Double-Stranded RNA Endothelial Cells to Apoptosis Induced by ...

The Journal of Immunology

IFN-␣ Sensitizes Human Umbilical Vein Endothelial Cells to Apoptosis Induced by Double-Stranded RNA1 William J. Kaiser, Jonathan L. Kaufman, and Margaret K. Offermann2 The ability of endothelial cells to mount an efficient antiviral response is important in restricting viral dissemination and eliminating viral infection from the endothelium and surrounding tissues. We demonstrate that dsRNA, a molecular signature of viral infection, induced apoptosis in HUVEC, and priming with IFN-␣ shortened the time between when dsRNA was encountered and when apoptosis was initiated. IFN-␣ priming induced higher levels of mRNA for dsRNA-activated protein kinase, 2ⴕ5ⴕ-oligoadenylate synthetase, and Toll-like receptor 3, transcripts that encode dsRNA-responsive proteins. dsRNA induced activation of dsRNA-activated protein kinase and nuclear translocation of transcription factors RelA and IFN regulatory factor-3 in IFN-␣primed HUVECs before the activation of intrinsic and extrinsic apoptotic pathways. These changes did not occur in the absence of dsRNA, and apoptosis resulting from incubation with dsRNA occurred much later when cells were not primed with IFN-␣. The entire population of IFN-␣-primed HUVECs underwent nuclear translocation of RelA and IFN regulatory factor-3 in response to dsRNA, whereas less than one-half of the population responded with apoptosis. When IFN-␣-primed HUVECs were coincubated with dsRNA and proteasome inhibitors, all HUVECs were rendered susceptible to dsRNA-induced apoptosis. These studies provide evidence that many endothelial cells that are alerted to the risk of infection by IFN-␣ would undergo apoptosis sooner in response to dsRNA than non-IFN-␣-primed cells, and this would enhance the likelihood of eliminating infected cells prior to the production of progeny virions. The Journal of Immunology, 2004, 172: 1699 –1710.

T

he innate immune system defends against microbial infection in part through the recognition of pathogen-associated molecular patterns, with dsRNA representing a molecular motif recognized during viral infections (1– 4). dsRNA activates multiple transcription factors, including IFN regulatory factor-3 (IRF-3)3 and NF-␬B (5– 8), and this affects the expression of many genes that are important in the innate antiviral response to viral infection. One component of the response to dsRNA is the secretion of type I IFN, thereby alerting neighboring cells to the risk of viral infection (8 –10). Type I IFNs, which include IFN-␣ and IFN-␤, induce expression of multiple genes that encode proteins that respond to dsRNA, including the dsRNA-activated protein kinase (PKR), 2⬘5⬘-oligoadenylate synthetase (2⬘5⬘-OAS), and Toll-like receptor 3 (TLR3) (10 –13). The higher expression of these genes enhances the likelihood of having a rapid response if cells encounter dsRNA as a consequence of viral infection. The vascular endothelium encounters a variety of different viruses that can infect endothelial cells (14 –18), but cellular and host defenses mount an effective response to most of these viruses, thereby preventing persistent infection of the vasculature. When the hemorrhagic virus Ebola infects endothelial cells, key compoWinship Cancer Institute, Emory University, Atlanta, GA 30322 Received for publication June 20, 2003. Accepted for publication November 5, 2003. 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. 1 This work was supported by U.S. Public Health Service Grants RO-1 CA79402 and P30 AR42687. J.L.K. is the recipient of the Edwin M. Crissey Fellowship in Hematology and Oncology at Emory University. 2 Address correspondence and reprint requests to Dr. Margaret K. Offermann, Winship Cancer Institute, Emory University, 1365-B Clifton Road NE, Atlanta, GA 30322. E-mail address: [email protected] 3 Abbreviations used in this paper: IRF, IFN regulatory factor; IAP, inhibitor of apoptosis; OAS, oligoadenylate synthetase; PKR, dsRNA-activated protein kinase; TLR, Toll-like receptor; c, cellular; BID, BH3 interacting domain death agonist; DEVD-AFC, N-acetylAsp-Glu-Val-Asp-7-amido-4-(trifluoromethyl)coumarin; zVAD-fmk, N-benzyloxycarbonyl-Val-Ala-Asp-(O-Me)fluoromethyl ketone.

Copyright © 2004 by The American Association of Immunologists, Inc.

nents of the innate antiviral machinery are blocked, leading to a loss of responsiveness to IFNs and to dsRNA (16, 19). This disruption of the innate antiviral response has tremendous biological consequences because endothelial cells are heavily infected with Ebola, yet the cells are unable to respond to the virus or signal to the immune system to generate an inflammatory response, and the infection is usually fatal (20). Some viruses enter a latent state in which only a small number of viral genes are expressed, including genes that restrict recognition and/or clearance by the innate antiviral machinery. Latent infection of endothelial cells with human herpesvirus 8 leads to Kaposi’s sarcoma (21, 22) and with CMV contributes to some cases of atherosclerosis and posttransplantation vascular disease (23–25). Endothelial cells are highly responsive to both synthetic and viral dsRNA, leading to a number of changes beneficial in the control of viral infection. For example, dsRNA induces the expression of the cellular adhesion molecules ICAM-1, VCAM-1, and E-selectin (26, 27), molecules that are important in adhesion and transendothelial migration of leukocytes that could help contain the viral infection (28, 29). dsRNA also induces expression of MHC I (19), a protein critical for the presentation of foreign Ags to CTL, contributing to the acquired immune response (30, 31). Endothelial cells also produce the cytokines IL-1␣, IL-1␤, IL-6, and type I IFN in response to dsRNA (26, 32, 33), thereby contributing to leukocyte maturation and alerting neighboring cells to the potential risk of viral infection. The dsRNA also reduces protein synthesis, most likely through the activation of the antiviral proteins PKR and 2⬘5⬘-OAS (26, 27). This restricts the expression of viral proteins, making it more difficult for the virus to make progeny virions. Some cells respond to viral infection with a form of programmed cell death called apoptosis. Apoptosis involves activation of the caspase family of proteases that dismantle the cell through the cleavage of specific proteins (34). For the host, apoptotic death of infected cells reduces viral spread in part through 0022-1767/04/$02.00

1700 preventing the production of progeny virions (35–38). Many different viruses encode proteins that block apoptosis, enhancing the ability of the virus to complete its life cycle (39 – 42). Viruses often disrupt the function of p53, a protein that is proapoptotic in a variety of cellular contexts (43– 46). Some viruses encode proteins that prevent activation of specific components of the apoptotic machinery. For example, human herpesvirus 8 encodes an inhibitor that blocks activation of caspase-8 (viral FLIP) and an inhibitor of that blocks activation of caspase-9 (viral Bcl-2) (47– 49). When apoptosis occurs before the expression of viral genes, production of progeny virions is prevented, restricting dissemination of the viral infection. Conversely, if apoptosis is delayed, then viral gene products are more likely to be expressed that can subvert the cellular response, enhancing the production of infectious virions and/or the establishment of persistent infection. Endothelial cells are more resistant than many types of cells to some agents that induce apoptosis (50, 51). Although endothelial cells are known to respond to dsRNA with multiple changes in gene expression and a reduction in protein synthesis (19, 26, 27), the role of apoptosis in the endothelial response to synthetic or viral dsRNA has not been reported. In this study, we demonstrate that multiple caspases are activated by dsRNA, and IFN-␣ priming dramatically accelerates their activation by dsRNA. IFN-␣ acts in part by inducing higher levels of dsRNA-responsive proteins that are in turn activated by dsRNA. All IFN-␣-primed cells respond to dsRNA with changes in transcription factors, whereas less than 50% of cells undergo apoptosis in response to dsRNA. The survival of a subpopulation of cells during prolonged incubation with dsRNA is eliminated by low concentrations of proteasome inhibitors, demonstrating that all HUVECs are sensitive to dsRNA-induced apoptosis when proteasome inhibitors are present, suggesting that survival is an acquired property.

dsRNA-INDUCED APOPTOSIS IN ENDOTHELIAL CELLS number of fragmented nuclei. At least 2500 total nuclei were counted for each condition. To assess the cellular localization of IRF-3 or the RelA subunit of NF-␬B, HUVECs were plated on collagen I-coated 8-well chamber slides (BD Biosciences, San Diego, CA). Following treatment, cells were fixed for 20 min in 4% paraformaldehyde in PBS and then incubated for 10 min in 0.2% Triton X-100 (Sigma-Aldrich) in PBS. Cells were blocked for 1 h in blocking buffer (PBS containing 3% BSA and 0.05% Tween 20) and then incubated for 2 h at room temperature with rabbit polyclonal Ab for IRF-3 (Active Motif, Carlsbad, CA) or RelA SC-109 (Santa Cruz Biotechnology, Santa Cruz, CA) diluted 1/100 in blocking buffer. Cells were washed several times with PBS and then incubated for 1 h at room temperature with Alexa Fluor 488 goat anti-rabbit Ab (Molecular Probes) diluted 1/800 in blocking buffer. Following multiple washes with PBS, cells were mounted with the ProLong Antifade kit (Molecular Probes).

RNA isolation and Northern blot analysis Total cellular RNA was isolated using RNA-Bee (Tel-Test, Friendswood, TX), according to the manufacturer’s instructions. Total cellular RNA was size fractionated using 1% agarose formaldehyde gels in the presence of 1 ␮g/ml ethidium bromide. The RNA was transferred to Hybond-N (Amersham Pharmacia Biotech) and covalently linked by UV irradiation using a Stratalinker UV cross-linker (Stratagene, La Jolla, CA). After a 1-h prehybridization in the absence of 32P-labeled DNA, the nylon was hybridized at 42oC overnight in ULTRAhyb (Ambion, Austin, TX) containing sheared salmon sperm DNA (Ambion) and ⬃100 ng of [32P]dCTP DNA (ICN Biomedicals, Aurora, OH) labeled at a sp. act. of ⱖ1 ⫻ 108 cpm/␮g. Blots were washed twice in 2⫻ SSC/1% SDS for 30 min at 55oC, then once in 0.2⫻ SSC/0.1% SDS for 30 min at 65oC. Autoradiography was performed with an intensifying screen at ⫺70oC. Blots were stripped using boiling water before rehybridization with subsequent probes. DNA probes were labeled using the random primer oligonucleotide method with the Prime It II kit (Stratagene). The PKR probe was a PCR fragment of the entire PKR open reading frame, the 2⬘-5⬘-OAS probe was the 1.2-kb EcoRI fragment of the human cDNA, the TLR3 probe was the 1.3-kb BamHI/XbaI fragment from IMAGE clone 4724599 (American Type Culture Collection, Manassas, VA), and GAPDH probe was a 1.2-kb EcoRI fragment of HHCMC32 (52).

Cell viability and FACS analysis of hypodiploid cells

Materials and Methods Cell culture HUVECs were obtained from the Emory Skin Diseases Research Core Center (Emory University). In this facility, HUVECs were prepared from umbilical veins, as previously described (19, 27). Cells were cultured in medium 199 supplemented with 20% FCS (HyClone Laboratories, Logan, UT), 16 U/ml heparin (ESI Pharmaceuticals, Cherry Hill, NJ), 50 ␮g/ml endothelial cell growth supplement (Biomedical Technologies, Stoughton, MA), 2 mM L-glutamine, 100 U/ml penicillin, and 100 U/ml streptomycin (Invitrogen, Carlsbad, CA), and grown at 37°C on tissue culture plates coated with 0.4% gelatin (Sigma-Aldrich, St. Louis, MO). Cells were passaged at confluency by splitting 1:4 and were used between passages 3 and 6. Where indicated, cells were treated with rIFN-␣ 2b (IFN-␣) from Schering-Plough (Kenilworth, NJ), IL-1␤ (Roche, Basel, Switzerland), TNF-␣ (Roche), TRAIL (Biomol, Plymouth Meeting, PA), Fas Ab CH-11 (Medical Biological Laboratories Naka-Ku, Nagoya, Japan), LPS from Salmonella abortus equi (Sigma-Aldrich), poly(I:C) (Amersham Pharmacia Biotech, Piscataway, NJ), benzyloxycarbonyl-Leu-Leu-leucinal (Calbiochem, La Jolla, CA), benzyloxycarbonyl-Ile-Glu(O-t-butyl-Ala-leucinal (Calbiochem), and N-benzyloxycarbonyl-Val-Ala-Asp-(O-Me) fluoromethyl ketone (zVAD-fmk) (Biomol).

Microphotography, quantitation of intact nuclei, and immunofluorescence Photography of HUVECs was performed with an Optronics Magnifire S99800 digital camera mounted on a Nikon Eclipse TE300 inverted microscope or on a Nikon Eclipse 800, both equipped with epifluorescence optics. For time-lapse photography, selected frames were extracted from a series compiled by Image Pro Plus software (Media Cybernetics, Silver Spring, MD). To identify fragmented nuclei, cells were washed with PBS and then incubated for 10 min in 70% ethanol containing 10 ␮g/ml Hoechst 33258 (Molecular Probes, Eugene, OR). To determine the number of intact nuclei, four random fields were photographed at ⫻40 magnification for each of triplicate samples prepared per experimental condition. The number of nuclei within each field was determined with Image Pro Plus software, and the reported values were adjusted by manually subtracting the

To determine the percentage of viable cells, adherent and nonadherent cells were collected and resuspended in PBS containing 0.04% trypan blue, and the percentage of cells that excluded trypan blue was directly counted with a hemacytometer. The percentage of hypodiploid cells was determined by collecting both adherent and nonadherent cells. Cells were fixed and permeabilized overnight in 70% ethanol, rinsed with PBS, and then incubated for 60 min at room temperature with 50 ␮g/ml propidium iodide in PBS, pH 7.3, 1 mg/ml RNase A, and 5 mM EDTA. Ten thousand events per sample were acquired on a BD Biosciences FACScan, and the percentage of cells in the subdiploid DNA content was measured on a histogram.

Immunoblot analysis For each sample, both adherent and nonadherent cells were collected. Cells were pelleted at 500 ⫻ g and then lysed at 4°C with agitation for 30 min in 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM Na2 EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na3VO4, and protease inhibitor mixture (BD PharMingen, San Diego, CA). For detection of phosphorylated proteins, lysis buffer was supplemented with Phosphatase Inhibitor Cocktail 1 and 2 (Sigma-Aldrich). The lysate was centrifuged at 12,000 ⫻ g for 5 min, and the supernatant fraction was collected. For cytosolic separation, digitonin fractionation was performed by resuspending the cell pellet in 0.025% digitonin (Calbiochem) lysis buffer (250 mM sucrose, 20 mM HEPES, pH 7.4, 5 mM MgCl2, 10 mM KCl, 1 mM EDTA, 1 mM EGTA, and protease inhibitor mixture). After 10-min incubation on ice, samples were centrifuged at 12,000 ⫻ g for 5 min, and the supernatant fraction containing cytosolic proteins was collected. Cellular extracts were combined with 3⫻ SDS loading buffer (187.5 mM Tris-HCl, pH 6.8, 6% SDS, 30% glycerol, 0.03% bromophenol blue, 125 mM DTT), size fractionated by SDS-PAGE on criterion gels (Bio-Rad, Hercules, CA), and then electrotransferred to Immobilon-P membranes (Millipore, Bedford, MA). Membranes were immunoblotted overnight with primary Abs that detect the following proteins (typically diluted 1/1000 in TBS-Tween with 5% milk): actin I-19, caspase-1 C-20, caspase-3 N-19, caspase-4 N-15, and caspase-9 H-83 (Santa Cruz Biotechnology); caspase-2 G310-1248, caspase-2 35, caspase-7 8-1-47, N-terminal DFF45, and cytochrome c (BD PharMingen); cleaved caspase-3 D175, caspase-6, caspase-8 1C12, caspase-9, eIF2␣,

The Journal of Immunology eIF2␣-phos-S51, PKR, and PKR-phos-T446 (Cell Signaling Tech, Beverly, MA); caspase-5 (StressGen Biotechnologies, Victoria, British Columbia, Canada); BH3 interacting domain death agonist (BID) P15 (BioSource International, Camarillo, CA); caspase-10 4C1 (Medical Biological Laboratories); and ␣-tubulin (Sigma-Aldrich). Anti-mouse IgG HRP (Bio-Rad), anti-rabbit IgG HRP (Bio-Rad), or anti-goat IgG HRP (Santa Cruz Biotechnology), diluted 1/10,000, were used as secondary Abs, and the immunoblots were visualized with the ECL system (Amersham Pharmacia Biotech).

Fluorescent enzymatic assay for caspase activity Adherent and nonadherent HUVECs were collected and pelleted at 500 ⫻ g. Cells were washed with PBS, pH 7.4, resuspended in lysis buffer (50 mM HEPES, pH 7.4, 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-1propanesulfonate, 1 mM DTT, 0.1 mM EDTA, and 0.1% Triton X-100), and centrifuged at 12,000 ⫻ g for 5 min. A fraction of the supernatant was diluted in assay buffer (50 mM HEPES, pH 7.4, 100 mM NaCl, 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 10 mM EDTA, and 10% glycerol) and allowed to equilibrate for 10 min. AcetylN-acetyl-Asp-Glu-Val-Asp-7-amido-4-(trifluoromethyl) coumarin (DEVDAFC) (Biomol) was added to a final concentration of 200 ␮M. A Bio-Tek

1701 Instruments (Winooski, VT) FL600 Microplate Fluorescence Reader equipped with a 400 nm excitation wavelength and a 508 nm emission wavelength filter set measured the relative fluorescence.

Results

IFN-␣ priming accelerates dsRNA-induced apoptosis Changes in cell number reflect the balance of cell proliferation and cell death. To investigate the effect of synthetic dsRNA poly(I:C) on endothelial cell proliferation and cell death, we incubated HUVECs in the presence of dsRNA for 4, 8, 12, or 24 h and then calculated the number of adherent cells with intact nuclei at each of these time points. Because IFN-␣ pretreatment modulates some responses to dsRNA in HUVECs (26), we compared the response to dsRNA in cells that were pretreated with IFN-␣ with non-IFN␣-pretreated cells. Under control conditions, the number of adherent cells increased during the first 12 h, and then cell numbers stabilized between 12 and 24 h as the cellular monolayer became

FIGURE 1. IFN-␣ pretreatment accelerates the apoptosis that results from incubation with dsRNA. HUVEC were incubated with standard medium or medium supplemented with 1000 U/ml IFN-␣ for 24 h (IFN-␣-pretreated). At time 0, the medium was replaced with nonsupplemented medium (control) or medium supplemented with 100 ␮g/ml poly(I:C) (dsRNA). A, Changes in cell number. The number of intact nuclei per ⫻100 microscopic fields was counted as described. Each data point represents the mean ⫾ SD for three separate samples that is presented relative to the number of cells in the 4-h control. B, Morphologic changes induced by dsRNA. Bright-field microscopic analysis at ⫻100 was done comparing cells that were not pretreated with IFN-␣ with those that were pretreated. The photography of cells incubated for 24 h was done after removal of nonadherent cells and rinsing with PBS (panels 1– 4). For the 4-h time point, the medium containing nonadherent cells was present for the bright-field examination (panels 5 and 6). Fluorescence microcopy using Hoechst 33258 to label the DNA was done after 4-h stimulation to assess the nuclear fragmentation (panels 7 and 8). C, Time-lapse photography. HUVECs were incubated with medium supplemented with 1000 U/ml IFN-␣ for 24 h. Fresh medium containing dsRNA was then added, and cells were photographed at the indicated times that were at 3-min intervals to demonstrate morphologic changes of apoptosis. Arrows in the first panel demarcate the location of the process. D, Quantitation of the percentage of apoptotic cells. IFN-␣-pretreated and nonpretreated cells were incubated in the absence (control) or presence of dsRNA for 8 h. Cells were fixed and then DNA stained with propidium iodide, and the DNA content was assessed by flow cytometry, as described. Each condition was done in triplicate. The mean percentage of hypodiploid events ⫾ SD for each condition is shown.

1702

dsRNA-INDUCED APOPTOSIS IN ENDOTHELIAL CELLS

more densely packed (Fig. 1A). HUVECs pretreated for 24 h withIFN-␣ without subsequent dsRNA stimulation were similar in number to control cells at all time points assessed. Incubation of HUVECs with dsRNA reduced the number of viable cells, but this reduction was not apparent until after 12 h in cells that were not pretreated with IFN-␣. There was a 20% reduction in cell number within 4 h of incubation with dsRNA in cells that were pretreated with IFN-␣, and the cell number reached a plateau by 8 –12 h without any further decrease. Thus, pretreatment with IFN-␣ dramatically shortened the time frame for dsRNA-induced cell death, but then cell numbers stabilized. Treatment of HUVECs with dsRNA induced a spindle-shaped appearance that was distinct from the cobblestone appearance of control cells or cells incubated with IFN-␣ (Fig. 1B). The morphology of dying HUVECs resembled apoptotic cell death and included extensive membrane blebbing, nuclear fragmentation, loss of contact with neighboring cells, and ultimately detachment/ exclusion from the endothelial cell monolayer (Fig. 1B, panel 6). Multiple nonadherent cells were seen within 4 h of incubation with dsRNA in cells that were pretreated with IFN-␣, but not in cells that were not pretreated (Fig. 1B, panels 5 and 6). Visualization of nuclear fragmentation with a bisbenzimide dye (Hoechst 33258) to stain cellular DNA further confirmed that a higher percentage of IFN-␣-pretreated cells vs nonpretreated cells was undergoing apoptosis in response to dsRNA at 4 h (Fig. 1B, panels 7 and 8). When the nonviable cells were removed after cells were incubated with dsRNA for 24 h, there was a relatively uniform distribution of surviving cells, but the cell density was less in the IFN-␣-pretreated cells compared with nonpretreated cells (Fig. 1B, panels 3 and 4). Morphologic changes characteristic of apoptosis were seen as early as 2 h after the addition of dsRNA and peaked by 4 h in cells that had been pretreated with IFN-␣, whereas they occurred much later (12–18 h) if cells were not incubated with IFN-␣ before dsRNA (data not shown). The increase in apoptosis was dependent on dsRNA and did not occur in response to IFN-␣ in the absence of dsRNA. Time-lapse photography taken at 3-min intervals beginning at 2 h and 20 min following the addition of dsRNA to IFN-␣-pretreated cells demonstrated that once morphologic changes were detectable within an individual cell, the process proceeded rapidly and was completed in less than 20 min (Fig. 1C). The decline in the number of intact nuclei per field plateaued at roughly 8 h in IFN-␣-primed cells treated with dsRNA, whereas the decline was just beginning at 8 h for nonprimed cells incubated with dsRNA (Fig. 1A). Thus, the 8-h time point was selected to illustrate changes in the percentage of apoptotic cells. Under control conditions, ⬃6% of cells displayed hypodiploid DNA content, and incubation with dsRNA for 8 h increased this percentage by ⬃2-fold (Fig. 1D). Pretreatment with IFN-␣ without subsequent dsRNA did not alter the percentage of apoptotic cells relative to the control, whereas cells that were primed with IFN-␣ and then treated with dsRNA displayed a 6-fold increase in the percentage of cells with hypodiploid DNA relative to control, with nearly 35% of cells hypodiploid.

TRAIL (57, 58). We focused on changes in expression of genes that enable the cell to detect dsRNA and found that incubation with IFN-␣ for 24 h induced higher levels of PKR, TLR3, and 2⬘5⬘OAS (Fig. 2A). We hypothesized that the higher levels of expression resulting from IFN-␣ priming would not be accompanied by greater activation of these proteins until cells encountered dsRNA. This was directly examined for PKR by examining the relationship between total PKR protein levels and its activation state. The activation state of PKR was determined both by examining phosphorylation of its substrate, eIF2␣, and by using an Ab that recognizes PKR phosphorylated at threonine 446 (T446), a residue located in the activation loop of the kinase domain of PKR (59). IFN-␣-primed cells had much higher protein levels for PKR than nonprimed cells, yet there was no increase in the phosphorylation of PKR at threonine 446 (T446) nor in the phosphorylation of its substrate eIF2␣ in the absence of dsRNA (Fig. 2B). Incubation of IFN-␣-primed cells with dsRNA induced a large increase in the phosphorylation of both PKR and eIF2␣ within 15 min, demonstrating that dsRNA-induced activation was very rapid. Cells that were not primed with IFN-␣ had much lower levels of PKR than the IFN-␣-primed cells, and incubation of nonprimed cells with dsRNA had minimal effect on the phosphorylation of PKR and its substrate eIF2␣ in this time frame. Thus, incubation with IFN-␣ increased the total amount of PKR, yet activation did not occur until cells were incubated with dsRNA. We did not directly measure 2⬘5⬘-OAS activation in the HUVECs, but its activation most likely followed a similar pattern to PKR because high levels of expression were induced by IFN-␣ (Fig. 2A), and it resembles PKR in its activation by binding dsRNA (60, 61).

IFN-␣ induces the expression of multiple genes that respond to dsRNA

FIGURE 2. Priming with IFN-␣ induces the expression of dsRNA-sensing gene products. A, Total cellular RNA was isolated from HUVECs incubated in the absence or presence of IFN-␣ for 24 h. The RNA (20 ␮g/lane) was size fractionated and analyzed for the indicated genes by Northern blot analysis. B, HUVEC were pretreated for 24 h in the absence or presence of IFN-␣ (1000 U/ml). At time 0, the medium was changed, and where indicated, dsRNA was added at 100 ␮g/ml. Cell extracts were prepared at either time 0 or 15 min after the addition of dsRNA, size fractionated, and analyzed by Western blot using Abs that specifically detected the phosphorylated forms of PKR and eIF2␣ and with Abs that detected total PKR and total eIF2␣.

We demonstrated that priming with IFN-␣ did not directly induce apoptosis, yet IFN-␣ sensitized HUVECs to dsRNA-induced apoptosis. IFN-␣ induces expression of multiple genes that could modulate the ability of dsRNA to induce apoptosis, including genes that encode proteins that are directly activated by dsRNA such as PKR and 2⬘5⬘-OAS (53–55), proteins that are critical for dsRNA-signaling pathways such as TLR3 (12, 13, 56), and proteins that directly participate in the apoptotic cascade such as

Pretreatment with IFN-␣ alters the kinetics and magnitude of caspase activation induced by dsRNA Apoptotic cell death is coordinated by the activity of members of the caspase family of cysteine proteases (34, 62, 63). Caspases characteristically cleave after aspartate residues, and each member has distinct substrate specificity (64, 65). To assess the activation

The Journal of Immunology state of caspases in HUVECs challenged with dsRNA as a function of time, we used the fluorogenic substrate optimal for effector caspases-3 and -7, acetyl-DEVD-AFC (65, 66). For these assays, the fluorescence of the substrate is quenched before cleavage, and an increase in fluorescence occurs in response to substrate cleavage. IFN-␣ priming alone did not alter the amount of caspase activity relative to that detected in control cell extracts, but IFN-␣ priming dramatically enhanced the amount of caspase activity that resulted from incubation with dsRNA (Fig. 3A). The caspase activity measured in response to dsRNA in the absence of IFN-␣ was less than that which occurred when cells were pretreated with IFN-␣. Caspase activity in IFN-␣-primed cells increased ⬎16-fold following 4 h of dsRNA stimulation compared with only a 3- to 4-fold increase in dsRNA-treated cells that had not been primed with IFN-␣. The robust level of caspase activity detected at the 4-h time point in response to dsRNA in the IFN-␣-pretreated cells progressively decreased when incubation with dsRNA was continued for 8 or 12 h as the number of cells undergoing apoptosis declined. The effect of varying the concentration of dsRNA was examined during a 4-h incubation with IFN-␣-pretreated HUVEC. There was some reduction in the amount of caspase activity if concentrations of dsRNA were reduced from 100 to 10 ␮g/ml or less, yet the level of activity was nonetheless 7-fold above control using 0.4 ␮g/ml dsRNA (Fig. 3B). Thus, dsRNA activated caspases when there was a 250-fold reduction in the concentration of dsRNA. Extracellular signals such as Fas ligand, TNF-␣, and TRAIL initiate apoptosis by binding to cell surface receptors and use Fasassociated death domain protein as an adaptor protein that can lead to the activation of caspase-8 in the appropriate cellular context (34). Success at initiating apoptosis through the extrinsic pathway is affected by multiple factors, including the levels of expression of the components of the cascade and the presence of proteins that block caspase activation. If IFN-␣ directly affects one or more of the components that are shared by the various death receptor pathways, then apoptosis in response to multiple agents that engage death receptors is likely to be enhanced by pretreatment with IFN-␣. In contrast, if priming with IFN-␣ exclusively enhances dsRNA-induced apoptosis, this would point to a mechanism that is more specific to dsRNA. We found that priming with IFN-␣ primarily sensitized HUVECs to apoptosis induced by dsRNA (Fig. 3C). Cells that were primed with IFN-␣ responded to dsRNA with a 16-fold increase in DEVDase activity, whereas the increase was only 3-fold in response to dsRNA of non-IFN-␣-primed cells. Priming with IFN-␣ did not enhance apoptosis induced by TRAIL or by Fas Ab, and it enhanced apoptosis by 2-fold or less in response to TNF-␣ and LPS, and hence was minor compared with the enhancement that occurred in response to dsRNA. This suggests that the mechanism by which IFN-␣ priming enhanced dsRNA-induced apoptosis was unlikely to be through alteration in a shared component of a death receptor cascade. Caspases are synthesized as precursor molecules that undergo proteolytic activation (34, 62). An increase in cleavage of DEVDAFC by cellular extracts is generally indicative of effector caspase activation, and this was confirmed by detection of caspase-3 maturation by immunoblot analysis (Fig. 3D). The caspase-3 proform was detected in all samples, but caspase-3 processing exclusively occurred following dsRNA stimulation. Peak caspase-3 processing was detected 4 h following dsRNA treatment of IFN-␣-primed cells and declined at later time points, consistent with changes in enzymatic activity detected in cellular extracts shown in Fig. 3A. Incubation with dsRNA in cells that were not pretreated with IFN-␣ also induced the appearance of caspase-3 cleavage products, but the appearance of these was much less and was delayed.

1703

FIGURE 3. dsRNA induces the activation of caspases in endothelial cells. HUVECs were incubated with standard medium or medium supplemented with 1000 U/ml IFN-␣ for 24 h (IFN-␣-pretreated). At time 0, the medium was replaced with nonsupplemented medium (control) or medium supplemented with poly(I:C) (dsRNA). The dsRNA was at 100 ␮g/ml in A, C, and D, and at the indicated concentration in B. For each assay, identical amounts of protein were used to assess the ability of cell extracts to cleave acetyl-DEVD-AFC, as described in Materials and Methods. The mean ⫾ SD from triplicate samples was determined for each condition in A–C. The fold increase was determined by comparison with the 4-h control in A–C. C, The ability of IFN-␣ pretreatment to enhance caspase activation by dsRNA (100 ␮g/ml), LPS (1 ␮g/ml), TNF-␣ (50 ng/ml), TRAIL (1 ␮g/ml), and Fas Ab (100 ng/ml) was compared in extracts prepared 4 h following stimulation. D, Conditions identical with those assayed in A were assessed by Western blot using an Ab that detects unprocessed caspase-3 (procasp-3) in combination with an Ab that detects cleaved forms of caspase-3 (cp). The size of the proteins was determined by comparison with m.w. markers (data not shown).

1704


The dsRNA-induced cleavage of caspase-3 in both the nonprimed cells and the IFN-␣-primed cells coincided with the onset of cell death that was observed in Fig. 1A. Multiple caspases are activated by dsRNA Current models posit that a hierarchy of caspase activation regulates apoptosis, with the diversity of apoptotic signals first activating large prodomain initiator caspases that subsequently target short prodomain effector or executioner caspases, such as caspase-3 (34, 62, 63). To identify caspases that might participate in dsRNA-induced apoptosis, we examined immunoblots for evidence of processing of multiple caspases. HUVECs expressed zymogen forms of caspases 1–10 (Fig. 4). There was no evidence of proteolytic processing of any of the caspases in IFN-␣-primed cells in the absence of incubation with dsRNA, whereas incubation of IFN-␣-pretreated cells with dsRNA for 4 h triggered proteolytic processing of caspases-2, -3, -6, -7, -8, -9, and -10. Peak levels of cleaved forms of these caspases and/or a reduction in the amount of proforms were seen at 4 h, with much less seen in extracts prepared after incubation with dsRNA for 8 and 12 h. The cleavage of these multiple caspases was accompanied by the cleavage of DFF45 and its isoform DFF35 (DFF45/35), classical effector caspase substrates (67, 68), confirming that caspases were not only processed, but were also actively cleaving downstream substrates. Not all caspases that were expressed in IFN-␣-pretreated HUVECs underwent proteolytic processing in response to incubation with dsRNA. Caspases-1, -4, and -5 are best known for their involvement in proinflammatory cytokine processing (69 –71). Caspases-1, -4, and -5 were expressed in endothelial cells, but unlike the other caspases examined, dsRNA did not induce the cleavage and/or proform disappearance, indicating that these caspases were not likely involved in dsRNA-induced apoptosis. Timing of dsRNA-induced caspase activation in IFN-␣-pretreated HUVECs Because all of the caspases cleaved during dsRNA-induced apoptosis were processed within 4 h, we monitored earlier time points to more precisely define the onset of caspase activation. We first measured the level of DEVDase activity in cellular extracts prepared hourly following the addition of dsRNA. dsRNA induced a 4-fold increase in caspase activity at 2 h, a 22-fold increase at 3 h, and a 16-fold at 4 h (Fig. 5A), indicating that peak effector caspase activity occurred between 2 and 4 h following the addition of dsRNA to IFN-␣-primed HUVECs. Extracellular signals induce apoptosis by initiating the formation of a death receptor complex that leads to caspase-8 activation, whereas intrinsic intracellular events trigger release of cytochrome c from mitochondria, leading to the formation of the apoptosome and activation of caspase-9 (34). Both apical caspases converge on the cleavage of caspase-3, and activated caspase-3 cleaves downstream substrates, including DFF45/35. We examined the onset of cleavage of caspases-3, -8, and -9, as well as DFF45/35 cleavage and the redistribution of cytochrome c to the cytosol to address the role of extrinsic vs intrinsic pathways in the apoptosis resulting from dsRNA in IFN-␣-primed HUVECs. Caspase-8 underwent cleavage into intermediate p43/p41 forms within 2 h, whereas the p18 subunit of fully matured caspase-8 was first detected at 2.5 h (Fig. 5B). The cleavage of caspase-3 and its substrates DFF45/35 were also detected 2 h following incubation with dsRNA and continued to increase at subsequent time points (Fig. 5B). The processing of caspase-9 occurred more slowly than the processing of caspases-3 and -8, but the release of cytochrome c from the mitochondria occurred within 2 h, suggesting that the apoptosome might be formed within this time frame. These data

FIGURE 4. Identification of caspases proteolytically processed during dsRNA-induced apoptosis in IFN-␣-pretreated cells. Cell extracts were prepared from cells that were pretreated with IFN-␣ for 24 h before incubation under control conditions or with dsRNA (100 ␮g/ml). Western blot analysis was performed using Abs for the indicated proteins. The mobility of caspases and their cleavage products was in agreement with their known sizes based on comparison with m.w. markers. The unprocessed form is indicated as the proform, and cleavage products are indicated as cp or as the fragment of the indicated m.w. Nonspecific cross-reactivity is indicated as ⴱ. The multiple isoforms of procaspase-10 and its cleavage products are indicated. Extracts were also assessed for cleavage of the caspase substrates DFF45/35 and for actin.

indicate that both intrinsic and extrinsic apoptotic pathways responded to dsRNA within 2 h in IFN-␣-primed HUVECs, and the changes in the apical caspases were accompanied by rapid activation of effector caspases, leading to the cleavage of the substrates DFF45/35.


1705 when caspase-8 cleaves the proapoptotic Bcl-2 family member BID, thereby recruiting the mitochondrial pathway by inducing the release of cytochrome c from the mitochondria (73, 74). Interestingly, cytochrome c release from the mitochondria and caspase-9 processing were less sensitive to inhibition by zVAD-fmk than the other caspases examined (Fig. 6A). BID was expressed in IFN-␣primed HUVECs, but dsRNA did not induce much cleavage of BID, suggesting it was not responsible for the release of cytochrome c from the mitochondria. Furthermore, zVAD-fmk blocked dsRNA-induced maturation of caspase-8, yet cytochrome c release from the mitochondria was not blocked. The zVAD-fmk also blocked dsRNA-induced processing of caspases-2 and -3, suggesting that they were also not required for dsRNA-induced release of cytochrome c. All IFN-␣-primed HUVECs respond to dsRNA with nuclear translocation of RelA and IRF-3, yet many cells survive by a process that is sensitive to proteasome inhibitors

FIGURE 5. Kinetics of dsRNA induced processing of caspases in IFN␣-pretreated HUVECs. HUVEC were either not pretreated or were pretreated with IFN-␣ (1000 U/ml) for 24 h. Cells were incubated in the absence (⫺) or presence (⫹) of dsRNA (100 ␮g/ml poly(I:C)) for the indicated times. A, Identical amounts of protein were used to assess the cleavage of acetyl-DEVD-AFC by caspases present in the cell extracts, as described in Materials and Methods. The mean ⫾ SD from triplicate samples was determined for each condition, and the fold increase was determined by comparison with the 0-h control. B, Whole cell extracts were used for Western blot analysis of all proteins, except cytochrome c and ␣-tubulin, which were assessed using cytosolic extracts isolated by digitonin fractionation, as described.

The mitochondrial release of cytochrome c in response to dsRNA in IFN-␣-primed cells is not prevented by zVAD-fmk In Fig. 5B, we demonstrated that incubating IFN-␣-primed endothelial cells with dsRNA led to the processing of caspases-8 and -9, suggesting that both death receptor and mitochondrial pathways were activated during dsRNA-induced apoptosis. We used the broad-spectrum caspase inhibitor zVAD-fmk to examine potential cross talk between the extrinsic and intrinsic apoptotic pathways (72). Stimulation of IFN-␣-primed HUVECs with dsRNA in the presence of zVAD-fmk led to a dose-dependent reduction in the appearance of cleaved forms of caspases-2, -3, and -8 (Fig. 6A), and this resulted in cell survival with no morphologic evidence of cell death (Fig. 6B). Extrinsic apoptotic signals can be amplified

In Fig. 1, we demonstrated that ⬃60% of IFN-␣-primed HUVEC survived during prolonged incubation with dsRNA. This raised the possibility that a subpopulation of HUVECs might be unresponsive to any dsRNA signaling. To test this possibility, we analyzed dsRNA-induced nuclear translocation of the transcription factors IRF-3 and RelA. When IFN-␣-pretreated HUVECs were incubated with dsRNA, the entire population of cells responded with nuclear accumulation of both IRF-3 and RelA, yet only a subset of the cells displayed processing of caspase-3 (Fig. 7A). Thus, the survival of some cells during incubation with dsRNA was not due to an overall lack of responsiveness to dsRNA. All IFN-␣-pretreated HUVECs responded to dsRNA with nuclear accumulation of the transcription factor IRF-3 and RelA, yet only a subset underwent apoptosis. Proteasomes play an important role in regulating many cellular processes that are induced by dsRNA, including gene transcription, cell cycle regulation, protein translation, and apoptosis (75–79). To determine whether the proteasome plays a role in the sensitivity of IFN-␣-pretreated HUVECs to dsRNA-induced apoptosis, we stimulated IFN-␣primed endothelial cells with dsRNA in the presence of either benzyloxycarbonyl-Leu-Leu-leucinal or benzyloxycarbonyl-IleGlu-(O-t-butyl)-Ala-leucinal, two distinct cell-permeable peptide aldehydes that inhibit the chymotryptic activity of the proteasome. Incubation of endothelial cells with dsRNA in the presence of either inhibitor led to the death of almost all cells within 8 h compared with the death of only ⬃45% when cells were incubated with dsRNA in the absence of proteasome inhibitor (Fig. 7B). Minimal cell death resulted from incubation with proteasome inhibitors in the absence of dsRNA, indicating that the proteasome inhibitors were not directly toxic to cells at the concentrations used. These studies were done using low concentrations of proteasome inhibitors that rendered all cells sensitive to dsRNA-induced cell death, but did not fully block nuclear accumulation of RelA induced by dsRNA (data not shown). These studies demonstrate that all endothelial cells can be rendered sensitive to dsRNA-induced apoptosis by proteasome inhibitors, and suggest that degradation of some protein(s) by the proteasome contributes to the survival of a significant percentage of cells during prolonged incubation with dsRNA.

Discussion Coordinating a specific and measured defense against microbial infection depends in part upon the innate immune system’s ability to recognize pathogen-associated molecular patterns, with dsRNA representing a molecular pattern common to most viral infections (11, 35, 80). Although dsRNA has been shown to induce apoptosis

1706


FIGURE 6. Caspase inhibition with zVAD-fmk blocks dsRNA-induced apoptosis, but not cytochrome c release from the mitochondria. IFN-␣-pretreated HUVECs were incubated in the absence or presence of dsRNA for 4 h. When present, zVADfmk was added at the indicated concentrations simultaneously with dsRNA. A, Both whole cell and cytosolic extracts were prepared and analyzed by Western blot with Abs to the indicated proteins and cleavage products (cp). Cytosolic extracts were used for assessment of cytochrome c and ␣-tubulin, and whole cell extracts were used for all other proteins. B, The morphology of the IFN-␣-pretreated cells incubated with the indicated additives (zVAD-fmk at 50 ␮M).

in some cell types, this study is the first to demonstrate that dsRNA induces apoptosis in endothelial cells. The sensitivity of these cells to apoptosis induced by dsRNA is in contrast to the relative resistance of endothelial cells to apoptosis induced by a variety of other agents. For example, TNF-␣, TRAIL, and LPS induce only modest levels of apoptosis in endothelial cell unless protein synthesis is inhibited (50, 81). Virally infected cells often secrete type I IFN (9, 11), and this would prepare endothelial cells for a rapid response if they become infected and/or encounter dsRNA. Apoptosis should occur more rapidly in response to viral dsRNA in IFN-␣primed cells than in naive cells, thereby killing many virally infected cells before the production of progeny virions. Our studies also suggest that some virally infected cells are likely to survive and express multiple dsRNA-induced genes, including cellular adhesion molecules, cytokines, and MHC I (19, 26, 27). The dsRNAinduced changes in the surviving cells could either contribute to an effective immune response that would eliminate the virus from the host, or if the host response is unsuccessful at eradicating the virus, a state of chronic inflammation could result. The ability of IFN-␣ to selectively sensitize endothelial cells to dsRNA-induced apoptosis is consistent with the ability of IFN-␣ to induce higher levels of expression of PKR, 2⬘5⬘-OAS, and TLR3, antiviral genes that are activated by dsRNA. The relationship between changes in protein levels and changes in activation state is

illustrated by changes in PKR. IFN-␣ induced high levels of PKR expression in HUVECs, yet there was very little enzymatic activity until cells were incubated with dsRNA. Activation of PKR occurred within 15 min of exposure to dsRNA, leading to autophosphorylation and an increase in the phosphorylation of its substrate eIF2␣. Although we did not directly measure changes in activation of 2⬘5⬘-OAS, we suspect that its activation accompanied the activation of PKR because the IFN-␣-primed cells expressed high levels of mRNA for 2⬘5⬘-OAS, and 2⬘5⬘-OAS is activated by a similar mechanism to PKR through the direct binding of dsRNA (82, 83). Less is known about how dsRNA activates TLR3, but recent studies provide evidence that TLR3 is required for dsRNA-induced activation of NF-␬B and of IRF-3 (56, 84), events that occurred within 30 min of incubation of IFN-␣-primed HUVECs with dsRNA and before the initiation of apoptosis. To characterize the core components of the apoptotic machinery involved in dsRNA-induced apoptosis, we examined 10 of the 11 known human caspases by immunoblot analysis and found evidence for the participation of 7 of the 10 caspases that were examined. We detected the processing and/or proform disappearance of the apical caspases-2, -8, -9, and -10 and of the effector caspases-3, -6, and -7. We were unable to detect any significant lag between the detection of processed forms of initiator caspases and the effector caspases or the cleavage of the caspase substrate


FIGURE 7. All IFN-␣-pretreated HUVECs respond to dsRNA with nuclear translocation of RelA and IRF-3 and are sensitive to dsRNA-induced cell death in the presence of proteasome inhibitors. A, HUVECs were pretreated with IFN-␣ for 24 h. At time 0, the medium was changed to medium lacking dsRNA (control) or medium containing poly(I:C) at 100 ␮g/ml (dsRNA). Cells were fixed after 30 min and analyzed for nuclear translocation of RelA or IRF-3 by immunofluorescent microscopy, whereas cells were fixed at 4 h for immunohistochemistry of cleaved caspase-3 with simultaneous labeling of nuclei with 4⬘,6⬘-diamidino-2-phenylindole. The arrow indicates the cell within the field that was immunoreactive for cleaved caspase-3. B, HUVECs were pretreated with IFN-␣ for 24 h. At time 0, the medium was changed to medium supplemented with the indicated additive. Cell viability was determined by trypan blue exclusion. Data represent the mean percentage of viable cells at 8 h ⫾ SD from triplicate determinations.

DFF45/35, indicating that the progression from activation of apical caspases to activation of effector caspases and substrate cleavage occurred rapidly. Consistent with this finding, the time interval between the first evidence of membrane changes and the extensive membrane blebbing within an individual cell was less than 20 min. This is consistent with recent reports documenting the rapidity of the apoptotic process (85). The peak cleavage of caspase-9 lagged behind the peak cleavage of caspases-8 and -3, yet cytosolic cytochrome c was detected in the same time frame as cleaved forms of caspase-8 and -3, indicating that the integrity of the outer mitochondrial membrane was disrupted, a process that is regulated by Bcl-2 family members (86, 87). Cytosolic cytochrome c contributes to the formation of the apoptosome, and caspase-9 does not require proteolytic processing to be enzymatically active (64, 88).

1707 The cleavage of BID by caspase-8 enables the death receptor pathway to recruit the mitochondrial apoptotic machinery in some cells, yet very little cleavage of BID occurred in response to incubation of IFN-␣-primed HUVECs with dsRNA. Furthermore, the dsRNA-induced redistribution of cytochrome c to the cytosol occurred when the broad-spectrum inhibitor of caspases, zVADfmk, blocked processing of caspase-8. The dsRNA-induced cleavage of caspases-2, -3, and -8 were all inhibited by zVAD-fmk, whereas release of cytochrome c and cleavage of caspase-9 were not prevented. This suggests that the intrinsic apoptotic pathway can be directly activated by dsRNA in IFN-␣-primed HUVECs. Studies are underway to determine whether apoptosis is exclusively initiated through a single apical caspase or whether more than one is directly activated by dsRNA in HUVECs. The mechanism by which dsRNA activates caspases and the apoptotic cascade is unknown but may be linked to changes in protein synthesis that result from activation of PKR and 2⬘5⬘-OAS. The inhibition of protein synthesis that results from activation of either PKR or 2⬘5⬘-OAS has been shown to be important for dsRNA-induced apoptosis in some cells (1, 89, 90). The phosphorylated eIF2␣ that increases in response to activated PKR decreases the initiation of protein synthesis by interfering with the GDP-GTP exchange required for a new round of initiation, but the magnitude of the inhibition varies in different cells, in part due to cell-type differences in the amount of eIF2B, the exchange factor for GDP (91–93). Activation of 2⬘5⬘-OAS by dsRNA interferes with protein synthesis by catalyzing the formation of 2⬘5⬘-oligoadenylates, and the oligoadenylates in turn activate RNase L, reducing protein synthesis by the destruction of some RNA. Cycloheximide is important for apoptosis induced by TNF-␣ and LPS in endothelial cells because it inhibits the synthesis of the antiapoptotic proteins cFLIP and Mcl-1 (50). In contrast, cycloheximide is not required for dsRNA-induced apoptosis in HUVECs, suggesting that the reduction in protein synthesis that occurs when IFN-␣-pretreated HUVECs are incubated with dsRNA might be sufficient to contribute to the process even though the reduction is much less than occurs in response to cycloheximide (26). TLR3 represents an additional dsRNA-responsive protein that might be involved. Studies are underway to address the specific roles of these candidate proteins in the observed responses to dsRNA in HUVECs. The majority of the apoptosis occurred within 2– 4 h of incubation with dsRNA in IFN-␣-primed HUVECs, whereas a smaller percentage of cells died between 8 and 12 h. Longer incubation of IFN-␣-primed HUVECs with dsRNA was not accompanied by any further decline in cell number. All HUVECs, including those that did not undergo apoptosis in response to dsRNA, responded to dsRNA with nuclear translocation of both RelA and IRF-3. We have previously shown that IFN-␣-primed HUVECs express ICAM-1, VCAM-1, IL-6, and other immunomodulatory genes when examined 24 h after incubation with dsRNA (26). Thus, while apoptosis is generally regarded as a mechanism for cell death that does not elicit an inflammatory response, the surviving endothelial cells express multiple proinflammatory genes (26, 27). This suggests that the endothelial cells that do not undergo apoptosis can recruit leukocytes to generate an immunologic response to viral infection. The survival of cells during prolonged incubation with dsRNA was eliminated when low concentrations of proteasome inhibitors were present. The sensitization of HUVECs to dsRNA-induced cell death by proteasome inhibitors is distinct from the protection from apoptosis that occurs when proteasome inhibitors are present during incubation with some other proapoptotic signals. For example, the proteasome inhibitor lactacystin prevents apoptosis in HUVECs in response to LPS and cycloheximide by preventing the

1708 degradation of the antiapoptotic proteins cellular (c)-FLIP and Mcl-1 (50). The endothelial cells used in our experiments were proliferating, and others have shown proteasome inhibitors are more toxic to proliferating endothelial cells than to quiescent cells (75). We thus used low concentrations of proteasome inhibitors that sensitized the entire population to dsRNA-induced cell death, but did not fully block activation of NF-␬B, demonstrating that the function of the proteasome was not fully disrupted. Possible mechanisms for the increased dsRNA-induced cell death include reduced expression of antiapoptotic genes and/or an inability to terminate proapoptotic signals. A reduction of NF-␬B activation might reduce the expression of NF-␬B-inducible antiapoptotic genes such as the inhibitors of apoptosis (IAPs) and c-FLIP (94, 95). Inhibition of the proteasome could also affect the function of IAPs because multiple IAPs are E3 ubiquitin-protein isopeptide ligases that are important in the ubiquination and proteasomal degradation of active caspases and of the proapoptotic protein Smac/ DIABLO (96 –98). These examples illustrate a few of the possible mechanisms that are being explored for how the activity of the proteasome might confer resistance to dsRNA-induced apoptosis. Endothelial cells that were not primed with IFN-␣ were much slower to respond to dsRNA with apoptosis than IFN-␣-primed cells and displayed no apparent decrease in cell number until cells were incubated with dsRNA for ⬎12 h. We have previously examined changes in gene expression in response to dsRNA in nonprimed HUVECs and have found that many genes are expressed at high levels within 4 h of incubation with dsRNA, including ICAM-1, VCAM-1, E-selectin, PKR, and 2⬘5⬘-OAS (26, 27). In addition, dsRNA induces secretion of type I IFNs, thereby preparing uninfected cells for potential infection (33). Thus, endothelial cells undergo many dsRNA-induced changes before apoptosis in non-IFN-␣-primed cells. The delay in the interval between when dsRNA is encountered and when apoptosis occurs increases the likelihood that viral production will be successfully completed, but other responses to dsRNA should be of benefit in the host response. Importantly, the type I IFN produced in response to dsRNA should be of benefit to uninfected neighboring cells by altering them to the risk of potential infection and modifying the apoptotic response, as shown in this study. Many DNA and RNA viruses generate sufficient dsRNA to activate PKR and/or 2⬘5⬘-OAS, and most viruses express viral genes that disrupt specific components of the innate antiviral machinery, providing evidence that the dsRNA-induced cellular responses are detrimental to most viruses (35–38, 40, 80, 99). Studies with synthetic dsRNA reveal the cellular response to viral infection unobscured by antagonistic viral genes. The knowledge of the ideal innate antiviral response elicited by synthetic dsRNA will be useful not only for understanding this important antiviral pathway, but also for exploring how viruses activate or subvert these responses.

Acknowledgments We thank Dr. Paula Vertino and members of her laboratory for reagents and helpful discussions.

References 1. Castelli, J., K. A. Wood, and R. J. Youle. 1998. The 2-5A system in viral infection and apoptosis. Biomed. Pharmacother. 52:386. 2. Williams, B. R. 1999. PKR; a sentinel kinase for cellular stress. Oncogene 18:6112. 3. Williams, B. R., C. S. Gilbert, and I. M. Kerr. 1979. The respective roles of the protein kinase and pppA2⬘ p5⬘ A2⬘ p5 A-activated endonuclease in the inhibition of protein synthesis by double stranded RNA in rabbit reticulocyte lysates. Nucleic Acids Res. 6:1335. 4. Medzhitov, R., and C. Janeway, Jr. 2000. Innate immune recognition: mechanisms and pathways. Immunol. Rev. 173:89.

dsRNA-INDUCED APOPTOSIS IN ENDOTHELIAL CELLS 5. Au, W. C., P. A. Moore, W. Lowther, Y. T. Juang, and P. M. Pitha. 1995. Identification of a member of the interferon regulatory factor family that binds to the interferon-stimulated response element and activates expression of interferoninduced genes. Proc. Natl. Acad. Sci. USA 92:11657. 6. Lin, R., C. Heylbroeck, P. M. Pitha, and J. Hiscott. 1998. Virus-dependent phosphorylation of the IRF-3 transcription factor regulates nuclear translocation, transactivation potential, and proteasome-mediated degradation. Mol. Cell. Biol. 18:2986. 7. Weaver, B. K., K. P. Kumar, and N. C. Reich. 1998. Interferon regulatory factor 3 and CREB-binding protein/p300 are subunits of double-stranded RNA-activated transcription factor DRAF1. Mol. Cell. Biol. 18:1359. 8. Visvanathan, K. V., and S. Goodbourn. 1989. Double-stranded RNA activates binding of NF-␬B to an inducible element in the human ␤-interferon promoter. EMBO J. 8:1129. 9. Maniatis, T., J. V. Falvo, T. H. Kim, T. K. Kim, C. H. Lin, B. S. Parekh, and M. G. Wathelet. 1998. Structure and function of the interferon-␤ enhanceosome. Cold Spring Harbor Symp. Quant. Biol. 63:609. 10. Kalvakolanu, D., and E. C. Borden. 1996. An overview of the interferon system: signal transduction and mechanisms of action. Cancer Invest. 14:25. 11. Samuel, C. E. 2001. Antiviral actions of interferons. Clin. Microbiol. Rev. 14:778. 12. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognition of double-stranded RNA and activation of NF-␬B by Toll-like receptor 3. Nature 413:732. 13. Miettinen, M., T. Sareneva, I. Julkunen, and S. Matikainen. 2001. IFNs activate Toll-like receptor gene expression in viral infections. Genes Immun. 2:349. 14. Friedman, H. M., E. J. Macarak, R. R. MacGregor, J. Wolfe, and N. A. Kefalides. 1981. Virus infection of endothelial cells. J. Infect. Dis. 143:266. 15. Harcourt, B., P. Rota, K. Hummel, W. Bellini, and M. Offermann. 1999. Induction of intercellular adhesion molecule 1 gene expression by measles virus in human umbilical vein endothelial cells. J. Med. Virol. 57:9. 16. Harcourt, B. H., A. Sanchez, and M. K. Offermann. 1999. Ebola virus selectively inhibits responses to interferons, but not to interleukin-1␤, in endothelial cells. J. Virol. 73:3491. 17. Kahl, M., D. Siegel-Axel, S. Stenglein, G. Jahn, and C. Sinzger. 2000. Efficient lytic infection of human arterial endothelial cells by human cytomegalovirus strains. J. Virol. 74:7628. 18. Schnittler, H. J., F. Mahner, D. Drenckhahn, H. D. Klenk, and H. Feldmann. 1993. Replication of Marburg virus in human endothelial cells: a possible mechanism for the development of viral hemorrhagic disease. J. Clin. Invest. 91:1301. 19. Harcourt, B. H., A. Sanchez, and M. K. Offermann. 1998. Ebola virus inhibits induction of genes by double-stranded RNA in endothelial cells. Virology 252:179. 20. Zaki, S. R., W.-J. Shieh, P. W. Greer, C. S. Goldsmith, T. Ferebee, J. Katshitshi, F. K. Tshioko, M. A. Bwaka, T. Swanepoel, P. Calain, et al. 1999. A novel immunohistochemical assay for the detection of Ebola virus in skin: implications for diagnosis, spread, and surveillance of Ebola hemorrhagic fever. J. Infect. Dis. 179:S36. 21. Boshoff, C., T. F. Schulz, M. M. Kennedy, A. K. Graham, C. Fisher, A. Thomas, J. O. McGee, R. A. Weiss, and J. J. O’Leary. 1995. Kaposi’s sarcoma-associated herpesvirus infects endothelial and spindle cells. Nat. Med. 1:1274. 22. Sturzl, M., C. Blasig, A. Schreier, F. Neipel, C. Hohenadl, E. Cornali, G. Ascherl, S. Esser, N. H. Brockmeyer, M. Ekman, et al. 1997. Expression of HHV-8 latency-associated T0.7 RNA in spindle cells and endothelial cells of AIDS-associated, classical and African Kaposi’s sarcoma. Int. J. Cancer 72:68. 23. Borchers, A. T., R. Perez, G. Kaysen, A. A. Ansari, and M. E. Gershwin. 1999. Role of cytomegalovirus infection in allograft rejection: a review of possible mechanisms. Transplant Immunol. 7:75. 24. Fish, K. N., C. Soderberg-Naucler, L. K. Mills, S. Stenglein, and J. A. Nelson. 1998. Human cytomegalovirus persistently infects aortic endothelial cells. J. Virol. 72:5661. 25. Degre, M. 2002. Has cytomegalovirus infection any role in the development of atherosclerosis? Clin. Microbiol. Infect. 8:191. 26. Harcourt, J. L., M. K. Hagan, and M. K. Offermann. 2000. Modulation of doublestranded RNA-mediated gene induction by interferon in human umbilical vein endothelial cells. J. Interferon Cytokine Res. 20:1007. 27. Offermann, M. K., J. Zimring, K. H. Mellits, M. K. Hagan, R. Shaw, R. M. Medford, M. B. Mathews, S. Goodbourn, and R. Jagus. 1995. Activation of the double-stranded-RNA-activated protein kinase and induction of vascular cell adhesion molecule-1 by poly(I).poly(C) in endothelial cells. Eur. J. Biochem. 232:28. 28. Tailor, A., and D. N. Granger. 2000. Role of adhesion molecules in vascular regulation and damage. Curr. Hypertens. Rep. 2:78. 29. Alexander, J. S., J. W. Elrod, and J. H. Park. 2001. Roles of leukocyte and immune cell junctional proteins. Microcirculation 8:169. 30. Gromme, M., and J. Neefjes. 2002. Antigen degradation or presentation by MHC class I molecules via classical and non-classical pathways. Mol. Immunol. 39:181. 31. Klein, J., and A. Sato. 2000. The HLA system: first of two parts. N. Engl. J. Med. 343:702. 32. Garfinkel, S., D. S. Haines, S. Brown, J. Wessendorf, D. H. Gillespie, and T. Maciag. 1992. Interleukin-1␣ mediates an alternative pathway for the antiproliferative action of poly(I.C) on human endothelial cells. J. Biol. Chem. 267:24375.

The Journal of Immunology 33. Einhorn, S., A. Eldor, I. Vlodavsky, Z. Fuks, and A. Panet. 1985. Production and characterization of interferon from endothelial cells. J. Cell. Physiol. 122:200. 34. Strasser, A., L. O’Connor, and V. M. Dixit. 2000. Apoptosis signaling. Annu. Rev. Biochem. 69:217. 35. Benedict, C. A., P. S. Norris, and C. F. Ware. 2002. To kill or be killed: viral evasion of apoptosis. Nat. Immun. 3:1013. 36. Cuconati, A., and E. White. 2002. Viral homologs of bcl-2: role of apoptosis in the regulation of virus infection. Genes Dev. 16:2465. 37. Hardwick, J., and D. Bellows. 2003. Viral versus cellular bcl-2 proteins. Cell Death Differ. 10:S68. 38. Ameisen, J. C., O. Pleskoff, J. D. Lelievre, and F. De Bels. 2003. Subversion of cell survival and cell death: viruses as enemies, tools, teachers and allies. Cell Death Differ. 10:S3. 39. Gale, M., Jr., B. Kwieciszewski, M. Dossett, H. Nakao, and M. G. Katze. 1999. Antiapoptotic and oncogenic potentials of hepatitis C virus are linked to interferon resistance by viral repression of the PKR protein kinase. J. Virol. 73:6506. 40. Barber, G. N. 2001. Host defense, viruses and apoptosis. Cell Death Differ. 8:113. 41. Hardwick, J. M. 2001. Apoptosis in viral pathogenesis. Cell Death Differ. 8:109. 42. Tschopp, J., M. Thome, K. Hofmann, and E. Meinl. 1998. The fight of viruses against apoptosis. Curr. Opin. Genet. Dev. 8:82. 43. Wang, J., J. D. Belcher, P. H. Marker, D. E. Wilcken, G. M. Vercellotti, and X. L. Wang. 2001. Cytomegalovirus inhibits p53 nuclear localization signal function. J. Mol. Med. 78:642. 44. Aguilar-Lemarroy, A., P. Gariglio, N. J. Whitaker, S. T. Eichhorst, H. zur Hausen, P. H. Krammer, and F. Rosl. 2002. Restoration of p53 expression sensitizes human papillomavirus type 16 immortalized human keratinocytes to CD95-mediated apoptosis. Oncogene 21:165. 45. Filippova, M., H. Song, J. L. Connolly, T. S. Dermody, and P. J. Duerksen-Hughes. 2002. The human papillomavirus 16 E6 protein binds to tumor necrosis factor (TNF) R1 and protects cells from TNF-induced apoptosis. J. Biol. Chem. 277:21730. 46. Friborg, J., Jr., W. Kong, M. O. Hottiger, and G. J. Nabel. 1999. p53 inhibition by the LANA protein of KSHV protects against cell death. Nature 402:889. 47. Cheng, E. H., J. Nicholas, D. S. Bellows, G. S. Hayward, H. G. Guo, M. S. Reitz, and J. M. Hardwick. 1997. A Bcl-2 homolog encoded by Kaposi sarcoma-associated virus, human herpesvirus 8, inhibits apoptosis but does not heterodimerize with Bax or Bak. Proc. Natl. Acad. Sci. USA 94:690. 48. Low, W., M. Harries, H. Ye, M. Q. Du, C. Boshoff, and M. Collins. 2001. Internal ribosome entry site regulates translation of Kaposi’s sarcoma-associated herpesvirus FLICE inhibitory protein. J. Virol. 75:2938. 49. Moore, P. S., and Y. Chang. 1998. Kaposi’s sarcoma-associated herpesvirusencoded oncogenes and oncogenesis. J. Natl. Cancer Inst. Monogr. 23:65. 50. Bannerman, D. D., J. C. Tupper, W. A. Ricketts, C. F. Bennett, R. K. Winn, and J. M. Harlan. 2001. A constitutive cytoprotective pathway protects endothelial cells from lipopolysaccharide-induced apoptosis. J. Biol. Chem. 276:14924. 51. Sata, M., and K. Walsh. 1998. Endothelial cell apoptosis induced by oxidized LDL is associated with the down-regulation of the cellular caspase inhibitor FLIP. J. Biol. Chem. 273:33103. 52. Adams, M. D., M. Dubnick, A. R. Kerlavage, R. Moreno, J. M. Kelley, T. R. Utterback, J. W. Nagle, C. Fields, and J. C. Venter. 1992. Sequence identification of 2,375 human brain genes. Nature 355:632. 53. Justesen, J., R. Hartmann, and N. O. Kjeldgaard. 2000. Gene structure and function of the 2⬘-5⬘-oligoadenylate synthetase family. Cell. Mol. Life Sci. 57:1593. 54. Tan, S. L., and M. G. Katze. 1999. The emerging role of the interferon-induced PKR protein kinase as an apoptotic effector: a new face of death? J. Interferon Cytokine Res. 19:543. 55. Turpaev, K., R. Hartmann, and J. Justesen. 1999. 2⬘-Adenylated derivatives of Ap3A activate RNase L. FEBS Lett. 457:9. 56. Jiang, Z., M. Zamanian-Daryoush, H. Nie, A. M. Silva, and X. Li. 2003. Poly I:C-induced TLR-3-mediated activation of NF-␬B and MAP kinases is through an IRAK-independent pathway employing components TLR3-TRAF6-TAK-1TAB2-PKR. J. Biol. Chem. 27:27. 57. Kumar-Sinha, C., S. Varambally, A. Sreekumar, and A. M. Chinnaiyan. 2002. Molecular cross-talk between the TRAIL and interferon signaling pathways. J. Biol. Chem. 277:575. 58. Vidalain, P. O., O. Azocar, H. Yagita, C. Rabourdin-Combe, and C. Servet-Delprat. 2001. Cytotoxic activity of human dendritic cells is differentially regulated by double-stranded RNA and CD40 ligand. J. Immunol. 167:3765. 59. Vattem, K. M., K. A. Staschke, and R. C. Wek. 2001. Mechanism of activation of the double-stranded-RNA-dependent protein kinase, PKR: role of dimerization and cellular localization in the stimulation of PKR phosphorylation of eukaryotic initiation factor-2 (eIF2). Eur. J. Biochem. 268:3674. 60. Desai, S. Y., and G. C. Sen. 1997. Effects of varying lengths of double-stranded RNA on binding and activation of 2⬘-5⬘-oligoadenylate synthetase. J. Interferon Cytokine Res. 17:531. 61. Gribaudo, G., D. Lembo, G. Cavallo, S. Landolfo, and P. Lengyel. 1991. Interferon action: binding of viral RNA to the 40-kilodalton 2⬘-5⬘-oligoadenylate synthetase in interferon-treated HeLa cells infected with encephalomyocarditis virus. J. Virol. 65:1748. 62. Hengartner, M. O. 2000. The biochemistry of apoptosis. Nature 407:770. 63. Konopleva, M., S. Zhao, Z. Xie, H. Segall, A. Younes, D. F. Claxton, Z. Estrov,

1709

64. 65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78. 79. 80. 81.

82.

83.

84.

85.

86.

87. 88. 89.

S. M. Kornblau, and M. Andreeff. 1999. Apoptosis: molecules and mechanisms. Adv. Exp. Med. Biol. 457:217. Donepudi, M., and M. G. Grutter. 2002. Structure and zymogen activation of caspases. Biophys. Chem. 101:145. Stennicke, H. R., M. Renatus, M. Meldal, and G. S. Salvesen. 2000. Internally quenched fluorescent peptide substrates disclose the subsite preferences of human caspases 1, 3, 6, 7 and 8. Biochem. J. 350:563. Thornberry, N. A., T. A. Rano, E. P. Peterson, D. M. Rasper, T. Timkey, M. Garcia-Calvo, V. M. Houtzager, P. A. Nordstrom, S. Roy, J. P. Vaillancourt, et al. 1997. A combinatorial approach defines specificities of members of the caspase family and granzyme B: functional relationships established for key mediators of apoptosis. J. Biol. Chem. 272:17907. Chen, D., R. A. Stetler, G. Cao, W. Pei, C. O’Horo, X. M. Yin, and J. Chen. 2000. Characterization of the rat DNA fragmentation factor 35/inhibitor of caspaseactivated DNase (short form): the endogenous inhibitor of caspase-dependent DNA fragmentation in neuronal apoptosis. J. Biol. Chem. 275:38508. Gu, J., R. P. Dong, C. Zhang, D. F. McLaughlin, M. X. Wu, and S. F. Schlossman. 1999. Functional interaction of DFF35 and DFF45 with caspase-activated DNA fragmentation nuclease DFF40. J. Biol. Chem. 274:20759. Martinon, F., K. Burns, and J. Tschopp. 2002. The inflammasome: a molecular platform triggering activation of inflammatory caspases and processing of proIL-␤. Mol. Cell 10:417. Kuida, K., J. A. Lippke, G. Ku, M. W. Harding, D. J. Livingston, M. S. Su, and R. A. Flavell. 1995. Altered cytokine export and apoptosis in mice deficient in interleukin-1␤ converting enzyme. Science 267:2000. Thornberry, N. A., H. G. Bull, J. R. Calaycay, K. T. Chapman, A. D. Howard, M. J. Kostura, D. K. Miller, S. M. Molineaux, J. R. Weidner, J. Aunins, et al. 1992. A novel heterodimeric cysteine protease is required for interleukin-1␤ processing in monocytes. Nature 356:768. Janicke, R. U., P. Ng, M. L. Sprengart, and A. G. Porter. 1998. Caspase-3 is required for ␣-fodrin cleavage but dispensable for cleavage of other death substrates in apoptosis. J. Biol. Chem. 273:15540. Li, S., Y. Zhao, X. He, T. H. Kim, D. K. Kuharsky, H. Rabinowich, J. Chen, C. Du, and X. M. Yin. 2002. Relief of extrinsic pathway inhibition by the Biddependent mitochondrial release of Smac in Fas-mediated hepatocyte apoptosis. J. Biol. Chem. 277:26912. Tafani, M., N. O. Karpinich, K. A. Hurster, J. G. Pastorino, T. Schneider, M. A. Russo, and J. L. Farber. 2002. Cytochrome c release upon Fas receptor activation depends on translocation of full-length bid and the induction of the mitochondrial permeability transition. J. Biol. Chem. 277:10073. Drexler, H. C., W. Risau, and M. A. Konerding. 2000. Inhibition of proteasome function induces programmed cell death in proliferating endothelial cells. FASEB J. 14:65. Chen, Z., J. Hagler, V. J. Palombella, F. Melandri, D. Scherer, D. Ballard, and T. Maniatis. 1995. Signal-induced site-specific phosphorylation targets I␬B␣ to the ubiquitin-proteasome pathway. Genes Dev. 9:1586. Palombella, V. J., O. J. Rando, A. L. Goldberg, and T. Maniatis. 1994. The ubiquitin-proteasome pathway is required for processing the NF-␬B1 precursor protein and the activation of NF-␬B. Cell 78:773. Sears, C., J. Olesen, D. Rubin, D. Finley, and T. Maniatis. 1998. NF-␬B p105 processing via the ubiquitin-proteasome pathway. J. Biol. Chem. 273:1409. Li, X. L., and B. A. Hassel. 2001. Involvement of proteasomes in gene induction by interferon and double-stranded RNA. Cytokine 14:247. Katze, M. G., Y. He, and M. Gale, Jr. 2002. Viruses and interferon: a fight for supremacy. Nat. Rev. Immunol. 2:675. Slowik, M. R., W. Min, T. Ardito, A. Karsan, M. Kashgarian, and J. S. Pober. 1997. Evidence that tumor necrosis factor triggers apoptosis in human endothelial cells by interleukin-1-converting enzyme-like protease-dependent and -independent pathways. Lab. Invest. 77:257. Rebouillat, D., and A. G. Hovanessian. 1999. The human 2⬘, 5⬘-oligoadenylate synthetase family: interferon-induced proteins with unique enzymatic properties. J. Interferon Cytokine Res. 19:295. Hovanessian, A. G. 1991. Interferon-induced and double-stranded RNA-activated enzymes: a specific protein kinase and 2⬘, 5⬘-oligoadenylate synthetases. J. Interferon Res. 11:199. Fitzgerald, K., S. McWhirter, K. Faia, D. Rowe, E. ZLatz, D. Golenbock, A. Coyle, S.-M. Liao, and T. Maniatis. 2003. IKKe and TBK1 are essential components of the IRF3 signaling pathway. Nat. Immun. 4:491. Rehm, M., H. Dussmann, R. U. Janicke, J. M. Tavare, D. Kogel, and J. H. Prehn. 2002. Single-cell fluorescence resonance energy transfer analysis demonstrates that caspase activation during apoptosis is a rapid process: role of caspase-3. J. Biol. Chem. 277:24506. Luo, X., I. Budihardjo, H. Zou, C. Slaughter, and X. Wang. 1998. Bid, a Bcl2 interacting protein, mediates cytochrome c release from mitochondria in response to activation of cell surface death receptors. Cell 94:481. Bernardi, P., V. Petronilli, F. Di Lisa, and M. Forte. 2001. A mitochondrial perspective on cell death. Trends Biochem. Sci. 26:112. Shi, Y. 2002. Mechanism of caspase activation and inhibition during apoptosis. Mol. Cell 9:459. Gil, J., J. Alcami, and M. Esteban. 1999. Induction of apoptosis by doublestranded-RNA-dependent protein kinase (PKR) involves the ␣ subunit of eukaryotic translation initiation factor 2 and NF-␬B. Mol. Cell. Biol. 19:4653.

1710 90. Zhou, A., J. Paranjape, T. L. Brown, H. Nie, S. Naik, B. Dong, A. Chang, B. Trapp, R. Fairchild, C. Colmenares, and R. H. Silverman. 1997. Interferon action and apoptosis are defective in mice devoid of 2⬘, 5⬘-oligoadenylate-dependent RNase L. EMBO J. 16:6355. 91. De Haro, C., R. Mendez, and J. Santoyo. 1996. The eIF-2␣ kinases and the control of protein synthesis. FASEB J. 10:1378. 92. Player, M. R., and P. F. Torrence. 1998. The 2-5A system: modulation of viral and cellular processes through acceleration of RNA degradation. Pharmacol. Ther. 78:55. 93. Saelens, X., M. Kalai, and P. Vandenabeele. 2001. Translation inhibition in apoptosis: caspase-dependent PKR activation and eIF2-␣ phosphorylation. J. Biol. Chem. 276:41620. 94. Stehlik, C., R. de Martin, I. Kumabashiri, J. A. Schmid, B. R. Binder, and J. Lipp. 1998. Nuclear factor (NF)-␬B-regulated X-chromosome-linked iap gene expression protects endothelial cells from tumor necrosis factor ␣-induced apoptosis. J. Exp. Med. 188:211.

dsRNA-INDUCED APOPTOSIS IN ENDOTHELIAL CELLS 95. Micheau, O., S. Lens, O. Gaide, K. Alevizopoulos, and J. Tschopp. 2001. NF-␬B signals induce the expression of c-FLIP. Mol. Cell. Biol. 21:5299. 96. MacFarlane, M., W. Merrison, S. B. Bratton, and G. M. Cohen. 2002. Proteasome-mediated degradation of Smac during apoptosis: XIAP promotes Smac ubiquitination in vitro. J. Biol. Chem. 16:16. 97. Huang, H., C. A. Joazeiro, E. Bonfoco, S. Kamada, J. D. Leverson, and T. Hunter. 2000. The inhibitor of apoptosis, cIAP2, functions as a ubiquitin-protein ligase and promotes in vitro monoubiquitination of caspases 3 and 7. J. Biol. Chem. 275:26661. 98. Suzuki, Y., Y. Nakabayashi, and R. Takahashi. 2001. Ubiquitin-protein ligase activity of X-linked inhibitor of apoptosis protein promotes proteasomal degradation of caspase-3 and enhances its anti-apoptotic effect in Fas-induced cell death. Proc. Natl. Acad. Sci. USA 98:8662. 99. Balachandran, S., P. C. Roberts, L. E. Brown, H. Truong, A. K. Pattnaik, D. R. Archer, and G. N. Barber. 2000. Essential role for the dsRNA-dependent protein kinase PKR in innate immunity to viral infection. Immunity 13:129.