(Bloomberg) -- The global push to use artificial intelligence to find new medicines faces a crucial test as one front-runner starts approaching late-stage trials for a drug discovered by algorithms. 

Insilico Medicine — which has headquarters in Hong Kong and New York — used AI to develop an experimental drug for the incurable lung disease idiopathic pulmonary fibrosis. The treatment is in mid-stage trials in the US and China with some results expected early 2025.

The results of Insilico’s trials are being closely watched in the drug industry because the company used AI to identify a new approach to fight against the deadly disease and produce a novel molecule to treat it. UK-based researcher Deep Pharma Intelligence says that the Insilico therapy is the global industry’s first fully AI-based preclinical candidate. 

While pharmaceutical companies worldwide are investing heavily in AI, there is still no clear evidence the technology can generate life-saving therapies. A string of other leading molecules that relied on AI have faced setbacks and Insilico’s could still fail in the process or take years to reach the market. At the same time, the implications of any success would be huge, opening the door for new and cheaper AI therapies that can save lives and cut costs for health systems. 

While the US has sought to ensure China doesn’t get the upper hand in AI, any success by Insilico would highlight the benefits of international scientific collaboration. The company already straddles several geopolitical fault-lines to accelerate drug discovery. Its data scientists are mostly in Canada and the United Arab Emirates, where Ukrainians and Russians work alongside each other. Much of its laboratory testing is outsourced in mainland China, while staff in Taiwan purse partnerships with other drug companies. 

Insilico stands out because it “has a new target identified using AI, a molecule discovered using AI and has made it through to mid-stage clinical trials,” founder and CEO Alex Zhavoronkov said. If the company’s lung disease drug is successfully developed, “that would change everything.”  

Zhavoronkov, a Latvian scientist, set up Insilico in 2014 in Baltimore, within Johns Hopkins University where he got his master’s in biotechnology. These days, he generally dresses casually - in dark jeans and a black shirt with his corporate logo — while flying around the globe to seek partnerships and push projects forward. 

By his reckoning, Insilico’s model — designed by AI and synthesized, laboratory tested in China — is similar to how Apple Inc. works. The iPhone maker designs products in California and relies heavily on workers in the world’s second-biggest economy to make parts and assemble them.  

The field of AI became hot in the health-care industry after Google parent Alphabet Inc.’s DeepMind unit used an AI program called AlphaFold to beat a biologist at predicting the shape of proteins, the basic building block of diseases. The intention is to use AI to make it cheaper and faster to find new drugs by eliminating much of the guesswork and hundreds of lab experiments typically required to identify promising molecules. 

Beyond the start-up world, even large drugmakers like Pfizer Inc., GSK Plc to Takeda Pharmaceutical Co. are now tapping AI. Morgan Stanley has estimated that over the next decade, the use of AI in early-stage drug development could translate into an additional 50 novel therapies worth more than $50 billion in sales. 

Insilico, which has raised over $400 million from backers including Warburg Pincus LLC and Lilly Asia Ventures, says the steps from finding a novel target to designing a medicine to make it ready for human testing took just about 18 months at a budget of around $2.7 million. That’s roughly a third of the time and a 10th of the cost it might have taken without algorithms. 

Insilico now has an autonomous AI-run robotic laboratory in Suzhou in China, outside of Shanghai, which runs experiments in place of humans to analyze the quality and composition of substances and feeds data back to its system to improve its AI models. The company also has an AI platform that predicts clinical trial outcomes, and is building another one that generates biologics and studies quantum computing, said Alex Aliper, president of Insilico.

Still, AI’s use in drug development remains nascent and is littered with failures that show the difficulties facing Insilico in getting its drug to market. UK-based Exscientia Plc’s partner Sumitomo Pharma Co. said in 2022 it abandoned a candidate for obsessive-compulsive disorder designed by AI after it didn’t achieve the expected criteria in the initial stage of trials. BenevolentAI’s atopic dermatitis candidate, also identified via algorithms, failed to reduce inflammation in a mid-stage trial, it said in April.

An editorial published in journal Nature in October cautioned that claims by drugmakers working with AI need independent verification and clinical trials — the same standard all work in the drug industry is held to — to determine whether they hold up. 

Side effects in human beings can also be hard to predict via algorithms, and Insilico’s success won’t be clear until it’s completed final trials. Christoph Kuppe, a professor of nephrology at RWTH Aachen University in Germany, who has published a study in Nature on kidney fibrosis, said data from animal experiments on Insilico’s drug look fine. Yet he worries about the long-term effects of the treatment on people. “With all the anti-inflammatory therapies, one would usually be worried about is infections, how the immune system is affected,” he said. 

Still, currently available therapies for fibrosis treat mainly symptoms, and Insilico’s approach is unique as it targets cells thought to be responsible for scarring of tissues, Kuppe said. 

Insilico hasn’t observed side effects related to the immune system in its drug candidate so far and it will continue to monitor the long-term safety and efficacy in clinical trials, Zhavoronkov said.

Insilico now has more than 30 programs in its pipeline with three other experimental medicines at a clinical stage, such as for cancer and Covid-19. All have been made using its AI platform. In September, it said it sold rights to develop and sell a potential cancer drug to California-based Exelixis Inc. for an upfront payment of $80 million and is co-developing another candidate for cancer with Shanghai Fosun Pharmaceutical Group Co., controlled by Chinese billionaire Guo Guangchang.

The company is “really becoming the global tier-one player in the AI drug discovery space,” said Min Fang, the managing director who leads the healthcare, consumer and internet franchise in China for Warburg Pincus. Its business model of leveraging AI to identify new targets, design novel compounds and partner with other pharma for development is “very attractive,” he said. 

©2023 Bloomberg L.P.