Inhibition of Infection and Replication of Human Herpesvirus 8 in Microvascular Endothelial Cells by Alpha Interferon and Phosphonoformic Acid

Abstract

Infection of endothelial cells with human herpesvirus 8 (HHV-8) is an essential event in the development of Kaposi's sarcoma. When primary microvascular endothelial cells (MECs) were infected with HHV-8 at a low multiplicity of infection, considerable latent replication of HHV-8 occurred, leading to a time-dependent increase in the percentage of virus-infected cells that was accompanied by cellular spindling and growth to a high density with loss of contact inhibition. Only a low percentage of MECs supported lytic replication of HHV-8 and produced infectious virus. Phosphonoformic acid blocked production of infectious virus but did not inhibit the rapid expansion of latently infected MECs. Pretreatment of MECs with alpha interferon (IFN-α) prior to infection effectively reduced HHV-8 viral gene expression, latent replication, and production of infectious virus. High levels of the double-stranded RNA activated protein kinase (PKR) were expressed in HHV-8-infected cells, and incubation with IFN-α increased PKR expression more in virus-infected cells than in uninfected cells. MECs that were immortalized with simian virus 40 large-T antigen differed from nonimmortalized MECs in their response to infection with HHV-8 and demonstrated that cells with elevated levels of expression of antiviral transcripts expressed viral transcripts at reduced levels. These studies demonstrate that MECs respond to HHV-8 with enhanced expression of cellular antiviral genes and that augmentation of innate antiviral defenses with IFN-α is a more effective strategy than inhibition of viral lytic replication to protect MECs from infection with HHV-8 and to restrict proliferation of virus-infected MECs.

Human herpesvirus 8 (HHV-8), also known as Kaposi's sarcoma (KS) herpesvirus, is present in KS lesions from all risk groups worldwide and is essential for the development of KS (18, 19). There is considerable evidence that the spindle cells that comprise the KS tumor are of endothelial origin (28, 36, 72). Most of the spindle cells in KS lesions express a small number of viral transcripts that reflect latent infection. These include latency-associated nuclear antigen (LANA), kaposin (also called T0.7), and a bicistronic mRNA that encodes v-cyclin (also called K-cyclin and cyclin K) and v-FLICE inhibitory protein (4, 26, 63, 77-79). As lesions progress, the number of spindle cells increases, and most cells continue to express latency-associated viral gene products (26, 79, 80). In addition, a low percentage of KS spindle cells and macrophages within KS lesions harbor HHV-8 that is undergoing lytic replication (10, 77, 80, 90). Lytic replication is required for production of infectious virus and is accompanied by expression of some viral proteins that have paracrine functions (11, 15, 56, 65). The majority of cells in KS lesions are latently infected. Nonetheless, the paracrine effects of viral genes expressed during the lytic cascade probably contribute to the angiogenesis and inflammatory response that are characteristic of KS lesions (14, 25, 55).

Most KS cells are latently infected with HHV-8 (28, 70, 77) and are not susceptible to pharmacologic agents such as ganciclovir and phosphonoformic acid (PFA) that specifically target the lytic DNA replication machinery of herpesviruses (43, 53). Hence, agents that target lytic replication have had limited success in the treatment of KS (49). Clinical studies provide evidence that cytokines affect the development and/or progression of KS. Infusions of recombinant tumor necrosis factor alpha or granulocyte colony-stimulating factor can worsen KS (1, 76), whereas alpha interferon (IFN-α) induces remissions in patients with adequate CD4 counts who do not have constitutional symptoms (45, 46). In addition, high levels of inflammatory and angiogenic cytokines are often present in KS lesions (61, 73). These findings suggest that clinical factors that alter cytokine levels are likely to affect the development or severity of KS in individuals who are infected with HHV-8.

A variety of different cell types, including endothelial cells (ECs), secrete type I IFN in response to viral infection, thereby alerting neighboring cells to the risk of viral infection (40, 52). Type I IFNs, which include IFN-α and IFN-β, signal through a shared receptor to induce expression of multiple genes that can be activated if the cells become infected. These genes include those for the double-stranded RNA (dsRNA)-activated protein kinase (PKR) and the 2′5′-oligoadenylate synthetase (2′5′-OAS) (40, 71). dsRNA is a molecular motif encountered during viral infection that activates these enzymes. Viral infection can also activate PKR through PACT, a stress-modulated cellular activator of PKR (64, 71). The greater expression of PKR, 2′5′-OAS, and other antiviral proteins that is induced by type I IFN enhances the likelihood of having a rapid antiviral response upon activation by virus infection.

Multiple viral gene products are expressed during lytic replication, thereby increasing the likelihood for activation of the innate antiviral machinery. In contrast, there are only a few viral gene products expressed during latent replication, and this decreases the potential for recognition by the innate immune system. HHV-8 encodes both lytic and latent proteins that can disrupt responses to IFN and IFN-induced gene products (20, 50, 67, 91). Thus, the success of cellular antiviral defenses in protecting ECs from infection with HHV-8 is affected not only by cellular proteins but also by viral proteins.

In this study, we investigated early events that occurred following infection of dermal microvascular ECs (MECs) with HHV-8. We demonstrate that latent replication of HHV-8 occurred when nonimmortalized MECs were infected at a low multiplicity of infection, and latently infected cells had a growth advantage over uninfected cells. HHV-8-infected MECs increased as a percentage of the population and grew to a high density with loss of contact inhibition, even when virus production was blocked by PFA. MECs responded to HHV-8 with elevated expression of cellular antiviral genes, and cells that were virus infected expressed more PKR in response to incubation with IFN-α than did uninfected cells. We also demonstrate that IFN-α protected MECs from infection with HHV-8 and reduced the ability of HHV-8 to undergo either latent or lytic replication in MECs.

MATERIALS AND METHODS

Cell culture.

Human dermal MECs were obtained from the Emory University Skin Diseases Research Core Center and prepared from human foreskins as previously described (33, 60). MECs from multiple donors were pooled to minimize the impact of genetic differences on the cellular response to infection with HHV-8. The simian virus 40 (SV40)-immortalized MECs (SV40-MEC) used in these experiments were the HMEC-1 cell line that was generated by transfection of SV40 large-T antigen into MECs (2). Cells were cultured either in MCDB 131 medium (Mediatech, Herndon, Va.) supplemented with 30% heat-inactivated human serum, 5 ng of epidermal growth factor per ml, 1 μg of hydrocortisone per ml, 16 U of heparin per ml, 50 μM cyclic AMP, 2 mM l-glutamine, 100 U of penicillin per ml, and 100 U of streptomycin per ml (Invitrogen, Carlsbad, Calif.) or in Clonetics endothelial Bullet medium (EGM-2-MV medium; Cambrex BioScience Walkersville, Inc., San Diego, Calif.) 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 nonimmortalized MECs were used between passages 3 and 6. Where indicated, cells were treated with 1,000 U of recombinant IFN-α2b (Schering-Plough, Kenilworth, N.J.) per ml 24 h prior to infection, or IFN-α was added to fresh medium 24 h prior to harvest. Where indicated, PFA (Sigma) was added at a final concentration of 0.5 mM upon removal of the inoculum and was present in the medium, which was changed every 3 to 4 days. A reporter cell line for HHV-8, T1H6 (38), was maintained in Dulbecco's modification of Eagle's medium (Mediatech) supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U of penicillin per ml, 10 U of streptomycin per ml, and hygromycin B (50 μg/ml) and split 1:10 every 3 to 4 days.

RNA isolation and Northern blotting.

Total cellular RNA was prepared with RNAzol B (Tel-Test, Inc., Friendswood, Tex.) in accordance with the manufacturer's recommendation. For Northern blot analysis, RNA (20 μg) was size fractionated on a 1% agarose formaldehyde gel in the presence of 1 μg of ethidium bromide per ml (74). The RNA was transferred to nitrocellulose and covalently linked by baking in vacuo for 2 h at 80°C and by UV irradiation with a UV cross-linker (Stratalinker; Stratagene, La Jolla, Calif.). Hybridizations were performed at 42°C overnight in 5× SSC (1× SSC is 150 mM NaCl plus 15 mM Na citrate)-1% sodium dodecyl sulfate (SDS)-5× Denhardt's solution-50% formamide-10% dextran sulfate-100 μg of sheared denatured salmon sperm DNA per ml. Approximately 1 × 10⁶ to 2 × 10⁶ cpm of labeled probe per ml (specific activity, >10⁸ cpm/μg of DNA) was used in each hybridization. Following hybridization, membranes were washed with a final stringency of 0.2× SSC in 0.1% SDS at 55°C. The nitrocellulose was stripped with boiling water prior to rehybridization with other probes. Autoradiography was performed with an intensifying screen at −70°C.

³²P labeling of probes.

Radiolabeling was done with an oligolabeling kit (Stratagene, La Jolla, Calif.) and ³²P dCTP in accordance with the manufacturer's recommendations. The templates for the HHV-8 genes were produced by PCR amplification of the published sequence with purified HHV-8 DNA from BCBL-1 cells. ICAM-1 (82), 2′5′-OAS (9), PKR (54), and glyceraldehyde 3-phosphate dehydrogenase were inserts from the open reading frames (ORFs) excised from cDNA clones.

Production of cell-free HHV-8 virions.

Infectious virus was prepared from BCBL-1 cells that were seeded at 2 × 10⁵/ml and cultured in the presence of 0.3 mM butyrate or 20 ng of tetradecanoyl phorbol acetate (TPA) per ml. BCBL-1 cells were pelleted by centrifugation at 400 × g for 5 min. The cell-free culture medium was filtered through a 0.45-μm-pore-size filter to remove cellular debris, and the filtrate was spun at 15,000 × g for 3 h. The pellet containing virions was resuspended in serum-free M199 medium, aliquoted, and stored at −80°C until use.

Quantitation of infectious HHV-8 with the T1H6 reporter cell system.

Infectious HHV-8 was quantitated with the T1H6 reporter cell line, which contains a regulator of transcription activation (RTA)-responsive element upstream of a β-galactosidase reporter gene (38). Briefly, 8 × 10⁴ T1H6 cells were seeded per well into 48-well plates. On the following day, samples undergoing analysis for infectious virus were filtered through a 0.45-μm-pore-size membrane and incubated with the T1H6 reporter cells in the presence of 8 μg of Polybrene per ml, centrifuged for 30 min at 400 × g, and incubated for 90 min at 37°C. The medium was changed, and cells were cultured for 3 days in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum and penicillin-streptomycin. At harvest, the cells were washed once with phosphate-buffered saline (PBS) and then freeze-thawed three times in 50 μl of PBS. The extracts were clarified by centrifugation at 12,000 × g for 3 min, and aliquots of the supernatant were assayed for β-galactosidase with a Luminescent β-galactosidase assay (Clontech) and a LUMIstar Galaxy luminometer (BMG LabTechnologies, Durham, N.C.). Infection of T1H6 cells with a dilution series of virus was used to generate a standard curve.

Preparation of viral DNA and quantitative real-time PCR.

HHV-8 virions from BCBL-1 cells were pelleted by centrifugation for 3 h at 15,000 × g. The pellet was resuspended in buffer consisting of 40 mM Tris-HCl (pH 7.4), 10 mM NaCl, 6 mM MgCl₂, and 10 mM CaCl₂. DNase (Promega RQ1 DNases) was added to a final concentration of 12 U/ml, and RNase (Boehringer Mannheim RNase) was added to a concentration of 6.25 μg/ml. This suspension was incubated for 1 h at 37°C, followed by inactivation of DNase by incubation at 65°C for 15 min. One-fourth volume of 5× lysis buffer (100 mM Tris-HCl [pH 7.4], 50 mM EDTA, 500 mM NaCl, 2.5% SDS) was added. Proteinase K was then added to a final concentration of 0.1 mg/ml. The reaction mixture was subjected to incubation at 55°C for 30 min and at 37°C overnight, followed by two phenol-chloroform extractions. Glycogen was added as a carrier, and viral DNA was precipitated with an equal volume of isopropanol, followed by an ethanol wash. The pellet was suspended in 10 mM Tris-HCl-1 mM EDTA. For quantitative PCR, an aliquot of the sample in a background of 100 ng of human umbilical vein EC DNA was added to the Taqman 2× Universal Master Mix (Perkin-Elmer, Norwalk, Conn.) and the ORF26 primer and fluorogenic probe set described by White and Campbell (87) in a total volume of 25 μl. PCR cycling conditions were 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min in an iCycler (Bio-Rad, Hercules, Calif.). A standard curve for HHV-8 genome equivalents was generated by including a dilution series of BCBL-1 DNA (10¹ to 10⁵ copies per reaction mixture).

HHV-8 infection of cells.

For preparation of RNA for Northern blot analysis, SV40-MECs or MECs were cultured in gelatin-coated 10-cm-diameter culture plates and infected when cells were approximately 80% confluent. Medium was removed, and cells were washed with PBS and infected with 2 ml of virus suspension in serum-free medium for 2 h with occasional rocking. Ten milliliters of complete medium consisting of MCDB 131 medium supplemented with 30% human serum, 5 ng of EGF per ml, 1 μg of hydrocortisone per ml, 16 U of heparin per ml, 50 μM cyclic AMP, 2 mM l-glutamine, 100 U of penicillin per ml, and 100 U of streptomycin per ml was added, and the medium was changed every 2 days. For studies characterizing viral protein expression and virus production, MECs were seeded in collagen-coated eight-well Biocoat chamber slides (BD Biosciences, Bedford, Mass.) and incubated with 250 μl of inoculum in the presence of 1% polyethylene glycol and 8 μg of Polybrene per ml for 2 h at 37°C in M199 medium. The inoculum was removed, and Clonetics EGM-2-MV medium was added and changed on days 1, 3, 6, 9, 11, and 13 of the time course.

Immunofluorescence.

Cells were washed with PBS, fixed in 4% paraformaldehyde in PBS at room temperature for 20 min, washed, permeabilized for 10 min with 0.1% Triton X-100, and then washed with PBS. Nonspecific immunofluorescence was reduced by incubating cells with 20% normal goat serum and 3% bovine serum albumin in PBS, followed by washing three times in 1.5% bovine serum albumin-PBS. The antibody for LANA was preadsorbed with uninfected human umbilical vein EC extract for 1 h at 37°C and clarified by centrifugation at 12,000 × g for 2 min prior to use to reduce nonspecific binding. Both single- and dual-immunofluorescence analyses were done with anti-LANA (anti-ORF73, 1:250; Advanced Biotechnologies, Columbia, Md.), anti-PKR (B-10, 1:100; Santa Cruz Biotechnology, Santa Cruz, Calif.), and anti-PPF (anti-ORF59, 1:500; Advanced Biotechnologies) monoclonal antibodies. Secondary antibodies (1:500) were goat anti-mouse, goat anti-rabbit, or goat anti-rat conjugated to Alexa Fluor 568 and Alexa Fluor 488 (Molecular Probes Inc., Eugene, Oreg.). The nuclei were counterstained with 4′,6′-diamidino-2-phenylindole (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 and epifluorescence optics. The emission patterns of the two or three fluorescent probes were collected separately and overlaid by MagnaFire software to create two- and three-color images. Quantitation of cell number and the percentage of cells expressing either LANA or PPF were done by direct counting of the total number of cells on the basis of DAPI counterstaining and the number of cells expressing LANA and PPF. For each sample, four random fields containing 50 to 250 cells per field were counted and presented as the mean ± the standard deviation.

RESULTS

Kinetics of viral gene expression and proliferation of infected MECs.

Differences in the amount of virus used to initiate infection of MECs with HHV-8 could have profound effects on the consequences of infection. The binding of viral particles to cell surface receptors activates the focal adhesion kinase, leading to a number of cellular changes that would not occur during latent replication (3, 59). Furthermore, viral infection stimulates many cells to secrete type I IFN and other cytokines (42, 48, 62, 88). Thus, differences in the percentage of cells that are initially infected might alter the susceptibility to secondary infection through paracrine factors. We initiated infection with low levels of infectious virus to explore the abilities of infected MECs to latently replicate and to produce infectious virus that could disseminate through the culture. MECs were infected with HHV-8 with inocula that differed by 10-fold. The cells were not passaged during these experiments, and the medium was changed every 2 to 3 days. MECs infected with HHV-8 were morphologically altered with an increase in cell spindling and loss of contact inhibition in a time- and dose-dependent manner (Fig. 1A). These changes occurred more rapidly when cells were infected with the larger inoculum. Spindling also occurred when non-HHV-8-infected cells were stimulated with TPA, TNF-α, or dsRNA (data not shown), but loss of contact inhibition and growth to a high density were exclusively seen in HHV-8-infected cells (Fig. 1 and data not shown).

FIG. 1.

Morphological changes in MECs infected with HHV-8. MECs were seeded onto collagen-coated eight-well chamber slides and incubated with 250 μl of M199 medium containing 1% polyethylene glycol, 8 μg of Polybrene per ml, and amounts of HHV-8 that differed by 10-fold (1× and 10× inocula) for 2 h at 37°C. The inoculum was then removed and replaced with Clonetics EGM2-MV medium that was changed on days 1, 3, 6, 9, 11, and 13 of the time course. Magnification, ×200.

Next, we examined the kinetics of latent and lytic cycle events in infected cultures by immunofluorescence analysis. LANA is critical for persistence of HHV-8 DNA in latently infected cells (6, 24). It accumulates in heterochromatin-associated nuclear bodies, participates in replicating the viral episome, and tethers HHV-8 DNA to chromosomes, coupling HHV-8 replication to cellular replication so that the virus is maintained in daughter cells (6). LANA was identified as a punctate nuclear antigen (Fig. 2A to D), and it was also closely associated with the chromatin of MECs undergoing mitosis (Fig. 2C). The DNA polymerase processivity factor (PPF) encoded by ORF59 is expressed exclusively during the lytic DNA replication cycle of HHV-8 (16, 17, 89, 92). Immunofluorescence microscopy revealed that there was a clear difference at 7 days postinfection (dpi) in the percentage of cells that expressed LANA (red), PPF (green), and a combination of LANA and PPF (yellow) in cultures that were inoculated with a 10× compared to a 1× inoculum of HHV-8, but this difference was nearly gone by day 14 (Fig. 2A). Bright-field microscopy revealed that infection led to formation of HHV-8-infected foci that were high in density with some detached cells at time points when infectious virus was detectable in the medium (e.g., Fig. 1, day 10, 10× inoculum). These regions were more strongly immunoreactive for PPF than were the surrounding areas (Fig. 2B), indicating that the distribution of cells undergoing lytic replication was not uniform. Areas of latent and lytic replication were sometimes present within the same microscopic field (Fig. 2D). The nuclei of some cells contained structures that had the morphological appearance of replication compartments and expressed PPF (Fig. 2D, arrow), a protein that is a component of HHV-8 replication compartments (89).

FIG. 2.

Immunofluorescence analysis of HHV-8 protein expression in MECs. MECs were inoculated on day 0 with HHV-8 titers that differed by 10-fold in accordance with the protocol described in the legend to Fig. 1. Dual-immunofluorescence analysis was done with antibodies for LANA (red) and PPF (green), and DNA was fluorescently labeled with DAPI (blue). Regions that express both LANA and PPF appear yellow. (A) Cells that were infected with the 1× and 10× inocula were fixed on days 7 and 14. Magnification, ×200. (B) Focus of cells undergoing lytic replication that has many cells expressing PPF. Magnification, ×200. (C) LANA-positive cell completing mitosis, with LANA localized to chromosomes in both daughter cells. Magnification, ×1,000. (D) MECs within the same high-power field harboring HHV-8 in different replicative phases (latent and lytic). A cell with a replication compartment that is characteristic of virus production is indicated by the arrow Magnification, ×1,000. (E and F) The percentages of cells expressing LANA (E) and PPF (F) in DAPI-positive cells were calculated at each time point by immunofluorescence analysis. The data are presented as the mean ± the standard deviation.

The differences in the percentage of cells that expressed viral proteins correlated with differences in the amount of inoculum at early time points but not at later time points (Fig. 2E and F). When cells were infected with the larger inoculum (10×), some LANA-positive cells were evident within 1 day of infection, whereas they were not seen until 4 dpi when the smaller inoculum (1×) was used (Fig. 2E). The percentage of cells that expressed LANA reached a plateau of approximately 74% by 10 dpi when infection was initiated with the 10× inoculum. In contrast, LANA sharply increased from 16 to 51% between days 10 and 14 when cells were infected with the 1× inoculum (Fig. 2E). During this time period, there was an increase in the expression of PPF to 5% (Fig. 2F). Dual-immunofluorescence analysis indicated that a much higher percentage of cells expressed LANA than PPF at all of the time points examined, and many of the cells that expressed PPF simultaneously expressed LANA (Fig. 2 and data not shown).

Quantitation of HHV-8 production.

The expression of PPF indicated that some infected MECs were in the lytic phase of viral DNA replication. A number of events need to be successfully completed to produce infectious virus. There is no standard plaque assay for rapid quantitation of infectious HHV-8. We thus compared two different methods for the ability to accurately measure the amount of infectious virus: quantitative real-time PCR (Q-PCR) to measure the amount of encapsidated viral DNA (87) and the infection of the T1H6 reporter cell line that responds to HHV-8 Rta with transactivation of an RTA-dependent β-galactosidase reporter (38). Aliquots from multiple independent viral preparations were used to infect MECs. The percentages of MECs that expressed LANA at 2 dpi were determined by immunofluorescence microscopy and then plotted against the genome equivalents of encapsidated viral DNA in the inoculum as determined by Q-PCR (Fig. 3A) or the relative luminescence measured with the T1H6 reporter cell assay (Fig. 3B). The amount of encapsidated viral DNA that was used to infect MECs did not correlate with the percentage of cells that expressed LANA following viral infection (correlation coefficient, 0.36; Fig. 3A). In contrast, the luminescence units that were measured with the T1H6 reporter cell line directly correlated with the percentage of MECs that expressed LANA following viral infection (correlation coefficient, 0.99; Fig. 3B). This indicates that the T1H6 reporter cell line can be reliably used to assess the amount of infectious virus in different samples and was used for subsequent studies. The detection of infectious virus in conditioned medium lagged significantly behind the expression of PPF. On day 10, MECs that were infected with the 1× and 10× inocula displayed large differences in the amounts of infectious virus present in conditioned medium (Fig. 3C). This difference was no longer obvious at 14 dpi, when cells infected with the smaller inoculum were releasing as much infectious virus as were cells infected with the larger inoculum. This demonstrates that under standard culture conditions, a smaller amount of inoculum primarily delays the kinetics of infection.

FIG. 3.

Quantitation of infectious HHV-8. Viral stocks were prepared from different batches of conditioned medium from TPA- or butyrate-stimulated BCBL-1 cells. Each stock was assayed for the amount of infectious virus with both Q-PCR of encapsidated virion DNA and an assay measuring the RTA-driven β-galactosidase expression that resulted from infection of the T1H6 reporter cell line. Aliquots of each stock were used to infect MECs. At 2 days postinfection, cells were fixed and the numbers of MECs positive for LANA expression were determined by immunofluorescence microscopy. (A) The number of genome equivalents per microliter of virus stock was compared to the percentage of LANA-positive cells. Correlation coefficient = 0.36. (B) The β-galactosidase activity of each sample is plotted as the number of luminescence units per microliter and is compared to the percentage of LANA-positive MECs. Each point represents a separate viral stock. Correlation coefficient = 0.99. (C) The amounts of infectious virus detected in conditioned medium from MECs at 10 and 14 dpi were determined with the T1H6 reporter cell system. Values are the mean fluorescence ± the standard deviation without subtraction of the background fluorescence.

Effects of DNA polymerase inhibitor.

The production of infectious virus by MECs infected with HHV-8 raised the possibility that the increase in the percentage of infected cells during the time course was in part due to secondary virus infection. The viral DNA polymerase inhibitor PFA reduced the production of infectious virus to undetectable levels (Fig. 4A). This did not prevent the increase in the percentage of LANA-positive cells (Fig. 4B and C). PFA had no detectable effect on the increase in the percentage of LANA-positive cells that resulted from infection with the 10× inoculum that plateaued with more than 70% of the cells infected with HHV-8 by day 10 (Fig. 4B). The presence of PFA slightly reduced the magnitude of the increase in the percentage of LANA-positive cells that occurred between days 10 and 14 from the 1× inoculum (Fig. 4C) but not from the 10× inoculum (Fig. 4B), suggesting that part of the increase observed with the 1× inoculum resulted from de novo infection by virus produced by HHV-8-infected MECs. Incubation with PFA had no significant effect on the number of uninfected cells, and the increase in the cell number that accompanied infection with HHV-8 was not significantly altered when PFA was present throughout the incubation period (see Fig. 6B). These results indicate that most of the increase in the percentage of MECs that expressed LANA was due to replication of latently infected cells, with only a minor increase resulting from infectious virus produced by the MECs.

FIG. 4.

Inhibition of production of infectious HHV-8 with PFA. MEC were inoculated with 1× or 10× HHV-8 in the presence or absence of PFA (0.5 mM). PFA was present in the medium throughout the incubation. Cells were cultured in endothelial Bullet medium after the 2-h inoculation period. (A) The amount of infectious HHV-8 in conditioned medium at 10 dpi with the 10× inoculum was measured with the RTA-responsive T1H6 reporter cell line and reveals that there was no detectable virus present in the medium containing PFA. (B and C) The percentage of cells that expressed LANA was determined by immunofluorescence and is shown as a function of time, comparing cells incubated in the absence and presence of PFA that were inoculated with the 10× (B) and 1× (C) HHV-8 inocula. For each condition, the T1H6 reporter cell assay confirmed that there was no detectable infectious virus in conditioned medium from cells that were incubated with PFA.

FIG. 6.

Changes resulting from IFN-α-pretreatment or incubation with PFA in HHV-8-infected MECs. MECs either received no supplements, were pretreated with IFN-α for 24 h (1,000 U/ml), or were incubated in the continuous presence of PFA (0.5 mM). Cells were infected with HHV-8 with inocula that differed by 10-fold and were cultured in Clonetics EGM2-MV medium. (A) Cells that were infected with the 10× inoculum were analyzed at 10 dpi by immunofluorescence analysis with antibodies to LANA (red) and PPF (green). Nuclei were counterstained with DAPI (blue). Original magnification, ×400. (B) The number of cells per high-power field (original magnification, ×1,000) at 14 dpi was determined by microscopic examination. The mean ± the standard deviation from triplicate determinations is shown.

Protection of MECs from HHV-8 with IFN-α pretreatment.

IFN-α is well recognized as a cytokine that augments antiviral defenses by inducing the expression of antiviral proteins. We explored the ability of IFN-α to protect MECs from infection with HHV-8 by incubating cells with IFN-α for 24 h prior to infection with concentrations of IFN-α that induce peak levels of expression of multiple IFN-α-responsive genes in ECs (34). The IFN-α was then removed, and cells were infected with the 1× and 10× inocula of HHV-8. Priming with IFN-α dramatically decreased the ability of cells to support infection with HHV-8 (Fig. 5A and B). IFN-α-pretreated cells experienced very little increase in the percentage of LANA-positive cells over the entire 14-day incubation period when they were infected with the 1× inoculum, whereas LANA expression progressively increased in untreated cells (Fig. 5B). Infectious virus was not detected in the conditioned medium from IFN-α-primed cells infected with the smaller inoculum at any time point over the 14-day incubation period, and cells did not express detectable PPF (data not shown). When IFN-α-pretreated MECs were infected with the 10× inoculum, there was very little increase in the percentage of LANA-positive cells until 10 dpi, when 12% were LANA positive, compared to 74% of the nonpretreated cells (Fig. 5A). The percentage of LANA-positive cells then sharply increased to 42% on day 14. Cells infected with the larger inoculum produced much less infectious virus than did untreated cells when examined at 10 dpi but returned to the control level by day 14 (Fig. 5C). These results demonstrate that IFN-α pretreatment inhibits latent viral replication and lytic viral production, and the protection from viral infection was more prolonged when fewer cells were initially infected.

FIG. 5.

Reduction of infection with HHV-8 by pretreatment of MECs with IFN-α. MECs were pretreated with IFN-α (1,000 U/ml) for 24 h. The medium was changed, and MECs were infected with either a 1× or a 10× inoculum of HHV-8. After incubation with infectious virus for 2 h, cells were cultured in Clonetics EGM2-MV medium. The percentages of LANA-positive cells over time in response to the 10× inoculum (A) and the 1× inoculum (B) are shown. LANA expression was assessed by immunofluorescence. (C) The amount of infectious virus in the conditioned medium was determined with the T1H6 reporter assay. The medium was changed 24 h prior to analysis at each time point. Thus, the amount of infectious virus in the conditioned medium reflects the amount made during the 24 h prior to harvest.

The data presented in Fig. 4 and 5 demonstrated that incubation of MECs with IFN-α prior to infection with HHV-8 more effectively reduced the number of HHV-8-infected MECs than did continuous incubation with PFA. Not only were there fewer LANA-positive cells when cells were pretreated with IFN-α than when they were incubated with PFA, but the level of LANA expression in HHV-8-infected MECs was considerably lower in IFN-α-pretreated cells (Fig. 6A). In addition, the percentage of cells that expressed PPF was severely reduced in IFN-α-pretreated MECs (Fig. 6A and data not shown). The abilities of IFN-α pretreatment to protect MECs from infection with HHV-8 and to inhibit the expansion of latently infected MECs were accompanied by differences in the morphology and growth pattern of MECs. As shown in Fig. 1, individual cells were easy to identify when cells were mock infected, but cells exceeded confluence when they were infected with HHV-8, making them difficult to count. We nonetheless were able to get an estimate of the impact of infection on cell number by directly counting cells by microscopic examination. Infections initiated with 1× and 10× inocula of HHV-8 led to twofold and threefold increases in cell density, respectively, by 14 dpi compared to that of uninfected cells (Fig. 6B). Continuous incubation with PFA did not affect the HHV-8-induced increase in cell density, consistent with the progressive increase in LANA-positive cells during incubation with PFA. In contrast, IFN-α pretreatment prevented the increase in cell density that resulted when cells were infected with the 1× inoculum, and it reduced, but did not fully block, the increase that resulted from infection with the 10× inoculum. Neither continuous PFA nor IFN-α pretreatment had any detectable effect on the morphology or number of uninfected cells (Fig. 6B and data not shown).

Expression of the antiviral protein PKR in HHV-8-infected MECs.

Primary nonimmortalized MECs were used for the studies shown in Fig. 1 to 6. Several groups have reported that immortalization of MECs increases the percentage of cells that support infection with HHV-8 (47, 57). MECs immortalized with SV40 large-T antigen (SV40-MECs) maintain the morphology and cell surface markers that are characteristic of MECs yet fail to senesce (2). We infected SV40-MECs with HHV-8 and examined viral gene expression and the expression of a cellular antiviral gene over a 7-day time course. The mRNA for both PAN and v-cyclin/vFLIP were readily detected at 2 dpi, remained at similar levels on day 4, and declined slightly on day 7 (Fig. 7A). An important component of the cellular response to viral infection involves the enhanced expression of cellular antiviral genes, including those that encode PKR and 2′5′-OAS. Expression of 2′5′-OAS reached peak levels at 4 dpi and remained elevated on day 7. Thus, infection of SV40-MECs with HHV-8 led to the expression of both lytic (PAN) and latent (v-cyclin/vFLIP) viral transcripts, as well as increased expression of a cellular antiviral gene (that for 2′5′-OAS).

FIG. 7.

Changes in expression of antiviral genes in response to infection with HHV-8. (A) SV40-MECs were incubated with HHV-8 in serum-free medium for 2 h, and then complete medium was added that consisted of MCDB 131 medium supplemented with 30% human serum, 5 ng of epidermal growth factor per ml, 1 μg of hydrocortisone per ml, 16 U of heparin per ml, 50 μM cyclic AMP, 2 mM l-glutamine, 100 U of penicillin per ml, and 100 U of streptomycin per ml. At the indicated times, cells were harvested and RNA was isolated, size fractionated (15 μg per lane), and subjected to Northern blot analysis for the indicated genes. Ethidium bromide staining of the RNA is also shown. (B) SV40-MECs and MECs were grown under identical conditions. A single stock of HHV-8 was added in parallel to the SV40-MECs and the MECs. Cells were incubated with virus in serum-free medium for 2 h. Complete medium from the same stock was then added to all of the cells. At the indicated times, cells were harvested and RNA was isolated, size fractionated (15 μg per lane), and subjected to Northern blot analysis for the indicated genes. Separate experiments were sources of the data shown in panels A and B.

Immortalization with SV40 large-T antigen disrupts several key components of the innate antiviral machinery, including the function of p53 and Rb (12). We directly compared nonimmortalized MEC to SV40-MECs in order to determine whether the immortalization altered the kinetics of HHV-8 gene expression and cellular antiviral gene expression. SV40-MECs and nonimmortalized MECs were infected with equal amounts of HHV-8 under identical culture conditions in an experiment separate from that shown in Fig. 7A. MECs that were obtained from multiple donors were pooled to reduce differences resulting from genetic differences of the MECs. Ethidium bromide staining illustrates equal loading of RNA for the SV40-MECs with lower loading of RNA for the MECs. The exposure of the Northern blot for each probe was optimized to illustrate the differences between SV40-MECs and MECs. The immortalized cells differed considerably from their nonimmortalized counterparts in the response to viral infection (Fig. 7B). Nonimmortalized MECs had much higher levels of expression of the HHV-8 kaposin, v-cyclin/vFLIP, and PAN transcripts than did SV40-MECs. Conversely, HHV-8 induced much higher levels of expression of the cellular antiviral genes for 2′5′-OAS and PKR in SV40-MECs than in nonimmortalized MECs. The SV40-MECs that had higher expression of antiviral genes had lower levels of viral gene expression, suggesting that the antiviral proteins were reducing the expression of viral genes in SV40-MECs. These studies demonstrate that immortalization of the MECs dramatically altered the consequences of viral infection. They also suggest that enhanced expression of cellular antiviral transcripts might contribute to the reduced expression of both latent and lytic viral transcripts. Nonimmortalized MECs were used for all other studies.

The elevated levels of the PKR and 2′5′-OAS mRNAs that were observed when MECs were infected with HHV-8 could result from direct induction of the cellular genes that encode these proteins by HHV-8 or from induction by type I IFN secreted by virus-infected cells. To distinguish between these possibilities, we used an immunofluorescence assay to determine whether elevation of PKR was a direct response to viral infection. LANA-positive cells expressed elevated PKR levels compared to surrounding uninfected cells (Fig. 8B). Thus, infection with HHV-8 directly induced higher levels of expression of PKR in MECs. When HHV-8-infected MECs were incubated with IFN-α for the final 24 h of a 7-day incubation period, levels of PKR increased more dramatically in the virus-infected cells than in the surrounding uninfected cells (Fig. 8D). In addition, when mock-infected cells were incubated with IFN-α for 24 h, the induced levels of PKR were much lower than the levels seen in virus-infected cells (Fig. 8C and D). This demonstrates that viral infection not only directly induced PKR expression but also enhanced PKR induction by IFN-α. IFN-α during the final 24 h of a 10-day incubation period did not decrease the percentage of cells that expressed LANA, but it caused a slight reduction in the percentage of cells that expressed PPF and reduced the amount of infectious virus by about 20% (data not shown). This indicates that incubation of virus-infected cells with IFN-α for 24 h has much less of an impact on the HHV-8 infection of MECs than does incubation with IFN-α for 24 h prior to infection.

FIG. 8.

Expression of PKR in HHV-8-infected MECs. MECs were infected with HHV-8 with the 10× inoculum. Cells were cultured in Clonetics EGM2-MV medium. Some cells were also incubated with IFN-α (1,000 U/ml) for the final 24 h of a 7-day incubation period. On day 7, cells were fixed and analyzed for expression of PKR or LANA by immunofluorescence assay. Nuclei were counterstained with DAPI (blue). Magnification, ×400.

DISCUSSION

These studies indicate that when less than 5% of nonimmortalized MECs were initially infected with HHV-8, there was a time-dependent increase in the percentage of virus-infected cells that was not due to secondary infection. The latently infected MECs increased as a percentage of the population because they had a growth advantage over uninfected cells. Latent proliferation of HHV-8-infected MECs should not trigger signaling events that result from binding or internalization of virions, events that include activation of focal adhesion kinase, phosphatidylinositol 3-kinase, and other intracellular signaling pathways (3, 59). Thus, our approach that initiated infection with small amounts of infectious virus would minimize the numbers of cells that respond with activation of intracellular signaling pathways and production of cytokines. In contrast to our approach, several other groups used infectious strategies that generated 50 to 90% LANA-positive ECs within 24 h (31, 44, 47, 83). They did not observe the time-dependent increase of LANA-positive cells. In our study, frequent mitoses were seen in MECs expressing LANA, and LANA was comparably distributed in daughter cells, consistent with LANA's role in replicating viral episomes and controlling episomal segregation in daughter cells (6, 7, 24, 37). Cultures infected with HHV-8 exceeded confluence, with two- to threefold more cells in HHV-8-infected cultures than in uninfected cultures. While PFA was effective at preventing virus production, it had no impact on the number of LANA-positive cells in the cultures infected with the 10× inoculum. This indicates that the proliferation of latently infected cells was primarily responsible for the increase. Incubation of PFA reduced the magnitude of the increase in the LANA-positive population at late time points when cells were infected with the 1× inoculum, suggesting that virus release and de novo infection could contribute to the expansion of the LANA-positive cells, but its contribution was minor. LANA stimulates entry into S phase through its interactions with glycogen synthase kinase 3β that contributes to the stabilization of β-catenin (30). LANA has been shown to be sufficient to enhance proliferation of human umbilical vein ECs when expressed through transfection (85). LANA disrupts the function of p53 and Rb (29, 41, 69). The disruption of Rb should lead to unregulated cell proliferation owing to enhanced expression of E2F-regulated genes, many of which are used for DNA replication and cell cycle progression (75, 84). Other viral genes that are expressed during latency probably contributed to the growth advantage of the latently infected cells. v-cyclin activates CDK-6, leading to the phosphorylation of Rb (21, 32), thereby complementing the ability of LANA to disrupt the cell cycle checkpoint that is normally regulated by Rb. v-FLICE inhibitory protein blocks apoptosis (8, 81), contributing to the evasion of cellular antiviral defenses.

Augmentation of cellular antiviral defenses by pretreatment with IFN-α prior to infection with HHV-8 protected most of the population of MECs from supporting infection with HHV-8. The few IFN-α-pretreated cells that were infected were so low in number and expanded so slowly that they were not observed until much later time points. In the LANA-negative cells, viral entry might have occurred and initiated a cellular response, resulting in an abortive infection or apoptosis. In IFN-α-pretreated MECs, there was a minimal increase in the percentage of virus-infected cells for at least 10 dpi. The duration of the protection was affected by the amount of virus used to infect the MECs. HHV-8 was able to overcome the cellular defenses more quickly when infection was initiated with a 10-fold larger inoculum, an inoculum that initially infected less than 5% of the untreated cells. We have previously shown that pretreatment with IFN-α enhances the expression of PKR and 2′5′-OAS in ECs, yet activation does not occur in response to IFN-α alone (39). Activation of PKR occurs within 15 min of incubation with dsRNA in IFN-α-primed ECs, and initiation of apoptosis occurs within 2 h. De novo infection with HHV-8 leads to a temporal cascade of viral gene expression, and lytic replication is dependent on the expression of Rta (58, 86). Virus-infected MECs express Rta within 2 h of infection (44). Expression of Rta is accompanied by the expression of complementary transcripts (51), and these might provide sufficient dsRNA to activate both PKR and 2′5′-OAS. This suggests that the protection from HHV-8 that occurred when MECs were pretreated with IFN-α might relate to activation of PKR and 2′5′-OAS and sensitization to dsRNA-induced apoptosis. Activation of PKR by dsRNA leads to phosphorylation of eIF2α, a translation factor whose phosphorylation blocks the initiation of mRNA translation (23, 40). Activation of 2′5′-OAS by dsRNA leads to formation of oligoadenylates in 2′5′ linkage that activate a latent RNase, RNase L (27). Together, these two enzymes decrease the overall rate of protein synthesis of both viral and cellular proteins. The IFN-α-pretreated cells that were infected with HHV-8 expressed much less LANA than did nonpretreated cells, perhaps because of reduced protein synthesis in response to activated PKR and/or 2′5′-OAS.

The majority of MECs that were infected with HHV-8 expressed LANA, and cells that expressed LANA had elevated PKR protein levels compared to those of neighboring cells that did not express LANA. This indicates that infection of MECs with HHV-8 induced expression of this cellular antiviral protein. PKR was expressed at higher levels in HHV-8-infected MECs than in uninfected MECs that were incubated with IFN-α for 24 h. Thus, infection with HHV-8 was a more potent inducer of PKR expression in MECs than was incubation with IFN-α alone. Infection with HHV-8 also induced expression of 2′5′-OAS mRNA. The heightened expression of IFN-responsive genes in HHV-8-infected MECs is consistent with microarray studies that demonstrated that IFN-regulated genes were a major group of genes that were expressed at elevated levels in HHV-8-infected MECs (66). The latently infected MECS that expressed elevated PKR proliferated over the 14-day course of the experiment, suggesting that the PKR and other antiviral proteins were not sufficient to completely prevent latent replication of HHV-8 in MECs. This could occur if latently infected MECs lacked activators of PKR or if viral proteins suppressed activation. The cells that expressed LANA responded to IFN-α with an increase in expression of the antiviral protein PKR. Incubation of virus-infected cells with IFN-α for 24 h did not affect the percentage of cells that expressed LANA, but there was a slight reduction in the percentage that expressed PPF. This is consistent with studies with HHV-8-infected BCBL-1 cells, where IFN-α induced expression of PKR and 2′5′-OAS in cells undergoing either latent or lytic replication, but cells that entered the lytic phase of viral replication were much more susceptible to apoptosis in response to IFN-α than were latently infected cells (67).

Immortalization of MECs with SV40 had a profound effect on the cellular response to infection with HHV-8. Higher levels of antiviral genes were induced by HHV-8 in SV40-MECs than in their nonimmortalized counterpart, and this was accompanied by reduced expression of latent and lytic viral transcripts compared to cells that expressed less PKR and 2′5′-OAS. This suggests that the heightened expression of cellular antiviral genes reduced the level of viral gene expression in infected cells. A MEC-derived cell line immortalized with human papillomaviruses E6 and E7 maintains HHV-8 during serial passage (57). Human papillomaviruses E6 and E7 disrupt many of the same regulatory processes that are disrupted by SV40 large-T antigen, including the function of p53 and Rb (12). Thus, alterations induced by viral immortalizing proteins need to be considered when using virus-immortalized cells to study infection with HHV-8.

Initiation of infection of MECs with low titers of HHV-8 is similar to the initiation of infection that is likely to occur in the development of KS lesions. HHV-8 can circulate in plasma and in leukocytes in some patients with KS (35), and detection of HHV-8 in the blood is associated with the development of new KS lesions (13, 68). The amount of virus that circulates is relatively small. Thus, it is likely that a few cells are initially infected and undergo latent replication to generate lesions composed predominately of latently infected cells (26, 28, 78, 80). IFN-α represents an effective treatment for KS in some patients (45, 46). Our studies provide evidence that IFN-α is especially effective at protecting uninfected MECs from infection with HHV-8. In addition, we demonstrate that IFN-α reduces the number of HHV-8-infected cells that are undergoing lytic replication. HHV-8-infected cells undergoing lytic replication represent a small fraction of the virus-infected cells within KS lesions (26, 79, 80). They nonetheless are thought to be important in the disease process by expressing viral proteins that contribute to the phenotype of KS lesions through paracrine mechanisms (5, 22). The reduction of latent and lytic replication by IFN-α, combined with the protection of uninfected MECs from infection with HHV-8, represents a likely mechanism by which IFN-α induces remissions of KS.