By Leonardo Cheng and Johnny Moseman

A team of Johns Hopkins engineers has discovered that machine learning can help predict efficient lipid nanoparticle designs, which can lead to improved vaccine development.

With the success of mRNA lipid nanoparticles (LNPs) in COVID vaccines, LNPs are becoming a popular means for gene therapies and development. However, design standards and rules of LNPs for such use are not well studied. More knowledge is needed on how LNP design impacts the delivery of their gene cargos to different cells to meet different treatment needs.

Addressing this challenge is a team of researchers led by Hai-Quan Mao, director, and core researcher at the Johns Hopkins Institute for NanoBioTechnology and professor of materials science and engineering at the Johns Hopkins Whiting School of Engineering. They have developed a machine learning model using curated DNA LNP data to gain insights from high-throughput screening studies to predict efficient LNP designs.

“Lipid nanoparticle carriers are not a ‘one-size-fits-all’ option. By combining machine learning with our LNP screening and optimization, we can customize designs to address specific therapeutic needs,” said Mao.

Their research was published in ACS Nano.

The team tested the delivery efficiency of over 1,000 LNP formulas in six types of cells, which include brain, kidney, eye, and cancer cells. The machine-learning platform identified critical relationships between nanoparticle composition and efficacy for each cell type. Then, they compared the similarities and differences among the identified LNP carriers for each cell type application.

The team found that the platform shows machine learning models can effectively predict the efficacy of new LNP formulations. It also provides one of the first quantitative approaches that identifies LNP design rules that allow for preferential gene delivery in different types of cells.

Gene and cell therapies show promise to fight a variety of diseases, from genetic disorders to cancer. However, reliable and targeted delivery of the therapeutic cargo to the specified cell and tissue targets remains elusive. The platform provides researchers with critical information about what makes one LNP design effective and can accelerate the development of LNP-based gene therapies.

“We plan to leverage this technology to conduct machine learning-guided and accelerated development of new lipid nanoparticle designs to deliver therapeutic gene cargos to cells of clinical interest,” said Leonardo Cheng, biomedical engineering PhD student and lead author of the paper.

This research is supported by the National Institute of Allergy and Infectious Diseases and the National Institute of Biomedical Imaging and Bioengineering.