Laminin-411 is a vascular ligand for MCAM and facilitates TH17 cell entry into the CNS

Abstract

TH17 cells enter tissues to facilitate pathogenic autoimmune responses, including multiple sclerosis (MS). However, the adhesion molecules involved in the unique migratory capacity of TH17 cells, into both inflamed and uninflamed tissues remain unclear. Herein, we characterize MCAM (CD146) as an adhesion molecule that defines human TH17 cells in the circulation; following in vitro restimulation of human memory T cells, nearly all of the capacity to secrete IL-17 is contained within the population of cells expressing MCAM. Furthermore, we identify the MCAM ligand as laminin 411, an isoform of laminin expressed within the vascular endothelial basement membranes under inflammatory as well as homeotstatic conditions. Purified MCAM-Fc binds to laminin 411 with an affinity of 27 nM, and recognizes vascular basement membranes in mouse and human tissue. MCAM-Fc binding was undetectable in tissue from mice with targeted deletion of laminin 411, indicating that laminin 411 is a major tissue ligand for MCAM. An anti-MCAM monoclonal antibody, selected for inhibition of laminin binding, as well as soluble MCAM-Fc, inhibited T cell adhesion to laminin 411 in vitro. When administered in vivo, the antibody reduced TH17 cell infiltration into the CNS and ameliorated disease in an animal model of MS. Our data suggest that MCAM and laminin 411 interact to facilitate TH17 cell entry into tissues and promote inflammation.

Figure 1. hMCAM identifies a small population of human memory T cells with the majority of the capacity to secrete IL-17.

(A) hMCAM is expressed on about 3–6% of circulating human CD4+ T cells, nearly exclusively in the CD45RO+ memory T cell pool. CD4+CD45RO+ memory T cells from five individual donors were further sorted into hMCAM− and hMCAM+ populations and stimulated for four days in the presence of anti-CD3 and anti-CD28 and supernatants were analyzed for IL-17. Each individual donor (n = 5) is represented by a different color (* indicates p<0.05). (B) Freshly isolated human CD4+ T cells were stained for hMCAM and coexpression of chemokine receptors, CCR6 and CCR7. (C) Human CD4+CD45RO+ T cells were sorted based on hMCAM expression, and stimulated with anti-CD3 and anti-CD28 in the absence of exogenous cytokines. Cells were collected after five days, following PMA/Ionomcycin stimulation with golgi inhibition for the final five hours. Cells were stained for intracellular IL17 and IFNγ. (D) Cells were sorted for hMCAM expression and stimulated as in C. They were then restained for MCAM expression, as well as intracellular IL17.

Figure 2. hMCAM+ T cells expand in response to cytokine stimulation, and specifically produce the majority of TH17 cytokines.

(A) Human CD4+CD45RO+ T cells were stimulated with anti-CD3 and anti-CD28 in the presence of the indicated cytokine(s). Cells were collected after five days, following PMA/Ionomcycin stimulation with golgi inhibition for the final five hours. Cells were stained for surface hMCAM and the percent hMCAM+ is shown (n = 4). Cells, were also stained for intracellular IFNγ (B) IL-17 (C) IL-22 (D) CCL20 (E) following cytokine stimulations and the percent cytokine positive within either the hMCAM− or hMCAM+ cell population is shown (n = 4 for each cytokine analysis). Data is representative of at least five individual donors. * indicates p<0.05.

Figure 3. hMCAM binds to a ligand in the ECM with identical staining to laminin α4.

Calcein labeled, hMCAM expressing MOLT 4 cells were preincubated with either isotype control (A) or anti-hMCAM (clone 17) (B) followed by incubation on tissue sections from healthy mice. After gentle washing of unbound cells, and mounting with DAPI, bound cells were visualized by fluorescent microscopy. Healthy mouse brain sections containing choroid plexus were stained with fluorescently labeled mMCAM-Fc protein and pan-laminin antibody. Staining of mMCAM-Fc was detected on choroid plexus (C) as well as the vasculature throughout the tissues (D). Fluorescently labeled mMCAM-Fc protein was preincubated with either isotype control (E) or anti-mMCAM (clone 15) (F) before addition to tissue sections of healthy mouse brain. Healthy mouse tissues were stained with fluorescently labeled mMCAM-Fc and CD31 (G) anti-mMCAM and pan laminin (H) or anti-mMCAM alone (I). Tissues from mice with active EAE were stained with fluorescently labeled mMCAM-Fc and pan-laminin (J) or mMCAM-Fc and an antibody specific to the α4 chain of laminin (K).

Figure 4. mMCAM colocalizes with laminin 411 on the choroid plexus, and shows no specific binding to tissues from LAMA4−/− mice.

(A) Staining of healthy mouse choroid plexus with anti-laminin, anti-laminin α4 and mMCAM-Fc (B) Staining of healthy brain tissues from LAMA4−/− mice with anti-laminin and mMCAM-Fc.

Figure 5. Laminin 411 functions as a ligand for MCAM.

(A) CHO cells transfected with hMCAM were incubated with recombinant laminin 411. Laminin binding to the surface of the cells was detected with an anti-laminin antibody (left panel), while no binding was detected to CHO cells lacking MCAM expression (center panel). Recombinant laminin 511 did not bind to MCAM expressing CHO cells, and binding of laminin 411 was specifically inhibited by preincubation of the cells with anti-hMCAM (clone 17, right panel). (B) Human CD4+CD45RO+ memory T cells were purified and incubated with TCR stimulation in the presence of TGFβ and IL1β to induce hMCAM expression (approximately 50% expressed hMCAM, similar to Figure 2A). Cells were incubated in plates coated with either laminin 411 or laminin 511 in the presence of either hMCAM-Fc or anti-hMCAM and specific binding of the cells to the laminin was determined. * indicates p<0.05. (C) Binding of recombinant human laminin 411 to immobilized hMCAM-Fc protein was detected by Biacore with a Kd of approximately 27 nM. Data is representative of three individual experiments. (D) mMCAM is expressed at high levels in nearly all CHO cells transfected with the mMCAM gene, and sorted for high expressors (blue histogram compared to shaded isotype control). However, soluble mMCAM protein does not bind to these cells at levels higher than either human IgG control or irrelevant Fc tagged protein, CTLA-4. Furthermore, no specific binding of mMCAM-His to mMCAM-Fc was detected by Biacore. (E) Laminin α4 was detected in both human MS lesions and normal appearing white matter.

Figure 6. Mouse T cells express MCAM following TH17 polarization, and blockade of MCAM influences disease progression in EAE.

(A) A population of PLP specific T cells was generated in vivo by immunization of SJL mice with PLP in CFA. Splenocytes from immunized mice were restimulated in vitro with PLP in the presence of the indicated cytokine(s), and mMCAM expression on CD4+ T cells was determined. (B) SJL mice were immunized with PLP/CFA as described. Two days after onset of disease, (between days 12 and 14 post immunization), and daily thereafter, the animals received either anti-mMCAM (clone 15) neutralizing antibody or isotype control antibody. Disease progression was scored, and body weights were monitored. * indicates p<0.05 by Wilcoxon's non-parametric test. Data represents the mean of 15 mice ± sem. Quantification of infiltrating mMCAM+ cells (C) from EAE induced mice treated with either isotype control or anti-mMCAM. **** indicates P<0.0001. Sections were scored as described in Methods S1. (D) T cells were isolated from the CNS of mice treated with either isotype control or anti-mMCAM and analyzed by flow cytometry for expression of CD4 and mMCAM. Graph indicates the percentage of CD4+ cells that are mMCAM+ in either treatment group. Each dot represents the percentage of mMCAM+ cells within the CD4+ T cell population in a single mouse. * indicates p<0.05.