Human Immunodeficiency Virus Replication in a Primary Effusion Lymphoma Cell Line Stimulates Lytic-Phase Replication of Kaposi’s Sarcoma-Associated Herpesvirus

Abstract

Human immunodeficiency virus (HIV) and Kaposi’s sarcoma-associated herpesvirus (KSHV) coinfect many individuals in North America and in parts of Africa. Infection with HIV is a leading risk factor for the development of Kaposi’s sarcoma (KS). In this study, we tested the hypothesis that HIV infection of common or adjacent cells would stimulate replication and spread of KSHV. Infection of a primary effusion lymphoma cell line by vesicular stomatitis virus type G-pseudotyped HIV type 1 led to a rapid induction of lytic-phase KSHV replication. Induction of lytic KSHV replication by HIV required active replication of HIV. The addition of the nucleoside reverse transcriptase inhibitor azidothymidine or the protease inhibitor indinavir to the culture prevented HIV spread and inhibited the associated induction of KSHV lytic replication. Lytic replication occurred in both HIV-infected and HIV-uninfected cells within the culture, and could be induced in uninfected cells via a soluble factor released from the HIV-infected cells. Transmission of infectious KSHV to an uninfected target cell was enhanced by HIV replication and was inhibited by antiretroviral drugs. These results may have implications for the pathogenesis and treatment of KS in individuals coinfected with KSHV and HIV.

Kaposi’s sarcoma (KS) is the most common neoplasm occurring in individuals with AIDS (18, 22, 23). In 1994, sequences of a novel herpesvirus now termed KS-associated herpesvirus (KSHV) or human herpesvirus 8 were identified within KS tissues by representational difference analysis (14). Since that time, work from multiple laboratories has established that virtually all KS tissues from both human immunodeficiency virus (HIV)-seropositive and -seronegative patients harbor KSHV sequences (3, 15, 27, 32). KSHV seroprevalence studies have indicated that persons at higher risk for KS have significantly higher KSHV infection rates than low-risk persons (29, 41, 43). KSHV infects the endothelium-derived spindle cells which are thought to be central to KS pathogenesis (2, 8, 9). These and other data have contributed to an emerging consensus that KSHV plays an important role in the pathogenesis of KS. KSHV has also been found in primary effusion lymphoma (PEL), a rare B-cell lymphoma most commonly seen in AIDS patients, and in multicentric Castleman’s disease (11, 42).

Many factors are likely to contribute to the pathogenesis of HIV-associated KS. Among these, Gallo and colleagues have described the important role of inflammatory cytokines in promoting the growth of KS spindle cells, which themselves release cytokines and angiogenic factors contributing to the development of KS (17, 20, 40). The HIV transactivator protein Tat has also been shown to influence the migration and growth of KS spindle cells (19). It is provocative that KSHV encodes homologs of cellular inflammatory cytokines such as vIL-6, vMIP-I, vMIP-II, and vMIP-III, as well as genes such as v-cyclin D, v-bcl-2, and v-GPCR, which may play a role in cellular proliferation. If indeed KSHV plays a causal role in KS and in PEL, then identification of factors which influence the spread and replication of this virus within the host are important to elucidate. One potential influence upon KSHV replication that has not been thoroughly studied is coinfection with HIV.

Here we present the results of experiments examining the effects of HIV infection upon replication of KSHV in a PEL cell line. This cell line, BC-3, was derived from a body cavity-based lymphoma in an HIV-negative patient and harbors KSHV in the absence of Epstein-Barr virus (4). We used pseudotyping of HIV to introduce the virus to the cells and observed that HIV could spread and replicate efficiently in this cell culture system. HIV replication potently induced lytic phase replication of KSHV in the culture and enhanced transmission of KSHV to uninfected target cells. HIV-induced KSHV lytic replication could be inhibited by antiretroviral drugs. We believe these results have bearing on the pathogenesis and management of KS in HIV-infected individuals.

MATERIALS AND METHODS

Cell lines.

BC-3 cells were obtained from the American Type Culture Collection (CRL-2227). 293_GN cells were obtained from Gary Nabel through the NIH/NIAID AIDS Reference and Reagent Program. PEL cell lines BC-3 and BCBL-1 and control cell lines Jurkat, MT-2, and Namalwa were maintained in RPMI 1640 medium (Atlanta Biologicals, Atlanta, Ga.) supplemented with 10% fetal calf serum (FCS), 2 mM l-glutamine, 100 U of penicillin per ml, and 100 μg of streptomycin per ml at 37°C in 5% CO₂. 293T and 293_GN cells were maintained in Dulbecco modified Eagle medium (DMEM; Atlanta Biologicals) supplemented with 10% FCS, 100 U of penicillin per ml, and 100 μg of streptomycin per ml at 37°C in 5% CO₂.

HIV-1 infection and pseudotyping with VSV-G.

Virus stocks were produced by calcium phosphate transfection of the infectious molecular clone pNL4-3 with or without the vesicular stomatitis virus envelope glycoprotein (VSV-G) expression plasmid pHCMV-G (45). Virion-containing supernatants were filtered through a 0.45-μm-pore-size HT Tuffryn membrane (Gelman Sciences, Ann Arbor, Mich.) and stored in aliquots at −80°C until needed. Virus stocks were normalized for virion content by a commercial assay for the virion major core protein p24 (Organon Teknika, Durham, N.C.). Infection of PEL cells with HIV-1 was achieved by incubating PEL cells with pseudotyped NL4-3 virus (20 ng of p24/10⁶ cells) for 2 h at 37°C. Cells were then washed in phosphate-buffered saline (PBS) one time and further incubated for 5 min at 37°C with 0.05% Trypsin-EDTA to ensure complete removal of pseudotyped NL4-3 virus on the surface of cells. Cells were washed twice with PBS and suspended in RPMI 1640 containing 10% FCS. Supernatants were sampled periodically to monitor HIV-1 virion production by using a commercial p24 antigen capture enzyme-linked immunosorbent assay (ELISA; Organon Teknika).

BC-3 infection with HIV by cocultivation.

Jurkat cells were infected overnight with HIV-1_NL4-3 (10 ng of p24/10⁶ cells), washed, and incubated at 37°C for 6 days in RPMI 1640 with 10% FCS. Cells were then harvested, washed twice in RPMI 1640, and subjected to gamma-irradiation with 10,000 rads for 30 min in a cesium-137 irradiator. Then, 10⁶ irradiated cells were cocultured with 10⁶ cells each of uninfected BC-3, BC-2, or Jurkat cells in separate wells of a 6-well tissue culture plate. Control wells contained 10⁶ irradiated HIV-1 infected Jurkat cells alone. Cellular supernatants were sampled and assayed for reverse transcriptase (RT) activity by using a microassay format. Briefly, 10 μl of supernatant was added to 20 μl of an RT reaction cocktail (50 mM Tris-HCl [pH 7.9], 75 mM KCl, 2 mM dithiothreitol, 5 mM MgCl₂, 25 μg of poly(rA)-poly(dT) per ml, 0.05% NP-40, 50 μCi of [³H]dTTP per ml) in a 96-well plate. The plates were sealed and incubated for 2 h at 37°C. Next, 10 μl of the reaction mix from each well was spotted onto DE-81 paper. The DE-81 paper was washed three times with 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) and once with 95% ethanol, dried, and incorporated radioactivity assayed in a Matrix gas scintillation counter (Packard Instruments, Meriden, Conn.).

Cell-free infection of target cells by using a Transwell assay.

BC-3 cells were infected with pseudotyped HIV-1_NL4-3 as described above. At day 21 postinfection, 10⁶ HIV-infected BC-3 cells were transferred to the lower chamber of a six-well dual chamber tissue culture plate containing a 0.4-μm (pore-size) membrane (Transwell; Costar, Cambridge, Mass.). Target Jurkat, MT-2, BC-3, or BC-2 cells (10⁶ each) were added to the upper chamber. Six days later, the target cells were removed, washed two times in PBS, and transferred to a separate tissue culture dish. The supernatants from each well were collected over a 2-week period, and the RT activity was assayed as described above.

RT-PCR detection of lytic-phase transcripts.

Open reading frame 29 (ORF29)-directed RT-PCR assay was performed as described previously (36). Total RNA (500 ng) was reverse transcribed by using RT-PCR beads (Pharmacia Biotech, Piscataway, N.J.) containing Moloney murine leukemia virus RT and random hexanucleotide primers. The reaction mixtures were incubated at 42°C for 30 min and an additional 5 min at 95°C to inactivate the RT. A total of 5% of this cDNA product was then added to a 97.5 μl of PCR reaction mixture containing 2.5 U of AmpliTaq Polymerase (Perkin-Elmer), 1 mM concentrations of each deoxynucleoside triphosphate, 2 mM MgCl₂, 1× PCR buffer, and 2 μl each of control ribosomal S9 protein primers (Clontech Laboratories, Inc., Palo Alto, Calif.) or published ORF29 primers spanning both sides of open reading frames of ORF29A (GCA CGT AGC CAA CTC CGT G) and ORF29B (GCA GGA AAC TCG TGG AGC G). After 35 cycles (1 min at 95°C, 1 min at 58°C, and 1 min at 72°C) of amplification, the PCR product was analyzed via electrophoresis on 1.5% agarose–ethidium bromide gels. In the inhibition experiments 50 nM indinavir (Merck and Co., Whitehouse Station, N.J.) or 2 μM azidothymidine (AZT) was added to the cells at the time of induction or infection with HIV-1. The concentration of the drugs used in the experiment was sufficient to inhibit HIV-1 replication in MT-2 cells over a 3-week period (data not shown).

Field inversion gel electrophoresis (FIGE) analysis of KSHV DNA.

BC-3 cells were subjected to no treatment, treatment with tetradecanoyl phorbol acetate (TPA) at 20 ng/ml, or infection with VSV-G pseudotyped HIV-1 as described above. At 48 h post-TPA treatment or post-HIV infection, the cells were harvested and suspended in cell suspension buffer (10 mM Tris [pH 7.2], 20 mM NaCl, 50 mM EDTA). After equilibration of the cell suspension to 50°C, the cells were molded into agarose plugs (4 × 10⁶/plug) by adding an equal volume of 2% InCert agarose (FMC Products, Rockland, Maine). Plugs were incubated for 12 to 18 h at 50°C in proteinase K reaction buffer (100 mM EDTA [pH 8.0], 0.2% sodium deoxycholate, 1% sodium lauryl sarcosine, 1 mg of proteinase K per ml), washed four times in wash buffer (20 mM Tris [pH 8.0], 50 mM EDTA], and stored at 4°C until the assay was performed. FIGE was performed by using FIGE MAPPER (Bio-Rad) as per the supplier’s instructions. Samples were run on 1% pulsed-field gel electrophoresis-certified agarose (Bio-Rad) gels in 0.5× Tris-borate-EDTA at 150 V (forward) or 50 V (reverse), ramped from 3 to 30 s for 20 h at 15°C with a peristaltic pump that circulated buffer constantly through the FIGE Mapper Cell (Bio-Rad). Two different DNA size markers, which included Lambda ladder PFG and mid-range PFG markers (New England Bio labs, Beverly, Mass.) were run on either side of the samples for precise size determination of the linear DNA. Gels were transferred to Nytran filters (Schleicher & Schuell, Keene, N.H.) and hybridized with a radiolabeled KSHV v-cyclin D-specific probe. After the wash steps, the blots were exposed to radiographic film.

Antibodies and immunoprecipitation.

ORF59 monoclonal antibody 11D1 was generously provided by Bala Chandran (University of Kansas Medical Center, Kansas City). Sheep polyclonal antisera directed against ORF26 was kindly provided by J. Victor Garcia (St. Jude’s Children’s Research Hospital, Memphis, Tenn.). Rabbit antimatrix antiserum was produced in our laboratory by injecting New Zealand White rabbits with recombinant HIV-1 matrix protein expressed in bacteria. For immunoprecipitation analysis, 10⁷ uninduced, TPA-induced, HIV-infected BC-3 or BCBL-1 cells were labeled with 500 μCi of [³⁵S]cysteine-[³⁵S]methionine (ICN, Irvine, Calif.) in RPMI 1640 medium deficient in cysteine and methionine, supplemented with 5% dialyzed fetal bovine serum. After labelling for 20 h at 37°C, cells were washed twice in PBS and solubilized with lysis buffer (0.05 M Tris-HCl [pH 7.4], 0.15 M NaCl, 1% glycerol, 1% Triton X-100, 2 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 5 μg of aprotinin per ml, and 1 μg each of leupeptin, pepstatin, and antipain per ml), sonicated, and centrifuged at 100,000 × g for 1 h. Equal volumes of cell lysates were precleared with protein A-Sepharose (Pharmacia) at 4°C for 1 h and then immunoprecipitated by using sheep polyclonal antibodies specific for KSHV minor capsid protein or monoclonal antibody specific for KSHV lytic cycle associated ORF59 protein. After 1 h of incubation with antibodies and protein A-Sepharose at 4°C, the precipitates were washed and then suspended in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer. Samples were analyzed by SDS-PAGE and autoradiography. In Transwell experiments, a total 3 × 10⁶ cells were placed in the inserts, with control or HIV-1-infected BC-3 cells in the bottom wells (3 × 10⁶). After 3 days of incubation, the cells in the inserts were labeled with 250 μCi of [³⁵S]Trans Label (ICN). Cells were solubilized and immunoprecipitated by using methods described above.

Immunofluorescence analysis for HIV and KSHV gene products.

Immunofluorescence analysis for KSHV gene products was performed by using procedures described previously (24). Briefly, 10⁷ uninduced, TPA-induced (20 ng/ml, 48 h), or HIV-1-infected (48 h) cells were washed in PBS, spotted on slides, and air dried under UV light in a laminar flow hood. Cells were fixed in cold 70% acetone for 10 min. Fixed cells were incubated with PBS containing 0.1% bovine serum albumin–0.1% Triton X-100 (PBS-BT) for 10 min, followed by incubation with ORF59 monoclonal antibody (1:10 in PBS-BT) or rabbit anti-HIV-1 matrix antibody for 1 h at 37°C. Cells were washed in PBS with three changes of buffer and incubated further with Fluorlink Cy3-labelled goat anti-mouse immunoglobulin G (IgG; Amersham Life Sciences, Arlington Heights, Ill.) or Cy2-labelled goat anti-rabbit IgG (1:1,000 in PBS-BT) for another 1 h at 37°C. After a washing in PBS, slides were mounted with Fluorsave reagent (Calbiochem, La Jolla, Calif.) and examined for specific fluorescence under 20× or 100× oil immersion objectives. In some experiments, spread of HIV and KSHV lytic replication within an infected BC-3 cell culture was monitored by immunofluorescence. To do this, BC-3 cells were infected with VSV-G-pseudotyped HIV-1_NL4-3 as described above (5 ng of p24/10⁶ cells) and a fraction of the infected cells collected at 1, 3, 5, 7, and 10 days postinfection for fixation and immunofluorescence analysis by using the antibodies described above. Fluorescent images were acquired by using a Zeiss fluorescence microscope equipped with a digital camera, and the numbers of cells stained with Cy2, Cy3, or both fluorophores were counted. Ten low-power fields (20× objective) were counted for each fluorophore at each time point.

KSHV transmission and inhibition by antiretroviral drugs.

293 cell infections with concentrated KSHV virions were performed according to methods described previously (36). Cell supernatants from uninduced, TPA-induced, or HIV-infected BC-3 cells were centrifuged at 100,000 × g for 2 h at 4°C. The viral pellets were suspended in serum-free DMEM (1/30 or original volume), and 1 ml of the concentrated virus was then used to infect 5 × 10⁵ cells (plated a day before) in a 10-cm dish for 8 h at 37°C. Cells were washed twice with PBS and further cultured in DMEM with 10% fetal bovine serum. Cells were treated with TPA (10 ng/ml) for 8 to 12 h prior to harvesting. Total RNA was extracted at 72 to 96 h after inoculation by using TRIzol (GIBCO Life Technologies). Antiretroviral drugs were added to BC-3 cells at the time of infection by pseudotyped HIV-1. AZT was added to the media to achieve a final concentration of 2 μM; indinavir was added to a concentration of 50 nM. Total RNA from KSHV infected 293_GN cells was used in an ORF29 RT-PCR assay according to the methods described above. The PCR products were separated via electrophoresis on a 1% agarose gel, transferred to nylon membrane (Zeta-Probe blotting membranes: Bio-Rad Laboratories, Hercules, Calif.). DNA on nylon membranes was UV cross-linked and hybridized to a probe generated by PCR by using a previously described nested pair of ORF29-specific oligonucleotides 29Bi (CTG ACG AGT TCA CGG ATG) and 29Ai (TAC ACG CGA CCC GGA GGA) (36). The probe was ³²P-labeled with the Rediprime II random prime labeling system (Amersham Life Sciences). Membranes were hybridized with labeled probe for 4 h at 65°C in a Hybrisol solution (Oncor, Gaithersberg, Md.). Membranes were washed briefly in 2× SSC after hybridization and by vigorous agitation for 15 min each in the following solutions: 2× SSC–0.1% SDS, 0.5× SSC–0.1% SDS, and 0.1× SSC–0.1% SDS. The last wash was performed at 65°C for high stringency. Membranes were exposed to Kodak Biomax-MS film.

RESULTS

HIV replication in a PEL cell line.

In order to facilitate the study of KSHV-HIV interactions, we sought to develop a cell culture system in which both viruses are present and in which HIV replication can take place. PEL cells are a convenient and well-studied source of KSHV (4, 12, 35). However, these cells are of the B-cell lineage and would not be expected to support HIV entry. Consistent with this description, our attempts to infect PEL cells with large amounts of cell-free CXCR4-tropic (X4) HIV or CCR5-tropic (R5) HIV failed (data not shown). Pseudotyping of HIV with VSV-G allows HIV to enter a wide variety of cell types (1, 33). Pseudotyped X4 virus was produced by cotransfection of 293T cells with HIV-1_NL4-3 proviral DNA and a VSV-G expression plasmid. Pseudotyped HIV was then applied to four PEL cell lines (BC-1, BC-2, BC-3, and BCBL-1) and two control B-cell lines (Daudi and Namalwa), and cultures were monitored for p24 antigen release. One cell line, BC-3, allowed significant productive HIV replication and spread in the culture (Fig. 1A). In some experiments, an initial peak in released virus was followed by a delay in virus release which coincided with significant BC-3 lysis (days 5 to 7, Fig. 1A), after which HIV replication in the culture increased. The infection of BC-3 cells by HIV was also confirmed by electron microscopy (data not shown). Daudi, Namalwa, and BCBL-1 cells were also infected with pseudotyped HIV-1, but released much lower levels of p24 (Fig. 1A). Interestingly, although the infection in BCBL-1 cells did not result in high-level p24 production, the pattern of p24 release also was suggestive of HIV spread within this PEL cell line (Fig. 1B). To test the hypothesis that cell-cell contact allowed the productive spread of HIV-1 in BC-3 cells, we next performed cocultivation experiments by using irradiated HIV-infected Jurkat T cells as the source of virus. Irradiated HIV-infected Jurkat cells maintained as controls in the absence of added cells failed to demonstrate release of detectable virus as measured by RT activity (Fig. 1C, solid triangles). Cocultivation of BC-2 cells with irradiated Jurkat cells similarly did not lead to release of detectable virus (Fig. 1C, crossed circles). As a positive control, cocultivation of irradiated HIV-infected Jurkat cells with uninfected Jurkat cells resulted in significant release of virus into the cellular supernatants (Fig. 1C, open squares). BC-3 cell cocultivation with irradiated HIV-infected Jurkat cells also resulted in release of significant amounts of virus into the supernatant, confirming the competency of BC-3 cells to propagate HIV-1 (Fig. 1C, solid diamonds). To confirm the previously described inability of cell-free virus to infect BC-3 cells and to further demonstrate production of HIV from BC-3 cells, Transwell experiments were performed. HIV-infected BC-3 cells were placed in culture wells and separated from target cells by a 0.4-μm-pore-size membrane. Target BC-3 and BC-2 cells failed to become infected with cell-free virus when exposed to released virus in this manner (Fig. 1D, solid diamonds and triangles, respectively). Target Jurkat cells and MT-2 cells were efficiently infected by virus which passed through the membrane (Fig. 1D, open squares and circles, respectively). These data indicate that infection of BC-3 cells may be initiated by either VSV-G pseudotyping of HIV or by cell-cell spread from infected T-cell lines and that spread of HIV within the BC-3 culture required viral transmission via cell-to-cell contact.

FIG. 1.

Productive replication of HIV-1 in PEL cells. (A) Release of p24 antigen from BC-3 and control cell lines infected with VSV-G-pseudotyped HIV-1. p24 release in the cellular supernatants was monitored by p24 antigen capture ELISA. (B) Release of p24 antigen from BCBL-1 cells. Data are from the same experiment as in panel A. Note different scale for p24 values. (C) Cocultivation of PEL cell lines with irradiated, HIV-infected Jurkat cells. A total of 10⁶ of the indicated target cells were cocultivated with 10⁶ gamma-irradiated and HIV-infected producer cells. The release of virus was measured by RT assay in counts per minute (cpm). Jurkat (I) indicates irradiated, HIV-infected controls alone. (D) Transwell experiment examining transmission of cell-free virus from HIV-infected BC-3 cells to target cells separated by a 0.4 μm membrane. Replication of HIV from susceptible or nonsusceptible target cells is indicated by RT activity (cpm).

HIV infection induces lytic-phase KSHV replication in BC-3 cells.

PEL cell cultures consist largely of cells harboring KSHV in latency, with a small minority of cells in untreated cultures undergoing lytic KSHV replication (38). Treatment with phorbol esters such as TPA rapidly induces lytic growth of KSHV, allowing analysis of KSHV replication and the production of infectious particles. To determine whether HIV infection altered the growth characteristics of KSHV, we infected BC-3 cells with pseudotyped HIV and monitored markers of lytic-phase replication over time. For comparison, TPA was used to induce KSHV lytic replication, and identical markers were assessed. TPA treatment resulted in the appearance of a spliced-gene RT-PCR signal (ORF29) representing lytic-phase replication (36). This marker was first apparent at 12 h by this assay and persisted at various time points for up to 48 h (Fig. 2A, TPA lanes). HIV infection also resulted in the appearance of this lytic-cycle marker, with a slight delay compared to TPA (Fig. 2A, HIV lanes). This marker was apparent at all times assessed thereafter, indicating continued activation of lytic KSHV replication by HIV.

FIG. 2.

Induction of lytic-phase KSHV replication by HIV infection. (A) Detection of KSHV ORF29 transcripts by RT-PCR. BC-3 cells were treated with TPA (left) or infected with pseudotyped HIV (right), and cells were harvested at 0, 6, 12, 24, 36, and 48 h posttreatment or postinfection. RNA was prepared from cell lysates, and a standard amount of total cellular RNA subjected to reverse transcription, followed by 32 cycles of PCR with ORF29-specific primers. The predicted ORF29 product is 300 bp and represents a spliced RNA product present only upon lytic replication of KSHV. Control RT-PCR reactions employing S9 rRNA primers are shown below. (B) Induction of linear KSHV DNA upon HIV infection. BC-3 cells were infected by pseudotyped HIV or treated with TPA as already described. Cells harvested at time 0 (control) or 24 or 48 h after treatment were molded into agarose plugs, separated on a native agarose gel by the method of Gardella (24), and probed for KSHV sequences by using a ³²P-labelled probe specific for the KSHV cyclin D homolog. The positions of linear and circular forms of KSHV DNA are indicated, with size markers (in kilobases) given on the left.

These results suggested that HIV replication in BC-3 cells induced active (productive) KSHV replication. We next analyzed the effect of HIV replication upon production of replicative forms of KSHV DNA. The predominant form of the KSHV genome in latency is in an extrachromosomal circular episome. Upon TPA treatment of PEL cells, the replicative, linear form of the KSHV genome is induced (37). BC-3 cells were therefore subjected to TPA treatment or to infection with pseudotyped HIV, and the viral DNA analyzed by FIGE followed by Southern blotting. TPA treatment allowed the detection of linear KSHV genome by 24 h after addition to the media (Fig. 2B, TPA lanes). HIV infection also resulted in the appearance of linear KSHV DNA (Fig. 2B, HIV lanes). This induction of linear DNA was not apparent at 24 h but was prominent at 48 h postinfection.

Lytic-phase KSHV proteins are produced after HIV infection.

In order to extend the evidence presented above that HIV infection of BC-3 cells results in the appearance of lytic-phase RNA and of linear viral DNA, we next examined two KSHV proteins associated with KSHV lytic replication. The ORF26 gene product is a 34-kDa protein derived from a viral sequence with strong homology to minor capsid proteins of other gammaherpesvirus subfamily members and has been used as a marker of lytic KSHV replication (34). The ORF59 gene product is a 50-kDa product of an early-late class KSHV transcript that is thought to be an accessory protein for viral DNA replication. This gene product is the target of antibodies that are widely used for monitoring KSHV lytic-phase replication (13). Using an antiserum to the ORF26 product, no specific band was apparent in uninduced BC-3 cells, while a specific band of the predicted size was apparent from lysates of cells stimulated with TPA or infected with HIV (Fig. 3A). Immunoprecipitation of the same lysates with control rabbit IgG failed to demonstrate this protein. Similarly, both HIV infection and TPA treatment resulted in the appearance of the ORF59 lytic-phase gene product when an antibody specific to this product was employed for immunoprecipitation (Fig. 3B). Thus, HIV infection of BC-3 cells results in the induction of lytic-phase KSHV transcripts, the induction of lytic-phase KSHV proteins, and the appearance of linear KSHV DNA.

FIG. 3.

Induction of KSHV lytic-phase protein synthesis by HIV infection. BC-3 or BCBL-1 cells were labelled with [³⁵S]cysteine-[³⁵S]methionine for 20 h and then harvested to generate control uninduced lysates for immunoprecipitation. Treatment with TPA or infection with pseudotyped HIV was carried out for 24 h prior to harvesting of samples represented in the TPA and HIV lanes. Cells were lysed, the nuclei were removed, and large nucleic acid fragments were pelleted by centrifugation prior to immunoprecipitation with antibodies specific for KSHV lytic gene products. Control antibodies matched for species and isotype but lacking specificity for KSHV products were utilized in parallel reactions to assess the specificity of the immunoprecipitation. (A) Induction of ORF26 gene product in BC-3 cells. Immunoprecipitation with rabbit antisera directed against the KSHV ORF26 gene product (minor capsid) is shown on the left. The results of immunoprecipitation with control rabbit IgG are shown on the right. Molecular mass markers are indicated on the left of the gel in kilodaltons. (B) Induction of ORF59 gene product in BC-3 cells. Immunoprecipitation was done with murine monoclonal antibody directed against the ORF59 gene product (left) or with control murine IgG (right). Molecular mass markers are indicated on the right in kilodaltons. (C) Induction of ORF59 gene product in BCBL-1 cells. Immunoprecipitation with the ORF59 monoclonal antibody was performed in BCBL-1 cells infected with VSV-G-pseudotyped HIV. HIV-infected BC-3 cells were immunoprecipitated identically and are presented on the left. Control lanes indicate immunoprecipitation of HIV-uninfected BC-3 and BCBL-1 cells. Molecular mass markers are indicated at the left in kilodaltons (lane M).

Low-level replication of HIV occurred in BCBL-1 cells infected with VSV-G-pseudotyped virus (Fig. 1B). To determine whether the induction of lytic replication of KSHV by HIV was limited to BC-3 cells alone or could be seen in additional PEL cell lines, the induction of the ORF59 gene product was analyzed in HIV-infected BCBL-1 cells. A 50-kDa band was detected via immunoprecipitation from HIV-1-infected BCBL-1 cells in a manner identical to that of BC-3 cells (Fig. 3C). The induction of KSHV lytic replication by HIV infection is thus not limited to BC-3 cells alone.

KSHV lytic replication occurs in HIV-infected and in HIV-uninfected BC-3 cells and increases over time in a spreading infection.

In order to assess the mechanism through which HIV replication in a PEL culture induces lytic KSHV replication, we asked whether lytic replication occurred only in those cells expressing HIV antigens (HIV and KSHV coinfected cells) or if cells expressing no HIV antigens also undergo induction. BC-3 cells were infected with VSV-G-pseudotyped HIV as previously described for the experiment in Fig. 1A. Immunofluorescence microscopy was employed in this analysis with an antibody to the ORF59 lytic gene product to monitor KSHV replication and an antiserum against the HIV matrix protein (MA) to monitor production of HIV antigens. First, the induction of KSHV lytic replication in the culture by TPA and by HIV infection was assessed. Untreated cells demonstrated a low level of lytic-phase KSHV antigen (<1%), while both TPA and HIV resulted in 20 to 30% of the culture undergoing lytic replication by 24 h (Fig. 4A to C). Dual staining for ORF59 and for HIV MA revealed three populations of cells: those coexpressing MA and ORF59 product, those expressing ORF59 alone, and those expressing MA alone. A representative high-power field is shown in Fig. 4D to F. In this field, two doubly stained cells are shown and are indicated by arrows. More ORF59-expressing singly staining cells are present than singly stained cells expressing MA. It should be noted that no information regarding spatial relationships of the stained cells in Fig. 4 can be derived, since these cells are grown in suspension and pelleted onto glass coverslips for immunostaining.

FIG. 4.

Immunofluorescence microscopy of HIV-infected BC-3 cells. Cells were fixed on glass coverslips, stained with primary antibodies and secondary antibodies as indicated, and photographed on a Zeiss epifluorescence microscope. Secondary antibodies used with ORF59 primary antibody are shown in red (Cy3); antimatrix antibodies were detected by Cy2 secondary antibodies and are shown in green. Panels A to C and G to L were photographed under the low-power (20×) objective; panels D to F were photographed under the high-power (100×) objective. (A) Untreated cells stained with murine anti-ORF59 antibody. (B) Cells collected after 24 h of TPA treatment and stained with murine anti-ORF59 antibody. (C) Cells collected 24 h postinfection with pseudotyped HIV and stained with anti-ORF59 antibody. (D) Detection of ORF59 in a high-power field; arrows indicate cells which are dually stained and are present in panels D, E, and F. (E) Detection of HIV matrix protein in the same high-power field. (F) Overlay of images from panels D and E reveals dually stained (arrows) and individually stained cells in the culture. (G) Detection of HIV matrix protein in a spreading infection of BC-3 cells at day 1 postinfection. (H) HIV matrix protein detection at day 3 postinfection. (I) HIV matrix protein detection at day 7 postinfection. (J) ORF59 detection in the same spreading HIV infection of BC-3 cells at day 1 postinfection. (K) ORF59 detection at day 3 postinfection. (L) ORF59 detection at day 7 postinfection.

To further examine the spread of HIV in the BC-3 culture, we employed similar immunofluorescence methods but utilized a lower inoculum of virus (5 ng of p24/10⁶ BC-3 cells). Cell lysis in the first 4 to 7 days was noted to be significantly decreased with this inoculum (data not shown). The number of HIV-infected cells increased steadily over time, as indicated by the numbers of green cells at days 1, 3, and 7 (Fig. 4G to I). The number of BC-3 cells undergoing lytic replication also increased early in the course of HIV infection and remained well above background levels at day 7 (Fig. 4J to L). To compare the numbers of singly and dually stained cells during this spreading infection, the numbers of stained cells per 10 low-power fields were counted. Consistent with the data presented in Fig. 4G to L, the number of cells infected with HIV increased steadily over time (Table 1). Interestingly, while HIV-infected cell numbers increased steadily as measured up to 10 days postinfection, the number of cells undergoing lytic replication decreased dramatically after day 7 (Table 1). At each time point, the numbers of dually stained cells was significantly less than the numbers of cells undergoing lytic replication. These data confirm that HIV infection of BC-3 cells resulted in a productive, spreading infection within the culture and demonstrate that lytic replication of KSHV occurs during the course of the spreading infection. Furthermore, the number of ORF59-expressing cells exceeds that of the HIV-infected cells, suggesting that additional factors such as soluble factors released from HIV-infected cells may stimulate KSHV lytic replication.

TABLE 1.

Immunofluorescence analysis of HIV spread and KSHV lytic replication in HIV-infected BC-3 cell culture

Culture	No. of cells at:
	Day 1	Day 3	Day 5	Day 7	Day 10
HIV positivea	16	303	502	950	1592
ORF59 positiveb	143	604	847	1125	25
Dual positivec	9	129	178	290	7

^aHIV-infected BC-3 cells were identified by immunofluorescence microscopy by using anti-MA polyclonal antibody as described in Materials and Methods. Shown are the total number of fluorescent cells detected in 10 low-power (200×) fields examined at each time point. 

^bORF59 positive cells were identified by immunofluorescence microscopy by using an anti-ORF59 monoclonal antibody as described in Materials and Methods. Shown are the total number of ORF59 positive cells detected in 10 low-power (200×) fields examined at each time point. 

^cCells which stained positive for both MA and ORF59 by using the methods described above were identified by overlaying the images obtained from 10 low-power (200×) fields from each time point and counting dually stained (yellow) cells. 

A soluble factor(s) released from HIV-infected BC-3 cells induces KSHV lytic-phase replication.

The doubly stained cells described above provide evidence that HIV infection of cells harboring KSHV can induce lytic replication within the individual coinfected cells. However, the predominance of cells expressing ORF59 in the absence of detectable HIV antigen suggested that mediators of lytic replication may be produced upon HIV infection, which then act upon the HIV-uninfected cell population. To address this hypothesis, we utilized a cell insert culture system in which HIV-infected BC-3 cells or HIV-infected MT-2 cells were placed in culture and separated from naive BC-3 cells in a lower chamber by a 0.4-μm-pore-size filter. The BC-3 cells in the lower chamber were then analyzed for the presence of markers of lytic KSHV replication after 24 h. As shown in Fig. 5, a lytic KSHV gene product (p50, the ORF59 gene product) was induced by a filterable mediator released by HIV-infected, but not HIV-uninfected, BC-3 cells. Remarkably, no induction of lytic replication occurred when MT-2 cells infected by HIV were placed in the upper chamber. The induction of lytic replication by the factor released from HIV-infected BC-3 cells was not due to HIV particles themselves, since pelleted HIV particles from infected BC-3 cells failed to induce p50 (data not shown).

FIG. 5.

Induction of KSHV lytic replication by a soluble factor released from HIV-infected PEL cells. BC-3 cells were infected with VSV-G-pseudotyped HIV and placed in the upper chamber of a culture plate insert with a 0.45-μm filter. MT-2 cells infected with HIV-1_NL4-3 were placed in the upper chamber of a separate insert. Untreated BC-3 cells were placed in the lower chambers. After 24 h, the inserts were removed and cell lysates were prepared from the cells in the lower chamber. Immunoprecipitation was performed with ORF59 monoclonal antibody as for Fig. 3. The cells in each upper chamber are indicated at the top of the lanes above the bar, with the target (BC-3 cells) below the bar. Minus signs indicate uninfected cells in the upper chamber; plus signs indicate HIV-infected cells. A control BC-3 well stimulated with TPA is shown in the far right lane. Molecular mass markers are indicated on the left of the figure in kilodaltons.

Antiretroviral drugs inhibit HIV-mediated induction of KSHV lytic replication.

HIV replication can be inhibited by drugs acting at one of several steps in the retroviral lifecycle. To test the hypothesis that inhibition of HIV replication would prevent the induction of KSHV replication described above, we chose two antiretroviral drugs commonly employed in the treatment of HIV-infected individuals, AZT and indinavir. AZT acts at an early step in the virus life cycle through inhibition of reverse transcription, while indinavir has no effect on early events but prevents the initiation of a second round of infection in culture by specifically inhibiting the viral protease. With the ORF29 RT-PCR product as a marker of KSHV lytic replication, AZT completely inhibited HIV-mediated lytic-phase induction of KSHV. In contrast, AZT had no effect upon KSHV replication in cells treated with TPA (Fig. 6, middle panels). A different effect was seen when the protease inhibitor indinavir was employed. No significant effect of indinavir was seen in TPA-stimulated cells. However, the induction of lytic replication in HIV-infected cells was inhibited by indinavir only at late time points after infection (36 and 48 h; Fig. 6, bottom panels). This result is consistent with the site of action of indinavir, since effects on replication in a spreading infection would be expected after 24 h. It is important to note that the initial induction of ORF29 message by HIV infection within indinavir-treated cells is only transient, since further rounds of HIV replication are prevented by this drug. This finding was confirmed by immunoprecipitation analysis, which demonstrated that the ORF59 gene product production in indinavir-treated and HIV-infected BC-3 cells was markedly decreased compared to untreated HIV-infected BC-3 cells (data not shown) and is further supported by the inhibition of KSHV transmission by indinavir (below).

FIG. 6.

Inhibition of HIV-induced but not TPA-induced KSHV lytic replication by antiretroviral drugs. VSV-G-pseudotyped HIV or treatment with TPA was utilized to induce KSHV lytic-phase replication, and results were monitored by ORF29 RT-PCR as previously described. Cells were harvested at 0, 6, 12, 24, 36, or 48 h posttreatment or postinfection. S9 rRNA primers were used as a control for the reverse transcription and PCR reactions. Shown are results after treatment with no antiretroviral drug (top panels), with AZT (2 μM) (middle panels), or with indinavir (50 nM) (bottom panels). In the drug-treated experiments, the drug was added at the time of TPA induction or HIV infection (time zero).

KSHV transmission is enhanced by HIV infection and inhibited by antiretroviral drugs.

Stimulation of some PEL cells, including BC-3 cells, by phorbol ester results not only in the induction of markers of lytic-phase replication but also in the production of infectious KSHV virions (38). Although cell culture systems for monitoring KSHV replication are thus far limited to low-level replication, passage through multiple rounds of replication can be achieved in cell culture and can be monitored by detection of KSHV transcripts in previously uninfected cells (36). We next sought to determine whether HIV infection of BC-3 cells resulted in the production of increased amounts of infectious virus as determined by enhanced transmission of KSHV to a target cell line. For this purpose, 293_GN, a cell line previously shown to propagate KSHV, was chosen as the target cell line. Uninduced, HIV-infected, and TPA-induced BC-3 cell supernatants were collected at 24 h posttreatment or postinfection, and virus particles were isolated by filtration and centrifugation. Resuspended virus was applied to 293_GN cells for 48 h prior to stimulating the targets with TPA. Transmission of infectious KSHV was then assessed by RT-PCR and Southern blotting by using primers and a specific probe for the ORF29 spliced product. A small amount of background transmission from uninduced cells, representing active replication in a minority of cells in uninduced culture, was detected in this manner (Fig. 7A, uninduced). However, transmission was significantly enhanced by HIV infection or by TPA treatment (Fig. 7A, HIV and TPA lanes). In a separate transmission experiment, BC-3 cells infected with HIV were simultaneously treated with the antiretroviral drugs AZT or indinavir. TPA treatment and HIV infection again resulted in readily detectable signal from target 293_GN cells, while the addition of either antiretroviral agent eliminated detectable transmission (Fig. 7B). Together with the results shown in Fig. 6, these results indicate that antiretroviral therapy can prevent HIV-mediated stimulation of KSHV lytic replication and can decrease the transmission of infectious KSHV virions to uninfected cells.

FIG. 7.

Analysis of KSHV transmission and the effect of antiretroviral drugs. KSHV lytic replication in BC-3 cells was uninduced or was induced by TPA treatment or HIV infection. At 24 h postinduction, cellular supernatants were harvested and filtered through a 0.45-μm filter. Viruses were pelleted at 100,000 × g for 30 min, resuspended in DMEM, and applied to 293_GN cells. After 48 h of incubation with virus, 293_GN cells were treated with TPA. Cell lysates were prepared, and an RT-PCR reaction for KSHV ORF29 was performed. RT-PCR products were separated on agarose gels, transferred to nylon membranes, and probed with a ³²P-labelled probe specific for ORF29. Pluses indicate RT-PCR reactions performed in the presence of RT, and minuses indicate control reactions performed in the absence of RT. (A) Transmission experiment indicating transmission by virus from uninduced, HIV-infected, and TPA-treated cells. (B) Transmission in the presence or absence of antiretroviral drugs. TPA and HIV lanes represent experiments performed as in panel A. “HIV + IND” indicates results of the addition of indinavir to BC-3 cells at the time of HIV infection at the same concentrations as those in Fig. 6. “HIV + AZT” indicates results of the addition of AZT to BC-3 cells. The rightmost lane is a probe of ORF29 RT-PCR product from TPA-induced BC-3 cells as a marker.

DISCUSSION

We report here that HIV replication within a cell culture harboring KSHV in latent phase can induce active replication of KSHV and lead to enhanced KSHV transmission. These results may have important implications for understanding the pathogenesis of KSHV infection in HIV-infected individuals. Although a causal role for KSHV in the pathogenesis of KS has not been proven, a substantial body of evidence indicates that KSHV plays a role in KS tumor development (reviewed in reference 2). The immune deficiency which develops in HIV-infected individuals is one factor which may contribute to KS tumor development in coinfected individuals. However, we suggest that HIV infection can influence KSHV gene expression and replication more directly, through alteration of KSHV gene expression and through enhanced transmission of KSHV to uninfected cells.

This report demonstrates that HIV can infect and replicate within tumor cells which harbor KSHV. A productive, spreading infection was initiated in this tissue culture system when HIV was introduced into cells by using envelope pseudotyping or by cell-to-cell spread. Cell-free virus was unable to initiate infection of BC-3 cells or other PEL cells, indicating that cell-to-cell transmission was responsible for the spread of HIV infection after its introduction. The replication of HIV also occurred in other PEL cell lines such as BCBL-1 after introduction of pseudotyped virus, but at levels less than that of BC-3 cells (Fig. 1B). The differences seen in transmission of X4 HIV between individual PEL cell lines are likely due to differences in cell-surface expression of HIV receptor and coreceptor molecules. We have detected high levels of CD4 expression on BC-3 cells by flow cytometry, while CXCR4 levels are very low (data not shown). In contrast, no detectable CD4 was found on the surface of BCBL-1 cells. The reason that cell-free nonpseudotyped HIV cannot infect BC-3 cells remains under investigation, but it may be related to the low levels of cell surface CXCR4. Here we have utilized the efficient spread of HIV in BC-3 cells to examine the effect of HIV replication upon induction of KSHV lytic replication. Lytic replication of KSHV was induced early in a spreading infection and continued until 7 days postinfection (Fig. 4).

It is important to note that lytic KSHV replication was induced both in cells infected with HIV and within HIV-uninfected cells within this culture system. Indeed, immunofluorescence microscopy revealed that many more cells were undergoing lytic replication than were infected with HIV (Fig. 4 and Table 1). Furthermore, a soluble factor produced from HIV-infected BC-3 cells and not from uninfected cells or from an immortalized T-cell line infected with the same HIV strain could stimulate KSHV lytic replication (Fig. 5). The identity of this soluble mediator of lytic replication is not known. However, these data suggest that HIV and KSHV may interact in complex ways. HIV infection of BC-3 cells may induce the production and release of a soluble KSHV gene product which then triggers a change from latent to lytic-phase replication in bystander cells. KSHV encodes a number of cytokine homologs which are produced during lytic replication and could potentially influence the activation state of surrounding cells, including vIL-6, vMIP-I, vMIP-II, and vIRF-1. Among these, vIL-6 has been shown to be involved in an autocrine fashion in controlling growth of PEL cells (but not in inducing lytic KSHV replication). Alternatively, the released soluble factor may be a cellular gene product whose production is influenced by HIV-KSHV interactions.

The HIV Tat protein has been suggested to play a role in KS tumor pathogenesis. Soluble Tat can be released from HIV-infected cells and taken up by nearby cells, including KS spindle cells (19). Tat has also been shown to activate KSHV replication when added to PEL cells or to peripheral blood mononuclear cells from some individuals with KS (25). However, in our experiments, HIV-infected MT-2 cells did not release a soluble factor capable of inducing lytic replication of KSHV in BC-3 cells, while HIV-infected BC-3 cells did. This would imply that, if Tat is the soluble factor, BC-3 cells release Tat more efficiently or somehow activate Tat in a manner not seen with infected T-cell lines. Although Tat is an important candidate gene product for future investigation, other HIV gene products will also require consideration. The differential inhibition of induction of KSHV replication seen with AZT versus indinavir provides some clues in this regard. While AZT treatment of HIV-exposed BC-3 cells completely inhibited lytic induction of KSHV, indinavir treatment did not inhibit induction until time points after 24 h of infection (Fig. 6). These data support the idea that HIV gene expression and not the entry process or VSV-G pseudotyping was required to induce KSHV replication. Furthermore, the effect on KSHV induction must require steps in HIV replication which occur between reverse transcription and virus assembly. Although this does not narrow the list of potential HIV gene products to analyze, it does support the hypothesis that active transcription of HIV gene products is required.

Our results demonstrating interruption of KSHV replication and transmission by antiretroviral therapy may have relevance to recent clinical reports of KS regression upon treatment with protease inhibitor-containing regimens (7, 10, 30, 39). KSHV transmission was prevented in our study by both AZT and indinavir treatment of HIV-infected BC-3 cells (Fig. 7). Although initial induction of KSHV lytic transcripts did occur in indinavir-treated cells, this finding was most striking within the first 24 h of infection and decreased thereafter (Fig. 6). We interpret this finding as indicating that HIV replication during the first round of infection initiated KSHV lytic replication in a limited number of cells but that indinavir treatment effectively prevented any further spread of HIV within the culture and thus prevented further induction of KSHV from occurring. The initial amount of KSHV induction indicated by the ORF29 message seen in indinavir-treated cells was not sufficient to stimulate production of infectious KSHV virions to levels above the sensitivity threshold of our KSHV transmission assay. In support of this interpretation is the low level of KSHV linear DNA seen 24 h postinfection of BC-3 cells, which increases thereafter (Fig. 2B, HIV lanes). It is likely that the amount of infectious KSHV produced and available at this 24-h time point is low, similar to that of one round of replication in indinavir-treated cells. Taken together, our results support the hypothesis that antiretroviral regimens which include RT and protease inhibitors may decrease production and spread of KSHV. If ongoing HIV replication contributes to KSHV gene expression and KSHV replication in coinfected human hosts, then inhibition of HIV replication should decrease KSHV replication. Furthermore, if alterations in KSHV replication or gene expression contribute to the maintenance and growth of KS tumors, then successful antiretroviral therapy could lead to arrest of tumor growth or tumor regression. Reports that antiretroviral therapy decreases KSHV viral load and that this decrease is associated with KS tumor regression support this contention (30, 44). It is important to note that in this model the effect of antiretroviral therapy on HIV replication indirectly reduces KSHV replication or gene expression and that by extension the anti-KS activity attributed to antiretroviral therapy is not due to effects of the drugs on KSHV itself.

HIV and KSHV coinfected cells have not been identified in coinfected individuals, and KS spindle cells are uniformly negative when examined for HIV DNA or RNA (21, 28). However, several cell types have been identified which are susceptible to both KSHV infection and HIV infection. The most prominent of these are monocytes, which have been clearly demonstrated to harbor KSHV in infected individuals and are infectable by R5 HIV strains (6). Furthermore, both viruses are commonly spread by the sexual route, can be detected in semen, and are thought to spread efficiently through anal intercourse (5, 16, 26, 31). We suggest that HIV and KSHV coinfection of common cells may occur in some cellular compartments and that this coinfection may alter the spread and viral load of KSHV within a coinfected individual. Detection of coinfected cells from patient samples may be difficult, especially if lytic KSHV replication and its attendant cell lysis rapidly follow. However, coinfection of common cells may not be required for HIV to exert important effects upon KSHV and upon KS tumors. Soluble factors released from adjacent HIV-infected cells may be sufficient to induce alterations in KSHV gene expression within monocytes or B lymphocytes harboring latent KSHV. It will be important in future studies to determine if HIV and KSHV infect common cells in vivo and to identify soluble mediators of KSHV lytic replication. Further studies with this cell culture model of HIV- and KSHV-infected cells may facilitate identification of additional factors relevant to the in vivo spread and replication of KSHV and to the pathogenesis of KS.