First Fully AI-Generated Drug Enters Phase II Clinical Trials for Idiopathic Pulmonary Fibrosis
核心洞察
Insilico Medicine (搜索)'s INS018_055 becomes the first drug with both AI-discovered target and AI-generated design to reach Phase II clinical trials for idiopathic pulmonary fibrosis (搜索) treatment.
The drug development process took just 30 months from target discovery to Phase I trials, approximately half the time of traditional drug discovery methods.
IPF affects about 100,000 people in the U.S. and typically leads to death within 2-5 years if untreated, with current treatments only slowing disease progression.
Insilico Medicine (搜索), a Hong Kong-based biotech startup, has achieved a significant milestone in artificial intelligence-driven drug discovery by advancing INS018_055, the first fully AI-generated drug, into Phase II clinical trials for treating idiopathic pulmonary fibrosis (搜索) (IPF). The company, which has secured more than $400 million in funding, represents a breakthrough in demonstrating AI's potential to accelerate pharmaceutical development.
Novel AI-Driven Discovery Process
The drug development journey began in 2020 when Insilico scientists applied their target discovery engine PandaOmics to analyze omics and clinical datasets related to tissue fibrosis (搜索). The AI system performed sophisticated gene and pathway scoring, identifying 20 potential targets through deep feature synthesis, causality inference, and de novo pathway reconstruction. A natural language processing engine analyzed millions of data files, including patents, research publications, grants, and clinical trial databases, to assess target novelty and disease association.
"It is the first fully generative AI drug to reach human clinical trials, and specifically Phase II trials with patients," said Alex Zhavoronkov, founder and CEO of Insilico Medicine (搜索). "While there are other AI-designed drugs in trials, ours is the first drug with both a novel AI-discovered target and a novel AI-generated design."
Following target identification, scientists deployed Chemistry42, the company's generative AI chemistry engine, which utilizes 500 predictive pre-trained models including transformer-based systems, GANs, and genetic algorithms. These models reward and punish generated molecules based on their ability to meet necessary conditions, including target engagement, metabolic stability, and penetrability. From 79 synthesized molecules, researchers selected the 55th compound, which demonstrated promise in improving fibrosis (搜索) with a favorable safety profile in mouse models.
Addressing Critical Medical Need
Idiopathic pulmonary fibrosis (搜索) affects approximately 100,000 people in the United States and five million worldwide. The chronic disease causes scarring in the lungs and progressive, irreversible decline in lung function. Patients typically die within two to five years following diagnosis, and current treatments primarily focus on slowing disease progression while often causing uncomfortable side effects.
"Insilico chose to target IPF because the Company wanted this lead program to set a precedent: simultaneous success in target discovery and drug candidate generation for a broad indication such as fibrosis (搜索)," according to the company's development rationale. The condition's implications in aging also influenced the strategic focus.
Accelerated Development Timeline
The preclinical candidate was selected in February 2021, just 18 months after the project began. Nine months later, the company announced first-in-human Phase I trials, achieving the milestone from novel target discovery to Phase I in under 30 months—approximately half the time required for traditional drug discovery methods.
INS018_055 demonstrated nanomolar IC50 values for target inhibition and showed activity against nine other fibrosis (搜索)-related targets. In vivo studies revealed the compound's ability to improve fibrosis in a Bleomycin-induced mouse lung fibrosis model, leading to enhanced lung function. The molecule also demonstrated a favorable safety profile in 14-day repeated mouse dose range-finding studies.
Clinical Trial Progress
After completing IND-enabling studies, Insilico initiated a first-in-human microdose trial in November 2021 in Australia with eight healthy volunteers. The results exceeded expectations, showing favorable pharmacokinetic and safety profiles that successfully demonstrated clinical proof-of-concept.
The drug advanced to Phase I clinical trials, enrolling 78 healthy volunteers in New Zealand across 10 cohorts consisting of five single ascending dose and five multiple ascending dose cohorts to determine maximum tolerated dose and establish dosage recommendations.
The current Phase II study is a randomized, double-blind, placebo-controlled trial taking place over 12 weeks in China. Insilico plans to expand the testing population to 60 subjects at 40 sites in the United States and China. "We expect to have results from the current Phase II trial next year," Zhavoronkov stated, noting that exact timing for future phases remains difficult to predict due to the disease's rarity and specific patient criteria requirements.
Broader AI Pipeline
Beyond INS018_055, Insilico has two other drugs partially generated by AI in clinical stages. One is a COVID-19 (搜索) drug in Phase I clinical trials, and another is a USP1 inhibitor for treating solid tumors (搜索) that recently received FDA approval to initiate clinical trials.
"When this company was launched, we were focused on algorithms—developing the technology that could discover and design new molecules," Zhavoronkov reflected. "I never imagined in those early days that I would be taking my own AI drugs into clinical trials with patients. But we realized that in order to validate our AI platform, we needed to not only design a new drug for a new target, but bring it into clinical trials to prove that our technology worked."
The company remains optimistic that INS018_055 will reach market and benefit patients within the next few years, pending successful completion of Phase II and subsequent Phase III studies with hundreds of participants.