Anal. Chem. 2000, 72, 3109-3115
Equilibrium and Kinetic Measurements of Muscarinic Receptor Antagonism on Living Cells Using Bead Injection Spectroscopy Peter S. Hodder,*,† Craig Beeson, and Jaromir Ruzicka
Department of Chemistry, University of Washington, Seattle, Box 351700, Washington 98195-1700
Bead injection spectroscopy (BIS) techniques are introduced for automated measurement of pharmacological antagonism by functional assay. Chinese hamster ovary cells that express the rat type 1 muscarinic receptor are cultured on microbeads and used as a renewable biological target for muscarinic receptor antagonist ligands. A flow injection instrument is used to reproducibly sample and capture the cells in a jet ring chamber. The effect of the antagonist pirenzepine on the carbachol-induced intracellular calcium response of the cells is measured with a fluorescence microscope photometry system. The BIS functional assay is used to quantify both equilibrium and kinetic pharmacological values for pirenzepine. In addition, two muscarinic receptor antagonists (pirenzepine and atropine) are assayed to compare their relative efficacy at diminishing the calcium response. Due to the precision of the automated fluid/bead handling protocols, and reproducibility of the measured calcium response, the quantification of useful pharmacological information from living cells by BIS techniques is demonstrated. Radioligand Binding and Functional Assays. An essential goal of pharmacological research is the elucidation of the interaction between a ligand (protein, low molecular weight molecule, DNA, RNA) and a target receptor. Radioligand binding (RLB) techniques are often used to determine equilibrium and kinetic binding constants for the receptor-ligand interaction. In a typical RLB assay, the ligand is labeled with a radioactive species equilibrated with a purified receptor, and radiometric measurements are performed to determine the amount of ligand bound to the receptor. Since a receptor-ligand interaction is governed by the law of mass action, a well-established set of analytical methods are available to quantitate both equilibrium and kinetic binding constants, as well as a variety of other parameters of pharmacological interest.1,2 With the advent of molecular and cell biology techniques, cellbased functional assays now complement RLB assays.3 In a typical functional assay, a ligand is equilibrated with cells that express a receptor of interest. If the ligand binds the receptor and activates † Present address: Merck & Co., 502 Louise Lane, North Wales, PA, 19454. (1) Sweetnam, P. M.; Caldwell, L.; Lancaster, J.; Bauer, C., Jr.; McMillan, B.; Kinnier, W. J.; Price, C. H. J. Nat. Prod. 1993, 56, 441-55. (2) Hulme, E. C. Receptor-ligand interactions: a practical approach; IRL Press at Oxford University Press: Oxford, England, New York, 1992. (3) Hertzberg, R. P. Curr. Opin. Biotechnol. 1993, 4, 80-4.
10.1021/ac991231v CCC: $19.00 Published on Web 06/20/2000
© 2000 American Chemical Society
a specific cellular biochemical reaction, the biological response can be measured through a variety of detection methods.4,5 By quantifying to what degree a receptor-ligand interaction serves to invoke (agonize) or diminish (antagonize) a pharmacological response, pharmacological parameters similar to those from RLB assays can be calculated.6 By using a living cell as a biosensor, functional assays have distinct advantages over RLB assays. Since functional assays can employ receptors that are in an unpurified native state, data obtained by functional assay have more physiological relevance: a biological response occurs only when a receptor is bound and activated. This contrasts to many RLB assays, where purification of the receptor is necessary, and a ligand’s functional activity cannot be directly determined. Also, typical RLB assays require additional steps to separate receptor-ligand complexes from free ligand and to measure nonspecific binding of the ligand to other materials in the assay vial (for a notable exception, see ref 7). Despite the advantages of functional assays over RLB assays, functional assays have unique technical challenges that stem from the complexity of living organisms. Since the sensor is living biological material, careful attention must be paid to its physiology if useful results are desired. In addition, many biological responses are transient, which requires precise coordination between delivery of ligand and measurement of response. Also important to consider is that the biological response is separated from the receptor-ligand interaction by several biochemical reactions; therefore, more complex methods of data analysis are necessary to extract quantitative pharmacological information from assay data.6 Since the automation of cell-based functional assays may remove some of these challenges, research has been focused on the development of sophisticated analytical instrumentation.8-10 BIS Techniques. Bead injection spectroscopy (BIS) is based upon the use of flow injection (FI) techniques to capture a solid (4) Haugland, R. P. Handbook of Fluorescent Probes and Research Chemicals, 4th ed.; Molecular Probes: Eugene, OR, 1996. (5) Baxter, G. T.; Bousse, L. J.; Dawes, T. D.; Libby, J. M.; Modlin, D. N.; Owicki, J. C.; Parce, J. W. Clin. Chem. 1994, 40, 1800-4. (6) Kenakin, T. P. Pharmacologic analysis of drug-receptor interaction, 3rd ed.; Lippincott-Raven: Philadelphia, 1997. (7) Major, J. S. J. Recept. Signal Transduction Res. 1995, 15, 595-607. (8) Melamed, M. R.; Lindmo, T.; Mendelsohn, M. L. Flow Cytometry and Sorting, 2nd ed.; Wiley-Liss: New York, 1990. (9) Coward, P. C.; S. D. H.; Wada, H. G.; Humphries, G. M.; Conklin, B. R. Anal. Biochem. 1999, 270, 242-8. (10) McConnell, H. M.; Owicki, J. C.; Parce, J. W.; Miller, D. L.; Baxter, G. T.; Wada, H. G.; Pitchford, S. Science 1992, 257, 1906-12.
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000 3109
phase in a specialized detector flow cell (e.g., a jet ring (JR) chamber), perfuse the captured solid phase with sample or reagent, and dispose of all material after assay.11 BIS techniques have demonstrated their utility at performing automated cell-based functional assays.12,13 Despite the success at this task, the BIS research published so far does not completely demonstrate its general applicability to cell-based functional assays. This is because one of the two applications of a functional assay has yet to be investigated viz., the measurement of pharmacological antagonism. Therefore, the aim of the research presented here was to develop and execute BIS functional assays that could quantitatively measure the equilibrium and kinetic binding properties of pirenzepine, a muscarinic receptor antagonist. In addition, the use of BIS to discriminate the muscarinic receptor antagonists pirenzepine and atropine was investigated. The reproducibility of the BIS technique at repetitively measuring biological responses provides the means necessary for extracting useful quantitative pharmacological information by functional assay. Also, the FI protocols developed allow precise temporal control of agonist and antagonist concentrations in the JR chamber, a feature that is critical for extraction of useful kinetic rate constants from functional assay data. EXPERIMENTAL SECTION Apparatus. The BIS instrument and photometry system used to measure functional antagonism has been previously described.12 Briefly, a FIAlab 3000 instrument (Alitea Instruments, Medina, WA) consisting of a syringe pump, peristaltic pump, and multiport valve was connected to a jet ring chamber. The syringe pump was used to sample and deliver both drugs and suspended biological material to the JR chamber. The peristaltic pump was used to evacuate the JR chamber contents. The JR chamber was affixed to the stage of an inverted epifluorescence microscope (Axiovert 100, Zeiss, Oberkochen, FRG). Fluorescence measurements were made possible by a dual-monochromator photometry system (PTI, South Brunswick, NJ). For this research, the JR chamber was modified so that a larger inner diameter (1.0 mm) stainless steel tube was used to trap cell-attached beads (Upchurch Scientific, Oak Harbor, WA). Dose-response and binding curve analysis was done by nonlinear regression statistics with a commercial software package (GraphPad, San Diego, CA). Other regression calculations were done with spreadsheet software. Materials and Reagents. The preparation of materials, reagents, and cells has been described in detail previously.12 All materials and reagents were acquired from Sigma (St. Louis, MO) unless otherwise specified below. To prevent clumping of cells, Pluronic F-68 was added to the assay buffer to a final concentration of 0.8% w/v. No significant differences were apparent in functional assay data due to the addition of this reagent. Chinese hmster ovary (CHO) cells transfected with the rat type 1 muscarinic receptor (m1) (CHO-m1-WT3, CRL-1985, American Type Tissue Culture Collection, Rockville, MD) were first cultured to confluence on Cytodex-2 microcarriers (Amersham-Pharmacia Biotech, Piscataway, NJ) using sterile tissue culture technique. Approximately 0.5 h prior to functional assay, cells were loaded with (11) Ruzicka, J.; Scampavia, L. Anal. Chem. 1999, 71, 257A-63A. (12) Hodder, P. S.; Ruzicka, J. Anal. Chem. 1999, 71, 1160-6. (13) Lahdesmaki, I., Ruzicka, J. Anal. Chem., 1999. 71. 5248-52.
3110
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
12 µM fura-2 AM (Molecular Probes, Eugene, OR). All experiments were conducted at room temperature. For KB determination, loading of cells with fura-2 took place in the presence of a specified concentration of pirenzepine (see below). Cells were then washed and suspended in assay buffer containing pirenzepine and aliquotted to a miniature roller flask that was connected via stainless steel tubing to the FIAlab instrument’s multiport valve. Standards consisting of mixtures of carbachol and pirenzepine were made in assay buffer as required. For the efficacy ranking experiments, standards containing mixtures of atropine (RBI, Natick, MA) and carbachol or pirenzepine and carbachol were prepared in a similar fashion. BODIPYFL, a green fluorescent pirenzepine analogue, was used for some tracer experiments (Molecular Probes). To measure cell viability, a calcein AM and ethidium homodimer-1 fluorescence-based assay was used (Molecular Probes). RESULTS AND DISCUSSION Equilibrium Measurement of Competitive Antagonism. Similar to classifying the strength of an weak acid by its Ka, pharmacologists rank the affinity of different antagonists by measuring their equilibrium dissociation constant (KB). To determine the KB of the competitive antagonist pirenzepine (pzp) with the CHO-m1 cells, the Schild regression was employed.6,14,15 The equation of the Schild regression is a line of unit slope, where the dose ratio (dr) is the ratio of the agonist concentrations that invoke the same biological response in the presence or absence of antagonist:
log(dr - 1) ) log[pzp] - log KB
(1)
If the slope of the regression is not significantly different from unity, then the y-intercept of the regression is the pKB of the antagonist, i.e., the negative logarithm of its equilibrium dissociation constant. Prior to Schild regression calculation, data for doseresponse curves are first collected in the absence of antagonist and then repeated in the presence of different concentrations of antagonist. For each dose-response curve, the effective concentration that produces 50% of the maximal response (EC50) is calculated, and EC50 dose ratios are determined. However, the requirement of equilibrium conditions for Schild analysis can complicate the determination of antagonist equilibrium dissociation constants by functional assay.6,15 Equilibrium Assay Development. Since calculation of Schild regression requires the collection of a considerable amount of calcium responses, cell viability was of some concern during the BIS assay. With the CHO-m1-WT3 cells, good (>90%) cell viability was observed as long as 6 h after loading with fura-2 AM. However, significant decreases in calcium responses were apparent in the assay data ∼3 h after loading. To eliminate artifact from the data and access higher sample throughput, a preliminary goal of the antagonist functional assay research was the development a more rapid BIS protocol. BIS protocols were developed that optimized both the throughput of the assay and assay flow rates for antagonist equilibration (14) Arunlakshana, O.; Schild, H. O. Br. J. Pharmacol. 1997, 120, 151-61; discussion 148-50. (15) Kenakin, T. P. Eur. J. Pharmacol. 1980, 66, 295-306.
Table 1. FIAlab Software Protocol, 1-mm-i.d. Tubing BIS functional assay step microbead delivery 1 2 3 microbead perfusion 1 2 microbead ejection 1 2 3 4
explanation of step
vol (µL)
flow rate (µL/s)
aspirate assay buffer aspirate microbead suspension deliver microbeads to JR chamber
700 250 950
100 50 9
aspirate drug mixture deliver mixture to JR chamber
300 300
75 4
1000 1000
100 175
aspirate assay buffer turn JR chamber solenoid on deliver assay buffer to JR chamber turn JR chamber solenoid off
with the receptor population. An important improvement to the BIS apparatus involved replacing the 0.8 -mm-i.d. stainless steel tubing in the JR chamber with 1-mm-i.d. tubing. From previous experiments, it was estimated that ∼300 beads (2 × 104-3 × 104 cells) were trapped in the JR chamber, forming a small column of ∼2 mm in height.12 Since the back pressure behind a bead column is directly proportional to column height,16 the larger diameter tubing allowed faster flow rates without increasing the amount of column loss due to excessive back pressure (Table 1). The higher reagent and sample delivery flow rates increased the throughput from 4 to 11 assays/h. The time necessary to collect enough data for a single dose-response curve (eight curve points, each point measured in triplicate for statistical purposes) decreased to ∼2 h, allowing the same batch of cells to be used for each curve. Another benefit of the larger inner diameter tube was that more beads (and therefore cells) were now in the microscope objective’s plane of focus, improving the signal-to-noise ratio in fluorescence measurements. In this way, the improved BIS assay protocol allowed rapid, reproducible measurements of calcium responses without sacrificing assay sensitivity or bead column stability. Equilibrium Assay Results. Previous automated functional assays with CHO-m1-WT3 cells consisted of simultaneously injecting the appropriate concentrations of pirenzepine and carbachol to the cells and measuring a modulated extracellular acidification response.17 However, analysis of this assay data yielded Schild regressions with slopes significantly larger than unity, which suggested that equilibrium had not been reached between pirenzepine and the muscarinic receptor population at the time of response measurement.15 Using published RLB assay data, it was determined that incubation times of 1-5 min were necessary for concentrations of 1-10 µM pirenzepine to equilibrate with the muscarinic receptor population.18 To eliminate this nonequilibrium artifact from BIS functional assay data, cells were loaded with fura-2 and handled thereafter in assay buffer that contained the appropriate concentration of pirenzepine for Schild analysis (0, 1, 3, or 10 µM pirenzepine). Since the time elapsed from fura-2 loading to functional assay was 0.5 h or longer, it was assumed that the antagonist was fully equilibrated with the receptor prior to calcium response measurement. (16) Willard, H. H.; Merritt, L. L.; Dean, J. A.; Settle, F. A. Instrumental Methods of Analysis, 7th ed.; Wadsworth Publishing Co.: Belmont, CA, 1988. (17) Baxter, G. T.; Young, M. L.; Miller, D. L.; Owicki, J. C. Life Sci. 1994, 55, 573-83. (18) Wang, J. X.; Mei, L.; Yamamura, H. I.; Roeske, W. R. J. Pharmacol. Exp. Ther. 1987, 242, 981-90.
Figure 1. (a) BIS equilibrium functional assay results. Calcium responses from carbachol (cch) stimulations in the presence of 3 µM pirenzepine (pzp), performed in triplicate. The data from this assay compose one of the dose-response curves in (b). FIAlab software protocol details in Table 1. (b) Dose-response curves: equilibrium functional assay data. Series of dose-response curves constructed for subsequent Schild regression. Note that the EC50 of doseresponse curves shifts to higher concentrations when the functional assay is conducted in the presence of higher antagonist concentrations.
The effect of the equilibration protocol was immediately apparent in BIS functional assay data (Figure 1a). As expected, dose-response curves constructed from the data showed EC50 shifts that increased with increasing antagonist concentration (Figure 1b). Schild regression of the dose ratio data yielded a slope not significantly different from unity (0.97 ( 0.21). Statistical analysis of the resulting y-axis intercept yielded a pKB of 7.7 ( Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
3111
1.4 (95% confidence limits). No pirenzepine pKB data based on Schild analysis were found in the published literature. However, the assay results compared favorably to a literature value of pKB (7.6) obtained from radioligand binding assays performed with membrane preparations purified from CHO-m1-WT3 cells.19 Kinetic Measurement of Competitive Antagonism. An expression that relates biological response to an antagonist’s apparent kinetic rate constant can be derived from the law of mass action.20 Dose ratios similar to those defined in eq 1 are written to express the concentration of receptors occupied by antagonist as a function of time:
ln
(
)
DReq - DRt ) -kappt DReq - DRo
(2)
Equation 2 is the equation of a straight line, where the negative of the slope is kapp, the apparent rate constant for the association of antagonist with receptor. The value DReq is the dose ratio of responses prior to antagonist addition and after the antagonist has come to equilibrium with the receptor, DRt is the dose ratio after incubation time t, and DRo is the dose ratio at time zero. To perform a kinetic analysis, a functional assay is executed with a full agonist, and a dose-response curve is constructed from the resulting data. Additional functional assays are then executed by first exposing the biological material to a specified concentration of antagonist for a certain time interval, followed by exposure to the full agonist in the presence of antagonist. The resulting diminished biological responses are then measured. Treatment of assay data is initiated by transforming the control dose-response curve into a straight line via conversion into probits, performing a linear regression, and calculating an EC50.21 Decay data are added to the probit plot, and lines parallel to the original dose-response data are drawn through each decay datum. The expected EC50 (probit ) 5) for each decay datum is then extrapolated from the graph. Dose ratios are calculated by dividing the EC50 probit values at any time t (i.e., in the presence of antagonist) by the EC50 probit value at time zero (i.e., in the absence of antagonist). The logarithm of the dose ratio term is graphed versus the time of incubation, and the kapp is calculated by linear regression. Kinetic Assay Development. For extraction of useful kapp values, accurate determination of antagonist/receptor incubation intervals was necessary. BIS techniques were therefore optimized to allow precise temporal control of the agonist and antagonist concentrations. In classical FI theory, increasing the flow rate of an FI assay makes an injected sample more rapidly reach its maximum concentration at the detector flow chamber .22 To establish the start time of the incubation interval, a fast-flow rate antagonist injection scheme was developed that produced a steep gradient of drug introduction to the JR chamber. However, as described above, extremely fast flow rates resulted in high bead column back pressure, which ejected beads out of the JR chamber (19) Buck, M. A.; Fraser, C. M. Biochem. Biophys. Res. Commun. 1990, 173, 666-72. (20) Ruffolo, R. R., Jr.; Patil, P. N. Blood Vessels 1979, 16, 135-43. (21) Goldstein, A. Biostatistics, an introductory text; Macmillan: New York, 1964. (22) Ruzicka, J.; Hansen, E. H. Flow injection analysis, 2nd ed.; J. Wiley: New York, 1988.
3112
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
prior to calcium response measurement. The fastest flow rate to deliver sample to the flow chamber without bead column loss becoming significant was found to be 9 µL/s. As expected, tracer curve analysis with a fluorescent pirenzepine analogue proved that a very steep gradient of antagonist was delivered to the cells in the JR chamber. Therefore, the time of antagonist injection was used as the starting time of incubation. The minimum amount of incubation time was set by the time required for the syringe to completely deliver the entire antagonist volume to the JR chamber. To increase the incubation time, the syringe pump was stopped for a specified time interval after delivery. An expression was written to quantify the time of incubation for this step:
tinc,sec ) 60 + tpause
(3)
where the total time of incubation (tinc) represents the sum of the time elapsed for delivery of antagonist to the JR chamber (60 s) plus the time the FIAlab protocol was on pause (tpause) prior to agonist delivery. Any additional incubation time that elapsed between the injection of carbachol and the measurement of the calcium response was also determined. To eliminate artifact that would result from nonequilibrium conditions between pirenzepine and the receptor population, the injected sample consisted of a mixture of carbachol and pirenzepine. Experiments showed that 25 s elapsed after injection of carbachol prior to induction of the calcium response. Therefore, the total incubation time prior to the measurement of the calcium response was the sum of tinc and 25 s. For the case where the carbachol/pirenzepine mixture was injected without any pirenzepine incubation step, a marked decrease in the calcium response was still observed. This signified that a fraction of the receptors had already been bound to pirenzepine prior to the induction of the calcium response. Since this calcium response took a substantial period of time to reach its peak (∼20 s), the incubation time for this measurement was declared to be the amount of time it took the response to reach half of its maximum, 15 s. The FIAlab instrument was modified so that sample tubes of antagonist, agonist/antagonist mixtures, and assay buffer were all connected via stainless steel tubing to its multiport valve (Figure 2). The individual FI protocol steps are shown in Table 2. During preliminary assays, it was found that high concentrations of pirenzepine would rapidly extinguish the calcium response, prohibiting the collection of enough decay data points for further pharmacological analysis. In contrast, the use of lower concentrations of antagonist required increased incubation times with the cells in order to measure a significant diminishment of the calcium response. Therefore, the time required to assay each sample would increase proportionally and decrease sample throughput. A concentration of 125 nM pirenzepine resulted in complete extinction after an incubation time of 5 min; this concentration was chosen for all further kinetic experiments. Kinetic Assay Results. To perform kinetic assay for a single incubation time, five drug samples were assayed. After trapping a fresh set of cells in the JR chamber, the first sample assayed was 1 µM carbachol, which was known from previous measurements to give a calcium response of probit ) 5. This calcium
Table 2. FIAlab Software Protocol Steps for BIS Kinetic Functional Assay BIS functional assay step microbead delivery 1 2 3 antagonist incubation 1 2 3 4 microbead perfusion 1 2 microbead ejection 1 2 3 4
explanation of step
vol (µL)
flow rate (µL/s)
aspirate assay buffer aspirate microbead suspension deliver microbeads to JR chamber
650 225 875
100 50 9
aspirate antagonist (or control) deliver to JR chamber
400 400
75 9
aspirate antagonist/agonist mixture deliver to JR chamber
500 500
75 5
aspirate assay buffer turn JR chamber solenoid on deliver assay buffer to JR chamber turn JR chamber solenoid off
1000
100
1000
175
delay (if necessary) t seconds
Figure 2. Schematic of the BIS kinetic functional assay apparatus. Apparatus is similar to that used for equilibrium functional assays. Note that all samples necessary for the entire assay are on the multiport valve.
response was used to calibrate the three next measured calcium responses. A fresh set of cells was then trapped and incubated with 125 nM pirenzepine or assay buffer containing no antagonist for a specified of time (60, 120, or 180 s). This was followed by delivery of the second sample, a mixture of 125 nM pirenzepine and 1 µM carbachol. After the diminished calcium response was measured, beads were flushed to waste, and the assay was repeated. Three iterations of cell capture, pirenzepine (or assay buffer) incubation, pirenzepine/carbachol mixture injection, and cell disposal were performed for statistical purposes. Finally, a fresh set of cells were trapped and again exposed to 1 µM carbachol for calibration purposes. Typical assay data for two different incubation times is shown in Figure 3. As can be seen in the figure, a reproducible decrease in the calcium response was measured, which was proportional to incubation time with the antagonist. To analyze the BIS kinetic assay data, a probit plot was constructed (Figure 4). The probit conversion of the carbachol dose-response curve with no pirenzepine present allowed a linear regression to be calculated with good fit (r2 ) 0.980). For each incubation time, the three decay data were averaged and normalized to the two 1 µM carbachol stimulations and then added to
Figure 3. (a) BIS kinetic functional assay data. First and last calcium responses result from 1 µM carbachol stimulation, with no prior incubation with pirenzepine. Other calcium responses result from simultaneous injection of 1 µM carbachol and 125 nM pirenzepine (15-s incubation). FIAlab software protocol described in Table 2. (b) BIS kinetic functional assay data. Calcium responses resulting from 1 µM carbachol stimulation, after an 80-s incubation with 125 nM pirenzepine. Stimulations performed in triplicate. First and last calcium responses result from 1 µM carbachol stimulation, with no prior pirenzepine incubation.
the probit plot. Lines parallel to the dose-response curve were graphed through each decay data point, and the responses at probit ) 5 extrapolated. Once the dose ratios were calculated, a plot of logarithmic dose ratios versus time was constructed. Regression analysis yielded a value for kapp ) 4.6 × 10-3 s-1. Standard error of the slope was calculated to be 2 × 10-4 s-1. Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
3113
Figure 4. Probit plot. BIS kinetic assay data. Dose-response curve for carbachol at far left (control), fit to straight line. Decay data points for pirenzepine incubation times of 15, 85, 145, and 205 s plotted and parallels drawn to carbachol dose-response curve. EC50 values for dose ratios come from extrapolated values at probit ) 5.
There are very few examples in the published literature where functional assays are used to determine association or dissociation rate constants for ligand-receptor interactions and none that determine binding constants for muscarinic receptor antagonists.6,20 This is most likely due to the fact that kinetic functional assays require precise temporal control of agonist or antagonist addition during the assay, which is difficult to achieve with pipetbased or manual techniques. Instead, the kapp of the BIS functional assay was compared to RLB assays conducted on purified receptor preparations. The concentration of pirenzepine used in the BIS assay and published values for k1 ad k2 were substituted into the law of mass action, and a kapp was calculated. Although published values manifest significant variability (between 10-5 and 10-2 s-1), the kapp from the BIS functional assay falls within the expected range.18,23,24 Discrimination of Muscarinic Receptor Antagonists. Assay data from previous BIS experiments gave greater insight into the calcium response elicited by muscarinic receptor agonists, showing significant differences in the kinetics of the elicited calcium response.12 Therefore, a BIS technique was developed to investigate the effect that muscarinic receptor antagonists had on the kinetics of the calcium response. Volumes of antagonist and agonist were aliquotted to sample vials, such that each vial had a different concentration of antagonist (atropine, 0.01, 0.1 and 1 µM; pirenzepine, 0.1, 1, 10, 100 µM), and the same concentration of agonist (carbachol, 10 µM). Similar to the functional assay protocol described previously,12 trapped cells were exposed to the different agonist/antagonist mixtures, and the calcium responses were recorded. As expected, assay data revealed that increasing the concentration of injected antagonist had the effect of decreasing the calcium response (Figure 5). In addition, the relative efficacies of the two antagonists at diminishing the calcium response was readily apparent: approximately 10-100 times higher concentrations of pirenzepine than atropine were required to extinguish the carbachol-induced calcium response. These values compare well with rankings based on RLB assays.19 More interestingly, closer inspection of the BIS peaks revealed that increasing the concentration of pirenzepine had the effect of changing the shape of the (23) Lin, S. C.; Olson, K. C.; Okazaki, H.; Richelson, E. J. Neurochem. 1986, 46, 274-9. (24) Birdsall, N. J.; Chan, S. C.; Eveleigh, P.; Hulme, E. C.; Miller, K. W. Trends Pharmacol. Sci. 1989, (Suppl.), 31-4.
3114 Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
Figure 5. Ranking of functional antagonism by BIS techniques. Different concentrations of the muscarinic receptor antagonist pzp or atropine (atr) simultaneously injected with 10 µM cch. Mixtures injected at 1 µL/s flow rates (FIAlab protocol steps similar to those in ref 12. Exact mixtures labeled on curves. Data show that higher concentrations of pirenzepine are required to diminish the calcium response than atropine.
calcium response. For example, the decay from the maximum of the pirenzepine-antagonized calcium response is more pronounced at higher concentrations. However, increasing the concentration of antagonist does not cause an increase in latency of the calcium response. Although we are unable to explain the exact nature of these phenomena, they could signify a more complex interaction between the antagonists and the cell-signaling pathways responsible for eliciting the calcium response or perhaps suggest the ability of the antagonists to inhibit the biphasic calcium response.25,26 Similar to the unique kinetic effects that were observed in BIS assays with muscarinic receptor agonists, this type of BIS functional assay should facilitate the use of novel methods to elucidate the complexities of intracellular signaling mechanisms.27 CONCLUSION Previous experiments have shown that BIS techniques can be used to rank different muscarinic receptor agonists by measuring how effectively they invoke a calcium response. This paper demonstrates the converse; viz., BIS techniques can also be used to rank different muscarinic receptor antagonists based upon how effectively they extinguish an agonist-induced calcium response. With the successful completion of both equilibrium and kinetic measurements of functional antagonism, the BIS functional assay methodology has been proven to be useful for the automation of fluorescence-based functional assays. The improvements made to the FI instrument and the FIAlab software protocols have allowed measurement of functional responses with faster sample throughput than the previous BIS functional agonism measurements. As (25) Lambert, D. G.; Burford, N. T.; Nahorski, S. R. Biochem. Soc. Trans. 1992, 20, 130-5. (26) Lauffenburger, D. A.; Linderman, J. J. Receptors: models for binding, trafficking, and signaling; Oxford University Press: New York, 1993. (27) Shea, L.; Linderman, J. J. Biochem. Pharmacol. 1997, 53, 519-30.
demonstrated by the BIS kinetic functional assay, more sophisticated fluid handling schemes are easily incorporated into functional assay protocols. This permits greater flexibility in the design of pharmacological assays.
Medical Sciences Grant R01 GM 45260. The authors are indebted to Prof. Neil Nathanson of the UW Dept. of Pharmacology for his thoughtful discussions and guidance in radioligand binding assay execution.
ACKNOWLEDGMENT The authors acknowledge the generous support of ZymoGenetics, Inc. (Seattle, WA) for use of their tissue culture facilities. This research was funded by National Institutes of Health General
Received for review October 29, 1999. Accepted April 13, 2000. AC991231V
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
3115
