Blood-spinal cord barrier disruption contributes to early motor-neuron degeneration in ALS-model mice

Abstract

Humans with ALS and transgenic rodents expressing ALS-associated superoxide dismutase (SOD1) mutations develop spontaneous blood-spinal cord barrier (BSCB) breakdown, causing microvascular spinal-cord lesions. The role of BSCB breakdown in ALS disease pathogenesis in humans and mice remains, however, unclear, although chronic blood-brain barrier opening has been shown to facilitate accumulation of toxic blood-derived products in the central nervous system, resulting in secondary neurodegenerative changes. By repairing the BSCB and/or removing the BSCB-derived injurious stimuli, we now identify that accumulation of blood-derived neurotoxic hemoglobin and iron in the spinal cord leads to early motor-neuron degeneration in SOD1(G93A) mice at least in part through iron-dependent oxidant stress. Using spontaneous or warfarin-accelerated microvascular lesions, motor-neuron dysfunction and injury were found to be proportional to the degree of BSCB disruption at early disease stages in SOD1(G93A) mice. Early treatment with an activated protein C analog restored BSCB integrity that developed from spontaneous or warfarin-accelerated microvascular lesions in SOD1(G93A) mice and eliminated neurotoxic hemoglobin and iron deposits. Restoration of BSCB integrity delayed onset of motor-neuron impairment and degeneration. Early chelation of blood-derived iron and antioxidant treatment mitigated early motor-neuronal injury. Our data suggest that BSCB breakdown contributes to early motor-neuron degeneration in ALS mice and that restoring BSCB integrity during an early disease phase retards the disease process.

Keywords: amyotrophic lateral sclerosis; neurodegeneration.

Fig. 1.

Early microvascular lesions in SOD1^G93A mice: acceleration by Warfarin and protection by 5A-APC. (A–E) IgG (A and B), hemoglobin (C), hemosiderin (D), and free iron (E) in the lumbar anterior horn of SOD1^G93A mice receiving from 35 to 95 d saline, 0.3–0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin (W), and/or 100 µg⋅kg⁻¹⋅d⁻¹ 5A-APC (APC), 100 mg⋅kg⁻¹⋅d⁻¹ deferoxamine mesylate (DFX), or 50 mg⋅kg⁻¹⋅d⁻¹ glutathione monoethyl ester (GSHE) with either saline or warfarin (0.4 mg⋅kg⁻¹⋅d⁻¹). B6SJL littermates and SOD1^WT mice received saline or warfarin (0.6 mg⋅kg⁻¹⋅d⁻¹). (F, Upper) Prussian blue-positive hemosiderin deposits (blue) and collagen IV-positive capillaries (brown) in lumbar cord anterior horn of 95-d-old B6SJL littermate and saline-treated SOD1^G93A mouse. (Lower) Perl’s-positive hemosiderin deposits (red) in 95-d-old SOD1^G93A mouse receiving saline or 0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin. Broken line denotes boundary of the anterior horn. (G) Collagen IV-positive capillaries (green) and CD235a-positive erythrocytes (red) in the lumbar anterior horn of 95-d-old B6SJL littermate receiving 0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin and SOD1^G93A mice receiving saline, 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin or 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin and 5A-APC (100 μg⋅kg⁻¹⋅d⁻¹). (H) Percentage of capillary-associated and noncapillary hemosiderin deposits in 95-d-old SOD1^G93A mice receiving saline or warfarin (0.6 mg⋅kg⁻¹⋅d⁻¹); n = 4–5 mice per group. (I) Positive correlation between the number of lumbar hemosiderin deposits and the degree of anticoagulation determined as international normalized ratio (INR) in 95-d-old SOD1^G93A mice receiving saline or warfarin (0.3–0.6 mg⋅kg⁻¹⋅d⁻¹). Individual data points from four to five animals per group; r, Pearson’s correlation. In B–E, mean ± SEM, n = 3–5 mice per group. *P < 0.05; NS, nonsignificant. In B–E, APC, DFX, and GSHE treatments were compared with the respective saline or warfarin treatments as indicated by broken lines.

Fig. 2.

Onset of motor impairment and prevention in SOD1^G93A mice with spontaneous and accelerated microvascular lesions. (A and B) Cumulative probability (A) and mean age (B) of motor impairment in SOD1^G93A mice treated with saline (n = 14) or 0.3 (n = 21), 0.4 (n = 15), and 0.6 (n = 16) mg⋅kg⁻¹⋅d⁻¹ warfarin (W) from day 35 postnatal. Acceleration of motor symptoms in days relative to saline is provided above each group. Values in B, mean ± SEM, *P < 0.05. Nontransgenic B6SJL littermates (n = 14) received 0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin. (C) Negative correlation between onset of motor impairment and number of lumbar hemosiderin deposits in SOD1^G93A mice treated with saline or warfarin (0.3–0.6 mg⋅kg⁻¹⋅d⁻¹). Individual data points are from four to five mice per group; r, Pearson’s correlation. (D and E) Cumulative probability (D) and mean age (E) of motor impairment in SOD1^G93A mice treated daily with saline (n = 14), 100 µg/kg 5A-APC (n = 15), 100 mg/kg DFX (n = 14), or 50 mg/kg GSHE (n = 14) from day 35 postnatal. Delay of motor symptoms in days relative to saline is provided above each group. Values in E, mean ± SEM, *P < 0.05 compared with saline. Nontransgenic B6SJL littermates (n = 14 per group) received APC, DFX, or GSHE. (F and G) Cumulative probability (F) and mean age (G) of motor impairment in SOD1^G93A mice treated with 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin (n = 15) and 100 µg⋅kg⁻¹⋅d⁻¹ 5A-APC (n = 14), 100 mg⋅kg⁻¹⋅d⁻¹ DFX (n = 14), or 50 mg⋅kg⁻¹⋅d⁻¹ GSHE (n = 14) from day 35 postnatal. Delay of motor symptoms in days relative to 0.4 mg/kg warfarin is provided above each group. Values in G, mean ± SEM, *P < 0.05 compared with 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin.

Fig. 3.

Early motor-neuron degenerative changes and prevention in SOD1^G93A mice with spontaneous and accelerated microvascular lesions. (A–C) ChAT-positive motor neurons (magenta) and SMI-311–positive neurites (green) (A) and quantification of motor neurons (B) and neuritic density (C) in lumbar spinal cord of 95-d-old nontransgenic littermates and SOD1^G93A mice treated with saline, 0.3–0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin (W), or 100 µg⋅kg⁻¹⋅d⁻¹ 5A-APC, 100 mg⋅kg⁻¹⋅d⁻¹ DFX, or 50 mg⋅kg⁻¹⋅d⁻¹ GSHE with saline or 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin. (D) Negative correlation between SMI-311–positive neurites and lumbar hemosiderin deposits of 95-d-old individual SOD1^G93A mice treated with saline or 0.3–0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin. r, Pearson’s correlation; n = 4–5 animals per group. (E) Ubiquitin-positive accumulates (green) in motor neurons (red, visualized with SMI-311) in the lumbar anterior horn in 95-d-old SOD1^G93A mice treated with saline or 0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin. B6SJL, a nontransgenic littermate control. (F) Quantification of ubiquitin accumulation in motor neurons in mice from B. (G) Positive correlation between motor-neuron ubiquitin accumulation and the number of lumbar hemosiderin deposits in SOD1^G93A mice treated with saline or 0.3–0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin. Each point is an individual data point; r, Pearson’s correlation; n = 3–5 mice per group. In B, C, and F, values are mean ± SEM; n = 3–5 mice per group. *P < 0.05; NS, nonsignificant. In B, C, and F, APC, DFX, and GSHE treatments were compared with the respective saline and warfarin treatments as indicated by broken lines.

Fig. 4.

Early oxidant stress and prevention in SOD1^G93A mice with spontaneous and accelerated microvascular lesions. (A) Oxidized protein carbonyls in lumbar cords of 95-d-old nontransgenic B6SJL controls and SOD1^G93A mice treated with saline, 0.3–0.6 mg⋅kg⁻¹⋅d⁻¹ warfarin (W), or 100 µg⋅kg⁻¹⋅d⁻¹ 5A-APC (APC), 100 mg⋅kg⁻¹⋅d⁻¹ DFX, or 50 mg⋅kg⁻¹⋅d⁻¹ GSHE with saline or 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin. (B) Positive correlation between free iron and protein carbonyls in mice from A. Each point is an individual data point; r, Pearson’s correlation; n = 3–4 mice per group. (C–E) Representative immunoblotting of human oxidized SOD1 detergent-insoluble species (Left) and detergent-insoluble SOD1 aggregates (Right) in lumbar cord of 95-d-old SOD1^G93A mice treated with saline, 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin, or 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin with 100 µg⋅kg⁻¹⋅d⁻¹ 5A-APC (C), 100 mg⋅kg⁻¹⋅d⁻¹ DFX (D), or 50 mg⋅kg⁻¹⋅d⁻¹ GSHE (E). (F) ChAT-positive motor neurons (red) and 3-nitrotyrosine (3-NT)-positive signal (green) in lumbar spinal cord of 95-d-old SOD1^G93A mice treated with saline, 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin or 0.4 mg⋅kg⁻¹⋅d⁻¹ warfarin and 100 μg⋅kg⁻¹⋅d⁻¹ 5A-APC. (G) Quantification of 3-NT–positive signal in mice from A. In A and G, mean ± SEM; n = 3–5 mice per group. *P < 0.05; NS, nonsignificant.

Fig. 5.

A schematic illustrating how early blood–spinal cord barrier (BSCB) breakdown initiates motor-neuron injury and how BSCB-directed treatments blocking different steps in the BSCB pathogenic cascade can prevent early motor-neuron injury.