Fragmentary Solutions Astex Therapeutics Puts the Pieces Together

Abstract

Sunesis Pharmaceuticals (www.sunesis.com), a 130 person company located in South San Francisco, focuses on cell-cycle inhibitors for cancer. Sunesis calls its fragment-based discovery approach “tethering” and is focusing on weak binding fragments by looking at the binding event itself. Sunesis also uses another approach called extended tethering, where fragments are linked by a covalent bond to the target protein. “In one instance you fish off the bank and the next you fish off the pier,” said Dr. James Wells, professor of pharmaceutical chemistry and cellular and molecular pharmacology at UCSF and Sunesis cofounder.Sunesis has three compounds in the pipeline; SNS-595, in phase II clinical trials; SNS-032, in phase I clinical trials; and SNS-314 (preclinical.) The company also has five strategic collaborations with Biogen Idec, Johnson & Johnson PRD, and Merck. Founded in 1998, Sunesis raised $200 million from various investors and went public in 2005. “I think it [fragment based discovery] is complementary to high-throughput screening,” said Robert McDowell, Ph.D, vice president, Discovery Chemistry.On the other side of the Bay, Plexxikon (www.plexxikon.com) is a 65 person company located in Berkeley. Plexxikon starts with fragments between 150 and 350 Da in molecular weight, then screens them using biochemical assays, looking for very weak activity. Subsequent cocrystallographic structural analysis sorts the wheat from the chaff. Sometimes, to get additional interaction information, Plexxikon employs diverse techniques like surface plasmon-enhanced Raman spectroscopy and NMR to weed out the false positives from each technique. “You use an assay just to identify a measure of biochemical activity at high concentration,” said Dean Artis, “We really let the crystals decide what is an interesting compound versus a confusing assay artifact. You get a very small, compact starting point.”Following this methodology, the company develops a “scaffold” that has binding affinity for multiple members of a protein family. Currently, Plexxikon uses several scaffolds to target each of three protein families (kinases, nuclear receptors, phosphodiesterases). Artis commented that Plexxikon has filed INDs for molecules generated using this approach. According to Artis, candidate compounds for preclinical testing can be generated in as little as 3–6 months. The company initiated collaborations with Genentech, Inc., in 2003 and Wyeth Pharmaceuticals in 2004.“I came to Plexxikon to make drugs,” Artis said. “This process gives you a very efficient path to do not just an exercise in technology, but to do drug design in a time frame that might actually help people sooner rather than later. Our first drug is in patients, and we've only been doing chemistry for four years.”What is the view of big pharma on this method? “We typically use it as the complementary method to HTS,” said Dr. Andreas Marzinzik, group leader, Integrated Lead Discovery Program, Novartis Institute of Biomedical Research, “we would really want to focus on targets where HTS delivers no promising hits.” Marzinzik cited, for example, non-ATP competitive inhibitors of kinases, which are hard to identify by high throughput screening. “The advantage is really that FBS technology is not prone to false positives.” According to Marzinzik, Novartis assembled a working group to advance fragment-based discovery technology in 2005. Novartis has an internal project and also collaborates with outside companies for specific projects. “It is definitely a tool for us here at Novartis,” Marzinzik said.It remains an open question, after the iterative optimizing of weak leads, if FBDD really shaves off that much time and effort from the discovery process. But it could pick up a few unprepossessing fragments that might grow up to be great drugs if they just got a bit of attention.