Importance of CXC Chemokine Receptor 2 in the Homing of Human Peripheral Blood Endothelial Progenitor Cells to Sites of Arterial Injury
Circulating endothelial progenitor cells (EPCs) may contribute to endothelial regeneration; however, the exact mechanisms of their arterial homing remain elusive. We examined the role of the angiogenic chemokine receptor CXCR2 in the homing of human EPCs. Isolated EPCs expressed CXCR2 together with kinase insert domain–containing receptor, CD31, vascular endothelial cadherin, and CXCR4. Adhesion assays under flow conditions showed that EPCs preferentially adhered to β2-integrin ligands, that firm arrest on fibronectin or fibrinogen was enhanced by the CXCR2 ligands CXCL1 or CXCL7, and that blockade of CXCR2 significantly reduced EPC adhesion on platelet-coated endothelial matrix. This was corroborated by the involvement of CXCR2 in EPC recruitment to denuded areas of murine carotid arteries ex vivo and in vivo. Notably, blocking CXCR2 inhibited the incorporation of human EPCs expressing CXCR2 at sites of arterial injury in athymic nude mice. Immunoreactivity for the β-thromboglobulin isoform CXCL7 was observed in murine platelets and denuded smooth muscle cells (SMCs) early after wire injury, and transcripts for CXCL7 and CXCL1 were detected in isolated human arterial SMCs. Human KDR+CXCR2+ cells showed better in situ adhesion to injured murine carotid arteries than KDR+CXCR2− cells, were predominantly CD14+, and improved CXCR2-dependent endothelial recovery after injury in nude mice. In conclusion, our data clearly demonstrate the importance of CXCR2 for the homing of circulating EPCs to sites of arterial injury and for endothelial recovery in vivo.
Adult peripheral blood contains a subpopulation of bone marrow (BM)-derived vascular precursor cells, which can differentiate into mature endothelial cells and have, therefore, been termed endothelial progenitor cells (EPCs).1,2 These circulating EPCs are characterized by expression of CD133, CD34, and vascular endothelial growth factor receptor-2 (VEGFR-2/kinase insert domain–containing receptor [KDR]).3,4 After an in vitro culture for 4 to 7 days, EPCs still express hematopoietic progenitor marker and are characterized by development of colony forming units and the positivity for marker of the endothelial lineage.1–5 Functional studies in animals have implied a contribution of ex vivo–expanded EPCs to therapeutic reendothelialization or neovascularization.4–7 A variety of growth factors, cytokines, and chemokines, but also apoptotic bodies from endothelial cells, may influence mobilization, homing, proliferation, and differentiation of EPCs.4,5,8 However, the exact signaling involved in the homing of EPCs to sites of endothelial injury remains incompletely understood. As a chemokine essential for the mobilization of EPCs and other progenitor cells (PCs) from the BM, stromal cell–derived factor-1α (SDF-1α/CXCL12) interacts with CXC chemokine receptor 4 (CXCR4) to mediate EPC recruitment along hypoxic gradients and can thereby contribute to neovascularization of ischemic tissue.9,10 As recent studies revealed that the CXCL12/CXCR4 axis is also crucial for tumor neoangiogenesis and for recruiting of smooth muscle PCs in vivo, its role in peripheral PC biology appears to be more universal.11,12 Notably, experiments using BM chimeras implied that CXCR4 and CXCL12 may be more important for the recruitment of BM-derived progenitors of smooth muscle cells (SMCs) to the site of arterial injury than for the recruitment of PCs, giving rise to endothelial cells.12 Evidence from rodent models revealed that the proangiogenic CXCR2 ligand keratinocyte-derived chemokine (KC/CXCL1) is not only expressed at sites of endothelial dysfunction, contributing to inflammatory and atherogenic monocyte recruitment, but is also involved in endothelial recovery after arterial injury.13–15 The human KC ortholog growth-regulated oncogene-α (GRO-α/CXCL1) is a potent angiogenic factor binding to CXCR2 and promoting tumor growth.16 However, the functional effects of CXCR2 and its ligands on human EPCs remain to be clarified. Hence, we explored and compared the contributions of CXCR2 and CXCR4 to the homing of isolated as well as circulating human EPCs in a context of vascular repair after endothelial injury. Here, we show that CXCR2 is expressed on EPCs and functions as an arrest receptor supporting their homing to sites of arterial injury and thereby contributing to endothelial recovery.
Materials and Methods
Cell Isolation and Culture
Mononuclear cells (MNCs) were separated by Biocoll density gradient centrifugation (Biochrom) from buffy coats of healthy human volunteers and plated (107/well) on fibronectin-coated 6-well plates in MV2 endothelial growth medium (PromoCell). Medium was changed at day 4, and cells were used for experiments at day 7. Human umbilical vein endothelial cells (HUVECs) cultured in endothelial cell growth medium (PromoCell) were used between passages 3 to 5. Human SMCs isolated from the arteria mammaria interna were provided by Dr R. Blindt. Adherent cells were detached using Accutase (PAA Laboratories). Human CD34+ cells were isolated from MNCs by magnetic separation (CD34 kit and autoMACS sorter, Miltenyi Biotec) with a purity of >90%.
RT-PCR Analysis of CXCR2 and CXCR4
Total RNA from EPCs, HUVECs, or SMCs was isolated using RNeasy Kit (Qiagen), and 1 μg of RNA was reverse-transcribed into cDNA using Mo-MLV RT (Invitrogen). RT-PCR was performed with 20 ng of cDNA using Taq DNA polymerase (Promega) for 30 cycles (30 seconds, 95°C; 30 seconds, 60°C; 1 minute, 72°C). Primers: CXCR4 (forward), 5′-CCTGCTGACTATTCCCG ACTTCATC-3′ and (reverse) 5′-CCAAGGAAAGCATAGAGGATGGG-3′; CXCR2 (forward), 5′-GGGCAACAATACAGCAAACT-3′ and (reverse) 5′-GCACTTAGGCAGGAGGTCTT-3′.
Flow Cytometry and Sorting
Cells were reacted with CXCR2 and CXCR4 monoclonal antibodies (mAbs) (MAB331 and MAB171; R&D Systems) and vascular endothelial (VE)-cadherin polyclonal antibody (Ab) (C-19; Santa Cruz Biotechnology), or respective isotype controls, and fluorescein isothiocyanate (FITC)–conjugated secondary Abs (Sigma-Aldrich) for 30 minutes on ice. EPCs were incubated with 10 μg/mL 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl-indocarbocyanine (DiI)-labeled acetylated LDL (CellSystems) and counter-stained with 10 μg/mL lectin–FITC (Sigma) after detachment and fixation (2% paraformaldehyde). EPCs, MNCs, or CD34+ cells pretreated with FcR blocking reagent (Miltenyi Biotec) were reacted with mouse anti-human VEGFR2–phycoerythrin (VEGFR2-PE) (FAB357P), CD14-PE (FAB3832P; R&D), CD31-FITC (MHCD3101; Caltag), CXCR2-FITC (6C6; BD Biosciences), CD34-PE (AC136, Miltenyi Biotec) mAbs, and respective isotype controls at recommended concentrations and fixed and analyzed in a FACSCalibur flow cytometer (Becton Dickinson). At least 10 000 gated cells were acquired and analyzed using CellQuestPro software.
For sorting, MNCs reacted with VEGFR2-PE/CXCR2-FITC mAbs were resuspended (107/mL) in HEPES-buffered Hank’s Balanced Salt Solution+0.5% BSA and 2 mmol/L EDTA. High-pressure sorting with purity precision mode was performed in a FACSAria sorter (Becton Dickinson) to separate and collect KDR+CXCR2+ or KDR+CXCR2− MNCs. The fraction of KDR+CXCR2+ MNCs was highest in the CD14+ monocyte light scatter gate R1 also used for collection.
In Vitro Adhesion Assays in Flow and Transmigration Assay
Laminar flow assays were as described.17 Dishes were coated with fibronectin (10 μg/mL), collagen I (80 μg/mL), human fibrinogen (10 μg/mL, Sigma-Aldrich, all in PBS), and soluble intercellular adhesion molecule-1 (20 μg/mL, in 10 mmol/L Tris, pH 9).18 Extracellular matrix (ECM) was prepared by carefully detaching confluent HUVECs. CXCL1 (R&D), CXCL7, or CXCL12 (PeproTech, all 100 ng/mL in PBS) was immobilized on fibronectin, fibrinogen, or collagen for 1 hour. EPC adhesion to BSA-coated plastic was negligible and not altered by coimmobilized chemokines (data not shown). Some EPCs were directly stimulated with CXCL1 or CXCL12 for 10 minutes or preincubated with neutralizing mAbs against CXCR2, CXCR4, or IgG2a isotype control (10 μg/mL each) for 30 minutes. Human platelets activated with 4 μmol/L thrombin receptor activating peptide (TRAP-6; Sigma-Aldrich) were immobilized on ECM for 30 minutes at 37°C. Dishes were assembled as the lower wall of a flow chamber on the stage of an Olympus IX50 microscope. EPCs or HUVECs were resuspended in HHMC medium (HEPES-buffered Hank’s Balanced Salt Solution, 1 mmol/L Mg2+/Ca2+, 0.5% BSA) and perfused (1.5 dyne/cm2) at 37°C. After 4 minutes, firmly adherent cells were quantified in multiple fields recorded with a JVC 3CCD video camera.17
Transmigration assays were performed using FluoroBlok inserts (8-μm pore size; Becton Dickinson). EPCs (5.0×104) stained with calcein/acetoxymethyl ester (Molecular Probes) were placed in fibronectin-coated inserts in medium 199 (Gibco) with 0.5% BSA in duplicate. Inserts were placed into wells of a 24-well plate containing 100 ng/mL CXCL1, CXCL7, or CXCL12 in medium 199+0.5% BSA. After 3 hours at 37°C, the fluorescence of cells migrated to the bottom surface of the filter was measured using an ELISA reader.
Ex Vivo Perfusion of Murine Carotid Arteries and EPC Recruitment After Injury In Vivo
After 1 week of atherogenic diet, carotid arteries of apolipoprotein E (ApoE)−/− mice (M&B, Ry Denmark) were subjected to wire-induced endothelial denudation and isolated for ex vivo perfusion after 24 hours.12,14,19 calcein/acetoxymethyl ester–labeled EPCs (5×105) pretreated with mAbs to CXCR2, CXCR4, or IgG2a isotype control were perfused through carotid arteries (5 μL/min), and adhesive interactions with the injured vessel wall were recorded using an Olympus BX51 microscope and stroboscopic epifluorescence illumination (Drelloscop 250). For in vivo assays, EPCs (106) and sorted KDR+CXCR2+ or KDR+CXCR2− MNCs (2.5×105) labeled with rhodamine-6G chloride (Molecular Probes) were injected intracardially, and their in situ recruitment to injured carotid arteries was analyzed in multiple fields using an Axiotech Vario 100 microscope (×10/×20 water immersion objective; Carl Zeiss).
In addition, EPCs labeled with CM-DiI (Molecular Probes), or CD14+KDR+CXCR2+ cells, were pretreated with blocking CXCR2 mAb or isotype and injected intracardially into athymic NMRI nude mice (Harlan Winkelmann, Borchen, Germany) 24 hours after wire injury. After 7 days, carotid arteries were harvested by in situ perfusion fixation for immunohistochemical analysis of homing and endothelial recovery. Animal studies were approved by local authorities and complied with German animal protection law.
Immunohistochemistry and RT-PCR for CXCLs
Serial sections from carotid arteries were stained with biotinylated CXCL7/NAP-2 polyclonal Ab (BAF393), CXCR2 mAb (MAB2164; R&D), or VE-cadherin polyclonal Ab (Santa Cruz Biotechnology); P-selectin mAb (Pharmingen), α-SM actin (SMA) mAb (clone 1A4; DAKO), and FITC-conjugated Ab or Vectastain ABC-AP (Vector Labs). Nuclei were stained by 4′,6-diamidino-2-phenylindole (DAPI). Human SMCs (passage 6 to 8) were treated with interferon-γ/tumor necrosis factor-α (100 U/mL; PeproTech) for 24 hours. RNA isolation and RT-PCR were as above. Primers: CXCL1 (forward), 5′-AGGGAATTCACCCCAAGAAC-3′ and (reverse) 5′-TAA CTATGGGGGATGCAGGA-3′; CXCL7 (forward), 5′-TGGAAACAACTCTAGCTCAGC CTTCTC-3′ and (reverse) 5′-TCCAGGCAGATTTTCCTCCCATCC-3′; CXCL8 (forward), 5′-TCTGCAGCTCTGTGTGAAGG-3′ and (reverse) 5′-AATTTCTGTGTTGGCGCAGT-3′. PCR products were confirmed by agarose gel electrophoresis and sequencing.
All data represent mean±SEM of >3 independent experiments. Data were analyzed by 2-tailed t test or 1-way ANOVA with Tukey’s test for multiple comparisons using GraphPad Prism 4.0 (San Diego, Calif). Differences with P<0.05 were considered statistically significant.
Characterization of Cultured Endothelial-Like Cells From Human Peripheral Blood
Peripheral blood–derived MNCs cultured under endothelial-specific conditions developed a spindle-shaped appearance and typical cell clusters at day 7 after the isolation (Figure 1A). As analyzed by flow cytometry (Figure 1A), these cells were positive for KDR (49.2±6.3%) and the endothelial-specific markers CD31 (92.6±3.2%) and VE-cadherin (94.0±0.6%). Adherent cells intensely took up acetylated LDL and bound endothelial-specific lectin (Figure 1A). In line with published data,1–9 these cells were therefore identified as early peripheral blood endothelial outgrowth or cultured EPCs. Moreover, RT-PCR analysis revealed the presence of mRNA transcripts for the chemokine receptors CXCR2 and CXCR4 in EPCs, whereas in HUVECs, CXCR4 but not CXCR2 transcripts were detected (Figure 1B). This expression profile was confirmed at the protein level by flow cytometric analysis of CXCR2 and CXCR4 surface expression (Figure 1C). EPCs were strongly positive for CXCR2 and CXCR4, whereas HUVECs showed a very weak staining for CXCR2 but were largely positive for CXCR4. As evidenced by specific mean fluorescence intensities, CXCR2 surface expression was substantially stronger on isolated EPCs than on HUVECs, whereas CXCR4 surface expression did not differ (Figure 1C). Thus, isolated EPCs at day 7 in culture were strongly positive for endothelial markers, as well as for CXCR2 and CXCR4.
Arrest of EPCs and HUVECs on ECM Components Under Laminar Flow
To compare the spontaneous adhesion profile of EPCs and HUVECs on various ECM components under flow conditions, we used a parallel wall flow chamber, as described.17 Cells were perfused at 37°C and the number of firmly adherent cells per millimeter squared was determined. Although firm adhesion of EPCs was significantly better supported on intercellular adhesion molecule-1 and on fibrinogen, HUVECs preferentially adhered to collagen I or to HUVEC matrix (Figure 2). Firm adhesion of EPCs and HUVECs on fibronectin did not significantly differ (Figure 2).
Arrest in Flow and Transmigration of EPCs Triggered by Chemokines
Using the flow chamber, we next analyzed the adhesion of EPCs induced by the CXCR2 ligands CXCL1 or CXCL7, or by the CXCR4 ligand CXCL12. The adhesion of EPCs in flow was significantly increased by CXCL1 or CXCL7 immobilized on fibronectin (Figure 3A), on fibrinogen (Figure 3B, representative images), or on collagen I (data not shown), whereas arrest was not enhanced by immobilized CXCL12 (Figure 3A and 3B). Direct stimulation of EPCs with either CXCL1 or CXCL12 significantly increased their arrest on fibronectin (data not shown).
To examine a contribution of CXCL1, CXCL7, or CXCL12 to directed transmigration of EPCs, a trans-well filter system was used. Notably, CXCL12 and to a lesser degree CXCL7 triggered EPC transmigration, whereas CXCL1 did not induce transmigration (Figure 3C).
CXCR2 Mediate Firm Adhesion of EPCs to Platelets In Vitro and Injured Arteries In Vivo
The presence of surface-adherent platelets is an early hallmark of arterial injury. Notably, platelets adherent on ECM also supported substantial arrest of EPCs under flow conditions in vitro. Whereas a neutralizing CXCR4 hardly reduced adhesion on platelet-coated ECM, a neutralizing mAb to CXCR2 significantly inhibited the adhesion of EPCs (Figure 4A).
To further assess whether early EPC recruitment after endothelial denudation involves CXCR2 and/or CXCR4, we performed ex vivo perfusion of carotid arteries isolated from hyperlipidemic ApoE−/− mice with EPCs 24 hours after wire injury using neutralizing mAbs against both receptors. When compared with isotype control, blocking CXCR4 significantly reduced firm arrest of EPCs (Figure 4B). Notably, however, blocking CXCR2 caused an even more pronounced decrease in EPC adhesion (P=0.001). Similar results were obtained for the adhesion of EPCs to injured carotid arteries after injection into ApoE−/− mice in vivo, where blocking either CXCR2 or CXCR4 reduced EPC recruitment (Figure 4C). Thus, the homing of EPCs to denuded areas of injured carotid arteries requires a CXCR2- and CXCR4-dependent activation step in vivo.
CXCR2 Mediates EPC Homing and Endothelial Repair In Vivo
To investigate the relevance of CXCR2 for homing of EPCs and endothelial regeneration, we performed experiments using infusion of DiI-labeled human EPCs pretreated with a blocking CXCR2 mAb or isotype control in athymic nude mice 24 hours after wire injury of the carotid artery. Indeed, fluorescence microscopy revealed an incorporation of human DiI-EPCs at sites of injury in carotid arteries of athymic nude mice. This incorporation was significantly reduced by pretreatment with a neutralizing CXCR2 mAb (Figure 5A). Staining for CXCR2 revealed a colocalization of incorporated DiI-EPCs (Figure 5B), which were found in direct contact to surface-adherent platelets or adjacent to neointimal SMCs (Figure 5C) early after wire injury.
Presence of CXCR2 Ligands in SMCs and Platelets
Whereas murine platelets do not express CXCL8, CXCL1 was not found to be expressed early after arterial injury.15,19 Platelets contain precursors for CXCL7, eg, β-thromboglobulin or connective tissue–activating peptide III (CTAP-III), which can be processed by cathepsin G.20 Because EPCs express cathepsins,21 we analyzed the expression of CXCL7 isoforms in carotid arteries of ApoE−/− mice 24 hours after wire injury. Staining of injured carotid arteries with an antibody against CXCR7 showed intense reactivity in freshly denuded medial SMCs, as well as in platelets adherent to the denuded vessel wall (Figure 6A). Expression of CXCL7 mRNA transcripts was also detectable in isolated human arterial SMCs but not in HUVECs (Figure 6B). This expression was constitutive and not further enhanced by costimulation with inflammatory cytokines. This was in contrast to other CXCR2 ligands, eg, CXCL1 or CXCL8, which were present in both SMCs and HUVECs (Figure 6B). Moreover, human arterial SMCs supported CXCR2-mediated adhesion of EPCs in flow (data not shown). Thus, platelets and vascular SMCs represent alternative source for functional CXCR2 ligands.
Functional Relevance of CXCR2 on Circulating KDR+ Cells for Arterial Homing In Vivo
To address whether CXCR2 is important for homing of circulating EPCs in vivo, we assessed the in situ adhesion of injected human KDR+CXCR2+ or KDR+CXCR2− cell subsets to injured carotid arteries of ApoE−/− mice. This revealed that the early adhesion of sorted KDR+CXCR2+ cells to the site of endothelial injury was significantly higher than that of KDR+CXCR2− cells (Figure 7A). To further characterize the expression of CXCR2 on EPC subpopulations in peripheral blood, we performed a flow cytometric subanalysis of human MNC subsets after double staining for KDR and CXCR2. This revealed that KDR+CXCR2+ cells were most prevalent (3.94±1.43%, n=4) within the CD14+ fraction but amounted to only 0.47±0.15% (n=9) within the CD14− fraction (Figure 7B). Because circulating EPCs have also been defined as CD34+KDR+ cells,1,3–5 we performed magnetic separation of human CD34+ cells with subsequent costaining for KDR and CXCR2. Notably, only 0.21±0.09% (n=3) of CD34+ cells were KDR+CXCR2+ (Figure 7C). Finally, we used the human CD14+CXCR2+KDR+ cell population pretreated with blocking CXCR2 mAb or isotype control for therapeutic infusion in athymic nude mice 24 hours after wire injury of carotid arteries to analyze reendothelialization by staining for luminal VE-cadherin after 7 days. Notably, endothelial recovery was significantly improved by infusion of isotype control-treated but not CXCR2 mAb–treated CD14+CXCR2+ KDR+ cells (Figure 7D). Given the importance of CXCR2 for the homing of circulating EPCs in vivo, CXCR2+KDR+ cells appear to represent a crucial subset of circulating EPCs with relevance for endogenous endothelial recovery.
Here we demonstrate that isolated human EPCs express high levels of the CXC chemokine receptor CXCR2, which was hardly expressed on more mature endothelial cells. Known as a functional arrest receptor,13 CXCR2 was important for triggering the recruitment of EPCs on ECM coimmobilized with chemokines or on ECM-adherent platelets in a laminar flow model in vitro. In contrast, the CXCL12/CXCR4 axis triggered adhesion on fibronectin, only on direct prestimulation of EPCs and appeared more crucial in mediating EPC transmigration. This was confirmed in vivo and by ex vivo perfusion of mouse carotid arteries after wire-induced denudation, where blocking CXCR2 more markedly diminished EPC recruitment to the site of injury than blocking CXCR4. Moreover, blocking CXCR2 significantly reduced the incorporation of infused human EPCs at sites of arterial injury in athymic nude mice. Early after endothelial denudation, we detected CXCL7 isoforms as additional CXCR2 ligands present in platelets and medial SMCs at the site of denudation. Finally, circulating KDR+CXCR2+ cells displayed more marked adhesion to injured carotid arteries than KDR+CXCR2− cells and for the CD14+ population a CXCR2-dependent capacity for endothelial regeneration, clearly underscoring the relevance of CXCR2 for homing of circulating EPCs and endothelial recovery in vivo.
In this study, we have systematically analyzed the adhesive properties of EPCs in vitro under flow conditions on adhesion molecules and ECM coimmobilized with chemokine or coated with platelets and in injured carotid arteries perfused ex vivo or in vivo. In line with findings that β integrins on EPCs are constitutively active,22 EPCs exhibited more substantial adhesion to β2-integrin ligands (intercellular adhesion molecule-1 or fibrinogen) than arguably more mature HUVECs, whereas preferential adhesion was not seen on fibronectin or collagen I. This may extend the functional impact of β2 integrins reported for the homing and neovascularization capacity of EPCs.22 The adhesion of EPCs was studied on platelet-coated endothelial ECM under shear flow as a suitable in vitro model for the injured vessel wall. Because recent studies23,24 revealed an importance of platelets for EPC differentiation in vitro, interactions of adult EPCs with platelets may also play an essential role for re-endothelialization after arterial injury. Indeed, experiments in GPIIb-deficient mice have implicated platelet aggregation in the homing of vascular PCs after arterial injury.25 Because inhibition of CXCR2 considerably reduced EPC adhesion on ECM-adherent platelets, and in injured carotid arteries ex vivo and in vivo, CXCR2 may represent an additional component crucial for EPC arrest after injury.
Murine CXCL1 is sequestrated on activated endothelium in response to inflammatory stimuli or during early native atherosclerosis,14,15 but also upregulated after wire-induced endothelial denudation in ApoE−/− mice,15,19 whereas blockade of KC in vivo delayed endothelial repair after wire injury.15 In light of our present data, this effect may be attributable to an attenuated adhesive capacity of EPCs for endothelial recovery. However, CXCL1 was not detectable in platelets or within the first 24 hours early after injury.15,19 Among other CXCR2 ligands, CXCL8 is not expressed in mice and CXCL5 does not serve as an arrest chemokine.26 Of note, the CXCR2 ligand CXCL7 was present both on luminal platelets and medial SMCs 24 hours after vascular injury, and in isolated human arterial SMCs. The expression in arterial SMCs was constitutive and in contrast to other CXCR2 ligands (CXCL1/8) not inducible by inflammatory cytokines, possibly bearing a homeostatic function for vascular integrity. Coimmobilized on fibronectin or fibrinogen, CXCL7 was able to induce EPC adhesion in flow.
Activated platelets at sites of endothelial denudation provide an additional source for β thromboglobulins (eg, CTAP-III) as precursors for CXCL7.20 Conversion of CTAP-III into CXCL7 occurs after proteolytic cleavage by cathepsin G and isolated EPCs have been shown to express higher amount of cathepsins than HUVECs.20,21,26 Hence, it is conceivable that EPCs may secrete cathepsin G to process CTAP-III to CXCL7 and to promote early CXCR2-dependent EPC arrest on denuded SMCs or adherent platelets. Notably, CXCR2 plays a crucial role in atherosclerotic lesion formation, as evidenced by the reduced susceptibility to atherosclerosis in mice reconstituted with CXCR2−/− BM.27 A recent study revealed that transfer of spleen-derived EPCs in ApoE−/− mice can destabilize atherosclerotic plaques,28 implying proatherogenic features of EPCs possibly recruited via CXCR2. Our data further unveil that mice repopulated with CXCR2−/− BM exhibit an impaired reendothelialization and consequently enhanced neointima formation after arterial wire injury (M.H., A.Z., E.A.L., C.W., unpublished data). This is in line with previous findings that blockade of the CXCR2 ligand CXCL1 delayed endothelial recovery and exacerbated lesion formation after injury.15 Taken together, CXCR2 appears crucial in controlling adhesion and influx of EPCs. Given the monocytic features displayed by adult EPCs and their elusive role in advanced atherosclerosis, an inhibition of CXCR2 signaling on EPCs or monocytes during advanced atherosclerosis may serve to stabilize atherosclerotic plaques, eg, caused by impaired EPC influx. Conversely, during early atherosclerosis or after endothelial denudation, CXCR2 may be beneficial for improving endothelial recovery or regeneration by promoting EPC homing.
Unlike CXCL1, CXCL12 was found to primarily induce transmigration of isolated human EPCs. Although EPC adhesion on fibronectin or ECM-adherent platelets did not involve CXCL12/CXCR4 in vitro, CXCR4 contributed to EPC recruitment in injured carotid arteries ex vivo and in vivo, possibly because of the presence of SMC-derived CXCL12.12 Conversely, CXCL7 but not CXCL1 induced EPC transmigration, confirming the notion that CXCL1 is specialized to serve as an arrest chemokines for monocytes.29 Thus, CXCL12 and to a lesser extent CXCL7 may be important for EPC mobilization, migration, and neovascularization. This supports data showing a role of CXCL12 in trafficking of BM PCs to ischemic tissue or in improving neovascularization.9,10,30 Besides EPCs, enhanced neovascularization is attributed to a recruitment of smooth muscle PCs, which requires the CXCL12/CXCR4 axis.12
Although EPCs do not encompass a homogeneous population and can be derived from various origins, they have been defined in vivo as cells expressing different levels of CD133, CD34, and KDR, depending on their location (BM or peripheral blood) and degree of maturation.1,3–5 To assess whether CXCR2 is essential for homing of circulating EPCs in vivo, we injected KDR+CXCR2+ or KDR+CXCR2− cells in ApoE−/− mice after wire injury. Notably, KDR+CXCR2+ cells were distinguished from KDR+CXCR2− cells by a high-adhesive capacity to sites of injury in vivo. Subanalysis of KDR+CXCR2+ cells revealed that the fraction of this subset was highest among CD14+ MNCs, possibly corresponding to the recently identified CD14+ CD34low cells,31 and lower among CD14− or CD34+ MNCs. When used for therapeutic injection, the CD14+KDR+CXCR2+ cell subset caused a CXCR2-dependent improvement of reendothelialization after arterial injury in nude mice. Selection of the rare subsets by fluorescence-activated cell sorting did not yield sufficient cell numbers, preventing further in vivo homing or recovery experiments as a certain limitation of our study. Hence, circulating KDR+CXCR2+ cells may represent an EPC subpopulation responding to CXCR2 ligands with increased adhesiveness for sites of endothelial damage or dysfunction in vivo, to facilitate endothelial recovery.
In conjunction with previous findings, the homing of EPCs appears to be a complex process involving integrins22,32 but also chemokine receptors, in particular CXCR2. Whereas CXCR4 is crucial for mobilizing vascular PCs, including EPCs, and for recruiting SMC progenitors contributing to neointima formation after injury,12 it is conceivable from our data that CXCR2 may be more important for homing of EPCs after endothelial injury to limit neointima formation. This is supported by our findings that CXCR2 ligands, namely CXCL7, are expressed or presented after wire injury and that blocking CXCL1 aggravates neointima formation by delaying endothelial recovery after denudation.15 As an additional mode of fine tuning, the levels of CXCR2 and CXCR4 expression on EPCs may be differentially regulated by arterial injury and/or concomitant hypoxia, as shown for CXCL12 induction.9,10 Taken together, these novel observations may help to improve our understanding of how EPC homing is regulated in the context of vascular repair and regeneration after endothelial injury.
We thank M. Garbe, S. Knarren, M. Roller, and S. Wilbertz for technical assistance.
Sources of Funding
This work was supported by the Deutsche Forschungsgemeinschaft (We1913/7-1) and the Interdisciplinary Center for Clinical Research “BIOMAT” (NTV B113-a).
Original received May 25, 2006; resubmission received October 9, 2006; revised resubmission received January 13, 2007; accepted January 18, 2007.
Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T, Witzenbichler B, Schatteman G, Isner JM. Isolation of putative progenitor endothelial cells for angiogenesis. Science. 1997; 275: 964–967.
Peichev M, Naiyer AJ, Pereira D, Zhu Z, Lane WJ, Williams M, Oz MC, Hicklin DJ, Witte L, Moore MA, Rafii S. Expression of VEGFR-2 and AC133 by circulating human CD34+ cells identifies a population of functional endothelial precursors. Blood. 2000; 95: 952–958.
Urbich C, Dimmeler S. Endothelial progenitor cells: characterization and role in vascular biology. Circ Res. 2004; 95: 343–353.
Kalka C, Masuda H, Takahashi T, Kalka-Moll WM, Silver M, Kearney M, Li T, Isner JM, Asahara T. Transplantation of ex vivo expanded endothelial progenitor cells for therapeutic neovascularization. Proc Natl Acad Sci U S A. 2000; 97: 3422–3427.
Hristov M, Erl W, Linder S, Weber PC. Apoptotic bodies from endothelial cells enhance the number and initiate the differentiation of human endothelial progenitor cells in vitro. Blood. 2004; 104: 2761–2766.
Guleng B, Tateishi K, Ohta M, Kanai F, Jazag A, Ijichi H, Tanaka Y, Washida M, Morikane K, Fukushima Y, Yamori T, Tsuruo T, Kawabe T, Miyagishi M, Taira K, Sata M, Omata M. Blockade of the SDF-1/CXCR4 axis attenuates in vivo tumor growth by inhibiting angiogenesis in a vascular endothelial growth factor-independent manner. Cancer Res. 2005; 65: 5864–5871.
Zernecke A, Schober A, Bot I, von Hundelshausen P, Liehn EA, Mopps B, Mericskay M, Gierschik P, Biessen EA, Weber C. SDF-1α/CXCR4 axis is instrumental in neointimal hyperplasia and recruitment of smooth muscle progenitor cells. Circ Res. 2005; 96: 784–791.
Zernecke A, Weber KS, Erwig LP, Kluth DC, Schroppel B, Rees AJ, Weber C. Combinatorial model of chemokine involvement in glomerular monocyte recruitment: role of CXC chemokine receptor 2 in infiltration during nephrotoxic nephritis. J Immunol. 2001; 166: 5755–5762.
Liehn EA, Schober A, Weber C. Blockade of keratinocyte-derived chemokine inhibits endothelial recovery and enhances plaque formation after arterial injury in ApoE-deficient mice. Arterioscler Thromb Vasc Biol. 2004; 24: 1891–1896.
Haghnegahdar H, Du J, Wang D, Strieter RM, Burdick MD, Nanney LB, Cardwell N, Luan J, Shattuck-Brandt R, Richmond A. The tumorigenic and angiogenic effects of MGSA/GRO proteins in melanoma. J Leukoc Biol. 2000; 67: 53–62.
Weber KS, Klickstein LB, Weber C. Specific activation of leukocyte beta2 integrins LFA-1 and Mac-1 by chemokines mediated by distinct pathways via the alpha subunit cytoplasmic domains. Mol Biol Cell. 1999; 10: 861–873.
Schober A, Zernecke A, Liehn EA, von Hundelshausen P, Knarren S, Kuziel WA, Weber C. Crucial role of the CCL2/CCR2 axis in neointimal hyperplasia after arterial injury in hyperlipidemic mice involves early monocyte recruitment and CCL2 presentation on platelets. Circ Res. 2004; 95: 1125–1133.
Urbich C, Heeschen C, Aicher A, Sasaki K, Bruhl T, Farhadi MR, Vajkoczy P, Hofmann WK, Peters C, Pennacchio LA, Abolmaali ND, Chavakis E, Reinheckel T, Zeiher AM, Dimmeler S. Cathepsin L is required for endothelial progenitor cell-induced neovascularization. Nat Med. 2005; 11: 206–213.
Chavakis E, Aicher A, Heeschen C, Sasaki K, Kaiser R, El Makhfi N, Urbich C, Peters T, Scharffetter-Kochanek K, Zeiher AM, Chavakis T, Dimmeler S. Role of β2-integrins for homing and neovascularization capacity of endothelial progenitor cells. J Exp Med. 2005; 201: 63–72.
Langer H, May AE, Daub K, Heinzmann U, Lang P, Schumm M, Vestweber D, Massberg S, Schonberger T, Pfisterer I, Hatzopoulos AK, Gawaz M. Adherent platelets recruit and induce differentiation of murine embryonic endothelial progenitor cells to mature endothelial cells in vitro. Circ Res. 2006; 98: e2–e10.
De Boer HC, Verseyden C, Ulfman LH, Zwaginga JJ, Bot I, Biessen EA, Rabelink TJ, Vanzonneveld AJ. Fibrin and activated platelets cooperatively guide stem cells to a vascular injury and promote differentiation towards an endothelial cell phenotype. Arterioscler Thromb Vasc Biol. 2006; 26: 1653–1659.
Massberg S, Konrad I, Schurzinger K, Lorenz M, Schneider S, Zohlnhoefer D, Hoppe K, Schiemann M, Kennerknecht E, Sauer S, Schulz C, Kerstan S, Rudelius M, Seidl S, Sorge F, Langer H, Peluso M, Goyal P, Vestweber D, Emambokus NR, Busch DH, Frampton J, Gawaz M. Platelets secrete SDF-1α and recruit BM-derived progenitor cells to arterial thrombi in vivo. J Exp Med. 2006; 203: 1221–1233.
Baltus T, von Hundelshausen P, Mause SF, Buhre W, Rossaint R, Weber C. Differential and additive effects of platelet-derived chemokines on monocyte arrest on inflamed endothelium under flow conditions. J Leukoc Biol. 2005; 78: 435–441.
George J, Afek A, Abashidze A, Shmilovich H, Deutsch V, Kopolovich J, Miller H, Keren G. Transfer of endothelial progenitor and BM cells influences atherosclerotic plaque size and composition in ApoE knockout mice. Arterioscler Thromb Vasc Biol. 2005; 25: 2636–2641.
Smith DF, Galkina E, Ley K, Huo Y. GRO family chemokines are specialized for monocyte arrest from flow. Am J Physiol Heart Circ Physiol. 2005; 289: H1976–H1984.
De Falco E, Porcelli D, Torella AR, Straino S, Iachininoto MG, Orlandi A, Truffa S, Biglioli P, Napolitano M, Capogrossi MC, Pesce M. SDF-1 involvement in endothelial phenotype and ischemia-induced recruitment of BM progenitor cells. Blood. 2004; 104: 3472–3482.
Romagnani P, Annunziato F, Liotta F, Lazzeri E, Mazzinghi B, Frosali F, Cosmi L, Maggi L, Lasagni L, Scheffold A, Kruger M, Dimmeler S, Marra F, Gensini G, Maggi E, Romagnani S. CD14+CD34low cells with stem cell phenotypic and functional features are the major source of circulating endothelial progenitors. Circ Res. 2005; 97: 314–322.
Blindt R, Vogt F, Astafieva I, Fach C, Hristov M, Krott N, Seitz B, Kapurniotu A, Kwok C, Dewor M, Bosserhoff AK, Bernhagen J, Hanrath P, Hoffmann R, Weber C. A novel drug-eluting stent coated with an integrin-binding cyclic Arg-Gly-Asp peptide inhibits neointimal hyperplasia by recruiting endothelial progenitor cells. J Am Coll Cardiol. 2006; 47: 1786–1795.