Development and Evaluation of Emulsion-Liposome ...

Copyright © 2006 American Scientific Publishers All rights reserved Printed in the United States of America

Journal of Nanoscience and Nanotechnology Vol. 6, 1–9, 2006

Development and Evaluation of Emulsion-Liposome Blends for Resveratrol Delivery Chi-Feng Hung1 , Jan-Kan Chen2 3 , Mei-Hui Liao4 , Huey-Ming Lo1 5 , and Jia-You Fang3 4 ∗ 1 School of Medicine, Fu Jen Catholic University, Taipei, Taiwan Department of Physiology, College of Medicine, Chang Gung University, Kweishan, Taoyuan, Taiwan 3 Research and Development Unit, Formosa Biomedical Technology Corporation, Taipei, Taiwan 4 Pharmaceutics Laboratory, Graduate Institute of Natural Products, Chang Gung University, Kweishan, Taoyuan, Taiwan 5 Department of Internal Medicine, Shin Kong Wu Ho-Su Memorial Hospital, Taipei County, Taiwan 2

Keywords: Resveratrol, Lipid Emulsions, Liposomes, Drug Delivery, Intimal Hyperplasia. 1. INTRODUCTION The results of several epidemiological studies have suggested that mortality from coronary heart disease can be decreased by the consumption of red wine. It is conceivable that resveratrol can play an important role in the prevention and therapy of cardiovascular diseases because it has been reported to inhibit platelet aggregation and coagulation, alter eicosanoid synthesis, and modulate lipoprotein metabolism.1 2 This compound also exhibits anticancer and anti-inflammatory activities.3 The oral bioavailability of resveratrol is poor because enterocytes efficiently metabolize resveratrol mainly by sulfate conjugation and glucuronidation,4 5 leading to an irrelevant in vivo effect by oral administration as compared to the powerful in vitro efficacy.6 Hence other routes such as a parenteral injection should be considered in order to obtain the therapeutic benefits. The design of resveratrol formulations for ∗

Author to whom correspondence should be addressed.

J. Nanosci. Nanotechnol. 2006, Vol. 6, No. 8

injection should overcome two shortcomings of resveratrol: its rapid elimination half-life in plasma (0.13 h) and its poor water solubility.7 8 Some solvents used for resveratrol loading such as dimethyl sulfoxide are inadequate and toxic and are thus unsuitable for clinical situations.9 Lipid emulsions and liposomes are potentially interesting drug delivery systems because of their ability to incorporate drugs with poor solubility within the dispersal phase. Direct contact of the drug with body fluids and tissues can be avoided, and the drug can slowly be released over a prolonged period of time.10 The aim of this present study was to develop formulations with both lipid emulsions and liposomes for resveratrol delivery. The formulations were prepared with coconut oil as the oil phase, soybean lecithin and Brij surfactants as the emulsifiers, and glycerol formal as a solubilizer. The mean droplet size, zeta potential, and drug encapsulation were evaluated to determine the physicochemical characteristics of these emulsion-liposome blends. This study utilized in vitro Franz cells to explore the influences of the

1533-4880/2006/6/001/009

doi:10.1166/jnn.2006.420

1

RESEARCH ARTICLE

Nano- and submicron-sized vesicles are beneficial for the controlled delivery of drugs. Resveratrol, the main active polyphenol in red wine, was incorporated into various combinations of emulsions and liposomes to examine its physicochemical characteristics and cardiovascular protection. The blends of emulsion-liposome were composed of coconut oil, soybean lecithin, glycerol formal, and non-ionic surfactants. Multiple systems were assessed by evaluating the droplet size, surface charge, drug encapsulation, release rate, and stability. The vesicle diameter of the systems ranged from 114 to 195 nm. The liposomal vesicles in the systems had smaller diameters (of 43 ∼ 56 nm) (F6 and F7). Drug encapsulation of ∼70% were achieved by the vesicles. The inclusion of resveratrol in these systems retarded the drug release in both the presence and absence of plasma in vitro. The emulsion-liposome blends which incorporated Brij 98 (F5) exhibited the slowest release at zeroorder for resveratrol delivery. Treatment using resveratrol in the blended formulations dramatically inhibited vascular intimal thickening, which was tested in an experimental model in which endothelial injury was produced in normal rat carotid arteries. Intraperitoneal injection of the multiple systems was associated with no or negligible liver and kidney toxicity. We concluded that encapsulation by the emulsion-liposome blends is a potent way to enhance the preventative and therapeutic benefits of resveratrol.

Development and Evaluation of Emulsion-Liposome Blends for Resveratrol Delivery

emulsion-liposome blends on the release of resveratrol in plasma. The in vivo inhibition of neointimal hyperplasia after arterial injury in a rat model was evaluated to validate the effect of the resveratrol formulations. The in vivo tolerance of the liver and kidneys was also determined.

2. MATERIALS AND METHODS 2.1. Materials Soybean lecithin (SL, Alcolec 40P® was purchased from American Lecithin Company (Oxford, CT, USA). Coconut oil and glycerol formal were obtained from Sigma Chemical (St. Louis, MO, USA). Brij 30 (tetraethylene glycol mono-n-dodecyl ether) and Brij 98 (polyoxyethylene 20 oleyl ether) were from Acros Organics (Geel, Belgium). Cellulose membranes (Cellu-Sep® T2, with a molecular weight cutoff of 6000∼8000) were supplied by Membrane Filtration Products (Seguin, TX, USA). All other chemicals and solvents were of analytical grade.

RESEARCH ARTICLE

2.2. Determination of Resveratrol Solubility An excess amount (5 mg) of resveratrol was added to 1 ml of selected buffers or oils as listed in Table I, and shaken reciprocally at 37 C for 24 h. The suspension was centrifuged at 10,000 rpm for 10 min, and the drug concentration in the supernatant was determined using high-performance liquid chromatography (HPLC) after the appropriate dilution. The HPLC system included a Hitachi L-7110 pump, a Hitachi L-7200 sample processor, and a Hitachi L-7400 UV/visible detector. A 25-cm-long, 4-mm inner diameter stainless RP-18 column (Merck, Darmstadt, Germany) was used. The mobile phase was a methanol: pH 2.6 aqueous solution adjusted by acetic acid (45:55) at a flow rate of 1.0 ml/min. The UV/visible detector was set at 310 nm. 2.3. Preparation of the Lipid Emulsion-Liposome Blends The aqueous and oil phases were separately prepared. The aqueous phase consisted of double-distilled water and emulsifiers, whereas the oil phase consisted of coconut oil Table I. Solubility of resveratrol in various vehicles. Vehicle pH 4 buffer pH 7.4 buffer pH 10 buffer Coconut oil Corn oil Olive oil Mineral oil Squalene Glycerol formal a Not detected. Each value represents the mean ± S.D. n = 6.

2

Resveratrol solubility (g/g) 11.44 ± 0.34 13.55 ± 0.80 21.43 ± 2.17 179.75 ± 8.28 85.72 ± 6.77 47.93 ± 7.27 —a —a >5000

Hung et al.

and resveratrol (0.2 % w/v of the final products). The two phases were heated separately to 60 C. The aqueous phase was mixed using a high-shear homogenizer (Pro 250, Pro Scientific, USA) for 10 min. The two phases were then combined and further homogenized for 10 min, and then subjected to probe sonicator (VCX 600, Sonics and Materials, USA) at 35 W for 20 min. The various formulations used in this study are listed in Table II. 2.4. Vesicle Size and Zeta Potential The mean particle size and zeta potential of the emulsions were measured by a laser scattering method (Nano ZS® 90, Malvern, UK). The emulsion-liposome blends were diluted 100-fold with double-distilled water before the measurement. The determination was repeated three times/sample for three samples. The stability of the emulsion-liposome blends was determined by monitoring the size and surface charge at 37 C as a function of time for 28 days. 2.5. Efficiency of Resveratrol Encapsulation The emulsion-liposome blends were centrifuged at 48 000 × g and 4 C for 30 min in a Beckman Optima MAX® ultracentrifuge (Beckman Coulter, USA) in order to separate the incorporated drug from the free form. The supernatants were analyzed by HPLC for the free form to determine the encapsulation percentage. 2.6. In Vitro Release Resveratrol release from the emulsion-liposome blends was measured using a Franz diffusion cell. The cellulose membrane was mounted between the donor and receptor compartments. The donor medium consisted of 1 ml of vehicle containing resveratrol. The receptor medium consisted of 10 ml of 30% ethanol in pH 7.4 buffer to maintain the sink condition during the experiments. The available diffusion area between cells was 1.539 cm2 . The stirring rate and temperature were kept at 600 rpm and 37 C, respectively. At appropriate intervals, 300-l aliquots of the receptor medium were withdrawn and immediately replaced with an equal volume of fresh buffer. The released amount of drug was determined by HPLC. The effect of plasma on the release characteristics was investigated by adding 2 ml of human plasma to the donor phase as release media. The experiment of adding 2 ml of normal saline served as the control. 2.7. Rat Balloon Injury Study Adult male Sprague-Dawley rats weighing 400∼500 g were obtained from the National Laboratory Animal Center (Taipei, Taiwan). Each animal was anesthetized by an intraperitoneal injection of chloral hydrate (40 mg/kg), followed by a longitudinal midline cervical incision that permitted exposure of the left common, external, and J. Nanosci. Nanotechnol. 6, 1–9, 2006

Hung et al.

Development and Evaluation of Emulsion-Liposome Blends for Resveratrol Delivery Table II. The composition and characterization of resveratrol formulations by vesicle size, zeta potential, and drug encapsulation. Code F1 F2 F3 F4 F5 F6 F7

Compositiona

Size (nm)

Zeta potential (mV)

Encapsulation (%)

10% COb + 5% SLc 20% CO + 5% SL 10% CO + 10% formal + 5% SL 10% CO + 10% formal + 5% SL + 3% Brij30 10% CO + 10% formal + 5% SL + 3% Brij98 10% formal + 5% SL 10% formal + 5% SL + 3% Brij98

196.1 ± 2.6 299.5 ± 17.3 195.2 ± 2.8 114.3 ± 2.6 147.3 ± 1.7 56.6 ± 0.5 43.4 ± 0.7

−68.8 ± 1.1 −59.7 ± 1.2 −69.2 ± 0.9 −54.6 ± 1.6 −54.7 ± 0.9 −53.0 ± 5.6 −43.4 ± 2.4

—d —d 63.0 ± 1.7 11.9 ± 0.9 71.5 ± 1.8 86.2 ± 2.2 67.7 ± 1.2

a

The ratio of liposome composition is weight ratio (%). CO, coconut oil. c SL, soybean lecithin d Not detected. Each value represents the mean ± S.D. (n = 3 for size and zeta potential, n = 4 for encapsulation). b

2.8. Histological Examination of the Rat Liver and Kidneys The tolerance of the liver and kidneys to resveratrol formulations was examined by histological observations. The aqueous solution or emulsion-liposome blends with resveratrol (3 mg/kg), which was three-times of the dose used in balloon injury study, were administered to rats via intraperitoneal injections every day for 21 days. After treatment, specimens of the liver and kidneys were taken for histological examination. Each specimen was fixed in a 10% buffered formaldehyde solution at pH 7.4 for at least 48 h. Each section was dehydrated using ethanol, embedded in paraffin wax, and stained with H&E. For each sample, three different sites were examined and evaluated under light microscopy (Eclipse 4000, Nikon, Japan). Photomicrographs of the three randomly selected sites of each sample were taken with a digital camera (Coolpix 950, Nikon, Japan). J. Nanosci. Nanotechnol. 6, 1–9, 2006

2.9. Erythrocyte Hemolysis Blood samples were obtained from a healthy donor by venipuncture and collected into test tubes containing 124 mM sodium citrate (1 volume of sodium citrate solution +9 volumes of blood). The erythrocytes were immediately separated by centrifugation at 2000 × g for 5 min and washed three times with 4 volumes of normal saline solution. Erythrocytes collected from 1 ml of blood were resuspended in 10 ml of normal saline. Immediately thereafter, 2.5 ml of 2% (w/v) dispersions of the emulsion-liposome blends and mixtures thereof in saline were incubated with 0.1 ml of the erythrocyte suspension. Incubations were carried out at 37 C with gentle tumbling of the test tube. After 1 h of incubation, the samples were centrifuged for 5 min at 2000 × g. The absorbance of the supernatant was measured at 415 nm to determine the percentage of hemolysis. Hemolysis induced with double distilled water was taken as 100%. 2.10. Statistical Analysis The statistical analysis of differences among the various treatments was performed using unpaired Student’s t-test. A 0.05 level of probability was taken as the level of significance.

3. RESULTS 3.1. The Solubility of Resveratrol in Various Vehicles Aqueous buffers and some oils were used in the solubility test as shown in Table I. It was found that resveratrol was practically insoluble in buffers with pH values of from 4 to 10. Glycerol formal exhibited the greatest solubility capacity for resveratrol. Coconut oil was superior to the other oils for enhancing resveratrol solubility. Mineral oil and squalene had the poor effect on solubility among the selected oils. 3

RESEARCH ARTICLE

internal carotid arteries. A 2F catheter was introduced through the external carotid artery into the common carotid artery. Then the balloon was inflated, and passed three times without rotation, along the common carotid artery. After deflating the balloon, the catheter was withdrawn, the external carotid artery was ligated, the neck incision was closed, and the rats were returned to their cages. The control group of rats was treated daily for 7 days prior to and 14 days following the balloon injury by intraperitoneal injections of saline. The emulsion-liposome blends with resveratrol (1 mg/kg) were also administered via intraperitoneal injections. Fourteen days after the balloon injury, animals were anesthetized by an intraperitoneal injection of chloral hydrate. The left carotid arteries were removed, fixed in 4% formaldehyde, and stained with hematoxylin and eosin (H&E) for light microscopy in a standard manner. The neointimal area and neointimal area/medial area ratio were quantitatively measured using image microscopy (Olympus BX51) and image measurement software (SPOT Application).


Released amount (µg/cm2)

Coconut oil was selected as the oil phase of the formulations for resveratrol. As depicted in Table II, the formulation composed of 10% coconut oil and 5% SL (F1) showed a droplet size of ∼200 nm. The droplet size increased (p < 0 05) as the volume of the oil phase increased from 10% to 20% (F2). The anionic fractions in SL were responsible for the negative zeta potential of F1 and the other systems as shown in Table II. When resveratrol was incorporated into the emulsions, the excess amount of drug formed white precipitates in the F1 and F2 systems. This possibly indicated insufficient drug loading with coconut oil. Hence glycerol formal was added to promote the solubility of resveratrol in these systems (F3). No precipitate formed after the addition of glycerol formal. A drug encapsulation efficiency of 63.0% was obtained using F3 (Table II). Changes in the vesicle size and zeta potential were not significant (p > 0 05) after incorporation of 10% glycerol formal in the systems (F3 vs. F1). The vesicles were possibly formed both at the oil-water interface (lipid emulsions) and in the aqueous phase (liposomes). The pH values of all formulations were ranged between 6.0 and 6.3. Brij 30 and Brij 98 were used as co-emulsifiers to strengthen the lipid layers formed by SL (F4 and F5). The addition of the non-ionic surfactants to the emulsionliposome blends led to an initial decrease in the droplet size (Table II). This incorporation also led to a decrease in the surface charges of the oil droplets. Visible deterioration of free oil was soon observed after preparation of the Brij 30-containing formulation (F4), accompanied by lower encapsulation of resveratrol as shown in Table II. On the other hand, the Brij 98-containing formulation (F5) showed high entrapment for loading resveratrol.

4

800

600

400

200

0 0

5

10

15

20

25

30

35

Time (h) Fig. 1. In vitro release of resveratrol across a cellulose membrane from an aqueous solution with 10% glycerol formal (control) and the lipid emulsion-liposome blends. Each value represents the mean and S.D. (n = 4).

presence of plasma decreased in the order of F5 < F3 < control, which was the same with that in the absence of plasma. 3.4. Rat Balloon Injury Study As shown in Figure 3A, the carotid artery not subjected to balloon injury served as the control and essentially exhibited no demonstrable neointima in the photomicrograph. Compared to the uninjured artery, there was a significant increase in intimal area and a decrease in the crosssectional area of the lumen in the injured artery, which received the injury and treatment with normal saline only

3.3. In Vitro Release The ability of the lipid emulsion-liposome blends to deliver resveratrol was examined by determining the drug release across a cellulose membrane. An aqueous solution incorporating 10% glycerol formal was used as the control. As shown in Figure 1, the release of resveratrol from the aqueous solution showed an initial burst, then leveled off after 12 h of administration. F3 retarded the resveratrol release and attained the same release amount as the aqueous solution after 24 h. The emulsion-liposome blend incorporating Brij 98 (F5) further slowed down resveratrol release (Fig. 1). In an attempt to determine the drug release in an in vivo status, the study was performed with both plasma and normal saline as release media, as it was anticipated that plasma proteins might have some effects on the release characteristics. As illustrated in Figure 2, resveratrol release from the formulations was lower in plasma than in normal saline. The trend of drug release in the

Control F3 F5 F6

1000

Control with plasma Control with normal saline F3 with plasma F3 with normal saline F5 with plasma F5 with normal saline

600

Released amount (µg/cm2)

RESEARCH ARTICLE

3.2. Physicochemical Characteristics of the Emulsion-Liposome Blends

Hung et al.

400

200

0 0

5

10

15

20

25

30

35

Time (h) Fig. 2. In vitro release of resveratrol across a cellulose membrane from an aqueous solution with 10% glycerol formal (control) and lipid emulsion-liposome blends in the presence of human plasma and normal saline. Each value represents the mean and S.D. (n = 4).

J. Nanosci. Nanotechnol. 6, 1–9, 2006

Hung et al.

Development and Evaluation of Emulsion-Liposome Blends for Resveratrol Delivery (A)

Neointima area ( µm2)

20000

15000

10000

* *

5000

*

0 No drug

Control

F3

F5

(B) 0.6

Neointima / Media

0.5 0.4 0.3

*

0.2

* *

0.1 0.0 No drug

(Fig. 3B). The intimal thickening that formed after the endothelial injury contained smooth muscle cells embedded in matrix. Evident breaks in the internal elastic lamina could be observed in some specimens. Intimal hyperplasia was evidently inhibited to a substantial degree with treatment using resveratrol in various formulations as shown in Figure 3C-3E. Compared to the normal saline treatment, a significant reduction (p < 0 05) was seen in both neointimal area and the neointima/media ratio with resveratrol in an aqueous solution (Fig. 4). This finding suggests that resveratrol may have a role in the treatment of restenosis. The inclusion of resveratrol in the multiple systems (F3 and F5) further reduced (p < 0 05) the proliferation. No significant difference (p > 0 05) was detected between the inhibitory effect on hyperplasia by the F3 and F5 formulations (Fig. 4). 3.5. In Vitro and In Vivo Tolerance of the Emulsion-Liposome Blends Since resveratrol shows strong affinity for the liver and kidneys,2 the histology of both organs was examined after 21 days of intraperitoneal administration of resveratrol. J. Nanosci. Nanotechnol. 6, 1–9, 2006

F3

F5

Fig. 4. Resveratrol inhibition of intimal thickening in the rat injury model. Rats were treated with normal saline or control, F3, and F5 formulations with resveratrol as detailed in “Methods” for 14 days after carotid injury. At sacrifice, arteries were sectioned and stained with H&E for morphologic analysis of (A) neointimal area and (B) neointimal/media ratio. ∗ p < 0 05. Each value represents the mean and S.D. (n = 3).

The drug dose used in the tolerance study was three-times higher than that in the balloon injury study to induce the possible toxicity as far as possible. Resveratrol in aqueous solution showed gross evidence of liver toxicity as manifested by focal lymphocytes in filtration in portal areas (Fig. 5A). Focally mild lymphocyte aggregations in liver lobules were also observed. This phenomenon was also shown in F3-treated rats (Fig. 5B). No significant hepatocyte injury but slight lymphocyte aggregation in liver lobules was seen in F5-treated rats (Fig. 5C). A similar result was observed in the histology of the kidneys. Both the aqueous solution and F3 showed interstitial lymphocyte infiltration near the glomerulus (Fig. 5D). This may indicate an inflammatory reaction occurred with the two formulations. Moreover, F3 produced lymphocyte infiltration in the renal tubules (Fig. 5E). There were generally no pathological changes in the kidney in the F5-treated group (Fig. 5F). To evaluate the safety of the emulsion-liposome blends themselves, the hemolytic activity was determined as shown in Table III. Resveratrol molecules were not incorporated in these formulations. The control group caused 5

RESEARCH ARTICLE

Fig. 3. Representative sections of H&E-stained carotid arteries with or without balloon injury. (A) Non-injured carotid artery; arteries treated by (B) normal saline, (C) resveratrol in aqueous solution with 10% glycerol formal (control), (D) resveratrol in F3, and (E) resveratrol in F5. (original magnification ×200).

Control

Development and Evaluation of Emulsion-Liposome Blends for Resveratrol Delivery (A)

Hung et al.

(A)

(B)

240

F3 F5

220

(C)

Size (nm)

200

(D)

180 160 140 120 100 0

(E)

5

10

15

20

25

20

25

Time (day)

(F)

(B)

–35

RESEARCH ARTICLE

Fig. 5. Representative sections of H&E-stained liver and kidney after intraperitoneal injection of resveratrol in various formulations for 21 days. (A) Liver after treatment with aqueous solution, (B) liver after treatment with F3, (C) liver after treatment with F5, (D) kidney after treatment with aqueous solution, (E) kidney after treatment with F3, (F) kidney after treatment with F5. (original magnification ×200).

significant hemolysis. In the present study, the hemolytic activity of glycerol formal was reversed when the emulsion-liposome blends were incorporated (Table III). Both F3 and F5 showed tolerable hemolysis ( 0 05) between the inhibitory ability of hyperplasia by F3 and F5 (Fig. 4). However, the inhibition by F3 was slightly higher than that by F5. Even if F5 showed desirable in vivo efficacy, if the release of resveratrol from F5 is too slow, this may lead to failure to exert its therapeutic potency (Figs. 1 and 2). Safety is an important prerequisite for injections. The liver and kidney toxicity induced following intraperitoneal injections of resveratrol formulations showed a trend of F5 < F3 < control. As compared to the other formulations, the drug was slowly released over a prolonged period from the Brij 98-containing vehicle (Figs. 1 and 2). The direct contact of the drug with the body fluids and tissues can be avoided,10 which may lead to a minimization of adverse effects (Fig. 5). The hemolytic potential of the injectable forms has generally been found to correlate with the severity of lesions.19 This relation may also apply to the toxicity in the peritoneal cavity with the intraperitoneal route. The control group without resveratrol exhibited significant hemolysis. This may have been due to the presence of 10% glycerol formal in the control. Glycerol formal undergoes slow hydrolysis and forms formaldehyde.26 Formaldehyde may induce hemolysis as reported previously.27 Erythrocytes can be protected from hemolysis by oil-in-water emulsions.28 The hemolytic activity of glycerol formal was inhibited when the multiple systems were incorporated. The higher hemolysis of F5 than F3 may have been due to the presence of Brij 98. A previous study suggested that non-ionic surfactants with structures similar to that of Brij can induce hemolysis by affecting the fluidity of the erythrocyte membrane.29 The overall results in the tolerance study indicated that treatment with the emulsionliposome blends of resveratrol was less toxic than the non-encapsulated form, suggesting the potential for therapeutic application. 8

Hung et al.

The long-term stability test for F3 and F5 showed that there was no significant change (p > 0 05) in vesicle size or zeta potential during 28 days. Surface potentials should play an important role in the stability due to electrostatic repulsion. Any factor which lowers the zeta potential may lead to instability and the resultant formation of large oil droplets.30 As a result, the emulsion-liposome blends exhibited excellent stability in normal use.

5. CONCLUSIONS Blends of lipid emulsions and liposomes with coconut oil, SL, glycerol formal, and resveratrol were developed in order to evaluate their therapeutic efficacy and safety. The liposomal size in the aqueous phase (F6 and F7) was smaller than the emulsion droplet size in the oil-water interfaces (F3 and F5). The incorporation of Brij 98 (F5) further retarded resveratrol release from the multiple systems. The in vivo results showed that resveratrol limited neointimal hyperplasia following arterial injury in rats. The inclusion of resveratrol in the emulsion-liposome blends further inhibited hyperplasia. The controlled-release ability and drug protection by the multiple systems may have contributed to the high efficacy of the restenosis treatment. The results of erythrocyte hemolysis suggested that the emulsion-liposome blends could reduce the toxicity of glycerol formal, in which resveratrol is highly soluble. Resveratrol included in the vesicles was also prevented from accumulating in the liver and kidneys, resulting in less toxicity. The vehicles developed in this study were beneficial to the preventative purposes and therapeutic applications of resveratrol.

References and Notes 1. M. Jang, L. Cai, G. O. Udeani, K. V. Slowing, C. F. Thomas, C. W. W. Beecher, H. H. S. Fong, N. R. Farnsworth, A. D. Kinghorn, R. G. Mehta, R. C. Moon, and J. M. Pezzuto, Science 175, 218 (1997). 2. L. Frémont, Life Sci. 66, 663 (2000). 3. J. Gusman, H. Malonne, and G. Atassi, Carcinogenesis 22, 1111 (2001). 4. M. I. Kaldas, U. K. Walle, and T. Walle, J. Pharm. Pharmacol. 55, 307 (2003). 5. S. Sale, R. D. Verschoyle, D. Boocock, D. J. L. Jones, N. Wilsher, K. C. Ruparelia, G. A. Potter, P. B. Farmer, W. P. Steward, and A. J. Gescher, Br. J. Cancer 90, 736 (2004). 6. D. M. Goldberg, J. Yan, and G. J. Soleas, Clin. Biochem. 36, 79 (2003). 7. J. Marier, P. Vachon, A. Gritsas, J. Zhang, J. Moreau, and M. P. Ducharm, J. Pharmacol. Exp. Ther. 302, 369 (2002). 8. B. Jannin, M. Menzel, J. Berlot, D. Delmas, A. Langon, and N. Latruffe, Biochem. Pharmacol. 68, 1113 (2004). 9. M. T. Lin, M. L. Yen, C. Y. Lin, and M. L. Kuo, Mol. Pharmacol. 64, 1029 (2003). 10. M. Jumaa, and B. W. Müller, Int. J. Pharm. 163, 81 (1998). 11. C. Washington, Adv. Drug Deliv. Rev. 20, 131 (1996). 12. M. Sznitowska, S. Janicki, E. Dabrowska, and K. ZurowskaPryczkowska, Eur. J. Pharm. Sci. 12, 175 (2001). 13. F. Liu, and D. Liu, Pharm. Res. 12, 1060 (1995).


Hung et al.


14. N. Pongcharoenkiat, G. Narsimhan, R. T. Lyons, and S. L. Hem, J. Pharm. Sci. 91, 559 (2002). 15. J. J. Wheeler, K. F. Wong, S. M. Ansell, D. Masin, and M. B. Ball, J. Pharm. Sci. 83, 1558 (1994). 16. F. Ishii and Y. Nagasaka, Colloid Surf. B: Biointerface 37, 43 (2004). 17. M. Jumaa and B. W. Müller, Eur. J. Pharm. Biopharm. 54, 207 (2002). 18. C. Weingarten, N. S. S. Magalhaes, A. Baszkin, S. Benita, and M. Seiler, Int. J. Pharm. 75, 171 (1991). 19. S. Bjerregaard, L. Wulf-Andersen, R. W. Stephens, L. R. Lund, C. Vermehren, I. Söderberg, and S. Frokjaer, J. Control. Rel. 71, 87 (2001). 20. S. Kawakami, F. Yamashita, and M. Hashida, Adv. Drug Deliv. Rev. 45, 77 (2000). 21. M. W. Liu, G. S. Roubin, and S. B. King, Circulation 79, 1374 (1989).

22. J. Burns, T. Yokota, H. Ashihara, M. E. J. Lean, and A. Crozier, J. Agric. Food Chem. 50, 3337 (2002). 23. R. S. Schwartz, N. A. Chronos, and R. Virmani, J. Am. Coll. Cardiol. 44, 1373 (2004). 24. J. Zou, Y. Huang, K. Cao, G. Yang, H. Yin, J. Len, T. Hsieh, and J. M. Wu, Life Sci. 68, 153 (2000). 25. H. Jung, L. Medina, L. García, I. Fuentes, and R. Moreno-Esparza, J. Pharm. Pharmacol. 50, 43 (1998). 26. R. E. Staples, Teratogenicity of Formaldehyde, in Formaldehyde Toxicity, edited by J. E. Gibson, NY Hemisphere Publishing Corporation (1983), pp. 51–59. 27. E. P. Orringer, and W. D. Mattern, N. Engl. J. Med. 294, 1416 (1976). 28. M. Jumaa, and B. W. Müller, Eur. J. Pharm. Sci. 9, 285 (2000). 29. E. Galembeck, A. Alonso, and N. C. Meirelles, Chem.-Biol. Interact. 113, 91 (1998). 30. J. Han, S. S. Davis, and C. Washington, Int. J. Pharm. 215, 207 (2001).

Received: 28 October 2005. Revised/Accepted: 17 January 2006.

RESEARCH ARTICLE


9