Featured Article - April 2010
Short description: Structural genomics has an important role to play in guiding development of specific anti-cancer drugs.
Protein kinases currently constitute the largest target family for cancer drug discovery. Ten small-molecule kinase inhibitors have been approved for cancer treatment, but only a few are reasonably specific. The others inhibit multiple pathways, resulting in off-target effects that relegate them to end-of-life treatments.
A commentary in Nature Chemical Biology by Stefan Knapp and colleagues from the Structural Genomics Consortium Oxford explains the challenges — and possible solutions — in obtaining highly specific kinase inhibitors and explains the important role structural genomics plays. 1
Kinase inhibitors are performing well in phase I clinical trials, with an attrition rate of 53%, which is respectable in comparison with the overall antitumour drug failure rate of 82% at this stage. The challenge now is to improve this rate by producing inhibitors with higher specificity and preventing cross-reactivity.
But both the publication record and the patents resulting from research in this area are skewed towards the 25% of human kinases that are reasonably well understood. Of the remainder, 25% have a completely unknown function and the remaining 50% have only minimal characterization. Yet these unstudied or barely studied kinases are frequently mutated in cancer and have been identified in kinome-wide RNA interference knockdown studies, suggesting that these targets have a pivotal role in disease development.
Many of these uncharacterized kinases have entered specificity screening, and the results so far indicate that they are strongly inhibited by current clinical or preclinical inhibitors, indicating a cross-reaction. In addition, common kinase inhibitors used as chemical probes in cell biology studies have been found to target several proteins at once, calling into doubt some of the conclusions of those studies. 2
This strong inhibition of distantly related kinases by a single drug is puzzling at first glance. Most of the inhibitors target the ATP-binding site but many of the kinases that are inhibited by the same compound have only 20% or less sequence similarity at this site. But from the structures that have been solved so far, it is clear that a few conserved interactions are mediating the binding of most inhibitors.
The key features of the kinase active site include a hydrophobic purine-binding cavity, a hinge segment that connects the two lobes of the kinase and a backbone hydrogen that anchors ATP-mimetic ligands. The active state of kinases is rigid and well conserved within the kinase family, leading to a high degree of cross-reactivity.
By contrast, the inactive state has a much wider range of conformations and is more promising for achieving high specificity — the anticancer drug lapatinib targets a unique inactive conformation, for example. It is likely that many inactive conformations exist.
High-resolution crystal structures would greatly help the rational design of selective kinase inhibitors. So far, 138 structures of human kinases are available and high-throughput structural genomics methods are contributing significantly to this drive.
The structural biology approach goes hand-in-hand with developments in high-throughput sequencing focused on a comprehensive characterization of the genetic mutations that occur in cancer. This combination of high-throughput approaches should boost kinase drug discovery and cancer treatment.