Liposome and niosome preparation using a membrane contactor for scale-up

Colloids and Surfaces B: Biointerfaces 94 (2012) 15–21

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Liposome and niosome preparation using a membrane contactor for scale-up Thi Thuy Pham, Chiraz Jaafar-Maalej, Catherine Charcosset ∗ , Hatem Fessi Université de Lyon, F-69622 Lyon, France

a r t i c l e

i n f o

Article history: Received 4 November 2011 Accepted 26 December 2011 Available online 21 January 2012 Keywords: Membrane contactor Ethanol injection Liposomes Niosomes Caffeine Spironolactone

a b s t r a c t The scaling-up ability of liposome and niosome production, from laboratory scale using a syringe-pump device to a pilot scale using the membrane contactor module, was investigated. For this aim, an ethanol injection-based method was applied for liposome and niosome preparation. The syringe-pump device was used for laboratory scale batches production (30 ml for liposomes, 20 ml for niosomes) then a pilot scale (750 ml for liposomes, 1000 ml for niosomes) were obtained using the SPG membrane contactor. Resulted nanovesicles were characterized in terms of mean vesicles size, polydispersity index (PdI) and zeta potential. The drug encapsulation efficiency (E.E.%) was evaluated using two drug-models: caffeine and spironolactone, a hydrophilic and a lipophilic molecule, respectively. As results, nanovectors mean size using the syringe-pump device was comprised between 82 nm and 95 nm for liposomes and between 83 nm and 127 nm for niosomes. The optimal E.E. of caffeine within niosomes, was found around 9.7% whereas the spironolactone E.E. reached 95.6% which may be attributed to its lipophilic properties. For liposomes these values were about 9.7% and 86.4%, respectively. It can be clearly seen that the spironolactone E.E. was slightly higher within niosomes than liposomes. Optimized formulations, which offered smaller size and higher E.E., were selected for pilot scale production using the SPG membrane. It has been found that vesicles characteristics (size and E.E.%) were reproducible using the membrane contactor module. Thus, the current study demonstrated the usefulness of the membrane contactor as a device for scaling-up both liposome and niosome preparations with small mean sizes. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Drug delivery systems using colloidal carriers made of phospholipids and non-ionic surfactants, called liposomes [1] and niosomes [2] respectively, have distinct advantages over conventional dosage forms [24]. The major advantage lies in their amphipathic nature, which allows the incorporation of both hydrophilic and hydrophobic drugs. These vectors can act as drug reservoirs; they may serve as a solubilizing matrix, as a local depot for sustained release or permeation enhancers of dermally active compounds [3]. The modification of the composition, the membrane structure or surface can adjust the drug release rate and improve the affinity for the target site [4]. Liposomes are enclosed spherical vesicles that are organized in one or several concentric phospholipidic bilayers with an aqueous core inside. They are formed upon the self-assembling of the phospholipid molecules in contact with water. Although liposomes could be of a great interest for drug delivery, there are still some

∗ Corresponding author at: Laboratoire d’Automatique et de Génie des Procédés (LAGEP), UMR 5007, CNRS, CPE, 43 bd du 11 Novembre, 691622 Villeurbanne Cedex, France. Tel.: +33 04 72 43 18 67; fax: +33 04 72 43 16 99. E-mail address: [email protected] (C. Charcosset). 0927-7765/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2011.12.036

problems associated with their physico-chemical stability (phospholipid hydrolysis or oxidation) and their industrial preparation. As a way out, alternatives to phospholipids vesicles have been formulated to avoid technical difficulties during the preparation and ensure a better stability of carriers. Niosomes, vesicles based on hydrated mixtures of cholesterol and non-ionic surfactant, were first reported in the seventies by researchers in the cosmetic industry. They behave in vivo like liposomes, by prolonging the circulation of entrapped drug and modifying its organ distribution [5]. Although the structure and properties of niosomes are similar to those of liposomes, they alleviate their disadvantages in terms of chemical stability and cost which make them more attractive and offering greater interest for industrial manufacturing. In literature, several methods were reported for the preparation of vesicles; the conventional ethanol injection technique, first described by [27], offers the advantage of simplicity, the absence of potentially harmful chemicals, forming small sized particles without any physical treatments, the minimum of technical requirement, and, further, the possibility of scale-up [28]. Hence, novel approaches, based on the principle of the ethanolinjection technique, were reported. The membrane contactor is one of approaches that can be directly scaled-up in order to obtain a large-scale production of vesicles, thereby avoiding intermediate studies. Membrane contactors, firstly applied for the preparation

16

T.T. Pham et al. / Colloids and Surfaces B: Biointerfaces 94 (2012) 15–21

of emulsions, precipitates, polymeric and lipidic nanoparticles [6] have known increasing interest and were recently reported for the liposomes preparation using SPG membranes [7] or a hollow fibre module [8]. There are, however, no reported studies for the preparation of niosomes by this technique. It has been proved that scale-up is more straightforward with membrane contactors [9]. That is, membrane operations usually scale linearly, so that a predictable increase in capacity is achieved simply. The aims of this work are to develop and optimize a novel preparation strategy, based on a membrane contactor and to investigate the scaling up ability from laboratory scale using the syringe-pump device to pilot scale. Drug loaded liposomal and niosomal formulations were prepared with two different drugmodels (spironolactone, a hydrophobic molecule and caffeine, a hydrophilic one). Laboratory batches were prepared and formulations were characterized in terms of size, PdI, zeta potential, morphology and encapsulation efficiency (E.E.). Then, optimized formulations were produced on a large scale, results were compared and scale-up ability of the membrane contactor was discussed. 2. Materials and methods 2.1. Materials 2.1.1. Reagents Caffeine was supplied by Sigma–Aldrich Chemicals (Saint Quentin Fallavier, France). Cholesterol, dicetyl phosphate (DCP) and surfactants: Sorbitan monostearate (Span® 60) and Polysorbate 60 (Tween® 60) were purchased from Sigma–Aldrich Chemicals (Saint Quentin Fallavier, France). The phospholipid Lipoid® E80 was obtained from Lipoïd GmbH (Ludwigshafen, Germany). Produced from egg yolk lecithin, it contained approximately 82% of phosphatidyl-choline and 9% of phosphatidylethanolamine. All reagents were acquired with their analysis certificate. Deionized water (resistivity of 18 M cm−1 ) was used. Ethanol (100%) was obtained from Carlo Erba Reagenti (Milano, Italy). The reagents were of analytical grade and used such as without further purification. 2.1.2. Syringe-pump The syringe infusion pump 22 was purchased from Harvard Apparatus (Holliston, MA, United States). 2.1.3. SPG membrane module Shirasu porous glass (SPG) tubular membranes were purchased from SPG Technology (Miyazaki, Japan). SPG membranes are prepared by phase-separated glass leaching in the Na2 –O–CaO–MgO–Al2 O3 –B2 O3 –SiO2 system, which is synthesized from volcanic ash, called Shirasu, used as the main raw material [10,11]. The SPG membrane dimensions were as follows: length 0.125 m, inner diameter 1 × 10−2 m and thickness 1 × 10−3 m leading to an active membrane surface of 3.9 × 10−3 m2 . In the present study, the hydrophilic SPG membrane mean pore size was 0.9 ␮m. 2.2. Methods 2.2.1. Liposome and niosome preparation 2.2.1.1. Membrane contactor based technique. The membrane contactor experimental set-up included a positive displacement pump (Filtron, France), a pressurized vessel (Millipore), equipped with a manometer M1 , connected on one side to a nitrogen bottle (Linde Gas, France) and on the other side to the SPG membrane with two manometers (placed at the inlet and outlet of the module) (Fig. 1).

For the liposome and niosome preparation, the required amounts of excipients and active principle ingredient (API) were dissolved in ethanol. The nanovector compositions are detailed in Table 1. In both cases, the ethanolic phase was placed in the pressurized vessel. The connecting valve to the nitrogen bottle was opened and the nitrogen pressure was set at a fixed level. The aqueous phase was then pumped through the membrane device. Experiments were carried out in an open loop configuration to avoid the recirculation of formed nanovesicles in the experimental set-up, which might lead to their degradation. Therefore, flowrates were set to allow the circulation of both aqueous and organic phases at the same time. The optimal parameters applied for liposome preparation were selected from our previous study using a similar SPG membrane [12]. The pressure of the dispersed phase and the flowrate of the aqueous phase were set to 2.4 bar and 0.85 l/min, respectively. For niosomes, these data were changed and were set to 1.7 bar for the dispersed phase pressure and 1.2 l/min for the aqueous phase flowrate. Thus, as the water arrived to the membrane device inlet, the valve connecting the pressurized vessel to the filtrate side of the device was opened so that the organic phase permeated through the pores into the aqueous phase. Spontaneous nanovesicle formation occurred as soon as the organic solution was in contact with the aqueous phase. The experiment was stopped when air bubbles started to appear in the tube connecting the pressurized vessel to the membrane module, indicating that the pressurized vessel was empty. Then, the suspension was stabilized for 15 min under magnetic stirring (RW 20, Ika-Werk). Experiments were conducted at 22 ± 2 ◦ C for liposomes, and at 60 ± 2 ◦ C for niosomes. Finally, the ethanol was removed by rotary evaporation under reduced pressure (Rotavapor R-144, Buchi, Flawil, Switzerland). Finally, the SPG membrane was regenerated at the end of each experiment. The washing was performed by flushing the module twice using 500 ml of water and 250 ml of ethanol. The membrane permeability was evaluated as the slope of the permeate flow rate versus transmembrane pressure. This parameter was measured at the beginning of each experiment and was checked to be around 90% of its initial value. 2.2.1.2. Syringe-pump based technique. Nanovesicles were prepared using a syringe-pump device according to a recent study carried in our laboratory [7]. Optimum preparation parameters were set as follows: aqueous to organic phase volume ratio = 2, aqueous phase stirring rate during solvent injection = 800 rpm. The injection rate was set at 2 ml/min. The required amounts of excipients and active substance (Table 1) were dissolved in ethanol. The resulting organic phase was injected by means of the syringe-pump in a defined volume of ultra pure water under magnetic stirring (RW 20, Ika-Werk). Spontaneous liposome or niosome formation occurred as soon as the organic solution was in contact with the aqueous phase. This formation became evident on the appearance of the characteristic opalescence of colloidal dispersions. The suspension was stabilized for 15 min under magnetic stirring. Then the ethanol was removed by evaporation under reduced pressure (Rotavapor R-144, Buchi, Flawil, Switzerland). 2.2.2. Liposome and niosome characterization 2.2.2.1. Size distribution and zeta potential. Nanovesicle size, polydispersity index (PdI) and zeta-potential were measured by Dynamic light scattering (DLS) using the Malvern Zetasizer Nanoseries (Malvern Instruments Zen 3600, Malvern, UK). Zeta potential was measured by Smoluchowski’s equation from electrophoretic mobility of vesicles [13].

T.T. Pham et al. / Colloids and Surfaces B: Biointerfaces 94 (2012) 15–21

17

Fig. 1. Schematic drawing of the membrane experimental set-up for nanovector preparation.

Each sample was diluted with water and analyzed in triplicate at 25 ◦ C. The data were collected using the DTS (nano) software (version 5.0) provided with the instrument. 2.2.2.2. Transmission electron microscopy (TEM). Nanovesicle morphology was examined using a CM 120 transmission electron microscope (Philips, Eindhoven, Netherlands) operating at an accelerating voltage of 80 kV. For sample preparation, a dispersion drop was applied to a carbon-coated copper grid and left for 1 min then the remaining liquid was removed using a filter paper. Negative staining using a 2% phosphotungstic acid solution (w/w, pH 7.1), was directly made on the deposit during 1 min. Finally, the excess of phosphotungstic solution was removed with a filter paper and dried samples were observed. 2.2.2.3. Entrapment efficiency. Vesicles encapsulation efficiency (E.E.%) was measured by separating free drug from vesicleentrapped drug using the ultracentrifugation technique. Briefly, a defined volume of the suspension was centrifuged for 50 min at 50,000 rpm and +4 ◦ C (OptimaTM Ultracentrifuge, Beckman Coulter, USA) in order to separate the unloaded drug. The free drug amount (FD) was determined in the supernatant. On the other hand, an equal volume of suspension has been used in order to assess the total amount of the drug (TD) which was measured

after having dissolved and disrupted vesicles in ethanol using an ultrasound bath for 10 min (Bandelin Sonorex, Schalltec GmbH, Germany). The drug encapsulation efficiency (E.E. %) was expressed as follows: E.E. =

TD − FD × 100 TD

Caffeine and spironolactone concentrations were measured using a spectrophotometer UV-Vis method (Shimadzu UV mini-1240 V, Kyoto, Japan) at  absorbance of 272 nm and 237 nm, respectively. Spectrophotometric analytical method was validated as usually required. 3. Results and discussion 3.1. Laboratory scale production using the syringe-pump device 3.1.1. Liposomes 3.1.1.1. Caffeine loaded-liposomes. In the first part of the study, caffeine loaded-liposomes were prepared using the syringe-pump device. The process parameters have been optimized in a former study [7]. The Lipoid E80 was selected in this study as it has been frequently reported for liposome preparation in our laboratory [7,8]. Table 2 summarizes the formulation parameters (concentration in

Table 1 Nanovesicle compositions. Liposomes Composition

Organic phase

Aqueous phase

Active principle ingredient (API) + Lipoïd E80 Cholesterol Ethanol Water

Niosomes Volume (ml)

Composition

Syringe-pump

Membrane contactor

10

250

20

500

Active principle ingredient (API) + Cholesterol DCP Ethanol Water

Volume (ml) Syringe-pump

Membrane contactor

2

100

18

900

18

T.T. Pham et al. / Colloids and Surfaces B: Biointerfaces 94 (2012) 15–21

Table 2 Liposomes prepared by the syringe-pump device: quantitative composition and characteristics. Phospholipid concentration (mg/ml)

Cholesterol concentration (%, w/w PL)

API concentration (mg/ml)

Mean size ± S.D. (nm)

Caffeine L20Ch10Ca L20Ch30Ca L40Ch10Ca L40Ch30Ca L60Ch10Ca L60Ch30Ca

20 20 40 40 60 60

10 30 10 30 10 30

10 10 10 10 10 10

89 82 90 89 95 95

Spironolactone L40Ch30Spi

40

20

3

Formulation

phospholipids, cholesterol and caffeine); the phospholipid concentration was increased from 20 to 60 mg/ml and cholesterol fraction from 10% to 30% (w/w phospholipid). It is important to mention that cholesterol was classically added to liposome formulation in order to stabilize and reduce the lipidic bilayer permeability. As can be seen in Table 2, it was found that liposome mean size was comprised between 80 and 99 nm and the polydispersity index (PdI) between 0.12 and 0.17. The slightly larger size was obtained at the higher phospholipid concentration, whereas, the cholesterol concentration had no influence on the liposome properties. Similar results were previously reported for unloaded liposomes prepared by the ethanol injection method [7]. These PdI data indicated a rather uniform size distribution of the prepared liposomes. A typical size distribution curve (formulation L40Ch30Ca) will be shown in a following section. The zeta potential values varied from −16 to −29 mV, negative values are usually recognized to indicate good stability of the suspension. On the other hand, caffeine encapsulation efficiency was increased from 3.8% to 9.7%, by increasing phospholipid concentration. The same result was also reported with cytarabine-loaded liposomes as encapsulation efficiencies increased with increased phospholipid concentration [7]. Furthermore, the cholesterol concentration affected the encapsulation efficiency, the higher encapsulation being obtained at the higher cholesterol concentration. According to these results, it may be suggested that higher phospholipid and/or cholesterol concentrations decrease the membrane permeability and thus are favourable to caffeine retention within vesicles. It has been suggested that caffeine may be entrapped within lipophilic bilayer and in the hydrophilic core [14]. In another study, caffeine entrapment efficiency was 16% (w/w) for liposomes prepared by sonication [26], this higher data may be attributed to the larger vesicles size. 3.1.1.2. Spironolactone loaded-liposomes. Spironolactone, a lipophilic active principle ingredient, was chosen as a second model drug. The formulation parameters as well as liposome characteristics are given in Table 2. The encapsulation efficiency of spironolactone within liposome (86.4%) was much higher than caffeine, the hydrophilic model drug. Similar value was previously found (93%) for spironolactone loaded liposomes prepared by the ethanol injection method using a hollow fibre membrane device [8]. This high encapsulation efficiency can be explained by the high lipophilicity of this molecule and therefore attributed to its good solubility and affinity to phospholipid bilayers. Authors also suggested that at least a part of spironolactone was adsorbed on the vesicle surface as the zeta potential value decreased from −43 mV (unloaded liposomes) to −23 mV (spironolactone-loaded liposomes); the rest of spironolactone could be entrapped within lipid bilayers. In the present study, the zeta potential value was about −24 mV which is coherent with the previous value reported by Laouini et al. [8]. It is

± ± ± ± ± ±

4 2 3 5 3 4

89 ± 3

PdI ± S.D.

0.12 0.14 0.17 0.15 0.15 0.16

± ± ± ± ± ±

0.02 0.02 0.01 0.01 0.01 0.01

0.14 ± 0.01

Zeta potential ± S.D. (mV)

−16 −20 −29 −27 −23 −27

± ± ± ± ± ±

1.3 0.3 0.4 0.8 0.7 0.6

−24 ± 1.1

E.E. ± S.D. (%)

3.8 6.5 5.6 6.9 9.0 9.7

± ± ± ± ± ±

0.1 0.2 0.2 0.4 0.1 0.2

86.4 ± 2

usually recognized that the entrapment efficiency of drugs within liposomes is affected by the hydrophilic/lipophilic character of the drug. As example, the encapsulation efficiency of dipropionate of beclomethasone, a higly lipophilic drug, was found equal to 98% [7]. But, further factors may affect this value such as the total phospholipid amount, the chain length, the bilayer additives (e.g. cholesterol, tocopherol), and the bilayers properties (gel or fluid state) [15]. 3.1.2. Niosomes 3.1.2.1. Caffeine loaded-niosomes. Niosomes are colloidal systems similar to liposomes in terms of structure, physical properties, and preparation methods. They offer advantages over liposomes, such as better stability, lower costs of the excipients, and larger range of surfactant classes available for the preparation [16,17]. In this section, niosomes were prepared using the syringepump device. The process parameters used for the ethanol injection method were taken from a previous study [18]. The organic to aqueous phase volumes ratio was set to 1/9 (2 ml organic phase and 18 ml aqueous phase). The parameters used in the various formulations as well as the characteristics of the prepared niosome (mean size, PdI, zeta potential and encapsulation efficiency) are summarized in Table 3. From previous publication, Tween® 60, compared to other surfactants, was found to give the highest encapsulation efficiency of a water soluble local antibiotic gentamicin sulphate [19]. It was also suggested that the length of the alkyl chain is a crucial factor of membrane permeability which clearly affected entrapment efficiency [17]. Accordingly, vesicles formulated using Tween® 60 (C18 stearyl chain) offer higher encapsulation efficiencies compared to surfactants such as Span 60® (C16 chain) and Brij® 35 (C12 lauryl chain). Otherwise, high encapsulation efficiency of Span® 60 based niosomes was observed with several API: calcein [20], colchicine [17], and ciclopirox olamine [18]. Hence, in this section, Tween® 60 and Span® 60 were selected for niosome production. A molar ratio of 1:1 of surfactant to cholesterol is usually included in most formulations for the formation of stable niosomes [17,19,20]. Incorporation of cholesterol at increasing amounts eventually causes modification to absence of gel-to-liquid phase transition of lipid bilayers. Therefore, cholesterol alters the fluidity of bilayers chains and increases the orientational order degree leading to a decreasing permeability. DCP, a charged molecule, is usually added to prevent niosome aggregation and improve the preparation stability by increasing surface charge of the vesicles [19,21]. A molar ratio of surfactant-cholesterol-DCP of 45:45:10 was selected according to a previous study [21]. Niosome mean size was comprised between 81 and 127 nm. The largest size was obtained using Span 60 as surfactant. The DCP, which is used to increase the niosome stability, did not affect the niosome size. The PdI was found between 0.12 and 0.21 indicating a rather uniform size distribution. The zeta potential values were

T.T. Pham et al. / Colloids and Surfaces B: Biointerfaces 94 (2012) 15–21

19

Table 3 Niosomes prepared by the syringe-pump device: quantitative composition and characteristics. Surfactant concentration (mM)

Cholesterol concentration (mM)

DCP concentration (mM)

Caffeine Spa-Ca Tw-Ca Tw-DCP-Ca

Span 60 Tween 60 Tween 60

105 105 105

105 105 105

– – 23.3

Spironolactone Tw-Spi Tween 60 Tw-DCP-Spi Tween 60

105 105

105 105

– 23.3

comprised between −23 and −54 mV, which indicated stable colloidal dispersion. The lowest values were measured when DCP was added to the formulations, as DCP is a negatively charged molecule. The same effect of DCP was reported for ciclopirox olamine-loaded niosomes [18]. On the other hand, the encapsulation efficiency of caffeine within niosomes ranged from 2.5% to 9.7%. It can be noted the E. E. within niosomes and liposomes were therefore in the same range. However, Tween 60 showed higher E.E. values; for this reason this surfactant was selected for the next experiments. Then, the presence of DCP also led to an increase in the caffeine E.E. The encapsulation of a water-soluble fluorescent marker (calcein) within niosomes gave highest encapsulation efficiency which was about 6% [20]. Moreover, highest caffeine encapsulation efficiencies within niosomes (30.4% and 81.7%) have been also reported; in this case authors applied the lipid hydration method for the niosome preparation which resulted on larger vesicle size (6–22 ␮m) [22]. Then, this larger size could explain the important E.E. values as it is usually recognized that it significantly increased with particles size [20]. 3.1.2.2. Spironolactone loaded-niosomes. Spironolactone loadedniosomes were prepared using the syringe-pump device. The formulations parameters and the niosome main characteristics are given in Table 3. Spironolactone loaded-niosomes had larger mean size and higher zeta potential (−47 mV) compared to caffeine loaded-niosomes. In addition, the encapsulation efficiency was higher for spironolactone (95.6%) than caffeine (9.7%). Although no phospholipd is present in the niosome formulations, it could be suggested that lipophilic drugs are also highly entrapped in niosome bilayers which may be explained by the particularity of niosomes bilayers structure including surfactant and cholesterol. Otherwise, the lower encapsulation efficiency was found equal to 70.1%, for niosomes prepared in the absence of DCP; which is similar to results observed in the previous section with caffeine loaded-niosomes. Thus, it can be concluded that niosomes offer higher E.E. (95.6%) than liposomes (86.4%). Similar results were previously reported in a comparative study of aceclofenac E.E. between liposome and niosome formulations. Aceclofenac niosomes prepared from Span 60:Cholesterol (7:6 molar ratio) showed higher entrapment efficiency than PC:Cholesterol liposomes at the same molar ratio. This result was explained by the solubilizing effect of the surfactant within niosomes membrane [25]. 3.2. Pilot scale batches using a membrane contactor module For liposomes scale-up, an optimized formulation was selected from previous batches and the ingredient amounts were multiplied by 25 (Table 4). For caffeine loaded liposome, L40Ch30Ca formulation was selected to be prepared using the membrane contactor device. The resulted mean size was 90 nm and the encapsulation efficiency 11% (compared to 89 nm and 6.9% using the syringe device). In addition,

Mean size ± S.D. (nm)

PdI ± S.D.

10 10 10

127 ± 4 89 ± 5 83 ± 3

0.19 ± 0.01 0.12 ± 0.01 0.12 ± 0.01

−23 ± 1.1 −25 ± 0.7 −54 ± 0.5

3 3

81 ± 5 112 ± 3

0.14 ± 0.02 0.21 ± 0.01

−29 ± 0.8 −47 ± 0.6

APIConcentration (mg/ml)

Zeta potential ± S.D. (mV)

E.E. ± S.D. (%)

2.5 ± 0.1 4.0 ± 0.2 9.7 ± 0.3 70.1 ± 2 95.6 ± 3

the PdI, zeta potential, and size distribution curve were very similar to those obtained with the syringe device (Fig. 2A). For spironolactone loaded-liposome, the formulation L40Ch30Spi was selected. The mean size was 93 nm and the encapsulation efficiency was 87.9% (compared to 89 nm and 86.4% with the syringe device). These results confirmed that the preparation characteristics obtained by the membrane module were very similar to those obtained with the syringe device although the volumes had been multiplied by 25. The mean dispersed phase flowrate was evaluated to 375 ml/mn during the experiment (250 ml of dispersed phase passed in 40 s). The flux was not measured versus time, thus no exact fouling data were available. However, this high flow rate suggested that

(A)

25

20

Number (%)

Surfactant

L40Ch30Ca (SP) L40Ch30Ca (MC)

15

10

5

0

0

100

200

300

400

Vesicle size (nm)

(B)

25

20

Number (%)

Formulation

Tw-DCP-Ca (SP) Tw-DCP-Ca (MC)

15

10

5

0

0

100

200

300

400

500

600

Vesicle size (nm) Fig. 2. Size distribution profiles of vesicles prepared using the syringe-pump (SP) device and the membrane contactor (MC): (A) caffeine-loaded liposomes and (B) caffeine-loaded niosomes.

20

T.T. Pham et al. / Colloids and Surfaces B: Biointerfaces 94 (2012) 15–21

Table 4 Liposomes prepared using the membrane contactor: quantitative composition and characteristics. Formulation

Phospholipid concentration (mg/ml)

Cholesterol concentration % (w/w PL)

Caffeine L40Ch30Ca (MC)

40

20

Spironolactone L40Ch30Spi (MC)

40

20

Mean size ± S.D. (nm)

PdI ± S.D.

10

90 ± 3

0.18 ± 0.01

−26 ± 1.2

3

93 ± 6

0.17 ± 0.01

−20 ± 0.6

API concentration (mg/ml)

Zeta potential ± S.D. (mV)

E.E. ± S.D. (%)

11 ± 0.3 87.9 ± 2

Table 5 Niosomes prepared using the membrane contactor: quantitative composition and characteristics. Formulation

Surfactant concentration (mM)

Cholesterol concentration (mM)

DCP concentration (mM)

Caffeine Tw-DCP-Ca (MC)

105

105

23.3

Spironolactone Tw-DCP-Spi (MC)

105

105

23.3

Mean size ± S.D. (nm)

PdI ± S.D.

10

111 ± 4

0.23 ± 0.01

−53 ± 0.5

3

115 ± 5

0.21 ± 0.01

−54 ± 0.7

API concentration (mg/ml)

very low fouling occurred, probably because the dispersed phase was mainly constituted by ethanol. Larger quantities of liposome batches could then be prepared. For niosomes, formulations were selected from the syringe preparation study to be scaled-up with the membrane contactor. The amounts and volumes were multiplied by 50 (Table 5). Therefore, the aqueous and organic phases used were equal to 900 and 100 ml, respectively. The formulations incorporating DCP were selected as they gave the higher encapsulation efficiencies. For caffeine loaded-niosomes, we selected the formulation TwDCP-Ca. Using the membrane contactor, the mean size and the encapsulation efficiency were found equal to 111 nm and 9.1%,

Zeta potential ± S.D. (mV)

E.E. ± S.D. (%)

9.1 ± 0.2 94.7 ± 3

respectively. With the syringe device, the mean size and the encapsulation efficiency were 83 nm and 9.7%, respectively. In addition, both preparations showed similar PdI, zeta potential, and size distribution curves (Fig. 2B). In addition, the Tw-DCP-Spi spironolactone loaded-niosome preparation was selected for scale-up. The mean size and encapsulation efficiency obtained with the membrane contactor, 115 nm and 94.7%, respectively, were again very close to the data obtained with the syringe device (112 nm and 95.6%, respectively). For both preparations, the mean flowrate during the experiment was 200 ml/min (the dispersed phase volume of 100 ml passed in nearly 30 s). Like for liposome preparation, these high flowrates

Fig. 3. Transmission electron microscopy images of vesicles: (A and B) liposomes and (C and D) niosomes.

T.T. Pham et al. / Colloids and Surfaces B: Biointerfaces 94 (2012) 15–21

suggested low fouling of the membrane as the dispersed phase was mainly constituted of ethanol. These results confirm that the membrane contactor is a suitable technique for niosome preparation at large scale. 3.3. TEM morphology observation Some preparations of liposomes and niosomes were observed by TEM (Fig. 3). Liposome images are given in Fig. 3A and B. As can be seen, liposomes were spherical shaped with multilayered membrane structure. Their size was estimated from TEM pictures at the range of 50–200 nm; these data were coherent with DLS measurements. Fig. 3C and D showed niosomes with a spherical structure and a mean size comprised between 50 and 100 nm. These values were also in the range of size evaluated by DLS. 4. Conclusion In this study, liposomes and niosomes were prepared by the ethanol injection technique. Several formulations were first prepared using a syringe device which required small volumes of both aqueous and organic phases. For liposome preparation, 20 ml of aqueous phase was used and 10 ml of organic phase. For niosome preparation, 18 ml was used for the aqueous phase and 2 ml for the organic phase. Then optimum formulations were selected to be prepared using membrane contactor, batch volumes have been multiplied by 25 and 50, for liposome and niosome preparations, respectively. A specific attention was devoted to the encapsulation efficiencies within liposomes and niosomes. For this aim, two model drugs were selected; caffeine and spironolactone. The encapsulation efficiency of caffeine was found around 10% and was probably the result of its low lipophilicity. Small amounts of caffeine are located inside the core and the bilayer of the vesicles. Higher encapsulation efficiencies were obtained for spironolactone, which is probably due to its better lipophilic properties. Higher amounts of this drug are then located inside the bilayers of liposomes or niosomes. Once optimized formulations were obtained with the syringe device, the preparations were scaled-up using a membrane technique. The dispersed phase is passed through the membrane pores and the vesicles (either liposomes or niosomes) are formed in the aqueous phase which circulates tangentially to the membrane surface. It was shown that the characteristics of the preparations were very similar using the syringe device and the membrane contactor. These results confirm that the membrane device is an attractive alternative for the preparation of liposomes and niosomes by the ethanol injection technique. The method is simple to perform and can be scaled-up further by using larger membrane area or membrane devices in parallel or in series. References [1] G. Betageri, M. Habib, Liposomes as drug carriers, Pharm. Eng. 14 (1994) 76–77.

21

[2] H. Schreier, J. Bouwstra, Liposomes and niosomes as topical drug carriers: dermal and transdermal drug delivery, J. Control. Release 30 (1994) 1–15. [3] J.Y. Fang, C.T. Hong, W. Chiu, Y.Y. Wang, Effect of liposomes and niosomes on skin permeation of enoxacin, Int. J. Pharm. 219 (2001) 61–72. [4] M. Tabbakhian, N. Tavakoli, M.R. Jaafari, S. Daneshamouz, Enhancement of follicular delivery of finasteride by liposomes and niosomes: 1. In vitro permeation and in vivo deposition studies using hamster flank and ear models, Int. J. Pharm. 323 (2006) 1–10. [5] K. Ruckmani, B. Jayakar, S.K. Ghosal, Nonionic surfactant vesicles (Niosomes) of cytarabine hydrochloride for effective treatment of leukemias: encapsulation, storage and in vitro release, Drug Dev. Ind. Pharm. 26 (2000) 217–222. [6] C. Charcosset, Membrane processes in biotechnology: an overview, Biotechnol. Adv. 24 (2006) 482–492. [7] C. Jaafar-Maalej, R. Diab, V. Andrieu, A. Elaissari, H. Fessi, Ethanol injection method for hydrophilic and lipophilic drug-loaded liposome preparation, J. Liposome Res. 20 (2010) 228–243. [8] A. Laouini, C. Jaafar-Maalej, S. Sfar, C. Charcosset, H. Fessi, Liposome preparation using a hollow fiber membrane contactor – application to spironolactone encapsulation, Int. J. Pharm. 415 (2011) 53–61. [9] A. Gabelman, S.T. Hwang, Hollow fiber membrane contactors, J. Membr. Sci. 159 (1999) 61–106. ´ M. Shimizu, T. Nakashima, Permeability of hydrophilic and [10] G.T. Vladisavljevic, hydrophobic Shirasu-porous-glass (SPG) membranes to pure liquids and its microstructure, J. Membr. Sci. 250 (2005) 69–77. ´ I. Kobayashi, M. Nakajima, R.A. Williams, M. Shimizu, T. [11] G.T. Vladisavljevic, Nakashima, Shirasu porous glass membrane emulsification: characterization of membrane structure by high-resolution X-ray microtomography and microscopic observation of droplet formation in real time, J. Membr. Sci. 302 (2007) 243–253. [12] C. Jaafar-Maalej, C. Charcosset, H. Fessi, A new method for liposome preparation using a membrane contactor, J. Liposome Res. 21 (2011) 213–220. [13] A. Sze, D. Erickson, L. Ren, D. Li, Zeta-potential measurement using the Smoluchowski equation and the slope of the current-time relationship in electroosmotic flow, J. Colloid Interface Sci. 261 (2003) 402–410. [14] A.M. Fadda, B.M. Baroli, A.M. Maccioni, C. Sinico, D. Valenti, F. Alhaique, Phospholipid-detergent systems: effects of polysorbates on the release of liposomal caffeine II, Farmaco 53 (1998) 650–654. [15] S.B. Kulkarni, G.V. Betageri, M. Singh, Factors affecting microencapsulation of drugs in liposomes, J. Microencapsul. 12 (1995) 229–246. [16] I.F. Uchegbu, S.P. Vyas, Non-ionic surfactant based vesicles (niosomes) in drug delivery, Int. J. Pharm. 172 (1998) 33–70. [17] Y. Hao, F. Zhao, N. Li, Y. Yang, K. Li, Studies on a high encapsulation of colchicine by a noisome system, Int. J. Pharm. 244 (2002) 73–80. [18] K.S. Shaikh, B. Chellampillai, A.P. Pawar, Studies on non-ionic surfactant bilayer vesicles of ciclopirox olamine, Drug Dev. Ind. Pharm. 36 (2010) 946–953. [19] G. Abdalbary, N. El-gendy, Niosome-encapsulated gentamicin for ophthalmic controlled delivery, AAPS Pharm. Sci. Technol. 9 (2008) 740–747. [20] A. Manosroi, P. Wongtrakul, J. Manosroi, H. Sakai, F. Sugawara, M. Yuasa, M. Abe, Characterization of vesicles prepared with various non-ionic surfactants mixed with cholesterol, Colloids Surf. B 30 (2003) 129–138. [21] Z.S. Bayindir, N. Yuksel, Characterization of niosomes prepared with various nonionic surfactants for paclitaxel oral delivery, J. Pharm. Sci. 99 (2010) 2049–2060. [22] P. Khazaeli, A. Pardakhty, H. Shoorabi, Caffeine-loaded niosomes: characterization and in vitro release studies, Drug Deliv. 14 (2007) 447–452. [24] C. Hu, D.G. Rhodes, Proniosomes:, A novel drug carrier preparation, Int. J. Pharm. 185 (1999) 23–35. [25] M. Nasr, S. Mansour, N.D. Mortada, A.A. Elshamy, Vesicular aceclofenac systems: a comparative study between liposomes and niosomes, J. Microencapsulation 25 (2008) 499–512. [26] E. Touitou, F. Levi-Schaffer, N. Dayan, F. Alhaique, F. Riccieri, Modulation of caffeine skin delivery by carrier design: liposomes versus permeation enhancers, Int. J. Pharm. 103 (1994) 131–136. [27] S. Batzri, E.D. Korn, Single bilayer liposomes prepared without sonication, Biochim. Biophys. Acta 298 (1973) 1015–1019. [28] A. Wagner, K. Vorauer-Uhl, G. Kreismayr, H. Katinger, The crossflow injection technique: an improvement of the ethanol injection method, J. Liposome Res. 12 (2002) 259–270.