Using AI to Advance Drug Delivery Systems for Glaucoma and Other Chronic Diseases

Newswise — Medical researchers at Johns Hopkins University Wilmar Eye Institute used artificial intelligence models and machine learning algorithms to determine which constituents of the amino acids that make up therapeutic proteins safely deliver therapeutics to animal eye cells. He said he was successful in predicting the most likely delivery.

A collaboration with researchers at the University of Maryland, the project hopes to advance new, more well-tolerated drug treatments for common chronic blinding eye diseases such as glaucoma and macular degeneration. 3000000 and about 20 million US people respectively. Current drug therapy for these diseases, which consists of multiple daily eye drops or frequent eye injections, is effective, but such delivery systems can be difficult to sustain and tolerate over the long term. , scientific efforts to develop delivery systems that bind to eye components are encouraged. It activates cells and safely extends the therapeutic effects of cell-borne drugs.

In 2020, the Food and Drug Administration approved an implantable device that can be placed in the eye to release glaucoma medication. Although the device worked longer than eye drops or injections, long-term use can cause eye cell death, forcing patients to return to eye drops and injections.

issued May 2 Nature Communications, a new study reveals the effectiveness of amino acids, also known as peptides or small proteins, in a model designed by artificial intelligence that binds to specific chemicals in rabbit eye cells and safely administers drugs over weeks. We have shown that it accurately predicts random sequences and reduces the need for Accommodates frequent and stringent treatment schedules. The research team specifically studied peptides that bind to melanin, the compound that gives the eye its color, which has the advantage of being widely present throughout the specialized structures of the cells of the eye.

Other studies investigating drug delivery using peptides have shown how effective this system is, but the research team wanted to find peptides that strongly bind to widespread ophthalmic compounds. He pointed out that he was thinking To that end, the researchers reasoned that rapid machine learning using artificial intelligence techniques could aid in the selection and prediction of effective peptide sequences to try. Dr. Laura EnsignMarcella E. Wal, Professor of Ophthalmology, Johns Hopkins University School of Medicine, and co-corresponding author of the paper.

The team started by feeding machine learning models with thousands of data points, including amino acid and peptide sequence features. These data helped computer models “learn” the chemical and binding properties of specific amino acid combinations and, in time, learned to predict candidate peptide sequences for drug delivery using melanin. .

An artificial intelligence model generated 127 peptides predicted to have varying abilities to penetrate specialized melanin-housing cells, bind to melanin, and be non-toxic to the cell. Of these 127 peptides, the model predicted the peptide called HR97 to have the highest binding success rate. The research team also confirmed the properties of these peptides, including good cellular uptake and binding, and no signs of cell death.

To test the model’s predictions, the researchers conjugated HR97 to brimonidine, a drug used to treat glaucoma by lowering intraocular pressure, and injected it into the eyes of adult rabbits. To determine the performance of HR97, the researchers measured levels of brimonidine in ocular cells by examining drug concentration in cells after administration of an drug delivery system. They found that large amounts of brimonidine were present for up to a month. This indicates that HR97 successfully penetrated the cells, bound to melanin and released the drug over a longer period of time. The researchers also confirmed that the intraocular pressure-lowering effect of brimonidine bound to HR97 lasted up to 18 days and showed no signs of irritation in the eyes of rabbits.

Future studies using artificial intelligence to predict peptides for drug delivery could have profound implications for other conditions involving melanin, Ensign said, targeting other specialized structures. It states that it can also be extended to

“We believe we are on our way to finding solutions that improve patient care and quality of life using drug delivery systems. It’s about creating things that can improve lives,” says Ensign.

In the future, researchers will need to extend the duration of action even further, test the AI ​​model’s success in predicting drug delivery with other drugs, and find ways to determine safety in humans, Ensign said. .

Other researchers who participated in this study are Henry Hsueh, Usha Rai, Watsala Liyanage, Yoo Chun Kim, Matthew Appell, Jahnavi Pejavar, Kirby Leo, Charlotte Davison, Patricia Kolodziejski, Ann Mozzer, HyeYoung Kwon, Maanasa Sista, Sri Vishnu Kiran Rompiccharla, Malia. Edwards, Ian Pisa, and Justin Haynes from the Johns Hopkins University School of Medicine. Nicole Anders and Avelina Hemingway of Johns Hopkins University Sidney Kimmel Comprehensive Cancer Center. Lenny Ti Chou and Michael Cummings of the University of Maryland.

This research was supported by funding from the NIH (R01EY026578, R01EY031041, P30CA006973 and S10OD020091), the Robert H. Smith Family Foundation, the Marcella E. Wall and Maryland E-Innovation Initiative Fund, and the Marcella E. Wall Professorship in Ophthalmology. was established. Research to Prevent Blindness, National Eye Institute Training Grant (T32EY007143), National Science Foundation Award (DGE-1632976), and National Center for Translational Science Advancement Grant (UL1TR001079).