The von Hippel-Lindau gene product inhibits renal cell apoptosis via Bcl-2-dependent pathways.

Previous studies have reported a protective role for the von Hippel-Lindau (VHL) gene products against pro-apoptotic cellular stresses, but the mechanisms remain unclear. In this study, we examined the role of VHL in renal cells subjected to chemical hypoxia, using four VHL-negative and two VHL-positive cell lines. VHL-negative renal carcinoma cells underwent apoptosis following chemical hypoxia (short-term glucose deprivation and antimycin treatment), as evidenced by morphologic changes and internucleosomal DNA cleavage. Reintroduction of VHL expression prevented this apoptosis. VHL-negative cells displayed a significant (greater than 5-fold) activation of caspase 9 and release of cytochrome c into the cytosol following chemical hypoxia. In contrast, VHL-positive cells showed minimal caspase 9 activation, and absence of cytochrome c release under the same conditions. Caspase 8 was only minimally activated in both VHL-negative and -positive cells. In addition, VHL-positive cells displayed a striking up-regulation of Bcl-2 expression (5-fold) following chemical hypoxia. Antisense oligonucleotides to Bcl-2 significantly down-regulated Bcl-2 protein expression in VHL-positive cells and rendered them sensitive to apoptosis. Overexpression of Bcl-2 in VHL-negative cells conferred resistance to apoptosis. Our results suggest that VHL protects renal cells from apoptosis via Bcl-2-dependent pathways.

Inactivation of the von Hippel-Lindau (VHL) 1 tumor suppressor gene is responsible for the polycystic kidney disease and renal cell carcinoma (RCC) characteristic of the hereditary VHL cancer syndrome and for the majority of sporadic RCC (reviewed in Ref. 1). The VHL-negative human RCC cells are tumorigenic in nude mice, and reintroduction of VHL abrogates their tumorigenicity (2,3), establishing a role for the VHL gene as a tumor suppressor. Two biologically active products exist for the VHL gene: full-length VHLp24(MPR) and VHLp18(MEA) created by translation initiation from an internal site (3). Biochemical studies have revealed that the VHL gene products interact with elongins (4), CUL-2 (5), and Rbx1 (6) and possess ubiquitin ligase-like activity (7), suggesting a role in directing substrates for proteasomal degradation. Other studies have proposed a gatekeeper role for VHL in kidney cells, because it has been shown to induce renal cell differentiation and growth arrest (8,9), and regulate the renal cell response to stress (9 -11). Indeed, reintroduction of VHL in VHL-negative RCC cells provides protection from the cytotoxic effects of serum withdrawal (9), glucose deprivation (11), and UV irradiation (10). VHL has also been shown to down-regulate the expression of hypoxia-inducible genes in RCC (reviewed in Ref. 1). However, the role of VHL in modulating the apoptotic effects of hypoxia remains unknown.
Apoptosis or programmed cell death is characterized by distinct morphologic changes consisting of cell shrinkage, nuclear condensation, and internucleosomal DNA fragmentation (13). It has been observed in an increasing array of renal disorders (14,15) and particularly mediates renal tubule cell death in polycystic kidney disease (16), hypoxic-ischemic injury (17), nephrotoxicity (15), and proteinuric injury (18). Thus renal tubule cell apoptosis has emerged as a final common pathway in response to a wide variety of cellular stresses experienced at an intensity below the threshold for necrosis. This has been particularly well documented in in vitro studies using chemical hypoxia, in which mild injury leads to partial ATP depletion and apoptosis, whereas severe injury results in complete ATP depletion and necrotic cell death (17). This is consistent with the notion that apoptosis is a process that requires energy and that intracellular ATP concentration plays a crucial role in the determination of cell death fate by apoptosis or necrosis (17).
In general, the caspase family of proteases constitutes the ultimate effector of apoptosis, and it is convenient to classify the major intracellular apoptotic pathways according to the type of specific pro-caspase that is activated (19,20). Activation of the initiator pro-caspase 8 results predominantly from signaling via integral membrane death receptors such as Fas and TNFR1 (21). On the other hand, activation of the initiator pro-caspase 9 is dependent primarily on mitochondrial signaling pathways involving members of the Bcl-2 family (19,22,23). The Bcl-2-dependent pathways have recently been implicated in the pathogenesis of renal cell apoptosis following hypoxic-ischemic injury (24 -27). The best-characterized events following activation of pro-apoptotic Bcl-2 family members, such as Bax, include induction of mitochondrial permeability transition, release of mitochondrial cytochrome c into the cytosol, and activation of pro-caspase 9 (28 -32). Once activated, both caspase 8 and 9 participate in a cascade that culminates in the activation of caspase 3, which cleaves several substrates resulting in chromosomal DNA fragmentation and cellular morphologic changes characteristic of apoptosis (20). In contrast, the anti-apoptotic Bcl-2 family members such as Bcl-2 and Bcl-xL play a pivotal and overriding protective role by preserving mitochondrial structure and function, preventing onset of mitochondrial permeability transition, and inhibiting the release of cytochrome c into the cytosol (19,23,29,30).
In this study, using modified protocols of in vitro chemical hypoxia (17,(33)(34)(35)(36)(37)39), we show that VHL-negative human 786-0 RCC cells undergo dose-dependent apoptosis. Reintroduction of VHL strikingly protected these cells from apoptosis. VHL-negative cells displayed a significant activation of caspase 9 and release of mitochondrial cytochrome c into the cytosol following chemical hypoxia. In contrast, VHL-positive cells showed only a very modest activation of caspase 9 and no cytochrome c release under the same conditions. Caspase 8 was only minimally activated in both VHL-negative and -positive cells. In addition, VHL-positive cells displayed a marked upregulation of Bcl-2 expression when compared with VHL-negative cells. Antisense oligonucleotides to Bcl-2 significantly down-regulated Bcl-2 protein expression in VHL-positive cells and rendered them sensitive to apoptosis. Overexpression of Bcl-2 in VHL-negative cells conferred resistance to apoptosis. Our results suggest that VHL protects renal cells from apoptosis via Bcl-2-dependent pathways. We speculate that loss of VHL may sensitize cells to the pro-apoptotic effects of hypoxia, which may drive the selection of cells that can evade cell death, leading to tumorigenesis.

EXPERIMENTAL PROCEDURES
Cell Lines and Culture-The parental VHL-negative renal carcinoma cell line 786-0 was obtained from the American Type Culture Collection (Rockville, MD). The VHL-positive cell lines used in this study have been previously characterized (3,8,10) and include those stably expressing the VHLp24(MPR) or VHLp18(MEA) isoforms in the expression vector pCR3 (Invitrogen, CA). The other VHL-negative cell lines include lines containing either the empty expression vector alone or the functionally compromised VHL deletion constructs VHL(MPR)del and VHL(MEA)del, as previously described (3,8,10). Both deletion constructs lack VHL amino acids 114 -178, which are critical to VHL subcellular localization (38) and tumor suppressor functions (3,8). Cells were seeded at high density and grown in six-well polystyrene plates (Corning, NY) in a humidified incubator (37°C, 5% CO 2 ) with Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (both from Life Technologies Inc., Gaithersburg, MD). Stable transfectants were maintained in medium containing 600 g/ml G418 (Life Technologies). Cells were typically fully confluent within 2 days.
Chemical Hypoxia-We modified previously described protocols of in vitro chemical hypoxia by substrate depletion and/or inhibition of oxidative phosphorylation (17,(33)(34)(35)(36)(37)39) in 786-0 RCC cells. On the second day post-confluence, cells were incubated in glucose-free DMEM (Life Technologies), complete DMEM with 1 M antimycin A (Sigma Chemical Co., St. Louis, MO), or both (glucose-free DMEM with 1 M antimycin A). The latter has been shown to result in rapid, partial, and reversible ATP depletion and has been used as an in vitro model of ischemic renal cell injury (17,39). We have previously shown that such partial ATP depletion results in apoptosis of cultured distal tubule (Madin-Darby canine kidney) cells, whereas greater degrees of ATP depletion leads to necrotic cell death (17). We monitored ATP levels using a luciferase-based assay kit (Sigma). Briefly, cells were analyzed after 4 and 12 h of chemical hypoxia or incubation in control medium. An equal number of cells were incubated in 500 l of ATP-releasing agent for 5 min, the sample was cleared of insoluble material by centrifugation, and equal volumes of the supernatant were added to an equal volume of ATP assay mix. ATP levels were measured with a luminometer and expressed as a percentage of control values.
Overall cell viability was monitored using a MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide)-based assay kit (Sigma), designed for the measurement of cell viability as a function of mitochondrial activity in living cells. Briefly, equal numbers of cells were analyzed after 12 h of chemical hypoxia or incubation in control medium. MTT solution (10% of culture volume) was added, and the cells were returned to the incubator for an additional 3 h. MTT solvent was then added in an amount equal to the culture volume, and absorbance was measured spectrophotometrically at 570 nm. Background absorbance measured at 690 nm was subtracted, and the values were expressed as percentage of control values.
Apoptosis Assays-Internucleosomal DNA fragmentation was detected after 12 h of incubation in glucose-free medium containing antimycin, by DNA laddering assay (17). Briefly, equal numbers of cells were resuspended in 500 l of lysis buffer (1% SDS, 25 mM EDTA, 1 mg/ml proteinase K, pH 8) and incubated overnight at 50°C. Ribonuclease A (10 mg/ml) was then added for an additional 2-h incubation at 37°C. The chromosomal DNA was extracted with phenol/chloroform, precipitated with ethanol, and analyzed by agarose gel electrophoresis followed by staining with ethidium bromide to reveal the fragmentation pattern. DNA fragmentation was confirmed in situ utilizing the TdTmediated dUTP nick end-labeling (TUNEL) assay (ApoAlert DNA fragmentation assay kit, CLONTECH, La Jolla, CA), by which fluorescein-dUTP incorporation at the free ends of fragmented DNA is visualized using fluorescent microscopy (17). Briefly, adherent cells grown on coverslips and subjected to control conditions or chemical hypoxia for 12 h were washed with PBS, and fixed with 4% formaldehyde/PBS for 30 min at 4°C. After permeabilization with 0.2% Triton X-100/PBS for 15 min at 4°C, cells were incubated with a mixture of nucleotides and TdT enzyme for 60 min at 37°C in a dark, humidified incubator. The reaction was terminated with 2ϫ SSC, the cells were washed with PBS, and the coverslips were mounted on glass slides with Crystal/mount (Biomeda, Foster City, CA). Fluorescent nuclei were detected by visualization with a microscope equipped with fluorescein filters (IX70, Olympus). To confirm the occurrence of apoptosis, the characteristic morphologic changes, including cell shrinkage and nuclear condensation were detected by phase microscopy.
Caspase Assays-Caspase 8 and caspase 9 activity assays were performed using the ApoAlert Caspase 8 and 9 fluorescence assay kits (CLONTECH), respectively. Briefly, equal numbers of cells from each of the six cell lines were examined before or after 12 h of chemical hypoxia induced by glucose deprivation and antimycin. Cells were incubated in cell lysis buffer for 10 min and centrifuged, and the supernatants were incubated in reaction buffer containing IETD-AFC (specific substrate for caspase 8) or LEHD-AMC (specific substrate for caspase 9) at 37°C for 1 h, and assayed using a fluorometer.
Western Blots-For the detection of Bcl-2, Bcl-xL, and tubulin, equal number of cells were washed with PBS, solubilized in 2ϫ SDS sample buffer, boiled, and subjected to SDS-polyacrylamide gel electrophoresis as previously described (17). For the detection of released cytochrome c, cytosolic extracts were prepared as previously described (32). Briefly, cells were washed with PBS and incubated for 30 min on ice in 300 l of lysis buffer (68 mM sucrose, 200 mM mannitol, 50 mM KCl, 1 mM EGTA, 1 mM dithiothreitol, and 1ϫ Complete (Roche Molecular Biochemicals, Indianapolis, IN) protease inhibitor). Cells were lysed by six passages through a 25-guage needle and centrifuged at 4°C at 800 ϫ g. The supernatant was centrifuged at 14,000 ϫ g for 10 min, and the resultant supernatant (cytosol) was subjected to Western blotting. All protein concentrations were determined by the Bradford assay (Bio-Rad, Hercules, CA), and equal amounts of protein were loaded (30 g in each lane). Two different antibodies each were employed for the detection of Bcl-2 and Bcl-xL, for confirmation of differential expression in the various cell lines. The antibodies to Bcl-2 included a monoclonal at 1:500 dilution (clone #7, Transduction Laboratories, Lexington, KY) and a second monoclonal at 1:500 dilution (clone #124, Upstate Biotechnology Inc., Lake Placid, NY). The antibodies to Bcl-xL included a monoclonal at 1:500 dilution (clone #44, Transduction Laboratories) and a polyclonal at 1:500 dilution (Upstate Biotechnology). Monoclonal antibody to cytochrome c (clone #C-7, Upstate Biotechnology) was used at 1:1000 dilution. Monoclonal antibodies to Bax, Bad, Bid, and Bak (Transduction Laboratories) were used according to the manufacturer's recommendations. Monoclonal antibody to ␣-tubulin (Sigma) was employed at 1:10,000 dilution for confirmation of equal protein loading. Immunodetection of transferred proteins was performed using enhanced chemiluminescence (Amersham Pharmacia Biotech, Arlington Heights, IL).
Bcl-2 Antisense Oligonucleotide Experiments-Antisense and sense 18-mer oligonucleotides directed to the translation initiation site of the human Bcl-2 transcript were synthesized as follows: antisense, 5Ј-TCTCCCAGCGTGCGCCAT-3Ј; and sense, 5Ј-ATGGCGCACGCTGG-GAGA-3Ј (40). Cells were washed with Opti-MEM (Life Technologies). Oligonucleotides (final concentration, 10 M) were incubated in 10 mg/ml LipofectAMINE reagent (Life Technologies) in Opti-MEM for 30 min at room temperature and then applied to the washed cells for 5 h at 37°C (41). After removal of the oligonucleotide-containing medium, cells were washed and incubated in standard medium for 24 h, subjected to chemical hypoxia for an additional 12 h, and analyzed for Bcl-2 protein expression and DNA fragmentation as described above. In parallel experiments, cells were co-transfected with an expression vector containing GFP (pEGFP, CLONTECH) to demonstrate efficiency of transfection.

VHL-negative 786-0 RCC Cells Undergo Apoptosis following
Chemical Hypoxia-786-0 RCC cells were exposed to chemical hypoxia (17,(33)(34)(35)(36)(37)39) by glucose depletion and/or inhibition of oxidative phosphorylation via antimycin treatment. Post-confluence cells maintained in glucose-free medium for a short duration displayed only a modest decrease in ATP content, as shown in Fig. 1 (85% of control after 12 h from four separate experiments). Comparable results were obtained for cells treated with 1 M antimycin A alone. However, cells incubated in glucose-free medium with the addition of 1 M antimycin A showed a sharp and significant drop in ATP levels (28% of control after 12 h from four separate experiments). Longer incubations in glucose-free medium with antimycin resulted in undetectable ATP levels, with the majority of cells showing signs of necrotic cell death by 24 h (not shown). Thus, we chose to further examine the cells after 12 h, with partial ATP depletion. The MTT-based assay for cell viability revealed that the majority of cells treated with either glucose deprivation alone or after antimycin alone remained viable after 12 h, as shown in Fig. 2. However, only 55% of the cells (mean from four separate experiments) remained viable following partial ATP depletion with 12 h of combined glucose deprivation and antimycin treatment. We next tested the hypothesis that apoptosis represents the major form of cell death in 786-0 RCC cells following brief periods of chemical hypoxia. Internucleosomal DNA fragmentation was detectable by the characteristic 180-bp laddering pattern in cells subjected to 12 h of combined glucose-free and antimycin incubation, as illustrated in Fig. 3. Apoptosis was confirmed by detection of fragmented chromosomal DNA by TUNEL assay and by the morphologic appear-ance of shrunken cells with condensed nuclei, both shown in Fig. 4.
Reintroduction of VHL Protects 786-0 RCC Cells from Apoptosis Induced by Chemical Hypoxia-Because VHL has been shown to provide protection from the cytotoxic effects of serum withdrawal (9), prolonged glucose deprivation (11), and UV irradiation (10), it was of interest to determine the role of VHL in modulating the apoptotic effects of chemical hypoxia. This was examined in four VHL-negative (parental 786-0 cells, or cells stably transfected with expression vector alone or with nonfunctional VHL deletion mutants) and in two VHL-positive (786-0 cells stably transfected with VHLp24MPR or VHLp18MEA) cell lines. All six cell lines were devoid of DNA fragmentation prior to treatment, as shown in Fig. 5. However, following chemical hypoxia for 12 h, all VHL-negative lines tested displayed significant DNA laddering, whereas the VHLpositive cells were strikingly protected from apoptosis (Fig. 5). This result was confirmed in three separate experiments.
Reintroduction of VHL Prevents Activation of Caspase 9 following Chemical Hypoxia-It was of interest to determine the mechanism underlying the protective effect of VHL noted above. Because the major intracellular apoptotic pathways may be classified depending on the type of initiator pro-caspase activated (19,20), we evaluated the activity of caspase 8 and 9. Specific assays revealed only a very modest (about 2-fold, from four separate experiments) increase in caspase 8 activity in all six cell lines tested, independent of the presence or absence of VHL (Fig. 6). In contrast, VHL-negative cells displayed a significant (about 6-fold, from four separate experiments) increase in specific caspase 9 activity following chemical hypoxia for 12 h, whereas the similarly treated VHL-positive cell lines showed only a minimal (about 2-fold) increase in caspase 9 activity (Fig. 7). These results suggested that the protective effect of VHL leads to a relative inhibition of caspase 9 activation, thereby implicating mitochondrial signaling pathways involving members of the Bcl-2 family (19,22,23).
Reintroduction of VHL Prevents Release of Cytochrome c following Chemical Hypoxia-It is well known that activation of caspase 9 usually requires the release of cytochrome c from the mitochondria into the cytosol (31,32). Because VHL-positive cells displayed a blunted caspase 9 activation following chemical hypoxia, we investigated cytochrome c in these cells. Subcellular fractionation and examination of the cytosolic fractions of the six cell lines following chemical hypoxia for 12 h showed that cytochrome c was easily detected in the cytosol of all VHL-negative cell lines (Fig. 8). In striking contrast, no cytochrome c was found in the cytosolic fractions of either of the two VHL-positive lines. These results provide an explanation for the marked activation of caspase 9 in VHL-negative cells, and further implicate mitochondrial pathways involving Bcl-2 family members in the protective effect of VHL.
VHL-positive Cells Display a Significant Up-regulation of Bcl-2 and Bcl-xL Expression-Given the differences in cytochrome c release noted above, we next tested for differential expression of Bcl-2 family members in VHL-negative and -positive cells following chemical hypoxia for 12 h. No differences in the expression of pro-apoptotic Bcl-2 family members (Bax, Bad, Bid, and Bak) was noted (not shown). However, the VHLpositive cells displayed a significant up-regulation of the antiapoptotic factors Bcl-2 and Bcl-xL that persists following chemical hypoxia (Fig. 8). These important findings were confirmed using two distinct antibodies to Bcl-2 or Bcl-xL, obtained from two different companies (not shown). Densitometric analysis of five separate blots showed an approximately 5-fold increase in Bcl-2, and a 2-fold increase in Bcl-xL expression (not shown). These findings suggest that VHL-mediated up-regulation of anti-apoptotic Bcl-2 family members may constitute an under-lying mechanism for protection from apoptosis induced by chemical hypoxia.
Down-regulation of Bcl-2 Abrogates the Protective Effect of VHL-Because the above studies implicated Bcl-2 up-regulation as a major mechanism of VHL-mediated cytoprotection, we directly tested this hypothesis using antisense oligonucleotides to down-regulate Bcl-2 expression in VHL-positive cells. Western blot analysis of VHL-positive cells following incubation with antisense Bcl-2 oligonucleotides and chemical hypoxia for 12 h showed a marked down-regulation of Bcl-2 expression, when compared with cells similarly treated but with sense Bcl-2 oligonucleotides (Fig. 9). Co-transfection with GFP revealed a transfection efficiency of ϳ60% (not shown). Furthermore, down-regulation of Bcl-2 abrogated the protective effect of VHL, and these cells were rendered markedly apoptotic following chemical hypoxia (Fig. 9). These results suggest a pivotal role for Bcl-2 in the VHL-mediated protection from apoptosis induced by chemical hypoxia in RCC cells.
Overexpression of Bcl-2 Protects VHL-negative Cells from Apoptosis-We obtained three independent VHL-negative cell lines overexpressing Bcl-2 by approximately 4-fold (Fig. 10). When compared with untransfected 786-0 cells, the cells overexpressing Bcl-2 were markedly resistant to apoptosis following chemical hypoxia (Fig. 10). These results confirm the central role for Bcl-2 in the regulation of apoptosis in VHLnegative 786-0 cells.

FIG. 8. VHL-positive cells display up-regulation of Bcl-2 and
Bcl-xL expression and absence of cytochrome c release following chemical hypoxia. Following treatment with a combination of glucose deprivation and antimycin for 12 h, an equal amount (30 g) of protein from each cell line was analyzed by Western blotting with antibodies to Bcl-2 and Bcl-xL (Transduction Laboratories), and antibody to tubulin (as a control for protein loading). Western analysis with additional Bcl-2 and Bcl-xL antibodies (Upstate Biotechnology) confirmed these results. For detection of cytochrome c release, cytosolic fractions were isolated as described under "Experimental Procedures." These fractions contained equal amounts of tubulin (not shown), but cytosolic cytochrome c was detected only in VHL-negative cells. Shown is a representative result from five experiments.

DISCUSSION
In the present study, we have demonstrated that VHL-negative 786-0 RCC cells undergo significant dose-dependent apoptosis following chemical hypoxia. Reintroduction of VHL strikingly protected these cells from apoptosis. VHL-negative cells displayed a marked (greater than 5-fold) activation of caspase 9 and release of mitochondrial cytochrome c into the cytosol following chemical hypoxia. In contrast, VHL-positive cells showed only a very modest activation of caspase 9 and no cytochrome c release under the same conditions. No significant changes were seen in caspase 8 activation in VHL-negative or -positive cells. In addition, VHL-positive cells displayed a 5-fold up-regulation of Bcl-2 expression when compared with VHLnegative cells. Antisense oligonucleotides to Bcl-2 significantly down-regulated Bcl-2 protein expression in VHL-positive cells and rendered them sensitive to apoptosis induced by chemical hypoxia. Overexpression of Bcl-2 in VHL-negative cells conferred resistance to apoptosis. Our results suggest that VHL protects renal cells from apoptosis via Bcl-2-dependent pathways.
Renal tubule cell apoptosis has been consistently observed after brief periods of in vivo renal ischemia (14,15) and may represent a direct mechanism by which tubule cells are damaged. Ischemic injury produced by chemical hypoxia has also been shown to induce apoptosis in vitro in LLC-PK1 cells (37), mouse proximal tubule cells (39), and Madin-Darby canine kidney cells (17). The latter two studies demonstrated that cells subjected to partial ATP depletion (25-70% of control) died by apoptosis, whereas ATP depletion below 15% of control resulted in necrosis. Our current findings in 786-0 RCC cells are in agreement with those studies and emphasize the concept that apoptosis is an energy-requiring process and that intracellular ATP concentration plays a crucial role in the determination of cell death fate by apoptosis or necrosis (42).
Although activation of endonucleases (37) has been shown to occur in the final "post-mortem" phase, the proximal pathways involved in the stimulus recognition, signal transduction, and effector phases of renal tubule cell apoptosis following ischemic injury or chemical hypoxia remain under active investigation. The Bcl-2-dependent pathways have recently been implicated in renal cell apoptosis following hypoxia-reperfusion injury both in vivo (24) and in vitro (26,27). In one kidney cell culture model, the combination of hypoxia and glucose deprivation resulted in ATP depletion, release of cytochrome c, and caspase activation, followed by apoptosis during reoxygenation in the continued absence of glucose (26). However, inclusion of glucose during hypoxia or during reoxygenation abrogated the ATP depletion and the cell injury, suggesting that ATP depletion (and not the hypoxia itself) was the cause of apoptosis activation (26). In addition, Bcl-2 overexpression prevented cytochrome c release without ameliorating the ATP depletion. In the present study, we have shown that VHL-negative RCC cells respond to chemical hypoxia-induced ATP depletion by releasing cytochrome c and activating caspase 9, culminating in apoptosis. We have made the novel observation that reintroduction of VHL is associated with Bcl-2 overexpression, inhibition of cytochrome c release, blunting of caspase 9 activation, and a remarkable protection from hypoxia-induced apoptosis. Down-regulation of Bcl-2 in VHL-positive RCC cells resulted in a return of their susceptibility to apoptosis similar to the VHLnegative state. Overexpression of Bcl-2 in VHL-negative cells conferred protection from apoptosis. Thus, using a different model of ATP depletion, our study has confirmed the pivotal role of Bcl-2 in renal cell apoptosis.
What is the pertinence of examining apoptosis induction by ATP depletion to renal cells or tumors? First, apoptosis is critical to normal renal cell development (14,15), and so is the VHL gene product (9). Second, apoptosis mediates renal tubule cell death in polycystic kidney disease (16), which is an important characteristic of the VHL syndrome phenotype (1). Third, apoptosis mediates renal tubule cell death following ATP depletion secondary to hypoxia (17), and local areas of hypoxia and apoptosis within VHL-negative tumors have been proposed by Gorospe et al. (11) to stimulate vascularization and angiogenesis, giving rise to the highly vascularized tumors characteristic of the VHL phenotype.
Multiple mechanisms probably underlie the protective effects of VHL from renal cell apoptosis following cellular stresses. In the VHL-negative renal carcinoma cell line UMRC6, Gorospe et al. (11) have reported that prolonged glucose deprivation (3 days) resulted in apoptosis, which was prevented by VHL reintroduction. In that study, VHL-negative cells also showed a greater accumulation of misfolded or misprocessed proteins, which were postulated to interfere with the general maintenance of cellular function and integrity (11). Pause et al. (5) have shown that VHL-negative 786-0 RCC cells continue to proliferate and undergo apoptosis following serum starvation and that reintroduction of VHL enables these cells to accumulate p27 (a cyclin-dependent kinase inhibitor) and exit the cell cycle upon serum withdrawal. In addition, our group has recently reported that VHL-negative 786-0 RCC cells undergo apoptosis following UV treatment. Reintroduction of VHL is not only associated with up-regulation of Bcl-2 and Bcl-xL but also with accumulation of cyclin-dependent kinase inhibitors p27 and p21 and protection from apoptosis secondary to UV irradiation (10). We have also recently documented stabilization of p27 and p21 in VHL-positive cells following chemical hypoxia. 2 Thus, the collective data indicate that VHL is required for elimination of abnormal proteins and for cell cycle exit following cellular stresses and that a failure of these mechanisms may lead to enhanced sensitivity to apoptotic stimuli. The present study is the first to directly address the intracellular apoptotic mechanisms involved.
How does reintroduction of VHL lead to an up-regulation of Bcl-2 and Bcl-xL expression? Although the mechanism remains unclear, several possibilities may be surmised. In a recent study with the Wilms' tumor suppressor gene (WT1), overexpression of WT1 in rhabdoid cell lines up-regulated Bcl-2 via transcriptional mechanisms and conferred resistance to a variety of apoptotic stimuli (43). The VHL gene may similarly influence Bcl-2 gene transcription, especially because VHL has been shown to transcriptionally regulate the expression of several genes (44), including vascular endothelial growth factor (12), GLUT1 (12), tyrosine hydroxylase (45), and transforming growth factor-␤1 (46). It is also possible, given its role in targeting proteins for ubiquitination (7), that VHL enhances the degradation of factors that may either destabilize Bcl-2 mRNA or facilitate Bcl-2 protein degradation. Alternatively, increased levels of Bcl-2 and Bcl-xL in VHL-positive cells may simply reflect the induction of renal cell differentiation and maintenance of the epithelial phenotype that have been attributed to VHL (8,47).
The concept that VHL, a tumor suppressor, can exert an anti-apoptotic effect may appear counterintuitive at first, although it has now been demonstrated following a variety of apoptotic stimuli, including serum withdrawal (5), glucose deprivation (11), UV irradiation (10), and chemical hypoxia (this study). We hypothesize that loss of VHL function may increase the sensitivity of cells to physiologic stresses, thereby providing selective pressure for cells to override apoptosis under these conditions. The clonal outgrowth of VHL-negative cells that have evaded death and possibly accumulated additional genetic changes may then contribute to cystogenesis and tumorigenesis. Such a concept has been recently reviewed by Lowe and Lin (48), and is not unique to pVHL. Indeed, a similar mechanism has been reported for the retinoblastoma tumor suppressor gene Rb. Loss of Rb results in increased apoptosis (49,50), and reintroduction of Rb into RbϪ/Ϫ cells can suppress apoptosis (51). Alternatively, it is also possible that VHL's function as a suppressor of apoptosis may not be related to its tumor suppressor activity but may instead represent a mechanism underlying its role as a gatekeeper in kidney cells, allowing for cell survival and cell differentiation.
In summary, we have shown that VHL protects renal cells from apoptosis via Bcl-2-dependent pathways. It will be important in future studies to examine apoptotic events immediately proximal to cytochrome c release (such as activation of Bax and induction of mitochondrial permeability transition) to further define the direct role of VHL in the Bcl-2-mediated cytoprotection. Elucidation of the interplay between VHL and Bcl-2 family members may lead to clues toward understanding the development and progression of renal cell carcinogenesis that characterizes the VHL-deficient state.