Transplantation of GABAergic Cells Derived from Bioreactor-Expanded Human Neural Precursor Cells Restores Motor and Cognitive Behavioral Deficits in a Rodent Model of Huntington's Disease

Abstract

Huntington's disease (HD) is a neurodegenerative disorder that is characterized by progressive dementia, choreiform involuntary movements, and emotional deterioration. Neuropathological features include the progressive degeneration of striatal γ-aminobutyric acid (GABA) neurons. New therapeutic approaches, such as the transplantation of human neural precursor cells (hNPCs) to replace damaged or degenerated cells, are currently being investigated. The aim of this study was to investigate the potential for utilizing telencephalic hNPCs expanded in suspension bioreactors for cell restorative therapy in a rodent model of HD. hNPCs were expanded in a hydrodynamically controlled and homogeneous environment under serum-free conditions. In vitro analysis revealed that the bioreactor-expanded telencephalic (BET)-hNPCs could be differentiated into a highly enriched population of GABAergic neurons. Behavioral assessments of unilateral striatal quinolinic acid-lesioned rodents revealed a significant improvement in motor and memory deficits following transplantation with GABAergic cells differentiated from BET-hNPCs. Immunohistochemical analysis revealed that transplanted BET-hNPCs retained a GABAergic neuronal phenotype without aberrant transdifferentiation or tumor formation, indicating that BET-hNPCs are a safe source of cells for transplantation. This preclinical study has important implications as the transplantation of GABAergic cells derived from predifferentiated BET-hNPCs may be a safe and feasible cell replacement strategy to promote behavioral recovery in HD.

Introduction

Huntington's disease (HD) is an inherited autosomal dominant neurodegenerative disorder of the human central nervous system (CNS) that is characterized by progressive dementia, choreiform involuntary movements, and emotional deterioration (8,69). Neuropathological features include astrogliosis, atrophy of the neostriatum, and the progressive degeneration of striatal γ-aminobutyric acid (GABA)/enkephalin-containing medium-sized spiny projection neurons to the lateral segment of the globus pallidus (1,47,66). In later disease stages, GABA/substance P/dynorphin-containing neurons that project to the substantia nigra pars reticulata and the medial segment of the globus pallidus are also affected (67).

Since there are no effective long-term treatments for HD, cell replacement therapies, such as transplantation of fetal striatal tissue to restore lost or degenerated neural circuitry, are currently being investigated (23,73). In order to undertake preclinical studies relevant to cell replacement therapies, animal models of HD have been established by performing striatal excitatory lesioning through direct injection of quinolinic acid (QA), a selective N-methyl-d-aspartate (NMDA) receptor agonist (7). Striatal QA lesioning selectively degenerates mediumsized spiny GABAergic, but not aspiny GABAergic, neurons mirroring the postmortem neuropathological characteristics of HD (7,39). Importantly, unilateral QA lesioning has been shown not to be fatal in the short- to long-term study. For example, Chen and colleagues have conducted a study in the unilateral QA lesion model of HD for a period of 6 months (17). Transplantation studies have shown that fetal striatal cell transplants (i) survive (42), (ii) restore levels of the neurotransmitter GABA (12), (iii) establish efferent and afferent connections with the host (22,87), (iv) form synaptic connections between host fibers and grafted neurons (87), and (v) alleviate both motor and cognitive deficits in rodents and nonhuman primates (26,59,62,64).

The promising results of grafting studies in animal models of HD have led to the application of neural transplantation in the clinical setting. However, clinical outcomes in fetal striatal cell transplantation studies have been inconsistent, which is likely due to factors such as donor tissue dissection and preparation techniques, transplantation target sites, immunosuppression strategies, and, importantly, patient selection criteria and follow-up (27,31,36). Nevertheless, successful cases have shown that transplanted human fetal striatal neural cells survive the implantation procedure, become integrated into the circuitry of the host basal ganglia, and are physiologically active (15,20,21,31,32,40). Furthermore, such grafts have been shown to be associated with recovery of striatofrontal cortical brain metabolism (3) and improvements in cognitive and motor function with no undesirable side effects (3,32). Whereas the transplantation of human fetal striatal cells has proven to be a promising avenue for neural cell restoration, this approach suffers from a number of drawbacks, including limited fetal tissue availability, potential for transmission of microbiological or infectious agents, variability in the tissue source, and ethical issues related to procurement of fetal tissue. Thus, if neural transplantation is to become a routine therapeutic option for the treatment of neurodegenerative diseases, an alternative and unlimited supply of transplantable cells is required. In addition, general risks associated with neural transplantation and neurosurgery such as subdural hematomas and intracerebral hemorrhages as well as psychiatric events need to be minimized (see, e.g., 38,52).

Considerable attention has now focused on the potential use of human neural precursor cells (hNPCs) as an alternative cell source in therapeutic brain repair strategies (29). A key step that will facilitate the clinical translation of hNPC-based therapies is the ability to generate these cells efficiently and reproducibly in vitro. Suspension bioreactors are scalable vessels in which the growth environment can be monitored, manipulated, and controlled. They have been employed successfully to expand populations of a number of different mammalian stem cells, thereby, providing an avenue by which large numbers of stem cells can be generated under clinically acceptable conditions. Of relevance here, we have previously described suspension bioreactor protocols that enable the large-scale and standardized expansion of clinical-grade NPCs for extended periods under carefully controlled conditions (4,5,45,77). Karyotype analysis and immunohistochemistry has revealed that the defining characteristics of the NPCs are not altered as a result of expansion (4,5). Clinically compliant screening and quality assurance protocols can be routinely implemented to ensure that the cells are nontransformed, nontumorigenic, and free of microbiological agents. Last, bioreactorexpanded hNPCs are suitable for long-term cryopreservation without altering their self-renewing capacity and multipotentiality (4,5). In our hands, transplantation studies indicate that undifferentiated mouse NPCs expanded within suspension bioreactors differentiate into astrocytes and are nontumorigenic (56,58). Thus, suspension bioreactor expansion protocols are adoptable to criteria and the demands required for a large supply of clinical-grade hNPCs for cell restoration strategies.

Preclinical studies involving the transplantation of undifferentiated NPCs into the QA-lesioned striatum have shown cell survival and migration as well as improvements in brain metabolism and motor functions (54,83–85). Furthermore, transplanted NPCs differentiate into neuronal nuclei (NeuN) and glial fibrillary acidic protein (GFAP)-expressing neurons and glia, respectively (54,83), with a fraction of NeuN-positive neurons coexpressing dopamine- and cyclic adenosine monophosphate (cAMP)-regulated neuronal phosphoprotein (DARPP-32) or glutamic acid decarboxylase (GAD), markers of medium spiny striatal neurons susceptible to degeneration in HD (83). These studies support the concept that NPCs possess characteristics that make them attractive for cell replacement therapies targeted toward neurodegenerative disorders such as HD.

While evidence suggests positive effects of transplanting undifferentiated NPCs in a rodent model of HD, a cellular basis of behavioral restoration remains elusive. Moreover, the presence of mitotically active NPCs poses a risk of aberrant cell growth leading to tumor formation (2,80). To this end, we have previously reported a method to differentiate bioreactor-expanded telencephalic (BET)-hNPCs into GABAergic cells suitable for neural transplantation (58). In the present study, we aim to investigate whether transplantation of GABAergic cells derived from BET-hNPCs improve motor and memory deficits in the QA lesion model of HD.

Materials and Methods

Experimental Design

A total of 30 female albino Wistar rats with unilateral QA lesions of the striatum were used in this study. The animals were randomly divided into three experimental groups; in Group 1, the animals were not transplanted with any cells (control, n = 10), in Group 2 (n = 10), the animals were transplanted with 800,000 undifferentiated BET-hNPCs, and in Group 3 (n = 10), the animals were transplanted with 800,000 differentiated BET-hNPCs. Following a period of recovery, each group underwent behavioral assessment to determine the efficacy of transplanting BET-hNPCs.

Cell Expansion Medium Preparation

The medium used to expand the cell population was developed at the Pharmaceutical Production Research Facility (PPRF; University of Calgary, Alberta, Canada). PPRF-h2 was generated by supplementing PPRF-m4 medium with 10 mg/L of leukemia inhibitory factor (Chemicon, Temecula, CA, USA), 1.0 mmol/L dehydroepiandrosterone (Steraloids, Newport, RI, USA), and 20 mg/L basic fibroblast growth factor (R&D Systems, Minneapolis, MN, USA). The PPRF-m4 culture medium included Dulbecco's modified Eagle's medium/nutrient mixture F12 (DMEM/F12; 1:1) (Invitrogen, Carlsbad, CA, USA), 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesul-fonic acid (HEPES) (Sigma-Aldrich, St. Louis, MO, USA), 0.6% glucose (Sigma-Aldrich), 1.73 g/L sodium bicarbonate (Sigma-Aldrich), 2.0 mM glutamine (Invitrogen), 0.023 g/L insulin (Sigma-Aldrich), 20 nM progesterone (Sigma-Aldrich), 9.0 mg/L putrescine (Sigma-Aldrich), 0.025 g/L transferrin (Sigma-Aldrich), 30 nM sodium selenite (Sigma-Aldrich), and 20 mg/L epidermal growth factor (Peprotech, Rocky Hill, NJ, USA). The medium was vacuum filtered through a 0.22-mm bottle-top filter (BD Falcon, Bedford, MA, USA) prior to use.

Cell Culture Equipment and Handling

The hNPCs used in all experiments were derived from the telencephalic region of the fetal human brain and were pooled from several fetuses. Briefly, human fetal cells (10 weeks) were obtained with maternal consent from HIV-, hepatitis B and C-, human T-cell lymphotrophic virus-, and syphilis-negative women undergoing elective abortions in the pregnancy termination unit of the Queen Elizabeth II Health Science Centre Hospital, Halifax, Nova Scotia, Canada. The cells were procured under the strict guidelines of a protocol approved by Dalhousie University and the Queen Elizabeth II Health Science Centre Hospital Ethical Review Boards. The telencephalic region of the forebrain was dissected under sterile conditions, and tissue samples of the specimen were sent for aerobic and anaerobic culture. The telencephalic region of the forebrain was then incubated in 1 ml of 0.25% trypsinethylenediaminetetraacetic acid (EDTA) (Invitrogen) at 37°C for 15 min, exposed to 10 ml of PPRF-h2 medium, spun at 1,000 rpm (refrigerated centrifuge: Beckman, model lGS-6R; rotor: GH3.8, S/N 96U22162 max. RPM 3750, Beckman Coulter Canada, Mississauga, ON, Canada) for 5 min, resuspended in 1 ml of PPRF-h2 medium, and gently triturated into a single cell suspension. Cell density and viability (≥98%) were determined with a hemocytometer using a standard 0.1% trypan blue exclusion test (Sigma-Aldrich). All cultures were maintained at 37°C in a humidified culture incubator with a 5% CO₂ atmosphere. The cells were then placed in expansion medium in a 25-cm² culture plate (Nunc, Penfield, NY, USA) (passage level 0). After 14 days in culture, the hNPCs were transferred from the culture plate into one T-75 flask containing 12.5 ml of fresh PPRF-h2 medium (passage level 1). After 5 days, the neurospheres were harvested and enzymatically dissociated into a single cell suspension using 0.25% trypsin-EDTA. Following hemo-cytometer cell counts, hNPCs were inoculated (100,000 cells/ml) into T-25 flasks each containing 5 ml of PPRF-h2 medium (passage level 2). The cultures were incubated at 37°C and 5% CO₂ for 14 days and fed every 5 days by replacing 40% of the spent medium with fresh medium. The cultures underwent an additional passage using the same subculture conditions. After 14 days, the passage level 3 hNPCs were harvested and enzymatically dissociated using 0.25% trypsin-EDTA. The resulting single cell suspension was inoculated at 100,000 cells/ml into two 125-ml bioreactors (Corning, Inc., Tewksbury MA, USA), each containing 100 ml of PPRF-h2 medium (now passage level 4). The bioreactors were placed on Thermolyne Cell-Gro slow-speed magnetic stirrers set at a speed of 100 rpm (Thermolyne, Dubuque, IA, USA). The inner surface of the bioreactor and the outer surface of the magnetic stir bar were siliconized with Sigmacote (Sigma-Aldrich) prior to use to prevent the cells from adhering. The cultures were fed every 5 days by replacing 40% of the medium with fresh medium. After 14 days in suspension culture, the aggregates at passage level 4 were harvested, placed in PPRF-h2 supplemented with 10% dimethyl sulfoxide (Sigma-Aldrich), and cryopreserved as aliquots in liquid nitrogen. To this point, the cells had spent a total of 61 days in culture.

After 1 month in cryostorage, each aliquot of hNPCs was quickly thawed in a 37°C water bath and placed in a 25-cm² T-flask (Nunc) containing 5 ml of PPRF-h2 medium. After 2 days, the hNPCs were harvested, triturated, inoculated into fresh medium (T-flasks) at 75,000 cells/ml, and then cultured for 14 days to ensure that they were fully recovered prior to the differentiation and transplantation procedures. Cell density and viability (≥98%) were determined with a hemocytometer using a standard 0.1% trypan blue exclusion test.

In Vitro Differentiation of BET-hNPC and Phenotypic Analysis

BET-hNPC neurospheres were collected and incubated in 0.25% trypsin-EDTA (Invitrogen) at 37°C for 15 min. Neurospheres were then exposed to 10 ml of differentiation medium consisting of DMEM supplemented with B27 (Invitrogen), N2 (Invitrogen), brain-derived trophic factor (BDNF, Regeneron, Terrytown, NY, USA), and valproic acid (Sigma-Aldrich), as previously described by Mukhida and colleagues (58). The neurospheres were then spun at 1,000 rpm for 5 min, resuspended in 1 ml of differentiation medium, and gently triturated into a single cell suspension. hNPC density and viability (≥98%) were determined with a hemocytometer using a standard 0.1% trypan blue exclusion test. Single cell suspensions of BET-hNPCs were then plated onto poly-l-lysine-coated 16-well chamber slides (BD Biosciences, San Jose, CA, USA) at 100,000 cells/ml. All cultures were maintained at 37°C in a humidified culture incubator (5% CO₂ atmosphere) for 7 days, with a 50% differentiation medium change every other day. Finally, the cells were exposed to 4% paraformaldehyde (PFA; Mallinckrodt, Inc., St. Louis, MO, USA) for 1 h prior to immunocytochemistry.

In Vivo Studies

All in vivo procedures were performed at the Cell Restoration Laboratory, Brain Repair Center, Dalhousie University, in accordance with the Guidelines of the Canadian Council of Animal Care and the Dalhousie University Animal Care Committee. A total of 30 adult female albino Wistar rats (Charles River, St. Constant, QC, Canada) (200–225 g) were housed with water and food available ad libitum in a colony room maintained on a 12:12 h light/dark cycle at constant temperature (19–22°C) and humidity (40–50%).

QA Lesioning

Excitotoxic animal models using intrastriatal injection of the NMDA receptor agonist QA replicate neurochemical and neuropathological changes, such as the selective degeneration of striatal medium-sized spiny GABAergic neurons, which are observed in HD. Moreover, QA lesioning of the striatum produces many of the cognitive and motor deficits symptomatic of HD that are consistent with neuronal loss (7,39).

Rats were anesthetized using an anesthetic drug mixture (2.0 ml/kg) of 25% ketamine hydrochloride (Ketalean, MTC Pharmaceuticals, Cambridge, ON, Canada), 6% xylazine (Rompun, Miles Canada, Inc., Etobicoke, ON, Canada), and 2.5% acepromazine maleate (Wyeth-Ayerst Canada, Montreal, QC, Canada) in 0.9% saline and positioned in a Kopf stereotaxic instrument (Kopf Instruments, Tujunga, CA, USA). A midline scalp incision followed by a burr hole was performed for the injection of QA using stereotactic surgery to deliver a total of 2 μl of 200 nmol QA (Sigma-Aldrich) dissolved in phosphate-buffered saline (PBS; GIBCO/Life Technologies, Grand Island, NY, USA) (pH = 7.4) into the striatum at the following coordinates from bregma: AP = +0.48 mm, ML = −3.0 mm, DV = −5.0 mm (dura) (63) with the incisor bar set at −3.3 mm. QA was delivered at a rate of 1.0 μl/min via a 10-μl Hamilton syringe (Reno, NV, USA) that was left in place for 5 min before retrieval. After surgery, the muscle and skin were closed with polyglactin sutures (Ethicon, Sommerville, NJ, USA), and animals were allowed to recover for 4 weeks prior to further testing.

Preparation of Undifferentiated BET-hNPCs for Transplantation

BET-hNPC neurospheres were collected and treated with 0.25% trypsin-EDTA at 37°C for 15 min. Neuro-spheres were then exposed to 10 ml of 0.05% DNase (Sigma-Aldrich, Oakville, ON, Canada, Cat# D5025)/ DMEM, spun at 1,000 rpm for 5 min, resuspended in 1 ml of 0.05% DNase/DMEM, and gently triturated into a single cell suspension. Undifferentiated NPC density and viability (≥98%) were determined with a hemocytometer using a standard 0.1% trypan blue exclusion test prior to transplantation.

Preparation of Differentiated BET-hNPCs for Transplantation

BET-hNPC neurospheres were collected and treated with 0.25% trypsin-EDTA at 37°C for 15 min. Neuro-spheres were then exposed to 10 ml of differentiation medium, spun at 1,000 rpm for 5 min, resuspended in 1 ml of differentiation medium, and gently triturated into a single cell suspension. BET-hNPC density and viability (≥98%) were determined with a hemocytometer using a standard 0.1% trypan blue exclusion test. The single cell suspensions were then plated on poly-l-lysine-coated 100-mm culture plates (BD Biosciences) at 100,000 cells/ml. All cultures were maintained at 37°C in a humidified culture incubator (5% CO₂ atmosphere) for 7 days, with a 50% differentiation medium change every other day. Following 7 days, differentiated BET-hNPCs were harvested, centrifuged at 1,000 rpm for 5 min, and finally resuspended in 1 ml of 0.05% DNase/DMEM. Cell density and viability (≥98%) were determined with a hemocytometer using a standard 0.1% trypan blue exclusion test prior to transplantation.

Cell Transplantation

A single cell suspension containing either 800,000 undifferentiated or differentiated BET-hNPCs was injected stereotactically into the striatum. Briefly, rats were anesthetized using an anesthetic mixture (2.0 ml/kg) of 25% ketamine hydrochloride, 6% xylazine, and 2.5% acepromazine maleate in 0.9% saline. Rats were then positioned in a Kopf stereotaxic instrument; a midline scalp incision followed by a burr hole was then performed. A total of 800,000 undifferentiated or differentiated BET-hNPCs (400,000 cells/μl) were then transplanted at a rate of 1 μl/min using a glass capillary (open diameter 50–70 μm) connected to a 2-μl Hamilton syringe that was left in place for 5 min before retrieval. Stereotactic coordinates were at the following coordinates from the dorsal surface from bregma: 1) AP = +0.0 mm, ML = −3.6 mm, DV = −5.2 and −4.3 mm; 2) AP = −0.3 mm, ML = −3.6 mm, DV = −5.2 and −4.3 mm, TB = −3.3 (63). After transplantation, the muscle and skin were closed with polyglactin absorbable sutures (Ethicon). All animals were immunosuppressed by daily intraperitoneal (IP) injection of cyclosporin (10 mg/kg IP; Sandimmune, Sandoz Pharmaceuticals, Princeton, NJ, USA), beginning 2 days prior to cell grafting and continuing until the conclusion of the experiment.

Behavioral Assessment

Numerous behavioral testing paradigms for animal models of HD are described in the literature (16,43,50,79,82,83,86). Among these behavioral tests, this study examined the amphetamine-induced rotational behavior and the cylinder tests, which have been routinely used in our previously published studies for assessment of motor deficits (6,55,57). In addition, the T-maze test was performed to examine spatial working memory.

Amphetamine-Induced Rotational Test

Animals were tested for asymmetrical rotational behavior via the use of an amphetamine challenge at 3.5 weeks following QA lesioning and at 2, 4, 6, 8, and 10 weeks posttransplantation using a computerized video activity monitor (Videomex, Columbus Instruments, Columbus, OH, USA) that record both clockwise- and counterclockwise-turning behavior. Animals were given an amphetamine challenge (5 mg/kg, IP; Sigma Aldrich Co. Ltd., Gillingham, Dorset, UK, Cat# A5880), and rotational behavior was assessed at 10 min intervals over a 90 min period.

Cylinder Test

Animals were tested for forelimb asymmetry (72) via the use of the cylinder test prior to QA lesioning at 3.5 weeks following QA lesioning and at 2, 4, 6, 8, and 10 weeks posttransplantation. The percentage of forelimb placements made on the wall of the plexiglass cylinder (Plastic World and Design, Dartmouth, NS, Canada) was determined during explorative behavior, with unilateral motor system injury following QA lesioning resulting in asymmetrical limb usage. A total of 100 forelimb placements were analyzed for all animals at each time point.

T-Maze (Spatial Working Memory

Animals were tested for spatial working memory via the use of the alternation T-maze test prior to QA lesioning at 3.5 weeks following QA lesioning and at 2, 4, 6, 8, and 10 weeks posttransplantation. Animals were trained in a perspex T-maze (Plastic World and Design) composed of a starting box closed by a sliding door (16.5 × 16.5 cm), a choice arm (70 × 10 cm), and two goal arms perpendicular to the choice arm (42.5 × 10 cm). At the distal end of each goal arm was a plastic food cup containing food reinforcement. Animals were placed in the closed starting box for 5 s and were then allowed to choose either goal arm, after which they were enclosed for a further 10 s. Animals were then placed in the starting box for a further 5 s, and another trial was undertaken. Animals were reinforced only if they entered the goal arm opposite to the one chosen on the previous trial. Animals were trained to perform the alternation T-maze test correctly on 10 consecutive trials with food reinforcement at baseline prior to QA lesioning. For each testing session, the percentage of correct alternations and the number of consecutive choices of the same goal arm (perseveration index) in 10 consecutive trials were analyzed for all animals at each time point.

Tissue Preparation and Immunohistochemistry

Animals were given an IP anesthetic overdose with a mixture (2.0 ml/kg) of 25% ketamine hydrochloride, 6% xylazine, and 2.5% acepromazine maleate in 0.9% saline and perfused transcardially with 250 ml of cold 0.1 M PBS, pH 7.4, followed by 250 ml of 4% PFA in 0.1 M PBS for 15 min. The brains were removed and postfixed overnight in 4% PFA in 0.1 M PBS at 4°C followed by cryoprotection overnight in 30% sucrose (Sigma-Aldrich Canada, Cat# 8411) in 0.1 M PBS at 4°C. Brains were then cut in 40-mm coronal sections on a Leitz freezing microtome (Leica Microsystems, Inc., Concord, ON, Canada) at −30°C and stored in Millonig's solution (16.88 mg/ml NaH₂PO₄H₂O, 3.86 mg/ml NaOH, 0.006% sodium azide in distilled water; Sigma-Aldrich Canada).

Immunohistochemical staining was performed to visualize the cell phenotypes and survival of BET-hNPCs within the graft site. Briefly, 40-μm sections were rinsed [3 × 5 min in 0.1 M phosphate buffer (PB)], quenched for 30 min at room temperature (RT) in 10% MeOH, 3% H₂O₂, and 0.1 M PB (Sigma-Aldrich Canada), and then rinsed in 0.1 M PB. The sections were then preincubated in a diluent of 5% normal horse serum (Dakocytomation Dako North America, Inc., Carpinteria, CA, USA), 0.3% Triton X-100 (Sigma-Aldrich) in 0.1 M PB at RT for 1 h. The sections were then placed in a diluent containing the primary antibody [mouse anti-human neuron-specific enolase (hNSE), 1:1,000, Chemicon] overnight at 4°C. The sections were then rinsed in 0.1 M PB and incubated in a diluent containing the secondary antibody (biotinylated horse anti-mouse IgG, 1:500; Vector Laboratories, Burlingame, CA, USA) for 1 h at RT. The sections were then rinsed in 0.1 M PB and incubated for 1 h with avidin-biotinylated peroxidase complex (1:500; Vector Elite ABC kit, Vector Laboratories) in 0.1 M PB. Sections were then rinsed in 0.1 M PB and underwent product visualization by the addition of 3,3′-diaminobenzidine (DAB: Sigma Aldrich Canada, Cat# 5905; 0.5 mg/ml in 0.1 M PB, 6.6 ml/20 ml of 30% H₂O₂). Sections were then rinsed in 0.1 M PB, mounted on gelatin-coated (Sigma Aldrich Canada, Cat# G-2625) slides (Fisher Scientific, Ottawa, ON, Canada), dehydrated, counterstained with 8% cresyl violet, and coverslipped with Entellan (Merck Chemicals, Billercia, MA, USA).

Immunofluorescence and confocal microscopy were also performed on tissue sections and in vitro cell culture chamber slides. Briefly, 40-mm sections and chamber slides were rinsed in 0.1 M PB, preincubated in a diluent of 8% normal goat serum (Life Technologies, Inc., Burlington, ON, Canada) in 0.3% Triton X-100 in 0.1 M PB at RT for 1 h. Sections and chamber slides were then placed in diluents containing primary antibodies: mouse anti-human-specific nuclear antigen (hNA; 1:2,000, Chemicon), mouse anti-Ki67 (1:500, Pharmingen, BD Biosciences), mouse anti-nestin (1:2,000, BD Pharmingen, BD Biosciences), mouse anti-glial fibrillary acidic protein (GFAP; 1:1,000, NovaCastra, Burlington, ON, Canada), rabbit anti-GFAP (1:1,000, Dako, Carpinteria, CA, USA), mouse anti-β-III-tubulin (1:1,000, Chemicon), mouse anti-GABA (1:1,000, Chemicon), rabbit anti-GABA (1:1,000, Sigma-Aldrich), rabbit anti-GAD (1:1,000, Sigma-Aldrich), rabbit anti-calretinin (1:2,000, Swant, Marly, Switzerland), mouse anti-GABA_A receptor β-subunit (1:10, Boehringer Mannheim Biochemica, Basel, Switzerland), rabbit anti-calbindin-D28k (1:1,000, Chemicon), rabbit anti-substance P (1:1,000, Immunostar, Hudson, WI, USA), rabbit anti-met enkephalin (1:1,000, Immunostar), rabbit anti-leu-enkephalin (1:1,000, Immunostar), rabbit anti-somatostatin (1:200, Incstar/Immunostar), mouse anti-parvalbumin (1:1,000, Sigma-Aldrich), mouse anti-oligodendrocyte marker O4 (1:200; Cedarlane, Burlington, ON, Canada; Cat# MAB345), rabbit anti-D2 dopamine receptor (1:200; Cedarlane; Cat# AB5084), rabbit anti-kainate/2-amino-3-(3-hydroxy-5-methyl-isoxazol-4-yl)propanoic acid (AMPA) glutamate receptor (GluR) 2/3 (1:500; Cedarlane, Cat# AB1506), and rabbit anti-NMDA R1 (postsynaptic glutamate receptor; 1:1,000, Cedarlane, Cat# ABN99) overnight at 4°C. The sections were then rinsed in 0.1 M PB and then incubated with the secondary antibodies goat anti-rabbit, Alexa-488 or 555 (1:500, Molecular Probes, Invitrogen, Eugene, OR, USA) and goat anti-mouse, Alexa-488 or 555 (1:500, Molecular Probes, Invitrogen) for 1 h at RT. In some experiments, DNA counterstain TO-PRO-3 (1 mM, Molecular Probes, Invitrogen) was applied for 15 min at RT. The sections were then mounted on gelatin-coated slides, dried, coverslipped with Vectashield mounting medium (Vector Laboratories), and viewed using confocal microscopy (Zeiss Canada, Toronto, ON, Canada).

The degree of antigen expression was assessed by optical density using a densitometry software program (Scion Image Beta4.03, National Institute of Health, Bethesda, MD, USA). The relative optical density of antigen expression in the transplanted striatum was compared to that of the unlesioned striatum in order to correct for nonspecific background staining.

Stereology

To estimate the graft volume of undifferentiated and differentiated hNPCs, unbiased stereological cell-counting software was used (Stereo Investigator, MicroBrightfield, Williston, VT, USA). Graft volume was assessed using the Cavalieri method on every 12th hNSE/cresyl violet section (Sigma Aldrich Canada, Cat# C5042) using a 10× objective lens. The graft volume was defined as that portion of the striatum containing hNSE cell bodies and fibers.

Statistical Analysis

Statistical analysis was carried out using GraphPad Prism software (Graphpad Software, Inc., San Diego, CA, USA). Variation in asymmetrical rotational behavior using an amphetamine challenge, forelimb asymmetry using the cylinder test, spatial working memory using the T-maze test, and the number of consecutive choices of the same goal arm (perseveration index) between each treatment group, and time posttransplantation was analyzed using two-way analysis of variance. One-way analysis of variance followed by a Tukey's multiple comparisons post hoc test was then used to identify variation in asymmetrical rotational behavior, forelimb asymmetry, spatial working memory, and perseveration at each time point posttransplantation between treatment groups.

Results

Bioreactor System Enables Standardized and Stable Expansion of hNPCs

hNPCs were inoculated as single cells into suspension culture bioreactors in which the hydrodynamic conditions were maintained by controlling the agitation rate at 100 rpm. The single cells formed spherical aggregates with an average size of 460±150 mm (n = 107) at 14 days postinoculation (Fig. 1A). After a short lag phase, the cells proliferated rapidly exhibiting a maximum specific growth rate of 0.009 h⁻¹ (cell population doubling time of approximately 78 h) (Fig. 1B). The average viability over this same period was 93%.

Figure 1.

Culture of human neural progenitor cell (hNPC) neurospheres. (A) Brightfield image of hNPC neurospheres sampled from a suspension bioreactor at 14 days postinoculation (scale bar: 200 μm). (B) Average cell density (×10⁵ cells/ml) and average viability (%) of hNPCs in suspension culture bioreactors. hNPCs were inoculated at 100,000 cells/ml (single cell suspension) into duplicate 125-ml bioreactors, each containing 100 ml of Pharmaceutical Production Research Facility (PPRF-h2) medium agitated at 100 rpm. The suspension bioreactors were incubated at 37°C in a humidified atmosphere containing 5% CO₂ in air. The cultures were fed every 5 days by replacing 40% of the medium with fresh medium.

Differentiated BET-hNPCs Express GABAergic Neuronal Markers

Immunofluorescent and confocal phenotypic analyses of undifferentiated BET-hNPCs (day 0) compared to 7 days of in vitro differentiation (Fig. 2) revealed a statistically significant decrease in the expression of proliferation marker Ki67 (37.04±1.51% vs. 4.66±0.67%, p < 0.001, data not shown) and NPC marker nestin (100±1.81% vs. 11.56±1.81%, p < 0.001) (Fig. 2A, E). In contrast, at day 7 postdifferentiation, there was a significant increase in neuronal marker β-III-tubulin (5.61±1.08% vs. 64.19±2.93%, p < 0.001) (Fig. 2C, G). Furthermore, differentiated BET-hNPCs expressed characteristic markers of striatal medium spiny inhibitory neurons with a significant increase in the expression of GABA (1.32±0.68% vs. 99.10±0.19%, p < 0.001) (Fig. 2D, H), the GABA_A receptor β-subunit (0.00±0.00% vs. 73.90±2.03%, p < 0.001) (Fig. 2L), GAD (0.27±0.2% vs. 99.00±0.35%, p < 0.001) (Fig. 2N), and the intracellular calcium-binding protein calretinin (0.00±0.00% vs. 58.82±2.43%, p < 0.001) (Fig. 2O). However, there was a significant decrease in the expression of the opioid peptide leu-enkephalin (43.90±3.38% vs. 7.28±1.74%, p < 0.001) (Fig. 2P). Confocal imaging unambiguously showed that differentiated BET-hNPCs coexpressed GABA and β-III-tubulin (Fig. 2H–J) as well as GABA and GABA_A receptor β-subunit (Fig. 2K–M), respectively. Interestingly, there was no alteration in expression of the astrocytic marker GFAP (1.03±0.38% vs. 1.30±0.31%, p > 0.05) (Fig. 2B, F).

Both undifferentiated and differentiated BET-hNPCs did not express the oligodendrocyte marker O4, the intracellular calcium-binding protein calbindin-D28k, the opioid peptide met-enkephalin, somatostatin, parvalbumin, substance P, the postsynaptic D2 dopamine receptor, and the postsynaptic glutamate receptor (data not shown).

Taken together, these results indicate that differentiation downregulates expression of precursor cell-associated markers and leads to the majority of BET-hNPCs expressing GABAergic neuronal phenotypic markers.

Figure 2.

Immunofluorescent and confocal images of undifferentiated BET-hNPCs. Undifferentiated bioreactor-expanded telencephalic (BET)-hNPCs expressed human nuclear antigen (hNA; red) and colabeled with (A) nestin (green), (B) glial fibrillary acidic protein (GFAP; green), (C) β-III-tubulin (green), and (D) γ-aminobutryic acid (GABA; green) (scale bar: 50 μm). Differentiated BET-hNPCs expressed hNA (red) and colabeled with (E) nestin (green), (F) GFAP (green), and (G) β-III-tubulin (green) (scale bar: 50 μm). Differentiated BET-hNPCs coexpressed (H) GABA (green) and (I) β-III-tubulin (red); (J) merged image (scale bar: 50 μm). They coexpressed (K) GABA (green) and (L) GABA_A receptor β-subunit (red); (M) merged image (scale bar: 20 μm). They also expressed other neuronal markers such as (N) glutamic acid decarboxylase (GAD; green) (scale bar: 50 μm), (O) calretinin (green) (scale bar: 20 μm), and (P) leu-enkephalin (green) (scale bar: 20 μm) that colabeled with hNA (red).

Transplantation of BET-hNPCs Promotes Functional Recovery

Following QA lesion, the animals were equally divided into three experimental groups: the animals with no cells (Group 1), the animals with 800,000 undifferentiated BET-hNPCs (Group 2), and the animals with 800,000 differentiated BET-hNPCs (Group 3).

Cylinder Test

Striatal QA lesioning resulted in a significant increase in forelimb asymmetry and right forelimb preference in Group 1 (63.76±3.18% vs. 51.20±0.44%, p < 0.01), Group 2 (65.70±2.58% vs. 49.20±0.74%, p < 0.01), and Group 3 (63.18±0.99% vs. 50.55±0.55%, p < 0.05) at week 0 posttransplantation when compared to baseline prelesion forelimb asymmetry (Fig. 3A–C).

There was a significant decrease in right forelimb asymmetry in Group 3 compared with Group 2 at 4 weeks (58.64±2.53% vs. 69.90±3.51%, p < 0.05), 6 weeks (57.09±2.26% vs. 67.78±3.58%, p < 0.05), 8 weeks (54.00±4.71% vs. 68.22±3.89%, p < 0.05), and 10 weeks posttransplantation (55.27±3.59% vs. 69.44±3.22%, p < 0.01) (Fig. 3C, #). However, there was no statistical difference in right forelimb preference between Group 1 and Group 3 at all time points, presumably due to the moderate percentage improvements in forelimb asymmetry, relatively large standard errors, and sample size. Groups 1 and 2 showed no significant difference in right forelimb asymmetry at all time points (p > 0.05).

Although there was a significant increase in forelimb asymmetry and right forelimb preference in Group 3 following QA lesioning (50.55±0.55% vs. 63.18±0.99%, p < 0.05) and at 2 weeks posttransplantation (50.55±0.55% vs. 64.27±2.19%, p < 0.05), the gradual recovery from right forelimb asymmetry was evidenced by no statistical difference in forelimb asymmetry at 4 weeks (50.55±0.55% vs. 58.64±2.53%, p > 0.05), 6 weeks (50.55±0.55% vs. 57.09±2.26%, p > 0.05), 8 weeks (50.55±0.55% vs. 54.00±4.71%, p > 0.05), and 10 weeks posttransplantation (50.55±0.55% vs. 55.27±3.59%, p > 0.05) when compared to baseline prelesion forelimb asymmetry (Fig. 3C, ϕ). In contrast, no significant difference was observed in fore-limb asymmetry in Groups 1 and 2 when compared to the lesioned pretransplantation rotational behavior at all time points (p > 0.05).

Figure 3.

Right forelimb asymmetry after quinolinic acid (QA) lesioning and BET-hNPC transplantation into the striatum at 2, 4, 6, 8, and 10 weeks posttransplantation. (A) Representative image of a QA-lesioned animal displaying forelimb asymmetry and right forelimb preference (circle). (B) Image of a QA-lesioned animal following transplantation of differentiated BET-hNPCs (Group 3) at 10 weeks displaying no forelimb asymmetry and no right forelimb preference (circles). (C) There was a significant decrease in right forelimb preference following transplantation of differentiated BET-hNPCs (Group 3: blue symbol) at 4, 6, 8, and 10 weeks when compared to undifferentiated BET-hNPCs (Group 2 : red symbol) (^#p < 0.05, ^##p < 0.001) (mean±SEM). Importantly, no significant difference was found in forelimb asymmetry following transplantation of differentiated BET-hNPCs (Group 3: blue symbol) at 4, 6, 8, and 10 weeks posttransplantation when compared to the prelesion forelimb asymmetry baseline (^ϕp > 0.05) (mean±SEM). (D) Asymmetrical rotational behavior following an amphetamine challenge (5 mg/ml) after QA lesioning and BET-hNPC transplantation into the striatum at 2, 4, 6, 8, and 10 weeks posttransplantation. There was a significant decrease in asymmetrical rotational behavior following transplantation of differentiated BET-hNPCs (Group 3: blue symbol) at 2, 4, 6, 8, and 10 weeks posttransplantation when compared to control animals (Group 1: black symbol) (**p < 0.01, ***p < 0.001), as well as animals transplanted with undifferentiated BET-hNPCs (Group 2 : red symbol) (^##p < 0.01, ^###p < 0.001) (mean±SEM). There was a significant decrease in rotational behavior following transplantation of differentiated BET-hNPCs (Group 3: blue symbol) at 2, 4, 6, 8, and 10 weeks when compared to baseline lesioned pretransplantation rotational behavior (^ϕϕϕp < 0.001) (mean±SEM). (E) Cognitive assessment using T-maze after QA lesioning and BET-hNPC transplantation into the striatum at 2, 4, 6, 8, and 10 weeks posttransplantation. There was a significant increase in alternation behavior following transplantation of differentiated BET-hNPCs (Group 3: blue symbol) at 2, 6, 8, and 10 weeks posttransplantation when compared to control animals (Group 1: black symbol) (*p < 0.05, **p < 0.01, ***p < 0.001), as well as at 2 weeks posttransplantation when compared to animals transplanted with undifferentiated BET-hNPCs (Group 2 : red symbol) (^#p < 0.05) (mean±SEM). Importantly, a significant increase was observed in alternation behavior following transplantation of differentiated BET-hNPCs (Group 3: blue symbol) at 6, 8, and 10 weeks when compared to the lesioned pretransplantation rotational behavior (^ϕp < 0.05, ^ϕϕϕp < 0.001) (mean±SEM). (F) There was a significant decrease in perseveration following transplantation of differentiated BET-hNPCs (Group 3: blue symbol) when compared to control animals (Group 1: black symbol) (*p < 0.05) or animals transplanted with undifferentiated BET-hNPCs (Group 2 : red symbol) (^#p < 0.05) (mean±SEM) at 2 weeks posttransplantation.

Amphetamine-Induced Rotational Test

There was a significant decrease in amphetamine-induced rotational behavior in Group 3 at 2 weeks (3.93±1.06 vs. 12.71±1.41 rotations/min, p < 0.001), 4 weeks (3.12±1.12 vs. 13.09±1.99 rotations/min, p < 0.001), 6 weeks (3.97±1.47 vs. 13.12±1.89 rotations/min, p < 0.01), 8 weeks (2.91±1.49 vs. 13.77±1.82 rotations/min, p < 0.01), and 10 weeks posttransplantation (3.81±1.86 vs. 12.62±1.33 rotations/min, p < 0.01) when compared to Group 1 (Fig. 3D, *). Furthermore, there was a significant decrease in rotational behavior in Group 3 at 2 weeks (3.93±1.06 vs. 14.04±1.35 rotations/min, p < 0.001), 4 weeks (3.12±1.12 vs. 13.91±2.00 rotations/min, p < 0.001), 6 weeks (3.97±1.47 vs. 14.12±2.42 rotations/min, p < 0.01), 8 weeks (2.91±1.49 vs. 14.47±2.38 rotations/min, p < 0.001), and 10 weeks posttransplantation (3.81±1.86 vs. 13.56±2.51 rotations/min, p < 0.01) when compared to Group 2 (Fig. 3D, #). No significant difference was found in rotational behavior between Groups 1 and 2 at all time points (p > 0.05).

Normalization of rotational behavior was evidenced by a significant decrease in rotational behavior in Group 3 at 2 weeks (3.93±1.06 rotations/min, p < 0.001), 4 weeks (3.12±1.12 rotations/min, p < 0.001), 6 weeks (3.97±1.47 rotations/min, p < 0.001), 8 weeks (2.91±1.49 rotations/min, p < 0.001), and 10 weeks posttransplantation (3.81±1.86 rotations/min, p < 0.001) when compared to baseline lesioned pretransplantation rotational behavior (13.96±1.46 rotations/min) (Fig. 3D, f). In contrast, no significant difference was observed in rotational behavior in Groups 1 and 2 when compared to the lesioned pretransplantation rotational behavior at all time points (p > 0.05).

T-Maze Test

A significant decrease in alternation behavior was observed following QA lesioning in all groups when compared to baseline prelesion alternation behavior: Group 1 (10.00±0.0 vs. 6.3±0.24 alternations, p < 0.001), Group 2 (10.00±0.0 vs. 6.15±0.31 alternations, p < 0.001), and Group 3 (10.00±0.0 vs. 6.05±0.24 alternations, p < 0.001) (Fig. 3E, at week 0 posttransplantation). There was a significant increase in alternation behavior in Group 3 when compared to Group 1 at 2 weeks (7.09±0.29 vs. 5.95±0.28 alternations, p < 0.05), 6 weeks (7.55±0.26 vs. 6.10±0.26 alternations, p < 0.01), 8 weeks (8.09±0.36 vs. 6.00±0.21 alternations, p < 0.001), and 10 weeks posttransplantation (8.09±0.33 vs. 6.15±0.42 alternations, p < 0.01) (Fig. 3E, *). Furthermore, alternation behavior in Group 3 was significantly increased when compared to Group 2 at 2 weeks posttransplantation (7.09±0.29 vs. 5.90±0.29 alternations, p < 0.05) (Fig. 3E, #). No significant difference was detected in alternation behavior between Groups 1 and 2 at all time points (p > 0.05).

Importantly, gradual improvement in alternation behavior was evidenced by a significant increase in correct alternation behavior in Group 3 at 6 weeks (7.55±0.26 alternations, p < 0.05), 8 weeks (8.09±0.36 alternations, p < 0.001), and 10 weeks posttransplantation (8.09±0.33 alternations, p < 0.001) when compared to the lesioned pretransplantation alternation behavior (Fig. 3E, f). There was no significant difference in alternation behavior in Groups 1 and 2 when compared to baseline lesioned pretransplantation alternation behavior at all time points (p > 0.05).

There was a significant increase in perseveration following QA lesioning in Group 1 (0.0±0.0 vs. 4.50±0.67, p < 0.001), Group 2 (0.0±0.0 vs. 3.85±0.37, p < 0.001), and Group 3 (0.0±0.0 vs. 5.41±0.67, p < 0.001) when compared to the prelesion perseveration (Fig. 3F). There was a significant decrease in perseveration in Group 3 when compared to Group 1 (3.73±0.36 vs. 6.00±0.72, p < 0.05) and Group 2 (3.73±0.36 vs. 5.75±0.64, p < 0.05) at 2 weeks posttransplantation (Fig. 3F, * and #, respectively). There was no significant difference in perseveration between Groups 1 and 2 at all time points (p > 0.05).

Collectively, results obtained from the behavioral assessment support our hypothesis that differentiation of BET-hNPCs to a GABAergic phenotype prior to transplantation is a prerequisite for behavioral recovery in the QA lesion model of HD.

Differentiated BET-hNPCs Form Striatal-Like Neuron Clusters

Immunohistochemical analysis revealed robust survival of undifferentiated and differentiated BET-hNPCs in all animals with extensive staining for hNSE in cell bodies and fibers. hNSE staining was not observed in control animals without transplantation of BET-hNPCs (Fig. 4A–D). hNSE-positive undifferentiated BET-hNPCs (Group 2) were widely distributed throughout the rostrocaudal axis of the striatum or observed to migrate within the corpus callosum (Fig. 4B). In contrast, differentiated BET-hNPCs (Group 3) were not as widely distributed throughout the rostrocaudal axis of the striatum, did not migrate within the corpus callosum, and had dense clusters of cells located at the primary transplantation site (Fig. 4C, D).

Stereological analysis of the graft volume was defined as that portion of the striatum containing hNSE-positive bodies and fibers. There was a significant increase in the graft volume in Group 2 when compared to Group 3 (3.01±0.18 mm³ vs. 1.94±0.19 mm³, p < 0.001).

Figure 4.

Brightfield images showing coronal sections stained with human neuron-specific enolase (hNSE) and cresyl violet at 10 weeks posttransplantation. Brightfield images showing coronal sections of (A) control (Group 1), (B) undifferentiated (Group 2), and (C) differentiated (Group 3) BET-hNPC grafts stained with hNSE and cresyl violet at 10 weeks posttransplantation (scale bar: 500 μm). (D) High-power image of a differentiated BET-hNPC graft (scale bar: 100 μm).

Differentiated BET-hNPCs Retain GABAergic Phenotype and Undergo Maturation Following Transplantation

Immunofluorescent and confocal analysis showed that the majority of differentiated BET-hNPCs (Group 3) at 10 weeks posttransplantation coexpressed hNA and β-III-tubulin (Fig. 5A–D), GAD (Fig. 5E–H), GABA (Fig. 5I–L), or calretinin (Fig. 5M–P), which indicates that they retained the GABAergic cell phenotype after transplantation. Some populations of differentiated BET-hNPCs were reminiscent of striatal neurons expressing the GABA_A receptor β-subunit (Fig. 6A–D), the intracellular protein substance P (Fig. 6E–H), and the presynaptic marker synaptophysin (Fig. 6I–L). Interestingly, differentiated BET-hNPC grafts appeared to be innervated by anti-kainate/2-amino-3-(3-hydroxy-5-methyl-isoxazol-4-yl)propanoic acid (AMPA) glutamate receptor (GluR) 2/3-positive fibers (Fig. 6M–P) and expressed D2 dopamine receptors (Fig. 6Q–T). These results suggest that differentiated BET-hNPCs undergo further maturation and integrate within the host environment following transplantation.

Figure 5.

Immunofluorescent and confocal images of differentiated BET-hNPCs (Group 3) grafted into the QA-lesioned striatum at 10 weeks posttransplantation. A majority of differentiated BET-hNPCs maintained a neuronal phenotype that expressed both (A) hNA (red) and (B) β-III-tubulin (green); (C) merged image (scale bar: 50 μm); (D) orthogonal image. Differentiated BET-hNPCs maintained other neuronal markers expressed by striatal neurons: (E) hNA (red) colabeled with (F) GAD (green); (G) merged image (scale bar: 50 μm); (H) orthogonal image. (I) hNA (red) colabeled with (J) GABA (green); (K) merged image (scale bar: 50 μm); (L) orthogonal image. (M) hNA (red) colabeled with (N) calretinin (green); (O) merged image (scale bar: 50 μm); (P) orthogonal image. Note that hNA immunoreactivity (red) was completely surrounded by (D) β-III-tubulin (green), (H) GAD (green), (L) GABA (green), and (P) calretinin (green) unambiguously demonstrating the immunoreactivity of transplanted BET-hNPCs rather than the host cells (scale bar: 50 μm).

Figure 6.

Immunofluorescent and confocal images of differentiated BET-hNPCs (Group 3) grafted into the QA-lesioned striatum at 10 weeks posttransplantation. Differentiated BET-hNPCs express striatal neuronal markers colabeling with (A) GABA (green) and (B) GABA_A receptor b subunit (red); (C) merged image (scale bar: 20 μm); (D) orthogonal image. They also colabel with (E) substance P (green) and (F) hNSE (red); (G) merged image (scale bar: 10 μm); (H) orthogonal image. Furthermore, differentiated BET-hNPCs express pre- and postsynaptic markers: (I) DNA-bound nuclear counterstain TO-PRO-3 (blue), (J) hNSE (green), and (K) synaptophysin (red); (L) merged image (scale bar: 20 μm); (M) TO-PRO-3 (blue), (N) hNSE (red), (O) anti-kainate/2-amino-3-(3-hydroxy-5-methyl-isoxazol-4-yl)propanoic acid (AMPA) glutamate receptor 2/3 (GluR 2/3; green); (P) merged image (scale bar: 50 μm); (Q) hNSE (red) and (R) dopamine D2 receptor (green); (S) merged image (scale bar: 20 μm); (T) orthogonal image.

Differentiated BET-hNPCs (Group 3) showed no evidence of cellular proliferation or tumor formation as evidenced by the lack of colabeling of Ki67 (Fig. 7A) with TO-PRO-3 and calretinin (Fig. 7B), while cells localized in the subventricular zone within the same section were positive for Ki67 (Fig. 7C). Furthermore, they displayed no expression of the astrocytic marker GFAP (Fig. 7D); however, sparse numbers expressed nestin (Fig. 7E). In contrast, undifferentiated BET-hNPCs (Group 2) were viable at 10 weeks posttransplantation, the majority of which differentiated into GFAP-expressing astrocytes that colabeled with hNA (Fig. 7F). Undifferentiated BET-hNPCs were not observed to express β-III-tubulin, GABA, or GAD and showed no evidence of cellular proliferation or tumor formation as assessed using Ki67 (data not shown).

Figure 7.

Immunofluorescent and confocal images of differentiated and undifferentiated BET-hNPCs grafted into the QA-lesioned striatum at 10 weeks posttransplantation. Differentiated BET-hNPCs did not express (A) cell cycle-associated protein Ki67, but colabeled with (B) DNA-bound nuclear counterstain TO-PRO-3 (blue) and calretinin (green) (scale bar: 20 μm). (C) Positive control for Ki67 (red) within the subventricular zone of the same animal (scale bar: 50 μm). (D) Orthogonal image shows that differentiated BET-hNPCs labeled for hNA (red) but did not express the astrocytic marker GFAP (green). (E) Orthogonal image indicates that sparse numbers of differentiated BET-hNPCs colabeled for hNA (red) and nestin (green). (F) Orthogonal image reveals that a majority of grafted undifferentiated BET-hNPCs differentiated into GFAP-expressing astrocytes (green) that colabeled with hNA (red).

Discussion

For neural cell transplantation to become a routine therapeutic option for the treatment of neurodegenerative diseases such as HD, there is a need to develop clinically acceptable methods of producing large quantities of neural cells in a standardized fashion. This study describes how a small number of hNPCs derived from a single procurement of fetal telencephalic tissue was serially expanded in suspension bioreactors and then differentiated into a highly enriched population of GABAergic neurons (>99%) that promoted significant motor and cognitive improvement following transplantation in a rodent model of HD.

Large-Scale Expansion of hNPCs and Efficient In Vitro Generation of GABAergic Neurons

The expansion of mammalian cell populations in suspension bioreactors holds numerous advantages over standard static culture systems. Bioreactors are much more amenable to efficient scale-up, provide a means to exert a high level of control over the culture environment, and exhibit a greater degree of reproducibility between cell production runs. In the current study, neurospheres were rapidly generated from hNPCs within suspension bioreactors under controlled conditions that maintained stable oxygen, nutrient, osmolality, pH, and culture hydrodynamics at levels that promoted optimal cell growth (45,74–78). Previously, our group has shown that bioreactor-expanded rodent NPCs can form neurospheres and exhibit large expansion ratios (10⁷ in 35 days) without loss of cell multipotency (56) or defining characteristics as determined by karyotype analysis and immunohistochemistry (45,74). We have also shown that suspension bioreactors can be used to rapidly expand hNPCs (a 36-fold expansion over a course of 18 days) while maintaining high viabilities (>90%) and neurogenic potential (4,5). Moreover, serial subculturing of these cells in standard suspension bioreactors resulted in an overall cell-fold expansion of 7.8 × 10¹³ after 140 days while maintaining nestin expression, multipotency, and stable chromosome number (5). This large cell expansion ratio and multipotentiality were reflected in the current study, where the stem cell marker nestin and the cell cycle-associated protein Ki67 were expressed in the majority of undifferentiated BET-hNPCs. Consequently, bioreactor-based expansion may be a feasible technique for the clinical-grade production of hNPCs.

In vitro differentiation of BET-hNPCs revealed a highly enriched population of GABAergic neurons that expressed β-III-tubulin (64.19±2.93%), GABA (99.10±0.19%), GAD (99.00±0.35%), calretinin (58.82±2.43%), and the GABA_A receptor β-subunit (73.90±2.03%). The expression of GAD in differentiated BET-hNPCs also excludes the possibility that GABA immunoreactivity was a result of in vitro uptake. The significant decrease in nestin and Ki67 following in vitro differentiation suggests that BET-hNPCs underwent neuronal phenotypic maturation (46). This result corroborates the decreases in phenotypic expression of nestin and Ki67 following differentiation of mouse telencephalic NPCs (46).

In contrast to the present study, other studies have shown that human and rodent NPCs differentiate into GABAergic cell phenotypes in vitro (10,44), albeit in low quantities when subjected to extended serial passaging (44). Specifically, these studies showed a decrease in the expression of β-III-tubulin and GABA, with a corresponding increase in GFAP expression at later passages (44), which is consistent with neuronal cell generation from hNPCs from the spinal cord (65). Furthermore, hNPCs did not express GABA after 7 days in culture (44).

Differentiated BET-hNPCs coexpressed GABA, GAD, calretinin, β-III-tubulin, enkephalin, and the GABA_A receptor-β subunit, which are typically expressed by striatal GABAergic neurons. Immunohistochemical studies have shown various populations of GABAergic medium-sized spiny and aspiny neurons within the neostriatum that coexpress calretinin with both GAD₆₇ and GABA (9,19,48,68). Specifically, calretinin and GABA colabel in approximately 74% of cells and colocalize with calbindin-D28k in approximately 4% of cells. Another study indicated that calretinin is not coexpressed with calbindin-D28k, parvalbumin, nicotinamide adenine dinucleotide phosphate-diaphorase, or choline acetyltransferase in the human striatum, indicating that calretinin-expressing neurons are a distinct class of neuron in the striatum (19). Consistent with those studies, the current study indicated that calretinin was expressed in approximately 58.82±2.43% of differentiated BET-hNPCs and that these cells did not coexpress calbindin-D28k, somatostatin, or parvalbumin.

Transplantation of BET-hNPC-Derived GABAergic Cells Promotes Functional Recovery in Huntingtonian Rats

The behavioral studies that were carried out in this study support our hypothesis that differentiation of BET-hNPCs to a GABAergic neuronal lineage prior to transplantation is critical to restore behavioral deficits in a rodent model of HD. This is highlighted by the fact that transplantation of undifferentiated BET-hNPCs did not improve any behavioral deficit in the current study.

The current study utilized the cylinder test to investigate forelimb asymmetry and right forelimb preference. There was a significant increase in forelimb asymmetry and right forelimb preference following QA lesioning. Nonetheless, this forelimb asymmetry and right forelimb preference was attenuated by 4 weeks and continued improving to full recovery thereafter following the transplantation of BET-hNPC-derived GABAergic cells. In support, transplantation of ciliary neurotrophic factor (CNTF)-treated hNPCs into the QA-lesioned striatum has previously been shown to partially improve motor behavior in the cylinder test (54). However, considering that the study indicated only 1.0±0.29% of the hNA-positive hNPCs coexpressed the neuronal marker NeuN, and no colabeling was found with GAD or DARPP-32, this suggests that behavioral improvement was caused by a mechanism other than restoration of circuit connectivity. Furthermore, as the study did not examine apomorphine- or amphetamine-induced rotational asymmetry, a comparison of the nigrostriatal benefit of these cells cannot be determined (54). In addition, fetal striatal tissue grafting has also been shown to substantially improve deficits in the skilled forelimb paw-reaching test (26,59), while another study utilizing the immortalized hNT cell line showed no functional improvement in the paw reaching and cylinder test when transplanted into a rodent model of HD (34).

Amphetamine-induced rotational asymmetry following unilateral striatal QA lesioning was substantially ameliorated by an average of 74.59±1.58% following transplantation of GABAergic cells derived from BET-hNPCs. Corroborating this finding, fetal striatal grafts significantly attenuate and normalize motor asymmetry in response to either apomorphine or amphetamine (26,41,49,60,61,64,70), with failure to affect apomorphineinduced asymmetrical behavior when grafts exhibit low levels of dopamine D1 and D2 receptors (41). This suggests that nigrostriatal dopaminergic input and postsynaptic receptors are critical for functional behavioral recovery. As the differentiated BET-hNPC grafts express synaptophysin and the dopamine D2 receptor, the host nigrostriatal dopaminergic system may be reinnervating the graft and making functional synaptic contacts. Further study utilizing anterograde and retrograde labeling will substantiate the connectivity of the graft–host neuronal system.

Spatial working memory as analyzed utilizing the T-maze test indicated a significant decrease in alternation behavior following QA striatal lesioning, demonstrating the critical role of the striatum in cognitive function (51,88). QA lesioning has been shown to lead to deficient retrieval of stored memories of visuospatial skills and impaired transfer of learning, consistent with striatal degeneration accounting for the cognitive abnormalities associated with HD (30). Following transplantation of BET-hNPC-derived GABAergic cells, there was a significant increase in alternation behavior compared to control and baseline lesioned pretransplantation alternation behavior, while undifferentiated BET-hNPCs did not result in significant improvement. Supporting our findings, fetal striatal grafts have been shown to significantly correct delayed alternation behavior to approximately 90%, whereas lesioned animals remain at approximately 50% (43), with repeated learning tasks improving deficits to near control values (11,53,62). Considering that BET-hNPC-derived GABAergic cells significantly improved alternation behavior in the current study as the time course progressed, a repeated learning task may be beneficial for graft maturation and functional recovery in a complex cognitive task. As such, the reestablishment of the appropriate afferent and efferent connections between the host brain and the graft may be an important factor, with environmental enrichment learning tasks a contributing factor to the plasticity and functional recovery of neural grafts (24,25). Clinical trials have also shown that transplantation of human fetal striatal neural cells is associated with recovery of striatofrontal cortical brain metabolism and improvements in cognitive and motor function (3,32).

It is noteworthy that the differentiated BET-hNPC graft site appeared to be populated with a GluR 2/3-positive fiber network. Whether this is indicative of possible reinnervation and reestablishment of the corticostriatal system, which is reported to play a critical role in learning (28,51), may only be determined through future tracer studies. Nonetheless, a study by Cicchetti and colleagues (20) has shown that glutamatergic cortical neurons project to transplanted striatal neurons following neural transplantation of striatal anlagen into the human striatum in a clinical trial of HD.

Immunohistochemical analyses revealed robust survival of BET-hNPCs following transplantation. Although there was an increase in the graft volume of undifferentiated BET-hNPCs compared with BET-hNPC-derived GABAergic cells, this was not due to cellular proliferation, as hNA did not colabel with Ki67. Instead, the majority of undifferentiated BET-hNPCs underwent in vivo differentiation into GFAP-positive astrocytes and were widely distributed throughout the rostrocaudal axis of the host striatum, and migrated within the corpus callosum. This cellular migration is likely accountable for the graft volume increase. This observation is in keeping with the report by McBride and colleagues that undifferentiated hNPCs show migration throughout the host brain upon transplantation in a rodent model of HD (54). Others also have found that undifferentiated NPCs undergo in vivo differentiation into primarily astrocytes when transplanted into nonneurogenic regions of the adult CNS (14,18,33,56,81). It has been suggested that glial restrictive cues for NPCs may predominate over neuronal differentiation in adult CNS (13).

In contrast, differentiated BET-hNPCs had dense clusters of cells located at the primary transplantation site and was devoid of apparent migration outside the target boundaries. Furthermore, differentiated BET-hNPCs were observed to form structures surrounding the primary graft site that were reminiscent of striatal-like neuron clusters observed following transplantation of cells from the embryonic ganglionic eminence or human striatal neuroblasts (35,37). Taken together, these findings imply that cellular migration per se is not associated with functional recovery. But rather, the integration of the transplanted BET-hNPC-derived GABAergic cells into the host structural architecture may be significant for restoration of damaged brain circuitry.

Analysis of differentiated BET-hNPC grafts revealed robust expression of GABA, GAD, calretinin, β-III-tubulin, and the GABA_A receptor-β subunit in the cell bodies or extensive fiber networks, indicating that phenotypes acquired during in vitro differentiation were maintained or upregulated posttransplantation. Differentiated BET-hNPCs also showed no evidence of GFAP or Ki67 expression, suggesting that these cells were not undergoing dedifferentiation following transplantation. In fact, populations of differentiated BET-hNPCs also expressed substance P and the synaptic marker synaptophysin, which were not observed following in vitro differentiation. These observations provide additional evidence that differentiated BET-hNPCs undergo maturation to a phenotype appropriate for the host brain region following transplantation.

In support, hNT cells significantly upregulate calretinin expression upon transplantation into the striatum (71). Nonetheless, further immunohistochemical studies will definitively determine if predifferentiated BET-hNPCs express striatal medium spiny neuron markers such as chicken ovalbumin upstream promoter 1 transcription factor (COUP-TF)-interacting protein 2 (CTIP2), forkhead box P1 (FOXP1), and DARPP-32 following transplantation. Although it is noteworthy, but beyond the scope of this study, the differentiated BET-hNPC graft site appears to be populated with a GluR 2/3-positive fiber network in addition to dopaminergic fibers while expressing postsynaptic dopamine D2 receptors. In order to achieve the cognitive and motor functional recovery discussed above, the graft may have formed appropriate synaptic contacts with the host brain via GABAergic interneurons, nigrostriatal dopaminergic, and corticoidstriatal glutamatergic systems. Future studies involving tracing techniques and receptor immunochemistry may confirm this hypothesis.

On a final note, clinical and preclinical studies not only suggest that there may be preferential need for interneurons in HD transplants but also that interneuron may be preferentially protected from disease-like neuronal degeneration. Specifically, a study by Cicchetti and colleagues (20) indicated that following neural transplantation of striatal anlagen into the human striatum, which ultimately demonstrated marginal and transient clinical benefits, grafts underwent disease-like neuronal degeneration with a preferential loss of projection neurons in comparison to interneurons (20,21). In addition to clinical trials, preclinical studies also indicate that striatal interneurons are vital for normal striatal graft development, integration, and functional recovery in a rodent model of HD (86). Consequently, future long-term studies will have to be undertaken to determine whether differentiated BET-hNPCs are susceptible to disease-like neuronal degeneration akin to the projection neurons described by Cicchetti and colleagues (20) or whether they retain dendritic arborization and cellular integration as well as continue to promote functional recovery in a QA model of HD.

Conclusion

This study provides evidence that small numbers of primary telencephalic hNPCs can be rapidly expanded to clinical quantities through multiple passages in a standardized suspension bioreactor system and yield cell populations that may have utility in new cell replacement therapies aimed at treating neurodegenerative conditions such as HD. Differentiated BET-hNPCs displayed no change in their potentiality after extended passaging and expressed immunohistochemical markers that resemble calretinin-expressing GABAergic neurons of the striatum. Furthermore, differentiated BET-hNPCs promoted functional recovery in the asymmetrical rotational behavior, forelimb asymmetry, and T-maze alternation tests, suggesting a restoration of information relays in corticobasal ganglia circuits. Further transplantation studies of differentiated BET-hNPCs will be needed in nonhuman primate models of HD to assess complex motor and cognitive function prior to clinical cell replacement therapy.