Assistant Managing Editor

Nimbus Discovery LLC came out from behind the clouds earlier this month when its computational drug discovery model attracted a high-profile investor: Microsoft's Bill Gates.

Gates is no stranger to biotech. He previously invested in ICOS Corp. (acquired by Eli Lilly and Co. in 2007), and the Bill & Melinda Gates Foundation has provided substantial funding for translational research and drug development in emerging territories. And last year, he committed $10 million to Schrödinger, a 20-year-old Columbia University spinout that offers a range of informatics and drug discovery software to the biotech and pharmaceutical industries.

It was his interest in Schrödinger that connected Gates to Cambridge, Mass.-based Nimbus. The small biotech was founded in May 2009 by Atlas Ventures, with help from Schrödinger, which wanted to put some of its technology in the hands of a drug discovery firm. Schrödinger had licensed use of its tools here and there, but "they found it frustrating that no company was taking advantage of the full suite of software," said Jonathan Montagu, vice president of business operations at Nimbus.

So Nimbus was created with that goal in mind. "We're essentially translating that [software] into medicine," he told BioWorld Today.

The technologies from Schrödinger include the use of cloud computing as well as the WaterMap technology, designed to improve drug potency and selectivity by taking into account water molecules at binding sites.

"Previous approaches have ignored the effects of water," Montagu said. But those water molecules can get trapped in proteins and make them unstable. It's really about understanding what drives drug binding and being able to "look into the inner workings of protein drugs," he added.

Nimbus' lead programs go after targets that historically have been difficult to hit. The company aims to have a development candidate ready by the end of this year and an investigational new drug application by the end of 2012 for a lead candidate against IRAK4 (interleukin-1 receptor-associated kinase 4).

A kinase that regulates immunity, IRAK4 is believed to be crucial in childhood to help protect against bacterial infections, but in adults it can be linked to autoimmune and auto-inflammatory diseases such as gout, lupus and inflammatory bowel disease. It's also a target in diffuse large B-cell lymphoma (DLBCL), specifically activated B-cell-like ABCDLBCL.

But IRAK4 is a difficult kinase to work with, said Rosana Kapeller, chief scientific officer, in part because the "binding site is very shallow, so it's hard to find a handle." But Nimbus' technology has allowed it to develop very selective compounds, she added. "It gives us a road map. We know exactly where we can go and where we cannot go."

Because ABCDLBCL is a more manageable indication for a small firm – and because it's possible to demonstrate clinical efficacy in Phase I testing – Nimbus will start with that indication first. Pending successful data, the firm likely will seek a partner to take the drug into further development.

Montagu said the firm aims to fund the drug only through the end of Phase I. "We'd like to try to exit during early development."

The same goes for its second program targeting acetyl-coA carboxylase (ACC), an enzyme associated with metabolism. Inhibiting the enzyme can actually mimic exercise, Kapeller said. making it an obvious target for treating obesity and diabetes. It also is emerging as a target in cancer metabolism, particularly in solid tumors and there has been "some really beautiful data in breast cancer," she told BioWorld Today.

But ACC is tricky because it's a bid enzyme with multiple domains. "The pocket is very big and so compounds end up very hydrophobic," Kapeller said. Nimbus' method focuses on an allosteric site, and the firm is working to select potential compounds for testing.

The company has a third program, though it has not yet disclosed the target and that work is at an earlier stage. Beyond that, "we have a portfolio of 20 targets to look at," thanks to the firm's work with Schrödinger, Kapeller said. Those programs can be developed down the road by Nimbus, or the firm might seek pharma collaborations.

"We're really trying to create a new drug discovery model," Montagu said. Besides its link to Schrödinger, Nimbus has something else to set it apart from traditional biotech start-ups. The firm is set up as a holding company, designed to keep each program in a separate entity. It's not quite a project-financing model, since the company's financing – like its recent seed round co-led by Gates and Schrödinger co-founder Richard Friesner – goes to the holding company and "and we push it down" into the individual projects, he explained.

"Normally investors put money in your company and hope 10 years later that somebody buys you," Montagu said. By placing each asset into a separate unit, Nimbus stands a better chance of being able to monetize each one through collaborations or acquisitions.

Only a few other firms, such as Adimab LLC and Ablexis LLC, have employed similar models, "but I think we'll see a lot more," he added.

Nimbus, which has two full-time employees, but about 30 external advisors and consultants, expects to remain virtual going forward. That "gives us a lot of flexibility," Kapeller said. "We can get the best talent, independent of geographic location."