Electroacupuncture Promotes the Differentiation of Transplanted Bone Marrow Mesenchymal Stem Cells Preinduced with Neurotrophin-3 and Retinoic Acid into Oligodendrocyte-Like Cells in Demyelinated Spinal Cord of Rats

Abstract

Transplantation of bone marrow mesenchymal stem cells (MSCs) promotes functional recovery in multiple sclerosis (MS) patients and in a murine model of MS. However, there is only a modicum of information on differentiation of grafted MSCs into oligodendrocyte-like cells in MS. The purpose of this study was to transplant neurotrophin-3 (NT-3) and retinoic acid (RA) preinduced MSCs (NR-MSCs) into a demyelinated spinal cord induced by ethidium bromide and to investigate whether EA treatment could promote NT-3 secretion in the demyelinated spinal cord. We also sought to determine whether increased NT-3 could further enhance NR-MSCs overexpressing the tyrosine receptor kinase C (TrkC) to differentiate into more oligodendrocyte-like cells, resulting in increased remyelination and nerve conduction in the spinal cord. Our results showed that NT-3 and RA increased transcription of TrkC mRNA in cultured MSCs. EA increased NT-3 levels and promoted differentiation of oligodendrocyte-like cells from grafted NR-MSCs in the demyelinated spinal cord. There was evidence of myelin formation by grafted NR-MSCs. In addition, NR-MSC transplantation combined with EA treatment (the NR-MSCs + EA group) reduced demyelination and promoted remyelination. Furthermore, the conduction of cortical motor-evoked potentials has improved compared to controls. Together, our data suggest that preinduced MSC transplantation combined with EA treatment not only increased MSC differentiation into oligodendrocyte-like cells forming myelin sheaths, but also promoted remyelination and functional improvement of nerve conduction in the demyelinated spinal cord.

Introduction

Multiple sclerosis (MS) mainly affects young adults. The disease is characterized by destruction of myelin in the central nervous system (CNS), that affects 2–150 per 100,000 humans (30). A hallmark pathological change of MS is demyelination, resulting in axonal damage in the CNS, which eventually causes permanent disability. Remyelination takes place spontaneously. However, because of poor proliferation and differentiation of oligodendrocyte precursor cells (OPCs), remyelination often fails during MS (24, 31). Thus, boosting remyelination through enhancing cell numbers and differentiation of OPCs is a therapeutic strategy to improve disability. Bone marrow mesenchymal stem cells (MSCs) have been suggested as a promising candidate for the treatment of MS (3, 9, 12) for their roles in immune modulation and neuroprotection (2, 6, 11, 16, 22, 50). The capacity of transdifferentiation into myelin-forming cells of MSCs seems to be an important mechanism involved in the therapeutic effect of MSCs (45). Indeed, our previous work showed that grafted MSCs, modified by the tyrosine kinase receptor type 3 (TrkC) gene in the injured spinal cord with neurotrophin-3 (NT-3) elevated by electroacupuncture (EA) could differentiate into oligodendrocyte-like cells (13). Therefore, we speculate that NT-3 and TrkC may be promoters for transdifferentiation of MSCs into neural cells.

NT-3 and its high-affinity receptor TrkC contribute greatly to the survival and differentiation of OPCs and formation of myelin sheath. OPCs and oligodendrocytes were reported to decline in NT-3 and TrkC knockout mice (26). NT-3 specifically increases the mature oligodendrocyte population and enhances remyelination in demyelination of adult rat CNS (23). However, primary MSCs do not express TrkC, and to overcome this shortcoming, we utilized all-trans retinoic acid (RA) to treat MSCs modified by the NT-3 gene. We further demonstrated that combination of NT-3 gene transfer and RA promotes MSCs expressing TrkC (52). Based on the above, we speculated that exogenous NT-3 in combination with RA in vivo would yield a similar result.

EA is a therapeutic approach in which a needle inserted into an acupoint is connected to a trace pulse current with the purpose of producing synthetic electric and needling stimulation (32). The NT-3 level declined obviously in the demyelinated spinal cord, but increased significantly after treatment by EA (21).

Arising from the above, we hypothesized that interaction between NT-3 and its receptor TrkC would significantly promote the differentiation of MSCs into oligodendrocyte precursor cells/oligodendrocyte-like cells. To test this hypothesis, we transplanted NT-3 and RA preinduced MSCs (NR-MSCs) into the demyelination area induced by ethidium bromide (EB) and then investigated whether EA treatment would promote endogenous NT-3 secretion in the demyelinated spinal cord. Here we report that increased NT-3 through EA could further enhance differentiation of grafted NR-MSCs into oligodendrocyte-like cells with the potential of forming myelin sheath as well as improvement of remyelination and nerve conduction in the spinal cord.

Materials and Methods

Culture and Preinduction of MSCs

MSCs were obtained as previously described (52) from green fluorescent protein (GFP) transgenic Sprague–Dawley (SD, male) rats (Osaka University, Osaka, Japan). Briefly, the rats (5–7 days old) were anesthetized. MSCs were flushed from the femurs with Dulbecco's modified Eagle's medium-low glucose (LG-DMEM; Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS; TBD Co., Tianjin, China), and then the cells were seeded at 1 × 10⁷/ml in a 50-ml tissue culture flask (Haimen Tianlong Experiment Instrument, Haimen, China) with LG-DMEM supplemented with 10% FBS, penicillin (100 units/ml; Guangzhou Pharmaceuticals Corporation, Guangzhou, China), and streptomycin (100 units/ml; Guangzhou Pharmaceuticals Corporation), and incubated at 37°C in a 5% CO₂ humidified atmosphere. The medium was changed at 48 h. Floating cells were discarded. Fresh medium was added and replaced every 3 days. When the adherent primary MSCs grew to 80% confluency (passage 0), the cells were passaged using 0.25% trypsin (Sigma-Aldrich, St. Louis, MO, USA)/0.02% EDTA (Sangon, Shanghai, China) at the ratio of 1:2. MSCs from passages 3–5 were used for all experiments.

The MSCs were preinduced with LG-DMEM containing NT-3 (20 ng/ml; Sigma-Aldrich) or RA (1 μM/L; Sigma-Aldrich) or NT-3 and RA.

Real-Time Quantitative Polymerase Chain Reaction (RTQ-PCR) Analysis for Transcription of TrkC Gene in MSCs

MSCs were divided into four groups (n = 4): (1) the Control group, treated with LG-DMEM + 10% FBS; (2) the RA group, treated with LG-DMEM + 10% FBS containing RA (1 μM/L); (3) the NT-3 group, treated with LG-DMEM + 10% FBS containing NT-3 (20 ng/ml); (4) the NT-3 + RA group, treated with LG-DMEM + 10% FBS containing NT-3 (20 ng/ml) and RA (1 μM/L). RNA was extracted from MSCs after being preinduced for 1 day and 3 days by using TRIzol reagent (Invitrogen, Grand Island, NY, USA). cDNA was synthesized using 1 μg of RNA by the cDNA Cycle Kit (DRR037S; TaKaRa Bio Inc., Otsu, Japan). Fifty nanograms of cDNA was used for each real-time PCR in TaKaRa SYBR Premix Ex Taq™ mixture (DRR041S; TaKaRa Bio Inc.) Total reaction volume was 20 μl. The reaction took place in a BioRad iCycler iQ5 PCR (BioRad, Hercules, CA, USA). Comparison was based on normalization with the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) transcription level. The following PCR conditions were used: 95°C for 30 s, 40 repeated cycles of 95°C for 5 s, followed by 60°C for 31 s. The primers used were as follows: TrkC forward 5′-ACCATGGCATCACTACACCTTC-3′; TrkC reverse 5′-CTTAGATTGTAGCACTCAGCCAG-3′. GAPDH forward 5′-CAAGTTCAACGGCACAGTCAA-3′; GAPDH reverse 5′-TGGTGAAGACGCCAGT AGACTC-3′.

In Situ Hybridization

MSCs were seeded and cultured on glass coverslips in a 24-well plate (Costar; Corning, Corning, NY, USA). Four groups of MSCs were cultured for 1 day. MSCs were fixed with 4% paraformaldehyde (Guangzhou Chemical Reagent Factory, Guangzhou, China) and 0.1% diethylpyrocarbonate (Invitrogen) for 30–60 min, washed three times with in situ hybridization PBS (Zhongshan Golden Bridge, Beijing, China), and incubated in blocking buffer (without probes; Boster, Wuhan, China) for 2–3 h at 42°C. MSCs were reacted with 20 μl of hybridization solution containing TrkC probes overnight at 42°C. Probes were digoxin-labeled TrkC-mRNA oligonucleotides (No. MK1651; Boster): 5′-AACTACACTTGGCTGGGC AAAGAGAGAC AT-3′; MSCs were washed with 2 × SSC (NaCl 17.6 g, C₆H₅O₇Na₃ 8.8 g in 1 L dH₂O; Guangzhou Chemical Reagent Factory) for 5 min × 3 times, 0.5 × SSC for 5 min × 3 times, and 0.2 × SSC for 5 min × 3 times. MSCs were incubated with biotin (Boster) for 60 min, followed by reaction with SABC-Cy3 (Boster) for 1 h at 37°C. After staining with Hochest 33342 (Sigma-Aldrich) for 15 min, MSCs were coverslipped and examined by fluorescence microscopy.

Western Blot Analysis

Four groups of MSCs (10⁶) were harvested after being preinduced for 1 day, 3 days, and 7 days. The samples were homogenized in ice-cold lysis buffer (7 M urea, 2 M thiourea, 4% 3-cholamidopropyl dimethylammonio-1-propanesulfonate, 0.5% Biolytes pH 5–8, 0.5% Triton X-100, and 1% cocktail; Beyotime, Nantong, China) using homogenizers and a sonication probe (Ningbo Scientz, Ningbo, China). Homogenates were centrifuged at 12,000 rpm (13362 × g, 5154D; Eppendorf, Hamburg, Germany) for 15 min at 4°C. The supernatant was collected and stored at −80°C for Western blot. The protein concentration was assayed by means of Bradford's method. Equal amounts of protein (50 mg) were then run and separated on a 10% SDS polyacrylamide gel (SDS-PAGE; CWBIO, Beijing, China) followed by electrophoresis (BioRad) for 1 h at 150 V. Proteins were transferred to polyvinylidene difluoride (Millipore, Billerica, MA, USA) membranes for 2 h at 250 mA. After blocking nonspecific binding sites with 5% nonfat milk (Amresco, Solon, OH, USA) in TBST [0.5% Tween 20 (Beyotime) in TBS (Tris 2.42 g; Genview, Houston, TX, USA; NaCl 8 g, Guangzhou Chemical Reagent Factory) in 1 L dH₂O] for 1 h at room temperature, the membranes were incubated with rabbit anti-TrkC (1:800; Cell Signaling, Boston, USA) and mouse anti-β-actin (1:1,000; Cell Signaling) overnight at 4°C, respectively. After washing, the membranes were treated with a horseradish peroxidase-conjugated secondary antibody (goat anti-rabbit or goat anti-mouse, 1:5,000; Jackson ImmunoResearch Labs, Inc., Beverly, MA, USA) for 2 h at room temperature and then washed thoroughly. The protein bands were visualized by enhanced chemiluminescence detection reagents (Applygen Technologies Inc., Beijing, China) and exposed onto X-films (Kodak, Rochester, NY, USA) for 1–5 min. Neonatal hippocampi were used as a positive control.

Induction of Demyelinating Lesion

Adult female SD (220–250 g) rats were supplied by the Experimental Animal Center of Sun Yat-sen University, China. All experimental protocols and animal handling procedures were approved by the Animal Care and Use Committee of Sun Yat-sen University and were consistent with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All efforts were made to reduce the use and suffering of animals. The rats used were housed in a temperature-controlled (24 ± 2°C) and light-controlled (12:12 light–dark cycle) room with free access to food and water. Prior to experimental manipulation, rats were allowed to acclimatize to the housing facilities and handled daily for at least 1 week. Focal demyelinating lesions of dorsal funiculus (DF) were induced by injection of EB into spinal cord according to previous reports (21, 53). Briefly, the rats were anesthetized with 1% sodium pentobarbital (40 mg/kg, IP; Merck, Darmstadt, Germany). A laminectomy was performed at the T9 vertebral level, and the spinal cord was exposed. After the dorsal dura was sheared open, 1 μl EB (0.1 mg/ml, in vehicle phosphate-buffered saline; Sigma-Aldrich) was stereotaxically injected into the DF of the T10 spinal cord at a depth of 1 mm using a microsyringe (Shanghai Anting plant microinjector, China). After surgery, all rats received an intramuscular injection of penicillin (160,000 U/ml/day) and were housed separately in individual cages with enriched soft bedding. Beam walking was performed every other day after the surgery and transplantation of MSCs.

A total of 41 rats were used for the entire study, and of these, eight rats (n = 4 for each group) were used for enzyme-linked immunosorbent assay (ELISA). After transplantation, 33 rats were used, and among these, 30 of them were for electrophysiology recording (n = 5 for each group). These rats were subsequently used for analysis in semithin sections and ultrathin sections (n = 5 for each group) and quantification of differentiated transplanted MSCs (n = 4 for each group). Of the 33 rats, two of them were for immunoelectron microscopy and one for immunohistochemical staining at 60 days.

Electroacupuncture (EA) Therapy

EA therapy was administered as described in our previous work (21, 33, 34). Briefly, 1 day after EB injection, EA stimulation was performed at Jizhong (GV6) and Zhiyang (GV9) of Governor Vessel (GV) every other day. GV6 and GV9 were located at the interval of T7–T8 and T11–T12 processus spinosus, respectively. A pair of stainless needles (Suzhou Medical Supplies, Suzhou, China) of 0.35-mm diameter were inserted into GV6 and GV9 to a depth of 5 mm but exterior to the vertebral canal. The needles were connected to the output terminals of an EA apparatus (Model G 6805-2, Shanghai Medical Electronic Apparatus Company). Alternating strings of dense-sparse frequencies (60 Hz for 1.05 s and 2 Hz for 2.85 s alternately) were used for EA. The intensity was adjusted to induce slight twitch of the hindlimbs (≤1 mA) with the intensity lasting for 20 min.

For the NR-MSCs + NP group, the nonpoint area (NP) was located at the interval of the T7–T8 and T11–T12 processus spinosus, 10–13 mm lateral to posterior median line. This area was treated using oblique insertion to a depth of about 10 mm subcutaneously.

Detection of NT-3 Expression by Enzyme-Linked Immunosorbent Assay (ELISA)

Eight rats were divided into two groups (n = 4 for each group): the EB (14 days) group (induction of demyelinating lesion by EB for 14 days) and the EB + EA (14 days) group (induction of demyelinating lesion by EB and EA for 14 days). The rats were anesthetized with 1% sodium pentobarbital (50 mg/kg, IP) and transcardially perfused with 200 ml of ice-cold 0.1 M PB [NaCl 8 g, KCl 2.2 g, Na₂HPO₄ 1.44 g, KH₂PO₄ 0.24 g (Guangzhou Chemical Reagent Factory) in 1 L dH₂O]. The segments containing the lesion areas were excised, placed on dry ice and weighed, and then mechanically homogenized in ice-cold 0.1 M PB. Homogenates were centrifuged at 14,000 rpm (18,188 × g, 5154D; Eppendorf), 4°C for 10 min and used for NT-3 ELISA according to the instructions of the manufacturer (NT-3 Emax ImmunoAssay System; Boster).

Animal Groups and Cell Transplantation

Three days after EB injection, rats were reanesthetized and divided into six groups: the EB (group 1), EA (group 2), MSCs (group 3), NR-MSCs (group 4), NR-MSCs + EA (group 5), and NR-MSCs + NP (group 6) groups. DMEM (1 μl) or MSCs (1 × 10⁵/μl) was directly injected into the EB lesion area of spinal cord. The EB and EA groups received 1 μl DMEM injection, respectively. The MSCs group was given 1 μl MSC (1 × 10⁵/μl) injection. The NR-MSCs, NR-MSCs + EA, and NR-MSCs + NP groups received 1 μl NR-MSC (1 × 10⁵/μl) injection, respectively. The NR-MSCs were preinduced with NT-3 and RA for 24 h. Before transplantation, 10 μg/ml Hoechst 33342 (Sigma-Aldrich) was used to label MSCs in the different treatment groups for 45 min. After surgery, the EA and RN-MSCs + EA groups received GV-point EA therapy for 29 days, respectively. The RN-MSCs + NP group received NP + EA therapy instead.

Electrophysiology

The electrophysiology was performed at 30 days based on the findings of our previous work (21, 53) that the model is normally adequately impaired by this time. Thirty days after the cell transplantation, 30 rats (n = 5 for each group) were deeply anesthetized with ketamine (40 mg/kg; Jiangsu Hengrui Medicine, Lianyungang, China) and 1% sodium pentobarbital (30 mg/kg) and fixed with a stereotaxic apparatus (RWD Life Science Co., Ltd., Shenzhen, China) to undergo electrophysiological testing. The scalp was sheared, and routine disinfection was carried out. The scalp was cut open along the longitudinal midline. We set anterior fontanel as the origin in order to determine the right sensorimotor cortex, located 2 mm behind the anterior fontanel, 2 mm to the right of the midline. The skull was drilled, and the sensorimotor cortex was exposed. Motor-evoked potentials (MEPs) were elicited by electrical stimulation of the sensorimotor cortex. MEPs were recorded simultaneously from the contralateral sciatic nerve using bipolar electrodes (Shanghai Huayi Medical Instrument). These electrodes were then connected to the BL-410E Data Acquisition Analysis System for Life Science (Taimeng, China). The Parameter Settings of the MEPs signal are as follows: gain parameter 2,000, time constant 0.01 s, and filtering 300 Hz. Each pulse stimulation is 50 ms in duration at a frequency of 5.1 Hz and with a 6-mV voltage density. Fifty MEP responses were collected for each rat of the right and left sensorimotor cortex. The amplitude and latency of recorded averages were analyzed.

Immunohistochemical Staining

Following the MEP recording, six groups of rats were perfused with 4% paraformaldehyde (Guangzhou Chemical Reagent Factory). The spinal cord was dissected, postfixed overnight in paraformaldehyde, and cryoprotected in 30% sucrose (Guangzhou Chemical Reagent Factory) at 4°C. Transverse cryostat sections (25-μm thickness) of spinal cord were cut and mounted on slides (Citotest Labware Manufacturing, Nanjing, China) coated with gelatin (Guangzhou Chemical Reagent Factory). The slides were rinsed with PBS, then blocked with 10% normal goat serum (Beijing Zhongshan Golden Bridge) for 30 min at 37°C and reacted with primary antibodies overnight at 4°C. Primary antibodies used were as follows: rabbit anti-rat TrkC (1:600; Cell Signaling), mouse anti-rat GFP (1:400; Chemicon, Temecula, CA, USA), rabbit anti-rat GFP (1:100; Beyotime), rabbit anti-rat NG2 (a marker for oligodendrocyte progenitor cells, 1:200; Chemicon), mouse anti-rat adenomatous polyposis coli (APC, a marker for mature oligodendrocytes, 1:200; Chemicon), rabbit anti-rat myelin basic protein (MBP, a marker for myelin, 1:200; Santa Cruz, Santa Cruz, CA, USA), chicken anti-rat myelin basic protein (1:200; Santa Cruz), and rabbit anti-rat neurofilament (NF, 1:300; Sigma-Aldrich). Subsequently, the sections were rinsed with PBS and incubated with their respective secondary antibodies for 1 h at 37°C. Cy3-conjugated goat anti-mouse IgG and Cy3-conjugated goat anti-rabbit IgG (1:800; Jackson ImmunoResearch Labs, Inc., West Grove, PA, USA), FITC-conjugated goat anti-mouse IgG, FITC-conjugated goat anti-rabbit IgG (1:200; Jackson ImmunoResearch Labs, Inc.), and Dylight 649-conjugated goat anti-chicken IgG (1:200; Jackson ImmunoResearch Labs, Inc.) were used as secondary antibodies, respectively. Finally, the sections were rinsed with PBS and coverslipped. The stained slides were examined with a Zeiss fluorescence or Zeiss confocal microscope (Zeiss, Oberkochen, Gemany), and images were captured.

Quantification of Differentiation of Transplanted MSCs

The spinal cords of the MSCs, NR-MSCs, NR-MSC + EA, and NR-MSC + NP groups (n = 4 for each group) were cut in transverse sections. Every fifth section was mounted on a gelatin-coated slide. Four transverse sections from each cell transplantation group were randomly selected for NG2 and APC immunofluorescent triple labeling to assess the differentiation of transplanted MSCs into oligodendrocyte progenitor-like cells and oligodendrocyte-like cells. The DF of spinal cord in each section was examined at 300× magnification and the image captured using a Zeiss fluorescence microscope. Two full views from different areas of the demyelinated DF were scrutinized, and the total number of NG2- and APC-positive cells, the number of NG2- or APC-positive cells with GFP and Hoechst 33342 labeling, and the total number of cells labeled with GFP and Hoechst 33342 were counted. The ratio of triple-labeled cells (NG2 or APC and GFP and Hoechst 33342) to the total GFP and Hoechst 33342-labeled cells was calculated and compared among different MSC transplantation groups.

Myelin Sheath Formation 60 Days After Grafted MSCs

The spinal cord of animals in the NR-MSCs + EA group (n = 1) was cut in longitudinal sections and immuno-fluoresence stained with MBP, NF, and GFP antibodies for MSCs grafted 60 days after transplantation.

Semithin Resin Sections and Ultrathin Sections

Following perfusion with 4% paraformaldehyde, the spinal cords in six groups of animals (n = 5 for each group) were postfixed in 2.5% glutaraldehyde (Guangzhou Chemical Reagent Factory) for 72 h. Two 3-mm blocks were cut from the cord and processed for semithin sectioning. After rinsing in PBS, the tissue blocks were fixed with 1% osmium tetraoxide (OsO₄; Zhongjingkeyi Technology, Beijing, China) for 1 h at room temperature in darkness. Dehydration was through a graded series of ethanol and absolute acetone (Guangzhou Chemical Reagent Factory). Finally the tissue blocks were immersed in 50% Epon 812 (Ted Pella, Redding, CA, USA) in acetone for 2 h and embedded in Epon 812 overnight and then hardened at 60°C for 48 h. Semithin sections (1 μm thickness) were cut and stained with toluidine blue (Amresco). Two sections, in which the lesion area included the entire DF, were selected from each block for myelin counting. Twenty sections were selected from each group (n = 5). The depth of the DF in the selected sections was divided into three layers (the superficial, middle, and deep layers) and examined at 400× magnification. Three types of myelin sheath, including the newly formed myelin (remyelination; thinner and lighter in color), degenerated myelin (loose, redundant sheaths that form loops), and normal myelin (compact, thick dark sheaths), within a unit area (100 μm²) in the respective layer were identified and enumerated. The total number of three types of myelin sheath from 12 unit areas per rat was calculated. Ultrathin sections were prepared, stained with uranyl acetate (Zhongjingkeyi Technology) and lead citrate (Zhongjingkeyi Technology), and then examined under a transmission electron microscope (Philips CM 10, Eindhoven, Holland).

Immunoelectron Microscopy

Twenty-nine days after EA therapy, the rats in the NR-MSCs + EA group (n = 2) were perfused intracardially with 0.9% saline followed by 500 ml of fixative consisting of 4% paraformaldehyde + 0.1% glutaraldehyde + 15% saturated picric acid (Guangzhou Chemical Reagent Factory) in 0.1 M phosphate buffer at pH 7.4. T8–T12 spinal cord segments were removed, postfixed for 4 h at 4°C in fresh fixative without glutaraldehyde and subsequently cut into 50-μm sagittal sections on a vibratome (Thermo Scientific, Pittsburgh, PA, USA). After cryoprotection in 10% and 20% sucrose for 30 min, respectively, vibratome sections were transferred into cryprotectant solution containing 30% sucrose and 10% glycerol (Guangzhou Chemical Reagent Factory) in 0.1 M PBS for 2 h at 4°C, followed by a quick freeze–thaw in liquid nitrogen 15 s. After washing with PBS, the sections were treated for 40 min with 20% goat serum to block nonspecific binding of the antibody. The sections were immunostained with rabbit anti-rat GFP (1:100; Beyotime) for 24 h at 4°C, followed by incubation with 1.4 nm immunogold-labeled goat anti-rabbit IgG antibody (Nanoprobes, Yaphank, NY, USA) overnight at 4°C, and then postfixed with 2% glutaraldehyde in 0.1 M PBS for 10 min. These sections were washed and then were silver enhanced with HQ silver Kit (Nanoprobes). The sections were then postfixed with 2% OsO₄ in 0.1 M PBS, dehydrated in a graded series of ethanol concentrations, and flat-embedded in Epon812. The tissue blocks were cut into ultrathin sections (70 nm thick) and collected on formvar-coated nickel slot grids (Zhongjingkeyi Technology) and stained with uranyl acetate. The sections were examined using a CM12 Philips electron microscope.

Data Analysis

Data were given as the means ± standard deviation. They were analyzed using one-way ANOVA or t-test by statistical software SPSS17.0 (Chicago, IL, USA). The Bonferroni test was applied to the post hoc test with one-way ANOVA. The significance level was set at p < 0.05.

Results

TrkC mRNA and Protein Expression in Preinduced MSCs

To assess transcriptional changes, RTQ-PCR was used to quantify TrkC mRNA in the control, NT-3, RA, and NT-3 + RA groups at 1 and 3 days after induction. The level of TrkC gene expression was significantly higher in the NT-3 + RA group (at 1 day) than in the control and other treatment groups at either time point, except for the 3-day NT-3 + RA group (p < 0.05) (Fig. 1A).

Figure 1.

TrkC mRNA and protein expression in preinduced MSCs. (A) The MSCs were cultured for 1 and 3 days. TrkC mRNA was measured by RTQ-PCR and normalized by the level of the housekeeping gene GAPDH. TrkC mRNA transcription level was significantly increased in the 1-day NT-3 + RA group. *p < 0.01; #*p < 0.05. (B1–B3) In situ hybridization of TrkC mRNA. MSCs from 1-day NT-3 + RA group with fluorescence-labeled nucleotide probe specific to TrkC mRNA (red, B1). (B3) The merge image of TrkC mRNA and Hoechst 33342 (Hoe). Scale bars: 40 μm (B1–B3). (C) NR-MSCs were analyzed for the presence of TrkC using Western blot at 1, 3, and 7 days after preinduction. TrkC was expressed in neonatal hippocampus (a positive control); however, it was not detected in any of the MSCs (C). (D1–E3) TrkC expression by immunohistochemical staining. (D1–D3) TrkC is not detected in the MSCs in the 7-day NT-3 + RA group, but TrkC, however, is expressed in the replication-defective recombinant adenoviral vector carrying the human TrkC gene (Ad-TrkC) infected MSCs (a positive control) (E1–E3). Scale bars: 20 μm (D1–E3).

To further verify whether TrkC mRNA was expressed in MSCs, we performed in situ hybridization with a nucleotide probe specific to the mRNA of TrkC. We found TrkC mRNA distributed in the cytoplasm and nucleus of MSCs at 1 day in the NT-3 + RA group (Fig. 1B1–B3); this was also observed in the MSCs of the other groups at 1 day (Supplementary Fig. 1; http://www.scribd.com/doc/214814123/Supplementary-Fig-1).

To assess the overall expression of TrkC protein, we performed Western blot on MSCs of the control, NT-3, RA, and NT-3 + RA groups at 1 day; the NT-3 + RA group at 3 and 7 days; and on positive control cells from neonatal hippocampi. A specific protein band of ~145 kDa, consistent with the predicted size of the TrkC protein, was found in cells of the neonatal hippocampus; however, it was not detected in MSCs (Fig. 1C).

In vitro analysis of TrkC expression by immunohistochemical staining showed that TrkC protein could not be detected in MSCs of the 7-day NT-3 + RA group (Fig. 1D1–D3), and in the MSCs of the other groups (Supplementary Fig. 2; http://www.scribd.com/doc/214814069/Supplementary-Fig-2), but replication-defective recombinant adenoviral vector carrying the human TrkC gene (Ad-TrkC) infected MSCs that acted as a positive control expressed TrkC protein (Fig. 1E1–E3).

Figure 2.

TrkC expression in NR-MSCs after transplantation in vivo and NT-3 expression in the demyelinated spinal cord tissue. (A1–A5) TrkC immunoexpression is detected in MSCs in the RN-MSCs + EA group. Immunostaining image of TrkC (A1), GFP (A2), and Hoechst 33342 (A3) was merged (A4). (A5) Overlay shows many cells coexpressing TrkC and GFP. Scale bars: 10 μm (1A–A5). (B) Pink box in the schematic diagram shows the figure site of (A1–A5) in the dorsal spinal cord. NT-3 protein in the demyelinated tissue was detected by ELISA 14 days after EB injection. (C) NT-3 level in the EB + EA group was higher than that in the EB group (p < 0.05).

TrkC Expressed in the Demyelination/Graft Area

MSCs did not express TrkC protein in vitro whether preinduced by NT-3 + RA or not. However, double staining of TrkC/GFP showed that MSCs of the RN-MSCs + EA group expressed TrkC in the demyelination/graft area of spinal cord at 30 days after transplantation (Fig. 2A1–A4), as well as others groups (Supplementary Fig. 3; http://www.scribd.com/doc/214814092/Supplementary-Fig-3). We examined the z-axis by high-power confocal microscopy. Sta cking images showed that TrkC and GFP-positive MSCs overlapped (Fig. 2A5). Together, these data suggest that MSCs can express TrkC protein, when transplanted subsequently into the demyelination area.

Figure 3.

Differentiation of MSCs into oligodendrocyte progenitor cells (OPCs) 30 days following transplantation. Cryostat sections from the demyelinated spinal cord 30 days after grafting of MSCs were stained with antibodies against NG2 (red) for OPCs and GFP (green) for grafted MSCs. Nuclear staining was by Hoechst 33342 (blue). The immunostained sections were observed under a fluorescence microscope. (A1–D4) NG2 and GFP double labeling shows the NG2-positive OPCs (yellow arrows) in MSCs (A1–A4), NR-MSCs (B1–B4), NR-MSCs + EA (C1–C4), and NR-MSCs + NP (D1–D4) groups. Scale bars: 10 μm (A1–D4). (E) Cell number counting indicates that the percentages of NG2/GFP-positive cells in the NR-MSCs + EA and NR-MSCs + NP groups were higher than that in the other two groups. *p < 0.01; #p < 0.05.

EA Improved NT-3 Level

NT-3 in the demyelination tissue of spinal cord was detected by ELISA 14 days after EB injection. NT-3 expression level in the EB + EA group was higher than that in the EB group (p < 0.05) (Fig. 2C). The result suggests that EA treatment can promote NT-3 level in the demyelinated tissue of spinal cord.

EA Promoted Differentiation of NT-3 and RA Preinduced MSCs In Vivo

Thirty days after transplantation, GFP-positive MSCs were predominantly observed in the demyelination/graft area. The MSCs appeared to be well integrated with the host tissue and displayed diverse morphological forms. Moreover, GFP-positive MSCs were found to migrate to other areas of the demyelination/graft area from the primary site of injection.

The grafted GFP-MSCs in all four groups expressed to varying extents OPC and oligodendrocyte markers (NG2 and APC, Figs. 3 and 4). Colocalization of different markers was confirmed by stacking images of confocal microscopy, which showed that expression of NG2, GFP, and Hoechst 33342 appeared to overlap (Supplementary Fig. 4; http://www.scribd.com/doc/214814107/Supplementary-Fig-4). Likewise, expression of APC, GFP, and Hoechst 33342 overlapped (Supplementary Fig. 5; http://www.scribd.com/doc/214814111/Supplementary-Fig-5). Importantly, data analysis showed that the percentage of NG2/GFP double-positive cells was higher in the NR-MSCs + EA group than that in the MSCs and NR-MSCs groups (Fig. 3E). Moreover, the percentage of APC/GFP double-positive cells in the NR-MSCs + EA group was highest compared with other groups (Fig. 4E). The results suggest that NT-3 + RA preinduced, and EA therapy were beneficial for differentiation of grafted MSCs into oligodendrocyte-like cells.

Figure 4.

Differentiation of MSCs into oligodendrocyte-like cells 30 days following transplantation. Cryostat sections from the demyelinated spinal cord 30 days after grafting of MSCs were stained with antibodies against APC (red) for oligodendrocytes and GFP (green) for grafted MSCs. Hoechst 33342 (blue) staining was for nucleus. The sections were examined under a fluorescence microscope. (A1–D4) APC and GFP double labeling shows APC-positive oligodendrocyte-like cells (yellow arrows) in the MSCs (A1–A4), NR-MSCs (B1–B4), NR-MSCs + EA (C1–C4), and NR-MSCs + NP (D1–D4) groups. Scale bars: 10 μm (A1–D4). (E) Note percentage of APC/GFP-positive cells in the NR-MSCs + EA group was significantly higher than that in the MSCs and NR-MSCs groups. *p < 0.05.

To determine whether oligodendrocyte-like cells from MSCs could form myelin sheaths, we performed MBP immunohistochemical staining. We did not find MBP/GFP double-positive cells in the MSCs and NR-MSCs groups (Fig. 5A4, B4), but they were present in the NR-MSCs + EA and NR-MSCs + NP groups (Fig. 5C4, D4). We next used MBP/NF/GFP triple staining to assess axon–glial interaction by confocal microscopy. Processes were observed projecting from differentiated myelin-forming cells surrounding host axons (Fig. 5E5) in the NR-MSCs + EA group, and MBP and GFP overlapped in the stacking image (Fig. 5E6). The results suggest that grafted MSCs could differentiate into oligodendrocyte-like cells and form myelin sheaths in the NR-MSCs + EA group. This was verified by immunoelectron microscopy using a GFP antibody. GFP-positive grafted cells were visible within the demyelination/graft area, and the GFP reaction products were localized in the cell nucleus and cytoplasm. Remarkably, a variable number of myelin sheaths were detected in the grafted MSCs (Fig. 5F1, F2). It is noteworthy that these myelin-forming cells lacked the basement membrane that is normally associated with Schwann cells. Additionally, distinct heterochromatin clumps were observed in oligodendrocyte (Fig. 5F1).

Figure 5.

Myelin sheath forming by MSCs at 30 days following transplantation. (A1–D4) Cryostat sections from the demyelinated spinal cord 30 days after grafting of MSCs and immunostained with antibodies against MBP (red) for myelinating oligodendrocyte and GFP (green) for MSCs grafted. Nuclear staining was by Hoechst 33342 (blue). (A1–D4) MBP and GFP double labeling shows that MBP and GFP-positive myelin-forming cells (yellow arrows) were observed in the NR-MSCs + EA (C1–C4) and NR-MSCs + NP (D1–D4) groups, but were not detected in the MSCs (A1–A4) and NR-MSCs (B1–B4) groups. Scale bars: 10 μm (A1–D4). (E1–E6) The sections were stained with antibodies against MBP (purple) for myelinating oligodendtrocye, NF (red) for axon, and GFP (green) for MSCs grafted. Nuclear staining was with Hoechst 33342 (blue). (E1–E6) Overlapped MBP (purple), NF (red), GFP (green), and Hoechst 33342 (blue) immunofluorescence image is seen in (E5, white arrows). Note MBP and GFP doublelabeled structure surrounded an NF positive axon (E5, arrow). The overlap of MBP and GFP was verified in the stacking image (E6). Scale bars: 10 μm (E1–E6). (F1–F2) By electron microscopy, grafted GFP-positive cells were identified at the demyelination/graft area in the demyelinated spinal cord. Profiles of myelin sheaths appeared to be surrounded by axons. (F1) GFP reaction products (gold particles and silver-enhanced, red arrows) were found in the nucleus and cytoplasm. (F2) A higher magnification image of boxed area in (F1). Scale bars: 1 μm (F1); 200 nm (F2).

NR-MSC Implants and EA Promoted Remyelination in and around Demyelination Area

Semithin sections of the demyelination/graft area of spinal cord were stained by toluidine blue. Myelin sheaths were classified into newborn myelin (remyelination), degenerated myelin, and normal myelin as reported previously (21, 36). All three kinds of myelin sheaths were present in large numbers in DF (Fig. 6A1, A2). By the electron microscopy, degenerated myelin was loose in composition and disorganized (Fig. 6A2, B1). The newly formed myelin (newborn myelin) was thinner than normal myelin (Fig. 6A2, B2). Normal myelin was compact in texture and smooth in outline (Fig. 6A2, B3). Statistical analysis showed that the number of newborn myelin was increased significantly in the NR-MSCs + EA group compared with the EB, EA, and MSC groups (p < 0.05, respectively). More normal myelin was also observed in the NR-MSCs + EA group compared with the EB and EA groups (p < 0.01, respectively) (Fig. 6C).

Figure 6.

Diverse morphological forms of myelin sheath and numerical comparison of three kinds of myelins. Cryostat sections from the demyelinated spinal cord 30 days after grafting of MSCs. The myelin structures were examined by light and electron microscopy. (A1) Semithin sections stained with toluidine blue showing a focal demyelinated area in the DF of spinal cord. Scale bars: 160 μm (A). (A2) A magnified view in the boxed area (red box) in (A). Degenerated myelin (red arrow), newborn myelin (black arrow), normal myelin (red arrowhead), and naked axon (black arrowhead). Scale bars: 10 μm (A2). (B1–B4) Electron micrographs showing three kinds of myelin sheath and naked axon. (B1) Degenerated myelin (red arrow); (B2) newborn myelin (black arrow); (B3) normal myelin (red arrowhead); (B4) naked axon (black arrowhead). Scale bars: 0.5 μm (B1–B4). (C) Quantitative analysis of three kinds of myelin: degenerated myelin, remyelination (newborn myelin), and normal myelin. *p < 0.01; #p < 0.05.

Motor-Evoked Potentials (MEPs)

To assess functional recovery of axonal conduction following remyelination, MEP responses (Fig. 7A–F) were used as an in vivo functional measurement in the DF of spinal cord (n = 5 for each group). The latencies of MEPs in the NR-MSCs + EA group were decreased significantly compared with the EB group (p < 0.01) (Fig. 7G). The amplitudes of MEPs in the NR-MSCs + EA group were significantly higher compared with the EB, EA, and MSCs groups (p < 0.05, respectively) (Fig. 7H).

Figure 7.

MEP recovery 30 days after transplantation of preinduced MSCs and EA treatment. (A–F) Showing a representative positive–negative-positive wave with a similar wave shape among all six groups. (A) the EB group, (B) the EA group, (C) the MSCs group, (D) the NR-MSCs group, (E) the NR-MSCs + EA group, and (F) the NR-MSCs + NP group. Comparison of the CMEP latency (G) and CMEP amplitude (H) among six groups. *p < 0.01; #p < 0.05.

Discussion

We reported previously that EA and the NT-3 gene-modified MSCs and could promote remyelination in the demyelinated spinal cord of rats (21, 53). Here we have extended our study to further explore whether EA combined with NR-MSC transplantation treatment would promote differentiation of MSCs in demyelinated spinal cord injury induced by EB. If so, we sought to ascertain whether this might be through upregulating the expression of TrkC in MSCs and increasing the concentration of NT-3 in the demyelination/graft area that might collectively participate in the myelin structure reconstruction and improve the nerve conduction of injured spinal cord. First, we found that the transcription of TrkC mRNA in MSCs was markedly increased at 1 day when MSCs were preinduced with NT-3 + RA. More importantly, grafted MSCs could express TrkC protein in vivo. Along with this, the NT-3 expression level was significantly higher in demyelinated spinal cord tissue after 14-day EA treatment. Remarkably, both NR-MSCs + EA group exhibited higher NG2-positive and APC-positive cells, and the increase of these cells was statistically significant after 30 days. Furthermore, some grafted MSCs in the NR-MSCs + EA and NR-MSCs + NP groups were marked by MBP. Immunoelectron microscope showed that GFP-positive MSCs could differentiate into oligodendrocyte-like cells and form myelin sheaths enclosing axons. Another notable feature was more newborn myelin and normal myelin in the NR-MSCs + EA group in comparison with other tested groups, and the difference was statistically significant. The NR-MSCs + EA group had shorter latency and higher amplitude in cortical MEPs compared with other tested groups, and likewise, the difference was statistically significant.

We have demonstrated in a previous study that combination of adenoviral vector-mediated NT-3 gene transfer and RA could initiate MSCs to transcribe TrkC mRNA (52). In the present study, we attempted to increase transcription of TrkC mRNA with treatment with exogenous NT-3 in combination with RA, instead of modifying the NT-3 gene. It is striking that TrkC mRNA expression level was increased by 2.5-fold when MSCs were preinduced with exogenous NT-3 + RA (the NT-3 + RA group) for 1 day. However, the transcription level of TrkC mRNA in the NT-3 and RA groups was relatively unaltered when compared with that in the control group. Zhao et al. reported that, without RA, brain-derived neurotrophic factor (BDNF), gene-modified MSCs, or exogenous BDNF could not cause MSCs to express O₄, nestin, or NeuN (54). This suggests that the promotion of transcription might be attributed to a synergistic effect of NT-3 and RA in combination. We did not find TrkC protein in MSCs in vitro, despite the fact that TrkC mRNA transcription was promoted by NT-3 + RA. Discrepancy between the expression of mRNA and protein has been well documented (17). Thus, in glial cells, CYP2D6 mRNA was found; on the other hand, CYP2D6 protein was absent (42). Protein translation is regulated by a variety of factors, for example, Vav1 mRNA expression was detected in several breast cancer cell lines, yet no detectable levels of Vav1 protein were observed due to cbl-c proteasomal degradation (41). The discrepancy in expression of TrkC mRNA and its protein as observed in this study may have the following explanations. One possible explanation would be that despite the improvement of TrkC mRNA transcription by NT-3 + RA induction in vitro, it might not have provided a sufficient environmental cue to initiate the translation of TrkC. Interestingly, TrkC protein was localized in MSCs induced by the same NT-3 + RA concentration in a three-dimensional gelatin sponge scaffold in a separate study by us (55), but its expression could not be detected in a two-dimensional culture in the present study. Moreover, Pisati et al. also reported that MSCs express TrkC when cocultured with slices of neonatal brain cortex, whereas primary cultures of MSCs do not (40). Althought TrkC protein was not detected in MSCs in vitro, it was expressed by demyelinated spinal cord at 30 days at the injected site after transplantation. This may be due to the influence of the local neural microenvironment on the grafted MSCs. For example, bone morphogenetic protein 2, whose expression is widespread in the nervous system (35, 37), induced expression of TrkC mRNA and responsiveness of neurons to NT-3 as measured by neurite outgrowth (51). NT-3 and TrkC have been implicated in oligodendrocyte differentiation of postnatal neural stem cells (20), and our earlier work demonstrated that TrkC gene-modified MSCs differentiating into oligodendrocyte-like cells were significantly increased compared to primary MSCs after grafted into transected spinal cord of rats for 10 weeks (13). In light of present results, it is suggested that MSCs express more TrkC mRNA when the cells were preinduced by NT-3 + RA, and this would result in a higher oligodendrocyte-like cell differentiation.

The present results have shown that NT-3 level was drastically increased in demyelinated spinal cord by EA. It is reasonable to suggest that the NT-3 enhanced would act on TrkC mRNA transcribing NR-MSCs grafted to the demyelination/graft area. Consequently, this would promote the differentiation of MSCs. In the present study, NT-3 concentration in the demyelinated spinal cord tissue increased by 25% after EA for 14 days. The result was consistent with our previous study (21). Neurotrophin secretion depends on multiple regulatory mechanisms. Releasing of NT-3 is determined by depolarization, calcium influx via N-type, L-type calcium channels, or via NMDA receptors and cAMP concentration and CaM kinase II activity (5, 29, 46). Vasoactive intestinal polypeptide (VIP) promotes neuronal differentiation through activity-dependent neurotrophic factor (ADNF), a protein secreted by VIP-stimulated astroglia.

ADNF causes secretion of NT-3 by neurons (8). Moreover, it has been demonstrated that EA could cause depolarization of presynaptic terminals of neurons (32) and improve intracellular concentration of calcium (48), cAMP (14), and VIP (19). The above may offer possible explanations for the underlying mechanisms of NT-3 elevation by EA. Furthermore, NT-3 may be synthesized and released by autocrine mode by the neural cells themselves (18).

Although MSCs can differentiate into oligodendrocyte-like cells in vitro (1, 27), there is no direct evidence to support that MSCs can differentiate into mature remyelinating oligodendrocytes in the demyelination/graft area of nervous tissue. We reported previously that exogenous NT-3 could promote the differentiation of TrkC-MSCs into oligodendrocyte-like cells by binding to its receptor TrkC. Moreover, the effect of NT-3 promoting the differentiation of TrkC-MSCs was prevented by application of K252a (a specific inhibitor of neurotrophin-related tyrosine kinase) (13). Arising from the present results, it is speculated that combination of NR-MSC transplantation with EA treatment could enhance the differentiation efficacy of the MSCs. Thus, the differentiation of NR-MSCs into OPCs (in the NR-MSCs + EA group) and oligodendrocyte-like cells (in the NR-MSCs + EA group) was significantly increased, concomitant to an elevation of NT-3 levels. Our data suggest that EA can raise the differentiation of grafted NR-MSCs by elevating NT-3 concentration in demyelinated spinal cord tissue. Moreover, in the NR-MSCs + EA and NR-MSCs + NP groups, MSCs could express MBP (the protein components of the myelin membrane). In addition, the results of confocal microscopy and immunoelectron microscopy further demonstrated that the NR-MSC-derived oligodendrocyte-like cells could form myelin sheaths wrapping the axons in the demyelination/graft area. The myelin sheaths wrapping the axons in the NR-MSCs + EA group remained evident 2 months after EA (Supplementary Fig. 6; http://www.scribd.com/doc/214814119/Supplementary-Fig-6). Our data suggest that the grafted MSCs in the NR-MSCs + EA group can differentiate into remyelinating oligodendrocyte-like cells. Kassis et al. reported that MSCs could express gala ctocerebroside and O₄ (oligodendrocytic marker) in chronic experimental autoimmune encephalomyelitis (28), but there was no evidence of myelin sheath associated with the grafted MSCs. The effects of NT-3 and TrkC and regulatory mechanism on the differentiation of MSCs into OPCs/oligodendrocyte-like cells are obscure. PI3-kinase (PI3K) pathway and cyclic AMP-response element-binding protein can play an important role in the survival and proliferation of OPCs by NT-3 (25, 38). Extra cellular signal-related kinase1/2, but not PI3K pathway, is required for NT- 3-induced oligodendrocyte differentiation of postnatal neural stem cells (20). Besides NT-3, EA could increase the level of platelet-derived growth factor (44), basic fibroblast growth factor (39), and insulin-like growth factor 1 (15), which could induce the MSC differentiation into oligodendrocyte-like cells (27, 47).

OPCs and subventricular zone-derived neural stem cells in CNS are the two sources of newly generated myelinating oligodendrocytes. Deficiencies in the number of the two kinds of cells and impaired OPC differentiation ability (31) result in failure of remyelination during MS. We showed previously robust mobilization of OPCs within the demyelination/graft area after EA or following grafting of NT-3 gene-modified MSCs (21, 53). Here we showed a sparsity of degenerated myelins but more newborn myelin and normal myelin in the demyelination/graft area in the NR-MSCs + EA group. Remyelination is essential for the recovery of nerve conduction in demyelinated spinal cord tissue. The electrophysiology results showed that the latencies of MEPs were decreased significantly in the NR-MSCs + EA group, while the amplitudes of MEPs were higher significantly in this group. The changes in latency and amplitude are in accord with the greater frequency of myelin sheaths in the NR-MSCs + EA group. Latency is a manifestation of spinal cord conduction velocity (7) and depends on morphological maturity of myelinated axons (10). The peak amplitude reflects the action potential (43) and correlates with axonal damage (49). Consequently, our data indicate that combination of NT-3 with grafting of NR-MSCs can reduce the myelin sheath damage, boost oligodendrocyte remyelination, and improve the nerve conduction. Bystander neuroprotective and neurotrophic properties are important therapeutic mechanisms of MSC transplantation for demyelination disease (4). Therefore, we speculate that the trophic factors secreted by NR-MSCs may make a contribution to enhance remyelination and that the contribution was strengthened by EA.

In summary, the present results have shown that combination of NR-MSC transplantation with EA treatment not only enhances MSC differentiation into OPCs and oligodendrocyte-like cells, which appear to form newborn myelins wrapping axons, but also promotes myelin repair. This may be the underlying cause of an increase in nerve conduction for functional improvement. Our results indicate that preinduced MSCs combined with EA treatment may be a new strategy for neural transdifferentiation of MSCs in treatment of multiple sclerosis.