A future in fragments

Astex’s research on drug fragments is taking it deep into the oncology field, as Emma Davies finds out.

Astex, a UK-based drug discovery and development company, is contemplating a public offering in the near future. With clinical trials for cancer compounds on the horizon and a sturdy technology to back up its science, Astex may well attract investors who have grown weary of investing in the risk-laden biotechnology industry.

Astex was founded in 1999 by Harren Jhoti, and Sir Tom Blundell and Chris Abell from Cambridge university. Jhoti had a background in commercial drug discovery, which included time spent as head of Glaxo Wellcome’s structural biology and bioinformatics group. He was utterly convinced that X-ray crystallography technology could be used to enhance the drug discovery process. Fortunately, Blundell and Abell shared his view and a vision soon emerged of novel, high-throughput X-ray crystallography equipment which could be used to determine rapidly protein-ligand crystal structures. It was not long before discussions with financial backers started; these were very promising and Jhoti was reassured that ’there was a good business opportunity in the making’.

As might be expected with a new company, the early days of Astex were not easy. Before the company had its own premises, Astex funded post-doctoral researchers in both Blundell’s and Abell’s labs (and to this day the university has a small equity stake in the company).

However, things progressed rapidly and it wasn’t long before the company had moved into buildings on the Cambridge Science Park. Four years later and Astex was already upgrading to purpose-built facilities on the Science Park. At the opening of the new buildings, Timothy Haines, Astex chief executive, described how the previous facilities had become very cramped and spilled over into mobile units. ’We now have a much better environment in which to pursue our goal of developing small molecule therapeutics using our fragment-based discovery approach,’ he exclaimed.

With a so-called Pyramid approach, the researchers use virtual screening to help identify small molecules which could feasibly fit in the active site of the protein that they have chosen to work on and to estimate the docking modes of the molecules in protein-binding sites. Any promising fragments are taken to the next stage, which involves a bit of wet chemistry: researchers isolate ultra-pure protein crystals and soak them in a solution containing a selected drug fragment in the hope that the molecule will wend its way to the enzyme’s active site. At this point, the high-throughput X-ray crystallography comes in and the fully automated system takes over, carrying out a full structural analysis of the crystal and anything that is locked inside it.

Proprietary software called AutoSolve analyses the electron density maps that are created by Pyramid to determine whether a drug fragment has slotted into an enzyme’s active site. The X-ray crystallography not only gives a clear picture of the initial fragment binding into the target active site but also indicates the orientation of the drug fragment and which interactions are responsible for the binding.

A positive binding result is known as a hit and is plugged back into Pyramid to help select extra functional groups that could improve binding or selectivity. The enhanced fragments then go back to X-ray crystallography for further screening. The trick is to add no redundant groups to the molecule.

All of the data from the various stages of the drug discovery process are put on an internal web site, which enables the teams of Astex researchers, from computational and medicinal chemists to biologists, to work closely together to generate lead compounds.

Astex made a ’landmark’ discovery of the first crystal structure of a human CYP450 (2C9) in December 2001. It was not long before the company was announcing that it had solved the structure of a second isoform, CYP 3A4. Then, last July, Astex researchers published a paper in Nature, describing the interactions between CYP 2C9 and warfarin, a widely-used blood-thinning drug (P A Williams et al, Nature, 424, 464). The enzyme can prevent warfarin from working by catalysing the hydroxylation of its active enantiomer. The paper reveals how Astex researchers determined the crystal structure of CYP 2C9 complexed with the anti-coagulant and discovered a warfarin binding pocket in the enzyme. It suggests that the protein ’may simultaneously accommodate multiple ligands during its biological function’ and that a drug molecule bound at the warfarin site would be ideally placed to make direct molecular interactions with another drug molecule interacting with the haem group.

Discovering the new warfarin binding site may not only have implications for understanding the mechanism used by the CYP450 family but may give some more insight into complex drug-drug interactions. The company is currently using its proprietary P450 structural information to help it to generate lead compounds with reduced metabolic problems.

The CYP450 work proved to be a bit of a money spinner, pulling in one deal after another, with Astex setting up collaborations with AstraZeneca, Aventis pharmaceuticals, Mitsubishi Pharma, and Fujisawa Pharma based on its proprietary CYP450 structures.

In March 2004, Astex set up an alliance with Boehringer Ingelheim (BI), a German pharmaceutical company, to discover new drugs in undisclosed therapeutic areas. Astex is using its technology to identify lead compound against BI targets and BI will do any clinical development. BI made an upfront payment to Astex, funds the research programmes and will provide milestone payments and royalties.

A year later, Astex signed up to another Institute of Cancer Research and Cancer Research Technology collaboration, this time with the Wellcome Trust. The collaboration is identifying and developing novel drug candidates against BRAF, a gene which can cause cancer when mutated. The mutated BRAF gene has been identified in about 70 per cent of malignant melanomas and 10 per cent of colon cancers. BRAF normally forms part of a chain of control switches that must be ’on’ for a cell to grow and divide. However, when mutated, BRAF is active all of the time and cells can multiply unchecked. The collaboration is trying to develop specific inhibitors to switch off mutated BRAF and prevent certain cancers growing.

In 2003 Astex also set up a multi-target drug alliance with Schering, a German pharmaceutical company, to discover novel drugs for a range of therapeutic indications, but focusing on orally available small molecule drugs to target tumours. Then, last October, Astex took itself further down the oncology road when it announced that it had raised ?23m in cash though a takeover of Berlin-based metaGen Pharmaceuticals, an oncology company. As part of the deal, metaGen’s owners, Schering and two venture capital firms, Apax Partners’ Funds and Hypo Vereinsbank, became shareholders in Astex. Haines said that ’the opportunity to merge with metaGen was compelling given that it provides Astex with the key assets to fuel the next stage of our growth including cancer drug development capabilities and a broader cash basis’.

As Astex steamrolls ahead, it is starting to develop its own drug pipeline and plans to start clinical trials on its lead oncology compounds in 2005 or early 2006. The aim is to take these drug compounds as far as Phase II development and then to partner with a big company.

As little as 10 years ago, training and job opportunities were widely perceived to be better in big pharmaceutical companies than in smaller firms, says David Rees, director of medicinal chemistry at Astex. However, he is currently witnessing a reversal of this trend, with many people now choosing smaller firms like Astex over pharmaceutical giants. Let’s hope that the stock market takes as kindly to the company when it goes public.