Insilico Medicine announced the nomination of a novel preclinical therapeutic candidate for treating COVID-19, designed using the generative chemistry AI platform Chemistry42. The new drug candidate is a 3CL protease inhibitor unique from existing drugs in its class because it can be rapidly produced. While this nomination is a potentially important development for the ongoing COVID-19 pandemic (if successful in clinical trials), what may be even more strategic is the conceptualization of how AI-driven drug discovery and development can be a rapid way of addressing potential future pandemics, with potentially more lethal pathogens than COVID-19.

The coronavirus that caused the global COVID-19 pandemic is a positive sense, single stranded RNA beta coronavirus, a member of Beta-CoV lineage B (subgenus Sarbecovirus), possessing high likelihood for human-to-human transmission. The RNA sequence is around 30 kb in length. There is an insightful technical report by Innophore, summarizing information about the coronavirus: its structure, active sites, sequence data, and so on.

The unprecedented speed at which the new coronavirus spread across the globe, disrupting economies and the global healthcare system, revealed the world's unpreparedness for these kinds of challenges. Healthcare facilities and hospitals were quickly overwhelmed. The world's inability to react quickly to a pandemic of this size and speed led to a detrimental output: more than half a billion cases of COVID-19 and more than 6 million deaths, according to WHO statistics.

One of the key reasons for the high death toll was the absence of an efficient antiviral remedy at the time the first cases of COVID-19 rapidly spread. The pharmaceutical industry has a notoriously slow business model, where a typical drug discovery program can take many years to successfully progress to a ready-to-prescribe drug. Another standard bottleneck is manufacturing and logistics, even if the innovative solution does become available quickly. For the most part, the industry was largely unable to come up with a rapid solution to stop the pandemic at the very beginning, so it quickly spiraled out of control. The antiviral options available at the time the pandemic struck proved to be of low effectiveness.

When dealing with events like the COVID-19 pandemic, time is paramount. The amount of infected people grows exponentially and they need efficient antivirals and vaccines sooner rather than later. Artificial intelligence (AI) is a proven way to dramatically accelerate research. No wonder, when the pandemic struck, people turned to artificial intelligence for various needs associated with the health crisis -- including drug discovery and vaccine development, optimizing production and logistics processes, epidemiologic modeling and prediction, high-throughput diagnostics of patients at scale, and sorting real world data for medical treatment demanding forecasting.

A vivid example of the power of AI-driven research was demonstrated by Moderna (NASDAQ: MRNA), one of the most "digitized" biotechs. A lucky convergence of technologies, such as mRNA itself and its delivery method via LNPs – supported by advanced digital tools and predictive AI algorithms – allowed Moderna to develop a successful COVID-19 vaccine in months. Using technology Moderna produced a vaccine in record time and more than 200 million doses of its vaccine were administered in the US alone.

But vaccination alone is not enough to tackle the crisis. Since the beginning of the pandemic, the global healthcare system has been in desperate need for efficient antivirals to treat patients infected by the coronavirus.

The first strategy offers a shorter path to the public, which is important. In fact, most of the immediate drug discovery efforts in the early stages of the pandemic were focused on drug repurposing of known clinically approved drugs and virtual screening for the molecules available from chemical libraries. However, this method proved to be of limited efficiency. For example, the IC50 of lopinavir, an HIV protease inhibitor, against the 3C-like protease was found to be approximately 50 micromolar, which was far from ideal.

The second strategy, de novo drug design, is a more complex process, but it can provide a much better eventual solution -- an efficient first-in-class or best-in-class antiviral that can save lives of COVID-19 patients with severe forms of the disease. In an attempt to come up with a bold solution for both the current pandemic and for future pandemics, Insilico Medicine chose to pursue this route. On January 28, 2020, Insilico utilized part of its generative chemistry pipeline Chemistry42 to design novel drug-like inhibitors of COVID-19 and began generation on January 30.

Insilico wanted to choose the right kind of target for the design of a corresponding molecule. Many potential therapeutics aimed at containing the spread of SARS-CoV-2 have targeted the S, or spike, protein, a surface protein that plays a vital role in viral entry into host cells, since that is the approach that was taken with both SARS and MERS coronaviruses.

However, according to Insilico's study in collaboration with Nanome, published back in 2020, two-thirds of the SARS-CoV-2 genome comprised non-structural proteins, such as the viral protease (the protein necessary for viral replication). Insilico therefore concluded that such alternative potential targets should not be overlooked, and decided to focus efforts on C30 Endopeptidase, also referred to as the 3C-like proteinase or coronavirus 3C-like protease (3CLP) or coronavirus main protease (Mpro ). 3CLP is a homodimeric cysteine protease and a member of a family of enzymes found in the Coronavirus polyprotein.

Image source: Wiki

In February 2020 Insilico published the first structures of small molecules targeting the key protein SARS-CoV-2 3C-like protease, and later selected a pool of compounds to be synthesized and tested using the company's own resources and the resources offered by its closest partners.

Three months later, Insilico Medicine published further results of its AI-driven COVID-19 drug discovery efforts, this time using 6W63 crystal structure of Mpro complexed with a non-covalent inhibitor. The company combined two molecular design approaches: ligand-based and crystal structure–based and published 10 representative structures for potential development. In collaboration with Nanome, Insilico conducted a medicinal chemistry evaluation of selected ligand-protein complexes using a virtual reality (VR) environment as a collaboration medium.

Today, following Insilico's initial pioneering efforts in designing COVID-19 antivirals, the company is announcing a new preclinical candidate (PCC) for the treatment of SARS-CoV-2 developed using Insilico's proprietary AI-driven generative chemistry approach. The drug candidate is a novel 3C-like (3CL) protease inhibitor.

The 3CL protease is a protein that is translated early in the COVID-19 viral replication process – without which the virus cannot replicate. Recent studies have shown that other drugs designed to inhibit this protease possessed robust efficacy and safety in clinical trials. Insilico Medicine's PCC is positioned to advance this class of small molecule inhibitors even further with a drug candidate that is orally available, effective in low dose, and can be synthesized quickly. It is also advantageous in that there is no need for co-dosing with Ritonavir. The PCC with novel small molecule structure was designed using Insilico's proprietary machine learning platform Pharma.AI capable of rapid, highly-automated end-to-end drug discovery.

Pharma.AI consists of three main pipelines: target discovery (PandaOmics), small molecule drug discovery (Chemistry42), and predictors of clinical trial outcomes (InClinico). This system is designed to achieve maximum automation of drug discovery processes for a broad range of human diseases. The small molecule drug discovery pipeline can be used to generate inhibitors of bacterial and viral protein targets. Multiple publications explaining the basic concepts and approaches in generative chemistry were published by the team. Since there is a known protease target for 2019-nCoV and its sequence and structure are also known, Insilico decided to apply only the generative chemistry pipeline Chemistry42 to perform this drug design project.

Despite the fact that Insilico did not have its own laboratory space for safely handling class 3 airborne pathogens like SARS-CoV-2, the program reached PCC stage in record time thanks to Insilico Medicine's highly sophisticated AI platform and extensive network of external partners and collaborators.

The key danger of highly contagious pathogens is the rapid dynamics of a disease spread in the population, causing high mortality rates. This was true with the devastating plague events in middle ages, the Great Influenza epidemic of the 1920s and with the ongoing COVID-19 pandemic. The ability to develop novel antiviral drugs or antibiotics rapidly, based on the relevant data about a novel pathogen, is the key necessity. This can be achieved with an AI-driven drug discovery approach, when timelines for the early drug design phase can be reduced to months or even days, as opposed to the multi-year process of "traditional" approaches.

Insilico Medicine has created and extensively validated the end-to-end AI-system, Pharma.AI, which is capable of accelerating drug discovery for various indications at a fraction of a typically incurred R&D cost. Since 2021, Insilico has delivered seven preclinical candidates discovered and designed using its AI platform in a variety of disease areas, including fibrosis, inflammation, and oncology, including one for the QPCTL immuno-oncology target in partnership with Fosun Pharma. The company has also successfully completed a Phase 0 microdose trial and entered a Phase I clinical trial with its first internally developed program for fibrosis.

The present nomination of the 3CL protease inhibitor to treat COVID-19 represents another proof of concept for Insilico's Pharma.AI platform and research approach, involving an internal team of medicinal chemists and an extensive network of external collaborators and research partners. This platform is optimized to address major health challenges of the future, including new pandemics.

Given the availability of modern technologies to quickly generate data about any novel pathogen (e.g. next-generation sequencing (NGS), Cryo-EM, etc.) -- it took just days for scientists to get full data about the genome of SARS-CoV-2 and make it available to the global research community to enable follow-up efforts. Soon it will become feasible to rely on advanced AI-driven target discovery and drug design systems, like Pharma.AI, for ensuring a more rapid and proactive response to any pandemics. The collective job of the drug discovery community is to prevent the same delayed reaction to emerging health threats as the world experienced in early 2020 with COVID-19.

However, only well-validated platforms, like Pharma.AI, can be regarded as a reliable line of defense against novel emerging pathogens. It takes years for a sophisticated AI-system to be developed, and it requires numerous projects to prove it can provide reliable output and high quality results -- novel efficient therapeutics.