Full text of DOI:10.1211/0022357001777379

J. Pharm. Pharmacol. 2000, 52: 1239±1246
Received April 3, 2000
Accepted April 4, 2000
# 2000 J. Pharm. Pharmacol.
Pharmacokinetic Interactions between HIV-1 Protease Inhibitors
in Rats: Study on Combinations of Two Kinds of HIV-1 Protease
Inhibitors
NOBUHITO SHIBATA, YASUHIRO MATSUMURA, HIROYUKI OKAMOTO, YUKO KAWAGUCHI,
AKIKO OHTANI, YUKAKO YOSHIKAWA AND KANJI TAKADA
Department of Pharmacokinetics, Kyoto Pharmaceutical University, Yamashina-ku, Kyoto 607-8414,
Japan
Abstract
The drug interactions between four human immune de®ciency virus (HIV-1) protease
inhibitors have been characterized by in-vitro metabolic studies using rat liver microsomal
fractions and in-vivo oral administration. In this study, a new HPLC analytical method
developed by us was used for the simultaneous determination of saquinavir and nel®navir
in rat plasma and microsomes.
The metabolic clearance rates (Vmax=Km) of saquinavir, nel®navir, and indinavir were
1
1
170Á9Æ 10Á9, 126Á1Æ 4Á4, and 73Á0Æ 2Á0 mL min (mg protein) , respectively. Ritonavir
was the strongest inhibitor with inhibition constants (Ki) of 1Á64 mM for saquinavir, 0Á95 mM
for indinavir, and 1Á01 mM for nel®navir. Nel®navir was the second strongest inhibitor with
Ki's of 2Á35 mM for saquinavir and 2Á14 mM for indinavir. Indinavir was the third strongest
inhibitor with Ki's of 2Á76 mM for nel®navir and 3Á55 mM for saquinavir. Saquinavir was the
weakest inhibitor for the other three HIV-1 protease inhibitors. After oral co-administration
in combination with another HIV-1 protease inhibitor, the AUCs of saquinavir, indinavir,
and nel®navir were signi®cantly increased compared with mono-treatment. The AUCs of
saquinavir were increased about 10Á1-, 3Á1- and 45Á9-fold in the presence of indinavir,
nel®navir and ritonavir, respectively. The AUCs of indinavir were increased about 6Á8-,
5Á9- and 9Á4-fold in the presence of nel®navir, saquinavir and ritonavir, respectively. The
AUCs of nel®navir were increased about 2Á2-, 6Á6- and 8Á5-fold in the presence of indi-
navir, saquinavir and ritonavir, respectively.
The in-vivo effects observed after co-administration of two kinds of HIV-1 protease
inhibitor were not always expected from in-vitro data, suggesting the presence of other
interaction processes besides metabolism in the liver. These results provide useful infor-
mation for the treatment of AIDS patients receiving combination therapy with two HIV-1
protease inhibitors.
The recent discovery of HIV-1 protease inhibitors
has introduced a new class of ®rst-line drug
therapies for mid-stage and advanced-stage
acquired immunode®ciency syndrome (AIDS)
patients (Chiba et al 1997). In particular, combi-
nation therapy with two kinds of reverse tran-
scriptase inhibitor and an HIV-1 protease inhibitor
has been found to be more effective than either
drug alone in reducing levels of HIV-1 RNA,
increasing CD4 cell counts and preventing the
death of AIDS patients (Hoetelmans et al 1998).
Currently, four HIV-1 protease inhibitors, indina-
vir, saquinavir, ritonavir and nel®navir, are used in
clinical practice (Williams & Sinko 1999). Recent
studies on the metabolic fate of the HIV-1 protease
inhibitors showed that the most in¯uential isozyme
involved in the metabolism of the protease inhibi-
tors is CYP3A4, with the isomers of CYP2C9 and
CYP2D6 also contributing (von Molttke et al
1998). For instance, ritonavir is the strongest inhi-
bitor in-vitro, with an estimated Ki from 0Á95 to
1Á64 mM to the other three HIV-1 protease inhibitors
in experiments with rat liver microsomes (Yamaji
Correspondence: N. Shibata, Department of Pharmacoki-
netics, Kyoto Pharmaceutical University, 5 Nakauchi-cho
Misasagi Yamashina-ku, Kyoto 607-8414, Japan.

1240
NOBUHITO SHIBATA ET AL
In-vitro metabolic studies using rat liver
microsomal fractions
et al 1999). Ritonavir also potently inhibited the
CYP-mediated metabolites of indinavir, saquinavir
and nel®navir in human liver microsomes (Kou-
driakova at al 1998). Combination therapy with two
kinds of HIV-1 protease inhibitor has showed a
potent clinical effectiveness in preventing tolerance
by HIV-1 (Barry et al 1997). Therefore, it is con-
sidered that combination therapy with two HIV-1
protease inhibitors will be introduced increasingly
for the treatment of AIDS patients in clinical
practice. To establish the optimum combination, it
is necessary to examine the effects of drug inter-
actions on the metabolism of each combination of
two HIV-1 protease inhibitors.
In this report, using our newly developed assay
method for saquinavir and nel®navir by HPLC, we
investigated the aspects of drug interactions
between combinations of two kinds of HIV-1 pro-
tease inhibitor in rat both in-vitro and in-vivo. We
will refer to an optimal combination of two HIV-1
protease inhibitors.
After the dislocation of neck vertebrae, rat liver (5±
10 g) was excised quickly and perfused with ice-
cold potassium chloride (KCl) solution (1Á15%
w=v). Hepatic microsomes were prepared by dif-
ferential ultracentrifugation (Hayes et al 1995).
Brie¯y, rat liver was homogenized in 5 volumes of
1Á15% w=v KCl in a glass-Te¯on homogenizer kept
on ice. The homogenates were centrifuged at 4ꢀC at
9000 g for 15 min and the supernatant were cen-
trifuged again at 105 000 g for 1 h. Final micro-
somal pellets were resuspended in 0Á1 M phosphate
1
buffer to a concentration of 8Á0 mg protein mL .
Microsomal protein concentrations were deter-
mined by method of Lowry et al (1951) using
bovine serum albumin as a standard. The metabo-
lism of HIV-1 protease inhibitor was measured in
an NADPH-generating system according to the
method speci®ed below. To a clean 15-mL conical
glass tube, 100 mL of 5 mM NADPH in 0Á1 M
phosphate buffer (pH 7Á4), 100 mL of 50 mM G6P in
0Á1 M phosphate buffer, 4 mL of 500 units of
G6PDH in 0Á1 M phosphate buffer, 100 mL of
50 mM MgCl2 in 0Á1 M phosphate buffer, 586 mL of
0Á1 M phosphate buffer and 10 mL of HIV-1 pro-
tease inhibitor solution were added. When the
inhibition experiment was performed, each 5 mL of
HIV-1 protease inhibitor solution was added. After
a 5-min pre-incubation at 37ꢀC in a water bath, the
metabolic reaction was initiated by adding 100 mL
Materials and Methods
Materials
Glucose-6-phosphate (G6P), G6P dehydrogenase
(G6PDH) and nicotinamide adenosine dinucleotide
phosphate (NADP) were obtained from Sigma
Chemicals (St Louis, MO). Saquinavir was kindly
supplied from Hoffman-LaRoche Laboratory, and
nel®navir was extracted from capsules and was
puri®ed by preparative HPLC. Acetonitrile (HPLC
grade) and diethyl ether were supplied by Kanto
Chemical Co. Inc. (Tokyo, Japan). Bovine serum
albumin and sodium carmellose (CMC-Na) were
obtained from Nacalai Tesque Co. (Kyoto, Japan).
All other reagents used were of analytical grade
and were used without further puri®cation. The
standard stock solutions of protease inhibitors were
prepared by dissolving in methanol at a con-
of rat liver microsomal suspension (®nal con-
1
centration: 0Á8 mg protein mL
of microsomal
suspension). After incubation for 10 min, the reac-
tion was stopped by the addition of 100 mL of ice-
cold 2 M K3PO4. The mixture was vortexed vigor-
ously and centrifuged at 14 000 g for 10 min to
precipitate protein. The resultant supernatant was
used for the HPLC analysis of saquinavir and nel-
®navir.
Enzyme kinetic analysis
1
centration of 250±1000 mg mL . These solutions
The initial reaction rate of HIV-1 protease inhibitor
in rat liver microsomes was determined under lin-
ear conditions. Kinetic parameters, Michaelis con-
stant (Km) and maximum reaction rate (Vmax), were
determined by Lineweaver-Burk plot. The esti-
mates of the inhibition constant (Ki) were deter-
mined by plotting the slopes of the Lineweaver-
Burk plot versus inhibitor concentrations (Fitz-
simmons & Collins 1997).
were used to prepare standards for the calibration
curves of HPLC analysis. The calibration curve
samples were prepared by adding known amounts
of protease inhibitors to appropriate materials.
Animals and preparation of oral test solutions
Male Wistar rats, about 10 weeks old, were
obtained from Nippon SLC Co. Ltd. (Hamamatsu,
Japan). The rats were housed in pairs at least seven
days under controlled environmental conditions;
general food and water were freely available. Each
HIV-1 protease inhibitor (5 g) was suspended with
10 mL of 2% (w=v) CMC-Na solution.
In-vivo pharmacokinetic studies using rats
Three or four rats, 350Æ 10 g, fasted overnight with
free access to water for at least 12 h, received an

PHARMACOKINETIC INTERACTIONS BETWEEN HIV-1 PROTEASE INHIBITORS
1
1241
oral dose of HIV-1 protease inhibitors. Rats were
anaesthetized with light ether anaesthesia. At
30 min before the administration of drug, 0Á2 mL of
blank blood samples were withdrawn from the
external left jugular vein by direct puncture using a
syringe equipped with a 26-gauge needle after
incision. The oral dose of each HIV-1 protease
each drugs was 0Á1 mg mL . The extraction and the
detection of indinavir were performed using our
method reported previously.
Pharmacokinetic analysis
A non-compartmental pharmacokinetic analysis
was applied to the plasma concentration±time data
using a computer program, WinHARMONY
(Yoshikawa et al 1998). The terminal elimination
rate constant, lz, was determined by a linear
regression of at least three data points from the
terminal portion of the plasma concentration±time
plots. The area under the plasma concentration±
time curve after oral administration, AUC, was
calculated using the linear trapezoidal rule up to the
last measured plasma concentration, Cp(last), and
extrapolated to in®nity using a correction term,
1
inhibitor was 20 mg kg , and drug suspensions
1
(0Á5 mg mL ) with 2% CMC-Na were orally
administered by a feeding tube within a con-
secutive 3-min interval. Then, 0Á2-mL samples of
blood were collected into heparinized centrifuging
tubes from the external left jugular vein at 0Á5, 1,
1Á5, 2, 3, 4, and 6 h after the injection. Plasma
samples were obtained by centrifuging blood
samples at 9000 g for 30 min, and were immedi-
ately frozen in a deep freezer at 80ꢀC until ana-
lysis.
namely Cp(last)lz
1. The area under the ®rst-
moment curve to the last measured plasma con-
centration (AUMC) was also calculated using the
linear trapezoidal rule and the addition of the
Extraction procedure and analytical conditions for
HIV-1 protease inhibitors
concentration term after the last measured point
The resultant aqueous layer (380 mL) from the rat
microsomal experiment and plasma (100 mL) were
diluted with phosphate buffer to make 1 mL of
sample solution. These specimens were extracted
with 3 mL of diethyl ether. The mixture was shaken
for 10 min at 9000 g. After the aqueous phase had
been frozen in a cold bath at 10ꢀC, the ether
extract was transferred to a clean test tube. The
organic liquid was evaporated to dryness at 50ꢀC
under a stream of nitrogen gas. The determinations
of saquinavir and nel®navir in sample specimens
were performed by an HPLC method. The liquid
chromatograph (Shimadzu LC-10AS, Kyoto,
Japan) was equipped with an automatic injector
(Toso Model 8020), a UV detector (Shimadzu
SPD-10A) and a chromatographic terminal (Chro-
matec, Kyoto, Japan). The analytical column for
saquinavir and nel®navir was a Chemcosorb
5-ODS-UH (4Á6 mm i.d. 6 250 mm). The column
was maintained at 65ꢀC for all separations. Elution
was carried out isocratically at a ¯ow-rate of
1
(t(last)) to in®nity, namely, t(last)Cp(last) z
l
‡
1
=
2
Cp(last) z . The terminal elimination half-life, t 2,
l
was determined by dividing ln2 by lz. The mean
residence time, MRT, was calculated as AUMC=
AUC. The apparent clearance, CLapp, was calcu-
1
lated by Doral AUC , where Doral represents the
oral dose.
Statistics
All values are expressed as meansÆ s.e.m. unless
otherwise noted. Statistical differences of the
means were assumed to be signi®cant when
P < 0Á05 (two-sided t-test).
Results
Figure 1 shows Lineweaver-Burk and Dixon plots
for the inhibition of saquinavir metabolism in rat
liver microsomes by nel®navir. The saquinavir
concentrations employed as the substrate were
0Á75±6Á0 mM, and the nel®navir concentrations as
the inhibitor were 7Á5±30Á0 mM. The kinetic para-
meter values (i.e., Michaelis constant, Km and
maximum velocity, Vmax) of saquinavir were
1
1Á1 mL min with a mixture of acetonitrile and
water (32Á5 : 67Á5, v=v). The mobile phase was
degassed before use. The detection of saquinavir
and nel®navir was performed at a UV wavelength
of 210 nm and 245 nm, respectively, and the
retention times of these drugs were 15 min and
12Á5 min, respectively. The ®nal extraction residues
were reconstituted with 200 mL of the mobile phase
and 100 mL of sample was injected to the HPLC
system. For the quantitative determination of drugs,
the absolute calibration method was used. The
correlation coef®cients for the calibration curves
were 0Á999 or better, and the detection limits for
estimated by a linear regression analysis and were
8Á31Æ 0Á18 mM and 1Á42Æ 0Á07 nmol min
1
(mg
protein) , respectively. The metabolic clearance
1
rate (Vmax=Km) of saquinavir was 170Á9Æ
1
1
10Á9 mL min
(mg protein)
(Figure 1A). With
the addition of nel®navir, the metabolism of
saquinavir was inhibited, and the inhibition

1242
NOBUHITO SHIBATA ET AL
Figure 1. Lineweaver-Burk (A) and Dixon (B) plots for inhibition of saquinavir metabolism by nel®navir in rat liver microsomes.
Lineweaver-Burk plot revealed a competitive interaction between saquinavir and nel®navir. The Ki value of nel®navir for
saquinavir is 2Á35 mM. A. r, saquinavir alone; j, with 7Á5 mM nel®navir; m, with 15 mM nel®navir; d, with 30 mM nel®navir. B. r,
0Á75 mM saquinavir; j, 1Á5 mM saquinavir; m, 3Á0 mM saquinavir; d, 6Á0 mM saquinavir.
Figure 2. Lineweaver-Burk (A) and Dixon (B) plots for inhibition of nel®navir metabolism by saquinavir in rat liver microsomes.
Lineweaver-Burk plot revealed a competitive interaction between nel®navir and saquinavir. The Ki value of saquinavir for
nel®navir is 5Á22 mM. A. r, saquinavir alone; j, with 7Á5 mM nel®navir; m, with 15 mM nel®navir; d, with 30 mM nel®navir. B. r,
0Á75 mM saquinavir; j, 1Á5 mM saquinavir; m, 3Á0 mM saquinavir; d, 6Á0 mM saquinavir.
1
constant, Ki of nel®navir was estimated to be
2Á35 mM (Figure 1B). Figure 2 shows Lineweaver-
Burk and Dixon plots for the inhibition of nel®navir
metabolism in rat liver microsomes by saquinavir.
The nel®navir concentrations employed as the
substrate were 0Á75±6Á0 mM, and the saquinavir
concentrations as the inhibitor were 5Á0±15Á0 mM.
The estimated values of Km and Vmax of nel®navir
(mg protein) (Figure 2A). With the addition of
saquinavir, the metabolism of nel®navir was also
inhibited, and the value of Ki was estimated to be
5Á22 mM for nel®navir (Figure 2B). As previously
reported (Yamaji et al 1999), the values of Km and
Vmax for indinavir were 4Á52Æ 0Á08 mM and
1
1
0Á33Æ 0Á03 nmol min
(mg protein) , respec-
tively, and the value of Vmax=Km was
1
1
1
(mg protein) . From these
were 5Á87Æ 0Á51 mM and 0Á74Æ 0Á05 nmol min
73Á0Æ 2Á0 mL min
1
(mg protein) , respectively. The value of
parameters, the metabolic clearance rates of
saquinavir, nel®navir and indinavir were deter-
1
Vmax=Km for nel®navir was 126Á1Æ 4Á4 mL min

PHARMACOKINETIC INTERACTIONS BETWEEN HIV-1 PROTEASE INHIBITORS
1243
Figure 4. Effect of HIV-1 protease inhibitors (indinavir,
nel®navir, ritonavir) on the plasma concentration-time pro®les
of saquinavir after co-administration to rats. r, saquinavir
alone; j, with indinavir; m, with nel®navir; d, with ritonavir.
The oral doses of saquinavir, indinavir, nel®navir and ritonavir
were 20 mg kg 1, respectively. Each symbol with bar repre-
sents the meanÆ s.e. of three or four experiments.
Figure 3. Summary of the inhibition constants, Ki's, of four
HIV-1 protease inhibitors in the rat liver microsomes. The
values of Ki, except for the mutual values between saquinavir
and nel®navir, were quoted from our previous report (Yamaji
et al 1999).
mined and the rank order of those values was
saquinavir > nel®navir > indinavir.
Figure 3 summarizes all Ki values between the
four HIV-1 protease inhibitors. In this ®gure, we
combined the Ki values obtained in both our pre-
vious study (Yamaji et al 1999) and in this study.
Ritonavir is a potent metabolic inhibitor of
CYP3A4, and several investigators have studied the
metabolic inhibition of ritonavir on saquinavir,
indinavir and nel®navir; however, none of these
HIV-1 protease inhibitors altered the metabolism of
ritonavir in-vitro or in-vivo (Koudriakova et al
1998). Our results (Figure 3) were consistent with
those in the literature, that is, ritonavir exhibited
smallest values of Ki: 1Á64 mM for saquinavir,
1Á01 mM for nel®navir and 0Á95 mM for indinavir.
The Ki values of nel®navir were in a middle posi-
tion of the four HIV-1 protease inhibitors: 2Á35 mM
for saquinavir and 2Á14 mM for indinavir. The Ki
values of indinavir for saquinavir and nel®navir
were 3Á55 mM and 2Á76 mM, respectively, and these
values were relatively large.
On the basis of in-vitro metabolic studies, we
examined the in-vivo pharmacokinetic interactions
between various combinations of two HIV-1 pro-
tease inhibitors. Figure 4 shows the effects of
indinavir, nel®navir and ritonavir on the pharma-
cokinetics of saquinavir after oral administration
and Figure 5 shows the effects of indinavir,
saquinavir and ritonavir on the pharmacokinetics of
nel®navir after oral administration to rats. The
pharmacokinetic parameters based on the non-
compartmental pharmacokinetic analysis are sum-
marized in Table 1, where the data of pharmaco-
kinetic parameters for indinavir after co-
administration with nel®navir, saquinavir or rito-
navir in our previous report (Yamaji et al 1999)
were included. When saquinavir, indinavir or nel-
Figure 5. Effect of HIV-1 protease inhibitors (saquinavir,
indinavir, ritonavir) on the plasma concentration-time pro®les
of nel®navir after co-administration to rats. r, saquinavir
alone; j, with indinavir; m, with nel®navir; d, with ritonavir.
The oral doses of saquinavir, indinavir, nel®navir and ritonavir
were 20 mg kg 1, respectively. Each symbol with bar repre-
sents the meanÆ s.e.m. of three or four experiments.
®navir were administered alone to rats: the max-
imum plasma drug concentration (Cmax) values,
were 0Á05Æ 0Á02, 0Á80Æ 0Á29 and 0Á09Æ 0Á01
1
1
=
mg mL ; the t 2 values were 2Á14Æ 0Á28, 0Á58
Æ 0Á06 and 1Á36Æ 0Á11 h; the AUCs were
1
0Á15Æ 0Á03, 1Á07Æ 0Á31 and 0Á17Æ 0Á02 mg h mL ;
and the CLapp values were 34Á9Æ 5Á2, 7Á8Æ 0Á1 and
1
32Á1Æ 2Á5 L h , respectively. As a result of co-
administration of ritonavir, the AUCs of saquinavir,
indinavir and nel®navir were signi®cantly
increased from 0Á15Æ 0Á03 to 6Á88Æ 1Á87
(P < 0Á01), from 1Á07Æ 0Á31 to 10Á01Æ 2Á88
(P < 0Á05)
and
from
0Á17Æ 0Á02
to
1
1Á44Æ 0Á41 mg h mL
(P < 0Á05), respectively,
while the CLapp was signi®cantly decreased
(P < 0Á01) from 34Á9Æ 5Á2 to 0Á9Æ 0Á3, from
7Á8Æ 0Á1 to 0Á5Æ 0Á1 and from 32Á1Æ 2Á5 to
1
1Á3Æ 0Á6 L h , respectively. When co-adminis-

1244
NOBUHITO SHIBATA ET AL
Table 1. Pharmacokinetic parameter values of saquinavir, indinavir and nel®navir in combination.
Pharmacokinetic parameter
1
1
1
CLapp (L h )
Cmax (mg mL
0Á05Æ 0Á02
)
AUC (mg h mL
0Á15Æ 0Á03
)
t12 (h)
MRT (h)
Saquinavir alone
‡ Indinavir
2Á14Æ 0Á28
1Á26Æ 0Á09
1Á96Æ 0Á54
2Á85Æ 1Á21
0Á58Æ 0Á06
0Á51Æ 0Á02**
0Á53Æ 0Á02**
2Á94Æ 0Á72*
1Á36Æ 0Á11
1Á20Æ 0Á09
1Á10Æ 0Á07
8Á67Æ 1Á96**
3Á90Æ 0Á65
2Á39Æ 0Á27
3Á17Æ 0Á80
5Á04Æ 1Á63*
1Á62Æ 0Á17
0Á91Æ 0Á06*
0Á81Æ 0Á04*
4Á85Æ 1Á25*
2Á95Æ 0Á26
2Á42Æ 0Á25
2Á26Æ 0Á06
13Á18Æ 3Á17**
34Á9Æ 5Á2
0Á61Æ 0Á02**
0Á10Æ 0Á02
1Á52Æ 0Á25**
0Á47Æ 0Á07
4Á2Æ 0Á7**
27Á1Æ 5Á5**
0Á9Æ 0Á3**
7Á8Æ 0Á1
‡ Nel®navir
‡ Ritonavir
2Á62Æ 0Á79*
0Á80Æ 0Á29
6Á88Æ 1Á87**
1Á07Æ 0Á31
Indinavir alone
‡ Nel®navir
‡ Saquinavir
‡ Ritonavir
7Á16Æ 1Á19**
6Á76Æ 1Á08**
4Á22Æ 2Á33*
0Á09Æ 0Á01
7Á27Æ 1Á92*
6Á27Æ 1Á27**
10Á01Æ 2Á88*
0Á17Æ 0Á02
0Á9Æ 0Á2*
1Á0Æ 0Á2*
0Á5Æ 0Á1**
32Á1Æ 2Á5
16Á6Æ 2Á2*
5Á4Æ 0Á7**
1Á3Æ 0Á6**
Nel®navir alone
‡ Indinavir
0Á15Æ 0Á03
0Á37Æ 0Á06
‡ Saquinavir
‡ Ritonavir
0Á39Æ 0Á05
1Á12Æ 0Á18
0Á33Æ 0Á10**
1Á44Æ 0Á41*
The data for indinavir after co-administration with nel®navir, saquinavir or ritonavir in our previous report (Yamaji et al 1999)
were included. Each value represents the meanÆ s.e.m. of three or four experiments. *P < 0Á05, **P < 0Á01, compared with
appropriate control.
tered with indinavir, the Cmax of saquinavir was
signi®cantly increased (P < 0Á01) from 0Á05Æ 0Á02
to 0Á61Æ 0Á02 mg mL 1 and the AUC also increased
peaks of saquinavir and nel®navir could not be
separated on the chromatograms. In this study,
good simultaneous separation of saquinavir and
nel®navir on a chromatogram was attained by use
of another kind of reversed-phase column, namely,
Chemcosorb 5-ODS-UH with a 30% carbon con-
1
from 0Á15Æ 0Á03 to 1Á52Æ 0Á25 mg h mL , while
the CLapp was markedly decreased (P < 0Á01) from
1
34Á9Æ 5Á2 to 4Á2Æ 0Á7 L h . However, the effect of
Ê
nel®navir on the plasma concentration±time pro®le
of saquinavir was less than that of indinavir,
namely, the Cmax and AUC of saquinavir were
increased about 2- and 3Á1-fold with co-adminis-
tration of nel®navir, although there were no dif-
ferences in values of CLapp. With co-administration
of saquinavir, the Cmax of nel®navir was increased
tent and silica gel of 60 A pore size. Therefore, we
could describe the intensity of interaction between
combinations of two kinds of HIV-1 protease
inhibitor: ritonavir=saquinavir, ritonavir=nel®navir,
ritonavir=indinavir, saquinavir=nel®navir, saquina-
vir=indinavir and nel®navir=indinavir.
The Eadie-Hofstee plots analysis for the meta-
bolism of saquinavir, indinavir and nel®navir
showed that only one enzyme took part in the
metabolism of these HIV-1 protease inhibitors by
rat liver microsomes in the concentration range
employed in this study. Therefore, the metabolic
competitive inhibition between these four HIV-1
protease inhibitors could be elucidated by their
metabolic clearance rates. Increases in plasma
concentrations of HIV-1 protease inhibitors as a
result of drug interaction between ritonavir and the
other three HIV-1 protease inhibitors is well
established and is believed to be due to inhibition
of CYP3A4-mediated metabolism by ritonavir.
Saquinavir, nel®navir and indinavir also have
inhibitory effects on CYP3A4-mediated metabo-
lism (Kempf et al 1997). The values of metabolic
1
from 0Á09Æ 0Á01 to 0Á39Æ 0Á05 mg mL and the
AUC also increased from 0Á17Æ 0Á02 to
1
1Á12Æ 0Á18 mg h mL , while the CLapp sig-
ni®cantly (P < 0Á01) decreased from 32Á1Æ 2Á5 to
1
5Á4Æ 0Á7 L h . However, the effect of indinavir on
the plasma concentration-time pro®le of nel®navir
was less than that of saquinavir. The effect of
nel®navir on the pharmacokinetic pro®les of indi-
navir were of a comparable degree to that of
saquinavir. The time to reach peak concentrations
of saquinavir and nel®navir tended to increase as
their AUCs were increased in the presence of other
HIV-1 protease inhibitors.
Discussion
clearance (Vmax=Km) for ritonavir were reported to
(Molla et al 1996). The
1
be 21Á7 mL min 1 mg
In our previous report on drug interactions of
HIV-1 protease inhibitors (Yamaji et al 1999), we
assayed the HIV-1 protease inhibitors in rat plasma
and liver microsomes by HPLC methods using a
reversed-phase column with a 20% carbon content
rank order of the Vmax=Km for the four HIV-1
protease inhibitors are saquinavir > nel®navir >
indinavir > ritonavir. Therefore, co-administration
of ritonavir would decrease the metabolic clearance
of saquinavir, nel®navir and indinavir. In addition,
the value of CLapp reported for ritonavir is
Ê
and silica gel with a pore size of 100 A (Chemco-
sorb 5-ODS-H). Using this column, however, the

PHARMACOKINETIC INTERACTIONS BETWEEN HIV-1 PROTEASE INHIBITORS
1245
1
0Á351 L h (Kempf et al 1997). The rank order of
transporter family, and is localized at the apical
surface of secretion in various tissues (Pajeva et al
1996). It appears to act as a general detoxi®cation
system protecting tissues from endogenous or
exogenous lipophilic compounds (Ecker & Chiba
1997). Alsenz et al (1998) con®rmed that saqui-
navir and ritonavir were both substrates for an
ef¯ux mechanism in Caco-2 cells, most likely P-
glycoprotein, which acts as a counter-transporter
for both drugs. In addition, Pro®t et al (1999)
reported that saquinavir is a substrate for P-glyco-
protein and that ritonavir, nel®navir and indinavir
modulate P-glycoprotein function in both human
lymphocytes and Caco-2 cells. These observations
indicate that ritonavir, saquinavir, nel®navir and
indinavir are all potent modulators for P-glyco-
protein. However, the degree of intensities of these
drugs as modulators of P-glycoprotein that is an
localized at the apical surface of the small intestine
are unclear. It is evidenced that the co-function of
P-glycoprotein and CYP3A4 in small intestine is an
important factor to oral drug delivery. P-glycopro-
tein increases the exposure to drug-metabolizing
enzymes and enhances intestinal metabolism of
drugs by prolonging the intercellular residence time
through the repetitive processes of excretion and
reabsorption (Benet et al 1996). The possible con-
tributions of P-glycoprotein in the intestinal meta-
bolism was further supported by the observations
that the time to reach peak concentrations of
saquinavir and nel®navir were increased as their
AUCs were increased in the presence of other HIV-
1 protease inhibitors. Consequently, to explain the
contradictions that exist between in-vivo and in-
vitro results of the interactions of nel®navir with
the other three HIV-1 protease inhibitors, it is
necessary to perform further detailed experiments
on the absorption process of these HIV-1 protease
inhibitors via P-glycoprotein traversing the plasma
membrane and CYP3A4 existing in the inside of
enterocytes.
CLapp was saquinavir > nel®navir > indinavir >
ritonavir, and this order is the same as that of
metabolic clearance in-vitro. Therefore, the co-
administration of ritonavir decreased the CLapp
values of saquinavir, nel®navir and indinavir, and
consequently increased their Cmax and t 2. More-
over, von Molttke et al (1998) reported that the
rank order of the inhibitory effects of these HIV-1
protease inhibitors on the 4-hydroxylation of tria-
zolam, which is mainly metabolized by CYP3A4 in
human hepatic microsomes, was ritonavir >
indinavir > nel®navir > saquinavir. Hence, it was
assumed that the rank order of inhibitory effects of
these drugs gained from our in-vitro data in rat liver
microsomes essentially agreed with the human in-
vitro data.
Using an HPLC method we newly developed, we
could make the drug interactions between the
combinations of two kinds of HIV-1 protease
inhibitor clear. Judging by the degree of inhibitory
intensity in the in-vitro rat liver microsomes, the
inhibitory effect for the metabolism of saquinavir
as a substrate was in the rank order of ritona-
vir > nel®navir > indinavir, while that for indinavir
as a substrate was in the rank order of ritona-
vir > nel®navir > saquinavir. Judging by the phar-
macokinetic parameters, the increasing effect for
the bioavailability of saquinavir after co-adminis-
tration in-vivo was in the rank order of ritona-
vir ꢁ nel®navir > indinavir, while that of indinavir
was in the rank order of ritonavir > nel®navir >
saquinavir. Accordingly, for combinations of
saquinavir with the other three HIV-1 protease
inhibitors (ritonavir, nel®navir, indinavir) or indi-
navir with the other three HIV-1 protease inhibitors
(ritonavir, nel®navir, saquinavir), the in-vitro
results re¯ect the in-vivo results. For nel®navir,
however, the inhibitory effect on the metabolism of
nel®navir as a substrate in-vitro was in the rank
order of ritonavir > indinavir > saquinavir. How-
ever, the enhancement of nel®navir bioavailability
after co-administration in-vivo was in the rank
order of ritonavir > saquinavir > indinavir. Clearly,
in the case of nel®navir with the other three HIV-1
protease inhibitors (ritonavir, saquinavir, indina-
vir), the in-vitro results did not re¯ect the in-vivo
results, suggesting the participation of other inter-
actions besides the metabolic process in the liver.
The oral bioavailability of saquinavir is variable
within and between patients (Noble & Faulds
1996). As one reason for this low oral bioavail-
ability, the absorption is decreased by an active
ef¯ux pump in the intestine such as P-glycoprotein.
P-glycoprotein is a 170-kDa transmembrane pro-
tein that is a member of the ATP binding cassette
1
=
With the clinical use of HIV-1 protease inhibi-
tors, the chemotherapy of AIDS patients has pro-
gressed day by day. Indeed, combination therapy
with two kinds of reverse transcriptase inhibitors
and an HIV-1 protease inhibitor, namely, hyper-
active anti-retroviral therapy, has markedly
improved the survival of AIDS patients (Finzi et al
1999). Combination therapy using HIV-1 protease
inhibitors provides greater viral suppression, and
several regimens with co-administration of two
kinds of protease HIV-1 inhibitors, including
quadruple therapy, are now undergoing clinical
trial (Barry et al 1997). Ritonavir, saquinavir and
indinavir are peptide-mimetic compounds, and
their metabolic clearances are very extensive in the

1246
NOBUHITO SHIBATA ET AL
Kempf, D. J., Marsh, K. C., Kumar, G., Rodrigues, A. D.,
Denissen, J. F., McDonald, E., Kukulka, M. J., Hsu, A.,
Granneman, G. R., Baroldi, P. A., Sun, E., Pizzuti, D., Plattner,
J. J., Norbeck, D. W., Leonard, J. M. (1997) Pharmacoknetic
enhancement of inhibitors of the human immunode®ciency
virus protease by coadministration with ritonavironavir. Anti-
microb. Agents Chemother. 41: 654±660
Koudriakova, T., Iatsimirskaia, E., Utkin, I., Gangl, E.,
Vouros, P., Storozhuk, E., Orza, D., Marinina, J., Gerber,
N. (1998) Metabolism of the human immunode®ciency
virus protease inhibitors indinavir and ritonavir by human
intestinal microsomes and expressed cytochrome
P45503A4=3A5: mechanism-based inactivation of cyto-
chrome P4503A by ritonavir. Drug Metab. Dispos. 26:
552±561
body. Therefore, chronic use of these compounds
will provide undesirable side-effects on the liver.
Nel®navir, however, is not a peptide-mimetic
compound, and is thought to exhibit less side-
effects on the liver. From this point of view, we
recommend nel®navir as a drug of ®rst choice
among these four HIV-1 protease inhibitors.
Moreover, taking both the in-vitro and in-vivo
results into consideration, the best combination of
two protease inhibitors for the therapy of AIDS
patients is ®rstly nel®navir=saquinavir, and sec-
ondly nel®navir=indinavir. This study provides
useful information for the treatment of AIDS
patients when they receive a combination therapy
of two kinds of HIV-1 protease inhibitors.
Lowry, O. H., Rosebrough, N. J., Farr, A. L., Randall, R. J.
(1951) Protein measurement with the folin phenol agent. J.
Biol. Chem. 193: 265±275
Molla, A., Korneyeva, M., Gao, Q., Vasavanonda, S., Schip-
per, P. J., Mo, H. M., Markowitz, M., Chernyavskiy, T., Niu,
P., Lyons, N., Hsu, A., Grannerman, G. R., Ho, D. D.,
Boucher, C. A. B., Leonard, J. M., Norbeck, D. W., Kempf,
D. J. (1996) Ordered accumulation of mutations in HIV
protease confers resistance to ritonavir. Nat. Med. 2:
760±766
Noble, S., Faulds, D. (1996) Saquinavir, a review of its
pharmacology and clinical potential in the management of
HIV infection. Drugs 52: 93±112
Pajeva, I. K., Wiese, M., Cordes, H. P., Seydel, J. K.
(1996) Membrane interactions of some catamphi-
philic drugs and relation to their multidrug-resistence-
reversing ability. J. Cancer Res. Clin. Oncol. 122:
27±40
References
Alsenz, J., Steffen, H., Alex, R. (1998) Active apical secretory
ef¯ux of the HIV protease inhibitors saquinavir and ritonavir
in Caco-2 cell monolayers. Pharm. Res. 15: 423±428
Barry, M., Gibbons, S., Back, D., Mulcahy, F. (1997) Protease
inhibitors in patients with HIV disease. Clin. Pharmacokinet.
32:194±209
Benet, L. Z., Wu, C. Y., Hebert, M. F., Wacher, V. J. (1996)
Intestinal drug metabolism and antitransport process: a
potential shift in oral drug delivery. J. Control. Release
39: 139±143
Pro®t, L., Eagling, V. A., Back, D. J. (1999) Modulation of P-
glycoprotein function in human lymphocytes and Caco-2
cell monolayers by HIV-1 protease inhibitors. AIDS 13:
1623±1627
Chiba, M., Hensleighh, M., Lin, J. H. (1997) Hepatic and
intestinal metabolism of indinavir, an HIV protease inhibi-
tor, in rat and human microsomes. Biochem. Pharm. 53:
1187±1194
Yamaji, H., Matsumura, Y., Yoshikawa, Y., Takada, K. (1999)
Pharmacokinetic interactions between HIV-protease inhibi-
tors in rats. Biopharm. Drug Dispos. 20: 241±247
Yoshikawa, Y., Kato, K., Sone, H., Takada, K. (1998)
Development and evaluation of noncompartmental pharma-
cokinetic analysis program `WinHARMONY' using
Visual BASIC language having a function of automatic
recognition v time pro®le. Jpn. J. Clin. Pharmacol. Ther.
29: 475±487
von Molttke, L. L., Greenblatt, D. J., Grassi, J. M., Granda, B.
W., Duan, S. X., Fogelmmann, S. M., Daily, J. P., Harmatz,
J. S., Shader, R. I. (1998) Protease inhibitors as inhibitors of
human cytochromes P450: high risk aassociated with rito-
navir. J. Clin. Pharmacol. 38: 106±111
Ecker, G., Chiba, P. (1997) Structure±activity-relationship
studies on modulators of the multidrug transporter P-glyco-
proteinÐan overview. Wien. Klin. Wochenschr. 107=22:
681±686
Finzi, D., Blankson, J., Siliciano, J. D. (1999) Latent infection
of CD4 ‡ T cells provides a mechanism for lifelong
persistence of HIV-1, even in patients on effective combina-
tion therapy. Nat. Med. 5: 512±519
Fitzsimmons, M. E., Collins, J. M. (1997) Selective biotrans-
formation of the human immunode®ciency virus protease
inhibitor saquinavir by human small-intestinal cytochrome
P4503A4. Drug Metab. Dispos. 25: 256±666
Hoetelmans, R. M. W., Reijers, M. H. E., Weverling, G. J.,
Kate, R. W., Wit, F. W. N. M., Mulder, J. W., Weigel, H.
M., Frissen, P. H. J., Lange, J. M. A. (1998) The effect of
plasma drug concentrations on HIV-1 clearance rate during
quadruple drug therapy. AIDS 12: F111±F115
Williams, G. C., Sinko, P. J. (1999) Oral absorption of the HIV
protease inhibitors: a current update. Adv. Drug Del. Rev.
39: 211±238