CRISPR Shows Preclinical Promise in Treating Alzheimer’s, Challenges Persist

DNA and genome editing_iStock, elenabs

Pictured: DNA and genome editing concept/iStock, elenabs

CRISPR-Cas9—the Nobel Prize–winning gene-editing system developed by Jennifer Doudna and Emmanuelle Charpentier—is on the cusp of breaking into clinical trials for Alzheimer’s disease, but experts note barriers that must be overcome and question whether scientists might be moving too quickly. 

In Amsterdam last month at the Alzheimer’s Association International Conference (AAIC), researchers from the University of California San Diego (UCSD) and Duke University School of Medicine presented two different approaches leveraging CRISPR to prevent and treat Alzheimer’s.

At UCSD, Brent Aulston’s team is using CRISPR to edit the amyloid precursor protein (APP), a transmembrane protein expressed in many tissues throughout the body but concentrated in the synapses of neurons.

Brent Aulston_UCSD
Brent Aulston

“We know from human populations that all the genetic causes of Alzheimer’s disease affect how APP is processed,” Aulston told BioSpace. “APP is essentially physically cut within the cell by one or two enzymes.”

The first of these enzymes is alpha secretase. When APP is cleaved by alpha-secretase, a protective molecule called secreted APP-alpha is released. “In healthy people, the majority of the APP is being cleaved to produce secreted APP-alpha,” Aulston said. However, in Alzheimer’s patients, APP is pushed toward the other cleavage pathway, mediated by beta-secretase. “This causes kind of the classical toxic fragments known as amyloid-beta peptides.” 

Using CRISPR tools, Aulston’s team is working to change the APP gene, “so that we get more alpha cuts and less of the beta cuts.” 

Aulston emphasized that the intention is not to knock out the entire APP protein. “APP is a very important protein in the brain. It has a lot of functions. If you take the APP protein out of a mouse, the mice are smaller, their brains are smaller, there’s more neuroinflammation [and] they’re cognitively compromised. We still want that protein expressed.”

So, the researchers designed their CRISPR to edit out only a small portion of APP, aiming to keep the resultant protein from coming into contact with beta-secretase. Testing out the approach in mice, the team showed that the animals had fewer amyloid-beta plaques and markers of brain inflammation but more of the neuroprotective APP.

“You keep all the physiological functions of APP . . . and you actually get an increase in secreted APP alpha,” Aulston said. The mice also showed improvement in behavioral and neural symptoms. “We believe this demonstrates that, in mice, our potential treatment strategy is both safe and efficacious,” Aulston said in the abstract presented at AAIC. “These results justify future studies aimed at getting APP CRISPR editing into human testing.”

But Rudolph “Rudy” Tanzi, a professor of neurology at Harvard Medical School and Massachusetts General Hospital who co-discovered the APP gene, told BioSpace he has concerns about using CRISPR to alter this protein. “For fun, we would say APP stands for ‘all-purpose protein’ because it does so many things,” he said. Among other roles, he noted that APP helps to stabilize and integrate synapses and helps blood to clot in the event of an injury.

Tanzi added that amyloid-beta acts as one of the brain’s main host defense mechanisms. In editing out the amyloid-beta domain, Tanzi said two things will be lost: an antimicrobial peptide property that protects against infection and a checks-and-balances system that quells synaptic excitation firing. “Without amyloid-beta, the cells will keep firing,” he said. Among other potential problems, overexcited neurons can trigger proteins in target muscle cells to become stressed, misfolded and non-functional.

Duke Targets APOE

On the opposite coast, researchers from Duke University School of Medicine are using CRISPR-Cas9 gene editing to downregulate the expression of the APOE ε4 gene, which is associated with an increased risk of Alzheimer’s. It is estimated that 15% to 25% of the population carry this allele, while up to 5% have two copies.

Ornit Chiba-Falek_Duke University
Ornit Chiba-Falek

Like Aulston, Ornit Chiba-Falek, professor in neurology, division chief of translational brain sciences at Duke and senior author of the study presented at AAIC, told BioSpace that the intention is not to entirely eliminate the APOE protein, which helps carry cholesterol and other types of fat in the bloodstream, but rather to fine-tune the levels of the different variants. APOE ε2, for example, is thought to provide some protection against Alzheimer’s, while APOE ε3, the most common variant, is believed to have no effect on the disease. The Duke team aims to specifically reduce the production of the protein encoded by APOE ε4.

The platform “demonstrated an efficient and precise editing of APOE ε4 and APOE expressions” in miniature brains derived from human induced pluripotent stem cell (hiPSC) from an Alzheimer’s patient and in humanized mouse models, the authors wrote in the abstract presented at AAIC. Importantly, they continued, the specific targeting of the ε4 allele “has resulted in no detectable editing of the ε3 allele in the isogeneic hiPSC-derived neurons.”

Chiba-Falek and Boris Kantor, associate research professor of neurobiology at Duke and the study's first author, co-founded a biotech company, CLAIRIgene, to further develop the platform, and Chiba-Falek said they are exploring collaboration opportunities with the pharmaceutical industry. Next up are IND-enabling studies, which she said will focus on safety, route of administration and durability.

Tough Translational Terrain

Having proven the APP-editing concept, Aulston said his team has climbed the first mountain. “Now, we’re at the base of the other mountain, which is the translational side,” he said. “As it turns out, the human brain is a little more difficult to treat than a small mouse.”

Gerold Schmitt-Ulms_University of Toronto
Gerold Schmitt-Ulms

Gerold Schmitt-Ulms, a professor at the University of Toronto studying Alzheimer’s, tauopathies and prion disorders, expressed his concerns about CRISPR’s translatability to neurological diseases. He told BioSpace the technology is “lagging behind” other gene therapy approaches in this space for two main reasons: immunogenicity and delivery challenges.

On the first point, Schmitt-Ulms said the adeno-associated virus (AAV) capsids used in the delivery of the CRISPR machinery will generate an immune response that can limit the effectiveness of the therapies. “Around 70% of us have had exposure to natural AAVs and therefore carry antibodies in our blood that would destroy AAV-based gene therapy vectors,” he said. In addition, “the body has never seen CRISPR enzymes. They’re like foreign agents, and when you express them, there will be immune responses raised to them.”

Most current gene therapy clinical trials for neurodegenerative diseases involve the delivery of a particular overexpressed protein, Schmitt-Ulms said. “These are proteins that the body knows. APOE ε2, for example, is known to the body and won’t raise an immune response.”

As for delivery, “To date, no one can deliver any of these gene therapies to every cell in the human brain,” Schmitt-Ulms said. “The best one can currently achieve is that a few percent of the brain cells take up the therapeutic virus. Often, that is not enough for the desired benefit to manifest.”

Given the challenges, Tanzi said he believes CRISPR should be used only in extreme cases, such as for people carrying two copies of APOE ε4.

“I think we’re moving a little too fast, honestly, into genomic editing,” he said. “As a geneticist, I kind of think about the genome the way the sailor thinks about the ocean. We respect it. We live with it. Once you think you know it, it can take you out.” 

Heather McKenzie is a senior editor at BioSpace. You can reach her at heather.mckenzie@biospace.com. Follow her on LinkedIn and Twitter @chicat08.

Back to news