Minocycline reduces remyelination by suppressing ciliary neurotrophic factor expression after cuprizone-induced demyelination

Authors

  • Tatsuhide Tanaka,

    Corresponding author
    1. Department of Functional Anatomy and Neuroscience, Asahikawa Medical University, Asahikawa, Hokkaido, Japan
    • Address correspondence and reprint requests to Tatsuhide Tanaka, Department of Functional Anatomy and Neuroscience, Asahikawa Medical University, Midorigaoka-higashi 2-1-1-1 Asahikawa, Hokkaido 078-8510, Japan. E-mail: ttanaka@asahikawa-med.ac.jp

    Search for more papers by this author
  • Koichi Murakami,

    1. Department of Functional Anatomy and Neuroscience, Asahikawa Medical University, Asahikawa, Hokkaido, Japan
    Search for more papers by this author
  • Yoshio Bando,

    1. Department of Functional Anatomy and Neuroscience, Asahikawa Medical University, Asahikawa, Hokkaido, Japan
    Search for more papers by this author
  • Shigetaka Yoshida

    1. Department of Functional Anatomy and Neuroscience, Asahikawa Medical University, Asahikawa, Hokkaido, Japan
    Search for more papers by this author

Abstract

Remyelination is disrupted in demyelinating diseases such as multiple sclerosis, but the underlying pathogenetic mechanisms are unclear. In this study, we employed the murine cuprizone model of demyelination, in which remyelination occurs after removal of the toxin from the diet, to examine the cellular and molecular changes during demyelination and remyelination. Microglia accumulated in the corpus callosum during weeks 2–4 of the cuprizone diet, and these cells remained activated 2 weeks after the change to the normal diet. To examine the role of microglia in remyelination, mice were treated with minocycline to inactivate these cells after cuprizone-induced demyelination. Minocycline treatment reduced the number of CC1-positive oligodendrocytes, as well as levels of myelin basic protein (MBP) and CNPase in the remyelination phase. The expression of CNTF mRNA in the corpus callosum increased after 4 weeks on the cuprizone diet and remained high 2 weeks after the change to the normal diet. Minocycline suppressed CNTF expression during the remyelination phase on the normal diet. Primary culture experiments showed that CNTF was produced by microglia in addition to astrocytes. In vitro, CNTF directly affected the differentiation of oligodendrocytic cells. These findings suggest that minocycline reduces remyelination by suppressing CNTF expression by microglia after cuprizone-induced demyelination.

Abbreviations used
CNTF

ciliary neurotrophic factor

CNTFR

CNTF receptor α

DMEM

dulbecco's modified Eagle's medium

IGF-1

insulin-like growth factor 1

LIFR

leukemia inhibitory factor receptor

LPS

lipopolysaccharide

MS

multiple sclerosis

PB

phosphate buffered

Oligodendrocytes and myelin play essential roles in the vertebrate CNS. Demyelination, associated with inherited and acquired diseases such as multiple sclerosis (MS), disrupts saltatory nerve conduction, leading to axonal degeneration and neurological disabilities (Dutta and Trapp 2011). Remyelination is a regenerative process in which new myelin sheaths are formed on demyelinated axons. However, this regenerative process is limited in MS, owing in part to the failure of adult oligodendrocyte precursor cells to differentiate into myelinating oligodendrocytes (Franklin 2002; Kuhlmann et al. 2008). For this reason, enhancing remyelination is an attractive therapeutic strategy for preventing axonal loss in the later stages of MS (Dubois-Dalcq et al. 2005; Miller and Mi 2007).

Cuprizone has been frequently used to study experimental demyelination and remyelination (Matsushima and Morell 2001). Cuprizone feeding leads to oligodendrocytic apoptosis and subsequent reversible demyelination in the corpus callosum. The reliable and nearly complete remyelination in this model allows researchers to study the mechanisms that underlie successful regeneration. When significant demyelination is detected in this model, an increased number of microglia and astrocytes are observed within the lesion (Matsushima and Morell 2001), indicating that these glial cells play a role in demyelination and remyelination.

Microglia, generally considered immune cells of the CNS, are involved in many types of inflammatory processes in the brain (Hanisch and Kettenmann 2007; David and Kroner 2011). Microglia have also been proposed to be involved in the initial stage of the pathogenetic process of MS (Van der Valk and De Groot 2000). However, Olah and colleagues have provided evidence for a role of microglia in remyelination from the onset of demyelination and during the remyelination process (Olah et al. 2012). Thus, there has been considerable debate in the field as to whether the microglial response is beneficial or detrimental for tissue protection and repair (David and Kroner 2011). It is possible that the properties of activated microglia differ between the demyelination and remyelination phases. Several studies have demonstrated that inhibition of microglial activation with minocycline prevents cuprizone-induced demyelination. These reports indicate that microglia play an active role in oligodendrocytic cell death and demyelination (Pasquini et al. 2007; Skripuletz et al. 2010). However, microglia are also known to support cell survival during tissue repair following injury to the CNS (David and Kroner 2011).

A number of cytokines and neurotrophic factors support the proliferation and differentiation of glial cells and regulate myelination (Althaus et al. 2008; Schmitz and Chew 2008). Indeed, the expression of many cytokines and neurotrophic factors are elevated during demyelination and remyelination induced by cuprizone feeding (Matsushima and Morell 2001; Gudi et al. 2011). Ciliary neurotrophic factor (CNTF) was originally discovered as a survival factor for chick ciliary neurons (Stöckli et al. 1989). CNTF has been shown to be produced by astrocytes following brain injury and promotes myelination and the survival of a variety of neuronal populations (Moore et al. 2011). However, the function of microglial-derived CNTF in remyelination after cuprizone-induced demyelination has not been examined. In this study, we investigated the role of activated microglia and CNTF in remyelination.

Materials and methods

Mice and experimental induction of demyelination

Female 8- to 10-week-old C57BL/6 mice were obtained from the Institute of Asahikawa Medical University. Demyelination was induced by feeding a diet containing 0.2% cuprizone (bis-cyclohexanone oxaldihydrazone; Sigma, St. Louis, MO, USA) mixed into a ground standard rodent chow. For remyelination, animals were returned to the normal diet after cuprizone treatment. Age-matched control mice were maintained on the same diet without cuprizone.

Minocycline treatment

Mice were injected intraperitoneally (i.P.) with minocycline (Sigma) or vehicle (phosphate-buffered saline) from the first day of normal diet after cuprizone feeding (Fig. 3a). Mice received 100 mg/kg minocycline twice a day for the first 2 days and once a day for the next 5 days, followed by 50 mg/kg minocycline for an additional 7 days until the mice were killed.

Cell culture

Primary cultures of microglial cells were prepared from C57BL/6 mice on post natal day 1 (P1) and cultured as described previously (Tanaka et al. 2009). To examine the effects of minocycline on microglia, cells were treated with minocycline (5 μg/mL) along with lipopolysaccharide (LPS) (200 ng/mL). For examination of CNTF secretion from microglia, the culture medium was changed to Opti-MEM (Life Technologies, Grand Island, NY, USA). The CNTF protein secreted into the culture medium was concentrated with trichloroacetic acid. CG-4-16 cells were kindly provided by Dr. T. Ogata (National Rehabilitation Center) and maintained in proliferation medium [Dulbecco's modified Eagle's medium (DMEM) and Ham's F12-containing 2% fatal calf serum (FCS), 5 μg/mL insulin, 16.1 μg/mL putrescine, 50 μg/mL apo-transferrin, 4.6 μg/mL d-galactose, and 8 ng/mL sodium selenite]. To promote differentiation, cells were seeded onto poly-D-lysine-coated dishes. After attachment, cells were gently washed and subsequently cultured in differentiation medium (DMEM and Ham's F12-containing 0.5% FCS, 5 μg/mL insulin, 16.1 μg/mL putrescine, 50 μg/mL apo-transferrin, 4.6 μg/mL d-galactose, and 8 ng/mL sodium selenite). To examine the effects of minocycline on CG4-16, cells were treated with minocycline (5 μg/mL) along with staurosporine (1 μM). Primary cultures of oligodendrocytes were prepared from C57BL/6 mice on E19. Briefly, mouse cerebrum was digested with 0.25% trypsin and 70 U DNase for 15 min, and passed through a 70-μm nylon mesh. The resultant cell suspension was diluted with DMEM supplemented with 10% fetal bovine serum, 50 units/mL penicillin and 50 μg/mL streptomycin, and seeded onto flasks. After 7–10 days of culture, the flasks were shaken by hand, which detached oligodendrocyte progenitor cells into the medium. The medium from flasks was spun and the resulting pellets were re-suspended in DMEM and Ham's F12-containing 0.5% FCS, 5 μg/mL insulin, 16.1 μg/mL putrescine, 50 μg/mL apo-transferrin, 4.6 μg/mL d-galactose, and 8 ng/mL sodium selenite.

Reagents for cell culture

Reagents used in this study were as follows: (LPS, Sigma), recombinant human CNTF (Peprotech, Rocky Hill, NJ, USA), minocycline (Sigma), and staurosporine (Sigma).

Tissue preparation

Mice were anesthetized with sodium pentobarbital (100 mg/kg, i.P.) and perfused transcardially with saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4). Brains were removed, post-fixed overnight in the same fixative, and then immersed in 30% sucrose in 0.1 M PB overnight. Brains were then frozen in powdered dry ice, embedded in Tissue-Tek OCT compound (Sakura Finetek, Torrance, CA, USA) and stored at −80°C prior to sectioning. Frozen 20-μm transverse sections of the brain were cut on a cryostat and mounted onto glass slides.

Immunohistochemistry

The sections were immersed in 0.1 M PB-containing 5% bovine serum albumin (BSA) and 0.3% Triton X-100 for 1 h. Anti-adenomatous polyposis coli (1 : 250; APC or CC1, Merck Millipore, Billerica, MA, USA), anti-GST-pi (1 : 250; BD Transduction Laboratories, Franklin Lakes, NJ, USA), anti-Iba1 (1 : 500; Wako, Osaka, Japan), anti-CD68 (1 : 100; AbD Serotec, Raleigh, NC, USA), and anti-CNTF (1 : 100; Millipore) antibodies were applied overnight at 4°C. Fluorescent dye Alexa Fluor 488-conjugated anti-rabbit IgG (1 : 1000; Life Technologies) or Alexa Fluor 568-conjugated anti-mouse IgG (1 : 1000; Life technologies) was used as the secondary antibody. The images were obtained on a confocal laser scanning microscope with Fluoview FV1000-D software (Olympus, Tokyo, Japan). The corpus callosum, between +1.10 mm and −0.22 mm of the bregma, was used in this study. Fluoview FV1000-D software was used to count the number of CC1-positive cells and to measure CD68 intensity in confocal images.

Immunocytochemistry

Cultured cells were fixed with 4% paraformaldehyde in 0.1 M PB for 1 h at 20°C and incubated with blocking solution containing 1% BSA and 0.1% Triton X-100 in phosphate-buffered saline for 1 h. Primary antibodies, anti-CNTF (1 : 100; Millipore), anti-Iba1 (1 : 500; Wako), and anti-CNTF receptor α (CNTFRα) (1 : 100; Santa Cruz, Santa Cruz, CA, USA) were applied overnight at 4°C. Fluorescent dye Alexa Fluor 488-conjugated anti-rabbit IgG (1 : 1000; Life technologies) or Alexa Fluor 568-conjugated anti-mouse IgG (1 : 1000, Life technologies) was used as the secondary antibody.

Western blotting

Samples were prepared as described previously (Tanaka et al. 2009). Protein concentration was measured using a bicinchoninic acid protein assay kit (Thermo Scientific, Rockford, IL, USA). Equal amounts of protein were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then transferred to a polyvinylidene difluoride membrane (Millipore). The blots were probed with anti-myelin basic protein (MBP) (1 : 10 000; Abcam, Cambridge, MA, USA), anti-2′,3′-cyclic nucleotide 3′ phosphodiesterase (CNPase) (1 : 1000; Sigma), anti-olig1 (1 : 1000; Millipore), anti-CNTF (1:1000; Millipore) and anti-actin (1 : 500; Santa Cruz) antibodies. Immunoreactive bands were visualized using horseradish peroxidase-conjugated anti-mouse and anti-rabbit IgG using enhanced chemiluminescence western blotting detection reagents (GE Healthcare, Piscataway, NJ, USA). Data were acquired in arbitrary densitometric units using Scion image software and transformed to percentages of densitometric levels obtained from scans of control samples visualized on the same blot. The density of actin bands was used to standardize the density of myelin-related proteins.

Rt-pcr

Total RNA was extracted using TRIzol (Life Technologies) and converted to cDNA with reverse transcriptase using oligo(dT)20 primers to prime AMV reverse transcriptase (Promega, Madison, WI, USA) according to the manufacturer's instructions. Specific DNAs were mixed and amplified with 0.5 μM of each PCR primer and cDNA. The following primers were used in this study: brain-derived neurotrophic factor (BDNF) sense primer, 5′-GCGGCAGATAAAAAGACTGC-3′; BDNF anti-sense primer, 5′-CTTATGAATCGCCAGCCAAT-3′; CD68 sense primer, 5′-TAGGACCGCTTATAGCCCAAGGA-3′; CD68 anti-sense primer, 5′-AGCAGTGCCATTTGTGGTGGGA-3′; CNPase sense primer, 5′-CCGGACACATAGTACCCGC-3′; CNPase anti-sense primer, 5′-TACGCCTCGGGAGAAGTCTG-3′; CNTF sense primer, 5′-TCGCAGAGCAATCACCTCTGAC-3′; CNTF anti-sense primer, 5′-ACATCCCTTGGAAGGTACGGT-3′; CNTFRα sense primer, 5′-TGGCAGCATGTGGAGCTCTCGGAT-3′; CNTFRα anti-sense primer, 5′-AGTATGGAGGGTCCTCCGCCGCT-3′; glyceraldehyde 3-phosphate dehydrogenase (GAPDH) sense primer, 5′-CTACATGGTCTACCTGTTCCAG-3′; GAPDH anti-sense primer, 5′-AGTTGTCATGGATGACCTTGG-3′; gp130 sense primer, 5′-TGAGGAGGCTAGTGGGACCACAT-3′; gp130 anti-sense primer, 5′-ACGGTCAGCTCTGTGCCAGTGA-3′; IGF-1 sense primer, 5′-ACAGCTGGACCAGAGACCCTTTGC-3′; IGF-1 anti-sense primer, 5′-GGCTTCAGTGGGGCACAGTA-3′; IL-1β sense primer, 5′-CTCCATGAGCTTTGTACAAGG-3′; IL-1β anti-sense primer, 5′-TGCTGATGTACCAGTTGGGG-3′; LIFR sense primer, 5′-GCATTTCAGGAAAATCGGCA-3′; LIFR anti-sense primer, 5′-CCTCTGATTGTGGCATTCCT-3′; MBP sense primer, 5′-ACTCACACACACGAGAACTACCCA-3′; MBP anti-sense primer, 5′-CCAGCTAAATCTGCTGAGGG-3′; PLP sense primer, 5′-GTTGTATGGCTCCTGGTGTTTGC-3′; PLP anti-sense primer, 5′-ACGGCGAAGTTGTAAGTGGCAGC-3′. PCR cycling conditions were 5 min at 96°C, then multiple cycles of 30 s at 96°C, 30 s at 60–63°C, and 30 s at 72°C. PCR products were separated by electrophoresis on 1.5% agarose gels and visualized using ethidium bromide. Real-time quantitative PCR was performed following the protocol in the Light Cycler SYBR Green kit (Roche, Mannheim, Germany).

Statistical analysis

Statistical analyses were performed using Student's t-test or one-way anova, and differences between treatment means were determined with the Tukey–Kramer test.

Results

Demyelination and spontaneous remyelination in the corpus callosum during cuprizone treatment

We used cuprizone to induce demyelination in C57BL/6 mice. A primary target of demyelination in this model is the corpus callosum. Although we confirmed that demyelination was induced by 6 weeks of the cuprizone diet (Figure S1), it is well known that partial spontaneous remyelination occurs between 3 and 6 weeks during acute demyelination induced by cuprizone (Matsushima and Morell 2001; Kipp et al. 2009). Therefore, it was necessary to examine in detail the time course of expression of myelin-related molecules to better understand the molecular mechanisms of demyelination and remyelination. qRT-PCR analysis demonstrated that mRNA levels of MBP and CNPase were decreased in the corpus callosum at 2 weeks of the cuprizone diet, and then increased afterward (Fig. 1a and b). Immunohistochemistry revealed that the number of CC1-positive oligodendrocytes were decreased at 2 weeks of the cuprizone diet to 38.7% of the pre-diet amount, and reached the lowest number, 17.4%, at 4 weeks (Fig. 1d–f and i). Thereafter, no further decrease was observed in our model (Fig. 1g–i). Similar results were obtained by GST-pi-staining (Fig. 1j–o). Therefore, we gave cuprizone for 4 weeks, and allowed remyelination afterward by removing cuprizone from the diet.

Figure 1.

Cuprizone feeding depletes oligodendrocytes from the corpus callosum at week 4, and afterward, spontaneous remyelination occurs. (a, b) Expression profile of MBP and CNPase mRNAs in the corpus callosum of mice during demyelination. Data are presented as the mean ± SEM of three to five independent experiments. *< 0.05, **< 0.01 compared with control. (c) Images (d–h and j–n) were taken from the red-boxed area. (d–h) Time course of CC1 (green) expression merged with DAPI (blue) in the corpus callosum. Arrows indicate CC1-positive cells. Scale bar, 50 μm. (i) Quantification of the density of CC1-positive cells in the corpus callosum at the indicated weeks. (j–n) Time course of GST-pi (green) expression merged with DAPI (blue) in the corpus callosum. Arrows indicate GST-pi-positive cells. Scale bar, 50 μm. (o) Quantification of the density of GST-pi-positive cells in the corpus callosum at the indicated weeks. Data are presented as mean ± SEM of three to five independent experiments. *< 0.05, **< 0.01.

Microglia accumulate in the corpus callosum during demyelination and remyelination

Cuprizone-induced demyelination is accompanied by the accumulation of microglia (Matsushima and Morell 2001). Indeed, in our cuprizone model, microglia accumulated within the first 2 weeks after exposure to cuprizone (Fig. 2b and e). When demyelination was at its peak, an increased number of microglia were observed in the lesion (Fig. 2c and e). Interestingly, the number of microglia did not decrease, even after we induced remyelination by removal of cuprizone (Fig. 2d and e), suggesting that the microglia remained in the corpus callosum during the remyelination phase. We therefore investigated the function of activated microglia in remyelination.

Figure 2.

Microglia accumulate at the lesion site. (a–d) Confocal images of the corpus callosum of cuprizone-treated mice stained for Iba1 (red) and DAPI (blue). Scale bar, 50 μm. (e) Quantification of the density of Iba1-positive cells in the corpus callosum at the indicated weeks. Data are presented as the mean ± SEM of three to five independent experiments. **< 0.01 compared with control.

We next examined the expression profiles of cytokines and neurotrophic factors during cuprizone-induced demyelination and remyelination. Recent studies indicate that interleukin-1β (IL-1β), insulin-like growth factor 1 (IGF-1) and CNTF increase in the corpus callosum in cuprizone-induced demyelination (Mason et al. 2001; Gudi et al. 2011). Therefore, we first examined the expression levels of these factors in our model. IL-1β and IGF-1 levels increased at 4 weeks of the cuprizone diet, but were reduced during remyelination (Figure S2a and b). In contrast, BDNF mRNA expression did not change at 4 weeks of cuprizone treatment (Figure S2d). CNTF mRNA levels increased during the demyelination phase and remained high during the middle phase of remyelination (Figure S2c). Thus, temporal differences in the expression of these factors were observed in the cuprizone-induced demyelination model, indicating that the expression of certain cytokines and neurotrophic factors are dynamically regulated during demyelination and remyelination.

Minocycline reduces levels of myelin-related factors

To examine the role of microglia, we treated mice with minocycline to inactivate microglia. As a preliminary experiment, we compared the effects of 100 mg/kg and 50 mg/kg minocycline, and concluded that the 100 mg/kg dose is necessary to suppress microglial activation in the corpus callosum (Figure S3). Indeed, there have been studies which used 100 mg/kg minocycline in mice (Herrmann et al. 2007; Orazizadeh et al. 2009). We therefore administered 100 mg/kg minocycline daily for the first 7 days from the beginning of the normal diet after cuprizone feeding, followed by 50 mg/kg minocycline for an additional 7 days until the mice were killed (Figure 3a). First, we confirmed the effect of minocycline on microglial inactivation in the corpus callosum. Immunohistochemical analysis revealed reduced expression of CD68, a marker of activated microglia/macrophages, in minocycline-treated mice compared with vehicle-treated mice (58.3 ± 4.3% of control), suggesting that minocycline inhibited microglial activation. (Fig. 3b and c). qRT-PCR showed that CD68 mRNA levels tended to be lower in minocycline-treated mice (55.3 ± 11.8% of vehicle-treated control) (Fig. 3d). However, minocycline did not change microglial morphology in the corpus callosum (Fig. 3b). We further examined whether the minocycline directly affects on cells other than microglia. First, to determine the effect of minocycline on astrocytes in vitro, we treated cultured astrocytes with minocycline. LPS-induced inducible nitric oxide synthase (iNOS) mRNA expression was reduced by minocycline at 500 μg/mL, but not at 200 μg/mL (Figure S4a). In contrast, 200 μg/mL minocycline had a significant effect on LPS-induced iNOS mRNA expression in microglia (Figure S4b), suggesting that astrocytes are less sensitive to minocycline than microglia.

Figure 3.

Microglial response and remyelination are reduced by minocycline administration in the remyelination phase. (a) Diagram of the scheme of saline and minocycline administration. (b) Confocal images of the corpus callosum of vehicle and minocycline-treated mice stained for CD68 (green) and Iba1 (red). Scale bar, 50 μm. (c) Quantification of the intensity of CD68-positive cells in the corpus callosum of vehicle and minocycline-treated mice. Data are presented as the mean ± SEM of five independent experiments. **< 0.01. (d) Expression level of CD68 mRNA in the corpus callosum of mice treated with minocycline. Data are presented as the mean ± SEM of five independent experiments. (e) Confocal images of the corpus callosum of vehicle and minocycline-treated mice stained for CC1 (green). Scale bar, 50 μm. (f) Quantification of the density of CC1-positive cells in the corpus callosum of vehicle and minocycline-treated mice. Data are presented as the mean ± SEM of five independent experiments. *< 0.05. (g, h): Expression profiles of mRNAs for myelin-related factors in the corpus callosum of mice treated with minocycline (mino) or vehicle (veh). Data are presented as the mean ± SEM of three to five independent experiments. *< 0.05, **< 0.01.

We next compared the number of oligodendrocytes and the expression of myelin-related proteins in the corpus callosum of minocycline and vehicle-treated mice. Minocycline treatment resulted in a significant decrease in the number of CC1-positive oligodendrocytes (204.4 ± 16.0/mm2) compared with mice treated with vehicle (258.9 ± 21.2/mm2; = 0.046) (Fig. 3e and f). As shown in Fig. 3g and h, 4 weeks of cuprizone diet decreased mRNA levels of MBP and CNPase. One week after the return to the normal diet, the expression of these mRNAs recovered in vehicle-treated mice. Minocycline treatment resulted in a significant inhibition of recovery of MBP (minocycline, 48.4 ± 6.2%; vehicle, 73.2 ± 9.9%) and CNPase (minocycline, 45.4 ± 10.6%; vehicle, 88.2 ± 7.0%) expression. Two-week administration of minocycline also inhibited recovery of the expression of MBP (minocycline, 84.1 ± 9.7%; vehicle, 119.3 ± 4.3%) and CNPase (minocycline, 110.1 ± 11.3%; vehicle, 156.0 ± 3.9%) mRNAs (Fig. 3g and h). In contrast, MBP and CNPase mRNA levels in vehicle-treated mice were similar to or above those in healthy controls. These results indicate that minocycline inhibits remyelination.

Microglia have also been shown to be involved in the pathogenesis of MS (Van der Valk and De Groot 2000), and Defaux and colleagues showed that minocycline promotes remyelination in rat brain cell cultures after interferon-γ and LPS-induced demyelination (Defaux et al. 2011). To examine whether microglial activation is important in demyelination or remyelination in the cuprizone model, we used different schemes of minocycline administration (Figure S5). We administered minocycline for 2 weeks after 2 weeks of the cuprizone diet instead of after 4 weeks of the cuprizone diet (Figure S5a). The 2-week cuprizone diet decreased mRNA levels of MBP and CNPase, and 2 weeks after the return to the normal diet, the expression of these mRNAs recovered in vehicle-treated mice. There were no significant differences between vehicle-treated and minocycline-treated mice in the mRNA levels of MBP and CNPase (Figure S5b and c). We also investigated the effects of a 3-week administration of minocycline along with 3-week cuprizone treatment (Figure S5d). Minocycline treatment did not result in significant differences in MBP, CNPase or proteolipid protein (PLP) mRNA levels compared with mice treated with vehicle (Figure S5e–g). Therefore, we sought to further clarify the positive role of microglial activation in remyelination after a 4-week cuprizone treatment period.

Minocycline reduces CNTF expression during the remyelination phase

We next examined the effects of minocycline on the expression of cytokines and neurotrophic factors during the remyelination phase after the 4-week cuprizone diet. Minocycline treatment in the early remyelination phase reduced mRNA levels of IL-1β and IGF-1 (Fig. 4a and b). However, at 2 weeks of remyelination combined with a 2-week minocycline-treatment period, no significant reductions in mRNA levels of IL-1β, IGF-1, or BDNF were observed compared to vehicle-treated mice (Fig. 4a–c). In contrast, CNTF mRNA expression increased at 4 weeks of the cuprizone diet and remained high 2 weeks after the return to the normal diet without minocycline treatment. Minocycline treatment in the remyelination phase caused a significant reduction in CNTF mRNA levels at 1 week (cuprizone 4 w + normal diet 1 w: minocycline, 89.8 ± 19.1% of the amount before cuprizone diet; vehicle, 352.2 ± 63.4%; = 0.017) and 2 weeks (cuprizone 4 w + normal diet 2 w: minocycline, 165.5 ± 49.0% of the amount before cuprizone diet; vehicle, 396.1 ± 60.9%; = 0.042) (Fig. 4d) after the return to the normal diet. Furthermore, immunohistochemistry revealed a significant reduction in the number of CNTF-positive cells in the 2-week minocycline-treated mice compared with vehicle-treated mice (minocycline, 124.4 ± 15.8 mm2; vehicle, 203.3 ± 29.0 mm2; = 0.044) (Fig. 4e and f). These results show that the expression of CNTF was suppressed by minocycline during the remyelination phase.

Figure 4.

Ciliary neurotrophic factor (CNTF) expression is reduced by minocycline during the remyelination phase. (a–d) Expression profile of cytokine and neurotrophic factor mRNAs in the corpus callosum of mice treated with minocycline (mino) or vehicle (veh). Data are presented as the mean ± SEM of three to five independent experiments. *< 0.05, **p < 0.01. N.S, not significantly different. (e) Immunohistochemical analysis of CNTF in the corpus callosum of vehicle and minocycline-treated mice. Scale bar, 50 μm. (f) Quantification of the density of CNTF-positive cells in the corpus callosum of vehicle and minocycline-treated mice. Data are presented as the mean ± SEM of five independent experiments. *p < 0.05.

CNTF is produced by microglia in addition to astrocytes

Previous studies have shown that CNTF is produced by astrocytes following brain injury, and supports the survival of a variety of neuronal populations, as well as myelination (Albrecht et al. 2003; Moore et al. 2011). We examined whether CNTF was produced by microglia in addition to astrocytes. We confirmed that CNTF mRNA was expressed by primary microglial cultures as well as by astrocytes (Fig. 5a). Western blot analysis detected CNTF protein in the supernatant of primary microglial cultures, and primary cultured cells were stained by an anti-CNTF antibody (Fig. 5b and c). LPS increased CNTF mRNA expression in primary microglia, and minocycline treatment tended to reduce this increase (Fig. 5d). Immunohistochemistry also revealed that CNTF was expressed by microglia in the corpus callosum during the remyelination phase (Fig. 5e).

Figure 5.

Ciliary neurotrophic factor (CNTF) is produced by microglia in addition to astrocytes. (a) Expression level of CNTF mRNA in primary astrocytes (astro) and microglia (micro). Representative data for CNTF mRNA are shown. (b) Western blot analysis of secreted CNTF from primary microglial cells (micro). The cell culture medium was changed to Opti-MEM, and secreted protein in the culture medium was concentrated with trichloroacetic acid. Concentrated media were analyzed by immunoblotting using antibodies against CNTF. Sciatic nerve lysates were used as positive controls. (c): Representative immunocytochemical labeling of primary microglia (Iba1, red), CNTF (green), and DAPI (blue). (d) Expression of mRNA for CNTF in primary microglia treated with 5 μg/mL minocycline (mino). Cells were treated with minocycline along with 200 ng/mL lipopolysaccharide. Data are presented as the mean ± SEM of three independent experiments. (e) Confocal images of the corpus callosum at the remyelination phase stained for CNTF (green), Iba1 (red), and DAPI (blue). Scale bar, 50 μm.

CNTF induces MBP and PLP expression in an oligodendrocytic cell line

Next, we examined the expression of CNTFRα and its cofactors in the demyelination and remyelination phases of cuprizone treatment. Although CNTF mRNA expression was induced in the corpus callosum by the 4-week cuprizone treatment, the expression of CNTFRα mRNA was not changed in either the demyelination or remyelination phase (Fig. 6a). Similar results were obtained for 130-kDa glycoprotein (gp130) and leukemia inhibitory factor receptor (LIFR), cofactors for CNTFRα (Fig. 6a). Subsequently, we examined whether CNTF could directly affect the differentiation of oligodendrocytes. We used an oligodendrocytic cell line, CG-4-16, that transforms into mature oligodendrocytic cells when cultured in glial development medium-containing 0.5% FCS. qRT-PCR and immunocytochemistry demonstrated that CNTFRα mRNA and protein were expressed in both immature and mature cells (Fig. 6b and c). Daily addition of 10 ng/mL recombinant CNTF to the maturation medium for 4 days, but not for 2 days, increased levels of MBP and PLP transcripts (Fig. 6d and e). Interestingly, CNPase mRNA expression was unaffected by recombinant CNTF (Fig. 6f). We also performed primary cultures and observed an increase in MBP-positive mature oligodendrocytes typically bearing a complex morphology during differentiation in the presence of recombinant CNTF (Fig. 6g).

Figure 6.

Ciliary neurotrophic factor (CNTF) induces MBP and PLP mRNA expression in an oligodendrocytic cell line. (a) Expression profile of mRNAs for CNTF, CNTF receptor α (CNTFRα), and cofactors in the corpus callosum of mice during demyelination (4 w) and remyelination (n2 w, normal diet 2 weeks). Representative data are shown. (b) Expression of CNTFRα mRNA in the oligodendrocytic cell line CG-4-16. Representative data are shown. (c) Representative immunocytochemical labeling of CNTFRα (green), NG2 (red), and DAPI (blue) in CG-4-16 cells. (d–f) Expression profile of MBP (d), PLP (e) and CNPase (f) in CG-4-16 cells treated with recombinant CNTF (10 ng/mL) for 2, 4, and 6 days (filled circle) or no treatment (open circle). Data are presented as the mean ± SEM of three independent experiments. *< 0.05, **< 0.01. (g) Representative immunocytochemical labeling of MBP in primary oligodendrocytes treated with recombinant CNTF (10 ng/mL) for 4 days or no treatment.

Discussion

There has been considerable debate as to whether microglial activation is favorable or unfavorable for tissue protection and repair (David and Kroner 2011). Li and colleagues reported a positive role of microglial activation in remyelination following demyelination induced by focal injection of ethidium bromide (Li et al. 2005). Other studies have shown a harmful role of microglia in cuprizone-treated mice. In a number of studies, minocycline was injected at or soon after commencement of the cuprizone diet (Pasquini et al. 2007; Skripuletz et al. 2010). However, the experimental design of these studies made it difficult to clearly discriminate the function of microglia in the demyelination and remyelination phases. In this study, we focused on microglia during the remyelination stage to determine whether they are important for remyelination. Therefore, it was crucial to distinguish the remyelination phase from the demyelination phase. Consequently, we examined the time course of the effect of cuprizone, and found that at 4 weeks of the cuprizone diet, mature oligodendrocytes increased in number, suggesting that remyelination was occurring by that time. Thus, we fed mice cuprizone to induce demyelination for 4 weeks, and then changed to a normal diet to induce remyelination.

To examine the role of microglia during remyelination, we treated mice with minocycline to inactivate microglia after 4-week cuprizone treatment. We employed 100 mg/kg minocycline daily for the first 7 days beginning on the day of the normal diet after cuprizone feeding, followed by 50 mg/kg minocycline daily until the animals were killed. Skripuletz and co-workers reported that 50 mg/kg minocycline reduced microglial reactivity in the cortex but not in the corpus callosum (Skripuletz et al. 2010). A high concentration of minocycline may be necessary in the corpus callosum to inactivate microglia, as we show in Figure S3. Mice did not display behavioral abnormalities, but showed modest weight loss when treated with 100 mg/kg minocycline (data not shown). Minocycline treatment started after 4 weeks on the cuprizone diet resulted in a significant inhibition in recovery of MBP and CNPase expression (Fig. 3g and h). When minocycline treatment was started from the beginning of or 2 weeks into the cuprizone diet, inhibition of myelin-related molecules was not detected (Figure S5a–c). In our model, the number of microglia in the lesion at 2 weeks of cuprizone treatment was lower than that at 4 weeks of cuprizone treatment (Fig. 2). It is possible that minocycline administration at 2 weeks of cuprizone treatment was too early to affect remyelination, owing to the lower microglial accumulation. Thus, we conjectured that by inhibiting microglial activation during the remyelination phase, we could reduce the number of oligodendrocytes and hinder remyelination (Fig. 3). Our results are in good agreement with the reports by Franklin's group showing that microglial/macrophage activation is critical for remyelination (Kotter et al. 2005; Li et al. 2005). In contrast, Defaux and colleagues showed that minocycline promotes remyelination in rat brain cell cultures after interferon-γ and LPS-induced demyelination (Defaux et al. 2011). There were many differences between their methods and ours, which may account for the differing findings. In their study, minocycline was added 1 h prior to treatment with interferon-γ and LPS. Their experimental design made it difficult to differentiate the functions of microglia between the demyelination and remyelination phases. Microglial properties may be drastically altered between the demyelination and remyelination phases. One of the beneficial functions of microglia is the removal of myelin debris during the repair process (Kreutzberg 1996; Hanisch and Kettenmann 2007; David and Kroner 2011). Previously, we showed that the clearance of debris by microglia allowed axonal growth from severed neurites, suggesting that removal of debris helps provide a favorable environment for regeneration (Tanaka et al. 2009).

Cytokines and neurotrophic factors expressed by microglia may also be important in remyelination. Although IL-1β and IGF-1 expression was suppressed in the early phase of remyelination by minocycline, there were no obvious differences in IL-1β, IGF-1, or BDNF levels between the vehicle and 2-week minocycline-treated groups (Fig. 4a–c). In contrast, we found a significant reduction in CNTF expression both in 1-week and 2-week minocycline-treated mice (Fig. 4d–f). Therefore, it appears that IL-1β, IGF-1 and CNTF are important in the early phase of remyelination, and that CNTF plays a major role thereafter as well. Previous studies have shown that CNTF is produced by astrocytes following brain injury and supports the survival of a variety of neuronal populations, as well as myelination (Moore et al. 2011). As shown in Fig. 5a, both astrocytes and microglia have the potential to express CNTF mRNA. Cai and co-workers reported that minocycline reduced astrocytic reactivity in the hippocampus of a rat model of vascular cognitive impairment (Cai et al. 2010). Consequently, we treated cultured astrocytes with minocycline, and found that astrocytes were affected by minocycline to a much lower extent than microglia (Figure S4a–b). We further examined the effect of minocycline on astrocytes in vivo. It is known that demyelination is accompanied by the appearance of microglia and astrocytes in cuprizone model. In our model, astrocytes were accumulated within the lesion when significant demyelination was detected. The number of astrocytes did not decrease as that of microglia did, even after we induced remyelination by removal of cuprizone (Figure S4c). We confirmed that minocycline did not affect glial fibrillary acidic protein (GFAP) expression in the corpus callosum during remyelination. Immunohistochemical analysis revealed that there were no differences the number of GFAP-positive cells between vehicle-treated mice and minocycline-treated mice in the corpus callosum (Figure S4d–e). We showed minocycline treatment resulted in a significant decrease the number of CC1-positive oligodendrocytes and expression of MBP and CNPase mRNA (Fig. 3e–h). However, it was recently reported that minocycline protects oligodendrocyte progenitor cells from oxygen/glucose deprivation in culture studies (Schmitz et al. 2012). We next examined the effect of minocycline on protection from cell death of oligodendrocyte progenitor cells. We confirmed that staurosporine-induced L-lactate dehydrogenase release in oligodendrocytic cell line, CG-4-16 was not blocked by minocycline treatment (Figure S6a). Furthermore, staurosporine-induced cell death of CG-4-16 was not protected even in the presence of minocycline (Figure S6b). We further examined the number of NG2-positive oligodendrocyte progenitor cells between vehicle-treated mice and minocycline-treated mice in the corpus callosum (Figure S6c). NG2-positive cells were not increased by minocycline administration in the remyelination phase. Minocycline treatment tended to reduce NG2-positive cells but there were not significantly difference between minocycline- and vehicle-treated mice. From this result, we further studied on microglial function in this research.

It has been reported that CNTF can promote oligodendrocyte survival and maturation by activating astrocytes (Albrecht et al. 2007; Nash et al. 2011). However, only a few studies have demonstrated a direct role of CNTF in myelin formation (Mayer et al. 1994; Stankoff et al. 2002). Mice lacking the CNTF gene exhibit a more severe experimental autoimmune encephalomyelitis phenotype (Linker et al. 2002), and over-expressing CNTF reduces demyelination and induces clinical recovery in encephalomyelitis (Lu et al. 2009), suggesting a beneficial role of this trophic factor in remyelination. In this study, we provide evidence that CNTF directly induces MBP and PLP mRNA expression in vitro (Fig. 6d and e). Furthermore, CNTF led to a drastic increase in MBP-positive mature oligodendrocytes that harbored complex processes in primary oligodendrocytes (Fig. 6g). Interestingly, CNPase mRNA was not changed in the presence or absence of CNTF (Fig. 6f). CNPase is known to be expressed in immature oligodendrocytes prior to myelin formation (Yin et al. 1997). Therefore, we examined whether CNPase mRNA was highly expressed at short duration exposure of CNTF. However, we did not obtain the prospective results (Data not shown). We further examined the expression of CNPase in protein level; however, there were not differences the expression level in the presence or absence of recombinant CNTF (Data not shown). Further work will need to address this question.

In summary, we show that minocycline treatment reduces remyelination, at least in part by suppressing CNTF expression after cuprizone-induced demyelination. Although it has been reported that CNTF is produced by astrocytes and promotes myelination (Moore et al. 2011), we demonstrate that CNTF is also produced by microglia. Further investigation of the mechanism of action of CNTF on oligodendrocytes may facilitate the development of new therapeutic strategies for the treatment of MS.

Acknowledgements

We thank Dr. T Ogata (National Rehabilitation Center) for the gift of the oligodendrocyte cell line, CG-4-16 cells, and Mr. T Sasaki, Mr. K Hazawa and Mr. T Nomura for technical assistance. This study was supported by a grant for research from the Noastec Foundation (to T.T.) and a Grant-in-Aid for young scientists (B) (to Y.B).

Ancillary