There are over 500 protein kinases encoded by the human genome, most of which are members of the same superfamily. This has created a plethora of potential targets that can be studied in a unified way, but has highlighted the difficulty in developing compounds that are capable of inhibiting one of these enzymes specifically. The development of Glivec has shown that inhibition of more than one protein kinase can sometimes be beneficial, allowing the same drug to have more than one therapeutic use. However, more frequently one would expect such a lack of specificity to give rise to unwanted or unacceptable side effects. The recent availability of large panels of protein kinases (e.g. Davies et al. 2000; Bain et al. 2003) has been of considerable help in assessing the specificities of protein kinase inhibitors, and it is to be expected that such panels will continue to expand and eventually include the entire repertoire of protein kinases.
Lack of specificity may also mean that the therapeutic effect of a drug is actually mediated by inhibition of another protein kinase and not by inhibition of the kinase for which it was originally developed. For example, inhibitors of the cell cycle regulator CDK2 have been developed that suppress the proliferation of tumour cells, but these compounds may actually exert their therapeutic effects by inhibiting other protein kinases, such as CDK7 and/or CDK9, which are regulators of RNA polymerase II. It is therefore unclear whether the effects of these compounds are really mediated via CDK2. In order establish that the therapeutic effect of a drug is mediated by inhibition of a particular protein kinase one needs to show that the effects of the drug disappear in cells that express a drug-resistant mutant of the protein kinase (Eyers et al. 1999). It is possible to convert protein kinases to drug-resistant forms by single amino acid replacements (Brown et al. 1995; Eyers et al. 1998) so that, as for other types of drug, the development of drug resistance is a potential hazard. Mutations in Abl that make it resistant to Glivec are the cause of relapse in patients with chronic myelogenous leukaemia (Gorre et al. 2001). However, resistance to Glivec is mainly seen in patients with the most advanced stage of this disease, where extensive genomic instability has already taken place.
Most of the protein kinase inhibitors developed thus far target the ATP-binding site and must therefore be of sufficient potency to compete with the millimolar concentrations of ATP that are present in the intracellular milieu. Clearly, it is possible to develop compounds with the requisite in vivo potency, as shown by the number of compounds undergoing human clinical trials. However, this remains a challenging problem, especially for protein kinases that bind ATP particularly tightly. Some of the most interesting protein kinase inhibitors developed thus far, including Glivec (Schindler et al. 2000) and the p38 MAP kinase inhibitor BIRB 796 (Pargellis et al. 2002), not only target the ATP-binding site, but also trigger structural changes that induce the inactive conformations of these protein kinases. Two other compounds, PD 98059 and U0126, do not target the ATP-binding site at all, but bind to the inactive conformation of MKK1, preventing it from being activated by the protein kinase Raf (Alessi et al. 1995; Davies et al. 2000). The development of more compounds that prevent one protein kinase from activating another may be a promising strategy for novel drug development in this area, since many of these enzymes are components of protein kinase 'cascades'. Another way of generating compounds that are not ATP-competitive would be to target the binding sites for protein substrates, a topic discussed by Lawrence (in Part 1).
There are about 150 protein phosphatase catalytic subunits encoded by the human genome, and they fall into three main superfamilies. The generation of compounds that discriminate between different protein phosphatases is therefore also a challenging one. However, in contrast to protein kinases, the option of targeting an ATP binding pocket does not exist. Moreover, the protein substrate-binding cleft can be very polar, as in the case of PTP1B (Kellie 2003). This has made it difficult to develop compounds that combine high potency with cell permeability. The only protein phosphatase inhibitor that has advanced to human clinical trials, cyclosporin, inhibits PP2B in an unusual way; it binds to the protein cyclophilin, and the cyclosporin-cy-clophilin complex then inhibits the protein phosphatase (Liu et al. 1991). As discussed earlier, it seems more likely that the future of drug discovery in this area may lie in targeting the regulatory subunits of serine/threonine-specific protein phosphatases.
Finally, it is important to mention that inhibitors of protein kinases are not only becoming important for the treatment of disease, but also as reagents for the study of cell signalling. The huge number of citations garnered by the publications that have introduced these compounds to the scientific community are a reflection of the widespread need for these compounds by the scientific community. For example, I was surprised to learn from the Institute for Scientific Information that the paper we published in 1995 with David Dudley and Alan Saltiel at Parke Davis on the mechanism of action of PD 98059 (Alessi et al. 1995) was the UK's most frequently cited original research paper over the past 10 years in the fields of biology and biochemistry, while our publication with Peter Young and John Lee at SmithKline Beecham on the specificity of SB 203580 (Cuenda et al. 1995), a prototypic p38 MAP kinase inhibitor, was the UK's sixth most cited original research paper over this period. Although many compounds are advertised for sale as 'specific protein kinase inhibitors', in practice many have turned out to inhibit so many protein kinases that conclusions drawn from their use are likely to be erroneous (Davies et al. 2000; Bain et al. 2003). The number of really useful protein kinase inhibitors that are available commercially is still rather limited, but the number will increase considerably over the next few years. I believe that pharmaceutical companies have much to gain from the discoveries that will be made by exploiting these compounds, and it is to be hoped that many more will be released for general use in the future.
Was this article helpful?