Data were collected using a Powerlab and Chart software ADIn

Data were collected using a Powerlab and Chart 5 software (ADInstruments, Bella Vista, NSW, Australia) and were analysed using GraphPad Prism 5 (GraphPad Software, San Diego, CA). Responses to ghrelin receptor agonists in phenylephrine, methoxamine, endothelin-1, U46619 and 60mM [K+] contracted vessels were measured as the % change in active wall tension from that immediately prior to their addition. Contractions to ghrelin receptor agonists were measured as a % of the peak force of contraction to phenylephrine (3µM for saphenous, 10µM for mesenteric artery) or 60mM [K+] (in basilar artery). Both cumulative and non-cumulative curves were fitted using three parameter models. Data are reported as mean±S.E.M or, for data generated by the three parameter models, the best-fit parameter determined by nonlinear regression and the standard error is presented; n is the number of animals. For the Schild plot a linear regression was performed using the EC50 values obtained from the phenylephrine concentration response curves. 1- or 2- way ANOVAs, followed by Dunnett׳s post-tests were used where indicated. P<0.05 was considered significant. The following compounds were used: ulimorelin and TZP102 from Tranzyme Pharma; capromorelin (CP424391) from Pfizer Pharmaceuticals, Sandwich, UK and synthesised by us; carbamylcholine chloride (carbachol), phenylephrine hydrochloride, methoxamine hydrochloride, propranolol hydrochloride, prazocin, idazoxan, atropine, glibenclamide, nifedipine, L-NAME, and ET-1 (from Sigma-Aldrich, Sydney, Australia); rat ghrelin (Auspep, Melbourne, Victoria, Australia); rat desacyl ghrelin (GL Biochem, China), AZP-531 (ChinaPeptides, Shanghai, China); YIL 781 and U46619 (Tocris, Bristol, UK). Stock solutions were prepared as follows; ghrelin, desacyl ghrelin, AZP-531 (all 1mM pH 5.6 acetic A 61603 hydrobromide mg buffer solution), U46619 (100µM 70% ethanol), glibenclamide (10mM in DMSO), nifedipine (1mM 70% ethanol), YIL 781 (10mM 10% DMSO). All other drugs used were prepared in dH2O. The carrier solvents, acetic acid buffer, DMSO and ethanol, added alone in the same amounts as in experiments with drugs, had no effects on arterial wall tension in the vessels studied.
Results
Discussion In a previous study of ulimorelin׳s effects, a vasodilatator effect on arteries preconstricted with phenylephrine was reported (Callaghan et al., 2014). Ghrelin receptor antagonists, JMV2959 and YIL 781 (Esler et al., 2007), did not inhibit this effect and ghrelin receptor expression was not detected in the arteries. In the current work, we observed that ulimorelin caused vasodilatation in saphenous arteries from rats, as well as wild type and Ghsr null mice, which supports the conclusion that vasodilatation is mediated through a receptor different to the ghrelin receptor. Two classes of ghrelin receptors, other than GHSR1a, have been pharmacologically identified at which ghrelin and compounds related to ghrelin act, the ghrelin receptor-like receptors (GRLRs), where ghrelin and desacyl ghrelin are both active, and the unacylated ghrelin (UAG) receptors at which desacyl ghrelin but not ghrelin is an effective agonist (Callaghan and Furness, 2014). The sites identified in the present study do not match GRLRs or UAG receptors as neither ghrelin nor desacyl ghrelin have a vasodilatator action in the arteries investigated. The potencies of ulimorelin at mediating the effects described in this study were quite different from those at the ghrelin receptors. At rat and human ghrelin receptors, ulimorelin has an EC50 in the range of 1–2nM (Hoveyda et al., 2011, Callaghan et al., 2014). For relaxation of preconstricted arteries, the EC50 of ulimorelin was 0.5–5µM and for contraction the EC50 was 10–30µM, providing further evidence that these effects are not mediated by the ghrelin receptor. The concentrations of ulimorelin we studied are within the range achieved in vivo following infusion of 160–600µg/kg ulimorelin which produced peak plasma concentrations in the range 6–14µM (Lasseter et al. 2008). However, the approximately 1000-fold lower potency at the off-target sites identified in the current study suggests that they can be avoided in therapeutic applications of ulimorelin targeting the ghrelin receptor.