Delivery of a GDNF Gene into the Substantia Nigra after a Progressive 6-OHDA Lesion Maintains Functional Nigrostriatal Connections

Experimental Neurology 166, 1–15 (2000) doi:10.1006/exnr.2000.7463, available online at http://www.idealibrary.com on

Delivery of a GDNF Gene into the Substantia Nigra after a Progressive 6-OHDA Lesion Maintains Functional Nigrostriatal Connections Dorothy A. Kozlowski,*,1 Bronwen Connor,*,2 Jennifer L. Tillerson,† Timothy Schallert,† and Martha C. Bohn* *Children’s Memorial Institute for Education and Research, Department of Pediatrics, Northwestern University Medical School, Chicago, Illinois 60614; and †Department of Psychology, University of Texas, Austin, Texas 78712 Received December 22, 1999; accepted April 12, 2000

eration has commenced. Thus, GDNF gene therapy may ameliorate the consequences of Parkinson’s disease through rescuing compromised dopaminergic neurons. © 2000 Academic Press Key Words: Parkinson’s disease; GDNF; adenoviral vector; gene therapy; nigrostriatal system; 6-OHDA.

The effects of delivering GDNF via an adenoviral vector (AdGDNF) 1 week after lesioning dopaminergic neurons in the rat substantia nigra (SN) with 6-hydroxydopamine (6-OHDA) were examined. Rats were unilaterally lesioned by injection of 6-OHDA into the striatum, resulting in progressive degeneration of dopaminergic neurons in the SN. One week later, when substantial damage had already occurred, AdGDNF or a control vector harboring ␤-galactosidase (AdLacZ) was injected into either the striatum or SN (3.2 ⴛ 10 7 PFU/␮l in 2 ␮l). Rats were examined behaviorally with the amphetamine-induced rotation test and for forelimb use for weightbearing movements. On day 30 postlesion, the extent of nigrostriatal tract degeneration was determined by injecting a retrograde tracer (FluoroGold) bilaterally into the lesioned striatum. Five days later, rats were sacrificed within 2 h of amphetamine injection to examine amphetamine-induced Fos expression in the striatum, a measure of dopaminergic-dependent function in target neurons. AdGDNF injection in the SN rescued dopaminergic neurons in the SN and increased the number of dopaminergic neurons that maintained a connection to the striatum, compared to rats injected with AdLacZ. Further support that these spared SN cells maintained functional connections to the striatum was evidenced by increased Fos expression in striatal target neurons and a decrease in amphetamine-induced rotation. In contrast to the effects observed in rats injected with AdGDNF in the SN, rats injected with AdGDNF in the striatum did not exhibit significant ameliorative effects. This study demonstrates that experimentally increasing levels of GDNF biosynthesis near the dopaminergic neuronal soma is effective in protecting the survival of these neurons and their function even when therapy is begun after 6-OHDA-induced degen-

INTRODUCTION

The therapeutic potential of glial cell line-derived neurotrophic factor (GDNF) for Parkinson’s disease (PD) is supported by numerous studies. Injection of GDNF protein into the lateral ventricles, striatum, or substantia nigra (SN) in rodent or primate models of PD has been reported to protect dopaminergic (DA) neurons in the SN, increase dopamine levels in the striatum and SN, and ameliorate behavioral deficits (4, 6, 14, 15, 17). While these initial studies suggest a promise for GDNF as a novel therapy, practical issues are evident when considering GDNF for human use where the injection of GDNF protein into the lateral ventricle may not effectively reach the dopaminergic neurons (22). Since PD results in progressive cell death, neurotrophic support is likely to be required for a long period of time. Therefore, safe, minimally invasive methods need to be developed to deliver these factors in a manner which provides the potential of affecting dopaminergic neurons in the brainstem specifically and chronically, in the absence of side effects. Since most neurotrophic factors are labile substances that do not readily cross the blood– brain barrier effectively, viral vectors provide a potential means for delivering GDNF to degenerating dopaminergic neurons in a way where the factor is continuously synthesized in the desired anatomical location. Previous studies using adenoviral (Ad) vectors and adenoassociated viral vectors (AAV) have shown that GDNF gene delivery prior to a 6-OHDA lesion of the nigrostriatal system protects against death of dopaminergic neurons in rat (3, 9, 10, 12, 29). Adenoviral GDNF also protects against the development of lesion-

1 Current address: Dept. of Biological Sciences, DePaul University, 2325 N. Clifton, Chicago, IL 60614. 2 Current address: Division of Pharmacology, Faculty of Medical and Health Science, Auckland University, New Zealand.

1

0014-4886/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

2

KOZLOWSKI ET AL.

induced behaviors (3, 11, 12). These observations suggest that if healthy dopaminergic neurons remain in the SN of patients with PD, GDNF gene delivery may slow or inhibit death of these neurons and prolong dopaminergic function. However, many dopaminergic neurons may already be compromised by the disease process when clinical symptoms become evident. This raises the question of whether GDNF gene delivery might have an ameliorative effect on a substantia nigra that has already incurred damage. To address this question, we delivered a GDNF gene 1 week after a progressive 6-OHDA lesion of the nigrostriatal pathway in rats and assessed the effects on dopaminergic neuronal survival and function. Rats received a unilateral partial lesion of the nigrostriatal pathway by injection of 6-OHDA into the striatum. This results in a progressive degeneration of dopaminergic neurons in the SN as reported previously by Sauer and Oertel (38). One week later, at which time significant damage to the dopamine terminals has already occurred, an Ad vector harboring human preproGDNF (AdGDNF) or the cellular marker gene, LacZ (AdLacZ), was injected. Vectors were injected near the lesion site in the striatum or above the SN to compare effects of GDNF gene delivery to the terminals or to the soma of the dopaminergic neurons. To study the effects of GDNF on survival of dopaminergic neurons and the maintenance of their nigrostriatal projections, the retrograde tracer FluoroGold (FG) was injected into the striatal lesion site at the end of the experiment and the number of FG-labeled neurons in the SN was counted. This strategy permitted an assessment of the effects of GDNF on dopaminergic neuronal survival without relying on the dopaminergic phenotype, as well as a specific assessment of the population of dopaminergic neurons that maintained axonal projections to the lesion site. In addition, GDNF effects on dopamine-dependent function were studied by examining amphetamine-induced Fos expression in target neurons in the striatum and by assessing behaviors known to be dependent on dopamine. METHODS AND MATERIALS

Animals Forty male Fischer 344 rats (200 –300 g) were housed in the Children’s Memorial Institute for Education and Research vivarium. Food and water were available ad libitum during a 12-h light/dark cycle and animals were treated according to institutional and NIH guidelines. All rats received a unilateral progressive 6-OHDA lesion and 1 week later were divided into five treatment groups, based on vector injection: AdLacZ into the SN (LacZ-SN, n ⫽ 7); AdLacZ into the striatum (LacZ-ST, n ⫽ 7); AdGDNF into the SN (GDNF-SN, n ⫽ 8); AdGDNF into the striatum

(GDNF-ST, n ⫽ 8); or no vector control (control, n ⫽ 8). Three rats were excluded from the study due to the absence of a 6-OHDA lesion as verified by no evident decrease in TH staining in both the striatum and the SN and no sign of a needle track in the striatum. Two rats died prior to the end of the study (at 2 and 3 weeks postlesion). This resulted in a decrease in group size for the following: AdLacZ into the striatum (AdLacZ-ST, n ⫽ 6); AdGDNF into the striatum (AdGDNF-ST, n ⫽ 7); AdGDNF into the SN (AdGDNF-SN, n ⫽ 7); and no vector control (Control, n ⫽ 6). The remaining groups remained as indicated above unless otherwise noted. Surgery 6-OHDA lesion. A modification of the progressive 6-OHDA lesion as described previously was used (10, 38). Adult male Fischer 344 rats were anesthetized with chloral hydrate–sodium pentobarbital anesthesia (149 and 31 mg/kg respectively, ip). Once anesthetized and shaved, animals were placed in a Stoelting stereotaxic apparatus. The skull was exposed and the bite bar was adjusted to level bregma and lambda in the horizontal plane. A burr hole was made unilaterally at ⫹1.0 A/P and 3.0 M/L from bregma. A progressive unilateral 6-OHDA lesion was obtained by injecting 6-OHDA-HBr (16 ␮g in 2.8 ␮l of 0.2 mg/ml ascorbic acid in 0.9% sterile saline) unilaterally into the striatum (⫺5.0 D/V) through the burr hole using a 10-␮l Hamilton syringe with a 30-gauge needle at 0.5 ␮l/min. The injection was done manually. After injection the needle was left in place for 5 min and retracted at 1 mm/min. The wound was sutured, infiltrated with lidocaine ointment (1%) and a triple antibiotic ointment, and the animal was kept warm until it recovered from anesthesia. Vector injection. Adenoviral vectors encoding human GDNF prepro-protein (AdGDNF) or the reporter enzyme ␤-galactosidase with a nuclear localization signal (AdLacZ) were prepared and titered as described previously (10). Vector titers were matched for infectious particles as follows: AdGDNF, 3.2 ⫻ 10 7 PFU, titer ⫽ 1 ⫻ 10 7 particles/␮l, particle ratio ⫽ 15; AdLacZ, 3.2 ⫻ 10 7 PFU, titer ⫽ 1 ⫻ 10 7 particles/␮L particle ratio ⫽ 33. One week after the unilateral progressive 6-OHDA lesion, animals were anesthetized (149 mg/kg chloral hydrate; 31 mg/kg sodium pentobarbital), shaved, and placed in a Stoelting stereotaxic holder. Two microliters of the vector was injected unilaterally using a 10-␮l Hamilton syringe at 0.5 ␮l/min using an automatic microinjector (Stoelting). The injections were made either into the striatum just posterior and medial to the 6-OHDA lesion (⫹0.1 A/P, 2.8 M/L, ⫺5.0 D/V) or just above the SN (⫺5.3 A/P, 1.8 M/L, ⫺7.4 D/V). The needle was left in place for 5 min and then withdrawn at 1 mm/min. The burr hole was covered

AdGDNF MAINTAINS DOPAMINERGIC FUNCTION AND CONNECTIONS

with gel foam, the overlying skin was sutured, and the animal was placed in its home cage to recover from anesthesia prior to being returned to the vivarium. Control animals were anesthetized, but did not receive any intracerebral vector injections. Fluorogold injection. To examine the remaining neural connections between the striatum and SN, a retrograde tracer (FG, 2%, Fluorochrome, Englewood, CO; 0.2 ␮l at 0.05 ␮l/min) was injected bilaterally into the striatum, 30 days after the 6-OHDA lesion, using the same coordinates used for the lesion (⫹1.0 A/P; 3.0 M/L, ⫺5.0 D/V). The overlying skin was sutured and animals were returned to the vivarium following recovery from anesthesia. Behavioral Testing Amphetamine-induced rotation. The degree of amphetamine-induced rotation was assessed on day 0, prior to the 6-OHDA surgery, to establish a baseline and assign surgery side. Typically, animals did not rotate in just one direction before a lesion, but there was a slight intrinsic bias. The lesion side was chosen to counteract the natural bias by placing it contralateral to the side of the natural rotation bias. On days 0, 21, and 35 postlesion, amphetamine-induced rotation was recorded by videotaping rats in plastic bowls (i.e., rotation chambers 38 cm in diameter and 18 cm deep), after being injected, ip, with 6.8 mg/kg d,l-amphetamine sulfate. The number of clockwise and counterclockwise turns during a 50-min period was counted and expressed as rotations/min. Forelimb use for weight-bearing movement. Forelimb use during exploration of vertical surfaces was observed and analyzed using a modification of previously described methods (12, 23, 40). On days 0, 14, 21, 28, and 35 postlesion animals were placed in a threesided Plexiglas platform and videotaped for 5 min. The videotaped spontaneous limb use was scored by a blinded observer. Instances of left, right, and simultaneous (both) forelimb use were scored during exploratory behavior along the wall of the platform both in horizontal and vertical planes and during landing on the floor after termination of wall contact. Ipsilateral (to the lesion) forelimb use was analyzed for wall behaviors and landing behaviors separately: ((ipsilateral wall/(ipsilateral ⫹ contralateral ⫹ both wall behaviors)); (ipsilateral land/(ipsilateral ⫹ contralateral ⫹ both land behaviors)). These two scores were added and divided by 2 to obtain ipsilateral forelimb use. Histology and Morphometry On day 35 postlesion animals were sacrificed by transcardial perfusion (0.9% saline followed by 4% paraformaldehyde in PBS) within 2 h of the last amphetamine injection used in the rotation test. The

3

brains were removed and placed in 4% paraformaldehyde overnight followed by cryoprotection in 30% sucrose. Brains were sliced coronally on a sliding microtome (50-␮m serial sections through the striatum and SN) and kept in sets of six, resulting in a set of sections each 300 ␮m apart. The sections were then stored in cryoprotectant at ⫺20°C. For immunocytochemistry, sections were stained without primary antibodies as negative controls. No immunoreactivity was evident in any of the sections for any of the primary antibodies. Fos and GDNF immunoreactivity (Fos-IR, GDNFIR) and morphometry. Amphetamine-induced expression of Fos⫹ nuclei in the striatum and virally mediated human GDNF expression in the striatum and SN were determined by immunocytochemistry. Sections obtained from either the striatum or the SN (every 300 ␮m) were incubated in 0.3% H 2O 2 in phosphate-buffered saline (PBS) for 15 min and blocked for 60 min at room temperature (RT) in 3% normal goat serum (NGS) and 2% bovine serum albumin (BSA) in 0.05% Triton-X (TX) in PBS. Sections were then incubated in the primary antibody (GDNF, polyclonal rabbit anti-GDNF from R&D Systems, 1:250; Fos, polyclonal anti-Fos from Santa Cruz sc-52-G, 1:1000) overnight at RT in 1% NGS, 1% BSA, and 0.05% TX in PBS. Sections were then incubated in a biotinylated goat anti-rabbit secondary antibody (Vector Laboratories, Burlingame, CA) diluted 1:500 in 1% NGS and 1% BSA in PBS for 2.5 h followed by incubation in avidin– peroxidase conjugate (Vectastain ABC Elite, Vector Laboratories) in PBS for 2 h. Visualization was with diaminobenzidine (DAB; 20 mg/ml), 0.8% nickel sulfate, 0.005% H 2O 2 in 50 mM sodium acetate, 10 mM imidazol buffer (pH 7.0). Several washes with 0.05% TX-100 in PBS were performed between each step. To quantify amphetamine-induced Fos expression in the striatum, sections anterior to the lesion site (700 – 900 ␮m), within the 6-OHDA injection site, at the striatal vector injection site, and 1.0 –1.5 mm posterior to the lesion site were analyzed. Fos expression in the striatum was measured in a 4-mm 2 area surrounding the injection site. Using NeuroLucida software (MicroBrightField, Colchester, VT) and a Nikon inverted microscope, a 4-mm 2 area was outlined at 188⫻ magnification using the “contour function” and a calibrated grid. At 952⫻ magnification, Fos⫹ nuclei within this area were counted in randomly selected grid areas (100 ␮m 2) every 240 ␮m using the “Meander Scan” function. Fos⫹ nuclei were counted in both the lesioned and the unlesioned striatum and were expressed as a percentage of the unlesioned striatum. Abercrombie’s correction was not performed in the measurement of Fos⫹ nuclei. Given the thickness of the sections in which the nuclei were being measured (50 ␮m) and the size of the nuclei being measured (5 ␮m) the correction would be negligible.

4

KOZLOWSKI ET AL.

To quantify the area of human GDNF expression, stained sections were analyzed by outlining the GDNF stained area (penumbra) using the Neurolucida morphometry system. The total area of GDNF-IR was obtained by adding the areas of GDNF expression in each section. The anterior–posterior extent of GDNF expression was also measured. Tyrosine hydroxylase immunoreactivity (TH-IR) and morphometry. TH-IR was used to visualize the decrease in tyrosine hydroxylase positive (TH⫹) fibers in the striatum. Sections obtained from the striatum (every 300 ␮m) were incubated in Tris-buffered saline (TBS) at RT with additional reagents as noted below. Sections were incubated for 15 min in 0.3% H 2O 2 in TBS, blocked for 20 min in 10% NGS and 0.5% TX in TBS, and incubated overnight with 1:2000 polyclonal rabbit anti-TH antibody (Chemicon, Temecula, CA) in 1% NGS and 0.3% TX-100 in TBS. Sections were then incubated in a biotinylated goat anti-rabbit secondary antibody (Vector Laboratories) diluted 1:500 in 1% NGS and 0.1% TX in TBS for 2.5 h followed by incubation in avidin–peroxidase conjugate (Vectastain ABC Elite, Vector Laboratories) in TBS for 2 h. Visualization was with DAB (20 mg/ml), 0.8% nickel sulfate, 0.005% H 2O 2 in 50 mM sodium acetate, 10 mM imidazole buffer (pH 7.0). Several washes with 0.1% TX in TBS were performed between each step. The extent of the loss of TH⫹ fibers in the striatum was measured using computer-assisted imaging techniques, NIH Image 1.60. The area of decreased TH-IR was determined visually, outlined, and quantified in all striatal brain sections, as previously described (12). Total lesion volume was obtained by adding the lesioned areas together and multiplying by the sampling interval (300 ␮m). The data are expressed as volume of reduced TH-IR in the striatum. The density of TH⫹ fibers in the area of decreased TH-IR in the striatum was also measured using optical densitometry techniques in NIH Image 1.60. The optical density of the area of decreased TH-IR in the striatum was measured in sections that contained either the 6-OHDA lesion site or the vector injection site. The optical density of the cortex was assigned a level of 0 while the density in the unlesioned striatum was assigned a level of 1.0. The section to be measured was then calibrated using this scale and the optical density measured. The data are expressed as optical density of TH-IR in the striatum. Immunocytochemistry for both TH and Fos. To examine the colocalization of TH⫹ fibers and Fos⫹ nuclei in the striatum, striatal sections (every 300 ␮m) were incubated in 0.3% H 2O 2 in PBS for 15 min and blocked for 40 min at RT in 3% NGS and 2% BSA in PBS/0.05% TX. Sections were then incubated in the first primary antibody overnight at RT in 1% NGS, 1% BSA, and 0.05% TX in PBS (polyclonal anti-Fos from Santa Cruz

sc-52-G; 1:1000). Sections were then incubated in a biotinylated goat anti-rabbit secondary antibody (Vector Laboratories; 1:500 in 1% NGS and 1% BSA in PBS) for 2.5 h followed by incubation in avidin–peroxidase conjugate (Vectastain ABC Elite, Vector) in PBS for 2 h. Visualization was with DAB (20 mg/ml), 0.8% nickel sulfate, 0.005% H 2O 2 in 50 mM sodium acetate, 10 mM imidazole buffer (pH 7.0). Several washes with 0.05% TX-100 in PBS were performed between each step. The sections were then placed overnight in TBS and on the following day incubated in the second primary antibody overnight at RT in 1% NGS, 0.3% TX in TBS (monoclonal anti-TH; Diasorin; 1:2000). Sections were then incubated in biotinylated anti-mouse secondary antibody (Vector Laboratories; 1:250 in 1% NGS, 0.3% TX in TBS) for 2.5 h followed by a 2-h incubation in avidin–peroxidase conjugate (Vectastain ABC Elite in TBS). Visualization was with Nova Red (Vector Laboratories). Several washes with 0.1% TX in TBS were performed between each step. In order to control for the cross-reactivity of Nova Red to the first primary antibody applied in the double label, a secondary antibody control study was performed. Using the immunocytochemical procedures detailed above, the first primary antibody was detected using DAB. Following this, the sections were incubated in a secondary biotinylated goat antimouse antibody (diluted 1:500) for 2.5 h at RT. The secondary antibody was then removed, and sections were washed in TBS and incubated for 2 h in avidin–peroxidase conjugate as previously described. The sections were then visualized using Nova Red. In the control sections, no staining for Nova Red was detected, indicating antibody specificity. Retrograde tracing of neurons in the SN. The number of remaining connections from the SN to the lesion site in the striatum was determined by counting the number of FG⫹ neurons in the SN. The number of FG⫹ neurons in the SN was counted at 952⫻ magnification every 300 ␮m in both the lesioned and the unlesioned hemisphere using the Neurolucida morphometry system. The data are expressed as the percentage of FG⫹ neurons remaining in the lesioned hemisphere compared to the unlesioned hemisphere. To verify that the FG⫹ neurons were in fact dopaminergic, SN sections were processed for immunocytochemistry using both a TH antibody and a FG antibody. Sections of SN were incubated in 0.3% H 2O 2 in TBS for 15 min and blocked in 10% NGS and 0.5% TX in TBS for 20 min. Sections were then incubated overnight at RT in a monoclonal TH antibody (Diasorin) diluted 1:2000 in 1% NGS and 0.3% TX in TBS followed by a 2.5-h incubation in biotinylated anti-mouse secondary antibody (Vector, 1:250 in 1% NGS, 0.3% TX in TBS) at room temperature. The sections were then incubated for 2 h at room temperature in Texas red– avidin solution (Vector, 1:2000 in TBS) and left over-

AdGDNF MAINTAINS DOPAMINERGIC FUNCTION AND CONNECTIONS

night at 4°C. Next these sections were incubated in a polyclonal FG antibody (Chemicon; 1:3000 in 1% NGS, 0.3% TX in TBS) overnight at RT. Sections were then incubated in a biotinylated anti-rabbit secondary antibody (Vector, 1:250 in 1% NGS, 0.3% TX in TBS) for 2.5 h at RT followed by a 2-h incubation at RT in fluorescein avidin DCS solution (Vector, 1:2000 in TBS). Double labeling in the SN was verified using confocal microscopy (Olympus Fluoview Confocal). The FG was examined using the FITC filter (excitation 488 nm; emission 510 –550 nm) and the TH was examined using the Texas red filter (excitation 568 nm; emission 590⫹ nm). Negative control sections were stained with only the Fos or the TH antibody alone or without both primary antibodies. Using these control sections it was verified that the fluorescent signal exhibited by one antibody in one channel could not be detected in the channel for the examination of the other antibody. In addition, there was no cross-reactivity of the second primary antibody to the first when examined using the secondary antibody control study described in the previous method for immunostaining for both Fos and TH. Host immune response—Hematoxylin and eosin. Sections containing the vector injection sites were stained with hematoxylin and eosin to assess host immune response. The intensity of host immune response was assessed by a blinded observer using a qualitative rating scale (12). The rating scale was from 1 to 5 (1 being least severe and 5 being most severe) and was based on the intensity and spread of inflammatory infiltrates throughout the neural tissue. A score was obtained for each rat and comparisons were made between injection site (striatum vs SN) and vector type (AdGDNF vs AdLacZ). X-gal histochemistry. AdLacZ transgene expression was visualized by histochemistry for X-gal (5bromo-4-chloro-3-indoyl-␤-D-galactopyranoside) which is converted to a blue reaction product by the LacZ gene product, ␤-galactosidase, as described previously (12). Brain sections were examined at 188⫻ magnification, and each brain section was scanned for areas containing blue nuclei. Once an area was identified, it was outlined using the “contour” function at 476⫻ magnification using NeuroLucida software (MicroBrightField, Colchester, VT). The magnification was then increased to 952⫻ and the “Meander Scan” function was engaged. This allowed the entire contour to be sampled in 180 ⫻ 180-␮m increments. Blue nuclei within each increment were marked with a designated cursor symbol using the mouse. This was continued until the entire contour area was sampled and a total number of blue cells in the section was obtained. The number of blue nuclei was summed and multiplied by the number of sections in the sampling interval to obtain an estimate of the number of transgene expressing cells in the entire infected brain area. Abercrom-

5

bie’s correction was not performed in the measurement of nuclei. Given the thickness of the sections in which the nuclei were being measured (50 ␮m) and the size of the nuclei being measured (5 ␮m) the correction would be negligible. The anterior to posterior extent of ␤-gal expression from the injection site was also measured. Statistical Analysis Morphological and weight data were analyzed using a one-way ANOVA for group with posthoc analysis conducted using Tukey–Kramer post hoc analyses. Behavioral data were analyzed with a repeated-measure two-way ANOVA for group and day of testing with day being the repeated measure. Post hoc analyses were conducted using simple mean comparisons between groups at specific test days. RESULTS

Fluorogold-Positive Neurons in the SN: Maintenance of Nigrostriatal Connections The number of neurons that maintained a nigrostriatal connection was examined by injecting a retrograde tracer into the striatum at day 30 postlesion, in the same area as the 6-OHDA lesion. Injecting the tracer at the end of the study resulted in the labeling of the subpopulation of dopaminergic nigral neurons exposed to 6-OHDA. Counting these neurons produced an estimate of the number of dopaminergic neurons that still maintained connections to the striatum, 30 days postlesion, and was not just a measure of surviving dopaminergic cell bodies in the SN. The percentage of FG⫹ cells in the lesioned SN was obtained by comparing cell counts to those in the contralateral unlesioned SN. A one-way ANOVA for group indicated a significant difference among the five groups (F(4, 27) ⫽ 5.2, P ⬍ 0.003, see Fig. 1). Post hoc analysis revealed that rats that received AdGDNF into the SN showed a significant increase in the percentage of FG⫹ cells in the lesioned SN (51.5 ⫾ 6.4%) compared to LacZ-ST (23.6 ⫾ 4.8%), LacZ-SN (24.9 ⫾ 3.7%), and control (28.8 ⫾ 2.4%) animals (Tukey–Kramer; P ⬍ 0.05). Animals injected with AdGDNF into the striatum showed a slight increase in the number of FG⫹ cells (36.9 ⫾ 5.8%); however, this was not significantly different from the control groups. To verify that labeled neurons in the SN were dopaminergic, the SN was processed for both FG and TH immunocytochemistry. Confocal microscopy was used to show that the FG⫹ neurons in the lesioned nigra rescued by GDNF were TH⫹ (see Fig. 2). This result demonstrates that AdGDNF injected into the SN 1 week postlesion maintains a significantly greater number of neuronal connections between striatum and SN at 4 weeks postvector injection and that these connections are dopaminergic.

6

KOZLOWSKI ET AL.

FIG. 1. FG⫹ neurons in the SN following injection of the retrograde tracer FG into the striatum at day 30 postlesion. The number of FG⫹ neurons in the unlesioned (A) and lesioned (B–D) hemisphere was counted and expressed as %FG⫹ cells in the lesioned SN. The 6-OHDA lesion resulted in 30% of FG⫹ cells in the SN which still maintained a connection to the striatal lesion 35 days postlesion (E). Injection of AdGDNF into the SN (B) increased the number of nigrostriatal connections to approximately 55% (*P ⬍ 0.003; E). Injection of AdGDNF into the ST resulted in slightly more FG⫹ cells than control, but the difference was not significant (C and E).

Fos Immunoreactivity in the Striatum The function of the striatal dopaminergic system was assessed using amphetamine-induced expression of Fos. In the striatum, enhancement of dopamine-induced neural activity results in the expression of Fos protein (8, 25, 33). The expression of the immediate early gene c-fos following the administration of a dopamine agonist such as amphetamine is thought to be due to activation of postsynaptic dopaminergic D1 receptors on striatonigral neurons and is associated with neural activity (13, 25, 35). In animals with striatal denervation, there is a decrease in amphetamine-in-

duced Fos expression (25, 33) that can be normalized by transplantation of dopaminergic tissue (8, 33). In this study, we asked whether AdGDNF can prevent the decrease in Fos expression seen following striatal denervation. Fos⫹ nuclei were measured in the lesioned and unlesioned striatum in two sections containing the 6-OHDA injection site, in one section 700 ␮m anterior and one 1–1.5 mm posterior to the 6-OHDA injection site, and at the site of striatal vector injection. A twoway ANOVA for group and site indicated a significant effect of group F(4, 15) ⫽ 16.67, P ⬍ 0.0001). Post hoc analysis revealed that at 35 days postlesion, a

AdGDNF MAINTAINS DOPAMINERGIC FUNCTION AND CONNECTIONS

7

FIG. 2. Immunocytochemistry for FG and TH in the lesioned SN of a rat injected with AdGDNF into the SN. The FG⫹ cells rescued by AdGDNF expressed both FG and TH suggesting that the rescued neurons were indeed dopaminergic.

striatal 6-OHDA lesion resulted in the expected decrease in Fos expression in the lesioned hemisphere compared to the unlesioned hemisphere at all levels examined. In the sections containing the 6-OHDA injection site, amphetamine-induced Fos expression in the lesioned hemisphere was decreased to about 50 – 60% of the unlesioned hemisphere. In animals injected with AdGDNF into the SN, Fos expression was significantly increased to 110% of the unlesioned hemisphere (P ⬍ 0.05, Fig. 3). AdGDNF into the striatum

and AdLacZ did not have any effect on amphetamineinduced Fos expression. This suggests that AdGDNF, when injected into the SN, maintains dopaminergic function in the striatum. TH Immunoreactivity in the Striatum The loss of the dopaminergic phenotype in the striatum 35 days following a progressive 6-OHDA lesion was examined by measuring the volume of decreased

FIG. 5. Immunocytochemistry for Fos and TH. (A) Negative control, striatal section stained without primary antibodies. (B) Low magnification image of the lesioned striatum showing the area of decreased TH-IR still containing Fos⫹ nuclei. (C) Higher magnification of the striatal area with decreased TH-IR. Note that the amount of Fos⫹ nuclei in this area is identical to the amount of Fos⫹ nuclei in an area with normal levels of TH-IR (D).

8

KOZLOWSKI ET AL.

FIG. 3. Amphetamine-induced Fos expression in the lesioned striatum. (A) The number of Fos⫹ nuclei were counted in the unlesioned and lesioned striatum in the 6-OHDA injection site. Lesioned-Control animals showed a 40% decrease in Fos expression in the lesioned striatum when compared to the unlesioned striatum as did rats injected with the control vector AdLacZ (B). Injection of AdGDNF into the SN prevented this decrease in Fos-expression (*P ⬍ 0.0001). Injection of AdGDNF into the ST showed no effect.

TH-IR throughout the entire striatum and by measuring the optical density of TH⫹ fibers in the area of decreased TH-IR in sections that contained either the 6-OHDA injection site or the vector injection site. There was no significant difference in the volume of decreased TH-IR among groups (F(4, 27) ⫽ 0.349; P ⫽ 0.84, Fig. 4, top). In addition, within the area of decreased TH-IR, there were no significant differences in the density of TH-IR fibers (F(4, 27) ⫽ 0.319; P ⫽ 0.86, Fig. 4, bottom). This suggests that although AdGDNF injected into the SN maintained dopaminergic function, it did not significantly affect TH-IR in the

dopaminergic fibers of the striatum. The lack of THimmunoreactive fibers may not be necessarily equivalent to lost dopaminergic innervation of the striatum or lost dopaminergic function. The finding that amphetamine-induced Fos expression was evident in striatal areas that did not contain TH⫹ fibers supports this concept (Fig. 5). Behavior Behavioral analysis was conducted using the amphetamine-induced rotation test and observations of

AdGDNF MAINTAINS DOPAMINERGIC FUNCTION AND CONNECTIONS

FIG. 4. The volume of decreased TH-IR in the lesioned striatum and the density of TH-IR fibers in the area of decreased TH-IR were measured. There were no significant differences among groups in either measure.

forelimb use for weight-shifting movements. In both tests, two separate control groups were assigned based on vector injections. Uninjected, lesioned animals (control; n ⫽ 6) were members of both control groups. Animals injected striatally with AdLacZ (n ⫽ 6) were placed in the control-ST group and animals injected with AdLacZ into the SN (n ⫽ 7) were placed in the control-SN group. Thus, rats treated with AdGDNF in the striatum were compared with the control-ST group and rats treated with AdGDNF in the SN were compared with the control-SN group. Statistical analysis of AdLacZ vs control animals for both behavioral measures indicated that there were no significant differences among groups (P ⬎ 0.05). In the amphetamine rotation test, one control rat was removed from the analysis due to a lack of movement (Control-ST, n ⫽ 12; Control-SN, n ⫽ 11). This rat just sat in the chamber and exhibited stereotypic movements such as head bobbing and chewing. In the observations of limb use rats were not included in the analysis if the total number of behaviors exhibited during the observation period was less than 5. Two animals from the Control group were excluded from

9

the analysis based on this criteria (Control-ST, n ⫽ 10; Control-SN, n ⫽ 11). Data from day 35 were not included in the analysis due to the large number of animals that did not meet the criteria (n ⫽ 10). These animals did not explore the testing environment; however, behavior in the home cage was not noticeably different. Therefore, limb use observations are presented from day 0 (prelesion) to day 28 postlesion. Amphetamine-induced rotation. An intraperitoneal injection of amphetamine results in an increase in the amount of dopamine in the striatum and an increase in behavioral activity such as locomotion, rearing, and stereotypic behaviors (41 for review). In unilaterally denervated animals amphetamine induces rotational behavior in the direction ipsilateral to the lesion, i.e., away from the area with greater dopamine levels (45). The induction of amphetamine-induced rotation has been used as an indication of the severity of dopaminergic denervation, has been associated with the level of dopamine release in the intact striatum (41 for review), and has been shown to vary depending upon the area of the striatum that is denervated (21). To determine if GDNF gene delivery prevented the development of behaviors associated with dopamine depletion, amphetamine-induced rotation was measured prior to lesioning with 6-OHDA and at 21 and 35 days postlesion and is presented as the number of ipsilateral rotations per minute (Fig. 6). Prior to the 6-OHDA lesion, all groups showed no significant rotation. Following the 6-OHDA lesion, the Control-SN, Control-ST, and AdGDNF-ST groups showed an increase in ipsilateral rotation that decreased over time. Injection of AdGDNF into the ST did not result in a decrease in rotation behavior (F(1, 16) ⫽ 0.180, P ⫽ 0.68; Fig. 6B). However, injection of AdGDNF into the SN 1 week after the 6-OHDA lesion prevented the development of amphetamine-induced rotation when compared to Control-SN animals (F(1, 17) ⫽ 4.734, P ⫽ 0.04; Fig. 6A). Therefore, AdGDNF injected into the SN not only protected striatal dopaminergic function but it also prevented the development of a dopamine-dependent behavioral deficit. Limb use for weight-shifting movements. The observation of forelimb use for weight-shifting movements is a measure of a non-drug-induced behavioral asymmetry. It has been shown that animals with unilateral damage to either the sensorimotor cortex or the nigrostriatal system preferentially use the forelimb ipsilateral to the lesion for exploring their environment (11, 12, 24, 40). This study also demonstrated that animals with a unilateral 6-OHDA lesion followed by no vector injection, injection of AdLacZ into either the striatum or the SN, or injection of AdGDNF into the striatum showed an increased preference for the ipsilateral forelimb as early as 14 days postlesion. This preference did not decrease by day 28 postlesion. In-

10

KOZLOWSKI ET AL.

jection of AdGDNF into the striatum did not result in any significant decrease in preference for ipsilateral forelimb use (F(1, 15) ⫽ 0.009, P ⫽ 0.92). Animals that were injected with AdGDNF into the SN showed a decreased preference for ipsilateral forelimb use when compared to Control-SN animals; however, this did not reach statistical significance (F(1, 16) ⫽ 3.32, P ⫽ 0.08) when analyzed with a two-factor repeated-measures ANOVA (Fig. 7). Transgene Expression Transgene expression was analyzed using X-gal histochemistry for rats injected with AdLacZ and GDNF immunocytochemistry for rats injected with AdGDNF. For animals injected with AdLacZ, the number of blue nuclei was counted in sections using the NeuroLucida morphometry system. This number was multiplied by 6 to obtain an estimate of the number of transgeneexpressing cells in the entire brain area sampled. The anterior to posterior extent of ␤-gal expression from the injection site was also measured. Results indicated an average of 11,813 (⫾6152) blue nuclei within a 2.5-mm (⫾0.6 mm) extent of SN. In the striatum an average of 24,752 (⫾8255) blue nuclei was counted within a 3.5-mm (⫾0.5 mm) extent. Although there were more blue cells in the striatally injected animals than in those injected in the nigra, there was not a significant difference in transgene expression between the two sites (F(1, 10) ⫽ 2.87, P ⫽ 0.12). Overall, transgene expression in this study was more extensive than in previously published studies by this laboratory (12). The vectors used were identical; however, the vector in this study was injected using a Stoelting automatic injector, whereas previously published data from this laboratory involved manual injection of the vector. GDNF-IR was measured using the Neurolucida morphometry system. Immunostaining revealed an area of GDNF immunoreactivity surrounding the injection site (Fig. 8). This was considered to be a measure of the extent of secretion of the GDNF protein at 1 month postinjection. To measure the GDNF-IR penumbra, striatal and nigral sections expressing GDNF-IR were examined under the microscope at 188⫻ magnification. The stained area was outlined using the Neurolucida system and an area and volume were obtained. The GDNF-IR penumbra in striatally injected animals encompassed a significantly larger area and volume than in animals injected above the nigra (F(1, 16) ⫽ 14.622, P ⬍ 0.001; area (mm 2) ⫽ 39.54 ⫾ 3.3 and 19.55 ⫾ 4.0, respectively; volume (mm 3) ⫽ 11.86 ⫾ 1.0 and 5.86 ⫾ 1.20 respectively). However, the anterior to posterior extent of GDNF-IR was identical in both groups (3.2–3.3 mm).

Host Immune Response The host immune response was measured qualitatively by examining H&E-stained sections for inflammatory infiltrates. The observer rated each injection site on a scale from 1 to 5, with 1 indicating a low level of inflammatory infiltrates surrounding the injection site to 5 indicating an extremely high level of infiltrates that permeated into the neural tissue surrounding the injection site. The scores obtained in this study ranged from 1 to 5 with most animals scoring between 2 and 4. There was no difference in the level of host immune response between injections of AdLacZ versus AdGDNF (Kruskal–Wallis P ⫽ 0.28) suggesting that both vectors resulted in the same level of host immune response. However, there was an increased immune response in the striatum when compared to the SN (score ⫽ 3.7 vs 2.2). This may be due to the striatum having received three injections, one of which was only 5 days prior to sacrifice, while the SN only received one injection. The three injections in the striatum may have produced a more extensive host immune response due to an additive effect. Effects of AdGDNF on Weight One of the side effects of GDNF protein injections intracerebrally or intraventricularly is significant weight loss (27, 28, 31). In this study, animal body weights were obtained at the beginning and end of the study. All treatment groups showed normal and similar levels of weight gain (52.75 ⫾ 2.54 g; F(4, 35) ⫽ 0.161, P ⫽ 0.956) during the experimental period, suggesting no deleterious effect of the AdGDNF injection on weight. DISCUSSION

Injection of AdGDNF into the SN, 1 week following 6-OHDA-induced damage of the nigrostriatal pathway, increased the number of dopaminergic neurons in the SN that maintained their connection to the striatum. Rats treated with AdGDNF in the SN also showed increased dopaminergic function in the striatum, as assessed by amphetamine-induced Fos expression, and did not develop the behavioral deficits observed in control rats. In contrast, injection of AdGDNF into the striatum resulted in no neuroprotection and no amelioration of behavioral deficits in this model. Our previous research demonstrated that injection of AdGDNF into either the striatum or the SN 1 week prior to a partial progressive 6-OHDA lesion also resulted in neuroprotection (10 –12). In aged rats, injection into the striatum, but not the SN, 1 week prior to lesioning, ameliorated behavioral deficits, decreased lesion size, and increased dopaminergic function (10 – 12). The apparent discrepancy between those studies

AdGDNF MAINTAINS DOPAMINERGIC FUNCTION AND CONNECTIONS

11

port of the AdGDNF vector with expression of the GDNF transgene in the DA neurons or through retrograde transport of the GDNF protein, secreted by cells in the striatum that expressed transgene (10, 11). In the present study, the extent of damage to the striatal fibers may have been too extensive at the time of vector injection, preventing the retrograde transport of GDNF protein or vector to the SN. Nonetheless, at day 35 postlesion, the number of FG⫹ neurons in the SN was slightly increased in rats that received AdGDNF into the striatum. It may be that striatal fibers recover over time and begin to be affected by the continuously synthesized GDNF, thus eliciting a slower recovery time course. Therefore, these findings do not completely rule out the striatum as a therapeutic site for gene therapy. They call for further exploration of the effectiveness of AdGDNF injected into the striatum at later time points or in other models of progressive degeneration of the SN, ones that do not involve immediate toxic destruction of striatal terminals such as the MPTP-treated mouse or monkey. The issue is important given that in most patients with PD dopaminergic fibers persist in the basal ganglia, thus providing a substrate for GDNF-induced neuroprotection.

FIG. 6. Amphetamine-induced rotation. (A) The number of rotations/minute was decreased in rats that received AdGDNF into the SN (*P ⬍ 0.04) but was not affected by injection of AdGDNF into the striatum (B).

and the results presented here can be explained by the temporal difference in the administration of the lesion relative to the delivery of the GDNF gene. In this study, AdGDNF was injected 1 week after an intrastriatal injection of 6-OHDA. It is well known that following an injection of 6-OHDA into the striatum there is immediate toxic damage to the dopaminergic fibers and axons in the striatal area followed by rapid degeneration of these terminals, which begins within 24 h postinjection. Mandel and colleagues demonstrated a loss of TH-IR (dopaminergic) fibers to the level of the globus pallidus 1 week after an intrastriatal 6-OHDA injection (30), the time point at which AdGDNF was injected into the striatum in the present study. We also have observed that dopaminergic terminals in the striatum cannot retrogradely transport FG at 30 min or 2 h after being exposed to 6-OHDA (Kozlowski et al., unpublished observations). The previously demonstrated effectiveness of striatal AdGDNF injection prior to lesioning may have been due to protection of these terminals against the acute toxic effects of the 6-OHDA. Moreover, neuroprotection may have resulted from continuous exposure of the DA neurons to high levels of GDNF for 1 week prior to the lesion. This could have occurred either through a retrograde trans-

FIG. 7. Ipsilateral forelimb use for weight-bearing movements prior to the 6-OHDA lesion (Day 0) and at different time points after the lesion. (A) There was a trend for a decrease in preference for ipsilateral forelimb use in animals treated with AdGDNF into the SN (P ⫽ 0.08). (B) AdGDNF into the striatum resulted in a preference for the ipsilateral forelimb similar to that in control animals.

12

KOZLOWSKI ET AL.

FIG. 8. GDNF-IR in the ST and SN 1 month postinjection of AdGDNF (35 days postlesion). The lesion did not interfere with expression of GDNF in either site.

The present study demonstrated that AdGDNF was effective only when injected into the degenerating SN. The degenerating nigra may provide an environment that is more receptive to AdGDNF. An upregulation of neurotrophic factors and their receptors has been demonstrated following neural injury, denervation, and seizures (7, 32). Specifically, following facial nerve injury in the rat there was an increase in mRNA expression of the signaling component of the GDNF receptor, c-ret and GFR-␣1 out to 21 days post injury (7). These findings suggest that injecting AdGDNF into the degenerating SN may have been more effective in this rescue paradigm due to the receptive nature of the degenerating nigral environment. In previous studies, injection into the SN 1 week prior to a progressive 6-OHDA lesion prevented cell death in the SN, but did not affect dopaminergic function or behavior (10, 12). However, this previously documented lack of improved dopaminergic function was demonstrated in aged rats (12) and has never been tested in young rats. Our observations are consistent, however, with studies in

young 6-OHDA-lesioned rats, which show improved behavioral function and protection of dopaminergic cell bodies with GDNF protein injections into the SN (2, 5, 18, 39). It is possible that AdGDNF injection into the SN of old rats may not have been able to induce functional and behavioral protection due to a decreased capacity for neural plasticity in the aged brain. Neuroprotection mediated by either GDNF protein or GDNF gene delivery has previously been measured by counting the number of TH-IR or prelabeled FG⫹ neurons surviving in the SN (2, 5, 6, 10 –12, 20, 36, 37, 39, 42, 43, 46). Although GDNF was found to increase survival of dopaminergic neurons, it is not clear whether these neurons maintained functional connections to the striatum. Using a prelabeling paradigm, Connor et al. (12) demonstrated that injections of AdGDNF into either the striatum or the SN resulted in significantly more FG⫹ neurons in the lesioned SN, suggesting that both sites resulted in neuroprotection. However, only injections of AdGDNF into the striatum resulted in behavioral amelioration, as well as an increase in dopaminergic terminals, suggesting that FG⫹ neurons protected by nigral injections of AdGDNF did not maintain nigrostriatal connections. The measure of neuroprotection in the present study focused on measuring the maintenance of nigrostriatal connections by injecting a retrograde tracer into the striatum at the end of the experiment. Consequently, surviving dopaminergic neurons that did not maintain a connection to the striatum were not labeled or counted. Using this method, it was demonstrated that the injection of AdGDNF into the SN maintained 52% of nigrostriatal connections (compared to the unlesioned SN) at 4 weeks postlesion. What is undetermined is whether the increase in nigrostriatal connections in this study is due to neuroprotection of remaining neurons or to the rescue of degenerating neurons. In the progressive lesion model, the number of prelabeled FG⫹ neurons decreases sharply from 94 to 59% during weeks 1 and 2 postlesion (38). Given this time course, it is likely that AdGDNF in this paradigm was performing a neuroprotective role rather than reversing the damage in compromised neurons. However, our data do not rule out the possibility that sprouting was induced in DA neurons. In addition to maintaining the nigrostriatal pathway, injection of AdGDNF into the SN resulted in the maintenance of dopaminergic function in the striatum as measured by amphetamine-induced expression of Fos. In normal rats, amphetamine-induced expression of Fos in the striatum relies on the activation of postsynaptic D1 receptors on striatonigral neurons in the striatum (1, 13, 25). Following a 6-OHDA lesion, there is a decrease in Fos expression, presumably due to the decrease in amphetamine-induced dopamine release in the lesioned hemisphere, since D1 receptors are not affected by intrastriatal 6-OHDA lesions (34).

AdGDNF MAINTAINS DOPAMINERGIC FUNCTION AND CONNECTIONS

This lesion-induced decrease in Fos induction can be reversed by intrastriatal grafts, presumably by their ability to increase dopamine levels (8). In the present study, AdGDNF may have maintained dopaminergic function in the striatum by increasing the release of dopamine. GDNF has been shown to increase dopamine levels and their metabolites in the SN of rats with complete 6-OHDA lesions of the medial forebrain bundle and in the MPTP monkey (16, 18, 26). In these models, GDNF did not affect basal levels of dopamine or its metabolites in the striatum. In normal rats, infusion of GDNF protein into the SN has been shown to increase dopamine turnover in the striatum (19). The progressive partial lesion results in a population of remaining dopaminergic fibers that have been shown to respond in a compensatory manner by increasing dopamine production (47). The effects of GDNF on dopamine and its metabolites in the progressive partial lesion have not been studied. The mechanisms by which AdGDNF in the SN could increase dopaminergic function in this model include stimulating compensatory release of dopamine from striatal fibers surrounding the lesion, increasing dopaminergic function indirectly through effects on other dopaminergic systems in the VTA or SN reticulata, or directly stimulating release of dopamine by rescued nigral neurons that project to the lesion site. These possibilities are currently being investigated using in vivo microdialysis. A novel finding of this study is the neuroprotection of nigrostriatal connections and dopaminergic function, without the neuroprotection of dopaminergic fibers in the striatum as assessed by TH immunocytochemistry. The FG⫹ neurons rescued by AdGDNF in the SN were obviously functional due to their capacity to uptake and transport FG. In addition, these neurons had TH⫹ cell bodies suggesting that they had the capacity to make dopamine, yet their terminals showed a level of TH immunoreactivity, similar to that seen in untreated lesioned rats. It is likely that the level of TH in the terminals of FG⫹/TH⫹ neurons, protected by AdGDNF, was very low, below the threshold of staining. Previous studies have demonstrated that the phenotypic marker, TH can be transiently lost and then rejuvenated by the administration of GDNF (15). It may be that the time point examined in this study preceded the rejuvenation of the phenotypic marker or that TH-IR in terminals may not be an unequivocal marker for dopaminergic function or fiber sprouting in the striatum. AdGDNF not only enhanced dopaminergic function in this study, but also prevented the development of DA-dependent behavioral deficits. Amphetamine-induced rotation was significantly decreased in animals that received AdGDNF into the SN. Amphetamineinduced rotation is influenced by asymmetries in the nigrostriatal dopamine system and reduced amphetamine-induced rotation may reflect an increase in do-

13

paminergic function. Although this concept has been challenged by investigators who demonstrated decreased rotation behavior without an increase in striatal dopamine, the reliability of this assay is improved when dopaminergic function is estimated via assessing levels of dopamine metabolites or Fos induction (44). A reduction in the preference for ipsilateral forelimb use was also evident; however, this was not statistically significant. Although forelimb use asymmetries are affected by nondopaminergic manipulation of the motor system, the forelimb test is very sensitive to dopamine depletion (Schallert, personal communication). AdGDNF may have maintained dopaminergic function in the lesioned terminals at a level which could be enhanced by stress such as amphetamine, but not high enough to ameliorate a non-drug-induced DA-sensitive test such as preference for ipsilateral limb use. It is also feasible that AdGDNF could be exerting its neuroprotective effects on other neurotransmitter systems, shown to be influenced by GDNF (31). Studies focusing on the effects of AdGDNF on other neurotransmitter systems are currently being explored. Many studies have demonstrated that GDNF can protect DA cell bodies. The crucial goal is to demonstrate that GDNF can maintain the function and efficacy of protected dopaminergic neurons. This study has demonstrated that an adenoviral vector harboring the GDNF gene injected into the SN can ameliorate dopaminergic function and protect dopaminergic neurons and their connections even when injected after damage to the nigrostriatal system has occurred. This study provides support for GDNF gene therapy as a possible therapeutic approach for maintaining dopaminergic function in the degenerating nigrostriatal system of humans with Parkinson’s disease. ACKNOWLEDGMENTS The authors acknowledge Ianina Filipovich, Angelica Oviedo, M.D., and Richard Anderson for their technical assistance and Dr. Beverly Davidson and the vector core at the University of Iowa for the production of the viral vectors. This project was supported by NIH Grants NS31957 and NS39267, the Medical Research Institute Council of Children’s Memorial Hospital (M.C.B.), and partial support from the Carver Foundation (Vector Core). The original AdGDNF was the gift of Genetic Therapy/Novartis.

REFERENCES 1.

Abrous, D. N., V. Bernard, M. Le Moal, B. Bloch, and J. P. Herman. 1996. Phenotype of striatal cells expressing c-Fos following amphetamine treatment of rats with intrastriatal dopaminergic grafts. Eur. J. Neurosci. 8: 2521–2529. 2. Beck, K. D., J. Valverde, T. Alexi, K. Poulsen, B. Moffat, R. A. Vandlen, A. Rosenthal, and F. Hefti. 1995. Mesencephalic dopaminergic neurons protected by GDNF from axotomy-induced degeneration in the adult brain. Nature 373: 339 –341. 3. Bilang-Bleuel, A., F. Revah, P. Colin, I. Locquet, J.-J. Robert, J. Mallet, and P. Horellou. 1997. Intrastriatal injection of an ad-

14

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

KOZLOWSKI ET AL. enoviral vector expressing glial cell-line derived neurotrophic factor prevents dopaminergic neuron degeneration and behavioural impairment in a rat model of Parkinson’s disease. Proc. Natl. Acad. Sci. USA 94: 8818 – 8823. Bjorklund, A., C. Rosenblad, C. Winkler, and D. Kirik. 1997. Studies on neuroprotective and regenerative effects of GDNF in a partial lesion model of Parkinson’s disease. Neurobiol. Dis. 4: 186 –200. Bowenkamp, K. E., A. F. Hoffman, G. A. Gerhardt, M. A. Henry, P. T. Biddle, B. J. Hoffer, and A. E. Granholm. 1995. Glial cell line-derived neurotrophic factor supports survival of injured midbrain dopaminergic neurons. J. Comp. Neurol. 355: 479 – 489. Bowenkamp, K. E., P. A. Lapchak, B. J. Hoffer, P. J. Miller, and P. C. Bickford. 1997. Intracerebroventricular glial cell linederived neurotrophic factor improves motor function and supports nigrostriatal dopamine neurons in bilaterally 6-hydroxydopamine lesioned rats. Exp. Neurol. 145: 104 –117. Burazin, T. C. D., and A. L. Gundlach. 1998. Up-regulation of GDNF-alpha and c-ret mRNA in facial motor neurons following facial nerve injury in the rat. Mol. Brain. Res. 55: 331–336. Cenci, M., P. Kalen, R. Mandel, K. Wictorin, and A. Bjorklund. 1992. Dopaminergic transplants normalize amphetamine and apomorphine-induced Fos expression in the 6-hydroxydopamine lesioned striatum. Neuroscience 46: 943–957. Choi-Lundberg, D. L., and M. C. Bohn. 1998. Applications of gene therapy to neurological diseases and injuries. In Stem Cell Biology and Gene Therapy (P. J. Quesenberry, G. S. Stein, B. Forget, and S. Weissman, Eds.), Wiley-VCH, New York. Choi-Lundberg, D. L., Q. Lin, Y.-N. Chang, Y. L. Chiang, C. M. Hay, H. Mohajeri, B. L. Davidson, and M. C. Bohn. 1997. Dopaminergic neurons protected from degeneration by GDNF gene therapy. Science 275: 338 –341. Choi-Lundberg, D. L., Q. Lin, T. Schallert, D. Crippens, B. L. Davidson, Y.-N. Chang, Y. L. Chiang, J. Qian, L. Bardwaj, and M. C. Bohn. 1998. Behavioral and cellular protection of rat dopaminergic neurons by an adenoviral vector encoding glial cell linederived neurotrophic factor. Exp. Neurol. 154: 261–275. Connor, B., D. A. Kozlowski, T. Schallert, J. L. Tillerson, B. L. Davidson, and M. C. Bohn. 1999. Differential effects of glial cell line-derived neurotrophic factor (GDNF) in the striatum and substantia nigra of the aged parkinsonian rat. Gene Ther. 6: 1936 –1951. Dragunow, M., and H. A. Robertson. 1987. Kindling stimulation induces c-fos protein in granule cells of the rat dentate gyrus. Nature 329: 441– 442. Gash, D. M., Z. Zhang, and G. Gerhardt. 1998. Neuroprotective and neurorestorative properties of GDNF. Ann. Neurol. 44: S121–S125. Gash, D. M., Z. M. Zhang, A. Ovadia, W. A. Cass, A. Yi, L. Simmerman, D. Russell, D. Martin, P. A. Lapchak, F. Collins, B. J. Hoffer, and G. A. Gerhardt. 1996. Functional recovery in Parkinsonian monkeys treated with GDNF. Nature 380: 252– 255. Gerhardt, G., W. A. Cas, P. Huettl, S. Brock, Z. Zhang, and D. M. Gash. 1999. GDNF improves dopamine function in the substantia nigra but not the putamen of unilateral MPTPlesioned rhesus monkeys. Brain Res. 817: 163–171. Hebert, M. A., and G. A. Gerhardt. 1997. Behavioural and neurochemical effects of intranigral administration of glial cell line-derived neurotrophic factor on aged Fischer 344 rats. J. Pharmacol. Exp. Ther. 282: 760 –768. Hoffman, A. F., C. G. van Horne, S. Eken, B. J. Hoffer, and G. A. Gerhardt. 1997. In vivo microdialysis studies of somatodendritic dopamine release in the rat substantia nigra: Effects of

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

unilateral 6-OHDA lesions and GDNF. Exp. Neurol. 147: 130 – 141. Hudson, J., A.-C. Granholm, G. Gerhardt, M. Henry, A. Hoffman, P. Biddle, N. Leela, L. Mackerlova, J. Lile, F. Collins, and B. Hoffer. 1995. Glial cell line-derived neurotrophic factor augments midbrain dopaminergic circuits in vivo. Brain Res. Bull. 36: 425– 432. Kearns, C., and D. Gash. 1995. GDNF protects nigral dopaminergic neurons against 6-hydroxydopamine in vivo. Brain Res. 672: 104 –111. Kirik, D., C. Rosenblad, and A. Bjorklund. 1998. Characterization of behavioral and neurodegeneration changes following partial lesions of the nigrostriatal dopamine system induced by intrastriatal 6-hydroxydopamine in the rat. Exp. Neurol. 153: 259 –277. Kordower, J. H., S. Palfi, E. Y. Chen, S. Y. Ma, T. Sendera, E. J. Cochran, E. J. Mufson, R. Enn, C. G. Goetz, and C. D. Comella. 1999. Clinicopathological findings following intraventricular glial-derived neurotrophic factor treatment in a patient with Parkinson’s disease. Ann. Neurol. 46: 419 – 424. Kozlowski, D. A., D. C. James, and T. Schallert. 1996. Usedependent degeneration of neuronal injury following unilateral sensorimotor cortex injury. J. Neurosci. 16: 4776 – 4786. Kozlowski, D. A., S. L. vonStuck, S. M. Lee, D. A. Hovda, and D. P. Becker. 1996. Behaviorally-induced contusions following traumatic brain injury: Use-dependent secondary insults. Soc. Neurosci. Abstr. 744.17. Labandeira-Garcia, J., G. Rozas, E. Lopez-Martin, I. Liste, and M. Guerra. 1996. Time course of striatal changes induced by 6-hydroxydopamine lesion of the nigrostriatal pathway, as studied by combined evaluation of rotational behaviour and striatal Fos expression. Exp. Brain Res. 108: 69 – 84. Lapchak, P., D. Araujo, D. Hilt, J. Sheng, and S. Jiao. 1997. Adenoviral vector-mediated GDNF gene therapy in a rodent lesion model of late stage Parkinson’s disease. Brain Res. 777: 153–160. Lapchak, P., D. Gash, F. Collins, D. Hilt, P. Miller, and D. Araujo. 1997. Pharmacological activities of glial cell line-derived neurotrophic factor (GDNF): Preclinical development and application to the treatment of Parkinson’s disease. Exp. Neurol. 145: 309 –321. Lapchak, P. A., P. J. Miller, F. Collins, and S. Jiao. 1997. Glial cell line-derived neurotrophic factor attenuates behavioural deficits and regulates nigrostriatal dopaminergic and peptidergic markers in 6-hydroxydopamine-lesioned adult rats: Comparison of intraventricular and intranigral delivery. Neuroscience 78: 61–72. Mandel, R., S. Spratt, R. Snyder, and S. Leff. 1997. Midbrain injection of recombinant adeno-associated virus encoding rat glial cell line-derived neurotrophic factor protects nigral neurons in a progressive 6-hydroxydopamine-induced degeneration model of Parkinson’s disease in rats. Proc. Natl. Acad. Sci. USA 94: 14083–14088. Mandel, R. J., D. Kirik, C. Rosenblad, and A. Bjorklund. 1999. Functional recovery after intrastriatal 6-OHDA lesions mediated by intrastriatal injections of recombinant AAV encoding rat GDNF. In Society for Neuroscience Meeting, Miami, FL, 1999. 131.16. Martin, D., G. Miller, T. Cullen, N. Fischer, D. Dix, and D. Russell. 1996. Intranigral or intrastriatal injections of GDNF: Effects on monoamine levels and behavior in rats. Eur. J. Pharmacol. 317: 247–256. Oyesiku, N. M., C. O. Evans, S. Houston, R. S. Darrell, J. S. Smith, Z. L. Fulop, C. E. Dixon, and D. G. Stein. 1999. Regional changes in the expression of neurotrophic factors and their receptors following acute traumatic brain injury in the adult rat brain. Brain Res. 833: 161–172.

AdGDNF MAINTAINS DOPAMINERGIC FUNCTION AND CONNECTIONS 33.

34.

35.

36.

37.

38.

39.

Paul, M. L., A. M. Graybiel, J. C. David, and H. A. Robertson. 1992. D1-like and D2-like dopamine receptors synergistically activate rotation and c-fos expression in the dopamine-depleted striatum in a rat model of Parkinson’s disease. J. Neurosci. 12: 3729 –3742. Przedborski, S., M. Levivier, H. Jiang, M. Ferreira, V. JacksonLewis, D. Donaldson, and D. M. Togasaki. 1995. Dose-dependent lesions of the dopaminergic nigrostriatal pathway induced by intrastriatal injection of 6-hydroxydopamine. Neuroscience 67: 631– 647. Robertson, H. A., M. R. Peterson, K. Murphy, and G. S. Robertson. 1989. D1-dopamine receptor agonists selectively activate striatal c-fos independent of rotational behavior. Brain Res. 503: 346 –349. Rosenblad, C., D. Kirik, B. Devaux, B. Moffat, H. S. Phillips, and A. Bjorklund. 1999. Protection and regeneration of nigral dopaminergic neurons by neurturin or GDNF in a partial lesion model of Parkinson’s disease after administration into the striatum or lateral ventricle. Eur. J. Neurosci. 11: 1554 –1566. Rosenblad, C., A. Martinez-Serrano, and A. Bjorklund. 1998. Intrastriatal glial cell line-derived neurotrophic factor promotes sprouting of spared nigrostriatal dopaminergic afferents and induces recovery of function in a rat model of Parkinson’s disease. Neuroscience 82: 129 –137. Sauer, H., and W. H. Oertel. 1994. Progressive degeneration of nigrostriatal dopamine neurons following intrastriatal terminal lesions with 6-hydroxydopamine a combined retrograde tracing and immunocytochemical study in the rat. Neuroscience 59: 401– 415. Sauer, H., C. Rosenblad, and A. Bjorklund. 1995. Glial cell line-derived neurotrophic factor but not transforming growth factor beta-3 prevents delayed degeneration of nigral dopaminergic neurons following striatal 6-hydroxydopamine lesion. Proc. Natl. Acad. Sci. USA 92: 8935– 8939.

40.

41.

42.

43.

44.

45.

46.

47.

15

Schallert, T., and J. Tillerson. 1999. Intervention strategies for degeneration of dopamine neurons in parkinsonism: Optimizing behavioral assessment of outcome. In Central Nervous System Diseases (D. F. Emerich, R. L. Dean III, and P. R. Sanberg, Eds.). Humana Press, Totowa, NJ. Schwarting, R., and J. Huston. 1996. The unilateral 6-hydroxydopamine lesion model in behavioural brain research analysis of functional deficits, recovery and treatments. Prog. Neurobiol. 50: 275–331. Shults, C. W., T. Kimber, and D. Martin. 1996. Intrastriatal injection of GDNF attenuates the effects of 6-hydroxydopamine. NeuroReport 7: 627– 631. Sullivan, A. M., J. Opacka-Juffry, and S. B. Blunt. 1998. Longterm protection of the rat nigrostriatal dopaminergic system by glial cell line-derived neurotrophic factor against 6-hydroxydopamine in vivo. Eur. J. Neurosci. 10: 57– 63. Tseng, J. L., E. E. Baetge, A. D. Zurn, and P. Aebischer. 1997. GDNF reduces drug-induced rotational behavior after medial forebrain bundle transection by a mechanism not involving striatal dopamine. J. Neurosci. 17: 325–333. Ungerstedt, U., and G. W. Arbuthnott. 1970. Quantitative recording of rotational behavior in rats after 6-hydroxydopamine lesions of the nigrostriatal dopamine system. Brain Res. 24: 485– 493. Winkler, C., H. Sauer, C. S. Lee, and A. Bjorklund. 1996. Short-term GDNF treatment provides long-term rescue of lesioned nigral dopaminergic neurons in a rat model of Parkinson’s disease. J. Neurosci. 16: 7206 –7215. Zigmond, M., E. Abercrombie, T. Berger, A. Grace, and E. Stricker. 1993. Compensatory responses to partial loss of dopaminergic neurons: Studies with 6-hydroxydopamine. In Current Concepts in Parkinson’s Disease Research (J. Schneider and M. Gupta, Eds.), pp. 99 –140. Hogrefe and Huber, Seattle.