Evaluation of Surface Charge Density and Surface Potential by

Sep 6, 2007 - Carmustine-loaded catanionic solid lipid nanoparticles with serotonergic 1B receptor subtype antagonist for in vitro targeted delivery t...
1 downloads 0 Views 274KB Size
Download PDF
11228

J. Phys. Chem. B 2007, 111, 11228-11236

Evaluation of Surface Charge Density and Surface Potential by Electrophoretic Mobility for Solid Lipid Nanoparticles and Human Brain-Microvascular Endothelial Cells Yung-Chih Kuo* and I-Chun Chen Department of Chemical Engineering, National Chung Cheng UniVersity, Chia-Yi, Taiwan 62102, Republic of China ReceiVed: April 13, 2007; In Final Form: July 2, 2007

Electrophoretic mobility, ζ potential, surface charge density, and surface potential of cacao butter-based solid lipid nanoparticles (SLN) and human brain-microvascular endothelial cells (HBMEC) were analyzed in this study. Electrophoretic mobility and ζ potential were determined experimentally. Surface charge density and surface potential were evaluated theoretically via incorporation of ion condensation theory with the relationship between surface charge density and surface potential. The results revealed that the lower the pH value, the weaker the electrostatic properties of the negatively charged SLN and HBMEC. A higher content of cacao butter or a slower stirring rate yielded a larger SLN and stronger surface electricity. On the contrary, storage led to instability of SLN suspension and weaker electrical behavior because of hydrolysis of ionogenic groups on the particle surfaces. Also, high H+ concentration resulted in excess adsorption of H+ onto HBMEC, rendering charge reversal and cell death. The largest normalized discrepancy between surface potential and ζ potential occurred at pH ) 7. For a fixed biocolloidal species, the discrepancy was nearly invariant at high pH value. However, the discrepancy followed the order of electrical intensity for HBMEC system at low pH value because mammalian cells were sensitive to H+. The present study provided a practical method to obtain surface charge properties by capillary electrophoresis.

1. Introduction Capillary electrophoresis is an effective separation technique developed recently. It has been applied commonly to biomedical analysis such as the detection of drugs, proteins, DNA, and RNA.1-4 Also, distribution of sialic acid on human erythrocytes5 and alteration in the charge characteristics of damaged cell membranes6 could be resolved by electrophoresis. In addition, fixed charge density of chondroitin sulfate and keratin sulfate7 on bovine knee chondrocytes was studied by electrophoretic mobility.8 Hence, electrophoresis was regarded as one of the standard methods for the understanding of surface charge on biological entities. Moreover, capillary electrophoresis was particularly valuable for physiological interests in both practice and theory. Similarly, ζ potential was usually used in the description of electrical traits for biocolloidal surfaces and interactions. Also, the difference between ζ potential and surface potential could be one of the key indices for the effect of mobile ionic species on the surface charge of biological cells. However, surface potential could not be determined straightforwardly by scientific instruments at the moment.9 For the transport of substances from the circulation system into the central nervous system, regulation by brain-microvascular endothelial cells (BMEC) was the dominant factor.10 Variation in BMEC membrane charge could be strongly influenced by the cellular pH value, which was closely related to microfluidic motion, physicochemical stimulation, and metabolic pathway.11 When external force caused hemorrhage or ischemia in brain injury, cellular edema, activation of anaerobic metabolism, and reduction in pH value and membrane potential would occur.12 Furthermore, an increase in Na+/H+ exchange * To whom correspondence should be addressed: tel 886-5-272-0411. ext 33459; fax 886-5-272-1206; e-mail [email protected].

could result from an increase in pH value around BMEC.12 Hence, a systematic investigation on the relation between pH value and BMEC membrane potential become inevitable. Biomimic solid lipid nanoparticle (SLN) was composed of a biocompatible lipid core with apposite lipophilicity and stability as colloidal drug delivery system.13 SLN was also an efficacious carrier for the transport of brain-targeted therapeutics across BMEC.14 However, a decrease in pH value caused a compressed electrical double layer and increased proton adsorption, rendering a decrease in surface charge and stability.15,16 In addition, uptake of SLN by cells might significantly affect the tissue performance.17-19 Thus, it is of critical importance to examine the electrical properties of SLN and SLN-incorporated BMEC for the understanding of SLN delivery into the brain. In this study, effects of pH value, content of cacao butter, stirring rate, and storage period on ζ potential and electrophoretic mobility of SLN were measured experimentally. Also, effects of pH value on ζ potential and electrophoretic mobility of human BMEC (HBMEC) and SLN-absorbed HBMEC were investigated. Electrophoretic mobility was converted into surface charge density by the theory concerning condensation of counterions on polyion.20 In addition, surface potential was calculated as an analytical function of surface charge density.21 Finally, discrepancy between surface potential and ζ potential of the SLN and HBMEC systems was discussed. 2. Experimental Section Preparation of SLN. SLN with various weight ratios of cacao butter to stearic acid was prepared by a microemulsion method14 with several modifications. Briefly, 8% (w/w) lipid including cacao butter (OCG Cacao Inc., Whitinsville, MA) and stearic acid (Fluka, Buchs, Switzerland) were melted at 72 °C.

10.1021/jp072876z CCC: $37.00 © 2007 American Chemical Society Published on Web 09/06/2007

Solid Lipid Nanoparticles and Endothelial Cells

J. Phys. Chem. B, Vol. 111, No. 38, 2007 11229

Figure 1. Electrical properties of SLN with various contents of cacao butter: (]) 0%, (0) 4%, and (O) 8%. Stirring rate was 500 rpm. (a) ζ potential; (b) electrophoretic mobility; (c) average surface charge density; (d) average surface potential. L-R-Phosphatidylcholine (PC; 5% w/w; Sigma, St. Louis, MO), 10% (w/w) taurocholate (Sigma), 7.5% (w/w) ethanol (Riedelde Hae¨n, Seelze, Germany), and 69.5% (w/w) deionized water (Barnstead, Dubuque, IA) were mixed with the melted lipid. One aliquot of the emulsified liquid was added into 10 aliquots of deionized water at various stirring rates at 3 °C for 15 min. After filtration through filter paper with 1 µm pores, 2% (w/v) D-trehalose (Sigma) was added into the filtrate. The suspension containing SLN was frozen at -80 °C for 3 h. SLN powders were obtained by lyophilization (Eyela, Tokyo, Japan) at -80 °C over 24 h. The cumulant Z-average diameter of SLN was determined by a Zetasizer 3000 HSA with photo correlation spectroscopy (Version 1.41, Malvern Instruments, Worcestershire, U.K.). Fluorescent SLN was synthesized by entrapping fluorescein isothiocyante- (FITC-) conjugated dextran 70000 (Sigma) in melted lipid. For the study of the electrical properties, 8 mg of SLN was suspended in 4 mL of Tris buffer (RDH, Seelze, Germany) with a specific pH value adjusted by hydrochloric acid (Hanawa, Osaka, Japan) and sodium hydroxide (Showa, Tokyo, Japan); that is, the density of SLN was 2 mg/mL. Magnification of HBMEC and Uptake of SLN. HBMEC (Biocompare, South San Francisco, CA) with a density of 2 × 104 cells/cm2 was seeded on a dish coated with human fibronectin (Sigma) in endothelial cell medium (ECM, Bio-

compare) containing 5% fetal bovine serum (FBS, Biocompare), 1% endothelial cell growth supplement (Biocompare), 1% antibiotic antimycotic solution (100×, Sigma), 0.5% gentamicin solution (Sigma), 1% MEM nonessential amino acid solution (Sigma), 1% L-glutamine (Sigma), and 1% sodium pyruvate (Sigma). HBMEC was then cultured over 6 days in a humidified 37 °C, CO2 incubator (NuAire, Plymouth, MN) with replacement of ECM every 2 days. After a wash by Dulbecco’s phosphate-buffered saline (DPBS, Sigma), HBMEC was detached by 0.4 mL of 0.025% trypsin-0.5 mM EDTA (Sigma) and equally allotted to three dishes. Culture dishes coated with rat tail collagen (Sigma) were employed beyond passage 6, and proliferated HBMEC for passages 6-16 was utilized in the present study. HBMEC could be frozen with ECM containing 10% FBS and 10% dimethyl sulfoxide (J. T. Baker, Phillipsburg, NJ) in an ultra-low-temperature freezer (Sanyo, Osaka, Japan) at -80 °C for 1 day and then stored in liquid nitrogen. HBMEC could be unfrozen in a 37 °C water bath within 1 min. Morphology of HBMEC was examined by a phase-contrast biological microscope (Motic, Richmond, British Columbia, Canada). Uptake of SLN by HBMEC was obtained through coculture of attached HBMEC with 0.01% (w/v) SLN in the CO2 incubator for 90 min. HBMEC was washed by DPBS twice to remove the freely suspended SLN before trypsin-EDTA

11230 J. Phys. Chem. B, Vol. 111, No. 38, 2007

Kuo and Chen

TABLE 1: Average Diameter, D, and Average Specific Surface Area, Asp, of SLN with Stirring Rate of 500 rpm cacao butter (%) D (nm) Asp (m2/g)

0

1

2

3

4

5

6

7

8

165.8 22.62

203.3 18.32

225.5 16.38

254.3 14.42

268.9 13.54

274.3 13.19

296.2 12.12

324.7 10.89

385.3 9.21

treatment. A fluorescent image of SLN could be observed by applying 0.01% (w/v) FITC-labeled SLN to the HBMEC culture. Briefly, HBMEC was seeded on a coverslip coated with rat tail collagen and human fibronectin. After cultivation with FITClabeled SLN, HBMEC was dehydrated by methanol (Mallinckrodt Baker, Phillipsburg, NJ) and treated by Triton-X-100 (Acros Organics, Geel, Belgium). For immunohistochemical staining of the cytoplasm, anti-human von Willebrand factor VIII (Sigma) and anti-rabbit IgG with FITC conjugate (Sigma) were employed.10 The fluorescent image of the uptake of SLN by HBMEC was obtained by use of argon laser and FITC filter at 458 nm (excitation) and 488 nm (emission) under a phasecontrast fluoromicroscope (Axioskop 2 plus, Zeiss, MunchenHallbergmoos, Germany). HBMEC was buffered in Tris with a specific pH value at a density of 6 × 104 cells/mL for ζ potential and 4 × 104 cells/mL for electrophoretic mobility.

Measurement of ζ Potential and Capillary Electrophoretic Mobility. ζ potentials of SLN and HBMEC were measured at 25 °C by a Zetasizer 3000 HSA. Biocolloidal suspension was slowly injected into the quartz tube to avoid bubble formation. Laser Doppler velocimetry, LDV, was employed in the Zetasizer, and the ζ potential was estimated by the Henry equation. Young’s interference fringes were obtained by two laser beams in LDV, and the frequency shift was converted to the particle velocity. Hence, the average electrical behavior of the sample was evaluated by Zetasizer. Also, the Zetasizer was applicable to determine the particulate positive/negative electricity, which could be utilized to adjust the direction of the electric field in capillary electrophoresis, especially near the isoelectric point (charge reversal). Electrophoresis of the samples was performed by a P/ACE-2100 with Gold data acquisition software (Beckman Coulter, Palo Alto, CA) followed by a UV detector (Beckman

Figure 2. Electrical properties of SLN with various stirring rates: (]) 1000, (0) 500, and (O) 250 rpm. Content of cacao butter was 6%. Panels a-d, same as Figure 1.

Solid Lipid Nanoparticles and Endothelial Cells

J. Phys. Chem. B, Vol. 111, No. 38, 2007 11231

Coulter) with absorbance wavelength of 214 nm. The internal surface of a capillary was activated by 0.1 N HCl before it was filled up with Tris buffer. Note that the pKa of Tris is 8.08, and the commonly encountered acidic buffers include DPBS with pKa1 of 2.14, pKa2 of 7.1, and pKa3 of 13.3, acetate with pKa of 4.75, and 2-(N-morpholino)ethane sulfonic acid (Mes) with pKa of 6.13.22 Here, only Tris was introduced to the electrokinetic measurement. In addition, coating of 5% poly(ethylene oxide) 600 000 (Aldrich, Milwaukee, WI) on the capillary wall was required to avoid the interference by electroosmosis. Electrophoretic mobility, µ, can be determined by µ ) (l/t)/(V/L), where l ) effective capillary length, t ) migration time, V ) applied electrical potential, and L ) total capillary length.8 For the present system, l ) 38.5 cm, L ) 45 cm, V ) 10 kV, temperature was 37 °C, and the inner and outer diameters of the capillary were 75 and 375 µm, respectively. Since separation of the charged particles with different electricity was performed by capillary electrophoresis, the intensity of a specific biocolloid could be observed. In this study, the results of Zetasizer 3000 HSA and capillary electrophoresis of P/ACE-2100 were consistent due to the high purity of specimens. Calculation of Surface Charge Density and Surface Potential. Surface charge density and surface potential were not easily detected by experimental instruments. However, the two surface charge properties could be evaluated from electrophoretic mobility. From theoretical results of counterion condensation,20 the potential distribution near a polyelectrolyte was obtained. By substitution of the potential distribution into the Smoluchowski equation, the amount of adsorbed counterions was correlated to electrophoretic mobility.23 Furthermore, potential on particulate surfaces could be estimated by an analytical expression for the surface potential as a function of surface charge density.21 The formulations are briefly described as follows: ηr(1 + κr) n µ )1Z Ze

[

]

e2a(a + b)Ca0 κ) 0rkBT

(1a)

σ0 )

4πr2

experimental investigation on the fraction of condensed ions, NaCl and KBr were introduced in the medium in the range of 10-100 mM according to the counterion-condensation theory.23 Hence, the current method was appropriate for the prediction of σ0 although electrolytic species were present. Furthermore, average surface potential, φ0, could be estimated by φ0 )

{

[

]

2 1 2(q - 1) ln (q - p) + + a X k3pq

[

φ0 ) (1b)

(2)

Here, (Z - n) was evaluated by eq 1a via experimentally determined µ and r. Thus, σ0 was obtained from eq 2. In an

]}

2 1 2 ln [(q + 1)/2] 4(q - 1) (2q + 1) X2 k32pq 2k32p3q3

1/2

where n ) number of counterions on SLN, Z ) number of monovalent charged groups on SLN, η ) viscosity of surrounding fluid, r ) particle radius, κ ) reciprocal of Debye screening length, e ) elementary charge, µ ) electrophoretic mobility, a ) cationic valence, b ) anionic valence, Ca0 ) bulk concentration of cations, r ) relative permittivity, 0 ) permittivity of a vacuum, kB ) Boltzmann constant, and T ) absolute temperature. Note that the charged groups of sodium taurocholate and L-R-phosphatidylcholine on SLN were SO3- and PO4-, which were monovalent anions.24 Counterions, which might be condensed on SLN surface, were monovalent Tris+, H+, and Na+. The charged groups on HBMEC were mainly COO- and NH3+, and counterions might be Tris+, H+, and Na+ for negatively charged cells or OH- and Cl- for positively charged cells. Since only monovalent Tris+, H+, OH-, Cl-, and Na+ existed in the medium, a and b were both unity in the present study. The effective charge e(Z - n) on SLN surface could be correlated with average surface charge density, σ0, as follows -e(Z - n)

Figure 3. Effect of storage at 4 °C on the variation in the diameter of SLN: (0) 3 or (O) 6 months. Content of cacao butter was 8%, and stirring rate was 250 rpm.

(3c)

-aeσ0 2k30rkBTκ

(3d)

X ) κr

(3e)

[(k - 2)k1 + 2k2] for k e 4 k

k1 )

(3b)

q ) (p2 + 1)1/2 p)

k3 )

eφ0 kBT

(3a)

2 {(k) [(k/2)2/(k-2) - 1]} 1/2

k2 )

2 k1/2

k ) 2 + 2b/a

(3f)

(3g)

(3h) (3i)

where φ0 ) average dimensionless surface potential, p ) average dimensionless surface charge density, X ) dimensionless radius, and k3 ) parameter of electrolyte valence. Note that eq 3a is a general formula valid for charged surfaces including specific adsorption of metal ions or charged polymer molecules on colloidal surfaces.

11232 J. Phys. Chem. B, Vol. 111, No. 38, 2007

Kuo and Chen

Figure 4. Electrical properties of SLN with various storage periods: (]) 1, (0) 3, or (O) 6 months. Content of cacao butter was 8%, and stirring rate was 250 rpm. Panels a-d, same as Figure 1.

The normalized discrepancy, DSZ, between average surface potential and average ζ potential, 〈ζ〉, was defined as DSZ )

|φ0 - 〈ζ〉| × 100 |φ0|

(4)

In addition, specific surface area, Asp, of SLN could be calculated by25 Asp )

3 surface area of SLN 4πr2 ) ) (density)(volume of SLN) (F)(4/3πr3) Fr

(5)

where F ) density of SLN. 3. Results and Discussion Effect of Internal Cacao Butter on Electrical Properties of SLN. Figure 1 presents the variation in ζ potential, electrophoretic mobility, average surface charge density, and average surface potential of SLN with various contents of cacao butter as functions of pH value of the medium. Here, average surface charge density and average surface potential were evaluated from the mean electrophoretic mobility. As revealed in Figure 1, for a constant content of cacao butter, the larger the pH value, the greater the absolute value of ζ potential, electrophoretic mobility, average surface charge density, and average surface potential. This behavior was consistent with the theoretical

prediction.26 Note that cacao butter was composed of linolenic acid, oleic acid, palmitic acid, and stearic acid. The surface groups on the negatively charged SLN-containing phospatidylcholine were mainly -N+(CH3)3 and -PO4-. The adsorption equilibrium constants of OH- onto -N+(CH3)3 and Na+ onto -PO4- were, respectively, 5.25 × 109 and 0.051.27 Thus, an increase in pH value led to an increase in the amount of negative charge on SLN surface. On the other hand, the adsorption equilibrium constants of Cl- onto -N+(CH3)3 and H+ onto -PO4were, respectively, 0.218 and 5.58 × 105.27 Hence, as pH value decreased, an electroneutralization of -PO4- occurred, rendering a decrease in the SLN surface electricity. Also, for a fixed pH value, the higher the content of cacao butter, the greater the absolute value of ζ potential, electrophoretic mobility, average surface charge density, and average surface potential. This was because the charged surface ingredients, including L-R-phosphatidylcholine and taurocholate, were constant in every batch of SLN. Since the amount of surfactants was constant during fabrication, the total number of ionogenic groups was fixed throughout the experiments. Besides, an increase in the content of cacao butter yielded an increase in the SLN diameter and a decrease in the specific surface area, as indicated in Table 1. Larger SLN produced less total surface area and more concentrated surface charge on the particle surface. Hence, a high content of cacao butter resulted in strong electrostatic properties. In a study on the electrophoresis of liposomes containing

Solid Lipid Nanoparticles and Endothelial Cells distearoylphosphatidylcholine, distearoylphosphatidylglycerol, and cholesterol, the migration time was determined by the particle size.28 In the present study, the rationale behind the relationship between the content of cacao butter and the SLN size could be the crystalline structure of stearic acid and cacao butter. Note that cacao butter consists of four fatty acids with different molecular lengths, leading to low crystallinity. Lipid core containing only stearic acid might produce more flawless crystals and smaller lattice intervals than that containing multiple ingredients. Hence, the SLN size was affected by the crystalline structure of the two components. Effect of Stirring Rate on Electrical Properties of SLN. Figure 2 shows the variation in ζ potential, electrophoretic mobility, average surface charge density, and average surface potential of SLN with various stirring rates as functions of pH value of the medium. As exhibited in this figure, for a fixed pH value, the faster the stirring rate, the smaller the absolute value of ζ potential, electrophoretic mobility, average surface charge density, and average surface potential. Stirring rate during cold-water formation of SLN denoted the dissipation force, which influenced the particle size. For the present fabrication process, the average SLN diameters were 266.8 nm for 1000 rpm, 296.2 for 500 rpm, and 342.1 nm for 250 rpm, suggesting that a faster stirring rate caused a larger dissipation and a smaller particle. This result was consistent with the behavior of Gelucire 44/14 and Compritol ATO 888 system.29 Here, the amount of every reagent was constant in each batch of SLN, and the overall functional groups were distributed on all SLN surfaces. Thus, a faster stirring rate yielded smaller colloids, more specific surface area, and a lower capacity of allocated charge on the outer layer of SLN, rendering weaker electrical characteristics of the particles. Effect of Storage on Electrical Properties of SLN. Variation in the difference between average diameter of SLN before and after storage at 4 °C as a function of pH value of the medium was shown in Figure 3. Here, ∆D ) DS - DF, where DS and DF are, respectively, the average diameter after storage and that of freshly prepared SLN. For clear presentation, data of 1-month storage were not shown in this figure because no severe change in SLN diameter was observed for pH ) 1-13. As implied in Figure 3, an obvious difference in the average diameter was obtained when pH e 4 for 3-month storage and pH e 6 for 6-month storage. This was because a low pH value caused less charge and thinner double-layer thickness, yielding weaker electrostatic repulsion between colloidal particles.30 Lack of electrical protection led to low stability of the suspension and the growth of SLN, and an increase in particle size was accompanied by a decrease in surface electricity, suggesting a structural deformation of the lipid core.31 Lipid crystallites could be divided into R, β, and β′ forms, where β′ form was the most compact configuration. However, storage led to an increase in the ratio of β form in the crystalline phases.31 After 6-month storage, an increase in the higher-melting-point β form was also concluded by the thermogram of differential scanning calorimetry.32 Thus, reorientation of the charged fatty acids in the internal crystals could cause lipid deformation, rendering the reduction in surface charge and the flocculation of SLN. A proper stability would be achieved if ζ potential was larger than 30 mV, in general. In the present study, ca. 90 nm increase in average diameter was observed over 6-month storage at pH ) 1, indicating that storage in a strong acidic environment would be improper for the negatively charged SLN. Figure 4 presents the variation in ζ potential, electrophoretic mobility, average surface charge density, and average surface

J. Phys. Chem. B, Vol. 111, No. 38, 2007 11233

Figure 5. Typical micrographs of HBMEC at 400×. (a) Optical image, where dark areas are cells; (b) fluorescent image of the SLN uptake by HBMEC, where green areas are cells and bright small stains are SLN.

potential of SLN with various storage periods as functions of pH value of the medium. As displayed in this figure, for a fixed pH value, the longer the storage period, the smaller the absolute value of ζ potential, electrophoretic mobility, average surface charge density, and average surface potential. The reasons for this result concerning ionogenic ingredients are elaborated below. Negatively charged L-R-phosphatidylcholine (PC) on SLN surface could be first hydrolyzed to become oleic acid and lysophosphatidylcholine.33 The latter was further decomposed to become palmitic acid and glycerophosphoric group. Note that the hydrolytic kinetics of PC depended on temperature, pH value, buffer concentration, and ionic strength of the medium.34 In the present study, SLN suspension was stored at 4 °C, and hydrolysis of PC could be appreciable. Also, hydrolysis of PC was accelerated in acidic medium. Furthermore, oleic acid and palmitic acid adsorbed on SLN surface, and negatively charged glycerophosphoric groups desorbed from the surface. Hence, the longer the storage period, the stronger the hydrolysis of PC, and this process led to a decreased negative charge on SLN. Tersely, hydrolysis of PC yielded redistribution of surfactants and functional groups on SLN surfaces, resulting in a decrease in ζ potential of SLN and an increase in SLN size. Effect of SLN Uptake on Electrical Properties of HBMEC. Morphology of HBMEC and typical image of fluorescent

11234 J. Phys. Chem. B, Vol. 111, No. 38, 2007

Kuo and Chen

Figure 6. Electrical properties of HBMEC incubated with SLN: (]) HBMEC, (0) HBMEC containing SLN with 1% cacao butter, and (O) HBMEC containing SLN with 7% cacao butter. SLN is prepared at 500 rpm. Panels a-d, same as Figure 1.

staining for the uptake of SLN by HBMEC after incubation are shown in Figure 5. As presented in Figure 5a, the structure of HBMEC was similar to capillary cells obtained from bovine35,36 and sheep.37 As displayed in Figure 5b, the bright little spots were SLN, and the relatively dark circles in the green fluorescence were the nuclei of HBMEC. It was evidenced that SLN adhered to HBMEC membrane or penetrated into both cytoplasm and nucleus. Although the charged surface layer of mammalian cells can be analyzed by electrokinetic theory, experimental study can play an important role in the understanding of electrical properties of cellular membrane.8 Also, ζ potential of a moving biological entities with structured or nonstructured surface was experimentally detectable. In fact, variation in ζ potential of biological cells in medium with various pH values was attractive to many biochemical physicists. For example, picocyanobacteria bore positive charge in medium with pH less than 5 because of absorption of protons by basic functional groups.38 In addition, the absolute value of ζ potential of physiologically active Bacillus subtilis was lower than that of physiologically inactive B. subtilis because the pH value near the cellular surface was lower than that in the bulk.39 Also, ζ potential and cell transmission of biomass were sensitive to pH value and ionic strength in expanded bed.40 Besides, the membrane thickness

and the radius of HBMEC were ca. 10 nm and ca. 10 µm, respectively; that is the correction factor was ca. 10-3 (ratio of the two).41 Hence, a model based on rigid particles was a good approximation to estimate the properties of HBMEC. Figure 6 shows the variation in ζ potential, electrophoretic mobility, average surface charge density, and average surface potential of HBMEC incubated with SLN as functions of pH value of the medium. During motion along the capillary at pH e 3, HBMEC died owing to low surface charge and long electrophoretic time. Thus, the mobility data by capillary electrophoresis of pH g 4 were presented. On the contrary, the Zetasizer could overcome this situation before HBMEC apoptosis, and the ζ potential could be obtained. Also, as revealed in Figure 6, the lower the pH value, the smaller the absolute value of ζ potential, electrophoretic mobility, average surface charge density, and average surface potential, as pH g 4. As suggested in Figure 6a, charge reversal occurred in the range of pH < 3. The charge on cell membrane was derived mainly from glycocalyx, which was composed of glycoprotein, glycolipid, proteoglycan, and endothelial nitric oxide synthase.42 These proteins were negatively charged in physiological conditions and mainly determined the electrostatic traits of HBMEC. Ionization of the carboxyl and amino groups on HBMEC became COO- and NH3+, depending on the pH value of the

Solid Lipid Nanoparticles and Endothelial Cells

J. Phys. Chem. B, Vol. 111, No. 38, 2007 11235

TABLE 2: Normalized Discrepancy between Average Surface Potential and Average ζ Potential of SLN and HBMEC Systems SLN cacao buttera (%)

HBMEC and SLN

stirring rateb (rpm)

storage timec (months)

cacao butter (%)

pH

0

4

8

1000

500

250

1

3

6

1 2 3 4 5 6 7 8 9 10 11 12 13

19.33 19.89 25.99 31.84 32.51 37.31 42.91 40.21 35.84 34.54 34.10 34.47 34.43

2.20 7.01 22.68 27.02 29.70 36.51 40.11 37.33 36.18 35.66 34.56 35.82 36.73

5.79 10.98 24.39 26.94 29.45 36.85 42.66 40.23 37.54 37.76 38.33 39.50 39.65

11.82 16.66 25.58 27.79 31.29 41.23 41.79 39.57 37.42 36.95 34.95 36.35 36.37

6.02 11.72 22.24 23.21 26.81 40.66 47.30 45.81 43.09 42.51 40.94 41.05 41.41

7.10 9.60 13.84 21.72 23.20 37.21 46.66 44.57 41.31 40.80 40.57 41.15 41.62

13.72 21.84 24.69 24.80 32.03 40.21 45.85 41.46 36.48 33.85 31.58 27.86 27.67

12.58 20.70 22.55 22.04 25.27 37.01 41.38 40.03 34.19 33.50 26.71 26.88 27.18

20.78 25.93 25.67 29.91 30.72 34.39 36.05 33.59 32.54 32.20 28.84 26.52 26.21

d

1

7

13.55 18.27 19.77 25.51 20.05 19.24 18.94 18.43 18.10 18.06

12.30 14.77 20.19 24.30 21.62 18.81 16.38 13.76 12.15 11.47

9.52 11.36 14.18 20.13 20.09 15.11 14.77 14.84 14.93 14.07

a Stirring rate was 500 rpm. b Cacao butter content was 6%. c Stirring rate was 250 rpm and cacao butter content was 8%. d Control group: HBMEC without the incubation with SLN.

medium. Hence, charge reversal occurred around pH valuea close to the isoelectric point. For example, absorption or dissociation of H+ with the functional groups, that is, the chargeregulation behavior, could cause cellular charge reversal, which was prevalent in mammalian cells.43,44 In addition, the size of the counterion might play an important role in the charge reversal of biocolloids.45 Charge reversal might finally cause HBMEC death. Also, for a fixed pH value, incubation with SLN led to a decrease in the electricity of HBMEC. Furthermore, the higher the content of cacao butter applied to the incubation, the smaller the absolute value of ζ potential, electrophoretic mobility, average surface charge density, and average surface potential. The rationale behind this result was that lysophosphatidylcholine hydrolyzed from PC induced an activation of surface signal factors, rendering a reduction in transendothelial electrical resistance and an increased permeability.46 Hence, hydrolysis of PC on SLN was the key reason for the reduction in cellular electricity, and a higher content of cacao butter produced a higher density of PC on SLN surface, leading to weaker electrostatic behavior. The proteinic groups on HBMEC might be carried away from the membrane phase during endocytosis of choline-containing SLN, causing a reduction in electricity.47 It was also inferred that the main route for the uptake of SLN by HBMEC was intermembrane transfer.14 Note that a higher content of cacao butter yielded a larger SLN, which might remove more charged ingredients in the HBMEC membrane and reduce the HBMEC electricity. Difference between Surface Potential and ζ Potential. Table 2 lists the normalized discrepancy between average surface potential and average ζ potential of the present SLN and HBMEC systems. As revealed in this table, the largest normalized discrepancy occurred at pH ) 7 because of the smallest amount of proton in the medium. Note that proton and hydroxyl were the potential determining ions, PDI, in the present systems, and PDI dominated the distribution of electrical potential. Also, the normalized discrepancy at a fixed pH value was almost irrelevant to the electrostatic properties among the studied SLN groups. On the other hand, for a fixed pH value in pH e 7, stronger HBMEC electricity by the uptake of SLN yielded a greater normalized discrepancy. The reason of this result might be related to the surface softness, that is, the thicknesses of surfactant zone and cell membrane.41 The sizes of phosphate group and choline group were ca. 1.8 and 2.0 nm, respectively,27 and the thickness of HBMEC membrane was ca. 30 nm.42 Thick surface layer on the external region of a biocolloid provided a charge buffer for a relatively structured

double layer, yielding an ordered normalized discrepancy according to the charge characteristics. The normalized discrepancy of HBMEC system was smaller than that of SLN, indicating a stronger charge-buffer function of the HBMEC membrane phase. Besides, the ordered normalized discrepancy in the range of pH e 7 for HBMEC system might be derived mainly from the sensitivity of HBMEC to H+. For fresh SLN, the normalized discrepancy was nearly invariant to pH value at pH g 9, and for stored SLN, the normalized discrepancy was nearly invariant to pH value at pH g 11. Also, the normalized discrepancy was almost invariant to pH value at pH g 10-11 for HBMEC system. The main reason for this observation was that the negative charge on the biocolloidal surfaces was saturated at high pH values, although the double-layer thickness was strongly compressed by sodium and hydroxyl groups. Besides, adequate particulate interaction and PC hydrolysis during storage shifted the invariability of the normalized discrepancy to a higher pH value. 4. Conclusions In summary, negatively charged SLN with various ratios of cacao butter to stearic acid were prepared. ζ potential and electrophoretic mobility of SLN, HBMEC, and the combination of the above two biocolloids were experimentally determined. Surface charge density and surface potential were calculated from electrophoretic mobility via theoretical models. For the SLN system, a higher content of cacao butter or a slower stirring rate yielded a larger SLN. A larger freshly prepared SLN led to a higher density of the ionogenic groups on particles and larger surface electricity. Over storage beyond 3 months at low pH value, obvious aggregation of SLN and hydrolysis of PC occurred, yielding a decrease in the charge amount on SLN. Furthermore, an increase in pH value of the medium caused an increase in the negative charge of both SLN and HBMEC systems. For SLN system, no charge reversal at low pH value was observed; on the contrary, charge reversal at pH e 3 for the HBMEC system occurred, leading to cell death. In addition, uptake of SLN caused a decrease in the HBMEC electricity, and uptake of SLN with higher content of cacao butter resulted in a more substantial decrease in the HBMEC electricity. For the deviation between surface potential and ζ potential, the medium of pH ) 7 caused the largest normalized discrepancy between the two potentials because of a thick double layer. For a fixed particulate species, high pH led to an almost invariant normalized discrepancy because of surface charge saturation.

11236 J. Phys. Chem. B, Vol. 111, No. 38, 2007 Since HBMEC was sensitive to H+, the order of normalized discrepancy at a fixed pH followed the order of electrical intensity of the HBMEC system at pH e 7. Acknowledgment. This work is supported by the National Science Council of the Republic of China. References and Notes (1) Glynn, J. R.; Belongia, B. M.; Arnold, R. G.; Ogden, K. L.; Baygents, J. C. Appl. EnViron. Microbiol. 1998, 64, 2572. (2) Tagliaro, F.; Turrina, S.; Smith, F. P. Forensic Sci. Int. 1996, 77, 211. (3) Vergnon, A. L.; Chu, Y. H. Methods 1999, 19, 270. (4) Li, Q.; Yeung, E. W. Appl. Spectrosc. 1995, 49, 1528. (5) Kawahata, S.; Ohshima, H.; Muramatsu, N.; Kondo, T. J. Colloid Interface Sci. 1990, 138, 182. (6) Zhang, P. C.; Keleshian, A. M.; Sachs, F. Nature 2001, 413, 428. (7) Van Damme, M. P. I.; Tiglias, J.; Nemat, N.; Preston, B. N. Anal. Biochem. 1994, 223, 62. (8) Kuo, Y. C.; Lin, T. W. J. Phys. Chem. B 2006, 110, 2202. (9) Lin, D. Q.; Brixius, P. J.; Hubbuch, J. J.; Thommes, J.; Kula, M. R. Biotechnol. Bioeng. 2003, 83, 149. (10) Kuo, Y. C.; Chung, C. Y. J. Chin. Inst. Chem. Eng. 2005, 36, 627. (11) Sipos, I.; Torocsik, B.; Tretter, L.; Adam-Vizi, V. Cell. Mol. Neurobiol. 2005, 25, 141. (12) Lo, E. H.; Singhal, A. B.; Torchilin, V. P.; Abbottd, N. J. Brain Res. ReV. 2001, 38, 140. (13) Westesen, K.; Bunjes, H.; Koch, M. H. J. J. Controlled Release 1997, 48, 223. (14) Kuo, Y. C.; Su, F. L. Int. J. Pharm. 2007, 340, 143. (15) Zimmernann, E.; Muller, R. H. Eur. J. Pharm. Biopharm. 2001, 52, 203. (16) Freitas, C.; Muller, R. H. J. Microencapsulation 1999, 16, 59. (17) Mullar, R. H.; Massen, S.; Schwarz, C.; Mehnert, W. J. Controlled Release 1997, 47, 261. (18) Miglietta, A.; Cavalli, R.; Bocca, C.; Gabriel, L.; Gasco, M. R. Int. J. Pharm. 2000, 210, 61. (19) Cavalli, R.; Zara, G. P.; Caputo, O.; Bargoni, A.; Fundaro, A.; Gasco, M. R. Pharmacol. Res. 2000, 42, 541. (20) Levin, Y.; Barbosa, M. C.; Tamashiro, M. N. Europhys. Lett. 1998, 41, 123.

Kuo and Chen (21) Hsu, J. P.; Kuo, Y. C. J. Colloid Interface Sci. 1995, 170, 220. (22) Li, S. F. Y. Capillary Electrophoresis: Principles, Practice and Applications; Elsevier: Amsterdam, 1992. (23) Fernan´dez-Nieves, A.; Fernan´dez-Barbero, A.; de las Nieves, F. J. Langmuir 2000, 16, 4090. (24) Fillery-Travis, A. J.; Foster, L. H.; Robins, M. M. Biophys. Chem. 1995, 54, 253. (25) Schubert, M. A.; Mu¨ller-Goymann, C. C. Eur. J. Pharm. Biopharm. 2005, 61, 77. (26) Huang, S. W.; Hsu, J. P.; Kuo, Y. C.; Tseng, S. J. J. Phys. Chem. B 2002, 106, 2117. (27) Kotyn´ska, J.; Figaszewski, Z. A. Biochim. Biophys. Acta 2005, 1720, 22. (28) Radko, S. P.; Stastna, M.; Chrambach, A. Anal. Chem. 2000, 72, 5955. (29) Quintanar-Guerrero, D.; Tamayo-Esquivel, D.; Ganem-Quintanar, A.; Alle´mann, E.; Doelker, E. Eur. J. Pharm. Sci. 2005, 26, 211. (30) Kuo, Y. C. J. Phys. Chem. B 2005, 109, 11727. (31) Freitas, C.; Muller, R. H. Int. J. Pharm. 1998, 168, 221. (32) Freitas, C.; Muller, R. H. Eur. J. Pharm. Biopharm. 1999, 47, 125. (33) Herman, C. J.; Groves, M. J. J. Pharm. Pharmacol. 1992, 44, 539. (34) Grit, M.; Underberg, W. J. M.; Crommelin, D. J. A. J. Pharm. Sci. 1993, 82, 362. (35) Fenart, L.; Casanova, A.; Dehouck, B.; Duhem, C.; Slupek, S.; Cecchelli, R.; Betbeder, D. J. Pharmacol. Exp. Ther. 1999, 291, 1017. (36) Kuo, Y. C.; Chen, H. H. Int. J. Pharm. 2006, 327, 160. (37) Craig, L. E.; Spelman, J. P.; Strandberg, J. D.; Zink, M. C. MicroVasc. Res. 1998, 55, 65. (38) Dittrich, M.; Sibler, S. J. Colloid Interface Sci. 2005, 286, 487. (39) Ahimou, F.; Paquot, M.; Jacques, P.; Thonart, P.; Rouxhet, P. G. J. Microbiol. Methods 2001, 45, 119. (40) Lin, D.-Q.; Brixius, P. J.; Hubbuch, J. J.; Thommes, J.; Kula, M.R. Biotechnol. Bioeng. 2003, 83, 149. (41) Ohshima, H. J. Colloid Interface Sci. 1994, 163, 474. (42) Adamson, R. H.; Clough, G. J. Physiol. 1992, 445, 473. (43) Hsu, J. P.; Kuo, Y. C. J. Chem. Soc., Faraday Trans. 1995, 91, 4093. (44) Kuo, Y. C. J. Chem. Phys. 2003, 118, 398. (45) Kuo, Y. C. Langmuir 2003, 19, 5942. (46) Huang, F.; Subbaiah, P. V.; Holian, O.; Zhang, J.; Johnson, A.; Gertzberg, N.; Lum, H. Am. J. Physiol. Lung Cell Mol. Physiol. 2005, 289, L176. (47) Abbott, N. J.; Romero, I. A. Mol. Med. Today 1996, 2, 106.