QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: jpet.102.047704v1 306/3/903    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Egger, L. A. Articles by Detmers, P. A. Search for Related Content PubMed PubMed Citation Articles by Egger, L. A. Articles by Detmers, P. A. Pubmed/NCBI databases Compound via MeSH Substance via MeSH Hazardous Substances DB (L)-PHENYLALANINE

INFLAMMATION AND IMMUNOPHARMACOLOGY

A Small Molecule ₄₁/₄₇ Antagonist Differentiates between the Low-Affinity States of ₄₁ and ₄₇: Characterization of Divalent Cation Dependence

Pharmacology (L.A.E., J.C., U.K., P.A.D.), High Throughput Screening (C.M.), Immunology and Rheumatology (G.V.R., E.M., R.A.M.), Medicinal Chemistry (L.S.L., S.E.d.L., D.N.Y., G.Y., W.K.H.), Drug Metabolism (D.C.D., C.E.R., M.A.W., A.N.J.), Merck & Co., Rahway, New Jersey; Aventis (J.A.S.), Bridgewater, New Jersey; and Biogen, Inc. (R.B.P., D.M.S., W.-C.L., M.A.C.), Cambridge, Massachusetts

Received December 4, 2002; accepted May 22, 2003.

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
An ₄₁/₄₇ dual antagonist, ³⁵S-compound 1, was used as a model ligand to study the effect of divalent cations on the activation state and ligand binding properties of ₄ integrins. In the presence of 1 mM each Ca²⁺/Mg²⁺, ³⁵S-compound 1 bound to several cell lines expressing both ₄₁ and ₄₇, but 2S-[(1-benzenesulfonyl-pyrrolidine-2S-carbonyl)-amino]-4-[4-methyl-2S-(methyl-{2-[4-(3-o-tolyl-ureido)-phenyl]-acetyl}-amino) pentanoylamino]-butyric acid (BIO7662), a specific ₄₁ antagonist, completely inhibited ³⁵S-compound 1 binding, suggesting that ₄₁ was responsible for the observed binding. ³⁵S-Compound 1 bound RPMI-8866 cells expressing predominantly ₄₇ with a K_D of 1.9 nM in the presence of 1 mM Mn²⁺, and binding was inhibited only 29% by BIO7662, suggesting that the probe is a potent antagonist of activated ₄₇. With Ca²⁺/Mg²⁺, ³⁵S-compound 1 bound Jurkat cells expressing primarily ₄₁ with a K_D of 18 nM. In contrast, the binding of ³⁵S-compound 1 to Mn²⁺-activated Jurkat cells occurred slowly, reaching equilibrium by 60 min, and failed to dissociate within another 60 min. The ability of four ₄₁/₄₇ antagonists to block binding of activated ₄₁ or ₄₇ to vascular cell adhesion molecule-1 or mucosal addressin cell adhesion molecule-1, respectively, or to ³⁵S-compound 1 was measured, and a similar rank order of potency was observed for native ligand and probe. Inhibition of ³⁵S-compound 1 binding to ₄₁ in Ca²⁺/Mg²⁺ was used to identify nonselective antagonists among these four. These studies demonstrate that ₄₁ and ₄₇ have distinct binding properties for the same ligand, and binding parameters are dependent on the state of integrin activation in response to different divalent cations.

₄ integrins are constitutively expressed on a variety of leukocytes and can bind to shared or distinct binding partners. ₄₁ is expressed on lymphocytes, eosinophils, and monocytes and mediates adhesion to vascular cell adhesion molecule-1 (VCAM-1) expressed on the endothelium and to the connecting segment-1 (CS-1) subdomain of human fibronectin in the extracellular matrix. ₄₁ and ₄₇ are coexpressed on peripheral blood leukocytes, and ₄₇ is highly expressed on a discrete subpopulation of gut-homing memory T and B lymphocytes, mediating lymphocyte adhesion within the vasculature of the gastrointestinal tract, where its major ligand, mucosal addressin cell adhesion molecule-1 (MAdCAM-1), is preferentially expressed on high endothelial venules (Butcher, 1999). Although both ₄₇ and ₄₁ can bind VCAM-1 and CS-1, ₄₁ does not bind MAdCAM-1, and ₄₇ binds to MAdCAM-1 with higher affinity than to VCAM-1 or to CS-1 (Berlin et al., 1993).

Key motifs for the binding of ₄₇ and ₄₁ to native ligands have been defined as leucine-aspartic acid-threonine in MAdCAM-1 (Viney et al., 1996), isoleucine-aspartic acid-serine in VCAM-1 (Wang et al., 1995), and leucin-aspartic acid-valine in CS-1 (Wayner and Kovach, 1992). Small molecule antagonists of ₄₇ that mimic the LDT motif have been described that block the binding of ₄₇-expressing cells to MAdCAM-Ig in the presence of Mn²⁺ (Carson et al., 1997; Shroff et al., 1998; Martin et al., 1999; Harriman et al., 2000; Egger et al., 2002). Similarly, antagonists of ₄₁ have been reported to block binding of Mn²⁺-activated (Jackson et al., 1997; Vanderslice et al., 1997; Lin et al., 1998; Hagmann et al., 2001; Muller et al., 2001) and unactivated ₄₁ (Chen et al., 1999, 2001) to ligand in vitro.

The essential role of cation-binding sites in regulating integrin function is known, but the coordination of each cation-binding site and the individual role of different metal cations is not well understood (Leitinger et al., 2000). All integrin -subunits have seven homologous 60-amino acid repeats at the N terminus that have been predicted to fold into a -propeller structure (Shimaoka et al., 2002), and three to four putative Ca²⁺ binding sites are located within repeats 4 through 7. A metal ion-dependent activation site motif is a unique Mg²⁺/Mn²⁺ binding site located in the I-domain of the -subunit, and divalent cation bound at this site has a structural role in coordinating the binding of ligand to the I-domain containing integrins. Although ₄ does not contain an I-domain, an I-like domain that contains a metal ion-dependent activation site-like motif is present in the -chain of all integrins. Although the effect of divalent cations on ₄₇-ligand interactions has not been extensively characterized, recent studies have shown that Ca²⁺ is essential to support rolling under shear flow, whereas Mg²⁺ can promote firm adhesion of cells expressing ₄₇ to MAdCAM-1 (de Chateau et al., 2001).

To assess the effect of divalent cations on the activation state of ₄ integrins expressed on human lymphocytes, we used a novel dual ₄₁/₄₇ antagonist, ³⁵S-compound 1 (Fig. 1), as a model ligand. A similar approach has been used to study multiple activation states of ₄₁ through their different affinities for a small molecule ligand (Chen et al., 1999, 2001), but the binding of a small molecule ligand to different activation states of ₄b₇ has not been described. These studies provide new information that ₄₁ and ₄₇ have distinct binding affinities for the same small molecule ligand, and binding is dependent on the state of integrin activation in response to different divalent cations.

View larger version (16K): [in this window] [in a new window]   Fig. 1. Structure of antagonists of 41/47 integrins. The structures of compounds 1, 2, 3, and 4 (Hagmann et al., 2001; Kopka et al., 2002; Lin et al., 2002), TR14035 (Sircar et al., 1999a), and BIO7662 (Chen et al., 2001) are shown. Chemical names are given in "Compounds".

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
Compounds. Compound 1, N-[N-benzenesulfonyl-4(R)-cyclopropylamino-2(S)-prolyl]-(L)-4-(2',6'-bismethoxyphenyl)phenylalanine; compound 2, N-{[N-(3,5-dichlorobenzenesulfonyl)]-2(S)-methylprolyl}(L)-4-(2',6'-bismethoxyphenyl)phenylalanine; compound 3, N-{[N-(3,5-dichlorobenzenesulfonyl)]-2(S)-methylprolyl}-(L)-4-(2',6'-bishydroxyphenyl)phenylalanine; and compound 4, N-{[N-(3,5-dichlorobenzenesulfonyl)]-2(R)-methylprolyl}-(D)-phenylalanine, were synthesized as described previously (Hagmann et al., 2001; Kopka et al., 2002; Lin et al., 2002) or by using similar methods. TR14035, N-(2,6-dichlorobenzoyl)-(L)-4-(2',6'-bis-methoxyphenyl) phenylalanine, was synthesized as described previously (Sircar et al., 1999a). BIO7662, 2S-[(1-benzenesulfonyl-pyrrolidine-2S-carbonyl)-amino]-4-[4-methyl-2S-(methyl-{2-[4-(3-o-tolyl-ureido)-phenyl]-acetyl}-amino) pentanoylamino]-butyric acid, was synthesized as described previously (Chen et al., 2001). The structures of all compounds are shown in Fig. 1. ³⁵S-Compound 1 was synthesized by reacting [³⁵S]PhSO₂Cl with an appropriately diprotected amine precursor in dichloromethane in the presence of ditertbutylmethylpyridine. After sequential deprotection of the resulting [³⁵S]sulfonamide intermediate, crude ³⁵S-compound 1 was purified by semipreparative HPLC and formulated as a solution in methanol with a concentration of 1.14 mCi/ml. The radiochemical purity of the compound was >95% as measured by reverse phase HPLC, and the specific activity was measured to be 722 Ci/mmol by liquid chromatography/mass spectrometry. The radioligand in MeOH was aliquoted and stored at -20°C. The binding affinity for the radiolabeled compound was indistinguishable from that of the unlabeled compound as measured by the ability of compound to block the binding of native ligands to cells expressing ₄₇ and ₄₁.

Antibodies and Cell Lines. The following purified monoclonal antibodies were obtained from BD PharMingen (San Diego, CA): 4B4 (mouse anti-human ₁), FIB27 (rat anti-mouse ₇ that cross-reacts with human ₇), and isotype controls (mouse IgG1, rat IgG2b). HP2/1 (mouse anti-human ₄) was obtained from Coulter/Immunotech (Hialeah, FL). The following cell lines were used: RPMI-8866 cells (human B cell line) obtained from John A. Wilkins (University of Manitoba, Winnipeg, Canada); Jurkat and HUT-78 (human T cell lines) from American Type Culture Collection (Manassas, VA); and K562/₄₇ cells, a stably transfected human erythroleukemia cell-line obtained from David J. Erle (University of California, San Francisco, San Francisco, CA).

Binding of Native Ligands to Cells Expressing ₄₇ and ₄₁. A ligand binding assay for Mn²⁺-activated ₄₇ has been described previously (Egger et al., 2002) and was performed by incubating RPMI-8866 cells (7.5 x 10⁵ cells/well) with <200 pM iodinated human MAdCAM-Ig. Similarly, a ligand binding assay for activated ₄₁ was performed by incubating Jurkat cells (5 x 10⁵ cells/well) expressing ₄₁ with <100 pM iodinated VCAM-Ig in the presence of Mn²⁺, as described previously (Hagmann et al., 2001). Purified VCAM-Ig and MAdCAM-Ig were labeled with ¹²⁵I using Bolton Hunter reagent and purified using HPLC gel filtration chromatography, and specific radioactivities were in excess of 1,100 Ci/mmol. Compounds were evaluated by incubating radioligand, compound (prepared in DMSO; <1% DMSO final concentration), cells, and binding buffer (25 mM HEPES, 150 mM NaCl, 3 mM KCl, 2 mM glucose, and 0.1% bovine serum albumin, pH 7.4) containing 1 mM MnCl₂ at 25°C for 30 min (₄₁ assays) or 45 min (₄₇ assays) in a 96-well Millipore (Bedford, MA) multiscreen MHVBN filtration plate. After filtration and a single wash with binding buffer, the filtration plates were dried and transferred to adaptor plates. After adding 100 µl of Microscint-20 (PerkinElmer Life Sciences, Boston, MA) to each well, the plates were sealed, placed on a shaker for 1 min, and counted on a PerkinElmer Top-Count. Wells containing cells + radioligand + 1 µM compound or DMSO alone served as controls to calculate 100 and 0% inhibition, respectively.

Binding of a Dual ₄₁/₄₇ Antagonist Probe, ³⁵S-Compound 1, to Cells Expressing ₄₇ or ₄₁. Equilibrium binding studies were performed by incubating either RPMI-8866 cells (7.5 x 10⁵ cells/tube) expressing ₄₇ or Jurkat cells (5 x 10⁵ cells/tube) expressing ₄₁ in binding buffer containing either 1 mM MnCl₂ or 1 mM each CaCl₂ and MgCl₂ with 0 to 30 nM (for ₄₇ studies) or 0 to 60 nM (for ₄₁ studies) ³⁵S-compound 1 in siliconized Eppendorf microfuge tubes for 1 h at 4°C. All ³⁵S-compound 1 binding studies with RPMI-8866 cells were conducted in the presence or absence of a specific ₄₁ antagonist, 100 nM BIO7662 (Chen et al., 2001), to selectively block ₄₁. The cells were pelleted by centrifugation at 20,000g for 3 min, washed twice with binding buffer at 4°C, transferred to a scintillation vial containing 5 ml of CytoScint (ICN Pharmaceuticals, Costa Mesa, CA), and cell-associated ³⁵S-compound 1 was measured by scintillation counting. Tubes containing cells + radioligand + 1 µM compound 1 or DMSO alone served as controls to calculate 100 and 0% inhibition, respectively. Data were analyzed by nonlinear regression to calculate B_max and K_D values.

Kinetic analysis of ³⁵S-compound 1 binding to cells expressing ₄ was performed by incubating RPMI-8866 cells (7.5 x 10⁵ cells/tube) expressing ₄₇ or Jurkat cells (5 x 10⁵ cells/tube) expressing ₄₁ in binding buffer containing either 1 mM MnCl₂ or 1 mM CaCl₂ and 1 mM MgCl₂ with 6.5 nM ³⁵S-compound 1 and 5 nM unlabeled compound 1 in siliconized Eppendorf microfuge tubes for 2 to 120 min at 4°C. RPMI-8866 cells were pretreated with 100 nM BIO7662 (Chen et al., 2001) as described above. Binding was terminated by adding 5 µM compound 1 at each time point. Cells were immediately transferred to an ice-bath for 10 min, pelleted by centrifugation at 20,000g for 3 min, and cell associated ³⁵S-compound 1 was determined by scintillation counting as described above.

When the rate of ³⁵S-compound 1 dissociation was evaluated, reaction mixtures were incubated with 6.5 nM ³⁵S-compound 1 and 5 nM unlabeled compound 1 for 1 h at 4°C, followed by the addition of 5 µM compound 1 for another 2 to 120 min. At each time point, the cells were pelleted by centrifugation, washed twice with 4°C binding buffer containing either 1 mM MnCl₂ or 1 mM each CaCl₂ and MgCl₂, and counted for cell-associated ³⁵S-compound 1 as described above. On rates (k_on), off rates (k_off), and K_D values for the binding of ³⁵S-compound 1 were determined from kinetic measurements. Prism 3.0 software was used to calculate k_obs (min^-1) and k_off (min^-1) values from the on and off rate binding curves, respectively: k_on (min^-1 nM^-1) = (k_obs - k_off)/[ligand], and K_D (M) = k_off/k_on.

A ³⁵S-compound 1 binding assay for activated or unactivated ₄ was performed by incubating either RPMI-8866 cells (7.5 x 10⁵ cells/well), K562/₄₇ cells (1 x 10⁵ cells/well), HUT-78 cells (5 x 10⁵ cells/well), or Jurkat cells (5 x 10⁵ cells/well) in binding buffer containing either 1 mM MnCl₂ or 1 mM CaCl₂ and 1 mM MgCl₂ with less than 150 pM ³⁵S-compound 1 for ₄₇ or ₄₁ studies. RPMI-8866 cells were pretreated with 100 nM BIO7662 (Chen et al., 2001) as described above. Test compounds were evaluated by incubating radioligand, compound, cells, and binding buffer at 25°C for 45 min in a 96-well multiscreen filtration plate on a shaking platform. After filtration and a single wash with binding buffer containing either 1 mM MnCl₂ or 1 mM CaCl₂ and 1 mM MgCl₂, the plates were processed and counted as described above for the binding of native ligand to cells expressing ₄. Nonspecific binding (NSB) was determined by the addition of 1 µM compound 1.

Quantitative FACS Analysis. A total of 10⁶ RPMI-8866 or Jurkat cells were incubated for 30 min on ice in FACS buffer (phosphate-buffered saline containing 1 mM each CaCl₂ and MgCl₂, 5% fetal bovine serum, 100 µg/ml goat IgG, and 0.05% sodium azide) containing saturating levels of the following phycoerythrin-conjugated antibodies: FIB504 rat anti-mouse ₇ (2.4 µg/ml; cross-reacts with human ₇), MAR4 mouse anti-human ₁ (80 µg/ml), 9F10 mouse anti-human ₄ (10 µg/ml), mIgG1 isotype, and rIgG2a isotype controls. All phycoerythrin-conjugated antibodies were obtained from BD PharMingen. Cells were washed in FACS buffer and resuspended in FACS buffer containing 1 µg/ml propidium iodide. Cells were analyzed by a FACScan flow cytometer (BD Biosciences, Franklin Lakes, NJ). Standardized quantum R-phycoerythrin microbeads (Flow Cytometry Standards Corp., Fishers, IN) were analyzed by flow cytometry and used to create a calibration curve that relates mean fluorescence intensities to molecules of equivalent soluble fluorescence for use in calculating receptor density values.

Statistical Analysis. Curve fits and statistics were performed using KaleidaGraph (Synergy, Reading, PA) and GraphPad Prism (GraphPad Software Inc., San Diego, CA) with a one-way analysis of variance (nonparametric test) followed by a Tukey's post test if overall P < 0.05, and a paired t test was used when comparing only two sets of data. Data were analyzed by nonlinear regression with an equation for one-site binding to calculate B_max and K_D values. Nonlinear regression analysis was used with equations for one-phase exponential association with no weighting or one-phase exponential decay with no weighting to obtain curve fits for association and dissociation plots, respectively. R² values were used as an indication of the goodness of fit, and both single- and double-phase exponential equations were compared to obtain the best fit. Double reciprocal plots were analyzed by linear regression analysis.

	Results

Top Abstract Materials and Methods Results Discussion References

 
Potency of Compound 1 in ₄₁ and ₄₇ Ligand Binding Assays. Compound 1 (Fig. 1) represents one of a structural class of potent ₄₁ antagonists (Hagmann et al., 2001). To determine whether compound 1 could also block ligand binding to ₄₇, the ability of this compound to inhibit binding of ¹²⁵I-hMAdCAM-Ig to RPMI-8866 cells in the presence of the divalent cation Mn²⁺ was evaluated using methods described previously (Egger et al., 2002). RPMI-8866 cells, a human B cell line, were chosen for the assay, because they express high levels of ₄₇ (60,000 ₄₇ receptors/cell), but low levels of ₄₁ (4,000 ₄₁ receptors/cell), as demonstrated by quantitative flow cytometry (Fig. 2A). In addition, the specificity of MAdCAM-Ig binding to ₄₇ on the RPMI-8866 cells has been confirmed by demonstrating that anti-₄ and anti-₇ mAbs block binding, whereas an anti-₁ mAb does not block binding (Egger et al., 2002). Furthermore, the low levels of ₄₁ expressed on the RPMI-8866 cell line do not bind VCAM-Ig in the presence of anti-₇ mAbs and1mMMn²⁺ or1mMCa²⁺/Mg²⁺ (data not shown). Thus, RPMI-8866 cells can be used to evaluate the potency of compounds in blocking binding of ₄₇, but not ₄₁, to MAdCAM-Ig. Compound 1 inhibited the binding of MAdCAM-Ig to ₄₇ with an IC₅₀ of 1.1 nM (Table 1).

View larger version (19K): [in this window] [in a new window]   Fig. 2. Surface expression of 41 and 47 on RPMI-8866 and Jurkat cell lines. Quantitative FACS analysis was used to determine the density of antibody binding sites for 41 and 47 expressed on the surface of the RPMI-8866 human B cell line (A) and the Jurkat human T cell line (B) as described under Materials and Methods. The data represent the mean number of receptors expressed per cell (rec/cell) for at least two independent experiments for each cell type. The isotype control, anti-4, anti-1, and anti-7 mAb-treated cell profiles are represented by the gray, black, dashed, and dotted lines, respectively.

View this table: [in this window] [in a new window]   TABLE 1 Activity of antagonists of 41/47 in ligand binding assays Compounds were tested using a 10-point titration in assays that measure the binding of either 125I-MAdCAM-Ig, 125I-VCAM-Ig, or 35S-compound 1 to cells expressing 4 to determine compound potency and specificity. Assays for activated 47 measured the binding of either 125I-MAdCAM-Ig to RPMI-8866 cells in the presence of 1 mM Mn2+ or the binding of 35S-compound 1 to RPMI-8866 cells in the presence of 1 mM Mn2+ and 100 nM BIO7662. Similarly, assays for activated 41 measured the binding of either 125I-VCAM-Ig or 35S-compound 1 to Jurkat cells in the presence of 1 mM Mn2+, whereas an unactivated 41 assay measured the binding of 35S-compound 1 to Jurkat cells in the presence of 1 mM Ca2+/Mg2+. Values represent the average of at least two independent experiments, and all values were within 95% confidence limits.

To compare the potency of compound 1 for blockade of ₄₁ under similar conditions, the ability of the compound to inhibit ¹²⁵I-hVCAM-Ig binding to Jurkat cells expressing ₄₁ in the presence of the divalent cation Mn²⁺ was performed as described previously (Egger et al., 2002). For this assay, Jurkat cells, a human T cell line, were chosen, because they express high levels of ₄₁ (90,000 ₄₁ receptors/cell), but low levels of ₄₇ (7,000 ₄₇ receptors/cell) (Fig. 2B). Specificity of VCAM-Ig binding to ₄₁ on Jurkat cells, was confirmed previously using anti-₄ and anti-₁ mAbs to completely abrogate binding, in the absence of inhibition by anti-₇ mAb (Egger et al., 2002). The low level of ₄₇ expressed on the Jurkat cell line does not bind MAdCAM-Ig in the presence of 1 mM Mn²⁺ or 1 mM Ca²⁺/Mg²⁺ (data not shown), indicating that Jurkat cells can be used to evaluate the potency of compounds in blocking binding of ₄₁, but not ₄₇, to VCAM-Ig. Compound 1 inhibited the binding of VCAM-Ig to ₄₁ with an IC₅₀ of 0.10 nM (Table 1). Thus, compound 1 is a potent dual ₄₁/₄₇ antagonist.

Divalent Cation-Dependent Binding of a Radiolabeled ₄₁/₄₇ Antagonist Probe, ³⁵S-compound 1, to RPMI-8866 Cells, K562/₄₇ Cells, and HUT-78 Cells. ₄-ligand interactions are dependent on divalent cations, which modulate the affinity state of the integrins for their ligands (Leitinger et al., 2000). To test the divalent cation dependence of binding, an assay was developed to measure the binding of ³⁵S-compound 1 to RPMI-8866 cells expressing activated or unactivated ₄₇ in the presence of the divalent cations Mn²⁺ or Ca²⁺/Mg²⁺, respectively (Fig. 3). In the presence of 1 mM Mn²⁺, RPMI-8866 cells bound ³⁵S-compound 1 with a ratio of specific to nonspecific binding of 12 (Fig. 3A). In the presence of 1 mM Ca²⁺/Mg²⁺, ³⁵S-compound 1 bound RPMI-8866 cells with a ratio of specific to nonspecific binding of 6. Specific binding was not observed in the presence of either 1 mM Ca²⁺ or 1 mM Mg²⁺ alone (data not shown). The inclusion of 10 mM EDTA in the binding reaction abrogated specific binding, as expected.

View larger version (28K): [in this window] [in a new window]   Fig. 3. Effect of divalent cations on the binding of 35S-compound 1 to RPMI-8866 cells. A, binding of 35S-compound 1 to RPMI-8866 cells was measured by pretreating cells with binding buffer containing one of the following: 1 mM Mn2+, 1 mM Ca2+ and 1 mM Mg2+, 1 mM Mn2+ plus 10 mM EDTA, or 1 mM Ca2+ and 1 mM Mg2+ plus 10 mM EDTA, or the corresponding binding buffer containing 100 nM BIO7662. Total binding was measured as described under Materials and Methods in the absence () or presence () of BIO7662, and addition of 1 µM unlabeled compound 1 was used to define NSB for cells treated with binding buffer alone () or binding buffer containing BIO7662 (). B and C, binding of 35S-compound 1 to K562/47 cells (B) or HUT-78 cells (C) was measured by pretreating cells with binding buffer containing either 1 mM Mn2+ or 1 mM Ca2+ and 1 mM Mg2+ in the absence () or presence () of 100 nM BIO7662. Total binding was determined as described above, and treatment with 1 µM unlabeled compound 1 was used to define NSB for cells treated with binding buffer alone () or binding buffer containing BIO7662 (). Statistics were generated with Prism by using a one-way analysis of variance (nonparametric test) with a Tukey's post test, and *, P < 0.001 when comparing total binding for each condition with the matched nonspecific binding control. The data represent the mean ± S.E.M. calculated from a total of four replicates from at least two independent experiments.

To determine whether the low level of ₄₁ present on RPMI-8866 cells (Fig. 2A) contributed to the observed binding of ³⁵S-compound 1, binding was measured with or without 100 nM BIO7662 added to specifically block ₄₁ (Fig. 3A). BIO7662, a highly selective inhibitor of ₄₁ (Chen et al., 2001; Pepinsky et al., 2002; Leone et al., 2003), was used for these studies because none of the available anti-₄ or anti-₁ neutralizing monoclonal antibodies blocked the binding of ³⁵S-compound 1 to ₄₁ even at concentrations up to 3.3 µg/ml (data not shown). BIO7662 (100 nM) was selected to completely saturate ₄₁ (IC₅₀ of 0.03 nM in the ₄₁/¹²⁵IVCAM-Ig binding assay; Table 1) without interfering with ₄₇ (IC₅₀ of 3.34 µM in the ₄₇/¹²⁵I-MAdCAM-Ig binding assay; Table 1). In the presence of 1 mM Mn²⁺ and 100 nM BIO7662, RPMI-8866 cells bound ³⁵S-compound 1 (input counts 150 pM) with a ratio of specific to nonspecific binding of 8 (Fig. 3A), and binding was reduced by 29% compared with cells not treated with BIO7662, indicating that, as expected, most of the total counts bound are due to binding to ₄₇. Total binding in the presence of 1 mM Ca²⁺/Mg²⁺ and BIO7662 was not significantly different from background binding, suggesting that low levels of unactivated ₄₁ expressed on RPMI-8866 cells are responsible for the binding of ³⁵S-compound 1 under these conditions. Comparison of the specific binding of ³⁵S-compound 1 to RPMI-8866 cells measured in the presence of 1 mM Ca²⁺/Mg²⁺ with or without BIO7662 was significantly different with a P value < 0.01. Furthermore, when incubating RPMI-8866 cells with 0 to 30 nM ³⁵S-compound 1 in the presence of 1 mM Ca²⁺/Mg²⁺ and 100 nM BIO7662, no significant binding above background was observed (data not shown). The results shown in Fig. 3A indicate that ₄₇ on RPMI-8866 cells requires Mn²⁺ to support binding of ³⁵S-compound 1 and that the unactivated state of the receptor does not support binding.

To determine whether the inability of ³⁵S-compound 1 to bind unactivated ₄₇ was dependent on cell type, we measured binding to K562 cells stably transfected with ₄₇ (K562/₄₇) in the presence of different divalent cations. As demonstrated by quantitative FACS analysis, K562/₄₇ cells express 200,000 copies/cell of ₄₁ and 200,000 copies/cell of ₄₇ (data not shown). In the presence of 1 mM Mn²⁺ and 100 nM BIO7662, K562/₄₇ cells bound ³⁵S-compound 1 with a 19-fold ratio of specific to nonspecific binding, and binding was reduced by 63% compared with cells not treated with BIO7662 (Fig. 3B). Thus, the binding of ³⁵S-compound 1 to activated K562/₄₇ cells is mediated by both ₄₁ and ₄₇, with each integrin having a similar contribution to the total binding. In the presence of 1 mM Ca²⁺/Mg²⁺, ³⁵S-compound 1 bound K562/₄₇ cells with a ratio of specific to nonspecific binding of 31-fold (Fig. 3B), but pretreating the cells with 100 nM BIO7662 abrogated binding (Fig. 3B), indicating that K562/₄₇ cell binding to ³⁵S-compound 1 is mediated solely by unactivated ₄₁. To extend the above-mentioned findings to another cell line, the effect of divalent cations on the binding of ³⁵S-compound to HUT-78 cells was evaluated. By FACS analysis, HUT-78 cells express equal proportions of both ₄₁ and ₄₇ (data not shown), and the relative receptor density of ₄₇ was similar to levels observed on RPMI-8866 cells (Erle et al., 1994). Mn²⁺-activated HUT-78 cells bound to ³⁵S-compound 1 with a 10- or 2-fold ratio of specific to nonspecific binding in the absence or presence of 100 nM BIO7662, respectively (Fig. 3C). In the presence of 1 mM Ca²⁺/Mg²⁺, ³⁵S-compound 1 bound unactivated HUT-78 cells with a ratio of specific to nonspecific binding of 9-fold, but failed to bind unactivated cells after pretreatment with 100 nM BIO7662 (Fig. 3C), indicating that unactivated ₄₁ is also completely responsible for HUT-78 cell binding to ³⁵S-compound 1. Thus, the binding properties of ³⁵S-compound 1 for different activation states of ₄ are not dependent on cell type or expression levels of ₄₁ relative to ₄₇, but rather are dependent on the activation state of the integrins.

Binding Kinetics of ³⁵S-Compound 1 to RPMI-8866 Cells. After three cell lines expressing both ₄₁ and ₄₇ were evaluated, subsequent studies on ₄₇ focused on characterizing the binding of ³⁵S-compound 1 to RPMI-8866 cells that predominantly express ₄₇. To determine whether an assay could be developed to study both the association and dissociation of unlabeled compounds from binding to activated ₄₇, equilibrium and kinetic studies for the binding of ³⁵S-compound 1 to ₄₇ were performed in the presence of Mn²⁺ with cells pretreated with 100 nM BIO7662 (Fig. 4). Binding of ³⁵S-compound 1 was dose-dependent and saturable with maximal binding to ₄₇ observed at approximately 10 nM. Assuming that each receptor bound one ligand, specific counts bound at saturation provided a direct measure of ₄₇ expression levels. Based on the calculation of B_max values, the RPMI-8866 cells used in these studies had approximately 43,000 copies of ₄₇/cell, which is in good agreement with the receptor density determined by quantitative FACS analysis (Fig. 2A).

View larger version (16K): [in this window] [in a new window]   Fig. 4. Equilibrium binding of 35S-compound 1 to RPMI-8866 cells in the presence of Mn2+. RPMI-8866 cells were incubated with increasing concentrations of 35S-compound 1 (0-30 nM) in binding buffer containing 1 mM Mn2+ and 100 nM BIO7662. Specific binding (cpm) was measured as described under Materials and Methods, and nonspecific binding was determined in the presence of 1 µM unlabeled compound 1. Bmax values (receptors/cell) were calculated by nonlinear regression analysis by using a one-site binding model (R2 value of 0.98). The data shown are representative of at least two independent experiments containing duplicate samples in each experiment.

The quadratic shape of the binding curve in Fig. 4 suggested that the apparent K_D for binding was small compared with the concentration of ₄₇ (360 pM) present in the binding assay, which is consistent with reports that measure binding of small molecule antagonists to ₄₁ (Chen et al., 1999, 2001). Based on this observation, it was evident that the saturation binding curves measured titration of the receptor to full occupancy, and a kinetic assessment of affinity was used to measure actual affinity constants. In the presence of 1 mM Mn²⁺ and 100 nM BIO7662, binding occurred rapidly with a k_on of 0.05 nM^-1 min^-1, reached equilibrium within 10 min, and reversed with a k_off of 0.09 min^-1 (Fig. 5). Thus, a K_D value of 1.9 nM was calculated based on the binding kinetics of compound 1 to activated ₄₇.

View larger version (18K): [in this window] [in a new window]   Fig. 5. On and off rates for 35S-compound 1 binding to RPMI-8866 cells in the presence of Mn2+. RPMI-8866 cells were incubated for the indicated times with 6.5 nM 35S-compound 1 and 5 nM unlabeled compound 1 in binding buffer containing 1 mM Mn2+ and 100 nM BIO7662. Specific binding (cpm) for compound association () or dissociation () was measured as described under Materials and Methods, with nonspecific binding determined in the presence of 1 µM unlabeled compound 1. kobs was determined at a ligand concentration of 11.5 nM by fitting the data to a one-phase exponential association equation (R2 value of 0.92). The indicated on rate (kon), off rate (koff) and KD value for the binding of 35S-compound 1 to 47 on RPMI-8866 cells were determined from the kinetic measurements by fitting the data to a one-phase exponential association or decay equation (R2 value of 0.99 for each curve). The data shown are representative of at least two independent experiments containing duplicate samples in each experiment.

Effect of Divalent Cation Concentration on the Binding of ³⁵S-Compound 1 to RPMI-8866 Cells. To elucidate the role of divalent cations on integrin activation, the effects of metal ion concentration on the binding of the ³⁵S-compound 1 to RPMI-8866 cells were examined. Cells were pretreated with 100 nM BIO7662 to block ₄₁, and binding was measured as a function of changing Mn²⁺ or Mg²⁺ concentrations from 0.03 to 100 mM (Fig. 6A). Mn²⁺ enhanced the binding of the ³⁵S-compound 1 to RPMI-8866 cells with an EC₅₀ of 0.5 mM, whereas Mg²⁺ only enhanced binding at higher concentrations (3-fold increase in specific binding at 5 mM; Fig. 6A).

View larger version (27K): [in this window] [in a new window]   Fig. 6. Binding of 35S-compound 1 to RPMI-8866 cells under different divalent cation conditions. For all experiments, RPMI-8866 cells were pretreated with 100 nM BIO7662 and incubated at room temperature for 45 min with 110 pM of 35S-compound 1. A, cells were incubated in the presence of increasing concentrations of Mn2+ () or Mg2+ (). Specific binding (cpm) was measured as described under Materials and Methods, and nonspecific binding was determined in the presence of 1 µM unlabeled compound 1. B, cells were incubated in the presence of 5 mM Mn2+ () or 5 mM Mg2+ () and increasing concentrations of Ca2+, and specific counts bound were determined as described above. C, cells were incubated in the presence of 0.2 mM Mn2+ (), 1 mM Mn2+ (), or 5 mM Mn2+ () and increasing concentrations of Ca2+. D, a double reciprocal plot of the data shown in C. By linear regression analysis, the three lines intersect approximately on the x-axis, indicative of noncompetitive inhibition. R2 values are 0.97 (5 mM Mn2+ with Ca2+), 0.99 (1 mM Mn2+ with Ca2+) and 0.99 (0.2 mM Mn2+ with Ca2+). All data shown are representative of at least two independent experiments containing duplicate samples in each experiment.

To determine whether Ca²⁺ can affect the ability of Mg²⁺ or Mn²⁺ to activate ₄₇, RPMI-8866 cells were treated with 100 nM BIO7662 and increasing concentrations of Ca²⁺ (0.04-5 mM) in combination with 5 mM Mn²⁺ or5mMMg²⁺, to achieve maximal activation and partial activation, respectively. In the presence of 5 mM Mn²⁺, Ca²⁺ at 1.25 mM and 5 mM inhibited the binding of ³⁵S-compound 1 to RPMI-8866 cells by 35 and 65%, respectively (Fig. 6B). Similarly, when combined with 5 mM Mg²⁺,Ca²⁺ at 1.25 and 5 mM inhibited the binding of ³⁵S-compound 1 to RPMI-8866 cells by 38 and 51%, respectively.

To further explore the interaction between Mn²⁺ and Ca²⁺, the binding of ³⁵S-compound 1 to ₄₇ was measured in the presence of increasing concentrations of Ca²⁺ (0.1-100 mM) in combination with three different fixed concentrations of Mn²⁺ (0.2, 1, and 5 mM). Ca²⁺ inhibited the binding of ³⁵S-compound 1 to RPMI-8866 cells with IC₅₀ values of 1.0, 4.0, and 14.0 mM, respectively. A double reciprocal plot of these data (Fig. 6D) indicated that the inhibition observed by Ca²⁺ in the presence of Mn²⁺ is noncompetitive in nature, and this lack of competition is between Ca²⁺ and Mn²⁺ rather than ligand.

Divalent Cation-Dependent Binding ³⁵S-Compound 1 to Jurkat Cells. To understand how divalent cations regulate the activation state of ₄₁, an assay was developed to measure the binding of ³⁵S-compound 1 to Jurkat cells expressing activated or unactivated ₄₁ in the presence of the divalent cations Mn²⁺ or Ca²⁺/Mg²⁺, respectively (Fig. 7). Binding of ³⁵S-compound 1 to Jurkat cells was dose-dependent and reached saturation at 5 nM. Because available anti-₄₁ mAb did not compete with the binding of ³⁵S-compound 1, the specificity of ³⁵S-compound 1 binding to ₄₁ was determined by direct competition with unlabeled compound 1 at 1 µM. Specific counts bound at saturation provided a direct measure of ₄₁ expression levels, and, based on this value, Jurkat cells used in these studies have approximately 64,000 copies (in Mn²⁺) and 57,000 copies (in Ca²⁺/Mg²⁺) of ₄₁/cell, which is in agreement with the receptor density determined by quantitative FACS analysis (Fig. 2B). Binding was dependent on the presence of divalent cations, and was blocked in the presence of 10 mM EDTA (data not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 7. Equilibrium binding of 35S-compound 1 to Jurkat cells expressing 41 in the presence of Mn2+ or Ca2+/Mg2+. Jurkat cells were incubated with increasing concentrations of 35S-compound 1 (0-54 nM) in binding buffer containing 1 mM Mn2+ (A) or 1 mM Ca2+/Mg2+ (B) and 100 nM BIO7662. Specific binding (cpm) was measured as described under Materials and Methods, with nonspecific binding determined in the presence of 1 µM unlabeled compound 1. Bmax values (receptors/cell) were calculated by nonlinear regression analysis by fitting the data to a one-site binding model (R2 value of 0.95 or 0.97 for A and B, respectively). The data shown are representative of at least two independent experiments.

As observed for the binding of ³⁵S-compound 1 to ₄₇, the apparent K_D for binding seemed to be small compared with the concentration of ₄₁ (350 and 320 pM, respectively) present in the binding assay (Fig. 7). The kinetic assessment of binding of compound 1 to ₄₁ in the presence of 1 mM Mn²⁺ indicated that binding occurred very slowly, reaching equilibrium by 60 min, and dissociation was not observed after another 60 min of incubation (Fig. 8A), or even after another 180 min (48,036 specific cpm bound after 240 min of total incubation time; data not shown). In the presence of 1 mM Ca²⁺/Mg²⁺, binding occurred with a k_on of 0.01 nM^-1 min^-1, reaching equilibrium within 10 min, and dissociation was rapid, with a k_off of 0.24 min^-1, resulting in a K_D of 18 nM (Fig. 8B). Because the saturation binding curves measured titration of the receptor to full occupancy, the K_D values obtained from the kinetic binding curves represent the actual binding affinity. Although binding of the protein ligand ¹²⁵I-VCAM-Ig requires an activated state of ₄₁, achieved by adding 1 mM Mn²⁺, ³⁵S-compound 1 is observed to bind both activated and unactivated states of ₄₁. Although similar association rates were observed for the binding of ³⁵S-compound 1 to unactivated and activated ₄₁, dissociation rates were dependent on the activation state of the integrin, with dissociation from the activated receptor indiscernible under the conditions of the assay.

View larger version (19K): [in this window] [in a new window]   Fig. 8. On and off rates for 35S-compound 1 binding to Jurkat cells in the presence of Mn2+ or Ca2+/Mg2+. A, Jurkat cells were incubated for the indicated times with 6.5 nM 35S-compound 1 and 5 nM unlabeled compound 1 in binding buffer containing 1 mM Mn2+ or 1 mM Ca2+/Mg2+. A, specific binding (cpm) for compound association () to or dissociation () from Mn2+-activated 41 was measured as described under Materials and Methods. B, specific binding (cpm) for compound association ()toor dissociation () from unactivated 41 was measured as outlined above and as described under Materials and Methods. kobs was determined at a ligand concentration of 11.5 nM, and the on rate (kon), off rate (koff) and KD values for the binding of 35S-compound 1 were determined from kinetic measurements by fitting the data to a one-phase exponential association or decay equation (R2 value of 0.86 and 0.99 for association and dissociation, respectively). The data shown are representative of at least two independent experiments containing duplicate values in each experiment.

Ability of Antagonists of ₄₁/₄₇ to Block the Binding of Native Ligand or ³⁵S-Compound 1 to RPMI-8866 or Jurkat Cells. After demonstrating that ³⁵S-compound 1 can be used to analyze ₄ interactions on both RPMI-8866 and Jurkat cell lines, we were interested in determining the potency of four ₄₁/₄₇ antagonists. Compounds were initially evaluated for their ability to block ¹²⁵I-MAdCAM-Ig binding to RPMI-8866 cells and ¹²⁵I-VCAM-Ig binding to Jurkat cells, in the presence of Mn²⁺ (Table 1). Compound 2, a recently described ₄₁ antagonist (Hagmann et al., 2001), and compound 3, a structurally related analog, inhibited binding to both ₄₁ and ₄₇ (Table 1). An inactive analog, compound 4 (Hagmann et al., 2001; Egger et al., 2002), did not inhibit binding when tested at concentrations up to 100 µM, demonstrating the importance of specific structural features to the activities of compounds 1, 2, and 3 (Kopka et al., 2002; Table 1). TR14035 (Fig. 1) has been reported to potently inhibit both ₄₁ and ₄₇ (Martin et al., 1999; Sircar et al., 1999a,1999b; Egger et al., 2002), and this was confirmed (Table 1).

After evaluating the potency of compounds in conventional ₄-ligand binding assays, we determined the potency of the same four compounds to block ³⁵S-compound 1 binding to RPMI-8866 cells in the presence of Mn²⁺ and 100 nM BIO7662. The concentration of ³⁵S-compound 1 used for the binding assay was maintained at less than 150 pM, based on an IC₅₀ of 400 pM for competition by unlabeled compound 1 (Table 1), and the results for the compounds tested are shown in Table 1. Although a small (5-fold or less) shift toward reduced potency in the ³⁵S-compound 1/₄₇ assay was observed, a similar rank order of compounds (cmpd 1 TR14035 > cmpd 2 > cmpd 3 >> cmpd 4) was maintained for the IC₅₀ values when the ³⁵S-compound 1/₄₇ and ¹²⁵IMAdCAM-Ig/₄₇ binding assays were compared.

The ability of the same four compounds to block ³⁵S-compound 1 binding to Jurkat cells expressing activated ₄₁ in the presence of Mn²⁺ was also measured. The concentration of ³⁵S-compound 1 used for the binding assay was maintained at less than 150 pM, based on an IC₅₀ of 2.5 and 2.3 nM for competition by unlabeled compound 1 with activated and unactivated Jurkat cells, respectively (Table 1), and the potency of compounds tested is shown in Table 1. Despite the overall shift toward reduced potency (26- to 79-fold less potent) in the ³⁵S-compound 1/₄₁ assay, the same rank order of compound IC₅₀ values (cmpd 1 = cmpd 2 = TR14035 > cmpd 3 >> cmpd 4) was observed for the ³⁵S-compound 1/₄₁ and ¹²⁵I-VCAM-Ig/₄₁ binding assays.

The same four compounds were titrated for inhibition of ³⁵S-compound 1 binding to Jurkat cells expressing unactivated ₄₁ in the presence of Ca²⁺/Mg²⁺. As described in Table 1, the activities of compounds 2 and 3 were within 2-fold of the values observed in the ³⁵S-compound 1/activated ₄₁ binding assay in the presence of Mn²⁺. As expected, compound 4 did not block binding of

³⁵S-compound 1 to unactivated ₄₁ at concentrations up to 1 mM (Table 1). Nonselective compounds, those that have relatively equal potency for activated and unactivated ₄₁, would be expected to follow the same rank order in both assays. Using this criterion, compounds 2 and 3 were identified as dual antagonists that are nonselective for ₄₁, whereas TR14035 had a 17-fold greater potency for inhibiting binding to activated ₄₁ than for unactivated ₄₁, making it a dual antagonist that is selective for activated ₄₁.

	Discussion

Top Abstract Materials and Methods Results Discussion References

 
³⁵S-Compound 1 was used as a model ligand to study divergent ₄₇- and ₄₁-ligand interactions in the presence of different metal cations. ³⁵S-Compound 1 bound both unactivated and activated states of ₄₁ that occur in suspension in the presence of Ca²⁺/Mg²⁺ or Mn²⁺, respectively, but only bound activated ₄₇. Binding kinetics revealed that ³⁵S-compound 1 dissociated from activated ₄₇ and unactivated ₄₁, but failed to dissociate from activated ₄₁ in the time frame observed. Although radiolabeled probes have been used to study the function of ₄₁-ligand interactions (Chen et al., 1999, 2001), this is the first report that a dual ₄₁/₄₇ antagonist can be used to elucidate the effect of divalent cations on both ₄₇- and ₄₁-ligand interactions, defining distinct binding properties of the probe for each integrin.

Integrins are known to exist in multiple affinity states in the presence of different divalent cations (Shimaoka et al., 2002). For many integrin-ligand interactions, including those of ₁,