Oncogene (2011) 30, 575–587 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc ORIGINAL ARTICLE Caspase-3 cleaves XIAP in a positive feedback loop to sensitize melanoma cells to TRAIL-induced apoptosis

MHo¨rnle1, N Peters1, B Thayaparasingham1,HVo¨rsmann1, H Kashkar2 and D Kulms1

1Institute of Cell Biology and Immunology, University of Stuttgart, Stuttgart, Baden-Wu¨rttemberg, Germany and 2Institute for Medical Microbiology, Immunology and Hygiene, Center of Molecular Medicine Cologne, University of Cologne, Cologne, Germany

Successful treatment of melanoma is still challenging, 57 kDa protein which appears to be ubiquitously because metastasis remain chemoresistant and radio- expressed in adult and fetal tissues (Hunter et al., resistant. Accordingly, combinational treatments invol- 2007). XIAP consists of three BIR domains (baculoviral ving death ligands are mandatory. In a recent study from IAP repeat 1–3), a RING-finger domain conferring our lab, the majority out of 18 melanoma cell lines ubiquitin (Ub) protein ligase (E3) activity and a recently remained resistant against treatment with the death ligand discovered evolutionarily conserved Ub-acceptor (UBA) TRAIL (tumor necrosis factor related apoptosis inducing domain (Gyrd-Hansen et al., 2008; Srinivasula and ligand). Resistance was shown to be mainly due to Ashwell, 2008). XIAP was identified as the only cellular incomplete processing of caspase-3 into catalytically protein being able to bind caspase-3, -7 and -9 directly inactive p21 by binding of the anti-apoptotic protein and to inhibit both the initiation and the execution X-linked inhibitor of apoptosis protein (XIAP). Co- phase of apoptosis (Deveraux et al., 1999). Given its role irradiation with sub-lethal ultraviolet (UV) B caused in apoptosis and its frequently elevated expression in depletion of XIAP resulting in synergistic sensitization of malignant cells, XIAP has been considered as a all but two melanoma cell lines to TRAIL. We show here promising therapeutic target in mammalian cancer. the XIAP depletion to essentially require initial caspase- Consequently, research efforts have lately been focused mediated cleavage, which promotes proteasomal degrada- on the development of drugs targeting XIAP (IAP tion of XIAP. Utilizing specific caspase inhibitors and inhibitors), which are currently being evaluated in small interfering RNA-mediated knockdown, we further clinical trials (Schimmer et al., 2006). identified caspase-3 to be responsible for performing the Pro-apoptotic second mitochondria-derived activator initial cleavage of XIAP after UVB treatment. Additional of caspase/direct inhibitor of apoptosis-binding protein evidence suggests an accelerated mitochondrial outer with low pI (Smac/DIABLO), being released from membrane permeabilization in response to co-treatment mitochondria during the onset of intrinsic apoptosis, is with TRAIL and UVB, which directs the release of XIAP known to functionally inhibit XIAP by distracting the antagonizing factors including Smac. Distraction of XIAP protein from caspases (Du et al., 2000; Sun et al., 2008). consequently liberates caspase-3 to autocatalytically Full autocatalytic processing of released caspases con- process into active p17. Activated caspase-3 cleaves XIAP sequently executes the apoptotic process, while the fate and further enhances its activation in a positive regulatory of XIAP is still discussed. Smac/DIABLO-dependent feedback loop. The molecular mechanism discovered here and -independent mechanisms have been described to appears to have broader implications, because cleavage of mediate ubiquitination and proteasomal degradation of XIAP was also shown to accompany cisplatin-induced XIAP thereby enhancing apoptosis (Fu et al., 2003; sensitization of melanoma cells to TRAIL. Yang and Du, 2004; Sohn et al., 2006; Liu et al., 2007; Oncogene (2011) 30, 575–587; doi:10.1038/onc.2010.434; Galban and Duckett, 2009). Vice versa, there are published online 20 September 2010 numerous reports implying E3 activity of XIAP to cause ubiquitination and immediate proteasomal degradation Keywords: caspase-3; XIAP; melanoma; apoptosis of pro-apoptotic XIAP counterparts, including cytosolic Smac/DIABLO (MacFarlane et al., 2002; Morizane et al., 2005; Rehm et al., 2006), caspase-9 (Morizane Introduction et al., 2005), caspase-7 (Creagh et al.,2004)andcaspase-3 (Suzuki et al.,2001;Chenet al., 2003), thereby directing The most intensively studied inhibitor of apoptosis them to proteasomal degradation. Taken together, the protein (IAP) family member, XIAP (X-linked IAP), is a cellular decision between pro- and anti-apoptotic re- sponses underlies the regulation of caspase activity mainly provided by a multilayered cross-talk between Correspondence: Dr D Kulms, Institute of Cell Biology and XIAP and caspases and intracellular regulator circuits Immunology, University of Stuttgart, Allmandring 31, Stuttgart, including mitochondria and proteasomes. Baden-Wu¨rttemberg 70569, Germany. E-mail: [email protected] Malignant melanoma represents one of the most Received 15 April 2010; revised 4 August 2010; accepted 16 August 2010; aggressive malignancies with a highly chemoresistant published online 20 September 2010 and radioresistant metastatic stage (Flaherty, 2006), XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 576 offering a poor prognosis for affected patients active p17 fragment (Figure 1b). To inquire whether (Lasithiotakis et al., 2006). Therefore, therapeutic XIAP depletion during apoptosis induction is dependent strategies combining chemotherapeutic drugs and tumor on proteasomal degradation, cells were treated with the selective death ligands are under investigation to proteasome inhibitor MG132 before TRAIL þ UVB improve cancer treatment (Ganten et al., 2006; Zhang stimulation. Strikingly, proteasomal inhibition regained et al., 2006; Koschny et al., 2007). An effective offence only a truncated 29 kDa form of XIAP, whereas no to malignant melanoma with the tumor selective ligand accumulation of the full-length protein could be TRAIL (tumor necrosis-related apoptosis inducing detected. This suggests proteolytic processing of XIAP ligand; Walczak et al., 1999) is curtailed by the fact to be responsible for generation of the 29 kDa XIAP that many melanoma cell lines are a priori resistant fragment, which is subsequently committed to protea- against TRAIL-induced apoptosis (Zeise et al., 2004). somal degradation (Figure 1c). This assumption was On investigating 18 melanoma cell lines, we could supported by the observation that apoptosis induc- recently show that TRAIL susceptibility was completely tion—as documented by poly (ADP-ribose) polymerase independent of the tumor progression stage, while only cleavage—clearly coincided with decrease in wt-XIAP, the minority of cell lines responded to TRAIL at all but with accumulation of the c-terminal cleavage (Thayaparasingham et al., 2009). Co-exposure to a non- product D1-241-XIAP in cells ectopically expressing apoptotic UV (ultraviolet) B dose was shown to highly both variants and treated with the proteasome inhibitor synergistically sensitize all melanoma cell lines to MG132 (Figure 1d). Of note, in primary melanocytes, TRAIL-induced apoptosis, providing an excellent tool XIAP depletion did not occur when using identical to study the mechanism underlying TRAIL resistance. TRAIL þ UVB doses and even remained absent when TRAIL insensitivity was based on incomplete caspase-3 treating these primary cells with a high dose of 500 ng/ processing into the catalytic inactive p21 fragment only, ml TRAIL and 400 J/m2 UVB (Figure 1c), proposing a while UVB-induced sensitization coincided with XIAP tumorselective mechanism. depletion, completion of caspase-3 activation yielding Based on biochemical cell-free analyses, caspases have the catalytically active p17, and consequently apoptosis been previously shown to be capable of cleaving XIAP induction (Thayaparasingham et al., 2009). Investigat- and the data provided identified a specific caspase- ing TRAIL resistance of melanoma cells and their cleavage motive in XIAP (Deveraux et al., 1999). synergistic sensitization by sub-lethal UVB, we revealed Accordingly, the contribution of caspases in XIAP a novel mechanism involving caspase-3-induced clea- processing was exemplarily investigated using the vage of XIAP as a prerequisite for its proteasomal melanoma cell line WM-115 (vertical growth phase). degradation. Our data elucidate the nodal point causing Application of pan caspase inhibitors (Q-Val-Asp(non- TRAIL resistance in melanoma and may consequently omethylated)-OPhMe)inhibitor (QVD); z-Val-Ala-Asp help to open new doors for tumorselective killing. (OMe)-FMK (zVAD)) completely prevented depletion of XIAP upon TRAIL þ UVB treatment and also circumvented stabilization of truncated XIAP upon Results concomitant proteasome inhibition (Figure 2a). Using specific inhibitors for caspase-6 (z-Val-Glu-Ile-Asp(OMe)- Caspase-mediated proteolytic cleavage of XIAP promotes FMK, (zVEID)), -8 (z-Ile-Glu-Thr-Asp(OMe)-FMK its proteasomal degradation and sensitization of (zIETD)), -3 (z-Asp-Glu-Val-Asp(OMe)-FMK, (zDEVD)) melanoma cells to TRAIL and -9 (z-Leu-Glu-His-Asp(OMe)-FMK, (zLEHD)), ana- Resistance of melanoma cells to TRAIL-induced lysis of XIAP depletion revealed that any of the tested apoptosis has been shown to be independent of the caspases could potentially be involved in XIAP processing, tumor progression stage but to correlate with XIAP- with caspase-9 being less effective (Figure 2a). Correspond- dependent inhibition of full caspase-3 processing ingly, caspase-9 inhibition had only a minor effect on (Thayaparasingham et al., 2009). Treating three mela- apoptosis inhibition compared with the other caspases noma cell lines representing the radial growth phase (Figure 2b). Following this line, ectopic expression of an (RGP), the vertical growth phase (VGP) and the XIAP variant in which the caspase cleavage motive was metastatic stage (MM) with 100 ng/ml TRAIL did not mutated(myc-XIAP-D242E)wasshowntobeprevented yield appreciable amounts of apoptosis, staying below from depletion by TRAIL þ UVB stimulation compared 10%. Co-irradiating cells with a sub-lethal UVB dose to myc-wt-XIAP (Figure 2c; short exposure), and thus (200 J/m2), however, resulted in a highly synergistic failed to show the 29 kDa cleavage product following co- apoptotic response after 16 h (Figure 1a). Strikingly, treatment with TRAIL þ UVB upon proteasomal inhibi- primary melanocytes remained fully resistant tion by MG132 (Figure 2c; long exposure). Correspond- (Figure 1a), demonstrating the impact of this treatment ingly, XIAP-D242E appeared to be more potent in as a new potent therapeutic protocol for melanoma. inhibiting apoptosis compared with XIAP-wt (Figure 2d). According to previous findings, caspase-3 processing froze at the level of the catalytically inactive p21 fragment 6 h after TRAIL treatment. Irradiation with Cleavage of XIAP is facilitated by caspase-3 sub-lethal UVB alone did not cause any caspase-3 Being aware of the fact that the so called specific caspase cleavage, whereas the co-treatment resulted in loss of inhibitors also show non-specific side effects by inhibit- XIAP and full caspase-3 processing into the catalytically ing other caspases additionally (McStay et al., 2008), we

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 577 70 WM-3211 (RGP) 60 WM-115 (VGP) WM-852 (MM) 50 prim. melanocytes

40

30 % apoptosis 20

10 R 0 Co TRAIL UVB TRAIL+UVB

(RGP) (VGP) (MM) WM-3211 WM-115 WM-852 melanocytes TRAIL: - ++- - +- + - + - + - +- + [6 h] UVB: - - ++- -++ - -++- -++[6 h] xIAP 33 kDa casp-3 21 kDa 17 kDa β-actin

(RGP) (VGP) (MM) low dose high dose WM-3211 WM-115 WM-852 prim. Mel prim. Mel TRAIL+UVB: - + + -++-++ -++- + + [6 h] MG132: - - +--+--+ --+--+ [-30 min] 55 kDa XIAP 29 kDa β-actin

TRAIL+UVB MG123: w/o 1 2 4 6 8 [h] PARP cleaved PARP myc-wt-XIAP endog. XIAP myc-Δ1-241-XIAP β-actin

Figure 1 Caspase-mediated proteolytic cleavage of XIAP promotes its proteasomal degradation and sensitizing of melanoma cells to TRAIL. (a) Primary melanocytes as well as melanoma cell lines WM-3211, WM-115 and WM-852 were treated with TRAIL (100 ng/ml) or UVB (200 J/m2) alone or in combination. After 16 h, apoptosis was determined using a Cell Death Detection enzyme-linked immunosorbent assay (ELISA). (b) Cells were treated as in a. After 6 h, processing of XIAP and caspase-3 was documented by western blot analysis. (c) Melanoma cells were pre-treated with MG132 (20 mM) for 30 min and subsequently stimulated with TRAIL (100 ng/ml) and UVB (200 J/m2). Primary melanocytes were treated with the same low doses or with a high TRAIL (500 ng/ml) and high UVB (400 J/m2) dose. After 6 h, processing of XIAP was documented by western blot analysis. (d) Cells were co-transfected with myc-wt-XIAP and the C-terminal XIAP fragment myc-D1-241-XIAP. After 24 h, transfectants pre-treated for 30 min with MG132 were stimulated with TRAIL and UVB for the indicated times. Protein status of endogenous XIAP, ectopic wt-XIAP and the D1-241-XIAP fragment was visualized by western blot analysis. In parallel, apoptosis induction was documented by poly (ADP-ribose) polymerase (PARP) cleavage. took two different approaches to further characterize fully inhibited not only upon zDEVD treatment but also the caspase being responsible for endogenous XIAP upon caspase-8 (zIETD) inhibition, whereas caspase-9 cleavage and apoptosis sensitization. We first scrutinized (zLEHD) inhibition had a minor effect on its activity the activation status as well as the cleavage pattern of as well as cleavage (Supplementary Figure S1A). As caspase-8 being activated after TRAIL treatment, caspase-8 activity in melanoma cells is extremely low caspase-9 becoming activated upon mitochondrial outer (Supplementary Figure S1B; see also Thayaparasing- membrane permeabilization and caspase-3 being di- ham et al., 2009), we rather focussed on caspase-8 rectly inhibited by XIAP upon putative selective processing, which was slightly reduced by caspase-3 inhibition. In this context, caspase-3 was shown to be inhibition (zDEVD) but not by caspase-9 inhibition

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 578 w/o +MG132 zVAD zVEID zVAD zVEID w/o zDEVD zLEHD w/o zDEVD zLEHD w/o QVD zIETD w/o QVD zIETD [-1 h] TRAIL+UVB: - + ++++++ - + + + + + + + [6 h] 55 kDa xIAP 29 kDa β-actin

90 w/o 80 +zVAD +zVEID 70 +zIETD 60 +zDEVD +zLEHD 50 40

% apoptosis 30 20 10 0 Co TRAIL+UVB

myc-wt-xIAP myc-xIAP (D242E) zVAD: - - ---++ - - ---++ MG132: - ---+-+ - ---+-+ UVB: - -+++ + + - -+++ + + TRAIL: - +++- ++- +++- ++ myc-xIAP (short exp.) endog. xIAP

(long exp.) 29 kDa β-actin

40 mock 35 xIAP-wt xIAP-D242E 30 25 20 15 % apoptosis 10 5 0 Co TRAIL UVB TRAIL+UVB Figure 2 Depletion of XIAP requires caspase-dependent cleavage. (a) WM-115 cells were pre-treated or not with MG132 for 30 min, followed by incubation with specific caspase inhibitors (20 mM) as indicated for another hour. After subsequent co-treatment with TRAIL (100 ng/ml) and UVB (20 J/m2) for 6 h, XIAP processing was documented by western blot analysis. (b) Cells were pre-treated with the specific caspase inhibitors for 1 h and apoptosis induction was determined 16 h after TRAIL, UVB and TRAIL þ UVB stimulation in a Cell Death Detection ELISA. (c) WM-115 cells overexpressing wt-XIAP or mutated XIAP-D242E were pre-treated for 1 h with pan caspase inhibitor zVAD and/or for 30 min with proteasome inhibitor MG132 and stimulated with either TRAIL or UVB alone or combining both stimuli. After 6 h, cleavage status of endogens and ectopically expressed XIAP was determined by western blot analysis. (d) Cells overexpressing wt-XIAP or mutated XIAP-D242E were stimulated with either TRAIL or UVB or both, stimuli and apoptosis determined 16 h later in a Cell Death Detection ELISA.

(zLEHD; Supplementary Figure S1B). Finally, inhibi- and activation (Supplementary Figure S1C). Accordingly, tion of caspase-8 (zIETD) as well as of caspase-3 caspase-9 seems to be least important in downstream (zDEVD) was shown to fully prevent caspase-9 processing caspase activation, which is in line with data shown in

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 579 si lacZ si casp-8 si lacZ si casp-9 TRAIL+UVB: - +-+[6 h] TRAIL+UVB: - +-+[6 h] xIAP xIAP p55 p47 casp-9 p35 p43 casp-8 β-actin p18 β -actin si lacZ si casp-3 TRAIL+UVB: - +-+[6 h] si lacZ si casp-6 xIAP TRAIL+UVB: -++ - [6 h] xIAP p32 casp-3 p33 casp-6 p21 p17 β -actin β-actin

si lacZ si xIAP w/o w/o zLEHD zIETD zVAD zDEVD [-1 h] w/o zVAD zDEVD zIETD zLEHD w/o w/o TRAIL+UVB: - + + + + + - - + + + + + [6 h] xIAP p32 casp-3 p21 p17 β-actin

mock xIAP-wt xIAP-D242E 10 10 10 9 9 9 8 8 8 7 7 7 6 6 6 5 5 5 4 4 4 3 3 3 2 n-fold caspase-3 activity 2 2 n-fold caspase-8 activity n-fold caspase-9 activity 1 1 1 0 0 0 Co TRAIL+UVB Co TRAIL+UVB Co TRAIL+UVB Figure 3 Cleavage of XIAP is facilitated by caspase-3. (a) Caspase-8, caspase-9, caspase-6 and caspase-3 were knocked down, using specific small interfering RNA (siRNA). After 72 h, cells were stimulated with TRAIL (100 ng/ml) and UVB (20 J/m2) for 6 h and endogenous XIAP status was documented by western blot analysis. (b) XIAP was knocked down using specific siRNA. After 72 h, WM-115 cells were pre-treated with the indicated caspase inhibitors for 1 h and stimulated with TRAIL (100 ng/ml) and UVB (20 J/m2). After 6 h, caspase-3 processing was documented by western blot analysis. (c) Cells overexpressing wt-XIAP or mutated XIAP-D242E were co-stimulated with TRAIL and UVB and activities of caspase-3, -8 and -9 were determined in fluorogenic caspase activation assay using specific substrates.

Figure 2a. As caspase-8-induced caspase-3 activation co-stimulation. In si-lacZ-treated cells, stimulation with is much stronger than vice versa (Supplementary TRAIL þ UVB caused depletion of XIAP and proces- Figure S1D), and in view of the fact that upon caspase- sing of any of the respective caspases (Figure 3a). 3 inhibition caspase-8 is still active but XIAP cleavage is However, only in cells lacking caspase-3, XIAP deple- fully inhibited (see Figure 2a), caspase-3 appears to be tion upon TRAIL þ UVB treatment was completely the most likely candidate in endogenously processing abolished (Figure 3a). XIAP to promote its subsequent proteasomal degradation. To exclude a possible role of XIAP in interfering with To finally prove this issue, in a different approach we the cross-talk between the initiator caspase-8 and -9 knocked down caspase-8, -9, -6 and -3, respectively, and and the effector caspase-3, we specifically knocked examined the endogenous XIAP level upon apoptotic down XIAP. Cells lacking XIAP showed pronounced

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 580 autocatalytic processing of caspase-3 into p17 upon after 6 h as documented in a cell death detection TRAIL þ UVB treatment (Figure 3b). Pan caspase enzyme-linked immunosorbent assay (Figure 4a). In inhibition and only the specific blockade of caspase-8 cells pre-treated with TRAIL alone for 16 h and fully abrogated caspase-3 processing, indicating caspase-8 subsequently exposed to sub-lethal UVB, the apoptotic to be the only initial activator of caspase-3. In contrast, response started little earlier after 4 h (Figure 4b), nicely inhibition of caspase-9 could not prevent caspase-3 coinciding with the onset of XIAP depletion and full processing in XIAP knockdown cells (compare Supple- caspase-3 processing (Figure 4c). Cytosolic Smac/ mentary Figures S1A–D). Only, caspase-3 inhibition DIABLO is known to antagonize the function of XIAP could arrest caspase-3 processing at the level of p21 even and ultimately results in enhanced caspase-3 activity, in the absence of XIAP, strongly indicating that in these which we show in turn to deplete XIAP. Therefore cell lines, final cleavage into p17 is an autocatalytic increased Smac release upon UVB exposure may be the process independent of caspase-9. In the inverse approach, underlying molecular mechanism which gives rise to overexpression of the non-cleavable XIAP mutant accelerated XIAP depletion and enhanced susceptibility (D242E), almost completely inhibited catalytic activity to TRAIL upon UVB exposure. Cell fractionation of caspase-3 (Figure 3c). Taken together, data strongly analysis showed a significant release of Smac as well as indicated that caspase-3-mediated cleavage of XIAP is cytochrome c from mitochondria upon TRAIL stimula- followed by autocatalytic processing into the active p17 tion only in conjunction with UVB exposure, being fragment, thereby playing a major role in UVB-induced indicative for mitochondrial outer membrane permea- synergistic sensitization of melanoma cells to TRAIL. bilization (Figure 4d). As cytochrome c triggers the formation of the apoptosome and caspase-9 activation, but our results so far excluded an impact of caspase-9 on Mitochondrial outer membrane permeabilization and XIAP cleavage, we focussed on the influence of Smac on XIAP antagonization promote its depletion by enhancing XIAP depletion upon UVB exposure. We therefore caspase-3 activation expressed cytosolic mature Smac lacking the mitochon- To furthermore investigate the molecular details causing drial targeting sequence (MTS), which was accom- synergistic induction of apoptosis, we analyzed the plished by using the Ub fusion system previously UVB-induced onset of cell death in TRAIL-treated cells. described to bypass the mitochondrial maturation of In co-treated cells, apoptosis was shown to be induced Smac (Hunter et al., 2003). The Ub-Smac fusion

50 50 TRAIL+UVB 16 h TRAIL

40 40

30 30

20 20 % apoptosis % apoptosis 10 10

0 0 [UVB] Co 1 h 2 h 4 h 6 h 8 h Co w/o 1 h 2 h 4 h 6 h 8 h UVB 0.5 h 16 h 0.5 h 16 h TRAIL + UVB 123456[h] + UVB TRAIL: - + + + ++++[16 h] 0.5 1 2 4 6 8 [h] cyto TRAIL: - + + + + + + + [16 h] cyto c xIAP mito

p32 cyto Smac/ DIABLO casp-3 mito p21 p17 β-actin β-actin hsp60

Figure 4 The onset of apoptosis upon UVB co-stimulation correlates with full caspase-3 processing and mitochondrial dysfunction. (a) WM-115 cells were co-treated with TRAIL (100 ng/ml) and UVB (20 J/m2) for the indicated time points and apoptosis induction was calculated in a Cell Death Detection ELISA. (b) WM-115 cells were treated for 16 h with TRAIL. Subsequently, cells were UVB exposed and onset of apoptosis determined at the indicated time points after UVB exposure. (c) Cells were treated as in a. At the indicated time points, caspase-3 processing and XIAP depletion was documented in whole-protein extracts and (d) Smac/DIABLO as well as cytochrome c release from mitochondria in cytosolic and mitochondrial protein extracts. Hsp60 served as loading control for mitochondrial extracts.

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 581 proteins contain an amino terminal human Ub fused to of the proteasome inhibitor MG132 also yielded the human wt-Smac and mature Smac (DMTS). When truncated 29 kDa fragment of XIAP (Figure 6d), expressed in the cytosol, Ub will be immediately cleaved indicating an identical molecular mechanism to underlie by an Ub-specific protease to generate the correct MTS the synergistic sensitization of melanoma cells to (mitochondrial wt-Smac) or DMTS (cytosolic mature TRAIL þ cisplatin-induced apoptosis. Smac) amino termini of Smac. In addition, a dsRed fluorescence tag was added to the C-terminus to allow for distinction within transfected cells. WM-115 cells were transfected with either mitochondrial Ub-wt-Smac, Discussion cytosolic Ub-DMTS-Smac or control Ub-DsRed (Kash- kar et al., 2006). Mitochondrial localization of Ub-wt- Owing to the early metastatic potential, successful Smac was illustrated in cells co-transfected with mito- treatment of melanoma still represents a challenge, pEGFP using immunofluorescent microscope, while because the aggressive metastasis appears to be widely Ub-DMTS-Smac and the control Ub-DsRed remained resistant against conventional treatment using chemo- cytosolic and distributed diffusely throughout the cells therapeutic drugs and radiotherapy. It is therefore (Figure 5a). Significantly, only overexpression of Ub- of utmost importance to understand the molecular DMTS-Smac appeared to enhance XIAP depletion and signaling network conveying resistance within melano- caspase-3 processing (Figure 5b), and was shown to ma cells in order to reveal common targets for effective significantly induce apoptosis in TRAIL only treated therapeutic treatment. TRAIL insensitivity of melano- cells (Figure 5c). This indicates that XIAP antagonizing ma has been shown to be based on the XIAP-mediated proteins such as cytosolic Smac/DIABLO potentiate the incomplete caspase-3 maturation. Combination treat- UVB-induced XIAP depletion and TRAIL sensitiza- ment of TRAIL with UVB sensitized melanoma cells tion. Although cytosolic Smac efficiently initiated XIAP which coincided with XIAP depletion, completion of depletion and caspase-3 activation, small interfering caspase-3 activation and efficient apoptotic cell death RNA-mediated knockdown of endogenous Smac failed (Thayaparasingham et al., 2009). The common under- to fully prevent TRAIL þ UVB-induced apoptosis lying mechanism included nuclear factor-kB-dependent (Figure 5d). This suggests that cytochrome c release repression of XIAP expression, resulting in depletion of also has an important role by inducing formation of this inhibitory protein over time. However, the mole- the apoptosome and preparing the cell for caspase-9 cular mechanism of XIAP depletion as a crucial event activation. Thus, blocking mitochondrial outer mem- taking place during early onset of apoptosis remained brane permeabilization by overexpression of bcl-2 unsolved. The present study aimed to elucidate how almost completely inhibited apoptosis induced by the UVB exposure contributes to XIAP depletion and combination of TRAIL and UVB (Figure 5e). synergistically enhances TRAIL-induced apoptosis in Together, our data showed that TRAIL stimulation melanoma cells. causes caspase-8-mediated initial caspase-3 processing The exact mechanisms of regulating XIAP expression generating the p21 caspase-3. Owing to strong XIAP in cells during apoptosis induction are controversially binding, further autocatalytic activation of caspase-3 discussed and appear to be cell type dependent. In this did not occur, which results in a lack of significant context, strong evidence exists involving proteasomal apoptotic outcome (Figure 6a). Upon co-stimulation degradation of XIAP (Yang et al., 2000; Dan et al., with another sub-lethal stimulus, namely UVB, the pro- 2004; Dohi et al., 2004; Silke et al., 2004, 2005; Shi et al., apoptotic response of mitochondria increased, and 2005; Hougardy et al., 2006; Sohn et al., 2006; Liou traces of XIAP-antagonizing factors including Smac/ et al., 2007; Liu et al., 2007). Besides the cellular DIABLO released from mitochondria– not sufficient to counterpart of XIAP, mitochondrial Smac/DIABLO induce apoptosis itself—initiate the distraction of XIAP (Liou et al., 2007), a number of different proteins were from caspase-3 and its maturation into active p17. shown to influence proteasomal degradation of XIAP, Active caspase-3 cleaves XIAP, thereby targeting it encompassing protein kinase c (PKC) (Shi et al., 2005), to proteasomal elimination (Figure 6b). To explore Notch (Liu et al., 2007), Grim (Silke et al., 2004) and whether the caspase-3-dependent cleavage of XIAP may Akt (Dan et al., 2004) as well as IAP family members have broader therapeutic implications, we co-treated cIAP 1 þ 2 (Silke et al., 2005; Cheung et al., 2008) and cells with TRAIL and a sub-lethal dose of the DNA survivin (Dohi et al., 2004). Vice versa, proteasomal damaging chemotherapeutic drug cisplatin which also inhibition in certain cell types was shown to favor preferentially induces the mitochondrial apoptotic path- downregulation of XIAP, thereby triggering apoptosis way (Liu et al., 2010). Whereas primary melanocytes (Leverkus et al., 2003; Inoue et al., 2006; Kashkar et al., remained fully resistant, WM-115 melanoma cells were 2007; Ndozangue-Touriguine et al., 2008; Blankenship shown to be resistant against 10 mM cisplatin and showed et al., 2009; Hussain et al., 2009). These contradictory only a moderate apoptotic response to 20 mM cisplatin. data have recently been expanded by several findings Combining both concentrations of cisplatin with displaying proteasome-independent ways of XIAP 100 mg/ml of TRAIL, however, caused synergistic inactivation, involving a tight XIAP-caspase cross-talk induction of apoptosis in these cells (Figure 6c), beyond simple protein–protein interaction. In this coinciding with XIAP depletion and full caspase-3 context, XIAP has previously been shown to contain processing (Figure 6d). Most importantly, application a caspase consensus site localized between BIR2 and

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 582 BIR3, which was recognized for cleavage by different mented to occur in myeloma cells upon bortezomib caspases in a cell-free system (Deveraux et al., 1999). treatment (Desplanques et al., 2009) or to take place Although, caspase-mediated XIAP cleavage was docu- in HeLa cells upon TRAIL þ MG132 application

Ub-DsRed Ub-Smac-DsRed Ub-ΔMTS-DsRed

merge

DAPI

GFP- mito

Smac- variants DsRed

vector(DsRed) Smac-wt Smac-ΔMTS TRAIL:- +- +- ++- - ++- [6 h] UVB:-++ - -++- -++- [6 h] xIAP

p32

casp-3 p21 p17 β-actin

cd e35 25 50 mock silacZ bcl-2 mock siSmac 30 Smac-wt 40 20 Smac-ΔMTS 25 30 20

15 15 20 % apoptosis

% apoptosis 10 10 10 5 % apoptosis

0 0 5 Co TRAIL UVB TRAIL+UVB Co TRAIL UVB TRAIL+UVB Smac/ bcl-2 DIABLO β α-tub. 0 -actin Co TRAIL siSmac: -++-+-+ - bcl-2: --++-+-+

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 583

TRAIL UVB TRAIL

caspase-8 mitochondria caspase-8

cytochrome c

Smac/ DIABLO caspase-9 caspase-9

xIAP caspase-3 xIAP caspase-3

cleavage

APOPTOSIS APOPTOSIS

45 WM-115 MG132 [-30 min] 40 prim. melanocyt es Cis:- - 10 20 10 20 10 20 [μM] TRAIL:-- + - + + + + [16 h] 35 XIAP (short exp.) 30 55 kDa * 25 XIAP (long exp.) 29 kDa 20 p32 % apoptosis 15 casp-3 p21 10 p17 β 5 R -actin 0

Co TRAIL Cis-10 Cis-20

TRAIL+Cis-10TRAIL+Cis-20 Figure 6 Suggested model for the positive feedback loop sensitizing melanoma cells to TRAIL-induced apoptosis. (a) Moderate caspase-8 activation but no caspase-9 activation occur in melanoma cells upon treatment with TRAIL. Downstream caspase-3 processing is arrested at the level of catalytically inactive p21 owing to XIAP binding, yielding only very slight apoptotic responses (below 10%). (b) Upon co-irradiation with sub-lethal UVB, the intrinsic mitochondrial amplification loop becomes activated over time, causing release of Smac/DIABLO and cytochrome c. Consequently, Smac-induced distraction of XIAP from caspase-3 causes initial caspase-3 processing into catalytically active p17. Low amounts of the fully activated caspase-3 then cleave XIAP for proteasomal degradation thereby initiating a positive amplification loop causing enhanced processing of caspase-3 as well as full processing of caspase-9 within the apoptosome, thereby synergistically enhancing apoptosis. (c) Primary melanocytes as well as WM- 115 melanoma cells were treated with TRAIL (100 ng/ml) or cisplatin (10 and/or 20 mM) alone or co-treated with both stimuli. After 16 h, apoptosis was determined in a Cell Death Detection ELISA. (d) Cells were pre-treated or not with the proteasome inhibitor MG132 (20 mM) for 30 min and subsequently stimulated with TRAIL or cisplatin as indicated. After 16 h, XIAP and caspase-3 processing was documented by western blot analysis. * ¼ non-specific protein band occurring upon long exposure.

Figure 5 Mitochondrial dysfunction contributes to XIAP depletion, caspase-3 processing and apoptosis enhancement. (a) Cells were cotransfected with mito-pEGFP and respective Smac-DsRed constructs (vector, Smac-wt, Smac-DMTS). After 24 h, nuclei of fixed cells were stainedwith40,6-diamidino-2-phenylindole (DAPI; 1 mg/ml) and localization of Smac was detected by immunofluorescence microscopy. Scale bars indicate 20 mm. (b) Cells transfected with the different Smac-DsRed constructs as in (a) were treated with TRAIL (100 ng/ml), UVB (20 J/m2)or both. After 6 h, XIAP depletion and caspase-3 processing was determined by western blot analysis. (c) Cells transfected as in a and b were left untreated or stimulated with TRAIL for 16 h and apoptosis was determined subsequently in a Cell Death Detection ELISA. (d) Cells transfected with either lacZ siRNA or Smac siRNA were stimulated with TRAIL or UVB or both stimuli. After 16 h, apoptosis induction was determined. (e) Cells expressing the empty pcDNA3 vector or bcl-2 were treated as in (d) and apoptosis induction was measured after 16 h.

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 584 (Hougardy et al., 2006) and to be antagonized by the to have broader implications, even in a clinical pan caspase inhibitor zVAD in keratinocytes (Leverkus perspective. et al., 2003), the underlying molecular mechanism remained illusive. Moreover, the impact of XIAP cleavage on the apoptotic process has never been Materials and methods addressed before. Using specific caspase inhibitors and small interfering RNA-mediated knockdown, we clearly Unless otherwise stated, results of Cell Death Detection demonstrate that in melanoma cells, caspase-3 is enzyme-linked immunosorbent assay and caspase activation responsible for endogenous XIAP cleavage as a pre- assays are presented as mean±s.d. of three independently requisite for proteasomal degradation and sensitization performed experiments. Western blot analysis and immuno- of melanoma cells to apoptosis following TRAIL fluorescent images show one representative out of three treatment. Using a XIAP mutant which can not be independently performed experiments. cleaved by caspase-3, we showed for the first time that specific cleavage of XIAP is a crucial event to obtain Cells and reagents enhanced TRAIL-induced apoptosis. The human melanoma cell lines WM-3211 (radial growth Stimulation of melanoma cells with sub-lethal UVB phase), WM-115 (vertical growth phase), WM-852 (MM) were alone did not induce apoptosis, and neither showed kindly provided by Dr Friedegund Meier, Department of caspase-3 processing nor XIAP depletion, as expected. Dermatology, University of Tu¨bingen. Cells were maintained in RPMI 1640 medium (Invitrogen, Karlsruhe, Federal Upon TRAIL stimulation alone, caspase-3 processing Republik of Germany (FRG)) supplemented with 10% fetal was terminated at the level of p21 because of binding by calf serum (PromoCell, Heidelberg, FRG) and 1% penicillin- XIAP. The lack of mitochondrial outer membrane streptomycin (Invitrogen) at a density of 5 Â 105 cells/ml in permeabilization in response to TRAIL treatment alone a humidified atmosphere of 5% CO2 at 37 1C. Primary and the failure of cytosolic XIAP antagonizing factors melanocytes from three patients were prepared form human resulted in constant caspase-3 inhibition by XIAP. foreskin (Edmondson et al., 1999), maintained in Melanocyte Onset of apoptosis upon co-exposure with sub-lethal Growth Medium (M2, Promocell) and assayed in passage 4. UVB coincided with mitochondrial membrane pertur- UVB irradiation was performed in phosphate-buffered saline bation and release of cytochrome c and Smac/DIABLO. using a bank of six fluorescent bulbs (TL12, Phillips, Although cytochrome c release may cause formation of Eindhoven, NL, USA), which emit most of the energy within the UVB range (290–320 nm), with an emission peak at the apoptosome, full caspase-9 activation is very likely 313 nm. Unless otherwise stated, cells were exposed to a sub- inhibited due to XIAP binding as well, because XIAP lethal UVB dose of 200 J/m2 intending not to alter cell viability was shown to bind caspase-9 with high affinity itself. Recombinant human iz-TRAIL protein, N-terminally (Deveraux et al., 1999). When, in parallel, Smac/ fused to a isoleucine-zipper motif in order to constitutively DIABLO is released from mitochondria, it induces build the trimerised active form (Ganten et al., 2006) was distraction of XIAP from caspase-3. Liberated caspase-3 kindly provided by Dr Henning Walczak, Department of then cleaves XIAP in a positive regulatory feedback Immunology, Division of Medicine, Imperial College London. loop, thereby quickly enhancing the proteasomal XIAP Cisplatin was administered at 10 or 20 mg/ml (Platinex; Bristol- turnover, amplifying its autocatalytic activation and Myers-Squibbs, Munich, FRG). The proteasome inhibitor consequently synergistically executing apoptosis via two MG132 (Merck, Darmstadt, FRG) was supplied at 20 mM, caspase inhibitors zVAD-fmk, zIETD-fmk, zLEHD-fmk, different pathways. Despite from execution of cell death zDEVD-fmk and zVEID-fmk (Merck) were added at 20 mM. via cleavage of death substrates, activated caspase-3 may now also trigger final activation of caspase-9 by cleaving XIAP bound to caspase-9 and/or by cleavage of Caspase activation assay Cells were harvested and lysed in lysis buffer (20 mM Tris, pH caspase-9 itself, thereby enhancing its catalytic activity 7.4; 200 mM NaCl; 1% Igepal and Complete protease inhibitor (Twiddy and Cain, 2007) and alleviating XIAP binding cocktail (Roche, Mannheim, FRG)) for 30 min on ice. After (Denault et al., 2007). Consequently, this positive amplifi- centrifugation, 15 mg cellular extract was combined with 50 ml cation loop requires both cytochrome c and Smac release, activity buffer (220 mM manitol; 68 mM sucrose; 10 mM Hepes, although Smac appears to be the primary trigger. pH 7.4; 5 mM Na-Pyruvat; 2 mM NaCl; 2 mM MgCl2; 2.5 mM Accordingly, cells ectopically expressing a cytosolic KH2PO4; 0.5 mM ethylene glycol tetraacetic acid; 1 mM and active variant of Smac (Smac-DMTS) showed dithiothreitol (DTT), 0.1 mM phenylmethylsulfonyl fluoride enhanced processing of caspase-3 and some XIAP and 20 mm specific caspase-3 (Ac-DEVD-AMC, Axxora, depletion following stimulation with TRAIL alone. Loerrach, FRG), caspase-9 (Ac-LEHD-AMC) or caspase-8 These findings are in good accordance with a recent (Ac-IETD-AMC) substrate. Respective substrate turnover was measured every 15 min over a 2 h time period in a FluoStar study showing that in MCF-7 cells, TRAIL-induced luminometer (BMG Lab Technologies, Offenburg, FRG) apoptosis is dependent on Smac/DIABLO release from with an exitation of 390 nm and an emission of 460 nm. mitochondria and XIAP inhibition, rather than from n-fold caspase activity was determined by calculating: emission cytochrome c release and caspase-9 activation (Maas of sample cells/emission of control cells. et al., 2010). As this positive regulatory feedback loop involving Plasmids, cloning and transfection caspase-mediated XIAP cleavage also appeared to Based on a myc-XIAP-pcDNA3 plasmid (Kashkar et al., synergistically sensitize melanoma cells to treatment 2007), the non-cleavable D242E mutant form of XIAP was with TRAIL and sub-lethal doses of cisplatin, it seems generated by site-directed mutagenesis using pfu-ultra poly-

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 585 merase (Stratagen, La Jolla, CA, USA) followed by DpnI 10 mM cytochalasin B; 220 mM mannitol; 68 mM sucrose and digestion (Fermentas Inc., Glen Burnie, MD, USA) according Complete) incubated on ice for 20 min and passed 10 times to the manufacturer’s instructions. Deletion mutant myc-D1- through a gauge needle (26G1/2). After centrifugation at 241-XIAP was generated using expand high fidelity PCR 800 r.p.m. for 10 min, cytosolic proteins and mitochondria system (Roche). Mitochondrial and cytosolic expression of were separated by centrifuging supernatants again at biologically active Smac variants (Smac-wt, Smac-DMTS) 13 000 r.p.m. for 10 min. A total of 60–80 mg of protein samples were performed using Ub fusion system (Kashkar et al., were subjected to 12–15% SDS–PAGE, blotted onto nitro- 2006), the original work was from Hunter (Hunter et al., 2003). cellulose membranes and incubated with antibodies directed Bcl-2 was ectopically expressed from pcDNA3 (Kashkar et al., against caspase-3, caspase-9, Smac, bcl-2 (#9662, #9502, 2005). #2954, #2872; Cell Signaling, Beverly, MA, USA), caspase-8, For ectopic expression, 6.5  106 WM-3211 cells were XIAP, poly (ADP-ribose) polymerase (#551248, #610716, electroporated with 25 mg of respective plasmids in 600 ml #551025; BD-Biosciences, San Jose, CA, USA) and cyto- RPMI þ 10% fetal calf serum þ 1.25% dimethyl sulfoxide and chrome c (BioCarta, San Diego, CA, USA; clone CTC05). stimulated 24 h later. Specific knockdown was facilitated by Equal loading was monitored by reprobing membranes with transfecting 6  105 cells with 20 pmol of the respective small an antibody against a-tubulin (DM1A, Neomarkers Inc., interfering RNA for caspase-8-50-GCAAUCUGUCCUUCC Freemount, CA, USA) or b-actin (Sigma, Munich, FRG). UGAATT-30, caspase-9-50-GAAGAUUGUGAACAUCUUC Horseradish peroxidase-conjugated secondary antibodies were TT-30, caspase-6-50-CGGCUCAUGGUACAUUCAATT-30, purchased from Amersham (Buckinghamshire, UK). Bands caspase-3-50-UGAUCUUACACGUGAAGAATT-30, Smac/ were visualized by applying a chemiluminescense SuperSignal DIABLO-50-CCGCCAGGAAUCACAUUCA-30, XIAP-50- detection system (PIERCE, Rockford, IL, USA). AUCCAUCCAUGGCAGAUUATT-30 or lacZ-50-GCGGCU GCCGGAAUUUACCTT-30 for control purposes, using Lipofectamine 2000 (Invitrogen) 72 h before stimulation. Immunofluorescence and microscopy Cells co-transfected with respective Smac-DsRed constructs, and mito-pEGFP (Clontech, Mountain View, CA, USA) were Determination of cell death grown on coverslips, washed with phosphate-buffered saline At 16 h after stimulation, cells were detached from dishes, and and fixed in 4% paraformaldehyde for 15 min at room apoptosis was measured using a Cell Death Detection enzyme- temperature. Nuclei were labeled with 40,6-diamidino-2-phe- linked immunosorbent assay (Roche). The enrichment of nylindole (1 mg/ml), and coverslips were mounted using mononucleosomes and oligonucleosomes released into the Fluoromount G (Southern Biotech, Birmingham, Alabama, cytoplasm of cell lysates is detected by biotinylated anti- USA). Immunofluorescence was detected by Plan-Apochromat histone- and peroxidase-coupled anti-DNA-antibodies and is  63/1.40 Oil DIC M27 objective lenses using the Apotome calculated as follows: absorbance of sample cells/absorbance setting of the Zeiss Cell Observer HS system (Carl Zeiss AG, of control cells. The enrichment factor of 2 corresponds to Goettingen, FRG) and the appropriate filter sets. Images were 10% apoptotic cells as determined by AnnexinV staining processed with Adobe photoshop. followed by fluorescence-activated cell sorting analysis.

Western blot analysis Cells were harvested and lysed in lysis buffer (50 mM Hepes, Conflict of interest pH 7.5; 150 mM NaCl; 10% glycerol; 1% Triton-X-100; 1.5 mM MgCl2;1mM ethylene glycol tetraacetic acid; 100 mM The authors declare no conflict of interest. NaF; 10 mM pyrophosphate; 0.01% NaN3 and Complete). After centrifugation, supernatants were collected and the protein content was determined by DC BioRad Protein assay kit (BioRad, Hercules, CA, USA). For mitochondrial/cyto- Acknowledgements solic fractionation, cells were lysed in cytosolic buffer (50 mM piperazine-N,N0-bis(2-ethansulfunic acid) (PIPES), pH 7.0; We thank the Deutsche Krebshilfe/108717 for funding this 50 mM KCl; 2 mM MgCl2;5mM ethylene glycol tetraacetic acid; project.

References

Blankenship JW, Varfolomeev E, Goncharov T, Fedorova AV, Dan HC, Sun M, Kaneko S, Feldman RI, Nicosia SV, Wang HG et al. Kirkpatrick DS, Izrael-Tomasevic A et al. (2009). Ubiquitin binding (2004). Akt phosphorylation and stabilization of X-linked inhibitor modulates IAP antagonist-stimulated proteasomal degradation of of apoptosis protein (XIAP). J Biol Chem 279: 5405–5412. c-IAP1 and c-IAP2(1). Biochem J 417: 149–160. Denault JB, Eckelman BP, Shin H, Pop C, Salvesen GS. (2007). Chen L, Smith L, Wang Z, Smith JB. (2003). Preservation of caspase-3 Caspase 3 attenuates XIAP (X-linked inhibitor of apoptosis subunits from degradation contributes to apoptosis evoked by protein)-mediated inhibition of caspase 9. Biochem J 405: 11–19. lactacystin: any single lysine or lysine pair of the small subunit is Desplanques G, Giuliani N, Delsignore R, Rizzoli V, Bataille R, sufficient for ubiquitination. Mol Pharmacol 64: 334–345. Barille-Nion S. (2009). Impact of XIAP protein levels on the Cheung H, Plenchette S, Kern CJ, Mahoney DJ, Korneluk RG. survival of myeloma cells. Haematologica 94: 87–93. (2008). The RING domain of cIAP1 mediates the degradation of Deveraux QL, Leo E, Stennicke HR, Welsh K, Salvesen GS, Reed JC. RING-bearing inhibitor of apoptosis proteins by distinct pathways. (1999). Cleavage of human inhibitor of apoptosis protein XIAP results in Mol Biol Cell 19: 2729–2740. fragments with distinct specificities for caspases. EMBO J 18: 5242–5251. Creagh EM, Murphy BM, Duriez PJ, Duckett CS, Martin SJ. (2004). Dohi T, Okada K, Xia F, Wilford CE, Samuel T, Welsh K et al. Smac/Diablo antagonizes ubiquitin ligase activity of inhibitor of (2004). An IAP-IAP complex inhibits apoptosis. J Biol Chem 279: apoptosis proteins. J Biol Chem 279: 26906–26914. 34087–34090.

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 586 Du C, Fang M, Li Y, Li L, Wang X. (2000). Smac, a mitochondrial Liu L, Yang C, Herzog C, Seth R, Kaushal GP. (2010). Proteasome protein that promotes cytochrome c-dependent caspase activation inhibitors prevent cisplatin-induced mitochondrial release of by eliminating IAP inhibition. Cell 102: 33–42. apoptosis-inducing factor and markedly ameliorate cisplatin ne- Edmondson SR, Russo VC, McFarlane AC, Wraight CJ, Werther GA. phrotoxicity. Biochem Pharmacol 79: 137–146. (1999). Interactions between growth hormone, insulin-like growth Liu WH, Hsiao HW, Tsou WI, Lai MZ. (2007). Notch inhibits factor I, and basic fibroblast growth factor in melanocyte growth. apoptosis by direct interference with XIAP ubiquitination and J Clin Endocrinol Metab 84: 1638–1644. degradation. EMBO J 26: 1660–1669. Flaherty KT. (2006). Chemotherapy and targeted therapy combina- Maas C, Verbrugge I, de Vries E, Savich G, van de Kooij LW, Tait SW tions in advanced melanoma. Clin Cancer Res 12: 2366s–2370s. et al. (2010). Smac/DIABLO release from mitochondria and XIAP Fu J, Jin Y, Arend LJ. (2003). Smac3, a novel Smac/DIABLO splicing inhibition are essential to limit clonogenicity of Type I tumor cells variant, attenuates the stability and apoptosis-inhibiting activity after TRAIL receptor stimulation. Cell Death Differ (doi: 10.1038/ of X-linked inhibitor of apoptosis protein. J Biol Chem 278: cdd.2010.39). 52660–52672. MacFarlane M, Merrison W, Bratton SB, Cohen GM. (2002). Galban S, Duckett CS. (2009). XIAP as a ubiquitin ligase in cellular Proteasome-mediated degradation of Smac during apoptosis: signaling. Cell Death Differ 17: 54–60. XIAP promotes Smac ubiquitination in vitro. J Biol Chem Ganten TM, Koschny R, Sykora J, Schulze-Bergkamen H, Buchler P, 277: 36611–36616. Haas TL et al. (2006). Preclinical differentiation between apparently McStay GP, Salvesen GS, Green DR. (2008). Overlapping cleavage safe and potentially hepatotoxic applications of TRAIL either alone motif selectivity of caspases: implications for analysis of apoptotic or in combination with chemotherapeutic drugs. Clin Cancer Res 12: pathways. Cell Death Differ 15: 322–331. 2640–2646. Morizane Y, Honda R, Fukami K, Yasuda H. (2005). X-linked Gyrd-Hansen M, Darding M, Miasari M, Santoro MM, Zender L, inhibitor of apoptosis functions as ubiquitin ligase toward mature Xue W et al. (2008). IAPs contain an evolutionarily conserved caspase-9 and cytosolic Smac/DIABLO. J Biochem 137: 125–132. ubiquitin-binding domain that regulates NF-kappaB as well as cell Ndozangue-Touriguine O, Sebbagh M, Merino D, Micheau O, survival and oncogenesis. Nat Cell Biol 10: 1309–1317. Bertoglio J, Breard J. (2008). A mitochondrial block and expression Hougardy BM, Maduro JH, van der Zee AG, de Groot DJ, van den of XIAP lead to resistance to TRAIL-induced apoptosis during Heuvel FA, de Vries EG et al. (2006). Proteasome inhibitor MG132 progression to metastasis of a colon carcinoma. Oncogene 27: sensitizes HPV-positive human cervical cancer cells to rhTRAIL- 6012–6022. induced apoptosis. Int J Cancer 118: 1892–1900. Rehm M, Huber HJ, Dussmann H, Prehn JH. (2006). Systems analysis Hunter AM, Kottachchi D, Lewis J, Duckett CS, Korneluk RG, of effector caspase activation and its control by X-linked inhibitor Liston P. (2003). A novel ubiquitin fusion system bypasses the of apoptosis protein. EMBO J 25: 4338–4349. mitochondria and generates biologically active Smac/DIABLO. Schimmer AD, Dalili S, Batey RA, Riedl SJ. (2006). Targeting J Biol Chem 278: 7494–7499. XIAP for the treatment of malignancy. Cell Death Differ 13: Hunter AM, LaCasse EC, Korneluk RG. (2007). The inhibitors of 179–188. apoptosis (IAPs) as cancer targets. Apoptosis 12: 1543–1568. Shi RX, Ong CN, Shen HM. (2005). Protein kinase C inhibition and Hussain AR, Ahmed M, Ahmed SO, Al Thari S, Khan AS, Razack S x-linked inhibitor of apoptosis protein degradation contribute to the et al. (2009). Proteasome inhibitor MG-132 mediated expression of sensitization effect of luteolin on tumor necrosis factor-related p27Kip1 via S-phase kinase protein 2 degradation induces cell cycle apoptosis-inducing ligand-induced apoptosis in cancer cells. Cancer coupled apoptosis in primary effusion lymphoma cells. Leuk Res 65: 7815–7823. Lymphoma 50: 1204–1213. Silke J, Kratina T, Chu D, Ekert PG, Day CL, Pakusch M et al. Inoue T, Shiraki K, Fuke H, Yamanaka Y, Miyashita K, Yamaguchi (2005). Determination of cell survival by RING-mediated regula- Y et al. (2006). Proteasome inhibition sensitizes hepatocellular tion of inhibitor of apoptosis (IAP) protein abundance. Proc Natl carcinoma cells to TRAIL by suppressing caspase inhibitors and Acad Sci USA 102: 16182–16187. AKT pathway. Anticancer Drugs 17: 261–268. Silke J, Kratina T, Ekert PG, Pakusch M, Vaux DL. (2004). Unlike Kashkar H, Deggerich A, Seeger JM, Yazdanpanah B, Wiegmann K, Diablo/smac, Grim promotes global ubiquitination and specific Haubert D et al. (2007). NF-kappaB-independent down-regulation degradation of X chromosome-linked inhibitor of apoptosis (XIAP) of XIAP by bortezomib sensitizes HL B cells against cytotoxic and neither cause apoptosis. J Biol Chem 279: 4313–4321. drugs. Blood 109: 3982–3988. Sohn D, Totzke G, Essmann F, Schulze-Osthoff K, Levkau B, Janicke Kashkar H, Seeger JM, Hombach A, Deggerich A, Yazdanpanah B, RU. (2006). The proteasome is required for rapid initiation of death Utermohlen O et al. (2006). XIAP targeting sensitizes receptor-induced apoptosis. Mol Cell Biol 26: 1967–1978. Hodgkin lymphoma cells for cytolytic T-cell attack. Blood 108: Srinivasula SM, Ashwell JD. (2008). IAPs: what’s in a name? Mol Cell 3434–3440. 30: 123–135. Kashkar H, Wiegmann K, Yazdanpanah B, Haubert D, Kronke M. Sun H, Stuckey JA, Nikolovska-Coleska Z, Qin D, Meagher JL, Qiu S (2005). Acid sphingomyelinase is indispensable for UV light-induced et al. (2008). Structure-based design, synthesis, evaluation, and Bax conformational change at the mitochondrial membrane. J Biol crystallographic studies of conformationally constrained Smac Chem 280: 20804–20813. mimetics as inhibitors of the X-linked inhibitor of apoptosis protein Koschny R, Walczak H, Ganten TM. (2007). The promise of (XIAP). J Med Chem 51: 7169–7180. TRAIL—potential and risks of a novel anticancer therapy. JMol Suzuki Y, Nakabayashi Y, Takahashi R. (2001). Ubiquitin-protein Med 85: 923–935. ligase activity of X-linked inhibitor of apoptosis protein promotes Lasithiotakis KG, Leiter U, Gorkievicz R, Eigentler T, Breuninger H, proteasomal degradation of caspase-3 and enhances its anti- Metzler G. (2006). The incidence and mortality of cutaneous apoptotic effect in Fas-induced cell death. Proc Natl Acad Sci melanoma in Southern Germany: trends by anatomic site and USA 98: 8662–8667. pathologic characteristics, 1976 to 2003. Cancer 107: 1331–1339. Thayaparasingham B, Kunz A, Peters N, Kulms D. (2009). Sensitiza- Leverkus M, Sprick MR, Wachter T, Mengling T, Baumann B, tion of melanoma cells to TRAIL by UVB-induced and Serfling E et al. (2003). Proteasome inhibition results in TRAIL NF-kappaB-mediated downregulation of xIAP. Oncogene 28: sensitization of primary keratinocytes by removing the resistance- 345–362. mediating block of effector caspase maturation. Mol Cell Biol 23: Twiddy D, Cain K. (2007). Caspase-9 cleavage, do you need it? 777–790. Biochem J 405: e1–e2. Liou JY, Matijevic-Aleksic N, Lee S, Wu KK. (2007). Prostacyclin Walczak H, Miller RE, Ariail K, Gliniak B, Griffith TS, Kubin M inhibits endothelial cell XIAP ubiquitination and degradation. et al. (1999). Tumoricidal activity of tumor necrosis factor-related J Cell Physiol 212: 840–848. apoptosis-inducing ligand in vivo. Nat Med 5: 157–163.

Oncogene XIAP cleavage by caspase-3 sensitizes melanoma to TRAIL MHo¨rnle et al 587 Yang QH, Du C. (2004). Smac/DIABLO selectively reduces the levels by ultraviolet-B radiation via downregulation of FLIP. J Invest of c-IAP1 and c-IAP2 but not that of XIAP and livin in HeLa cells. Dermatol 123: 746–754. J Biol Chem 279: 16963–16970. Zhang XD, Wu JJ, Gillespie S, Borrow J, Hersey P. (2006). Yang Y, Fang S, Jensen JP, Weissman AM, Ashwell JD. (2000). Human melanoma cells selected for resistance to apoptosis Ubiquitin protein ligase activity of IAPs and their degradation in by prolonged exposure to tumor necrosis factor-related proteasomes in response to apoptotic stimuli. Science 288: 874–877. apoptosis-inducing ligand are more vulnerable to necrotic cell Zeise E, Weichenthal M, Schwarz T, Kulms D. (2004). Resistance of death induced by cisplatin. Clin Cancer Res 12: human melanoma cells against the death ligand TRAIL is reversed 1355–1364.

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

Oncogene