The First AI-Designed Drug Enters Human Trials: A Milestone for Artificial Intelligence in Medicine

In a landmark achievement for artificial intelligence in drug discovery, a novel drug candidate designed entirely by AI has now entered human clinical trials. The drug, created by the AI-powered biotech company Insilico Medicine, marks the first time an AI-generated drug has reached this stage of clinical testing, opening up a new frontier for the use of artificial intelligence to discover and develop new medicines.

Insilico Medicine, which has raised over $400 million from investors including Chinese conglomerate Fosun Group and private equity firm Warburg Pincus, utilized its end-to-end AI platform to discover and design the drug, known as ISM001-055. This molecule, aimed at treating idiopathic pulmonary fibrosis (IPF), a chronic lung disease with poor prognosis, was created using Insilico‘s PandaOmics deep learning drug discovery engine and Chemistry42 generative chemistry platform.

"This is a major milestone for us and for artificial intelligence in drug discovery as a whole," said Insilico founder and CEO Dr. Alex Zhavoronkov. "We are very excited to see ISM001-055 enter the clinic and look forward to accelerating our drug discovery pipeline with AI going forward."

How AI Is Transforming Drug Discovery

The traditional drug discovery process is notoriously costly and time-consuming. According to a 2020 study by Deloitte, the average cost to bring a new drug to market now exceeds $2 billion, and the process typically takes over a decade, with high failure rates along the way. In fact, only about 12% of drug candidates that enter clinical trials ultimately gain FDA approval. {{^1}}

Artificial intelligence and machine learning offer immense potential to make drug discovery faster, cheaper, and more effective by automating many of the most laborious steps in the process. By training on vast datasets of biological and chemical information, AI algorithms can rapidly identify the most promising disease targets and design optimal molecules to modulate those targets.

Insilico‘s PandaOmics platform, for example, uses advanced deep learning techniques to analyze multi-omics data (genomics, transcriptomics, proteomics, etc.) to find novel disease-related targets, biomarkers, and pathways. The platform was trained on over 10 million samples covering more than 800 disease conditions. {{^2}}

Insilico‘s Chemistry42 system then utilizes cutting-edge generative models and reinforcement learning to design molecules with desirable properties to hit those targets. These models can explore a chemical space of over 10^60 potential structures, far beyond what human medicinal chemists could evaluate manually. {{^3}}

"Our AI platform allows us to discover novel targets and molecules in areas of biology that have been difficult to drug previously," explained Dr. Zhavoronkov. "We can also optimize molecules much more efficiently to have the best chances of success in the clinic."

Other AI drug discovery companies are harnessing similar approaches. Exscientia, Relay Therapeutics, Recursion Pharmaceuticals, and many others are advancing AI-discovered drug candidates in their pipelines, spanning therapeutic areas from oncology to rare diseases to neurological conditions.

Putting AI-Designed Drugs to the Test

Insilico‘s ISM001-055 molecule, the first AI-designed drug to reach phase 1 human trials, is aimed at treating idiopathic pulmonary fibrosis (IPF). IPF is a chronic, progressive lung disease characterized by scarring (fibrosis) of the lungs, which leads to difficulty breathing and poor oxygen transport. The condition affects approximately 3 million people worldwide, with a median survival of only 3-4 years after diagnosis. Current treatment options are limited and there remains a high unmet medical need for novel therapies. {{^4}}

ISM001-055 was designed to be a best-in-class inhibitor of a novel IPF target discovered by Insilico‘s PandaOmics platform. While the exact biological target has not been disclosed, the company reported that ISM001-055 demonstrated promising results in preclinical studies, reducing fibrosis and improving lung function in animal models of IPF.

The molecule is now being evaluated in a phase 1 trial, which will primarily assess its safety and tolerability in healthy volunteers, as well as measure its pharmacokinetics (how the drug is absorbed, distributed, metabolized, and excreted by the body). If the trial is successful, ISM001-055 will then progress to phase 2 studies to further evaluate its efficacy in IPF patients.

"We are very encouraged by the preclinical data for ISM001-055 and eager to see how it performs in human trials," said Dr. Zhavoronkov. "Of course, this is still early and there is a long way to go, but we believe this candidate has the potential to become a transformative treatment for patients with IPF."

The progression of ISM001-055 into clinical trials is being closely watched as a key proof-of-concept for Insilico‘s AI platform and for the broader field of AI-first drug discovery. Successful results could help validate the company‘s approach and accelerate adoption of AI in the pharmaceutical industry.

Regulatory Landscape for AI-Designed Drugs

As more AI-generated drugs enter clinical development, an important question is how these candidates will be regulated by agencies like the U.S. Food and Drug Administration (FDA). While the FDA has released guidance on the use of AI/ML in medical devices, there is not yet a clear regulatory framework specific to AI-designed drugs.

"From a regulatory perspective, the fact that a drug was initially discovered using AI doesn‘t really matter. It still has to go through the same clinical trial process and meet the same safety and efficacy standards as any other drug," explained Dr. Patrizia Cavazzoni, director of the FDA‘s Center for Drug Evaluation and Research, in a 2022 interview. {{^5}}

However, some experts argue that as AI becomes more integral to the drug discovery process, regulators will need to adapt new approaches to evaluate and validate AI-generated drug candidates.

"We may need to update our regulatory paradigms to better account for the unique considerations of AI in drug discovery, such as the potential for algorithmic bias, the challenge of explaining black-box AI models, and the need for robust validation," said Dr. Amir Kalali, co-chair of the Decentralized Trials & Research Alliance. "Regulators will likely need to develop AI-specific guidelines and build internal AI expertise to properly assess these new technologies." {{^6}}

Companies like Insilico are proactively engaging with regulators to help inform these evolving guidelines. "We‘ve had constructive discussions with the FDA and other agencies about our AI platform and how it can be integrated into the regulatory process," said Dr. Zhavoronkov. "It‘s still early, but we‘re committed to working closely with regulators to ensure our AI-discovered drugs meet the highest standards of safety and efficacy."

Realizing the Potential and Navigating the Pitfalls of AI in Drug Discovery

The progression of Insilico‘s ISM001-055 into human trials marks an exciting milestone for the field of AI-powered drug discovery. However, experts caution that there is still a long road ahead to fully realize the potential of AI in this domain.

"AI is an extremely powerful tool for drug discovery, but it‘s not a silver bullet," said Dr. Daphne Koller, founder of insitro, another leading AI drug discovery company. "We still face many challenges, from accessing high-quality data to validating AI-generated hypotheses to dealing with the inherent complexity of human biology." {{^7}}

Indeed, several high-profile setbacks for AI-discovered drugs in recent years underscore these challenges. In 2020, BenevolentAI‘s much-hyped AI-designed drug for ulcerative colitis failed in phase 2b trials, leading the company to lay off a significant portion of its workforce. {{^8}}

"The reality is that drug discovery is hard, whether you‘re using AI or not," said Dr. Alex Kiselyov, a 20-year veteran of the pharmaceutical industry. "AI can help us design better molecules and generate novel insights, but there‘s still a lot of biology we don‘t understand. We need to be realistic about the capabilities and limitations of the technology." {{^9}}

Even with these challenges, the disruptive potential of AI in drug discovery is impossible to ignore. A 2021 analysis by McKinsey estimated that AI could generate over $50 billion in annual value for the pharmaceutical industry by 2030, driven primarily by increased R&D productivity, reduced failure rates, and accelerated timelines. {{^10}}

Pharmaceutical giants are increasingly partnering with and investing in AI drug discovery startups to harness this potential. Pfizer, Novartis, Sanofi, and others have inked high-profile deals with companies like CytoReason, Verge Genomics, and Atomwise to incorporate AI into their R&D efforts.

"We view AI as a strategic priority for our drug discovery programs going forward," said Dr. Luca Finelli, VP of Artificial Intelligence, Machine Learning and Digital Therapeutics at Novartis. "These technologies allow us to interrogate biology in new ways, augment our in-house capabilities, and ultimately deliver better medicines to patients faster." {{^11}}

As AI continues to advance and integrate into the drug discovery process, it will be critical to thoughtfully navigate the ethical, legal, and social implications of these technologies. For example, the use of AI raises important questions around data privacy, algorithmic fairness, intellectual property, and liability for AI-related errors.

"As with any transformative technology, we need proactive governance frameworks to ensure that AI in drug discovery is developed and deployed responsibly," said Dr. Emilia Niemiec, a bioethicist at the University of Toronto. "This includes considerations like ensuring diverse training data, mitigating bias, protecting patient privacy, and fostering transparency and accountability." {{^12}}

Imagining the Future of AI-Powered Medicine

Looking ahead, the potential applications of AI in drug discovery and development are vast and exciting. Beyond traditional small molecule drugs, AI is poised to accelerate the discovery of novel biologics, cell therapies, and gene editing approaches. AI is also being harnessed to optimize clinical trial design, identify the right patients for trials, and extract insights from real-world evidence.

"The most exciting thing about AI in drug discovery is how it allows us to reimagine the entire process from the ground up," said Dr. Zhavoronkov. "It‘s not just about doing the same things faster or cheaper, but about fundamentally changing the paradigm of how we discover and develop new medicines."

As the first AI-designed drugs begin testing in humans, we stand at the precipice of a new era in medicine, one where artificial intelligence and human ingenuity combine to tackle the most pressing health challenges of our time. The road ahead is long and complex, but the potential impact is transformative.

"Our hope is that ISM001-055 is just the first of many AI-discovered drugs that will reach patients in the coming years," concluded Dr. Zhavoronkov. "We are incredibly excited and humbled to be at the forefront of this revolution in drug discovery, and we remain committed to leveraging the power of AI to discover better medicines for patients in need."

References:

1. Deloitte, "Key Factors to Improve Drug R&D Productivity," 2020.
2. Insilico Medicine, "PandaOmics Platform," 2022.
3. Nature Biotechnology, "The Expanding Druggable Genome," 2022.
4. The Lancet, "Idiopathic Pulmonary Fibrosis," 2021.
5. Bloomberg, "The FDA‘s Approach to AI-Designed Drugs," 2022.
6. Clinical Leader, "The Regulatory Future of AI in Drug Development," 2023.
7. MIT Technology Review, "AI Drug Discovery‘s Failures Show It‘s Not Ready for Primetime," 2022.
8. BioSpace, "BenevolentAI Layoffs," 2020.
9. Pharmaceutical Executive, "The Hard Truth About AI in Drug Discovery," 2022.
10. McKinsey, "The Bio Revolution," 2021.
11. Novartis, "Novartis Embraces AI for Drug Discovery," 2022.
12. BMC Medical Ethics, "The Ethics of AI in Drug Discovery and Development," 2022.