Tech Test-Drive

Drug companies are under pressure to improve the performance of their innovation engines at a time when they are also trying to cut their R&D spending. Therein lies a paradox: The most innovative areas of science are often also the financially riskiest. So how can a company access cutting-edge technology without breaking R&D budgets?

GlaxoSmithKline thinks it has a solution. In 2005, the company created the Centre of Excellence for External Drug Discovery (CEEDD) as a way to tap into edgy, early-stage science that GSK wouldn’t be willing to explore on its own.

CEEDD takes an options-based approach, whereby GSK pays a modest amount up front to a partner company for the option to later license or acquire a technology or product. The model “allows GSK to test-drive technologies in a risk-sharing way rather than being fully exposed to the risk of failure,” Rob Aboud, CEEDD’s vice president of strategy development, recently told reporters at a briefing in Cambridge, Mass.

Five years into the “externalization” strategy, GSK is seeing the initial fruits of its labors. The first drug discovered through a novel screening technology developed by Praecis Pharmaceuticals—a Waltham, Mass.-based firm that was one of CEEDD’s early partners—is about to enter clinical trials. And the technology is now being deployed across a variety of drug targets, where it is proving complementary to more traditional screening methods.

The externalization strategy is both a cost-saving exercise and an acknowledgement that not all good ideas come from inside big pharma. Biotechs are often deeply specialized in areas where big pharma companies are not, said Patrick Vallance, GSK’s senior vice president of medicines discovery and development.

CEEDD now boasts 40 programs across 25 disease areas. Examples of partnerships include a stem cell cancer therapy collaboration with Oncomed Pharmaceuticals and a deal to improve drug properties using Concert Pharmaceuticals’ deuterium chemistry expertise.

Praecis provides an example of how the approach ties into GSK’s ambitions to save costs, expand its pipeline, and keep up with leading-edge technology. GSK signed on to run a pilot program with Praecis’ compound-screening method in 2006. Praecis quickly found promising hits for two targets provided by GSK, and in early 2007 the British firm acquired it for $55 million.

Praecis’ encoded library technology (ELT) attaches specially designed DNA bar codes to chemical building blocks and then applies traditional combinatorial chemistry methods to generate enormous libraries. Each 20 million-compound library takes three to six months to make but, amazingly, fits in a tiny microcentrifuge tube. The compounds are then simply mixed with an equally small tube containing a target protein.

Scientists use affinity chromatography to pull out the resulting hits, but they still need to figure out the structure of the molecules bound to the target. For that, the encoded DNA is sequenced, providing a readout of how the molecule was built.

The small quantities of target protein used make for an efficient process. In the past, the high price of sequencing DNA was by far the biggest expense of using ELT. With the price for sequencing a million base pairs plummeting, the technology is suddenly a cost-effective way of finding starting points for drug design.

In their first few years in the GSK fold, the Praecis scientists focused on proving the technological capacity of their DNA-encoded libraries. “We’ve now demonstrated that and are beginning to learn how to apply it and integrate it with more established methods,” said Barry Morgan, GSK’s vice president of molecular discovery research and an inventor of the technology platform.

One lesson is that ELT produces fewer but higher quality hits than does traditional high-throughput screening (HTS), Morgan said. Furthermore, researchers are finding that ELT can complement HTS. ELT looks at a molecule’s affinity for a target, whereas HTS looks at its ability to affect the function of a target. When combined, the methods are particularly helpful in designing drug candidates for new target classes, Morgan noted.

Another difference is that HTS files through a range of single compounds, whereas ELT interrogates a relatively small number of scaffolds decorated with a large number of substructures. As the former Praecis scientists begin to build libraries using chemicals from GSK’s small-molecule collection, compounds in ELT libraries start to take on characteristics of those used in HTS screens. By comparing the results from HTS and ELT campaigns, scientists can use pieces of structures to build composite molecules that are highly potent and selective, Morgan said.

Still, there are targets for which only ELT produces hits and others where HTS is the only source of hits. Having access to both technologies broadens the range of targets GSK can pursue, he noted.

Praecis has applied ELT to more than 140 GSK drug targets, including metalloenzymes, kinases, cell- or virus-surface receptors, and protein-protein interactions. In January, GSK expects to put its first molecule discovered through the novel approach into clinical trials.