Nonpeptidic ligands for peptide-activated G protein-coupled receptors

Jade S. Blakeney, Robert C. Reid, Giang T. Le, David P. Fairlie

Research output: Contribution to journalReview ArticleResearchpeer-review

73 Citations (Scopus)

Abstract

G protein-coupled receptors (GPCRs) are the largest family of proteins considered to be viable targets for pharmaceuticals, but less than 5% of these membrane-spanning cell-surface proteins are currently regulated by drugs in man. A major challenge in the design and development of druggable lead compounds for this receptor class is the creation of new nonpeptidic chemical entities with structural diversity and receptor selectivity. Native peptide ligands for GPCRs provide important clues to the design of nonpeptidic ligands but they tend to bind in different ways to, or at subtly different sites on, any given receptor. Since many different GPCRs appear to recognize recurring structural fragments in their potent nonpeptidic ligands, the recombination of complementary fragments from compounds already known to be bioavailable drugs might be a successful approach to new drug leads, though with associated risks of losing receptor selectivity. To provide a more comprehensive basis for assessing the status and future prospects of GPCR- directed drug discovery, we have assembled a diverse portfolio of 577 nonpeptidic chemical structures known to interact with 76 peptide-activated GPCRs, and reported on their chemical composition, selective affinities for GPCRs, agonist/antagonist potencies, and pharmacological properties. We emphasized their advantages as drugs (over peptide ligands), described some of their origins, and identified some more promising chemical fragments that provide important clues for developing new nonpeptidic ligands for GPCRs through fragment assembly, privileged structure elaboration, and natural product modification. This portfolio of compounds provides a valuable guide to druggable fragments for lead compound discovery, and not only for GPCR target proteins. It may enable researchers to identify similarities in scaffolds, substituents, and other fragments across different compound and receptor classes. Peptide-activated GPCRs have historically been considered difficult pharmaceutical targets because of their large interacting protein-protein surfaces which are difficult to attack with small organic compounds. While this might have once been a reasonable consideration for agonist development, antagonists usually only need to interfere either with small 'hot spots' to block receptor activation by endogenous agonists, or with key interactions that define high-affinity binding. There are also now numerous examples of small molecule synthetic agonists, inverse agonists, and allosteric regulators that exert their functions through binding to only very small regions of a GPCR. It seems likely that the often huge receptor-binding domains of endogenous protein and peptide agonists that target GPCRs have evolved to confer high affinity and especially receptor selectivity for signal transduction via GPCRs. The actual activating or effector domain of GPCR ligands is however frequently a very small segment, often of low affinity alone, that is anchored in proximity of its ligand-binding site on a GPCR by a higher affinity GPCR-binding domain of the agonist that has no direct receptor-activating role in signal transduction. Clearly, based on the content of this review alone, peptide-activated GPCRs are certainly tractable and very promising targets for small nonpeptidic therapeutic candidates. Of the several hundred known peptide-activated GPCRs, most academic and industry researchers have necessarily focused on investigating the regulation of just one or a few GPCRs and usually for a single disease area, for which they painstakingly glean insights on structure - function relationships toward the development of a potent and selective drug. We decided to review the field laterally in a generic way, partly because we do not think this has been done previously to anything like the extent of this review, and partly because we wished to examine numerous targets to see if there were common approaches used successfully (and unsuccessfully) to develop drugs, as well as less common approaches that have not been used widely. In the present perspective survey, there are indeed a number of strikingly common structural features of ligands for GPCRs, as well as some promising directions in drug discovery that have not been fully exploited for many GPCRs. For peptide-activated GPCRs, the most obvious historical approach to drug discovery has been to examine the peptides that activate the receptor. This has often been the starting point for developing nonpeptidic drugs, with the peptide backbone structure and side chain composition affording valuable clues for the generation of constrained peptides, cyclic peptides, peptide fragments fused to nonpeptidic units, and peptidomimetics designed and iteratively optimized to make ligands more potent, more selective, less peptidic, and more druggable. Such peptidic compounds have been valuable mechanistic probes of biological processes, helped to validate therapeutic targets, and sometimes have proven to be valuable drugs in their own right. We have previously overviewed such ligands for over 100 peptide-activated GPCRs and thus do not mention them herein. 11,514,515 Generally, such peptidomimetics have ultimately had to be converted to more nonpeptidic compounds to generate cheaper drugs with improved bioavailability, stability, favorable pharmacokinetic and pharmacodynamic profiles. In the present review, we have instead focused on over 500 nonpeptidic ligands for over 70 peptide-activated GPCRs for the potential treatment of a large and diverse range of diseases and briefly commented on their pharmacological properties. Unlike peptide ligands, which can bind to the extracellular N-terminus and/or the three extracellular loops of GPCRs, the vast majority of nonpeptidic ligands for GPCRs primarily target the cell membrane, localizing to the hydrophobic transmembrane helices that are fairly conserved throughout the GPCR families. The most common approach used to identify active compounds has been high-throughput assaying of natural product and synthetic compound libraries resulting in 'hits' that are usually of micromolar affinity for a receptor. Mostly, this effort has involved random compound screening, but more recent efforts have concentrated on more focused structural libraries either based on known active compound classes, drug-like structures, or smaller units that could conceivably be linked through fragment-based assembly. Combinatorial chemistry (or parallel synthesis) has probably been of best value for generating drug fragments or optimizing hits into more potent/selective 'leads', and of less value in directly finding high value hits or leads unless extensive structural diversity can be readily incorporated into the chemistry. More recently, homology models of receptors have become more useful for ligand docking, with ligand binding sites being assigned through combinations of docking studies and effects of site-directed receptor mutagenesis on structure - activity relationships for affinities/activities of tested ligands. This work, like the evaluation of ligand affinity/potency, can be compromised if the receptor is transfected into a particular cell type, because this process artificially increases the number and density of receptors, and thus, ligand activities rarely correspond in magnitude to their effects on whole cells with native expression levels of GPCRs. It is also not yet clear whether transfection alters downstream activity profiles of agonists. Very recently, it has become more widely appreciated that ligand binding to a GPCR is frequently species-dependent and that residues vary in a GPCR between species. Such variations, by influencing ligand affinity/potency, can provide valuable information for identifying the ligand binding site on a GPCR. Polymorphisms even within human populations can complicate the situation even further. Virtual screening of in silico compound libraries has recently also become more viable, and a cheaper alternative to finding hits, as homegrown computer software programs have improved to some extent allowing more accurate ligand fitting and scoring, though there are still no really effective commercial software programs available for this endeavor. In any case, ligand docking is inherently compromised by the lack of precise structural information on GPCRs and so modelling necessarily lacks the degree of accuracy required for confident predictions of appropriate drug composition. Another approach to hit finding is based on the detection of common 'scaffolds', termed "privileged structures",8,14,25,48,49,51-57,64,70,193,337,516-521 in ligands that bind to different GPCRs. Even a cursory inspection of the nonpeptidic compounds assembled in this review reveals many components that are shared by ligands for different GPCRs. Just a few of these are benzimidazoles (10, 17, 147, 212, 319, 430), benzimidazolepyridines (2-6, 14, 18, 19), benzodiazepines (56, 57, 202-206, 565, 566), benzofurans (475, 482, 523), benzothiophenes (431), biphenylimidazoles (2-14, 17, 18), biphenyltetrazoles (7-14), diaminopyridines (58-60), dihydropyridines (48, 438-440), dihydroquinoxalinone (53-55), imidazoles (1, 8, 12, 15, 16, 100, 101, 213, 289, 379), imidazopyridines (2-6, 14, 21, 146, 236, 237), imidazolopyrimidinone (270), indanes (98, 230-232, 278, 279, 283, 284, 440, 457, 472, 525), indazoles (320, 321, 530), indoles (77, 93-96, 207, 209, 211, 212, 252, 253, 297-299, 317, 318, 428- 430, 442, 473, 488, 500, 511, 528, 529, 535, 553), lactams (128, 263, 276, 277, 303, 368), quinazolin(on)es (184, 185, 207, 208, 247, 331, 332, 448-450, 507), quinolines (33-35, 37-39, 322-328, 506), piperazines (38, 57, 80, 83, 84, 152-155, 343-350, 358, 359, 374-376, 381-390, 410, 438, 477-479, 486, 520, 524, 525, 555, 556) pyridazin(on)es...

Original languageEnglish
Pages (from-to)2960-3041
Number of pages82
JournalChemical Reviews
Volume107
Issue number7
DOIs
Publication statusPublished - 1 Jul 2007
Externally publishedYes

Cite this