Inverse Agonism and Neutral Antagonism at Wild-Type and Constitutively Active Mutant Delta Opioid Receptors

Home , Intrinsic activity, Inverse agonist

0022-3565/05/3131-410–421 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 313, No. 1 U.S. Government work not protected by U.S. copyright 77321/1195209 JPET 313:410–421, 2005 Printed in U.S.A.

P. Tryoen-To´th, F. M. Dećaillot, D. Filliol, K. Befort, L. H. Lazarus, P. W. Schiller, H. Schmidhammer, and B. L. Kieffer Institut de Geńe´ tique et de Biologie Molećulaire et Cellulaire, Centre National de la Recherche Scientifique/Institut National de la Sante´ et de la Recherche Me´dicale/Universite´ Louis Pasteur, Illkirch, France (P.-T.T., D.F., K.B., B.L.K.); Department of Pharmacology and Biological Chemistry, Mount Sinai School of Medicine, New York, New York (F.M.D.); Medicinal Chemistry Group, Laboratory of Computational Biology and Risk Analysis, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina (L.H.L.); Laboratory of Chemical Biology and Peptide Research, Clinical Research Institute of Montreal, Montreal, Quebec, Canada (P.W.S.); and Department of Pharmaceutical Chemistry, Institute of Pharmacy, University of Innsbruck, Innsbruck, Austria (H.S.) Downloaded from Received September 9, 2004; accepted December 3, 2004

ABSTRACT The delta opioid receptor modulates nociceptive and emotional methoxybenzyl]-N,N-diethylbenzamide], and inverse agonist, ICI jpet.aspetjournals.org behaviors. This receptor has been shown to exhibit measurable 174864 (N,N-diallyl-Tyr-Aib-Aib-Phe-Leu), were also included. In a spontaneous activity. Progress in understanding the biological screen using wild-type and CAM M262T delta receptors, naltrin- relevance of this activity has been slow, partly due to limited dole (NTI) and close derivatives were mostly inactive, and TIPP characterization of compounds with intrinsic negative activity. behaved as an agonist, whereas Dmt-Tic-OH and N,N(CH3)2- Here, we have used constitutively active mutant (CAM) delta Dmt-Tic-NH2 showed inverse agonism. The two latter compounds receptors in two different functional assays, guanosine 5Ј-O- showed negative activity across 27 CAM receptors, suggesting (3-thio)triphosphate binding and a reporter gene assay, to test that this activity was independent from the activation mechanism. potential inverse agonism of 15 delta opioid compounds, orig- These two compounds also exhibited nanomolar potencies in at ASPET Journals on January 24, 2020 inally described as antagonists. These include the classical dose-response experiments performed on wild-type, M262T, antagonists naloxone, naltrindole, 7-benzylidene-naltrexone, Y308H, and C328R CAM receptors. TICP(⌿) exhibited strong and naltriben, a new set of naltrindole derivatives, H-Tyr-Tic-Phe- inverse agonism at the Y308H receptor. We conclude that the ⌿ ⌿ Phe-OH (TIPP) and H-Tyr-Tic [CH2N]Cha-Phe-OH [TICP( )], as stable N,N(CH3)2-Dmt-Tic-NH2 compound represents a useful well as three 2Ј,6Ј-dimethyltyrosine-1,2,3,4-tetrahydroquinoline- tool to explore the spontaneous activity of delta receptors, and 3-carboxylate (Dmt-Tic) peptides. A reference agonist, SNC 80 NTI and novel derivatives behave as neutral antagonists. [(ϩ)-4-[(␣R)-␣-((2S,5R)-4-Allyl-2,5-dimethyl-1-piperazinyl)-3-

Evidence that a number of G protein coupled receptors past years, and the existence of high basal activity has been pro- (GPCRs) show spontaneous activity has accumulated in the posed for more than 60 GPCRs (Seifert and Wenzel-Seifert, 2002). Key tools to investigate receptor basal signaling are ligands that block this activity, the so-called inverse agonists (Milligan, 2003). This work was supported by the Human Frontier Science Program, by the In contrast to neutral antagonists, which compete with agonists Institut National de la Sante´et de la Recherche Me´dicale, by the Universite´ Louis Pasteur, by the Association de la Recherche pour le Cancer, by the but are devoid of biological activity per se, inverse agonists are Institut UPSA de la Douleur, by the Mission Interministe´rielle de Lutte contre endowed with intrinsic negative activity (Strange, 2002). Differen- la Drogue et la Toxicomanie, and by National Institute on Drug Abuse (Na- tional Institutes of Health-National Institute on Drug Abuse Grant DA05010). tial activity of inverse agonists and neutral antagonists was re- Article, publication date, and citation information can be found at ported in recombinant systems and was also evidenced in physio- http://jpet.aspetjournals.org. doi:10.1124/jpet.104.077321. logical situations (de Ligt et al., 2000).

ABBREVIATIONS: GPCR, G protein-coupled receptor; ICI 174864, N,N-diallyl-Tyr-Aib-Aib-Phe-Leu; CAM, constitutively active mutant; NLX, naloxone; NTI, naltrindole; NTB, naltriben; BNTX, 7-benzylidene-naltrexone; HS-378, 17-cyclopropylmethyl-4,5␣-epoxy-14␤-methylindolo[2Ј,3Ј:6,7]morhinan-3-ol hydrochloride; HS-414, 17-allyl-4,5␣-epoxy-14␤-ethoxyindolo[2Ј,3Ј:6,7]morphinan-3-ol methanesulfonate; HS-464, 17-cyclopropylmethyl-4,5␣-epoxy- 14␤-ethoxyindolo[2Ј,3Ј:6,7]morphinan-3-ol hydrochloride; HS-510A, 17-cycloproplymethyl-4,5␣-epoxy-14␤-ethoxybenzofuro[2Ј,3Ј:6,7]morphinan-3-ol salicylate; HS-531, 17-allyl-4,5␣-epoxy-14␤-methoxy-1Ј-methylindolo[2Ј,3Ј:6,7]morphinan-3-ol hydrochloride; HS-595, 17-cyclopropylmethyl-4-me- Ј Ј ␤ ⌿ ⌿ Ј Ј thoxybenzofuro[2 ,3 :6,7]morphinan-3,14 -diol; TIPP, H-Tyr-Tic-Phe-Phe-OH; TICP( ), H-Tyr-Tic [CH2N]Cha-Phe-OH; Dmt-Tic, 2 ,6 -dimethyltyrosine-1,2,3,4-tetrahydroquinoline-3-carboxylate; SNC 80, [(ϩ)-4-[(␣R)-␣-((2S,5R)-4-Allyl-2,5-dimethyl-1-piperazinyl)-3-methoxybenzyl]-N,N-diethylbenzamide]; GTP␥S, guanosine 5Ј-O-(3-thio)triphosphate; WT, wild type; FCS, fetal calf serum; SEAP, secreted alkaline phosphatase; FK, forskolin. 410 Inverse Agonism at Constitutively Active Delta Receptors 411

The concept of functional inverse agonism was pioneered cluded the prototypic delta agonist SNC 80 [(ϩ)-4-[(␣R)-␣- with delta opioid receptors, which belong to the opioid recep- ((2S,5R)-4-Allyl-2,5-dimethyl-1-piperazinyl)-3-methoxyben- tor family comprising also ␮, ␬, and opioid receptor-like re- zyl]-N,N-diethylbenzamide], as well as the reference inverse ceptors (Kieffer, 1997). The finding that ICI 174864 (N,N- agonist ICI 174864 in all experiments. diallyl-Tyr-Aib-Aib-Phe-Leu-OH) (Cotton et al., 1984), This is the first study that compares the functional activity originally considered as delta antagonist, was able to de- of a wide range of delta compounds using a set of several crease basal GTPase activity in NG108–15 neuroblastoma CAM delta receptors. Our data show that the highly delta

cells, whereas other antagonists had no effect in this func- selective ligand N,N (CH3)2-Dmt-Tic-NH2 is most potent and tional assay, was the first demonstration of negative intrinsic efficacious at all the tested mutant receptors and seems to be activity at an endogenously expressed GPCR (Costa and a better inverse agonist than the reference ICI 174864 com- Herz, 1989). The existence of delta receptor spontaneous pound. TICP(⌿) also showed inverse agonism, mainly at one activity was further confirmed using different functional as- mutant receptor. NTI derivatives were otherwise inactive, says in cell lines expressing either the endogenous (Szekeres leaving NTI and closely related analogs as neutral antago- and Traynor, 1997) or the recombinant delta receptor ex- nists. pressed in HEK 293 (Chiu et al., 1996; Labarre et al., 2000; Zaki et al., 2001), CHO (Hosohata et al., 1999), rat-1 fibro- blast (Mullaney et al., 1996; Merkouris et al., 1997), GH3 Materials and Methods

(Liu and Prather, 2002), or C6 glioma cells (Neilan et al., Opioid Ligands. [3H]Diprenorphine (50 Ci/mmol) was obtained Downloaded from 1999). From all these studies, ICI 174864 appeared to show from PerkinElmer Life and Analytical Sciences (Boston, MA). Com- reliable and potent negative activity and was taken as the mercial delta opioid compounds were as follows: SNC 80, ICI 174864, reference molecule for delta inverse agonism. ICI 174864, and NLX (all from Sigma/RBI, Natick MA), NTB, NTI, and BNTX however, is a peptidic ligand that shows low potency and (Tocris Cookson Inc., Ellisville, MO), and TIPP, where Tic is 1,2,3,4 ␮ ␬ limited delta selectivity toward and opioid receptors (Ki tetrahydroisoquinoline (Phoenix Pharmaceuticals, Belmont, CA). The delta compound TICP[⌿] was prepared as described by Schiller ratios 1:155 and 1:344; Corbett et al., 1993). Therefore, the jpet.aspetjournals.org identification of novel stable agonists, with improved phar- et al. (1996). Dmt-Tic molecules were synthesized as reported by macological properties, is highly desirable to study the bio- Salvadori et al. (1995). The naltrindole derivatives were synthesized as described by Schutz et al. (2002). logical activity of delta receptors in physiological context Expression of Wild-Type and Mutant Receptors. For opioid (Filliol et al., 2000; Kieffer and Gave´riaux-Ruff, 2002). ligand and [35S]GTP␥S binding assays, the wild-type (WT) human Basal activity of GPCRs is not easily measurable, and the delta receptor, the CAM M262T and the CAM C328R human delta search for inverse agonists has been facilitated by the con- receptors (De´caillot et al., 2003) were stably expressed in HEK 293

struction of constitutively active mutant (CAM) receptors cells (American Type Culture Collection, Manassas, VA). Cells were at ASPET Journals on January 24, 2020 where potency and efficacy of inverse agonists are enhanced transfected at about 70% confluency with WT or mutated delta (Rossier et al., 1999). In a previous study, we have used a receptor-expressing vectors (pcDNA3, 8 ␮g/10-cm dish) using the random mutagenesis approach to identify human delta CAM Jet-PEI reagent (Polytransfection, Strasbourg, France) according to receptors (Decaillot et al., 2003). Here, we have investigated the manufacturer’s protocol. Forty-eight hours after transfection, the the in vitro pharmacological activity of a series of delta an- culture medium, Dulbecco’s modified Eagle’s medium (Invitrogen) supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, CA) tagonists from distinct structural families at these CAM re- and gentamycine (Invitrogen), was replaced by fresh medium con- ceptors. Using two different functional assays, we have ex- taining the selecting agent G418 (Invitrogen) at a 0.5 mg/ml concen- amined properties of a number of commercially available tration. Cells were grown at 37°C in the presence of 5% CO2. Resis- delta antagonists. tant colonies appeared after 2 weeks of G418 application. Cellular We have tested four alkaloid opioids that have been clas- clones were amplified and screened for receptor expression by [3H]di- sically defined as antagonist compounds and are widely used prenorphine binding on whole cells (0.75 and 1.5 nM, 2 ϫ 105 cells/ in pharmacological studies. These compounds are naloxone well). Clones with highest expression levels were amplified for fur- (NLX), considered the reference opioid antagonist, as well ther experiments. The Y308H mutant receptor was expressed as naltrindole (NTI), naltriben (NTB), and 7-benzylidene- transiently in HEK 293 cells. Transfection was performed in Dulbec- naltrexone (BNTX) that exhibit reasonable delta selectivity co’s modified Eagle’s medium with 10% FCS with antibiotics, and cells were harvested after 48 h. For the secreted alkaline phospha- (Corbett et al., 1993) and are structurally related. We also tase (SEAP) assay, all WT and CAM receptors were expressed tran- have included delta compounds that have been developed siently in HEK 293 cells as previously described (Befort et al., 2001) more recently in the search for compounds with better delta with a slightly modified procedure. Briefly, HEK 293 cells were selectivity (see Fig. 1). Among them was a new series of NTI plated in 96-well plates (Biocoat; Falcon, Cowley, UK) at a density of derivatives (Schmidhammer et al., 1998) including HS-378, 35,000 cells/well on the day before transfection. Cells were cotrans- HS-414, HS-464, HS-510A, HS-531, and HS-595. H-Tyr-Tic- fected at about 70% confluency with the reporter gene pCRE-SEAP Phe-Phe-OH (TIPP) (Schiller et al., 1992), which was the first (1 ␮g/well; BD Biosciences Clontech, Palo Alto, CA) and WT or peptide from a new class of highly delta selective antagonists, mutated delta receptor plasmid DNA (0.15 ␮g/well) using Jet-PEI ⌿ ⌿ reagent. Thirty-two hours after transfection, cells were serum- and H-Tyr-Tic [CH2N]Cha-Phe-OH [TICP( )], a more hy- drophobic and stable derivative (Schiller et al., 1999), were starved overnight and subjected to the SEAP test. Ligand Binding Assay. HEK 293 cells expressing WT and CAM also tested. Further we have studied three compounds from Ј Ј receptors stably or transiently were harvested for membrane prepara- the Dmt-Tic family, where 2 ,6 methylation of the N-termi- tion according to the described procedure (Befort et al., 2001). Mem- nal tyrosine residue (Dmt) is combined to the heteroaromatic branes were resuspended in 50 mM Tris HCl, pH 7.4, and 0.32 M residue Tic, leading to a class of pseudodipeptides with high sucrose buffer, aliquoted, and stored at Ϫ80°C until use. Protein con- delta selectivity and antagonist properties (Salvadori et al., centration was assayed by the Bradford method (Bio-Rad, Hercules, 1995; Bryant et al., 1998, 2003). Importantly, we have in- CA). For saturation experiments, 5 ␮g of membrane proteins were 412 Tryoen-To´thetal. Downloaded from jpet.aspetjournals.org at ASPET Journals on January 24, 2020

Fig. 1. Ligands and receptors. Structures of delta compounds studied as potential inverse agonists. These include the classical antagonists NLX, NTI, BNTX, and NTB, a set of NTI derivatives, TIPP and TICP[⌿] from the TIPP family (Schiller et al., 1999), as well as three compounds from the Dmt-Tic pseudopeptide family (Bryant et al., 2003). Inset, CAM delta receptors used for the inverse agonist screen (M262T), as well as for dose-response curves (M262T, Y308H, and C328R). Black dots, position of activating mutations on the receptor, displayed in a serpentine representation. The 24 other CAM receptors used to test inverse agonist activity of the most potent compounds have been detailed earlier (De´caillot et al., 2003). diluted in 50 mM Tris HCl, pH 7.4, in a final volume of 250 ␮l and Analytical Sciences). Microplates were dried, then wells were covered incubated with variable concentrations (0.02–3 nM) of [3H]diprenor- by 40 ␮l of scintillation cocktail (PerkinElmer Life and Analytical Sci- phine for1hat25°C. Nonspecific binding was determined in the ences) and counted in a Microplate Scintillation/Luminescence Counter presence of 10 ␮M naloxone. For competition studies, 5 ␮g of membrane (Canberra Industries, Meriden, CT) using the TopCount-NXT software. proteins was incubated with 0.7 (C328R), 1 (WT and M262T), or 7 Assays were performed in triplicates for saturation experiments and in 3 (Y308H) nM [ H] diprenorphine, in the presence of variable concentra- duplicates for competition studies. Kd and Ki values of opioid ligands tions (1 pM to 100 ␮M) of competing opioid ligand for 1h at 25°C. were calculated using the Prism software (GraphPad Software Inc., San Membranes were washed with cold 50 mM Tris-HCl, pH 7.4, buffer by Diego, CA). filtration through 0.1% polyethylenimine-presoaked microplate filters [35S]GTP␥S Binding Assay. Opioid ligand-stimulated (Unifilter GF/B) using a Filtermate Harvester (PerkinElmer Life and [35S]GTP␥S binding was performed using the same membranes as Inverse Agonism at Constitutively Active Delta Receptors 413

above. [35S]GTP␥S binding was carried out essentially as described lot et al., 2003). [3H]Diprenorphine affinity was unchanged at previously (Befort et al., 1999). Briefly, 10 ␮g of membrane proteins this mutant receptor (Table 1). To examine whether the was incubated for1hat25°C without (basal) or with opioid ligands M262T mutation alters binding affinities of the delta ligands (3 pM to 10 ␮M) in the assay buffer containing 50 mM Hepes, pH 7.6, under study, we determined Ki values for each compound 5 mM MgCl , 100 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 0.1% 2 (Table 1). Affinity changes never exceeded 4-fold, and with bovine serum albumin, 30 ␮M GDP, and 0.2 nM [35S]GTP␥S. For the exception of ICI 174864, all compounds exhibited similar background determination, 10 ␮M nonradioactive GTP␥S was used. Incubation mixture was filtered as above (ligand binding section) nanomolar affinity for WT and M262T mutant receptors. except that microplate filters were presoaked in water. Plates were Functional activity of all the compounds was, therefore, dried, and radioactivity was counted as above. Assays were per- tested. 35 ␥ formed in triplicates. EC50 values of the various opioid ligands were We first used the classical [ S]GTP S binding assay that determined with the Prism software using nonlinear regression measures receptor-mediated G protein activation occurring and one-site competition. For each experiment, ligand-stimulated at an early stage of the signaling cascade. In this assay, [35S]GTP␥S binding values were normalized to the basal value [35S]GTP␥S binding is increased in the presence of agonist (absence of ligand) defined as 100%. and decreased in the presence of inverse agonist. We treated SEAP Assay. Cells were cotransfected using receptor and pCRE- membrane preparations from the two stable cell lines (WT SEAP expression vectors as described above. Thirty-two hours after and M262T) with each delta opioid ligand, and results are transfection, cells were serum-starved (0.1%, instead of 10% FCS) shown in Fig. 2 (top panel). The agonist SNC 80 increased overnight. Forty-eight hours after transfection, cells were stimulated 35 ␥ Downloaded from with forskolin (FK) alone (10 ␮M) or with forskolin and the opioid basal [ S]GTP S binding about 2-fold at both WT and mu- ligand to be tested. Treatments with all ligands were performed in tant receptors. As expected, ICI 174864, the reference delta the presence of 0.1% FCS for 5 to 6 h. Aliquots of culture medium inverse agonist, decreased the basal signal. In this set of supernatants (15 ␮l) were transferred into black 96-well plate, di- experiments, the negative activity of ICI 174864 was low at luted into 150 mM NaCl and 40 mM Tris-HCl, pH 7.2, buffer and WT and significant at the M262T mutant receptor, consistent heat-inactivated at 65°C for 20 min. The enzymatic reaction was with the expected higher efficacy of inverse agonists at CAM performed in the presence of assay buffer (666 mM diethanolamine,

receptors (De´caillot et al., 2003). No significant modification jpet.aspetjournals.org 3.3 mM MgCl , and 6.6 mM L-homoarginine, pH 9.8), enhancer 35 2 of basal [ S]GTP␥S binding could be detected for NLX, NTI, solution (Tropics, Bedford, MA), and 0.4 mM SEAP chemilumines- BNTX, NTB, and TIPP at WT and M262T receptors. Naltrin- cent substrate (Tropics) in a final volume of 150 ␮l. Luminescence was quantified 15 min after substrate addition (TopCount-NXT, dole derivatives (HS compounds) showed no inverse agonist Canberra Industries). To compare data from different experiments activity at any receptor and even showed a trend toward for each compound and to allow comparison across mutant receptors, agonism at the WT receptor. Finally the compounds ⌿ the SEAP signals obtained in the presence of FK and opioid com- TICP( ), Dmt-Tic-OH, and N,N(CH3)2-Dmt-Tic-NH2 showed

pounds were normalized relative to signal obtained in the presence of significant inverse agonism at preparations expressing the at ASPET Journals on January 24, 2020 FK alone, defined as 100%. A prototypic agonist (SNC 80) and the M262T receptor. The three compounds also showed detect- reference inverse agonist (ICI 174864) were included in each exper- able, although not always significant, negative activity at the iment. Dose response experiments were performed with agonist or WT receptor. Altogether, therefore, the three latter com- ␮ inverse agonist concentrations ranging from (3 pM to 100 M). pounds appeared as the most efficacious inverse agonists in Assays were performed in triplicate, and EC values were calcu- 50 this assay. lated using the Prism software as above. Statistics. In opioid ligand screening experiments (Fig. 2), one- To further examine bioactivity of the delta opioid com- way analysis of variance with protected least significant difference pounds, we used the more sensitive high-throughput SEAP Fischer posteriori test was used to evaluate significant effects of the reporter gene assay, where the endpoint measurement is compounds (p Ͻ 0.05). performed downstream from G protein activation along the cAMP pathway. In this assay, both a cAMP-responsive Results reporter gene and the receptor are transiently coexpressed in HEK 293 cells. FK stimulation directly activates the Screening for Inverse Agonism at WT and M262T reporter gene, and the amount of gene product (SEAP) is Mutant Delta Receptors. In our previous search for CAM measured using a chemiluminescent substrate. Typically, delta receptors, we have identified 30 point-mutated recep- an opioid agonist will decrease the FK-stimulated reporter tors with high spontaneous activity (De´caillot et al., 2003). gene activity (SEAP activity), whereas an inverse agonist The mutant M262T exhibited highest responsivity to ICI will enhance this activity (De´caillot et al., 2003). Data are 174864 in the reporter gene assay and was, therefore, chosen shown in Fig. 2 (bottom panel). The delta opioid agonist, as a screening tool in this study. Also, because the mutation SNC 80, decreased the SEAP signal at levels correspond- is located on the cytoplasmic face of the sixth transmembrane ing to 50% (WT) and 70% (M262T) of levels observed in the domain and is, therefore, distant from the binding site, we absence of agonist. ICI 174864 increased the SEAP signal reasoned that ligand binding should be preserved for most at cells expressing both WT (120% above basal) and M262T compounds under study. First, we established stable cell (130% above basal) receptors. As in the [35S]GTP␥S bind- lines expressing the WT or M262T receptors. Receptor ex- ing assay, this activity better detected at the mutant re-

pression levels were determined by Scatchard analysis of ceptor. NLX, BNTX, Dmt-Tic-OH, and N,N(CH3)2-Dmt- 3 Ϯ [ H]diprenorphine binding. Bmax values were 4.38 0.91 Tic-NH2 exhibited significant inverse agonist activity at ϭ Ϯ pmol/mg for the WT receptor (n 2) and 1.57 0.35 pmol/mg the M262T receptor, whereas only N,N(CH3)2-Dmt-Tic- ϭ 35 ␥ for the M262T receptor (n 2). Basal [ S]GTP S binding at NH2 markedly increased SEAP activity at the WT recep- the mutant receptor was consistently higher than WT tor. TICP(⌿), which showed inverse agonism in the (142.1 Ϯ 7.8% of WT, n ϭ 3) despite a 2.7-fold lower expres- [35S]GTP␥S assay, was inactive in the SEAP assay. Con- sion level, confirming that M262T is a CAM receptor (De´cail- versely, TIPP, which was inactive in [35S]GTP␥S assay, 414 Tryoen-To´thetal.

Fig. 2. Screening for inverse agonist activity. Fifteen delta compounds were tested on WT (left) and M262T mutant (right) recep- Downloaded from tors. Delta opioids include, from left to right: the reference agonist SNC 80 (left black bars), the reference inverse agonist ICI 174864 (right black bars), commercial alkaloids (NLX, NTI, BNTX, and NTB) and peptide (TIPP), the TIPP analog TICP(⌿), three Dmt-Tic compounds, and six novel NTI derivatives (gray bars). Top panel, li- jpet.aspetjournals.org gand-induced [35S]GTP␥S binding. Ligand concentrations were 10 ␮M (except SNC 80 at 1 ␮M). [35S]GTP␥S binding activities were normalized to basal activity (without ligand) defined as 100% (horizontal line). Data are means Ϯ S.E.M. from three to five experiments performed in triplicate and using at least two independent membrane preparations. Bottom panel, ligand-induced at ASPET Journals on January 24, 2020 SEAP activation. Ligand concentrations were at 1 ␮M (except ICI 174864 at 10 ␮M) in the presence of 10 ␮M FK. SEAP activities were normalized to signal obtained with FK alone, defined as 100% and indicated by a horizontal line. Data are means Ϯ S.E.M. from three to nine independent transfection experiments performed in triplicate. Activities of the tested compounds were compared with basal values using the Student’s t test. One star, p Ͻ 0.05; two stars, p Ͻ 0.01; three stars, p Ͻ 0.001 (one-way analysis of variance).

exhibited strong agonism in the SEAP test at both recep- agonist effect for HS-595. NTI and NTB were inactive at tors. Finally, no significant effect of HS compounds could both receptors as well. be observed at WT and M262T receptors, except for a slight Taken together, screening data from both functional as- Inverse Agonism at Constitutively Active Delta Receptors 415

TABLE 1 Binding affinities of delta compounds to WT and mutant delta receptors Competition experiments were carried out on HEK 293 cell membrane preparations expressing receptors either stably (WT, M262T, C328R) or transiently (Y308H) and using ͓3 ͔ Ϯ variable concentrations of unlabeled ligands to displace H diprenorphine. Data are means S.E.M. from n experiments performed in triplicate (Kd) or in duplicate (Ki) using at least two different membrane preparations.

WT M262T Y308H C328R Compounds Ki nKi nKi nKi n nM nM nM nM ͓3 ͔ Ϯ Ϯ Ϯ Ϯ H Diprenorphine (Kd) 0.39 0.06 2 0.72 0.21 2 8.00 1.22 5 0.44 0.06 3 SNC 80 9.41 Ϯ 3.07 3 9.62 Ϯ 2.59 3 438 Ϯ 21 5 5.73 Ϯ 1.84 4 ICI 174864 434 Ϯ 61 4 1308 Ϯ 247 3 Ͼ4800 2 567 Ϯ 53 6 NLX 66.20 Ϯ 0.71 3 59.11 Ϯ 2.92 2 NTI 0.30 Ϯ 0.15 3 0.12 Ϯ 0.02 3 BNTX 3.22 Ϯ 0.58 3 6.10 Ϯ 1.48 3 NTB 0.10 Ϯ 0.02 3 0.23 Ϯ 0.04 3 TIPP 2.90 Ϯ 0.66 3 7.03 Ϯ 3.05 3 TICP͓␺͔ 4.62 Ϯ 0.64 3 3.91 Ϯ 0.27 3 21.82 Ϯ 1.53 4 7.20 Ϯ 1.14 6 H-Dmt-Tic-OH 13.13 Ϯ 1.71 3 5.20 Ϯ 1.85 3 200 Ϯ 7 4 13.74 Ϯ 4.12 7 Ϯ Ϯ H-Dmt-Tic-NH2 9.41 1.13 3 9.43 1.16 3 Ϯ Ϯ Ϯ Ϯ N,N(CH3)2-Dmt-Tic-NH2 2.50 0.76 3 1.81 0.66 3 18.23 3.48 4 1.90 0.41 4

HS-378 0.40 Ϯ 0.01 2 0.90 Ϯ 0.03 3 Downloaded from HS-414 2.10 Ϯ 0.72 3 4.13 Ϯ 1.04 3 HS-464 0.11 Ϯ 0.03 3 0.40 Ϯ 0.09 3 HS-510A 0.12 Ϯ 0.03 3 0.41 Ϯ 0.15 3 HS-531 0.42 Ϯ 0.08 2 0.80 Ϯ 0.09 2 HS-595 0.22 Ϯ 0.01 2 0.41 Ϯ 0.05 2

says and from both WT and M262T CAM receptors led to effect of ICI 174864 at most CAM receptors, suggesting jpet.aspetjournals.org the following conclusions: TIPP and HS-595 show either no that N,N(CH3)2-Dmt-Tic-NH2 is a more efficacious inverse activity or agonist activity; NTI, NTB, and the other HS agonist. As for this compound, the closely related com- derivatives show no significant activity in any test, despite pound Dmt-Tic-OH behaved as a delta inverse agonist at their high affinity for the receptors; NLX and BNTX show most CAM receptors (Fig. 3, middle panel). For Dmt-Tic- inverse agonism in one test and for one receptor only; and OH, highest levels of activation were obtained at mutant as for ICI 174864, Dmt-Tic-OH, N,N(CH3)2-Dmt-Tic-NH2, receptors T134A, T213S, V283A, and C328R. When data ⌿ and TICP( ) behave as inverse agonists. The latter com- from this ligand were directly compared with those of at ASPET Journals on January 24, 2020 pounds were, therefore, studied further. N,N(CH3)2-Dmt-Tic-NH2, the overall efficacy appeared Inverse Agonist Activity of TICP(⌿), Dmt-Tic-OH, lower (data not shown). Also, the inverse agonist activity of and N,N(CH3)2-Dmt-Tic-NH2 across a Series of CAM Dmt-Tic-OH was lower than the reference compound ICI Receptors. We examined the negative functional activity of 174864 for mutants D21G, N169S, W274R, and E323K, ⌿ ICI 174864, TICP( ), Dmt-Tic-OH, and N,N(CH3)2-Dmt-Tic- which was not the case for N,N(CH3)2-Dmt-Tic-NH2. Dmt- NH2. First, a control experiment was performed where empty Tic-OH, therefore, seemed to be a less efficient inverse ⌿ HEK 293 cells (transfected with the empty pcDNA3 plasmid) agonist than N,N(CH3)2-Dmt-Tic-NH2. TICP( ) was inac- were subjected to these compounds (10Ϫ5 M) in both tive at most mutants (data not shown). The compound [35S]GTP␥S binding and SEAP assays. All compounds were showed detectable activity at mutant receptors M262T and inactive (data not shown), indicating that their activity is C328R. Interestingly, the inverse agonist effect of TICP(⌿) mediated by delta receptors. was remarkably active at the Y308H receptor, reaching We investigated whether ICI 174864, TICP(⌿), Dmt-Tic- 220% of the FK-stimulated signal (Fig. 3, bottom panel).

OH, and N,N(CH3)2-Dmt-Tic-NH2, which show inverse ag- This was the highest value observed across all mutants onist activity at WT and M262T mutant receptors, would and for all ligands. also inhibit the high spontaneous activity of other CAM Together, these data show that ICI 174864, TICP(⌿), Dmt- delta receptors independently from the position of the mu- Tic-OH, and N,N(CH3)2-Dmt-Tic-NH2 all exhibit inverse ago- tation. To test the activity of the compounds at a large set nism at several CAM delta receptors. This observation con- of mutant receptors, we used the SEAP assay that does not firms that these compounds are endowed with robust require membrane preparations. We expressed 27 CAM negative intrinsic activity, which favors the inactive receptor delta receptors from our previous study (Dećaillot et al., conformation, even though distinct activation mechanisms 2003) and measured the modulation of SEAP responses are likely to be involved in the different mutants (Dećaillot et ⌿ after TICP( ), Dmt-Tic-OH, and N,N(CH3)2-Dmt-Tic-NH2 al., 2003). The four compounds show variable efficacies de- stimulation. Each compound was used at a saturating pending on the CAM receptors. At this stage, comparing the concentration (1 ␮M) and systematically compared with compounds is difficult because these differences could be due ICI 174864 (10 ␮M). Maximal effects are shown in Fig. 3. either to distinct intrinsic functional activities or because As in our first experiment (Fig. 2), the most active inverse mutations alter binding affinities of the compounds to vari- agonist was N,N(CH3)2-Dmt-Tic-NH2. This compound en- able extends. Therefore, we focused our attention on the hanced the FK-induced SEAP signal at all other CAMs three CAM receptors M262T, Y308H, and C328R, which ex- (Fig. 3, top panel). For the stronger effects, the rank of hibited the highest basal activities (see our previous work, order was M262T Ͼ T213S ϭ T134A ϭ Y308H Ͼ C328R ϭ Dećaillot et al., 2003), and where inverse agonists showed E323K ϭ V283A. This effect was more important than the best activities in this study (Fig. 3). 416 Tryoen-To´thetal. Downloaded from

Fig. 3. Inverse agonist activity of delta compounds across

CAM delta receptors. N,N(CH3)2-Dmt-Tic-NH2 (top), Dmt- Tic-OH (middle), and TICP(⌿) (bottom) were tested at a 1 ␮M concentration and compared with ICI 174864 (10 ␮M) in each experiment. Activating mutations are distributed all along the receptor sequence, and mutant receptors are jpet.aspetjournals.org presented from N-terminal (left) to C-terminal (right) mutations. Structural and functional properties of these mutants have been described earlier (De´caillot et al., 2003). Ligand activities were quantified by the SEAP reporter gene assay. SEAP activities were normalized to the FK signal alone, defined as 100% and represented by an horizontal line. Histograms of three to four independent experiments with triplicate transfections are shown. at ASPET Journals on January 24, 2020

Dose Responses to Dmt-Tic Compounds at WT, residue is centrally located within the ligand binding pocket M262T, Y308H, and C328R Mutant Receptors. To fully and known to influence opioid affinity (Befort et al., 1996, 1999). characterize agonism and inverse agonism of the most active Affinities to Y308H, however, remained measurable, and dose delta compounds, we first determined their binding affinities responses were also performed for this mutant. at the different receptors (Table 1). We have shown above Further, we determined potencies and efficacies using the that opioid binding at M262T delta receptor is comparable two functional assays at the selected receptors. For [35S]GTP␥S Ϯ with WT receptor. Similarly, Ki values were unchanged at binding assays, we used stable expression for WT (Bmax, 4.38 ϭ Ϯ the C328R mutant, consistent with the notion that M262 and 0.91 pmol/mg protein; n 2), M262T (Bmax, 1.57 0.35 ϭ C328 residues are located on the cytoplasmic face of the recep- pmol/mg protein; n 2), C328R mutants (Bmax, 5 pmol/mg tor and unlikely involved in ligand binding (De´caillot et al., protein) and transient expression for Y308H receptor (Bmax, 2003). In contrast, binding affinities were substantially de- 7.99 Ϯ 1.26 pmol/mg protein; n ϭ 2). For the SEAP assay, creased at the Y308H receptor. This was expected because this receptors were expressed transiently, as before. Inverse Agonism at Constitutively Active Delta Receptors 417

The agonist SNC 80 increased [35S]GTP␥S binding and 48.6 Ϯ 4.3% from basal) compared with the M262T receptor Ϯ decreased the SEAP response at WT and M262T receptors, as (Emax, 71.0 4% from basal). At the Y308H mutant, SNC 80 35 ␥ Ϯ in our previous experiments. In the [ S]GTP S assay, po- activity was low in the SEAP assay (Emax, 85.3 1.1% from Ϯ Ϯ ϭ tencies (EC50) of SNC 80 were similar at WT (EC50, 2.79 basal; EC50, 100.41 62.21 nM; n 3) and undetectable in ϭ Ϯ ϭ 35 ␥ 0.85 nM; n 3) and M262T (EC50, 0.48 0.14 nM; n 4) the [ S]GTP S binding test, partly due to the fact that the receptors. In the SEAP assay also, EC50 values were compa- Y308H mutation strongly impairs SNC 80 binding (46-fold Ϯ ϭ 35 ␥ rable at WT (EC50, 0.17 0.06 nM; n 5) and M262T decrease in affinity, see Table 1) and that [ S]GTP S bind- Ϯ ϭ mutant (EC50, 0.16 0.01 nM; n 4) receptors. In the ing is less sensitive compared with SEAP assay. SNC 80 was 35 ␥ [ S]GTP S binding assay, maximal responses (Emax) were also inactive on mutant C328R (data not shown), although in Ϯ similar for the two receptors (WT, Emax, 187.8 6%; M262T, this case, SNC 80 binding affinity was not impaired (see Ϯ Emax, 174.5 5.9% from basal binding). In the SEAP assay, Table 1). From our previous data (De´caillot et al., 2003), it is the agonist efficacy was higher at the WT receptor (Emax, likely that the strong spontaneous activity of both Y308H Downloaded from jpet.aspetjournals.org

Fig. 4. Dose responses for inverse agonist activity of at ASPET Journals on January 24, 2020

N,N(CH3)2-Dmt-Tic-NH2. Dose response curves were established on WT, M262T, Y308H, and C328R CAM receptors (top to bottom) using [35S]GTP␥S binding (left) or the SEAP reporter gene assay (right). Responses to the reference agonist and inverse agonist (SNC 80 and ICI 174864)

are displayed together with responses to N,N(CH3)2-Dmt- Tic-NH2. For purpose of clarity only one representative experiment is show in each experimental condition. EC50 and efficacy values from several experiments are indicated in Table 2 and in text. In the [35S]GTP␥S binding assay, membranes were incubated with the ligand at concentrations ranging from 3.16 pM to 100 ␮M. [35S]GTP␥S binding activities were normalized to basal activity (absence of ligand), defined as 100%. For the SEAP assay, transfected cells were incubated for 5 h with FK alone (10 ␮M) or with various concentrations of the ligand (3 pM to 100 ␮M). SEAP activity values were normalized to the FK signal (100%). SNC 80 was ineffective at the C328R mutant. 418 Tryoen-To´thetal.

and C328R mutants limits further modulation by an agonist. M262T receptor (64 Ϯ 11% and 193 Ϯ 2%, n ϭ 2), as earlier. Whether other delta opioid agonists are able to activate these Both these activities were inhibited in the presence of NTI (1 mutant receptors remains to be tested. Representative dose ␮M), with SEAP signal values close to values obtained with responses for SNC 80 activity at WT, M262T, and Y308H FK alone (102 Ϯ 9% and 104 Ϯ 2%, n ϭ 2, respectively). mutants are shown in Fig. 4. Similar results were obtained for the WT receptor (data not The reference inverse agonist ICI 174864 decreased shown). [35S]GTP␥S binding and increased SEAP responses at WT Dose Responses of TICP(⌿) at WT, M262T, Y308H, and mutant receptors (Figs. 4 and 5). In the [35S]GTP␥S and C328R Mutant Receptors. Although inverse agonist Ϯ assay, ICI 174868 efficacies (Emax) were 76.9 2.4% (WT, activity was detected at saturating concentrations (Figs. 2 n ϭ 2), 73.9 Ϯ 2.4% (M262T, n ϭ 4), 92.8 Ϯ 2.2% (Y308H, n ϭ and 3), and although affinities at all the receptors were in the Ϯ ϭ 2), and 68.5 2.6% (C328R, n 2) of basal values. Emax in nanomolar range (Table 2), we could not obtain reproducible Ϯ ϭ Ϯ 35 ␥ the SEAP test were 147.5 21.5% (WT, n 3), 170.1 9.5 EC50 values in [ S]GTP S binding dose-response experi- (M262T, n ϭ 5), 174.2 Ϯ 12.7% (Y308H, n ϭ 3), and 159.5 Ϯ ments at WT, M262T, and C328R receptors (data not shown). ϭ 21.8% (C328R, n 3) of FK-stimulated signal. Potencies However, Emax values were consistent across experiments could not be determined accurately, reflecting the extremely (WT, 79.1 Ϯ 1.9%, n ϭ 2; M262T, 77.3 Ϯ 1.8%, n ϭ 7; C328R, low affinity of ICI 174864 compound toward WT and mutant 73.8 Ϯ 3.6%, n ϭ 3), demonstrating negative effect of this delta receptors (Ͼ400 nM, Table 1) compared with the other compound at these mutants. Interestingly, TICP(⌿) was ex-

delta compounds. tremely active at the Y308H receptor (Fig. 5). In the Downloaded from 35 ␥ ⌿ Potencies and efficacies of N,N(CH3)2-Dmt-Tic-NH2 are [ S]GTP S binding assay, TICP( ) decreased the basal sig- shown in Table 2. Inverse agonism of this compound was nal by 66.2 Ϯ 5.6% (n ϭ 5). In the SEAP assay, the enhance- significant at the WT receptor in the two functional assays. ment of FK activation reached 227.7 Ϯ 12.3% (n ϭ 4) of basal Efficacies were consistently higher at CAM receptors com- level. Therefore, in both functional assays, maximal effects of pared with WT, reaching 63.8 Ϯ 2.5% and 248.5 Ϯ 16.7% of TICP(⌿) at this mutant were comparable with those obtained 35 ␥ basal values in the [ S]GTP S binding and SEAP assays, with N,N(CH3)2-Dmt-Tic-NH2 and corresponded to maximal jpet.aspetjournals.org respectively (see Y308H receptor). Potencies paralleled affin- changes that could be obtained under our experimental con- Ϯ 35 ␥ ities (Table 1) for all the receptors and were in the low ditions. EC50 values were 20. 7 5.6 ([ S]GTP S binding) Ϯ nanomolar range. The inverse agonist activity of N,N(CH3)2- and 17.4 2.2 (SEAP) nM and were comparable with the Ki ⌿ Dmt-Tic-NH2 is illustrated by representative dose-response value at this receptor (Table 1). In conclusion, TICP( ) curves displayed in Fig. 4, where functional activity of this showed spectacular inverse agonism at the Y308H receptor compound is compared with those of the reference agonist and modest inverse agonism at the WT and other CAM

SNC 80 and inverse agonist ICI 174864. receptors. at ASPET Journals on January 24, 2020 The closely related Dmt-Tic-OH compound showed similar properties (Table 2). Potencies and efficacies were compara- Discussion ble with those of N,N(CH3)2-Dmt-Tic-NH2 at the WT receptor but tended to be lower at mutant receptors, particularly in Inverse agonism at delta opioid receptors was previously the SEAP assay. Together with our previous observation of demonstrated in various cell lines expressing either native or an overall lower efficacy across the 27 mutant receptors (Fig. wild-type recombinant receptor. Here, we have used a collec- 3), it seemed that inverse agonism was slightly less robust for tion of mutant receptors with high basal activity as another this compound. Adding to this, results from Dmt-Tic-OH approach to examine inverse agonist properties of delta com- were generally less reliable across experiments (data not pounds. We have examined delta ligands from distinct struc-

shown), probably due to lower stability of the compound that tural families, and our data show that N,N(CH3)2-Dmt-Tic- forms a diketopiperazine (Capasso et al., 1995; Balboni et al., NH2 exhibits consistent and potent inverse agonism at WT 1997) and subsequently loses activity (Balboni et al., 1997). and mutant receptors, TICP(⌿) displays negative activity in

N,N(CH3)2-Dmt-Tic-NH2 remains, therefore, the best candi- a receptor-dependent fashion, and NTI and related deriva- date for future inverse agonist studies, particularly in phys- tives are essentially neutral compounds. iological preparations. An innovative aspect of this study is the use of CAM delta Finally, we tested whether the inverse agonist activity of receptors to better assess inverse agonism of delta opioid

N,N(CH3)2-Dmt-Tic-NH2 could be reversed by a neutral an- compounds. Indeed, when the 15 compounds were screened tagonist. In a new set of experiments, we observed that SNC for negative activity, stronger activities were detected at the ␮ ␮ 80 (1 M) and N,N(CH3)2-Dmt-Tic-NH2 (1 M), respectively, M262T mutant compared with WT receptor. This demon- decreased and increased the FK-induced SEAP signal at the strates that, as was shown for other GPCRs (Rossier et al.,

Fig. 5. Dose responses for inverse agonist activity of TICP(⌿) at the Y308H receptor. TICP(⌿) activity was measured in [35S]GTP␥S binding experiment (left) and in the SEAP assay (right). Responses to the reference agonist and inverse agonist (SNC 80, ICI 174864) are also represented for comparison. [35S]GTP␥S binding activities as well as SEAP responses were normalized to basal activity ϭ (absence of ligand 100%). EC50 and efficacy values from four ([35S]GTP␥S) and four (SEAP) independent experiments are indicated in Results. Inverse Agonism at Constitutively Active Delta Receptors 419

TABLE 2 Potencies and efficacies of Dmt-Tic compounds to enhance the FK-stimulated SEAP signal and decrease basal ͓35S͔GTP␥S binding at WT and mutant delta receptors

EC50 and maximal activation values are calculated from dose-response curves for each ligand in both functional assays. Representative experiments for N,N(CH3)2-Dmt- ͓35 ͔ ␥ Tic-NH2 are shown in Fig. 4. Maximal responses to the compounds were normalized to basal S GTP S binding or FK-stimulated SEAP response obtained in the absence of ligand, defined as 100%. Values are means Ϯ S.E.M. from n separate experiments performed in triplicate.

N,N(CH3)2-Dmt-Tic-NH2 H-Dmt-Tic-OH

EC 50 Efficacy n EC 50 Efficacy n

nM % nM % ͓35S͔GTP␥S WT 4.60 Ϯ 3.48 86.45 Ϯ 4.49 2 4.85 Ϯ 1.63 88.28 Ϯ 2.85 3 M262T 2.30 Ϯ 0.62 75.91 Ϯ 3.53 4 8.50 Ϯ 4.54 82.63 Ϯ 0.66 5 Y308H 70.04 Ϯ 13.44 63.81 Ϯ 2.53 3 163 Ϯ 43 72.46 Ϯ 1.80 3 C328R 0.68 Ϯ 0.25 68.94 Ϯ 3.46 3 0.46 Ϯ 0.02 86.24 Ϯ 0.05 2 SEAP WT 2.73 Ϯ 0.72 134.01 Ϯ 9.51 3 3.17 Ϯ 1.41 134.32 Ϯ 9.50 4 M262T 6.70 Ϯ 2.41 173.72 Ϯ 21.80 4 47.93 Ϯ 18.15 147 Ϯ 84 Y308H 65.54 Ϯ 15.76 248 Ϯ 17 4 311 Ϯ 40 187 Ϯ 73 C328R 3.55 Ϯ 0.85 185 Ϯ 13 3 78.49 Ϯ 32.98 156 Ϯ 15 3 Downloaded from 1999; Behan and Chalmers, 2001), a CAM delta receptor is a same three Dmt-Tic compounds in a [35S]GTP␥S binding better detector of inverse agonism than the WT receptor. study using a HEK 293 cell line stably expressing the WT Confirming this, inverse agonist efficacies of the various com- delta receptor. They showed neutral antagonism, partial ago-

pounds were generally comparable at the WT receptor, nism, and full inverse agonism for Dmt-Tic-NH2, Dmt-Tic- whereas partial and full inverse agonism could be distin- OH, and N,N(CH3)2-Dmt-Tic-NH2, respectively. Both their

guished at the several CAM delta receptors, when full dose data and our study, therefore, put forward N,N(CH3)2-Dmt- jpet.aspetjournals.org responses were performed (see Figs. 4 and 5). Tic-NH2 as the more active delta inverse opioid agonist. As The inverse agonist activity of ICI 174864 at the WT re- many other Dmt-Tic opioids, this compound is delta selective Ϯ ␮ Ϯ ceptor appeared modest in our expression system, compared (Ki delta, 0.31 0.05 nM; Ki , 511.4 74.6 nM; Salvadori et with previous studies that have used other cellular models al., 1995). In addition, because the N-methylation augments (see Introduction and references therein). Replacing Naϩ by hydrophobicity and prevents spontaneous cyclization (Bryant et ϩ K ions, for example, which successfully increased ICI al., 2003), N,N(CH3)2-Dmt-Tic-NH2 is considered a stable com-

174864 activity in the case of excitable cells such as pound (Bryant et al., 2003). Altogether, therefore, N,N(CH3)2- at ASPET Journals on January 24, 2020 NG108-15 (Costa and Herz, 1989; Szekeres and Traynor, Dmt-Tic-NH2 appears to be the best pharmacological tool to 1997) or GH3 cells (Liu and Prather, 2002), was ineffective in explore the spontaneous activity of delta receptors. our HEK 293 cells. Therefore, when HEK 293 cells are used Peptides from the TIPP family showed unanticipated ac- as host cells, CAM receptors represent a clear advantage for tivities. TIPP was originally developed as a delta opioid an- the study of inverse agonism. tagonist (Schiller et al., 1992). Consistent with this observa- Another interesting finding was that the negative activity tion, TIPP was inactive in our [35S]GTP␥S binding assay. of the Dmt-Tic compounds was significant at many distinct However, in more recent studies, Martin et al. (2001) re- CAM receptors, independently from the localization of the ported that TIPP could show agonistic properties in several mutation. Evidence that different mutations may produce cell lines using cAMP measurements. In addition, the latter multiple activate conformations has been suggested by us authors showed that, in GH3 cells, TIPP exhibited inverse and others on the basis of signaling or trafficking studies (see agonist or agonist properties, depending on whether the end- Whistler et al., 2002 or De´caillot et al., 2003, among others). point measurement was performed early (receptor binding) Our finding that Dmt-Tic compounds are active at all the or late (cAMP response) along the signaling cascade (Martin tested CAM receptors, with mutations distributed through- et al., 2002). In line with these observations, we found ago- out the receptor proteins, strongly suggests that these com- nistic properties for TIPP in the SEAP reporter gene assay pounds are able to inhibit the several potentially distinct that is sensitive to cAMP levels. Together with the study by active conformations (De´caillot et al., 2003). Dmt-Tic mole- Martin et al., therefore, our data suggest that TIPP may cules should, therefore, represent reliable and potent inverse activate receptors to a small extend, and that the agonistic agonists in distinct experimental contexts. nature of TIPP is revealed after signal amplification only. The three pseudopeptides from the Dmt-Tic family (Bryant This activity, however, may not be easily detectable since et al., 2003) all showed negative activity in the [35S]GTP␥S Pineyro et al. (2001) showed no TIPP-induced modifications binding screening experiment. The reason why inverse ago- of cAMP levels when the delta receptor was expressed in

nism of Dmt-Tic-NH2 was not confirmed in the SEAP assay, HEK 293 cells. Differences between the study of Pineyro et despite a better sensitivity of this assay, remains open. Dmt- al. and our study may arise from distinct assay conditions.

Tic-OH and N,N(CH3)2-Dmt-Tic-NH2 otherwise remained Whether TIPP behaves as agonist or inverse agonist under strong inverse agonists throughout the study, the latter ex- physiological conditions remains to be explored. ⌿ hibiting consistently slightly better activity. N,N(CH3)2- TICP( ) was derived from TIPP by reduction of the peptide 2 3 3 Dmt-Tic-NH2, therefore, appears as a full inverse agonist, bound between Tic and Phe and saturation of the Phe with an efficacy comparable with that of ICI 174864 and with aromatic ring, leading to a chemically stable pseudopeptide ␮ a much higher potency (nanomolar instead of micromolar). with high delta selectivity (Ki delta, 0.259 nM; Ki ,1050 nM; Labarre et al. (2000) compared functional activity of the Schiller et al., 1996). Like TIPP, TICP(⌿) was reported to be 420 Tryoen-To´thetal. a potent antagonist (Schiller et al., 1999). In their study Befort K, Filliol D, De´caillot FM, Gave´riaux-Ruff C, Hoehe MR, and Kieffer BL (2001) A single nucleotide polymorphic mutation in the human mu-opioid receptor assessing functional activities of TIPP (see above), Pineyro et severely impairs receptor signaling. J Biol Chem 276:3130–3137. al. (2001) also examined TICP(⌿) and reported that TICP(⌿) Befort K, Tabbara L, Kling D, Maigret B, and Kieffer BL (1996) Role of transmembrane residues of the ␦-opioid receptor in ligand recognition. J Biol Chem 271: acted as a full inverse agonist (similar to ICI 174864) for the 10161–10168. cAMP response. In our study, TICP(⌿) showed only modest Befort K, Zilliox C, Filliol D, S Y and Kieffer BL (1999) Constitutive activation of negative activity at the WT delta receptor, although we used delta opioid receptor by mutations in transmembrane domain III and VII. J Biol Chem 274:18574–18581. a similar expression system and a similar endpoint measure- Behan DP and Chalmers DT (2001) The use of constitutively active receptors for ment. Clearly, the detection of negative activity is dependent drug discovery at the G protein-coupled receptor gene pool. Curr Opin Drug Discov Dev 4:548–560. on the exact assay conditions. Bryant SD, Jinsmaa Y, Salvadori S, Okada Y, and Lazarus LH (2003) Dmt and TICP(⌿) nevertheless exhibited inverse agonism at the opioid peptides: a potent alliance. Biopolymers 71:86–102. Bryant SD, Salvadori S, Cooper PS, and Lazarus LH (1998) New D-opioid antagonists mutant Y308H receptor in our study. In fact, not only as pharmacological probes. Trends Pharmacol Sci 19:42–46. TICP(⌿) but also the two Dmt-Tic compounds behaved as Capasso S, Sica F, Mazzarella L, Balboni G, Guerrini R, and Salvadori S (1995) Acid full inverse agonists at this receptor, whereas ICI 174864 catalysis in the formation of dioxopiperazines from peptides containing tetrahydroisoquinoline-3-carboxylic acid at position 2. Int J Pept Protein Res 45:567–573. activity was low. The Y308H receptor, therefore, seemed Chiu TT, Yung LY, and Wong YH (1996) Inverse agonistic effect of ICI-174,864 on particularly responsive to peptides from the TIPP and the cloned delta-opioid receptor: role of G protein and adenylyl cyclase activation. Mol Pharmacol 50:1651–1657. Dmt-Tic families. The reason why TICP(⌿) showed strong Corbett AD, Paterson SJ, and Kosterlitz HW (1993) Selectivity of ligands for opioid inverse agonism at this receptor and, more generally, why receptors, in Handbook of Experimental Pharmacology, Opioids I (Herz A ed) pp 645–679, Springer-Verlag, Berlin, Germany. the pseudopeptides were highly active at the Y308H recep- Costa T and Herz A (1989) Antagonists with negative intrinsic activity at ␦ opioid Downloaded from tor remains unexplained. In contrast to mutations M262T receptors coupled to GTP-binding proteins. Proc Natl Acad Sci 86:7321–7325. Cotton R, Giles MG, Miller L, Shaw JS, and Timms D (1984) ICI 174864: a highly and C328R, this mutation is located centrally within the selective antagonist for the opioid ␦-receptor. Eur J Pharmacol 97:331–332. binding pocket and drastically modifies opioid binding (Be- de Ligt RA, Kourounakis AP, and AP IJ (2000) Inverse agonism at G protein-coupled fort et al., 1996, 1999). Specific atomic interactions may receptors: (patho)physiological relevance and implications for drug discovery. Br J Pharmacol 130:1–12. take place between these inverse agonists and the signal- Decaillot FM, Befort K, Filliol D, Yue S, Walker P, and Kieffer BL (2003) Opioid receptor random mutagenesis reveals a mechanism for G protein-coupled receptor ing core of the receptor (Tm 3/Tm 6/Tm 7; Decaillot et al., jpet.aspetjournals.org activation. Nat Struct Biol 10:629–636. 2003) that strongly favor the inactive receptor conforma- Filliol D, Ghozland S, Chluba J, Martin M, Matthes HW, Simonin F, Befort K, tion. Three-dimensional modeling studies, in the future, Gaveriaux-Ruff C, Dierich A, LeMeur M, et al. (2000) Mice deficient for delta- and mu-opioid receptors exhibit opposing alterations of emotional responses. Nat Genet may clarify specific structural features within the 25:195–200. TICP(⌿)-Y308H receptor complex. Hosohata K, Burkey TH, Alfaro-Lopez J, Hruby VJ, Roeske WR, and Yamamura HI Labarre et al. (2000) showed weak partial inverse agonism (1999) (2S, 3R) TMT-L-Tic-OH is a potent inverse agonist at the human delta opioid receptor. Eur J Pharmacol 380:R9–R10. for the NTI derivative HS-378. Our data showed no activity Kieffer BL (1997) Molecular aspects of opioid receptors, in The Pharmacology of Pain (Dickenson A and Besson JM eds) pp 281–303, Springer-Verlag, Berlin, Germany.

for this compound, as well as for a set of five closely related at ASPET Journals on January 24, 2020 Kieffer BL and Gave´riaux-Ruff C (2002) Exploring the opioid system by gene knock- HS compounds, at the CAM M262T receptor. We even ob- out. Prog Neurobiol 66:285–306. served weak partial agonism at the WT receptor for most of Labarre M, Butterworth J, St-Onge S, Payza K, Schmidhammer H, Salvadori S, Balboni G, Guerrini R, Bryant SD, and Lazarus LH (2000) Inverse agonism by dmt-Tic them. Together, it seems that NTI and NTI derivatives (HS analogues and HS 378, a naltrindole analogue. Eur J Pharmacol 406:R1–R3. compounds) are weakly active at the delta receptor, and this Liu JG and Prather PL (2002) Chronic agonist treatment converts antagonists into inverse agonists at delta-opioid receptors. J Pharmacol Exp Ther 302:1070–1079. activity could be slightly negative (Labarre et al., 2000), Martin NA, Ruckle MB, VanHoof SL, and Prather PL (2002) Agonist, antagonist and neutral, or slightly positive (this study) depending on the inverse agonist characteristics of TIPP (H-Tyr-Tic-Phe-Phe-OH), a selective delta- experimental system. Compared with ICI 174864 and the opioid receptor ligand. J Pharmacol Exp Ther 301:661–671. Martin NA, Terruso MT, and Prather PL (2001) Agonist activity of the delta- two active Dmt-Tic compounds, these compounds could, antagonists TIPP and TIPP-psi in cellular models expressing endogenous or trans- therefore, be considered as neutral ligands. fected delta-opioid receptors. J Pharmacol Exp Ther 298:240–248. Merkouris M, Mullaney I, Georgoussi Z, and Milligan G (1997) Regulation of spon- In conclusion, we have investigated the functional activity taneous activity of the delta opioid receptor: studies of inverse agonism in intact of several peptidic and nonpeptidic delta opioid compounds cells. J Neurochem 69:2115–2122. Milligan G (2003) Principles: extending the utility of [35S]GTP␥S binding assays. across delta receptors with variable levels of spontaneous Trends Pharmacol Sci 24:87–90. ␦ activity. Our data indicate that the N,N(CH3)2-Dmt-Tic-NH2 Mullaney I, Carr IC, and Milligan G (1996) Analysis of inverse agonism at the opioid receptor after expression in Rat-1 fibroblasts. Biochem J 315:227–234. compound represents a useful inverse agonist to modulate Neilan CL, Akil H, Woods JH, and Traynor JR (1999) Constitutive activity of the the spontaneous activity of delta receptors in recombinant delta opioid receptor expressed in C6 glioma cells: identification of non-peptide and possibly physiological situations. Our data also show delta-inverse agonists. Br J Pharmacol 128:556–562. Pineyro G, Azzi M, De Lean A, Schiller P, and Bouvier M (2001) Short-term inverse- that TIPP peptides are endowed with intriguing properties agonist treatment induces reciprocal changes in delta-opioid agonist and inverse- that deserve further exploration in different experimental agonist binding capacity. Mol Pharmacol 60:816–827. Rossier O, Abuin L, Fanelli F, Leonardi A, and Cotecchia S (1999) Inverse agonism conditions. NTI derivatives remain essentially neutral in our and neutral antagonism at alpha(1a)- and alpha(1b)- adrenergic receptor sub- experiments, and the most selective compounds (HS ligands) types. Mol Pharmacol 56:858–866. Salvadori S, Attila M, Balboni G, Bianchi C, Bryant SD, Crescenzi O, Guerrini R, may serve as neutral antagonists in future experiments. Picone D, Tancredi T, Temussi PA, et al. (1995) Delta opioidmimetic antagonists: prototypes for designing a new generation of ultraselective opioid peptides. Mol Med 1:678–689. Acknowledgments Schiller PW, Nguyen TMD, Weltrowska G, Wilkes BC, Marsden BJ, Lemieux C, and We thank Christophe Iderne, Mathieu Ballie´, and Gabriel Recht Chung NN (1992) Differential stereochemical requirements of ␮ vs.Ѩ opioid receptors for ligand binding and signal transduction: development of a class of potent for important contribution; Severo Salvadori for the generous gift of and highly Ѩ-selective peptide antagonists. Proc Natl Acad Sci USA 89:11871– Dmt-Tic compounds; and Dominique Massotte for critical reading of 11875. the manuscript. Schiller PW, Weltrowska G, Berezowska I, Nguyen TM, Wilkes BC, Lemieux C, and Chung NN (1999) The TIPP opioid peptide family: development of delta antagonists, delta agonists and mixed mu agonist/delta antagonists. Biopolymers 51:411– References 425. Balboni G, Guerrini R, Salvadori S, Tomatis R, Bryant SD, Bianchi C, Attila M, and Schiller PW, Weltrowska G, Nguyen TM-D, Lemieux C, Chung NN, Zelent B, Wilkes Lazarus LH (1997) Opioid diketopiperazines: synthesis and activity of a prototypic BC, and Carpenter KA (1996) Peptides: chemistry, structure and biology, in class of opioid antagonists. Biol Chem 378:19–29. Proceedings of the 14th American Peptide Symposium (Kaumaya TP, Hodges RS Inverse Agonism at Constitutively Active Delta Receptors 421

eds) 1995 June 18–23; Columbus, OH. pp 609–611, Mayflower Scientific Ltd., guanosine-5Ј-O-(3-[35S] thio)triphosphate to NG108–15 cell membranes: charac- Kingswinford, United Kingdom. terization of agonist and inverse agonist effects. J Pharmacol Exp Ther 283:1276– Schmidhammer H, Krassnig R, Greiner E, Schultz J, White A, and Berzetei-Gurske 1284. IP (1998) Synthesis and biological evaluation of 14-alkoxymorphinans: XV. Novel Whistler JL, Gerber BO, Meng EC, Baranski TJ, von Zastrow M, and Bourne HR delta opioid receptor antagonists with high affinity and selectivity in the 14- (2002) Constitutive activation and endocytosis of the complement factor 5a recep- alkoxy-substituted indolomorphinan series. Helv Chim Acta 81:1064–1069. tor: evidence for multiple activated conformations of a G protein coupled receptor. Schutz J, Dersch CM, Horel R, Spetea M, Koch M, Meditz R, Greiner E, Rothman Traffic 3:866–877. RB, and Schmidhammer H (2002) Synthesis and biological evaluation of 14- Zaki PA, Keith DE Jr, Thomas JB, Carroll FI, and Evans CJ (2001) Agonist-, alkoxymorphinans: XVII. Highly delta opioid receptor selective 14-alkoxy- antagonist- and inverse agonist-regulated trafficking of the delta-opioid receptor substituted indolo- and benzofuromorphinans. J Med Chem 45:5378–5383. correlates with, but does not require, G protein activation. J Pharmacol Exp Ther Seifert R and Wenzel-Seifert K (2002) Constitutive activity of G-protein-coupled 298:1015–1020. receptors: cause of disease and common property of wild-type receptors. Naunyn Schmiedebergs Arch Pharmacol 366:381–416. Address correspondence to: Kieffer Brigitte, Institut de Ge´ne´tique et Bi- Strange PG (2002) Mechanisms of inverse agonism at G-protein-coupled receptors. ologie Mole´culaire et Cellulaire, 1 rue Laurent Fries, B.P. 1042, 67404 Illkirch Trends Pharmacol Sci 23:89–95. Cedex, France. E-mail: [email protected] Szekeres PG and Traynor JR (1997) Delta opioid modulation of the binding of Downloaded from jpet.aspetjournals.org at ASPET Journals on January 24, 2020