Visualization of chemokine binding sites on human brain microvessels

Abstract

The chemokines monocyte chemoattractant protein-1 (MCP-1) and macrophage inflammatory protein-1alpha (MIP-1alpha) aid in directing leukocytes to specific locales within the brain and spinal cord during central nervous system inflammation. However, it remains unclear how these chemokines exert their actions across a vascular barrier, raising speculation that interaction with endothelial cells might be required. Therefore, experiments were performed to determine whether binding domains for these chemokines exist along the outer surface of brain microvessels, a feature that could potentially relay chemokine signals from brain to blood. Using a biotinylated chemokine binding assay with confocal microscopy and three-dimensional image reconstruction, spatially resolved binding sites for MCP-1 and MIP-alpha around human brain microvessels were revealed for the first time. Binding of labeled MCP-1 and MIP-1alpha could be inhibited by unlabeled homologous but not heterologous chemokine, and was independent of the presence of heparan sulfate, laminin, or collagen in the subendothelial matrix. This is the first evidence of specific and separate binding domains for MCP-1 and MIP-1alpha on the parenchymal surface of microvessels, and highlights the prospect that specific interactions of chemokines with microvascular elements influence the extent and course of central nervous system inflammation.

Figure 1

MCP-1 and MIP-1α binding to human brain microvessels. Microvessels purified from human brain cortex were reacted with either biotinylated MCP-1 (a) or biotinylated MIP-1α (b), and then visualized with avidin-fluorescein. Negative controls for microvascular staining included microvessels coincubated with MCP-1 (c) or biotinylated MIP-1α (d) and their corresponding, specific antibodies. Positive controls for detection of chemokine binding sites show PBM, known to possess receptors for MCP-1 and MIP-1α, reactive with biotinylated MCP-1 (e) and biotinylated MIP-1α (f). Negative controls for PBM staining included biotinylated MCP-1 (g) and biotinylated MIP-1α (h) along with their respective antibodies. Bars, 40 μm.

Figure 2

Topological distributions of MCP-1 and MIP-1α binding along human brain microvessels. Microvessels were first reacted with biotinylated chemokines/avidin-fluorescein, fixed, then subsequently immunoreacted with monoclonal anti–Factor VIII antibody/rhodamine anti–mouse IgG, to detect endothelial cells. Samples were viewed by confocal microscopy, z-series were obtained, and three-dimensional images were reconstructed as described in Materials and Methods. MCP-1 (a–d) and MIP-1α (e–h) staining are indicated in green, and immunostained endothelial cells appear red. Images in a–d and e–h represent different orientations of a given microvessel, displaying a nearly 360° distribution of the binding sites for the two chemokines. Bars, 40 μm. Images in d and h represent portions of b and e, respectively, that have been enlarged by computer.

Figure 4

Saturable binding of MCP-1 and MIP-1α along brain microvessels. Total labeled chemokine binding at increasing biotinylated chemokine concentrations was assessed as described in Materials and Methods. Nonspecific binding was similarly determined in the presence of 50-fold excess of unlabeled chemokine, and subtracted from total binding values to give specific binding (mean pixel intensity).

Figure 3

Competition of MCP-1 and MIP-1α binding to brain microvessels. Microvessels were incubated with a constant concentration of either biotinylated MCP-1 or biotinylated MIP-1α in the presence of increasing concentrations of unlabeled chemokine (indicated as x-fold greater than biotinylated chemokine). Relative chemokine binding was determined, and corrected for nonspecific binding as described in Materials and Methods. Values are plotted as percentages of maximal binding obtained in the absence of unlabeled competitor ± standard error. (a) Biotinylated MCP-1 plus unlabeled MCP-1; (b) biotinylated MCP-1 plus unlabeled MIP-1α; (c) biotinylated MIP-1α plus unlabeled MIP-1α; and (d) biotinylated MIP-1α plus unlabeled MCP-1.

Figure 5

Codistribution of MCP-1 and MIP-1α binding with perivascular markers along brain microvessels. Microvessels were incubated with biotinylated chemokines (either MCP-1, left column, or MCP-1α, right column) and the indicated antibodies (top to bottom). Chemokine binding is displayed in green, and antibody reactions are in red. Insets show three-dimensional reconstructions, obtained as described in Materials and Methods. For orientation purposes, asterisks indicate the direction looking into the lumen of each microvessel fragment. Anti-CD68 specifically stains monocyte-derived macrophages. Arrowheads denote staining of the individual matrix components, which appear outside the domains of chemokine binding. Bars, 40 μm.

Figure 6

Heparinase I treatment of human brain microvessels. Freshly isolated brain microvessels were exposed to heparinase I treatment and then subjected to biotinylated chemokine binding (MCP-1 and MIP-1α) and immunofluorescent detection of heparan sulfate as described in Materials and Methods. Green fluorescence indicates chemokine binding and red fluorescence indicates heparan sulfate distribution. Boxed areas at the top right are three-dimensional renderings of optical cross sections through individual microvessels, looking into the vessel lumen (denoted by asterisks). (a and b) Chemokine binding to control microvessels not treated with enzyme; (c and d) chemokine binding to heparinase I–treated samples. Despite nearly complete removal of perivascular heparan sulfate by pretreatment of microvessels with heparinase I, binding of both MCP-1 and MIP-1α is still readily detected. Bars, 40 μm.

Figure 7

Chemokine binding to heparinase I–treated human brain microvessels. Brain microvessels were treated as described in Fig. 5, and quantitatively evaluated for their extent of chemokine binding ± standard error (as detailed in Materials and Methods). **P < 0.001; *P < 0.05.

Figure 8

Chemokine receptor expression in human brain microvessels. Brain microvessels were immunostained with anti-CCR1, anti-CCR2, anti-CCR5, and isotype control antibody (IC), all followed by fluorescein-conjugated secondary antibody, or with just secondary antibody alone as a negative control (NC). Bars, 80 μm.