Widespread and efficient transduction of spinal cord and brain following neonatal AAV injection and potential disease modifying effect in ALS mice

Abstract

The architecture of the spinal cord makes efficient delivery of recombinant adeno-associated virus (rAAV) vectors throughout the neuraxis challenging. We describe a paradigm in which small amounts of virus delivered intraspinally to newborn mice result in robust rAAV-mediated transgene expression in the spinal cord. We compared the efficacy of rAAV2/1, 2/5, 2/8, and 2/9 encoding EGFP delivered to the hindlimb muscle (IM), cisterna magna (ICM), or lumbar spinal cord (IS) of neonatal pups. IS injection of all four capsids resulted in robust transduction of the spinal cord with rAAV2/5, 2/8, and 2/9 vectors appearing to be transported to brain. ICM injection resulted in widespread expression of EGFP in the brain, and upper spinal cord. IM injection resulted in robust muscle expression, with only rAAV2/8 and 2/9 transducing spinal motor and sensory neurons. As proof of concept, we use the IS paradigm to express murine Interleukin (IL)-10 in the spinal cord of the SOD1-G93A transgenic mouse model of amyotrophic lateral sclerosis. We show that expression of IL-10 in the spinal axis of SOD1-G93A mice altered the immune milieu and significantly prolonged survival. These data establish an efficient paradigm for somatic transgene delivery of therapeutic biologics to the spinal cord of mice.

Figure 1

Demonstration of the three routes of rAAV injection used in this study. For the purpose of demonstration, 1 µl of 1% Evans blue dye was used followed by immediate visualization of the extent and localization of the blue dye. IS injection of 1 µl dye into the lumbar spinal cord ta a depth of 1–2 mm (a) resulted in coverage of the lower half of the spinal cord (d). ICM injection site is located under the skull and is best targeted when the pups head is bent forward (b,d). The brainstem and fourth ventricle are stained with the dye following injection of the dye into a depth of 1.5 mm (inset, b). For targeting the hindlimb muscle, IM injection was performed (c,d) at a depth of 3–5 mm, resulting in localized dye in the thigh region. (e) Representative fluorescence micrographs of formalin fixed entire spinal cords of unininjected mice (Control) or mice injected neonatally with rAAV2/1, 2/5, 2/8, or 2/9 expressing EGFP. IS injection of all serotypes resulted in robust and widespread expression of EGFP. ICM and IM injections of rAAV2/8 and rAAV2/9 displayed widespread EGFP expression, while injection of rAAV2/1 and rAAV2/5 via the ICM and IM routes showed either low levels of EGFP fluorescence or no detectable expression, respectively. Scale bar, 2 mm. n = 3–4/group. rAAV, recombinant adeno-associated virus; IS, lumbar spinal cord; ICM, cisterna magna; IM, intramuscular.

Figure 2

Comparative spinal cord biodistribution patterns of different rAAVs injected via IS, ICM, or IM routes. EGFP expression in cervical and lumbar spine in mice injected with rAAV-EGFP via IS, ICM, and IM routes on neonatal day P0. Uninjected tissue served as control. The 3-week old IS cohort showed EGFP expression spread throughout the spinal cord for all the serotypes tested. ICM injected cohorts show robust EGFP expression for rAAV2/8 and 2/9, but weak and infrequent expression for 2/5 or 2/1. All IM injected mice show widespread hindlimb muscle transduction and widespread but weak transgene expression in rAAV2/8 and rAAV2/9 injected spinal cords while no detectable EGFP expression was seen in either rAAV2/1 or rAAV2/5 injected spinal cords. Scale bar, 500 µm (Spinal cord), 200 µm (muscle). n = 3–4/group. rAAV, recombinant adeno-associated virus; IS, lumbar spinal cord; ICM, cisterna magna; IM, intramuscular.

Figure 3

Comparative cellular tropism of spinal cord targeted rAAVs following IS delivery in neonatal mice. EGFP expression in the spinal cords of mice injected with different rAAV serotypes via IS route on neonatal day P0. Representative tricolor merged fluorescent photomicrograph from 3-week-old mice spinal cord sections were generated following co-labeling with anti EGFP antibody (488 nm—green), DAPI (350 nm—blue) and cell type specific marker antibodies (568 nm—red) for neurons (NeuN), motor neurons (ChAT), and astrocytes (GFAP). Scale bar, 50 µm. rAAV, recombinant adeno-associated virus; IS, lumbar spinal cord.

Figure 4

Differential transduction patterns of the frontal cortex by rAAVs delivered via IS, ICM, or IS routes. EGFP expression in the frontal cortex of mice injected with different rAAV serotypes via IS, ICM, or IM on neonatal day P0. Uninjected tissue served as control. Representative sections from 3-week-old mice show EGFP staining in the frontal cortex area of the brain. Scale bar, 500 µm. n = 3–4/group. rAAV, recombinant adeno-associated virus; IS, lumbar spinal cord; ICM, cisterna magna; IM, intramuscular.

Figure 5

Extension of life span in rAAV2/1-IL-10 injected SOD1G93A mice. (a) IS administration of rAAV2/1-IL-10 in neonatal mice increases life span of SOD1G93A mice without affecting disease onset. Kaplan–Meier curve for survival shows that IL-10 expression extends the life span by 15.2% over the control group. No difference in median survival was noted between the rAAV2/1-EGFP injected mice and control (uninjected) mice. (b) IL-10 injected SOD1G93A mice maintained their body weights as compared to age-matched controls during their prolonged survival. Body weights of mice are graphically presented from day 70 postnatal till day 160, when the entire control group succumbed to paralysis. IL-10 expression did not significantly changed the body weight of non-Tg mice; however, the IL-10 injected SOD1G93A mice showed consistent trend of lower weight compared to controls (P = 0.1). n = 12, 7, 22 for uninjected, rAAV2/1-EGFP, and rAAV2/1-IL-10 nontransgenic cohort; n = 16, 7, 14 for uninjected, rAAV2/1-EGFP, and rAAV2/1-IL-10 SOD1G93A cohort. (c) IL-10 levels were quantified by ELISA of soluble spinal cord lysates from the rostral and caudal parts from 3 weeks old SOD1 G93A mice injected with rAAV-IL-10 on neonatal day P0. n = 5–10 mice, P < 0.05, t-test. (d) GFAP and Iba-1 immunohistochemistry showed increased immunoreactivity in the spinal cords of end-stage SOD1G93A (G93A±) mice as compared to age-matched nontransgenic (NTg) controls. IL-10 overexpression did not significantly modify this pathology. n = 5/group. (e) Representative anti-GFAP immunoblot and intensity analysis of immunoreactive GFAP protein levels (mean ± SEM), normalized to β-actin showing that SOD1 mice have elevated levels of GFAP as compared to controls. There were no significant changes in GFAP levels in mice overexpressing IL-10 compared to genotype-matched controls (n = 3/group; P > 0.05, t-test). (f) Nanostring RNA profiling of spinal cords from nontransgenic and SOD1G93A mice showed that multiple immune mediators were affected in response to IL-10. A representative set of immune genes that were most significantly affected are depicted here after adjusting for a false discovery rate of 0.1%. For the complete array data, please see Supplementary Table S2. n = 3–4/group; P < 0.001, 1 way Anova. rAAV, recombinant adeno-associated virus; IS, lumbar spinal cord; IL, interleukin.

Figure 6

Schematic depiction of rAAV transduction in the brain and spinal cord following IM, ICM, and IS injection. Data based on EGFP expression patterns from this study has been summarized to depict the relative biodistribution properties of different rAAV serotypes transduced via IM, ICM, and IS injection routes. (diamond symbol) rAAV2/1; (filled circle) rAAV2/5; (squate symbol) rAAV2/8; (triangle symbol) rAAV2/9. Please see Table 1 for details. rAAV, recombinant adeno-associated virus; IS, lumbar spinal cord; ICM, cisterna magna; IM, intramuscular.