Artificial intelligence sheds light on how some brains resist Alzheimer’s memory loss

Some older adults develop the physical brain changes associated with Alzheimer’s disease but never experience memory loss or cognitive decline. A new study published in Acta Neuropathologica Communications reveals a specific genetic pattern that explains this natural protection against the disease. The research also introduces a new mouse model that mimics this resilient state, offering a path toward treatments that might prevent memory loss before symptoms appear.

Alzheimer’s disease is a progressive brain disorder that slowly destroys memory and thinking skills. Inside the brains of people with the condition, clumps of proteins called amyloid plaques and twisted fibers known as tau tangles accumulate. These protein build-ups disrupt communication between brain cells and eventually cause the cells to die.

For many years, medical professionals assumed that the presence of these plaques and tangles automatically led to dementia. Autopsy studies and brain scans eventually revealed a different reality. About one-fifth to one-third of older adults possess a high volume of these protein deposits but remain entirely sharp and clear-minded until the end of their lives.

Doctors call this condition asymptomatic Alzheimer’s disease. People with this condition represent a distinct biological state of cognitive resilience rather than simply being in an early stage of the illness. Their brains seem to possess built-in defenses that block the physical protein blockages from destroying their mental faculties.

Understanding exactly how these individuals maintain their memory has proven difficult. Progress stalled because researchers lacked a way to parse the massive amounts of genetic data from human brains. They also lacked an animal model that accurately recreated this specific condition of having brain pathology without mental decline.

A team of researchers at the University of California San Diego set out to solve this puzzle. The project was led by Debashis Sahoo, an associate professor of pediatrics and computer science, and Sushil K. Mahata, an adjunct professor of medicine. They wanted to uncover the biological mechanisms that decouple physical brain damage from cognitive failure.

To tackle the data problem, the researchers turned to an artificial intelligence framework called a Boolean Network Explorer. This computer model allowed them to analyze genetic information from thousands of human brain samples. Unlike standard methods that look for simple correlations, this approach looks for stable, directional relationships between genes that remain consistent across different people and stages of the disease.

Using this system, the team identified a distinct genetic pattern, or fingerprint, of 40 specific genes. This signature accurately separated healthy aging brains from those with symptomatic Alzheimer’s disease. The genes involved were heavily tied to functions like cell inflammation and the transport of chemical messengers within the brain.

The researchers tested this 40-gene fingerprint against 35 independent groups of human data to ensure it was accurate and reliable across different studies and brain regions. They also looked closely at which specific types of brain cells were driving these genetic changes. They found that astrocytes, a type of support cell in the brain, showed the most prominent alterations in their genetic activity.

Once they had a reliable human genetic signature, Sahoo and Mahata applied it to genetic data from various laboratory mice. They wanted to see if any existing animal models matched the genetic state of human asymptomatic Alzheimer’s disease. They found their match in a specific group of genetically modified mice.

The key to this discovery involved a protein called Chromogranin A. This protein is normally found inside the secretory granules of brain cells, which are tiny pouches that cells use to store and release chemical messengers. People with Alzheimer’s disease often have elevated levels of this protein in their cerebrospinal fluid.

The research team had previously engineered mice that were missing the gene responsible for making Chromogranin A. In the new study, they bred these mice with another type of mouse that is prone to developing destructive tau protein tangles. They then evaluated these new mice using both behavioral tests and microscopic brain examinations.

The behavioral and microscopic tests revealed an unexpected contrast between male and female mice. The male mice without Chromogranin A developed severe tau tangles in their brains, matching the physical damage seen in typical Alzheimer’s disease. Despite this extensive physical damage, these male mice navigated mazes and completed memory tests just as well as perfectly healthy mice.

This decoupling of brain damage and memory loss means the male mice closely mimic human asymptomatic Alzheimer’s disease. The female mice without Chromogranin A demonstrated an even stronger form of protection. They did not develop the destructive tau tangles at all, and they kept their memory and learning abilities completely intact.

“Even when the brain shows clear signs of Alzheimer’s, some people stay mentally sharp,” Mahata said in a press release.

To understand why the female mice were so protected, the researchers looked closely at their synapses using high-powered electron microscopes. Synapses are the microscopic gaps where brain cells connect and communicate. In typical Alzheimer’s disease, the tiny clear vesicles that carry chemical messages across these gaps are destroyed early in the disease process.

The female mice without Chromogranin A retained a dense, healthy supply of these clear messenger vesicles. Their brain cell connections looked almost identical to those of healthy control mice. This preserved brain architecture likely explains why the female mice maintained their cognitive abilities so well.

The research team also looked at how tau tangles spread through different parts of the brain cells. In standard disease models, tau tangles invade both the dendrites, which receive signals, and the axons, which send signals. The female mice without Chromogranin A managed to suppress the formation of these tangles in both cellular compartments.

By removing Chromogranin A, the researchers essentially activated a latent protective system in the brain. The team noted that this protection appears to rely heavily on biological sex. Recognizing these sex differences could help scientists tailor future treatments depending on whether a patient is male or female.

While the study offers a new way to study cognitive resilience, it does have a few limitations. The researchers primarily examined the hippocampus and the prefrontal cortex, which are brain areas heavily involved in memory and decision making. They did not analyze other brain regions that are also affected in the early stages of Alzheimer’s disease, such as the basal forebrain.

Future studies will need to explore whether this protective mechanism works uniformly across the entire brain. The researchers also do not yet know the exact biological reasons behind the differences between the male and female mice. They plan to investigate whether sex hormones, chromosome differences, or alternative cellular structures are responsible for the added protection seen in females.

The team also intends to look beyond pure genetic data in their upcoming experiments. They want to study the actual proteins and metabolic chemicals in the brain to get a more complete picture of how resilience works. Combining these different types of biological data will help them understand how cellular changes directly affect animal behavior.

Ultimately, this new computational and experimental framework offers a fresh perspective on Alzheimer’s disease. By shifting the focus away from late-stage brain damage and toward natural protective mechanisms, scientists might discover new ways to intervene. The goal is to develop therapies that mimic this natural resilience, keeping patients’ minds sharp even if they carry the physical markers of the disease.

The study, “AI guided discovery of a murine model of asymptomatic Alzheimer’s disease,” was authored by Suborno Jati, Sahar Taheri, Satadeepa Kal, Subhash C. Sinha, Brian P. Head, Sushil K. Mahata and Debashis Sahoo.