Comparison of the Confluence-Initiated Neurogenic Differentiation Tendency of Adipose-Derived and Bone Marrow-Derived Mesenchymal Stem Cells

Adipose-derived mesenchymal stem cells (ADSCs), which tended to neurogenically differentiate spontaneously after achieving high confluence, were observed. Human ADSCs reaching 80% confluence were cultured in DMEM without an inducing factor for 24 h and then maintained in DMEM plus 1% FBS medium for 7 days. The neurogenic, adipogenic, and osteogenic genes of the factor-induced and confluence-initiated differentiation of the ADSCs and bone marrow-derived mesenchymal stem cells (BMSCs) at passages 3 to 5 were determined and compared using RT-qPCR, and the neurogenic differentiation was confirmed using immunofluorescent staining. In vitro tests revealed that the RNA and protein expression of neuronal markers, including class III β-tubulin (TUBB3), microtubule-associated protein 2 (MAP2), neurofilament medium polypeptide (NEFM), neurofilament heavy polypeptide (NEFH), and neurofilament light polypeptide (NEFL), had been enhanced in the confluence-initiated differentiation of the ADSCs. In addition, the expressions of neurotrophins, such as the nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and glial cell-derived neurotrophic factor (GDNF), were also elevated in the confluence-initiated differentiation of the ADSCs. However, the confluent ADSCs did not show a tendency toward spontaneous adipogenic and osteogenic differentiation. Moreover, compared with the confluent ADSCs, the tendency of spontaneous neurogenic, adipogenic, and osteogenic differentiation of the confluent human bone marrow mesenchymal stem cells (BMSCs) was not observed. The results indicated that ADSCs had the potential to spontaneously differentiate into neuron-like cells during the confluent culture period; however, this tendency was not observed in BMSCs.


Introduction
Over the past two decades, mesenchymal stem cells (MSCs) have been studied extensively. Owning the abilities of self-renewal and multi-lineage differentiation [1,2], MSCs are considered to be a promising therapeutic treatment for organ repair and regeneration. MSCs are known to be isolated from several organs, such as bone marrow [3,4], adipose tissues [5], dental tissues [6], umbilical cord blood [7], and the liver [8]. Adipose-derived mesenchymal stem cells (ADSCs), which can be gained from adipose tissue, are easily and numerously available and can be provided as an alternative for tissue engineering and cell All the methods were carried out in accordance with relevant guidelines and regulations, and all the adipose tissue was collected with the patients' informed consent. About 2-3 × 10 6 cells were seeded and cultured in DMEM containing 10% FBS and antibiotic-antimitotic solution (Corning life science, New York, NY, USA) for the cell culture. The fresh growth medium was changed every two days, and the cells were propagated every four days at a 1:5 split before the growth reached 80% of the confluence.

Confluence-Initiated Differentiation
For spontaneous differentiation, the 80% confluent cells at passages 5-6 were cultured in a culture medium without the inducing factor for 24 h. The cells were then maintained in the DMEM plus 1% FBS medium for seven days.

Osteogenic Induction Medium (OIM)-Induced Differentiation
The OIM included a culture medium supplemented with 10% FBS, 50 µg/mL ascorbic acid-2 phosphate (Nacalai, Kyoto, Japan), 0.01 µM dexamethasone (Sigma-Aldrich), and 1 mM b-glycerol phosphate (Sigma-Aldrich), all of which were used to induce osteogenic differentiation. The medium was changed every two days for a period of seven days.

RT-qPCR Analysis
Total RNA was extracted using the Trizol reagent (Invitrogen). For cDNA synthesis, reverse transcription reactions and synthesis were carried out by the iScript TM cDNA Synthesis Kit (Bio-Rad ® , Hercules, CA, USA). The cDNA was synthesized by the M-MuLV reverse transcriptase, and PCR was performed with specific primers and the Fast SYBR ® Green Master Mix (Applied Biosystems, Foster City, CA, USA). All the sequences of the primers are listed in Table 1. Each cycle consisted of the following steps: the initial denaturation step for 10 min at 95 • C, denaturation for 3 s at 95 • C, and annealing for 30 s at 60 • C (40 cycles). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was detected as an internal control. The results were analyzed using the software supplied with a machine utilizing the comparative C T (∆∆C T ) method.

Immunofluorescence (IF) Staining and Intensity Quantification
Primary antibodies against microtubule-associated protein 2 (MAP2; GeneTex Inc., Irvine, CA, USA), class III β-tubulin (TUBB3; Santa Cruz Biotechnology, Santa Cruz, CA, USA), neurofilament medium polypeptide (NEFM; OriGene Technologies, Medical Center Drive, Rockville, MD, USA), and neurofilament light polypeptide (NEFL; OriGene Technologies) were added on the slides at an appropriate dilution for IF staining. The slides were then incubated with the secondary antibodies with green fluorescence protein against the primary antibodies for fluorescence analysis. All the slides were counterstained with 4,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) for nuclear (blue) fluorescent staining. Total florescence intensity in one field was measured by the Image-Pro Plus (v4.5.0.29, Media Cybernetics, Silver Spring, MD, USA), and the average intensity from 120 to 140 cells in six fields was calculated for each slide. The average florescence intensities of individual cells in the images were also calculated.

Statistical Analysis
Statistical analysis was performed using Prism (version 5.03, GraphPad, La Jolla, CA, USA.). The Mann-Whitney U test was applied for the two-group comparison and one-way analysis of variance (ANOVA) for the multi-group comparison. All the results were presented as means ± standard errors. A p value of less than 0.05 was regarded as significant.

Confluence-Initiated and Factor-Induced Neurogenic Differentiation Effects on Human ADSCs In Vitro
Spontaneous factor-induced neurogenic differentiation was observed by detecting the neuronal marker gene expressions of the ADSC differentiated cells by RT-qPCR. Neuron-associated gene expressions, such as TUBB3 ( Figure 1A), MAP2 ( Figure 1B), NEFM ( Figure 1C), and NEFH ( Figure 1D), were monitored for seven days. The expressions of TUBB3, MAP2, and NEFH were all significantly enhanced on day 3. The TUBB3 expression on day 1 was significantly higher than that on day 0. The MAP2 expressions on days 5 and 7 were also significantly increased when compared with those on day 0. In addition, the expression of NEFM was significantly elevated on day 5. These genes were highly expressed in the confluence-initiated differentiation of the ADSCs within seven days, which indicated the neurogenic differentiation of ADSCs could spontaneously happen at high confluence. sion on day 1 was significantly higher than that on day 0. The MAP2 expressions on days 5 and 7 were also significantly increased when compared with those on day 0. In addition, the expression of NEFM was significantly elevated on day 5. These genes were highly expressed in the confluence-initiated differentiation of the ADSCs within seven days, which indicated the neurogenic differentiation of ADSCs could spontaneously happen at high confluence. The time course experiments of detecting neuron-associated gene expressions include class Ⅲ βtubulin (TUBB3) (A), microtubule-associated protein 2 (MAP2) (B), neurofilament medium polypeptide (NEFM) (C), and neurofilament heavy polypeptide (NEFH) (D) in the spontaneously differentiated ADSCs after high confluence was achieved in the cultures by RT-qPCR. The value of each gene expression was normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Relative fold changes of the values from day 1 to day 7 were compared with the value of day 0 (without induction) as one. The values are expressed as mean ± S.D. with three experimental replicates. Statistical significance was determined using the Mann-Whitney U test. "*" represents p < 0.05.

Comparison of Confluence-Initiated Neurogenic Differentiation on the ADSCs and BMSCs
To determine the confluence-initiated neurogenic differentiation on the ADSCs and BMSCs, the neuron-associated gene expressions in the factor-induced and confluence-initiated neurogenic differentiation of the ADSCs and BMSCs were compared by RT-qPCR. TUBB3 ( Figure 2D) were all highly expressed in the confluenced-initiated differentiated ADSCs but not in the BMSCs. The expression of TUBB3 of the confluence-initiated differentiation of the ADSCs on days 1 and 3 was significantly higher than that of the BMSCs. MAP2 and NEFH on day 3 and NEFH and NEFM on day 5 in the confluence-initiated differentiation of the ADSCs were significantly higher than that of the BMSCs. The neuron-associated protein expressions, including TUBB3 ( Figure Figure 3D), were detected in the factor-induced neurogenic differentiation of the ADSCs through IF analysis. These proteins were also observed to be upregulated in the confluence-initiated differentiation of the ADSCs ( Figure 3E-H). However, the neurogenic differentiation of the BMSCs could not be induced by confluence since all genes ( Figure 2) and proteins ( Figure 3I-L) were not upregulated. According to the results of quantifying the total immunofluorescent intensity ( Figure 3M) and the fluorescence of each cell (Figure 3N), the intensities of the four proteins in the confluence-initiated differentiation of the ADSCs were significantly higher than those of the BMSC group, except the result of the fluorescent intensity of the MAP2 normalized with cells. Besides, the expression of neurotrophins on day 7, including the nerve growth factor (NGF) ( Figure 4A), brain-derived neurotrophic factor (BDNF) ( Figure 4B), and glial cell-derived neurotrophic factor

Comparison of Confluence-Initiated Neurogenic Differentiation on the ADSCs and BMSCs
To determine the confluence-initiated neurogenic differentiation on the ADSCs and BMSCs, the neuron-associated gene expressions in the factor-induced and confluenceinitiated neurogenic differentiation of the ADSCs and BMSCs were compared by RT-qPCR. TUBB3 (Figure 2A Figure 3N), the intensities of the four proteins in the confluence-initiated differentiation of the ADSCs were significantly higher than those of the BMSC group, except the result of the fluorescent intensity of the MAP2 normalized with cells. Besides, the expression of neurotrophins on day 7, including the nerve growth factor (NGF) ( Figure 4A), brain-derived neurotrophic factor (BDNF) (Figure 4B), and glial cell-derived neurotrophic factor (GDNF) ( Figure 4C), was significantly upregulated in the confluence-initiated differentiation of the ADSCs compared with that of the control group on day 0. (GDNF) ( Figure 4C), was significantly upregulated in the confluence-initiated differentiation of the ADSCs compared with that of the control group on day 0. Statistical significance to compare confluence-initiated differentiation of the ADSCs and BMSCs was determined using one-way ANOVA analysis. "*" represents p < 0.05. "**" represents p < 0.01. "***" represents p < 0.001.   The mean of the value that was normalized to the cell numbers in the field was calculated. All the values are expressed as mean ± S.D. with three experimental replicates. Statistical significance to compare the multiple groups was determined using one-way ANOVA analysis. "*" represents p < 0.05. "**" represents p < 0.01.  . Neurotrophic gene expressions of the ADSCs during factor-induced and confluence-initiated neurogenic differentiation. Neurotrophic gene expressions, including the nerve growth factor (NGF) (A), brain-derived neurotrophic factor (BDNF) (B), and cell-derived neurotrophic factor (GDNF) (C) in the factor-induced (factor) and confluence-initiated (confluence) differentiated ADSCs on day 7 were detected by RT-qPCR. The value of each gene expression was normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Relative fold changes of the values of the factorinduced and confluence-initiated differentiated ADSCs were compared with the value of the control (without induction) as one. The values are expressed as mean ± SEM with three experimental replicates. Statistical significance was determined using the one-way ANOVA test. "*" represents p < 0.05. "**" represents p < 0.01. "***" represents p < 0.001.

Comparison of the ADSCs and BMSCs in Confluence-Initiated Adipogenic and Osteogenic Differentiation
The adipogenic and osteogenic differentiation in the ADSCs and BMSCs was also determined by RT-qPCR. Peroxisome proliferator-activated receptor γ (PPARγ) and lipoprotein lipase (LPL) genes were detected to verify the adipogenic differentiation in the ADSCs (Figure 5A,B) and BMSCs ( Figure 5E,F). Alkaline phosphatase (ALP) and collagen . Neurotrophic gene expressions of the ADSCs during factor-induced and confluence-initiated neurogenic differentiation. Neurotrophic gene expressions, including the nerve growth factor (NGF) (A), brain-derived neurotrophic factor (BDNF) (B), and cell-derived neurotrophic factor (GDNF) (C) in the factor-induced (factor) and confluence-initiated (confluence) differentiated ADSCs on day 7 were detected by RT-qPCR. The value of each gene expression was normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Relative fold changes of the values of the factorinduced and confluence-initiated differentiated ADSCs were compared with the value of the control (without induction) as one. The values are expressed as mean ± SEM with three experimental replicates. Statistical significance was determined using the one-way ANOVA test. "*" represents p < 0.05. "**" represents p < 0.01. "***" represents p < 0.001.

Comparison of the ADSCs and BMSCs in Confluence-Initiated Adipogenic and Osteogenic Differentiation
The adipogenic and osteogenic differentiation in the ADSCs and BMSCs was also determined by RT-qPCR. Peroxisome proliferator-activated receptor γ (PPARγ) and lipoprotein lipase (LPL) genes were detected to verify the adipogenic differentiation in the AD-SCs ( Figure 5A,B) and BMSCs ( Figure 5E,F). Alkaline phosphatase (ALP) and collagen type I (COL I) genes were used to determine the osteogenic differentiation of the ADSCs (Figure 5C,D) and BMSCs ( Figure 5G,H). Adipogenic and osteogenic genes were highly expressed in the factor-induced differentiation of the ADSCs and BMSCs, which explained that fact that differentiation could be induced by a factor-inducing medium. However, there were significantly fewer adipogenic genes in the differentiated ADSCs and BMSCs in the confluent situation. In the confluence-initiated osteogenic differentiation, neither ALP nor COL I was high in the differentiated ADSCs and BMSCs. collagen type Ⅰ (COL Ⅰ) (D,H) of the ADSCs and BMSCs, after the factor-induced (factor) and confluence-initiated (confluence) differentiation on day 7 were detected by RT-qPCR. The value of each gene expression was normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Relative fold changes of the values of the factorinduced and confluence-initiated differentiated ADSCs were compared with the value of the control (without induction) as one. The values are expressed as mean ± SEM with three experimental replicates. Statistical significance was determined using the one-way ANOVA test. "*" represents p < 0.05. "**" represents p < 0.01. "***" represents p < 0.001.

Discussion
In this study, we found that the confluent ADSCs at passages 3 to 5 without additional induction would express neurogenic genes and proteins that showed the tendency toward spontaneous neurogenic differentiation, but not adipogenic and osteogenic differentiation. However, spontaneous neurogenic, adipogenic, and osteogenic differentiation of the confluent BMSCs at the same passages as the ADSCs was not observed in this study.
Stem cells have an inherent propensity to differentiate into different lineages according to their origins. Tendon-derived stem cells from tendon tissue had the potential of spontaneous tendongenic differentiation [44]. ADSCs from early passages (P1-P3) were Figure 5. Detection of adipose-related and osteocyte-related gene expressions of the ADSCs and BMSCs during confluenceinitiated differentiation. Adipose-related genes, including peroxisome proliferator-activated receptor γ (PPARγ) (A,E) and lipoprotein lipase (LPL) (B,F), and osteocyte-related genes, including alkaline phosphatase (ALP) (C,G) and collagen type I (COL I) (D,H) of the ADSCs and BMSCs, after the factor-induced (factor) and confluence-initiated (confluence) differentiation on day 7 were detected by RT-qPCR. The value of each gene expression was normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Relative fold changes of the values of the factor-induced and confluence-initiated differentiated ADSCs were compared with the value of the control (without induction) as one. The values are expressed as mean ± SEM with three experimental replicates. Statistical significance was determined using the one-way ANOVA test. "*" represents p < 0.05. "**" represents p < 0.01. "***" represents p < 0.001.

Discussion
In this study, we found that the confluent ADSCs at passages 3 to 5 without additional induction would express neurogenic genes and proteins that showed the tendency toward spontaneous neurogenic differentiation, but not adipogenic and osteogenic differentia-tion. However, spontaneous neurogenic, adipogenic, and osteogenic differentiation of the confluent BMSCs at the same passages as the ADSCs was not observed in this study.
Stem cells have an inherent propensity to differentiate into different lineages according to their origins. Tendon-derived stem cells from tendon tissue had the potential of spontaneous tendongenic differentiation [44]. ADSCs from early passages (P1-P3) were determined to have better abilities to differentiate into adipocytes spontaneously [45]. Moreover, the potential of spontanous adipogenic, chondrogenic, and osteogenic differentiation on ADSCs maintained in a proliferation medium for 14 days without further passaging was also found [46]. However, spontaneous differentiation might not be beneficial to MSCs. MSCs from the placenta went through aging and spontaneous osteogenic differention upon routine in vitro cultative expansion, and the supplementation of the basic fibrobliast growth factor (bFGF) could enhance proliferation and suppress spontaneous differentiation. One explanation is that the physiological niches of MSCs would prevent spontaneous differentiation and promote self-renewal [47]. Extensive passaging (>passages 10) and low density seeding were reported to cause spontaneous osteogenic differentiation of ADSCs and their decreased adipogenic potential, the results of which might link to cell senescence [48]. Rat BMSCs were also found to have the tendency of spontaneous osteogenic differentiation after being cultured for 21 days without osteogenic induction [49]. However, the cultative passages of the ADSCs and BMSCs were not longer than five times, and the culture period for each passage was not longer than seven days in this study. The MSCs were still young enough to keep the stemness and prevent spontaneous differentiation. These might be the reasons why the results of this study did not exhibit spontaneous adipogenic and osteogenic differentiation of the ADSCs and BMSCs.
In this study, we found that the confluent human ADSCs at passages 3 to 5, which had been continuously cultured for seven days, had the tendency toward spontaneous neurogenic differentiation. The cause of spontaneous neurogenic differentiation in the ADSCs might be due to cell confluence, which was the reason why the differentiation was "confluence-initiated". ADSCs cultured at high density highly expressed growth factors that are associated with differentiation [39]. In addition, ADSCs from young donors exhibited high baseline expression levels of neural and glial markers [50]. The source of the ADSCs in this study was from females under 30 years old. We suggest that spontaneous neurogenic differentiation of ADSCs from young donors could be induced by cell confluence, and without a long-term culture in vitro. One explanation suggested that MSCs had the tendency toward neurogenic differentiation due to the origin of the neuropithelium or neural crest [51]. Another reason was that the cellular stress of the in vitro artificial culture enviroment might cause spontaneous neurogenic marker expressions of MSCs [52]. However, the mechanism needs to be studied further.
MSCs can secrete trophic factors, such as hepatocyte growth factor (HGF), stromal derived factor-1 (SDF-1), insulin-like growth factor (IGF-1), epithelial growth factor (EGF), nerve growth factor (NGF), transforming growth factor-alpha (TGF-α), and tissue angiogenesis vascular endothelial growth factor (VEGF), which enhance cell survival. VEGF, TGF-β, and β-FGF, which play important roles in wound healing, contribute to angiogenesis, granulation tissue formation, and re-epithelialization [42]. The growth factors may be stimulated by the use of ADSCs. On the other hand, platelet-rich plasma (PRP) is a source of growth factors, such as IGF-1, VEGF, etc., that can also be applied to wound healing [12]. The confluence and differentiation of ADSCs may be affected when ADSCs and PRP are co-administered. Neurotrophins, including NGF, BDNF, and GDNF, help the growth and function of neurons, and they were detected in the conditioned medium of the spontaneously differentiated ADSCs to confirm the neurogenic differentiation potential in this study. In the confluent population of ADSCs, neurotrophins secreted by the spontaneous neurogenic differentiated ADSCs may easily enhance other peripheral ADSCs toward neural differentiation. However, one limitation of this study is that it is still unknown whether confluence-initiated spontaneous neurogenic differentiation could make MSCs become functional neurons or not. Further studies are needed. In a previous study, primary rat BMSCs were observed to spontaneously differentiate into neural precursor cells after a six-week culture [53]. Another study declared rat BMSCs were also found to self-differentiate into neuron-like cells, and the researchers proposed that increased NGF and the gene expressions of the receptors TrkA and TrkB in rat BMSCs without external induction might cause self-differentiation [54]. However, confluence-initiated spontaneous neurogenic differentiation of the BMSCs was not observed in this study. Due to the fact that the BMSCs used in this study were delinked from the donors, we could not reveal the correlation of donor age and spontaneous neurogenic differentiation in the BMSCs.
The spontaneous neurogenic differentation potential of ADSCs might be increased due to cellular confluence, but the lack of evidence suggests that mature neurons could be derived from spontaneous differentiated ADSCs. This study reveals the possibility of the clinical application for confluent ADSCs to be injected into a nerve conduit to treat nerve injury. Compared with BMSCs, highly confluent ADSCs may tend to exhibit spontaneous neural differentiation that benefits nerve repair.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper.