The determination of the biophysical aspects of drug

Transcription

1 pubs.acs.org/ac Probing Biochemical Mechanisms of Action of Muscarinic M3 Receptor Antagonists with Label-Free Whole Cell Assays Huayun Deng, Chaoming Wang,, Ming Su, and Ye Fang*, Biochemical Technologies, Science and Technology Division, Corning Inc., Corning, New York 14831, United States NanoScience Technology Center, Department of Mechanical, Materials and Aerospace Engineering, University of Central Florida, Orlando, Florida 32816, United States ABSTRACT: Binding kinetics of drugs is increasingly recognized to be important for their in vivo efficacy and safety profiles. However, little is known about the effect of drug binding kinetics on receptor signaling in native cells. Here we used label-free whole cell dynamic mass redistribution (DMR) assays under persistent and duration-controlled stimulation conditions to investigate the influence of the binding kinetics of four antagonists on the signaling of endogenous muscarinic M3 receptor in native HT-29 cells. Results showed that DMR assays under different conditions differentiated the biochemical mechanisms of action of distinct M3 antagonists. When co-stimulated with acetylcholine, tiotropium, a relatively slow binding antagonist, was found to selectively block the late signaling of the receptor, suggesting that acetylcholine attains its binding equilibrium faster than tiotropium does, thereby still being able to initiate its rapid response until the antagonist draws up and fully blocks the signaling. Furthermore, DMR assays under microfluidics allowed estimation of the residence times of these antagonists acting at the receptor in native cells, which were found to be the determining factor for the blockage efficiency of M3 receptor signaling under duration-controlled conditions. This study demonstrates that DMR assays can be used to elucidate the functional consequence of kinetics-driven antagonist occupancy in native cells. The determination of the biophysical aspects of drug target interactions is essential to the understanding of all biological processes and to improve the efficiency of drug discovery and development. Traditionally, equilibrium potency or efficacy obtained in a closed system is used to screen and prioritize drug molecules. However, drug target interactions in vivo often occur far from equilibrium conditions, 1 and poor in vivo efficacy has been recognized to be a primary source of the high drug attrition rate encountered in the recent years, largely due to the poor predicative power of in vitro results. 2 In recent years, increasing data suggest that the kinetics of drug target interactions play an important role in determining efficacy, safety, duration of action, tolerability, indication, and therapeutic differentiation A fast on-rate tends to improve efficacy of inhibitor drugs when the concentration of endogenous agonist is high, while a slow off-rate increases drug target residence time, thus resulting in elongated duration of action. 4,10 In consideration of the common rapid binding of the same family of drugs for a given target, the drug target residence time is considered to be a critical indicator of their in vivo efficacy and toxicity profiles. 4 This is elicited by tiotropium, a long lasting anti-bronchiodilatation drug whose biochemical mechanism of action is believed to be related to its slow dissociation from the muscarinic M3 receptor in the lungs. 13,14 The biochemical mechanism of action describes the kinetics-related mechanism of action of the drugs; 3 this is different from the molecular mechanism of action, or mode of action, which classifies drugs based on their biological effects. The drug target residence time (the reciprocal of K off ) is the time required for an 63% dissociation of the drug target complex 4 and is often measured using biochemical kinetics assays and/or end-point functional assays. 11 Here we applied label-free dynamic mass redistribution (DMR) assays under different conditions to probe the kinetics-driven interactions of representative antagonist drugs with an endogenous M3 receptor in native HT-29 cells. The DMR assay is a phenotypic assay that measures receptor activation and signaling in real time at the whole cell level The recent development of microfluidic DMR assays has added additional revenue to differentiate the biochemical mechanisms of the action of drugs Thus, it is possible to investigate the functional consequence of kinetics-driven antagonist occupancy in native cells using DMR assays under different conditions. We observed that at the whole cell level the blockage of acetylcholine-activated M3 receptor signaling by these antagonists is correlated well with their residence times. EXPERIMENTAL SECTION Materials. Atropine, cholera toxin (CTx), and pertussis toxin (PTx) were purchased from Sigma Chemical Company (St. Louis, MO). Acetylcholine, ipratropium bromide, Y27632, U73122, and N-methyl-scopolamine (NMS) were obtained Received: June 1, 2012 Accepted: September 6, 2012 Published: September 6, American Chemical Society 8232

2 from Tocris Bioscience Co. (St. Louis, MO). Tiotropium was purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Except for acetylcholine, which was dissolved in water, all other drugs were prepared in 10 mm dimethyl sulfoxide (DMSO) and diluted using the assay buffer (1 times Hank s balanced salt solution (HBSS) buffer, 20 mm Hepes, ph 7.1) to the indicated concentrations. Epic 384-well cell culture compatible biosensor microplates and inserts were obtained from Corning Inc. (Corning, NY). Polydimethylsiloxane (PDMS) was obtained from Dow Chemical (Midland, MI). Fabrication of Microfluidic Biosensor Devices. The microfluidic device was fabricated using a method described previously. 19,20 In brief, the designed features on the chrome mask were transferred onto silicon wafers using a standard photolithographic process. A PDMS prepolymer solution containing a mixture of PDMS oligomers and a reticular agent with a 10:1 mass ratio (Sylgard 184 Kit, Dow Chemical) was cast onto patterned silicon wafers and cured at room temperature for about 24 h to minimize shrinkage after curing. Next, the PDMS replicas were carefully peeled away from the wafers. After inlet and outlet holes had been punched, the PDMS replicas were aligned and reversibly bonded onto the top of the biosensor inserts. The microfluidic biosensor device obtained has a 4 4 array of functional resonant waveguide grating (RWG) biosensors, each biosensor having a dimension of 2 2 mm and the array having a footprint that is compatible with a standard 384-well microplate. The biosensor array was made onto a glass substrate using a microfabrication process developed in house. 22 The microfluidic device contains a 4 4 array of microfluidic chambers, each having three inlets and one outlet and the microchamber array covering a 4 8 array of biosensors; therefore, only one biosensor is centered in each microchamber, and the adjacent biosensor is sacrificed. The distance from an inlet to the outlet is 9 mm; the central width of the chamber is 5 mm. To ensure an appropriate cell culture and to minimize the effect of any nonspecific absorption of ligand molecules to the top PDMS surface of the microchamber on the cells, the height of the microchannel was set to 200 μm. The total volume required to fill the chamber was 6 μl. Cell Culture. A human colorectal adenocarcinoma HT-29 cell line was obtained from American Type Cell Culture (Manassas, VA) and passaged using the complete medium (McCoy s 5A medium modified and supplemented with 10% fetal bovine serum, 4.5 g/l glucose, 2 mm glutamine, and 100 μg/ml penicillin and streptomycin) at 37 C with 5% CO 2. The native cells were passed with trypsin/ethylenediaminetetraacetic acid when approaching 90% confluence to provide a new maintenance culture on T-75 flasks and an experimental culture on the biosensor microplates or chambers. The confluency for all cells at the time of the assays was 95%. For cell culture within the microfluidic biosensor device, the device, tubing (Tygon S-54-HL, Saint-Gobain Performance Plastics, Akron, OH), and syringes (500 ml, gastight 1700 series, Hamilton, Reno, NV) were first sanitized with 70% ethanol. Afterward, cells suspended in 6 μl of the complete medium were injected into each chamber. Cells were allowed to seed via a 30 min incubation at room temperature. Then tubing was plugged into the microchamber inlets and connected to syringes that were connected to a syringe pump (model SP230IW; World Precision Instruments, Sarasota, FL). The biosensor device with cultured cells was then maintained within a Petri dish with a cover, wherein extra cell medium within the dish was used to keep the desired humidity to avoid evaporation of the medium. The cells were cultured at 37 C with 5% CO 2, and perfusion with the complete media was started at a flow rate of 5 μl/h. After being cultured overnight, the cells reached a confluency of about 95%. For cell culture in regular Epic biosensor microplates, cells suspended in the complete medium were directly seeded into each well of the microplates. After overnight culturing, the cells reached a confluency of 95%. Quantitative Real-Time Polymerase Chain Reaction (qrt-pcr). Total RNA was extracted from the HT-29 cells using an RNeasy mini kit (Qiagen, catalog no ). Oncolumn DNase digestion was performed using an RNase-free DNase set (Qiagen, catalog no ) to eliminate genomic DNA contamination. The concentration and quality of total RNA were determined using a Nanodrop 8000 (Thermo Scientific). Customized PCR-array plates for 352 G proteincoupled receptor (GPCR) genes and reagents were ordered from SABiosciences (Qiagen, Valencia, CA). About a 1 μg total of RNA was used for each 96-well PCR-array. The PCR-array was performed on an ABI 7300 Real-Time PCR system following the manufacturer s instructions. DMR Assays in Microplates. DMR assays in microplates were considered to be under a persistent stimulation scheme, where cells are exposed to a compound throughout the assays. These assays were performed using a conventional Epic system (Corning Inc.), which is a wavelength-interrogation reader system tailored for RWG biosensors in microtiter plates. 23,24 This system consists of a temperature-control unit, an optical detection unit, and an on-board liquid handling unit operated by robotics. The detection unit is centered on integrated fiber optics and enables kinetic measures of cellular responses with a time interval of 15 s. For DMR assays in the microplate, the confluent cells were washed twice and maintained with the assay buffer (HBSS) and then incubated within the reader system for 1 h. Afterward, a 2 min baseline was established. Compound solutions were then transferred to the biosensor wells using the onboard liquid handling device, and the cell responses were recorded in real time. For antagonist washout experiments, the cells were first incubated with an antagonist for 10 min and then washed five times using the buffer. After further incubation within the system for 75 min, a 2 min baseline was established, the agonist solutions were then added, and the cellular response was monitored. All studies were carried out with at least three replicates unless specifically mentioned. DMR Assays under Microfluidics. A recently developed RWG biosensor imager 22 was used for all DMR assays under microfluidics. This imager is based on a swept wavelength interrogation scheme. In brief, a light beam from a swept tunable light source was directed to focus on the 4 4 microfluidics biosensor device. A high-speed complementary metal-oxide semiconductor (CMOS) digital camera was used to record the escaped and reflected resonant lights with an effective spatial resolution of 25 μm. During a single cycle of wavelength sweep from 825 to 840 nm total, 150 spectral images were acquired every 3 s, and the spectral image stack was processed into a sensor resonant wavelength or a DMR image in real time, leading to a temporal resolution of 3 s. The DMR image was obtained after the starting resonant wavelengths at all locations were normalized to zero. The wavelength sweeping was performed stepwise in 100 pm every 20 ms. To minimize assay variability and the impact of time lag across the biosensor, cells located only within the 8233

3 central area of a biosensor (0.2 2 mm) were sampled to generate an averaged response. For DMR assays under microfluidics, microchambers having confluent cells were connected to three independently operated syringes using Tygon O tubing. Initially, cells were perfused with the assay buffer at a flow rate of 1 μl/min until a baseline signal was established and normalized to zero. With the switch of operational pumps, the cells were then stimulated with a ligand for a specific period of time, followed by perfusion with the assay buffer to remove the ligand. Finally, the cells were stimulated again with a continuous flow of acetylcholine solution. The three independently operated inlets enable continuous perfusion of cells during the assay without introducing any abrupt changes of shear stress or laminar flow perturbation inside the microfluidic chamber. All DMR signals were background corrected. RESULTS AND DISCUSSION We were focused on characterizing the biochemical mechanisms of action of four antagonist drugs, atropine, ipratropium, N-methyl-scopolamine (NMS) and tiotropium, at the endogenous M3 receptor in HT-29 cells using multiple DMR assays (Figure 1). The choice of antagonists was based on their distinct binding kinetics reported in the literature. 25 Acetylcholine, the natural agonist of muscarinic receptors, was used as the reference agonist. The signaling pathways, downstream of the acetylcholine-induced M3 receptor activation, were determined on the basis of the effect of the pretreatment of HT-29 cells with a variety of chemical probe molecules on the characteristics of the acetylcholine-induced DMR. The competitive inhibition of the acetylcholine response by an antagonist was determined using a DMR co-stimulation assay, wherein the two ligands were added together. Time-dependent inhibition of the acetylcholine response by antagonists was performed using a two-step DMR assay, wherein the antagonist was preincubated with cells for different periods of time. The functional recovery of the antagonist-preoccupied receptors was assayed under both static (i.e., in microplates) and perfusion (i.e., with microfluidics) conditions, wherein an antagonist washout step was incorporated between the antagonist and agonist stimulation steps. In consideration of the RWG biosensor sensitivity to thermal fluctuation, and thorough washing of cells in the microplate often causes unwanted perturbations in cells; the two stimulation steps under the static condition have to be separated for a sufficiently long time (>60 min). In contrast, coupled with the relatively low perfusion rate, microfluidics offers precise control of the duration of ligand stimulation and functional recovery of antagonist-preoccupied receptors, without introducing any obvious perturbations during the DMR measurements. Thus, the microfluidic assays are better suited for the determination of the drug residence times than the assays in the microplate. First, we used qrt-pcr to examine the expression of five muscarinic receptors in native HT-29 cells. Results showed that HT-29 cells only express mrnas at a moderate level for the M3 receptor (cycle threshold, Ct, 26.0), but little or no mrnas for M1 (Ct, 38.1), M2 (Ct, 35.0), M4 (Ct, 32.7), and M5 (undetected). As controls, the Ct values for hypoxanthine phosphoribosyltransferase 1 and glyceraldehyde-3-phosphate dehydrogenase in HT-29 cells were found to be 22.0 and 15.9, respectively. This expression pattern was consistent with pharmacological studies reported in the literature. 26,27 Literature mining suggests that HT-29 cells express a very small 8234 Figure 1. DMR assays under different conditions. (a) DMR agonist assay in the microplate wherein only the agonist is introduced to the cells. (b) DMR co-stimulation assay in the microplate wherein both agonist and antagonist are introduced together. (c) Two-step DMR antagonist assay in the microplate wherein an agonist solution is added after the pretreatment of the cells with the antagonist. (d) Three-step DMR assay in the microplate wherein a washing step is introduced between the first antagonist treatment and the second agonist stimulation step. (e) DMR under microfluidics wherein the first antagonist solution 1 is depleted using perfusion with the buffer 2, followed by the agonist solution 3. The white box indicates the dimension and location of the biosensor, while the black box indicates the area where the cellular responses are measured. The colored solution pattern is obtained using numerical modeling, wherein a buffer solution, 2 μl in total, is sandwiched between the first antagonist and the second agonist solution; all flow rates are 1 μl/ min. There is little diffusion occurring when the solution passes through the microchamber. However, the perfusion with the buffer is very effective at depleting the antagonist molecules. amount of acetylcholine esterase, an enzyme that degrades acetylcholine. 28,29 Second, we performed acetylcholine dose responses in HT- 29 cells using microplate-based DMR assays. The results showed that acetylcholine triggered a dose-dependent and saturable DMR in HT-29 cells (Figure 2a), leading to EC 50 values of ± 9.8, ± 17.5, and ± 14.3 nm (six independent measurements, n = 12), based on its amplitude at 3, 8, and 50 min poststimulation, respectively (Figure 2b). These results suggest that HT-29 cells endogenously express functional M3 receptor, whose activation led to a robust DMR signal. Third, we examined the origin of the acetylcholine-induced DMR in HT-29 cells using a microplate-based DMR assay in combination with various chemical probes. The main results are summarized in Figure 3. Blockage of phospholipase C with U73122 led to suppression, but not complete inhibition, of the acetylcholine DMR, indicating that its response contains a contribution from the G αq pathway. Permanent inhibition of G αi by ADP ribosylation of a Cys residue and decoupling of the G protein from the receptor with PTx only suppressed the early DMR event of acetylcholine; a similar result was obtained for

4 control, each antagonist alone was found to result in a negligible DMR response. Figure 2. DMR dose responses of acetylcholine in HT-29 cells. (a) Real-time DMR signals and (b) DMR amplitudes at 3, 8, and 50 min poststimulation [indicated by the dotted lines in (a)] as a function of acetylcholine doses. Data represent the mean ± the standard deviation (n = 12). Figure 4. Dose-dependent inhibition of the acetylcholine DMR by different antagonists under the co-stimulation condition in the microplate. The DMR amplitude at 50 min poststimulation of 4 μm acetylcholine was plotted as a function of antagonist doses. Data represent the mean ± the standard deviation from two independent measurements (n = 4). Figure 3. Cellular mechanisms of the acetylcholine DMR in HT-29 cells. Real-time DMR signals of 4 μm acetylcholine without any pretreatment (control) and after pretreatment with PTx (100 ng/ml, overnight), CTx (200 ng/ml, overnight), U73122 (25 μm, 1 h), and Y27632 (25 μm, 1 h). Data represent the mean ± the standard deviation from three independent measurements (n = 48). We further performed the dose responses of acetylcholine in the presence of an antagonist with a series of fixed doses under the co-stimulation condition in the microplate. Results showed that the presence of all four antagonists dose dependently caused a right shift in the acetylcholine dose response, with little impact on its maximal response (Figure 5a). Schild the pretreatment with the Rho kinase (ROCK) inhibitor Y27632, indicating that the early DMR originated from G αi and also involves ROCK activity. However, permanent activation of G αs by ADP ribosylation of an Arg residue with CTx increased the early response but suppressed the late response of acetylcholine. In consideration of the sensitization of the G αi pathway and the heterologous desensitization of G αq by the elevated basal level of camp induced by the CTx pretreatment, 30 the sensitivity of the acetylcholine to CTx treatment indicates that its early response is partly due to the G αi pathway, and its late response is partly due to the G αq pathway. Together, these results suggest that the acetylcholine DMR originated from multiple pathways, including G αq, G αi, and ROCK pathways. Fourth, we performed a microplate-based DMR costimulation assay to determine the dose-dependent inhibition of the acetylcholine DMR by an antagonist. The assay was performed in the microplate by adding the two ligands together, wherein the acetylcholine dose was fixed to be 4 μm, a dose close to its EC 80, to trigger DMR. Cellular responses at 50 min poststimulation were used to generate the dose inhibition curves. Results showed that all four antagonists dose dependently inhibited the acetylcholine DMR, leading to apparent EC 50 values of 3.64 ± 0.21, 0.83 ± 0.04, 0.54 ± 0.04, and 0.16 ± 0.01 nm (n = 4) for atropine, ipratropium, NMS, and tiotropium, respectively (Figure 4). This potency rank order was consistent with their known binding profile. 25 As a Figure 5. Schild analysis of M3 receptor antagonists under a costimulation condition in the microplate. (a) Acetylcholine DMR dose responses in the presence of ipratropium at different fixed doses. The DMR amplitudes at 50 min poststimulation were plotted. (b) Schild plots for the four antagonists. Data represent the mean ± the standard deviation from two independent measurements (n = 4). analysis suggests that the four antagonists exhibited apparent pk B values of 8.83, 9.99, 10.38, and 10.19, with slopes of 1.08, 1.0, 1.0, and 1.04 for atropine, ipratropium, NMS, and tiotropium, respectively (Figure 5b). Together, these results obtained under the co-stimulation condition suggest that all four antagonists were competitive with the binding of acetylcholine. Fifth, we determined time-dependent inhibition of the DMR of 4 μm acetylcholine by different antagonists, considering that different antagonists pose distinct on rates. This was done by preincubating the cells in the microplate with an antagonist for a specific period of time, followed by stimulation with acetylcholine. The results showed that different antagonists gave rise to distinct time-dependent inhibitions of the acetylcholine response (Figure 6). For atropine and ipratropium, their dose inhibition curves were clearly right shifted 8235

5 Figure 6. Time-dependent inhibition of the acetylcholine DMR by antagonists under the co-stimulation condition in the microplate. The DMR amplitude at 50 min poststimulation of 4 μm acetylcholine was plotted as a function of antagonist doses. The antagonists were used to pretreat the cells for a specific period of time (0, 30, or 60 min). Data represent the mean ± the standard deviation from two independent measurements (n = 4). as the preincubation time increased; similar but less pronounced was for that of NMS. Conversely, tiotropium exhibited the opposite trend: its potency to block the acetylcholine response increased as the preincubation time increased. Such a time-dependent inhibition curve is indicative of the unique binding kinetics of tiotropium relative to acetylcholine, compared with other M3 antagonists. Both slow on- and off-rates of tiotropium may contribute to this. Sixth, we compared the effects of the four antagonists on the acetylcholine dose response without and with a washing step between the antagonist and agonist stimulation steps in the microplate. To reach saturated fractional occupancy, each antagonist was assayed at a dose that was 100 times its respective equilibrium binding constant; the dose was 20, 20, 5, and 1 nm for atropine, ipratropium, NMS, and tiotropium, respectively. The K d value was reported to be 0.18, 0.16, 0.038, and nm for atropine, ipratropium, NMS, and tiotropium, respectively. 25 We chose 90 min to separate the two steps, considering the time required to reach equilibrium binding of antagonists and the effect of washing on the baseline of the biosensor response due to the mismatch in temperature between solutions as well as the sensitivity of cells to solution exchange in the microplate. Results showed that without the washing, the 90 min preincubation of the cells with atropine, ipratropium, or NMS all caused a marked right shift of the acetylcholine dose curve (Figure 7a), and all three antagonists markedly suppressed the early response, but not the late response, of acetylcholine (Figure 7b). Conversely, 1 nm tiotropium almost completely blocked the response of acetylcholine up to 1 mm (Figure 7, panels a and b). On the other hand, with the washing, the pretreatment with atropine, ipratropium, or NMS had little effect on both the potency and efficacy of acetylcholine, and tiotropium had little effect on the potency of acetylcholine but greatly suppressed its maximal response (Figure 7, panels c and d). It is worth noting that in this assay, the cells were preincubated with each antagonist for Figure 7. Effect of washing on the inhibition profiles of M3 receptor antagonists. (a) Acetylcholine DMR amplitudes at 50 min poststimulation as a function of antagonists, wherein the cells were continuously pretreated with the antagonists for 90 min. (b) Maximal acetylcholine DMR signals in the presence of distinct antagonists without washing. (c) Acetylcholine DMR amplitudes at 50 min poststimulation as a function of antagonists, wherein the cells were pretreated with the antagonists for 10 min and then washed five times with the assay buffer and further incubated for 75 min. (d) Maximal acetylcholine DMR signals with washing. For (b and d), the acetylcholine dose was 32, 128, 128, and 256 μm and 1 mm for the cells pretreated with the assay buffer (control), 20 nm atropine, 20 nm ipratropium, 5 nm NMS, and 1 nm tiotropium, respectively. Data represent the mean ± the standard deviation [n = 4 and 16 for (a and c) and (b and d), respectively]. 10 min, followed by washing for 3 min, and incubation with the assay buffer for 75 min. Upon the basis of their known binding characteristics obtained using radioligand binding assays, 25 it was estimated that at the dose examined, atropine, ipratropium, and NMS should reach maximal receptor occupancy, while tiotropium should only reach an 80% receptor occupancy with incubation for 10 min. However, the maximal acetylcholine response in the tiotropium-pretreated cells was much smaller than that predicted by operational theory (see discussion below). This may, at least in part, be due to the differences in binding characteristics of living cells versus membrane preparations and partly due to continuous binding and rebinding during the antagonist washout step. Nonetheless, these results suggest that atropine, ipratropium, and NMS were short-acting antagonists, and tiotropium exhibited long-acting antagonism. Seventh, we compared the acetylcholine DMR characteristics in the presence of distinct M3 antagonists. The data were obtained in the microplate. Results showed that acetylcholine at 4 μm did not trigger any clear DMR after pretreatment with 512 nm NMS for three different periods (Figure 8a). Similar results were observed for atropine and ipratropium. Conversely, 4 μm acetylcholine still resulted in a small positive DMR when the doses of tiotropium greatly exceeded its binding affinity under the co-stimulation condition (Figure 8b), and it led to a very small but detectable negative DMR when high doses of tiotropium were used to pretreat the cells for 60 min (Figure 8c). The pathway deconvolution study suggests that the small 8236

6 Figure 8. Effect of antagonists on the acetylcholine DMR characteristics. (a) Acetylcholine DMR signals after pretreatment with the buffer (control) and 512 nm NMS for different times (0, 30, and 60 min). (b and c) The DMR of 4 μm acetylcholine (b) cotreated or (c) pretreated with tiotropium at different doses for 60 min. (d) DMR induced by the co-stimulation with 4 μm acetylcholine and 16 nm tiotropium. The cells were without any pretreatment (control) or pretreated with PTx (100 ng/ml, overnight), CTx (200 ng/ml, overnight), U73122 (25 μm, 1 h), and Y27632 (25 μm, 1 h). Data represent the mean ± the standard deviation [n = 16 for (a and d), n = 4 for (b and c)]. positive DMR induced by co-stimulation of 4 μm acetylcholine and 16 nm tiotropium was mostly sensitive to the Rho kinase inhibitor Y27632 but less sensitive to CTx, PTx, or U73122 (Figure 8d). These results suggest that among the four antagonists tested, tiotropium is unique in its low effectiveness to block Rho kinase activity downstream of the activation of the endogenous M3 receptor. A recent molecular dynamic stimulation study suggests that both agonists and antagonists often follow the same well-defined, dominant pathway, binding to the β 1 - and β 2 -adrenergic receptors. 31 Since the receptor activator acetylcholine attains its binding equilibrium faster than tiotropium does, receptor activation by acetylcholine is initially possible after addition of both compounds until the antagonist draws up and fully blocks its signaling. The relatively low efficiency of tiotropium to block the acetylcholine-triggered rapid signaling event related to the Rho kinase activity may not be directly related to its clinical features, since tiotropium is applied as a prophylactic medicine. However, it is very likely that certain slow-binding antagonist drugs for other GPCRs may also behave similarly. Therefore, this phenomenon may still be clinically relevant. Lastly, we determined the residence time of the four antagonists at the endogenous M3 receptor in HT-29 cells using DMR assays under microfluidics. The use of microfluidics can minimize the rebinding related to mass transport during washing. The rebinding is believed to contribute to the longacting effects of certain ligands. 32,33 For DMR under microfluidics, the cells were first exposed to an antagonist for 10 min. The antagonist was then removed through perfusion with the assay buffer for 10 or 60 min. Afterward, the cellular response to 4 μm acetylcholine was recorded and used to calculate the drug target residence time. Results showed that the continuous buffer perfusion for 10 min almost completely prevented the inhibitory effect of atropine (Figure 9a) but only partially prevented the inhibitory effect of 100 nm ipratropium Figure 9. Effect of antagonists on the acetylcholine DMR characteristics under microfluidics. (a) Acetylcholine DMR signals after perfusion with the buffer (control), 1 nm atropine, or 100 nm atropine for 10 min, followed by perfusion with the buffer for 9 min. (b) Acetylcholine DMR signals after perfusion with the buffer (control) or 1 or 100 nm ipratropium for 10 min, followed by perfusion with the buffer for 9 min. (c) Acetylcholine DMR signals after perfusion with the buffer (control), 100 nm NMS, or 100 nm tiotropium for 10 min, followed by perfusion with the buffer for 10 min. (d) Acetylcholine DMR after perfusion with the buffer (control), 1 nm or 100 nm ipratropium for 10 min, followed by perfusion with the buffer for 60 min. The acetylcholine dose was 4 μm. The negative control (buffer) was obtained by continuous perfusion with the buffer. The perfusion rate was 2 μl/min for all. Data represent the mean ± the standard deviation (n = 4 for all). (Figure 9b), 100 nm NMS, or 100 nm tiotropium (Figure 9c). Conversely, perfusion with the buffer for 60 min completely prevented the inhibitory effect of ipratropium at 1 or 100 nm (Figure 9d). To estimate the drug residence time, we applied the Black and Leff operational theory to first determine the efficacy of acetylcholine to trigger the DMR signal and then to stimulate the acetylcholine DMR dose responses as a function of spare receptors. This would provide a basis to calculate the percentage of antagonist occupancy, so it is possible to estimate the drug residence time through comparison with the experimental data obtained using DMR under microfluidics. Considering the buffer perfusion time-dependent recovery of acetylcholine responses, we used the acetylcholine responses after perfusion with the antagonist at 100 nm for 10 min, followed by the buffer perfusion for 10 min, as the basis to calculate the drug target residence time. To do so, we first determine the efficacy of acetylcholine using the Black and Leff model of operational agonism. 34 It defines agonist [A] response as 35 [A] τemax response = [A]( τ + 1) + KA (1) where [A] is the agonist concentration; E max, the maximal response of acetylcholine in HT-29 cells; K A, the equilibrium dissociation constant of acetylcholine; τ, anefficacy term equal to the ratio of the receptor density to K e ; and K e, the efficiency of signal transduction by the ligand receptor complex. The K A value of acetylcholine binding to the endogenous M3 receptor in HT-29 was unknown. Thus, we used a K A value of 4200 nm 8237

7 for acetylcholine, which was determined on the basis of a competitive radiobinding assay. 36 We fit the dose response with the operational mode leading to an estimation of the τ value for acetylcholine, which was found to be 24.3 ± 1.9 (six independent measurements, each in duplicate). The dose response as a function of the receptor level was then simulated using the operational model (Figure 10a). The Figure 10. Drug target residence time determination. (a) Simulated sensitivity of acetylcholine to the spare receptors (in percentage). (b) Relationship of the acetylcholine DMR in the antagonist treatment duration-controlled cells with their off- and on-rates. reduced spare receptors were used to mimic the blockage by the preoccupied antagonists. Results showed that the apparent potency and efficacy of acetylcholine decreased as the total number of receptors decreased. On the basis of the acetylcholine DMR amplitude at 50 min poststimulation in the cells preexposed to distinct antagonists, each at 100 nm for a controlled duration, we first estimated the percentage of the receptor occupancy by each antagonist. We then calculated their off-rate and residence time at the endogenous M3 receptor. The offrate was estimated to be 0.066, 0.022, , and min 1 and the residence time 15, 46, 149, and 1365 min for atropine, ipratropium, NMS, and tiotropium, respectively. The off-rate was within 2 4 times the values reported in the literature. 25 Interestingly, the percentage of the acetylcholine DMR signal in the cells pre-exposed to 100 nm antagonists for the controlled duration, obtained after being normalized to the positive control, was found to be in the general log linear relationship with the off-rate, but not the on-rate, of the four antagonists (Figure 10b). This result is largely in agreement with previous findings that the duration of the action of these M3 antagonists observed in vivo is mostly determined by their off-rates, rather than their on-rates. 4,13,14 CONCLUSIONS In conclusion, DMR assays under both persistent and durationcontrolled stimulation conditions have been used to study the kinetics-dependent blockage of the signaling of the endogenous M3 receptor in native HT-29 cells by four antagonists. The acetylcholine-activated M3 receptor was found to trigger multiple signaling pathways. Among the four antagonists examined, tiotropium displayed kinetics-dependent selectivity to block the M3 receptor signaling. Tiotropium was found to be less effective at blocking the early DMR event than at blocking the late DMR event of acetylcholine, when it was used to costimulate the receptor with acetylcholine. The early DMR event of acetylcholine was mostly sensitive to the Rho kinase activity. Tiotropium also gave rise to a left-shifted potency to block the late acetylcholine DMR as the preincubation time increased, a trend that was opposite to those of the other three antagonists. DMR assays under microfluidics allowed us to estimate the residence time of these antagonists at the endogenous receptors in native cells. This study highlights the power of DMR assays to differentiate biochemical mechanisms of action of drugs in live cells. AUTHOR INFORMATION Corresponding Author * fangy2@corning.com. Tel: Author Contributions H.D. performed most DMR assays in the microplate and analyzed the data. C.W. performed all DMR assays under microfluidics and a portion of DMR assays in the microplate and analyzed the data. Y.F. and M.S. conceived of the idea. Y.F. designed the experiments, analyzed all data, and wrote the manuscript. All authors have given approval for the final version of the manuscript. H.D. and C.W. contributed equally to this work. Notes The authors declare the following competing financial interest(s): Y.F. and H.D. are employees and stockholders of Corning Inc. Epic is a Corning commercial product. REFERENCES (1) Gleeson, M.; Hersey, A.; Montanari, D.; Overington, J. Nat. Rev. Drug Discovery 2011, 10, (2) Pammolli, F.; Magazzini, L.; Riccaboni, M. Nat. Rev. Drug Discovery 2011, 10, (3) Swinney, D. C. Nat. Rev. Drug Discovery 2004, 3, (4) Copeland, R. A.; Pompliano, D. L.; Meek, T. D. Nat. Rev. Drug Discovery 2006, 5, (5) Tummino, P. J.; Copeland, R. A. Biochemistry 2008, 47, (6) Zhang, R.; Monsma, F. Curr. Opin. Drug Discovery Dev. 2009, 12, (7) Copeland, R. A. Expert Opin. Drug Discovery 2010, 5, (8) Lu, H.; Tonge, P. J. Curr. Opin. Chem. Biol. 2010, 14, (9) Zhang, R.; Monsma, F. Expert Opin. Drug Discovery 2010, 5, (10) Keighley, W. Drug Discovery World 2011, 12 (3), (11) Vauquelin, G. MedChemComm 2012, 3, (12) Singh, J.; Petter, R. C.; Baillie, T. A.; Whitty, A. Nat. Rev. Drug Discovery 2011, 10, (13) Disse, B.; Speck, G. A.; Rominger, K. L.; Witek, T. J., Jr.; Hammer, R. Life Sci. 1999, 64, (14) Barnes, P. J. Chest 2000, 117, 63S 66S. (15) Fang, Y.; Ferrie, A. M.; Fontaine, N. H.; Yuen, P. K. Anal. Chem. 2005, 77, (16) Kenakin, T. Nat. Rev. Drug Discovery 2009, 8, (17) Verrier, F.; An, S.; Ferrie, A. M.; Sun, H.; Kyoung, M.; Fang, Y.; Benkovic, S. J. Nat. Chem. Biol. 2011, 7, (18) Fang, Y. Expert Opin. Drug Discovery 2011, 6, (19) Goral, V.; Jin, Y.; Sun, H.; Ferrie, A. M.; Wu, Q.; Fang, Y. PLoS One 2011, 6, e (20) Goral, V.; Wu, Q.; Sun, H.; Fang, Y. FEBS Lett. 2011, 585, (21) Zaytseva, N.; Miller, W.; Goral, V.; Hepburn, J.; Fang, Y. Appl. Phys. Lett. 2011, 96, (22) Ferrie, A. M.; Wu, Q.; Fang, Y. Appl. Phys. Lett. 2010, 97, (23) Fang, Y.; Ferrie, A. M.; Fontaine, N. H.; Mauro, J.; Balakrishnan, J. Biophys. J. 2006, 91, (24) Fang, Y. Sensors 2007, 7, (25) Dowling, M. R.; Charlton, S. J. Br. J. Pharmacol. 2006, 148, (26) Kopp, R.; Lambrecht, G.; Mutschler, E.; Moser, U.; Tacke, R.; Pfeiffer, A. Eur. J. Pharmacol. 1989, 172,

8 (27) Koenig, J. A.; Edwardson, J. M. Mol. Pharmacol. 1996, 49, (28) Park, S. E.; Kim, N. D.; Yoo, Y. H. Cancer Res. 2004, 64, (29) Pettersson, A.; Nilsson, L.; Nylund, G.; Khorram-Manesh, A.; Nordgren, S.; Delbro, D. S. Eur. J. Pharmacol. 2009, 609, (30) Tran, E.; Fang, Y. J. Recept. Signal Transduction Res. 2009, 29, (31) Dror, R. O.; Pan, A. C.; Arlow, D. H.; Borhani, D. W.; Maragakis, P.; Shan, Y.; Xu, H.; Shaw, D. E. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, (32) Vauquelin, G. Expert Opin. Drug Discovery 2010, 5, (33) Vauquelin, G.; Charlton, S. J. Br. J. Pharmacol. 2010, 161, (34) Black, J. W.; Leff, P. Proc. R. Soc. London, Ser. B 1983, 220, (35) Kenakin, T. ACS Chem. Biol. 2009, 4, (36) Cheng, K.; Khurana, S.; Chen, Y.; Kennedy, R. H.; Zimniak, P.; Raufman, J. P. J. Pharmacol. Exp. Ther. 2002, 303,