Role of plasma kallikrein–kinin system activation in synovial recruitment of endothelial progenitor cells in experimental arthritis

Authors


Abstract

Objective

To examine whether activation of the plasma kallikrein–kinin system (KKS) mediates synovial recruitment of endothelial progenitor cells (EPCs) in arthritis.

Methods

EPCs were isolated from Lewis rat bone marrow, and expression of progenitor cell–lineage markers and functional properties were characterized. EPCs were injected intravenously into Lewis rats with arthritis, and their recruitment and formation of de novo blood vessels in inflamed synovium were evaluated. The role of plasma KKS was examined using a plasma kallikrein inhibitor (EPI-KAL2) and an antikallikrein antibody (13G11). A transendothelial migration assay was used to determine the role of bradykinin and its receptor in EPC mobilization.

Results

EPCs from Lewis rats exhibited a strong capacity to form tubes and vacuoles and expressed increased levels of bradykinin type 2 receptor (B2R) and progenitor cell markers CD34 and Sca-1. In Lewis rats with arthritis, EPCs were recruited into inflamed synovium at the acute phase of disease and formed de novo blood vessels. Inhibition of plasma kallikrein by EPI-KAL2 and 13G11 significantly suppressed synovial recruitment of EPCs and hyperproliferation of synovial cells. Bradykinin stimulated transendothelial migration of EPCs in a concentration-dependent manner. This was mediated by B2R, as demonstrated by the finding that knockdown of B2R with silencing RNA completely blocked bradykinin-stimulated transendothelial migration. Moreover, bradykinin selectively up-regulated expression of the homing receptor CXCR4 in EPCs.

Conclusion

These observations demonstrate a novel role of plasma KKS activation in the synovial recruitment of EPCs in arthritis, acting via kallikrein activation and B2R-dependent mechanisms. B2R might be involved in the mobilization of EPCs via up-regulation of CXCR4.

The pathogenesis of rheumatoid arthritis (RA) is critically dependent on neovascularization of the synovium, which occurs before clinical symptoms appear (1). Synovial neovascularization creates a direct conduit by which circulating leukocytes that exacerbate inflammation may enter the joint and provides nutrients to hyperproliferative synovium. Although neovascularization in adult tissue was previously thought to occur exclusively through angiogenesis, it was recently noted that circulating endothelial progenitor cells (EPCs) are recruited to inflamed synovium and participate in synovial neovascularization in arthritis (2, 3). EPCs, which are derived from adult bone marrow, have a strong capacity to home to inflamed and ischemic tissue and differentiate into vessel-forming endothelial cells. The mechanism for EPC homing in arthritis, however, remains unknown, and several important issues need to be addressed, e.g., how EPCs are regulated under the inflammatory conditions in arthritis, the cellular and molecular mechanisms by which EPCs home to inflamed synovium, how EPCs differentiate into new blood vessels, and the extent to which EPCs contribute to synovial neovascularization. Understanding the mechanisms that control the process of EPC homing is important, because it will enable the identification of key molecules as novel targets for pharmacologic blockade by specific inhibitors and receptor antagonists in arthritis.

The plasma kallikrein–kinin system (KKS) plays an important role in the pathogenesis of arthritis (4, 5). The KKS consists of 4 plasma proteins: factor XII, factor XI, prekallikrein, and high molecular weight kininogen (HK) (6). HK is responsible for the binding of this system complex to cell membrane or the contact surface, which is necessary for the system's assembly and activation. Upon activation, plasma kallikrein cleaves HK to release bradykinin, a nonapeptide. The inflammatory response in arthritis includes an acute phase with edema, pain, and neutrophil migration, all of which are known to be associated with plasma kallikrein activation and release of bradykinin. In patients with RA, elevated levels of plasma kallikrein and bradykinin are detected in synovial fluid and are positively correlated with the degree of joint pain and inflammation (7, 8).

The notion that the KKS is important in arthritis pathogenesis is supported by several observations in animal models. HK in Lewis rats has a mutation (Ser511Asn) that renders it susceptible to cleavage by plasma kallikrein, and the administration of the streptococcal cell wall polymer peptidoglycan–polysaccharide (PG-PS) induces a systemic inflammatory response, including arthritis, in Lewis rats but not in other strains such as Buffalo and Fischer rats (9). Along with synovitis and joint erosion in Lewis rats, plasma prekallikrein and HK levels are decreased, likely due to consumption of the precursor proteins following KKS activation. A specific kallikrein inhibitor prevents arthritis and the systemic complications in the PG-PS model, by blocking activation of the KKS (10). The B/N/Ka rat strain has a severe deficiency of plasma kininogen due to a single point mutation, Ala163Thr, which results in defective secretion of kininogen from the liver. Kininogen-deficient rats on a Lewis genetic background exhibit attenuated acute and chronic inflammatory arthritis (11), demonstrating that plasma KKS and HK cleavage product kinins are key mediators of inflammatory disease.

Taken together, the above findings show that activation of the KKS is critical for the pathogenesis of arthritis in the Lewis rat model. Thus, this model is appropriate for use in investigating the potential link between KKS activation and the dynamic recruitment of EPCs into inflamed synovium during the process of arthritis. The aim of this study was to determine whether plasma KKS activation is associated with synovial recruitment of EPCs in arthritis. We first established a method to evaluate the recruitment of EPCs into inflamed synovium in the Lewis rat model of arthritis. Subsequently, we investigated the effect of specific plasma kallikrein inhibition on synovial recruitment of EPCs and tested whether bradykinin and bradykinin type 2 receptor (B2R) are involved in the mobilization of EPCs.

MATERIALS AND METHODS

Animals.

Eight-week-old Lewis rats (Charles River) weighing 180–200 gm were used in this study. Green fluorescent protein (GFP)–transgenic Lewis rats were provided by Dr. Eiji Kobayashi (Jichi Medical School, Shimotsuke, Japan). Rats were maintained under climate-controlled conditions in a 12-hour light/dark cycle. The animals were fed standard rodent chow and water ad libitum. The health status of the animals was monitored in accordance with the guidelines of the Institutional Animal Care and Use Committee.

Isolation of EPCs from Lewis rats.

Bone marrow cells were aspirated from femurs and tibias of 8-week-old female Lewis rats and GFP-transgenic Lewis rats. Mononuclear cells (MNCs) were isolated by density-gradient centrifugation (Histopaque 1083; Sigma) for 30 minutes at 400g. Rat CD34+ bone marrow MNCs were purified using mouse anti-rat CD34 antibody (Santa Cruz Biotechnology) and a CELLectionPan Mouse IgG Kit (Invitrogen). They were resuspended in endothelial cell growth culture medium 2 (EGM-2; Cambrex) with 10% fetal bovine serum (FBS), cultured in tissue culture plates coated with type I collagen (Invitrogen), and incubated with 5% CO2 at 37°C. The cells were expanded in large scale. Fluorescent EPCs, isolated from GFP-transgenic Lewis rats or labeled with carboxyfluorescein diacetate–succinimidyl ester (Molecular Probes), were used in the assays. Rat lung microvessel endothelial cells (rLMECs; VEC Technologies) were used as a control.

Measurement of messenger RNA (mRNA) expression by reverse transcription–polymerase chain reaction (RT-PCR) and quantitative real-time RT-PCR.

Expression of mRNA was measured by RT-PCR as previously described (12). Briefly, total RNA was prepared using TRIzol reagent (Invitrogen). RNA (a total of 100 ng) was used as template in a one-step RT-PCR procedure (SuperScript One-Step RT-PCR with Platinum Taq; Invitrogen). RT for complementary DNA synthesis was accomplished with 30-minute incubation at 50°C, which was followed by PCR cycling as follows: initialization step at 94°C for 2 minutes, followed by 30 cycles of denaturation at 94°C for 15 seconds, annealing at 55°C for 30 seconds, and extension at 72°C for 1 minute, using primers at 0.2 μmoles/liter. The RT-PCR products were identified by 4% agarose gel electrophoresis. Primer sequences were as follows: B2R forward 5′-CCATCTCTCCACCTGCATTG-3′, reverse 5′-CGTCTGGACCTCCTTGAACT-3′; B1R forward 5′-TCTTCCTGGTGGTGGCTATC-3′, reverse 5′-CGTTCAACTCCACCATCCTT-3′; von Willebrand factor (vWF) forward 5′-CACAGGTAGCACACATCACT-3′, reverse 5′-CTCAAAGTCTTGGATGAAGA-3′; CD34 forward 5′-GCCCAGTCTGAGGTTAGGCC-3′, reverse 5′-ATTGGCCTTTCCCTGAGTCT-3′; CD31 forward 5′-GCCCTGTCACGTTTCAGTTT-3′, reverse 5′-CTGCAATGAGCCCTTTCTTC-3′; Sca-1 forward 5′-CGGTCATTCAGACCACACACAG-3′, reverse 5′-TGGGTTGAAGTTCTCGCTCTTG-3′; CD14 forward 5′-CTTGTTGCTGTTGCCTTTGA-3′, reverse 5′-CGTGTCCACACGCTTTAGAA-3′; CD144 forward 5′-AGGACGTGGTGCCAGTAAAC-3′, reverse 5′-CTGTGATGTTGGCGGTATTG-3′; CXCR4 forward 5′-CTGCATCATCATCTCCAAGC-3′, reverse 5′-GGAAAGGATCTTGAGGCTGG-3′; α4 integrin forward 5′-CCCAGGCTACATCGTTTTGT-3′, reverse 5′-ATGGGAGTGAGGATGTCTCG-3′; CD11b forward 5′-TTACCGGACTGTGTGGACAA-3′, reverse 5′-AGTCTCCCACCACCAAAGTG-3′; E-selectin forward 5′-TTTTTGGCACGGTATGTGAA-3′, reverse 5′-AGGTTGCTGCCACAGAGAGT-3′. A primer pair for β-actin (G5740; Promega) was used as a control.

Quantitative real-time PCR was performed using the Maxima SYBR Green/ROX qPCR Master Mix kit according to the instructions of the manufacturer (Fermentas). Primer sequences were as follows: CD31 forward 5′-ATCCTGTCGGGTAACGATGTA-3′, reverse 5′-CTTCGGAGACTGGTCACAATG-3′; CD34 forward 5′-TTAAGGGAGACATCAAATGTTCA-3′, reverse 5′-GCTAGATTCAAGGAGCATACACTA-3′; vWF forward 5′-CAATACGGAAGCATCAATACCA-3′, reverse 5′-CCAAGGAGGTATCCATGATGAT-3′; Sca-1 forward 5′-AGACGGCAAAAGTCAGGTTAC-3′, reverse 5′-AACCTGTATCATTTGCCCTCATC-3′; B1R forward 5′-CAGGTGAAGCTGTGAGCTCTT-3′, reverse 5′-AAGAAGCAGATAGTGATGACGAA-3′; B2R forward 5′-AAGGACGATCCTCACTCGTCT-3′, reverse 5′-TCTGAAAAGGTCCCGTTATGA-3′; β-actin forward 5′-ACGTTGACATCCGTAAAGACC-3′, reverse 5′-GCCACCAATCCACACAGAGT-3′. In an ABI 7500 system (Applied Biosystems), PCR was conducted for 45 cycles of 95°C for 15 seconds and 60°C for 1 minute, preceded by an initial step at 95°C for 10 minutes. The results were analyzed using SDS software, version 2.1.

Three-dimensional vasculogenesis assay.

We evaluated the vasculogenic capacity of Lewis rat EPCs in culture within 3-dimensional collagen gels, as recently described (13). Briefly, EPCs at passage 2 were resuspended in endothelial cell basal medium 2 (EBM-2) (2 × 106 cells/ml) and mixed with neutralized type I collagen (4 mg/ml; BD Biosciences) at a volume ratio of 1:1. One hundred microliters of mixture was transferred to one well of a 96-well culture plate and incubated for 20 minutes at 37°C to allow gel formation. The gel matrices were then overlaid with EGM-2 containing growth factor and 10% FBS and cultured for 48 hours. After fixation with 2% paraformaldehyde, the gels were stained with 0.1% toluidine blue (Sigma) and visualized under a microscope (Eclipse TE300; Nikon)

Transendothelial migration of EPCs.

Transendothelial migration was analyzed using 96-well Transwell filters (3.0 μm polycarbonate membrane; Corning Costar). Rat LMECs were cultured on collagen-coated Transwell filters until becoming confluent, and then fluorescent EPCs (3.0 × 104) in EBM-2 containing 2% FBS were placed on top of the rLMEC monolayer. EBM-2 containing 2% FBS and bradykinin (Sigma) was added to the lower compartment. After incubation for 18 hours at 37°C, the filters were fixed with 2% paraformaldehyde, and fluorescent EPCs that had migrated into the lower side of the filters were visualized with a fluorescence microscope and counted in random microscopic fields.

Construction of small interfering RNA (siRNA) and transfection of EPCs with siRNA.

The sequences selected as the targeting region were as follows: for rat B2R 5′-CCGCACTGGAGAACATCTTTGTCCT-3′, and for CXCR4 5′-GGATAACTACTCCGAAGAA-3′. A nonsilencing sequence (5′-AACCTGCGGGAAGAAGTGG-3′) was used as a control (control siRNA). The siRNA oligos labeled with fluorescein were synthesized by Invitrogen. Transfection of EPCs was performed with HiPerfect (Qiagen).

Rat models of arthritis and implantation of EPCs.

In the PG-PS–induced arthritis model, female specific pathogen–free Lewis rats received a single intraperitoneal injection of PG-PS (15 μg of rhamnose/gm body weight; BD Lee Laboratories) (11). In the collagen-induced arthritis (CIA) model, female specific pathogen–free Lewis rats received 2 injections of 250 μg bovine type II collagen in Freund's complete adjuvant (Chondrex) at the base of the tail, 1 week apart. The severity of arthritis was assessed by measuring ankle joint diameter with a digital caliper (Ultra-Call Mark III; F. V. Fowler). The mean of triplicate measurements of each hind paw was recorded every day during the first week and then every other day during the remainder of the protocol. Change in joint diameter (in mm) from baseline (day 0) was calculated. Fluorescent rEPCs (1 × 107) in a volume of 300 μl were administered to recipient animals via tail vein injection. After specified periods of time, the animals were killed by administration of CO2. Hind limb joints were harvested and fixed in buffered formalin.

Histopathologic and immunohistochemical analyses.

Paws were decalcified in formic acid (Fisher Scientific) or EDTA. The tissue was embedded in paraffin, and the sections were stained with hematoxylin and eosin for microscopic examination. Some sections were stained with polyclonal anti-vWF antibody (Abcam) plus 0.2% Triton X-100 (Sigma) and Alexa Fluor 633–labeled goat anti-rabbit antibody (Molecular Probes). The sections were photographed under a Leica TCS SP5 confocal laser scanning microscope, and images were processed using Adobe Photoshop 9.0.

Quantitative analysis of synovial recruitment of EPCs during the acute phase of arthritis.

The paraffin-embedded, 10-μm–thick ankle joint sections were stained with Syto Red Fluorescent Nucleic Acid Stain (Molecular Probes) for visualizing total synovial cells including the implanted fluorescent EPCs. The sections were examined under a TCS SP5 confocal microscope. Total synovial cells and EPCs in synovial tissue specimens were enumerated.

Statistical analysis.

Mean ± SEM results from experiments performed at least 3 times were calculated. Statistical comparisons were performed by Student's t-test (for comparisons of 2 groups only) or one-way analysis of variance and Student-Newman-Keuls test (for comparisons of multiple groups). P values less than 0.05 were considered significant.

RESULTS

Characterization of rat EPCs.

In this study, we first isolated EPCs from Lewis rats. Bone marrow cells were cultured on type I collagen surfaces, and multiple colonies often appeared after 7 days. A typical colony with cobblestone-like morphology was observed on day 7 and day 12 (Figure 1) and was subsequently collected for large-scale expansion. EPCs are often characterized by the progenitor capacity of tube and vacuole formation and expression of progenitor and endothelial markers (13). In an in vitro angiogenesis assay we recently established (13), EPCs formed vacuoles and tubes (Figure 1B), suggesting that EPCs possess the endothelial progenitor capacity for differentiation. To characterize whether the collected cells expressed the markers for endothelial cell and progenitor cell lineage, we measured their mRNA expression by RT-PCR, because the availability of antibodies that recognize rat antigens is limited. Rat LMECs were used as a control for differentiated endothelial cell lineage. Total RNA was extracted from EPCs and rLMECs at passage 3. As shown in Figure 1C, rEPCs and rLMECs expressed similar levels of mRNA for certain endothelial cell markers, such as CD144 (VE-cadherin), CD31, and vWF. EPCs exclusively expressed the hematopoietic progenitor cell markers Sca-1 and CD34, and they did not express the monocyte-lineage marker CD14. Both EPCs and rLMECs expressed B2R but not B1R.

Figure 1.

Isolation and characterization of rat endothelial progenitor cells (EPCs). A, Appearance of a colony with cobblestone-like morphology (arrows) after CD34-positive bone marrow cells isolated from female Lewis rats were cultured on collagen surfaces for 7 days (left). The colony had become large after 12 days of culture (right). Bars = 100 μm. B, Formation of vacuoles (V) and tubes (T) after the isolated EPCs were cultured for 48 hours in the presence of 25 ng/ml vascular endothelial growth factor in 3-dimensional (3-D) type I collagen gels. Bar = 50 μm. C, Endothelial cell– and progenitor cell–lineage markers of EPCs characterized by reverse transcription–polymerase chain reaction (RT-PCR). Total RNA was extracted from rat EPCs (rEPCs) and rat lung microvessel endothelial cells (rLMECs) at passage 3. vWF = von Willebrand factor; B2R = bradykinin type 2 receptor. D, Formation of multiple colonies of green fluorescent protein (GFP)–EPCs isolated from GFP-transgenic Lewis rat bone marrow. Original magnification × 100. E, Quantification of mRNA for endothelial cell– and progenitor cell–lineage markers of GFP-EPCs by real-time RT-PCR (n = 3). Results were normalized against mRNA levels in rLMECs, and relative quantification between rLMECs and EPCs was calculated using the equation 2 − ΔΔCt. Values are the mean ± SEM.

We also isolated EPCs from bone marrow of GFP-transgenic Lewis rats. In these animals the formation of original colonies appeared on day 7 (Figure 1D). A colony was collected for expansion, and its expression of mRNA for endothelial cell– and progenitor cell–lineage markers was quantitated by real-time RT-PCR. Compared with rLMECs, EPCs expressed higher levels of the progenitor cell–lineage markers CD34 (mean ± SEM relative expression 8.13 ± 0.03 versus 1.00 ± 0.01) and Sca-1 (2.00 ± 0.01 versus 1.00 ± 0.01), but lower levels of the endothelial cell–lineage markers CD31 (0.72 ± 0.05 versus 1.00 ± 0.07) and vWF (0.13 ± 0.08 versus 1.00 ± 0.04). Interestingly, B2R expression by EPCs was 4-fold higher than B2R expression by rLMECs (3.90 ± 0.03 versus 1.00 ± 0.03), and neither EPCs nor rLMECs expressed B1R (Figure 1E).

EPCs home to inflamed synovium and form new blood vessels in Lewis rats with arthritis.

KKS activation is closely associated with the pathogenesis of arthritis induced by PG-PS in Lewis rats (10). In the present study, rats developed an acute arthropathy within 72 hours after a single intraperitoneal injection of PG-PS (Figure 2A). Light microscopic examination of hematoxylin and eosin–stained ankle sections revealed inflamed and hyperplastic synovium with joint destruction, MNC infiltration, and pannus formation (Figure 2B). In the PG-PS–induced arthritis model, acute inflammation usually resolves within 7 days and is followed by a secondary chronic phase, which begins ∼14 days after injection. We used this model to investigate synovial recruitment of EPCs during the acute phase and new vessel formation during the chronic phase. Venous injection of fluorescent EPCs was performed on day 1 after administration of PG-PS. On day 5, although fluorescent EPCs were not detected in the synovial tissue of disease-free ankle joints, they had accumulated in inflamed synovium (Figure 2C).

Figure 2.

Homing of endothelial progenitor cells (EPCs) to inflamed synovium. Female specific pathogen–free Lewis rats were administered a single intraperitoneal injection of saline or peptidoglycan–polysaccharide (PG-PS) (15 μg of rhamnose/gm body weight) on day 0. A, Photographs of the hind paws 5 days after injection of saline or PG-PS. B, Representative hematoxylin and eosin–stained sections, showing features of destructive arthritis with mononuclear cell infiltration (arrows) and advanced pannus (P) after PG-PS treatment. C, Findings in the synovium of rats in which fluorescent EPCs were injected into a tail vein 1 day after administration of PG-PS, and in the inflamed synovium of arthritic rats without implantation of EPCs. Rats were killed on day 5, and hind limbs were fixed in 4% paraformaldehyde and decalcified. Paraffin-embedded sections were viewed and photographed under a fluorescence microscope. Original magnification × 200.

To assess vessel formation by implanted EPCs in the inflamed synovium of arthritic rats on day 12, sections from the ankle joints were stained with rabbit anti-vWF antibody and Alexa Fluor 633–labeled goat anti-rabbit antibody. As shown in Figure 3, the implanted fluorescent EPCs formed blood vessels, which were positive for anti-vWF staining. The recruitment of implanted EPCs to inflamed joints and formation of new blood vessels suggest that EPCs home to inflamed synovium during the acute phase of arthritis and participate in the formation of new blood vessels.

Figure 3.

Involvement of endothelial progenitor cells (EPCs) in the formation of blood vessels in inflamed synovium. A, To observe de novo vessel formation by EPCs in the arthritic joint, female Lewis rats were administered a single intraperitoneal injection of peptidoglycan–polysaccharide (15 μg of rhamnose/gm average body weight) on day 0, and fluorescent EPCs were injected into a tail vein 1 day later. Rats were killed on day 12, and hind limbs were fixed in 4% paraformaldehyde and decalcified. Paraffin-embedded sections were stained with rabbit anti–von Willebrand factor (anti-vWF) antibody and Alexa Fluor 633–labeled goat anti-rabbit antibody. The sections were photographed under a confocal laser scanning microscope, and images were processed using Adobe Photoshop 9.0. New blood vessels formed by fluorescent EPCs stained positive for anti-vWF (arrows in d). a, b, and c show a green fluorescence channel image, a red fluorescence channel image, and a brightfield image, respectively; d shows an overlay of a, b, and c. Bar = 60 μm. B, Higher-magnification views of the boxed area in panel d of part A are shown. Bar = 20 μm.

Inhibition of plasma kallikrein by EPI-KAL2 and 13G11 attenuates synovial recruitment of EPCs during the acute phase of arthritis.

The differentiation of EPCs into new vessels in synovium during the chronic phase of arthritis suggests that EPCs are recruited during the acute phase. Synovial recruitment of implanted EPCs was quantified 6 days after injection of PG-PS. Paraffin-embedded ankle joint sections were stained with Syto Red Fluorescent Nucleic Acid Stain for visualizing total synovial cells. As shown in Figure 4A, total synovial cells and fluorescent EPCs in the synovial tissue could be viewed and enumerated. This method allowed us to quantitatively evaluate synovial hyperplasia and synovial recruitment of EPCs.

Figure 4.

Inhibition of endothelial progenitor cell (EPC) recruitment into inflamed synovium by the plasma kallikrein inhibitor EPI-KAL2. A, Female Lewis rats were administered a single intraperitoneal injection of peptidoglycan–polysaccharide (PG-PS) on day 0, and fluorescent EPCs were injected into a tail vein 1 day later. Rats were killed on day 5. Paraffin-embedded sections of ankle joints were stained with Syto Red, and homing of EPCs to inflamed synovium was assessed. Total synovial cells (red) and EPCs (green) (arrows) were photographed under a fluorescence microscope. b = bone; GFP = green fluorescent protein. B, Two groups of Lewis rats (n = 6 per group) were administered an intraperitoneal injection of PG-PS on day 0, and fluorescent EPCs were injected into a tail vein 1 day later. Beginning on day 0, one group (control) received phosphate buffered saline (PBS) (circles) and the other received EPI-KAL2 (10 mg/kg body weight) (squares) twice daily, and changes in joint diameter from that measured on day 0 were determined. Values are the mean ± SEM. ∗ = P < 0.01; ∗∗ = P < 0.005, versus EPI-KAL2–treated animals. C, Two groups of female Lewis rats (n = 4 per group) were treated as described in B and then killed on day 5. Paraffin-embedded ankle joint sections were stained with Syto Red and photographed under a fluorescence microscope. Ten high-power fields (hpf; 400×) were evaluated per section, and total synovial cells and fluorescent EPCs were enumerated. Values are the mean ± SEM.

In the rat model of arthritis, activation of the KKS can be modulated by inhibition of plasma kallikrein (10). In this study we first used a potent plasma kallikrein inhibitor, EPI-KAL2 (Ki <1 nM), which is the Kunitz domain inhibitor specific for plasma kallikrein (14). As seen in Figure 4B, joint swelling in control rats with untreated arthritis started 24 hours after PG-PS injection, with the increase in joint diameter becoming marked on day 2 and peaking on day 3 (mean ± SEM 1.78 ± 0.23 mm higher than baseline), followed by a decrease. In contrast, the PG-PS–injected rats treated with EPI-KAL2 did not develop any significant paw swelling during the acute phase of disease (days 2–4), although mild changes occurred during the chronic phase after day 3 (Figure 4B). These data indicate that EPI-KAL2 prevented joint swelling (arthritis) throughout the acute phase of the disease and greatly ameliorated joint inflammation during the chronic phase. Concomitantly, treatment with EPI-KAL2 significantly reduced the total number of synovial cells (by ∼50%) and the number of recruited EPCs (by almost 75%) (Figure 4C).

Additionally, we reexamined the effect of plasma kallikrein inhibition using an inhibitory antikallikrein antibody, 13G11 (15), in a CIA model. Joint swelling in Lewis rats started 14 days after the first injection of collagen and reached maximal levels on day 20. Treatment with 13G11 significantly inhibited joint swelling over the course of observation (Figure 5A). Moreover, administration of 13G11 in this CIA model significantly reduced the total number of synovial cells (by ∼40%) and the number of recruited EPCs (by almost 55%), compared with untreated controls (Figure 5B).

Figure 5.

Attenuation of arthritis severity and endothelial progenitor cell (EPC) recruitment into inflamed synovium by the antikallikrein antibody 13G11. A, Female Lewis rats were immunized with a mixture of type II collagen and Freund's complete adjuvant on days 0 and 7. Beginning on day 7, rats (n = 6 per group) received intraperitoneal injections of phosphate buffered saline (PBS) (control) or 13G11 (8 mg/kg body weight) every other day, and changes in joint diameter from that measured on day 0 were determined. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.001, versus PBS-treated animals. B, Two groups of female Lewis rats (n = 4 per group) were treated as described in A. Fluorescent EPCs were injected into a tail vein on day 14, and the animals were killed on day 18. Hind limbs were fixed and decalcified, and paraffin-embedded ankle joint sections were stained with Syto Red and photographed under a fluorescence microscope. Total synovial cells and fluorescent EPCs (per high-power field [hpf]) in the inflamed synovium were enumerated as described in Figure 4. Values are the mean ± SEM.

Bradykinin interaction with B2R up-regulates homing receptor CXCR4 expression and stimulates transendothelial migration of EPCs.

Upon activation, plasma kallikrein cleaves HK to release bradykinin. Transendothelial migration is an essential step in EPC homing to sites of inflammatory and ischemic tissue (16). To further investigate whether and how plasma KKS activation regulates EPC homing, we measured bradykinin-mediated transendothelial migration of EPCs. As EPCs expressed the bradykinin receptor B2R but not B1R (Figures 1C and D), we knocked down B2R, using an siRNA approach. Figure 6A shows that B2R mRNA expression in EPCs was markedly down-regulated by specific siRNA. In a transendothelial migration assay using 96-Transwell filters, bradykinin increased the transmigration of EPCs over the differentiated rLMECs, in a concentration-dependent manner (Figure 6B). Before and after the transmigration, the integrity of the mature endothelial cell monolayer was maintained (results available from the corresponding author upon request). The transmigration of EPCs expressing B2R siRNA was significantly decreased (Figure 6B), suggesting that B2R is required for bradykinin-mediated transendothelial migration of EPCs.

Figure 6.

Bradykinin (BK) type 2 receptor (B2R) mediates transendothelial migration of endothelial progenitor cells (EPCs) by up-regulation of CXCR4 expression. A, EPCs were transfected with control small interfering RNA (siRNA) (lane 1) and B2R siRNA (lane 2) for 48 hours, and expression of mRNA for B2R, B1R, and β-actin (control) was examined by reverse transcription–polymerase chain reaction (RT-PCR). B, After transfection with fluorescein-labeled siRNA for 48 hours, EPCs were seeded onto confluent rat lung microvessel endothelial cells (rLMECs) on Transwell filters. Bradykinin was added to the lower compartment at 100 nM or 1,000 nM or no further treatment was administered, and migration assay was performed as described in Materials and Methods. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01, versus control (no addition of bradykinin). C, After transfection with siRNA for 48 hours, EPCs were starved with 1% serum for 6 hours and then left unstimulated or stimulated for 6 hours with bradykinin at 100 nM or 1,000 nM. The expression of homing receptor mRNA was analyzed by RT-PCR; findings indicated that bradykinin-induced up-regulation of CXCR4 expression is dependent on B2R. Results shown are representative of 3 independent experiments. D, After transfection with siRNA for 48 hours, expression of mRNA for CXCR4 and β-actin (control) was examined by RT-PCR. Results shown are representative of 3 independent experiments. E, After transfection with fluorescein-labeled control siRNA or CXCR4 siRNA for 48 hours, EPCs were seeded onto confluent rLMECs. Bradykinin was added to the lower compartment at 100 nM or 1,000 nM or no further treatment was administered, and migration assay was performed as described in Materials and Methods. Values are the mean ± SEM. ∗ = P < 0.005; ∗∗ = P < 0.001, versus control (no addition of bradykinin).

Further, we examined whether bradykinin modulates the expression of major homing receptors, including CXCR4, α4 integrin, E-selectin, and CD11b. We found that bradykinin selectively up-regulated the expression of CXCR4 (Figure 6C), but not of the other 3 receptors (results not shown). Knockdown of B2R prevented the up-regulation of CXCR4 (Figure 6C), indicating that the effect of bradykinin likely occurs via B2R. To further examine whether bradykinin mediates transendothelial migration through up-regulation of CXCR4, the expression of CXCR4 mRNA after B2R siRNA transfection was examined by RT-PCR. Compared with expression observed after control siRNA transfection, CXCR4 expression was found to be reduced after transfection with B2R siRNA (Figure 6D). Down-regulation of CXCR4 inhibited transendothelial migration of EPCs mediated by bradykinin at 100 nM and 1,000 nM (Figure 6E).

DISCUSSION

There is a growing body of evidence indicating that bone marrow–derived EPCs contribute to synovial neovasculogenesis, a critical step in the development of arthritis (2, 3). To date, our understanding of the molecular mechanisms by which EPCs are recruited from the circulation to inflamed tissue remains very limited. Experimental and clinical observations have demonstrated that the KKS plays a critical role in arthritis pathogenesis, but the molecular and cellular mechanisms by which this system mediates arthritis remain unknown. Upon activation of the KKS, plasma kallikrein cleaves HK to liberate bradykinin; the remaining activation product is HKa, the 2-chain cleaved HK. Recently we demonstrated that HKa regulates EPC functions in vitro (13, 17), revealing the connection between KKS activation and EPC biology. In the present study we have identified another role of the KKS in the homing function of EPCs. Our results indicate that 1) in the Lewis rat model of arthritis, EPCs are recruited to inflamed synovium and form new vessels; 2) inhibition of plasma kallikrein by EPI-KAL2 and 13G11 attenuates synovial recruitment of EPCs; 3) EPCs isolated from Lewis rats constitutively express B2R, by which bradykinin stimulates transendothelial migration of EPCs; and 4) bradykinin selectively up-regulates expression of the homing receptor CXCR4 via B2R. Our current results demonstrate a novel link between synovial neovasculogenesis and activation of the plasma KKS in the pathogenesis of arthritis and reveal a previously undescribed mechanism for the homing of EPCs to inflamed synovium.

The successful isolation of EPCs from Lewis rats and establishment of the model for detecting synovial recruitment of implanted EPCs allowed us to document a role of the KKS in EPC homing in arthritis. Since EPCs express high levels of B2R (Figure 1E), we hypothesized that activation of the KKS, via the release of bradykinin, stimulates synovial recruitment of EPCs in arthritis. Indeed, the plasma kallikrein inhibitor EPI-KAL2 and the antikallikrein antibody 13G11 attenuated the recruitment of EPCs to the inflamed synovium (Figures 4 and 5), demonstrating that activation of the KKS mediates the homing of EPCs to inflamed synovium in arthritis. These inhibitors exert their effect, at least in part, through inhibition of bradykinin production in plasma during the acute phase of arthritis.

A previous study by investigators at our institution indicated that a selective antagonist of B2R attenuates arthritis (18), suggesting that bradykinin is an important mediator of arthritis and its effect occurs mainly through B2R. A recent study has demonstrated that B2R mediates the homing process of EPCs (19). Results of the present investigation enabled us to define the role of B2R in transendothelial migration of EPCs, a critical step in the EPC homing process. Bradykinin stimulated transendothelial migration of EPCs in vitro, and this was blocked by down-regulation of B2R expression (Figure 5). During the process of joint inflammation, a chemokine gradient is established, directing the homing of bone marrow–derived circulating EPCs. We now provide novel evidence that the interaction between bradykinin and B2R contributes to the homing of EPCs into inflamed synovium with neovascularization potential. Since levels of bradykinin are elevated in the synovial fluid of arthritic animals and patients with RA (20), high expression of B2R in EPCs enables them to sense kinin gradients and to facilitate cell migration and invasion of EPCs. CXCR4 is a critical receptor for stem/progenitor cell homing, and bradykinin selectively up-regulates CXCR4 via activation of B2R (Figure 5), thereby stimulating the synovial recruitment of EPCs. Thus, EPCs are likely a previously unidentified target for bradykinin in the setting of acute arthritis.

The continuous recruitment of EPCs to inflamed synovium and their excessive utilization for neovascularization result in the depletion of circulating EPCs in RA patients, which may dampen vascular repair (21). We have recently reported that HKa, another product of HK cleavage by plasma kallikrein, inhibits vasculogenic differentiation of EPCs and accelerates the onset of their senescence (13, 17). Therefore, our current observation that inhibition of plasma kallikrein suppresses synovial recruitment of EPCs suggests that plasma kallikrein might serve as a therapeutic target in arthritis, with dual benefits: inhibition of synovial neovasculogenesis and improvement of circulating EPC function. Further in-depth investigations into the cellular and molecular mechanisms by which the KKS regulates EPC homing and differentiation should improve our understanding of arthritis pathogenesis.

Taken together, these findings demonstrate for the first time that EPC homing to inflamed synovium during the acute phase of arthritis is associated with activation of the plasma KKS, which underlies a novel role of this system in the pathogenesis of arthritis. During the development and progress of arthritis, plasma kallikrein activation results in the cleavage of HK, which in turn causes release of bradykinin. Bradykinin up-regulates the expression of CXCR4, thereby mediating the transendothelial migration of EPCs and their homing to inflamed synovium, in which they participate in synovial neovascularization. The current results not only provide new information for understanding the molecular and cellular mechanisms of synovial neovasculogenesis, but also support the hypothesis that therapeutic approaches targeting KKS activation have considerable potential in arthritis.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Wu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Dai, Agelan, Yang, Zuluaga, Sexton, Colman, Wu.

Acquisition of data. Dai, Yang, Wu.

Analysis and interpretation of data. Dai, Sexton, Colman, Wu.

ADDITIONAL DISCLOSURES

Author Sexton is an employee of Dyax.

Acknowledgements

We would like to thank Dr. Eiji Kobayashi for providing breeding pairs of GFP-transgenic Lewis rats, and Susan Elizabeth Seta, Christina Hu, and Zhanli Xie for excellent technical assistance.

Ancillary