Induction of B Cell Hyperplasia in Simian Immunodeficiency Virus–Infected Rhesus Macaques with the Simian Homologue of Kaposi's Sarcoma–Associated Herpesvirus

A bone marrow (BM) aspirate was collected in preservative-free ammonium heparin (1,000 U/ml) from a 2 y, 202-d-old captive-reared rhesus macaque 1 mo after experimental inoculation with a macrophage-tropic variant of SIV_mac239. The BM aspirate was dispersed, and isopynic gradient-purified (Ficoll-Paque; Pharmacia) BM mononuclear cells were seeded into T-25 culture flasks and grown in the presence of endothelial serum-free media (GIBCO BRL) supplemented with 10% fetal bovine serum, 1% l-glutamine, 1% penicillin/streptomycin/neomycin, and 30 μg/ml endothelial cell growth supplement (GIBCO BRL). Primary BM mononuclear cell cultures developed cytopathic effects (CPE) after 10–12 d and were rapidly frozen in liquid N₂ and thawed, and then supernatants were clarified by centrifugation and filtered through a 0.45-μM membrane. The filtered BM mononuclear cell culture extracts were inoculated on primary rhesus monkey fibroblast cultures 24. Fibroblast cultures developing CPE were scraped free into medium, pelleted at 400 g, washed in PBS, and suspended in cold Ito and Karnovsky's fixative (2.5% glutaraldehyde, 0.5% picric acid, 1.6% paraformaldehyde, 0.005% ruthenium red) in 0.1 M sodium cacodylate buffer, pH 7.4, for 2 h. Ito and Karnovsky's–fixed cells were washed in cacodylate buffer, postfixed in 1% OsO₄ and 0.8% K₃Fe (Cn)₆ in cacodylate buffer for 1 h, rinsed in distilled H₂O, and prestained in 2% aqueous uranyl acetate for 1 h. Fixed and prestained cells were dehydrated in a graded series of acetone, embedded in Epon 812 epoxy resin, polymerized at 60°C, and sectioned at 60 nm on an MT 5000 ultramicrotome. Copper grid–mounted sections were stained with lead citrate and uranyl acetate and viewed on a Phillips 300 electron microscope.

Cell-free virus isolated from the primary BM mononuclear cell culture was serially passaged three times in primary rhesus fibroblasts to generate viral stocks, and the stocks were titered on primary rhesus fibroblasts before animal infection studies. For the isolation of vDNA, infected primary rhesus fibroblasts from two 850 cm² roller bottles were harvested and the cell debris removed by low speed centrifugation. The cell pellet was resuspended in culture medium, sonicated to release intracellular virus particles, and centrifuged to pellet cell debris. Virus was pelleted from supernatants by high speed centrifugation in a Beckman JA-14 rotor (Beckman Instruments, Inc.) for 1 h at 9,000 rpm and purified through a six-step sorbitol gradient ranging from 20 to 70% spun in a Beckman SW41 rotor for 2 h at 18,000 rpm. Gradient-purified virus was diluted in balanced buffered salts solution and pelleted through a 20% sorbitol cushion. The virus pellet was resuspended in Tris/EDTA buffer (TE; 10 mM Tris/HCl, pH 8.0, and 1 mM EDTA) and lysed in TE with 0.6% SDS and proteinase K (2 mg) at 37°C for 5 h. vDNA was isolated by CsCl₂ gradient centrifugation in a Beckman Ti 75 rotor at 38.4 × 10³ rpm for 72 h, collected, and dialyzed against TE.

PCR amplification of the herpesvirus polymerase gene was performed on purified vDNA using the following degenerated herpesvirus (HV) primers: HV-5, 5′-GCTCTAGATT(C,T)GA(C,T)AT(A,C,T)GA (A,G)TG(CT)-3′; and HV-6, 5′-GGATTCGG(G,A)AA(C,T) TC(G,A)TA(A,C,G,T)AC(A,C,G,T)TC-3′. The conditions for PCR amplification were as follows: 94°C for 2 min (1 cycle); 94°C for 1 min, 37°C for 1.5 min, 3-min ramp to 72°C, 72°C for 1.5 min (10 cycles); 94°C for 1 min, 50°C for 1 min, 72°C for 1.25 min (30 cycles); and 72°C extension for 10 min (1 cycle). Each PCR reaction used 0.1 μg of DNA, 50 pmol of each primer, 1 U of Vent polymerase, 40 μM of each deoxynucleotide triphosphate, 10 mM KCl, 10 mM Tris/HCl, pH 8.8, 10 mM (NH₄)₂SO₄, and 0.1% Triton X-100 in a final volume of 50 μl. Amplification was performed in a Perkin-Elmer 2400 Thermocycler. PCR products were run out on a 1% agarose gel; DNA fragments were isolated and restriction endonuclease digested with XbaI and EcoRI for cloning into pSP73 (Promega Corp.). Included in the degenerate PCR assay were DNA samples from rhesus CMV- and rhesus Epstein-Barr virus (EBV)-like lymphocryptovirus–infected cells 24,25. DNA sequencing was performed with Applied Biosystems (ABI) PRISM Dye Terminator Cycle Sequencing Ready Reaction Kits with AmpliTaq DNA polymerase as per the manufacturer's instructions. The DNA sequence was then determined on an ABI 373A sequencer in the Molecular Biology Core at the Oregon Regional Primate Research Center (ORPRC, Beaverton, OR). All plasmids were sequenced in both directions to verify DNA sequences. Open reading frames (orfs) were identified with MacVector (version 5.0; Oxford Molecular Group), and amino acid sequences were aligned by the CLUSTAL method 26 used with the MEGALIGN program provided with DNASTAR biocomputing software for the Macintosh.

Viral genomic DNA (5 μg) was restriction endonuclease digested with EcoRI, run out on a 1% agarose gel, and transferred to nitrocellulose. For use as a probe, the KSHV thymidylate synthase (TS) gene was generated by PCR amplification from DNA isolated from BCBL-1 cells 27 with oligonucleotide primers corresponding to a specific region of the KSHV genome encoding TS (nucleotides 20,236–20,650) 28 using standard PCR conditions. Immobilized vDNA was probed with random-primed, [³²P]dCTP-labeled KSHV TS for 16 h at 42°C. Hybridized membranes were then washed using conditions of moderate stringency (0.2× SSC and 0.1% SDS at 50°C) for 1 h, and bound probe was visualized by exposing DuPont-NEN reflection film to the washed membrane at −80°C with a DuPont-NEN reflection screen. The corresponding vDNA fragment found to hybridize was isolated and further digested with restriction endonuclease BamHI, which yielded a 3.4-kb fragment that was found to hybridize to the KSHV TS probe. This fragment was cloned into pSP73 and subjected to DNA sequence analysis as described above. Orfs were identified using MacVector software for the Macintosh, and the orfs were compared with SwissProt and PIR protein databases for homologous orfs with BLASTX. Alignments to identified proteins were generated either by the GAP program from the Wisconsin GCG analysis package (version 9.1; Oxford Molecular Group) or by the CLUSTAL program provided with the DNASTAR program. Sequences for RRV 17577 IL-6, TS, and MIP are available from EMBL/GenBank/DDBJ under accession number AF087411.

DNA extractions from peripheral blood and tissues were performed with a Puregene DNA isolation kit essentially as described by the manufacturer (Gentra Systems). PCR amplification for detecting the presence of the RRV genome was performed with the following oligonucleotide primers: vMIP-1, 5′-CCTATGGGCTCCATGAGC-3′; and vMIP-2, 5′-ATCGTCAATCAGGCTGCG-3′. PCR amplification for detecting the rhesus EBV genome was accomplished with oligonucleotide primers derived from published sequences of the latent membrane protein (LMP)-1 29. PCR amplification of the rhesus β-globin gene was used as a control for the amount and quality of DNA used for each reaction 30. With the exception of β-globin, the conditions for PCR were 94°C for 2 min (1 cycle); 94°C for 0.5 min, 50°C for 0.5 min, 72°C for 0.5 min (30 cycles); and 72°C extension for 5 min (1 cycle). For the rhesus β-globin PCR, the annealing temperature was raised to 55°C. Each PCR reaction used 0.1 μg of total DNA, 50 pmol of each primer, 1 U of Vent polymerase, 40 μM of each deoxynucleotide triphosphate, 10 mM KCl, 10 mM Tris/HCl, pH 8.8, 10 mM (NH₄)₂SO₄, 2 mM MgSO₄, and 0.1% Triton X-100 in a final volume of 50 μl. The PCR reactions were run out on a 1% agarose gel, transferred to nitrocellulose, and probed with a [³²P]ATP-labeled oligonucleotide primer specific for vMIP (vMIP-3, 5′-ATATTAAACACTCGCCGC-3′), rhesus EBV LMP-1, or β-globin. Hybridizations were performed overnight at room temperature in 6× SSC, 0.1% SDS, and 10 mg/mL Escherichia coli tRNA. Southern blots were then washed with 2× SSC and 0.1% SDS twice at room temperature, followed by two washes for 1 h in 2× SSC and 0.1% SDS at 47°C (vMIP), 45°C (β-globin), or 52°C (rhesus EBV LMP-1), and bound probe was visualized by exposing DuPont-NEN reflection film to the washed membrane at −80°C.

RRV 17577–infected cells were solubilized with 0.5% NP-40 and 1% sodium deoxycholate in PBS and clarified in a Beckman SW28 rotor at 23,500 rpm for 1 h at 4°C. The clarified supernatant was used as antigen for coating 96-well Maxisorp plates (500 ng/well; Nunc, Inc.). ELISAs were then performed essentially as described 31.

All aspects of the animal studies were performed according to institutional guidelines for animal care and use at the ORPRC. Rhesus monkeys (identification numbers 18483, 18503, 18540, and 18570) were inoculated intravenously with cell-free supernatants containing the equivalent of 5 ng of SIV_mac p27 prepared from Cos-1 cells transfected with a full length SIV_mac239 molecular clone 32. Before experimental SIV inoculation, PBMCs from the four monkeys were prescreened for in vitro susceptibility to SIV_mac239 infection essentially as described 33. 8 wk after SIV infection, animals 18483 and 18570 were inoculated intravenously with 5 × 10⁶ PFU of RRV 17577. Two additional animals, 19092 and 19286, were also inoculated with 5 × 10⁶ PFU of RRV 17577 as RRV-only controls. At sequential time points, animals were sedated with ketamine hydrochloride (10 mg/kg body weight) and given complete physical examinations. Venipunctures were performed, and EDTA- or preservative-free ammonium heparin–anticoagulated blood samples were collected and separated into plasma and PBMCs for analysis. Serial peripheral LN biopsies were performed, and PBMCs and dispersed LN mononuclear cells were cultured in 1640 medium (GIBCO BRL) supplemented with 20% fetal bovine serum, 1% l-glutamine, and 1% penicillin/streptomycin/neomycin to determine if continuously replicating or transformed cells were present. Plasma was monitored for SIV p27 during the first 4 wk after infection using the SIV p27 antigen capture ELISA (Coulter Immunology). Duplicate plating of 2 × 10⁵ PBMCs or LN mononuclear cells on primary rhesus fibroblasts and PCR analysis of PBMCs and LN mononuclear cell DNA was used for RRV isolation and the detection of RRV DNA, respectively. Animals that became moribund were killed, and complete necropsies were performed. Tissues were snap-frozen in liquid N₂ for PCR analysis or fixed in neutral-buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin for light microscopy. Fresh PBMCs and dispersed mononuclear cells from lymphoid tissues were collected for leukocyte subset analyses, primary culture, and virus isolation via cocultivation with primary rhesus fibroblasts or cryopreserved for PCR analysis.

Alterations in subsets were monitored sequentially after experimental inoculations by FACS™ analysis (Becton Dickinson). T lymphocyte subsets were labeled with OKT4 (CD4; Ortho Diagnostic Systems, Inc.) and B9.11 (CD8, Coulter Immunology) mAbs. B lymphocytes were labeled with B-Ly-1 (CD20; Coulter Immunology) mAb. The CD23 and CD40 B cell activation markers were identified with EBVCS4 (provided by Dr. B. Sugden, University of Wisconsin, Madison, WI) and G28-5 (provided by Dr. J. Ledbetter, XCYTE Therapies, Seattle, WA) mAbs, respectively 34,35. Total serum Ig (IgG and IgM) concentrations and κ and λ light ratios in total serum Ig were determined by nephelometry, and the Coombs' test (direct antiglobulin) was performed to determine if hemolytic anemia in one of the animals was immune mediated (Quest Diagnostics, Inc.).

Primary cell cultures established with BM mononuclear cells from macaque 17577 developed CPE in the fibroblast-like adherent cells after 10–12 d in culture and were harvested for passage. Primary rhesus monkey fibroblast cultures were inoculated with filtered extracts from the primary BM cultures. 10–12 d after inoculation, fibroblast cultures developed CPE characterized by granular focal lesions followed by cell rounding and sloughing that generated holes in the monolayer (Fig. 1 a). Transmission electron microscopic examination of the infected fibroblasts revealed 150–200-nm virus particles with characteristic dense cores resembling herpesviruses in the extracellular spaces as well as immature and mature virus particles in the nuclei and cytoplasm, respectively (Fig. 1 b). This herpesvirus isolate was subsequently passaged, purified, and titered on primary rhesus fibroblasts, and vDNA was purified from cell-free virions for molecular characterization.

The initial molecular characterization of the virus was conducted using degenerate PCR analysis targeting a conserved region of the vDNA polymerase gene (FDIEC-VYEFP) from beta- and gammaherpesviruses. Purified vDNA, which was capable of producing infectious virus from CaCl₂-transfected fibroblasts, was subjected to PCR, and the resulting fragments were cloned and sequenced. As controls, we included DNA from rhesus CMV– and rhesus EBV–infected cells as well as DNA from uninfected cells. This comparative analysis revealed that the virus was a gammaherpesvirus, as the deduced amino acid sequence was more closely related to rhesus EBV than rhesus CMV (Table). Inclusion of amino acid sequences from other known gammaherpesviruses, such as KSHV and HVS, supported this assignment. Moreover, comparative analysis with the recently described RRV strain H26-95 revealed that the herpesvirus isolate recovered from rhesus macaque 17577 was an RRV.

To further characterize RRV strain 17577, we screened vDNA, restriction endonuclease digested with EcoRI, by Southern blot analysis for the presence of a TS gene, an unspliced cellular gene homologue that is conserved in KSHV and HVS but not EBV. Using a KSHV TS gene fragment as a probe, a single DNA fragment was found to hybridize under conditions of moderate stringency (data not shown). This DNA fragment was subsequently isolated and further digested for subcloning and DNA sequence determination.

DNA sequence analysis of a 3.4-kb fragment revealed three orfs, two with striking homology to KSHV and a single orf having 100% identity with RRV and limited homology to cellular IL-6 (Table). One orf encodes a homologue of KSHV MIP-II with 30.9% amino acid sequence identity, as defined by the GAP program provided with the Wisconsin GCG analysis package, and the second orf encodes a homologue of both cellular and vTS, with 66 and 64.6% amino acid sequence identity with KSHV and HVS, respectively. The third orf revealed 100% identity with RRV strain H20-95 IL-6, supporting our initial finding that the 17577 virus is an RRV.

Rhesus macaque 17577 was humanely killed 503 d after SIV infection and found to have multicentric LPD involving both lymphoid and nonlymphoid tissues. Attempts to propagate continuously replicating or transformed cells from mononuclear cells dispersed from BM and LNs were unsuccessful. Histopathologically, there was marked follicular lymphoid hyperplasia in LNs, spleen, and BM. The follicular elements were dominated by giant, often irregularly shaped, secondary, reactive germinal centers. Follicular architecture in the cortices of LNs was accentuated by expanded cellular interfollicular zones dominated by plasma cells and vascular and stromal elements (Fig. 2, a and b). The medullary cords were densely populated by plasma cells, and arborizing cords of plasma cells extended into the paracortical zones. Follicular lymphoid hyperplasia was particularly striking in the BM (Fig. 2 c). The LPD lesions in nonlymphoid tissues were principally nodular, pleomorphic lymphoid masses composed of cytologically normal lymphocytes, plasma cells, and macrophage-like cells that effaced normal tissue architecture (Fig. 2d and Fig. e).

To determine if the LPD lesions in macaque 17577 harbored RRV, oligonucleotide primers specific to the RRV MIP homologue were designed to detect vDNA by PCR amplification and Southern blot analysis. By semiquantitative PCR analysis, vDNA sequences were found to be present at >590 copies/0.1 μg of BM-derived DNA (Fig. 3 a). As most animals held in captivity are naturally infected with RRV 18,36, BM aspirates were obtained from other animals to determine if RRV is a common resident of BM. BM samples were collected from two SIV-infected animals without peripheral lymphadenopathy, and two additional samples were obtained from animals that were killed due to disease unrelated to SIV infection or LPD. Analysis of these samples revealed no detectable signal, suggesting that RRV DNA in the BM-derived DNA from these four animals was either absent or below the limits of detection (Fig. 3 b). Finally, the BM-derived DNA from macaque 17577 was analyzed for rhesus EBV–specific sequences, as rhesus EBV has been shown to be associated with LPD in SIV-infected animals 37. No signal corresponding to rhesus EBV DNA sequences was detected (Fig. 3 c), indicating that rhesus EBV was not present in this lesion.

To test whether RRV strain 17577 can cause disease in its natural host, six juvenile rhesus macaques were selected for experimental inoculation studies. All six macaques were ∼1.5 yr old and shown to be negative for RRV infection over a 6-mo period by absence of RRV-specific serum antibody as measured by whole-virus ELISA, failure to detect RRV sequences in peripheral blood leukocytes by PCR analysis, and failure to isolate RRV from PBMCs cocultured with primary rhesus fibroblasts. As >75% of immunocompetent animals in our colony are seropositive for RRV by 3 yr of age and do not develop spontaneous LPD, we initially inoculated four of the animals with SIV_mac239 to provide a background of immunosuppression before experimental RRV inoculation. 8 wk after SIV infection, two of the four animals inoculated with SIV (18483 and 18570) were inoculated intravenously with 5 × 10⁶ PFU of cell-free RRV 17577. The two remaining SIV-infected animals (18503 and 18540) were maintained as SIV-only infection controls. Two additional animals (19092 and 19286) were infected with RRV 17577 to serve as RRV-only infection controls.

Sequential FACS™ analysis of PBMCs from the experimental animals revealed limited CD4⁺ lymphocyte depletion (mean values = 900–1,100/μl) 2 wk after SIV infection in the SIV-infected animals, followed by a rebound and sustained CD4⁺ lymphocyte counts in the normal range (mean values = 1,500–2,000/μl), whereas CD8⁺ lymphocyte counts increased slightly compared with the preinoculation time points. Significant differences in the number of CD20⁺ B lymphocytes were observed among the three groups of animals. As early as 6 wk after RRV infection, the number of CD20⁺ B lymphocytes increased dramatically in the two RRV/SIV-infected macaques (mean values = 2,900–3,360/μl) and persisted for 34 wk in animal 18483 and 14 wk in animal 18570 (Fig. 4, a and b). In contrast, the SIV-infected control animals exhibited persistently low numbers of CD20⁺ B lymphocytes (mean values = 200–300/μl) beginning 2 wk after SIV infection (Fig. 4c and Fig. d). The number of circulating CD20⁺ B lymphocytes increased slightly 2 wk after infection (mean values = 950–1,000/μl) in the RRV-only controls (Fig. 4e and Fig. f). Further FACS™ analysis revealed that the PBMCs from the two RRV/SIV-infected macaques expressed the CD40 activation marker coincident with CD20 but not CD23, which is expressed after rhesus EBV infection (reference 38; data not shown).

RRV was readily recovered from PBMCs of the two RRV/SIV-infected animals as early as 4 wk after RRV infection in macaque 18570 and 8 wk after infection in macaque 18483. Subsequently, positive virus cultures were obtained repeatedly from these animals for the duration of the study (Fig. 4, a and b). The PBMCs from the two RRV/SIV-infected animals were also shown to harbor RRV DNA 4 wk after RRV inoculation in macaque 18483 and 14 wk after infection in macaque 18570, as determined by PCR and Southern blot analysis for the vMIP gene. Interestingly, no substantial difference in cell-associated viral load was discernible in PBMCs throughout the course of infection by PCR or virus isolation (data not shown). The two SIV-only control macaques were consistently negative for RRV by both methods of detection (Fig. 4c and Fig. d). The RRV-only control macaques were positive for RRV by both methods of detection 2 wk after infection and remained positive through 6 wk after infection (Fig. 4e and Fig. f).

The humoral response to RRV infection was assessed sequentially using a whole-virus ELISA. One of the two RRV/SIV-infected animals failed to develop a measurable antibody response to RRV, whereas a low antibody response to RRV was detected in the second RRV/SIV-infected macaque 4 wk after RRV infection (Fig. 4, a and b). Both of the SIV-only controls remained seronegative for RRV for the duration of the study (Fig. 4c and Fig. d), whereas the two RRV-only controls developed strong sustained anti-RRV antibody responses beginning 2 wk after infection (Fig. 4e and Fig. f).

Early symptoms of SIV infection included fever, mild skin rash, and malaise, coincident with primary SIV p27 antigenemia in the four SIV-infected macaques by 2 wk after SIV infection. 10 wk after RRV infection, the two RRV/SIV-infected macaques developed marked peripheral lymphadenopathy, which persisted for >56 wk in macaque 18483 and 31 wk in macaque 18570. Additionally, the two RRV/SIV-infected macaques developed marked, persistent splenomegaly with spleens ∼10 times normal size. 30 wk after RRV infection, macaque 18570 developed severe hemolytic anemia that was unresponsive to treatment with prednisolone and was killed 286 d after RRV infection. At necropsy, there was marked generalized lymphadenopathy, hepatomegaly, splenomegaly, diffuse erythroid hyperplasia in the BM, and membranous glomerulonephritis. Analysis of necropsy tissues for vDNA indicated that virus was present in all lymphoid organs (Fig. 5). Of particular interest is the absence of vDNA in the BM, which also did not exhibit lymphoid hyperplasia observed in the BM of macaque 17577 from which RRV 17577 was isolated. Only slight transient peripheral LN enlargement was seen in the SIV- and RRV-only control macaques, and splenomegaly was not observed.

Histopathologic examination of peripheral LN biopsy samples revealed marked differences between the three groups. 27 wk after RRV infection, LNs from the two RRV/SIV-infected macaques had marked angiofollicular lymphoid hyperplasia characterized by giant, secondary, reactive germinal centers that were frequently irregular in shape and lacked distinct mantle zones. Medullary cords and paracortical areas were infiltrated by sheets of plasma cells. The interfollicular zones were accentuated and relatively hypocellular, with increased vascularity and plasma cell infiltration (Fig. 6, a and b). The lymphoid hyperplasia was persistent and multicentric. Virtually all of the LNs recovered from macaque 18570 at necropsy were profoundly enlarged. The morphologic changes were similar to those observed in the LN biopsy 27 wk after RRV infection, with the exception that the necropsy LN samples contained a mix of both secondary reactive and tertiary regressive germinal centers, with vascular hyalinization in most LNs. Plasma cells dominated the medullary cords, paracortex, and interfollicular zones. The spleen of macaque 18570 also contained large, irregularly shaped lymphoid follicles that lacked mantles and were composed of a mosaic pattern of hyalinized cell-poor areas and cellular areas of normally proliferating germinal center lymphocytes. Cords and islands of mature plasma cells were prevalent in the red pulp (data not shown). In contrast, LNs from the control SIV animals, collected 35 wk after SIV infection, had profoundly atrophied lymphoid follicles and hypocellular paracortices with condensed stromal elements characteristic of SIV-induced lymphoid atrophy (Fig. 6 c; reference 39). LN biopsies from the two RRV control macaques 9 wk after infection had normal histologic features (Fig. 6 d).

Immunophenotyping of LN mononuclear cells revealed significant differences in the number of CD20⁺ B lymphocytes in the LNs from the two RRV/SIV-infected macaques compared with LNs from macaques in the SIV- and RRV-only control groups. The four animals in the two control groups had twofold increases in the number of LN CD20⁺ B lymphocytes compared with preinoculation samples, whereas the two RRV/SIV-infected animals had fourfold increases in the number of CD20⁺ B lymphocytes compared with the preinoculation control samples and twofold increases over the postinfection samples from the two control groups (Fig. 7). Attempts to culture continuously replicating or transformed cells from dispersed LN mononuclear cells and PBMCs were unsuccessful.

Coinfection with SIV and RRV resulted in a marked increase in serum IgG (Table). Two RRV/SIV experimentally infected macaques, as well as macaque 17577 from which RRV strain 17577 was recovered, developed marked hypergammaglobulinemia (serum IgG >2,000 mg/dl). κ and λ Light chain levels in total plasma Ig increased from mean preinfection values of 96 and 121 mg/dl in 17577 and the two RRV/SIV experimentally infected macaques to 149 and 212 mg/dl, respectively. The pre- and postinfection mean κ and λ light chain ratios were 0.8 and 0.7, respectively, indicating that the increase in plasma light chain concentrations was not associated with appreciable skewing. Postinfection total plasma Ig κ and λ levels decreased in the single SIV-only animal tested (18503), also without appreciable skewing. Terminal serum from macaque 18570 was also positive in the direct antiglobulin (Coombs') test, strongly suggesting that the hemolytic anemia in this animal was immune mediated. Modestly elevated serum IgG (1,233 mg/dl) was noted in SIV-only control macaque 18540. Increases in serum IgG were not observed in the second SIV-only control macaque or the two RRV-only control macaques (serum IgG <700 mg/dl). Normal polyclonal electrophoretic profiles were obtained from all serum samples (data not shown).

In this study, we describe the isolation of RRV strain 17577 from an SIV-infected rhesus macaque that eventually developed LPD. Like the RRV isolate previously described by Desrosiers et al. 18, strain H26-95, the isolate described here encodes a vIL-6–like cytokine gene. Moreover, this isolate also contains a vMIP-like homologue with amino acid sequence identity to KSHV vMIP-II. Sequence analysis of the entire RRV strain 17577 genome has recently been completed and reveals that the genome is essentially colinear with KSHV and possesses similar genes with the potential for transformation, regulation of apoptosis, or growth control as KSHV 40. In particular, RRV possesses a gene, referred to as R1, in the same position as saimiri transforming protein and K1 of HVS and KSHV, respectively. BLAST analysis with R1 shows that the NH₂ terminus is similar to that of the Fc receptor, suggesting that R1 is a transmembrane receptor and therefore may be involved in signal transduction and transformation. RRV stain 17577 is also >99% identical at the DNA level to the regions of the RRV strain H26-95 genome available for analysis. The majority of base differences are found in the midregion of orf 8, glycoprotein B, and predicted to encode both conserved and nonconserved changes in the protein. Geographical variation in KSHV is increasingly documented in association with different forms of KS and KSHV infection rates and raises the possibility that sequence variation may relate to differences in virulence or transmission 41,42,43,44,45. Interestingly, genotypic differences in glycoprotein B of human CMV are postulated to be associated with increased virulence in BM transplant recipients 46. This same region has not been analyzed from different isolates of KSHV derived from AIDS-related KS, classical/endemic KS, or LPD, and the significance of the sequence variation in the RRV glycoprotein B gene is unknown.

RRV strain 17577 was recovered from the BM of a naturally infected rhesus macaque that developed profound, persistent angiofollicular peripheral lymphadenopathy after experimental infection with a macrophage-tropic variant of SIV_mac239. Subclinical multicentric involvement of lymphoid tissues and solid organs was found at necropsy 503 d after SIV infection. PCR analysis of BM tissue exhibiting lymphocytic hyperplasia revealed that RRV DNA was present in high copy number but not in BM from normal macaques or in macaque 18570. We speculate that the difference in viral load in the BM of animal 18570 may be due to the lack of LPD in the BM, compared with the florid LPD lesions observed in the BM of macaque 17577.

As a causal role for RRV in the development of these lesions cannot be established by PCR analysis alone, we undertook experimental inoculation studies in rhesus macaques to determine whether RRV is just a ubiquitous agent in nonhuman primates or a pathogen capable of inducing disease manifestations in its natural host. The absence of B cell hyperplasia in the SIV- or RRV-only controls and the presence of high RRV loads in hyperplastic lymphoid tissues from SIV and RRV–coinfected macaques indicate that B cell hyperplasia is strongly associated with RRV. A role for RRV in B cell hyperplasia or LPDs is consistent with the biology of most or all other gammaherpesviruses (KSHV, EBV, MHV-68, HVS). The index macaque and the macaques experimentally infected with SIV_mac239 and RRV 17577 developed profound, persistent, multicentric angiofollicular lymphadenopathy resembling the multicentric plasma cell form of MCD either in the absence or presence of HIV infection 47. The hypergammaglobulinemia in the index macaque and both experimental RRV/SIV-infected macaques and the fatal immune-mediated hemolytic anemia in one of the two experimentally RRV/SIV-infected macaques are also clinical features that closely parallel those observed in KSHV-infected AIDS patients and HIV-1–seronegative patients with MCD 48,49,50.

Both SIV and HIV-1 induce in vivo B lymphocyte hyperactivation, as shown by hypergammaglobulinemia, circulating immune complexes, and sometimes elevated levels of autoantibodies 49,51,52. However, the consort of profound hypergammaglobulinemia, immune-mediated hemolytic anemia, and progressive lymphadenopathy reported here are unusual disease manifestations in SIV_mac-infected macaques, and few cases are reported 20,39,53,54. The RRV infection status in these animals is not known, as RRV was only recently identified. The absence of hypergammaglobulinemia, immune-mediated hemolytic anemia, and persistent angiofollicular lymphadenopathy in our SIV_mac239-only and RRV-only infection control macaques suggest that RRV coinfection may be an important factor in the immunopathogenesis of retrovirus-induced immunodeficiency disease in macaques. This cofactor role has long been recognized for EBV in human AIDS patients, and a similar role is evolving for KSHV.

Despite the significant evidence of B cell hyperplasia in the peripheral blood, LNs, and spleens of the RRV/SIV coinfected animals, they do not appear to have overt clonal abnormalities in their B cells, nor were we able to establish continuously replicating primary B lymphocyte cultures from them. This differs considerably from rhesus EBV infection, where we have been able to generate continuously replicating cultures of rhesus EBV–positive B cells in vitro from PBMCs collected from naturally infected macaques (our unpublished results). The evolution of clonal abnormalities in LPDs is complex 55, and although monoclonality was not evident in the LPD manifested in the RRV/SIV-coinfected animals, minor clonal abnormalities may be present. Recent studies suggest that differential expression of rhadinovirus cytokine and oncogene homologues may play an important role in the development of B cell hyperplasia and other abnormal cellular proliferations 15,48,56,57. Although one of the experimental RRV/SIV-infected macaques developed fatal immune-mediated hemolytic anemia, none of the macaques developed lymphoma, and continuously replicating lymphoid cells could not be obtained from the RRV-positive, hyperplastic lymphoid tissues or peripheral blood. Immortalized KSHV⁺ cell lines have also not been established from healthy KSHV⁺ carriers or KS and MCD patients, with the exception of a continuously replicating KSHV⁺/EBV⁻ cell line obtained from the peripheral blood of an HIV-seronegative patient with PEL who exhibited a phenotype similar to PEL-derived cell lines 58. Thus, like KSHV, RRV may require other transforming events or cofactors for lymphomas to develop.

Consistent with the association between high RRV load and B lymphocyte hyperplasia reported here, we have recently shown that healthy RRV carrier macaques harbor RRV in their CD20⁺ B lymphocytes in vivo 59. Recent studies support the hypothesis that KSHV vIL-6 plays an important role in the pathogenesis of PEL and MCD 48,57, and a similar role for the RRV IL-6–like cytokine is under investigation. Preliminary analysis of the RRV IL-6–like cytokine reveals that it can functionally substitute for cellular IL-6 in IL-6 bioassays 60. IL-6 is a cytokine with several functions, including the ability to activate B cells into immunoglobulin-producing cells. More importantly, IL-6 dysregulation appears to be a factor in the development of MCD and multiple myeloma 61,62.

The experimental infection protocol used here, SIV_mac239 infection before RRV infection, was designed to provide RRV a replicative advantage, as most herpesviruses, e.g., CMV and EBV, manifest disease in immunocompromised hosts. Moreover, serological analysis of the animals at ORPRC and other primate facilities reveal that >90% of immunocompetent animals are seropositive for RRV. Despite the high prevalence rates, few animals develop LPD, implying that the host immune response is capable of controlling the virus infection. What role SIV_mac239 infection plays in the disease manifestations reported here is not clear. One possible mechanism is that SIV-induced immunosuppression may be important for sustained RRV infection and replication. Some have proposed a similar cofactor role for HIV-1 in AIDS-related KS and PEL, whereas others believe, at least in the histogenesis of KS, that HIV-1 invokes an inflammatory response, promoting an increase in inflammatory cytokines that sustain KS lesions 15. We believe that RRV is likely an opportunistic infection in the SIV-infected animals, given the absence of similar disease manifestations in animals experimentally inoculated with RRV or SIV alone.

Finally, in addition to providing an animal model for some aspects of KSHV pathogenesis, coinfection studies such as those described here should be important for AIDS pathogenesis. Preliminary results suggest that the presence of RRV attenuates SIV replication. The KSHV MIP-II protein has been shown to inhibit CCR-dependent HIV infection in vitro 63,64. Like HIV, genetically divergent strains of SIV utilize CCR5 and other orphan seven-transmembrane receptors for viral entry 65,66,67,68,69. With the availability of manipulatable molecular clones, we expect that RRV will be extremely useful as an animal model for studying aspects of both KSHV and AIDS pathogenesis. One significant advantage of RRV is that it can be propagated and plaque titered in cultured cells, allowing molecular dissection of genes postulated to be related to pathogenesis. Studies are in progress to determine the role of the vIL-6 homologue and other viral genes in KSHV-induced disease and in this animal model of B cell hyperplasia.

We thank Drs. Ann B. Hill, David Johnson, Jay Nelson, and Robert P. Searles for helpful discussions and Ms. Lori Boshears for assistance with the manuscript.

This work was supported by Public Health Service grants RR00163 and CA75922.