Short Duration of Elevated vIRF-1 Expression during Lytic Replication of Human Herpesvirus 8 Limits Its Ability To Block Antiviral Responses Induced by Alpha Interferon in BCBL-1 Cells

Cell culture.

BCBL-1 (an HHV-8-positive, EBV-negative PEL cell line), JS-EBV (an EBV-immortalized B-lymphoblastoid cell line), and BL-41 (an HHV-8-negative, EBV-negative B cell line) (NIH AIDS Research and Reference Program, Rockville, Md.) were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U of penicillin per ml, and 10 U of streptomycin per ml. A reporter cell line for HHV-8, T1H6 (27), was maintained in Dulbecco's modified Eagle's medium (Cellgro, Herndon, Va.) supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U of penicillin per ml, 10 U of streptomycin per ml, and 50 μg of hygromycin B per ml and split 1:10 every 3 to 4 days.

Preparation of rabbit antiserum to vIRF-1.

Rabbit antiserum to vIRF-1 was generated against full-length His-tagged bacterially expressed vIRF-1 at the Pocono Rabbit Farm and Laboratory (Canadensis, Pa.). By Western blot analysis, the polyclonal antiserum to vIRF-1 at a 1:1,000 dilution did not cross-react with recombinant IRF-1, IRF-2, or extracts from cells not infected with HHV-8 (data not shown). However, it reacted with in vitro translated-vIRF-1 and with a protein of identical molecular mass that corresponded to the predicted molecular mass of vIRF-1 in HHV-8-positive BCBL-1 cells (the predicted molecular mass of vIRF-1 is 48,465 kDa).

Northern blot analysis.

Total cellular RNA was isolated from BCBL-1 cells by using Trizol (GibcoBRL, Grand Island, N.Y.) and was size fractionated on a 1% agarose-formaldehyde gel. RNA was transferred to nitrocellulose and covalently linked by UV irradiation with a Stratalinker (Stratagene, La Jolla, Calif.) and by baking in vacuo for 2 h at 80°C. DNA probes for vIRF-1 (74), Rta (54), viral interleukin-6 (vIL-6) (72), v-cyclin (72), 2′,5′-OAS (2), and PKR (26) were radiolabeled by using an oligolabeling kit (Stratagene) with [³²P]dCTP according to the manufacturer's recommendations. Hybridization was performed at 42°C in 5 ×SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), 1% sodium dodecyl sulfate (SDS), 5× Denhardt's solution, 100 μg of denatured salmon sperm DNA per ml, 50% formamide, 10% dextran sulfate, and radiolabeled DNA probe. The nitrocellulose was washed with a final stringency of 0.2× SSC in 0.1% SDS at 55°C. Blots were serially probed for the genes indicated, and the nitrocellulose was stripped with boiling water prior to rehybridization with other probes.

Western blot analysis.

Western blot analysis was performed with total cellular lysate from BCBL-1 cells that were harvested and lysed in cell lysis buffer (Cell Signaling Technology, Beverly, Mass.) with protease inhibitor cocktail (BD PharMingen, San Diego, Calif.). Total protein was size fractionated by SDS-12.5% polyacrylamide gel electrophoresis. Proteins were transferred from the acrylamide gel to polyvinylidene difluoride membranes. The membranes were blocked with 5% dry nonfat milk and then probed with antibodies against vIRF-1 at a 1:1,000 dilution, against vIL-6 at 1:80, and against α-tubulin at 1:1,000. After the secondary antibody reaction, the membranes were washed in TBS-Tween (0.02 M Tris [pH 7.6], 0.1 M sodium chloride, 0.05% Tween 20), visualized by using the enhanced chemiluminescence reaction (Amersham Pharmacia) as specified by the manufacturer, and subjected to autoradiography. Semiquantitative analysis of proteins was performed with National Institutes of Health image software. The antibody to α-tubulin was from Sigma (St. Louis, Mo.) and the vIL-6 antibody was a rabbit polyclonal antibody kindly provided by Patrick Moore and Yuan Chang.

ISH.

In situ hybridization (ISH) was performed on sections from cell pellets that were fixed overnight in 4% paraformaldehyde-0.01 M phosphate-buffered saline (pH 7.4) and embedded in paraffin. The sections were digested with proteinase K (75 μg/ml) for 10 min at 37°C. Both sense and antisense riboprobes were prepared from linearized pcDNA3/vIRF-1 (74) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (American Type Culture Collection, Manassas, Va.) plasmids by using digoxigenin-UTP (Roche Molecular Diagnostics/Boehringer Mannheim, Indianapolis, Ind.) according to the manufacturer's protocol. Antisense GAPDH, detecting an endogenous RNA target, was used as a positive control probe, while sense vIRF-1 and sense GAPDH were negative control probes. Hybridization was performed at 50°C for 18 to 24 h with riboprobes (10 ng/ml) in 50% formamide hybridization cocktail (Ameresco, Solon, Ohio). Sections were washed with a final stringency of 0.1× SSC at 37°C and then treated with RNase A (2.5 μg/ml) (Sigma) for 20 min. Alkaline phosphatase-conjugated antidigoxigenin antibody (1:500; Roche Molecular Diagnostics/Boehringer Mannheim) and nitroblue tetrazolium chloride grade III/5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt (Sigma) substrate were used for detection, followed by nuclear red counterstain. JS-EBV was used as a negative cellular control.

IFA.

Cytospin preparations from BCBL-1 cells were fixed in ice-cold acetone and air dried. Both single and dual immunofluorescence (IFA) analyses were done with the anti vIRF-1 (1:1,000), anti-PKR (B-10, 1:100; Santa Cruz Biotechnology), anti-PML (PG-M3, 1:500; Santa Cruz Biotechnology), anti-PPF (ORF 59 clone 11D1, 1:500; Advanced Biotechnologies, Columbia, Md.), anti-γ-tubulin (GTU-88, 1:1000; Sigma), anti-p53 (DO-1, 1:100, Santa Cruz Biotechnology), and anti-LANA (ORF 73, clone LN53, 1:500; Advanced Biotechnologies) antibodies. Secondary antibodies were goat anti-mouse, goat anti-rat, and goat anti-rabbit conjugated to Alexa Fluor 568 or Alexa Fluor 488 (Molecular Probes Inc., Eugene, Oreg.). The nuclei were counterstained with 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI) (Molecular Probes Inc.), and cells were mounted with the ProLong antifade kit (Molecular Probes). Imaging was performed with a Nikon Eclipse E-800 microscope equipped with an Optronics MagnaFire S99800 digital camera or with a Zeiss axioplasm laser-scanning confocal microscope with a Zeiss X100 1.3 oil emersion objective. The emission patterns of the two or three fluorescent probes were collected separately and were overlaid by use of MagnaFire software to create two- and three-color images.

Fluorescence ISH (FISH) analysis.

BCBL-1 cells were cytospun onto positively charged glass slides (ThermoShandon, Pittsburgh, Pa.), fixed in ice-cold acetone, air dried, and stained with vIRF-1 antibodies as described for IFA. The cells were then fixed with 10% paraformaldehyde at room temperature for 10 min; washed with phosphate-buffered saline twice; dehydrated sequentially for 1 min in ice-cold 70, 85, and 100% ethanol; and air dried. To denature the DNA, the cells were treated with 70% formamide in 2× SSC (pH 7.0 to 8.0) at 73°C for 7 min. The specimens were dehydrated at room temperature in the ethanol series and left to air dry. The CEP8 SpectrumOrange DNA probe (10 μl) to AT-rich alpha satellite sequences in the centromere region of chromosome 8 from the CEP8 SpectrumOrange direct-labeled fluorescent DNA probe kit (Vysis, Inc., Downers Grove, Ill.) was placed directly on the samples and covered with a glass coverslip that was sealed with rubber cement. The hybridization was carried out at 42°C in a humidified chamber for 4 h. The cells were washed with 0.4× SSC for 5 min at 73°C and with 2× SSC-0.1% NP-40 for 1 min at room temperature, followed by counterstaining with DAPI. Images were captured and processed as described for IFA.

Assessment of cell viability and apoptosis.

Trypan blue exclusion and direct counting with a hemacytometer were used to determine cell viability. Apoptosis was determined by using the DeadEnd fluorometric terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) system (Promega Corp., Madison, Wis.). As a negative control, BCBL-1 cells were incubated in the absence of terminal deoxynucleotidyltransferase enzyme. Cells were counterstained with propidium iodide containing DNase-free RNase A. For each sample, 10,000 events were acquired on a FACScan flow cytometer (Becton Dickinson, San Jose, Calif.).

Assay for infectious HHV-8 production with the T1H6 cell line.

HHV-8 infection of T1H6 cells activates the lacZ gene under control of the PAN promoter in a sensitive and quantitative manner (27). By using this cell line, the amounts of infectious HHV-8 in medium obtained from BCBL-1 cultures under different conditions were measured as described previously (27). Briefly, 8 × 10⁴ T1H6 cells were seeded in each well of a 48-well plate. The following day, medium from BCBL-1 cultures (200 μl) was added to each well in the presence of 8 μg of Polybrene per ml. Three days later, cells were harvested, and their β-galactosidase activities were measured by luminescent β-galactosidase assay (Clontech, Palo Alto, Calif.) with a LUMIstar Galaxy luminometer (BMG LabTechnologies, Durham, N.C.). A dilution series of infectious virus was used as a standard curve, and recombinant β-galactosidase was also used to determine the linear range of the assay.

Time course of vIRF-1 expression.

To elucidate the effects of vIRF-1 on IFN-induced changes during the course of viral infection, we first analyzed the time course of vIRF-1 expression by Western blot analysis. BCBL-1 cells that were not incubated with TPA expressed low levels of vIRF-1 at all time points examined (Fig. 1A). There was a slight increase in vIRF-1 expression at 8 and 24 h after TPA treatment, whereas peak levels were achieved by 48 h following incubation with TPA and remained elevated for at least 96 h. As a control, levels of α-tubulin were assessed and found to decrease slightly at the time points at which vIRF-1 increased.

To assess changes at an individual cell level, the numbers of cells that expressed high levels of vIRF-1 were counted by IFA. The percentage was not above the control level of 1 to 4% when cells were incubated with TPA for 24 h, but the percentage increased to approximately 30% of cells when incubation with TPA was continued for 48 h (Fig. 1B). When IFN-α was present during incubation with TPA, there was no increase in vIRF-1 expression, indicating that IFN-α blocked the ability of TPA to induce high-level vIRF-1 expression.

Unique localization of vIRF-1.

The cellular localization and morphological appearance of vIRF-1 within BCBL-1 cells varied among the conditions examined. Intense immunoreactivity for vIRF-1 was most frequently localized in the nucleus, but some cells contained high levels of vIRF-1 in the cytoplasm, including cells with high levels of nuclear and cytoplasmic vIRF-1 and others with high levels of vIRF-1 predominately in the cytoplasm (Fig. 2A to D). Some cells had multiple nuclear elements, each expressing vIRF-1 at high levels (Fig. 2B). The pattern of vIRF-1 expression in cells that were incubated with IFN-α alone did not differ from that in untreated cells (Fig. 2A and C), whereas cells that were simultaneously incubated with TPA and IFN-α had a distinct pattern of vIRF-1 that was characterized by large clumps of nuclear vIRF-1 and extensive punctuate vIRF-1 in the cytoplasm (Fig. 2D). In situ hybridization of nonstimulated cells demonstrated that only a few cells expressed very high levels of vIRF-1 mRNA, whereas vIRF-1 mRNA was below the level of detection in the majority of the population (Fig. 2E).

In addition to the strong immunoreactivity for vIRF-1 that was seen in a subset of cells, punctate immunoreactivity for vIRF-1 in the nucleus was observed in nearly all BCBL-1 cells (Fig. 3A). The punctate vIRF-1 protein found in all cells was distinct from the strong immunofluorescence for vIRF-1 that was associated with lytic replication (e.g., Fig. 3A has a single strongly positive cell surrounded by cells with the punctate nuclear vIRF-1 immunofluorescence). We hypothesized that vIRF-1 might be present in PML bodies, because vIRF-1 specifically binds to p53, p300, and CREB (37, 45, 60, 61), proteins that can be found in PML bodies (33, 70), and the punctate nature and nuclear localization of the vIRF-1 immunoreactivity suggested that it was specific. Dual IFA with antibodies to vIRF-1 and PML demonstrated complete colocalization of PML and vIRF-1 in latently infected cells producing yellow immunofluorescence (Fig. 3B), whereas confocal microscopy demonstrated that BCBL-1 cells that expressed vIRF-1 at a high level during lytic replication had vIRF-1 both within and outside the PML bodies (Fig. 3C, green cell). When cells were simultaneously examined for expression of vIRF-1 and p53, strong expression of p53 primarily occurred when HHV-8 was within the lytic phase of viral replication, with cells expressing high levels of vIRF-1 also expressing elevated levels of p53 (Fig. 3D). Confocal microscopy revealed that both proteins colocalized to some punctate areas in the nucleus (Fig. 3D, yellow dots in inset), but the levels of both proteins were so high that the majorities of p53 and vIRF-1 were not colocalized (Fig. 3D). Confocal microscopy revealed that the majority of BCBL-1 cells expressed PML without expressing any detectable p53 (Fig. 3E). Colocalization of p53 and PML was detected in cells that expressed p53, and the amount of p53 present far exceeded the amount of PML. Not all viral proteins examined colocalized with PML. For example, LANA and PML protein did not colocalize in any cell (Fig. 3F). Thus, vIRF-1, but not p53, was found associated with PML of all BCBL-1 cells, whereas p53 was colocalized with PML in BCBL-1 cells supporting lytic replication of HHV-8. The levels of vIRF-1 that were in PML bodes were low compared to the levels in some cells undergoing lytic replication of HHV-8, and cells expressing elevated levels of vIRF-1 also expressed elevated levels of p53. All subsequent studies exploring induction of vIRF-1 and the consequences of its expression during lytic replication focused on cells that had vIRF-1 levels that far exceeded the levels restricted to the PML bodies.

Centrosome patterns in multinucleated BCBL-1 cells.

Many, but not all, cells that expressed high levels of vIRF-1 protein had a multinucleated appearance. There are three potential explanations for the multinucleated appearance. (i) vIRF-1 has been reported to alter chromatin structure, preventing binding of propidium iodide to condensed DNA (37). Thus, vIRF-1 might be preventing access of DAPI to the DNA in some regions of the nucleus, thereby generating multiple nuclear lobes within a single nucleus. (ii) Some cells within the lytic phase might have undergone cellular DNA replication and nuclear division without cytokinesis, leading to multinucleated cells. (iii) Alternatively, abnormal chromosomal segregation of incompletely duplicated chromosomes could generate micronuclei. To explore these possibilities, we used γ-tubulin to identify centrosomes within the cells, since γ-tubulin is an essential component of centrosomes and centrosomes are required for appropriate chromosomal segregation during nuclear division (30-32, 44, 46). None of the cells that expressed high levels of vIRF-1 had more than one centrosome (red dots close to the nucleus), and many of the cells that expressed high levels of vIRF-1 lacked any detectable centrosome, even when they appeared to be multinucleated (Fig. 4A and data not shown). This is in contrast to the case for non-TPA-stimulated cells, which contained either one or two centrosomes in all cells except in spontaneously lytic cells that strongly expressed vIRF-1 and were deficient in centrosomes. The presence of multiple nuclear elements without multiple centrosomes in some vIRF-1-expressing cells suggested that there might not be multiple nuclei present. To assess this possibility, FISH with a probe for sequences in the centromere region of chromosome 8 was performed. Multinucleated cells should have one or two copies of chromosome 8 in each nucleus, whereas a hyperlobated nucleus or a micronucleus would not contain chromosome 8 in each element. IFA followed by FISH demonstrated that each nuclear element that expressed high levels of vIRF-1 contained genetic material from chromosome 8 (Fig. 4B), providing evidence that each nuclear element contained replicated chromosomes and hence had undergone nuclear division with appropriate chromosomal segregation. This indicates that there was a defect in cytokinesis that led to the formation of multinucleated cells during lytic replication of HHV-8. It has not yet been determined which viral gene(s) is responsible for these changes.

Short duration of vIRF-1 expression.

We were surprised by the relatively low percentage of cells that expressed high levels of vIRF-1 during prolonged incubation with TPA. This could occur if only a low percentage of cells were entering the lytic cascade. Alternatively, it could occur if high levels of vIRF-1 were expressed for only a short time during lytic replication. We explored these possibilities by comparing expression of vIRF-1 with that of PPF (encoded by ORF 59), a delayed-early viral protein expressed exclusively during lytic replication (9, 10). Cells that were not incubated with TPA included very few cells that expressed elevated levels of vIRF-1 or PPF (Fig. 5A), but low levels of vIRF-1 were present in the PML bodies as seen under higher magnification (data not shown). The percentage of cells that expressed high levels of vIRF-1 increased in response to incubation with TPA and stabilized after 2 days of TPA, whereas the percentage that expressed PPF continued to increase, with the majority of cells expressing PPF after 5 days of incubation with TPA (Fig. 5A). The majority of cells that expressed high levels of vIRF-1 also expressed PPF, whereas many cells expressed PPF without having elevated levels of vIRF-1. The high percentage of cells that expressed PPF indicated that most BCBL-1 cells entered the lytic cascade in response to prolonged incubation with TPA, yet not all cells within the lytic cascade expressed vIRF-1. This suggested that the expression of vIRF-1 might be of shorter duration than the expression of PPF.

We then assessed the consequences of the removal of TPA after 48 h of stimulation on the expression of vIRF-1 and PPF in the population, looking at expression of each protein individually by IFA. Most cells in the population expressed high levels of PPF in response to incubation with TPA for 48 h, and this percentage did not decline when TPA was withdrawn and incubation was continued for 24 h in medium lacking TPA (Fig. 5B). In contrast, there was a decline in the percentage of cells that expressed high levels of vIRF-1 when TPA was withdrawn (Fig. 5B and 6A). This suggested that cells continued to be recruited into the lytic phase of viral replication when incubation with TPA was continued, whereas withdrawal of TPA prevented additional cells from entering the lytic cascade yet allowed cells within the cascade to progress. The decline that occurred in the percentage of cells that expressed vIRF-1 when TPA was withdrawn suggested that high-level vIRF-1 was expressed relatively early in the lytic cascade and did not persist as cells progressed to later stages of the cascade, whereas PPF expression was of longer duration. The percentage of cells that expressed high levels of vIRF-1 declined from approximately 30 to 10 to 12% over a 24-h period when TPA was removed from cells that had been primed for 48 h (Fig. 6A). When IFN-α was added to cells after the TPA was removed, the percentage of cells that expressed high levels of vIRF-1 declined more rapidly.

The rapid decline in the percentage of cells that expressed high levels of vIRF-1 suggested that the protein was not stable. This hypothesis was confirmed when TPA-stimulated BCBL-1 cells were incubated with cycloheximide to inhibit protein synthesis. This led to a rapid decline in vIRF-1 expression in TPA-stimulated BCBL-1 cells and revealed that the protein half-life was less than 3 h (Fig. 6B). In contrast, levels of α-tubulin were more stable during incubation with cycloheximide. Thus, vIRF-1 protein levels would rapidly decline unless de novo synthesis of vIRF-1 was occurring.

Inhibitory effect of vIRF-1 on responses to IFN-α in BCBL-1 cells.

Based on the short half-life of vIRF-1 and the expression of high levels of vIRF-1 for only a short portion of the lytic cascade, we considered that vIRF-1 might not be expressed at high levels for a duration sufficient to protect cells from IFN-α-induced changes. To explore this possibility, we used IFA to determine whether cells that expressed high levels of vIRF-1 in response to TPA were resistant to induction of PKR protein by IFN-α. The percentage of cells that expressed high levels of vIRF-1 increased in response to incubation with TPA for 48 h (Fig. 7A and B). Incubation with IFN-α induced high levels of PKR protein expression in cells that lacked elevated levels of vIRF-1 but not in cells that expressed elevated levels of vIRF-1 (Fig. 7C). Examination using filters that exclusively measured PKR expression confirmed that PKR protein was not expressed when levels of vIRF-1 were elevated (data not shown), providing evidence that vIRF-1 offered some protection from IFN-α-induced changes.

Consequences of TPA priming for gene induction by IFN-α.

Cells that expressed high levels of vIRF-1 showed some protection from IFN-α-induced changes, yet the percentage of TPA-stimulated cells that expressed high levels of vIRF-1 was never more than 35%, even though most cells entered the lytic cascade. This suggested that many of the cells within the lytic cascade were likely to respond to IFN-α with increased expression of antiviral genes. This issue was explored by using Northern blot analysis to compare changes in gene expression resulting from incubation with IFN-α in nonprimed and TPA-primed cells. As expected, TPA-primed cells expressed much higher levels of the lytic cycle mRNAs (Rta, vIRF-1, and vIL-6) than did cells that were not primed with TPA (Fig. 8A). In addition, there was an increase in the level of the latency-associated transcript v-cyclin, but this increase was minor compared to the increase of the lytic-associated genes. Importantly, both TPA-primed and nonprimed cells had similar increases in PKR and 2′,5′-OAS mRNAs when incubated with IFN-α for 8 h, demonstrating that cells harboring HHV-8 in the lytic and latent phases responded comparably to IFN-α as a population, even if individual cells expressing high levels of vIRF-1 or other vIRFs were more resistant. The product of the viral gene Rta is both necessary and sufficient for inducing HHV-8 to enter the lytic cascade (40, 63). Removal of TPA for 24 h reduced the amount of Rta expression by more than 80%, whereas it was undetectable if cells were incubated with IFN-α for 24 h after TPA was removed. Incubation of TPA-primed cells with IFN-α for 24 h also decreased the expression of vIRF-1 and vIL-6 by at least 90% compared to the same time point in the absence of IFN-α, but the levels of PKR and 2′,5′-OAS did not decline to the same extent. Incubation of non-TPA-primed cells with IFN-α did not induce increased expression of vIL-6 mRNA, in contrast to the induction of vIL-6 by IFN-α that was reported to occur in BCP-1 cells (12). Western blot analysis confirmed that TPA induced high levels of both vIRF-1 and vIL-6 protein expression, and IFN-α was insufficient to induce vIL-6 in nonprimed BCBL-1 cells (Fig. 8B). Furthermore, incubation of TPA-primed cells with IFN-α accelerated the decline in vIRF-1 protein levels. From these studies, we can conclude that (i) prolonged incubation with TPA led to the recruitment of additional cells into the lytic phase of viral replication, whereas removal of the TPA decreased the recruitment, leading to a decline in vIRF-1 and Rta expression; (ii) despite the high-level vIRF-1 expression, the population of TPA-stimulated cells remained as responsive to IFN-α as nonprimed cells; and (iii) incubation of TPA-primed cells with IFN-α reduced the expression of vIRF-1 mRNA and protein and the expression of other lytic-associated genes, indicating that cells within the lytic cascade were not fully protected from IFN-α.

Decreased proliferation and apoptosis in response to TPA priming.

The loss of centrosomes in cells undergoing lytic replication of HHV-8 combined with the expression of replication-associated protein, a viral protein that blocks cellular DNA replication (68, 71), should prevent cellular replication as HHV-8 progresses through the lytic phase. We explored this issue by examining changes in cell number over a 48-h period for TPA-primed and unprimed BCBL-1 cells. Cells that were not induced into the lytic cycle increased in number by over fourfold over 48 h, whereas there was a twofold increase in the number of TPA-primed cells (Fig. 9A). When TPA-primed cells were incubated with IFN-α, there was no increase in cell number during the 48-h incubation. Incubation of nonprimed cells with IFN-α had less of an impact on cell number than priming with TPA, with cells increasing threefold in number over the 48-h incubation. Thus, IFN-α reduced the number of TPA-primed cells more effectively than the number of nonprimed cells.

We used the TUNEL assay to assess the role of apoptosis in the changes in cell number that occurred in response to TPA and IFN-α. We found a greater-than-twofold increase in the percentage of cells undergoing apoptosis at 24 h after the completion of TPA priming (Fig. 9B and C). When TPA-primed cells were incubated with IFN-α for 24 h, the percentage of TUNEL-positive cells increased to more than 40%, whereas incubation with IFN-α had very little effect on amount of apoptosis that occurred in nonprimed cells in this time frame (approximately 10% in the absence and presence of IFN-α). These results demonstrate that many BCBL-1 cells that entered the lytic cascade in response to TPA priming withdrew from the cell cycle and underwent apoptosis, and the level of apoptosis was doubled if TPA-primed cells were incubated with IFN-α. In contrast, incubation with IFN-α for 24 h marginally increased the amount of apoptosis in latently infected cells.

Effect of IFN-α on production of infectious HHV-8 in TPA-stimulated BCBL-1 cells.

Many cells within the lytic cascade responded to IFN-α with enhanced expression of antiviral genes and enhanced apoptosis, suggesting that neither vIRF-1 nor other viral proteins blocked responses to IFN-α in many BCBL-1 cells that entered the lytic cascade. This suggested that IFN-α might have a profound effect on production of infectious virus by preventing successful completion of the process. Alternatively, the disruption of responses to IFN-α that occurred when high levels of vIRF-1 were transiently expressed during the lytic cascade might have been sufficient to ensure production of infectious virus. To explore these possibilities, we quantitated the amount of infectious virus produced by TPA-stimulated BCBL-1 cells and examined the impact of IFN-α on the process. Production of infectious virus was measured by using an Rta-dependent reporter cell line whose transcriptional activation was directly proportional to the amount of infectious HHV-8, as previously reported (27) and as shown in Fig. 10C. In the absence of IFN-α, infectious virus began to appear in conditioned medium 3 days following incubation with TPA and climbed to peak levels by day 5 (data not shown). We thus examined changes in cell number and the amount of infectious virus present in conditioned medium at 4 and 5 days following TPA stimulation. When BCBL-1 cells were incubated with TPA for 4 and 5 days, the number of viable cells was reduced by approximately 80% compared to control cells grown in the absence of TPA (Fig. 10A). In contrast, incubation with IFN-α in the absence of TPA reduced the cell number by less than 20% (data not shown). The reduction in cell number that occurred in response to incubation with TPA was expected, since entry into the lytic cascade blocked cellular proliferation and induced apoptosis, as shown in Fig. 9. Despite the low number of viable cells, infectious virus was readily detectable in the conditioned medium from TPA-stimulated cells and continued to increase between days 4 and 5, whereas there was minimal infectious virus in conditioned medium from nonstimulated cells on either day even though the number of viable cells was much higher (Fig. 10B). When IFN-α was present throughout the incubation with TPA, all TPA-induced accumulation of extracellular virus was eliminated. There was also a reduction in the number of viable cells, but the impact of IFN-α on the TPA-induced changes in cell number were minor compared to the impact on TPA-induced changes in infectious virus production (compare TPA to TPA+ IFN-α in Fig. 10A and B). When the addition of IFN-α was delayed until 24 or 48 h after the addition of TPA, the number of surviving cells approached the number that resulted from incubation with TPA alone (Fig. 10A), yet the production of infectious virus was still severely compromised (Fig. 10B). When the addition of IFN-α was delayed until 48 h after TPA stimulation, then extracellular infectious virus appeared, but the amount was small and levels did not increase between days 4 and 5, unlike the case for TPA-stimulated cells in the absence of IFN-α. Thus, IFN-α was effective at preventing production of infectious virus, especially when present early in the lytic cascade.