2.1. The Isolation and Culture of Rat NSPCs

The methods for the culture and expansion of NSPCs have been previously described [29]. In brief, E18 cerebral anlage tissue from transgenic GFP-expressing Sprague-Dawley (SD) rats was mechanically dissociated in basal medium (DMEM/F12). Then, the dissociated cells were collected by centrifugation, and the cell suspension was cultured in DMEM containing 1% B27 (Invitrogen), basic fibroblast growth factor (bFGF, 20 ng/ml; Invitrogen), epidermal growth factor (EGF, 20 ng/ml; Invitrogen), and N2 (10 μl/ml; Invitrogen) (Figure S1A). On the day of transplantation, to identify NSPCs in vitro, the cells from passage 3 (Figure S1B) were immunostained with Nestin, the result showed strong and stable emission of the green fluorescent signal and formed neurospheres (Figure S1C), and the neurospheres expressed Nestin (Figure S1D). To observe the differentiation potential of the NSPCs, the cells at passage 3 were cultured in a medium containing 1% fetal bovine serum without EGF or bFGF. We detected the cells expressing the neuronal marker MAP-2 (Figure S1E), the astrocytic marker glial fibrillary acidic protein (GFAP) (Figure S1F), and the oligodendrocyte marker O1 (Figure S1G) seven days later. This result indicated that the transplanted NSPCs have the capacity to differentiate to neurons, astrocytes, and oligodendrocytes. Before transplantation, the NSPCs were prepared as 1.5 × 10⁵ viable cells per microliter and 2.5 × 10⁴ viable cells per microliter in 0.1 M PBS after digestion with 0.05% trypsin/DMEM and washing with basal medium (DMEM/F12).

2.2. Spinal Cord Injury and Transplantation

Healthy adult female SD rats (220-250 g, n = 85) purchased from Beijing Vital River Laboratory Animal Technology Corporation were used in this study. All experimental protocols and animal handling procedures were approved by the Beijing Neurosurgical Institute Laboratory Animal Ethics Committee in China. All surgeries were performed under anesthesia with intraperitoneal (i.p.) injections of 10% chloral hydrate (0.3 ml/100 g of body weight). After inducing deep anesthesia and incising the skin, subcutaneous tissue, and muscle, a laminectomy was performed to expose the T10 spinal cord region under an operating microscope. The dura was cut with an 11-blade scalpel, and then, a 2 mm long section of the spinal cord was completely removed at the T10 level using iridectomy scissors and microscope forceps under the dura. The operating board was spun, and the operating microscope was adjusted to clear residual fibers. After complete SCI, 10 μl of cell suspension (1.5 × 10⁵ viable cells per microliter in the high-dose group, n = 30; 2.5 × 10⁴ viable cells per microliter in the low-dose group, n = 30) was immediately microinjected into the lesion cavity using a Hamilton syringe. Subjects injected with 10 μl of PBS were used as the operation control (SCI group, n = 25). After the surgical incisions were sutured, the subjects were placed on a heating pad until fully awake and then received 0.9% saline i.p. (3 ml per subject). Postoperative care included the administration of penicillin once a day for three days to prevent infection and manual pressing of the bladder twice a day until recovery of automatic micturition. The subjects that underwent NSPC transplantation were not immunosuppressed due to the homogeneity with the NSPCs, and the NSPC-transplanted rats did not incur notable immune responses compared to those in the control group throughout the entire experiment.

2.3. Behavioral Analysis

Locomotor function was evaluated using the Basso-Beattie-Bresnahan (BBB) locomotor rating scale [30]. The open-field environment was a molded plastic wading pool with a nonslippery surface (100 cm in diameter and 21 cm in height) [30]. The BBB scores were assessed weekly after the surgery. Seven subjects were randomly selected for testing, which lasted for approximately 4 min and was recorded using a digital video camera (Sony). The hindlimb movement scores were assessed by two independent observers blinded to the group identities.

2.4. Electrophysiology

Motor potentials were evoked and recorded according to a previously described protocol [31, 32]. The main difference in our procedure was the i.p. injection of 10% chloral hydrate (0.3 ml/100 g of body weight) [33]. To record the motor-evoked potentials (MEPs), one needle electrode was placed in the tibialis anterior muscle (the cathode) and another was placed subcutaneously at the level of the foot pad (the anode). To induce the MEPs (after central stimulation), one needle electrode was placed subcutaneously at the level of the lower jaw (the anode) and another was placed at the cranial level (the cathode). For the ground, an electrode was placed subcutaneously in the lumbar region. The electrophysiological signals were recorded by an electromyography system (Neuro-MEP-Micro, Russia) with bandpass filtering from 2 Hz to 10 kHz. The pulse duration used throughout the experiments was 0.1 ms. To induce the MEPs, stimulation with an intensity of 25 mA was applied to the needle electrode at the cranial level. MEP measurements were performed in the three groups at 8 weeks after the operation (n = 7 animals per group).

2.5. Neuronal Tracing

To label the corticospinal tract (CST) fibers, 4 μl of biotinylated dextran amine (BDA, 10%, 10,000 MW; Invitrogen) was bilaterally injected into the motor cortices at eight sites (0.5 μl per site) at 8 weeks postsurgery (n = 3 per group) [34]. Two weeks later, the subjects were perfused with 4% paraformaldehyde (PFA). To trace the functional transsynaptic neural pathways by labeling the descending CST fibers, 1% biotinylated cholera toxin B subunit (CTB, Invitrogen, C-34779, Molecular Probes), which can transfer across synapses to second- and third-order neurons [35, 36], was injected into bilateral L2/L3 spinal white matter using a Hamilton syringe (1 μl per site) in three subjects from each group. Seven days postinjection, the subjects were perfused with 4% PFA, and the spinal cord was subjected to CTB staining.

2.6. Histology and Immunohistochemistry

After being deeply anesthetized, all animals were perfused with 4% PFA, and spinal cord segments containing a lesion or graft site were obtained. The samples were placed in 4% PFA overnight at 4°C and then transferred to 30% sucrose solution (with 0.1 M PBS) and incubated overnight. The spinal cord samples were embedded in OCT compound and frozen in n-hexane (-80°C) after being processed. Horizontal cryosections, 10 μm thick, were cut using a cryostat microtome (Leica). The sections were washed three times with 0.01 M PBS and then incubated with the primary antibody chicken anti-GFP (1 : 200, Thermo Fisher Scientific, to label GFP-expressing NSPCs), rabbit anti-MAP-2 (1 : 50, CST, to label mature neurons), rabbit anti-NeuN (1 : 200, Abcam, to label mature neurons), mouse anti-βIII tubulin (Tuj1, 1 : 200, CST, to label immature and mature neurons), mouse anti-O1 (1 : 100, Millipore, to label oligodendrocytes), rabbit antisynaptophysin (SYN, 1 : 200, CST, to label presynaptic terminals), rabbit antineurofilament (NF, 1 : 1000, Abcam, to label axons), mouse anti-GFAP (1 : 300, CST, to label astrocytes), mouse anti-MBP (1 : 200, Abcam, to label oligodendrocytes and myelin), and mouse anticholine acetyltransferase (CHAT, 1 : 200, Abcam, to label spinal cord motor neurons and motor axons) overnight at 4°C. Then, after being washed three times with 0.01 M PBS, the samples were incubated for 2 hours at 4°C with Alexa Fluor 488-, 594-, and 647-conjugated goat antibodies (1 : 250, Thermo Fisher Scientific) or Alexa Fluor 594-conjugated streptavidin (1 : 250, Invitrogen) to label BDA-traced CST axons or CTB-traced neurons. After the sections were washed again with 0.01 M PBS, mounting medium with DAPI (nuclear staining; ZSGB-BIO, China) was added to the sections, which were then cover slipped. All samples were permeabilized with 0.03% Triton X-100 and blocked for 1 hour at room temperature with 10% normal horse serum. Finally, sections were examined and imaged by confocal laser scanning microscopy (Zeiss, Airyscan LSM880, Germany).

2.7. Quantitative Immunohistochemical Analyses

To quantitatively analyze the survival of GFP-positive grafted cells in the lesion area, three subjects each were sacrificed in the high- and low-dose groups at 1 week postgrafting. Then, one for every ten in the whole series of sections from each rat was selected for immunostaining with GFP and DAPI. One random image within the rostral, central, and caudal areas of the SCI site was, respectively, imaged in each section at 400x magnification using an LSM880 system. Finally, the numbers of GFP-positive cells in each experiment rat were quantified using the Image-Pro Plus software (Media Cybernetics, Silver Spring, MD).

To quantify axonal regeneration in vivo, four rats were sacrificed in every group, and one for every ten in the whole series of sections from each rat was selected. After all the selected sections were immunolabeled for NFs, one random image within the rostral, central, and caudal areas of the SCI graft site was, respectively, imaged in each section at 200x magnification using an LSM880 system. Then, the number of NF-positive axons in every image was quantified using the Image-Pro Plus software (Media Cybernetics, Silver Spring, MD). In brief, an image was chosen, and NF-positive nerve fibers were converted into an area of interest (AOI); then, the pixel area of the AOI was automatically calculated by the software [37, 38]. Finally, we statistically analyzed the area of NF-positive fibers in the rostral, central, and caudal regions of the lesion site in each rat from the three groups (n = 4 per group).

To quantitatively analyze the area of MAP-2-positive neurons in the lesion core, one random area in the central region of the SCI graft site was, respectively, imaged in each section at 200x magnification using an LSM880 system (ten sections per rat, n = 4 per group). The Image-Pro Plus software was used as described above to evaluate the proportion of MAP-2-positive neurons in the lesion core.

To quantify the percentage of cells that were positive for GFAP, MAP-2, and MBP in the regenerated tissue at the SCI (2 mm) site at 8 weeks postsurgery, three sequential sections for every ten in the whole series of sections from each animal were selected for GFAP, MAP-2, and MBP immunostaining (n = 4 rats per group). One random area within the rostral, central, and caudal regions of the SCI graft site was, respectively, imaged in each section at 200x magnification using an LSM880 system. We assessed the mean area of cells positive for GFAP, MAP-2, and MBP in each section using the Image-Pro Plus software, as described above. Then, we quantitatively analyzed the area of cells positive for each marker in the total area of regenerated tissue, i.e., the sum of the areas stained positive for GFAP, MAP-2, and MBP, in one rat.

2.8. RNA Isolation and RT-PCR

To compare the mRNA expression levels of the three groups at 1, 3, and 7 days postsurgery (n = 3 per group), total RNA was extracted from 3 mm long spinal cord samples, which included the lesion area, using the TRIzol Reagent (Life Technologies). To synthesize complementary DNA (cDNA), reverse transcription was performed using a High-Capacity cDNA Reverse Transcription kit (Life Technologies) according to the manufacturer’s instructions. The gene expression levels are listed in supplemental Table 1. Gene expression was quantified using an Applied Biosystems 7500 Fast System (Life Technologies). The data are presented as fold-change values.

2.9. Transmission Electron Microscopy (TEM)

To detect remyelination and axonal regeneration, subjects (n = 3 per group) randomly selected from each experiment group were perfused with 4% PFA at 8 weeks postsurgery. Then, the samples containing a lesion or graft site were obtained and postfixed with 1% osmium tetroxide, dehydrated, and embedded in Durcupan resin. To obtain axial ultrathin sections in the lesion area, the samples were cut from the lesion core using an ultramicrotome (Leica EM UC7). Then, axial ultrathin sections were examined under a transmission electron microscope (HITACHI, H-7650, Japan).

2.10. Statistical Analysis

All statistical analyses were performed using the GraphPad Prism 6.0. All data are presented as the mean ± SEM. Correlations between the transplantation dose and the number of surviving NSPCs were assessed using the Pearson correlation coefficients and linear regression analyses. One-way ANOVA with Fisher’s least significant difference (LSD) test was used to assess the comparisons among the three groups. If equal variances were found, the LSD test was applied; otherwise, the Kruskal-Wallis test and Dunnett’s T3 test were used. Two-group comparisons were made by using Student’s t-tests. The significance level was set at p < 0.05.

3.1. High-Dose NSPC Transplantation Improved the Injury Microenvironment and Significantly Increased NSPC Survival and Integration

In this study, we used qPCR to examine whether the number of transplanted cells affects the harmful microenvironment; different numbers of NSPCs were transplanted into the SCI site immediately after complete injury at the T10 level. High-dose NSPC transplantation notably increased the expression of anti-inflammatory cytokines, such as interleukin- (IL-) 10 and transforming growth factor- (TGF-) β, in addition to attenuating the expression level of tumor necrosis factor- (TNF-) α, at 3 days postsurgery (Figures 1(a) and 1(b)). Furthermore, in the high-dose group, we observed significant increases in the expression levels of several neurotrophic factors, including glial cell-derived neurotrophic factor (GDNF), brain-derived neurotrophic factor (BDNF), insulin-like growth factor- (IGF-) 1, and vascular endothelial growth factor- (VEGF-) A but not neurotrophic factor- (NT-) 3 or nerve growth factor (NGF) (Figure 1(c)). However, there were no significant differences in the cytokine mRNA expression levels between the low-dose and SCI groups.

Next, we conducted immunofluorescence analysis to examine whether high-dose NSPC transplantation enhanced the engraftment and survival of NSPCs in severe SCI to complement the apoptosis and loss of grafted cells. Immunostaining showed greater GFP-positive cell survival and integration in the lesion area in the high-dose group than in the low-dose group and more grafted cells distributed in the rostral and caudal areas of the lesion and migrated into the host spinal cord in the high-dose group (Figures 1(d) and 1(e)). The Pearson correlation analysis revealed a remarkable positive relationship between the transplantation dose and the number of surviving NSPCs after severe SCI at 7 days postgrafting (Figure 1(f), r = 0.9899).

3.2. High-Dose NSPC Transplantation Significantly Promoted Motor Functional Recovery and Electrophysiological Recovery in Rats with Complete SCI

To examine the influence of the NSPC transplantation dose on functional recovery, we quantitatively evaluated functional improvements every week after surgery using the BBB open-field locomotor scale in subjects with complete SCI at the T10 level treated with either low- or high-dose NSPC transplantation. Two weeks postsurgery, we found greater hindlimb locomotion recovery in the high-dose groups than in the SCI group; there were no differences between the high- and low-dose groups (Figure 2(a)). At five weeks postsurgery, a significant improvement in the BBB scores was observed in the high-dose group compared with the low-dose group (Figure 2(a)), and a significant improvement in the BBB scores was observed in the low-dose group compared to the SCI group (Figure 2(a)). The BBB score in the high-dose group reached a functional plateau at 8 weeks postsurgery at approximately 6.4, significantly higher than the score 3.1 of 1.9 in the low-dose and SCI groups, respectively (n = 7 rats per group, p < 0.01).

To further determine the functional status of the severed spinal cord, we analyzed the MEPs at 8 weeks postsurgery. In the high- and low-dose groups, after a 25 mA stimulation was applied at the cranial level, we detected an electrical signal at the tibialis anterior muscle. However, no signal was detected in the SCI group (Figure 2(b)). To compare the MEP recovery of the high- and low-dose groups, we quantitatively analyzed the amplitude and latency of the MEPs. We found that the signals in the high-dose group exhibited greater amplitudes and shorter latencies than those in the low-dose group (Figures 2(c) and 2(d)). Taken together, these data further confirm the superior functional recovery achieved in the high-dose group.

3.3. Abundant NF-Positive Neuronal Fibers Regenerated at the Injury Site in the High-Dose Group

Many previous studies have demonstrated that only a few neuronal fibers regenerate at the injury site following complete SCI, especially in the central region of the injury site [37]. To confirm whether high-dose NSPC transplantation increased the distribution of neuronal fibers at the injury site, we used longitudinal tissue sections containing the injury site from the three groups to stain for NFs. The results showed greater NF-positive fiber penetration and occupation in the entire lesion area in the high-dose group than in the low-dose and SCI groups, notably in the lesion core (Figures 3(a)–3(c)). We further quantitatively analyzed the number of NF-positive fibers in the rostral, central, and caudal regions of the lesion area. The data showed a significantly enhanced distribution of NF-positive fibers in the rostral, central, and caudal regions of the lesion in the high-dose group compared with the other two groups (Figure 3(e)). Interestingly, there were more NF-positive fibers in the central region of the lesion in the low-dose group than in the SCI group, although NF-positive fibers were rarely observed in the central region (Figure 3(e)). Furthermore, by triple-immunostaining for NF, GFP, and DAPI, we demonstrated that most of the NF-positive fibers in the lesion area were derived from grafted GFP-expressing NSPCs (Figure 3(d)).

3.4. High-Dose NSPC Transplantation Significantly Increased the Expression of Neural Marker MAP-2 at the Spinal Cord Transection Site

To test whether enhancing the NSPC transplantation dose increased the distribution of neurons at the injury site, we applied MAP-2 and DAPI fluorescence immunolabeling to a horizontal spinal cord section from each of the three groups. Representative images showed a significant difference in the distribution of MAP-2-positive neurons between the high- and low-dose groups and between the NSPC transplantation groups and the SCI group over the entire lesion area, notably in the center (Figures 4(a)–4(c)). Quantitative analyses showed an increased MAP-2-positive area in the central region of the lesion in the NSPC transplantation groups compared with the SCI group. Importantly, the MAP-2-positive area in the center of the injury site was significantly greater in the high-dose group than in the low-dose and SCI groups (Figure 4(d)). In addition, we chose four subjects from each group for a quantitative analysis of the proportion of the MAP-2-positive area of the total lesion area, which was the sum of the injury site areas positive for MAP-2, GFAP, and MBP, in each subject at 8 weeks postsurgery. The data revealed a significantly improved percentage of neurons in the lesion tissue in the high-dose group compared with the low-dose and SCI groups (Figure 4(e)). However, we also found that the percentage of neurons in the lesion tissue was not enhanced in the low-dose group compared with the SCI group, although the distribution of MAP-2-positive neurons in the central region of the lesion was increased (Figure 4(d)). By immunolabeling with GFP, MAP-2, and DAPI, we demonstrated the presence of many MAP-2-positive neurons derived from GFP-expressing grafted NSPCs (Figure 4(f)).

3.5. Grafted NSPC-Derived Oligodendrocytes Contributed to the Remyelination of Growing Axons in the High-Dose Group

In our study, more myelination was observed by light microscopy at the spinal cord transection site in the high-dose group than in the low-dose group (Figures 5(a)–5(c)). Using TEM analysis, we detected many myelinated fibers regenerated in the high-dose group and few myelinated fibers regenerated in the low-dose group; almost no regenerated myelinated axons were observed in the lesion area in the SCI group (Figures 5(d)–5(f)). These results suggest that some engrafted cells differentiated into mature oligodendrocytes, forming thick sheaths to wrap growing axons at the SCI site in the NSPC transplantation groups, as recent reports have indicated that neural progenitor cells grafted into lesions can contribute to remyelination [39]. Therefore, we utilized GFP, MBP, and DAPI immunolabeling at the SCI graft site, and the immunohistochemical analyses revealed many GFP/MBP double-positive oligodendrocytes in the lesion area (Figure 5(g)). Based on these results, we added NF staining and observed some GFP/MBP double-positive myelination of the GFP/NF double-positive fibers (Figure 5(h)). Furthermore, for each group, we quantified the proportion of MBP-positive cells in the total area positive for three cell type-specific markers in the lesion. Of the regenerated tissue, approximately 21.3% of the area was positive for MBP in the high-dose group, and this value was significantly higher than that for the low-dose (6.2%) and SCI groups (3.2%). Surprisingly, statistical analysis showed a significant difference between the low-dose (6.2%) and SCI groups (3.2%), meaning that low-dose NSPC transplantation also appeared to induce oligodendrocyte differentiation and remyelination (Figure 5(i)).

3.6. High-Dose NSPC Transplantation Significantly Alleviated Glial Scar Formation at the Injury Site

We also observed glial scarring and the formation of a cavity in the lesion area. However, a remarkably different phenomenon was observed in the high-dose group at 8 weeks postgrafting. At 8 weeks postgrafting, immunohistochemical analysis of GFAP revealed that the GFAP-positive astrocyte area was decreased over the whole lesion area in the high-dose group compared with the low-dose group and especially with the SCI group (Figures 6(a)–6(c)). To investigate whether the glial scar reduction was associated with the high-dose NSPC transplantation, four subjects from each group were sacrificed for quantification of the GFAP-positive astrocyte area in the lesion. As expected, the results showed significantly less GFAP-positive glial scarring in the lesion in the high-dose group. In contrast, massive glial scarring was observed in the low-dose group, with no significant difference from that observed in the SCI group (Figure 6(d)). In addition, we quantified and analyzed the proportion of GFAP-positive astrocytes in the regenerated tissue structure at the injury site. The proportion (only 58.8%) was significantly lower in the high-dose group than in the low-dose (84.7%) and SCI groups (90.9%) (Figure 6(e)).

3.7. Grafted NSPC-Derived Neurons Formed a Nascent Functional Neural Network in a Relayed Manner

Indeed, we observed more functional recovery and neuronal tissue regeneration in the high-dose group than in the other two groups. To further explore the underlying mechanism relating the functional recovery and the abundant neuronal tissue regeneration, we investigated whether functional structure formation occurred at the injury site. At 8 weeks after high-dose NSPC grafting, the immunolabeling analyses revealed that many grafted NSPC-derived neurons expressed a characteristic marker of spinal motor neurons, CHAT (Figure 7(b)). In addition, we also observed a graft-derived, GFP-labeled, and Tuj1-positive neuron expressing a synaptic marker of synaptic junctions at the lesion area, suggesting synaptic connectivity among the grafted neurons (Figure 7(a)). To confirm whether CST fibers were regenerated across the injury site in the high-dose group, 10% BDA was injected into eight sites per hemisphere to trace corticospinal axons at 8 weeks postsurgery. Two weeks later, the immunohistochemical analyses revealed that no BDA-positive CST fibers had regenerated across the injury site, and only some BDA-labeled fibers at the rostral border of host-graft interface were detected. Interestingly, we observed the formation of synaptic-like structures between the BDA-labeled CST fibers and GFP-expressing grafted cells at the rostral host-injury border based on immunostaining for GFP/BDA/SYN/DAPI (Figure 7(d)). Altogether, these results suggested that grafted GFP-labeled cells formed synaptic connectivity with grafted cells and with the host spinal cord.

As nascent synaptic networks were developed at the injury site in the high-dose group, we next investigated whether the nascent synaptic networks could deliver signals across the transection site. Biotinylated CTB, a bacterial toxin for neuronal tracing, can transfer across synapses to second- and third-order neurons, allowing functional neural networks to be traced in the spinal cord. At 8 weeks postgrafting, 1% CTB was injected into the spinal cord of subjects at the L2/L3 level. One week later, the subjects were sacrificed. By multiple immunostaining, we found that CTB was clearly present in caudal CHAT-positive spinal motor neurons and passed through neuronal fibers and arrived at spinal motor neurons at the T8/T9 level. Notably, CTB was also detected in grafted GFP-expressing NSPC-derived CHAT-positive motor neurons at the lesion core in the high-dose group (Figures 7(g)–7(i)). In the SCI group, CTB was almost entirely absent in the lesion area and in the rostral spinal cord over the injury site.