Although numerous studies have implicated the IAPs (inhibitor of apoptosis proteins) in the control of apoptotic cell death, analyses of murine Iap-targeted cells have not revealed significant differences in their susceptibility to apoptosis. In the present study, we show that, under defined conditions, murine cells lacking XIAP (X-linked inhibitor of apoptosis) and c-IAP (cellular IAP) 2, but not c-IAP1, exhibit heightened apoptotic sensitivity to both intrinsic and extrinsic apoptotic stimuli.
The IAP (inhibitor of apoptosis protein) family are thought to play a variety of physiological roles in addition to their initially described function as suppressors of programmed cell death. IAPs were first discovered in the genomes of baculoviruses, and cellular orthologues were subsequently identified in an evolutionarily diverse range of organisms . IAPs are thought to exert their pro-survival effects primarily through the direct binding and inhibition of caspases, a family of aspartate-specific cysteine proteases that are the executioners of cell death. Although several members of the IAP family have been shown to perturb caspase activity, XIAP (X-linked inhibitor of apoptosis) is a more potent inhibitor of caspases than any other family member. Nevertheless, two related IAP proteins, c-IAP (cellular IAP) 1 and c-IAP2, have both been described as exhibiting anti-apoptotic activity [2,3].
XIAP contains three BIR (baculoviral IAP repeat) domains, which are the defining elements of the IAPs, as well as a C-terminal RING (really interesting new gene) domain, which has been shown to catalyse the ubiquitination of target proteins through its role as an E3 ubiquitin ligase. Two domains of XIAP are responsible for direct high-affinity binding to caspases. A region N-terminal to the second central BIR domain binds to effector caspases 3 and 7, whereas the most C-terminal BIR domain (BIR3) is specific for the binding and inhibition of caspase 9. Ectopic expression of XIAP has been shown to suppress cell death induced through either the receptor-mediated or the mitochondrial pathways, and experimental evidence strongly supports a critical role for caspase binding in this suppression (reviewed in ). XIAP is itself regulated by several inhibitory proteins, including Smac (second mitochondrial-derived activator of caspase)/DIABLO (direct IAP-binding protein with low pI), a nuclear-encoded protein which is localized to mitochondria in healthy cells, but which is released into the cytosol during apoptosis. Smac/DIABLO binds to XIAP through the same domain utilized by XIAP to bind caspases, leading to a displacement of the IAP–caspase interaction and the triggering of caspase-dependent cell death [4–7].
The characterization of genetically modified mice has frequently revealed profound insights into the function of the gene product and, in many cases, has shed light on the pathogenesis of human diseases in which the orthologous gene is targeted. Interestingly, although XIAP has been well characterized as a potent inhibitor of caspases 3, 7 and 9 in vitro and in human cell lines  as described above, primary cells derived from Xiap (also known as Miha)-null mice have been reported not to show an increased sensitivity to apoptotic stimuli . Additionally, although c-IAP2 does not appear to contain the structure required for apoptotic inhibition , murine macrophages lacking c-IAP2 are sensitive to death in an inflammatory environment . In the present study, we examined these apparently paradoxical findings in more detail by using IAP-deficient cells from matched littermate controls and defined apoptotic conditions, and found that cells deficient in XIAP and c-IAP2 exhibited a heightened sensitivity to pro-death signals.
All cells were cultured in DMEM (Dulbecco's modified Eagle's medium) (Mediatech) supplemented with 10% (v/v) FBS (fetal bovine serum) (Mediatech), 2 mM glutamine (Gibco) and 1% penicillin/streptomycin (Gibco) at 37 °C with 5% CO2. XIAP (day 12.5), c-IAP1 (day 14.5) and c-IAP2 (day 14) MEFs (mouse embryonic fibroblasts) were isolated from individual embryos from timed matings of XIAP wild-type males with XIAP heterozygous females or two c-IAP1 heterozygous mice following standard procedures. MEFs were prepared similarly from c-IAP2-deficient mice (kindly provided by Dr Robert Korneluk, Department of Pediatrics, University of Ottawa, ON, Canada). After two passages, wild-type and KO (knockout) embryos were transformed by serial infection with lentiviruses expressing Ras and E1A (early region 1A) . Experiments were performed with male XIAP wild-type and KO cells, female c-IAP1 wild-type and male c-IAP1 KO cells.
Lungs from male littermate XIAP wild-type and KO mice were removed, minced and shaken at 37 °C in RPMI 1640 medium supplemented with 5% (v/v) FBS, 1 mg/ml collagenase A (Roche) and 20 units/ml DNase (Sigma). The suspensions were then expelled through a needle (18-gauge) ten times and suspended in HBSS (Hanks balanced salt solution) (Mediatech). Red blood cells were lysed, and the remaining cells were placed in culture medium and allowed to grow. All experiments were performed using cells grown to between passages 2 and 4.
XIAP KO mice  were back-crossed in the C57BL/6 mouse strain for at least 12 generations. c-IAP1-null mice were generated as described previously . All mice were housed under specific pathogen-free conditions within the animal care facility at the University of Michigan. All experiments were approved by the The University of Michigan Committee on the Use and Care of Animals.
Reconstitution of XIAP MEFs
Stable reconstitution of Xiap-null MEFs was accomplished by infection with a lentivirus expression system (S. Galbán, C. Hwang, J. M. Rumble, K. A. Oetjen, C. W. Wright, A. Boudreault, J. Durkin, J. W. Gillard, J. B. Jaquith, S. J. Morris and C. S. Duckett, unpublished work) encoding either full-length XIAP or a D148A/W310A double mutant generated by site-directed mutagenesis.
Cells were treated in triplicate as indicated with medium, recombinant murine TNF (tumour necrosis factor) (Roche) and CHX (cycloheximide) (Sigma) or etoposide (Bristol-Meyers Squibb) for 24 h. Floating cells were collected and combined with adherent cells treated with trypsin/EDTA (Mediatech), and all were resuspended in PI (propidium iodide) buffer [2 μg/ml PI (Sigma) and 1% BSA (Sigma) in 1× PBS] for flow cytometry. Data were collected on a Beckman Coulter Cytomics FC-500 machine and analysed using FlowJo (Treestar).
Whole-cell lysates were prepared using RIPA lysis buffer [1% NP-40 (Nonidet P40), 0.5% sodium deoxycholate, 0.1% SDS, 1 mM DTT (dithiothreitol) and 1 mM PMSF in 1× PBS] supplemented with protease inhibitors. Samples were resolved by SDS/PAGE (4–12% gradient gels) (Invitrogen), transferred on to nitrocellulose membranes (Invitrogen) and blocked in 5% (w/v) non-fat dried skimmed milk powder in Tris-buffered saline containing 0.1% Tween (Bio-Rad). Membranes were incubated at room temperature (23 °C) for 1 h or overnight at 4 °C with the following antibodies: anti-(cleaved caspase 3) antibody (1:1000 dilution; Cell Signaling Technology), anti-XIAP antibody (1:1000 dilution; BD Pharmingen), anti-β-actin antibody (1:5000 dilution; Sigma) and anti-rIAP (rat IAP) antibody (1:1000 dilution; a gift from Dr Robert Korneluk). Secondary horseradish-peroxidase-conjugated anti-mouse, anti-rabbit or anti-rat antibodies (all at 1:5000 dilution; GE Healthcare) were used for 1 h at room temperature. ECL® (enhanced chemiluminescence) (GE Healthcare) and Kodak XAR film were used for visualization purposes.
siRNA (short interfering RNA)
Oligonucleotides were obtained from Invitrogen with the following sequences: c-IAP2 #1, 5′-GAGGCUUGCAAAGCUCAAAGGCAUG-3′, c-IAP2 #2, 5′-UAGAUCAUCUGACUCCUCCUCCUCG-3, and control, 5′-GCGACAAUUGCAAGUAGUCACCAUA-3. Two serial transfections were performed 24 h apart in 12-well plates with Lipofectamine™ 2000 (Invitrogen), using 4 μl of the oligonucleotide, and cells were treated as indicated 30 h after the second transfection.
RESULTS AND DISCUSSION
To address whether murine XIAP may modulate apoptosis, fibroblasts isolated from lungs of Xiap-null mice and control littermates were treated with TNF, a ligand commonly used to stimulate the extrinsic apoptotic pathway. In cell-culture systems, TNF does not kill without the addition of the protein synthesis inhibitor CHX, in part because TNF is generally thought to induce the transcription of genes that suppress apoptosis [13–15]. Interestingly, as shown in Figure 1(A), lung fibroblasts lacking XIAP were highly sensitive to treatment with TNF and CHX compared with their littermate counterparts, suggesting that XIAP does, in fact, suppress apoptosis in murine cells.
XIAP modulates apoptosis in murine fibroblasts
Murine XIAP is closely related to the human protein and contains all of the critical components for performing a function similar to that of its human counterpart . However, previous studies using Xiap-null MEFs did not reveal a sensitivity to TNF . To examine further the responsiveness of Xiap-deficient pulmonary fibroblasts, we evaluated the role of murine XIAP in receptor-mediated death using matched littermate MEFs. Intriguingly, the concentration of CHX used previously to sensitize cells to TNF-induced death was found to induce killing under these experimental conditions, even in the absence of TNF (results not shown). Using a lower concentration of CHX (0.1 μg/ml), death was potentiated through the TNF receptor without cell death occurring with CHX alone. As shown in Figure 1(B), Xiap-null MEFs were significantly more sensitive to treatment than their wild-type counterparts. Reintroduction of wild-type murine XIAP into the KO cell line demonstrated that this effect was wholly dependent on XIAP, since these cells were protected to approximately the same level as the wild-type cells. Furthermore, reintroduction of a mutated form of XIAP which cannot inhibit caspases (D148A/W310A) into the KO cell line was unable to protect the cells to the same degree as the wild-type protein. These studies were also performed in primary MEFs as well as MEFs generated from distinct embryos with essentially the same results (results not shown). These results further support those obtained from the pulmonary fibroblasts, suggesting that murine XIAP does modulate the apoptotic threshold in a similar caspase-dependent manner to the human protein.
The lack of XIAP in genetically targeted mice has been suggested to be compensated for by c-IAP1 and c-IAP2 overexpression . Therefore the levels of these proteins in MEFs from IAP-null mice were examined. As shown in Figure 2(A), the level of c-IAP1 was found to be higher in transformed fibroblasts lacking XIAP, consistent with initial reports . As described previously , c-IAP2 was up-regulated in c-Iap1-null MEFs, the only cell lysates in which c-IAP2 could be detected. Additionally, XIAP protein levels appeared to be unaffected by the absence of either c-IAP1 or c-IAP2.
Lack of XIAP or c-IAP2 sensitizes cells to TNF-induced apoptosis, whereas loss of c-IAP1 does not
Since changes in the expression of c-IAPs were observed, the possible contribution of these proteins in protecting cells from receptor-mediated death was also examined. MEFs from mice deficient in c-Iap1 or c-Iap2 and corresponding littermates were treated with TNF and CHX to induce apoptosis. Under these experimental conditions, c-IAP1 deficiency did not affect the amount of death observed in response to TNF and CHX treatment (Figure 2B). Interestingly, cells lacking c-IAP2 were significantly more sensitive to TNF- and CHX-induced death than their wild-type counterparts, suggesting that, unlike c-IAP1, c-IAP2 may play a role in protection from apoptosis under these experimental conditions.
The function of XIAP in inhibiting cell death is to specifically block the activation of caspases, leading us to examine caspase activation in IAP-deficient MEFs treated with TNF and CHX. After 8 h of treatment, a significant amount of cleaved caspase 3 was observed in wild-type MEFs by immunoblotting (Figure 2C). This level increased dramatically in TNF- and CHX-treated Xiap-deficient MEFs, consistent with the notion that XIAP is responsible for blocking the cleavage and activation of caspase 3. Additionally, reconstitution with wild-type murine XIAP blocked caspase 3 cleavage to a greater degree than the endogenous protein at this time point. Supporting the viability studies, no difference was observed in the amount of caspase 3 cleavage between c-Iap1-null MEFs and littermate control cells. Caspase 3 activation in the c-IAP2 MEFs also corroborated the findings from the viability studies, showing increased cleavage in treated c-Iap2-null MEFs.
Ectopic expression of human XIAP has also been shown to be protective against intrinsic or mitochondrial cell death , so the intrinsic apoptotic pathway in IAP-deficient MEFs was also examined. As shown in Figure 3(A), Xiap-null cells were more sensitive to etoposide-induced death than wild-type MEFs, and this phenotype was reversed by reconstitution with wild-type XIAP. Expression of D148A/W310A XIAP was unable to protect against this stimulus in XIAP-deficient fibroblasts. Considered together, these results indicate that murine XIAP not only modulates the threshold for mitochondrial death, but also that it is similar to human XIAP, with the caspase-binding activity likely to be the primary anti-apoptotic function of the protein. The activity of the c-IAPs in the mitochondrial death pathways was also tested with etoposide. MEFs lacking c-IAP1 were not found to be more sensitive to etoposide-induced death, but instead were found to be slightly more resistant than wild-type (Figure 3B). However, similarly to the effects observed with TNF-induced death, c-IAP2 deficiency resulted in a significant sensitivity to etoposide-induced death (Figure 3C), indicating that c-IAP2 can protect from both mitochondrial and receptor-mediated apoptosis.
Xiap- and c-Iap2-null MEFs are sensitive to etoposide-induced apoptosis
Since c-IAP2 was observed to be up-regulated in c-Iap1-null fibroblasts, we investigated whether c-IAP2 was protecting these cells from death to compensate for the loss of c-IAP1. Using RNAi (RNA interference), we were able to knockdown expression of c-IAP2 in c-Iap1-null MEFs (Figure 4, inset) and evaluate the death response to TNF plus CHX or to etoposide. As shown in Figure 4, when the amount of c-IAP2 present is reduced, c-Iap1-null MEFs were more sensitive to death induced by both stimuli. However, the sensitization induced in these cells was no greater than that observed in the c-Iap2-null cells with intact c-IAP1, which suggested that c-IAP1 did not contribute to apoptotic resistance.
Knockdown of c-IAP2 in c-Iap1-null MEFs sensitizes them to apoptosis
Previous studies have suggested that murine XIAP, in contrast with the human protein, may be dispensable for protection against apoptosis, since mice deficient in XIAP did not display any immediately obvious defects in apoptosis . The studies described in the present paper demonstrate that murine XIAP is capable of inhibiting caspase activation to modulate apoptosis. As in human cells, Xiap-deficient mouse cells are more sensitive to apoptosis, and this is likely to be the result of an increased activation of the effector caspase 3. This is supported by the results showing that mutation of the caspase-binding residues results in the same phenotype as a complete lack of protein. This was observed for apoptosis induced by both the intrinsic (etoposide) and extrinsic (TNF) pathways.
Although a lack of murine XIAP was found to result in sensitivity to apoptosis, c-IAP1 deficiency did not affect apoptosis. Interestingly, cell death was increased in the absence of c-IAP2, in parallel with a rise in caspase 3 activation in these cells. This suggested that c-IAP2 may be functioning similarly to XIAP, although it is as yet unclear what the mechanism of the inhibition of apoptosis might be. One possibility relates to previous studies which showed that a lack of c-IAPs can increase the production of and sensitivity to TNF [17–19]. This, however, does not account for the sensitivity to etoposide-induced death seen in c-IAP2-deficient MEFs. The similar sensitivity to both intrinsic and extrinsic death signals suggests that c-IAP2 may act at a point where the pathways converge, which could be at the level of caspase 3 activation. However, it has been shown previously that c-IAP2 lacks the residues required to inhibit caspase activation directly . Alternatively, c-IAP2 might affect the function of XIAP in a transcription-independent manner through the IAP inhibitor Smac/DIABLO, for which c-IAP2 is a ubiquitin ligase . Smac/DIABLO is released from the mitochondria upon apoptotic signaling and binds to XIAP to prevent its association with caspases, allowing apoptosis to proceed [4–7]. It is possible that c-IAP2 normally acts as a homoeostatic regulator of spontaneously released Smac/DIABLO, and, in the absence of c-IAP2, Smac/DIABLO is better able to neutralize XIAP, abrogating its ability to modulate caspase activation.
In summary, we find that murine XIAP deficiency renders cells more sensitive to apoptosis within a range of concentrations of apoptotic stimuli, both receptor-mediated and mitochondrial, suggesting that it functions similarly to its human homologue. The strength of the apoptotic signal appears to be vital, since modulation of cell death by endogenous XIAP is quickly overwhelmed by greater concentrations of apoptotic stimuli. This may explain the discrepancy with previous studies, and suggests that, with further exploration using more sensitive in vivo systems, the function of murine XIAP may be elucidated further. The present study suggests that the Xiap-null mice will be an important tool with which to study human disorders related to cell death deregulation.
We thank Dr John Silke (Department of Biochemistry, LaTrobe University, VIC, Australia) for his gift of cDNA encoding mouse XIAP. We are grateful to Josh Stoolman for help in the construction of plasmids for reconstitution of Xiap-null MEFs, Dr María S. Soengas (Department of Dermatology, University of Michigan) for providing E1A and Ras plasmids for the transformation of MEFs, and Dr Robert Korneluk (Department of Pediatrics, University of Ottawa, ON, Canada) for providing the anti-rIAP antibody and c-Iap2-null mice. Many thanks also to John Wilkinson, Ezra Burstein and the members of the Duckett laboratory for critical reading of this manuscript and thoughtful discussions.
early region 1A
fetal bovine serum
inhibitor of apoptosis protein
baculoviral IAP repeat
direct IAP-binding protein with low pI
mouse embryonic fibroblast
second mitochondrial-derived activator of caspase
tumour necrosis factor
X-linked inhibitor of apoptosis
This work was supported in part by the University of Michigan Biological Scholars Program, Department of Defense IDEA Award W81XWH-04-1-0891, National Institutes of Health grant GM067827 and the Sandler Foundation Award (to C. S. D.), Cancer Biology Training Grant CA09676 from the National Institutes of Health (to R. A. C.), BMRC Post-doctoral Award from the University of Michigan (to C. W. W.) and a Canadian Institute of Health Research Grant MOP37850 (to P. A. B.). C. S. D. is a consultant for Aegera Therapeutics Inc. (Montreal, QC, Canada), and P. A. B. is a founder and shareholder of Aegera Therapeutics Inc.