The Pharmacology of TD-8954, a Potent and Selective 5-HT4 Receptor Agonist with Gastrointestinal Prokinetic Properties

This study evaluated the in vitro and in vivo pharmacological properties of TD-8954, a potent and selective 5-HT4 receptor agonist. TD-8954 had high affinity (pKi = 9.4) for human recombinant 5-HT4(c) (h5-HT4(c)) receptors, and selectivity (>2,000-fold) over all other 5-hydroxytryptamine (5-HT) receptors and non-5-HT receptors, ion channels, enzymes and transporters tested (n = 78). TD-8954 produced an elevation of cAMP in HEK-293 cells expressing the h5-HT4(c) receptor (pEC50 = 9.3), and contracted the guinea pig colonic longitudinal muscle/myenteric plexus preparation (pEC50 = 8.6). TD-8954 had moderate intrinsic activity in the in vitro assays. In conscious guinea pigs, subcutaneous administration of TD-8954 (0.03–3 mg/kg) increased the colonic transit of carmine red dye, reducing the time taken for its excretion. Following intraduodenal dosing to anesthetized rats, TD-8954 (0.03–10 mg/kg) evoked a dose-dependent relaxation of the esophagus. Following oral administration to conscious dogs, TD-8954 (10 and 30 μg/kg) produced an increase in contractility of the antrum, duodenum, and jejunum. In a single ascending oral dose study in healthy human subjects, TD-8954 (0.1–20 mg) increased bowel movement frequency and reduced the time to first stool. It is concluded that TD-8954 is a potent and selective 5-HT4 receptor agonist in vitro, with robust in vivo stimulatory activity in the gastrointestinal (GI) tract of guinea pigs, rats, dogs, and humans. TD-8954 may have clinical utility in patients with disorders of reduced GI motility.

INTRODUCTION 5-Hydroxytryptamine (5-HT) plays a critical role in coordinating gastrointestinal (GI) transit (Hansen and Skadhauge, 1997;Grider et al., 1998;Jin et al., 1999;Baker, 2005). In response to intestinal stretching and mucosal stimulation, 5-HT is released from enterochromaffin cells of the mucosal epithelium, and promotes peristalsis via activation of intrinsic primary afferent neurons located in the submucous plexus (Kirchgessner et al., 1992;Foxx-Orenstein et al., 1995;Gershon and Tack, 2007). Of the 5-HT receptors believed to influence GI motility (e.g., 5-HT 1A , 5-HT 2A , 5-HT 2B , 5-HT 3 , and 5-HT 4 ; Beattie and Smith, 2008), the 5-HT 4 receptor subtype is considered particularly important, both physiologically and pathophysiologically (Kadowaki et al., 1996;Kim and Camilleri, 2000;Baker, 2005). The peristaltic reflex, for example, is dependent on activation of 5-HT 4 receptors on intrinsic primary afferent neurons, interneurons, and motor neurons within the gut wall, which results in the coordinated release of acetylcholine, substance P, and calcitonin gene-related peptide. The release of these agents is associated with synchronized contraction and relaxation of GI smooth muscle, and propulsion of luminal contents (Jin et al., 1999;Gershon and Tack, 2007). Additionally, activation of 5-HT 4 receptors in the smooth muscle of the human distal colon, and on enteric neurons or enterocytes promotes a direct relaxant effect and fluid secretion, respectively, further supporting GI transit (Hillier et al., 1994;Hansen and Skadhauge, 1997).
Agents interacting with 5-HT 3 and 5-HT 4 receptors have provided meaningful benefit to patients with GI functional disorders (Sanger, 2008). The GI prokinetic activity of 5-HT 4 receptor agonists, such as tegaserod (Zelnorm®), cisapride (Propulsid®), velusetrag (TD-5108), prucalopride (Resolor®), and mosapride has been demonstrated in a variety of species (Jin et al., 1999;Briejer et al., 2001b;Inui et al., 2002;Ji et al., 2003;Manini et al., 2009), and clinical efficacy has been established in patients with irritable bowel syndrome with constipation (IBS-C), chronic idiopathic constipation, functional dyspepsia, or gastroparesis (Deruyttere et al., 1987;Muller-Lissner, 1987;Abell et al., 1991;Camilleri, 2001;Johanson, 2004;Patel et al., 2004;Camilleri et al., 2008;Goldberg et al., 2010). Cisapride and tegaserod were used widely to treat upper and lower GI disorders of reduced motility, respectively, although their clinical efficacy in many patients was modest (Kellow et al., 1995;Evans et al., 2004), possibly reflecting their interactions with receptors other than the 5-HT 4 subtype (Briejer et al., 1995;Beattie et al., 2004;Beattie and Smith, 2008;De Maeyer et al., 2008). The clinical use of cisapride and tegaserod is now restricted on the basis of cardiovascular safety concerns (Barbey et al., 2002;Pasricha, 2007). Cisapride was associated with serious cardiac arrhythmias including ventricular tachycardia, ventricular fibrillation, and torsades de pointes as a result of its potent cardiac human ether-à-go-go-related gene (hERG) potassium channel inhibitory activity (Mohammad et al., 1997). The marketing of tegaserod was suspended in the United States due to concerns of an increased risk of serious ischemic cardiovascular events, although two recent epidemiological studies failed to identify such an association (Anderson et al., 2009;Loughlin et al., 2010). The newer generation, highly selective 5-HT 4 receptor agonists, velusetrag and prucalopride, produce robust clinical efficacy in chronic idiopathic constipation patients, have been well tolerated to date, and offer renewed hope for patients experiencing disorders of reduced GI motility (Camilleri et al., 2008;Sanger, 2009;Goldberg et al., 2010;Manabe et al., 2010). Prucalopride is approved in Europe for the treatment of women with chronic idiopathic constipation who have not responded adequately to laxatives.
In this study, the in vitro and in vivo pharmacodynamic properties of a structurally novel 5-HT 4 receptor agonist, amino]methyl}-piperidin-1-ylmethyl)piperidine-1carboxylic acid methyl ester; Figure 1), have been investigated. The preclinical activity of several standard 5-HT 4 agonists was evaluated in parallel for comparison to that of TD-8954.

MATERIALS AND METHODS
All animal experiments were conducted in accordance with the principles of good laboratory animal care provided by the Institutional Animal Care and Use Committees of Theravance, Inc. (rodent studies) or Drug Research Laboratories (dog study). The study protocol and consent form for the human single ascending dose study were reviewed by the clinical site's Institutional Review Board, and each volunteer provided written informed consent prior to initiation of study procedures. The study was conducted at a single site in accordance with the United States Code of Federal Regulations and the principles of the Declaration of Helsinki.
Binding data were analyzed by non-linear regression analysis using GraphPad Prism™ software (GraphPad Software, Inc., San Diego, CA, USA) and a three-parameter model for one-site competition. The pK i (negative decadic logarithm of K i ) values for test compounds were calculated from the best-fit IC 50 values, and the K d value of the radioligand, using the Cheng-Prusoff equation [Cheng and Prusoff, 1973; where [L] = radioligand concentration], and are reported as the mean ± SEM. Standard compounds were evaluated in parallel with TD-8954; the pK i values were consistent with those reported previously (Vickery et al., 2007;Smith et al., 2008).

Whole cell cAMP accumulation
Whole cell cAMP accumulation assays were performed as described previously  using a homogeneous radioimmunoassay (Flashplate Adenylyl Cyclase Activation Assay System; Perkin Elmer Life Sciences, Boston, MA, USA). HEK293-h5-HT 4(c) cells were lifted in Versene and collected by centrifugation (1,200 × g, 5 min) in phosphate-buffered saline (PBS). The cell pellet was resuspended gently in warm "stimulation buffer" (provided in the assay kit) and diluted to 5 × 10 5 cells/mL. Cells (25,000 per well) were incubated with test compound (10 pM to 100 μM) for 15 min at 37˚C in a 96-well Flashplate, in a total volume of 0.1 mL. In antagonist inhibition studies, cells were preincubated in the absence or presence of GR113808 (0.6, 1.7, or 5 nM) for 20 min at 37˚C prior to the addition of the test agonist. After the incubation period, [ 125 I]-cAMP was added in 100˚μL of ice-cold "detection buffer" to each well, according to the manufacturer's instructions. Bound radioactivity was quantified by scintillation counting and the amount of cAMP produced was extrapolated from a cAMP standard curve.
Data were analyzed by non-linear regression analysis with GraphPad Prism™ using the three-parameter sigmoidal concentration-effect model (slope constrained to unity). The potency of test agents was reported as a mean (±SEM) pEC 50 value (negative decadic logarithm of the effective concentration producing 50% of the maximum response), and the intrinsic activity (IA) as a mean (±SEM) percentage of the maximum 5-HT-evoked response. The pEC 50 and IA values for standard compounds (tested in parallel with TD-8954) were consistent with those reported previously (Vickery et al., 2007;Smith et al., 2008). Schild regression analysis was used to determine pK b values for GR113808 in the antagonist inhibition studies for 5-HT and TD-8954 (Kenakin, 1997). Concentration ratios (CR) were calculated as the ratio of the EC 50 values in the presence and absence of antagonist. The log[CR-1] was plotted against the log[GR113808], resulting in a linear relationship with the slope not significantly different from unity. The slope was therefore constrained to unity, and the X-intercept was extrapolated to provide a measure of the pK b value for GR113808.

5-HT 4 receptor selectivity
Off-target selectivity screening was conducted at Theravance, Inc. or at a contract research organization (CEREP, Paris, France). Conventional radioligand binding studies were conducted using, in the majority of assays, membranes prepared from cell lines transfected with the respective human recombinant receptor, ion channel, or transporter. The percent inhibition of specific binding by TD-8954, at a single concentration (1 μM, in duplicate), was determined. Conventional whole cell voltage-clamp techniques were used to examine the interaction of TD-8954 (3 μM) with neuronal (rat Na v 1.2) and cardiac (human Na v 1.5) voltage-gated sodium channels and hERG potassium channels (see Smith et al., 2006Smith et al., , 2008.
Tissues were washed three times (at 0, 15, and 30 min after mounting) and then challenged (at 45 min) with 5-HT at a concentration (0.3 μM) previously established to evoke a maximal contractile response. Once the contraction had reached its maximum, tissues were washed four times (2 min between each wash), and once more 10 min later. An additional priming challenge of 5-HT (0.3 μM) was made 15 min later. Following further washing prior to, and after a third 5-HT (0.3 μM) challenge, a cumulative concentration-response curve to 5-HT, tegaserod, mosapride, cisapride, prucalopride, or TD-8954 was constructed. TD-8954 concentration-response curves were also constructed in tissues previously exposed for 15 min to the selective 5-HT 4 receptor antagonist, piboserod (0.3 μM; Gaster et al., 1995) or vehicle. Changes in colonic isometric force were recorded by means of a force transducer (World Precision Instruments, model Fort 100), amplifier (Astro Med., model S48), and a data acquisition system (Biopac MP100, Acknowledge™ Waveform Acquisition and Analysis software). Contractile responses to test compounds were normalized to those of the last, primed 5-HT (0.3 μM) response in each tissue. Data were analyzed by non-linear regression analysis with GraphPad Prism using the four-parameter sigmoidal concentration-effect model. The pEC 50 and IA (i.e., the maximum compound-evoked response expressed as a percentage of the fully primed 5-HT maximum) were derived and reported for each test agent.

GUINEA PIG COLONIC TRANSIT
Adult, male Dunkin Hartley guinea pigs (220-300 g, Harlan, Chicago, IL, USA) were acclimated to their holding room (temperature controlled at 21 ± 1˚C and 12:12 h light-dark cycle commencing at 7 a.m.) for approximately 1 week prior to surgery. A standard guinea pig diet (7006, Harlan Teklad) and drinking water were provided ad libitum. Guinea pigs were anesthetized with isoflurane (2-3%, in oxygen) in an induction chamber. Maintenance of anesthesia was achieved with isoflurane (2-3%) administered via a nose cone. The mid-scapular area and abdomen were shaved and cleansed with betadine and 70% isopropanol. A small incision was made in the lower abdomen to expose the proximal colon. After isolation, a small incision was made in the proximal colon (approximately 2 cm from the cecum) and a cannula consisting of micro-renathane (MRE-040) tubing with a 2-cm silicone rubber tip (RenaSil™; 0.047 OD × 0.025 ID) was introduced and advanced approximately 2 cm toward the aboral end. A purse-string suture (Ethicon, 6-0 silk) was used to anchor the cannula in the colon and Baytril antibiotic (2.27%) was then applied topically to the colonic surgical site. The muscle layer was closed with a 4-0 Vicryl suture. The cannula was then secured to the nearby musculature with a 6-0 silk suture, tunneled under the skin and exteriorized at the midscapular region. The cannula was flushed and locked with sterile saline (Baxter), sealed with a sterile stainless steel pin, and secured to the back of the neck with a wound clip. The incisions in the peritoneum and abdomen were cleansed of blood and closed with a 3-0 Ethilon suture (Ethicon). Subcutaneous (s.c.) warmed lactated-Ringer's solution (3 mL) was administered immediately after surgery in addition to an intramuscular injection of the opioid analgesic, Buprenex® (buprenorphine, 0.03 mg/kg).
At least 5 days after surgery, guinea pigs were assigned randomly to a study group. Animals were dosed with test agent or vehicle (2 mL/kg s.c.), and 5 min later, each guinea pig was gently restrained and a non-absorbable marker (0.2 mL) was infused into the proximal colon via the implanted cannula. The marker consisted of 6 g of carmine red dye per 15 mL of carboxymethyl cellulose (0.5%). The study personnel were blinded to the treatment that each animal received. Animal cages were visually inspected for the presence of excreted red fecal pellets at 30-min intervals until each guinea pig had excreted pellets containing the red marker, or until 10 h had lapsed from the time of the marker injection. In the case that an animal failed to produce red fecal pellets within 10 h, the animal was left overnight in a clean cage and inspected the following morning. If excretion of dye occurred overnight, a value of 10 h was assigned. The whole colonic transit was defined as the time that lapsed between marker injection and the appearance of dye in the feces. Data for each treatment group were expressed as a mean percent increase (±SEM) in colonic transit time relative to vehicle-treated animals. Differences between treatment groups were determined using one-way analysis of variance www.frontiersin.org (ANOVA) with a Dunnett's post hoc test (p < 0.05 considered to be statistically significant).

RAT ESOPHAGEAL RELAXATION
The technique of digital sonomicrometry was used (Adelson and Million, 2004;Armstrong et al., 2006). Adult, male Sprague-Dawley rats (250-350 g, Harlan, Chicago, IL, USA) were acclimated to the colony room (temperature controlled at 21 ± 1˚C and 12:12 h light-dark cycle commencing at 7 a.m.) for at least 5 days prior to investigation. Standard rat diet (Harlan Teklad) and drinking water were available ad libitum. Rats were anesthetized with isoflurane (2-3%) in an induction chamber and anesthesia was then maintained with isoflurane (2-3%) via a nose cone for the duration of each experiment. Animals were placed, in a supine position, on a heated pad to maintain body temperature at 37-38˚C [monitored rectally with a sensor (Physitemp BAT-12)]. A midline incision was made in the skin and muscle layers of the abdomen, and the stomach and esophagus were exposed. A small incision was made in the upper duodenum approximately 1 cm from the pyloric sphincter, to permit intraduodenal (i.d.) administration of test agents. A micro-renathane catheter (MRE-40 with a 1 cm RenaSil rubber tip) was inserted approximately 1.5 cm into the duodenum via the incision and closed with 6-0 silk purse-string suture. Two piezoelectric crystals (1 mm diameter; Sonometrics Corp.) were gently glued, in a longitudinal orientation, to the distal esophagus (1 cm from the lower esophageal sphincter) using Vetbond tissue adhesive. The inter-crystal distance was approximately 2 mm. The wires connecting the crystals to the measurement device (Sonometrics Corp. TRX series 8) were exteriorized through the abdominal incision site, which was then closed with 4-0 Ethilon suture.
Following surgery, baseline esophageal tone was allowed to stabilize over approximately 30 min, prior to drug administration. The settings for the Sonometrics system were fixed within the Sonoview software (Sonometrics Corp., version 3.2.1) as follows: sampling rate = 99.4 Hz, transmit pulse = 375 ns, inhibit delay = 1.2-1.5 mm, velocity of sound through biological tissue = 1.59 mm/μs. Drug vehicle, followed by increasing doses of test compounds (0.5 mL/kg) were administered cumulatively via the duodenal cannula. Each dose was administered only when the esophageal response to the preceding dose had reached a maximum (typically 15-20 min). Changes in inter-crystal distance (in mm) from resting levels were averaged for each treatment group. The data were fitted to a sigmoidal dose-response relationship with variable slope, and an ED 50 value (i.e., the dose resulting in 50% of the maximum response) was calculated using Prism graphics software 3.0 (GraphPad, Inc.).

DOG GASTROINTESTINAL CONTRACTILITY
Five adult, female beagle dogs (6.5-8.5 kg, Marshall BioResources, North Rose, NY, USA), were maintained in a temperature controlled holding room at 21-24˚C and 12:12 h light-dark cycle commencing at 7 a.m. for at least 7 days prior to surgery. Dogs were fed at 8 a.m. and 6 p.m. with commercial dog food, while water was available ad libitum. Under pentobarbital anesthesia (30 mg/kg i.v. for induction, 1-2 mg/kg i.v. for maintenance), a strain gauge transducer (R. B. Products, Stillwater, MN, USA, calibrated with weights of 10-200 g) was sutured on to the antrum of the stomach (1 cm above the pylorus), duodenum, and jejunum (10-12 cm distal to the ligament of Treitz). Strain gauges were positioned to record circular muscle activity (Bass and Wiley, 1972). The strain gauge wiring was anchored securely to abdominal muscle with four stainless steel sutures. The abdominal incision was then closed around the wiring assembly. All dogs were treated with a fentanyl (25 μg) patch, which was applied 1 day prior to surgery and removed 3-4 days post-surgery. At least 2 weeks were allowed to lapse after surgery, before testing.
Dogs were fasted for 18 h prior to each study. The animals were positioned in slings, and the strain gauges were connected to a Grass (model 7P) polygraph to allow recording of the GI motility patterns. In fasted beagles, three phases of activity are evident in the antrum, duodenum and jejunum: quiescence followed by a pre-burst and then a burst period. After recording data for a full motility cycle, and confirming commencement of a quiescent period, each dog was dosed orally with TD-8954 (0.01 and 0.03 mg/kg), tegaserod (0.1 and 0.3 mg/kg), or vehicle (1 mL/kg over 5 s), followed by a 10 mL flush of water. After dosing, GI motility was monitored for 3 h. Dogs were allowed at least 4 days between treatments, the sequence of which was based on a Latin square design.
The area under the contractile force-time curve (AUC) of GI muscle contractions was measured directly from the polygraph tracings by means of a Numonics™ 2210 digitizing tablet in conjunction with an IBM 80486 computer and Sigma Scan™ analysis software (Jandel Scientific). Data were expressed as mean values (±SEM) for each treatment group.

HUMAN SINGLE ASCENDING DOSE STUDY
A Phase 1, double-blind, randomized, placebo-controlled, single ascending dose study was performed, consisting of seven sequential cohorts. In each cohort, eight healthy male and female subjects (18-50 years old) were randomized such that six subjects received a single oral dose of TD-8954 (0.1, 0.5, 1, 2, 5, 10, or 20 mg) and two subjects received placebo. The lowest dose of TD-8954 was formulated as an aqueous solution, while the other doses were administered as a powder in capsule. All subjects were confined to the clinical research unit from day 1 (admission) until after the 48-h post-dose safety assessments, and then returned on day 7 for a follow-up visit. A bowel diary was used to record the date and time of each bowel movement after dosing. The bowel diary was reviewed daily and before discharge. Blood samples were collected to assess the pharmacokinetics of TD-8954, and safety and tolerability were closely monitored.

MATERIALS
Standard biochemical and tissue culture reagents were purchased from Sigma-Aldrich (St Louis, MO, USA) and Invitrogen (Carlsbad, CA, USA) respectively. [ 3 H]GR113808 was purchased from Amersham Biosciences (Newark, NJ, USA). The Flashplate Adenylyl Cyclase Activation Assay System was purchased from Perkin-Elmer (Boston, MA, USA). TD-8954, prucalopride, mosapride, and piboserod were synthesized at Theravance, Inc. GR113808, tegaserod and cisapride were purchased from Tocris Cookson (Ellisville, MO, USA), Apin Chemicals (Abingdon, Oxon, UK) and Sequoia Research Products (Pangbourne, UK), respectively. Carbachol, 5-HT, and ketanserin were purchased from Sigma-Aldrich. For in vitro radioligand binding and cAMP accumulation studies, stock agonist solutions (10 mM) were prepared in DMSO, diluted to 400 μM with 50 mM HEPES (pH 7.4) at 25˚C, containing 0.1% BSA, and serial dilutions prepared in the same buffer. For isolated tissue studies, 10 mM solutions, prepared in DMSO, were diluted serially in sterile water. For in vivo studies, TD-8954 and prucalopride were dissolved in 5% dextrose in distilled water (D5W) or 0.9% saline, while tegaserod and mosapride were prepared in 10% sulfobutyl ether-beta cyclodextrin (10% SBE/CD), and cisapride was dissolved in 10% SBE/CD containing citrate (20 mM). Doses were expressed with respect to the free base weights of each compound.

5-HT 4 RECEPTOR SELECTIVITY
TD-8954 was >2,000-fold selective for h5-HT 4(c) receptors over other 5-HT receptors ( Table 2), and all non-5-HT receptors, transporters, ion channels, and enzymes tested ( Table 3). TD-8954 (3 μM) had no effect on hERG potassium currents (n = 6 cells) while cisapride (20 nM) was associated with a mean inhibition of 65% in the same cells. Exposure to TD-8954 (3 μM) for 3 min had no effect on the magnitude of the inward rat Na v 1.2a or human Na v 1.5 sodium currents (n = 3 cells for each).

GUINEA PIG COLONIC TRANSIT
In vehicle (2 mL/kg s.c.)-treated guinea pigs, the mean time taken for excretion of the first fecal pellet containing red dye was typically between 220 and 310 min. Following s.c. dosing, TD-8954, tegaserod, cisapride, mosapride (each at 0.03-3 mg/kg) and prucalopride (0.03-10 mg/kg) increased colonic transit, reducing the time taken for excretion of the dye, compared to vehicle-treated animals (Figure 3), although statistical significance (p < 0.05, oneway ANOVA with a Dunnett's post hoc test) was achieved only for all of the TD-8954 doses and for the 0.3 and 3 mg/kg prucalopride www.frontiersin.org doses. TD-8954 was more potent than tegaserod, prucalopride, cisapride, and mosapride, being significantly active at the lowest dose tested (0.03 mg/kg). At 0.03 mg/kg, TD-8954 had already achieved its maximum effect.

DOG GASTROINTESTINAL CONTRACTILITY
The quiescent phase of the antrum, duodenum, and jejunum motility cycle generally lasted for 50-60 min before transitioning into the pre-burst period (muscle contractions of gradually increasing magnitude occurring at random, with a duration of 30-60 min), followed by the burst period (vigorous and frequent contractions, with a duration of 5-15 min). Following oral administration of vehicle (1 mL/kg), there was little or no change in the activity of the antrum, duodenum, or jejunum; the expected motility patterns characteristic of fasted beagles were maintained throughout the observation period (Figures 5 and 6). TD-8954 (0.01 and 0.03 mg/kg) and tegaserod (0.1 and 0.3 mg/kg) produced increases in contractility in the antrum, duodenum, and jejunum ( Figure 5). The onset of contractile activity with TD-8954 occurred typically within 10 min of dosing (Figure 6).
Comparison of the activities of TD-8954 and tegaserod indicated that TD-8954 was statistically significantly more potent than tegaserod in the antrum, duodenum, and jejunum following oral administration (p < 0.05, ANOVA, followed by Dunnett's post hoc test).

HUMAN SINGLE ASCENDING DOSE STUDY
In healthy human subjects, GI prokinetic effects of TD-8954 (0.1-20 mg) were observed (Figure 7). The number of bowel movements from 0 to 24 h after each TD-8954 dose was increased significantly relative to placebo (p < 0.03 upon comparison of each TD-8954 treated group and placebo, based on Wilcoxon rank sum test). Compared to placebo, each TD-8954 dose was associated with a statistically significant reduction in the time to first bowel movement (p < 0.05 for all treatment groups comparing difference in survival function between each TD-8954 treated group and placebo obtained by log-rank test). www.frontiersin.org

DISCUSSION
Disorders of reduced GI motility, such as IBS-C, chronic idiopathic constipation, gastroparesis, post-operative ileus, and functional dyspepsia have a significant impact on the quality of life of affected individuals. Prior to their removal from the market, cisapride and tegaserod provided some relief to patients afflicted with these disorders (Prather et al., 2000;Evans et al., 2004;Ford et al., 2009). Limitations in their efficacy have been attributed to a lack of selectivity for the 5-HT 4 receptor subtype De Maeyer et al., 2008). Cisapride is a potent 5-HT 2A and 5-HT 2B receptor antagonist, while tegaserod has affinity for, and/or antagonist potency at, 5-HT 1B , 5-HT 2A , and 5-HT 2B receptors (Buchheit et al., 1995;Beattie et al., 2004;De Maeyer et al., 2008). Interactions of cisapride or tegaserod with non-5-HT receptors have also been proposed to underlie cardiac arrhythmic or ischemic adverse events. It is now well established that cisapride can induce ventricular tachycardia, ventricular fibrillation, and torsades de pointes as a result of potent cardiac hERG potassium channel inhibitory activity, particularly when cytochrome P 450 3A4 substrates are co-administered (Mohammad et al., 1997). The perceived risk of cardiovascular ischemic events with tegaserod, initially identified upon reviewing clinical trial data, has been questioned recently (Anderson et al., 2009;Loughlin et al., 2010). It is clear, however, that a significant unmet medical need remains for new therapeutic agents that will be efficacious and well tolerated in patients with functional GI disorders. One approach has been the development of selective 5-HT 4 receptor agonists, such as velusetrag (Beattie et al., 2008a;Smith et al., 2008;Goldberg et al., 2010), prucalopride (Briejer et al., 2001a;Camilleri et al., 2008), and TD-8954 (Beattie et al., 2008b), which, it is believed, will provide robust efficacy with acceptable tolerability and safety for patients. TD-8954, the subject of this study, is one of the most potent and selective 5-HT 4 receptor agonists described to date. The in vitro data demonstrated that TD-8954 had high affinity and potency at the human 5-HT 4 receptor (pK i and pEC 50 values of 9.4 and 9.3, respectively). In each assay, TD-8954 was significantly more potent than each of the comparator 5-HT 4 agonists tested in parallel, an observation consistent with that made previously in experiments with human isolated colonic circular muscle preparations (Beattie et al., 2008b). Activation of the 5-HT 4 receptor by TD-8954 was confirmed in vitro by antagonist inhibition, where the GR113808 pK b value was consistent with the [ 3 H]-GR113808 binding pK d and antagonist pK b values reported previously against 5-HT and tegaserod , providing further evidence of a 5-HT 4 receptor-mediated elevation of cAMP. TD-8954 had moderate IA at human recombinant 5-HT 4 receptors and endogenous 5-HT 4 receptors in the guinea pig colonic LMMP. Optimal target activation to induce a physiological response is dependent upon, amongst other things, the IA of an agonist and the receptor reserve in the target tissue (Grimwood and Hartig, 2009). While TD-8954 had moderate 5-HT 4 receptor agonist IA in vitro, robust in vivo GI prokinetic or 5-HT 4 agonist activity, to at least the same degree as standards with higher IA in vitro, was noted in guinea pigs, rats, and dogs. The data suggest that a particularly low level of agonist activity in the in vitro assays (e.g., such as that evident with mosapride) is required before reduced efficacy is observed in vivo, consistent with a large 5-HT 4 receptor reserve. TD-8954 had >2,000-fold selectivity for the human 5-HT 4(c) receptor over all other targets tested. Notably, TD-8954, unlike tegaserod or cisapride, had no affinity for 5-HT 1B , 5-HT 2A , 5-HT 2B , or 5-HT 3 receptors. Also, in contrast to cisapride, TD-8954 had no inhibitory effect at the hERG potassium channel.
The high 5-HT 4 receptor agonist potency of TD-8954 was also evident in vivo. TD-8954 (0.03-3 mg/kg s.c.) produced a statistically significant increase in colonic transit in conscious guinea pigs. Following TD-8954 administration and injection of carmine red dye into the proximal colon, the time for excretion of the first fecal pellet containing the marker was markedly reduced. The potent prokinetic activity of TD-8954 in this guinea pig model is in keeping with the proposed role of 5-HT and 5-HT 4 receptor activation in promoting GI motility (Muller-Lissner, 1987;Jin et al., 1999;Briejer et al., 2001b;Ji et al., 2003). The TD-8954-induced relaxation of the rat esophagus in this study is also consistent with agonist activity at the 5-HT 4 receptor (Triggle et al., 1988;Reeves et al., 1991). The technique of digital sonomicrometry provides a sensitive method to demonstrate the 5-HT 4 receptor agonistmediated changes in rat esophageal tone (Armstrong et al., 2006). In conscious, fasted dogs, TD-8954 (10 and 30 μg/kg) produced a dose-dependent increase in contractility of the antrum, duodenum, and jejunum following oral administration. TD-8954 was clearly more potent than tegaserod throughout the canine GI tract, consistent with its superior 5-HT 4 agonist potency in vitro and oral pharmacokinetic properties (Theravance, Inc., data on file) in this species. Considering the published data from similar models (Gullikson et al., 1993;Briejer et al., 2001b;Tazawa et al., 2002), the findings of this study are entirely consistent with activation of 5-HT 4 receptors by TD-8954 in the canine GI tract.
Based on the positive preclinical pharmacodynamic effects of TD-8954, a single ascending dose study was performed in healthy human subjects. The data demonstrated that TD-8954 was associated with an increase in bowel movement frequency and a reduction in the time to first stool compared to placebo. The high potency noted in the preclinical assays was also noted clinically; a dose as low as 0.1 mg was associated with a prokinetic effect, and 0.5 mg produced a maximal response. The free plasma C max and AUC 0-24 values for TD-8954 following a single dose of 0.5 mg to humans are 3.4 and 36.2 nM/h, respectively. Assuming that the free concentration of TD-8954 in plasma is equivalent to that at its site of action in the GI tract, mean and maximal 5-HT 4 receptor occupancies in the 24-h period following a 0.5 mg dose should be approximately 80 and 90%, respectively. This apparent requirement for a high level of receptor occupancy to achieve a maximal agonist response (Grimwood and Hartig, 2009) is consistent with the moderate IA of TD-8954 demonstrated preclinically in the in vitro assays.

CONCLUSION
TD-8954 is a potent and selective 5-HT 4 receptor agonist in vitro with robust in vivo GI activity in guinea pigs, rats, dogs, and notably humans. As a result of its demonstrated prokinetic activity in healthy human subjects, TD-8954 may have value in the treatment of patients with disorders of reduced GI motility.