Download intestinal metabolism: the role of enzyme localization in phenol

0090-9556/02/3005-586–594$7.00
DRUG METABOLISM AND DISPOSITION
Copyright © 2002 by The American Society for Pharmacology and Experimental Therapeutics
DMD 30:586–594, 2002
Vol. 30, No. 5
575/975194
Printed in U.S.A.
INTESTINAL METABOLISM: THE ROLE OF ENZYME LOCALIZATION IN PHENOL
METABOLITE KINETICS
PRAJAKTI A. KOTHARE1
CHERYL L. ZIMMERMAN
AND
Department of Pharmaceutics, University of Minnesota, Minneapolis, Minnesota
(Received September 12, 2001; accepted January 18, 2002)
This article is available online at http://dmd.aspetjournals.org
ABSTRACT:
ance of phenol was 1.11 ml/min. Sulfation was the predominant
metabolic pathway after vascular administration of phenol. After
luminal dosing, the intestinal intrinsic clearances of phenol at the
low and high doses were 7.29 ⴞ 1.39 (n ⴝ 4) and 3.55 ⴞ 1.16 ml/min
(n ⴝ 3), respectively, indicating saturation at the higher dose.
Moreover, there was a decrease in the area under the curve ratio
(metabolite/phenol) at the high luminal dose. Luminal administration, in general, produced greater glucuronidation. These data and
STELLA simulations suggest that enzyme localization at both the
cellular and tissue levels has a significant influence on intestinal
metabolism.
Physiological models of hepatic extraction have been used quite
successfully to predict in vivo clearance from in vitro data. The well
stirred model (Pang and Rowland, 1977) is most frequently used,
owing in part to its simplicity. However, a similar application of the
well stirred model to the intestine was less successful (Thummel et al.,
1997). A possible reason for its failure might be the polarized expression of drug-metabolizing enzymes within the intestine (Chowdhury
et al., 1985; Watkins, 1997). Most enzymes involved in intestinal
elimination, such as the UDP-glucuronosyltransferases (UGTs2), cytochrome P450 3A (CYP3A), and efflux pumps such as P-glycoprotein, are expressed in the epithelial cells (Chowdhury et al., 1985;
Watkins, 1997). On the other hand, the vasculature (“serosa”) is
devoid of significant enzymatic activity. This seems to refute the
premise of the intestine as a well stirred organ.
Because the intestine can receive compounds luminally and vascularly, elimination could occur both during and after absorption, phe-
nomena called pre- and postabsorptive intestinal elimination (Routledge and Shand, 1979). The polarized intestinal tissue might allow
drug administered perorally to have better access to these eliminating
sites than that administered vascularly, leading to greater intestinal
extraction and possible qualitative differences in metabolite kinetics,
as well. Thus, intestinal clearance is potentially a function of intrinsic
clearance as well as the ability of the drug to diffuse to the epithelial
eliminating site, i.e., its diffusional clearance. An assessment of these
factors led Gwilt and colleagues (1988) to model the intestine as three
compartments: the lumen, the metabolically active mucosa, and the
blood in equilibrium with the gut tissue, which can be considered the
serosa (Gwilt et al., 1988). These authors proposed that differences in
pre- and postabsorptive intestinal elimination could arise due to a
“diffusional barrier” at the mucosa-serosa interface. A diffusional
barrier is not necessarily a membranous barrier per se but could be an
increased diffusional path length as a consequence of the polarized
localization of intestinal enzymes.
Phenol, the model substrate used in the present study, is extensively
metabolized by the rat gut (Cassidy and Houston, 1984). It has a
simple biotransformational fate, being largely converted to a sulfate
and a glucuronide with essentially all of the metabolites being renally
eliminated (Capel et al., 1972). Phenol is glucuronidated by members
of the UGT1A family (Inoue et al., 1999) and sulfated by the phenol
sulfotransferases (Oddy et al., 1997). The UGTs show considerable
intestinal expression in rats (Mojarrabi and Mackenzie, 1998; Inoue et
al., 1999; Tukey and Strassburg, 2000). The UGTs display a decreasing expressional gradient in epithelial cells from the villus to the crypt
(Chowdhury et al., 1985) and are localized between the nuclear and
apical membrane of the epithelial cell (Inoue et al., 1999). The active
site of the UGTs is found within the endoplasmic reticulum (ER)
We acknowledge the financial support of the International Student Work Opportunity Program at the University of Minnesota. The work described in this
article was carried out in partial fulfillment of the requirements for a Ph.D. at the
University of Minnesota (P.A.K.).
1
Current Address: Eli Lilly & Co., Lilly Laboratory for Clinical Research, 550
North University Boulevard, Indianapolis IN 46202.
2
Abbreviations used are: UGTs, uridine diphosphoglucuronosyltransferases;
IPI, in situ perfused intestine; SMA, superior mesenteric artery; PST, phenol
sulfotransferase; CL, clearance; HPLC, high-performance liquid chromatography;
ER, endoplasmic reticulum; AUC, area under the curve; 6AC, 6-aminocarbovir.
Address correspondence to: Dr. Cheryl L. Zimmerman, College of Pharmacy,
University of Minnesota, 308 Harvard Street Southeast, Minneapolis, MN 55455.
E-mail: [email protected]
586
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
The influence of enzyme localization and blood flow on intestinal
elimination was evaluated in rats. Phenol was administered vascularly (⬃1400 and 2500 ␮g) and luminally (intrajejunal bolus doses
of ⬃100 and 1000 ␮g) to the recirculating in situ perfused intestine.
The portal effluent and the reservoir were sampled. The intestinal
extraction ratios for phenol at the low and high vascular doses
were (mean ⴞ S.D., n ⴝ 3) 0.09 ⴞ 0.02 and 0.11 ⴞ 0.01, respectively.
The perfusion flow rate was also varied from 5 to 12 ml/min at a
vascular dose of ⬃2500 ␮g of phenol. The organ clearance at the
lowest flow rate significantly exceeded those at the higher flow
rates. The presence of a diffusional barrier at the mucosa-serosa
interface was suggested. The calculated mean diffusional clear-
587
ENZYME LOCALIZATION AND INTESTINAL METABOLISM
lumen (Burchell and Coughtrie, 1989). The phenol sulfotransferases
(PSTs) are also found in the intestinal mucosa (Dunn and Klaasen,
1998) and are cytosolic enzymes (Burchell and Coughtrie, 1997).
Thus, both enzymes display some form of tissue and cellular localization.
The objectives of the present study were to establish a relationship
between intestinal organ clearance and its putative determinants,
namely, diffusional clearance, intrinsic clearance, and intestinal blood
flow, and to evaluate the role of enzyme localization on intestinal
metabolism. Phenol was administered vascularly and luminally to the
recirculating in situ perfused intestine (IPI) and the IPI with luminal
dosing. The influence of tissue and cellular localization of drugmetabolizing enzymes on intestinal organ clearance and metabolite
kinetics was studied.
Materials and Methods
CLgw,org ⫽
Dose
AUCres
(1)
where AUCres is the area under the concentration-time curve of the drug in the
reservoir from 0 to ⬁ (Rowland et al., 1973). The AUCres from 0 to the last
experimental time point, t, was calculated by the linear trapezoidal rule. The
concentration-time profile was fit exponentially with KaleidaGraph (version
3.5; Abelbeck/Synergy, Reading, PA) to obtain a terminal slope, ␭. The AUC
from the last time point to infinity was calculated by dividing the concentration
at the last time point by ␭. The two partial AUCs were added to obtain the
AUC from 0 to infinity. The intestinal extraction ratio, Egw, was calculated as
Egw ⫽
CLgw,org
Qres
(2)
where Qres was the perfusate flow rate.
After luminal dosing, the preabsorptive clearance was calculated based on
an equation developed from the model in Fig. 1. In the figure, Cres, Cgw, and
Cgw, in were the concentrations of the drug in the reservoir, the gut wall, and
entering the gut wall, respectively. Cpv was the concentration of the drug
leaving the gut wall. The rate of drug entry from the lumen to the gut wall was
given by dX/dt. The preabsorptive clearance calculated was the intrinsic
clearance (CLgw) of the gut. The concentration entering the gut wall, Cgw,in,
was considered equivalent to the concentration leaving the reservoir, Cres. The
concentration leaving the gut wall, Cpv, was assumed to be equivalent to the
concentration within the gut wall, Cgw. This assumption is justified for phenol
because it has a very high free fraction (approximately 86%) in a similar
perfusion medium (Ballinger et al., 1995); hence, the free drug may be
considered to be in a rapid equilibrium between the tissue water and the
perfusate.
A mass balance equation for the rate of change of drug amount with time
within the gut wall (dgw/dt) was obtained:
dgw
⫽ Rate in ⫺ Rate out
dt
⫽
dX
⫹ Qres 䡠 Cgw, in ⫺ CLgw 䡠 Cgw ⫺ Qres 䡠 Cpv
dt
Substituting the assumptions of the model in the above equation:
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
Phenol, phenyl glucuronide, p-cresol, sodium pentobarbital, glucose, bovine
serum albumin (25% in Tyrode’s solution), and sulfatase (type H-I from Helix
pomatia) were all obtained from Sigma-Aldrich (St. Louis, MO). Dextran T-40
was obtained from Amersham Biosciences (Piscataway, NJ). All other chemicals were reagent grade or better. Packed human red blood cells were obtained
from the American Red Cross (St. Paul, MN).
Vascular Dosing: The Recirculating IPI. Male Sprague-Dawley rats, 250
to 275 g (Harlan, Indianapolis, IN), were used. Surgery was conducted with
minor modification of a procedure described previously (Hirayama et al.,
1989). The perfusate was Krebs-Henseleit bicarbonate buffer (pH 7.4), bovine
serum albumin, Dextran T-40, and washed human red blood cells in appropriate proportions (Hirayama et al., 1989), and the superior mesenteric arterial
perfusate flow rate was 7.5 ml/min (Davies and Morris, 1993). The effluent
perfusate from the portal vein was returned to the reservoir. A small incision
was made in the jejunum to accommodate a glass cannula, and luminal fluid
was collected during the course of the experiment. The viability of the perfused
organs was continuously monitored by means of blood pressure at the superior
mesenteric artery (SMA) and pO2 in the effluent portal perfusate. Phenol was
administered at vascular doses of 1400 and 2500 ␮g to the reservoir. Samples
(300 ␮l) were taken from the reservoir and portal vein at 3, 8, 15, 30, 45, and
60 min. In addition, a reservoir sample taken before the start of the experiment
was used to calculate the dose. Samples were centrifuged at 13,000g for 2 min
(Fisher Microcentrifuge, model 250B; Fisher Scientific, Pittsburgh, PA). The
supernatant was stored at ⫺20°C until further analysis.
Luminal Dosing: In Situ Perfused Intestine with Luminal Dosing.
Surgery was performed on anesthetized male Sprague-Dawley rats to initiate a
blank IPI. The blank vascular perfusate was recirculated until the preparation
was stabilized at 37°C. The jejunum was ligated 2 to 3 cm from the ligament
of Trietz, and a loose ligature was placed 15 cm distal to it. A small incision
was made in the jejunum at this point, and a metal feeding tube was secured
in it by tightening the ligature around it. The hub of the feeding tube was
attached to a glass syringe containing the dose, and this was introduced as a
bolus at time 0. Samples (300 ␮l) were taken from the reservoir at 0, 10, 25,
55, and 70 min and from the portal effluent at 2, 5, 10, 30, 55, and 70 min.
Sample handling was as described above. At the end of the experiment, the
jejunal loop was flushed with air and saline. The luminal fluid collected was
used to determine the amount in the lumen at the end of the experiment and,
when subtracted from the original dose, gave the amount absorbed during the
course of the experiment. Two luminal doses of phenol (100 and 1000 ␮g)
were administered.
Analytical Methods. Phenol samples were analyzed by HPLC with fluorescence detection. The HPLC setup consisted of a Beckman Gold 126 system
(Beckman Coulter, Inc., Fullerton, CA) with Gold Nouveau integration software (version 1.6; Beckman Coulter) and a RF 530 fluorescence detector
(Shimadzu Corp., Kyoto, Japan). A reversed-phase C18 Supelcosil (4.6 ⫻ 250
mm; 5 ␮ particle size; Supelco, Bellefonte, PA) column was used. The mobile
phase consisted of buffer and acetonitrile (75:25 parts by volume) delivered at
a flow rate of 1 ml/min with isocratic elution. The buffer was 0.05 M KH2PO4
(Fisher Scientific) with 5 mM tetrabutylammonium phosphate (Regis Technologies, Inc., Morton Grove, IL) as an ion-pairing agent. The pH of the buffer
was adjusted to 5.5 with 1% phosphoric acid. The excitation and emission
wavelengths were 260 and 305 nm, respectively.
Standard curves of 1 to 80 ␮g/ml and 2.5 to 100 ␮g/ml were constructed for
phenol and phenyl glucuronide, respectively. The internal standard was pcresol (50 ␮g/ml). The retention times for phenyl glucuronide, phenol, and
p-cresol were 3.9, 8.3, and 13.7 min, respectively. A phenyl sulfate standard
was not commercially available; hence, an enzymatic incubation with sulfatase
was used. An aliquot of 200 ␮l of the sample supernatant was split into two
100-␮l portions. One part was assayed for phenol and the glucuronide as
described above. The other 100 ␮l was incubated with 40 ␮l of sulfatase (5000
U/ml) for 16 h in a shaking water bath at 37°C. The difference between the
phenol concentration before and after sulfatase incubation yielded the concentration of phenyl sulfate. Because a phenyl glucuronide standard was available,
it could be shown that the sulfatase preparations were devoid of glucuronidase
activity. There was no decrease in phenyl glucuronide concentrations in
perfusate samples undergoing sulfatase incubation (data not shown).
The HPLC assay for phenol and phenyl glucuronide was validated for interand intraday precision (% coefficient of variation, % CV) and accuracy. These
were found to be within acceptable limits, with a % CV and bias of less than
10%, except for the lowest concentration of the phenyl glucuronide, which had
a % CV of approximately 15% (Kothare, 2001).
Data Analysis. The actual vascular dose administered was calculated by
multiplying the initial concentration of drug in the reservoir by the volume of
the reservoir. Due to the surgical setup, the gut was the only eliminating organ
in the “circuit”. Thus, the postabsorptive gut clearance, CLgw, org, could be
calculated by
588
KOTHARE AND ZIMMERMAN
FIG. 2. The diffusional model of intestinal clearance.
FIG. 1. Schematic representation of the model used to calculate intestinal
clearance after luminal dosing.
In the figure,Cres is the concentration of drug in the reservoir; Cgw is the
concentration of drug in the gut wall; Cgw,in is the concentration of drug entering the
gut wall; Cpv is the concentration of drug in the portal vein; Qres is the perfusate flow
rate; CLgw is the intestinal intrinsic clearance. X represents the dose, dX/dt is the
rate of drug entry from the lumen to the gut wall, and 夡 represents a sampling site.
Vgw 䡠
dCpv dX
⫽
⫹ Qres 䡠 Cres ⫺ CLgw 䡠 Cpv ⫺ Qres 䡠 Cpv
dt
dt
where Vgw, the volume of gut wall, was taken as 11 ml/250 g rat (Davies and
Morris, 1993).
Multiplying both sides of the equation by dt and integrating with respect to
time from 0 to t:
Vgw 䡠
冕
scribed in the preceding section regarded the gut wall as a single compartment,
and the pre- and postabsorptive intestinal clearances calculated were the
intrinsic clearance and organ clearance by the gut, respectively. To get a
clearer understanding of the physiological determinants of intestinal clearance,
a more rigorous theoretical analysis was done; the model employed is shown
in Fig. 2. The gut wall was divided into two compartments, the eliminating
mucosa and the non-eliminating serosa. A diffusional clearance, CLdiff, governed mass transfer between the two compartments. The intrinsic clearance
from the mucosa, CLgw, was responsible for intestinal elimination. Both
vascular and luminal dosing cases were evaluated. The detailed derivations are
provided in Appendix I.
Based on this analysis, after vascular dosing of drug, the fraction escaping
gut wall clearance, fg, was given by:
fg ⫽
Q 䡠 共CLdiff ⫹ CLgw)
CLdiff 䡠 CLgw ⫹ CLdiff 䡠 Q ⫹ CLgw 䡠 Q
(4)
Cpv at t⫽t
where
dCpv
Cpv at t⫽0
⫽
冕
X at t⫽t
dX ⫹ Qres 䡠
X at t⫽0
冕
t
0
Cres 䡠 dt ⫺ CLgw 䡠
冕
t
Cpv 䡠 dt ⫺ Qres 䡠
0
冕
t
Cpv 䡠 dt
0
Vgw[Cpv,t ⫺ Cpv,0] ⫽ [Xt ⫺ X0] ⫹ QresAUCres,t ⫺ CLgwAUCpv,t ⫺ Qres AUCpv,t
where AUCx,t was the AUC in compartment “x ” from time 0 to t, calculated
by the linear trapezoidal rule; X⬘ ⫽ [Xt ⫺ X0 ] was the amount of drug
absorbed from the gut lumen from time 0 to t; Cpv,t was the drug concentration
in the portal vein at time “t ” minutes. The intestinal intrinsic clearance (CLgw)
was obtained by rearranging the above equation to obtain:
CLgw ⫽
X⬘ ⫹ Qres(AUCres,t ⫺ AUCpv,t) ⫺ Vgw 䡠 Cpv,t
AUCpv,t
(3)
All statistical analyses involved analysis of variance and were conducted with
StatView (version 4.01; SAS Institute, Cary, NC).
Vascular Dosing of Phenol with Varying Vascular Flow Rates. Experiments were conducted at flow rates of 5 (n ⫽ 3), 10 (n ⫽ 2), and 12 (n ⫽ 2)
ml/min with 2500 ␮g vascular doses of phenol. In addition, data were used
from the aforementioned study in vascularly dosed rats with a flow rate of 7.5
ml/min. Flow rates of less than 5 ml/min were attempted but were unsuccessful
because the mesenteric vessels appeared to collapse.
Theoretical Evaluation of Intestinal Clearance. The data analyses de-
Q ⫽ intestinal blood flow
CLdiff ⫽ diffusional clearance of the drug between the serosa and mucosa
CLgw ⫽ intrinsic clearance by the gut
Thus, the fraction escaping gut wall clearance, fg (and therefore, Egw and the
intestinal organ clearance, CLgw,org from eq. 2), is a complex function of the
blood flow to the organ, the diffusional clearance, and the intrinsic clearance
of the drug. It is interesting to consider three limiting conditions of eq. 4.
Case 1: If there is no diffusional barrier (CLdiff ⬎⬎ CLgw)
fg ⫽
Q
Q ⫹ CLgw
(4a)
This is the situation where there is a perfusion rate-limited distribution of the
drug, and the diffusional clearance has no effect on the extraction ratio. This
equation is essentially the same as that for the well stirred model of the liver.
Case 2: If there is a diffusional barrier at the mucosa-serosa interface (CLdiff
⬍⬍ Q)
fg ⫽ 1
(4b)
In such a case, there is a diffusion rate-limited distribution of drug and no
extraction results.
Case 3: For a drug with high intrinsic clearance and poor diffusion (CLgw
⬎⬎ CLdiff)
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
The amounts of the drug in the reservoir, serosa, and mucosa are given by Xr, Xs,
and Xm, respectively. The perfusate flow rate is Q. 夡 represents a sampling site. The
diffusional clearance between the serosa and mucosa is CLdiff. The intrinsic clearance of the drug by the gut is CLgw. After luminal dosing, X represents the dose and
dX/dt is the rate of drug entry from the lumen into the mucosa.
ENZYME LOCALIZATION AND INTESTINAL METABOLISM
589
FIG. 4. A schematic of STELLA Model II dividing the gut wall based on the
cellular compartmentalization of intestinal enzymes.
fg ⫽
Q
Q ⫹ CLdiff
(4c)
In this case, the extent of extraction is determined by the diffusional clearance
between the mucosal and serosal spaces.
Also based on the diffusional model in Fig. 2, after luminal dosing, the AUC
measured to infinity in any compartment yielded the intrinsic clearance, CLgw
(Appendix IB)
AUCm ⫽ AUCs ⫽ AUCr ⫽
D
CLgw
(5)
The above model highlights the potential importance of diffusional barriers in
intestinal extraction. Cases 2 and 3 clearly display situations where a large
difference between pre- and postabsorptive elimination might result from the
existence of a diffusional barrier at the serosa-mucosa interface.
Diffusional Clearance of Phenol. After vascular dosing in the diffusional
model, the AUC to infinity in the reservoir is given by eq. A– 8 (in Appendix
IA). A rearrangement of that equation was used to calculate the diffusional
clearance of phenol
CLdiff ⫽
D䡠CLgw 䡠 Q
AUCr 䡠 CLgw 䡠 Q ⫺ D 䡠 CLgw ⫺ D 䡠 Q
(6)
where
CLdiff
D
Q
AUCr
CLgw
⫽
⫽
⫽
⫽
⫽
diffusional clearance of the drug
vascular dose of drug
blood flow rate
AUC to infinity of parent in the reservoir
intrinsic clearance by the gut (previously calculated from the
luminal dosing of phenol).
STELLA Simulations. To get a better understanding of the metabolite
kinetics of phenol, simulations were done with STELLA simulation software
(Macintosh version 4.0; HPS Systems, Hanover, NH). An initial model (Model
I) was built that assumed the gut to be a single compartment. Because this did
not satisfactorily simulate the experimental results, a more complex model
(Model II) was built further dividing the gut wall based on the cellular
localization of the conjugating enzymes. The schematics of the models used
are provided in Figs. 3 and 4.
Results
Viability. The perfusate pressure at the SMA and the portal pO2
were used to assess viability. For both drugs, a standard deviation of
not more than 5% from the mean was seen during the course of the
experiment (Kothare, 2001), indicating viability of the preparations.
In the studies with different flow rates, the SMA pressure showed a
rising trend with increasing flow rate, indicating an increasing resistance within the blood vessels. The pO2 levels also increased with
flow, indicating more anaerobic conditions presumably due to a
shorter residence time.
Vascular Dosing. Table 1 summarizes the postabsorptive intestinal
metabolism of phenol. A low “postabsorptive” extraction ratio was
obtained after vascular dosing for phenol, and there was no statistical
difference between the extraction ratios obtained at the high and low
vascular doses. Figure 5 shows the concentration-time profiles of
phenol at the low vascular dose. The phenol concentration in the
reservoir declined in a biphasic manner. The higher vascular dose of
phenol revealed a similar pattern (Kothare, 2001).
The two metabolites of phenol showed different shapes in the
concentration-time curves (Fig. 5). The sulfate appeared early and
then reached an asymptotic value. The glucuronide, on the other hand,
appeared later and ascended slowly. An analysis of the luminal
secretions showed that neither phenol nor the conjugates were secreted to a large extent (⬍1% of dose) (Kothare, 2001).
At the high vascular dose of phenol, the portal AUC ratio (sulfate/
phenol), 0.30 ⫾ 0.06, was statistically greater ( p ⬍ 0.05) than the
portal AUC ratio (glucuronide/phenol), 0.13 ⫾ 0.02. Similarly, at the
low vascular dose, the portal AUC (sulfate/phenol), with a value of
0.29 ⫾ 0.14, statistically exceeded the portal AUC ratio (glucuronide/
phenol) at 0.02 ⫾ 0.02. This confirmed sulfation as the predominant
metabolic pathway after vascular dosing.
Luminal Dosing. In sharp contrast to the vascular dosing case, the
low luminal dose of phenol was more extensively metabolized, as
evidenced by comparing the concentration-time profiles (Fig. 6) and
the value of the AUC ratio (metabolite/phenol) from each route to
those in Fig. 5. At the low luminal dose of phenol, the intestinal
intrinsic clearance approached the perfusate flow rate (Table 2).
Glucuronidation was the predominant metabolic pathway in the majority of rats, with the glucuronide levels exceeding those of the
sulfate at most time points. At the low luminal dose of phenol, the
portal AUC ratio (glucuronide/phenol) was 1.18 ⫾ 0.52 whereas the
portal AUC ratio (sulfate/phenol) was 0.74 ⫾ 0.35. The AUC ratio
(metabolite/phenol) overall exceeded that after vascular dosing, indicating greater metabolism.
The distinct metabolic profiles seen with vascular dosing of phenol
were repeated with the luminal dose (Fig. 6). Phenyl sulfate again
reached an asymptote. In contrast, phenyl glucuronide continued its
ascending trend even at later time points when most of the phenol (the
assumed driving force) was all but gone. This trend was repeated in
the concentration-time profiles at the higher luminal dose of phenol as
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
FIG. 3. A schematic of STELLA Model I assuming a single gut wall
compartment.
590
KOTHARE AND ZIMMERMAN
TABLE 1
Extraction ratios and intestinal organ clearances after vascular dosing of phenol
Data are expressed as mean ⫾ S.D. of n ⫽ 3.
Dose
CLgw,org
␮g
ml/min
1462 ⫾ 285
2513 ⫾ 239
0.66 ⫾ 0.16
0.84 ⫾ 0.09
Egw
0.09 ⫾ 0.02
0.11 ⫾ 0.01
Data are represented as mean ⫾ S.D. of n ⫽ 3.
TABLE 2
Intestinal intrinsic clearance for phenol after luminal dosing
Data are expressed as mean ⫾ S.D.
FIG. 5. Concentration-time profiles of phenol (f), phenyl glucuronide (F), and
phenyl sulfate (‚) after the low vascular dose of phenol from the reservoir and
the portal samples.
Data are represented as mean ⫾ S.D. of n ⫽ 3.
well (Kothare, 2001). However, at the high luminal dose, the intrinsic
intestinal clearance was reduced by half, indicating saturation of
metabolism (Table 2). At the higher luminal dose of phenol, the portal
AUC ratio (glucuronide/phenol) was 0.18 ⫾ 0.05 whereas the portal
AUC ratio (sulfate/phenol) was 0.15 ⫾ 0.10. Both routes showed a
statistically lower AUC ratio (metabolite/phenol) at the higher luminal
dose, indicating that saturation was occurring in both pathways.
Based on eq. 6, the mean diffusional clearance (CLdiff) of phenol
could be calculated. Because CLgw was determined from luminal
dosing and the AUCr was determined after vascular dosing, it was
necessary to use the doses from the two routes of administration that
generated similar “effective” concentrations at the enzyme site. Keeping this in mind, the high vascular dose of phenol (⬃2000 ␮g), when
divided by the reservoir volume (200 ml), yielded an effective concentration of 10 ␮g/ml. The low luminal dose (⬃100 ␮g), when
divided by the rat intestinal volume of ⬃11 ml for a 250 g rat (Davies
Dose
n
␮g
116 ⫾ 16
1044 ⫾ 238
CLgw
ml/min
4
3
7.29 ⫾ 1.39a
3.55 ⫾ 1.16a
a
A one-way analysis of variance indicated that the intrinsic clearance was significantly
higher at the low luminal dose, p ⬍ 0.05.
and Morris, 1993), yielded a concentration of ⬃9 ␮g/ml. The initial
effective concentrations generated by the high vascular dose and the
low luminal dose were, therefore, comparable. The diffusional clearance (CLdiff) of phenol was 1.11 ml/min, which was much smaller
than the perfusion flow rate. This is similar to Case 2 (eq. 4b) and is
a possible explanation for the negligible postabsorptive extraction
seen.
Effect of Flow Rate on Clearance. The results of varying the flow
rate are summarized in Table 3. A Fisher’s post hoc analysis revealed
that the postabsorptive extraction ratio, Egw, at the lowest flow rate (5
ml/min) was significantly greater than that at each of the higher flow
rates ( p ⬍ 0.05). The gut extraction ratios at each of the higher flow
rates (7.5–12 ml/min) were statistically similar to each other. The
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
FIG. 6. Concentration-time profiles of phenol (f), phenyl glucuronide (F), and
phenyl sulfate (‚) after the low luminal dose of phenol from the reservoir and
the portal samples.
ENZYME LOCALIZATION AND INTESTINAL METABOLISM
591
TABLE 3
A comparison of the intestinal organ clearance of phenol at various flow rates
Q
Dose
CLgw,org
ml/min
␮g
ml/min
5
7.5
10
12
3033 ⫾ 212a
2513 ⫾ 239
2546d
4272
0.71 ⫾ 0.08b
0.84 ⫾ 0.09
1.04
1.29
Egw
0.14 ⫾ 0.02c
0.11 ⫾ 0.01
0.11
0.11
Data expressed as mean ⫾ S.D. (n ⫽ 3).
The intestinal clearance at 5 ml/min was statistically lower than that at each of the higher
flow rates as determined by a Fisher’s post hoc analysis ( p ⬍ 0.05).
c
The extraction ratio at 5 ml/min was statistically greater than that at each of the higher flow
rates as determined by a Fisher’s post hoc analysis ( p ⬍ 0.05).
d
Data expressed as mean (n ⫽ 2).
a
b
Discussion
This study aimed to experimentally evaluate the role of enzyme
localization in pre- and postabsorptive intestinal elimination. A thorough theoretical assessment of a diffusional model of intestinal elimination was conducted. Alternate models, including those based on the
segregation of intestinal blood flow, have also been evaluated and will
be published in due course.
The postabsorptive intestinal extraction ratio of phenol was very
FIG. 7. Glucuronide concentration-time profiles from the reservoir samples at
each flow rate.
Data are expressed as mean ⫾ S.D. (n ⫽ 3) at flow rates of 5 (f) and 7.5 (E)
ml/min; mean (n ⫽ 2) at flow rates of 10 (‚) and 12 (F) ml/min.
FIG. 8. Simulated concentration-time profiles in the reservoir after luminal
dosing from Model I.
small. It was indeed much lower than the values reported in the
literature after intraduodenal dosing of phenol (Cassidy and Houston,
1984). After vascular dosing, the drug apparently has limited access to
the apically expressed metabolizing enzymes. From a diffusional
perspective, the effective path length of a vascularly administered
drug molecule to the enzyme site would be magnified severalfold due
to the tortuosity of its pathway from the vasculature to the villus tip.
As a result, its “effective diffusivity” after vascular administration
could be much lower than that after luminal administration.
A theoretical analysis revealed that after vascular dosing, the fraction escaping gut wall metabolism (as well as the intestinal organ
clearance) was a complex function of intestinal blood flow, intrinsic
clearance of the drug by the gut, and the diffusional clearance of the
drug between the serosa and the mucosa. The calculated diffusional
clearances of phenol were much smaller than the intestinal blood flow.
In such a situation, the theoretical model predicted that the fraction
escaping gut wall metabolism would be unity. Indeed, the experimental results showed negligible postabsorptive intestinal extraction ratios. In other words, there was a diffusional barrier to the metabolism
of the vascularly administered drug at the mucosa-serosa interface.
Because the vascular blood flow is also an important determinant of
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
intestinal organ clearance, CLgw,org, at 5 ml/min was also significantly
lower than that at each of the higher flow rates. No obvious trends in
the phenol concentration-time profiles were seen (Kothare, 2001). The
AUC ratio (metabolite/phenol) at each of the flow rates showed that
sulfation remained the major metabolic pathway at each of the flow
rates and there were no obvious flow-dependent trends in the overall
formation of the metabolite (Kothare, 2001). However, a comparison
of the glucuronide concentration-time profiles from the reservoir
samples (Fig. 7) revealed that there was a greater lag time in the
appearance of the glucuronide at 5 ml/min compared with each of the
higher flow rates.
STELLA Modeling. There is no evidence in the literature to
suggest sequential metabolism for either of the conjugates of phenol
in the rat. Therefore, in a recirculating system, as long as substrate is
present the conjugates should display a biphasic accumulation curve
(an initial rapid formation phase followed by subsequent slower
accumulation phase). This was not observed for either of the metabolite concentration-time profiles. In addition, concentration-time profiles of the two conjugates showed distinctly different shapes.
STELLA simulation software was used to get a better mechanistic
understanding of the metabolite kinetics. Simulations were conducted
for the vascular and luminal dosing case; however, only the luminal
dosing case is discussed here. The equations used to set up the models
will be provided by the authors upon request (also published in
Kothare, 2001). An initial model, termed “Model I” (Fig. 3), assumed
that formation was the rate-limiting step in the appearance of the
conjugates in the perfusate. As expected, the simulated reservoir
concentration-time profile yielded a biphasic ascending curve (Fig. 8)
in contrast to the observed reservoir concentration-time profile (Fig.
6). Therefore, the more complicated Model II (Fig. 4) was developed.
The gut wall was divided into cytosolic and ER compartments. The
simulated reservoir concentration-time profiles now generated (Fig. 9)
were more consistent with the experimentally observed profiles in Fig.
6. Also, perhaps more interestingly, the concentration-time profile of
the glucuronide in the various compartments (Fig. 10) showed a
significant accumulation of the glucuronide within the ER lumen.
Thus, the simulations suggested a rate-limiting step involved in the
efflux of the glucuronide as well as a metabolic recycling of the
sulfate.
592
KOTHARE AND ZIMMERMAN
FIG. 9. Simulated concentration-time profiles in the reservoir after luminal
dosing from Model II.
Note that the concentrations in the reservoir and the ER are on different y-axis
scales.
intestinal clearance, the effect of changing vascular flow rate on
intestinal organ clearance was evaluated. The lowest flow rate of 5
ml/min yielded statistically greater extraction than the higher flow
rates, presumably due to a greater residence time within the tissue, but
the extraction ratio was still very low. Because the diffusional clearance calculated was much smaller than the vascular flow rate, Case 2
(eq. 4b) of the theoretical model was applicable and predicted that the
entire dose would escape gut wall metabolism. As predicted, changes
in blood flow had little effect on the extraction ratio. The lowest flow
rate was associated with a greater lag time in the appearance of the
glucuronide compared with the higher flow rates.
In marked contrast to the vascular dosing case, after the low luminal
dose of phenol, the concentrations of phenol approached zero by the
end of the experiment. Moreover, the AUC ratios (metabolite/phenol)
after luminal dosing exceeded those observed after vascular dosing. In
addition, the intrinsic gut clearance at the low dose approached the
intestinal blood flow, qualifying phenol as a high clearance drug
(Cassidy and Houston, 1984).
Similar differences between luminal and vascular dosing were
reported earlier from this lab (Wen et al., 1999) with 6-aminocarbovir
(6AC), a prodrug of the carbocyclic nucleoside, carbovir. The conversion of 6AC to carbovir is carried out by adenosine deaminase,
which is also mucosally expressed (Chinsky et al., 1990). Both the
6AC and the phenol examples clearly refute the premise of the
intestine as a well stirred organ; hence, models of intestinal extraction
contingent on a diffusion rate-limited distribution of drug seem more
appropriate. From a practical standpoint, the loss of this heterogeneity
in disruptive systems like cell layers, enterocytes, and microsomes
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
FIG. 10. Simulated concentration-time profile of the glucuronide in the reservoir
and ER compartment based on Model II.
probably contributes to their frequent failure to predict in vivo intestinal clearance.
Two aspects of the metabolite kinetics of phenol were notable—the
distinct shapes of each of the metabolite curves and the route dependence of the formation of the primary metabolite. Because the literature does not suggest sequential metabolism of either conjugate,
these curves may be more reflective of enzyme localization at the
cellular level. In theory, the plateau observed with the sulfate reservoir
concentration profile could indicate either a saturation in its formation
or a metabolic recycling to the parent. Sulfate ions were constantly
supplied to the intestine by way of the Krebs-Henseleit buffer in the
vascular perfusate; hence, a depletion of the cosubstrate seems unlikely. Intracellular metabolic recycling to the parent phenol might
explain the plateauing of the sulfate curve. In fact, the reservoir sulfate
profiles showed a double peak effect, which would be in keeping with
such a recycling.
After luminal administration, the glucuronide concentrations continued to rise even as phenol concentrations (the presumed driving
force) approached zero. Furthermore, there was a decrease in AUC
ratio (metabolite/phenol) and intrinsic clearance in going from the low
to high luminal dose. This may indicate a saturable process being
involved in the release of the glucuronide from the cell.
STELLA simulation software was used to gain a better understanding of the metabolite kinetics of phenol. Model I assumed formation
rate-limited metabolite kinetics in a “closed” recirculating system.
Predictably, the metabolites showed biphasic ascending concentration
curves, unlike the experimentally observed results.
Thus, the more complicated Model II was developed. This divided
the gut wall based on the cellular localization of enzymes into cytosolic and membrane (endoplasmic reticulum)-bound compartments.
Conceivably, phenol entering the gut epithelium would have immediate access to the cytosolic PSTs but would have to diffuse across the
lipophilic ER membrane to access the UGTs. The sulfate formed
could either recirculate unchanged or metabolically recycle to the
parent phenol (Tan and Pang, 2001). Metabolic (“futile”) recycling of
the sulfate could explain the plateau observed in its concentrationtime profile. Phenol’s access to the UGTs in the ER lumen was
modeled as a diffusional process. By setting the diffusional clearance
of the glucuronide to be smaller than its formation clearance, a
diffusional barrier to the glucuronide was modeled. The simulated
reservoir glucuronide profile (Fig. 9) now more closely resembled the
experimental observations. Interestingly, this resulted in a significant
accumulation of the glucuronide within the ER lumen. This could
explain the continued appearance of the glucuronide despite the
declining phenol concentration.
A putative rate-limiting step in intestinal glucuronidation release
was also suggested by a study with salicylamide (Barr and Riegelman,
1970). Salicylamide was administered in vitro to rabbit everted gut
sacs and by in vivo perfusion of an intestinal loop with collection of
mesenteric vein effluent. In the in vitro experiments, whereas the
salicylamide dose was increased severalfold, the amount of glucuronide formed remained constant. Furthermore, in the in vivo experiments, there was an appearance of the glucuronide even though the
salicylamide concentrations were declining. With the recognition that
a member of the multidrug resistance-associated protein family is
expressed in the intestine (Peng et al., 1999; Mottino et al., 2000), the
capacity-limited efflux of the glucuronide seems reasonable.
The striking “route” dependence in the major metabolic pathway is
intriguing. Conceivably, phenol readily encountered the cytosolic
PSTs but required time to diffuse across the lipid-rich barrier of the
ER to access the active site of the UGTs. The phenyl glucuronide,
being more polar and much larger than its aglycone, would experience
593
ENZYME LOCALIZATION AND INTESTINAL METABOLISM
Appendix IA: Effect of a Diffusional Barrier on Intestinal
Clearance After Vascular Dosing
Based on the model depicted in Fig. 2, the following mass transfer
equations defined the rates of change of drug amount in each compartment after vascular dosing.
In the reservoir:
dXr
⫽ Q 䡠 Cs ⫺ Q 䡠 Cr
dt
(A-1)
CLgw
AUCr ⫽
D 䡠 (CLdiff 䡠 CLgw ⫹ CLdiff 䡠 Q ⫹ CLgw 䡠 Q)
CLdiff 䡠 CLgw 䡠 Q
(A-8)
Since the instantaneous extraction ratio, Egw, across the intestine after
vascular dosing is defined as
Egw ⫽ 1 ⫺
Cs
Cr
From 0 to ⬁ this may be expressed as
Egw ⫽ 1 ⫺
AUCs
AUCr
(A-9)
The fraction escaping gut wall metabolism, fg, is therefore:
fg ⫽ 1 ⫺ Egw ⫽
AUCs
AUCr
(A-10)
Substituting A-7 and A-8 in A-10:
fg ⫽
Q 䡠 (CLdiff ⫹ CLgw)
CLdiff 䡠 CLgw ⫹ CLdiff 䡠 Q ⫹ CLgw 䡠 Q
(A-11)
(A-3)
dXm dX
⫽
⫺ CLgw 䡠 Cm ⫺ CLdiff 䡠 Cm ⫹ CLdiff 䡠 Cs
dt
dt
⫽
⫽
⫽
⫽
⫽
⫽
⫽
⫽
amount of drug in the reservoir,
amount of drug in the serosa,
amount of drug in the mucosa,
perfusate flow rate,
concentration of drug in the serosa,
concentration of drug in the reservoir,
concentration of drug in the mucosa,
diffusional clearance of drug between the mucosa and
serosa,
⫽ intrinsic clearance of the drug from the mucosa.
Integrating the above equations with respect to time from 0 to ⬁,
eqs. A-4 through A-6 are obtained:
Q 䡠 AUCs ⫺ Q 䡠 AUCr ⫽ ⫺D
(A-7)
From the model in Fig. 2, the rate of change of drug after luminal
dosing in the reservoir and serosal compartment is still described by
eqs. A-1 and A-2. The rate of change of drug in the mucosa after
luminal dosing is given by the following equation:
In the above equations:
Xr
Xs
Xm
Q
Cs
Cr
Cm
CLdiff
D 䡠 (CLdiff ⫹ CLgw)
CLdiff 䡠 CLgw
AUCs ⫽
(A-2)
In the mucosa:
dXm
⫽ CLdiff 䡠 Cs ⫺ CLdiff 䡠 Cm ⫺ CLgw 䡠 Cm
dt
Equations A-4 through A-6 were solved simultaneously to yield
explicit expressions for AUCs and AUCr.
Appendix IB: Effect of a Diffusional Barrier on Intestinal
Clearance After Luminal Dosing
In the serosal compartment:
dXs
⫽ Q 䡠 Cr ⫺ CLdiff 䡠 Cs ⫹ CLdiff 䡠 Cm ⫺ Q 䡠 Cs
dt
AUCs ⫽ AUC 0 to ⬁ of drug in the serosa,
AUCr ⫽ AUC 0 to ⬁ of drug in the reservoir,
AUCm ⫽ AUC 0 to ⬁ of drug in the mucosa.
where dX/dt ⫽ rate of drug entry from the lumen.
Integrating eqs. A-1, A-2, and A-12 with respect to time from 0 to
⬁, eqs. A-13, A-14, and A-15 are obtained to describe mass transfer
in the reservoir, serosa, and mucosa, respectively:
Q 䡠 AUCs ⫺ Q 䡠 AUCr ⫽ 0
(A-13)
(⫺Q ⫺ CLdiff) 䡠 AUCs ⫹ Q 䡠 AUCr ⫹ CLdiff 䡠 AUCm ⫽ 0
(A-14)
CLdiff 䡠 AUCs ⫹ (⫺CLdiff ⫺ CLgw) 䡠 AUCm ⫽ ⫺D
(A-15)
The above equations were solved simultaneously to obtain the AUC in
each compartment. The AUC in each compartment was exactly the
same and is given in eq. A-16:
(A-4)
AUCm ⫽ AUCr ⫽ AUCs ⫽
(⫺Q ⫺ CLdiff) 䡠 AUCs ⫹ Q 䡠 AUCr ⫹ CLdiff 䡠 AUCm ⫽ 0
(A-12)
D
CLgw
(A-16)
(A-5)
References
CLdiff 䡠 AUCs ⫹ (⫺CLdiff ⫺ CLgw) 䡠 AUCm ⫽ 0
In the above equations
(A-6)
Ballinger LN, Cross SE, and Roberts MS (1995) Availability and mean transit times of phenol
and its metabolites in the isolated perfused rat liver: normal and retrograde studies using tracer
concentrations of phenol. J Pharm Pharmacol 47:949 –956.
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
an even greater diffusional barrier to its efflux from the ER lumen.
Therefore, sulfation would be expected to be the primary metabolic
pathway, as seen after vascular administration. In sharp contrast,
luminal administration of phenol yielded the glucuronide as the primary metabolite, in addition to a much greater overall extraction. This
could be indicative of greater and perhaps more prolonged access of
the perorally administered phenol to the apically placed UGTs. In
other words, in the case of luminal dosing, the effect of cellular
localization of the UGTs was offset by their tissue localization. The
route dependence of the metabolic profile raises the possibility that
xenobiotic metabolism (or carcinogen bioactivation) might be influenced by the route of exposure.
In conclusion, in light of the clear influence of enzyme localization
at the cellular and tissue level on intestinal clearance, the division of
the intestine into the lumen, mucosa, and serosa might in itself not be
adequate to model intestinal clearance. The mucosal compartment
could be further divided into a compartment for membrane-bound
enzymes like the UGTs and CYPs and another for cytosolic enzymes
like the sulfotransferases. Thus, models of the intestine that are
contingent on a diffusion rate-limited rather than perfusion ratelimited distribution of compounds need further investigation.
594
KOTHARE AND ZIMMERMAN
that are predominantly expressed in human colon. Biochem Biophys Res Commun 247:704 –
709.
Mottino AD, Hoffman T, Jennes L, and Vore M (2000) Expression and localization of multidrug
resistant protein mrp2 in rat small intestine. J Pharmacol Exp Ther 293:717–723.
Oddy EA, Manchee GR, Freeman NM, Ward MA, and Coughtrie MWH (1997) Purification and
characterization of a canine liver phenol sulfotransferase. Drug Metab Dispos 25:1205–1210.
Pang KS and Rowland M (1977) 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 enzymatic activity on hepatic drug clearance.
J Pharmacokinet Biopharm 5:625– 653.
Peng K, Cluzeaud F, Bens M, Huyen JDV, Wioland MA, Lacave R, and Vandewalle A (1999)
Tissue and cell distribution of the multidrug resistance-associated protein (MRP) in mouse
intestine and kidney. J Histochem Cytochem 47:757–767.
Routledge PA and Shand DG (1979) Presystemic drug elimination. Annu Rev Pharmacol Toxicol
19:447– 468.
Rowland M, Benet LZ, and Graham GG (1973) Clearance concepts in pharmacokinetics.
J Pharmacokinet Biopharm 1:123–136.
Tan E and Pang KS (2001) Sulfation is rate limiting in the futile cycling between estrone and
estrone sulfate in enriched periportal and perivenous rat hepatocytes. Drug Metab Dispos
29:335–346.
Thummel KE, Kunze KL, and Shen DD (1997) Enzyme-catalyzed processes of first-pass hepatic
and intestinal drug extraction. Adv Drug Deliv Rev 27:99 –127.
Tukey RH and Strassburg CP (2000) Human UDP-glucuronosyltransferases: metabolism, expression and disease. Annu Rev Pharmacol Toxicol 40:581– 616.
Watkins PB (1997) The barrier function of CYP3A4 and p-glycoprotein in the small bowel. Adv
Drug Deliv Rev 27:161–170.
Wen Y, Remmel RP, and Zimmerman CL (1999) First-pass disposition of (⫺)-6-aminocarbovir
in rats. I. Prodrug activation may be limited by access to enzyme. Drug Metab Dispos
27:113–121.
Downloaded from dmd.aspetjournals.org at ASPET Journals on June 17, 2017
Barr WH and Riegelman S (1970) Intestinal drug absorption and metabolism II. Kinetic aspects
of intestinal glucuronide conjugation. J Pharm Sci 59:164 –168.
Burchell B and Coughtrie MWH (1989) UDP-Glucuronosyltransferases. Pharmacol Ther 43:
261–289.
Burchell B and Coughtrie MWH (1997) Genetic and environmental factors associated with
variation of human xenobiotic glucuronidation and sulfation. Environ Health Perspect 105:
739 –747.
Capel ID, French MR, Millburn P, Smith RL, and Williams RT (1972) The fate of [14C]phenol
in various species. Xenobiotica 2:25–34.
Cassidy MK and Houston JB (1984) In vivo capacity of hepatic and extrahepatic enzymes to
conjugate phenol. Drug Metab Dispos 12:619 – 624.
Chinsky JM, Ramamurthy V, Fanslow WC, Ingolia DE, Blackburn MR, Shaffer KT, Higley HR,
Trentin JJ, Rudolph FB, Knudsen TB, and Kellems RE (1990) Developmental expression of
adenosine deaminase in the upper alimentary tract of mice. Differentiation 42:172–183.
Chowdhury JR, Novikoff PM, Chowdhury NR, and Novikoff AB (1985) Distribution of
UDP-glucuronosyltransferase in rat tissue. Proc Nat Acad Sci USA 82:2990 –2994.
Davies B and Morris T (1993) Physiological parameters in laboratory animals and humans.
Pharm Res (NY) 10:1093–1095.
Dunn RT II and Klaassen CD (1998) Tissue-specific expression of rat sulfotransferase messenger
RNAs. Drug Metab Dispos 26:598 – 604.
Gwilt PR, Comer S, Chaturvedi PR, and Waters DH (1988) The influence of diffusional barriers
on presystemic gut elimination. Drug Metab Dispos 16:521–526.
Hirayama H, Xu X, and Pang KS (1989) Viability of the vascularly perfused, recirculating rat
intestine and intestine-liver preparations. Am J Physiol 257:G249 –G258.
Inoue H, Yokota H, Taniyama H, Kuwahara A, Ogawa H, Kato S, and Yuasa A (1999)
1-Naphthol-␤-D-glucuronides formed intraluminally in rat small intestine mucosa and absorbed into the colon. Life Sci 65:1579 –1588.
Kothare PA (2001) Factors Affecting the Intestinal Extraction of Compounds. Ph.D. thesis,
University of Minnesota, Minneapolis, MN.
Mojarrabi B and Mackenzie PI (1998) Characterization of two UDP glucuronosyltransferases