The humanized anti-α4 integrin Ab Natalizumab is an effective treatment for relapsing-remitting multiple sclerosis. Natalizumab is thought to exert its therapeutic efficacy by blocking the α4 integrin-mediated binding of circulating immune cells to the blood-brain barrier (BBB). As α4 integrins control other immunological processes, natalizumab may, however, execute its beneficial effects elsewhere. By means of intravital microscopy we demonstrate that natalizumab specifically inhibits the firm adhesion but not the rolling or capture of human T cells on the inflamed BBB in mice with acute experimental autoimmune encephalomyelitis (EAE). The efficiency of natalizumab to block T cell adhesion to the inflamed BBB was found to be more effective in EAE than in acute systemic TNF-α-induced inflammation. Our data demonstrate that α4 integrin-mediated adhesion of human T cells to the inflamed BBB during EAE is efficiently blocked by natalizumab and thus provide the first direct in vivo proof of concept of this therapy in multiple sclerosis.

In multiple sclerosis (MS)3 and in its animal model experimental autoimmune encephalomyelitis (EAE), circulating immune cells get access to the CNS where they start the molecular events leading to inflammation, edema formation, and demyelination, all of which set the ground for the development of the disabling clinical picture of the disease. Interaction of circulating immune cells with the endothelial blood-brain barrier (BBB) is thus a critical step in the pathogenesis of EAE and MS. α4β1 integrin was identified as mediating T cell adhesion to inflamed vessels in frozen sections of EAE brains in vitro (1), and Abs blocking α4 integrins were shown in a variety of animal models to prevent the development of EAE (summarized in Ref. 2). Based on these findings, the humanized monoclonal anti-α4 integrin Ab natalizumab was developed for the treatment of MS with the idea of targeting α4 integrin-mediated extravasation of inflammatory cells into the CNS. In clinical trials, natalizumab proved to be highly beneficial in reducing MS disease activity regarding both clinical parameters as well as magnetic resonance intensity measurements of disease activity (3, 4). The rare risk of progressive multifocal leukoencephalopathy ( = 260) and an increased rate of herpes recrudescence in patients receiving natalizumab (natalizumab packaging insert; Biogen Idec) suggest however, that the Ab may have broader immunosuppressive effects. In fact, besides mediating T cell extravasation, α4 integrins have been demonstrated in animal models to be involved in multiple immune cell functions including T cell activation and polarization (5), retention of memory T cells in their niches (6), and the localization and maturation of hematopoietic stem cells (7). Thus, better understanding of to what extent the drug natalizumab prevents extravasation of circulating human T cells into the CNS in MS patients would greatly improve our understanding of how natalizumab may exert its immunosuppressive effects.

To directly visualize the effect of natalizumab on the interaction of human T cells with the inflamed BBB in vivo, we investigated human T cell interaction with the inflamed spinal cord white matter microcirculation in SJL mice with EAE by intravital fluorescence videomicroscopy. Human T cells were found to initiate contact by rolling or capture and to firmly adhere to the inflamed spinal cord microvasculature in vivo. Whereas natalizumab almost completely blocked the firm adhesion of human T cells to the BBB, it surprisingly left T cell rolling and capture unaffected. Effective blocking of T cell adhesion to the inflamed BBB was seen in mice with EAE as a model for the neuroinflammatory setting of MS. In contrast, natalizumab only partially and transiently reduced T cell interaction with TNF-α-stimulated spinal cord microvessels in SJL mice. Our study provides the first direct in vivo proof of concept that natalizumab does inhibit T cell extravasation into the CNS by specifically inhibiting T cell adhesion to the inflamed BBB.

Female SJL mice were obtained from Harlan and used at the age of 12 wk. Animal experiments were performed in accordance with the requirements of the local government. EAE was induced by immunization with proteolipid protein aa 139–151 exactly as described (8). Experiments were performed with mice suffering from clinical EAE at days 10–11 postimmunization with clinical scores of 0.5 (limp tail) and 1 (hind leg paraparesis).

Human recombinant TNF-α was a gift from D. Maennel (University of Regensburg, Regensburg, Germany) and natalizumab was from S. Goelz (Biogen Idec). Human IgG4 κ Ab was purchased from Sigma-Aldrich.

Binding of soluble mouse and human VCAM-1-Fc fusion proteins (R&D Systems) to Jurkat T cells was investigated by FACS analysis exactly as described (9). Briefly, 2 × 106 Jurkat T cells were incubated for 30 min at room temperature in PBS, 0.1% BSA, 0.2 mM Mn2+ with increasing concentrations of soluble recombinant human or mouse VCAM-1/Fc in the presence or absence of the mouse anti-human α4 integrin Ab 21/6, the mouse anti-human CD3 (BD Pharmingen), or the mouse anti-human CD18 (Ancell). After washing, FITC-labeled goat anti-human IgG (Vector Laboratories) was applied and fluorescence staining was analyzed on a FACScalibur flow cytometer (BD Biosciences).

Jurkat T cells were repeatedly sorted for high α4 integrin expression. PBMCs were isolated from buffy coats of healthy donors obtained from a blood bank (Inselspital, Bern, Switzerland) and cultured overnight in RPMI 1640 and 20% FCS. Preparation of PBMCs one day before the assay was necessary due to time constraints that did not allow us to perform all of the experimental procedures within one working day. The protocol used has been approved by the Bernese Cantonal Ethical Review Board (KEK No. 17/05). T cells were purified from PBMCs by negative magnetic selection according to the manufacturer’s protocol (Pan T cell isolation kit II from Miltenyi Biotec), fluorescently labeled with 125 nM calcein-AM (Molecular Probes), and incubated with natalizumab or human IgG4 at 170 μg per 4 × 106 T cells in 300 μl of physiological saline (0.9% (v/w) NaCl) 20 min prior to T cell infusion and injection with the cells. Natalizumab binding to human T cells is shown in supplemental Fig. 2.4 The natalizumab dosage used was deduced from the dosage of natalizumab in MS patients. MS patients receive natalizumab once a month as an infusion of 300 mg of injection per patient irrespective of the patient’s body weight. Dosage adjustment for mice was therefore based on blood volume; considering a blood volume of 4–5 liters for humans and ∼2. 5 ml for a SJL mouse of 25 g, the corresponding amount of natalizumab is between 120 and 187 μg per mouse. Based on the average blood volume of mice used in the study, the injection of 170 μg per mouse was chosen.

Surgical preparations, intravital microscopy by epi-illumination techniques, and quantitative analysis of the spinal cord microcirculation was performed exactly as described (10, 11, 12) using a custom-made Mikron IVM-500 fluorescence microscope connected to a silicon-intensified target camera (Dage-MTI). Spinal cord microvasculature was visualized with 1% tetramethylrhodamine isothiocyanate (TRITC)-conjugated dextran (0.1 ml of TRITC-dextran, m.w. = 155,000; Sigma-Aldrich). Systemic injection of three separate 100-μl aliquots of fluorescently labeled T cells was achieved by infusion via a catheter placed into the right carotid artery below a ligation, in the direction of the aortic arch. Observations were made using × 4, ×10, and × 20 long-distance working objectives. Real time microscopic images were recorded using a digital videocassette recorder (DSR-11; Sony) for later off-line analysis, which was performed exactly as described (10, 13). Similarly, hemodynamic parameters were analyzed and calculated based on Hagen-Poiseuille equations precisely as described (10).

Mann-Whitney U statistics were used for comparisons between different data sets. Asterisks indicate significant differences (∗, p < 0.05; ∗∗, p < 0.01; and ∗∗∗, p < 0.005). For the analysis of the firm adhesion of T cells, quantitative data are given as mean values ± SD calculated from the values in each animal. For analysis of differences between the groups, a Mann-Whitney U test was performed.

Inhibition of α4 integrin-mediated leukocyte trafficking across the inflamed BBB is thought to be the underlying concept of the therapeutic effect of natalizumab in reducing inflammatory and clinical signs in MS. However, direct in vivo evidence for natalizumab-mediated inhibition of leukocyte trafficking across the BBB is still missing. Imaging techniques allowing study of the interaction of individual circulating white blood cells with the CNS microvascular wall in the brain or spinal cord of humans are not available to date. Therefore, it is desirable to investigate immune cell interaction with the inflamed BBB in an MS-like neuroinflammatory setting in a model organism. In this study we specifically investigated the effect of natalizumab on the interaction of human T cells with the inflamed BBB, as we have recently shown that T cells but not myeloid cells critically rely on β1 integrins to accumulate in the CNS during EAE, suggesting that T cells are the main target of the anti-VLA-4 therapy (12).

It is well established that the structural requirements for the binding of human and mouse α4β1 integrin to mouse VCAM-1 are the same involving a cluster of amino acids in domain 1 of mouse VCAM-1, which is identical with the α4β1-binding site in human VCAM-1 (14). Based on these previous observations, we compared the binding of soluble mouse and human recombinant VCAM-1-Fc fusion proteins to human Jurkat T cells expressing high α4β1 integrin levels by FACS analysis. Increasing concentrations of mouse and human soluble VCAM-1 were found to bind with similar affinities to Jurkat T cells in an α4β1 integrin-dependent manner (Fig. 1 A).

The comparability of human α4β1 integrin-mediated binding to human and mouse VCAM-1 and the observation that circulating human T cells can extravasate across the inflamed BBB into the CNS parenchyma in mice suffering from EAE (Fig. 1, B and C) prompted us to perform live imaging of the interaction of human T cells with the BBB in a mouse model of MS to study the effect of natalizumab. By performing intravital fluorescence videomicroscopy of the spinal cord white matter microvasculature in SJL mice with EAE, we were able to directly visualize the multistep interaction of human T cells with the inflamed BBB under physiological shear forces in vivo in a neuroinflammatory setting modeling MS (10).

After visualization of the vascular system by the injection of TRITC-dextran, fluorescently labeled human T cells were infused and could readily be observed passing through the inflamed spinal cord white matter microvessels (supplemental video 1). An average of 22% of the total number of control IgG4-pretreated human T cells observed to pass through a given microvessel over a period of 1 min made initial contact with the inflamed spinal cord microvessels. This initial contact of human T cells with the inflamed BBB was characterized by T cell rolling along the vascular wall with reduced velocity or by transient capture, i.e., the abrupt stop of the T cells on the vessel wall for up to 7 s (Fig. 2,a and supplemental video 1A). Surprisingly, natalizumab did not alter the G protein-independent rolling or capture of human T cells to the inflamed BBB during EAE as compared with control T cells pretreated with human IgG4, indicating that α4 integrins are not dominantly involved in mediating the initial contact of T cells with the inflamed BBB in MS in vivo (Fig. 2,a and supplemental video 2A). This observation is in apparent contrast to our previous observation made under noninflammatory conditions, where the α4 integrin was observed to mediate the G protein-independent capture of freshly activated encephalitogenic T cell blasts to low levels of constitutively expressed VCAM-1 in the noninflamed spinal cord microvasculature in healthy mice (10). Thus, up-regulation of endothelial adhesion molecules other than VCAM-1 on the inflamed BBB in EAE seems to allow for α4 integrin-independent initial contact of human T cells with the inflamed microvascular wall. In accordance with this, we recently demonstrated that β1 integrin-deficient T cells are able to capture and roll on inflamed CNS microvessels during EAE (12). Because P-selectin is found to be up-regulated in inflamed CNS microvessels during EAE (8) and can mediate T cell rolling via P-selectin glycoprotein ligand-1 (15, 16), the inhibition of T cell rolling in inflamed CNS microvessels during EAE might require the simultaneous blockage of several adhesion mechanisms. Natalizumab was, however, found to strongly reduce the firm adhesion of human T cells to the BBB in EAE by 70% as determined 10 min after cell infusion compared with controls (supplemental videos 1B and 2B). This potent natalizumab-mediated reduction of human T cell adhesion to the inflamed spinal cord microvascular wall was sustained 30, 60, and 120 min after T cell infusion (Fig. 2 b), providing evidence that in the presence of α4 integrin-blocking natalizumab, T cells are unable to maintain G protein-dependent, integrin-mediated firm contact with the inflamed BBB in vivo. Our observations therefore provide the first live visual proof that human T cells use α4 integrins to bind to the inflamed BBB in vivo and that natalizumab blocks T cell entry into the CNS by specifically inhibiting the firm adhesion of human T cells to the BBB in a neuroinflammatory setting modeling MS.

To define whether natalizumab generally inhibits the adhesion of human T cells to the inflamed BBB in vivo, we next modeled an acute systemic inflammation by i.v. application of TNF-α in SJL mice 4 h before an investigation of human T cell interaction with the acutely stimulated spinal cord white matter microvascular wall. Blood vessels were again visualized by TRITC-dextran. Upon infusion, fluorescently labeled human T cells were found to initiate contact with the inflamed spinal cord microvascular wall by rolling and capture, similarly as observed in EAE (Fig. 3,a and supplemental video 3A). Interestingly, under acute TNF-α-stimulated conditions, natalizumab significantly reduced the percentage of T cells rolling along the activated spinal cord white matter microvascular wall but did not inhibit T cell capture (supplemental video 4A). Taken together, natalizumab therefore again failed to significantly inhibit the initial contact of human T cells to the TNF-α stimulated BBB, confirming that under inflammatory conditions α4 integrins are not required to initiate the interaction of human T cells with the BBB in vivo. In contrast to the observations made in EAE, natalizumab only partially and transiently (at 10 min but no longer at 30 or 60 min after infusion) reduced the firm adhesion of human T cells to the TNF-α-stimulated BBB in vivo, suggesting that the drug is less efficient in inhibiting T cell interaction with the acutely inflamed BBB (Fig. 3 b and supplemental videos 3B and 4B). The different inhibitory capacity of natalizumab in EAE vs TNF-α induced inflammatory conditions may be due to different α4 integrin ligand densities expressed on the BBB upon stimulation with TNF-α vs during the complex neuroinflammatory situation in EAE. Additionally, differences in hemodynamic flow parameters in the spinal cord microvasculature in SJL mice with EAE vs TNF-α injected SJL mice may have an impact on natalizumab-mediated inhibition of T cell interaction with the BBB. Although neither inflammatory condition by itself produced significantly different hemodynamic flow parameters compared with those determined by us in the spinal cord white matter microcirculation of healthy SJL mice previously (10), direct comparison of the hemodynamic flow parameters between the spinal cord microvasculature under EAE or TNF-α-stimulated conditions revealed significant differences (supplemental Fig. 3). Specifically, the mean velocity of circulating T cells, wall shear rates, and wall shear stress were found to be significantly lower in spinal cord microvessels during EAE when directly compared with those in mice injected with TNF-α (supplemental Fig. 3). These findings suggest that lower shear forces may favor α4 integrin-mediated T cell adhesion to the inflamed CNS microvasculature, and thus natalizumab can efficiently block this step of the multistep T cell migration cascade across the inflamed BBB during EAE.

Taken together, our data provide the first direct in vivo proof of concept of natalizumab-mediated inhibition of stable adhesion of human T cells to the inflamed BBB in a mouse model of MS. Surprisingly, natalizumab-mediated inhibition of T cell adhesion to the inflamed BBB was found to be efficient in EAE modeling the neuroinflammation in MS, but not in acute systemic inflammation induced by TNF-α. Thus, our data suggest that the clinical efficacy of anti-α4 integrin therapy in MS is at least in part due to the inhibition of α4 integrin-mediated T cell adhesion (and surprisingly not T cell rolling or capture) to the BBB, thus preventing T cell entry into the CNS.

We owe special thanks to Heidi Tardent for excellent microsurgical preparations and to Jens Stein for patient advice about performing and analyzing the intravital fluorescence videomicroscopy experiments. We thank Biogen Idec for the gift of natalizumab and Frédérique Bard and Ted Yednock (Elan Pharmaceuticals) for very fruitful discussions.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


This work has been supported by the National Multiple Sclerosis Society of the United States, the Swiss Multiple Sclerosis Society, and the Bern University Research Foundation. C.C. has been funded by the Fondation pour la Recherche Médicale, the Association pour la Recherche sur la Sclérose en Plaques, and the National Multiple Sclerosis Society.


Abbreviations used in this paper: MS, multiple sclerosis; BBB, blood-brain barrier; EAE, experimental autoimmune encephalomyelitis; TRITC, tetramethylrhodamine.


The online version of this article contains supplemental material.

Yednock, T. A., C. Cannon, L. C. Fritz, F. Sanchez-Madrid, L. Steinman, N. Karin.
. Prevention of experimental autoimmune encephalomyelitis by antibodies against α4β1 integrin.
Engelhardt, B., R. M. Ransohoff.
. The ins and outs of T-lymphocyte trafficking to the CNS: anatomical sites and molecular mechanisms.
Trends Immunol.
Polman, C. H., P. W. O'Connor, E. Havrdova, M. Hutchinson, L. Kappos, D. H. Miller, J. T. Phillips, F. D. Lublin, G. Giovannoni, A. Wajgt, et al
. A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis.
N. Engl. J. Med.
Rudick, R. A., W. H. Stuart, P. A. Calabresi, C. Confavreux, S. L. Galetta, E. W. Radue, F. D. Lublin, B. Weinstock-Guttman, D. R. Wynn, F. Lynn, et al
. Natalizumab plus interferon β1a for relapsing multiple sclerosis.
N. Engl. J. Med.
Mittelbrunn, M., A. Molina, M. M. Escribese, M. Yanez-Mo, E. Escudero, A. Ursa, R. Tejedor, F. Mampaso, F. Sanchez-Madrid.
. VLA-4 integrin concentrates at the peripheral supramolecular activation complex of the immune synapse and drives T helper 1 responses.
Proc. Natl. Acad. Sci. USA
Sixt, M., M. Bauer, T. Lammermann, R. Fassler.
. β1 integrins: zip codes and signaling relay for blood cells.
Curr. Opin. Cell Biol.
Bungartz, G., S. Stiller, M. Bauer, W. Muller, A. Schippers, N. Wagner, R. Fassler, C. Brakebusch.
. Adult murine hematopoiesis can proceed without β1 and β7 integrins.
Döring, A., M. Wild, D. Vestweber, U. Deutsch, B. Engelhardt.
. E- and P-selectin are not required for the development of experimental autoimmune encephalomyelitis in C57BL/6 and SJL Mice.
J. Immunol.
Yednock, T. A., C. Cannon, C. Vandevert, E. G. Goldbach, G. Shaw, D. K. Ellis, C. Liaw, L. C. Fritz, L. I. Tanner.
. α4β1 integrin-dependent cell adhesion is regulated by a low affinity receptor pool that is conformationally responsive to ligand.
J. Biol. Chem.
Vajkoczy, P., M. Laschinger, B. Engelhardt.
. α4-integrin-VCAM-1 binding mediates G protein-independent capture of encephalitogenic T cell blasts to CNS white matter microvessels.
J. Clin. Invest.
Engelhardt, B., P. Vajkoczy, M. Laschinger.
. Detection of endothelial/lymphocyte interaction in spinal cord microvasculature by intravital videomicroscopy.
Methods Mol. Med.
Bauer, M., C. Brakebusch, C. Coisne, M. Sixt, H. Wekerle, B. Engelhardt, R. Fassler.
. β1 integrins differentially control extravasation of inflammatory cell subsets into the CNS during autoimmunity.
Proc. Natl. Acad. Sci. USA
Stein, J. V., G. Cheng, B. M. Stockton, B. P. Fors, E. C. Butcher, U. H. von Andrian.
. L-selectin-mediated leukocyte adhesion in vivo: microvillous distribution determines tethering efficiency, but not rolling velocity.
J. Exp. Med.
Renz, M. E., H. H. Chiu, S. Jones, J. Fox, K. J. Kim, L. G. Presta, S. Fong.
. Structural requirements for adhesion of soluble recombinant murine vascular cell adhesion molecule-1 to α4β1.
J. Cell Biol.
Kerfoot, S. M., M. U. Norman, B. M. Lapointe, C. S. Bonder, L. Zbytnuik, P. Kubes.
. Reevaluation of P-selectin and α4 integrin as targets for the treatment of experimental autoimmune encephalomyelitis.
J. Immunol.
Battistini, L., L. Piccio, B. Rossi, S. Bach, S. Galgani, C. Gasperini, L. Ottoboni, D. Ciabini, M. D. Caramia, G. Bernardi, et al
. CD8+ T cells from patients with acute multiple sclerosis display selective increase of adhesiveness in brain venules: a critical role for P-selectin glycoprotein ligand-1.

Supplementary data