Comparisons of neurotrophic effects of mesenchymal stem cells derived from adipose tissue, bone marrow, and cranial bone on chronic spinal cord injury

Cell-based therapies with mesenchymal stem cells (MSCs) are considered as promising strategies for spinal cord injury (SCI). MSCs have unique characteristics due to difference in the derived tissues. However, relatively few studies have focused on differences in the therapeutic effects of MSCs derived from different tissues. Here, the therapeutic effects of adipose tissue-derived MSCs (aMSCs), bone marrow-derived MSCs (bMSCs), and cranial bone-derived MSCs (cMSCs) on chronic SCI model rats were compared. Methods was analyzed were established weight-drop motor function was evaluated from before injury to 4 weeks after Endogenous neurotrophic factor and neural repair factor expression in spinal cord (SC) tissue were examined by real-time PCR and western blot analyses. Furthermore, the neurotrophic effects (i.e., neurite formation and elongation) of each MSC type were veried by co-culture with NG108-15 neural cells. PE-conjugated immunoglobulin (Ig)G1 FITC-conjugated IgG1 Biolegend) used as an isotype FITC-conjugated antibody (Ab) against CD44 and PE-conjugated Abs against CD29, CD90, were used as MSC markers. FITC-conjugated Ab against CD45 and PE-conjugated Ab against CD34 were used as hematopoietic markers. Flow cytometry was conducted using the BD FACSVerse™ Flow Cytometer and data acquisition and analyses were performed using BD FACSuite™ software (both, BD Biosciences).


Background
The main causes of spinal cord injury (SCI) include tra c accidents, tumors, and degenerative diseases.
SCI is a serious burden to healthcare systems because of the large patient population [1] and remaining severe dysfunction. The acute phase of SCI is characterized by contusion, laceration, stretch, compression, or direct massive destruction of the spinal cord (SC). Subsequent second injuries such as in ammation and cytotoxic edema, lead to the apoptosis of neurons at the lesion site [2][3][4][5]. The chronic phase of SCI is characterized by the formation of a glial scar after gliosis. Although the formation of a glial scar border prevents further expansion of tissue damage in the acute phase, glial scar formation in the chronic phase limits for axonal regrowth [6,7]. However, even under such circumstances, synapse plasticity and axonal sprouting are promoted by various molecules [8]. Thus, it is important to identify the mechanisms that regulate neural regeneration to promote functional recovery after SCI. Although physical rehabilitation [9] and drugs [10,11] have been used to treat SCIs, there is currently no effective treatment strategy.
Among the many strategies for the SCI treatment, cell-based therapy is considered one of the most important and promising. Mesenchymal stem cells (MSCs) are especially appealing because of the potential of self-renewal and multi-lineage differentiation, secretion of humoral factors, migratory capacity, and ability to promote endogenous neurogenesis [12][13][14]. MSCs, which are isolated mainly from the bone marrow [15][16][17], have been investigate in numerous clinical trials [15,18]. Other than the bone marrow, MSCs can be isolated from adipose tissue, dental pulp, and Wharton jelly [19,20]. Previous studies have identi ed differences in the characteristics of MSCs derived from various tissue [20][21][22].
Our group has focused on the cranial bone which originates from the neural crest as a new source of MSCs. Shinagawa et al. showed that cranial bone-derived MSCs (cMSCs) have a higher neurogenic potential than bone marrow-derived MSCs (bMSCs) [23], whereas Abiko et al. found differences in the therapeutic effects between bMSCs and cMSCs with the use of an ischemic stroke model [24]. Thus, unique characteristics of MSCs derived from various tissues may have different therapeutic effects due to their unique characteristics. Although it is necessary to choose optimal MSCs for each central nervous system (CNS) disease and stage of illness, the timing of treatment with MSCs for SCI was mainly in the acute or subacute phase in previous studies [25][26][27]. In addition, relatively few reports have investigated differences in the therapeutic effects of MSCs derived from various tissued on chronic SCI. Hence, further studies are needed to investigate the effects and underlying mechanisms of various MSCs for the treatment for chronic SCI is an important issue. Therefore, the present study aimed to compare the neurotrophic effects of adipose tissue-derived MSCs (aMSCs), bMSCs, and cMSCs in chronic SCI model rats.

Methods
Isolation and culture of aMSCs, bMSCs and cMSCs For isolation of MSCs, we collected abdominal adipose tissue, femur, tibia, and cranial bone from adult female Sprague-Dawley (SD) rats (body weight, 250-300g). The adipose tissues collected were digested with Collagenase-A type AFA (560 U/mg; Worthington Biochemical Corporation, Lakewood, NJ, USA) and ltered. The extracts were centrifuged and then seeded into tissue culture dishes (Sumitomo Bakelite Co., Tokyo, Japan) containing culture medium, consisting of low-glucose Dulbecco's modi ed Eagle's medium (DMEM; Sigma-Aldrich Corporation, St Louis, MO, USA), supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scienti c, Waltham, MA, USA), penicillin (100 U/mL), and streptomycin (100 µg/mL, both from Sigma-Aldrich). The adhesion of aMSCs was detected the next day after seeding. In the present study, bMSCs and cMSCs were established as described previously [24]. Brie y, bone marrow was collected from the femur and tibia, ltered, and the resulting cells were seeded into culture dishes. Once the cells had adhered to the bottom of the culture dishes at about 5-7 days, non-adherent cells were eliminated by changing the culture medium. The adherent cells were used as bMSCs. After removing muscle, periosteum, and dura matter, cranial bones were seeded into culture dishes. The adhesion of cMSCs was detected at 5-7days after seeding. Cells were maintained at 37°C under an atmosphere of 5% CO 2 /95% air and the culture medium was changed every 3 days after the detection of cell adhesion. The aMSC, bMSC, and cMSC collected were passaged to more than 80% con uence. Multi-lineage cell differentiation aMSCs, bMSCs, and cMSCs at passage 3 or 4 were used for differentiation into osteoblasts, adipocytes, or neurons. To induce osteogenic differentiation, cells were cultured in Mesenchymal Stem Cell Osteogenic Differentiation Medium (PromoCell GmbH, Heidelberg, Germany) for 21 days, which changes to the medium every 3 or 4 days. To con rm calcium deposition, the cells were stained with alizarin red S solution (Sigma-Aldrich). To induce adipogenic differentiation, cells were cultured in Mesenchymal Stem Cell Adipogenic Differentiation Medium (PromoCell) for 14 days, which change to the medium was changed every 3 or 4 days. Afterward, the cells were nally stained with oil red O solution (Wako Pure Chemical Industries, Osaka, Japan) to con rm the formation of lipid droplets. To induce neural differentiation, cells were cultured under modi ed neural differentiation conditions with the use of neural conditioning medium and neural differentiation medium, as described in previous reports [23,24].
Reverse transcription and real-time polymerase chain reaction (PCR) of MSCs At con uence, the cultured cells after reached con uency were collected in PBS. Total RNA was extracted using the NucleoSpin ® RNA kit (MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany), in accordance with the manufacturer's protocol. The puri ed RNA was reverse transcribed into complementary deoxyribonucleic acid (cDNA) using the ReverTra Ace-α-™ reverse transcription kit (Toyobo Co., Ltd., Osaka, Japan). Using cDNA as a template, real-time PCR was performed with the 7900 HT Real-Time PCR system (Applied Biosystems, Carlsbad, CA, USA), in accordance with the manufacturer's instructions. Real-time PCR ampli cation and analyses of rat brain-derived neurotrophic factor (Bdnf), glial cell-derived neurotrophic factor (Gdnf), broblast growth factor 2 (Fgf2), nerve growth factor (Ngf), and sortilin 1, which is also known as neurotrophin 3 (Sort1/Nt-3), were performed using the FastStart™ Universal Probe Master Mix (Roche, Basel, Switzerland) and TaqMan™ Gene Expression Assays (Applied Biosystems). Actin beta (Actb) was used as the loading control to normalize the relative quantity of the speci c messenger ribonucleic acid (mRNA) in each sample. The TaqMan Gene Expression Assays used in this study are listed in Table 1. Surgical procedure Adult female SD rats (body weight, 250-300 g) were used to construct a SCI model with the weightdropping method [26,28]. The rats were anesthetized and a midline linear incision was made over the thoracic (Th) 9-11 spinous processes. After exposing the laminae of Th9-11, laminectomy was carried out at Th10. An impactor rod was set on the surface of the SC at Th10 and a cylindrical brass weight (10 g) was dropped onto the impactor to create a SC contusion was made with a force of 50 g/cm. Following the contusion, the skin was sutured to close the lesion. After the surgical procedure, the rats were administered antibiotics for 5 days postoperatively to prevent infection. The bladders of the SCI rats were compressed manually twice daily until autonomic bladder function had recovered su ciently. After the SCI, the rats also received passive joint motion exercises daily to prevent contracture of after the hindlimb joints.
Experimental groups and cell transplantation SCI rats were assigned to one of the following four treatment groups (n = 10 each): only PBS group; aMSCs group; bMSCs group; cMSCs group. Rats in the aMSC, bMSC, and cMSC groups were injected intravenously with 1.0 × 10 6 MSCs in 300 µL of PBS at 4week post the SCI. To identify the transplanted MSCs in the recipient SCI rats, the MSCs were labeled with PKH26 (Sigma-Aldrich) just before transplantation.

Motor functional assessment of SCI model rats
The inclined plane test and the Basso-Beattie-Bresnahan locomotor rating scale (BBB scale) were used to evaluate hind limb function. The BBB scale is a 22-point scale that systematically follows the recovery of hind-limb function, and ranges from a score of 0, indicative of no observed movement of the hind-limb, to a score of 21, normal ambulation [29]. The inclined plane test assesses the maximum angle at which the rats can hold a position for 5 s on an inclined plane [30]. Motor function was assessed just before the SCI, on day 1 post the SCI, and weekly until sacri ce at week 8 post the SCI.

Detection of transplanted cells in SC
To detect the PKH26-labeled transplanted cells, SCs were collected and cryoprotected. Rats were euthanized by deep anesthesia and transcardial perfusion of 4% paraformaldehyde. The xed SCs that were removed from the vertebral columns, soaked in 30% sucrose solution, embedded in Tissue-Tek O.C.T compound (Sakura Finetechnical Co., Tokyo, Japan), and frozen in liquid nitrogen. Afterward, the spinal columns were sliced at a thickness of 10 μm with a cryostat and the sections were encapsulated by VECTASHIELD Mounting Medium with DAPI (H-1200, Vector Laboratories, Burlingame, CA, USA) to stain the cell nuclei. The encapsulated sections were examined under a multifunctional microscope (KEYENCE Co.).

real-time PCR analysis of SC tissue
Five rats from each group were sacri ced 4 weeks after MSC transplantation. RNA from the SCs was isolated with the ISOGEN RNA extraction reagent (Nippon Gene, Tokyo, Japan) by centrifugation and reverse transcribed into cDNA, as described above. Real-time PCR analyses of rat Bdnf, tropomyosin receptor kinase B (TrkB), growth-associated protein 43 (Gap-43), and synaptophysin (Syn) were performed using TaqMan Gene Expression Assays. Actb was used as the loading control to normalize the relative quantity of the speci c mRNA in each sample. The TaqMan Gene Expression Assays used in this study are listed in Table 2. Western blotting analysis After RNA extraction, total protein was extracted from the remaining lysate and concentrations were measured with the Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA, USA). Proteins (20 μg per lane) were loaded into the lanes of a 15% polyacrylamide gel, separated by electrophoresis, and transferred to nitrocellulose membranes (HybondTM-ECL; GE Healthcare, Little Chalfont, UK). Membranes were blocked with a blocking buffer (20 mM Tris-HCl [pH 7.4], 137 mM NaCl, 0.1% Tween-20, 1% bovine serum albumin) for 60 min at room temperature and then incubated overnight at 4°C with primary Abs against BDNF (GTX132621; dilution 1:2000; GeneTex Inc., Irvine, CA, USA), GAP-43 (GTX127937; dilution 1:10000; GeneTex), and SYN (ab32127; dilution 1:5000; Abcam, Cambridge, UK). After washing with Trisbuffered saline with Tween 20, the membranes were incubated with horseradish peroxidase (HRP)conjugated anti-rabbit IgG (PI-1000; dilution 1:2000; Vector Laboratories) as the secondary Ab. The immunoreaction was visualized using either the SuperSignal® West Pico Chemiluminescent Substrate (Pierce, Rockford, IL, USA). Images were acquired with the VersaDoc™ imaging system (Bio-Rad). The membranes were then stripped with a solution of 62.5 mM Tris-HCl, 2% sodium dodecyl sulfate, and 100 mM 2-mercaptoethanol (pH 6.8), and labeled with HRP-conjugated anti-Actb Ab (A3854; Sigma-Aldrich). The expression levels of the target proteins were quanti ed by measuring the band densities using ImageJ software (version 1.53; National Institutes of Health, Bethesda, MD, USA). Actb was used as the loading control for data normalization.

Immunohistochemical analysis
For immunohistochemical analyses, the xed SCs were embedded in para n using standard methods. Para n blocks were sliced into 10-μm-thick sections, which were rehydrated using xylene and ethanol solutions. After treatment with antigen retrieval solution at 98°C for 15 min, nonspeci c binding was blocked with 1% bovine serum albumin in PBS for 60 min. Afterward, the sections were incubated overnight with rabbit anti-SYN Ab (dilution, 1:500; Abcam) at 4°C, followed by FITC-conjugated goat antirabbit IgG (F9887; dilution, 1:100; Sigma-Aldrich) for 60 min at room temperature. Then, the stained sections were incubated overnight with mouse anti-Tuj1 Ab (GTX631836, 1:250, GeneTex) at 4°C, followed by Rhodamine Red-X-A niPure Fab Fragment Donkey Anti-Mouse IgG (715-297-003; dilution, 1:100; Jackson ImmunoResearch Europe Ltd., Suffolk, UK) for 60 min at room temperature. The stained sections were encapsulated by VECTASHIELD Mounting Medium with DAPI (Vector Laboratories) to stain the cell nuclei. The encapsulated sections were examined under a multifunctional microscope (KEYENCE Co.).
Co-culture of MSCs and NG108-15 neural cells aMSCs, bMSCs, and cMSCs at passage 3 were seeded at 20,000 cells per cm 2 in the wells of sixwell plates (Sumitomo Bakelite). As a negative control (no co-culture), only the MSC growth medium was added to the well. After 24 h, the NG108-15 cells formed a monolayer at a density of 2,000 cells and the co-culture was continued in DMEM/HAM's F12 (Wako Pure Chemical Industries), supplemented with 1% FBS (Thermo Fisher Scienti c), penicillin (100 U/mL), streptomycin (100 µg/mL, both from Sigma-Aldrich) for an additional 48 h. Afterward, the cells were xed in 4% (w/v) paraformaldehyde. NG108-15 neurite outgrowth was assessed by uorescent immunocytochemical analysis. After blocking with 1% bovine serum albumin for 60 min at room temperature, the cells were incubated overnight with mouse anti-Tuj1 Ab (dilution, 1:1000; GeneTex) incubated at 4 °C overnight followed by Alexa Fluor 488conjugated anti-mouse IgG Ab (H+L) (dilution, 1:500; Molecular Probes Europe BV) as the uorescent secondary Ab for 60 min at room temperature. DAPI (dilution, 1:800; Kirkegaard & Perry Laboratories) was used to stain the cell nuclei. The stained cells were examined under a multifunctional microscope (KEYENCE Co.) and neurite analysis was performed using ImageJ software with the NeuronJ plugin [31].

Characteristics of MSCs from different tissues
Flow cytometry was performed to detect surface markers of the aMSCs, bMSCs, and cMSCs. All MSCs were positive for CD29, CD44, and CD90, which are known as MSC positive markers, whereas negative for CD34 and CD45, which are known as MSC negative markers. The expression trends of the immune phenotype were similar for each MSC type ( Table 3). The potential of the MSCs to differentiate into osteoblasts, adipocytes and neurons was investigated. After differentiation, the aMSCs, bMSCs, and cMSCs were positively stained with Arizarin Red-S staining, Oil Red-O staining, and anti-β-tubulin III Ab (Fig. 1A-C). The data are presented as the mean + standard deviation (SD) of three independent experiments. aMSCs, adipose tissue-derived MSCs; bMSCs, bone marrow-derived MSCs; cMSCs, cranial bone-derived MSCs; SD, standard deviation.

Motor functional improvements of chronic SCI rats
The BBB score and inclined plane test were used to evaluate functional improvement after the SCI by MSC transplantation. At 24 h post SCI, most rats showed acute accid paralysis with no spontaneous movement of the hind limbs. From day 7 post the SCI, spontaneous coarse movement of the hind limbs was observed, including gradual recovery. However, there was almost no improvement from day 21 post the SCI. MSCs were transplanted at day 28 post the SCI and were evaluated continuously over a period of 28 days after transplantation. The results of two-way (10 time points × 4 groups) ANOVA revealed signi cant differences among the groups (p < 0.001). Post-hoc analyses were performed at each time point to compare differences among the groups. Based on the BBB score, functional improvements than in the cMSC group were superior to those of the PBS group from day 49 post the SCI (p < 0.05) (Fig. 3A).
The results of the inclined plane test demonstrated that rats in the cMSC group were able to maintain a signi cantly higher angle than in the PBS group from day 42 post the SCI (days 42 and 56: p < 0.01; day 49, p < 0.05) (Fig. 3B). Collectively, these results indicate that the greatest functional improvement after the SCI was obtained with cMSCs.

Localization of transplanted cells in SC
There were no uorescent cells in the spinal gray matter of the PBS group (Fig. 4A). On the other hand, PKH26-labeled aMSCs, bMSCs, and cMSCs injected intravenously were detected in the spinal gray matter of each group at day 28 post transplantation ( Fig. 4B-D).

Effects of transplantation of different MSCs on neurotrophic and neural plasticity factor expression in s SC tissue
To clarify the effects of transplanted MSCs in SC tissue, mRNA expression levels of Bdnf, a receptor of BDNF, and neuronal plasticity factors were evaluated. The results of one-way ANOVA revealed signi cant differences in the mRNA expression levels of Bdnf, TrkB, and Gap-43 in among the different groups (p < 0.05). The results of post-hoc analyses showed that Bdnf and Gap-43 expression levels were signi cantly higher in the cMSC group than those in the PBS and aMSC groups (p < 0.05) (Fig. 5A, C). TrkB expression was also signi cantly higher in the cMSC group than in the PBS group (p < 0.05) (Fig. 5B). On the other hand, there were no signi cant differences in Syn expression levels among the groups (Fig. 5D).
The protein expression levels of BDNF, GAP-43 and SYN were also evaluated. The results of one-way ANOVA revealed signi cant differences in the protein expression levels of BDNF and SYN among the different groups (p < 0.05). The results of post-hoc analysis showed that SYN expression was signi cantly higher in the cMSC group than in the PBS group (p < 0.05) (Fig. 6C), whereas BDNF expression tended to be higher in the cMSC group, but not signi cantly (p = 0.0644 vs. the PBS group; p = 0.0728 vs. the aMSC group; p = 0.0812 vs. the bMSC group) (Fig. 6A). On the other hand, there were no signi cant differences in GAP-43 expression among the groups (Fig. 6B).
Next, the localization of SYN was investigated, as SC tissue showed signi cant changes in protein expression levels. In the motor neurons located in the anterior horn of the SC stained with Tuj1, SYN was con rmed in the contoured part of the neurons in each group. In the PBS group, SYN was sparsely present on the neurons, whereas the SYN content was relatively high in the contoured part of the neurons in each group transplanted with MSCs ( Fig. 7A-D). Especially in the cMSC group, the staining intensity of SYN was strong over the entire contour of the neuron (Fig. 7D).

Neurotrophic effects of different MSCs in vitro
To compare the neurotrophic effects of aMSCs, bMSCs, and cMSCs on neurite outgrowth, different MSCs were co-cultured with NG108-15 cells. After 48 h of co-culture, the NG108-15 cells were immunostained and the following quantitative parameters were evaluated (Fig. 8A): (1) the percentage of neurite-like processes; (2) the number of neurites expressed per cell; (3) the average neurite length (µm); and (4) the longest neurite length in each group (µm). One-way ANOVA revealed signi cant differences in all parameters among the groups (p < 0.05) and post-hoc analyses showed that the values of all parameters were signi cantly greater for NG108-15 neural cells co-cultured with cMSCs (p < 0.05) than those of NG108-15 neural cells cultured alone (control). Moreover, the percentage (71.80 %) of neurite-like processes was greater for co-cultured with cMSCs than that of the control (47.12 %; p < 0.01) and those co-cultured with bMSCs (58.47 %; p < 0.05) (Fig. 8B). Regarding the number of neurites expressed per cell, co-cultured with cMSCs had more neurites than the control in comparison with control (1.62 vs. 0.92/cell, respectively, p < 0.05) (Fig. 8C). Regarding the average neurite length, co-cultured with cMSCs had longer neurites than the control (58.61 µm vs. 45.42 µm, respectively, p < 0.05) (Fig. 8D). Likewise, longest neurite length was greater for co-cultured with cMSCs than for the control (408.32 µm vs. 205.68 µm, respectively, p < 0.05) (Fig. 8E).

Discussion
The aim of the present study was to investigate the neurotrophic effects of transplanted MSCs derived from adipose tissue, bone marrow, and cranial bone on functional improvement after SCI.
The results of gene expression analyses revealed differences in the expression patterns of the neurotrophic factors among the MSCs derived from different tissues. In cMSCs, Bdnf, Ngf, and Sort1 (Nt-3) expression levels were higher than in aMSCs and bMSCs. On the other hand, Gdnf and Fgf2 expression levels were higher in aMSCs than in bMSCs and cMSCs. Previous studies have reported that the expression levels of neural crest-related genes are stronger in cMSCs than in bMSCs [23,24]. Sakai et al.
observed that dental pulp-derived MSCs (dpMSCs) originating from the same neural crest as the cranial bone have higher expression of BDNF and NT-3 than bMSCs [32]. In addition, Mead et al. demonstrated that dpMSCs have high expression levels of BDNF, NGF, and NT-3, whereas aMSCs have higher expression levels of FGF2 than bMSCs and dpMSCs [33]. The trends in the present study were consistent with previous reports. Thus, the high expression levels of Bdnf, Ngf, and Nt-3 in cMSCs were considered to be related to the origin of the neural crest, which differs from mesoderm germ layers originating from other tissues.
The transplantation effects of MSCs derived from different tissues on motor function were investigated in rat models of chronic SCI. In the cMSC group, the motor function score improved signi cantly from 2 or 3 weeks after transplantation, but the changes were negligible in the other groups, suggesting that the neurotrophic factors of cMSCs can improve motor function after a SCI.
The results of SC tissue analyses showed that BDNF and GAP-43 mRNA expression and SYN protein expression were signi cantly higher in the cMSC group than in the PBS and aMSCs groups, with a similar tendency in BDNF protein expression. Furthermore, the intensity of SYN immunostaining of the contours of the motor neurons in the anterior horn of the gray matter was weak in the PBS group, but strong in the cMSC group. In the CNS, BDNF plays important roles in neurogenesis such by promoting synaptic plasticity, myelination, and axonal growth [34][35][36]. Previous studies have reported that MSCs enhanced endogenous BDNF expression in neural tissue [37][38][39][40], suggesting that this effect may be mediated by humoral factors released by MSCs [41]. NGF has been reported to increase BDNF expression in neurons via the TrkA/ERK/CREB pathway [42], and Wang et al. demonstrated that transplantation of NT-3expressing MSCs enhances BDNF expression in the injured SC [43]. Therefore, the neurotrophic factors produced by cMSCs may have contributed to the increased expression of endogenous Bdnf in the cMSC group.
GAP-43 is a neuron-speci c protein that is abundant in axon growth cones [44]. In the injured SC, GAP-43 promotes axon regeneration, new axon germination, and neurite outgrowth [45]. The SYN protein is abundant in presynaptic vesicles and is an important factor re ecting synaptogenesis and plasticity [46].
GAP-43 was reportedly involved in the synaptic plasticity and nerve repair of a SCI model [47], and uctuations in SYN expression in injured spinal motor neurons have been suggested as important indicators associated with the loss and recovery of motor function [48,49]. As changes in GAP-43 and SYN expression levels in the injured SC have signi cant effects on improving motor function, an increase in these factors in the cMSC group may have contributed to improved motor function. In previous studies, BDNF enhanced GAP-43 and SYN expression in neurons through the Trk receptor and subsequent ERK/CREB signaling [50,51]. The promotion of GAP-43 and SYN expression may be mediated by endogenous BDNF of SC tissue that is increased by cMSC transplantation.
It has been con rmed that MSCs are present in the gray matter of the SC tissue after transplantation. Rooney et al. showed that transplanted MSCs survived in the SC for long periods without differentiation and continued to supply neurotrophic factors [52]. Therefore, transplanted MSCs may have direct effects on neurons. Based on these previous reports, in vitro neurite outgrowth experiments using co-culture were conducted to examine the neurotrophic effects on neural cells. Co-culturing with MSCs derived different tissues to assess the tendency of neurite formation showed that cMSCs signi cantly promoted neurotrophic effects. Previous studies have shown that neurite formation and elongation are mediated by neurotrophic factors such as BDNF, NGF, and GDNF [40,[53][54][55]. As expected, co-culture with MSCs expressing these factors promoted neurite growth. In particular, cMSCs were more effective than other MSCs because of the greater number of factors and the higher expression levels of factors bene cial to nerve growth.
The results of the present study showed that cMSCs can promote functional recovery in the chronic phases of CNS diseases, suggesting that such effects are mediated by neurotrophic factors. The expected roles of MSCs in developing treatment strategies for SCI, such as suppressing in ammatory responses and apoptosis in the acute phase, and reducing glial scar formation and promoting nerve growth after the subacute phase, vary depending on the stage. In addition, previous studies reported that multiple doses of MSCs produced a more remarkable effect [56,57]. As these reports focused on MSCs derived from a single tissue, it is important to select MSCs that are most suitable for a particular stage.
Hence, and in the future studies are warranted to identify an appropriate combination of MSCs for administration at each stage of SCI.

Conclusions
In conclusion, the results of the present study showed that cMSCs express an abundance of neurotrophic factors as compared with MSCs derived from other tissues. Further, cMSCs contributed to functional recovery in rat models of chronic SCI by promoting endogenous neurotrophic and neural plasticity factors expression. Even though further studies of cMSCs for the treatment of other diseases and various stages of illness are needed, the results of this study con rmed the e cacy of cMSCs in cell-based therapy for chronic SCI.

Consent for publication
Not applicable.

Availability of data and materials
All the datasets used and/or analyzed during this study are available from the corresponding authors on reasonable request.
Competing interests LY is a director and shareholder of Space Bio-Laboratories Co., Ltd. (SBL) and YK is a president and shareholder of SBL. The interest con icts of this research have been approved by the Con ict of Interest Management Committee. By regularly reporting research progress to the Con icts of Interest Management Committee, we will maintain fairness regarding the interests of this research. All other authors have no personal nancial or institutional interest in any of the drugs, materials, or devices described in this article and declare that they have no competing interests.

Funding
This work was supported in part by Grants-in-Aid for Scienti c Research from the Japan Society for the Promotion of Science (JSPS KAKENHI grant no. 18K10709).

Authors' contributions
TO participated in the design of the study, carried out the in vitro and in vivo experiments, analyzed and interpreted the data, and drafted the manuscript. YM participated in the study design and coordination, carried out the in vivo experiments, and analyzed the data. TK participated in the design of the study, carried out the in vitro and in vivo experiments, and analyzed and interpreted the data. KN participated in the design of the study, analyzed and interpreted the data, and performed the statistical analyses.TM participated in the in vivo experiments, analyzed and interpreted the data, and secured nancial support. YK participated in the design of the study and analyzed and interpreted the data. KK participated in the coordination, helped to draft, the nal version of the manuscript, and secured nancial support. LY conceived the study, participated in its design and coordination, analyzed and interpreted the data, helped to draft the nal version of the manuscript, and secured nancial support. All authors read and approved the nal version of the manuscript for publication.