Download Concentration-Effect Relationship of l

0022-3565/98/2863-1152$03.00/0
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS
Copyright © 1998 by The American Society for Pharmacology and Experimental Therapeutics
JPET 286:1152–1158, 1998
Vol. 286, No. 3
Printed in U.S.A.
Concentration-Effect Relationship of l-Propranolol and
Metoprolol in Spontaneous Hypertensive Rats after
Exercise-Induced Tachycardia1
LENA BRYNNE, MATS O. KARLSSON and LENNART K. PAALZOW
Department of Pharmacy, Division of Biopharmaceutics and Pharmacokinetics, Uppsala University, Uppsala, Sweden
Accepted for publication April 24, 1998
This paper is available online at http://www.jpet.org
The beta antagonists are among the most widely prescribed drugs for treatment of diverse cardiovascular diseases such as hypertension, angina pectoris, ischemic heart
disease and arrhythmias. Propranolol was the first clinically
used beta receptor antagonist and is the standard to which
other beta antagonists are compared. A number of beta antagonists have been developed with different properties, such
as relative affinity for beta-1 and beta-2 receptors, intrinsic
sympathomimetic activity, membrane-stabilizing activity,
differences in lipid solubility and general pharmacokinetic
disposition properties. Despite their long use, the pharmacodynamic relationships have often been poorly characterized,
mainly due to the restricted concentration ranges used in
clinical studies. Other factors that have hampered the characterization of pharmacodynamic characteristics have been
data with high variability and the use of dose rather than
concentration.
The beta-1 and beta-2 receptors are present in most tisReceived for publication December 22, 1997.
1
This study was supported in part by the Swedish Pharmaceutical Society,
Sweden. This study has been presented at the meeting of the Annual American
Pharmaceutical Society (AAPS), Boston, 1997.
l-propranolol and dl-metoprolol, respectively. The slope in the
linear model was steeper for l-propranolol than for dl-metoprolol, 28.9 6 2.8 and 4.48 6 0.39 beats z ml z (min z mg)21,
respectively. Plasma protein binding of l-propranolol was saturable. The unbound IC50 for l-propranolol was 1.14 6 0.27
ng/ml. The concentration-effect relationship of l-propranolol
was altered at higher plasma concentrations, due to saturable
protein binding. The Imax and the linear concentration-effect
relationship may be interpreted as a specific b-antagonist effect and a membrane-stabilizing effect, respectively. Using exercise-induced tachycardia as a pharmacodynamic endpoint,
to study the effect of b-antagonists in spontaneous hypertensive rats, seems to give reliable results and can be a useful
model to extrapolate to humans.
sues, the former is dominant in myocardial tissues and the
latter is more common in peripheral vessels and bronchial
tissue. The effect of beta-1 blockade is a reduction of myocardial contractility and heart rate, resulting in a decreased
cardiac output, although beta-2 blockade causes bronchial
constriction and decreased vascular tone (Hoffman and
Lefkowitz, 1996). In the absence of sympathetic stimuli,
there is only a poor correlation between drug plasma concentration and the decrease in heart rate (Hager et al., 1981).
This is not surprising, because heart rate at rest is regulated
by the parasympathetic system, whereas exercise increases
the adrenergic innervated activity. Many models have been
developed to quantify the reduction in heart rate caused by
beta antagonists after a controlled pharmacological (e.g., isoproterenol) or physiological (e.g., exercise) stimulus. It is
important to state that these two methods do not yield equivalent mechanistic responses. Heart rate reduction by an isoproterenol stimulus is mediated by both beta-1 and beta-2
receptors (Pringle et al., 1988), whereas exercise is purely
beta-1 mediated (Arnold et al., 1985).
Exercise-induced tachycardia is the most widely used
method of measuring the drug effect in humans (Wellstein et
ABBREVIATIONS: AGP, a1-acid glycoprotein; SHR, spontaneous hypertensive rat.
1152
Downloaded from jpet.aspetjournals.org at ASPET Journals on June 15, 2017
ABSTRACT
The concentration-effect relationship of l-propranolol and dlmetoprolol were investigated in spontaneous hypertensive rats
using reduction in exercise-induced tachycardia as a pharmacodynamic endpoint. The influence of protein binding on the
effect relationship was also assessed. The rats were assigned
to treatment or placebo groups, where each group received
three randomly selected consecutively increasing steady-state
infusions. Different pharmacodynamic effect models were fitted
to the data, using nonlinear mixed effect modeling. The data
were best described by a combined effect model, with a sum of
an ordinary Imax and a linear model. At the lower concentration
range, the ordinary Imax model dominated, although at higher
concentrations, the effect was linearly related to the antagonist
concentration. The Imax were 83 6 6 and 103 6 6 beats z min21
and the IC50 were 18.1 6 4.3 and 50.6 6 15.2 ng/ml for
Pharmacodynamics of b-Blockers
1998
Materials and Methods
Animals and drugs. Male SHR weighing 300 6 10 g (S.D.) were
used (Möllegaard, Ejby, Denmark). The animals were housed under
standardized conditions; room temperature 22 6 1°C (S.D.), humidity 55 6 5% (S.D.) and controlled light (7:00 A.M. to 7:00 P.M.) with
free access to food and water. At least 1 wk before start of experiment, the rats were acclimatized to these conditions. During the
experiment, the rats were kept in individual cages (CMA/120, CMA,
Solna, Sweden). The study was approved by the Animal Ethics
Committee of the University of Uppsala.
Crystalline l-propranolol hydrochloride (99% purity) and dl-metoprolol tartrate (Sigma Chemical Co., St. Louis, MO) were completely
dissolved in a physiological saline solution (Pharmacia & Upjohn AB,
Stockholm, Sweden).
Surgical procedure and drug infusion. Two days before drug
administration, two indwelling polyethylene catheters (PE50, Intramedic, KEBO, Spånga, Sweden) were implanted under light ether
anesthesia. The left carotid artery was used to monitor heart rate
and to sample blood while the study drugs were administered
through the right jugular vein. The animals were trained during
these two days to be accustomed to the handling and equipment.
The study was conducted in two separate parts. In each part, the
rats were randomly assigned to receive drug or placebo treatment,
and were given three consecutive intravenous infusions to steadystate (Harvad Apparatus Syringe Infusion Pump 22; B & K, Sollentuna, Sweden), where l-propranolol was given to 17 rats, dl-metoprolol to 13 and the corresponding placebo groups were five and six
animals each. The steady-state levels were randomly picked among
different aimed steady-state levels. Although, they were always attained from between the lowest up to the highest target concentration level. The desired steady-state plasma concentrations (Css)
were; 20, 70, 200, 400, 800, 1400, 2200, 3000 and 3500 ng/ml for
l-propranolol and 50, 100, 200, 400, 800, 1600, 3200, 6400, 12800 and
25600 ng/ml for dl-metoprolol.
A technique using two consecutive intravenous infusions (Wagner,
1974) was implemented so that different pseudo steady-state plasma
concentrations were rapidly attained. The following equations for
calculating the infusion rates (Q1 and Q2) were used:
Q1 5
Q2 5
Q2
12e
2bzT
Dose z C ss
A/ a 1 B/ b
(1)
(2)
where A, B, a and b are pharmacokinetic macro-constants (Gibaldi
and Perrier, 1982). The pharmacokinetic parameters used to calculate the rate of infusion required for a given steady-state concentration were derived from Terao and Shen (1983) and Vermeulen et al.
(1993). A 30 min (T) loading infusion was followed by a maintenance
infusion over 2 hr and 20 min, for each of the three desired levels.
The drug solution was given in such a concentration that the rats
received less than 3 ml during the total 8.5-hr infusion.
Effect measurements and blood sampling. Mean arterial
blood pressure was measured by a pressure transducer (Statham
P23 DC, Grass Instruments Co., Quincy, MA). Heart rate was captured from the pressure signal, which triggered a tachograph
(7P4DE) and the signals were recorded by a polygraph (Grass model
7). The recorded data were instantly converted in a MacLab interface
(ADInstruments, Castle Hill, Australia) and fed into a Macintosh LC
IIvi computer. The data were stored on a hard disk for off-line
analysis, using the MacLab software Chart, version 3.2.6 (AdInstruments, Castle Hill, Australia).
Exercise-induced tachycardia was obtained by having the animals
run in a motorized wheel for 10 min (6 m z min21). Mean arterial
blood pressure and heart rate were measured continuously for 15
min after exercise. In each rat, two effect measurements were recorded before start of infusion (baseline) and at each concentration
level (effect values taken between 10 to 15 min postexercise). A mean
value of each was used in the pharmacodynamic analysis.
Arterial blood samples, 200 ml, were collected in heparinized Eppendorf tubes. A total of 1200 ml was drawn from each animal. Two
blood samples were drawn at each steady-state level after the effect
measurement, and a mean value was used in the pharmacodynamic
analysis. The samples were centrifuged at 7200 3 g for 10 min
whereby plasma was separated and frozen immediately (270°C)
until analysis.
Protein binding. The binding of l-propranolol to rat plasma was
determined by equilibrium dialysis. Pooled plasma obtained from
eight untreated rats that had undergone the same surgical procedure previously described, including time in the restraining cage,
was frozen (270°C) pending equilibrium dialysis. The plasma sample was adjusted to pH 7.40 using carbogen gas and spiked in
triplicate with known amounts of l-propranolol to achieve the following concentrations: 20, 60, 100, 200, 300, 400, 600, 1000, 2000 and
3500 ng/ml. The Teflon chambers were filled with 0.8 ml plasma and
0.8 ml of 0.13 M phosphate buffer (pH 7.40). Before dialysis, the
Spectrapor 4 membranes (Spectrum Medical Industries Inc., Houston, TX; cut-off value of 12–14,000) were soaked in buffer for 2 to 3
hr. The cells were kept rotated at 37°C for 4.5 hr, by which time
equilibrium has been reached. Volume shift was assessed by weighing the plasma/buffer solutions before and after equilibrium dialysis.
a1-Acid glycoprotein and drug analysis. The AGP concentrations were determined by using a modification of a previous published method (QR method) (Imamura et al., 1994). The analytical
system consisted of a pump (ESA-580 Chelmsford, MA), a Triathlon
autoinjector (Sparc Holland Emmen, the Netherlands), a fluorescence detector (Shimadzu RF-551, Kyoto, Japan), and an integrator
(Schimadzu C-R5A Chromatopac). The detector was set at excitation
and emission wavelengths of 496 nm and 590 nm, respectively. A
1/15 M phosphate buffer (pH 7.40) was used as mobile phase with a
flow rate of 0.5 ml/min. All volumes used were one-fourth of that
described in the original method, when preparing standard curves
and samples. Standards and samples were placed in the autosampler, where 50 ml were injected. The difference in height of the
Downloaded from jpet.aspetjournals.org at ASPET Journals on June 15, 2017
al., 1992), but corresponding animal studies have been few.
SHR have been shown to be an appropriate animal model in
studying essential hypertension in humans, and was used in
this study to evaluate the concentration-effect relationship of
two beta antagonists, using exercise-induced tachycardia as a
pharmacodynamic endpoint.
One of the most widely prescribed beta antagonists, metoprolol, was selected, using propranolol as a reference. Because of the large difference in disposition between the two
enantiomers of propranolol (Walle et al., 1988), the l-enantiomer was selected in this study. There is also a large difference in protein binding between the two beta antagonists.
Propranolol is extensively bound (95%) to plasma proteins, a
binding that decreases at higher concentrations (Smits and
Struyker-Boudier, 1979), whereas the more hydrophilic
metoprolol is negligibly bound (12%) (Brogden et al., 1977).
Consequently, when an effect model is evaluated over a large
concentration interval, the impact of protein binding on the
concentration-effect relationship has to be considered.
Our objectives were to determine the concentration-effect
relationships of l-propranolol and dl-metoprolol in conscious
SHR, and to evaluate the use of exercise-induced tachycardia
as a pharmacodynamic endpoint for future studies. The influence of plasma protein binding of l-propranolol was also
determined.
1153
1154
Brynne et al.
Vol. 286
Epred ij 5 E 0, i 2 m i z C ss, ij
(3)
Epred ij 5 E 0, i 2
I max, i z C ss, ij
IC 50, i 1 C ss, ij
(4)
Epredij is the jth observed effect at the corresponding steady-state
plasma concentration (Css,ij) of the ith individual, E0 the baseline
effect, Imax the maximal effect and IC50 the concentration at halfmaximal effect and m is the slope of the linear relationship between
Epred and Css. As the data suggested an extension of the ordinary
Imax model, dual effect models (Paalzow and Edlund, 1979; Holford
and Sheiner, 1982; Paalzow, 1990) were also considered. Both a
linear (equation 5) and an ordinary Imax model (equation 6) were
tested as the dual effect.
Epred ij 5 E 0, i 2
Epred ij 5 E 0, i 2
I max, i z C ss, ij
IC 50, i 1 C ss, ij
I max1, i z C ss, ij
IC 51, i 1 C ss, ij
2 m i z C ss, ij
2
I max2, i z C ss, ij
IC 52, i 1 C ss, ij
(5)
(6)
The last model did not significantly improve the objective function
value, and therefore the more simple dual effect model (equation 5)
was chosen to describe the data.
The effect after placebo administration was evaluated separately
using nonlinear mixed effect modeling, estimating the deviation
from baseline values. No significant change in heart rate was found
in the control groups of propranolol (n 5 5) and metoprolol (n 5 6).
However, the data from the control rats were incorporated into the
data analysis because they contain information about baseline and
residual variability. All data were first analyzed separately for each
drug. Simultaneous fits of data from both drugs were performed, as
well as to total and unbound propranolol, to evaluate differences
between the estimated population pharmacodynamic parameters.
The protein binding of l-propranolol was best described by the
following equation using naive pooling in the nonlinear regression
analysis:
C 5 Cu 1
N z C u z Pt
1 1 Ka z Cu
(7)
where C and Cu represents total and unbound l-propranolol concentrations, respectively. N is the number of binding sites per mole
protein, Pt is the total concentration of binding protein, i.e., AGP, and
Ka is the association constant between drug and binding site. The
unbound fraction was calculated as the ratio of unbound and total
drug concentrations. Individual unbound concentrations were calculated by using equation 7.
The nonlinear regression analysis was performed by using both
the first-order (FO) and first-order conditional estimation (FOCE)
methods in the software package NONMEM, version V (Beal and
Sheiner, 1992). The NONMEM program is based on a statistical
model that explicitly takes into account both the inter-animal variability in the parameters as well as intra-animal residual error. All
data will be fitted simultaneously although preserving the individuality. This approach is particular suitable in cases where some of
the concentration-effect curves are not completely defined, as all
available information may provide reasonable estimates of the missing parts. The graphic analysis was performed using the Xpose
package (Jonsson and Karlsson, 1998) running under Splus, version
3.3 (Statistical Sciences, 1993). An additive error model (equation 8)
was used to characterize the residual error:
Eobs ij 5 Epred ij 1 e ij
(8)
where Eobsij and Epredij are the jth observed and predicted effect for
the ith individual. eij represents the residual error of variance for the
ith individual, its distance between the jth observation and prediction. The values for eij are assumed to be symmetrically distributed,
Downloaded from jpet.aspetjournals.org at ASPET Journals on June 15, 2017
chromatograms, with and without QR solution, gave the fluorescence
of AGP alone and was used as the response in the standard curves.
Rat AGP showed a lower fluorescence response than human AGP.
Human AGP standards were used to check the linearity of the
system and rat AGP standards were used in the standard curve. The
method was highly reproducible with a interday variability of ,8%.
The limit of quantification was 0.08 mg/ml with a coefficient of
variation of 17% (n 5 6).
The l-propranolol and dl-metoprolol concentrations in plasma and
phosphate buffer were determined using high-performance liquid
chromatography with fluorescence detection, as described by Rutledge and Garrick (1989), with some modifications. The analytes
were extracted from 100 ml samples and standards in a 10-ml roundbottomed test tubes. Internal standard, 50 ml, (12 mM metoprolol
tartrate or 4 mM propranolol HCl in distilled water) was added to the
100-ml sample followed by 100 ml sodium hydroxide (4 M) and mixed
briefly. Three mLs of diethyl ether was then added to the mixture.
The contents of the test tube were vortexed for 2 min and centrifuged
at 1400 3 g for 10 min. The organic phase (upper layer) was transferred into a 10-ml conical glass tube and evaporated to dryness
under a steam of oxygen free nitrogen, at 35°C. The residue was
dissolved in 150 ml of 5 mM H2SO4, vortexed and then centrifuged
(10 min 2000 3 g) before injection. In the protein binding study, the
phosphate buffer samples were diluted with 5 mM H2SO4 and the
plasma samples with blank plasma, when necessary, up to a total
volume of 100 ml, to end up within the range of the standard curve.
All solvents and reagents were of analytical grade.
The chromatographic equipment consisted of a pump (Shimadzu
LC-9A, Kyoto, Japan), a mBondapak colonn (C18 column, 30 cm 3 3.9
mm i.d., Waters Associates Milford, MA), a fluorescence detector
(Shimadzu RF-535 or Jasco FP-920 Tokyo, Japan) and an integrator
(Schimadzu C-R5A Chromatopac). The excitation and emission
wavelengths were 235 nm and 335 nm using Shimadzu RF-535 for
l-propranolol, and 280 and 305 nm using Jasco FP-920 for dl-metoprolol, respectively, in the animal study. For the protein binding
analysis, the l-propranolol plasma and phosphate concentrations
were analyzed using Jasco FP-920 with excitation and emission
wavelengths of 222 and 350 nm, respectively. All sample injection
was performed using a CMA/200 autoinjector (CMA, Solna, Sweden)
fitted with a 100-ml sample loop. The composition of the mobile phase
was a 65:35 mixture of solution A and B, respectively, when running
plasma samples and changed to 70:30 when running samples of
phosphate buffer (pH 7.40). Solution A contains 25 ml of l-heptanesulfonic acid (5 mM) in 1 liter acetic acid (1%) and solution B contains
25 ml of l-heptanesulfonic acid (5 mM) in 1 liter acetonitrile. The flow
rate was 2.0 ml/min. Retention times for plasma samples were 2.7
and 4.7 min for metoprolol and propranolol.
Peak area ratios of drug/internal standard were calculated and
used in single-level calibration curves drawn through origin with
replicates (n 5 5) of the highest calibrator. The standard curves were
linear within the range of 1.8 to 400 ng/ml and 3.9 to 800 ng/ml for
l-propranolol and dl-metoprolol, respectively. The interday variation
for l-propranolol in the concentration range 1.8 to 330 ng/ml was
#5% and the accuracy varied between 92 to 100%. The interday
variation for metoprolol in the concentration range 9.8 to 490 ng/ml
was #11% and the accuracy varied between 90 and 100%. The
absolute recovery was between 100% (1.8 ng/ml) to 96% (326 ng/ml)
for l-propranolol, although it varied between 112% (3.9 ng/ml) and
93% (780 ng/ml) for dl-metoprolol. The limit of quantification was 1.8
ng/ml for l-propranolol and 3.9 ng/ml for dl-metoprolol, respectively.
Data analysis. Population models describing the concentrationeffect relationships of the two b-antagonists were performed by relating the reduction of induced exercise-tachycardia directly to the
measured steady-state plasma concentrations (time-independent) of
the drugs. Both a linear (equation 3) and an ordinary Imax model
(equation 4) were first fit to the data.
Pharmacodynamics of b-Blockers
1998
with mean zero and variance s2. A proportional error model was used
in the protein binding study:
Cobs 5 Cpred ~ 1 1 e !
(9)
where Cobs and Cpred are the observed and predicted total plasma
concentration. The interindividual variability in the pharmacodynamics is described using an exponential variance model, assuming
log-normal distribution of the pharmacodynamic parameters:
P i 5 P̃ z exp~ h i !
(10)
Results
Protein binding. The relationship between unbound fraction and total plasma concentration of l-propranolol was described by a binding model for different independent sets of
equivalent binding sites (fig. 1). The number of binding site
was 1.08 6 0.07 and the association constant was 0.736 6
0.059 mM21 at an AGP concentration of 22.6 mM (MW
43,500). The fraction unbound of l-propranolol was 5.3 6
0.5% (mean 6 S.D.) and linear in the 60 to 1400 ng/ml
concentration range (fig. 1). At more than 1400 ng/ml, the
free fraction increased, and at 3500 ng/ml, it was 10 6 0.7%
(mean 6 S.D.). No results were obtained for the lowest stud-
ied concentration (20 ng/ml) because of assay limitations. No
apparent volume shift was seen during the 4.5-hr equilibration time.
Concentration-effect relationship. The baseline variation between the two treatment groups was not significantly
different. The maximal observed decrease in heart rate for
l-propranolol and dl-metoprolol was 168 and 216 beats z
min21, respectively. At concentrations higher than 3000
ng/ml of l-propranolol, acute bradycardia occurred, and consequently the highest steady-state level has been excluded in
this evaluation.
The same results were obtained when data were analyzed
separately or simultaneously, and therefore the latter is presented here. The concentration-effect relationships of the two
drugs was best described by a combined effect model, represented as the sum of an ordinary Imax model and a linear
model. The individual baseline normalized effect values,
transformed after the modeling, are shown in figure 2. The
pharmacodynamic parameter estimates are presented in table 1. No significant differences between l-propranolol and
dl-metoprolol were found in either the Imax or the IC50 values. The Imax values were 21.3% for l-propranolol and 26.0%
for dl-metoprolol, and their corresponding IC50 values were
18.1 ng/ml (69.8 nM) and 50.6 ng/ml (94.6 nM). A significant
difference in the linear slopes (P , .001) was found between
the two drugs. The concentration interval at Imax, where a
plateau occurred, was longer for dl-metoprolol than for
l-propranolol.
Figure 3 illustrates the simulations of the partial and the
combined concentration-effect relationships for l-propranolol. The inter-animal variability was low (,12%) for all parameters except for the IC50 values. The estimated IC50
values with 95% confidence intervals were predicted to be
18.1 (4 –74) and 1.14 (0.28 – 4.67) ng/ml for total and unbound
l-propranolol and 50.6 (15–173) ng/ml for dl-metoprolol. The
intra-animal variability was 13 beats z min21. The first-order
(FO) and first-order conditional estimation (FOCE) methods
performed equally, and therefore only the results of the
former are presented.
Simulation of the total and unbound l-propranolol and
dl-metoprolol concentration-effect relationships are illustrated in figure 4A. The corresponding two parts of the concentration-effect relationships are given in figures 4B and C.
The IC50u value of l-propranolol was 1.14 6 0.27 ng/ml (4.40
nM). The IC50u value of dl-metoprolol was calculated to be
44.5 ng/ml (83.3 nM), assuming linear protein binding and
that the unbound fraction of dl-metoprolol is 88% (Brogden et
al., 1977). About a 20-fold difference in the IC50u values was
found when the two drugs were compared. There were large
differences between the slopes of the linear model; 343 6 51
and 28.9 6 2.8 beats z ml z (min z mg)21 for unbound and total
l-propranolol and 4.48 6 0.39 beats z ml z (min z mg)21 for
dl-metoprolol.
Discussion
Fig. 1. Nonlinear plasma protein binding of l-propranolol (mean 6 S.D.)
described by a binding model for different independent set of equivalent
binding sites.
In earlier pharmacodynamic studies of b-antagonists, simplified concentration-effect relationships such as linear or
log-linear models have frequently been used, mainly due to
the small dose/concentration range studied in humans. In
this study, we clearly demonstrate that the concentrationeffect relationship of l-propranolol and metoprolol is more
Downloaded from jpet.aspetjournals.org at ASPET Journals on June 15, 2017
in which Pi is the value of the ith individual of an arbitrary pharmacodynamic parameter, P̃ is the population mean parameter value,
and exp(hi) expresses the (random) difference between Pi and P̃. The
values for hi are assumed to be independently multivariate distributed, with mean zero and diagonal variance-covariance matrix V
with diagonal elements (v12, . . . , vm2). The values of the population
parameters u, s2, and V are estimated from the data, where u is the
population parameter estimate.
Discrimination between different models was calculated via comparison of the objective function values (22log likelihood) by NONMEM and by visual inspection of the goodness of fit plots. The
difference between the objective function values for two hierarchical
models is approximately chi-square distributed and may consequently be used for model selection purposes. In this study, P , .05
was used as the significance level. All data are given as mean 6 S.E.
unless otherwise stated.
1155
1156
Brynne et al.
Vol. 286
TABLE 1
Pharmacodynamic estimates obtained by population analysis of concentration-effect data after l-propranolol and dl-metoprolol administration to
SHR (mean 6 S.E.)
E0 (beats z min21)
l-Propranolol n 5 22
Total
Unbound
dl-Metoprolol na 5 19
Total
Imax (beats z min21)
IC50 (ng z ml21)
m (beats z ml z (min z mg)21)
fu (%)
a
a
b
c
388 6 4
388 6 4
283 6 6
289 6 7
18.1 6 4.3
1.14 6 0.27
28.9 6 2.8
343 6 51
5.3 6 0.5b
396 6 6
2103 6 6
50.6 6 15.2
4.48 6 0.39
88c
All animals.
Free fraction at the linear relationship.
Data taken from Brogden et al., 1977.
Fig. 3. Simulated concentration-effect relationship of l-propranolol,
showing both the combined and partial effect models.
complex and better described by a combined relationship of
at least two different components. This was obtained by
using a well-defined stimulus (exercise-induced tachycardia)
in combination with a large drug concentration interval.
When using a large dose/concentration interval, saturation
processes might be revealed, such as nonlinear protein binding. Differences in the free fraction are of minor importance
in experiments with only one compound, but when a highly
bound agent is compared to a virtually unbound drug, this
needs to be accounted for. The protein binding in this study
was linear at low plasma l-propranolol concentrations, which
is in agreement with previous findings in Wistar rats (Chindavijak et al., 1988). An increase in free fraction has earlier
been found for dl-propranolol in SHR (Smits and StruykerBoudier, 1979), where the nonlinearity started at around 100
ng/ml in rats with a normal AGP level. From the literature it
is known that the AGP level can increase 6-fold 2 days after
surgery (Yasuhara et al., 1983; Lin et al., 1987), suggesting
that the nonlinearity in this study would occur at higher drug
concentrations.
In addition to surgery, infections, inflammations and
stress, are important conditions that will increase the AGP
concentration. It has been shown that the AGP level has a
large influence on both the pharmacokinetics and the apparent pharmacodynamics of propranolol when total plasma
concentrations are used in the evaluation (Yasuhara et al.,
1983), and consequently the levels of AGP is essential to
quantify and explain interindividual variability. The saturation of the protein binding in our study affects the response of
l-propranolol at high plasma concentrations. It increases the
steepness of the concentration-effect relationship at high
plasma levels, suggesting a nonlinear increase of the effect at
Downloaded from jpet.aspetjournals.org at ASPET Journals on June 15, 2017
Fig. 2. The concentration-effect relationships of l-propranolol and dl-metoprolol, observed (F) and predicted (——).
1998
high doses. During saturated protein binding the half-life is
likely to increase, because the relative change in the volume
of distribution is expected to be larger than in clearance
(Evans et al., 1973). In addition, the b-antagonists at these
high concentration levels will decrease the blood flow and
thereby the clearance, which suggest that an even longer
half-life should be observed. The consequence on duration of
action is difficult to interpret, but changes in kinetics is
suggested to be larger than the difference in the effect slope.
The maximal response after exercise-induced heart rate in
humans occurred at plasma dl-propranolol concentrations
from 100 ng/ml and above (Hager et al., 1981; Lalonde et al.,
1987; Sowinski et al., 1995), which is in agreement with our
data. The Imax value was 21.3% for l-propranolol in this
1157
study, which is lower in comparison values reported from
human studies, 33.3 to 38.3%, where dl-propranolol was
given orally (Lalonde et al., 1987; Sowinski et al., 1995). The
estimated IC50 value in our study, 18.1 ng/ml (l-propranolol),
was within the range reported in these human studies (10.2–
24.4 ng/ml, dl-propranolol). The unbound IC50 values in
these human studies were 1.29 to 2.77 ng/ml, which is similar
to our observation (1.14 ng/ml). The corresponding values
reported for dl-metoprolol are in agreement with the present
findings (Abrahamsson et al., 1990).
In a study performed on dogs, both propranolol and metoprolol equipotently reduced exercise-induced heart rate in a
similar fashion (Åblad et al., 1980). In our study, we did not
find large differences in the pharmacodynamics between the
two drugs either. However, a 20-fold difference appears when
the IC50 values are corrected for protein binding, making
propranolol the more potent drug. The difference in IC50
values of different studies could be due to the exercise test
used, i.e., duration, power and type. As the drugs are competitive antagonists, a higher degree of exercise (agonist) will
shift the effect relationship toward higher concentrations
(Wellstein et al., 1985).
Except for the large difference in protein binding between
propranolol and metoprolol, they also differ in such factors as
their respective affinities to the beta-1 and beta-2 receptors,
membrane stabilizing activity and lipophilicity. Selectivity is
a relative rather than an absolute property, because increasing doses of b1-selective antagonists results in a dose-related
beta-2 adrenoceptor blockade (Lipworth et al., 1991) and
because higher doses of a beta-2 adrenoceptor selective antagonist will reduce exercise-induced tachycardia (Harry et
al., 1988). The different roles of beta-1 and beta-2 receptors in
the genesis of cardiovascular responses have proved to be
more complex, because of their coexistence in the heart
(Brodde et al., 1983). The rat’s left ventricle contains about
74% beta-1 and 26% beta-2 adrenergic receptors (Vago et al.,
1984). Due to the high density of beta-1 receptors in the
heart, one would expect that they would contribute more to
the positive inotropic and chronotropic responses than would
beta-2 (Molenaar and Summers, 1987), and that the latter
would play only a minor role in the heart rate alteration.
A major effect of beta antagonist therapy is increased AV
nodal conduction time resulting in a prolonged AV nodal
refractoriness (increased PR interval). Hence, beta antagonists are useful in terminating reentrant arrhythmias that
involve the AV node and in controlling ventricular response
in atrial fibrillation or flutter (Roden, 1996). As it is difficult
to standardize and use arrhythmias as pharmacodynamic
endpoints, many studies have been performed in vitro. Pruett
and coworkers (1977, 1980) found that d- and dl-propranolol
shortened action potential duration with similar potency at a
concentration as low as 100 ng/ml. The antiarrhythmic activity correlated better with the tissue propranolol concentration, and when this was taken into account, the minimum
plasma concentration expected to produce an electrophysiologic effect in patients was 150 ng/ml. This was confirmed
when 40% of the patients with ventricular arrhythmias responded to the drug at plasma concentrations in excess of 150
ng/ml, a level associated with a high degree of beta receptor
blockade (Woosely et al., 1979). When d-propranolol was
given separately, it suppressed ventricular arrhythmias both
Downloaded from jpet.aspetjournals.org at ASPET Journals on June 15, 2017
Fig. 4. The predicted effect models showing total and unbound plasma
concentration; (——) dl-metoprolol, (– – –) l-propranolol and (• – • – )
unbound l-propranolol. The combined effect model (A) for both drugs
shows similar Imax values (B) but large differences in the linear slopes (C)
between the drugs.
Pharmacodynamics of b-Blockers
1158
Brynne et al.
Acknowledgments
The authors thank Mrs. Britt Jansson for her analytical expertise
and Beatrice Nilsson and Marika Segerström for hard work during
their undergraduate studies. The authors thank the section of PreClinical Pharmacokinetics, Pharmacia & Upjohn AB, Uppsala, Sweden, for placing the protein binding equipment at our disposal. Janet
Wade, PhD, is cordially thanked for discussion of the manuscript.
References
Åblad B, Carlsson E, Johansson and Regårdh C-G (1980) Metoprolol, in Pharmacology of Antihypertensive Drugs (Scriabine A ed) pp 247–262, Raven Press, New
York.
Abrahamsson B, Lücker P, Olofsson B, Regårdh C-G, Sandberg A, Wieselgren I and
Bergstrand R (1990) The relationship between metoprolol plasma concentration
and beta1-blockade in healthy subjects: A study on conventional metoprolol and
metoprolol CR/ZOK formulations. J Clin Pharmacol 30:S46 –S54.
Arnold JMO, O’Connor PC, Riddell JG, Harron DWG, Shanks RG and McDevitt DG
(1985) Effects of the b2-adrenoceptor antagonist ICI 118,551 on exercise tachycardia and isoprenaline-induced b-adrenoceptor responses in man. Br J Clin Pharmacol 19:619 – 630.
Beal SL and Sheiner LB eds (1992) NONMEM Users Guides. NONMEM Project
Group, University of California at San Francisco, San Francisco.
Brodde O-E, Karad K, Zerkowski H-R, Rohm N and Reidemeister JC (1983) Coexistence of b1- and b2-adrenoceptors in human right atrium. Circ Res 53:752–758.
Brogden RN, Heel RC, Speight TM and Avery GS (1977) Metoprolol: A review of its
pharmacological properties and therapeutic efficacy in hypertension and angina
pectoris. Drugs 14:321–348.
Chindavijak B, Belpaire FM, De Smet F and Bogaert MG (1988) Alteration of the
pharmacokinetics and metabolism of propranolol and antipyrine elicited by indwelling catheters in the rat. J Pharmacol Exp Ther 246:1075–1079.
Evans GH, Nies AS and Shand DG (1973) The disposition of propranolol. III.
Decreased half-life and volume of distribution as a result of plasma binding in
man, monkey, dog and rat. J Pharmacol Exp Ther 186:114 –122.
Gibaldi M and Perrier D (1982) Multicompartment models. Pharmacokinetics, 2nd
ed pp 45–111, Marcel Dekker, Inc., New York.
Hager WD, Pieniaszek HJ, Perrier Jr BA, Mayersohn M and Goldberger V (1981)
Assessment of beta blockade with propranolol. Clin Pharmacol Ther 30:283–290.
Harry JD, Norris SC, Percival GC and Young J (1988) The dose in humans at which
ICI 188,551 (a selective b2-adrenoceptor blocking agent) demonstrates blockade of
b1-adrenoceptors. Clin Pharmacol Ther 43:492– 498.
Hoffman BB and Lefkowitz RJ (1996) Catecholamines, sympathomimetic drugs, and
adrenergic receptor antagonists, in Goodman & Gilman’s The Pharmacological
Basis of Therapeutics, 9th ed (Hardman JG, Limbird LE, Molinoff PB, Ruddon RW
and Gilman AG eds) pp 199 –248, McGraw-Hill, Inc., New York.
Holford NHG and Sheiner LB (1982) Kinetics of pharmacologic response. Pharmacol
Ther 16:143–166.
Imamura H, Maruyama T, Okabe H, Shimada H and Otagiri M (1994) A simple and
rapid fluorometric determination method of a1-acid glycoprotein in serum using
quinaldine red. Pharm Res 11:566 –570.
Jonsson EN and Karlsson MO (1998) Xpose—an S-PLUS based population pharmacokinetic/pharmacodynamic model building aid for NONMEM. Comp Methods
Prog Med, in press.
Lalonde RL, Straka RJ, Pieper JA, Bottorff MB and Mirvis DM (1987) Propranolol
pharmacodynamic modeling using unbound and total concentrations in healthy
volunteers. J Pharmacokin Biopharm 15:569 –582.
Lin T-H, Sugiyama Y, Sawada Y, Suzuki Y, Iga T and Hanano M (1987) Effect of
surgery on serum a1-acid glycoprotein concentration and serum protein binding of
dl-propranolol in phenobarbital-treated and untreated rats. Drug Metab Dispos
15:138 –140.
Lipworth BJ, Irvine NA and McDevitt DG (1991) A dose-ranging study to evaluate
the b1-adrenoceptor selectivity of bisoprolol. Eur J Clin Pharmacol 40:135–139.
Molenaar P and Summers RJ (1987) Characterization of beta-1 and beta-2 adrenoceptors in guinea pig atrium: Functional and receptor binding studies. J Pharmacol Exp Ther 241:1041–1047.
Murray KT, Reilly C, Koshakji RP, Roden DM, Lineberry MD, Wood AJJ, Siddoway
LA, Barbey JT and Woosely RL (1990) Suppression of ventricular arrhythmias in
man by d-propranolol independent of beta-adrenergic receptor blockade. J Clin
Invest 85:836 – 842.
Paalzow LK (1990) Pharmacodynamics, in Comprehensive Medical Chemistry, Biopharmaceutics (Hansch C, Sammes PG and Taylor JB eds) pp 333–358, Pergamon
Press, Oxford.
Paalzow LK and Edlund PO (1979) Multiple receptor responses: A new concept to
describe the relationship between pharmacological effects and pharmacokinetics of
a drug: Studies on clonidine in the rat and cat. J Pharmacokin Biopharm 7:495–
510.
Pringle TH, Riddell JG and Shanks RG (1988) Characterization of the betaadrenoceptors which mediate the isoprenaline-induced changes in finger tremor
and cardiovascular function in man. Eur J Clin Pharmacol 35:507–514.
Pruett JK, Walle T and Bagwell EE (1977) Association of in vivo and in vitro
propranolol levels in canine Purkinje fibers with antiarrhythmic effects. J Pharmacol Exp Ther 202:45–54.
Pruett JK, Walle T and Walle UK (1980) Propranolol effects on membrane repolarization time in isolated canine Purkinje fibers: Threshold tissue content and the
influence of exposure time. J Pharmacol Exp Ther 215:539 –543.
Roden DM (1996) Antiarrhythmic drugs, in Goodman & Gilman’s The Pharmacological Basis of Therapeutics, 9th ed (Hardman JG, Limbird LE, Molinoff PB,
Ruddon RW, Gilman AG eds) pp 839 – 874, McGraw-Hill, Inc., New York.
Rutledge RD and Garrick C (1989) Determination of metoprolol and its a-hydroxide
metabolite in serum by reversed-phase high-performance liquid chromatography.
J Chromatogr Sci 27:561–565.
Smits JFM and Struyker-Boudier HAJ (1979) Propranolol in conscious spontaneously hypertensive rats. II. Disposition after subcutaneous and intracerebroventricular administration. Naunyn-Schmiedeberg Arch Pharmacol 309:19 –24.
Sowinski KM, Burlew BS and Johnson JA (1995) Racial differences in sensitivity to
the negative chronotropic effects of propranolol in healthy men. Clin Pharmacol
Ther 57:678 – 683.
Statistical Sciences Inc. (1993) S-PLUS Reference Manual, version 3.3. StatSci, a
division of Mathsoft Inc., Seattle, WA.
Terao N and Shen DD (1983) Alterations in serum protein binding and pharmacokinetics of l-propranolol in the rat elicited by the presence of an indwelling venous
catheter. J Pharmacol Exp Ther 227:369 –375.
Vago T, Bevilacqua M, Dagani R, Meroni R, Frigeni G, Santoliß C and Norbiato G
(1984) Comparison of rat and human left ventricle beta-adrenergic receptors:
Subtype heterogeneity delineated by direct radioligand binding. Biochem Biophys
Res Comm 121:346 –354.
Vermeulen AM, Belpaire FM, De Smet F, Vercruysse I and Bogaert MG (1993) Aging
and the pharmacokinetics and metabolism of metoprolol enantiomers in the rat. J
Gerontology 48:B108 –B114.
Wagner JG (1974) A safe method for rapidly achieving plasma concentration plateaus. Clin Pharmacol Ther 16:691–700.
Walle T, Webb JG, Bagwell EE, Walle UK, Daniell HB and Gaffney TE (1988)
Stereoselective delivery and actions of beta receptor antagonists. Biochem Pharmacol 37:115–124.
Wellstein A, Palm D, Pitschner HF and Beiz GG (1985) Receptor binding of propranolol is the missing link between plasma concentration kinetics and the effect-time
course in man. Eur J Pharmacol 29:131–147.
Wellstein A, Woosley RL and Schwartz SL (1992) Antiarrhythmic drugs and cardiac
glycosides, in The In Vivo Study of Drug Action (van Boxtel CJ, Holford NHG,
Danhof M eds) pp 277–301, Elsevier Science Publishers B. V., Amsterdam.
Woosley RL, Kornhauser D, Smith R, Reele S, Higgins SB, Nies AS, Shand DG and
Oates JA (1979) Suppression of chronic ventricular arrhythmias with propranolol.
Circulation 60:819 – 827.
Yasuhara M, Fujiwara J, Kitade S, Katayama H, Okumura K and Hori R (1983)
Effect of altered plasma protein binding on pharmacokinetics and pharmacodynamics of propranolol in rats after surgery: Role of a1-acid glycoprotein. J Pharmacol Exp Ther 235:513–520.
Send reprint requests to: Dr. Lena Brynne, Department of Pharmacy,
Division of Biopharmaceutics and Pharmacokinetics, Box 580, S-751 23 Uppsala, Sweden.
Downloaded from jpet.aspetjournals.org at ASPET Journals on June 15, 2017
through non-beta-mediated effect (includes membrane-stabilizing activity) and beta-mediated effect (Murray et al., 1990).
Propranolol and metoprolol display a large difference in
partition coefficient, where propranolol is much more lipophilic (logD 20.2) than metoprolol (logD 0.98). The large
difference in lipophilicity could explain the difference in
membrane-stabilizing activity, due to membrane dissolving
capacity. The steeper slope of propranolol should then indicate that it would display more membrane-stabilizing activity as compared to metoprolol, which is in agreement with
the literature (Hoffman and Lefkowitz, 1996). Concluding
that the linear component in the dual effect model could be
membrane-stabilizing activity.
In conclusion, using spontaneous hypertensive rats to
study the concentration-effect relationship of beta-blockers,
with exercise-induced tachycardia as the pharmacodynamic
endpoint, gives reliable results and may be a useful model to
extrapolate to humans. The combined Imax and linear concentration-effect relationship can be interpreted as a specific
beta antagonist effect and a membrane-stabilizing effect, respectively.
Vol. 286