EphrinB2 knockdown in cervical spinal cord preserves diaphragm innervation in a mutant SOD1 mouse model of ALS

Abstract

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease characterized by motor neuron loss. Importantly, non-neuronal cell types such as astrocytes also play significant roles in disease pathogenesis. However, mechanisms of astrocyte contribution to ALS remain incompletely understood. Astrocyte involvement suggests that transcellular signaling may play a role in disease. We examined contribution of transmembrane signaling molecule ephrinB2 to ALS pathogenesis, in particular its role in driving motor neuron damage by spinal cord astrocytes. In symptomatic SOD1^G93A mice (a well-established ALS model), ephrinB2 expression was dramatically increased in ventral horn astrocytes. Reducing ephrinB2 in the cervical spinal cord ventral horn via viral-mediated shRNA delivery reduced motor neuron loss and preserved respiratory function by maintaining phrenic motor neuron innervation of diaphragm. EphrinB2 expression was also elevated in human ALS spinal cord. These findings implicate ephrinB2 upregulation as both a transcellular signaling mechanism in mutant SOD1-associated ALS and a promising therapeutic target.

Keywords: ALS; astrocyte; ephrin; human; motor neuron; mouse; neuroscience; respiratory.

Figure 1.. EphrinB2 expression was increased in ventral horn astrocytes.

SOD1^G93A mouse ventral horn cervical spinal cord tissue immuinostained for ephrinB2 at 60 days (b), 120 days (c), endstage (d), and WT mouse age-matched control (a), scale bar: 200 µm. Quantification of ephrinB2 expression within the ventral horn shows a progressive increase in expression over time compared to WT controls (e). Endstage ephrinB2 expression in the thoracic (f, g) and lumbar (h–k) regions; scale bar: 200 µm, 100 µm, respectively. Endstage SOD1^G93A mouse cervical spinal cord tissue co-immuostained for ephrinB2 (l, n, o, q) and neuronal and astrocyte lineage-specific markers NeuN (m–n) and GFAP (p–q), respectively; scale bar: 30 µm. Analysis in panels A-K: n=3–4 mice per genotype and per time point; 1–2 females and 2 males per condition. Analysis in panels L-O: n=3 mice per genotype and per time point; 1 female and 2 males per condition.

Figure 2.. Lenti-shRNA injection reduced ephrinB2 expression in cervical spinal cord astrocytes.

We injected 60-day-old SOD1^G93A mice with Lenti-GFP control or Lenti-shRNA-Efnb2 into cervical ventral horn (a). Injections were made bilaterally into six sites throughout C3-C5 to target the PhMN pool (b).Thirty µm transverse tissue sections of the cervical spinal cord show robust expression of the GFP reporter bilaterally within the ventral horn (c); scale bar: 500 µm. Cell lineage of viral transduction was assessed using three markers: neuronal marker NeuN, oligodendrocyte lineage marker Olig2, and astrocyte marker GFAP. Spinal cord tissue collected from SOD1^G93A injected with Lenti-GFP vector was sectioned at 30 µm and immunostained for NeuN (d-f), Olig2 (g–i) and GFAP (j–l); scale bar: 150 µm. Quantification of transduction lineage was assessed by counting the total numbers of GFP+ cells that were co-labeled with each lineage-specific marker and expressing this as a percentage of the total number of GFP+ cells (m). We assessed the amount of knockdown achieved by the Lenti-shRNA-Efnb2 vector by immunostaining endstage SOD1^G93A cervical spinal cord tissue with an anti-ephrinB2 antibody in both Lenti-GFP control (o, q) and Lenti-shRNA-Efnb2 (p, r) tissue; scale bar: 100 µm. Knockdown was quantified by counting the total number of GFP+ cells expressing ephrinB2+ within the cervical ventral horn (n). Analyses in all panels: n=3 mice per condition; 1 female and 2 males per condition.

Figure 3.. EphrinB2 knockdown protected cervical spinal cord motor neurons and preserved functional innervation of the diaphragm in SOD1^G93A mice.

SOD1^G93A mice injected with Lenti-GFP control or Lenti-shRNA-Efnb2 were cresyl violet stained and MN counts were performed at 117 days of age. Thirty µm transverse cervical spinal cord tissue sections were stained with cresyl violet (a); scale bar: 250 µm. The dotted box outlines the ventral horn and area of the image shown in (c). MN populations within the ventral horn were quantified (b). Representative images show a greater loss of MNs in Lenti-Control (c) compared to the Lenti-shRNA-Efnb2 (d) group; scale bar: 100 µm. SOD1^G93A mice injected with Lenti-GFP control or Lenti-shRNA-Efnb2 were assessed in vivo for PhMN-diaphragm innervation by electrophysiological analysis at 117 days of age. CMAP amplitudes were recorded from each hemi-diaphragm following ipsilateral phrenic nerve stimulation. Representative traces of Lenti-GFP (e) and Lenti-shRNA-Efnb2 (f) recordings show a larger CMAP amplitude in the Lenti-shRNA-Efnb2 group. Quantification of maximal CMAP amplitude shows significant preservation in the Lenti-shRNA-Efnb2-treated group compared to control (g). Analysis in panels A-D: n=4 mice per condition; 2 females and 2 males per condition. Analysis in panels E-G: n=4 mice per genotype and per time point; 2 females and 2 males per condition.

Figure 4.. Knockdown of ephrinB2 in the cervical spinal cord ventral horn did not extend survival or delay onset of disease in SOD1^G93A mice.

Biweekly weights of Lenti-shRNA-Efnb2 and Lenti-Control mice were recorded until endpoint sacrifice (a), and disease onset was measured for each animal when there was a 10% drop in body weight (b). Biweekly individual hindlimb grip strengths were assessed for each animal (c), and disease onset was recorded when the animal had a 10% decline in hindlimb grip strength (d). Each animal was also tested for forelimb grip strength (e), and disease onset was recorded when the animal had a 10% decline in forelimb grip strength (f). Weights, forelimb grip strength and hindlimb grip strength were taken biweekly starting one week prior to injection of Lenti-shRNA-Efnb2 or Lenti-GFP control, and all force measurements plotted were the average force (lb) of all animals combined in each group. Disease duration was determined by time from weight onset to endstage (g), hindlimb disease onset to endstage (h), and forelimb diseaseonset to endstage for each animal (i). Disease duration was also measured from the time of forelimb disease onset to time of hindlimb disease onset (j). Survival was measured as the day each animal reached endstage, which was determined by the righting reflex (k). Analyses in all panels: n=8–10 mice per genotype and per time point; 4–5 females and 4–5 males per condition.

Figure 5.. EphrinB2 knockdown preserved morphological innervation of the diaphragm in SOD1^G93A mice.

Diaphragm muscles were labeled with SMI-312 (green), SV2, (green) and alpha-Bungarotoxin (blue). Representative images of fully-innervated (a), partially-denervated (b) and completely-denervated ((c): arrowheads denote completely-denervated NMJs) NMJs are shown; all scale bars: 30 µm. Compared to SOD1^G93A mice treated with Lenti-GFP (d), animals injected with Lenti-shRNA-Efnb2 (e) showed greater preservation of PhMN innervation of the diaphragm NMJ. Quantification revealed a significant increase in the percentage of fully-innervated NMJs (f) and a decrease in the percentage of completely-denervated NMJs (g) in the Lenti-shRNA-Efnb2 group compared to Lenti-GFP controls. The percentage of partially-denervated NMJs was not statistically different between the two groups (h). Analyses in all panels: n=4 mice per condition; 2 females and 2 males per condition.

Figure 6.. EphrinB2 upregulation in human spinal cord from ALS donors with an SOD1 mutation.

Immunoblotting analysis on postmortem samples from human ALS donors with an SOD1 mutation and also non-diseased human samples. In the lumbar enlargement, there was a large increase in ephrinB2 protein expression in the SOD1 mutation ALS samples compared to the non-diseased controls (top blot). GFAP immunoblotting on the same lumbar spinal cord samples shows a robust increase in GFAP protein levels in the two samples with increased ephrinB2 expresison (middle blot). There was not increased ephrinB2 expression in a disease unaffected region in these same three ALS donor samples, as ephrinB2 protein levels were not elevated in the frontal cortex (top blot). Immunoblot for total protein (bottom blot). Demographic information: Donor 1 – death at 67 years; male; non-ALS; Donor 2 – death at 70 years; male; non-ALS; Donor 3 – death at 70 years; female; non-ALS; Donor 4 – death at 42 years; female; SOD1-D102H mutation; absence of C9orf72 repeat expansion; Donor 5 – death at 55 years; male; SOD1-A4V mutation; absence of C9orf72 repeat expansion; Donor 6 – death at 58 years; male; SOD1-V87A mutation; absence of C9orf72 repeat expansion.