Glial promoter selectivity following AAV-delivery to the immature brain

Abstract

Recombinant adeno-associated virus (AAV) vectors are versatile tools for gene transfer to the central nervous system (CNS) and proof-of-concept studies in adult rodents have shown that the use of cell type-specific promoters is sufficient to target AAV-mediated transgene expression to glia. However, neurological disorders caused by glial pathology usually have an early onset. Therefore, modelling and treatment of these conditions require expanding the concept of targeted glial transgene expression by promoter selectivity for gene delivery to the immature CNS. Here, we have investigated the AAV-mediated green fluorescent protein (GFP) expression driven by the myelin basic protein (MBP) or glial fibrillary acidic protein (GFAP) promoters in the developing mouse brain. Generally, the extent of transgene expression after infusion at immature stages was widespread and higher than in adults. The GFAP promoter-driven GFP expression was found to be highly specific for astrocytes following vector infusion to the brain of neonates and adults. In contrast, the selectivity of the MBP promoter for oligodendrocytes was poor following neonatal AAV delivery, but excellent after vector injection at postnatal day 10. To extend these findings obtained in naïve mice to a disease model, we performed P10 infusions of AAV-MBP-GFP in aspartoacylase (ASPA)-deficient mouse mutants presenting with early onset oligodendrocyte pathology. Spread of GFP expression and selectivity for oligodendrocytes in ASPA-mutants was comparable with our observations in normal animals. Our data suggest that direct AAV infusion to the developing postnatal brain, utilising cellular promoters, results in targeted and long-term transgene expression in glia. This approach will be relevant for disease modelling and gene therapy for the treatment of glial pathology.

Figure 1. Promoter selectivity targets transgene expression to specific neural cell types in vitro and in the adult brain.

AAV vectors (1×10⁹ vg) were used to express GFP driven by the indicated promoters in enriched primary oligodendrocyte cultures (A–C). For in vivo studies vectors (2×10⁹ vg) were injected in the striatum of adult mice (D–F). Representative images of results by double-immunocytochemistry for GFP (green) and cell-type specific markers (red) illustrate promoter selectivity. A, In primary cultures AAV-CBA-GFP expressed in all NeuN⁺ neurons. In addition, NeuN-negative astrocytes (top in picture) showed GFP immunoreactivity. B, AAV-MBP-GFP-mediated GFP-expression was restricted to ASPA⁺ oligodendrocytes in vitro. C, AAV-GFAP-GFP transduction resulted in GFP immunoreactivity limited to cultured GFAP⁺ astrocytes. D, CBA promoter-controlled GFP expression was highly specific to neurons in vivo. E, The MBP promoter was selective for forebrain oligodendrocytes. F, The GFAP promoter drove GFP specifically in astrocytes. Representative results from three independent experiments are shown. Bars: A–C, 50 µm; D–F, 100 µm.

Figure 2. Quantification of glial promoter selectivity and vector spread after AAV delivery to adult mice.

Three weeks prior to analysis 2×10⁹ vg AAV-MBP-GFP (A–E) or AAV-GFAP-GFP (F–J) were injected into the dorsal striatum of adult mice (n = 3). Higher magnification views of the striatum following detection of GFP with either ASPA (A,F), NeuN (B,G) or ALDH1L1 (C,H) was performed to identify the GFP⁺ cell-types. Quantification of results after AAV-MBP-GFP injection showed the vast majority of transgene-expressing cells were ASPA⁺ oligodendrocytes, followed by a smaller fraction representing NeuN⁺ neurons (D). In contrast, AAV-GFAP-GFP-mediated transgene expression was strictly astrocytic (I). The vector spread, determined by monitoring transgene expression in the rostro-caudal extension, was comparable for both vectors (E,J). Arrows in A,B,H indicate co-labelling. Bars: 50 µm.

Figure 3. Astrocytic transgene expression after neonatal delivery of AAV-GFAP-GFP.

AAV (2×10⁹ vg) was administered to the striatum of newborn mice. Brains (n = 3) were analyzed three weeks later for GFP expression (green) in combination with cell-type specific markers (red). Immunoreactivities of GFP with ASPA (A), or NeuN (B) segregated. C, Co-staining with ALDH1L1 identified GFP⁺ cells as astrocytes (arrows). D, Quantitative comparison of neural populations expressing GFP shows selectivity of the GFAP promoter in astrocytes. E, Percentage of GFP⁺ cells among individual neural populations in the target region showing transgene expression in 50% of astrocytes. In contrast, only negligible numbers of oligodendrocytes or neurons expressed the transgene. Bar: 50 µm.

Figure 4. Spread of GFP expression after AAV-GFAP-GFP delivery to the neonatal striatum.

The vector (2×10⁹ vg) was delivered unilaterally to the striatum (n = 3). Three weeks later brains were processed into 40 µm sections, immunostained for GFP and every 8^th section was used to determine the area showing GFP immunoreactivity. The graph shows quantitative results after plotting the area covered by GFP immunoreactivity as a function of the distance from the injection site. An arrowhead labels the approximate injection site. Bar: 1 mm.

Figure 5. Quantification of MBP promoter selectivity after intrastriatal AAV delivery to neonates or P10 animals.

AAV-MBP-GFP (2×10⁹ vg) was delivered to the striatum of neonates (A–E) or P10 animals (F–J). Brains (n = 3) were analyzed three weeks later for GFP expression (green) in combination with cell-type specific markers (red). After neonatal delivery transgene expression was observed to various degrees in oligodendrocytes (A), neurons (B), and astrocytes (C). Quantitative comparison of neural populations expressing GFP shows strongest activity of the MBP promoter in astrocytes (D). Percentage of GFP⁺ cells among individual neural populations in the target region showing transgene expression in 30% of astrocytes. Only few oligodendrocytes or neurons expressed the transgene (E). After P10 delivery transgene expression was predominantly observed in oligodendrocytes (F), but not in neurons (G), or astrocytes (H). Quantitative comparison of neural populations expressing GFP shows high specificity of the MBP promoter for oligodendrocytes (I). Percentage of GFP⁺ cells among individual neural populations in the target region showing transgene expression in the majority of oligodendrocytes. Other cell-types showed negligible percentage of GFP⁺ cells (J). Arrows in A–C and F indicate co-labelling. Bars: 50 µm.

Figure 6. Spread of GFP expression after AAV-MBP-GFP delivery to the neonatal or P10 striatum.

Vectors (2×10⁹ vg) were delivered unilaterally to the striatum (n = 3) at P0 (left) or P10 (right). Three weeks later brain sections were immunostained for GFP and every 8^th section was used to determine the area showing GFP immunoreactivity. The graph shows quantitative results after plotting the area covered by GFP immunoreactivity as a function of the distance from the injection site. An arrowhead labels the approximate injection site. The spread of transgene expression is comparable after neonatal or P10 delivery. Bars: 1 mm.

Figure 7. Promoter specificity and volume of GFP-expression after AAV injection at different postnatal stages.

A, For each promoter and time point used, the proportion of neurons, oligodendrocytes, and astrocytes was calculated as a percentage of the total number of cells expressing GFP (n = 3). B, Summary of the percentage of GFP-expressing cells relative to the three different neural populations in the target area. Contrary to neonatal AAV-delivery, the MBP promoter robustly targeted the oligodendrocyte population following injection at P10 and P90 (MBP-P0: 3.0±0.2%, MBP-P10: 68.3±9.2%, MBP-P90: 53.3±12.5%). The GFAP promoter resulted in robust transgene expression in the astrocyte population regardless of the time point of AAV-injection (GFAP-P0: 52.0±5.6%, GFAP-P90: 65.4±3.1%). C, Volume showing GFP expression after AAV-MBP-GFP or AAV-GFAP-GFP delivery. Vector delivery at early time points result in higher efficacy compared to the adult stage (MBP-P0: 23.5±1.7 mm³, MBP-P10: 23.1±1.8 mm³, MBP-P90: 8.3±1.2 mm³, GFAP-P0: 26.7±5.8 mm³, GFAP-P90: 5.8±1.2 mm³). D, Vector spread relative to the whole brain volume (MBP-P0: 5.9±0.4%, MBP-P10: 4.6±0.4%, MBP-P90: 1.7±0.2%, GFAP-P0: 6.7±50.7%, GFAP-P90: 1.2±0.2%). p<0.001, 2-way ANOVA and Bonferroni post-test.

Figure 8. The MBP promoter is selective for oligodendrocytes in a mouse model of the leukodystrophy Canavan Disease.

AAV-MBP-GFP was administered to the striatum of ASPA-deficient mice at P10 and brains analysed three weeks later to determine spread and selectivity of transgene expression. Immunodetection of GFP⁺ cells (A) and counterstaining for NeuN (B) revealed robust GFP-staining in subcortical white matter tracts. While GFP immunoreactivity (green) mostly segregated from neurons (red), the merged image revealed some co-expressing cells in the neocortex (C). High power images of the striatum showed GFP-expression exclusively in white matter oligodendrocytes (D–F). Shown are representative results of three independent experiments. Bars: C, 1 mm; F, 50 µm.