World’s First AI-Designed COVID Drug Starts Clinical Trials -  The Star | Virus World | Scoop.it

Believed to be the world’s first COVID-19 therapeutic created by generative AI, the drug will soon begin clinical trials, the Canadian-founded company announced. For the first time ever, a COVID-19 drug entirely designed by generative AI has been approved for human use and will soon begin clinical trials in China, the Canadian-founded company announced Thursday. The drug, named ISM3312, was discovered in a matter of months by multinational biotechnology company Insilico Medicine. Believed to be the world’s first COVID-19 therapeutic created by generative AI, one of the company’s CEOs claims the drug is already “as perfect as it can get,” according to initial tests. “I just don’t think it’s possible to make a better one, period,” said Alex Zhavoronkov, the Canadian founder and co-CEO of Insilico Medicine, told the Star. “The molecule is so stable... it is an absolutely novel molecule not found in any of the libraries” of known therapeutics. ISM3312 showed great promise during preclinical studies done on mice and isolated cells, able to significantly lower viral load in lung tissue and reduce lung inflammation. It can be taken orally or inhaled, is effective even at low doses and potentially has few side effects, the company said in a release. “If your molecule is very tolerable, and if you make very little of it available in the body and it still works, it’s perfect, right?” Zhavoronkov said.

 

ISM3312 works by inhibiting a protein called 3CL protease — a critical factor in viral replication and a popular target for anti-COVID drugs. Unlike similar therapeutics, however, Zhavoronkov said his drug works on a very broad spectrum — showing efficacy not only against all current COVID variants, but also coronaviruses other than SARS-CoV-2. As such, it may possess the ability to resist future mutations, providing a solution to drug-resistant strains, he continued. “You can also use it potentially for prevention,” Zhavoronkov said — because the 3CL protease target is only present in viruses, it should be safe to take when otherwise healthy. The company started work on the drug in early 2020, back near the start of the pandemic. They began by slotting the 3D structure of 3CL protease into one of their in-house AI programs; within a matter of days, the model was able to generate about 20 different potential drug candidates from scratch. In contrast, traditional drug discovery pipelines can take years or even decades to find one suitable therapeutic. “The beauty of generative AI (in drug discovery) is instead of searching for a needle in a haystack, it generates a bunch of perfect needles for you,” Zhavoronkov said. “Then you just have to choose from those perfect needles — if you want the longer one or the one that is more stable.” By February of 2020, Insilico had published their first set of drug candidates online; by April, they’d filed a related patent application. But it wouldn’t be until recently that they’d receive approval to start clinical trials. “Unfortunately, at that time, you probably remember everybody was running around with their heads on fire, right?” Zhavoronkov said.

 

Insilico, still a relatively small company at the time, had difficulty finding a partner interested in their product. “Some big pharmas thought that COVID was going to go away very quickly, some big pharmas thought vaccines will solve the problem and some big pharmas didn’t even believe there was a pandemic for a while,” he continued. “And at that time we were not exactly very well funded.” The project was further stalled in 2021 due to the limited availability of COVID-19 viruses for use in testing, with governments and big pharmaceutical companies given priority. Insilico used the delay to refine and improve on their drug candidates, eventually narrowing the list down to just one: ISM3312. While frustrating, Zhavoronkov admitted the delay inadvertently led to the discovery of his “perfect” drug and its subsequent approval for human use by the China National Medical Products Administration. That said, had all the resources been available at the start, Zhavoronkov said “I can probably do it all again in under a year.” The company hasn’t given a start date or timeline for the clinical trials, though it promises to start “soon.” “There remains an urgent need for conducting well-designed clinical trials for scientific evaluation of potential COVID-19 therapies such as ISM3312,” said Dr. Sujata Rao, Insilico’s senior vice-president and head of clinical development, in a release. “We are committed to exploring ongoing unmet medical needs to include subgroups under-represented in COVID-19 trials as well as in patients who may have variants that are more resistant to antiviral drugs.” Insilico isn’t the only company utilizing AI to treat COVID-19, but they may be the furthest along. For example, company Exscientia AI limited is currently in the discovery phase of a new COVID drug using a different AI system, in a project wholly backed by the Bill and Melinda Gates Foundation. Given the promise of their new drug, Zhavoronkov said it may be a “very long time,” if ever, before Insilico looks at making another COVID-19 therapeutic. “It might make sense developing maybe one more drug for COVID, also AI generated. But that would come much later — it will take a while for resistance to develop against mine.”