Published online before print July 24, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Immunology, Weizmann Institute of Science, Rehovot, Israel;
Department of Hematology and Bone Marrow Transplantation, Chaim Sheba Medical Center, Tel-Hashomer, Israel;
|| Institute for Research in Biomedicine, Bellinzona, Switzerland;
Laboratory for Immunological and Molecular Cancer Research, Third Medical Department with Hematology, Medical Oncology, Hemostaseology, Rheumatology and Infectiology of the Paracelsus Medical University Salzburg, Salzburg, Austria; and
Department of Internal Medicine V, University of Heidelberg, Germany
1 Correspondence: Department of Immunology, The Weizmann Institute of Science, Rehovot, 76100, Israel. E-mail: ronen.alon{at}weizmann.ac.il
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: adhesion molecules • cell trafficking • shear • endothelium • migration
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4β1) and LFA-1 (Lß2). Endothelial-displayed CXCL12 can rapidly stimulate VLA-4 and LFA-1 adhesiveness to their respective endothelial ligands, VCAM-1 and ICAM-1, and thereby allow these integrins to mediate lymphocyte and CD34+ cell arrest on endothelial surfaces under disruptive shear forces [8
, 11
]. The primary CXCL12-binding receptor is CXCR4, and until recently, the CXCL12–CXCR4 relationship was thought to be exclusive [12 ]. CXCL12 and CXCR4 share multiple functional similarities, and CXCR4–/– and CXCL12–/– mice exhibit embryonic lethality at approximately the same point in fetal development [13 14 15 ]. Recently, an additional CXCL12-binding chemokine receptor, CXCR7 [RDC1, ChemoCentryx chemokine receptor 2 (CCXCKR2), ChemoCentryx Inc., Mountain View, CA, USA], has been identified and was suggested to mediate CXCR4-like functions in subsets of leukocytes [16 ]. CXCR7 possesses high sequence similarity with known chemokine receptors but displays some modifications in the Asp-Arg-Tyr-Leu-Arg-Arg/Ile-Val motif, which is critical for G-protein coupling by conventional GPCRs [17 , 18 ]. Like CXCR4, CXCR7 is highly conserved between human and mouse [19 , 20 ], and CXCR7–/– mice die rapidly after birth [21 ]. CXCR7 binds to CXCL12 and to the IFN-inducible T cell chemoattractant (CXCL11) [16 , 22 ]. The affinity of CXCL12 to CXCR7 is approximately tenfold higher than the affinity of CXCL12 to CXCR4 [16 , 22 ]. Membrane-associated CXCR7 is expressed on many tumor cell lines, activated endothelial cells, and early fetal liver cells, but its expression on the surface of hematopoietic and immune cells has been disputed [22 ]. Despite its high affinity to CXCL12, the role of CXCR7 in CXCL12-dependent cell motility and chemotaxis has also been a subject of debate [16 , 22 ]. A study published by Balabanian et al. [16 ] suggested that CXCL12 signals through CXCR7 expressed on the surface of primary T cells and that CXCR7 participates in lymphocyte motility. In contrast, another recent study reported that CXCL12 binding to CXCR7 does not elicit calcium mobilization in breast cancer cells and argued that CXCR7 is not involved in CXCL12-mediated cell migration [22 ].
Our earlier results suggested a crucial role for CXCR4 and Gi protein signaling in motility and rapid integrin activation of human PBL and CD34+ progenitor cells interacting with CXCL12 [9 ]. The identification of CXCR7 as a novel CXCL12-binding receptor and the suggestion that CXCL12 signals through CXCR7 to promote chemotaxis of T cells [16 ] prompted us to investigate the contribution of this novel receptor to CXCL12-mediated motility and integrin activation in human T lymphocytes. Compared with CXCR4, we detect only minute CXCR7 levels on the surface of T lymphocytes, and blocking CXCR4 on these cells eliminated all CXCL12-triggered motility and integrin activation processes, ruling out an intrinsic signaling activity of CXCR7 toward CXCL12. Nevertheless, a significant pool of intracellular CXCR7 got occupied by CXCR7-blocking mAb or a high-affinity, small molecule antagonist, and this severly impaired the ability of cell-surface CXCR4 to rapidly activate lymphocyte integrin adhesiveness under shear flow conditions. This novel crosstalk between CXCR7 and surface CXCR4 was not essential, however, for CXCL12-stimulated motility of T lymphocytes nor for general CXCR4 signaling to proximal targets such as Akt and ERK. Similar results were obtained with CD34+ cells. This is a first indication that nonsignaling GPCRs, which are largely excluded from the cell surface, can be targets for extracellular antagonists and suggests that CXCR7 blocking can be therapeutically useful for selective antagonism of CXCR4-mediated integrin activation rather than for general CXCR4 inhibition by direct, CXCR4-specific blockers.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4 HP1/2 mAb were kindly provided by Dr. Blake Pepinsky (Biogen, Cambridge, MA, USA). R-PE-conjugated anti-CXCR4 mAb (clone 12G5) was purchased from BioLegend (San Diego, CA, USA). The anti-CXCR7-blocking mAb, clone 9C4, is described elsewhere [23
]. The anti-CXCR7-blocking mAb, clone 11G8 (purified or R-PE-conjugated), and the low molecular weight compounds CCX733 and CCX266 (see Supplementary Fig. 2 and Supplementary Table 1) were the kind gift of Dr. Bretton Summers (ChemoCentryx Inc.) [22
]. Both molecules are members of the genus of compounds disclosed in U.S. Patent Application US20070167443. The anti-CXCR7 mAb, clone 358426, was purchased from R&D Systems but was not confirmed to be a blocking mAb and was merely used for staining. The anti-early endosome antigen 1 (EEA1) mAb TL12006 was from BD Transduction Labs (San Jose, CA, USA). Rabbit anti-linker for activation of T cells (LAT; #9166) was from Cell Signaling (Danvers, MA, USA). The β2-blocking antibody TS1/18 (IgG1) was provided by Timothy Springer (Harvard University, Cambridge, MA, USA). p44/42 and Akt antibodies were purchased from Cell Signaling Technology (Beverly, MA, USA).
Cells
Human PBMC were isolated from the whole blood of healthy donors [24
]. T lymphocytes were purified and cultured further in LPS-free RPMI/10% FCS for 15–18 h before use, as described [11
]. Results were also confirmed on freshly isolated T lymphocytes. Umbilical cords and cord blood (CB) were obtained from healthy donors. All cells were obtained after informed consent in accordance with procedures approved by the Human Ethics Committee of the Weizmann Institute of Science (Rehovot, Israel). CD34+ cells were purified with MACS (Miltenyi Biotec, Bergisch Gladbach, Germany) after standard separation of mononuclear cells from CB. HUVECs were cultured as described [9
].
Flow cytometry and immunofluorescence
Surface-expressed CXCR7 was identified by staining with 10 µg/ml anti-CXCR7 mAb (11G8, 9C4, or 358426) and Alexa488-conjugated secondary antibodies (Invitrogen, Carlsbad, CA, USA) at 4°C and analyzed on a FACScan (Becton Dickinson, Mountain View, CA, USA) compared with appropriate isotype controls. For detection of intracellular CXCR7 by immunofluorescence, cells were fixed with 4% paraformaldehyde/2% sucrose, permeabilized with saponin (0.2%), and stained with anti-CXCR7 mAb (11G8, 9C4, or 358426) and Alexa488-conjugated secondary antibodies. Where indicated, CXCR4 was stained with R-PE-conjugated anti-CXCR4 mAb in the presence of excess of mouse Igs. Serial Z-stacked (0.4 µM/section) confocal imaging was performed with a Delta Vision microscope (Applied Precision, Issaquah, WA, USA) with a 60x oil-dipping objective, numerical aperture (NA) 1.42. Deconvolution and three-dimensional image analysis were performed using Softworx (Applied Precision).
To assess mAb binding to the intracellular CXCR7 pool, intact cells were pretreated with 5 µg/ml R-PE-conjugated anti-CXCR7 mAb (clone 11G8) or R-PE-conjugated IgG1 isotype control for 30–120 min at 37°C in the presence or absence of CXCR7 blockers. Cells were washed in PBS at 4°C, fixed with 4% paraformaldehyde/2% sucrose, and subjected to flow cytometry. For calculation of the total CXCR7 pool, T cells were fixed with 4% paraformaldehyde/2% sucrose, permeabilized with 0.2% saponin, stained with the same R-PE-conjugated anti-CXCR7 mAb or corresponding isotype control, and subjected to flow cytometry under identical settings. The ratio between CXCR7 staining by the anti-CXCR7 mAb obtained in nonpermeabilized cells to the CXCR7 staining of prepermeabilized cells represented the relative access of the intracellular CXCR7 pool to the anti-CXCR7 mAb. All flow cytometry data were analyzed using the FlowJo 6.4 software (Tree Star Inc., Ashland, OR, USA).
In vitro shear flow assays
Laminar flow adhesion assays were performed as described [25
]. Plates were coated with protein A followed by coimmobilization of heat-inactivated or intact chemokine and subsequent overlay with ICAM-1-Fc [26
]. For preparation of sVCAM-1/chemokine substrates, sVCAM-1 was mixed in coating medium (PBS, 20 mM sodium bicarbonate, pH 8.5, 2 µg/ml HSA) and adsorbed overnight at 4°C, alone or with the indicated amounts of intact or heat-inactivated chemokines. Plates were blocked with HSA (20 µg/ml). VCAM-1-coating densities were comparable when the ligand was coimmobilized with heat-inactivated or intact chemokine (data not shown). HUVECs were preseeded on fibronectin-coated plates for 24 h, TNF--stimulated for 4 h (0.2 ng/ml, 5 U/ml; R&D Systems), and overlaid with CXCL12 where indicated. T cells were washed in 5 mM EDTA and suspended in HBSS containing 2 mg/ml BSA, 10 mM HEPES, pH 7.4. For CXCR7 and CXCR4 inhibition experiments, cells were pretreated with 10 µg/ml anti-CXCR7, 2.5 µM CCX733 or CCX266, 5 µM AMD3100, or 10 µg/ml anti-CXCR4 mAb (12G5) for 30 min, unless otherwise indicated. All flow experiments were performed at 37°C as described [27
]. Cellular interactions with the adhesive substrates were identified by computerized tracking of individual cell motion, and categories (transient, rolling, rolling followed by arrest, and immediate arrest) and frequencies of cell tethers were determined as described [11
]. Tethers were defined as transient if cells attached briefly (less than 2 s) to the substrate and as arrests, if immediately arrested and remaining stationary for at least 3 s of continuous flow. To assess adhesiveness of lymphocyte-surface CXCR4 during subsecond contacts, T lymphocytes were perfused over surface-bound, anti-GPCR mAb (12G5, 0.2 µg/ml) coated with inactivated or active CXCL12 or control chemokines. The frequency and type of T lymphocyte tethers served as a measure of spontaneous or chemokine-modulated CXCR4 adhesiveness to the mAb-coated surface.
Motility and chemotaxis assays
For motility assays on immobilized CXCL12, motility chamber microslides (ibidi, Integrated BioDiagnostics, Munich, Germany) were coated with 200 ng/ml CXCL12, washed, and blocked with HSA. T lymphocytes were added and immediately videotaped with Softworx 3.5 (Applied Precision) using a 40x/0.95 NA differential interference contrast objective. All experiments were performed at 37°C. Cells that exhibited leading and trailing edges for at least 1 min were considered polarized and were subdivided into "polarized stationary" or "polarized motile" categories depending on their ability to move for at least 30 µm from their original position during 10 min of tracking.
Chemotaxis assays of T lymphocytes toward CXCL12 were performed in RPMI 1640 with 0.5% BSA using transwell culture inserts with 5 µm pore size (Costar, Cambridge, MA, USA) for 2 h at 37°C in 5% CO2. Chemotaxis assays of CD34+ cells were performed as described [28 ].
Western blots
Untreated or pretreated cells were stimulated with CXCL12 as indicated. Cells were solubilized [29
], and lysates were separated by SDS-PAGE. Immunoreactive bands were visualized using HRP-conjugated secondary antibody and the ECL system (Sigma Chemical Co.).
Statistical analysis
For statistical comparison between groups, the paired, two-tailed Student’s t-test was used. Analyses were performed using the statistics tool of Origin 7.5 (OriginLab Corp., Northampton, MA, USA).
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (15K): [in a new window] |
Figure 1. CXCR7 expression in primary T lymphocytes. (A) CXCR7 and CXCR4 surface expression of resting T cells was determined by flow cytometry using the anti-CXCR7 mAb 358426 and the anti-CXCR4 mAb 12G5, respectively, and anti-mouse Alexa488 secondary antibodies. For detection of intracellular CXCR7, cells were fixed, permeabilized, and stained with the same antibodies. The fluorescence histograms show the relative fluorescence intensity of the cells stained with anti-CXCR7 and anti-CXCR4 (black lines) compared to the corresponding isotype control (tinted histograms). (B) Immunofluorescence stainings of representative, permeabilized T cells reacted with anti-CXCR7 (labeled with Alexa488-conjugated secondary antibody), anti-CXCR4 (directly labeled with PE), and 4',6-diamidino-2-phenylindole (DAPI; blue), as described in Materials and Methods. (C) Immunofluorescence stainings of permeabilized T cells reacted with anti-CXCR7 and anti-EEA1 mAb as well as DAPI, as described in Materials and Methods.
|
inhibitor PTX (Fig. 2C)
. Collectively, our data argue against any role of CXCR7 in early CXCR4–Gi signaling to ERK and PI-3K/Akt in rapid CXCL12–CXCR4–Gi-mediated, random T cell motility and in T cell chemotaxis toward a CXCL12 gradient.
 View larger version (51K): [in a new window] |
Figure 2. Effect of CXCR7 blockage on CXCL12-triggered motility and CXCR4 signaling under shear-free conditions. (A) Chemotaxis toward CXCL12 of T lymphocytes pretreated with the CXCR7-blocking mAb, 9C4 or 11G8, with a low molecular weight CXCR7 antagonist, or with the CXCR4 blocker AMD3100, as described in Materials and Methods. Results show the percentage of migrated cells (mean±range of duplicate samples, representative for three experiments). con, Control. (B) T lymphocytes were pretreated with CXCR7 blockers or AMD3100, and the cell phenotype and motility were monitored for 10 min on immobilized CXCL12. At least 120 cells were monitored in each group. Results are representative of two independent experiments with each CXCR7 inhibitor. (C) T cells were pretreated with various CXCR7 blockers, AMD3100, or pertussis toxin (PTX; 100 ng/ml overnight). Cells were then stimulated with 100 ng/ml CXCL12 for the indicated time-points, and lysates were prepared. Western blots were performed using antibodies against phospho-Akt, phospho-ERK1/2, and total p44/42.
|
50% (Fig. 3A)
. A second CXCR7-blocking mAb, clone 11G8, showed comparable blocking activity (data not shown, and see 
Fig. 6
below). Thus, in the absence of functional CXCR4, CXCR7, on its own, failed to mediate any LFA-1 activation by CXCL12. Nevertheless, optimal LFA-1 activation by CXCL12 required intact CXCR4 and CXCR7. In contrast to its interference with CXCR4-triggered LFA-1 activation, CXCR7 blocking did not interfere with LFA-1-mediated arrest triggered by other chemokines that do not bind CXCR4 (Fig. 3 B and C)
or to spontaneous chemokine-independent LFA-1 adhesiveness to high-density ICAM-1 (Fig. 3A
, right). Similarly to LFA-1 activation, in situ CXCL12-induced activation of VLA-4 led to augmented arrest of T cells on VCAM-1, which was highly susceptible to CXCR4 blocking with AMD3100 (Fig. 3D)
. Reminiscent of its inhibition in CXCL12 triggering of LFA-1 activation, the CXCR7-blocking mAb 9C4 inhibited VLA-4-mediated arrest on immobilized CXCL12 by 50% (Fig. 3D)
. As observed with LFA-1, CXCR7 blockage did not interfere with spontaneous VLA-4 adhesiveness to VCAM-1 (Fig. 3D
, right) and did not affect VLA-4 stimulation by irrelevant integrin-stimulatory chemokines that do not bind CXCR4, such as CCL21 and CXCL9 (Fig. 3 E and F)
. Having observed the potent inhibitory capacity of the CXCR7-blocking mAb 9C4 on CXCL12 triggered from LFA-1- and VLA-4-mediated adhesion, we next addressed the ability of a similar CXCR7-blocking treatment to interfere with T lymphocyte interactions with inflamed endothelium under shear flow. T lymphocytes were perfused over TNF--activated HUVEC, which mediate lymphocyte rolling through their E-selectin but fail to stimulate integrin activation and lymphocyte arrest on its ligands ICAM-1 and VCAM-1 via CXCR4, unless overlaid with CXCL12 [8
, 11
]. Consistently with the findings made on isolated ICAM-1 and VCAM-1, the majority of CXCL12-triggered lymphocyte arrests on TNF--activated HUVEC was inhibited by CXCR7 blocking (Fig. 3G)
, whereas all CXCL12-triggered arrests were eliminated by combined blocking of VLA-4 and LFA-1 (ref. [11
], and data not shown). Thus, intact lymphocyte CXCR7 is required for optimal CXCL12–CXCR4-triggered activation of VLA-4 and LFA-1 on T cells arrested on inflamed endothelium under shear flow.
 View larger version (36K): [in a new window] |
Figure 3. CXCR7 inhibition by two anti-CXCR7-blocking antibodies specifically inhibits CXCL12-dependent LFA-1 and VLA-4 activation at subsecond contacts. (A, left) T cells were pretreated with anti-VLA-4 (HP1/2) as a negative control antibody (Ab con), anti-CXCR7 mAb (9C4), anti-CXCR4 mAb (12G5), or AMD3100 (CXCR4 inhibitor). Cells were perfused for 1 min at 0.5 dyn/cm2 over ICAM-1 coimmobilized (imm.) with heat-inactivated (–) or functional (+) CXCL12. Arresting cells are expressed in frequencies of flowing cells passing close to the substrate. *, P = 0.0048, for 9C4-treated versus untreated cells. (Right) Tethering frequency on high-density ICAM-1 (coated at a tenfold higher density than in the main figure) after 9C4 pretreatment. All data are expressed as the mean ± range of two fields in view and depict one experiment representative of four. (B) Cells were pretreated with 9C4 and perfused for 1 min at 0.5 dyn/cm2 over ICAM-1 coimmobilized with heat-inactivated (–) or functional (+) CXCL9. Data are representative for two experiments. ICAM-1 coating was adjusted to obtain similar CXCL9-dependent LFA-1 activation as in A. (C) Pretreated cells were perfused over ICAM-1 coimmobilized with heat-inactivated (–) or functional (+) CCL21. ICAM-1 density was identical to that tested in A. Data are expressed as the mean ± range of two fields in view and show one experiment representative of four. (D, left) T cells were pretreated with anti-CXCR7 (9C4), anti-CXCR4 (12G5), or AMD3100. Cells were perfused for 1 min at 0.75 dyn/cm2 over VCAM-1 coimmobilized with heat-inactivated (–) or functional (+) CXCL12. Categories of interactions (tethers) are expressed as frequencies of cells in direct contact with the substrate. *, P = 0.011, for 9C4-treated versus control cells. (Right) Tethering frequency on high-density VCAM-1 (coated at a fivefold higher concentration than in the main figure). Data are expressed as the mean ± range of two fields in view and depict one experiment representative of six. (E) Cells were pretreated with 9C4 and perfused for 1 min at 0.75 dyn/cm2 over VCAM-1 coimmobilized with heat-inactivated (–) or functional (+) CXCL9. (F) Pretreated cells were perfused over VCAM-1 coimmobilized with heat-inactivated (–) or functional (+) CCL21. Data are expressed as the mean ± range of two fields in view and represent one experiment representative of three. (D–F) VCAM-1-coating densities were maintained constant, independent of the coimmobilized chemokine. (G) Effect of anti-CXCR7 blocking on T cell interactions with TNF--stimulated HUVEC. Cells pretreated with the anti-CXCR7 mAb 9C4 were perfused over the HUVEC monolayer at 0.75 dyn/cm2. Where indicated, HUVEC were preoverlaid with CXCL12. Categories of interactions were determined in frequencies of interacting cells within the cell flux in direct contact with the monolayer. Data are expressed as the mean ± range of two fields in view and represent one experiment representative of three.
|
 View larger version (23K): [in a new window] |
Figure 4. CXCR7 blockage by a low molecular weight antagonist specifically inhibits CXCL12-dependent LFA-1 and VLA-4 activation during subsecond contacts and arrest of T cells on inflamed endothelium. (A, left) T lymphocytes were pretreated with CCX733, CCX266 (control compound), or AMD3100. Cells were perfused for 1 min at 0.5 dyn/cm2 over ICAM-1/CXCL12. Arrests are expressed as frequencies of cells in direct contact with the substrate. (Right) Tethering frequency on ICAM-1/CXCL9 after 9C4 pretreatment. Data are expressed as the mean ± range of two fields in view and depict one experiment representative of three. (B, left) Cells, pretreated with CCX733, CCX266, or AMD3100, were perfused over VCAM-1/CXCL12 at 0.75 dyn/cm2. (Right) Tethering frequency on VCAM-1/CXCL9 after CCX733 pretreatment. Data are expressed as the mean ± range of two fields in view and depict one experiment representative of three. (C) Untreated cells or cells pretreated with CCX733 or AMD3100 were perfused over TNF--stimulated HUVEC at 0.75 dyn/cm2. HUVEC were overlaid with CXCL12 where indicated. Categories of interactions were determined as described before.
|
 View larger version (19K): [in a new window] |
Figure 5. CXCR7 blockage attenuates CXCR4 adhesiveness stimulated by surface-bound CXCL12. (A, left) A scheme describing the determination of CXCL12-stimulated CXCR4 adhesiveness to a surface-immobilized anti-CXCR4 mAb, which was coimmobilized with CXCL12, and the frequency and strength of interacting T lymphocytes were determined at 0.5 dyn/cm2. (Right) Frequency of CXCR4-mediated T cell interactions (transient or firm) with surface-immobilized anti-CXCR4 mAb under shear flow. T cells were pretreated with anti-CXCR4 (12G5) or isotype control mAb and perfused over a surface coated with the anti-CXCR4 mAb 12G5 (0.2 µg/ml) and CXCL12 or an irrelevant chemokine (CCL21). (B) Anti-CXCR7 treatment impairs CXCL12-stimulated CXCR4 adhesiveness on T cells. Tethering frequency of T cells to anti-CXCR4 mAb coimmobilized with CXCL12 or CCL21. T cells were left intact or pretreated for 2 h with different anti-CXCR7 mAb or with the small molecule antagonist CCX733. T cells were also preincubated with 100 µM AG490 or with a carrier solution (–) for 2 h. The right-most bar depicts interactions of T cells pretreated for 60 min with the actin cytoskeleton disrupting drug latrunculin (0.5 µM). (C) Effects of JAK/STAT inhibition on CXCL12-mediated integrin activation. T cells were preincubated with AG490 or with a carrier solution (–) for 2 h. Cells were perfused for 1 min at 0.75 dyn/cm2 over ICAM-1 (left) or VCAM-1 (right) coimmobilized with heat-inactivated (–) or functional (+) CXCL12. All values are the mean ± range of two fields in view. (A and B) Experiments are each representative of two independent runs.
|
 View larger version (25K): [in a new window] |
Figure 6. CXCR7-mediated modulation of integrin activation by CXCL12 requires prolonged exposure of intracellular CXCR7 to CXCR7 antagonists. (A) T cells were pretreated with the CXCR7-blocking mAb 11G8 or the CXCR7 antagonist CCX733 for 30 min or 120 min and perfused over ICAM/CXCL12, with or without a prior washing. The figure shows percent inhibition relative to inhibition by CXCR4 blockage (AMD3100 pretreatment). (B) PE-labeled antibody gains access to a small fraction of the intracellular CXCR7 pool following prolonged incubation. Cells were pretreated for 30 or 120 min with PE-conjugated anti-CXCR7 clone 11G8 in the presence or absence of competitive CCX733 or CCX266 (control compound) to confirm the specificity of anti-CXCR7 binding to the intracellular CXCR7, washed, fixed, and subjected to flow cytometry. The PE signal was determined as described in Materials and Methods. Data shown represent the mean ± range of two independent experiments.
|
We next addressed the ability of the small molecule CXCR7 antagonist to interfere with T lymphocyte arrest on inflamed endothelium under shear flow in the absence or the presence of CXCL12. T lymphocytes were perfused over TNF--activated HUVECs. As shown in Figure 4C
, neither the CXCR7 nor the CXCR4 antagonists interfered with selectin-mediated lymphocyte capture and rolling on TNF--activated HUVEC. However, when CXCL12 was absorbed apically on the endothelial monolayer, the majority of T cells tethered to the endothelial surface immediately arrested or came to final arrest after several seconds of rolling (Fig. 4C
and Supplementary Movie 1). Consistent with our findings about the isolated VLA-4 and LFA-1 ligands (Fig. 4 A and B)
, essentially all CXCL12-triggered, integrin-mediated T cell arrests on TNF--activated HUVEC were also eliminated upon T cell CXCR7 blockage by CCX733 or by CXCR4 blockage with AMD3100 (Fig. 4C
and Supplementary Movie 2). As CXCR7 was blocked solely on the T cell partner, we could exclude any contribution from potential inhibition of endogenously expressed CXCR7 of the HUVEC surface. Collectively, our results suggest that the lymphocyte CXCR7 plays an important role in CXCL12–CXCR4-triggered activation of VLA-4 and LFA-1 adhesiveness to inflamed endothelium under shear flow.
CXCR7 blockage interferes with CXCL12-stimulated CXCR4 adhesiveness developed at subsecond contacts
We have previously shown that immobilized CXCL12, when occupying CXCR4 within subsecond time-frames under shear flow, can potently increase the adhesiveness of the GPCR to immobilized mAb in an actin-dependent manner [11
, 34
] (Fig. 5A
). Furthermore, the ability of immobilized CXCL12 to rapidly trigger T cell CXCR4 adhesiveness in this nonphysiological assay closely correlated with the ability of the chemokine to activate T cell integrins under shear flow [11
, 34
]. We therefore determined next whether CXCR7 inhibition in T lymphocytes also interferes with the ability of CXCL12 to increase CXCR4 adhesiveness. As expected, CXCR4 adhesiveness to the 12G5 mAb was augmented rapidly upon T cell interaction with immobilized CXCL12 but not with an irrrelevant chemokine (CCL21; Fig. 5A
). Notably, blocking of CXCR7 by two mAbs, 9C4 and 11G8, as well as by the small compound antagonist dramatically impaired this CXCL12-triggered CXCR4 adhesiveness (Fig. 5B)
. Furthermore, disruption of the actin cytoskeleton completely abrogated this step without affecting CXCR4 expression levels (Fig. 5B
, and data not shown). These data therefore suggest a role of CXCR7 in the earliest CXCL12-induced CXCR4 adhesiveness critical for rapid integrin activation by surface CXCL12 under shear flow [11
, 34
]. This CXCL12-stimulated CXCR4 adhesiveness could be the result of CXCR4 dimerization. As CXCR4 dimerization by CXCL12 involves JAK/STAT signaling pathways [35
], we next asked whether CXCL12-triggered CXCR4 adhesiveness requires intact JAK-STAT. Surprisingly, the JAK2 inhibitor AG490 did not affect CXCL12-triggered CXCR4 adhesiveness (Fig. 5B)
. Accordingly, JAK2 inhibition did not affect CXCL12-stimulated LFA-1 or VLA-4 activation by immobilized CXCL12 (Fig. 5C)
and thus, could not mimic the effects of CXCR7 blockage on CXCR4-mediated integrin activation. Collectively, our data suggest a modulatory role of CXCR7 in CXCL12-stimulated CXCR4 adhesiveness, an actin-dependent rearrangement event that correlates strongly with the ability of CXCR4 to mediate upon occupany by suface-bound CXCL12, a Gi-dependent, JAK/STAT-independent activation of LFA-1 and VLA-4 integrins under shear flow.
The CXCR4 regulatory CXCR7 subset is a small fraction of the intracellular CXCR7 pool and is inhibited irreversibly only after prolonged incubation with CXCR7 blockers
Our staining data suggested that a major portion of lymphocyte CXCR7 is intracellular. To find what fraction of this pool is made available to the function-blocking mAb, we followed the kinetics of inhibition of CXCL12-mediated integrin activation after T cell pre-exposure to the function-blocking CXCR7 reagents for 10- to 120-min periods. A 10-min preincubation with anti-CXCR7 mAb or the CXCR7 antagonist failed to inhibit integrin activation by CXCL12 (data not shown), suggesting that it is the intracellular pool of CXCR7 that is targeted by the CXCR7-blocking reagents shown in previous parts to attenuate CXCL12-triggered integrin activation. Pretreatment of the T cells for 30 min was, however, sufficient to achieve maximal inhibition of CXCL12-mediated integrin activation, similar to inhibition obtained after 120 min (Fig. 6A
). Interestingly, whereas at 30 min, this high-degree inhibition was fully reversed by washing out the various blockers, after 120 min incubation, inhibition became essentially irreversible (Fig. 6A)
. This irreversible inhibition was not, however, the result of CXCR4 or CXCR7 degradation, as indicated by lymphocyte staining (data not shown). Notably, whereas intracellular CXCR7 was reversibly bound by the anti-CXCR7 mAb during the first 30-min period of incubation (Fig. 6B)
, during 120 min incubaion, the anti-CXCR7 mAb got access and irreversibly bound to 10–15% of the intracellular CXCR7 pool (Fig. 6B)
. This access was specifically dependent on anti-CXCR7 binding to CXCR7, as evident by its complete abrogation in the copresence of the CXCR7 antagonist CCX733 (Fig. 6B)
. This competition on the access of the the anti-CXCR7 mAb to its intracellular target molecules confirmed the specificity of our observations and argued against nonspecific internalization of the anti-CXCR7 by pinocytosis or FcR-mediated internalization. Our findings suggest collectively that a small fraction of the intracellular CXCR7, once occupied with CXCR7 mAb or antagonists, can no longer facilitate CXCR4-mediated LFA-1 and VLA-4 activation by CXCL12. Prolonged cell exposure to the various CXCR7 blockers (i.e., 2 h) exposes larger fractions of the intracellular CXCR7 to irreversible inhibition and thereby, irreversibly suppresses CXCL12-triggered signaling to integrins.
Intact CXCR7 is required for CXCL12-triggered VLA-4 activation in CD34+ cells but not for chemotaxis toward CXCL12
Hematopoietic progenitors expressing the CD34 marker use CXCL12 for a variety of migratory processes and can rapidly arrest on VCAM-1 in response to VLA-4-activating, CXCL12-triggered signals [9
, 10
, 36
]. Similar to T cells, the vast majority of CXCR7 on CD34+-enriched cells was intracellular (Fig. 7A
). To assess the role of this CXCR7, CB CD34+ cells were allowed to interact with a VCAM-1-coated surface coimmobilized with CXCL12. As observed with T lymphocytes, CXCR7 blockage by the function-blocking anti-CXCR7 mAb 9C4 or the small compound antagonist CCX733 markedly inhibited CXCL12-induced, VLA-4-mediated arrests under shear flow (Fig. 7B)
. CXCR7 blockage did not affect, however, the chemotaxis of these CD34+ cells toward CXCL12 (Fig. 7C)
. Thus, intact CXCR7 on CD34+ cells is required for rapid VLA-4 activation by CXCL12 under shear stress conditions but not for the motility of these cells toward CXCL12.
 View larger version (18K): [in a new window] |
Figure 7. CXCR7 blockage inhibits CXCL12-induced arrests of human CB CD34+ cells on VCAM-1. (A) Surface expression of CXCR7 was determined by flow cytometry with anti-CXCR7 antibody (9C4) and secondary antibodies. For detection of intracellular CXCR7, cells were fixed with paraformaldehyde/sucrose, permeabilized with saponin, and subsequently stained with anti-CXCR7 mAb and secondary antibodies. Histograms show one experiment representative of two donors. (B) CD34+ cells were pretreated with anti-CXCR7 (9C4, left) or the small molecule CXCR7 inhibitor CCX733 (right) and perfused for 1 min at 0.75 dyn/cm2 over VCAM-1 coimmobilized with CXCL12. (C) Chemotaxis of CD34+ cells pretreated with anti-CXCR7 (9C4, left) or with the CXCR7 inhibitor CCX733 (right) toward CXCL12. (B and C) Results show the mean and range of duplicate samples.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Characterizing the CXCR7 expression on T lymphocytes, we detected only minute levels of CXCR7 on the surface of human T lymphocytes, in contrast to a previous report by Balabanian and coworkers [16 ]. Instead, we found considerable levels of intracellular CXCR7 nearby the plasma membrane in these cells, a fraction of which is contained within early endosomes, based on colocalization analysis with the early endosomal marker EEA1. We also determined that only a minute fraction of the total CXCR7 can be present on the surface of resting T cells at steady-state conditions, where it becomes available to staining by an extracellular anti-CXCR7 and to binding by small molecule CXCR7 antagonists. Notably, when we used the fixation technique used in the Balabanian study instead of conventional flow cytometry of intact T lymphocytes, we did observe higher CXCR7 surface staining as reported by Balabanian et al. [16 ]. This staining may be attributed to partial permeabilization of the plasma membrane by the fixative, which rendered the intracellular CXCR7 pool accessible to antibody staining.
The main conclusion arising from our multiple functional assays is that T lymphocytes as well as CD34+ cells do not use their CXCR7 to transduce CXCL12-dependent Gi signaling, motility, or chemotaxis. Indeed, all of these CXCL12-triggered processes were blocked entirely by CXCR4 inhibition and were fully retained upon prolonged CXCR7 blockage. Our results are consistent with previous data in other cellular systems, indicating that CXCL12 binding to CXCR7 does not mediate calcium mobilization or chemotaxis [21 , 22 ]. As a result of polar alterations in its Asp-Arg-Tyr motif, CXCR7 might not be coupled in a classical way to G proteins but rather, may function as an atypical chemokine receptor similar to Duffy antigen receptor for chemokines [37 ], D6 [38 , 39 ], or CCXCKR [40 ]. However, none of these silent receptors has been shown to modulate physiological activity triggered by an extracellular chemokine, whereas CXCR7 appears to modulate a physiological stimulatory activity of CXCL12 towards multiple integrins. Our results are therefore a first example of a silent receptor to a specific chemokine (i.e., CXCL12), which modulates a specialized signaling function of a second GPCR for the same chemokine. A portion of the CXCR7 may give rise to CXCR7/CXCR4 complexes with a specialized role in integrin activation rather than in CXCR4-mediated motility. Indeed, CXCR7 and CXCR4 form functional heterodimers in human embryo kidney 293 cells with higher responsiveness to CXCL12 than CXCR4 alone [21 ], although the existence of such heterodimers in primary lymphocytes cannot be demonstrated experimentally by current methodologies.
The ability of CXCR4 to activate integrins depends on its adhesiveness in the context of immobilized CXCL12 under shear stress conditions, a property found by us to be independent and therefore upstream to Gi protein signaling [34 ]. CXCR7 blockage impaired CXCL12-triggered CXCR4 adhesiveness, identifying CXCR7 as a direct modulator of CXCL12-induced CXCR4 rearrangement events preceding CXCR4-mediated Gi activation. These rearrangements depend on intact actin cytoskeleton but not on the JAK/STAT pathway, previously implicated in an essential CXCR4 dimerization step prior to Gi signaling [35 ]. Correspondingly, JAK/STAT signaling was not required for integrin activation by CXCL12 under shear flow, ruling out a role for a JAK-STAT-dependent CXCR4 dimerization in the CXCL12-stimulated CXCR4 rearrangements that precede CXCR4-mediated Gi activation and determine integrin stimulation by CXCL12 under shear flow. CXCL12, when coimmobilized with L-selectin ligands, can markedly destabilize L-selectin-mediated rolling adhesions also via a Gi-independent, CXCR4-dependent mechanism [41 ]. However, this CXCL12-dependent activity was insensitive to CXCR7 inhibition (manuscript in preparation). These data collectively suggest that the CXCR7 crosstalk with CXCR4 identified in this study is necessary for CXCR4 to maintain critical adhesiveness to CXCL12 without which rapid downstream signaling to VLA-4 and LFA-1 cannot proceed. Integrin activation was recently shown by us to require simultaneous application of external shear forces on the integrin and on its activating GPCR [42 ]. The dependence of CXCR4 adhesiveness on an intact actin cytoskeleton may arise from basal or CXCL12-enhanced CXCR4 anchorage to the cytoskeleton essential for the GPCR to tolerate these shear forces during rapid integrin activation. Our findings suggest that these CXCL12-stimulated CXCR4 rearrangements, which take place under shear stress conditions rather than global CXCL12-stimulated signaling, triggered in shear-free environments, are the main target for inhibition by the antagonist-occupied CXCR7. Importantly, none of the CXCR7-blocking reagents enriched CXCR7 expression on the plasma membrane, and so, their capacity to inhibit CXCL12-triggered CXCR4 signaling to integrins was not the result of enhanced CXCR7 translocation to the plasma membrane.
One of the intriguing findings of our study is that the potent inhibition of CXCR4-mediated integrin activation was observed upon the occupancy of a small fraction of the submembranal pool of CXCR7. The kinetics of inhibition and the staining data support the notion that the intracellular CXCR7 pool is made available to the externally introduced reagents via a small, constantly recycling subset that is present at steady-state at low levels on the outer leaflet of the plasma membrane. Nevertheless, we could not assess the involvement of this putative CXCR7 recycling in the novel CXCR7 crosstalk with CXCR4, as blockers of endosomal acidification processes involved in recycling, such as chloroquine, altered CXCR4 expression and therefore, could not be used in our experimental setting. Nevertheless, an externally applied anti-CXCR7 mAb incubated for 120 min with T cells, was found to irreversibly block CXCR4-mediated integrin activation. This irreversible inhibition was achieved after mAb binding to less than 15% of the total CXCR7 pool. Taken together, our findings suggest that although only a minor fraction of the submembranal CXCR7 pool, apparently associated with early endosomes, persistently recycles to the plasma membrane once occupied by a blocker, this pool is restricted from facilitating CXCL12-mediated CXCR4 rearrangments critical for rapid signaling to lymphocyte integrins under shear flow but not for all other CXCR4 signaling functions, which are elicited by soluble or immobilized CXCL12 under shear-free conditions.
In conclusion, we demonstrate a novel crosstalk between CXCR7 and CXCR4, which is essential for rapid CXCL12-triggered integrin activation involved in lymphocyte and CD34+ cell arrest on endothelial surfaces expressing integrin ligands but is dispensable for all other testable, CXCL12-stimulated motility and signaling processes. This is a first demonstration that inhibition of an atypical, intracellular chemokine receptor can interfere effectively with a specialized, physiological function of another signaling chemokine receptor abundantly expressed on the cell surface. Our study also suggests that GPCRs largely excluded from the cell surface can still serve as therapeutic targets for small and large extracellular antagonists. Circulating leukocytes may express other intracellular pools of atypical chemokine receptors such as, e.g., D6 [43 ]. Whether the crosstalk between intracellular CXCR7 and surface CXCR4 is a unique case or an example of a broader set of similar inter-GPCR communications [44 45 46 ] involved in integrin activation processes remains to be determined in future studies. At any rate, CXCR7 blockers can be potentially used as a new class of inhibitors to attenuate integrin-stimulatory properties of CXCR4 without affecting other essential, CXCR4-mediated signaling and motility processes.
Received February 6, 2008; revised June 17, 2008; accepted June 25, 2008.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4 integrin clustering by immobilized chemokines stimulates leukocyte tethering and rolling on endothelial vascular cell adhesion molecule 1 under flow conditions J. Exp. Med. 192,495-506 4 integrin avidity but not of leukocyte function-associated antigen-1 avidity to endothelial ligands under shear flow requires cholesterol membrane rafts J. Biol. Chem. 277,40027-40035 triggers CXCR4 receptor dimerization and activates the JAK/STAT pathway FASEB J. 13,1699-1710This article has been cited by other articles:
 |
|
 |
|
  M. Tarnowski, K. Grymula, R. Reca, K. Jankowski, R. Maksym, J. Tarnowska, G. Przybylski, F. G. Barr, M. Kucia, and M. Z. Ratajczak Regulation of Expression of Stromal-Derived Factor-1 Receptors: CXCR4 and CXCR7 in Human Rhabdomyosarcomas Mol. Cancer Res., January 1, 2010; 8(1): 1 - 14. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Campana, L. Bosurgi, M. E. Bianchi, A. A. Manfredi, and P. Rovere-Querini Requirement of HMGB1 for stromal cell-derived factor-1/CXCL12-dependent migration of macrophages and dendritic cells J. Leukoc. Biol., September 1, 2009; 86(3): 609 - 615. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. A. Zabel, Y. Wang, S. Lewen, R. D. Berahovich, M. E. T. Penfold, P. Zhang, J. Powers, B. C. Summers, Z. Miao, B. Zhao, et al. Elucidation of CXCR7-Mediated Signaling Events and Inhibition of CXCR4-Mediated Tumor Cell Transendothelial Migration by CXCR7 Ligands J. Immunol., September 1, 2009; 183(5): 3204 - 3211. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Levoye, K. Balabanian, F. Baleux, F. Bachelerie, and B. Lagane CXCR7 heterodimerizes with CXCR4 and regulates CXCL12-mediated G protein signaling Blood, June 11, 2009; 113(24): 6085 - 6093. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. E. Luker, M. Gupta, and G. D. Luker Imaging chemokine receptor dimerization with firefly luciferase complementation FASEB J, March 1, 2009; 23(3): 823 - 834. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
