In Simple Terms
Some people have a change in one copy of a gene that helps the brain develop, leaving the other copy unable to do enough on its own. In lab-grown brain cells, researchers used a tool to make the working copy more active, and some features moved closer to those in healthy cells. The work is an early lab study, not a treatment for people.
When One Copy Isn’t Enough
Many genes come in two copies, and when one copy is impaired, a cell may make less of the gene’s product. In some cases, the healthy copy cannot produce enough on its own to maintain the gene’s normal activity. This condition is called haploinsufficiency, or gene-dosage insufficiency. The healthy copy is still present, but it is being asked to do nearly as much as both copies—and may not be able to keep up.
The researchers noted that newly arising, or de novo, mutations of this kind that have substantial effects account for about 15% of the genetic liability for autism spectrum disorder. That figure refers to a share of genetic susceptibility; it does not mean that 15% of autistic people have these same mutations, or that autism has a single cause. Autism is complex, and this study focused on a specific group of genetic changes that can reduce the activity of an important gene.
The team chose CHD8 because it helps regulate the activity of other genes involved in nervous-system development. SCN2A, meanwhile, plays a central role in the ability of nerve cells to send electrical signals. The genes do different jobs, but both are strongly linked to autism, and reduced activity in either can produce measurable effects in cell models.
The study was led by Daniel Geschwind, a physician-scientist at the David Geffen School of Medicine at the University of California, Los Angeles. George Chen, now at Children’s Hospital Los Angeles, was the first author. Researchers from the New York Genome Center and New York University also contributed. The team set out to answer a practical question: instead of trying to repair the impaired copy, could they carefully increase the activity of the one that remains functional?
Turning Genes On Without Cutting DNA
The CRISPR-based tool used in the study relies on a version of the Cas9 protein that has been disabled so it cannot cut DNA. The researchers attached it to an activating factor called p300, allowing them to guide the system to regulatory sites that help switch genes on. The aim was to increase gene expression—that is, to prompt the cell to make more of the gene’s product—without cutting the DNA sequence itself.
Rather than directing the tool to promoters near the beginning of each gene, the team focused on enhancers associated with gene activity in the developing brain. Enhancers are regulatory regions that can affect how strongly a gene is switched on, even though they are not part of the instructions for making a protein. The researchers identified enhancers linked to CHD8 and SCN2A, then tested the approach in excitatory neurons and forebrain organoids.
The choice of target reflected how sensitive these genes can be to dosage. Increasing their activity too much could itself cause problems, particularly when cells need gene expression to stay within a suitable range. The researchers reasoned that targeting regulatory regions linked to the brain might allow a more selective increase than activating a gene through its promoter. Even so, the strategy still needs careful measurement and monitoring to determine how much activity it adds.
The experiments used human pluripotent stem cells, which the team turned into excitatory neurons, as well as organoids that model parts of the forebrain. An organoid is a three-dimensional cell model grown in the lab that lets researchers study some aspects of development and organization. It is not a complete brain and cannot reproduce everything that happens in a person’s body. Neurons grown in two-dimensional cultures provide a different kind of measurement, and their results should not be confused with findings from organoids.
What Changed in the CHD8 and SCN2A Models?
In organoids carrying CHD8 mutations, the mutations were associated with greater-than-usual growth and increased multiplication of progenitor cells—cells that can later give rise to neurons. After the gene was activated, organoid size decreased in most of the treated cell lines. Their gene-expression patterns also shifted toward those observed in healthy models.
A smaller organoid is not, by itself, proof that the model became completely normal or that every effect of the mutation disappeared. It does, however, mark a change in an abnormal feature seen in the model, and it was accompanied by changes in gene activity. Considering both kinds of observations matters: organoid size alone cannot reveal what is happening inside its cells, while a change in gene expression alone does not prove that cell function has improved.
In CHD8-mutant neurons grown in two-dimensional cultures, indicators associated with neuronal maturation increased after activation. This is a different result from the change in organoid size and offers another angle on the findings: did the approach affect only the growth of a three-dimensional model, or did individual neurons also show changes in their characteristics? The experiment found improvement in those indicators, but that alone cannot show how the changes might translate to brain development in people.
For SCN2A models, the researchers focused on the ability of neurons to produce electrical signals. After gene activity was increased, electrical signaling in patient-derived cells moved closer to the level recorded in healthy reference cells. The assessment therefore went beyond asking whether the gene’s expression had risen: it also tested a functional feature directly related to SCN2A’s role in nerve signaling.
However, the conditions for producing neuronal activity were not identical across all cell lines. Compared with HUES66 cells, patient-derived neurons required a stronger electrical current to sustain repeated firing under the conditions tested by the team. The researchers interpreted this difference as a baseline variation in the cells’ intrinsic excitability—not, by itself, as evidence that the functional rescue had failed.
This finding is a reminder that cell models are not interchangeable copies of one another. Differences between two cell lines can affect the starting point against which a response is measured, even when researchers target the same gene in the same way. That is why results need to distinguish the effect of gene activation from the variation that was already present between cell lines, rather than treating every difference as proof of complete success or complete failure.
Promising Results, With Clear Limits
The team found that directing the activating tool to enhancers produced a sustained increase in CHD8 and SCN2A expression in mutant cells and cortical organoids. This increase was accompanied by gene-expression changes and changes in some features associated with the mutations. These findings support the idea that boosting the remaining functional copy may be worth exploring when reduced gene activity is part of the problem.
But “sustained” does not mean the effect lasted forever. The impact of activation declined over time, an important limitation if the eventual goal is to maintain increased activity for a long period. The researchers also described the increase as moderate, rather than a dramatic jump in expression. More work is needed to find out whether the effect can be made stronger and longer-lasting while keeping gene activity within an appropriate range.
The study also left open the question of when to activate the genes. Normal gene-expression patterns change during development, so finding that increased expression affected a cell model does not tell us whether starting the intervention after the expected peak in gene activity would produce the same result. Timing could matter, but its effects must be tested directly before practical conclusions can be drawn.
The approach also faces challenges in delivering the CRISPR components to the intended cells, improving the design of the tools, and considering alternative Cas enzymes. These are not minor technical details. Success in lab-grown cells does not show that the system can be delivered safely and efficiently to the brain, or that it would work at an appropriate level in living human tissue.
It is important, then, to distinguish a demonstration of a biological principle from the development of a treatment. The study did not test an intervention in people, nor did it show that activating CHD8 or SCN2A reduces autism-related traits in patients. Instead, it showed that changing the activity of two genes in human cell models shifted some measurements toward patterns closer to those in healthy reference models. A move toward clinical use would require further work on effectiveness, duration, timing, delivery, and control of the degree of activation.
The findings matter because they point to a different way of approaching some mutations. Rather than changing the DNA or replacing a gene, the tool tries to make better use of the copy that is still working. That idea could be valuable when gene activity can be increased in a controlled way, but it cannot be assumed to work for every gene or every mutation. Because genes differ in their roles and sensitivity to dosage, each case needs to be assessed on its own.
The Bottom Line
A study led by researchers at the University of California, Los Angeles, offers early evidence that activating enhancers linked to CHD8 and SCN2A can raise the activity of those genes in human brain-cell models. The change was accompanied by improvements in several genetic, developmental, and electrical measurements. These included smaller organoids in most CHD8-mutant models, higher levels of indicators of maturation in some neurons, and SCN2A patient-derived models whose excitability moved closer to the reference level.
The effect weakened over time, the increase in gene activity was moderate, and important questions remain about delivery, timing, and how tightly activation can be controlled. The results are not a ready-made autism treatment. They are a laboratory step toward testing a new way to compensate for reduced activity in certain genes. The crucial next question is whether the approach can be developed into something safe and effective beyond cell models.