A common assumption about developmental brain conditions is that treatment must happen very early, before the brain finishes its major stages of development.
A new laboratory study challenges that idea by showing that a disrupted brain signal could still be improved in adult mice.
The research comes from the Institute for Basic Science in South Korea and was led by neuroscientist Eunjoon Kim. It appeared in the peer-reviewed journal Nature Communications on May 29, 2026.
At the center of the study is a communication system used by brain cells. Brain cells pass messages across tiny junctions called synapses, and many different receptors help receive those messages.
One of them is the NMDA receptor. It helps brain connections strengthen or weaken as we learn, form memories and adapt to new experiences.
Scientists have found unusually low NMDA receptor activity in several experimental models of brain disorders. These include models related to autism, schizophrenia and some forms of intellectual disability.
That does not mean low NMDA activity explains all autism. Autism spectrum disorder includes many different developmental pathways, genetic influences and combinations of traits, so researchers increasingly look for treatments aimed at particular biological subgroups.
For an NMDA receptor to work, it needs glutamate as well as a helper molecule such as glycine. Glycine therefore offers researchers one possible way to adjust the strength of this signaling system.
Past attempts have tried to raise glycine by blocking a protein called GlyT1, which normally removes glycine from spaces around cells. The problem is that GlyT1 is present in many parts of the nervous system, including areas that help control essential functions.
That broad distribution can make the target difficult to use precisely. The Institute for Basic Science team instead investigated another transporter, Slc6a20a, whose human counterpart is called SLC6A20.
Slc6a20a also moves glycine, but its distribution in the brain makes it an interesting alternative. It is found in areas such as the cortex and hippocampus, which are heavily involved in thinking, memory and complex behavior.
The researchers tried reducing Slc6a20a rather than adding glycine directly. If less glycine is removed, more may remain available to help NMDA receptors work.
They used a technology called an antisense oligonucleotide, or ASO. An ASO is a short, specially designed piece of genetic material that can reduce the amount of a chosen protein made by cells.
Similar types of genetic medicines are already used for some human diseases, which makes ASOs an important area of medical research. However, every new ASO still requires extensive testing for its own safety and effectiveness.
The team first studied mice with changes in Shank2 and Shank3. These genes help organize proteins at synapses, allowing brain cells to send and receive signals efficiently.
Changes in the human SHANK2 and SHANK3 genes have been linked with autism and other developmental conditions. SHANK3 changes are also a major cause of Phelan-McDermid syndrome.
In the mouse models, NMDA receptor signaling was weaker than normal. After the animals received the Slc6a20a-targeting ASO, this signaling improved.
The researchers also reported improvements in several behaviors measured in the mice. Depending on the genetic model, these included social interaction, communication-related behavior and excessive repetitive grooming.
The results differed somewhat between the Shank2 and Shank3 models. That detail matters because it shows that even when two conditions share a signaling problem, they may not respond in exactly the same way.
Perhaps the most thought-provoking part of the study was the age of the animals. The researchers were able to produce improvements in adult mice rather than treating only very young animals.
This raises the possibility that some changes in brain function are not permanently fixed after development. If the same principle eventually applies to humans, certain biological problems might remain treatable later in life.
The researchers also looked closely at what happened inside brain cells after treatment. Instead of finding sweeping changes in how much protein the cells contained, they found corrections in the way many proteins were chemically regulated.
Cells often control proteins by attaching small chemical groups to them. One important process is phosphorylation, which can change whether a protein is active, where it goes inside a cell or how it interacts with other proteins.
The ASO shifted abnormal phosphorylation patterns involving synapses and NMDA receptor pathways toward a healthier state. This suggests that the treatment may work by improving how an existing network operates rather than rebuilding the brain from scratch.
The team then tested the same basic strategy using human cells. They created cortical organoids carrying SHANK2 or SHANK3 mutations using CRISPR gene editing.
Organoids are sometimes described as miniature organs, but that description can be misleading. Brain organoids are simplified groups of cells grown in a dish and cannot reproduce consciousness, behavior or the full structure of a human brain.
Even with those limitations, they are useful because researchers can study human nerve cells carrying specific genetic changes. In this study, the organoids showed weakened NMDA receptor activity.
When the scientists reduced human SLC6A20 with an ASO, NMDA receptor function moved closer to normal. This provided an important bridge between the mouse findings and human biology.
A single administration also continued to show effects in mice for at least eight weeks, and the researchers reported no detectable adverse effects during that period. Longer and more detailed safety studies would still be necessary before human testing.
The study has several strengths. It used more than one autism-related genetic model, examined both behavior and brain signaling, investigated molecular changes, and reproduced a key effect in human-derived organoids.
But its limitations are equally important. Mouse social behavior is not human autism, organoids are incomplete models, and the study cannot tell us whether changing SLC6A20 would improve meaningful outcomes in autistic people.
It also remains unclear who might benefit most. Autism is not a single molecular disorder, so people without NMDA receptor underactivity may receive little benefit or could respond differently.
The best interpretation is therefore not that scientists have discovered an autism treatment ready for patients. Rather, they have identified a specific biological target that successfully corrected a defined signaling problem across several laboratory models.
If future studies confirm its safety and identify the right patients, SLC6A20 inhibition could become one route toward more personalized treatment for conditions involving weak NMDA receptor signaling. For now, the work is a promising proof of concept and an intriguing sign that some developmental brain changes may remain more flexible in adulthood than previously assumed.
If you care about health, please read studies that vegetarian diet may increase your depression risk, and Vitamin D could help reduce depression symptoms.
For more health information, please see recent studies about why pizza is a very addictive food, and MIND diet could improve cognitive health in older people.
Source: Institute for Basic Science.
A common assumption about developmental brain conditions is that treatment must happen very early, before the brain finishes its major stages of development.
A new laboratory study challenges that idea by showing that a disrupted brain signal could still be improved in adult mice.
The research comes from the Institute for Basic Science in South Korea and was led by neuroscientist Eunjoon Kim. It appeared in the peer-reviewed journal Nature Communications on May 29, 2026.
At the center of the study is a communication system used by brain cells. Brain cells pass messages across tiny junctions called synapses, and many different receptors help receive those messages.
One of them is the NMDA receptor. It helps brain connections strengthen or weaken as we learn, form memories and adapt to new experiences.
Scientists have found unusually low NMDA receptor activity in several experimental models of brain disorders. These include models related to autism, schizophrenia and some forms of intellectual disability.
That does not mean low NMDA activity explains all autism. Autism spectrum disorder includes many different developmental pathways, genetic influences and combinations of traits, so researchers increasingly look for treatments aimed at particular biological subgroups.
For an NMDA receptor to work, it needs glutamate as well as a helper molecule such as glycine. Glycine therefore offers researchers one possible way to adjust the strength of this signaling system.
Past attempts have tried to raise glycine by blocking a protein called GlyT1, which normally removes glycine from spaces around cells. The problem is that GlyT1 is present in many parts of the nervous system, including areas that help control essential functions.
That broad distribution can make the target difficult to use precisely. The Institute for Basic Science team instead investigated another transporter, Slc6a20a, whose human counterpart is called SLC6A20.
Slc6a20a also moves glycine, but its distribution in the brain makes it an interesting alternative. It is found in areas such as the cortex and hippocampus, which are heavily involved in thinking, memory and complex behavior.
The researchers tried reducing Slc6a20a rather than adding glycine directly. If less glycine is removed, more may remain available to help NMDA receptors work.
They used a technology called an antisense oligonucleotide, or ASO. An ASO is a short, specially designed piece of genetic material that can reduce the amount of a chosen protein made by cells.
Similar types of genetic medicines are already used for some human diseases, which makes ASOs an important area of medical research. However, every new ASO still requires extensive testing for its own safety and effectiveness.
The team first studied mice with changes in Shank2 and Shank3. These genes help organize proteins at synapses, allowing brain cells to send and receive signals efficiently.
Changes in the human SHANK2 and SHANK3 genes have been linked with autism and other developmental conditions. SHANK3 changes are also a major cause of Phelan-McDermid syndrome.
In the mouse models, NMDA receptor signaling was weaker than normal. After the animals received the Slc6a20a-targeting ASO, this signaling improved.
The researchers also reported improvements in several behaviors measured in the mice. Depending on the genetic model, these included social interaction, communication-related behavior and excessive repetitive grooming.
The results differed somewhat between the Shank2 and Shank3 models. That detail matters because it shows that even when two conditions share a signaling problem, they may not respond in exactly the same way.
Perhaps the most thought-provoking part of the study was the age of the animals. The researchers were able to produce improvements in adult mice rather than treating only very young animals.
This raises the possibility that some changes in brain function are not permanently fixed after development. If the same principle eventually applies to humans, certain biological problems might remain treatable later in life.
The researchers also looked closely at what happened inside brain cells after treatment. Instead of finding sweeping changes in how much protein the cells contained, they found corrections in the way many proteins were chemically regulated.
Cells often control proteins by attaching small chemical groups to them. One important process is phosphorylation, which can change whether a protein is active, where it goes inside a cell or how it interacts with other proteins.
The ASO shifted abnormal phosphorylation patterns involving synapses and NMDA receptor pathways toward a healthier state. This suggests that the treatment may work by improving how an existing network operates rather than rebuilding the brain from scratch.
The team then tested the same basic strategy using human cells. They created cortical organoids carrying SHANK2 or SHANK3 mutations using CRISPR gene editing.
Organoids are sometimes described as miniature organs, but that description can be misleading. Brain organoids are simplified groups of cells grown in a dish and cannot reproduce consciousness, behavior or the full structure of a human brain.
Even with those limitations, they are useful because researchers can study human nerve cells carrying specific genetic changes. In this study, the organoids showed weakened NMDA receptor activity.
When the scientists reduced human SLC6A20 with an ASO, NMDA receptor function moved closer to normal. This provided an important bridge between the mouse findings and human biology.
A single administration also continued to show effects in mice for at least eight weeks, and the researchers reported no detectable adverse effects during that period. Longer and more detailed safety studies would still be necessary before human testing.
The study has several strengths. It used more than one autism-related genetic model, examined both behavior and brain signaling, investigated molecular changes, and reproduced a key effect in human-derived organoids.
But its limitations are equally important. Mouse social behavior is not human autism, organoids are incomplete models, and the study cannot tell us whether changing SLC6A20 would improve meaningful outcomes in autistic people.
It also remains unclear who might benefit most. Autism is not a single molecular disorder, so people without NMDA receptor underactivity may receive little benefit or could respond differently.
The best interpretation is therefore not that scientists have discovered an autism treatment ready for patients. Rather, they have identified a specific biological target that successfully corrected a defined signaling problem across several laboratory models.
If future studies confirm its safety and identify the right patients, SLC6A20 inhibition could become one route toward more personalized treatment for conditions involving weak NMDA receptor signaling. For now, the work is a promising proof of concept and an intriguing sign that some developmental brain changes may remain more flexible in adulthood than previously assumed.
