Disposition of naproxen, naproxen acyl glucuronide and its rearrangement isomers in the isolated perfused rat liver

Article abstract of DOI:10.1080/00498250110052715

1. An isolated perfused rat liver (IPRL) preparation was used to investigate separately the disposition of the non-steroidal anti-inflammatory drug (NSAID) naproxen (NAP), its reactive acyl glucuronide metabolite (NAG) and a mixture of NAG rearrangement isomers (isoNAG), each at 30 μg NAP equivalents ml^-1 perfusate (n = 4 each group). 2. Following administration to the IPRL, NAP was eliminated slowly in a log-linear manner with an apparent elimination half-life (t_1/2) of 13.4 ± 4.4 h. No metabolites were detected in perfusate, while NAG was the only metabolite present in bile in measurable amounts (3.9 ± 0.8% of the dose). Following their administration to the IPRL, both NAG and isoNAG were rapidly hydrolysed (t_1/2 in perfusate = 57 ± 3 and 75 ± 14 min respectively). NAG also rearranged to isoNAG in the perfusate. Both NAG and isoNAG were excreted intact in bile (24.6 and 14.8% of the NAG and isoNAG doses, respectively). 3. Covalent NAP-protein adducts in the liver increased as the dose changed from NAP to NAG to isoNAG (0.20 to 0.34 to 0.48% of the doses, respectively). Similarly, formation of covalent NAP-protein adducts in perfusate were greater in isoNAG-dosed perfusions. The comparative results suggest that isoNAG is a better substrate for adduct formation with liver proteins than NAG.

Full text of DOI:10.1080/00498250110052715

2001,
31,
6, 309- 319
no.
xenobiotica,
vol.
Disposition of naproxen, naproxen acyl
glucuronide and its rearrangement isomers in
the isolated perfused rat liver
A. LO, R. S. ADDISON*, W. D. HOOPER and R. G. DICKINSON
Centre for Studies in Drug Disposition, The University of Queensland, Clinical Sciences
Building, Royal Brisbane Hospital, Qld 4029 Australia
Received 16 November 2000
1. An isolated perfused rat liver (IPRL) preparation was used to investigate
separately the disposition of the non-steroidal anti-in ammatory drug (NSAID) naproxen
(NAP), its reactive acyl glucuronide metabolite (NAG) and a mixture of NAG
1
rearrangement isomers (isoNAG), each at 30 g NAP equivalents ml perfusate (
each group).
4
n
m
ˆ
2. Following administration to the IPRL, NAP was eliminated slowly in a log-linear
manner with an apparent elimination half-life of 13.4 4.4 h. No metabolites were
t1
2
…
†
§
=
detected in perfusate, while NAG was the only metabolite present in bile in measurable
amounts (3.9 0.8% of the dose). Following their administration to the IPRL, both NAG
§
and isoNAG were rapidly hydrolysed (
in perfusate 57 3 and 75 14 min
t1=2
ˆ
§
§
respectively). NAG also rearranged to isoNAG in the perfusate. Both NAG and isoNAG
were excreted intact in bile (24.6 and 14.8% of the NAG and isoNAG doses, respectively).
3. Covalent NAP-protein adducts in the liver increased as the dose changed from
NAP to NAG to isoNAG (0.20 to 0.34 to 0.48% of the doses, respectively). Similarly,
formation of covalent NAP-protein adducts in perfusate were greater in isoNAG-dosed
perfusions. The comparative results suggest that isoNAG is a better substrate for adduct
formation with liver proteins than NAG.
Introduction
Conjugation with glucuronic acid is a major metabolic pathway for most drugs
containing a carboxylic acid functional group. Acyl glucuronides are intrinsically
reactive, able to undergo a number of reactions including hydrolysis, rearrange-
ment and formation of adducts with biological macromolecules (Faed 1984,
Dickinson 1993, Fenselau 1994). Hydrolysis, catalysed by hydroxide ion,
-
-
glucuronidases, esterases or serum albumin, leads to the liberation of the parent
aglycone. Rearrangement, via hydroxide ion-catalysed acyl migration, yields the
2-, 3- and 4- positional isomers, which are resistant to hydrolysis by
-
O
-
glucuronidases (but not by hydroxide ion). In contrast to the 1- -acyl glucur-
O
onide, the positional isomers can undergo transient ring opening, with the
production of a reactive aldehyde intermediate which can react with amino groups
on proteins. In this rearrangement/glycation mechanism of drug-protein adduct
formation, both the aglycone and glucuronic acid elements of the conjugated
xenobiotic are incorporated into the adduct (Smith
. 1986, Ding
et al
. 1993,
O
—
et al
Williams and Dickinson 1994). If rearrangement does not occur, the 1- -acyl
—
—
glucuronide can react directly with nucleophilic
SH,
OH and NH2
* Author for correspondence: e-mail:
R.Addison@medicine.uq.edu.au
ISSN 0049-825 4 print/ISSN 1366-5928 online
2001 Taylor & Francis Ltd
Xenobiotica
#
http://www.tandf.co.uk/journals
DOI: 10.1080/0049825011005271 5

310
et al.
A. Lo
CH
CH
3
3
COOH
COOH
COOH
HO
H
O
H
O
CH
O
3
desmethyl-NAP-
phenolic glucuronide
HO
OH
NAP
phenolic
O-dealkylation
acyl-glucuronidation
*
glucuronidation
sulfonation
CH
3
COOH
COOH
OH
CH
3
O
O
H
O
C
HO
H
desmethyl-NAP
acyl-
OH
HO
CH
O
3
CH
3
COOH
NAG
H
glucuronidation
HO SO
3
COOH
OH
CH
3
O
desmethyl-NAP-sulfate
O
C
2-, 3- and 4-O-
positional isomers
(isoNAG)
O
H
OH
HO
HO
2-, 3- and 4-O-
positional isomers
desmethyl-NAG
Figure 1. Metabolism of NAP via dealkylation and conjugation. *Not yet demonstrated in human
or animal studies.
groups on proteins, resulting in loss of the glucuronic acid moiety (i.e. transacyla-
tion mechanism) (van Breeman and Fenselau 1985, Williams and Dickinson 1994,
Ding
. 1995). Both mechanisms of adduct formation have been shown to occur
et al
for carboxylate drugs, with many examples coming from the non-steroidal anti-
in ammatory drugs (NSAIDs).
Naproxen (NAP, 6-methoxy- -methyl-2-naphthaleneacetic acid) is an
¬
NSAID primarily metabolized in humans by acyl glucuronidation to form
naproxen acyl glucuronide (NAG) and by -dealkylation to form 6- -desmethyl-
O
O
naproxen (DMN), which in turn can be further metabolized to glucuronide and
sulphate conjugates (Vanggard-Anderson and Hansen 1992, Vree . 1993,
et al
. 1989) (Ž gure 1). Limited data are available on irreversible binding of
Kiang
et al
NAG with proteins. Bischer
(1995) described the covalent binding of the
et al.
diastereomers of NAG (as the 1- -acyl glucuronide) to human serum albumin,
O
and human and rat plasma proteins
, and reported a higher extent of adduct
in vitro
formation with the -enantiomer in all cases. Additionally, Iwaki
. (1998)
et al
R
reported that,
, human serum albumin forms adducts with NAG (via the 1-
in vitro
- and 2- -isomers) through both the transacylation and glycation mechanisms,
O
O
with the latter forming more slowly but showing greater stability than the former.
No data are available with respect to the formation of covalent adducts of DMN-
glucuronides with proteins.
Drug or xenobiotic molecules that possess both carboxy and phenolic func-
tional groups can be subject to glucuronidation at either or, perhaps, both groups.
The latter reactions, little described for drug molecules, would form novel ‘di-
glucuronide’ metabolites. This novel metabolism has been previously described in
the rat for the NSAID di unisal, which forms a ‘di-glucuronide mixture’
comprised of the phenolic glucuronides of the 2-, 3- and possibly 4- -positional
O
isomers of di unisal acyl glucuronide (King and Dickinson 1991). However, we

311
Naproxen glucuronides in the perfused rat liver
found no evidence to support appreciable phenolic glucuronidation of the 1- -acyl
O
glucuronide. Furthermore, acyl glucuronidation of di unisal phenolic glucuronide
did not occur. Interestingly, the formation of the phenolic glucuronide of DMN
has not been described in human or animal studies.
Our interest in NAP was ultimately directed to the possibility of the formation
of novel di-glucuronide metabolites of DMN, due to the presence of both carboxy
and phenolic groups in this Phase I metabolite of NAP. In order to gain a better
overall understanding of the metabolism of NAP in the rat, we have investigated
its metabolic disposition following administration to the isolated perfused rat liver
(IPRL) in the form of parent drug, its acyl glucuronide conjugate (NAG), and a
mixture of acyl migration rearrangement isomers (isoNAG), with speciŽ c aims of
(1) probing preferential mechanisms of covalent adduct formation and (2) explor-
ing hepatic disposition of DMN.
Materials and methods
Chemicals and reagents
1
NAP, as the -enantiomer, -glucuronidase (
crude extract, 134800 units ml
,
S
Helix pomatia
-
containing 4000 units sulphatase ml 1), -glucose and bovine serum albumin (fraction V) were from
d
Sigma Chemical Co. (St Louis, MO, USA). Methanol, acetonitrile (HPLC grade), 1-chlorobutane,
ethyl acetate, hexane and diethyl ether (AR grade) were from Mallinckrodt (Melbourne, Australia).
Naphthyl acetic acid (NAA) was from BDH chemicals (Sydney, Australia), and iso urane was supplied
by Abbott Australasia Pty Ltd (Sydney, Australia). All other reagents used were of AR grade purity.
DMN was prepared by re uxing a solution of NAP in 4 HCl (1:5 w/v) for 5h. The solution was
m
cooled over ice water, the precipitate recrystallized in boiling water, and Ž ltered and dried. DMN was
formed with a yield of 84% and was shown to be of high purity by HPLC.
IPRL experiments
Animal experiments were approved by the Animal Experimentation Ethics Committee of the
University of Queensland. Male Sprague-Dawley rats (300-350 g) were obtained from the Herston
Medical Research Centre (Brisbane, Australia).
The IPRL experiments were conducted in recirculating mode as described by Wang and Dickinson
(1998). After an initial equilibration of 40min, perfused livers meeting viability criteria (bile  ow
5
l min 1; portal vein pressure 40mmHg; perfusate pH maintained at 7.4 0.1; adequate oxygen
5
4
m
§
consumption as indicated by the di erence in colour between perfusate entering (bright red) and leaving
€
(dark brownish-red) the liver; consistency of colour of perfused tissue) were dosed with either NAP,
NAG or isoNAG (3mg NAP equivalents/100ml perfusate), administered as a bolus dose to the
reservoir. Serial perfusate samples (500 l) were collected predose and at 5, 10, 20, 30, 60, 90, 120, 150
m
and 180 min after dosing, and immediately centrifuged. The supernatant was immediately frozen on dry
ice and stored at 20 C until analysis. Bile samples were collected predose and half-hourly over the
8
perfusion period. The bile was collected on ice in pretared vessels and the acyl glucuronide was
stabilized by titrating the pH to 5.0, using small volumes of 0.1 phosphoric acid. These samples were
m
also frozen on dry ice and stored at 20 C until analysis. At the completion of perfusion, the remaining
8
perfusate was centrifuged and the supernatant frozen. The liver was rinsed with saline, dried, weighed,
homogenized with 2 vols (w/v) 0.1 phosphate bu er, pH 4.5, and frozen at 20 C until analysis.
m
8
€
Biosynthesis of glucuronides
NAG was isolated from the bile of male Sprague-Dawley rats dosed with NAP, as described for
di unisal (King and Dickinson 1991). Following recovery from surgery, the rats were injected via the
1
jugular vein with three doses of NAP (50 mg kg in 0.1 NaHCO ) at 1-h intervals. Bile was collected
m
3
in tared vessels over ice and was titrated to pH 5.0 with 1 phosphoric acid to stabilize the NAG. The
m
bile samples were adjusted to pH 1.0-1.5 with 2 HCl, and extracted three times with 3 vols 1-
m
chlorobutane to remove any NAP present. The NAG was then extracted three times with 3 vols ethyl
acetate, the extracts combined and evaporated at 37 C under a gentle stream of air. NAG was obtained
8
as a viscous green syrup and was reconstituted in water (10 ml). The pH was adjusted to 5.0 with small
additions of 0.1 NaOH. The purity of the bile extract was checked by HPLC and found to contain
m
only small amounts ( 5%) of NAP and isoNAG.
<

312
et al.
A. Lo
IsoNAG was prepared as an unspeciŽ ed mixture of acyl migration isomers by rearrangement of
NAG under mild alkaline conditions. Brie y, the puriŽ ed NAG solution was adjusted to pH 8.5 with
2
NaOH and incubated at 37 C for 5 h. Under these conditions, acyl migration is the dominant
m
8
reaction, although some hydrolysis of the NAG and isoNAG being generated can also occur. The
reaction was monitored by taking serial hourly samples for analysis by HPLC. After 5 h, -glucur-
-
onidase was added to the solution of isoNAG, to a Ž nal concentration of 1000 units ml 1, and incubated
overnight at 37 C to hydrolyse any residual NAG. The Ž nal solution of isoNAG was readjusted to pH
8
1.0-1.5 with 2 HCl and any residual NAP removed by extraction three times with 3 vols 1-
m
chlorobutane. The isoNAG solution was then extracted three times with 3 vols ethyl acetate, the
extracts combined and evaporated at 37 C under a gentle stream of air. IsoNAG was obtained as a thick
8
green syrup and was reconstituted in water (10 ml), and the pH adjusted to 5.0 with small additions of
0.1 NaOH. The purity of the isoNAG solution was checked by HPLC and found to be 96%.
m
>
HPLC analyses
The analytical system consisted of a model 510 HPLC pump, a model 481 UV detector and a WISP
712 autosampler (Waters Associates, Milford, MA, USA). The analytical column was a 4- m Novapak
m
C18 cartridge column (100 8 mm i.d.) housed in an RCM 8 10 compression unit, and was preceded
£
£
by a Bondapak C18 Corasil guard column (Waters). The mobile phase consisted of water (1000 ml),
m
acetonitrile (670 ml) and glacial acetic acid (2ml), and the  ow rate was 2ml min 1. Using these
conditions, NAP eluted at 13.3 min, NAG at 2.8 min (as a single peak well resolved from positional
isomers), the 4- -positional isomer of NAG as a single peak at 2.0 min and the 2- - and 3- -positional
O
O
O
isomers as non-baseline-resolved doublets with peaks between 3.3 and 3.8 min. UV detection at 225 nm
was used in all cases. All quantitative analyses were performed in conjunction with standard curves
constructed from blank samples of an appropriate matrix, spiked with increasing concentrations of
NAP. Standard curves had a coe cient of determination of 0.99 in all cases.
5

Analysis of non-covalently bound drug and metabolites
Perfusate samples obtained from the NAP-, NAG- and isoNAG-dosed IPRL experiments were
analysed for NAP and its metabolites as follows. Aliquots of perfusate samples (100 l) were added to
m
1
150 l 20 g ml NAA in acetonitrile containing 4% (v/v) acetic acid. After vortexing and centrifuga-
m
m
tion, 20 l samples were injected into the HPLC system. For analysis of the bile samples, aliquots (20 l)
were diluted with 150 l 200 g ml NAA in acetonitrile containing 4% (v/v) acetic acid and 230 l 0.1
phosphate bu er, pH 4.5. Aliquots (10 l) of the prepared bile samples were injected into the HPLC
m
m
m
1
m
m
m
m
€
system.
Analysis of non-covalently bound drug species in tissue was conducted as follows. Liver homogenate
(1 g) was added to 4 vols acetonitrile containing 4% acetic acid (v/v), and the samples vortexed and
centrifuged (3000 ). The supernatant was removed and evaporated at 37 C under a gentle stream of air.
g
8
To the residues, 1 ml NaOH (1 ) was added and the samples were then incubated at 80 C to allow
hydrolysis of glucuronides. Internal standard (50 l 20 g ml NAA in water) was then added and the
samples extracted twice with 4 ml hexane/diethyl ether (1:1 v/v). After gentle mixing and centrifugation,
the organic layer was removed and evaporated under a gentle stream of air at 37 C. The residue was
reconstituted in 100 l mobile phase, and 30 l was injected into the HPLC system.
m
8
1
m
m
8
m
m
Analysis of covalently bound drug metabolites
NAP-protein adducts were measured in Ž nal perfusate and liver homogenate samples as follows.
Samples of perfusate (1 ml) or liver homogenate (1 g) were added to 4 vols acetonitrile containing 4%
(v/v) acetic acid to precipitate proteins. After mixing (30 s) and centrifugation (3000 ), the supernatant
g
was discarded, and the pellet resuspended and washed with methanol/diethyl ether (3:1 v/v, 4ml). This
washing procedure was repeated seven times to remove non-covalently bound drug and metabolites
from the protein pellets. After vortexing and centrifugation, the last wash was analysed to conŽ rm that
the protein pellet was free of non-covalently bound drug. The pellet was dried under a gentle stream of
air and digested overnight in 1 ml 1 NaOH at 80 C. The digest was acidiŽ ed with 200 l 10 HCl, and
m
m
8
m
1
50 l 20 g ml NAA in water was added. The sample was then extracted twice with 4 ml hexane/
diethyl ether (1:1 v/v), and the combined organic layers dried down under a gentle stream of air at 37 C.
The residue was reconstituted in 100 l mobile phase, and 30 l was injected into the HPLC system.
m
m
8
m
m
Data analysis
The apparent perfusate half-lives of NAP, NAG and isoNAG were determined from the slope of the
log concentration-time proŽ le using linear regression analysis. Statistical comparisons between perfu-
sion experiments were performed using unpaired Student’s -tests.
t

313
Naproxen glucuronides in the perfused rat liver
Results
The concentration-time proŽ le of NAP following its administration to the
isolated perfused rat liver preparation is shown in Ž gure 2. NAP concentrations in
serial perfusate samples showed a slow linear decrease over the perfusion period,
with an elimination half-life ( ), calculated between 30 and 180 min of perfusion,
t1 2
of 13.4 4.4 h. At 3 h, the r=ecovery of unchanged NAP from perfusate was
§
75.8 12.0% of the dose (table 1). No metabolites of NAP were measurable in
§
serial or 3-h perfusate samples. Analysis of bile samples revealed NAG to be the
only metabolite present in measurable amounts (3.9 0.8% of the dose), with the
§
majority excreted between 30 and 60 min after dosing. Covalent NAP-protein
adducts were found in the liver, accounting for 0.20 0.06% of the dose, while
§
liver homogenate contained 1.5 0.3% of the dose as non-bound NAP derived
§
species. No NAP-protein adducts were found in the perfusate above the detection
limit of the assay.
Rapid hydrolysis of NAG occurred after its administration to perfused livers
(Ž gure 2), with a
, calculated over the essentially linear 5-30 min period, of
t1=2
11.3 1.4 min. Its elimination thereafter (60-180 min) was slower, although
§
essentially linear, with a
57 2 2 7 min. NAP appeared rapidly in perfusate
t1=2 ˆ
§
: :
during the Ž rst hour of perfusion, followed by a more gradual rise from 60 to
180 min. NAG also rearranged rapidly to form isoNAG, with the concentration in
perfusate at 5 min after dosing being the same as that of NAP. However, isoNAG
was also rapidly eliminated under the perfusion conditions with a
t1 2
ˆ
=
23 6 5 1 min, and fell below quantiŽ able limits 60 min after dosing. At the
§
:
:
end of perfusion, only 1.6% of administered NAG was recovered in perfusate,
Table 1. Recovery of NAP-derived species after administration of NAP, NAG and isoNAG at 3mg
NAP equivalents to the isolated perfused rat liver preparation.
% Recovery of dosea
NAP-
protein
adducts
Non-bound
NAP-derived
species
NAP
NAG
isoNAG
Total
NAP dosed
Perfusate
Bile
Liver
Total
75.8 12.0
-
<
-
-
75.8 12.0
§
§
<
<
<
-
b
3.9 0.8
3.92 0.80
§
§
<
c
-
-
0.20 0.06
1.52 0.27
1.71 0.33
§
§
§
75.8 12.0
3.9 0.8
0.20 0.06
1.52 0.27
81.4 12.1
§
§
§
§
§
NAG dosed
Perfusate
Bile
Liver
Total
60.9 8.3
1.6 0.6
0.65 0.10
63.4 9.0
§
§
§
§
<
<
24.6 7.3d
3.0 0.86
-
-
27.8 8.2d
§
§
§
<
-
-
-
0.34 0.03d 2.33 0.86
2.7 0.8d
§
§
§
60.9 8.3
25.8 6.5d
3.0 0.86
0.99 0.12d 2.33 0.86
92.9 2.2
§
§
§
§
§
§
IsoNAG dosed
Perfusate
Bile
Liver
Total
55.8 5.0d
5.8 1.7
1.53 0.14e
-
-
63.1 6.7
§
§
§
§
<
<
-
14.8 3.4e
-
14.8 3.4d
§
§
<
-
-
0.48 0.12d 7.65 1.91d;e 8.1 2.0d;e
§
§
§
55.8 5.0d
20.6 1.8e
2.02 0.25d;e 7.65 1.91d;e 86.1 5.2
§
§
§
§
§
a Values are means SD
4 .
†
n
§
…
ˆ
b Below the limit of the assay.
c No measurement made.
d SigniŽ cantly di erent
0 05 from the corresponding value from NAP-dosed perfusions.
0 05 from the corresponding value from NAG-dosed perfusions.
p
€
€
…
…
†
<
<
:
:
e SigniŽ cantly di erent
p
†

314
et al.
A. Lo
(A)
(B)
(C)
Figure 2. Concentration-time proŽ les of NAP, NAG and isoNAG after dosing the recirculating
isolated perfused rat liver preparation with (A) NAP, (B) NAG and (C) isoNAG at 3 mg NAP
equivalents. Results are mean SD
4 .
n
§
… ˆ †

315
Naproxen glucuronides in the perfused rat liver
while NAP in perfusate accounted for 60.9 8.3% of the dose (table 1). NAG was
§
rapidly excreted in bile, with 24.6 7.3% of the dose being recovered, signiŽ cantly
§
higher than the recovery observed for this species in NAP-dosed IPRL. The
majority of this excretion occurred in the Ž rst 90 min after dosing. Covalent NAP-
protein adducts were higher in the liver of NAG-dosed IPRL compared with
NAP-dosed preparations, accounting for 0.34 0.03% of the dose, while recovery
§
of non-bound drug was also found in higher levels (2.3 0.9% of the dose). NAP-
§
protein adducts were also detected in low levels in the perfusate at the end of the
perfusion period (0.65% of the dose).
Following its administration to IPRL preparations, isoNAG decreased in serial
perfusates (Ž gure 2) with an initial (0-30 min)
slower elimination phase (60-180 min) with
the isoNAG dose to the parent drug was evident, with the NAP concentration in
Ž nal perfusate being similar to that obtained in NAG-dosed perfusions. Recoveries
of isoNAG and NAP from perfusate were 5.8 1.7% and 55.8 5.0% of the dose
16 7 1 0 min, followed by a
t1 2
t1=2 ˆ
ˆ
§
: :
=
75 0 14 5 min. Hydrolysis of
§
:
:
§
§
respectively (table 1). Serial bile samples predominantly contained isoNAG
(14.8 3.4% of the dose), with no other metabolites present above the detection
§
limit of the assay.
The formation of NAP-protein adducts in liver was signiŽ cantly higher
0 05 in both NAG-dosed (recovery 0.34 0.03% of the dose) and iso-
p
…
†
ˆ
§
< :
NAG-dosed (0.48 0.12% of the dose) IPRL than that observed in those dosed
§
with the parent drug (table 1). Non-bound NAP-derived species were higher
0 05 in isoNAG dosed IPRL (7.7 1.9% of the dose) than in NAP- and
NAG-dosed IPRL experiments. The perfusate at the end of the 3-h isoNAG-
p
…
†
§
< :
dosed preparations also showed the highest concentration of covalent NAP-
protein adducts, with a mean recovery of 1.53 0.14% of the dose.
§
Discussion
The carboxylate NSAID NAP is metabolized in humans and animals to its acyl
glucuronide conjugate NAG (Vree
. 1993). Also documented is demethylation
et al
to DMN, which bears both carboxy and phenolic functions, and which can
therefore be conjugated variously by acyl glucuronidation and sulphonation
(Kiang
. 1989, Vanggard-Anderson and Hansen 1992, Vree
. 1993), and
et al
et al
perhaps even by novel diconjugation (as shown earlier for the carboxy/phenolic
NSAID di unisal; King and Dickinson 1991). One of our longer term goals was to
explore hypotheses regarding diconjugation mechanisms. The present study was
concerned not only with establishing the detailed
disposition of NAP (as a
in vivo
prelude to establishing that of DMN), but also to probe intrahepatic transport of
and protein adduct formation from NAG and, more speciŽ cally, its acyl migration
rearrangement isomers (i.e. isoNAG). The role of NAG and isoNAG as reactive
substrates for covalent modiŽ cation of plasma protein
has already been
in vitro
studies are lacking.
established (Iwaki
. 1998). However,
et al
in vivo
The present study administered NAP, its preformed acyl glucuronide (NAG),
and a prepared mixture of NAG rearrangement isomers (isoNAG), to the IPRL.
After administration of NAP, its elimination from perfusate was very slow
(apparent
vivo
13 h, almost three times slower than that observed in the rat
et al.
t1 2 ˆ ¹
in
1972). NAG in bile was the only metabolite quantiŽ able, and
; Runke=l
accounted for 4% of the dose. By comparison, when preformed NAG was
¹

316
et al.
A. Lo
administered at the same dose, systemic hydrolysis to parent NAP was the major,
and systemic rearrangement to isoNAG the minor, disposition pathway observed
in perfusate (Ž gure 2). The perfusate proŽ les of the isoNAG generated
in situ
appeared to be biphasic and essentially parallel to those of its precursor NAG.
Following dosing of preformed isoNAG itself to perfused livers, hydrolysis to
NAP was the dominant pathway discerned from perfusate analyses. As is known
generally, an acyl glucuronide is not reformed during rearrangement amongst its
isomers (Ž gure 1) (Spahn-Langguth and Benet 1992, Fenselau 1994). Therefore,
after dosing to perfusions, NAG will not be reformed spontaneously from isoNAG
by rearrangement; any ‘new’ NAG must originate from glucuronidation of NAP
released by hydrolysis of isoNAG. However, this must be minimal given the
results obtained after administration of NAP itself to perfused livers. In this
context it is noteworthy that 2 and 6% of the administered NAG or isoNAG
<
doses respectively were recovered as such in perfusate at 3 h, in contrast to the
recovery of NAP at 3 h (table 1).
The perfusate proŽ les obtained were not dissimilar to those reported earlier
following dosage of the salicylate NSAID di unisal, its acyl glucuronide, and
rearrangement isomers to the IPRL (Wang and Dickinson 1998). Thus, while
hydroxide ion-catalysed rearrangement seems to be the predominant mechanism
of degradation of acyl glucuronides in physiological bu ers
, hydrolysis
in vitro
, presumably due to the combined e ects of hydro-
€
seems to predominate
in vivo
xide ion, esterases and -glucuronidases (Smith
€
. 1990, King and Dickinson
et al
-
1991, Williams and Dickinson 1994). Interestingly, the ‘terminal’ elimination half-
life of NAG in the IPRL system, where the perfusate contains 1% BSA, was
57.2 2.7 min, slightly longer than those reported by Bischer
found that NAG degraded with half-lives of 47 and 27 min in phosphate bu er
. (1995) who
et al
§
€
containing 0.3 and 3.0% human serum albumin respectively. The -glucuronidase-
-
resistant rearrangement isomers, isoNAG, were, like NAG, eliminated from
perfusate in a facile manner, though less rapidly than NAG. This was not
unexpected, as acyl glucuronide isomers have previously been shown to be more
stable in the IPRL (Andrews and Dickinson 1998, Wang and Dickinson 1998) and
in the rat
(King and Dickinson 1991).
in vivo
Perfused livers dosed with NAP yielded only 4% of the dose as the acyl
¹
glucuronide NAG in bile in 3 h (table 1), indicating the relatively poor capacity of
the IPRL to metabolize NAP. However, after dosing with preformed NAG, 25%
¹
of the dose was recovered (as NAG) in bile in 3 h, the great majority ( 80%) of this
>
being recovered in the Ž rst 1.5 h of perfusion. For perfused livers dosed with
isoNAG, 15% of the dose was recovered (as isoNAG) in bile in 3 h, with the
¹
major part again being in the Ž rst 1.5 h. The only explanation for these results is
that considerable portions of intact NAG and isoNAG in perfusate are taken up by
hepatocytes at the sinusoidal membrane, transported intrahepatically to the
canalicular membrane, and excreted into bile. Such a transport phenomenon
will be involved in the recovery of 3% of the NAG dose as isoNAG in bile
(most likely subsequent to rearrangement of NAG in perfusate, though rearrange-
ment of NAG subsequent to its biliary excretion will also play some role). A small
portion of the NAG dose recovered in bile as NAG will be produced by
intrahepatic glucuronidation and biliary excretion of NAP, generated by hydro-
lysis of NAG in perfusate.

317
Naproxen glucuronides in the perfused rat liver
Small choleretic e ects were observed (data not shown) in this study during the
€
Ž rst 1.5 h following dosing with NAG and isoNAG, but not after dosing with
NAP. This is consistent with the relative recovery data for NAG or isoNAG in
bile after the various doses, as choleretic e ects are known to accompany biliary
€
excretion of the acyl glucuronides of valproate (Dickinson
1982, Andrews and
et al.
Dickinson 1998) and di unisal (Wang and Dickinson 1998). The observed
choleretic e ects coincided with the period of major excretion of NAG and
€
isoNAG in bile (i.e. in the Ž rst 1.5 h after dosing).
The perfusate AUC0
istration to the IPRL were 4370 436, 480 130 and 953 113 g minml
for NAP, NAG and isoNAG following their admin-
3h
1
,
m
§
§
§
respectively. Covalent NAP-protein adducts in perfusate (notionally with BSA),
measured at 3 h, comprised 0, 0.65 and 1.53% of the doses, respectively. If
exposure of perfusate to NAP-derived species is measured by AUC0 3h, then
NAP itself cannot be a candidate substrate for covalent adduct formation, though
clearly both NAG and isoNAG can act in this capacity. For comparison here, it is
instructive to remember that NAG can spontaneously form isoNAG, though the
reverse does not happen. Thus, an exposure of perfusate proteins to
1
953 113 g min ml
yielded 1.53 0.14% of the isoNAG dose as covalent
m
§
§
adducts (i.e. a ‘conversion rate’ of 1.605 10 3 min 1). After dosage with NAG,
£
perfusate proteins were exposed to spontaneously generated isoNAG
(AUC0
123 28 g min ml 1). If the same adduct yield determined above
3h
m
ˆ
§
applied, this exposure would account for 0.20% of the 0.65% yield of adducts
following dosage with NAG. On this basis, it could be suggested that isoNAG is a
slightly better substrate (via the glycation mechanism) than NAG (presumably
via
the transacylation mechanism, ‘conversion rate’ 9.409 10 4 min 1) for covalent
ˆ
£
NAP-protein adduct formation in perfusate of the IPRL.
Covalent NAP-protein adducts were also found in liver, with the extent of
formation increasing as the dose changed from NAP to NAG to isoNAG (table 1).
This trend is consistent with our earlier observations for di unisal, its acyl
glucuronide and rearrangement isomers following dosing to the IPRL (Wang
and Dickinson 1998). It is di cult to estimate the intrahepatic exposure to NAG

and/or isoNAG after each of the three dose types. However, some comparative
measure might be taken from the biliary excretion data. Thus, the values for the
sum of NAG and isoNAG excreted in bile after dosing with NAP, NAG and
isoNAG were 3.9 0.8, 27.6 8.2 and 14.8 3.4% of the doses, respectively (i.e.
§
§
§
ratio 1:7:4). This corresponds with a ratio of 1:1.7:2.4 for the respective recoveries
of the doses as adducts in liver. This is a somewhat coarse analysis, but it may
serve to illustrate the point that isoNAG appears to be a somewhat better substrate
for adduct formation in liver, in agreement with the observations discussed above
for adduct formation in perfusate.
The relative levels of non-covalently bound NAP-derived species in liver tissue
at the end of the 3-h perfusion period (table 1) were quite intriguing. Thus, 1.5, 2.3
and 7.7% of the doses of NAP, NAG and isoNAG respectively were recovered in
liver as non-covalently bound species (measured as NAP after alkaline hydrolysis
of liver extracts). It is presumed that the liver at 3 h would be in equilibrium with
the drug species circulating in perfusate, and it would further be expected that the
more lipophilic species (in this case NAP, as compared with NAG and isoNAG)
would achieve relatively higher equilibrium concentrations. However, the con-
1
centrations of NAP in perfusate at 3 h were 22.7, 18.3 and 16.7 g ml following
m

318
et al.
A. Lo
dosing with NAP, NAG and isoNAG respectively, i.e. the opposite trend to that
found for non-covalently bound NAP-derived species in liver. There are, of
course, residual concentrations of NAG and isoNAG at 3 h after dosing with
1
these species (0.5 and 1.7 g NAP equivalent ml respectively), though these
m
concentrations are considerably lower than the concurrent concentrations of NAP
in these perfusions. The data appear to point to an unexplained avid but non-
covalent association of isoNAG with liver tissue.
Conclusion
This study has shown that NAP is only very slowly glucuronidated in the
IPRL. Its acyl glucuronide, NAG, undergoes both hydrolysis and rearrangement,
is taken up intact by the liver and excreted in bile, and serves as a substrate for
covalent adduct formation with both perfusate and liver proteins. The rearrange-
ment isomers, isoNAG, also undergo hydrolysis, uptake and biliary excretion, and
seem to be even better substrates for adduct formation.
References
Andrews, J. A.
and
1998, Metabolic discrimination between valproic acid
Dickinson, R. G.,
glucuronide and its rearrangement isomers in the perfused rat liver.
Proceedings of the
, 5, P2-38.
Australasian Society of Clinical and Experimental Pharmacologists and Toxicologists
Bischer, A., Zia Amirhosseini, P., Iwaki, M., McDonagh, A. F.
-
and
1995,
Benet, L. Z.,
Stereoselective binding properties of naproxen glucuronide diastereomers to proteins.
, 23, 379-395.
Pharmacokinetics and Biopharmaceutics
Journal of
1993, Acyl glucuronide conjugates: reactive metabolites of nonsteroidal anti-
Dickinson, R. G.,
in ammatory drugs.
, 36, 157-162.
Proceedings of the Western Pharmacological Society
and
1982, An osmotic
, 32, 241-247.
Dickinson, R. G., Harland, R. C., Kaufman, S. N., Lynn, R. K.
explanation for valproic acid induced choleresis in the rat.
Ding, A., Ojingwa, J. C., McDonagh, A. F., Burlingame, A. L.
Gerber, N.,
Arzneimittelforschung
and 1993, Evidence of
Benet, L. Z.,
covalent binding of acyl glucuronides to serum albumin via an imine mechanism as revealed by
tandem mass spectrometry.
3801.
, 90, 3797-
Proceedings of the National Academy of Sciences, USA
-
and
1995,
Benet, L. Z.,
Ding, A., Zia Amirhosseini, P., McDonagh, A. F., Burlingame, A. L.
Reactivity of tolmetin glucuronide with human serum albumin: IdentiŽ cation of binding sites
and mechanisms of reaction by tandem mass spectrometry.
369-376.
, 23,
Drug Metabolism and Disposition
1984, Properties of acyl glucuronides: Implications for studies of the pharmacokinetics
Faed, E. M.,
and metabolism of acidic drugs.
, 15, 1213-1249.
Drug Metabolism and Disposition
1994, Acyl glucuronides as chemically reactive intermediates.
Fenselau, C.,
Pharmacology
Iwaki, M., Ogiso, T., Inagawa, S.
2- -acyl glucuronides of naproxen and their covalent binding to human serum albumin.
Handbook of Experimental
, 112, 367-389.
and
1998,
regioselective stability of -1- - and
Kakehi, K.,
In vitro
O
-
O
Journal
, 88, 52-57.
of Pharmaceutical Sciences
Kiang, C. H., Lee, C.
-
and
1989, Isolation and identiŽ cation of 6-desmethyl-naproxen
Kushinsky, S.,
sulfate as a new metabolite of naproxen in human plasma.
43-48.
, 17,
Drug Metabolism and Disposition
and
1991, Studies on the reactivity of acyl glucuronides: I. Phenolic
Dickinson, R. G.,
King, A. R.
glucuronidation of isomers of di unisal acyl glucuronide in the rat.
42, 2289-2299.
,
Biochemical Pharmacology
and
1972, Absorption, distribution,
Forchielli, E.,
Runkel, R., Chaplin, M., Boost, G., Segre, E.
metabolism and excretion of naproxen in various laboratory animals and human subjects.
Journal
, 61, 703-708.
of Pharmaceutical Sciences
Smith, P. C., McDonagh, A. F.
and
1986, Irreversible binding of zomepirac to plasma
Benet, L. Z.,
protein
and
.
, 77, 934-939.
in vitro
Smith, P. C., McDonagh, A. F.
of zomepirac glucuronide and its covalent binding to plasma proteins in the guinea pig.
in vivo Journal of Clinical Investigation
and 1990, E ect of esterase inhibition on the disposition
Benet, L. Z.,
€
Journal of
, 252, 218-224.
Pharmacology and Experimental Therapeutics

319
Naproxen glucuronides in the perfused rat liver
-
and
1992, Acyl glucuronides revisited: is the glucuronidation
Benet, L. Z.,
Spahn Langguth, H.
process a toxiŽ cation as well as a detoxiŽ cation mechanism?
, 24, 5-47.
Drug Metabolism Reviews
and
1985, Acylation of albumin by 1- -acyl glucuronides.
O Drug
Van Breeman, R. B.
Metabolism and Disposition
and
Fenselau, C.,
, 13, 318-320.
-
1992, Simultaneous quantitative determination of
Hansen, S. H.,
Vanggard Anderson, J.
naproxen, metabolite 6- -desmethylnaproxen and their Ž ve conjugates in plasma and urine
O
samples by high performance liquid chromatography on dynamically modiŽ ed silica.
Journal of
, 577, 325-333.
Chromatography
Vree, T. B., van den Biggelaar Martea, M., Verwey van Wissen, C. P. W. G. M., Vree, M. L.
1993, The pharmacokinetics of naproxen, its metabolite
-
-
and
-
Guelen, P. J. M.,
desmethylnaproxen, and their acyl glucuronides in humans. E ect of cimetidine.
O
British
€
, 35, 467-472.
Journal of Clinical Pharmacology
and 1998, Disposition and covalent binding of di unisal and di unisal
Wang, M.
Dickinson, R. G.,
acyl glucuronide in the isolated perfused rat liver.
, 26, 98-104.
Drug Metabolism and Disposition
and
1994, Studies on the reactivity of acyl glucuronides - IV.
Dickinson, R. G.,
Williams, A. M.
Modulation of reversible and covalent interaction of di unisal acyl glucuronide and its isomers
with human plasma protein , 47, 457-467.
.
in vitro Biochemical Pharmacology

Copyright of Xenobiotica is the property of Taylor & Francis Ltd and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.