3 pages due by 20 hrs(pete)

Desensitization of the Human Motilin Receptor by Motilides

Leen Thielemans, Inge Depoortere, Jason Perret, Patrick Robberecht, Yaoquan Liu, Theo Thijs, Chris Carreras, Emmanuel Burgeon, and Theo L. Peeters Gut Hormone Lab, Center for Gastroenterological Research, Department of Pathophysiology, Katholieke Universiteit Leuven, Leuven, Belgium (L.T., I.D., T.T., T.L.P.); Department of Biological Chemistry and Nutrition, Faculty of Medicine, Université Libre de Bruxelles, Brussels, Belgium (J.P., P.R.); Euroscreen, Brussels, Belgium (E.B.); and Kosan Biosciences, Hayward, California (Y.L., C.C.)

Received November 30, 2004; accepted March 9, 2005

ABSTRACT Tachyphylaxis may have contributed to the failure of the moti- lide ABT-229 [N-ethyl, N-methyl 4�� deoxy erythromycin (EM)-B enolether] in clinical trials. We compared the desensitizing po- tency of structurally related motilides [EM-A, EM-A enolether (ME4), N-ethyl, N-methyl EM-A (ME36), EM-B enolether (ME67), N-ethyl, N-methyl EM-A enolether (EM523), ABT-229 and 4�� deoxy EM-A enolether (KOS1326)] in a Chinese hamster ovary (CHO)-K1 cell line expressing the human motilin receptor (MTLR) and in rabbit duodenal segments. CHO-MTLR cells were preincubated with motilides prior to stimulation with mo- tilin. The negative logarithm of the preincubation concentration reducing the maximal motilin-induced Ca2� flux to 50% was calculated (pDC50). Internalization was visualized in CHO-K1 cells containing an enhanced green fluorescent protein (EGFP)- tagged MTLR and quantified in binding experiments. The con- tractile response of repeated stimulations was measured in duodenal segments. In CHO-MTLR cells, the pDC50 was ABT-

229 (8.78) � motilin (7.77) � EM-A (4.78), different from their order of potency to induce Ca2� release (pEC50): motilin (9.39) � ABT-229 (8.46) � EM-A (7.11). In cells with the EGFP- tagged MTLR, ABT-229 decreased membrane fluorescence by 25 � 2% compared with 16 � 2% for motilin and 8 � 2% for EM-A. Binding studies confirmed that EM-A did not induce MTLR internalization (residual binding 96 � 4% compared with motilin, 31 � 3% and ABT-229, 21 � 1%). Comparison of the pDC50 and pEC50 values of the other motilides ME4 (5.90; 8.08), ME67 (6.03; 8.12), ME36 (3.32; 6.62), EM-523 (6.02; 8.22), and KOS1326 (7.32; 8.14) suggested that the strong desensitizing properties of ABT-229 are mostly related to the removal of the 4�-OH of the cladinose sugar. The decline of the contractile response in duodenal segments correlated with the pDC50. The ability to desensitize and internalize the MTLR is not only determined by potency. This may be an important criterion for the development of a clinically useful compound.

The finding that the antibiotic erythromycin-A (EM-A) inter- acts with the motilin receptor (Peeters et al., 1989) became clinically relevant when it was discovered that EM-A acceler- ated delayed gastric emptying in patients with diabetic gastro- paresis (Janssens et al., 1990). As a result, more potent motor- stimulating EM derivatives were developed without antibiotic activity: the motilides (Omura et al., 1987). Recently the clinical development of the most promising motilide ABT-229 was stopped because of its failure to improve symptoms in patients with functional dyspepsia with normal or delayed gastric emp- tying (Talley et al., 2000) and in patients with type 1 diabetes

mellitus (Talley et al., 2001). The authors of the studies with ABT-229 concluded that there was no future for drugs with the prokinetic profile of the motilides, but others have argued that ABT-229 had undesirable properties that may be avoided in new motilides (Tack and Peeters, 2001; Camilleri, 2002). The most important property in this respect is the development of tachyphylaxis. Previous studies described a decreased contrac- tile response of rabbit colonic myocytes due to down-regulation of motilin receptors after chronic oral treatment (Bologna et al., 1993) and after short-term intravenous treatment with eryth- romycin A (Depoortere et al., 1991), the parent compound from which ABT-229 was derived. One-month treatment of rabbits with ABT-229 also resulted in a down-regulation of motilin receptors and a reduced contractile response to ABT-229 and motilin (Depoortere et al., 1999). Tachyphylaxis was also ob- served after treatment of healthy subjects with ABT-229. A single dose of ABT-229 increased gastric emptying after the

This work was supported by grants from the Flemish Foundation for Sci- entific Research (FWO G.0144.04) and the Belgian Ministry of Science (GOA 03/11 and IAP P5/20).

Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.

doi:10.1124/jpet.104.081497.

ABBREVIATIONS: EM-A, erythromycin-A; CHO, Chinese hamster ovary; MTLR, motilin receptor; EGFP, enhanced green fluorescent protein; ABT-229, N-ethyl, N-methyl 4�� deoxy EM-B enolether; MTLR-EGFP, EGFP-tagged motilin receptor; PBS, phosphate-buffered saline; BSA, bovine serum albumin; PCR, polymerase chain reaction; ACh, acetylcholine; ME4, EM-A enolether; ME36, N-ethyl, N-methyl EM-A; ME67, EM-B enolether; EM523, N-ethyl, N-methyl EM-A enolether; KOS1326, 4�� deoxy EM-A enolether.

0022-3565/05/3133-1397–1405$20.00 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 313, No. 3 Copyright © 2005 by The American Society for Pharmacology and Experimental Therapeutics 81497/1199571 JPET 313:1397–1405, 2005 Printed in U.S.A.

1397

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

first meal, but emptying of a second meal was not affected, although ABT-229 plasma levels were still high (Verhagen et al., 1997).

Recently, we have shown that the Ca2� fluxes induced by peptidyl and nonpeptidyl motilin agonists in a Chinese ham- ster ovary (CHO) cell line expressing the cloned human mo- tilin receptor and the Ca2� indicator apoaequorin (CHO- MTLR) correlate strongly with the contractile response of these compounds in rabbit duodenal strips (Thielemans et al., 2002). In the present study, we used the aequorin-based luminescence assay to study the desensitization of the moti- lin receptor by ABT-229 in more detail. To clarify whether desensitization is a common characteristic of all motilides, the desensitizing properties of several motilides with differ- ent structural features were compared. To determine the mechanism underlying desensitization, receptor internaliza- tion following application of ABT-229 was quantified by re- ceptor binding studies and visualized by using an enhanced green fluorescent protein (EGFP)-tagged motilin receptor (MTLR-EGFP).

Materials and Methods Materials

Norleucine13-porcine-motilin ([Nle13]po-motilin) was purchased at Eurogentec (Namur, Belgium) and will further be referred to as motilin. Erythromycin-A (EM-A), EM-A enol ether (ME4), N-ethyl, N-methyl EM-A (ME36), EM-B enol ether (ME67), and N-ethyl, N-methyl EM-A enol ether (EM-523) were obtained from Prof. J. Hoogmartens (Laboratory of Pharmaceutical Chemistry, University of Leuven, Belgium). N-Ethyl, N-methyl 4� deoxy EM-B enol ether (ABT-229) was a gift from Dr. P. Lartey (Lartey et al., 1995; Abbott Laboratories, Abbott Park, IL). Macrolide KOS1326 was synthesized by Kosan Biosciences, Inc. (Hayward, CA).

Cell Culture

A CHO-K1 cell line stably expressing the human motilin receptor and the mitochondrially targeted apoaequorin was obtained from Euroscreen SA (Brussels, Belgium). The cells were cultured in Ham’s F-12 containing 10% fetal bovine serum, 100 �g/ml streptomycin, 2.5 �g/ml amphotericin B (Invitrogen, Carlsbad, CA), 400 �g/ml G418, and 5 �g/ml puromycin (Sigma-Aldrich, St. Louis, MO) and spliced once a week with 5 mM EDTA in PBS.

Calcium Measurements: Aequorin Luminescence Assay

Suspended cells (5 � 106 cells/ml) were loaded with coelenterazine h (5 �M) (Molecular Probes, Leiden, The Netherlands) at room temperature for at least 4 h to reconstitute active aequorin. Cells were then diluted 10-fold with BSA medium (Dulbecco’s modified Eagle’s medium/Ham’s F-12 with Hepes, without phenol red, 0.1% BSA, 100 IU/ml penicillin, 100 �g/ml streptomycin, and 2.5 �g/ml amphotericin B) and 100 �l/well was injected (50,000 cells) into a 96-well plate containing increasing concentrations of the test com- pound diluted in BSA medium.

Immediately after injection of cells, the emitted light was mea- sured using the Microlumat plus luminometer (Berthold Technolo- gies, Bad Wildbad, Germany) for 20 s. The intensity of the emitted light was integrated using the Winglow software (Berthold Technol- ogies), yielding for each well one value representative of the emitted light and, hence, of the stimulation of the motilin receptor by the agonist present in the well. This value was expressed as a percentage of maximal stimulation obtained with Triton X-100 (0.9%). All values were corrected for background by subtracting the blanc value (� BSA medium). The negative logarithm of the concentration producing half of the maximal response (pEC50) was calculated from the dose-

response curves by linear interpolation. All experiments were per- formed in duplicate, and each compound was tested at least two times. Results are represented as mean � standard error of the mean (S.E.M.).

To assess desensitization, cells were pretreated for at least 1 h with different concentrations of agonist during loading, washed, and centrifuged before dose-response curves were established with moti- lin. The effect of pretreatment on the maximal response was as- signed as the effect of desensitization. The average of the maximal plateau values of the dose-response curve to motilin under desensi- tizing conditions was expressed as a percentage of maximal response under control conditions. The potency of a compound to induce de- sensitization was calculated after plotting the remaining maximal responses as a function of the concentration of the compound used to desensitize. A sigmoid concentration-response curve was fitted through the data by nonlinear regression analysis using the Graph- Pad Prism software (GraphPad Software Inc., San Diego, CA), and a pDC50 value and its standard error of the best-fit value (negative logarithm of the concentration producing half-maximal desensitiza- tion) was calculated. For each compound, � values were calculated by subtraction of pEC50 values with pDC50 values. The square root of the sum of squares of the standard deviation (S.E.M. for pEC50) and the standard error of the best-fit value (S.E. for pDC50) resulted in the standard deviation of these � values. A pooled standard deviation (sp) was derived from which a t value was calculated for the compar- ison of two � values, with n1 � n2 � 2 degrees of freedom using:

t0 � x� 1 � x� 2

sp �1n1 � 1n2 .

Receptor Binding Studies

Membrane Preparation. CHO-K1 cells expressing the MTLR were cultured until 90% confluence was reached. After removal of medium, cells were scraped from the plates in Ca2�/Mg2�-free PBS. After centrifugation for 3 min at 1500 g, pellets were resuspended in a buffer containing 15 mM Tris-HCl pH 7.5, 2 mM MgCl2, 0.3 mM EDTA, and 1 mM EGTA and homogenized in a glass homogenizer. The crude membrane fraction was collected by two consecutive cen- trifugation steps at 40,000g for 25 min separated by a washing step in the same buffer. The final pellet was resuspended in a buffer containing 7.5 mM Tris-HCl pH 7.5, 12.5 mM MgCl2, 0.3 mM EDTA, 1 mM EGTA, and 250 mM sucrose and flash frozen in liquid nitro- gen. The protein content was determined by the Folin method (Lowry et al., 1951).

Cell Membrane Receptor Binding. Competition binding assays were performed by incubating CHO-MTLR membranes (8 �g of protein/tube) in a final volume of 0.1 ml with 125I-motilin (0.3 nM) for 60 min at 31°C. Unlabeled motilin, motilin fragments, and motilides were used as competitors at concentrations ranging from 10�12 to 10�4 M in binding buffer (25 mM Hepes pH 7.4, 1 mM CaCl2, 5 mM MgCl2, 0.5% BSA). After incubation, the samples were filtered on GF/B filters, washed, and counted in a gamma counter. All values were corrected for nonspecific binding determined by addition of an excess (10�5 M) of unlabeled motilin. IC50 values were determined by nonlinear regression using a single-site model (PRISM; GraphPad Software Inc.). Each compound was tested three times in duplicate.

Construction and Expression of the MTLR Coding Sequence as an N-Terminal Fusion Protein with the EGFP

The cDNA for the full coding sequence inserted in the pcDNA3.1 vector (provided by Euroscreen SA) was amplified by PCR using the proof reading Taq polymerase Pfu Turbo (Stratagene, La Jolla, CA). The 5� primer contained the XhoI restriction site followed by a Kozak consensus sequence preceding the ATG initiation codon and 18 gene- specific nucleotides. The 3� primer contained a 5� EcoRI restriction site followed by the 26 terminal nucleotides of the coding sequence,

1398 Thielemans et al.

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

the TGA stop codon being mutated to remove the translation termi- nation signal and provide an in phase sequence with the EGFP coding sequence that follows immediately the multiple cloning site in the pEGFP plasmid (BD Biosciences Clontech, Palo Alto, CA). The PCR product was digested by DpnI restriction enzyme to remove parental target DNA, purified on QIAquick PCR purification col- umns (QIAGEN GmbH, Hilden, Germany), and digested by XhoI and EcoRI. The pEGFP plasmid was digested in the same manner. The digestion products were purified on QIAquick PCR purification col- umns and an aliquot of the resulting 50 �l was run on a 0.8% Tris acetate EDTA agarose gel. A fraction of the digestion products, i.e., the motilin coding sequence and the pEFGP vector, were mixed and dried in a speed vacuum concentrator. The dried DNA was resus- pended, ligated with T4 DNA ligase (Promega, Madison, WI), and transformed into TOP10 One Shot Ultracompetent Escherichia coli bacterial host cells according to manufacturer’s instructions (Invitro- gen, Carlsbad, CA). Several colonies were picked; miniprep DNA was prepared using the GFX Micro Plasmid Preparation kit (Amersham Biosciences Inc., Piscataway, NJ), and the presence of an insert was verified by gel electrophoresis. Three clones were checked by auto- mated DNA sequencing on an ABI machine using BigDye Termina- tor chemistry. One clone was then amplified using the GenElute Endotoxin-Free Midi Prep kit (Sigma-Aldrich). Transfection of 20 �g of purified plasmid DNA was carried out by electroporation (Electro- porator II; Invitrogen) according to the manufacturer’s instructions into CHO-PAM28 cells—CHO cells stably expressing the aequorine gene (Euroscreen SA). After selection by G418 (Invitrogen), 24 indi- vidual clones were picked and further amplified for functional char- acterization.

Visualization of Endocytosis of the MTLR-EGFP

CHO cells containing the MTLR-EGFP were grown on four-well coverglass chambers until 80% confluency. After washes with ice- cold BSA medium, digital pictures were made with an Olympus camera on an inverted Nikon microscope (40� oil) equipped with a fluorescence unit and filtered with filtercube B-2A (EX BP540 –580, DM RK595, EM BA 600 – 610). BSA medium containing 10�5 M compound of interest was added, after which the cells were gently shaken at 4°C for 1 h for equilibration binding, followed by 1 h incubation at 37°C. Internalization was stopped by washing the cells with ice-cold PBS. Pictures of the same cells were taken. For quan- tification, the fluorescent grayscale pictures were converted to black and white. The percentage black and white at the membrane and in the cytosol was calculated for 10 cells per condition before and after stimulation using Scion Image (Scion Corporation, Frederick, MD) and compared with control. Each condition was evaluated in three different experiments. Means were compared using an unpaired t test (PRISM; GraphPad Software Inc.).

Quantification of MTLR-EGFP Internalization

The capacity of different compounds to promote internalization was examined by adding the compound of interest (10�5 M) to the CHO cells containing the MTLR-EGFP in 12-well plates for 60 min at 37°C. Surface-bound ligands were removed with a gentle acid wash (50 mM sodium citrate, 0.2 mM sodium phosphate, 90 mM NaCl, and 0.1% bovine serum albumin, pH 5.0; 10 min, 4°C), which does not affect subsequent receptor binding, and then a radiorecep- tor binding assay was performed (125I-motilin, 0.3 nM, 4 h at 4°C) to measure receptors remaining at the cell surface. Internalized recep- tors were expressed as a percentage loss of cell surface binding compared with cells not exposed to the compound of interest. Means � S.E.M. were compared using an unpaired t test (PRISM; GraphPad Software Inc.).

Contractility Studies

Integral rabbit duodenal segments (1.5–2 cm) were vertically sus- pended in tissue baths containing Hepes buffer pH 7.4 (11.6 mM

Hepes, 11.5 mM glucose, 137 mM NaCl, 5.9 mM KCl, 1.2 mM CaCl2, and 1.2 mM MgCl2) continuously gassed with 100% O2 and kept at 37°C. Contractions were recorded isotonically using HP 7DCDT- 1000 transducers from Hewlett Packard (Palo Alto, CA) and a dis- placement control unit obtained from Janssen Scientific Instrument Division (Beerse, Belgium). Signal output was connected to a re- corder and to a computer and sampled for digital analysis using the WinDaq data acquisition system and a DI-2000 PGH card (Dataq Instruments, Akron, OH).

Strips were equilibrated in the tissue bath until a stable response to acetylcholine (ACh, 0.1 mM) was obtained. Strips were then stim- ulated with either ABT-229 (69 nM), KOS1326 (0.17 �M), EM-A (14.1 �M), ME4 (0.44 �M), or EM-523 (66 nM) at a concentration corresponding to five times their potency to induce contractions. After 4 min, strips were washed out two times and washed again 2�, 6�, and 10� after the first wash out. The washing procedure was then repeated every 10 min until 1 h after the application of the agonists. Strips were then stimulated for a second time with the compound of interest and washed again according to the procedure described above. After the third application, strips were not washed out but a supramaximal dose of motilin (1 �M) followed by a supramaximal dose of ACh (0.1 mM) was applied. Desensitization was expressed as a percentage of the response of the agonists obtained after the first application. The response to motilin at the end of the experiment was expressed relative to the response to ACh. All experiments were approved by the Ethical Committee of the University of Leuven.

Results Desensitization of the Motilin Receptor. Desensitiza-

tion of the motilin receptor was induced by prestimulation of CHO-MTLR cells with different concentrations (10�12, 10�9, 10�8, 10�7, 10�6, or 10�5 M) motilin. This reduced the max- imal response in a subsequent concentration response curve to motilin to 97.2 � 1.3, 83.9 � 7.8, 64.9 � 1.3, 3.8 � 1.3, 2.8 � 0.9, and 1.6 � 1.3%, respectively (Fig. 1).

Next, motilin’s ability to desensitize was compared with that of the motilides ABT-229 and EM-A. These compounds are structurally related but differ at four positions, as illus- trated in Fig. 2.

Both agonists were tested at 10�5 M, a concentration suf- ficient to evoke a maximal Ca2� luminescent response. How- ever, following pretreatment at this concentration, there was

Fig. 1. Concentration-dependent desensitization of the motilin receptor. Pretreatment of CHO-MTLR cells with motilin at the concentrations indicated decreased the maximal Ca2� response of a subsequent concen- tration-response curve to motilin. The maximal response of the second stimulation is expressed as the percentage of control (no pretreatment) and is the mean � S.E.M. of at least three experiments in duplicate.

Differential Motilin Receptor Desensitization 1399

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

a marked difference in the maximal response to a subsequent stimulation to motilin, which was reduced to 65.8 � 1.8% for EM-A and to �0.2 � 0.4% for ABT-229. The desensitizing capacity of ABT-229 but not of EM-A was therefore compa- rable with that of motilin.

To investigate which structural elements underlie the stronger desensitizing potency of ABT-229 in comparison to EM-A, several EM-A derivatives were tested, containing one or two ABT-229-like modifications (Fig. 2). The modifications include the presence of an enol configuration (ME4, ME67, EM-523, and KOS1326), the presence of an ethyl group at the amine group of the desosamine sugar (ME36, EM-523), and the lack of a hydroxyl at position 12 of the lactone ring (ME67) or at position 4� of the cladinose sugar (KOS1326).

Figure 3 shows that compounds with an enol configuration such as ABT-229, ME4, EM-523, ME67, and KOS1326 in- duced a more profound desensitization at 10�5 M than EM-A. In contrast, desensitization was seemingly absent with ME36, which compared with EM-A, contains only a modifi- cation at the amine group of the desosamine sugar.

However, the enol configuration also increases the potency of motilides to induce contractions in muscle strips and to induce a Ca2� flux in the CHO-MTLR cells (Depoortere et al., 1989; Thielemans et al., 2002). Therefore, although the pre- treatment concentration was sufficient to induce a maximal

response for all compounds tested, one could argue that less potent compounds may require higher pretreatment concen- trations. To fully discriminate between potency and ability to desensitize, we therefore established a “concentration re- sponse curve of desensitization” for each motilide. Figure 4 shows the effect of pretreatment of the CHO-MTLR cells with increasing concentrations of motilin, EM-A, or ABT-229 on the maximal response to a subsequent concentration re- sponse curve to motilin.

From these curves, the pretreatment concentration which reduces the maximal response to 50% of the control value was calculated. The negative logarithm of this concentration, pDC50 (�log[desensitizing concentration 50%]) reflects the potency to desensitize. The values for all motilides tested are shown in Table 1. This table also lists the potencies of these compounds to induce Ca2� fluxes (pEC50 values) in un- treated cells (Thielemans et al., 2002).

Although ABT-229 is 8.5 times less potent than motilin in activating the cells (pEC50: 8.46 � 0.08 versus 9.39 � 0.03), it is 10.2 times more potent in inducing desensitization (pDC50: 8.78 � 0.03 versus 7.77 � 0.08). Thus, potency does not seem to solely determine the ability to desensitize. To further illustrate this point, Table 1 shows the difference between the pDC50 and pEC50 values, referred to as � value.

Five compounds have a similar � value (motilin, 1.62 �

Fig. 2. Structural features of the motilides used. Shown are the structures of EM-A and ABT-229 with the differences between them either in gray or bold. The table indicates which of the structural modifications present in ABT-229 are also present in the different motilides.

1400 Thielemans et al.

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

0.15; EM-A, 2.33 � 0.28; ME4, 2.18 � 0.42; ME67, 2.09 � 0.25; and EM-523, 2.20 � 0.28) indicating that there is a relationship between potency and desensitizing ability (Pear- son’s r � 0.9054). However, for ABT-229 and KOS1326, the � value is significantly lower than in these compounds (p 0.01) and for ME36, it is significantly higher (p 0.01).

Binding of Motilides with the MTLR. Because both activation and desensitization result from the interaction of the ligand with its receptor, we investigated whether the higher than expected ability of ABT-229 and KOS1326 to desensitize could be explained by a higher affinity for the motilin receptor. The pIC50 values (Table 1) calculated from the displacement curves did not reveal an increased affinity for ABT-229 or KOS1326.

Comparison of MTLR-EGFP and Wild-Type Recep- tor. Because desensitization involves receptor internaliza- tion, an EGFP-tagged motilin receptor was generated and stably expressed in CHO cells to follow receptor trafficking. The cells also expressed the Ca2� indicator apoaequorin

(MTLR-EGFP), which allowed us to determine whether GFP interfered with signal transduction and desensitization.

For the wild-type receptor and the MTLR-EGFP, the order of potency to induce Ca2� release is: motilin�ABT- 229�EM-A (Fig. 5A). The absolute potency (pEC50) of all compounds is 9.17 � 1.14 (motilin), 8.12 � 0.17 (ABT-229), and 6.45 � 0.08 (EM-A) in the MTLR-EGFP and is somewhat lower than in the MTLR.

Pretreatment of MTLR-EGFP with 10�5 M motilin, ABT- 229, or EM-A reduced the response to a subsequent stimula- tion with motilin to 18.7 � 2.5, 2.9 � 1.3, and 99.2 � 8.5%, respectively (Fig. 5B). These effects are less pronounced than in the wild-type receptor, but the discrepancy between acti- vation and desensitization remains because ABT-229, al- though less potent than motilin, desensitizes most at 10�5 M.

Internalization of the EGFP-Tagged MTLR. In cells expressing the EGFP-tagged receptor, changes in the fluo- rescence distribution after stimulation with the ligands at 10�5 M were studied by fluorescence microscopy. Images obtained before and after stimulation are shown in Fig. 6A.

Before the start of the experiment, MTLR-EGFP was prin- cipally localized at the plasma membrane. After 1-h stimu- lation at 37°C in the presence of the motilin agonists, the intensity of the signal at the cell surface was reduced, espe- cially for motilin and ABT-229 and was concentrated in the cytosol.

The images were analyzed and the percentage change in fluorescence in the cytosol and in the membrane owing to stimulation was calculated (Fig. 6B). All agonists induced a significant cytosolic increase and a significant decrease of membrane fluorescence compared with control (p 0.005). The cytosolic increases were similar for motilin, ABT-229, and EM-A (22 � 2, 25 � 2, and 20 � 2%, respectively, p 0.0001 versus control). In contrast, ABT-229 (25 � 2%) de- creased membrane fluorescence significantly more compared with motilin (16 � 2%, p 0.01), which in turn induced a more significant decrease compared with EM-A (8 � 2%, p 0.01).

Radioligand Binding with the EGFP-Tagged MTLR. To validate the semiquantitative data of the analysis of the fluorescence distribution, the more quantitative method of receptor binding was used to measure ligand-induced endo- cytosis. Radioligand binding with 125I-motilin was performed after pretreatment of MTLR-EGFP with 10�5 M motilin, ABT-229, or EM-A for 20 and 60 min at 37°C (Fig. 7). Resid- ual binding after 20-min prestimulation was 44.9 � 2.3% (motilin), 30.8 � 3.8% (ABT-229), and 124.3 � 2.0% (EM-A) and was further reduced to 30.8 � 3.3% (p � 0.0029), 21.3 � 0.8% (p 0.0001), and 96.1 � 4.4% (p 0.0001), respectively, after prestimulation for 60 min.

In line with the desensitization data, internalization after prestimulation with ABT-229 was higher (p � 0.029) than after stimulation with motilin or EM-A, although all com- pounds are equipotent at the concentration used for pre- stimulation. Curiously, prestimulation with EM-A for 20 min induced a significant increase in binding (124.3 � 2.0%; p 0.0001).

Contractility Studies. The desensitizing properties of ABT-229, KOS1326, EM-A, ME4, and EM-523 were also com- pared in vitro in the tissue bath by determining the contrac- tile response of segments of rabbit duodenum to three appli-

Fig. 3. Comparison of desensitization induced by different motilides. Cells were first stimulated with a motilide at 10�5 M (x-axis), and the maximal response to a subsequent concentration-response curve to mo- tilin was determined and expressed as a percentage of control (no pre- treatment). Each value represents the mean � S.E.M. of at least two experiments in duplicate.

Fig. 4. Concentration response curves of desensitization induced by mo- tilin, ABT-229, and EM-A. Cells were pretreated with different concen- trations of the test compounds (x-axis), and the reduction in maximal response to a subsequent concentration response curve to motilin was calculated and expressed as a percentage of control (y-axis). Each value is mean � S.E.M. of at least two experiments in duplicate.

Differential Motilin Receptor Desensitization 1401

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

cations with a 1-h interval. Figure 8 shows that for all compounds the contractile response decreases, but the effect was most pronounced with ABT-229; the third application resulted in a contraction that was only 8.92 � 0.96% of its first application. EM-A is at the other end with 86.50 � 4.22% (n � 3). Compounds EM-523 and ME4 showed a com- parable decrease after the third application (56.58 � 4.71 and 55.20 � 8.86%, respectively, n � 4).

Discussion Our study shows that motilides desensitize the motilin

receptor to different extents in relation to their ability to activate the release of Ca2�. However, this does not apply to all motilides since ABT-229 has a much stronger desen- sitizing potency than may be expected from its activity or from its binding affinity suggesting that for some motil- ides, activation and desensitization are partially de- coupled. The more profound desensitization was also translated in an enhanced internalization of the motilin

receptor to the cytosol following stimulation with ABT-229 as was evidenced by radioligand binding and by using an EGFP-tagged MTLR. A comparison of the effect of struc- tural changes on changes in activity and desensitization suggests that the removal of the hydroxyl at position 4� of the cladinose sugar contributes most to the strong desen- sitizing properties of ABT-229, although a combination with other structural changes may be required. One may question the relevance of the model we developed to study desensitization because CHO cells, although convenient, may have different activation pathways than gastrointes- tinal smooth muscle. However, the data we obtained in the tissue bath using the classical model to study the potency of motilides, the contractile response of the rabbit duode- num, corroborate the conclusions of the cellular model. We suggest that the strong desensitizing properties of ABT- 229 may have contributed to its clinical failure and that a cellular model may be a convenient way to develop motil- ides without desensitizing properties.

TABLE 1 Summary of the experimental parameters for the different motilides

Test Method pEC50 � S.E.M. Luminescence pIC50 � S.E.M. Binding pDC50 � S.E. Desensitization pEC50–pDC50 � S.E.�

Motilin 9.39 � 0.03 8.55 � 0.05 7.77 � 0.08 1.62 � 0.15 ABT-229 8.46 � 0.08 7.23 � 0.10 8.78 � 0.03 �0.32 � 0.16 EM-A 7.11 � 0.05 5.51 � 0.10 4.78 � 0.26 2.33 � 0.28 ME36 6.62 � 0.02 4.65 � 0.06 3.32 � 0.15 3.30 � 0.16 ME4 8.08 � 0.01 6.27 � 0.13 5.90 � 0.42 2.18 � 0.42 ME67 8.12 � 0.10 6.49 � 0.10 6.03 � 0.10 2.09 � 0.25 EM-523 8.22 � 0.01 6.59 � 0.06 6.02 � 0.28 2.20 � 0.28 KOS1326 8.14 � 0.16 7.16 � 0.08 7.32 � 0.0002 0.82 � 0.28

Fig. 5. A, comparison of the Ca2� release evoked after concentration-dependent activation of wild type (top) and MTLR-EGFP (bottom) by motilin and motilides. Concentration-response curves of motilin, ABT-229, and EM-A are expressed as percentage of Triton X-100 values. B, comparison of desensitization of wild-type (f) and EGFP-tagged MTLR (�) by motilin and motilides. The response to motilin after pretreatment with a ligand at 10�5

M is expressed as percentage versus control (no pretreatment). Values are mean � S.E.M. of three experiments in duplicate.

1402 Thielemans et al.

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

As a general rule, ligands cause both receptor-mediated signaling and signal desensitization. In the case of G protein- coupled receptors, a ligand signals by activating its receptor, which in turn activates a second messenger (G protein) inside the cell. G protein-coupled receptor desensitization involves

the uncoupling of receptors from their heterotrimeric G pro- teins, resulting in the internalization (endocytosis) of recep- tors to endosomes and down-regulation (Ferguson, 2001).

The events of activation and desensitization have been linked in many receptor systems, among which the � adre- noceptor has been characterized best (Benovic et al., 1988). Our data confirm that there is indeed a correlation between the ability to activate and desensitize, e.g., EM-A is less potent than the enol ether derivatives ME4, EM-523, and ME67 and also induces less desensitization. However, the results with ABT-229 suggest a decoupling of activation and desensitization. Such discrepancies have also been described for the �-opioid receptor, which is related to the motilin receptor, in the sense that both receptors can be stimulated by a peptide (motilin versus enkephalin) and by a naturally occurring nonpeptide (motilides versus morphine, metha- done, etorphine, etc.). The desensitization of the �-opioid receptor mediated by methadone and L-�-acetyl methadone was more pronounced than for morphine and was dispropor- tionate to their efficacies (Yu et al., 1997). Furthermore, functional desensitization of the monkey D1A dopamine re- ceptor cannot be predicted reliably from the agonist potency for stimulating adenylate cyclase (Lewis et al., 1998).

The general paradigm that activation and internalization are coupled is based on studies on �2-adrenergic and musca- rinic receptors demonstrating that partial agonists cause less internalization than full agonists and the amount of receptor internalization caused by an agonist generally correlates with coupling efficiency to G proteins (Toews and Perkins, 1984; Thompson and Fisher, 1990; Szekeres et al., 1998). Again, exceptions have been described (Mahan et al., 1985; Barak et al., 1994). Perhaps the best evidence that G protein coupling is not required for receptor endocytosis is provided by the observation that in S49 murine lymphoma cell lines, which either lack Gs� or have point mutations preventing receptor/G protein interactions, �2-adrenergic internaliza- tion in response to agonist stimulation is normal (Mahan et al., 1985).

In our data, the discrepancy between activation and desen- sitization is translated into a discrepancy between activation and internalization. Similarly, observations with the �- and �-opioid receptor in transfected cells and enteric neurons in intact animals show that both peptidyl (enkephalin) and

Fig. 6. A, fluorescence imaging of MTLR-EGFP trafficking in CHO-K1 cells before (top panel) and after (bottom panel) stimulation with agonists (10�5 M) for 1 h at 37°C. Before treatment, fluorescence is mainly located at the cell surface, whereas relocation to the cytosol took place after treatment with motilin, EM-A, and ABT-229. B, quantification of the change in fluorescence distribution between the cell surface and the cytosol after stimulation with motilin, EM-A, or ABT-229. Three different sets of cells were stimulated with the same ligand. The change in fluo- rescence was measured in 10 cells per set. Values were expressed as mean � S.E.M. #, p 0.0001 versus control; �, p 0.01 versus ABT-229; ��, p 0.01 versus motilin and p 0.0001 versus control.

Fig. 7. Residual binding of 125I-motilin at the surface of MTLR-EGFP cells pretreated with motilin agonists at 10�5 M for 20 or 60 min. Results are expressed as percentage residual binding versus nontreated cells (mean � S.E.M., n � 3 in duplicate). ���, p 0.0001 versus control and ��, p 0.005; �, p 0.05 versus motilin.

Fig. 8. Contractility study. Repeated applications of motilides to rabbit duodenal muscle strips result in decreased contractility. For each motil- ide, this is expressed as percentage of the contractility that occurred during the first application.

Differential Motilin Receptor Desensitization 1403

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

nonpeptidyl (etorphine) agonists stimulate receptor endocy- tosis in the expected manner. Remarkably, the high affinity �-opioid receptor agonist morphine does not cause detectable internalization, even at concentrations that strongly inhibit adenylate cyclase (Keith et al., 1996; Sternini et al., 1996).

Our results indicate that desensitization and internaliza- tion of the motilin receptor are coupled in line with the paradigm. However, recently it was shown in the same cel- lular context that morphine does desensitize the �-opioid receptor without causing internalization (Borgland et al., 2003). Therefore, it has been suggested that receptor endo- cytosis is a mechanism for resensitization and down-regula- tion rather than desensitization (Tsao et al., 2001). As the motilin receptor in our study is stimulated for 1 h to cause desensitization, it cannot be excluded that the activity that is still measured is the sum of desensitization, internalization, resensitization, and down-regulation.

All of these apparently diverse effects can be accounted for in terms of the “ensemble theory” (Kenakin, 2002a,b). This theory states that a receptor can adopt numerous conforma- tions, each unfolding and exposing different regions to the intracellular apparatus and each associated with specific functions (activation, desensitization, internalization, etc.). These conformations may differ, even for ligands with similar affinity and related structures, and may lead, for example, to different desensitizing properties despite similar activation potencies. Apparently our data illustrate this point because ABT-229, structurally closely related to EM-523, ME4, and ME67 and with a comparable potency, causes stronger de- sensitization. Our data also suggest that the presence of an ethyl at the amine group of the desosamine sugar, as in ME36, reduces the desensitizing ability, whereas removal of the hydroxyl at position 4� of the cladinose sugar, as in ABT-229 and KOS1326 but not at position 12 of the lactone ring as in ME-67, causes an increase.

Internalization was also measured by receptor binding studies. Prestimulation with EM-A induced receptor up-reg- ulation. An increase in receptor density after ligand stimu- lation has been demonstrated for A1 adenosine receptors after antagonist treatment (Ciruela et al., 1997), and inverse agonists can up-regulate histamine H2 receptors (Alewijnse et al., 1998), cannabinoid receptors (Bouaboula et al., 1999), and opioid receptors (Zaki et al., 2001). Both agonist and antagonist treatments have been shown to up-regulate dopa- mine receptors (Filtz et al., 1994; Zhang et al., 1994; Geurts et al., 1999).

Our data support the hypothesis that tachyphylaxis might underlie the disappointing outcome of motilides in clinical trials. Internalization of the motilin receptor might explain why the high blood levels of ABT-229 were not able to induce gastric contractions and, hence, accelerate gastric emptying of a second meal in healthy subjects (Verhagen et al., 1997). However, the effects of ABT-229 have not just been described as nonexistent; moreover a worsening of symptoms has been reported by patients with functional dyspepsia (Talley et al., 2000) and diabetic gastroparesis (Talley et al., 2001). This might lead to the conclusion that ABT-229 is still active (Camilleri, 2002) after a 4-week treatment, apparently con- tradicting the occurrence of tachyphylaxis. This paradox might be resolved if one considers the long-term action of ABT-229 as the induction of desensitization rather than ac- tivation. Indeed, when motilin receptors are not present on

the cell surface, even endogenous motilin will not be able to induce contractions. The observation that less antral contrac- tions after a second meal occur in healthy volunteers treated with ABT-229 compared with placebo (Verhagen et al., 1997) is consistent with this hypothesis. A down-regulation of mo- tilin receptors has also been shown in rabbits that had been treated with ABT-229 for 4 weeks (Depoortere et al., 1999). Whether this underlies the worsening of symptoms is still to be proven, as gastric emptying and antral contractility have not been measured at the end of both studies by Talley et al. (2000, 2001). Besides gastric emptying, impaired accommo- dation and hypersensitivity are also considered as underly- ing pathophysiological mechanisms that can be targeted for treatment of dyspeptic patients. A higher tone, increased intensity, and frequency of spontaneous contractions in the proximal stomach and a reduction of meal-induced relaxation have been described for erythromycin (Bruley des Varannes et al., 1995; Distrutti et al., 1999; Piessevaux et al., 2001). It has therefore been pointed out that the worsening of symp- toms might also be due to an effect of ABT-229 on gastric fundus relaxation (Tack and Peeters, 2001). Combining both hypotheses, the adverse effects might be caused by (acute) spatial nonspecific action of ABT-229 in decreasing fundus relaxation and by (chronic) functional nonspecific action in decreasing antral motility by tachyphylaxis.

In conclusion, our data suggest that it should be possible to develop motilides with high potency and less desensitizing ability. Therefore, it is indicated that before attempting large-scale clinical trials, the desensitizing potency of a mo- tilide should be evaluated.

References Alewijnse AE, Smit MJ, Hoffmann M, Verzijl D, Timmerman H, and Leurs R (1998)

Constitutive activity and structural instability of the wild-type human H2 recep- tor. J Neurochem 71:799 – 807.

Barak LS, Tiberi M, Freedman NJ, Kwatra MM, Lefkowitz RJ, and Caron MG (1994) A highly conserved tyrosine residue in G protein-coupled receptors is required for agonist-mediated beta 2-adrenergic receptor sequestration. J Biol Chem 269: 2790 –2795.

Benovic JL, Bouvier M, Caron MG, and Lefkowitz RJ (1988) Regulation of adenylyl cyclase-coupled beta-adrenergic receptors. Annu Rev Cell Biol 4:405– 428.

Bologna SD, Hasler WL, and Owyang C (1993) Down-regulation of motilin receptors on rabbit colon myocytes by chronic oral erythromycin. J Pharmacol Exp Ther 226:852– 856.

Borgland SL, Connor M, Osborne PB, Furness JB, and Christie MJ (2003) Opioid agonists have different efficacy profiles for G protein activation, rapid desensiti- zation and endocytosis of mu-opioid receptors. J Biol Chem 278:18776 –18784.

Bouaboula M, Dussossoy D, and Casellas P (1999) Regulation of peripheral canna- binoid receptor CB2 phosphorylation by the inverse agonist SR 144528. Implica- tions for receptor biological responses. J Biol Chem 274:20397–20405.

Bruley des Varannes S, Parys V, Ropert A, Chayvialle JA, Roze C, and Galmiche JP (1995) Erythromycin enhances fasting and postprandial proximal gastric tone in humans. Gastroenterology 109:32–39.

Camilleri M (2002) Motilin agonists and dyspepsia: throwing out the baby with the bath water. Gut 51:612; author reply 612– 613.

Ciruela F, Saura C, Canela EI, Mallol J, Lluis C, and Franco R (1997) Ligand- induced phosphorylation, clustering and desensitization of A1 adenosine recep- tors. Mol Pharmacol 52:788 –797.

Depoortere I, Peeters TL, Matthijs G, Cachet T, Hoogmartens J, and Vantrappen G (1989) Structure-activity relation among macrolide antibiotics in inducing contrac- tions and in displacing bund motilin in rabbit duodenum. J Gastrointest Motil 1:150 –159.

Depoortere I, Peeters TL, and Vantrappen G (1991) Effect of erythromycin and of octreotide on motilin receptor density in the rabbit. Regul Pept 32:85–94.

Depoortere I, Verlinden M, Thijs T, Van Assche G, and Peeters TL (1999) The motilide ABT-229 selectively downregulates motilin receptors in different tissues. Evidence for motilin receptor subtypes. Gastroenterology 116:A1061.

Distrutti E, Azpiroz F, Soldevilla A, and Malagelada JR (1999) Gastric wall tension determines perception of gastric distention. Gastroenterology 116:1035–1042.

Ferguson SS (2001) Evolving concepts in G protein-coupled receptor endocytosis: the role in receptor desensitization and signaling. Pharmacol Rev 53:1–24.

Filtz TM, Guan W, Artymyshyn RP, Facheco M, Ford C, and Molinoff PB (1994) Mechanisms of up-regulation of D2L dopamine receptors by agonists and antag- onists in transfected HEK-293 cells. J Pharmacol Exp Ther 271:1574 –1582.

Geurts M, Hermans E, and Maloteaux JM (1999) Enhanced striatal dopamine D(2)

1404 Thielemans et al.

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from

receptor-induced [35S]GTPgammaS binding after haloperidol treatment. Eur J Pharmacol 382:119 –127.

Janssens J, Peeters T, Vantrappen G, Tack J, Urbain J, De Roo M, Muls E, and Bouillon R (1990) Improvement of gastric emptying in diabetic gastroparesis by erythromycin. Preliminary studies [see comments]. N Engl J Med 322:1028 –1031.

Keith DE, Murray SR, Zaki PA, Chu PC, Lissin DV, Kang L, Evans CJ, and von Zastrow M (1996) Morphine activates opioid receptors without causing their rapid internalization. J Biol Chem 271:19021–19024.

Kenakin T (2002a) Drug efficacy at G protein-coupled receptors. Annu Rev Pharma- col Toxicol 42:349 –379.

Kenakin T (2002b) Efficacy at G-protein-coupled receptors. Nat Rev Drug Discov 1:103–110.

Lartey PA, Nellans HN, Faghih R, Petersen A, Edwards CM, Freiberg L, Quigley S, Marsh K, Klein LL, and Plattner JJ (1995) Synthesis of 4�-deoxy motilides: identification of a potent and orally active prokinetic drug candidate. J Med Chem 38:1793–1798.

Lewis MM, Watts VJ, Lawler CP, Nichols DE, and Mailman RB (1998) Homologous desensitization of the D1A dopamine receptor: efficacy in causing desensitization dissociates from both receptor occupancy and functional potency. J Pharmacol Exp Ther 286:345–353.

Lowry DH, Rosenbrough MJ, Farr AL, and Randall RJ (1951) Protein measurement with the Folin fenol reagent. J Biol Chem 193:732–737.

Mahan LC, Koachman AM, and Insel PA (1985) Genetic analysis of beta-adrenergic receptor internalization and down- regulation. Proc Natl Acad Sci USA 82:129 – 133.

Omura S, Tsuzuki K, Sunazuka T, Marui S, Toyoda H, Inatomi N, and Itoh Z (1987) Macrolides with gastrointestinal motor stimulating activity. J Med Chem 30: 1941–1943.

Peeters TL, Matthijs G, Depoortere I, Cachet T, Hoogmartens J, and Vantrappen G (1989) Erythromycin is a motilin receptor agonist. Am J Physiol 257:G470 –G474.

Piessevaux H, Tack J, Wilmer A, Coulie B, Geubel A, and Janssens J (2001) Per- ception of changes in wall tension of the proximal stomach in humans. Gut 49:203–208.

Sternini C, Spann M, Anton B, Keith DE, Bunnett NW, von Zastrow M, Evans C, and Brecha NC (1996) Agonist-selective endocytosis of mu opioid receptor by neurons in vivo. Proc Natl Acad Sci USA 93:9241–9246.

Szekeres PG, Koenig JA, and Edwardson JM (1998) The relationship between agonist intrinsic activity and the rate of endocytosis of muscarinic receptors in a human neuroblastoma cell line. Mol Pharmacol 53:759 –765.

Tack J and Peeters T (2001) What comes after macrolides and other motilin stimu- lants? Gut 49:317–318.

Talley NJ, Verlinden M, Geenen DJ, Hogan RB, Riff D, McCallum RW, and Mack RJ (2001) Effects of a motilin receptor agonist (ABT-229) on upper gastrointestinal symptoms in type 1 diabetes mellitus: a randomised, double blind, placebo con- trolled trial. Gut 49:395– 401

Talley NJ, Verlinden M, Snape W, Beker JA, Ducrotte P, Dettmer A, Brinkhoff H, Eaker E, Ohning G, Miner PB, et al. (2000) Failure of a motilin receptor agonist (ABT-229) to relieve the symptoms of functional dyspepsia in patients with and without delayed gastric emptying: a randomized double-blind placebo-controlled trial. Aliment Pharmacol Ther 14:1653–1661.

Thielemans L, Depoortere I, Vanden Broeck J, and Peeters TL (2002) The motilin pharmacophore in CHO cells expressing the human motilin receptor. Biochem Biophys Res Commun 293:1223–1227.

Thompson AK and Fisher SK (1990) Relationship between agonist-induced musca- rinic receptor loss and desensitization of stimulated phosphoinositide turnover in two neuroblastomas: methodological considerations. J Pharmacol Exp Ther 252: 744 –752.

Toews ML and Perkins JP (1984) Agonist-induced changes in beta-adrenergic recep- tors on intact cells. J Biol Chem 259:2227–2235.

Tsao P, Cao T, and von Zastrow M (2001) Role of endocytosis in mediating down- regulation of G-protein-coupled receptors. Trends Pharmacol Sci 22:91–96.

Verhagen MA, Samsom M, Maes B, Geypens BJ, Ghoos YF, and Smout AJ (1997) Effects of a new motilide, ABT-229, on gastric emptying and postprandial antroduodenal motility in healthy volunteers. Aliment Pharmacol Ther 11:1077– 1086.

Yu Y, Zhang L, Yin X, Sun H, Uhl GR, and Wang JB (1997) Mu opioid receptor phosphorylation, desensitization and ligand efficacy. J Biol Chem 272:28869 – 28874.

Zaki PA, Keith DE, Thomas JB, Carroll FI, and Evans CJ (2001) Agonist-, antago- nist- and inverse agonist-regulated trafficking of the delta-opioid receptor corre- lates with, but does not require, G protein activation. J Pharmacol Exp Ther 298:1015–1020.

Zhang LJ, Lachowicz JE, and Sibley DR (1994) The D2S and D2L dopamine receptor isoforms are differentially regulated in Chinese hamster ovary cells. Mol Pharma- col 45:878 – 889.

Address correspondence to: T. L. Peeters, Gut Hormone Lab, Gasthuisberg O & N, B-3000 Leuven, Belgium. E-mail: [email protected]

Differential Motilin Receptor Desensitization 1405

at A S

P E

T Journals on O

ctober 11, 2020 jpet.aspetjournals.org

D ow

nloaded from