UFRGS neuroscientist Eduardo Zimmer’s group employed systems biology to validate the effectiveness of current models in studying the disease and the results are promising.

Art by Clarice Wenzel

By Pedro Lira

In Brazil, 1.7 million seniors (60+) face dementia. Alzheimer’s disease alone drives 55% of cases, according to data from the Brazilian Alzheimer’s Association. This highlights the urgent need for new treatments, spurring science to explore diverse research strategies.

In a breakthrough for Alzheimer’s research, a group led by neuroscientist Eduardo Zimmer from the Federal University of Rio Grande do Sul confirmed a 92% similarity between mouse models and human disease manifestation in a study published in the January issue of iScience. This brings us closer to understanding Alzheimer’s disease and validates the effectiveness of current models.

First of all, it’s important to understand one factor that makes this scenario complex: there are two types of Alzheimer’s. One is sporadic, which accounts for around 95% of cases and usually affects people over the age of 65, and is multifactorial, i.e. a combination of environmental and genetic events. The other is genetic Alzheimer’s, which accounts for only 5% of cases and has almost 100% penetrance, i.e., if the person has any alterations in the APP, PSEN1, and PSEN2 genes, they will develop the disease, usually around the age of 40.

Understanding the complexity of Alzheimer’s disease begins with recognizing its two main types: sporadic and genetic. The former, which accounts for about 95% of cases, usually affects people over the age of 65 and is caused by a combination of environmental and genetic factors. The latter accounts for only 5% of cases but is highly predictable: if a person inherits certain mutations in the APP, PSEN1, and PSEN2 genes, they will almost certainly develop the disease, typically around 40.

Researchers used the predictability of genetic Alzheimer’s disease to develop the field’s primary models decades ago. The rationale is simple: if certain genetic mutations cause Alzheimer’s in humans, introducing those same mutations into rodents should replicate the disease process. Ideally, these models would exhibit symptoms such as memory loss and the buildup of beta-amyloid protein between neurons, disrupting communication and impairing brain function.

However, the scientific community is questioning the relevance of current Alzheimer’s models. Because they’re based on the rarer genetic type (only 5% of cases), they may not fully represent the far more common sporadic form of the disease. This disease is complex, influenced by behavioral and environmental factors not captured in existing models.

The limitations of existing models have spurred the search for better alternatives. In 2016, the National Institutes of Health (NIH), a U.S. government agency, funded the MODEL-AD consortium, a collaboration of research centers dedicated to developing models that more accurately mimic sporadic Alzheimer’s disease. A key difference in MODEL-AD’s approach is how they handle the beta-amyloid protein. Instead of introducing the specific mutation that causes plaque buildup in genetic Alzheimer’s, they “humanize” the animal’s beta-amyloid gene. Essentially, they modify the APP gene to produce a human-like protein without the mutation. This allows researchers to study how modifiable risk factors, such as obesity or high blood pressure, interact with the human-like protein and potentially contribute to the development of Alzheimer’s disease, mirroring the complex interplay observed in humans.

This innovative strategy has transformed Alzheimer’s research, enabling the creation of models that closely resemble the sporadic disease that affects most patients. However, the critical question remains: How well do these models reflect human disease at the molecular level?

This question motivated Eduardo Zimmer, a biochemist and professor in the Department of Pharmacology at UFRGS. His team used next-generation genetic sequencing to analyze postmortem brain samples from Alzheimer’s patients and animal models. Their evaluation focused on identifying similarities and differences between them. The results of the cross-species comparison are presented in a recent article published in the journal iScience.

Initially, the similarity was shockingly low. The sporadic model showed only 9% similarity to the disease in human brains—far below what researchers anticipated after decades of drug development based on this pattern. “When we saw that the animal models were not mimicking the disease, we realized that this would cause great confusion in the literature,” says Zimmer.

Building on this knowledge, the researchers explored a new approach: systems biology, specifically the subfield that focuses on how genes act in an orchestrated way to participate in different biological processes. “Everything in nature works in networks,” explains Marco Antônio De Bastiani, a postdoctoral researcher in the group and co-author of the paper. This understanding highlights the importance of analyzing biological processes holistically rather than focusing on individual genes in isolation.

“We then repeated the experiments but focused on analyzing the relevant biological processes involved, such as neuroinflammation,” De Bastiani continues. “Functionally, a group of genes associated with cerebral energy metabolism will be very similar between humans and rodents, even if the genes are not identical.”

From this new perspective, the similarity numbers increased dramatically. The sporadic disease model showed a 92% similarity to human Alzheimer’s disease. While the models may seem inadequate when focused solely on individual genes, they excel when examining biological processes. “In other words, 92% of the biological processes altered in human Alzheimer’s are also altered in this rodent model. This validates its use for studying different aspects of the disease,” says Zimmer.

Echoing the sentiment of British statistician George Box that “all models are wrong, but some are useful,” Zimmer says that “because it is necessary to simplify biological complexity to create models, they are rarely right, but they are still fundamental to understanding questions of nature. That’s why the search for new Alzheimer’s models is important, but it’s also encouraging to see how effective our current ones can be,” he concludes.