Abstract
Antagonists of the vanilloid receptor TRPV1 (transient receptor potential vanilloid type 1) have been reported to produce antihyperalgesic effects in animal models of pain. These antagonists, however, also caused concomitant hyperthermia in rodents, dogs, monkeys, and humans. Antagonist-induced hyperthermia was not observed in TRPV1 knockout mice, suggesting that the hyperthermic effect is exclusively mediated through TRPV1. Since antagonist-induced hyperthermia is considered a hurdle for developing TRPV1 antagonists as therapeutics, we investigated the possibility of eliminating hyperthermia while maintaining antihyperalgesia. Here, we report four potent and selective TRPV1 modulators with unique in vitro pharmacology profiles (profiles A through D) and their respective effects on body temperature. We found that profile C modulator, (R,E)-N-(2-hydroxy-2,3-dihydro-1H-inden-4-yl)-3-(2-(piperidin-1-yl)-4-(trifluoromethyl)phenyl)acrylamide (AMG8562), blocks capsaicin activation of TRPV1, does not affect heat activation of TRPV1, potentiates pH 5 activation of TRPV1 in vitro, and does not cause hyperthermia in vivo in rats. We further profiled AMG8562 in an on-target (agonist) challenge model, rodent pain models, and tested for its side effects. We show that AMG8562 significantly blocks capsaicin-induced flinching behavior, produces statistically significant efficacy in complete Freund's adjuvant- and skin incision-induced thermal hyperalgesia, and acetic acid-induced writhing models, with no profound effects on locomotor activity. Based on the data shown here, we conclude that it is feasible to modulate TRPV1 in a manner that does not cause hyperthermia while maintaining efficacy in rodent pain models.
The vanilloid receptor TRPV1 emerged as a therapeutic target for pain based on the fact that 1) TRPV1 agonists cause pain (Szallasi and Blumberg, 1999; Jones et al., 2004), 2) TRPV1 expression is up-regulated during painful conditions (Ji et al., 2002; Matthews et al., 2004), 3) TRPV1 knockout mice display attenuated pain behaviors (Caterina et al., 2000; Davis et al., 2000; Keeble et al., 2005), and 4) TRPV1 antagonists reverse pain behavior in rodent models of pain (Walker et al., 2003; Gavva et al., 2005b; Ghilardi et al., 2005; Honore et al., 2005; Rami et al., 2006). We have recently reported the identification of AMG 517, a selective and orally bioavailable TRPV1 antagonist, as a clinical candidate. AMG 517 acted as a potent antagonist of TRPV1 in vitro, blocked capsaicin-induced flinch in rats (an “on-target” challenge model), and acted as an antihyperalgesic in a model of inflammatory pain (CFA-induced thermal hyperalgesia) (Doherty et al., 2007; Gavva et al., 2007a). Similar to other TRPV1 antagonists, AMG 517 caused hyperthermia in rodents, dogs, and monkeys, but hyperthermia attenuated upon repeated administration of AMG 517 (Gavva et al., 2007a). During phase I clinical trials in healthy volunteers, AMG 517 elicited a marked but reversible hyperthermia. Furthermore, AMG 517 administered after third molar extraction, a surgical cause of acute pain, elicited long-lasting hyperthermia, with body temperature surpassing 40°C, demonstrating that TRPV1 blockade elicits an undesirable magnitude of hyperthermia in some individuals (Gavva et al., 2008).
Due to the undesirable hyperthermia observed for AMG 517 in humans, we examined the effect of TRPV1 antagonists on body temperature regulation more closely and sought possible ways to eliminate this on-target effect. Initially, we examined the effects of several TRPV1 antagonists represented by various chemotypes, including cinnamides, ureas, amides, benzimidazoles, and piperazine carboxamides, on rat body temperature. We found that all chemotypes that we have tested cause 0.5 to 1.5°C increase in body temperature, suggesting that antagonist-induced hyperthermia is not chemotype-specific but rather occurs because of TRPV1 blockade in vivo (Gavva et al., 2007b). Furthermore, TRPV1-selective antagonists (AMG0347 and AMG 517) did not cause hyperthermia in TRPV1 knockout mice, demonstrating that entire hyperthermic effect was TRPV1-mediated (Steiner et al., 2007). The second approach that we examined was to study the effect of peripherally restricted TRPV1 antagonists. We postulated that we may be able to separate the analgesic effect of TRPV1 antagonists from their hyperthermic effect if the TRPV1 antagonists were excluded from the central nervous system. Unfortunately, this approach also did not prove to be successful because peripheral restriction of antagonists did not eliminate hyperthermia, suggesting that the site of action is predominantly outside of the blood-brain barrier (Gavva et al., 2007b; Tamayo et al., 2008). The site of action for antagonist-induced hyperthermia to be present outside of blood-brain barrier was further confirmed by TRPV1 desensitization experiments (with resiniferatoxin) that demonstrated that visceral TRPV1 channels were responsible for antagonist-induced hyperthermia (Steiner et al., 2007). The final approach to eliminate TRPV1 antagonist-induced hyperthermia was to evaluate the compounds that display differential pharmacologies (i.e., compounds that differentially modulate distinct modes of in vitro TRPV1 activation such as capsaicin, pH 5, and heat) in vivo.
Here, we describe the differential pharmacology of four TRPV1 modulators and their effects on body temperature. We further characterized AMG8562, an orally bioavailable TRPV1 modulator and its effects on pain behavior in rodent models.
Materials and Methods
Reagents. All the cell culture reagents were purchased from Invitrogen (Carlsbad, CA). All TRPV1 antagonists used in this study were synthesized at Amgen Inc. (Thousand Oaks, CA) (Fig. 1).
Agonist-Induced 45Ca2+ Uptake Assay. Two days before the assay, cells were seeded in Cytostar 96-well plates (GE Healthcare, Chalfont St. Giles, UK) at a density of 20,000 cells/well. The activation of TRPV1 and TRPV2 was followed as a function of cellular uptake of radioactive calcium (45Ca2+; MP Biomedicals, Irvine, CA). Capsaicin (0.5 μM), pH 5, and heat (45°C) were used as agonists for TRPV1, and 2-APB (200 μM) was used as an agonist for TRPV2. To determine the ability of compounds to block agonist activation of TRPV1, compounds were incubated with CHO cells expressing the TRP channel for 2 min before the addition of agonist and 45Ca2+, and cells were washed after further incubation of 2 min to determine the 45Ca2+ uptake. All the antagonist 45Ca2+ uptake assays were conducted as reported previously (Gavva et al., 2007a) and had a final 45Ca2+ concentration of 10 μCi/ml. Radioactivity was measured using a MicroBeta Jet (PerkinElmer Life and Analytical Sciences, Boston, MA). Data were analyzed using GraphPad Prism 4.01 (GraphPad Software Inc., San Diego, CA).
Luminescence Readout Assay for Measuring Intracellular Calcium. Stable CHO cell lines expressing human TRPA1, TRPM8, TRPV3, and TRPV4 were generated using tetracycline-inducible T-REx expression system from Invitrogen. To enable a luminescence readout based on intracellular increase in calcium (Le Poul et al., 2002), each cell line was also cotransfected with pcDNA3.1 plasmid containing jelly fish aequorin cDNA. Twenty-four hours before the assay, cells were seeded in 96-well plates, and TRP channel expression was induced with 0.5 μg/ml tetracycline. On the day of the assay, culture media were removed, and cells were incubated with assay buffer (Ham's F-12 containing 30 mM HEPES for TRPA1, TRPM8, and TRPV3; Ham's F-12 containing 30 mM HEPES, 1 mM CaCl2, and 0.3% bovine serum albumin for TRPV4) containing 15 μM coelenterazine (P.J.K, Kleinblittersdorf, Germany) for 2 h. Antagonists were added for 2.5 min before addition of an agonist. The luminescence was measured by a charge-coupled device camera based FLASH-luminometer built by Amgen, Inc. The following agonists were used to activate TRP channels: 80 μM allylisothiocyanate for TRPA1, 1 μM icilin for TRPM8, 200 μM 2-APB for TRPV3, and 1 μM4α-PDD for TRPV4. Compound activity was calculated using GraphPad Prism 4.01 (GraphPad Software Inc.).
Animals. Adult male Sprague-Dawley rats weighing 150 to 250 g or CD-1 mice weighing 20 to 25 g (Charles River Laboratories, Inc., Wilmington, MA) were allowed at least 1 week of acclimation in the Amgen's Association for Assessment and Accreditation of Laboratory Animal Committee-approved animal care facility before being used. In one telemetry study, rats (n = 6/group; JVC from Taconic Farms, Germantown, NY) were treated with various TRPV1 antagonists in 100% dimethyl sulfoxide (1 ml/kg i.v.) or vehicle control. In subsequent telemetry, efficacy, and side effect studies in rats (n ≥ 6/group), various doses of TRPV1 antagonists in a dosing volume of 5 ml/kg in an Ora-Plus/5% Tween 80 suspension p.o. or gabapentin (Sigma-Aldrich, St. Louis, MO) in saline 2 ml/kg i.p. or vehicle controls were administered at the treatment times specified for each model. Mice (n ≥ 8/group) were treated with vehicle, TRPV1 antagonist, or ibuprofen (Sigma-Aldrich) in a dosing volume of 10 ml/kg in an Ora-Plus/5% Tween 80 suspension p.o. at the doses and treatment times specified for each model. Blood samples were collected immediately following behavioral testing for pharmacokinetic analysis.
As a standard operation procedure, capsaicin-induced flinch and pain models were conducted in a blinded manner by randomization of dosing groups, inclusion of a positive control, and experimenters recording behavioral endpoints not participating in dosing of compounds.
Capsaicin-Induced Flinch Model. AMG8562 at 0.1, 0.3, 1, and 3 mg/kg or AMG8163 at 3 mg/kg (used as the positive control) or vehicle was administered 2 h before right hind paw intraplantar injection of 0.5 mg/25 ml capsaicin (Sigma-Aldrich; in a volume of 25 μl of 5% ethanol in phosphate-buffered saline without Ca2+ and Mg2+). Immediately following the injection of capsaicin, the number of hind paw flinches was recorded over a 1-min period by investigators fully blinded to the randomized treatment conditions.
Radiotelemetry in Rats. Rats were single-housed upon arrival to our animal facility. Before implanting the radiotelemetry probe (model ER-4000 PDT; Mini Mitter, Bend, OR), rats were lightly anesthetized using isoflurane (IsoFlo; Abbott Laboratories, Abbott Park, IL) at a concentration of 2% isoflurane at 2 l/min oxygen flow. While positioned in right lateral recumbence, the left lateral abdominal wall was clipped and cleaned with Betadine solution (Purdue Frederick Company, Stamford, CT) followed by 70% alcohol in water. A 3 to 4-mm incision was made through the skin and abdominal wall, into which a sterilized probe was inserted. The surgical site was closed with absorbable suture material (4-0 Vicryl; Ethicon Inc., Somerville, NJ) and silk suture (4-0; Ethicon, Inc.). Animals were returned to home caging for a 3-day recovery period before the drug treatment experiments.
On the day of the experiment, single-housed animals were placed on the radiotelemetry receivers and acclimated to the test room for at least 1 h before the baseline recording. The temperature in the room used for radiotelemetry experiments was maintained at 20 ± 2°C. Baseline recording of core body temperature was recorded for 30 min before the drug administration. Rats (n = 6/group) were either treated with vehicle or TRPV1 antagonist, and the body core temperature recordings were continued for 2 to 4 h.
CFA-Induced Thermal Hyperalgesia. After 3 days of full habituation to the testing equipment, baseline thermal latencies were recorded in the rats using the modified Hargreaves' thermal stimulating apparatus (University of California San Diego, La Jolla, CA) followed by injection of CFA into the right hind paw (50 μl of 0.1% CFA). Twenty-one hours after the CFA injection, animals were dosed (p.o.) with vehicle or AMG8562 at 30 or 100 mg/kg. AMG8163 at 3 mg/kg was used as the positive control. Two hours after drug dosing (23 h after CFA injection), paw withdrawal latencies were measured by investigators fully blinded to the randomized treatment conditions.
Acetic Acid-Induced Writhing in Mice. Mice were habituated to the testing room in their home cages overnight. Behavioral observations occurred in clear plastic testing cylinders 30 cm in height and 30 cm in diameter, with opaque walls between cylinders, such that mice were not visible to each other (Langford et al., 2006). TRPV1 antagonist, vehicle, or 200 mg/kg ibuprofen was injected 1 h before acetic acid injection. Immediately following i.p. injection of 0.6% acetic acid in 10 ml/kg saline, the number of writhes (characteristic lengthwise abdominal constrictions with torso elongations) was counted for 30 min.
Open-Field Activity in Mice and Rats. Rats were habituated in a reversed light cycle room for at least 1 week and to the testing room for 1 h before dosing. Mice were housed in a regular light cycle room and habituated overnight to the dark testing room in their home cages. TRPV1 antagonist or vehicle was injected 1 h before placing mice or 1.5 h before placing rats in the open-field apparatus. Open-field activity was measured using a system that counts interruptions of a set of photobeams (Hamilton-Kinder, Poway, CA) for the course of 60 min. To begin a session, animals are removed from the home cage and placed individually into an independent Plexiglas box (16 × 16 × 15-inch box) surrounded by a frame consisting of 32 photocells (16Y and 16X) that track the movement of the animal. Photobeam breaks are used as an indication of activity and are broken in to the following parameters per minute: basic movements (beam breaks), distance traveled (centimeters), time spent (seconds), and number of repetitive beam breaks (i.e., stereotypic movement).
Statistical Analysis. Behavioral results were analyzed using one-way analysis of variance (ANOVA) with Dunnett's multiple comparisons post hoc test for significance relative to vehicle. Multiple ANOVAs (one for each time point) were used for analysis of body core temperature data. The development of thermal hyperalgesia in baseline versus CFA- or skin incision-treated rats versus vehicle as well as the effect of gabapentin versus saline in spinal nerve ligation-induced mechanical allodynia was compared using Student's t test. Logarithmic transformations were made of von Frey threshold data in spinal nerve ligation-induced mechanical allodynia as described above. Statistical calculations and graphs were made using Graph-Pad Prism 5.01 (GraphPad Software Inc.).
Results
TRPV1 Antagonists Exhibit Differential Pharmacology. It has been reported that some TRPV1 antagonists exhibit differential pharmacology at distinct modes of activation (McIntyre et al., 2001; Gavva et al., 2005a, 2007b). We have previously defined two profiles of TRPV1 antagonists: group A that blocks all modes of activation and group B that blocks capsaicin and heat but not proton activation (Gavva et al., 2005a). Here, we evaluated several compounds for their ability to modulate three distinct modes of TRPV1 activation in agonist-induced 45Ca2+ uptake assays using CHO cells stably expressing the rat TRPV1 channels (Fig. 2). AMG8163 potently inhibited 0.5 μM capsaicin, pH 5, and heat (45°C) activation of rat TRPV1 and represents “profile A” or group A (Table 1; Fig. 2, A–C) (Gavva et al., 2007b). AMG8563 inhibited capsaicin and heat activation but only partially blocked pH 5 activation similar to capsazepine and AMG0610 that represent “profile B” or “group B” antagonists (Gavva et al., 2005a) (Fig. 2, D–F). AMG8562 blocked capsaicin activation, did not affect heat activation, and potentiated pH 5 activation, thus representing a new modulation profile, “profile C” (Fig. 2, G–I). AMG7905 inhibited capsaicin activation and potentiated both heat and pH 5 activation of rat TRPV1, representing yet another new modulation profile, “profile D” (Fig. 2, J–L). Both AMG8562 and AMG7905 by themselves did not induce 45Ca2+ uptake into CHO cells expressing the rat TRPV1 channels at physiological pH (pH 7.2), indicating that they are not partial agonists. In addition to blocking capsaicin activation, all four compounds inhibited putative endogenous ligand [anandamide and N-arachidonyldopamine (NADA)] activation of TRPV1 (Table 1).
Potentiators of Proton Activation Do Not Cause Hyperthermia. We have recently demonstrated that TRPV1 antagonists that represent profiles A and B cause transient hyperthermia (Gavva et al., 2007b). Since we are interested in dissecting the pharmacology of TRPV1 antagonists to evaluate whether hyperthermia can be separated from their ability to act as antihyperalgesics, we tested all four molecules that represent profiles A through D for their effect on body temperature. In the first experiment, AMG8163, AMG8563, and AMG8562 (representing profiles A, B, and C, respectively) were administered intravenously to rats implanted with radiotelemetry probes to monitor body temperature (Fig. 3). Overall, there were significant effects of these compounds on body temperature (Fig. 3A) from 10 min after dosing (F3,20 = 14.0; p < 0.01) through 120 min after dosing (F3,20 = 42.3; p < 0.01). Both AMG8563 and AMG8163 produced a significant 1.0–1.5°C increase in body temperature as predicted for profile A and B antagonists (Gavva et al., 2007b). AMG8163 caused a maximal increase in temperature relative to vehicle of 1.1°C at 110 min (vehicle, 37.4 ± 0.2°C; AMG8163, 38.7 ± 0.1°C; p < 0.01). AMG8563 caused a maximal increase in temperature of 1.2°C at 40 min (vehicle, 37.3 ± 0.2°C; AMG8563, 38.5 ± 0.2°C; p < 0.01). AMG8562, however, produced a significant decrease of 1.0°C in body temperature at 20 min after administration (vehicle, 37.2 ± 0.1°C; AMG8562, 36.2 ± 0.3°C; p < 0.01).
We further profiled AMG8562 at 1, 3, 10, and 30 mg/kg doses by oral administration to rats to assess dose dependence and maximal temperature effects using the same route used for in vivo efficacy studies (Fig. 3B). During the first three time points after dosing, there is no significant effect of AMG8562 on body temperature (F4,25 = 0.65–2.5; p > 0.05). At 40 and 70 min after dosing, there is a significant effect of AMG8562 (F4,25 = 3.4 and 2.3, respectively; p < 0.05), with significance relative to vehicle of only the 30-mg/kg dose (p < 0.01). At 80 and 160 min after dosing, all four doses of AMG8562 significantly decrease body temperature (F4,25 = 3.8 and 3.3, respectively; p < 0.05). This decrease in body temperature was considered modest because the greatest change from vehicle was only –0.6°C (after 1 mg/kg at 160 min, vehicle group temperature was 37.5 ± 0.1°C; compared with the 30 mg/kg dose of AMG8562, group temperature was 36.8 ± 0.1°C). Oral administration of AMG8562 produced a smaller decrease in body temperature than intravenous administration. We assume that intravenous administration might have resulted in greater exposure compared with oral administration that resulted in a larger decrease in body temperature. Profile C compounds such as AMG8562 may cause decrease in body temperature as opposed to the hyperthermia elicited by profile A and B compounds, AMG8163 and AMG8563, respectively.
We also profiled AMG7905 at 0.3, 1, 3, 10, and 30 mg/kg doses by oral administration to assess dose dependence and maximal temperature effects (Fig. 3C). Beginning at 30 min after dosing and throughout the testing period, AMG7905 significantly and dose-dependently decreased body temperature at 30 mg/kg (F6,52 = 3.7; p < 0.01). The maximal effect occurred at 160 min after dosing (F6,52 = 40.0; p < 0.01), with vehicle group temperature of 37.1 ± 0.1°C and 30 mg/kg AMG7905-administered group temperature of 34.2 ± 0.4°C. Since AMG7905 potentiated both heat and pH 5 activation and caused a relatively greater magnitude of decrease in body temperature relative to vehicle, we concluded that potentiation of both heat and pH 5 seem to result not only in a lack of hyperthermia but also actually a marked hypothermia. TRPV1 modulator-induced hypothermia returned to baseline values within 12 h.
AMG8562 Is a Selective Modulator of TRPV1 with Acceptable Pharmacokinetic Properties. Since AMG8562 did not cause hyperthermia in rats, we evaluated its ability to modulate activation of closely related TRP channels in cell-based assays that measure agonist-induced increases in intracellular calcium in CHO cells recombinantly expressing the appropriate TRP channel. The IC50 value for AMG8562 was >20 μM against allyl isothiocyanate-activated TRPA1, icilin-activated TRPM8, 2-APB-activated TRPV2, and TRPV3. AMG8562 IC50 value against 4α-PDD-activated TRPV4 was approximately 3 μM (Fig. 4). In these assays, ruthenium red served as a positive control for TRPA1 (IC50 value of 367 nM), TRPV2 (IC50 value of 2028 nM), and TRPV4 (IC50 value of 24 nM). Compound M8-B and compound V3-H served as positive controls for TRPM8 (IC50 value of 5 nM) and TRPV3 (IC50 value of 72 nM), respectively. Similar to AMG8562, AMG7905 was also found to be selective for TRPV1 (IC50 against TRPA1, TRPM8, and TRPV1–4 was >4 μM). Similarly, AMG8562 IC50 value against human TRPA1, TRPM8, TRPV3, and TRPV4 was >4 μM (data not shown). We acknowledge that most of the selectivity data generated here were with exogenous ligands, and they may or may not necessarily correlate with yet to be discovered endogenous ligands for each of the TRP channels.
Interestingly, AMG8562 profile at human TRPV1 is different compared with its effect on rat TRPV1. AMG8562 behaved like a profile B antagonist of human TRPV1, i.e., blocked capsaicin and heat activation fully and partially blocked pH 5 activation of human TRPV1 (Fig. 5).
Furthermore, in a selectivity screen conducted by Cerep (Seattle, WA), a 10 μM concentration of AMG8562 did not show significant binding (≥45% inhibition) to any of the targets that included receptors, enzymes, and ion channels (Table 2). AMG8562 demonstrated acceptable metabolic stability in rat liver microsomes (CLin vitro = 61 μl/min/mg) and superior in vivo pharmacokinetic properties in rats (t1/2 = 4.4 h, CLin vivo = 0.85 liter/h/kg, and Foral = 61% at 5 mg/kg); hence, it was deemed suitable for oral administration during in vivo studies.
AMG8562 Potently Blocks Capsaicin-Induced Flinch Behavior in Rats. First, we evaluated the ability of AMG8562 to block capsaicin-induced flinch because other TRPV1 antagonists that potently block capsaicin activation in vitro block capsaicin-induced flinching in vivo (Seabrook et al., 2002; Gavva et al., 2007a). AMG8562 significantly and dose-dependently decreased capsaicin-induced flinching in rats (F7,80 = 15.7; p < 0.01; Fig. 6A), with a 100% blockade of flinching following the 3 mg/kg. Baseline number of flinches in the vehicle-administered group was 10.8 ± 1.7 flinches. The minimally effective dose to block capsaicin-induced flinch for AMG8562 was 1 mg/kg, which corresponds to a minimally effective plasma concentration of 120.2 ± 7.9 ng/ml (p < 0.05). The positive control, 3 mg/kg AMG8163, showed 100% blockade of flinching in the same experiment.
AMG8562 Significantly Reduces CFA-Induced Thermal Hyperalgesia. Several TRPV1 antagonists have been shown to block hyperalgesia induced by CFA (Pomonis et al., 2003; Walker et al., 2003; Gavva et al., 2005b; Honore et al., 2005). Since AMG8562 modulates TRPV1 differentially compared with the molecules used in the studies mentioned above, we evaluated AMG8562 effect on CFA-induced thermal hyperalgesia (Fig. 6B). Before CFA injection, baseline latency of all rats combined was 10.4 ± 0.1 s. Twenty-three hours after CFA injection, rats exhibited an increase in thermal sensitivity as shown by the decrease in latency in the vehicle-treated rats (3.6 ± 0.3 s) relative to their baseline latency (t22 = 21.4; p < 0.01). AMG8562 significantly reduced CFA-induced thermal hyperalgesia (F3,44 = 3.9; p > 0.05) at 30 and 100 mg/kg (p < 0.05). The corresponding plasma concentrations immediately after behavioral data collection were 4153 ± 172 and 7699 ± 281 ng/ml, respectively (Fig. 6B). Latency to respond to the thermal stimulus was increased to 5.8 ± 0.7 s following 30 mg/kg AMG8562 and to 5.8 ± 0.5 s following 100 mg/kg AMG8562. The maximal effect, expressed as percentage of reduction of the “window of hyperalgesia” (window being defined as vehicle treated post-CFA latency minus baseline latency), is 27 ± 9%. As shown previously (Gavva et al., 2007a), the positive control AMG8163 significantly reduced CFA-induced thermal hyperalgesia (p < 0.01), with latency increased to 6.2 ± 0.5 s, constituting a 35 ± 7% effect in this experiment with 31 ± 4 ng/ml plasma concentration.
AMG8562 Significantly Reduces Plantar Skin Incision-Induced Thermal Hyperalgesia. TRPV1 antagonists such A-425619 and AMG8163 have been reported to reverse thermal hyperalgesia induced by surgical incision (Honore et al., 2005; Magal et al., 2005). Hence, we evaluated and compared AMG8562 effects with AMG8163 in the same model (Fig. 6C). Before skin incision, baseline latency of all rats combined was 11.3 ± 0.1 s. Two hours postplantar skin incision, rats exhibited an increase in thermal sensitivity as shown by the decrease in latency in the vehicle-treated group to 3.6 ± 0.3 s relative to their baseline latency (t22 = 23.6; p < 0.01). AMG8562 significantly reduced thermal hyperalgesia in postincision rats (F3,44 = 5.6; p < 0.01). Latency to respond to the thermal stimulus was increased to 5.5 ± 0.6 s following 30 mg/kg and to 5.9 ± 0.5 s following 100 mg/kg (p < 0.05), thus constituting a significant reversal of hyperalgesia following both doses (p < 0.05), with maximal effect of 29 ± 7%. Corresponding plasma concentrations immediately following collection of behavioral data were 1786 ± 75 and 5918 ± 155 ng/ml, respectively (Fig. 6C). As shown previously (Magal et al., 2005), the positive control AMG8163 significantly reduced thermal hyperalgesia relative to vehicle-treated rats (p < 0.01), with latency increased to 6.6 ± 0.7 s, constituting a39 ± 9% effect in this experiment and with corresponding plasma concentration of 20 ± 1 ng/ml.
AMG8562 Reduces Acetic Acid-Induced Writhing. Since some TRPV1 antagonists have been reported to block acetic acid-induced writhing (Tang et al., 2007), we evaluated AMG8562 in this model (Fig. 6D). In mice, intraperitoneally injected acetic acid caused 36.4 ± 7.8 writhes. AMG8562 significantly reduced the number of writhes (F4,39 = 2.7; p < 0.05; Fig. 6D). The number of writhes observed in mice administered with 10, 30, and 100 mg/kg AMG8562 was 19.22 ± 5.3, 14.8 ± 5.7, and 15.0 ± 5.9, respectively. The corresponding plasma concentrations were 879 ± 111, 3024 ± 268, and 11029 ± 1094 ng/ml, respectively. Although there is a trend of decreased acetic acid-induced writhing relative to vehicle for all treatment groups, only 30 mg/kg AMG8562 and the positive control ibuprofen showed statistically significant reduction (p < 0.05). The percentage of reduction in writhing counts was 59% following 30 mg/kg AMG8562. Overall variability was high in this assay, with S.E.M. accounting for 21% of mean writhing counts in the vehicle group (compared with S.E.M. accounting for 9.2 and 7.2% of mean latency in the CFA- and skin incision-induced thermal hyperalgesia assays, respectively), thus decreasing the power to detect significant drug effects.
AMG8562 Does Not Affect Most Open-Field Activity Parameters in Mice and Rats. Because it is possible that compounds that reduce locomotor activity could seem to be “false-positive” antihyperalgesics or analgesics in pain models, we evaluated AMG8562 in open-field locomotor activity tests. Total distance traveled following vehicle treatment was 19,297 ± 2012 cm in mice and 15,212 ± 1980 cm in rats. There was no significant effect of 100 mg/kg (corresponding plasma concentration was 5416 ± 445 ng/ml) AMG8562 on total distance traveled in mice (F4,33 = 1.2; p > 0.05; Fig. 7A). There was no significant effect of 100 mg/kg (corresponding plasma concentration was 5659 ± 734 ng/ml) AMG8562 on total distance traveled in 60 min in rats (F3,28 = 0.31; p > 0.05; Fig. 7B). There was a significant 17.5 ± 1.6% increase in rest time in rats following 100 mg/kg AMG8562 (F3,28 = 3.0; p < 0.05) but not in mice (F3,26 = 1.5; p > 0.05) (data not shown). In both rats and mice, there was no significant effect on any other open-field parameters analyzed, including rest time, rearing (counts or seconds), fine movements, basic movements, and immobility (data not shown). These results suggest that AMG8562 efficacy in rat or mouse pain models is not confounded by locomotor impairment or sedation.
Discussion
TRPV1 antagonists are being considered as next generation pain therapeutics because of their antihyperalgesic effects in rodent models of inflammation and cancer (for reviews, see Immke and Gavva, 2006; Szallasi et al., 2007). In the present study, we have characterized four compounds that exhibited differential modulation of TRPV1 activation by capsaicin, pH 5, and heat. Similar to previous findings (Gavva et al., 2007b), profile A (AMG8163) and profile B (AMG8563) modulators of TRPV1 caused hyperthermia in rats. Here, we have identified two additional groups of modulators: profile C (AMG8562) that blocks capsaicin, but not heat activation, and potentiates pH 5 activation of TRPV1; and profile D (AMG7905) that blocks capsaicin activation and potentiates both heat and pH 5 activation of TRPV1. Interestingly, AMG8562 did not cause hyperthermia, whereas AMG7905 caused marked hypothermia. In addition, other compounds that exhibit the same profile as AMG7905 also caused marked hypothermia (data not shown). Relative to the AMG8562 effects, the marked hypothermia of 3°C elicited by AMG7905 was considered undesirable; hence, it was decided that profile D is not suitable for evaluation of its antihyperalgesic effects in pain models. Because our goal was to evaluate the feasibility of separating the hyperthermic effect of TRPV1 modulation from the antihyperalgesic effects, we have further characterized AMG8562 in rodent models of pain and show that AMG8562 acts as an antihyperalgesic and does not cause hyperthermia in rats. It is important to note here that AMG8562 behaved as a profile B antagonist at human TRPV1, a profile that was reported to cause hyperthermia (Gavva et al., 2007b). Hence, it is expected that AMG8562 would cause hyperthermia in humans. The feasibility of profile C modulation of both rat and human TRPV1 by a single small molecule is yet to be demonstrated.
Previously, we demonstrated that competitive antagonists of capsaicin that block all modes of TRPV1 activation block pH 5 and heat activation allosterically (Gavva et al., 2005a). This suggested that the allosteric inhibition of pH 5 and heat activation depends on the ability of the compound to lock the channel conformation in the closed or nonconducting state (Gavva et al., 2005a). In fact, compounds such as SB-366791 seemed to promote conformations that result in potentiation of pH 5 activation (Gavva et al., 2005a). All four compounds described here interact with the capsaicin binding pocket on TRPV1 and are potent antagonists of capsaicin activation. However, based on their ability to allosterically potentiate pH 5 and/or heat activation, we have defined two new profiles (profiles C and D). For example, AMG8562 (profile C compound that blocks capsaicin activation and potentiates pH 5 activation) may be considered as a competitive antagonist at the capsaicin binding pocket and as a positive allosteric modulator for proton activation. Similarly, AMG7905 (profile D compound that blocks capsaicin activation and potentiates both pH 5 and heat activation) may be considered as a competitive antagonist at the capsaicin binding pocket and as a positive allosteric modulator for both proton and heat activation.
Since AMG8562 is a selective modulator of rodent TRPV1, did not cause hyperthermia, and is orally bioavailable, we evaluated it by oral administration in an on-target challenge model (capsaicin-induced flinch), CFA- and surgical incision-induced thermal hyperalgesia in rats, and acetic acid-induced writhing in mice. As expected for potent capsaicin antagonists (Gavva et al., 2007a,b), AMG8562 blocked capsaicin-induced flinching potently. Similar to previous reports of TRPV1 antagonists (Gavva et al., 2005b, 2007a; Honore et al., 2005; Magal et al., 2005), AMG8562 partially reversed thermal hyperalgesia in inflammatory (CFA) and acute (surgical incision) pain models as well as acetic acid-induced writhing. Overall, these studies with AMG8562 suggest that hyperthermic effects can be separated from antihyperalgesic effects of TRPV1 modulation. Mechanisms of the AMG8562 antihyperalgesic effect in CFA and surgical incision models and the blockade of acetic acid-induced writhing by AMG8562 may differ. For example, AMG8562 blocks capsaicin, anandamide, and NADA activation, and by inference it also blocks the other putative endogenous ligands, such as 12-hydroperoxyeicosatetraenoic acid and oleolyldopamine. It is possible that the increased concentrations of endogenous ligands and phospholipase C-mediated activation of TRPV1 may predominantly contribute to inflammation or injury-related hyperalgesia. Therefore, the antihyperalgesic effect of AMG8562 in inflammatory and injury-induced pain models may represent blockade of the endogenous activation mechanisms of TRPV1 in vivo. The efficacy in the acetic acid-induced writhing in mice could be mediated by AMG8562-induced desensitization of the visceral TRPV1 channels, since 1) AMG8562 potentiates pH 5 activation of both rat and mouse TRPV1, 2) the mechanisms of acetic acid-induced writhing may include activation of visceral TRPV1 channels, and 3) visceral TRPV1 channels are tonically activated in vivo (Steiner et al., 2007). Identification of more profile C compounds, evaluating them in different pain models, and detailed mechanistic studies in vivo should reveal more accurate mechanism(s) of action for this new class of compounds.
TRPV1-selective antagonist reversal of inflammation-induced hyperalgesia ranged from 35 to 50% in our experimental conditions (Magal et al., 2005; Gavva et al., 2007a; Wang et al., 2007; this study), which were conducted in a blinded manner that involved randomization of dose groups and investigators that measured hyperalgesia not participating in dosing of compounds. Since these conditions eliminate subjectivity to the greatest degree possible and give the most stringent results, we believe that TRPV1 partially contributes to inflammatory pain.
The minimally effective dose of AMG8163 to block capsaicin-induced flinch was 0.03 mg/kg, which corresponds to plasma concentration of 2.6 ng/ml (Gavva et al., 2007a), and the minimally effective dose of AMG8562 to block capsaicin-induced flinch was 1 mg/kg, which corresponds to plasma concentration of 120 ng/ml (this study; Fig. 5). There is a 46-fold higher plasma concentration requirement of AMG8562 for comparable efficacy. However, AMG8163 is only approximately 2-fold more potent at TRPV1, and protein binding of AMG8562 and AMG8163 are comparable. One possible reason for the efficacy difference may be that AMG8562 blocks only putative endogenous ligands and AMG8163 blocks all modes of TRPV1 activation; therefore, a higher dose of AMG8562 may be required to show comparable efficacy. Alternately, differential distribution of AMG8163 and AMG8562 into organs and tissues, such as central nervous system, may contribute to different dose requirements for each compound. The minimally effective dose of AMG8163 to cause hyperthermia was 0.1 mg/kg (Tamayo et al., 2008). It is impressive that AMG8562 did not cause hyperthermia up to 100-mg/kg dose that is 1000-fold higher compared with AMG8163. An important point to note here is that AMG8562 blocked TRPV4 (IC50 value of ∼3 μM) in addition to modulating TRPV1, and it is not known whether TRPV4 antagonism contributed to lack of hyperthermia or to antihyperalgesia. Detailed characterization of additional profile C compounds should provide better understanding of efficacy (antihyperalgesia) and hyperthermia relationship (or lack thereof) for profile C compounds.
The unique profile of AMG8562 indicates that it is feasible to eliminate hyperthermia while preserving antihyperalgesia by differential modulation of distinct modes of TRPV1 activation. Although all four molecules studied here were identified based on their ability to block capsaicin activation of TRPV1, it is their differential pharmacology at pH 5 and heat activation that helped us to discover that profile C modulation of TRPV1 does not cause hyperthermia (summarized in Table 3). Potentiation of pH 5 activation alone seems to negate the classic TRPV1 blockade (profile A or B)-mediated hyperthermia. Furthermore, potentiation of both pH 5 and heat activation of TRPV1 causing marked hypothermia in vivo suggests profile D modulators act as functional agonists with respect to their effects on body temperature.
The TRPV1 blockade elicited hyperthermia is a major hurdle for the development of first generation TRPV1 modulators (TRPV1-selective profile A antagonists) as therapeutics because i) pharmacological blockade of TRPV1 elicited marked hyperthermia in healthy humans (Gavva et al., 2008), 2) TRPV1 antagonist-elicited hyperthermia did not show clear attenuation in humans (Gavva et al., 2008), and 3) some individuals are more susceptible to marked and persistent hyperthermia after TRPV1 blockade (Gavva et al., 2008). The potential of developing the second generation modulators such as profile C compounds as therapeutics has yet to be evaluated. Before considering profile C modulators for clinical development, further studies are required to understand 1) whether the modest efficacy and high exposure requirement of AMG8562 are common to profile C compounds; 2) whether it is possible to discover more efficacious profile C compounds, and whether they cause hyperthermia; 3) whether profile C compounds cause toxicities by damaging the viscera or gastric mucosa; and 4) whether it is possible to invent a compound that exhibits profile C modulation of both rodent and human TRPV1.
Acknowledgments
We thank Yunxin Bo, Lillian Liao, Markian Stec, Chenera Balan, Partha Chakrabarti, and Phi Tang for synthesis of AMG8163, AMG8562, AMG8563, and AMG7905.
Footnotes
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
doi:10.1124/jpet.107.132233.
-
ABBREVIATIONS: TRPV1, transient receptor potential vanilloid type 1; AMG 517, N-(4-[6-(4-trifluoromethyl-phenyl)-pyrimidin-4-yloxy]-benzothiazol-2-yl)-acetamide; CFA, complete Freund's adjuvant; AMG0347, (E)-N-(7-hydroxy-5,6,7,8-tetrahydronaphthalen-1-yl)-3-(2-(piperidin-1-yl)-6-(trifluoromethyl)pyridin-3-yl)acrylamide; TRPV2, transient receptor potential vanilloid type 2; 2-APB, 2-aminoethoxydiphenyl borate; CHO, Chinese hamster ovary; TRPA1, transient receptor potential ankyrin 1; TRPM8, transient receptor potential melastatin 8; TRPV3, transient receptor potential vanilloid type 3; TRPV4, transient receptor potential vanilloid type 4; TRP, transient receptor potential; 4α-PDD, 4α-phorbol-12,13-didecanoate; AMG8562, (R,E)-N-(2-hydroxy-2,3-dihydro-1H-inden-4-yl)-3-(2-(piperidin-1-yl)-4-(trifluoromethyl)phenyl)acrylamide; AMG8163, tert-butyl-2-(6-([2-(acetylamino)-1,3-benzothiazol-4-yl]oxy)pyrimidin-4-yl)-5-(trifluoromethyl)phenylcarbamate; ANOVA, analysis of variance; AMG8563, (S,E)-N-(2-hydroxy-2,3-dihydro-1H-inden4-yl)-3-(2-(piperidin-1-yl)-4-(trifluoromethyl)phenyl)acrylamide; M8-B, N-(2-aminoethyl)-N-((3-(methyloxy)-4-((phenylmethyl)oxy)phenyl)methyl)-2-thiophenecarboxamide; V3-H, 1-(((5-chloro-1,3-benzothiazol-2-yl)thio)acetyl)-8-methyl-1,2,3,4-tetrahydroquinoline; AMG7905, N-(6-(2-(cyclohexylmethylamino)-4-(trifluoromethyl)phenyl)pyrimidin-4-yl)benzo[d]thiazol-6-amine compound; A-425619, 1-isoquinolin-5-yl-3-(4-trifluoromethyl-benzyl)-urea; SB-366791, 4′-chloro-3-methoxycinnamanilide; 5-HT, 5-hydroxytryptamine; CGS 21680, 2-[p-(2-carboxyethyl)phenethylamino]-5′-N-ethylcarboxamidoadenosine; RX 821002, 2-(2-methoxy-2,3-dihydrobenzo(1,4)dioxin-2-yl)-4,5-dihydro-1H-imidazole; CGP 12177, 4-[3-[(1,1-dimethylethyl)amino]-2-hydroxypropoxy]-1,3-dihydro-2H-benzimidazol-2-one; CP 55940, 5-(1,1-dimethylheptyl)-2-(5-hydroxy-2-(3-hydroxypropyl)cyclohexyl)phenol; WIN 55212-2, (R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl) pyrrolo-[1,2,3-d,e]-1,4-benzoxazin-6-yl]-1-naphthalenyl-methanone; SCH 23390, R-(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine; U 69593, (+)-(5α,7α,8β)-N-methyl-N-[7-(1-pyrrolidinyl)-1-oxaspiro[4.5]dec-8-yl]benzeneacetamide; NADA, N-arachidonyldopamine.
-
↵1 Current affiliation: Abraxis Bioscience Inc., Marina Del Rey, California.
-
↵2 Current affiliation: St. Jude Medical, Camarillo, California.
-
↵3 Current affiliation: University of Texas Medical Branch, Galveston, Texas.
-
↵4 Current affiliation: Pfizer Inc., Kalamazoo, Michigan.
-
↵5 Current affiliation: Allergan Inc., Irvine, California.
-
↵6 Current affiliation: Celgene, Summit, New Jersey.
- Received September 28, 2007.
- Accepted April 16, 2008.
- The American Society for Pharmacology and Experimental Therapeutics