Similarities and Differences in RANTES- and (AOP)-RANTES–triggered Signals: Implications for Chemotaxis

HEK-293 cells (ATCC TIB202) were obtained from the American Type Culture Collection, and CCR5-transfected HEK-293 cells were kindly donated by Dr. J. Gutierrez (Dept. of Immunology and Oncology, Centro Nacional de Biotecnología, Madrid, Spain). Antibodies used include rabbit anti-JAK1 and anti-JAK2 (Upstate Biotechnology, Inc.), anti-Gα_i and anti-STAT5 (Santa Cruz Biotechnology); monoclonal anti-PTyr (4G10; Upstate Biotechnology Inc.), anti-PTyr (PY20) (Transduction Laboratories), and anti–β₂-microglobulin (PharMingen). Anti-CCR5 mAb, CCR5-02, CCR5-03, and anti-CXCR4 mAb were generated in our laboratory as described (Mellado et al., 1997). EGF-stimulated A-431 cell lysates were obtained from Upstate Biotechnology Inc. and recombinant human RANTES and SDF-1α were from Peprotech Inc.

Cells were centrifuged (250 g, 10 min, room temperature), plated in V-bottom 96-well plates (2.5 × 10⁵ cells/well), and incubated with 50 μl/ well biotin-labeled mAb (5 μg/ml, 60 min, 4°C). Cells were washed twice in PBS with 2% BSA and 2% FCS, and centrifuged (250 g, 5 min, 4°C). FITC-labeled streptavidin (Southern Biotechnologies Associates, Inc.) was added, incubated (30 min, 4°C) and plates were washed twice. Cell-bound fluorescence was determined in a Profile XL flow cytometer at 525 nm (Coulter Electronics).

Changes in intracellular calcium concentration were monitored using a fluorescent probe (Fluo-3AM; Calbiochem). Cells (2.5 × 10⁶ cells/ml), were resuspended in RPMI containing 10% FCS, 10 mM Hepes, and incubated with 10 μl/10⁶ cells of Fluo-3AM (300 μM in DMSO, 15 min, 37°C) (Frade et al., 1997a). After incubation, cells were washed, resuspended in complete medium containing 2 mM CaCl₂, and maintained at 4°C until just before chemokine addition to minimize membrane trafficking and to eliminate spontaneous Ca²⁺ entry. Calcium mobilization in response to 10 nM RANTES or (AOP)-RANTES and 20 nM SDF-1α was determined at 37°C in an EPICS XL flow cytometer at 525 nm. It includes background level stabilization and determination of the probe loading level for each sample. Only samples with a similar load, as assessed by ionophore-induced Ca²⁺ mobilization, were used (5 μg/ml ionomycin; Sigma Chemical Co.).

CCR5-transfected HEK-293 cells (0.5 × 10⁶ cells/ml) were starved for 120 min at 37°C, 5% CO₂ in RPMI 1640 containing 0.1% BSA. Cells (0.25 × 10⁶ cells in 0.1 ml) were placed in the upper well of 24-well, 8-μm pore size transmigration chambers (Transwell; Costar Corp.) and RANTES or (AOP)-RANTES (diluted in 0.6 ml RPMI containing 0.25% BSA) were added to the lower well. Plates were incubated (240 min, 37°C, 5% CO₂) and the cells that migrated to the lower chamber were counted, as described (Frade et al., 1997a).

RANTES- or (AOP)-RANTES–stimulated cells (20 × 10⁶) were lysed in a detergent buffer (20 mM triethanolamine, pH 8.0, 300 mM NaCl, 2 mM EDTA, 20% glycerol, 1% digitonin, with 10 μM sodium orthovanadate, 10 μg/ml leupeptin, and 10 μg/ml aprotinin) for 30 min at 4°C with continuous rocking and then centrifuged (15,000 g, 15 min). Immunoprecipitations were performed essentially as described earlier (Mellado et al., 1997). Protein extracts precleared by incubation with 20 μg of anti–mouse IgM-agarose (Sigma Chemical Co.) or protein A–Sepharose (60 min, 4°C) were centrifuged (15,000 g, 1 min), immunoprecipitated with the appropriate antibody (5 μg/sample, 120 min, 4°C), followed by anti–mouse IgM-agarose or by protein A–Sepharose if the first antibody was derived from rabbit serum. Immunoprecipitates or protein extracts were separated in SDS-PAGE and transferred to nitrocellulose membranes. Western blot analysis was performed as described (Mellado et al., 1997), using 2% BSA in TBS as blocking agent for the antiphosphotyrosine analyses. When stripping was required, membranes were incubated for 60 min at 60°C with 62.5 mM Tris-HCl, pH 7.8, containing 2% SDS and 0.5% β-mercaptoethanol. After washing with 0.1% Tween 20 in TBS for 2 h, membranes were reblocked, reprobed with the appropriate antibody, and developed as above. In all cases, protein loading was carefully controlled using a protein detection kit (Pierce Chemical Co.) and, when necessary, by reprobing the membrane with the immunoprecipitating antibody.

Serum starved CCR5 transfected HEK-293 cells (20 × 10⁶) were stimulated with 10 nM RANTES, 10 nM (AOP)-RANTES, or 20 nM SDF-1α for 1 min at 37°C. The reaction was terminated by addition of 1 ml of cold PBS and centrifugation (30 s, 15,000 g). After washing twice with cold PBS, 10 μl of 100 mM disuccinimidyl suberate (DSS; Pierce Chemical Co.) was added for 10 min at 4°C with continuous rocking. The reaction was terminated by adding 1 ml of cold PBS and washing three times. The pellet was lysed for 60 min and immunoprecipitated as above.

Immunofluorescence experiments were performed essentially as described (Nieto et al., 1997). Briefly, 2 × 10⁶ T lymphoblasts in 500 μl complete medium were allowed to adhere to fibronectin (FN)-coated coverslips in 24-well plates (Costar Corp.). Cytokines and chemokines (10 ng/ml) were added, and cells were incubated (37°C, 5% CO₂). After 30 min, cells were fixed with 3.7% formaldehyde in PBS for 10 min at room temperature, and stained with specific mAb plus a FITC-labeled rabbit anti– mouse IgG F(ab′)₂ (Pierce Chemical Co.). The proportion of CCR5 polarization was calculated by direct cell count (n = 400–500) of 10 random fields for each condition using a photomicroscope with 60× oil immersion objectives (Labophot-2; Nikon Inc.). Preparations were photographed on Ektachrome 400 film. Images were acquired with a high performance CCD camera (Cohu) coupled to the microscope and connected to a workstation (model Q550CW; Leica Imaging Systems, Ltd.). Images were visualized, processed, and stored using QFISH software (version V1.01; Leica Imaging Systems, Ltd.) and printed on a color printer (Phaser 440; Tektronix Inc.).

We developed anti–human CCR5-specific mAb using synthetic peptides corresponding to the NH₂-terminal domain of this receptor (amino acids 13–28), a CCR5-specific sequence not present in other chemokine receptors. These mAbs recognize the CCR5 in human PBMC (not shown) and in CCR5 stably transfected HEK-293 cells, in flow cytometry (Fig. 1 A), as well as in Western blot, and immunoprecipitation analysis (Fig. 1 B). No binding is observed in mock- or CCR2-transfected HEK-293 cells, or in any other cell tested that does not express CCR5.

In response to RANTES or (AOP)-RANTES, these CCR5-transfected HEK-293 cells mobilize calcium (Fig. 2 A) and are desensitized to a second stimulation. However, only RANTES triggers migration in these cells (Fig. 2 B). PTX treatment abrogates both calcium release and cell migration in response to RANTES and (AOP)-RANTES (Fig. 2 A), whereas no effect was observed following incubation with cholera toxin (not shown). This is consistent with other studies showing that RANTES downstream signals in other cell lines and T cells are coupled to PTX-sensitive G proteins (Bacon et al., 1995b), and with coupling of other chemokine receptors to G_i in transfected HEK-293 cells (Aragay et al., 1998; Mellado et al., 1998), indicating the utility of this cell line in the study of signaling through GPCR.

We have described the ability of RANTES to polarize peripheral blood T cells, such that the CCR5 receptor migrates to the cell's leading edge while the ICAM-1, ICAM-3/ CD43, CD44 molecules localize at the uropod (Nieto et al., 1997). Therefore, we tested the ability of both of these ligands to induce polarization in peripheral T cell blasts. Time-course studies of ICAM-3 redistribution to the cell uropod of FN-adhered T lymphoblasts showed that RANTES induced ICAM-3 redistribution in a significant proportion of the cells as early as 15 min after stimulation. Maximum redistribution, seen at 30 min, persisted until at least 90 min of incubation with the chemokine. In contrast, (AOP)-RANTES–induced ICAM-3 polarization showed a much slower, weaker response, with a peak at 30 min that vanished thereafter (Fig. 3 A).

We next studied the effects of these ligands on the polarization of CCR5 to the advancing front of migrating T lymphoblasts, a phenomenon that may be involved in the chemotaxis mechanism. We found that whereas RANTES triggered redistribution of CCR5 molecules to the leading edge of the cells, membrane CCR5 expression was not detectable in (AOP)-RANTES–stimulated polarized T lymphoblasts bearing ICAM-3 redistributed to the uropod (Fig. 3 B). The downregulation of the (AOP)-RANTES– triggered chemokine receptor results in a net inhibition of CCR5 redistribution to the leading edge of T lymphoblasts stimulated with this chemokine derivative, compared to those treated with RANTES (Fig. 3 C). This concurs with previous observations showing that (AOP)-RANTES induces CCR5 downregulation by blocking receptor recycling (Mack et al., 1998).

When CCR5-transfected HEK-293 cells were stimulated with RANTES or (AOP)-RANTES, a 38-kD protein phosphorylated in tyrosine residues was initially identified as the CCR5 receptor by immunoblot (not shown). Thereafter, lysates from transfected HEK-293 cells were immunoprecipitated with anti-CCR5, and Western blots developed with anti-PTyr antibodies (Fig. 4). An increase in CCR5 receptor phosphorylation in tyrosines is seen as early as 60 s after stimulation. At this time no differences were observed following treatment with either stimulus, indicating that (AOP)-RANTES not only binds to CCR5, but also signals through this receptor. The early, weak tyrosine phosphorylation level may occur because only one tyrosine (Tyr 126 in the DRY motif) was being phosphorylated, analogous to the case of Tyr 139 in the CCR2 receptor (Mellado et al., 1998). Similar results are observed in the inverse experiment, in which anti-PTyr antibody immunoprecipitates are analyzed in Western blot with anti-CCR5 receptor antibody (not shown). Again, tyrosine phosphorylation is detected in CCR5 after stimulation with either RANTES or (AOP)-RANTES. Nevertheless, significant differences are observed in tyrosine phosphorylated CCR5 after longer stimulation periods. Whereas CCR5 phosphorylation persists in RANTES-treated cells after 15 min of treatment, this is not the case after (AOP)- RANTES stimulation, reinforcing the observation that there are differences in the signaling pathways activated by these chemokines.

To ascertain which kinase is responsible for the rapid CCR5 chemokine receptor phosphorylation, CCR5-transfected HEK-293 cells were stimulated with RANTES or (AOP)-RANTES, and cell lysates immunoprecipitated with anti-CCR5, or with anti–β₂-microglobulin as an isotype-matched antibody control. The use of a specific anti-JAK1 antibody identified a 130-kD protein in the anti-CCR5 immunoprecipitates (Fig. 5 A). Furthermore, the JAK1 tyrosine kinase is phosphorylated on tyrosine residues after stimulation with either ligand, since anti-PTyr immunoprecipitates can be developed with anti-JAK1 antibodies in Western blot (Fig. 5 B). JAK1 association to the CCR5 receptor takes place as early as 30 s after RANTES or (AOP)-RANTES stimulation (Fig. 5 A). Small amounts of JAK1 were also found associated to the CCR5 receptor in the absence of added ligand, consistent with receptor phosphorylation in the absence of exogenous ligand in CCR5-transfected HEK-293 cells (Fig. 4, A and B).

Neither immunoprecipitation of cell lysates with isotype- matched control antibodies to other cell proteins such as β₂-microglobulin nor immunoprecipitation in mock-transfected HEK-293 cells (not shown) revealed the presence of JAK1 (Fig. 5 A). This rules out nonspecific protein association to membrane components under these experimental conditions. The rapid association of JAK1 to the CCR5 receptor triggered by these two chemokines, but not by others tested (MCP-1 and SDF-1α), suggests a role for this tyrosine kinase in early receptor phosphorylation following ligand stimulation. Interestingly, while MCP-1 promotes JAK2 association to the receptor in CCR2-transfected HEK-293 cells, neither JAK2 nor JAK3 associates to CCR5 after RANTES or (AOP)-RANTES activation of CCR5-transfected HEK-293 cells (data not shown).

To identify the downstream signaling pathway activated by the JAK1 kinase, we tested its association to members of the STAT transcription factor family. We analyzed JAK1 kinase-activated STAT transcription factors in anti-CCR5 immunoprecipitates, and found that STAT5 associates to the receptor complex in response to both ligands (Fig. 5 C). The association of STAT5 correlates in time with JAK1 phosphorylation and binding. Furthermore, the precipitated STAT5 is phosphorylated on tyrosine residues, showing it is in the activated state (Fig. 5 D). Immunoprecipitates of RANTES-stimulated cell lysates with isotype-matched control antibodies do not contain STAT5. These results were further validated by the reverse experiments, in which anti-STAT5 antibodies were used for immunoprecipitation and Western blots developed with either anti-PTyr or anti-CCR5 antibodies (not shown). Therefore, binding of both RANTES and (AOP)-RANTES to the CCR5 receptor induces receptor and JAK1 phosphorylation and STAT5 binding. Independent of the fact that other STAT transcriptional factors may be implicated in CCR5-mediated signaling, our data show that no major differences are observed in RANTES or (AOP)-RANTES activation of the JAK/STAT pathway.

Similar to growth factor–induced dimerization of tyrosine kinase receptors, some GPCR, including chemokine receptors like CCR2 and CXCR4, undergo ligand-induced dimerization (Hebert et al., 1996; Rodríguez-Frade et al., 1999; Vila-Coro, A.J., J.M. Rodríguez-Frade, A. Martín de Ana, M.C. Moreno-Ortiz, C. Martínez-A., and M. Mellado, manuscript submitted for publication). We have now tested whether RANTES and (AOP)-RANTES trigger CCR5 dimerization. DSS-mediated cross-linking in CCR5-transfected HEK-293 cells was carried out after RANTES or (AOP)-RANTES stimulation. In CCR5-transfected HEK-293 cells, but not in mock-transfected cells, a high molecular mass receptor species (76 kD) was observed following both stimuli. This band corresponds to the expected molecular mass of two CCR5 molecules, as assessed by immunoprecipitation with anti-CCR5 antibodies and Western blot developed with anti-CCR5 antibodies (Fig. 6 A). When CCR5-transfected HEK-293 cells were stimulated with SDF-1α, no CCR5 dimerization was observed under these experimental conditions (Fig. 6 A), despite the fact that HEK-293 cells express a functional CXCR4 receptor as determined by flow cytometry and SDF-1α–induced calcium mobilization assays (Fig. 6 B). When the same experiment was done using anti-CXCR4 mAb for immunoprecipitation and Western blot, a high molecular mass species (84 kD) of the CXCR4 receptor is observed only after SDF-1α, but not after RANTES or (AOP)-RANTES stimulation (Fig. 6 B).

All together, these data indicate specific ligand-induced dimerization, as SDF-1α is not able to induce CCR5 dimerization, while RANTES or (AOP)-RANTES are not able to induce CXCR4 dimers. In addition, association of CCR5 with other membrane components, including other chemokine receptors, can be discarded under these experimental conditions, as cells treated with SDF-1α render only CXCR4 dimers, and those treated with RANTES and (AOP)-RANTES render only CCR5 dimers. High molecular mass species containing CCR5 and CXCR4 simultaneously are not detected, as shown when cells are stimulated with SDF-1α or RANTES and (AOP)-RANTES, and when immunoprecipitation and Western blot are performed using anti-CCR5 or anti-CXCR4 antibodies, respectively. Both RANTES and (AOP)-RANTES, like other chemokines and similar to members of the large cytokine family, can trigger early signaling events such as receptor dimerization, followed by recruitment of tyrosine kinases that phosphorylate both the receptor and the STAT transcription factors.

The PTX dependence of RANTES- and (AOP)- RANTES–triggered Ca²⁺ mobilization indicates Gα_i involvement in this process. It has been demonstrated recently that chemokine receptor–mediated chemotaxis is triggered by the βγ subunit of Gα_i (Neptune and Bourne, 1997). In MCP-1–mediated chemotaxis, we have shown that Gα_i associates rapidly to the CCR2 receptor, releasing βγ, which is responsible for activating migration and cell polarization. Thus, we investigated whether the differences in chemotaxis triggering reflect differences in RANTES- and (AOP)-RANTES–mediated Gα_i-CCR5 receptor association.

CCR5- or mock-transfected HEK-293 cells were stimulated with RANTES or (AOP)-RANTES, immunoprecipitated with anti-CCR5 or isotype-matched control antibodies, followed by a Western blot with anti-Gα_i antibodies. After both RANTES and (AOP)-RANTES stimulation, Gα_i associated to CCR5 (Fig. 7 A). Whereas association persisted for longer than 15 min when cells were RANTES-stimulated, Gα_i is dissociated from the receptor after 5 min of (AOP)-RANTES treatment (Fig. 7 A). The CCR5 expression level was controlled in both immunoprecipitates by stripping and reblotting membranes with anti-CCR5 antibody (Fig. 7 A). These data concur with the similar Ca²⁺ mobilization promoted by both of these ligands (Fig. 2 A), as Gα_i association to CCR5 is unaltered at the times employed for these experiments (1–3 min). Rapid Gα_i dissociation from the CCR5, induced only by (AOP)-RANTES, implies a role for G_i in later chemokine-triggered events such as chemotaxis. This result concurs with data showing the importance of βγ release from G_i in chemotaxis (Neptune and Bourne, 1997).

The activation of p¹²⁵FAK and the kinase ZAP-70 has been described in T cells after RANTES activation (Bacon et al., 1996). Therefore, we tested whether p¹²⁵FAK activation coincides with its association to the CCR5. CCR5-transfected HEK-293 cells were stimulated with 10 nM RANTES or (AOP)-RANTES, cell lysates were immunoprecipitated with anti-p¹²⁵FAK or an isotype-matched control antibody and analyzed in Western blot for the presence of CCR5. RANTES, but not (AOP)- RANTES, promotes the association of this kinase to the receptor, reaching maximum binding after 15 min (Fig. 7 B). Even in the absence of ligand, residual association between these two proteins was observed. The lack of p¹²⁵FAK association and the rapid Gα_i dissociation from the CCR5 receptor in the case of (AOP)-RANTES suggest that association of these proteins with the receptor is crucial in triggering the signaling pathways that lead to the different effects promoted by these two chemokines.

Interest is rapidly growing in a broad view of the chemokine function, not only because of significance in HIV infection, but also because chemokines and their receptors are expressed by a wide variety of nonhematopoietic cells (Horuk et al., 1997). Because of the relevance of chemokines in numerous pathologies, it appears of the utmost importance to devise therapies based on the administration of chemokines that lack chemotactic and leukocyte-activating properties, but are able to act as receptor antagonists. These are the basic characteristics of (AOP)- RANTES, a RANTES analogue that has been chemically modified at its NH₂ terminus, although very little is known of the mechanism underlying its action.

We have generated anti-CCR5 mAbs and used them to characterize the signaling pathways activated by RANTES and (AOP)-RANTES. The results show that ligand-induced CCR5 dimerization occurs when these chemokines signal target cells for activation. In the response to cytokines, receptor dimerization triggers a pathway involving tyrosine kinases and transcriptional factors. Only the JAK1 tyrosine kinase responds to RANTES and to (AOP)-RANTES by rapidly associating to CCR5 and is phosphorylated as soon as 30 s after binding, indicating that JAK1 activation is virtually simultaneous with its association to CCR5. In the cytokine receptors, activation and association of JAK kinase to the receptor creates docking sites for SH2-containing proteins such as STAT, leading to their phosphorylation, followed by activation of gene transcription (Ihle, 1996). After binding of either of the chemokine ligands studied here, STAT5 is also associated to the CCR5, in accordance with the role of the JAK kinases in transducing signals from hematopoietic growth factor receptors (O'Shea, 1997). Our data concur with a recent report showing rapid activation of certain STAT transcription factors in T cells after RANTES and MIP-1α stimulation, although in that case association to the CCR5 receptor was not analyzed (Wong and Fish, 1998). Together, these results indicate that chemokine-mediated activation of GPCR leads to signal transduction, which invokes intracellular phosphorylation intermediates used by other cytokine receptors. These earlier events are equally activated by both RANTES and (AOP)-RANTES.

Both RANTES and (AOP)-RANTES induce CCR5 dimerization, a phenomenon that requires specific interaction between a given chemokine and its respective receptor. When experiments were performed using SDF-1α as a ligand neither CCR5 dimers nor CCR5-CXCR4 heterodimers were observed, whereas CXCR4 dimers are present in anti-CXCR4 immunoprecipitates. Direct interaction between two CCR5 molecules has also been demonstrated by Benkirane et al. (1997), even in the absence of ligand stimulation. The functional significance of dimerization was suggested by Hebert et al. (1996) using the epitope tagging approach, showing that agonist stimulation of the β₂-adrenergic receptor stabilized the dimeric state of the receptor. The present data indicate that although dimerization is a primary event following ligand interaction with the chemokine receptor, it does not necessarily initiate identical transduction events. (AOP)- RANTES thus triggers Ca²⁺ mobilization as does RANTES, but promotes neither cell migration nor CCR5 redistribution to the leading edge in cell polarization. The rapid Gα_i association to CCR5 induced by both ligands explains the similar Ca²⁺ mobilization response and lack of response in PTX-pretreated cells.

It has been shown for a number of GPCR that different ligands of a given receptor can induce different responses. This is the case for CXCR2, from which IL-8 and NAP-2 elicit different responses, although they bind with equivalent affinity to this receptor (Ben-Baruch et al., 1997). It is feasible that RANTES and (AOP)-RANTES induce distinct conformational changes in the CCR5. In fact, it has been shown that different structural requirements are needed in GPCR to activate different pathways (Hunyady et al., 1995; Arai et al., 1997a). In the case of angiotensin II receptor changes have been described in internalization kinetics that do not alter other functional responses (Hunyady et al., 1995).

Whereas Gα_i association to CCR5 persists as long as 15 min after RANTES treatment, it dissociates from CCR5 by 5 min after (AOP)-RANTES stimulation. This variance must give rise to a clear difference in the availability of free, active βγ subunits by affecting some ligand-induced responses. In fact, G_i βγ subunit release is known to be important in chemokine receptor–mediated chemotaxis (Arai et al., 1997b; Neptune and Bourne, 1997). Another conspicuous difference is the lack of late tyrosine phosphorylation in CCR5 after (AOP)-RANTES treatment, in contrast to the continuous tyrosine phosphorylation of CCR5 observed following RANTES induction. This may occur because of tyrosine phosphorylation mediated by other tyrosine kinases. The activation of both p¹²⁵FAK and ZAP-70 has been described after RANTES stimulation (Bacon et al., 1996). Here we demonstrate association of p¹²⁵FAK, which is related to cytoskeletal proteins and migration phenomena (Sieg et al., 1998), to the receptor following RANTES but not (AOP)-RANTES stimulation. The failure of (AOP)-RANTES to maintain G_i association for long (15 min) periods concurs with the lack of association and activation of p¹²⁵FAK and the inactivation of subsequent events. Both activation of G_i and p¹²⁵FAK, as well as chemotaxis, occur after RANTES activation, but are absent when cells were treated with (AOP)-RANTES, indicating a possible relationship between these events.

Chemotaxis involves several phenomena including: changes in cell shape, integrin affinity, and integrin recycling at the cell's leading edge (Condeelis, 1993; Lawson and Maxfield, 1995). These events appear to be mediated by G protein activation, as well as phosphorylation signals through chemokine receptors (Bokoch, 1995). In fibroblasts and smooth muscle cells, the regulation of intracellular FAK and its interaction with the cytoskeletal proteins α-actinin, talin, and vinculin have also been given much attention with respect to cell polarization and migration (Clark and Brugge, 1995; Miyamoto et al., 1995). As a consequence of RANTES stimulation, T cells undergo a polarization process in which the CCR5 receptor redistributes to the leading cell edge and p¹²⁵FAK associates to the receptor. Both events, implicated in the late response to chemokines, are required for chemotaxis and are triggered by RANTES. Early (AOP)-RANTES–induced signaling events, such as Ca²⁺ mobilization, result in weak, transient cell polarization, and redistribution of the ICAM-3 adhesion receptor to the T lymphocyte uropod. This antagonist does not mediate cellular chemotaxis. Recycling of CCR5 molecules to the cell membrane, which probably occurs at the advancing front of the cell and is inhibited by (AOP)- RANTES (Mack et al., 1998), appears to be critical for the mechanism of directional cell migration towards a chemokine gradient. This is further supported by our data showing the failure of CCR5 to cluster at the leading edge of (AOP)-RANTES–stimulated cells. Therefore, these results suggest that during the chemotaxis process early signals induced by chemokines through receptor dimerization may not be sufficient to induce cell migration. Signals triggered by sustained association of the G protein to the chemokine receptor may also regulate other events necessary for the chemotactic response, such as integrin affinity or actin polymerization at the leading cell edge. Our data also support the idea that p¹²⁵FAK kinase association to the receptor is not involved in ligand-induced CCR5 internalization, since (AOP)-RANTES is as efficient as RANTES in promoting internalization (Mack et al., 1998), without inducing p¹²⁵FAK association.

The data presented here can be incorporated in a model in which signaling through chemokine receptors directly involves early signals that occur in the first few minutes after ligand binding, including receptor dimerization; association and activation of JAK tyrosine kinases, and activation of STAT transcriptional factors, as well as late signals such as cell polarization, activation, and association of p¹²⁵FAK kinase to the chemokine receptor (Fig. 8). The fact that chemokine receptors activate the JAK/STAT pathway and that dimerization may be a general mechanism for chemokine activity adds a new perspective to understanding how the multiplicity of chemokine functions are achieved. Furthermore, it suggests an interesting new objective for therapeutic intervention in chemokine-associated pathologies, including inflammation and AIDS.

Internalization of chemokine receptors is a phenomenon that can be dissociated from their role as HIV-1 coreceptors (Aramori et al., 1997; Howard et al., 1998). Although the ability of the coreceptor to internalize is not required for HIV entry, it contributes to the HIV suppressive effect of CC and CXC chemokines (Amara et al., 1997). (AOP)-RANTES is a potent inhibitor of HIV-1 infectivity, an effect that is suggested because it enhanced CCR5 internalization and inhibition of recycling (Mack et al., 1998). As internalization is an agonist-driven event, (AOP)-RANTES as well as other NH₂-terminal modified chemokines should be able to activate their receptors. The differences between RANTES- and (AOP)-RANTES– induced internalization may be related to increased affinity of (AOP)-RANTES for the CCR5 receptor (Mack et al., 1998), although differences in the signaling pathways activated by these two ligands have not been reported. Here we correlate the known effects of (AOP)-RANTES with a signaling cascade and suggest that the dimerization of chemokine receptors is an alternative mechanism for chemokine-induced HIV-1 inhibition.

Receptor dimerization is also the mechanism through which antibodies to the CCR2 NH₂-terminal domain block HIV-1 infection. In fact, the anti-CCR2 mAb CCR2-01 that blocks HIV-1 infection (Frade et al., 1997b), but not MCP-1–induced responses (Frade et al., 1997a), triggers CCR2 dimerization (Mellado, M., manuscript in preparation). The concept that receptor dimerization is required for prevention of HIV-1 infection, besides including the nonsignaling mechanism indicated by (AOP)-RANTES, incorporates the model based on receptor internalization, which does not take place in the absence of dimerization. For both receptor desensitization and internalization, we have shown the requirement for GRK2 association and phosphorylation in serine/threonine residues of the COOH- terminal receptor domain. This in turn associates with arrestin and clathrin, and undergoes internalization (Aragay et al., 1998). Dimerization is a necessary event for the formation of the receptor/GRK-2/arrestin/clathrin complex. This model consolidates all available experimental evidence, and offers novel prospects for screening new drugs that may prevent HIV-1 infection and/or inflammatory processes.

We thank Drs. R. Geha (Harvard Medical School, Boston, MA) and J.C. Gutierrez-Ramos (Millennium Pharmaceuticals, Boston, MA) for critical reading of the manuscript, Dr. J. Gutiérrez for CCR5 transfectants, L. Gómez for help with animals, M.C. Moreno and I. López for FACS^® analysis, and C. Bastos and C. Mark for secretarial and editorial assistance, respectively.

This work was supported by grants from the Spanish Dirección General de Investigaciones Cientificas, European Union, the Spanish CICyT and the Regional Government of Madrid (CAM), and Pharmacia & Upjohn. The Department of Immunology and Oncology was founded and is supported by the Spanish Research Council (CSIC) and Pharmacia & Upjohn.

DSS

disuccinimidyl suberate

FAK

focal adhesion kinase

FN

fibronectin

GPCR

G protein–coupled receptors

JAK

Janus associated kinase

PTX

pertussis toxin