Artificial intelligence links gut bacterial metabolites to Alzheimer’s progression

In a new study published in cell reportResearchers have developed a sophisticated systems biology approach that combines artificial intelligence (AI), genetics and multi-omics analysis to explore how metabolites produced by gut bacteria affect Alzheimer’s disease.

The study identifies specific receptors in humans that interact with these metabolites, potentially opening new avenues for therapeutic intervention. This significant discovery could lead to the development of new drugs that target these interactions, raising hope for treating or even preventing Alzheimer’s disease.

Alzheimer’s disease is a progressive neurodegenerative disease that primarily affects older adults and is characterized by a decline in cognitive functions such as memory and reasoning. It is characterized by the accumulation of beta-amyloid plaques and tau protein tangles in the brain, which interfere with neurological function and lead to cell death.

The exact cause of Alzheimer’s disease is not fully understood, but it is believed to involve genetics, lifestyle and environmental factors that affect the brain over time. As the disease progresses, it can severely impact daily life and independence, making it one of the most common causes of dementia in older adults.

Previous research has shown that the gut bacteria of people with Alzheimer’s disease change as the disease progresses. Metabolites produced by these bacteria can affect brain health and may contribute to the development of disease. However, the specific pathways by which these metabolites exert their effects remain largely a mystery.

This gap in understanding prompted this new study, which aimed to identify the interactions between these metabolites and the human receptors they affect. The study was conducted by Feixiong Cheng and his team, bringing together experts from the Cleveland Clinic Genome Center, Roruvo Center for Brain Health, and the Center for Microbiome and Human Health.

The researchers used machine learning algorithms to analyze more than one million potential metabolite-receptor pairs to predict the interactions most likely to affect disease. Genetic data, including Mendelian randomization, complement these predictions by assessing causality and receptor involvement.

“Intestinal metabolites are key to many physiological processes in our bodies, and each key corresponds to a lock on human health and disease,” Cheng said. “The problem is that we have tens of thousands of receptors and thousands of metabolites in our system, so manually figuring out which key goes into which lock is slow and expensive. That’s why we decided to use artificial intelligence.

The study also involved experimental validation using neurons from Alzheimer’s patients, in which the effects of specific metabolites on tau protein levels, a key biomarker of progression of the disease, were tested. This multifaceted approach allows researchers to map important interactions within the gut-brain axis, thereby revealing potential therapeutic targets for Alzheimer’s disease.

One of the most striking results of the study was the identification of specific G-protein-coupled receptors (GPCRs) that interact with metabolites produced by gut bacteria. The researchers focused on orphan GPCR receptors whose natural activators were unknown and found that certain metabolites activate these receptors. This finding is of particular interest because it opens new avenues for drug development to target these receptors to potentially modulate their activity to benefit disease prevention or alleviation.

Among the metabolites studied, phenylethylamine and agmatine stood out for their effects on the tau protein, which is involved in the neurodegenerative features of Alzheimer’s disease. Studies have shown that these metabolites can significantly alter the levels of phosphorylated tau protein in neurons of Alzheimer’s disease patients. Agmatine, in particular, exhibits protective effects by reducing harmful tau phosphorylation, suggesting that it may be a potential candidate for therapeutic development.

The application of machine learning models is critical for predicting interactions between more than one million metabolite-receptor pairs. This high-throughput approach not only simplifies the identification of relevant interactions but also enhances understanding of the complex mechanisms by which gut microbiota influence brain health. By integrating genetic analysis and experimental data, the researchers were able to test these predictions and refine their understanding of the gut-brain axis in the context of Alzheimer’s disease.

While promising, the study’s authors acknowledge some limitations. The complexity of the gut-brain axis means these findings are preliminary and require further validation through experimental and clinical studies. Future studies are needed to confirm these interactions in organisms and explore the therapeutic potential of modulating these pathways.

Furthermore, studies have primarily focused on biochemical interactions at the molecular level without considering the broader physiological and environmental factors that may influence these processes in living systems.

Nonetheless, this study provides a valuable framework for understanding how gut bacterial metabolism affects brain health and disease. The implications of these findings extend beyond Alzheimer’s disease, as these methods and insights may be applicable to other neurological and systemic diseases affected by the gut microbiota.

“We focused specifically on Alzheimer’s disease, but metabolite-receptor interactions play a role in almost all diseases involving gut microbes,” Cheng said. “We hope our approach will provide a framework to advance the entire field of metabolism-related diseases and human health.”

The study is titled “Systematic characterization of the multi-omics landscape between gut microbial metabolites and GPCRome in Alzheimer’s disease” and is authored by Yunguang Qiu, Yuan Hou, Dhruv Gohel, Yadi Zhou, Jielin Xu, Marina Bykova, Yuxin Yang, James B. Leverenz, Andrew A. Pieper, Ruth Nussinov, Jessica ZK Caldwell, J. Mark Brown and Feixiong Cheng.