Journal of Controlled Release 210 (2015) 160–168

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

G5-PEG PAMAM dendrimer incorporating nanostructured lipid carriers enhance oral bioavailability and plasma lipid-lowering effect of probucol Rong Qi a,b,⁎, Yan-zhi Li a,b,d, Cong Chen a,b, Yi-ni Cao a,b, Mao-mao Yu a,b, Lu Xu a,b, Bing He c, Xu Jie a,b, Wen-wen Shen a,b, Yu-nan Wang a,b, Mallory A. van Dongen e, Guo-qing Liu a,b, Mark M. Banaszak Holl e, Qiang Zhang c, Xue Ke d a

Peking University Institute of Cardiovascular Sciences, Peking University Health Science Center, Peking University, Beijing 100191, China Key Laboratory of Molecular Cardiovascular Sciences, Ministry of Education, China c School of pharmaceutical sciences, Peking University, Beijing 100191, China d School of pharmacy, China Pharmaceutical University, Nanjing 210009, China e Department of Chemistry, University of Michigan, Ann Arbor, MI 48109-1055, USA. b

a r t i c l e

i n f o

Article history: Received 18 November 2014 Received in revised form 8 April 2015 Accepted 20 May 2015 Available online 21 May 2015 Keyword: PEG-PAMAM dendrimers Nanostructured lipid carriers Probucol Water solubility Oral bioavailability Lipid-lowering effect

a b s t r a c t This work aimed to improve the oral bioavailability and plasma lipid-lowering effect of probucol (PB) by constructing a combined drug delivery system (CDDS) composed of nanostructured lipid carrier (NLC) and PEGylated poly(amidoamine) dendrimer (PEG-PAMAM). PEG-PAMAM with dendrimer generations of 5 (G5PEG) or 7 (G7-PEG) were incorporated in PB-NLCs to form PB-CDDSs, PB-NLCs/G5-PEG and PB-NLCs/G7-PEG. The resultant two kinds of PB-CDDSs were characterized by particle size, zeta potential, drug encapsulation efficacy, PB release rates, and physical stability. Formulation effects of NLC and CDDS on the cellular uptake of hydrophobic drug were explored in Caco-2 cells by fluorescent Cy5 dye as a hydrophobic drug model. Furthermore, in vivo pharmacokinetics of the PB-CDDS composed of G5-PEG and PB-NLCs were investigated in a low density lipoprotein receptor knockout (LDLr−/−) mouse model, including plateau plasma PB concentrations after oral administration of multiple doses, and bioavailability after oral administration of a single dose of different PB formulations. In addition, lipid-lowering effect of PB-NLCs/G5-PEG was studied. The results indicate that both G5-PEG and G7-PEG significantly improved aqueous solubility of PB. The two PB-CDDSs exhibited similar particle size (around 150 nm) as PB-NLCs, but slower PB burst release rate, higher total PB release amount, and better particle morphology and storage stability than PB-NLCs. In comparison with traditional NLC, CDDS dramatically enhanced cellular uptake of Cy5 into Caco-2 cells. In vivo results demonstrate that PB-NLCs/G5PEG had the highest plateau plasma PB concentration and oral bioavailability, and the greatest cholesterollowering effect in comparison with PB suspensions and PB-NLCs. Therefore, G5-PEG incorporating NLC can be exploited as a promising drug delivery system to improve oral bioavailability and lipid-lowering effect of PB. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Poorly water soluble drugs with low or variable oral bioavailability are complicated to formulate for medical development, and also have limited therapeutic effects in clinic. Probucol (PB) is a phenolic compound with antioxidative, hypocholesterolemic and anti-inflammatory effects [1], and is used for prevention and treatment of cardiovascular diseases in clinic [2,3]. Its poor water solubility (only 5 ng·mL−1) [4] results in its weak and variable absorption in gastrointestinal tract, and hence undesirable oral bioavailability of less than 10%, which dramatically affects its pharmacodynamics and clinical applications. This problem has not been solved by commercial PB tablets so far. Therefore, ⁎ Corresponding author at: 38 Xueyuan Road, Peking University Institute of Cardiovascular Sciences, Peking University Health Science Center, Beijing 100191, China. E-mail address: [email protected] (R. Qi).

http://dx.doi.org/10.1016/j.jconrel.2015.05.281 0168-3659/© 2015 Elsevier B.V. All rights reserved.

it is essential to develop a more effective formulation to enhance PB's oral bioavailability. Some PB nano-formulations, such as solid dispersions [5] and nanoself-emulsifying delivery systems [6] have been studied and reported. These nano-formulations improved PB water solubility and cellular uptake to some extent, but many problems remained to be solved, including low PB loading capacity, storage stability, carrier aging, potential toxicity from residual organic solvent, and gastrointestinal mucosal injury caused by surfactants. Nanostructured lipid carriers (NLCs), a novel nano-scaled drug delivery system developed from solid lipid nanoparticles (SLNs) [7], has a matrix structure composed of physiologically compatible and biodegradable solid and liquid lipids. Incorporation of liquid lipid leads to crystal disorder and interruption of solid lipid molecules, and therefore, increases drug loading and decreases leakage of the drugs during storage [8]. Incorporation of proper amount of liquid lipid has no influence

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

on the solid core of the NLC nanoparticles, but ensures constant release of drugs. Dendrimers represent a novel class of macromolecules which can be used as drug delivery carriers to improve characteristics of the loaded drugs, including increased water solubility, prolonged circulation time, improved stability in in vivo surroundings, and targeting to disease tissues [9,10]. In addition, researchers found that dendrimers can promote drug transport across the gastrointestinal tract [10]. The concept of combined drug delivery system (CDDS) in which two different biomaterials or drug delivery systems are incorporated together to overcome drawbacks of traditional delivery systems has been put forward since the beginning of this century [11]. CDDS composed of dendrimers and liposomes not only increases encapsulation efficacy of the loaded drugs, but also modifies their release rates [12,13]. Our previous work demonstrates that PAMAM dendrimer modified nanoliposomes could significantly increase absorption, bioavailability [14] and pharmacodynamic effects [15] of the hydrophobic drugs. However, the limited drug loading capacity of liposomes (usually less than 10%) makes liposome formulation inconvenient to administrate to the patients, especially for those drugs with high therapeutic doses, such as PB with its clinical dosage being 500 mg·d−1. Since the drug loading capacity of NLCs (about 25%) is much greater than liposomes, CDDS composed of PEGylated PAMAM dendrimers (PEG-PAMAM) and NLC were constructed in this work, and in vitro effects of the CDDS on drug release, storage stability, and cellular uptake of PB in Caco-2 cells were evaluated. In addition, the in vivo pharmacokinetics and cholesterollowering effect of G5-PEG incorporating PB-NLCs were examined and compared with traditional PB-NLCs in a low density lipoprotein receptor knockout (LDLr−/−) hyperlipidemia mouse model. The PAMAM dendrimers used in this study are comprised of an ethylenediamine core and four amidoamine dendron arms, and terminated with amine groups, and therefore, bear positive charges [16]. For PEG-PAMAM, 10% terminal amine groups are replaced by PEG (MW 5000) [17], which not only confers the PEG-PAMAM higher biocompatibility and lower toxicity than PAMAM, but also leaves enough positive charges on the PEG-PAMAM to facilitate their binding to negatively charged lipid compartments of the PB-NLCs through static interaction [18]. Two different generations of PEG-PAMAM were employed since the more flexible G5 PEG-PAMAM (G5-PEG) and more rigid G7 PEG-PAMAM (G7-PEG) might have different effects on PB solubility and oral absorption. 1.1. Materials and animals PAMAM dendrimers obtained from Dendritech, Inc. were purified (G5-NH2, G7-NH2) or prepared (G5-PEG, G7-PEG) using previously published methods [17]. Stearic acid, glycerin monostearate and oleic acid were purchased from Changwei Pharmaceutical Excipients Technology Co., Ltd. (Shanghai, China). Lecithin (injection class, PC-57T) was purchased from Kewpie Corporation (Tokyo, Japan). Probucol was purchased from Otsuka Corporation (Chiyodakum, Tokyo, Japan). Methanol (Chromatography Grade) was purchased from Xihua Special Reagent Factory (Tianjin, China). Acetonitrile (Chromatography Grade) was purchased from Merck (Germany). Fetal Bovine Serum (FBS), Dulbecco's Modified Eagle's Medium (DMEM, 4.5 g·L−1 dayglucose), non-essential amino acids and 0.25% Trypsin-0.02% EDTA were purchased from GIBCO (USA). Hank's Balanced Salt Solution (HBSS, containing Ca2+ and Mg2+) was purchased from Maichen technology Co., Ltd (Beijing, China). Penicillin, streptomycin, Pluronic F68, Fluoresce dye Cy5, MTT and all other reagents were purchased from Sigma-Aldrich (Beijing, China). Low density lipoprotein receptor knockout (LDLr−/−) mice, male, weight of 18–25 g, 8–10-week-old, were provided by Gene Modified Animal Platform of Peking University Institute of Cardiovascular Sciences. The Laboratory Animal Care Principles (NIH publication no. 85–23, revised 1996) were followed, and the experimental protocol

161

was approved by Animal Care Committee, Peking University Health Science Center. All mice were raised under a 12-hour light/dark cycle with free access to food and water. 1.2. Determination of PB concentration PB concentration was determined by High Performance Liquid Chromatography (HPLC) using a Grace Smart C18 column (150 mm × 4.6 mm, filled with 5 μm particles). The mobile phase was acetonitrile/ water (95/5, v/v) and the flow rate was 1.0 mL/min at room temperature. Signal was detected at 242 nm. The calibration curve of peak area (A) against PB concentration (C/μg·mL−1) was calculated in an equation: A = 52,264 × C-10,785 Correlation coefficient (R2) of the equation was 0.9999 within a range of PB concentrations from 10.6 ng·mL−1 to 109 μg·mL−1. 1.3. Solubilization effects of PAMAM dendrimers on PB PAMAM dendrimers (G5-NH2, G5-PEG, G7-NH2, or G7-PEG) with a charge concentration of 174 nM or 696 nM (4 × 174 nM) were mixed with excess amount of PB (about 5 mg) in tubes and diluted with pure water to a constant volume of 10 mL. The test mixtures were then mechanically shaken (Ronghua SHA-C, China) at 200 r·min−1 for 48 h at 37 °C. Samples were taken for 1 mL every 12 h, and centrifuged (Eppendorf 5810R, Germany) at 10,450 g for 10 min. PB concentrations in supernatant were determined by HPLC. 1.4. Preparation of PB-NLCs and PB-CDDSs Emulsion-evaporation and low temperature-solidification techniques were applied to prepare PB-NLCs. Briefly, PB (25 mg), solid lipid mixed by glycerin monostearate (20 mg) and stearic acid (20 mg), lecithin (40 mg) and oleic acid (20 mg) were dissolved in ethanol (5 mL) to obtain a mixture. The resultant mixture was heated to 75 °C in a water bath. The hot mixture was then injected drop by drop into 3% Pluronic F68 solution at the same temperature and under a stir rate of 1000 r·min−1 to obtain the final mixture (30 mL). The resultant o/w emulsions were stirred for 30 min at 75 °C by a mechanical agitator to strengthen the emulsification. The PB-NLCs were obtained by solidifying the o/w emulsions in an ice water bath using a mechanical stir rate of 1000 r·min−1 (MS-H280-Pro, China) for 1 h. PB-CDDSs were constructed according to the same procedure as described above in the preparation of PB-NLCs, except for adding defined amount of G5-PEG or G7-PEG into Pluronic F68 solution, and the obtained PB-CDDSs were described as PB-NLCs/G5-PEG or PB-NLCs/G7-PEG, respectively. PEG-PAMAMs at three charge concentrations of 174 nM, 696 nM (4 × 174 nM) or 1044 nM (6 × 174 nM) were used to prepare PB-CDDSs. 1.5. Characterization of PB-NLCs and PB-CDDSs The size, zeta-potential and PDI of PB-NLCs and PB-CDDSs were measured by Malvern Zetasizer Nano-ZS (UK). Encapsulation efficacy and drug loading percentages in PB-NLCs and PB-CDDSs were determined by a filtrating membrane method. Briefly, the unloaded free PB was separated by filtering PB-NLCs or PB-CDDSs through 0.22 μm filtration membrane. Filtrate containing loaded PB in the formulations was de-emulsified by methanol, and PB concentrations were then determined and described as Cencap. The total PB concentrations (described as C0) in the formulations were determined by directly de-emulsifying the formulation with methanol without filtration. Encapsulation efficacy and drug loading percentages were calculated according to the following equations:

Encapsulation efficacyðEEÞ=% ¼ Cencap=C 0  100%

162

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

Drug loadingðDLÞ=% ¼ Cencap=W lipid  100%:

and test drug release values at the time point of i. If f2 b 50, the two profiles were considered statistically different [19].

where Cencap represents the concentration of PB loaded in NLCs or CDDSs, C0 represents total concentration of PB in NLCs or CDDSs, and Wlipid represents total mass of lipids in the formulation, including stearic acid, glycerin monostearate, lecithin and oleic acid. In addition, the morphologies of the PB formulations were observed by Cryo transmission electron microscope (TEM, FEI TECNAI F30, Netherland), and thermograms were done by differential scanning calorimeter (DSC-60, Shimadzu, Japan).

1.7. Storage thermal stability of the PB formulations

1.6. In vitro drug release rates

1.8. Cell culture of Caco-2 cells

Drug release rates from PB-NLCs, PB-NLCs/G5-PEG and PB-NLCs/G7PEG were determined by a dialysis method. The three PB formulations with a volume of 0.5 mL were placed in a dialysis bag (MW cut off 10 KD), and were diluted with 2.5 mL of simulated intestinal liquid (pH 6.8 phosphate buffer solution containing 0.5% (w/v) SDS). After both ends were clamped, the dialysis bag was placed in 100 mL of the simulated intestinal liquid as a release medium. The release system was stirred (100 r·min−1) at 37 °C, and at the designed time points (0.08, 0.17, 0.33, 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 10, 12 and 24 h), 3 mL release medium was removed for analysis and along with the addition of an equal volume of fresh media to keep sink condition. The samples were centrifuged (Eppendorf 5810R, Germany) at 10,450 g for 10 min and PB concentrations present in the release media were determined by HPLC. Cumulative PB release percentages of the formulations were calculated according to the following equation:

Caco-2 cells with passages from 30 to 50 were maintained in cell culture flasks using DMEM supplemented with 10% fetal bovine serum, 1% nonessential amino acids, 100 U·mL−1 of penicillin and 100 μg·mL−1 of streptomycin, and were cultured at 37 °C in an atmosphere of 5% CO2 and 95% relative humidity. Cell culture medium was changed every other day. The cells at 70–90% confluence were passaged at 1:3 dilution ratio by using 0.25% trypsin/EDTA solution.

2 6 Cn 6 4L=V 2 þ

The PB-NLCs were stored at 4 °C or room temperature (RT), and samples were withdrawn from the stored PB-NLCs at Day 0, Day 1, Day 5 and Day 10 to determine the size and encapsulation efficacy. Stability of the PB-CDDSs at a storage condition of 4 °C was studied to evaluate the effects of PEG-PAMAM incorporation on the stability of PB-NLCs.

1.9. Cytotoxicity of the two PB-CDDSs in Caco-2 cells The cytotoxicity of the two PB-CDDSs with respect to Caco-2 cells was tested using an MTT assay [20]. Briefly, PB-NLCs/G5-PEG and PBNLCs/G7-PEG at a dendrimer charge concentration of 174 nM, 696 nM (4 × 174 nM) or 1044 nM (6 × 174 nM) were prepared with HBSS (containing Ca2+, Mg2+) as a water phase. Caco-2 cells were seeded on 96well plates at a density of 10,000 cells per well. After growing to be confluent, the cells were incubated with the PB-CDDSs for 6 h. The PBCDDSs were then removed and replaced with 100 μL of 0.5 mg·mL−1 MTT. After a four-hour incubation period, the optical density (OD) of each sample was measured at 490 nm by a Microplate Reader (Model 550, Bio-Rad, USA). Four duplicates were done for each PB-CDDS sample. Parallel experiments consisting of wells without seeded cells were employed to obtain a background and additional control measurements were obtained using cells treated only with HBSS but without the PBCDDSs. Cell viability (%) = (OD of samples − OD of background) / (OD of control − OD of background) × 100%. The material was considered cytotoxic if cell viability was less than 90%.

3 ðC n−1 þ …… þ C 2 þ C 1 Þ  V 17 L=V 2 7  100% 5 V2

where Cn represents PB concentrations in the sample taken at the designed time points; L represents theoretical drug dosage (the unit of L/V2 agrees with that of Cn); V1 represents the constant sample volume at the designed time point; V2 represents the total volume of release medium. Similarity factor f2 between the two PB release profiles was calculated 3 2

1.10. Preparation of Cy5-NLCs and Cy5-CDDSs

7 6 100 ffi7 as follows: f 2 ¼ 50  lg6 5 where n is the number of 4rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi n ∑i¼1 ðTi−RiÞ2 1þ

Cy5 is a hydrophobic fluorescent dye with a molecular weight of 694.90. Since it has similar property as PB, Cy5 was chosen in this study to evaluate the formulation effects of NLCs and CDDSs on the

n

sampling time points, i is the time points, and Ri and Ti are the reference

7

***

6

***

5 4

8

***

8

Solubility of PB/µg· mL -1

B

1×174nM 4×174nM

***

3 2 1 0.04 0.02

Solubility of PB/µg· mL -1

A

G5-NH2 G5-PEG G7-NH2 G7-PEG

6

4

2

0 12h

24h

36h

48h

0.00 G5-NH2

G5-PEG

G7-NH2

G7-PEG

-2 Fig. 1. Effects of generations, PEGylation, charge concentrations and incubation time of PAMAM dendrimers on PB aqueous solubility (n = 3). (A) Data were collected from 24 h incubation of PB with PAMAM dendrimers at a charge concentration of 174 nM or 4 × 174 nM. (B) Dendrimers were used at a charge concentration of 4 × 174 nM. ***p b 0.001.

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

DL/%

102.5 103.7 97.8 101.3 ± 3.1

23.3 25.7 24.7 24.6 ± 1.2

150 100 50

7PE G LC s/ G PB -N

PB -N

LC s/ G 5PE G

0

B 0

Zeta potential/mV

-5 -10 -15

7PE G LC s/ G

PB -N LC s/ G 5PE G

-20

C 150

Unmodified 1×174nM 4×174nM

100

50

PE G 7-

5PE G

0

PB -N LC s/ G

Twelve 8–10-week-old male LDLr −/− mice were divided into 3 groups based on their body weight. The PB suspensions, PB-NLCs or PB-NLCs/G5-PEG at a PB dose of 65 mg·kg−1 was orally administrated to the mice for one time. Then, blood samples of mice from each group were taken orbitally at the time points of 0.5, 1, 2, 3, 4, 6, 9 and 12 h. PB concentrations in the plasma were determined by HPLC. The profiles of plasma PB concentrations (C) versus time points (t) from each group were drawn, and pharmacokinetic parameters, such as peak time Tmax and concentration Cmax, were calculated

EE/%

−22.7 −22.8 −17.2 −20.9 ± 3.2

200

Encapsulation efficacy/%

1.14. PB bioavailability after oral administration of a single dose of the PB formulations

Zeta/mV

0.364 0.359 0.373

A

1.13. Steady plasma PB concentrations On the last two days of the four-week experiment, blood samples of mice from each group were taken orbitally before drug administration. Plasma was separated by centrifuging (Eppendorf 5810R, Germany) at 1672 g for 10 min at 4 °C and then treated by a direct precipitation method. Briefly, 1 mL methanol was added into 50 μL plasma and the mixture was vortexed for 3 min. The resultant mixture was centrifuged (Eppendorf 5810R, Germany) at 10,450 g for 10 min at 4 °C. PB concentrations in supernatant were determined by HPLC as steady valley PB concentrations. To calibrate profile of PB plasma concentrations (C) versus peak area (A), 5 μL PB stock solutions were aliquoted in certain volumes to get different PB concentrations, and were then diluted into 200 μL blank plasma. After being vortexed for 3 min, 1 mL of methanol was added, and the resultant mixture was handled according to the above procedure. A calibrating profile was regressed by weighting least square method with PB plasma concentrations (C) versus peak area (A), and equation was as follows: A = 2239.67 × C-800.97 (C/μg·mL−1),R2 = 0.9943.

PDI

150.2 145.1 147.5 147.6 ± 2.6

according to a regression equation; AUC0→ t was calculated from area under the profile of C–t by a trapezoidal method. Relative bioavailability (Fr) was calculated by comparing AUC0→ t of the two tested formulations (PB-NLCs and PB-NLCs/G5-PEG) with that of PB suspension: Fr % = [AUC0→ t(test)/AUC0→ t(reference)] × 100%.

1.12. In vivo cholesterol-lowering effects of the PB formulations Sixteen 8–10-week-old male LDLr−/− mice were divided into four groups based on body weight and fast plasma total cholesterol (TC) level. The four groups of mice were orally administrated with saline (NS), PB suspension, PB-NLCs or PB-NLCs/G5-PEG (the charge concentration of G5-PEG was 696 nM) twice a day for 4 weeks in company with feeding on a high fat diet containing 20% fat and 0.5% cholesterol. A PB dosage of 130 mg·kg−1·d−1 was applied in PB suspension and the two PB formulations, which equals to a clinical human adult dosage of PB (500 mg·d−1). At the end of 4 weeks, all the mice were weighed and their fast plasma TC was determined.

Size/nm

PB -N

Caco-2 cells were seeded at a density of 1 × 106 cells per well on 6-well plates. After growing to be confluent, the cells were incubated with Cy5, Cy5-NLCs, Cy5-NLCs/G5-PEG or Cy5-NLCs/G7-PEG for 3 h. At the end of incubation, the cells were washed twice with ice-cold PBS buffer, and then trypsinized with 0.25% trypsin/EDTA solution. The resultant cell suspensions were collected in tubes and centrifuged at 1000 r·min− 1 for 5 min to get cell pellets. The cell pellets were washed with ice-cold PBS for three times, and then resuspended with 0.4 mL HBSS (containing Ca2 +, Mg2 +). Fluorescence intensity of Cy5 being taken up into the cells was determined by flow cytometry at an excitation wavelength of 646 nm and an emission wavelength of 664 nm.

1 2 3 Mean ± SD

PB -N LC s/ G

1.11. Cellular uptake of Cy5 in Caco-2 Cells

Table 1 Reproducibility of PB-NLCs preparation (n = 3).

Particle size/nm

cellular uptake of hydrophobic drug into Caco-2 cells as measured by flow cytometry. Cy5-NLCs and Cy5-CDDSs were prepared according to the same procedure as described above in the preparation of PB-NLCs and PB-CDDSs, except for replacing PB with Cy5. In Cy5-CDDSs, the final charge concentration of PEG-PAMAMs was 696 nM (4 × 174 nM).

163

Fig. 2. Characterization of PB-NLCs, PB-NLCs/G5-PEG and PB-NLCs/G7-PEG on (A) particle size, (B) zeta potential and (C) drug encapsulation efficacy (EE%).

164

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

1.15. Statistic analysis Data were presented as mean ± standard deviation (SD). To determine statistical significance of p b 0.05, unpaired two-tailed Student's t-test was performed. To determine the statistical significance of three different PB formulations, including in vitro drug release, stability, cytotoxicity and in vivo cholesterol-lowering effects, one-way ANOVA test was used (Graphpad Prism, Version 5).

and generation. At a charge concentration of 4 × 174 nM, both PAMAMs and PEG-PAMAMs significantly enhanced PB water solubility. At an equal charge concentration, G7 had greater solubilization effect than G5 (p b 0.001), and PEG-PAMAMs were better than PAMAMs (p b 0.001). Fig. 1B shows that solubilization effects of PAMAMs and PEG-PAMAMs had a saturation time point, and that maximum solubilization occurred at 24 h after incubation. The decrease of solubilization might be due to the degradation of PB after long time incubation at 37 °C [13].

2. Results and discussions 2.2. Reproducibility of PB-NLCs preparation 2.1. Enhanced solubility of PB by PAMAM dendrimers Saturated water solubility of PB was very low [4], approximately 5 ng·mL−1. Fig. 1A shows that solubilization effects of PAMAMs and PEG-PAMAMs on PB were proportional to their charge concentration

Three batches of PB-NLCs were prepared in parallel. The average size of PB-NLCs was about 147.6 ± 2.6 nm, EE was 101.3 ± 3.1%, and mean PB loading percentage was 24.6 ± 1.2% (Table 1). Therefore, the formulation and preparation procedure of PB-NLCs had good reproducibility.

Fig. 3. Cryo TEM pictures, DSC thermograms and cumulative release profiles of the PB formulations. (A) Cryo TEM pictures of Blank NLCs, PB-NLCs and G5-PEG modified NLCs (from left to right); (B) DSC thermograms of PB, Blank NLCs, physical mixture of PB and blank NLCs, PB-NLCs and G5-PEG modified PB-NLCs (from up to down); (C) Cumulative release profiles of the PB formulations of PB-NLCs, PB-NLCs/G5-PEG and PB-NLCs/G7-PEG (n = 3). Charge concentration of PEG-PAMAMs in PB-CDDSs was 696 nM.

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

A

B 400

##

100

Encapsulation efficacy/%

** Particle size/nm

165

300

**

200

100

0

4 °C RT 80

###

***

60

40

***

20

0

0

1

5

0

10

1

5

10

Time/Day

Time/Day

Fig. 4. Stability of PB-NLCs at different storage temperature and time. After the designed interval days, samples were taken for determination of size (A) and EE% (B) of PB-NLCs (n = 3). **p b 0.01, ***p b 0.001 vs Day 0 for the same temperature; ##p b 0.01, ###p b 0.001.

2.3. Characterization of the PB formulations Incorporating either G5-PEG or G7-PEG in PB-NLCs almost had no effect on particle size and encapsulation efficacy (EE%) of PB-NLCs (Fig. 2A and C), but decreased the absolute zeta potential of PB-NLCs, especially at high dendrimer charge concentration of 696 nM (4 × 174 nM), because the negative charges of NLCs were partially neutralized by the positive charges of PEG-PAMAMs. Literature showed that CDDS composed of PAMAM and liposomes improved EE% of the loaded drug [21]. In this study, incorporation of PEG-PAMAM with NLCs was not able to further increase EE% of the PB-NLCs, which was already as high as 101.3 ± 3.1%.

2.4. Morphology and DSC thermograms of the PB formulations The particle morphology of Blank NLCs, PB-NLCs and G5-PEG modified PB-NLCs was observed by Cryo TEM, and pictures are shown in Fig. 3A. The particles of the blank NLCs were transparent because of non-loading of PB. Compared to those of PB-NLCs, particles of G5-PEG modified PB-NLCs presented more spherical shape but with the similar particle size as the PB-NLCs, both of which were around 100–200 nm. DSC thermograms of pure PB, Blank NLCs, physical mixture of PB and blank NLCs, PB-NLCs and PB-NLCs/G5-PEG are presented in Fig. 3B. Crystalline PB showed a sharp endothermic peak at 128 °C. The melting peak of blank NLCs took place at 54 °C. A small melting peak of PB (125 °C) appeared in the thermogram of physical mixture of PB and blank NLCs, suggesting that crystalline PB existed in the mixture.

A ***

Particle size/nm

##

** **

160 140 120 120 0 0

1

Time/Day

5

Encapsulation efficacy/%

###

180

2.5. In vitro drug release from the different PB formulations As shown in Fig. 3C, PB release rate from PB-NLCs reached up to 50% over the first 2 h, and 70% at 24 h. PB released from the PB-CDDSs was slower in the first 2 h (45%, p b 0.01), but faster at 24 h (80%, p b 0.01) compared with that of PB-NLCs. A previous study of PEG-PAMAM modified liposomes demonstrated more sustained drug release rate than unmodified liposomes [22], which was consistent with these results. Similarity factor f2 between PB-NLCs and PB-NLCs/G5-PEG was calculated as 45.70 (p b 0.01), while PB-NLCs and PB-NLCs/G7-PEG were 47.53 (p b 0.01); however, f2 between PB-NLCs/G5-PEG and PB-NLCs/ G7-PEG was 87.88. These results indicate that cumulative release profiles were different between PB-NLCs and PB-CDDSs, but not between the two PB-CDDSs. 2.6. Storage thermal stability of the PB formulations Fig. 4 shows that the physical stability of PB-NLCs was significantly influenced by storage temperature and time, and in general, PB-NLCs had better stability at 4 °C than at RT. When the storage time reached 10 days, more significant particle aggregation was observed at RT than 4 °C, more dramatic thermal motion of the particles at RT might be an

B

## 200

However, the endothermic peak of PB totally disappeared in the formulation of PB-NLCs, indicating that PB presented in the NLC particles as either an amorphous state or a molecularly dispersed state but not a crystalline state. G5-PEG modification did not change the thermogram of PB-NLCs.

100

## ##

90

**

80

* *

## ##

***

** **

PB-NLCs PB-NLCs/G5-PEG PB-NLCs/G7-PEG

70 60 60 0 0

1

5

Time/Day

Fig. 5. Stability of PB-NLCs, PB-NLCs/G5-PEG and PB-NLCs/G7-PEG at 4 °C. At the designed time points, samples were taken for determination of size (A) and EE% (B) of three PB formulations (n = 3). Charge concentration of PEG-PAMAMs used in CDDSs was 696 nM. *p b 0.05, **p b 0.01, ***p b 0.001 vs Day 0 for the same formulation; ##p b 0.01, ###p b 0.001.

166

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

with that at Day 0 (Fig. 5). Compared with PB-NLCs at Day 5, the size of PB-NLCs/G5-PEG and PB-NLCs/G7-PEG expanded only by 9.2% and 14.8%, respectively (p b 0.01); and EE% decreased only by 7.9% and 8.6%, respectively (p b 0.01). These results indicate that PB-CDDSs successfully retard the size expansion and PB degradation or leakage from the NLCs (Fig. 5). Therefore, the physical stability of PB-NLCs was significantly improved by incorporation of PEG-PAMAMs in PB-NLCs.

2.7. Cytotoxicity of the two PB-CDDSs on Caco-2 cells Fig. 6A demonstrates that Caco-2 cell viability is less than 90% when G7-PEG was incorporated in PB-NLCs at a charge concentration of 6 × 174 nM. At charge concentrations of 4 × 174 nM or less, all PBCDDSs treatments resulted in the cell viabilities of greater than 90%. Therefore, PEG-PAMAMs at a charge concentration of 696 nM (4 × 174 nM) were incorporated into PB-NLCs to form the PB-CDDSs for the following experiments.

2.8. Cellular uptake of Cy5-NLCs and Cy5-CDDSs in Caco-2 cells Fig. 6B shows that Cy5-NLCs had a similar particle size as PB-NLCs, and G5-PEG modification did not significantly change the particle size of Cy5-NLCs, which was consistent with that of the PB-NLCs/G5-PEG. The results indicate that hydrophobic dye Cy5 could replace PB in cellular experiment to observe uptake of PB-NLCs and PB-CDDSs. Fig. 6C shows that NLC formulation significantly enhanced Cy5 uptake in Caco-2 cells (p b 0.001). In addition, Cy5-NLCs/G5-PEG, compared with Cy5-NLCs/G7-PEG and Cy5-NLCs alone, showed the greatest enhancement (p b 0.001) in Cy5 uptake in Caco-2 cells, while Cy5-NLCs/G7-PEG had no difference with Cy5-NLCs. Comparing Fig. 1 with Fig. 6C, although G7-PEG was more effective than G5-PEG in improvement of PB water solubility, it was less effective in enhancement of cellular uptake of Cy5-NLCs into Caco-2 cells than G5-PEG. Due to a larger hydrophobic interior space, increased flexibility resulting from PEGylation, as well as entrapping more PB molecules by more PEG chains on every dendrimer molecule, G7-PEG was more effective than G5-PEG in improvement of PB water solubility. Nevertheless, transmembrane transport of G7-PEG was much less than G5-PEG in Caco-2 cells, perhaps because higher generation, greater molecular weight, and more rigid structure reduced the ability of G7-PEG to permeate through cell membranes as compared to G5-PEG [23]. More Cy5 internalized in Caco-2 cells might indicate more efficient in vivo absorption of PB-NLCs/G5-PEG than PB-NLCs/ G7-PEG. Therefore, PB-NLCs/G5-PEG was used for in vivo studies to explore the effects of CDDSs on oral bioavailability and pharmacodynamics of PB.

Fig. 6. Cytotoxicity and cellular uptake of the CDDSs in Caco-2 cells. (A) Cytotoxicity of PBNLCs/G5-PEG and PB-NLCs/G7-PEG at different dendrimer charge concentrations (n = 4); (B) Comparison in particle size of PB-NLCs and Cy5-NLCs with or without G5-PEG modification (n = 3); (C) Formulation effects of NLCs and CDDSs on Cy5 uptake in Caco-2 cells (n = 3). ***p b 0.001 vs Cy5; ###p b 0.001 vs Cy5-NLCs; $$$ p b 0.001 vs Cy5-NLCs/4(G7-PEG). For B & C, PEG-PAMAMs at a charge concentration of 696 nM was used in preparation of the CDDSs.

important cause; more significant decrease of EE% at RT might be resulted from the leakage or degradation of PB. When stored at 4 °C for 5 days, the size of PB-NLCs expanded by 21.5% (p b 0.001), and EE% decreased by 13.5% (p b 0.001) compared

A

B ### ##

800

NS PB PB-NLCs PB-NLCs/G5-PEG

30

TC/mg· dL-1

#

** 400

**

200

0 baseline

Week 4

body weight of mice/g

#

***

600

20

10

0 baseline

Week 4

Fig. 7. Plasma cholesterol level and body weight of LDLr−/− mice after a four-week oral administration of the PB formulations (n = 4). (A) Plasma cholesterol-lowering effects of PB raw suspensions, PB-NLCs and PB-NLCs/G5-PEG. *p b 0.05, **p b 0.01, ***p b 0.001 vs baseline for the same group; #p b 0.05, ##p b 0.01, ###p b 0.001. (B) Body weight of the mice from each group.

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

167

Fig. 8. Plasma PB concentrations after oral administration of the PB formulations to LDLr−/− mice (n = 4). (A) Steady plasma PB concentrations after administration of the formulations for 4 weeks. *p b 0.05, ***p b 0.001 vs PB; ###p b 0.001 vs PB-NLCs. (B) Profiles of PB plasma concentrations versus time after a single dose.

2.9. In vivo cholesterol-lowering effects of the PB formulations Hypercholesterolemia of LDLr−/− mice can be induced by a high fat diet [24]. As is seen in Fig. 7A, after being fed with a high fat diet for 4 weeks, plasma TC level of the LDLr −/− mice in NS treated group rose to 600 mg·dL− 1 (p b 0.001), indicating successful modeling of hyperlipidemia. Plasma TC level of the mice in PB-NLCs treated group was significantly lower than that in NS treated group (p b 0.01), but PB-NLCs/G5-PEG treated group had the best plasma cholesterollowering effects among all the three PB formulations treated groups (p b 0.001 vs NS treated group; p b 0.05 vs PB suspension treated group). It could be concluded that NLC formulation significantly improved lipid-lowering effects of PB and this effect was enhanced further by G5-PEG incorporating NLCs. In addition, all mice of the four groups in the experimental conditions showed normal appetite and activity with no obvious change of body weight (Fig. 7B), which indicates that lipidlowering effects of the PB formulations were not caused by other adverse effects on the mice, but resulted from pharmacodynamic effects of PB enhanced by NLC and CDDS formulations. 2.10. Steady plasma PB concentration after multiple doses of the PB formulations Plasma PB concentration reached a plateau after multiple dose administration of PB or the PB formulations for almost 4 weeks, which was considered as the steady plasma concentration. Fig. 8A shows that steady PB plasma concentration from PB-NLCs treated group was significantly higher than that from PB suspensions treated group (p b 0.05). Moreover, PB-NLCs/G5-PEG treated group had the highest PB plasma concentration among all the three groups, which was significantly higher than PB-NLC treated group (p b 0.001).

(Fig. 8B) demonstrate that both PB-NLCs and PB-NLCs/G5-PEG significantly enhanced PB oral absorption and plasma concentration, but the CDDS formulation was more effective than the traditional NLCs, which explained very well why cholesterol-lowering effects (Fig. 7A) of the PB formulations ranked as an order of PB-NLCs/G5-PEG N PBNLCs N PB suspensions. Table 2 shows that the bioavailability of PB-NLCs and PB-NLCs/G5PEG was significantly higher than that of PB suspensions by 68.7% (p b 0.05) and 114.4% (p b 0.05), respectively. The pharmacokinetic parameters, such as AUC0→ 12h, Cmax and Tmax, were analyzed by variance analysis using SPSS software. The results indicate that AUC0→ 12h and Cmax were significantly different (p b 0.05) between PB suspensions and PB-NLCs or PB-NLCs/G5-PEG, among which PB-NLCs/G5-PEG had the highest values (p b 0.05). In vitro cellular uptake experiments showed that PEG-PAMAMs incorporating NLCs promoted cellular uptake of hydrophobic fluorescent dye more effectively than traditional NLCs, which agreed well with our in vivo pharmacodynamic and pharmacokinetic results. Therefore, it could be concluded that NLCs improved oral bioavailability of PB significantly, and G5-PEG incorporating NLCs strengthened this effect further (p b 0.05). PAMAM dendrimers have been reported to act as absorption enhancers to improve the gastrointestinal absorption of drugs [25] up to 11 times, while traditional enhancers including surfactants, bile salts, chelating agents and fatty acids only increase absorption about 4 times [26]. The reasons why PB-NLCs/G5-PEG promoted PB oral absorption effectively were as follows: 1) the PB-NLCs/G5-PEG slowed down PB burst release rate but increased its total cumulative release (Fig. 3C); 2) G5-PEG incorporation improved storage stability of PBNLCs (Fig. 5); 3) G5-PEG facilitated PB absorption on gastrointestinal mucosa (Fig. 6C) [27]. All these factors contributed to the better lipidlowering effects and higher oral bioavailability of PB-NLCs/G5-PEG as compared to PB-NLCs.

2.11. PB C–t profiles after a single dose of the PB formulations

3. Conclusions

The plateau plasma PB concentration (Fig. 8A) and 12 h bioavailability after oral administration of a single dose of the PB formulations

CDDSs composed of PEG-PAMAMs and NLCs retard PB burst release, enhance the total cumulative PB release, improve the cellular uptake of PB, and significantly enhance thermal stability, thus improve storage. Compared to the traditional NLCs, PB-NLCs/G5-PEG significantly enhances oral bioavailability and lipid-lowering effects of PB. Therefore, CDDS composed of G5-PEG incorporating NLC could be a potential oral delivery system for the hydrophobic drug PB.

Table 2 Main pharmacokinetic parameters of PB after a single dose administration of PB suspension, PB-NLCs and PB-NLCs/G5-PEG (*p b 0.05 vs PB; #p b 0.05 vs PB-NLCs). n = 4.

Tmax/h Cmax/μg·mL−1 AUC0→ 12h/μg·mL−1·h Fr/%

PB suspensions

PB-NLCs

PB-NLCs/G5-PEG

4.0 3.05 ± 0.40 23.43 ± 4.44 100

4.0 5.19 ± 0.73* 39.53 ± 9.43* 168.7

4.0 6.08 ± 0.51*,# 50.23 ± 10.75*,# 214.4

Acknowledgement This work was supported by the National Natural Science Foundation of China (No. 81270368) and National Basic Research Program of

168

R. Qi et al. / Journal of Controlled Release 210 (2015) 160–168

China (No. 2009CB930300). We also thank Professor ZHANG Na and her research group from Shandong University for technique support of NLC preparation. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jconrel.2015.05.281. References [1] S. Yamashita, Y. Matsuzawa, Where are we with probucol: a new life for an old drug? Atherosclerosis 207 (2009) 16–23. [2] M.M. Buckley, K.L. Goa, A.H. Price, R.N. Brogden, Probucol. A reappraisal of its pharmacological properties and therapeutic use in hypercholesterolaemia, Drugs 37 (1989) 761–800. [3] S. Yamashita, H. Hbujo, H. Arai, M. Harada-Shiba, S. Matsui, M. Fukushima, Y. Saito, T. Kita, Y. Matsuzawa, Long-term probucol treatment prevents secondary cardiovascular events: a cohort study of patients with heterozygous familial hypercholesterolemia in Japan, J. Atheroscler. Thromb. 15 (2008) 292–303. [4] Y.T. Naomi Yagi, Harumi Kenmotsu, Hitoshi Sekikawa, Masahiko Takada, Dissolution behavior of probucol from solid dispersion systems of probucol polyvinylpyrrolidone, Chem. Pharm. Bull. 44 (1996) 241–244. [5] Y. Kubo, Y. Terashima, N. Yagi, H. Nochi, K. Tamoto, H. Sekikawa, Enhanced bioavailability of probucol following the administration of solid dispersion systems of probucol-polyvinylpyrrolidone in rabbits, Biol. Pharm. Bull. 32 (2009) 1880–1884. [6] E. Ishida, C. Managit, S. Kawakami, M. Nishikawa, F. Yamashita, M. Hashida, Biodistribution characteristics of galactosylated emulsions and incorporated probucol for hepatocyte-selective targeting of lipophilic drugs in mice, Pharm. Res. 21 (2004) 932–939. [7] X. Lin, X. Li, L. Zheng, L. Yu, Q. Zhang, W. Liu, Preparation and characterization of monocaprate nanostructured lipid carriers, Colloids Surf. A Physicochem. Eng. Asp. 311 (2007) 106–111. [8] E.B. Souto, S.A. Wissing, C.M. Barbosa, R.H. Muller, Development of a controlled release formulation based on SLN and NLC for topical clotrimazole delivery, Int. J. Pharm. 278 (2004) 71–77. [9] R. Esfand, D.A. Tomalia, Poly(amidoamine) (PAMAM) dendrimers: from biomimicry to drug delivery and biomedical applications, Drug Discov. Today 6 (2001) 427–436. [10] E.R. Gillies, J.M. Frechet, Dendrimers and dendritic polymers in drug delivery, Drug Discov. Today 10 (2005) 35–43. [11] G. Purohit, T. Sakthivel, A.T. Florence, Interaction of cationic partial dendrimers with charged and neutral liposomes, Int. J. Pharm. 214 (2001) 71–76. [12] K. Gardikis, C. Tsimplouli, K. Dimas, M. Micha-Screttas, C. Demetzos, New chimeric advanced Drug Delivery nano Systems (chi-aDDnSs) as doxorubicin carriers, Int. J. Pharm. 402 (2010) 231–237. [13] K.N. Kontogiannopoulos, A.N. Assimopoulou, S. Hatziantoniou, K. Karatasos, C. Demetzos, V.P. Papageorgiou, Chimeric advanced drug delivery nano systems (chi-aDDnSs) for shikonin combining dendritic and liposomal technology, Int. J. Pharm. 422 (2012) 381–389.

[14] R. Qi, H. Zhang, L. Xu, W. Shen, C. Chen, C. Wang, Y. Cao, Y. Wang, M.A. van Dongen, B. He, S. Wang, G. Liu, M.M. Banaszak Holl, Q. Zhang, G5 PAMAM dendrimer versus liposome: a comparison study on the in vitro transepithelial transport and in vivo oral absorption of simvastatin, Nanomed. Nanotechnol. Biol. Med. 11 (2015) 1141–1151. [15] Q. Ma, Y. Han, C. Chen, Y. Cao, S. Wang, W. Shen, H. Zhang, Y. Li, M.A. van Dongen, B. He, M. Yu, L. Xu, M.M. Banaszak Holl, G. Liu, Q. Zhang, R. Qi, Oral absorption enhancement of probucol by PEGylated G5 PAMAM dendrimer modified nanoliposomes, Mol. Pharm. 12 (2015) 665–674. [16] M. Jedrych, K. Borowska, R. Galus, B. Jodlowska-Jedrych, The evaluation of the biomedical effectiveness of poly(amido)amine dendrimers generation 4.0 as a drug and as drug carriers: a systematic review and meta-analysis, Int. J. Pharm. 462 (2014) 38–43. [17] R. Qi, Y. Gao, Y. Tang, R.R. He, T.L. Liu, Y. He, S. Sun, B.Y. Li, Y.B. Li, G. Liu, PEGconjugated PAMAM dendrimers mediate efficient intramuscular gene expression, AAPS J. 11 (2009) 395–405. [18] P. Kolhe, E. Misra, R.M. Kannan, S. Kannan, M. Lieh-Lai, Drug complexation, in vitro release and cellular entry of dendrimers and hyperbranched polymers, Int. J. Pharm. 259 (2003) 143–160. [19] J.Z. Duan, K. Riviere, P. Marroum, In vivo bioequivalence and in vitro similarity factor (f2) for dissolution profile comparisons of extended release formulations: how and when do they match? Pharm. Res. 28 (2011) 1144–1156. [20] M.V. Berridge, A.S. Tan, Characterization of the cellular reduction of 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT): subcellular localization, substrate dependence, and involvement of mitochondrial electron transport in MTT reduction, Arch. Biochem. Biophys. 303 (1993) 474–482. [21] A. Papagiannaros, K. Dimas, G.T. Papaioannou, C. Demetzos, Doxorubicin-PAMAM dendrimer complex attached to liposomes: cytotoxic studies against human cancer cell lines, Int. J. Pharm. 302 (2005) 29–38. [22] K. Gardikis, S. Hatziantoniou, M. Bucos, D. Fessas, M. Signorelli, T. Felekis, M. Zervou, C.G. Screttas, B.R. Steele, M. Ionov, M. Micha-Screttas, B. Klajnert, M. Bryszewska, C. Demetzos, New drug delivery nanosystem combining liposomal and dendrimeric technology (liposomal locked-in dendrimers) for cancer therapy, J. Pharm. Sci. 99 (2010) 3561–3571. [23] K.M. Kitchens, M.E. El-Sayed, H. Ghandehari, Transepithelial and endothelial transport of poly (amidoamine) dendrimers, Adv. Drug Deliv. Rev. 57 (2005) 2163–2176. [24] C.L. Chang, C. Torrejon, U.J. Jung, K. Graf, R.J. Deckelbaum, Incremental replacement of saturated fats by n−3 fatty acids in high-fat, high-cholesterol diets reduces elevated plasma lipid levels and arterial lipoprotein lipase, macrophages and atherosclerosis in LDLR−/− mice, Atherosclerosis 234 (2014) 401–409. [25] D.S. Goldberg, N. Vijayalakshmi, P.W. Swaan, H. Ghandehari, G3.5 PAMAM dendrimers enhance transepithelial transport of SN38 while minimizing gastrointestinal toxicity, J. Control. Release 150 (2011) 318–325. [26] Y. Lin, T. Fujimori, N. Kawaguchi, Y. Tsujimoto, M. Nishimi, Z. Dong, H. Katsumi, T. Sakane, A. Yamamoto, Polyamidoamine dendrimers as novel potential absorption enhancers for improving the small intestinal absorption of poorly absorbable drugs in rats, J. Control. Release 149 (2011) 21–28. [27] K.M. Kitchens, R.B. Kolhatkar, P.W. Swaan, H. Ghandehari, Endocytosis inhibitors prevent poly(amidoamine) dendrimer internalization and permeability across Caco-2 cells, Mol. Pharm. 5 (2008) 364–369.