Male C57BL/6 mice aged between 5 and 6 weeks, acquired from the Experimental Animal Center, Chinese Academy of Sciences (Shanghai, China), were used in the present study. To induce demyelination, mice were fed with standard rodent chow containing 0.2% CPZ powder (Sigma-Aldrich, St. Louis, MO, USA) for 5 weeks. After 5 weeks' induction, CPZ was removed and replaced with standard diet for 2 weeks allowing for spontaneous remyelination. All protocols were performed and approved in accordance with the National Institutes of Health Guide for the Care and the Animal Research Welfare Council of School of Basic Medical Science of Fudan University (20140266-086). All endeavors were made as far as possible to reduce the sacrifice of animals and relieve their sufferings during the experiments.

EA treatment started from the first day of week 5 and lasted for 3 weeks. Mice received 30 min EA treatment once every other day. Two acupoints in the governor vessel (GV) were adopted, which is Baihui (GV20) and Zhiyang (GV9). A pair of stainless needles with diameters of 0.3 mm (Suzhou Medical Supplies, Suzhou, P.R. of China) were obliquely and subcutaneously inserted into GV9 and GV20 to a depth of 5 mm or so but exterior to the harnpan and vertebral canal. The pair of needles were then connected with an output terminal of an EA apparatus (HANS Acupoint Nerve Stimulator, LH202H, Beijing, P.R. of China), with alternating strains of dense-sparse frequencies of 2/15 Hz and stimulating current of 3–4 mA. Mice were restrained by an apparatus made by our lab specially for the EA treatment. All other mice groups have been restrained the same way. In our preliminary study, we have set up a group of sham-EA. Mice in sham-EA group received the same acupuncture treatment without electric stimulation.

The assessment of mice locomotor coordination was performed by beam walking test (Skripuletz et al., 2010). In brief, mice were trained to traverse a 1.5 cm narrow wood beam to reach a so-called safety box with paddings inside, which created a relatively safe environment encouraging mice to traverse the beam. Mice were placed on one end of the 100 cm long beam (horizontally 60 cm above the platform), and the traversing times were recorded. To reduce errors, the time mice spent on both ends (10 cm of each end) were neglected, only the time spent in the middle 80 cm of the beam were recorded. As displayed in Video 1, we started recording at the moment hind limbs of mice passing the START line on the beam, and end recording at the moment fore limbs of mice touching the END line on the beam. Mice received two consecutive trials in each test and the mean time of two trials was used for statistical analysis (cut-off time 60 s). Thick cotton cushions were placed under the beam in case mice slip down.

Pole test was performed according to previous reports with minor adjustments (Lin et al., 2013). In brief, mice were placed tenderly head-up facing the apex of a vertical round wood pole (diameter: 1 cm; height: 50 cm) with gauze-wrapped rough surface enabling mice to grab. The time mice climbing over the apex of pole with head facing down and body in a vertical position (turn-back time), and the time mice descending to the bottom of the pole after climbing over (touch-down time) were recorded for analysis (Video 3). The cut-off time was 30 and 60 s for turn-back time and touch-down time, respectively. In each test, each mouse conducted two consecutive trials and the average time of two trials was recorded for statistical analysis. Thick cotton cushions were placed under the pole in case mice fall off.

Mice were gently anesthetized with saline solution of pentobarbital sodium (70 mg/kg, i.p.) and perfused with 4% paraformaldehyde. The cerebra were resected and serial 4 μm paraffin sections were stained with hematoxylin and eosins (HE) or luxol fast blue (LFB) to assess demyelination within the area of CC. For HE staining, the process was conducted following our previous protocol (Wang et al., 2012). For LFB staining, paraffin tissues were stained in LFB/cresyl violet overnight at 55°C, followed by washing in 95% ethanol and double-distilled water to remove redundant dye. Then, the white matter of the sections was distinguished from the gray matter in a lithium carbonate solution for about 15 s (until easily distinguishable). Later on, the sections were washed by double-distilled water and 75% ethanol.

After drying in 100% propylene glycol, brain tissues were stained with 0.5% Red Oil O solution (Sigma, USA) at a temperature of 60°C for 6 min. After that the brain slices were incubated with 85% propylene glycol for 2 min following by rinsing. Nuclei were stained with haema-toxylin (Sigma, USA).

Paraffin sections (4 μm) were used for MBP (1:200, Millipore, AB980) and dMBP (1:2000, Millipore, AB5864) staining to determine the degree of myelination and the production of degraded myelin debris respectively. Frozen sections (25 μm) were prepared for staining with Iba1 (1:400, Wako, Japan), Arg1 (1:200, Santa Cruz, USA), iNOS (1:200, Santa Cruz, USA), CD206 (1:100, R&D, USA), CD16/32 (1:500, BD, USA), and O4 (1:50, Millipore, USA) to determine the recruitment of microglia into the area of CC and the phenotype of the recruited microglia. The sections were first washed three times and blocked with 4% goat serum in 0.3% Triton X-100 for 3 h at room temperature followed by incubation with antibodies described above at 4°C for 12 h followed by Alexa Fluor 488 or 555 goat anti-rabbit secondary antibodies (1:1000, Invitrogen, USA) at 37°C. All sections were treated with Fluorescence Decay Resistant Medium with DAPI before being covered with coverslip. The images were captured by a multiphoton laser scanning confocal microscopy system (Olympus Fluoview FV1000) or common fluorescence microscope (Leica DMI6000).

Western blot was performed to quantify the expression of MBP. The brain was removed and CC tissue was quickly resected on ice. The CC was ultrasonically homogenized in radioimmunoprecipitation assay lysis buffer (RIPA buffer, Beyotime, Shanghai, China) followed by 12,000 rpm centrifugation for 10 min and the supernatant was collected for western blot analysis. Equal amounts of protein samples (20 μg of total protein) were analyzed by SDS-PAGE with primary antibodies being either rabbit anti-MBP (1:500, Millipore, AB980) or anti-GAPDH (1:10000, Proteintech, HRP-60004). Proteins were detected via incubation with horseradish peroxidase-conjugated secondary antibodies and an ECL chemiluminescence detection system (Tanon, Shanghai, China). The images were obtained using an ImageQuant LAS4000 mini image analyzer (GE Healthcare, Buckinghamshire, UK).

Total RNA was prepared using the Qiagen RNeasy kit. Libraries were prepared using the NEBNext Library Prep Kit (New England Biolabs) according to the manufacturer's instructions. Library quality was assessed using a Bioanalyzer (Agilent) and then the samples were sequenced on the Illumina Hiseq 2000 with a goal of 30 million reads per sample. Raw FASTQ files were aligned using PRADA and FPKM values obtained using Cuffllinks for gene expression analysis.

To evaluate the mRNA expression of genes related to oligodendrocyte lineage and myelination, the corpus callosum were dissected 7 weeks after the first day of cuprizone feeding; total RNA were isolated using Trizol reagent (Invitrogen, USA). The relative abundance of target mRNAs were then quantified using SYBR Green qRT-PCR detection (Light Cycler 96 real-time PCR detection system, Roche, Switzerland). The primer sequences of target mRNA are listed in Table 1. The housekeeping gene, HPRT, was used as an internal reference for standardization of the analysis. Relative quantification was performed by determination of the n-fold differential expression with the 2^−ΔΔCt method and is expressed as relative fold change compared to HPRT. Melting curves were used to establish the purity of the amplified band. The PCR products were sequenced to confirm identity.

Western blot bands and immunofluorescence staining (integrated optical density, IOD) were analyzed using ImageJ software. The quantitative statistical graphs were created based on the relative fold change.

All quantitative data were presented as mean ± standard error of the mean (S.E.M.). Statistical analysis and graphs were obtained using Graphpad 5.0 software. Differences between groups were analyzed with Student's t-test and one-way analysis of variance (ANOVA) followed by LSD post-test or Bonferroni post-test. Values of P < 0.05 were regarded as the criteria of significance.

During the first 10 days after CPZ administration, we observed obvious weight loss of CPZ fed mice compared to mice with normal diet (Figure 1A). Two weeks later, CPZ fed mice gradually regained weight, but still did not weigh as much as the normal group. CPZ feeding can lead to extensive demyelination in many regions of brain including hippocampus, cerebellum, cortex, and with corpus callosum (CC) the most vulnerable one, which results in motor coordination impairment. To evaluate the motor coordinative function, beam walking test and pole test were performed. Mice exposed to CPZ exhibited severe coordinative locomotor dysfunction with the peak at and around week 4 (Figures 1B–D; Videos 1–4). As shown in the videos, in beam test CPZ fed mice tended to be more hesitant traversing the beam with more fumbles and pauses; in pole test, CPZ fed mice had more difficulty in climbing over the apex of the pole. However, the touch-down time in pole test was not affected by CPZ (Data sheet 1, Figure S1). No mice fell off or slipped over from the beam or pole during our behavioral tests. CC is previously reported as the most vulnerable region during CPZ treatment (Steelman et al., 2012). Thus, we evaluated the histomorphological change of CC region by HE and LFB staining and observed conspicuous demyelination in CC after CPZ treatment for 5 weeks (Figures 1E,G). Myelin basic protein (MBP) is the main component of oligodendrocyte as well as myelin within CC. Therefore, we further detected the amount of MBP in CC by western blot and verified obvious decrease of MBP in CPZ fed mice (Figure 1F). Those results clearly indicate a successful establishment of CPZ induced demyelinating model.

To examine the effect of EA on the neurological behaviors of mice with demyelination, we performed beam walking test and pole test, which aim to examine motor coordinative function in association with ataxia, which is one of the main clinical features of MS patients. Previously, we compared the effect of EA treatment and sham-EA treatment (described in Section Materials and Methods), and our preliminary results indicate that sham-EA treatment exert little effect on remyelination (Data sheet 1, Figure S2). Thus, to reduce unnecessary sacrifice of mice under the consideration of animal welfare, we excluded sham-EA group for the following experiments. EA treatment started at the beginning of week 5 and lasted for 3 weeks until mice were sacrificed at the end of week 7 (Figure 2A). Baihui (GV20) and Zhiyang (GV9) acupoints were used for EA treatment (Figure 2B). Mice receiving EA treatment exhibited a better performance in beam walking test (Figure 2C) and turn-back time in pole test (Figure 2D). Two weeks EA administration already effectively decreased the traverse time in beam walking test (Figure 2C); and the traverse time in beam test and turn-back time in pole test completely reverted back to normal by the end of the treatment. However, the touch-down time in pole test, which mainly indicating the muscular strength of forelimbs, was not affected either by CPZ (Figure S1) or EA treatment (data not shown). These results prompt us to conclude that EA modulates neurobehavioral dysfunction associated with demyelination, and that muscle strength of forelimbs was not affected either by CPZ or EA treatment.

CC is the most evident and notable region of demyelination in CPZ model. Thus, to evaluate the effect of EA on remyelination, we applied RNA-seq to the transcriptome of CC tissue from normal mice, CPZ fed mice and CPZ + EA treated mice at the end of week 7. We focused on the expression profiles of genes relating to myelin components, nodal/paranodal proteins, and some transcription factors that play a role in oligodendrocyte development. The RNA-seq results (Data sheet 2) revealed remarkably reduced expression of most genes in CPZ group, while EA treatment reversed the reduction and stimulated the expression of most of those myelin-related genes, notably CNP, PLP1/2, CLDN11, and CNTNAP3 (Figure 3A). To further confirm the results, we applied RT-PCR experiments to detect the mRNA expression of genes encoding proteins of myelin components and the results were basically in accordance with RNA-seq results (Figures 3B–E). Myelin basic protein (MBP) is the main component of myelin sheath, and the anti-MBP antibody is one of the most commonly used markers to examine the amount of myelin and the integrity of an intact myelin structure; CNPase is also a myelin protein which is exclusively expressed by lineage of oligodendrocytes in the CNS. To further verify the efficacy of EA in remyelination, we examined the expression of MBP and CNPase in CC by using western blot. At week 5, we did not observe obvious difference in the expression of MBP or CNPase between CPZ group and EA group; however, EA treatment for 3 weeks caused significantly increased protein expression of MBP and CNPase in CC region compared to CPZ group (Figures 3F–H). Similar results were also obtained from immunofluorescence experiments. Since demyelination in corpus callosum is uneven (more severe in caudal part), we detected demyelination in both caudal and rostral part of corpus callosum (Figure 3I) with MBP and O4. At week 7, CPZ fed mice showed decreased myelin integrity compared to normal mice as detected by MBP and O4 staining, whereas EA can significantly enhance the expression of MBP and O4 and rebuild dense myelin sheaths (Figures 3J,K). Taken together, these results confirm the efficacy of EA on remyelination.

In CPZ demyelinating model, demyelination is followed by accumulation of degraded myelin debris and activation of glial cells (Praet et al., 2014). It is reported that the accumulation of myelin debris is detrimental to CNS axonal remyelination by inhibiting the differentiation of OPC (Kotter et al., 2006). Thus, the removal and clearance of degraded myelin debris, generated during the process of demyelination, is a critical step for remyelination process (Lampron et al., 2015). To test whether the effect of EA on remyelination is related to the clearance of myelin debris, an antibody which specifically binds degraded myelin basic protein (dMBP) was used. This dMBP antibody is an important tool to study myelin debris, which has previously been reported to bond a certain MBP epitope that is only accessible in areas of myelin degeneration (Matsuo et al., 1997). Consistent with previous study (Cantoni et al., 2015), we were still able to find a certain amounts of degraded myelin debris in CC compared to normal mice at week 7, when 2 weeks after CPZ has already been withdrawn. However, 3 weeks of EA treatment significantly reduced the existence of myelin debris (Figures 4A,B). To further confirm this result, we performed red oil O staining to detect the presence of lipid-rich myelin debris. Degraded myelin debris tend to display dark-red deposits and clumps in red oil O staining (Weinger et al., 2011). We found that EA treatment for 3 weeks also reduced dark-red clumps as compared to CPZ group at week 7 (Figure 4C). Thus, we confirm that EA treatment promotes clearance of degraded myelin debris, which is beneficial for subsequent axonal remyelination.

Compelling evidence suggests that microglia, especially M2-polarized cells, confer neuroprotection in MS through enhancing clearance and phagocytosis of degraded myelin debris, as well as recruiting OPCs to promote remyelination (Jurevics et al., 2002; Mikita et al., 2011; Vaknin et al., 2011; Olah et al., 2012). In this regard, we reasoned that the decrease of degraded myelin debris and the promotion of remyelination resulting from EA treatment could be related to the regulatory effect of EA on microglia. To prove this hypothesis, microglia in areas of CC were assessed by staining Iba1. The CPZ model is a demyelinating model of non-autoimmune character with intact brain-blood barrier hardly ever allowing large amount of peripheral monocytes/macrophages to infiltrate the CNS (McMahon et al., 2002; Remington et al., 2007). Additionally, previous studies rarely distinguished brain-resident microglia from macrophages of peripheral origin, with the two cell types being phenotypically nearly the same. For this reason, in this study we refer to microglia/macrophages as microglia.

At week 7, when EA treatment was in its 3rd week, more microglia assembled into CC in EA group than CPZ group (Figures 5A,B). Notably, we found the morphology of microglia assembled into CC of EA group was quite different from that of CPZ group as well as the microglia outside the CC in EA group (Figures 5A,C,D). Microglia in CPZ group and in areas outside the CC of EA group were mostly ramified microglial cells, while microglia assembled into CC of EA group looked more like phagocytic microglia for these cells generally displaying an ameboid, plump shape, indicating an active phagocytic process taking place in CC after EA treatment, which supposedly is the cause for enhanced clearance of myelin debris. To further prove it, we checked the gene expression relating to phagocytosis according to the gene annotation of Postitive Regulation of Phagocytosis (GO: 0050766) from AmiGo2 (http://amigo.geneontology.org/amigo/term/GO:0050766). A large number of phagocytosis-related genes were up-regulated after EA treatment (Figure 5E). Taken together, these results prove that EA facilitates microglia assembling into corpus callosum exerting phagocytic function.

It has been reported that microglia tend to change their phenotype toward M2 phenotype during the process of myelin debris clearance, and microglia with M2 property exert protective effect in demyelination (Kotter et al., 2006; Neumann et al., 2008; Skripuletz et al., 2013). Therefore, we conducted immunofluorescence staining to detect expression of markers of M2 and M1 phenotype on microglia in CC. The expression of CD16/32 and iNOS (M1 markers) were elevated in both CPZ group and EA group due to exposure to cuprizone, but there is no difference of their expression between CPZ and EA group (Figures 6B,C,E). On the contrary, we found significantly elevated expression of CD206 and Arg1 (M2 markers) after EA treatment compared to CPZ group (Figures 6B,D,E). Since we observed elevated expression of Iba1 in corpus callosum at week 7 (Figures 5A,B), the increasd expression of M2 markers could also result from the increased number of microglia assembling in CC. So we then compared the CD206/CD16/32 ratio in different groups and we found a higher CD206/CD16/32 ratio after EA treatment. These results indicate that more microglia in corpus callosum change their phenotype to M2 phenotype after EA treatment (Figure 6F).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.