Comparison of independent screens on differentially vulnerable motor neurons reveals alpha-synuclein as a common modifier in motor neuron diseases

Abstract

The term "motor neuron disease" encompasses a spectrum of disorders in which motor neurons are the primary pathological target. However, in both patients and animal models of these diseases, not all motor neurons are equally vulnerable, in that while some motor neurons are lost very early in disease, others remain comparatively intact, even at late stages. This creates a valuable system to investigate the factors that regulate motor neuron vulnerability. In this study, we aim to use this experimental paradigm to identify potential transcriptional modifiers. We have compared the transcriptome of motor neurons from healthy wild-type mice, which are differentially vulnerable in the childhood motor neuron disease Spinal Muscular Atrophy (SMA), and have identified 910 transcriptional changes. We have compared this data set with published microarray data sets on other differentially vulnerable motor neurons. These neurons were differentially vulnerable in the adult onset motor neuron disease Amyotrophic Lateral Sclerosis (ALS), but the screen was performed on the equivalent population of neurons from neurologically normal human, rat and mouse. This cross species comparison has generated a refined list of differentially expressed genes, including CELF5, Col5a2, PGEMN1, SNCA, Stmn1 and HOXa5, alongside a further enrichment for synaptic and axonal transcripts. As an in vivo validation, we demonstrate that the manipulation of a significant number of these transcripts can modify the neurodegenerative phenotype observed in a Drosophila line carrying an ALS causing mutation. Finally, we demonstrate that vector-mediated expression of alpha-synuclein (SNCA), a transcript decreased in selectively vulnerable motor neurons in all four screens, can extend life span, increase weight and decrease neuromuscular junction pathology in a mouse model of SMA. In summary, we have combined multiple data sets to identify transcripts, which are strong candidates for being phenotypic modifiers, and demonstrated SNCA is a modifier of pathology in motor neuron disease.

Fig 1. Intermuscular variability in the levels of neuromuscular junction pathology in the Smn^2B/- mouse model of SMA.

A) Schematic diagram showing location of muscles which were innervated by either vulnerable or resistant motor neurons in a mouse model of SMA. Vulnerable muscles, as defined by increased neuromuscular junction (NMJ) pathology, include: external and internal oblique; transversus abdominis; and rectus abdominis. Resistant muscles, as defined by a low level of NMJ pathology, include: levator auris longus; auricularis superior; and adductor auris longus. B) Confocal micrographs showing NMJs with the pre-synaptic terminal labeled with antibodies against neurofilament (NF; Green) and synaptic vesicle protein 2 (green) and the muscle endplate labeled with alpha-bungarotoxin (red) from rectus abdominis and auricularis superior muscles. Note that in the wild-type abdominal and Smn^2B/- cranial muscle, all endplates appear fully innervated where each endplate is covered by the pre-synaptic terminal labeled with SV2 and NF. In the rectus abdominis from the Smn^2B/-, mouse there is evidence of significant NMJ pathology, as evidenced by endplates lacking a pre-synaptic terminal (white arrow heads). Scale bar = 40μm.

Fig 2. Comparison of 4 independent screens on differentially vulnerable motor neurons reveals a large number of common transcriptional changes.

Scatter plot showing the fold change of transcripts which were differentially expressed in differentially vulnerable motor neurons in the RNAseq performed by Murray et al., 2015 [16] and in the microarray study on differentially vulnerable motor neuron performed by Brockington et al., 2013 [17] (green), Kaplan et al., 2014 [19] (blue) and Hedlund et al., 2010 [18](red). Numbers denote number of number of common transcriptional screens within each quadrant of the plot. Note that the majority of changes occur with a common directional change i.e. fall within the bottom left or top right quadrant of the graph.

Fig 3. In vivo validation of transcripts in a Drosophila model of ALS8.

A) Images show eyes from wild-type, DVAP-P58S (carrying ALS8 patient mutation), DVAP-P58S Dfd (AL8 patient mutation with decreased expression of Dfd, the Drosophila homologue of Hoxc4) and DVAP-P58S Dfd (AL8 patient mutation with decreased expression of Nrbp1) Drosophila. Note the decrease in eye size observed in DVAP-P58S flies. This phenotype was supressed by decreased expression of dfd, and enhanced by decreased expression of Nrbp1. B) Bar chart (Mean ± SEM) showing the area (mm²) of the eye in Drosophila lines which over or under express specified transcripts in DVAP-P58S flies. *** P<0.001, **P<0.01 by ANOVA with Holm-Sidak’s post hoc test. N = approx. 12 flies per group with each value reflecting average of 2 eyes per fly. C) Table details the transcripts which were identified as modifiers of the eye phenotype in DVAP-P58S flies, denoting the Drosphila official gene symbol, the official gene symbol of the mouse homologue and the directional change observed in vulnerable motor neurons in the independent transcriptional screens [–19]. For those transcripts which were decreased in vulnerable motor neurons, their expression was increased in DVAP-P58S flies, and for those transcripts which were increased in vulnerable motor neurons, their expression was decreased in DVAP-P58S flies.

Fig 4. Overexpression of SNCA ameliorate phenotype and neuromuscular junction pathology in the Smn^2B/- mouse model of SMA.

A, B) Kaplan Maier plot (A) and weight curve (B) showing profile of control (Smn^2B/+; black) and untreated Smn^2B/- (red) compared to mice treated with a low dose (1e¹¹; blue) or high dose (3e¹¹; green) of AAV9 expressing SNCA. Note that while the low dose only increases weight gain (Student’s t-test p < 0.0001, the high dose of AAV9-SNCA significantly increased weight (Student’s t-test p<0.0001) and lifespan in the Smn^2B/- mouse model (Mantel-Cox Survival Curve Comparison Test p = 00027). C) Confocal micrographs showing NMJs with the pre-synaptic terminal labeled with antibodies against neurofilament (NF; red) and synaptic vesicle protein 2 (red) and the muscle endplate labeled with alpha-bungarotoxin (green) from the transversus abdominis muscle from P18 mice. Note the presence of fully (arrowhead) and partially (arrow) denervated endplates in the untreated Smn^2B/- mouse which were less commonly observed in the Smn^2B/- mouse treated with high dose AAV9-SNCA. Scale bar = 20μm. D) Bar chart ± SEM showing the increase in the percentage of fully occupied endplates in untreated Smn^2B/- mice (black bars) compared to Smn^2B/- treated with high dose AAV9-SNCA. *** P <0.001 by Mann-Whitney U test where n = 4/8 mice/muscles per group.