Intravascular AAV9 preferentially targets neonatal neurons and adult astrocytes

Abstract

Delivery of genes to the brain and spinal cord across the blood-brain barrier (BBB) has not yet been achieved. Here we show that adeno-associated virus (AAV) 9 injected intravenously bypasses the BBB and efficiently targets cells of the central nervous system (CNS). Injection of AAV9-GFP into neonatal mice through the facial vein results in extensive transduction of dorsal root ganglia and motor neurons throughout the spinal cord and widespread transduction of neurons throughout the brain, including the neocortex, hippocampus and cerebellum. In adult mice, tail vein injection of AAV9-GFP leads to robust transduction of astrocytes throughout the entire CNS, with limited neuronal transduction. This approach may enable the development of gene therapies for a range of neurodegenerative diseases, such as spinal muscular atrophy, through targeting of motor neurons, and amyotrophic lateral sclerosis, through targeting of astrocytes. It may also be useful for rapid postnatal genetic manipulations in basic neuroscience studies.

Figure 1

Intravenous injection of AAV9 leads to widespread neonatal spinal cord transduction. Cervical (a–d) and lumbar (e–l) spinal cord sections ten-days following facial-vein injection of 4×10¹¹ particles of scAAV9-CB-GFP into postnatal day-1 mice (n=8). GFP-expression (a,e,i) was predominantly restricted to lower motor neurons (a,e,i) and fibers that originated from dorsal root ganglia (a,e). GFP-positive astrocytes (i,l, arrows) were also observed scattered throughout the tissue sections. Lower motor neuron and astrocyte expression were confirmed by co-localization using choline acetyl transferase (ChAT) (b, f,j) and glial fibrillary acidic protein (GFAP) (c,g,k), respectively. Merged images of GFP, ChAT and GFAP (d,h,l). A z-stack image (i–l) of the area within the box in h, shows the extent of motor neuron (arrow heads) and astrocyte (arrows) transduction within the lumbar spinal cord. Scale bars, 200 μm (d,h), 20 μm (l).

Figure 2

In situ hybridization of spinal cord sections from neonate and adult injected animals demonstrates that cells expressing GFP are transduced with scAAV9-CB-GFP. Negative control animals injected with PBS (a–b) showed no positive signal (n=2). However, antisense probes for GFP demonstrated strong positive signals for both neonate (n=3) (c) and adult (n=3) (e) sections analyzed. No positive signals were found for the sense control probe in neonate (d) or adult (f) spinal cord sections. Tissues were counterstained with Nuclear Fast Red for contrast while probe hybridization is in purple.

Figure 3

Widespread GFP-expression 21-days following intravenous injection of 4×10¹¹ particles of scAAV9-CB-GFP to postnatal day-1 mice. GFP localized in neurons and astrocytes throughout multiple structures of the brain (n=6) as depicted in: (a) striatum (b) cingulate gyrus (c) fornix and anterior commissure (d) internal capsule (e) corpus callosum (f) hippocampus and dentate gyrus (g) midbrain and (h) cerebellum. All large panels show GFP (green) and dapi (blue) merged images. Insets of selected regions show high magnification merged images of GFP (green), NeuN (red) and GFAP (blue) labeling. Schematic representations depicting the approximate locations of each image throughout the brain are shown in (Supplementary Fig. 6). Higher magnification images of select structures are available in (Fig. 4). Scale bars, 200 μm (a); 50 μm (e); 100 μm (b–d,f–h)

Figure 4

High-magnification of merged GFP (green) and dapi (blue) images of brain regions following neonate (n=8) (a–d) or adult (n=6) (e–f) intravenous injection of scAAV9-CB-GFP. Astrocytes and neurons were easily detected in the striatum (a), hippocampus (b) and dentate gyrus (c) following postnatal day-1 intravenous injection of 4×10¹¹ particles of scAAV9-CB-GFP. Extensive GFP-expression within cerebellar Purkinje cells (d) was also observed. Pyramidal cells of the hippocampus (e) and granular cells of the dentate gyrus (f) were the only neuronal transduction within the brain following adult tail vein injection. In addition to astrocyte and neuronal transduction, widespread vascular transduction (f) was also seen throughout all adult brain sections examined. Scale bars, 200μm (e); 100μm (f), 50μm (a–d)

Figure 5

Intravenous injection of AAV9 leads to widespread predominant astrocyte transduction in the spinal cord and brain of adult mice. GFP-expression in the cervical (a–d) and lumbar (e–h) spinal cord as well as the brain (m–p) of adult mice 7-weeks after tail vein injection of 4×10¹² particles of scAAV9-CB-GFP (n=6). In contrast to postnatal day-1 intravenous injections, adult tail vein injection resulted in almost exclusively astrocyte transduction. GFP (a,e), ChAT (b,f) and GFAP (c,g) demonstrate the abundance of GFP-expression throughout the spinal grey matter, with lack of co-localization with lower motor neurons and white matter astrocytes (d,h). Co-localization of GFP (i), excitatory amino acid transporter 2 (EAAT2) (j), and GFAP (k) confirm that transduced cells are astrocytes (l). Tail vein injection also resulted in primarily astrocyte transduction throughout the brain as seen in the cortex (m–n), thalamus (o) and midbrain (p). All brain panels show GFP (green) and dapi (blue) merged images. Neuronal GFP-expression in the brain was restricted to the hippocampus and dentate gyrus (m–n, Fig. 4 e–f). Schematic representations depicting the approximate locations of each image throughout the brain are available in. (Supplementary Fig. 7) Scale bars, 10 μm (l),200 μm (d,h,p)