Subscribe to eTOC

At the Bench-Familial Alzheimer's Disease
CRISPR Corrects Mutation in Neurons Generated from Skin Cells of Patients with Familial Alzheimer's Disease


Using CRISPR gene editing, researchers were able to modify a mutant gene associated with familial Alzheimer's disease and observed normalization of amyloid beta and electrophysiological deficits.

An international group of researchers was able to correct a mutation associated with familial Alzheimer's disease (AD), and they observed, in vitro, the normalization of the cells' electrophysiological properties and production of amyloid-beta (Abeta).

The research team used fibroblasts from three sisters, two of whom carried a mutation in the presenilin 2 (PSEN2) gene. They converted them first into induced pluripotent stem cells (iPSCs), which are capable of being further transformed into any mature cell, and then into basal forebrain cholinergic neurons (BFCNs). BFCNs in the adult brain are among the first to develop pathological signs associated with AD. They then used a form of gene editing — CRISPR/Cas9 — to correct the mutation.

The electrophysiological abnormalities observed and corrected in BFCNs associated with the PSEN2 mutation may be a contributor to Alzheimer's pathology, independent of the role of Abeta, the researchers wrote in the study, published online October 27 in Acta Neuropathologica Communications. Previous studies, they noted, indicate that both PSEN1 and PSEN2 can function as ion channels that might underlie the electrical changes observed in the new study.

The researchers plan to implant the corrected BFCNs into the brains of mice with the PSEN2 mutation to see if they will result in behavioral improvements in the mice. It is unknown if such a method will succeed, but at the very least, they said, the paper demonstrates an improved in vitro protocol to generate human BFCNs from iPSCs, cutting in half the time necessary to do so.

The refined protocol was overseen by Scott Noggle, PhD, senior vice president for research at the New York Stem Cell Research Foundation, and one of the two senior authors of the study.

The other senior author, Samuel E. Gandy, MD, PhD, Mount Sinai Professor of Alzheimer's Disease Research and professor of neurology and psychiatry at Mount Sinai School of Medicine in New York, told Neurology Today: “Now that we have these cells whose integrity we believe to be crucial to normal short-term memory function, we're going to try with some mouse models, maybe also with aged dogs with canine cognitive dysfunction, to see whether implanting these corrected BFCNs will have any benefit in learning behavior.”

“It is conceivable, although I'm not sure it's likely, that autologously generated BFCNs implanted into the hippocampus may slow decline or help sustain memory function longer,” Dr. Gandy said. “I don't at all think we're talking about a cure, but we might be able to extend independence.”


An earlier collaborative study from Dr. Noggle and colleagues led to the publication of a paper published in 2014 describing the molecular profile of iPSC-derived neural progenitors taken from subjects with a mutation in the PSEN1 gene. That paper showed unexpectedly increased expression of the pro-inflammatory gene nucleotide-binding oligomerization domain, leucine rich repeat and pyrin domain containing 2 (NLRP2).

“This led us to examine NLRP2 expression in our PSEN2 mutant lines and employ CRISPR/Cas9 to investigate if activation of the inflammasone was tightly linked to the pathogenic mutation in PSEN2,” they wrote in the new paper.

While the research team did not find altered expression of NLRP2 in gene-corrected PSEN2 lines, they did observe differences in the excitability of the BFCNs that were related to the mutation and could be corrected with gene editing.


DR. SAMUEL E. GANDY: “It is conceivable, although Im not sure its likely, that autologously generated BFCNs implanted into the hippocampus may slow decline or help sustain memory function longer. I dont at all think were talking about a cure, but we might be able to extend independence.”

In the new study, the research took fibroblasts from three sisters. Two were positive for the PSEN2 mutation and were experiencing cognitive decline; the third, who served as a control, had the wild-type form of the gene and was cognitively healthy. The mutation, associated since 1995 with autosomal dominant early onset familial Alzheimer's disease, has previously been shown to raise the Abeta 42-43/40 ratio, thereby promoting assembly of Abeta oligomers and fibrils.

The researchers transformed the three sets of fibroblasts into iPSC lines. Maturation of stem cells into various neuronal subtypes involves the action of transcription factors that control expression of region-specific and/or neuronal subtype-specific proteins. As the brain develops, BFCNs become differentiated from a small hill of tissue known as the medial ganglionic eminence (MGE).

In the current paper, Drs. Gandy and Noggle found that the new protocol enables them to accelerate the appearance of both MGE-specific and BFCN-specific proteins. They were also able to record mature action potentials in neurons as early as day 38 in culture.

“Action potentials are very sensitive evidence of electrical firing across certain synapses,” Dr. Gandy explained. “The more robust the action potential signal, the more efficiently the cells can be used for highly sensitive, high-throughput screening for compounds that modulate BFCN function.”

As expected, the BFCNs with the PSEN2 mutation displayed the typically pathological changes in Abeta secretion, including a two-fold increase in the Abeta 42/40 ratio, a 50 percent increase in the amount of secreted Abeta-40, and a 2.5-fold increase in Abeta 42 species in the conditioned media from PSEN2 neural progenitors.

Unexpectedly, however, they observed electrophysiological changes in the mutant BFCNs. Not only were fewer maximum number of spikes generated in response to a depolarizing current, but the height of the action potential was also decreased. No prior study had ever described the electrophysiological properties of PSEN2 mutant BFCNs, the study noted.

Upon correcting the mutation with CRISPR/Cas9, not only did the increased Abeta 42/40 normalize, but so too did the electrophysiological deficits.


In interviews with Neurology Today, several neuroscientists praised the technical innovations described in the paper. They expressed doubts, however, that implanting CRISPR-corrected BFCNs in the basal forebrains of mice or people could have any beneficial effect.

“I am excited to look into the basal forebrain cholinergic protocol that they describe,” said Tracy Young-Pearse, PhD, an associate professor of neurology at Brigham and Women's Hospital and the Harvard Stem Cell Institute. But, she added, “I don't think Alzheimer's disease is a good candidate for cell replacement. Unlike in Parkinson's disease, where a small number of localized cells are lost, there is widespread loss of neurons throughout the brain in AD.”

She also expressed some skepticism that the electrophysiological deficits observed by Dr. Gandy's group are a direct cause of AD pathology, independent of the role of Abeta.

“It's hard to disentangle the Abeta phenotypes from the electrophysiological changes observed; the two are intimately linked,” Dr. Young-Pearse said. “We know that Abeta 42 affects synapse number and electrophysiological activity, and that synapse loss is one of the strongest correlates of clinical impairment in AD. So, I do think that electrophysiology is very important in the disease, but that defects are likely mediated through Abeta and tau.”

In a paper published in Stem Cell Reports within days of Dr. Gandy's, she examined neurons derived from iPSCs of patients with an amyloid precursor protein (APP) mutation, and found fate-specific effects of the mutation, depending on which types of neurons they became. Her group also corrected the mutation with CRISPR.

Dave R. Schubert, PhD, professor and laboratory head of the cellular neurobiology laboratory at the Salk Institute for Biomedical Sciences, said he believes that the electrophysiological defects associated with familial Alzheimer's disease mutations may play an important role in AD clinical pathology. But the findings needed to be confirmed in more than just two sisters.

More importantly, he said, “Is this finding relevant to most patients without a familial form of AD? I don't think so. One could make the argument that basal forebrain cells are the first to have problems in sporadic AD. That could be. But then another question is whether you could use this technique to make these BFCNs from patients' own skin cells and then somehow inject them into patients and make them well again. I don't think that has any chance at all, at least not in the near future.”

He described as highly unlikely that the implanted neurons would integrate into the basal forebrain, sprouting the myriad axons and connections of healthy BFCNs.

Lawrence Goldstein, PhD, director of the University of California, San Diego Stem Cell Program, said the hypothesis that the electrophysiological deficits observed in Dr. Gandy's paper could be an independent, correctable cause of AD is reasonable but as yet unproved.

“The burden is on them to extend the study and see where it leads,” he told Neurology Today. “The basal forebrain neurons project all over the cortex with very complex projection patterns. Nobody knows if you can reproduce that by transplantation. I think that will be really hard.”

Dr. Gandy said he agreed with the view that treatment of AD by implanting replacement neurons may be unlikely. “I don't disagree that it's far-fetched,” he said, “but it is at least worth trying in some animal models to see if implanting BFCNs helps the learning behaviors of a mouse.”


The correction of a mutation associated with familial Alzheimer's disease would have been impossible just five years ago, when a remarkable new method for gene editing, called CRISPR/Cas9, was first described in the journal Science.

The history of the method, however, dates back much earlier, to a chance observation of an E. coli gene. Writing in the December 1987 edition of the Journal of Bacteriology, Japanese researcher Yoshizumi Ishino, PhD, described what he called an “unusual structure” of 29 nucleotides that repeated, letter for letter, five times. Nothing like it had ever been seen before. “The biological significance of these sequences is not known,” he wrote.

Other researchers soon began spotting other similarly repeated sequences, first in the tuberculosis bacterium, then in prokaryotic organisms known as archaea. In 2001, these mysterious repeats were named “clustered regularly interspaced short palindromic repeats,” abbreviated as CRISPR. (They are palindromic because the sequence is identical whether read from front to back or from back to front.)

Soon, researchers began to figure out that the repeats represented a previously unrecognized immune defense system that bacteria and archaea employed to remember and protect against the genetic signature of viruses they had previously encountered. The CRISPR sections were like copy-and-paste devices that could surround any snippet of the foreign DNA they chose. Along with an enzyme known as Cas9, they could cleave, copy, and paste the viral snippet into the bacterial DNA, so that it could be recognized in case of future attacks by the same virus.

Cribbing the technique from nature, the CRISPR/Cas9 system for inserting or removing genes was first described in a seminal paper published in the journal Science in August of 2012, led by molecular biologist Jennifer Doudna, PhD, of the University of California, Berkeley, and Emmanuelle Charpentier, PhD, now director of the Max Planck Institute for Infection Biology in Berlin.

Drs. Doudna and Charpentier described the use of CRISPR/Cas9 only in bacteria, and just six months later, a group led by Feng Zhang, PhD, at the Broad Institute in Cambridge, MA, reported its use in human cells, again in the journal Science.

Even as a legal battle has since ensued over which group deserves the patent rights, CRISPR/Cas9 has become the go-to method for a wide variety of gene-editing experiments, primarily because of the speed and simplicity of the method compared to prior gene-editing methods.

Just last year, in the journal Nature, the first use of the method involving an Alzheimer's-related gene was described. Both the amyloid precursor protein and presenilin 1 mutations were inserted into human induced pluripotent stem cells, which were then grown into cortical neurons that displayed disease-associated defects in vitro.

Since then, at least four other papers have now described the use of CRISPR/Cas9 to explore and correct mutations associated with familial Alzheimer's disease. How far the powerful new method for editing genes will go in the study and treatment of AD and other disorders is anybody's guess.

—Dan Hurley