Verseon Corp – A New Twist on In Silico Drug Design

By Michael Goodman

March 18, 2016 | For the past decade, a few companies have been pursuing the application of quantum mechanics to the modeling of ligand/receptor interactions. San-Francisco-based Verseon Corporation may have a platform in physics-based computational drug design that could change the economics and the hit quality of drug discovery.

Launched in 2002 by the founders of Pulsent Corp, a video compression specialist, Verseon is predicated on the idea that a more accurate method of modeling protein-drug interactions could replace the large and costly process of high throughput screening (HTS) by exploring a larger chemical space and producing novel compounds with higher affinity for their target.

Corporate compound collections represent about 4-6 million distinct compounds—a fraction of the drug-like chemical space that could potentially be synthesized. Moreover, because of the presence of analogues and variations on core chemical structures, corporate collections are not very chemically diverse. Verson holds that there are simply not enough medicinal chemists alive to make a significant dent in the accessible drug-like chemical space by making more compounds. Nor is there enough computing power to solve protein-drug systems by brute force. More efficient computational models are needed.

The Platform

Verseon’s proprietary discovery platform consists of two basic components: the Molecule Creation Engine (VMCE) and the Molecule Modeling Engine (VMME).

The VMCE generates large numbers of virtual, novel, drug-like, and synthesizable compounds. It also provides synthesis guidelines for each virtual compound in the form of a recipe for manufacture that can be modified by medicinal chemists. For a typical disease program, Verseon’s platform screens about 10⁸ virtual compounds from the VMCE against a target protein. It models this interaction through the VMME. Using sophisticated optimization algorithms in an attempt to bypass many unfavorable binding states and thus lower the computational load, the VMME assesses the likelihood of each Verseon virtual compound to bind to the (virtual) target protein in an (virtual) aqueous environment. Why in water? Because it makes up approximately 60% of the human body.

The VMME typically performs more than 10¹¹ compute operations per tested compound against the target, capturing all conformational changes in the binding process. It does so across a large, dedicated private computing cloud. In order to predict the binding mode and estimate the binding affinity with reasonable accuracy, it is necessary to model various factors—electrostatics, solvation, hydrogen bonding—that contribute to the affinity between ligand and receptor.

This is where quantum mechanics comes in. As Verseon co-founder and CEO Adityo Prakash puts it: “We had to build the physics insights over 12 years. This is not Big Data; this is hard science. How do you model how something will bend, flex, vibrate, twist—what the actual binding affinity will be, what the entropic effects will be? That’s how you get to the heart of how it’ll really behave in nature. This is key. If you can replace the live experiment with a very high fidelity computer simulation, which is what has hamstrung the field for the last few decades, then you can actually do this.”

The output of the VMME is a collection of hits prioritized by binding affinity. The VMCE can be used to generate a cluster of closely related variants to prevent the entry of single agents into the downstream discovery process. Next, the Verseon hit is sent to the medicinal chemists for investigation and potential synthesis. Med chem selects a subset of hits to submit for synthesis. Once synthesized, Verseon hits are subjected to a battery of in vitro and ex vivo assays; the survivors are promoted to lead candidate status and then undergo preclinical pharmacokinetic, toxicity, and efficacy studies. The lab work is seamlessly integrated with the front-end computational design.

The decision as to which compounds to advance further is helped by having multiple lead candidates of differing chemotypes per each program.

The platform has been rigorously validated. Prakash and his co-founders went around showing it to people like Robert Karr, then SVP of R&D Strategy at Pfizer and John Leonard, then global head of R&D at pre-split Abbott Labs. (Both Karr and Leonard sit on Verseon’s scientific advisory board.) Those men basically proposed hundreds of test cases based on experiments they’d done in which they selected a drug and a target, and challenged the platform to predict how they would bind. In each case, they knew the outcome beforehand. And in each case, the platform’s prediction matched the experiment.

Kennie Merz, Ph.D., director of Michigan State University’s Institute for Cyber Enabled Research and a veteran of Pharmacopeia’s R&D group (acquired by Ligand in 2008) is impressed by Verseon’s progress, but has a caveat. Verseon has been reluctant to publish until its IP protection is locked in. In the absence of a peer-reviewed publication, or external validation of its platform (e.g., the D3R Grand Challenge) Merz says—quoting the Missouri license plate—“Show me.” 

The Portfolio

Verseon’s portfolio includes three programs, all novel, wholly owned products of its platform, in various stages of discovery and preclinical development. All are small molecules targeting moderate-large population diseases.

• An antithrombotic agent against stroke/venous thromboembolism/atrial fibrillation/acute coronary syndrome.
• A plasma kallikrein inhibitor against diabetic retinopathy and diabetic macular edema.
• A non-VEGF angiogenesis inhibitor against solid tumors.

The anticoagulant is the most advanced. Currently in IND enabling studies, Prakash expects it to enter the clinic this year.  He adds, “The anticoagulant candidates were identified and produced at a fraction of the cost of a typical drug discovery program. They represent novel chemical matter that can’t be found in existing chemical libraries.”

The current crop of novel oral anticoagulants (NOACs)—Xarelto, Pradaxa, Eliquis—carry the risk of major bleeding as well as serious side effects including rebound stroke or heart attack on cessation of therapy. Verseon’s anticoagulants, because of their unique chemical structure and novel mechanism for direct thrombin inhibition, have been shown in preclinical animal models—both efficacy tests and bleeding risk tests—to have a superior safety profile to the current NOACs and a substantially reduced risk of major bleeding.

The Company

Verseon was enabled by continuing improvements in the price-performance of computing. The company’s three founders are all physicists; expertise in R&D planning, discovery biology, business development, and computational chemistry fill out the management team. The animating idea was that physicists—essentially “outsiders”—could bring a fresh perspective to the problems that have bedeviled drug research.

The three founders invested in the company at its inception. Karr, Leonard and a few other undisclosed investors provided external funding later on, once the platform was validated. The British investor Neil Woodford helped finance Verseon’s $100 million debut on the London AIM market in May 2015 (AIM is the London Stock Exchange’s international market for growth companies, similar to the U.S. NASDAQ).

The AIM listing ran counter to the flow of European biotechs flocking to the U.S. exchanges. Though not as liquid as the NASDAQ, Prakash said the AIM market was right for Verseon in other ways. Some of its major investors were UK-based and had restrictions about investing in private companies. Verseon had no interest in having to comply with Sarbanes-Oxley or the onerous reporting requirements of U.S. exchanges. AIM seemed more hospitable to early-stage companies bent on steady, systematic growth.  “For now, for us, it was the right platform. We’ve got a lot of investors with a fairly long time-horizon for their investment that matched our outlook on the company’s growth arc. And we’ve always the option of moving to another exchange down the road.” said Prakash.