A subsequent modeling study by a different group was

A subsequent modeling study by a different group was based on the bovine rhodopsin homology model using the 1L9H construct [38]. However, the site-directed mutagenesis data and docking suggested that the niacin-binding pocket is different to that described in the aforementioned report, but similar to the ligand-binding pocket of most class A GPCRs. The presumed binding pocket consists of residues mainly from TMH3, TMH5, TMH6, and ECL2. Interestingly, Arg111 has also been recognized as a critical residue for binding by functioning as a basic anchor point for the WIN 18446 component of ligands. An example of this model-directed structure–activity relationship (SAR) efforts is described in Section 5.3vide infra.
Five major in vitro assays have been employed to evaluate GPR109A agonists: a 3H-niacin-binding displacement assay and a guanine nucleotide exchange functional assay on the GPR109A-expressing CHO cell membranes [25], a calcium mobilization assay [37], a forskolin-stimulated cyclic adenosine monophosphate (cAMP) accumulation assay in CHO-hGPR109A cells [27], and a lipolysis inhibition assay in human adipocytes [29]. To evaluate the in vivo therapeutic index (TI) of GPR109A agonists regarding FFA reduction and vasodilation, mouse [39], rat, and dog [40], [41] models have been developed. GPR109A agonists caused dose- and time-dependent vasodilation in these models. The TI can be defined as the threshold drug level that induces flushing over the IC50 for FFA reduction. In mice and rats, FFA reduction and vasodilation were measured as a function of plasma drug levels in two separate experiments. Changes in ear perfusion were established using laser Doppler flowmetry to measure the magnitude of vasodilation and were used to estimate the threshold plasma level of drug required to trigger vasodilation. Compared to the rat model, the dog model allowed measurement of FFA suppression, vasodilation, and compound blood levels in the same conscious dogs after oral compound administration. Vasodilation was established by measuring changes in red color values in the ear using a spectrocolorimeter. Because FFA reduction and flushing were measured in the same animal, it was possible to establish TI on a dose basis as well. Thus the TI was defined as the threshold dose at which vasodilation occurred in a subset of the animals divided by the dose which gave maximal FFA suppression. Based on these criteria, the TI of niacin is 0.5–0.7 and the TI of acipimox is 1–1.4 in rats and dogs, which indicates that it is difficult to mechanistically separate FFA reduction and vasodilation effects for these two drugs. This is consistent with the evidence (Section 2, vide supra) that both FFA suppression and vasodilation are GPR109A-dependent events [34].
Medicinal Chemistry The discovery of the high-affinity niacin receptor, GPR109A, has spawned significant interest from academia as well as the pharmaceutical industry in searching for novel and potent agonists of this receptor [42], [43], [44]. Medicinal chemistry efforts toward development of GPR109A agonists have led to several classes of potent modulators including full orthosteric, partial orthosteric, and positive allosteric agonists.
Conclusions With well-established animal models to evaluate vasodilation and FFA reduction, several compounds were profiled in vivo and they indeed displayed improved TIs relative to niacin. Furthermore, these models appeared to correlate to humans as two candidates including both MK-0354 and INCB-19062 showed excellent FFA reduction accompanied by minimal flushing in humans. At least four clinical candidates including GSK-256073 [112], MK-0354 [102], MK-1903 [113], and INCB-19062 [91], and one preclinical candidate MK-6892 [77] have been reported. Neither the structure of GSK-256073, nor the clinical data, has been reported. In Phase II clinical trials, neither GPR109A partial agonist MK-0354, nor the full agonist MK-1903 showed substantial lipoprotein effects, and both candidates were discontinued. INCB-19062 is targeted to a type II diabetes indication based upon the related role of FFA to insulin sensitization in type II diabetes, and the robust FFA lowering effect observed in a Phase I clinical trial devoid of FFA rebound.