Study of the influence of ascorbyl palmitate and amiodarone in the stability of unilamellar liposomes

Abstract

Amiodarone (AMI) is a low water-solubility drug, which is very useful in the treatment of severe cardiac disease. Its adverse effects are associated with toxicity in different tissues. Several antioxidants have been shown to reduce, and prevent AMI toxicity. The aim of this work was to develop and characterize Dimyristoylphosphatidylcholine (DMPC) liposomal carriers doped with ascorbyl palmitate (Asc₁₆) as antioxidant, in order to either minimize or avoid the adverse effects produced by AMI. The employment of liposomes would avoid the use of cosolvents in AMI formulations, and Asc₁₆ could minimize the adverse effects of AMI. To evaluate the partition and integration of AMI and Asc₁₆ in lipid membranes, penetration studies into DMPC monolayers were carried out. The disturbance of the liposomes membranes was studied by generalized polarization (GP). The stability of liposomes was evaluated experimentally and by means of the Derjaguin-Landau-Verwey-Overbeek (DLVO) theory. The size particle and zeta potential (ζ) values of the liposomes were used for application in calculations for attractive and repulsive forces in DLVO theory. In experimental conditions all of these vesicles showed stability at time 0, but only DMPC + Asc₁₆ 10% + AMI 10% liposomes kept their size stable and ζ during 28 days. These results are encouraging and suggest that such systems could be suitable for AMI delivery formulations.

• Amiodarone
• ascorbyl palmitate
• DLVO theory
• generalized polarization
• Gibbs monolayers
• liposomes

Introduction

Amiodarone (2-butylbenzofuran-3-yl 4-(2-diethylaminoethoxy)-3,5-di-iodophenyl ketone) (AMI) (M.M: 645.3) is a cationic amphiphilic molecule at neutral and acidic pHs (pKa ∼6.6 at 25 °C) (Boury et al., 2001) (Figure 1a: AMI molecule). It is a class III antiarrhythmic drug and is used in the control of ventricular and supraventricular arrhythmias, including those associated with Wolff-Parkinson-White (WPW) syndrome. AMI is usually administrated through intravenous infusion (Sweetman, 2009, p. 1211).

Figure 1. Chemical structure of (a) amiodarone, and (b) ascorbyl palmitate.

Adverse cardiovascular effects and others are common with AMI. Many of them are dose-related and reversible with reduction in doses (Sweetman, 2009, p. 1211). It was suggested in reference (Sweetman, 2009, p. 1212), that amiodarone-induced phospholipidosis (PLD) and this fact may explain some of these adverse effects. The PLD is a storage disorder characterized by the excessive accumulation of phospholipids in tissues (Anderson & Borlak, 2006). The mechanism of the PLD is not entirely clear. However, it was suggested (Vereckei et al., 1993) that oxidative stress, a process that generates free radicals, is involved in the pathogenesis produced by AMI. Moreover, external factors, such as light, can exacerbate the AMI side-effects. Photosensitivity reactions, produced by light exposure in presence of the drug, are also common. This is a skin reaction produced by accumulation of AMI and light action that generates a reactive oxygen species and consequently develops the peroxidation of lipid chains in cell membranes and other intracellular structures (Stratton & Liebler, 1997). It has been reported in literature (Kachel et al., 1990) that photoprotective agents such as beta-carotene and vitamin E may reduce reactive oxygen species and thereby reduce the AMI phototoxicity. Kachel et al. (1990) reported that vitamin E reduces pulmonary cytotoxicity produced by AMI. It was also reported that vitamin E and a sylimarina, a type of flavonoid, decrease PLD induced by AMI (Agoston et al., 2003). Other authors (Ray et al., 2000), reported that proanthocyanidins from grape seeds are antioxidants potentially used in the treatment of PLD induced by AMI. The use of antioxidants is theoretically a form to either repair or minimize photosensitivity and other reactions produced by PLD.

Ascorbyl palmitate (Asc₁₆) (Figure 1b) is a good antioxidant in model cell systems (Lo Nostro, 1997). It is known that Asc₁₆ retains the same radical-scavenging properties of ascorbic acid and their antioxidant efficiency is comparable to that of other natural reducing agents, such as carotenes, polyphenols, and tocopherols (Sweetman, 2009, p. 1629). Their amphiphilic nature allows these vitamin C derivatives to form aggregates that provide an ideal environment for the solubilization of hydrophobic and sensitive drugs that might otherwise be easily degraded and oxidized when exposed to light, heat, dissolved oxygen, and other radical-producing species (Lo Nostro et al., 2003, Benedini et al., 2011b). The antioxidant characteristics of Asc₁₆ mean that it could be used to reduce or avoid AMI side-effects.

Due to the low aqueous solubility of amiodarone to develop an injectable formulation, surfactants and cosolvents addition are required. Then, marketed formulations are compounded with amiodarone hydrochloride (50 mg/ml), polysorbate 80 (Tween 80, 100 mg/ml) and benzyl alcohol (20 mg/ml) generally in 3 ml ampoules (Benedini et al., 2010). In addition, the excipients make amiodarone solubilization in aqueous solution possible, and can also contribute to the appearance of other kinds of side-effects such as negative inotropy and hypotension (Gough et al., 1982; Yasaka et al., 1979).

The AMI adverse effects and the use of cosolvents to solubilizate it, lead to unwavering search for drug carriers systems less harmful to humans. Systems that fulfill these principles are liposomes, given that these colloidal carriers are formed by phospholipids and other biocompatible components that do not produce adverse effects. Liposomes can be made entirely from natural substances and therefore are non-toxic, biodegradable and non-immunogenic (Lasic, 1998). A binary phospholipid-Asc₁₆ liposome system is proposed here as suitable to be used as a stable drug carrier where both components coexist with a reasonably favorable interaction. This behavior is supported by previous studies where Asc₁₆ show a great tendency to integrate into phospholipid monolayers adding stability to these supramolecular structures (Mottola et al., 2013). The aim of this work was to develop liposomes formed by phospholipids and Asc₁₆ with characteristics assessed by extrapolation from monolayer adsorption/penetration studies. AMI favorable incorporation to the binary phospholipid-Asc₁₆ membranes was evidenced in monolayer penetration studies. Both phospholipid-Asc₁₆ and phospholipid-Asc₁₆-AMI liposome systems have been studied by generalized polarization (GP). Liposome dispersions are not stable thermodynamically (Lasic, 1998), but measurements of stability were performed and, during 28 days, the liposomes containing DMPC + Asc₁₆ 10% + AMI 10% showed stability in either their size and z potential. This can be attributed to an increase of negative charge induced by Asc₁₆ and a structured effect induced by AMI to the ternary liposomes that keep them in a stable dispersion and decrease their aggregation. Furthermore, it is important to remark that this behavior is in accord with the predictions performing by DLVO theory. This theory explains the stability of lyophobic colloidal systems given that two independent types of forces govern the interaction between similar colloidal particles immersed in polar (especially aqueous) solutions: Attractive van der Waals forces and repulsive electrostatic forces (Derjaguin & Landau, 1941; Verwey & Overbeek, 1948). Here DLVO theory has been extended by introducing the hydrated radius of the ions adsorbed onto the liposome surfaces. Liposomes size, size distribution and zeta potentials (ζ) data were applied in DLVO theory.

Experimental section

Materials

Ascorbyl palmitate (Asc₁₆) was purchased from Fluka (Italy) (MW: 414.53), Amiodarone (M.M: 645.3) was from Parafam (Argentine). Dimyristoylphosphatidylcholine (DMPC) was provided by Avanti Polar Lipids Inc. (Alabaster, AL) and Laurdan was from Molecular Probes (Eugene, OR). All the reagents were of analytical grade (99% pure) and used without further purification. Solvents were of the highest commercial purity available. The water was purified by using a Milli-Q (Millipore, Billerica, MA) system, giving a product with a resistivity of ∼18.5 MΩ/cm.

Methods

Adsorption of AMI to bare air/water interface

After the injection of the drug (75 μl of AMI solution) in ethanol to the aqueous subphase of a Teflon trough under continuous stirring (trough volume: 15 ml; final subphase concentration unless indicated was 15.5 μM of AMI), the changes in surface pressure and surface potential at constant area were registered as a function of time, while a Gibbs monolayer was established. A control with ethanol alone was used to check that surface pressure was not affected. Surface pressure was determined by a Pt plate using the Wilhelmy method, and the Volta surface potential (ΔV) was monitored using an air-ionizing ²⁴¹Am surface electrode placed ∼5 mm above the interface and a Ag/AgCl/Cl^- (3M) reference electrode dipped into the subphase. The equipment used was a homemade circular Langmuir balance controlled by an electronic unit (Monofilmetter). All experiments were performed at 23 ± 2 °C.

Penetration of AMI to phospholipid monolayers

Penetration time curves were performed by deposition of a chloroformic solution of lipids at the air/water interface, in order to form a phospholipid monolayer previous to AMI injection in the subphase. The phospholipid monolayers were initially pure DMPC or DMPC/Asc₁₆ (80:20). The aqueous solutions used as subphase was: Saline solution pH ∼5.5 (NaCl 145 mM).

Vesicle preparation

DMPC, Asc₁₆, Laurdan and AMI (when accounted) were placed in appropriate amounts in glass tubes (all reagents were dissolved in chloroform-methanol, 1:1). These solutions were dried under N₂, to achieve a uniform film. It was kept under N₂ for about 1 h. Then, 300 μl of buffer (10 mM Hepes, NaCl y MgCl₂ a pH 7.4) were added to get a 1:2 (lipid: buffer, μg:μl) ratio. Solutions were heated by 5 min at a temperature above lipid transition and next each tube was vortexed 1 min, yielding multilamellar liposomes (MLV). To standardize the size of vesicles, the tubes were sonicated 30 min. Finally, 700 μl of buffer were added and MLV were filtered to form unilamellar liposomes with an extrusion syringe (15 times).

DMPC + Asc₁₆ liposomes and DMPC + Asc₁₆ + AMI liposomes

Binary DMPC-Asc₁₆ liposomes, and ternary DMPC-Asc₁₆-AMI liposomes were made. The concentrations and proportions of reagents used in preparing the different liposomes are presented in Table 1. Liposomes with Laurdan were prepared to determine GP and without the probe to ζ and size particle measurements.

Table 1. Composition of the tested systems.

Liposomes	DMPC (μmol)	DMPC (μg)	Asc16 (μmol)	Asc16 (μg)	AMI (μmol)	Laurdan (μmol)
DMPC (control)	0.22	150	–	–	–	–
DMPC + Asc16 10%	0.22	150	0.022	9.12	–	–
DMPC + Asc16 20%	0.22	150	0.044	18.24	–	–
DMPC + Asc16 30%	0.22	150	0.066	27.36	–	–
DMPC + Laurdan (0.6 μM)	0.22	150	–	–	–	0.6 × 10^−3
DMPC + Asc16 10% + Laurdan (0.6 μM)	0.22	150	0.022	9.12	–	0.6 × 10^−3
DMPC + Asc16 20% + Laurdan (0.6 μM)	0.22	150	0.044	18.24	–	0.6 × 10^−3
DMPC + Asc16 30% + Laurdan (0.6 μM)	0.22	150	0.066	27.36	–	0.6 × 10^−3
DMPC + Asc16 10% + AMI 10%	0.22	150	0.022	9.12	0.022	0.6 × 10^−3
DMPC + Asc16 10% + AMI 10% + Laurdan (0.6 μM)	0.22	150	0.022	9.12	0.022	0.6 × 10^−3
DMPC + Asc16 10% + AMI 20%	0.22	150	0.022	9.12	0.044	0.6 × 10^−3
DMPC + Asc16 10% + AMI 20% + Laurdan (0.6 μM)	0.22	150	0.022	9.12	0.044	0.6 × 10^−3

AMI, Amiodarone; DMPC, Dimyristoylphosphatidylcholine; Asc₁₆, ascorbyl palmitate.

Generalized polarization (GP)

Static state fluorescence spectroscopy measurements were carried out using a spectrofluorimeter SLM 4800 (SLM Instruments, Urbana, IL) and quartz cells 5 × 5 mm. The temperature was controlled by a circulating bath of thermostated water (Haake, Darmstadt, Germany). The temperature range was between 5 and 50 °C. The samples were excited at a λ = 360 nm, and spectra were collected in the range from 390–500 nm. From excitation spectra generalized polarization values (GP) were computed according to the following: where I₄₃₄ and I₄₉₀ correspond to fluorescence intensity values of the probe Laurdan to the emission wavelengths of 434 nm and 490 nm. These wavelengths are characteristic of gel and liquid-crystalline phases, respectively (Parasassi et al., 1991). Transition temperatures (T_t) were obtained as the zero point in the second derivative (inflection point) of the experimental GP vs. temperature curves.

Size particle and zeta potential (ζ)

The particle size, the polydispersity index (PDI) and ζ of the liposomes were determined using a Zeta sizer Nano ZS90 (Malvern Instruments, Malvern, UK). PDI is a parameter that defines the size particle distribution, more strictly, it is a dimensionless measure of the amplitude of the particle size (International Organization for Standardization [ISO] 13321, 1996). It ranges from 0–1 and is calculated according to ISO 13321 standards (ISO, 1996, Zielinska et al., 2008). PDI values between 0.1 and 0.2 indicate that the samples are monodisperse (Tantra et al., 2010).

Stability prediction by DLVO

Virtually no heterogeneous dispersed system is completely stable. Colloidal particles system follows a continuous Brownian motion and therefore particles continuously collide with each other. In these collisions the particles can be repelled or associated in clusters and eventually attain coalescence and phase separation.

The surface charge of the particles, besides keeping them in suspension, modifies ions distribution in aqueous solution. Ions with the same charge are repelled by the surface of the particle and others with opposite charge are attracted to form a double layer.

The liposome dispersions have a high surface energy associated with the large surface area. They are thermodynamically unstable and tend naturally to cluster or merge in order to reduce its area and surface free energy.

In most of colloidal systems, liposomes in this case, the interaction forces are modeled from the DLVO theory. Details of this theory applied to liposomes are given in references (Derjaguin & Landau, 1941; Verwey & Overbeek, 1948) and the Supplementary Material, available online.

Stability measurements

In this section, ζ and size particle were measured as detailed in the point 2.2.5 during 28 days to carry out the stability of liposomes of: DMPC, DMPC + Asc₁₆ 10%, DMPC + Asc₁₆ 10% + AMI 10%, and DMPC Asc₁₆ 10% + AMI 20%. The samples were kept at 4 °C between measurements. All the measures were performed in triplicate. The stability of the liposomes of DMPC and DMPC + Asc₁₆ 10% was evaluated as a control.

Results and discussion

Adsorption on and penetration of AMI into phospholipid interfaces

We studied AMI adsorption to bare air/water interface to determinate its surface properties. When a concentrated solution of a moderately soluble amphiphile is dissolved in the bulk of an aqueous solution, the adsorption of these amphiphilic molecules to the air/water interface results in a Gibbs monolayer, evidenced by a decrease of surface tension (increase of surface pressure) and changes in surface potential (ΔV) (Figure 2a). As opposed to Langmuir monolayers (Gaines, 1966), Gibbs monolayers reflect the equilibrium between the bulk monomeric form of the amphiphile and the organized film at the interface. Figure 2(a) shows that AMI auto-organizes forming a Gibbs monolayer when diluted in the aqueous subphase, reaching surface pressure values of 34 ± 2 mN/m above a subphase concentration of 15 μM (see insert in Figure 2a). For a regular amphiphile the adsorption kinetic curve can be described by a hyperbolic function (Maget-Dana, 1999; Vollhardt & Fainerman, 2010; Yaminsky et al., 1996). In this case, we observed a moderate overpressure phase and a slow recovery to equilibrium values (Figure 2a). The auto-organization of the cationic amphiphile at the interface should result in a positively charged surface and the establishment of a double layer of ions, which repeals protons increasing the surface pH. Therefore, partial neutralization of the AMI molecules at the interface may occur (pKa 6.6) explaining the moderate dropping in the surface pressure values observed. A similar effect was previously described for Asc₁₆ adsorption to air/water interface (Mottola et al., 2013). The increase in surface potential was also registered as a function of time showing a fast increase and stabilization to high positive values (600 ± 20 mV, Figure 2a see also inset). This high value (much higher than the 50 ± 15 mV measured for Asc₁₆ monolayers and also than the 400 ± 20 mV observed for DMPC monolayers [Mottola et al., 2013]) may reflect the positive charged nature of the film besides other properties (the dipole moment of the AMI molecules, the induced dipole orientation of water molecules at the sub-interface and the double layer potential (Benedini et al., 2011a; Maget-Dana, 1999).

Figure 2. Surface pressure-time of Gibbs monolayers and penetration of AMI to phospholipid or phospholipid/Asc₁₆ monolayers. (a) Gibbs monolayer of AMI formed by adsorption to a bare air/saline solution surface and (b) penetration of AMI to DMPC monolayer (full lines) or DMPC/Asc₁₆ (80:20) (dashed lines) at ∼30–32 mN/m. The black and gray lines and symbols in (a) and (b) correspond to surface pressure and surface potential time curves, respectively, after injection of AMI in the subphase. Final AMI subphase concentration 15.5 μM. The inset in (a), shows the final surface pressure and potential dependence with the subphase AMI concentration. The inset in (b) shows cut-off curves of AMI penetration to DMPC (black dots and black line) or DMPC/Asc₁₆ (80:20) (grey dots and grey line). The arrow in (b) indicates the time of AMI injection in the subphase. The curves shown are chosen from a set of duplicates that differ in less than 2 mN/m.

In order to explore the interaction of AMI with biomembranes, we performed studies of penetration of AMI to phospholipid monolayers. The zwitterionic lipid dimiristoyl-phosphatidylcholine (DMPC) organized in Langmuir monolayer was used as biomembrane model. At the experimental temperature DMPC monolayers have the ability to organize in a liquid-expanded phase at all surface pressures until collapse (Antunes-Madeira et al., 1995). AMI showed a good ability to penetrate DMPC monolayers (Figure 2b). It induced an increase in surface pressure reaching values of 42 ± 2 mN/m (when initial DMPC pressure is 30 ± 2 mN/m). This implies an interaction pressure of 8 ± 2 mN/m. The surface potential registries of the penetration experiments showed a moderate increase with AMI insertion, reaching 500 ± 20 mV.

To assess the possibility of using binary DMPC/Asc₁₆ liposomes as drug carriers, we further extended our studies to the penetration of AMI into DMPC/Asc₁₆ mixed monolayers. Former studies showed that the introduction of 20 mol% of Asc₁₆ to DMPC monolayers induces a phase transition (from liquid-expanded to liquid-condensed) at ∼35 mN/m with the occurrence of condensed domains in the micrometer scale that corresponds to a low area fraction (≤20%). However, the transition does not lead to a complete condensed film and the two phases merges at ∼45 mN/m to a new phase with intermediate (expanded) properties (Mottola et al., 2013). From an examination of Figure 2(b) we can conclude that the presence of 20 mol% Asc₁₆ into the DMPC monolayer slightly affect AMI penetration, showing an interaction pressure of 9 ± 2 mN/m (at an initial surface pressure of 30 mN/m). However, when we analyzed the increase in surface pressure (Δπ) induced by AMI penetration at different initial surface pressures (see inset Figure 2b) we found a favorable insertion to DMPC/Asc₁₆ monolayers at low surface molecular packing. This difference was enhanced when we analyzed the intercept to the Δπ axis. This parameter is called Δπ₀ and gave values of 36 mN/m (r²= 0.95) for pure DMPC monolayers and 44 mN/m (r²= 0.99) when containing 20 mol % of Asc₁₆. This difference may reflect favorable electrostatic interaction between the anionic Asc₁₆ and the cationic AMI amphiphiles. On the other hand, both monolayers showed similar exclusion pressures (43–45 mN/m). This implies that AMI can penetrate phospholipid membranes to surface pressures above the pressures expected for lipid bilayers (∼30 mN/m).

Biophysical state of membranes of DMPC liposomes containing Asc₁₆ by generalized polarization

The amphiphilic fluorescent probe (Laurdan) adsorbs to the polar region of a lipid interface and shows its solvatochromic properties (Parasassi et al., 1991). The generalized polarization (GP) is a parameter that accounts for the change in its spectroscopic properties. This value is dependent on the amount of water in the membrane, which in turn is directly related to the polarity of the surface, and indirectly to the order thereof. The structural disturbance of the bilayer of pure DMPC in presence of Asc₁₆ was determined by calculating GP of the Laurdan probe at different temperatures and is plotted in Figure 3(a).

Figure 3. Changes of the generalized polarization (GP) parameter of the fluorescent probe of Laurdan depending on temperature of unilamellar liposomes from: (a) Pure DMPC (▪), DMPC + Asc₁₆ 10% (•), DMPC + Asc₁₆ 20% (▴) and DMPC + Asc₁₆ 30% (▾). (b) Pure DMPC (▪), DMPC + Asc₁₆ 10% + AMI 10% (▪) and DMPC + Asc₁₆ 10% + AMI 20% (▪). The arrow indicates the increase of the membrane structuration by addition of increasing amounts of AMI.

The curves in Figure 3(a) showed a drop of GP when the sample is heated above ∼20 °C indicating the beginning of a phase transition for pure DMPC liposomes. The introduction of a second component would reduce GP values if altering the gel structure of DMPC and/or increase them when affecting the liquid-crystaline phase. As it can be deduced from Figure 3(a), the presence of Asc₁₆ did not affect significantly the GP curve and therefore, it did not disturb the DMPC gel or liquid-crystalline phases. From the inflection point in the GP vs. temperature curves, the transition temperature (T_t) was obtained for both pure DMPC and with increasing concentrations of Asc₁₆. These values are reported in Table 2.

Table 2. Transition temperatures (T_t) vs. generalized polarization (GP) of liposomes.

DMPC liposomes	Tt (°C)
Control	25.1 ± 0.8
+Asc16 10%	26.1 ± 1.7
+Asc16 20%	27.4 ± 1.9
+Asc16 30%	25.9 ± 1.3
+Asc16 10% + AMI 10%	27.1 ± 1.1
+Asc16 10% + AMI 20%	24.7 ± 2

AMI, Amiodarone; DMPC, Dimyristoylphosphatidylcholine; Asc₁₆, ascorbyl palmitate.

The T_t calculated for liposomes formed by DMPC and the three different ratios of Asc₁₆ (10, 20 and 30%) did not differ from the T_t value of pure DMPC lipid bilayer that was used as control.

Biophysical state of membranes of DMPC liposomes containing Asc₁₆ and AMI by generalized polarization

Figure 3(b) shows GP vs. T curves for DMPC with Asc₁₆ and AMI systems in the two tested proportions. A separation of the curves from the phase transition zone was observed, which becomes more evident towards the disordered phase (indicated by an arrow in Figure 3b). However, there were not significant shifts of the phase transition temperature in respect to those of pure DMPC (T_t ∼ 23 °C) (see Table 2). This separation of the curves towards higher GP indicates a non-homogeneous, uncooperative system and that the presence of AMI caused a small “structuring effect” on the disordered phase. Similar effect was reported in reference (Antunes-Madeira et al., 1995), where DMPC liposomes containing AMI were studied and its effect was evaluated by fluorescence polarization. Both in reference (Antunes-Madeira et al., 1995) and in this work, an increase in the content of AMI in the liposome affects only the fluid phase of the bilayer but not the gel phase. This result may indicate that AMI is insoluble in the gel phase and segregates in small islands but dissolves in the fluid phase thus disturbing its structure. It is worth noting that the presence of Asc₁₆ results in a similar effect that was reported for binary phospholipids + AMI liposomes. This indicates that this antioxidant does not generate a significant perturbation of the membrane structure of the liposome.

Zeta potential (ζ) and particle size measurements for DMPC liposomes containing Asc₁₆ or Asc₁₆ and AMI

We further characterized some physical properties of binary and ternary liposomes. Table 3 shows ζ and the average size (diameter) of particles with their respective standard deviations. In all cases we found a single peak in the particle size distribution (monomodal distribution) indicating a structural stability of the bilayer membranes. Liposomes fusion and/or fission would be evidenced by a polidisperse population. ζ reflects the net charge of the liposome plus its hydration shield and its behavior when exposed to an electric field. In most electrolyte solutions, as it is the case in our control liposomes, the ζ values for phosphatidylcholine (PC) becomes negative because of the adsorption of anions located among the zwitterionic dipoles of the polar head groups. The reported values for ζ of pure DMPC liposomes in solution span the −5 to −25 mV range (Alonso et al., 1995), depending on the ionic strength of the medium, in agreement with our findings.

Table 3. Zeta potential (ζ) and particle size of the tested systems are shown. All results were taken at day “0”. The experiments were performed in triplicate at 28 °C.

DMPC liposomes	ζ (mV)	Average particle size (nm)	PDI
Control	−5.94 ± 0.1	178 ± 3	0.1 ± 0.02
+Asc16 10%	−10.6 ± 0.2	195 ± 4	0.12 ± 0.05
+Asc16 20%	−14.8 ± 0.3	159.1 ± 15	0.1 ± 0.02
+Asc16 30%	−16.5 ± 0.5	182.5 ± 39	0.13 ± 0.03
+Asc16 10% + AMI 10%	−8.35 ± 0.1	185.6 ± 7	0.1 ± 0.06
+Asc16 10% + AMI 20%	−5.33 ± 0.2	191.3 ± 8	0.18 ± 0.05

PDI, polydispersity index; AMI, Amiodarone; DMPC, Dimyristoylphosphatidylcholine; Asc₁₆, ascorbyl palmitate.

Increasing the Asc₁₆ content in liposomes did not cause a significant change in their size and conducted to more negative ζ values. The results shown in Table 3 provide a direct evidence of Asc₁₆ incorporation to the membranes, since under the experimental pH (∼6) Asc₁₆ is negatively charged. This result agrees with the monolayer experiments that show a favorable integration of Asc₁₆ to phospholipid membranes as said above. At low concentrations Asc₁₆ samples develops lamellar structures, either gel or liquid crystal. Consequently, its packing parameter is also close to unity (Palma et al., 2002), and it was expected that its inclusion in the liposomes did not alter the basic structure. On the other hand, the provision of a phospholipid environment to the Asc₁₆ molecules allows maintaining the flexibility properties of the liquid-crystalline liposomes, a desirable property for drug carrier function.

Table 3 also shows ζ and size of the liposomes with the addition of increasing amounts of AMI. It should be noted that ζ shifts to lower negative values with increased proportions of AMI. This is because AMI is positively charged in the experimental pH (∼6). Therefore, it is expected that if the AMI proportion is increased, the ζ potential must become less negative and consequently can destabilize the system. As can be seen in Table 3, the average size of the liposomes is relatively stable with less than 0.2 PDI values indicating a monodisperse population (Tantra et al., 2010).

Stability study of liposomes by DLVO theory

As shown in the previous paragraph, the liposome charge neutralization by the addition of AMI may affect the stability of the liposome suspension by reducing electrostatic repulsion between vesicles and promoting their fusion. To analyze this possibility we used the DLVO theory for the liposomes in study. Equation 6.3 in the Supplementary Material (available online) was employed to determine V_R without correction for counterion hydration and the results are presented in Figure 1 in the online Supplementary Material. This Equation was used for pure DMPC systems and with addition of 10, 20 and 30% of Asc₁₆. The resulting V_DLVO potential is monotonically attractive so the liposomes should be unstable to aggregation. Since this does not occur, corrections not considered in the original DLVO theory are needed.

To take hydration forces into account, a new term was included in Equation 6.3 in the online Supplementary Material, obtaining Equation (2) (Sabín et al., 2006) . For this model, it was considered that there is a layer of hydrated counterions covering the surface of liposomes, as described in (Ohshima et al., 1982):

Δ being the radius of the adsorbed hydrated ions.

In the case of liposomes of phosphatidylcholine (PC), which generally have low surface charge (<50 mV), V_DLVO takes the following form (Sabin, 2007):

The results are represented in Figure 4, where a secondary minimum is observed in liposomes containing Asc₁₆. This study also evidences that the higher the Asc₁₆ content, the more pronounced liposome stability is predicted.

Figure 4. DMPC liposomes with Asc₁₆. DLVO potential (V_DLVO) vs. distance with considering the counterions’ hydration. Liposome compositions were as follows: DMPC + Asc₁₆ 30% (Δ), DMPC + Asc₁₆ 20% (▾), DMPC + Asc₁₆ 10% (○), and pure DMPC (•). The arrow indicates the secondary minimum.

There also are some considerations regarding to the Hamaker constant, A, used to carry out the computations. This constant describes the van de Waals interaction between two bodies. It is in the range of 1 × 10⁻²¹ – 1 × 10⁻²⁰J. In reference (Ohshima et al., 1982), in their study of the system dipalmitoyl-phosphatidylcholine (DPPC) in water Ohshima et al. (1982) estimated the Hamaker constant in (3.6 ± 0.8) × 10⁻²¹J.

We took the value of this constant as that of linear hydrocarbons having 16 or more carbons, since we have worked with Asc₁₆ (containing a hydrocarbon chain of 16 carbon atoms) and DMPC (hydrocarbon chains of 14 carbon atoms). The Hamaker constant has two contributions as described in Equation (4).

The zero frequency component of the Hamaker constant (A_ν = 0) is screened by the presence of electrolytes, while the non-zero-frequency component (A_ν > 0) is not affected by the presence of electrolytes but it is affected by the retardation effect. These components were taken as A_ν = 0 = −0.8 × 10⁻²¹J and A_ν > 0≈ 6 × 10⁻²¹J as shown in reference (Sabin, 2007, p. 190), in the absence of electrolyte A≈ 5.2 × 10⁻²¹J. This value is similar to hexadecane interacting with hexadecane through water, which was found experimentally, A₁₂₁ = (3–6) × 10⁻²¹J and reported in the literature (Israelachvili, 1991, pp. 196–197).

Retardation correction for A_ν > 0 was not considered, since it becomes significant at distances of about 5 nm or more and can be reduced to half when distance is 10 nm (Sabin, 2007, pp. 260–287), as shown in Supplementary Figure 2, available online, where A is plotted as a function of the distance between the particles. As shown in Figure 4, V_DLVO increased as we approach 1 nm, and a minimum appears at distances less than 4 nm. Therefore no correction for retardation was taken.

To compare the dependence of A with distance (x), it was assumed that between 5 and 10 nm, A decays linearly from 6 × 10⁻²¹J to 3 × 10⁻²¹. It should be noted that the decay of A actually has a more complex function with the distance than assumed here and there are ways to calculate it numerically solving the Lifshitz’s equation (Israelachvili, 1991, pp. 196–197). However, the influence of this reduction did not affect the overall results.

The effect of changes in concentration of additives on V_DLVO curves may be seen in Figure 4. The onset of the potential barrier is at a very short distance from the liposome surface (×≤4 nm); at this point retardation effects are not important, and then further refinement was not considered necessary.

It can be seen that the addition of Asc₁₆ increases negatively ζ and the stability of the vesicles, because of the incorporation of the negatively charged Asc₁₆. This charge is neutralized by the addition of AMI. In Figure 5, it is shown that the V_DLVO values decrease with the increase of the AMI proportion in the liposomes. Thus, at high concentrations of AMI in the liposomes ζ shifts to zero, and even if the hydration repulsive forces are considered in the calculation V_DLVO becomes less than k_BT when AMI concentration is above 10%, and thus liposomes must become unstable and susceptible to aggregation.

Figure 5. DLVO potential (V_DLVO) vs. distance between the particles. (▪): DMPC + Asc₁₆ 10%, (▽) DMPC + Asc₁₆ 10% + AMI 10% and (•) DMPC + Asc₁₆ 10% + AMI 20%. The arrow indicates the secondary minimum.

It has been determined the value of ζ (−8.29 mV) in which the maximum value of the potential barrier of the liposomes is 1k_BT, and therefore below this value the system could aggregate. This is shown in Supplementary Figure 3, available online. The proportion of AMI which gives this result in liposomes having 10% Asc₁₆ is 12.7%. This is shown in Supplementary Figure 4 (available online).

Since the studied liposomes were stable in all cases (even containing 20% AMI) at time 0, besides DLVO forces, there must be an additional process of stabilization, which is given by the hydrophilic character of the surface of vesicles. These repulsive hydration forces (Israelachvili, 1991, pp. 196–197) can be much stronger than the DLVO potential barrier, especially at very short distance between the approaching surfaces (distances less than 5–10 molecular diameters of the solvent, about 1–2 nm in water). These forces arise from the need to dehydrate the surfaces when the particles are close enough as to fall in the primary minimum. Since this process is energetically very costly, the system remains stable.

Experimental stability study of liposomes

To consider the possibility that liposomes doped with AMI could be applied in drug delivery systems, we studied their stability by measurement of the size particle (nm) and zeta potential (ζ) (mV), as a function of time (days). Figure 6(a) shows that pure DMPC liposomes modified their size in the first 5 days and kept the reached size thereafter. Liposome formed by DMPC + Asc₁₆ kept their size stability until day 11 and then the size increase was less than 18% of the original size. DMPC + Asc₁₆ 10% + AMI 10% showed a stable size during all of experimental time. Finally, for DMPC + Asc₁₆ 10% + AMI 20% liposomes, the size particle was stable until day 11, and then started to increase up to 50% more than the original size (∼200 nm). These results agree with the prediction made by DLVO modified theory. Related with the zeta potential analysis, the chart plotted in Figure 6(b) shows that the values of this parameter did not vary a large amount during the experimental time for all the samples. It is important to remark that DMPC + Asc₁₆ 10% + AMI 20% liposomes modified their size particle and ζ around the day 11, shifting this last parameter to more positive values. This result agrees with the previous results presented in the section entitled Stability study of liposomes by DLVO theory, because these liposomes were stable at time 0 but not during all the time of the experiments. Furthermore, using DLVO theory for liposomes of DMPC and 10% of Asc₁₆, the addition of 12.7% of AMI is enough to reach 1k_BT, which represent the limit of AMI amount allowed in these liposomes. The liposomes with 20% of AMI were stabilized initially by other forces, as hydration forces, that likely are more important at the beginning of the experiments than DLVO repulsive and attractive forces. However, these initial stabilization forces are not energetically favorable and then, as time passed, an increment in liposome size resulted.

Figure 6. Liposomes stability as a function of time (days) by means of (a) size particle (nm) and (b) zeta potential (ζ) (mV) Liposome compositions were as follows: (○) DMPC, (▪) DMPC + Asc₁₆ 10%, (▽) DMPC + Asc₁₆ 10% + 10% AMI, and (•) DMPC + Asc₁₆ 10% + AMI 20%.

Conclusion

Binary liposomes composed by phospholipid and Asc₁₆ behaves as a system suitable for drug delivery function. From monolayer experiments we conclude that AMI shows a good surface activity (adsorption to bare air/water surface reaching equilibrium surface pressures of ∼34 mN/m). AMI showed a moderate but favorable propensity to penetrate phospholipid monolayers (interaction pressure 8 mN/m) that is slightly improved by the presence of 20 mol% Asc₁₆. The incorporation of Asc₁₆ molecules to liposomes and the establishment of a negatively charged surface were confirmed by ζ measurement of binary liposomes, which evidenced an increase of a negative potential with Asc₁₆. Furthermore, the size of the binary liposomes showed monomodal distribution that suggests stability of the system. Liposome experiments support that AMI can be incorporated to these antioxidant liposomes when they are present in a liquid-disordered state: For all tested systems (DMPC + Asc₁₆, with and without AMI) GP values at temperatures below T_t were very similar in all studied conditions. However, above T_t, liposomes with AMI showed an increase in GP values depending on its concentration, indicating a structuring effect in the liquid crystalline phase of the system. Initially, the addition of AMI to liposomes formed by DMPC and Asc₁₆ did not change the size and stability of the liposomes. However, the incorporation of the cationic AMI molecules neutralizes the negative surface charge of the binary DMPC + Asc₁₆ liposomes. Due to the low electrostatic repulsion between vesicles, liposomes stability could be affected; but this has been justified by the application of the DLVO theory modified by hydration and correction of the Hamaker constant. Nevertheless, DMPC + Asc₁₆ 10% + AMI 10% liposomes were stable during 28 days, not only at time 0, keeping their size and zeta potential without important changes. In summary, the results shown in this work support the application of the ternary liposomes composed of phospholipids + Asc₁₆ + AMI as a drug delivery system. As far as we know, such a system, not previously reported in literature, appears attractive for use as a base in carrier systems for drugs that require an antioxidant activity to avoid or minimize AMI adverse effects. This is a first step in the development of these systems. A deeper study might be needed in order to focus future works to reach in vivo studies.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

This work was supported by grants from the Universidad Nacional del Sur (N°: 24/Q046), Universidad Nacional de Córdoba (N°: 05/C578), Consejo Nacional de Investigaciones Científicas y Técnicas de la República Argentina (CONICET) and FONCyT (Argentina) (N°: 11220100100072). SA, MLF, PVM, and SDP are Career Investigators of CONICET. LB has a fellowship of CONICET.