|Home | About | Journals | Submit | Contact Us | Français|
C-X-C motif ligand12 (CXCL12) and its receptor, CXCR4, play an important role in hematopoietic stem/progenitor cells’ (HSPCs) migration from and retention within the bone marrow. HSPCs are very selective in their chemotactic response and undergo chemotaxis only in response to CXCL12. In addition to CXCR4, HSPCs express receptors for various other chemokines; however, the role of these receptors is not well understood. Freshly isolated CD34+ cells (highly enriched for HSPCs) from cord blood (CB) express low levels of C-C motif receptor 5 (CCR5); however, if the cells were washed with acidic buffer prior to antibody staining, to remove any ligand bound to CCR5, then nearly 80% of CD34+CB cells were found to express CCR5 on the cell surface. Although none of the CCR5 ligands investigated in this study (CCL3, CCL4 and CCL5) induced chemotaxis, at relatively high concentration they transiently enhanced CXCL12 mediated chemotaxis of CD34+CB cells. In contrast, CXCL12 mediated adhesion of cells to VCAM-1 coated surfaces was reduced if CD34+CB cells were pretreated with these CCR5-ligands for 15 mins. The effect of these chemokines on CXCL12 mediated responses was not at the level of CXCR4 expression, but on downstream signaling pathways elicited by CXCL12. Pretreatment with CCR5 chemokines enhanced CXCL12 mediated Akt phosphorylation, but down-modulated calcium flux in CD34+CB cells. Modulation of CXCL12 mediated responses of CD34+ cells by CCR5 chemokines provides a possible mechanism that underlies movement of HSPCs during inflammation.
At steady-state, hematopoietic stem and progenitor cells (HSPCs) are largely confined within bone marrow which provides the most suitable environment for maintenance of both hematopoietic stem cells and proliferating progenitor cells. Although the exact mechanism and the molecular interactions that underlie the retention of HSPCs in bone marrow are not clearly understood, their retention depends on adhesive interactions between HSPCs and stomal cells in the marrow. Based on various studies it is evident that the C-X-C ligand12 (CXCL12-also known as stromal derived factor-1)/ CXCR4 axis plays a pivotal role in retention of HSPCs in the bone marrow (1–4). At steady-state only a few HSPCs are found in the circulation, however, disruption of the CXCL12/ CXCR4 axis using pharmacological agents leads to increased release of HSPCs from bone marrow into circulation (3,4).
Among various chemokines and chemokine receptors, the role of CXCL12 and CXCR4 in modulating chemotactic response and adhesion to extracellular matrix of HSPCs has been widely studied (5–7). HSPCs are selective in their chemotactic response and undergo chemotaxis only in response to CXCL12 (8). It is however known that besides CXCR4, HSPCs express receptors for various other chemokines (8–10); the potential role of the other chemokine receptors in regulating HSPCs function(s) remain(s) largely unknown.
CXCR4 is a G-protein coupled receptor (G-PCR) and desensitization of G-PCRs is a physiologically important and complex process that participates in the turning-off of G-PCRs (11). Homologous desensitization involves turning-off of G-PCR when activated with its ligand, whereas heterologous desensitization refers to processes whereby the activation of one G-PCR can result in the inhibition of another heterologous G-PCR to signal (12). During certain inflammatory states there is an increase in various chemokines (13–15), including CCR5-ligands, and also in the circulating pool of HSPCs (16–18). It is therefore important to investigate if there is any cross-talk between CXCR4 and other chemokine receptors in HSPCs. It is possible that elevated levels of certain chemokines may modulate CXCL12 mediated responses of HSPCs.
We studied the effects of CCL3, CCL4 and CCL5, chemokines that share CCR5 as a common receptor, on CXCL12 mediated responses of CD34+CB cells. We show that although CCR5 responsive chemokines do not induce chemotaxis of CD34+CB cells, at relatively high concentrations they accelerate CXCL12 mediated chemotaxis. In contrast, they down modulate CXCL12 mediated adhesion of CD34+CB cells to VCAM-1. This demonstrates cross talk between CCR5 and CXCR4, which may have implications in migration of HSPCs during inflammation.
Human cord blood (CB) was obtained after informed consent with institutional review board approval and CD34+CB cells were isolated as described previously (7).
Chemotaxis assays were performed using 24-well chemotaxis chambers, pore size 5.0µM (Corning Costar, Cambridge, MA). CD34+CB cells in 100 µl of chemotaxis medium (IMDM+1% BSA) were added to the top well of a transwell chamber. Chemotaxis medium alone (600µl) or containing 200 ng/ml of CXCL12 was added to bottom well. To evaluate the effect of CCR5 responsive chemokines on CXCL12 mediated chemotaxis, CCL3, CCL4, or CCL5 (R&D, Minneapolis, MN) was added to the top well along with CD34+CB cells. These chemokines were formulated and stored in conditions described by the manufacturer to retain their maximum activity. Dose response analysis of the CCR5 ligands in modulating CXCL12 mediated chemotaxis of CD34+CB cells was performed and 1000ng/ml of CCR5 ligands was found to be the optimal dose. Therefore, for all the assays described in this study 1000 ng/ml CCR5 ligand has been used unless otherwise indicated. In some cases these chemokines alone were added in the bottom well. Chemotaxis of CD34+ CB cells was evaluated for 30 or 90 mins at 37°C in 5% CO2 humidified atmosphere. At the end of chemotaxis assay, cells that had migrated to the lower chamber as well as input cells were counted for 30 secs using FACScan under identical flow conditions. In a few experiments, CCR5 responsive chemokines were added to both the upper and lower well and the movement of cells was assessed as described above. To assess whether CCR5 binding chemokines modulate CXCL12 mediated chemotaxis across a wide concentration range of CXCL12, CD34+CB cells along with CCL4 (1000ng/ml) were added to the top well and various concentrations of CXCL12 (0–1000ng/ml) was added to the bottom well of transwell chamber and chemotaxis was evalauted as described above. To confirm that CCL3, CCL4 and CCL5 were modulating CXCL12 mediated responses via CCR5, CD34+CB cells were pretreated with either anti-CCR5 blocking antibody (R&D, Mineapolis, MN) (20µg/ml) or control antibody for 45 mins and then cells were added to the top well along with CCL3, CCL4, and CCL5, and chemotaxis towards CXCL12 was assessed as described above. To determine the specificity of the role of CCR5 chemokines on CXCL12 mediated chemotaxis we adopted two approaches. Firstly, heat inactivated (HI) CCL4 (10 min at 100°C) or CCL4 was added to the top well along with CD34+CB cells and the ability of HI-CCL4 and CCL4 to modulate CXCL12 mediated chemotaxis was evaluated as described above. Secondly, in a few experiments, CD34+CB cells were added to the top well along with Met-RANTES (1000ng/ml), a CCR5 and CCR1 antagonist (19) alone or in the presence of CCL5 (1000ng/ml) and chemotaxis of the cells towards CXCL12 was assessed as described above.
Chemokine receptor internalization was studied as described (20). Cells were incubated at 37°C for various time periods with 1000 ng/ml CCL3, CCL4 or CCL5. After washing the cells once with acidic glycine buffer (pH 2.7) (21), followed by a wash in phosphate buffered saline (PBS), receptor expression on the cell surface was determined using fluorescein isothiocynate (FITC)-conjugated CCR5 antibody and allophycocyanin (APC)-conjugated CXCR4 antibody (R&D, Minneapolis, MN) in combination with phycoerythrin (PE)-conjugated CD34 antibody (Miltenyi Biotec). Background fluorescence was evaluated using corresponding PE, APC and FITC-conjugated isotype control antibodies.
CCR5 and CXCR4 expression was also visualized by confocal microscopy. Cells were fixed and permeabilized by adding 0.5 ml of cytofix/cytoperm solution (BD Biosciences). Cells were incubated at 37°C for 10 mins and washed twice using Perm/wash buffer (BD Biosciences). Cells were stained with anti-CCR5-FITC and anti-CXCR4-APC antibodies and examined by Zeiss confocal microscope.
Freshly enriched CD34+CB cells were washed, resuspended in FACScan buffer [PBS (phosphate-buffered saline) + 2 mM ethylenediaminetetraacetic acid (EDTA) + 2% fetal calf serum (FCS)], incubated for 5 minutes at room temperature with normal human serum (1:20 dilution), and then incubated with a recommended volume of anti-human antibodies: CD34 conjugated to FITC (Miltenyi), CXCR4 conjugated to PE (R&D) and CD38 conjugated to APC (BD Biosciences). Cells were incubated for 30 minutes at 4°C, washed, and fixed in 2% paraformaldehyde prior to analysis by FACScan using CellQuest software (BD Biosciences Immunocytometry Systems, San Jose, CA).
Calcium (Ca2+) flux induced by CXCL12 in CD34+CB cells was studied by flow cytometry (22). An equal volume of Fluo-3AM (stock concentration 2 mM; Molecular Probes, Eugene, OR) and Pluronic acid (stock concentration 20% wt/vol; Molecular Probes) was mixed just before use. Cells were washed and resuspended in IMDM + 2% BSA, and Fluo-3AM/Pluronic acid mix was added for a final Fluo-3/AM concentration of 4 µM. After incubation for 45 minutes at room temperature, cells were washed in Ca2+ flux assay buffer [Hanks Balanced Salt solution (HBSS) containing 20 mM HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) and 0.2% bovine serum albumin (BSA), pH 7.4] to remove extracellular dye, incubated for 10 minutes at room temperature, and analyzed by FACScan. Background fluorescence of each sample was measured, and CCL3, CCL4, CCL5 (1000ng/ml) or CXCL12 (200ng/ml) was then added to samples. Samples were quickly mixed by vortex, and Ca2+ influx was recorded using FACScan machine. The data was analyzed using FlowJo software (Tree Star Inc., Ashland, OR). In parallel experiments, Fluo-3AM loaded CD34+CB cells were sequentially stimulated with CCR5 ligand followed by CXCL12. Cells were treated at first with CCR5 ligand and Ca2+ influx was recorded for 120 secs; thereafter, the cells were stimulated with CXCL12 and Ca2+ influx recorded for another 120 secs.
F-actin polymerization assays were performed as described previously (23) with a few modifications. In brief, cells were resuspended in IMDM supplemented with 0.1% BSA at 106 cells/mL and pretreated with CCR5 ligands for 5 mins. CXCL12 (200 ng/mL) was added to the cell suspension. At various time intervals, cells were permeabilized and fixed using 0.2 mL permeabilizing/fixing solution (Cytofix/Cytoperm; BD Biosciences Pharmingen) and stained with 0.1 mL Phalloidin-rhodamine solution (4 × 10−7 M; Molecular Probes, Eugene, OR). Cells were incubated for 10 minutes at 37°C, and washed twice with wash buffer (BD Biosciences Pharmingen, Palo Alto, CA). Pellets were resuspended in 500 µL 2% paraformaldehyde solution. Fluorescence was measured using FACScan and mean fluorescence was calculated using Cell Quest software.
Adhesion assays were performed in 96-well plates (high-binding; Costar, Cambridge, MA) coated with purified VCAM-1 (R&D, Mineapolis, MN 10 µg/mL) overnight at 4°C in PBS. To block nonspecific binding sites, plates were subsequently incubated for another 2 hours at 37°C with 2% BSA in PBS. CD34+CB cells were pretreated with medium alone or CCL3, CCL4 or CCL5 for the indicated periods, and then added to wells and incubated at 37°C in a humidified atmosphere containing 5% CO2 for 30 minutes in the absence or presence of CXCL12 (200ng/ml). After incubation, nonadherent cells were removed, and the wells were washed 2 to 3 times with medium. Adherent cells were recovered by detaching them using Cell Dissociation Buffer (Gibco).
Freshly isolated CD34+ CB cells were kept in IMDM+2%BSA for 60 mins and then pretreated with CCL3, CCL4 or CCL5 (1000ng/ml) for 5 mins or left untreated. After 5 mins, the cells were stimulated with CXCL12 (200ng/ml) for a further 3 mins. A portion of the cells were kept unstimulated to study the background level of activation of the signaling pathways. The cells were spun down, washed and lysed using NP40-based lysis buffer (7). Whole cell protein lysates were prepared and western blot done as described (7).
Student 2-tailed t test was used for statistical analysis. The level of significance is indicated by P value.
We first examined CCR5 expression on CD34+CB cells by staining the cells with anti-CCR5 antibody (clone 45523) conjugated with FITC and examining them using flow cytometry. A small percentage of freshly isolated CD34+CB cells express CCR5 on the cell surface (Fig. 1Aiii). However, if CD34+CB cells were fixed and permeabilized prior to staining with anti-CCR5 antibody (clone 45523), greater than 80 % (range 70–95%) of CD34+CB cells were found to express high levels of CCR5 (Fig. 1Aiv). This observation suggested that a significant proportion of CD34+CB cells have a pool of intracellular CCR5. The intracellular pool of CCR5 was confirmed by examining fixed and permeabilized CD34+CB cells stained with anti-CCR5 (clone 45523) antibody under confocal microscope (Fig. 1B). Chemokine receptors have been reported to exist in antigenically distinct conformations in various cell types (24). Therefore, to rule out the possibility that low level of CCR5 expression detected on cell surface is due to inability of clone 45523 to recognize the epitope on CCR5 due to the conformation of CCR5 in which it exists on CD34+CB cells, we next used a panel of CCR5 antibodies against different epitopes on the CCR5 receptor to further evaluate expression of CCR5 on freshly isolated CD34+CB cells. As shown in Fig. 1C, the proportion of CD34+CB cells expressing CCR5 varied depending upon the clone of antibody used for detection. CCR5 positive CD34+ cells were detected mainly by clones 45523 (2CCR5Ab) and 45549 (4CCR5Ab), both in non-permeabilized and permeabilized cells. The level of CCR5 expression detected by both of these clones was similar. Both of these antibody clones are multi-domain reactive and require the presence of multiple extracellular loops of the CCR5 receptor for full activity (25). Very few CD34+ CB cells were found to be CCR5 positive when stained with anti-CCR5 antibody-clone 45502 (1CCR5Ab) which reacts with the N-terminal extracellular domain of the CCR5 receptor as well as with clone 45531 (3CCR5Ab), the epitope for which lies in the second half of extracellular loop2 (ECL2) of CCR5.
Earlier studies have shown that CXCL12 mediated signaling through CXCR4 can be modulated by CCR5 chemokines in B and T- cells (26,27). To assess whether CCR5 responsive chemokines could modulate CXCL12 mediated responses of CD34+CB cells, we first examined the effect of CCR5 chemokines on CXCL12 mediated chemotaxis of CD34+CB cells. We evaluated the effect of various concentrations of CCL4 on chemotaxis of CD34+CB cells towards CXCL12 As shown in Fig. 2A, chemotaxis of CD34+CB cells towards CXCL12 was significantly enhanced in the presence of CCL4 only when CCL4 was present at relatively high concentration (1000 ng/ml). A similar increase in CXCL12 induced chemotaxis was observed when CCL4 was added either to the top well alone or to both top and bottom wells of the transwell chamber (Fig. 2B). We next examined the effect of other CCR5 ligands, namely, CCL3 and CCL5, on chemotaxis of freshly isolated CD34+CB cells towards CXCL12. Similar to CCL4, these chemokines also enhanced chemotaxis of CD34+CB cells towards CXCL12 (Fig. 2C). However, none of these chemokines alone, in the absence of CXCL12, induced chemotaxis in CD34+CB cells; chemotaxis in response to these chemokines was similar to background levels that were generally in the 0.5–2% range (data not shown). Intriguingly, the stimulatory effect of these chemokines on CXCL12 mediated chemotaxis of CD34+CB cells was observed only for short term, up to 30 mins (Fig. 2D), and this effect was lost if chemotaxis was allowed to proceed for a longer time period. Thus, CCR5 responsive chemokines accelerate the initial chemotactic response of CD34+CB cells towards CXCL12. Furthermore, the enhancing effect on CXCL12 mediated chemotaxis by CCL4 was observed across a broad concentration range of CXCL12 (Fig. 2E).
Enhanced chemotaxis was observed at relatively high concentration of CCR5 ligands. Therefore, to verify that the effects observed are due to specific action of CCR5 binding ligands, the effect of heat-inactivated CCR5 chemokine at 1000ng/ml on CXCL12 mediated chemotaxis was evaluated. As shown in Fig.2F, heat inactivated chemokine did not have any effect on CXCL12 mediated chemotaxis of CD34+CB cells, demonstrating that denaturation of CCL4 abolishes it ability to modulate CXCL12 mediated chemotaxis. Thus, intact and functional CCR5 binding chemokine is required for the observed effect. Although CCL3, CCL4 and CCL5 can signal through more than one receptor, CCR5 is the only known common receptor shared by these chemokines. Therefore, the effects of these chemokines are likely to be mediated through interaction with CCR5. To verify the role of CCR5, CD34+CB cells were pretreated with anti-CCR5 blocking antibody and the effect of anti-CCR5 antibody on chemotaxis evaluated. As shown in Fig. 2F, CCR5 antibody abrogated the enhancing effect of CCL4 on CXCL12 mediated chemotaxis of CD34+ cells. In addition we also evaluated the effect of Met-RANTES, a CCR5 and CCR1 antagonist (19), on enhanced CXCL12 mediated chemotaxis by CCR5 ligands. Unlike CCR5 ligands, Met-RANTES did not have any effect on CXCL12 mediated chemotaxis (Fig. 2G). Moreover, in the presence of Met-RANTES, the ability of CCL5 to enhance CXCL12 mediated chemotaxis was abrogated (Fig. 2G).
We next examined whether this effect was reproduced in a more primitive hematopoietic cell compartment. CD34++/CD38lo cells identify primitive hematopoietic populations (28). We found that when chemotaxis was evaluated after 30 mins, enhancement of CXCL12 mediated chemotaxis by the CCR5-ligands was also apparent in the more immature subset, in the CD34++/CD38lo cells (Fig. 2H).
In addition to chemotaxis, CXCL12 also modulates adhesion of HSPCs to VCAM-1(6). Retention of HSPCs in the bone marrow is dependent on adhesive interaction between HSPCs and the stromal cells. Adhesion by the VLA-4-VCAM-1 axis plays a role in retention of HSPCs in the bone marrow (29,30). Since CXCL12 is known to activate VLA-4 on HSPCs and CXCL12 attenuates VLA-4-VCAM-1 interaction (6), w e examined the effect of CCL3, CCL4 and CCL5 on CXCL12 stimulated adhesion of CD34+CB cells to VCAM-1. CD34+CB cells were pretreated with CCL3, CCL4 or CCL5 (1000ng/ml) for 15 mins and then the ability of these cells to adhere to VCAM-1 coated surface in the presence or absence of CXCL12 was examined. As reported earlier (6) CXCL12 enhanced adhesion of CD34+CB cells to VCAM-1 coated surface (Fig. 3A). CCR5 chemokines did not have any significant effect on adhesion of CD34+CB cells to VCAM-1 (Fig.3A). However, unlike chemotaxis, pretreatment with the CCR5 chemokines significantly abrogated CXCL12 mediated enhanced adhesion of CD34+CB cells to VCAM-1 coated surfaces (Fig 3A). Down-modulation of CXCL12 mediated adhesion by these chemokine was observed if the cells were pretreated with the CCR5 ligands at 1000 ng/ml for 15 mins and lasted for 30 mins of pretreatment (Fig. 3B). HI-CCL4 did not have any effect on adhesion (data not shown). Pretreatment for 60 mins did not have much, if any, effect on CXCL12 mediated adhesion of CD34+CB cells (Fig. 3B).
Since adhesion of CD34+CB cells to VCAM-1 was decreased by pre-exposure of these cells to CCR5 chemokines, we next investigated whether this effect was due to change in VLA4 expression. As shown in Fig. 3C, CD34+CB cells express VLA-4 and the level of expression of VLA-4 was not changed upon exposure to CCR5 chemokines.
To gain insight into the mechanism as to how CCR5 responsive chemokines modulate CXCL12 mediated responses in CD34+CB cells, we examined the effect of exposure to CCR5 ligands on CCR5 and CXCR4 expression in these cells. CD34+CB cells were exposed to 1000ng/ml of CCL3, CCL4 or CCL5 for 1 and 30 mins and CCR5 expression on the cell surface was studied by staining the cells with anti-CCR5 antibody (clone 45523) conjugated with FITC. 1000 ng/ml of CCR5-ligands was used since this dose was found to be an optimal dose for modulation of CXCL12 mediated chemotaxis (Fig. 2A). Since the epitope for CCR5 antibody (clone 45523) overlaps with the ligand binding site on CCR5 receptor (25), to assure that antibody binding to the receptor is not affected due to receptor occupancy by the ligand, the cells were washed in acidic buffer prior to staining with antibody. Wash in acidic buffer removes ligand bound to receptor (21). Surprisingly, after acid wash we could detect greater than or equal to 80% of CD34+CB cells expressing CCR5 on the cell surface without or upon exposure to CCL4 for 1 min (Fig. 4Aii and iii) and also to CCL3 and CCL5 (data not shown). It has been earlier reported that CD34+ cells produce low level of CCR5 ligands that bind to CCR5 receptors on the cells leading to their internalization, thus explaining the low level of CCR5 expression on the surface of CD34+ cells (10). However, it is also possible that the low level of CCR5 detected on the surface of freshly isolated CD34+CB cells is due to receptor occupancy by the ligand, that interferes with binding of antibody to the receptor, leading to underestimation of CCR5 expression on freshly isolated CD34+ cells. Indeed, the selective effect of acid wash on surface expression of CCR5 but not on CXCR4 expression on CD34+CB cells (Fig.4Bi and ii) demonstrates that the majority of CD34+CB cells express CCR5 on the cell surface. To rule out the possibility that the process of acid wash induces a conformation change in CCR5 leading to increased detection by the CCR5 antibody, we added CCL4 (100ng/ml) to the acid washed CD34+CB cells and then evaluated whether in the presence of the ligand there is change in detectable level of CCR5 on acid washed CD34+CB cells. As shown in Fig 4C i–v, in the presence of CCL4, CCR5 expression detected on acid washed CD34+CB cells is significantly reduced and also, the cells which were CCR5 positive, had lower detectable level of CCR5 (than acid washed cells in the absence of ligand). These findings demonstrate that binding of ligand to CCR5 receptor does indeed interfere with its detection by antibody and increased expression of CCR5 on acid washed compared to freshly isolated CD34+CB cells is not due to conformational change in CCR5, but due to removal of bound ligand to the receptor. Thus, the surface expression of CCR5 on freshly isolated CD34+CB cells is underestimated due to ligand occupancy. Moreover, since CXCR4 expression on CD34+CB cells (Fig 4B i and ii) was not altered upon acid wash, this confirmed that increased CCR5 expression level observed upon acid wash was not due to any generalized damage to the cell membrane. At 30 min post exposure of CD34+CB cells to CCL4, CCR5 expression level on the surface as well as the percentage of CD34+CB cells expressing CCR5 was reduced (Fig. 4Aiv). Down modulation of CCR5 expression at 30 min post CCL4 exposure, is likely due to internalization of the CCR5 receptor (31).
We also evaluated the effect of exposure to CCR5 chemokines on CXCR4 expression on CD34+CB cells. As shown in the Fig. 4B, treatment with CCL4 (as well as CCL3 and CCL5-data not shown) did not have any significant effect on CXCR4 expression on CD34+CB cells thus suggesting that modulation of CXCL12 mediated responses by CCR5 chemokines is not due change in CXCR4 expression.
CXCL12 is known to induce transient calcium mobilization in CD34+ cells (5,7) and calcium flux regulates adhesion (32–34). We, therefore, examined whether the CCR5 chemokines had any effect on CXCL12 induced calcium mobilization in CD34+ cells. As reported earlier (7,35), CXCL12 stimulation of CD34+ cells led to calcium mobilization (Fig 5Ai–iii). None of the CCR5 chemokines investigated in this study alone, led to any significant calcium mobilization in CD34+CB cells. However, if the cells were pre-stimulated with any one of the CCR5 chemokines prior to CXCL12 stimulation, then calcium mobilization induced by CXCL12 was significantly reduced (Fig. 5A i–iii).
Chemokines induce F-actin polymerization (7), that plays an important role in chemotaxis (36). Since CCR5 chemokines increase chemotaxis of CD34+ cells to CXCL12, we investigated whether these chemokines affected CXCL12 mediated F-actin polymerization. As shown in Fig 5B, CXCL12 induced F-actin polymerization was significantly enhanced when the cells were pretreated with CCL3, CCL4 or CCL5.
Since CCR5 responsive chemokines altered chemotactic responses of CD34+CB cells to CXCL12 without altering CXR4 expression level, we hypothesized that downstream signals emanating from CXCR4 may be affected by pre-exposure of CD34+CB cells to CCR5 chemokines. CXCL12 has been shown to activate both Akt and Erk and both of these pathways have been implicated in chemotactic response (7,37,38). Therefore, we examined whether CCR5 chemokines have any effect on CXCL12 mediated Erk or Akt phosphorylation in CD34+CB cells. CD34+CB cells were pre-exposed to CCR5 chemokines for 5 mins and further stimulated with CXCL12 for 3 mins. As reported earlier (7,37), stimulation of CD34+CB cells with CXCL12 increased Erk1/2 phosphorylation (Fig. 5Ci). Although stimulation of CD34+CB cells with CCR5 ligands induced mild Erk1/2 phosphorylation, this was significantly lower than CXCL12 induced Erk1/2 phosphorylation (Fig. 5Ci). Moreover, CXCL12 induced Erk1/2 phosphorylation was not affected by pre-exposure of the cells to CCR5 chemokines (Fig. 5Ci). However, Akt phosphorylation in response to CXCL12 was significantly enhanced by pre-exposure of CD34+CB cells to CCR5 chemokines (Fig. 5Cii). CCR5 chemokines by themselves did not induce any significant Akt phosphorylation in these cells.
Since CXCL12 mediated Akt phosphorylation was significantly increased by pre-exposure of CD34+CB cells to CCR5-ligand, we considered the possibility that increased Akt phosphorylation underlies the enhanced chemotactic response of CD34+CB cells to CXCL12, when pre-exposed to CCR5 ligands. To test this possibility, CD34+CB cells were pretreated with LY294002 (2 µM), a PI-3 kinase inhibitor, prior to chemotaxis assay. As shown in Fig. 6A, pretreatment with LY294002, abrogated the enhanced CXCL12 mediated chemotaxis observed in CCL4 pretreated CD34+CB cells. In addition LY294002 also reduced CXCL12 induced F-actin polymerization in CCL4 pretreated CD34+CB cells (Fig. 6B). Similar effect was also observed when CCL3 and CCL5 were used (data not shown).
HSPCs are extremely selective in their chemotactic response and CXCL12 is the only chemokine that induces chemotaxis of HSPCs (8). In addition to CXCR4, the receptor for CXCL12, HSPCs also express receptors for a variety of other chemokines, either on their surface at low levels, or intracellularly (8–10). However, the functional significance of the expression of other chemokine receptors on these cells is not well understood. During inflammation, levels of various chemokines including CCR5 ligands, are elevated (39) and there is a concomitant increase in circulating pool of HSPCs (16,17). In this study we demonstrate that CCR5 responsive chemokines, CCL3, CCL4 and CCL5, modulate CXCL12 mediated chemotaxis and adhesion of CD34+CB cells. Specifically, we found that these chemokines enhance chemotaxis of CD34+CB cells towards CXCL12; however, this effect is short-lasting; suggesting that the CCR5 ligands accelerate the initial chemotactic response. The effect of these chemokines on CXCL12 mediated chemotaxis could be blocked by anti-CCR5 antibody thus confirming the role of CCR5 in this process. Increased/accelerated chemotaxis towards CXCL12 in the presence of the CCR5-responsive chemokines was dependent on PI-3 kinase activation. In contrast to chemotaxis, CXCL12 mediated adhesion of CD34+CB cells to VCAM-1 was inhibited by these CCL3, CCL4 and CCL5. A diagrammatic representation of the processes determined in this study is shown in Fig. 7.
The three CCR5 responsive chemokines investigated in this study did not induce chemotaxis by themselves; however, when present at high concentration (1000ng/ml) they significantly enhanced short-term chemotaxis of CD34+CB cells towards CXCL12. Expression of CXCR4 remained unchanged upon exposure of CD34+CB cells to CCR5-ligands, suggesting that the effects of these CCR5-ligands are downstream of CXCR4. Although CCR5 receptor expression was detected in only 1–21% of freshly isolated CD34+ CB cells, upon fixing and permeabilizing the cell, greater than 80% of the CD34+CB cells expressed abundant CCR5 receptors. Furthermore, staining with a panel of anti-CCR5 antibodies revealed that both in permeabilized and non-permeabilized CD34+CB cells, CCR5 expression could be detected using two antibody clones (45523, 45549). These antibodies are multi domain reactive, the epitope for which includes ECL2 as well as other residues on CCR5 (25). It has been reported by others that CD34+CB cells produce low levels of CCR5 ligands and therefore, since the cells are exposed to ligand, this leads to internalization of the CCR5 in CD34+CB cells (10). Since the epitopes of both of the antibody clones (45523, 45549) overlap with the ligand binding region on CCR5 receptor (25), it was possible that the apparent level of CCR5 expression on CD34+CB cells was underestimated since receptor occupancy by ligand could directly interfere with antibody binding to receptor or indirectly due to steric hindrance. Indeed, we found that if the CD34+CB cells were washed with acidic buffer to remove bound ligand prior to staining with antibody, then CCR5 expression was detected in the majority of the CD34+CB cells. The use of acid wash to remove ligand from receptor has been used previously to study receptor expression (21). However, there was a possibility that the acid wash affected cell membrane integrity and made the cells leaky; therefore, the CCR5 detected upon acid wash could represent the intracellular pool of CCR5. This possibility was ruled out based on the finding that the expression of CXCR4 receptor was unchanged upon acid wash. Moreover, in the presence of CCR5 ligand, detection of CCR5 receptor in acid washed CD34+CB cells was significantly reduced. This suggests that CCR5 receptor is not detected in freshly isolated CD34+ cells due to receptor occupancy by ligand.
Although, freshly isolated CD34+CB cells appeared to express low levels of CCR5 receptors, the interaction between CCR5 ligands and CCR5 was evident since exposure of CD34+CB cells to the CCR5 chemokines resulted in time-dependent internalization of the CCR5 receptor. Phosphorylation of receptor upon ligand occupancy has been shown to be important for receptor internalization (43). Time dependent CCR5 internalization upon exposure to CCR5 ligands (Fig. 4Aiv compared to Fig. Aii) suggests that CCR5 is expressed on the cell surface and the CCR5 ligands bind to the receptor leading to intracellular events resulting in receptor internalization. In addition to action through their receptor(s), chemokines can also manifest biological functions by binding to glycoproteins (40). Also, at a higher concentration, a chemokine can interact with other chemokines and form heteromers and these heteromers can evoke biological responses (41, 42). The possibility of CCL3, CCL4 and CCL5 binding to glycoprotein and/heteromerization of these chemokines with CXCL12 for manifestation of the biological effects observed in this study can not be completely ruled out. However, our finding that Met-RANTES, a CCR5 and CCR1 antagonist, abrogated CCR5 ligand induced up-regulation of CXCL12 mediated chemotaxis, without having any effect on CXCL12 mediated chemotaxis by itself, suggests a role for CCR5 in enhanced CXCL12 mediated chemotaxis by CCR5 ligands. The role of CCR5 in manifestation of the biological effects of these chemokines was further established by using anti-CCR5 blocking antibodies which were found to abrogate enhanced CXCL12 mediated chemotaxis of CD34+CB cells by CCL4.
The CXCR4 chemokine receptor is a G-PCR that triggers multiple intracellular signals in response to its ligand, CXCL12. CXCL12 mediated chemotaxis involves multiple activation pathways and cooperation of several cytoplasmic domains of CXCR4 (7,37,44,45). CXCL12 stimulation leads to calcium mobilization, activation of Erk and PI-3 kinase pathways (37,38,44–46). The role of the individual signaling pathways in CXCL12 induced chemotaxis is not well understood. While PI-3 kinase and activation of its downstream target, Akt, have been implicated in chemotaxis (37), they are not the only players that regulate chemotaxis (37,38,46). A recent study has suggested that activation of PI-3 kinase accelerates initial chemotaxis (47), perhaps by helping directional sensing (48,49); but once the gradient of chemokine is established it is then dispensable for chemotaxis. In addition to the PI-3 kinase-Akt pathway, Erk1/Erk2 also regulates chemotaxis (7,37). Interestingly, pretreatment of CD34+CB cells with CCL3, CCL4 or CCL5, led to increased Akt phosphorylation upon CXCL12 stimulation compared to CD34+CB cells stimulated with CXCL12 alone and this was accompanied with an increased F-actin polymerization. Inhibition of PI-3 kinase, upstream of Akt, by LY294002, completely reversed CCR5 ligand-induced increased CXCL12 mediated chemotaxis of CD34+CB cells. This was also accompanied by decreased F-actin polymerization. However, the effect of LY294002 on CXCL12 alone mediated chemotaxis was less profound and F-actin polymerization was also not affected significantly (data not shown). Thus, it appears that enhanced CXCL12 mediated chemotaxis of CD34+CB cells observed in the presence of CCR5 ligands requires Akt activation.
Unlike chemotaxis, adhesion to VCAM-1, a ligand for VLA-4, in response to CXCL12 was inhibited by CCR5 ligands. Adhesion to VCAM-1 could be altered either due to increased VLA-4 expression or due to integrin-dependent cell avidity due to G-PCR activation (32,50). Our data shows that CCR5 chemokines did not alter VLA-4 expression on CD34+CB cells. However, pretreatment with CCR5 chemokines abrogated calcium flux induced by CXCL12 stimulation. Cytoplasmic calcium elevation has been shown to play an important role in increasing cell binding (33). Elevation of intracellular calcium induces a high affinity binding state of VLA-4 (34). Therefore, reduced calcium flux in CD34+CB cells pretreated with CCR5 chemokines may, in part, explain the abrogation of adhesion to VCAM-1 in CD34+CB cells pretreated with CCR5 responsive chemokines.
Our findings that show that CCR5 responsive chemokines have different effects on CXCL12 induced Akt phosphorylation (increased) and calcium flux (decreased) (see model in Fig. 7). It has been noted by others as well that stimulation of CXCR4 with CXCL12 results in calcium mobilization as well as PI-3 kinase activation but these pathways are independently activated (44). Moreover, calcium mobilization is uncoupled from CXCL12 mediated chemotaxis (51). While both intracellular loop (ICL) 2 and ICL3, and the carboxy tail of CXCR4 is required for CXCL12 mediated chemotaxis, ICL2 is dispensable for G(αi) mediated responses, including calcium flux (45). It remains unclear how ligand-activated CCR5 exerts its effects on CXCL12 mediated signaling pathways in CD34+CB cells, especially since CCR5 ligands alone, other than causing CCR5 internalization, did not stimulate either calcium flux or significant phosphorylation of Akt and Erk1/2 in CD34+CB cells. It is possible that in the presence of its ligand, CCR5 interacts with CXCR4 and thereby modulates CXCL12 stimulated signaling pathways. Previous studies had ruled out interaction between CXCR4 and CCR5 (52), however a recent study has shown that CCR5 can hetero-dimerize with CXCR4 in activated T cells at the immunological synapse (53). Based on the latter study it is possible that CCR5 and CXCR4 can hetero-dimerize under specific activation conditions and in a cell-type dependent manner, and that CCR5 ligands can variously affect CXCL12 mediated signaling pathways. In our study we found that in CD34+ CB cells, CCR5 ligands selectively down-modulate CCR5 expression without affecting CXCR4. This finding would argue against a stable CCR5-CXCR4 heterodimer in CCR5 ligand activated CD34+ CB cells. It is possible that in the presence of relatively high concentrations of CCR5 ligand, there is some sort of interaction between CCR5 and CXCR4, leading to conformational changes in CXCR4 resulting in modulation of various CXCL12 mediated signaling pathways. Alternatively, it is possible that CCR5 ligands activate protein kinases that can phosphorylate CXCR4 receptor or modify downstream targets, thus altering signaling pathways activated upon CXCL12 binding without affecting internalization of CXCR4 (54–56).
Homologous desensitization of CXCR4 by CXCL12 plays an important role in regulating CXCL12 mediated responses. In addition to homologous desensitization, heterologous modulation of CXCR4 has also been described (10,27). Heterologous modulation of CXCR4 by other chemokine receptors in CD34+ cells may play an important role during inflammation when levels of various chemokines are elevated. During inflammation, HSPCs are mobilized into the circulation (16,17). Unlike G-CSF induced mobilization of HSPCs, mobilization of HSPCs during inflammation and in response to chemokines is a very rapid and transient event (16,17). Our study shows that CXCL12 mediated responses can be modulated by CCR5 ligands; although the amount of CCR5 ligand required for modulating CXCL12 mediated responses was found to be relatively high. During inflammation, chemokine levels are elevated in serum and chemokine concentrations in the microenvironment could be significantly higher than that measured in serum. It would be interesting in the future to investigate whether the findings made in this study are involved mechanistically in mobilization of HSPCs during inflammation and in response to chemokines.
These studies were supported by U.S. Public Health Service Grants: RO1 HL 67384 and RO1 HL 56416 to HEB, and a project in PO1 HL 53586 to HEB.