New approaches to AI, cell biology may have uncovered a novel drug target for COVID-19: Humans themselves

By combining artificial intelligence and an extensive atlas of protein biology, Cyclica is taking new approaches to searching for old drugs that could fight COVID-19—and, now, it’s helped find what could become a first-in-class candidate.

While other drug and vaccine developers hold up the body’s ACE2 receptor as the lock and the coronavirus’s spike protein as the key that opens up the cell for infection, Cyclica has focused on everything on the other side of the door.

First, parsing the inner workings of human cells—including the mechanisms hijacked by the virus—gave researchers more opportunities to intervene in the entire process.

At the same time, the company is inverting the typical process of computer-aided drug discovery: Instead of taking a single protein target and generating hundreds of never-before-seen molecules that may fit the bill, Cyclica takes known compounds and puts them up against each of the thousands of proteins used by the human body. 

“We flipped the problem on its head and came at it with a different strategy,” Cyclica co-founder and CEO Naheed Kurji said in an interview. “Instead of going narrow and optimizing on one protein, we looked at the entire proteome.” 

These simulations can give a panoramic view of the effects of a particular drug, from its potential success against a disease to any unintended toxicities, Kurji added.

RELATED: AI's hunt for the molecule to stop COVID-19

The combination of these methods helped lead the company to a match: Novartis’ lung cancer therapy Tabrecta, approved by the FDA last year for non-small cell tumors with a particular genetic mutation tied to the cellular signals that can fuel the growth of cancer cells.

Tabrecta’s antiviral potential stems from connections to pathways that help govern acute immune system responses. The drug was plucked from a database of about 10,000 molecules, with most either approved by the FDA or studied in mid- to late-phase clinical trials. 

“We did an all-against-all analysis,” Kurji said, "and specifically focused on the proteins related to infectious diseases like COVID, both on the human side and the viral side.”

RELATED: Cyclica banks $17M to decentralize drug discovery, speed drug development

But Toronto-based Cyclica isn’t just tackling the problem from a different direction. The company hopes to surround it.

In early 2020, Cyclica launched its own “COVID stimulus program” with the goal of building a full-stack AI approach to drug repurposing and enlisted several organizations to help. 

With Ryerson University, the University of Toronto, the Vector Institute, the Hospital for Sick Kids and St. Michael’s Hospital, researchers completed a series of experiments showing Tabrecta could act as a brake on the coronavirus’s infectivity and reproduction. What’s more, the entire approach could be mutation-proof.

“By targeting human proteins necessary for the coronavirus lifecycle, we are developing therapeutic strategies applicable to future strains or variants,” said Bo Wang, Ph.D., lead AI scientist of the Vector Institute, which provided data on potential human targets.

The consortium’s findings are being published early on the preprint server BioRxiv and will be made available for open-source research before they're submitted for peer review.

RELATED: Could a lymphoma drug work better against COVID-19 than Gilead's remdesivir?

“COVID has highlighted how important repurposing discoveries are to the future of medicine,” added Costin Antonescu, Ph.D., an associate biology professor at Ryerson University. “These drugs move to clinical trials and through regulatory approval quickly and can therefore reach patients faster. We’ve found a drug that has demonstrated robust antiviral activity in our model systems and we look forward to getting this data in the hands of pharma.”

Meanwhile, Cyclica’s proteome-wide AI approach could apply to just about any disease, infectious or not, said Kurji. It can also help head off potential complications or side effects that may not become apparent until later—and more expensive—stages of development. All told, the project is helping establish a program that could be useful in a post-pandemic world.

“Infectious diseases are becoming more and more rampant,” he said. “If anything, I think COVID has been a catalyst to the need for rapid, high-quality innovation in the pharmaceutical and biotechnology space.”

This is especially true at the intersection of AI, new computational techniques and biological experiments, Kurji said.

“The silver lining of COVID is that it put an entire spotlight on innovation efforts in our space, and vaccines were brought to the market in 18 months. The entire world rallied around science, and I don't think that's going to stop.”