Lentivirally Delivered Glial Cell Line-Derived Neurotrophic Factor Increases the Number of Striatal Dopaminergic Neurons in Primate Models of Nigrostriatal Degeneration

MATERIALS AND METHODS

Experiment 1: effect of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine on striatal tyrosine hydroxylase-immunoreactive cells. This experiment was performed to confirm previous findings now using stereological counting procedures to show that MPTP-treated monkeys have increased numbers of TH-ir striatal neurons (Betarbet et al., 1997). Four rhesus monkeys (4–5 years of age) received two unilateral intracarotid injections of MPTP separated by 1 month according to a protocol described previously (Kordower et al., 2000) (Table1). Briefly, animals were tranquilized with ketamine (10 mg/kg, i.m.) and then maintained on an anesthetic plane with isoflurane (1–2%). The animals were put in the supine position. For each injection, a right-sided incision was made along the medial edge of the sternocleidomastoid muscle. The carotid sheath was opened and the common carotid artery, internal jugular vein, and vagus nerves were identified. The common carotid was exposed below the carotid bifurcation. The external carotid artery was then ligated. A 27 gauge butterfly needle was inserted into the common carotid artery in a direction retrograde to blood flow; for each injection, 20 ml of saline containing 3 mg of MPTP-HCl was infused at a rate of 1.33 ml/min (15 min). After the infusion was completed, 3 ml of saline was delivered, and then the incision was closed. Two additional untreated rhesus monkeys served as unoperated controls.

Experiment 2: lentiviral delivery of glial cell line-derived neurotrophic factor to aged monkeys. This experiment investigated the effects of in vivo GDNF gene delivery on the number of striatal TH-ir cells in aged rhesus monkeys. The details of procedures performed on these monkeys have been reported previously (Kordower et al., 2000). In this experiment, four aged rhesus monkeys (24–27 years of age) received intrastriatal injections of lentivirus encoding for GDNF, and three monkeys received control injections of lentivirus encoding for β-galactosidase (β-gal) (Table 1). All animals received six lentivirus injections targeted for the caudate nucleus (n = 2 injections; 5 and 10 μl), putamen (n = 3; 10, 10, and 5 μl), and substantia nigra (n = 1; 5 μl). Coordinates were based on magnetic resonance imaging as described previously (Kordower et al., 2000). Monkeys were killed 3 months after the surgery. The construction of the vectors has been described previously (Kordower et al., 1999,2000; Deglon et al., 2000). Three months after the lentivirus treatment, all animals were killed.

Experiment 3: lentiviral delivery of glial cell line-derived neurotrophic factor to monkeys treated with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine. This experiment tested the hypothesis that GDNF delivered via a lentiviral vector would increase the number of TH-ir striatal neurons in MPTP-treated monkeys. Eight monkeys comprised this experiment, the methodological details of which have been reported previously (Kordower et al., 2000). Briefly, monkeys received a single injection of MPTP (3 mg) into the right carotid artery followed 1 week later by injections of GDNF delivered by a lentiviral vector (lenti-) (n = 4) or lenti-β-gal (n = 4) into the right caudate nucleus, putamen, and substantia nigra. The injection scheme was identical to that described above for aged monkeys. Three months after the lentivirus treatment, all animals were killed.

Preparation of tissues. Rhesus monkeys were killed with an overdose of pentobarbital, perfused transcardially with saline, and postfixed with 4% paraformaldehyde solution in phosphate buffer (0.1m), pH 7.4, at 4°C. The brains were removed immediately and cut into coronal blocks for postfixation. Tissue blocks were then washed in a series of cold graded sucroses and sectioned (40 μm) in the coronal plane on a freezing microtome.

Immunohistochemistry. Sections through the forebrain and midbrain were processed for the visualization of TH, DAT, or GDNF immunoreactivity using modifications of procedures published previously (Kordower et al., 1999, 2000). Briefly, endogenous peroxidase activity was eliminated with a 20 min incubation in 0.1 msodium periodate. Background staining was then inhibited by a 1 hr incubation in 3% normal serum and 2% bovine serum albumin (BSA), after which the tissue was incubated for 48 hr in the primary antibody. The concentrations of the primary antibodies were as follows: TH (1:20,000; Chemicon, Temecula, CA), DAT (1:500; generously provided by Dr. Alan Levey, Emory University, Atlanta, GA), or GDNF (1:250; Transduction Laboratories, Lexington, KY). Using a labeled antibody procedure, sections were then sequentially incubated in the appropriate biotinylated IgG (1:200; Vector Laboratories, Burlingame, CA) for 1 hr, washed in a Tris-buffered saline solution containing 0.05% Triton X-100, and incubated in “Elite” avidin–biotin complex (ABC) (1:1000; ABC kits; Vector Laboratories) for 75 min. The chromogen solution that completed the reaction consisted of 0.05% diaminobenzidine and 0.005% H₂O₂. For the GDNF staining, the immunohistochemical reaction product was intensified by adding 2.5% nickel II sulfate to the chromogen solution. Sections were mounted on gelatin-coated slides, dehydrated through graded alcohols (70, 95, and 99%), cleared in xylenes, and coverslipped with Permount (Fisher Scientific, Pittsburgh, PA). Additional sections were processed simultaneously in an identical manner except for the deletion of the primary antibody or substitution of the primary antibody with an irrelevant IgG matched for protein concentration. No specific immunoreactivity was seen for any stain under these conditions.

Histological analysis. In experiment 1, stereological counts of striatal TH-ir neurons were performed bilaterally on MPTP-treated animals and unilaterally in control animals. In experiments 2 and 3, stereological counts of TH-ir, GDNF-ir, or DAT-ir neurons were performed unilaterally on the side of the lenti-GDNF or lenti-β-gal infusions. For all experiments, stereological counts were performed using a computerized optical disector system. This method allows for an unbiased stereological quantification of a defined portion of a structure independent of its size, shape, orientation, tissue shrinkage, or anatomical level (Emborg et al., 1998). The computerized optical disector system consists of a computer-assisted image analysis, a microscope, a computer-controlled x, y, z motorized stage, a stereological software program (MicroBrightField Inc., Colchester, VT), and a high-sensitivity video camera. Before each series of measurements, the instrumentation was calibrated. The striatum was outlined under low magnification (1.25×). Five percent of the outlined region was quantified using a disector counting frame and a systematic random-sampling design. The average section thickness for each section and the antibody penetration throughout the entire tissue section were determined empirically. The total number of striatal neurons was quantified using a high magnification (100×) planapo oil immersion objective with a 1.4 numerical aperture. Under the disector principle, at least 200 striatal neurons were identified in each case. Once the top of the section was in focus, the z-plane was lowered 1–2 μm. Care was taken to ensure that the top and bottom forbidden planes were not included in the analysis. The total number of TH-ir striatal neurons (N) was calculated using the following formula: N =N_V ×V_Striatum, whereN_V is the numerical density andV_Striatum is the volume of the striatum as determined by the Cavalieri principle (1966).

Confocal double immunofluorescence of tyrosine hydroxylase with neuronal-specific nuclear protein. Based on tissue availability, double-immunofluorescence experiments were performed with tissue from experiment 2 to determine whether the TH-ir striatal cells coexpressed the neuronal marker neuronal-specific nuclear protein (NeuN) or the trophic factor GDNF. Sections were first incubated in a blocking solution (5% normal goat serum, 2% BSA, and 0.3% Triton X-100 in TBS, pH 7.4) for 1 hr to inhibit background staining. Then sections were incubated in primary rabbit polyclonal anti-TH (1:1000; Chemicon) for 24 hr at 4°C. After three washes, sections were incubated in secondary goat anti-rabbit IgG coupled to the fluorescent marker Cy2 (1:200; Jackson ImmunoResearch, West Grove, PA) for 1 hr. After completion of the reaction, sections were washed and background staining was inhibited by incubating the sections in a blocking solution containing 5% normal horse serum, 2% BSA, and 0.3% Triton X-100 in TBS, pH 7.4. Then sections were incubated in mouse monoclonal anti-NeuN (1:1000; Chemicon) for 24 hr at 4°C followed by incubation in goat anti-mouse IgG coupled to the fluorescent marker rhodamine (1:200, Jackson ImmunoResearch) for 1 hr at room temperature.

Confocal double immunofluorescence of tyrosine hydroxylase with glial cell line-derived neurotrophic factor. To assess double labeling of TH-ir cells with the neurotrophic marker GDNF, sections were immersed in the blocking solution as described above for 1 hr and then incubated in primary mouse monoclonal anti-TH (1:10,000; Incstar, Stillwater, MN) for 24 hr at 4°C. After three washes, sections were incubated in secondary goat anti-mouse IgG coupled to the fluorescent marker Texas Red (1:200; Jackson ImmunoResearch) for 1 hr at room temperature. Then sections were blocked with 5% normal horse serum, 2% BSA, and 0.3% Triton X-100 in TBS, pH 7.4, and incubated in goat polyclonal anti-GDNF (1:200; Chemicon) for 24 hr at 4°C and in biotinylated secondary donkey anti-goat IgG (1:200; Vector Laboratories) coupled to the fluorescent marker Cy2 for 1 hr at room temperature. Statistical analysis. The data were analyzed using a two-way ANOVA followed by a Fisher's PLSDpost hoc test when there was a significant overall ANOVA (Statview 4.0; Abacus Concepts, Calabasas, CA).

RESULTS

DISCUSSION

The present series of experiments demonstrated that the number of TH-ir striatal cells in nonhuman primates increases in response to dopaminergic depletion; this effect is enhanced by the lentiviral delivery of GDNF. Double-labeling studies in lentivirus-treated animals revealed that these TH-ir cells coexpressed NeuN-ir, indicating that cells expressing this dopaminergic marker were neurons. In addition, TH-ir neurons coexpressed GDNF-ir, indicating that the induction of the TH phenotype may be occurring via an autotrophic mechanism. It should be noted that the increase in DAT-ir cells after GDNF treatment was similar to that observed for TH-ir and that the numbers of TH-ir and DAT-ir cells were strongly correlated. These data suggest that the effects of dopamine depletion and GDNF administration provide concomitant influences on the genes responsible for TH and DAT expression.

A previous study found increases in TH-ir striatal neurons after the unilateral or bilateral nigrostriatal lesions engendered by MPTP in nonhuman primates (Betarbet et al., 1997). This study quantified TH-ir neurons in four anatomically matched sections and did not have tissue prepared in a manner that was compatible with a stereological analysis. Stereological counts of TH-ir neurons in the present study confirmed the increase in neuronal number after MPTP-induced dopaminergic denervation. Indeed, in the study by Betarbet et al. (1997), bilateral increases in TH-ir striatal neurons were observed after unilateral MPTP injections. Stereological analyses in the present study confirmed this surprising effect as well. The fact that an increase in TH-ir striatal neurons was observed on the “intact side” strongly indicates that the increases seen on the side with lesions were authentic and not attributable to the fact that they were easier to identify and count after removal of the TH-ir striatal neuropil. However, it cannot be ruled out that the increase in TH-ir striatal neurons seen on the intact side may have resulted from a small loss of nigrostriatal neurons that occurred as a result of MPTP crossing over to the contralateral side after the unilateral intracarotid injection of MPTP. In this regard, a small loss of TH-ir has been reported previously in monkeys on the side contralateral to the MPTP injection (Bankiewicz et al., 1999). If MPTP crossover mediated the increase in TH-ir neurons on the intact side, this would suggest that just a small loss of striatal dopamine is capable of eliciting this plasticity response.

It should be noted that the number of TH-ir striatal neurons in the control MPTP/lenti-β-gal-treated animals in experiment 3 was less than what was observed in experiment 1, in which monkeys were treated with MPTP alone. There are a number of factors that could have contributed to this difference in neuronal number. In experiment 1, the monkeys were exposed to multiple MPTP injections and were killed 5 months after surgery, whereas the monkeys in experiment 3 received a single MPTP injection and were killed 3 months postoperatively. Thus, repetitive injections combined with a longer postoperative time course could be responsible for this discrepancy, In addition, technical issues could have played a role in this difference. In each of the three experiments, tissue from control and experimental animals was histochemically processed together. However, tissues from the different experiments were processed independently, using different lots of antibody and different reagents. The intensity of staining was much greater in general through all dopaminergic regions in the tissues processed in experiments 1 and 2 compared with experiment 3; this likely contributed to the differences in absolute numbers. The critical comparisons to be concerned with in this study are the differences between conditions within an experiment rather than the absolute number of neurons responding to dopamine depletion and lenti-GDNF treatment.

The molecular mechanism underlying this plasticity response remains to be elucidated. Endogenous trophic factor release by glia in response to MPTP treatment has been reported as one potential stimulator of the normally quiescent TH gene (Francis et al., 1995) and may account for the increase in TH-ir cells. Regardless of whether this mechanism is in effect, the present data suggest that the increase of TH-ir cells is a compensatory response to striatal dopamine depletion. However, the fact that nonhuman primates become symptomatic so quickly after intracarotid MPTP administration suggests that this response is insufficient to sustain functional effects in these animals.

In this study, lenti-GDNF dramatically increased the number of TH-ir striatal neurons in both aged monkeys and monkeys rendered parkinsonian via MPTP. Aged rhesus monkeys undergo phenotypic losses of nigrostriatal dopaminergic markers, as evidenced by the fact that 50% of nigral neurons fail to express TH and 30% fail to express DAT (Emborg et al., 1998). Intracarotid injection of MPTP induces a more robust and frank degeneration of nigrostriatal neurons, in which often up to 90% of nigrostriatal neurons are lost (Bankiewicz et al., 1999). It is interesting that lenti-GDNF treatment was equipotent in increasing the number of TH-ir striatal neurons in MPTP-treated animals, in which the lesion was more severe relative to aged monkeys; in aged monkeys the lesion is more modest and does not produce frank neuronal degeneration. This suggests that the underlying status of the host system does not influence the ability of lenti-GDNF to modify the expression of TH-ir striatal neurons.

The mechanism by which GDNF increases the number of TH-ir neurons is not yet known. In the nigrostriatal dopaminergic system, GDNF has been shown to signal through a multireceptor complex composed of a novel glycosylphosphatidylinositol-anchored GDNF receptor and the receptor tyrosine kinase product of the c-ret proto-oncogene (Durbec et al., 1996; Trupp et al., 1996). The expression of this receptor complex is maximal during early postnatal development but it is at very low levels in the adult striatum (Trupp et al., 1997). As stated previously, the vast majority of TH-ir cells coexpressed GDNF-ir, suggesting that this trophic factor functioned in an autotrophic manner. Interestingly, there appears to be specificity in this response, because not all cells that were successfully transfected with the lentivirus expressed TH-ir. Unfortunately, sufficient tissue sections were not available to analyze the expression of ret and GDNF receptor α after GDNF delivery and to address some of these issues of mechanisms. Such studies are fertile ground for further investigation.

Some potential mechanisms that mediate the augmentation of striatal TH-ir neurons have been proposed. One hypothesis is that existing GABAergic striatal interneurons are converted to a TH-ir phenotype (Betarbet et al., 1997). This theory is based on a morphology analysis and colocalization of TH-ir with GAD-ir in the MPTP primate model of PD. In the present experiment, double labeling of TH-ir and GAD-ir were inconclusive (data not shown). One other potential mechanism may involve neural stem cells that migrate and differentiate to TH-ir-positive neurons that result from dopamine depletion and/or trophic factor release (Svendsen and Smith, 1999). This later hypothesis has important implications for neural plasticity, but it needs to be confirmed experimentally.

In summary, the present study revealed that the number of TH-ir striatal cells in nonhuman primates increased in response to dopaminergic depletion; this effect was further magnified by GDNF treatment. In aged monkeys receiving lentivirus encoding for GDNF, many TH-ir cells coexpressed GDNF-ir and NeuN-ir staining, indicating that these neurons are converted to a phenotype directly by the trophic effects of GDNF. In both aged and MPTP-treated monkeys, lenti-GDNF increases the number of TH-ir neurons by more than sevenfold, illustrating the potency of this trophic factor for dopaminergic neurons other than those arising from the ventral midbrain. In this role, GDNF may mediate plasticity in the dopamine-depleted primate brain and may serve to compensate for cell loss by converting striatal neurons to a dopaminergic phenotype. Additional studies are needed to determine whether these TH-ir striatal cells are extant striatal neurons or are generated from differentiated progenitor cells, and whether TH-ir cells are capable of producing dopamine and exerting functional effects.