Gene therapy in the central nervous system: Direct versus indirect gene delivery

Over the last decade, the combination of molecular biology and cell transplantation techniques has given rise to a powerful method for gene therapy. The implantation of genetically modified cultured cells has been extensively used in the central nervous system (CNS) in various experimental models of neurologic disorders. More recently, viral and chemical methods have been developed to further efforts to shuttle transgenes into the relatively inaccessible brain. Adenoviral and liposomal synthetic vectors carry transgenes into neural tissue in situ and are beginning to show promise as new methods for CNS therapy. © 1995 Wiley-Liss, Inc.     

Immune problems in central nervous system cell therapy

Transplantation of cells and tissues to the mammalian brain and CNS has revived the interest in the immunological status of brain and its response to grafted tissue. The previously held view that the brain was an absolute “immunologically privileged site” allowing indefinite survival without rejection of grafts of cells has proven to be wrong. Thus, the brain should be regarded as a site where immune responses can occur, albeit in a modified form, and under certain circumstances these are as vigorous as those seen in other peripheral sites. Clinical cell transplant trials have now been performed in Parkinson’s disease, Huntington’s disease, demyelinating diseases, retinal disorders, stroke, epilepsy, and even deafness, and normally are designed as cell replacement strategies, although implantation of genetically modified cells for supplementation of growth factors has also been tried. In addition, some disorders of the CNS for which cell therapies are being considered have an immunological basis, such as multiple sclerosis, which further complicates the situation. Embryonic neural tissue allografted into the CNS of animals and patients with neurodegenerative conditions survives, makes and receives synapses, and ameliorates behavioral deficits. The use of aborted human tissue is logistically and ethically complicated, which has lead to the search for alternative sources of cells, including xenogeneic tissue, genetically modified cells, and stem cells, all of which can and will induce some level of immune reaction. We review some of the immunological factors involved in transplantation of cells to CNS.     

Human neural transplantation

Great advances in neurobiology have resulted from 100 years of neural transplantation research. In the last 20 years, there has been a focus on using neural transplantation to repair the damaged central nervous system (CNS) utilising experimental animal models of various human neurodegenerative disease and CNS injury. Since 1985, there has been a rapid proliferation of adrenal medullary autograft transplantation to the caudate nucleus of humans with Parkinson's disease. However, this operation proved to be unsuccessful and was associated with unacceptable morbidity. Implantation of human fetal mesencephalon into patients with severe parkinsonism has supplanted the adrenal operation and has produced promising results, with some patients reported to improve markedly and some evidence of graft survival noted on positron emission tomography (PET). Host tissue recovery appears to be an important mechanism for this clinical improvement. The optimal technique is to use three to four fetuses from induced abortions of 6.5 to 8 weeks gestation, with multiple stereotactic implants into the putamen and caudate nucleus. Many biological questions still remain and the community remains troubled by the ethical problems of using fetal tissue obtained from abortions. This procedure is still experimental and should be restricted to a few centres with excellence in cell and molecular biology. A multicentre study is needed to more carefully evaluate CNS transplantation. Cloned neural precursor cells or immortalized embryonic cell lines genetically modified to manufacture selected growth factors or neurotransmitters may offer an alternative to the use of human fetal tissue. Much more experimental animal research is necessary before transplantation can be used to treat other CNS maladies.     

细胞移植对脑出血脑损伤的保护作用

目前用于脑出血治疗研究的主要细胞有：神经干细胞、遗传工程神经干细胞、骨髓间质干细胞、脐血细胞、胚胎干细胞、重组细胞、微囊化人工细胞等，本文就前5种细胞移植在出血性脑损伤中的保护作用进行综述。干细胞移植在脑出血性动物身上的实验所取得的良好效果，显示细胞移植重建损伤的脑组织、改善脑功能成为治疗脑出血疾病的必然途径。临床实验也取得了一定效果，但要真正应用于临床目前还有许多问题要解决：要想准确的定向诱导细胞分化，需要更深入的研究局部微环境，细胞因子及基因对干细胞分化的作用；细胞移植促进脑出血功能的恢复，其确切的机制还需要进一步研究 ；诱导分化的神经干细胞是否能象正常神经细胞一样能分泌神经递质，是否具有复杂的电生理等。     

Genetically modified NT2N human neuronal cells mediate long-term gene expression as CNS grafts in vivo and improve functional cognitive outcome following experimental traumatic brain injury

Human Ntera-2 (NT2) cells can be differentiated in vitro into well-characterized populations of NT2N neurons that engraft and mature when transplanted into the adult CNS of rodents and humans. They have shown promise as treatments for neurologic disease, trauma, and ischemic stroke. Although these features suggest that NT2N neurons would be an excellent platform for ex vivo gene therapy in the CNS, stable gene expression has been surprisingly difficult to achieve in these cells. In this report we demonstrate stable, efficient, and nontoxic gene transfer into undifferentiated NT2 cells using a pseudotyped lentiviral vector encoding the human elongation factor 1-alpha promoter and the reporter gene eGFP. Expression of eGFP was maintained when the NT2 cells were differentiated into NT2N neurons after treatment with retinoic acid. When transplanted into the striatum of adult nude mice, transduced NT2N neurons survived, engrafted, and continued to express the reporter gene for long-term time points in vivo. Furthermore, transplantation of NT2N neurons genetically modified to express nerve growth factor significantly attenuated cognitive dysfunction following traumatic brain injury in mice. These results demonstrate that defined populations of genetically modified human NT2N neurons are a practical and effective platform for stable ex vivo gene delivery into the CNS.     

Brain transplantation of genetically engineered human neural stem cells globally corrects brain lesions in the mucopolysaccharidosis type VII mouse

In the present study, we investigated the feasibility of using human neural stem cells (NSCs) in the treatment of diffuse central nervous system (CNS) alterations in a murine model of mucopolysaccharidosis VII (MPS VII), a lysosomal storage disease caused by a genetic defect in the beta-glucuronidase gene. An immortalized NSC line derived from human fetal telencephalon was genetically engineered to overexpress beta-glucuronidase and transplanted into the cerebral ventricles of neonatal MPS VII mouse. Transplanted human NSCs were found to integrate and migrate in the host brain and to produce large amount of beta-glucuronidase. Brain contents of the substrates of beta-glucuronidase were reduced to nearly normal levels, and widespread clearing of lysosomal storage was observed in the MPS VII mouse brain at 25 days posttransplantation. The number of engrafted cells decreased markedly after the transplantation, and it appears that the major cause of the cell death was not the immune response of the host but apoptotic cell death of grafted human NSCs. Results showed that human NSCs would serve as a useful gene transfer vehicle for the treatment of diffuse CNS lesions in human lysosomal storage diseases and are potentially applicable in the treatment of patients suffering from neurological disorders.     

Identification and early characterization of genetically modified NGF-producing neural stem cells grafted into the injured adult rat brain

OBJECTIVE: To investigate whether genetically modified mouse neural stem cells (NSC) expressing recombinant human nerve growth factor (rhNGF) and transplanted in chemically injured rat brain, can survive and eventually acquire phenotypic characteristics of early nerve cells. METHODS: Stably high expression of rhNGF in NSC was obtained by a new lentivirus-mediated expression system. To test the effectiveness of hNGF secreted by rhNGF-NSC, hereby we performed either a bioassay for neurite outgrowth in PC12 rat cells or immunoblot analysis for TrkA, the high-affinity NGF receptor, from engineered NSC. rhNGF and mock-NSC were grafted into adult injured rats striatum and 3 days later, animals were killed, and brains were removed and examined by immunohistochemical analysis. RESULTS: The results showed that rhNGF-producing NSC cultured for extended period of time release bioactive hNGF in the culture media which promotes PC12 neuronal differentiation and correlates with the up-regulation of TrkA. rhNGF-NSC transplanted into the injured brain can survive, produce hNGF and induce the expression of NGF receptors, p75(NTR) and TrkA. Discussion:In vitro and in vivo experiments confirmed the ability of rhNGF-NSC to secrete bioactive hNGF. Our data provide by means of genetically modified rhNGF-producing NSC, a useful experimental tool to test the potential clinical effectiveness of trophic factors relevant to central nervous system (CNS).     

Human neural stem cells genetically modified for brain repair in neurological disorders

Existence of multipotent neural stem cells (NSC) has been known in developing or adult mammalian CNS, including humans. NSC have the capacity to grow indefinitely and have multipotent potential to differentiate into three major cell types of CNS, neurons, astrocytes and oligodendrocytes. Stable clonal lines of human NSC have recently been generated from the human fetal telencephalon using a retroviral vector encoding v‐myc. One of the NSC lines, HB1.F3, carries normal human karyotype of 46XX and has the ability to self‐renew, differentiate into cells of neuronal and glial lineages, and integrate into the damaged CNS loci upon transplantation into the brain of animal models of Parkinson disease, HD, stroke and mucopolysaccharidosis. F3 human NSC were genetically engineered to produce L‐dihydroxyphenylalanine (L‐DOPA) by double transfection with cDNA for tyrosine hydroxylase and guanosine triphosphate cylohydrolase‐1, and transplantation of these cells in the brain of Parkinson disease model rats led to L‐DOPA production and functional recovery. Proactively transplanted F3 human NSC in rat striatum, supported the survival of host striatal neurons against neuronal injury caused by 3‐nitropro‐pionic acid in rat model of HD. Intravenously introduced through the tail vein, F3 human NSC were found to migrate into ischemic lesion sites, differentiate into neurons and glial cells, and improve functional deficits in rat stroke models. These results indicate that human NSC should be an ideal vehicle for cell replacement and gene transfer therapy for patients with neurological diseases. In addition to immortalized human NSC, immortalized human bone marrow mesenchymal stem cell lines have been generated from human embryonic bone marrow tissues with retroviral vectors encording v‐myc or teromerase gene. These immortalized cell lines of human bone marrow mesenchymal stem cells differentiated into neurons/glial cells, bone, cartilage and adipose tissue when they were grown in selective inducing media. There is further need for investigation into the neurogenic potential of the human bone marrow stem cell lines and their utility in animal models of neurological diseases.     

Current concepts in central nervous system regeneration

A dictum long-held has stated that the adult mammalian brain and spinal cord are not capable of regeneration after injury. Recent discoveries have, however, challenged this dogma. In particular, a more complete understanding of developmental neurobiology has provided an insight into possible ways in which neuronal regeneration in the central nervous system may be encouraged. Knowledge of the role of neurotrophic factors has provided one set of strategies which may be useful in enhancing CNS regeneration. These factors can now even be delivered to injury sites by transplantation of genetically modified cells. Another strategy showing great promise is the discovery and isolation of neural stem cells from adult CNS tissue. It may become possible to grow such cells in the laboratory and use these to replace injured or dead neurons. The biological and cellular basis of neural injury is of special importance to neurosurgery, particularly as therapeutic options to treat a variety of CNS diseases becomes greater.     

Transplants of cells genetically modified to express neurotrophin-3 rescue axotomized Clarke's nucleus neurons after spinal cord hemisection in adult rats

To test the idea that genetically engineered cells can rescue axotomized neurons, we transplanted fibroblasts and immortalized neural stem cells (NSCs) modified to express neurotrophic factors into the injured spinal cord. The neurotrophin-3 (NT-3) or nerve growth factor (NGF) transgene was introduced into these cells using recombinant retroviral vectors containing an internal ribosome entry site (IRES) sequence and the beta-galactosidase or alkaline phosphatase reporter gene. Bioassay confirmed biological activity of the secreted neurotrophic factors. Clarke's nucleus (CN) axons, which project to the rostral spinal cord and cerebellum, were cut unilaterally in adult rats by T8 hemisection. Rats received transplants of fibroblasts or NSCs genetically modified to express NT-3 or NGF and a reporter gene, only a reporter gene, or no transplant. Two months postoperatively, grafted cells survived at the hemisection site. Grafted fibroblasts and NSCs expressed a reporter gene and immunoreactivity for the NGF or NT-3 transgene. Rats receiving no transplant or a transplant expressing only a reporter gene showed a 30% loss of CN neurons in the L1 segment on the lesioned side. NGF-expressing transplants produced partial rescue compared with hemisection alone. There was no significant neuron loss in rats receiving grafts of either fibroblasts or NSCs engineered to express NT-3. We postulate that NT-3 mediates survival of CN neurons through interaction with trkC receptors, which are expressed on CN neurons. These results support the idea that NT-3 contributes to long-term survival of axotomized CN neurons and show that genetically modified cells rescue axotomized neurons as efficiently as fetal CNS transplants.     

Transplants of cells engineered to produce GABA suppress spontaneous seizures

PURPOSE: Cell transplantation into the brain is an aggressive clinical alternative. The hopes of treating diseases like intractable temporal lobe epilepsy have been subdued because the preclinical successes thus far have shown only slowing of epileptogenesis, or suppression of electrically induced seizures. Because the hallmark of epilepsy is spontaneous seizures, the clinical relevance of these studies has been questioned. The purpose of this study was to establish that cells genetically engineered to produce gamma-aminobutyric acid (GABA) could suppress spontaneous seizures in an accepted model of temporal lobe epilepsy. METHODS: Conditionally immortalized neurons were engineered to produce GABA under the control of tetracycline. These cells were transplanted into the substantia nigra of spontaneously seizing animals. After transplantation, the animals were monitored for 3 days immediately after surgery and again for 3 days beginning 7-8 days after surgery. Seizures and epileptiform spikes were recorded and later analyzed with detection software combined with video monitoring. RESULTS: Animals that received genetically engineered GABA-producing cells had significantly fewer spontaneous seizures than did animals that received control cells, or animals that received GABA-producing cells plus doxycycline at the observation period starting 1 week after transplantation. A significant suppression of epileptiform spikes also was noted between the group that received GABA-producing cells and the group that received the same cells but were given doxycycline. The engineered cells show evidence of integration with the host but limited survival. CONCLUSIONS: These data demonstrate that genetically engineered cells have the ability to suppress spontaneous seizures when transplanted into seizure-modulating nuclei. This is an important step toward defining a clinical potential for this approach in epilepsy. The fact that the gene of interest can be regulated suggests that individualizing transplant therapy may be possible.     

脑源性神经营养因子基因工程细胞脑内移植治疗帕金森病大鼠的实验研究

目的观察脑源性神经营养因子（ＢＤＮＦ）基因工程成肌细胞脑内纹状体移植对帕金森病大鼠的治疗作用。方法建立逆转录病毒介导的ＢＤＮＦ表达质粒并转染成肌细胞,筛选阳性细胞进行脑内移植。结果基因工程成肌细胞脑内移植可明显提高帕金森病大鼠黑质酪氨酸羟化酶（ＴＨ）阳性神经元的存活率,使纹状体多巴胺含量明显增加,动物的旋转行为改善约５０％,并持续２个月之久。结论脑源性神经营养因子基因工程成肌细胞脑内纹状体移植可明显改善动物的旋转行为并可促进黑质神经原的存活,为帕金森病的治疗提供了一种新的有效的治疗方法     

Migration of cells from primary transplants of allo- and xenografted foetal striatal tissue in the adult rat brain

Primary neural cells derived from human xenografts migrate extensively following transplantation into the adult rat CNS. However, it is unknown whether cells from allografts have the same capability to migrate within the adult rat brain. Moreover, it is unclear whether human-derived cells migrate to this extent as an inherent property of being in a xenograft environment, or whether it is due to the large size of the developed human brain compared with the adult rat brain. In order to address these issues we have designed an experimental paradigm to investigate the potential for cells derived from grafts of primary rat, mouse and human foetal striatal tissue to migrate following intrastriatal transplantation in an adult rat model of Huntington's disease (HD). Green fluorescent protein (GFP)-expressing rat and mouse donors and an antibody specific to human nuclear antigen enabled identification of graft-derived cells within the host brain, and double-labelling with GFP and neuronal nuclear antigen or immunostaining with human-specific tau identified graft-derived neurons. Twelve weeks post-transplantation, cells had migrated throughout the host in all groups; however, human cells and neurons had migrated significantly more than rat or mouse cells. These results demonstrate that neural cells derived from allografts are capable of migrating in the adult rat CNS and that the extent of migration is most likely determined by the size of the mature donor adult brain. This has important implications for the use of allo- and xenogeneic tissue as a source for transplantation in treating diffuse neurodegenerative disorders such as HD.     

Neural-Targeted Gene Therapy for Rodent and Primate Hemiparkinsonism

Expression of the rate-limiting enzyme for catecholamine biosynthesis, tyrosine hydroxylase (TH), via retrovital and plasmid expression vectors improved the efficacy of conditionally immortalized nigral neural cells in ameliorating rodent and nonhuman primate models of Parkinson's disease through neural transplantation. No improvement in rotational behavior occurred when sham transplants or nondopaminergic transplants were performed. Transplantation of the temperature-sensitive immortalized parental nigral neural line with a TH expression vector resulted in improvement for at least 2 months. Improvement was accompanied by HPLC evidence of increased -DOPA production and immunocytochemical evidence of TH in the transfected cells increased over that of the parental line. No tumor formation was detected. These results suggest that: (1) temperature-sensitive immortalized neural cells may be genetically engineered successfully to improve their efficacy for the treatment of parkinsonism; and (2) a change in -DOPA production, as opposed to growth factor production or other factors, is likely to account for the observed improvement, since the parental and derived lines differ by a single gene.     

Neural Stem Cell Transplantation as a Therapeutic Approach for Treating Lysosomal Storage Diseases

Treating the central nervous system manifestations of subjects with neuropathic lysosomal storage diseases remains a major technical challenge. This is because of the low efficiency by which lysosomal enzymes in systemic circulation are able to traverse the blood brain barrier into the central nervous system. Intracranial transplantation of neural stems cells genetically modified to overexpress the respective deficient enzymes represents a potential approach to addressing this group of diseases. The unique properties of neural stem cells and progenitor cells, such as their ability to migrate to distal sites, differentiate into various cell types and integrate within the host brain without disrupting normal function, making them particularly attractive therapeutic agents. In addition, neural stem cells are amenable to ex vivo propagation and modification by gene transfer vectors. In this regard, transplanted cells can serve not only as a source of lysosomal enzymes but also as a means to potentially repair the injured brain by replenishing the organ with healthy cells and effecting the release of neuroprotective factors. This review discusses some of the well-characterized neural stem cell types and their possible use in treating neuropathic lysosomal storage diseases such as the Niemann Pick A disease.     

Expression of brain natriuretic peptide by human bone marrow stromal cells

Bone marrow stromal cells (BMSC) have been shown to generate neural cells under experimental conditions in vitro and following transplantation into animal models of stroke and traumatic CNS injury. Hastened recovery from the neurological deficit has not correlated with structural repair of the lesion in the stroke model. Secretory functions of BMSC, such as the elaboration of growth factors and cytokines, have been hypothesized to play a role in the enhanced recovery of neurological function. Using gene expression arrays, real time RT-PCR and radioimmunoassay, we have found that brain natriuretic peptide (BNP) is synthesized and released by BMSC at physiologically relevant levels in vitro. BNP, like its close homolog atrial natriuretic peptide (ANP), exerts powerful natriuretic, diuretic and vasodilatory effects. We speculate that transplanted BMSCs facilitate recovery from brain and spinal cord lesions by releasing BNP and other vasoactive factors that reduce edema, decrease intracranial pressure and improve cerebral perfusion.     

Nurr1基因在体外培养大鼠骨髓源性神经干细胞的表达

目的获得表达孤儿核受体（Nurr1）基因的骨髓源性神经干细胞（BMSCs-NSCs）。方法构建携带Nurr1基因的重组腺相关病毒（AAV）载体AAV-pcDNA3．1-Nurr1，提取质粒，用脂质体转染法转染大鼠BMSCs-NSCs，并用RT-PCR和免疫细胞化学方法检测阳性细胞。结果经酶切鉴定和DNA测序，证实得到了序列正确的重组pAAV-Nurr1，获得了Nurr1阳性的BMSCs-NSCs。结论重组AAV携带的Nurr1基因能够在BMSCs-NSCs中表达，本研究为进一步探讨将该基因工程细胞用于帕金森病基因治疗提供了可能性，     

Long‐term fate of grafted hippocampal neural progenitor cells following ischemic injury

The adult CNS has a very limited capacity to regenerate neurons after insult. To overcome this limitation, the transplantation of neural progenitor cells (NPCs) has developed into a key strategy for neuronal replacement. This study assesses the long‐term survival, migration, differentiation, and functional outcome of NPCs transplanted into the ischemic murine brain. Hippocampal neural progenitors were isolated from FVB‐Cg‐Tg(GFPU)5Nagy/J transgenic mice expressing green fluorescent protein (GFP). Syngeneic GFP‐positive NPCs were stereotactically transplanted into the hippocampus of FVB mice following a transient global cerebral ischemia model. Behavioral tests revealed that ischemia/reperfusion induced spatial learning disturbances in the experimental animals. The NPC transplantation promoted cognitive function recovery after ischemic injury. To study the long‐term fate of grafted GFP‐positive NPCs in a host brain, immunohistochemical approaches were applied. Confocal microscopy revealed that grafted cells survived in the recipient tissue for 90 days following transplantation and differentiated into mature neurons with extensive dendritic trees and apparent spines. Immunoelectron microscopy confirmed the formation of synapses between the transplanted GFP‐positive cells and host neurons that may be one of the factors underlying cognitive function recovery. Repair and functional recovery following brain damage represent a major challenge for current clinical and basic research. Our results provide insight into the therapeutic potential of transplanted hippocampal progenitor cells following ischemic brain injury. © 2014 Wiley Periodicals, Inc.     

Inducible expression of FGF2 by a rat oligodendrocyte precursor cell line promotes CNS myelination in vitro

Transplantation of glial cells into the central nervous system (CNS) may be a promising approach for the treatment of myelin disorders such as multiple sclerosis (MS). Myelination by transplantation of oligodendrocyte precursors has been obtained in different animal models of demyelination. A strategy to favor CNS remyelination is to enrich the lesioned areas in growth factors to stimulate the quiescent population of oligodendrocyte precursors. In this context, we have developed a genetically modified CG4 cell line (CG4-FGF2), which are able to release significant amounts of fibroblast growth factor 2 (FGF2) in a controlable fashion in vitro. The data presented here demonstrate that upon induction with Dox, CG4-FGF2 cells retain their capacity to differentiate in vitro. Additionally, we provide evidence that FGF2 release by engineered cells enhance proliferation and migration of cells of the oligodendrocyte lineage without preventing them to differentiate and myelinate axons in vitro.     

移植分泌神经营养素3的人胚神经干细胞在脊髓损伤大鼠体内存活及对运动功能的影响

BACKGROUND： Many methods have been attempted to repair nerves following spinal cord injury, including peripheral nerve transplantation, Schwann cell transplantation, olfactory ensheathing cell transplantation, and embryonic neural tissue transplantation. However, there is a need for improved outcomes.
OBJECTIVE： To investigate the repair feasibility for rat spinal cord injury using human neural stem cells （hNSCs） genetically modified by lentivirus to express neurotrophin-3.
DESIGN, TIME AND SETTING： In vitro cell biological experiment and in vivo randomized, controlled genetic engineering experiment were performed at the Third Military Medical University of Chinese PLA and First People＇s Hospital of Yibin, China from March 2006 to December 2007.
MATERIALS： A total of 64 adult, female, Wistar rats were used for the in vivo study. Of them, 48 rats were used to establish models of spinal cord hemisection, and were subsequently equally and randomly assigned to model, genetically modified hNSC, and normal hNSC groups. The remaining 16 rats served as normal controls.
METHODS： hNSCs were in vitro genetically modified by lentivirus to secrete both green fluorescence protein and neurotrophin-3. Neurotrophin-3 expression was measured by Western blot. Genetically modified hNSC or normal hNSC suspension （5 × 10^5） was injected into the rat spinal cord following T10 spinal cord hemisection. A total of 5μL Dulbecco＇s-modified Eagle＇s medium was infused into the rat spinal cord in the model grop. Transgene expression and survival of transplanted hNSCs were determined by immunohistochemistry. Motor function was evaluated using the Basso, Beattie, and Bresnahan （BBB） scale.
MAIN OUTCOME MEASURES： The following parameters were measured： expression of neurotrophin-3 produced by genetically modified hNSCs, transgene expression and survival of hNSCs in rats, motor function in rats.
RESULTS： hNSCs were successfully genetically modified by lentivirus to stably express neurotrophin-3. The transplanted hNSCs primarily gathered at, or around, the injection site two weeks following transplantation, and gradually migrated towards the surrounding tissue. Transplanted hNSCs were observed 7.0-8.0 mm away from the injection site. In addition, hNSCs were observed 10 weeks after transplantation. At week 4, BBB locomotor scores were significantly greater in the genetically modified hNSC and normal hNSC groups, compared with the model group （P 〈 0.05）, and scores were significantly greater in the genetically modified hNSC group compared with the normal hNSC group （P 〈 0.05）.
CONCLUSION： hNSCs were genetically modified with lentivirus to stably secrete neurotrophin-3. hNSCs improved motor function recovery in rats following spinal cord injury.