The challenge of using models to research Alzheimer’s – 03/01/2024 – Fundamental Science

In Brazil, 1.7 million people aged 60 or over have some type of dementia. Alzheimer’s disease alone accounts for 55% of cases, according to data from the Brazilian Alzheimer’s Association. The demand for new treatments encourages science to launch different research strategies.

The research group led by neuroscientist Eduardo Zimmer, from the Federal University of Rio Grande do Sul, is closer to better understanding the disease: in an article published in January in the magazine iScience, scientists analyzed animal models – mice, in this specific case – currently used to study Alzheimer’s, and checked whether they actually represent how the disease manifests itself in humans. And the answer is positive: the similarity between the model of the most common type of Alzheimer’s was 92% with its human manifestation.

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

One of the main strategies to investigate this condition are models that mimic the effects of the disease on the human brain in animals. Scientific models, that is, are physical and/or mathematical representations of a context, event, process or idea system of interest with the aim of emulating a real situation (such as a disease) in a simplified and controllable situation. There are several types of scientific models, and animal models using mice are just one example.

Based on knowledge about genetic Alzheimer’s, which is simpler to predict, researchers developed the main models used in the area a few decades ago. The logic is: if people with such genetic changes will develop Alzheimer’s, genetically modified rodents with the same changes also develop symptoms, such as memory failure, beta-amyloid deposits – the accumulation of this protein between neurons prevents the transmission of signals , damaging neural activity –, and others.

These models, however, have been questioned by the scientific community. This is because these studies are limited to the specificities of just one type of Alzheimer’s, just the genetic one, which represents only 5% of diagnosed patients. At the other end, sporadic disease affects the majority of the population and has other layers of complexities, such as behavioral and environmental factors.

This question opened the doors to a search for new models. In 2016, a consortium of research centers was formed — MODEL-AD — funded by the National Institutes of Health, a US government agency, with the aim of developing different models that more reliably mimic the sporadic disease. For example, instead of inserting into the animal the human mutation that causes the deposit of beta-amyloid plaques, MODEL-AD “humanizes” the beta-amyloid protein gene without adding a mutation associated with genetic Alzheimer’s. In other words: researchers simply alter the APP gene so that it decodes a human-like protein. Thus, modifiable risk factors, such as obesity, high blood pressure and others, and genetic ones, are introduced into animals, making it possible to observe how the disease develops, just as it would occur in the human brain.

This strategy has revolutionized Alzheimer’s studies because it makes it possible to create models of the sporadic disease, which affects the majority of patients. But to what extent are these models similar, at the molecular level, to human disease?

It was this question that motivated Eduardo Zimmer, biochemist and professor in the Department of Pharmacology at UFRGS, to investigate the topic. Using next-generation genetic sequencing, the neuroscientist and his research group collected brains postmortem from Alzheimer’s patients and animal models, and evaluated the similarities and differences between them. The article published in the magazine iScience brings the results of this comparison between species.

Initially, the similarity was found to be tenuous. The sporadic model showed only 9% similarity with the disease in human brains, a percentage far from what was expected by researchers who have been developing drugs and therapies based on this pattern for decades. “When we saw that the animal models were not mimicking the disease, we realized that this would cause a lot of confusion in the literature,” says Zimmer.

Knowing this, the researchers had another idea. According to the principles of systems biology, a subarea of ​​this field of knowledge, genes act in an orchestrated way to participate in different biological processes. “Everything in nature works in networks”, explains Marco Antônio De Bastiani, postdoctoral fellow in the group, who also wrote the article. This means that analyzing biological processes would be better than analyzing an isolated gene.

“So we did the same tests, but analyzing the biological processes involved, such as neuroinflammation,” continues De Bastiani. “In functional terms, a group of genes associated with brain energy metabolism will be very similar between humans and rodents, even if the genes themselves are not identical.”

From this new perspective, the similarity numbers rose a lot. The sporadic disease model had 92% similarity with human Alzheimer’s disease. Focusing only on the gene, the models are bad, but analyzing biological processes, they are great. “In other words: the model shows that 92% of the biological processes altered in human Alzheimer’s are altered in that rodent model as well. So it serves to investigate different aspects of the disease”, explains Zimmer.

He often quotes British statistician George Box, who said that “all models are wrong, but some are useful.” “As we always need to simplify biological knowledge to make models, they are in fact almost never right, but they are still fundamental for understanding questions in nature”, says Zimmer. “That’s why the search for new Alzheimer’s models is important, but it’s also exciting to see how the ones we currently have can be very effective,” he concludes.

*

Pedro Lira is a journalist and social media at Instituto Serrapilheira.

The Fundamental Science blog is edited by Serrapilheira, a private, non-profit institute that promotes science in Brazil. Sign up for the Serrapilheira newsletter to keep up to date with news from the institute and the blog.