Liposomal formulations of lipophilic compounds

Abstract

The present invention relates to the preparation of liposomes with enhanced loading capacity for pharmaceutically and/or diagnostically active agents and/or cosmetic agents which are substantially solubilized by the liposomal membranes, to liposome dispersions with enhanced stability with respect to release of the active agent and/or cosmetic agent from the liposomes obtainable by the process, and to pharmaceutical or cosmetic compositions comprising said stabilized liposome dispersions. The preparation may involve dehydration and rehydration steps of liposome dispersions which may be carried out by spray drying.

Claims

1. A process of preparing a stable liposomal preparation comprising a taxane, the process comprising: i) providing a stressed liposomal preparation comprising at least one saccharide, wherein liposomes of the stressed liposomal preparation comprise a cationic lipid and the stressed liposomal preparation comprises a positive zeta potential in about 0.05 M KCl solution at about pH 7.5 at room temperature, wherein an osmolar gradient is generated to provide the stressed liposomal preparation, ii) incubating the stressed liposomal preparation with a taxane, thereby obtaining a stressed liposomal preparation comprising a taxane incorporated into the liposomal membrane, iii) separating unsolubilised taxane from the stressed liposomal preparation, and iv) dehydrating by spray drying the stressed liposomal preparation to obtain a dehydrated liposomal preparation, and v) optionally, rehydrating the dehydrated liposomal preparation to obtain a stressed liposomal preparation, wherein the osmolar gradient maintains a stress on the liposomal preparation and stabilizes the taxane in the liposomal preparation and wherein the at least one saccharide is retained in the liposomes to maintain the osmolar gradient before and after rehydration.

2. The process according to claim 1, wherein the taxane is paclitaxel or a derivative thereof.

3. The process of claim 1, wherein separating unsolubilised taxane from the stressed liposomal preparation comprises filtration or centrifugation.

4. The process of claim 1, wherein the stressed liposomal preparation of step (ii) comprises DOTAP, DOPC, and paclitaxel.

5. The process of claim 4, wherein the stressed liposomal preparation of step (ii) comprises liposomes containing DOTAP, DOPC, and paclitaxel in a molar ratio of about 50:47:3.

6. The process of claim 1, wherein the saccharide is selected from mono-, di-, oligo- or poly-saccharides.

7. The process of claim 6, wherein the saccharide is trehalose.

8. The process of claim 1, further comprising rehydrating the dehydrated liposomal preparation comprising the taxane, wherein the rehydrated liposomal preparation comprises liposomes with homogeneous size distribution.

9. The process of claim 1, wherein one hour after rehydration, the liposomal preparation has a PI of less than 0.3, less than 0.25, or less than 0.20.

10. The process of claim 1, wherein the stressed liposomal preparation comprises a lower concentration of the saccharide in a free aqueous phase outside the liposomes as compared to a concentration of the saccharide encapsulated in an aqueous phase inside the liposomes.

11. The process of claim 1, wherein the osmolar gradient is generated by diluting the liposomal preparation with an aqueous medium to obtain the stressed liposomal preparation.

12. The process of claim 1, wherein the osmolar gradient is generated by dialyzing the liposomal preparation against an aqueous medium to obtain the stressed liposomal preparation.

13. The process of claim 1, wherein a range of the osmolar gradient maintaining the stress on the liposomal preparation is a concentration difference of between about 5% to about 30% (w/w).

Description

FIGURE LEGENDS

(1) FIG. 1: Amount of paclitaxel solubilised by liposome preparations with a lipid concentration of 10 mM and a total trehalose concentration of 10% (w/w). The liposomes were obtained from concentrated precursor formulations by dilution with water. The abscissa shows the initial concentration, i.e., the data at 10% trehalose were obtained from a suspension which was not diluted, and the data at 40% trehalose were obtained from an undiluted suspension which was originally 40% (w/w) in trehalose, and 40 mM in lipid, and which was diluted 1+3 with water.

(2) FIG. 2: Count rate of 10 mM DOTAP/DOPC liposomes obtained from a concentrated formulation of 30 mM lipid in 30% (w/w) trehalose. The solution was diluted with water/trehalose mixtures to different overall trehalose concentration between 30% and 10% (w/w). All formulations had the same final lipid concentration of 10 mM, but the trehalose concentration gradient between encapsulated and free aqueous phase was as indicated by the x-axis.

EXAMPLES

(3) Loading of Paclitaxel to Liposomes with Different Trehalose Gradients

(4) Summary

(5) The effect of osmolar gradients on the partitioning of paclitaxel in liposomes at equilibrium with a saturated aqueous phase was investigated. Liposomes with different osmolar gradients were produced and incubated to paclitaxel crystals. All formulations had the same composition; more precisely, the lipid concentration and the trehalose concentration were c.sub.lipid=10 mM and c.sub.trehalose=10% (w/w). Some of the formulations were prepared and extruded at higher lipid and trehalose concentration, and diluted with water after extrusion to the final concentration. In this way, the free aqueous phase was diluted, but the encapsulated aqueous phase was not diluted (neglecting swelling effects and solute exchange through defects). An osmolar gradient between the encapsulated and the free aqueous phase was established, which increased with increasing dilution. The so formed liposomes were incubated with dry paclitaxel and the amount of paclitaxel which was solubilised by the liposomes was determined. A monotonous increase of solubilised paclitaxel with increasing dilution (concentration gradient) was found. The results indicate that the amount of paclitaxel which partitions in the liposome membrane at equilibrium increases with increasing osmolar gradient.

(6) TABLE-US-00001 MATERIALS Paclitaxel, Lot 06/150 Cedarburg Pharmaceuticals DOTAP, Lot MBA 113 Merck Eprova DOPC, Lot G181PC49 Avanti Polar Lipids Water, Milli-Q-Synthesis Millipore Trehalose-Dihydrate, highly pure Senn Chemicals Chloroform, p.a. Merck Acetonitrile, HPLC grade (ACN) Merck Tetrahydrofuran, HPLC grade (THF) Merck Ammonium acetate, p.a. Merck Trifluoroacetic acid, p.a. Merck Syringe filter minisart Sartorius 0.2 m pore size, 25 mm diameter Membrane: Cellulose acetate HPLC System 1100 Agilent Degasser (G1379A) Binary Pump (G1312A) Thermostated autosampler (Autoinjektor G1329A, Thermostat G1330B) Thermostated column compartment (G1316A) Diode array detector (G1315B) or variable wavelength detector (G1314A) ChemStation for LC 3D, Rev. A.09.01 Extruder, 10 ml Northern Lipids Zetasizer 3000 Malvern Instruments

(7) Methods

(8) Preparation of Empty Liposomes

(9) DOTAP/DOPC-formulations (1:1 ratio), or formulations comprising only DOTAP or DOPC, were prepared by the film method. The required amounts of lipids were weighed into a round flask and dissolved in chloroform. The solvent was evaporated to dryness in a rotary evaporator (Heidolph, Germany) at a pressure of about 150 mbar at a temperature of about 40 C. for about 15 minutes. The film was dried at 10 mbar over 60 minutes and subsequently hydrated in a trehalose solution in water by gently shaking the flask. Amounts of lipids, trehalose concentration and volume of trehalose solution were chosen to result in suspensions comprising lipid concentrations between 10 mM to 40 mM and trehalose concentrations between 9.8% and 39.2% (w/v) for DOTAP/DOPC-formulations and between 10 mM to 30 mM lipid concentration and between 9.8% and 29.4% (w/v) trehalose concentrations for DOTAP or DOPC only formulations. The resulting suspensions of multilamellar liposomes were extruded five times through a polycarbonate membrane of a pore size of 200 nm at a pressure of about 5 bar. After extrusion, the suspensions were diluted with water to yield suspensions with a lipid concentrations of 10 mM and a total trehalose concentration of 9.8% (w/v).

(10) Paclitaxel Loading

(11) 5 ml of the suspensions comprising empty liposomes prepared as described above were added to 2.6 mg dry paclitaxel (corresponding to a theoretical paclitaxel concentration of 600 M) in 15 ml Falcon tubes. The batches were stirred for 1 h at room temperature (magnetic stirrer).

(12) After stirring, non liposomal bound paclitaxel was separated by filtration of 2 ml of each batch through a syringe filter (Sartorius minisart, 0.2 m, cellulose acetate membrane). Paclitaxel- and lipid concentration in the resulting filtrates was subsequently analysed by HPLC.

(13) Analytical Methods

(14) Determination of Paclitaxel Content

(15) Samples were diluted in ACN/THF/2 mM ammonium acetate 48/18/34 (v/v/v). Stationary phase: LiChroCART 250-4; LiChrospher 60, RP-select B length 250 mm, ID: 4 mm, particle size 5 m Mobile phase: ACN/THF/2 mM ammonium acetate 32/12/56 (v/v/v) Flowrate: 1 ml/min Temperature column compartment: 35 C. Detector wavelength: 229 nm Injected volume: 10 l Runtime: 40 min

(16) Determination of Lipid Content

(17) The lipid content of the batches before and after filtration was analysed by HPLC to monitor a potential loss of liposomal material by the filtration process. Samples were diluted in ACN/water 50/50. Stationary phase: Phenomenex Luna 5 C8(2) 100 , 150 mm2 mm Mobile phase: Acetonitrile with 0.1% TFA, water with 0.1% TFA

(18) Gradient Lipid Determination;

(19) TABLE-US-00002 time (min) ACN (%) 0 50 4.12 50 7.06 75 14.13 100 21.20 100 23.56 50 30.00 50 Flowrate: 0.4 ml/min Temperature column compartment: 45 C. Detector wavelength: 205 nm Injected volume: 5 l Runtime: 30 min

(20) Results

(21) DOTAP/DOPC-Formulations

(22) In FIG. 1 the amount of paclitaxel which was solubilised by incubation with different DOTAP/DOPC liposome formulations is shown. All formulations contained the same amount of lipid (liposomes) and trehalose. They were obtained from different more concentrated formulations by dilution with water. Except of the absolute concentration, all originating formulations were equivalent and they were treated equally. The abscissa gives the initial trehalose concentration, before dilution. Because in all cases the final overall trehalose concentration was 10%, the trehalose gradient increases with increasing values of the abscissa. As can be seen, the amount of paclitaxel which was solubilised by the liposomes increased monotonously with the trehalose gradient (dilution). The loading capacity of the liposomes increased with increasing trehalose gradient.

(23) DOTAP- and DOPC-Formulations

(24) The following table shows the amount of paclitaxel solubilised by DOTAP and DOPC liposome formulations (single components) in dependence of the initial trehalose concentration used for preparation:

(25) TABLE-US-00003 TABLE 1 DOTAP- and DOPC-formulations Paclitaxel c.sub.0 trehalose concentration (M) (%) DOTAP DOPC 9.8 163 159 12.3 192 167 14.7 225 192 17.2 297 224 19.6 289 217 24.5 297 289 29.4 339 221

(26) Also for the liposomes from pure lipid a clear dependence on the solubilisation capacity from the trehalose gradient was found. For DOTAP formulations, an increase of the paclitaxel loading capacity with an increasing trehalose concentration difference inside and outside the liposome comparable to the results for DOPTAP/DOPC formulations was observed. The effect was less pronounced liposomes consisting of 100% DOPC.

(27) 2 Loading of Paclitaxel to Liposomes with Different Trehalose Gradients After Spray-Drying

(28) 2.1 Summary

(29) Aim of this example was to test if the positive effect of the osmotic gradient on the paclitaxel loading to liposomes is present if the liposomes are spray dried between formation and adjustment of the trehalose gradient. DOTAP/DOPC liposomes were prepared at two different concentrations, namely 10 mM lipid in 10% (w/w) trehalose solution and 20 mM lipid in 20% (w/w) trehalose solution. Both formulations were spray dried at the respective concentration. The spray dried powders were both reconstituted with water to a lipid concentration of 10 mM and a corresponding trehalose concentration of 10% (w/w). The liquid formulations were exposed to paclitaxel as described above, and the amount of solubilised paclitaxel was determined. It was found, that the formulation which was spray dried from the double concentrated state (20 mM lipid/20% (w/w) trehalose) formulation solubilized more paclitaxel than the one, which was spray dried form the single concentrated state (10 mM lipid, 10% w/w trehalose).

(30) The results indicate that the trehalose distribution inside/outside the liposomes was not affected by spray drying under the selected conditions. After reconstitution of the formerly double concentrated product, liposomes with a trehalose concentration gradient were obtained, correspondingly to the effect of direct dilution of the liquid formulation.

(31) 2.2 Methods

(32) Liposome Formation

(33) The formulations were prepared by ethanol injection. The appropriate amounts of lipid solution in ethanol (200 mM DOTAP-Cl, 188 mM DOPC) were injected under stirring into a solution of trehalose in water. The trehalose concentration was 20% (w/w) for the 20 mM liposomes and 10% (w/w) for the 10 mM liposomes. The required amount of lipid solution in ethanol was about 2.5 ml/l for the 10 mM formulation and 5 ml for the 20 mM formulation.

(34) The resulting polydisperse liposomes formulations were extruded five times across polycarbonate membranes of 200 nm pore size at a pressure of about 5 bar.

(35) Spray Drying

(36) Spray drying was performed with a Niro SD micro spray dryer using a two fluid nozzle. Spraying conditions were as follows: Outlet temperature=100 C., inlet temperature=145 C., feed rate=340 g/h, atomizer gas rate 2.3 kg/h, drying gas rate 30 kg/h.

(37) Reconstitution

(38) The dry powders, both from the previously 10 mM and the previously 20 mM formulation, were reconstituted with water to the lipid concentration of 10 mM.

(39) Paclitaxel Loading Assay

(40) Paclitaxel loading to the liposomes was performed as described in Example 1.

(41) 2.3 Results

(42) The results as obtained from the paclitaxel loading to the reconstituted powders are shown in table 2. As can be seen the formulation which was spray dried at 20 mM lipid concentration solubilised much more paclitaxel than the formulation with the initial lipid concentration of 10 mM. It is concluded, that the elevated paclitaxel loading for the formerly 20 mM formulation was due to an osmolar gradient between encapsulated and free aqueous phase, which was not present in the formerly 10 mM formulation. Spray drying and reconstitution of the dry powder did not lead to trehalose equilibration between the interior and the exterior aqueous phase and therefore the osmolarity of the encapsulated aqueous phase was higher in case of the formerly 20 mM formulation.

(43) Table 2: Solubilisation of paclitaxel by formulations obtained by reconstitution of spray dried powders. The lipid and trehalose concentration was identical in both cases (10 mM lipid, 10% w/w trehalose), but before spray drying one formulation was 10 mM lipid, 10% trehalose and the other formulation was 20 mM lipid, 20% trehalose.

(44) TABLE-US-00004 Initial lipid concentration Concentration of of the formulation solubilised paclitaxel 10 mM (PD_L_07030) 164 (M) 20 mM (PD_L_07031) 340 (M)

(45) 3. Stability of Loaded Liposomes

(46) 3.1 Summary

(47) To evaluate the question, whether liposome preparations with an inside/outside trehalose gradient not only have a higher loading capacity, but also a greater stability with regard to the release of paclitaxel, the release of paclitaxel from the liposomes was traced as a function of time. Formulations with a relatively high paclitaxel fraction, namely 5 mol %, and with different trehalose gradients between 0% and 20% (w/w) were prepared. Liposomes comprising a trehalose gradient did not show any substantial paclitaxel release within the tested period of 21 days, while in liposomes without a trehalose gradient, the retained fraction of paclitaxel decreased to less than 1%.

(48) 3.2 Method

(49) DOTAP/DOPC Formulations

(50) DOTAP/DOPC liposomes comprising about 5 mol % paclitaxel (for exact values see table), 10 to 30 mM lipids, and 9.8% to 29.4% (w/v) trehalose were prepared according to the above described film method by adding the respective amount of lipids and paclitaxel to the chloroform solution. Subsequently the 30 mM batches were diluted to a lipid concentration of 10 mM (overall trehalose concentration 9.8% w/v) with water.

(51) The samples were stored at 4 C. and the paclitaxel content of the liposomes was determined after 0, 1, 5, 14, and 21 days by the above described method using filtration and HPLC analysis.

(52) 3.3 Results

(53) The results are summarized in Table 3. The concentration of paclitaxel (m) retained in the liposomes is shown. The lipid concentration was 10 mM, therefore, the paclitaxel concentration of 100 M corresponds to a molar concentration with respect to lipid of 1%.

(54) In the formulation without trehalose gradient, the retained trehalose fraction monotonously decayed of to a value of less that 100 M (less than 1 mol % with respect to lipid) after 21 days. In contrary, with trehalose gradient the retained paclitaxel did not fall below 400 M. No monotonous decay was observed in that case, in other words, it appears that the value of about 400 M represents a physically stable state of paclitaxel in the liposomes.

(55) TABLE-US-00005 TABLE 3 Retention of paclitaxel in DOTAP/DOPC liposomes. trehalose concentration gradient 0 10% 20% Liposomally retained paclitaxel (M) before filtration t (d) after filtration 470 444 425 0 453 446 411 1 438 428 407 5 391 429 420 14 90 427 412 21 79 409 400

(56) The final fractions of retained paclitaxel are similar to the values as obtained from the loading assay for equivalently treated liposomes. Therefore, the data from the loading assay can be taken as predictive for the stability limit of loaded liposomes. If no other effects take place, the numbers from the loading assay will give information about the amount of paclitaxel which is retained by the liposomes under the given conditions. As a further conclusion from the present examples, it appears that the trehalose gradient, and the improved stability, is fully maintained for several days. In the present case, the effect was maintained for 21 days.

(57) 4 Methods for Determination of Trehalose Gradients in Liposome Preparations In Situ

(58) 4.1 Summary

(59) Concentrated liposome formulations in trehalose at a concentration c.sub.1 were prepared and diluted either with water or with trehalose solution in different ratios to obtain media with trehalose concentration c.sub.2, where c.sub.2c.sub.1. All formulations had the same final lipid concentration of 10 mM. The formulations were analysed based on local changes of optical properties (refractive index). Count rates of dynamic light scattering measurements were used to demonstrate the changes of scattering intensity. With increasing trehalose gradient, c.sub.1-c.sub.2, the count rate of dynamic light scattering (PCS) measurements monotonously increased.

(60) 4.2 Method

(61) Dynamic light scattering was measured with a goniometer BI-200SM from Brookhaven Instruments (Holtsville USA). Measurements were performed with a 30 mW laser of a of 641 nm wavelength at an angle of 90. For data analysis inverse Laplace transformation with optimize regularization techniques was performed.

(62) 4.3 Samples

(63) DOTAP/DOPC (molar ration 1:1) liposomes with a total lipid concentration of 30 mM were prepared in a solution of 30% trehalose. The 30 mM lipid 30% trehalose liposome preparation was extruded across 200 extrusion membranes. Subsequently, the liposomes were diluted with water, 30% trehalose solution or mixtures thereof in ratios as indicated in the table. A: 30 mM liposomes in 30% trehalose B: 30% trehalose in water C: Water

(64) TABLE-US-00006 TABLE 4 Dilution protocol for 10 mM lipid formulations in an aqueous phase with trehalose at concentrations between 10% and 30% (w/w) # Vol. A Vol. B Vol. C Final composition 1 1 2 10 mM liposomes in 30% trehalose 2 1 1.5 0.5 10 mM liposomes in 25% trehalose 3 1 1 1 10 mM liposomes in 20% trehalose 4 1 0.5 1.5 10 mM liposomes in 15% trehalose 5 1 2 10 mM liposomes in 10% trehalose

(65) The lipid concentration in the final preparation was always 10 mM, but the overall trehalose concentration varied between 10% (dilution with water) and 30% (dilution with 30% trehalose solution). With the initial trehalose concentration of 30%, this resulted in a numerical trehalose concentration gradient between 0% (total concentration=30%) and 20% (total concentration=10%). One hour after dilution, Dynamic light scattering measurement was performed.

(66) 4.4 Results

(67) FIG. 2 shows results of the dynamic light scattering measurements. The count rate is given as a function of trehalose gradient. The count rate monotonously increased with increasing trehalose gradient. This can be correlated to the increase of refractive index gradient and swelling effects on increasing trehalose gradient. It is well known, that liposomes can act as ideal osmometers, and osmolar gradients can be determined from light scattering and light absorption properties under suitable conditions (de Gier 1993; Cabral, Hennies et al. 2003). Besides the intensity from quasi elastic light scattering, also other light scattering techniques as well as absorption or turbidimetry measurements can be used. The present observations indicate that analyzing the count rate in dynamic light scattering measurements can be used as a tool to control success of trehalose gradient formation for a given formulation. Furthermore the data confirm that the change of the osmolarity of the aqueous medium of a liposomal suspension renders the physical properties of the liposomes.

(68) 5 Influence of a Trehalose Gradient on the Physical Stability of Liquid DOTAP/DOPC Liposome Formulations of Paclitaxel

(69) 5.1 Summary

(70) In this example the stabilizing effect of trehalose gradients on DOTAP/DOPC liposome formulations of paclitaxel as shown by Example 3 was further investigated. Liposomes were prepared at a concentration of 30 mM (in 32% w/w trehalose solution), diluted with different trehalose/water solutions, and paclitaxel release after mechanical stress was determined.

(71) It was found, that the physical stability increased with increasing trehalose gradient. The findings confirmed the stabilizing effect of trehalose gradients on paclitaxel comprising liposomes also for processing at pilot scale.

(72) 5.2 Methods and Materials

(73) Liposome Manufacturing

(74) Liposomes were produced by the ethanol injection technique. Briefly, a solution of 200 mM DOTAP and 188 mM DOPC (total lipid concentration 388 mM) was injected under stirring at a temperature of 2-8 C. into the aqueous phase (8.11 ml of lipid solution in ethanol for 100 ml of aqueous phase) to yield polydisperse liposomes with a lipid concentration of about 30 mM. For the aqueous phase a solution of 32.1% w/w trehalose dihydrate with 184.5 M citric acid was selected.

(75) Extrusion was performed as indicated at a pressure of 3 bar with polycarbonate membranes of 220 nm pore size. Sterile fitration was performed as indicated using milipak 20 sterile filters or durapore membranes (Millipore, Molsheim, France).

(76) Concentration Gradients

(77) The initial 30 mM liposome formulations in 30% (w/w) trehalose solution (PD-L-09111) was diluted with water to different final lipid and trehalose concentrations.

(78) TABLE-US-00007 TABLE 5 Dilution of tested samples PD-L-09111 Water C.sub.lipid C trehalose c trehalose Name (ml) (ml) (mM) (% w/w) (% w/w) PD-L-09111 100 0 30 30 0 PD-L-09112 70 70 15 15 15 PD-L-09113 90 54 18.8 18.8 11.2 PD-L-09114 100 0 30 30 0 PD-L-09115 80 64 16.7 16.7 13.3 PD-L-09116 100 40 21.4 21.4 8.6 PD-L-09119 100 24 25 25 5

(79) Stability Testing

(80) The samples were put on a shaker and agitated at 150 rpm at 25 C. or at 2-8 C., respectively. After 24 and 48 hours the samples were analyzed and the amount of paclitaxel retained in the liposomes was determined.

(81) Determination of Paclitaxel Retention/Release in the Liposomes Preparations

(82) Paclitaxel retention in liposomes was investigated by filtration of the liposome preparations in order to remove paclitaxel crystals from the liposome product (as described in Example 3). The remaining paclitaxel was quantified by HPLC analysis. Additionally, optical microscopy was used to investigate the samples for paclitaxel crystals.

(83) 5.3 Results

(84) Results are given in Tables 6-12. For simplicity, the initial concentration of trehalose is approximated as 30% (w/w). Concentration gradients as depicted are calculated from the initial trehalose concentration and the dilution factor. The actual concentration gradients between encapsulated and free aqueous phase will have the same trend, but the absolute values may differ slightly from those given in the tables.

(85) As can be seen, the stability increases with increasing trehalose gradient. The amount of liposomally retained paclitaxel increases and less paclitaxel crystals are observed. The stability is higher at 5 C. compared to 25 C.

(86) TABLE-US-00008 TABLE 6 Stability at 0% trehalose gradient PXL lost on Particle number Temp time filtration Crystals in >1 M >1 M >25 M ( C.) (h) (%) microscopy PD-L-09111 30 0 5 24 <5 No PD-L-09111 30 0 5 24 8 Yes PD-L-09111 30 0 5 48 8 Yes PD-L-09111 30 0 5 48 19 Yes PD-L-09111 30 0 25 24 86 Yes PD-L-09111 30 0 25 24 86 Yes PD-L-09111 30 0 25 48 85 Yes PD-L-09111 30 0 25 48 85 Yes

(87) TABLE-US-00009 TABLE 7 Stability at 5% trehalose gradient C.sub.tre C.sub.lipid (% Temp time PXL lost on Crystals in Sample (mM) w/w) ( C.) (h) filtration (%) microscopy PD-L-09119 25 5 5 24 <5 No PD-L-09119 25 5 5 24 <5 No PD-L-09119 25 5 5 48 <5 No PD-L-09119 25 5 5 48 <5 No PD-L-09119 25 5 25 24 29 Yes PD-L-09119 25 5 25 24 44 Yes PD-L-09119 25 5 25 48 70 Yes PD-L-09119 25 5 25 48 69 Yes

(88) TABLE-US-00010 TABLE 8 Stability at 8.6% trehalose gradient C.sub.tre C.sub.lipid (% temp time PXL lost on Crystals in Sample (mM) w/w) ( C.) (h) filtration (%) microscopy PD-L-09116 21.4 8.6 5 24 <5 No PD-L-09116 21.4 8.6 5 24 <5 No PD-L-09116 21.4 8.6 5 48 <5 No PD-L-09116 21.4 8.6 5 48 <5 No PD-L-09116 21.4 8.6 40 24 <5 No PD-L-09116 21.4 8.6 40 24 <5 No PD-L-09116 21.4 8.6 40 48 <5 Yes PD-L-09116 21.4 8.6 40 48 <5 Yes

(89) TABLE-US-00011 TABLE 9 Stability at 11.2% trehalose gradient C.sub.tre C.sub.lipid (% temp time PXL lost on Crystals in Sample (mM) w/w) ( C.) (h) filtration (%) microscopy PD-L-09113 18.8 11.2 5 24 <5 No PD-L-09113 18.8 11.2 5 24 <5 No PD-L-09113 18.8 11.2 5 48 <5 No PD-L-09113 18.8 11.2 5 48 <5 No PD-L-09113 18.8 11.2 40 24 <5 No PD-L-09113 18.8 11.2 40 24 <5 No PD-L-09113 18.8 11.2 40 48 <5 No PD-L-09113 18.8 11.2 40 48 <5 No

(90) TABLE-US-00012 TABLE 10 Stability at 13.3% trehalose gradient C.sub.tre C.sub.lipid (% temp time PXL lost on Crystals in Sample (mM) w/w) ( C.) (h) filtration (%) microscopy PD-L-09115 16.7 13.3 5 24 <5 No PD-L-09115 16.7 13.3 5 24 <5 No PD-L-09115 16.7 13.3 5 48 <5 No PD-L-09115 16.7 13.3 5 48 <5 No PD-L-09115 16.7 13.3 40 24 <5 No PD-L-09115 16.7 13.3 40 24 <5 No PD-L-09115 16.7 13.3 40 48 <5 No PD-L-09115 16.7 13.3 40 48 <5 No

(91) TABLE-US-00013 TABLE 11 Stability at 13.3% trehalose gradient C.sub.tre C.sub.lipid (% temp time PXL lost on Crystals in Sample (mM) w/w) ( C.) (h) filtration (%) microscopy PD-L-09115 16.7 13.3 5 24 <5 No PD-L-09115 16.7 13.3 5 24 <5 No PD-L-09115 16.7 13.3 5 48 <5 No PD-L-09115 16.7 13.3 5 48 <5 No PD-L-09115 16.7 13.3 40 24 <5 No PD-L-09115 16.7 13.3 40 24 <5 No PD-L-09115 16.7 13.3 40 48 <5 No PD-L-09115 16.7 13.3 40 48 <5 No

(92) TABLE-US-00014 TABLE 12 Stability at 15% trehalose gradient C.sub.tre C.sub.lipid (% temp time PXL lost on Crystals in Sample (mM) w/w) ( C.) (h) filtration (%) microscopy PD-L-09112 15 15 5 24 <5 No PD-L-09112 15 15 5 24 <5 No PD-L-09112 15 15 5 48 <5 No PD-L-09112 15 15 5 48 <5 No PD-L-09112 15 15 40 24 <5 No PD-L-09112 15 15 40 24 <5 No PD-L-09112 15 15 40 48 <5 No PD-L-09112 15 15 40 48 <5 No

(93) 6 Stability of Spray-Dried Paclitaxel-Loaded Liposomes

(94) 6.1 Methods and Materials

(95) Materials were used as described in the previous examples.

(96) Liposomes consisting of DOTAP, DOPC and paclitaxel (molar ratio 50/47/3) were formed by the ethanol injection techniques as described above. The paclitaxel was solubilised with the lipids in the ethanol solution. Liposomes at a concentration of 20 mM lipid in 20% w/w trehalose (batch PD-L-09031) and 10 mM lipid in 10% trehalose (batches MDG09.108-08-001 and PD-L-09032) were prepared. The liposomes were extruded five times across polycarbonate membranes of 200 nm pore size and sterile filtrated as described above.

(97) After preparation of the liposomal suspensions, the liposomes were dehydrated. Batches PD-L-09031 and PD-L-09031 were spray dried in a Niro SD-Micro spray dryer with spray drying parameters as described in Example 2.

(98) Batch MDG09.108-08-001 was dehydrated by freeze drying, using a Epsilon 2-12D (Christ) freeze drying unit. The liposomal suspension was kept at 4 C. for 1 hour and frozen at 40 C. for about 5 hours. After freezing, temperature was increased to 16 C. and primary drying was performed at a pressure of 0.1 bar for 90 hours. For secondary drying, the temperature was increased to 20 C., while pressure was reduced to 0.01 bar.

(99) The dry powders were reconstituted with water to a lipid concentration of 10 mM in 10.5% (w/w) trehalose. The resulting liquid liposome products were investigated one hour after reconstitution and 24 hours after reconstitution with dynamic lights scattering measurements using a using a Malvern Zetasizer 1000HSA, Series DTS5101 (Settings: Analysis=mono modal, Dilation=1.2; Order of fit=3; Point selection First=18; Last Point Selection=By Number 22; Point weighting=quatric; Attenuator=16; Viscosity 1.200 cp; Refractive Index=1,348; Number of Measurements=3; Delay Between Measurements=0; Measurement Duration=Auto) to determine Z.sub.ave and PI. Before measurement the samples were diluted ten-fold with 10.5% (w/w) trehalose dehydrate solution.

(100) 6.2 Results

(101) The results are shown in Table 13.

(102) The findings for the formulation which was spray dried at the same lipid and trehalose concentration as in the rehydrated product were substantial different from the results for the product which had been sprayed from double concentrated liquid feed and a trehalose gradient had been generated upon rehydration. The formulation without concentration gradient displayed significantly higher Z.sub.ave and PI values and increased within 24 hours after reconstitution, while the formulation with concentration gradient did not show such increase. The increase in Z.sub.ave and PI is considered to be related to paclitaxel release from the formulation without concentration gradient, which was less stable. The data are in accordance with the results of Example 2, where more paclitaxel could be loaded to liposomes which had been obtained after spray drying at double concentration and subsequent generation of a trehalose gradient. Spray drying of the paclitaxel liposome formulations at higher trehalose concentrations and subsequent generation of a trehalose gradient improves the stability of the formulation after reconstitution.

(103) In comparison to the spray dried samples, the freeze dried formulation showed a much higher PI already after reconstitution.

(104) TABLE-US-00015 TABLE 13 Comparison of dehydration and rehydration methods 1 h 24 h Batch c.sub.trehalose Z.sub.ave (nm) PI Z.sub.ave (nm) PI MDG09.108-08-001 0% 170.3 0.480 175.4 0.493 PD-L-09032 0% 167 0.331 260 0.65 PD-L-09031 10% 160.8 0.203 160.5 0.199

(105) 7 Large Scale Manufacturing and Spray Drying

(106) 7.1 Material

(107) 7.1.1 Basic Materials USP Semi-Synthetic Paclitaxel API, Phyton Biotech, Lot CP209N0014 DOTAP-CI, Merck Eprova AG, Lot MBA-020 DOPC, Avanti Polar Lipids Inc., Lot GN181 PC-12 ,-Trehalose Dihydrate High Purity (Low Endotoxine), Ferro Pfanstiehl, Lot 33205A Ethanol absolute EP, Nova Laboratories Art.-Nr. A4478B Citric acid monohydrate EP/USP, Nova Laboratories Art.-Nr. V290 Water for injection, Nova Laboratories Art.-Nr. A15210C

(108) 7.1.2 Equipment Injection capillary ID: 2 mm Filter cartridge Memtrex PC 0.2 pm from GE, Article No. MPC92O5FHV, Lot. 60240937 Sterile filter Opticap XL4 with 0.22 pm Duraporemembrane from Millipore, Article Nr. KVGLAO4TT3 Lot.: COCA1 0972 Formulation-Vessel (Nova Laboratories Ltd.) Extrusion-Vessel (Nova Laboratories Ltd.) Bioburden-Reduction-Vessel (Nova Laboratories Ltd.) Holding-Vessel (Nova Laboratories Ltd.) Peristaltic pump (Nova Laboratories Ltd.) 20 L-pressure vessel with standpipe (Nova Laboratories Ltd.) Butterfly-vents (Nova Laboratories Ltd.) ASD-1 aseptic spraydryer (GEA Niro S/A, Copenhagen Denmark)

(109) 7.2 Methods

(110) 7.2.1 Production of Liquid Formulation

(111) 7.2.1.1 Preparation of Organic Solution

(112) For batch 001, 349.3 g DOTAP-CL were dissolved in solved in 700 g absolute ethanol and stirred for approximately 4 h. 369.5 g DOPC were dissolved in 700 g absolute ethanol and stirred for approximately 3 h. Subsequently the two lipid solutions were joined and added to 25.617 g paclitaxel. The resulting organic solution was stirred for about 2 hours and finally adjusted to a total weight of 2122.5 g by the addition of absolute ethanol. Batches 002 to 004 were prepares accordingly

(113) 7.2.1.2 Preparation of Aqueous Solution

(114) For batch 001, 10819 g of trehalose dehydrate were added to about 20 kg of water for injection in a formulation vessel and stirred at 700 rpm for 90 min. Subsequently 1.258 g citric acid monohydrate were added and stirred until complete dissolution. The final volume of the aqueous solution was adjusted to 34.53 kg and stirred for another 10 min. Batches 002 to 004 were prepared accordingly

(115) 7.2.1.3 Ethanol Injection

(116) For batch 001, the organic solution was injected into the aqueous solution by means of a peristaltic pump with an injection rate of about 250 g/min. During injection, the solution was stirred at about 500 rpm. After the injection was finished the solution was stirred for 2 min at 600 rpm and subsequently for 1 min at 700 rpm. During the whole injection process the temperature was kept below 8 C. Batches 002 to 004 were stirred at 550 rpm during injection with no additional stirring thereafter.

(117) 7.2.1.4 Extrusion

(118) 8 extrusions over a 5 filter cartridge with a 0.2 pm polycarbonatemembrane were performed. The filter cartridge was ventilated with 0.4-0.5 bar each, and extrusion was performed with a pressure of 3.0 bar. Temperature was kept below 8 C.

(119) 7.2.1.5 Dilution

(120) After the 8th extrusion the weight of the formulation was determined. Based on the formulation's density of 1.106 g/ml the amount of water was calculated which was required to obtain a 20 mM formulation (based on total lipid concentration), which corresponds to a 1:1.5 dilution. For batch 001, the required amount of water for injection was added at 1.66 l/min by means of a peristaltic pump through a 2 mm ID capillary while the solution was stirred at about 500 rpm. The added water for injection had been chilled to below 8 C. before adding to the formulation. For batches 002 to 004 dilution was performed at 0.62 l/min to 0.83 l/min at a stirring speed of about 600 rpm.

(121) 7.2.1.6 Reduction of Bioburden

(122) Before the reduction of bioburden (1'st sterile filtration), the OpticapXL4 filter was washed with 20 L water for injection at a pressure of about 0.5 bar. Filling and ventilation of the filter was performed gravimetrically. The filtration was performed at a pressure of 2.5 bar, whereby the pressure was applied promptly. Temperature of the formulation was kept below 8 C.

(123) 7.2.1.7 Sterile Filtration

(124) Before the sterile filtration, the OpticapXL4 filter was washed with 20 L water for injection at a pressure of about 0.5 bar. Ventilation of the filter was performed at 0.5 bar. The filtration was performed at a pressure of 2.5 bar, whereby the pressure was applied promptly. Temperature of the formulation was kept below 8 C.

(125) 7.2.2 Spray-Drying

(126) The liposomal formulation was spray dried in a ASD-1 aseptic spray-dryer (GEA Niro S/A, Copenhagen Denmark). Batch 001 was dried as single batch (Run 1), whereas batches 002 to 004 were sprayed sequentially in a continuous fashion (Run 2). For spray-drying a two-fluid nozzle, nitrogen as drying gas, and the following parameters were used:

(127) TABLE-US-00016 TABLE 14 spray drying settings Parameter Set point Drying gas rate 80 kg/h Atomizer-gas pressure 3 bar Atomizer-gas rate 3 kg/h Outlet temperature 95 C. Feed rate 2 L/h Feed temperature 0 C.-30 C.

(128) 7.3. Stability of Liposomes

(129) 7.3.1. Method

(130) Dehydrated liposomal compositions from Run1 were rehydrated in water for injection to a total lipid concentration of 10 mM, thus the reconstitution conditions further increase the osmolar gradient of the preparation which had been 20 mM before dehydration. For comparison, a corresponding liposomal composition (Reference Batch) which was prepared without dilution and had been dehydrated by lyophilisation (as for example disclosed in WO 2004/002468). The amount of paclitaxel (including paclitaxel degradation products) retained in the liposomes was determined according to the method described in Example 1.3 after reconstitution of the liposomes and after 24 hours at 25 C. The percentage of paclitaxel retained and filterable (crystallised) paclitaxel was calculated based on the total paclitaxel present in the preparations.

(131) 7.3.2. Results

(132) TABLE-US-00017 TABLE 15 Release of paclitaxel from products T0 After 24 h at 25 C. Liposomally Filterable Liposomally Filterable Preparation retained paclitaxel retained paclitaxel Time paclitaxel [%] [%] paclitaxel [%] [%] Run1 99.56 0.44 99.89 0.11 99.93 0.07 100.09 0.09 99.67 0.33 99.87 0.13 Reference 99.63 0.37 98.88 1.12 Batch 99.48 0.52 98.68 1.32 99.72 0.28 99.16 0.84

(133) The data show that liposomal preparations prepared in the absence of an osmolar gradient release paclitaxel faster. This can already be observed after a relatively short time span of 24 h.

(134) 7.4. Analysis of Particle Size and Polydispersity

(135) 7.4.1. Methods

(136) Particle size (zaverage) and polydispersity index (PI) were determined by PCS (173 diffraction) using a Zetasizer Nano ZS (Malvern Instruments). In brief, sampled from Run1 (corresponding to Batch 1) and Run2 (corresponding to Runs 2-4) and a sample form Reference Batch (see above) were resuspended in water to a total lipid concentration of 10 mM. For measurement the samples were diluted ten fold with 10.5% trehalose dehydrate solution (w/w). The samples were stored for 24 hours and 25 C. and measured again.

(137) The following settings were used for measurement and data analysis: Measurement type=Size; Sample: Material=Polystyrene latex, RI: 1.590; Absorption: 0.01; Dispersant=10.5% Trehalose, Temperature: 25 C., Viscosity 1.200 cP RI: 1.342; General options=Mark-Houwink parameters; Temperature=25 C., Equilibration time: 5 minutes; Cell=DTS0012-Disposable sizing cuvette; Measurement: Number of runs=15; Run duration (seconds)=100; Number of measurements=3; Delay between measurements; Advanced=Fixed Position at Position 4.65, Attenuator 6; Analysis Parameters: Analysis mode=General; Cumulants analysis: Order of fit=3; Weighting scheme=Quadratic; Cumulants point selection: Automatic first point=Yes; Last point selection method=Cut-off; Fraction of signal=0.1; Dilation=1.2; Display range: Lower limit=0.6; Upper limit=10000; Filtering: Filter factor=75; Multimodal-analysis: Result transformation=Mie; Use result transformation=Yes; Weighting scheme=Quadratic; Resolution=normal; Multimodalpoints selection; Automatic first point=Yes; Last point selection method=Cut-off Fraction of signal=0.01; Dilation=1.2; Size classes: Number of size classes=70; Lower limit=0.4; Upper limit=10000; Thresholds: Lower Threshold=0.05; Upper Threshold=0.01; Filter factor=default.

(138) 7.4.2. Results

(139) The results are shown in Table 16:

(140) TABLE-US-00018 Sample zaverage (nm) PI Reference Batch 133 0.337 Run 1 Batch 1 143 0.16 Run 2 Batch 2 138 0.15 Batch 3 143 0.17 Batch 4 144 0.19

(141) The formulations manufactured with an osmotic gradient and dehydrated by spray-drying (Runs 1 & 2) displayed a very similar zaverage but significantly lower PI values compared to the reference sample manufactured without an osmotic gradient and dehydrated by lyophilisation. Thus the product produced by the process described above is more homogeneous than a product produced by a conventional process.

(142) 7.5 Ultracentrifugation Characterisation

(143) 7.5.1. Method

(144) Analysis of ultracentrifugation was performed by Nanolytics (Potsdam, Germany).

(145) Each sample was reconstituted in H.sub.2O and a 1:1 mixture of H.sub.2O: D.sub.2O to a lipid concentration of 10 mM and equilibrated for one hour at room temperature. Subsequently the samples were diluted 1:1 with the respective solvent. After another hour of equilibration the samples were subjected to ultracentrifugation. 400 l of the respective liposomal dispersion were subjected to a titanium ultracentrifugation cuvette with an optical path of 12 mm. The samples were centrifuged in a Optima XL-I analytical ultracentrifuge (Beckmann-Coulter, Palo Alto) using an An50Ti 8-place rotor (Beckmann-Coulter, Palo Alto) equipped with Rayleigh interference optics at 20000 rpm and 25 C. During the centrifugation, the concentration profile along the radial coordinate was by means of the refractivity gradient within the solution. Samples were measured as duplicates.

(146) 7.5.2. Data Analysis

(147) 7.5.2.1 Definition of the Sedimentation Coefficient

(148) The primary indicator in analytical ultracentrifugation is the sedimentation coefficient defined as follows:

(149) $\begin{array}{cc}\frac{m\left(1-\stackrel{\_}{}\right)}{f}=\frac{u}{{}^{2}r}s& \left(1\right)\end{array}$

(150) Wherein u is the sedimentation velocity of the particle, m the mass of the particle, is the specific volume, f is the friction term and is the density of the solvent.

(151) Determination of the Sedimentation Coefficient

(152) The sedimentation coefficient is calculated directly from the measured data without further assumptions according to:

(153) $\begin{array}{cc}\ln \frac{r}{r_m}=s{}^{2}dt& \left(2\right)\end{array}$

(154) Wherein r is the distance to the rotation axis, and r.sub.m is the meniscus. The run time integral .sup.2dt is determined by the measuring equipment.

(155) 7.5.2.3 Sedimentation Coefficient Distribution

(156) Instead of a single sedimentation coefficient at a specific radius, the whole r-axis can be transformed into an s-axis. The fringe shift at the respective position is proportional to the mass concentration of the particle species present there, so that the measures amplitude can be taken as y-coordinate. However it is required to correct the y-coordinate with regard to the radial dilution. By including the correction term the following function g(s) is obtained, giving the mass concentration of the particle species sedimenting with velocity s:

(157) $\begin{array}{cc}g\left(s\right)=\frac{1}{c_0}\frac{dc}{dt}{\left(\frac{r}{r_m}\right)}^{2}r{}^{2}dt& \left(3\right)\end{array}$

(158) The concentration c, respectively c0 is given in units fringe shift, wherein a fringe equals a full phase, thus a light and a dark line of the interference pattern. The size is direct proportional to the concentration given in g/l in case the refractive index increment can be assumed to be equal for all particle species, which is the case for the present samples which are chemical uniform material which is simply present in a polydisperse distribution:

(159) $\begin{array}{cc}dc=d.Math..Math.\frac{1}{l\frac{dn}{dc}}& \left(4\right)\end{array}$

(160) Wherein is the concentration in fringe shift units, is the wavelength of the laser, l the width of the cuvette, and dn/dc is the refractory index increment of the solute in a given solvent. Due to this proportionality the concentration in equation (3) can be stated directly as mass concentration.

(161) g(s) (or its integrated form G(s) can in principle be calculated scan by scan coordinate transformation and calculation of the y-coordinate according to equation (3); the overall result would be obtained by averaging the mainly scans, which are mainly redundant. Thus it is reasonable to perform a global data fitting over all scans. Thereby time and space independent noise is isolated and further statistic noise is partitioned to obtain the best g(s) is obtained from the entire measurement data. SedFit v12.4. software from Peter Schuck was used for the fitting.

(162) 7.5.2.4 Interpretation of Sedimentation Coefficient Distribution

(163) The sedimentation coefficient distribution in the form of the function g(s) already gives information on the from of the distribution; usually it is desirable to transform the primary measurement parameter s into the diameter or mass of the particle. The sedimentation coefficient is related to the molar mass by the SVEDBERG Equation

(164) $\begin{array}{cc}M=\frac{\mathrm{sRT}}{D\left(1-\stackrel{\_}{}\right)}& \left(5\right)\end{array}$
via the diffusion coefficient. For globular objects, as for the present liposomes, the diffusion coefficient can be replaced by the diameter of a sphere, whereby it has to be considered, that the liposome is filled with water, which does not contribute to the sedimentation but to the friction.

(165) If the diffusion coefficient in equation 5 is replaced by the STOKES-EINSTEIN-Equation
D=6R.sub.h(6)
and considers the diameter d=2Rh for a sphere and a volume fraction of water ; the SVEDBERG Equation becomes

(166) $\begin{array}{cc}d=\sqrt{\frac{18s}{\left(1-\right)\left({}_s-\right)}}& \left(7\right)\end{array}$
wherein is the viscosity of the solvent and .sub.s is the density of the solute.

(167) For the conversion of the sedimentation coefficient distribution into a size distribution it is required that the swelling and density of the liposomes remain constant or are given as a distribution dependent on the sedimentation coefficient. Thus the density is determined experimentally.

(168) 7.5.2.5 Analysis of the Density Variation

(169) The density of sedimenting particles can be determined by analytical ultracentrifugation in two solvents with differing density. In the solvents having a lower density, a particle normally exhibits a smaller s-valuethe particle sediments slower, if the density difference to the surrounding solvent decreases. Both s-values, which describe the same particle, fulfil equation (7) with the parameters for the respective solvent. For the two solvents (index 1 and 2) the following applies:

(170) $\begin{array}{cc}\frac{{}_1{s}_1}{\left[{}_1+\left(1-\right){}_s\right]-{}_1}=\frac{{}_2{s}_2}{\left[{}_2+\left(1-\right){}_s\right]-{}_2}& \left(8\right)\end{array}$

(171) The anhydrous density of the particle .sub.s and the swelling parameter on both sides of the equation are identical, since they refer to the same object. Identical particles are defined as elements of the sedimentation coefficient distribution with identical y-coordinate G(s).

(172) Thus equation (8) can be rearranged and simplified to obtain the anhydrous density:

(173) 0$\begin{array}{cc}{}_s=\frac{{}_1{s}_1{}_2-{}_2{s}_2{}_1}{s_1{}_1-s_2{}_2}& \left(9\right)\end{array}$

(174) The diameter can be derived according to the following equation:

(175) $\begin{array}{cc}d=\sqrt{\frac{18\left({}_1{s}_1-{}_2{s}_2\right)}{\left(1-\right).Math.\left({}_1-{}_2\right)}}& \left(10\right)\end{array}$

(176) To solve equation (10), independent information such as the diameter of the particles is required, which can be determined experimentally by dynamic light scattering experiments as described above.

(177) 7.5.3. Results

(178) TABLE-US-00019 TABLE 17 Anhydrous density of liposomal preparation Anhydrous density Manufacturing Run Sample (g/ml) Run 1 Sample 1 1.183 Run 2 Sample 1 1.174 Sample 2 1.154 Sample 3 1.223

REFERENCES

(179) Antonietti, M. and S. Forster (2003). Vesicles and liposomes: A self-assembly principle beyond lipids. Advanced Materials 15(16): 1323-1333. Bangham A. D., Standish M. M., Watkins, J. C. (1965). Diffusion of univalent ions across the lamellae of swollen phospholipids, J. Mol. Biol. 13: 238-52. Cabral, E. C. et al. (2003). Preparation and characterization of diacetylene polymerized liposomes for detection of autoantibodies. J Liposome Res 13(3-4): 199-211. De Gier, J. (1993). Osmotic behaviour and permeability properties of liposomes. Chem Phys Lipids 64(1-3): 187-196. Ertel, A., A. G. Marangoni, et al. (1993). Mechanical properties of vesicles. I. Coordinated analysis of osmotic swelling and lysis. Biophysical Journal 64(2): 426-434. Evans, D. F. and Wennerstrom H. (1994). The Colloidal Domain, VHC Publishers, Inc., New York, pp 48-49. Goormaghtigh, E. and G. A. Scarborough (1986). Density-based separation of liposomes by glycerol gradient centrifugation. Anal Biochem 159(1): 122-31. Gregoriadis G. (1995). Engineering liposomes for drug delivery: Progress and problems. Trends in biotechnology 13 (12): 527-537. Hallett, F. R., J. Marsh, et al. (1993). Mechanical properties of vesicles. II. A model for osmotic swelling and lysis. Biophysical Journal 64(2): 435-442. Huang, C. H., Charlton, J. P. (1971) Determination of partial specific volumes by sedimentation velocity method, The Journal of Biological Chemistry, 246(8): 2555-2560. Koppel, D. E. (1972). Analysis of macromolecular polydispersity in intensity correlation spectroscopy: the method of cumulants. Journal of Chemical Physics 57(11): 4814-4820. New et al. (1990). Liposomes. A Practical Approach. Oxford University Press. Pages 33-104).