Differentiation ability of rat‐mesenchymal stem cells from bone marrow and adipose tissue to neurons and glial cells

Mesenchymal stem cells (MSCs) are multipotent cells and can differentiate into neurons and glial cells. In vitro differentiation would be done by the addition of various factors. There remains no comparison for the differentiation of MSCs from rat bone marrow (rBMMSCs) and adipose tissue (rATMSCS) into neurons and glial cells with basic fibroblast growth factor (bFGF), epidermal growth factor (EGF), and brain‐derived neurotrophic factor (BDNF). The aims of this study were to investigate the effect of bFGF, EGF, and BDNF supplementation on the differentiation ability of rBMMSCs and rATMSCs into neurons and glial cells. MSCs were cultured with bFGF and EGF for 4 days and then BDNF was added until day 8. Characterization of MSCs before and after induction was carried out by observing the cell morphology and several cell markers. Flowcytometry analysis was performed for MSCs markers (CD90, CD29) and neurons and glial cell markers (A2B5, Beta‐III‐tubulin, PSAN‐CAM); while MAP‐2, a neuron marker, was analyzed by immunocytochemistry. Induction of both types of MSCs showed MAP‐2‐positive cells, decreased MSCs markers, and in rBMMSCs showed increased neuron markers. The number of neuron marker positive cells in rBMMSCS was higher than rATMSCs. This study showed that the addition of bFGF, EGF, and BDNF to the medium induced rBMMSCs into neurons and glial cells, but the conditions were not optimal for rATMSC as judged by the expression of neural markers (A2B5, Beta‐III‐tubulin, PSAN‐CAM, andMAP‐2).


Introduction
The Mesenchymal stem cells (MSCs) are undifferentiated cells that have the ability of selfrenewal and differenti ate into other cells (Dominici et al. 2006). MSCs are also multipotent; MSCs are capable of proliferating and differ entiating into several constituent cells of the body such as osteoblasts, chondrocytes, adipocytes, neurons, and glial cells. The Mesenchymal stem cells (MSCs) can differen tiate into various cell types provides opportunities in cell utilization as a therapy cell, especially in neurodegener ative diseases (Halim 2010). Neurons and glial cells de rived from MSCs would offer an advantage for cell therapy in the future for the regeneration of neurons and glial cells in neurodegenerative diseases like spinal cord injury (Qu and Zhang 2017).
The Mesenchymal stem cells (MSCs) can be obtained from adult tissues such as bone marrow, adipose tissue, pe ripheral blood, cord blood, tendons, and ligaments (Sand haanam et al. 2013). MSCs from the bone marrow can be obtained in small amounts and are more invasive (Bagha ban Eslaminejad et al. 2008). MSCs from adipose tissue are easily obtained in large quantities through liposuction and easily propagated in vitro (Ikegame et al. 2011). MSCs from bone marrow and adipose tissue are also able to dif ferentiate into osteoblasts, chondrocytes, adipocytes, neu rons, and glial cells (Safford et al. 2004; Rebelatto et al. 2008; Tohill et al. 2004. Induction of MSCs differentiation into neurons and glial cells can be done by adding various neurotrophic fac tors, growth factors, and chemicals in vitro. Several stud ies have reported that nerve growth factor (NGF), brain derived neurotrophic factor (BDNF), and neurotrophin 3 (NT3) are important neurotrophic factors in inducing the proliferation and differentiation of neural stem cells (Naghdi et al. 2009; Chen et al. 2014). In addition, sev eral growth factors such as epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), insulinlike growth factor1 (IGF1), glial growth factor (GGF), and bone mor phogenetic protein4 (BMP4) supplementation to the cul ture medium can induce MSCs to neurons and glial cells (Tohill et al. 2004; Liang et al. 2013; Guan et al. 2014).
The MSCs differentiation into neurons and glial cells in vitro can be characterized by the expression of neu rons and glial cells markers. A2B5 is an immature glial marker in the brain area of the subventricular zone (SVZ) that will develop into astrocytes and oligodendrocytes; and it is widely expressed in embryonic and neonatal neural networks (Dietrich et al. 2002; van Strien et al. 2014. PSANCAM is a neuron progenitor cell or glial progen itor cell marker during brain development (neurogenesis) (Zhang and Jiao 2015). PSANCAM is also reported to be a marker of the development and migration of neurons and synapse formation in the immature nervous system (Quartu et al. 2008). BetaIIItubulin (Tuj1) is a neuron marker that begins to be expressed in the embryonic phase of brain development and often found in postmitotic neu rons that are still immature (und Halbach 2007). MAP2 is a cytoskeletal protein needed for the proliferation, de velopment, differentiation, and maintenance of neurons (Soltani et al. 2005; Liu et al. 2001).
Previous studies have reported the role of combina tions of several growth factors and neurotrophic factors in differentiation into neurons and glial cells. MSCs se crete the growth factors and neurotrophic factorlike EGF, bFGF, and BDNF in the conditioned medium (Wilkins et al. 2009; Pawitan 2014. bFGF plays a role in cell growth, differentiation, and survival (Yang et al. 2008). EGF also plays a role in maintaining cell proliferation and differentiation. The addition of both growth factors (EGF and bFGF) into MSCs culture medium is also reported to increase cell proliferation and support MSCs differentia tion into neurons and glial cells (Radtke et al. 2009; Hu et al. 2013. BDNF plays a role in the differentiation, de velopment, survival of neuron stem cells, axon regener ation, and synapse formation (Lim et al. 2008). The ad dition of growth factors (bFGF, EGF, and BDNF) to the culture medium can be used to induce MSCs differenti ation into neurons and glial cells. Nevertheless, the role of these three factors in the differentiation of MSCs from bone marrow and adipose tissue is unknown. Therefore, this study aimed to investigate the effect of bFGF, EGF, and BDNF supplementation on the differentiation ability of rBMMSCs and rATMSCs into neurons and glial cells by observing their markers (A2B5, BetaIIItubulin, PSAN CAM, and MAP2) to provide basic theoretical data for further research.

Materials and Methods
The study was conducted at the Stem Cell Laboratory, Centre for Research and Development of Biomedical and Basic Health Technology (CRDBBHT), National Health Research and Development Institute (NIHRD), Ministry of Health of the Republic of Indonesia from March to Oc tober 2017.

Isolation and culture of rat MSCs from bone marrow and adipose tissue
The male Sprague Dawley (SD) rats at age of 23 months and weighed 200300 g were used in this study. The animals were obtained from the Animal Laboratory, CRDBBHT, NIHRD, Ministry of Health of the Republic of Indonesia. The animals were handled in compliance with the regulation of local animal welfare facility rules. Isolation of rat adipose tissue and bone marrow from fe mur and tibia bones were performed under anesthesia with ketamine (75100 mg/kg) in mixture with xylazine (510 mg/kg) via intraperitoneal injection. This procedure has been approved by the ethical committee of NIHRD. Isolation of MSCs from the femur and tibia bones of rat (rat bone marrow MSCs, rBMMSCs) was done by cutting off each bone and flushing modification methods by a modified method of Rinendyaputri and Noviantari (2015), by flushing the bones with a syringe of MEM cul ture medium (Sigma) supplemented with 10% fetal bovine serum (FBS) (Gibco), sodium bicarbonate (Sigma), non essential amino acids 1% (Sigma), mercaptoethanol 0.1 mM (Sigma), and gentamicin (Sigma). The cells were in cubated in a 5% CO 2 incubator (Heracell Vios 160i) at 37°C . The culture medium was replaced after MSCs attached in 23 d (Rinendyaputri and Noviantari 2015).
Isolation of MSCs from rat adipose tissue (rATMSCs) was done by washing the adipose tissue in phosphate buffered saline (PBS) supplemented with 1% gentamicin in a petri dish. Adipose tissue was chopped into pieces (±23 mm 2 ) using sterile scissors and inserted into a con ical tube containing a 0.075% type I collagenase solution (Gibco) with a ratio of 1:2 of adipose tissue to collage nase solution. The mixture was incubated in a water bath at 37°C for 1 h with shaking every 5 min. After incuba tion, the mixture was centrifuged at 1400 rpm for 10 min, then the supernatant was removed, and PBS was added to the pellet. The suspension was filtered with a 70 µm cell strainer (or electroplated tea filters). The filtered cell suspension was centrifuged again at 1400 rpm for 10 min and the supernatant was removed. The pellets were resus pended with a culture medium (MEM supplemented with 10% FBS, sodium bicarbonate, 1% nonessential amino acids, 0.1 mM mercaptoethanol, gentamicin, and Gluta MAX [Gibco]). MSCs from adipose tissue were cultured in an incubator with 5% CO 2 at 37°C. The medium was replaced every 23 d. Cells that have been confluent 80% are ready to passage.
When MSCs were nearly confluent (7080%), the cells were passaged with trypsinEDTA (Gibco). The medium was removed. Cells were washed twice with PBS. TrypsinEDTA was added to the cells, and the cells were incubated for 5 min at 37°C. The cells were suspended in the culture medium. The number of cells was counted with a Neubauer hemacytometer cell counting chamber af ter trypan blue staining. The cells were replated in a 12 well plate (1×10 5 cells per well), and cells were incubated at 37°C in the culture medium. The medium was replaced every 23 d.

Induction of MSCs neural differentiation with bFGF, EGF, and BDNF
Induction methods of MSCs neural differentiation were modified from Anghileri et al. (2008) by changing Dulbecco's modified Eagle's medium with Neurobasal medium, omitting retinoic acid (RA) to induce MSCs, and omitting the isolation of spherical floating aggregates af ter incubation with bFGF and EGF. The MSCs at the pas sage (P3) were harvested by trypsinization and then were centrifuged. The pellet was resuspended in the induction medium on a 12well plate (in duplicate). The induc tion consisted of 2 stages: (1) the cells were cultured in Neurobasal medium supplemented with bFGF and EGF (20 ng/mL) containing 10% FBS, 2% B27 supplement (Gibco), antibiotic and antimycotic, and GlutaMAX for 4 d; (2) the cells were cultured with the same medium as (1) with the addition of BDNF (20 ng/mL) until day 8. The cells were incubated in a 5% CO 2 incubator at 37°C. The induction medium was replaced every 23 d. Induced MSCs were observed under an inverted micro scope (Nikon ECLIPSE TE2000U) with NISElements F Imaging Software. Flowcytometry analysis of A2B5, betaIIItubulin, and PSANAM; and immunocytochem istry of MAP2 were conducted on day 8.

Flowcytometry
In this study, rBMMSCs and rATMSCS were character ized with CD29+, CD90+, and CD45 (Biolegend). In duced MSCs were characterized by using A2B5, PSAN CAM+ (Miltenyi Biotec), and BetaIIItubulin (Biole gend) markers according to the instruction kit. The flow cytometry analysis was performed by using a BD Accuri C6 Plus. Flowcytometry data analyzed with Flowjo V10: population gated by total MSC, the single cells; and the markers of MSCs, neurons, and glial cells. The control population was stained with isotype antibodies. Gates were drawn based on matched isotype control cocktails.

MAP-2 immunocytochemistry of induced MSCs
Immunocytochemistry of neural differentiationinduced MSCs was adopted from Stephanie et al. (2013)  Positive and negative controls were included in every staining protocol. Positive control for MAP2 immuno cytochemistry was primary neuron culture from rat brain. MAP2positive cells were indicated by brown color in the cytoplasmic and nucleus area. The negative control was obtained by omitting MAP2 primary antibody.

Data analysis
Data on percentage of MSCs, neuron, and glial marker positive and negative were analyzed using Statistical Prod uct and Service Solution (SPSS) 16. A comparison be tween groups was performed using a ttest with a 95% con fidence level (P<0.05).

Results and Discussion
Isolated MSCs from bone marrow and adipose tissue showed morphology dominated by cells such as fibrob last cells, then cells became confluent and monolayer (Fig  ure 1A and 1B). Immunophenotype analysis of rBMMSCs and rATMSCs at passage 3 (P3) before and after induc tion using flowcytometry showed that more than 50% of cells were CD29+ and CD90+ (MSCs markers) and less than 3% cells were CD45+ (hematopoietic marker) (Fig  ure 2A). After differentiation induction, the percentages of CD90, and CD29 positive were decreased. Decreases in the percentage of rBMMSCs markers (CD29 and CD90) positive cells before and after induction were as 1.95, and 1.18 fold, respectively. The decrease in the percentage of rATMSCs markers (CD29) before and after induction was as 1.66 fold, while the percentage of CD90positive cells in rATMSCs was slightly increased. The comparison between the percentage of CD29 of both MSCs showed a significant difference. Based on the paired sample t test, CD29 had a significant difference with the pvalue of 0.007 for rBMMSCs and 0.004 for rATMSCs before and after induction. Comparison between rBMMSCs and rATMSCs after induction showed that CD29 was signifi cantly different (P<0.05).
Induced MSCs showed morphology change after in duction with bFGF, EGF, and BDNF. Both MSCs showed characteristics of neuronlike morphology with a con densed nucleus, contracted cytoplasm (blue arrow) with two or three cellular processus like bipolar neurons and multipolar neurons (Figures 1C and 1D). MAP2 expres sion in induced MSCs was observed on day 8 after the ad dition of bFGF, EGF, and BDNF. MAP2 positive cells , and CD90 between rBMMSCs and rATMSCs before and after differentiation induction. The percentage of CD45 and CD29 from rBMMSCs and CD45 from rATMSCs showed a significant difference before and after induction (n=3, three independent experiments in duplicate measurement for each experiment). Percentage of CD29 and CD90 after induction between rBMMSCs and rATMSCs showed a significant difference (n=3, three independent induction experiments in duplicate measurement for each experiment). *P<0.05. (b-e) Representative figures of flowcytometry analysis. Immunophenotype of MSCs marker of rBMMSCs before induction (b) and after induction (c). Immunophenotype of MSCs of rATMSCs before induction (d) and after induction (e). Each panel consists of (1) isotype control and (2) sample.
showed brown color in the cytoplasmic and nucleus area (red arrow) ( Figure 1E and 1F). MAP2negative cells ap peared as cells with bluish staining nucleus or hematoxylin stained nucleus ( Figure 1G). After induction with bFGF, EGF, and BDNF, rB MMSCs showed an increased percentage of positive cells for the glial progenitor maker, A3B5, and the immature neuron markers, betaIIItubulin, and PSANCAM. In creases in the percentage of neuron markers A2B5, beta IIItubulin, and PSANCAMpositive cells before and af ter induction of rBMMSCs were as 1.34, 1.53, and 1.47 fold, respectively. But, rATMSCs show decreases in the percentage of neuron markers A2B5, betaIIItubulin, and PSANCAM before and after induction of rATMSCs were as 1.66, 1.2, and 2.03fold, respectively ( Figure 3A). Us ing a paired ttest, the comparison between the percentage of A2B5 of rBMMSCs before and after induction showed a significant difference. The percentage of A2B5 from rB MMSCs had a significant difference with pvalue of 0.032.

Discussion
Isolation and culture of rat bone marrow and rat adipose tissue MSCs were successfully performed, and isolated MSCs showed a high percentage of MSCs markers (CD90 and CD29) with a low percentage of hematopoietic sur face markers (CD45). These results are consistent with the MSCs criteria based on the International Society for Cel lular Therapy (ISCT), which defines MSCs as cells able to attach to a plastic culture container, positive for CD90, CD73, CD105, and CD44, and negative for CD34, CD45, HLADR, and CD11b.1.
In this research, we used MSCs at passage 3 because in this passage, cells show high proliferation and high mul tilineage differentiation capacity. The proliferation of the cells as determined by the cumulative population doubling level was observed at its peak on passage 3, and the pro liferation stopped after passage 5 (Lee et al. 2013).
From our research, we have confirmed that rATMSCs and rBMMSCs express neuron markers spontaneously be fore induction. Deng et al. (2006) reported that although in the absence of specialized induction reagents, MSCs spontaneously express certain neuron phenotype markers. MSCs are positive for several neuronspecific proteins, in cluding βIII tubulin (12%) and NFM (13.2%); negative for PSANCAM, a surface protein expressed on migra tory neuroblasts; positive for the astrocytespecific pro tein, S100β (15%); but negative for the astrocyte interme diate filament proteins, GFAP, and vimentin (Deng et al. 2006).
In this study, retinoic acid (RA) was removed from the neural differentiation induction medium of Anghileri et al. (2008) because we wanted to know the role of the three growth factors (EGF, bFGF, and BDNF) exclusively in the differentiation of MSCs into neurons and glial cells to pro vide basic theoretical data for further research. Anghileri et al. (2008) used 30 d for induction and obtained neuron differentiation in 57% of ATMSC. In this study, we in duced MSCs with only 3 growth factors (bFGF, EFG, and BDNF) for 8 d, and obtained 15% of cells differentiated into neurons and glial cells. This efficiency is comparable to those reported previously. Jeon et al. (2007) reported that MSCs cultured in neuron preinduction medium con taining a combination of growth factors NT3 (30 ng/mL) and bFGF (10 ng/mL) for 45 d followed by an induc tion medium containing NT 3 (30 ng/mL) and BDNF (10 ng/mL) for 7 d produces Nestinpositive cells (4.7 ± 0.8% preinduction and 14.2 ± 2.0% postinduction).
In the present research, we omitted the isolation of spherical floating aggregates (neurosphere) after incuba tion with bFGF and EGF to make an easy and more effi cient differentiation protocol. There are several studies of differentiation MSCs to neurons without neurosphere iso lation steps. Guan et al. (2014) reported that rat MSCs from bone marrow differentiated to neurons by adding combinations of growth factors (EGF, bFGF, IGF1, and NT3) without neurosphere isolation (Guan et al. 2014). Ikegame et al. (2011) also reported that mice MSCs from bone marrow and adipose tissue differentiated to neurons and glial cells after 48 h with several chemical compounds without neurosphere isolation. The percentage of neu ron marker from MSCs from adipose tissue was 40±6% (MAP2), 14±2% (NeuN), and 23±5% (Nestin) (Ikegame et al. 2011).
In the differentiation of MSCs into neurons and glial, there are a variety of different markers according to the stages of differentiation (Rushing and Ihrie 2016). PSA NCAM and A2B5 are markers of neurogenesis as a marker of an immature neuron and glial cells. This is in accor dance with the previous research conducted by Czarnecka et al. (2017), which states that MSCs from the human um bilical cord can differentiate into neurons and glial cells us ing commercial neuron mediums (MSC Neurogenic Dif ferentiation Medium) which produces NCAM+ (50,83 ± 3.01%) and A2B5+ (19.97 ± 1.70%). In this study, a de crease in the percentage of neuron and glial markers may be due to the differentiation of MSCs that are already in the mature neuron stage with positive MAP2 markers.
Our current result differs from that of Ikegame et al. (2011), in which MSCs from adipose tissue showed a bet ter efficacy in treating ischemic stroke in mice by injecting MSCs from adipose tissue or bone marrow. They reported that several tissue regenerative factors like vascular en dothelial cell growth factor (VEGF), angiopoietin1, and hepatocyte growth factor (HGF) from mouse ATMSCs are higher than BMMSCs Ikegame et al. (2011). Assessment of secreted growth factors from rBMMSCs and rATMSCS should be done to compare concentrations of growth and neurotrophic factors in the conditioned medium of both MSCs. Further, add more growth factors and characteri zation of neurons and glial cells induced from rBMMSCs and rATMSCS should also be done with another neural markers, i.e,. Nestin, NeuN, and GFAP.

Conclusions
The addition of bFGF, EGF, and BDNF to the medium induces rBMMSCs into neurons and glial cells, but the medium induction is not optimal for rATMSC by the expression of neural markers (A2B5, BetaIIItubulin, PSANCAM, and MAP2) and needs to be improved.