Temperature-controlled content release from liposomes encapsulating Pluronic F127

Journal of Controlled Release 76 (2001) 27–37 www.elsevier.com / locate / jconrel

Temperature-controlled content release from liposomes encapsulating Pluronic F127 Parthapratim Chandaroy, Arindam Sen, Sek Wen Hui* Molecular and Cellular Biophysics Department, Roswell Park Cancer Institute, Buffalo, NY 14263, USA Received 12 March 2001; accepted 24 May 2001

Abstract Temperature-dependent internal content release from liposomes was examined using di-oleoylphosphatidylcholine (DOPC) / cholesterol liposomes with encapsulated Pluronic F127 molecules. The interaction of Pluronic F127 with the lipid bilayer at elevated temperature causes the release of encapsulated contents. Content release was measured using fluorescent markers of two different sizes: small, carboxyfluorescein (CF), and large, bovine serum albumin-conjugated fluorescein iso-thiocyanate (BSA-FITC). Release of CF was studied using fluorescence de-quenching, while that of BSA-FITC was studied using fluorescence emission quenching due to fluorescence resonance energy transfer (FRET). Temperaturecontrolled complete internal content release was achieved at a precise temperature by controlling the concentration of the encapsulated Pluronic. Increasing cholesterol % in the liposome composition resulted in a sharper transition with temperature in content release. The onset temperature of content release increased with decrease in Pluronic concentration. For the same Pluronic concentration, the onset temperature also depended on the size of the encapsulated marker and was higher for larger markers. We have established that onset of content release is determined by the critical micellar temperature (CMT) of the Pluronic. Temperature-sensitive liposomes, made stealth using di-stearoyl(polyethylene glycol 5000) phosphatidylethanolamine (DSPEG5000PE) in conjunction with Pluronic F127, had similar temperature sensitivity and efficiency in content release compared to the non-stealth liposomes.  2001 Published by Elsevier Science B.V. Keywords: Temperature-controlled release; Pluronic F127; Fluorescence de-quenching; Fluorescence energy transfer

1. Introduction Liposomes have been used extensively in the past decades as drug carriers [1]. In order to be used as efficient carriers they have to meet certain criteria. Some of those criteria are — evade the mononuclear phagocyte system (MPS) to prolong the circulation half-life (t 1 / 2 ), and release of the encapsulated drug *Corresponding author. Tel.: 11-716-845-8595; fax: 11-716845-8683. E-mail address: [email protected] (S.W. Hui).

only at the targeted site. The use of sterically stabilized liposomes has increased the liposome circulation time considerably [2,3]. Fusogenic liposomes have been developed [4,5] for cytoplasmic delivery of membrane-impermeable molecules. Controlled release of internal content from a suitably designed stimulus-sensitive liposome can be achieved by using various stimuli, such as temperature [6–9], pH [10–12] and light [13,14]. Different approaches have been used to produce temperature-sensitive liposomes for controlled release, such as use of the phase transition property of

0168-3659 / 01 / $ – see front matter  2001 Published by Elsevier Science B.V. PII: S0168-3659( 01 )00429-1

28

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

the constituent lipids [9]. One of the most effective methods of modification so far has been the use of polymers [7,8,15]. Polymers interact with the liposome membrane lipids and above a certain temperature or concentration can cause release of the internal content. One of the polymers used has been poloxamer, which can cause moderate to severe release around physiological temperature [16,17]. Poloxamers are polyethylene oxide (PEO)–polypropylene oxide (PPO)–polyethylene oxide tri-block co-polymers of different molecular weights. The hydrophobic PPO group in the middle links the two hydrophilic PEO groups. The amphiphilic nature of the poloxamer renders itself extremely useful in various applications as emulsifiers and stabilizers [18]. In an aqueous environment, poloxamers at a given concentration would remain as individual (non-associated) co-polymers, from here on termed as ‘monomers’, at temperatures below their critical micellar temperature (CMT). Above the CMT the molecules become more lipophilic, and form micelles with hydrophobic PPO groups at the core of the micelle. Poloxamers of different molecular weights and with different hydrophilic–lipophilic balance (HLB) have different CMTs [19]. This monomer-to-micellar transition process is extremely temperature-sensitive. With a slight change of temperature, the corresponding critical micellar concentration (CMC) may change by several orders of magnitude [20]. Temperature-controlled content release from phosphatidylcholine (PC) liposomes coated with a copolymer of N-isopropylacrylamide (NIPAM) has been attempted by Kim et al. [7]. Results from that study show that the extent of release is quite low (up to |35%) without the aid of gel-to-liquid crystalline phase transition. We intend to improve on such a system by using a poloxamer instead of the NIPAM copolymer. Another study [8] shows the use of poly(NIPAM) coated di-oleoylphosphatidylethanolamine (DOPE) liposomes to achieve temperaturetriggered content release. This study relies on the stability imparted by the copolymer associated with DOPE at temperatures below the lower critical solution temperature (LCST) of the copolymer. Being a non-bilayer-forming lipid, DOPE does not form stable liposomes without the copolymer at the temperatures studied. The system becomes unstable at temperatures above the LCST due to a reduction

in the stabilizing effect of the copolymer. We take a different approach to design temperature-sensitive liposomes, using a poloxamer. Poloxamers would not associate with the liposome bilayer at temperatures below the CMT. Above CMT, they would partition into the bilayer, causing defects in the bilayer, leading to eventual disruption of the bilayer. This method can be applied to most bilayer-forming lipids. Moreover, we tested controlled release also in stealth liposomes to lay the foundation for its application in vivo. We used Pluronic F127 (M.W. |12,600, PEO 98 – PPO 67 –PEO 98 ), a poloxamer, in this study for its high molecular weight, desired HLB and a low CMT around the physiological temperature. Here we report the temperature-dependent release of encapsulated tracer molecules of different molecular weights from liposomes of different lipid compositions. We used different mol% of cholesterol as well as different concentration of the poloxamer molecules to study the effect on such release. Both small (CF, M.W. |376) and large (BSA-FITC, M.W. |66,000) markers were used to study the effect of molecular size on their release. We also tested the release of the same markers from stealth liposomes containing distearoyl (polyethylene glycol 5000) phosphatidylethanolamine (DS(PEG5000)PE).

2. Materials and methods

2.1. Materials Pluronic F127 was a gift from BASF (Mount Olive, NJ, USA) as free samples. All lipids, di-oleoyl phosphatidylcholine (DOPC), cholesterol and DS(PEG5000)PE, were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Fluorescein iso-thiocyanate conjugated with bovine serum albumin (BSA-FITC) and Texas Red conjugated with Dextran (Dextran-TR, M.W. 70,000) were purchased from Sigma (St. Louis, MO, USA). Triton X-100 was obtained from Kodak (Rochester, NY, USA). All reagents used were of analytical grade.

2.2. Liposome preparation All liposomes were made with Egg PC and different mol% of Pluronic F127. Multi-lamellar

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

vesicles (MLV) and large unilamellar vesicles (LUV) were made for different experiments. Pluronic F127 was mixed with buffered marker solution, where needed, and entrapped inside the vesicles. Lipids, in chloroform, were mixed in a roundbottomed flask and dried under a gentle stream of nitrogen gas to form a thin layer on the flask wall. The film was dried further in a vacuum chamber, for 3 to 4 h, to remove any remaining solvent. MLVs were formed by first re-suspending the dry lipid film with buffered dye solution, with or without Pluronic F127, followed by vortexing. LUVs were formed by extruding the MLV solution through a 0.2 mm polycarbonate filter (Millipore, Bedford, MA, USA) for 15 times or more. All the liposomes were prepared and kept inside a cold room (48C) until used in the experiment. Within experimental error, the amounts of CF or BSA-FITC encapsulated are not dependent on the F127 concentration, as measured by the fluorescence after complete lysis of liposomes by Triton X-100 after the experiment.

2.3. CF release assay Measurement of CF release is a widely used method to determine liposome permeability [21]. Fluorescence of CF, at 100 mM concentration, is self-quenching, and release of the marker in the environment increases the fluorescence due to dilution de-quenching. A solution of 100 mM CF (with 60 mM NaCl and 5 mM phosphate buffer), with appropriate % (w / v) of Pluronic F127, was added to the dry DOPC with different mol% of cholesterol to form the MLVs. The liposomes were then extruded to form LUVs, as described previously in the ‘Liposome preparation’ section. Untrapped CF and the LUVs were separated using a Sephadex G-50 column. An elution buffer comprising 217 mM sucrose and 5 mM phosphate was used in the column to balance the internal osmotic pressure of the liposomes. All the preparation steps were performed inside a cold room (48C). For each experimental sample, 250 ml of the liposome fraction was further diluted, using the same elution buffer as above, to a final solution volume of 3 ml. Fluorescence intensity of the CF was measured using a SLM 8000 fluorimeter. The initial fluorescence (Ex. 492 nm, Em. 518 nm) intensity of a

29

sample at 48C was recorded. The sample temperature during measurement in the fluorimeter was maintained at the desired level using an adjustable thermostat-controlled heating / cooling unit. All other samples used in the same experiment were kept at their respective desired temperatures in water baths. Before measurement, each sample was kept for 15 min in the fluorimeter chamber to bring it to thermal equilibrium. At and beyond this time, the fluorescence readings had reached steady values, indicating an equilibrium release was achieved. Fluorescence of the CF is also temperature dependent. Thus, the fluorescence intensities obtained in the experiments were corrected for any temperature effect. Increase in fluorescence intensity can only be converted to % release of CF if all the measured concentration values fall on the linear part of the fluorescence de-quenching curve. Our observation from a CF calibration curve suggests that such a conversion would be valid in this case. After the release measurement, a 15 ml solution of 10% Triton X-100 was added to the liposome solution in order to completely lyse the liposomes. Fluorescence intensity was measured again after lysis. The % CF release value of a sample at temperature t was calculated using the equation % CF release 5 (IS 2 I0 ) /(IT 2 I0 )*100%

(1)

where IS is the fluorescence intensity value of the sample at temperature t, I0 is the fluorescence intensity value of the sample at 48C, and IT is the total fluorescence intensity value of the sample at temperature t measured after complete lysis of the liposomes using Triton X-100.

2.4. BSA-FITC release assay The release of BSA-FITC from liposomes was measured by donor fluorescence quenching due to fluorescence resonance energy transfer (FRET) [22]. We did not use the fluorescence de-quenching of BSA-FITC because high concentrations of BSAFITC lead, to some extent, to aggregation of the encapsulating MLVs. Lipid solutions comprising DOPC and cholesterol (50:50 by mole) were used to form the MLVs. We used 50 mol% cholesterol, since that composition gave the best encapsulation of BSA-FITC among all the different compositions. A

30

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

solution of 5.5 mg / ml of donor BSA-FITC (with 100 mM NaCl and 5 mM phosphate buffer) at a nonquenching concentration, with appropriate % (w / v) of Pluronic F127, was added to the dry lipids to form the MLVs. The liposomes were then extruded to form LUVs, as described previously in the ‘Liposome preparation’ section. In order to remove any untrapped BSA-FITC molecules, the LUVs were separated by a Sephadex G-100 column, and an appropriate fraction was collected and used for the assay. An elution buffer comprising 125 mM NaCl and 5 mM phosphate was used in the column to balance the internal osmotic pressure of the liposomes. All the preparation steps were performed inside a cold room (48C). For each sample, 50 ml of the liposome fraction was further diluted, using the same elution buffer as above, to a final solution volume of 3 ml. The fluorescence intensity spectrum of BSA-FITC was measured and recorded using a SLM 8000 fluorimeter. The excitation wavelength was kept at 492 nm, the excitation maximum of the donor fluorophore FITC, throughout the experiment. All the fluorescence emission intensity readings were taken at the emission maximum of the donor BSA-FITC (519 nm). The spectra of samples at 48C were first recorded. After recording the initial fluorescence emission intensity of a sample of BSA-FITC alone, 100 ml of a solution of the acceptor Dextran-TR (1 mg / ml) was added to the sample. The fluorescence emission intensity of the sample containing both the donor and the acceptor was then recorded. A 10 ml solution of 10% Triton X-100 was added to the sample solution in order to completely lyse the liposomes. Fluorescence intensity was recorded again after lysis. The procedure was repeated for samples treated at other temperatures. The efficiency (E) of donor quenching due to FRET was calculated according to [22] E 5 [1 2 (Fda /Fd )]*100

(2)

where Fda is the fluorescence intensity of the donor in the presence of the acceptor and Fd is the fluorescence intensity of the donor in the absence of the acceptor. The efficiency of donor quenching can be converted to % release when there are enough acceptor molecules present in the medium to allow energy

transfer from each donor molecule released in the medium. Since the concentration of TR in the sample solution was more than 350 times the maximum concentration of FITC, the above assumption for such conversion would be valid. Thus, we can calculate the % BSA-FITC release of a sample at temperature t using the equation % BSA-FITC release 5 (ES 2 Eini ) / (ETriton 2 Eini )*100%

(3)

where ES is the efficiency of donor quenching of the sample at temperature t, ETriton is the efficiency of donor quenching of the sample at temperature t after complete lysis of the liposomes using Triton X-100, and Eini is the efficiency of donor quenching of the sample at 48C. The fluorescence intensity values were corrected for the background and inner filter effect. The sample temperature was maintained during measurement at the desired level by a thermostat-controlled cuvette holder inside the fluorimeter. All other samples, waiting to be measured, were kept at their respective desired temperature in water baths. Before measurement, each sample was kept for 15 min in the fluorimeter chamber to bring it to thermal equilibrium.

3. Results At high temperatures, polymer-induced content release is quite low for egg PC and some saturated PC liposomes [7]. Our study with egg PC liposomes containing Pluronic F127 showed only a content release of up to 35% (data not shown), supporting the observation of Kim et al. [7]. Previous studies have found that content release, due to the presence of poloxamers, is less for a lipid composition of mostly saturated lipids in comparison to that for a lipid composition of mostly unsaturated lipids [17,23,24]. Thus, one possible explanation for the low release of these Pluronic F127-containing vesicles could be that egg PC, with its near 50:50 composition of saturated / unsaturated lipids, behaves somewhere intermediate between gel and fluid phase lipids. Our choice of host lipid (DOPC) has more unsaturated lipid than egg PC. We have also used

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

31

cholesterol in our lipid composition to reduce release from the control vesicles.

significant release starts at lower temperatures with increasing % of Pluronic content.

3.1. Release of CF from DOPC (with 5 mol% cholesterol) LUVs

3.2. Effect of DS( PEG5000)PE on the release of CF

This experiment shows the effect of different weight % of encapsulated Pluronic F127 on the release of CF at different temperatures (Fig. 1). The control (0% Pluronic) sample shows minimal release throughout the experimental temperature range. All the other four curves represent samples containing encapsulated Pluronic: 0.02, 0.04, 0.08 and 0.16% (w / v; 0.625, 1.25, 2.5 and 5 mol% of lipid in the initial mixture). The sample containing 0.02% (w / v) Pluronic starts showing significant release of about 35% at 308C, and increases gradually to 64% at 458C. The sample containing 0.04% (w / v) Pluronic shows considerable release of about 23% at 228C. The release increases to 60% at 308C and further to 77% at 458C. The sample containing 0.08% (w / v) Pluronic has a release of 23% at 228C. The % release jumps to 76% at 308C and then levels off. The 0.16% (w / v) Pluronic sample shows severe release (60%) at 228C. The % release subsequently increases to a maximum of 87%. There is a general trend in the content release of these samples. CF release is higher with increasing % of Pluronic content. In addition,

Stealth liposomes are effective in increasing the liposome circulation time [2,3] giving the liposomes increased opportunity to deliver internal content at the target site. Therefore, it is important to test content release in a stealth liposome system, simulating in vivo application. This experiment aims to compare the internal content release of DOPC (with 5 mol% cholesterol) LUVs in the presence and absence of PEG-conjugated lipids. One mole percent DS(PEG5000)PE, enough to make a bilayer stealth [25], was added to DOPC and cholesterol to form PEG-containing LUVs. The results are shown in Fig. 2. Samples were made either without PEG lipid (triangular symbols), or with 1 mol% DS(PEG5000)PE (square symbols). The two ‘control’ curves (with filled symbols) represent liposomes containing no Pluronic F127, while the two ‘F127’ curves (with open symbols) represent liposomes with 0.08% (w / v) of encapsulated Pluronic. Although both ‘control’ samples show an increase in release with temperature, the % release is much lower in

Fig. 1. % CF release from DOPC LUVs at different temperatures. All LUVs are made of DOPC and 5 mol% cholesterol. The encapsulated solution comprises CF and different % (w / v) Pluronic F127. The ‘control’ sample (m) has no Pluronic (0%). The ‘F127’ samples are: 0.02% (n), 0.04% (앳), 0.08% (s) and 0.16% (h). The error bar represents variations among at least three repeating samples.

Fig. 2. % CF release from DOPC LUVs at different temperatures. All LUVs are made of DOPC and 5 mol% cholesterol. The PEG samples have an additional 1 mol% of DS(PEG5000)PE in the lipid composition. The encapsulated solution comprises CF alone (control) or CF with 0.08% (w / v) Pluronic F127 (F127). The ‘control’ samples are: no PEG (m) and with PEG (j). The ‘F127’ samples are: no PEG (n) and with PEG (h). The error bar represents variations among at least three repeating samples.

32

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

comparison to the ‘F127’ samples. The ‘control’ sample, with PEG lipids, shows 30% release at 458C. Both the ‘F127’ samples show release characteristics similar to each other. The ‘F127’ sample without PEG lipid shows about 20% release at 228C. This is followed by a significant increase in release (76%) at 308C, leveling off subsequently. The ‘F127’ sample with PEG lipid has a 38% release at 228C. The release gradually increases with increasing temperature, reaching a maximum of about 80%.

3.3. CF release from DOPC LUVs containing 50 mol% cholesterol High mol% of cholesterol (up to 50%) is frequently used in making liposomes for drug delivery [26,27]. Cholesterol helps in attaining increased stability of the liposome as well as reduced leakage of the encapsulated drug [28]. For our purpose, we want to see whether an increase in the amount of cholesterol in the lipid composition affects the CF release from LUVs. All samples consist of LUVs made from a lipid composition of DOPC / cholesterol (50:50 by mole). The encapsulated CF solution contains different % (w / v) of Pluronic F127 for four different samples (0, 0.04, 0.08 and 0.16% w / v). The results are plotted in Fig. 3. The ‘control’ sample, without Pluronic, shows very little release. The CF release in this sample is only about 8% at 468C. At a concentration of 0.04% (w / v) of encapsulated Pluronic, the release is minimal, up to 308C, after which the content release increases abruptly to 70% at 348C. At 388C, the internal content release reaches a plateau of 86%. For the sample containing 0.08% (w / v) of encapsulated Pluronic, the release is small until 268C. The release reaches a value of 84% at 348C and then levels off. The sample with 0.16% (w / v) Pluronic has a 46% CF release at 228C. It reaches a value of about 90% at temperatures of 268C and above. Fig. 4 shows a comparison between the CF release experiments with 5 and 50 mol% cholesterol. The ‘control’ samples, without F127, are very similar to each other. Each of the F127 samples contains 0.08% (w / v) of Pluronic F127. The ‘F127’ sample with 50 mol% cholesterol shows a sharper transition than that with 5 mol% cholesterol.

Fig. 3. % CF release from DOPC LUVs at different temperatures. All LUVs are made of DOPC / cholesterol (50:50 by mol%). The encapsulated solution comprises CF and different % (w / v) Pluronic F127. The ‘control’ sample (m) has no Pluronic (0%). The ‘F127’ samples with different % (w / v) of Pluronic are: 0.04% (앳), 0.08% (s) and 0.16% (h). The error bar represents variations among at least three repeating samples.

3.4. Release of BSA-FITC All the experiments performed so far in our study involve release of the marker CF, which has a molecular weight of 376. A small marker like CF

Fig. 4. A comparison of temperature-dependent CF release from DOPC LUVs with different % of cholesterol in the lipid composition. The LUVs are made of DOPC / cholesterol (95:5 or 50:50 by mol%). The encapsulated solution comprises CF with 0 or 0.08% (w / v) Pluronic F127. The ‘control’ samples, without Pluronic, contain: 5% cholesterol (m) and 50% cholesterol (j). The ‘F127’ samples, with 0.08% (w / v) of Pluronic, contain: 5% cholesterol (n) and 50% cholesterol (h). The error bar represents variations among at least three repeating samples.

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

can escape even through small structural defects in the bilayer. That would not be the case for a much larger molecule, which needs considerably larger pores in the membrane for its release. Thus, release of a large marker would indicate the presence of large defects, even total disruption, of the liposome membrane. As a large marker our choice was BSAFITC, which has a molecular weight of |66,000, two orders of magnitude larger than CF. This experiment intends to show the temperature-dependent release of encapsulated BSA-FITC molecules due to the presence of different % (w / v) of encapsulated Pluronic F127 (Fig. 5). All samples consist of LUVs made from a lipid composition of DOPC and cholesterol (50:50 by mole) encapsulating the BSAFITC solution with different % (w / v) of Pluronic. The ‘control’ curve represents a sample with no encapsulated Pluronic. The other three curves correspond to samples with 0.04, 0.08 and 0.16% (w / v) of encapsulated Pluronic, respectively. At 228C, all four samples show a small % release, ranging from 6 to 12%. The ‘control’ sample has a small % release in the experimental temperature range, the maximum being 19% at 468C. The sample containing 0.04% (w / v) of Pluronic shows a low % release up to 388C. After that the content release jumps to 84% at 428C

Fig. 5. % BSA-FITC release from DOPC LUVs at different temperatures. All LUVs are made of DOPC / cholesterol (50:50 by mol%). The encapsulated solution comprises BSA-FITC and different % (w / v) Pluronic F127. The ‘control’ sample (m) has no Pluronic (0%). The ‘F127’ samples contain different % (w / v) of Pluronic: 0.04% (앳), 0.08% (s) and 0.16% (h). The error bar represents variations among at least three repeating samples.

33

and levels off. The sample with 0.08% (w / v) of Pluronic shows similar release characteristics, except that the sudden increase in % release is at 348C, instead of 388C as observed in the case of the 0.04% (w / v) sample. At 228C, the 0.16% (w / v) Pluronic sample leaks 6% and then increases to about 17% at 268C. Eventually the release increases to about 86% at 308C, subsequently leveling off. All the % release values obtained so far indicate that the onset temperature is governed by the concentration of encapsulated Pluronic. According to our hypothesis, the onset of content release is triggered by the association of the Pluronic molecules with lipid bilayers at temperatures above the CMT. We may now compare our experimentally obtained onset temperature points with experimentally obtained CMT values for free Pluronic F127 at corresponding concentrations [20]. We only used results obtained from experiments with liposomes made of DOPC (with 50% cholesterol), since we have both CF and FITC data using this composition. The onset values for our experimental results were calculated using the inflection point (i.e. mid-point) analysis of the curves with different % (w / v) of Pluronic. Fig. 6 shows such a comparison. For any Pluronic F127 concentration, CF and BSA-FITC onset temperatures are, respectively, lower and higher than the known CMT values [20]. It could be expected that the onset temperature for the CF

Fig. 6. Comparison of inflection points of experimental content release data for CF and BSA-FITC, for samples made of DOPC / cholesterol (50:50 by mol%), at different % (w / v) of Pluronic F127. The corresponding CMTs are obtained from previously published results [20]. The curves represent: onset for CF release (j), CMT value [20] (앳) and onset for BSA-FITC release (m).

34

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

system would be lower than that of the BSA-FITC system, since CF would be released through smaller bilayer defects than BSA-FITC. For BSA-FITC, complete release is expected to take place at a higher temperature than CMT because such release needs significant bilayer disruption.

4. Discussion The goal of designing temperature-sensitive liposomes is the release of internal content, e.g. drug, at the targeted site in response to a precise external thermal stimulation. The temperature-sensitive association of Pluronic F127 with the liposome membranes makes the integrity of liposomes thermally sensitive. The earlier conception of poloxamer association with lipid membranes was that the poloxamers merely adsorb to the membrane surface [23]. Further studies indicated a possible insertion of the hydrophobic moieties into the lipid bilayer [29–31]. Fig. 7 shows a schematic representation of the possible modes of the Pluronic F127 molecule associating with a liposome membrane. There are two possibilities by which a Pluronic F127 molecule can incorporate itself in the bilayer. One is through the bilayer, while the other is anchoring the PPO part of the Pluronics inside the core but remaining on the same side of the bilayer as the temperature is raised. Above the CMT, Pluronic molecules would start to

incorporate themselves into the bilayer. When a sufficient amount of Pluronic is incorporated into the bilayer walls of liposomes, the bilayer can be significantly weakened or disrupted, forming large defects, leading to the release of contents. This forms the basis of temperature-induced content release. The disruption of the liposomes results also in the release of Pluronic F127 into the suspension medium. The released F127 was too diluted to dissolve liposomes even at the highest concentration of F127 encapsulated (0.16%). We could detect liposomes (slightly smaller than the 200 nm diameter liposomes before heat treatment) by negative stain electron microscopy and by dynamical light scattering in high temperature treated samples. These liposomes could be the product of resealing and / or reformation that took place at the CMT. The incomplete disruption may be the reason that, for all F127 concentrations, the maximum release reached only 80–90%. Nevertheless, the CMT transition is very sharp, giving a sharp temperature threshold for content release. In all of our experiments, we encapsulated the Pluronic F127 molecules along with the marker solution, instead of co-solubilizing Pluronic molecules with the lipid during the preparation of the liposomes. The advantage of this approach is that by using an aqueous solution of the Pluronic of known concentration, we can control the concentration of Pluronic in the encapsulated solution exactly. In contrast, using the co-solubilization approach, we

Fig. 7. A schematic diagram of the possible modes of association of Pluronic F127 molecules with the lipid bilayer. The inset shows a magnified version with Pluronic molecules either penetrating through the bilayer or anchored to the bilayer interior.

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

would not be able to control the exact concentration of the encapsulated Pluronic.

4.1. Effect of DS( PEG5000)PE We found that the presence of DS(PEG5000)PE does not alter the thermal sensitivity of our liposomes. The result of this experiment (Fig. 2) shows some apparently intriguing information: both curves, with and without PEG lipid, show similar release characteristic. Since the PEG lipid is also present on the inner side of the LUV membrane, the steric repulsion from the surface-anchored PEG on that side is expected to prevent Pluronic molecules from approaching the bilayer, resulting in an inhibition of content release. However, our result indicates that such stealthing is ineffective for Pluronic molecules. At 1 mol% concentration, the surface-anchored PEG, in the mushroom configuration, should completely shield the bilayer surface against cell adhesion and protein adsorption [25]. However, Pluronic F127, being much smaller than cells and proteins, may not be sterically hindered by the PEG molecules. Penetration of certain molecules [32], including polymers such as poly(2-ethylacrylic acid) [31], through a PEG protection layer has been observed and explained in terms of possible polymer–polymer interactions [31]. Thus, the similar behavior of samples with and without PEG is not entirely unexpected.

4.2. Effect of cholesterol At 50 mol% cholesterol (Fig. 3) the transition in release is considerably sharper than liposomes with 5 mol% cholesterol (Fig. 1). The % release for the control sample with 50% cholesterol is lower than that of the control sample in the experiment using 5% cholesterol (Fig. 1), especially at temperatures below the CMT. This is probably due to the presence of an increased amount of cholesterol in the lipid composition rendering the bilayer more stable and less prone to casual release of internal content when subjected to the low level attack by Pluronic monomers. This is in agreement with the interpretation that cholesterol in lipid bilayers increases the elastic area expansion modulus, thereby helping to resist polymer binding and intercalation into the bilayer [31,33]. However, at temperatures greater than the

35

CMT of F127, the CF releases caused by all F127 concentrations tested are slightly higher for samples containing 50% cholesterol than for those containing 5% cholesterol. The bilayer disruption in this case is by a different mechanism that is more severe than polymer binding and intercalation, and a higher percentage of cholesterol may favor such a disruption process because of the less ideal mixing of cholesterol and DOPC [34].

4.3. Effect of molecular weight of encapsulated molecules The result of the BSA-FITC experiment (Fig. 5) shows a similar trend in release characteristics as seen in the CF experiment (Fig. 3). The transition is sharper in the BSA-FITC experiment, which is expected since a large molecule like BSA-FITC would not leak at lower temperatures when the bilayer defects are likely to be smaller. Also, the transition is taking place at higher temperatures than for the CF experiment (Fig. 6). One could argue that the critical solution / micelle transition of F127 is not exactly first degree, that a certain percentage of polymer is weakly interacting with the bilayer even below the CMT, thus causing the leakage of CF but not BSA-FITC. The release of the latter awaits the completion of the critical solution / micelle transition when the entire population of F127 is available to disrupt the bilayer. However, one must bear in mind that the CMT values for F127 were obtained from polymer–water systems without the presence of lipid bilayers. The presence of liposomes may alter the balance such that the exact CMT as published is not exactly relevant, and the transition mid-point and width are lipid dependent. In this case, the correlation of CMT with content release results is more complex. We chose BSA-FITC, as a large fluorescently labeled molecule, over other candidates (FITC-Dextrans of M.W. 40,000 or 70,000) simply because the latter caused severe aggregation of the lipid vesicles, even at lower concentrations. We chose a very large molecule (Dextran-TR, M.W. |70,000) as our acceptor to avoid any back-diffusion of the acceptor to the interior of the LUV. The sharp transitions in the curves showing nominal to nearcomplete % release indicate possible disruption of the liposome bilayer.

36

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37

5. Conclusion We have shown that the onset for content release from liposomes encapsulating Pluronic F127 is determined by the CMT of the poloxamer. Using this relation, we may set the content release at any given temperature by adjusting the Pluronic concentration. By varying the concentration of the encapsulated Pluronic F127, we have established that the onset temperature increases with decreasing Pluronic concentration. Within the range of concentration used, we have achieved temperature-controlled complete internal content release from DOPC vesicles at precisely defined temperatures. Increased cholesterol % in the lipid composition produced a sharper transition in content release. The onset temperature also depends on the size of the encapsulated molecules. Lastly, we have demonstrated that this system would work equally well using stealth liposomes.

[8]

[9]

[10]

[11]

[12]

[13] [14]

Acknowledgements [15]

We thank Dr. Paschalis Alexandridis for his valuable insight and suggestions regarding this work. This work was partially supported by grant GM30969 to S.W.H. from the National Institutes of Health.

[16]

[17]

References [18] [1] D.D. Lasic (Ed.), Liposomes: From Physics To Applications, Elsevier, Amsterdam, 1993. [2] M.C. Woodle, D.D. Lasic, Sterically stabilized liposomes [review], Biochim. Biophys. Acta 1113 (2) (1992) 171–199. [3] G. Blume, G. Cevc, Molecular mechanism of the lipid vesicle longevity in vivo, Biochim. Biophys. Acta 1146 (2) (1993) 157–168. [4] R.M. Straubinger, N. Duzgunes, D. Papahadjoupoulos, pHsensitive liposomes mediate cytoplasmic delivery of encapsulated macromolecules, FEBS Lett. 179 (1) (1985) 148– 154. [5] J. Connor, L. Huang, Efficient cytoplasmic delivery of a fluorescent dye by pH-sensitive immunoliposomes, J. Cell Biol. 101 (2) (1985) 582–589. [6] J.N. Weinstein, R.L. Magin, M.B. Yatvin, D.S. Zaharko, Liposomes and local hyperthermia: selective delivery of methotrexate to heated tumors, Science 204 (1979) 188–191. [7] J.-C. Kim, S.K. Bae, J.-D. Kim, Temperature sensitivity of liposomal lipid bilayers mixed with poly(N-iso-

[19]

[20]

[21]

[22]

propylacrylamide-co-acrylic acid), J. Biochem. 121 (1997) 15–19. K. Kono, R. Nakai, K. Morimoto, T. Takagishi, Thermosensitive polymer-modified liposomes that release contents around physiological temperature, Biochim. Biophys. Acta 1416 (1999) 239–250. G.R. Anyarambhatla, D. Needham, Enhancement of the phase transition permeability of DPPC liposomes by incorporation of MPPC: a new temperature-sensitive liposome for use with mild hyperthermia, J. Liposome Res. 9 (4) (1999) 491–506. M.B. Yatvin, W. Kreutz, B.A. Horwitz, M. Shinitzky, pHsensitive liposomes: possible clinical implications, Science 210 (1980) 1253–1255. M. Maeda, A. Kumano, D.A. Tirrell, H 1 -induced release of contents of phosphatidylcholine vesicles bearing surface bound polyelectrolyte chains, J. Am. Chem. Soc. 110 (1988) 7455–7459. M. Zignani, D.C. Drummond, O. Meyer, K. Hong, J.-C. Leroux, In vitro characterization of a novel polymeric-based pH-sensitive liposome system, Biochim. Biophys. Acta 1463 (2) (2000) 383–394. C. Pidgeon, C.A. Hunt, Light sensitive liposomes, Photochem. Photobiol. 37 (5) (1983) 491–494. V.C. Anderson, D.H. Thompson, Triggered release of hydrophilic agents from plasmalogen liposomes using visible light or acid, Biochim. Biophys. Acta 1109 (1) (1992) 33–42. H. Hayashi, K. Kono, T. Takagishi, Temperature-dependent associating property of liposomes modified with a thermosensitive polymer, Bioconjug. Chem. 9 (3) (1998) 382–389. M. Jamshaid, S.J. Farr, P. Kearney, I.W. Kellaway, Poloxamer sorption on liposomes: comparison with polystyrene latex and influence on solute efflux, Int. J. Pharm. 48 (1988) 125–131. M.C. Woodle, M.S. Newman, F.J. Martin, Liposome leakage and blood circulation: comparison of adsorbed block copolymers with covalent attachment of PEG, Int. J. Pharm. 88 (1992) 327–334. P. Alexandridis, Poly(ethylene oxide) poly(propylene oxide) block copolymer surfactants [review], Curr. Opin. Colloid Int. Sci. 2 (5) (1997) 478–489. P. Alexandridis, T.A. Hatton, in: V. Pillai, D.O. Shah (Eds.), Dynamic Properties of Interfaces and Association Structures, AOCS Press, Champaign, IL, 1996, pp. 231–262, Chapter 12. P. Alexandridis, J.F. Holzwarth, T.A. Hatton, Micellization of poly(ethylene oxide)–poly(propylene oxide)–poly(ethylene oxide) triblock co-polymers in aqueous solutions — thermodynamics of copolymer association, Macromolecules 27 (9) (1994) 2414–2425. J.N. Weinstein, S. Yoshikami, P. Henkart, R. Blumenthal, W.A. Hagins, Liposome–cell interaction: transfer and intracellular release of a trapped fluorescent marker, Science 195 (4277) (1977) 489–492. K.-H. Cheng, T. Wiedmer, P.J. Sims, Fluorescence resonance energy transfer study of the associative state of membranebound complexes of complement proteins C5b-8, J. Immunol. 135 (1) (1985) 459–464.

P. Chandaroy et al. / Journal of Controlled Release 76 (2001) 27 – 37 [23] M.A. Khattab, S.J. Farr, G. Taylor, I.W. Kellaway, The in vitro characterization and biodistribution of some non-ionic surfactant coated liposomes in the rabbit, J. Drug Target. 3 (1) (1995) 39–49. [24] S.M. Moghimi, C.J.H. Porter, L. Illum, S.S. Davis, The effect of poloxamer-407 on liposome stability and targeting to bone marrow: comparison with polystyrene microspheres, Int. J. Pharm. 68 (1991) 121–126. [25] H. Du, P. Chandaroy, S.W. Hui, Grafted poly-(ethylene glycol) on lipid surfaces inhibits protein adsorption and cell adhesion, Biochim. Biophys. Acta 1326 (2) (1997) 236–248. [26] M.S. Webb, T.O. Harasym, D. Masin, M.B. Bally, L.D. Mayer, Sphingomyelin-cholesterol liposomes significantly enhance the pharmacokinetic and therapeutic properties of vincristine in murine and human tumor models, Br. J. Cancer 72 (4) (1995) 896–904. [27] I. Ogihara-Umeda, S. Kojima, Cholesterol enhances the delivery of liposome-encapsulated gallium-67 to tumors, Eur. J. Nucl. Med. 15 (9) (1989) 612–617. [28] C. Kirby, J. Clarke, G. Gregoriadis, Effect of cholesterol content of small unilamellar liposomes on their stability in vivo and in vitro, Biochem. J. 186 (1980) 591–598. [29] K. Kostarelos, P.F. Luckhamand, Th.F. Tadros, Addition of (tri-)block copolymers to phospholipid vesicles: a study of

[30]

[31]

[32]

[33]

[34]

37

the molecular morphology and structure using hydrophobic dye molecules as bilayer probes, J. Colloid Interface Sci. 191 (2) (1997) 341–348. J.D. Castile, K.M.G. Taylor, G. Buckton, A high sensitivity differential scanning calorimetry study of the interaction between poloxamers and dimyristoylphosphatidylcholine and dipalmitoylphosphatidylcholine liposomes, Int. J. Pharm. 182 (1) (1999) 101–110. J.K. Mills, G. Eichenbaum, D. Needham, Effect of bilayer cholesterol and surface grafted poly(ethylene glycol) on pH-induced release of contents from liposomes by poly(2ethylacrylic acid), J. liposome Res. 9 (2) (1999) 275–290. D. Needham, T.J. McIntosh, D.V. Zhelev, Surface chemistry of the sterically stabilized PEG-liposomes: general principles, in: A. Janoff (Ed.), Liposomes Rational Design, Marcel Dekker, New York, 1998, pp. 13–62. D. Needham, J.K. Mills, G. Eichenbaum, Interaction between poly(2-ethylacrylic acid) and lipid bilayer membranes: effects of cholesterol and grafted poly(ethylene glycol), Faraday Discuss.: Molecular Interactions of Biomembranes 111 (1998) 103–110. S.W. Hui, N.-B. He, Molecular organization in cholesterol– lecithin bilayers by x-ray and electron diffraction measurements, Biochemistry 22 (1983) 1159–1164.