N-type Calcium Channel in the Pathogenesis of Experimental Autoimmune Encephalomyelitis*

One of the family of voltage-gated calcium channels (VGCC), the N-type Ca2+ channel, is located predominantly in neurons and is associated with a variety of neuronal responses, including neurodegeneration. A precise mechanism for how the N-type Ca2+ channel plays a role in neurodegenerative disease, however, is unknown. In this study, we immunized N-type Ca2+ channel α1B-deficient (α1B−/−) mice and their wild type (WT) littermates with myelin oligodendrocyte glycoprotein 35–55 and analyzed the progression of experimental autoimmune encephalomyelitis (EAE). The neurological symptoms of EAE in the α1B−/− mice were less severe than in the WT mice. In conjunction with these results, sections of the spinal cord (SC) from α1B−/− mice revealed a reduction in both leukocytic infiltration and demyelination compared with WT mice. No differences were observed in the delayed-type hypersensitivity response, spleen cell proliferation, or cytokine production from splenocytes between the two genotypes. On the other hand, Western blot array analysis and RT-PCR revealed that a typical increase in the expression of MCP-1 in the SC showed a good correlation with the infiltration of leukocytes into the SC. Likewise, immunohistochemical analysis showed that the predominant source of MCP-1 was activated microglia. The cytokine-induced production of MCP-1 in primary cultured microglia from WT mice was significantly higher than that from α1B−/− mice and was significantly inhibited by a selective N-type Ca2+ channel antagonist, ω-conotoxin GVIA or a withdrawal of extracellular Ca2+. These results suggest that the N-type Ca2+ channel is involved in the pathogenesis of EAE at least in part by regulating MCP-1 production by microglia.

Multiple sclerosis (MS) is an immune-mediated disorder of the CNS leading to a progressive decline of motor and sensory functions and permanent disability. The pathogenesis of MS consists of two phases, inflammation and neurodegeneration (1). The rodent model of experimental autoimmune encephalomyelitis (EAE) 3 is most commonly studied as a model for MS. Although no single animal model mimics a human autoimmune disease completely, EAE resembles the pathology found in MS, including infiltration of activated peripheral inflammatory cells into the CNS, loss of myelin, and axonal damage (2,3). MS and EAE are considered T helper (Th) 1 cell-dependent diseases because of the association between disease severity and the expression of Th1 cytokines IFN-␥ and IL-12 in the CNS, cerebrospinal fluid, and leukocytes (4 -6). In addition, activated macrophages are the predominant leukocytes in CNS infiltrates of afflicted animals and patients, similar to the infiltrates that characterize Th1-dependent delayed type hypersensitivity (DTH) (7). Recently, Th17 cells producing IL-17 have also been implicated in a variety of autoimmune disorders in both mice and humans (8,9). Once inflammatory cells have spread to the CNS, the immune responses induce demyelination and axonal loss (1).
Voltage-gated calcium channels (VGCCs) are localized in the plasma membrane, and VGCC-mediated Ca 2ϩ entry into cells is essential for a wide variety of physiological events. Structurally, VGCCs are composed of an oligomeric complex of proteins consisting of five subunits: ␣ 1 , ␣ 2 , ␤, ␥, and ␦. The ␣ 1 subunit is the pore-forming component and functions as the voltage sensor (10). The subunits are classified by their pharmacological and electrophysiological properties as P-, Q-, N-, L-, R-, and T-type (11). The ␣ 1B (Ca V 2.2) gene encodes the ␣ 1 subunit of the N-type Ca 2ϩ channel, which is specifically expressed in the nervous system and is predominantly coupled to synaptic transmission. In addition, the N-type Ca 2ϩ channel might play critical roles in the process of neurodegeneration, and its inhibition reduces neuronal damage and improves behavioral outcomes accompanied with traumatic brain injury (12,13) and ischemia (14). In MS and EAE, the N-type Ca 2ϩ channel is reported to accumulate within the axons of demyeli-nating lesions (3). The pathophysiological role of N-type Ca 2ϩ channel in MS/EAE, however, is not known. To understand the pathophysiological function of the N-type Ca 2ϩ channel in EAE, a genetic approach seems quite useful. Mice lacking the ␣ 1B subunit, a functional component of the N-type Ca 2ϩ channel, exhibit a normal lifespan without apparent behavioral abnormalities (15). Some groups have reported that these animals show altered responses to pain (16 -18), glucose homeostasis (19), and sleep (20); however, neurodegeneration has not been examined thus far in these mice. In this study we induced EAE in ␣ 1B Ϫ/Ϫ mice and their wild type (WT) littermates with myelin oligodendrocyte glycoprotein 35-55 (MOG  ) to elucidate how the N-type Ca 2ϩ channel affects the pathology of EAE.

EXPERIMENTAL PROCEDURES
Experimental Animals-All animal procedures conformed to the Japanese regulations for animal care and use, following the Guidelines for Animal Experimentation of the Japanese Association for Laboratory of Animal Science, and were approved by the Animal Care and Use Committee of Eisai Co., Ltd. and Chiba University. Mice with a non-functional ␣ 1B subunit of the N-type Ca 2ϩ channel were previously generated by gene-targeting methods using a TTS ES cell line derived from an F1 embryo between CBA and C57BL/6 (15). For the present experiments, ␣ 1B Ϫ/Ϫ female mice (backcrossed generation into C57BL/6NCrlCrlj) were kept in specific pathogen-free facilities and maintained at constant temperature (22°C Ϯ 2°C) and humidity (55 Ϯ 5%) with a 12:12 h light:dark cycle. The mice had free access to a normal diet (MF, Nippon CLEA, Inc., Tokyo) and water. The ␣ 1B Ϫ/Ϫ and WT littermate female mice (8 -12 weeks) were used for most experiments.

Identification of Genotype in a 1B
Ϫ/Ϫ Mice-Genotypes were identified by PCR from DNA of tail biopsies. The primers used to screen the ␣ 1B wild type, heterozygous, or homozygous genotypes were NP6 (5Ј-TGGCACCTTATGCCTTGCAC-GGTGCCTGCG-3Ј), NP8 (5Ј-GGTCGAGATGGCTTGC-GGGACCCGTTGGGA-3Ј), and AGN2 (5Ј-GCCTGCTTG-CCGAATATCATGGTGGAAAAT-3Ј), corresponding to nucleotides of the cytoplasmic repeat II-III linker of the ␣ 1Bsubunit and inserted into phosphoglycerol kinase promoter neomycin resistance gene sequences. PCR were carried out with a first extension at 94°C for 2 min; followed by 35  Induction and Assessment of EAE-For the induction of EAE, mice were immunized subcutaneously in the four flanks with 100 g of MOG 35-55 emulsified 1:1 in complete Freund's adjuvant (Difco, Detroit, MI) supplemented with 500 g of Mycobacterium tuberculosis H37Ra on day 0. In addition, 30 ng of pertussis toxin (List Biological Laboratories, Campbell, CA) was injected intravenously on days 0 and 2. Mice were weighed and observed for signs of EAE daily. Scoring was as follows: 0, no disease; 0.5, partial tail paralysis; 1, complete tail paralysis; 1.5, decline in the righting reflex; 2, impairment of the righting reflex; 2.5, hindlimb weakness; 3, one hindlimb paralysis; 3.5, both hindlimb paralysis; 4, one forelimb paralysis; 4.5, both forelimb paralysis; 5, moribund or dead.
Histological Analysis-Mice were anesthetized with pentobarbital and perfused with heparin in saline followed by 10% formalin. The spinal cord (SC) was dissected out and fixed in 10% formalin. Paraffin-embedded sections were stained with Luxol fast blue for visualization of demyelination and hematoxylin-eosin (HE) for visualization of inflammatory infiltrate. The degree of demyelination was determined by scoring as follows: 0, no demyelination; 1, mild demyelination; 2, moderate demyelination; 3, severe demyelination. The leukocytic infiltration was quantified as the percentage of the total SC area using BZ-9000 Analyzer (Keyence, Osaka).
Delayed-type Hypersensitivity-DTH responses to MOG  were measured at 13 days post-immunization (dpi). 20 g of MOG  in PBS was injected subcutaneously into the ear.  Before the challenge and 24 h afterward, footpad swelling was measured using a dial thickness gauge. Spleen Cell Proliferation-Spleen cells were harvested from mice at 15 dpi and cultured in 96-well plates at 2 ϫ 10 5 cells/well in culture medium containing various concentrations of MOG 35-55 , 2 g/ml of anti-mouse CD3, or 2 g/ml of concanavalin A. 0.5 Ci of [ 3 H]thymidine was added to each well on the fifth day, and cells were harvested 16 h later. Radioactivity was measured using a BETA PLATE liquid scintillation counter (PerkinElmer Life Sciences). The culture medium for spleen cells consisted of RPMI1640 (Sigma) supplemented with 10% FBS, 500 M 2-mercaptoethanol, penicillin (100 units/ml), and streptomycin (100 g/ml).
Cytokine Production Assay-For cytokine production assays, mice were immunized with MOG  according to the method used for EAE induction without the pertussis toxin injection. Pertussis toxin injection is essential for EAE induction by increasing the permeability of the blood-brain barrier (21). In our experience, however, spleen cells from mice suffering from EAE occasionally show a reduction in cytokine production in response to stimuli. At 14 dpi, spleen cells were prepared by plating in 96well round-bottom plates (2 ϫ 10 5 cells/well) and stimulated with MOG 35-55 or 2 g/ml of antimouse CD3 plus anti-mouse CD28. The resulting supernatants were collected after 72 h and stored at Ϫ20°C until tested. An ELISA was performed to detect the concentration of each cytokine according to the manufacturer's instructions (IL-12, Biolegend, San Diego, CA; IFN-␥, BD Biosciences, San Jose, CA; and IL-17, R&D Systems, Minneapolis, MN). Western Blot Array Analysis for Cytokines-Mice immunized with or without MOG 35-55 for 10 days were anesthetized with pentobarbital and perfused with heparin in saline. The SC was dissected out and homogenized in a lysis buffer (20 mM Tris, pH 7.6, 0.5% Nonidet P-40, 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 2 mM DTT, 0.5 mM PMSF, 20 mM ␤-glycerophosphate, 1 mM sodium vanadate, and 1 mg/ml of leupeptin). The homogenates were then centrifuged at 14,000 ϫ g for 30 min at 4°C. The resulting supernatants (400 g) were subjected to the Mouse Cytokine Antibody Array 3 (RayBiotech Inc., Norcros, GA), which detects 62 cytokines according to the manufacturer's protocol.
Detection of Chemokine mRNA Transcripts-Mice immunized with or without MOG 35-55 for 10 or 14 days were anesthetized with pentobarbital and perfused with heparin in saline, and the SC was dissected out. Total RNA was prepared from the SC using ISOGEN (Wako Chemicals, Tokyo) according to the manufacturer's instructions. Single strand cDNA was synthesized from prepared RNA (1 g), using Moloney MLV reverse transcriptase (Invitrogen) and oligo(dT) primer (Invitrogen) in a total volume of 20 l. The resulting cDNA sample (1 l) was subjected to PCR amplification of mouse MCP-1, MIP-1␥, MIP-2, or MIP-3␤ cDNA using specific primers (sense primer, 5Ј-AGCACCAGCCAACTCTCACTG-3Ј, antisense primer, 5Ј-AGAAGT-GCTTGAGGTGGTTGTG-3Ј for MCP-1; sense primer, 5Ј-TCATTC-TTACAACTGCTCTTG-3Ј, antisense primer, 5Ј-TGAGTTTTGC-TCCAATCTTTC-3Ј for MIP-1␥; sense primer, 5Ј-CACCAACC-ACCAGGCTACAG-3Ј, antisense primer, 5Ј-TTTGACCGCCCTT-GAGAGTG-3Ј for MIP-2; sense primer, 5Ј-GACTGCTGCCTGTC-TGTGAC-3Ј, antisense primer, 5Ј-TCTTCAGTCTTCGGATGATG-3Ј for MIP-3␤). As an internal control, GAPDH cDNA was amplified using specific primers (sense primer, 5Ј-ACCACAGTCCATGC-CATCAC-3Ј, antisense primer, 5Ј-TCCACCACCCTGTTGCTGTA-3Ј). Using these primers, PCR was performed at 94°C for 5 min, followed by individual cycles at 94°C for 30 s, individual temperature for 60 s, 72°C for 45 s, with an extension step of 7 min at 72°C at the end of the last cycle. The individual annealing temperatures, cycle numbers, and product sizes were as follows: 62°C, 34 cycles and 465-bp for MCP-1; 57°C, 34 cycles and 314-bp for MIP-1␥; 59°C, 34 cycles and 167-bp for MIP-2; 59°C, 34 cycles and 181-bp for MIP-3␤; and 54°C, 27 cycles and 452-bp for GAPDH. To evaluate leukocytic infiltration into the SC, each proximal portion of the SC used for total RNA preparation was fixed in 4% paraformaldehyde, 0.1 M sodium phosphate buffer (pH 7.4) and stored at 4°C. The SC was then cut into 20-m thick sections and stained with 4Ј,6-diamino-2phenylindole (DAPI). In case of elucidating the correlation between inflammatory progression and chemokine (MCP-1 and MIP-1␥) expression, mice immunized with MOG 35-55 for 10 or 14 days were anesthetized with pentobarbital and perfused with heparin in saline, and the SC was dissected out and stored at Ϫ20°C. Each proximal portion of the SC was fixed and confirmed whether or not leukocytes infiltrated into the SC by DAPI staining. Then, total RNA was prepared from the SC without and with leukocytic infiltration at 10 and 14 dpi, respectively, and the synthesized cDNA was subjected to real time PCR. All primers were obtained from Applied Biosystems (Ccl2, MCP-1; Mm00441242_m1, Ccl9, MIP-1␥; Mm00441260_m1 and GAPDH, customized). Signal values of each gene were normalized by the expression of GAPDH.
Immunohistochemical Analysis-The SC fixed in 4% paraformaldehyde, 0.1 M sodium phosphate buffer (pH 7.4) was cut into 20-m thick sections, which were placed on poly-L-lysine-coated slides. The sections were subjected to double staining with rabbit anti-MCP-1 polyclonal antibody (Abcam, Cambridge, MA) in combination with mouse anti-glial fibrillary acidic protein (GFAP) monoclonal antibody (Cy3-conjugated, Sigma), goat anti-Iba1 polyclonal antibody (Abcam), or mouse anti-CD3 monoclonal antibody (BD Biosciences). In determination of the morphology of activated microglia, a rabbit anti-Iba1 polyclonal antibody (Wako Chemicals, Tokyo) was used. In some experiments, the sections were subjected to double staining with a goat anti-MIP-1␥ polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and an anti-CD3 antibody (BD Biosciences).
Preparation of Microglia and Astrocytes-Microglia were prepared as a primary mixed culture with astrocytes. The brain cortex was dissected from neonatal WT and ␣ 1B Ϫ/Ϫ mice (1-3 days old) on a child platform under sterile conditions. Tissue was collected in Leibovitz 15 medium. The cortical tissue was minced and trypsinized for 30 min (0.25% trypsin with 10 mM HEPES buffer, 0.6% glucose, antibiotic-antimycotic (Invitrogen), and DNase I) at 37°C. The cortical tissues were dissociated into single cells by gentle trituration in culture medium (DMEM/F-12 supplemented with 10% FBS and antibiotic-antimycotic) and filtered through a nylon mesh filter. Suspended cells from two mouse brains were added to each culture flask (75 cm 2 ) with 15 ml of culture medium. Cultures were maintained at 37°C in a humidified atmosphere (5% CO 2 ). Culture medium was replaced after 2 days and twice per week thereafter. After 14 days in vitro, microglia were harvested from the mixed glial cell culture by shaking the flasks at 150 rpm for 2 h at 37°C, which is a purification step for microglia. Microglia and the resulting monolayer astrocytes were then replated for an assay in 96-well plates containing Microglia Culture Medium (SUMITOMO BAKELITE, Tokyo) and DMEM supplemented with 10% FBS, respectively.
In Vitro Activation of Microglia and Astrocytes-Activation of microglia and astrocytes was performed by stimulation with LPS (Sigma) or a cytokine mixture of TNF-␣ and IFN-␥ at equal concentrations (PeproTech, London) for 24 h. Microglia from WT mice were stimulated with various concentrations of LPS or the cytokine mixture to determine the maximal dose of each agent for MCP-1 induction. To elucidate the effects of -conotoxin, GVIA (-CgTx) on MCP-1 production in activated microglia, the microglia from WT mice were stimulated with 0.1 g/ml of LPS or the cytokine mixture (1 ng/ml of TNF-␣ plus 1 ng/ml of IFN-␥) in the presence or absence of 1 M -CgTx. In some experiments, microglia from WT and ␣ 1B Ϫ/Ϫ mice were stimulated with 0.1 g/ml of LPS or the cytokine mixture (1 ng/ml of TNF-␣ plus 1 ng/ml of IFN-␥). To elucidate the role of extracellular Ca 2ϩ in MCP-1 production, microglia from WT mice were stimulated with 0.1 ng/ml of LPS in DMEM or Ca 2ϩ -free DMEM (Invitrogen). The supernatants were collected 24 h after stimulation and subjected to an ELISA for MCP-1 according to the manufacturer's instructions (BIOSOURCE, Camarillo, CA). The number of microglia or astrocytes per well was detected by using the Cell Counting Kit-8 (DOJINDO, Kumamoto).

Preparation of Microglia from the SC of Adult Mice-WT and ␣ 1B
Ϫ/Ϫ mice were anesthetized with 10% urethane and per-fused with heparin in saline to clear the intravascular blood cells. The SC was extracted by flushing the spinal canal with saline and then minced, trypsinized, and dissociated into single cells as described above. The cells were suspended in ice-cold 75% Percoll (GE Healthcare; 9 volumes of Percoll were mixed with 1 volume of 10ϫ PBS to yield a stock isotonic Percoll solution, and Percoll densities were obtained by dilution with PBS). 3.3 ml of cell suspension was gently applied on 5 ml of ice-cold 25% Percoll using a Pasteur pipette. The density gradient was centrifuged at 800 ϫ g (slow acceleration and no brake) for 25 min at 4°C. After centrifugation the cells at the interface were collected and washed with PBS. The cells were cultured for 1 day in Microglia Culture Medium before the experiments.
Determination of ␣ 1B Expression on Microglia and Astrocytes-Cultured microglia and astrocytes on Lab-Tek chamber (Thermo Fisher Scientific Inc., Waltham, MA) were fixed with 4% paraformaldehyde for 20 min and subjected to double staining with an anti-Ca v 2.2 antibody (Chemicon, Temecula, CA) in combination with goat anti-Iba1 polyclonal antibody (Abcam) or FITCconjugated anti-mouse CD11b (BD Biosciences) and guinea pig anti-glutamate-aspartate transporter (GLAST) polyclonal antibody (Millipore, Billerica, MA). After incubating with fluorescent dye-conjugated second antibodies, localization of ␣ 1B was observed by a fluorescent microscopy and a confocal laser scanning microscope (Axioplan and LSM510, Zeiss, Oberkochen). In detection of ␣ 1B mRNA expression, total RNA was prepared from cultured microglia and astrocytes of the two genotypes, and RT-PCR was performed according to a previous report with a slight modification (22). Using specific primers for mouse ␣ 1B (sense primer, GTGGACAACCTTGCCAATG, antisense primer, GCATG-GTCACAATGTAATG) and KOD-FX DNA polymerase (TOYOBO, Osaka), PCR was performed at 98°C for 2 min, followed by 35 cycles at 98°C for 15 s, 54°C for 30 s, 68°C for 30 s, with an extension step of 7 min at 68°C at the end of the last cycle. The PCR product sizes were 1270 and 1273 bp for splicing variants of ␣ 1B .
Statistics-Data are expressed as the mean Ϯ S.E. Statistical analysis was conducted using the software package SAS 8.1 (SAS Institute Japan Ltd., Tokyo). Statistical significance was determined by Student's t test or analysis of variance followed by Student's t test, and p values of Ͻ0.05 were considered to be significant.

Neurological Symptoms of EAE-WT and ␣ 1B
Ϫ/Ϫ mice were immunized with MOG  and monitored up to day 58. As shown in Fig. 1A, ␣ 1B Ϫ/Ϫ mice showed less severe neurological symptoms than WT mice at all time points. This significant difference in the course of disease between WT and ␣ 1B Ϫ/Ϫ mice was also observed for male mice (data not shown). Likewise, the body weight loss in WT mice was significantly more severe than in ␣ 1B Ϫ/Ϫ mice (Fig. 1B). Clinical parameters of this and other experiments are summarized in Table 1. The mean maximal score of ␣ 1B Ϫ/Ϫ mice was significantly lower than that of WT mice, although there was no clear difference in the incidence of the disease between the two genotypes. The day of EAE onset in ␣ 1B Ϫ/Ϫ mice tended to be slower than that in WT mice.
Histopathological Features of SC with EAE-In ␣ 1B Ϫ/Ϫ mice, demyelination and vacuolation of the thoracic cord were less apparent compared with WT mice in the chronic phase (Fig. 2, A-D). Accordingly, the semi-quantitative scoring of histological features revealed that the degree of demyelination in ␣ 1B Ϫ/Ϫ mice was significantly decreased compared with that in WT mice (Fig. 2E). To evaluate the histopathology of SCs in the acute phase, both ␣ 1B Ϫ/Ϫ and WT mice were immunized with MOG  and sacrificed at 15 dpi when the peak of the acute attack was observed. At this point sections of the SC from most of the WT mice revealed widespread perivascular leukocyte infiltrates, accompanied with severe demyelination (Fig. 3, A  and B). On the other hand, sections of ␣ 1B Ϫ/Ϫ SCs showed only focal perivascular leukocytic infiltrates and very slight demyelination (Fig.  3, C and D). The ratio of the leukocyte infiltration area to the whole SC area was significantly reduced in ␣ 1B Ϫ/Ϫ mice compared with the WT mice (Fig. 3E).
Peripheral Immunological Responses against MOG 35-55 -To elucidate the mechanisms underlying the attenuated symptoms of EAE in ␣ 1B Ϫ/Ϫ mice, differences in the potency of the immune responses to MOG  between ␣ 1B Ϫ/Ϫ and WT mice were examined. In the above acute phase experiment, we evaluated the DTH response before sacrifice. At 13 dpi, MOG  was subcutaneously injected into the ear, and ear swelling was measured. As shown in Fig. 4A, a DTH response was induced in ␣ 1B Ϫ/Ϫ mice, the extent of which was comparable with that in WT mice. The timing of the challenge with MOG  to the ear, namely, either before or after EAE onset, was irrelevant to DTH generation. In either case, the DTH response to MOG  was equivalent between ␣ 1B Ϫ/Ϫ and WT mice (data not shown). Moreover, we evaluated the DTH response of both genotypes to another antigen, DNP-KLH. This antigen caused swelling in ␣ 1B Ϫ/Ϫ mice equivalent to that seen in WT mice after a challenge on the footpad (Fig. 4B).
To further define the immunological function of both WT and ␣ 1B Ϫ/Ϫ mice, we examined the proliferation of spleen cells isolated from WT and ␣ 1B Ϫ/Ϫ mice at 15 dpi. Both anti-CD3and concanavalin A-induced proliferation of T lymphocytes generated from the ␣ 1B Ϫ/Ϫ cells in vitro were similar to those for WT cells (Fig. 4, C and D). On the other hand, the ␣ 1B Ϫ/Ϫ cells showed a tendency toward greater proliferative responses to MOG 35-55 compared with WT cells (Fig. 4E).
We collected supernatants from cultured spleen cells and determined the production of cytokines such as IFN-␥, IL-12, and IL-17 by spleen cells from WT and ␣ 1B Ϫ/Ϫ mice sensitized with MOG  . As shown in Fig. 5, stimulation of spleen cells with MOG 35-55 induced IFN-␥, IL-12, and IL-17 secretion in a dose-dependent manner. The ␣ 1B Ϫ/Ϫ cells secreted a greater amount of each cytokine compared with the WT cells, although there was no statistical difference in cytokine production between the two genotypes. This showed a good correlation with the results of MOG 35-55 -induced splenocyte proliferation (Fig. 4E). On the other hand, stimulation of spleen cells with anti-CD3 effectively induced the production of all cytokines in both WT and ␣ 1B Ϫ/Ϫ cells, without any statistical difference. Expression Profiles of Cytokines-In EAE, chemokines produced in the CNS can contribute to the recruitment of inflammatory mononuclear cells from the periphery to sites of injury, leading to neuroinflammation (23,24). To further elucidate the mechanisms underlying the attenuated symptoms of EAE in ␣ 1B Ϫ/Ϫ mice, we comprehensively investigated differences in expression of cytokines involving chemokines in the SC between WT and ␣ 1B Ϫ/Ϫ mice at 10 dpi. As shown in Fig. 6, A and B, several cytokines were induced by immunization with MOG  . We focused on MCP-1, MIP-1␥, MIP-2, and MIP-3␤, which are indicated by a double-line square in Fig. 6B. An analysis by densitometer revealed that the expression level of these chemokines in WT mice immunized with MOG 35-55 was at least 2-fold higher than in ␣ 1B Ϫ/Ϫ mice (Fig. 6C). We next investigated the mRNAs of MCP-1, MIP-1␥, MIP-2, and MIP-3␤ in the SC at 10 and 14 dpi, around the onset of EAE in the WT and ␣ 1B Ϫ/Ϫ mice, respectively. The typical induction of MCP-1 mRNA was observed in individual SC named 1 and 3 from WT mice at 10 dpi and 1 and 2 from WT mice at 14 dpi. For the ␣ 1B Ϫ/Ϫ mice, MCP-1 mRNA was typically induced in the SC named 1 at 14 dpi. In accordance with these results, the induction pattern of MIP-1␥ mRNA in the SC for both genotypes was quite similar to that seen for MCP-1 mRNA. In the case of MIP-2 mRNA, the induction pattern in the SC for WT mice was well correlated with that of MCP-1 mRNA. The induction pattern of MIP-2 mRNA in the SC for ␣ 1B Ϫ/Ϫ mice, however, was different from that of MCP-1 mRNA. The induction pattern of MIP-3␤ mRNA in the SC for both genotypes showed no similarity to those of MCP-1, MIP-1␥, and MIP-2 mRNAs (Fig. 7A). We confirmed that leukocytic infiltration was present in SCs 1 and 3 from WT mice at 10 dpi; 1 and 2 from WT mice at 14 dpi; and 1 from ␣ 1B Ϫ/Ϫ mice at 14 dpi based on DAPI staining (data not shown), indicating a good correlation between MCP-1 and MIP-1␥ expression and leukocytic infil-tration in the SC. To further define the correlation between those chemokines expression and inflammatory progression, we examined the mRNA expression of MCP-1 and MIP-1␥ in the SC of the two genotypes without and with leukocytic infiltration at 10 and 14 dpi, respectively. The expression ratio of each chemokine mRNA in the SC with infiltration at 14 dpi versus that without the infiltration at 10 dpi was calculated. As shown in Fig. 7B, the marked induction in expression of MCP-1 mRNA (48-fold) was observed in the SC of WT mice, the ratio of which was significantly decreased in the case of ␣ 1B Ϫ/Ϫ mice (13-fold). The expression of MIP-1␥ mRNA was also induced in the SC of WT mice (23-fold), the ratio of which showed a tendency to decrease in case of ␣ 1B Ϫ/Ϫ mice (15-fold). Then, the localization of MCP-1 and MIP-1␥ in the SC was then investigated by immunohistochemistry.
As shown in Fig. 8, A and B, MCP-1-like immunoreactivity (LI) was primarily observed in GFAP-positive cells, astrocytes, in the control SC of both WT and ␣ 1B Ϫ/Ϫ mice. On the other hand, localization of MCP-1-LI showed good correlation with Iba1-positive cells, either microglia or macrophages, in the marginal area of the leukocytic infiltratation region in the SC for WT mice with EAE progression (Fig. 8C). For the SC from ␣ 1B Ϫ/Ϫ mice, MCP-1-LI also co-localized with Iba1-LI in the marginal area of the leukocytic infiltration region. The intensity  OCTOBER 22, 2010 • VOLUME 285 • NUMBER 43 of MCP-1-LI, however, was clearly suppressed compared with the WT mice (Fig. 8D). The population of Iba1-positive cells in the marginal area of the leukocytic infiltratation region was comparable between the two genotypes (Fig. 8, C and D). A portion of the MCP-1-LI was co-localized with CD3-positive infiltrating cells (Fig. 8E). Iba1-positive cells were distributed not only in the leukocytic infiltration area, but also extensively in its marginal area of the SC (Fig. 8, F and G). Likewise, most Iba1-positive cells showed morphology with an enlarged cell body and short processes (Fig. 8H). The GFAP-LI tended to divide the MCP-1-LI at the marginal area of the leukocytic infiltration region (Fig. 8I). On the other hand, the MIP-1␥-LI was restricted to the leukocytic infiltration region and mostly co-localized with CD3-LI (Fig. 8J).

N-type Ca 2؉ Channel in EAE
MCP-1 Secretion from Microglia-Activated microglia are known to secrete a variety of proinflammatory cytokines and chemokines and to regulate inflammatory responses in the CNS (25). Because we hypothesized that activated microglia could be a plausible source of MCP-1, we investigated whether the N-type Ca 2ϩ channel contributes to MCP-1 production in activated microglia. Both LPS and cytokines (TNF-␣ plus IFN-␥) induced MCP-1 production in microglia in a dose-dependent manner (Fig. 9, A and B). The specific N-type Ca 2ϩ channel antagonist -CgTx significantly inhibited LPS-and cytokine-induced MCP-1 release from microglia of WT mice (Fig. 9C). In addition, MCP-1 secretion induced by LPS and cytokines from microglia of ␣ 1B Ϫ/Ϫ mice was significantly less than that of WT mice (Fig. 9D). As expected, -CgTx has no inhibitory effect on LPS-induced MCP-1 secretion from microglia of ␣ 1B Ϫ/Ϫ mice, indicating that -CgTx works specifically on the N-type Ca 2ϩ channel (data not shown). MCP-1 secretion induced by LPS from microglia was significantly sensitive to extracellular Ca 2ϩ (Fig. 9E). On the other hand, MCP-1 secretion from astrocytes was also induced by both LPS and cytokines. The activity of cytokineinduced MCP-1 release per cell from microglia was higher than that from astrocytes. In contrast, the activity of LPS-induced MCP-1 release pre-cell from microglia was less than that from astrocytes (Fig. 9F). In either case, the activated microglia was potentially equivalent to the activated astrocytes as the source for MCP-1. As shown in Fig. 9G, the ␣ 1B -LI was identified on Iba1-positive cells, microglia, of WT mice but not ␣ 1B Ϫ/Ϫ mice, indicating that the N-type Ca 2ϩ channel is expressed on microglia. The z-stack analysis showed colocalization of ␣ 1B -LI and CD11b-LI in the dual labeled cells. The ␣ 1B -LI was identified on GLAST-positive cells, astrocytes of WT mice but not ␣ 1B Ϫ/Ϫ mice, indicating that the N-type Ca 2ϩ channel is expressed also on astrocytes. The z-stack analysis showed the colocalization of ␣ 1B -LI and GLAST-LI in the dual labeled cells (Fig. 9H). In agreement with these results, ␣ 1B mRNA was detected in microglia and astrocytes of WT mice but not ␣ 1B Ϫ/Ϫ mice (Fig.  9I). We further estimated a significant difference in the cytokine-induced MCP-1 production between the two genotypes in microglia from the SC of adult mice (Fig. 9J).

DISCUSSION
Numerous studies have demonstrated that the calcium ion is critical in the neuronal damage seen in neurodegenerative disease models such as trauma (26), ischemia (27), and EAE (28). Hence, we investigated whether the N-type Ca 2ϩ channel is involved in the pathology of EAE. Our data clearly show that ␣ 1B Ϫ/Ϫ mice could be more resistant to MOG 35-55 -induced EAE compared with WT mice (Fig. 1 and Table 1). This suggests participation of the N-type Ca 2ϩ channel in EAE. At first, we predicted that the amelioration of EAE in the ␣ 1B Ϫ/Ϫ mice might result from repression of axonal damage. Kornek et al. (3) reported accumulation of the N-type Ca 2ϩ channel in dystrophic axons of EAE rats, suggesting that calcium influx through this channel is a possible mechanism of axonal degeneration in EAE. We also confirmed the decrease in neuronal cell death in ␣ 1B Ϫ/Ϫ mice, but not WT mice, in brain damaged models, namely, cold-injury of the cortex and direct injection of IL-1␤ in the hippocampus. 4 Histopathological analysis, however, revealed that leukocyte infiltration in the SC during the acute phase of EAE in ␣ 1B Ϫ/Ϫ mice was suppressed compared with that in WT mice (Fig. 3). Infiltration of inflammatory cells into the SC is well known to occur prior to the clinical symptoms of EAE, and its range affects the severity of EAE (29,30). Thus, the N-type Ca 2ϩ channel can contribute to EAE at a much earlier stage than axonal damage, raising the possibility that the role of the N-type Ca 2ϩ channel in EAE is associated with alterations in the antigen-specific immune responses. To address to this notion, differences in the DTH response, a proliferation assay of splenocytes, and a cytokine production assay from splenocytes between WT and ␣ 1B Ϫ/Ϫ mice were examined. A DTH response was observed in ␣ 1B Ϫ/Ϫ mice induced by MOG  , similar to observations in WT mice. A proliferation assay showed no impairment in the responsiveness of T cells to MOG  . Rather, the proliferative response to MOG 35-55 of T cells from ␣ 1B Ϫ/Ϫ mice was greater than that of T cells from WT mice (Fig. 4). The production of Th1 and Th17 cytokines, which are closely related to EAE, from ␣ 1B Ϫ/Ϫ mice was comparable with that from WT mice. Furthermore, the MOG 35-55induced production of Th1/Th17 cytokines in ␣ 1B Ϫ/Ϫ cells tended to be greater than that in WT cells (Fig. 5), which was consistent with the MOG 35-55 -induced T cell proliferation (Fig. 4). These results clearly indicate that ␣ 1B Ϫ/Ϫ mice have a normal immune response against MOG  and suggest that the attenuation of EAE symptoms in ␣ 1B Ϫ/Ϫ mice cannot be explained by peripheral immunodeficiency. This idea was also supported by the comprehensive analysis for cytokine production in MOG 35-55 -stimulated lymph node cells (supplemental Fig. S1). The expression pattern of MOG 35-55 -induced cytokines in lymph node cells from WT mice was quite similar to that from ␣ 1B Ϫ/Ϫ mice. Furthermore, the induction ratio of each cytokine was also equivalent between the two genotypes except IL-3.
Not only peripheral immune cells, but also CNS glial cells such as microglia and astrocytes, are important for the development of MS and EAE (31)(32)(33). Activated glia produce nitric oxide, proinflammatory cytokines, and chemokines that contribute to peripheral leukocyte recruitment, oligodendrocyte cell death, and axonal damage (32,34,35). Indeed, the suppression of glial cell activity by a small molecule compound leads to the amelioration of EAE (24). To further elucidate whether the mechanism underlying the attenuated symptoms of EAE in ␣ 1B Ϫ/Ϫ mice is attributable to some event in the SC, we comprehensively investigated changes in the expression of cytokines/chemokines in the SC (Fig. 6). The expression of 15 molecules in total was increased by MOG  immunization in WT mice than in ␣ 1B Ϫ/Ϫ mice. We then identified MCP-1, MIP-1␥, MIP-2, and MIP-3␤ as plausible candidates to connect N-type Ca 2ϩ channel signals to the progression of EAE, given that their induction in WT mice was more than twice as high as that in ␣ 1B Ϫ/Ϫ mice. In addition, ingenuity pathway analysis (IPA, Redwood, CA), a recently developed informatics tool, supported the possible relationship of these chemokines to the progression of EAE (data not shown). Among these candidates, the induction of MCP-1 and MIP-1␥ expression was clearly present in the SC where the infiltration of leukocytes was observed for both genotypes. Likewise, the induction ratio of MCP-1 and MIP-1␥ in the SC of WT mice was decreased in the case of ␣ 1B Ϫ/Ϫ mice (Fig. 7). Thus, we focused on MCP-1 and MIP-1␥ and evaluated the localization of these chemokines in the SC. The localization of MIP-1␥-LI was restricted to the leukocytic infiltrate area and was mostly seen in infiltrating T lymphocytes (Fig. 8J). We confirmed that MIP-1␥-LI was also observed in CD4-positive Th cells (data not shown). This finding was supported by a previous study (36) and suggests that the expression of MIP-1␥ may be attributed to T lymphocytes recruited to the SC by some other chemokine/cytokine. On the other hand, MCP-1-LI was localized to astrocytes under resting conditions in the SC with similar intensity between the two genotypes (Fig. 8, A and B). Once an inflammatory response is initiated in the SC, MCP-1 could be predominantly expressed in Iba1-positive cells at a higher level in WT mice than in ␣ 1B Ϫ/Ϫ mice (Fig. 8, C and D). In such a state of the SC, MCP-1-LI was also partially observed in the T lymphocyte infiltrates (Fig. 8E). The population of Iba1-positive cells correlated with MCP-1-LI, however, was much larger than that of CD3-positive cells, indicating that Iba1-positive cells could be the primary source of MCP-1 near the onset of EAE (Fig. 8, E-G). Furthermore, the decrease in intensity of MCP-1-LI in Iba1-positive cells in the SC of ␣ 1B Ϫ/Ϫ mice suggests that the N-type Ca 2ϩ channel is possibly involved in MCP-1 production in Iba1-positive cells, at least in the marginal area of the leukocytic infiltrate region.
Although Iba1 is expressed in both microglia and macrophages (37), we considered the majority of the Iba1-positive cells in this case to be microglia based on the following criteria: 1) Iba1-positive cells were observed in the white matter of the SC in a widespread manner near the infiltrating area of leukocytes, suggesting that the infiltrated macrophages could not cover the entire area and were restricted to the area where the leukocytes were present; and 2) most of the Iba1-positive cells showed a hypertrophied or amoeboid morphology typical of that observed in activated microglia (Fig. 8H), as described in a previous study (38). Furthermore, we confirmed that the N-type Ca 2ϩ channel is not expressed, at least on inflammatory peritoneal macrophages, using an anti-Ca v 2.2 antibody, 4 indicating no participation by macrophages in the N-type Ca 2ϩ channel-regulated MCP-1 production in Iba1-positive cells.
The fact that the MCP-LI observed in astrocytes under resting conditions disappeared from astrocytes in the marginal area of leukocytic infiltration is of interest, although the exact reason for this observation is still unknown. The activated astrocytes having N-type Ca 2ϩ channel also release MCP-1 at least in vitro (Fig. 9, F, H, and I). A recent report suggests that astrocytederived MCP-1 plays a neuroprotective role as a mediator (39). How astrocyte-derived MCP-1 affects EAE progression is also of interest.
MCP-1 is also one of the chemokines released from glial cells and can attract various types of cells including monocytes, T lymphocytes, and dendritic cells (40). Previous studies have demonstrated the expression of MCP-1 in the CNS of patients with MS (41)(42)(43). In EAE, the role of MCP-1 has been well characterized by neutralization with an anti-MCP-1 antibody or DNA vaccination, MCP-1 knock-out mice, and CCR2 (receptor for MCP-1) knock-out mice (44 -48). In these studies, the inhibition of MCP-1 decreased the clinical severity of EAE in conjunction with a reduction in leukocyte infiltration into the CNS. Taken together with the present results, the importance of activated microglia-derived MCP-1 in the progression of EAE is clear. The involvement of Ca 2ϩ signals in the regulation of MCP-1 expression is not known. We investigated this point using primary cultured microglia from WT and ␣ 1B Ϫ/Ϫ mice. The genetic disruption of the functional subunit of the N-type Ca 2ϩ channel ␣ 1B , treatment of microglia from WT mice with -CgTx, or extracellular Ca 2ϩ withdrawal significantly decreased MCP-1 production in activated microglia. Indeed, the N-type Ca 2ϩ channel is expressed on microglia from WT mice (Fig. 9). These results clearly demonstrate that the N-type Ca 2ϩ channel can participate in MCP-1 production in activated microglia. However, extracellular Ca 2ϩ withdrawal showed more inhibitory effect on MCP-1 release from activated microglia compared with the inhibition of N-type Ca 2ϩ channel by genetic disruption or -CgTx. This fact tempts us to consider that other Ca 2ϩ entry pathways are also likely to be important. Further studies will be necessary to determine the precise mechanism by which the N-type Ca 2ϩ channel regulates MCP-1 production in microglia in response to inflammatory stimuli such as LPS and cytokines. The present study, however, identified a novel role for the N-type Ca 2ϩ channel in regulating the secretion of MCP-1 from microglia. In general, the role of the N-type Ca 2ϩ channel is considered to be restricted to the regulation of neuronal function (15)(16)(17)(18). Thus, FIGURE 9. Involvement of the N-type Ca 2؉ channel in MCP-1 production in microglia. Microglia from the cortex of neonatal WT mice were stimulated with the indicated concentrations of LPS (A) or a cytokine mixture (TNF-␣ plus IFN-␥ in equal amounts) (B). C, microglia from WT mice were stimulated with 0.1 g/ml of LPS or cytokines (1 ng/ml of TNF-␣ plus 1 ng/ml of IFN-␥) in the absence (black bar) or presence (white bar) of -CgTx (1 M). Supernatants were collected 24 h after stimulation and subjected to an ELISA for MCP-1. Data are shown as mean Ϯ S.E. (n ϭ 3). *, p Ͻ 0.05; **, p Ͻ 0.01 (Student's t test for unpaired values). D, cultured microglia from neonatal WT (black bar) and ␣ 1B Ϫ/Ϫ (KO, white bar) mice were stimulated with 0.1 g/ml of LPS or cytokines (1 ng/ml of TNF-␣ plus 1 ng/ml of IFN-␥), and the supernatants were subjected to an ELISA for MCP-1 in the same manner as described above. Data are shown as mean Ϯ S.E. (n ϭ 3). **, p Ͻ 0.01 (Student's t test for unpaired values). E, microglia from WT mice were stimulated with 0.1 ng/ml of LPS in the presence (black bar) or absence (white bar) of extracellular Ca 2ϩ , and the supernatants were subjected to an ELISA for MCP-1 in the same manner as described above. Data are shown as mean Ϯ S.E. (n ϭ 3). **, p Ͻ 0.01 (Student's t test for unpaired values). F, microglia (black bar) or astrocytes (white bar) from WT mice were stimulated with 0.1 ng/ml of LPS or cytokines (1 ng/ml of TNF-␣ plus 1 ng/ml of IFN-␥), and the supernatants were subjected to an ELISA for MCP-1 in the same manner as described above. Each bar is expressed as MCP-1 production per cell. Data are shown as mean Ϯ S.E. (n ϭ 4). **, p Ͻ 0.01 (Student's t test for unpaired values). Cultured microglia (G) or astrocytes (H) express the functional subunit ␣ 1B of the N-type Ca 2ϩ channel. Fixed cultured microglia or astrocytes were subjected to double staining with an anti-␣ 1B (Ca v 2.2) antibody and an anti-Iba1 antibody, an anti-CD11b antibody (G), or an anti-GLAST antibody (H), followed by a reaction with fluorescent dye-conjugated second antibody. Z-stack analysis was performed by a confocal laser scanning microscope. I, total RNA was prepared from cultured microglia and astrocytes of the two genotypes (WT and ␣ 1B Ϫ/Ϫ ), and RT-PCR was performed. PCR products were subjected to 1.2% agarose gel electrophoresis and visualized by staining with ethidium bromide. J, microglia from the SC of adult WT (black bar) and ␣ 1B Ϫ/Ϫ (KO, white bar) mice showed a difference in MCP-1 production similar to microglia from the neonatal cortex. Data are shown as mean Ϯ S.E. (n ϭ 3). **, p Ͻ 0.01 (Student's t test for unpaired values). OCTOBER 22, 2010 • VOLUME 285 • NUMBER 43 the present study gives new insight into the function of the N-type Ca 2ϩ channel as a regulator of inflammatory responses and suggests that the N-type Ca 2ϩ channel might be a target for drug discovery for MS.