Cancer Diagnosis and Therapy
Suberoylanilide hydroxamic acid (SAHA) overcomes multidrug resistance and induces cell death in P-glycoprotein-expressing cells
Article first published online: 8 MAR 2002
Copyright © 2002 Wiley-Liss, Inc.
International Journal of Cancer
Volume 99, Issue 2, pages 292–298, 10 May 2002
How to Cite
Ruefli, A. A., Bernhard, D., Tainton, K. M., Kofler, R., Smyth, M. J. and Johnstone, R. W. (2002), Suberoylanilide hydroxamic acid (SAHA) overcomes multidrug resistance and induces cell death in P-glycoprotein-expressing cells. Int. J. Cancer, 99: 292–298. doi: 10.1002/ijc.10327
- Issue published online: 11 APR 2002
- Article first published online: 8 MAR 2002
- Manuscript Accepted: 11 JAN 2002
- Manuscript Revised: 7 JAN 2002
- Manuscript Received: 16 NOV 2001
- National Health and Medical Research Council of Australia
- Anti-Cancer Council of Victoria
- Wellcome Trust
- Austrian Science Fund. Grant Numbers: SFB-F204, P14482
- Tyrolean Cancer Research Institute (TCRI) Tiroler Landeskrankeanstalten Aktiengesellschaft
- Tyrolean Cancer Aid Society
- histone deacetylase;
- multidrug resistance;
Multidrug resistance (MDR) mediated by the ATP-dependent efflux protein P-glycoprotein (P-gp) is a major obstacle to the successful treatment of many cancers. In addition to effluxing toxins, P-gp has been shown to protect tumor cells against caspase-dependent apoptosis mediated by Fas and tumor necrosis factor receptor (TNFR) ligation, serum starvation and ultraviolet (UV) irradiation. However, P-gp does not protect against caspase-independent cell death mediated by granzyme B or pore-forming proteins (perforin, pneumolysin and activated complement). We examined the effects of the chemotherapeutic hybrid polar compound suberoylanilide hydroxamic acid (SAHA) on P-gp-expressing MDR human tumor cell lines. In the CEM T-cell line, SAHA, a histone deacetylase inhibitor, induced equivalent death in P-gp-positive cells compared with P-gp-negative cells. Cell death was marked by the caspase-independent release of cytochrome c, reactive oxygen species (ROS) production and Bid cleavage that was not affected by P-gp expression. However, consistent with our previous findings, SAHA-induced caspase activation was inhibited in P-gp-expressing cells. These data provide evidence that P-gp inhibits caspase activation after chemotherapeutic drug treatment and demonstrates that SAHA may be of value for the treatment of P-gp-expressing MDR cancers. © 2002 Wiley-Liss, Inc.
Many chemotherapeutic drugs induce death of their target cells by activating physiologic apoptotic pathways. Two functionally separable, yet molecularly linked intracellular death pathways requiring a family of cysteine aspases (caspases) have been identified.1, 2 One pathway requires ligation and oligomerization of death receptors such as Fas and tumor necrosis factor (TNF) receptor I to initiate the activation of membrane-proximal caspases such as caspase-8 and –10, which in turn cleave and activate effector caspases such as caspase-3 and -7. The other pathway requires disruption of the mitochondrial membrane, usually by pro-apoptotic Bcl-2 proteins, the dissipation of transmembrane potential (ΔΨm) and the release of mitochondrial proteins including cytochrome c. Cytochrome c functions with Apaf-1 to induce activation of caspase-9, thereby initiating the apoptotic caspase cascade. Many commonly used chemotherapeutic agents such as doxorubicin, vincristine and cisplatin can induce death of their tumor targets by initiating the mitochondrial apoptotic pathway.3, 4
Multidrug resistance (MDR) is a major obstacle to the successful treatment of many human cancers. A hallmark of MDR is the expression of the ABC transporter molecule P-glycoprotein (P-gp). Typically, P-gp functions to confer MDR by actively effluxing chemotoxins from cells, thereby preventing intracellular accumulation and subsequent apoptotic cell death.5 In addition, we and others have recently shown that P-gp can inhibit the activation of caspases and prevent apoptosis induced by diverse stimuli such as serum starvation, ligation of death receptor molecules Fas and TNFR and UV irradiation.6–9 However, our initial studies revealed that P-gp could not protect against the cytotoxic granule proteins perforin and granzyme B, which elicit cell death in a caspase-independent manner.7 Given that combination therapies using P-gp antagonists and conventional chemotherapeutics have been unsuccessful,10 we speculated that drugs that could induce caspase independent-cell death might be effective in eliminating MDR tumors. We therefore sought to identify and characterize agents that may induce cell death in this manner.
The hybrid polar compound (HPC) hexamethylene bisacetamide (HMBA) can induce cellular differentiation,11 and we have demonstrated that HMBA can cause equivalent caspase-independent cell death in P-gp-expressing (P-gp-positive) and non-P-gp-expressing (P-gp-negative) human leukemia and colon carcinoma cells.9 Apoptosis was accompanied by caspase-independent release of cytochrome c and downregulation of Bcl-2. From these findings we speculated that HPCs might represent a class of chemotherapeutic agent that may be effective in eliminating P-gp-expressing MDR tumors. Although HMBA has been used clinically to treat acute myelogenous leukemia (AML), successful treatment has been limited due to the relatively high doses required (5 mM and above) and the subsequent toxicity and adverse side effects.12 However, second-generation HPCs including suberoylaninide hydroxamic acid (SAHA) and m-carboxycinnamic bishydroxamide (CBHA), which are also capable of inducing differentiation, have now been developed.13
As a differentiation agent, SAHA is 2,000 times more potent than HMBA, but, unlike HMBA, SAHA is also a strong inhibitor of histone deacetylase (HDAC) activity.13 Acetylation of histones is a key process in activating transcription14 and SAHA has been reported to selectively induce expression of the p21WAF1/CIP1 cyclin-dependent kinase inhibitor to effect cell cycle arrest.15 Furthermore, SAHA has been shown to induce p53-independent apoptosis of U937 human leukemia cells.16 We have recently identified important molecular events necessary for SAHA-induced cell death. SAHA mediated the cleavage and activation of the pro-apoptotic Bcl-2 family member Bid, resulting in the release of cytochrome c and ROS from the mitochondria into the cytosol.17 Although caspase activation was induced by SAHA, caspase activity was not necessary for SAHA-mediated cell death or cleavage of Bid. Importantly, Bcl-2 could completely inhibit mitochondrial damage, ROS production and cytochrome c release and subsequently inhibit SAHA-mediated cytotoxicity.17 These studies demonstrated the importance of the mitochondria in mediating SAHA-induced cell death. In contrast, CBHA, which is also a HDAC-inhibitor, has been shown to induce death of neuroblastoma cells that was attenuated by addition of the caspase inhibitor ZVAD-fmk in short-term assays.18 Addition of CBHA resulted in increased expression of Fas and Fas ligand and Glick and colleagues18 speculate that the death receptor pathway may be involved in CBHA-induced cell death.
In animal models, SAHA has low toxicity, can inhibit tumor growth in vivo19 and is currently in human phase I clinical trials. Given the molecular characteristics of SAHA-mediated cell death and the therapeutic potential of this drug, we investigated whether SAHA could induce apoptosis in P-gp-expressing cell lines. We have analyzed the molecular death pathways used by SAHA in P-gp-expressing and -non-expressing tumor cell lines. SAHA induced equivalent death of P-gp-positive and P-gp-negative CEM, LoVo and K562 cells. Although caspase activation was greatly inhibited in P-gp-positive cells, the molecular events associated with SAHA-induced cell death, cleavage of Bid, permeabilization of the mitochondrial membrane and increased ROS production were not affected. These data indicate that P-gp can inhibit drug-induced caspase activation but may not affect other death pathways mediated by certain chemotherapeutic drugs such as SAHA. Although SAHA induced death in all cell types studied, the amount of K562 death was significantly lower than in the other cell lines and these cells preferentially underwent a cell cycle arrest in the G1/G0 phase. This was in contrast to CEM cells, which did not arrest in G1/G0 and rapidly underwent apoptosis. Thus, SAHA can induce equivalent cell death and/or cell cycle perturbations in P-gp-positive and P-gp-negative cells. Inhibition of tumor cell growth, either by inhibiting proliferation and/or inducing cell death, are desirable therapeutic properties for chemotherapeutic drugs. We therefore propose that SAHA may be effective for the treatment of P-gp-expressing MDR tumors.
MATERIAL AND METHODS
The acute T-cell leukemia cell line, CEM-CCRF and its doxorubicin-selected and P-gp-positive derivative CEM-P-gp have been described previously.20 Colon carcinoma LoVo cells and LoVo cells resistant to adriamycin (LoVo-P-gp), and K562 and Re vincristine selected K562-P-gp cells have been previously described.9 CEM cells overexpressing human Bcl-2 were a gift from David Huang (Walter and Eliza Hall Institute, Melbourne, Victoria, Australia). All cells were grown in RPMI-1640 supplemented with 10% (vol/vol) FCS, 2 mm/L glutamine, 100 units/mL penicillin and 100 μg/mL streptomycin (GIBCO, Grand Island, NY). Cells were cultured for 4–24 hr with 0–10 μM SAHA, 100–500 ng/mL doxorubicin, 100–500 ng/mL vincristine or 10–100 ng/mL anti-Fas antibody clone CH-11 (Upstate Biotechnology, Lake Placid, NY). Vincristine was obtained from Phillip Kantharidis, (Peter MacCallum Cancer Institute, East Melbourne, Victoria, Australia). SAHA was kindly provided by Dr. Victoria Richon (Sloan Kettering Cancer Center, New York, NY). To inhibit the activation of caspases, cells were pretreated for 60 min with peptidyl fluoromethylketones (ZFA-fmk or ZVAD-fmk; final 0–40 μmol/L; Enzyme Systems Products, Dublin, CA).
Cytotoxicity and viability assays
Cell death was assessed by 51Cr release assay as previously described.21 The spontaneous release of 51Cr was determined by incubating the target cells with medium alone. The maximum release was determined by adding SDS to a final concentration of 5%. The percent specific lysis was calculated as follows: 100 × [(experimental release − spontaneous release)/(maximum release − spontaneous release)]. Cells were cultured at 2 × 105 cells/mL in the presence or absence of cell death stimuli for varying times. Trypan blue exclusion assays were performed as previously described.9 In all assays, 150–300 cells were counted for each data point and data were calculated as the mean ± SE of triplicate samples and are representative of at least 3 separate assays. The number of apoptotic or dead cells (blue cells) was expressed as a percentage of the total cell number. Cells (2 × 105) treated with various apoptotic stimuli were assessed for DNA content by propidium iodide (PI) as previously described.9
Clonogenic assays and propidium iodide staining
Western blot analysis
Cells (2 × 105) were lysed in 50 μL of ice-cold Nonidet P-40 lysis buffer as previously described.9 Protein determinations were performed using Bradford reaction. Proteins (10–20 μg) were separated on 10, 12 or 15% SDS polyacrylamide gels and electroblotted onto nylon membranes. Blots were probed with anti-human caspase-3 mAb (Transduction Laboratories, Lexington, KY), anti-human PARP (Boehringer Mannheim, Indianapolis, IN), anti-human α-tubulin monoclonal antibody (MAb; Sigma, St. Louis, MO) and anti-human Bid MAb (Xiaodong Wang, Howard Hughes Medical Institute and Department of Biochemistry, University of Texas Southwestern Medical Center, Dallas, TX) and visualized by enhanced chemiluminescence (Amersham, Buckinghampshire, UK).
Cells (1 × 106) were treated for 24 hr in the presence of 0 or 2.5 μM SAHA or 100 ng/mL vincristine. Cells were washed twice in ice-cold PBS and lysed in Nonidet P-40 lysis buffer without protease inhibitors. Lysate (50 μg) was incubated with Ac-Asp-Glu-Val-Asp-pNA (Ac-DEVD-pNA) substrate (1.25 mM final; Bachem, Bubendorf, Switzerland) in BAADT buffer, pH 7.3, containing 0.1 M HEPES and 0.05 M CaCl2 at 37°C and optical density (OD; 750 λ/405 absorbance) was read every 30 min. Relative fold activity was determined by dividing OD readings of samples containing drug (SAHA or vincristine) by OD of control untreated samples.
Cytochrome c release
Cytosolic extracts from cells (2 × 105) treated for 24 hr with 2.5 μM SAHA were prepared as previously described.9 Cytosolic proteins (250 μg) were immunoprecipiated with cytochrome c MAb clone 6H.2 (0.5 μg/mL final; PharMingen, San Diego, CA) at 4°C for 1 hr and then overnight with protein A sepharose beads (30 μL of 50% w/v slurry). Beads were washed 5× in buffer containing 30 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X-100, 10% glycerol, 0.1 mM AEBSF. Beads were resuspended in denaturing sample buffer, separated on a SDS-15% polyacrylamide gel and transferred onto nylon membranes. Blots were probed with anti-human cytochrome c MAb clone 7.H.2 (PharMingen).
Cells (2 × 105) were washed twice in 37°C 5 mM HEPES-buffered saline (pH 7.4). Cells were then resuspended in 37°C 5 mM HEPES-buffered saline alone or containing 10 ng/mL H2DCFDA dye (C-400, Molecular Probes, Eugene, OR) and incubated at 37°C for 15 min. Cells were then washed with ice-cold HEPES saline and placed on ice. Fluorescence of oxidized C-400 was measured using a FACscalibur FACScan at with an excitation wavelength of 480 nm and an emission wavelength of 525 nm.
Preparation of histones and analysis of histone acetylation
Histones were prepared from 2 × 107 cells. Acidic extraction of core histones was performed from whole cells with 0.3 M HCl after prior extraction of linker histones and HMG proteins with 5% perchloric acid. For the extraction procedures with 5% perchloric acid as well as 0.3 M HCl, the pellet was taken up in 20 mL for a first extraction step and the extraction was repeated with 10 mL of the respective acids. In each extraction step the pellets were subjected to homogenization in a Dounce homogenizer by 10 up-and-down strokes. After incubation for 30 min on ice, the homogenate was centrifuged for 15 min at 20,000g, 4°C. The supernatants of the 2 extractions with 0.3 M HCl were united and the core histones precipitated with 25% trichloric acid (final concentration). After incubation on ice for at least 1 hr, the precipitate of the core histones was collected by centrifugation at 20,000g, for 20 min at 4°C. The pellet was washed with HCl-acetone and dried in vacuo. Electrophoretic separation of the acetylated forms of histone H4 was performed in acidic urea-Triton X-100 polyacrylamide gels (12% T, 2.6% C, 8 M urea).22
SAHA induces equivalent cell death in P-gp-positive and P-gp-negative tumor cell lines
To investigate whether SAHA could induce cell death in a number of different paired P-gp-expressing cell lines, P-gp-positive and P-gp-negative CEM (T-cell leukemia), LoVo (colon carcinoma) and K562 (erythroblastoma) cells were treated for 24 hr with 0–10 μM SAHA or the chemotherapeutic agent vincristine as a control for P-gp function. SAHA induced equivalent cell death in a dose-dependent manner in both P-gp-positive and P-gp-negative CEM cells in a 24 hr short-term 51Cr release assay (Fig. 1a). To confirm that our 51Cr release assays accurately measured cell survival, we employed a sensitive clonogenic assay demonstrating that both P-gp-positive and P-gp-negative CEM cells were equally sensitive to SAHA, whereas only P-gp-negative cells were sensitive to vincristine (Fig. 1b).
Treatment of P-gp-positive and P-gp-negative LoVo cell lines with SAHA also resulted in a dose-dependent loss of cell viability (Fig. 2a). Only the P-gp-negative Lovo cells were sensitive to vincristine-induced death, demonstrating the functionality of P-gp in these cells (Fig. 2a). In contrast to CEM and LoVo cells, K562 and K562-P-gp cells were relatively insensitive to SAHA-induced cell death (Fig. 2b), although the effect was similar for both P-gp-positive and P-gp-negative K562 cells. These cells were not inherently resistant to all apoptotic stimuli, as K562 cells were sensitive to treatment with vincristine (Fig. 2b).
Analysis of the cell cycle profiles of CEM and K562 cells treated with SAHA revealed that CEM cells accumulate in the G2/M and sub-G1 (apoptotic) fractions, with approximately 32% of CEM and 38% of CEM-P-gp cells undergoing apoptosis at 24 hr (Fig. 3). Vincristine induced a similar accumulation of cells within the sub-G1 and G2/M fractions in CEM cells, whereas CEM-P-gp cells remained relatively unaffected. By contrast, K562 and K562-P-gp cells accumulate in G1/G0 and have only a slight increase in apoptotic cells compared with background levels. Clearly not all cell death stimuli induce a G1/G0 arrest in K562 cells, as vincristine induced accumulation in the sub-G1 and G2/M fractions. These results confirm the data shown in Figure 1 and suggest that SAHA can induce cell death or cell cycle arrest within G1/G0 depending on the cell type.
SAHA-induced caspase 3 activation is inhibited in P-gp-positive CEM cells
Previous studies have shown that P-gp inhibits activation of caspases during Fas-mediated apoptosis and thereby prevents cell death.7 Since we and others have demonstrated that SAHA can induce the activation of caspases3, 16, 17 we assessed whether caspase-3 activation was altered in P-gp expressing cells. P-gp-positive and P-gp-negative CEM cells were treated for 24 hr with 2.5 μM SAHA or 100 ng/mL vincristine and caspase-3 activity was determined by Western blot (Fig. 4) and DEVDase activity assays (Fig. 5). As shown in Figure 4, P-gp-negative CEM cells treated with either SAHA or vincristine displayed cleavage of the DNA repair enzyme and caspase-3 substrate, poly ADP-ribose polymerase (PARP; 48.8% cleavage with SAHA; 34.8% cleavage with Vin), as well as caspase-3 cleavage (Fig. 4, lanes 1–3). In contrast, CEM-P-gp cells treated with SAHA displayed reduced PARP cleavage (12.6% cleavage) and no detectable caspase-3 cleavage (Fig. 4, lanes 4–6). Neither PARP nor caspase-3 was cleaved in CEM-P-gp cells treated with vincristine.
DEVDase activity assays performed on lysates from cells treated for 24 hr with 2.5 μM SAHA (Fig. 5a) or 100 ng/ml vincristine (Fig. 5b) confirmed that the rate of caspase-3 activation was significantly reduced in P-gp-expressing CEM cells. The low-level caspase-3 activity seen in SAHA-treated CEM-P-gp cells (Fig. 5a) correlated with the minimal cleavage in PARP seen in Figure 4 (lane 5). By contrast, caspase activity in vincristine-treated CEM-P-gp cells was completely abolished, which also correlates with the PARP cleavage profile seen in Figure 4. The enhanced effect of P-gp on vincristine-induced PARP and caspase cleavage may be due to the ability of P-gp to efflux vincristine out of the CEM cells and also to inhibit caspase activation. To demonstrate that SAHA-induced caspase activation could be inhibited by caspase inhibitors, CEM and CEM-P-gp cells were pretreated with the poly-caspase inhibitor ZVAD-fmk and then cultured with SAHA for 24 hr (Fig. 5c). As expected, ZVAD-fmk completely inhibited all caspase-3 activity in CEM and CEM-P-gp cells, demonstrating that ZVAD-fmk is an effective inhibitor of caspase activation during SAHA-mediated cell death. Furthermore, a clear difference in the rate of caspase-3 acitivity in SAHA-treated CEM and CEM-P-gp cells was observed (Fig. 5c). Therefore, these studies provide evidence that in addition to inhibiting caspase activation mediated by death receptor ligation, P-gp can also inhibit drug-induced caspase-3 activity.
The data presented above indicate that caspase-3 activity is not necessary for SAHA-induced cell death. To determine whether other caspases may be required for SAHA-induced death of CEM or CEM-P-gp cells, we pretreated the cells with the poly-caspase inhibitor ZVAD-fmk or the control cathepsin inhibitor ZFA-fmk, before incubating the cells with SAHA and performing clonogenic assays. As shown in Figure 6, inhibition of caspase activity did not affect SAHA-induced cell death in either cell line. As expected, CEM (Fig. 6a) but not CEM-P-gp cells (Fig. 6b) were sensitive to apoptosis induced by vincristine. Importantly, ZVAD-fmk significantly affected vincristine-induced cell death (Fig. 6a), demonstrating the effectiveness of this inhibitor in our system. Using an extremely sensitive cell death assay, these data demonstrate that inhibition of caspase activation, either by addition of small peptide inhibitors or by overexpression of P-gp, does not affect cell death induced by SAHA.
SAHA inhibits histone deacetylation activity in both CEM and CEM-P-gp cells
SAHA has previously been reported to be a potent inducer of differentiation and an effective HDAC inhibitor.23 It has been proposed that SAHA induces differentiation by increasing transcription, specifically of the p21waf1/cip1 gene, by specifically inhibiting histone deacetylation at the p21waf1/cip1 locus.15 We have demonstrated that SAHA-induced cell death is dependent on new gene transcription and translation,17 indicating that HDAC inhibition is also necessary for the death-inducing activity of SAHA. We therefore endeavored to determine whether SAHA could equivalently inhibit histone deacetylation in CEM and CEM-P-gp cell lines. Histones 2B and 4 (H2B and H4) from P-gp-positive and P-gp-negative CEM cells treated with SAHA displayed a much higher degree of acetylation compared with untreated controls (data not shown). These results indicate that SAHA functions as a potent HDAC inhibitor independently of P-gp status.
SAHA-mediated cell death marked by cytochrome c release, ROS production and BID cleavage is not affected by P-gp
We and others have demonstrated that mitochondrial membrane disruption plays an essential role in caspase-independent cell death.9, 24–26 We have determined that cell death induced by SAHA is mediated by the cleavage and activation of the pro-apoptotic Bcl-2 family member Bid, mitochondrial membrane perturbation and production of ROS.17 We therefore assessed SAHA-induced death of CEM and CEM-P-gp cells to determine whether P-gp could inhibit any of these events.
SAHA induces caspase-independent Bid cleavage upstream of mitochondrial perturbation, resulting in release of cytochrome c and ROS into the cytosol.17 We therefore assayed for the cleavage of SAHA-induced Bid in P-gp-expressing CEM cells. P-gp-positive and P-gp-negative CEM cells were cultured for 24 hr with 2.5 μM SAHA or 100 ng/mL vincristine in the presence or absence of ZVAD-fmk or ZFA-fmk and Bid cleavage was assessed by Western blot (Fig. 7a). Bid was cleaved in CEM cells treated with both SAHA and vincristine and in CEM-P-gp cells treated with SAHA (Fig. 7a). SAHA-induced Bid cleavage was not inhibited by ZVAD-fmk; however, ZVAD-fmk did completely abolish vincristine-mediated cleavage of Bid. These results taken together with the caspase 3 activity assays (Figs. 4, 5) provide strong evidence that SAHA-induced Bid processing is mediated by a non-caspase protease that is not affected by P-gp overexpression.
To assess the release of cytochrome c from the mitochondria into the cytosol, we performed immunoprecipitation/Western blots on cytosolic extracts from CEM and CEM-P-gp cells treated for 24 hr with 2.5 μM SAHA. Cytochrome c was present in cytosolic lysates from CEM and CEM-P-gp cells cultured with 2.5 μM SAHA for 24 hr but was not detected in untreated control cells (Fig. 7b). In addition, cytochrome c release occurred independently of caspase activation since it was not blocked by ZVAD-fmk (Fig. 7b).
The production of ROS has been implicated in mitochondrial membrane depolarization and cell death,27 and inhibition of ROS production can significantly affect SAHA-induced apoptosis.17 SAHA induced a dose-dependent increase in the production of ROS within 6 hr (data not shown) in both CEM and CEM-P-gp cells, with a maximum production at 24 hr (Fig. 8). Interestingly, treatment of cells with vincristine did not increase ROS production. SAHA-induced ROS production in CEM and CEM-P-gp cells was not affected by ZVAD-fmk (data not shown). Taken together, these data demonstrate that irrespective of P-gp expression or caspase activity, SAHA can induce the cleavage and activation of Bid, perturbation of the mitochondrial membrane and release of ROS resulting in target cell death.
In addition to its drug efflux activity, P-gp can inhibit caspase-3 activation and apoptosis induced by ligation of Fas and TNF death receptors.7, 8 It is now evident that chemotherapeutic drugs of diverse structure and specificity can induce the death of target cells by activating physiologic apoptotic pathways and that drugs such as doxorubicin and vincristine require caspase activation for full cytotoxic function.8 We therefore hypothesized that drugs that are not P-gp substrates and function in the absence of caspase activity would be capable of inducing death of P-gp-positive MDR tumor cells. The HDAC inhibitor SAHA activates a novel cell death pathway characterized by the cleavage and activation of the BH3-only Bcl-2 family member Bid, mitochondrial membrane perturbation and release of mitochondrial cytochrome c and ROS into the cytosol.17 In our study we demonstrate that SAHA can act equivalently on P-gp-positive and P-gp-negative cells of lymphoid, myeloid and epithelial origin to induce cell death or growth arrest and that P-gp had no effect on the ability of SAHA to act as a histone deacetylase inhibitor. Consistent with our previous data using Fas ligation as the death stimulus, SAHA-induced caspase-3 activation was inhibited in P-gp-expressing cells. These results, taken together with our previous reports,7–9 provide definitive evidence that P-gp inhibits caspase-3 activation mediated either via ligation of death receptors or perturbation of the mitochondria.
We have recently shown that mitochondrial perturbation is essential to SAHA-mediated cell death.17 Therefore, we examined the effects of P-gp on key events that occur upstream and downstream of SAHA-induced mitochondrial membrane damage and subsequent cell death. P-gp had no effect on the cleavage and activation of Bid, the release of cytochrome c or the production of ROS. Therefore, although P-gp could inhibit SAHA-induced caspase activation, it has no effect on key upstream events that occur before (Bid cleavage) or after (cytochrome c release and ROS production) mitochondrial membrane depolarization. Thus, chemotherapeutic agents that can induce caspase-independent cell death mediated by mitochondrial perturbation may be more effective in treating MDR tumors than conventional caspase-dependent chemotherapeutics.
The K562 erythroblastoma cell line was significantly less sensitive to death induced by SAHA and more prone to cell cycle arrest than CEM T cells or LoVo colon carcinoma cells. In response to treatment with SAHA for 24 hr, a small proportion of K562 cells underwent apoptosis but most accumulated in the G1/G0 phase of the cell cycle. No difference in the degree or kinetics of histone acetylation was observed between CEM and K562 cells treated with SAHA (Ruefli and Johnstone, unpublished observations). We demonstrated that other chemotherapeutic drugs such as vincristine are equivalently cytotoxic in K562, CEM and LoVo cells and it is presently unclear why SAHA preferentially induces cell cycle arrest rather than death in these cells. Others have demonstrated that K562 cells are relatively insensitive to the cytotoxic action of other HDAC inhibitors such as sodium butyrate and trichostatin A compared with cytokine-dependent cell lines treated with the same drugs.28 In that case, suppression of cytokine signaling appeared to be the primary cytotoxic role of these drugs, which did not affect cytokine-independent K562 cells. It has also been demonstrated that upregulation of the cyclin-dependent kinase inhibitor p21WAF1/CIP1 could reduce the cytotoxic action of the HDAC inhibitor azelaic bishydroxamic acid (ABHA), possibly by inhibiting caspase-3 activation.29 However, we find that SAHA induces p21WAF1/CIP1 upregulation in both CEM and K562 cells and that phosphorylation of the tumor suppressor protein and indirect target of p21WAF1/CIP1 pRb is similarly suppressed in both cell types (Ruefli and Johnstone, unpublished observations).
HDAC inhibitors represent a novel class of chemotherapeutic agents that induce differentiation and cell death in many transformed cell types and may lead to more successful treatment of cancers. It has been clearly demonstrated that SAHA and other hydroxamic acid-based HPCs increase accumulation of core histones by inhibition of HDAC activity.18, 23, 30 In the present study, accumulation of acetylated histones was demonstrated in both CEM and CEM-P-gp cells, indicating that P-gp has no effect on the ability of SAHA to regulate transcription. However, it is evident that not all HDAC inhibitors function in the same manner. Indeed, the SAHA-related compound CBHA has been demonstrated to induce expression of cell surface Fas and possibly mediate cell death via activation of the caspase-dependent death receptor pathway.18 In contrast, we do not observe increased expression of death receptors or their ligands after SAHA treatment (data not shown) and, as presented herein, caspase activity is not necessary for SAHA function.
Our data demonstrate that P-gp can inhibit the death effector pathways required by some drugs (i.e., doxorubicin, vincristine) but not by others (i.e., SAHA). SAHA activates a novel apoptotic pathway not reliant on caspase activation but dependent on new gene transcription to induce Bid cleavage and mitochondrial membrane disruption. It now remains to identify the cellular genes targeted for expression by SAHA and other HDAC inhibitors and to characterize the molecular mechanisms by which the protein products of these genes induce apoptosis and/or differentiation. Our results strongly suggest that certain HPCs such as SAHA may hold potential as therapies for P-gp-expressing MDR tumors.
We thank Dr. J. Trapani and Dr. S. Russell for critically reviewing our manuscript and Dr. X. Wang and Dr. V. Richon for reagents. R.W.J. is a Wellcome Trust Senior Research Fellow and M.J.S. is a Principal Research Fellow of the National Health and Medical Research Council of Australia. D.B. was supported by grant SFB-F204 from the Austrian Science Foundation and by the TCRI, which is supported by the Tiroler Landeskrankeanstalten Aktiengesellschaft (TILAK), the Tyrolean Cancer Aid Society, various businesses, financial institutions and the people of Tyrol.