Gleevec (or Glivec outside the US) (imatinib) is a good example of a drug that has become appreciated for its wider uses some time after appearing on the market for an orphan disease. Originally it was approved for chronic myelogenous leukaemia (CML), a rare but devastating condition that begins in the often asymptomatic chronic phase, and over the course of several years progresses through an accelerated phase and ultimately to a blast crisis, by which time it is associated with a very high mortality rate. From diagnosis, the median survival period is just over five years.
Gleevec is now also approved for nine indications including gastrointestinal stromal tumour (GIST); but these are still relatively rare conditions. The recent report that the same drug has effects in an animal model of type I diabetes could take the drug into another dimension. Or could it?
Since the discovery in the mid-1990s of the compound class from which imatinib was chosen, the profile of its known enzymatic activities has widened. Imatinib was initially identified in a screen of phenylamino-pyrimidines in a PDGF-R (platelet derived growth factor receptor) suppressor assay by investigators at Novartis. Imatinib is however also an inhibitor of a different tyrosine kinase, bcr-abl, and this enzyme is involved in a specific genetic abnormality associated with CML. The first report of the association came from two researchers at the University of Pennsylvania School of Medicine, who showed that in CML, parts of chromosome 9 and chromosome 22 switch places. This alteration soon became known as the Philadelphia (Ph) chromosome (from where the discovery was made) and could be detected in approximately 95% of patients with CML. The Ph chromosome produces the bcr-abl kinase enzyme that enhances tyrosine kinase activity, and changes the cell’s normal genetic instructions. So far so good, but why does imatinib possess additional enzymatic activity, and what are the implications of this?
Imatinib actually affects the activity of three enzymes. Apart from bcr-abl, the other two are c-kit and PDGF-R. All three are tyrosine kinases, and in each case imatinib occupies the active site that is normally occupied by the substrate ATP, inhibiting the enzyme’s phosphorylating activity. Since all kinases bind ATP, there is bound to be a strong similarity in this site across kinases, so imatinib’s multiple activities are not unexpected. Beyond these three enzymes, the compound is considered to be relatively selective with respect to other kinases.
Actually, the complex profile of imatinib is likely to be responsible for its utility in GIST. C-kit gene gain of function mutations are important in the pathogenesis of GISTs. Evidence links imatinib’s clinical efficacy against GIST with specific mutations in c-kit. Furthermore, GIST patients with mutations in PDGF-R are responsive to imatinib. Some of the other approvals of imatinib (for instance for dermatofibrosarcoma) are linked to the ability of imatinib to inhibit PDGF-R.
So, back to diabetes. First, the animal model results are not the first to associate imatinib with a therapeutic effect in diabetic conditions. Some years ago, the drug was found to improve fasting glucose levels in diabetic CML patients. Second, the observation that imatinib also improves type II diabetes, suggest that the mechanism involved could be shared between the two conditions (type I and type II). One possibility is that the drug counteracts diabetes by maintaining ß-cell function by promoting a state similar to ischaemic preconditioning, and involving (among other things) NF-{kappa}B activation. Interestingly, the protective effect of imatinib was shared by sunitinib, which possesses a quite different profile of tyrosine kinase inhibitory activities. Sunitinib is a multi-targeted receptor tyrosine kinase inhibitor that is approved for the treatment of renal cell carcinoma and imatinib-resistant GIST. Its mechanism involves inhibition of PDGF-R and c-kit but not bcr-abl.
On the face of it, the enlargement of the putative treatment population for imatinib from CML, which affects a few thousand people per year, to diabetes, which affects hundreds of millions of people, could have a dramatic effect on the commercial potential for Glivec. However, a deeper analysis suggests this enlargement is unlikely to occur.
When Gleevec was introduced as an orphan medication, it was priced accordingly, at $32,000 per year for a 400 mg/day dose. As time has passed, the list of approvals for the product has grown considerably, however, the new indications for which it has been approved are still rare diseases. The expansion into a much more common condition like diabetes would take it into another realm, so that if it were to be used widely, it would place enormous strain on health systems worldwide. As it is currently earning around $3.7 billion a year for Novartis, its originator is not likely to pursue a new indication that puts current revenue at risk with any great fervour. The base composition of matter patent for imatinib is not due to expire in the US until 2015, so for the moment the opportunity to expand the product’s indications rests solely with the originator.
Aside from the commercial and financial issues, another reason why Glivec will not be used for diabetes lies in the heightened hurdles for cardiovascular safety now demanded of anti-diabetic medication. Increasing concern about the harmful effects of new drugs for diabetes is exemplified by the meta-analysis of the investigational PPAR modulator muraglitazar, the development of which was discontinued. As a result, the FDA has suggested that all new drugs (of whatever mechanism) will need to demonstrate cardiovascular safety before approval, either by an explicit failure to show a signal for cardiovascular risk, or through long-term clinical trials.
Although a direct association of imatinib with cardiovascular disease has not been demonstrated, there is concern in general about about cardiotoxicity as a a significant complication of cancer treatment with kinase inhibition. Clearly, there is a significant downside risk associated with the enlarged testing required to demonstrate the null hypothesis, namely that imatinib does not impair cardiovascular safety in diabetics.
This case history demonstrates a number of lessons about new indications for existing drugs. First, imatinib has clearly benefited from a wider range of clinical applications beyond the first approval. Most of these are oncological conditions, which have helped to lift a product that many in Novartis considered commercially marginal during its discovery and development, to blockbuster status. Second, much of the expanded utility of imatinib derives from a multiplicity of biological activities, even though it was originally optimised for potency and tyrosine kinase selectivity. Third, the identification of a new indication for an existing molecule, even if it is in an area of considerable commercial attraction, does not necessarily give rise to an attractive commercial opportunity. Drug repurposing is so much more than new indication discovery.
Along the way, there is another interesting point. Drug repurposing often falls into two categories, namely whether the new use that may be identified for a drug is ‘on-‘ or ‘off-target’, namely whether it involves the known biological activity of the drug, or whether it involves a new, unknown activity. Most drug repurposed discoveries (probably between 70 and 90% [Lipinski C, personal communication]) do not uncover a new biological activity. For this proportion, it is the new association between the existing biology and the secondary indication that is novel. In other words, repurposing, in addition to offering the risk, cost and time advantages that have been previously espoused, offers the opportunity to address a particular medical condition in a new way; and therefore developments arising from this approach typically result a first-in-class new medicine.
To put this in perspective, of the 20 new molecular entities introduced in 2007, only five of them (plus one monoclonal antibody) represented novel mechanisms, and the remainder of the entrants either represented ‘me-too’ or ‘follow-on’ approaches, where agents affecting the drug target(s) or compounds in the same class had already received regulatory approval; or they were ‘low-risk’ actives such as pro- or ante-drugs, or stereoisomers of existing entities.
When put in this light, drug repurposing is a lot more than ‘new tricks for old dogs’. In addition to its value in producing new pharmaceutical products more efficiently, it shines a bright new light into areas of discovery research based on new mechanisms which we may follow for decades to come.