Diverse Pharmacology of Allosteric Ligands for the FFA3 Short Chain Free Fatty Acid Receptor

Brian Hudson1, Hannah Murdoch1, Laura Jenkins1, Trond Ulven2, Graeme Milligan1. 1University of Glasgow, Glasgow, UK, 2University of Southern Denmark, Odense, Denmark

Introduction: Short chain fatty acids (SCFAs) are produced through fermentation of non-digestible carbohydrates by the microbiota in the gut. The identification of two GPCRs, FFA2 and FFA3, activated by SCFAs stimulated interest in the role of these receptors linking gut microbiota to health. However, a dearth of pharmacological tools selective for FFA2 vs FFA3 has made disentangling the functions of these receptors challenging. To address this we synthesized three compounds: 4-(furan-2-yl)-5-oxo-N-(o-tolyl)-5,6,7,8-tetrahydroquinoline-3-carboxamide (1); 4-(2-bromophenyl)-2-methyl-5-oxo-N-(o-tolyl)-1,4,5,6,7,8-hexahydroquinoline-3-carboxamide (2); and 2-methyl-5-oxo-4-(3-phenoxyphenyl)-N-(o-tolyl)-1,4,5,6,7,8-hexahydroquinoline-3-carboxamide (3); based on a chemical series reported in the patent literature to have FFA3 activity [1], and explored their activity at human FFA2 and FFA3.

Methods: Flp-In T-Rex 293 cells engineered to inducibly express human FFA2 or FFA3 in response to doxycycline were utilized in functional assays measuring [35S]GTPγS incorporation, extracellular signal regulated kinase 1/2 (ERK) phosphorylation, and inhibition of forskolin stimulated cAMP following previously described protocols [2].

Results: Testing compounds 1-3 at FFA3 demonstrated that while 1 was an agonist with pEC50 values of: 5.65 ± 0.07 in [35S]GTPγS; 5.37 ± 0.16 in pERK; and 5.77 ± 0.05 in cAMP; neither 2 nor 3 produced a measurable agonist response. 1 was significantly more potent (p<0.001) than the SCFA agonist propionate (C3) in each assay and was a full agonist yielding similar efficacy to C3 in each assay. Compound 1 was selective for FFA3, displaying no activity at FFA2 in the [35S]GTPγS assay. To determine whether 1 bound to the orthosteric site of FFA3 its function at mutant forms of the receptor with Arg residues (5.39 and 7.35) important for SCFA binding [3] mutated to Ala was examined. While mutation of either Arg residue eliminated C3 response, neither potency nor efficacy of 1 was altered, indicating 1 likely is an allosteric agonist of FFA3. The ability of 1 to allosterically modulate C3 function was assessed using the pERK assay. Increasing concentrations of 1 enhanced C3 potency, yielding an overall cooperativity (αβ) value of 41. Reverse experiments showed that C3 likewise enhanced the potency of 1 (αβ = 72), demonstrating reciprocal allosterism. Finally, we considered whether 2 and 3 are also allosteric modulators of FFA3. Indeed, 2 was found to enhance the potency of C3 (αβ = 6.0) in the [35S]GTPγS assay, suggesting that this compound is a positive allosteric modulator. In contrast, while 3 increased the potency of C3, it also decreased C3 efficacy, yielding an overall negative cooperativity value for αβ of 0.77 (α = 105 and β = 0.0073), indicating 3 has a complex mixed mode of allosteric action. Despite this, since a maximum concentration (30 μM) of 3 did result in an 89% decrease in C3 efficacy, this compound may prove useful as a functional FFA3 antagonist.

Conclusion: Selective allosteric modulators of FFA3 with diverse pharmacology that may be useful in uncovering the function of this receptor have been identified.

[1] Leonard et al. 2006 PCT Int. Appl. WO2006052566;

[2] Hudson et al. 2012 FASEB J 26:4951;

[3] Stoddart et al. 2008 J Biol Chem 283:32913