|Home | About | Journals | Submit | Contact Us | Français|
Neural progenitor cells (NPs) have shown several promising benefits for the treatment of neurological disorders. To evaluate the therapeutic potential of human neural progenitor cells (hNPs) in amyotrophic lateral sclerosis (ALS), we transplanted hNPs or growth factor (GF)-expressing hNPs into the central nervous system (CNS) of mutant Cu/Zn superoxide dismutase (SOD1G93A) transgenic mice. The hNPs were engineered to express brain-derived neurotrophic factor (BDNF), insulin-like growth factor-1 (IGF-1), VEGF, neurotrophin-3 (NT-3), or glial cell-derived neurotrophic factor (GDNF), respectively, by adenoviral vector and GDNF by lentiviral vector before transplantation. Donor-derived cells engrafted and migrated into the spinal cord or brain of ALS mice and differentiated into neurons, oligodendrocytes, or glutamate transporter-1 (GLT1)-expressing astrocytes while some cells retained immature markers. Transplantation of GDNF- or IGF-1-expressing hNPs attenuated the loss of motor neurons and induced trophic changes in motor neurons of the spinal cord. However, improvement in motor performance and extension of lifespan were not observed in all hNP transplantation groups compared to vehicle-injected controls. Moreover, the lifespan of GDNF-expressing hNP recipient mice by lentiviral vector was shortened compared to controls, which was largely due to the decreased survival times of female animals. These results imply that although implanted hNPs differentiate into GLT1-expressing astrocytes and secrete GFs, which maintain dying motor neurons, inadequate trophic support could be harmful and there is sexual dimorphism in response to GDNF delivery in ALS mice. Therefore, additional therapeutic approaches may be required for full functional recovery.
Amyotrophic lateral sclerosis (ALS) is a late-onset neurodegenerative disease characterized by progressive and lethal motor neuronal death. The disease's hallmark is a selective degeneration and loss of motor neurons in the brain and spinal cord. Most cases (90%) are sporadic and about 10% are inherited (familial ALS). Mutations in the Cu/Zn superoxide dismutase (SOD1) gene are the most common form of inherited ALS, accounting for ~20% of all familial types. This has led to the generation of transgenic mice and rats carrying SOD1G93A that have many characteristics of human ALS pathology, and these are currently the most widely used models of disease (Lepore and Maragakis, 2007). Although several hypotheses have been proposed to explain mutant SOD1-mediated pathogenesis, the exact mechanism responsible for motor neuron degeneration remains unknown (Bruijn et al., 2004; Boillee et al., 2006; Pasinelli and Brown, 2006). However, recent studies in the mutant SOD1 mouse suggest that the disease is non-cell autonomous to motor neurons (Clement et al., 2003), cells neighboring motor neurons such as mutant astrocytes and microglia strongly accelerate disease progression (Beers et al., 2006; Boillee et al., 2006; Yamanaka et al., 2008b), and mutant SOD1 in cell types other than motor neurons and oligodendrocytes contribute to the initiation of motor neuron degeneration (Yamanaka et al., 2008b). Among several possibilities, mutant astrocytes could lead to glutamate-dependent excitotoxicity due to loss of glutamate transporter-1 (GLT1) [excitatory amino acid transporter-2 (EAAT2) in humans] (Howland et al., 2002), enhance inflammatory response from microglia (Yamanaka et al., 2008b), be unable to regulate glutamate receptor subunit 2 expression in motor neurons (Van Damme et al., 2007), or release 1 or more unknown toxic factors that accelerate the death of embryonic stem cell-derived motor neurons (Di Giorgio et al., 2007; Nagai et al., 2007).
Many clinical trials and studies have been undertaken to find effective therapies for ALS, but proven methods still do not exist. Recently, some trials of effective growth factor (GF) delivery have been reported in ALS model based on the assumption that the provision of GF, which could be trophic for motor neurons, would be of benefit in motor neuron disorders. Injection of adeno-associated virus (AAV) encoding insulin-like growth factor-1 (IGF-1) gene into muscle slowed disease progression in ALS mice (Kaspar et al., 2003), and delivery of a lentivirus encoding VEGF extended survival of ALS mice (Azzouz et al., 2004). However, current viral vectors cannot alter dosage or discontinue therapy and have significant limitations for reaching widespread areas of the CNS. In addition, retrograde transport may be affected early in the disease (Williamson et al., 1999), and larger animals such as monkeys may be less efficient at taking back the protein to the spinal cord, thus limiting this approach in patients. Another approach for GF delivery is the direct infusion of trophic molecules into the spinal cord or brain, which yields widespread distribution throughout the CNS. Continuous intracerebroventricular (ICV) infusion of VEGF or IGF-1 extended survival time in ALS mice, and a phase I clinical trial using the same strategy showed a modest benefit in patients without adverse effects (Nagano et al., 2005a, b; Storkebaum et al., 2005). However, direct delivery of GF in vivo has been hindered by limited bioavailability, inability to penetrate gray matter, and a relatively short half-life.
Neural progenitor cells (NPs) possess a range of therapeutic properties for neural repair: functional neuronal or glial replacement, delivery of therapeutic gene products synthesized by stem cells, mitigation of toxic or inflammatory components of the neural environment, and replacement of multiple neural elements that define a given CNS region (Snyder et al., 1995; Flax et al., 1998; Imitola et al., 2004; Park et al., 2002, 2006a, b; Lee et al., 2007). The human neural progenitor cells (hNPs) can be expanded in long-term cultures, and differentiate into neural cell types and integrate into host neural circuitry after transplantation into the developing and adult rodent CNS (Svendsen et al., 1997; Flax et al., 1998; Imitola et al., 2004; Kim et al., 2006; Lee et al., 2007). The hNPs can also be engineered to release glial cell-derived neurotrophic factor (GDNF) and survive transplantation into the lumbar spinal cord of the SOD1G93A rat. These cells continue to release GDNF within the degenerative environment of the spinal cord and show a protective effect on motor neuron cell death (Klein et al., 2005; Suzuki et al., 2007a). However, this was not through the differentiation of donor-derived cells into GLT1-expressing astrocytes around the surviving motor neurons, and GDNF-expressing hNP grafts did not extend survival times in ALS rats. Therefore, in the present study, simultaneous combined strategies of hNP transplantation that could not only replace non-neuronal cells but also effectively deliver GF to the CNS were performed in ALS mice. Both replacement of GLT1-expressing astrocytes and delivery of GF such as IGF-1 or GDNF maintain dying motor neurons. However, improvement in motor performance and extension of lifespan were not observed. These results indicate possible hNP treatment strategies to maintain dying motor neurons for ALS, however, other therapeutic approaches are required for full functional recovery.
The titer of each adenovirus was about 2-6 × 1012 pfu/ml. To demonstrate and optimize adenoviral infection on hNPs, we infected hNPs with AdGF at different MOI ranges between 0 and 250 pfu per cell. When cells were infected with AdGF at 80 MOI, nearly 100% of hNPs appeared to be GFP-positive and healthy. The amount of secreted brain-derived neurotrophic factor (BDNF), GDNF, and IGF-1 from 1 × 106 cells for 48 h was 75.67, 82.13, and 84.68 ng, respectively. In contrast, AdVEGF was used at 8 MOI because transplantation of hNPs expressing high amounts of VEGF induced death in some experimental mice. The amount of secreted VEGF was 9.04 ng from 1 × 106 cells for 48 h. The titer of lentivirus was 1.08 × 105 TU/µl, and the cells were infected at 3 MOI. According to ELISA, hNPs-LtGDNF secreted 31.89 ng of GDNF from 1 × 106 cells during 48 h under proliferative conditions.
We performed immunocytochemistry to characterize hNPs-AdGF and hNPs-LtGDNF under differentiation conditions. Transduction of the viral vector was confirmed by GFP and GF expression (Figure 1). The majority of viral vector-infected hNPs (~90%) expressed immature cell marker nestin, and many of them expressed more than 1 of the cell type-specific proteins studied, which suggests that cells are double-labeled with glial and neuronal markers are multipotent progenitors (Kim et al., 2006) (data not shown). For example, according to types of infected viruses and secreting GF, about 30-90% of cells expressed GFAP, a protein that is expressed by astrocytes and neural stem cell/progenitor cells, and about 25-45% of cells expressed TUJ1, a neuronal marker. However, only a low percentage (~1%) of cells expressed O4, an immature oligodendrocyte marker (data not shown). These results indicate that a major portion of virus-infected GF-secreting hNPs still remain immature state under differentiation conditions in vitro.
The animals were sacrificed at 2 or 4 weeks following transplantation and the end point of the disease. In all experimental groups, hNPs and GF-expressing hNPs showed migration and engraftment throughout the spinal cord or brain when implanted via the cisterna magna or cerebral lateral ventricles, respectively (Figure 2--5).5). In vivo distribution of the donor-derived cells was identified by the expression of GFP, human cell-specific markers (hHSP27, hNuc and hGFAP), or BrdU. GFP-expressing donor-derived cells in vivo indicate that they also express GF simultaneously. Donor-derived cells migrated from the injection site to the adjacent brain stem and whole spinal cord when injected into the cisterna magna. GFP-expressing donor derived cells appeared to be migrating toward the adjacent brain stem from the fourth ventricle and engrafted within the medulla when mice were analyzed 2 weeks after transplantation (Figure 2A and B). Some donor-derived cells might migrate through the subarachnoid space and become engrafted into the white matter of the cervical or thoracic spines (Figure 2C-E). In a few cases, some engrafted cells could be found in the area of the lumbar spine of ALS mice. Moreover, human-specific GFAP-positive cells showed migration from white matter to gray matter of the cervical or thoracic spine of mice 4 weeks following transplantation (Figure 2F-H).
To investigate the patterns of differentiation of donor-derived cells, the brain stem and whole spinal cord of ALS mice were analyzed by immunohistochemistry at 4 weeks following transplantation. About 20% of implanted hNPs could be found to differentiate into TUJ1-positive neurons in the cervical or thoracic spine of mice (n = 10) (20.4 ± 1.3%, mean ± SEM) (Figure 3A-C), however, a major portion of the engrafted cells remained hNestin-expressing immature cells in the spinal cord (n = 10) (50.6 ± 1.5%) (Figure 3G-I). Some donor-derived cells also differentiated into GFAP-expressing astrocytes (n = 10) (28.5 ± 1.7%) but there was hardly any CNPase-positive oligodendrocytes (less than 1%) (data not shown). When hNPs-LtGDNF was injected into the mice, only ~1% of donor-derived cells differentiated into TUJ1-positive neurons (Figure 3D-F), however, more than 50% of implanted cells expressed hGFAP in the brain stem and spinal cord of the mice (n = 8) (53.5 ± 2.1%) (Figure 3J-L). In addition, ~30% and ~5% of GFAP-expressing donor-derived astrocytes expressed EAAT2 and EAAT1, respectively, in the cervical or thoracic spine of the mice (n = 5) (30.8 ± 1.1% and 4.7 ± 0.5%) (Figure 3M-O). These results suggest that hNPs-LtGDNF-derived astrocytes could protect host motor neurons by reducing glutamate mediated excitotoxicity.
Among all transplantation groups, GDNF- and IGF-1-expressing hNP implantation showed survival and trophic effects on motor neurons of the spinal cord of mice. When mice were analyzed 4 weeks following hNPs-LtGDNF transplantation, some donor-derived astrocytes could be found to express EAAT2 and GDNF in the cervical or thoracic spine (Figure 4A-F). To investigate whether GDNF- and/or EAAT2-expressing donor-derived astrocytes could attenuate motor neuron loss and induce trophic effects on motor neurons, we counted the number of choline acetyltransferase (Chat), NF, or NeuN-positive cells and observed the size of their cell bodies in the thoracic spine of both the hNPs-LtGDNF and vehicle-injected groups. By 4 weeks post grafting (~103-day-old animals), there was a significant (~23%) reduction in the number of motor neurons in the vehicle-injected group (Figure 4J) compared to the hNPs-LtGDNF-injected group (Figure. 4G-I). The average number of motor neurons per longitudinal section of the thoracic spine in the vehicle and hNPs-LtGDNF injected groups (n = 5) was 423 ± 32 and 550 ± 25, respectively. The hGFAP-positive donor-derived cells were located adjacent to host motor neurons and their cell bodies in the gray matter of thoracic spine, and the size of motor neuron cell bodies appeared to be bigger than that of the control group (Figure 4G-J). In case of hNPs-AdIGF-1 transplantation, BrdU-positive donor-derived cells were shown to be located juxtapositionally to the host motor neurons in the thoracic spine of ALS mice, and the host motor neurons appeared to have larger cell bodies, longer neuronal processes, and enhanced neurites branching and elongation compared to vehicle-injected controls (Figure 5).
We performed the wire maneuver and rota-rod tests to evaluate the effects of hNPs and GF-expressing hNP transplantation on motor function in ALS mice. We also observed changes in body weight in all transplantation groups and vehicle-injected controls from before transplantation up to death (data not shown). Overall, there was no significant difference in motor performance and body weight transition among the hNPs and hNPs-AdGF transplantation groups and vehicle-injected controls although hNPs-AdGDNF implanted mice showed short-term transient improvement in motor performance during the early period following implantation (Figure 6A and B). The number of mice examined in the vehicle, hNPs, hNPs-AdBDNF, AdIGF-1, hNPs-AdVEGF, and hNPs-AdGDNF transplanted groups was 128, 16, 18, 22, 29, and 46, respectively. In the hNPs-LtGDNF transplanted group (n = 40), motor performance in the rota-rod test was better than in the vehicle-injected controls (n = 37) at 3 and 4 weeks following transplantation, however, after then, performance became worse compared to the control group (Figure 6C). When the results of the rota-rod test were analyzed by gender of the mice, male mice in the hNPs-LtGDNF transplanted group had better performance during the early period following transplantation than male mice in the control group whereas the late deterioration in motor performance in the hNPs-LtGDNF transplanted group was largely due to worsened performance in female mice (Figure 6D and E). However, there was no difference on the wire maneuver test between both groups (Figure 6F).
To test the role of transplantation of hNPs and GF-expressing hNPs on disease outcome, we observed the lifespan in all transplantation groups and vehicle-injected controls. There were no statistical differences among hNPs (n = 16) and hNPs-AdGF-transplanted mice (n = 115) and vehicle-injected mice (n = 124) in survival duration (hNPs, 137.1 ± 4.1 d; hNPs-AdBDNF, 139.2 ± 4.1 d, n = 18; hNPs-AdIGF-1, 134.5 ± 2.9 d, n = 22; hNPs-AdVEGF, 133.5 ± 1.6 d, n = 29; hNPs-AdGDNF, 135.8 ± 1.9 d, n = 46; vehicle, 133.8 ± 3.9 d) (Figure 7A). In contrast, hNPs-LtGDNF suspension transplantation significantly decreased overall survival. The average survival times in the vehicle and hNPs-LtGDNF groups were 143.3 ± 2.9 d (n = 21) and 132.8 ± 2.0 d (n = 23) when injected into the cisterna magna and 142.5 ± 2.5 d (n = 16) and 135.5 ± 2.2 d (n = 17) when injected into both the cisterna magna and cerebral lateral ventricle, respectively (Figure 7B). Both data show statistically significant differences between the hNPs-LtGDNF and control groups (P < 0.05).
In order to evaluate gender difference on survival duration, we separately analyzed survival times of male and female mice in both the hNPs-LtGDNF and control groups. In male mice, the mean survival times of the vehicle (n = 19) and hNPs-LtGDNF groups (n = 24) were 136.0 ± 2.2 d and 130.8 ± 1.8 d, respectively, when injected into the cisterna magna or both the cisterna magna and cerebral lateral ventricle. There was no statistically difference between both groups (Figure 7C). However, in female mice, the mean survival times of the vehicle (n = 18) and hNPs-LtGDNF groups (n = 16) were 150.3 ± 2.2 d and 138.6 ± 2.2 d, respectively, and there was a significantly difference between both groups (P < 0.05) (Figure 7D). Therefore, the negative effect of hNPs-LtGDNF transplantation into ALS mice on survival time was mainly due to the shortened survival period in female mice.
Transplantation of neural stem/progenitor cells has shown a range of therapeutic potential for neurological disorders because engrafted cells could replace lost or dysfunctional neurons and glias, deliver various growth factors or other therapeutic gene products that expressed intrinsically or engineered ex vivo to be overexpressed by the stem/progenitor cells, mitigate toxic or inflammatory components of the neural environment, replace multiple neural elements that define a given CNS region, and integrate seamlessly into host neural circuitry.
Previous studies of motor neuron disease models showed that transplantation of embryonic or neural stem cell-derived neural precursors replaced motor neurons, and established some connection in part with peripheral targets in a functional manner (Wu et al., 2002; Kerr et al., 2003; Harper et al., 2004; Deshpande et al., 2006). However, this strategy still has many limitations associated with appropriate motor neuron differentiation, axonal extension into the peripheral nervous system, formation of functional neuromuscular junction, and establishment of a functional neural network with the host nervous system.
Other cell transplantation approaches that did not replace injured motor neurons, including those with enhanced growth factor expression, also showed some improvement in ALS models (Aebischer et al., 1996; Garbuzova-Davis et al., 2002, 2003; Corti et al., 2004, 2007; Llado et al., 2004; Xu et al., 2006; Yan et al., 2006; Martin and Liu, 2007). Moreover, administration of growth factors was reported to have some beneficial roles on disease outcome in ALS rodent models (Federici and Boulis, 2006). Thus, stem/progenitor cell-based ex vivo gene therapy, employing cells genetically engineered to express growth factors, could provide not only appropriate missing neurons or glias but also simultaneously trophic factors for motor neurons in ALS models after transplantation. However, when GDNF-expressing human neural progenitor cells were directly transplanted into the lumbar spinal cord, there were no benefits in motor performance and animal survival although they provided limited neuroprotection (Klein et al., 2005; Suzuki et al., 2007a).
The diffuse distribution of grafted cells throughout the CNS via direct transplantation of stem/progenitor cells into cerebrospinal fluid (CSF) may be a potential alternative to intraparenchymal injection because ALS is characterized by relatively selective degeneration of motor neurons in the brain and spinal cord. Previous studies have demonstrated that neural stem/progenitor cells can be delivered to the brain and spinal cord via intraventricular and intrathecal injection because of their responsiveness to signals derived from the injured and degenerating CNS (Flax et al., 1998; Park et al., 2002, 2006a, b; Kerr et al., 2003; Lepore and Maragakis, 2007). Therefore, in this study, we initially hypothesized that the intrathecal injection of hNPs and GF-expressing hNPs by adenoviral vector through the cistern magna would show decreased motor neuron death and increase animal survival times. However, these cells overall do not improve motor behavior and animal survival although some donor-derived cells distribute diffusely in the brain stem and spinal cord and differentiate into neurons and astrocytes but largely retain the immature cell marker nestin and show limited protective effects on motor neurons. The only positive result is that GDNF-expressing hNP-implanted mice show short-term improvement in motor performance. As our next experiment, we transplanted GDNF-expressing hNPs by lentiviral vector because the expression of adenovirally delivered GDNF might be transient and insufficient, and hNPs-LtGDNF appear to have gained more glial cells-specific differentiation capacity, especially astrocytes, according to in vitro and in vivo differentiation experiments. However, as a result, motor performance and extension of lifespan were not improved but instead worsened in female mice.
As described earlier, intrathecal and/or ventricular injection of cells have several advantages, such as wide distribution of administered cells, easy application, and noninvasive methods. In this study, some transplanted hNPs-LtGDNF could survive and migrate into host tissues, differentiate into neurons and glias, including astrocytes expressing EAAT2 and GDNF, and exert neuroprotective influences on neighboring host motor neurons. However, many cells remained in the ventricular region and CSF tract of the spinal cord until 4 weeks after transplantation, and their engraftments and integrations seemed to be very limited. Because of the low nutrient level of CSF, the survival rate of transplanted cells might be low (Habisch et al., 2007). Furthermore, disruption of the blood-brain barrier and blood-spinal cord barrier in SOD1G93A transgenic mice was reported (Garbuzova-Davis et al., 2007). Thus, use of high doses of immune suppressants or combined immune suppressants would be required to increase the viability of donor-derived cells affected by immune rejection and chances for functional recovery.
In ALS mice models, axonopathy is known to induce loss of neural connection to the muscle at the neuromuscular junction long before motor neuron degeneration and initiation of symptoms (Fischer, et al., 2004; Pun et al., 2006). Transplantation of GDNF-expressing human neural precursor cells in ALS rat models induced survival of motor neurons in the spinal cord, but did not reduce the number of denervated muscle end plates (Suzuki et al., 2007). In addition, a recent study showed that when motor neurons were transplanted, axonal outgrowth and attachment to the muscle could be achieved by using GDNF-secreting cells implanted into the sciatic nerve (Deshpande et al., 2006). These findings suggest that combinatorial approaches of the protection of motor neuron cell bodies and nerve terminals in the muscle may lead to potential improvements in limb function and survival of ALS mice.
In this study, the effects of transplantation of hNPs-LtGDNF in motor performance and survival of ALS mice were negative. The main reason was worsened motor performance and survival in female mice. Late disease onset and slow progression of female SOD1 mice has been reported (Suzuki et al., 2007b), and the protective role of estrogen in female mice appear to influence disease progression (Choi et al., 2008). Interestingly, estrogen stimulates the expression of GDNF in the developing hypothalamus (Ivanova et al., 2002). Therefore, excessive expression of GDNF in the CNS due to hNPs-LtGDNF transplantation could be one of the reasons for the negative effects on motor performance and survival of female SOD1 mice. We also observed some irritable behaviors, such as jumping, squeaking, and turning, in female mice after cells were injected through both the cisterna magna and intracerebral ventricle. The delivery of growth factors for a long period in the CNS could be harmful. In patients with Parkinson's disease, side effects of centrally delivered GDNF have been reported (Kordower et al., 1999; Nutt et al., 2003). As we observed in mice transplanted with NPs-AdVEGF, the delivery of low-dose VEGF would be better than a high dose in Parkinson's models (Yasuhara et al., 2005). Therefore, modulation of expression level of growth factors in vivo seems to be very important. However, the precise mechanisms of the negative effects of hNPs-LtGDNF transplantation of female SOD1 mice in this study need to be clarified.
In chimeric mutant SOD1 mice, the aberrant function of non-neuronal cells (microglia and astrocytes) appears to contribute to disease progression progression (Clement et al., 2003; Boillee et al., 2006; Yamanaka, et al., 2008a). Chimeric mutant SOD1 animals suggest that increasing the proportion of healthy wild-type non-neuronal cells is related to the prolongation of animal survival times. Selective reduction of mutant SOD1 from astrocytes results in prolongation of disease duration (Yamanaka, et al., 2008a). Astrocytes in ALS patients and animal models have been revealed to have functional abnormalities in glutamate transport, specifically GLT1 (EAAT2) (Rothstein et al., 1995; Howland et al., 2002). Loss of this astroglial protein is known to cause excitotoxic motor neuron degeneration (Rothstein et al., 1996). Furthermore, ALS astrocytes alter the expression of dendritic glutamate (AMPA) receptors in motor neurons, also making them more susceptible to excitotoxicity (Lepore et al., 2008). More recent studies also suggest that human mutant SOD1 carrying astrocytes release several toxic factors, including the activation of NOX2 to produce oxygen radicals, which induce embryonic stem cell-derived motor neuron degeneration (Di Giorgio et al., 2007, 2008; Nagai et al., 2007; Marchetto, et al., 2008). Thus, these results indicate that any approach to replace altered astroglial function may be of therapeutic benefit. A recent study showed that focal transplantation of glial-restricted precursors into the cervical spine of ALS models efficiently differentiate into GLT1-expressing normal astrocytes, attenuate motor neuron loss, slow declines in motor functions, and extend survival and disease duration of animals (Lepore et al., 2008). In the present study, some transplanted hNPs-LtGDNF also differentiated into EAAT2- and/or EAAT1-expressing astrocytes in the brain stem and spinal cord of mice, and appeared to play a role in attenuating motor neuron death. However, compared to the study mentioned above, intrathecal transplantation of GDNF-expressing human neural progenitors in this study showed decreased engraftment and survival of donor-derived cells in the spinal cord, inaccurate targeting to key motor neuron pools that ultimately affect survival in ALS models, and inefficient differentiation into functional mature astrocytes that reduce excitotoxic motor neuron death.
We used SOD1 transgenic mice (Jackson Laboratories, Bar Harbor, ME) expressing the human SOD1G93A gene (G93A; Gly93 → Ala mutation) for the ALS model. The mutation has a dominant gain-of-function effect. Male transgenic mice were bred with background-matched B6C3/F1 wild-type females. The progeny were genotyped by PCR at 4-5 weeks after birth as previously described (Gurney, 1994), and maintained as SOD1G93A heterozygotes for subsequent studies. All transgenic ALS mice received the same care and housing and were evenly distributed to control and transplantation groups.
Human fetal tissue from cadavers at 13 weeks of gestation was obtained with full parental consent and approval of the research ethics committee of Yonsei University College of Medicine, Seoul, Korea. The methods of acquisition conformed to NIH and Korean government guidelines. The telencephalic brain tissue was dissected, dissociated in trypsin, and seeded into tissue culture-treated 100-mm plates (Corning) with serum-free growth medium. Mitogenic stimulation was achieved by adding fibroblast growth factor-2 (FGF2; R&D, Minneapolis, MI) and leukemia inhibitory factor (LIF; Sigma, St. Louis, MO). The concrete manners of the hNP culture as neurospheres are described in a previously published paper (Kim et al., 2006).
We constructed recombinant adenoviral vectors bearing human BDNF, IGF-1, VEGF, NT-3, or GDNF under the control of the constitutive cytomegalovirus (CMV) promoter, and placed these in tandem with a sequence of humanized version of the recombinant green fluorescent protein (hrGFP; Stratagene, La Jolla, CA) gene under the control of an internal ribosomal entry site (IRES) (Supplemental Data Figure S1A). Adenoviral vectors constructed in this fashion were designated as AdBDNF, AdIGF-1, AdVEGF, AdNT-3, AdGDNF, or AdGF as a whole, and produced in conformity with the AdEasy Adenoviral Vector System manual (Stratagene). Infectious recombinant viruses were purified by CsCl gradient centrifugation and titrated on 293 cells by Tissue Culture Infecting Dose 50 (TCID50; QBiogene).
We constructed a lentiviral vector carrying the GDNF gene, which was designated as LtGDNF. LtGDNF was produced by calcium-precipitate-transfection (transient transfection) method using vectors that had been obtained from Trono lab (http://tronolab.epfl.ch/) in accordance with the method offered by the manufacturer. The backbone of transfer vector is pWPI (Supplemental Data Figure S1B). The packaging and envelope plasmids are psPAX2 and pMD2G, respectively. Enhanced green fluorescent protein (eGFP) is a reporter gene. The transfection was conducted to the 293FT cell line and produced viral vectors were concentrated by ultracentrifugation under the sucrose cushion. The viral titer was checked as transducing units (TU) by FACS analysis method offered by Trono lab.
For adenoviral vector transduction, hNPs were seeded onto cell culture dishes with serum-free growth medium containing FGF2, LIF, and heparin. One hour later, recombinant adenoviruses encoding BDNF, IGF-1, VEGF, NT-3, or GDNF were added to each well with a multiplicity of infection (MOI) of 80. The medium was replaced with fresh one the next day, and transplantation of cells into mice was conducted 3 days following viral infection. Conditioned media were collected from adenoviral infected-hNP cultures to determine the levels of secreted adenovirus-derived BDNF, IGF-1, VEGF, NT-3, or GDNF via ELISA as previously described (Woodhall et al., 2001).
The transduction method of lentiviral vector encoding GDNF into hNPs was the same as for adenovirus. The titer of lentivirus was 3 MOI, and the level of GDNF secreted from hNPs was checked by PCR and ELISA. To confirm the safety of lentiviral infection, we had observed the survival, cellular morphology, and GFP and GDNF expression patterns of lentivirus-infected hNPs for at least 10 weeks prior to transplantation. In addition, neural cell types and morphologies of lentivirus-infected hNPs under proliferation and differentiation conditions were observed and identified by immunocytochemistry.
Adenovirus encoding GF or lentivirus encoding GDNF infected hNPs (designated as hNPs-AdGF or hNPs-LtGDNF) were prelabeled by ex vivo exposure to bromodeoxyuridine (BrdU, 3 µM/L; Sigma) for 5 days before transplantation. After anesthesia with an intraperitoneal injection of ketamine (50 mg/kg) and xylazine (10 mg/kg), the mice were fixed on a Benchmark digital stereotaxic instrument (myNeuroLab.com, St. Louis, MO). Ten µl of hNPs, hNPs-AdGF, or hNPs-AdGDNF suspended in H-H buffer (1 × Hanks' balanced salt solution, 10 mM HEPES, pH 7.4; GIBCO) were slowly injected into the cisterna magna of ALS mice using a 26-gauge needle about 75 days after birth. In some ALS mice, hNPs-LtGDNF suspension was injected not only into the cisterna magna but also the bilateral cerebral ventricles (4 µl into each cerebral ventricle). The ventricular injection sites were 0.9 mm lateral and 0.1 mm anterior to the bregma and 2 mm deep from the dura. H-H buffer was injected into the littermate heterozygote mice in the control group. Cyclosporine was injected (10 mg/kg) intraperitoneally daily from the day before transplantation to the day of death or sacrifice in all hNP transplantation groups and vehicle-injected controls.
To evaluate impairment in motor function in ALS mice, we measured body weight and motor performance using a wire-maneuver and rota-rod (Columbus Instruments, Columbus, OH) test once a week from the week before hNP transplantation. The maximum time allowed to an animal was 180 s in the wire-maneuver test and 300 s in the rota-rod test. Survival analysis was also conducted along with the behavioral tests. Mouse mortality was determined by the day of death or inability to right itself within 30 s when the mouse was placed on its side. During the behavior test, we also evaluated whether hNP transplantation could have an influence on the normal behavior of mice.
Using various cell type-specific markers to characterize the virus-infected hNPs in vitro, we performed immunocytochemistry with the following antibodies: anti-hNestin (human-specific Nestin, 1:200; Chemicon, Temecula, CA), anti-GFAP (glial fibrillary acidic protein, 1:1500; Dako, Carpinteria, CA), anti-hGFAP (human-specific GFAP, 1:500; Sternberger, Lutherville, MD), anti-Tuj1 (neuronal class β-Tubulin III, 1:500; Covance, Berkeley, CA), anti-NeuN (neuronal nuclear protein, 1:40; Chemicon), anti-CNPase (2'3'-cyclic nucleotide 3'-phosphodiesterase, 1:100; Chemicon), anti-O4 (oligodendrocyte marker O4, 1:200; Chemicon), anti-GFP (1:300; Invitrogen), anti-Chat (choline acetyltransferase, 1:200; Chemicon), anti-PDGFR (platelet-derived growth factor receptor, 1:100; Santa Cruz Biotechnology, Inc., Santa Cruz, CA), FITC-conjugated anti-BrdU (1:20; Roche, Mannheim, Germany), biotin-conjugated anti-hHSP27 (human-specific heat shock protein 27, 1:40; Stressgen, Ann Arbor, MI), anti-EAAT1 (1:100; Santa Cruz Biotechnology), anti-EAAT2 (1:100; Santa Cruz Biotechnology), and anti-GDNF (1:50; Santa Cruz Biotechnology).
After transplantation, immunohistochemical analysis of brain and spinal cord tissue of ALS mice model was performed. We first perfused the mice with chilled 1× PBS and 4% paraformaldehyde (PFA; Sigma) in 0.1M PIPES buffer, and then the brain and spinal cord were removed and post-fixed. These tissues were cut in 16-µm sections on cryostat and stained with the antibodies described above. The stained sections were mounted and examined by immunofluorescence microscopy (B×51; OLYMPUS, Tokyo, Japan).
Because transgenic mice have different characters according to litter cage, data were analyzed primarily by unpaired t-test. Survival analysis was conducted by Kaplan-Meier method. All data are shown as mean ± standard error of the mean (SEM).
Supplemental Data include a figure and can be found with this article online athttp://e-emm.or.kr/article/article_files/SP-41-7-05.pdf.
This research was supported by a grant from Stem Cell Research Center (SC-4130) and the Korean Science and Engineering Foundation through the National Core Research Center for Nanomedical Technology.