How Does Artificial Intelligence Help in Drug Development?

The Long Journey from Idea to Drug

Drug development is an expensive, inefficient process full of risks and failures. It is estimated that up to 86% of candidate drugs developed between 2000 and 2015 did not reach their intended target. The use of artificial intelligence (AI) could significantly improve, accelerate, and reduce the cost of this process. It is also likely to enable the development of more completely new molecules – first-in-class. Currently, numerous molecules developed with the help of AI and machine learning techniques are in various stages of clinical research. Preliminary results look promising.

Drugs for Common and Rare Diseases

Some biopharmaceutical start-ups use AI to design new molecules. For example, low molecular weight inhibitors targeting cancer, idiopathic pulmonary fibrosis, atopic dermatitis, ulcerative colitis, and other inflammatory and autoimmune diseases are in phase I-II studies. More AI-designed drugs, such as those against tuberculosis, malaria, or COVID-19, are in preclinical development.

How to Reach the Target Faster

In January 2023, Insilico Medicine announced positive results of its phase I study, which tested the safety and pharmacokinetics of the low molecular weight inhibitor INS018_055 for the treatment of idiopathic pulmonary fibrosis, a devastating disease causing scarring of lung tissue. This is the first molecule designed by AI after identifying a new target protein. In June 2023, the FDA approved the molecule's progress to a double-blind, placebo-controlled phase II study involving a total of 60 adult patients. Particularly impressive is the acceleration of the process enabled by AI – clinical development of INS018_055 started only in February 2021.

Smart Assistants

Generative AI can be used not only for designing specific molecules but also for predicting which group of patients will best respond to treatment with these molecules and why. In practice, extensive genome, transcriptome, and proteome analyses should be possible using AI, identifying biomarkers that determine the likelihood of treatment response.

The American biopharmaceutical start-up Recursion uses AI to analyze data from millions of experiments and billions of microscope images generated by robots in its laboratory. The company has low molecular weight drugs in phase II and III clinical trials targeting genetically determined oncological and rare diseases, such as familial adenomatous polyposis or neurofibromatosis type 2. The founder of the company, Chris Gibson, compares his platform to Google Maps for biology. “Biology and chemistry are so vast and complex. The goal is not to find everything. The goal is to find something really good and develop it,” he adds. So far, it seems that AI is quite successfully helping with this.

(este)

Sources:
1. Arnold C. Inside the nascent industry of AI-designed drugs. Nat Med 2023; 29 (6): 1292–1295, doi: 10.1038/s41591-023-02361-0.
2. National Library of Medicine. NCT05938920. Study evaluating INS018_055 administered orally to subjects with idiopathic pulmonary fibrosis (IPF). ClinicalTrials.gov, 2023 Dec 26. Available at: www.clinicaltrials.gov/study/NCT05938920