Predicting quick success in the world of drug discovery is no guarantee, but Predix Pharmaceuticals Inc. thinks it may be on the right path.

Incorporated in November 2000 under the moniker Bio-IT, the company sprung from technology developed at Tel Aviv University in Israel. Focusing on G protein-coupled receptors that have yet to be crystallized, technology developer Orin Becker, in conjunction with Martin Karplus, worked to tackle the protein-folding problem with GCPRs.

"Using a computer, no one has been able to figure why proteins assume the 3-dimensional structures they form," newly appointed CEO Michael Kauffman told BioWorld Today. "In order to do efficient drug development, you really have to know what the drug target looks like in 3-dimensional space. And to date, no one has been able to know that for GCPRs. So there's a huge market - currently GCPRs do around $27 billion - but no one knows what the 3-D structure looks like for any of these targets."

Becker developed the computational technology named Predict that Tel Aviv University licensed to the Woburn, Mass.-based firm. Through the predictive algorithm technology, Predix works to discover lead candidates for GPCR-related diseases based on computer-generated 3-D models of GCPRs. And Kauffman said Predict gets the job done fast - it involves a four-month process to produce hits and another month to screen them.

"Within five months we can start a program, come up with a 3-D structure and then prove that it has biological activity," said Kauffman, who took over the helm at Predix after leaving his vice president of medicine post at Cambridge, Mass.-based Millennium Pharmaceuticals Inc.

It didn't take long for Predix to prove its technology's proof of principle. After licensing the technology, the company spent 18 months developing a suite of algorithms that would do in silico screening and score virtual hits.

"In the last six months we have come out with our first round of what we would call virtual drug candidates," Kauffman said. "The company basically has hit the jackpot."

He said that in that two-year period, the Predict technology produced 3-D structures for 36 GCPRs in eight different families. The company said that coupling the structures with its RISS in silico screening process increases hit rates from the 10 percent to 25 percent range, significantly better than high-throughput screening hit rates of less than a tenth of a percent.

"We start with libraries that are about 1 million virtual compounds, but only about a hundred or fewer compounds actually will be picked by the algorithm, and then we send those off for actual screening assays," Kauffman said. "So we only screen about a hundred compounds."

To date Predix has developed three lead programs, courtesy of an accurate identification process. The screens identified for the NK1 candidate a 56nM compound, a 139nM compound for the 5-HT1A program, and a 278nM compound for 5-HT4.

"Frankly, we were a little shocked to see this," Kauffman said. "Anything less than 500nM for a hit puts you a year and a half into the process, in our estimation, so we're basically at a lead stage."

He said the NK1 compound, which is focused on asthma and depression, would enter animal studies in the next two months. Predix plans to study the 5-HT1A compound's role in depression and anxiety, while the 5-HT4 candidate will be studied in gastrointestinal disorders and Alzheimer's disease. Kauffman said Predix is targeting NK1's clinical debut in 2004.

But in addition to Predix's rapid development pace, the technology allows for an efficient use of manpower. Kauffman said two employees do the virtual computational work, the company outlicenses the screening for a relatively nominal fee, and three chemists then take over the program.

As part of its business plan, Predix already has entered into a development collaboration with an unnamed pharmaceutical partner. Under the agreement, the partner selected a GCPR, to which Predix will apply its modeling and virtual screening process. The partner then will perform binding assays.

"We're looking to partner with two to three additional pharmaceutical and biotech companies, but we're moving our own programs ahead into the clinic," Kauffman said. "The company intends to bring its products through Phase IIa - proof of efficacy - and then outlicense them to large pharmaceutical partners."

Kauffman said the company, which completed its Series A round of financing at the end of last year, is in the process of seeking its second round. Such funding could be directed toward internal growth in its medicinal chemistry and drug development divisions, situated in the U.S. Predix already has added a manufacturing director, but still is searching for a preclinical development director and expects to hire clinical personnel early next year. Predix maintains a computational work facility in Ramat Gan, Israel, a unit Kauffman said appears in place.

"This opportunity really was too good to pass up," he said. "In previous jobs I have tried to make some major innovations in clinical development, as well as to try to link the discovery and preclinical phases much more into the clinical development phases - putting it all together."