A mixture of amino acids inhibits propranolol metabolism in
perfused rat livers. To obtain mechanistic information about the interaction, a related but less tissue-bound drug, metoprolol, was used
to determine Vmax and
KM for parent drug and two metabolites in the
presence and absence of amino acids. Six groups of 4 livers from 24 male Sprague-Dawley rats were perfused in the single-pass mode at 3 ml/min/g liver for 130 min with oxygenated buffer containing 3.74, 4.49, 5.61, 7.48, 18.7, or 44.9 µM metoprolol. From 50 to 90 min, a
balanced amino acid mixture was included in the buffer. Samples of
liver effluent taken every 5 min were analyzed by HPLC for metoprolol
and two metabolites, 
-hydroxymetoprolol and
O-demethylmetoprolol. Steady-state concentrations of drug
determined before, during, and after amino acids were used to determine
Vmax and apparent KM
values by nonlinear curve-fitting under each condition. Amino acids
reversibly reduced the Vmax values of
metoprolol and both metabolites by ~50% without significantly
affecting apparent KM values. As a result,
large increases in availability occurred, especially at low metoprolol
inlet concentrations (>90%). Amino acids also increased oxygen
consumption until the effluent buffer was almost depleted. Possible
mechanisms influencing Vmax include direct
inhibition of metabolic enzymes by amino acids or cosubstrate (NADPH or
oxygen) limitation. Amino acid-mediated pericentral oxygen depletion in
the hepatic sinusoids could result in inhibition of drug-metabolizing
enzymes, and is consistent with a reduction of
Vmax and oxygen depletion in the effluent
buffer during amino acid coinfusion. We postulate that one or more of
these mechanisms could contribute to the interaction between food and
high first-pass drugs observed in humans.
 |
Introduction |
Rac-metoprolol
is one of several drugs known to exhibit increased oral availability
when coadministered with a high-protein meal, including propranolol,
propafenone, labetolol, zuclopenthixol, and dixyrazine (1). Although
this food interaction was first observed two decades ago, its mechanism
remains to be fully elucidated, despite intensive study (1). The
mechanism is likely to be complex, with contributions from more than
one of the physiological responses to food combining to cause a net
increase in AUCoral.1 Because drugs that show
the "food effect" are almost completely absorbed from the
gastrointestinal tract, it is generally agreed that the interaction is
due to a reduction in first-pass metabolism (1). Studies to date on
propranolol have concentrated on the liver, although recent studies in
rabbits suggest the possible contribution of intestinal metabolism (2).
A reduction in hepatic first-pass metabolism could occur through
changes to hepatic blood flow, plasma protein binding, or metabolic
activity. The hypothesis that the increase in AUCoral of
propranolol could be caused by a transient increase in hepatic blood flow (3) was questioned, because flow changes could not account
for the magnitude of the increase (4, 5). It is also unlikely that food
causes an increase in the unbound fraction of propranolol in plasma
(6). Experiments in humans to test for transient changes in metabolic
activity have shown that food causes inhibition of presystemic primary
conjugation of propranolol (7), although data relating to the more
important phase I pathways are inconclusive (8, 9). Nevertheless,
simulations have indicated that propranolol availability would be most
sensitive to changes in apparent Vmax and
somewhat sensitive to changes in apparent KM,
both of which contribute to intrinsic clearance (10).
To explore further whether nutrients could inhibit propranolol
metabolism, amino acids were coinfused into rat livers perfused with
buffer containing propranolol (11). Global inhibition of metabolism was
observed, indicating that at concentrations achieved after a high
protein meal, amino acids could inhibit all of the pathways of
propranolol metabolism and that the degree of metabolic inhibition was
related to the concentration of amino acids in the buffer. Hepatic
tissue binding prevented further exploration of the mechanism using
propranolol as a model drug. Pilot studies showed, however, that
metoprolol is much less extensively tissue bound and its metabolism was
inhibited by amino acids, making it an appropriate model drug for
mechanistic studies involving the measurement of Michaelis-Menten
parameters. The metabolism of metoprolol is depicted in fig.
1. We describe herein experiments that confirm our pilot
observations in perfused rat livers and that focus on amino acid
effects on the apparent Vmax and
KM of metoprolol metabolism. As part of the
liver viability assessment, effluent buffer O2 content was
measured.
Materials and Methods
Chemicals.
(±)-Metoprolol tartrate and nadolol (internal standard) were purchased
from Sigma Chemical Co. (St. Louis, MO). H119/66 p-OH benzoate (-hydroxymetoprolol), H105/22 p-OH benzoate
(O-demethylmetoprolol), and H117/04 HCl (metoprolol acid)
were gifts from Astra Hässle (Mölndal, Sweden). Solvents
were HPLC grade (BDH, Toronto, Canada). All other chemicals used were
of analytical grade.
Animals.
Livers used in the study were from 24 male Sprague-Dawley rats
(198-250 g, Charles River, St. Constant, Quebec, Canada) maintained on
standard laboratory chow and water ad libitum in accordance with the guidelines of the Canadian Council on Animal Care. Animals were randomly assigned to six groups of four in a parallel design, one
group for each of six concentrations of metoprolol in the perfusion
medium (3.74, 4.49, 5.61, 7.48, 18.7, and 44.9 µM). Perfusions were
performed sequentially according to the random assignments.
Surgical Procedure.
After anesthetizing the rat with methoxyflurane (Pitman-Moore,
Mississauga, Ontario, Canada), the liver was isolated and perfused in
the single-pass mode with oxygenated Krebs bicarbonate buffer in a
temperature- and humidity-controlled cabinet as previously described
(12, 13). During the surgical procedure, there was no prolonged period
of anoxia because perfusion was started immediately after cannulation
of the portal vein. The entire surgical procedure was conducted rapidly
(<20 min) to minimize the chance of an unsuccessful preparation. Liver
viability was assessed by: 1) measurement of the
O2 tension in the liver effluent using a biological
O2 monitor (YSI model 5300; YSI, Inc., Yellow Springs, OH)
equipped with a micro-O2 probe attached to a low-flow cell;
2) the physical appearance of the liver; and 3)
the maintenance of metoprolol steady-state concentrations.
Liver Perfusions.
The study was designed both to evaluate the apparent Michaelis-Menten
parameters and to measure the effect of a balanced mixture of amino
acids on apparent Vmax and
KM values for metoprolol and its metabolites.
Metoprolol in Krebs bicarbonate buffer was infused at 3 ml/min/g liver
into different livers at various concentrations over the range of
saturation of metabolism. Because the liver outlet steady-state
concentration reflected the velocity of metabolism or formation of each
metabolite, plots of velocity vs. concentration were
constructed from which the apparent Vmax and
KM for metoprolol and each metabolic pathway
were calculated. To assess the effect of amino acids, the metoprolol
and/or metabolite steady-state concentrations in the effluent were
perturbed by adding a balanced mixture of amino acids to the buffer.
Based on the measurement of portal venous concentration of
phenylalanine (250 µM) in dogs after a high-protein meal (14), a
balanced mixture of amino acids in the perfusion buffer-medium was
added as a dilution of Aminosyn* II 10% Amino Acids Injection (Abbott
Laboratories, Montreal, Canada). Final concentrations of
constituent amino acids in the inlet buffer were as previously
described (11). Velocity vs. concentration plots were again
constructed, and the apparent Vmax and
KM values in the presence of amino acids were
calculated for comparison with the preamino acid values. Once
steady-state liver outlet concentrations had been measured in the
presence of amino acids, buffer without amino acids was perfused again
to verify the reversibility of the amino acid interaction.
Time 0 was defined by the initiation of perfusion with Krebs
bicarbonate buffer containing metoprolol at 1 of the 6 inlet concentrations. A blank perfusion buffer sample was collected before
time 0; after time 0, perfusate samples were collected over 20 sec at
1, 2.5, 5, 7.5, 10 min, and then every 5 min until 130 min. After the
initial 50-min control phase, during which metoprolol and metabolite
concentrations in the effluent buffer had been at steady state for
~30 min, the perfusion medium was switched to Krebs buffer containing
amino acids in addition to metoprolol. The metoprolol and amino acids
perfusion was continued for 40 min, then the perfusion medium was
returned to the original Krebs buffer containing only metoprolol for 40 min. During the perfusion, the temperature and O2 tension
in the effluent were recorded when samples were taken. Samples of liver
effluent were analyzed for metoprolol and metabolites immediately
following each perfusion experiment.
Analytical Method.
After addition of the internal standard (50 µl of a 6.46 µM nadolol
solution in water) to 1.0 ml of the perfusate sample, 50 µl of the
mixed sample was directly injected onto the HPLC system consisting of a
Waters model 510 Pump, a Shimadzu model SIL-9A Autoinjector, a NovaPak
C18 8 × 100 mm, 4-µm particle size column, an ABI
Spectroflow 980 Programmable Fluorescence detector (excitation
wavelength 224 nm; no emission filter), and a Shimadzu model C-R3A
Chromatopac Integrator. The mobile phase consisted of
water:acetonitrile:triethylamine (91:9:0.3, v/v/v), adjusted to pH 3.0 with orthophosphoric acid and pumped at 3.0 ml/min. Run time was 22 min. The assay was validated according to generally accepted criteria
(15).
Data Analysis.
When the single-pass perfusion system was at steady-state, and the
distribution/binding processes of the drug in the liver preparation
were essentially complete, the rate of drug loss across the liver
equaled the rate of metoprolol elimination (i.e. the rate of metabolism). Css and
tss were defined as previously described (12).
The hepatic tissue binding of metoprolol was calculated from the
effluent steady-state metoprolol data as previously described (12).
Hepatic disposition of drug was described by standard steady-state mass
balance equations for drug elimination by an organ (16). Parameters
calculated from Cin,
Cout, and Q included E, CL (ml/min) and F. The 
(nmol/min/g liver) was
calculated from Q (ml/min/g liver),
Cin, and Cout (µM), and
the formation rates of -hydroxymetoprolol and
O-demethylmetoprolol from metoprolol (nmol/min/g liver) were
calculated from Q and the corresponding Cout,m as previously described (17).
The apparent Vmax and KM
values for metoprolol metabolism and each metabolic pathway were
estimated by nonlinear curve-fitting (PCNONLIN V4.0; SCI Software,
Lexington, KY) to the Michaelis-Menten equation using a weighting
factor of 1/. The Gauss-Newton with Levenberg and Hartly fitting
algorithm was used (18). Two different measures were used for
[S] in the liver water: 1)
Cout, which corresponds to liver concentration
according to the well-stirred model; and 2) the logarithmic
mean of the inlet and outlet metoprolol concentration (Cin 
Cout)/ln(Cin/Cout),
which is the average concentration of the drug in liver water according
to the parallel tube model (17).
A repeated-measures ANOVA was used to determine whether coinfusion of
amino acids significantly affected steady-state effluent concentrations
of metoprolol and its metabolites. Comparison of steady-state
concentrations in the effluent among the initial control phase (from 0 to 50 min), amino acid coinfusion phase (from 50 to 90 min) and after
the removal of amino acid phase (from 90 to 130 min), and comparisons
of the pharmacokinetic parameters between the initial control and amino
acid coinfusion conditions were made by paired t test
(p 
0.05).
O2 Uptake.
Although it was originally intended that O2 concentration
values in the effluent perfusate be used only for liver viability assessment, the measurements were appropriate to determine
O2 uptake by the livers over time. Given that the buffer
was equilibrated with 95% O2/5% CO2 and the
solubility of O2 in 37°C Ringers is 1.0755 µmol/ml
(19), the influent O2 concentration was ~1.02 µmol/ml.
The O2 monitor readout was in percentage saturation; the
influent buffer was calibrated to 100%. O2 content in the effluent buffer was thus calculated as the readout percentage saturation × 1.02 µmol/ml. For each liver, the mean
O2 content of the effluent buffer during each phase of the
experiment was calculated (i.e. 5-50 min, 55-90 min, and
95-130 min). O2 uptake during each phase was calculated as
where O2in is the influent concentration of
O2, O2out is the mean effluent concentration of
O2 during a particular phase of the experiment, and
Q is the buffer flow rate for a particular liver expressed
as ml/min/g liver.
We have observed that the O2 uptake rate varies
considerably between livers. Although the addition of metoprolol
increases O2 uptake slightly compared with blank buffer,
the effects of even high concentrations of drug were deemed to be
negligible compared with the interindividual variability of
O2 uptake. Therefore, for the purpose of comparing
O2 uptakes between the phases of the experiments, data from
all the livers were pooled. Data were compared using paired
t tests (p < 0.05).
| |
Results |
Assay Validation.
Separation of metoprolol (MW free base, 267.4) and its three
metabolites -hydroxymetoprolol (free base MW, 283.4),
O-demethylmetoprolol (free base MW, 253.4), and metoprolol
acid (zwitterion MW, 267.4) was achieved with no interferences from
endogenous compounds (fig. 2). Retention times for
-hydroxymetoprolol, O-demethylmetoprolol, metoprolol
acid, nadolol (internal standard, free base MW 309.4), and metoprolol
were 2.7, 3.2, 3.85, 5.55, and 15.5 min, respectively.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 2.
HPLC chromatograms of metoprolol and three
metabolites obtained after 50-µl injections of (A) drug-free
perfusate; (B) a spiked sample; and (C) an effluent perfusate sample
spiked with nadolol (internal standard).
Peak identities (retention times) are: 1,
-hydroxymetoprolol (2.70 min); 2,
O-demethylmetoprolol (3.20 min); 3, metoprolol acid (3.85 min); 4, nadolol (5.55 min); and 5,
metoprolol (15.5 min).
|
|
The calibration curve for metoprolol was based on the unweighted least
squares regression of the peak area ratios of metoprolol to nadolol
(internal standard) from 0.374 to 74.8 µM. Calibration curves for
-hydroxymetoprolol (0.48-11.9 µM),
O-demethylmetoprolol (0.51-10.3 µM), and metoprolol acid
(0.75-18.7 µM) were plotted as the peak height ratios of each
compound to nadolol against the corresponding concentrations of
analytes. Calibration curves were linear over the concentration ranges
examined, and their correlation coefficients were 
0.999. Blank
samples of effluent liver buffer showed no interfering peaks with the
analytes of interest. Variability was assessed by preparing replicate
samples of each analyte at three concentrations covering the range from the limit of quantitation. The validated limits of quantitation were
the lower concentrations of the ranges reported herein. For intraday
variability, six replicates were prepared on 1 day; for interday
variability, two replicates were prepared on six separate days.
Accuracy for the analytical method was within 5% of the nominal
concentration, except for the lowest concentration of metoprolol acid
(16%). In most cases, intraday and interday coefficients of variation
were <5%, except for metoprolol intraday variation (15%) and
metoprolol interday variation (11%) at their lowest concentrations.
Steady-State.
The time for the effluent concentrations of metoprolol,
-hydroxymetoprolol, and O-demethylmetoprolol to reach
steady-state were calculated for each of the six different inlet
concentrations of metoprolol (table 1). Except in one
case, the average time for metoprolol to reach steady-state was <20
min, regardless of the metoprolol inlet concentrations; the effluent
concentrations of -hydroxymetoprolol and
O-demethylmetoprolol also quickly reached steady-state. The
concentration of the metabolite metoprolol acid continued to rise in
the effluent and did not reach steady-state during the perfusion.
Hepatic Tissue Binding of Metoprolol.
Metoprolol hepatic tissue binding (nmol/g liver), calculated from the
metoprolol concentrations in the effluent over the time 0-tss, increased as the metoprolol inlet
concentration was increased (table 1). But, when the metoprolol inlet
concentration was below 7.48 µM, the tissue binding remained
relatively constant.
Metoprolol and Its Metabolites in the Effluent.
Table 2 lists the steady-state concentrations of
metoprolol, -hydroxymetoprolol and O-demethylmetoprolol
in the effluent during the three phases of liver perfusion, including
"Pre-Amino Acids" (from 0 to 50 min as the control), "Amino
Acids" (from 50 to 90 min), and "Post-Amino Acids" (from 90 to
130 min).
View this table:
[in this window]
[in a new window]
|
TABLE 2
Steady-state metoprolol, -hydroxymetoprolol, and
O-demethylmetoprolol concentrations (mean ± SD, µM) in the
effluent at each concentration
|
|
At all six metoprolol inlet concentrations, the metoprolol effluent
concentrations during amino acid coinfusion were significantly higher
(range: 23-97%) than during the control phase (0-50 min). The
increases were relatively larger at low than at high inlet concentrations (table 2). Post-amino acid effluent concentrations returned toward pre-amino acid levels, but did not seem to recover completely, although only in the cases of 5.61 and 7.48 µM inlet concentrations were there significant differences between the post- and
pre-amino acid periods.
For -hydroxymetoprolol, the effluent concentrations during amino
acid coinfusion were significantly lower (range: 26-54%) than during
the control period at all six metoprolol inlet concentrations (table
2). The decreases were relatively larger at high metoprolol inlet
concentrations than at lower concentrations. Except for when the
metoprolol inlet concentrations were 4.49 and 5.61 µM, -hydroxymetoprolol concentrations rose significantly when amino acid
coinfusion was terminated. The post-amino acid concentrations returned
to within 8% to 19% of pre-amino acid level, remaining significantly
lower than the pre-amino acid effluent levels, except when the
metoprolol inlet concentrations were 3.74 and 4.49 µM. This failure
to return completely to the pre-amino acid levels is consistent with
the observation of slowly declining effluent concentrations of
-hydroxymetoprolol in pilot studies with livers not coinfused with
amino acids.
O-demethylmetoprolol effluent concentrations decreased
significantly (15-48%) when metoprolol was coinfused with the amino acids, for all metoprolol inlet concentrations except 3.74 and 5.61 µM. The decreases were relatively larger at higher metoprolol inlet
concentration (table 2). After removal of amino acids, the
O-demethylmetoprolol concentrations increased toward control levels.
Effect of Amino Acids on Pharmacokinetic Parameters.
Considering only the 50-min control period in all the livers,
metoprolol E and CL decreased by >60% as the
inlet concentrations of metoprolol were increased from 3.74 to 44.9 µM (table 3). Similarly, over the range of inlet
concentrations, the formation CL of -hydroxymetoprolol
and O-demethylmetoprolol decreased by 62 and 40%,
respectively. F increased by 167% as the inlet
concentration of metoprolol was increased from 3.74 to 44.9 µM. When
metoprolol inlet concentrations were between 3.74 and 5.61 µM, the
parameters were similar. But, when the metoprolol inlet concentrations
were 7.48 µM or more, large decreases in E and
CL, as well as marked increases in F, occurred.
View this table:
[in this window]
[in a new window]
|
TABLE 3
Pharmacokinetic parameters of metoprolol and metabolites at each
concentration in the absence (control) and presence (AA) of amino
acidsa
|
|
During the amino acid coinfusion period, E and CL
of metoprolol decreased by 72%, as the inlet concentration of
metoprolol was increased from 3.74 to 44.9 µM (table 3). Similarly,
the formation CL of -hydroxymetoprolol and
O-demethylmetoprolol decreased by 73% and 68%,
respectively. Metoprolol F in the presence of the amino
acids increased by 67% as the inlet concentration was increased from
3.74 to 44.9 µM. As with the control period, changes in these
parameters were greater when metoprolol inlet concentrations exceeded
7.48 µM.
Amino acid coinfusion decreased metoprolol E and
CL by 35-54%, compared with the control phase, over the
range of inlet concentrations. Similarly, the formation CL
of -hydroxymetoprolol decreased significantly by 28-55%.
O-demethylmetoprolol formation CL decreased by
17-50%, but only at metoprolol inlet concentrations exceeding 5.61 µM. Amino acids increased F by 21-93% over the six inlet
concentrations. The increase in F was greater at lower
metoprolol inlet concentrations.
Estimation of Apparent Vmax and
KM and the Effect of Amino Acids.
Amino acids significantly decreased the apparent
Vmax values for metoprolol,
-hydroxymetoprolol, and O-demethylmetoprolol (table
4). Metoprolol apparent Vmax
decreased by 50% if substrate concentrations were calculated based on
the well-stirred model and by 55% if based on the parallel tube model.
Similarly, the apparent Vmax value for
-hydroxymetoprolol formation decreased by 53% (well-stirred) or
56% (parallel tube). The apparent Vmax of
O-demethylmetoprolol decreased by 53% (well-stirred) or
61% (parallel tube).
View this table:
[in this window]
[in a new window]
|
TABLE 4
Apparent Vmax (nmol/min/g liver) and KM (nM)
for metoprolol metabolism, -hydroxylation, and O-demethylation
estimated by nonlinear curve-fitting based on the assumptions of the
well-stirred and parallel tube models
|
|
There were no consistent changes in the apparent
KM in the presence of the amino acids. In 4 of
the 6 cases, the apparent KM values declined in
the presence of the amino acids, but the decreases were not
statistically significant. The standard errors of the
KM estimates were high; this would have
prevented detection of any real differences between the control and
amino acid conditions. The assumptions of the parallel tube model
resulted in slightly higher estimated Vmax and
KM values than those of the well-stirred model.
A weighting factor of 1/ was used in the curve-fitting because the
velocity of metabolism varied more between livers at higher than at
lower concentrations (figs. 3 and 4).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 3.
Curve fits of rate of metoprolol metabolism
vs. outlet metoprolol concentration without and with amino acid (A.A.)
coinfusion ( , metoprolol only;  , metoprolol + amino acids),
based on the assumptions of the well-stirred model.
|
|
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 4.
Curve fits of rate of metoprolol metabolism
vs. logarithmic average metoprolol concentration without and with amino
acid (A.A.) coinfusion (, metoprolol only; , metoprolol + amino acids), based on the assumptions of the parallel tube model.
|
|
O2 Uptake.
In every liver, O2 uptake increased upon coinfusion of
amino acids so that the effluent buffer was almost depleted. The mean O2 content of the effluent buffer declined from 0.166 ± 0.089 µmol/ml during the initial phase of the experiments to
0.053 ± 0.058 µmol/ml during amino acid coinfusion (compared
with the inlet concentration of 1.02 µmol/ml). After the amino acid
coinfusion was stopped, the mean effluent buffer O2 levels
reached a peak of 0.37 µmol/ml at 110 min, then declined to 0.31 µmol/ml at the end of the experiment (130 min).
The mean buffer flow rate was 3.47 ± 0.38 ml/min/g liver; the
O2 delivery rate to the livers thus averaged 3.54 ± 0.39 µmol/min/g liver. The mean O2 uptake was 2.96 ± 0.44 µmol/min/g liver during the pre-amino acid phase, increasing
to 3.36 ± 0.44 µmol/min/g liver during amino acid coinfusion.
The mean O2 uptake after amino acid coinfusion averaged
2.56 ± 0.73 µmol/min/g liver. Paired t tests
indicated that the increase in O2 uptake with amino acid coinfusion and the return toward control values after amino acid coinfusion were statistically significant (p < 0.05).
| |
Discussion |
Analytical Method.
Direct-injection HPLC methods with fluorescence detection using
C18 columns have been used for the measurement of
metoprolol and metabolite concentrations in human urine (20, 21), but neither method detected O-demethylmetoprolol. A method that
allowed separation and detection of all three major metabolites has
been described with conditions suitable for urine and microsomal
preparations (22). This method was used as the basis for development of
the present method. The conditions were modified to allow good
separation of the analytes from interfering peaks present in directly
injected rat liver perfusate (fig. 2). The present method also provides the appropriate compromise in sensitivity to allow quantitation of
metoprolol and its metabolites when metoprolol is present at up to
50-fold higher concentration than one of the metabolites.
The Model System.
Our study with propranolol revealed that a balanced mixture of amino
acids could globally inhibit its metabolism, but revealed little about
the mechanism of this inhibition (11). To gain more mechanistic
information, our approach was to determine whether the inhibition
occurs as a result of changes to the affinity or capacity (or both) of
the metabolic enzymes. This required determination of the apparent
KM (affinity) and Vmax
(capacity) values of the metabolic pathways. To be able to observe both
direct and indirect mechanisms, we conducted this study using isolated,
perfused rat livers.
A weakness of propranolol as a model drug is that tissue binding
prevents steady-state from being reached, except at very high,
saturating inlet concentrations. The hepatic tissue binding of
metoprolol was at least 6-fold lower than the 2 µmol/g liver reported
for propranolol and varied with inlet concentration (table 1) (12, 23).
Even at nonsaturating inlet concentrations as low as 3.74 µM,
steady-state was reached within 20 min (table 1). Thus, studies at
steady-state could be conducted at concentrations covering the range of
saturation of metabolism. This is an important advantage of metoprolol
as a model drug.
Metoprolol is extensively metabolized in the human liver with ~85%
of the administered dose recovered in urine as unchanged metoprolol
(10%), -hydroxymetoprolol (10%), and metoprolol acid (65%) (24)
(fig. 1). In rats, ~95% of a 100 mg/kg (374 µmol/kg) oral dose was
recovered in urine as unchanged metoprolol (4.5%), -hydroxymetoprolol (25%), O-demethylmetoprolol (3.5%),
and metoprolol acid (62%) (25). Thus, the metabolism of metoprolol in
rats is similar to what is observed in humans. Metoprolol acid is a secondary metabolite, occurring as an oxidation product of
O-demethylmetoprolol (26). Therefore, in addition to the
parent drug, all three of these metabolites were measured in the liver
effluent. During the initial 50-min control phase, metoprolol,
-hydroxymetoprolol, and O-demethylmetoprolol approached
steady-state levels in the effluent within 20 min, regardless of the
inlet concentration used (table 1). Metoprolol acid concentrations,
however, continued to rise throughout the perfusion experiments for
unknown reasons. Therefore, it was excluded from the calculations in
this study. We speculate that because this metabolite is more polar
than the others (it is a form of amino acid and therefore exists as a
zwitterion), it may not readily cross membranes into the effluent.
Alternatively, it may be secreted into the bile, which was not
collected separately in these experiments.
Michaelis-Menten Kinetics of Metoprolol.
E declined by 62% from 0.73 to 0.28 when the metoprolol
inlet concentration was increased from 3.74 to 44.9 µM, thus
indicating that the inlet concentrations spanned much of the range of
saturation of metabolism (table 3). The experimental design, whereby
different livers were used for each point of the vs.
[S] curve gives rise to a lack of precision in the
measurements not seen with microsomal preparations. Differences between
rat liver preparations resulted in scattered data, especially at lower
metoprolol inlet concentrations (figs. 3 and 4); thus, the
computer-processed curve-fitting did not converge on precise values for
apparent KM (table 4). The mean apparent
KM values 3.6 (well-stirred) and 8.8 (parallel
tube) µM are lower than the apparent KM value
of 39 µM for metoprolol metabolism by microsomes reported by
Arfwidsson et al. (25), likely due to differences in
the experimental conditions. The Vmax values
from microsomes and the perfused liver preparation cannot be directly
compared, but if a microsomal protein content of 45 mg protein/g liver
is assumed (27), then the reported Vmax for
microsomes of 1.28 nmol/mg protein/min may be converted to a value of
~58 nmol/min/g liver. The estimates from the perfused livers compared
favorably at 38 (well-stirred) and 46 (parallel tube) nmol/min/g liver
(table 4). Using the same conversion factor of 45 mg protein/g liver,
the apparent KM and Vmax
values for -hydroxylation were 18 µM and 13 nmol/min/g liver and
for O-demethylation were 18 µM and 7.7 nmol/min/g liver in
rat liver microsomes (24). These values compare more favorably with
this study than the parent drug data (table 4). The parallel tube model
gave somewhat higher estimates of apparent KM
and Vmax than did the well-stirred model; these
higher estimates were in general in closer agreement with the
microsomal data. It should be noted that the values for
KM and Vmax may depend to
some extent on the range of concentrations of drug used in the
experiments, because in humans both high-affinity, low-capacity and
low-affinity, high-capacity enzyme systems for each pathway have been
identified (28). These may also exist in the rat, although they have
not yet been reported.
Effect of Amino Acids on Metoprolol Metabolism.
Meal protein is digested and transported into the portal circulation as
amino acids (14), so they were used in the perfusion buffer to mimic a
high-protein meal. In pilot studies, a mixture of amino acids inhibited
metoprolol metabolism similarly to propranolol metabolism (11). Whereas
the main objective of this study was to determine whether amino acids
would affect the apparent KM and/or
Vmax of metoprolol, a second objective was
to determine the sensitivity of metoprolol availability to amino acids
at nonsaturating vs. saturating concentrations of
metoprolol.
The increase in effluent metoprolol concentrations, coupled with the
corresponding reductions in -hydroxymetoprolol and
O-demethylmetoprolol levels indicated that metoprolol
metabolism was inhibited (table 2); no apparent effect of the amino
acids on hepatic tissue binding of metoprolol was observed, because the
new steady-state was reached immediately upon coinfusion of amino
acids. The recovery toward pre-amino acid effluent concentrations of
metoprolol and its metabolites indicated that the amino acid effect was
reversible. The -hydroxymetoprolol levels failed to recover
completely after removal of the amino acids; however, under control
conditions, its concentration in the effluent slowly declined over the
course of the perfusion. The effect of amino acids was temporary, so it
was likely that no irreversible alteration or damage to the hepatic
metabolic enzyme system occurred.
Nonlinear curve-fitting of metabolism/formation velocity vs.
metoprolol concentration in the liver water revealed that amino acids
reduced Vmax by ~50%, but had no significant
effect on the apparent KM of metoprolol and its
two metabolic pathways (table 4). Therefore, amino acids can reversibly
reduce the capacity for oxidative metabolism, ruling out competitive
inhibition as a mechanism. With competitive inhibition, changes to
KM only would have been observed. It was evident
that the hepatic elimination model used in the curve-fitting did not
influence the estimated effect of the amino acids on the apparent
Vmax and KM (table 4, figs. 3 and 4).
In an earlier study, we observed only a small amino acid-mediated
increase in F of propranolol when it was infused at
saturating concentrations, but predicted a larger increase if we had
been able to observe the system at low (nonsaturating) concentrations (11). A principal advantage of metoprolol as a model drug is its low
tissue binding, because we were able to observe the interaction at low,
more clinically relevant steady-state concentrations. Over the entire
range of inlet concentrations, coinfused amino acids caused significant
decreases in metoprolol E and CL, as well as the
formation clearances of -hydroxymetoprolol and
O-demethylmetoprolol. At the highest inlet concentration,
the observed increase in F was only 21%, similar to the
observations with propranolol, but at the lowest inlet concentration,
F increased by 93% from 0.27 to 0.53 (3.74 µM, table 3).
Across the group of drugs affected by food, the average increase in AUC
may be <50%, but increases of >100% may occur in individual
patients. None of the mechanisms proposed to date has been shown to
have the capability to produce sufficiently large F changes
to explain the entire effect (29). The high sensitivity of F
at low inlet concentrations of metoprolol in the perfused rat liver
shows that metabolic inhibition under conditions that have relevance
for the human clinical situation can cause increases in bioavailability
similar in magnitude to those observed in humans.
Possible Mechanisms for the Effect of Amino Acids.
This study indicates that the inhibition of metoprolol metabolism in
the perfused rat liver during coinfusion of amino acids occurs through
a reduction in the capacities of the oxidative pathways, implying that
the dominant component of the mechanism may be direct noncompetitive or
uncompetitive inhibition of cytochrome P450, such as by allosteric
effects, or it may be indirect, such as by limiting the supply of
cosubstrates for oxidation, NADPH, or O2.
No reports were found relating to direct interactions between amino
acids and cytochrome P450 enzymes. These should be observable in
microsomal systems, however. Direct interactions would be expected to
be somewhat specific toward enzyme forms rather than global, however,
and all of the pathways of both metoprolol and propranolol are
similarly affected (11). Although more than one enzyme seems to be
involved in each metabolic pathway of metoprolol metabolism, our
observations on metoprolol could be attributable to a single high-affinity, low-capacity enzyme system. In human microsomal experiments, CYP2D6 catalyzes both -hydroxylation and
O-demethylation with similar apparent
KM values to those observed in the present rat
liver perfusion study, whereas the low-affinity, high-capacity pathway(s) were linear in the concentration range used in the present
study and could remain unaffected (30). Direct inhibition of a single
CYP enzyme is not consistent with our previous observations on
propranolol, however. Not only ring-hydroxylation, but
N-dealkylation and deamination are inhibited by amino acids;
these metabolic reactions seem to be catalyzed by different enzymes
(31).
A continuous supply of NADPH is required as a cofactor for cytochrome
P450-dependent drug oxidation. NADPH is generated by multienzyme
systems that exist in several intracellular compartments. Indirect
regulation of metabolism by limitation of the supply of NADPH may occur
in vivo, and this may be an important control mechanism in
drug oxidation since maintenance of the redox state of the NADP: the
NADPH couple is a highly regulated process in intact cells (32). In
fasted rats, competing reactions for substrate and cofactor play a
major role in determining the availability of NADPH for mixed-function
oxidation. In this study, however, the ample supply of glucose in the
perfusion buffer (11 mM) should have provided sufficient NADPH
via the pentose phosphate shunt to support cytochrome P450
activity. It is uncertain whether the amino acid mixture used would
affect the redox state of the hepatocytes, because metabolic reactions
involving both production and utilization of reducing equivalents may
occur. In addition, the production of reduced NADPH and NADH is
indirectly dependent on intracellular O2 tension (33). When
O2 tension is low, the reduced form may be favored.
Although the NADPH supply could have been limiting, the conditions used
seem unlikely to cause its depletion. Other mechanisms of metabolic
inhibition are more plausible, but this mechanism cannot yet be ruled
out.
Among the reports related to the food effect on high first-pass drugs,
we have encountered none indicating that nutrients may cause sufficient
depletion of hepatic O2 that it could be a limiting
cofactor for drug metabolism. Nevertheless, coadministration of amino
acids caused both decreased Vmax values for the
parent drug and metabolites and an increase in hepatic O2
consumption sufficient to deplete almost completely the buffer of
O2 in every liver. O2 limitation could have a
role within the hepatocyte if other reactions with higher affinities
were to compete successfully with cytochrome P450 for O2. A
much larger effect could be expected, however, if O2 were
to be used before it reached the cells containing the drug-metabolizing
enzymes. Thus, the distribution of enzymes within the zones of the
hepatic acinus may be an important determinant of the amount of
O2 reaching the cells containing the drug-metabolizing enzymes. It is known that an O2 gradient exists between the
portal and central venous sides of the liver lobule (33). The influent O2 concentration in vivo is ~85 µM; the
difference across the lobule is ~50 µM. Handling of amino acids,
including transport into the cells and gluconeogenesis, is an
O2-consuming process that is conducted predominantly by the
hepatocytes in the periportal zone of the acinus (34). The cytochrome
P450 enzymes that perform drug metabolism are distributed downstream,
mainly in the pericentral region of the acinus (32, 35). Therefore,
when both amino acids and drug are present in the portal vein, the
location of the metabolic enzymes would dictate that amino acid
metabolism has priority, and this process may deplete O2
supplies before sinusoidal blood reaches the pericentral zone.
Propranolol clearance is less during antegrade perfusion than during
retrograde perfusion of rat livers under hypoxic conditions, indicating
that propranolol metabolism is sensitive to the local O2
concentration in the pericentral region where most of the oxidative
enzymes are located (36). It is important to note that this observation
was made under slightly hypoxic conditions, wherein no extraoxidative
load other than by drug metabolism was placed on the liver. Therefore
the reported data cannot indicate whether increasing O2
utilization in the periportal region would increase the sensitivity of
propranolol clearance to hypoxia. Although it seems likely that placing
an oxidative load upstream from a sensitive system could inhibit metabolism through O2 depletion, this remains to be
demonstrated.
Although the conditions used in this study were not designed to
illustrate the effect of O2 depletion by amino acids, they are comparable with those used in other liver perfusion studies. The
pre-amino acid O2 consumption of ~3 µmol/min/g liver is
somewhat higher than the 2.3 µmol/min/g liver reported to be usual in
liver perfusion studies (37). This figure is also higher than the 1.5-2.0 µmol/min/g liver rate of O2 consumption reported
in an early study on the effect of hypoxia on propranolol elimination (38), but somewhat lower than the 3.5 µmol/min/g liver consumption reported by the same group in a more recent study on the sensitivity of
propranolol elimination to hypoxia (39). In the latter study, authors
indicate that a higher O2 delivery rate (6.4 µmol/min/g liver) is more physiological than lower rates (we used 3.5 µmol/min/g liver). They observed an immediate, direct, and reversible relationship between O2 delivery and propranolol clearance below a
delivery rate of 6 µmol/min/g liver. In our study, amino acids caused
almost complete O2 extraction from the buffer, and the
studies cited herein indicate that it is likely that most of the
O2 was extracted upstream from the enzymes responsible for
metoprolol metabolism.
As yet, we cannot be sure how relevant our results may be to the
in vivo situation in humans, but studies in humans indicate that meal protein (in contrast to meal carbohydrate or fat) results in
O2 depletion from the hepatic effluent, despite a large
increase in hepatic blood flow (40, 41). This is consistent with the general observation that the food effect on high first-pass drugs is
associated particularly with high protein meals. Under these circumstances, the reduced hepatic arterial flow and enhanced oxidative
metabolism of amino acids in the periportal cells almost certainly led
to reduced O2 availability in the perivenous zone, wherein
most of the cytochrome P450 activity resides. Sensitivity of human
hepatocytes and their cytochrome P450 activities to reduced O2 levels is unknown. Nevertheless, data from rats indicate
that propranolol clearance is very sensitive to O2 levels
(39), and the maximum effect of a protein meal on hepatic
O2 consumption occurred close to the time of peak drug
levels in the liver (10).
An amino acid-mediated decrease in metabolic capacity of metoprolol and
other affected drugs, such as propranolol, is consistent with many, but
not all, of the observations related to the food effect. In particular,
recent observations that small increases in the propranolol AUC were
observed in both humans and dogs when they were shown food but not
allowed to ingest it indicate that other mechanisms may contribute to
the interaction (42). It has been stated on numerous occasions that the
food effect seems to be complex. We propose that one or more of the
mechanisms discussed herein could contribute to this phenomenon.
Further experimentation will be required to establish the importance of
each.
We thank Astra Hässle for supplying metoprolol and metabolites
for use in this study, and Dr. Dion Brocks for his helpful comments on
the manuscript.
This study was supported by the Medical Research Council of
Canada, Development Grant DG-405 and by the Heart and Stroke
Foundation of Saskatchewan. This work has been presented in part as an
abstract [ISSX Proc. 8, 311 (1995)].
| 1.
|
A. Melander,
D. Lalka, and
A. McLean:
Influence of food on the presystemic metabolism of drugs.
Pharmacol. Ther.
38,
253-267 (1988)[Medline].
|
| 2.
|
P. Du Souich,
H. Maurice, and
L. Heroux:
Contribution of the small intestine to the first-pass uptake and systemic clearance of propranolol in rabbits.
Drug. Metab. Dispos.
23,
279-284 (1995)[Abstract].
|
| 3.
|
A. J. McLean,
P. J. McNamara,
P. duSouich,
M. Gibaldi, and
D. Lalka:
Food, splanchnic blood flow, and bioavailability of drugs subject to first-pass metabolism.
Clin. Pharmacol. Ther.
24,
5-10 (1978)[Medline].
|
| 4.
|
C. K. Svensson,
D. J. Edwards,
P. M. Mauriello,
S. H. Barde,
A. C. Foster,
R. A. Lanc,
E. Middleton, and
D. Lalka:
Effect of food on hepatic blood flow: implications in the "food effect" phenomenon.
Clin. Pharmacol. Ther.
34,
316-323 (1983)[Medline].
|
| 5.
|
M. W. Modi,
J. M. Hassett, and
D. Lalka:
Influence of posture on hepatic perfusion and the presystemic biotransformation of propranolol: simulation of the food effect.
Clin. Pharmacol. Ther.
44,
268-274 (1988)[Medline].
|
| 6.
|
J. Feely,
J. Nadeau, and
A. J. J. Wood:
Effects of feeding on the systemic clearance of indocyanine green and propranolol blood concentrations and plasma binding.
Br. J. Clin. Pharmacol.
15,
383-385 (1983)[Medline].
|
| 7.
|
H. Liedholm and
A. Melander:
Concomitant food intake can increase the bioavailability of propranolol by transient inhibition of its presystemic primary conjugation.
Clin. Pharmacol. Ther.
40,
29-36 (1986)[Medline].
|
| 8.
|
A. J. McLean,
C. Isbister,
A. Bobik,
B. Pharm, and
F. J. Dudley:
Reduction of first-pass hepatic clearance of propranolol by food.
Clin. Pharmacol. Ther.
30,
31-34 (1981)[Medline].
|
| 9.
|
H. Liedholm,
B. E. Wahlin, and
A. Melander:
Mechanisms and variations in the food effect on propranolol bioavailability.
Eur. J. Clin. Pharmacol.
38,
469-475 (1990)[Medline].
|
| 10.
|
H. A. Semple,
Y. K. Tam, and
R. T. Coutts:
A computer simulation of the food effect: transient changes in hepatic blood flow and Michaelis-Menten parameters as mediators of hepatic first pass metabolism and bioavailability of propranolol.
Biopharm. Drug Dispos.
11,
61-76 (1990)[Medline].
|
| 11.
|
H. A. Semple and
F. Xia:
Interaction between propranolol and amino acids in the single-pass isolated, perfused rat liver.
Drug Metab. Dispos.
24,
794-798 (1995).
|
| 12.
|
H. A. Semple and
F. Xia:
Propranolol disposition in the single-pass isolated, perfused rat liver in the presence and absence of insulin and glucagon.
Drug Metab. Dispos.
22,
822-826 (1994)[Abstract].
|
| 13.
|
L. L. Miller:
Technique of isolated rat liver perfusion. In
"Isolated Liver Perfusion and Its Applications" (I. Bartosek,
A. Guaitani and
L. L. Miller, eds.), pp. 11-52. Raven Press, New York, 1973.
|
| 14.
|
D. H. Elwyn:
The role of the liver in regulation of amino acid and protein metabolism. In
"Mammalian Protein Metabolism" (H. N. Munro, ed.), pp. 523-557. Academic Press, New York, 1970.
|
| 15.
|
V. P. Shah,
K. K. Midha,
S. Dighe,
I. J. McGilveray,
J. P. Skelly,
A. Yacobi,
T. Layloff,
C. T. Viswanathan,
C. E. Cook,
R. D. McDowall,
K. A. Pittman, and
S. Spector:
Analytical methods validation: bioavailability, bioequivalence, and pharmacokinetic studies.
J. Pharm. Sci.
81,
309-313 (1992).
|
| 16.
|
M. Rowland and
T. Tozer:
"Clinical Pharmacokinetics." Williams & Wilkins, Baltimore, 1995.
|
| 17.
|
K. S. Pang and
M. Rowland:
Hepatic clearance of drugs. I. Theoretical considerations of a "well-stirred" model and a "parallel tube" model. Influence of hepatic blood flow, plasma and blood cell binding, and the hepatocellular enzyme activity on hepatic drug clearance.
J. Pharmacokinet. Biopharm.
5,
625-653 (1977)[Medline].
|
| 18.
| M. Davies and I. J. Whitting: A modified form of Levenberg's
correction. In "Numerical Methods for Non-linear Optimization,"
chap. 12. Academic Press, New York, 1972.
|
| 19.
| "YSI Model 5300 Biological Oxygen Monitor Instruction Manual."
Yellow Springs Instrument Company, Inc., Yellow Springs, OH.
|
| 20.
|
K. Balmér and
Y. Zhang:
Lagerström, and B. E. Persson: Determination of metoprolol and two major metabolites in plasma and urine by column liquid chromatography and fluorometric detection.
J. Chromatogr.
417,
357-365 (1987)[Medline].
|
| 21.
|
J. Godbillon and
M. Duval:
Determination of two metoprolol metabolites in human urine by high performance liquid chromatography.
J. Chromatogr.
309,
198-202 (1984)[Medline].
|
| 22.
|
M. S. Lennard:
Quantitative analysis of metoprolol and three of its metabolites in urine and liver microsomes by high-performance liquid chromatography.
J. Chromatogr.
342,
199-205 (1985)[Medline].
|
| 23.
|
J. H. Anderson,
R. C. Anderson, and
L. S. Iben:
Hepatic uptake of propranolol.
J. Pharmacol. Exp. Ther.
206,
172-180 (1978)[Abstract/Free Full Text].
|
| 24.
|
M. S. Lennard,
H. K. Crewe,
G. T. Tucker, and
H. F. Woods:
Metoprolol oxidation by rat liver microsomes: inhibition by debrisoquine and other drugs.
Biochem. Pharmacol.
35,
2757-2761 (1986)[Medline].
|
| 25.
|
A. Arfwidsson,
K. O. Borg,
K.-J. Hoffmann, and
I. Skanberg:
Metabolism of metoprolol in the rat in vitro and in vivo.
Xenobiotica
6,
691-711 (1976)[Medline].
|
| 26.
|
K. O. Borg,
E. Fellenius,
R. Johansson, and
M. Wallborg:
Pharmacokinetic studies of metoprolol-(3H) in the rat and the dog.
Acta Pharmacol. Toxicol.
36,
104-115 (1975)[Medline].
|
| 27.
|
J. B. Houston:
Utility of in vitro drug metabolism data in predicting in vivo metabolic clearance.
Biochem. Pharmacol.
47,
1469-1479 (1994)[Medline].
|
| 28.
|
S. V. Otton,
H. K. Crewe,
M. S. Lennard,
G. T. Tucker, and
H. F. Woods:
Use of quinidine inhibition to define the role of the sparteine/debrisoquine cytochrome P450 in metoprolol oxidation by human liver microsomes.
J. Pharmacol. Exp. Ther.
247,
2427 (1988).
|
| 29.
|
Y. K. Tam:
Individual variation in first-pass metabolism.
Clin. Pharmacokinet.
25,
300-328 (1993)[Medline].
|
| 30.
|
M. Kim,
D. D. Shen,
A. C. Eddy,
W. L. Nelson, and
L. K. Roskos:
Inhibition of the enantioselective oxidative metabolism of metoprolol by verapamil in human liver microsomes.
Drug Metab. Dispos.
21,
309-317 (1993)[Abstract].
|
| 31.
|
S. Fujita,
S. Umeda,
Y. Funae,
S. Imaoka,
H. Abe,
R. Ishida,
T. Adachi,
M. Masuda,
A. Kazusaka, and
T. Suzuki:
Regio- and stereoselective propranolol metabolism by 15 forms purified cytochromes P-450 from rat liver.
J. Pharmacol. Exp. Ther.
264,
226-233 (1993)[Abstract/Free Full Text].
|
| 32.
|
R. G. Thurman,
F. C. Kauffman, and
J. Baron:
Biotransformation and zonal toxicity. In
"Regulation of Hepatic Metabolism" (R. G. Thurman,
F. C. Kauffman and
K. Jungermann, eds.), pp. 321-382. Plenum, New York, 1986.
|
| 33.
|
R. G. Thurman,
S. Ji, and
J. J. Lemasters:
Lobular oxygen gradients: possible role in alcohol-induced hepatotoxicity. In
"Regulation of Hepatic Metabolism" (R. G. Thurman,
F. C. Kauffman and
K. Jungermann, eds.), pp. 293-320. Plenum, New York, 1986.
|
| 34.
|
S. Seifter and
S. Englard:
Energy metabolism. In
"The Liver: Biology and Pathobiology" (I. M. Arias, ed.), pp. 323-364. Raven Press Ltd., New York, 1994.
|
| 35.
|
K. S. Pang and
J. A. Terrell:
Retrograde perfusion to probe the heterogeneous distribution of hepatic drug metabolizing enzymes in rats.
J. Pharmacol. Exp. Ther.
216,
339-346 (1981)[Abstract/Free Full Text].
|
| 36.
|
S. L. Elliott,
D. J. Morgan,
P. W. Angus,
H. Ghabrial,
R. G. Watson, and
R. A. Smallwood:
The effect of hypoxia on propranolol clearance during antegrade and retrograde flow in the isolated perfused rat liver preparation.
Biochem. Pharmacol.
45,
573-578 (1993)[Medline].
|
| 37.
|
K. S. Pang:
Liver perfusion studies in drug metabolism and drug toxicity. In
"Drug Metabolism and Drug Toxicity" (J. R. Mitchell and
M. G. Horning, eds.), pp. 331-352. Raven Press, New York, 1984.
|
| 38.
|
D. B. Jones,
G. W. Mihaly,
R. A. Smallwood,
L. K. Webster,
D. J. Morgan, and
N. P. Madsen:
Differential effects of hypoxia on the disposition of propranolol and sodium taurocholate by the isolated perfused rat liver.
Hepatology
4,
461-466 (1984)[Medline].
|
| 39.
|
S. L. Elliott,
D. J. Morgan,
P. W. Angus,
H. Ghabrial, and
R. A. Smallwood:
Sensitivity of propranolol elimination to hypoxia in the isolated perfused rat liver preparation.
Clin. Exp. Pharmacol. Physiol.
20,
27-33 (1993)[Medline].
|
| 40.
|
T. Brundin and
J. Wahren:
Influence of protein ingestion on human splanchnic and whole-body oxygen consumption, blood flow, and blood temperature.
Metabolism
43,
626-632 (1994)[Medline].
|
| 41.
|
M. Dauzat,
M. Lafortune,
H. Patriquin, and
G. Pomier Layrargues:
Meal induced changes in hepatic and splanchnic circulation: a noninvasive Doppler study in normal humans.
Eur. J. Appl. Physiol.
68,
373-380 (1994).
|
| 42.
|
J. M. Power,
D. J. Morgan, and
A. J. McLean:
Effects of sensory (teasing) exposure to food on oral propranolol bioavailability.
Biopharm. Drug Dispos.
16,
579-589 (1995)[Medline].
|
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Abstract
Introduction
Abstract
