Synthetic di-sulfated iduronic acid attenuates asthmatic response by blocking T-cell recruitment to inflammatory sites

Tie2-Ext1^iKO Mice Show Diminished T-Cell Accumulation and Inflammatory Response in the Lung in a Mouse Model of OVA-Induced Asthma.

Recently, we have developed Tie2-Ext1^iKO mouse line in which embryonic lethality of homozygous Ext1^−/− mouse is avoidable. Indeed, there was little expression of endothelial heparan sulfate in the lungs of Tie2-Ext1^iKO mice (Fig. 1A). To determine the role of endothelial heparan sulfate in asthma, we sensitized and challenged these mice with OVA allergen. In OVA-challenged control mice, periodic acid–Schiff (PAS) staining showed significant mucus accumulation in airway epithelium. In addition, a large number of CD3⁺ T cells and eosinophils were detected in the peribronchial regions (Fig. 1B). In contrast, there was little infiltration of CD3⁺ T cells and eosinophil cells, as well as mucus secretion in Tie2-Ext1^iKO mice. There was also a significant reduction of total cell number including lymphocytes, macrophages, and eosinophils in BALF in Tie2-Ext1^iKO mice (Fig. 1C). Moreover, whereas BALF in control mice contained high levels of Th2 cytokines IL-5 and IL-13 but not the Th1 cytokine IFN-γ, only low levels of IL-5 and IL-13 could be detected in that of Tie2-Ext1^iKO (Fig. 1D), suggesting that OVA-triggered Th2 activation was abolished in Tie2-Ext1^iKO mice.

To determine the effect of deficiency in heparan sulfate on airway responsiveness, methacholine (MCh) responsiveness was assessed in OVA-challenged Tie2-Ext1^iKO mice. Compared with OVA-challenged control mice, OVA-challenged Tie2-Ext1^iKO mice showed significantly less response to MCh at 48 mg/mL (P = 0.0033) (Fig. 1E). Taken together, these results indicate that endothelial heparan sulfate is essential for lymphocyte accumulation at sites of allergic inflammation in the lung, and that the reduced accumulation is associated with less airway responsiveness.

Interruption of Lymphocyte Trafficking Pathway Attenuates Asthmatic Response.

We have previously reported that endothelial heparan sulfate is required for lymphocyte recruitment to secondary lymph nodes. Here, we hypothesized that the attenuation of the OVA induced asthmatic response in Tie2-Ext1^iKO mice is due to blocking of lymphocyte recruitment from regional lymph nodes. Previous studies have demonstrated that the homeostatic lymphocyte trafficking pathway was impaired in L-selectin ligand-deficient [GlcNAc6ST-1/2 double knockout (G6ST DKO)] mice (21, 22). To evaluate the contribution of homeostatic lymphocyte trafficking to the lung, G6ST DKO mice were challenged with OVA. Similar to Tie2-Ext1^iKO mice, G6ST DKO mice had significantly fewer CD3⁺ T cells and eosinophil cells around the airway walls compared with control mice (Fig. 2A). Moreover, G6ST DKO mice BALF contained significantly fewer eosinophils and lymphocytes (Fig. 2B). These data support that accumulation of lymphocytes in lung tissue is mainly contributed by their recruitment from regional lymph nodes.

Glycan Microarray Showed Chemokine CCL20 Binds Short Heparin-Like Sulfated Oligosaccharide.

Our previous studies have indicated that heparan sulfate is important to maintain morphologic localization of the homeostatic chemokine CCL21 (20). Here, the results from OVA-challenged Tie2-Ext1^iKO asthmatic mice led us suggest that heparan sulfate also interacts with inflammation-related chemokines such as CCL20. Thus, we analyzed the binding profile of CCL20 to a heparin-like carbohydrate library, which contains carbohydrates with various sulfation patterns ranging from monosaccharides to hexasaccharides (Fig. 3A) by our glycan microarray system (18, 19). The screening results showed that CCL20 strongly bound to hexasaccharides 1, 2, and moderately bound to hexasaccharide 5 and tetrasaccharide 7. This indicates that the affinity of CCL20 toward heparan sulfate becomes higher with increasing sulfations levels (Fig. 3 B and C). Notably, CCL20 exhibited high affinity to disaccharide 10 and even monosaccharide 11. Observed fluorescent intensities of those carbohydrates were comparable to 5-kDa natural heparin.

A Unique and Unnatural Monosaccharide Di-S-IdoA Inhibits Heparan Sulfate–CCL20 Interaction.

Monosaccharide 11 is Di-S-IdoA, which contains two axial sulfate groups. Because Di-S-IdoA is an unnatural synthetic monosaccharide that has never been isolated from a natural source, we were prompted to further characterize its interaction with CCL20 in more detail. The data from surface plasmon resonance (SPR) showed that CCL20 binds to immobilized Di-S-IdoA in the micromolar range (K_D = 2.9 × 10⁻⁶ M) (Fig. 4A). We next analyzed the inhibitory effect of Di-S-IdoA. Interaction between immobilized hexasaccharides 1 and CCL20 was blocked by Di-S-IdoA in a concentration-dependent manner and 500 µM Di-S-IdoA inhibited at a comparable level of 5,000 µM heparin (Fig. 4B), whereas nonsulfated IdoA showed no such inhibitory activity at 5 mM. We next synthesized Di-S-IdoA with aminopentyl linker (Figs. S1 and S2). Interestingly, in the cultured F2 cell, which is known to express endogenous heparan sulfate, Di-S-IdoA also interfered with the CCL20–heparan sulfate interactions in a dose-dependent manner (Fig. 4C). These results suggest that Di-S-IdoA is an effective as a functional inhibitor of CCL20 chemokine activity. To next study the specificity of Di-S-IdoA, the inhibitory effect of Di-S-IdoA on the bindings between the various proteins and endothelial cells was assayed. It is known that CCL21 (23, 24), IL-8 (23, 25), L-selectin (26, 27), and complement component 5a (C5a) (28, 29) are involved both in the binding to heparin/heparan sulfate in vitro and in the asthma pathogenesis. The result showed that Di-S-IdoA did not block the attachment of CCL21 to mouse endothelial F2 cells, whereas heparin efficiently blocked (Fig. 4D). Di-S-IdoA significantly blocked the binding of L-electin to F2 cells. However, Di-S-IdoA showed even higher inhibitory effect than heparin in IL-8 binding. In this experimental model, C5a did not show any bindings to F2 cells. Those results indicate Di-S-IdoA has unique binding preferences distinct from heparin.

Inhibitory Effect of Di-S-IdoA in OVA-Challenged WT Mice.

To test the possibility of Di-S-IdoA as an inhibitor of CCL20 activity in vivo, we injected either PBS or Di-S-IdoA into tail vein right before the reexposure of OVA in WT mice. Pretreatment of Di-S-IdoA decreased the mucus secretion and recruitment of CD3⁺ T cells in the lung tissues (Fig. 5A). We also observed a reduced total leukocyte number in Di-S-IdoA–treated and neutralizing anti-CCL20 antibody-treated mice but not nonsulfated IdoA (Non-S-IdoA)-treated mice (Fig. 5B). These results suggest that i.v. administered Di-S-IdoA can access at the inflamed postcapillary venule, where it inhibits the activity of chemoattractants.

Because CCL20 has been demonstrated to be mainly synthesized from airway epithelial cells, we hypothesized that administration of Di-S-IdoA through intranasal inhalation would interfere the localization and presentation on the vascular endothelial cell surface in inflamed lung. To test the accessibility of Di-S-IdoA through the lung alveoli, we administered Di-S-IdoA by inhalation before OVA challenge in WT mice. Indeed, we found robust expression of CCL20 on the vascular endothelial cell wall in the OVA-challenged WT mice. In contrast, there was little expression of CCL20 in the Di-S-IdoA–pretreated mice, although heparan sulfate is present on vascular surface (Fig. 5C). These data suggest that CCL20 presentation on the endothelial cells at the lung inflammatory site was blocked by Di-S-IdoA administration. Indeed, the number of total leukocytes in BALF was dramatically decreased in 100 µg of Di-S-IdoA treatment (P < 0.05) compared with saline-treated mice (Fig. 5D). Blood tests from the WT mice treated with Di-S-IdoA showed no side effect in liver function, kidney function, and blood cell count (Table S1). Taken together, our results demonstrated that inhalation administration of Di-S-IdoA was effective in reducing airway inflammation in allergen-challenged WT mice.

Reagents and Animals.

Tie2-Ext1^iKO mice were generated and maintained as reported (20). To induce the deletion of endothelial heparan sulfate, mice of 4-wk age were treated with doxycycline for 3 consecutive weeks before experiments. All protocols for animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the Sanford-Burnham Medical Research Institute, and were conducted in accordance with National Institutes of Health guidelines. OVA was purchased from Sigma, heparan sulfate antibody 3G10 from US Biological, recombinant human CCL20 from R&D and Shenandoah Biotech, recombinant human CCL21 from Sino Biological, recombinant human IL-8 from Cell Sciences, recombinant mouse L-selectin from Life Technologies, recombinant human C5a from BioVision, eosinophil anti-MBP antibody (Dr. James Lee, Mayo Clinic, Scottsdale, AZ) (37), anti-CD3ε antibody from Santa Cruz, anti-macrophage inflammatory protein 3α (CCL20) antibody from Abcam, neutralizing anti-CCL20 antibody and anti-CCL21 antibody from R&D, anti-His antibody from Clontech, and anti-von Willebrand factor antibody from Dako.

OVA-Induced Model of Asthma in Mice.

For sensitization, mice were intraperitoneally administrated a suspension containing 50 µg of OVA and 0.5 mg of aluminum hydroxide (Sigma) in 200 µL of PBS on day 0 and day 12. On days 23, 25, and 27, the mice were briefly anesthetized with isoflurane and challenged intranasally with 20 µg of OVA in 50 µL of PBS through the nose. For studies with a therapeutic intervention, 100 µg of Di-S-IdoA or Non-S-IdoA, or 50 µg of neutralizing anti-CCL20 antibody was administered either i.v. or inhalationally 30 min before each OVA challenge. Twenty-four hours after the last OVA challenge, the mice were killed and the lung bronchoalveolar lavage was performed by infusion of 1 mL of 0.1% PBS through the trachea with Surflash IV catheters (Terumo). The lung lavage was subjected to leukocyte counting by a cytometer, and each subpopulation was analyzed by cytospin followed by a stain with Hema-3 (Fisher).

Immunohistochemistry.

Mouse lung lobes were fixed with 4% (wt/vol) PFA followed by paraffin embedding and histological examination by H&E staining and PAS staining. For evaluation of immune infiltration into lung inflammatory sites, tissues were stained by anti-CD3ε antibody and anti-MBP antibody described above.

Airway Responsiveness to MCh.

Airway responsiveness to MCh was assayed 24 h after the final OVA challenge in anesthetized, intubated, and ventilated mice as previously described (38). Intubated and ventilated mice were anesthetized intraperitoneally. The dynamic airway resistance was determined in mice exposed to nebulized PBS or MCh at 0, 3, 24, and 48 mg/mL.

Glycan Microarray.

Microarrays were fabricated as previously reported (39). Microarray slide was incubated with 100 µL (5 µg) of rhCCL20 (R&D Systems) in HBS-N containing 0.01% Tween 20 for 1 h at room temperature under mild shaking, washed three times with HBS-N, and dried by centrifugation. The slide was then reacted with 100 µL (2 µg) of anti-hCCL20 goat IgG (R&D Systems) for 1 h, followed by incubation with anti-goat Alexa Fluor 594 antibody (Invitrogen) for 1 h. Slides were scanned with a GenePix 4300A microarray scanner and analyzed by GenePix Pro software. Binding of CCL20 to immobilized heparin-like oligosaccharide 1 in the presence of Non-S-IdoA and Di-S-IdoA was observed in the same manner. The amino group of each monosaccharide sample was acylated before use by treatment with a solution of acetic anhydride–triethylamine–methanol (2:3:5, vol/vol/vol) followed by condensation and lyophilization. The inhibition solution contained 5 mM 5-kDa heparin (Santa Cruz Biotechnology), 5 mM Non-S-IdoA, or 5, 50, and 500 µM Di-S-IdoA.

Immunohistochemistry.

For heparan sulfate staining, tissue sections were treated with heparatinase at 37 °C for 1 h before incubation with anti-heparan sulfate 3G10 antibody. Alexa Fluor 568 anti-mouse IgG2b was used for secondary reaction. Vascular endothelial cells were visualized by anti-von Willebrand factor antibody (1:100 dilution). Lung endothelial CCL20 could not be visualized by the general fluorescent staining protocol. Instead, localization of blood vessel CCL20 was examined by i.v. injection of anti-CCL20 antibody (25 µg per mouse) through mouse tail vein. Thirty minutes after the injection, lungs were collected and frozen in OCT compound (Tissue Tek). Frozen tissue sections were then incubated with Alexa Fluor 568 anti-rabbit IgG (Invitrogen).

To avoid the cross-reaction in double staining with CCL20, sections were blocked with anti-rabbit IgG antibody and protein A. Then the sections were incubated with anti-von Willebrand factor antibody, followed by incubation with FITC–anti-rabbit IgG-Fc antibody (Jackson ImmunoResearch).

Statistical Analyses.

All data are showed as means ± SD. Unpaired two-tailed Student t test was used for statistical analyses. We considered P values of less than 0.05 as statistically significant. Degrees of statistical significance are presented as *P < 0.05, **P < 0.01, and ***P < 0.001.