Pharmacokinetic study of all‐trans‐retinoyl‐β‐d ‐glucuronide in Sprague–Dawley rats after single and multiple intravenous administration(s)

Pharmacokinetic Study of All-Trans-Retinoyl-b-D-Glucuronide in Sprague±Dawley Rats After Single and Multiple Intravenous Administration(s) HAI-SHU LIN,1 ARUN B. BARUA,2 JAMES A. OLSON,2 KERWIN SIEW YANG LOW,3 SUI YUNG CHAN,1 MEI LENG SHOON,3 PAUL C. HO1 1

Department of Pharmacy, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260

2

Department of Biochemistry, Biophysics, and Molecular Biology, Iowa State University, Ames, Iowa 50011

3

Department of Pharmacology, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260

Received 28 December 2000; revised 5 June 2001; accepted 8 June 2001

ABSTRACT: All-trans-retinoyl-b-D-glucuronide (RAG) is an endogenous active metabolite of all-trans-retinoic acid (ATRA). In the present study, the pharmacokinetics of RAG was examined after the administration of a single intravenous does (5, 10, or 15 mmol/kg) and of multiple daily intravenous doses (5 mmol/kg) to rats for 8 days. The plasma concentrations of RAG and ATRA were measured by a reverse-phase HPLC method. A rapid distribution phase of approximately 1 h was observed in all of the rats after single or multiple doses. Thereafter, RAG was eliminated through a ®rst-order process, in accord with a typical two-compartment ®rst order pharmacokinetic pro®le. After single intravenous doses, the AUC of RAG increased proportionally with the dose and the clearance remained unchanged within the tested doses. There was no statistical signi®cant difference in distribution rate constants from central compartment to peripheral compartment (K12) and from peripheral compartment to central compartment (K21) between different doses. However, as the dose increased from 5 mmol/kg to 10 mmol/kg, the volume of distribution at the steady state (Vss) and the volume of peripheral compartment (Vp) decreased signi®cantly ( p < 0.05) from 1.290  0.269, 0.928  0.232. L/kg to 0.961  0.149, 0.647  0.107 L/kg, respectively. Vss and Vp at a dose of 15 mmol/kg (0.924  0.187, 0.698  0.165 L/kg) were not signi®cantly different from that at 10 mmol/kg. Thus, RAG might saturate the tissue-binding sites at higher doses. ATRA was detected as a metabolite of RAG at low levels (usually < 0.05 mM) only in the ®rst 2 h after intravenous administration. RAG clearly was not extensively hydrolyzed to ATRA in our study. After multiple daily intravenous administration of RAG, the clearance (Cl) and the elimination rate constant (K10) remained unchanged (p > 0.05), indicating that long-term daily administration of RAG did not induce its accelerated metabolism. However, K12, Vp, and Vss declined signi®cantly ( p < 0.05) from 1.67  0.54 hÿ1, 0.928  0.232 L/kg, and 1.290  0.269 L/kg to 0.96  0.48 hÿ1, 0.494  0.147 L/kg, and 0.818  0.187 L/kg, respectively. Therefore, long-term daily dosing of RAG seemed to decrease its distribution pro®le. Although the AUC of RAG did not change signi®cantly after multiple dosing, the AUC of ATRA after RAG dosing signi®cantly declined ( p < 0.05) from 0.032  0.019 mM
h to 0.010  0.006 mM
h. The decline in the AUC of ATRA might re¯ect an increase in its uptake by tissue and/or in its metabolism. Because enhanced clearance is not associated with RAG after multiple Correspondence to: P.C. Ho (Telephone: (65) 8742651; Fax: (65) 7791554; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 90, 2023±2031 (2001) ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

2023

2024

LIN ET AL.

administrations, RAG could be considered as an alternate to ATRA in appropriate clinical applications. ß 2001 Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 90:2023±2031, 2001

Keywords:

RAG; ATRA; intravenous injection; pharmacokinetics; leukemia

INTRODUCTION Acute promyelocytic leukemia (APL), the M3 subtype according to the French±American± British classi®cation, accounts for about 10% of all myeloid leukemias.1±5 APL is characterized by the failure of promyelocytes to undergo normal differentiation into mature granulocytes.1±5 In most patients, APL is associated with a reciprocal chromosomal translocation between the 15th and 17th chromosome.1,3,4 In clinical practice, alltrans-retinoic acid (ATRA, vitamin A acid, tretinoin), an endogenous oxidative metabolite in the vitamin A pathway, has been used either singly or in combination with other compounds to treat APL.1±6 Compared with traditional antineoplastic chemotherapy, which is both toxic to normal cells and to leukemic cells, ATRA can selectively induce the terminal differentiation or apoptosis of leukemic cells without causing marrow hypoplasia or exacerbation of the frequently occurring hemorrhagic syndrome.1 Most newly diagnosed APL patients go into complete remission at the beginning of ATRA treatment.1±6 However, with continuous daily administrations of ATRA, the level of ATRA in the plasma declines rapidly, such that most patients develop drug resistance to ATRA and many relapse.1±3,5,6 Furthermore, about 25% of patients develop the toxic ``retinoic acid syndrome,'' where the sequelae can be fatal.1,3 As ATRA is almost insoluble in aqueous solution, the intestinal absorption of orally administered ATRA is affected both by the pH and, by the fatty acid composition of the intraluminal bile.6 Therefore, plasma concentrations following an oral dose of ATRA are highly variable.7,8 These drawbacks limit the applications of ATRA in the treatment of APL and other cancers. All-trans-retinoyl-b-D-glucuronide (RAG) was ®rst identi®ed as an endogenous conjugated metabolite of all-trans-retinoic acid (ATRA) in bile, after large oral doses of ATRA were given to many different species.9,10 RAG is present in human plasma at low concentrations (5±17 nM) even without ATRA dosing.9±11 RAG supports both growth and epithelial cell differentiation.9,10 In clinical trials carried out in the United States and India, RAG was approximately equipotent to

ATRA in the topical treatment of acne without showing the toxic side effects of ATRA.9,10,12,13 Similar to ATRA, RAG effectively inhibits the growth and induces the terminal differentiation of the acute myeloblastic leukemic cell lines, HL60 (M2 subtype) and NB4 (M3 subtype).9,10,14±16 However, compared to ATRA, RAG is much more water soluble and much less toxic to NB4 and HL60 cells, even at very high concentrations (10ÿ5 M).9,16 Thus, the clinical side effects of RAG are predictably less than those of ATRA. Most importantly, RAG does not detectably bind to cellular retinoic acid binding proteins (CRABPs),9,17 which are the carriers in the transport of ATRA to its metabolizing enzymes.2,6 The induction both of CRABPs and its cytochrome P450 (CYP) metabolizing enzymes might well lead to clinical retinoic acid resistance.3,6 Because the current information about the pharmacokinetics of RAG is sparse, we decided to explore the detailed pharmacokinetics of RAG after single and multiple intravenous administration(s) into rats.

EXPERIMENTAL SECTION Special Precautions To prevent the isomerization of retinoids, all laboratory procedures involving the manipulation of retinoids were executed in a dimly lit environment.6 Materials All-trans-retinoyl-b-D-glucuronide (RAG) was synthesized by a previously published method.18 All-trans-retinoic acid (ATRA) was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). 2-Hydroxypropyl-b-cyclodextrin (HPbCD) (MW: approx. 1386) was obtained from Roquette Co. (62136 Lestrem, France). Ammonium acetate and sodium chloride were purchased from Sigma Chemical Co. (St. Louis, MO). HPLC grade methanol and acetonitrile were obtained from Fisher Scienti®c Co. (Fair Lawn, NY). Milli-Q water was used to prepare all aqueous solutions used in this study.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

ALL-TRANS-RETINOYL-b-D-GLUCURONIDE IN SPRAGUE±DAWLEY RATS

Analytical Assay RAG and ATRA were detected and quantitated by Shimadzu HPLC system. The system con®guration included a SCL-10A VP system controller, a SPD-10A VP UV-VIS detector and a LC-10AT VP isocratic pump. This HPLC system was controlled by an IBM Personal Computer 300GL. Separation of retinoids was made on a reversed-phase column (ODS-Hypersil, 5 mm, 250  4 mm, Hewlett Packard, Waldbronn, Germany), which was protected by a precolumn ®lter (Rheodyne Inc., Cotati, CA) and a guard column (ODS-Hypersil, Hewlett Packard, Waldbronn, Germany). The mobile phase consisted of acetonitrile-methanol-2% w/v ammonium acetate (40:40:20 v/v), and the ¯ow rate was 1.5 mL/min. In this system, RAG eluted at 4.7 min, and ATRA at 6.7 min. All detection was made at ambient temperature. Acetonitrile (200 mL) and methanol (200 mL) were added to each 100 mL of plasma sample to precipitate proteins. After vortexing for 1 min, and centrifuging for 5 min at 4,000 rpm, 200 mL of the supernatant was injected onto the HPLC system for the assay of RAG and ATRA. The inbetween run coef®cients of variation of RAG and ATRA were all less than 7%. RAG and ATRA were calibrated against reference standards. The standard curves of RAG and ATRA were linear (r2 > 0.99) over the calibration ranges (RAG: 0.1±10 mM, ATRA 0.02±0.15 mM). This HPLC assay method was a slight modi®cation of our previously reported methods.19,20 Preparation of the Dosing Solution The aqueous solubility of RAG is about 0.21 mM (our unpublished data), which is not high enough to achieve the concentration for the preparation of intravenous injection used in this study. Thus, hydroxypropyl-b-cyclodextrin (HPbCD), an excipient that can increase the aqueous solubility of retinoids,19,20 as well as many other hydrophobic drugs, was applied. The RAG solution for intravenous injection was prepared as below: 10 mg RAG (21 mmol) was suspended in 10 mL 0.9% saline containing 20% (w/v) HPbCD. The suspension was then shaken on a horizontal rotary shaker for 1 h at a speed of 200 rpm at ambient temperature. The suspension was ®nally ®ltered through a 0.22-mm disc ®lter (Millipore Co., USA). The ®ltrate was analyzed by HPLC to determine its RAG concentration. This method was similar to the preparation to ATRA±cyclodextrin complex as described in our previous studies.19,20 The RAG

2025

solution for pharmacokinetic study was freshly prepared daily. Animals Adult male Sprague±Dawley rats weighing 240  23 g (mean  SD) were obtained from Laboratory Animal Center of the university. Each rat was housed in a single cage and kept on a 12 h light/dark cycle. Food and water were provided ad libitum. On the day before the study, an indwelling catheter was inserted into the femoral vein under anesthesia. Intravenous doses were given the blood samples were drawn via this catheter. The surgical procedures were conducted in accord with guidelines for the humane use of animals in scienti®c research. A total of four groups of rats were studied. Each group contained 10 rats. Rats in group 1, 2, and 3 received a single intravenous dose of RAG (5, 10, or 15 mmol/kg, respectively), blood samples in all these three groups were collected at 10, 20, 30, and 40 min at 1, 1.5, 2, 3, 4, 5, and 6 h after the intravenous injection, extra blood samples were also collected at 7 h in groups 2 and 3, and 8 h in group 3. Rats in group 4 received daily intravenous doses of 5 mmol/kg for 8 days. After the eighth dose, blood samples were collected through the same time schedule as group 1. 0.3 mL heparin (50 U)saline was used to ¯ush the catheter after each intravenous injection and blood sampling. Blood samples were centrifuged at 4,500 rpm for 5 min; plasma samples were then collected, and kept frozen at ÿ808C until HPLC assay. Pharmacokinetic Analysis All pharmacokinetic parameters were estimated with the software, WinNonlin version 1.0 (Scienti®c Consulting Inc., Lexington, KY). Because a rapid distribution phase of RAG was observed in the ®rst hour after intravenous dosing in all of the rats, pharmacokinetic parameters were generated by ®tting the plasma concentration  time data to the classical two-compartment ®rst-order open model (the model was supplied by WinNonlin version 1.0) using nonlinear least-squares curve ®tting with a weighting factor of 1/Y2 as described previously.20 The AUC of ATRA was calculated by a noncompartmental method. Statistical Analysis Statistical analysis was performed by using SPSS 10.0 (SPSS Inc. Chicago, IL). Data were expressed

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

2026

LIN ET AL.

as mean values  standard deviations (SD). Statistical comparisons between groups 1, 2, and 3 were made by using one-way ANOVA with the post hoc Tukey test. The pharmacokinetic parameters of groups 1 and 4 were compared by twotailed, independent-samples, t-test. A value of p < 0.05 was adopted to indicate statistical signi®cance.

RESULTS The mean plasma RAG concentration±time curve after intravenous administration(s) is shown in Figure 1. Semilogarithmic plots of RAG concentration  time exhibited the characteristic twocompartment ®rst-order elimination kinetics in all of the rats from groups 1±4. The pharmacokinetic parameters of RAG after single or multiple intravenous administration(s) are listed in Table 1. These parameters were obtained by ®tting the plasma RAG concentration data of individual rat to the two-compartment ®rst-order open model. The correlation coef®cients of each individual ®tting were all greater than 0.97, and most of them were greater than 0.99. After single intravenous administration, the AUCs of RAG in groups 1, 2, and 3 (dose: 5, 10,

Figure 1. Pharmacokinetics of RAG after single or multiple intravenous administration(s). Key: (*) 15 mmol/kg, single dose; (~) 10 mmol/kg, single dose; (!) 5 mmol/kg, single dose; (&) 5 mmol/kg, after eight multiple daily administrations. Error bar represents SD.

and 15 mmol/kg, respectively) were 7.29  1.06, 13.8  0.769, and 21.6  4.2 mM
h, respectively. The AUC values were proportional to the doses. After normalizing the AUCs with dose, no statistically signi®cant difference ( p > 0.05) was found among groups 1, 2, and 3. The elimination constants (K10), the elimination half lives (t1/2 (K10)) and the clearance (Cl) were all similar among

Table 1. Pharmacokinetic Parameters of RAG After Intravenous Administration(s) Parametersa ÿ1

Dose (m mol
kg ) Vc(L
kgÿ1) Vp(L
kgÿ1) Vss(L
kgÿ1) K10(hÿ1) K12(hÿ1) K21(hÿ1) AUC (mM
h) t1/2 (K10)(h) a (hÿ1) b (hÿ1) t1/2a(h) t1/2b(h) A (mM) B (mM) Cl(L
hÿ1
kgÿ1) AUMC (mM
h2) MRT (h) AUCATRA(mM
h)

Group 1

Group 2

5 0.347  0.084 0.928  0.232 1.290  0.269 2.09  0.46 1.67  0.54 0.621  0.168 7.29  1.06 0.351  0.082 4.02  1.04 0.315  0.083 0.182  0.043 2.32  0.54 13.8  3.16 1.19  0.31 0.698 0.092 13.7  4.18 1.86  0.403 0.032  0.019

10 0.312  0.059 0.647  0.107b 0.961  0.149b 2.43  0.54 1.38  0.38 0.649  0.115 13.8  0.769b 0.297  0.055 4.07  0.941 0.386  0.061 0.177  0.034 1.83  0.29b 31.0  7.6b 2.36  0.42b 0.729  0.042 18.2  3.6 1.32  0.22b 0.084  0.019b

a

Pharmacokinetic parameters are represented by mean  SD (n ˆ 10). Indicated p < 0.05 between this group and group 1. Indicated p < 0.05 between this group and group 2.

b c

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

Group 3 15 0.298  0.063 0.698  0.165b 0.924  0.187b 2.43  0.27 1.46  0.33 0.698  0.165 21.6  4.2b,c 0.288  0.032 4.18  0.61 0.403  0.070b 0.169  0.024 1.76  0.29b 48.4  10.9b,c 4.12  1.77b,c 0.717  0.132 27.9  6.0b,c 1.29  0.145b 0.125  0.032b,c

Group 4 5  8 days 0.324  0.067 0.494  0.147b 0.818  0.187b 2.16  0.46 0.96  0.48b 0.610  0.195 7.50  1.13 0.336  0.079 3.34  0.99 0.395  0.106 0.227  0.076 1.90  0.62 14.8  2.8 1.14  0.42 0.680  0.101 9.06  2.39b 1.20  0.22b 0.010  0.006b

ALL-TRANS-RETINOYL-b-D-GLUCURONIDE IN SPRAGUE±DAWLEY RATS

groups 1, 2, and 3 (p > 0.05). There was no statistical signi®cant difference in the distribution rate constants from the peripheral compartment to the central compartment (K21) and from the central compartment to the peripheral compartment (K12) in groups 1±3. As the dose increased, the volumes of central compartment (Vc) in groups 1, 2, and 3 were found to be 0.347  0.084, 0.312  0.059, and 0.298  0.063 L/kg, respectively. Although no statistically signi®cant differences existed between these three groups, the Vc tended to decrease with the increased dose. The volumes of distribution at steady state (Vss) and the volumes of the peripheral compartment in groups 1, 2, and 3 were 1.290  0.269, 0.928  0.232; 0.961  0.149, 0.647  0.107; and 0.924  0.187, 0.698  0.165 L/kg, respectively. Statistically signi®cant differences (p < 0.05) were found between group 1 and either groups 2 or 3; while no signi®cant differences (p > 0.05) were observed between groups 2 and 3. After administration of a single dose, similar to the previous ®nding in NMRI mice.21 ATRA was identi®ed as a metabolite of RAG in the ®rst 2 h (Figure 2). However, unlike the study in NMRI mice, ATRA existed at low concentrations (usually < 0.05 mM) and was not detectable in every sample. The AUC values of ATRA in groups 1, 2, and 3 were 0.032  0.019, 0.084  0.019, and 0.125  0.032 mM
h, respectively. ATRA was not detectable in most of the plasma samples at 3 h after intravenous administration. Thus, RAG clearly was not extensively hydrolyzed to ATRA upon its intravenous administration to Sprague± Dawley rats.

2027

After eight daily intravenous administrations of RAG, its pharmacokinetics still followed the two-compartment ®rst order elimination model. The AUC, Cl, K10, t1/2 (K10) in group 4 remained unchanged in comparison to group 1 ( p > 0.05). Thus, long-term daily administration of RAG did not induce accelerated metabolism. K21 remained unchanged ( p > 0.05), implying that long-term treatment did not alter its distribution rate constant from the peripheral compartment to the central compartment. However, the K12, Vp, and Vss declined from 1.67  0.54 hÿ1, 0.928  0.232 L/ kg, and 1.290  0.269 L/kg to 0.96  0.48 hÿ1, 0.494  0.147 L/kg and 0.818  0.187 L/kg, respectively (p < 0.05), indicating multiple daily dosing of RAG changed its distribution pro®le to the peripheral compartment. After the distribution phase, the concentration of RAG in group 4 declined faster than that in group 1. In group 4, 6 h after dosing, RAG was undetectable in plasma in most of the rats while it was detectable in most of the rats in group 1 at that time point. Also, the RAG plasma concentrations in group 1 at 3, 4, and 5 h were higher than those in group 4 ( p < 0.05). These re¯ected to the changes in distribution of RAG. Despite these differences, the AUC of RAG did not change signi®cantly after multiple dosing (7.29  1.06 versus 7.50  1.13 mM
h in groups 1 and 4, respectively). In group 1, ATRA was mainly detected with 2 h after intravenous dosing; while in group 4, ATRA was detected only in the ®rst hour after dosing (Figure 2). The AUC of ATRA after RAG dosing signi®cantly declined from 0.032  0.019 mM
h in group 1 to 0.010  0.006 mM
h in group 4 (p<0.05). The decline in the AUC of ATRA might well re¯ect its accelerated metabolism upon repeated dosing, which has been well demonstrated previously.6,22,23

DISCUSSION

Figure 2. Plasma ATRA kinetics after single or multiple intravenous administration(s) of RAG Key: (*) 15 mmol/kg, single dose; (~) 10 mmol/kg, single dose; (!) 5 mmol/kg, single dose; (&) 5 mmol/kg, eight daily administrations. Error bar represents SD.

The pharmacokinetics of ATRA have been studied extensively in rats, monkeys, and humans.6,20,22 ±24 After the administration of a single intravenous dose, ATRA is eliminated through a saturable process (one-compartment Michaelis±Menten).20,22 ±24 Because of its dosedependent pharmacokinetics, the AUC of ATRA does not increase proportionally with dose.24 However, after 12 daily oral treatments of ATRA in rats, its intravenous pharmacokinetic pro®le

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

2028

LIN ET AL.

changed to a ®rst-order process.23 In monkeys, eight daily intravenous doses of ATRA increased the clearance by about 50%.22 CRABP also increased threefold after only 3 days of ATRA administration,22 and CYP are also induced by ATRA.6,25 These inductions of CRABP, a carrier of ATRA to its metabolizing enzymes, and of CYP in all likelihood markedly contribute to the decline in its plasma concentration.6,22 Thus, the autoinduced accelerated metabolism of ATRA may well cause its pharmacokinetic pro®le to change from Michaelis±Menten to ®rst order, thereby enhancing the decline in drug plasma concentration after multiple dosing and reducing its clinical ef®cacy.22,23 The issue at hand is whether or not RAG follows a pharmacokinetic pro®le similar to that of ATRA. Because the absorption of orally administered RAG is limited in vitamin A-suf®cient rats,10,18,26 pharmacokinetic studies of RAG have only been conducted in the past after intraperitoneal injection into rats,27 subcutaneous21,28 and intravenous injections21 into mice. In the intravenous study in mice,21 only ®ve blood samples were taken in a 6 h period after the intravenous injection of RAG (5 mmol/kg), thereby limiting the pharmacokinetic analysis. Thus, the present study, in which 10 or more time points were collected and ®tted to an appropriate model, is the ®rst detailed pharmacokinetic study of RAG after intravenous administration. In the current study, RAG was found to be eliminated through a kinetic pro®le different from its parent compound, ATRA. The pharmacokinetic characteristics indicated that the elimination of RAG was dose-independent within our tested range (5, 10, and 15 mmol/kg). After a single intravenous dose, RAG was eliminated through a two-compartment ®rst-order process. The Cl, K10, and t1/2 (K10) were found to be dose independent within the tested dosing range. Clinically, the dose-independent elimination could be one of the advantages because it will be easy to maintain the plasma drug level within the therapeutic range. The AUC, Cl, K10, and t1/2 (K10) in group 4 (dose: 8  5 mmol/kg) remained unchanged after an 8-day treatment with RAG (p > 0.05), indicating that its rate of metabolism was not accelerated by multiple dosing. Similar results were obtained in Harland ICR female mice; namely, one singlepoint plasma concentration 2 h after dosing of RAG did not change during 56 days of daily subcutaneous dosing.28 RAG does not bind either to CRABPs17 or to nuclear retinoic acid recep-

tors29 and, in so far as we are aware, does not induce either CRABPs or CYP. Thus, the different kinetic pro®les of RAG and ATRA after multiple dosing may well depend on the distinct properties of these two compounds. Although there were no statistical signi®cant differences in Vc, k12, k21 among groups 1±3; we did ®nd there was a trend of slight decreases in Vc and k12 (Vc was 0.347  0.084, 0.312  0.059, and 0.298  0.063 L/kg; k12 was 1.67  0.54, 1.38  0.38, and 1.46  0.33 hÿ1, respectively from group 1 to 3) and a trend of increase in k21 (the k21 was 0.621  0.168, 0.649  0.115, and 0.698  0.165 hÿ1, respectively from group 1 to 3). Because Vss is equal to:   k12 Vc
1 ‡ k21 the combined effects of the decrease in Vc and k12, and the increase in k21 would result in the signi®cant decrease in Vss, despite the respective differences in Vc, k12, and k21 were statistically insigni®cant. Vp and Vss decreased signi®cantly (p < 0.05) from group 1 to the other two groups (2 and 3), although Vp and Vss in the latter two groups were not signi®cantly different. One possible explanation for the decline in Vp and Vss is that the binding of RAG in peripheral tissue might be saturable at higher doses. Thus, once these tissue binding sites for RAG are saturated, RAG remains in the circulation and Vp and Vss decreases. The binding of warfarin to tissue seems to follow a similar course.30 King et al. found that as the dose of warfarin increased from 2 to 5 mg, the apparent volume of distribution (V/F) decreased from 21 to 12 L.30 Although the Vp and Vss of RAG decreased with an increase in dose, the extent to change in its Vp and Vss was not as large as that found in warfarin. After eight daily doses of RAG, K12 signi®cantly decreased by 42% from its value in group 1, while the Vp and Vss decreased by about 47 and 37%, respectively. These could also be explained by the decrease in tissue binding at the peripheral compartment. However, further study on the tissue distribution of RAG after single and multiple dosings are required to verify this postulation. In this study, RAG was not extensively hydrolyzed to RA. The enzyme presumably responsible for RAG hydrolysis is b-glucuronidase, which is primarily localized in the lysosomes of cells.9

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

ALL-TRANS-RETINOYL-b-D-GLUCURONIDE IN SPRAGUE±DAWLEY RATS

Thus, the water-soluble RAG may well not be accessible to the enzyme. In most other in vivo and in vitro studies, the rate of hydrolysis of RAG to ATRA was also slow.9,16,18,28 In NMRI mice, however, RAG is hydrolyzed to ATRA much more rapidly than in other rodent model.21 Indeed, the AUC of ATRA after a single intravenous dose of RAG (5 mmol/kg) to NMRI mice was even higher than that after a single intravenous dose of ATRA (5 mmol/kg).21 In our study, ATRA was only detected in the ®rst 2 h and only at very low concentrations (< 0.05 mM). In contrast, the range of Cmax for RAG was 15±52.5 mM. After eight daily administrations of RAG, the AUC of ATRA in group 4 was only one-third of that in group 1 (Table 1). The uptake of ATRA by tissues and/or its metabolism thus seem to be enhanced by prior treatment with RAG, even though the resultant plasma ATRA concentrations were low. In contrast, the daily subcutaneous injection of RAG (30 mmol/kg) to Harland ICR female mice for 12 days did not in¯uence the plasma ATRA concentrations at 2 h.28 As the NMRI mouse, species differences may exist between the Sprague-Dawley rat and the Harland ICR mouse. In addition, it would be dif®cult for them to judge whether the elimination of ATRA was changed during multiple dosings through the comparison of one single point of drug plasma concentration. There are always demands to ®nd new retinoids with higher therapeutic ef®cacy and lower toxicity. 9-cis-Retinoic acid (9-cis-RA), an isomer of ATRA, has been proposed to be a substitute of ATRA, because it was found in some previous in vitro studies to be more potent than ATRA in inducing myeloid differentiation and also the apoptosis of ATRA resistant APL cell line, UF1.31±34 However, clinical studies indicated that although, 9-cis-RA could induce complete remission, it still failed to reverse clinically acquired retinoid resistance in acute promyelocytic leukemia.35 9-cis-RA also has the side effects for example, ``retinoic acid syndrome,'' similar to ATRA.36 Furthermore, daily oral administration of 9-cis-RA induced its accelerated metabolism and the plasma drug concentration declined substantially.35,37 The mechanism of the differentiation induction of RAG is still not clear. It has been suggested that RAG might work per se or as a transport agent of ATRA from the cytosol to the nucleus.9 Although its activity might be attributed to ATRA, RAG has its own advantages as a prodrug of ATRA. The

2029

metabolism pathway of ATRA has been well studied. As illustrated in a previously published ®gure,2 the oxidation of ATRA catalyzed by CYP 450 enzymes mainly takes place in smooth endoplasmic reticulum. This process is facilitated by cellular retinoic acid binding protein (CRABPs). Because CYP 450 and CRABP and both inducible and become much more abundant after treatment with ATRA, most of the ATRA in the cytoplasm will be bonded by CRABP and transported to CYP 450. Thus, only a very little amount of ATRA can enter the nucleus and take action on cell differentiation. This may be one of the major mechanisms of clinical ATRA resistance. However, RAG lacks the binding af®nity to CRABP17,29 and its transportation to CYP 450 is avoided. Furthermore, in human leukemic cells, HL 60, RAG was found to be taken up well and is localized mainly in the nuclei.9 Thus, it may work well as slow release precursor of ATRA in the nuclei.9 Gallup et al. showed that the 50% effective doses (ED50) of RAG relative to the differentiation, growth, and cytotoxicity toward HL 60 cells were 1.4, 3.5, and > 10 mM, respectively, compared with 0.8, 1.0, and 1.6 mM for ATRA.9,15 Although the differentiation and growth inhibition effect of RAG is two to three times weaker than ATRA, its cytotoxicity is about 10-fold weaker than ATRA. As it has been well documented that the therapeutic effect of ATRA is attributed to its differentiation effect, but not the cytotoxicity, effect, the decreased cytotoxicity of RAG would be an advantage causing reduced side effects. In the present study, the clearance of RAG was found to be affected by daily treatments. Thus, RAG might be advantageously used in place of other retinoids in the treatment of leukemia and other retinoid-responsive clinical conditions. It warrants con®rming the therapeutic applications of RAG in further studies.

ACKNOWLEDGMENTS The authors thank Ms. Q. Y. Zhou, Mr. W. J. Sam, and Mr. Y. Chen for their help during the experiment and the preparation of manuscript. The present study was supported by the Academic Research Fund R148-000-021-112 offered by the National University of Singapore. The ®rst author would like to thank the National University of Singapore for the offer of a Research Scholarship. The synthesis of RAG was supported by NIHDK39733 to J.A. Olson.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

2030

LIN ET AL.

REFERENCES 1. Chen Z, Wang ZY, Chen SJ. 1997. Acute promyelocytic leukemia: Cellular and molecular basis of differentiation and apoptosis. Pharmacol Ther 76:141±149. 2. Warrell RP, Dethe H, Wang ZY, Degos L. 1993. Acute promyelocytic leukemia. N Engl J Med 329:177±189. 3. Gillis JC, Goa KL. 1995. Tretinoin-A review of its pharmacodynamic and pharmacokinetic properties and use in the management of acute promyelocytic leukaemia. Drugs 50:897±923. 4. Lin RJ, Egan DA, Evans RM. 1999. Molecular genetics of acute promyelocytic leukemia. Trends Genet 15:179±184. 5. Tallman MS, Andersen JW, Schiffer CA, Appelbaum FR, Feusner JH, Ogden A, Shepherd L, Willman C, Bloom®eld CD, Rowe JM, Wiernik PH. 1997. All-trans-retinoic acid in acute promyleocytic leukemia. N Engl J Med 337:1021±1028. 6. Regazzi MB, Iacona I, Gervasutti C, Lazzarino M, Toma S. 1997. Clinical pharmacokinetics of tretinoin. Clin Pharmacokinet 32:382±402. 7. Adamson PC, Pitot HC, Balis FM, Rubin J, Murphy RF, Poplack DG. 1993. Variability in the oral bioavailability of all-trans-retinoic acid. J Natl Cancer Inst 85:993±996. 8. Lazzarino M, Regazzi MB, Corso A. 1996. Clinical relevance of all-trans retinoic acid pharmacokinetics and its modulation in acute promyelocytic leukemia. Leuk Lymphoma 23:539±543. 9. Formelli F, Barua AB, Olson JA. 1996. Bioactivities of N-(4-hydroxyphenyl) retinamide and retinoyl beta-glucuronide. FASEB J 10:1014±1024. 10. Barua AB. 1997. Retinoyl beta-glucuronide: A biologically active form of vitamin A. Nutr Rev 55:259±267. 11. Barua AB, Olson JA. 1986. Retinoyl beta-glucuronide: An endogenous compound of human blood. Am J Clin Nutr 43:481±485. 12. Gunning DB, Barua AB, Lloyd RA, Olson JA. 1994. Retinoyl b-glucuronide: A nontoxic retinoid for topical treatment of acne. J Dermatol Treat 5:181±185. 13. Goswami BC, Baishya B, Barua AB, Olson JA. 1999. Topical retinoyl beta-glucuronide is an effective treatment of mild to moderate acne vulgaris in Asian Indian patients. Skin Pharmacol Appl 12:167±173. 14. Zile MH, Cullum ME, Simpson RU, Barua AB, Swartz DA. 1987. Induction of differentiation of human promyelocytic leukemia cell line HL-60 by retinoyl glucuronide, a biologically active metabolite of vitamin A. Proc Natl Acad Sci USA 84:2208± 2212. 15. Gallup JM, Barua AB, Furr HC, Olson JA. 1987. Effects of retinoid beta-glucuronides and N-retinoyl

16.

17. 18.

19.

20.

21.

22.

23.

24.

25.

26.

27. 28.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001

amines on the differentiation of HL-60 cells in vitro. Proc Soc Exp Biol Med 186:269±274. Janick±Buckner D, Barua AB, Olson JA. 1991. Induction of HL-60 cell differentiation by watersoluble and nitrogen-containing conjugates of retinoic acid and retinol. FASEB J 5:320±325. Mehta RG, Barua AB, Olson JA, Moon RC. 1992. Retinoid glucuronides do not interact with retinoid binding proteins. Int J Vitam Nutr Res 62:143±147. Becker B, Barua AB, Olson JA. 1996. All-transretinoyl beta-glucuronide: New procedure for chemical synthesis and its metabolism in vitamin Ade®cient rats. Biochem J 314:249±252. Lin HS, Chean CS, Ng YY, Chan SY, Ho PC. 2000. 2-Hydroxypropyl-beta-cyclodextrin increases aqueous solubility and photostability of all-transretinoic acid. J Clin Pharm Ther 25:265±269. Lin HS, Chan SY, Low KSY, Shoon ML, Ho PC. 2000. Kinetic study of a 2-hydroxypropyl-b-cyclodextrin-based formulation of all-trans-retinoic acid in Sprague-Dawley rats after oral or intravenous administration. J Pharm Sci 89:260±267. Nau H, Elmazar MMA, Ruhl R, Thiel R, Sass JO. 1996. All-trans-retinoyl-beta-glucuronide is a potent teratogen in the mouse because of extensive metabolism to all-trans-retinoic acid. Teratology 54:150±156. Adamson PC, Boylan JF, Balis FM, Murphy RF, Godwin KA, Gudas LJ, Poplack DG. 1993. Time course of induction of metabolism of all-transretinoic acid and the up-regulation of cellular retinoic acid-binding protein. Cancer Res 53:472± 476. El Mansouri S, Tod M, Leclerq O, Petitjean G, Perret G, Porthault M. 1995. Time- and dosedependent kinetics of all-trans-retinoic acid in rats after oral or intravenous administration(s). Drug Metab Dispos 23:227±231. Adamson PC, Balis FM, Smith MA, Murphy RE, Godwin KA, Poplack DG. 1992. Dose-dependent pharmacokinetics of all-trans-retinoic acid. J Natl Cancer Inst 84:1332±1335. Sonneveld E, van den Brink CE, van der Leede BJM, Schulkes PKAM, Petkovich M, van der Burg B, van der Saag PT. 1998. Human retinoic acid (RA) 4-hydroxylase (CYP26) is highly speci®c for alltrans-RA and can be induced through RA receptors in human breast and colon carcinoma cells. Cell Growth Differ 9:629±637. Barua AB, Duitsman PK, Olson JA. 1998. The role of vitamin A status in the conversion of all-trans retinoyl beta-glucuronide to retinoic acid in male Sprague±Dawley rats. J Nutr Biochem 9:8±16. Barua AB, Olson JA. 1989. Chemical synthesis of all-trans [11-3H] retinoyl b-glucuronide and its metabolism in rats in vivo. Biochem J 263:403±409. Sidell N, Sawatsri S, Connor MJ, Barua AB, Olson JA, Wada RK. 2000. Pharmacokinetics of chroni-

ALL-TRANS-RETINOYL-b-D-GLUCURONIDE IN SPRAGUE±DAWLEY RATS

29.

30.

31. 32.

33.

cally administered all-trans-retinoyl-b-glucuronide in mice. Biochem Biophys Acta 1502:264±272. Sani BP, Barua AB, Hill DL, Shih TW, Olson JA. 1992. Retinoyl b-glucouronide: Lack of binding to receptor proteins of retinoic acid as related to biological activity. Biochem Pharmacol 43:919± 922. King SY, Joslin MA, Raudibaugh K, Pieniaszek HJ Jr., Benedek ZH. 1995. Dose-dependent pharmacokinetics of warfarin in healthy volunteers. Pharm Res 12:1874±1877. Bollag W, Majewski S, Jablonska S. 1994. Cancer combination chemotherapy with retinoidsÐ Experimental rationale. Leukemia 8:1453±1457. Kizaki M, Ikeda Y, Tanosaki R, Nakajima H, Morikawa M, Sakashita A, Koef¯er HP. 1993. Effects of novel retinoic acid compound, 9-cisretinoic acid, on proliferation, differentiation, and expression of retinoic acid receptor-alpha and retinoid-x receptor-alpha RNA by hl-60 cells. Blood 82:3592±3599. Sakashita A, Kizaki M, Pakkala S, Schiller G, Tsuruoka N, Tomosaki R, Cameron JF, Dawson MI, Koef¯er HP. 1993. 9-cis-Retinoic acidÐEffects

34.

35.

36.

37.

2031

on normal and leukemic hematopoiesis in vitro. Blood 81:1009±1016. Sato A, Imaizumi M, Kizaki M, Saito T, Yoshinari M, Suzuki H, Iinuma K. 1998. 9-cis Retinoic acid-induced differentiation and apoptosis of acute promyelocytic leukemia cell line, UF-1, with resistance of all-trans retinoic acid. Exp Hematol 26:727±727. Miller WH, Jakubowski A, Tong WP, Miller VA, Rigas JR, Benedetti F, Gill GM, Truglia JA, Ulm E, Shirley M, Warrell RP. 1995. 9-cis Retinoic acid induces complete remission but does not reverse clinically acquired retinoid resistance in acute promyelocytic leukemia. Blood 85:3021±3027. Soignet SL, Benedetti F, Fleischauer A, Parker BA, Truglia JA, Crisp MR, Warrell RP. 1998. Clinical study of 9-cis retinoic acid (LGD1057) in acute promyelocytic leukemia. Leukemia 12:1518±1521. Wadler S, Schwartz EL, Anderson P, Runowicz CD, Chuang L, Del Priore G, Hochster H, Goldberg G, Fields A, Loewen G, Haynes H. 1999. Preliminary phase II clinical and pharmacokinetic study of 9-cis retinoic acid in advanced cervical cancer. Cancer J Sci Am 5:165±170.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 90, NO. 12, DECEMBER 2001