Neuronal expression of the transcription factor Gli1 using the Talpha1 alpha-tubulin promoter is neuroprotective in an experimental model of Parkinson's disease

Abstract

Nigrostriatal neurons degenerate during Parkinson's disease. Experimentally, neurotoxins such as 6-hydroxydopamine (6-OHDA) in rodents, and MPTP in mice and non-human primates, are used to model the disease-induced degeneration of midbrain dopaminergic neurons. Glial-cell-derived neurotrophic factor (GDNF) is a very powerful neuroprotector of dopaminergic neurons in all species examined. However, recent reports have indicated the possibility that GDNF may, in the long term and if expressed in an unregulated manner, exert untoward effects on midbrain dopaminergic neuronal structure and function. Although GDNF remains a powerful neurotrophin, the search for alternative therapies based on alternative and complementary mechanisms of action to GDNF is warranted. Recently, recombinant adenovirus-derived vectors encoding the differentiation factor Sonic Hedgehog (Shh) and its downstream transcriptional activator (Gli1) were shown to protect dopaminergic neurons in the substantia nigra pars compacta from 6-OHDA-induced neurotoxicity in rats in vivo. A pancellular human CMV (hCMV) promoter was used to drive the expression of both Shh and Gli1. Since Gli1 is a transcription factor and therefore exerts its actions intracellularly, we decided to test whether expression of Gli1 within neurons would be effective for neuroprotection. We demonstrate that neuronal-specific expression of Gli1 using the neuron-specific Talpha1 alpha-tubulin (Talpha1) promoter was neuroprotective, and its efficiency was comparable to the pancellular strong viral hCMV promoter. These results suggest that expression of the transcription factor Gli1 solely within neurons is neuroprotective for dopaminergic neurons in vivo and, furthermore, that neuronal-specific promoters are effective within the context of adenovirus-mediated gene therapy-induced neuroprotection of dopaminergic midbrain neurons. Since cell-type specific promoters are known to be weaker than the viral hCMV promoter, our data demonstrate that neuronal-specific expression of transcription factors is an effective, specific, and sufficient targeted approach for neurological gene therapy applications, potentially minimizing side effects due to unrestricted promiscuous gene expression within target tissues.

Figure 1

Construction and molecular characterization of the novel RAd vectors, RAd-Tα1-lacZ, and RAd-Tα1-Gli1. RAds were generated by homologous recombination after cotransfection into 293 cells of the shuttle plasmid encoding the promoter transgene cassette, and pBHG10, containing all necessary sequences of the adenoviral genome, apart from the E1 and E3 deletions, needed to construct a recombinant vector. The diagram to show the genetic structure of corresponding plasmids, the results of the restriction analysis, and Southern blot, used for the construction of RAd-Tα1-LacZ are all shown in (a). For the restriction analysis and Southern blot, lanes are as follows: lane 1, molecular size markers from a HindIII digest of λ DNA; lane 2, RAd-Tα1-LacZ; lane 3, shuttle plasmid pΔE1-Tα1-LacZ; and lane 4, pBHG10, all digested with BamHI. Next to the restriction analysis is the corresponding Southern blot using a specific probe for the Tα1 promoter sequence, as illustrated in the schematic drawing. The diagram to show the genetic structure of corresponding plasmids, the results of the restriction analysis, and Southern blot, used for the construction of RAd-Tα1-Gli1 are all shown in (b). For the restriction analysis and Southern blot, lanes are as follows: lane 1, molecular size DNA markers; lane 2, shuttle plasmid pΔE1-Tα1-Gli1; lane 3, RAd-Tα1-Gli1, all digested with EcoRI. Next to it is the corresponding Southern blot showing the results utilizing a probe recognizing sequences within transgene Gli1, as illustrated.

Figure 2

Expression of nuclear β-galactosidase in the neocortex, striatum, and SNpc following the injection of RAd-Tα1-lacZ into the striatum. RAd-Tα1-lacZ(ns) was injected into the striatum, and 7 days later, animals were perfused fixed and immunoreacted with antibodies recognizing β-galactosidase protein. Labeled nuclei were detected both in the striatum (A), cingular/somatosensory cortex (B), periolfactory cortex (C, D), and the SNpc. Thus, nuclear β-galactosidase immunoreactivity was detected within the striatum, and within cells projecting to the striatum. To determine the identity of transduced cells, tissues were double immunoreacted for β-galactosidase and cell-type specific markers, as shown in Figure 3.

Figure 3

Identification of cells expressing nuclear β-galactosidase as neurons in the neocortex, striatum, and SNpc following the injection of RAd-Tα1-lacZ(ns) into the striatum. TH labels dopaminergic neurons and NeuN reacts with most neuronal cell types throughout the brain. (a) and (d) show nuclear β-galactosidase, visualized with Texas red, and (b, e and h) show neurons marked with NeuN and TH, respectively, each detected with secondary antibodies labeled with Fluorescein. Note the different labeling pattern of TH – it labels the cytoplasm and neurites – and NeuN, staining mainly the nuclei. (c) is the overlay of (a and b); (f) is the overlay of (d and e); (i) is the overlay of (g and h). Scale bar in (g) is 100 μm. The number of double labeled cells in each case was >95% in each of the tissues. The <5% of cells not expressing the neuronal marker were not identified further.

Figure 4

Labeling of nigrostriatal neurons with FG and the stereological quantitative unbiased stereological estimation of their numbers. Retrograde labeling of nigrostriatal dopaminergic neurons with FG, their lesioning with intrastriatal 6-OHDA, and their double immunoreaction with antibodies to TH is shown in (A). (B) shows the results of the quantitative unbiased stereological analysis, demonstrating that the number of double-labeled neurons for FG and TH is equally revealed if all sections are counted, or if only every sixth one (*) is analyzed.

Figure 5

Neuroprotection of SNpc cell bodies following the injection of RAd vectors expressing GDNF or Gli1 into the striatum. All four treatment groups are represented: the left panels, (a, c, e, and g) show the ipsilateral side exposed to RAd at a dose of 6 × 10⁷ IU and subsequent 6-OHDA-induced neurodegneration. In this paradigm, as FG is administered preceding 6-OHDA-induced neurodegneration, microglial cells on the ipsilateral side are labeled with FG, possibly by phagocytosing degenerated neurons. Microglial cells made visible by the engulfed FG can be distinguished morphologically from the polygonally shaped neurons and their long neurites by a smaller cell body and fine reticular processes. The small microglial cells never expressed TH immunoreactivity. The right panels (b, d, f, and h) show the contralateral untreated control side, only exposed to FG and vehicle. On this side exclusively neurons are labeled with FG. Scale bar in (h) is 100 μm and applies to images (a–h). (i) shows the quantitation of the survival of nigral dopaminergic neurons containing FG.

Figure 6

Effects of RAd-GDNF (a), RAd-Gli1 (hCMV-Gli1) (b), or RAd-Tα1-Gli1 (c) on the density of striatal TH-immunoreactive fibers in the striatum. Injection site, area of denervation, and the boundary with normal striatum are illustrated. The asterisks overlies the injection site, the red broken line indicates the boundary between the denervation caused by 6-OHDA (to the left of the red line), and the remaining unaffected striatum (to the right). Only in animals injected with RAd-GDNF did we detect a halo of immunoreactive fibers (to the left of the broken blue line; shown in more detail in Figure 7). No such halo was detected in any of the animals injected with the vectors expressing Gli1. The scale bar in (c, 200 μm) applies to all images.

Figure 7

Neuropathology of TH-immunoreactive fibers following the injection of RAd vectors expressing GDNF into the striatum of animals subsequently injected with the neurotoxin 6-OHDA. (a) shows a larger representation of the area surrounding the needle tract in the ipsilateral striatum of an animal treated with 6 × 10⁷ IU RAd-GDNF and subsequent 6-OHDA. TH was detected immunohistochemically. Five distinctive zones surrounding the needle tract can be distinguished microscopically. 1 is the site of injection filled with macrophages, 2 corresponds with an area of necrosis, 3 shows neurites protected from 6-OHDA neurotoxicity, 4 demonstrates the effect of 6-OHDA toxicity on DA neurons 28 days after its injection and 5 represents an area unaffected by administration of all previously mentioned substances. (a–f) show examples of the axonmorphology in the different zones. Of particular note are swollen dystrophic axon terminals only found in zone 3, see image (c), and in the transition between zone 4 and 5, see image (e). Scale bar in (f) applies to images (b)–(f) and measures 20 μm.