Researchers at EPFL have created the first 3D Lipid Brain Atlas,
mapping 539 distinct chemical territories in the mouse brain using mass spectrometry and machine learning, revealing how lipids organize to influence brain function and disease.
A team led by Giovanni D’Angelo, Luca Fusar Bassini, and Gioele La Manno at EPFL has constructed the first detailed 3D map of lipid distribution across the mouse brain, identifying 539 chemically distinct regions called “lipizones.” These territories, defined by unique lipid combinations, provide a new framework for understanding brain organization and its role in conditions like Alzheimer’s and depression, according to a study published in Nature.
Mapping Lipid Patterns in the Mouse Brain
The researchers used mass spectrometry imaging to analyze 172 lipids across 109 brain slices from 11 mice, generating 7 million data points. Machine learning tools transformed these into a 3D model, revealing lipizones that align with known brain structures but also expose previously hidden chemical boundaries. So scientists have had only a blurry picture of how lipids differ from one part of the brain to another,
D’Angelo said, emphasizing the gap this atlas fills.
The atlas shows lipids functioning like molecular postal codes, linking distant but related brain regions. For example, clusters of nerve cell bodies were found to share lipid signatures with distant fiber terminals, a connection missed by gene-based maps. White matter, once thought uniform, emerged as a patchwork of distinct lipid zones, suggesting greater chemical diversity in myelin-producing cells than previously understood.
Researchers integrated high-resolution mass spectrometry imaging with sophisticated machine-learning algorithms to examine around 7 million molecular data points across the entire mouse brain. Instead of categorizing brain regions based solely on anatomy or gene expression, the team concentrated on the specific lipid combinations found in various areas. This analysis revealed 539 distinct metabolic territories, or “lipizones,” each defined by a characteristic lipid signature.
The study also uncovered dynamic lipid changes during pregnancy. Pregnant mice showed widespread shifts in lipid composition, with the cortex and white matter undergoing significant reorganization. Levels of galactosyl ceramide, a key myelin component, rose sharply, indicating the brain may rewire itself to meet physiological demands. Together, these shifts suggest the brain retunes its own wiring to meet the demands of pregnancy,
La Manno said.
Uncovering Brain Chemistry in Disease and Pregnancy
The atlas, now freely available online at https://lbae-v2.epfl.ch/, offers a reference for studying lipid disruptions in diseases like Alzheimer’s and depression. While the mouse map cannot directly apply to humans, the methodology could reveal similar patterns in the human brain. What transfers to humans isn't the map, it's the method,
said Fusar Bassini.
The Lipid Brain Atlas challenges traditional views of brain organization, highlighting lipid chemistry as a critical, underexplored layer of neural structure. By combining lipid data with genetic maps, researchers can better understand how chemical changes affect brain function over time.
This work opens new avenues for investigating the molecular mechanisms behind neurological disorders and the brain’s adaptability. Scientists now have a tool to explore how lipid patterns shift in aging and disease. The Atlas reveals that the brain constantly adjusts its lipids to meet the body's changing needs,
D’Angelo noted.
Expanding Understanding of Brain Chemistry
The results establish a framework for examining how the brain’s chemical structure varies across regions, cell types, genders, physiological conditions, and possibly disease. What was once seen as a simple white matter landscape is now revealing itself as a complex molecular mosaic, with pregnancy demonstrating how significantly this landscape can be altered.
Scientists have spent decades studying the brain’s cells, connections, genes, and proteins. Yet another essential part of brain biology has received far less attention: lipids. These fatty molecules form the membranes around neurons, support myelin insulation that shields nerve fibers, and are crucial for communication between brain cells. Traditional imaging methods struggle to differentiate individual lipid molecules, creating a significant gap in understanding the brain’s chemical organization.
The research team examined 172 different lipids in 109 brain slices collected from 11 mice. This produced around 7 million measurements, which researchers integrated into a detailed 3D model using machine learning tools created for the project. The resulting atlas identified 539 chemically unique regions, termed “lipizones,” each with a specific lipid combination forming a distinct chemical signature. Many of these regions correspond to established brain structures and cell types. Others reveal unexpected relationships between groups of nerve cell bodies and distant regions connected by their fibers.
In pregnant mice, lipids underwent significant shifts across multiple regions, with changes more extensive than the typical differences seen between males and females. The research also discovered that the lipid makeup of a small brain tissue sample holds sufficient data to identify its original region. Essentially, the tissue’s chemical makeup acted as a molecular fingerprint, a unique identifier for its location in the brain.
The Lipid Brain Atlas is a foundational resource for studying lipids in brain development, physiology, and pathology. It provides a framework for exploring how lipid patterns change in response to physiological states, such as pregnancy, and how these changes might relate to neurological conditions. Researchers can now begin asking another fundamental question: What is the chemical identity of this region?
The answer appears to be encoded, at least in part, in its lipids.