Gene Editing Delays Disease Onset in ALS Model Mice

The infamous gene editor CRISPR can now count SOD1 among its successes. Mutant forms of the gene cause amyotrophic lateral sclerosis (ALS), and transgenic animals harboring such variants also succumb to an ALS-like disease. Researchers led by David Schaffer at the University of California, Berkeley, infected these mice with a virus loaded with components CRISPR needs to target mutant SOD1. As described December 20 in Science Advances, it suppressed the gene in a hefty proportion of motor neurons in the spinal cord, and the mice developed disease a month later than expected. However, the treatment was no magic bullet: It failed to target astrocytes, and did not slow progression of the disease once it started.

  • A virus loaded with CRISPR machinery dampened expression of human mutant SOD1 in mice.
  • Astrocytes were unaffected.
  • The treatment delayed the onset of motor symptoms and neurodegeneration, and extended lifespan.

While key improvements, such as broadening the number of targeted cell types, are in the works, this proof-of-concept study was the first to demonstrate that injecting the CRISPR apparatus into the blood can successfully target a disease gene in neurons, and prolong survival.

Fashioned from an ancient bacterial defense system, CRISPR (clustered regularly interspaced short palindromic repeats) scuppers gene expression. Complementary guide RNAs steer the Cas9 nuclease to the target DNA, whereupon it induces double-stranded breaks that lead to insertions or deletions (indels) that derail transcription of the gene (Sep 2014 news series). CRISPR can also be used to insert desired sequences and correct mutations, rather than simply halting transcription. Researchers have rapidly optimized the tool, improving its specificity and slimming down the size of its components. The technology has potential to thwart expression of pathological variants, or even correct them, making it an attractive strategy for familial diseases caused by known mutations.

SOD1 fits the bill. Antisense oligonucleotides have been used to dampen SOD1 expression in mouse models of ALS, but injecting them into the spinal cord has yet to trigger reductions in CSF SOD1 in ALS patients (May 2013 news). 

First author Thomas Gaj and colleagues wanted to test CRISPR. They turned to the G93A-SOD1 mouse, which expresses 25 copies of the mutated human SOD1. Rather than attempting to correct the G93A mutation, the researchers aimed to knock down expression of the gene. They designed multiple candidate guide RNAs that would bind human, but not mouse, SOD1. They also employed a cut-down version of the Cas9 nuclease from Staphylococcus aureus—SaCas9—which is small enough to squeeze into an AAV vector along with its guide RNA. They tested the combinations in mouse neuroblastoma cell lines transfected with human SOD1-G93A, and selected one that knocked down expression of the transgene by 92 percent.

The researchers next injected this vector intravenously into newborn SOD1-G93A mice. They also used a guide RNA specific for the mouse Rosa26 locus as a control. An AAV9 strain that had been optimized to cross the blood-brain barrier delivered SaCas9 into 74 percent of choline acetyltransferase (ChAT)-positive neurons in the ventral horn, according to immunohistochemistry. However, astrocytes in the same region expressed little, if any, SaCas9. Astrocytes also make SOD1, and it contributes to neurodegeneration in mice (Apr 2007 news; Ilieva et al., 2009). 

SOD1 Take Down. CRISPR (green) targeting hSOD1 (yellow) suppresses the protein (right), whereas one targeting a control locus (left) does not. [Courtesy of Gaj et al., Science Advances, 2017.]

Western blot analysis of spinal cord lysate revealed that compared with the control CRISPR, hSOD1-targeted CRISPR reduced expression of the protein three- and 2.5-fold in the lumbar and thoracic regions, respectively. Oddly, levels of the transgenic protein were not affected in the cervical region of the spinal cord, despite robust expression of SaCas9 in motor neurons there. The treatment did not affect expression of mouse SOD1. The researchers also sequenced hSOD1-G93A transgenes in lumbar and thoracic spinal cord tissue (which contains motor neurons in addition to non-transduced nerve and glial cells), and found a seven- and 14-fold rise in indels—CRISPR’s smoking gun. In agreement with the western blot data, the researchers did not detect a significant presence of indels in cervical spinal cord.

Would halting SOD1-G93A expression in a subset of motor neurons provide a therapeutic benefit to the mice? Yes, according to several measures. Using peak weight to mark disease onset, the researchers found that the treatment delayed disease by an average of 33 days, and prolonged survival by a similar amount. Notably, the treatment did nothing to slow disease progression: Following disease onset, animals treated with hSOD1-targeted CRISPR succumbed just as quickly as control animals. Staining spinal cord sections from end-stage mice revealed 50 percent more motor neurons in treated animals than in controls. However, astrocyte numbers appeared unaffected by the treatment, and many of the cells harbored inclusion bodies loaded with mutant SOD1.

Schaffer told Alzforum that the findings demonstrate CRISPR’s potential as a treatment for ALS, particularly in patients harboring an autosomal-dominant mutation in genes such as SOD1 or C9ORF72. However, he acknowledged room for optimization. Most importantly, researchers will need to design viral vectors that efficiently infect astrocytes. SOD1 expression in the glial cells reportedly exacerbates disease, and the lack of delivery to these cells likely explains why the treatment only delayed disease onset without slowing progression, Schaffer said (Oct 2011 news). Schaffer told Alzforum that his lab is currently developing a library of millions of AAVs with an assortment of random mutations, in the hopes of finding one with a penchant for both neurons and glia.

Robert Brown, University of Massachusetts, Amherst, pointed out that once a cell is infected with AAV, it will continue to express the transgenes for years. He thinks while this might be a desirable outcome in the case of RNA-based silencing strategies, which require indefinite silencing, it is both unnecessary and potentially hazardous in the case of CRISPR. Tweaking the virus to express CRISPR components transiently will be necessary, he said. Schaffer told Alzforum that his lab is working on generating such “short pulse” viruses.

Ultimately, using CRISPR to correct disease-causing mutations, rather than wiping out expression of genes entirely, will be the most favorable therapeutic strategy, said Brown. “We all agree that correcting somatic mutations is the Holy Grail, and CRISPR takes us in that direction.”—Jessica Shugart

Series Citations

  1. CRISPR: A New Tool For Gene Editors 12 Sep 2014

News Citations

  1. Paper Alert: Antisense Oligonucleotide Therapy Safe for ALS? 9 May 2013
  2. Glia—Absolving Neurons of Motor Neuron Disease 23 Apr 2007
  3. Believe It: Astrocytes Kill Neurons in New ALS Model 10 Oct 2011

Research Models Citations

  1. SOD1-G93A (hybrid) (G1H)

Paper Citations

  1. . Non-cell autonomous toxicity in neurodegenerative disorders: ALS and beyond. J Cell Biol. 2009 Dec 14;187(6):761-72. PubMed.

No Available Further Reading