Novel Cannabinoid Receptor 2 (CB2) Low Lipophilicity Agonists Produce Distinct cAMP and Arrestin Signalling Kinetics without Bias

Abstract

Cannabinoid Receptor 2 (CB2) is a promising target for treating inflammatory diseases. We designed derivatives of 3-carbamoyl-2-pyridone and 1,8-naphthyridin-2(1H)-one-3-carboxamide CB2-selective agonists with reduced lipophilicity. The new compounds were measured for their affinity (radioligand binding) and ability to elicit cyclic adenosine monophosphate (cAMP) signalling and β-arrestin-2 translocation with temporal resolution (BRET-based biosensors). For the 3-carbamoyl-2-pyridone derivatives, we found that modifying the previously reported compound UOSS77 (also known as S-777469) by appending a PEG2-alcohol via a 3-carbomylcyclohexyl carboxamide (UOSS75) lowered lipophilicity, and preserved binding affinity and signalling profile. The 1,8-naphthyridin-2(1H)-one-3-carboxamide UOMM18, containing a cis configuration at the 3-carboxamide cyclohexyl and with an alcohol on the 4-position of the cyclohexyl, had lower lipophilicity but similar CB2 affinity and biological activity to previously reported compounds of this class. Relative to CP55,940, the new compounds acted as partial agonists and did not exhibit signalling bias. Interestingly, while all compounds shared similar temporal trajectories for maximal efficacy, differing temporal trajectories for potency were observed. Consequently, when applied at sub-maximal concentrations, CP55,940 tended to elicit sustained (cAMP) or increasing (arrestin) responses, whereas responses to the new compounds tended to be transient (cAMP) or sustained (arrestin). In future studies, the compounds characterised here may be useful in elucidating the consequences of differential temporal signalling profiles on CB2-mediated physiological responses.

Keywords: CB2; beta-arrestin 2; cannabinoid; cannabinoid receptor agonists; chemistry; cyclic AMP; drug development; kinetics; pharmaceutical; receptor; signal transduction.

Figure 1

Binding affinities of compounds of interest at human CB2 or CB1 determined from radioligand competition binding assays (pK_d for CP55,940, pK_i for other compounds). Data are mean ± SEM from three to four independent experiments, except for UOSS75 at CB1 (n = 2). CB1 pK_i values for all compounds, other than CP55,940, are estimates (see main text and Table 1). The arrow for UOMM18 in the CB1 set indicates no detectable binding to the lowest possible affinity constant measurable in our assay (pK_i < 4). CB2 affinity was statistically compared between UOSS77 and its novel derivatives; * represents p < 0.05.

Figure 2

cAMP levels in response to varying concentrations of compounds of interest co-stimulated with 5 μM forskolin in cells transfected with human CB2 or control plasmid (“No R”). Biosensor measurements over a 20 min drug treatment were averaged, then normalised to vehicle-treated without forskolin (V; 0%) and vehicle-treated with forskolin (F; 100%) conditions. pEC₅₀ and E_max are displayed in boxes for test compounds. E_max is expressed as absolute magnitude percentage decrease from 100% (ΔE_max), such that larger ΔE_max represents greater efficacy. CP55,940 has pEC₅₀ 8.0 ± 0.1 and ΔE_max 35.1% ± 3.5%. Plotted and parameter data are mean ± SEM from three to four independent experiments. Asterisks and hashes represent a significant difference from the parameters for CP55,940 and UOSS77, respectively. ## and *** represent p < 0.01 and < 0.001, respectively.

Figure 3

cAMP signalling efficacy (E_max; top panels) and potency (pEC₅₀; bottom panels) at different time-points during 20 min stimulation with compounds of interest and 5 μM forskolin. Parameters were derived from concentration response curves generated at corresponding time-points with data normalised to vehicle-treated without forskolin (V; 0%) and vehicle-treated with forskolin (F; 100%), as in Figure 2 and Figure S1, such that smaller cAMP measurements indicate greater efficacy (inhibition of cAMP synthesis). Data are mean ± SEM from three to four independent experiments. Horizontal lines across the top of plots indicate results of within-drug statistical comparisons to the reference time-point (maximal efficacy), which is indicated by the dataset symbol. Solid capped lines indicate significant difference from the reference time-point (p < 0.05), whereas dotted lines indicate no significant difference from the reference time-point.

Figure 4

β-arrestin-2 plasma membrane translocation induced by varying concentrations of compounds of interest, in cells transfected with human CB2 or control plasmid (“No R”). Biosensor measurements over a 20 min drug treatment were averaged, then normalised to vehicle-treated (V; 100%). pEC₅₀ and ΔE_max (% increase above V) are displayed in boxes for test compounds. CP55,940 has pEC₅₀ 7.2 ± 0.1 and ∆E_max 17.6% ± 0.8%. Plotted and parameter data are mean ± SEM from three independent experiments. Asterisks and hashes represent a significant difference to the parameters of CP55,940 and UOSS77, respectively. #, ** and *** represent p < 0.05, 0.01 and 0.001, respectively.

Figure 5

β-arrestin-2 plasma membrane translocation efficacy (E_max; top panels) and potency (pEC₅₀; bottom panels) at different time-points during a 20 min stimulation with compounds of interest. Parameters were derived from concentration response curves generated at corresponding time-points with data normalised to vehicle-treated (as in Figure 4 and Figure S2), then 100% subtracted (V; 100%), implying that larger E_max values indicate proportionally greater efficacy than the vehicle measurement. Data are mean ± SEM from three independent experiments. Horizontal lines across the top of plots indicate results of within-drug statistical comparisons to the reference time-point (first robust concentration response curve), which is indicated by the dataset symbol. Solid capped lines indicate significant difference from the reference time-point (p < 0.05), whereas dotted lines indicate no significant difference from the reference time-point.

Figure 6

Signalling bias relative to CP55,940. Bias quantified as ΔΔLogR for cAMP signalling (positive; “cAMP Bias”) versus β-arrestin-2 translocation (negative; “Arrestin Bias”). (A) Bias for mean responses over a 20 min stimulation. (B) Bias at different time points throughout a 20 min stimulation. Data represents mean ± SEM from three to four independent experiments for each pathway.