Intrathecal delivery of recombinant AAV1 encoding hepatocyte growth factor improves motor functions and protects neuromuscular system in the nerve crush and SOD1-G93A transgenic mouse models

Abstract

Amyotrophic lateral sclerosis (ALS) is a fatal neuromuscular disease resulting from motor neuron degeneration that causes muscle weakness, paralysis, and eventually respiratory failure. We investigated whether recombinant adeno-associated virus encoding human hepatocyte growth factor (rAAV-HGF) could generate beneficial effects in two mouse models with neuromuscular problems when intrathecally delivered to the subarachnoid space. We chose AAV serotype 1 (rAAV1) based on the expression levels and distribution of HGF protein in the lumbar spinal cord (LSC). After a single intrathecal (IT) injection of rAAV1-HGF, the protein level of HGF in the LSC peaked on day 14 and thereafter gradually decreased over the next 14 weeks. rAAV1-HGF was initially tested in the mouse nerve crush model. IT injection of rAAV1-HGF improved mouse hindlimb strength and rotarod performance, while histological analyses showed that the length of regenerated axons was increased and the structure of the neuromuscular junction (NMJ) was restored. rAAV1-HGF was also evaluated in the SOD1-G93A transgenic (TG) mouse model. Again, rAAV1-HGF not only improved motor performance but also increased the survival rate. Moreover, the number and diameter of spinal motor neurons (SMNs) were increased, and the shape of the NMJs restored. Data from in vitro motor cortical culture experiments indicated that treatment with recombinant HGF protein (rHGF) increased the axon length of corticospinal motor neurons (CSMNs). When cultures were treated with an ERK inhibitor, the effects of HGF on axon elongation, protein aggregation, and oxidative stress were suppressed, indicating that ERK phosphorylation played an important role(s). Taken together, our results suggested that HGF might play an important role(s) in delaying disease progression in the SOD1-G93A TG mouse model by reducing oxidative stress through the control of ERK phosphorylation.

Keywords: Adeno-associated virus (AAV); Amyotrophic lateral sclerosis (ALS); Corticospinal motor neuron (CSMN); Hepatocyte growth factor (HGF); Oxidative stress.

Fig. 1

IT administration of rAAV1-HGF promoted functional recovery in the sciatic nerve crush mouse model. a 4.12 × 10⁸ GC of rAAV1, 2, 5, or 6-HGF were intrathecally injected into C57BL/6 mice at postnatal day 60 (P60). The LSCs were collected and total proteins were isolated 4 weeks after injection, followed by ELISA specific to human HGF (hHGF). For statistical analysis, one-way ANOVA was performed, followed by Tukey’s post-hoc test. ^****p < 0.0001 for rAAV1 vs. other serotypes. b 1.4 × 10⁹ GC of four serotypes were intrathecally injected into 2-month-old C57BL/6 mice. The LSCs were collected 4 weeks after injection. Tissues were fixed, followed by IHC assay using an antibody specific to GFP (green). The boundary between white and grey matter is distinguished by dotted lines, and white matter is indicated by white arrows. c 5 × 10⁹ GC of rAAV1-HGF were intrathecally injected into 2-month-old C57BL/6 mice. The LSCs were collected 1, 2, 4, 8, 12, and 16 weeks after IT injection and subjected to ELISA for hHGF. For statistical analysis, two-way ANOVA was performed, followed by Sidak’s post-hoc test. The p-value between the two groups was 0.0002. d C57BL/6 mice at P60 were intrathecally injected with 5 × 10⁹ GC of rAAV1-C or rAAV1-HGF. One week later, LSCs were collected and subjected to IHC assay. Antibodies specific to ChAT (green) and NeuN (red) were used to label SMNs, together with those for p-Met (magenta). In low magnification panels, the boundary between white and grey matter is distinguished by dotted lines. In merge panels, p-Met-expressing SMNs are indicated by white arrows. e The proportion of SMNs expressing p-Met per total SMNs was measured and represented as a bar graph. For statistical analysis, Student’s t-test was performed. ^****p < 0.0001. In bar graphs, values are represented as mean ± SEM. Scale bar: b = 100 μm, d = 100 μm for low magnification panels and 20 μm for the others

Fig. 2

IT delivery of rAAV1-HGF facilitated the regeneration of the sciatic nerves and TA muscles after sciatic nerve crush. a-h Using 2-month-old C57BL/6 mice, sciatic nerve crush was induced, and 5 × 10¹¹ GC of rAAV1-C or rAAV1-HGF were intrathecally injected into the LSC. Behavioral tests were performed once a week for 28 days (a-b). For IHC assay, the sciatic nerves and TA muscles were collected 5 days after nerve crush (c-f). a Hindlimb strength was measured 1, 7, 10, 14, 21, and 28 days after nerve crush. Value at day 0 was measured before nerve crush. b Mean latency to fall from the rotarod was measured 1, 7, 10, 14, 21, and 27 days after nerve crush. Value at day 0 was measured before nerve crush. In Fig. 2a-b, two-way ANOVA was performed, followed by Tukey’s post-hoc test. c Representative images of the sciatic nerves from mice treated with rAAV1-C or rAAV1-HGF after nerve crush. Antibodies specific to TUJ1 and SCG10 were used as markers for neurons and regenerating axons, respectively. Crush sites are indicated by dotted lines, and the tip of SCG10 signals is indicated by white arrows. Proximodistal direction is indicated by dotted lines. d After sciatic nerve crush, lengths of regenerated nerves were measured using Fiji software and represented as a bar graph. For statistical analysis, Student’s t-test was performed. e-f An antibody specific to UCHL1 was used as a marker for presynaptic terminals, whereas that of α-bungarotoxin (α-BTX) was used for postsynaptic end plates [9]. The shapes of NMJs were analyzed to determine whether they were pretzel-shaped or distorted (e). The integrity of NMJs was determined by measuring to what degree presynaptic terminals merged with postsynaptic end plates (f). g-i After sciatic nerve crush, the shape (g) and integrity (h) of NMJs and average α-BTX area (i) were measured. For graphs, values are represented as mean ± SEM. In Fig. 2g-h, two-way ANOVA was performed, followed by Tukey’s post-hoc test. In Fig. 2g, ^****p < 0.0001 for modified NMJ, ^####p < 0.0001 for pretzel-shaped NMJ. In Fig. 2h, ^**p < 0.01 and ^****p < 0.0001 for fully innervated NMJ, ^####p < 0.0001 and n.s. > 0.05 partially innervated NMJ. In Fig. 2i, one-way ANOVA was performed, followed by Tukey’s post-hoc test. For the remaining bar graphs, ^*p < 0.05, ^**p < 0.005, ^***p < 0.001, ^****p < 0.0001. Scale bars: c = 200 μm; d-e = 20 μm

Fig. 3

IT delivery of rAAV1-HGF ameliorated disease progression and prolonged the survival rate of SOD1-G93A TG mice. a-f SOD1-G93A TG mice at P60 were intrathecally injected with 5 × 10¹¹ GC of rAAV1-C or rAAV1-HGF. Behavioral tests including those for forelimb strength (a), hindlimb strength (b), rotarod (c), and hanging wire (d) were performed, and body weight (e) was measured once a week until P158. In Fig. 3a-e, one-way ANOVA was performed followed by Tukey’s post-hoc test at each time point. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001. f Survival curve was drawn based on Mantel-Cox and Gehan-Breslow-Wilcoxon tests for comparing two groups. In the Gehan-Breslow-Wilcoxon test, p value was 0.0115. Median survival days were 145.5 for the rAAV1-C-treated group, and 155 for the rAAV1-HGF-treated group. For graphs, values are represented as mean ± SEM

Fig. 4

IT delivery of rAAV1-HGF led to histological improvements of SMNs and NMJs in SOD1-G93A TG mice. a-g non-TG or TG mice at P60 were intrathecally injected with rAAV1-C or rAAV1-HGF. The LSCs (a-d) and TA muscles (e-f) were collected at P100, followed by IHC assay. a Representative image of SMNs. Antibodies specific to ChAT and NeuN were used to label SMNs. b Number of SMNs per ventral horn was counted and represented as a bar graph. c Proportion of SMNs per DAPI-positive cells was counted and represented as a bar graph. d Diameter of SMNs were measured and represented as a bar graph. In, Fig. 4b-d, one-way ANOVA was performed, followed by Tukey’s post-hoc test. e-f The shape and integrity of NMJs were determined as mentioned in Fig. 2 and represented as a bar graph. In Fig. 4e-f, two-way ANOVA was performed, followed by Tukey’s post-hoc test. In Fig. 4e, ^****p < 0.0001 for modified NMJ and ^####p < 0.0001 for pretzel-shaped NMJ. In Fig. 4f, ^****p < 0.0001 for fully innervated NMJ and ^####p < 0.0001 for partially innervated NMJ. g Relative mass of TA was calculated using total body mass and represented as a bar graph. For bar graphs, values are represented as mean ± SEM. For statistical analysis, one-way ANOVA was performed, followed by Tukey’s post-hoc test. For bar graphs, ^*p < 0.05, ^**p < 0.005, ^***p < 0.001, n.s. > 0.05. Scale bar: a = 50 μm

Fig. 5

Treatment with rHGF enhanced the axonal outgrowth of CSMNs in motor cortical cultures. a-d SOD1-G93A TG mice were sacrificed at P3. After dissociating motor cortices, 20,000 cells were seeded on 24-well plates. Three days later, cells were fixed and subjected to immunocytochemistry (ICC) assay. a Representative image of CSMNs. Antibodies specific to UCHL1 (green) and Ctip2 (red, not shown) were used to label CSMNs. Serum-free media was used as a control medium (C). Cells were visualized with confocal laser scanning microscopy. The axon length of CSMNs was measured using Fiji software (yellow line). The axon length of CSMNs was measured after treatment with PHA665752, an inhibitor for Met (b) or rHGF (c) or rHGF plus PHA665752 (d). For bar graphs, values are represented as mean ± SEM. In Fig. 5b-d, one-way ANOVA was performed, followed by Tukey’s post-hoc test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001. Scale bar: a = 50 μm

Fig. 6

Inhibition of ERK, PI3K, and p38 resulted in decreased axonal outgrowth of CSMNs. a-d The axon length of CSMNs was measured and represented as a bar graph after treatment with U0126 (ERK inhibitor) (a), LY294002 (PI3K inhibitor) (b), SB203580 (p38 inhibitor) (c), and SP600125 (JNK inhibitor) (d). For bar graphs, values are represented as mean ± SEM. In Fig. 6a-d, one-way ANOVA was performed, followed by Tukey’s post-hoc test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001, n.s. > 0.05

Fig. 7

Phosphorylation of ERK was increased by treatment with HGF in the LSC and motor cortical cultures. a-c Motor cortices from non-TG or TG mice at P3 were collected. After dissociation, cells were seeded on six-well plates with 3.2 × 10⁶ cells/well. a Cells were treated with 100 ng/ml of rHGF, and 5 days later, total proteins were isolated, followed by Western blot analysis. b-c Cells were treated with 1 μM of PHA665752 or 10 μM of U0126. After 30 min, cells were treated with 100 ng/ml of rHGF, and five days later, total proteins were isolated followed by Western blot analysis. DMSO was used as a negative control. d The axon length of CSMNs was measured and represented as a bar graph after co-treatment with rHGF and U0126. For statistical analysis, one-way ANOVA was performed, followed by Tukey’s post-hoc test. e-f non-TG or TG mice at P60 were intrathecally injected with rAAV1-C or rAAV1-HGF. The LSCs were collected, and total proteins were extracted at P100, followed by Western blot analysis (e). Relative levels of phosphorylated ERK were measured using Fiji software and represented as a bar graph (f). For bar graphs, values are represented as mean ± SEM. For statistical analysis, one-way ANOVA was performed, followed by Tukey’s post-hoc test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001

Fig. 8

Treatment with rHGF reduced levels of protein aggregation and oxidative stress induced by mutant SOD1. a-b ICC assay was performed as described in Fig. 5a. Antibodies specific to UCHL1 and Ctip2 were used to label CSMNs, together with those for hSOD1 (magenta). C: Treated with SFM. In the hSOD1-stained panels, the cell boundaries of the CSMNs were outlined based on the UCHL1 signals, except for the panel in the second row, where the cell boundaries are ambiguous. Proportion of hSOD1-positive cells per DAPI-positive cells was counted and represented as a bar graph (b). For statistical analysis, one-way ANOVA was performed, followed by Tukey’s post hoc test. ^****p < 0.0001 for non-TG SFM vs. TG SFM, ^###p < 0.001 for TG SFM vs. TG HGF, ⁺⁺p < 0.01 for TG HGF vs. TG HGF + PHA665752 (or TG HGF + U0126). c 6.4 × 10⁴ cells were seeded on 96-well plates. Cellular ROS levels were measured using a ROS detection kit and represented as a bar graph. Pyocyanin was used as an inducer of ROS, whereas N-acetyl-L-cysteine (NAC) was employed as a scavenger of ROS. For statistical analysis, one-way ANOVA was performed, followed by Tukey’s post-hoc test. ^****p < 0.0001 for SFM vs. pyocyanin (or NAC or HGF), ^##p < 0.01 for HGF vs. HGF + PHA665752, ^###p < 0.001 for HGF vs. HGF + U0126. d The axon length of CSMNs was measured and represented as a bar graph. For bar graphs, values are represented as mean ± SEM. For statistical analysis, one-way ANOVA was performed, followed by Tukey’s post-hoc test. ^****p < 0.0001 for non-TG SFM vs. TG SFM, ^####p < 0.0001 for TG SFM vs. TG NAC (or TG HGF), ⁺⁺⁺⁺p < 0.0001 for TG HGF vs. TG HGF + pyocyanin. Scale bar: a = 20 μm