www.elsevier.com/locate/brainres

Brain Research 1026

Research report

Neural progenitor cell transplants promote long-term functional

recovery after traumatic brain injury

Deborah A. Sheara,b,1, Matthew C. Tatec,1, David R. Archerd, Stuart W. Hoffmane,f,

Verne D. Hulceb, Michelle C. LaPlacac, Donald G. Steine,f,*

aDepartment of Psychology, Emory University, Atlanta, GA, USAbField Neurosciences Institute, Saginaw, MI, USA

cDepartment of Biomedical Engineering, Georgia Institute of Technology/Emory University, Atlanta, GA, USAdDepartment of Pediatrics, Emory University, Atlanta, GA, USA

eDepartment of Emergency Medicine, Emory University, Atlanta, GA, USAfDepartment of Neurology, Emory University, Atlanta, GA, USA

Accepted 28 July 2004

Available online 15 September 2004

Abstract

Studies demonstrating the versatility of neural progenitor cells (NPCs) have recently rekindled interest in neurotransplantation methods

aimed at treating traumatic brain injury (TBI). However, few studies have evaluated the safety and functional efficacy of transplanted NPCs

beyond a few months. The purpose of this study was to assess the long-term survival, migration, differentiation and functional significance of

NPCs transplanted into a mouse model of TBI out to 1 year post-transplant. NPCs were derived from E14.5 mouse brains containing a

transgene-expressing green fluorescent protein (GFP) and cultured as neurospheres in FGF2-containing medium. Neurospheres were injected

into the ipsilateral striatum of adult C57BL/6 mice 1 week following unilateral cortical impact injury. Behavioral testing revealed significant

improvements in motor abilities in NPC-treated mice as early as 1 week, and the recovery was sustained out to 1 year post-transplant. In

addition, mice receiving NPC transplants showed significant improvement in spatial learning abilities at 3 months and 1 year, whereas an

intermediate treatment effect on this behavioral parameter was detected at 1 month. At 14 months post-transplant, GFP+ NPCs were observed

throughout the injured hippocampus and adjacent cortical regions of transplanted brains. Immunohistochemical analysis revealed that the

majority of transplanted cells co-labeled for NG2, an oligodendrocyte progenitor cell marker, but not for neuronal, astrocytic or microglial

markers. In conclusion, transplanted NPCs survive in the host brain up to 14 months, migrate to the site of injury, enhance motor and

cognitive recovery, and may play a role in trophic support following TBI.

D 2004 Elsevier B.V. All rights reserved.

Theme: Disorders of the nervous system

Topic: Transplantation

Keywords: Traumatic brain injury; Rotorod; Spatial learning; Morris water maze; Neural stem cell; NG2; Oligodendrocyte progenitor cell; Neurosphere

0006-8993/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.brainres.2004.07.087

* Corresponding author. Brain Research Laboratory, Evans Building

Room 261B, 1648 Pierce Drive, Emory University, Atlanta, GA 30322,

USA. Tel.: +1 404 712 2540; fax: +1 404 727 2388.

E-mail address: dstei04@emory.edu (D.G. Stein).1 We wish to acknowledge that the first and second authors share equal

credit for work completed in this study.

1. Introduction

Traumatic brain injury (TBI) is a significant clinical

problem in the United States, yet few effective strategies for

treating it have emerged [38]. The disappointing outcomes

of numerous clinical trials examining pharmacological

treatments after TBI [8] may be due to a number of

complex secondary events subsequent to the initial trau-

(2004) 11–22

D.A. Shear et al. / Brain Research 1026 (2004) 11–2212

matic insult, including elevated inflammation, excitotox-

icity, demyelination and ischemia. These secondary modes

of damage have been shown to contribute to delayed cell

death and prolonged functional deficits [14,25,33]. Thus, a

long-term strategy may be required to effect a clinically

significant solution. In this context of long-term treatment

after TBI, strategies incorporating sustained release of

trophic factors and encapsulated cell therapy have demon-

strated feasibility [2,21], but may face clinical limitations

such as volume constraints and the inability of donor cells to

interact with the host tissue. As an alternative to pharmaco-

logical strategies and encapsulated cell therapies, direct

transplantation of neural cells into the injured brain may

provide for long-term survival and integration that could

mediate functional recovery through mechanisms such as

bulk trophic support, cell–cell mediated repair and replace-

ment of cells lost by injury.

The success of neural transplantation fundamentally

hinges on the choice of cell type. Autologous primary cells

are perhaps the most attractive choice from an immuno-

logical standpoint, but are limited with regard to availability

and capacity to generate enough cells for therapy. Many cell

lines have been generated to provide a virtually infinite

supply of cells for research and clinical therapies [10,13],

but cells lines are associated with concerns about tumoro-

genesis and potential immune reactions.

Neural progenitor cells (NPCs) present an interesting

approach for mediating repair and rescue of the host tissue

after injury, because it is relatively easy to generate large

quantities of cells in vitro, and because of their inherent

ability to adapt to signals from host cells and the

extracellular environment [12,15,18,26,42]. Further, the

multipotential nature of NPCs may be of particular

importance after trauma, where the introduction of multiple

cell types may be necessary for repair and regeneration of

injured tissue [20,28,36]. Recent studies indicate that

transplanted NPCs may respond to signals present in the

injured brain by differentiating into neurons and glia

according to transplant location [3,5], and it has been

demonstrated that these different phenotypes can promote

repair and restoration of function after TBI [23,37].

However, the functional mechanisms underlying trans-

plant-mediated recovery are not clear and the long-term

effects of these NPC transplants following TBI remain

unknown. Thus, the objective of the present study was to

evaluate the long-term survival, migration, differentiation

and functional significance of FGF2-responsive NPCs

transplanted into a mouse model of TBI.

2. Materials and methods

2.1. Subjects

Male C57BL.6J mice (8 weeks; Jackson, Bar Harbor,

ME) were housed individually in plastic cages and kept on a

12-h light–dark cycle (lights on from 0700–1900 h). Food

and water were available ad libitum. Thirty-five animals

were randomly assigned to the following five groups: (1)

injury, no cells (INJURY, n=6), (2) injury+vehicle (INJ/

VEH, n=6), (3) injury+neural progenitor cells (INJ/NPC,

n=7), (4) craniectomy, no cells (CRAN, n=6) and (5) sham

operation, no cells (SHAM, n=6). Behavioral testing was

conducted during the first 2 h of the nocturnal cycle (i.e.,

1900–2100 h). All procedures involving animals conformed

to guidelines set forth in the Guide for the Care and Use of

Laboratory Animals (U.S. Department of Health and

Human Services, Pub no. 85-23, 1985) and were approved

by the Emory University Institutional Animal Care and Use

Committee.

2.2. Experimental injury model

Unilateral contusions of the lateral frontoparietal cortex

were created using a controlled cortical impact (CCI)

device as previously described [38]. Mice were anesthe-

tized with ketamine/xylazine (90 and 10 mg/kg) and

mounted in a stereotaxic device. Using aseptic techniques,

a sagittal incision was made in the scalp and the fascia

retracted to expose the cranium. A 4-mm craniectomy was

made over the left frontoparietal cortex with a standard

tissue biopsy punch (center: �1.0 mm AP, +2.0 mm ML

from bregma). After removal of the bone, the injury was

produced by activating a pneumatic piston (tip diameter=3

mm) positioned 158 from vertical in the coronal plane to

achieve perpendicular impact of the cortical tissue to a

depth of 1 mm (velocity=6 m/s, duration=150 ms).

Duration and velocity were verified in real time via a

linear variable differential transformer (Macrosensors,

Pennsauken, NJ). Following the injury, the wound cavity

was thoroughly cleaned and all bleeding stopped before the

fascia and scalp were sutured closed. In all experiments,

animals were coded for surgery and treatment to prevent

experimenter bias during behavioral testing and histological

examination.

2.3. Isolation of neural progenitor cells

Pregnant B6-TgN (h-act-EGFP) osbY01 mice (a gift

from Dr. Masaru Okabe, Osaka University, Osaka, Japan)

were sacrificed and embryos were isolated at gestational

day 14.5 by caesarian section. Following removal of the

skull layer, parasagittal cuts were made in each hemi-

sphere. The ganglionic eminence was carefully separated

from the underlying tissue and mechanically dissociated

into a single cell solution in cold Hank’s balanced salt

solution (HBSS, pH 7.40, Sigma, St. Louis, MO). Cells

were maintained in suspension culture in neurosphere

medium (serum-free DMEM/F12 medium containing

insulin [25 Ag/ml], transferrin [100 Ag/ml], putrescine [60

AM], sodium selenite [30 nM], progesterone [20 nM] and

glucose [0.6%]). Human recombinant FGF2 (20 ng/ml)

D.A. Shear et al. / Brain Research 1026 (2004) 11–22 13

was added every 3–4 days to maintain the cells as

proliferating spheres that were passaged every 7–10 days,

as density merited.

2.4. Transplantation surgery

One week following CCI injury, all mice were reanes-

thetized using ketamine/xylazine (90 and 10 mg/kg) and

placed in a stereotaxic device. NPCs (passage 5 or lower)

were harvested from suspension culture and resuspended in

neurosphere medium (without FGF2) for transplantation.

Approximately 100,000 cells were stereotactically trans-

planted as neurospheres into the striatum ipsilateral to the

injury (coordinates: AP +1.0 mm, ML +1.5 mm from

bregma, DV �2.5 mm from dura) via a 10-Al Hamilton

syringe (26-gauge needle). Neurospheres were injected in 3

Al of medium and injections were driven by a syringe pump

(Sage Instruments M362, Cambridge, MA), which was

calibrated to achieve a constant flow rate of 0.25 Al/min.

INJ/VEH animals received injections of medium only.

Following all injections, the needle was left in place for 5

min before being slowly withdrawn.

2.5. Behavioral testing

2.5.1. Rotorod

A rotorod apparatus (Columbus Instruments, Columbus,

OH) was used to measure motor coordination and balance.

Prior to surgeries, mice were trained on the rotorod at a

constant speed (10 rpm) for a maximum of 60 s, and the

latency to fall off the rotorod within this time period was

recorded. Mice received four training trials per day with a

60-s inter-trial interval for 3 consecutive days to obtain a

baseline level of performance.

Following CCI injury, motor abilities were assessed on

the rotorod at 1 day prior to NPC transplant, and at 1, 2, 9,

12, 20 and 48 weeks post-transplant. On testing days, the

mice received two trials at each of four speed levels (5, 10,

15 and 20 rpm), and the mean latency to fall off the rotorod

for the two trials at each speed level was recorded and used

in subsequent analysis.

2.5.2. Morris water maze

Mice were tested for spatial learning in a Morris water

maze (MWM) paradigm at three different time points (1, 3

and 12 months) over a 1-year post-transplant period. The

water maze apparatus used in this task consisted of a white,

circular tank (90 cm deep, 133 cm diameter) filled with

opaque water (20F1 8C, nontoxic Artistak white tempera

paint) to a depth of 64 cm. A submerged white platform was

placed in the northeast quadrant. During each of the three

testing periods (i.e., 1, 3 and 12 months), animals were given

two trials per day for 14 consecutive days, with a 30-min

inter-trial interval. At the start of each trial, the subject was

placed in the pool (snout facing the pool wall) at one of four

randomly determined starting positions (north, east, south,

west). The mouse was allowed to swim freely in the pool

until it located the hidden platform or until 90 s had elapsed.

If the mouse failed to find the platform within 90 s, it was

manually guided through the water and placed on the

platform. At the end of each trial, the mouse was left on

the platform for 20 s. The latency to escape the water and the

distance swum were detected and recorded for each trial

using motion detection hardware/software (San Diego

Instruments, San Diego, CA). Average swim speed measures

were also calculated from the final trial of each testing period

(1, 3 and 12 months post-transplant) in order to differentiate

motor dysfunction from learning impairment.

At 1 month post-transplant, mice were tested on a

missing platform probe trial to assess exploratory behavior.

The test consisted of removing the platform and recording

the swim path of each mouse over a 60-s period while it

searched for the missing platform. The distance swum,

time spent in each quadrant and total path length were

recorded.

One day after the missing platform probe trial, a visible

platform MWM test was used to assess whether spatial

learning deficits detected in injured mice could be attributed

to impairments in escape motivation, or visual and/or motor

abnormalities. For this test, the white platform was replaced

with a black one (same location) raised 0.5 cm above water

level, and latency to locate the visible platform was recorded

in a single 60-s trial.

2.6. Histology

At 14 months post-transplant, animals were deeply

anesthetized and transcardially perfused with 10 ml 0.1

M phosphate-buffered saline (PBS, pH 7.4) followed by 20

ml 4% paraformaldehyde in 0.1 M phosphate buffer (PB,

pH 7.4). Brains were removed, post-fixed in paraformalde-

hyde for 12 h, and cut in 50-Am coronal sections on a

vibratome. The sections were immediately processed or

kept at 4 8C in cryoprotectant medium (30% sucrose, 30%

ethylene glycol and 1% PVP-40 in 0.1 M PBS, pH 7.40).

The first series of sections was mounted on gel-coated

slides and stained for Nissl bodies with thionine to

determine placement and extent of the injury. Remaining

series were processed for immunofluorescent detection of

specific phenotypic markers.

2.6.1. Immunofluorescence

For fluorescent antibody labeling, free-floating sections

were preincubated in 0.1 M TBS (pH 7.4) containing 20%

normal serum and 0.5% TritonX-100 for 2 h at 4 8C.Sections were then incubated in the primary antibodies

overnight at 4 8C. All primary antibodies were diluted in 0.1

M TBS containing 20% normal serum. For double or triple

labeling, sections were incubated in primary antibodies from

different species simultaneously and in secondary antibodies

sequentially. Primary antibodies used included markers for

GFP (mouse and rabbit anti-GFP, Molecular Probes,

D.A. Shear et al. / Brain Research 1026 (2004) 11–2214

Eugene, OR, and chicken anti-GFP, Chemicon, Temecula,

CA), mature neurons (NeuN and neurofilament 70, mono-

clonal, Chemicon), astrocytes (GFAP, monoclonal, Chem-

icon), macrophage/microglia (F480, monoclonal, Serotec,

Raleigh, NC) mature oligodendrocytes (CNPase, polyclo-

nal, Chemicon) and oligodendrocyte progenitor cells

(OPCs) (NG2, polyclonal, Chemicon). Following primary

antibody incubation, sections were exposed to the appro-

priate fluorophore-conjugated secondary antibody solution

for 90 min at 4 8C. Sections were rinsed in TBS, mounted

on gelatin-coated slides, coverslipped with Vectashield

mounting medium (Vector, Burlingame, CA), and evaluated

via confocal microscopy (Zeiss LSM 510, Stuttgart,

Germany) for survival and phenotypic differentiation.

2.6.2. Lesion reconstruction and NPC quantification

Lesion volume and both ipsilateral and contralateral

hippocampal formation volumes were quantified from five

evenly spaced coronal sections ranging from �0.70 to

�2.30 mm AP (relative to bregma). The perimeters of the

necrotic cavity and the borders of the ipsilateral and

contralateral HC were traced on digitized images using a

CoolSNAP-Pro camera (Roper Scientific PVCAM 4.5

capture program, Media Cybernetics, Carlsbad, CA) con-

nected to an Olympus BH-2 microscope (Olympus America,

Melville, NY). Area measurements of outlined brain regions

were computed using Image Pro Plus software (version

4.5.1.22, Media Cybernetics) and converted to total lesion

and hippocampal volume estimates (mm3) via integration of

lesion area versus distance curves.

Surviving GFP+ NPCs were counted using confocal

microscopy over a coronal series spanning approximately

+1.00 anterior to �3.20 posterior to bregma, with sections

spaced 200 Am apart. For each brain section quantified, a

single image was captured (20� magnification) and the total

number of GFP+ cells, and the location of each cell (cortex,

HC or other) were recorded from each image (over 99% of

all surviving GFP+ cells were in the cortex or HC). In

addition, the percentage of total, cortical, or hippocampal

GFP+ cells that co-labeled with phenotypic markers was

quantified.

2.7. Statistical analysis

Unless otherwise noted, all histological and behavioral

data were analyzed using either one-way or mixed-

factorial analysis of variance (ANOVA) followed by Fisher

PLSD post-hoc tests ( pb0.05). The Pearson r correlation

coefficient ( pb0.05) was calculated to assess potential

linear relationships between NPC survival and behavioral

parameters.

No significant differences were detected between CRAN

and SHAM-operated animals on any behavioral or histo-

logical parameter. Animals in these two groups were

therefore pooled together, and seven were randomly selected

to serve as a CONTROL (n=7) group for statistical analysis.

3. Results

3.1. Behavior

3.1.1. Rotorod

Functional recovery of motor behavior mediated by the

transplants was assessed using a rotorod task. Since our goal

was to assess potential NPC-mediated effects for 1 year

following transplantation, testing dates were spaced farther

apart as the study progressed to minimize potential

acclimation effects and optimize task sensitivity. Based on

pilot work done in our laboratory, we chose a fixed-speed

version of rotorod testing. In the present study, this protocol

resulted in sensitive measures of sustained motor impair-

ment in CCI-injured mice out to 1 year post-injury.

Prior to surgery, all mice showed learning of the rotorod

test and reached a stable baseline level of performance

within 3 days, as measured by an increase in the mean

duration to maintain balance on the rotorod (Fig. 1).

Throughout the study, CONTROL mice consistently main-

tained balance on the rotorod for the maximum latency of 60

s on all rotation speeds, except 20 rpm, when they would

occasionally fall off. Conversely, all injured groups dis-

played significant motor impairment at 10 rpm (F3,22=3.32,

pb0.05), 15 rpm (F3,22=12.57, pb0.05) and 20 rpm

(F3,22=13.00, pb0.05) when re-tested one day prior to

NPC-transplant surgery (6 days post-injury). There was no

overall difference between groups at 5 rpm at this time point

(F3,22=2.12, pN0.05).

Following transplantation surgery, a significant differ-

ence between groups at each speed was confirmed by a

mixed-factorial ANOVA (4 groups�6 post-transplant trial

blocks), which revealed a significant main effect at 5 rpm

(F3,22=5.20, pb0.05), 10 rpm (F3,22=4.50, pb0.05), 15 rpm

(F3,22=16.70, pb0.05) and 20 rpm (F3,22=15.20, pb0.05).

Overall, NPC-treated mice performed significantly better on

this task than did either the INJURY or INJ/VEH groups,

and the INJ/VEH group displayed the greatest motor

impairment (Fig. 1).

Notably, at 15 rpm, the NPC-mediated recovery on the

rotorod was observed as early as 1 week post-transplant and

was sustained out to 1 year post-transplant (Fig. 1).

Moreover, at this speed, with the exception of the 9-week

time point, the NPC mice were indistinguishable statistically

from the SHAM group from 1 week to 1 year post-

transplant. However, NPC animals did not fully perform to

control levels, since improvement in motor abilities was not

present when animals were tested at 20 rpm (Fig. 1).

3.1.2. Morris water maze

Potential NPC-mediated effects on injury-induced spa-

tial learning deficits were assessed in the MWM task. We

used a testing paradigm that called for two trials per day

for 14 consecutive days per testing period in order to

optimize task sensitivity, and we recorded both swim

speeds (latencies) and the distance taken to find the

D.A. Shear et al. / Brain Research 1026 (2004) 11–22 15

platform. Overall, CCI injury produced sustained cognitive

impairment in this task at 1 month (F3,22=3.90, pb0.05), 3

months (F3,22=3.90, pb0.05) and 12 months (F3,22=5.35,

pb0.05) post-transplant. NPC transplants facilitated spatial

learning, albeit not to control levels, during the initial

MWM testing at 1 month post-transplant, and significantly

improved retention on this task at 3 months (Fig. 2). At 12

months post-transplant, INJ/NPC animals continued to

perform significantly better than the non-transplanted

INJURY animals.

At 1 month post-transplant, the platform was removed

after the last trial and a probe test was performed to assess

NPC-mediated effects on exploratory behavior. Quantifica-

tion of the duration of exploratory time spent in the quadrant

that previously contained the hidden platform revealed that

while the uninjured CONTROL group spent the majority of

the time in exploratory behavior, the non-transplanted injury

groups spent significantly less time exploring this quadrant

(F3,22=11.71, pb0.05) (Fig. 2). Further, injured animals

receiving NPC transplants demonstrated a significant

attenuation of injury-induced exploratory behavior impair-

ment ( pb0.05).

Following the 1-month testing period, the visible

platform MWM test was used to assess whether cognitive

deficits in injured mice were a result of impaired attention,

vision, or motor performance. As illustrated in Fig. 2,

injured and control groups did not differ in overall level of

performance (F3,22=5.35, pN0.05), indicating that injury-

induced deficits in the MWM test were not due to

nonspecific sensory impairment but were indeed a result

of cognitive impairment. Further support for this finding

comes from the observation that average swim speeds and

swim distances calculated for the final day of MWM

testing at 1, 3 and 12 months post-transplant were not

significantly different among the experimental groups

( pN0.05).

3.2. Histology

The CCI injury model produced consistent lesions of

the left parietotemporal cortex extending from approx-

imately 0.7 mm anterior to 3.0 mm posterior from bregma

(Fig. 3A). Results of paired sample t-tests comparing the

ipsi- and contralateral HC revealed significant ipsilateral

HC degeneration in all injured groups ( pb0.05, Fig. 3B).

NPC transplants did not affect either necrotic cavity size

(Fig. 3B) or HC degeneration (Fig. 3C) despite substantial

NPC migration and survival throughout the injured regions

(Fig. 4).

Fig. 1. Rotorod performance. Following training and prior to any surgery

(PRE-INJ), all mice used in this study displayed a stable, baseline level of

performance on the rotorod, as quantified by latency to fall off of the rotorod.

Further, all injured groups showed similar motor deficits one day prior to

transplant surgery (PRE-TP). Although treatment effects were observed at 5

and 10 rpm, of particular interest is the observation that at 15 rpm the NPC-

mediated recovery on the rotorod was observed as early as 1 week post-

transplant and was sustained out to 1 year post-transplant. Moreover, at this

speed, with the exception of the 9-week data, the NPC-treated mice are

indistinguishable statistically from the sham group from 1 week to 1 year

post-transplant. However, note that NPC-transplanted animals did not

perform to control levels since improvement in motor abilities was not

present when animals were tested at 20 rpm. Data points represent means;

error bars=S.E.M.; *pb0.05 CONTROL compared to INJURY group;ypb0.05 compared to INJ/VEH; zpb0.05 compared to INJ/NPC.

Fig. 2. Morris water maze performance. Overall, cognitive impairment in both INJURYand INJ/VEH-injected mice was sustained out to 1 year post-transplant.

However, the INJ/NPC group demonstrated enhanced spatial learning relative to the INJURYand INJ/VEH groups during initial acquisition of MWM testing at

1 month post-transplant and further improved retention of this task at 3 months and 1 year post-transplant. At 1 month post-transplant, probe test analysis

confirmed impairment of spatial cognition in both INJURY and INJ/VEH mice (e.g. spent less time searching the quadrant in which the platform had been

submerged during training), whereas NPC-treated mice showed significant improvement in this task as compared to the INJURY group. In contrast, all groups

performed equally well in locating the visible platform on day 15, indicating that observed cognitive dysfunction was not resulting from visuospatial

abnormalities. Error bars represent S.E.M.; *pb0.05 compared to CONTROL; ypb0.05 compared to INJ/NPC.

D.A. Shear et al. / Brain Research 1026 (2004) 11–2216

At 14 months post-transplant, the NPC transplants

appeared to be well tolerated by the host brain and did

not produce morphological or behavioral abnormalities. The

majority of surviving GFP+ transplanted cells at 1 year post-

transplant were homogeneously distributed throughout the

injured HC and surrounding white matter tracts and

exhibited a highly ramified morphology (Fig. 4A,D). Sparse

numbers of GFP+ cells were occasionally observed in other

Fig. 3. Lesion reconstruction at 14 months post-transplant. The controlled cortical impact model produced consistent necrosis of the left parietotemporal cortex

extending from approximately +0.70 mm AP to �3.00 mm AP from bregma (Fig. 3A). NPC transplants did not affect necrotic cavity size (B) or hippocampal

(HC) degeneration (C; shows comparison between ipsilateral and contralateral HC; *pb0.05 compared to intact hemisphere).

D.A. Shear et al. / Brain Research 1026 (2004) 11–22 17

areas, such as the periventricular region of the lateral

ventricles and subcortical brain regions. No evidence of

abnormal proliferation or tumor formation was associated

with transplants at 14 months. Interestingly, when NPC

transplants were performed in healthy adult brains and

followed over a period of 6 months, the cells did not migrate

significantly, but remained as a dense deposit within the

striatum surrounding the injection needle track (Fig. 4E).

There was a moderate correlation between the location

of surviving NPCs and overall MWM performance at

both 1 (r7=�0.80, pb0.05) and 3 months (r7=�0.74,

pb0.05) post-transplant. Specifically, at these two time

points we observed an inverse relationship between

percentage of total NPCs in the HC and latency to find

the MWM platform, suggesting that increased NPC

survival in the HC is related to improved acquisition

and retention of this spatial learning task (Fig. 4B,C).

There was no correlation between remaining NPCs and

performance on the rotorod task.

The majority of the GFP+ transplanted cells exhibited

multiple branched processes with small cell bodies (~10–15

Am in diameter) and did not express antigens associated

with neurons, astrocytes, microglia or mature oligodendro-

cytes (Fig. 5). Rather, the majority of surviving GFP+ NPCs

(84+4%) expressed NG2, a cell surface proteoglycan found

on adult OPCs (Fig. 6). Notably, a number of host cells in

the brains of injured animals, distinguished from trans-

planted cells by the lack of GFP expression, also expressed

NG2, suggesting that the transplanted NPCs had merged

with the host glial response to injury (Fig. 6A–F).

4. Discussion

An important feature of any potential therapy for TBI is

the improvement of functional outcomes following injury.

In the present study, we first assessed motor function and

showed that (1) our CCI injury model results in motor

deficits sustained out to 1 year post-injury and (2) NPC

transplants enhanced recovery from injury-induced motor

impairment. NPC-mediated motor recovery was observed as

early as 1 week and sustained up to 1 year post-transplant,

indicating a long-term effect of the grafts on functional

recovery. (3) Despite the 14-month survival of some NPC

cells, there was no evidence of extensive proliferation or

tumor formation during this period of time. These results

extend previous findings that have shown rapid, albeit

incomplete, motor recovery during the initial few weeks

following transplantation of immortalized or stromal cell

transplants following TBI [32,37,38,40,42].

Fig. 4. NPC distribution at 14 months post-transplant. GFP+ transplanted cells were distributed primarily throughout the injured HC and adjacent cortical areas

(A). Moderate correlations were detected between the location of surviving NPCs and overall MWM performance at 1 and 3 months post-transplant (B and C,

respectively) suggesting that increased NPC survival in the HC is related to improved spatial learning performance in the MWM. Confocal micrograph showing

distribution of NPCs in the HC of an injured animal at 14 months post-transplant (D) and in the striatum of an uninjured animal at 6 months post-transplant (E).

Note that, in the absence of injury-induced cues, the transplanted NPCs did not migrate away from the striatal site of implantation. Inset shows a primary GFP+

neurosphere in culture prior to transplantation (scale bar=100 Am). HC=hippocampus, Fi=fimbria, STR=striatum; scale bars=100 Am (D) and 20 Am (E).

D.A. Shear et al. / Brain Research 1026 (2004) 11–2218

The present study provides a clear demonstration that

primary NPC transplants can accelerate the recovery of

spatial performance and cognitive performance following

TBI. We also observed a correlation between spatial

learning performance and the location of surviving NPCs

at 14 months post-transplant, indicating that migration of

NPCs into the ipsilateral hippocampus may be required for

cognitive improvement. Our previous work investigating

the time course of migration of transplanted NPCs in the

injured brain revealed that at 1 week post-transplant, cells

were primarily in the striatum, an area associated with

motor coordination, and 3 months later cells had migrated

into the injured HC [40], an area implicated in spatial

learning. These previous neuroanatomical data, along with

the functional data in the present study demonstrating

motor recovery at 1 week and cognitive recovery at 3

months, suggest a relationship between the spatial location

of transplanted cells and improvements in functional

modalities associated with the occupied region.

Immunohistochemical analysis of transplanted brains at 1

year post-transplant revealed that the majority of the trans-

planted GFP+ NPCs showed co-expression of NG2, a cell

surface proteoglycan found on adult OPCs. NG2+ OPCs

comprise approximately 5–8% of the adult CNS glial

population [13] and are antigenically and morphologically

distinct from neurons, astrocytes, oligodendrocytes, micro-

glia and macrophages. Adult OPCs have been shown to

contact neurons and axons at synapses and nodes of Ranvier,

respectively, and interact with astrocytes and macrophages

within the glial scar at lesion sites, suggesting that they are

specialized to monitor and respond to changes in the integrity

of the CNS [9,11]. Thus, adult OPCs may represent a novel

glial cell population in the adult CNS [30,31].

In the present study, host NG2+ glia were present in the

uninjured brains, where they showed slightly higher levels

of NG2 expression in areas lining the lateral ventricles and

in the HC [4]. In response to CCI injury, host NG2+ cells

were observed in the injured HC and adjacent cortical areas,

Fig. 5. Immunohistochemical evaluation of GFP+ NPCs in the injured HC at 14 months post-transplant. Horizontal rows represent individual and merged

images taken from the same tissue section. Confocal micrographs showing labeling for GFAP (A), NeuN (B), Neurofilament (E), CNPase (H), GFP (C, F, I)

and merged images (D, G, J) demonstrate that transplanted cells did not differentiate into mature neurons, astrocytes or oligodendrocytes. Scale bars=50 Am(A–D) and 20 Am (E–J).

D.A. Shear et al. / Brain Research 1026 (2004) 11–22 19

where they sustained elevated levels of NG2 expression out

to 14 months post-injury. These findings in the context of

TBI are consistent with previous work demonstrating NG2+

cell proliferation in the contused spinal cord [16,27] and

Fig. 6. Phenotype of GFP cells. At 14 months post-transplant, approximately 84%

a cell surface proteoglycan found on adult OPCs. Confocal micrographs at low (A

NPCs (B, E) and merged images (C, F). Scale bars=100 Am (A–C) and 20 Am (

suggest that a response initiated by host NG2+ cells may be

an integral component of the response to brain injury. In the

current study, the spatial distribution within the glial scar,

the retention of NG2+ expression for an extended time after

(F4%) of surviving GFP+ NPCs showed co-expression of the NG2 antigen,

–C) and high (D–F) power magnification show NG2+ cells (A, D), GFP+

D–F).

D.A. Shear et al. / Brain Research 1026 (2004) 11–2220

injury (14 months post-transplant), the lack of differ-

entiation of transplanted cells into CNPase+ adult oligoden-

drocytes, and the demonstration of long-lasting functional

recovery, suggest that in our model of brain injury the NPC

transplants merged with and aided host glial-mediated repair

processes. The possibility that transplanted cells expressing

NG2 may represent latent OPCs or other glial cell types

cannot be excluded [16], and future studies elucidating the

distribution and functional role of NG2-expressing cells in

the intact and injured CNS may yield further insight into

transplant-mediated functional recovery following TBI.

Our finding that approximately 85% of surviving GFP+

NPCs differentiated along the NG2+ lineage was surprising

given the multipotential capacity of this population of NPCs

in vitro [40]. One explanation could be that primary NPCs

derived from the ganglionic eminence have a propensity to

differentiate along the NG2+ lineage when transplanted

within the striatum. Evidence for this hypothesis comes

from the observation that one subpopulation of early OPCs

originates in the ganglionic eminence, which gives rise to

the striatum and basal ganglia in the rodent CNS [35].

Conversely, it has been demonstrated that primary NPCs are

sensitive and respond to signals within the host environ-

ment, and that upregulation of NG2 in response to CNS

injury occurs as early as 3 days, and peaks at 7 days, post-

injury [21]. Accordingly, since NPCs were transplanted into

injured brains during a time of peak NG2 expression (7 days

post-injury), we theorize that the transplanted NPCs may

augment endogenous repair mechanisms by merging with

the host glial response to injury. We also observed that

NPCs transplanted into uninjured brains did not proliferate

or migrate significantly (up to 6 months post-transplant),

which supports the idea that the appearance of transplanted

cells as NG2+ glia in the injured HC is not the result of an

intrinsic NPC property or preferential path of migration, but

a capacity of the transplanted cells to detect and respond to

the dynamic environment of the injured brain. Nevertheless,

comparing different timing and/or locations of post-injury

NPC transplantation could provide important insights

regarding specific factors that influence the phenotypic

differentiation and migratory behavior of these cells.

The mechanisms underlying the beneficial effects of

transplanted NPCs on motor and cognitive behavior in this

study may involve the release of neurotrophic factors by the

transplanted cells [6,17,19,23,24,26,29,39] and/or OPC-

mediated remyelinating mechanisms [22,27,34,36,41,43].

Our finding that NPC transplants are capable of mediating

motor recovery as early as 1 week after implantation,

coupled with the lack of CNPase (a marker of myelinating

cells) expression in transplanted cells at 1 year post-

transplant, provides strong support for a neurotrophic

mechanism of action, though the possibility of transplanted

cells contributing indirectly to host remyelination efforts

remains a possibility. Other researchers have also demon-

strated that cell transplants into the brain can enhance

functional outcomes through their production of trophic

factors instead of through the re-formation of specific

neuronal connections. For example, Borlongan et al. [7]

recently report that peripherally injected human umbilical

cord blood cells, which enter the brain but do not cross the

blood–brain barrier, increased brain levels of neurotrophic

factors and reduced the size of the cerebral infarcts caused

by middle cerebral artery occlusions.

Independent of the specific mechanisms of action

promoted by transplanted NPCs, a number of observations

made in the present study have important implications for

transplantation strategies aimed at treating CNS injury. First,

the ability of transplanted primary NPCs to survive,

consistently migrate to specific loci, and promote functional

recovery out to 1 year post-transplant without evidence of

tumor formation, suggests that long-term transplantation of

multipotential cells, cultured and expanded in vitro, may

present a safe and reproducible therapy for treating CNS

damage. Second, the ability to achieve sustained multi-

modal functional recovery without introduction or replace-

ment of neurons suggests that strategies aimed at neuronal

replacement may not be critical or sufficient for promoting

sustained functional recovery following complex, chronic

neurological disorders such as TBI. At least for traumatic

brain injuries, the research from our laboratory and others

suggests that the pre- or post-transplantation differentiation

of progenitor cells into neuronal phenotypes is not a

prerequisite to obtain functional recovery. Thus, approaches

such as pharmacological intervention, encapsulated cell

systems, genetically modified cells, or glial transplants may

indeed have merit for potential therapies in brain and spinal

cord injury. In addition, the ability of transplanted NPCs to

migrate long distances to areas of damage (a feature of

NPCs that has been demonstrated previously [1]) may prove

to be particularly attractive for treating CNS disorders where

the damage is diffuse or not amenable to direct delivery of

transplants. Finally, the observation that transplanted NPCs,

though capable of forming all the major CNS cell

populations, may uniformly adapt to and merge with the

normal host response to injury, demonstrates another

important capacity of NPCs that may be taken advantage

of in future neural transplantation strategies.

With regard to future clinical applications of cell trans-

plantation into the CNS, one of the most important issues is

cell source and preparation. Xenogeneic sources have the

advantage of being able to provide large quantities of cells,

but there are concerns of immune rejection and transmission

of zoonotic diseases. Conversely, allogeneic or autologous

cell sources are more desirable from an immunologic

perspective but may not provide adequate cell numbers to

be clinically useful for neurologic applications. One

potential solution to the problem of cell availability of

autologous or allogeneic transplants is in vitro expansion of

cells in culture prior to transplantation. In vitro expansion

offers the potential advantages of an almost infinite supply

of cells and the ability to manipulate cells genetically or

phenotypically in a controlled manner, two important facets

D.A. Shear et al. / Brain Research 1026 (2004) 11–22 21

of developing any potential large-scale clinical applications.

Potential disadvantages of culturing cells prior to trans-

plantation include increased cost, delayed time to transplant

and the possibility of altering cell function (e.g., migration,

differentiation, etc.) such that transplant function may not be

optimal. The current study provides evidence that expansion

of cells prior to transplantation within a controlled in vitro

environment may provide a safe and reliable method of cell

preparation prior to transplantation, and that cultured NPCs

can produce long-term functional benefits in the injured

CNS without tumor formation.

In summary, our data suggest that NPCs transplanted into

an intact region of an injured brain survive at least 1 year

post-transplant, migrate to areas of damage and produce no

overt behavioral or morphological abnormalities such as

tumor formation. Further, we demonstrate that NPC trans-

plants can promote multi-modal functional recovery after

injury and that the recovery is not a consequence of direct

replacement of injured neurons, but rather suggests a

mechanism of repair involving trophic support. Better

understanding of the specific mechanisms of NPC trans-

plant-mediated repair may help to develop novel strategies

for treating TBI and other complex CNS disorders.

Acknowledgments

We would like to thank Dr. Masaru Okabe (Osaka

University, Osaka, Japan) for the generous donation of

EGFP transgenic mice, and Dr. Howard Reese (Dept. of

Neurology, Emory University) for assistance with confocal

microscopy. Funding for this study was provided by

Research Gifts from Field Neurosciences Institute and

Genre (to DGS), Georgia Tech/Emory Biotechnology

Research Center, and a NSF GRF (to DAS).

References

[1] K.S. Aboody, A. Brown, N.G. Rainov, K.A. Bower, S. Liu, W. Yang,

J.E. Small, U. Herrlinger, V. Ourednik, P.M. Black, X.O. Breakefield,

E.Y. Snyder, Neural stem cells display extensive tropism for

pathology in adult brain: evidence from intracranial gliomas, Proc.

Natl. Acad. Sci. U.S.A. 97 (2000) 12846–12851.

[2] P. Aebischer, M. Goddard, A.P. Signore, R.L. Timpson, Functional

recovery in hemiparkinsonian primates transplanted with polymer-

encapsulated PC12 cells, Exp. Neurol. 126 (1994) 151–158.

[3] Y. Benninger, S. Marino, R. Hardegger, C. Weissmann, A. Aguzzi,

S. Brandner, Differentiation and histological analysis of embryonic

stem cell-derived neural transplants in mice, Brain Pathol. 10 (2000)

330–341.

[4] D.E. Bergles, J.D. Roberts, P. Somogyi, C.E. Jahr, Glutamatergic

synapses on oligodendrocyte precursor cells in the hippocampus,

Nature 405 (2000) 187–191.

[5] L.M. Bjorklund, R. Sanchez-Pernaute, S. Chung, T. Andersson, I.Y.

Chen, K.S. McNaught, A.L. Brownell, B.G. Jenkins, C. Wahlestedt,

K.S. Kim, O. Isacson, Embryonic stem cells develop into functional

dopaminergic neurons after transplantation in a Parkinson rat model,

Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 2344–2349.

[6] C.V. Borlongan, C.D. David, P.R. Sanberg, CNS entry of peripherally

injected umbilical cord blood cells is not required for neuroprotection

in stroke, Stroke (2004) (in press).

[7] C.V. Borlongan, Y. Tajima, J.Q. Trojanowski, V.M. Lee, P.R. Sanberg,

Transplantation of cryopreserved human embryonal carcinoma-

derived neurons (NT2N cells) promotes functional recovery in

ischemic rats, Exp. Neurol. 149 (1998) 310–321.

[8] M.R. Bullock, B.G. Lyeth, J.P. Muizelaar, Current status of neuro-

protection trials for traumatic brain injury: lessons from animal

models and clinical studies, Neurosurgery 45 (1999) 207–217

discussion 217–220.

[9] A.M. Butt, J. Kiff, P. Hubbard, M. Berry, Synantocytes: new

functions for novel NG2 expressing glia, J. Neurocytol. 31 (2002)

551–565.

[10] L.S. Chen, J. Ray, L.J. Fisher, M.D. Kawaja, M. Schinstine, U.J.

Kang, F.H. Gage, Cellular replacement therapy for neurologic

disorders: potential of genetically engineered cells, J. Cell. Biochem.

45 (1991) 252–257.

[11] M.R. Dawson, J.M. Levine, R. Reynolds, NG2-expressing cells in the

central nervous system: are they oligodendroglial progenitors? J.

Neurosci. Res. 61 (2000) 471–479.

[12] F.H. Gage, Mammalian neural stem cells, Science 287 (2000)

1433–1438.

[13] F.H. Gage, L.J. Fisher, Intracerebral grafting: a tool for the neuro-

biologist, Neuron 6 (1991) 1–12.

[14] T.A. Gennarelli, Mechanisms of brain injury, J. Emerg. Med. 11

(Suppl. 1) (1993) 5–11.

[15] J.A. Gray, H. Hodges, J. Sinden, Prospects for the clinical application

of neural transplantation with the use of conditionally immortalized

neuroepithelial stem cells, Philos. Trans. R. Soc. Lond., B Biol. Sci.

354 (1999) 1407–1421.

[16] L.L. Jones, Y. Yamaguchi, W.B. Stallcup, M.H. Tuszynski, NG2 is a

major chondroitin sulfate proteoglycan produced after spinal cord

injury and is expressed by macrophages and oligodendrocyte

progenitors, J. Neurosci. 22 (2002) 2792–2803.

[17] J.P. Kesslak, M. Nieto-Sampedro, J. Globus, C.W. Cotman, Trans-

plants of purified astrocytes promote behavioral recovery after frontal

cortex ablation, Exp. Neurol. 92 (1986) 377–390.

[18] V.G. Kukekov, E.D. Laywell, O. Suslov, K. Davies, B. Scheffler, L.B.

Thomas, T.F. O’Brien, M. Kusakabe, D.A. Steindler, Multipotent stem/

progenitor cells with similar properties arise from two neurogenic

regions of adult human brain, Exp. Neurol. 156 (1999) 333–344.

[19] R. Labbe, A. Firl Jr., E.J. Mufson, D.G. Stein, Fetal brain transplant:

reduction of cognitive deficits in rats with frontal cortex lesions,

Science 221 (1983) 470–472.

[20] S.W. Levison, J.E. Goldman, Multipotential and lineage restricted

precursors coexist in the mammalian perinatal subventricular zone, J.

Neurosci. Res. 48 (1997) 83–94.

[21] O. Lindvall, P. Odin, Clinical application of cell transplantation and

neurotrophic factors in CNS disorders, Curr. Opin. Neurobiol. 4

(1994) 752–757.

[22] S. Liu, Y. Qu, T.J. Stewart, M.J. Howard, S. Chakrabortty, T.F.

Holekamp, J.W. McDonald, Embryonic stem cells differentiate into

oligodendrocytes and myelinate in culture and after spinal cord

transplantation, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 6126–6131.

[23] D. Lu, Y. Li, A. Mahmood, L. Wang, T. Rafiq, M. Chopp, Neural and

marrow-derived stromal cell sphere transplantation in a rat model of

traumatic brain injury, J. Neurosurg. 97 (2002) 935–940.

[24] A. Mahmood, D. Lu, L. Wang, M. Chopp, Intracebral transplantation

of marrow-stromal cells cultured with neurotrophic factors promotes

functional recovery in adult rats subjected to traumatic brain injury, J.

Neurotrauma. 19 (2002) 1609–1617.

[25] T.K. McIntosh, K.E. Saatman, R. Raghupathi, D.I. Graham, D.H.

Smith, V.M. Lee, J.Q. Trojanowski, The Dorothy Russell Memorial

Lecture. The molecular and cellular sequelae of experimental

traumatic brain injury: pathogenetic mechanisms, Neuropathol. Appl.

Neurobiol. 24 (1998) 251–267.

D.A. Shear et al. / Brain Research 1026 (2004) 11–2222

[26] R. McKay, Stem cells in the central nervous system, Science 276

(1997) 66–71.

[27] D.M. Mctigue, P. Wei, B.T. Stokes, Proliferation of NG2-positive cells

and altered oligodendrocyte numbers in contused rat spinal cord, J.

Neurosci. 21 (2001) 3392–3400.

[28] E.A. Milward, C.G. Lundberg, B. Ge, D. Lipsitz, M. Zhao, I.D.

Duncan, Isolation and transplantation of multipotential populations of

epidermal growth factor-responsive, neural progenitor cells from the

canine brain, J. Neurosci. Res. 50 (1997) 862–871.

[29] M. Nieto-Sampedro, J.P. Kesslak, R. Gibbs, C.W. Cotman, Effects of

conditioning lesions on transplant survival, connectivity, and function.

Role of neurotrophic factors, Ann. N. Y.A. Acad. Sci. 495 (1987)

108–119.

[30] A. Nishiyama, A. Chang, B.D. Trapp, NG2+ glial cells: a novel glial

cell population in the adult brain, J. Neuropathol. Exp. Neurol. 58

(1999) 1113–1124.

[31] A. Nishiyama, M. Yu, J.A. Drazba, V.K. Tuohy, Normal and reactive

NG2+ glial cells are distinct from resting and activated microglia, J.

Neurosci. Res. 48 (1997) 299–312.

[32] M.F. Philips, G. Mattiasson, T. Wieloch, A. Bjorklund, B.B.

Johansson, G. Tomasevic, A. Martinez-Serrano, P.M. Lenzlinger, G.

Sinson, M.S. Grady, T.K. McIntosh, Neuroprotective and behavioral

efficacy of nerve growth factor-transfected hippocampal progenitor

cell transplants after experimental traumatic brain injury, J. Neuro-

surg. 94 (2001) 765–774.

[33] J.E. Pierce, D.H. Smith, J.Q. Trojanowski, T.K. McIntosh, Enduring

cognitive, neurobehavioral and histopathological changes persist for

up to one year following severe experimental brain injury in rats,

Neuroscience 87 (1998) 359–369.

[34] S. Pluchino, A. Quattrini, E. Brambilla, A. Gritti, G. Salani, G. Dina,

R. Galli, U. Del Carro, S. Amadio, A. Bergami, R. Furlan, G. Comi,

A.L. Vescovi, G. Martino, Injection of adult neurospheres induces

recovery in a chronic model of multiple sclerosis, Nature 422 (2003)

688–694.

[35] S. Rakic, N. Zecevic, Early oligodendrocyte progenitor cells in the

human fetal telencephalon, Glia 41 (2003) 117–127.

[36] M.S. Rao, Multipotent and restricted precursors in the central nervous

system, Anat. Rec. 257 (1999) 137–148.

[37] P. Riess, C. Zhang, K.E. Saatman, H.L. Laurer, L.G. Longhi, R.

Raghupathi, P.M. Lenzlinger, J. Lifshitz, J. Boockvar, E. Neugebauer,

E.Y. Snyder, T.K. McIntosh, Transplanted neural stem cells survive,

differentiate, and improve neurological motor function after exper-

imental traumatic brain injury, Neurosurgery 51 (2002) 1043–1052

discussion 1052–1054.

[38] I. Roberts, G. Schierhout, P. Alderson, Absence of evidence for the

effectiveness of five interventions routinely used in the intensive care

management of severe head injury: a systematic review, J. Neurol.

Neurosurg. Psychiatry 65 (1998) 729–733.

[39] D.G. Stein, Practical and theoretical issues in the uses of fetal brain

tissue transplants to promote recovery from brain injury, in: S. Finger,

T.E. LeVere, R. Almli, D.G. Stein (Eds.), Brain Injury and Recovery,

Plenum, New York, 1988, pp. 249–272.

[40] M.C. Tate, D.A. Shear, S.W. Hoffman, D.G. Stein, D.R. Archer, M.C.

LaPlaca, Fibronectin promotes survival and migration of primary

neural stem cells transplanted into the traumatically injured mouse

brain, Cell Transplant 11 (2002) 283–295.

[41] M. Watanabe, Y. Toyama, A. Nishiyama, Differentiation of prolif-

erated NG2-positive glial progenitor cells in a remyelinating lesion, J.

Neurosci. Res. 69 (2002) 826–836.

[42] S.R. Whittemore, E.Y. Snyder, Physiological relevance and functional

potential of central nervous system-derived cell lines, Mol. Neurobiol.

12 (1996) 13–38.

[43] M.S. Windrem, N.S. Roy, J. Wang, M. Nunes, A. Benraiss, R.

Goodman, G.M. McKhann 2nd, S.A. Goldman, Progenitor cells

derived from the adult human subcortical white matter disperse and

differentiate as oligodendrocytes within demyelinated lesions of the

rat brain, J. Neurosci. Res. 69 (2002) 966–975.