The Role of Antiapoptotic Bcl-2 Family Members in Endothelial Apoptosis Elucidated with Antisense Oligonucleotides*

In this study, we utilized potent antisense oligonucleotides to examine the role of two Bcl-2 family members found in human umbilical vein endothelial cells (HUVEC). The first, A1, is thought to be a TNF-α-inducible cytoprotective gene, and the second, Bcl-XL, is constitutively expressed. Inhibition of the constitutive levels of Bcl-XL caused 10–25% of the cell population to undergo apoptosis and increased the susceptibility of cells to treatment with low concentrations of staurosporin or ceramide. The caspase inhibitor benzyloxycarbonyl-Val-Ala-Asp(OMe)-CH2 prevented DNA fragmentation and ΔYm loss caused by Bcl-XL inhibition or Bcl-XL inhibition combined with staurosporin. However, disruption of ΔYm caused by Bcl-XL inhibition combined with ceramide treatment was not inhibited by benzyloxycarbonyl-Val-Ala-Asp(OMe)-CH2, although DNA fragmentation was completely prevented. Taken together, these results demonstrate a direct protective role for Bcl-XL under normal resting conditions and under low level apoptotic challenges to HUVEC. Furthermore, Bcl-XL protects cells from caspase-dependent and -independent mechanisms of ΔYm disruption. In contrast to Bcl-XL, A1 inhibition did not show a marked effect on the susceptibility of HUVEC to undergo apoptosis in response to TNF-α, ceramide, or staurosporin. These results demonstrate that although A1 may be a cytoprotective gene induced by TNF-α, it is not primarily responsible for HUVEC resistance to this cytokine.

The signaling pathways involved in apoptosis have been intensely studied in the last few years due to their role in many disease states including cancer and neurodegenerative disorders. One central player is the Bcl-2 family of proteins, which can either promote cell survival (Bcl-2, Bcl-XL, A1, Mcl-1, and Bcl-W) or promote cell death (Bax, Bak, Bcl-XS, Bad, Bid, Bik, Bim, Hrk, Bok) (reviewed in Ref. 1). The relative amounts or equilibrium between these pro-and antiapoptotic proteins influence the susceptibility of cells to apoptosis. As such, many cancer cells up-regulate the antiapoptotic members as a mechanism to guard against programmed cell death. Although much work has been carried out studying the protective role of Bcl-2 and Bcl-XL in response to many different apoptotic stimuli, relatively little is known about the other antiapoptotic family members or the specific role of each isoform.
Endothelial cells play a pivotal role in modulating the inflammatory response. Stimulation of endothelial cells by tumor necrosis factor ␣ (TNF-␣) 1 activates the transcription factor NF-B, leading to induction of proinflammatory genes. TNF-␣ is also known to trigger programmed cell death, which in most cells is controlled by the simultaneous up-regulation of cytoprotective genes (Refs. 2-4; reviewed in Ref. 5). Because this anti-death activity requires de novo protein synthesis, cells resistant to the cytotoxic effects of TNF-␣, including endothelial cells, can be rendered sensitive by pretreatment with either RNA or protein synthesis inhibitors. To date, several TNF-␣inducible cytoprotective genes have been identified and have been shown through overexpression techniques to protect cells, at least in part, from death induced by TNF-␣ and cycloheximide/actinomycin D. This growing family represents a diverse range of proteins and includes manganous superoxide dismutase (6), plasminogen activator inhibitor type 2 (7), the zinc-finger protein A20 (8), the cyclin-dependent kinase inhibitor p21 (9), xiap, a member of the IAP family (10), IEX-1L, a protein of unknown function (11), the Bcl-2 family member A1 (12), and combinations of c-IAP-1, c-IAP-2, Traf1, and TRAF2 (13).
We initiated studies aimed at addressing the role of A1 as a TNF-␣-inducible cytoprotective gene and comparing its function to Bcl-XL, which is expressed constitutively in human umbilical vein endothelial cells (HUVEC). In order to help address the importance of each isoform in normal tissue, highly selective and potent antisense inhibitors were utilized in this study. We show that inhibition of A1 alone does not render HUVEC susceptible to TNF-␣-induced apoptosis. However, inhibition of Bcl-XL caused HUVEC to undergo apoptosis and sensitized cells to treatment with either ceramide or staurosporin. Furthermore, Bcl-XL protected cells from both caspasedependent and -independent mechanisms of ⌬Ym disruption.

MATERIALS AND METHODS
Cells and Reagents-HUVEC were obtained from Clonetics (San Diego, CA) and cultivated in endothelial growth medium supplemented with 10% fetal bovine serum. Cells were used between passages 2 and 5. Opti-MEM, Lipofectin reagent, and fetal bovine serum were purchased from Life Technologies, Inc. TNF-␣ was obtained from R & D Research (Minneapolis, MN), Bcl-X antibody and the ribonuclease protection assay were from Pharmingen (San Diego, CA). Caspase 3 antibody was from Transduction Laboratories (Lexington, KY), and Mito-Tracker Orange CMTMRos was purchased from Molecular Probes (Eugene, OR). SDS-polyacrylamide gels were obtained from NOVEX (San Diego, CA). Staurosporin, C6-ceramide, benzyloxycarbonyl-Val-Ala-Asp(OMe)-CH 2 (Z-VAD.fmk), and carbonyl cyanide m-chlorophenylhydrazone were purchased from Calbiochem. Hybond Nϩ nylon membranes and the ECL Plus detection kit were from Amersham Pharmacia Biotech.
Oligonucleotide Synthesis and Sequences Used-2Ј-O-Methoxyethyl chimeric antisense oligonucleotides were utilized for all experiments.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The A1 antisense oligonucleotide and its control contained phosphodiester linkages in the 2Ј-O-methoxyethyl regions and phosphorothioate linkages in the 2Ј-deoxynucleotide regions, which support RNase H activity in cells (14). The Bcl-XL antisense oligonucleotide and its control contained phosphorothioate linkages throughout. All oligonucleotides were synthesized using an Applied Biosystems 380B automated DNA synthesizer and purified as described previously (15,16).
Oligonucleotide and TNF-␣ Treatment of HUVEC-HUVEC were treated with oligonucleotides at ϳ80% confluence (17,18). Briefly, cells were washed three times with prewarmed (37°C) Opti-MEM. Oligonucleotides were premixed with 10 g/ml Lipofectin reagent in Opti-MEM at the desired concentration and applied to washed cells. HUVEC were incubated with the oligonucleotides for 4 h at 37°C, after which the medium was removed and replaced with standard growth medium. For experiments carried out with A1 antisense oligonucleotide, the cells were allowed to recover 16 -18 h after oligonucleotide treatment, prior to induction with TNF-␣ (10 ng/ml), C6-ceramide (5 M), staurosporin (1 nM), or Z-VAD.fmk (100 M). Because Bcl-XL is expressed constitutively, treatment with TNF-␣ (10 ng/ml), C6-ceramide (5-10 M), staurosporin (1-2 nM), or Z-VAD.fmk (100 M) was delayed 24 h after oligonucleotide treatment. This delay period allowed time for reduction in Bcl-XL protein prior to the addition of apoptotic stimuli.
Northern Analysis-Total cellular RNA was isolated either 4 h after TNF-␣ induction (A1 antisense experiments), or 24 -48 h after oligonucleotide treatment (Bcl-XL antisense experiments) utilizing Quiagen RNeasy Kit. Isolated RNA was separated on a 1% agarose/formaldehyde gel, transferred to a Hybond Nϩ nylon membrane over-night, and UV-cross-linked (Stratalinker 2400, Stratagene). Blots were hybridized several hours with random primed 32 P-labeled probes (Bcl-XL) or single-stranded PCR 32 P-labeled probes (A1) as described by Bednarczuk et al. (19). Both A1 and Bcl-XL cDNA's were generated by reverse transcription-polymerase chain reaction carried out with HUVEC total RNA. Polymerase chain reaction was subsequently utilized to amplify bases 22-684 of the human A1 gene and bases 34 -856 of the human Bcl-XL gene. The primer utilized to generate the single-stranded A1 probe was 5Ј-AGAAGTATGTGTTGGCAATCGT-3Ј. All Northern blots were reprobed with random primed 32 P-labeled human G3PDH cDNA to confirm equivalent loading of RNA samples.
Western Analysis-Samples for Western analysis were prepared using 400 l of radioimmune precipitation extraction buffer/100-mm plate. The protein concentration for each sample was determined by Bradford analysis. Equal amounts of total protein were separated on a 14% SDS-polyacrylamide minigel, and proteins were transferred to polyvinylidene difluoride membranes. Membranes were treated for a minimum of 2 h with specific primary antibody followed by incubation with secondary antibody conjugated to horseradish peroxidase. Visualization was carried out with the ECL Plus detection kit.
Hypodiploidy Apoptosis Assay-Cells were treated with oligonucleotides and apoptosis inducers as described. Both floating and adherent cells were harvested 24 -48 h after treatment; washed once with 1 ml of PBS, 5 mM EDTA; and fixed with 1 ml of 70% EtOH while vortexing gently. Fixed cells were stored at 4°C for 1 h to several days. Cells were pelleted by centrifugation, washed once with 1.0 ml of PBS, 5 mM EDTA, and resuspended with 0.3-1.0 ml Propidium Iodide mix (250 g/ml propidium iodide, 5 g/ml RNase A, 1 ϫ PBS, and 5 mM EDTA). After incubation in the dark for 1 h at room temperature, the cells were analyzed on a Becton Dickinson FACscan, and apoptotic cells (sub-G 1 population) were quantified.
Mitochondrial Transmembrane Potential (⌬Ym)-To evaluate the ⌬Ym, cells were incubated with the cationic lipophilic dye MitoTracker orange CMTMRos (150 nM) for 15 min at 37°C in the dark (20). Control cells were simultaneously treated with 50 M of the protonophore, carbonyl cyanide m-chlorophenylhydrazone, which disrupts ⌬Ym. Both adherent and floating cells were collected, washed once with PBS, 2% bovine serum albumin, and fixed in 1 ml of PBS containing 4% paraformaldehyde for 15 min while shaking at room temperature. Fixed cells were stored in the dark at 4°C for 1 day prior to analysis by flow cytometry. The percentage of cells with decreased fluorescence, and therefore with reduced ⌬Ym (compared with control cells not treated with oligonucleotides or apoptotic inducers), were determined and indicated on the bar graphs as percentage of cells gated. Cells treated with carbonyl cyanide m-chlorophenylhydrazone generally gave greater than 70% cells gated and showed a shift of 0.5 log units to a lower fluorescence.
Caspase 3 Activity-Caspase 3 activity was measured using Caspase 3 colorimetric assay kit from CLONTECH (Palo Alto, CA). Briefly, whole cell lysates were obtained, protein concentrations were determined, and equal amounts of protein were added to 50 M substrate (Asp-Glu-Val-Asp-p-nitroanilide) in assay buffer. Samples were incubated at room temperature and analyzed using a spectrophotometer (OD 418).

RESULTS
Bcl-2 Family Members in HUVEC-Ribonuclease protection assay was utilized to identify Bcl-2 family members present in basal and TNF-␣-treated HUVEC. High levels of the antiapoptotic members Bcl-XL and Mcl-1 were observed in both basal and induced cells, while Bcl-2 was not detected (Fig. 1). In contrast, the antiapoptotic member A1 was found in low abundance in basal HUVEC but increased ϳ10-fold following TNF-␣ treatment. Maximal induction of A1 was observed ϳ4 h after treatment. Interestingly, Bcl-XL was down-regulated approximately 20% upon treatment with TNF-␣.
a Accession numbers for A1 and Bcl-XL are U29680 and Z23115, respctively.
Inhibition of A1 mRNA through Antisense Oligonucleotide Treatment-An effective antisense oligonucleotide directed toward A1 was identified from a screen targeting 14 distinct positions on the A1 mRNA, extending from the 5Ј-UTR to the 3Ј-UTR using methods as described previously (21). All oligonucleotides were 20 nucleotides in length containing chimeric 2Ј-O-methoxyethyl modifications on a mixed backbone of phosphodiester and phosphorothioate linkages ( Table I). The most effective A1 antisense oligonucleotide was directed toward the 3Ј-UTR, immediately downstream from the stop codon, and was utilized in all subsequent experiments. Fig. 2 shows the concentration-dependent decrease of the A1 mRNA with increasing amounts of A1 antisense oligonucleotide, while a mismatch control oligonucleotide failed to show a reduction, demonstrating a sequence specific inhibition. The apparent IC 50 for the A1 antisense compound was less than 10 nM. Treatment of HUVEC with 50 nM of the A1 antisense oligonucleotide reduced the A1 mRNA to basal levels, completely inhibiting the TNF-␣-induced component. Reductions in A1 mRNA to basal levels were observed up to 30 h after antisense treatment (data not shown).
Inhibition of Bcl-XL mRNA and Protein through Antisense Oligonucleotide Treatment-A potent, sequence-specific antisense oligonucleotide directed toward Bcl-XL has previously been identified and shown to decrease Bcl-XL mRNA and protein levels in both human keratinocytes and A549 cells. 2 Here we tested its ability to inhibit Bcl-XL mRNA in HUVEC. As shown in Fig. 3, treatment of HUVEC with 100 nM of the Bcl-XL oligonucleotide decreased the Bcl-XL mRNA levels to ϳ5% of control, with an apparent IC 50 less than 20 nM. In comparison, a mismatch control oligonucleotide did not reduce the Bcl-XL mRNA levels. The Bcl-XL antisense compound is targeted to the coding region of the gene and is a chimera of 2Ј-O-methoxyethyl modification with a phosphorothioate backbone (Table I). It should be noted that the Bcl-XL antisense oligonucleotide utilized here targets both the long and short forms of Bcl-X (Bcl-XL and Bcl-XS). However, HUVEC contain such low levels of Bcl-XS mRNA, either basally or after induction with TNF-␣ (see Fig. 1), that the contribution of Bcl-XS to the total Bcl-X mRNA levels detected in our experiments is insignificant.
Bcl-XL protein expression was also examined by Western analysis. As shown in could represent changes in the phosphorylation state of Bcl-XL. If the Bcl-XS protein were present, it should migrate well below the Bcl-XL band under our conditions. Thus, the Bcl-XL antisense compound was able to decrease Bcl-XL protein levels to roughly 5% of amounts in untreated cells.
Specificity of Antisense Oligonucleotide Treatment-In order to demonstrate the specificity of A1 and Bcl-XL antisense oligonucleotides, HUVEC were treated with A1 antisense oligonucleotide, A1 control, Bcl-XL antisense oligonucleotide, or Bcl-XL control, in the presence or absence of TNF-␣. Northern analysis was carried out, and the membrane was probed sequentially with A1, Bcl-XL, and G3PDH-specific probes. The A1 antisense compound was effective in reducing A1 mRNA without affecting the Bcl-XL mRNA. Likewise, the Bcl-XL antisense compound specifically reduced Bcl-XL mRNA without affecting A1 mRNA (Fig. 5). In addition, there was no affect on either target with the control oligonucleotides. These results demonstrate that the antisense oligonucleotides utilized here can specifically inhibit the target mRNA without affecting other Bcl-2 family members.
HUVEC Viability after Inhibition of Endogenous A1 and Bcl-XL Protein: Basal Versus TNF-␣ Induction-Because A1 is induced by TNF-␣ treatment, its role in protecting HUVEC from TNF-␣-induced apoptosis has been suggested (12). To address this point, cells were treated with 50 nM A1 antisense oligonucleotide or control oligonucleotide and subsequently induced with TNF-␣. Apoptotic cells with fragmented DNA were identified by flow cytometry analysis. As shown in Fig. 6A, inhibition of the TNF-␣-induced A1 mRNA did not significantly affect the number of apoptotic cells observed after TNF-␣ treatment. However, cells treated with a combination of TNF-␣ and cycloheximide, which prevents the up-regulation of all TNF-␣inducible cytoprotective genes, resulted in over 50% cell death under these conditions. These results suggest that A1 is not solely responsible for protection of HUVEC from TNF-␣-induced programmed cell death.
Similar experiments were also carried out with the Bcl-XL antisense compound. Inhibition of Bcl-XL protein in basal HU-VEC caused 10 -25% of the cell population to undergo apoptosis (Fig. 6B). Thus, Bcl-XL may play an important role in the basal viability of endothelial cells. On the other hand, inhibition of the Bcl-XL protein did not render cells more sensitive to TNF-␣ treatment (i.e. the difference between Bcl-XL and control remained the same in the presence of TNF-␣), indicating that Bcl-XL is not critical in protecting cells from this apoptotic stimuli.
Because inhibition of either A1 or Bcl-XL protein did not render HUVEC more susceptible to TNF-␣ induced apoptosis, we investigated whether simultaneous inhibition of both proteins would result in increased apoptosis. Therefore, HUVEC were treated with both A1 and Bcl-XL antisense compounds, allowed to recover for 24 h, and then treated with or without TNF-␣. Fig. 6C shows that simultaneous inhibition of A1 and Bcl-XL did not significantly render HUVEC more susceptible to TNF-␣ treatment.
Ability of A1 Antisense Oligonucleotide and Bcl-XL Antisense Oligonucleotide to Sensitize HUVEC to Treatment with Ceramide or Staurosporin: Effect on DNA Fragmentation and Mitochondrial Transmembrane Potential-Next we tested the sensitivity of HUVEC treated with the A1 or Bcl-XL antisense oligonucleotide to treatment with the protein kinase inhibitor staurosporin or the lipid second messenger ceramide. Both of these agents have been shown to induce apoptosis in many cell types including endothelial cells (22,23). Treatment of HUVEC FIG. 5. Specificity of antisense oligonucleotide treatment. Cells were treated with 50 nM of A1 antisense, A1 control, Bcl-XL antisense, or Bcl-XL control oligonucleotides, followed by a recovery period of 24 h. Indicated cells were treated with TNF-␣ for 4 h, and all cells were harvested for total RNA. Northern analysis was carried out, and the membrane was probed sequentially with A1-, Bcl-XL-, and G3PDHspecific probes.

FIG. 6. Effect of A1 and Bcl-XL inhibition on DNA fragmentation (basal versus TNF-␣ induction).
A, HUVEC were treated with Lipofectin alone (Ⅺ), 50 nM A1 antisense (f), or 50 nM control oligonucleotide (_) as indicated. Cells were allowed to recover for 16 h, induced with TNF-␣, cycloheximide (10 g/ml), or TNF-␣ plus cycloheximide, followed by an additional 24-h incubation. Cells were harvested and analyzed by fluorescence-activated cell sorting for hypodiploidy as described under "Materials and Methods." B, HUVEC were treated with Lipofectin alone (Ⅺ), 50 nM Bcl-XL antisense oligonucleotides (f), or control oligonucleotide (_) as indicated. After a 24-h period, TNF-␣ was added to the indicated cells, followed by an additional 24-h incubation. C, HUVEC were treated with Lipofectin alone (Ⅺ), combinations of Bcl-XL-and A1-specific antisense oligonucleotides (50 nM/each) (f), or control oligonucleotides (50 nM/each) (_). Cells were allowed to recover for 24 h, induced with TNF-␣ as indicated and incubated for an additional 24-h period. Results shown are average of duplicate samples. with 1 nM staurosporin or 5 M C6-ceramide alone did not significantly affect the cellular DNA fragmentation. However, cells with decreased levels of Bcl-XL protein showed increased sensitivity toward these low concentrations of apoptotic stimuli (Figs. 7 and 8A). In contrast, cells pretreated with the A1 antisense compound showed only a slight increase in DNA fragmentation after treatment with TNF-␣ in the presence of either staurosporin or ceramide (Fig. 9A). As shown in Fig. 9A, HUVEC were more sensitive to ceramide treatment when carried out in the presence of TNF-␣ as described previously (23).
One of the major mechanisms of Bcl-2-mediated cytoprotection is by protecting mitochondrial function (reviewed in Refs. 24 and 25). Therefore, we investigated the effect of A1 and Bcl-XL inhibition on the ⌬Ym. A reduction in ⌬Ym has been shown to accompany apoptosis in many experimental models and can be monitored by the potential-sensitive, fluorescent dye CMTMRos. To confirm that CMTMRos was sensitive to mitochondrial transmembrane depolarization, control cells were treated with 50 M of the mitochondrial uncoupling agent carbonyl cyanide m-chlorophenylhydrazone. As shown in Fig.  8C, treatment of HUVEC with Bcl-XL antisense compound resulted in a reduction in the ⌬Ym. This reduction was poten-tiated by the combination of Bcl-XL inhibition and treatment with either staurosporin or ceramide. Thus, endogenous Bcl-XL may function by maintaining the mitochondrial integrity of resting HUVEC, where loss of Bcl-XL results in greater susceptibility to mitochondrial dysfunction. In contrast to these results, inhibition of A1 did not effect the ⌬Ym, after treatment with either TNF-␣ alone or TNF-␣ in combination with ceramide or staurosporin (Fig. 9C).

Involvement of Caspases in HUVEC Apoptosis
Resulting from Bcl-XL Inhibition-HUVEC were treated with the A1 or Bcl-XL antisense oligonucleotides or controls, in the presence or absence of Z-VAD.fmk, a broad-based cell-permeable caspase inhibitor. As shown in Fig. 8B, Z-VAD.fmk completely inhibited the DNA fragmentation resulting from Bcl-XL inhibition alone or from Bcl-XL inhibition in the presence of ceramide or staurosporin. Similar results were also obtained using Z-DEVD-.fmk, a known inhibitor of caspases 3, 7, and 8 (data not shown). Besides affecting DNA fragmentation, Z-VAD.fmk also inhibited ⌬Ym loss induced by Bcl-XL antisense oligonucleotide or Bcl-XL antisense oligonucleotide and staurosporin (Fig.  8D). However, inhibition of ⌬Ym was not observed when cells deficient in Bcl-XL protein were treated with ceramide. Bcl-XL antisense oligonucleotide plus ceramide-treated cells were not viable as shown by PI exclusion (data not shown), although DNA fragmentation was completely prevented.
Consistent with caspase inhibition data, proteolysis of procaspase 3 was also observed in Bcl-XL-deficient cells and cells treated with staurosporin. This reduction in procaspase 3 was inhibited by Z-VAD.fmk (Fig. 10A). In addition, separate experiments demonstrated an increase in caspase 3 activity following inhibition of Bcl-XL or Bcl-XL inhibition in the presence of staurosporin. The increased caspase 3 activity correlated with the proteolysis of procaspase 3 observed in the Western blot (Fig. 10B). DISCUSSION To date, much work has been carried out studying the protective role of Bcl-2 and Bcl-XL, while relatively little is known about the other antiapoptotic Bcl-2 family members A1, Bcl-W, and Mcl-1. One unresolved question has been the specific function or molecular mechanism of each isoform, since many cells express more than one. Through ribonuclease protection assay analysis, we show that endothelial cells express Bcl-XL and Mcl-1 constitutively, while A1 is found in very low amounts in basal cells but can be up-regulated by TNF-␣. Therefore, we initiated studies aimed at addressing and comparing the normal physiological role between two of these family members, A1 and Bcl-XL, by utilizing specific antisense oligonucleotides to inhibit the endogenous levels of each protein.
Although the identity of the survival factor(s) responsible for halting the TNF-␣ death pathway remains to be firmly established, a number of recent reports have identified several candidates. One such candidate, A1, was originally identified in phorbol ester-stimulated endothelial cells and shown to be induced by the inflammatory cytokines TNF-␣ and interleukin-1 (27). Further experiments demonstrated that endothelial cells expressing high levels of A1, through retrovirus-mediated transfer, were protected against cell death initiated by TNF-␣ plus actinomycin D (12). This protection was also observed with overexpression of Bcl-XL (12). However, others have reported that overexpression of either A1 or Bcl-XL did not rescue HU-VEC from death induced by TNF-␣ plus cycloheximide, although both effectively inhibited apoptosis induced by TNF-␣ plus ceramide. (23). To help clarify the role of A1 as a TNF-␣inducible cytoprotective gene, we asked whether inhibition of A1 would render cells sensitive to TNF-␣-induced apoptosis. We demonstrate that treatment of HUVEC with a highly specific and potent antisense oligonucleotide directed toward A1 completely inhibited the TNF-␣-inducible component of the A1 mRNA. However, this inhibition did not render cells significantly more susceptible to apoptosis by TNF-␣. Furthermore, cells with decreased levels of Bcl-XL, or both Bcl-XL and A1, continued to show resistance to TNF-␣ treatment. Thus, A1 and Bcl-XL do not appear to play critical roles in protecting endothelial cells from TNF-␣.
In contrast, inhibition of the endogenous levels of Bcl-XL resulted in a substantial amount of cell death in basal cells. In addition, Bcl-XL-deficient cells were rendered sensitive to concentrations of ceramide and staurosporin that were without effect in cells expressing normal levels of Bcl-XL. Thus, Bcl-XL may function physiologically as a protective gene that, under normal resting conditions, helps endothelial cells maintain their quiescent phenotype and perform their normal barrier functions even in the presence of low levels of stimulants (26). This would be critical for endothelial cells in vivo, which are constantly assaulted by noxious compounds such as oxidized low density lipoproteins or cytokines. Our results are consistent with overexpression data indicating that Bcl-XL protects cells from assault from a wide range of apoptotic stimulants, including ceramide (12,28) and staurosporin (29). It is interesting to note that as the concentrations of staurosporin and ceramide increased, the differences observed between normal and Bcl-XL-inhibited HUVEC slowly disappeared (data not shown). These results suggest that endogenous Bcl-XL can protect HUVEC from apoptotic stimuli up to a threshold level after which its presence or absence does not effect the outcome. This may make sense physiologically, since a high level of assault, causing substantial damage to a cell, would be desirable to end in cell death.
One of the major mechanisms of Bcl-2-mediated cytoprotection is by protecting mitochondrial function (reviewed in Refs. 24 and 25). More specifically, these proteins are thought to function through maintaining the mitochondrial permeability transition pore opening (and therefore protecting ⌬Ym) and preventing release of the caspase activators, AIF and cytochrome c, into the cytosol (25). In support of these functions, several groups have now reported that overexpression of Bcl-XL or Bcl-2 confers protection upon mitochondria, making it more difficult for many stimuli to induce PT pore opening and release of AIF and cytochrome c (25, 30 -32).
Consistent with a role in the maintenance of mitochondria integrity, we show that inhibition of endogenous Bcl-XL protein levels caused a reduction in ⌬Ym. This reduction was further potentiated with the combination of Bcl-XL inhibition plus ceramide or staurosporin. However, the observed ⌬Ym loss showed different responses to caspase inhibitors. In cells treated with Bcl-XL antisense oligonucleotide or Bcl-XL antisense oligonucleotide plus staurosporin, Z-VAD.fmk prevented apoptosis, within the time frame studied, as shown by a lack of DNA fragmentation and maintenance of ⌬Ym. However, Z-VAD.fmk was unable to protect cells against ⌬Ym loss and subsequent death induced by Bcl-XL antisense oligonucleotides plus ceramide, although DNA fragmentation was completely blocked.
These results can be explained in view of the current models. When cytochrome c is released into the cytosol, it can bind to Apaf-1, resulting in the activation of caspase 9 and subsequently caspase 3. It is now apparent that in some systems, cytochrome c release occurs prior to loss in ⌬Ym and that the subsequent activation of caspases is required for ⌬Ym loss, resulting in a circular amplification loop (Refs. 30 and 33-35; reviewed in Ref. 36). Thus, in this model, disruption of ⌬Ym is prevented by caspase inhibitors as was observed with Bcl-XL inhibition and Bcl-XL inhibition plus staurosporin.
Alternatively, in a number of other models, inhibition of caspases with Z-VAD.fmk fails to prevent the disruption of mitochondrial membrane function, despite the fact that Z-VAD.fmk does prevent the acquisition of the many biochemical hallmarks including DNA fragmentation (37)(38)(39)(40). Thus, in these systems caspase activation is downstream of mitochondria dysfunction, and furthermore, mitochondrial disruption is sufficient to cause cell death, through apoptosis or necrosis. This model would be consistent with our observations with Bcl-XL inhibition plus ceramide. Furthermore, a recent report demonstrated that GD3 ganglioside synthesis, which can be triggered by cell-permeating ceramides, resulted in disruption of ⌬Ym and induced apoptosis in a caspase-independent fashion (37).
Taken together, our data are consistent with endogenous Bcl-XL functioning by maintaining mitochondrial integrity of resting HUVEC, where loss of Bcl-XL results in greater susceptibility to mitochondrial dysfunction. The protection of ⌬Ym afforded by Bcl-XL functions in both caspase-dependent and caspase-independent mechanisms. These results demonstrate a definitive role for Bcl-XL in these processes and are complementary to overexpression data obtained previously.

FIG. 10. Inhibition of Bcl-XL results in procaspase 3 cleavage and increased caspase 3 activity that is inhibited by Z-VAD.fmk.
Cells were treated with Bcl-XL-specific or control antisense oligonucleotide (50 nM), and 24 h later were treated with staurosporin and/or Z-VAD.fmk. After an additional 24-h incubation, whole cell lysates were harvested, and protein concentrations were determined. A, procaspase 3 protein levels were determined by Western blot, loading equal amounts of protein in each lane. B, in a separate experiment, caspase 3 activity was determined using 50 M DEVD-p-nitroanilide substrate and recording OD 418.