Targeting misfolded SOD1 for ubiquitin-mediated degradation as a therapeutic strategy for ALS

Amyotrophic lateral sclerosis (ALS) is a devastating, progressive neurodegenerative disease characterised by the loss of motor neurons, resulting in muscle paralysis and ultimately death, usually within 3 years of diagnosis. The prevalence of ALS is 4-8 people per 100,000, with these numbers predicted to increase as the global population ages. In most cases of ALS (90-95%), there is no family history of the disease, and the patient is classified as having the sporadic form. In 5-10% of cases, however, a family history is present, inherited mutations are identified in a rapidly growing list of genes and patients are classified as having familial ALS. While the underlying cause of ALS remains elusive, the pathological hallmark for both sporadic and familial ALS is the aggregation of proteins into large, insoluble deposits in the cytoplasm of degenerating neurons. It is still unknown if the aggregates themselves are toxic to cells, however, the misfolding of protein is likely an initial pathological event. This observation provides a valuable basis in the search for diagnostic tools and treatments.

The first gene linked to ALS was superoxide dismutase 1 (SOD1), encoding an enzyme responsible for converting harmful superoxide radicals to molecular oxygen and hydrogen peroxide. There are currently over 200 ALS-associated mutations identified in SOD1. These mutations are thought to interfere with the folding of SOD1 protein into its correct three-dimensional conformation, resulting in the accumulation of the misfolded form into insoluble aggregates - a toxic gain-offunction. Currently the leading therapeutic strategy to reduce SOD1 protein levels is gene silencing through RNAi, however, this approach cannot differentiate between mutant forms of the protein and the correctly folded and functioning wild-type form. Therefore, the overarching aim of this study was to design a therapeutic tool that could specifically reduce the misfolded form of SOD1, preserving the wild-type protein. Modelled on proteolysis targeting chimeric small molecules (PROTACs), the BioPROTAC approach employs a single chain variable fragment derived from previously validated humanised antibodies that specifically recognise misfolded forms of SOD1. The scFv is fused by flexible linker to a modified E3 ligase to facilitate the ubiquitylation of mutant SOD1 and its subsequent proteasomal degradation.

Chapter 2 details the design and in vivo testing of a panel of seven anti-SOD1 scFvs, fused to a truncated form of the E3 ligase, C-terminal Hsc-70 interacting protein (CHIP). The seven BioPROTACs were assessed for their efficacy at reducing soluble and aggregated SOD1mut. Across multiple cell lines, the panel of BioPROTACs reduced the amount of soluble and insoluble SOD1mut-EGFP fluorescence compared to controls. Degradation via the proteasome was confirmed by an MG132 inhibitor assay. Microscopy showed the BioPROTACs co-localising with SOD1mut and binding was confirmed by co-immunoprecipitation.

In Chapter 3, the BioPROTAC was further optimised by evaluating a panel of eight different E3 ligases for their ability to degrade misfolded SOD1, using a similar suite of cell-based functional assays. Three of the eight ligase chimeras had soluble expression across the nucleus and cytoplasm, and could effectively reduce soluble and aggregated misfolded SOD1. The culmination of this work was the selection of the most effective E3 ligase combined with the most effective scFv to generate a lead BioPROTAC called “MisfoldUbL” as it specifically reduces misfolded SOD1 and acts via the ubiquitin ligase pathway.

In Chapter 4, the in vivo therapeutic efficacy of MisfoldUbL was investigated in an animal model. CRISPR/Cas9 technology was used for the targeted insertion of the MisfoldUbL transgene under the human pan-neuronal, synapsin 1 promoter. These mice were crossed with SOD1G93A mice and offspring were monitored from 50 days of age with weight, neurological score and motor function assessed until end-stage. In compound transgenic SOD1G93A/MisfoldUbL mice, MisfoldUbL attenuated the ALS phenotype, preventing weight loss, protecting motor function and delaying disease progression. In female mice, onset of disease was also delayed. Post-mortem analysis of an age-matched cohort at 90 days revealed an increase in the amount of soluble SOD1 and a concomitant decrease in the amount of insoluble SOD1 in the brains but not spinal cords of SOD1G93A/MisfoldUbL mice. The MisfoldUbL transgene also conferred protection on motor neurons in the ventral lumbar cord in this cohort.

These promising results provide proof-of-concept supporting the therapeutic potential of BioPROTAC-style protein molecules for the targeted degradation of misfolded SOD1, and warrant further investigations into delivery in a more clinically relevant context.