Intercellular Adhesion Molecule-1 Expression Is Required on Multiple ...

The Journal of Immunology

Intercellular Adhesion Molecule-1 Expression Is Required on Multiple Cell Types for the Development of Experimental Autoimmune Encephalomyelitis1 Daniel C. Bullard,* Xianzhen Hu,† Trenton R. Schoeb,* Robert G. Collins,§ Arthur L. Beaudet,¶ and Scott R. Barnum2†‡ Many members of the Ig superfamily of adhesion molecules, such as ICAM-1 and VCAM-1, have been implicated in the pathogenesis of multiple sclerosis. Although it is well-established that VCAM-1/VLA-4 interactions can play important roles in mediating CNS inflammatory events in multiple sclerosis patients and during the development of experimental allergic encephalomyelitis (EAE), the contributions of ICAM-1 are poorly understood. This is due in large part to conflicting results from Ab inhibition studies and the observation of exacerbated EAE in ICAM-1 mutant mice that express a restricted set of ICAM-1 isoforms. To determine ICAM-1-mediated mechanisms in EAE, we analyzed ICAM-1 null mutant mice (ICAM-1null), which express no ICAM-1 isoforms. ICAM-1null mice had significantly attenuated EAE characterized by markedly reduced spinal cord T cell infiltration and IFN-␥ production by these cells. Adoptive transfer of Ag-restimulated T cells from wild-type to ICAM-1null mice or transfer of ICAM-1null Ag-restimulated T cells to control mice failed to induce EAE. ICAM-1null T cells also showed reduced proliferative capacity and substantially reduced levels of IFN-␥, TNF-␣, IL-4, IL-10, and IL-12 compared with that of control T cells following myelin oligodendrocyte glycoprotein 35–55 restimulation in vitro. Our results indicate that ICAM-1 expression is critical on T cells and other cell types for the development of demyelinating disease and suggest that expression of VCAM-1 and other adhesion molecules cannot fully compensate for the loss of ICAM-1 during EAE development. The Journal of Immunology, 2007, 178: 851– 857.

M

ultiple sclerosis (MS)3 is considered primarily a T cell-mediated autoimmune disease, with self-reactivity directed against several myelin-derived Ags, including myelin basic protein and myelin oligodendrocyte glycoprotein (MOG). Many other cell types, such as macrophages, mast cells, glial cells, and ␥␦ T cells, as well as various blood-borne and membrane-anchored effector molecules (e.g., Ab, complement, cytokines/chemokines, and others) also contribute to MS pathogenesis and inflammation (1–5). A hallmark feature of MS is the trafficking of Ag-specific T cells and macrophages into the CNS, where they initiate inflammation and destruction of oligodendrocytes and neurons. The movement of these inflammatory cells into the CNS is regulated by leukocyte/endothelial cell adhesion pro-

*Department of Genetics, †Department of Microbiology, and ‡Department of Neurology, University of Alabama at Birmingham, Birmingham, AL 35294; and §Department of Pediatrics and ¶Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX 77030 Received for publication July 7, 2006. Accepted for publication October 26, 2006. 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. 1 This work was supported by National Multiple Sclerosis Society Grant RG-3437A-6 and National Institutes of Health Grants NS46032 (to S.R.B.) and RR017009 (to D.C.B.). 2 Address correspondence and reprint requests to Dr. Scott R. Barnum, Department of Microbiology, University of Alabama at Birmingham, 845 19th Street South, Bevill Biomedical Research Building 842, Birmingham, AL 35294. E-mail address: [email protected] 3 Abbreviations used in this paper: MS, multiple sclerosis; MOG, myelin oligodendrocyte glycoprotein; EAE, experimental autoimmune encephalomyelitis; CDI, cumulative disease index; LFB, Luxol fast blue.

Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 www.jimmunol.org

teins and chemoattractant/activating molecules. Studies using CNS inflammatory model systems, including the MS model experimental autoimmune encephalomyelitis (EAE), strongly suggest that the adhesion molecules VLA-4 and its ligand VCAM-1, as well as LFA-1, play an integral part in this process, while the E- and P-selectin interactions with P-selectin glycoprotein ligand 1 do not significantly contribute or serve redundant functions in promoting leukocyte adhesion in the CNS microvasculature (reviewed in Refs. 6 and 7). However, the roles of other adhesion molecules, especially ICAM-1, remain controversial. ICAM-1 has long been implicated in the pathogenesis of MS and EAE. For example, increased expression of ICAM-1 has been shown on endothelial cells, microglia, and astrocytes in active MS or EAE lesions (8 –15). Several reports have also demonstrated, using in vitro adhesion assays, that ICAM-1 is important for T cell adhesion and transendothelial migration through cytokine-stimulated brain endothelial cell monolayers (16 –20). In addition, ICAM-1-dependent engagement and signaling in lymphocytes, blood-brain barrier endothelial cells, microglia, and astrocytes can activate effector mechanisms that promote the progression of neuroinflammation (10, 21). Finally, genetic analyses have provided evidence of an association between a single nucleotide polymorphism in the ICAM-1 gene and MS development in several populations (22–24). Previously published studies have varied widely with respect to the role of ICAM-1 in the initiation and progression of EAE. Most reports have described results using inhibitory ICAM-1 mAbs to block interactions of this adhesion molecule with its ligands. Although several of these studies have shown a protective outcome in active EAE (25–27), an almost equal number have reported no protection or increased severity of disease (28 –30). In contrast, loss, or inhibition of the ICAM-1 ligands LFA-1 and Mac-1 in

852 EAE models have, in most cases, shown reduction in clinical severity and partial protection against the development of CNS inflammation. Finally, Samoilova et al. (31) performed EAE using Icam1tm1Bay mutant mice, which do not express the full-length form of ICAM-1, but do express several of the alternatively spliced isoforms of ICAM-1. Surprisingly, these mice developed a more severe course of EAE in the chronic phase of the disease, suggesting that expression of one or more alternatively spliced forms of ICAM-1 may augment CNS inflammatory events in this model. We report here results of EAE studies using ICAM-1 null mutant mice (ICAM-1null), which are deficient in all ICAM-1 isoforms. Complete absence of expression of all forms of this adhesion molecule resulted in markedly attenuated disease with little to no cellular infiltration or demyelination compared with wild-type mice. In addition, adoptive transfer experiments demonstrated that ICAM-1 expression is required on multiple cell types for the development of EAE. These findings, combined with our observations that loss of ICAM-1 resulted in reduced T cell proliferation and cytokine production in ICAM-1null mice consistent with poor T cell priming, indicate that in the absence of ICAM-1, T cells are inadequately activated to induce disease.

Materials and Methods Mice Mice containing a null mutation for ICAM-1 were generated by gene targeting in 129/Sv-embryonic stem cells as described previously (32). Briefly, a targeting construct designed to delete the coding region of the ICAM-1 locus was made as shown in Fig. 1. This construct was electroporated into AB2.1 embryonic stem cells and transformed clones were selected based on resistance to unomycin. Genomic DNA was obtained from selected clones and homologous recombination was confirmed using restriction digestion with SpeI and XbaI followed by Southern blotting with the 5⬘ and 3⬘ probes. Two clones confirmed to carry the desired mutation were injected into C57BL/6 blastocytes and transferred into foster mothers. The ICAM-1 mutation was then backcrossed onto the C57BL/6 strain for at least seven generations (The Jackson Laboratory). ICAM-1null mice are phenotypically normal and require no extraordinary husbandry. Inbred C57BL/6 mice were used as controls for all experiments. All studies were performed with approval from the University of Alabama at Birmingham (UAB) Institutional Animal Care and Use Committee.

Induction of active and transferred EAE For active EAE, control and ICAM-1null mice were immunized with MOG peptide35–55 as described (33). MOG peptide was synthesized by standard 9-fluorenyl-methoxycarbonyl chemistry and was ⬎95% pure as determined by reversed-phase-HPLC (Biosynthesis). Onset and progression of EAE symptoms was monitored daily using a standard clinical scale ranging from 0 to 6 as follows: 0, asymptomatic; 1, loss of tail tone; 2, flaccid tail; 3, incomplete paralysis of one or two hind limbs; 4, complete hind limb paralysis; 5, moribund; 6, dead. Only mice with a score of at least 2 (flaccid tail) for ⬎2 consecutive days were judged to have onset of EAE. For each animal a cumulative disease index (CDI) was calculated from the sum of the daily clinical scores observed between days 7 and 30. For transferred EAE, spleens of control or ICAM-1⫺/⫺ donors were removed, 2–3 wk following induction of active EAE, and prepared as previously described (33). Briefly, cells were stimulated for 24 h with MOG peptide (20 ␮g/ml) in the presence of freshly irradiated naive splenic APCs. IL-2 (20 U/ml) was added and after an additional 24 h of culture, T cells were purified by gradient centrifugation using Ficoll. Passive EAE was induced by injecting ⬃5 ⫻ 106 purified T cells derived from wild-type mice into ICAM-1null mice or by injecting ⬃5 ⫻ 106 purified T cells derived from ICAM-1null mice into wild-type mice. In both cases, ⬃5 ⫻ 106 purified T cells derived from wild-type mice were injected in wild-type mice as a control to monitor disease development.

ICAM-1 DEFICIENCY INHIBITS EAE

FIGURE 1. Generation of ICAM-1null mice. A, The genomic locus of ICAM-1 is deleted from the BamHI site 5⬘ of exon 1 to the SalI site 3⬘ of the poly(A) site in exon 7. B, The locus after targeting has the ICAM-1 coding sequence replaced by a puromycin-resistance cassette (PURO) transcribed in the opposite direction. The lines under the genomic locus show the 3⬘ and 5⬘ external probes. Restriction enzymes: Sp, SpeI; B, BamHI; X, XbaI; S, SalI.

with Luxol fast blue for evaluation of demyelination. The extent of inflammation and demyelination was scored based on lesion size (0 – 4) and lesions were evaluated for lymphocyte accumulation, neutrophil infiltration, demyelination, axonal degeneration, and gliosis (0 – 4). Tissues were evaluated without identification as to experimental group. Severity scores were calculated as the mean over all segments of the products of the intensity scores multiplied by the extent scores for each lesion characteristic (inflammation, axonal degeneration, gliosis, and demyelination). The means of the individual lesion characteristic severity scores were summed to give the overall severity score.

Isolation and flow cytometric analysis of leukocytes from spinal cords Spinal cords were removed from control and ICAM-1null mice with active EAE (days 12–15) after perfusion with PBS, ground through a cell strainer, washed in PBS, resuspended in 40% Percoll, and layered on 70% Percoll. After centrifugation at 2000 rpm (room temperature, 25 min), cells at the interface were removed, washed in PBS, and stained as described. Cells obtained from spinal cords were incubated with anti-CD16/32 (24G2, FcR block) to prevent nonspecific staining. Spinal cord leukocytes were stained with anti-CD4-FITC (GK1[CR1].5), antiCD-8-PE (53-6.7), antiCD45-FITC (30F11), anti-IL-TNF-␣-PE (MP6-XT22), and antiIFN-␥-FITC (XMG1.2), all from eBioscience. Stained cells and forward scatter were analyzed using a FACSCalibur and the data were analyzed using CellQuest software (BD Biosciences).

T cell proliferation and cytokine and chemokine production Ag-specific T cell proliferation assays were performed as previously described (33). Single-cell suspensions from spleens obtained 14 days after EAE induction were cultured in 96-well plates at 5 ⫻ 105 cells/well with increasing concentrations of MOG35–55 peptide in triplicate. After 48 h, cultures were pulsed with [3H]thymidine for an additional 18 h and incorporation of thymidine was measured. The in vitro cytokine assays were performed essentially as described for the proliferation assay. Duplicate cultures were either left untreated or stimulated with MOG peptide alone (5 ␮g/ml). Culture supernatants were collected at 48 h for use in cytokine

Histopathology Mice with actively induced EAE were sacrificed at 13 and 25 days p.i. by CO2 inhalation, and spinal columns were removed, fixed in 10% buffered formalin, and paraffin embedded. Sections (5-␮m thick) from the cervical, thoracic, and lumbar spinal cord were cut and either stained with H&E for overall lesion evaluation and characterization of inflammatory responses or

FIGURE 2. The clinical course of active EAE is markedly attenuated in ICAM-1null mice. Active EAE was induced with MOG peptide 35–55 and signs of disease were scored for 30 days as described in Materials and Methods. Results shown are the daily mean clinical score for wild-type (n ⫽ 11) and ICAM-1null mice (n ⫽ 12) from three experiments.


853

Table I. EAE signs in wild-type mice and ICAM-1null mice

Wild type (n ⫽ 11) ICAM-1null (n ⫽ 12)

CDIa

Disease Onsetb

Disease Incidencec

58 11

16 days

100 75

a

CDI is the mean of the sum of daily clinical scores observed between days 7 and 30. Disease onset is defined as the first day of 2 consecutive days with a clinical score of two or more. c Disease incidence is defined as the percentage of mice that displayed any clinical signs of disease. b

ELISAs. ELISA kits for murine cytokines (IFN-␥, TNF-␣, IL-2, IL-4, IL10, IL-12, IL-17, and TGF-␤) were purchased from R&D Systems. Each assay was performed according to the manufacturer’s instructions. Cytokine production by cultures of wild-type and ICAM-1null cells is reported as the percent of wild-type cytokine production. The data are pooled from two separate experiments.

Statistics Statistical significance between control and ICAM-1null mice for active and transferred EAE experiments was calculated using the Mann-Whitney U test; for proliferation assays, the Student t test was used. Results of evaluations for inflammation and demyelination were analyzed using ANOVA for main effects and Tukey’s test for pairwise mean comparisons.

Results Generation of ICAM-1null mice ICAM-1null mice were generated as described in Materials and Methods (Fig. 1). To verify the ICAM-1null phenotype, genomic DNA from wild-type and ICAM-1null mice was digested with BamHI and mini Southern blots were probed with a mouse ICAM-1 cDNA fragment up to the SalI site in exon 7. Bands representing the expected DNA fragments were detected for wild

FIGURE 4. Leukocyte subsets in spinal cords of control and ICAM1null mice with EAE are markedly different. A, Leukocytes isolated from spinal cords of control (n ⫽ 5) and ICAM-1null mice (n ⫽ 5) as described in Materials and Methods were immunostained for CD45. The infiltration of CD45⫹ leukocytes at day 12 p.i. was markedly reduced in ICAM-1null mice compared with controls. B, Leukocytes isolated from spinal cords of control (n ⫽ 5) and ICAM-1null mice (n ⫽ 5) as described in Materials and Methods were immunostained for CD4 and CD8. The numbers represent a normalized percentage of CD4⫹ and CD8⫹ T cells infiltrating the spinal cord of ICAM-1null mice compared with control mice. The results shown are from cells pooled within each group of mice.

type; no bands were observed in samples from ICAM-1null mice (data not shown). To verify a lack of ICAM-1 protein expression in ICAM-1null mice, sections of lung were subjected to immunohistochemical staining using a rat polyclonal anti-ICAM-1 Ab and serum levels of soluble ICAM-1 were analyzed by ELISA. As expected, ICAM-1 protein was not detected in ICAM-1null mice by either method (data not shown). Deletion of ICAM-1 significantly attenuates active EAE

FIGURE 3. Leukocyte infiltration and demyelination are reduced in ICAM-1null mice in EAE. Spinal cords from wild-type and ICAM-1null mice were obtained at 25 days postimmunization, fixed in 10% buffered formalin, and paraffin embedded. Sections from the cervical, thoracic, and lumbar regions (5 ␮m) were stained with H&E or Luxol fast blue (LFB) and scored as described in Materials and Methods. A, Representative section from a wild-type mouse stained with H&E. Arrow indicates widespread cellular infiltration and inflammation. B, Section from the same specimen as in A stained with LFB. Arrows indicate regions of significant demyelination. C, Representative section from an ICAM-1null mouse stained with H&E. Note the lack of cellular infiltration and inflammation. D, Section from the same specimen as in C stained with LFB. Little to no demyelination was observed throughout the white matter. Original magnification, ⫻10.

To determine the role of ICAM-1 in EAE, we induced active EAE using MOG35–55 peptide. We scored wild-type and ICAM-1null mice for signs of disease for 30 days. Based on our scoring criteria, which requires that mice have a score of 2 or more for 2 consecutive days, none of the ICAM-1null mice developed EAE (Fig. 2, Table I). However, 75% of the ICAM-1null mice developed tail weakness that generally presented for a few days and then recovered. The CDI for ICAM-1null mice was markedly lower than that of control mice (11 vs 58, respectively, Table I) and the overall course of disease was significantly reduced ( p ⫽ 0.0018, MannWhitney U test). Cellular infiltration and demyelination in ICAM-1null mice with EAE We next performed histopathological analysis on spinal cords of wild-type and ICAM-1null mice with active EAE to determine the extent and nature of the cellular infiltrate and the amount of demyelination between the two groups of mice. Representative spinal cord sections from wild-type mice obtained at 25 days after disease

854


FIGURE 5. The clinical course of adoptively transferred EAE is markedly attenuated in ICAM-1null mice. A, Transferred EAE was induced in wild-type (n ⫽ 3) and ICAM-1null (n ⫽ 5) mice by injecting encephalitogenic T cells (5 ⫻ 106) derived from wild-type mice with active EAE as described in Materials and Methods. Results shown are the daily mean clinical score from two separate experiments. B, Transferred EAE was induced in wild-type (n ⫽ 7) mice by injecting T cells (5 ⫻ 106) derived from ICAM-1null mice induced for active EAE. As a control, transferred EAE was induced in wild-type mice (n ⫽ 5) by injecting encephalitogenic T cells (5 ⫻ 106) derived from wild-type mice with active EAE. Results shown are the daily mean clinical score from three separate experiments.

induction had extensive cellular infiltration in the meninges and white matter (Fig. 3A) with corresponding demyelination (Fig. 3B). Sections obtained from ICAM-1null mice had no cellular infiltration, inflammation, axonal degeneration, and demyelination throughout the spinal cord, compared with wild-type mice (Fig. 3, C and D). The overall score for these parameters for all regions of the spinal cord in ICAM-1null mice was zero while wild-type mice had a score of 15.8 ( p ⬍ 0.05). We also analyzed for leukocyte infiltration early in EAE development (days 12–15 p.i.) and observed that total CD45⫹ leukocyte infiltration in the spinal cords of ICAM-1null mice was markedly reduced (Fig. 4A). Furthermore, of the few leukocytes that did infiltrate the spinal cords of ICAM-1null mice, there were fewer CD4⫹ and CD8⫹ T cells compared with wild-type mice (Fig. 4B). CD4⫹ and CD8⫹ T cell infiltration was reduced 55 and 87%, respectively, in ICAM-1null compared with wild-type mice. However, when normalized for the differences in overall CNS infiltration between the two groups of mice, CD4⫹ and CD8⫹ T cell infiltration was reduced 98 and 99.6%, respectively, in ICAM-1null compared with wild-type mice. These data demonstrate that the absence of ICAM-1 significantly prevents trafficking of leukocytes to the CNS.

To determine whether ICAM-1 deficiency on Ag-specific T cells would fail to induce or result in attenuated disease, we performed transferred EAE using MOG-sensitized T cells from ICAM-1null mice. Wild-type mice receiving ICAM-1null T cells also failed to develop EAE. Only one mouse developed very mild disease, which resolved within a few days of onset. As expected, transfer of Ag-specific T cells from wild-type to wild-type mice as an internal control resulted in clinical signs of EAE within 6 days (Fig. 5B, Table II). ICAM-1null T cells have reduced proliferative capacity and an altered cytokine profile compared with wild-type T cells To test the possibility that attenuated active and transferred EAE in ICAM-1null mice could be due to a poor proliferative capacity of ICAM-1null T cells, we performed in vitro proliferation assays as previously described in a “criss-cross” fashion (33). Stimulation of MOG-sensitized T cells from wild-type mice with APCs from ICAM-1null mice with various concentrations of MOG, resulted in robust T cell proliferation (Fig. 6, f). In contrast, MOG-sensitized T cells from ICAM-1null mice stimulated with wild-type APCs

Transfer of wild-type MOG-sensitized T cells to ICAM-1null mice or ICAM-1null MOG-sensitized T cells to wild-type mice does not induce EAE We next induced EAE by adoptively transferring MOG-sensitized T cells from wild-type mice to ICAM-1null mice. To our surprise, we observed no development of disease in ICAM-1null recipient mice while control mice readily developed EAE (Fig. 5A, Table II). Table II. Transferred EAE signs in wild-type mice and ICAM-1null mice

WT ⬎ WT (n ⫽ 5) WT ⬎ ICAM-1null (n ⫽ 5) WT ⬎ WT (n ⫽ 5) ICAM-1null ⬎ WT (n ⫽ 7) a

CDIa

Disease Onsetb

57 3 53 1.4

8 days 14 days

Disease Incidencec

100 28 100 14

CDI is the mean of the sum of daily clinical scores observed between days 7 and 30. Disease onset is defined as the first day of 2 consecutive days with a clinical score of two or more. c Disease incidence is defined as the percent of mice that displayed any clinical signs of disease. b

FIGURE 6. ICAM-1 expression is required on T cells, but not APCs for T cell proliferation. T cells were enriched by nylon-wool adherence from the spleens of wild-type (n ⫽ 4) or ICAM-1null mice (n ⫽ 4) undergoing active EAE. Wild-type T cells were cocultured with irradiated ICAM-1null splenic APCs plus MOG peptide (0.125– 4 ␮g/ml) and ICAM-1null T cells were cocultured in an identical fashion except that wild-type APCs were used. The cells were pulsed with [3H]thymidine and harvested at 18 h for determination of radioisotope incorporation. The results shown are expressed as the mean ⫾ SEM of fold-induction of wild-type or ICAM-1null T cell proliferation relative to background proliferation.


855

Discussion

FIGURE 7. Splenic ICAM-1null T cells produce a unique repertoire of cytokines during EAE. A, Encephalitogenic T cells enriched by nylon-wool adherence from the spleens of wild-type (n ⫽ 4) or ICAM-1null mice (n ⫽ 4) undergoing active EAE, were cocultured with irradiated splenic APCs from naive donors and stimulated with MOG peptide (1 ␮g/well). Supernatants were collected 48 h after stimulation and assayed by ELISA to quantitate production of each cytokine. The mean production of each cytokine is presented as the percent change of wild-type level. ⴱ, Significant changes in the levels compared with wild-type mice (p ⬍ 0.05 in all cases). B, Production of IFN-␥ in CD4⫹ and CD8⫹ T cells isolated from the spinal cords of control and ICAM-1null mice with active EAE. Leukocytes isolated from spinal cords of control (n ⫽ 5) and ICAM-1null mice (n ⫽ 5) as described in Materials and Methods were immunostained for CD4, CD8, and IFN-␥. The data shown are derived from gating on IFN-␥-producing cells. The results shown are from cells pooled within each group of mice.

resulted in significantly reduced proliferation ( p ⫽ 0.015) compared with that of wild-type T cells and ICAM-1null APCs (Fig. 6, 䡺). MOG-sensitized ICAM-1null T cells proliferated poorly when paired with APCs isolated from ICAM-1null mice (data not shown). The severely attenuated EAE in ICAM-1null mice may also be consistent with a shift to a Th2 cytokine repertoire. Fig. 7A shows that splenic ICAM-1null T cells did not produce elevated levels of IL-4 or other anti-inflammatory cytokines, such as IL-10 or TGF-␤, compared with wild-type mice. IL-4 and IL-10 levels were ⬍20 and 50%, respectively, of that seen in wild-type mice. The levels of several Th1 and proinflammatory cytokines produced by ICAM-1null T cells, including IFN-␥, TNF-␣, and IL-12 were dramatically reduced (Fig. 7A). The amount of IL-2 and IL-17 produced by ICAM-1null T cells was also significantly reduced but to a lesser degree than observed for IFN-␥ or TNF-␣. We also examined for the production of IFN-␥ by CD4⫹ and CD8⫹ T cells in the spinal cord of wild-type and ICAM-1null mice during the acute phase of EAE development (15 days p.i.). Consistent with the low levels of IFN-␥ produced by splenic T cells in the same time period, we observed a substantially lower percentage of IFN-␥-producing CD4⫹ and CD8⫹ T cells in ICAM-1null mice (Fig. 7B).

The expression of ICAM-1 on endothelial cells, leukocytes, neurons, and glial cells positions this adhesion molecule as a significant contributor to many facets in the pathogenesis of EAE. The results of this study clearly demonstrate that ICAM-1 plays a central role in the development of demyelinating disease. Our findings contrast with several previous EAE studies using inhibitory antiICAM-1 Abs or Icam1tm1Bay mice (25–31). At least for the Ab studies, these contrasting findings may be due to a number of reasons, including differential Ab specificity, insufficient dosage, differences in the timing of Ab delivery, and methodological differences in the induction or characterization of EAE. The different EAE phenotypes observed in the Icam1tm1Bay and ICAM-1null mice may be due in part to the exclusive expression of alternatively spliced isoforms in Icam1tm1Bay mice. To date multiple isoforms of ICAM-1 have been described in mice and humans (34 – 39), and their functions, changes in expression, and relative expression on the various cell types involved in the development of EAE are unknown. All ICAM-1 isoforms have the binding site for LFA-1 (in the first Ig domain), but only three isoforms have the Mac1-binding site (in the third Ig domain) (34). This difference alone suggests that subsets of ICAM-1 isoforms may mediate unique and perhaps unknown functions. However, it is not clear how differential expression of ICAM-1 isoforms modulates the severity and phenotypic features of EAE. Our observations in both the active and adoptive transfer EAE models suggest that ICAM-1 expression on multiple cell types is necessary for the development of CNS inflammation and demyelination. Perhaps the most dramatic outcome of our studies was the severe inhibition of leukocyte emigration into the spinal cords of ICAM-1null mice both in the acute and chronic phase of the disease. By both histopathology and flow cytometry, we observed little to no infiltration, remarkable preservation of myelin and axons throughout the white matter of the spinal cord and no inflammation or gliosis. These findings suggest that ICAM-1 expression on CNS endothelium is necessary for efficient adhesion and recruitment of leukocytes during the development of EAE and that other endothelial-expressed adhesion molecules, such as VCAM-1, cannot compensate for the loss of ICAM-1 expression in ICAM-1null mice to mediate these adhesive steps. This may be due to ICAM-1 and VCAM-1 acting at different steps in the adhesion cascade. Support for this possibility comes from several previous studies of CNS endothelial cell adhesion events using in vitro flow chamber assays or intravital microscopy. These studies demonstrated that endothelial-expressed ICAM-1 is important for both firm adhesion and transendothelial migration of T cells while VCAM-1 is important for early adhesion events, but not transendothelial migration (reviewed in Refs. 6 and 7). Another model supported by our data is one in which ICAM-1 and VCAM-1 overlap at the firm adhesion step, but are not completely interchangeable or redundant. Thus, loss or inhibition of either adhesion molecule during EAE would lead to significant disease attenuation. The marked inhibition of EAE in ICAM-1null mice may also be due, in part, to reduced proliferation of MOG-specific T effector cells. We observed a significant reduction in proliferation of ICAM-1null T cells when stimulated with wild-type APCs (Fig. 6 and data not shown). This result indicates that ICAM-1 on T cells is also required as an important costimulatory signal. This finding is supported by our observation that wild-type T cells proliferate efficiently in the absence of ICAM-1 on APCs (Fig. 6). Our results do not exclude a role for ICAM-1 on APC-mediated costimulation, but rather raise the possibility that ICAM-1 expression of T cells may also be critical for modulating the effector T cell response. In

856 fact the severe EAE observed in Icam1tm1Bay mice (31) could be due, in part, to altered ICAM-1-mediated costimulation as a result of differential ICAM-1 isoform expression. We would argue that failed or altered T cell priming is the major contributor to the attenuated EAE phenotype we observed in ICAM-1null mice. Faulty T cell priming could lead to the atypical production of cytokines seen during EAE in ICAM-1null mice relative to control mice. ICAM-1null mice do not express a cytokine profile consistent with any current EAE cytokine paradigm. The levels of IFN-␥, TNF-␣, and IL-12 in ICAM-1null mice were 10 –25% of those seen in control mice while IL-2 and IL-17 levels were 60 –70% that of control mice (Fig. 7). Interestingly, cytokine production by splenic T cells in Icam1tm1Bay mice was somewhat different from that of ICAM-1null mice. Both IL-2 and IFN-␥ levels were reduced in both ICAM-1null and Icam1tm1Bay mice; however, IL-10 expression was markedly reduced in ICAM-1null mice but increased in Icam1tm1Bay mice relative to control mice. Given that IL-10 is thought to be involved in the homeostatic regulation of the autoreactive T cell repertoire (40 – 42), elevated IL-10 levels are not consistent with the severe disease phenotype reported in Icam1tm1Bay mice. IL-4 levels were also significantly reduced in ICAM-1null mice suggesting little to no generation of Th2 cells in these mice. This is an unusual finding since previous studies have shown that IL-4 is important for modulating EAE severity (43, 44). These data bring into question the value of correlating changes in splenic T cell cytokine production to disease phenotype, particularly when examining a very limited range of cytokines. The regulation of ICAM-1 expression by glial cells has been studied extensively, and it is well established that expression increases in CNS inflammatory diseases (reviewed in Refs. 10 and 45). The absence of ICAM-1 on glial cells during EAE would affect a myriad of effector functions including activation and retention of infiltrating leukocytes, phagocytosis through interaction with Mac-1 and p150/95 and, the generation of inflammatory mediators. Loss of ICAM-1 expression on astrocytes alone contributes, in part, to better outcome in spinal cord injury (46) and EAE (J. Hu, J. Bethea, and S. R. Barnum, unpublished observations). In ICAM-1null mice, ICAM-1-mediated functions on glial cells are undoubtedly prevented. However, we feel that the loss of ICAM-1 glial cell functions, are secondary to reduced leukocyte infiltration and poor T cell priming as a mechanistic explanation for attenuated disease in ICAM-1null mice. The results we report here demonstrate that ICAM-1 is essential for the development of EAE and we would argue that it plays a similar role in MS. Therapeutic approaches to demyelinating disease in animal models and humans have focused primarily on inhibition of ␣4 integrins and LFA-1 (26, 28, 47–50). Our results argue that inhibitory approaches blocking ICAM-1-mediated functions warrant significant investigation in MS. Support for this argument comes from a recent study in which EAE was inhibited by treatment with a Staphylococcus aureus, extracellular adherence protein which binds to ICAM-1 (51). Extracellular adherence protein functions as a bacterial anti-inflammatory molecule by blocking ICAM-1-␤2-integrin interactions and does so with greater potency than Abs to ICAM-1 or ␤2-integrin receptors (52). The caveat for an ICAM-1-based therapeutic approach, as well as other adhesion molecules, in demyelinating disease is that the host may be immunocompromised if ICAM-1 functions on all cell types are inhibited. Our results, however, suggest that targeting ICAM-1 functions on T cells alone may provide significant protection from demyelination.


Acknowledgments Histology services were provided by the UAB Comparative Pathology Laboratory. We thank Dr. Sherry Smith and Jillian Adams for help with the flow cytometry studies and Dr. Chander Raman for critical reading of the manuscript.

Disclosures The authors have no financial conflict of interest.

References 1. Sospedra, M., and R. Martin. 2005. Immunology of multiple sclerosis. Annu. Rev. Immunol. 23: 683–747. 2. Rajan, A. J., Y. L. Gao, C. S. Raine, and C. F. Brosnan. 1996. A pathogenic role for ␥␦ T cells in relapsing-remitting experimental allergic encephalomyelitis in the SJL mouse. J. Immunol. 157: 941–949. 3. Secor, V. H., W. E. Secor, C. A. Gutekunst, and M. A. Brown. 2000. Mast cells are essential for early onset and severe disease in a murine model of multiple sclerosis. J. Exp. Med. 191: 813– 822. 4. Hemmer, B., B. Kieseier, S. Cepok, and H. P. Hartung. 2003. New immunopathologic insights into multiple sclerosis. Curr. Neurol. Neurosci. Rep. 3: 246 –255. 5. Owens, T. 2003. The enigma of multiple sclerosis: inflammation and neurodegeneration cause heterogeneous dysfunction and damage. Curr. Opin. Neurol. 16: 259 –265. 6. Engelhardt, B. 2006. Immune cell migration across the blood-brain barrier: molecular mechanisms and therapeutic targeting. Future Neurol. 1: 47–56. 7. Engelhardt, B. 2006. Molecular mechanisms involved in T cell migration across the blood-brain barrier. J. Neural Transm. 113: 477– 485. 8. McMurray, R. W. 1996. Adhesion molecules in autoimmune disease. Semin. Arthritis Rheum. 25: 215–233. 9. Bo, L., J. W. Peterson, S. Mork, P. A. Hoffman, W. M. Gallatin, R. M. Ransohoff, and B. D. Trapp. 1996. Distribution of immunoglobulin superfamily members ICAM-1, -2, -3, and the ␤2 integrin LFA-1 in multiple sclerosis lesions. J. Neuropathol. Exp. Neurol. 55: 1060 –1072. 10. Lee, S. J., and E. N. Benveniste. 1999. Adhesion molecule expression and regulation on cells of the central nervous system. J. Neuroimmunol. 98: 77– 88. 11. Dopp, J. M., S. M. Breneman, and J. A. Olschowka. 1994. Expression of ICAM-1, VCAM-1, L-selectin, and leukosialin in the mouse central nervous system during the induction and remission stages of experimental allergic encephalomyelitis. J. Neuroimmunol. 54: 129 –144. 12. Matsuda, M., N. Tsukada, C. S. Koh, T. Iwahashi, K. Shimada, and N. Yanagisawa. 1994. Expression of intercellular adhesion molecule-1 and lymphocyte function-associated antigen-1 in the spinal cord of rats during acute experimental allergic encephalomyelitis. Autoimmunity 19: 15–22. 13. Gimsa, U., S. A. Wolf, D. Haas, I. Bechmann, and R. Nitsch. 2001. Th2 cells support intrinsic anti-inflammatory properties of the brain. J. Neuroimmunol. 119: 73– 80. 14. Sipkins, D. A., K. Gijbels, F. D. Tropper, M. Bednarski, K. C. Li, and L. Steinman. 2000. ICAM-1 expression in autoimmune encephalitis visualized using magnetic resonance imaging. J. Neuroimmunol. 104: 1–9. 15. Carrithers, M. D., I. Visintin, S. J. Kang, and C. A. Janeway, Jr. 2000. Differential adhesion molecule requirements for immune surveillance and inflammatory recruitment. Brain 123(Pt. 6): 1092–1101. 16. Reiss, Y., G. Hoch, U. Deutsch, and B. Engelhardt. 1998. T cell interaction with ICAM-1-deficient endothelium in vitro: essential role for ICAM-1 and ICAM-2 in transendothelial migration of T cells. Eur. J. Immunol. 28: 3086 –3099. 17. Reiss, Y., and B. Engelhardt. 1999. T cell interaction with ICAM-1-deficient endothelium in vitro: transendothelial migration of different T cell populations is mediated by endothelial ICAM-1 and ICAM-2. Int. Immunol. 11: 1527–1539. 18. Lyck, R., Y. Reiss, N. Gerwin, J. Greenwood, P. Adamson, and B. Engelhardt. 2003. T-cell interaction with ICAM-1/ICAM-2 double-deficient brain endothelium in vitro: the cytoplasmic tail of endothelial ICAM-1 is necessary for transendothelial migration of T cells. Blood 102: 3675–3683. 19. Greenwood, J., C. L. Amos, C. E. Walters, P. O. Couraud, R. Lyck, B. Engelhardt, and P. Adamson. 2003. Intracellular domain of brain endothelial intercellular adhesion molecule-1 is essential for T lymphocyte-mediated signaling and migration. J. Immunol. 171: 2099 –2108. 20. Lebedeva, T., M. L. Dustin, and Y. Sykulev. 2005. ICAM-1 co-stimulates target cells to facilitate antigen presentation. Curr. Opin. Immunol. 17: 251–258. 21. Greenwood, J., S. Etienne-Manneville, P. Adamson, and P. O. Couraud. 2002. Lymphocyte migration into the central nervous system: implication of ICAM-1 signalling at the blood-brain barrier. Vascul. Pharmacol. 38: 315–322. 22. Nejentsev, S., M. Laaksonen, P. J. Tienari, O. Fernandez, H. Cordell, J. Ruutiainen, J. Wikstrom, T. Pastinen, S. Kuokkanen, J. Hillert, and J. Ilonen. 2003. Intercellular adhesion molecule-1 K469E polymorphism: study of association with multiple sclerosis. Hum. Immunol. 64: 345–349. 23. Cournu-Rebeix, I., E. Genin, G. Lesca, A. Azoulay-Cayla, N. Tubridy, E. Noe, M. Clanet, G. Edan, F. Clerget-Darpoux, G. Semana, and B. Fontaine. 2003. Intercellular adhesion molecule-1: a protective haplotype against multiple sclerosis. Genes Immun. 4: 518 –523. 24. Mycko, M. P., M. Kwinkowski, E. Tronczynska, B. Szymanska, and K. W. Selmaj. 1998. Multiple sclerosis: the increased frequency of the ICAM-1 exon 6 gene point mutation genetic type K469. Ann. Neurol. 44: 70 –75.

The Journal of Immunology 25. Archelos, J. J., S. Jung, M. Maurer, M. Schmied, H. Lassmann, T. Tamatani, M. Miyasaka, K. V. Toyka, and H. P. Hartung. 1993. Inhibition of experimental autoimmune encephalomyelitis by an antibody to the intercellular adhesion molecule ICAM-1. Ann. Neurol. 34: 145–154. 26. Kawai, K., Y. Kobayashi, M. Shiratori, G. Sobue, T. Tamatani, M. Miyasaka, and Y. Yoshikai. 1996. Intrathecal administration of antibodies against LFA-1 and against ICAM-1 suppresses experimental allergic encephalomyelitis in rats. Cell. Immunol. 171: 262–268. 27. Morrissey, S. P., R. Deichmann, J. Syha, C. Simonis, U. Zettl, J. J. Archelos, S. Jung, H. Stodal, H. Lassmann, K. V. Toyka, et al. 1996. Partial inhibition of AT-EAE by an antibody to ICAM-1: clinico-histological and MRI studies. J. Neuroimmunol. 69: 85–93. 28. Kobayashi, Y., K. Kawai, H. Honda, S. Tomida, N. Niimi, T. Tamatani, M. Miyasaka, and Y. Yoshikai. 1995. Antibodies against leukocyte functionassociated antigen-1 and against intercellular adhesion molecule-1 together suppress the progression of experimental allergic encephalomyelitis. Cell. Immunol. 164: 295–305. 29. Willenborg, D. O., R. D. Simmons, T. Tamatani, and M. Miyasaka. 1993. ICAM1-dependent pathway is not critically involved in the inflammatory process of autoimmune encephalomyelitis or in cytokine-induced inflammation of the central nervous system. J. Neuroimmunol. 45: 147–154. 30. Rose, J. W., C. T. Welsh, K. E. Hill, M. K. Houtchens, R. S. Fujinami, and J. J. Townsend. 1999. Contrasting effects of anti-adhesion molecule therapy in experimental allergic encephalomyelitis and Theiler’s murine encephalomyelitis. J. Neuroimmunol. 97: 110 –118. 31. Samoilova, E. B., J. L. Horton, and Y. Chen. 1998. Experimental autoimmune encephalomyelitis in intercellular adhesion molecule-1-deficient mice. Cell. Immunol. 190: 83– 89. 32. Dunne, J. L., R. G. Collins, A. L. Beaudet, C. M. Ballantyne, and K. Ley. 2003. Mac-1, but not LFA-1, uses intercellular adhesion molecule-1 to mediate slow leukocyte rolling in TNF-␣-induced inflammation. J. Immunol. 171: 6105– 6111. 33. Szalai, A. J., S. Nataf, X. Z. Hu, and S. R. Barnum. 2002. Experimental allergic encephalomyelitis is inhibited in transgenic mice expressing human C-reactive protein. J. Immunol. 168: 5792–5797. 34. King, P. D., E. T. Sandberg, A. Selvakumar, P. Fang, A. L. Beaudet, and B. Dupont. 1995. Novel isoforms of murine intercellular adhesion molecule-1 generated by alternative RNA splicing. J. Immunol. 154: 6080 – 6093. 35. Giorelli, M., A. De Blasi, G. Defazio, C. Avolio, L. Iacovelli, P. Livrea, and M. Trojano. 2002. Differential regulation of membrane bound and soluble ICAM 1 in human endothelium and blood mononuclear cells: effects of interferon ␤-1a. Cell Commun. Adhes. 9: 259 –272. 36. Werner, A., S. Martin, J. C. Gutierrez-Ramos, and G. Raivich. 2001. Leukocyte recruitment and neuroglial activation during facial nerve regeneration in ICAM1-deficient mice: effects of breeding strategy. Cell Tissue Res. 305: 25– 41. 37. Ochietti, B., P. Lemieux, A. V. Kabanov, S. Vinogradov, Y. St-Pierre, and V. Alakhov. 2002. Inducing neutrophil recruitment in the liver of ICAM-1-deficient mice using polyethylenimine grafted with Pluronic P123 as an organspecific carrier for transgenic ICAM-1. Gene Ther. 9: 939 –945. 38. van Den Engel, N. K., E. Heidenthal, A. Vinke, H. Kolb, and S. Martin. 2000. Circulating forms of intercellular adhesion molecule (ICAM)-1 in mice lacking membranous ICAM-1. Blood 95: 1350 –1355.

857 39. Robledo, O., A. Papaioannou, B. Ochietti, C. Beauchemin, D. Legault, A. Cantin, P. D. King, C. Daniel, V. Y. Alakhov, E. F. Potworowski, and Y. St-Pierre. 2003. ICAM-1 isoforms: specific activity and sensitivity to cleavage by leukocyte elastase and cathepsin G. Eur. J. Immunol. 33: 1351–1360. 40. Anderson, A. C., J. Reddy, R. Nazareno, R. A. Sobel, L. B. Nicholson, and V. K. Kuchroo. 2004. IL-10 plays an important role in the homeostatic regulation of the autoreactive repertoire in naive mice. J. Immunol. 173: 828 – 834. 41. Kuchroo, V. K., A. C. Anderson, H. Waldner, M. Munder, E. Bettelli, and L. B. Nicholson. 2002. T cell response in experimental autoimmune encephalomyelitis (EAE): role of self and cross-reactive antigens in shaping, tuning, and regulating the autopathogenic T cell repertoire. Annu. Rev. Immunol. 20: 101–123. 42. Zhang, X., D. N. Koldzic, L. Izikson, J. Reddy, R. F. Nazareno, S. Sakaguchi, V. K. Kuchroo, and H. L. Weiner. 2004. IL-10 is involved in the suppression of experimental autoimmune encephalomyelitis by CD25⫹CD4⫹ regulatory T cells. Int. Immunol. 16: 249 –256. 43. Furlan, R., P. L. Poliani, P. C. Marconi, A. Bergami, F. Ruffini, L. Adorini, J. C. Glorioso, G. Comi, and G. Martino. 2001. Central nervous system gene therapy with interleukin-4 inhibits progression of ongoing relapsing-remitting autoimmune encephalomyelitis in Biozzi AB/H mice. Gene Ther. 8: 13–19. 44. Bright, J. J., and S. Sriram. 2001. Immunotherapy of inflammatory demyelinating diseases of the central nervous system. Immunol. Res. 23: 245–252. 45. Benveniste, E. N., B. S. Huneycutt, P. Shrikant, and M. E. Ballestas. 1995. Second messenger systems in the regulation of cytokines and adhesion molecules in the central nervous system. Brain Behav. Immun. 9: 304 –314. 46. Brambilla, R., V. Bracchi-Ricard, W. H. Hu, B. Frydel, A. Bramwell, S. Karmally, E. J. Green, and J. R. Bethea. 2005. Inhibition of astroglial nuclear factor ␬B reduces inflammation and improves functional recovery after spinal cord injury. J. Exp. Med. 202: 145–156. 47. Yednock, T. A., C. Cannon, L. C. Fritz, F. Sanchez-Madrid, L. Steinman, and N. Karin. 1992. Prevention of experimental autoimmune encephalomyelitis by antibodies against ␣4␤1 integrin. Nature 356: 63– 66. 48. Baron, J. L., J. A. Madri, N. H. Ruddle, G. Hashim, and C. A. Janeway, Jr. 1993. Surface expression of ␣4 integrin by CD4 T cells is required for their entry into brain parenchyma. J. Exp. Med. 177:57– 68. 49. Gordon, E. J., K. J. Myers, J. P. Dougherty, H. Rosen, and Y. Ron. 1995. Both anti-CD11a (LFA-1) and anti-CD11b (MAC-1) therapy delay the onset and diminish the severity of experimental autoimmune encephalomyelitis. J. Neuroimmunol. 62: 153–160. 50. Miller, D. H., O. A. Khan, W. A. Sheremata, L. D. Blumhardt, G. P. Rice, M. A. Libonati, A. J. Willmer-Hulme, C. M. Dalton, K. A. Miszkiel, and P. W. O’Connor. 2003. A controlled trial of natalizumab for relapsing multiple sclerosis. N. Engl. J. Med. 348: 15–23. 51. Xie, C., P. Alcaide, B. V. Geisbrecht, D. Schneider, M. Herrmann, K. T. Preissner, F. W. Luscinskas, and T. Chavakis. 2006. Suppression of experimental autoimmune encephalomyelitis by extracellular adherence protein of Staphylococcus aureus. J. Exp. Med. 203: 985–994. 52. Chavakis, T., M. Hussain, S. M. Kanse, G. Peters, R. G. Bretzel, J. I. Flock, M. Herrmann, and K. T. Preissner. 2002. Staphylococcus aureus extracellular adherence protein serves as anti-inflammatory factor by inhibiting the recruitment of host leukocytes. Nat. Med. 8: 687– 693.