INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

An important scientific goal has been the development of small-animal models of human immunodeficiency virus type 1 (HIV) infection that simulate the stages of HIV disease in humans (32, 39, 46, 53). Even though severe combined immunodeficiency (SCID) mice transplanted with fetal human tissue (e.g., thymus or peripheral blood lymphocytes [reviewed in reference 47]) provide the means to study specific organ pathology induced by HIV in vivo, they do not fully represent the complex immunopathogenesis of HIV disease, which involves different tissues and occurs over an extended period of time, resulting in AIDS (48). Of the small animals tested as alternative models, rabbits have been proposed to be the most promising in certain respects (53). Infection with HIV, defined as the replication of detectable virus in the host and the development of antibodies to HIV, has been well documented in native rabbits in vivo (14, 20, 26, 36, 49). However, immunosuppression in HIV-infected rabbits has been observed inconsistently, and other clinical signs have been detected in rabbits only upon exposure to both human T-cell leukemia virus type 1 and HIV (36). The absence of AIDS-like symptoms in these animals may be explained by the low level of viral replication. Viral replication has been observed in native rabbit cells in vitro but never at levels approaching that in human cells (29, 37, 57). The production of mature and infectious viral particles relies on the accurate interplay of regulatory HIV proteins with cellular host factors (4, 12, 28, 30, 44, 54, 59). Unlike those in murine cells (56), cellular host factors in rabbit cells may support regulatory HIV proteins critical for viral transcription, suggesting that the main barrier to the replication cycle occurs before transcription, presumably at the level of viral entry (8).

Efficient HIV entry has long been recognized to require the human cell surface protein CD4 (43). However, expression of human CD4 does not render cells in mice and rabbits entirely permissive to the HIV replication cycle (10, 18, 42). The recent discovery that human chemokine receptors are essential cofactors for HIV entry into cells might explain the viral entry block in these cells (3, 9, 16, 17, 19). The principal chemokine receptors for HIV entry are CCR5 (3, 16, 17, 58), which mediates viral entry into macrophages (macrophagetropic viruses), and CXCR4 (19), which mediates entry into transformed T-cell lines (T-cell line-tropic viruses). Macrophagetropic viruses (using CCR5) predominate in HIV-positive patients over a long period of time (51) and may be responsible for sexual transmission of HIV (13, 41, 50, 60). In contrast, T-cell line-tropic viruses (using CXCR4) emerge in at least half of patients later in the course of HIV disease and have been associated with an acceleration of the immunodeficiency (11, 27, 33).

These findings suggest the possibility that a small-animal model of HIV disease could be generated by the expression of human CD4 and coreceptors in transgenic lines. However, in mice such transgenes were recently found to be insufficient to support HIV replication, which is likely to be the result of postentry restrictions (6). Alternatively, since earlier work suggested that the main restriction in rabbit cells may be at the level of viral entry, rabbit cells expressing human chemokine receptors and CD4 might permit robust viral replication and spread. The present study was performed to evaluate the feasibility of transgenic rabbits expressing human CCR5 and CD4 as a small-animal model of HIV infection. Using transfected cell lines, we compared the permissivity of rabbit-, mouse-, and human-derived cells for HIV infection after introduction of human CCR5 and CD4.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells. HeLa cells expressing human CD4 (HeLa-CD4, provided by B. Chesebro) (52) or both CD4 and CCR5 (HeLa-CD4/CCR5, provided by D. Kabat) (52) and 3T3 cells (a murine fibroblast cell line, provided by D. Littman) (16) expressing human CD4 (3T3-CD4) or both CD4 and CCR5 (3T3-CD4/CCR5) were grown as previously described (16, 52). To generate rabbit cells expressing human CD4, parental SIRC cells (a rabbit fibroblast-like cell line derived from the cornea of a normal rabbit with no detectable reverse transcriptase activity; American Type Culture Collection [ATCC CCL-60], Rockville, Md.) were transfected by the calcium phosphate method with an expression vector (pcDNA3; Invitrogen, Carlsbad, Calif.) encoding human CD4 and a neomycin resistance gene (pCD4neo [25]) and were selected in culture medium with 400 µg of neomycin (Life Technologies, Grand Island, N.Y.) per ml. To introduce human CCR5, SIRC-CD4 cells were transfected with an expression vector (LPXsrf-CCR5; see below) encoding an epitope-tagged form of human CCR5 (5) and a puromycin resistance gene and were selected in medium containing 0.5 µg of puromycin per ml. Clones for surface expression of either CD4 or both CD4 and CCR5 were screened by flow cytometry (see below). Human peripheral blood mononuclear cells (PBMC) from random donors and rabbit PBMC were recovered by Ficoll-Hypaque density gradient centrifugation (Histopaque 1077; Sigma Diagnostics). Human and rabbit PBMC were activated overnight by incubation in RPMI 1640 medium (Mediatech) with phytohemagglutinin (PHA; 5 mg/ml; Sigma) or PHA-M (2 ml of rehydrated PHA-M per 100 ml of culture medium; Life Technologies), respectively, per ml overnight. Human macrophages were isolated as previously described by Miller et al. (45).

Construction of the CCR5 expression plasmid. Plasmid pcDNA3-CCR5 (5) was digested with HindIII and XhoI, which released the epitope-tagged CCR5 fragment. This fragment was then ligated into the linearized retroviral vector LPXsrf, which had been digested with HindIII and SalI. LPXsrf (provided by A. DeFranco) is a Moloney murine leukemia virus-based retroviral vector containing a puromycin resistance gene.

Flow cytometry. Fluorescence-activated cell sorting (FACS) was performed as previously described (5), using a phycoerythrin-conjugated monoclonal antibody (MAb) against CD4 (Leu-3a; Becton Dickinson, San Jose, Calif.) and a murine MAb against the epitope tag of CCR5 (anti-FLAG MAb M1; Eastman Kodak Co., New Haven, Conn.).

Transfection. Transfection to assess CD4 downregulation by Nef was performed with SIRC-CD4 cells. Transfections for assessing auxiliary HIV gene function were done in SIRC-CD4, HeLa-CD4, and 3T3-CD4 cells in triplicate. All transfections were performed by the calcium phosphate method (Profection Mammalian Transfection Systems, Madison, Wis.). Testing was initiated 48 h after transfection as described below.

Viruses. Viral stocks of YU-2 and NL4-3 were obtained by transfecting 293T cells with the molecular clones pYU-2 and pNL4-3 (from B. Hahn and M. Martin via the NIH AIDS Research and Reference Reagent Program) (1, 40). Viral stocks of BaL (from R. Gallo, S. Gartner, and M. Popovic via the NIH AIDS Research and Reference Reagent Program) (23) were grown on macrophages, and viral stocks of JR-CSF (provided by B. Chesebro) (35) were expanded in PHA-activated PBMC from healthy donors. YU-2, BaL, and JR-CSF are macrophagetropic viruses which predominantly use CCR5 for viral entry. YU-2, in addition, can enter cells by using CCR3 (3, 9, 16, 52). NL4-3 has a tropism for T cells and uses predominantly CXCR4 for cell entry (52). High-titer HIV aliquots with viral capsid p24 values greater than 150 ng/ml were used for all inoculations.

Histology. Infected cells were fixed in 2% paraformaldehyde for 15 min, washed, and dried at room temperature. Cells were subsequently stained with hematoxylin-eosin and examined by light microscopy.

Northern blotting. Northern blot analysis was performed as described previously (38), using an -³²P-labeled fragment of full-length YU-2.

p24 measurements. The intracellular p24 assay measuring cell entry of HIV was performed as previously described (5). For secreted p24 antigen, 1 day after plating, SIRC, HeLa, and 3T3 cells expressing CD4 or both CD4 and CCR5 into a 96-well plate (10⁴ cells/well) were inoculated with HIV strains. Twenty-four hours after infection, the cells were washed three times to remove exogenous virus and the medium was replaced. Culture supernatants were harvested 2 days after infection and analyzed for p24 antigen content.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Nef, Tat, and Rev functions in rabbit, human, and murine cells. To determine whether SIRC cells support the functions of select regulatory HIV proteins, we studied the activities of Nef, Tat, and Rev. First, Nef promotes the endocytosis and lysosomal degradation of cell surface CD4 in human cells (22) through mechanisms that are incompletely understood (2, 24). This function of Nef was studied by transfecting SIRC-CD4 cells with expression vectors encoding Nef and cell surface CD8. Concomitantly supplied plasmids are typically taken up by the same cells. Thus, after staining with a MAb against CD8, gating on CD8⁺ cells selected cells most likely to be transfected with Nef. CD8⁺ cells were subsequently analyzed for CD4 downregulation. CD4 expression was decreased by approximately 60% in SIRC-CD4 cells transfected with a Nef-encoding vector, which is similar to the decrease in human-derived cells (21) (Fig. 1A).

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Rabbit cells support Nef, Tat, and Rev function. (A) To estimate CD4 downregulation by Nef, a vector encoding Nef and a vector encoding CD8 were cotransfected into SIRC-CD4 cells. Surface protein expression was detected by staining with a MAb against CD4 or CD8 followed by FACS analysis. Cotransfection was performed because plasmids typically enter the same cells. By gating on CD8+ cells, mainly cells transfected with pNef were analyzed for CD4 expression. (B) Tat function was measured in cells cotransfected with a reporter construct (pLTR-CAT) and a vector encoding Tat protein. (C) Rev function was measured in cells cotransfected with a reporter construct (pDM128) and a vector encoding Rev protein. All analyses were performed in triplicate. OD(405), optical density at 405 nm.

SIRC-CD4 cells transfected by pNL4-3 produce infectious viruses. To determine if SIRC cells can support production of mature HIV virions independently of viral entry, SIRC-CD4 cells were transfected with the proviral molecular clone pNL4-3 to circumvent the cellular entry step. SIRC-CD4 cells clearly supported viral production, although at levels somewhat lower than that of HeLa-CD4 cells transfected with pNL4-3 (Fig. 2A). In contrast, no p24 antigen was detected upon transfection of murine 3T3-CD4 cells (Fig. 2A). Functional integrity of virions produced by rabbit cells was shown by the production of p24 antigen of human PBMC exposed to supernatants from transfected rabbit cell cultures (Fig. 2B), while PBMC exposed to supernatant from transfected murine cells showed no signs of infection. Thus, rabbit cells, unlike murine cells, are able to carry out the basic postentry activities to produce mature virions. This finding implies that the major barrier to the HIV replication cycle in rabbit cells occurs before reverse transcription.

View larger version (12K): [in this window] [in a new window]   FIG. 2.   Rabbit cells produce infectious viral particles independently of viral entry. (A) Cells were transfected by the molecular clone pNL4-3. Two days later, supernatants were analyzed for the production of viral capsid p24 antigen. (B) To assess the functional integrity of virions produced by transfection, human PBMC were exposed to the supernatants of transfected cells, and p24 levels in supernatants of the PBMC were analyzed at different time points.

SIRC cells expressing human CD4 and CCR5 form multinucleated giant cells upon exposure to CCR5-dependent viral strains. To examine directly the effects of chemokine receptor and CD4 expression in rabbit cells on the HIV replication cycle, we generated SIRC cell lines that stably expressed CCR5 and CD4. Cell surface expression of these proteins was verified by flow cytometry (data not shown).

View larger version (100K): [in this window] [in a new window]   FIG. 3.   CCR5-dependent induction of syncytia in SIRC-CD4/CCR5 cells infected by BaL. Hematoxylin-eosin staining was performed 2 days after infection.

Viral gene expression in SIRC-CD4/CCR5 cells infected by YU-2. To assess the ability of rabbit cells to support viral gene expression upon infection by HIV, we measured viral transcripts in the cytoplasm of infected cells by Northern blot analysis. SIRC-CD4/CCR5 cells infected with YU-2 displayed the expected 2-, 4-, and 9-kb classes of viral mRNA, indicating intact viral promoter, transactivation, and splicing functions (Fig. 4). No viral transcripts were detected in cells expressing only CD4 that were exposed to HIV. The intensity of the hybridization signals was similar with RNAs obtained from infected rabbit or human cells expressing CD4 and CCR5 (Fig. 4).

View larger version (47K): [in this window] [in a new window]   FIG. 4.   HIV-specific hybridization transcripts obtained from total RNA of infected rabbit and human cell cultures. Two days after infection with YU-2, total RNA was isolated, resolved by electrophoresis, and transferred onto a Zeta GT probe. The blot was probed with an -32P-labeled fragment of full-length YU-2. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

View larger version (27K): [in this window] [in a new window]   FIG. 5.   Intracellular viral gene expression indicates vigorous viral replication. After fixation and permeabilization, SIRC cells infected with BaL were stained with MAb against intracellular p24 (p24i) antigen.

Production of mature virions by SIRC-CD4/CCR5 cells upon exposure to various CCR5-dependent viral strains. To assess the permissiveness of rabbit cells expressing CD4 and CCR5 to various CCR5-dependent viral strains, we also assessed viral capsid p24 production in the supernatant of infected cell cultures. SIRC-CD4/CCR5 cells and HeLa-CD4/CCR5 cells showed abundant secretion of p24 upon infection with BaL, YU-2, and JR-CSF, all of which are viral strains that use CCR5 as a coreceptor (Fig. 6A). The amounts of p24 antigen produced were similar in rabbit and human cells but varied depending on the viral strain used for inoculation. No infection of rabbit cells by NL4-3 was observed (Fig. 6A), which is consistent with the lack of human CXCR4 on these cells. No p24 antigen was produced by murine cells infected by these viruses, even in the presence of human CD4 and CCR5. These results indicate that the expression of CCR5 and CD4 on SIRC cells confers general susceptibility to infection with CCR5-dependent HIV strains but not with CXCR4-dependent strains.

View larger version (24K): [in this window] [in a new window]   FIG. 6.   Extracellular viral gene expression and production of mature virions in rabbit cells. (A) Two days after infection of cells with various HIV strains, culture supernatants were harvested and analyzed for p24 antigen. (B) To test the functional integrity of virions produced, human PBMC were exposed to the supernatant of transfected cells, and p24 levels in supernatants of the PBMC were analyzed at different time points.

Primary rabbit PBMC electroporated with pYU-2 produce infectious virus. To examine whether primary rabbit PBMC are able to produce infectious virus upon bypassing the entry step, activated rabbit PBMC were electroporated with pYU-2. Supernatants of three of six electroporations showed detectable p24 antigen production with a yield of 0.05 to 1.5 ng/ml, consistent with a very low transfection efficiency of primary rabbit PBMC. To assess whether virion particles produced by rabbit PBMC are functional, human PBMC were exposed to these supernatants. Indeed, these cultures produced abundant p24 antigen over time (Fig. 7), indicating that rabbit PBMC are able to support postintegration steps in the HIV replication cycle, leading to release of infectious virions.

View larger version (11K): [in this window] [in a new window]   FIG. 7.   Production of mature virions by primary rabbit PBMC. Human PBMC were exposed to supernatant of rabbit PBMC, which had been electroporated with pYU-2 2 days before. As demonstrated by the abundant production of p24 antigen in human PBMC, virions generated by rabbit PBMC were infectious.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study provides a clear demonstration that expression of human CCR5 in rabbit cells abolishes the entry block for HIV and permits HIV replication at levels similar to those in human cells. In addition, the virions produced by HIV-infected SIRC-CD4/CCR5 cells are intact and infectious as determined by several criteria. These findings imply that a transgenic rabbit model for CCR5 and CD4 may be feasible, although it is impossible to predict the levels of viral replication that would be achieved in vivo or the extent of disease that would be associated with such an infection.

The rabbit cell line that we used has no inherent susceptibility to infection by HIV. Nonetheless, this rabbit cell line appeared to have the necessary host-specific factors to support the regulatory HIV proteins that are critical for infectivity. First, Nef downregulated CD4 in SIRC-CD4 cells to an extent similar to that reported for human and murine cells (21), consistent with the notion that Nef-mediated downregulation of CD4 is neither species nor tissue specific (21). Second, the HIV LTR in SIRC cells was clearly transactivated by Tat, although somewhat less efficiently than in HeLa cells. Other reports have suggested that Tat function is intact in native rabbit T cells (8), implying that at least certain rabbit tissues express the cellular host factors needed for Tat function. Third, the expression of early and late viral gene products requires host cell factors to interact appropriately with Rev and its responsive elements in viral RNA (12, 54). The present transfection data indicated that Rev constructs functioned properly in rabbit cells, albeit somewhat less efficiently than in HeLa cells. Insignificant levels of Tat and Rev function were seen in murine cells in these experiments, as previously reported (56). Thus, our findings confirm and extend the observations of intact function of Tat and Rev previously reported in other rabbit cell lines (8) and, in addition, demonstrate intact Nef function in rabbit cells.

Evaluating the activity of HIV proteins is useful for delineating potential barriers to HIV replication in a cell line, but it provides no information about the ability of the cells to produce infectious viral particles. To test the ability of rabbit cells to support the HIV replication cycle independently of viral entry, we transfected the cells with the molecular clone pNL4-3. Indeed, consistent with earlier studies, they produced viral capsid p24 antigen at a lower level than did HeLa cells (8). The difference in p24 antigen production may reflect lower transfection efficiency in rabbit cells than in HeLa cells or lower efficiency of HIV regulatory proteins. Importantly, the virions produced by this method were infectious, as demonstrated by the positive infection of human PBMC by viral stocks prepared from these transfected cells. No viral capsid p24 was detected upon transfection of murine cells, and as expected, human PBMC exposed to supernatant from those cells were not infected.

The key experiments in our study were to determine if coexpression of CD4 and CCR5 would render rabbit cells permissive to HIV infection, including viral entry and production of virions capable of spreading infection. Indeed, SIRC-CD4/CCR5 cells infected with BaL, a macrophagetropic virus, formed massive multinucleated giant cells (syncytia) and showed complete disruption of the cellular structure, as was observed in infected HeLa-CD4/CCR5 cells. Similar destruction of the cell architecture was also observed after infections by the CCR5-dependent strains, JR-CSF and YU-2, but not after infection with the CXCR4-dependent strain, NL4-3. No cytopathic effects were observed in SIRC-CD4 cells infected with BaL. Thus, these results demonstrate that the main barrier to efficient HIV replication in rabbit cells is at the level of viral entry and is removed by the expression of human cell surface molecules. Our observations also substantiate the notion (3, 9, 16, 17, 52) that syncytium formation is induced not only by T-cell line-tropic HIV strains but also by macrophagetropic strains in the context of appropriate cognate coreceptors.

Northern analysis of SIRC-CD4/CCR5 and HeLa-CD4/CCR5 cells infected by YU-2 showed similar intensities of viral transcripts, which implies that similar amounts of viral transcripts were produced in these cells. In contrast to our findings, transcripts from HIV-infected native rabbit T cells were reported to be of much lower intensity than those from infected human T cells (8). We believe this difference reflects inefficient viral entry into native rabbit T cells rather than a lack of cellular factors supporting Rev function, although tissue-specific differences in host cell factors cannot be excluded as a factor in these studies.

Staining for intracellular viral capsid p24 antigen of BaL-infected SIRC-CD4/CCR5 cells also confirmed vigorous viral gene expression, which indicates permissivity to the HIV replication cycle and corroborates the histological findings. In contrast, neither SIRC-CD4 cells nor murine 3T3-CD4/CCR5 cells exhibited significant p24 expression.

The general susceptibility of SIRC CD4/CCR5 cells to CCR5-dependent strains was also verified by measuring p24 antigen production by cells infected by JR-CSF, YU-2, and BaL. As expected, no infection could be documented in SIRC-CD4/CCR5 cells infected with NL4-3, a T-cell line-tropic virus that uses CXCR4, indicating the selective susceptibility of these cells to CCR5-dependent strains. Since levels of virus production as assessed by p24 antigen levels were similar in rabbit and human cells, the lower efficiency of regulatory HIV protein function in rabbit cells does not appear to be critical for overall infectivity. The production of viral capsid p24 antigen in SIRC-CD4/CCR5 cells was 100 to 4,000 times higher (depending on the viral stain) than that in transformed rabbit T-cell lines lacking appropriate human chemokine receptors (8).

The final and most crucial test of virion integrity is the ability of virions to infect other cells. This was demonstrated by their successful passage to human PBMC. Thus, there are no absolute blocks or restrictions to the HIV replication cycle and the spread of infectious HIV virions in rabbit cells. One report has described the production of immature virions by primary rabbit PBMC infected by HIV (55). It is more likely that virions infecting primary rabbit PBMC lacking human chemokine receptors and human CD4 will be targeted to an unusual endocytic pathway, resulting in their disruption. Indeed, in the present study, electroporation experiments using primary rabbit PBMC confirmed that these cells are able to support postintegration steps in the HIV replication cycle. Thus, we have no reason to conclude that rabbit PBMC have an intrinsic inability to carry out the basic activities required for HIV replication.

In conclusion, viral replication in rabbit cells expressing human CCR5 and CD4 in vitro approaches the level seen in human cells and is markedly higher than that in murine cells. HIV virions produced by these cells are able to support spreading infection and successfully infect human PBMC. These findings suggest that transgenic rabbits expressing CCR5 and CD4 may support reasonable levels of HIV replication in vivo, and we speculate that such animals might serve as a useful small-animal model for HIV disease. A small-animal model that simulates selected stages of HIV transmission and/or disease pathogenesis in humans would be invaluable for the study of HIV disease, mechanisms of transmission and pathogenesis, and treatment or prevention strategies.