authorship. (Epoxo), a naturally found a , b epoxyketone peptide, Received 29 March 2006; revised 27 July 2006; accepted 1 August 2006; which is a highly selective and irreversible inhibitor of published online 18 September 2006 the chymotryptic-, tryptic- and post-glutamyl peptidyl PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1682 hydrolytic-like activities of the proteasome (Meng et al., Results 1999), and was first identified and isolated from a species of actinomycetes on the basis of its antitumor activity Inhibition of proteasome function using Epoxo induces (Hanada et al., 1992). Unlike several widely used PIs, caspase activity and apoptosis Epoxo is highly specific for the proteasome without To investigate the ability of Epoxo to inhibit proteaso- affecting the activity of other non-proteasomal pro- mal activity and induce apoptosis in human SH-SY5Y teases including calpain, trypsin and cathepsin B (Meng neuroblastoma cells we treated cells with 50 nM Epoxo et al., 1999). for varying time periods. Western blot analysis of mono- In addition to the post-translational modification and poly-ubiquitinylated proteins demonstrated a time- of proapoptotic proteins, inhibition of proteasomal dependent increase in the levels of ubiquitinylated function may also regulate several transcription-depen- proteins, characteristic of proteasome dysfunction and dent processes. Indeed, the activity of several transcrip- stress, especially following 16 and 24 h treatment with tion factors, including p53 and nuclear factor kappa B 50 nM Epoxo (Figure 1a). The increased levels of (NF-kB) (Chowdary et al., 1994; Palombella et al., ubiquitinylated proteins was followed by an activation 1994), are known to be modulated by proteasome of DEVDase activity, indicative of the activation of activity. This study was undertaken to analyse the the effector caspases -3 and -7 during this process underlying transcriptional mechanisms leading to (Figure 1b). Epoxo also led to the release of cytochrome apoptosis triggered by inhibition of the proteasome c from mitochondria (Figure 1c) and increased levels of system in cancer cells. Our data demonstrates apoptosis as detected by annexin V binding (Figure 1d). that although the two Bcl-2-homology domain 3-only Further analysis using Hoechst staining revealed con- (BH3-only) genes p53 upregulated modulator of apoptosis densation and fragmentation of most nuclei. Coincu- (PUMA) and Bim are transcriptionally activated by bation with the pan-caspase inhibitor, zVAD.fmk, proteasome inhibition, p53-dependent activation of inhibited these nuclear morphological changes indicat- PUMAplays a predominant role in this type of cell ing that caspases are critically involved in proteasome death. inhibition-induced apoptosis (Figure 1e).

Figure 1 Epoxo induces proteasomal stress and apoptosis in human SH-SY5Y neuroblastoma cells. (a) Increase in the level of ubiquitinylated proteins following Epoxo treatment. SH-SY5Y cells were treated with vehicle (DMSO) for 24 h or Epoxo (50 nM) for the indicated time periods. Whole-cell lysates were analysed by Western blotting. Membranes were probed with an antibody recognizing mono- or poly-ubiquitinylated proteins. a-Tubulin served as a loading control. Similar results were obtained in two separate experiments. (b) Time course of Epoxo induced caspase-3 like activity. SH-SY5Y cells were treated with vehicle (DMSO) for 24 h or Epoxo (50 nM) for indicated time periods. Following treatment cells were harvested and whole-cell lysate prepared and used to monitor caspase-3 like activity by measuring cleavage of the fluorogenic substrate Ac-DEVD-AMC (10 mM). Data are means7s.e. from n ¼ 3 separate experiments. Different from control (Con): *, Po0.05. (c) Immunofluoresence showing the redistribution of cytochrome c from mitochondria following treatment with 50 nM Epoxo for 24 h. (d) Quantification of cells undergoing apoptosis following treating with 50 nM Epoxo for the indicated time periods. Cells were stained with using Annexin V-FITC and PI and analysed by flow cytometry. (e) Hoechst staining of SH-SY5Y cells treated with DMSO, Epoxo (50 nM) with or without or pre-treatment with zVAD.fmk (100 mM). Condensation and fragmentation of the nuclear chromatin is clearly evident in the Epoxo-treated cells.

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1683 Microarray analysis of transcriptional response to Table 1 (continued ) proteasome inhibition by Epoxo Target genes Affymetrix Up/down- In an effort to identify apoptotic pathways involved in accession no. regulated Epoxo-induced cell death we utilized high-density microarrays to investigate the transcriptional changes NAG-1 (TGF-beta occurring after Epoxo treatment. A16 h treatment was superfamily) IL-15 1036_at + chosen as a suitable time point to analyse differences in NIK 974_at + gene expression owing to the fact that at this time point, TGFbIIR 1814_at + inhibition of the proteasome was clearly evident IL-6 receptor 160032_at + (Figure 1a) but preceded the major increase in caspase PKC zeta 362_at + CAMK2G (Ca/calmodulin- 32105_f_at + activity and apoptosis (Figure 1b, d). The microarray dep. protein kinase II) experiment was performed in duplicate with the thres- Smad6 1955_s_at À hold for differentially expressed up- and downregulated IL-7 1159_at À genes set at >2.8-fold/o2.8-fold (median value of both MAPKK6 2014_s_at À samples) and >3.0-fold/ 3.0-fold in at least one of the o Apoptosis regulators duplicate samples. Using this approach we identiﬁed 296 clusterin 36780_at + genes upregulated and 209 downregulated genes follow- Fas/Apo-1 1441_s_at + ing Epoxo treatment. Table 1 provides a selective list of TRAIL-R2 34892_at + Bim 31611_s_at + Caspase 7 38281_at + Table 1 Differentially expressed genes classiﬁed into distinct functional PTDSR (phosphatidylserine 35851_g_at + subgroups receptor) PUMA1700_at + Target genes Affymetrix Up/down- accession no. regulated Cytoprotective factors Heme Oxygenase 1 33802_at + Cytoplasmic chaperones CAP-2 (antiproteinase 2) 36312_at + HSP70B 35965_at + osteopontin 34342_s_at + HSP70-2 31692_at + glutathione peroxidase 770_at + HSP40 752_s_at + GCLM (gamma-glutamyl- 33163_r_at + HSP-B1 36785_at + cysteine synthetase) HSP-H1 34398_at + NQO1 (NAD(P)H-quinone 38066_at + DNAJB4 33533_at + oxireductase 1) DNAJB2 37365_at + glutathione reductase 35130_at + DNAJB6 41234_at + Nna1 34212_at +

Transcriptional regulators Proteasome function CHOP (C/EBP zeta) 39420_at + sequestosome 1 40898_at + ATF3 1774_at + UbcH2 369_s_at + ATF2 670_s_at + 26S proteasome subunit 11 1191_s_at + c-Jun 32583_at + RFPL3 (RET ﬁnger protein- 31941_s_at + c-fos 1916_s_at + like 3) NF-IL6 (C/EBP beta) 38354_at + 26S proteasome subunit S1 1314_at + C/EBP gamma 39219_at + 26S proteasome subunit p55 1192_at + mdm2 1880_at + UFD1L (ubiquitin fusion- 811_at + c-myc 37724_at À degrad. 1 like protein) n-myc 35158_at À 26S proteasome subunit 13 32211_at + AP-2 39668_at À UBE2C 1651_at + Nrf3 34282_at À Mxi (MAX interactor 1) 33966_at À Abbreviations: EBP, enhancer binding protein; HSP, heat-shock c-myb 2042_s_at À protein; MAPKK, mitogen-activated protein kinase kinase. NFIB (CCAAT-box binding 41229_at À transcription factor)

Cell cycle regulators Rad 1776_at + differentially expressed genes organized into distinct Mad 1774_at + Waf-1 2031_s_at + functional groups. The data from the microarray GADD45 1911_s_at + experiment were conﬁrmed for selected genes using CyclinT2 32055_g_at + quantitative real-time polymerase chain reaction (PCR) PLK2 (polo-like kinase 2) 41544_at + (Supplementary Figure 1; Figure 2). Acomprehensive PLK1 (polo-like kinase 1) 37228_at À list of all differentially expressed genes is included in aurora kinase A34852_g_at À CDC2 delta T 33324_s_at À Supplementary Tables 1 and 2 (Supplementary Data). CDC2 1803_at À cyclin A1943_at À Genes involved in protein folding in the cytosol. Consis- mad2 37282_at À tent with disturbances in protein turnover and the Signal transduction accumulation of protein aggregates within the cytosol 4-1BBL (TNFSF9) 35000_at + was the transcriptional activation of several genes 1890_at + involved in aiding protein folding. In fact, heat-shock

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1684

Figure 2 Epoxo induces transcriptional activation of PUMA and Bim genes. (a) Time course examining the expression of PUMA and Bim mRNAfollowing Epoxo in SH-SY5Y neuroblastoma cells as determined by real-time qPCR analysis. Data are from n ¼ 3 independent samples expressed as mean7s.e. (b) Kinetics of PUMAand p53 (left panel) and BimEL (right panel) protein expression following Epoxo-induced proteasomal stress in SH-SY5Y neuroblastoma cells. Whole-cell extracts of cells treated for the indicated times with 50 nM Epoxo or vehicle (DMSO) were analysed by Western blotting. Membranes were probed with either speciﬁc antibodies to PUMA, p53 or Bim. a-Tubulin and actin served as a loading control. Similar results were obtained in two separate experiments.

protein (Hsp)70B, a potent molecular chaperone, was downregulated genes after proteasome inhibition were identified as the highest induced gene (>700-fold from members of the c-Myc family of transcription factors microarray data). Other members of this group induced which control expression of genes involved in cell include Hsp40, Hsp-B1, as well as several DNAJB growth and metabolism (Dang, 1999). chaperones. Significantly, all the chaperones identified are involved in protein folding within the cytosol and Cell cycle regulation. Several genes involved in the are not associated with protein folding within the control of the cell cycle were found to be differentially endoplasmic reticulum (ER). This differs from the expressed. These included cyclin T2 and the p53 target transcriptional response to ER stress, which we have genes Waf1 and Mdm2, suggesting that Epoxo treat- previously profiled in the identical cellular system ment is associated with induction of p53-dependent (Reimertz et al., 2003). Here, proteasome inhibition cell signaling and induction of cell cycle arrest. Given does appear to induce immunoglobulin heavy chain- that p53 expression levels and signaling is regulated binding protein/78 kDa glucose-regulated protein ex- by ubiquitination by Mdm2, the identification of pression although the impairment of ER-associated p53-dependent targets is consistent with inhibition of protein degradation (ERAD) may result in the accumu- proteasomal function. lation of misfolded proteins within the ER and the subsequent induction of ER stress response. Antioxidant defence. Consistent with previously published data that suggest proteasomal stress is associated Genes involved in proteasomal function. The microarray with increases in the production of reactive oxygen data identified differential expression of several genes species (Kikuchi et al., 2003), is the transcriptional involved in the function of the proteasome. These activation of several genes involved in the defence included genes encoding the 26S proteasome subunits against oxidative stress. These include genes involved in 11, 13, S1 and p55. Indeed, previous studies have the regeneration of the antioxidant protein glutathione demonstrated transcriptional induction of proteasomal (glutathione reductase) as well as the antioxidant genes, subunits following proteasomal inhibition (Meiners heme oxygenase-1 (HO-1)andglutathione peroxidase. et al., 2003). Proteasomal inhibition induces expression of the Transcriptional regulators. Proteasomal stress resulted BH3-only proteins Bim and PUMA in increased expression of several members of the C/ Alarge functional group of upregulated genes were those enhancer binding protein (EBP) family of transcription involved in the regulation of apoptosis. Amongst those factors including C/EBP-homologous protein (CHOP), identified in this group included the death receptors Fas/ C/EBP beta and C/EBP gamma. Similarly, two mem- Apo1 and TRAIL-R2 as well as several members of the bers of the ATF transcription factor family, ATF-2 Bcl-2 family of proteins including the BH3-only mem- and ATF-3, were also induced. In addition, the micro- bers, Bim and PUMA. Transcriptional activation of array analysis also revealed a transcriptional activation BH3-only proteins plays a central role in the induction of both c-Jun and c-fos. The most prominently of apoptosis by a plethora of stimuli by creating an

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1685 imbalance between proapoptotic and antiapoptotic Bcl-2 proteins resulting in the release of cytochrome c and apoptosis (Villunger et al., 2003a; Willis and Adams, 2005). Because of the evidence of cytochrome c release in our model and in previous studies employing PIs (Wagenknecht et al., 2000; Mitsiades et al., 2002), we further characterized the transcriptional induction of Bim and PUMA following Epoxo treatment. Real-time quantitative PCR (qPCR) demonstrated an early transcriptional induction of PUMA mRNA, which was evident following 8 h treatment and peaked at 16 h (Figure 2a, left panel) in SH-SY5Y neuroblastoma cells. In contrast, the induction of Bim mRNAshowed a more delayed kinetics with a steady increase over the entire time period of 24 h (Figure 2a, right panel). This was reflected on a protein level with increased expression of PUMAevident following 16 h treatment with Epoxo, correlating with a maximal increase in the expression of p53 at 16 and 24 h (Figure 2b, left panel). Western blotting also revealed an increased expression of the Bim protein following inhibition of the proteasome (Figure 2b, right panel), before the increase in mRNA expression. This apparent discrepancy is likely caused by a decreased turnover of the BimEL protein after inhibition of the proteasome (Figure 3b, right panel), as BimEL has previously been shown to be targeted for degradation by the ubiquitin-proteasome pathway (Ley et al., 2003). Based on these observations, we conclude that enhanced Bim expression is triggered by two distinct mechanisms after proteasome inhibition: post-translational protein stabilization and a delayed induction of Bim transcription. In order to delineate whether PUMAor Bim expression were required during proteasome inhibition-induced apoptosis we isolated primary mouse embryonic fibroblasts (MEFs) from PUMAand Bim knockout mice and treated them with Epoxo. Loss of PUMAor Bim expression did not affect levels of proteasomal stress with similar levels of ubiquitinylated proteins evident in wild type (WT), PUMAÀ/À and Figure 3 (a) WT, PUMAÀ/À and BimÀ/À MEFs were treated BimÀ/À cells following Epoxo treatment (Figure 3a). with 50 nM Epoxo for 0–24 h. Whole-cell lysates were analysed by However, caspase activity was significantly attenuated in Western blotting and probed with an antibody recognizing mono- PUMAÀ/À MEFs, although loss of Bim expression had or poly-ubiquitinylated proteins. Actin served as a loading control. no significant effect (Figure 3b). To quantify the number (b) WT, PUMAÀ/À and BimÀ/À MEFs were treated with 50 nM Epoxo for 24 h. Following treatment cells were harvested and of cells undergoing apoptosis, we performed Hoechst whole-cell lysate prepared and used to monitor caspase-3 like staining 24 h post-Epoxo treatment. As observed with activity by measuring cleavage of the fluorogenic substrate Ac- the caspase activity assay, levels of apoptosis were DEVD-AMC (10 mM). Data are means7s.e. from n ¼ 4–8 inde- markedly reduced in PUMAÀ/À MEFS whereas loss of pendent samples from the same litter. *Po0.05 compared to WT Bim expression afforded no significant protection treated. (c) WT, PUMAÀ/À or BimÀ/À MEFs were treated with 50 nM Epoxo for 24 h. Cells were stained live with Hoechst 33258 (Figure 3c). Taken together these results suggest that and the percentage of cells with apoptotic nuclei was analysed as PUMAmay be a major contributor to proteasome described in Materials and Methods. Data are means7s.e. from inhibition-induced apoptosis. n ¼ 4 cultures per treatment. Experiments were performed twice with similar results. *Po0.05 compared to WT treated. Induction of PUMA is p53-dependent PUMAhas been identified as a key mediator of the p53- dependent apoptotic response (Nakano and Vousden, PUMAinduction following proteasomal stress we 2001; Yu et al., 2001). However, induction of PUMA utilized luciferase reporter assays employing a WT and following growth factor withdrawal (Han et al., 2001) a mutant PUMA promoter, where the p53 binding site and ER stress (Reimertz et al., 2003) has been is disrupted by point mutations (Han et al., 2001). demonstrated to occur in a p53-independent manner. Treatment of HeLa cells with 50 nM Epoxo for 24 h In order to investigate whether p53 was required for induced significant elevation of luciferase activity

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1686 indicative of activation of the PUMA promoter Epoxo for various time periods and analysed whole-cell (Figure 4a). Use of the construct with the mutated p53 lysates for DEVDase activity. Both WT and PUMAÀ/À binding site dramatically reduced the ability of Epoxo to cells experienced similar levels of proteasomal stress as activate the PUMA promoter, suggesting that PUMA assessed by Western blotting of ubiquitinylated proteins expression during proteasomal stress is largely p53- (Figure 5a). Time course analysis of DEVDase activity dependent (Figure 4a). In order to further characterize following treatment with Epoxo revealed that the WT this response we treated WT and p53À/À HCT116 cells HCT cells induced a time-dependent activation of with 50 nM Epoxo. Western blotting demonstrated that caspases with significant levels of caspase activity the levels of mono- and poly-ubiquitinylated proteins detectable following 12 h treatment. Interestingly, the following Epoxo treatment were comparable in both PUMAÀ/À cells displayed delayed kinetics of caspase WT and p53À/À HCT cells demonstrating that both cell activation following Epoxo treatment with significant types experienced similar amounts of proteasomal levels of caspase activation only detectable following 16 h stress, independent of their p53 status (Figure 4b). On treatment. Indeed, the caspase activity in PUMAÀ/À a mRNAlevel WT HCT cells showed an early induction cells was significantly attenuated compared to the levels of PUMA mRNAfollowing Epoxo treatment. How- in WT cells especially at the 16–24 h time point ever, this induction of PUMA was effectively ablated in (Figure 5b). In order to further delineate the role of p53À/À HCT cells demonstrating that PUMA induction p53 and PUMAduring proteasomal stress-induced during proteasomal stress is p53-dependent. In contrast, apoptosis we examined caspase activity and apoptosis the expression of Chop mRNAwas induced to similar in WT, p53À/À and PUMAÀ/À HCT cells treated with levels in both cell types irrespective of p53 status 50 nM Epoxo. Both caspase activity and apoptosis were (Figure 4c). significantly attenuated in both p53À/À and PUMAÀ/À HCT cells compared to HCT WT controls (Figure 5c Both p53 and PUMA are required for proteasomal and d). To further investigate the potential role of stress-induced apoptosis Bim in the residual Epoxo-induced apoptosis observed To further characterize the requirement for PUMA in p53À/À and PUMAÀ/À HCT cells, we transiently during Epoxo-induced cell death in human cancer cells knocked down Bim expression in WT, p53À/À and we treated WT and PUMAÀ/À HCT cells with 50 nM PUMAÀ/À HCT cells by RNAinterference. As

Figure 4 PUMAinduction after proteasome inhibition is p53-dependent ( a) HeLa cells were transiently transfected with either a WT 0.9 kB fragment of the PUMApromoter or p53 mutated PUMApromoter as described in the Materials and methods. At16–24 h post- transfection cells were treated with 50 nM Epoxo or DMSO (Control). Twenty-four hours later luminescence activity was monitored and normalized to the cotransfected RL-TK-luc. (b) WT and p53À/À HCT cells were treated with 50 nM Epoxo or DMSO (Control) for 16 h. Whole-cell lysates were analysed by Western blotting and probed with an antibody recognizing mono- or poly- ubiquitinylated proteins. a-Tubulin served as a loading control. Similar results were obtained in two separate experiments. (c) WT and p53À/À HCT cells were treated with DMSO (Control) or 50 nM Epoxo for 8 h. The expression of PUMA and Chop mRNAwas determined by real-time qPCR analysis. Data are from n ¼ 3 independent samples expressed as mean7s.e.

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1687

Figure 5 PUMAexpression is required for full caspase activation and apoptosis following proteasomal stress ( a) WT and PUMAÀ/À HCT cells were treated with 50 nM Epoxo or vehivle (DMSO) for 16 h. Whole-cell lysates were analysed by Western blotting and probed with an antibody recognizing mono- or poly- ubiquitinylated proteins. a-Tubulin served as a loading control. Similar results were obtained in two separate experiments. (b) Kinetics of caspase activation in HCT WT and PUMAÀ/À cells. Following treatment with vehicle or 50 nM Epoxo caspase-3 like activity in whole-cell lysates was measured by monitoring cleavage of the fluorogenic substrate Ac-DEVD-AMC. Data are means7s.e. from n ¼ 3 separate experiments; *Po0.05 compared to controls, #Po0.05 compared to HCT WT cells. (c) Caspase activity in WT, p53À/À and PUMAÀ/À cells. Cells were treated with 50 nM Epoxo for 24 h. Following treatment caspase-3 like activity in whole-cell lysates was measured by monitoring cleavage of the fluorogenic substrate Ac-DEVD-AMC. Data are means7s.e. from n ¼ 3 separate experiments; *Po0.05 compared to WT treated cells. (d) p53 and PUMA deficiency is associated with reduced levels of apoptosis. Flow cytometric analyses of cell death following Epoxo treatment in HCT WT and PUMAÀ/À. Cells were incubated with 50 nM Epoxo or vehicle (DMSO) for 24 h. Annexin-V-positive cells were measured by flow cytometry. Data are means7s.d. from n ¼ 4 cultures. Experiment was repeated twice with similar results. *Po0.05 compared to WT-treated cells. (e) siRNA-mediated knockdown of Bim expression downmodulates Epoxo-triggered Bim induction. WT HCT cells were transfected with control or Bim-specific siRNAs 24 h before treatment with 50 nM Epoxo for an additional 24 h. Whole-cell lysates were analysed by Western blotting and probed with an anti-Bim antibody. Actin served as a loading control. (f) Silencing of Bim expression does not enhance cellular resistance against Epoxo-induced cell death. WT HCT, HCT PUMAÀ/À and HCT p53À/À cells were transfected with control or Bim-specific siRNAs for 24 h before treatment with 50 nM Epoxo for an additional 24 h. Annexin-V- positive cells were measured by flow cytometry. demonstrated in Figure 5e, Bim short interfering RNA is required for Epoxo-induced caspase activation and (siRNA) dramatically reduced the increase in Bim cell death. expression following Epoxo treatment. Similar to the data in the BimÀ/À MEFs, knockdown of Bim expression did not afford any protection from Epoxo- Discussion induced cell death in WT HCT cells (Figure 5f). Interestingly, knockdown of Bim expression also did Despite abundant evidence for the therapeutic potential not confer any detectable protection from Epoxo- of PIs in a variety of malignancies, the relevant signaling induced cell death in the absence of PUMA or p53 pathways leading to apoptosis triggered by proteasome expression (Figure 5f). Taken together these results inhibition in cancer cells are not clearly defined. In this demonstrate that the p53/PUMApathway, but not Bim study, we demonstrate that treatment of neuroblastoma

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1688 cells with the highly potent and selective PI Epoxo transcriptional activation of the proapoptotic BH3-only impairs protein degradation and subsequently induces gene PUMA. Interestingly, ATF3, a recently identified activation of caspases and apoptosis. In our model positive regulator of p53 stability (Yan et al., 2005), was prolonged proteasome inhibition was associated with also found to be transcriptionally induced and may the concomitant induction of two members of the represent a positive feed back loop further amplifying proapoptotic BH3-only family, PUMAand Bim. Inter- p53 activation during proteasomal stress. Of note, it has estingly, PUMAinduction occurred in a p53-dependent been previously reported that PUMAaccounts for most manner and cells deficient in PUMAexpression were of the p53-dependent apoptosis in the HCT116 colon significantly protected from Epoxo-induced caspase carcinoma cells employed in this study (Yu et al., 2001, activation, suggesting a requirement for PUMAduring 2003). In our experiments, the extent of protection proteasomal stress-induced apoptosis. Transcriptional against Epoxo-induced cell death was very similar in activation and increased expression of BH3-only pro- p53-deficient and PUMA-deficient HCT116 cells, sug- teins is central in determining cell fate in response to a gesting that p53 mediates cell death largely through variety of stimuli, functioning as activation enablers and PUMAinduction, and that other transcriptional, transducing the cellular stress signal to other proapop- proapoptotic targets of p53 such as Fas (Owen-Schaub totic Bcl-2 family members such as Bax and Bak (for et al., 1995) and Trail-R2 (Wu et al., 1997) may play a review see Willis and Adams (2005)). less pronounced role, although they may additionally Owing to the broad number of substrates of the sensitize cells to therapies involving death ligands. proteasome, a significant number of players have been Interestingly, the most potently downregulated gene previously implicated in the proapoptotic activity of PIs, following proteasome inhibition was the oncogenic among them NF-kB, p53, Bcl-2 and caspases (for review transcription factor, c-myc, which may have a role see Adams (2004b)). NF-kB is constitutively activated in in the antitumorigenic effect of PIs. Furthermore, many cancer cells and is associated with cancer c-myc has been shown to have an inhibitory effect on development and progression (Karin, 2006). NF-kB p53-dependent target gene expression and apoptosis potently inhibits apoptosis by activation of antiapopto- (Ceballos et al., 2005). tic target genes such as Bcl-xL and IAPs)(Luoet al., Consistent with previous reports we found increased 2005). As IkBa the endogenous inhibitor of NF-kBis expression of Bim following inhibition of the protea- continuously degraded via the proteasome pathway, PIs some (Nikrad et al., 2005), however, our data argues act as indirect inhibitors of NF-kB (Karin and Ben- against a significant role of Bim in our model. Neriah, 2000). Indeed, inactivation of NF-kB has been Interestingly, in agreement with our findings a recent suggested to play a major role in the antitumor effect of study demonstrated increased Bim expression following the PI, Bortezomib (Velcade/PS-341), in multiple proteasome inhibition, although siRNA-mediated myeloma (Hideshima et al., 2002) and melanoma cells knockdown of Bim afforded no protection (Anan (Amiri et al., 2004). Despite these observations, inhibi- et al., 2006). However, it remains plausible that tion of NF-kB is not required to sensitize hepatocellular increased expression of Bim in response to PIs may be carcinoma cells, keratinocytes and lymphoma cells to of therapeutic benefit when used in combination with apoptosis (Kurland and Meyn, 2001; Leverkus et al., other agents such as TRAIL and paclitaxel (Nikrad 2003; Ganten et al., 2005). Our microarray analyses in et al., 2005; Tan et al., 2005). In our model a lack of SH-SY5Y neuroblastoma cells did not detect noticeable PUMAexpression was sufficient to offer significant expression changes in NF-kB-dependent cell death protection, suggesting that Bim induction is not able to regulators, such as Bcl-2, Bcl-xL, and the IAPs (data compensate for PUMAdeficiency as has been reported not shown), and subsequent real-time qPCR analysis in for other BH3-only proteins in some models (Villunger HCT colon cancer cells also failed to detect any et al., 2003a, b). However, although loss of Bim significant changes in the expression of these genes expression afforded no protection in the cell lines (Supplementary Figure 2). These data suggest that employed in this study it is still possible that PUMA inhibition of NF-kB might not play a major role in and Bim have overlapping action during proteaso- apoptosis induced by PIs in a significant number of mal-stress-induced apoptosis in certain cell types which cancer types. may be best addressed by studying cells from Similar controversy exists on the correlation between BimÀ/ÀPUMAÀ/À double knockout mice. Interest- the apoptosis-inducing efficiency of PIs and the p53 ingly, several recent studies have also demonstrated a status of various types of cancer cells, mainly owing to concomitant induction of Bim and PUMAfollowing the fact that this correlation seems to be cell type- treatment with the oxidative stress-inducing agents dependent. Although PIs have been shown to induce 6-OHDA(Biswas et al., 2005) and arsenite (Wong p53-independent apoptosis in glioma cells (Wagen- et al., 2005). In both scenarios the inhibition of PUMA knecht et al., 1999; Yin et al., 2005), the majority of expression was found to be sufficient to dramatically reports suggest that PIs induce a p53-dependent form of reduce the extent of apoptosis, whereas loss of Bim apoptosis in most cancer cell types, such as leukemia expression incurred no advantage with equivalent levels cells (Shinohara et al., 1996; Masdehors et al., 2000), as of cell death seen in WT and BimÀ/À cells (Biswas et al., well as melanoma and myeloma cells (Qin et al., 2005). 2005; Wong et al., 2005). Hence, scenarios exist where Consistent with previous studies we observed a promi- both PUMAand Bim are induced, but PUMAplays a nent increase of p53 levels, which was associated with more prominent role for the execution of cell death.

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1689 Curiously, these functional studies argue against a GeneArray Scanner (Agilent Technologies, Palo Alto, CA, recent study suggesting that a Bim BH3 peptide may USA) were performed as described in the Affymetrix Gene be a more potent inducer of cytochrome c release and Expression Analysis Technical Manual. The Affymetrix apoptosis than its PUMAcounterpart (Kuwana et al., Microarray Suite 5.0 was used to analyse the relative 2005), suggesting that potency of BH3-only proteins abundance of each gene from the average difference of intensities. For each individual gene, the signal ratio between may also be dependent on functional interactions perfect match and mismatch probe cells was used to determine with other molecules within the dynamic environment its ‘Absolute Call’, indicating whether the corresponding of a cell. transcript was present (P), absent (A), or marginal (M). For We have previously shown that cell death triggered by analysis of differential target gene expression, the ‘Difference ER stress also requires the transcriptional upregulation Call Decision Matrix’ was employed. Individual transcript of PUMA(Reimertz et al., 2003). Defects in protea- levels were ranked as either increased (I), marginally increased some-mediated protein degradation trigger similar (MI), decreased (D), marginally decreased (MD) or not cellular stress responses as observed during ER stress, changed (NC). To be considered as upregulated, the Absolute such as attenuation of protein synthesis and induction of Call of individual transcripts had to be present, and the transcription factors of the ATF and C/EBP families Difference Call had to be increased in both duplicate samples. To be considered as downregulated, the Absolute Call of (Reimertz et al., 2003) suggesting PUMAexpression is a individual transcripts had to be present in the control samples general response to mediate cell death during prolonged (DMSO), and the Difference Call had to be decreased in both defects in protein quality control and protein degrada- duplicate samples. tion. Our data emphasizes the fundamental importance of the p53/PUMApathway in determining the level of Real-time qPCR sensitivity of cancer cells to PIs. Total RNAwas extracted using the RNeasy mini Kit (Qiagen). First strand cDNAsynthesis was performed using 2 mg of total RNAas template and Moloney murine leukemia virus reverse Materials and methods transcriptase (Invitrogen) primed with 50 pmol of random hexamers. Quantitative real-time PCR was performed Materials using the LightCycler (Roche Diagnostics, Basel, Switzerland) Caspase substrate acetyl-DEVD-7-amido-4-methylcoumarin and the QuantiTech SYBR Green PCR kit (Qiagen) as (Ac-DEVD-AMC) and inhibitor z-Val-Ala-Asp (O-methyl)- per manufacturer’s protocol. Specific primers for each fluoromethyl ketone (zVAD.fmk) was from Bachem (St gene analysed were designed using Primer3 software Helen’s, UK). Epoxo and Hoechst 33258 were purchased (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). Sense from Sigma-Aldrich (Tallaght, Dublin, Ireland). All other and antisense primers were: ATCTCAGTGCAATGGCTTCC chemicals came in analytical grade purity from Sigma-Aldrich and CAATGCATTCTCCACACCAG for Bim; GACGACCTC (Tallaght, Dublin, Ireland). AACGCACAGTA and CACCTAATTGGGCTCCATCT for PUMA; GCCGAGAGAAAACAGTCCAG and GCCGAGA Cell lines and culture GAAAACAGTCCAG for WAF1; CCCCAAGATCCTGAAA Human SH-SY5Y neuroblastoma cells, HeLa, HCT116 WT, CAGA and CCGTTGCTGGACTGGATTAT for cJun; HCT116 p53À/À and HCT116 PUMAÀ/À cells were grown in GGCAAGGAGCTGAACAAGAG and GATAGTGGCGTTC Rosewell Park Memorial Institute medium 1640 medium CTCTGGA for Hsp70b; GGTCCTGTCTTCAGATGAAAA supplemented with 10% heat-inactivated fetal calf serum, TG and CTTGGTGCAGATTCACCATTC for Chop. Each 2mM glutamine, 100 U/ml penicillin, and 100 mg/ml strepto- primer pair was tested with a logarithmic dilution of a cDNAmix to generate a linear standard curve, which was used to calculate mycin, in a humidified atmosphere of 5% CO2 in air at 371C. Primary MEFs were isolated from WT, PUMAÀ/À or BimÀ/À the primer pair efficiency. The PCR reactions were performed in in a C57/BL/6 background from E14.5 day embryos using 20 ml volumes with following parameters: 951Cfor15min standard methods and were maintained in Dulbecco’s mod- followed by 40 cycles of 941C for 20 s, 591Cfor20s,721Cfor ified Eagle’s medium with 10% heat-inactivated fetal calf 20 s. The generation of specific PCR products was confirmed by serum, 100 U/ml penicillin, and 100 mg/ml streptomycin, melting curve analysis and gel electrophoresis. The data was o analysed using the Lightcycler Software 4.0 with all samples in a humidified atmosphere of 5% CO2 in air at 37 C. For all experiments fibroblasts were used between passages 1–3. normalized to b-actin. The generation of the PUMAÀ/À and BimÀ/À mice have been described previously (Bouillet et al., 1999; Villunger et al., Sodium dodecylsulfate–polyacrylamide gel electrophoresis and 2003a). Western blotting Preparation of cell lysates and Western blotting was carried Microarray analysis out as described (Reimertz et al., 2001). The resulting blots SH-SY5Y neuroblastoma cells were treated with 50 nM Epoxo were probed with a mouse monoclonal antiubiquitin antibody or vehicle (0.1% dimethylsulfoxide (DMSO)) for 16 h. Total (Affiniti, Victoria, Canada) diluted 1:1000, a mouse mono- RNAof duplicate samples was extracted using the RNeasy clonal anti-PUMA/Bbc3 antibody (Han et al., 2001) diluted Midi Kit (Qiagen, Hilden, Germany). Complementary RNA 1:1000, a rabbit polyclonal anti-Bim antibody (AAP-330; was synthesized from 40 mg of the total RNAby using the StressGen, Victoria, Canada) diluted 1:1000, or a mouse Superscript Choice Kit (Invitrogen, Paisley, UK). The cRNA monoclonal anti-a-tubulin antibody (clone DM 1A; Sigma, was prepared and biotin-labeled by in vitro transcription (Enzo Dublin, Ireland), diluted 1:5000. Horseradish peroxidase Biochemical, Farmingdale, NY, USA). Labeled cRNA was conjugated secondary antibodies (Pierce, Northumberland, fragmented and hybridized for 16 h at 451C to an HG-U95Av2 UK) were detected using SuperSignal West Pico Chemi- array (Affymetrix, Santa Clara, CA, USA). Hybridization, luminescent Substrate (Pierce) and imaged using a FujiFilm washing and staining, as well as scanning of the gene chips in a LAS-3000 imaging system (Fuji, Sheffield, UK).

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1690 Determination of caspase-3-like protease activity CUACCUUCUCGG(dTdT)-30; Bim antisense 50-CCGAGAA DEVDase (Caspase-3-like) activity was determined fluoro- GGUAGACAAUUG(dTdT)-30; control sense 50-UUCUCCG metrically using carbobenzoxy-Asp-Glu-Val-Asp-7-amino-4- AACGUGUCACGU(dTdT)-30; and control antisense 50-AC methyl-coumarin (DEVD-AMC) as substrate. Cleavage of GUGACACGUUCGGAGAA(dTdT)-30. Twenty-four hours DEVD-AMC to liberate free AMC was monitored by post-transfection cells were treated with 50 nM Epoxo or measuring fluorescence after 1 and 2 h intervals. Protein vehicle (DMSO) as control for 24 h. content was determined using the Pierce Coomassie Plus Protein assay reagent (Perbio, Northumberland, UK). Caspase Flow cytometry activity is expressed as change in fluorescent units per hour and Following treatments cells were stained with annexin V/ per microgram protein. propidium iodide (BioVision, Mountain View, CA, USA) as per manufacturer’s instructions. Samples were analysed Transient transfections and luciferase reporter gene assays immediately by flow cytometry. In all cases, a minimum of Cells were transiently transfected in 24-well plates using 104 events per sample were acquired. Flow cytometric analyses Metafectene (Biontex) as per manufacturer’s instruction. were performed on a CyFlow ML (Partec, Mu¨ nster, Germany) Transfection complexes consisted of 190 ng of a plasmid followed by analysis using FloMax software. containing 0.9 kb of the PUMApromoter (a kind gift from Dr T Chittenden) along with 10 ng of a plasmid encoding for the Renilla luciferase (phRL-TK-luc; Promega, Southampton, Statistics UK) to normalize for transfection efficiency between experi- Data are given as means7s.e. For statistical comparison, ments. At 16–24 h post-transfection cells were treated with t-test or one-way analysis of variance followed by Tukey Epoxo or vehicle for 24 h. Luciferase activity was mea- test were employed using SPSS software (SPSS GmbH sured using the Dual-Glo Luciferase Assay System (Promega) Software, Munich, Germany). P-values smaller than 0.05 were and resultant luminescence monitored using a Berthold considered to be statistically significant. Luminometer.

Hoechst staining of nuclear chromatin Abbreviations Cells cultured on 24-well plates were stained live with Hoechst 33258 (Sigma) at a ﬁnal concentration of 1 mg/ml. After incubation for 10 min, nuclear morphology was observed using PUMA, p53 upregulated modulator of apoptosis; GRP78, immunoglobulin heavy chain-binding protein/78 kDa glucose- an Eclipse TE 300 inverted microscope (Nikon, Du¨ sseldorf, Germany) and a 20 dry objective. For each time point and regulated protein; DMSO, dimethylsulfoxide; CHOP, C/EBP- treatment, a total number of 300 cells were analysed for homologous protein; Epoxo, epoxomicin; WT, wild type. apoptotic morphology in three subﬁelds of each culture. All experiments were performed at least twice with similar results. Acknowledgements

Gene silencing with siRNA The authors wish to thank Drs B Vogelstein, L Zhang, and The annealed siRNAduplexes were purchased from Sigma J Yu for p53 and PUMA-deﬁcient cells and Drs P Bouillet Proligo (Paris, France). HCT116 cells were transfected and J Adams for gifts of knockout mice. This study was with 100 nM of the appropriate duplex using Metafectene supported by grants from the DFG (PR 338/9-3 and 9-4) to (Biontex, Martinsried/Planegg, Germany) as per manufacturer’s JHMP and DK and Science Foundation Ireland (03/RP/B344) instructions. Sequences used were: Bim sense 50-CAAUUGU to JHMP.

References

Adams J. (2002). Proteasome inhibitors as new anticancer Bouillet P, Metcalf D, Huang DC, Tarlinton DM, Kay TW, drugs. Curr Opin Oncol 14: 628–634. Kontgen F et al. (1999). Proapoptotic Bcl-2 relative Adams J. (2004a). The proteasome: a suitable antineoplastic Bim required for certain apoptotic responses, leukocyte target. Nat Rev Cancer 4: 349. homeostasis, and to preclude autoimmunity. Science 286: Adams J. (2004b). The proteasome: a suitable antineoplastic 1735–1738. target. Nat Rev Cancer 4: 349–360. Breitschopf K, Zeiher AM, Dimmeler S. (2000). Ubiquitin- Aghajanian C, Soignet S, Dizon DS, Pien CS, Adams J, mediated degradation of the proapoptotic active form of Elliott PJ et al. (2002). Aphase I trial of the novel bid. Afunctional consequence on apoptosis induction. J Biol proteasome inhibitor PS341 in advanced solid tumor Chem 275: 21648–21652. malignancies. Clin Cancer Res 8: 2505–2511. Ceballos E, Munoz-Alonso MJ, Berwanger B, Acosta JC, Amiri KI, Horton LW, LaFleur BJ, Sosman JA, Richmond A. Hernandez R, Krause M et al. (2005). Inhibitory effect of (2004). Augmenting chemosensitivity of malignant melano- c-Myc on p53-induced apoptosis in leukemia cells. Microarray ma tumors via proteasome inhibition: implication for analysis reveals defective induction of p53 target genes and bortezomib (VELCADE, PS-341) as a therapeutic agent upregulation of chaperone genes. Oncogene 24: 4559–4571. for malignant melanoma. Cancer Res 64: 4912–4918. Chauhan D, Hideshima T, Anderson KC. (2005). Proteasome Anan A, Baskin-Bey ES, Bronk SF, Werneburg NW, inhibition in multiple myeloma: therapeutic implication. Shah VH, Gores GJ. (2006). Proteasome inhibition induces Annu Rev Pharmacol Toxicol 45: 465–476. hepatic stellate cell apoptosis. Hepatology 43: 335–344. Chowdary DR, Dermody JJ, Jha KK, Ozer HL. (1994). Biswas SC, Ryu E, Park C, Malagelada C, Greene LA. Accumulation of p53 in a mutant cell line defective in the (2005). Puma and p53 play required roles in death evoked in ubiquitin pathway. Mol Cell Biol 14: 1997–2003. a cellular model of Parkinson disease. Neurochem Res 30: Dang CV. (1999). c-Myc target genes involved in cell growth, 839–845. apoptosis, and metabolism. Mol Cell Biol 19: 1–11.

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1691 Ganten TM, Koschny R, Haas TL, Sykora J, Li-Weber M, Meng L, Mohan R, Kwok BH, Elofsson M, Sin N, Crews CM. Herzer K et al. (2005). Proteasome inhibition sensitizes (1999). Epoxomicin, a potent and selective proteasome hepatocellular carcinoma cells, but not human hepatocytes, inhibitor, exhibits in vivo antiinflammatory activity. Proc to TRAIL. Hepatology 42: 588–597. Natl Acad Sci USA 96: 10403–10408. Han J-w, Flemington C, Houghton AB, Gu Z, Zambetti GP, Mitsiades N, Mitsiades CS, Poulaki V, Chauhan D, Lutz RJ et al. (2001). Expression of bbc3, a pro-apoptotic Fanourakis G, Gu X et al. (2002). Molecular sequelae BH3-only gene, is regulated by diverse cell death and of proteasome inhibition in human multiple myeloma cells. survival signals. Proc Natl Acad Sci USA 98: 11318–11323. 99: 14374–14379. Hanada M, Sugawara K, Kaneta K, Toda S, Nishiyama Y, Nakano K, Vousden KH. (2001). PUMA, a novel proapopto- Tomita K et al. (1992). Epoxomicin, a new antitumor agent tic gene, is induced by p53. Mol Cell 7: 683–694. of microbial origin. J Antibiotics, 1746–1752. Nikrad M, Johnson T, Puthalalath H, Coultas L, Adams J, Hideshima T, Chauhan D, Richardson P, Mitsiades C, Kraft AS. (2005). The proteasome inhibitor bortezomib Mitsiades N, Hayashi T et al. (2002). NF-kappa B as a sensitizes cells to killing by death receptor ligand TRAIL therapeutic target in multiple myeloma. J Biol Chem 277: via BH3-only proteins Bik and Bim. Mol Cancer Ther 4: 16639–16647. 443–449. Huang H-k, Joazeiro CAP, Bonfoco E, Kamada S, Owen-Schaub LB, Zhang W, Cusack JC, Angelo LS, Leverson JD, Hunter T. (2000). The inhibitor of apoptosis, Santee SM, Fujiwara T et al. (1995). Wild-type human p53 cIAP2, functions as a ubiquitin-protein ligase and promotes and a temperature-sensitive mutant induce Fas/APO-1 in vitro monoubiquitination of caspases 3 and 7. J Biol Chem expression. Mol Cell Biol 15: 3032–3040. 275: 26661–26664. Palombella VJ, Rando OJ, Goldberg AL, Maniatis T. (1994). Karin M. (2006). Nuclear factor-[kappa]B in cancer develop- The ubiquitinproteasome pathway is required for processing ment and progression. Nature 441: 431. the NF-[kappa]B1 precursor protein and the activation of Karin M, Ben-Neriah Y. (2000). Phosphorylation meets NF-[kappa]B. Cell 78: 773. ubiquitination: the control of NF-[kappa]B activity. Annu Papandreou CN, Daliani DD, Nix D, Yang H, Madden T, Rev Immunol 18: 621–663. Wang X et al. (2004). Phase I trial of the proteasome Kikuchi S, Shinpo K, Tsuji S, Takeuchi M, Yamagishi S, inhibitor bortezomib in patients with advanced solid tumors Makita Z et al. (2003). Effect of proteasome inhibitor on with observations in androgen-independent prostate cancer. cultured mesencephalic dopaminergic neurons. Brain Res J Clin Oncol 22: 2108–2121. 964: 228. Park DJ, Lenz HJ. (2004). The role of proteasome inhibitors in Kurland JF, Meyn RE. (2001). Protease inhibitors restore solid tumors. Ann Med 36: 296–303. radiation-induced apoptosis to Bcl-2-expressing lymphoma Qin JZ, Ziffra J, Stennett L, Bodner B, Bonish BK, cells. Int J Cancer 96: 327–333. Chaturvedi V et al. (2005). Proteasome inhibitors trigger Kuwana T, Bouchier-Hayes L, Chipuk JE, Bonzon C, NOXA-mediated apoptosis in melanoma and myeloma cells. Sullivan BA, Green DR et al. (2005). BH3 domains of Cancer Res 65: 6282–6293. BH3-only proteins differentially regulate Bax-mediated Reimertz C, Kogel D, Lankiewicz S, Poppe M, Prehn JH. mitochondrial membrane permeabilization both directly (2001). Ca(2+)-induced inhibition of apoptosis in human and indirectly. Mol Cell 17: 525–535. SH-SY5Y neuroblastoma cells: degradation of apoptotic Leverkus M, Sprick MR, Wachter T, Mengling T, Baumann B, protease activating factor-1 (APAF-1). J Neurochem 78: Serfling E et al. (2003). Proteasome inhibition results in 1256–1266. TRAIL sensitization of primary keratinocytes by removing Reimertz C, Kogel D, Rami A, Chittenden T, Prehn JH. the resistance-mediating block of effector caspase matura- (2003). Gene expression during ER stress-induced apoptosis tion. Mol Cell Biol 23: 777–790. in neurons: induction of the BH3-only protein Bbc3/PUMA Ley R, Balmanno K, Hadfield K, Weston C, Cook SJ. and activation of the mitochondrial apoptosis pathway. J (2003). Activation of the ERK1/2 signaling pathway Cell Biol 162: 587–597. promotes phosphorylation and proteasome-dependent de- Richardson PG, Barlogie B, Berenson J, Singhal S, gradation of the BH3-only protein, Bim. J Biol Chem 278: Jagannath S, Irwin D et al. (2003). Aphase 2 study of 18811–18816. bortezomib in relapsed, refractory myeloma. N Engl J Med Luo JL, Kamata H, Karin M. (2005). IKK/NF-kappaB 348: 2609–2617. signaling: balancing life and death – a new approach to Shinohara K, Tomioka M, Nakano H, Tone S, Ito H, cancer therapy. J Clin Invest 115: 2625–2632. Kawashima S. (1996). Apoptosis induction resulting from Marshansky V, Wang X, Bertrand R, Luo H, Duguid W, proteasome inhibition. Biochem J 317(Part 2): 385–388. Chinnadurai G et al. (2001). Proteasomes modulate balance Suzuki Y, Nakabayashi Y, Takahashi R. (2001). Ubiquitin- among proapoptotic and antiapoptotic Bcl-2 family mem- protein ligase activity of X-linked inhibitor of apoptosis bers and compromise functioning of the electron transport protein promotes proteasomal degradation of caspase-3 and chain in leukemic cells. J Immunol 166: 3130–3142. enhances its anti-apoptotic effect in Fas-induced cell death. Masdehors P, Merle-Beral H, Maloum K, Omura S, Proc Natl Acad Sci USA 98: 8662–8667. Magdelenat H, Delic J. (2000). Deregulation of the ubiquitin Tan T-T, Degenhardt K, Nelson DA, Beaudoin B, Nieves- system and p53 proteolysis modify the apoptotic response in Neira W, Bouillet P et al. (2005). Key roles of BIM-driven B-CLL lymphocytes. Blood 96: 269–274. apoptosis in epithelial tumors and rational chemotherapy. Meiners S, Heyken D, Weller A, Ludwig A, Stangl K, Cancer Cell 7: 227. Kloetzel PM et al. (2003). Inhibition of proteasome activity Villunger A, Michalak EM, Coultas L, Mullauer F, Bock G, induces concerted expression of proteasome genes and Ausserlechner MJ et al. (2003a). p53- and drug-induced de novo formation of Mammalian proteasomes. J Biol Chem apoptotic responses mediated by BH3-only proteins Puma 278: 21517–21525. and Noxa. Science 302: 1036–1038. Melino G. (2005). Discovery of the ubiquitin proteasome Villunger A, Scott C, Bouillet P, Strasser A. (2003b). Essential system and its involvement in apoptosis. 12: 1155. role for the BH3-only protein Bim but redundant roles for

Oncogene PUMA and proteasome inhibition induced apoptosis CG Concannon et al 1692 Bax, Bcl-2, and Bcl-w in the control of granulocyte survival. Wu GS, Burns TF, McDonald III ER, Jiang W, Meng R, Blood 101: 2393–2400. Krantz ID et al. (1997). KILLER/DR5 is a DNAdamage- Wagenknecht B, Hermisson M, Eitel K, Weller M. (1999). inducible p53-regulated death receptor gene. Nat Genet 17: Proteasome inhibitors induce p53/p21-independent apopto- 141–143. sis in human glioma cells. Cell Physiol Biochem 9: 117–125. Yan C, Lu D, Hai T, Boyd DD. (2005). Activating Wagenknecht B, Hermisson M, Groscurth P, Liston P, transcription factor 3, a stress sensor, activates p53 by Krammer PH, Weller M. (2000). Proteasome inhibitor- blocking its ubiquitination. EMBO J 24: 2425–2435. induced apoptosis of glioma cells involves the processing of Yin D, Zhou H, Kumagai T, Liu G, Ong JM, Black KL et al. multiple caspases and cytochrome c release. J Neurochem 75: (2005). Proteasome inhibitor PS-341 causes cell growth 2288–2297. arrest and apoptosis in human glioblastoma multiforme Willis SN, Adams JM. (2005). Life in the balance: how BH3- (GBM). Oncogene 24: 344–354. only proteins induce apoptosis. Curr Opin Cell Biol 17: 617. Yu J, Wang Z, Kinzler KW, Vogelstein B, Zhang L. Wong HK, Fricker M, Wyttenbach A, Villunger A, Michalak (2003). PUMAmediates the apoptotic response to EM, Strasser A et al. (2005). Mutually exclusive subsets p53 in colorectal cancer cells. Proc Natl Acad Sci USA of BH3-only proteins are activated by the p53 and c-Jun 100: 1931–1936. N-terminal kinase/c-Jun signaling pathways during cortical Yu J, Zhang L, Hwang PM, Kinzler KW, Vogelstein B. (2001). neuron apoptosis induced by arsenite. Mol Cell Biol 25: PUMAinduces the rapid apoptosis of colorectal cancer 8732–8747. cells. Mol Cell 7: 673–682.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene