All animal experiments were performed according to the Guidelines for the Protection of Experimental Animals issued by the Japanese Government, the US National Institutes of Health, and the Society for Neuroscience. Heterozygous breeder pairs of twitcher (twi/+) were originally purchased from Jackson Laboratory (Bar Harbor, ME). Twi/twi and normal age-matched siblings (+/+) were identified by genotyping with genomic DNA extracted from the clipped tails by use of a Puregene DNA Isolation Kit (Gentra Systems, Minneapolis, MN). Genotyping was performed as previously reported 17.

The following primary antibodies were used: phycoerythrin (PE)-conjugated anti-TNFα (1:50; PharMingen, San Diego, CA), mouse monoclonal anti-myelin basic protein (MBP) antibody (1:200; Sternberger Monoclonals Incorporated, Lutherville, MA), rabbit polyclonal anti-rat-pi-form of glutathione-S-transferase (pi-GST) antibody (1:1000; MBL, Nagoya, Japan), rabbit polyclonal anti-cow glial fibrillary acidic protein (GFAP) antibody (prediluted; DAKO, Glostrup, Denmark), biotinylated Ricinus communis-agglutinin-1 (RCA-1) (50 μg/ml; Vector Laboratories, Burlingame, CA), and rabbit polyclonal NG2 chondroitin sulfate proteoglycan (NG2) antibody (1:200; Chemicon International Inc., Temecula, CA). Biotinylated Ricinus communis-agglutinin-1 (RCA-1) (50 μg/ml) was purchased from Vector Laboratories (Burlingane, CA).

Brains from twi/twi and +/+ mice killed at PND 20, 30, and 40 (n = 3 for each timing period) were immunostained for TNFα. The mice were perfused with cold physiological saline under deep inhalation anesthesia with sevoflurane, and the isolated brains were quickly frozen in liquid nitrogen. For routine histochemical staining, mice (n = 3 for each groups) were perfused with physiological saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4). The brain was removed, postfixed and embedded in paraffin blocks. Luxol fast blue (LFB)-periodic acid Schiff (PAS) staining was performed on the paraffin sections of twi/twi and +/+ at PND 40 for evaluation of neuropathology.

For the determination of mRNA levels, groups of twi/twi and +/+ (n = 3 each timing period) were killed at PND 20, 30, and 40 under appropriate anesthesia. The brains were then removed, divided into the cerebrum and cerebellum/brain stem, and quickly frozen in liquid nitrogen.

Frozen sections were fixed at 4°C in acetone and incubated with PE-conjugated rat anti-mouse TNFα antibody for 48 h. For double labeling with RCA-1 and anti-TNFα, TNFα-stained sections were reacted with biotinylated RCA-1 for 30 min at room temperature, and then with avidin-D-fluorescein isothiocyanate isomer (avidin-FITC; Vector Laboratories), diluted 1:1000 with PBS, for 30 min. For NG2 immunostaining, after blocking with 0.3% Triton-X100 for 1 h, frozen sections were incubated with anti-NG2 antibody for 12 h at 4°C, and incubated with Alexa 488-conjugated anti-rabbit IgG (H+L) (1:400; Molecular Probes, Inc., Eugene, OR) for 2 h.

Paraffin sections were used for immunostaining for MBP and pi-GST, and terminal deoxynucleotidyltransferase (TdT)-mediated dUTP nick end labeling (TUNEL). For immunocytochemistry, sections on glass slides were incubated serially with mouse anti-MBP or rabbit anti-pi-GST antibody, biotinylated goat anti-mouse or anti-rabbit immunoglobulins (Vector Laboratories), and avidin-biotin complex by using an ABC elite kit (ABC; Vector Laboratories). Immunoreactions were visualized by immersing the slides in a 0.03% H₂O₂ solution in 50 mM Tris-HCl (pH 7.6) containing 0.05% diaminobenzidine tetrahydrochloride (DAB) and 0.25% nickel ammonium sulfate. Twi/twi and +/+ at PND 40 were subjected to TUNEL staining. Nuclei with DNA fragmentation were detected by using an in situ apoptosis detection kit (Takara Biomedicals, Osaka, Japan). Briefly, after pretreatment with 0.1% trypsin for 15 min at 37°C, sections were reacted with TdT, dNTPs, and FITC-labeled dUTP for 90 min at 37°C, followed by horseradish peroxidase (HRP)-conjugated anti-FITC antibody overnight at 4°C. The immunoproduct was visualized with the same protocol described above.

To identify the type of TUNEL-positive cells, we combined the staining for pi-GST, GFAP and RCA-1 with the TUNEL procedure. After TUNEL staining, sections were incubated with PBS containing 0.3% TritonX-100 and 10% normal goat serum for 30 min and then with rabbit anti rat-pi GST antibody, rabbit anti-cow GFAP antibody or biotinylated RCA-1 at 4°C overnight. The procedures were basically the same as described above except for the use of ABC-alkaline phosphatase and naphthol AS-BI phosphate coupled with hexazotized new fuchsin (Merck, Darmstadt, Germany) as a chromogen.

Total RNA was isolated from the quick-frozen brains with Isogen (Nippon gene, Toyama, Japan). The random 9-mers-primed cDNA was prepared with an RNA-LA-PCR Kit (Takara Shuzo, Kyoto, Japan) and 2 μg of total RNA.

A LightCycler PCR and detection system (Roche Diagnosis, Mannheim, Germany) was used for the amplification and quantification of mRNA for TNFα, TNFR1, TNFR2 and glycerol aldehyde-3-phosphate dehydrogenase (G3PDH) as previously described 18. G3PDH served as an internal control. The sequence-specific primers used were as follow: TNFα forward primer: 5'-AGTGACAAGCCTGTAGCCCACG-3', TNFα reverse primer: 5'-TTTCTCCTGGTATGAGATAGC-3', TNFR1 forward primer: 5'-CTAAACAGCAGAACCGAGTGT-3', TNFR1 reverse primer: 5'-AGATACGTAGAGTGTCCTTGG-3', TNFR2 forward primer: 5'-ATAAAGCCACACCCACAACCT-3', TNFR2 reverse primer: 5'-CATCTCCCTGCCACTCACAA-3', G3PDH forward primer: 5'-TGAACGGGAAGCTCACTGG-3', and G3PDH reverse primer: 5'-TCCACCACCCTGTTGCTGTA-3'. The constructs, used to create a standard curve, were made by cloning each amplified fragment into the Hind III site of a pGEM vector (Promega, Madison, WI). The number of copies was calculated by plotting a dilution series on this standard curve in each PCR experiment. For amplification detection, the LightCycler DNA Master Hybridization Probes Kit was used. Quantification of TNFα mRNA was performed by conducting 50 cycles of repeated denaturation (1 s at 89°C), annealing (5 s at 58°C), and enzymatic chain extension (10 s at 72°C). The PCR amplification conditions for G3PDH were 40 cycles of repeated denaturation (1 s at 87°C), annealing (5 s at 57°C), and enzymatic chain extension (10 s at 72°C). Quantification of TNFR1 and TNFR2 mRNAs was made by using 50 cycles of repeated denaturation (1 s at 89°C), annealing (5 s at 58°C), and enzymatic chain extension (10 s at 72°C). Duplicated PCR products were evaluated by melting curve analysis.

Ibudilast was a generous gift from Kyorin Pharmaceutical Co. Ltd. (Tokyo, Japan). After dissolved to a concentration of 1 mg/ml in physiological saline containing 10% v/v of polyoxyethylene hydrogenated castor oil 60 (HCO60), ibudilast (10 mg/kg) was injected intraperitoneally daily into three twi/twi from PND 15 to PND 40, and five twi/twi from PND 30 to PND 45. For controls, the same volume of HCO 60 was injected into two twi/twi from PND 15 to PND 40 and four twi/twi from PND 30 to PND 45. The density of TUNEL-positive cells in the demyelinating lesion in twi/twi, treated from PND 30 to PND 45 was calculated by using MacSCOPE software (Mitani Co, Fukui, Japan). Two independent neuropathologists examined the LFB-PAS-stained coronal sections (four sections per mouse) at the level of the optic chiasm and at the cerebellopontine angles containing the paraflocculus in a double-blind manner and scored the severity of demyelination from 0 to 5. 0: no demyelination, 1: slight demyelination, 2: less than 25% of the areas are occupied by a demyelination focus, 3: 25% ~ 50% of the areas occupied, 4: 50 ~ 75% of the areas occupied, 5: more than 75% of the areas occupied. The scores were average of two examiners' evaluations.

The cDNA probe for TNFα comprised a 268-bp PCR fragment (forward primer; 5'-GATGGGTTGTACCTTGTCTACTCC-3' and reverse primer; 5'-CTAAGTACTTGGGCAGATTGACCT-3') from the mouse TNFα, and was subcloned into a pGEM-T Easy vector (Promega, Madison, Wisconsin). In situ hybridization was carried out by using manual capillary action technology with a Microprobe staining system (Fisher Scientific International, Hampton, NH) as previously described 1920. First, brain sections (10 μm) were deparaffinized with Auto Dewaxer (Research Genetics, Huntsville, AL). The sections were rinsed in Auto Alcohol, Universal Buffer, and Immuno/DNA buffer (Research Genetics). Predigestion by proteinase K (15 μg/ml; Sigma-Aldrich, St Louis, MO) was performed to increase the tissue penetration of the probe. After this digestion, the tissue sections were treated with Immuno/DNA buffer. The DIG-labeled cRNA probe was diluted to 0.5 μg/ml with Brigati probe diluent (Research Genetics), 50% deionized formamide, and 50% dextran sulfate. The probe solution was heated at 90°C to denature the cRNA structures and applied to the slides. The hybridization of tissue and probe was done at 50°C for three hours. After hybridization, the slides were washed in 2 × SSC containing nonionic detergent. The detection of the DIG-labeled RNA was performed by using the Genius DNA labeling and detection kit (Roche Diagnostics). For counterstaining, neutral red was applied.

Student's t test was performed by using Stat View software (SAS Institute, Cary, NC). p< 0.05 was considered as significant.

The level of TNFα mRNA was the same in both cerebellum and cerebrum of the +/+ at any age examined. In the cerebrum, the level of TNFα-mRNA in twi/twi was almost the same as that in +/+ until PND 30, however, it increased to become approximately 15 times higher at PND 40 than that of +/+. In the cerebellum, there was no difference in the TNFα mRNA level between twi/twi and +/+ at PND 20, however, its level increased significantly in twi/twi after PND 30, becoming 40 times higher in twi/twi than +/+ at PND 40 (Fig. 1A).

Next, we investigated the levels of TNFR1 and TNFR2. In the +/+ cerebellum, the level of TNFR1 mRNA was constant throughout all the ages examined, whereas in the twi/twi cerebellum, it significantly increased with the progression of demyelination, becoming 50 times higher than that in +/+ at PND 40. In contrast, mRNA for TNFR2 increased in twi/twi only after PND 40, when compared with that for +/+ (Fig. 1B).

Immunocytochemical analysis revealed that TNFα-immunoreactive cells were not recognized at PND 20 (Fig. 1C) in twi/twi. However, many TNFα-immunoreactive cells were found in the cerebral white matter, brain stem and cerebellar white matter (CWM) at PND 30 (Fig. 1D) and 40 (Fig. 1E). On the other hand, TNFα-immunoreactive cells were not detected anywhere in the +/+brain even at PND 40 (Fig. 1F). These data were compatible with the data of the quantitative RT-PCR.

The morphological characteristics of TNFα-positive cells were an irregular cellular contour and lack of delicate processes, reminiscent of ameboid microglia/macrophages. Furthermore, TNFα-positive cells were positive for RCA-1, a marker for macrophage (arrows in Fig. 2A), but negative for pi-GST, a marker for OLs, or GFAP, a marker for astrocytes (data not shown), confirming those cells to be microglia/macrophages. In the twi/twi brain, both TNFα-positive cells and TUNEL-positive cells were most abundant in the CWM (Fig. 2B, C) and in the spinal trigeminal tract (sp5) in the superior midbrain (Fig. 2E, F). The majority of TUNEL-positive cells were also positive for pi-GST (arrowheads in Fig. 2C, F, I), identifying them as OLs (inset in Fig. 2C). These lesions of the cerebellum were most severely demyelinated judged by MBP immunostaining (Fig. 2D, G). In contrast, in the corpus callosum, where demyelination was milder than in the cerebellum, only a few TNFα-positive cells were detected (Fig. 2H – J).

To investigate whether the inflammatory response in microglia/macrophages contributes to the demyelination in twi/twi, we administered a phosphodiesterase inhibitor, ibudilast, to twi/twi. Two out of five twi/twi treated from PND 30 revealed strikingly milder clinical symptoms (Fig. 3A). Even at PND 45, two of ibudilast-treated twi/twi from PND 30 could move smoothly despite mild hindlimb paralysis, and showed less severe tremor and ataxia than vehicle-treated twi/twi. These mice were bigger than vehicle-treated twi/twi, as they had less weight loss (Fig. 3B). In contrast, ibudilast-treated twi/twi from PND 15 showed neither apparent clinical improvement nor elongation of lifespan, however, their body weights were heavier than those of vehicle-treated twi/twi.

The signal for TNFα mRNA obtained by in situ hybridization was recognized in the cells with small nuclei in the CWM and sp5 of vehicle-treated twi/twi (inset in Fig. 4A), corresponding to the presence of TNFα-immunoreactivity in the microglia. This signal was significantly reduced in the ibudilast-treated twi/twi (Fig. 4B, D). The number of TUNEL-positive cells was decreased in the CWM in ibudilast-treated twi/twi (Fig. 4F, H) compared with that of the vehicle-treated mice (Fig. 4E, G). TUNEL-positive cells were decreased in other regions such as the 8^th nerve (8 n) and sp5 in ibudilast-treated twi/twi than in vehicle-treated mice (Fig. 5, the upper bar graph).

LFB-PAS staining revealed that the demyelination was remarkably suppressed in the ibudilast-treated mice from PND 30 (Fig. 4J, L) compared with the vehicle-treated ones (Fig. 4I, K), as shown in the score of demyelination (Fig. 5, lower bar graph). From these lines of evidence, we concluded that the demyelination and clinical symptoms were reduced with inhibition of TNFα in twi/twi.