Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Chemokines are a large family of cytokines that direct leukocyte migration (Baggiolini, 1998). Chemokines are believed to play a beneficial role in host defense against infection and a harmful role in those diseases exhibited by pathologic inflammation (Locati and Murphy, 1999; Rossi and Zlotnik, 2000). Besides their well established role in leukocyte trafficking, chemokines have been implicated in angiogenesis, tumor growth, and metastasis (Locati and Murphy, 1999; Rossi and Zlotnik, 2000). Neutralization of chemokine receptors is reported to inhibit metastasis (Muller et al., 2001), and chemokines themselves have proven to be effective as immunologic anticancer agents (Dilloo et al., 1996; Braun et al., 2000; Hromas et al., 2000; Nomura and Hasegawa, 2000; Sharma et al., 2000; Vicari et al., 2000). Chemokines are also mediators of inflammatory tissue destruction in a variety of human diseases, such as rheumatoid arthritis, atherosclerosis, myocardial infarction, and adult respiratory distress syndrome (Furie and Randolph, 1995; Strieter et al., 1996). For example, CCL3 and CXCL2 levels correlate with the severity of rheumatoid arthritis (Strieter et al., 1996). Interleukin-8, CCL2, and CXCL2 have been implicated in neutrophil-mediated reperfusion injury after myocardial infarction (Kukielka et al., 1995). CCL5, CCL7, and CCL11 levels are elevated in bronchial epithelium during asthma (Locati and Murphy, 1999).

CCL21 (isolated by us as Exodus-2 and by others as 6-Ckine/secondary lymphoid-tissue chemokine/T cell activator protein-4) (Hedrick and Zlotnik, 1997; Hromas et al., 1997; Nagira et al., 1997; Tanabe et al., 1997) is the most potent regulator of T cell chemotaxis yet described. In addition, the adhesion of circulating T cells to the lymph node high venule endothelium is induced by CCL21. CCL21 has been shown to activate surface lymphocyte function-associated antigen-1 on rolling T cells to induce its interaction with endothelial ICAM-1, immobilizing the T cell on the endothelium before chemotaxis (Campbell et al., 1998; Gunn et al., 1998; Cyster, 1999).

Evidence indicates that CCL21 also regulates the colocalization of lymphocytes and antigen-presenting cells in secondary lymphoid organs (Campbell et al., 1998; Gunn et al., 1998; Cyster, 1999). Secondary lymphoid organs (lymph nodes, spleen, tonsils, and Peyer's patches) act as the initiation site of antigen-mediated adaptive immune responses. T cells survey secondary lymphoid organs, permitting activated dendritic cells in the secondary lymphoid organs to come into contact with many T cells (Campbell et al., 1998; Gunn et al., 1998; Cyster, 1999). The appropriate expression of CCL21 in lymph nodes stimulates the migration of T cells and dendritic cells to specific regions of the node, where antigen presentation can occur.

Evidence also indicates that chemokines are presented to leukocytes by heparin-like glycosaminoglycans (GAGs) on the endothelial cell surface of venules (Webb et al., 1993; Hoogewerf et al., 1997; Locati and Murphy, 1999; Ali et al., 2000; Patel et al., 2001). Binding of the basic chemokines to acidic GAGs may also maintain a local chemokine concentration gradient within inflamed tissue toward which the leukocyte can migrate (Webb et al., 1993; Hoogewerf et al., 1997; Locati and Murphy, 1999; Ali et al., 2000; Patel et al., 2001). However, heparin has been shown to down-regulate leukocyte adherence, migration, and recruitment to a site of injury or inflammation (Lever et al., 2000; Perretti and Page, 2000). In addition, heparin was also shown to inhibit the pathogenesis of inflammatory diseases, including asthma, emphysema, adult respiratory distress syndrome, primary skin allograft rejection, myocardial infarction, rheumatoid arthritis, and inflammatory bowel disease (Naparstek et al., 1993; Nelson et al., 1993; Gaffney and Gaffney, 1996; Hodak et al., 1998; Stefanidou et al., 1999; Tyrrell et al., 1999; Yanaka et al., 2000). In this study, it was found that the ability of CCL21 to stimulate T cell adhesion and chemotaxis was inhibited by heparin, heparan, and low-molecular-weight heparins. This inhibition of T cell migration could be exploited therapeutically in autoimmune T cell infiltrative diseases.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Heparin Column Elution. One microgram of human CCL19, CCL20, or CCL21 was bound to a 1-ml HiTrap heparin-Sepharose column (Amersham Biosciences, Piscataway, NJ) and then eluted with 0.1 M stepwise fractions of KCl between 0.1 and 1.0 M KCl. One percent of each fraction was then run on a 4 to 20% SDS-polyacrylamide gel. The protein was then transferred to a polyvinylidene difluoride membrane, probed with goat anti-human CCL19, CCL20, or CCL21 (R & D Systems, Minneapolis, MN), followed by anti-goat IgG-horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA), and detected with ECL-Plus (Amersham Biosciences). Protein concentration was then quantified by absorbance at 280 nm.

Chemotaxis Assays. All chemokines were purchased commercially from R & D Systems, reconstituted in phosphate-buffered saline, and stored per manufacturer instructions. Chemotaxis assays using 24-well transwells (Corning Inc., Corning, NY) were performed as we described previously (Hromas et al., 1997). Briefly, the chemokine CCL19 (also known as Exodus-3, Epstein-Barr virus-induced molecule-1-ligand chemokine, MIP-3, and CK11), CCL20 (also known as Exodus-1, MIP-3, and liver and activation-regulated chemokine), or CCL21 was added to 1 ml of RPMI 1640 medium supplemented with 10% fetal bovine serum in the lower chamber. For these experiments, 200 ng/ml chemokine was used on the basis of previous experiments that showed that, at this concentration, all three chemokines resulted in the migration of normal human adult peripheral blood T cells, which was above the EC₅₀ for each chemokine and yet submaximal (Christopherson et al., 1999). Half a million normal human adult peripheral T cells in 200 µl of identical medium without chemokine was added to the upper chamber of the transwell (6.5-mm diameter, 5-µm pore size, polycarbonate membrane). Transwells were then incubated for 3 h at 37°C, 5% CO₂.

Adhesion Assays. Static adhesion assays were performed as we described previously (Campbell et al., 1998). Briefly, ICAM-1 was coated on glass slides, and lymphocytes were allowed to settle for 10 min on the surface of the slide. CCL21 was added, and nonadherent cells were washed away. Adherent cells were then fixed and counted.

Heparin-Induced Anticoagulation. Heparin-induced anticoagulation was examined by measuring changes in activated partial thromboplastin times (APTT), according to manufacturer instructions (Abbott Laboratories). Heparin-induced changes in APTT obtained on pooled normal human plasma were obtained at 0, 0.1, 0.2, and 0.3 U/ml heparin in the presence of 0, 100, 500, and 1000 ng/ml CCL21.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Heparin Affinity of CCL21. CCL21 is constitutively expressed in cells lining the high endothelial venules of normal lymph nodes. It mediates the adherence and migration of T cells into the secondary lymphoid organs (Campbell et al., 1998; Gunn et al., 1998; Cyster, 1999). It is proposed that heparin and related GAGs anchor chemokines on the endothelium at sites of inflammation, maintaining local chemokine concentration gradients and presenting them to circulating leukocytes (Webb et al., 1993; Hoogewerf et al., 1997). Whether heparin could function in a similar manner for CCL21 was examined and compared with CCL19 and CCL20. The affinity of the chemokines for heparin was tested by eluting the chemokine bound to a heparin-Sepharose column with 0.1 M stepwise fractions of KCl between 0.1 and 1.0 M KCl. CCL21 protein was found to bind strongly to heparin, eluting at 0.8 M KCl (Fig. 1). CCL19 and CCL20 do not bind as tightly, eluting at 0.5 to 0.6 and 0.6 M KCl, respectively.

View larger version (39K): [in this window] [in a new window]   Fig. 1.   Affinity of CCL19, CCL20, and CCL21 for heparin. Human CCL21 protein binds tightly to heparin. CCL21 elutes from the heparin-Sepharose column at the 0.8 M KCl fraction as detected by Western analysis. CCL19 and CCL20 bind less tightly to heparin, eluting at 0.6M KCl.

Glycosaminoglycan Inhibition of T Cell Migration and Adhesion. To test the effect of GAG binding on CCL21 function, chemotaxis assays were performed using peripheral blood T cells. CCL19 (also known as Exodus-3, Epstein-Barr virus-induced molecule-1-ligand chemokine, MIP-3, and CK11) and CCL20 (also known as Exodus-1, MIP-3, and liver and activation-regulated chemokine), two other related T cell-active chemokines, were included in these assays for comparison (Hedrick and Zlotnik, 1997; Hromas et al., 1997; Nagira et al., 1997; Tanabe et al., 1997; Campbell et al., 1998). CCL19 also uses CCR7, the same receptor as CCL21. Chemotactic responses to 200 ng/ml CCL19, CCL20, and CCL21 in the absence of heparin were 28.6 ± 0.8, 18.2 ± 0.7, and 34.2 ± 0.9%, respectively. Low concentrations (0.1 U/ml) of heparin, easily achievable in vivo, reduced CCL21-mediated T cell chemotaxis by 68% (Fig. 2) (P = 0.013). No reduction in CCL19- or CCL20-induced T cell migration occurred at this physiologic concentration of heparin. However, at the highest concentrations of heparin, which are supraphysiologic, the chemotactic ability of all three chemokines was inhibited. At 10.0 U/ml heparin, 92 ± 1.4% of CCL21 activity is lost, 70 ± 2.5% of CCL19 activity is lost, and 85 ± 0.7% of CCL20 activity is lost (P < 0.001, P = 0.032, and P = 0.043, respectively). Thus, the inhibition of CCL21 chemotactic activity when bound to a GAG implies that this particular chemokine cannot be functionally presented by GAGs as proposed.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Inhibitory effect of heparin on the T cell chemotactic properties of CCL19, CCL20, and CCL21. Low concentrations of heparin inhibit CCL21-induced migration of T cells. At 0.1 U/ml heparin, 68 ± 1.4% of CCL21 chemotactic activity is lost. At 1 U/ml heparin, 92 ± 1.4% of CCL21 activity is lost, 85 ± 0.7% of CCL20 activity is lost, and 70 ± 2.5% of CCL19 activity is lost.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Effect of heparinase on the inhibitory properties of heparin toward CCL21-induced chemotaxis. Heparinase, a glycosylase that degrades heparin into its component-sulfated saccharides, abrogates the inhibition of CCL21 chemotaxis by heparin.

View larger version (54K): [in this window] [in a new window]   Fig. 4.   Inhibition of CCL21-induced T cell chemotaxis by heparin-like GAGs. Varying doses of heparin, heparan, and the low-molecular-weight heparin drugs enoxaparin, dalteparin, and tinzaparin inhibit CCL21-induced peripheral blood T cell migration in transwell chemotaxis assays.

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Inhibition of CCL21-induced T cell adhesion by heparin-like GAGs. CCL21 is an extremely potent inducer of T cell adhesion. With heparin and low-molecular-weight heparins at doses achievable in vivo, T cell adhesion is inhibited.

Effect of CCL21 on Heparin-Induced Anticoagulation. Since heparin significantly inhibited CCL21 chemotactic activity, we next tested whether CCL21 inhibited the anticoagulant activity of heparin. Heparin-induced anticoagulation, as examined by measuring APTT, was unaffected by the presence of CCL21 (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The strong affinity of CCL21 for heparin suggests that heparin and related GAGs would be logical candidates to present endothelial CCL21 to leukocytes bearing its receptor (CCR7), as has been previously shown for other chemokines (Webb et al., 1993; Hoogewerf et al., 1997; Locati and Murphy, 1999; Ali et al., 2000; Patel et al., 2001). However, at physiologic concentrations of heparin and related GAGs, the ability of CCL21 to stimulate chemotaxis and adhesion was markedly inhibited. This inhibition was preferential for CCL21; 100-fold high concentrations of heparin were required to inhibit CCL19 or CCL20. Thus, at physiologic concentrations of GAGs, other chemokines besides CCL21 could still be functionally presented by binding to GAGs (Webb et al., 1993; Hoogewerf et al., 1997; Locati and Murphy, 1999; Ali et al., 2000; Patel et al., 2001). It is possible that the unique inhibition of CCL21 is mediated by the long, highly basic carboxyl-terminal extension of CCL21 that other chemokines lack (Hromas et al., 1997). CCL21 is notably more basic than CCL19 or CCL20 and requires greater concentrations of KCl to elute it from a heparin-Sepharose column. CCL19, CCL20, and CCL21 have estimated charges (pH 7.0) of 7.0, 8.3, and 17.0, respectively, calculated using isolated amino acid pKa values (Stryer, 1995). GAGs could bind this extension and sterically inhibit interaction with the CCL21 receptor.

A large amount of data supports the concept of heparin as an effective anti-inflammatory agent. Heparin has been shown to inhibit the pathogenesis of inflammatory diseases, including asthma, emphysema, adult respiratory distress syndrome, delayed-type hypersensitivity, primary skin allograft rejection, myocardial infarction, rheumatoid arthritis, and inflammatory bowel disease (Gaffney and Gaffney, 1996; Tyrrell et al., 1999; Lever et al., 2000; Perretti and Page, 2000). Previous studies have also shown that heparin has resulted in the reduction of severity of diseases such as rheumatoid arthritis (Gaffney and Gaffney, 1996) and lichen planus (Hodak et al., 1998; Stefanidou et al., 1999) in humans and acute graft versus host disease in mouse models (Naparstek et al., 1993). The pathology of each of these diseases involves local T cell invasion and damage to the involved tissue. Heparin has also been shown to inhibit Rantes binding to CCR5 (Martin et al., 2001), and it inhibits tumor necrosis factor--induced leukocyte adherence and migration into surrounding tissue (Lever et al., 2000). Thus, the anti-inflammatory activity of heparin may be due to its ability to inhibit leukocyte migration into diseased tissue. Beyond this, the mechanism for the anti-inflammatory activity of heparin has not been fully elucidated.

One mechanism for heparin inhibition of leukocyte adhesion that has been proposed is the binding of heparin to L-selectin and P-selectin adhesion molecules, which inhibits leukocyte adhesion (Nelson et al., 1993; Yanaka et al., 2000). Selectins were not used in the adhesion assays performed in this study to exclude this complicating factor. Heparin-like GAGs still inhibited the adhesion of T cells induced by CCL21. It is therefore likely that heparin-like polysaccharides have the ability to inhibit T cell migration and adhesion by tightly binding to CCL21 directly (Fig. 1). Pretreatment of T cells with heparin did not alter their chemotactic ability, suggesting that GAGs do not inhibit CCL21-induced adhesion and chemotaxis by binding to CCR7. It should be noted that blocking T cell access to selectins would not explain the inhibition of CCL21-induced T cell chemotaxis by GAGs. Thus, the steric inhibition of CCL21 interaction with CCR7 seems the most likely explanation for the finding that heparin and related GAGs have anti-inflammatory properties.

Inflamed venule endothelial cells in T cell infiltrative autoimmune diseases expressing CCL21 may act as a point of regulation for T cell migration into these tissues. GAGs, such as heparan and, to a lesser extent, heparin from degranulated mast cells, are present on the endothelial cell surface (Metcalfe et al., 1997; Kussie et al., 1999; Vlodavsky et al., 1999; Dempsey et al., 2000; Lanzavecchia and Sallusto, 2000). The presence of GAGs at the site of CCL21 expression could indicate that a regulatory loop governing T cell migration may exist on the surface of endothelial cells lining blood vessels. A balance between CCL21 production and the presence of heparin-like GAGs may regulate CCL21 activity and thereby T cell migration. In normal tissues, the balance between GAGs and CCL21 favors the GAGs, such that there is enough GAG to bind and inhibit whatever CCL21 is produced. In T cell inflammatory diseases, this balance favors CCL21, such that there is not enough GAG to bind and inhibit all available CCL21. Under this circumstance, enough available CCL21 is present on the endothelial cell surface to interact with the CCR7 receptor on rolling T cells, causing the cell to adhere to the surface of the venule and migrate into the inflamed tissue.

Another level to this endothelial regulation of T cell migration is likely. Glycosylases are produced locally that can degrade GAGs into their component saccharides, which we found did not inhibit CCL21. For example, the presence of heparinase is markedly increased in endothelial cells during inflammation (Kussie et al., 1999; Vlodavsky et al., 1999; Dempsey et al., 2000; Lanzavecchia and Sallusto, 2000). The increased concentrations of these glycosylases at the endothelium could degrade GAGs, which were found here to destroy the inhibition of CCL21, and further shift the migratory balance toward CCL21 activity during an inflammatory state. It has also been proposed that soluble GAGs can displace immobilized GAGs bound to chemokines (Kuschert et al., 1999). It is therefore possible that a secondary regulatory mechanism may exist between soluble and immobilized GAGs and chemokine activity. Thus, these data raise the possibility of the existence of a complex autoregulatory loop governing T cell migration at the endothelial surface between CCL21 and GAGs.

Heparin-induced anticoagulation, as examined by measuring changes in the APTT, was unaffected by the presence of CCL21. This indicates that even though heparin clearly affects the activity of CCL21, CCL21 does not affect the anticoagulation activity of heparin. There are two possibilities for why this occurs, and both may be active simultaneously. One possibility is that CCL21 does not bind to heparin at the site where heparin interacts with antithrombin III, allowing heparin to still inhibit coagulation while it is bound by CCL21. The second possibility is that heparin, as a large multimer, has multiple active sites that CCL21 cannot bind completely. All CCL21 could be bound to heparin, but not all heparin-active sites could be bound to CCL21. Much higher concentrations of CCL21 would be necessary to bind all the active sites and block the anticoagulant properties of heparin.

It was shown here that GAGs, including low-molecular-weight heparins, markedly inhibited CCL21-induced T cell adherence and chemotaxis. This raises the possibility that the endothelial regulation of T cell migration could be targeted for therapy of these T cell infiltrative diseases, which are in need of new treatment initiatives. The pathology of a number of other diseases relies on aberrant T cell infiltration, such as rheumatoid arthritis, inflammatory bowel disease, and organ transplant rejection. Low-molecular-weight heparins have an excellent clinical safety record, including a decreased risk of heparin-induced thrombocytopenia. Thus, they may be effective agents at treating these T cell infiltrative diseases by interfering with the abnormal recruitment of T cells from the circulation to sites of pathologic inflammation by endothelial cells expressing CCL21. The reduction in T cell migration into involved tissue would reduce the T cell-inflicted damage seen in those diseases. Randomized clinical trials are indicated to address this possibility.