R.N.T. Coffey is an AstraZeneca Newman Scholar in Urology.
Resistance to caspase-dependent, hypoxia-induced apoptosis is not hypoxia-inducible factor-1 alpha mediated in prostate carcinoma cells
Article first published online: 17 FEB 2005
Copyright © 2005 American Cancer Society
Volume 103, Issue 7, pages 1363–1374, 1 April 2005
How to Cite
Coffey, R. N. T., Morrissey, C., Taylor, C. T., Fitzpatrick, J. M. and Watson, R. W. G. (2005), Resistance to caspase-dependent, hypoxia-induced apoptosis is not hypoxia-inducible factor-1 alpha mediated in prostate carcinoma cells. Cancer, 103: 1363–1374. doi: 10.1002/cncr.20918
- Issue published online: 18 MAR 2005
- Article first published online: 17 FEB 2005
- Manuscript Accepted: 12 NOV 2004
- Manuscript Revised: 28 OCT 2004
- Manuscript Received: 15 JUN 2004
- Irish Cancer Society
- Mater College grant
- Science Foundation Ireland
- Wellcome Trust
- prostate carcinoma;
- androgen independence
Hypoxia occurs in association with cancer development, the result being a more aggressive and metastatic cancer phenotype. Hypoxia, which activates hypoxia-inducible factor-1 alpha (HIF-1α), is associated with a number of cellular changes including increased apoptotic resistance. The authors hypothesized that HIF-1α is central to the cell's ability to resist apoptosis induced during the hypoxia selection process.
PWR-1E, LNCaP, LNCaP-HOF, PC-3, and DU-145 cells were cultured in normoxic and hypoxic conditions. Apoptosis was assessed by propidium iodide DNA staining. Cleavage of specific substrates was used to assess caspase activity and Western blotting was used to assess mitochondrial release of cytochrome c and second mitochondria-derived activator caspase (SMAC)/Diablo. A dominant negative HIF-1α construct was transfected into the PC-3 and LNCaP cells to block HIF-1α activity.
PC-3 and DU-145 were resistant to apoptosis induced by exposure to hypoxia, but the PWR-1E and LNCaP cells were susceptible. This induction of apoptosis in the LNCaP cells was caspase dependent but independent of cytochrome c release. Blocking the activity of HIF-1α had no effect on increased apoptotic susceptibility in the PC-3 cells. LNCaP-HOF cells, which were resistant to hypoxia-induced apoptosis, showed no increase in HIF-1α expression or activity.
Apoptotic resistance is already established in cells that survive a hypoxic insult and whereas increased HIF-1α activity may be essential for the development of a more aggressive cancer phenotype, it may not be responsible for the initial selection of an apoptotic resistance phenotype. Cancer 2005. © 2005 American Cancer Society.
Androgen withdrawal is associated with the induction of apoptosis of prostate epithelial cells.1 Advanced prostate carcinoma is resistant to this effect, leading to the selection of an androgen-independent cancer. Many features are associated with this resistance—from hypersensitivity of the androgen receptor to the increased expression of antiapoptotic proteins.2 How these cells change into this androgen-independent phenotype is largely unknown, but genetic and environmental factors have been demonstrated.3
Increasing evidence suggests that androgen withdrawal is associated with increased hypoxia in the prostate gland, which drives its regression. This is modeled in the rat ventral prostate where glandular regression after castration is associated with an ischemic/hypoxic environment caused by a reduction in blood flow.4 Tissue hypoxia is also a characteristic feature of solid tumors and is the net result of a tumor outgrowing its original blood supply. As hypoxia is an integral feature of the formation of any solid tumor, an ability to adapt to a low oxygen supply and increase the apoptotic threshold is likely to be one of the critical factors in determining whether cancer cells survive in their tumor microenvironment.5 In relation to prostate carcinoma, hypoxia has been demonstrated to occur by direct measurements of PO2 levels using Eppendorf microelectrodes (Eppendorf AG, Hamburg, Germany) in radical prostatectomy samples.6 Immunohistochemical studies have also investigated biomarkers indicative of hypoxia.7 These biomarkers include vascular endothelial growth factor (VEGF),8 the alpha subunit of hypoxia-inducible factor-1 (HIF-1α), and many of the genes involved in the glycolytic pathway that promote metabolic adaptation to reduced oxygen pressure.9
Increased tumor tissue hypoxia is associated with a poor therapeutic response.10, 11 Evidence from in vitro studies demonstrates that tumor cells cultured under hypoxic conditions are resistant to a number of apoptotic-inducing agents.12–14 This has been associated with a stabilization of the mitochondrial membrane via alterations in the Bcl-2 family of proapoptotic and antiapoptotic proteins.
Evidence also suggests that hypoxia is capable of inducing apoptosis in a caspase-dependent manner through a mitochondrial signaling event that involves the release of cytochrome c and subsequent activation of caspase 9 through the formation of the apoptosome.15, 16 The reason for this differential response between cells to hypoxia is unknown.
The current study addresses the effects of hypoxia on the apoptotic signaling of the PWR-1E, LNCaP, PC-3, and DU-145 prostate carcinoma cell lines. It also determines a role for HIF-1α in prostate carcinoma cell resistance to hypoxia-induced cell death. As the development and progression of prostate carcinoma are associated with apoptotic resistance, the mechanism leading to inhibition of hypoxia-induced apoptosis has important therapeutic implications.
MATERIALS AND METHODS
The human prostate carcinoma cell lines PWR-1E, LNCaP, PC-3, and DU-145 were purchased from the American Tissue Type Culture Collection (Rockville, MD). LNCaP-HOF cells were a gift from Helmut Klocker (Department of Urology, University of Innsbruck, Innsbruck, Austria). Cells were cultured in RPMI-1640 medium supplemented with 0.5% glucose, 10% heat-inactivated fetal bovine serum, 50 U/mL penicillin, 50 μg/mL streptomycin, and 2 mM L-glutamine (GIBCO Life Technologies, Gaithersburg, MD). These cell lines were grown routinely and passaged in 75-cm3 vented tissue culture flasks at 37 °C in a humidified atmosphere of 5% CO2. Hypoxia was achieved by culturing cells either on tissue culture 6-well plates or in 75-cm3 vented tissue culture flasks at 37 °C in hypoxic chambers (Coy Instruments, Ann Arbor, MI) at 1 ± 0.5% O2 over a 72-hour period.
Detection of Apoptotic Events by Flow Cytometry
Apoptosis was quantified as the percent events with hypodiploid DNA as assessed by propidium iodide (PI) incorporation as previously described.17 Briefly, cells (1 × 106) were centrifuged at 200 ×g for 10 minutes, and resuspended in 500 μL of hypotonic fluorochrome solution (50 μg/mL PI, 3.4 mM sodium citrate, 1 mM Tris, 0.1 mM ethylenediaminetetraacetic acid [EDTA], 0.1% Triton X-100) and incubated on ice for 15–30 minutes before analysis. A minimum of 5000 events were collected and analyzed. Decreased PI intensity represented increased double-stranded DNA cleavage. PI viability assays also were performed to distinguish between intact cellular membranes and disrupted membranes of apoptotic and necrotic cells, respectively. All measurements were performed with the same instrument settings on an Epics XL-MCL Coulter Elite flow cytometer (Becton Dickinson BioSciences, San Jose, CA). Apoptosis also was confirmed by annexin V binding according to the manufacturer's guidelines (R & D Systems, Minneapolis, MN).
Mitochondrial and Cytosolic Preparations
Cytochrome c release from the mitochondria to the cytosol was determined by a modified protocol previously described.16 Briefly, 5 × 106 cells were incubated on ice for 5 minutes in 1 mL of ice-cold cell lysis and mitochondria intact (CLAMI) buffer containing 200 g/mL digitonin. Cells were then centrifuged (1020 ×g for 5 minutes at 4 °C) and supernatants containing cytosolic protein were stored at −80 °C. The pellets were then incubated on ice for 10 minutes in universal precipitation buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA, 2 mM ethyleneglycol-bis(β-aminoethyl)-N,N,N,N,-tetraacetic acid (EGTA), 0.2% Triton X-100, 0.3% IGEPAL CA630 (noniodet P-40 [NP-40]), 1 μM aprotinin, 1 μM leupeptin, and 10 μM phenylmethanesulfonyl fluoride [PMSF]). Samples were then centrifuged (10,000 ×g for 10 minutes at 4 °C) and supernatants containing the mitochondrial fraction were stored at −80 °C. Protein samples from each fraction were then quantified and standardized and 50 μg was used in Western blot analysis.
Nuclear and Cytosolic Preparations
To determine the translocation of HIF-1α from the cytosol to the nucleus, 5 × 106 cells suspended in 1 mL of Buffer A (10 mM 4-(2-hydroxyethyl)-piparazineethane sulfonic acid (HEPES), pH 7.5, 1.5 mM MgCl2, 10 mM KCl, 250 mM sucrose, 0.5% NP-40, 1 μM dithiothreitol [DTT], 1 μM leupeptin, 1 μM aprotinin, 10 mM PMSF, 1 μM pepstatin, and 1 mM DTT) were incubated on ice for 5 minutes. Cells were then centrifuged and the supernatant containing cytosolic protein was stored at −80 °C. The pellets were then washed once more in 250 μL Buffer A, centrifuged (20,000 ×g for 20 seconds at 4 °C), and the pellets resuspended in 75 μL Buffer B (10 mM HEPES, pH 8.0, 1.5 mM MgCl2, 1 mM NaCl, 2 mM EDTA, 1 μM leupeptin, 1 μM aprotinin, 10 mM PMSF, 1 μM pepstatin, and 1 mM DTT). Samples were incubated on ice for 10 minutes and centrifuged (20,000 ×g 15 minutes at 4 °C) and the supernatant containing the nuclear protein was stored at −80 °C. Protein samples from each fraction were then quantified and standardized to 50 μg, which was analyzed by Western blotting.
Western Blot Analysis
Protein fractions (50 μg per sample) were boiled in 1 × loading buffer for 5 minutes and electorphoresed on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels. Protein samples were then electrophoretically transferred to Immobilon-P membrane and stained using either anti-cytochrome c (Pharmingen, San Diego, CA), anti–HIF-1-α (Cambridge Biosciences, Rockville, MD), anti-magnesium superoxide dismutase (MnSOD) (StressGen Biotechnologies, Vancouver, British Columbia, Canada), or anti–cIAP-2 (R & D Systems). Equal loading for cytocolic and mitochondrial fractions was determined by restaining the membranes with a monoclonal antibody to β-actin (Becton Dickinson BioSciences) and equal loading for nuclear fractions was confirmed by staining with Coomassie blue. The immobilized protein samples were then incubated with the respective horseradish peroxidase (HRP) secondary antibodies (Becton Dickinson BioSciences) and the signal was detected using the enhanced chemiluminescence substrate system for detection of HRP (Amersham Pharmacia Biotech, Arlington Heights, IL).17
Enzyme-Linked Immunosorbent Assay Detection of Vascular Endothelial Growth Factor
Supernatants were collected and assessed for VEGF by standard enzyme-linked immunosorbent assay analysis according to the manufacturer's instructions (R & D Systems). All results were correct to total cellular protein in the well (ng/mg protein) as assessed above.
Detection of Caspase Activity
Cell lysates were prepared from cells (10 × 106) using a caspase isolation buffer (25 mM HEPES, pH 7.8, 5 mM MgCl2, 1 mM EDTA, 10μM leupeptin, 5 μM pepstatin, 100 μM PMSF, and 10 μM DTT) and caspase isolation buffer (100 mM HEPES, pH 7.5, 10% sucrose, 0.1% 3-[3-cholamidopropyl) dimethylammonio]-1-propane sulfonate (CHAPS), 10 μM leupeptin, 5 μM pepstatin, 100 μM PMSF, and 10 μM DTT). Aliquots of the lysates (40 μL) were diluted in caspase incubation buffer (40 μL) containing either 20 μM of caspase 3 (Ac-DEVD-AMC), caspase 8 (Ac-IETD-AMC), or caspase 9 (Ac-LEHD-AMC) fluorogenic substrates (BIOMOL Int'l; Plymouth, PA). Samples were read initially in a 96-well cytofluorometer II (Perkin Elmer Biosciences, Warrington, UK) at 380 nm excitation and 480 nm emission and a further reading was obtained after 1 hour of incubation. Results were expressed as the specific activity 460 nm/μg protein.17
Dual-Luciferase Reporter Assay
For transfections, cells were seeded in duplicate into replicate wells of 6-well plates (Falcon, Becton Dickinson) at a density of 2 × 105 cells per well. When cells reached 60–80% confluence, they were cotransfected with either 1 μg pGL-3-control (SV40 promoter and enhancer) or 1 μg of the pGL-3-reporter containing the PGK HRE cloned in along with 0.1 μg pRL-CMV control plasmid (Promega, Madison, WI). The PGK HRE sequence (TGTCACGTCCTGCACGACGCGAG) was cloned into the pGL-3-promoter as a trimer in the positive orientation. Expression of renilla luciferase from the pRL-CMV vector was used as an internal control for transfection efficiency, allowing the direct comparison of data from replicate experiments. Control cells were cotransfected with 1 μg of the pGL-3-control vector (Promega) and 0.1 μg of the pRL-CMV vector. The pGL-3-control contains both the SV40 promoter and enhancer sequences driving luciferase-positive expression. This was used as an internal control to monitor both the strength of expression from the various enhancers compared with the native SV40 sequence and the specificity of the transcriptional of each enhancer to each condition used. A dominant negative HIF-1α contruct (pcDNA3.1HIF1α), which lacks the transactivation domain, was used to block endogenouus HIF-1α activity. Empty pHook (1 μg) was used as a control for the No TAD-induced HIF-1α blocking experiments. Cells were transfected using the lipofectamine reagent in OPTIMEM medium (Gibco Life Technologies) according to the manufacturer's instructions. Both the pcDNA3.1HIF1α and the pGL-3-reporter containing the PGK HRE were kindly donated by Kaye Williams (School of Pharmacy and Pharmaceutical Sciences, University of Manchester, Manchester, UK).
Transfection was carried out for 16 hours. The medium was then changed to standard growth medium and 1 of the replicate plates was exposed to standard culture conditions (95% air, 5% CO2) for 24 hours. The other plate was placed in the hypoxia chamber and continually exposed to 1% oxygen in a nitrogen, 5% CO2 mix for 24 hours. After 24 hours, 1 plate from each condition was harvested immediately and lysed with 1 × passive lysis buffer (Promega) and firefly and renilla luciferase were analyzed simultaneously utilizing the Dual-Luciferase reporter assay system (Promega) according to the manufacturer's approved protocols and reagents.
The luminescent signal obtained from the positive expression of firefly luciferase was divided by that obtained for the renilla luciferase to normalize transfection efficiency. To calculate fold induction, the normalized firefly signal for each vector under each treatment condition was divided by the normalized value after 24 hours of aerobic exposure. General transcription/translation effects that may have arisen as a consequence of the treatments used were corrected by relating the fold differences obtained to the output from the pGL-3-control vector in air versus the conditions used. This was nominally assigned a value of 1.
Statistical analysis was carried out using one-way analysis of variance with Student–Newman correction. Significance was assumed for values of P < 0.05, and results are expressed as mean ± the standard deviation.
Hypoxia-Mediated Apoptosis is Dependent on Specific Caspase Activity
To study the effects of hypoxia on apoptosis, PWR-1E, LNCaP, PC-3, and DU-145 cells were cultured under 2 fixed oxygen concentrations (1% O2 [hypoxia] and 21% O2 [normoxia] ± 0.5%). Apoptosis was assayed by subGo PI incorporation using flow cytometry over a 72-hour period. The PWR-1E and LNCaP cells displayed a time-dependent increase in their rates of apoptosis under 1% oxygen with maximum apoptosis occurring at 72 hours. However, the PC-3 and DU-145 cells were resistant to hypoxia-induced apoptosis (Fig. 1A). DNA fragmentation assessed by DNA ladders on gel electrophoresis was also observed at the 72-hour time point (data not shown). Apoptosis was confirmed by annexin V staining for the LNCaP cells (Fig. 1B), which is an early change leading to 85 ± 12% apoptosis at 72 hours. This will lead ultimately to DNA fragmentation.
To further confirm apoptosis and determine a mechanism for its induction, a series of caspase activity assays were performed under hypoxic conditions using fluorogenic substrates specific for caspase 8/10, caspase 9, and caspase 3/7. As shown in Figure 2A, LNCaP cells grown under 1% O2 displayed significant increases in the activity of caspase 8/10 and caspase 3/7 but no detectable caspase 9 activity was observed over 72 hours. Both caspases 8/10 and 3/7 were activated in a similar time-dependent manner with maximum activity peaking as early as 8 hours and remaining elevated for 72 hours. The PC-3 cell line displayed no significant increase in caspase activity for caspases 8/10, 3/7, or 9 (Fig. 2B). This demonstrated that the PC-3 cells were resistant to hypoxia-induced apoptosis. To further confirm the role of the caspases, a pan-caspase inhibitor, z-VAD-fmk, was added before culturing the cells under hypoxic conditions and left in the culture medium for the duration of hypoxia treatment. z-VAD-fmk treatment had a significant protective effect on the rates of hypoxia-induced apoptosis both at 48 and 72 hours at 1% O2 (Fig. 2C).
Hypoxia-Induced Apoptosis is Associated with the Late Release of the Mitochondrial Death Factor SMAC/Diablo and Subsequent Mitochondrial Permeability
Mitochondrial release of cytochrome c and the subsequent activation of caspase 9 through the formation of the apoptosome have been well documented as a signaling event during apoptosis.18 The lack of detectable caspase 9 activity during hypoxia-induced apoptosis in the LNCaP cell line suggested that the mitochondria may not be playing an active role in the signaling for hypoxia-induced apoptosis. To investigate this further, we initially looked at the effects of hypoxia on the mitochondrial trans-membrane potential (ΔΨm). Both LNCaP (Fig. 3A) and PC-3 (Fig. 3B) cells displayed no alteration in ΔΨm during the early time points of hypoxia (1% O2) as determined by their continued ability to uptake DiOC6 during hypoxia, when maximum caspase activity was shown. The LNCaP cells did display a decrease in DiOC6 uptake at the 72-hour time point at 1% O2. The finding that this occurred at a later time suggests that this decrease in ΔΨm was a consequence of cell apoptosis and not a true signaling event leading to caspase activation.
To investigate further the role of the mitochondria in signaling LNCaP apoptosis, mitochondrial and cytosolic fractions were isolated over a 72-hour period under 1% O2. No cytochrome c protein was detected in the cytosolic fractions (Fig. 3C). This helps to explain the lack of caspase 9 activity in the LNCaP cell line, despite undergoing apoptosis. These cells did, however, display a time-dependent increase in cytosolic levels of SMAC/Diablo in response to hypoxia (Fig. 3C) at 24 hours with no detectable amounts at 72 hours, possibly due to degradation inducted by the high level of apoptosis. We used MnSOD as a mitochondrial marker protein, as it localized specifically to the mitochondria. No MnSOD was detectable in the cytosolic fractions, which indicated that there was no mitochondrial contamination of the cytosolic fractions (Fig. 3C). The PC-3 cell line displayed no expression of cytochrome c and SMAC/Diablo protein in the cytosolic fractions under hypoxia for any of the time points as determined by Western blotting (data not shown), which would confirm the lack of apoptosis.
Hypoxia Induced an Increase in Nuclear Expression of Hypoxia-Inducible Factor-1α and Hypoxia-Inducible Factor-1 Transcription
The HIF-1α complex plays a central role in the adaptive response to reduced oxygen by initiating transcription of target genes through promoter regions that have specific hypoxia-response elements (HRE) sequence enhancers. HIF-1α activity can be determined indirectly by looking for the expression of its hypoxic-inducible alpha subunit in cellular nuclear isolates. As previously demonstrated, PC-3 cells displayed a constitutive expression of HIF-1α under normoxia due to a gene amplification of the locus containing the HIF-1A gene, with no expression in the LNCaP cells. However, both the LNCaP and PC-3 cells displayed a time-dependent increase in nuclear protein expression of HIF-1α in response to hypoxia when compared with normoxia (Fig. 4A). The finding that HIF-1α expression still increased under hypoxia in the PC-3 cell line suggested that HIF-1α activation through the stabilization of the alpha subunit still occurred in response to hypoxia. Compared with the PC-3 cells, the LNCaP cells displayed lower levels of nuclear HIF-1α.
To confirm that increased nuclear expression of HIF-1α correlated with increased HIF-1α activity, a series of HIF-1α reporter assays were performed for all cell lines. All cell lines displayed an increase in HRE-mediated reporter activity in response to hypoxia when using the pGL-3 HRE plasmid in normoxia versus hypoxia. The PWR-1E, LNCaP, and DU-145 cell lines underwent an increase in firefly (reporter) activity when standardized with the control renilla luciferase (Fig. 4B). PC-3 cells also underwent an increase in firefly (reporter) activity. however, the basal rate of reporter activity under normoxia was higher than that in the LNCaP cell line when standardized per microgram protein of cell lysate. No significant increase was observed for the pGL-3-control vector in normoxia versus hypoxia for the LNCaP and PC-3 cell lines (0.9 and 1.1-fold difference, respectively).
To confirm other downstream products of HIF-1α activity, we assessed VEGF (Fig. 4C) and cIAP-2 (Fig. 4D) protein expression in the LNCaP and PC-3 cells after different exposure times to hypoxia. Hypoxia increased LNCaP expression of VEGF and cIAP-2 at 48 hours (Fig. 4C,D) in accordance with the increase in apoptosis in the LNCaP cells. There was an increase in the detection of a cleaved product at approximately 40 kilodalton with maximum expression at 72 hours.
PC-3 cells had basally higher levels of VEGF, which were not significantly increased over time (Fig. 4C). As has been previously demonstrated, there was also an increase in the basal expression of cIAP-2 with some increase in expression after 48 and 72 hours exposure to hypoxia.
Hypoxia-Inducible Factor-1 Alpha is not Associated with Resistance to Apoptosis
To determine whether the constitutive expression of HIF-1α was responsible for conferring resistance to hypoxia-induced apoptosis in the PC-3 cell line, a dominant negative HIF-1α construct (pcDNA3.1HIF1α), lacking a functional TAD domain, was used to transfect the PC-3 cells. The dominant negative HIF-1α construct was capable of blocking HRE reporter activity in response to hypoxia (Fig. 5A). This decrease in HRE activity did not have any effect on the rates of apoptosis in the PC-3 cells in response to hypoxia (Fig. 5B). To determine if blocking HIF-1α responses in the LNCaP cells would further increase their response to hypoxia, similar blocking experiments were carried out. Transfection with the dominant negative construct blocked HRE reporter activity in the LNCaP cells, but had no effect on increasing apoptosis in response to hypoxia at 24 hours (Fig. 6). To further explore the possibility that HIF-1α transcription factor expression contributed to apoptotic resistance, we examined the effects of hypoxia in LNCaP-HOF cells. These cells developed androgen independence in response to androgen ablation19 and were shown to be resistant to the effects of hypoxia (Fig. 7A) compared with the parental LNCaP cells. There was no increase in basal HIF-1α expression in the LNCaP-HOF cells compared with the parental LNCaP cells (data not shown). In addition, there was no difference when comparing their HRE reporter activity in response to hypoxia between the two cell types (Fig. 7B).
The mechanisms for the selection of a more aggressive tumor phenotype are largely unknown but the creation of a hypoxic environment may play a role.4 Where hypoxia is associated with the induction of cell death, some cells adapt to the stress and become more resistant to injury or apoptosis.12, 13, 20 HIF-1α transcribes antiapoptotic proteins including the inhibitor of apoptosis (IAP) proteins, specifically cIAP-2,20 which represents important survival factors for prostate carcinoma cells.21–23 Acute hypoxia also has been shown to increase the aggressiveness and survival of prostate carcinoma cells.24 What mediates this differential effect is unknown.
The current study demonstrates that hypoxia induces apoptosis in the PWR-1E and LNCaP cells. In the LNCaP cells, this occurs via an increase in caspase 8 and 3 activity independent of cytochrome c release and mitochondrial disruption. A range of mechanisms are associated with hypoxia-induced apoptosis—from the classical proposed mitochondrial pathway of cytochrome c release, formation of the apoptosome, and activation of caspase 916, 25 to decoupling of mitochondrial cytochrome c release independent of caspase activation.26 However, ultimately, all result in the activation of the caspase cascade. Where the current study demonstrates a cytochrome c-independent mechanism, there might be a role for SMAC release from the mitochondria, which was shown in the LNCaP cells after ≤ 72 hours of hypoxia. This would support the concept that other mitochondrial proteins have a role in initiating the apoptotic cascade. SMAC has been shown to neutralize the inhibitors of apoptosis proteins, thus facilitating the induction of apoptosis.27 Recent studies also have suggested that SMAC release is one of the prerequisites for caspase activation in prostate carcinoma cells.28 The metastatic PC-3, DU-145, and LNCaP-HOF cell lines were resistant to the effects of hypoxia, which could be a mechanism by which hypoxia selects a more aggressive and metastatic cancer phenotype that is associated with poor prognosis.24, 29 It has been shown also that PC-3 cell resistance to apoptosis is further increased after hypoxia.30 These effects are mediated through the increased expression of antiapoptotic factors including the IAP31 and Bcl-213, 32 family of proteins.
It is well established that prostate carcinoma exists in a hypoxic environment,6, 7, 33 which is ultimately associated with an increase in the expression of HIF-1α.7, 34 Expression of HIF-1α has been associated with the aggressiveness of prostate carcinoma cell lines,21, 35 with HIF-1α–stimulated transcription lowest in PrEC and LNCaP cells and highest in PC-3 and highly metastatic PC-3M cell lines.35 These alterations in transcription have been shown to decrease c-myc and cyclin D1 expression, altering cell cycle control. Studies have demonstrated that HIF-1α promotes the expression of a number of survival proteins, including the IAP proteins.20, 31, 36 This is supported in the current study, which demonstrates that hypoxia increases cIAP-2 expression in the LNCaP and, to a lesser extent, in the PC-3 cells, which already have a higher basal level of cIAP-2. There is a corresponding cleavage of cIAP-2 in the LNCaP cells, which undergo apoptosis and would confirm a role for the caspases. PC-3 cells may be resistant to the apoptotic signals of hypoxia due to the higher basal expression of cIAP-2. Recent studies have implicated the IAP in prostate carcinoma cell resistance to apoptosis21, 23 and in the development and progression of prostate carcinoma.22 In addition, increases in VEGF expression represent an important switch in progression, which leads to a more aggressive phenotype.34 We hypothesized that the HIF-1α transcription factor is central to the differential susceptibility to apoptosis that is seen in different prostate carcinoma cells. We initially supported this hypothesis as the PC-3 cells, which were resistant to hypoxia-induced apoptosis, have an increased basal expression of HIF-1α, as has been demonstrated previously.7 However, on preventing the increase in HIF-1α activity, with a dominant negative construct, the PC-3 cells did not increase their susceptibility to hypoxia-induced apoptosis. As the PC-3 cells are a more advanced cancer cell line with other genetic and cellular alterations, they may already have established mechanisms that made them resistant to apoptosis, including overexpression of the IAPs.21, 35 We also examined the effects of this dominant negative construct in the LNCaP cells. These did not demonstrate any further increase in their apoptotic susceptibility to hypoxia after a decrease in the HIF-1α response. These results demonstrate that hypoxia stabilization of the HIF-1α transcription factor may not be the only factor contributing to apoptotic resistance. We finally examined the effects of hypoxia on LNCaP-HOF cells, which were passaged from the androgen-dependent LNCaP parental cells to an androgen-independent phenotype. These cells demonstrate a resistance to hypoxia-induced apoptosis but have no increase in the basal expression of HIF-1α or response to hypoxia. Recent data from our laboratory have demonstrated that this effect is due to an altered apoptotic phenotype at the level of the mitochondria (unpublished data).
The current study demonstrates that resistance to hypoxia-induced apoptosis may be due to an already altered phenotype and is not dependent on hypoxia stabilization of HIF-1α and its associated transcription of proteins. This independence from HIF-1α may only be specific to apoptosis, as there is significant evidence to suggest a positive correlation between HIF-1α expression and a more aggressive tumor,21, 37 including solid tumor growth.38 It has been well documented that HIF-1α has a number of key regulatory functions in the tumor, including angiogenesis, glycolysis, growth factor signaling, immortalization, genetic instability, tissue invasion, and metastasis.39 The findings of the current study have important implications in any strategy that might target HIF-1α to prevent apoptotic resistance in prostate carcinoma.
The authors acknowledge Dr. Kaye Williams, School of Pharmacy and Pharmaceutical Sciences, University of Manchester, for supplying the pcDNA3.1 HIF-1α and pGL-3-reporter. They also thank Professor Helmut Klocker, Department of Urology, University of Innsbruck, Austria, for the LNCaP-HOF cells.