Gene Therapy (1998) 5, 1650–1655 1998 Stockton Press All rights reserved 0969-7128/98 $12.00 http://www.stockton-press.co.uk/gt Astrocyte-specific expression of tyrosine hydroxylase after intracerebral gene transfer induces behavioral recovery in experimental Parkinsonism J Segovia1, P Vergara1 and M Brenner2 1Departamento de Fisiologı´a, Biofı´sica y Neurociencias, Centro de Investigacio´n y de Estudios Avanzados del IPN, Me´xico; and 2Stroke Branch, National Institute of Neurological Disorders and Stroke, NIH, Bethesda, MD, USA Parkinson’s disease is a neurodegenerative disorder in a rat model of Parkinson’s disease produced by lesion- characterized by the depletion of dopamine in the caudate ing the striatum with 6-hydroxydopamine. Following putamen. Dopamine replacement with levodopa, a precur- microinjection of the transgene into the denervated stria- sor of the neurotransmitter, is presently the most common tum as a DNA–liposome complex, expression of the trans- treatment for this disease. However, in an effort to obtain gene was detected by RT-PCR and TH protein was better therapeutic results, tissue or cells that synthesize observed specifically in astrocytes by using double-labeling catecholamines have been grafted into experimental ani- immunofluorescence for GFAP and TH coupled with laser mals and human patients. In this paper, we present a novel confocal microscopy. Efficacy was demonstrated by sig- technique to express tyrosine hydroxylase (TH) in the nificant behavioral recovery, as assessed by a decrease in host’s own astrocytes. This procedure uses a transgene in the pharmacologically induced turning behavior generated which the expression of a TH cDNA is under the control of by the unilateral denervation of the rat striatum. These a glial fibrillary acidic protein (GFAP) promoter, which con- results suggest this is a valuable technique to express fers astrocyte-specific expression and also increases its molecules of therapeutic interest in the brain. activity in response to brain injury. The method was tested Keywords: tyrosine hydroxylase; glial fibrillary acidic protein; promoter; intracerebral gene transfer; Parkinson’s disease; 6-hydroxydopamine (6-OHDA) Introduction in response to dopaminergic agonists. This model was used to demonstrate the potential of catecholamine-pro- Parkinson’s disease is a neurodegenerative disorder ducing grafts in 1979,3,4 and since then the amount of caused by the death of the dopaminergic neurons of the experimental work in this area has grown at an extraordi- substantia nigra. It is generally accepted that levodopa (L nary pace in attempts to produce transplants that are 3–4-dihydroxyphenylalanine), a precursor of dopamine safer, more readily available, longer lasting and express (DA) that can readily cross the blood–brain barrier, is the more strongly. Transplants used include fetal nigral most effective agent to control the symptoms of Parkin- tissue and adrenal medulla, as well as primary cells and son’s disease. Most Parkinson’s patients have a good cell lines genetically engineered to express tyrosine initial response to levodopa, but after a few years become hydroxylase (TH), the limiting enzyme in catecholamine subject to adverse effects, which include dyskinesia, fluc- synthesis (for reviews see Refs 5 and 6). Various methods tuations of efficacy (on–off effect), freezing, mental of gene transfer have also been explored to engineer the 1 changes and loss of efficacy. Transplantation of cat- cultured cells, from classical transfections to the use of echolamine-producing tissue has received considerable retroviral, herpes virus and adenoviral vectors (for attention as an alternative therapy that delivers DA reviews see Refs 7 and 8). More recently, direct transfer directly to the striatum, sparing other tissues from of TH transgenes into the brain has been accomplished adverse effects of DA stimulation and metabolism, and using viral delivery systems,9–11 or liposome complexes.12 avoiding the drug peaks and valleys of pharmacological In general, these experiments have demonstrated that administration by providing a relatively constant source. intracerebral grafts survive within the adult brain, and The most used rodent model of Parkinson’s disease is that they induce behavioral recovery that is consistently the destruction of nigral dopaminergic neurons by the associated with increases in dopaminergic markers, such 2 stereotactic injection of 6-hydroxydopamine (6-OHDA). as TH immunoreactivity and synthesis and release of DA. When this lesion is performed unilaterally, rats will turn These results also provide further evidence that DA replacement is sufficient to ameliorate some of the symp- toms of the disease. The success in experimental animals Correspondence: J Segovia, Departamento de Fisiologı´a, Biofı´sica y Neuro- ciencias, Centro de Investigacio´n y de Estudios Avanzados del IPN, Av has led to the use of brain transplants to treat human Instituto Polite´cnico Nacional 2508, Me´xico, 07300, D.F. patients. Both adrenal tissue, from the same patient, and Received 18 February 1998; accepted 8 July 1998 fetal tissue have been transplanted into Parkinson’s Astrocyte-specific expression of TH in experimental Parkinsonism J Segovia et al 1651 patients, with varied degrees of success.13–16 Never- theless, safety, availability and duration and level of DA production remain paramount issues. In this article, we present a new system which addresses these issues. The method involves engineering the host’s own astrocytes to produce DA by introducing a transgene consisting of a TH cDNA under the control of an astrocyte-specific promoter derived from the glial fibrillary acidic protein (GFAP) gene. Using the host’s cells to produce DA avoids complications of rejection and of obtaining tissues for implant. The GFAP promoter also confers several advantages. By confining expression to astrocytes, it targets a presumably long-lived cell popu- lation, and one that can be expected to be more resistant than neurons to the oxidative stress that is a byproduct of DA metabolism.17 The GFAP promoter also has a built- in regulatory system that responds to reactive gliosis by increasing its transcriptional activity.18 Glial scars are observed in the substantia nigra of Parkinson’s patients, as well as in the caudate putamen and globus pallidus in advanced cases of Parkinson’s and in autosomal domi- Figure 1 Effect of gene transfer on turning behavior. The graph shows nant parkinsonism.19–21 Thus it is a reasonable expec- the turns per minute in response to apomorphine as a percentage of that tation that the activity of a GFAP promoter-driven TH observed before gene transfer (Pre). Rats were tested at 1, 2 and 3 weeks transgene will be increased in response to the progression after lipofection with either pGfa2-cLac or pGfa2-TH. Data presented are means; bars are s.e.m. *P Ͻ 0.05 as compared with Pre (ANOVA, fol- of the disease. lowed by Dunnett’s multiple comparison test). The efficacy of this system was tested by lipofection of the GFAP-TH transgene (pGfa2-TH) into the striatum of the 6-OHDA lesioned rat. The lipofection procedure was weeks after the gene transfer procedure. This result indi- used for this feasibility test as it avoids the possible toxic cates that the TH-containing plasmid induces behavioral effects of the more commonly used viral transfection sys- recovery in an experimental model of Parkinson’s tems.12,22,23 We report that the lipofected TH transgene is disease. expressed in the lesioned striatum, and that its expression is restricted to astrocytes. Furthermore, there is a signifi- Vector expression cant behavioral recovery of the lesioned rats. In order to demonstrate that the behavioral recovery can be attributed to the expression of the gfa2-TH transgene, Results we used RT-PCR to assay for the presence of TH mRNA in rats that underwent the gene transfer procedure. Three Behavioral effect days after the the third behavioral test (3 weeks after in The potential of the gfa2-TH transgene for treatment of vivo lipofection), rats were killed and total cellular RNA Parkinson’s disease was tested using the rat model in was separately isolated from each striatum. As can be which striatal lesions are produced by 6-OHDA (see observed in Figure 2, a signal was obtained that corre- Materials and methods). Following the lesioning, rats sponds to TH mRNA in the side ipsilateral to the 6- that turned at least five times per minute in response to OHDA lesion, where the pGfa2-TH vector–liposome apomorphine and seven times per minute in response to complex was injected, whereas there was no signal from amphetamine were selected for the gene transfer experi- the non-lesioned side. No TH mRNA was detected either ments, as we have previously shown that these responses in the lesioned or control striata of rats that received the indicate a greater than 95% decrease in striatal dopamine pGfa2-cLac vector (data not shown). and an almost complete disappearance of TH in the stria- tum ipsilateral to the lesion.24,25 One to 2 weeks after the Tyrosine hydroxylase and ␤-galactosidase protein behavioral testing, rats were subjected to the gene trans- expression fer procedure. Rats were randomly assigned to two dif- Immunohistochemistry was used to ascertain if the TH ferent experimental groups. The test group received mRNA was translated into protein and if so, if its 10 ␮g of the pGfa2-TH plasmid and the control group expression was localized to astrocytes. Coronal sections received 10 ␮g of plasmid pGfa2-cLac, in which the lacZ close to the injection sites were immunostained for TH gene replaces the TH cDNA. To increase the volume of and GFAP using secondary antibodies tagged with striata subjected to the gene transfer procedure, the fluorescein and with rhodamine, respectively. This dou- DNA–liposome complexes were injected in two different ble labeling technique allowed us to determine the co- positions separated by 1.5 mm in the antero-posterior localization of both markers by merging the images axis.