Synaptic transmission of neural stem cells seeded in 3-dimensional PLGA scaffolds

Biomaterials 30 (2009) 3711–3722

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Synaptic transmission of neural stem cells seeded in 3-dimensional PLGA scaffolds Yi Xiong a, Yuan-Shan Zeng a, b, *, Chen-Guang Zeng c, Bao-ling Du a, Liu-Min He c, Da-Ping Quan c, Wei Zhang a, Jun-Mei Wang a, Jin-Lang Wu d, Yan Li a, Jun Li e, f a

Division of Neuroscience, Department of Histology and Embryology, Zhongshan School of Medicine, Sun Yat-sen University, Guangzhou 510080, China Research for Stem Cell Biology and Tissue Engineering, Sun Yat-sen University, Guangzhou 510080, China c School of Chemistry and Chemical Engineering, Sun Yat-sen University, Guangzhou 510127, China d Department of Electron Microscope, Institute of Spinal Cord Injury, Zhongshan School of Medicine, Sun Yat-sen University, Guangzhou 510080, China e Department of Neurology, Wayne State University School of Medicine, Detroit, MI 48201, USA f John D. Dingell VA Medical Center, Detroit, MI 48201, USA b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 March 2009 Accepted 19 March 2009 Available online 17 April 2009

To explore therapeutic potential of engineered neural tissue, we combined genetically modified neural stem cells (NSCs) and poly(lactic acid-co-glycolic acid) (PLGA) polymers to generate an artificial neural network in vitro. NSCs transfected with either NT-3 or its receptor TrkC gene were seeded into PLGA scaffold. The NSCs were widely distributed and viable in the scaffold after culturing for 14 days. Immunoreactivity against Map2 was detected in >70% of these grafted cells, suggesting a high rate of differentiation toward neurons. Immunostaining of synapsin-I and PSD95 revealed formation of synaptic structures, which was also observed under electron microscope. Furthermore, using FM1-43 dynamic imaging, synapses in these differentiated neurons were found to be excitable and capable of releasing synaptic vesicles. Taken together, our artificial PLGA construct permits NSCs to differentiate toward neurons, establish connections and exhibit synaptic activities. These findings provide a biological basis for future application or transplantation of this artificial construct in neural repair. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Neural stem cell Transplantation PLGA Synapse NT-3 TrkC

1. Introduction Spinal cord injury (SCI) is a highly prevalent medical problem, which may be resulted from motor vehicle accidents or other trauma [1]. For instance, there is an annual incidence of 40 SCIs per million populations in the USA (http://www.spinalcord.uab.edu/). Victims of SCI are often afflicted by severe neurological disabilities, yet there has been no effective treatment up to date. The pathophysiological processes in SCI are multifactorial, involving blood vessel rupture, ischemia, edema, metabolic derangement, and free radicals formation in acute phase, and followed by axonal degeneration/regeneration, loss of glial cells, demyelination/remyelination, and formation of cavities in the injured site [2]. These devitalized tissues have been proposed to be replaced by exogenous neural tissues to restore functions. Toward this end, effort has been made to transplant neural tissues to mitigate neurological disabilities in SCI [3].

* Correspondence to: Yuan-Shan Zeng, Department of Histology and Embryology, Zhongshan School of Medicine, Sun Yat-sen University, 74 Zhongshan 2nd Road, Guangzhou 510080, China. Tel./fax: þ86 20 8733 1452. E-mail address: [email protected] (Y.-S. Zeng). 0142-9612/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2009.03.046

Neural stem cell (NSC) are the cells that can continuously selfrenew and differentiate into both neuronal and glial lineages, including neuron, astrocyte, and oligodendrocyte [4,5]. These features render therapeutic potential for neurological diseases [6], such as SCI [3,7–11]. However, effectiveness of NSC transplantation in the SCI is limited, particularly when descending/ascending pathways are separated by a gap. Grafted NSCs alone are often incapable of forming a neural network to bridge the gap in the injured spinal cord [7,10,11]. To circumvent this problem, engineered tissues, using poly(Llactic acid) (PLLA) or poly(lactic acid-co-glycolic acid) (PLGA) polymers, have been investigated to be used as a structural frame of the bridge. Tissue compatible polymers have been used to carry exogenous cells (such as Schwann cells or NSCs) into SCI animal model and promote recovery of hindlimb motility [3,12–14]. This approach appears sustainable in long term. For instance, in hemisection SCI model of rat, implantation of PLGA scaffold with NSCs into the injured site resulted in a functional improvement for one year. The transplantation reduced tissue loss and glial scarring [3]. PLLA scaffold-Schwann cells construct also improved the hindlime motor recovery in completed transection model [13]. Both PLLA and PLGA appear biocompatible in the injured site of spinal cord [3,12–14], and can be easily manufactured. PLGA

3712

Y. Xiong et al. / Biomaterials 30 (2009) 3711–3722

scaffold is biodegradable and may prevent scarring and cyst formation in the SCI animal models [3,15]. While disassociated single-cells may have compromised viability [16], neurospheres of NSCs seeded into scaffolds of the polymers showed improved survival and differentiation [16,17]. However, the biological basis for the therapeutic effect of NSC–polymer construct is still unclear. Regeneration of axons only passes through the injury epicenter, but not through the injury core after the construct is transplanted into the SCI animal model [3]. Whether NSCs are able to form functional synapses and neuronal network in the construct, an indispensable pre-requisite for re-establishing effectively functioning connections in the injured spinal cord, is unknown. Neurotrophins (NTs) are known to play important roles in neural survival, differentiation, and neurite outgrowth [18,19]. NSCs transfected with NT-3 yield a high percentage of differentiation toward neurons [20,21]. This percentage may be further increased when NSCs over-expressing NT-3 are mixed with NSCs expressing tyrosine receptor kinase C (TrkC), the NT-3 receptors [20]. Therefore, the present study engineers a PLGA neural construct containing grafted NSCs. The grafted NSCs transfected with either NT-3 or TrkC were mixed to optimize their differentiation toward neurons. This model is then allowed us to examine the formation of neural network in the artificial neural construct. 2. Materials and methods 2.1. Preparation of PLGA scaffolds Macroporous PLGA scaffold was synthesized as described in our earlier work [22]. Briefly, PLGA with an 75:25 monomer ratio (D,L-lactide:glycolide) was synthesized by ring-opening polymerization using Sn (Oct)2 as catalyzer and dodecenol as initiator. The average molecular weight of PLGA copolymer was 1.22  105 (Mn GPC). To obtain different pore sizes, the polymer concentration was increased from 2.5% to 20%. Sodium chloride as porogen was added into the polymer solution with PLGA/NaCl in a weight ratio of 1:9. PLGA scaffold formed variable sizes of pore from a few mm to 200 mm that were suitable for seeded neurospheres with diameters of 100–300 nm. PLGA rods with longitudinal parallel-channels were fabricated by an injection molding, combined with thermally induced phase separation. The lumens of the mold were pretreated with chlorotrimethylsilane, and placed into the freezer at 40  C for at least one hour. Five% (w/v) PLGA solution in 1,4-dioxane was injected into the cold mold quickly with a syringe, and kept the injection pressure until the polymer solution at the injection port of syringe was completely frozen. The mold was placed in the 40  C freezer for another two hours. The scaffolds were then lyophilized under 0.940 mbar at 0  C for at least four days. The polymer scaffold was trimmed into a rod shape with two cm in length and five mm in diameter. The rods were sterilized by 70% alcohol for 10 min, rinsed with sterilized phosphate buffered solution (PBS, pH 7.4) for 30 min, and were stored in a desiccator. Upon seeding NSCs into the PLGA for culture, the PLGA rod was cut into two mm thickness slices in transverse for the seeding. 2.2. NSCs preparation and identification NSCs were prepared as described previously [8]. Briefly, three-to-five-days-old Sprague–Dawley (SD) rat pups were anesthetized. Whole hippocampi were

dissected and dissociated in D-Hanks’ balanced salt solution (HBSS). After centrifuging at 1000 rpm for five minutes, the supernatant was removed. Pellet was resuspended in five ml basal medium including: DMEM/F12 (1:1) containing B27 supplement (20 ml/ml, Gibco, CA, USA) and bFGF (20 ng/ml, Invitrogen, CA, USA). The cells were plated onto 75-ml culture flasks. The medium was replaced every three days. Typically, the cells grew as suspending neurospheres and were passaged approximately once per week. To confirm the neurospheres were nestin-positive cells, cultures were fixed with 4% formaldehyde in 0.1 M phosphate buffer (pH 7.4) for 30 min at room temperature and rinsed in PBS, labeled with monoclonal antinestin (Table 1), a marker of neural precursor cell, followed by incubation with FITCconjugated anti-mouse IgG (1:1000, Jackson Immunological Research, PA, USA). The slides were examined under fluorescence microscope. 2.3. NSCs transfection and seeding in PLGA scaffold Recombinant adenoviral (Adv) vectors (Ad-NT-3 and Ad-TrkC) were produced as described in our previous study [20]. Ad-LacZ (gift from Dr. Huang WL) was used as control [10,11,20]. Neurospheres were infected with Ad-LacZ, Ad-NT-3 or Ad-TrkC at a multiplicity of infection (MOI) of 50. The cells were resuspended in five ml fresh basic medium and plated onto 25 ml-culture flasks for 48 h. Five experimental groups were established: NSCs only; LacZ–NSCs; TrkC–NSCs; NT-3-NSCs; NT-3/TrkC– NSCs (co-culture of NT-3-NSCs mixed with TrkC–NSCs). The transfection efficiency with these viral vectors has been published multiple times in the past [10,11,20]. Before seeding, we dissociated neurospheres mechanically and counted the number of cells on the cell board. We extracted 20 ml cell suspension and drop one drop on the board. Under light microscope, cells (A single neurosphere was regarded as one cell) were counted in the grids. Cells or neurospheres were also counted when they were on the right or upper border of the rectangular grid. The number of cells was calculated using a formula ((the total number of four quadrants of grids/ 100)400  104). For the NT-3/TrkC group, we first counted the cell number for both NT-3 and TrkC group, and then mixed them in 1:1 ratio. Difficulties were encountered initially to seed neurospheres into the scaffold. To circumvent this problem, 2.4  106 cells in 20 ml culture medium (including 1:1 DMEM/F12 and 10% fetal bovine serum) (TBD, Co, Tianjin, China) were placed on the top of PLGA slice. The seeding was facilitated by placing a Waterman filter paper (#1) underneath the slice to gently suck cells into the pores. The slices were incubated in 35 mm culture dish for 14 days. The culture medium was replaced every two days. In some experiments for the NT-3/TrkC group, 100 nM K252a (Calbiochem, Darmstadt, Germany), an inhibitor of neurotrophin-related tyrosine kinase, was added into the culture medium [21]. 2.4. Live–dead staining To evaluate the viability of the cells grafted into the PLGA, NSCs only, LacZ and NT-3/TrkC were seeded into PLGA slices with two mm thickness. After 14 days culture, the slices were rinsed with 0.1 M phosphate buffer (pH 7.4) for 30 min. The slices were incubated in two ml 0.1 M phosphate buffer containing 2 mM of calceinAM and 4 mM of ethidium homodimer (EthD-III) (Viability/Cytotoxity Assay Kit for Live & Dead Animal Cells, Biotium, USA) for one hour at 37  C. The slices were rinsed three times (20 min for each), and then fixed with 4% formaldehyde in 0.1 M phosphate buffer (pH 7.4) for 30 min. Transverse sections with 20 mm thickness were cut in succession using a cryostat. The 1st to the 5th sections were defined as peripheral sections and the 46th to the 50th as center section. The live-cells stained with calcein-AM show green color and the dead cells stained with EthD-III show red color under fluorescent microscope. Cell death rate was calculated by determining the percentage of EthD-1-positive cells over total cell number. For each experimental group, at least five fields of each section (including 4 corners and one centre) were imaged. The final cell death rate was derived from the average of three sets of experiments.

Table 1 Primary antibodies. Antibodies

Source

Species

Type

Dilution

Reference

Nestin Map2 GFAP MOSP PSD95 Synapsin-I p-C-jun b III tubulin p-p38 ChAT 5-HT Glutamate GABA

Chemicon Sigma Sigma Chemicon Sigma Sigma Santa Cruz Sigma Cell Signaling Chemicon Sigma Boster Boster

Mouse Mouse Rabbit Mouse Mouse Rabbit Mouse Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit

Monoclonal IgG Monoclonal IgG Polyclonal IgG Monoclonal IgG Monoclonal IgG Polyclonal IgG Monoclonal IgG Polyclonal IgG Polyclonal IgG Polyclonal IgG Polyclonal IgG Polyclonal IgG Polyclonal IgG

1:500 1:500 1:500 1:1000 1:500 1:1000 1:500 1:500 1:1000 1:500 1:500 1:200 1:200

Jahr et al., 2003 Binder et al., 1986 Debus et al., 1983 Mu et al., 1994 Kornau et al., 1995 Stone et al., 1994 Yatsushige et al., 2005 Zhang et al., 2006 Ge et al., 2005 Nunes-Tavares et al., 2000 Peressini et al., 1984 Ma et al., 2007 Li et al., 2005

Y. Xiong et al. / Biomaterials 30 (2009) 3711–3722 2.5. Immunocytochemistry (ICC) Synaptogenesis, neurotransmitter and phosphorylation of C-jun were determined using ICC staining with antibodies against specific proteins. ICC has been described in our previous publications [10,11]. In brief, the scaffolds were fixed with 4% formaldehyde in 0.1 M phosphate buffer (pH 7.4) for 30 min. Transverse sections of 30 mm thickness were cut using a cryostat. The sections were incubated with primary antibodies mixed in 0.3% Triton X-100 overnight at 4  C, followed by incubation with secondary antibodies of FITC- or Cy3-conjugated anti-mouse or anti-rabbit IgG (1:1000, Jackson Immunological Research). The slides were examined under fluorescence microscope. All primary antibodies are listed in Table 1. 2.6. Western blot The cell–PLGA construct was cultured for 14 days and washed once with cold PBS. The constructs were triturated using a fire-polished glass pipette with 0.25%

3713

trypsin in 0.03% EDTA and rinsed three times. Cells were collected and lysed in RIPA buffer (1  PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, plus protease inhibitor PMSF). Cells were incubated for 30 min on ice. The cell debris was pelleted by centrifugation at 4  C at 14,000 rpm for 30 min. The protein concentration in cell lysates was determined by BCA protein assay kit (Ding Guo, Co, Beijing, China). Equal amounts of total protein were loaded on a 10% polyacrylamide gel, separated by gel electrophoresis, and transferred onto a PVDF membrane (Millipore, MA, USA). The membranes were blocked with 5% non-fat milk in TBST (25 mM Tris–HCl, 0.15 M NaCl, and 0.1% Tween 20) for one hour at room temperature, and incubated with rabbit anti-bIII tubulin, and rabbit anti-phospho-p38, followed by anti-rabbit HRPconjugated IgG (1:2000, Santa Cruz, CA, USA). The membranes were then reacted with ECL western blot substrate kit (Pierce, Illinois, USA) before exposure. For the bIII tubulin detection, positive control was proteins extracted from neonatal rat brain. The signal intensity of each protein on Western blot was measured using TotalLab (Nonlinear Dynamics, Co, USA). All experiments were repeated three times and values were analyzed statistically with SPSS 11.5 software.

Fig. 1. Neurosphere and PLGA scaffold: (a) a neurosphere was visualized under light microscope. (b) Nuclei in the neurosphere were labeled by Hoechst-33324. (c) Cell bodies were stained with antibodies against nestin. (d) SEM of a transverse section of PLGA scaffold shows 16 tubes (arrow). There are numerous pores with variable diameters (asterisk) between tubes. (e) Longitudinal section of PLGA was imaged under a high magnification of SEM. Arrow points to the radial channel that extends from the scaffold tube (asterisk). (f) A neurosphere (asterisk) is attached to the wall of PLGA tube. Some cells migrate out of the neurosphere (arrow). (g) In the PLGA scaffold, a cell extends processes (arrow) from the cell body. Scale bar in (a) ¼ 20 mm in (a–c).

3714

Y. Xiong et al. / Biomaterials 30 (2009) 3711–3722

2.7. Ultrastructural observation PLGA scaffolds and cell–PLGA constructs were fixed in 2% glutaraldehyde for 90 min, osmicated with 1% Osmic acid for one hour, and dehydrated with a graded concentration of ethanol. The samples were coated with platinum. PLGA scaffold and cell morphology were examined under the scanning electron microscope (SEM) (Philips XL30 FEG) at 10 kV. For transmission EM (TEM), NT-3/TrkC scaffolds were fixed with 2.5% glutaraldehyde at 4  C for 30 min and osmicated with 1% osmic acid for one hour. Scaffolds were dehydrated through graded ethanols and embedded in Epon overnight, followed by polymerization at 60  C for 48 h. Acetone or epoxypropane was avoided since PLGA was dissolvable in these organic solvents. Ultrathin sections were obtained using ultramicrotome (Reichert E, Co, Vienna, Austria) and examined under TEM (Philips CM 10, Eindhoven, Holland).

2.8. Detection of synaptic activity NSCs in scaffolds were cultured for 14 days. Cells were loaded with 10 mM styryl dye (N-3-triethylammonmpropyl)-4-(4-(dibutylamino) styryl) (FM1-43, Invitrogen, CA, USA) under high [Kþ] (50 mM) environment and incubated (37  C and 0.5% CO2) for 10 min [23]. High [Kþ] stimulated the recycling of endocytic synaptic vesicles that contained the FM1-43. The scaffolds were rinsed three times (15–20 min for each) with culture medium in the absence of FM1-43, which would bring the endocytic/exocytotic activities down to the basal level. This also eliminated nonspecific labeling of cytoplasmic membrane, but kept synaptic vesicles still labeled by FM1-43. The cells were then unloaded by depolarization again with high [Kþ]. Release of FM1-43 labeled synaptic vesicles was imaged with the OLYMPUS FV500 fluorescent microscope. A control experiment to assess non-specific bleaching of fluorescence was performed separately and showed in the Supplementary data.

2.9. Statistical analysis For quantification of cell types in any given experiment, at least five random fields were selected and photographed (under 20 lens). The percentage of positive cells was determined relative to the total number of Hoechst 33342 (Sigma, Missouri, USA)-labeled cell nuclei. We counted about 500 cells in every group. Statistical analyses were performed using ANOVA. The significant level was set at 0.05.

3. Results 3.1. Isolation and culture of NSCs NSCs were isolated from the whole hippocampi of rat pups. They were observed under the phase contrast microscope after five-day culture. The cells were aggregated as free-floating neurospheres (Fig. 1a). To determine whether the spheres were neurospheres, their nuclei were labeled by Hoechst33324 (blue in Fig. 1b), and cells were immunostained with antibodies against nestin, a marker for neural precursors. Approximate two-thirds of cells were nestinpositive in the surface of neurospheres (green in Fig. 1c). Cells in the core of neurospheres were often not stained probably due to hypoxic or dysplasia resulted from rapid proliferation and tight contact between cells [17]. The unstained cells in the core were masked by the nestin-positive cells on the surface, but still could be visualized by adjusting the microscopic focus. These NSCs were then grafted into the PLGA scaffolds. 3.2. Geometric features of PLGA scaffolds To determine whether PLGA scaffolds are able to form an optimal structural environment for seeded NSCs, we studied the architecture of the scaffolds by SEM. The average diameter of each PLGA slice was 5 mm with a thickness of 2 mm. Longitudinal tubes in each slice were of 0.5 mm in diameter (arrow in Fig. 1d). Numerous pores with diameters varying from a few mm to 200 mm were dispersed in the PLGA scaffold (asterisk in Fig. 1d). High magnification of SEM of longitudinal PLGA section showed many channels (arrow in Fig. 1e) radially extended from the tube (asterisk in Fig. 1e) with many small pores distributed on their walls. Neurospheres cultured for 14 days were visible in the PLGA scaffolds by

Fig. 2. Fluorescence microscopic images of the PLGA/NSCs slices stained with calcein-AM and EthD-III staining. Viable cells were labeled by the green fluorescence generated from the esterase hydrolysis of a membrane-permeant dye, calcein-AM. Dead cells were marked by the red fluorescence of a membrane-impermeant DNA marker, EthD-III. Arrow points to dead cell in NSCs only (a) LacZ (b) and NT-3/TrkC (c) group. Asterisk marks live-cell. (d) The percentage of dead cells was low and not significantly different between the periphery and center of the PLGA slices in all three groups (p > 0.05; n ¼ 3). Scale bar ¼ 20 mm in (a–c).

Y. Xiong et al. / Biomaterials 30 (2009) 3711–3722

SEM (asterisk in Fig. 1f) and often attached to the wall of tube. Numerous cells migrated out of the neurospheres along the wall of tubes (arrow in Fig. 1f). Cells adhered on the wall exhibited multiple processes (arrow in Fig. 1g). Taken together, these findings suggest that PLGA scaffolds are able to form numerous tunnels with variable directions, an environment permitting grafted NSCs to adhere, extend, and migrate. 3.3. Survival and differentiation of NSCs in PLGA slices To detect whether NSCs were viable in the PLGA slices, cells grafted into the slices were cultured for 14 days, and stained with calcein-AM and EthD-III. Viable cells exhibit green fluorescence (asterisk in Fig. 2a–c) that was generated by the esterase hydrolysis of a membrane-permeant dye, calcein-AM. Dead cells were marked by red fluorescence (arrow in Fig. 2a–c) from a membrane-impermeant DNA marker, EthD-III. The percentages of dead cell were not significantly different between peripheral and central sections in all

3715

experimental groups, NSCs only (1.39  0.46% periphery; 1.43  0.51% center), LacZ (1.57  0.59%; 1.22  0.34%) and NT-3/ TrkC (1.86  0.47%; 1.58  0.67%; p > 0.05) (Fig. 2d). When comparison was made between different experimental groups, the difference was not significant (p > 0.05). These findings suggest that survival of grafted cells is not affected by their location or types of transfection. To determine whether transfected NSCs in the PLGA slices were differentiated, slices were cultured in 35 mm dishes for 14 days, and cut into 30 mm-thick transverse sections. We stained cells with different cell type markers and demonstrated NSCs differentiated into all three major CNS cell types: neurons (red in Fig. 3a–c), astrocytes (red in Fig. 3d–f), and oligodendrocytes (red in Fig. 3g–i). Most cells migrated out of the neurospheres and widely distributed in the slices, but there were still a few cells remaining in the channels. At day 14, 77.11% of 545 counted cells were Map2 positive in the NT-3/TrkC group (Fig. 3c). This percentage was the highest among all groups (Table 2), whereas only 13.71% of 590 counted

Fig. 3. NSCs were cultured in the PLGA slices for 14 days. Cells were immunostained with markers for neurons (Map2; arrow in a–c; red), astrocytes (GFAP; arrow in d–f; red) and oligodendrocytes (MOSP; arrow in g–i; red). A majority of cells expressed Map2 (b, c) in the NT-3- and NT-3/TrkC groups. Cells positive for GFAP or MOSP were much less. In contrast, astrocytes were abundant in the NSCs-(d). Nuclei were stained by Hoechst-33324 (blue). Scale bar ¼ 20 mm in (a–i).

3716

Y. Xiong et al. / Biomaterials 30 (2009) 3711–3722

Table 2 Comparison of cell-types (mean  s.d., %) among different groups. Groups

n

Map2 þ (%)

GFAP þ (%)

MOSP þ (%)

NSCsa LacZ-NSCsb TrkC-NSCsc NT-3-NSCsd NT-3/TrkCe

3 3 3 3 3

13.71  4.24 23.29  4.73 26.14  2.75 68.51  3.89 77.11  4.99

72.56  3.97 53.04  4.67 49.61  4.31 18.57  1.41 18.06  2.76

13.58  1.81 21.45  3.64 23.38  2.59 12.08  2.77 4.82  0.87

One-way ANOVA Test was used to show the statistical difference. Map2: P < 0.05 a vs b; a vs c; a vs d; a vs e; b vs d; b vs e; c vs d; c vs e; d vs e; P > 0.05 b vs c; GFAP: P < 0.05 a vs b; a vs c; a vs d; a vs e; b vs d; b vs e; c vs d; c vs e; P > 0.05 b vs c; d vs e; MOSP: P < 0.05 a vs b; a vs c; a vs e; b vs d; b vs e; c vs d; c vs e; d vs e; P > 0.05 a vs d; b vs c;

cells were Map2 positive in the NSCs group (Fig. 3a), and 23.29% of 522 counted cells in the LacZ-group (Fig. 3a; Supplementary Fig. S1), 26.14% of 546 counted cells in the TrkC group (Fig. 3b; Supplementary Fig. S1), and 68.51% of 579 counted cells in the NT-3 group (Fig. 3b). GFAP-positive astrocytes (Fig. 3d–f) and MOSPpositive oligodendrocytes (Fig. 3g–i) were also examined. Percentage of either GFAP-or MOSP-positive cells was the lowest in the NT-3/TrkC group (Table 2). This finding is consistent with the highest percentage of Map2 positive cells in the NT-3/TrkC group. The ratio of Map2-postive cells were agreeable with bIII-tubulin Western blot, in which expression of tubulin was significantly higher in the NT-3/TrkC and NT-3 group than that in other groups (p < 0.05; Fig. 4a and b). NT-3 is known to bind its receptor TrkC and activates a cascade of downstream signaling molecules, including phosphorylation of p38 [24]. This signaling pathway has been suggested to play a role in neuronal differentiation of NSCs [24]. To determine whether this

signaling pathway is activated in our grafted NSCs, cells in slices were harvested and subjected to Western blot analysis. The NT-3/ TrkC group had the highest level of phosphorylated p38 (Fig. 4c and d), compared with other groups. This increase of phosphorylated p38 was prevented by application of K252a, a specific inhibitor of neurotrophin-related tyrosine kinase. These results suggested that there is an activation of TrkC/p38 signaling in the NT-3/TrkC cells. Taken together, microenvironment in our scaffolds permits NSCs to differentiate into neurons. This differentiation is further improved when NT-3 and TrkC are simultaneously present. 3.4. Synaptogenesis of NSCs To determine whether NSCs in the PLGA scaffolds develop synapses, double-immunostaining for synapsin-I and postsynaptic density-95 (PSD95), markers for pre- and post-synapse, was carried out. Since NSCs only differentiate into neuronal and glial cell lineages in culture [25], PSD95 and synapsin-I should be specific for synapses. PSD95 was detected in the cell bodies, but synapsin-I expressed in both perikaryon and neurites (Fig. 5a–e). In the NT-3/ TrkC group, there were three patterns of PSD95 and synapsin-I expression: (a) cells positive for both PSD95 and synapsin-I (left inset in Fig. 5e and enlarged in Fig. 5f; yellow in overlay); (b) cells with positive PSD95-staining in the body but receiving synapsinpositive neurites or extending neurites with positive synapsin (right inset in Fig. 5e and enlarged in Fig. 5g); (c) cells positive for either PSD95 or synapsin. Overall, cells positive for synapsin-I appeared more than cells expressing PSD95 in the NT-3 group (Fig. 5d). In contrast, the number of synapsin-positive cells was comparable to the number of PSD95-positive cells in the NT-3/TrkC group. Moreover, the percentages of the cells with synapses were significantly higher in the NT-3/TrkC group, compared with other

Fig. 4. Cells collected from the slices were lysed, and proteins were extracted for Western blot analysis. (a) Levels of bIII tubulin expression appeared higher in the NT-3- and NT-3/ TrkC-groups than those in the NSCs, LacZ-, or TrkC group. Proteins extracted from the rat brain were used as a positive control (brain). (b) The level of proteins on the western blot was quantified by chemiluminescence and normalized by the loading control (GAPDH). The tubulin level was significantly higher in the NT-3/TrkC- and NT-3-group than those in other groups (*p < 0.05; n ¼ 3 for each group). (c) The same experiment was also performed with antibodies against phosphorylated p38. Compared with other groups, the level of phosphorylated p38 in the NT-3/TrkC group appeared higher. This increased level of phosphorylated p38 was eliminated by K252a. a-tubulin was used as loading control. (d) Quantitative analysis showed that level of phosphorylated p38 in the NT-3/TrkC was higher than that in any other groups (*p < 0.05; n ¼ 3 for each group; **NT-3/TrkC vs. NT-3/TrkC þK252a, p < 0.05).

Y. Xiong et al. / Biomaterials 30 (2009) 3711–3722

3717

Fig. 5. Culture cells in the PLGA slices were immunostained with antibodies against pre- and postsynaptic marker (PSD95 vs. synapsin-I). The nuclei were labeled by Hoechst-33324 (blue). (a) Only a few cells expressed synapsin-I and PSD95 in the NSCs group. PSD95 was localized in the cell body (right inset in a; scale bar ¼ 10 mm); whereas synapsin-I expressed in both cell body and neurite (left inset in a; scale bar ¼ 15 mm). (b) Synapsin-I or PSD95-positive cells were sparse in the LacZ- and TrkC group (c). (d) Synapsin-I or PSD95-positive cells appeared to increase in the NT-3 group, but only for PSD95-positive cells. (e) In the NT-3/TrkC group, the cells express both PSD95 and synapsin-I (left inset and enlarged in f), or PSD95 in the body but receiving synapsin-positive neurites or extending neurites with positive synapsin (right inset and enlarged in g), or positive for either PSD95 or synapsin. (h) Cells with staining of PSD95 and synapsin were manually counted. Compared with other groups, positively stained cells (synapsin-Iþ or PSD95þ) in the NT-3/TrkC group were significantly higher than any other groups (*p