Neural Stem Cells Overexpressing Arginine Decarboxylase Improve Functional Recovery from Spinal Cord Injury in a Mouse Model

Current therapeutic strategies for spinal cord injury (SCI) cannot fully facilitate neural regeneration or improve function. Arginine decarboxylase (ADC) synthesizes agmatine, an endogenous primary amine with neuroprotective effects. Transfection of human ADC (hADC) gene exerts protective effects after injury in murine brain-derived neural precursor cells (mNPCs). Following from these findings, we investigated the effects of hADC-mNPC transplantation in SCI model mice. Mice with experimentally damaged spinal cords were divided into three groups, separately transplanted with fluorescently labeled (1) control mNPCs, (2) retroviral vector (pLXSN)-infected mNPCs (pLXSN-mNPCs), and (3) hADC-mNPCs. Behavioral comparisons between groups were conducted weekly up to 6 weeks after SCI, and urine volume was measured up to 2 weeks after SCI. A subset of animals was euthanized each week after cell transplantation for molecular and histological analyses. The transplantation groups experienced significantly improved behavioral function, with the best recovery occurring in hADC-mNPC mice. Transplanting hADC-mNPCs improved neurological outcomes, induced oligodendrocyte differentiation and remyelination, increased neural lineage differentiation, and decreased glial scar formation. Moreover, locomotor and bladder function were both rehabilitated. These beneficial effects are likely related to differential BMP-2/4/7 expression in neuronal cells, providing an empirical basis for gene therapy as a curative SCI treatment option.


Introduction
Spinal cord injury (SCI) currently has no curative therapy. Patients with SCI immediately lose all subdamaged motor, sensory, and autonomic nervous system functions. Secondary processes occurring at the injury site can worsen SCI [1,2]. Furthermore, bladder sphincter relaxation is absent, leading to urinary retention [3].
Characterized by an initial mechanical trauma to the spinal cord, SCI results in breakdown of the blood-brain barrier, activation of glial cells, and necrosis [4]. A secondary signaling cascade causes a cyclic increase in inflammatory cytokines that leads to apoptosis and progressive oligodendrocyte loss, eventually resulting in demyelination and axonal degeneration [5]. Moreover, the adult central nervous system (CNS) has poor trophic support and a growth inhibitory environment that is hostile to endogenous spinal cord regeneration [6,7].
Bone morphogenetic proteins (BMPs) refer to approximately 15 growth-regulating polyfunctional cytokines from the transforming growth factor beta (TGF-β) superfamily; they are widely expressed in both intact and injured spinal cords [46][47][48]. Previous studies using SCI animal models have suggested that BMP-2/4/7 have similar expression patterns in neurons and neuroglial cells. Moreover, expression patterns are closely related to improving motor function [46,49]. Normally expressed in the intact spinal cord, various BMP ligands and receptors are rapidly up-or downregulated after injury. The functions of several BMPs are well studied. For instance, BMP-4 promotes astrocytic differentiation while inhibiting neuronal and oligodendrocyte differentiation [50,51]. BMP-7 inhibits oligodendrocyte cell death and increases neuronal survival in SCI [52][53][54].
In this context, we hypothesized that transplanting recombinant human ADC-murinecortex-derived NPCs (hADC-mNPCs) is a more effective SCI treatment than current methods. We therefore sought to explore NPC-based combination approaches that emphasize ADC gene therapy for reconstructing damaged spinal neural circuits and improving SCI recovery. Additionally, because agmatine and ADC are strongly linked, we aimed to confirm whether ADC is also involved in controlling BMP expression after SCI. Our study should greatly benefit the development of curative therapy for these debilitating injuries.

Treatment with hADC-mNPC Led to Enhanced Cell Survival around the SCI Lesion Site
The results of PKH-26 staining indicated that mNPCs, pLXSN-mNPCs, and hADC-mNPCs were clustered together and migrated from the injection site (thoracic vertebra [T]8 and T10) to the lesion site (T9) at around 1 week post-transplantation ( Figure 1A). The PKH-26 signal was significantly higher at the epicenter in hADC-mNPC-transplanted mice than in mNPC-transplanted mice or pLXSN-mNPC-transplanted mice ( Figure 1B). Additionally, the hADC-mNPC group had significantly higher fluorescence intensity than the mNPC and pLXSN-mNPC groups ( Figure 1B). Agmatine plays an important role in the production and inhibition of neuronal and glial cells after SCI through its regulation of bone morphogenetic protein (BMP)-2/4/7 expression. Specifically, agmatine administration post-SCI increases BMP-2/7 expression in neurons and oligodendrocytes while decreasing BMP-4 expression in astrocytes [43,46]. Bone morphogenetic proteins (BMPs) refer to approximately 15 growth-regulating polyfunctional cytokines from the transforming growth factor beta (TGF-β) superfamily; they are widely expressed in both intact and injured spinal cords [46][47][48]. Previous studies using SCI animal models have suggested that BMP-2/4/7 have similar expression patterns in neurons and neuroglial cells. Moreover, expression patterns are closely related to improving motor function [46,49]. Normally expressed in the intact spinal cord, various BMP ligands and receptors are rapidly up-or downregulated after injury. The functions of several BMPs are well studied. For instance, BMP-4 promotes astrocytic differentiation while inhibiting neuronal and oligodendrocyte differentiation [50,51]. BMP-7 inhibits oligodendrocyte cell death and increases neuronal survival in SCI [52][53][54].
In this context, we hypothesized that transplanting recombinant human ADC-murine-cortex-derived NPCs (hADC-mNPCs) is a more effective SCI treatment than current methods. We therefore sought to explore NPC-based combination approaches that emphasize ADC gene therapy for reconstructing damaged spinal neural circuits and improving SCI recovery. Additionally, because agmatine and ADC are strongly linked, we aimed to confirm whether ADC is also involved in controlling BMP expression after SCI. Our study should greatly benefit the development of curative therapy for these debilitating injuries.

Transplantation of hADC-mNPCs Attenuated Astrocyte Clustering at The Lesion Site
We used GFAP immunoreactivity to quantify astrocytes in mNPC, pLXSN-mNP and hADC-mNPC mouse tissue sections. By 2 weeks post-transplantation, hADC-mNP mice had a significantly smaller GFAP-positive area at the lesion site than mNPC an pLXSN-mNPC mice (Figure 2A). The hADC-mNPC group also had a significantly small glial scar area than the mNPC and pLXSN-mNPC groups ( Figure 2B). Figure 2. hADC overexpression in mNPCs reduces glial scar volume 2 weeks after transplantati into compression-lesioned spinal cords of adult mice. (A) Immunohistochemistry of longitudin spinal cord sections stained with antibodies against GFAP at T8, T9, and T10 at 2 weeks after tran plantation of hADC overexpressing mNPCs (hADC-mNPCs), empty retroviral overexpressi mNPCs (pLXSN-mNPCs), or mNPCs alone. The red box was the site of the glial scar lesion, and t middle of the glial scar is indicated by a red asterisk. Scale bar = 5 mm. (B) Lesion areas are signi cantly larger in mice with mNPC transplantation (45% of the area shown in part A) than transpla tations of pLXSN-mNPCs (40% of the area shown in A) and hADC-mNPCs (19% of the area show in A). Data are presented as means ± SEM. (n = 5 per sample, *** p < 0.001).

Transplantation of hADC-mNPCs Promoted Endogenous mNPC Differentiation into Oligodendrocytes following SCI
We examined neuronal (MAP-2), oligodendrocyte (Olig-2), and astrocyte (GFA marker protein expression at 1, 2, and 5 weeks after cell transplantation. Immunoblottin results found that MAP-2 protein expression was generally higher in the hADC-mNP group than in the mNPC or pLXSN-mNPC groups ( Figure 3A), and significantly so by weeks post-transplantation ( Figure 3B). At 2 weeks, GFAP expression was significant higher in the mNPC and pLXSN-mNPC groups than in the hADC-mNPC group; wh hADC-mNPC mice still expressed less GFAP at 1 and 5 weeks, this difference was n significant ( Figure 3C). Moreover, at 2 weeks post-transplantation, GFAP expression d creased in hADC-mNPC mice compared with that in mNPC and pLXSN-mNPC mi  spinal cord sections stained with antibodies against GFAP at T8, T9, and T10 at 2 weeks after transplantation of hADC overexpressing mNPCs (hADC-mNPCs), empty retroviral overexpression mNPCs (pLXSN-mNPCs), or mNPCs alone. The red box was the site of the glial scar lesion, and the middle of the glial scar is indicated by a red asterisk. Scale bar = 5 mm. (B) Lesion areas are significantly larger in mice with mNPC transplantation (45% of the area shown in part A) than transplantations of pLXSN-mNPCs (40% of the area shown in A) and hADC-mNPCs (19% of the area shown in A). Data are presented as means ± SEM. (n = 5 per sample, *** p < 0.001).

Transplantation of hADC-mNPCs Promoted Endogenous mNPC Differentiation into Oligodendrocytes Following SCI
We examined neuronal (MAP-2), oligodendrocyte (Olig-2), and astrocyte (GFAP) marker protein expression at 1, 2, and 5 weeks after cell transplantation. Immunoblotting results found that MAP-2 protein expression was generally higher in the hADC-mNPC group than in the mNPC or pLXSN-mNPC groups ( Figure 3A), and significantly so by 5 weeks post-transplantation ( Figure 3B). At 2 weeks, GFAP expression was significantly higher in the mNPC and pLXSN-mNPC groups than in the hADC-mNPC group; while hADC-mNPC mice still expressed less GFAP at 1 and 5 weeks, this difference was not significant ( Figure 3C). Moreover, at 2 weeks post-transplantation, GFAP expression decreased in hADC-mNPC mice compared with that in mNPC and pLXSN-mNPC mice ( Figure 3C). Finally, Olig-2 expression increased over time in all groups. However, it was notably higher in the hADC-mNPC group than in the mNPC group at 1 week ( Figure 3D). ( Figure 3C). Finally, Olig-2 expression increased over time in all groups. However, it was notably higher in the hADC-mNPC group than in the mNPC group at 1 week ( Figure 3D).

Transplantation of hADC-mNPCs Preserved and Enhanced Remyelination in the In Spinal Cord
In all experimental groups, SCI caused severe demyelination at the lesion site from LFB staining indicated that hADC-mNPCs preserved myelination at 6 we SCI, with that group possessing more myelin sheaths than the mNPC and pLXSN groups ( Figure 5A-C). Ultrastructural observations using TEM (×10,000) confirm findings. Additionally, TEM-based morphological examination showed that th mNPC group had significantly thicker myelin sheaths than the other two group exhibited axonal profiles with poor myelination ( Figure 5D-F). Thus, oligoden differentiated from hADC-mNPCs have the capacity to form mature myelin she remyelinate axons.

Transplantation of hADC-mNPCs Preserved and Enhanced Remyelination in the Injured Spinal Cord
In all experimental groups, SCI caused severe demyelination at the lesion site. Results from LFB staining indicated that hADC-mNPCs preserved myelination at 6 weeks after SCI, with that group possessing more myelin sheaths than the mNPC and pLXSN-mNPC groups ( Figure 5A-C). Ultrastructural observations using TEM (×10,000) confirmed these findings. Additionally, TEM-based morphological examination showed that the hADC-mNPC group had significantly thicker myelin sheaths than the other two groups, which exhibited axonal profiles with poor myelination ( Figure 5D-F). Thus, oligodendrocytes differentiated from hADC-mNPCs have the capacity to form mature myelin sheaths and remyelinate axons.

Transplantation of hADC-mNPCs Regulated BMP Expression in SCI Mice
We measured protein concentrations in tissue samples at 1, 2, and 5 weeks after tra plantation to determine how BMP signaling changed ( Figure 6A). Western blotting resu showed that BMP-2 protein expression in the hADC-mNPC group increased significan compared with levels in the mNPC group at 2 weeks post-transplantation ( Figure 6 Additionally, BMP-7 expression was higher in the hADC-mNPC group than in pLXSN-mNPC and mNPC groups at 1, 2, and 5 weeks post-transplantation ( Figure 6 Conversely, the hADC-mNPC group had noticeably lower BMP-4 expression than bo groups at all three time points ( Figure 6C).

Transplantation of hADC-mNPCs Regulated BMP Expression in SCI Mice
We measured protein concentrations in tissue samples at 1, 2, and 5 weeks after transplantation to determine how BMP signaling changed ( Figure 6A). Western blotting results showed that BMP-2 protein expression in the hADC-mNPC group increased significantly compared with levels in the mNPC group at 2 weeks post-transplantation ( Figure 6B). Additionally, BMP-7 expression was higher in the hADC-mNPC group than in the pLXSN-mNPC and mNPC groups at 1, 2, and 5 weeks post-transplantation ( Figure 6D). Conversely, the hADC-mNPC group had noticeably lower BMP-4 expression than both groups at all three time points ( Figure 6C).

Transplantation of hADC-mNPCs Improved Locomotor and Bladder Functional Recovery Following SCI
At 1 day post-SCI, all animals exhibited complete hindlimb paralysis (BMS score = 0). At 1 week post-SCI, animals were either capable of slight ankle movements or none at all. At 1 week after transplantation, hADC-mNPC mice were able to step on the floor with the soles of their feet, whereas mNPC and pLXSN-mNPC mice rarely showed such behavior. By 3 weeks, the hADC-mNPC group showed occasional plantar stepping, while the mNPC and pLXSN-mNPC groups sometimes exhibited plantar placement without weight support, although walking remained impossible. The hADC-mNPC group also significantly differed from the control group in terms of extensive angle movement.

Transplantation of hADC-mNPCs Improved Locomotor and Bladder Functional Recovery following SCI
At 1 day post-SCI, all animals exhibited complete hindlimb paralysis (BMS score = 0). At 1 week post-SCI, animals were either capable of slight ankle movements or none at all.
At 1 week after transplantation, hADC-mNPC mice were able to step on the floor with the soles of their feet, whereas mNPC and pLXSN-mNPC mice rarely showed such behavior. By 3 weeks, the hADC-mNPC group showed occasional plantar stepping, while the mNPC and pLXSN-mNPC groups sometimes exhibited plantar placement without weight support, although walking remained impossible. The hADC-mNPC group also significantly differed from the control group in terms of extensive angle movement.
At 4 weeks post-SCI, the hADC-mNPC group could not coordinate body movement but could occasionally rotate the top of the paw to take steps, whereas such movements were nearly absent in mNPC and pLXSN-mNPC mice. The hADC-mNPC group differed significantly from control and mNPC groups in this measure.
At 5 weeks post-SCI, the hADC-mNPC group stepped parallel to the floor on initial contact, occasionally took steps coordinated with body movement, and rotated the back of the feet; in contrast, body movement was minimal in mNPC and pLXSN-mNPC mice, and walking continued to be absent. The differences between hADC-mNPC, pLXSN-mNPC, mNPC, and control groups were significant.
At 6 weeks post-SCI, the hADC-mNPC group had recovered almost all control of body motion, stepping consistently by rotating the back of the foot, stepping parallel to the floor on initial contact, and exerting force at the tail end. The mNPC and pLXSN- At 4 weeks post-SCI, the hADC-mNPC group could not coordinate body movement but could occasionally rotate the top of the paw to take steps, whereas such movements were nearly absent in mNPC and pLXSN-mNPC mice. The hADC-mNPC group differed significantly from control and mNPC groups in this measure.
At 5 weeks post-SCI, the hADC-mNPC group stepped parallel to the floor on initial contact, occasionally took steps coordinated with body movement, and rotated the back of the feet; in contrast, body movement was minimal in mNPC and pLXSN-mNPC mice, and walking continued to be absent. The differences between hADC-mNPC, pLXSN-mNPC, mNPC, and control groups were significant.
At 6 weeks post-SCI, the hADC-mNPC group had recovered almost all control of body motion, stepping consistently by rotating the back of the foot, stepping parallel to the floor on initial contact, and exerting force at the tail end. The mNPC and pLXSN-mNPC groups occasionally performed plantar steps and moved their paws in coordination with the body. Overall, BMS scores and behavior differed significantly between the hADC-mNPC, pLXSN-mNPC, mNPC, and control groups ( Figure 7A).
The result of daily manual urine collection confirmed that all groups expelled a similar amount of urine from day 1 post-SCI to day 7. By 2 weeks, manually collected urine content was twice as low in the hADC-mNPC group as in the control, mNPC, and pLXSN-mNPC groups. These data demonstrate that bladder function returned more quickly to the hADC-mNPC group than to the other groups ( Figure 7B). Urinary capacity returned to all groups after 2 weeks, and assisted urination was stopped.
The result of daily manual urine collection confirmed that all groups expelled a similar amount of urine from day 1 post-SCI to day 7. By 2 weeks, manually collected urine content was twice as low in the hADC-mNPC group as in the control, mNPC, and pLXSN-mNPC groups. These data demonstrate that bladder function returned more quickly to the hADC-mNPC group than to the other groups ( Figure 7B). Urinary capacity returned to all groups after 2 weeks, and assisted urination was stopped.

Discussion
Agmatine has well-studied neuroprotective properties [31,36,[55][56][57] that could enhance the use of stem cell transplantation to treat CNS injuries, including SCI. Stem cell transplantation is intended to replace lost cells and provide trophic support to increase host neuron survival, along with host-mediated regeneration, repair, and plasticity [13,58,59]. Previously, we showed that agmatine enhanced neurogenesis of adult NPCs in the subventricular zone, and hADC transfection increased NPC differentiation [41,42,60]. Other studies have demonstrated that agmatine plays nerve protection and regeneration roles in various diseases [29,30,37,39,46,55,[61][62][63]. Here, we followed up on our previous study to confirm the therapeutic effects of hADC-overexpressing NPCs on SCI.
We first selected an SCI phase for cell implantation. The three phases of SCI (based on elapsed time after injury and pathophysiological criteria) are acute, subacute, and chronic. The acute phase lasts about 48 h immediately after initial hemorrhage, physical damage, and vascular damage: the injury results in ion imbalance, neurotransmitter accumulation (excitation toxicity), inflammation, edema, bleeding, ischemia, and cell necrosis [64][65][66][67][68][69][70]. The subacute phase is characterized by a phagocytic response and reactive astrocytes; the latter causes stellate glial scar formation, which prevents nerve tissue regeneration and majorly impedes recovery [10,65,[67][68][69][70][71][72]. Chronic SCI refers to the presence of symptoms for at least one year, along with a permanent cessation of neuronal impulse

Discussion
Agmatine has well-studied neuroprotective properties [31,36,[55][56][57] that could enhance the use of stem cell transplantation to treat CNS injuries, including SCI. Stem cell transplantation is intended to replace lost cells and provide trophic support to increase host neuron survival, along with host-mediated regeneration, repair, and plasticity [13,58,59]. Previously, we showed that agmatine enhanced neurogenesis of adult NPCs in the subventricular zone, and hADC transfection increased NPC differentiation [41,42,60]. Other studies have demonstrated that agmatine plays nerve protection and regeneration roles in various diseases [29,30,37,39,46,55,[61][62][63]. Here, we followed up on our previous study to confirm the therapeutic effects of hADC-overexpressing NPCs on SCI.
We first selected an SCI phase for cell implantation. The three phases of SCI (based on elapsed time after injury and pathophysiological criteria) are acute, subacute, and chronic. The acute phase lasts about 48 h immediately after initial hemorrhage, physical damage, and vascular damage: the injury results in ion imbalance, neurotransmitter accumulation (excitation toxicity), inflammation, edema, bleeding, ischemia, and cell necrosis [64][65][66][67][68][69][70]. The subacute phase is characterized by a phagocytic response and reactive astrocytes; the latter causes stellate glial scar formation, which prevents nerve tissue regeneration and majorly impedes recovery [10,65,[67][68][69][70][71][72]. Chronic SCI refers to the presence of symptoms for at least one year, along with a permanent cessation of neuronal impulse conduction in the spinal cord. Most such cases occur due to spinal cord deformation or vascular ischemia from trauma, tumors, and infections. During chronic SCI, nerve defects do not heal, leading to disorders such as convulsions, joint contractions, sensory inaction, and sphincter-movement abnormalities [67,69,[72][73][74]. A characteristic of this phase is the development of a syrinx after scars have formed [65]. We chose the subacute stage for transplantation because nerves are more likely to regenerate in this earlier period than during the chronic stage. Furthermore, the subacute stage is when treatments such as cell transplantation are the most needed, and actual clinical trials largely consist of this patient subpopulation. Several studies have also shown that NSC transplantation in the subacute stage has the best therapeutic effect. We thus considered our experimental timing to be the most clinically effective and relevant.
A week after transplantation, we confirmed that transplanted cells were concentrated in the lesion site. We also confirmed that the hADC-mNPC group had more transplanted cells at the damage site than the other two groups (mNPCs and pLXSN-mNPCs) ( Figure 1A,B). In our previous study, we confirmed that agmatine expression under oxidative damage increased significantly in the hADC-hMSC group compared with reference agmatine expression [45]. Additionally, we have shown that transplanted hADC-hMSCs had higher survival, proliferation, and migration than hMSCs alone, leading to enhanced functional recovery after SCI [45]. Increased agmatine secretion in response to an inflammatory environment is thus the most plausible explanation for higher graft survival in the hADC-mNPC group after SCI.
Next, we evaluated whether hADC-mNPC transplantation promotes tissue repair in SCI. Glial scar formation decreased significantly in hADC-mNPC mice compared with that in mNPC and pLXSN-mNPC mice, indicating that hADC overexpression in mNPCs enhances tissue repair and explaining the functional improvement in these animals at 2 weeks after cell transplantation (Figure 2A,B). The hADC-mNPC group also expressed lower levels of reactive astrocyte proteins ( Figure 3C).
We examined the growth and differentiation of transplanted hADC-mNPCs to better understand how they promote functional recovery and tissue repair. We observed that neuronal markers markedly increased in the hADC-mNPC group at 5 weeks posttransplantation compared with the mNPC and pLXSN-mNPC groups ( Figure 3B), a result confirmed through immunostaining ( Figure 4C). Thus, hADC-mNPCs appeared to induce neuronal differentiation after they were transplanted.
Differentiation to oligodendrocytes also occurred the most frequently among hADC-mNPCs ( Figures 3D and 4I), consistent with our previous study. After SCI, axon remyelination is important for functional recovery and is thought to depend on oligodendrocyte progenitor cells that give rise to nascent remyelinating oligodendrocytes [16,[75][76][77]. Therefore, increasing the number of oligodendrocytes increases the likelihood of recovering locomotor function [78]. Our findings suggest that transplanted hADC-mNPCs induce the differentiation of oligodendrocytes rather than astrocytes. Multiple staining and imaging experiments revealed enhanced myelin sheaths in hADC-mNPCs at 6 weeks post-SCI, along with remyelinated or mature axons ( Figure 5C,F). In contrast, the mNPC and pLXSN-mNPC groups possessed numerous degenerating myelinated axons and microglia/macrophage cells, along with noticeably fewer intact myelin sheaths ( Figure 5D,E).
We confirmed changes to BMP-2/4/7 expression after cell transplantation following SCI. BMP-2/7 expression was higher in the hADC-mNPC group than in the other groups at 1, 2, and 5 weeks after cell transplantation. This difference was significant at 2 weeks post-transplantation, when BMP-4 expression was also significantly lower in hADC-mNPC mice. Our results corroborate previous research demonstrating that increased BMP-2/7 expression after agmatine administration significantly affects neuron and oligodendrocyte production [46]. Furthermore, BMP-4 inhibits astrocytes and increases oligodendrocytes, a function that can be mediated by hADC-generated agmatine [46,79]. We therefore propose that hADC affects BMP-4 regulation through agmatine production, reducing the formation of reactive astrocytes and glial scars after SCI and eventually playing a major role in axonal regeneration.
Functional recovery following SCI depends on myelin preservation and remyelination [80,81]. Here, we clearly observed that hADC-mNPC mice had more myelin sheaths at their lesion sites than the mNPC and pLXSN-mNPC mice. We therefore concluded that hADC-mNPCs promote the regeneration of disintegrating axons when implanted in the damaged spinal cord. We also used BMS scores to verify whether hADC-mNPC transplantation actually improves post-SCI functional recovery. Our results indicated that mouse BMS scores improved significantly at 3-4 weeks after hADC-mNPC transplantation compared with pLXSN-mNPC transplantation. By 5 weeks post-transplantation, hADC-mNPC mice were close to normal mice in terms of behavior. This improved locomotor function confirmed our prior results using agmatine treatment and hADC-hMSC transplantation after SCI [45,46,63].
Because SCI patients lose the ability to urinate, residual urine often remains to causes secondary infections and complications in the urinary tract and kidneys [78]. In SCI mice, self-voiding is similarly impossible up to 2 weeks after surgery. Here, we observed that the hADC-mNPC mice had less urine residue in the bladder and recovered self-voiding function faster.
Taken together, our behavioral findings are consistent with the hADC-hMSC results we previously reported, suggesting that hADC overexpression in mNPCs is a potential alternative for SCI treatment. To advance clinical applications, further research is needed to clarify the pharmacological mechanism of ADC in SCI. We also recommend that future studies provide a detailed analysis of upstream and downstream BMP pathways after hADC-hMSC transplantation, thus providing insight on the regulation of BMP signaling in human CNS disease.

Construction of Recombinant Retrovirus pLXSN Containing Human Arginine Decarboxylase Gene
Recombinant retroviral vectors (pLXSN) are multicloning sites that use a long terminal repeat (LTR) promoter to clone and regulate downstream genes. Following previous methods [41,42], retroviral pLXSN (K1060, Clontech, San Jose, CA, USA) vectors containing the recombinant hADC gene were transfected into mNPCs. First, full-length hADC cDNA (Gen-Bank accession number AY325129) was PCR-amplified and ligated to pLXSN ( Figure 8B). The vectors were then cloned in E. coli DH5α competent cells (Takara, Japan) and identified with restriction analysis. Next, hADC-expressing and empty pLXSN plasmids containing neomycin resistance genes were transfected into the retroviral packaging cell line PT67 (ATCC, UK) using Lipofectamine 2000 (Sigma-Aldrich, St. Louis, MO, USA) ( Figure 8C). The optimal concentration for hADC and pLXSN resistant clone selection was achieved by adding G-418 (Sigma, USA) to Dulbecco's modified Eagle's medium (DMEM) (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, USA). Cell cultures were maintained for 1 week in a 5% CO 2 humidified atmosphere at 37 • C. To determine viral titers, virus-containing medium was first filtered through a 0.45 µm poly sulfonic filter (Sartorius AG, Bohemia, NY, USA), then added with a polybrene reagent (Sigma, USA) to the NIH/3T3 cell line (ATCC, UK) for infection. Clones with the highest titer were selected and stored at −70 • C until use [43]. After 1 week of culture, mouse-derived cortical NPCs (mNPCs) were infected with empty pLXSN and hADC-containing pLXSN. After 24 h of incubation with empty or hADC-containing pLXSN, the medium was replaced with the mNPC culture medium and maintained for another week. The subsequent experiments used mNPCs infected with hADC (hADC-mNPCs), mNPCs infected with pLXSN (pLXSN-mNPCs), and noninfected control mNPCs ( Figure 8D).

Construction of Recombinant Retrovirus pLXSN Containing Human Arginine Decarboxylase Gene
Recombinant retroviral vectors (pLXSN) are multicloning sites that use a long terminal repeat (LTR) promoter to clone and regulate downstream genes. Following previous methods [41,42], retroviral pLXSN (K1060, Clontech, San Jose, CA, USA) vectors containing the recombinant hADC gene were transfected into mNPCs. First, full-length hADC cDNA (GenBank accession number AY325129) was PCR-amplified and ligated to pLXSN ( Figure 8B). The vectors were then cloned in E. coli DH5α competent cells (Takara, Japan) and identified with restriction analysis. Next, hADC-expressing and empty pLXSN plasmids containing neomycin resistance genes were transfected into the retroviral packaging cell line PT67 (ATCC, UK) using Lipofectamine 2000 (Sigma-Aldrich, St. Louis, MO, USA) ( Figure 8C). The optimal concentration for hADC and pLXSN resistant clone selection was achieved by adding G-418 (Sigma, USA) to Dulbecco's modified Eagle's medium (DMEM) (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, USA). Cell cultures were maintained for 1 week in a 5% CO2 humidified atmosphere at 37 °C. To determine viral titers, virus-containing medium was first filtered through a 0.45 µ m poly sulfonic filter (Sartorius AG, Bohemia, NY, USA), then added with a polybrene reagent (Sigma, USA) to the NIH/3T3 cell line (ATCC, UK) for infection. Clones with the highest titer were selected and stored at −70 °C until use [43]. After 1 week of culture, mouse-derived cortical NPCs (mNPCs) were infected with empty pLXSN and hADC-containing pLXSN. After 24 h of incubation with empty or hADC-containing pLXSN, the medium was replaced with the mNPC culture medium and maintained for another week. The subsequent experiments used mNPCs infected with hADC (hADC-mNPCs), mNPCs infected with pLXSN (pLXSN-mNPCs), and noninfected control mNPCs ( Figure 8D).

Animal Model of Compression SCI
Studies were conducted on male ICR mice (8 weeks old, 28 ± 5 g: Samtako, Osan,

Animal Model of Compression SCI
Studies were conducted on male ICR mice (8 weeks old, 28 ± 5 g: Samtako, Osan, Republic of Korea). All animal experiments were performed in accordance with the Korean Food and Drug Administration guidelines. Protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of Yonsei Laboratory Animal Research Center (YLARC) (permit number 2010-0350). All mice were maintained in the specific pathogenfree facility of the YLARC under controlled conditions (23 • C, 12 h:12 h light: dark cycle).
Mice were intramuscularly anesthetized with a combination of Zoletil 50 (0.6 mg/kg; Virbac, Carros, France) and Rumpun (0.4 mg/kg; Bayer, Leverkusen, Germany). Body temperatures were monitored with a rectal probe and maintained at 36.5-37.5°C using heating pads. Thoracic vertebra (T) 8 to T10 was laminectomized without damaging the dura mater, and the spinal cord at T9 was subjected to compression (15 g/mm 2 ) for 1 min using a bilateral microclamping clip (Fine Science Tools, Vancouver, BC, Canada). The bladder was manually pressed twice daily until spontaneous voiding occurred, and any hematuria or urinary tract infection was treated with ampicillin (1 mg/kg: Sigma, St. Louis, MO, USA) daily for 1 week. Food and water were freely accessible in the cages.

PKH-26 Labeling of Transplanted Cells
The mNPCs, pLXSN-mNPCs, and hADC-mNPCs were labeled with red fluorescent PKH-26 (2 × 10 −8 mol/L culture medium; Sigma, St. Louis, MO, USA), following the manufacturer's protocol. Briefly, detached cells were washed with serum-free medium and resuspended in 1 mL of dilution buffer. The cell suspension was mixed with an equal volume of labeling solution containing PKH-26 and incubated for 5 min at 23 • C. The fluorescent dye has an aliphatic reporter molecule that integrates into the cell membrane via selective partitioning ( Figure 9A). After 2 mL of serum was added to terminate the reaction, cells were washed three times with culture medium and observed under a fluorescent microscope (CKX53, Olympus, Westborough, MA, USA).
bladder was manually pressed twice daily until spontaneous voiding occurred, and any hematuria or urinary tract infection was treated with ampicillin (1 mg/kg: Sigma, St. Louis, MO, USA) daily for 1 week. Food and water were freely accessible in the cages.

PKH-26 Labeling of Transplanted Cells
The mNPCs, pLXSN-mNPCs, and hADC-mNPCs were labeled with red fluorescent PKH-26 (2 × 10 −8 mol/L culture medium; Sigma, St. Louis, MO, USA), following the manufacturer's protocol. Briefly, detached cells were washed with serum-free medium and resuspended in 1 mL of dilution buffer. The cell suspension was mixed with an equal volume of labeling solution containing PKH-26 and incubated for 5 min at 23 °C. The fluorescent dye has an aliphatic reporter molecule that integrates into the cell membrane via selective partitioning ( Figure 9A). After 2 mL of serum was added to terminate the reaction, cells were washed three times with culture medium and observed under a fluorescent microscope (CKX53, Olympus, Westborough, MA, USA). ; cellular nuclei were stained with DAPI (blue). (B) All mNPC, pLXSN-mNPC, and hADC-mNPC transplantations were performed 1 week after SCI. Transplanted cell retention was assessed in spinal cord explants at 1, 2, and 5 weeks after cell transplantation. Assessments of forearm function were performed before and after injury, and weekly following transplantation. Bladder function analysis was performed simultaneously twice a day until 2 weeks after SCI.

Locomotor Recovery and Bladder Function Assessments
Hindlimb locomotor recovery was assessed in an open field test using the nine-point Basso Mouse Scale (BMS) [82]. Scores 0-2 reflect complete absence of ankle movement to greater ankle movement; scores 3-4 correspond to improvement in step and plantar placement; scores 5-8 corresponds to paws in standing position, hindlimb-forelimb coordination, and trunk stability; and score 9 indicates normal locomotion with trunk stability. The final BMS score was the average of each group member's left and right hind legs. For 1 week before surgery, all animals were exercised once a day at the same time to confirm normal mobility. Additionally, BMS tests were performed weekly by two scorers blinded to experimental conditions, starting from 1 week before SCI and ending at 6 weeks after SCI. Starting from SCI until 14 d postinjury (DPI), bladders were manually stimulated twice daily (n = 10 per group) until mice regained normal autonomic bladder function (approximately 11-15 DPI). Retained urine from each mouse was collected and measured, both in the morning and evening sessions (12 h interval) until 14 DPI ( Figure 9B).

Glial Scar Formation Analysis
To quantify glial scar area in the mNPC, pLXSN-mNPC, and hADC-mNPC mice, tissue sections (20 mm) were obtained 2 weeks after SCI and sequentially immunoreacted with GFAP antibody (1:500, Thermo, Waltham, MA, USA) at 4 • C overnight. Subsequently, tissue sections were incubated with the appropriate biotinylated secondary antibodies. Immunostaining was performed using an ABC kit (Vector, Burlingame, CA, USA), followed by reaction with 3,39-diaminobenzidine tetra hydrochloride (DAB, Sigma, St. Louis, MO, USA). Negative controls lacked the primary antibody. Glial scar area was obtained from measuring GFAP-positive regions around the lesion site (with central cavity and the number of reactive astrocytes) in Image J (National Institutes of Health, Bethesda, MD, USA).

Luxol Fast Blue Staining
Myelinated tissue was visualized using Luxol fast blue (LFB) staining. Spinal cord tissue sections (n = 5 per group) were rinsed in PBS and serially dehydrated in ethanol solutions (70%, 95%, and 100%) for 30 min each. Sections were then placed in 0.1% LFB solution (Sigma, USA) for oven incubation at 56 • C for 4 h. Excess stain was rinsed with 95% ethanol. To differentiate staining, tissue sections were incubated in 0.05% lithium carbonate solution and counterstained with 0.1% cresyl violet solution (Sigma, USA) for 30 s.

Assessing Ultrastructural Spinal Cord Changes
Post-SCI microstructural changes in the myelin sheath were assessed with transmission electron microscopy (TEM). Briefly, mice were perfused with normal saline, followed by a solution containing 2% glutaraldehyde and 4% PFA. After thermal stress for 12 h, each sample was fixed with 2% glutaraldehyde-PFA in 0.1 M PBS for 2 h and washed three times for 30 min in 0.1 M PBS. They were then postfixed with 1% OsO 4 dissolved in 0.1 M PBS for 2 h, dehydrated in an ascending ethanol series (50-100%), and infiltrated with propylene oxide. Specimens were embedded using a Poly/Bed 812 Embedding Kit (Polysciences, Warrington, PA, USA) and polymerized at 60 • C in an electron microscope oven (TD-700, DOSAKA, Japan) for 24 h. After incubation, sections (350 nm) were sliced and stained with toluidine blue to confirm embedding quality under a light microscope. Thinner [70 nm] sections were then sliced in a LEICA Ultracut UCT Ultra-microtome (Leica Microsystems, Germany) and counter-stained with 7% uranyl acetate and lead citrate for 20 min. These sections were observed under a TEM (JEM-1011, JEOL, Tokyo, Japan) at an acceleration voltage of 80 kV.

Statistical Analysis
All data are presented as means ± standard error. One-way analysis of variance (ANOVA) with post hoc Tukey's HSD test was used to determine between-group variation in SPSS 18.0. Significance was set at p < 0.05.

Conclusions
Our results suggest that hADC gene overexpression is an effective way to enhance the therapeutic potential of cell therapy for SCI. Transplanting hADC-mNPCs in an SCI mouse model improved locomotor and bladder function, decreased initial glial scar formation, induced remyelination via oligodendrogenesis, and increased neuronal differentiation of transplanted cells. To improve the degree and speed of functional recovery, the optimal stage for cell transplantation must be determined through further research. In particular, evidence from comparative data will be extremely useful. Future studies should also confirm the effects of ADC overexpression through different means, such as transplanting other virus types or even nonviruses that include the ADC gene.