INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The initial stage in lentiviral infection involves the binding of the viral envelope glycoprotein (Env) to a molecule on the surface of the target cell. The primary high-affinity binding receptor for human immunodeficiency virus (HIV) is CD4 (9, 26), a member of the immunoglobulin supergene family of molecules. However, binding of the viral glycoprotein to CD4 is insufficient for infection to proceed (29); for virus-cell fusion to occur, the target cell must also express an accessory molecule or coreceptor. The principal coreceptors for HIV infection have now been identified as members of the seven-transmembrane domain (7TM) superfamily of molecules. Syncytium-inducing (SI) T-cell line-tropic strains of virus require coexpression of the -chemokine receptor CXCR4 for infection (19), whereas non-syncytium-inducing (NSI) strains of virus require coexpression of the -chemokine receptor CCR5 for infection (1, 6, 10, 13, 14). In addition, other chemokine receptors such as CCR2b and CCR3 (6, 13, 41, 48), the receptor encoded by human cytomegalovirus US28 (39, 41), and the orphan receptor STRL33 (28) can function as coreceptors for HIV infection. More recently, additional members of the 7TM superfamily have been identified as coreceptors for infection with simian immunodeficiency virus (SIV). Two of these receptors, termed Bonzo and BOB, support infection with not only SIV but also HIV type 2 (HIV-2) and macrophage-tropic or dualtropic (both macrophage- and T-cell-tropic) strains of HIV-1 (11). Bonzo has subsequently been identified as being identical to STRL33 (28), whereas BOB is identical to GPR15 (21). A subsequent study has demonstrated that an additional molecule, designated GPR1 (30), can function as a coreceptor for SIV (18). Thus, a diverse range of 7TM molecules which can support infection with primate lentiviruses have now been identified.

The selective usage of chemokine receptors as coreceptors for infection by HIV and SIV is borne out by the sensitivity of the viruses to inhibition by chemokines. Infection with viruses which use CCR5 can be inhibited by the -chemokines RANTES, MIP-1, and MIP-1 (7, 14), whereas those which use CXCR4 can be inhibited by stromal cell-derived factor (SDF-1) (3, 36). Although infection of primary macrophages by certain primary NSI viruses is not inhibited reproducibly by the -chemokines RANTES, MIP-1, and MIP-1 (14, 33, 44), analogs of the -chemokines such as AOP-RANTES that inhibit HIV infection with an increased potency, inhibit infection of both peripheral blood mononuclear cells (PBMC) and primary macrophages, and do not trigger signalling via G proteins coupled to the chemokine receptor have been developed (47). Therefore, with the development of SDF-1 derivatives analogous to AOP-RANTES, it may be possible to generate therapeutic agents that are effective at inhibiting not only the NSI strains of HIV found in early infection but also the SI strains of virus which appear late in infection with the progression to AIDS.

Feline immunodeficiency virus (FIV) induces an AIDS-like illness in its natural host, the domestic cat (38). A proportion of primary isolates of FIV can be readily adapted to grow and form syncytia in the Crandell feline kidney (CrFK) cell line (45), analagous to the isolation of SI variants of HIV. Sequencing of the env gene from CrFK-tropic viruses would suggest that the principal determinant of CrFK tropism is an increase in charge of the V3 loop of the envelope glycoprotein (45, 51), further strengthening the analogy between CrFK-tropic strains of FIV and SI strains of HIV. While the primary high-affinity binding receptor for FIV remains elusive, recent studies have demonstrated a role for the feline homolog of CXCR4 in infection with CrFK-tropic strains of FIV (53, 56). Given that the appearance of CXCR4-dependent SI variants of HIV in the peripheral blood of HIV-infected individuals accompanies the progression to AIDS (8), the ability to study the role of such CXCR4-dependent strains of virus in disease pathogenesis is of obvious interest. Moreover, as it appears that several strains of SIV show preferential usage of CCR5 and not CXCR4 for infection (5, 11, 18), then FIV infection of the domestic cat is the only animal model described to date in which the contribution of CXCR4-dependent viruses to the pathogenesis of AIDS may be studied in the natural host of the virus.

In this study, we investigated the nature of the interaction between FIV and the chemokine receptor CXCR4. Given the high degree of amino acid sequence homology between human and feline CXCR4 (56), we examined the interaction between human SDF-1 and feline CXCR4. We have found that human SDF-1 binds specifically to feline CXCR4 and inhibits infection with FIV. We demonstrate that SDF-1 can upregulate CXCR4 expression with a corresponding enhancement of infection and that this effect can be mimicked by treatment of the cells with the phorbol ester phorbol myristate acetate (PMA). Moreover, infection of interleukin-2 (IL-2)-dependent T cells with FIV was resistant to the inhibitory effects of SDF-1, suggesting the existence of a CXCR4-independent mechanism of infection in these cells. These data suggest that the mechanism of infection with FIV bears striking similarities to infection with HIV and that the study of FIV infection of the domestic cat may provide a valuable insight into the pathogenesis of AIDS.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Antibodies and reagents. Recombinant human SDF-1, MCP-3, RANTES, MCP-1, MIP-1, MIP-1, and IL-8 were from PeproTech, Inc., Rocky Hill, N.J., or as described elsewhere (22). Synthetic human SDF-1 was obtained from I. Clarke-Lewis, University of British Columbia, Vancouver, British Columbia, Canada. Recombinant human RANTES, MIP-1, and MIP-1 were kindly provided by T. Wells, Glaxo SA, Geneva, Switzerland. Anti-human CXCR4 monoclonal antibodies R&D 44702 and R&D 44717 were a generous gift from Monica Tsang, R&D Systems, Minneapolis, Minn. Phorbol myristate acetate (PMA) was obtained from Calbiochem, Nottingham, United Kingdom. Immunoaffinity-purified FIV envelope glycoprotein from the PET strain of FIV (FIV_PET) was prepared as described previously (25). All culture media and supplements were obtained from Life Technologies, Paisley, United Kingdom. ¹²⁵I-SDF-1 and - (specific activity, 2,200 Ci/mol) were from Dupont NEN, Boston, Mass.

Cell lines and viruses. U87 cells expressing feline CXCR4 stably from the pRep4 vector (U87-feCXCR4 cells) have been described previously (56). U87 cells expressing human CXCR4 (U87-huCXCR4 cells) were obtained from N. Landau, Aaron Diamond AIDS Research Center, New York, N.Y. U87-feCXCR4, AH927 (40), and CrFK cells were maintained in Dulbecco's modification of minimal essential medium supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, sodium pyruvate (0.11 mg/ml), penicillin (100 IU/ml), and streptomycin (100 µg/ml) (DMEM). F422 (42), T3 (35), Q201 (52), and Mya-1 (32) cells were maintained in RPMI 1640 medium supplemented with 10% FBS, 2 mM glutamine, penicillin (100 IU/ml), streptomycin (100 µg/ml), and 5 × 10⁵ M 2-mercaptoethanol (RPMI medium). Q201 and Mya-1 cells were maintained in RPMI medium supplemented with recombinant human IL-2 (100 IU/ml). The culture media for U87-feCXCR4 and U87-huCXCR4 were supplemented with hygromycin (10 µg/ml). The CrFK-tropic virus FIV_PET was prepared from a culture of CrFK cells persistently infected with the F14 molecular clone of FIV_PET (37).

Inhibition of FIV infection by chemokines. CrFK cells were plated in 48-well tissue culture plates at 2 × 10⁴ cells per well in DMEM and allowed to adhere overnight. Chemokines were diluted to working concentration in DMEM, added to the cells in a total volume of 100 µl per well, and incubated for 1 h at 37°C. Ten CrFK syncytium-forming units of virus in 50 µl was added per well and incubated for a further 1 h. The supernatant was then aspirated, the wells were washed twice with DMEM, and maintenance chemokine was added at the appropriate final concentration. The cells were cultured for 4 days and then fixed and stained, and the syncytia were enumerated by light microscopy. Supernatants were stored from each well for the detection of viral p24 by enzyme-linked immunosorbent assay (ELISA) (FIV antigen test kit; IDEXX, Portland, Maine). Blocking assays using the feline IL-2-dependent T-cell line Mya-1 were performed in 96-well round-bottom plates. Cells (10⁵) were seeded into each well in 50 µl of RPMI. Chemokines were adjusted to double the final working concentration, added in 100 µl of RPMI medium to each well, and incubated for 1 h at 37°C. The final concentration of SDF-1, RANTES, IL-8, MIP-1, and MCP-1 was 1 µg/ml, while the MIP-1-MIP-1-RANTES mixture (MMR) consisted of 330 ng of each chemokine per ml. Virus was added in 50 µl per well and incubated for a further 1 h at 37°C. Finally, the cells were washed four times by centrifugation of the plate at 1,000 rpm followed by aspiration of the culture medium. The cells were then resuspended in fresh RPMI medium containing chemokines at the appropriate concentration and cultured at 37°C for 4 or 7 days, when samples of supernatants were collected for viral p24 ELISA.

Flow cytometry. Flow cytometric analyses were performed essentially as described previously (55). Briefly, adherent cells were removed from the culture plastic by incubation with trypsin-EDTA, washed once by centrifugation through DMEM supplemented with 10% FBS, and resuspended in phosphate-buffered saline (PBS) supplemented with 1.0% bovine serum albumin and 0.1% sodium azide. The cells were then incubated with either anti-feline/human CXCR4 antibody 44717 or isotype-matched (immunoglobulin G2b) antibody 44702, which does not recognize feline CXCR4, for 30 min at 4°C, washed twice by centrifugation, and then incubated for a further 30 min with fluorescein isothiocyanate-conjugated F(ab')₂ fragment of goat anti-mouse immunoglobulin G. Samples were washed twice by centrifugation and then analyzed on a Coulter EPICS Elite flow cytometer, 5,000 events being acquired in LIST mode for each sample.

Binding studies. The SDF-1 and SDF-1 binding studies were performed as described previously (22). SDF-1 was generated by peptide synthesis as described previously (22). U87-CXCR4 cells were seeded in 24-well plastic cell culture plates and grown until confluent. The culture medium was aspirated and replaced with PBS containing radiolabeled chemokine (500 pM) in the presence or absence of increasing concentrations of unlabeled chemokines or FIV envelope glycoprotein at room temperature for 30 min. The incubation was terminated by aspiration of the supernatant; the cells were then washed once with PBS and solubilized by the addition of 500 µl of 25% (wt/vol) sodium dodecyl sulfate. The lysate was then transferred to scintillation vials for counting in a gamma counter. Nonspecific binding was determined in the presence of 1 µM unlabeled chemokine, and each experiment was performed in duplicate. The binding data were curve fitted by using the IGOR software package (Wavemetrics Inc., Lake Oswego, Oreg.) to determine the binding affinity (K_D), number of sites, and amount of nonspecific binding.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Binding of human SDF-1 to feline CXCR4. The anti-human CXCR4 antibody 12G5 inhibits fusion between FIV-infected feline cells and human CXCR4-expressing cells but does not recognize feline CXCR4 (56). To examine the role of CXCR4 in infection of feline cells with FIV, we asked whether sufficient amino acid sequence homology existed between human and feline CXCR4 to permit SDF-1 binding to feline CXCR4. U87 cells stably transfected with feline CXCR4 were exposed to ¹²⁵I-labeled human SDF-1 (Fig. 1a) or SDF-1 (Fig. 1b) in the presence of increasing concentrations of the unlabeled ligand. The cells were incubated for 1 h, and then free chemokine was aspirated. The cells were then lysed, and bound chemokine was measured in a gamma counter. Homologous competition binding studies followed by Scatchard analysis revealed that SDF-1 bound to the feline CXCR4-transfected cells, but not feline CD9-transfected cells, with a K_D of 12.0 ± 5.3 nM and that SDF-1 bound with a K_D of 10.4 ± 1.5 nM. Therefore, feline CXCR4 acts as a high-affinity binding receptor for human SDF-1 and SDF-1.

View larger version (17K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   FIG. 1.   SDF-1 and - binding to feline CXCR4 and FIV gp120 competition. (a) Homologous competition binding curve of 125I-SDF-1 to feline CXCR4 in U87 cells. The insert shows Scatchard analysis of binding (KD = 12.0 ± 5.2 nM; representative experiment [n = 3]). Total cpm = 92,000; background cpm = 37,000. (b) Homologous competition binding curve of 125I-SDF-1 to feline CXCR4 in U87 cells. The insert shows Scatchard analysis of binding (KD = 10.4 ± 1.5 nM; representative experiment [n = 2]). Total cpm = 37,000; background cpm = 13,500. (c) Heterologous competition binding of FIV gp120 and 125I-SDF-1 to feline CXCR4 in U87 cells. The insert shows Scatchard analysis of binding (KD = 1.06 ± 0.14 nM [n = 2]). Total cpm = 76,000; background cpm = 20,000.

Inhibition of FIV infection by SDF-1. Having demonstrated that SDF-1 bound with high efficiency to feline CXCR4, we investigated whether infection of feline cells by FIV could be inhibited by SDF-1. CrFK cells were infected with FIV_PET in the presence of increasing concentrations of either recombinant human SDF-1 (Fig. 2a) or synthetic human SDF-1 (Fig. 2b). Four days postinfection, the cells were fixed and stained and the number of syncytia per well was quantified by light microscopy. Increasing concentrations of either recombinant or synthetic SDF-1 led to a dose-dependent inhibition of syncytium formation, with approximately 50% inhibition at 12.5 ng/ml and 80 to 100% inhibition of infection being achieved at concentrations of 0.5 to 1.0 µg/ml. These concentrations correlate well with the effective concentrations observed for inhibition of HIV infection by SDF-1 (3, 36). Culture supernatants were collected from the cells treated with synthetic human SDF-1 and analyzed for viral p24 production by ELISA. SDF-1 treatment led to a dose-dependent decrease in p24 production (Fig. 2c), in good agreement with the assay for syncytium formation, and complete inhibition of p24 production was achieved at a concentration of 1 to 2 µg/ml. No inhibition of infection was observed in CrFK cells treated with human MCP-3, RANTES, MCP-1, MIP-1, MIP-1, or IL-8 at similar concentrations (not shown). Moreover, in a separate experiment we found that the inhibitory activity of SDF-1 against FIV infection could be partially neutralized by a polyclonal anti-SDF-1 antiserum (data not shown).

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Inhibition of FIVPET infection of CrFK cells by SDF-1. CrFK cells were cultured overnight in 48-well plates and incubated in the presence of increasing concentrations of either recombinant human SDF-1 (a) or synthetic human SDF-1 (b) for 1 h prior to infection with FIVPET. Four days postinfection, the cells were fixed and stained and the number of syncytia per well was determined. Results are the means of triplicate wells. Supernatants were collected from the cells treated with synthetic human SDF-1 and analyzed for viral p24 by ELISA (c). Incubation with increasing concentrations of SDF-1 led to a dose-dependent reduction in viral p24 release.

Modulation of feline CXCR4 expression by PMA and SDF-1. The interaction between SDF-1 and human CXCR4 leads to down-regulation of the receptor (2, 23, 46). Previous studies on the IL-8 and MCP-1 receptors have indicated that internalization of chemokine receptors can occur via endocytosis (17, 43), and colocalization of CXCR4 and transferrin in early endosomes (2) suggests that down-regulation of CXCR4 by SDF-1 is mediated by endocytosis. To determine whether a functional interaction occurred between human SDF-1 and feline CXCR4, we next investigated the down-modulation of feline CXCR4 by SDF-1 and PMA. Previously we demonstrated that the anti-human CXCR4 antibody 12G5 failed to recognize feline CXCR4 (56). We therefore screened a newly developed panel of anti-human CXCR4 antibodies for reactivity against feline CXCR4. We identified three monoclonal antibodies (R&D 44701, 44717, and 44718) that reacted with the feline T-lymphoma cell lines 3201 (97.3% positive), T3 (89.8% positive), and F422 (99.5% positive) and the adherent cell line CrFK (37.0% positive) (Fig. 3). The three antibodies were indistinguishable with respect to fluorescence intensity and percentage positive on each of the cell lines. Northern blotting analysis confirmed that each of these cell lines expressed high levels of CXCR4 mRNA (data not shown and previous results [56]). In contrast, the AH927 cell line did not appear to express CXCR4 (1.5%), in agreement with previous findings that these cells do not express CXCR4 mRNA (56). Surprisingly, the Mya-1 cell line was not recognized by any of the three monoclonal antibodies despite expressing significant levels of CXCR4 mRNA. Indeed, Mya-1 mRNA was used as the source of CXCR4 mRNA for the generation of the feline CXCR4 cDNA clone (56). Similar findings were observed with a second IL-2-dependent feline T-cell line, Q201. Given that these antibodies recognize cells transfected with feline CXCR4 (50), the data suggest either that feline CXCR4 is not expressed at the cell surface in these cell lines or that the epitope recognized by these antibodies is masked.

View larger version (25K): [in this window] [in a new window]   FIG. 3.   CXCR4 expression on feline cell lines. Flow cytometric analysis of CXCR4 expression on feline cells was performed with the anti-human CXCR4 antibody R&D 44718; 5,000 events were collected for each sample in LIST mode. Histograms illustrate percentage positive (open) and the isotype-matched control (filled) which was used to set an analysis gate in which <1.0% cells were positive.

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Modulation of CXCR4 expression on feline cell lines following treatment with SDF-1 or PMA on F422, T3, and Ho6T1 (derivative of CrFK cell line) cells. The feline T-lymphosarcoma cell lines T3 and F422 and the adherent cell line CrFK were exposed to SDF-1 (1 µg/ml) or PMA (10 ng/ml) overnight and then analyzed by flow cytometry for CXCR4 expression with R&D 44717. Histograms illustrate treated cells (open) relative to untreated cells (filled) stained with R&D 44717; 5,000 events were collected for each sample.

View larger version (26K): [in this window] [in a new window]   FIG. 5.   Enhancement of infection of CrFK cells following treatment with SDF or PMA. CrFK cells were incubated with SDF (1 µg/ml) or PMA (50 ng/ml) overnight, washed, and then infected with FIVPET. Four days postinfection, supernatants were collected and analyzed for viral p24 by ELISA (a). The cells were fixed and stained, and the number of syncytia per well was determined (b).

View larger version (15K): [in this window] [in a new window]   FIG. 6.   Time course of CXCR4 up-regulation on CrFK cells by PMA. CrFK cells were incubated with PMA (50 ng/ml) for 0, 1, 3, 12, 24, or 48 h. The cells were then subcultured; half were replated and infected with FIVPET, while the remainder were analyzed for CXCR4 expression by flow cytometry (a). The data represent mean number of positive cells relative to the isotype-matched control () and the mean fluorescence intensity (). Four days after infection with FIVPET, the cells were fixed and stained and the number of syncytia per well was determined (b). Supernatants were collected, and viral p24 was quantified by ELISA (c). Results represent a typical experiment and are the means of three estimations.

Loss of CXCR4 expression following FIV infection. Infection with the CXCR4-dependent vcp strain of HIV-2 leads to down-regulation of surface expression of CXCR4 (16). We therefore investigated the effect of FIV infection on CXCR4 expression on CrFK. CrFK cells persistently infected with either FIV_PET or FIV_GL8, or uninfected control cells, were stained with the antibody R&D 44717 or an isotype-matched control antibody, and reactivity was analyzed by flow cytometry (Fig. 7). While 68.7% of the uninfected CrFK cells were positive for CXCR4 expression, CXCR4 expression was markedly reduced on the FIV_GL8- and FIV_PET-infected cells; in each case, only 2.0% of the cells were CXCR4 positive (isotype-matched control stained cells were 1.7 and 1.9% positive, respectively), suggesting either down-regulation of CXCR4 expression or elimination of CXCR4-expressing cells from the culture. The data provide further evidence for the interaction between FIV and CXCR4 in CrFK cells.

View larger version (15K): [in this window] [in a new window]   FIG. 7.   Down-regulation of CXCR4 in FIV-infected CrFK cells. CrFK cells persistently infected with a CrFK-adapted stock of either FIVGL8 (a) or FIVPET (b) were analyzed by flow cytometry for expression of CXCR4. Infected cells (open) are shown relative to uninfected CrFK cells (shaded) analyzed in parallel.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

FIV can be selected for growth and syncytium formation in the feline cell line CrFK, and isolates which grow in this cell line have an extended cell tropism including nonfeline cells such as the human cell line HeLa. Previously, we found that syncytium formation by CrFK-tropic FIV was mediated by the chemokine receptor CXCR4. In this study, we extend our previous observations and demonstrate that human SDF-1 and SDF-1 bind with a high affinity to feline CXCR4. Furthermore, human SDF-1 inhibited FIV infection of CrFK cells efficiently, leading to a dose-dependent reduction in both the number of syncytia and viral p24 production. In contrast, the human -chemokines RANTES, MIP-1, MIP-1, MCP-1, and MCP-3 or the -chemokine IL-8 had no effect on FIV infection of CrFK cells. It is possible that the amino acid sequences of the human -chemokines RANTES, MIP-1, and MIP-1 are sufficiently divergent from their feline homologs to render them nonfunctional in assays using feline cells, and indeed preliminary reports suggest that the feline homologs of MIP-1 and MIP-1 display only 75.3 to 79.6% and 73.9 to 88.0% homology to those of other species (15). However, given that SDF-1 inhibited FIV infection of CrFK cells completely, the data suggest that CrFK-adapted strains of FIV are analogous to T-cell line-adapted strains of HIV such as LAI or NL4-3 and that CrFK cells can be considered to be analogous to HeLa-CD4 cells, where infection is mediated exclusively by CXCR4 and is inhibited completely by SDF-1.

SDF-1 had no effect on infection of the IL-2-dependent T-cell line Mya-1 with CrFK-tropic FIV, implying the existence of a CXCR4-independent mechanism of infection and suggesting that CrFK-tropic viruses are in fact dualtropic. By analogy, infection of human PBMC with an HIV-1 recombinant pseudotyped with the envelope glycoprotein from an obligate CXCR4-dependent virus (HXBc2) was shown to be inhibited by SDF-1 (3); thus, if CrFK-tropic FIV was restricted to usage of CXCR4 alone (as with HIV HXBc2), we would expect to see inhibition of infection by SDF-1 irrespective of the cell type. In contrast, infection with a dualtropic virus such as 89.6 is inhibited by SDF-1 only when the target cell expresses CXCR4 alone; if the target cell expresses both CXCR4 and CCR5, then SDF-1 does not inhibit infection (12).

Intriguingly, CrFK cells were recognized by the cross-species-reactive anti-CXCR4 antibodies whereas Mya-1 cells were not, despite CXCR4 mRNA being abundant in both cell lines (56). Such anomalies have been described previously for HIV, where the anti-CXCR4 antibody 12G5 failed to block fusion mediated by LAI envelope in U87.CD4 cells transfected with CXCR4 (31). As recent studies have suggested that the epitope recognized by the monoclonal antibody 12G5 is masked in complexes between CD4 and CXCR4 but can be revealed by treatment of the cells with the anti-CD4 antibody Q4120 (34), the epitope recognized by the anti-feline CXCR4 antibodies may be masked on Mya-1 and Q201 cells. Alternatively, Mya-1 and Q201 cells may have a high turnover of CXCR4 at the cell surface, and thus although CXCR4 is expressed at the cell surface, it may be rapidly internalized. Recent studies have demonstrated that CXCR4 undergoes constitutive internalization in several human cell lines and that the rate of internalization is enhanced by the addition of phorbol ester and SDF-1 (46). Furthermore, greater than 90% of SDF-1 bound to the human cell line Jurkat is internalized within 2 h via CXCR4 (23). Future studies will analyze the kinetics of feline CXCR4 expression in order to ascertain the role of CXCR4 internalization in susceptibility and resistance to FIV infection.

FIV envelope glycoprotein competed with human SDF-1 for binding to feline CXCR4. Previous studies have suggested that CXCR4 and CD4 form a complex with envelope glycoprotein from T-tropic, SI strains of HIV (27), while CCR5 and CD4 form complexes with the envelope glycoprotein from macrophage-tropic, NSI strains of HIV (49, 57). In this study, we demonstrate that FIV envelope glycoprotein competes specifically with SDF for binding to feline CXCR4, confirming that there is a direct interaction between the virus and CXCR4. Furthermore, we have found that CXCR4 expression is down-regulated on FIV-infected CrFK cells. These findings parallel the CD4-independent binding of HIV T-tropic envelope glycoprotein to CXCR4 on hNT neurons (22), HIV-1 IIIB infection of hNT cells (20), and the usage of CXCR4 by CD4-independent strains of HIV-2, vcp and ROD-B (16). Of the coreceptors identified to date for HIV, only CXCR4 appears to be capable of supporting infection in a CD4-independent fashion (16). Recent studies have suggested that envelope glycoproteins from CCR5-dependent strains of HIV-2 and SIV can interact directly with CCR5 in a CD4-independent fashion, displacing the chemokine MIP-1 in the process (24). However, optimal binding and membrane fusion still require the formation of a trimolecular CCR5-CD4-gp120 complex (24, 49, 57). It has been proposed that Envs from HIV-2 and SIV may interact directly with chemokine receptors with a higher affinity than Envs from HIV-1 and that it is this reduced requirement for an interaction with CD4 that has enabled CD4-independent strains of HIV-2 to arise (24). FIV Env competed with SDF-1 for binding to CXCR4 with an efficiency similar to that for the competition between HIV-2 ST Env and MIP-1 for CCR5 (24). Thus, these data provide further support for CXCR4-dependent strains of FIV resembling CD4-independent infection with HIV and suggest a common mechanism of infection by the primate and feline lentiviruses.

Given that previous studies had demonstrated that CXCR4 down-regulation contributed to the antiviral activity of SDF-1 against HIV infection (2), we looked at the effects of SDF-1 and PMA on FIV infection of CrFK cells. Surprisingly, overnight treatment of CrFK cells with SDF-1 or PMA led to a marked up-regulation of CXCR4 expression and a concomitant increase in susceptibility to infection with FIV. CXCR4 up-regulation on human T cells has been observed following stimulation with either phytohemagglutinin or IL-2 (4). In this study, we have found that on CrFK cells, but not T3 or F422 cells, prolonged treatment with SDF-1 or PMA had a similar effect. Thus, FIV infection of CrFK cells can be either inhibited or enhanced by treatment with SDF-1, depending on the incubation time. These findings provide a possible explanation for many of the conflicting data regarding the inhibition of HIV infection by chemokines; the inhibitory effects of chemokines on HIV infection may be entirely dependent on the target cell type, and indeed, in assay systems based on heterogeneous cell populations such as PBMC, it is likely that the net effect may be the result of conflicting inhibitory and enhancing actions of the chemokine. Accordingly, the inhibitory activity of SDF against FIV infection in CrFK cells may be the sum of steric hindrance of the interaction between FIV_PET gp120 and CXCR4 and up-regulation of CXCR4 expression by SDF.

Our studies demonstrate that CXCR4 is the principal receptor for CrFK-tropic strains of FIV. Future studies will investigate the prevalence of such SI strains of FIV in infected cats and establish whether such viruses are more prevalent in cats which progress to AIDS. Understanding the contribution of these viruses to the immunodeficiency induced following FIV infection of the domestic cat may help to elucidate the pathogenesis of AIDS in HIV-infected individuals.