pharmacokinetics of dietary cancer chemopreventive compound dibenzoylmethane in rats and the impact...

BIOPHARMACEUTICS & DRUG DISPOSITIONBiopharm. Drug Dispos. 32: 65–75 (2011)

Published online 16 December 2010 in Wiley Online Library

(wileyonlinelibrary.com) DOI: 10.1002/bdd.734

Pharmacokinetics of Dietary Cancer ChemopreventiveCompound Dibenzoylmethane in Rats and the Impact ofNanoemulsion and Genetic knockout of Nrf2 on itsDisposition

Wen Lina,y, Jin-Liern Honga, Guoxiang Shena, Rachel T. Wua, Yuwen Wangb, Mou-Tuan Huangc,Harold L. Newmarkc, Qingrong Huangb, Tin Oo Khora, Tycho Heimbachd, and Ah-Ng Konga,�,y

aDepartment of Pharmaceutics, Ernest-Mario School of Pharmacy, Rutgers, The State University of New Jersey, 160 Frelinghuysen Road,

Piscataway, NJ 08854, USAbDepartment of Food Science, School of Environmental and Biological Sciences, Rutgers, The State University of New Jersey, 65 Dudley Road,

New Brunswick, NJ 08901, USAcDepartment of Chemical Biology, Ernest Mario School of Pharmacy, Rutgers, The State University of New Jersey, 65 Dudley Road, New

Brunswick, NJ 08901, USAdDMPK - Translational Sciences (TS), Novartis Institutes for BioMedical Research, One Health Plaza, East Hanover, NJ 07936, USA

ABSTRACT: The pharmacokinetic disposition of a dietary cancer chemopreventive compounddibenzoylmethane (DBM) was studied in male Sprague-Dawley rats after intravenous (i.v.) and oral(p.o.) administrations. Following a single i.v. bolus dose, the mean plasma clearance (CL) of DBMwas low compared with the hepatic blood flow. DBM displayed a high volume of distribution (Vss).The elimination terminal t1/2 was long. The mean CL, Vss and AUC0�N/dose were similar betweenthe i.v. 10 and 10 mg/kg doses. After single oral doses (10, 50 and 250 mg/kg), the absolute oralbioavailability (F*) of DBM was 7.4%–13.6%. The increase in AUC was not proportional to the oraldoses, suggesting non-linearity. In silico prediction of oral absorption also demonstrated low DBMabsorption in vivo. An oil-in-water nanoemulsion containing DBM was formulated to potentiallyovercome the low F* due to poor water solubility of DBM, with enhanced oral absorption. Finally, toexamine the role of Nrf2 on the pharmacokinetics of DBM, since DBM activates the Nrf2-dependentdetoxification pathways, Nrf2 wild-type (1/1) mice and Nrf2 knockout (�/�) mice were utilized.There was an increased systemic plasma exposure of DBM in Nrf2 (�/�) mice, suggesting that theNrf2 genotype could also play a role in the pharmacokinetic disposition of DBM. Taken together,the results show that DBM has low oral bioavailability which could be due in part to poor watersolubility and this could be overcome by a nanotechnology-based drug delivery system andfurthermore the Nrf2 genotype could also play a role in the pharmacokinetics of DBM. Copyright r2010 John Wiley & Sons, Ltd.

Key words: dibenzoylmethane; nanotechnology-based drug delivery system; Nrf2 genotype

Introduction

Dibenzoylmethane (DBM, 1,3-diphenyl-propane-dion), as shown in Figure 1, is a natural phyto-chemical found as a minor constituent in the rootof licorice [1]. It is a beta-diketone compoundyThese two authors contributed equally to this work.

*Correspondence to: Ah-Ng Kong, Department of Pharma-ceutics, Ernest-Mario School of Pharmacy, Rutgers, The StateUniversity of New Jersey, 160 Frelinghuysen Road, Piscat-away, NJ 08854, USA.

E-mail: [email protected]

Received 27 June 2010Revised 19 August 2010

Accepted 26 October 2010Copyright r 2010 John Wiley & Sons, Ltd.

with a variety of anti-cancer activities. DBM hasbeen reported to suppress the 7,12-dimethylbenz[a]anthracene (DMBA)-induced and estradiol (E2)-induced mammary tumorgenesis both in vivoand in vitro [2–6]. The potential mechanisms ofpreventing DMBA-induced mammary tumor byDBM include its inhibition of DMBA metabolismand the formation of DMBA–DNA adducts, aswell as facilitating the excretion of DMBA-relatedtoxic metabolites by inducing Phase II detoxifica-tion enzymes [3,4]. DBM also has been shown toinhibit E2-induced cell proliferation in humanbreast cancer cells and mouse mammary glandsthrough the E2–ER–ERE dependent pathway [2].Moreover, DBM has also been used as a sunscreento prevent the injurious effects of overexposure tosolar radiation. It has anti-carcinogenic effects inTPA-induced skin inflammation/tumor promo-tion in a mouse model [5]. Induction of Phase I/Phase II enzymes was considered as one majormechanism of how DBM could block/retard thecarcinogenesis process. In a rat model, both Phase IIenzymes including glutathione transferases andNAD(P)H: quinine reductase were induced by theconsumption of 1% DBM diet [6]. In addition,DBM was also able to modulate Phase I enzymescytochromes P450 (CYP) 1A1, 1A2 and 1B1 [7,8].Dibenzoylmethane has also been found to inhibitcell growth and to induce cell cycle arrest inhuman prostate cancer cells [9] and human coloncancer cells [10]. A recent investigation in ourlaboratory revealed that DBM could significantlyinhibit intestinal and colon tumorigenesis inApcMin/1 mice [11,12].

Although DBM has been recognized as a verypromising cancer chemopreventive agent, to ourknowledge, the pharmacokinetics of DBM hasnot been reported. Recently, an HPLC-UV analy-tical method to quantify the concentrations ofDBM in rat plasma was developed and validated[13]. Using this method, the in vivo pharmacoki-netics of DBM were evaluated following i.v. andoral administrations in rats. In order to enhancethe drug absorption across the gastrointestinal

tract following oral administration, a nanoparticleemulsion containing DBM was developed andhigher bioavailability (F�) value was obtainedwith this nanoemulsion.

A recent study has found that DBM activatesNrf2-dependent signaling pathways [14]. Nrf2belongs to the CNC (Cap-N-Collar) family oftranscription factors and possesses a highly con-served basic region-leucine zipper (bZip) struc-ture. Its important role in the regulation of theexpression of many mammalian drug metabo-lizing/detoxifying, transporters (e.g. multidrugresistance-associated proteins, MRP and organicanion-transporting polypeptide, OATPs) andantioxidant enzymes under oxidative or electro-philic stress has been verified in differentNrf2-deficient mice experiments, in which theexpression of these enzymes are dramaticallyattenuated and the Nrf2 knockout (�/�) miceare more susceptible to carcinogen-inducedcarcinogenesis (reviewed in [15–18]). Our presentstudy shows that Nrf2 has a substantial impacton the in vivo disposition of DBM in mice.

Materials and Methods

Chemicals

1,3-Diphenyl-1,3-propanedione (dibenzoylmethane,DBM) and internal standard (I.S.) 1-(5-chloro-2-hydroxy-4-methylphenyl)-3-phenyl-1,3-propane-dione (CHMPP), uridine 5-diphosphoglucuronicacid trisodium salt (UDPGA), nicotinamideadenine dinuclotide phosphate (NADP), glucose6-phosphate, glucose-6-phosphate dehydrogen-ase, MgCl2 were purchased from Sigma-Aldrich(St Louis) at a purity of more than 98%.Acetonitrile and methanol were all HPLC gradefrom Fisher Scientific. Trifluoroacetic acid (TFA)was spectrophotometric grade (499%, Aldrich,WI, USA). Ethyl acetate was purchased fromSigma with a purity of 99.9%. Other chemicalsused in this study were all of analytical gradeunless specified.

Intravenous and oral administrations of DBM torats

Male Sprague-Dawley rats weighing 250–300 gwith implanted jugular vein cannulae were

O O

Figure 1. Chemical structure of dibenzoylmethane

W. LIN ET AL.66

Copyright r 2010 John Wiley & Sons, Ltd. Biopharm. Drug Dispos. 32: 65–75 (2011)DOI: 10.1002/bdd

purchased from Hilltop Laboratory Animals Inc.(Scottdale, PA USA). The rats were housed at theAnimal Care Facility of Rutgers Universityunder 12 h light–dark cycles with free access tofood and water. Upon arrival, the rats weregiven AIN-76A diet (Research Diets, NJ, USA)free of antioxidant and acclimatized for 3 days.The rats (n 5 3 or 4) were fasted overnight andgiven DBM at doses of 10, 50 and 250 mg/kgin a vehicle of Cremophor EL/Tween-80/ethanol/water (2:1:1:6, v/v/v/v) or nanoemul-sion (250 mg/kg only) by oral gavages (p.o.).A separate group of rats (n 5 3 or 4) was alsogiven DBM at doses of 10 and 20 mg/kg in thesame vehicle as the intravenous (i.v.) routethrough the jugular vein cannulae. Bloodsamples (200 ml) were collected at 2 (i.v. only),7.5, 15, 30 min, 1, 1.5, 2, 4, 6, 8, 12, 24, and 36 hpost-dose. Plasma was separated immediately bycentrifugation and stored at �801C until analysis.

Blood cell partitioning

The blood to plasma partitioning ratio (CB/CP)for DBM was performed in fresh, pooled ratblood obtained from Bioreclamation Inc (Hicks-ville, NY). DBM (1000 mg/ml) in methanol wasspiked into blood at a ratio of 1:100 to give afinal concentration of 10 mg/ml. Samples wereincubated at 371C in a shaking water bath for 2 h.Plasma was separated immediately by centrifu-gation and analysed by HPLC-UV detectionmethod. This experiment was carried out intriplicate.

Intra-peritoneal and oral administrations ofDBM to mice

Age-matched (10–15 weeks old) male Nrf2 wildtype C57BL/6J (1/1) mice and Nrf2 knockoutC57BL/6J (�/�) mice were used. Mice werebred at the Nelson Animal Facility (wild type)and the Animal Facility of Laboratory for CancerResearch (Nrf2 knockout) at Rutgers University.After weaning, the mice were housed at RutgersUniversity Animal Facility at Gordon Roadunder 12 h light/dark cycles with free access tofood and water. The mice were given AIN-76Adiet 3 days before the experiments. The studywas designed by a composite blood samplingapproach. For oral administration, the mice were

fasted overnight and given DBM at a dose of500 mg/kg in a vehicle of Cremophor EL/Tween-80/ethanol/water (2:1:1:6, v/v/v/v) by oralgavages. Mice were also given DBM at dose of100 mg/kg in the same vehicle by intraperitoneal(i.p.) injection. Blood samples were collected attime points of 2 (i.p. only), 7.5, 15, 30 min, 1, 1.5,2, 4, 6, 8, 12, 24 and 36 h post-dose by heartpuncture. Three mice were used for each timepoint in the pharmacokinetic studies. Plasma wasseparated immediately by centrifugation andstored at �801C until analysis.

Preparation of DBM in nanoparticle vehicle

The components of the emulsion were DBM/Tween 20/glycerol monooleate medium-chaintriglycerides (MCT)/water (2:3:3:15:77, v/v/v/v).DBM was first dissolved in the MCT. The aqueousphase was mixed to ensure complete dissolutionof all additives. Subsequently, the MCT and thewater mixture were mixed by stirring for�1 min.After mixing, the crude emulsion was homo-genized by high speed homogenizer (24000 rpm)for �1 min. Finally, the emulsion was passedthrough a high pressure homogenizer (1500 bar)for 10 cycles (70 nm). The hydrodynamic oildroplet diameter in DBM-containing nanoemul-sion was measured using a light scatteringmethod with ZetaPALS system (Holtville, NY).The average particle size was found to be100–110 nm.

HPLC analysis of DBM

Rat plasma concentrations of DBM were deter-mined by a validated HPLC-UV detectionmethod [13]. Briefly, a 50 ml aliquot of the plasmasamples was spiked with the internal standard(CHMPP) working solution and extracted twicewith 200 ml ethyl acetate/methanol (95:5, v/v)solution. The organic layer was separated bycentrifugation and evaporated to dryness undernitrogen gas at room temperature. The residuewas reconstituted in 100 ml of acetonitrile/water(50:50, v/v) solution, and 20 ml was injected ontoa HPLC column. Analyses were performedusing a Shimadzu HPLC system at 41C with areverse phase column (GeminiTM C18 column,150� 2.0 mm, 5 mm, Phenomenex, Torrance, CAUSA) protected with a SecurityGuardTM

PHARMACOKINETICS OF DIBENZOYLMETHANE 67


cartridge system (Phenomenex) and a 0.45 mmin-line filter. The binary gradient mobile phase(mobile phase A: water/methanol (80:20, v/v)with 0.1% TFA and mobile phase B: acetonitrilewith 0.1% TFA) were pumped at a flow rate of0.2 ml/min. The UV detector was set at a singlewavelength of 335 nm. The Class-VP softwareversion 7.1.1 (Shimadzu, MD USA) was usedfor instrument control and data analysis. Theretention time for DBM and internal standard(CHMPP) were 21.4 and 24.0 min, respectively.The lower limit of quantification for DBMwas 0.05 mg/ml and the limit of detection was10 ng/ml. The linear calibration curves wereobtained in the concentration range of 0.05–20 mg/ml. The intra- and inter-day precisionand accuracy determination of quality controlsamples were below 15%.

Pharmacokinetic analysis

The DBM plasma concentration data wereanalysed by a non-compartmental analysis usingWinNonlin 4.0 software (Pharsight, CA USA) toobtain the pharmacokinetic parameters. The peakplasma concentrations (Cmax) and times theyoccurred (Tmax) were recorded. The area underthe plasma concentration versus time curve(AUC0-t) from time zero to the time of lastmeasured concentration (Clast) was determinedusing the log-linear trapezoidal rule. The AUCzero to infinite (AUC0�N) was obtained by theaddition of AUC0-t and the extrapolated areacalculated by Clast/kel. The elimination rateconstant (kel) was estimated from the slope ofterminal phase and the half-life (t1/2) wascalculated using the equation t1/2 5 ln2/kel. Themean residence time (MRT) was calculated asAUMC0�N/AUC0�N, where AUMC0�N is thearea under the first moment versus time curve.Clearance (CL) was estimated by Dose/AUC0�N

from i.v. administration. The volume of distribu-tion at steady state (Vdss) was calculated as CL �MRTIV. The maximum plasma concentration(Cmax) and the time to reach the Cmax wereobtained directly from the plasma concentra-tion–time profile. The absorption rate (ka) wascalculated using Gastroplus software (Simula-tionplus, Lancaster, CA) by simultaneous fittingwith a three-compartment model. The absolute

bioavailability (F�) was determined by the ratio ofthe dose-normalized AUC0�N following oral andintravenous administration. Deconvolution ana-lysis was performed in conjunction with a three-compartmental analysis of DBM data obtainedfrom with the i.v. administration at 10 mg/kg.

In silico prediction of physico-chemical propertiesof DBM

The prediction of the physico-chemical proper-ties of DBM was performed in GastroPlusTM

software, version 6.0 (Simulationplus, Lancaster,CA). The chemical structure was compiled withISIS Draw software, version 2.5 (MDL Informa-tion System, Los Altos, CA).

Preparation of liver microsomes

Pooled Nrf2 wild type (1/1) or Nrf2 knockout(�/�) C57BL/6J mice livers (3 g) were thawed in12 ml potassium phosphate buffer (0.1 M, pH 7.4,0.125 M KCl, 1.0 mM EDTA and 0.25 M sucrose),cut into small pieces with a blade and homo-genized with a rotor homogenizer (10 strikes) onice. Homogenates were then centrifuged at 41Cfor 20 min using a Soval SS34 rotor (9000 rpm,about 10000� g), the supernatant was pouredinto a new ultra-centrifugation tube and ultra-centrifugation was performed at 43000 rpm(100000� g) using the Ti90 rotor on a Beckmanultracentrifugation machine for 90 min at 41C.The supernatant was stored at �801C as thecytosol fraction. The pellet was washed withwashing buffer (0.1 M Tris-HCl, pH 7.4, 0.125 M

KCl and 1 mM EDTA) once and ultracentrifugedfor 60 min with the same conditions describedabove. The resulting pellet (microsome) wasresuspended in 1 ml of stocking buffer (0.25 M

sucrose, 0.1 mM ETDA), and the microsomeswere stored at �801C after protein concen-tration determination, until further analyses asdescribed below.

In vitro liver microsomes stability assay

The stability of DBM was examined in pooledmouse microsomes in the presence of NADPHand UPDGA. Mouse liver microsomes (1 mgprotein/ml) in 100 mM potassium phosphatebuffer (pH 7.4) were preincubated with DBM(50 mM, final concentration) for 5 min in a water

W. LIN ET AL.68


bath with a shaking speed of 100 rpm at 371C.The reactions were initiated by the additionof an NADPH-generating system (at a finalconcentration of 1 mM NADP1, 10 mM glucose6-phosphate, 2 U/ml glucose-6-phosphate dehy-drogenase, 10 mM magnesium chloride), withUDPGA (4 mM), with the final volume of thereaction mixture of 500 ml. Reactions were carriedout for 1 h. A 50 ml of reaction mixture wassampled at different time points (0, 2, 5, 10, 20, 30and 45 min) during the incubation and mixedwith 200 ml ethyl acetate/methanol mixture (95:5,v/v) twice to extract DBM. After centrifugation,the pooled supernatants were dried under astream of nitrogen gas and reconstituted in 100 mlof acetonitrile/water (50:50, v/v) before subject-ing to HPCL/UV analysis as described earlier.

Statistical analysis

Statistical analysis was conducted using a one-way ANOVA followed by Bonferroni’s test orStudent’s t-test using SigmaPlot, V11. Statisticalsignificance was set at po0.05.

Results and Discussion

The dietary compound DBM is a minor con-stituent in licorice and possesses a potentanti-carcinogenesis effect. However, many basicproperties of this compound, such as thepharmacokinetics of DBM, are still unknown.In the present study, the dose-dependent plasmasystemic exposure and oral bioavailability (F�) ofDBM were investigated in a rat model at a doserange of 10–250 mg/kg. As the predicted solubi-lity of DBM is very low (22 mg/ml), the bioavail-ability of DBM was also tested in the nanoparticleemulsion.

The mean plasma concentration–time profilesof DBM after an i.v. administration of 10 and20 mg/kg in rats are shown in Figure 2A. Thepharmacokinetic parameters determined by anon-compartmental analysis are summarized inTable 1. Following a single i.v. administration at adose of 10 and 20 mg/kg, the AUC0�N valueswere 16.577.85 and 27.972.40 mg �h/ml, respec-tively. The dose-normalized plasma concentra-tion–time profiles, presented in Figure 2B, weresuperimposable. In addition, the mean values of

dose-normalized AUC0�N were 1.65 and 1.40(mg h/ml)/(mg/kg) at doses of 10 and 20 mg/kg,which is not significantly different between 10

(A)

(B)

Figure 2. (A) The mean plasma concentration–time profiles ofdibenzoylmethane after intravenous administration. Ratswere dosed intravenously with 10 or 20 mg/kg of DBM. Dataare expressed as mean 7 SD, n 5 4. (B) The dose-normalizedmean plasma concentration–time profiles of DBM in rats afterintravenous administration. Rats were dosed with 10 or20 mg/kg of DBM

Table 1. Pharmacokinetics characteristics of dibenzoyl-methane after intravenous (i.v.) administration

Parameter Dose (mg/kg)

10 20

Number of animals 4 4CL, l/h/kga 0.7170.32 0.7270.07VSS, l/kga 6.7571.85 7.3574.26AUC0�N, mg h/ml 16.577.85 27.972.40AUC0�N /Dose, (mg h/ml)/(mg/kg)a 1.6570.79 1.4070.12Terminal t1/2, ha 14.475.00 12.075.34MRT, ha 10.473.33 10.276.10

aNo significant difference from corresponding parameters between

two groups (p40.05).



and 20 mg/kg (p 4 0.05). These results suggestthat DBM follows linear pharmacokinetics withinthe dose range tested in male rats. Following asingle i.v. administration at doses of 10 and20 mg/kg, DBM was eliminated with a systemicplasma clearance of 0.7170.32 and 0.7270.07 l/h/kg, respectively. The terminal t1/2 was14.475.00 and 12.075.34 h, respectively. Theplasma clearance and half-life of DBM remainedunchanged with increased doses, again suggest-ing linear pharmacokinetics of DBM in rats afteran i.v. administration. The steady state volume ofdistribution (Vdss) of DBM was 6.7571.85 and7.3574.26 l/kg at doses of 10 and 20 mg/kg,respectively. The values are greater than totalbody water of 0.67 l/kg, suggesting that DBMmay be widely distributed into tissues. Theextensive tissue distribution of DBM may resultfrom its structural properties of lipophilicity(log P 2.7). The mean residence time (MRT) valueof DBM after the i.v. 10 mg/kg was calculated tobe 10.473.33 h, which was similar to that of the20 mg/kg dose (10.276.10 h). Overall, DBMshows very similar pharmacokinetic patternsbetween the 10 and 20 mg/kg due to its linearpharmacokinetics Figure 3.

The blood-to-plasma concentration ratio(CB/CP) of DBM in the rats was 0.9370.19, aftera 2 h incubation at 371C. Dibenzoylmethaneappears to be equally distributed betweenplasma and blood cells in the rats. The systemicblood clearance of DBM was calculated to beabout 12.7 ml/min/kg. Therefore, the hepatic

extraction ratio (EH) was estimated in the rangeof 0.16–0.25 using literature values of hepaticblood flow in the rat (QH, about 55–80 ml/min/kg) [19,20], suggesting DBM is a low-clearancecompound with a poor extraction through theliver in Sprague-Dawley rats.

The mean plasma concentration–time profilesof DBM after single oral administrations of 10, 50and 250 mg/kg in rats are shown on Figure 4.The non-compartmental analysis was performedto obtain the basic pharmacokinetic parametersas listed in Table 2. Following an oral adminis-tration of 10, 50 and 250 mg/kg, DBM reachedCmax of 0.3470.04, 1.5070.41 and 2.8870.74mg/mlat Tmax of 2.0071.35, 2.0070.00 and 3.3371.15 h,respectively. The AUC0�N values were 2.2570.33,

Figure 3. The mean plasma concentration–time profiles ofdibenzoylmethane after oral administration. Rats were dosedby oral gavage with 10, 50 or 250 mg/kg of DBM. Data areexpressed as mean7SD, n 5 3 or 4

20

25Regular emulsion

Nanoemulsion

15

Fab

s (%

)

5

10

F

00 8 16 24 32 40 48

Time (hr)

(A)

(B)

Figure 4. (A) The mean plasma concentration–time profiles ofdibenzoylmethane after oral administration in nanoemulsionor regular emulsion with 250 mg/kg of DBM. Data areexpressed as mean7SD, n 5 4. (B) DBM cumulative fractionabsorbed (Fabs) after oral administration in nanoemulsion orregular emulsion with 250 mg/kg

W. LIN ET AL.70


9.4671.62, 30.476.89 mg h/ml at doses of 10, 50and 250 mg/kg, respectively. The increase of Cmax

and AUC0�N were less than dose-proportional,suggesting possible of absorption-limited phar-macokinetic properties of DBM at higher oraldoses. This absorptive mechanism of DBM mayinvolve a capacity-limited uptake transport sys-tem and may reach saturation at the highest dose.Thus, the absorption could be the rate-limitingstep when DBM is administered orally. Based ondata from oral and i.v. administrations of 10 mg/kg, the absorption rate constant ka was estimatedto be 0.026 h�1. This result suggests that absorp-tion of DBM was relatively slow in the rats withthe current emulsion vehicle. The value of ka

dropped to 0.00570.00015 h�1 at a dose level of250 mg/kg, suggesting a slower absorption athigh dose. The slow absorption process couldresult from the precipitation of DBM from vehiclein the GIT. Such an event might be more signi-ficant at high dose levels, which cause loweroral bioavailability. However, further futurein vitro experiments are required to prove theseassumptions.

Following single oral doses of DBM at 10, 50and 250 mg/kg, respectively, the absolute bioa-vailability (F�) of DBM was estimated to be13.6%, 11.5% and 7.38%, respectively (Table 2).The F� of DBM declined with increasing doses,and the reduction of F� was statistically signifi-cant as the dose increased. A reduction of F�

might result from the lower fraction beingabsorbed at the higher dose as discussed above.This low oral F� of DBM in the rats might berelated to its poor absorption in the GIT.

The fraction absorbed (Fabs) is determinedmainly by the solubility and stability of the drugin the GIT and the permeability across the intes-tinal membrane. In silico prediction of in vivoabsorption of DBM was performed using soft-ware GastroPlus (Simulations, CA). Dibenzoyl-methane is a lipophilic drug with a Log P valueof 2.7 and Log D value of 2.72 at pH 6.8,calculated by GastroPlus. The pKa of DBM waspredicted to be 8.64, so it is a weak acid thatmight precipitate in the stomach. The solubilityof DBM was predicted to be as low as 22 mg/ml atpH 6.8.

The predicted absorption of DBM by Gastro-plus software in rat jejunum was about 17.6%and the total absorption in the GI tract was about36.0% at a dose of 100 mg/kg (Figure 5). Theresults from in silico prediction further confirmed

Table 2. Pharmacokinetics characteristics of dibenzoylmethane after oral administration

Parameter Dose (mg/kg)

10 50 250 250 (nanoemulsion)

Number of animals 4 3 4 4Tmax, h 2.0071.35 2.0070.00 4.072.0 1.6770.29Cmax, mg/ml 0.3470.04 1.5070.41 3.0270.66 4.0071.07AUC0�N, mg h/ml 2.2570.33 9.4671.62 30.476.89 94.2724.8b

AUC0�N/Dose, (mg h/ml)/(mg/kg) 0.22570.033 0.19970.032a 0.12270.027a 0.37770.099b

Apparent terminal t1/2, h 1173.0 9.372.0 1178.0 1078.2MRT, h 7.672.3 9.472.0 1274.3 1076.8Ka, h-1 0.02670.044 0.02970.046 0.000570.00015 0.01470.018F�, % 13.672.3 11.571.94a 7.3871.66a 22.876.0b

aSignificant difference from corresponding parameters in comparison to 10 mg/kg in regular emulsion (po0.05).bSignificant difference from corresponding parameters between groups of 250 mg/kg in regular emulsion versus nanoemulsion (po0.05).

Figure 5. In silico prediction of compartmental absorption ofdibenzoylmethane in GI tract after oral administration inregular emulsion with 100 mg/kg of DBM in rats



the notion that poor absorption in GIT wouldcause the low bioavailability. In order to identifywhether the low absorption was caused by eitherlow permeability or low solubility, parametersensitivity analysis was performed with Gastro-plus software. The result demonstrated that thefraction absorbed of DBM was equally sensitiveto permeability and solubility (data not shown).

Since it would be difficult to improve thepermeability of DBM at this point, in the presentstudy, the drug was formulated in nanoemulsionmade with monooleate medium-chain triglycer-ides in order to enhance the absorption of DBMin this study. The mean plasma concentration–time profiles after oral administration of DBMwith the nanoemulsion and the regular emulsionat 250 mg/kg are shown in Figure 4A with somerelevant pharmacokinetic parameters listed inTable 2. The mean plasma concentration ofDBM with the nanoemulsion reached its peakat �1.7 h (Tmax), compared with a Tmax of 3.3 hwith the regular emulsion. The ka value increasedto 0.01470.018 h�1 with nanoparticle formulationcompared with 0.000570.00015 h�1 with theregular emulsion. These results indicate thatDBM administered in the nanoemulsion wasabsorbed much faster than in the regular emul-sion. The plasma concentrations of DBM withnanoemulsion were numerically higher thanthose with the regular emulsion at all investi-gated time points post oral dosing. The AUC0�N

values of orally dosed DBM in nanoemulsionand regular emulsion were 94.2724.8 and30.476.89 mg h/ml, respectively. Statistically, theAUC of DBM in nanoemulsion was significantlyhigher than that in the regular emulsion (Table 2).Compared with the bioavailability of 7.38% ofDBM in regular emulsion upon oral administra-tion, the oil-in-water (o/w) nanoemulsions ofDBM afforded the bioavailability of 22.8%, whichwas 3-fold higher than the regular emulsion ofDBM. Next, a deconvolution analysis was per-formed using WinNonlin software. As shown inFigure 4B, the fraction absorbed in GIT was about23.4% with nanoparticle emulsion, comparedwith 10.8% with the regular emulsion. Further-more, the slope of the absorption curve with thenanoparticle emulsion was steeper than theregular emulsion, indicating a faster absorptionrate. The nanoparticle formulation of DBM might

have resulted in better solubility, thus couldachieve improved absorption in GIT. The resultslend support to our initial speculation that thepoor absorption of DBM in GIT could havecontributed considerably to the low F� in rats.

The findings suggested that the nanoparticleemulsion is more effective in increasing theabsorption and exposure of DBM in rats. Thenanoparticle drug delivery system for DBMseems to be effective in increasing the absorptionin rats and increasing exposure. With increasinginterest in DBM as a potential cancer chemopre-ventive agent, there is a growing necessity toaddress certain factors about its chemical andpharmacokinetics properties, such as its pooraqueous solubility, resulting in low ka andbioavailability. Therefore, a new drug deliverysystem was proposed that uses nanotechnologyin an effort to overcome the possible problem ofDBM precipitating out of the drug deliveryvehicle due to its poor solubility in water andto increase the absorption of DBM in the GIT. Theresults of this study indicate that the nanoparticleemulsion may be a promising strategy to deliverDBM more efficiently.

Similar to many isothiocyanate cancer chemo-preventive compounds, the pharmacodynamic/chemopreventive activity of DBM appears to beclosely related to its activation of Nrf2-mediateddrug metabolizing/detoxifying, and antioxidantenzymes. However, the impact of Nrf2 knockoutgenotype on the pharmacokinetics of drugs hasnot been investigated. This is important particu-larly for the long-term therapeutic use of chemo-preventive compounds which are Nrf2activators. Therefore, the pharmacokinetics ofDBM were compared in Nrf2 (1/1, wildtype)and Nrf2 (�/�, knockout) mice. The DBM meanplasma concentration–time profiles are shown inFigure 6A. The basic pharmacokinetics para-meters obtained from the non-compartmentanalysis are listed in Table 3. The terminal half-life of DBM in Nrf (1/1) mice was 6 h, whichwas shorter than that in the rats. After an i.p.administration of DBM at 100 mg/kg, the con-centration of DBM reached its peak of 4.56 mg/ml(20.4 mM) at 15 min. After an oral administrationof DBM at 500 mg/kg, the concentration of DBMreached its peak of 17.4 mg/ml (77.5 mM) at 15 min(Figure 6A). In the Nrf2 (�/�) mice, compared

W. LIN ET AL.72


with that in the Nrf2 wild type mice, not only thepeak concentration, mean plasma concentrations,but also the mean plasma concentrations at theearlier time points were always lower than theirrespective concentrations in the wild type mice(Figure 6B). This interesting phenomenon sug-gests that DBM has poorer oral absorption inNrf2 knockout mice than that in the wild typeC57BL6J mice. However, in contrast to the earlytime points, the mean plasma concentrations ofDBM at later time points were consistentlyhigher in the Nrf2 knockout mice (Figure 6B),suggesting that the Nrf2 knockout mice mightmetabolize DBM slower than in the wild typemice. Similarly, the mean plasma concentrationsof DBM were always higher in the Nrf2 knockoutmice than that in the wild type mice after the i.p.administration (Figure 6A). This phenomenonwas also accompanied by a 2-fold increase in theapparent terminal half-life and a 2.3-fold increaseof the plasma AUC0�N (Table 3) in Nrf2 knock-out mice after the i.p. administration. Since thei.p. administration avoids the small intestinalabsorption as well as the metabolic processes ascommonly encountered after the oral adminis-tration, these results suggest that the Nrf2knockout mice might have poorer intestinalabsorption and slower systemic eliminationprocess for DBM compared with that in theNrf2 wild type mice. It has been shown pre-viously that sulforaphane, another potent activa-tor of Nrf2 [21–24], induces cytochrome P450enzymes, therefore the Nrf2 deficiency couldpossibly result in a decrease of Vmax in the overallenzyme kinetics involved in the metabolism ofDBM in the liver and or extra-hepatic, resultingin a slower systemic elimination process.

10

100 Nrf2 wildtype PO 500 mg/kgNrf2 knockout PO 500 mg/kgNrf2 wildtype IP 100 mg/kgNrf2 knockout IP 100 mg/kg

0.1

1

0 5 10 15 20 25 30 35 400.01

Time (h)

Nrf2 wildtype PO 500 mg/kg

20

30 Nrf2 knockout PO 500 mg/kgNrf2 wildtype IP 100 mg/kgNrf2 knockout IP 100 mg/kg

10DB

MD

BM

ug

/mL

ug

/mL

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.10

Time (h)

90

100

110wildtype liver microsomeknockout liver microsome

50

60

70

80

0 10 20 30 40 50 60 7040

50

Time (min)

f C

on

tro

lec

enta

ge

of

Pre

(A)

(B)

(C)

Figure 6. (A) The mean plasma concentration–time profiles ofdibenzoylmethane in Nrf2(1/1) and Nrf2(�/�) C57BL/6Jmice after oral (p.o.) and intraperitoneal (i.p.) administration.Data are expressed as mean 7 SD, n 5 3 or 4. (B) DBM plasmaconcentrations at early time points (from 2 min to 1 h).(C) Genetic impact of Nrf2 on the stability of DBM (50mM)in mouse liver microsomes

Table 3. Pharmacokinetic parameters of dibenzoylmethane in Nrf2(1/1) and Nrf2(�/�) C57BL/6J mice after an oral (500 mg/kg)and intraperitoneal (100 mg/kg) administration by a non-compartment analysis

Route p.o. i.p.

Genotype Nrf2(1/1) Nrf2(�/�) Nrf2(1/1) Nrf2(�/�)Number of animals 3 3 3 3Cmax, mg/ml 23.2 17.4 4.56 6.86Tmax, h 0.25 0.25 0.25 0.25AUC, mg h/ml 64.6 85.6 (m 1.3)a 9.62 22.1 (m 2.3)a

AUC0�N/Dose, (mg h/ml)/(mg/kg) 0.129 0.171 0.096 0.221Apparent terminal t1/2, h 6.09 8.52 2.35 5.04MRT, h 4.73 6.36 2.76 5.15

aFold increase over the Nrf2(�/�) mice.



The two major Phase II conjugating enzymesulfotransferase SULT1B1 and UDP-glucuronyl-transferase UGT2B5 were shown to be down-regulated by 2.5- and 2.3-fold in the liver of Nrf2knockout mice (unpublished data in our labora-tory). Dibenzoylmethane appears to be hydro-xylated by cytochrome P450 enzymes and thenforms conjugates by sulfation and glucuronida-tion (unpublished observations). Since formingmore water-soluble Phase II conjugates afterPhase I metabolism is one of the most importantpathways to eliminate xenobiotics (such as toxinsand drugs), the impaired expression of SULT1B1and UGT2B5 in the Nrf2 knockout mice livermight cause significant accumulation of DBMphase I metabolite (hydroxylated DBM);increased concentration of Phase I reactionproducts may also slow down the metabolismof DBM by Phase I enzymes (product inhibition).To validate this speculation, a stability assay ofDBM (50 mM) was next performed in liver micro-somes from Nrf2 (1/1) and Nrf2 (�/�) mice.As shown in Figure 6C, in the presence ofNADPH and UDPGA, the Nrf2(�/�) livermicrosomes metabolized DBM at 50 mM slowerthan the Nrf2 (1/1) liver microsomes. After anoral administration of 500 mg/kg DBM, DBMconcentration were in the range 31–95 mM within4 h after dosing, therefore, 50 mM DBM waschosen for testing. Using Gastroplus simulation,the liver concentration of DBM was estimated tobe about 2-fold higher than its plasma concentra-tions (data not shown). The impaired livermicrosomal activity in Nrf2 knockout mice athigher DBM concentration may cause moreaccumulation of DBM in Nrf2 knockout mice,which could be the explanation for the higherconcentrations of DBM observed in Nrf2 knock-out mice at later time points. Taken together, thechanges in expression of Phase II conjugatingenzymes caused by genetic knockout of Nrf2gene could contribute to alteration of the phar-macokinetics profile of DBM in mice, whichresulted in an overall increase of DBM plasmaexposure (AUC0�N) in the Nrf2 knockout mice.

In conclusion, the pharmacokinetics of DBMwas studied after a single i.v. and oral adminis-tration in rats. Dibenzoylmethane demonstratedlinear pharmacokinetics in the dose range10–20 mg/kg following i.v. administration. It is

a low CL natural dietary compound with arelatively long half-life. Dibenzoylmethane hasa low oral bioavailability (below 13%) andappears to be distributed quite extensively tothe tissues in rats. A higher F� of DBM wasachieved with a nanoparticle emulsion, whichappears to overcome its poor absorption in theGIT. The potential differential gene expression ofPhase II conjugating enzymes caused by thedeficiency of Nrf2 gene in mice could lead toalteration of the pharmacokinetic profiles ofDBM, which results in an overall increase inDBM plasma level and exposure.

Acknowledgements

This study was supported in part by NIH grantsR01-CA094828 and R01-CA118947. The authorsthank Pin-Wen Wang at Novartis for her assis-tance in statistics as well as the reviewers for thevery constructive comments during the review ofthis manuscript.

References

1. Fukai T, Nomura T. Five isoprenoid-substitutedflavonoids from Glycyrrhiza eurycarpa. Phytochem-istry 1994; 35: 515–519.

2. Lin CC, Tsai YL, Huang MT et al. Inhibition ofestradiol-induced mammary proliferation by di-benzoylmethane through the E2-ER-ERE-depen-dent pathway. Carcinogenesis 2006; 27: 131–136.

3. Lin CC, Lu YP, Lou YR et al. Inhibition by dietarydibenzoylmethane of mammary gland prolifera-tion, formation of DMBA-DNA adducts in mam-mary glands, and mammary tumorigenesis inSencar mice. Cancer Lett 2001; 168: 125–132.

4. Lin CC, Ho CT, Huang MT. Mechanistic studieson the inhibitory action of dietary dibenzoyl-methane, a beta-diketone analogue of curcumin, on7,12-dimethylbenz[a]anthracene-induced mammarytumorigenesis. Proc Natl Sci Counc Repub China B2001; 25: 158–165.

5. Huang MT, Lou YR, Xie JG et al. Effect of dietarycurcumin and dibenzoylmethane on formation of7,12-dimethylbenz[a]anthracene-induced mam-mary tumors and lymphomas/leukemias inSencar mice. Carcinogenesis 1998; 19: 1697–1700.

6. Singletary K, MacDonald C, Iovinelli M, Fisher C,Wallig M. Effect of the beta-diketones diferuloyl-methane (curcumin) and dibenzoylmethane onrat mammary DNA adducts and tumors induced

W. LIN ET AL.74


by 7,12-dimethylbenz[a]anthracene. Carcinogen-esis 1998; 19: 1039–1043.

7. MacDonald CJ, Ciolino HP, Yeh GC. Dibenzoyl-methane modulates aryl hydrocarbon receptorfunction and expression of cytochromes P50 1A1,1A2, and 1B1. Cancer Res 2001; 61: 3919–3924.

8. Dinkova-Kostova AT, Talalay P. Relation ofstructure of curcumin analogs to their potenciesas inducers of Phase 2 detoxification enzymes.Carcinogenesis 1999; 20: 911–914.

9. Jackson KM, DeLeon M, Verret CR, Harris WB.Dibenzoylmethane induces cell cycle deregula-tion in human prostate cancer cells. Cancer Lett2002; 178: 161–165.

10. Pan MH, Huang MC, Wang YJ, Lin JK, Lin CH.Induction of apoptosis by hydroxydibenzoyl-methane through coordinative modulation ofcyclin D3, Bcl-X(L), and Bax, release of cyto-chrome c, and sequential activation of caspases inhuman colorectal carcinoma cells. J Agric FoodChem 2003; 51: 3977–3984.

11. Shen G, Khor TO, Hu R et al. Chemoprevention offamilial adenomatous polyposis by natural diet-ary compounds sulforaphane and dibenzoyl-methane alone and in combination in ApcMin/1mouse. Cancer Res 2007; 67: 9937–9944.

12. Khor TO, Yu S, Barve A et al. Dietary feeding ofdibenzoylmethane inhibits prostate cancer intransgenic adenocarcinoma of the mouse prostatemodel. Cancer Res 2009; 69: 7096–7102.

13. Shen G, Hong JL, Kong AN. Development andvalidation of an HPLC method for the determina-tion of dibenzoylmethane in rat plasma andits application to the pharmacokinetic study.J Chromatogr B Analyt Technol Biomed Life Sci2007; 852: 56–61.

14. Thimmulappa RK, Rangasamy T, Alam J, Biswal S.Dibenzoylmethane activates Nrf2-dependentdetoxification pathway and inhibits benzo(a)pyrene induced DNA adducts in lungs. Med Chem2008; 4: 473–481.

15. Khor TO, Huang MT, Prawan A et al. Increasedsusceptibility of Nrf2 knockout mice to colitis-associated colorectal cancer. Cancer Prev Res 2008;1: 187–191.

16. Shen G, Kong AN. Nrf2 plays an important rolein coordinated regulation of Phase II drugmetabolism enzymes and Phase III drug trans-porters. Biopharm Drug Dispos 2009; 30: 345–355.

17. Kobayashi M, Yamamoto M. Molecular mechan-isms activating the Nrf2-Keap1 pathway ofantioxidant gene regulation. Antioxid Redox Signal2005; 7: 385–394.

18. Zhang DD. Mechanistic studies of the Nrf2-Keap1signaling pathway. Drug Metab Rev 2006; 38: 769–789.

19. Davies B, Morris T. Physiological parameters inlaboratory animals and humans. Pharm Res 1993;10: 1093–1095.

20. Grundy JS, Eliot LA, Foster RT. Extrahepatic first-pass metabolism of nifedipine in the rat. BiopharmDrug Dispos 1997; 18: 509–522.

21. Hanlon N, Okpara A, Coldham N, Sauer MJ,Ioannides C. Modulation of rat hepatic andpulmonary cytochromes P450 and phase IIenzyme systems by erucin, an isothiocyanatestructurally related to sulforaphane. J Agric FoodChem 2008; 56: 7866–7871.

22. Hu R, Hebbar V, Kim BR et al. In vivo pharma-cokinetics and regulation of gene expressionprofiles by isothiocyanate sulforaphane in therat. J Pharmacol Exp Ther 2004; 310: 263–271.

23. Paolini M, Perocco P, Canistro D et al. Induction ofcytochrome P450, generation of oxidative stressand in vitro cell-transforming and DNA-dama-ging activities by glucoraphanin, the bioprecursorof the chemopreventive agent sulforaphane foundin broccoli. Carcinogenesis 2004; 25: 61–67.

24. Yoxall V, Kentish P, Coldham N et al. Modulationof hepatic cytochromes P450 and phase IIenzymes by dietary doses of sulforaphane in rats:Implications for its chemopreventive activity. Int JCancer 2005; 117: 356–362.



pharmacokinetics of dietary cancer chemopreventive compound dibenzoylmethane in rats and the impact...

Documents