Survival and Differentiation of Human Embryonic Stem Cell-Derived Neural Precursors Grafted Spinally in Spinal Ischemia-Injured Rats or in Naive Immunosuppressed Minipigs: A Qualitative and Quantitative Study

Abstract

In previous studies, we have demonstrated that spinal grafting of human or rat fetal spinal neural precursors leads to amelioration of spasticity and improvement in ambulatory function in rats with spinal ischemic injury. In the current study, we characterize the survival and maturation of three different human embryonic stem (ES) cell line-derived neural precursors (hNPCs) once grafted into ischemia-injured lumbar spinal cord in rats or in naive immunosuppressed minipigs. Proliferating HUES-2, HUES-7, or HUES-9 colonies were induced to form embryoid bodies. During the nestin-positive stage, the rosettes were removed and CD184⁺/CD271^-/CD44^-/CD24⁺ population of ES-hNPCs FAC-sorted and expanded. Male Sprague–Dawley rats with spinal ischemic injury or naive immunosuppressed Gottingen–Minnesota minipigs received 10 bilateral injections of ES-NPCs into the L2–L5 gray matter. After cell grafting, animals survived for 2 weeks to 4.5 months, and the presence of grafted cells was confirmed after staining spinal cord sections with a combination of human-specific (hNUMA, HO14, hNSE, hSYN) or nonspecific (DCX, MAP2, CHAT, GFAP, APC) antibodies. In the majority of grafted animals, hNUMA-positive grafted cells were identified. At 2–4 weeks after grafting, double-labeled hNUMA/ DCX-immunoreactive neurons were seen with extensive DCX⁺ processes. At survival intervals of 4–8 weeks, hNSE⁺ neurons and expression of hSYN was identified. Some hSYN-positive terminals formed putative synapses with the host neurons. Quantitative analysis of hNUMA⁺ cells at 2 months after grafting showed comparable cell survival for all three cell lines. In the presence of low-level immunosuppression, no grafted cell survival was seen at 4.5 months after grafting. Spinal grafting of proliferating pluripotent HUES-7 cells led to consistent teratoma formation at 2–6 weeks after cell transplantation. These data show that ES-derived, FAC-sorted NPCs can represent an effective source of human NPCs to be used in CNS cell replacement therapies.

Introduction

Transient spinal cord ischemia is a serious complication associated with aortic cross clamping (i.e., surgical procedure to replace aortic aneurysm). The resulting neurological deficit is typically permanent and is clinically presented as paraparesis or fully developed spastic or flaccid paraplegia (9, 26, 31). Consistent with the clinical picture of spinal ischemic injury, experimental animal studies show similar functional deficit. Using cat, dog, rabbit, or rat spinal ischemia models, it was demonstrated that, after episodes of injurious intervals of spinal ischemia, there is progressive loss/appearance of (i) ambulatory motor function and (ii) increased muscle tone (rigidity) and stretch reflex activity (spasticity) (7, 10, 22, 24, 30, 40). Histopathological analysis of ischemia-injured spinal cord segments show that, in animals with spasticity, there is selective loss of small inhibitory γ-aminobutyric acid (GABA)ergic neurons while in animals with flaccidity there is widespread neuronal degeneration, including the small interneuronal population as well as α-motoneurons (10, 19, 40). Consistent with the functional role played by injured segmental inhibitory neurons, electrophysiological analysis of spinal segmental reflex activity in animals with spasticity shows (i) exacerbated electromyography (EMG) response and increase in ankle resistance during computer-controlled ankle dorsiflexion (i.e., spasticity) (20), (ii) increased Hoffman's (H) reflex activity and loss of rate-dependent depression (15, 22), and (iii) appearance of rigidity (continuous increase in muscle tone) (10, 20). In accordance with the hypothesized role of the lost GABAergic system in the evolution of local segmental hyperreflexia, pharmacological studies have demonstrated in animals with spinal ischemic injury effective suppression of spasticity and rigidity after spinal delivery of baclofen (GABA B receptor agonist), tiagabine (GABA uptake inhibitor) (15), or 2-amino-3-(5-methyl-3-oxo-1,2- oxazol-4-yl)propanoic acid (AMPA) receptor antagonist NGX424 (13).

Several experimental studies have demonstrated effective neurogenesis in brain in adult naive rodents or after a variety of pathological stimuli including brain ischemia (2, 16–18). In some studies, the magnitude of neuronal repopulation after experimental stroke has been shown to correlate with the degree of neurological recovery (4, 32). In contrast to brain, however, no neurogenesis in naive adult spinal cord has been demonstrated (8, 11). At present, there is no alternative pharmacological/neurotrophic treatment that would lead to mobilization of endogenous spinal neural precursors in an adult-injured spinal cord and would lead to effective neuronal repopulation. Accordingly, segment-targeted cell replacement represents the only available therapeutic alternative at present to achieve replacement of spinal ischemia or spinal trauma-evoked neuronal loss.

In our previous studies, we have demonstrated that spinal grafting of human NT neurons [derived from the NTera2 (NT2) teratocarcinoma cell line] or rat or human neural precursors into previously ischemia-injured spinal cord segments leads to a progressive amelioration of spasticity and recovery of motor function (6, 21). This functional benefit was associated with long-term graft survival, neuronal differentiation of grafted neural precursors and correlated functionally with improvement in motor-evoked potentials. In a more recent study, we have demonstrated comparable survival and neuronal differentiation of human spinal neural precursors at 6–7 weeks after lumbar grafting in naive immunosuppressed minipigs (33). These data have validated the potential use of cell replacement therapy in a well-defined spinal ischemia injury model characterized by selective interneuronal loss.

In the present study, we compare the engraftment properties and differentiation profile of human neural precursors derived from three different human embryonic stem cell lines (HUES-2, HUES-7, HUES-9) once grafted into the lumbar spinal cord of spinal ischemia-injured Sprague–Dawley (SD) rats or in naive immunosuppressed Gottingen–Minnesota minipigs. The degree of cell engraftment and maturation was validated by immunofluorescence staining using human-specific antibodies and coupled with quantitative analysis of surviving grafted human cells using unbiased stereological principles. In a separate study, the development of immune tolerance to grafted embryonic stem cell-derived neural precursor cells (ES-NPCs) and the teratoma formation after grafting of pluripotent or sorted ES-NPCs between 3 weeks to 6 months after cell grafting was studied.

Materials and Methods

All procedures involving rat spinal grafting studies were approved by the Animal Care Committee at the University of California, San Diego. Male Sprague–Dawley rats (300–350 g) were obtained from Harlan (Indianapolis, IN) and were housed in standard cages with corncob bedding. Animals had access to food and water ad libitum and were housed separately after surgery. A 12-h light/dark cycle (lights on at 7:00 AM) was used throughout the study.

All minipig grafting studies were carried out under protocols approved by the Institutional Animal Care and Use Committee of the Czech Academy of Sciences and were in compliance with the Association for Assessment of Laboratory Animal Care guidelines for animal use. Minipigs resulting from crossbreeding of Minnesota and Gottingen strains (both sexes; 18–23 kg) were obtained from the Institute of Animal Physiology and Genetics, Czech Academy of Sciences, Libechov, Czech Republic.

All studies were performed in such a manner as to minimize group size and animal suffering.

Induction and Isolation of Neural Precursors From Human Embryonic Stem Cells (ES-NPCs)

Proliferating HUES-2, HUES-7, or HUES-9 (Harvard University) colonies were used to induce embryonic bodies (EBs) using previously reported methods (37). After 7–14 days, EBs were dissociated by mechanical trituration and plated into low-adherence dishes in human embryonic stem cell (hESC) medium without fibroblast growth factor 2 (FGF2) for approximately 7 days. Next, EBs were plated onto poly-ornithine/laminin (Sigma)-coated dishes in DMEM/F12 plus N2 (Invitrogen). Rosettes were then dissociated with accutase (Chemicon) and plated again onto coated dishes with the same medium but now with FGF2. Populations of NPCs were achieved after one to two passages with accutase in the same conditions. To enrich NPCs, cells were fluorescence activated cell sorted (FACS) and CD184⁺, CD44/271^-, and CD24/CD15⁺ cell populations further expanded on poly-ornithine/laminin (Sigma)-coated dishes or in nonadherent Nunclon flasks to form neurospheres. For expansion, the bFGF as a sole mitogen was used (10 ng/mg) (38). A subpopulation of sorted NPCs (HUES-7 NPCs) was infected with synapsin-enhnaced green fluorescent protein (EGFP) lentivirus (HIV1; 10 M.O.I.) to facilitate the confirmation of neuronal identity of spinally grafted cells.

To induce in vitro neuronal differentiation, NPCs were treated with 10 ng/ml brain-derived growth factor (BDNF), 20 ng/ml glial-derived growth factor (GDNF) and 0.5 mM dibutyryl cyclic AMP for 2–3 weeks and then fixed with 4% paraformaldehyde for immunofluorescence staining or used for whole-cell patch clamp recording.

Multiparameter Intracellular FACS

Proliferating, previously sorted NPCs were dissociated at different passages to a single-cell suspension with accutase (BD Biosciences), and cell suspensions were fixed with fixation buffer and permeablized with perm buffer III (both from BD Biosciences) according to the manufacturer's recommendations. Permeabilized cells were resuspended in stain buffer (BD Biosciences) and were stained with the following antibodies: sex-determining region y Box 1-phycoerythrin (Sox1 PE), Sox2 peridinin-chlorophyll-protein complex-cyanine 5.5 (PerCP-Cy5.5), Nestin V450, and Nanog Alexa Fluor 647 (all from BD Biosciences). Gating was calculated based on corresponding isotype controls (all from BD Biosciences). Compensation was calculated using anti-mouse Ig CompBead Plus (BD Biosciences) that was stained with primary antibodies according to the manufacturer's recommendations. Cells were analyzed on a BD LSRII flow cytometry system.

Electrophysiology

Patch Clamp Recordings

Whole-cell patch recordings were performed from sorted HUES-7 NPCs that were infected with HIV1-channelrhodopsin (ChR2)-yellow fluorescent protein (YFP) lentivirus (39) (generous gift from Karl Deisseroth, Stanford) and differentiated for 3 weeks prior to recording. The recording micropipettes (tip resistance 4–6 MΩ) were filled with internal solution: 135 mM K-gluconate, 4 mM MgCl₂, 10 mM HEPES, 10 mM EGTA, 4 mM Mg-ATP, and 0.2 mM Na-GTP (pH 7.4). Recordings were made using a MultiClamp 700B amplifier and Digidata 1440A interface (Molecular Devices). Signals were filtered at 10 kHz and sampled at 10 kHz. The whole-cell capacitance was fully compensated. The bath was constantly perfused with fresh HEPES-buffered saline: 140 mM NaCl, 5 mM KCl, 10 mM HEPES, 1 mM EGTA, 3 mM MgCl₂, 10 mM glucose (pH 7.4). For current-clamp recordings, cells were clamped at a range of −60 to −80 mV. For voltage-clamp recordings, cells were clamped at −60 mV. Cells were visualized using an OLYMPUS BX51W1 fixed-stage upright microscope. All recordings were performed at room temperature.

Optical Stimulation

To excite ChR2-YFP ion channels, the fiber coupled blue laser module FC-473-300-MM (RGBLase LLC, CA, USA) with split fiber (for safe red guiding laser) and collimator optics was used (central wavelength 473 nm, optical power output 300 mW, fiber type 200 μm). The laser module was triggered through external trigger by recording software (Clampex, Molecular Devices). Cells were exposed to laser pulses of different durations (5–500 ms, typically 20–30 ms) in both voltage and current clamp modes.

Induction of Spinal Ischemic Injury in Rat

Transient spinal cord ischemia (10 min) was induced as previously described (30). Briefly, in isoflurane (1.5–2%)-anesthetized SD rats, a 2F Fogarty catheter (Am. V. Muller, CV 1035; Baxter, Inc., Irvine, CA, USA) was passed through the left femoral artery to the descending thoracic aorta to the level of the left subclavian artery. Distal arterial pressure (i.e., below the level of aortic occlusion) was monitored by cannulation of the tail artery (PE-50). Spinal cord ischemia was induced by inflation of the intra-aortic balloon catheter (0.05 ml of saline) and concurrent systemic hypotension (40 mmHg) induced by blood withdrawal [10.5–11 cc into a heated (37°C) external reservoir] via a 20-gauge polytetrafluoroethylene catheter placed in the left carotid artery. The efficacy of the occlusion was demonstrated by an immediate and sustained drop in distal blood pressure. After 10-min ischemia, the balloon was deflated, and the blood was reinfused. When the arterial blood pressure was stabilized (20–30 min after reflow), the arterial lines were removed and wounds were closed. After ischemia, the recovery of motor function was assessed in 2-day intervals using a modified 21-point open field locomotor scale (3). Only animals with Basso, Beattie, and Bresnahan (BBB) score of 0–4 (i.e., corresponding with chronic paraplegia) at 1–2 months after ischemia were used in the transplantation study.

Induction of C2 Hemicontusion in Immunodeficient (Athymic) Rats

Athymic, male Sprague–Dawley rats (250–275 g) were anesthetized with isoflurane (1.5% maintenance, in room air) and via sharp and blunt dissection the C2 vertebra was exposed and clamped in a stereotaxic frame (Stoelting, CO). A unilateral (left) laminectomy of C2 was performed with a dental drill, and a unilateral contusion injury was made using a New York Impactor (1 mm tip, 12.5 mm drop, 10 g), injuring the most lateral 1 mm of the spinal cord just caudal to the C2 dorsal root. The animal was then removed from the spinal clamps, the surgical site washed with a penicillin/streptomycin solution, and the muscle and skin were closed in several layers.

Preparation of hES-NPCs for Spinal Implantation

A summary of all experimental animal groups receiving spinal cell implants is provided in Table 1.

Table 1.

Summary of Experimental Groups Receiving Spinal Cell Implant

Experimental Model	Cell Line (Single Cell or Spheres)	No. of Spinal Injections/Cell Density	Postgrafting Survival	Immunosupp.
Spinal ischemia I (SD rat; n = 42) L2–L5 spinal grafts	HUES-2, Hues-7 (single cell suspension) HUES-9 (neurospheres)	5 bilateral/0.5 μl each 7,500–10,000 cells/inj. 5 bilateral/0.5 μl each 5–15 spheres/inj.	2 weeks to 4.5 months	+ (±)( * )
Spinal ischemia II (immunodeficient rat; n = 4) L2–L5 spinal grafts	Hues-7-NPCs (Syn-EGFP+) (single cell suspension)	5 bilateral/0.5 μl each 7,500–10,000 cells/inj.	2 months	N/A
Spinal cervical (C2) hemicontusion (immunodeficient rat; n = 6) C2 spinal grafts	Hues-7-NPCs (single cell suspension)	4 injections/0.5 μl each 7,500/inj.	6 months	N/A
Naive control (Gottingen–Minnesota minipigs; n = 6) L2–L5 spinal grafts	Hues-7-NPCs (single cell suspension)	5 bilateral/6 μl each 90,000–120,000 cells/inj.	4–8 weeks	+
Spinal ischemia (SD rat; n = 9) L3–L6 spinal grafts	Hues-7-pluripotent (single cell suspension)	5 bilateral/0.5 μl each 500, 2500 or 5,000 cells/inj.	2–6 weeks	+

*

In this experimental group animals were initially immunosuppressed by combined treatment with Prograf and mycophenolate mofetil (see p. 8 for details) and then continue to receive daily Prograf injection (1 mg/kg SC) for additional 4 months.

Previously sorted, proliferating ES-NPCs cultured on poly-ornithine/laminin (Sigma)-coated dishes were treated with accutase (Chemicon) for 10–15 min, and a single cell suspension was prepared by repeated (10–15 times) trituration using a 10-cc plastic pipette. The cell suspension was then treated with DNAse solution for 5–10 min. To remove cell debris, the cell suspension was then placed on the top of an ovomucoid gradient (Wortington, NJ) and centrifuged at 70 × g. Cells were then washed 2x in PBS, and the cell pellet was resuspended in calcium-free PBS. Some rats were grafted with ES-NPC-derived neurospheres. Before grafting, the neurosphere suspension was filtered through 40-μm plastic mesh without previous enzymatic treatment. It was estimated that, on average, 5–15 neurospheres (20–40 μm in diameter) were injected in a single injection.

For minipig grafting experiments, aliquots of previously frozen NPC cell suspension were used. On day of grafting, cells were thawed and washed 3x in PBS at 70 × g.

For all cell lines (freshly harvested or previously frozen), the viability of the cells was tested using a fluorescein diacetate/propiodium iodide or Trypan blue extrusion test. On average, 75–90% of viable cells were counted for all three cell lines used. After the viability test was completed, cells were stored in hibernation buffer at 4°C. Just before initiation of cell grafting, the single cell or neurosphere suspension was aspirated into a glass capillary (tip diameter: 100–120 μm) (for rat grafting) or 30-gauge needle (for pig grafting) using a 50-μl Hamilton syringe (see the following paragraphs for details).

Spinal Cord Implantation of hES-NPCs in Rats and Minipigs

Spinal Ischemia-Injured SD Rats

Animals with previous ischemic injury were anesthetized with 1.5–2% isoflurane, placed into a spinal unit apparatus (Stoelting, Wood Dale, IL, USA), and a partial laminectomy of Th12-L1 vertebra was performed using a dental drill (exposing the dorsal surface of L2–L5 spinal segments). Using a glass capillary connected to a microinjector (Stoelting), rats received 10 bilateral injections of the ES-NPCs (5 on each side; 0.5 μl/injection; 15,000–20,000 cells/μl; n = 10–16 for each cell line) in hibernation buffer. The duration of each injection was 60 s followed by a 30-s pause before capillary withdrawal. The center of the injection was targeted into the central gray matter (laminae V–VII) (distance from the dorsal surface of the spinal cord at L3 level: 1 mm) (14). The rostrocaudal distance between individual injections ranged between 300 and 500 μm. After implantation, the incision was cleaned with 3% H₂O₂ and penicillin/streptomycin mixture and closed in two layers. All animals were immunosuppressed with mycophenolate mofetil (MFF) (30 mg/kg/day IP) for the initial 7 days, and then MFF treatment was stopped. MFF treatment was combined with Prograf (3 mg/kg/day IP) for the initial 14 days. At 14 days, the dose of Prograf was lowered to 1 mg/kg/day IP and continued for the rest of survival. Immunosuppressive treatment was initiated 1 day before cell grafting. Animals were sacrificed at 2 weeks, 1 month, 2 months, and 4.5 months (n = 2–4 for each time point).

C2 Hemicontusion-Injured Athymic Rats

Three days after C2 hemicontusion, rats (n = 6) were placed under isoflurane anesthesia (1.5%), and the C2 vertebra was exposed and clamped, as described above. Spinal injections were made using a 33-gauge stainless steel needle and a manually controlled microinjector (Kopf Instruments). Four injections were made into the lesion site (0.5 μl/injection; 15,000 cells/μl). After transplantation, wounds were closed and animals were allowed to survive for 6 months.

Gottingen–Minnesota Minipigs

Animals were prepared for spinal cell grafting as previously described (33). Briefly, minipigs (both sexes; 18–23 kg; n = 6) were premedicated with intramuscular azaperonum (2 mg/kg) and atropine (1 mg/kg; Biotika, SK) and then induced with ketamine (20 mg/kg IV). After induction, animals were intubated with a 2.5F tracheal tube. Anesthesia was maintained with 1.5% isoflurane. Oxygen saturation was monitored throughout the procedure using a pulse oximeter (Nellcor Puritan Bennett, Inc., Ireland). After induction of anesthesia, animals were placed into supine position with all four extremities fixed to the operating table. The left jugular vein was exposed and catheterized with an 18-gauge central venous catheter (Certofix Mono V 330; B Braun, Germany). The end of the catheter was externalized on the side of the neck. Animals were then placed into prone position and prepared for spinal cell grafting. To immobilize the lumbar spinal cord, animals were mounted into a spinal immobilization apparatus and the lumbar portion of the animal was lifted 5” above the operating table to eliminate respiration-caused spinal cord pulsation. A dorsal laminectomy of L2–L5 vertebrae, corresponding to L3–L6 spinal segments, was then performed, and epidural fat was removed using cotton swabs. The dura was left intact. To deliver cells, the XYZ manipulator (Stoelting, Wood Dale, IL, USA) was used and mounted directly to the operating table. A Hamilton syringe with a 30-gauge needle was then mounted into the manipulator and connected to the microinjector (Stoelting) using PE-50 tubing. To connect the PE-50 tubing to the Hamilton syringe, the plunger was removed and one end of the PE-50 tubing was inserted 1 cm into the syringe and sealed with silicone. Animals then received a total of 10 injections (5 on each side; 6 ml/injection; 15,000–20,000 cells/ml) targeted into the intermediate zone (lamina VII) of L3–L6 segments. The distance between individual injections was 1–1.5 mm. All surgical interventions followed rigid aseptic procedures. All materials were subjected to autoclaving or gas sterilization.

Immediately after jugular vein catheterization, the animals received a bolus (0.1 mg/kg) injection of Prograf⁸ followed by continuous Prograf infusion (0.05 mg/kg/day) using externally mounted 5–7 day infusion pumps (Baxter Infusor) secured in a custom-made “minipig jacket.” Prograf® plasma concentration was monitored at 2-day intervals for the first 10–14 days and then at 7-day intervals for the duration of the study using radioimmunoassay (RIA). Targeted plasma concentration was 15–20 ng/ml. After cell grafting, animals survived for 4–8 weeks.

In Vivo Teratoma Analysis in SD Rats After Spinal Grafting of Pluripotent HUES-7 Cells

Sprague–Dawley rats with previous ischemic injury were anesthetized with 1.5–2% isoflurane, placed into a spinal unit apparatus (Stoelting, Wood Dale, IL, USA) and a partial laminectomy of the Th12-L1 vertebra was performed using a dental drill (exposing the dorsal surface of L2–L5 spinal segments). A single cell suspension was prepared from proliferating ES (HUES-7) colonies using collagenase and cells used for spinal grafting as previously described. All animals received five bilateral injections with increasing cell densities (500 μm apart) targeted into the intermediate zone (lamina VII) of L2–L5 segments. Three groups were studied: Group A, n = 3, 500 cells/injection; Group B, n = 3, 2,500 cells/injection; or Group C, n = 3, 5,000 cells/injection. As in the previous study, all animals were immunosuppressed with MFF (30 mg/kg/day IP) for the initial 7 days and then MFF treatment was stopped. MFF treatment was combined with Prograf (3 mg/kg/day IP) for the initial 14 days. At 14 days, the dose of Prograf was lowered to 1 mg/kg/day IP and continued for the rest of survival. After cell grafting, animals survived for 3–6 weeks and were then perfusion-fixed with 4% paraformaldehyde. The presence of teratoma was analyzed in transverse hematoxylin and eosin (H&E)-stained spinal cord sections.

Perfusion Fixation and Tissue Processing in Spinally Grafted Rats or Minipigs

At the end of the survival periods, animals were anesthetized with pentobarbital (100 mg/kg IP) and transcardially perfused with heparinized saline (100 ml, rats; 5 L, minipigs) followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB; 500 ml, rats; 5 L, minipigs). The spinal cords were dissected and postfixed in the same fixative for 1–3 days at 4°C. After postfixation, tissue was cryoprotected in graded sucrose solutions (10%, 20%, and 30%).

Immunofluorescence Staining

A standard immunofluorescence protocol was followed. Briefly, after cryoprotection, frozen coronal spinal cord sections (20–30 μm) were cut. Free-floating sections were placed in PBS (0.1 M; pH 7.4) containing 5% normal goat or donkey serum (NGS, DS) and 0.2% Triton X-100 (TX) for 2 h at room temperature to block nonspecific protein activity. This was followed by overnight incubation at 4°C with primary human-specific antibodies: mouse anti-nuclear protein/h-nuc (hNUMA; 1:100; Chemicon Int., Temecula, CA, USA), mouse anti-neuron-specific enolase (hNSE; 1:200; Novocastra Laboratory), mouse anti-synaptophysin (hSYN; 1:1,000; Chemicon Int.), mouse anti-nestin (hNESTIN; 1:1,000; Chemicon Int.), mouse anti-glial fibrillary acidic protein (hGFAP; 1:500; OriGene), rat anti-axonal neurofilament (HO14; 1:200; gift from Virginia Lee, University of Pennsylvania School of Medicine). These antibodies were combined in multiple labeling experiments with human nonspecific antibodies: rabbit anti-GFAP (1:500; Chemicon Int.), goat anti-doublecortin (DCX; 1:100; Chemicon Int.), mouse anti-microtubule-associated protein 2 (MAP2; 1:1,000; Chemicon Int.), rabbit anti-Ki67 (1:1,000; Abcam), mouse anti-β-III tubulin (TUJ1; 1:1,000; Millipore, Inc.), goat anti-choline acetyltransferase (CHAT; 1:200; Millipore, Inc.), mouse anti-tyrosine hydroxylase (TH; 1:500; Millipore, Inc.), rabbit anti-ionized calcium binding adaptor molecule 1 (Iba1; 1:1,000; Wako, Inc.), mouse anti-antigen presenting cells [APC (CD4/CD8); 1:500; Oncogene, Inc.]. After incubation with primary antibodies, sections were washed three times in PBS and incubated with fluorescent-conjugated secondary goat anti-rabbit, goat anti-mouse, or donkey anti-goat antibodies (Alexa 488, 594, 680; 4 μl/ml; Molecular Probes, Eugene, OR, USA). All blocking and antibody preparations were made in 0.1 M PBS/0.2% TX/5% NGS or 5% DS. For general nuclear staining, DAPI (3 μl/ml) was added to the final secondary antibody solutions. After staining, sections were mounted on slides, dried at room temperature, and covered with Prolong antifade kit (Molecular Probes). Stained sections were analyzed and photographed with an epifluorescence microscope (AX70; Olympus) and confocal microscope (Fluoview 1000, Olympus). The same immunofluorescence staining protocols were used for cultured hES-NPCs. In addition, the following antibodies were used for in vitro immunofluorescence staining: mouse anti-octamer binding transcription factor 4 (Oct-4; 1:500; Millipore, Inc.), rat anti-stage-specific embryonic antigen-3 (SSEA-3; 1:500; Millipore, Inc.), rabbit anti-Sox2 (1:500; Millipore, Inc.).

Quantification of Ki67 Expression in hNUMA⁺-Grafted Cells

To quantify Ki67/hNUMA⁺ cells, four spinal cord sections taken at intervals of a minimum of 100 μm from the grafted regions were used. The total number of hNUMA-positive and Ki67-positive nuclei were counted using the Image Pro software (Media Cybernetics).

Stereological Quantification of Grafted Cells

The total number of grafted cells immunoreactive for the hNUMA antibody was estimated using stereological, unbiased, and systematic sampling (14, 36). Four animals from each experimental group grafted with single cell suspension of HUES-2, HUES-7, and HUES-9-derived NPCs were used at 4 weeks after cell grafting. Each 10th previously stained section (30 mm thick) taken from grafted L2–L5 grafted segments was used for stereological quantification after applying a fractionator sampling scheme. The optical images (1 mm thick) were obtained with a Leica DMLB microscope using a 100x oil objective with numerical aperture 1.3. The total number of grafted cells was then calculated by applying the fractionator formula N = Q x 1/hsf x 1/asf x 1/ssf, where N is a total number of positive nuclei, Q is sum of cells counted, hsf is the height sampling fraction, asf is area sampling fraction, and ssf is slice sampling fraction. The calculated coefficient of error (CE) values ranged between 0.05 and 0.08. Individual cell nuclei were counted using an optical Dissector (Ellipse, SK). 3D reconstruction from Z-stack confocal optical images (0.5 mm thick) was done utilizing Improvision Volocity software Version 5.0.4.

Results

Effective Generation of ES-Derived Neural Precursors In Vitro

Proliferating human ES colonies (HUES-7) cultured on a feeder layer of mitomycin C-treated mouse embryonic fibroblasts showed consistent expression of pluripotent markers such as OCT4 and SSEA-3 (Fig. 1a–f). At the early stage of colony differentiation the expression of nestin was typically seen (Fig. 1h, i) and followed by progressive appearance of TUJ1-immunoreactive cells localized at the periphery of individual colonies (Fig. 1k, l). Sorted (CD184⁺, CD44^-, CD271^-, CD24⁺, and CD15^LO/HI) ES-NPCs cultured on ornithine and laminin surface in the presence of FGF-2 showed a typical multipolar morphology and were expanded for 10–15 passages (Fig. 1m–o). A comparable generation of NPCs from HUES-2 and HUES-9 cell lines was seen. However, the proliferation rate of sorted NPCs, if compared to HUES-7-derived NPCs, was slower (about 1–1.5 doublings/week).

Figure 1.

(a–f) Proliferating HUES-7 colonies cultured on mouse embryonic fibroblasts showed consistent expression of pluripotent markers such as OCT4 and SSEA-3. (g–l) During the initial phase of ES colony differentiation, a progressive increase in number of hNESTIN and TUJ1-immunoreactive cells localized primarily at the periphery of colonies can be seen. (m–o) FAC-sorted (CD184⁺, CD44/271^-, and CD24/CD15⁺) proliferating NPCs were expanded for up to 10–15 passages. During expansion, proliferating cells displayed a typical multipolar morphology. (p–s) Intracellular FACS analysis of passage 15 postsort HUES-7 NPCs stained with fluorochrome-conjugated antibodies to neural stem cell markers SOX1, SOX2, and hNESTIN, and the pluripotent stem cell marker Nanog. A dim background staining of 10–20% higher than isotype control with the N31-355 human Nanog antibody was seen. In all isolated NPC populations, a small percentage of NPCs are hNESTIN^POS/SOX2^POS/SOX1^NEG/LOW. Percents are indicated for each identified cell population.

We assessed the purity of proliferating previously sorted NPCs by intracellular FACS (Fig. 1p–s). HUES-7 NPC passage 15 postsort was highly positive for the NPC markers Nestin, Sox2 and Sox1 and did not express the pluripotency marker Nanog above background levels that are consistent with the N31-355 antibody. These data demonstrate that sorted NPCs are highly pure for multiple passages postsort and do not contain residual undifferentiated hESCs.

Electrophysiological Characterization of ES-Derived Neurons In Vitro

For electrophysiological recordings, proliferating NPCs were induced with cell culture media containing cAMP, GDNF, and BDNF for 2–3 weeks prior to recording. In induced NPC cultures, a consistent presence of DCX⁺/hNUMA⁺ neurons was seen (Fig. 2a–c).

Figure 2.

(a–c) HUES-7-derived, previously FAC-sorted NPCs were induced for 2–3 weeks with dibutyryl cyclic AMP, GDNF, and BDNF. After 4% paraformaldehyde fixation, cells were stained with DCX, hNUMA, and hNESTIN antibody. A high density of DCX⁺ neurons with subpopulation of hNESTIN-positive cells can be seen. For whole-cell patch-clamp recording, HIV1-SYN-ChR2-YFP-infected ES (HUES-7)-derived neurons previously cultured for 3 weeks in differentiating medium were used. (d) A brightfield microphotograph of patched cell. (e) Autofluorescence image (same cells as seen in brightfield) demonstrating YFP expression in HIV1-SYN-ChR2-YFP-infected neuron at 3 weeks after lentivirus infection. (f) Voltage-clamp recording at holding membrane potential (MP) of −60 mV, with step changes of MP (200 ms long) from hyperpolarized states to depolarized MP (from −100 to +40 mV) with the step of 20 m V. Note that almost no currents in hyperpolarized states are present. Fast inward (Na⁺) and persistent outward (K⁺) currents in depolarized membrane potentials, characteristic of neuronal cells can be seen. (g) Current-clamp recording (membrane potential −60 mV) with current pulses (300 ms) from −40 to +30 pA with step of 10 pA. Note that clear action potentials are triggered by depolarizing current pulses. (h) Voltage-clamp recording at holding MP of −60 mV with exposure to 200-ms-long pulse of blue laser light. Inward current indicates functional light-sensitive ChR2 cationic channels in the cell membrane. Note the presence of a pronounced and fast initial part of inward current followed by a smaller but persistent inward current response. (i) Current clamp recording at holding MP of −65 mV with repetitive exposures of 100-ms-long pulses of blue laser light in frequency of 2 Hz. Each light impulse triggered action potential as a result of fast membrane depolarization by ChR2 opening. In this particular case with prolonged 100-ms-long light impulse, each action potential is followed by unusual “elbowlike” depolarization caused by prolonged ChR2 cationic channels opening (while blue light is on). Note consistency in timing and shape of evoked responses.

Whole-cell recordings were performed on 6 ES (HUES-7)-derived neurons previously infected with HIV1-ChR2-YFP (ChR2-channel rhodopsin) lentivirus (Fig. 2e).

All recorded neurons showed negative resting membrane potential (RMP), close to normal RMP in neuronal cells (-40 to −70 mV). All recorded cells exhibited pronounced fast inward current on depolarization in voltage clamp mode (Fig. 2f). In current clamp mode, depolarization triggered action potentials (Fig. 2g). These signs demonstrate the neuronal phenotype of recorded ES-derived neurons.

To test the degree and functionality of channel rhodopsin-expressing neurons, cells were exposed to pulses of blue laser light of different duration. All laser pulses evoked membrane currents. In voltage clamp mode (Fig. 2h), the laser pulse gave rise to inward current, with a fast and pronounced initial part (more than −100 pA) and smaller and slower but persistent later part of inward current. In current clamp mode, the same pulses evoked depolarizations, and pulses longer than 4 ms triggered reliable action potentials. With longer duration of laser pulses, the spikes were accompanied by an unusual elbow-like form of depolarization, likely caused by continually opened ChR2 cationic ion channels, while the laser light was still on. Responses to laser excitation were reliable and consistent in shape and occurrence. Figure 2i shows responses to 100 ms lasting laser pulses, in frequency of 2 Hz.

Survival and Maturation of ES-NPCs in Rat Spinal Cord at 2–8 Weeks After Transplantation

Using human-specific antibodies, HUES-9-derived NPCs grafted spinally as neurospheres were identified in targeted spinal cord regions at 2–4 weeks after grafting. At both time points, intense hNESTIN-immunoreactive grafted cells were readily identified (Fig. 3a). No CHAT-immunoreactive neurons were seen within the grafted regions (Fig. 3b, c). Costaining with TH antibody showed numerous host-derived TH⁺ fibers crossing the hNESTIN⁺ grafts (Fig. 3a–c, yellow arrows). Costaining with hNUMA, DCX, and TUJ1 antibody showed numerous terminally differentiated grafted DCX⁺ neurons, which were also TUJ1-immunoreactive (Fig. 3d–h). In general, only limited migration of grafted cells was seen (Fig. 3h, yellow arrows), and individual grafts retained their spherical shape (Fig. 3a, b, yellow dashed circle).

Figure 3.

Human ES (HUES-9)-derived spinal neural precursors grafted into lumbar spinal cord in the form of neurospheres show long-term survival, neuronal differentiation, but only limited migration into the host parenchyma. (a–c) Using a glass capillary, rats with previous ischemic injury received 10 bilateral injections of ES-NPC neurospheres targeted into the intermediate zone of L2–L5 spinal segments. Lumbar spinal cord sections from cell-grafted animals at 4 weeks after cell grafting immunostained for human nestin (hNESTIN) showed a dense population of hNESTIN⁺ immature neural precursors at the core of transplanted neurospheres. Costaining of the same sections with CHAT and TH antibody showed no CHAT⁺-immunoreactive cells within grafted areas. Numerous TH⁺ processes were identified crossing the hNESTIN⁺ grafts (b, yellow arrow). (d–h) Triple staining with hNUMA, DCX, and TUJ1 antibody showed a dense population of grafted TUJ1 and DCX-immunoreactive neurons. Only limited migration of hNUMA⁺ cells outside of grafted neurospheres was seen (h, yellow arrow).

HUES-7-derived NPCs grafted spinally as a single cell suspension and analyzed 2–3 weeks after transplantation have shown a similar neural differentiation pattern as seen in neurosphere-NPC-grafted animals with numerous hNUMA/DCX-immunoreactive grafted neurons (not shown). At 8 weeks after grafting, a more pronounced degree of neuronal maturation was identified and was expressed by the presence of markers which are typical for near mature or mature neurons. Double staining with human specific synaptophysin (hSYN) and DCX antibody showed immunoreactive grafted human neurons in targeted spinal cord laminae (LIV-IX) (Fig. 4a–c). In the same areas, staining with human-specific NSE (hNSE) antibody showed a dense population of hNSE/MAP2-immunoreactive neurons (Fig. 4d–f).

Figure 4.

Human ES (HUES-7)-derived spinal neural precursors grafted into lumbar spinal cord of rats with spinal ischemic injury in form of single cell suspension show long-term survival, neuronal differentiation, and migration into the host parenchyma. (a–c) Using a glass capillary, rats with previous ischemic injury received 10 bilateral injections of ES-NPCs delivered as a single cell suspension and targeted into the intermediate zone of L2–L5 spinal segments. Lumbar spinal cord sections from cell-grafted animals at 8 weeks after cell grafting immunostained for human synaptophysin (hSYN) and doublecortin (DCX) revealed a high density of grafted DCX-immunoreactive neurons (a). In the same region an intense population of hSYN punctata-like structures were identified (b). (d–f) Staining of the adjacent sections with human-specific NSE (hNSE) antibody, a marker of near mature or mature neurons and MAP2, showed intense hNSE immunoreactivity in grafted MAP2-positive neurons. (g–j) Confocal analysis of spinal cord section taken from an animal grafted with synapsin-GFP-labeled HUES-7 NPCs at 2 months after cell grafting. Sections were stained with DCX antibody and show a near complete colocalization of both proteins. (k–m) Confocal analysis of 0.3–0.5 μm optical images of double-stained sections with hSYN/DCX antibody showed a spatial colocalization of hSYN punctate-like immunoreactivity with DCX⁺ processes.

In contrast to neurosphere-grafted animals, several solitary synapsin-GFP-labeled neurons, which were also DCX-immunoreactive and which migrated outside of the graft, were also identified in the ventral (lamina VIII–IX) gray matter (Fig. 4g–j). Costaining with Ki67 (mitotic marker) and hNUMA antibody showed only occasional (less than 2%) Ki67/hNUMA⁺ cells identified at 8 weeks after grafting (not shown).

Confocal analysis of 0.3- to 0.5-μm-thick optical images taken from sections double-stained with hSYN and DCX antibody showed colocalization of hSYN-immunoreactive punctata with DCX-positive processes (Fig. 4k–m).

Stereological quantification of NUMA⁺ cells showed similar survival of HUES-2, HUES-7, and HUES-9-derived NPCs at 2 months postgrafting (Table 2).

Table 2.

Quantification of hNUMA-Positive Cells in Rats Grafted With HUES-2, HUES-7, and HUES-9 Cell Lines and Surviving for 2 Months

Subject No.	HUES-2	HUES-7	HUES-9
No. 1 a	3,857	5,752	3,940
No. 2 a	5,233	4,690	3,112
No. 3 a	4,876	3,212	4,734
No. 4 a	4,967	2,988	3,489
Average b	4,733 ± 603	4,160 ± 1302	3,818 ± 697

a

Data for individual subject represent an average cell counts/injection site.

b

Data expressed as average ± SD.

Qualitative analysis of grafted HUES-7 and HUES-2-derived NPCs once grafted in a form of single cell suspension showed no detectable difference and a comparable distribution of grafted NUMAQ⁺/DCX⁺ neurons was seen (not shown).

Survival and Maturation of Spinally Grafted ES-NPCs in Naive Immunosuppressed Minipigs

Staining of transverse spinal cord sections taken from L2–L5 segments of grafted minipigs at 23 days after grafting showed a similar cell engraftment pattern as was seen in rat spinal cord. Numerous double-stained hNUMA/DCX⁺-immunoreactive neurons were identified in grafted regions (Fig. 5a–e). Costaining with MAP2 antibody showed that virtually all hNUMA/DCX⁺ neurons were also MAP2 immunoreactive (Fig. 5c–e).

Figure 5.

Human ES (HUES-7)-derived spinal neural precursors grafted into lumbar spinal cord of naive immunosuppressed minipigs in form of single cell suspension show similar differentiation profile as seen in rat spinal ischemia model. Using a 30-gauge stainless steel needle mounted on a XYZ micromanipulator, minipigs received 10 bilateral injections of ES-NPCs delivered as a single cell suspension and targeted into the intermediate zone of L3–L6 spinal segments. All animals were immunosuppressed by a continuous infusion of Prograf (0.05 mg/kg/day) delivered through a chronically placed intrajugular vein catheter coupled to an externally mounted 5-day infusion pump. Lumbar spinal cord sections were analyzed at 3 weeks to 2 months after cell grafting. (a–e) Staining of transverse spinal cord sections with hNUMA, DCX, and MAP2 antibody showed a dense population of triple-labeled grafted cells at 3 weeks after grafting. Confocal analysis showed that virtually all grafted hNUMA/DCX-immunoreactive neurons were also MAP2 immunoreactive. (f–i) Using human specific axonal neurofilament antibody (HO14), which does not cross react with porcine tissue, numerous HO14⁺ axons were seen in grafted regions at 2 months after cell transplantation. Some of the HO14⁺ axons showed colocalization with DCX (g, h, yellow arrow) and showed projections towards CHAT⁺ α-motoneurons (i). (j) To identify proliferating cells, sections were stained with Ki67 antibody. Less than 1% of grafted hNUMA⁺ cells were found to colocalize with Ki67 immunoreactivity. (k–m) Triple labeling with hNUMA, DCX, and GFAP antibody showed that a majority of grafted cells were DCX positive.

At 49 days after grafting, a more advanced stage of neuronal maturation was seen. Using human-specific axonal neurofilament antibody (HO14), which does not cross-react with the porcine tissue, extensive axonal sprouting was detected in hNUMA-positive grafted regions (Fig. 5f–h). Numerous HO14⁺ axons projecting towards CHAT-immunoreactive host α-motoneurons were identified (Fig. 5i). Probing for glial phenotype in grafted hNUMA-positive cells showed only occasional GFAP⁺ astrocytes, and the majority of grafted cells in the core of individual grafts were DCX positive (Fig. 5k–m). To identify proliferating grafted cells, sections were double-stained with hNUMA and Ki67 antibodies. Only occasional (less than 1%) double-stained cells were identified and the majority of hNUMA⁺ cells were Ki67 negative (Fig. 5j, yellow arrows).

Progressive ES-NPC Graft Rejection in Long-Term Grafted Ischemic Rats

We next analyzed survival of grafted HUES-7-NPCs in rats with previous ischemic injury at 4.5 months after cell grafting. Animals were continuously immunosuppressed with a low dose of Prograf (1 mg/kg SC per day). In contrast to short survival periods, no persisting surviving cells were detected. Massive infiltration with activated microglia/macrophages and presence of hypertrophic astrocytes around the needle tract was seen (Fig. 6a–c). In addition, infiltration with T lymphocytes (CD4/8) in the same regions was detected (Fig. 6c, insert).

Figure 6.

Increased local spinal parenchymal inflammation and lack of grafted cell survival in spinal ischemia-injured rats immunosuppressed with low dose of tacrolimus (1 mg/kg/day IP) at 4.5 months after ES-NPC grafting. (a–c) Transverse spinal cord section taken from the lumbar spinal cord in ischemic rats at 4.5 months after cell grafting and stained with IBA1 (IB1), GFAP, or CD4/8 antibodies. A massive infiltration with IBA1-immunoreactive cells (indicative of activated microglia and macrophages) in the areas of reactive astrocytosis can be seen. In the same regions, CD4/8-positive T-lymphocytes were identified (c, insert).

Formation of Teratomas in Animals Receiving Spinal Grafts of Pluripotent HUES-7 Cells

In Life Phase Neurological Assessment

In all animals receiving cell densities of 2,500 or 5,000/injection, a progressive deterioration of motor function, was seen at 2–3 weeks after grafting when all animals progressed from spastic paraplegia to flaccid paraplegia and developed urinary retention. Two of three animals receiving 500 cells/injection showed partial flaccidity with no signs of recovery at 6 weeks after grafting.

Macroscopical and Histological Analysis

(i) Macroscopic examination of dissected spinal cords revealed a tumor-“like” structure(s) in all neurologically impaired animals. (ii) Histological analysis of transverse spinal cord sections stained with H&E confirmed near complete loss of spinal cord integrity and the presence of teratomas (Fig. 7a–f), as evidenced by the presence of ectoderm (neuroepithelium; pigmented neuroepithelium) (Fig. 7b–d, yellow and red arrows), mesoderm (muscle) (Fig. 7e, red arrows), and endoderm [glandular epithelium; (Fig. 7f, yellow arrow) and respiratory epithelium; Fig. 7f, green arrow)] derivatives.

Figure 7.

(a–f) Formation of teratomas in animals receiving spinal grafts of pluripotent HUES-7 cells. Transverse spinal cord section (hematoxylin and eosin stain) taken from an animal receiving 5,000 cells/injection (total of five bilateral injections were delivered between L2 and L5 spinal segments) of pluripotent HUES-7 cells and surviving for 3 weeks. (a) Near complete loss of spinal cord integrity resulting from expanding teratomas can be seen. (b–f) Ectoderm (neuroepithelium; pigmented neuroepithelium; b–d, yellow and red arrows), mesoderm (muscle; e, red arrows), and endoderm (glandular epithelium; F, yellow arrow and respiratory epithelium; F, green arrow) derivatives can be readily identified. (g–i) Long-term (6 months) engraftment of sorted HUES-7 NPCs without teratoma formation in immunodeficient rats with previous C2 hemicontusion. Transverse spinal cord sections taken from an animal after C2 hemicontusion and grafted with HUES-7 NPCs (7,500 cells/injection) at 3 days after trauma. Sections were double-stained with human-specific GFAP and NSE antibody. (h) Near complete cavity filling with transplanted NSE⁺ neurons can be seen. (g–i) In the same region, numerous hGFAP⁺ astrocytes which migrated into the host tissue can also be identified (g, red arrows).

Long-Term (6-Month) Engraftment of Sorted HUES-7 NPCs with No Teratoma Formation in Immunodeficient Rats with Previous C2 Hemicontusion

Analysis of transverse spinal cord sections taken from the epicenter of cell-grafted C2 spinal cord hemicontusion and stained with human-specific GFAP and NSE antibody showed a near complete cavity filling with transplanted cells. The core of the graft was found to have a high density of hNSE⁺ grafted neurons with a dense hGFAP⁺ population of human astrocytes intermixed with hNSE⁺ neurons (Fig. 7g–i). Numerous hGFAP astrocytes that migrated into the host tissue were also identified (Fig. 7g, red arrows). No teratoma-like structures were seen in any of the six grafted animals.

Discussion

Effective Expansion of FAC-Sorted Human Embryonic Stem Cell-Derived Neural Precursors In Vitro

In our present study, neural precursors used for spinal grafting were FAC-sorted from induced human ES cell colonies after expansion of the rosette-derived NPCs. Populations of CD184⁺/CD271^-/CD44^-/CD24⁺ hNPCs were further expanded for more than 15 passages in the presence of bFGF used as a solo mitogen. There was no detectable change in proliferation rate and neuronal populations capable of generating action potentials in vitro were successfully induced even in high (12–15) passage numbers. For spinal grafting in the minipig model, cells were frozen using a programmable freezer, thawed just before in vivo grafting and then spinally grafted without any additional subculturing. These data demonstrate that, using this sorting, expansion, and freezing protocol, it is possible to effectively generate a homogenous population of transplantable ES-derived NPCs that maintain their neurogenic potential.

Previous studies from other laboratories have demonstrated a comparable FAC-sorting protocol for isolation of CD15⁽⁺⁾/CD29^(HI)/CD24^(LO) neural precursors from human embryonic stem cells (27).

Lack of Teratoma Formation After Spinal Grafting of Sorted ES-NPCs

To address the issue of purity of sorted ES-NPCs and the degree of pluripotent cell contaminants with potential teratoma-forming properties, two separate experiments were performed. First, we have demonstrated consistent teratoma formation at 2–3 weeks after spinal grafting of pluripotent HUES-7 cells once immunosuppressed animals received cell densities of 500–5,000 per injection. Second, we have analyzed spinal cord sections from immunodeficient rats at 6 months after grafting of sorted HUES-7-derived NPCs (7,500 cells/injection) and showed no teratoma formation. Jointly, these data indicate that sorted ES-NPCs have a high degree of purity and that the level of pluripotent cell contaminants, if present, is not sufficient for teratoma formation. However, it is likely that much longer survival periods (up to 18–24 months) will be required in the in vivo teratoma studies to provide definitive safety data on FAC-sorted ES-NPCs.

Spinally Grafted ES-NPCs Show a Similar Differentiation/Maturation Profile as Seen After Spinal Grafting of Fetal Human Spinal Neural Precursors

Current experimental data using rodent models (mice and rats) show that in contrast to brain there is no evidence that neurogenesis takes place in naive or injured adult spinal cord (8, 11). These data indicate that at present the only viable alternative to achieve an effective neuronal repopulation in spinal injury models characterized by local neuronal loss (i.e., spinal ischemic injury or localized spinal trauma) will be a region-specific cell replacement approach. In previous studies, using a rat spinal cord ischemia model, we have demonstrated that spinal grafting of rat or human fetal spinal cord neural precursors or postmitotic human NT neurons is associated with progressive improvement in ambulatory motor function over the course of 2 weeks to 3 months after cell grafting. Grafted human spinal neural precursors showed preferential neuronal differentiation characterized by the initial DCX expression in grafted neurons and then followed by the progressive appearance of human-specific NSE and synaptophysin immunoreactivity at intervals of 2–3 months after grafting (6, 21).

In our current study, a similar differentiation–maturation profile of ES-derived NPCs was seen in both the rat spinal ischemia and C2 contusion models at 2–6 months after grafting and in naive immunosuppressed minipigs at 4–8 weeks. In early stages after grafting (2–4 weeks), prominent DCX immunoreactivity was seen in terminally differentiated grafted neurons. This was followed by the appearance of hNSE⁺ neurons and intense human-specific synaptophysin immunoreactivity identified in the same regions. These maturation and neuron differentiation patterns are thus similar to what we have seen after spinal grafting of human fetal spinal neural precursors in the same rat spinal ischemia model. Similarly, in the minipig model, extensive neuronal differentiation and axonal sprouting was identified at 6–7 weeks after cell transplantation (6, 21, 33). Experimental data from other laboratories have demonstrated similar long-term survival and neuronal differentiation of human ES-blastocyst or fetal tissue-derived neural or oligodendrocyte precursors grafted into (i) neostriatum of immunodeficient naive rats (25), (ii) striatum of 6-hydroxydopamine (6-OHDA) or quinolinic acid-lesioned adult rats (5, 29, 34), or (iii) C5 trauma-injured spinal cord or dorsal root ganglion (DRG)-lesioned induced cavity model in rats (1, 28).

Lack of Immune Tolerance to Grafted ES-NPCs in Long-Term Low-Dose Immunosuppressed Rats

In our previous study, we have demonstrated effective survival of human fetal spinal cord-derived neural precursors or human hNT neurons at 2–3 months after spinal grafting. In those studies, Prograf (1 mg/kg/day SC) was used as the sole immunosuppression drug. In our more recent study using SOD1^G93A rats, we have demonstrated that, using the same cell line, it is necessary to use initial combined immunosuppressive therapy (mycophenolate mofetyl + Prograf; 7 days of combined therapy followed by Prograf treatment only) in order to achieve consistent cell survival at intervals of 2–2.5 months after grafting (12). To test the development of immune tolerance in our current study, a group of spinal ischemia-injured rats was allowed to survive for 4.5 months after spinal ES-NPC grafting while being continuously immunosuppressed with a low dose (1 mg/kg/day) of Prograf until sacrifice. No grafted cell survival was seen in any animal. Instead, a heavy infiltration with inflammatory elements such as microglia/macrophages and CD4/8⁺ T lymphocytes was seen in previously NPC-grafted spinal cord regions. These data demonstrate that our current combined immunosuppressive protocol with the initial use of high dose of Prograf (3 mg/kg/day) for 14 days is not sufficient to induce immune tolerance or long-term graft survival even if followed by a continuous low dose of Prograf (1 mg/kg/day) immunosuppression. Previous studies have demonstrated that continuous plasma levels of Prograf around 15 ng/ml are required for long-term survival of grafted islet cells in mice and that such levels cannot be achieved after single daily bolus subcutaneous delivery of 1–5 mg/kg of Prograf (23, 35).

An alternative explanation for the lack of graft survival is that grafted ES-NPCs cannot survive for extended periods after grafting in rodent spinal cord. To address this question, we grafted spinally the same ES-NPC line (HUES-7-derived) in immunodeficient rats with previous C2 contusion injury and analyzed graft survival at 6 months after grafting. In all grafted animals, consistent presence of grafted cells was seen. Jointly, these data show that (i) a modification in the immunosuppression protocol (i.e., BID drug delivery), (ii) the use of biodegradable Prograf formulations that provide long-lasting steady drug release (23, 35), or (iii) the use of immunodeficient rats will be required in order to achieve consistent grafted ES-NPC cell survival at extended periods after cell grafting. Such long-term graft survival (more than 3 months) is required for the expected formation of synaptic contact between grafted ES-derived neurons and neurons of the host and is therefore essential to conduct long-term correlative studies to identify functional incorporation of grafted ES-NPCs.

Summary

Our current data demonstrate that human embryonic stem cell-derived neural precursors engraft and mature once grafted into previously ischemia-injured lumbar or C2-hemicontused spinal cord in rats or in lumbar spinal cord of naive immunosuppressed minipigs. The maturation–differentiation profile of grafted cells during the first 2 months after grafting is similar as previously seen for human fetal spinal cord-derived neural precursors. These data show that ES-derived, FAC-sorted neural precursors represent a potential cell source to be used in cell replacement therapies in a variety of neurodegenerative disorders, such as brain or spinal ischemic or traumatic injury or in amyotrophic lateral sclerosis.