Article pubs.acs.org/Langmuir
Understanding the Mechanism of Enzyme-Induced Formation of Lyotropic Liquid Crystalline Nanoparticles Linda Hong,† Stefan Salentinig,† Adrian Hawley,§ and Ben J. Boyd*,†,‡ †
Drug Delivery, Disposition and Dynamics, Monash Institute of Pharmaceutical Science, and ‡ARC Centre of Excellence in Convergent Bio-Nano Science and Technology, Monash Institute of Pharmaceutical Sciences, Monash University, Parkville Campus, 381 Royal Parade, Parkville, VIC 3052, Australia § SAXS/WAXS Beamline, Australian Synchrotron, 800 Blackburn Rd., Clayton, VIC 3150, Australia S Supporting Information *
ABSTRACT: Liquid crystalline nanoparticles have shown great potential for application in fields of drug delivery and agriculture. However, optimized approaches to generating these dispersions have long been sought after. This study focused on understanding the mechanism of formation of cubosomes during the recently reported enzymatic approach and extending the approach to alternative lipid types other than phytantriol. The chain length of digestible lipids was found to influence the effectiveness of triglycerides in disrupting the equilibrium cubic phase structure to form the emulsion precursor. In general, a greater hydrophobicity of the triglyceride required a lower concentration to inhibit liquid crystal structure formation. Selachyl alcohol was also examined due to its nondigestible trait and ability to form the inverted hexagonal phase. Digestion of its precursor emulsion formed hexosomes analogous to the phytantriol-based systems. Finally, the assumption that fatty acids liberated during digestion needed to partition out of the nondigestible lipids for the re-formation of the phase structure was found to be untrue. Their ionization state, however, did have an effect on the resulting nanostructure, and this unique property could potentially provide a useful attribute for oral drug delivery systems.
■
INTRODUCTION
not yet clear how readily such processes may be scaled up for production. With the advent of highly-ordered structures observed in breast milk,20 there is interest in the cooperation between lipidbased formulations and enzymes. It was first noted by Patton et al. in simulated stomach conditions that the digestion of fat droplets could evolve liquid crystalline structure;21 however, they lacked the analytical techniques to confidently identify any cubic or hexagonal mesophases. The recently proposed enzymatic approach to form cubosomes22 (Figure 1) expanded on this to create a liquid crystalline delivery system. The justification is attributed to better stability for the shelf life and for sustained in vivo drug release. The process involves an emulsion precursor formulation that is enzymatically converted to cubosomes, as opposed to the direct dispersion of lipid in the top-down approach. The precursor is an emulsion consisting of digestible lipids that disrupt the structures formed by lyotropic nondigestible lipids in excess water. The emulsion is either an inverse micellar phase (L2) or a simple unstructured emulsion. To “activate” the formulation, the digestible lipids are digested with lipase enzyme. The resulting hydrophilic digestion products are thought to partition toward the aqueous
Dispersions of self-assembled lipid systems, including the bicontinuous cubic phase (V2) and the hexagonal phase (H2), termed cubosomes and hexosomes, respectively, have gained immense interest in recent years. They have been applied to vaccine delivery,1 drug delivery,2−5 biosensing,6 and food applications.7,8 This is attributed to their ability to control content release and encapsulate ingredients of varying polarities.8−10 Their novelty also comes with the ability to modify size and internal structure,11 in particular for the cubic phase systems which presents many different space groups such as Pn3m, Im3m, and Ia3d structures.12 Even so, the systems are not yet used in commercial drug delivery products, partly due to colloidal instability, and consequently, there are ongoing efforts to improve the formulation and production methods. Current methods to generate cubosomes include the topdown13 and bottom-up approach.14 These methods are not efficient at generating cubosomes. They can produce vesicular byproducts which are believed to eventually transition into cubosomes.15,16 Unfortunately, this can also confound interpretation of characterization and drug release kinetics.17 Additionally, they can be energy-intensive which risks the integrity of the internal structure of the cubosome14 as well as induce degradation of active ingredients.17 The auxiliary solvent evaporation18 and spray drying19 methods have since been developed to circumvent the use of high-energy processes. It is © 2015 American Chemical Society
Received: May 11, 2015 Revised: May 30, 2015 Published: June 1, 2015 6933
DOI: 10.1021/acs.langmuir.5b01615 Langmuir 2015, 31, 6933−6941
Article
Langmuir
Figure 1. Chemical structures on the left illustrate the nondigestible lipids and the digestible lipids which are the components of the precursor emulsion. The diagram on the right is the schematic of the enzymatic process. The triglyceride in the unstructured emulsion is digested by the lipase, allowing the digestion products, fatty acids, to potentially partition out of the lipid mass. The remaining PHYT can then reassemble into cubosomes in the excess aqueous medium.
bulk phase.22 This leaves the nondigestible lipids to selfassemble into their liquid crystalline structure. The technique is advantageous over former methods as it removes the need for handling any semisolid bulk material and avoids the use of highenergy processes.22 While it is true that the emulsion still necessitates energy for dispersing the lipids, it is considerably less than the energy required for breaking down the bulk lipid phase and can be formed by simple shaking. Constituents of this precursor emulsion include a nondigestible and digestible lipid. The digestible lipid, such as a triglyceride, has at least one ester linkage cleavable by lipase. Triglycerides are unable to form liquid crystalline structures and contain three ester bonds, releasing up to three fatty acids when hydrolyzed by lipase.23 Most long chain triglycerides are cleaved at the sn-1 and sn-3 positions, leaving the sn-2 position intact due to regioselectivity. This results in two fatty acids and one monoglyceride.23 For tributyrin, lipase action results in 3 mol of butyric acid and 1 mol of glycerol.24 On the contrary, the nondigestible lyotropic lipid lacks any ester groups and is capable of forming structured phases in excess water. Phytantriol (PHYT) is known to form the V2 cubic phase upon hydration25 and thus provides cubosomes when dispersed. It has been demonstrated to maintain this structure in the presence of lipase, thereby providing a sustained drug release effect.26 Selachyl alcohol (SA) is another nondigestible lipid but instead forms the H2 hexagonal phase.27 Similar to PHYT, its structural integrity of the phase persists in vivo during drug release.28 From the recent communication introducing the enzymatic approach to form cubosomes,22 there was a decreasing trend in the amount of triglyceride required to disrupt the structure of the PHYT cubic phase structure as the length of the triglyceride chain was increased. Triacetin, with two carbons in its corresponding fatty acid, required a concentration of at least 90% w/w of the total lipid mass to disrupt the V2 cubic phase. With longer chains such as tripropionin (three carbons) and tributyrin (four carbons), the concentrations decreased to 30 and 15%, respectively. Consequently, using longer chain triglycerides such as tricaprylin (eight carbons) and trilaurin (12 carbons) should prove more efficient at disrupting the phase structure at relatively lower concentrations.
Besides lipid composition, the distribution of fatty acids was another factor considered in relation to the mechanism of the enzymatic approach. Because of the hydrophilic nature of short chain fatty acids, they were purported to partition out of the lipid globules and into the bulk aqueous phase. In doing so, it allows the formation of the internally nanostructured particles. This hypothesized mechanism was not actually proven analytically. The impact of altering the pH of the digested precursor system, which is expected to alter the ionization of the fatty acids and influence their disposition in the system, was not investigated either. Therefore, to better understand the mechanism and general applicability of the enzyme-induced formation of cubosomes and hexosomes, synchrotron-based small-angle X-ray scattering (SAXS) was used. First, the impact of longer chain length triglycerides on the structures formed in the precursor emulsions was determined. The resultant phase diagram revealed the optimal compositions for precursor systems. Optimized precursors were digested using the previously reported in vitro lipolysis model coupled to the flow-through cell29 to allow for real-time monitoring of the structures evolved. The distribution between free fatty acids between the bulk aqueous phase of the digested formulations and those in the particles were determined using the ultrafiltration method30 and assayed using high performance liquid chromatography (HPLC).
■
MATERIALS AND METHODS
Phytantriol (3,7,11,15-tetramethylhexadecane-1,2,3-triol) was obtained from DSM Nutritional Products Ltd., Singapore, with purity of 97% and selachyl alcohol from Hai Hang Industries, Shang Hai, China. Tributyrin (C4TG), trilaurin (C12TG), and tripropionin (C3TG) were purchased from TCI, Tokyo, Japan. Tricaprylin (C8TG) and triacetin (C2TG) were obtained from Sigma-Aldrich, St. Louis, MO. Tris buffer, used as the digestion media, was composed of 150 mM sodium chloride (Chem-Supply, SA, Australia), 50 mM trizma maleate (Sigma-Aldrich, St. Louis, MO), 5 mM calcium chloride dihydrate (Ajax Finechem, NSW, Australia), 6 mM sodium azide (Merck Schuchardt OHG, Darmstadt, Germany), and 35 mM sodium hydroxide pellets (Ajax Chemicals, NSW, Australia). The pH was readjusted to 6.5 using sodium hydroxide and hydrochloric acid solutions, which is the optimal pH for activity of pancreatic lipase.31 Pluronic F108 (Sigma, France) was the colloidal stabilizer for the precursor emulsions. Pancreatin (Southern Biological, VIC, Australia) 6934
DOI: 10.1021/acs.langmuir.5b01615 Langmuir 2015, 31, 6933−6941
Article
Langmuir
Figure 2. Compilation of the SAXS profiles for phytantriol/tributyrin (C4TG) mixtures at varying concentrations. L2 phase was observed at 15% w/ w C4TG. In the 3% w/w profile, hexosomes with two slightly different lattice parameters were observed, noted by the lower intensity peaks adjacent to the √3 and √4 peaks. There is a region of uncertainty between 10 and 15% w/w C4TG about the exact location of the phase boundary. The phase boundaries were determined in the same way for each different nondigestible and digestible lipid combination, and the data are summarized in Figure 3. was the enzyme used, and the inhibitor was 4-bromophenylboronic acid, 4-BPBA (Aldrich, China). Fatty acids, caprylic acid (Sigma-Aldrich, St. Louis, MO), butyric acid (Sigma-Aldrich, St. Louis, MO), and lauric acid (Fluka, Switzerland) were used to make up the standard solutions for HPLC. HPLC grade methanol (Merck KGgA, Darmstadt, Germany) and trifluoroacetic acid (HiperSolv, England) were used in the mobile phase for HPLC. Water from a Milli-Q purification system (Billerica, MA) was used in formulations and for HPLC. All chemicals were used without further purification. Preparation of Precursor Formulations. Precursor emulsions consisting of one nondigestible lipid and one digestible lipid were combined at the required w/w ratios at 60 °C. The lipid component totaled 10% w/w of the total formulation while the remainder was Tris buffer containing 1% w/v Pluronic F108. The final mixture (17 mL) was vortexed and then sonicated via a Misonix Ultrasonic liquid processor S-4000 (USA) at 25 amplitude for 1 min, pulsing on for 4 s and off for 1 s. In Vitro Digestion of Optimal Precursor Formulations. Digestions were performed with a Metrohm potentiometric titrator (Switzerland), using the Metrohm iUnitrode Pt 1000 (Switzerland) as the pH probe. Conditions were optimized to 0.7 μL/min for the flow rate, magnetic stirrer at setting 6, and the basic solution was 0.2 M sodium hydroxide. Temperature was maintained at 37 °C with a water bath. Each digestion was run for 40 min with a preceding 5 min buffer time to readjust the pH to 6.5. The enzyme was added approximately 200 s after the readjustment time of the pH. At the end of digestion, 10 mL of the digested formulation was added to 90 μL of 0.5 M 4BPBA to inhibit further lipolysis, vortexed, and stored at 4 °C until further use. Software that controlled the digestion apparatus and processed its data was the accompanying program, Tiamo, version 2.3. The default composition for all digestions, was 16 mL of formulation and 1 mL of enzyme supernatant (∼7190 TBU) to give a total volume of 17 mL. Enzyme was prepared at 2.00 g for 5 mL of Tris buffer at pH 6.5 and then centrifuged (Allegra X-22R centrifuge) at 2360g for 15 min at room temperature. Assaying Free Fatty Acids Postdigestion. The pH of inhibited digested solutions were readjusted to pH 6.5 using the Denver Instrument pH conductivity meter model 220 (Australia), and then approximately 1 mL was transferred to an Eppendorf tube for subsequent analysis. The same was repeated for pH 8.0, 5.0, 4.0, and 3.0 using up to 10 M sodium hydroxide and 10 M hydrochloric acid to adjust the pH. The solutions at each pH were then filtered using the Amicon stirred ultrafiltration cell, model 8010 (USA), with Ultracel 30 kDa
ultrafiltration discs (Millipore, USA) to separate the liquid crystalline nanoparticles from the filtrate which contained the free fatty acids. The first four drops were discarded, and the filtrate of the bulk was analyzed.30 High-pressure liquid chromatography (HPLC) was used to quantify the amount of fatty acid in digested formulations.32 Details of the assay are provided in the Supporting Information. Small-Angle X-ray Scattering (SAXS). SAXS analyses were performed on the SAXS/WAXS beamline at the Australian Synchrotron.33 The wavelength of the X-ray beam was 1.1271 Å (11 keV), and a q range of 0.03 < q < 1.68 Å−1 was attained using a sampleto-detector distance of 567 mm. The two-dimensional scattering patterns were acquired over an active area of 169 × 179 mm2 using a Pilatus 1M detector, with the dimensions of each pixel 172 μm. Data were then converted to a one-dimensional scattering function I(q) using the in-house-developed software package scatterBrain. For equilibrium precursor formulations with varying triglyceride content, an aliquot (200 μL) was transferred to a 96-well plate, mounted vertically in the beam and the internal structure of the dispersed particles determined from a 2 s acquisition. For dynamic digestion studies using the in vitro lipolysis model, silicon tubing (total volume
