p34^SEI-1 Inhibits Apoptosis through the Stabilization of the X-Linked Inhibitor of Apoptosis Protein: p34^SEI-1 as a Novel Target for Anti–Breast Cancer Strategies

The recently identified p34^SEI-1-encoding SEI-1 gene positively regulates cell division by binding to cyclin-dependent kinase (CDK4; ref. 1). Previous studies have shown that the SEI-1 gene is highly expressed in carcinomas from pancreatic (2), lung (3), and ovarian tissues (4), suggesting that p34^SEI-1 functions as an oncoprotein. Interestingly, recent studies have shown that p34^SEI-1 is capable of oncogenic transformation of mouse fibroblasts (5). However, it is not clear whether p34^SEI-1 regulates apoptosis through the suppression of proapoptotic proteins or via the activation of antiapoptotic proteins.

Previous studies have established that the prosurvival protein Akt/PKB phosphorylates and subsequently stabilizes the X-linked inhibitor of apoptosis protein (XIAP; ref. 6). Other studies have shown that Notch inhibits apoptosis by directly preventing XIAP ubiquitination (7). These findings suggest that XIAP plays a key role in antiapoptotic functions and cell survival. The XIAP is one of the best-characterized members of the IAP family, which consists of at least eight proteins (8). Specifically, XIAP contains three baculovirus IAP repeat (BIR) domains and one RING-finger domain, which exerts E3 ligase activity (9). The BIR 2 domain directly binds to and inhibits active caspase-3 and caspase-7, whereas the BIR 3 domain associates with active caspase-9 (10). However, XIAP-mediated inhibition of caspase-3 and caspase-7 is abolished by Smac/DIABLO, which binds to the BIR 2 domain of XIAP (11). Previous studies have shown that XIAP is overexpressed in various breast cancer cells (12) and diverse malignant cells (13), indicating that XIAP may prove useful as a target for cancer therapies. With this in mind, various research groups have sought to identify inhibitors of XIAP and its BIR domains.

In this study, we show that p34^SEI-1 prevents the ubiquitination and degradation of XIAP through a direct association with the BIR2 domain. Our results reveal that p34^SEI-1 use s a novel apoptosis-inhibiting mechanism and that this protein represents a promising target for new antitumor therapies.

Cell culture, plasmids, and transfection. Human breast cancer cell lines as well as the human embryonic kidney line, 293, were cultured in DMEM (Life Technologies Bethesda Research Laboratories). The green fluorescent protein (GFP)-tagged p34^SEI-1 plasmid and retroviral vectors were obtained by subcloning p34^SEI-1 cDNA (i.e., provided by Dr. Rikiro Fukunaga, Graduate School of Medicine, Osaka University, Osaka, Japan) into the pEGFP-N1 (Invitrogen) and pBabe-retroviral vectors, respectively. Human cDNA encoding XIAP and 10 mutant variations were provided by Dr. C.S. Duckett (University of Michigan Medical School, Ann Arbor, MI). We also constructed three XIAP mutants using the PCR-based Quik Change site-directed mutagenesis kit (14). The XIAP^C202A, XIAP^C213A, and XIAP^K208A constructs (i.e., in which Cys-202, Cys-213, and Lys-208, respectively, were replaced with alanine). Lipofect-AMINE 2000 (Invitrogen) was used for transient transfection.

Cell death analysis. Control cells and cells expressing either SEI-1, a p34^SEI-1 small interfering RNA (siRNA) knockdown or a scrambled siRNA control were seeded at 10⁵ cells per 60-mm dish, 24 h before treatment, in the presence of various apoptotic stimuli (i.e., staurosporin, etoposide, cisplatin). The cells were then incubated for 18 h.

Immunoprecipitation and in vivo ubiquitination assays. To analyze the interaction between endogenous p34^SEI-1 and XIAP, cell lysates were prepared and immunoprecipitated with either anti–SEI-1 or anti-XIAP antibodies. Immunoprecipitation analysis was performed as previously described (7). In preparation for in vivo ubiquitination assays, cells were transfected with either GFP-tagged SEI-1 or HA-tagged XIAP. After immunoprecipitation with an anti-HA antibody, ubiquitin adducts were detected via immunoblot analysis using an anti-ubiquitin antibody.

RNA interference. Human breast cancer cells were transiently transfected with scrambled siRNA or p34^SEI-1 siRNA (i.e., 150 pmol siRNA per 60-mm dish; using p34^SEI-1 sequence 5′-CCGAAUUGGACUACCUCAUdTdT-3′). Scrambled control siRNA (i.e., 5′-GAAGCAGUCGCAGUGAAG AdTdT-3′) was obtained from Proligo LLC.

Immunofluorescence and immunohistochemistry analyses of p34^SEI-1. For colocalization of XIAP and p34^SEI-1, immunofluorescence staining was performed as previously described (15). To investigate the relationship between overexpression of p34^SEI-1 and the development of breast cancer, we conducted an immunohistochemical analysis of p34^SEI-1 expression in 60 tissue samples from patients with breast cancer. Immunohistochemical analysis was performed as previously described (16).

Immunoblot analysis. Immunoblot analysis was performed as previously described (17). Protein was probed with either anti-GFP, anti-HA, anti–caspase-7, anti-ubiquitin, anti–γ-tubulin (Santa Cruz Biotechnology), anti–cleaved caspase-9, anti–cleaved caspase-3, anti-XIAP (Cell Signaling Technology), or anti-p34^SEI-1 (AXXORA LLC) antibody.

p34^SEI-1 inhibits the apoptotic response to various stimuli. To determine if the differential induction of apoptosis by various stimuli (e.g., staurosporine, etoposide, and cisplatin) was dependent on the level of p34^SEI-1, we first assayed the expression of p34^SEI-1 in multiple lines of human breast cancer cells (Fig. 1A). The majority of the breast cancer cells displayed high levels of p34^SEI-1 expression, with the exception of the SK-BR 3 and Hs578T cell lines. So we examined whether p34^SEI-1–expressing cells were resistant to apoptotic inducer. Hs578T and MCF-7 cells were transfected with GFP-tagged p34^SEI-1 or the GFP retroviral vector. Cells expressing p34^SEI-1 or control GFP were treated with various stimuli (Fig. 1B,, i and ii). Cells expressing p34^SEI-1 were resistant to each of these apoptotic stimuli, whereas the control GFP cells were not resistant. To further explore the antiapoptotic function of p34^SEI-1, we examined the effects of p34^SEI-1 silencing using siRNAs. The population of dead cells increased in p34^SEI-1 siRNA–treated cells, relative to the population in scrambled siRNA–treated cells, after exposure to various stimuli (Fig. 1C,, i and ii), or TRAIL (Supplementary Fig. S1). To examine the mechanisms involved in the apoptosis-inhibiting effect of p34^SEI-1, we investigated whether p34^SEI-1 might affect the levels of several apoptotic related proteins. Expression of the antiapoptotic proteins cIAP-1 or cIAP-2 was not altered in p34^SEI-1-expressing 293 cells (Fig. 1D). However, XIAP expression increased in p34^SEI-1-expressing 293 cells (Fig. 1D). In addition, the population of dead cells and cleavage of procaspase-3 decreased in parallel with the up-regulation of XIAP in p34^SEI-1-siRNA–treated cells, as well as in p34^SEI-1-expressing 293 cells (Supplementary Figs. S2 and S3). Thus, the cleaved forms of caspase-3 and caspase-7 were inhibited via the up-regulation of XIAP, implying that p34^SEI-1 inhibits apoptosis by preventing the degradation of XIAP.

p34 ^SEI-1 inhibits the ubiquitination of XIAP and prevents subsequent degradation. As shown in Fig. 1, the degradation of XIAP was not induced in p34^SEI-1-expressing cells. As protein stability is controlled in part by the ubiquitin-proteasome pathway (18), we used the proteasome inhibitor, MG132, to determine if XIAP degradation required proteasome-mediated protein degradation in control cells and p34^SEI-1-expressing cells. In the absence of MG132, expression of XIAP increased significantly in cells expressing p34^SEI-1. When cells were treated with MG132, XIAP levels increased slightly upon p34^SEI-1 expression (Fig. 2A). We further explored whether XIAP turnover might be modulated by p34^SEI-1 in the presence of a protein synthesis inhibitor (cycloheximide). Expression of XIAP increased remarkably and was sustained by the expression of p34^SEI-1; however, XIAP expression gradually decreased in control cells after cycloheximide treatment (Fig. 2B,, i). We then explored XIAP levels in MCF7 cells treated with p34^SEI-1-siRNA. Expression of XIAP decreased rapidly in p34^SEI-1-siRNA–treated cells, relative to scrambled siRNA-treated cells (Fig. 2B,, ii), indicating that p34^SEI-1 stabilizes XIAP by preventing proteasome-dependent protein degradation. A previous study showed that XIAP degradation occurs in response to apoptotic stress (19). Thus, we examined whether XIAP degradation induced by apoptotic stimuli might be inhibited in a p34^SEI-1-dependent manner. Based on our findings thus far, we treated control cells and p34^SEI-1-expressing cells with etoposide. In control cells, endogenous XIAP was ubiquitinated and subjected to proteasomal degradation; however, these effects did not occur in cells expressing p34^SEI-1 (Fig. 2C). These results suggest that p34^SEI-1 significantly inhibits XIAP ubiquitination in response to apoptotic stress. Ubiquitination of XIAP was also not detected upon expression of p34^SEI-1 (Fig. 2D). A recent study reported that XIAP ubiquitination is enhanced by Smac/Diablo (14). Thus, we examined whether p34^SEI-1 might inhibit the binding of Smac/Diablo and XIAP and might thus prevent ubiquitination of XIAP by Smac/Diablo (Supplementary Fig. S4). Surprisingly, Smac/Diablo did not bind to XIAP in the presence of p34^SEI-1 and thus was unable to promote ubiquitination of XIAP by Smac/Diablo. These results suggest that p34^SEI-1 strongly inhibits XIAP ubiquitination and prevents subsequent XIAP degradation.

p34^SEI-1 binds directly to the BIR 2 domain of XIAP. We confirmed that the ubiquitination of XIAP is inhibited by p34^SEI-1 (Fig. 2); however, the mechanism by which this protein prevents ubiquitination remains unclear. Therefore, we examined the interaction between p34^SEI-1 and XIAP. We detected GFP-tagged p34^SEI-1 in HA-XIAP immunoprecipitates from 293 cells that had been cotransfected with constructs expressing these proteins (Fig. 3A). We also examined the interaction between endogenous XIAP and p34^SEI-1 in MCF-7 cells (Fig. 3B). Based on these results, we attempted to map the XIAP domains involved in direct binding to p34^SEI-1. Previous studies have shown that XIAP contains three BIR domains and one RING-finger domain, each of which has the potential to perform apoptosis-related functions (20). Hence, we used a construct expressing full-length XIAP (HA-XIAP), as well as constructs containing mutant variations of XIAP domains (HA-BIR1, HA-BIR2, HA-BIR3, and HA-RING; Fig. 3C) to determine which domains of XIAP interact with p34^SEI-1. After cotransfecting 293 cells with each construct and GFP-tagged p34^SEI-1, we observed GFP-tagged p34^SEI-1 in immunoprecipitates containing either HA-XIAP or HA-BIR2. However, p34^SEI-1-GFP did not interact with HA-BIR1, BIR3, or RING (Fig. 3D), suggesting that p34^SEI-1 interacts specifically with the BIR2 domain.

p34^SEI-1 is significantly overexpressed in human breast cancer tissues. As the overexpression of p34^SEI-1 renders breast cancer cells resistant to apoptotic stimuli, whereas repression of p34^SEI-1 has the opposite effect, we assessed p34^SEI-1 expression in formalin-fixed and paraffin-embedded samples obtained from surgical resections of 60 patients with breast cancer. As shown in Fig. 4A,, i, expression of p34^SEI-1 was faint or absent in normal breast tissue and weak in inflammatory cells. Interestingly, immunopositive staining of p34^SEI-1 was observed in breast cancer cells (Fig. 4A,, ii) and significant p34^SEI-1 overexpression was detected in 32 (53.3%) of the 60 breast cancer tissue samples (Fig. 4B). Similarly, XIAP expression was very low in normal breast tissue but strong in breast cancer cells (Fig. 4A,, iii and iv). Overexpression of XIAP was detected in 30 (50.0%) of the 60 breast cancer tissue samples (Fig. 4B). We examined the coexpression of p34^SEI-1 and XIAP in breast cancer cells and found that XIAP was expressed in 25 (78.1%) of the 32 p34^SEI-1-positive samples (Fig. 4C). In addition, the proteins were coexpressed in five of eight human breast cancer cell lines (Supplementary Fig. S6), suggesting that p34^SEI-1 and XIAP are significantly overexpressed in the tissues of patients with breast cancer.

Our findings show that p34^SEI-1 confers resistance to various apoptotic stimuli on human breast cancer cell lines and 293 cells (Supplementary Figs. S1, S2, and S3; Fig. 1). Furthermore, our data reveal that the antiapoptotic effects of p34^SEI-1 result from the up-regulation of XIAP. A recent study aligned the amino acid sequences of various proteins containing the second BIR domain (21), revealing differences in the sequences of XIAP and other proteins. Thus, we examined the binding specificity of p34^SEI-1 using XIAP mutants in which certain amino acid residues that are present only in the second BIR domain of XIAP were replaced with alanine (Supplementary Fig. S5A). Interestingly, p34^SEI-1 bound weakly to the XIAP mutants (Supplementary Fig. S5B). These results suggest that the specificity of p34^SEI-1 may be determined by specific amino acid sequences that exist only in the second BIR domain of XIAP. However, further studies are necessary to explore the details of this mechanism.

In response to apoptotic stimuli, XIAP undergoes proteasome-mediated degradation in cancer cells (19) and consequently promotes apoptosis and caspase activation. Recent studies have shown that the increased stability of XIAP enhances its antiapoptotic activity (7, 22). In the present study, we found that p34^SEI-1 inhibited the ubiquitination of XIAP, which is normally induced by apoptotic stimuli. In turn, this increased XIAP levels in p34^SEI-1-expressing cells (Fig. 2). These results suggest that p34^SEI-1 expression stabilizes XIAP and subsequently confers resistance to apoptosis. This model is supported by our demonstration of specific binding between p34^SEI-1 and XIAP (Fig. 3). Our results also show that p34^SEI-1 binds to the BIR2 domain of XIAP (Fig. 3D), but does not interact with other domains of XIAP in the absence of BIR2. However, the relationship between p34^SEI-1 and other BIR-interacting partners [e.g., Smac/Diablo (23)] remains unclear. These interactions may be influenced by different affinities of these proteins or by other unknown mechanisms. Further research is needed to determine the effects of p34^SEI-1 on XIAP antagonists. However, our findings show that p34^SEI-1 stabilizes XIAP by binding to the BIR2 domain.

The p34^SEI-1 protein rendered cells resistant to apoptotic stimuli and was highly expressed in 32 of 60 samples of human breast cancer examined in this study (Fig. 4). However, this protein is weakly expressed in normal tissues. Furthermore, a significant relationship between p34^SEI-1 and XIAP expression was observed in tissue samples from patients with breast cancer (Fig. 4C). Thus, p34^SEI-1 may contribute to tumor cell survival by increasing the stability of XIAP and consequently counteracting various apoptotic stimuli. However, as p34^SEI-1 expression was unrelated to lymph node metastasis and clinical stage (Supplementary Tables S1–5), it is likely that p34^SEI-1 participates solely in the early stages of breast cancer development. Thus, therapeutic strategies that interfere with the expression or function of p34^SEI-1 represent promising targets for anticancer drugs and are particularly attractive adjuvant treatments for breast cancer.

No potential conflicts of interest were disclosed.

Grant support: Research Center for Women's Diseases, at the Korea Science and Engineering Foundation and by a grant from Sookmyung Women's University (2007).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Colin S Duckett, of the University of Michigan Medical School, for kindly providing the XIAP plasmids; and Dr. Rikiro Fukunaga, of the Osaka University of Japan, for providing SEI-1 cDNA.