In vivo transduction of murine cerebellar Purkinje cells by HIV-derived lentiviral vectors

BR A IN RE S E A RCH 1 0 82 ( 20 0 6 ) 1 1 –2 2

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s

Research Report

In vivo transduction of murine cerebellar Purkinje cells by HIV-derived lentiviral vectors Takashi Torashimaa,b,d , Shigeo Okoyamac , Tomoyuki Nishizakid , Hirokazu Hiraia,b,⁎ a

Innovative Brain Science Project, Advanced Science Research Center, Kanazawa University, Kanazawa 920-8640, Japan PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan c Laboratory of Neuroanatomy, Center for Biomedical Research and Education, Graduate School of Medical Science, Kanazawa University, Kanazawa 920-8640, Japan d Department of Physiology, Hyogo College of Medicine, 1-1 Mukogawa-cho, Nishinomiya 663-8501, Japan b

A R T I C LE I N FO

AB S T R A C T

Article history:

Cerebellar Purkinje cells are key elements in motor learning and motor coordination, and

Accepted 22 January 2006

therefore, it is important to clarify the mechanisms by which Purkinje cells integrate

Available online 6 March 2006

information and control cerebellar function. Gene transfer into neurons, followed by the assessment of the effects on neural function, is an effective approach for examining gene

Keywords:

function. However, this method has not been used fully in the study of the cerebellum

Lentivirus

because adenovirus vectors, the vectors most commonly used for in vivo gene transfer, have

HIV

very low affinity for Purkinje cells. In this study, we used a human immunodeficiency virus

Purkinje cell

(HIV)-derived lentiviral vector and examined the transduction profile of the vector in the

Cerebellum

cerebellum. A lentiviral vector carrying the GFP gene was injected into the cerebellar cortex.

Bergmann glia

Seven days after the injection, Purkinje cells were efficiently transduced without significant

Stellate cell

influence on the cell viability and synaptic functions. GFP was also expressed, though less efficiently, in other cortical interneurons and Bergmann glias. In contrast to reported findings with other viral vectors, no transduced cells were observed outside of the cerebellar cortex. Thus, when HIV-derived lentiviral vectors were injected into the cerebellar cortex, transduction was limited to the cells in the cerebellar cortex, with the highest tropism for Purkinje cells. These results suggest that HIV-derived lentiviral vectors are useful for the study of gene function in Purkinje cells as well as for application as a gene therapy tool for the treatment of diseases that affect Purkinje cells. © 2006 Elsevier B.V. All rights reserved.

⁎ Corresponding author. Innovative Brain Science Project, Advanced Science Research Center, Kanazawa University, Kanazawa 920-8640, Japan. Fax: +81 76 265 2489. E-mail address: [email protected] (H. Hirai). 0006-8993/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2006.01.104

12

BR A IN RE S EA RCH 1 0 82 ( 20 0 6 ) 1 1 –22

Abbreviations: AAV, adeno-associated virus CF, climbing fiber EPSC, excitatory postsynaptic current FIV, feline immunodeficiency virus GFAP, glial fibrillary acidic protein HEK, human embryonic kidney HIV, human immunodeficiency virus HSV1, herpes simplex virus type 1 MSCV, murine embryonic stem cell virus NeuN, neuron-specific nuclear protein PBS, phosphate-buffered saline PF, parallel fiber SCA, spinocerebellar ataxia TU, transducing units VSV-G, vesicular stomatitis virus G protein

1.

Introduction

The cerebellum plays substantial roles in motor coordination and motor learning (Ito et al., 2002). Damage of the cerebellum causes numerous motor coordination problems that include oculomotor disturbances, postural instability, and gait and limb ataxia. Cerebellar Purkinje cells, the sole outputs from the cerebellar cortex, are key elements regulating the cerebellar function. Purkinje cells receive two excitatory inputs from parallel fibers (PFs), axons of granule cells and climbing fibers (CFs), axons of inferior olivary neurons, whose excitatory activity is modulated by two sorts of inhibitory interneurons, stellate cells and basket cells (Figs. 1a, b). Transfer of exogenous genes into cerebellar Purkinje cells would allow us to dissect the molecular mechanisms regulating various physiological events in Purkinje cells such as dendritic development, synapse formation, and synaptic plasticity; however, gene transfer into Purkinje cells has remained a big challenge. In the late 1990s, two reports demonstrated LacZ gene expression in Purkinje cells using adenovirus vectors, but the transduction efficiency was very low: only a few Purkinje cells were β-galactosidase-positive despite the transduction of thousands of glial cells (Hashimoto et al., 1996; Terashima et al., 1997), suggesting that adenovirus vectors are not efficiently incorporated into Purkinje cells. Since 2000, neurotrophic vectors derived from adenoassociated virus (AAV), herpes virus and immunodeficiency virus were used for gene delivery into Purkinje cells. Recent in vivo gene expression experiments using those viral vectors revealed that the efficiency of transduction of Purkinje cells by each virus depends on the injection site. With herpes simplex virus type 1 (HSV1) vectors, injection into the inferior olive, but not into the cerebellar cortex, causes efficient transduction of Purkinje cells (Agudo et al., 2002): the vectors are delivered from the inferior olive to the deep cerebellar nuclei and then to Purkinje cells by retrograde axonal transport. On the other hand, AAV vectors injected into the cerebellar cortex are

directly taken up by Purkinje cells and transduce them efficiently (Alisky et al., 2000; Kaemmerer et al., 2000). These results suggest that both HSV1 and AAV vectors appear to be effective means for transferring genes into Purkinje cells; however, these vectors have several disadvantages. For example, herpes vectors could cause an inflammatory response that would affect the viability of transduced cells, and the insert size in AAV vectors is limited to only ∼4 kb. Alisky et al. (2000) tested a lentiviral vector derived from feline immunodeficiency virus (FIV), whose insert capacity (∼8 kb) is two times as large as that of AAV vectors, to transfer a reporter gene into Purkinje cells and showed that the FIVderived vectors injected into the cerebellar cortex transduced Purkinje cells, like HSV1 and AAV vectors. Furthermore, unlike herpes virus-based vectors, lentiviral vectors are known to elicit no or minimal inflammatory response. Thus, lentiviral vectors seem to be a promising tool for gene transfer into Purkinje cells. Recently, a human immunodeficiency virus (HIV)-derived lentiviral vector has been increasingly used for gene transfer into neurons (de Almeida et al., 2001; Desmaris et al., 2001; Dittgen et al., 2004; Naldini et al., 1996). Despite this trend, there has been no report yet on gene delivery into cerebellar neurons using HIV-derived vectors. Therefore, in the present study, we tested the potential of HIV-derived lentiviral vector to transduce cerebellar Purkinje cells and found a unique feature of this viral vector in transducing neurons in the cerebellar cortex.

2.

Results

2.1. Cerebellar cells transduced by HIV-derived lentiviral vectors Lentiviral vectors carrying the GFP gene (Fig. 1c) were generated in HEK 293 T cells, and the virus-containing medium was concentrated by ultracentrifugation. We used the viral vectors

BR A IN RE S E A RCH 1 0 82 ( 20 0 6 ) 1 1 –2 2

13

Fig. 1 – Schema of a sagittal view of a mouse cerebellum depicting a viral injection site (a) and neural circuit in the cerebellum (b). Viral solution was injected into lobule 6 at the cerebellar vermis. The cerebellar cortex receives input from several nuclei in the brain stem, including the pontine nuclei (Pn). Neurons in those nuclei send axons (mossy fibers; MF) to the granule cells (GC). Parallel fibers (PF), axons of granule cells, make excitatory synapses with Golgi cells (Go) and dendritic spines of Purkinje cells (PC). Golgi cells, in turn, inhibit the activity of granule cells. Purkinje cells receive another excitatory input from neurons of the inferior olivary nuclei (IO) via climbing fibers (CF). The excitation of Purkinje cells is modulated by two inhibitory interneurons, stellate cells (St) and basket cells (Ba). Purkinje cells, the sole output from the cerebellar cortex, project their axons to neurons in the deep cerebellar nuclei (DCN). BG, Bergmann glia; ML, molecular layer; PCL, Purkinje cell layer; GCL, granule cell layer. (c) A diagram depicting the HIV-derived lentiviral vector construct used in this study (Hanawa et al., 2002). To remove tat-dependent transcription, the U3 region of LTR is deleted. The Δgag (deleted gag sequence) is included to enhance the packaging efficiency. For improved viral-mediated gene transfer, the central polypurine tract (cPPT) and central termination sequence (CTS) were added in Δpol (deleted pol sequence). RRE; rev response element. The MSCV promoter was derived from the chimeric LTR sequence of murine stem cell virus (MSCV) (Hawley et al., 1994). GFP; green fluorescent protein.

prepared from 32 independent cultures for injection. Based on the viral titers, the vector preparations were categorized into 3 groups, 108–109 (n = 11 cultures, mean ± SEM; 6.3 ± 0.8 × 108), 109– 1010 (n = 10 cultures, 3.2 ± 0.6 × 109) and more than 1010 (n = 11 cultures, 1.9 ± 0.2 × 1010) transducing unit (TU)/ml. Six microliters of the solution containing the lentiviral vector was injected stereotaxically into the molecular layer at lobule 6 of the cerebellar cortex (Fig. 1a). Seven days after the injection, mice were sacrificed for assessing transduction. Fluorescence stereoscopic microscope analysis showed that GFP was efficiently and selectively expressed from lobule 4–5 to lobule 8 in the cerebellum (Fig. 2a). Examination of the sagittal sections at the cerebellar vermis showed that most of the GFP-expressing cells were localized in the molecular layer and Purkinje cell layer of the cerebellar cortex (Fig. 2b). The enlarged view of Fig. 2b shows that the majority of transduced cells were Purkinje cells (Fig. 2c), where GFP was detected in cell bodies through the dendritic arbors to the spines (Figs. 2d, e) as well as the axonal extensions (Fig. 2b arrows). GFP expression was also observed in Bergmann glial cells immunolabeled for GFAP (Fig. 2f) and parvalbumin-expressing

stellate and basket cells (Fig. 2g). GFP-positive cells were observed also in the granule cell layer although the number was much less, compared with the molecular and Purkinje cell layers. Some of those cells were identified as Golgi interneurons because they were immunolabeled for mGluR2, a marker of Golgi cells (Fig. 2h) (Geurts et al., 2001; Neki et al., 1996b). Surprisingly, the lentiviral vector failed to transduce granule cells: we examined 700 GFP-expressing cells in the granule cell layer of 7 mice which received injection of high-titer-viral vectors (2.1 ± 0.7 × 1010 TU/ml, n = 3 culture), but none of them were immunolabeled for NeuN, a granule cell marker (Figs. 2i, j). In contrast to the in vivo findings, the lentiviral vector efficiently transduced cultured granule cells (Fig. 2k). These results indicate that the HIV-derived lentiviral vector has no or extremely low infectious capability toward granule cells in vivo.

2.2. Transduction profile of HIV-derived lentiviral vector in the cerebellum We next examined the spread of the injected solution and the transduction profile of the viral vectors from different titer

14

BR A IN RE S EA RCH 1 0 82 ( 20 0 6 ) 1 1 –22

Fig. 2 – Selective expression of GFP in the cerebellum by HIV-derived lentiviral vectors. The lentiviral vector with a titer of 8.9 × 108 TU/ml was injected to lobule 6 of the mouse cerebellum. (a) A stereoscopic image of GFP fluorescence (upper) and the merged image with a whole brain (lower) 7 days after the injection. Numbers on the cerebellum represent the lobule number, and the arrow shows the injection site (lobule 6). (b) A stereoscopic photo of a sagittal section of the cerebellar vermis, in which the major GFP signal is observed in the molecular layer and Purkinje cell layer. The linear GFP signal in the medullary substance (arrows) is from Purkinje cell axons projecting to the deep cerebellar nuclei. (c–e) A confocal microscopic image shows that Purkinje cells are the major cell types transduced by HIV-derived lentiviral vectors (c). GFP was expressed in the cell bodies and dendritic arbors of Purkinje cells (d). An enlarged view of the boxed area in (d) shows the transfer of GFP to the dendritic spines (e). (f–h) In addition to Purkinje cells, Bergmann glias (f), stellate/basket cells (g), and Golgi cells (h) were also transduced. Bergmann glias (f) were immunolabeled for GFAP: a process of a Bergman glia is indicated by arrows. Stellate and basket cells (g) were labeled with anti-parvalbumin antibody (arrows). Golgi cells in the granule cell layer were immunostained for mGluR2 (h). (i, j) Failure of the transduction of in vivo granule cells by HIV vectors. Granule cells in the cerebellar section were immunostained for NeuN. No GFP-expressing cells were labeled for NeuN. (k) Transduction of granule cells in culture. In dissociated cerebellar neuronal culture, granule cells were efficiently transduced by HIV vectors (arrows).

categories by assessing the transduced areas and the proportion of transduced cell types in the cerebellar cortex. Fluorescence of Fluoro-ruby was observed almost throughout lobules 6–8 (Figs. 3a, b). The fluorescence was detected from lobule 3 to lobule 10, although the proportion of the Fluororuby-labeled area was less in the lobules distant from the injection site (Fig. 3d, n = 5 mice). Transduced areas, which were encompassed by the Fluoro-ruby-positive area, became larger as the titers of the injected virus solutions became higher: about 30% of the area in lobule 6 was GFP-positive

when viral vectors with titers of 6.3 ± 0.8 × 108 TU (n = 11) were injected (n = 20 mice), while more than 50% (n = 20 mice) and 80% (n = 20 mice) of the areas were GFP-positive when viral vectors with titers of 3.2 ± 0.6 × 109 (n = 10) and 1.9 ± 0.2 × 1010 (n = 11) TU/ml, respectively, were used (Figs. 3c, d). We then examined the number of transduced cells in lobule 6. When the viral titers were increased from the order of 108 (2.4 × 108 TU/ml) to the order of 109 (2.4 × 109 TU/ml), the number of GFP-positive cells was significantly increased (P < 0.01, Fig. 4a, inset graph). At over 109 TU/ml, however,

BR A IN RE S E A RCH 1 0 82 ( 20 0 6 ) 1 1 –2 2

15

types, including Purkinje cells, Bergmann glia, stellate/basket cells and Golgi cells, were not influenced by the viral titers used for injection (Fig. 4b).

2.3. No significant effects of the viral infection on motor coordination and the viability of Purkinje cells Purkinje cells are vulnerable to being damaged substantially by subtle insults (Sarna and Hawkes, 2003). To test whether the HIV vectors employed in this study affect the viability of infected neurons and eventually the motor coordination, we conducted the rotarod test and electrophysiological analysis using mice 7–9 days after the injection of high titer viruses (2.4 × 1010 TU/ml). The results of the rotarod test showed no significant alteration in the motor performance between virus-injected mice and non-injected littermates (Fig. 5a). In patch clamp analysis, we first examined the passive membrane properties of infected (GFP-expressing) Purkinje cells (Fig. 5b). Resting membrane potential, input resistance

Fig. 3 – Spread of the injected solution and viral-titer-dependent expansion of transduced areas. Fluoro-ruby alone or together with the virus vectors was injected into lobule 6 as described in Materials and methods. (a) Schematic drawing showing the numbering of cerebellar lobules in a sagittal section of the vermis. (b) A typical fluorescence image showing the spread of Fluoro-ruby 2 h after the injection. (c) Typical GFP fluorescence images showing the viral-titer-dependent transduction. The titer of the injected viral vectors is shown below each image. (d) Quantitative data summarizing the GFP-expressing area as well as the spread of the injection in the cerebellum. The percentage of the area labeled with Fluoro-ruby in each lobule was measured (n = 5 mice) and the mean (gray area) ± SEM (vertical bars) is indicated on the graph. The transduced area (%) in each lobule was assessed only in the molecular layer and Purkinje cell layer because more than 90% of the transduced cells were present in those layers, whereas a substantial part of the fluorescence in the medulla was from GFP in the Purkinje cell axons targeting to the deep cerebellar nuclei. Data were obtained from 3 different groups of mice (20 mice each) classified according to the injected viral titers: 108–109 (average ± SEM, 6.3 ± 0.8 × 108, n = 11 independent cultures), 108–109 (3.2 ± 0.6 × 109, n = 10) and more than 1010 (average ± SEM, 1.9 ± 0.2 × 1010, n = 11) TU/ml.

the transduction was almost saturated: there was no significant difference in the numbers of transduced cells between the viral vectors with 2.4 × 109 TU/ml (n = 4 mice) and 2.4 × 1010 TU/ml (n = 4 mice, Fig. 4a, inset graph). The increase in the number of transduced cells mainly reflects the increase in the number of transduced Purkinje cells and Bergmann glia (Fig. 4a). On the other hand, the ratios of GFP-positive cell

Fig. 4 – Number of transduced cells and their cell type classification in the cerebellum that received injection of viral vectors with different titers: 2.4 × 108 (n = 4 mice, gray columns), 2.4 × 109 (n = 4 mice, hatched columns), and 2.4 × 1010 (n = 4 mice, filled columns) TU/ml. The number of GFP-expressing cells was counted in areas close to the injection site. (a) The inset graph shows the total number of GFP-expressing cells counted in 5 different fields (0.95 mm2/ field) under a ×20 objective. Asterisks show significant differences (*P < 0.05, **P < 0.01). (b) The ratio (%) of GFP-expressing cell types was determined in each viral titer group, based on their morphology as well as immunostaining for cell-type-specific markers, GFAP, parvalbumin and mGluR2. There were no significant differences in the respective cell types among the 3 different titer groups. PC; Purkinje cell, BG; Bergmann cell, ST/BA; stellate cell and basket cell, UC; unidentified cell.

16

BR A IN RE S EA RCH 1 0 82 ( 20 0 6 ) 1 1 –22

and membrane conductance of Purkinje cells in transduced cells were similar to the values in neighboring control noninfected cells (Table 1). We then tested whether synaptic transmission was altered upon viral infection. PF-mediated and CF-mediated EPSCs in infected Purkinje cells were evoked similarly as in control Purkinje cells (Fig. 5c), and there were no significant differences in either PF- or CF-EPSC kinetics (10– 90% rise time and decay time constant, Table 1).

2.4. cortex

Absence of transduced cells outside of the cerebellar

Previously, injection of the AAV or FIV vectors into the cerebellar cortex was shown to transduce neurons in the deep cerebellar nuclei (Alisky et al., 2000). To verify the potential of HIV vectors, we examined the GFP expression in the deep cerebellar nuclei. The neuronal cell bodies in the deep cerebellar nuclei were labeled with NeuroTrace™ 530/615. Confocal microscopy revealed that GFP was localized through axons of Purkinje cells to their terminals (Figs. 6a, b) but not in postsynaptic cell bodies (Figs. 6c, d), suggesting that unlike AAV and FIV vectors (Alisky et al., 2000), the HIV-derived lentiviral vector, at least in our experimental conditions, does not transduce neurons in the deep cerebellar nuclei. The cerebellar cortex receives afferents from neurons in the extracerebellar nuclei, which include the pontine nuclei and inferior olivary complex (Figs. 1a, b). Those neurons could be transduced if the viral vector were taken up from the nerve terminals localized in the cerebellar cortex and transferred retrogradely to their soma. Though AAV and FIV vectors have not been examined yet, adenovirus vectors injected into the cerebellar cortex were shown to transduce neurons in the pontine nuclei, reticulotegmental nucleus of the pons, and inferior olivary complex (Terashima et al., 1997). Then we examined whether HIV-derived lentiviral vectors could transduce neurons in those nuclei. To label neurons in the brain stem nuclei, which are synaptically connected with neurons in the cerebellar cortex, viral vectors were injected together with Fluoro-ruby (Schmued et al., 1990), a neuronal tracer, into the cerebellar cortex. Fig. 7 shows the Fluoro-ruby/GFP (a, c, e, g) and cresyl violet-stained (b, d, f, h) images of the inferior olivary complex (a–d) and pontine nuclei/reticulotegmental nucleus Fig. 5 – Lack of significant influence on motor coordination, Purkinje cell viability and synaptic functions following infection of the HIV-derived vectors. Mice were examined 7–9 days after the viral injection. (a) Results of rotarod test. Rotation speed was 30 rpm/min. There were no significant differences between the virus-injected mice and the non-injected littermates. (b) A phase contrast (left) and the corresponding GFP fluorescence (right) images of a sagittal slice of the cerebellar vermis for patch clamp analysis. A stimulation electrode (Stim) was placed on the molecular layer to record PF-mediated synaptic transmission from a GFP-expressing Purkinje cell patch-clamped with a recording electrode (Rec). ML; molecular layer, PCL; Purkinje cell layer, GCL; granule cell layer. (c) Typical examples of PF–Purkinje cell (PF–PC) and CF–Purkinje cell (CF–PC) EPSCs from an infected (left) or a neighboring non-infected (right) Purkinje cell. PF–PC and CF–PC EPSCs were recorded at a holding potential of −70 mV and −10 mV, respectively.

Table 1 – Electrophysiological properties of infected and neighboring control non-infected Purkinje cells

Resting membrane potential (mV) Input resistance (MΩ) Membrane capacitance (pF) PF-PC 10%–90% rise time (ms) Decay time constant (ms) CF-PC 10%–90% rise time (ms) Decay time constant (ms)

Infected cells

Non-infected cells

−63.4 ± 2.4

−64.5 ± 2.3

95.4 473.2 2.7 25.2 0.67 14.0

104.9 476.7 2.8 23.1 0.63 15.0

± ± ± ± ± ±

7.4 21.6 0.3 2.8 0.03 0.9

± ± ± ± ± ±

11.3 29.5 0.2 1.9 0.02 1.1

Data (mean ± SEM) were obtained from 10 infected and 7 noninfected cells (3 mice). There were no significant differences (P > 0.2).

BR A IN RE S E A RCH 1 0 82 ( 20 0 6 ) 1 1 –2 2

17

Fig. 6 – Lack of transduced neurons in the deep cerebellar nuclei. Sagittal sections of the cerebellum 7 days after the virus injection were stained for Nissl substance with NeuroTrace™530/615 (red). The fluorescence image (a) represents the deep cerebellar nuclei, where GFP-positive Purkinje cell axons were projecting. The presented area corresponds to the boxed area in the inset of panel (a). The enlarged image (b) shows that axons of Purkinje cells terminate on the neurons of the deep cerebellar nuclei. Further enlarged images (c, d) confirm that GFP is localized in the axons of Purkinje cells and their presynaptic terminals, but not in postsynaptic neurons.

(e–h) 7 days after the injection. Cell bodies were strongly labeled with Fluoro-ruby in the inferior olivary complex (Figs. 7a, c), pontine nuclei and reticulotegmental nucleus (Figs. 7e, g), whereas no GFP expression was observed in those areas. Detailed analysis of the whole brain sections prepared 7 days and 14 days after the viral injection revealed that GFP expression was strictly limited to the cells in the cerebellar cortex and never detected in other regions in the CNS.

3.

Discussion

Adenovirus vectors are often used to deliver genes into neurons. For example, neurons in the cerebral cortex, hippocampus, and striatum are effectively transduced by adenoviral vectors (Ehrengruber et al., 2001; Pi et al., 2004; Zheng et al., 2005). In contrast, their potential to transduce Purkinje cells is very limited: transgene expression was restricted to Purkinje cells along the injection needle tract (Hashimoto et al., 1996; Terashima et al., 1997). Recently, vectors derived from HSV1 (Agudo et al., 2002), AAV (Alisky et al., 2000; Kaemmerer et al., 2000), and FIV (Alisky et al., 2000) were found to be much more effective than adenoviral vectors to transfer genes into Purkinje cells. In addition to those viral

vectors, HIV-derived lentiviral vectors have been increasingly used for gene transfer into neurons: however, there have been no reports of studies testing their potential to transduce Purkinje cells. The results of our study revealed a unique feature of HIV-derived lentiviral vectors when injected into the cerebellar cortex. We found that the viral vector injected into cerebellar cortex efficiently transduced Purkinje cells without significant influencing the viability and synaptic functions. We think that high tropism of the HIV-derived vectors for Purkinje cells is due to not only their large surface for picking up the virus but also to the high affinity of vesicular stomatitis virus G protein (VSV-G) for the Purkinje cell membrane because chemical modification of the VSV-G envelope of HIV vectors markedly reduced the transduction efficiency for Purkinje cells without affecting the tropism for HELA or HEK cells (TT and HH, unpublished data). In areas close to the injection site, the number of transduced cells was almost saturated when the viral titer was over 109 TU/ml, while efficiently transduced areas were significantly larger when virus solutions with titers over 1010 TU/ml were used. On the other hand, the proportion of Purkinje cells in all the transduced cells was not influenced when the viral titers were more than 108 TU/ml. Although the efficiency was not high, GFP expression was observed also in

18

BR A IN RE S EA RCH 1 0 82 ( 20 0 6 ) 1 1 –22

Fig. 7 – Lack of transduced neurons in the inferior olivary complex and pontine nuclei. Pairs of low-power (a, b, e, f) and high power (c, d, g, h) micrographs of Fluoro-ruby-positive neurons in the inferior olivary complex (IO) and pontine nuclei (Pn)/ reticulotegmental nucleus of pons (RtTg) as indicated. The left panels are Fluoro-ruby fluorescence images overlaid with GFP fluorescence. Nissl-stained images of the same positions are presented in the right panels. Inset photos in panels a and e are Nissl-stained sagittal sections, in which boxed areas are enlarged and presented in panels a, b and e, f, respectively. No GFP signal was observed in those brain stem nuclei.

stellate cells, basket cells, Bergmann glia, and Golgi cells, whereas no granule cells were transduced. Extracortical neurons in the deep cerebellar nuclei, pontine nuclei, reticulotegmental nucleus of the pons and inferior olivary complex, which receive projections from Purkinje cells or send axons to the cerebellar cortex, were not transduced. The cerebellar neuron types transduced by the HIV-derived vector were basically the same as those transduced by an FIVderived one (Alisky et al., 2000), but there were some

differences. One interesting finding was that, although the number was very limited, some granule cells were transduced in the previous study using FIV-derived vectors, but not by ours using HIV-derived vectors. Another discrepancy is the infection of Bergmann glia. In contrast to our results, in which a substantial number of Bergmann glia was transduced by HIV-derived vectors (Figs. 2 and 4), FIV-derived vectors exclusively transduced neurons in the cerebellum (Alisky et al., 2000). As the cellular tropism of viruses is basically

BR A IN RE S E A RCH 1 0 82 ( 20 0 6 ) 1 1 –2 2

determined by envelope glycoproteins, the discrepancy of the cellular tropism between HIV- and FIV-derived viruses may be due to the difference of VSV-G strains. Actually, viral vector pseudotyped with the Chandipura strain of the VSV-G envelope was shown to transduce a higher percentage of glial cells in the striatum compared to the Indiana strain VSVG-pseudotyped vector (Wong et al., 2004). Except in the cerebellar cortex, FIV-derived vectors transduced neurons in the deep cerebellar nuclei (Alisky et al., 2000). The authors injected cholera toxin subunit b together with the viral vector to define the limits to which the virus solution extended and found that the injected solution did not reach the deep cerebellar nuclei. Thus, the FIV-derived vectormediated transduction of neurons in the deep cerebellar nuclei does not seem to be mediated by the direct diffusion of the virus to the nuclei. However, FIV-derived vectors as well as HIV vectors cannot be transferred transsynaptically because they are replication incompetent, and therefore, secondary virus particles are never produced from infected cells. One possible explanation of the transduction of neurons in the deep cerebellar nuclei is that FIV-viruses, but not cholera toxin subunit b, reached the deep cerebellar nuclei by spillover and directly infected the neurons. Injection of adenoviral vectors into the cerebellar cortex was shown to cause transduction of precerebellar neurons, including those in the pontine nuclei, reticulotegmental nucleus of the pons and inferior olivary complex (Terashima et al., 1997). This was achieved by retrograde axonal transport through mossy fibers or climbing fibers. Whether vectors derived from lentivirus and AAV were transported retrogradely and transduced neurons in those nuclei had remained to be clarified. We examined that possibility here by assessing the GFP expression in those nuclei following the lentiviral injection to the cerebellar cortex. While neuronal cell bodies in both nuclei were clearly labeled by Fluoro-ruby that was injected together with viral vectors, no GFP signal was observed in those nuclei (Fig. 7). Although the mechanism is currently unclear, we presume that the viral vectors were not transported retrogradely to the soma of the brain stem nuclei because the present lentiviral vectors were pseudotyped with VSV-G, which is considered to bind to membrane phospholipids of all mammalian cells (Burns et al., 1993; Wong et al., 2004), and the murine embryonic stem cell virus (MSCV) promoter is generally active in most mammalian cells (Hawley et al., 1994). Further study will resolve this issue. To express a foreign gene in a specific subset of cells, a transgenic approach in combination with a cell-type-specific promoter is often used. Previously, using L7 promoter, which drives gene expression specifically in Purkinje cells (Hashimoto et al., 1996), we produced transgenic mice (Hirai et al., 2005a). However, as the activity of the L7 promoter is quite low, 6 out of 8 transgenic lines exhibited very low expression levels of the transgene (less than 2–3% of the native protein level, H.H., unpublished data). A similar result was reported by Ichise et al. (2000). Thus, the transgenic approach requires long duration and much housing space to obtain lines with suitable transgene expression levels. In the viral-vectormediated gene transfer, high levels of gene expression can be induced rapidly in Purkinje cells at any time: this allows us to express wild-type or various kinds of mutant proteins such

19

as dominant-negative forms or proteins lacking phosphorylation sites or signaling-molecule-binding domains in the in vivo cerebellum. In addition to use for research, the viral-vector-mediated gene delivery system can be applied to clinical gene therapy. Purkinje cells are particularly vulnerable and affected by a variety of insults (Sarna and Hawkes, 2003), in which the autosomal dominant spinocerebellar ataxias (SCAs) are potentially amenable to gene therapy. As recently many genes possessing therapeutic potential against the SCAs have been identified (Cummings et al., 1998; Kobayashi et al., 2002; Matsumoto et al., 2004; Miller et al., 2005), HIV vectormediated expression of those genes in Purkinje cells of patients could be a promising strategy against the SCAs. Our study has shown that when injected into the cerebellar cortex, HIV-derived lentiviral vectors transduce cells only in the cerebellar cortex, with highest tropism for Purkinje cells. Thus, HIV-derived lentiviral vectors are useful tools for selective gene delivery into Purkinje cells, and application of these vectors to basic and clinical studies will facilitate the elucidation of the molecular mechanisms regulating cerebellar functions as well as the development of efficient and safe gene therapy protocols against diseases that affect Purkinje cells.

4.

Experimental procedures

All procedures for the care and treatment of animals were carried out according to NIH guidelines, and the experimental protocol was approved by the institutional animal care committee of Kanazawa University.

4.1.

Virus preparation

VSV-G pseudotyped lentiviral vectors provided by St. Jude Children's Research Hospital (Hanawa et al., 2002) was used in this study. The vectors were designed to express GFP under the control of the MSCV promoter (Hawley et al., 1994) (Fig. 1c). The virus vector was produced by cotransfection of human embryonic kidney (HEK) 293 T cells with a mixture of four plasmids using a calcium phosphate precipitation method. Briefly, cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 50 U/ml penicillin G and 50 μg/ml streptomycin (pH 7.35) at 37 °C in a 5% CO2 atmosphere. Cells were plated at 2–4 × 106 cells in a 10-cm culture dish 24 h before transfection. The four plasmid mixture consisted of 6 μg of pCAGkGP1R, 2 μg of pCAG4RTR2, 2 μg of pCAG-VSV-G, and 10 μg of vector plasmid pCL20c MSCV-GFP. The plasmids were mixed and diluted to a total volume of 450 μl with double-distilled water, after which 50 μl of 2.5 M CaCl2 was added and mixed well. Then 500 μl of 2× HEPES-buffered saline (280 mM NaCl, 50 mM HEPES, 1.5 mM Na2HPO4 [pH 7.05]) was added dropwise while vortexing, and the precipitate was immediately added to the cultures. Sixteen hours after transfection, the cells were washed with phosphate-buffered saline (PBS) twice and then cultured for an additional 24 h. The medium containing vector particles was harvested 40 h

20

BR A IN RE S EA RCH 1 0 82 ( 20 0 6 ) 1 1 –22

after transfection. The medium samples were filtered through 0.22-μm membranes and centrifuged at 25,000 rpm for 90 min. The virus samples were finally suspended in 45 μl of PBS (pH 7.4), frozen in aliquots, and stored at −80 °C. The titers of virus stocks were measured by transducing HeLa cells as previously described (Hanawa et al., 2002). Serial dilutions of the vector preparations were added to 105 HeLa cells growing in a monolayer in six-well plates in a total volume of 2 ml of DMEM containing Polybrene (6 μg/ml). After 16 h, the medium was replaced by fresh medium, and culturing was continued for another 3 days. Titers were routinely determined using a volume of vector preparation that yielded linear, dose-dependent transduction of target cells with a level not in excess of 20%.

4.2.

Cerebellar injection

C57Bl/6 mice between 6 and 12 weeks of age were used. Each mouse was injected intraperitoneally with sodium pentobarbital (40 mg/kg body weight) to produce deep anesthesia. The mouse was mounted in a stereotactic frame, and its head was shaved. A midline sagittal incision was made, and the cranium over the cerebellar vermis was exposed. A burr hole was drilled 5–6 mm caudal from the bregma. A tip of a Hamilton syringe attached to a micropump (UltramicroPump II; World Precision Instruments (WPI), Sarasota, FL, USA) was placed in the molecular layer of the cerebellar vermis (lobule 6), and 6 μl of the virus solution was injected at a rate of 200 nl/ min using a microprocessor-based controller (Micro4; WPI). This volume and speed were used because, in addition to soaking into the tissue, the virus solution spread over different lobules through the subarachnoidal space, leading to the transduction in larger areas. The syringe was left in place for an additional 2 min before it was withdrawn. Then the scalp was sutured, the mouse was kept on a heating pad until it had recovered from the anesthesia, and then it was returned to a standard cage.

4.3.

Injection of Fluoro-ruby

Fluoro-ruby (Invitrogen, Carlsbad, CA, USA) solution (10% in PBS) alone or together with a suspension of the HIV-derived lentiviral vectors were injected into lobule 6 of the cerebellar vermis. Mice were sacrificed 2 h after the injection to define the spread, while transduced areas and cell types were examined 7days after the injection.

4.4.

Cerebellar neuronal culture

Dissociated neuronal cultures from postnatal day 0 mice pups were prepared as described previously (Hirai and Launey, 2000).

4.5. Immunohistochemistry and acquisition of brain images For assessing the spread of the injection and transduced areas in the cerebellum, mice were anesthetized and perfused transcardially with 4% paraformaldehyde plus 2% picric acid in 0.1 M sodium phosphate buffer (pH 7.4). The cerebellum was

removed and postfixed overnight at 4 °C followed by cryoprotection in 30% sucrose. The fluorescence images of whole brains were obtained using a cooled CCD-camera (Keyence VB-7000, Osaka, Japan) attached to a fluorescence stereoscopic microscope (Keyence VB-G05). To determine the areas of GFP expression, transduced cell types and the number of each cell type, the cerebellum was cut into 100-μm sagittal sections using a microslicer (DOSAKA DTK-1000, Kyoto, Japan), and the sections were subjected to immunostaining with mouse monoclonal anti-parvalbumin (P-3088, SigmaAldrich, St. Louis, MO, USA), mouse monoclonal anti-glial fibrillary acidic protein (GFAP; MAB360, Chemicon, Temecula, CA, USA), mouse monoclonal anti-neuron-specific nuclear protein (NeuN; MAB377, Chemicon) or mouse monoclonal anti-mGluR2 (Gift from Prof. R. Shigemoto, National Institute for Physiological Sciences, Japan) (Neki et al., 1996a) antibodies. The primary antibodies for parvalbumin, GFAP, NeuN, and mGluR2 were used at a concentration of 1:2000, 1:400, 1:200, and 1:500, respectively. Sections were incubated overnight at room temperature, rinsed twice in PBS and then incubated with Alexa fluor 568-conjugated secondary antibody (10 μg/ml; Invitrogen) for 3 h at room temperature. Immunofluorescence was evaluated using a confocal laserscanning microscope (LSM 5 PASCAL; Zeiss, Oberkochen, Germany). To verify the anterograde transport of viral vectors to the deep cerebellar nuclei or precerebellar nuclei, brains were cut sagittally on a freezing microtome at a thickness of 30 μm. The sections were counterstained with cresyl violet or a fluorescent probe which specifically binds to Nissl substance (NeuroTrace™530/615; Invitrogen). For cresyl violet staining, sections were stained with 0.2% cresyl violet, dehydrated through a graded series of ethanol (70%, 95%, 100%, 100%), placed in xylene, and coverslipped using DPX mountant.

4.6. Quantitative analysis of transduced areas and cell counts Measurements were done on the sagittal sections of the cerebellar vermis. Because more than 90% of transduced cells were localized in the molecular layer and Purkinje cell layer, the percent ratio of transduction in the molecular layer and Purkinje cell layer was assessed in each lobule under a ×5 objective. For cell counting, several areas (0.95 mm2/area) in lobule 6 were randomly selected under a ×20 objective. Serial confocal-scanning images of GFP fluorescence and Alexa-fluor 568 immunofluorescence every 2-μm optical sectioning from the bottom to the top of the slice were obtained and reconstructed three-dimensionally, and the numbers of transduced Purkinje cells, Bergmann glial cells, basket/stellate cells, Golgi cells, and granule cells in the slice were estimated. Cells that expressed GFP, but were not stained with any of the antibodies for parvalbumin, GFAP, mGluR2, or NeuN, were classified as “unknown”. Then the ratios of respective transduced cell types were determined.

4.7.

Rotarod

Mice were tested 7 days after the viral injection. Their littermates that did not receive viral injection were used as

BR A IN RE S E A RCH 1 0 82 ( 20 0 6 ) 1 1 –2 2

controls. The Rota-Rod Treadmill (Muromachi kikai, Tokyo, Japan) consists of a gritted plastic rod (3 cm in diameter, 10 cm long) flanked by two large round plates (50 cm in diameter). The time the mouse remained on the rod (rotation speed, 30 rpm). was measured. A maximum of 120 s was allowed to test each animal.

4.8.

Electrophysiology

Sagittal slices (250 μm thick) of cerebellum were prepared, and whole-cell patch-clamp recordings were made from visually identified Purkinje cells at room temperature (26 °C) using an upright microscope (Zeiss Axioskop) as described previously (Hirai et al., 2005b). The slices were continuously superfused with an extracellular solution containing 124 mM NaCl, 2.5 mM KCl, 1.25 mM NaH2PO4, 1.5 mM MgSO4, 2 mM CaCl2, 26 mM NaHCO3, 20 mM glucose, and 0.1 mM picrotoxin, bubbled with 95% O2 and 5% CO2. Patch pipettes had a resistance of 3 to 4 MΩ in the intracellular solution containing 60 mM CsCl, 10 mM Cs D-gluconate, 20 mM TEA-Cl, 20 mM BAPTA, 4 mM MgCl2, 4 mM ATP, 0.4 GTP, and 30 HEPES (pH 7.3, adjusted with CsOH). Stimulation pipettes (5–10 μm tip diameter) were filled with the standard saline and used to apply square pulses for focal stimulation (duration, 10 μs; amplitude, 20–220 μA). Purkinje cells were clamped at −80 mV for recording PF excitatory postsynaptic currents (EPSCs) and at −10 mV for recording CF-EPSCs. PFs and CFs were stimulated in the molecular layer and the granule cell layer, respectively, 50–100 μm away from the Purkinje cell soma. An EPC-7 amplifier (HEKA Instruments, Lambrecht, Germany) and pCLAMP8 software (Molecular Devices, Sunnyvale, CA) were used for recording and data analysis. Signals were filtered at 2 kHz and digitized at 4 kHz (Digidata 1320A, Molecular Devices).

Acknowledgments We thank Dr. N. Yamada, M. Ohbayashi, S. Yanagisawa, A. Higashi, and C. Koyama for the technical assistance, and Prof. R. Shigemoto for the anti-mGluR2 antibody. The lentiviral vector was kindly provided by St. Jude Children's Research Hospital and George Washington University. This work was supported in part by the Mochida Memorial Foundation for Medical and Pharmaceutical Research, the Naito Memorial Foundation, the Mitani Research Foundation, the Brain Science Foundation, Japanese Grants-in-aid for Scientific Research provided by the Japan Society for the Promotion of Science and the 21st Century Center of Excellence Program of the Ministry of Education, Culture, Sports, Science and Technology of Japan (H. H.).

REFERENCES

Agudo, M., Trejo, J.L., Lim, F., Avila, J., Torres-Aleman, I., Diaz-Nido, J., Wandosell, F., 2002. Highly efficient and specific gene transfer to Purkinje cells in vivo using a herpes simplex virus I amplicon. Hum. Gene Ther. 13, 665–674. Alisky, J.M., Hughes, S.M., Sauter, S.L., Jolly, D., Dubensky Jr., T.W.,

21

Staber, P.D., Chiorini, J.A., Davidson, B.L., 2000. Transduction of murine cerebellar neurons with recombinant FIV and AAV5 vectors. NeuroReport 11, 2669–2673. Burns, J.C., Friedmann, T., Driever, W., Burrascano, M., Yee, J.K., 1993. Vesicular stomatitis virus G glycoprotein pseudotyped retroviral vectors: concentration to very high titer and efficient gene transfer into mammalian and nonmammalian cells. Proc. Natl. Acad. Sci. U. S. A. 90, 8033–8037. Cummings, C.J., Mancini, M.A., Antalffy, B., DeFranco, D.B., Orr, H.T., Zoghbi, H.Y., 1998. Chaperone suppression of aggregation and altered subcellular proteasome localization imply protein misfolding in SCA1. Nat. Genet. 19, 148–154. de Almeida, L.P., Zala, D., Aebischer, P., Deglon, N., 2001. Neuroprotective effect of a CNTF-expressing lentiviral vector in the quinolinic acid rat model of Huntington's disease. Neurobiol. Dis. 8, 433–446. Desmaris, N., Bosch, A., Salaun, C., Petit, C., Prevost, M.C., Tordo, N., Perrin, P., Schwartz, O., de Rocquigny, H., Heard, J.M., 2001. Production and neurotropism of lentivirus vectors pseudotyped with lyssavirus envelope glycoproteins. Mol. Ther. 4, 149–156. Dittgen, T., Nimmerjahn, A., Komai, S., Licznerski, P., Waters, J., Margrie, T.W., Helmchen, F., Denk, W., Brecht, M., Osten, P., 2004. Lentivirus-based genetic manipulations of cortical neurons and their optical and electrophysiological monitoring in vivo. Proc. Natl. Acad. Sci. U. S. A. 101, 18206–18211. Ehrengruber, M.U., Hennou, S., Bueler, H., Naim, H.Y., Deglon, N., Lundstrom, K., 2001. Gene transfer into neurons from hippocampal slices: comparison of recombinant Semliki Forest Virus, adenovirus, adeno-associated virus, lentivirus, and measles virus. Mol. Cell. Neurosci. 17, 855–871. Geurts, F.J., Timmermans, J., Shigemoto, R., De Schutter, E., 2001. Morphological and neurochemical differentiation of large granular layer interneurons in the adult rat cerebellum. Neuroscience 104, 499–512. Hanawa, H., Kelly, P.F., Nathwani, A.C., Persons, D.A., Vandergriff, J.A., Hargrove, P., Vanin, E.F., Nienhuis, A.W., 2002. Comparison of various envelope proteins for their ability to pseudotype lentiviral vectors and transduce primitive hematopoietic cells from human blood. Mol. Ther. 5, 242–251. Hashimoto, M., Aruga, J., Hosoya, Y., Kanegae, Y., Saito, I., Mikoshiba, K., 1996. A neural cell-type-specific expression system using recombinant adenovirus vectors. Hum. Gene Ther. 7, 149–158. Hawley, R.G., Lieu, F.H., Fong, A.Z., Hawley, T.S., 1994. Versatile retroviral vectors for potential use in gene therapy. Gene Ther. 1, 136–138. Hirai, H., Launey, T., 2000. The regulatory connection between the activity of granule cell NMDA receptors and dendritic differentiation of cerebellar Purkinje cells. J. Neurosci. 20, 5217–5224. Hirai, H., Miyazaki, T., Kakegawa, W., Matsuda, S., Mishina, M., Watanabe, M., Yuzaki, M., 2005a. Rescue of abnormal phenotypes of the delta2 glutamate receptor-null mice by mutant delta2 transgenes. EMBO Rep. 6, 90–95. Hirai, H., Pang, Z., Bao, D., Miyazaki, T., Li, L., Miura, E., Parris, J., Rong, Y., Watanabe, M., Yuzaki, M., Morgan, J.I., 2005b. Cbln1 is essential for synaptic integrity and plasticity in the cerebellum. Nat. Neurosci. 8, 1534–1541. Ichise, T., Kano, M., Hashimoto, K., Yanagihara, D., Nakao, K., Shigemoto, R., Katsuki, M., Aiba, A., 2000. mGluR1 in cerebellar Purkinje cells essential for long-term depression, synapse elimination, and motor coordination. Science 288, 1832–1835. Ito, M., 2002. Historical review of the significance of the cerebellum and the role of Purkinje cells in motor learning. Ann. N. Y. Acad. Sci. 978, 273–288. Kaemmerer, W.F., Reddy, R.G., Warlick, C.A., Hartung, S.D., McIvor, R.S., Low, W.C., 2000. In vivo transduction of cerebellar Purkinje

22

BR A IN RE S EA RCH 1 0 82 ( 20 0 6 ) 1 1 –22

cells using adeno-associated virus vectors. Mol. Ther. 2, 446–457. Kobayashi, T., Tanaka, K., Inoue, K., Kakizuka, A., 2002. Functional ATPase activity of p97/valosin-containing protein (VCP) is required for the quality control of endoplasmic reticulum in neuronally differentiated mammalian PC12 cells. J. Biol. Chem. 277, 47358–47365. Matsumoto, M., Yada, M., Hatakeyama, S., Ishimoto, H., Tanimura, T., Tsuji, S., Kakizuka, A., Kitagawa, M., Nakayama, K.I., 2004. Molecular clearance of ataxin-3 is regulated by a mammalian E4. EMBO J. 23, 659–669. Miller, V.M., Nelson, R.F., Gouvion, C.M., Williams, A., RodriguezLebron, E., Harper, S.Q., Davidson, B.L., Rebagliati, M.R., Paulson, H.L., 2005. CHIP suppresses polyglutamine aggregation and toxicity in vitro and in vivo. J. Neurosci. 25, 9152–9161. Naldini, L., Blomer, U., Gage, F.H., Trono, D., Verma, I.M., 1996. Efficient transfer, integration, and sustained long-term expression of the transgene in adult rat brains injected with a lentiviral vector. Proc. Natl. Acad. Sci. U. S. A. 93, 11382–11388. Neki, A., Ohishi, H., Kaneko, T., Shigemoto, R., Nakanishi, S., Mizuno, N., 1996a. Metabotropic glutamate receptors mGluR2 and mGluR5 are expressed in two non-overlapping populations of Golgi cells in the rat cerebellum. Neuroscience 75, 815–826. Neki, A., Ohishi, H., Kaneko, T., Shigemoto, R., Nakanishi, S., Mizuno, N., 1996b. Pre- and postsynaptic localization of a

metabotropic glutamate receptor, mGluR2, in the rat brain: an immunohistochemical study with a monoclonal antibody. Neurosci. Lett. 202, 197–200. Pi, R., Yin, W., Zheng, S., Qiu, P., Zhou, J., Guo, W., Su, T., Yan, G., 2004. Adenoviral mediated transfer of TIMP-3 partially prevents glutamate-induced cell death in primary cultured cortical neurons of the rat. Brain Res. Mol. Brain Res. 127, 136–139. Sarna, J.R., Hawkes, R., 2003. Patterned Purkinje cell death in the cerebellum. Prog. Neurobiol. 70, 473–507. Schmued, L., Kyriakidis, K., Heimer, L., 1990. In vivo anterograde and retrograde axonal transport of the fluorescent rhodamine-dextran-amine, Fluoro-Ruby, within the CNS. Brain Res. 526, 127–134. Terashima, T., Miwa, A., Kanegae, Y., Saito, I., Okado, H., 1997. Retrograde and anterograde labeling of cerebellar afferent projection by the injection of recombinant adenoviral vectors into the mouse cerebellar cortex. Anat. Embryol. (Berl.) 196, 363–382. Wong, L.F., Azzouz, M., Walmsley, L.E., Askham, Z., Wilkes, F.J., Mitrophanous, K.A., Kingsman, S.M., Mazarakis, N.D., 2004. Transduction patterns of pseudotyped lentiviral vectors in the nervous system. Mol. Ther. 9, 101–111. Zheng, J.S., Tang, L.L., Zheng, S.S., Zhan, R.Y., Zhou, Y.Q., Goudreau, J., Kaufman, D., Chen, A.F., 2005. Delayed Gene Therapy of glial cell line-derived neurotrophic factor is efficacious in a rat model of Parkinson's disease. Brain Res. Mol. Brain Res. 134, 155–161.