- Email: [email protected]
Memory effects of monolayers and vesicles formed by the non-ionic surfactant, 2C18E12
Journal of Colloid and Interface Science 316 (2007) 741–750 www.elsevier.com/locate/jcis
Memory effects of monolayers and vesicles formed by the non-ionic surfactant, 2C18E12 D.J. Barlow, C.M. Hollinshead, R.D. Harvey, L. Kudsiova, M.J. Lawrence ∗ Department of Pharmacy, Franklin Wilkins Building, King’s College London, 150 Stamford Street, London SE1 9NH, UK Received 20 April 2007; accepted 6 August 2007 Available online 11 August 2007
Abstract The behaviour of monolayers and bilayers formed by the dialkyl chain non-ionic surfactant, 1,2-di-O-octadecyl-rac-glycerol-3-ω-methoxydodecaethylene glycol (2C18 E12 ) in water at 297 K has been investigated. Using a surface film balance (or Langmuir trough) the compression– expansion cycle of the 2C18 E12 monolayer was found to be reversible when compressed to surface pressures (π ) less than 42 mN m−1 . Compression of 2C18 E12 monolayer to π greater than 42 mN m−1 above this resulted in a considerable hysteresis upon expansion with the π remaining high relative to that obtained upon compression, suggesting a time/pressure dependent re-arrangement of 2C18 E12 molecules in the film. Morphology of the 2C18 E12 monolayer, investigated using Brewster angle microscopy, was also found to depend upon monolayer history. Bright, randomly dispersed domains of 2C18 E12 of approximately 5 µm in size were observed during compression of the monolayer to π less than 42 mN m−1 . At π of 42 mN m−1 and above, the surfactant film appeared to be almost completely ‘solid-like.’ Regardless of the extent of compression of the monolayer film, expansion of the film caused formation of chains or ‘necklaces’ of individual surfactant domains, with the extent of chain formation dependent upon pressure of compression of the monolayer and the length of time held at that pressure. Irreversible effects on 2C18 E12 vesicle size were also seen upon temperature cycling the vesicles through their liquid–crystalline phase transition temperature with vesicles shrinking in size and not returning to their original size upon standing at 298 K for periods of more than 24 h. No comparable hysteresis, time, pressure or temperature effects were observed with the monolayer or vesicles formed by the corresponding phospholipid, disteaorylphosphatidylcholine, under identical conditions. The effects observed with 2C18 E12 are attributed to the ability of the polyoxyethylene head group to dehydrate and intrude into the hydrophobic chain region of the mono- and bilayers. These studies have important implications for the use of the vesicles formed by 2C18 E12 as drug delivery vehicles. © 2007 Elsevier Inc. All rights reserved. Keywords: Non-ionic surfactant; Monolayers; Vesicles; Langmuir isotherm; Brewster angle microscopy; Variable temperature turbidity measurements
1. Introduction In order rationally to design new surfactants to produce vesicles with suitable properties for drug delivery it is necessary to develop an understanding of the relationship between surfactant molecular structure and aggregate architecture and behaviour. The most informative structural studies performed to date have involved the use of neutron scattering experiments and in particular neutron specular reflection and small angle neutron scattering measurements, both used in combination with * Corresponding author.
E-mail addresses: [email protected], [email protected] (M.J. Lawrence). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.08.014
isotopic substitution. By these means it has been possible to determine the detailed structure of the monolayer of the nonionic polyoxyethylene surfactant, 2C18 E12 at the air/water interface both in the presence and absence of cholesterol and the corresponding phospholipid, disteaorylphosphatidylcholine (DSPC) [1–3] as well as the molecular architecture of the vesicles formed from this surfactant [4,5]. The data obtained from these studies were used to rationalise the disappointingly low encapsulation efficiency of the vesicles formed by 2C18 E12 and guided the re-formulation of the vesicles using added cholesterol and DSPC to give considerably increased encapsulation efficiency [6]. During these studies, however, it was clear that the monolayers and bilayers (vesicles) formed by 2C18 E12 did not behave in a reversible manner demonstrating considerable
742
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
time/pressure dependent properties. In the studies reported here we use a combination of techniques to explore this behaviour with a view to understanding the potential of the vesicles formed by 2C18 E12 as drug delivery vesicles.
The non-ionic surfactant, 1,2-di-O-octadecyl-rac-glycerol3-ω-methoxydodecaethylene glycol (2C18 E12 ) was synthesised and characterised according to Lawrence et al. [7]. Disteaorylphosphatidylcholine (DSPC) was purchased from Avanti Polar Lipids (USA). Chloroform and ethanol of HPLC-grade quality were supplied by Rathburn Chemicals Ltd. (UK). Water was firstly doubly distilled and purified using an Elgastat Maxima purifier (Elga, UK) to a resistance of at least 18 M. All chemicals were of the highest grade available and used as supplied.
and to reduce airborne contamination. P-polarised light (supplied by an NDYAG laser operating at a wavelength of 532 nm) was emitted by the BAM onto the water surface at an angle of incidence of the Brewster angle for water and the light reflected from the surface collected by two achromatic lenses and detected with a CCD camera. The CCD camera converted the reflectivity signal from the sample into a video image. Calibrations were performed by obtaining images of the bare water surface prior to spreading the monolayer. These ‘background images’ were subtracted from the sample images taken subsequently in order to obtain sharper pictures of the monolayer films. The spatial resolution of the BAM is approximately 2 µm. Two different magnifications were employed during the study, namely a low magnification using a ×10 objective where the images obtained represented a sample size of 430 µm (horizontal field of view) × 530 µm (vertical field of view) and a higher magnification (×20 objective), where the images represented a sample of size 215 µm × 265 µm.
2.2. Langmuir film
2.4. Vesicle preparation
The Langmuir trough (Nima Technology Ltd., Coventry, UK) was scrupulously cleaned by first soaking in neutral Decon, thereafter repeatedly rinsing with ultra-pure water and finally wiping the trough’s inner surfaces with first chloroform, then ethanol and then chloroform again. The Wilhelmy plate was a piece of filter paper which, prior to attachment to the Langmuir trough’s microbalance, was degreased by soaking in chloroform, and dried by allowing the solvent to evaporate. The trough was filled with water and the surface of the water cleaned of any surface active impurities by sweeping the Langmuir trough’s surface with its moveable Teflon barrier followed by suction using a Pasteur pipette attached to a vacuum pump. The temperature of the aqueous sub-phase was controlled at 297 ± 0.1 K by water circulation from a thermostat. Once the surface of the water was clean, a small aliquot (typically between 10 to 100 µL) of a 2 mg/mL chloroform solution of 2C18 E12 or DSPC was carefully added dropwise to the surface using a Hamilton microsyringe. 20 min was allowed for the chloroform to evaporate before the film was compressed. Isotherms were obtained by compressing and expanding the film at a rate of 10 mm min−1 . For the Brewster angle microscope studies, the barrier was moved to the desired surface pressure and held at that surface pressure while the monolayer was visualised. This process was repeated a number of times using the same monolayer and with freshly prepared monolayer films.
Vesicles were prepared by hydration of a thin film of 2C18 E12 or DSPC (25 mg deposited from evaporation of a chloroform solution). Hydration of the surfactant film was achieved by continuous agitation with 5 mL of solvent at 338 K for 30 min (>10 K above the Tc of the 2C18 E12 or DSPC) [8,9]. The resultant suspension was sonicated in a Decon bath sonicator for 60 min and then ultrasonicated for a further 15 min at 328 K using a Lucas Dawes probe sonicator fitted with a tapered microtip (operating at 15% of its maximum output). Any titanium particles shed from the probe sonicator were removed by centrifugation. The sizes of the vesicles thus prepared were assessed (after appropriate dilution) by laser light scattering using a Malvern Autosizer fitted with a 10 mW helium–neon laser. Dilution of the vesicles for size determination by light scattering was necessary to avoid inter-vesicular interactions giving a false estimate of vesicle size. Dilution of the vesicles was not considered to alter the level of surfactant hydration [2,4].
2. Experimental 2.1. Materials
2.3. Brewster angle microscopy (BAM) The BAM (BAM2plus, Nanofilm Technologie, GmbH, Goettingen, Germany) was mounted above the Langmuir trough, to allow for collection of images of the morphology of the 2C18 E12 monolayer. The Langmuir trough (monolayer prepared as described above) and the BAM were both housed in a cabinet to minimise disruption to the monolayer by air currents
2.5. Turbidity studies Vesicle suspensions of 2C18 E12 or DSPC for spectrophotometric measurement were prepared at a concentration of 5 mg mL−1 . The vesicle suspensions were transferred to a standard 1-cm fused silica cuvette and were incubated at their respective starting temperatures for 30 min prior to measurement. All turbidity measurements were recorded as absorbance at 500 nm (Aapp (500)) using a Perkin Elmer Lambda 5 spectrophotometer, with slit widths of 0.2 nm. The spectrophotometer was equipped with a Peltier thermostatic cell holder, controlled by a PTP-1 Peltier temperature programmer (accuracy ±0.1 K). The temperature within the cell was monitored using a Comark thermocouple positioned at approximately half the depth of the vesicle suspension. Absorbance (Aapp (500)) was measured, during repeated heating and cooling cycles carried
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
out at a rate of 0.6 K min−1 (every 1 K). During each experiment, three heating and cooling cycles were carried out, with 30 min temperature equilibration at both the upper and lower temperatures prior to the beginning of a heating or cooling run. The non-ionic 2C18 E12 vesicles were heated and cooled within the range of 303–328 K while DSPC vesicles were examined over the range 313–333 K. The optical density versus heating curve thus obtained was converted to a simple first derivative by taking the difference in the absorbance (Aapp (500)) between adjacent points using the Galactic Industries Grams/32 AI (version 6.0) computer program.
743
3.1.1. π–A isotherms 2C18 E12 is a double-chained surfactant that forms an insoluble film at the air–water interface at 297 K. Fig. 1 shows the π–A isotherm obtained for 2C18 E12 after a single compression cycle but derived from the mean of individual measurements on three films. The π–A isotherm exhibited by 2C18 E12 was extremely reproducible (with a standard deviation in surface pressure of less than 0.5 mN m−1 ) and was found to be independent of spreading volume (i.e. number of molecules deposited on the surface) when added to surface pressures of less than ∼40 mN m−1 . As can be seen from Fig. 1, the π–A isotherm obtained for 2C18 E12 exhibits three distinct regions suggesting a likely conformational change/re-arrangement of the 2C18 E12 molecules within the film upon compression. Up to surface pressures of 42 mN m−1 , the 2C18 E12 film is expanded with a non-zero surface pressure being detected even at very large areas per molecule. Furthermore the film is well behaved in that it was possible to maintain surface pressure and area steady for periods of at least several hours with no hysteresis of the π–A isotherm being observed upon subsequent expansion of the film. Both of these observations suggest no loss of 2C18 E12 to the aqueous sub-phase as has been observed for films prepared from much longer chain pegylated distearoylphosphatidylethanolamines (DSPE), such as DSPE-PEG45 and DSPE-PEG110 with DMPE [10]. In this context it should be noted that Kuhl et al. [11] have reported that DSPE-PEG90 , DSPE-PEG350 and DSPC-PEG750 all form stable monolayers
when present as sole component, although it has been reported that DSPE-PEG750 forms micelles when present at greater than 60 mol% [12] and therefore some hysteresis of the film might be anticipated. In contrast DSPE-PEG350 forms stable bilayers at 100 mol% [12]. A surface pressure of 42 mN m−1 (πc ) marks the start of a large plateau region similar in appearance to the expanded/condensed co-existence region reported for monolayers of pegylated distearoylphosphatidylethanolamines [11] albeit at far higher surface pressures than those recorded DSPE-PEG350 and DSPC-PEG750 (both of which show plateau regions commencing at surface pressures of ∼20 mN m−1 [11,13]. Upon further compression of the 2C18 E12 film to an area per 2C18 E12 molecule of about ∼40 Å2 , the surface pressure exerted by the film suddenly rises and the film collapses at a surface pressure of around 65 mN m−1 and a limiting area per molecule (determined from the intercept of the extrapolated steep rising part 2 of the isotherm with the abscissa) of 37.4 ± 1.3 Å , which is anomalous as it is less than the cross sectional area of two hydrocarbon chains. Upon expansion of a 2C18 E12 film previously compressed to a surface pressure of greater than πc but less than the collapse pressure of ∼65 mN m−1 , a large hysteresis was observed in the π–A isotherm with the surface pressure of the film remaining high at large areas per 2C18 E12 molecule. This hysteresis persisted for at least several hours and was variable in its extent, being dependent upon the compression pressure above πc . Again there was no evidence of loss of 2C18 E12 to the aqueous sub-phase. Fig. 2 shows the isotherm obtained for DSPC under exactly the same conditions as those used for 2C18 E12 . The isotherm obtained here for DSPC is in agreement with those reported for this lipid in the literature [14–16]. There are clear differences in the isotherms obtained for DSPC and 2C18 E12 , the most noticeable difference being that the DSPC isotherm does not show any distinct phase transition, with the exception of the transition from a gaseous phase, which was present at pressures below 1 mN m−1 , to the liquid expanded phase. The film was seen to collapse at a surface pressure of greater than 60 mN m−1 . At pressures below the collapse pressure no hysteresis was observable. Again the film was well behaved in that it was possible to hold the film without significant change in its properties for at least several hours.
Fig. 1. Surface pressure—isotherm of 2C18 E12 on a sub-phase of water at 298 K.
Fig. 2. Surface pressure—isotherm of DSPC on a sub-phase of water at 298 K.
3. Results 3.1. Monolayer studies
744
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
(a)
(b)
(c)
(d)
(e) Fig. 3. BAM images of monolayers of 2C18 E12 on a sub-phase of water at 298 K and held at a surface pressure of (a) 15 mN m−1 , (b) 20 mN m−1 , (c) 25 mN m−1 , (d) 30 mN m−1 and (e) 40 mN m−1 (×10 objective).
3.1.2. BAM images BAM studies were performed to visualise the film formed by 2C18 E12 at the air/water interface at a range of surface pressures. Fig. 3 shows a series of representative pictures of the 2C18 E12 film upon initial compression up to and including a π of 40 mN m−1 , using a ×10 objective. As can be seen the 2C18 E12 film is clearly visualized as bright, welldefined domains of approximately circular shape (even when examined under a ×20 objective, data not shown) and of similar size (∼5 µm). These circular domains were fairly evenly spaced over the surface of the water. Although Tsukanova and Salesse [17] have described the formation of similar looking domains of DSPE-PEG5000 molecules spread on a water subphase at a relatively low surface pressure of 13.6 mN m−1
using BAM Kuhl et al. [11] reported the absence of any domain formation or structuring in films formed by DSPE-PEG90 , DSPE-PEG350 and DSPE-PEG750 . (Note here that the polyoxyethylene head groups in the surfactant used in the present study approximately correspond to PEG500 .) In the present study, compression of the 2C18 E12 film only served to pack the circular domains closer together. At a surface pressure of 40 mN m−1 an almost complete, ‘solid-like’ surfactant film was observed. At surface pressures below πc , there were small areas of the film in which the domains seemed to have annealed into chains (or ‘necklaces’) rather than being regularly spaced; such organisation is seen in Fig. 4a, in a film compressed to 15 mN m−1 . This annealing had the result that some parts of the water sur-
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
(a)
(a)
(b)
(b)
745
Fig. 4. BAM images of monolayers of 2C18 E12 on a sub-phase of water at 298 K and held at a surface pressure of 15 mN m−1 (a) ×10 objective and (b) ×20 objective.
(c)
Fig. 5. BAM image of a monolayer of 2C18 E12 on a sub-phase of water at 298 K and held at a surface pressure of 45 mN m−1 (×20 objective).
face were quite bare of surfactant whilst other areas were more densely packed with ‘necklaces’ of surfactant domains. Upon careful examination of the domains at higher magnification (Fig. 4b) it was clear that the individual domains comprising these necklaces were still approximately circular and of about 5 µm diameter. It must be stressed, however, that the regions of the monolayer in which these chains were observed represented only a very small proportion (<5%) of the water surface. Fig. 5 shows a 2C18 E12 film compressed to beyond πc (to π 45 mN m−1 ). Here again the surfactant film appears quite close-packed, although it is difficult to conclude anything about the ‘state’ of the 2C18 E12 molecules comprising the film in this
Fig. 6. BAM image of monolayers of 2C18 E12 on a sub-phase of water at 298 K at a surface pressure of (a) 36 mN m−1 , (b) held at a surface pressure of 36 mN m−1 for 45 min and (c) compressed to 40 mN m−1 after being held at a surface pressure of 36 mN m−1 for 45 min ×10 objective).
region since the images look no different from those obtained for films compressed to a surface pressure less than 42 mN m−1 except that the surface of the film does not seem as uniformly smooth as those seen in Fig. 3, suggesting perhaps that there are inhomogeneities in the thickness of the film compressed to 45 mN m−1 . During the BAM studies it became clear that there were both time and pressure effects on the appearance of the film, even at surface pressures below πc . Fig. 6 demonstrates the effects of time on the 2C18 E12 film when compressed to π less than πc . In Fig. 6a, the monolayer is seen upon initial compression to a surface pressure of 36 mN m−1 while Fig. 6b shows the same film after being held at that surface pressure for 45 min.
746
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
(a)
(a)
(b)
(b)
Fig. 7. BAM image of monolayers of 2C18 E12 on a sub-phase of water at 298 K (a) compressed to 34 mN m−1 and re-opened to 30 mN m−1 and (b) compressed to 34 mN m−1 then re-opened to 19 mN m−1 (×20 objective).
(It should be noted that these images were chosen as they appeared to be representative of the entire surface.) It can be seen that over time, the film of evenly spaced domains does change in appearance with the domains merging to form ‘necklaces,’ and leading to some patches of water sub-phase appearing to be entirely devoid of surfactant molecules. Fig. 6c shows the same monolayer as in Fig. 6b after being compressed to 40 mN m−1 where the film appears to be in a similar solid state to that obtained upon direct compression of the film to 40 mN m−1 (Fig. 3). Fig. 7 shows images of a 2C18 E12 film, which was first compressed to 40 mN m−1 (less than πc ) and then a ‘reverse isotherm’ obtained (i.e. a gradual expansion of the barriers) to 30 mN m−1 (Fig. 7a) or to 19 mN m−1 (Fig. 7b). In these images, the domains are annealed as shown in Fig. 6. Again these images were selected as they were representative of the entire surface and there were no homogeneous regions of isolated domains similar to those present at similar pressures as shown in Fig. 3. These figures provide visual evidence of a hysteresis in the morphology of the film, in that upon compression, the domains become strongly annealed and do not separate again upon decompression, although it should be noted that under these conditions no hysteresis of the monolayer was observed via the film balance experiments. Images of the DSPC monolayer at the air/water interface, obtained using Brewster angle microscopy at a range of sur-
(c) Fig. 8. BAM image of monolayers of DSPC on a sub-phase of water at 298 K and held at a surface pressure of (a) 0.5 mN m−1 , (b) 17 mN m−1 and (c) 50 mN m−1 (×10 objective).
face pressures are shown in Fig. 8. As can be seen from Fig. 8 the images obtained for DSPC were quite different from those observed for 2C18 E12 . Fig. 8a shows the image of the DSPC monolayer at a surface pressures 0.5 mN m−1 when the film was in its gaseous state according to its surface pressure. Already, however, domains of condensed phase are visible on the sub-phase, although the film was very fluid and the domains (which are smaller than those observed with 2C18 E12 ) therefore hard to image. Upon compression, more domains were clear and the mobility of the film decreased. Similar images have been obtained for DSPC when spread on low concentrations of electrolyte [18,19]. No hysteresis of the DSPC film was observed upon expansion of the film.
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
747
3.2. Vesicles 3.2.1. Turbidity studies In common with phospholipid vesicles, upon first heating of the 2C18 E12 vesicles through their phase transition (i.e. from 303 K up to a temperature of 328 K), two large decreases in absorbance measured at 500 nm (Aapp (500)) were seen to occur at ∼316 and ∼320 K (Fig. 9a). By analogy with phospholipids, it was assumed that the first decrease was attributable to a pre-transition or gel to ripple phase (Tpre ), while the second was due to the liquid–crystalline phase transition (Tc ) [20]. The changes observed upon the first heating cycle were not reversible upon cooling as there was no increase in Aapp (500) until ∼316 K, whence there was a relatively rapid increase in Aapp (500) followed by a much slower rise in Aapp (500). However, even at 303 K, the Aapp (500) of the vesicular suspension had not recovered its original value prior to heating. Furthermore although the curves obtained in the second and third heating cycles were superimposable, they were not identical to the first heating curve as the extent of the first decrease in Aapp (500) was very much reduced. The extent of the second Aapp (500) decrease initiated at ∼320 K, was however unaffected by the change in initial turbidity at the start of the heating run. The second and third cooling curves were superimposable on the first. Fig. 9b shows the mean of the three heating and the three cooling curves. Fig. 9c gives the simple derivative of the mean heating and cooling curves shown in Fig. 9b and were used to determine the temperatures of the Tpre and Tc which were found to be 317 ± 0.5 and 322 ± 0.5 K. Upon cooling, the only peak seen was at 315.5 ± 0.5 K. For a population of 2C18 E12 vesicles with mean diameters for which the standard deviation is low, as is the case here, the decrease observed in turbidity upon heating can be reasonably assumed to coincide with a decrease in diameter of the vesicles. Similarly, because upon cooling the vesicles back to the starting temperature, Aapp (500) did not restore the Aapp (500) to its original value, it would seem that the vesicles had not returned to their original size. In order to confirm this apparent reduction in vesicle size after the first heating/cooling cycle, a sizing study was performed before and after the vesicles had been incubated at 333 K for 30 min. As can be seen from Table 1, even after 24 h incubation at 298 K, the vesicles had not returned to their original size. (Vesicle size changes over longer recovery periods were not examined.) This phenomenon may be influenced by both the original size (and lamellarity) of the vesicles. The reduction in size of 2C18 E12 vesicles observed may indicate that the vesicles require annealing for long periods below 303 K prior to use. Fig. 10a shows that for a suspension of DSPC liposomes, there were reproducible decreases in Aapp (500) in response to increasing temperature. There was a small decrease after 323 K and a larger decrease after 326 K. The second decrease in Aapp (500) was reversed upon cooling of the vesicle suspension (Fig. 10b), however, the first drop in Aapp (500) was missing from the cooling cycle curve. The data used in Fig. 10b were converted by calculating a simple difference derivative, into the curves shown in Fig. 10c. The changes in Aapp (500) observed
(a)
(b)
(c) Fig. 9. Temperature dependence of the optical density at 500 nm of 2C18 E12 vesicles in water (a) superimposed results of three heating cycles on the same sample, (b) mean results of 4 heating and cooling cycles and (c) simple derivative curve calculated from the mean heating and cooling curves in (b).
748
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
Table 1 Diameters of 2C18 E12 vesicles, both before and at various times after incubation at 333 K Experimental treatment
Vesicle diameter (nm)
1 h post sonication at 298 K 10 min post 30 min incubation at 333 K 1 h at 298 K 2 h at 298 K 3 h at 298 K 16 h at 298 K 24 h at 298 K
529 ± 14.1 390 ± 9.1 392 ± 9.7 428 ± 8.9 440 ± 17.4 442 ± 15.1 441 ± 9.9
Note. Light scattering measurements were performed at 298 K after appropriate temperature equilibration (n− = 3 ± s.d.).
in DSPC liposome suspensions are thus seen in Fig. 10c as two peaks (at 324 ± 0.5 K and 327 ± 0.5 K, respectively). These values accord well with the values obtained from DSC studies for the Tpre and Tc , these being 322.1 ± 2.9 K and 327.5 ± 1.5 K, respectively [21]. Fig. 10a clearly shows that although the temperatureinduced turbidity changes at the main phase transition of DSPC were reversible, a hysteresis was observed for the pre-transition, during the cooling cycle. This phenomenon has previously been observed for suspensions of DPPC vesicles [20,22–24]. Equilibration of the vesicle suspension 10 K below Tpre was sufficient to ensure recovery from the hysteresis in the pre-transition (Fig. 10a), which is thought to be maintained at elevated equilibration temperatures [23,24]. In contrast to the vesicles formed from 2C18 E12 there are no significant change in the size of the vesicles prepared from DSPC prior to and after three heating/cooling cycles (data not shown).
(a)
4. Discussion Lipid molecules containing relatively short chain hydrophilic polymers such as polyethylene glycol have attracted much interest because of their ability to produce sterically stabilised vesicles suitable for drug delivery and drug targeting applications. Before discussing the results of the present study, it is worth considering earlier studies on related pegylated molecules. Most of the related studies have examined the behavior of poly(ethylene) glycols (PEG) of varying molecular weights (typically 350–5000 D) when conjugated to the terminal amine group of distearoylphosphatidylethanolamine (DSPE). Significantly, these molecules differ from the type investigated in the present study in that the PEG head group is attached to the glycerol backbone via a negatively charged phosphidylethanolamine group, while in 2C18 E12 the PEG chain is directly linked to the glycerol moiety. Of the pegylated molecules studied to date only those containing short PEG chains, namely DSPE-PEG90 , DSPEPEG350 , DSPE-PEG750 have been reported to form ‘stable’ monolayers when used as pure component, whereas lipids with longer pegylated chains such as DSPE-PEG2000 have only been reported to form relatively stable, well-behaved monolayers when combined with the phospholipid, dimyristoylphosphatidylethanolamine (DMPE) [10]. However, unlike the films formed by 2C18 E12 no report exists of the film formed by these
(b)
(c) Fig. 10. Temperature dependence of the optical density at 500 nm of DSPC vesicles in water (a) superimposed results of three heating cycles on the same sample, (b) mean results of 6 heating and cooling cycles and (c) simple derivative curve calculated from the mean heating and cooling curves in (b).
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
short chain pegylated lipid molecules exhibiting any hysteresis. While the π–A isotherm obtained for 2C18 E12 films reported in the present study exhibited some similarity to those recorded for DSPE-PEG350 and DSPE-PEG750 , in that at low surface pressures the film was expanded and underwent a conformational change/re-arrangement at intermediate surface pressures, a number of differences were obvious. In particular the plateau region present in the 2C18 E12 isotherm commenced at a much higher surface pressure of 42 mN m−1 compared to the plateau reported for DSPE-PEG350 and DSPE-PEG750 which began at a surface pressure of approximately 20 mN m−1 . In addition, the limiting area per molecule of DSPE-PEG350 , DSPE-PEG750 , estimated from the π–A isotherms presented by Kuhl et al. [11], 2 was greater than 50 Å2 , much larger than the 37.4 ± 1.3 Å determined here for 2C18 E12 . Interestingly, at a surface pressure of 42 mN m−1 , the films formed by DSPE-PEG350 and DSPE-PEG750 were reported to be in the solid phase [11]. It is clear from the BAM studies presented here that 2C18 E12 is also present as a close packed film at a surface pressure of 42 mN m−1 (i.e. at the start of the plateau region), although at this surface pressure 2C18 E12 exhibits a far greater area per molecule of ∼110 Å2 (determined from neutron reflectivity studies [25]) than DSPE-PEG350 and DSPE-PEG750 . Although this value appears to be very large, and much greater than the values obtained for pegylated lipids, experimental data has shown that 2C18 E12 does indeed form vesicles. Once a complete monolayer is formed, any further compression of the film normally leads to a rapid increase in surface pressure of the film and then collapse of the film. Significantly, for 2C18 E12, neutron reflectivity studies suggest that when compressed to areas per molecule of less than ∼110 Å2 , the 2C18 E12 film begins to form islands of bi- or multilayers (as evidenced by the appearance of off-specular neutron reflection) and that furthermore these islands of bi- or multilayers do not disappear upon expansion of the film [25]. The formation of bior multilayers upon compression is unusual and is undoubtedly a consequence of dehydration upon compression of at least, some of the PEG chains of 2C18 E12 coupled with the ability of the dehydrated, methoxy capped polyethylene glycol chains to intrude into and mix with the hydrophobic region of the film; methoxy capped polyoxyethylene glycol is well known to be totally miscible with hydrocarbon solvents. This mixing of the hydrophilic polyoxyethylene head group and the stearoyl chains is most likely the explanation for the very small area per molecule recorded for 2C18 E12 at its collapse. Furthermore the dehydration of the PEG chains upon compression and their subsequent mixing with the stearoyl chains is the likely cause of the hysteresis (or memory effects) seen in the π–A isotherm because in order for the polyoxyethylene chains to re-enter the aqueous phase they must become re-hydrated, a process that may take some time due to the energetically favourable mixing of the hydrophilic methoxy capped polyoxyethylene head group and the stearoyl chains. While some overlap of the hydrophobic and hydrophilic PEG head groups has been previously reported for short pegylated chain lipids when in their solid-like state [11], the extent
749
of overlap might be anticipated to be far less when a charged phosphate group is present coupled with the presence of a free hydroxyl group. Certainly the free hydroxyl group will oppose the extensive intrusion of the PEG head groups into the hydrophobic chain region. As a consequence therefore in order for the short pegylated chain lipids to reduce the lateral pressure experienced upon compression, the molecules undergo a conformational change, which involves a conformational change of the polyoxyethylene head group to the brush conformation. Interestingly Tsukanova and Salesse [17] report that monolayers formed from the longer pegylated chain lipid DSPE-PEG5000 collapsed upon compression to form structures the thickness of bilayers. When visualising the 2C18 E12 film using the BAM it was clear that there were obviously time and pressure dependent changes in the morphology of the film as evidenced by the formation of necklaces, when the film was held at surface pressures both below and above the plateau region. Interestingly however the formation of the necklaces below πc had no apparent effect on the π–A isotherm, only when compressed to pressures greater than 42 mN m−1 was any hysteresis observable in the π–A isotherm which suggests that although the individual surfactant domains did over time become associated to form the necklaces, possibly due to dehydration and intermixing of the polyoxyethylene chains in the overlap region between domains, the conformation of the 2C18 E12 molecules did not significantly change. History dependent effects were also observed for the vesicles formed by 2C18 E12 upon temperature cycling the vesicles. Upon heating 2C18 E12 , the polymer becomes increasingly dehydrated and as a consequence some of the polyoxyethylene chains intrude into the hydrocarbon region of the bilayer (in much the same way as the PEG did upon compression of the 2C18 E12 film). The result of this intrusion is that upon cooling the surfactant behaves as if it possesses shorter hydrocarbon chains and therefore the hydrophobic chain region ‘solidifies’ at a lower temperature than anticipated from the surfactant’s molecular formula. In the present study the chains solidified at 315.5 K as opposed to the 322 K at which they melted. 5. Conclusion Although temperature and pressure dependent hysteresis effects have been previously noted in monolayers and vesicles prepared from water-insoluble surfactants such as phospholipids, what is unusual in the present study is the extent of the hysteresis exhibited by 2C18 E12 . The reason for this hysteresis is most likely the intermixing of the hydrophobic tails and hydrophilic polyethylene glycol chains of the surfactant; methoxy-capped polyethylene glycol is totally miscible with hydrocarbons. Such extensive mixing of the hydrophobic tails and hydrophilic head group of a surfactant is unusual and may have important consequences for the molecular architecture and therefore properties of the aggregates formed by 2C18 E12 and related polyethylene glycol surfactants. For example upon increasing temperature, the increased intermixing of the hydrophobic and hydrophilic head group regions of the vesicle-
750
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
forming surfactant 2C18 E12 would be expected greatly to reduce the barrier function of the vesicular bilayer, resulting in the rapid release of any entrapped solute. It is proposed that, by replacing the polyethylene glycol head group of a vesicle forming non-ionic surfactant with a polymer with little or no miscibility with hydrocarbon, virtually no intermixing of the hydrophobic and hydrophilic regions will occur and therefore the barrier function will not be as seriously compromised upon increasing temperature, and as a consequence entrapped solute will be released in a more controlled manner. By careful selection of the nature of the head group used to prepare the surfactants it should therefore be possible to tailor the properties of the aggregates.
[7] [8] [9] [10] [11]
[12] [13] [14]
References [15] [1] D.J. Barlow, G. Ma, M.J. Lawrence, J. Webster, G. Penfold, Langmuir 11 (1995) 3737–3741. [2] D.J. Barlow, G. Ma, J. Penfold, J.R.P. Webster, M.J. Lawrence, Langmuir 13 (1997) 3800–3806. [3] R.D. Harvey, R.K. Heenan, D.J. Barlow, M.J. Lawrence, Langmuir 20 (2004) 9282–9290. [4] G. Ma, D.J. Barlow, M.J. Lawrence, R.K. Heenan, P. Timmins, J. Phys. Chem. B 104 (2000) 9081–9085. [5] R.D. Harvey, R.K. Heenan, D.J. Barlow, M.J. Lawrence, Chem. Phys. Lipids 133 (2005) 27–36. [6] R.D. Harvey, D.J. Barlow, A.P.R. Brain, A.F. Drake, L. Kudsiova, M.J. Lawrence, R.K. Heenan, The effect of electrolyte on the encapsulation
[16] [17] [18] [19] [20] [21] [22] [23] [24] [25]
efficiency of vesicles formed by the non-ionic surfactant, 2C18 E12 (2007), submitted for publication. M.J. Lawrence, S.M. Lawrence, S. Chauhan, D.J. Barlow, Chem. Phys. Lipids 82 (1996) 89–100. M.J. Lawrence, S. Chauhan, S.M. Lawrence, D.J. Barlow, STP Pharma Sci. 6 (1996) 49–60. D. Marsh, Handbook of Lipid Bilayers, CRC Press, Boca Raton, FL, USA, 1990. T. Baekmark, G. Elender, D.D. Lasic, E. Sackmann, Langmuir 11 (1995) 3975–3987. T.L. Kuhl, J. Majewski, P.B. Howes, K. Kjaer, A. von Nahmen, K.Y.C. Lee, B. Ocko, J.N. Israelachvili, G.S. Smith, J. Am. Chem. Soc. 121 (1999) 7682–7688. A.K. Kenworthy, S.A. Simon, T.J. McIntosh, Biophys. J. 68 (1995) 1903– 1920. J.P. Coffman, C.A. Naumann, Macromolecules 35 (2002) 1835–1839. J.B. Li, J. Krägel, A.V. Makievski, V.B. Fainermann, R. Miller, H. Möhwald, Colloids Surf. A 142 (1998) 355–360. A.B. Serfis, R. Katzenberger, K. Williams, N. Tran, J. Colloid Interface Sci. 215 (1999) 356–363. T.H. Chou, I.-M. Chu, Colloids Surf. A 211 (2002) 267–274. V. Tsukanova, C. Salesse, Macromolecules 36 (2003) 7227–7235. I. Kubo, S. Adachi, H. Maeda, A. Seki, Thin Films 393 (2001) 80–85. M. Cárdenas, T. Nylander, B. Jönsson, B. Lindman, J. Colloid Interface Sci. 286 (2005) 166–175. P.N. Yi, R.C. MacDonald, Chem. Phys. Lipids 11 (1973) 114–134. R. Koynova, M. Caffrey, Biochim. Biophys. Acta 1376 (1998) 91–145. M.B. Abramson, Biochim. Biophys. Acta 225 (1971) 167–170. N.O. Petersen, S.I. Chan, Biochim. Biophys. Acta 509 (1978) 111–128. T.Y. Tsong, M.I. Kanehisa, Biochemistry 16 (1977) 2674–2680. G. Ma, Structural studies of non-ionic surfactant monolayers and vesicles, Ph.D. thesis, University of London, 1997.
Memory effects of monolayers and vesicles formed by the non-ionic surfactant, 2C18E12 D.J. Barlow, C.M. Hollinshead, R.D. Harvey, L. Kudsiova, M.J. Lawrence ∗ Department of Pharmacy, Franklin Wilkins Building, King’s College London, 150 Stamford Street, London SE1 9NH, UK Received 20 April 2007; accepted 6 August 2007 Available online 11 August 2007
Abstract The behaviour of monolayers and bilayers formed by the dialkyl chain non-ionic surfactant, 1,2-di-O-octadecyl-rac-glycerol-3-ω-methoxydodecaethylene glycol (2C18 E12 ) in water at 297 K has been investigated. Using a surface film balance (or Langmuir trough) the compression– expansion cycle of the 2C18 E12 monolayer was found to be reversible when compressed to surface pressures (π ) less than 42 mN m−1 . Compression of 2C18 E12 monolayer to π greater than 42 mN m−1 above this resulted in a considerable hysteresis upon expansion with the π remaining high relative to that obtained upon compression, suggesting a time/pressure dependent re-arrangement of 2C18 E12 molecules in the film. Morphology of the 2C18 E12 monolayer, investigated using Brewster angle microscopy, was also found to depend upon monolayer history. Bright, randomly dispersed domains of 2C18 E12 of approximately 5 µm in size were observed during compression of the monolayer to π less than 42 mN m−1 . At π of 42 mN m−1 and above, the surfactant film appeared to be almost completely ‘solid-like.’ Regardless of the extent of compression of the monolayer film, expansion of the film caused formation of chains or ‘necklaces’ of individual surfactant domains, with the extent of chain formation dependent upon pressure of compression of the monolayer and the length of time held at that pressure. Irreversible effects on 2C18 E12 vesicle size were also seen upon temperature cycling the vesicles through their liquid–crystalline phase transition temperature with vesicles shrinking in size and not returning to their original size upon standing at 298 K for periods of more than 24 h. No comparable hysteresis, time, pressure or temperature effects were observed with the monolayer or vesicles formed by the corresponding phospholipid, disteaorylphosphatidylcholine, under identical conditions. The effects observed with 2C18 E12 are attributed to the ability of the polyoxyethylene head group to dehydrate and intrude into the hydrophobic chain region of the mono- and bilayers. These studies have important implications for the use of the vesicles formed by 2C18 E12 as drug delivery vehicles. © 2007 Elsevier Inc. All rights reserved. Keywords: Non-ionic surfactant; Monolayers; Vesicles; Langmuir isotherm; Brewster angle microscopy; Variable temperature turbidity measurements
1. Introduction In order rationally to design new surfactants to produce vesicles with suitable properties for drug delivery it is necessary to develop an understanding of the relationship between surfactant molecular structure and aggregate architecture and behaviour. The most informative structural studies performed to date have involved the use of neutron scattering experiments and in particular neutron specular reflection and small angle neutron scattering measurements, both used in combination with * Corresponding author.
E-mail addresses: [email protected], [email protected] (M.J. Lawrence). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.08.014
isotopic substitution. By these means it has been possible to determine the detailed structure of the monolayer of the nonionic polyoxyethylene surfactant, 2C18 E12 at the air/water interface both in the presence and absence of cholesterol and the corresponding phospholipid, disteaorylphosphatidylcholine (DSPC) [1–3] as well as the molecular architecture of the vesicles formed from this surfactant [4,5]. The data obtained from these studies were used to rationalise the disappointingly low encapsulation efficiency of the vesicles formed by 2C18 E12 and guided the re-formulation of the vesicles using added cholesterol and DSPC to give considerably increased encapsulation efficiency [6]. During these studies, however, it was clear that the monolayers and bilayers (vesicles) formed by 2C18 E12 did not behave in a reversible manner demonstrating considerable
742
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
time/pressure dependent properties. In the studies reported here we use a combination of techniques to explore this behaviour with a view to understanding the potential of the vesicles formed by 2C18 E12 as drug delivery vesicles.
The non-ionic surfactant, 1,2-di-O-octadecyl-rac-glycerol3-ω-methoxydodecaethylene glycol (2C18 E12 ) was synthesised and characterised according to Lawrence et al. [7]. Disteaorylphosphatidylcholine (DSPC) was purchased from Avanti Polar Lipids (USA). Chloroform and ethanol of HPLC-grade quality were supplied by Rathburn Chemicals Ltd. (UK). Water was firstly doubly distilled and purified using an Elgastat Maxima purifier (Elga, UK) to a resistance of at least 18 M. All chemicals were of the highest grade available and used as supplied.
and to reduce airborne contamination. P-polarised light (supplied by an NDYAG laser operating at a wavelength of 532 nm) was emitted by the BAM onto the water surface at an angle of incidence of the Brewster angle for water and the light reflected from the surface collected by two achromatic lenses and detected with a CCD camera. The CCD camera converted the reflectivity signal from the sample into a video image. Calibrations were performed by obtaining images of the bare water surface prior to spreading the monolayer. These ‘background images’ were subtracted from the sample images taken subsequently in order to obtain sharper pictures of the monolayer films. The spatial resolution of the BAM is approximately 2 µm. Two different magnifications were employed during the study, namely a low magnification using a ×10 objective where the images obtained represented a sample size of 430 µm (horizontal field of view) × 530 µm (vertical field of view) and a higher magnification (×20 objective), where the images represented a sample of size 215 µm × 265 µm.
2.2. Langmuir film
2.4. Vesicle preparation
The Langmuir trough (Nima Technology Ltd., Coventry, UK) was scrupulously cleaned by first soaking in neutral Decon, thereafter repeatedly rinsing with ultra-pure water and finally wiping the trough’s inner surfaces with first chloroform, then ethanol and then chloroform again. The Wilhelmy plate was a piece of filter paper which, prior to attachment to the Langmuir trough’s microbalance, was degreased by soaking in chloroform, and dried by allowing the solvent to evaporate. The trough was filled with water and the surface of the water cleaned of any surface active impurities by sweeping the Langmuir trough’s surface with its moveable Teflon barrier followed by suction using a Pasteur pipette attached to a vacuum pump. The temperature of the aqueous sub-phase was controlled at 297 ± 0.1 K by water circulation from a thermostat. Once the surface of the water was clean, a small aliquot (typically between 10 to 100 µL) of a 2 mg/mL chloroform solution of 2C18 E12 or DSPC was carefully added dropwise to the surface using a Hamilton microsyringe. 20 min was allowed for the chloroform to evaporate before the film was compressed. Isotherms were obtained by compressing and expanding the film at a rate of 10 mm min−1 . For the Brewster angle microscope studies, the barrier was moved to the desired surface pressure and held at that surface pressure while the monolayer was visualised. This process was repeated a number of times using the same monolayer and with freshly prepared monolayer films.
Vesicles were prepared by hydration of a thin film of 2C18 E12 or DSPC (25 mg deposited from evaporation of a chloroform solution). Hydration of the surfactant film was achieved by continuous agitation with 5 mL of solvent at 338 K for 30 min (>10 K above the Tc of the 2C18 E12 or DSPC) [8,9]. The resultant suspension was sonicated in a Decon bath sonicator for 60 min and then ultrasonicated for a further 15 min at 328 K using a Lucas Dawes probe sonicator fitted with a tapered microtip (operating at 15% of its maximum output). Any titanium particles shed from the probe sonicator were removed by centrifugation. The sizes of the vesicles thus prepared were assessed (after appropriate dilution) by laser light scattering using a Malvern Autosizer fitted with a 10 mW helium–neon laser. Dilution of the vesicles for size determination by light scattering was necessary to avoid inter-vesicular interactions giving a false estimate of vesicle size. Dilution of the vesicles was not considered to alter the level of surfactant hydration [2,4].
2. Experimental 2.1. Materials
2.3. Brewster angle microscopy (BAM) The BAM (BAM2plus, Nanofilm Technologie, GmbH, Goettingen, Germany) was mounted above the Langmuir trough, to allow for collection of images of the morphology of the 2C18 E12 monolayer. The Langmuir trough (monolayer prepared as described above) and the BAM were both housed in a cabinet to minimise disruption to the monolayer by air currents
2.5. Turbidity studies Vesicle suspensions of 2C18 E12 or DSPC for spectrophotometric measurement were prepared at a concentration of 5 mg mL−1 . The vesicle suspensions were transferred to a standard 1-cm fused silica cuvette and were incubated at their respective starting temperatures for 30 min prior to measurement. All turbidity measurements were recorded as absorbance at 500 nm (Aapp (500)) using a Perkin Elmer Lambda 5 spectrophotometer, with slit widths of 0.2 nm. The spectrophotometer was equipped with a Peltier thermostatic cell holder, controlled by a PTP-1 Peltier temperature programmer (accuracy ±0.1 K). The temperature within the cell was monitored using a Comark thermocouple positioned at approximately half the depth of the vesicle suspension. Absorbance (Aapp (500)) was measured, during repeated heating and cooling cycles carried
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
out at a rate of 0.6 K min−1 (every 1 K). During each experiment, three heating and cooling cycles were carried out, with 30 min temperature equilibration at both the upper and lower temperatures prior to the beginning of a heating or cooling run. The non-ionic 2C18 E12 vesicles were heated and cooled within the range of 303–328 K while DSPC vesicles were examined over the range 313–333 K. The optical density versus heating curve thus obtained was converted to a simple first derivative by taking the difference in the absorbance (Aapp (500)) between adjacent points using the Galactic Industries Grams/32 AI (version 6.0) computer program.
743
3.1.1. π–A isotherms 2C18 E12 is a double-chained surfactant that forms an insoluble film at the air–water interface at 297 K. Fig. 1 shows the π–A isotherm obtained for 2C18 E12 after a single compression cycle but derived from the mean of individual measurements on three films. The π–A isotherm exhibited by 2C18 E12 was extremely reproducible (with a standard deviation in surface pressure of less than 0.5 mN m−1 ) and was found to be independent of spreading volume (i.e. number of molecules deposited on the surface) when added to surface pressures of less than ∼40 mN m−1 . As can be seen from Fig. 1, the π–A isotherm obtained for 2C18 E12 exhibits three distinct regions suggesting a likely conformational change/re-arrangement of the 2C18 E12 molecules within the film upon compression. Up to surface pressures of 42 mN m−1 , the 2C18 E12 film is expanded with a non-zero surface pressure being detected even at very large areas per molecule. Furthermore the film is well behaved in that it was possible to maintain surface pressure and area steady for periods of at least several hours with no hysteresis of the π–A isotherm being observed upon subsequent expansion of the film. Both of these observations suggest no loss of 2C18 E12 to the aqueous sub-phase as has been observed for films prepared from much longer chain pegylated distearoylphosphatidylethanolamines (DSPE), such as DSPE-PEG45 and DSPE-PEG110 with DMPE [10]. In this context it should be noted that Kuhl et al. [11] have reported that DSPE-PEG90 , DSPE-PEG350 and DSPC-PEG750 all form stable monolayers
when present as sole component, although it has been reported that DSPE-PEG750 forms micelles when present at greater than 60 mol% [12] and therefore some hysteresis of the film might be anticipated. In contrast DSPE-PEG350 forms stable bilayers at 100 mol% [12]. A surface pressure of 42 mN m−1 (πc ) marks the start of a large plateau region similar in appearance to the expanded/condensed co-existence region reported for monolayers of pegylated distearoylphosphatidylethanolamines [11] albeit at far higher surface pressures than those recorded DSPE-PEG350 and DSPC-PEG750 (both of which show plateau regions commencing at surface pressures of ∼20 mN m−1 [11,13]. Upon further compression of the 2C18 E12 film to an area per 2C18 E12 molecule of about ∼40 Å2 , the surface pressure exerted by the film suddenly rises and the film collapses at a surface pressure of around 65 mN m−1 and a limiting area per molecule (determined from the intercept of the extrapolated steep rising part 2 of the isotherm with the abscissa) of 37.4 ± 1.3 Å , which is anomalous as it is less than the cross sectional area of two hydrocarbon chains. Upon expansion of a 2C18 E12 film previously compressed to a surface pressure of greater than πc but less than the collapse pressure of ∼65 mN m−1 , a large hysteresis was observed in the π–A isotherm with the surface pressure of the film remaining high at large areas per 2C18 E12 molecule. This hysteresis persisted for at least several hours and was variable in its extent, being dependent upon the compression pressure above πc . Again there was no evidence of loss of 2C18 E12 to the aqueous sub-phase. Fig. 2 shows the isotherm obtained for DSPC under exactly the same conditions as those used for 2C18 E12 . The isotherm obtained here for DSPC is in agreement with those reported for this lipid in the literature [14–16]. There are clear differences in the isotherms obtained for DSPC and 2C18 E12 , the most noticeable difference being that the DSPC isotherm does not show any distinct phase transition, with the exception of the transition from a gaseous phase, which was present at pressures below 1 mN m−1 , to the liquid expanded phase. The film was seen to collapse at a surface pressure of greater than 60 mN m−1 . At pressures below the collapse pressure no hysteresis was observable. Again the film was well behaved in that it was possible to hold the film without significant change in its properties for at least several hours.
Fig. 1. Surface pressure—isotherm of 2C18 E12 on a sub-phase of water at 298 K.
Fig. 2. Surface pressure—isotherm of DSPC on a sub-phase of water at 298 K.
3. Results 3.1. Monolayer studies
744
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
(a)
(b)
(c)
(d)
(e) Fig. 3. BAM images of monolayers of 2C18 E12 on a sub-phase of water at 298 K and held at a surface pressure of (a) 15 mN m−1 , (b) 20 mN m−1 , (c) 25 mN m−1 , (d) 30 mN m−1 and (e) 40 mN m−1 (×10 objective).
3.1.2. BAM images BAM studies were performed to visualise the film formed by 2C18 E12 at the air/water interface at a range of surface pressures. Fig. 3 shows a series of representative pictures of the 2C18 E12 film upon initial compression up to and including a π of 40 mN m−1 , using a ×10 objective. As can be seen the 2C18 E12 film is clearly visualized as bright, welldefined domains of approximately circular shape (even when examined under a ×20 objective, data not shown) and of similar size (∼5 µm). These circular domains were fairly evenly spaced over the surface of the water. Although Tsukanova and Salesse [17] have described the formation of similar looking domains of DSPE-PEG5000 molecules spread on a water subphase at a relatively low surface pressure of 13.6 mN m−1
using BAM Kuhl et al. [11] reported the absence of any domain formation or structuring in films formed by DSPE-PEG90 , DSPE-PEG350 and DSPE-PEG750 . (Note here that the polyoxyethylene head groups in the surfactant used in the present study approximately correspond to PEG500 .) In the present study, compression of the 2C18 E12 film only served to pack the circular domains closer together. At a surface pressure of 40 mN m−1 an almost complete, ‘solid-like’ surfactant film was observed. At surface pressures below πc , there were small areas of the film in which the domains seemed to have annealed into chains (or ‘necklaces’) rather than being regularly spaced; such organisation is seen in Fig. 4a, in a film compressed to 15 mN m−1 . This annealing had the result that some parts of the water sur-
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
(a)
(a)
(b)
(b)
745
Fig. 4. BAM images of monolayers of 2C18 E12 on a sub-phase of water at 298 K and held at a surface pressure of 15 mN m−1 (a) ×10 objective and (b) ×20 objective.
(c)
Fig. 5. BAM image of a monolayer of 2C18 E12 on a sub-phase of water at 298 K and held at a surface pressure of 45 mN m−1 (×20 objective).
face were quite bare of surfactant whilst other areas were more densely packed with ‘necklaces’ of surfactant domains. Upon careful examination of the domains at higher magnification (Fig. 4b) it was clear that the individual domains comprising these necklaces were still approximately circular and of about 5 µm diameter. It must be stressed, however, that the regions of the monolayer in which these chains were observed represented only a very small proportion (<5%) of the water surface. Fig. 5 shows a 2C18 E12 film compressed to beyond πc (to π 45 mN m−1 ). Here again the surfactant film appears quite close-packed, although it is difficult to conclude anything about the ‘state’ of the 2C18 E12 molecules comprising the film in this
Fig. 6. BAM image of monolayers of 2C18 E12 on a sub-phase of water at 298 K at a surface pressure of (a) 36 mN m−1 , (b) held at a surface pressure of 36 mN m−1 for 45 min and (c) compressed to 40 mN m−1 after being held at a surface pressure of 36 mN m−1 for 45 min ×10 objective).
region since the images look no different from those obtained for films compressed to a surface pressure less than 42 mN m−1 except that the surface of the film does not seem as uniformly smooth as those seen in Fig. 3, suggesting perhaps that there are inhomogeneities in the thickness of the film compressed to 45 mN m−1 . During the BAM studies it became clear that there were both time and pressure effects on the appearance of the film, even at surface pressures below πc . Fig. 6 demonstrates the effects of time on the 2C18 E12 film when compressed to π less than πc . In Fig. 6a, the monolayer is seen upon initial compression to a surface pressure of 36 mN m−1 while Fig. 6b shows the same film after being held at that surface pressure for 45 min.
746
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
(a)
(a)
(b)
(b)
Fig. 7. BAM image of monolayers of 2C18 E12 on a sub-phase of water at 298 K (a) compressed to 34 mN m−1 and re-opened to 30 mN m−1 and (b) compressed to 34 mN m−1 then re-opened to 19 mN m−1 (×20 objective).
(It should be noted that these images were chosen as they appeared to be representative of the entire surface.) It can be seen that over time, the film of evenly spaced domains does change in appearance with the domains merging to form ‘necklaces,’ and leading to some patches of water sub-phase appearing to be entirely devoid of surfactant molecules. Fig. 6c shows the same monolayer as in Fig. 6b after being compressed to 40 mN m−1 where the film appears to be in a similar solid state to that obtained upon direct compression of the film to 40 mN m−1 (Fig. 3). Fig. 7 shows images of a 2C18 E12 film, which was first compressed to 40 mN m−1 (less than πc ) and then a ‘reverse isotherm’ obtained (i.e. a gradual expansion of the barriers) to 30 mN m−1 (Fig. 7a) or to 19 mN m−1 (Fig. 7b). In these images, the domains are annealed as shown in Fig. 6. Again these images were selected as they were representative of the entire surface and there were no homogeneous regions of isolated domains similar to those present at similar pressures as shown in Fig. 3. These figures provide visual evidence of a hysteresis in the morphology of the film, in that upon compression, the domains become strongly annealed and do not separate again upon decompression, although it should be noted that under these conditions no hysteresis of the monolayer was observed via the film balance experiments. Images of the DSPC monolayer at the air/water interface, obtained using Brewster angle microscopy at a range of sur-
(c) Fig. 8. BAM image of monolayers of DSPC on a sub-phase of water at 298 K and held at a surface pressure of (a) 0.5 mN m−1 , (b) 17 mN m−1 and (c) 50 mN m−1 (×10 objective).
face pressures are shown in Fig. 8. As can be seen from Fig. 8 the images obtained for DSPC were quite different from those observed for 2C18 E12 . Fig. 8a shows the image of the DSPC monolayer at a surface pressures 0.5 mN m−1 when the film was in its gaseous state according to its surface pressure. Already, however, domains of condensed phase are visible on the sub-phase, although the film was very fluid and the domains (which are smaller than those observed with 2C18 E12 ) therefore hard to image. Upon compression, more domains were clear and the mobility of the film decreased. Similar images have been obtained for DSPC when spread on low concentrations of electrolyte [18,19]. No hysteresis of the DSPC film was observed upon expansion of the film.
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
747
3.2. Vesicles 3.2.1. Turbidity studies In common with phospholipid vesicles, upon first heating of the 2C18 E12 vesicles through their phase transition (i.e. from 303 K up to a temperature of 328 K), two large decreases in absorbance measured at 500 nm (Aapp (500)) were seen to occur at ∼316 and ∼320 K (Fig. 9a). By analogy with phospholipids, it was assumed that the first decrease was attributable to a pre-transition or gel to ripple phase (Tpre ), while the second was due to the liquid–crystalline phase transition (Tc ) [20]. The changes observed upon the first heating cycle were not reversible upon cooling as there was no increase in Aapp (500) until ∼316 K, whence there was a relatively rapid increase in Aapp (500) followed by a much slower rise in Aapp (500). However, even at 303 K, the Aapp (500) of the vesicular suspension had not recovered its original value prior to heating. Furthermore although the curves obtained in the second and third heating cycles were superimposable, they were not identical to the first heating curve as the extent of the first decrease in Aapp (500) was very much reduced. The extent of the second Aapp (500) decrease initiated at ∼320 K, was however unaffected by the change in initial turbidity at the start of the heating run. The second and third cooling curves were superimposable on the first. Fig. 9b shows the mean of the three heating and the three cooling curves. Fig. 9c gives the simple derivative of the mean heating and cooling curves shown in Fig. 9b and were used to determine the temperatures of the Tpre and Tc which were found to be 317 ± 0.5 and 322 ± 0.5 K. Upon cooling, the only peak seen was at 315.5 ± 0.5 K. For a population of 2C18 E12 vesicles with mean diameters for which the standard deviation is low, as is the case here, the decrease observed in turbidity upon heating can be reasonably assumed to coincide with a decrease in diameter of the vesicles. Similarly, because upon cooling the vesicles back to the starting temperature, Aapp (500) did not restore the Aapp (500) to its original value, it would seem that the vesicles had not returned to their original size. In order to confirm this apparent reduction in vesicle size after the first heating/cooling cycle, a sizing study was performed before and after the vesicles had been incubated at 333 K for 30 min. As can be seen from Table 1, even after 24 h incubation at 298 K, the vesicles had not returned to their original size. (Vesicle size changes over longer recovery periods were not examined.) This phenomenon may be influenced by both the original size (and lamellarity) of the vesicles. The reduction in size of 2C18 E12 vesicles observed may indicate that the vesicles require annealing for long periods below 303 K prior to use. Fig. 10a shows that for a suspension of DSPC liposomes, there were reproducible decreases in Aapp (500) in response to increasing temperature. There was a small decrease after 323 K and a larger decrease after 326 K. The second decrease in Aapp (500) was reversed upon cooling of the vesicle suspension (Fig. 10b), however, the first drop in Aapp (500) was missing from the cooling cycle curve. The data used in Fig. 10b were converted by calculating a simple difference derivative, into the curves shown in Fig. 10c. The changes in Aapp (500) observed
(a)
(b)
(c) Fig. 9. Temperature dependence of the optical density at 500 nm of 2C18 E12 vesicles in water (a) superimposed results of three heating cycles on the same sample, (b) mean results of 4 heating and cooling cycles and (c) simple derivative curve calculated from the mean heating and cooling curves in (b).
748
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
Table 1 Diameters of 2C18 E12 vesicles, both before and at various times after incubation at 333 K Experimental treatment
Vesicle diameter (nm)
1 h post sonication at 298 K 10 min post 30 min incubation at 333 K 1 h at 298 K 2 h at 298 K 3 h at 298 K 16 h at 298 K 24 h at 298 K
529 ± 14.1 390 ± 9.1 392 ± 9.7 428 ± 8.9 440 ± 17.4 442 ± 15.1 441 ± 9.9
Note. Light scattering measurements were performed at 298 K after appropriate temperature equilibration (n− = 3 ± s.d.).
in DSPC liposome suspensions are thus seen in Fig. 10c as two peaks (at 324 ± 0.5 K and 327 ± 0.5 K, respectively). These values accord well with the values obtained from DSC studies for the Tpre and Tc , these being 322.1 ± 2.9 K and 327.5 ± 1.5 K, respectively [21]. Fig. 10a clearly shows that although the temperatureinduced turbidity changes at the main phase transition of DSPC were reversible, a hysteresis was observed for the pre-transition, during the cooling cycle. This phenomenon has previously been observed for suspensions of DPPC vesicles [20,22–24]. Equilibration of the vesicle suspension 10 K below Tpre was sufficient to ensure recovery from the hysteresis in the pre-transition (Fig. 10a), which is thought to be maintained at elevated equilibration temperatures [23,24]. In contrast to the vesicles formed from 2C18 E12 there are no significant change in the size of the vesicles prepared from DSPC prior to and after three heating/cooling cycles (data not shown).
(a)
4. Discussion Lipid molecules containing relatively short chain hydrophilic polymers such as polyethylene glycol have attracted much interest because of their ability to produce sterically stabilised vesicles suitable for drug delivery and drug targeting applications. Before discussing the results of the present study, it is worth considering earlier studies on related pegylated molecules. Most of the related studies have examined the behavior of poly(ethylene) glycols (PEG) of varying molecular weights (typically 350–5000 D) when conjugated to the terminal amine group of distearoylphosphatidylethanolamine (DSPE). Significantly, these molecules differ from the type investigated in the present study in that the PEG head group is attached to the glycerol backbone via a negatively charged phosphidylethanolamine group, while in 2C18 E12 the PEG chain is directly linked to the glycerol moiety. Of the pegylated molecules studied to date only those containing short PEG chains, namely DSPE-PEG90 , DSPEPEG350 , DSPE-PEG750 have been reported to form ‘stable’ monolayers when used as pure component, whereas lipids with longer pegylated chains such as DSPE-PEG2000 have only been reported to form relatively stable, well-behaved monolayers when combined with the phospholipid, dimyristoylphosphatidylethanolamine (DMPE) [10]. However, unlike the films formed by 2C18 E12 no report exists of the film formed by these
(b)
(c) Fig. 10. Temperature dependence of the optical density at 500 nm of DSPC vesicles in water (a) superimposed results of three heating cycles on the same sample, (b) mean results of 6 heating and cooling cycles and (c) simple derivative curve calculated from the mean heating and cooling curves in (b).
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
short chain pegylated lipid molecules exhibiting any hysteresis. While the π–A isotherm obtained for 2C18 E12 films reported in the present study exhibited some similarity to those recorded for DSPE-PEG350 and DSPE-PEG750 , in that at low surface pressures the film was expanded and underwent a conformational change/re-arrangement at intermediate surface pressures, a number of differences were obvious. In particular the plateau region present in the 2C18 E12 isotherm commenced at a much higher surface pressure of 42 mN m−1 compared to the plateau reported for DSPE-PEG350 and DSPE-PEG750 which began at a surface pressure of approximately 20 mN m−1 . In addition, the limiting area per molecule of DSPE-PEG350 , DSPE-PEG750 , estimated from the π–A isotherms presented by Kuhl et al. [11], 2 was greater than 50 Å2 , much larger than the 37.4 ± 1.3 Å determined here for 2C18 E12 . Interestingly, at a surface pressure of 42 mN m−1 , the films formed by DSPE-PEG350 and DSPE-PEG750 were reported to be in the solid phase [11]. It is clear from the BAM studies presented here that 2C18 E12 is also present as a close packed film at a surface pressure of 42 mN m−1 (i.e. at the start of the plateau region), although at this surface pressure 2C18 E12 exhibits a far greater area per molecule of ∼110 Å2 (determined from neutron reflectivity studies [25]) than DSPE-PEG350 and DSPE-PEG750 . Although this value appears to be very large, and much greater than the values obtained for pegylated lipids, experimental data has shown that 2C18 E12 does indeed form vesicles. Once a complete monolayer is formed, any further compression of the film normally leads to a rapid increase in surface pressure of the film and then collapse of the film. Significantly, for 2C18 E12, neutron reflectivity studies suggest that when compressed to areas per molecule of less than ∼110 Å2 , the 2C18 E12 film begins to form islands of bi- or multilayers (as evidenced by the appearance of off-specular neutron reflection) and that furthermore these islands of bi- or multilayers do not disappear upon expansion of the film [25]. The formation of bior multilayers upon compression is unusual and is undoubtedly a consequence of dehydration upon compression of at least, some of the PEG chains of 2C18 E12 coupled with the ability of the dehydrated, methoxy capped polyethylene glycol chains to intrude into and mix with the hydrophobic region of the film; methoxy capped polyoxyethylene glycol is well known to be totally miscible with hydrocarbon solvents. This mixing of the hydrophilic polyoxyethylene head group and the stearoyl chains is most likely the explanation for the very small area per molecule recorded for 2C18 E12 at its collapse. Furthermore the dehydration of the PEG chains upon compression and their subsequent mixing with the stearoyl chains is the likely cause of the hysteresis (or memory effects) seen in the π–A isotherm because in order for the polyoxyethylene chains to re-enter the aqueous phase they must become re-hydrated, a process that may take some time due to the energetically favourable mixing of the hydrophilic methoxy capped polyoxyethylene head group and the stearoyl chains. While some overlap of the hydrophobic and hydrophilic PEG head groups has been previously reported for short pegylated chain lipids when in their solid-like state [11], the extent
749
of overlap might be anticipated to be far less when a charged phosphate group is present coupled with the presence of a free hydroxyl group. Certainly the free hydroxyl group will oppose the extensive intrusion of the PEG head groups into the hydrophobic chain region. As a consequence therefore in order for the short pegylated chain lipids to reduce the lateral pressure experienced upon compression, the molecules undergo a conformational change, which involves a conformational change of the polyoxyethylene head group to the brush conformation. Interestingly Tsukanova and Salesse [17] report that monolayers formed from the longer pegylated chain lipid DSPE-PEG5000 collapsed upon compression to form structures the thickness of bilayers. When visualising the 2C18 E12 film using the BAM it was clear that there were obviously time and pressure dependent changes in the morphology of the film as evidenced by the formation of necklaces, when the film was held at surface pressures both below and above the plateau region. Interestingly however the formation of the necklaces below πc had no apparent effect on the π–A isotherm, only when compressed to pressures greater than 42 mN m−1 was any hysteresis observable in the π–A isotherm which suggests that although the individual surfactant domains did over time become associated to form the necklaces, possibly due to dehydration and intermixing of the polyoxyethylene chains in the overlap region between domains, the conformation of the 2C18 E12 molecules did not significantly change. History dependent effects were also observed for the vesicles formed by 2C18 E12 upon temperature cycling the vesicles. Upon heating 2C18 E12 , the polymer becomes increasingly dehydrated and as a consequence some of the polyoxyethylene chains intrude into the hydrocarbon region of the bilayer (in much the same way as the PEG did upon compression of the 2C18 E12 film). The result of this intrusion is that upon cooling the surfactant behaves as if it possesses shorter hydrocarbon chains and therefore the hydrophobic chain region ‘solidifies’ at a lower temperature than anticipated from the surfactant’s molecular formula. In the present study the chains solidified at 315.5 K as opposed to the 322 K at which they melted. 5. Conclusion Although temperature and pressure dependent hysteresis effects have been previously noted in monolayers and vesicles prepared from water-insoluble surfactants such as phospholipids, what is unusual in the present study is the extent of the hysteresis exhibited by 2C18 E12 . The reason for this hysteresis is most likely the intermixing of the hydrophobic tails and hydrophilic polyethylene glycol chains of the surfactant; methoxy-capped polyethylene glycol is totally miscible with hydrocarbons. Such extensive mixing of the hydrophobic tails and hydrophilic head group of a surfactant is unusual and may have important consequences for the molecular architecture and therefore properties of the aggregates formed by 2C18 E12 and related polyethylene glycol surfactants. For example upon increasing temperature, the increased intermixing of the hydrophobic and hydrophilic head group regions of the vesicle-
750
D.J. Barlow et al. / Journal of Colloid and Interface Science 316 (2007) 741–750
forming surfactant 2C18 E12 would be expected greatly to reduce the barrier function of the vesicular bilayer, resulting in the rapid release of any entrapped solute. It is proposed that, by replacing the polyethylene glycol head group of a vesicle forming non-ionic surfactant with a polymer with little or no miscibility with hydrocarbon, virtually no intermixing of the hydrophobic and hydrophilic regions will occur and therefore the barrier function will not be as seriously compromised upon increasing temperature, and as a consequence entrapped solute will be released in a more controlled manner. By careful selection of the nature of the head group used to prepare the surfactants it should therefore be possible to tailor the properties of the aggregates.
[7] [8] [9] [10] [11]
[12] [13] [14]
References [15] [1] D.J. Barlow, G. Ma, M.J. Lawrence, J. Webster, G. Penfold, Langmuir 11 (1995) 3737–3741. [2] D.J. Barlow, G. Ma, J. Penfold, J.R.P. Webster, M.J. Lawrence, Langmuir 13 (1997) 3800–3806. [3] R.D. Harvey, R.K. Heenan, D.J. Barlow, M.J. Lawrence, Langmuir 20 (2004) 9282–9290. [4] G. Ma, D.J. Barlow, M.J. Lawrence, R.K. Heenan, P. Timmins, J. Phys. Chem. B 104 (2000) 9081–9085. [5] R.D. Harvey, R.K. Heenan, D.J. Barlow, M.J. Lawrence, Chem. Phys. Lipids 133 (2005) 27–36. [6] R.D. Harvey, D.J. Barlow, A.P.R. Brain, A.F. Drake, L. Kudsiova, M.J. Lawrence, R.K. Heenan, The effect of electrolyte on the encapsulation
[16] [17] [18] [19] [20] [21] [22] [23] [24] [25]
efficiency of vesicles formed by the non-ionic surfactant, 2C18 E12 (2007), submitted for publication. M.J. Lawrence, S.M. Lawrence, S. Chauhan, D.J. Barlow, Chem. Phys. Lipids 82 (1996) 89–100. M.J. Lawrence, S. Chauhan, S.M. Lawrence, D.J. Barlow, STP Pharma Sci. 6 (1996) 49–60. D. Marsh, Handbook of Lipid Bilayers, CRC Press, Boca Raton, FL, USA, 1990. T. Baekmark, G. Elender, D.D. Lasic, E. Sackmann, Langmuir 11 (1995) 3975–3987. T.L. Kuhl, J. Majewski, P.B. Howes, K. Kjaer, A. von Nahmen, K.Y.C. Lee, B. Ocko, J.N. Israelachvili, G.S. Smith, J. Am. Chem. Soc. 121 (1999) 7682–7688. A.K. Kenworthy, S.A. Simon, T.J. McIntosh, Biophys. J. 68 (1995) 1903– 1920. J.P. Coffman, C.A. Naumann, Macromolecules 35 (2002) 1835–1839. J.B. Li, J. Krägel, A.V. Makievski, V.B. Fainermann, R. Miller, H. Möhwald, Colloids Surf. A 142 (1998) 355–360. A.B. Serfis, R. Katzenberger, K. Williams, N. Tran, J. Colloid Interface Sci. 215 (1999) 356–363. T.H. Chou, I.-M. Chu, Colloids Surf. A 211 (2002) 267–274. V. Tsukanova, C. Salesse, Macromolecules 36 (2003) 7227–7235. I. Kubo, S. Adachi, H. Maeda, A. Seki, Thin Films 393 (2001) 80–85. M. Cárdenas, T. Nylander, B. Jönsson, B. Lindman, J. Colloid Interface Sci. 286 (2005) 166–175. P.N. Yi, R.C. MacDonald, Chem. Phys. Lipids 11 (1973) 114–134. R. Koynova, M. Caffrey, Biochim. Biophys. Acta 1376 (1998) 91–145. M.B. Abramson, Biochim. Biophys. Acta 225 (1971) 167–170. N.O. Petersen, S.I. Chan, Biochim. Biophys. Acta 509 (1978) 111–128. T.Y. Tsong, M.I. Kanehisa, Biochemistry 16 (1977) 2674–2680. G. Ma, Structural studies of non-ionic surfactant monolayers and vesicles, Ph.D. thesis, University of London, 1997.