Call it the Goldilocks approach to designing cancer drugs. Opinion has been growing that if traditional chemotherapies kill tumour cells indiscriminately but are highly toxic, and if newer drugs developed to pinpoint specific protein kinases in tumours provide a dramatic response, but in a smaller proportion of patients than hoped, then a balance might be achieved by designing drugs that hit not so many of these kinases as to be toxic but not so few as to be ineffective.

So what does the FDA approval of the first of these new multikinase inhibitors — Bayer and Onyx's Nexavar (sorafenib), approved for renal cell carcinoma at the end of 2005 — mean for the 'magic bullet' approach that brought us successes such as Gleevec (imatinib; Novartis)? Many researchers feel that because tumours are often complex mixtures of cells in various stages and states it is better to develop drugs that inhibit more than one kinase. “Many researchers would now think that no one target alone is probably going to be useful in all tumours of a class due to the great deal of heterogeneity that exists as well as the multiple ways a tumour cell can overcome any given insult,” says Roy Herbst, Professor of Medicine and Chief of Thoracic Medical Oncology at the University of Texas MD Anderson Cancer Center.

Others say such statements are premature. “We can't say there should be a shift towards creating multikinase inhibitors at the expense of more specific inhibitors as we don't yet know whether the former drugs work by inhibiting more than one target in the tumour,” says Carlos Arteaga, Professor of Medicine and Cancer Biology at Vanderbilt University, Nashville, Tennessee. Data on Nexavar suggest that its effects in renal tumours are likely to be due to inhibition of the vascular endothelial growth factor (VEGF) receptor, despite the fact that the drug was initially developed as a RAF kinase inhibitor, says Arteaga.

“It is compelling to think that you could get a better response if you hit two targets instead of one,” says Charles Sawyers, Professor of Medicine at University California at Los Angeles. “I'm not against the idea; I just want to see the evidence that this is true.” Gleevec, which Sawyers was instrumental in developing, turned out to be a multikinase inhibitor. The drug was designed to target ABL in chronic myeloid leukaemia, but inhibits another kinase, KIT, in gastrointestinal stromal tumours. “No one planned that in advance, I don't know how you really could,” says Sawyers.

There are too many unknowns now to say how many targets future cancer drugs should block, says Charles Baum, Clinical Oncology Leader at Pfizer. “Evidence suggests that one candidate we are developing called sunitinib, which inhibits the VEGF and PDGF receptors, provides added therapeutic benefit in renal carcinoma than inhibiting VEGF alone,” says Baum. “But for cases where we really know the specific molecular target well and there's one target that's driving the disease, then a single targeted agent is probably appropriate.”

Understanding the molecular events that drive tumours and how any given drug is therefore working is crucial, but our knowledge in this area is far from complete, says Herbst. “We need to figure out what pathways lead to a drug response/improved survival, cause resistance or a lack of effect to a drug, and from this we can then hopefully understand what combination of treatments will have a better therapeutic effect,” says Herbst. Identifying the driving mutations in the more common cancers is still a challenge. And despite improvements in techniques, getting workable biopsies of tumours is tricky.

But just as the approval of Gleevec allowed researchers to investigate how the drug works and how resistance occurs, so Nexavar's approval, and possibly more to come (see box), will help researchers pinpoint the important mechanisms in tumour response to multikinase inhibitors. How the future of cancer drug development will be determined is perhaps best summed up by the title of Herbst's targeted therapy programme at the university: 'From the lab to the clinic and back again'.