Mutations in superoxide dismutase 1 (SOD1) cause paralysis in familial amyotrophic lateral sclerosis and promote its misfolding into neurotoxic aggregates. Previous studies have shown that mice expressing the ALS-causing G85R variant of SOD1 develop paralysis much faster after intraspinal injection of spinal homogenates from paralyzed G85R SOD1 mice. These findings, and other studies in cell models, established the prionoid templating properties of misfolded mutant SOD1. Previously, however, we noted that the widely used Gur1-G93A SOD1 mice, which express at high levels and develop paralysis by 6 months of age, were resistant to seeding by homogenates from paralyzed G93A mice. A line of G93A mice that express at very low levels (VLE-G93A) were responsive to seeding but at low efficiency. The poor susceptibility of G93A-SOD1 mice to seeding was not what we expected if prion-like propagation is essential to SOD1 ALS pathogenesis. In our prior, studies seeding homogenates from paralyzed G93A-SOD1 mice were injected into the spine of newborn mice, leading us to question whether older G93A SOD1 mice might be more susceptible to seeding. Here, we establish that adult VLE G93A SOD1 mice (up to 12 months of age) injected intrathecally with seeding homogenates containing misfolded G93A or G85R SOD1 developed accelerated motor neuron disease efficiently. Thus, we demonstrate that both the route and age of inoculation can influence the efficiency of SOD1 seeding to induce motor neuron disease in VLE G93A-SOD1 mice. These data, together with our earlier reports, suggest that prion-like templating contributes to disease progression in SOD1-ALS.

Study Design

The primary goal of the study was to determine whether older VLE G93A-SOD1 mice may be more susceptible to prion-like seeding by intrathecal administration of seeding homogenates. We injected VLE G93A mice with homogenates from paralyzed G93A SOD1 mice or pooled homogenates from paralyzed mice expressing G85R-SOD1 fused to yellow fluorescent protein (YFP) that had been made paralyzed by injection of homogenates from paralyzed G93A SOD1 mice (designated P2-G93A seeding pool). Initially, we challenged 12-month-old mice with both seeding homogenates and then subsequently determined whether younger VLE G93A-SOD1 mice were also susceptible, using the P2-G93A pool. Controls for these experiments included uninjected VLE G93A-SOD1 mice and mice injected with a pooled homogenate from young asymptomatic naïve 2-month-old Line 230 mice (Y230 pool). Pathological analyses of these animals produced image data for Figures 2 - 4 of the accompanying publication.

In prelude to the G93A mouse study, we first sought to compare the relative seeding efficacy of different injection routes using a highly permissive line of homozygous mice expressing G85R-SOD1:YFP (Line 230 model). The seeding-homogenates for these studies were either the control  Y230-pool or a pool of homogenates from seeded paralyzed G85R-SOD1:YFP mice. We compared the incubation time to paralysis following lumbar intraspinal (ISP) injection in P0 newborn mice to lumbar intrathecal injection in 2 month old Line 230 mice. Pathological analyses of these animals produced image data for Figure 1 of the accompanying publication.

            For most of these injections, we sought to inject mice from 2 to 3 litters. The homozygous 230 mice naturally produced larger cohorts. In the VLE G93A study, each cohort receiving the G93A or P2-G93A pool was between 8 and 10 transgene positive animals. Only one litter of VLE G93A mice was injected with the 230 pool and Y230 pool inoculum, producing smaller cohorts of transgene positive animals.

            The primary humane endpoint for all injections was overt paralysis of at least one hindlimb. In a subset of mice, other body condition factors such as hydration or muscle atrophy were considered in decisions to euthanize and harvest tissues from mice that were obviously weak but not yet paralyzed. Laboratory staff were not blind to treatment or genotype, but veterinary staff that assisted in determining endpoint were blind. Mice that were recommended for euthanasia due to fighting wounds or tumor development were excluded from the study. No mice that received seeding injections for data in Figure 1 were excluded from the study. For the data in Figure 2, there were mice that had to be excluded post-injection; Group 1 – 2 mice and Group 3 – 1 mouse.  

            Mice that were identified as distressed and recommended for euthanasia but did not otherwise exhibit signs of paralysis were excluded from statistical analyses of incubation periods. For data in Figure 2B there were the following exclusions: Group 1 – 1 animal, Group 2 – 2 animals, Group 3 – 1 animal, Group 5 – 2 animals, Group 7 – 3 animals.   

Neuropathology

Detailed methods for tissue collection and neuropathology are described in Ayers et al ²³. For imaging YFP fluorescence, paraffin-embedded tissues were cut at 5 μm thickness and then deparaffinized and mounted on coverslips with water; images were visualized by epifluorescence Olympus BX60 microscope. After imaging, the coverslips were removed, and the slides were stained by Campbell-Switzer silver stain as previously described ^23,34. Ubiquitin staining was performed as described previously ³⁹. Paraffin sections were deparaffinized, steamed in citrate buffer (10 mM citrate, pH 6.0, with 0.05% Tween-20) for 20 minutes. After rinsing the slides with tap water, the slides were blocked with 3% normal goat serum in PBS containing Tween 20, and then incubated overnight with ubiquitin antibody (mouse monoclonal MCA-Ubi-1, 1:1000; Encor Biotechnology, Gainesville, FL). After quenching with 0.03% hydrogen peroxide in PBS-T, sections were treated with a biotinylated anti-mouse secondary antibody (1:500; Vector Laboratories, Newark, CA), followed by ABC-HRP and DAB (KPL, Seracare, MD) detection. Slides were counterstained with hematoxylin. Images were acquired at 20× using an Aperio ScanScope XT (Leica Biosystems), processed in ImageScope, and edited in GIMP ((GNU Image Manipulation Program v3.0.4: gimp.org) for cropping and scale bar addition based on pixel calibration.