To whom correspondence should be addressed: Laboratory of Molecular Signaling and Apoptosis, Dept. of Biologic and Materials Sciences, University of Michigan, 1011 N. University Ave., Ann Arbor, MI 48109-1078. Tel.: 734-615-4386; Fax: 734-647-2110;
* This study was supported by National Institutes of Health Grants R01DE013848 and R01DE15964 (to C.-Y. W.) and T32-DE0757 (to A. M. F.). 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. The on-line version of this article (available at http://www.jbc.org) contains supplemental Figs. S1 and S2.
Cisplatin is one of the most common DNA-damaging agents used for treating patients with solid tumors such as squamous cell carcinoma (SCC). Unfortunately, significant levels of resistance in SCC cells emerge rapidly following cisplatin treatment. Here we report that the proteasome inhibitor PS-341, the representative of a new class of chemotherapeutic drugs, was capable of inducing apoptosis in cisplatin-resistant SCC cells via the endoplasmic reticulum stress. PS-341 stimulated the phosphorylation of PERK and the unfolded protein response, resulting in the induction of the transcription factor ATF-4. Importantly, the Bcl-2 homology domain 3-only (BH3-only) protein Noxa was found to be strongly induced in cisplatin-resistant SCC cells by PS-341 but not by cisplatin. The knock-down of Noxa using small interference RNA significantly abolished PS-341-mediated apoptosis in SCC cells. Using eIF2α mutant mouse embryonic fibroblasts, we found that functional eIF2α played an essential role in PS-341-induced Noxa expression. Taken together, our novel findings reveal a direct link between PS-341-induced endoplasmic reticulum stress and the mitochondria-dependent apoptotic pathway and suggest that PS-341 may be utilized for overcoming cisplatin-resistance in human SCC.
The proteasome inhibitor PS-341 (also known as Velcade or Bortezomib), represents a new class of chemotherapeutic drugs that has been demonstrated to induce apoptosis independently or synergistically with conventional cancer therapy in both solid and hematological tumors. PS-341 is a dipeptidyl boronic acid derivative that specifically inhibits the function of the 26 S proteasome (
). The 26 S proteasome is a large multisubunit ATP-dependent threonine protease complex found in the cytoplasm and nucleus of all eukaryotic cells; its principle function is to degrade proteins via the ubiquitin pathway (
). Many proteasomal substrates are known to mediate pathways that are dysregulated in human cancers. Also, the 26 S proteasome pathway degrades the misfolded or unfolded proteins and thereby maintains normal cellular functions (
). The ER stress subsequently induces a coordinated cellular response called the unfolded protein response (UPR). Although the UPR inhibits general protein translation to reduce the protein loads in the ER lumen, it also specifically up-regulates a set of genes such as molecular chaperons to alleviate the ER stress. Most studies have focused on how the UPR as well as the signaling molecules in UPR activated by the ER stress provides a protection against apoptosis (
). However, in ER stress-associated human chronic diseases, the ER stress eventually activates apoptotic signaling pathways to induce apoptosis, thereby resulting in tissue destruction. Although caspase-12 has been found to play a critical role in the ER stress-mediated apoptosis (
), the molecular signaling pathway that is associated with the ER stress-mediated apoptosis is not fully understood.
We have recently found that the proteasome inhibitor PS-341 potently induced apoptosis in head and neck SCC cell lines through the induction of ER stress in addition to the inhibition of the pro-survival nuclear factor-kappa B (NF-κB) signaling pathway (
). SCC cells comprise >90% of all malignancies diagnosed in the oral cavity and the head and neck region and is a tremendous public health challenge around the world. The 5-year survival rate for patients with head and neck cancer is one of the lowest of any major cancers and has remained un-improved over the last 20 years (
). Cisplatin has been a cornerstone chemotherapy treatment for patients with SCC. Unfortunately, significant numbers of resistant SCC cells emerge rapidly following initial cisplatin treatment. Thus, many retrospective clinical studies reported that cisplatin was unable to increase survival when administered as a monotherapy or in combination with surgery or radiotherapy (
). Cisplatin is a platinum-containing compound that induces DNA damage by binding to DNA and forming adducts that impair de novo synthesis. Mechanisms of cisplatin resistance are known to be dependent on a myriad of factors, including loss of p53, aberrant levels of cell cycle regulators such as cyclin D1, and Bcl-2 family proteins (
). Initial reductions in tumor burden are overshadowed by the ability of cisplatin-resistant cells to proliferate and emerge as recurrent or metastatic diseases for which the median survival plummets to ∼6 months (
), we were interested in investigating the potential of PS-341 to induce apoptosis in cisplatin-resistant SCC cell lines. Moreover, we wished that the studies might further help us to understand the molecular mechanism by which PS-341 induced apoptosis in SCC cells. We found that PS-341 was able to overcome cisplatin resistance and induce apoptosis in a panel of cisplatin-resistant SCC cell lines. Unlike cisplatin, PS-341 induced the markers of the ER stress, such as Gadd34, phosphorylation of the ER-resident stress-sensing kinase PERK, and a strong induction of the transcription factor ATF-4. Although the pro-apoptotic BH3-only protein Noxa was induced by both cisplatin and PS-341 in cisplatin-sensitive SCC cells, Noxa was strongly induced by PS-341, but not cisplatin, in cisplatin-resistant cells. The knock-down of Noxa by siRNA was able to significantly reduce the ability of PS-341 to induce apoptosis in cisplatin-resistant SCC cells. Taken together, our results elucidate a novel apoptotic signaling pathway that is activated by PS-341 through the ER stress and suggest that PS-341 may be utilized for treating cisplatin-resistant cancers.
Reagents and Cell Culture—PS-341, kindly provided by Millennium Pharmaceuticals, Inc., was dissolved in Me2SO to a concentration of 10 mm and stored at –20 °C; the concentration of Me2SO in culture never exceeded 1:5000. Cisplatin (cis-diammineplatinum(II) dichloride) was purchased from Sigma. UMSCC cell lines were all obtained from Dr. Thomas Carey at The University of Michigan and were cultured in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum and penicillin-streptomycin (Invitrogen). Wild-type (S/S) and mutant (A/A) eIF2-α mouse embryonic fibroblasts (MEFs) were cultured in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum and penicillin-streptomycin and supplemented with minimal essential medium amino acids solution and minimal essential medium non-essential amino acids solution (Invitrogen).
Western Blot Analysis—Cells (2 × 106) were plated in 10-cm tissue culture dishes 16–24 h before treatment with PS-341, cisplatin, or vehicle control. Cells were treated, unless otherwise indicated, with 1.0 μm PS-341 or 50 μm cisplatin. Whole cell lysates were prepared using modified radioimmune precipitation assay buffer containing phenylmethylsulfonyl fluoride and protease inhibitors (Sigma-Aldrich). 25- to 80-μg aliquots of lysates were resolved on 8% or 12% SDS-polyacrylamide gels and transferred to a polyvinylidene difluoride membrane (Immun-Blot, Bio-Rad) with a Bio-Rad semi-dry transfer apparatus. Membranes were blocked with 5% milk for 1 h at room temperature and probed with primary antibodies overnight at 4 °C. Antibodies were acquired from the following manufacturers: Puma, Phospho-PERK, Caspase-9, and Caspase-12 from Cell Signaling; Caspase-3, ATF-4 (CREB-2), Gadd34, and Gadd153/CHOP from Santa Cruz Biotechnology; Noxa from Abcam, Inc.; monoclonal α-tubulin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were from Sigma and Chemicon International, respectively. Following incubation with horseradish peroxidase-conjugated secondary antibodies, membranes were visualized by using an enhanced chemiluminescence reagent (Pierce).
Northern Blot Analysis and Real-time Reverse Transcription-PCR—Total RNAs were harvested from cells using TRIzol (Invitrogen) according to the manufacturer’s instructions. For Northern analysis, 5- to 10-μg aliquots of total RNAs were resolved on 1.5% agarose formaldehyde gels and transferred to a Zeta-Probe GT Genomic Tested Blotting (nitrocellulose) membrane (Bio-Rad) overnight. The membranes were hybridized with 32P-labeled cDNA probes and exposed to autoradiographic film. The cDNA template used to generate the Northern probe was generated using the SuperScript First-Strand Synthesis System (Invitrogen) according the manufacturer’s protocol. The human Noxa cDNA was obtained from ATCC. Northern probes were released with NotI/SalI. The human Puma expression vector pCEP4-HA-PUMA was a gift from Bert Vogelstein at the Johns Hopkins Oncology Center. Real-time RT-PCR using Cyber Green (Applied Biosystems) was carried out with an Applied Biosystems ABI 770 sequence analyzer. The real-time PCR primers for mouse Noxa were 5′-CGCCAGTGAACCCAACG-3′ (forward) and 5′-TTATGTCCGGTGCACTCCAC-3′ (reverse).
siRNA and Plasmid Transfection—The mouse ATF-4 mammalian expression vector was provided by David Ron at the New York University School of Medicine. Expression plasmid or an empty vector control (0.8 μg) was transfected into HNSCC cells for 4 h with Lipofectamine 2000 according to the manufacturer’s protocol. Trypan blue exclusion assay was performed 24 and 48 h following transfection to determine the percentage of cell death. As a transfection control, an expression plasmid containing green fluorescent protein was co-transfected into the cells (0.3 μg). For siRNA transfection, cells (3.5 × 105) were plated in 6-cm tissue culture dishes the day before transfection. siRNA (100 nm) was transfected into HNSCC cells overnight with Oligofectamine diluted in Opti-MEM (Invitrogen) according to the manufacturer’s protocol. 48 h following transfection cells were treated with 1.0 μm PS-341 as indicated. siRNA for Puma (NM_014417, targeting sequence: 5′-GGAGGGUCCUGUACAAUCUUU-3′) and Noxa (NM_021127, targeting sequence: AAACUGAACUUCCGGCAGAUU-3′) were synthesized at Dharmacon; luciferase control siRNA was also from Dharmacon.
DNA Ladder—PS-341- or cisplatin-treated cells were lysed and treated with 100 μg/ml proteinase K for 2 h at 50°C. DNA was extracted three times with phenol-chloroform mixed 1:1 and three times with chloroform alone. Precipitated DNAs were washed with 70% ethanol, and 5 μg of each sample was resolved on a 1.5% agarose gel.
PS-341 Induces Apoptosis in Cisplatin-resistant SCC Cell Lines—Recent work from our laboratory demonstrated that, unlike other common chemotherapeutic drugs, PS-341 was able to induce ER stress and apoptosis in SCC cell lines (
). Based on this finding, we hypothesized that PS-341 might be able to induce apoptosis in cisplatin-resistant SCC cell lines. To test this hypothesis, we utilized two cell lines, UMSCC-5PT and UMSCC-10BPT, both of which have been cultured to resistance with increasing concentrations of cisplatin to mimic the clinical dilemma of recurrent cisplatin-resistant SCC (
). As shown in Fig. 1A, whereas a dose of cisplatin as high as 50 μm only weakly induced cell death, both cell lines were very sensitive to PS-341-induced cytotoxicity. Additionally, we also found that UMSCC-1 cells that naturally displayed high resistance to cisplatin were sensitive to PS-341. Fig. 1 (B and C) revealed that PS-341 induced cell death in UMSCC-5PT cells in a time- and dose-dependent fashion. Moreover, Annexin-V and propidium iodide staining and fluorescence-activated cell sorting analysis also confirmed that PS-341 strongly induced cell death in cisplatin-resistant UMSCC-5PT cells (supplemental Fig. S1). As a control, we found that both cisplatin and PS-341 could induce apoptosis in cisplatin-sensitive UMSCC-23 and UMSCC-5 cells, indicating that cisplatin was functional (Fig. 1D). To determine whether PS-341 killed cisplatin-resistant cells via an apoptotic mechanism, we harvested genomic DNAs from SCC cells treated with PS-341 or cisplatin. As shown in Fig. 1E, both cisplatin and PS-341 induced DNA fragmentations in cisplatin-sensitive UMSCC-5 cells. Whereas the DNA laddering was not observed in cisplatin-treated UMSCC-5PT cells, PS-341 strongly induced DNA fragmentation in these cells. Consistently, we found that caspase-9 and -3 in cisplatinresistant SCC cells were activated by PS-341 but not by cisplatin (Fig. 1F). Taken together, these results suggest that PS-341 could induce apoptosis in cisplatin-resistant SCC cells.
PS-341 Induces the ER Stress in Cisplatin-resistant SCC Cells—Based on our previous findings that PS-341 induced the ER stress in SCC cells (
), we hypothesized that the ability of PS-341 to induce apoptosis in cisplatin-resistant SCC cell lines might also be preceded by the ER stress. As shown in Fig. 2 (A and B, top panel), the stimulation of PS-341, but not cisplatin, induced the phosphorylation of the ER membrane-resident stress-sensing kinase PERK, a hallmark of ER stress (
), in cisplatin-resistant SCC cells. It is known that activated PERK can phosphorylate eIF2α to attenuate protein synthesis and to regulate the expression of the UPR-dependent genes during the ER stress. Importantly, studies from knock-out mice have found that PERK is required for increased translation of the transcription factor ATF-4 in response to the ER stress, which plays a critical role in the induction of UPR-dependent genes (
). Consistently, Western blot analysis demonstrated that ATF-4 was induced by PS-341, but not cisplatin, in SCC cells (Fig. 2, A and B). Also, the UPR regulatory protein GADD-34 was induced by PS-341 but not cisplatin (Fig. 2, A and B). Interestingly, we found that, in contrast to MEFs, the UPR-associated CHOP was negligibly induced by PS-341 in SCC cells (data not shown). Taken together, these data suggest that PS-341, but not cisplatin, is able to induce the ER stress in cisplatin-resistant SCC cells.
PS-341 Induces Expression of the BH3-only Protein Noxa in Cisplatin-resistant SCC Cells—Having established that PS-341 was able to induce markers of ER stress in cisplatin-resistant SCC cells, we sought to identify downstream factors responsible for the induction of apoptosis. Elegant studies by Scorrano et al. (
) have demonstrated that both Bax and Bak also played an essential role in the ER stress-induced apoptosis. Indeed, we found that the deletion of Bax and Bak also provided protection against PS-341-mediated apoptosis in MEFs. However, both Bax and Bak were not induced by PS-341 in cisplatin-resistant cells (data not shown). Recently, two insightful studies reported that the ER stress-inducing agent thapsigargin could up-regulate the BH3-only Bcl-2 family member Puma in some cells by unknown mechanisms (
). The BH3-only subset of the Bcl-2 family has at least eight members, including Noxa and Puma, and is currently understood to induce apoptosis by promoting the release of cytochrome c from mitochondria (
). Therefore, we first examined whether Puma was induced by PS-341. Western blot analysis found that PS-341 weakly or negligibly induced Puma in cisplatin-sensitive UMSCC-23 cells (Fig. 3A) and cisplatin-resistant UMSCC-5PT cells (Fig. 3B) and UMSCC-10BPT cells (Fig. 3C). Interestingly, compared with PS-341, cisplatin strongly induced Puma in cisplatin-resistant UMSCC-5PT cells, but not in UMSCC-10BPT cells (Fig. 3, B and C). These results suggest that Puma might not be a critical pro-apoptotic protein induced by PS-341. Since another BH3-only protein Noxa was reported to be highly inducible upon apoptotic stimulation, we also examined whether Noxa was induced by PS-341 in SCC cells. As shown in Fig. 3A, Western blot analysis found that Noxa was significantly induced by PS-341 in cisplatin-sensitive UMSCC-23 cells. Importantly, Noxa was also strongly induced by PS-341 in cisplatin-resistant UMSCC-5PT, UMSCC-10BPT, and UMSCC-1 cells (Fig. 3, B–D). Because PS-341 might block the degradation of Noxa at the protein level, we also performed Northern blot analysis to examine whether PS-341 induced Noxa mRNA expression. As shown in Fig. 3 (E and F), Noxa mRNA was rapidly induced by PS-341 in UMSCC-23 and UMSCC-5PT cells. In summary, our results suggest that PS-341 might induce Noxa to overcome cisplatin-resistance in SCC cells.
PS-341-induced UPR Plays a Critical Role in PS-341-induced Noxa Expression—When the ER membrane-resident kinase PERK becomes phosphorylated in response to the ER stress, it phosphorylates eIF2α, which leads to a general inhibition of protein synthesis (
) have found that eIF2α phosphorylation played a critical role in the induction of UPR-dependent genes and promoted cell survival in response to ER stress. To further investigate the role of UPR in PS-341-mediated apoptosis, we utilized MEFs that express a mutant form of eIF2α (eIF2αA/A), which cannot be phosphorylated by PERK. Interestingly, we found that eIF2αA/A MEFs were significantly resistant to PS-341-mediated apoptosis compared with wild-type MEFs (eIF2αS/S) (Fig. 4A). Although caspase-12 was processed in both cells, the activation of caspase-3 was significantly delayed in eIF2αA/A MEFs (Fig. 4B). Furthermore, we observed that PS-341 was unable to induce ATF-4 or the ER stress pro-apoptotic transcription factor Gadd153/CHOP in eIF2-α mutant cells compared with wild-type cells by Western blot analysis (Fig. 4C). Because there were no commercial antibodies for mouse Noxa available, a real-time RT-PCR for Noxa was performed. As shown in Fig. 4D, our real-time RT-PCR revealed an induction of Noxa mRNA in eIF2αS/S MEFS but not eIF2αA/A MEFs in a time-dependent manner. This finding demonstrated that functional eIF2α and an intact UPR response were necessary in MEFs for PS-341 to induce accumulation of Noxa.
Noxa Plays a Critical Role in PS-341-induced Apoptosis in Cisplatin-resistant HNSCC Cells—To further determine whether Noxa played a critical role in the induction of apoptosis in cisplatin-resistant SCC cells, we first examined whether overexpression of Noxa could induce apoptosis by transient transfection in which GFP was utilized as a reporter. As shown in Fig. 5A, ectopic expression of Noxa significantly induced apoptosis in UMSCC-5PT cells 24 h after transfection of Noxa. To determine whether Noxa played an essential role in PS-341-induced apoptosis in cisplatin-resistant SCC cells, we utilized siRNA to knock down Noxa expression. Noxa siRNA, but not control siRNA for luciferase, significantly reduced PS-341-induced Noxa expression by 70–80% (Fig. 5B). As shown in Fig. 5 (C and D), the knock-down of Noxa significantly provided protection against PS-341-mediated apoptosis in both cisplatin-resistant UMSCC-5PT and UMSCC-10BPT cells. Moreover, we also examined whether the knock-down of Puma affected PS-341-induced apoptosis in UMSCC-5PT cells. As shown in Fig. 5 (E and F), the reduction of Puma expression did not inhibit PS-341-mediated apoptosis. Our results suggest that the induction of Noxa by PS-341 plays a critical role in the induction of apoptosis in cisplatin-resistant SCC cells.
In this study we demonstrated that PS-341 potently induced cell death in a panel of cisplatin-resistant SCC cell lines. Activation of caspases 9 and 3 and DNA fragmentation were observed in cisplatin-resistant SCC cells treated with PS-341, indicating the activation of the intrinsic apoptotic pathway. Consistent with our previous observation that PS-341 could induce ER stress in cisplatin-sensitive SCC cells (
), we found that PS-341, but not cisplatin, was able to induce markers of ER stress in cisplatin-resistant SCC cells. We observed phosphorylation of the ER membrane-resident stress-sensing kinase PERK and an accumulation of ATF-4 and Gadd34 following treatment with PS-341. Most importantly, we identified that Noxa played a critical role in the ER stress-mediated apoptosis induced by PS-341, providing a molecular link between the ER stress and activation of apoptosis.
Previously, it has been demonstrated that the ER stress stimulates caspase-12 activation in MEFs, which appeared to be mitochondria-independent (
). Currently, it is not clear whether there is a caspase-12-like caspase activated by PS-341 in human SCC cells. Thus, we cannot rule out that PS-341 may also activate a mitochondria-independent pathway to induce apoptosis in SCC cells. Interestingly, studies by Scorrano et al. (
) have found that Bax and Bak could be detected in ER, and there is a link between the ER stress and the mitochondrial pathway. The deletion of Bax and Bak provided protection against the ER-stress mediated apoptosis. Noxa has been found to function upstream of Bax and Bak to promote cytochrome c release from mitochondria to the cytosol, resulting in the activation of caspase-9 (
). In our studies, Western blot analysis with whole cell lysates prepared from cisplatin-sensitive SCC cells treated with PS-341 or cisplatin revealed an accumulation of both Noxa and Puma. Whereas Puma was weakly or negligibly induced by PS-341 in UMSCC-5PT and UMSCC-10BPT cells, we found that cisplatin induced Puma accumulation in UMSCC-5PT cells, although these cells are resistant to cisplatin, indicating that Puma may not be critical for PS-341 to overcome cisplatin resistance. Additionally, it is possible that specific inhibitors of Puma that may be highly expressed in SCC cells suppressed the pro-apoptotic activities of Puma. In contrast, we found that PS-341, but not cisplatin, strongly induced Noxa in cisplatin-resistant SCC cells, indicating that the ability of PS-341 to up-regulate Noxa might be an important mechanism by which it was able to overcome cisplatin resistance. Moreover, the depletion of Noxa by siRNA significantly compromised PS-341-mediated apoptosis in cisplatin-resistant SCC cells, validating that Noxa is critical for PS-341-mediated apoptosis. However, the knock-down of Noxa did not completely abolish PS-341-induced apoptosis. It was possible that PS-341 might induce apoptosis in SCC cells by inducing other BH3-only proteins. Although we did not observe that PS-341 induced other BH3-only proteins by gene expression profile (
), it is possible that PS-341 may induce the accumulation of BH3-only proteins by inhibiting proteasomal degradations in SCC cells. Several studies have reported that PS-341 could induce the level of Bik proteins to promote apoptosis (
) reported that PS-341 also induced Noxa in mantle-cell lymphoma cells independent of p53. Previously, there were several reports that PS-341 could induce apoptosis in human cancer cells in the absence of p53 (
). However, there are also some studies suggesting that the UPR may induce pro-apoptotic genes to induce apoptosis. For example, the deletion of the ER stress-responsive transcription factor CHOP in a mouse model has been found to inhibit cholesterol-induced apoptosis in macrophages via the UPR (
). Although CHOP was strongly induced by PS-341 in MEFs, it was weakly or negligibly induced by PS-341 in SCC cells. Interestingly, we found that PS-341 strongly induced ATF-4 in SCC cells, suggesting that CHOP is not responsible for PS-341-mediated apoptosis in cisplatin-resistant SCC cells. Using eIF2αA/A MEF cells that have an alanine substitution at serine 51 of eIF2α, the target of activated PERK, we found that PS-341 was unable to induce ATF-4 and CHOP/Gadd153. Importantly, we observed that PS-341 potently up-regulated Noxa mRNA in eIF2αS/S MEFs but not in eIF2αA/A MEFs, suggesting the necessity of eIF2α phosphorylation prior to Noxa induction. Moreover, we found that the knock-down of ATF-4 by siRNA also significantly reduced the up-regulation of Noxa by PS-341 (supplemental Fig. S2). Taken together, our results suggest that, although the PERK-eIF2α signaling pathway can promote cell survival, it can also potently induce apoptosis by inducing the pro-apoptotic genes in response to the ER stress.
Although dramatic initial reductions in tumor burden have been observed in patients receiving cisplatin therapy, resistant neoplastic cells very often proliferate and present as re-current or metastatic diseases, which grimly translates to survival rates of about 6 months for head and neck cancer patients (
). The ability to resist or develop resistance to cisplatin has been observed in several human cancers, including HNSCC, ovarian, and non-small cell lung cancers, which has fueled an intense search for novel adjuvant strategies. Given the fact that the chemoresistance is a significant problem in cancer therapy, our results suggest that PS-341 may offer a novel alternative for treating recurrent cancer patients. Interestingly, we found that Noxa was induced by cisplatin in cisplatin-sensitive SCC cells, but not in cisplatin-resistant SCC cells, which suggests that Noxa may play a critical role in cisplatin-mediated apoptosis. Noxa may be an important biomarker for predicting the chemosensitivity of cisplatin for patients with SCC.
We thank Dr. Thomas Carey for SCC cell lines and Lin Tao for PCR analysis.