Launch of Insilico’s Phase II Program Highlights Generative AI Momentum
14 Aug 2023
Phase 1Orphan DrugPhase 2Clinical Result
Pictured: Abstract design of computer code/iStock, carloscastilla
Last month, Hong Kong–based Insilico Medicine entered a drug fully created with generative artificial intelligence into Phase II clinical trials.
“To the best of our knowledge, this is the first drug discovered and developed using generative AI to have reached this clinical stage of development,” Sujata Rao, chief medical officer at Insilico, told BioSpace.
INS018_055, a potentially first-in-class anti-fibrotic small molecule inhibitor, is being evaluated to treat idiopathic pulmonary fibrosis (IPF). The first patients have been dosed in the Chinese cohort, while enrollment in the U.S. is expected to begin in the second half of this year with a target of 60 subjects across both countries. The FDA granted INS018_055 Orphan Drug designation in February 2023.
When Insilico was founded in 2014, “deep learning systems had just begun to outperform humans in image recognition,” Mona Flores, global head of medical AI at NVIDIA, told BioSpace. “The breakthroughs triggered a surge in AI interest, but most projects focused on imaging, voice and text. Training and validating deep neural networks to analyze those types of data takes days”—a relatively long time in the AI space.
Thus, nearly a decade later, AI technology is just beginning to hit primetime in drug development, with Insilico’s Phase II as one reflection of that success.
“So far, only a few AI-discovered drugs have entered clinical trials [and] most of them are in Phase I clinical trials,” said Chris Meier, managing director and partner at the Boston Consulting Group. “Reaching Phase II is, therefore, an important milestone. Going forward, we are likely to see other AI-discovered drugs entering clinical trials, including Phase II . . . and beyond.”
Enabling Target Discovery
Part of what makes the use of AI so attractive in drug development is its ability to reduce time and cost. For example, Rao said it took Insilico approximately 18 months from target discovery to IND-enabling studies, “which is much faster than typical drug discovery.”
Sujata Rao, Insilico
Insilico’s AI engine, called the Pharma.AI suite, involves three different platforms: PandaOmics, which uses a knowledge graph with target disease associations generated by a natural language processing model and develops algorithm options to find the best potential therapeutic targets; Chemistry42, which enables a workflow from novel target identification to be predictive in de novo studies; and inClinico, which enables the prediction of the outcome of clinical trials.
To develop INS018_055, the company used this system to assess the target within the context of millions of data files accumulated over several decades, including patents, research publications, grants and databases for clinical trials, and determined that it was likely to perform well against three key measures: novelty, confidence and ease of pursuing the drug commercially.
In January, Insilico reported positive data from a New Zealand–based Phase I trial of INS018_055 in healthy volunteers. In line with the company’s preclinical modeling, the data showed the drug was safe and tolerable with a favorable pharmacokinetic profile.
Rao declined to identify the drug’s target, only referring to it as Target X. She did share that the target is a key regulator of three pathways relevant to fibrosis: YAP/TAZ, TGF-β and Wnt.
The primary objectives of the Phase IIa program are to test the safety, tolerability, PK pro preliminary efficacy of INS018_055 over 12 weeks. In addition to the experimental drug, some patients will be on background therapy, such pirfenidone or nintedanib—the standard of care for patients with IPF—while some will receive INS018_055 alone. In parallel, the company is using its AI platform to develop a new inhalable formulation of the IPF candidate.
As Insilico pursues its own therapy, the company’s AI platforms have attracted interest from a growing number of biopharma firms in recent years. In November 2022, the company inked a multiyear research deal with Sanofi worth up to $1.2 billion to advance drug development candidates for six targets. Earlier that year, Fosun Pharma and Insilico entered a collaboration to advance multiple drug targets. Additional partners include Teva Pharmaceutical, Boehringer Ingelheim, Astellas Pharma, Janssen, Merck KGaA, and Pfizer.
Pictured: Insilico lab space/Courtesy of Insilico Medicine
The Evolving AI Space
Insilico is not alone in developing AI technologies that could support drug development. Another big player in this space is NVIDIA. In March 2023, the company unveiled NVIDIA Inception, a free program that includes large language models and generative AI services that have already been adopted by many life sciences technology and pharma firms.
Amgen was among the early users and was able to reduce the time it usually takes to train custom models for molecule screening and optimization from three months to just “a few weeks,” Peter Grandsard, executive director of biologics therapeutic discovery at Amgen’s Center for Research Acceleration by Digital Innovation, told BioSpace. Insilico has also taken advantage of NVIDIA’s technology, developing its Pharma.AI suiteon NVIDIA’s graphics processing units. Other users include Atomwise, Evozyne, Relation Therapeutics, Alchemab Therapeutics, Peptone and Recursion.
Over the past five years, AI technology has improved substantially, Meier said—especially the use of generative AI. “We have seen an acceleration in the number of AI-discovered drug molecules,” he said. “In particular, the number of small molecule drugs discovered with the help of AI is growing strongly.”
This number reached 158 in 2021, up from 119 in 2020 and 121 in 2019, according to data collected by his firm, Meier said.
By accelerating these early stages of drug development, AI is enabling biopharma firms to take on steps that have traditionally fallen to academics, Anat Cohen-Dayag, president and CEO of Israel-based Compugen, told BioSpace. Drug target discovery is usually done in academia and moved into drug development after 10 to 15 years of research around the target, she explained. “That’s what is unique at Compugen: we start from square one.”
Ana Mulero is a freelance writer based in Puerto Rico. She can be reached at anacmulero@outlook.com and @anitamulero on Twitter.
For more details,please visit the original website
The content of the article does not represent any opinions of Synapse and its affiliated companies. If there is any copyright infringement or error, please contact us, and we will deal with it within 24 hours.
Organizations
Indications
Hot reports
Get started for free today!
Accelerate Strategic R&D decision making with Synapse, PatSnap’s AI-powered Connected Innovation Intelligence Platform Built for Life Sciences Professionals.
Start your data trial now!
Synapse data is also accessible to external entities via APIs or data packages. Leverages most recent intelligence information, enabling fullest potential.