Gastrointestinal Hormonal Responses on GPR119 Activation in Lean and Diseased Rodent Models of Type 2 Diabetes

ORIGINAL ARTICLE Gastrointestinal hormonal responses on GPR119 activation in lean and diseased rodent models of type 2 diabetes

S Patel1,3,9, OJ Mace1,4,9, IR Tough2,9, J White1,5, T-A Cock1,6, U Warpman Berglund1,7, M Schindler1,8 and HM Cox2

BACKGROUND: G protein-coupled receptor 119 (GPR119) has emerged as a potential target for the treatment of type 2 diabetes (T2D) and tool compounds have been critical in the evaluation of GPR119 functions. METHODS: We synthesised a novel small-molecule GPR119 agonist, PSN-GPR119, to study GPR119 signalling activities in cells overexpressing GPR119. We measured GPR119-stimulated peptide hormone release from intestinal loops and oral glucose tolerance in vivo from lean (C57BL/6J mouse or Sprague-Dawley (SD) rat) and diabetic (ob/ob mouse or ZDF rat) models. To evaluate the direct effects of GPR119 agonism on gastrointestinal (GI) tissue, we measured vectorial ion transport (measured as ISC; short-circuit current) across rodent GI mucosae and from normal human colon specimens. RESULTS: GPR119 activation by PSN-GPR119 increased cAMP accumulation in hGPR119-overexpressing HEK293 cells (EC50, 5.5 nM), stimulated glucagon-like peptide 1 (GLP-1) release from GLUTag cells (EC50,75nM) and insulin release from HIT-15 cells (EC50,90 nM). In vivo, PSN-GPR119 improved glucose tolerance by ~50% in lean mice or rats and ~60% in the diabetic ob/ob mouse or ZDF rat models. Luminal addition of PSN-GPR119 to isolated loops of lean rat small intestine stimulated GLP-1, glucose insulinotropic peptide (GIP) and peptide YY (PYY) release under basal (5 mM) and high glucose (25 mM) conditions. Activation of GPR119 also reduced intestinal ion transport. Apical or basolateral PSN-GPR119 addition (1 μM) to lean or T2D rodent colon mucosae reduced ISC levels via PYY-mediated Y1 receptor agonism. The GPR119 response was glucose sensitive and was abolished by Y1 receptor antagonism. Similarly, in human colon, mucosa PSN-GPR119 acted via a Y1-speciﬁc mechanism. CONCLUSIONS: Our results show that functional GPR119 responses are similar in lean and diabetic rodent, and human colon; that GPR119 stimulation can result in glucose lowering through release of intestinal peptide hormones and that PSN-GPR119 is a useful tool compound for future studies. International Journal of Obesity (2014) 38, 1365–1373; doi:10.1038/ijo.2014.10

Keywords: GPR119; intestinal responses; incretins; peptide YY and Y1 receptors; glucagon-like peptide 1; type 2 diabetes rodent models

INTRODUCTION insulin-secreting cell lines and on rodent islets, which have been 5,8–10 New improved anti-diabetic therapies would be advantageous to characterised extensively. Importantly, insulin secretion address the increasing epidemic of type 2 diabetes (T2D) and mediated by GPR119 is glucose dependent and agonism will obesity. G protein-coupled receptor 119 (GPR119) is primarily not cause the hypoglycaemia associated with other diabetic 11,12 fi expressed in gastrointestinal (GI) enteroendocrine cells and agents. Furthermore, GPR119 expression is signi cantly β 1–3 fi increased in the islets of obese diabetic db/db mice when pancreatic cells in rodents and humans, attracting signi cant 13 interest as a therapeutic target for T2D. GPR119 agonists in vivo compared with islets from lean models adding credence to the have shown impressive lowering of postprandial glucose in rodent potential therapeutic value of this receptor. In addition, GPR119 3–5 fi models of T2D and GPR119 agonists have also demonstrated agonists have greater oral ef cacy as compared with intravenous beneficial effects on body weight,2 offering potential advantages efficacy, suggesting that GPR119 may control glycaemia via 12 to obese T2D patients over therapeutics already in the market. As additional incretin-based mechanisms. such, new GPR119 molecules have advanced into the clinic for the The secretion of endogenous anti-diabetic hormones, including treatment of T2D.6,7 glucose insulinotropic peptide (GIP) and glucagon-like peptide 1 9,14 The ability of GPR119 agonism to improve glucose tolerance (GLP-1), is a consequence of GPR119 agonism in vivo. In and reduce body weight shows that GPR119 biology warrants purified enteroendocrine cells and isolated murine intestinal further investigation. GPR119 agonists have direct effects on L cells, GLP-1 secretion is mediated by signalling pathways

1Prosidion Ltd, Oxford, UK and 2King's College London, Wolfson Centre for Age-Related Diseases, London, UK. Correspondence: Professor HM Cox, King’s College London, Wolfson Centre for Age-Related Diseases, Hodgkin Building, Guy’s Campus, Newcomen Street, London SE1 1UL, UK. E-mail: [email protected] 3Current address: Bioelectronics R&D, GlaxoSmithKline, Gunnels Wood Road, Stevenage, SG1 2NY, UK. 4Current address: Heptares Therapeutics, Broadwater Road, Welwyn Garden City, AL7 3AX, UK 5Current address: Argenta Discovery, 8/9 Spire Green Centre, Flex Meadow, Harlow, CM19 5TR, UK 6Current address: Pharmaxis Ltd, 20 Rodborough Road, Locked Bag 5015, Frenchs Forest, NSW 2086, Australia 7Current address: Science for Life Laboratory, Department of Medical Biochemistry & Biophysics, Karolinska Institute, S-171 21 Stockholm, Sweden 8Current address: Astra Zeneca R&D Mölndal, Cardiovascular & Metabolic Diseases iMed, Pepparedsleden 1, SE-431 83 Mölndal, Sweden 9These authors contributed equally to this work. Received 3 May 2013; revised 7 January 2014; accepted 14 January 2014; accepted article preview online 23 January 2014; advance online publication, 18 February 2014 Pharmacology of a novel GPR119 agonist, PSN-GPR119 S Patel et al 1366 involving L- and Q-type calcium channels.15 Additionally, rat ileum Oral glucose tolerance test (OGTT) treated in situ with an endogenous ligand for GPR119, oleoy- Animals were fed ab libitum and had free access to water and were lethanolamide (OEA) induced GLP-1 release.12 In humans, activa- maintained on a 12-h light:dark cycle. Oral glucose tolerance was assessed in overnight-fasted lean Sprague-Dawley (SD) rats (12 weeks old, n =8),lean tion of GPR119 by administration of a jejunal bolus of 2-oleoyl – fi C57BL/6J mice (11 12 weeks old, n = 10), Zucker diabetic fatty (ZDF) rats (12 glycerol (a component of fatty acid digestion with af nity for weeks old, n = 8) or obese ob/ob (leptin-deficient obese, B6.V-Lepob/J) mice 16 GPR119) stimulated GLP-1 release, implicating GPR119 as a fat (11–12 weeks old, n = 10) following acute treatment of PSN-GPR119. PSN- sensor. GPR119 (10 or 30 mg kg−1) or vehicle (20% 2-hydroxy propyl-β-cyclodextrin) The L-cell-derived hormone, peptide YY (PYY), is predominantly was administered by oral gavage 60 min (t = -60 min) before the glucose −1 expressed in the ileum and the colon,17 and circulating PYY bolus that was delivered by oral gavage (2 g kg ; t = 0 min). Blood samples – increases or decreases after feeding or fasting, respectively.18 PYY (t =0 250 min after glucose) were analysed using a commercially available 19 glucose oxidase assay. Plasma glucose is presented as absolute values (mM) activates Y1 and Y2 receptors to inhibit upper GI motility and and reactive area under the curve (AUC ) was calculated by subtraction 20,21 reactive inhibits water and electrolyte secretion in human colon, of the baseline AUC before glucose. and its product, PYY(3–36) appears to reduce appetite.22 In support of the anti-secretory role of PYY in the colon, activation GPR119 expression in intestinal regions of GPR119 mechanisms results in anti-secretory effects via As GPR119 transcript expression is very low (L cells represent o1% of the 23 endogenous PYY-activating epithelial Y1 receptors. Postprandial total mucosal cell population) intestinal RNA samples were pooled from blood glucose was also controlled by PYY, and both responses five human male donors (BioChain, Newark, CA, USA), whereas individual were absent from PYY-/- mice or their tissues.23 samples from six male SD rats and three male C57BL/6J mice were We synthesised PSN-GPR119 (N-(2-Hydroxyethyl)-4-{3-[1-(3- analysed separately. Real-time PCR was performed using pre-validated isopropyl-[1,2,4] oxadiazol-5-yl)piperidin-4-yl]propoxy}-2-methyl- primer / FAM-labelled GPR119 probe sets (Applied Biosystems, Paisley, benzamide (patent number, WO 2008/081205 A1), a potent UK) corresponding to GenBank accessions NM_178471.2 (human), small-molecule GPR119 agonist, and validated its potential to NM_181751.2 (mouse) and BD274917.1/ Celera rCT63637.0 (rat). Data were normalised to the housekeeping genes, hypoxanthine improve glucose tolerance in rodent models of diabetes and phosphoribosyl transferase I and hydroxyl-methylbilane synthase, obesity. Our hypothesis was that GPR119 signalling involved calculated using ΔCt. secretion of K-cell- and L-cell-derived hormones and that the mechanisms involved were consistent in lean and diseased rodent models of diabetes and obesity. In order to define the GPR119 Isolated rat intestinal perfusion and measurement of GIP, GLP-1 mechanisms in GI mucosa, we exploited the fact that PYY inhibits and PYY levels epithelial ion secretion in mammalian intestine and compared Male SD rats (250 g; Charles River, Margate, UK) were fed ad libitum, had the downstream PYY-mediated pathways of PSN-GPR119. We free access to water and were maintained on a 12-h light:dark cycle. Loops of whole small intestine were isolated and perfused in single-pass mode as extended these mechanistic studies to assess PSN-GPR119 activity previously described.25 Briefly, the first 30 min was a control period during in human colonic mucosal specimens. which Krebs–Henseleit (KH; in mM: 120 NaCl, 4.5 KCl, 1 MgSO4,1.8 Na2HPO4, 0.2 NaH2PO4, 25 NaHCO3, 1.25 CaCl2 and 5 D-glucose) was perfused at a flow rate of 1.6 ml min−1. After 30 min, the loop was perfused fi MATERIALS AND METHODS with a second experimental perfusate (identical to the rst except for the addition of PSN-GPR119). During this period, loops were either perfused Ethical approval with KH containing PSN-GPR119 (0.1–100 μM)or5mM glucose (osmotically Animal procedures were approved by the UK Home Office and undertaken balanced to 25 mM sugar with mannitol) for 30 min, then switched to an in accordance with the Animals (Scientific Procedures) Act (1986). Human identical perfusate containing PSN-GPR119 and 25 mM glucose. Serosal colon specimens were obtained from consenting patients undergoing samples were assayed for glucose, total GIP (tGIP; Millipore Ltd, Oxford, bowel resection surgery with the approval of Guy’s and St Thomas’ NHS UK), active GLP-1 (aGLP-1; Epitope Diagnostics Inc, San Diego, CA, USA) and Foundation Trust, in accordance with The Human Tissue Act (2004). total PYY (tPYY; Phoenix Pharmaceuticals, Burlingame, CA, USA) using commercially available ELISAs, as previously described.25 Tissue viability was indicated by maintenance of active glucose transport throughout the −1 GPR119-transfected HEK293 cell cAMP assay experiment. AUCreactive was calculated from time-courses in pg ml (g dry −1 ⩾ Stimulated cAMP accumulation was measured using an engineered weight) × min and presented as mean ± s.e.m. (n 4). HEK293 cell line expressing GPR119 as previously described.24 Monolayers were exposed to PSN-GPR119 (0.01–100 nM) for 30 min at 37 °C in buffer Mucosal electrophysiology containing 1% dimethyl sulphoxide. Cells were lysed and cAMP was Human colonic specimens from nine consenting patients undergoing determined (AlphaScreen cAMP kit, Perkin Elmer, Cambridge, UK). intestinal resection for primary malignancy, or GI regions from rodents (Charles River; lean SD rats or C57BL/6J mice and diseased ZDF rats or ob/ ob mice) were stripped of smooth muscle and adjacent pieces of mucosa Measurement of GLP-1 secretion from GLUTag cells were bathed in KH buffer, mounted on Ussing chambers and voltage- GLUTag cells were cultured in Dulbecco's modified Eagle medium clamped at 0 mV, as described in detail previously.21,23,26 Electrogenic − M 2 containing 5.5 m glucose. Cells were seeded onto 24-well plates coated ion transport was measured as short-circuit current (ISC;inμAcm ). with Matrigel (BD Biosciences, Oxford, UK) and used at 60–80% confluency. PSN-GPR119 was added to the apical (for potency measurement) or They were washed twice in nutrient-free bathing solution supplemented basolateral reservoir, whereas peptides were added basolaterally. with 0.1 mM diprotin A and 0.1 % (w/v) BSA and incubated with PSN- A concentration of 1 μM PSN-GPR119 gave near-maximal responses and GPR119 (0.1 nM–3 μM) for 2 h at 37 °C. Media was assayed for GLP-1 using was used to stimulate other GI areas. In mouse mucosa, vasoactive an active GLP-1 ELISA (Linco Research Inc., St Charles, MO, USA). intestinal polypeptide (VIP, 10 nM) was used to pre-stimulate anion 23,26 secretion, to optimise subsequent Gαi-coupled responses. PYY (10 nM) was a control activator of Y1 receptor-mediated responses through Measurement of insulin secretion from HIT-T15 cells Gαi-coupled reductions in epithelial cAMP and consequent attenuation of - 27,28 HIT-T15 cells were cultured in RPMI1640 with 10% . Cells were seeded in Cl secretion. Y1 responses were blocked with BIBO3304 ((R)-N-[[4- 96-well plates and cultured for 24 h before assay. Cells were pre-incubated (aminocarbonylamino methyl)-phenyl]methyl]-N2-(diphenylacetyl)-argini- for 30 min in Krebs-Ringer-HEPES buffer (119 mM NaCl, 4.7 mM KCl, 2.5 mM namide trifluoroacetate; 300 nM) and endogenous GLP-1 effects were CaCl2, 1.2 mM MgSO4,25mM NaHCO3, 1.2 mM KH2PO4,10mM HEPES) with abolished with 1 μM exendin(9–39). Mucosae were treated with both 0.1% BSA. PSN-GPR119 was added to buffer supplemented with 5.6 mM antagonists to block both endogenous peptide responses. The glucose glucose and incubated for 30 min. Media was assayed for insulin by ELISA dependence of agonist responses was tested using KH buffer (including + (Mercodia, Uppsala, Sweden). 11.1 mM glucose) or buffer containing mannitol (11.1 mM). The Na -glucose

Figure 1. PSN-GPR119 signiﬁcantly improves oral glucose tolerance in lean C57BL/6J mice and SD rats, compared with ob/ob mice and ZDF rat models. Left time-courses depict dosing of PSN-GPR119 (either 10 or 30 mg kg−1, p.o.) at t = −60 min and glucose (3 g kg−1)att = 0 min, and right glucose AUC reactive are calculated by subtracting the baseline glucose AUC before glucose delivery. Each point or bar is the mean ± 1s. e.m. and statistical differences compared with vehicle controls were **Po0.01, ***Po0.001. cotransporter 1 (SGLT1) inhibitor phloridzin (50 μM) was used to show Data analysis that only apical replacement of glucose reduced SGLT1 activity. The AUC data from perfused intestinal loops were calculated by α reactive 2-agonist UK14 304 (5-bromo-N-(4,5-dihydro-1H-imidazol-2-yl)-6-quinoxa- subtracting the baseline 5 mM glucose (no compound) period from the μ linamine, 1 M, basolateral) was used to reveal non-Y receptor Gαi-coupled PSN-GPR119-treated period. Statistical signiﬁcance was determined using reductions in ISC in rodent and human colon. unpaired Student’s t-test where; *Po0.05, **Po0.01 and ***Po0.001

© 2014 Macmillan Publishers Limited International Journal of Obesity (2014) 1365 – 1373 Pharmacology of a novel GPR119 agonist, PSN-GPR119 S Patel et al 1368 $ $$ versus 5 mM glucose (no compound) control or; Po0.05, Po0.01 and $$$ Po0.001 versus PSN-GPR119 (5 mM glucose). Plasma glucose was measured following in vivo glucose tolerance tests in millimolar. Comparisons were performed using one-way analysis of variance with Dunnett’s post hoc test; **Po0.01 and ***Po0.001 versus vehicle. ISC data was analysed using GraphPad Prism, version 5.0 (GraphPad, La Jolla, CA, USA). Single comparisons were performed using Student’s t-test, whereas multiple comparisons utilised one-way analysis of variance with Dunnett’s post-hoc test; Po0.05 was considered statistically different.

RESULTS GPR119 agonist activity, GLP-1 and insulin secretion in neuro- endocrine cell lines

PSN-GPR119 is a potent GPR119 agonist exhibiting an EC50 of 5.5 ± 2.0 nM for cAMP accumulation in hGPR119-transfected HEK293 cells (Supplementary Figure S1A). In GLUTag cells, PSN- GPR119 increased GLP-1 secretion (EC50, 75 ± 19 nM, Supple- mentary Figure S1B) and induced insulin secretion from HIT-T15 cells (EC50, 90 ± 21 nM; Supplementary Figure S1C).

Oral glucose tolerance tests In lean C57BL/6J mice, glucose tolerance following an oral glucose challenge improved dose-dependently after acute PSN-GPR119 (10 or 30 mg kg−1, p.o.) treatment (Figure 1). Blood glucose AUC was significantly lower following treatment with PSN-GPR119 compared with vehicle-treated animals (Po0.01; Figure 1). Despite a higher starting blood glucose concentration, the AUC for the glucose excursion following oral glucose challenge in ob/ ob mice was significantly decreased, notably to similar levels observed in the C57BL/6J mice (Po0.001; Figure 1). Blood glucose levels following 30 mg kg−1 PSN-GPR119 treatment in lean SD and ZDF rats were significantly decreased by ~ 50% and ~ 65%, respectively, again significantly when compared with vehicle (Po0.001 andPo0.01 respectively; Figure 1). Figure 2. Concentration responsiveness of PSN-GPR119 in lean SD rat small intestine perfused with 5 or 25 mM glucose. Rat small GPR119 mRNA expression in intestinal regions intestine was challenged luminally with 5 mM glucose in KH buffer GPR119 transcripts were found in small intestinal and colonic for 30 min to equilibrate the preparation. Afterwards it was perfused regions from mice, rats and humans (Supplementary Figures S2A for a further 30 min with an identical KH buffer containing either 5 and C). In lean C57BL/6J mice and SD rats, GPR119 expression was or 25 mM and PSN-GPR119 (0.03–10 μM). Serosal secretion samples highest in the distal gut and declined towards the duodenum, were analysed for total GIP, active GLP-1 and total PYY as indicated using commercially available ELISA kits. AUC reactive was calculated whereas human GPR119 expression was highest in the duodenum by subtracting the baseline AUC from the PSN-GPR119-treated AUC. with slightly lower levels in jejunum, ileum and colon Each bar is the mean ± 1 s.e.m. from six separate experiments and (Supplementary Figure S2C). statistical differences are: *Po0.05, **Po0.01 and ***Po0.001 $ $$ versus 5 mM glucose (no compound) control or Po0.05, Po0.01 $$$ o GPR119 agonist induced GI hormone secretion from isolated loops and P 0.001 versus PSN-GPR119 (5 mM glucose). of rat intestine Small intestines from fed male SD rats were perfused luminally for receptor-mediated anti-secretory responses23 using lean and 30 min with PSN-GPR119 in 5 or 25 mM glucose to mimic pre- and diseased colonic tissue. post-prandial conditions. A concentration of 5 mM glucose was fi present as an energy source and to mimic the availability of First, lean C57BL/6J mice were signi cantly lighter than ob/ob mice of the same age (Po0.001), whereas the latter displayed glucose from the plasma before a meal. PSN-GPR119 concentra- fi o tion dependently stimulated tGIP, aGLP-1 and tPYY secretion in signi cantly higher fasting blood glucose (P 0.001, Figure 3a) demonstrating obesity and hyperglycaemia. Apical PSN-GPR119 was the presence of 5 and 25 mM glucose (Po0.001 for all peptides equipotent in C57BL/6J colon (EC =97.5nM (38.5–246.9 nM)) and after 1 and 10 μM PSN-GPR119; Figure 2). Consistent with the 50 ob/ob colon (EC =103.0nM (50.5–210.3 nM); Figure 3b). Lean SD GLUTag model, ~ EC values with 5 mM glucose for tGIP, aGLP-1 50 50 rats out-weighed their ZDF counterparts (Po0.001) but the latter and tPYY were 1.1, 0.9 and 1.1 μM, respectively, and were not exhibited significantly higher fasting blood glucose levels (Po0.001; significantly different in the presence 25 mM glucose. PSN-GPR119 and glucose stimulation of GI hormone secretion appeared to act Figure 3c). PSN-GPR119 potency was however similar in colonic additively. mucosa from SD and ZDF rats (EC50 values were 126.6 nM (51.5–311.3 nM) versus 255.0 nM (34.8 nM–1.87 μM), respectively), and there were no significant differences in efficacy (Figure 3d). GPR119 agonism in lean and diseased rodent colon PSN-GPR119 sensitivity was greatest in the descending colon To investigate whether GPR119 signalling mechanisms were intact from lean and obese mice (Po0.05; Supplementary Figure S3). during metabolic disease, we exploited the finding that GPR119 There was no difference in PSN-GPR119 response size after apical activation causes endogenous PYY release and local epithelial Y1 or basolateral addition in any GI area. Subsequent PYY (10 nM)

Figure 3. GPR119 responses in C57BL/6J and ob/ob mice, SD and ZDF rats. (a and c) Weight and blood glucose measurements for the mice and rats used in these studies, and (b and d) concentration–response curves for PSN-GPR119 in colonic mucosa from C57BL/6J or ob/ob mice (b) and SD or ZDF rats. EC50 values are given in the text. Each bar or point is the mean ± s.e.m from the n numbers shown in parenthesis. ***Po0.001 (b and d) n = 4–5 and 4, respectively. responses followed a similar pattern in each model (data not GPR119 agonism in human colon mucosa shown). In contrast, PSN-GPR119 responses were comparable PSN-GPR119 added apically (Figure 6) or basolaterally (data not throughout the small and large intestine from SD and ZDF rats shown) caused long-lasting reductions in ISC (that were similar to (data not shown). PYY responses) with an EC50 of 71.3 (21.5–166.2) nM (Figure 6a). As Second, the receptor pharmacology was similar in lean and seen in rodent colon mucosa, PSN-GPR119 responses were diseased rodent colon. Initially, Y1 receptor tone was revealed blocked by the Y1 antagonist, BIBO3304 (Po0.05; Figure 6b), 21,29 using BIBO3304, a competitive Y1 antagonist, as seen previously in which also revealed Y1 tone, as seen previously. Furthermore, 26,29 fi 2 C57BL/6J colon. BIBO3304 signi cantly increased basal ISC co-addition with the Y2 antagonist, BIIE0246 ((S)-N -[[1-[2-[4-[(R,S)-5, levels (Po0.01), but the degree to which this occurred was 11-dihydro-6(6 h)-oxodibenz [b,e]azepin-11-yl]-1-piperazinyl]-2-oxoethyl] o reduced in ob/ob colon (P 0.05, Figure 4a). Y1 antagonism cyclopentyl]acetyl]-N-[2-[1,2-dihydro-3,5(4H)-dioxo-1,2-diphenyl-3H-1, abolished subsequent PSN-GPR119 responses in C57BL/6J and 2,4-triazol-4-yl]ethyl]-argininamide) raised basal ISC levels further ob/ob mouse colon (Po0.01, Figure 4a) and responses to exogenous than BIBO3304 alone, and the combination of the two Y PYY in all groups (Po0.05). In C57BL/6J and ob/ob tissues, there antagonists abolished subsequent PSN-GPR119 (Po0.05; were small exendin 4 increases in ISC, which were blocked by the Figure 6b) and PYY responses, but had no effect on UK14 304 – o GLP-1 receptor antagonist, exendin(9 39) (P 0.01, Figure 4a). α2-mediated responses anti-secretory responses (data not shown). In colon mucosa from SD and ZDF rats, Y1-mediated tone was Finally, PSN-GPR119 responses were glucose sensitive (Po0.05; similar, but interestingly a larger PSN-GPR119 response was Figure 6c) and the effects of SGLT1 inhibitor, phloridzin were also observed in ZDF tissue (Po0.05, Figure 4b). Consistent with the reduced (Po0.05) when apical glucose was replaced by mannitol. C57BL/6J and ob/ob data, PSN-GPR119 responses were abolished As seen in rodent mucosa, PYY-Y1 responses were not glucose by BIBO3304 and there were small GLP-1 responses that were sensitive in human colon. blocked by exendin(9–39) (Po0.05 in ZDF colon; Figure 4b). Third, GPR119 responses were glucose-sensitive in mucosae from lean and diseased rodents. Apical replacement of glucose by mannitol reduced apical PSN-GPR119 responses significantly DISCUSSION (Po0.01) and subsequent phloridzin activity (Po0.01; Figure 5), Few studies have shown intestinal efficacy with selective GPR119 in line with the apical localisation of electrogenic SGLT1. agonists in normal and T2D rodent models9,11,12,30 and as yet Basolateral PSN-GPR119 responses were also glucose-sensitive none have compared this with efficacy in human intestinal tissue. (data not shown). Interestingly, phloridzin responses were Our study shows that luminal delivery of a GPR119 agonist significantly larger in ZDF compared with SD colon (Po0.05; stimulates K- and L-cell activity acutely in rat intestine. PSN- Figure 5b). Control PYY-Y1 and UK14 304 α2-mediated responses GPR119 was highly potent, stimulating insulin and GLP-1 from responses were glucose-insensitive. secretory cell lines. As GPR119-activated insulin secretion is well

Figure 4. Blockade of Y1 receptors inhibits GPR119 agonism and PYY responses in descending colon mucosa from, in (a) mouse, (b) rat models. Vehicle (0.1% dimethyl sulphoxide, ﬁrst bar) or Y1 antagonist BIBO3304 (300 nM, BIBO) was added in the absence or presence of 1 μM exendin (9–39) (BIBO+9–39). After vasoactive intestinal polypeptide (VIP) (30 nM, data not shown), PSN-GPR119 (1 μM, apical), PYY (10 nM) and exendin 4 (100 nM) were added sequentially and their anti-secretory effects pooled to give the means ± 1 s.e.m. (n, numbers in parenthesis). Control responses to exendin 4 (100 nM, Ex 4) are also included. Statistical differences between vehicle and antagonist-pretreated agonist mean responses are as shown. *Po0.05, **Po0.01, ***Po0.001. Differences between lean and diseased rodent tissue responses are shown by a vertical line, *Po0.05.

established, we set out to determine the relative importance of that the contribution of each hormone towards glucose lowering this receptor on gut peptide release and to evaluate the peptide in lean and diseased models may differ and in fact recent data in pharmacology involved in human colon. GPR119 expression was mice indicates L-cell activity is more important for acute highest in mouse and rat colon, consistent with L-cell distribution regulation of glucose homeostasis.36 Unfortunately, we were along rodent GI tract. Interestingly, we observed a more uniform unable to determine the relative contribution of GIP, GLP-1 or PYY GPR119 distribution along the human GI tract with slightly to the improvement in glucose tolerance on our models. However, elevated expression in the duodenum. GPR119 levels may be the net effect is that PSN-GPR119 ameliorates postprandial blood higher in the upper GI tract of humans relative to rodents; glucose concentrations in rodent models of diabetes and obesity however, the source of the human gut RNA was pooled from indicating its potential to improve glucose tolerance in vivo. multiple donors (because GPR119 transcript expression is low, Luminal perfusion of PSN-GPR119 into loops of rat small L cells accounting for o1% of the total mucosal cell population), intestine caused concentration-dependent increases in tGIP, whereas the rodent intestinal RNA was sourced from individual aGLP-1 and tPYY secretion (an effect we have also seen with animals, which may also account for some differences. luminal OEA25). Peptide secretion was further enhanced by In vivo, oral PSN-GPR119 improved glucose tolerance signifi- elevating luminal glucose from 5 to 25 mM, demonstrating the cantly in lean and diseased rodents. This is consistent with other existence of further capacity for the K and L cells to secrete GIP, studies showing that GPR119 agonism (acute or chronic) leads to GLP-1 and PYY. GPR119 agonism and glucose acted in an additive effective glycaemic control.3,8,9,11,14,30 PSN-GPR119 (30 mg kg−1) manner, consistent with the recent report that GPR119 agonists was as efficacious (50–60% glucose lowering in an oral glucose stimulate GLP-1 secretion glucose independently.37 The notion tolerance test) in lean, as it was in rodents with T2D. The secretion that intestinal GPR119 agonism leads to an incretin effect is profiles of GIP and GLP-1 in patients with T2D are well established; supported by studies showing that luminal OEA application in vivo levels of GIP in systemic blood are normal in pre- and post- caused significant GLP-1 release compared with i.v. administration prandial conditions. However, patients with T2D are insensitive to of OEA at a 200-fold higher concentration.12 circulating GIP.31,32 Circulating levels of GLP-1, in contrast are Another in-house small-molecule GPR119 agonist, PSN632408, reduced in T2D.32 The exact mechanism of L-cell impairment is has been shown to stimulate L-cell-derived PYY responses in unknown, but it is not thought to be via increased clearance of human and murine colon.23 Apical or basolateral PSN632408 GLP-1.33 Exogenous administration has shown that the insulino- caused equivalent responses, through PYY activation of epithelial tropic activity of GIP is defective, whereas that of GLP-1 remains in Y1 receptors. PSN632408 was less potent than PSN-GPR119 in the T2D. The insulinotropic, glucagonostatic and gastric emptying present study, exhibiting an EC50 of 5.7 μM in mouse colonic 32,34,35 actions of GLP-1 are preserved in T2D. These data indicate mucosa, whereas in human colonic mucosa, 10 μM PSN632408

Figure 5. Glucose sensitivity of apical PSN-GPR119 (1 μM) in the presence of glucose (11.1 mM) both sides or following replacement with mannitol (11.1 mM) on the apical side only in C57BL/6J mice (a) and SD rats (b). Subsequent phloridzin (Phlor, 50 μM, apically) and PYY (10 nM) responses are also shown. Values are the mean − 1 s.e.m. from n numbers as shown. Statistical differences between responses obtained with glucose both sides versus mannitol apically, are shown, *Po0.05, **Po0.01, ***Po0.001. There were no differences between lean control and diseased mucosal responses except for phloridzin responses, which were significantly larger in ZDF than SD rat colon mucosa, *Po0.05 (vertical line). caused a maximal mucosal response23 similar to that achieved the presence of glucose, indicating that GPR119 agonism should here with 0.3–1 μM PSN-GPR119. reduce the risk of hypoglycemia in a diabetic population. To evaluate the efficacy of PSN-GPR119 during diabetes and However, in colonic preparations GPR119 agonism was not obesity, we used ZDF rats and ob/ob mice as diseased counter- abolished when glucose was removed, suggesting that PSN- parts to lean SD rats and C57BL/6J mice. Consistent with a 10-fold GPR119 may exert some glucose-independent activity, as seen for increase in potency over PSN632408, PSN-GPR119 (potency at GPR119 agonism in GLUTag cells and murine primary colonic 37 mouse GPR119 receptors of 7.9 versus 0.18 μM, respectively) cultures, but not in insulin-secreting pancreatic cells. We also exhibited similar increased potency (5.7 versus 0.1 μM) in C57BL/6J observed larger phloridzin responses in ZDF compared with SD colon mucosa. GPR119-induced PYY-Y1 signalling was consistent colon, potentially revealing alterations in colonic SGLT1 expression in diseased and lean rodent tissue and the responses to caused by disease. Normally, glucose is cleared from the lumen in PSN-GPR119 were equivalent in each gut region. The lack of the upper small intestine, so glucose reaching the colon in ZDF apparent sidedness to PSN-GPR119 mucosal responses was most rats would manifest as larger responses to phloridzin. Upregula- likely due to the high permeability of PSN-GPR119, which tion of colonic SGLT1 activity indicates malabsorption of glucose precluded determination of the exact location of GPR119 (that in the ZDF model. is, apical alone or both apical and basolateral membranes). Consistent with our in vivo studies, downstream peptide If GPR119 receptors are apically targeted, then their stimulation by signalling pathways (PYY-mediated) were clearly functionally gut-restricted agonists should limit off-target effects and minimise intact during disease, as the responses to PSN-GPR119 in both potential detrimental effects, for example, on fatty acid meta- diabetic ZDF and obese and diabetic ob/ob models were bolism in skeletal muscle.38 maintained, in line with our hypothesis. The same PSN-GPR119- Interestingly, we observed reduced Y1 tone in ob/ob colon, initiated PYY-Y1 signalling occurred in human colon, and was potentially a consequence of reduced PYY analogue, neuro- comparable to the glucose-sensitive PSN632408 signalling 23 peptide Y and Y1 colonic expression. Imai and co-workers have described previously. These findings suggest that intestinal shown reduced neuropeptide Y and Y1 expression in pancreatic GPR119 stimulation has the potential for glucose lowering in man 39 islets of ob/ob mice. Additionally, GPR119 responses (with 1 μM and may be comparable to that observed in rodent T2D models PSN-GPR119) were slightly larger in ZDF versus SD colon. Although here and in other recent studies.30,40,41 GPR119 expression is known to alter in disease and is increased in In summary, PSN-GPR119 is a potent GPR119 tool compound islets from obese rats,13 it would be interesting to determine the that significantly lowers post-prandial glucose in lean and receptor levels in ZDF and SD rat intestine. In lean and diseased diseased models of T2D; luminal delivery of PSN-GPR119 to mucosae, PSN-GPR119 responses were increased consistently in isolated loops of rat small intestine stimulates secretion of GIP,

© 2014 Macmillan Publishers Limited International Journal of Obesity (2014) 1365 – 1373 Pharmacology of a novel GPR119 agonist, PSN-GPR119 S Patel et al 1372 mimic more closely, the success of bariatric surgery in treating T2D through a gut-based therapeutic approach. GPR119 compounds are already in the clinic for the treatment of T2D and human data have been reported7,42 ( http://clinicaltrials.gov) showing modest glucose-lowering activity.

CONFLICT OF INTEREST The authors declare no conﬂict of interest.

ACKNOWLEDGEMENTS SP, OJM, JW, MS, UWB and T-AC were employees of Prosidion Ltd at the time of the study, and we thank RenaSci Ltd (Nottingham, UK) for conducting the oral glucose tolerance test studies.

AUTHOR CONTRIBUTIONS SP and HMC were responsible for the conception and design of the experiments and writing of the article. OJM, IRT and JW were also responsible for collection, analysis, interpretation and presentation of the data. UWB and T-AC were responsible for the oral glucose tolerance test studies. MS was responsible for reviewing the article.

REFERENCES 1 Fredriksson R, Hoglund PJ, Gloriam DE, Lagerstrom MC, Schioth HB. Seven evolutionarily conserved human rhodopsin G protein-coupled receptors lacking close relatives. FEBS Lett 2003; 554:381–388. 2 Overton HA, Babbs AJ, Doel SM, Fyfe MC, Gardner LS, GriffinGet al. Deorphanization of a G protein-coupled receptor for oleoylethanolamide and its use in the discovery of small-molecule hypophagic agents. Cell Metab 2006; 3: 167–175. 3 Chu ZL, Jones RM, He H, Carroll C, Gutierrez V, Lucman A et al. A role for beta-cell-expressed G protein-coupled receptor 119 in glycemic control by enhancing glucose-dependent insulin release. Endocrinology 2007; 148: 2601–2609. 4 Semple G, Fioravanti B, Pereira G, Calderon I, Uy J, Choi K et al. Discovery of the first potent and orally efficacious agonist of the orphan G protein-coupled receptor 119. J Med Chem 2008; 51: 5172–5175. 5 Yoshida S, Ohishi T, Matsui T, Shibasaki M. Identification of a novel GPR119 agonist, AS1269574, with in vitro and in vivo glucose-stimulated insulin secretion. Biochem Biophys Res Commun 2010; 400:437–441. 6 Jones RM, Leonard JN, Buzard DJ, Lehmann J. GPR119 agonists for the treatment of type 2 diabetes. Expert Opin Ther Pat 2009; 19:1339–1359. 7 Katz LB, Gambale JJ, Rothenburg PL, Vanapalli SR, Vaccaro N, Xi L et al. Effects of JNJ-38431055, a novel GPR119 receptor agonist, in randomized, double-blind, placebo-controlled studies in subjects with type 2 diabetes. Diabetes Obes Metab 2012; 14:709–716. 8 Chu ZL, Carroll C, Chen R, Alfonso J, Gutierrez V, He H et al. N-oleoyldopamine Figure 6. The effect of PSN-GPR119 on normal human colon mucosa enhances glucose homeostasis through the activation of GPR119. Mol Endocrinol (a) concentration–response relationship for PSN-GPR119 (added 2010; 24:161–170. apically, n = 3) and (b)Y receptor blockade (300 nM BIBO3304; 1 9 Flock G, Holland D, Seino Y, Drucker DJ. GPR119 regulates murine glucose +BIBO) inhibits GPR119 agonism (1 μM PSN-GPR119, apical) ± aY 2 homeostasis through incretin receptor-dependent and independent mechanisms. receptor antagonist (1 μM BIIE0246; +BIBO & BIIE). (c) Glucose Endocrinology 2011; 152:374–383. sensitivity of apical PSN-GPR119 (1 μM) responses in the presence 10 Ning Y, O'Neill K, Lan H, Pang L, Shan LX, Hawes BE, Hedrick JA. Endogenous and of glucose both sides (11.1 mM) or, with apical mannitol (11.1 mM)in synthetic agonists of GPR119 differ in signalling pathways and their effects on place of glucose. Subsequent phloridzin (Phlor, 50 μM, apically) and insulin secretion in MIN6c4 insulinoma cells. Br J Pharmacol 2008; 155: PYY (100 nM) responses are also shown. Values are the mean ± 1 s.e. 1056–1065. m. from n numbers in parenthesis. Statistical differences are shown, 11 Chu ZL, Carroll C, Alfonso J, Gutierrez V, He H, Lucman A et al. A role for intestinal *Po0.05. endocrine cell-expressed G protein-coupled receptor 119 in glycemic control by enhancing glucagon-like peptide-1 and glucose-dependent insulinotropic pep- GLP-1 and PYY; and GPR119 agonism activates anti-secretory tide release. Endocrinology 2008; 149: 2038–2047. 12 Lauffer LM, Iakoubov R, Brubaker PL. GPR119 is essential for oleoylethanolamide- responses through PYY-mediated Y1 receptor activation in colonic mucosa of lean and T2D rodent models. PSN-GPR119 activity in induced glucagon-like peptide-1 secretion from the intestinal enteroendocrine 58 – human colon was comparable to that of lean and T2D rodent L cell. Diabetes 2009; : 1058 1066. models and we conclude that L-cell-derived PYY, in addition to 13 Sakamoto Y, Inoue H, Kawakami S, Miyawaki K, Miyamoto T, Mizuta K, Itakura M. Expression and distribution of Gpr119 in the pancreatic islets of mice and rats: GIP and GLP-1, may contribute towards glucose lowering in man. predominant localization in pancreatic polypeptide-secreting PP-cells. Biochem The benefits of releasing more than one endogenous peptide Biophys Res Commun 2006; 351:474–480. from the enteroendocrine cell population in a glucose-dependent 14 Lan H, Vassileva G, Corona A, Liu L, Baker H, Golovko A et al. GPR119 is required manner could offer a therapeutic advantage to modulate for physiological regulation of glucagon-like peptide-1 secretion but not for insulinotropic pathways with reduced risk of hypoglycemic, and metabolic homeostasis. J Endocrinol 2009; 201:219–230.

International Journal of Obesity (2014) 1365 – 1373 © 2014 Macmillan Publishers Limited Pharmacology of a novel GPR119 agonist, PSN-GPR119 S Patel et al 1373 15 Rogers GJ, Tolhurst G, Ramzan A, Habib AM, Parker HE, Gribble FM, Reimann F. 30 Zhang M, Feng Y, Wang J, Zhao J, Li T, He M et al. High-throughput screening for Electrical activity-triggered glucagon-like peptide-1 secretion from primary GPR119 modulators identiﬁes a novel compound with anti-diabetic efﬁcacy in murine L-cells. J Physiol 2011; 589: 1081–1093. db/db mice. PLoS One 2013; 8: e63861. 16 Hansen KB, Rosenkilde MM, Knop FK, Wellner N, Diep TA, Rehfeld JF et al. 31 Nauck M, Stockmann F, Ebert R, Creutzfeldt W. Reduced incretin effect in type 2 2-Oleoyl Glycerol is a GPR119 agonist and signals GLP-1 release in humans. J Clin (non-insulin-dependent) diabetes. Diabetologia 1986; 29:46–52. Endocrinol Metab 2011; 96: E1409–E1417. 32 Nauck MA, Heimesaat MM, Orskov C, Holst JJ, Ebert R, Creutzfeldt W. Preserved 17 Sundler F, Ekblad E, Håkanson R. Localisation and colocalisation of GI peptides. In: incretin activity of glucagon-like peptide 1 [7-36 amide] but not of synthetic Handbook of Experimental Pharmacology Gastrointestinal Regulatory Peptides human gastric inhibitory polypeptide in patients with type-2 diabetes mellitus. Brown DR (ed) Springer-Verlag: Berlin, Germany, 1993; 106:pp1–28. J Clin Invest 1993; 91: 301–307. 18 Chan JL, Stoyneva V, Kelesidis T, Raciti P, Mantzoros CS. Peptide YY levels are 33 Brubaker PL. Minireview: update on incretin biology: focus on glucagon-like decreased by fasting and elevated following caloric intake but are not regulated peptide-1. Endocrinol 2010; 151: 1984–1989. by leptin. Diabetologia 2006; 49: 169–173. 34 Willms B, Werner J, Holst JJ, Orskov C, Creutzfeldt W, Nauck MA. Gastric 19 Tough IR, Forbes S, Tolhurst R, Ellis M, Herzog H, Bornstein J, Cox HM. Endogenous emptying, glucose responses, and insulin secretion after a liquid test meal: peptide YY and neuropeptide Y inhibit colonic ion transport, contractility and effects of exogenous glucagon-like peptide-1 (GLP-1)-(7-36) amide in type 2

transit differentially via Y1 and Y2 receptors. Br J Pharmacol 2011; 164:66–79. (noninsulin-dependent) diabetic patients. J Clin Endocrinol Metab 1996; 81: 20 Liu CD, Aloia T, Adrian TE, Newton TR, Bilchik AJ, Zinner MJ et al. Peptide YY: a 327–332. potential proabsorptive hormone for the treatment of malabsorptive disorders. 35 Drucker DJ. Minireview: the glucagon-like peptides. Endocrinology 2001; 142: Am Surg 1996; 62: 232–236. 521–527. 21 Cox HM, Tough IR. Neuropeptide Y, Y1, Y2 and Y4 receptors mediate Y agonist 36 Pedersen J, Ugleholdt RK, Jorgenson SM, Windelov JA, Grunddal KV, Schwartz TW responses in isolated human colon mucosa. Br J Pharmacol 2002; 135: 1505–1512. et al. Glucose metabolism is altered after loss of L cells and α-cells 22 Batterham RL, Cohen MA, Ellis SM, Le Roux CW, Withers DJ, Frost GS et al. but not influenced by loss of K cells. Am J Physiol Endocrinol Metab 2012; 304: Inhibition of food intake in obese subjects by peptide YY3-36. N Engl J Med 2003; E60–E73. 349: 941–948. 37 Lan H, Lin HV, Wang CF, Wright MJ, Xu S, Kang L et al. GPR119 agonists mediate 23 Cox HM, Tough IR, Woolston AM, Zhang L, Nguyen AD, Sainsbury A, Herzog H. GLP-1 secretion from mouse enteroendocrine cells through glucose-independent Peptide YY is critical for acylethanolamine receptor Gpr119-induced activation of pathways. Br J Pharmacol 2012; 165: 2799–2807. gastrointestinal mucosal responses. Cell Metab 2010; 11:532–542. 38 Cornall LM, Mathai ML, Hryciw DH, McAinch AJ. Is GPR119 agonism an 24 Overton HA, Babbs AJ, Doel SM, Fyfe CT, Gardiner LS, GriffinGet al. appropriate treatment modality for the safe amelioration of metabolic diseases? Deorphanization of a G protein-coupled receptor for oleoylethanolamide and its use Expert Opin Investig Drugs 2013; 22:487–498. in the discovery of small molecule hypophagic agents. Cell Metab 2006; 3:167–175. 39 Imai Y, Patel HR, Hawkins EJ, Doliba NM, Matschinsky FM, Ahima RS. Insulin 25 Mace OJ, Schindler M, Patel S. The regulation of K- and L cell activity by GLUT2 secretion is increased in pancreatic islets of neuropeptide Y-deficient mice. and the calcium-sensing receptor CasR in rat small intestine. J Physiol 2012; 590: Endocrinology 2007; 148: 5716–5723. 2917–2936. 40 Semple G, Ren A, Fioravanti B, Periera G, Calderon I, Choi K et al. Discovery of 26 Cox HM, Pollock EL, Tough IR, Herzog H. Multiple Y receptors mediate pancreatic fused bicyclic agonists of the orphan G-protein coupled receptor GPR119 with in polypeptide responses in mouse colon mucosa. Peptides 2001; 22: 445–452. vivo activity in rodent models of glucose control. Bioorg Med Chem Letts 2013; 21: 27 Cox HM, Cuthbert AW. Neuropeptide Y antagonises secretagogue evoked 3134–3141. chloride transport in rat jejunal epithelium. Pflugers Arch 1988; 413:38–42. 41 Ansarullah Yu L, Holstein M, DeRuyter B, Rabinovitch A, Guo Z. Stimulating β-cell 28 Cox HM, Cuthbert AW, Håkanson R, Wahlestedt C. The effect of neuropeptide Y regeneration by combining a GPR119 agonist with a DPP-IV inhibitor. PLoS One and peptide YY on electrogenic ion transport in rat intestinal epithelia. J Physiol 2013; 8: e53345. 1988; 398:65–80. 42 Goodman ML, Dow J, van Vliet AA, Hadi S, Karbiche D, Lockton JA. The novel 29 Tough IR, Holliday ND, Cox HM. Y(4) receptors mediate the inhibitory responses of GPR119-receptor agonist PSN821 shows glucose lowering and decreased energy pancreatic polypeptide in human and mouse colon mucosa. J Pharmacol Exp Ther intake in patients with T2DM after 14 days treatment. Diabetes 2011; 60: 2006; 319:20–30. A273 (Abstract).

Supplementary Information accompanies this paper on International Journal of Obesity website (http://www.nature.com/ijo)