Virtual mouse uses AI to test nanomedicine without living animals, offering ethical alternative

AI-Powered Tool Transforms Nanomedicine Development

Empa researcher Jimeng Wu has created a virtual mouse that could reshape how scientists develop nanomedicine, particularly for challenging conditions like brain tumor treatment. The AI-powered tool predicts how nanoparticles distribute across organs without requiring living animals, addressing both ethical concerns and practical limitations in current research methods 1

1

. Wu, a doctoral student in Empa's "Nanomaterials in Health" and "Technology and Society" labs, built this physiologically based pharmacokinetic model (PBPK) using data from 18 peer-reviewed mouse studies conducted by various research teams 2

2

.

The virtual mouse represents a significant alternative to animal testing by calculating nanoparticle distribution in the liver, kidneys, lungs, and spleen based on measurable properties such as size, coating, and surface charge. Unlike traditional PBPK models calibrated for single substances, Wu's AI model for nanomedicine adapts its parameters to different nanoparticles through a multivariate linear regression model—a machine learning approach that gives it flexibility across various nanomaterials 1

1

.

Tackling Drug Delivery Challenges at the Blood-Brain Barrier

The innovation addresses a critical challenge in biomedical research: delivering chemotherapy drugs across the blood-brain barrier. This highly selective filter protects the brain but blocks most medications, making brain tumors particularly difficult to treat. Nanoparticles—about 500 times smaller than the diameter of human hair—can potentially act as delivery vehicles that transport chemotherapeutical drugs through this barrier without causing damage 2

2

.

However, finding the right nanomaterial requires extensive testing. Nanoparticles distribute differently throughout the body depending on their shape, material composition, and size, accumulating in different organs with varying effects. Researchers have traditionally administered various nanomaterials to mice and examined their distribution and side effects—a process that is complex, time-consuming, expensive, and raises ethical issues under Swiss animal welfare legislation 1

1

.

Safe and Sustainable by Design Principles

The virtual mouse enables researchers to virtually test which nanoparticles suit specific tasks before manufacturing them, providing a decision-making aid before costly clinical trials begin. "This AI-supported screening tool allows researchers to virtually test which type of nanoparticles are best suited for a specific task before they even manufacture these particles," Jimeng Wu explains 2

2

.

Peter Wick, who supervises Wu's doctoral thesis alongside colleague Bernd Nowack, notes the model contributes to Safe and Sustainable by Design (SSbD) principles by increasing safety of new materials or therapies before development. However, he acknowledges limitations: the training dataset remains limited to 18 peer-reviewed papers with sufficient data quality, as many studies don't describe nanoparticle properties in adequate detail. The team now aims to feed the virtual mouse additional study data to increase prediction reliability, with a long-term goal to "shorten the process of developing nanomedicine materials all the way to their use as a drug in patients, while ideally being able to avoid animal testing" 1

1

.

Future Applications in Human Research

Wu's future research focuses on a "bridge strategy" to transfer her in silico model principles to human research by embedding them in a human PBPK model. This could expand analysis beyond the current four organs to include sensitive target organs, potentially investigating how certain nanoparticles cross barriers in human bodies 1

1

. The development signals a shift toward computational methods that could accelerate nanomedicine development while reducing reliance on animal experiments, though researchers must watch how expanded datasets improve model accuracy and whether regulatory bodies accept virtual screening as sufficient evidence for clinical advancement.