P-glycoprotein (P-gp) is a multidrug resistance protein that is especially interesting as target for drug development because it is the ABC transporter that has been most strongly linked to MDR and to determining the pharmacokinetic properties of drugs. Like all multidrug resistance proteins, it is part of a system that provides immunity to chemicals, which is normally protective to the organism. However, it can become problematic when it limits the bioavailability and usefulness of therapeutic drugs. Selection experiments using different cytotoxic agents have shown that at least 12 ABC transporters may play a role in MDR, but P-gp, multidrug resistance associated protein (MRP1), and breast cancer resistance protein (BCRP) are the transporters most strongly linked to MDR. They are the targets most likely to cause significant attenuation of the drug resistant phenotype if agents could be developed to modulate their function. There has been a strong causative link established between development of MDR and over ex
Among these, P-gp confers the strongest resistance to the widest variety of compounds, and its ex
It is reasonable to believe that inhibition of P-gp would be well tolerated based on results of studies of knockout mice which show that this enzyme is important for the health of an organism, but is not essential. P-gp knockout mice are viable, healthy, fertile, and appear identical to mice with normal P-gp ex
P-gp has the typical four domain arrangement consisting of two NBDs and two TMDs and undergoes the same large conformational changes that are associated with ATP binding and hydrolysis, drug binding and transport, and cross-talk between the domains and subdomains. Its abnormally broad substrate specificity is mediated by multiple, distinct drug binding sites within its TMDs that may be able to bind more than one substrate at a time. These regions have been observed to rearrange during substrate transport.
The soluble NBDs project into the cytosol and are connected to the TMDs by ICDs, as is typical. The NBDs share 80% sequence identity with each other, and approximately 40% identity to NBDs of similar transporters, but may not be completely equivalent in function since mutations in certain motifs do not produce the same effects in either NBD and the function of the protein is altered when some parts are exchanged between the NBDs. Each NBD is necessary for hydrolysis to occur as shown by experiments where mutations in critical binding domains in either NBD or vanadate trapping at a single NBD can completely abrogate ATPase activity in the protein. Despite the numerous events that must happen in this cycle, it has been demonstrated that ATP hydrolysis in Pgp obeys Michaelis-Menten kinetics.
It is already well established that compounds capable of inhibiting P-gp increase drug accumulation in cells in vitro and restore their toxicity. Altered levels of P-gp ex
Previous attempts to develop clinically useful P-gp inhibitors have produced three generations of compounds. First generation inhibitors were mostly compounds already used as drugs for other purposes that were discovered to inhibit P-gp in cells in culture and in in vitro assays. For in vitro use they still used and are very effective, but clinical trials of these compounds as co-therapeutics failed. Many of them worked by competing with drugs for access to drug binding sites and so were most likely pumped out of the cell like other substrates, which can translate into high doses, serious side effects and dose-limiting toxicity. For example, verapamil is a calcium channel blocker used to treat heart conditions that was discovered to restore cytotoxicity to resistant cells by increasing drug accumulation. Even though it can be administered safely to patients at concentrations effective for treatment of heart conditions, dose-limiting cardio toxicity ensured that maximal achievable concentrations were considerably lower than required to inhibit P-gp function. Toxicities were frequently encountered with inhibitors from this group that caused them to fail in clinical trials.
A second generation of compounds was then developed based on the structures of first generation inhibitors using quantitative structure activity relationships (QSAR). These had fewer toxic side effects and were more potent inhibitors of P-gp. Concentrations of these agents sufficient to inhibit P-gp could be reached in the plasma, but still none passed phase II clinical trials. The biggest problem with this generation of potential drugs was that their lack of specificity for P-gp that frequently led to toxic pharmacokinetic interactions when they were combined with chemotherapeutic agents necessitating decreases in doses of the drugs they were co-administered with. They were frequently seen to simultaneously inhibit both P-gp and detoxifying enzymes such as cytochrome P450-3A which are coordinately regulated in excretory organs181 and have highly overlapping substrate specificities. This impaired both drug elimination and metabolism and was not well tolerated by patients.
In the third generation, researchers started generating novel inhibitors that were optimized for interactions with P-gp through in silico methods and combinatorial chemistry. These compounds were up to 1000 times more potent than earlier generations. They are able to modulate P-gp at concentrations far below the concentrations at which they become toxic and they could discriminate between P-gp and other closely related enzymes. Tariquidar (XR9576), one of the best of these compounds is currently in Phase II clinical trials but is showing only limited success in reversing MDR. It is clear though that inhibition of P-gp is possible and can potentiate the effects of other drugs or significantly change their pharmacokinetics and bioavailability. However, any further efforts to discover inhibitors of P-gp will have to take into account the lessons learned from these first generations about what type of complications can be encountered from interactions with certain regions within the protein.