Catanionic vesicles from an amphiphilic prodrug molecule: a new concept for drug delivery systems

Abstract

Toxicity and low entrapment efficiency are the main problems for pharmaceutical applications of catanionic vesicles. In order to minimize surfactant toxicity, increase the drug loading content and simultaneously reduce the need for tedious chemical synthesis, we use oleic acid as the biocompatible surfactant, which reacts with the selected drug molecules (amlodipine) to produce an amphiphilic prodrug molecule for the straightforward fabrication of catanionic vesicles. The prodrug molecules are easily obtained by proton transfer between amlodipine and oleic acid molecules. The characterization of prodrug molecules and their aggregation behaviours in aqueous solutions are investigated by using Fourier transform infrared spectrophotometry (FTIR), ¹H-nuclear magnetic resonance (¹H-NMR), differential scanning calorimetry (DSC), surface tension measurement, transmission electron microscopy (TEM), dynamic light scattering (DLS), conductivity and zeta potential (ζ). The results demonstrate that vesicles could be easily formed with the prodrug amphiphilic molecules dispersed in aqueous solutions. Particularly, the drug release behaviour of the as-prepared catanionic vesicles exhibits excellent sustained drug release properties, which demonstrates their promising application in the newly designed drug delivery system.

---

Introduction

The term “vesicle” refers to spherical containers formed in aqueous solution by the self-assembly of amphiphilic molecules.¹ The core of the vesicle contains water while its shell is a bilayer of the amphiphilic molecules. In vesicle dispersions, the hydrophobic parts of the molecules are shielded from the aqueous solvent while the hydrophilic headgroups experience maximum contact with water.^2,3 Such unique vesicle structure has been achieved using ionic surfactants,⁴ non-ionic surfactants,^5,6 natural lipids,¹ block copolymers^7,8 or supramolecular amphiphiles.^9,10 Vesicles are of technological interest for potential application in various areas such as the cosmetic industry, nanostructured systems,^7,11 bioseparations and sensing,^12,13 and drug–gene delivery systems.^14,15 Especially in drug delivery systems, vesicles (liposomes) can simultaneously encapsulate hydrophilic drug molecules in the aqueous lumen of the vesicles and hydrophobic drugs in the hydrophobic membrane of the vesicles, which vastly improves the pharmacokinetics and biodistribution of drugs that suffer from poor solubility, poor stability and unwanted toxicity. However, conventional liposomes are rather difficult to prepare and generally have limited shelf life duration. A new class of self-assembled amphiphilic aggregates termed “catanionic” vesicles has been developed recently.⁴ Catanionic vesicles can be spontaneously formed upon mixing aqueous solutions of cationic and anionic surfactants. Such types of vesicular structures are formed mainly through the electrostatic association of oppositely-charged polar headgroups, which considerably decreases the effective headgroup area and benefits the vesicle formation according to the packing parameter theory.^16,17 Catanionic vesicles are analogous to liposomes but have several advantages over them. The major advantage lies in their ease of preparation, long-term stability, resultant ease of storage and relatively low cost, which correspond well to industrial requirements.¹⁸ Moreover, their size, charge, and permeability can be readily adjusted by varying the temperature, concentration, molar ratio and chain lengths of the two surfactants.¹⁹ Thus, catanionic vesicles provide potential promise for the design of a new drug delivery system. For example, Ghosh and Dey²⁰ prepared N-lauroylsarcosinate and N-cetylpyridinium chloride catanionic vesicles to entrap a water-soluble dye. In the presence of 10 mol% of cholesterol, the mixed surfactant vesicles exhibited leakage of the encapsulated calcein dye, showing potential application in pH-triggered drug release. Dew et al. incorporated catanionic aggregates in gel formulations.²¹ Prolonged drug release and reduced skin penetration rate were achieved, which indicated that the application of formulations containing catanionic vesicles was a future possibility. However, the previous application of catanionic vesicles for drug delivery systems concerned the loading of drugs into catanionic vesicles formed by two surfactants, which results in an uncertain low drug-loading content. Therefore, it is of great interest to find a straightforward drug delivery system for application in the biomedical area.

To date, toxicity and low entrapment efficiency are still the main problems regarding catanionic vesicles. This greatly limits the application of catanionic vesicles in drug delivery systems. To be effectively used for drug delivery, catanionic vesicles must have the essential features: biocompatible, biodegradable, non-toxic, non-immunogenic and non-carcinogenic. In order to minimize surfactant toxicity, increase the drug loading content and reduce the need for tedious chemical synthesis encountered by the conventional catanionic vesicles and pharmacosomes, herein we use oleic acid as the biocompatible surfactant which reacts with the selected drug molecules to result in an amphiphilic prodrug molecule for the straightforward fabrication of catanionic vesicles. The synthesized molecule is not only a prodrug with bioactivity but also an amphiphilic molecule useful for the formation of catanionic vesicles. This new kind of drug-participating catanionic vesicle is named “catanionic pharmacosome” according to the definition of “catanionic vesicle” and “pharmacosome.” The amphiphilic prodrug molecules are easily obtained by proton transfer between amlodipine and oleic acid molecules. To our knowledge, this is the first time that a catanionic pharmacosome has been obtained directly from an amphiphilic prodrug molecule synthesized from proton transfer between the drug and surfactant molecules. Moreover, directly using a drug as part of a carrier can substantially reduce the amount of inert carrier materials and largely increase the drug loading content.²² Various techniques such as Fourier transform infrared spectrophotometry (FTIR), ¹H-nuclear magnetic resonance (¹H-NMR), dynamic light scattering (DLS), surface tension measurement, transmission electron microscopy (TEM), conductivity and zeta potential (ζ) measurements are performed to characterize the formation of prodrug molecules and their aggregation behaviour in the aqueous solution. Specifically, the drug release behaviour of the as-prepared catanionic pharmacosome dispersions is studied. The results indicate that the as-prepared catanionic pharmacosomes have an excellent sustained drug release behaviour. The present study provides a new concept for drug delivery systems, which might open a new door for drug delivery applications.

Experimental section

Materials

Amlodipine was purchased from the Huameihua Technology Group Corporation, Wuhan, China. Oleic acid (Tianjin Damao Chemical Regent Factory) and dichloromethane (Sinopharm Chemical Reagent Co., Ltd, China) were of analytical grade and used without further purification. Double-distilled water was used in all experiments.

Synthesis of amphiphilic prodrug molecules and preparation of catanionic pharmacosomes

For the synthesis of prodrug molecules, equimolar amounts of oleic acid and amlodipine were dissolved in dichloromethane (Scheme 1). Subsequently, the mixture was allowed to react for 6 h under gentle stirring at room temperature. Finally, the mixture was evaporated to dryness at 40 °C under vacuum in a rotary evaporator. To fully remove the solvent residues, the product was further flushed for about 30 minutes with nitrogen and stored under vacuum overnight. The desirable amount of prodrug was dissolved in water with immediate sonication for 5 min at room temperature, after which, bluish catanionic pharmacosome dispersions were formed. All experiments were performed at neutral pH except the in vitro drug-release study (phosphate buffer solution: pH 1.2, pH 5.8, pH 7.4).

Scheme 1 Synthesis of amphiphilic prodrug molecules.

Methods.
Fourier transform infrared spectrophotometry (FTIR). FTIR spectroscopy (Bio-Rad Laboratories, Nicolet 6700 FTIR Spectrometer, USA) was used to analyze the formation of amlodipine prodrug molecules. FTIR spectra of oleic acid, amlodipine and the prodrug molecule were obtained with the KBr plate technique.
¹H-NMR measurements. The NMR data were recorded on a Bruker Avance 600 spectrometer operating at 600 MHz at room temperature. Amlodipine, oleic acid and amlodipine prodrug were freshly dissolved in 0.5 mL DMSO-d₆ in 5 mm diameter NMR tubes. Chemical shifts were reported in ppm with respect to tetramethylsilane.
Differential scanning calorimetry (DSC). DSC measurements were taken with a CDR-4P (Shanghai Precision and Scientific Instrument Co., Ltd. China). Weighed samples of 10 mg were placed in aluminum pans and the samples were scanned from 30 °C to 300 °C. The heating rate was 10 °C min⁻¹.
Surface tension measurement. The surface tension was measured with a Processor Tensionmeter-K12 (Krüss Co.) using the Wilhelmy-plate. The precise degree of the measurement is 0.01 mN m⁻¹. The concentration of dispersions changed from 3 × 10⁻⁶ mol L⁻¹ to 2 × 10⁻³ mol L⁻¹. All of the experiments were carried out at 25 °C.
Transmission electron microscopy (TEM). The morphologies of the aggregates were investigated by TEM (JEM-100CXII, Japan). The dispersions were adsorbed onto carbon-coated copper grids (200 mesh). Excess liquid was removed using a piece of filter paper. Then the dispersion was negatively stained with a drop of 2.0 wt% phosphotungstic acid solution for 10 seconds. The excess liquid was removed. The specimens were dried and observed using a transmission electron microscope at room temperature.
Size and size distribution. The size of the aggregates was determined using a DLS (Delsa™ Nano Submicron Particle Size Analyzer (A53878), Beckman Coulter Inc., USA). The samples were located in cuvettes at room temperature. The light-scattering cells were rinsed twice with acetone to ensure dust-free conditions before use. The data were analyzed using the CONTIN method.
Conductivity. Electrical conductivity measurements were performed with a DDS-11A conductivity meter using a glass electrode at room temperature. The distilled water conductivity was 3.0 μS cm⁻¹ at room temperature.
In vitro release studies. In vitro release studies of catanionic pharmacosomes were evaluated using a dynamic dialysis technique. A certain amount of catanionic pharmacosome dispersion was sealed in a dialysis membrane bag with molecular weight cut-off of ∼3500 Da (Solarbio), and incubated at 37.0 ± 0.5 °C in different phosphate buffer solutions (PBS, pH 1.2, pH 5.8, pH 7.4) with a total volume of 55 mL. The phosphate buffer solutions were composed of desirable amounts of NaH₂PO₄, NaOH and water. The solution was continuously stirred at 100 rpm. At selected time intervals, 1 mL of solution was withdrawn from the release media and replaced by the same amount of fresh PBS. The solution was assayed by a UV spectrophotometer (TU-1810, Beijing Purkinje General Instrument Company Limited, China) at a wavelength of 366 nm, which is a typical absorbance peak of amlodipine. For comparison, the release of the pure amlodipine in the release medium was also investigated.
Drug-release kinetics. The kinetics of amlodipine release from the catanionic pharmacosomes were determined by fitting the curves (% release against time) to distinct models (summarized in Table 1). The criterion for selecting the most appropriate model was based on a goodness-of-fit test.²³

Table 1 Equations for different drug release models

Model	Equation^a	Note
a Where Q is the amount of drug released in time (t), t is time, k0, k1, kH and k are the release constants of their corresponding equations.
Zero-order equation	Q = k0t
First-order equation	ln(100 − Q) = k1t	“(100 − Q)” is the amount of drug remaining at time (t).
Higuchi equation	Q = kHt^1/2
Ritger–Peppas equation	lnQ = a + nlnt (Q = kt^n)	“a” is a constant comprising the structural and geometric characteristics of the formulation; “n” is the release exponent.

---

Results and discussion

Chemical structure of amphiphilic amlodipine prodrug molecules

FTIR was performed to analyze the formation of the amphiphilic prodrug molecules. Fig. 1 shows the FTIR spectra of pure amlodipine (a), oleic acid (b), and amlodipine prodrug (c). In the amlodipine spectrum, the sharp band at 3391 cm⁻¹ is attributed to the stretching vibration of –NH₂, the intense peak at 1687 cm⁻¹ is derived from the existence of C–O stretching and the band at 1284 cm⁻¹ also exhibits the presence of C–O stretching. For the spectrum of oleic acid, the broad feature from 3500 cm⁻¹ to 2500 cm⁻¹ is characteristic of the O–H stretching band of the acid and the broadness of this band is caused by intramolecular hydrogen bonding. The characteristic carbonyl band appears at 1711 cm⁻¹. The two absorption peaks that appear at 1412 cm⁻¹ and 1285 cm⁻¹ are due to O–H bending and C–O stretching. The O–H in-plane and out-of-plane bands appear at 1465 cm⁻¹ and 938 cm⁻¹, respectively. The –CH₂ symmetric and asymmetric stretching vibrations are detected from 2854 cm⁻¹ to 2926 cm⁻¹. In the spectrum of prodrug, firstly the sharp band of –NH₂ disappears and asymmetric stretch of –NH₃⁺ appears with the band at 3300 cm⁻¹ in the FTIR spectrum. Another characteristic band of –NH₃⁺ is seen as a strong band at 1610 cm⁻¹. Our analysis is supported by the study of Foley and Enescu²⁴ who observed this mode (δ_asNH₃⁺) at 1614 cm⁻¹ for cysteine zwitterions. Umbrella-like vibrations of –NH₃⁺ with a wave number of 1398 cm⁻¹ appeared in the FTIR spectrum, which corresponds well with the reported value.²⁵ Secondly, the disappearance of the O–H stretching band, bending vibration, in-plane and out-of-plane bands further proves the transfer of protons. Thirdly, compared with pure oleic acid, a new band at 1569 cm⁻¹, corresponding to the characteristic infrared adsorption of carboxylate (–COO⁻), appears after the formation of amphiphilic prodrug molecules. All these characteristic absorption bands confirm that amlodipine molecules accepted protons donated by oleic acid molecules, thus forming the amlodipine prodrug molecules. The structure of the prodrug molecule is further confirmed by the ¹H-NMR and DSC techniques. The results are shown in Fig. S1 and Fig. S2 in the ESI.^†

Fig. 1 FTIR spectra of amlodipine (a), oleic acid (b), the amlodipine prodrug (c).

Surface activity of amphiphilic amlodipine prodrug molecules

According to our above analysis, the synthesized amlodipine prodrug molecule contains the drug cationic ion and the surfactant anionic ion. In order to understand its behaviour in aqueous solutions, the surface activity of the prodrug was studied using surface-tension measurements obtained from using the Wilhelmy method. Fig. 2 shows the aqueous surface tension as a function of prodrug concentration. From the figure it can be seen that surface tension (γ) decreases with the increase of amlodipine prodrug concentration. From the γ–c curve the critical aggregation concentration (CAC) can be determined to be about 6×10⁻⁵ mol L⁻¹, which is smaller than those obtained with conventional surfactants (about 10⁻³ mol L⁻¹∼10⁻⁴ mol L⁻¹). From this concentration point the amphiphilic prodrug molecules start to form aggregates spontaneously in the aqueous solution. The inset in Fig. 2 shows the hydrodynamic diameter (D_h) distributions of micelles with prodrug concentration at c = 8 × 10⁻⁵ mol L⁻¹. The D_h centered at the maximum intensity peak is around 7 nm, which corresponds to the value of the drug-participitating micelle. The water surface tension can be decreased to about 32 mN m⁻¹, suggesting that the amlodipine prodrug has extremely high surface active properties. It is noteworthy that both amlodipine and oleic acid cannot effectively decrease the water surface tension.

Fig. 2 Surface tension (γ) versus amlodipine prodrug concentration, the inset is hydrodynamic diameter (D_h) distributions of micelles with prodrug concentration at c = 8 × 10⁻⁵ mol L⁻¹.

Morphology of the aggregates

TEM was performed to characterize the morphology of the aggregates for the samples with different concentrations of amlodipine prodrug above CAC from 2 × 10⁻⁴ mol L⁻¹ to 1 × 10⁻² mol L⁻¹. The results are shown in Fig. 3. Within the concentration range of 2 × 10⁻⁴ mol L⁻¹ to 2 × 10⁻³ mol L⁻¹, the TEM results clearly show that single spherical unilamellar vesicles are formed in the dispersions of prodrug with a diameter of approximately 200 nm (Fig. 3a and 3b). This is also demonstrated by the bluish color of these samples, which is a typical color for vesicle fluids. The amphiphilic prodrug molecules contain –NH₃⁺ and –COO⁻ headgroups and the two hydrophobic tails from amlodipine and oleic acid molecules. It is a double-tailed zwitterionic surfactant, which is known to be able to form spontaneous vesicles in the aqueous solution⁴ according to the theory proposed by Israelachvili et al.²⁶ The formation of vesicles is illustrated in Fig. 3e. The protoned amlodipine molecule is a component of the bilayer. The aggregation behaviour of the prodrug molecule in the aqueous solution is somehow similar to the salt-free catanionic surfactant system. The presence of micelles is confirmed by dynamic light scattering (DLS), in solutions with concentrations that are higher than CAC but lower than the critical vesicle concentration (CVC, 2 × 10⁻⁴ mol L⁻¹). With the increase of prodrug concentration to 5 × 10⁻³ mol L⁻¹, rod-shaped aggregates appear (Fig. 3c and Fig. 3d) whose length ranges from 300 nm to 1 μm and cross-section is approximately 30–40 nm. The rod-shaped aggregates are formed by the fusion of the vesicles. According to the reference, the vesicles are first arranged in an ordered array and then fused through neck formation.²⁷ Finally, smoothing of the fused vesicles provides the rod-shaped multilamellar vesicles. Therefore, the state of amlodipine prodrug molecules in the solution experiences the following transitions with the increase of its concentration: the prodrug monomer → micelles → unilamellar vesicles → rod-shaped multilamellar vesicles. Hence, the concentration plays a crucial role in structural transition. It also should be noted that the pure amlodipine cannot dissolve in water and no vesicles appeared in the pure amlodipine aqueous solution. The amlodipine prodrug molecules contain both the polar groups of –NH₃⁺ and –COO⁻ and the hydrophobic chain from the oleic acid molecules. Hence, it is easy to understand the aggregation behaviour of the prodrug molecules in the aqueous solution.

Fig. 3 TEM images of the aggregates for the samples with different concentrations of prodrug (a) c = 2 × 10⁻⁴ mol L⁻¹ (b) c = 2 × 10⁻³ mol L⁻¹ (c) c = 5 × 10⁻³ mol L⁻¹ (d) c = 1 × 10⁻² mol L⁻¹ and the illustration of the vesicle formation (e).

Size and size distribution of the aggregates

Size plays an important role in the drug delivery system. Nanoparticles larger than 400 nm in diameter are known to be easily and rapidly captured by the reticuloendothelial systems and cannot maintain stable circulation in the bloodstream.^28,29 Thus the size, and size distribution, of the prepared prodrug dispersions solution are characterized by DLS. The results are shown in Fig. 4. The mean hydrodynamic diameter (D_h) centered at the maximum intensity peak is around 230 nm for the vesicular samples of c = 5 × 10⁻⁴ mol L⁻¹, 1 × 10⁻³ mol L⁻¹ and 2 × 10⁻³ mol L⁻¹, which totally supports the TEM results in the corresponding concentration. The D_h is largely increased to ∼540 nm and ∼845 nm for the samples of c = 5 × 10⁻³ mol L⁻¹ and c = 1 × 10⁻² mol L⁻¹, respectively, which correspond to the size of the rod-shaped multilamellar vesicles shown in the TEM images (Fig. 3c and 3d). Therefore, the size of the present spherical drug-participating vesicular dispersions is desirable for passive targeting drug delivery.

Fig. 4 Hydrodynamic diameter (D_h) distributions of the aggregates with different concentrations of the amlodipine prodrug.

The synthesized amlodipine prodrug molecules simultaneously contain –NH₃⁺ and –COO⁻ groups. Thus, the as-prepared catanionic pharmacosomes should be a pH-responsive assembly. In order to understand the pH-responsive property of the catanionic pharmacosomes, DLS is performed to measure the size of the aggregates for the sample of c = 2 × 10⁻³ mol L⁻¹ at different pH values. The results are shown in Fig. 5. It can be seen that D_h is largely dependent on the pH of the solution. When the pH value is below 7.0 the D_h is increased to a micrometer: around 1000 nm and 1300 nm for the measured samples of pH 5.8 and pH 1.2, respectively. While the pH value is larger than 7.0, the D_h of the catanionic pharmacosome is about 230 nm. The large size at lower pH value can be explained as follows. At low pH values, the carboxyl groups of the prodrug molecules could accept the abundant proton. In this case amlodipine prodrug molecules dissociate, leading to the release of oleic acid molecules. The oleic acid molecules separate from the water as small oil droplets due to its insolublility in water. At high pH values the size of the aggregates changes very little, indicating that the amlodipine prodrug molecules are stable in the alkaline aqueous solution. However, some protons in –NH₃⁺ are neutralized by –OH⁻, which is proved by the decrease of zeta potential with the rise of pH, for example, −9.85 mV for the sample of c = 2 × 10⁻³ mol L⁻¹ at pH 7.4 and −68.2 mV at pH 10.0. Therefore, the change of the solution’s pH value would result in the assembly and disassembly of the prodrug molecules, leading to the controlled release of the drug molecules from the catanionic pharmocosomes.

Fig. 5 The D_h of the aggregates for the sample of c = 2 × 10⁻³ mol L⁻¹ at different pH values.

Conductivity of the catanionic pharmacosome dispersions

The conductivities of the catanionic pharmacosome dispersions with different prodrug concentrations are shown in Fig. 6. There are two stages in the plot. Firstly, the increase of concentration the conductivity increases rapidly until to c = 2 × 10⁻³ mol L⁻¹, and then the conductivities keep nearly constant from c = 2 × 10⁻³ mol L⁻¹ to c = 1 × 10⁻² mol L⁻¹. The electrical conductivity of the solution is attributed to the conductivities of charged aggregates and free ions in the aqueous solution. In the present study, we prepared catanionic pharmacosomes from salt-free aqueous solutions according to Bronsted–Lowry's acid–base theory; therefore the concentration of charged aggregates plays a dominant role. At the beginning, the number of aggregates increases with the increase of the prodrug concentration, which results in the rise of the conductivity. Subsequently, the conductivity changes little from c = 2 × 10⁻³ mol L⁻¹ to c = 1 × 10⁻² mol L⁻¹, which could be associated with the formation of rod-shaped multilamellar vesicles in the solution.

Fig. 6 The conductivities of the dispersions with different prodrug concentrations.

In vitro release studies

The catanionic pharmacosome contains 59.1 wt% amlodipine, as calculated from the molecular structure in Scheme 1. This high drug-loading content is an important advantage of catanionic pharmacosomes over conventional nanoparticles and liposomes, whose drug-loading contents are usually below 20%. The drug-release profiles of catanionic pharmocosome dispersions in the phosphate-buffered solution (PBS, pH 7.4, the typical medium used to mimic blood fluid) at different concentrations are shown in Fig. 7. For comparison, the drug release profile of pure amlodipine is also presented in Fig. 7. As shown in the figure, the pure amlodipine is rapidly released into the PBS medium and reaches about 93% in 6 hours. However, for all of the catanionic pharmacosome dispersions, the release amounts only reach approximately 20% in 6 hours. The final release rate of amlodipine from the catanionic pharmacosome dispersions in phosphate buffered solution (pH 7.4) is significantly lower than that of the pure drug solution. This clearly indicates that the amount of the free drug released from the prodrug-constructing catanionic pharmacosome dispersions is effectively retarded. In this case, firstly, some prodrug molecules move from the vesicles to the solution because there always existed a dynamic equilibrium between the prodrug monomer and their participating-vesicles; subsequently, the prodrug molecules will disassociate into the amlodipine and oil acid molecules in the aqueous solution; thirdly the drug molecules diffuse from the dialysis bag into the PBS solution. Thus, the transport of the free drug molecules across the dialysis bag reduces the free drug concentration in the dialysis, leading to the gradual release of drug molecules from the catanionic pharmacosomes to the solution. Therefore, the release rate of the drug molecules could be well controlled. All the catanionic pharmacosome dispersions show a better sustained release profile, in that the total release amount is less than 50% within the whole studied time. The in vitro release results thus clearly reveal that this kind of catanionic pharmacosome could be a suitable carrier for the controlled release of amlodipine. It also can be seen from the figure that the drug release rate for the sample of c = 2 × 10⁻³ mol L⁻¹ was slightly faster than those from the samples of larger concentration, c = 5 × 10⁻³ mol L⁻¹ and c = 1 × 10⁻² mol L⁻¹. This may be caused by structural differences between the samples. As discussed previously for the TEM images (Fig. 3), the single spherical unilamellar vesicles are formed in the sample of c = 2 × 10⁻³ mol L⁻¹, while the rod-shaped aggregates exist in the samples of c = 5 × 10⁻³ mol L⁻¹ and c = 1 × 10⁻² mol L⁻¹.

Fig. 7 Cumulative releases of samples with different concentrations of pH 7.4 PBS. (a) pure drug (b) c = 2 × 10⁻³ mol L⁻¹ (c) c = 5 × 10⁻³ mol L⁻¹ (d) c = 1 × 10⁻² mol L⁻¹.

In order to better understand the effect of the release media on the drug release kinetics and further study the interaction mechanism between amlodipine and oleic acid molecules, the other two release media (pH 1.2, HCl and pH 5.8, PBS) were chosen to be studied via in vitro release. The percentage cumulative drug release from the as-prepared catanionic pharmacosomes (c = 2 × 10⁻³ mol L⁻¹) in different release media is shown in Fig. 8. It can be seen that the pH of the release medium has a significant effect on the drug release. Compared with the drug release in pH 7.4 PBS, the release rate is much faster in pH 1.2 and pH 5.8 solutions, which is somewhat like the release behaviour of the pure drug solution. In the two acid solutions, the release amount reaches 73% and 50% in 6 hours, respectively. However, in 7.4 PBS, the cumulative release percentage is only about 20% in 6 hours. This is mainly due to the following reasons: the prodrug is synthesized by the proton transfer between the amlodipine and the oleic acid molecules, thus its stability in the solution is largely dependent on the pH value of the solution. As discussed in the size and size distribution section, in acidic media the amlodipine prodrug molecules will be dissociated into amlodipine molecules and oleic acid molecules, which results in faster drug release. However, the prodrug molecules are stable in the neutral pH solution, thus the catanionic pharmacosome is stable in the solution. As a result, the release rate of amlodipine from the catanionic pharmacosome in the neutral medium is slower than that in the acidic medium.

Fig. 8 Cumulative release of the samples with prodrug concentration at c = 2 × 10⁻³ mol L⁻¹ in different release media, PBS (pH 7.4), PBS (pH 5.8), and HCl aqueous solution (pH 1.2).

Drug release kinetics

Mathematical models such as zero-order, first-order, and Higuchi's and Ritger–Peppas models³⁰ are usually used to describe the kinetics of the drug release from the test formulation. The zero-order equation describes the systems where the drug release rate is independent of time and drug concentration of the dissolved substance. The zero-order mechanism ensures that a steady amount of drug is released over time, minimizing potential drug peak–trough fluctuations and side effects, while maximizing the amount of time that the drug concentrations remain within the therapeutic window (efficacy). This is particularly important for 24 hour delivery of drugs with a narrow therapeutic index. The first-order equation describes systems where the drug release rate depends on its concentration. First-order drug release is the dominant extended release profile found in the pharmaceutical industry today. The Higuchi equation suggests that the drug release is controlled by diffusion. Analysis of the experimental data using the Ritger–Peppas equation and interpretation of the release exponent (n), provides a better understanding of the mechanisms for the controlled release. As reported in the previous literature, when n approximates 0.43, a Fickian diffusion controlled release is implied. However, for n = 1, the drug release occurs as an apparent zero-order mechanism (transport case II), which is not dependent on time. Values of n between 0.43 and 1 are an indication of anomalous transport (not Fickian).^31,32 The criterion for selecting the most appropriate model was based on a goodness-of-fit test.²³ Representative regression coefficients (R) obtained by fitting experimental release data to distinct models are shown in Table 2.

Table 2 The regression equation of amlodipine release from different aggregate dispersions in vitro

Sample	Model	Equation^a	R
a Where Q is the amount of drug release, t is the time and R is the regression coefficient.
Pure amlodipine	Zero-order kinetics	Q = 37.949 + 9.266t	0.791
	First-order kinetics	ln(100 − Q) = 4.137 − 0.371t	0.956
	Higuchi equation	Q = 18.395 + 32.679t^1/2	0.918
2×10^−3 mol L^−1	Zero-order kinetics	Q = 16.764 + 0.193t	0.868
	First-order kinetics	ln(100 − Q) = 4.411 + 0.00299t	0.897
	Higuchi	Q = 8.396 + 3.408t^1/2	0.958
	Ritger–Peppas	lnQ = 2.012 + 0.398lnt	0.973
5×10^−3 mol L^−1	Zero-order kinetics	Q = 13.118 + 0.121t	0.789
	First-order kinetics	ln(100 − Q) = 4.456 + 0.00157t	0.804
	Higuchi	Q = 7.250 + 2.227t^1/2	0.906
	Ritger–Peppas	lnQ = 1.711 + 0.386lnt	0.965
1×10^−2 mol L^−1	Zero-order kinetics	Q = 13.624 + 0.122t	0.861
	First-order kinetics	ln(100 − Q) = 4.455 − 0.00162t	0.884
	Higuchi	Q = 8.319 + 2.158t^1/2	0.951
	Ritger–Peppas	lnQ = 2.156 + 0.286lnt	0.990

---

For the samples with different concentrations, it is observed that the R value is the largest when fitted to the Ritger–Peppas equation as opposed to the other equations, which indicates a Ritger–Peppas release from these aggregates. As shown in Table 2, the release exponents are 0.398, 0.386 and 0.286 for the sample of c = 2 × 10⁻³ mol L⁻¹, c = 5 × 10⁻³ mol L⁻¹ and c = 1 × 10⁻² mol L⁻¹, respectively. Thus, the values of the release exponents indicate a Fickian diffusion-controlled release for the amlodipine from different vesicle dispersions. As for the release of pure amlodipine, the R value is largest when fitted to the first-order kinetics equation as opposed to the other equations, which indicates first-order release kinetics from the pure amlodipine. Although some processes may be clearly classified as either diffusion- or erosion-controlled, drug release is mostly governed by both mechanisms. However, these results may provide a better understanding of the mechanism of the drug release from the drug-loaded vesicles.

Conclusions

In conclusion, the aggregation behaviour of an amphiphilic prodrug molecule synthesized from proton transfer between the drug molecules and the surfactant molecules was investigated. The synthesized molecule is not only a prodrug with bioactivity but also a zwitterion surfactant with amphiphilic properties. The results demonstrate that the present amphiphilic amlodipine prodrug molecules self-assemble into catanionic pharmacosomes (bioactive catanionic vesicles) in the aqueous solution. The drug release behaviour results indicate that the as-prepared catanionic pharmacosomes exhibit excellent sustained drug release properties. The direct use of drugs as part of a carrier has its unique advantages. For example, it can minimize surfactant toxicity, increase the drug loading content and reduce the need for tedious chemical synthesis encountered by the conventional catanionic vesicles and pharmacosomes. The present study provides a new concept for drug delivery systems, which might open a new door for drug delivery applications. Further studies are ongoing to extend the present concept to various kinds of drugs.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (NSFC, No.20803044, 21173127), the Natural Science Foundation of Shandong Province (ZR2011BQ003) and the Independent Innovation Foundation of Shandong University ( IIFSDU, 2012TS099).

References

1. D. D. Lasic, Applications of liposomes in Handbook of Biological Physics, ed. R. Lipowsky and E. Sackmann, Elsevier, Amsterdam, 1993, vol. 1, p. 493.
2. S. H. Tung, H.Y. Lee and S. R. Raghavan, J. Am. Chem. Soc., 2008, 130, 8813.
3. J. H. Lee, Ph.D. Thesis, University of Maryland, 2006.
4. E. W. Kaler, A. K. Murthy, B. E. Rodriguez and J. A. N. Zasadzinski, Science, 1989, 245, 1371.
5. I. F. Uchegbu and S. P. Vyas, Int. J. Pharm., 1998, 172, 33.
6. K. M. Surinder, J. Neha and K. Gurpreet, Colloids. Surf., B: Biointerfaces., 2011, 87, 173.
7. D. E. Discher and A. Eisenberg, Science, 2002, 297, 967.
8. F. Wang, Y. C. Wang, L. F. Yan and J. Wang, Polymer., 2009, 50, 5048.
9. K. Liu, C. Wang, Z. Li and X. Zhang, Angew. Chem., Int. Ed., 2011, 50, 4952.
10. X. Zhang and C. Wang, Chem. Soc. Rev., 2011, 40, 94.
11. N. Hassan, J. M. Ruso and Á. Piñeiro, Langmuir, 2011, 27, 9719.
12. D. Xu and Q. Cheng, J. Am. Chem. Soc., 2002, 124, 14314.
13. S. Egli, M. G. Nussbaumer, V. Balasubramanian, M. Chami, N. Bruns, C. Palivan and W. Meier, J. Am. Chem. Soc., 2011, 133, 4476.
14. T. M. Allen and P. R. Cullis, Science., 2004, 303, 1818.
15. A. Boudier, P. Castagnos, E. Soussan, G. Beaune, H. Belkhelfa, C. Ménager, V. Cabuil, L. Haddioui, C. Roques, I. Rico-Lattes and M. Blanzat, Int. J. Pharm., 2011, 403, 230.
16. H. Yin, Y. Lin, J. Huang and J. Ye, Langmuir., 2007, 23, 4225.
17. K. K. Karukstis, C. A. Zieleniuk and M. J. Fox, Langmuir, 2003, 19, 10054.
18. X. Wang, E. J. Danoff, N. A. Sinkov, J. H. Lee, S. R. Raghavan and D. S. English, Langmuir, 2006, 22, 6461.
19. B. Tah, P. Pal, M. Mahato and G. B. Talapatra, J. Phys. Chem. B., 2011, 115, 8493.
20. S. Ghosh and J. Dey, J. Colloid. Interface. Sci., 2011, 358, 208.
21. N. Dew, K. Edwards, J. Eriksson, K. Edsman and E. Björk, Colloids. Surf. B: Biointerfaces, 2012, 89, 53.
22. Y. Shen, E. Jin, B. Zhang, C. J. Murphy, M. Sui, J. Zhao, J. Wang, J. Tang, M. Fan, E. V. Kirk and W. J. Murdoch, J. Am. Chem. Soc., 2010, 132, 4259.
23. M. Bamba, F. Puisieusx, J. P. Marty and J. T. Carstensen, Int. J. Pharm., 1979, 2, 307.
24. S. Foley and M. Enescu, Vib. Spectrosc., 2007, 44, 256.
25. L. Szabó, V. Chis, A. Pìrnau, N. Leopold, O. Cozar and Sz. Orosz, J. Mol. Struct., 2009, 924, 385.
26. J. Israelachvili, D. Mitchell and B. E. Ninham, J. Chem. Soc., Faraday Trans., 1976, 72, 1525.
27. S. Maity, P. Jana, S. K. Maity and D. Haldar, Langmuir., 2011, 27, 3835.
28. M. Yokoyama, J. Artif. Organs, 2005, 8, 77.
29. F. Danhier, O. Feron and V. Préat, J. Controlled Release, 2010, 148, 135.
30. P. L. Ritger and N. A. Peppas, J. Controlled Release, 1987, 5, 37.
31. P. L. Ritger and N. A. Peppas, J. Controlled Release, 1987, 5, 23.
32. N. Lin, J. Huang, P. R. Chang, L. Feng and J. Yu, Colloids. Surf. B: Biointerfaces., 2011, 85, 270.

---

Footnote

† Electronic Supplementary Information (ESI) available: H NMR and DSC spectra of the amphiphilic prodrug molecule. See DOI: 10.1039/c2ra20653f

---