NEURAL DIFFERENTIATION OF ADIPOSE-DERIVED STEM CELLS BY INDIRECT CO-CULTURE WITH SCHWANN CELLS

To investigate whether adipose-derived stem cells (ADSCs) could be subject to neural differentiation induced only by Schwann cell (SC) factors, we co-cultured ADSCs and SCs in transwell culture dishes. Immunoassaying, Western blot analysis, and RT-PCR were performed (1, 3, 7, 14 d) and the co-cultured ADSCs showed gene and protein expression of S-100, Nestin, and GFAP. Further, qRT-PCR disclosed relative quantitative differences in the above three gene expressions. We think ADSCs can undergo induced neural differentiation by being co-cultured with SCs, and such differentiations begin 1 day after co-culture, become apparent after 7 days, and thereafter remain stable till the 14th day.


INTRODUCTION
In 2000, Woodbury et al. were the first to induce bone marrow stromal cells (BMSCs) to differentiate into the neuronal phenotype in vitro, expressing neuron-specific enolase, NeuN, neurofilament-M, tau, trKA, and Nestin (Woodbury et al., 2000).Since then, the possibility of obtaining neural cells from BMSCs, as a new biological concept with potential clinical application, has been extensively investigated (Chopp et al., 2002;Hermann et al., 2004;Mahmood et al., 2004;Zurita et al., 2004Zurita et al., , 2006;;Bossolasco et al., 2005).Bone marrow stromal cells would be an ideal source for nerve tissue engineering if the procedures to harvest them were not so painful and invasive, and if the procedures could yield enough BMSCs more easily (Zuk et al., 2001).
Adipose-derived stem cells (ADSCs) are isolated from adipose tissue, which is derived from embryonic mesoderm.These ADSCs are self-renewable and can differentiate along several mesenchymal tissue lineages, including adipocytes, osteoblasts, myocytes, chondrocytes, endothelial cells, and cardiomyocytes (Zuk et al., 2001;Gimble et al., 2007).Such cells can also be induced into neuro-like cells or a Schwann cell phenotype in vitro (Safford et al., 2002;Ning et al., 2006;Kingham et al., 2007), and intra-cerebral transplantation of human ADSCs can improve neurological deficits after cerebral ischemia in rats (Kang et al., 2003).In addition, liposuction is a common and safe surgical procedure, while subcutaneous adipose tissue is abundant, readily accessible, and relatively expendable, so enough ADSCs can be obtained with minimal risk (Fraser et al., 2006).Adipose-derived stem cells seem more suitable than BMSCs for nerve tissue engineering as alternative cell sources of Schwann cells (SCs).
In all methods of inducing BMSC neural differentiation in vitro, co-culture seems a simple and effective approach.After performing direct coculture of SCs and BMSCs, Zurita suggested that intercellular contacts could play an important role in the process of BMSC neural differentiation (Zurita et al., 2005).However, according to previous studies (Ying et al., 2002;Alvarez et al., 2003;Wang et al., even if without intercellular contacts (Zurita et al., 2007).This means that only by means of cell factors secreted by SCs, BMSCs can also differentiate into neuro-like cells successfully.But until now, there were no reports about whether ADSCs could realize neural differentiation only by co-culture with SCs.To solve this problem, we applied transparent polycarbonate membranes with pores of 1.0 µm to impede contact between SCs and ADSCs and performed an indirect co-culture, hoping to induce ADSC neural differentiation and investigate the regularity of such differentiation at different time points (1, 3, 7, 14 d) after co-culture.

MATERIAL AND METHODS
All animal experiments described in this report were approved by the Ethical Guideline Committee for Animal Care of West China College of Medical Sciences, Sichuan University.

Isolation, culture, and identification of ADSCs and SCs
Adipose-derived stem cells were isolated from subcutaneous adipose tissue of the inguinal region of adult Sprague-Dawley rats.The subcutaneous fat was carefully dissected and then enzymatically dissociated for 60 min at 37°C using 0.1% (w/v) collagenase type I (Sigma, UK) and passed through a 75-μm filter to remove undissociated tissues, after which it was neutralized by the addition of alpha-Modified Eagle Medium (α-MEM; Hyclone, USA) containing 20% (v/v) fetal bovine serum (FBS, Hyclone, USA) and centrifuged at 1000×g for 8 min.The stromal cell pellet was resuspended in α-MEM containing 20% (v/v) FBS (Hyclone, USA) with 1% (v/v) penicillin/streptomycin solution.Cultures were maintained at subconfluent levels in an incubator at 37°C with 5% CO 2 and passaged with trypsin/EDTA (Sigma, UK) when required.Schwann cells were isolated from the bilateral sciatic nerves of 1-to 2-day-old Sprague-Dawley rat pups according to previous reports (Zurita et al., 2005) and cultured in α-MEM containing 20% (v/v) FBS (Hyclone, USA) with 2 μM forskolin (Sigma, UK), 20 ng/ml basic fibroblast growth factor (bFGF; PeproTech Ltd., UK), and 1% (v/v) penicillin/strep-tomycin solution.The culture medium was changed every 72 h.
Adipose-derived stem cells at passage 3 were identified by immunocytochemistry with antibodies against Stro-1, CD90, CD10, CD146, CD44, and CD34 (1:100, Santa Cruz, CA, USA).The multiple differentiation potential of these ADSCs was also investigated as previously described (Kingham et al., 2007).For osteogenic induction, cultures were treated with 100 μg/ml ascorbate, 0.1 μM dexamethasone, and 10 mM β-glycerophosphate for 3 weeks.Cells were then fixed with 4% (w/v) paraformaldehyde for 30 min at room temperature, washed three times with phosphate-buffered saline (PBS), and then incubated with 1% (w/v) Alizarin Red solution to stain for calcium deposition.For adipogenic induction, cells were treated with 0.5 mM isobutylmethylxanthine (IBMX), 1 µM dexamethasone, 10 µM insulin, and 200 µM indomethacin for 3 weeks.Cells were then fixed with 10% (v/v) formalin for 60 min, washed in H 2 O, and stained for lipid clusters with Oil Red O. Stem cells were identified in living cultures on the basis of cell soma and nuclear morphology.They were also identified by immunocytochemical labeling for S100 protein with antibody against S100 (1:100, Santa Cruz, CA, USA).The cover slips were examined using an Olympus IX 70 inverted fluorescence microscope.

Co-culture of ADSCs and SCs
To co-culture ADSCs and SCs, we used transwell culture dishes with a polycarbonate membrane (Millipore, USA).The pore size of membranes in our experiment was 1.0 µm, which is much smaller than size of the ADSC body (approximately 80-160 µm) and therefore greatly inhibits the migration of ADSCs into the lower chamber.Schwann cells (6×10 4 cells/cm 2 , in 3 ml of culture medium) were cultured in the well, and ADSCs (3×10 4 cells/cm 2 ) were cultured on the permeable membrane support, with co-cultures maintained in α-MEM containing 20% (v/v) FBS, so that both types of cells were exposed to the same culture media conditions, without cell contacts (Zurita et al., 2007).Meanwhile, ADSCs cultured on the permeable membrane support without SCs in the lower chamber were taken as control groups.At different time points (1, 3, 7, 14 d) after co-culture, all the membranes were cut down from the insert and mounted on glass cover slips for immunostaining.
Protein from co-cultured ADSCs and ADSCs of control groups was extracted with a commercially available protein assay kit (Bio-Rad, UK).After determination of protein concentrations, 20 μg of the protein was mixed with an equal amount of 2 × sodium dodecyl sulfate (SDS) loading buffer for Western blot analysis as previously reported (Kingham et al., 2007).Proteins were denatured at 99°C for 5 min, then resolved at 120 V on 15% (for S100) or 10% (for GFAP and Nestin) sodium dodec-yl sulfate polyacrylamide gels.Following electrophoretic transfer to polyvinylidene difluoride membranes, membranes were blocked for 1 h in 5% (w/v) non-fat dry milk in TBS-Tween (10 mM Tris pH 7.5, 100 mM NaCl, 0.1% (v/v) Tween).The membranes were then incubated overnight at 4°C with primary antibodies, including mouse antibodies to S100 (1:200, Santa cruz, CA, USA), mouse antibodies to nestin (1:200, Santa Cruz, USA), mouse antibodies to GFAP (1:200, Santa Cruz, USA), and GAPDH (1:500, Santa Cruz, USA).After washing with TBS-Tween, membranes were incubated with horseradish peroxidase-conjugated secondary antibody (goat anti-mouse, 1:1000; Santa Cruz, USA) at room temperature for 1 h.The loading control was GAPDH antibody.The blots were developed via enhanced chemiluminescence (Pierce, Rockford, IL, USA) and were digitally scanned (Bio-Rad, CA, USA).

RNA reverse transcriptase polymerase chain reaction (RT-PCR) and quantitative real-time PCR (qRT-PCR) analysis
Total RNA was extracted from all of the specimens at different time points (1, 3, 7, 14 d) using the Trizol Reagent (Invitrogen, USA) according to the manufacturer's instructions.The yield and quality of the RNA were assessed by measuring the absorbance at 260 and 280 nm, followed by electrophoresis on 3% (w/v) agarose gels.

Statistical analyses
All data were expressed as the mean ± SEM.Statistical calculations were performed using the SPSS software (SPSS, Chicago, IL, USA).Inter-group comparisons were made by one-way analysis of variance (ANOVA), after which Tukey's honestly significant difference test (Tukey HSD) was adopted to perform multiple inter-group comparisons.Values of P less than 0.05 were considered to be significant.

RESULTS
Immunocytochemical assessment of ADSC stem cell markers showed positive for Stro-1, CD90, CD10, CD146, and CD44, but totally negative for CD34 (Fig. 1).As there were no specific markers, ADSCs were identified on the basis of a complex immune phenotype characterized by the lack of a hematopoietic stem cell marker such as CD34 and expression of a number of surface molecules, including CD90, CD10, CD146, CD44, and STRO-1 (Ning et al., 2006;Krampera et al., 2007).To investigate the mul-   tiple differentiation potential of cells isolated from subcutaneous adipose tissue, they were treated with agents known to induce differentiation into osteoblastic and adipogenetic phenotypes.Osteogenic differentiation was confirmed by the production of calcium deposits detected with Alizarin Red (Fig. 1) and adipogenic differentiation by the presence of Oil Red O-positive lipid clusters (Fig. 1).With respect to morphology, SCs under a phase contrast microscope had a typical elongated spindle shape, and immunocytochemical assessment showed positive for S100 protein with high purity (Fig. 2).
The morphology of co-cultured ADSCs began to change as early as 24 h after co-culture, showing the elongated spindle shape characteristic of SCs, and co-cultured ADSCs with that shape became more and more numerous with the passage of time.Nestin protein is a neural progenitor cell marker, while both S-100 and GFAP protein are glial markers; they are all cell markers for identifying SCs (Ning et al., 2006;Kingham et al., 2007).Our immunocytochemical assessment of co-cultured ADSCs showed positive Fig. 3. Morphological and immunocytochemical assessments of co-cultured ADSCs.With respect to morphology, co-cultured AD-SCs showed the elongated spindle shape characteristic of Schwann cells as early as 24 h after co-culture (a), while at 3, 7, and 14 days after co-culture, co-cultured ADSCs with that shape became more and more numerous (e, i, m).Immunocytochemical assessments of co-cultured ADSCs at different time points (1, 3, 7, 14 d) showed positive for S-100 (b, f, j, n), Nestin (c, g, k, o), and GFAP (d, h, l, p), but the percentage of positive cells could not be counted precisely because the transparency of the polycarbonate membranes was not good enough, so we could not perform precise quantitative analyses only on the basis of immunocytochemical pictures.Scale bar = 100 µm.
for S-100, Nestin, and GFAP protein at all time points (1, 3, 7, 14 d), but the percentage of positive cells could not be counted precisely because the transparency of the polycarbonate membranes was not good enough, so we could not perform a precise quantitative analysis using only immunocytochemical pictures (Fig. 3).On the other hand, cultured ADSCs in the control groups kept their original small round or polygonal shape without changing to an elongated spindle shape, and these undifferentiated ADSCs showed negative for S-100, Nestin, and GFAP protein all the time.To confirm the results obtained by immunocytochemistry, Western blotting was performed (Fig. 4).Lyzates of co-cultured ADSCs at 1, 3, 7, and 14 days -but not ADSCs from the control groups -showed a GFAP-immunoreac- Lyzates of co-cultured ADSCs at 1, 3, 7, and 14 days and lyzates of ADSCs from control groups were blotted for S100, Nestin, and GFAP proteins (GAPDH was used as a loading control).Fig. 5. RT-PCR gel electrophoresis images.In above figures, all the co-cultured ADSCs at 1, 3, 7, and 14 days show positive strips for gene S100, Nestin, and GFAP in different degrees, while the ADSCs from control groups show no positive strips for S100, Nestin, and GFAP.We use GAPDH as a reference for assessing the PCR efficiency.For the relative quantities of S100 and Nestin gene expression, analyses showed no statistical differences between 1 and 3 d and between 7 and 14 d (P > 0.05), but the results for 7 or 14 d had significant statistical differences compared to those for 1 or 3 d (P < 0.05); for the relative quantities of GFAP gene expression, there were no statistical differences at 1, 3, 7, and 14 d (P > 0.05).
tive band corresponding to a molecular weight of 50 kDa.Nestin and S100 proteins were also detected in co-cultured ADSCs, but were absent in ADSCs from the control groups.
We also performed RT-PCR, and the results were consistent with immunocytochemistry and western blotting (Fig. 5).Subsequently, quantitative real-time PCR was performed to investigate the relative quantitative differences in gene expression of the above three genes at different time points (1, 3, 7, 14 d).For S100 and Nestin, analyses showed no statistical differences between 1 and 3 d and between 7 and 14 d (P > 0.05), but results for 7 and 14 d had significant statistical differences compared to those for 1 and 3 d (P < 0.05), which meant that the relative quantities of S100 and Nestin gene expression were higher at 7 and 14 d than at 1 and 3 d (Tables 2 and 3).For GFAP, there were no statistical differences at 1, 3, 7, and 14 d (P > 0.05), which meant that the relative quantity of GFAP gene expression maintained a stable level from 1 to 14 days without apparent fluctuation (Table 4).To make the results clear and direct, we used a bar graph to show relative quantitative differences at different time points (1, 3, 7, 14 d) for S100, Nestin, and GFAP (Fig. 6).

DISCUSSION
Our studies showed that ADSCs can differentiate into a Schwann cell phenotype only as a result of being co-cultured with SCs indirectly, excluding the possibility of cell fusion.Previous studies reported that BMSCs can be induced to undergo neural differentiation in the presence of certain cell growth factors such as BDNF, NGF, bFGF, PDGF, etc. (Sanchez et (Sanchez et al., 2000;Tohill et al., 2004;Tao et al., 2005;Caddick et al., 2006).Recently, two papers reported that when Recently, two papers reported that when supplemented with certain cell growth factor compounds, ADSCs can also be induced to differentiate into a Schwann cell phenotype (Kingham et al., 2007;Krampera et al., 2007).Adult SCs can secrete BDNF and NGF (Yamamoto et al., 1993), which makes it possible for stem cells to differentiate into neurolike cells when co-cultured with SCs.Zurita et al. performed indirect co-culture of SCs and BMSCs in transwell culture dishes and demonstrated neural differentiation of BMSCs (Zurita et al., 2007).We observed that ADSCs began morphologic change as early as 1 day after co-culture, expressing S100, Nestin, and GFAP to judge from immunocytochemical staining and RT-PCR analysis, and such expressions lasted until two weeks later.These results showed that neurotrophic SC factors can induce rapid and stable neural differentiation of ADSCs, which is in keeping with previous research (Tao et al., 2005;Kingham et al., 2007;Zurita et al., 2007).On the basis of the above analyses, we believe that co-culture represents a simple and effective way of inducing neural differentiation of ADSCs.
In our studies, S100 protein, a special cell marker of Schwann cells, showed a relative quantity of gene expression that climbed from the 1 st to 7 th day, then remained stable until the 14 th day.Nestin, a neural progenitor marker, exhibited the same regularity as S100 gene expression.As for GFAP, a glial marker, although we could see directly from the bar graph (Fig. 6) that the relative quantities of its gene expression increased continually from the 1 st to 14 th day, there were no statistical differences between 1, 3, 7, and14 d.The S100 protein, Nestin, and GFAP are all recognized as cell surface markers of SCs (Zurita et al., 2005(Zurita et al., , 2007;;Caddick et al., 2006;Kingham et al., 2007).Their strong gene expression in our co-cultured ADSCs indicated neural differentiation toward a Schwann cell phenotype.Analyzing the above results, we conclude that although ADSCs began neural differentiation as early as 1 day after co-culture, significant neural differentiation appeared at least 7 days later, and such differentiations can last at least two weeks.In other words, we needed at least 7 days to co-culture ADSCS with SCs in vitro before the co-cultured ADSCs could be transplanted into areas with peripheral nerve defects in vivo for nerve repair.Zurita et al. suggested previously that in vitro neural differentiation of BMSCs is unnecessary when BMSCs are used in grafting procedures for nervous system repair (Zurita et al., 2007).However, further investigation is needed to establish whether prior in vitro neural differentiation of ADSCs would be more effective for nerve repair in vivo.

Fig. 2 .
Fig. 2. Identification of SCs.a.Immunocytochemical assessment of Schwann cells showing positive for S100 protein, b. morphology of Schwann cells under phase contrast microscope showing typical elongated spindle shape.Scale bar = 100 µm.

Fig. 4 .
Fig. 4. Western blot analysis of Schwann cell-related proteins.Lyzates of co-cultured ADSCs at 1, 3, 7, and 14 days and lyzates of ADSCs from control groups were blotted for S100, Nestin, and GFAP proteins (GAPDH was used as a loading control).

Table 2 .Fig. 6 .
Fig. 6.Relative differences in S100, Nestin, and GFAP gene expression of co-cultured ADSCs.In the graph, bar height represents the relative quantity of gene expression (mean + SEM).For the relative quantities of S100 and Nestin gene expression, analyses showed no statistical differences between 1 and 3 d and between 7 and 14 d (P > 0.05), but the results for 7 or 14 d had significant statistical differences compared to those for 1 or 3 d (P < 0.05); for the relative quantities of GFAP gene expression, there were no statistical differences at 1, 3, 7, and 14 d (P > 0.05).
RAD, USA).During the real-time PCR process, the fluorescence signal was collected during annealing steps.The value of Ct (cycle of threshold) was calculated for further statistical analysis.In order to examine the efficiency of real-time PCR, standard curves were established with serial dilution of sample RNA (500 ng; S100, GFAP, Nestin, GAPDH, 10 × dilution).Total RNA of ADSCs in control groups was taken as a non-template control, which was included in all experiments and performed with RT-PCR to evaluate DNA contamination of reagents used for amplification.None of the experiments resulted in a positive signal from the non-template controls.

Table 1 .
Primer sets for RT-PCR and quantitative real-time PCR analysis.For. and Rev. indicate forward and reverse primers, respectively.All primers were purchased from Takara Biological Technology Co. Ltd, (Dalian, China).