Loss of spatacsin function alters lysosomal lipid clearance leading to upper and lower motor neuron degeneration

Abstract

Mutations in SPG11 account for the most common form of autosomal recessive hereditary spastic paraplegia (HSP), characterized by a gait disorder associated with various brain alterations. Mutations in the same gene are also responsible for rare forms of Charcot-Marie-Tooth (CMT) disease and progressive juvenile-onset amyotrophic lateral sclerosis (ALS). To elucidate the physiopathological mechanisms underlying these human pathologies, we disrupted the Spg11 gene in mice by inserting stop codons in exon 32, mimicking the most frequent mutations found in patients. The Spg11 knockout mouse developed early-onset motor impairment and cognitive deficits. These behavioral deficits were associated with progressive brain atrophy with the loss of neurons in the primary motor cortex, cerebellum and hippocampus, as well as with accumulation of dystrophic axons in the corticospinal tract. Spinal motor neurons also degenerated and this was accompanied by fragmentation of neuromuscular junctions and muscle atrophy. This new Spg11 knockout mouse therefore recapitulates the full range of symptoms associated with SPG11 mutations observed in HSP, ALS and CMT patients. Examination of the cellular alterations observed in this model suggests that the loss of spatacsin leads to the accumulation of lipids in lysosomes by perturbing their clearance from these organelles. Altogether, our results link lysosomal dysfunction and lipid metabolism to neurodegeneration and pinpoint a critical role of spatacsin in lipid turnover.

Keywords: Lipids; Lysosome; Motor neuron; Neurodegeneration; SPG11.

Copyright © 2017 The Authors. Published by Elsevier Inc. All rights reserved.

Figures

Fig. 1

Spg11 knockout causes a progressive motor deficit in mice. (A). Diagram showing the genomic structure of the Spg11 gene (top), the targeting vector (middle) and the targeted locus upon excision of the neomycin resistance cassette and action of the Cre-recombinase (bottom). The mutations introduced were c.6052C > T (p.Arg2018*), corresponding to c.6091C > T (p.Arg2031*) in human and c.6061C > T (p.Gln2021*), corresponding to c.6100C > T (p.Arg2034*) in human. (B) Western blot of brain protein extracts showing the absence of spatacsin in Spg11−/− samples and the strong decrease in the levels of spastizin. Representative image of three independent experiments. * : non-specific band. (C and D), Pictures showing an Spg11+/+ (C) and Spg11−/− mouse (D) at 16 months of age. Knockout mice had an abnormal posture and kyphosis of the spine.

Fig. 2

Spg11 knockout mice develop a progressive spastic and ataxic gait disorder. (A and B) Gait angle sketch (A) and values recorded (B) during a forced walk on a treadmill. The gait angle decreased in Spg11−/− mice from four months of age (n ≥ 12 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05 and ***p ≤ 0.001). (C) Step sequence regularity values recorded during a forced walk on a treadmill. Spg11−/− mice exhibited motor coordination impairment from eight months of age (n ≥ 12 animals/genotype/age; Kruskal-Wallis test; **p ≤ 0.01 and ***p ≤ 0.001). (D and E) Rotarod duration (D) and maximum speed (E). The duration and maximum speed of Spg11−/− in accelerated rotarod testing was less than those of heterozygous and control mice at as early as six weeks of age (n ≥ 12 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05, **p ≤ 0.01 and ***p ≤ 0.001). (F and G) The Y-maze principle (F) and spontaneous alternation values (G). Knockout mice did not show a preference between the visited and the unknown arm of the maze from four months of age (n ≥ 12 animals/genotype/age; Kruskal-Wallis test; ***p ≤ 0.001). This altered cognitive behavior underlines a spatial learning or a memory deficit. (H and I) Fear conditioning. Percentage of time spent in a frozen posture on day 1 before and after electrical shocks (H). There was no obvious learning deficit during the conditioning of the mice at any age. Percentage of time spent in a frozen posture on day 2, without any electrical shocks, after conditioning (I). Although the task was learned, knockout mice spent less time in a frozen position from eight months of age, indicating a cognitive and memory deficit (n ≥ 10 animals/genotype/age; Kruskal-Wallis test; **p ≤ 0.01 and ***p ≤ 0.001).

Fig. 3

Loss of spatacsin causes severe cortical damage. (A and B) Brain coronal slices of Spg11+/+ (A) and Spg11−/− (B) mice at 16 months of age. GFAP (green; astrocyte marker), and NeuN (magenta; neuronal marker) immunostaining. Scale bars: 1 mm. (C) Sagittal view of mouse brain showing the location of the coronal cut sections between Bregma 0.90 and 0.60 mm. (D–F) Brain (D), cortex (E) and corpus callosum (F) surface-areas. The surface-areas of the cortex and corpus callosum decreased from eight and four months, respectively, prior to general brain atrophy (n ≥ 5 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05 and ***p ≤ 0.001). (G) Diagram of the primary motor cortex layers between Bregma 0.90 and 0.60 mm. (H and I) NeuN (magenta) and GFAP (green) immunostaining in the primary motor cortex of Spg11+/+ (H) and Spg11−/− (I) mice at 16 months of age. Scale bars: 200 μm. (J and K) Quantification of total NeuN-positive cells in the primary motor cortex layers I + II + III (J) and V + VI (K). Immunostaining revealed a significant reduction in the number of neurons from layers I–VI of the motor cortex in knockout mice from eight months of age. (n ≥ 10 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; **p ≤ 0.01 and ***p ≤ 0.001). (L) Western blot of 16-month-old Spg11+/− and Spg11−/− brain samples showing the presence of cleaved (active) caspase-3 in the absence of spatacsin. (M and N) Quantification of total GFAP-positive cells in the primary motor cortex layers I + II + III (M) and V + VI (N). Neuron loss was accompanied by marked astrogliosis (n ≥ 10 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; ***p ≤ 0.001).

Fig. 4

Loss of spatacsin causes cerebellar and hippocampal impairment. (A) Sagittal view of mouse brain showing the location of the coronal cut sections between Bregma − 6.00 and − 6.30 mm. (B) The surface-area of the cerebellum was lower in the knockout than in the control mice from eight months. (n ≥ 5 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05). (C) The number of Purkinje cells was lower from eight months of age in knockout mice (n ≥ 5 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; ***p ≤ 0.001). (D and E) Cerebellar sections immunostained for GFAP (green; astrocyte marker), Calbindin (magenta, Purkinje cell marker), and Hoechst-33258 (blue, nucleus marker) revealed a severe loss of Purkinje cells in knockout mice (E) compared to control mice (D). Scale bars: 100 μm. PCL: Purkinje Cell Layer. (F) Sagittal view of mouse brain showing the location of the coronal cut sections between Bregma − 1.70 and − 2.00 mm. (G–H) NeuN (magenta; neuronal marker) and GFAP (green; astrocyte marker) immunostaining in the hippocampus of Spg11+/+ (G) and Spg11−/− (H) mice at 16 months of age. Scale bars: 0.5 mm. CA1: cornu Ammonis 1 region, CA2: cornu Ammonis 2 region, CA3: cornu Ammonis 3 region, DG: Dentate Gyrus region. (I) Hippocampal surface-area. The surface-area of the hippocampus was smaller in knockout than control mice at all ages (n ≥ 5 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05). (J–L) CA1 (J), CA2 + CA3 (K), and dentate gyrus (L) granule layer surface-areas. The surface-area of CA1 and the dentate gyrus granule layer were affected by atrophy from four months of age in knockout mice compared to control mice (n ≥ 5 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05).

Fig. 5

Spg11−/− mice show spinal cord atrophy and axonal degeneration of the corticospinal tract. (A and B) Toluidine blue-stained transverse slices of Spg11+/+ (A) and Spg11−/− (B) mouse cervical spinal cord at 16 months of age. Scale bars: 0.5 mm. GM: gray matter, WM: white matter. (C–F) Surface-areas of the white and gray matter of the cervical (C and E) and lumbar (D and F) spinal cord at eight and 16 months of age. At eight months, the white matter surface-area of the lumbar spinal cord was lower in Spg11−/− mice than controls and at 16 months of age, the whole spinal cord of knockout mice was atrophied (n ≥ 5 slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05 and **p ≤ 0.01). (G and H) Toluidine blue-stained transverse slices of Spg11+/+ (G) and Spg11−/− (H) mouse corticospinal tract at 16 months of age. Scale bars: 10 μm. (I and J) Toluidine blue-stained transverse slices of Spg11+/+ (G) and Spg11−/− (H) mouse cuneate fasciculus at 16 months of age. Scale bars: 10 μm. (K and L) Distribution of corticospinal tract axons according to their section surface-area at eight (I) and 16 (J) months of age. Knockout mice displayed a lower percentage of large surface-area axons than control mice from eight months of age (n ≥ 5 slices/animal and n ≥ 10 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05 and ***p ≤ 0.001). (M) Distribution of cuneate fasciculus axons according to their section surface-area at 16 months of age (n ≥ 5 slices/animal and n ≥ 1 0 animals/genotype/age; Kruskal-Wallis test). (N) Number of dystrophic axons in a 10,000 μm2 area of the corticospinal tract at 4, 8, and 16 months of age (n ≥ 5 slices/animal and n ≥ 10 animals/genotype/age; Kruskal-Wallis test; **p ≤ 0.01 and ***p ≤ 0.001). (O) Number of dystrophic axons in a 10,000 μm2 area of the cuneate fasciculus at 4, 8, and 16 months of age (n ≥ 5 slices/animal and n ≥ 10 animals/genotype/age; Kruskal-Wallis test).

Fig. 6

Loss of spatacsin leads to spinal motor neuron degeneration and amyotrophy. (A and B) ChAT (red; spinal motor neuron marker) immunostaining of lumbar spinal cord slices of Spg11+/+ (A) and Spg11−/− (B) mice at 16 months of age. Scale bars: 50 μm. (C–F) Quantitative analysis of the total number (C) and cell body area (D–F) of motor neurons per ventral horn in the lumbar spinal cord of Spg11+/+, Spg11+/− and Spg11−/− mice at 4, 8, and 16 months of age. Knockout mice displayed fewer large spinal motor neurons than controls from eight months. (n ≥ 5 slices/animal and n ≥ 10 animals/genotype/age; Kruskal-Wallis test; **p ≤ 0.01 and ***p ≤ 0.001). (G–I) Neuromuscular junction labeling with α-bungarotoxin (Red, postsynaptic marker) and anti-neufilament plus anti-synaptophysin antibodies (green, presynaptic markers), representing a normal Pretzel-like plaque (G) and abnormal fragmented plaques (H and I). Scale bars: 10 μm. (J) Neuromuscular junction (NMJ) surface-area in the soleus of Spg11+/+, Spg11+/−, and Spg11−/− mice at eight and 16 months of age. (n ≥ 200 NMJ/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05; **p ≤ 0.01). The NMJ surface-area of Spg11−/− mice was significantly less than in control mice. (K) Percentage of fragmented NMJ in the soleus of Spg11+/+, Spg11+/−, and Spg11−/− at 8 and 16 months of age. (n ≥ 200 NMJ/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05). Spg11−/− mice exhibited a higher percentage of fragmented neuromuscular junctions than control mice. (L and M) Hematein/Eosin-stained transverse slices of the medial part of the Spg11+/+ (L) and Spg11−/− (M) soleus at 16 months of age. Scale bars: 400 μm. (N and O) Total number (N) and cross-section area (O) of soleus myofibers from Spg11+/+, Spg11+/−, and Spg11−/− mice at eight and 16 months of age (n ≥ 3 muscle slices/animal and n ≥ 5 animals/genotype/age; Kruskal-Wallis test; *p ≤ 0.05). (P) Grip test. Spg11−/− mice showed a significant loss of muscular strength in both forelimbs and hindlimbs from 12 months of age compared to control mice (n ≥ 12 animals/genotype/age; Kruskal-Wallis test; **p ≤ 0.01 and ***p ≤ 0.001).

Fig. 7

Accumulation of autofluorescent particles colocalised with Lamp1 in cortical motor neurons of Spg11−/− mice. (A) Lamp-1 (magenta; lysosome marker) immunostaining and autofluorescent material (green) in Spg11+/+ and Spg11−/− cortical motor neurons from six-week and eight-month-old animals. Confocal microscopy images show autofluorescent particles surrounded by Lamp-1-positive staining in Spg11−/− neurons. Scale bars: 10 μm. (B) p62 (magenta) immunostaining and autofluorescence (green) in Spg11+/+ and Spg11−/− cortical motor neurons at six weeks, eight months, and 16 months of age. Confocal microscopy images show autofluorescent particles colocalizing with p62-positive staining only in the cortical motor neurons of 16-month-old Spg11−/− mice. Scale bars: 10 μm. (C) Western blot showing levels of LC3I, LC3-II, and tubulin in extracts from Spg11+/+ and Spg11−/− mouse cortices taken from animals of different ages. Quantification of LC3-II band intensities normalized against tubulin levels. The graph shows the mean ± SEM values. n = 3 independent samples.

Fig. 8

Loss of spatacsin promotes the accumulation of lipids in lipofuscine-like structures in neurons. (A) Electron micrographs of cortical neurons from Spg11+/+ and Spg11−/− mice at two or 16 months. Red lines indicate the plasma membrane of neurons. Arrowheads indicate low-density structures consistent with lipid droplets. Scale bars: 5 μm (upper panels) and 1 μm (lower panels). (B) Electron micrographs of spinal cord motor neurons from two-month old Spg11+/+ and Spg11−/− mice. Red lines indicate the plasma membrane of motor neurons. White squares indicate the zone shown at higher magnification in the insets. Scale bars: 5 μm. (C) Cathepsin D immunoelectron microscopy revealed with diaminobenzidine (DAB), showing the presence of DAB precipitates in lipofuscin-like structures (red arrowheads). Note the presence of DAB precipitates in a lysosome (red arrow). Asterisks indicate low-density structures compatible with lipid droplets. Scale bar: 500 nm. (D) Electron microscopy images of cortical neurons in the brain of an SPG11 patient (duration of disease: 10 years; age at death: 32 years), showing the accumulation of lipofuscin (arrowheads). The patient had the typical clinical features of SPG11 and carried, in trans, the heterozygous mutations c.2358_2359delinsTT (p.Glu786_Gly787delinsAspfs*) in exon 13 and c.4868delT (p.Leu1623Tyrfs*17) in exon 28. Post-mortem delay was 48 h, explaining the presence of vacuoles in the tissue. Scale bar: 2 μm.

Fig. 9

Spatacsin loss prevents the egress of lipids from lysosomes in mouse embryonic fibroblasts. (A). Electron micrographs of fibroblasts obtained from Spg11+/+ or Spg11−/− embryos. The inset shows a higher magnification of the lysosomes indicated by the arrowheads. Note the presence of osmiophilic, undigested material in the lysosomes of Spg11−/− cells. (B). Quantification of lysosome density and the mean number of lysosomes containing undigested material per μm2. The graph shows the means ± SEM. n > 20 cells in two independent experiments. t-test; ***p ≤ 0.001. (C). Lamp1 Immunostaining (green) and Nile red staining (magenta) of primary cultures of mouse embryonic fibroblasts cultured for 0, 2, or 24 h in an amino acid-poor medium (HBSS). After 24 h in HBSS, the Nile red staining was concentrated in lipid droplets. Scale bar: 2 μm. (D). Quantification of the colocalization of Nile red staining with the lysosomal marker Lamp1 in cells cultured for 0, 2, or 24 h in an amino acid-poor medium (HBSS). The graph shows the means ± SEM. n > 50 cells in five independent experiments. Two-way ANOVA followed by the Holm-Sidak test; *p ≤ 0.05. (E). Quantification of the number and size of lipid droplets. Graphs show the means ± SEM (n > 50 cells in five independent experiments). Two-way ANOVA followed by the Holm-Sidak test; *p ≤ 0.05; ***p ≤ 0.001.

Fig. 10

Spatacsin loss promotes lipid accumulation in lysosomes in primary cultures of cortical neurons. (A). Electron micrographs of primary cultures of cortical neurons derived from Spg11+/+ or Spg11−/− embryos. The inset shows a higher magnification of the lysosomes indicated by arrowheads. Note the presence of undigested material in the lysosomes of Spg11−/− neurons. Also, note the large number of lysosomes containing undigested material, appearing in black, in Spg11−/− neurons. Scale bar: 2 μm. (B). Quantification of the mean number of lysosomes containing undigested material per μm2. The graph shows the means ± SEM (n > 20 cells). t-test; ***p ≤ 0.001. (C). Lamp 1 immunostaining (green) and Nile red staining (magenta) of primary cultures of cortical neurons. The arrowheads indicate lipid droplets appearing as very bright structures upon Nile red staining. (D). Quantification of the colocalization of Nile red staining with the lysosomal marker Lamp1. The graph shows the means ± SEM (n > 70 cells in four different experiments). t-test; *p ≤ 0.05. (E). Number of lipid droplets per neuron in primary cultures of cortical neurons. The graph shows the means ± SEM (n > 70 cells in four independent experiments). t-test; **p ≤ 0.005.