- Email: [email protected]
Formulation of liposomes associated with recombinant interleukin-2: effect on interleukin-2 activity
Biomedicine & Pharmacotherapy 58 (2004) 162–167 www.elsevier.com/locate/biopha
Dossier: Drug delivery and drug efficacy
Formulation of liposomes associated with recombinant interleukin-2: effect on interleukin-2 activity Yann Pellequer, Michel Ollivon, Gillian Barratt * UMR CNRS 8612, Université Paris-Sud, 5, rue Jean-Baptiste Clément, 92296 Châtenay-Malabry, France Received 9 October 2003; received in revised form 3 December 2003; accepted 4 December 2003
Abstract Association of the cytokine interleukin-2 (rIL-2) within liposomes could prolong its circulating half-life and thus reduce side-effects and improve its efficacy in cancer and AIDS treatment. The effects of physical procedures used in liposome preparation on the biological activity of rIL-2 were determined. While heating to 50 °C reduced the activity of IL-2 in the CTLL-2 proliferation assay by 50%, sonication, either bath or probe, was less detrimental. The combination of all three treatments resulted in only 10% loss of activity. Probe sonication led to the appearance of dimers which were stable under reducing conditions. Small unilamellar and large unilamellar liposomes were formed, respectively, by probe sonication or extrusion of multilamellar vesicles of dipalmitoylphosphatidylcholine hydrated in the presence of rIL-2. A high proportion of the rIL-2 was associated with the vesicles. However, the biological activity of the liposome-associated rIL-2 was reduced 7- to 10-fold compared with control rIL-2. rIL-2 dimers were formed on contact with lipid, even without sonication. We can conclude that the association of rIL-2 with lipid masks its access to its cell-surface receptor at least under cell culture conditions. © 2004 Elsevier SAS. All rights reserved. Keywords: Interleukin-2; Liposomes; CTLL-2 proliferation
1. Introduction The production of recombinant interleukin-2 (rIL-2) has cleared the way for large-scale clinical applications, notably in oncology and AIDS treatment because of this cytokine’s central role in the regulation of the immune response. However, rIL-2 is a 15-kDa protein which is rapidly eliminated by renal filtration and serious side effects, including capillary leak syndrome, are associated with the systemic administration of the high doses of rIL-2 necessary to attain effective concentrations. Different systems such as minipellets [17], supramolecular biovectors [5], microspheres [16] and liposomes [2,3,11–13], have been tested for their ability to increase efficiency and reduce side effects by increasing the biological half-life of rIL-2.
Abbreviations: DPPC, dipalmitoylphosphatidylcholine; LUV, large unilamellar vesicles; MLV, multilamellar vesicles; OG, N-octyl b,Dglucopyranoside; RIL-2, recombinant human interleukin-2; SUV, small unilamellar vesicles. * Corresponding author. E-mail address: [email protected] (G. Barratt). © 2004 Elsevier SAS. All rights reserved. doi:10.1016/j.biopha.2003.12.008
Liposomes are interesting systems for drug delivery because of their capacity to include hydrophobic molecules within their bilayer membranes. This would be advantageous for rIL-2 association because the non-glycosylated protein is very poorly water-soluble. Liposomes have been employed as a delivery system for rIL-2 in systemic [2,3,11] as well as in locoregional immunotherapy [12,13]. High association of rIL-2 with small unilamellar liposomes prepared from dipalmitoylphosphatidylcholine (DPPC) has already been observed [8]. This synthetic, saturated lipid would be expected to yield liposomes which are more stable in storage (resistant to oxidation) and in biological medium (resistant to destabilisation by plasma proteins and lipoproteins) than those prepared from natural phospholipids such as egg phosphatidylcholine. This would make them more suitable as controlled release systems; however, liposome formation must be performed at a temperature above the phase transition temperature of 41 °C, which could be detrimental to proteins, causing denaturation and loss of activity. The procedures used to reduce liposome size, essential to ensure their prolonged presence in the circulation, might also affect proteins. The aim of this work was to test the effect of formulation conditions on the physical integrity and biological activity of rIL-2.
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
2. Materials and methods 2.1. Materials DPPC was purchased from Avanti Polar Lipids (Alabaster, AL, USA). N-octyl b,D-glucopyranoside (OG) was obtained from Sigma Chemicals Co, (St. Louis, MI, USA). Recombinant IL-2 (rIL-2) with a specific activity of 1.1 × 107 U/mg protein (RU 49637) produced in Escherichia coli was provided by Roussel-Uclaf (France). In this work, rIL-2 was purified to remove dextran used as a lyophilisation stabiliser as described in Pellequer et al. [19]. 2.2. Stability of rIL-2 under conditions of liposome preparation rIL-2 at 250 µg/ml in citric acid solution (9 g/l, pH 3) was maintained either in a water bath at 50 °C (1 or 15 min) or 1 min in a bath sonicator (at 20 or 50 °C) (Branson 2200, Branson Ultrasonics Corporation, Danbury, USA) or subjected to a probe sonication using a Vibracel 600 W ultrasound generator (Bioblock Scientific) at about 90 W power and 50% cycle time for five sonication cycles of 2 min, each separated by 1 min of rest (at 38 °C or after 1 min in a bath sonicator at 50 °C or without heating or bath sonication). Titanium particles generated by the probe sonication were removed from the liposome suspension by centrifugation at 4500 rpm, for 5 min at 10 °C (rotor No12158, angle rotor 6 × 30 ml, maximum radius 7.9 cm, min. radius 2.3 cm, angle 30°) in a Nalgene® tube (Nalge Company, Rochester, NY, USA). rIL-2 integrity and biological activity were determined by SDS-PAGE and bioassay, respectively, as described below. 2.3. Liposome preparation Liposome preparation is described in Pellequer et al. [19]. Briefly, DPPC (10 mg) was dissolved in chloroform (about 0.5 ml) and deposited as a film on a flat-bottomed flask by evaporation under nitrogen followed by overnight vacuum drying. This film was hydrated with 3 ml rIL-2 solution (250 µg/ml, pH 3). The multilamellar vesicles (MLV) formed were homogenised by bath sonication (1 min at 50 °C). Size reduction of liposomes was carried out by two different procedures: sonication or extrusion, as follows. Small unilamellar vesicles (SUV) were prepared by probe sonication at 38 °C at about 90 W power and 50% cycle time. Five sonication cycles of 2 min were each separated by 1 min of rest. Titanium particles generated by sonication were removed from the liposome suspension by centrifugation at 4500 rpm, for 5 min at 10 °C as described above. Large unilamellar vesicles (LUV) were prepared at 50 °C by successive extrusions of the MLV suspension under nitrogen pressure (maximum pressure = 3000 kPa). The latter was passed twice through each of successively 0.8, 0.4, 0.2, 0.1, 0.005 µm polycarbonate filters (Poretics Corporation, Livermore, USA).
163
The mean particle diameter was estimated by quasi-elastic light scattering (Nanosizer N4MD, Coulter Electronics Inc., Florida, USA) using the “Unimodal” method of date processing. The samples were analyzed at 25 °C; the viscosity and refractive index of the aqueous suspending medium were taken to be 0.890 mPa s and 1.3325, respectively. The mean diameters of liposomes obtained by sonication and extrusion were 52 and 108 nm (the standard deviations of one normally distributed dispersion were 10 and 14 nm), respectively. 2.4. Quantification of rIL-2 associated with liposomes rIL-2 was assayed by a method base on the intrinsic fluorescence of its single tryptophan residue as described by Pellequer et al. [19]. Briefly, liposomes were separated from free rIL-2 by exclusion chromatography on Sephadex® G75 Superfine swollen in citric acid solution (9 g/l, pH 3). The elution of rIL-2 liposomes and free rIL-2 was monitored by measuring the fluorescence with a four-channel fluorometer Spex F1T11I (equipped with Peltier-cooled photomultipliers and a 450 W xenon lamp) controlled by DM 3000 software (Spex industries Inc, Edison, USA). Fluorescence emission was measured at 90°. Samples were maintained at 25 °C. rIL-2 elution was monitored at 320 nm (tryptophan fluorescence emission) after excitation at 290 nm, while liposomes were detected at the same time by light scattering at 290 nm. Two hundred microlitres of liposomes (1:1 dilution) were disrupted and solubilised by 2.25 mg of OG. Samples were analysed by fluorescence (kexc = 290 nm; kem = 320 nm). The fluorescence intensity of a micellar solution (1.2 mg DPPC, 58 mg OG, 5 ml water) was used as a blank. Each result is the mean of three measurements. A calibration curve was obtained by dilution of rIL-2 in the micellar solution of DPPC/OG/water. 2.5. rIL-2 activity assay by CTLL-2 proliferation 2.5.1. CTLL-2 cell line The rIL-2 dependent mouse T cell line (CTLL-2 ECACC No 93042610) was cultured in 500 ml RPMI 1640 medium with 100 IU/ml penicillin, 100 µg/ml streptomycin, 5 mM sodium pyruvate, 10% foetal calf serum, 0.5 µl/ml b-mercaptoethanol, and 0.01 ml/ml non-essential amino acids at 37 °C in 5% CO2. The cell line was maintained with 10 IU/ml rIL-2. Dilutions of rIL-2 from the concentrated stock solution were made in the CTLL-2 culture medium containing serum. 2.5.2. MTT assay Cells were harvested by centrifugation, washed twice in phosphate buffered saline (PBS) and adjusted to 105 cells/ml in the CTLL-2 medium without rIL-2. rIL-2 or liposomes containing rIL-2 (200 UI/ml) was diluted (eight twofold dilutions) in flat bottomed 96-well microtiter plates (TPP, Trasadingen, Switzerland). Each dilution was made in six replicates. The CTLL-2 cells were added (100 µl) to the rIL-2
164
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
dilutions. MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) (Sigma, St. Louis, MO, USA) was diluted to 5 mg/ml in sterile PBS and filtered once through a 0.22 µm filter. After 48 h, the diluted MTT was added to the wells as 20 µl/well. Four hours after at 37 °C, the plates were centrifuged and the supernatants were eliminated. Two hundred microlitres of the extraction solution (0.1 g/ml sodium dodecyl sulphate, 0.5 ml/ml dimethyl formamide) were added to each well. The microtiter plates were incubated to overnight to dissolve reduced MTT crystals. Absorbance values were measured on a microtiter plate reader at 570 nm with a reference wavelength of 620 nm. The dose–response curves were plotted as mean absorbance against rIL-2 concentration measured in the fluorescence assay. The concentration giving 50% of the maximum response was calculated and expressed as a fraction of that of standard rIL-2 assayed in the same experiment. The ED50 of the standard rIL-2 varied depending on the passage number of the cell line. Each IL-2 preparation was assayed only once, but since they were always compared with standard rIL-2 assayed in identical conditions, we consider that the relative activities are accurate. 2.6. rIL-2 SDS-PAGE rIL-2 or liposomes samples were boiled in Laemmli sample buffer [15] and run on 12% SDS-PAGE under reducing (10% b-mercaptoethanol) conditions. Proteins in the gel were stained with silver nitrate according to the method described in Ref. [21]. Bands were compared with standard molecular weight markers (Gibco, Cergy-Pontoise, France).
3. Results and discussion 3.1. Activity of rIL-2 after ultrasound and/or temperature treatment 3.1.1. Cell proliferation assay Before studying the activity of rIL-2 associated with liposomes, it was necessary to determine the effect of the physical treatments used in the preparation of liposomes on the integrity of the protein. Since it is necessary to prepare liposomes above the phase transition temperature of the lipids used (in this case 41 °C for DPPC), the effect of heating to 50 °C for 1 and 15 min under the same conditions as those described in Section 2.3 was tested. Table 1 shows the ratio of rIL-2 activity in a cell proliferation test, based on the ED50 values for rIL-2 treated in various ways compared with control rIL-2. The ability of rIL-2 to maintain the growth of CTLL-2 cells was reduced by 50% by heating, regardless of the duration. Ultrasound treatment is used both to help form MLV and for size reduction to SUV. We observed that neither bath nor probe sonication had a dramatic effect on the rIL-2 molecule, since the activity was, respectively, 75% and 80% of the
Table 1 Effect of physical treatments used in liposome preparation on the biological activity of IL-2 Physical treatments T = heating T = heating T = 20 °C T = heating T = 20 °C T = heating T = heating
t = 1′ t = 15′ Bath sonicator t = 1′ Bath sonicator t = 1′ Probe sonicator t = 15′ Probe sonicator t = 15′ Bath and probe sonicator
IL-2 treated activity/IL-2 standard activity 0.5 0.5 0.75 0.8 0.8 0.8 0.9
rIL-2 (250 µg/ml) underwent various treatments as described in Section 2. Biological activity was assessed by CTLL-2 proliferation. The results are expressed as a ratio of the concentrations of treated and untreated rIL-2 which provoked half-maximal cell growth.
control. Surprisingly, this activity was not further reduced when sonication was carried out at 50 °C. rIL-2 which had undergone a combination of all physical treatments used to make SUV, bath sonication followed by probe sonication, both at elevated temperature, showed activity which was almost as high as that of control rIL-2. Thus, against all expectations, the effects of heating and of ultrasound energy were not additive, but seemed to counteract each other. 3.1.2. SDS-PAGE analysis In an attempt to explain these results the molecular state of the rIL-2 was analysed by SDS-PAGE. The ability of nonglycosylated rIL-2 to form homodimers has already been described in the literature [1,4,6,9,10,18]. In our previous work [19], we observed higher molecular weight bands in rIL-2 preparations by both gel exclusion chromatography and SDS-PAGE. Control rIL-2 showed a minor band at 25 kDa (Fig. 1). This was not altered by elevated temperature or bath sonication. On the other hand, treatments which included probe sonication showed increased intensity of this band. Since each lane was loaded with the same amount of protein as determined by the fluorescence assay, this increase is not due to selective destruction of the monomer but to an association brought about by the treatment. It should be noted that these gels were silver-stained and not subjected to Western blotting, so we have no formal proof that the higher molecular weight band is in fact a dimer of rIL-2. However, the fact that the molecular weight was twice that of the main band strongly suggests that this is the case. The gels were run in reducing conditions, which means that this dimerisation cannot be attributed to intermolecular disulphide bridges. Kaplan et al. [9] also observed multimers of rIL-2 in reducing conditions and suggested that they might be formed by transglutamination. They also proposed that self-associated rIL-2 had increased in-vitro activity based on the observation that the radioiodinated cytokine, which did not associate in this way, was less effective in a lymphocyte proliferation assay [9]. We did not observe a direct correlation between rIL-2 activity in the CTLL-2 (Table 1) and the appearance of
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
165
Fig. 1. SDS-PAGE of rIL-2 after various physical treatments. rIL-2 (250 µg/ml) underwent various treatments as described in Section 2. Samples 7.5 µg were loaded onto a 12% gel and run under reducing conditions. Proteins were revealed by silver staining. (1) Molecular weight markers; (2) control rIL-2; (3) bath sonication, 1 min without heating; (4) heating to 50 °C for 1 min; (5) heating for 15 min; (6) bath sonication with heating; (7) probe sonication with heating; (8) bath and probe sonication with heating; (9) probe sonication without heating.
dimers (Fig. 1) after the different treatments of rIL-2. However, our bioassay differs from that used by Kaplan et al. [9] in that they used a human T cell clone whereas we used the mouse CTLL-2 which undergoes apoptosis in the absence of rIL-2. Thus, any “sustained release” effect of self-associated rIL-2 as observed by Kaplan et al. [10] would be masked by the fact that if the cells are not exposed to the cytokine at the beginning of the incubation period they will become committed to apoptosis and will not be able to be stimulated for proliferation at later times. Another difference between our work and that of Kaplan et al. [9] is the exact sequence of the rIL-2 used. Our protein, from Roussel-Uclaf, is the natural IL-2 sequence, while that used by Kaplan was from the Chiron Corporation with two amino acid differences. Small differences in primary structure can affect the hydrophobicity of the protein and its self-aggregation behaviour [4]. In particular, the Chiron rIL-2 was found to aggregate in solution at neutral pH, leading to a change in the intrinsic fluorescence [6]. We did not observe any non-linearity of our fluorescence calibration curve; however, the protein was always maintained at pH 3, at which it shows maximal solubility in aqueous media. This is in contrast to the results of Advant et al. [1] who observed more extensive aggregation of the Chiron protein at pH 3.5 than at pH 6.5 and 8.2 by analytical ultracentrifugation. We can speculate that heating causes a denaturation of the rIL-2 molecule without self-association which reduces its ability to bind to its receptors and thereby stimulate CTLL-2 cells. On the other hand, ultrasound energy seems to perturb the molecule in a way which is conducive to forming dimers. This may protect rIL-2 from thermal denaturation. In the CTLL-2 assay, self-associated rIL-2 may provide sustained release over the 48-h incubation period and compensate for the loss of activity of monomeric rIL-2 on heating. 3.2. Activity of rIL-2 associated with liposomes rIL-2 was incorporated into liposomes prepared from DPPC. Although the results presented above suggest that sonication does not inactivate the cytokine to a great extent, extrusion was also used as a size-reduction technique because this procedure is more easily standardised and appli-
cable to large-scale production. The biological activity of liposome-associated rIL-2 was measured before and after separation of non-associated rIL-2 by gel chromatography. The concentrations of rIL-2 were measured by the fluorescence assay. 3.2.1. Cell proliferation assay Fig. 2 shows the dose–response curves in the CTLL-2. Liposome-associated rIL-2 was much less effective in maintaining cell growth than the free cytokine. Determination of the ED50 values revealed that for the non-separated samples, rIL-2 associated with sonicated and extruded liposomes was, respectively, 7 and 9.5 times less active than control rIL-2. When free rIL-2 was removed by gel filtration, the ratios were 7 and 11 times. Control experiments (data not shown) indicated that liposomes were not toxic for CTLL-2 cells. The results presented above show that loss of rIL-2 activity during the preparation of the liposomes cannot explain the loss of activity. The association of rIL-2 with the liposomes was very high. Gel exclusion chromatography revealed only a small peak corresponding to free rIL-2 for extruded liposomes and no detectable second peak for sonicated ones [19]. This is consistent with the observation that the rIL-2 activity did not change on separation of sonicated liposomes while it dropped after extruded liposomes were separated, suggesting that the free rIL-2 was more active than the associated. High association of recombinant rIL-2 with liposomes has already been observed by several authors [7,8,14] and can be attributed to interaction of the hydrophobic non-glycosylated protein with the liposome bilayer. This interaction could prevent recognition of rIL-2 molecules by the cell-surface receptor. The fact that association with liposomes protects rIL-2 against enzymatic degradation [8,20] suggests that the protein is at least partially embedded in the phospholipid bilayer. The reduction of rIL-2 activity on liposome association was less marked with the sonicated liposomes than with the extruded ones. This is consistent with the observation than probe sonication counteracted the effect of temperature on rIL-2, since extruded liposomes are subjected to bath sonication and heating, but not to probe sonication, therefore, a smaller proportion of the rIL-2 in the sonicated liposomes
166
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
0,8 Standard IL2 for sonication 0,7
Sonicated liposomes (x 0.14) Sonicated liposomes without free IL-2 (x 0.14)
0,6
Standard IL2 for extrusion
A (570-620 nm)
Extruded liposomes (x 0.10) 0,5
Extruded liposomes without free IL-2 (x 0.09)
0,4
0,3
0,2
0,1
0 0,1
1
10 IL-2/well (IU/ml)
100
1000
Fig. 2. Biological activity of liposome-associated rIL-2. Liposomes were prepared and separated from non-associated rIL-2 as described in Section 2. Protein concentrations were measured by fluorescence. Serial twofold dilutions of liposomes and standard rIL-2 were incubated with CTLL-2 for 48 h, after which the MTT conversion test was performed. Values are the mean ± S.D. of the net optical density in six replicate wells (optical density measured using a 620-nm high-pass filter subtracted from that measured using a 570-nm high-pass filter). The concentration giving 50% of the maximum response was calculated and expressed as a fraction of that of standard rIL-2 assayed in the same experiment: these values are given in parentheses in the legend.
would be inactivated. It was also observed that the percentage encapsulation of rIL-2 was higher in sonicated liposomes. If the association with lipids is saturable, the excess rIL-2 may be located in the aqueous compartment of the liposome and may be released more easily to bind to CTLL-2 cells. 3.2.2. SDS-PAGE analysis In order to investigate the effect of the lipid environment on the self-association of rIL-2, we incubated the protein with preformed liposomes (prepared by sonication in the same conditions as described above) and then performed SDS-PAGE. Fig. 3 shows that the intensity of the band at 25 kDa increased after incubation with DPPC liposomes.
Fig. 3. SDS-PAGE of rIL-2 after incubation with liposomes. rIL-2 (820 µl at 916 µg/ml) was incubated for 60 min at room temperature with 2180 µl DPPC liposomes prepared by extrusion (10 mg lipid). Samples 5 µg of rIL-2 or an equivalent amount of liposomes alone were loaded onto a 12% gel and run under reducing conditions. Proteins were revealed by silver staining. (1) Molecular weight markers; (2) control rIL-2; (3) liposomes alone; (4) rIL-2 incubated with liposomes.
Liposomes without protein showed no silver-stained bands. Therefore, it seems that in the absence of any physical treatment, the hydrophobic environment promotes the formation of dimers. However, the nature of these objects may not be the same as those produced by ultrasonication. Therefore, it is impossible to speculate about the biological activity of these dimers.
4. Conclusion This study has shown that some physical treatments used in the formulations of liposomes are not extremely detrimental to rIL-2 and that a high percentage association of the cytokine with small liposomes can be obtained. We can conclude that liposome-associated rIL-2 is less active than free rIL-2 because the access of the protein to its receptor is sterically hindered by the lipid. rIL-2 release from the liposomes in the 48-h culture system is probably too slow to prevent a large proportion of the cells from dying by apoptosis. However, this sort of formulation may be useful as a slow release vehicle for interleukin 2 in vivo. A study of the release of rIL-2 from liposomes under cell culture conditions is now under way. Another issue that will be addressed is the formulation of liposomes which remain in the circulation for an extended period, for example by including poly(ethylene glycol)-grafted lipids in the bilayer.
References [1]
Advant SJ, Braswell EH, Kumar CV, Kalonia DS. The effect of pH and temperature on the self-association of recombinant human interleukin-2 as studied by equilibrium sedimentation. Pharm Res 1995;12:637–41.
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167 [2]
Anderson PM, Katsanis E, Sencer SF, Hasz D, Ochoa AC, Bostrom B. Depot characteristics and biodistribution of interleukin-2 liposomes: importance of route of administration. J Immunother 1992;12:19.
[3]
Bergers JJ, Den Otter W, Dullens HF, Kerkvliet CT, Crommelin DJ. Interleukin-2-containing liposomes: interaction of interleukin-2 with liposomal bilayers and preliminary studies on application in cancer vaccines. Pharm Res 1993;10:1715.
[4]
Bergmann CA, Landmeier BJ, Kaplan DR. Phase separation analysis of recombinant interleukin 2. Mol Immunol 1991;28:99–105.
[5]
Castignolles N, Betbeder D, Ioualalen K, Merten O, Leclerc C, Samain D, et al. Stabilization and enhancement of interleukin-2 in vitro bioactivity by new carriers: supramolecular biovectors. Vaccine 1994;12:1413.
[6]
Fleischmann JD, Wentworth D, Valencic F, Imbembo AL, Koehler KA. Interleukin-2 self-association. Biochem Biophys Res Commun 1988;152:879–85.
[7]
Fleury L. Paris Sud: Thesis University; 1993.
[8]
Fleury L, Ollivon M, Dubois JL, Puisieux F, Barratt G. Preparation and characterization of dipalmitoylphosphatidylcholine liposomes containing interleukin-2. Braz J Med Biol Res 1995;28:519–29.
[9]
Kaplan D, Smith D, Huang R, Yildirim Z. Self-association of interleukin 2 bound to its receptor. Faseb J 1995;9:1096–102.
[10] Kaplan DR. Delivery of interleukin 2 for immunotherapy. J Chromatogr B 1994;662:315–23. [11] Kedar E, Gur H, Babai I, Samira S, Even-Chen S, Barenholz Y. Delivery of cytokines by liposomes: hematopoietic and immunomodulatory activity of interleukin-2 encapsulated in conventional liposomes and long-circulating liposomes. J Immunother 2000;23: 131.
167
[12] Khanna C, Anderson PM, Hasz DE, Katsanis E, Neville M, Klausner JS. Interleukin-2 liposome inhalation therapy is safe and effective for dogs with spontaneous pulmonary metastases. Cancer 1997;79: 1409. [13] Koppenhagen FJ, Balemans LTM, Steerenberg PA, Jagmont TM, Den Otter W, Storm G. The design of a pharmaceutically acceptable liposomal formulation of recombinant interleukin-2 (rIL-2) for locoregional anticancer immunotherapy. J Liposome Res 1999;9:313. [14] Koppenhagen FJ, Visser AJ, Herron JN, Storm G, Crommelin DJ. Interaction of recombinant interleukin-2 with liposomal bilayers. J Pharm Sci 1998;87:707–14. [15] Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680. [16] Liu LS, Liu SQ, Ng S, Heller J, Froix M. The potential application of alginate/chitosan porous microsphere loaded with interleukin-2 in tumour immunotherapy. Proceed Intern Symp Control Rel Bioact Mater 1995;22:542. [17] Matsuoka J, Sakagami K, Shiozaki S, Uchida S, Fujiwara T, Gohchi A, et al. Development of an interleukin-2 slow delivery system. ASAIO Trans 1988;34:729. [18] Moya G, Gonzalez LJ, Huerta V, Garcia Y, Morera V, Perez D, et al. Isolation and characterization of modified species of a mutated (Cys125-Ala) recombinant human interleukin-2. J Chromatogr A 2002;971:129–42. [19] Pellequer Y, Ollivon M, Barratt G. Methodology for assaying recombinant interleukin-2 associated with liposomes by combined gel exclusion chromatography and fluorescence. J Chromatogr B 2003; 783:151–62. [20] Vaage J, Mayhew E. Immunotherapy of a mouse mammary carcinoma by sustained peritumor release of IL-2. Int J Cancer 1991;47:582–5. [21] Wray W, Boulikas T, Wray VP, Hancok R. Silver staining of proteins in polyacrylamide gels. Anal Biochem 1981;118:197.
Dossier: Drug delivery and drug efficacy
Formulation of liposomes associated with recombinant interleukin-2: effect on interleukin-2 activity Yann Pellequer, Michel Ollivon, Gillian Barratt * UMR CNRS 8612, Université Paris-Sud, 5, rue Jean-Baptiste Clément, 92296 Châtenay-Malabry, France Received 9 October 2003; received in revised form 3 December 2003; accepted 4 December 2003
Abstract Association of the cytokine interleukin-2 (rIL-2) within liposomes could prolong its circulating half-life and thus reduce side-effects and improve its efficacy in cancer and AIDS treatment. The effects of physical procedures used in liposome preparation on the biological activity of rIL-2 were determined. While heating to 50 °C reduced the activity of IL-2 in the CTLL-2 proliferation assay by 50%, sonication, either bath or probe, was less detrimental. The combination of all three treatments resulted in only 10% loss of activity. Probe sonication led to the appearance of dimers which were stable under reducing conditions. Small unilamellar and large unilamellar liposomes were formed, respectively, by probe sonication or extrusion of multilamellar vesicles of dipalmitoylphosphatidylcholine hydrated in the presence of rIL-2. A high proportion of the rIL-2 was associated with the vesicles. However, the biological activity of the liposome-associated rIL-2 was reduced 7- to 10-fold compared with control rIL-2. rIL-2 dimers were formed on contact with lipid, even without sonication. We can conclude that the association of rIL-2 with lipid masks its access to its cell-surface receptor at least under cell culture conditions. © 2004 Elsevier SAS. All rights reserved. Keywords: Interleukin-2; Liposomes; CTLL-2 proliferation
1. Introduction The production of recombinant interleukin-2 (rIL-2) has cleared the way for large-scale clinical applications, notably in oncology and AIDS treatment because of this cytokine’s central role in the regulation of the immune response. However, rIL-2 is a 15-kDa protein which is rapidly eliminated by renal filtration and serious side effects, including capillary leak syndrome, are associated with the systemic administration of the high doses of rIL-2 necessary to attain effective concentrations. Different systems such as minipellets [17], supramolecular biovectors [5], microspheres [16] and liposomes [2,3,11–13], have been tested for their ability to increase efficiency and reduce side effects by increasing the biological half-life of rIL-2.
Abbreviations: DPPC, dipalmitoylphosphatidylcholine; LUV, large unilamellar vesicles; MLV, multilamellar vesicles; OG, N-octyl b,Dglucopyranoside; RIL-2, recombinant human interleukin-2; SUV, small unilamellar vesicles. * Corresponding author. E-mail address: [email protected] (G. Barratt). © 2004 Elsevier SAS. All rights reserved. doi:10.1016/j.biopha.2003.12.008
Liposomes are interesting systems for drug delivery because of their capacity to include hydrophobic molecules within their bilayer membranes. This would be advantageous for rIL-2 association because the non-glycosylated protein is very poorly water-soluble. Liposomes have been employed as a delivery system for rIL-2 in systemic [2,3,11] as well as in locoregional immunotherapy [12,13]. High association of rIL-2 with small unilamellar liposomes prepared from dipalmitoylphosphatidylcholine (DPPC) has already been observed [8]. This synthetic, saturated lipid would be expected to yield liposomes which are more stable in storage (resistant to oxidation) and in biological medium (resistant to destabilisation by plasma proteins and lipoproteins) than those prepared from natural phospholipids such as egg phosphatidylcholine. This would make them more suitable as controlled release systems; however, liposome formation must be performed at a temperature above the phase transition temperature of 41 °C, which could be detrimental to proteins, causing denaturation and loss of activity. The procedures used to reduce liposome size, essential to ensure their prolonged presence in the circulation, might also affect proteins. The aim of this work was to test the effect of formulation conditions on the physical integrity and biological activity of rIL-2.
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
2. Materials and methods 2.1. Materials DPPC was purchased from Avanti Polar Lipids (Alabaster, AL, USA). N-octyl b,D-glucopyranoside (OG) was obtained from Sigma Chemicals Co, (St. Louis, MI, USA). Recombinant IL-2 (rIL-2) with a specific activity of 1.1 × 107 U/mg protein (RU 49637) produced in Escherichia coli was provided by Roussel-Uclaf (France). In this work, rIL-2 was purified to remove dextran used as a lyophilisation stabiliser as described in Pellequer et al. [19]. 2.2. Stability of rIL-2 under conditions of liposome preparation rIL-2 at 250 µg/ml in citric acid solution (9 g/l, pH 3) was maintained either in a water bath at 50 °C (1 or 15 min) or 1 min in a bath sonicator (at 20 or 50 °C) (Branson 2200, Branson Ultrasonics Corporation, Danbury, USA) or subjected to a probe sonication using a Vibracel 600 W ultrasound generator (Bioblock Scientific) at about 90 W power and 50% cycle time for five sonication cycles of 2 min, each separated by 1 min of rest (at 38 °C or after 1 min in a bath sonicator at 50 °C or without heating or bath sonication). Titanium particles generated by the probe sonication were removed from the liposome suspension by centrifugation at 4500 rpm, for 5 min at 10 °C (rotor No12158, angle rotor 6 × 30 ml, maximum radius 7.9 cm, min. radius 2.3 cm, angle 30°) in a Nalgene® tube (Nalge Company, Rochester, NY, USA). rIL-2 integrity and biological activity were determined by SDS-PAGE and bioassay, respectively, as described below. 2.3. Liposome preparation Liposome preparation is described in Pellequer et al. [19]. Briefly, DPPC (10 mg) was dissolved in chloroform (about 0.5 ml) and deposited as a film on a flat-bottomed flask by evaporation under nitrogen followed by overnight vacuum drying. This film was hydrated with 3 ml rIL-2 solution (250 µg/ml, pH 3). The multilamellar vesicles (MLV) formed were homogenised by bath sonication (1 min at 50 °C). Size reduction of liposomes was carried out by two different procedures: sonication or extrusion, as follows. Small unilamellar vesicles (SUV) were prepared by probe sonication at 38 °C at about 90 W power and 50% cycle time. Five sonication cycles of 2 min were each separated by 1 min of rest. Titanium particles generated by sonication were removed from the liposome suspension by centrifugation at 4500 rpm, for 5 min at 10 °C as described above. Large unilamellar vesicles (LUV) were prepared at 50 °C by successive extrusions of the MLV suspension under nitrogen pressure (maximum pressure = 3000 kPa). The latter was passed twice through each of successively 0.8, 0.4, 0.2, 0.1, 0.005 µm polycarbonate filters (Poretics Corporation, Livermore, USA).
163
The mean particle diameter was estimated by quasi-elastic light scattering (Nanosizer N4MD, Coulter Electronics Inc., Florida, USA) using the “Unimodal” method of date processing. The samples were analyzed at 25 °C; the viscosity and refractive index of the aqueous suspending medium were taken to be 0.890 mPa s and 1.3325, respectively. The mean diameters of liposomes obtained by sonication and extrusion were 52 and 108 nm (the standard deviations of one normally distributed dispersion were 10 and 14 nm), respectively. 2.4. Quantification of rIL-2 associated with liposomes rIL-2 was assayed by a method base on the intrinsic fluorescence of its single tryptophan residue as described by Pellequer et al. [19]. Briefly, liposomes were separated from free rIL-2 by exclusion chromatography on Sephadex® G75 Superfine swollen in citric acid solution (9 g/l, pH 3). The elution of rIL-2 liposomes and free rIL-2 was monitored by measuring the fluorescence with a four-channel fluorometer Spex F1T11I (equipped with Peltier-cooled photomultipliers and a 450 W xenon lamp) controlled by DM 3000 software (Spex industries Inc, Edison, USA). Fluorescence emission was measured at 90°. Samples were maintained at 25 °C. rIL-2 elution was monitored at 320 nm (tryptophan fluorescence emission) after excitation at 290 nm, while liposomes were detected at the same time by light scattering at 290 nm. Two hundred microlitres of liposomes (1:1 dilution) were disrupted and solubilised by 2.25 mg of OG. Samples were analysed by fluorescence (kexc = 290 nm; kem = 320 nm). The fluorescence intensity of a micellar solution (1.2 mg DPPC, 58 mg OG, 5 ml water) was used as a blank. Each result is the mean of three measurements. A calibration curve was obtained by dilution of rIL-2 in the micellar solution of DPPC/OG/water. 2.5. rIL-2 activity assay by CTLL-2 proliferation 2.5.1. CTLL-2 cell line The rIL-2 dependent mouse T cell line (CTLL-2 ECACC No 93042610) was cultured in 500 ml RPMI 1640 medium with 100 IU/ml penicillin, 100 µg/ml streptomycin, 5 mM sodium pyruvate, 10% foetal calf serum, 0.5 µl/ml b-mercaptoethanol, and 0.01 ml/ml non-essential amino acids at 37 °C in 5% CO2. The cell line was maintained with 10 IU/ml rIL-2. Dilutions of rIL-2 from the concentrated stock solution were made in the CTLL-2 culture medium containing serum. 2.5.2. MTT assay Cells were harvested by centrifugation, washed twice in phosphate buffered saline (PBS) and adjusted to 105 cells/ml in the CTLL-2 medium without rIL-2. rIL-2 or liposomes containing rIL-2 (200 UI/ml) was diluted (eight twofold dilutions) in flat bottomed 96-well microtiter plates (TPP, Trasadingen, Switzerland). Each dilution was made in six replicates. The CTLL-2 cells were added (100 µl) to the rIL-2
164
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
dilutions. MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) (Sigma, St. Louis, MO, USA) was diluted to 5 mg/ml in sterile PBS and filtered once through a 0.22 µm filter. After 48 h, the diluted MTT was added to the wells as 20 µl/well. Four hours after at 37 °C, the plates were centrifuged and the supernatants were eliminated. Two hundred microlitres of the extraction solution (0.1 g/ml sodium dodecyl sulphate, 0.5 ml/ml dimethyl formamide) were added to each well. The microtiter plates were incubated to overnight to dissolve reduced MTT crystals. Absorbance values were measured on a microtiter plate reader at 570 nm with a reference wavelength of 620 nm. The dose–response curves were plotted as mean absorbance against rIL-2 concentration measured in the fluorescence assay. The concentration giving 50% of the maximum response was calculated and expressed as a fraction of that of standard rIL-2 assayed in the same experiment. The ED50 of the standard rIL-2 varied depending on the passage number of the cell line. Each IL-2 preparation was assayed only once, but since they were always compared with standard rIL-2 assayed in identical conditions, we consider that the relative activities are accurate. 2.6. rIL-2 SDS-PAGE rIL-2 or liposomes samples were boiled in Laemmli sample buffer [15] and run on 12% SDS-PAGE under reducing (10% b-mercaptoethanol) conditions. Proteins in the gel were stained with silver nitrate according to the method described in Ref. [21]. Bands were compared with standard molecular weight markers (Gibco, Cergy-Pontoise, France).
3. Results and discussion 3.1. Activity of rIL-2 after ultrasound and/or temperature treatment 3.1.1. Cell proliferation assay Before studying the activity of rIL-2 associated with liposomes, it was necessary to determine the effect of the physical treatments used in the preparation of liposomes on the integrity of the protein. Since it is necessary to prepare liposomes above the phase transition temperature of the lipids used (in this case 41 °C for DPPC), the effect of heating to 50 °C for 1 and 15 min under the same conditions as those described in Section 2.3 was tested. Table 1 shows the ratio of rIL-2 activity in a cell proliferation test, based on the ED50 values for rIL-2 treated in various ways compared with control rIL-2. The ability of rIL-2 to maintain the growth of CTLL-2 cells was reduced by 50% by heating, regardless of the duration. Ultrasound treatment is used both to help form MLV and for size reduction to SUV. We observed that neither bath nor probe sonication had a dramatic effect on the rIL-2 molecule, since the activity was, respectively, 75% and 80% of the
Table 1 Effect of physical treatments used in liposome preparation on the biological activity of IL-2 Physical treatments T = heating T = heating T = 20 °C T = heating T = 20 °C T = heating T = heating
t = 1′ t = 15′ Bath sonicator t = 1′ Bath sonicator t = 1′ Probe sonicator t = 15′ Probe sonicator t = 15′ Bath and probe sonicator
IL-2 treated activity/IL-2 standard activity 0.5 0.5 0.75 0.8 0.8 0.8 0.9
rIL-2 (250 µg/ml) underwent various treatments as described in Section 2. Biological activity was assessed by CTLL-2 proliferation. The results are expressed as a ratio of the concentrations of treated and untreated rIL-2 which provoked half-maximal cell growth.
control. Surprisingly, this activity was not further reduced when sonication was carried out at 50 °C. rIL-2 which had undergone a combination of all physical treatments used to make SUV, bath sonication followed by probe sonication, both at elevated temperature, showed activity which was almost as high as that of control rIL-2. Thus, against all expectations, the effects of heating and of ultrasound energy were not additive, but seemed to counteract each other. 3.1.2. SDS-PAGE analysis In an attempt to explain these results the molecular state of the rIL-2 was analysed by SDS-PAGE. The ability of nonglycosylated rIL-2 to form homodimers has already been described in the literature [1,4,6,9,10,18]. In our previous work [19], we observed higher molecular weight bands in rIL-2 preparations by both gel exclusion chromatography and SDS-PAGE. Control rIL-2 showed a minor band at 25 kDa (Fig. 1). This was not altered by elevated temperature or bath sonication. On the other hand, treatments which included probe sonication showed increased intensity of this band. Since each lane was loaded with the same amount of protein as determined by the fluorescence assay, this increase is not due to selective destruction of the monomer but to an association brought about by the treatment. It should be noted that these gels were silver-stained and not subjected to Western blotting, so we have no formal proof that the higher molecular weight band is in fact a dimer of rIL-2. However, the fact that the molecular weight was twice that of the main band strongly suggests that this is the case. The gels were run in reducing conditions, which means that this dimerisation cannot be attributed to intermolecular disulphide bridges. Kaplan et al. [9] also observed multimers of rIL-2 in reducing conditions and suggested that they might be formed by transglutamination. They also proposed that self-associated rIL-2 had increased in-vitro activity based on the observation that the radioiodinated cytokine, which did not associate in this way, was less effective in a lymphocyte proliferation assay [9]. We did not observe a direct correlation between rIL-2 activity in the CTLL-2 (Table 1) and the appearance of
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
165
Fig. 1. SDS-PAGE of rIL-2 after various physical treatments. rIL-2 (250 µg/ml) underwent various treatments as described in Section 2. Samples 7.5 µg were loaded onto a 12% gel and run under reducing conditions. Proteins were revealed by silver staining. (1) Molecular weight markers; (2) control rIL-2; (3) bath sonication, 1 min without heating; (4) heating to 50 °C for 1 min; (5) heating for 15 min; (6) bath sonication with heating; (7) probe sonication with heating; (8) bath and probe sonication with heating; (9) probe sonication without heating.
dimers (Fig. 1) after the different treatments of rIL-2. However, our bioassay differs from that used by Kaplan et al. [9] in that they used a human T cell clone whereas we used the mouse CTLL-2 which undergoes apoptosis in the absence of rIL-2. Thus, any “sustained release” effect of self-associated rIL-2 as observed by Kaplan et al. [10] would be masked by the fact that if the cells are not exposed to the cytokine at the beginning of the incubation period they will become committed to apoptosis and will not be able to be stimulated for proliferation at later times. Another difference between our work and that of Kaplan et al. [9] is the exact sequence of the rIL-2 used. Our protein, from Roussel-Uclaf, is the natural IL-2 sequence, while that used by Kaplan was from the Chiron Corporation with two amino acid differences. Small differences in primary structure can affect the hydrophobicity of the protein and its self-aggregation behaviour [4]. In particular, the Chiron rIL-2 was found to aggregate in solution at neutral pH, leading to a change in the intrinsic fluorescence [6]. We did not observe any non-linearity of our fluorescence calibration curve; however, the protein was always maintained at pH 3, at which it shows maximal solubility in aqueous media. This is in contrast to the results of Advant et al. [1] who observed more extensive aggregation of the Chiron protein at pH 3.5 than at pH 6.5 and 8.2 by analytical ultracentrifugation. We can speculate that heating causes a denaturation of the rIL-2 molecule without self-association which reduces its ability to bind to its receptors and thereby stimulate CTLL-2 cells. On the other hand, ultrasound energy seems to perturb the molecule in a way which is conducive to forming dimers. This may protect rIL-2 from thermal denaturation. In the CTLL-2 assay, self-associated rIL-2 may provide sustained release over the 48-h incubation period and compensate for the loss of activity of monomeric rIL-2 on heating. 3.2. Activity of rIL-2 associated with liposomes rIL-2 was incorporated into liposomes prepared from DPPC. Although the results presented above suggest that sonication does not inactivate the cytokine to a great extent, extrusion was also used as a size-reduction technique because this procedure is more easily standardised and appli-
cable to large-scale production. The biological activity of liposome-associated rIL-2 was measured before and after separation of non-associated rIL-2 by gel chromatography. The concentrations of rIL-2 were measured by the fluorescence assay. 3.2.1. Cell proliferation assay Fig. 2 shows the dose–response curves in the CTLL-2. Liposome-associated rIL-2 was much less effective in maintaining cell growth than the free cytokine. Determination of the ED50 values revealed that for the non-separated samples, rIL-2 associated with sonicated and extruded liposomes was, respectively, 7 and 9.5 times less active than control rIL-2. When free rIL-2 was removed by gel filtration, the ratios were 7 and 11 times. Control experiments (data not shown) indicated that liposomes were not toxic for CTLL-2 cells. The results presented above show that loss of rIL-2 activity during the preparation of the liposomes cannot explain the loss of activity. The association of rIL-2 with the liposomes was very high. Gel exclusion chromatography revealed only a small peak corresponding to free rIL-2 for extruded liposomes and no detectable second peak for sonicated ones [19]. This is consistent with the observation that the rIL-2 activity did not change on separation of sonicated liposomes while it dropped after extruded liposomes were separated, suggesting that the free rIL-2 was more active than the associated. High association of recombinant rIL-2 with liposomes has already been observed by several authors [7,8,14] and can be attributed to interaction of the hydrophobic non-glycosylated protein with the liposome bilayer. This interaction could prevent recognition of rIL-2 molecules by the cell-surface receptor. The fact that association with liposomes protects rIL-2 against enzymatic degradation [8,20] suggests that the protein is at least partially embedded in the phospholipid bilayer. The reduction of rIL-2 activity on liposome association was less marked with the sonicated liposomes than with the extruded ones. This is consistent with the observation than probe sonication counteracted the effect of temperature on rIL-2, since extruded liposomes are subjected to bath sonication and heating, but not to probe sonication, therefore, a smaller proportion of the rIL-2 in the sonicated liposomes
166
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167
0,8 Standard IL2 for sonication 0,7
Sonicated liposomes (x 0.14) Sonicated liposomes without free IL-2 (x 0.14)
0,6
Standard IL2 for extrusion
A (570-620 nm)
Extruded liposomes (x 0.10) 0,5
Extruded liposomes without free IL-2 (x 0.09)
0,4
0,3
0,2
0,1
0 0,1
1
10 IL-2/well (IU/ml)
100
1000
Fig. 2. Biological activity of liposome-associated rIL-2. Liposomes were prepared and separated from non-associated rIL-2 as described in Section 2. Protein concentrations were measured by fluorescence. Serial twofold dilutions of liposomes and standard rIL-2 were incubated with CTLL-2 for 48 h, after which the MTT conversion test was performed. Values are the mean ± S.D. of the net optical density in six replicate wells (optical density measured using a 620-nm high-pass filter subtracted from that measured using a 570-nm high-pass filter). The concentration giving 50% of the maximum response was calculated and expressed as a fraction of that of standard rIL-2 assayed in the same experiment: these values are given in parentheses in the legend.
would be inactivated. It was also observed that the percentage encapsulation of rIL-2 was higher in sonicated liposomes. If the association with lipids is saturable, the excess rIL-2 may be located in the aqueous compartment of the liposome and may be released more easily to bind to CTLL-2 cells. 3.2.2. SDS-PAGE analysis In order to investigate the effect of the lipid environment on the self-association of rIL-2, we incubated the protein with preformed liposomes (prepared by sonication in the same conditions as described above) and then performed SDS-PAGE. Fig. 3 shows that the intensity of the band at 25 kDa increased after incubation with DPPC liposomes.
Fig. 3. SDS-PAGE of rIL-2 after incubation with liposomes. rIL-2 (820 µl at 916 µg/ml) was incubated for 60 min at room temperature with 2180 µl DPPC liposomes prepared by extrusion (10 mg lipid). Samples 5 µg of rIL-2 or an equivalent amount of liposomes alone were loaded onto a 12% gel and run under reducing conditions. Proteins were revealed by silver staining. (1) Molecular weight markers; (2) control rIL-2; (3) liposomes alone; (4) rIL-2 incubated with liposomes.
Liposomes without protein showed no silver-stained bands. Therefore, it seems that in the absence of any physical treatment, the hydrophobic environment promotes the formation of dimers. However, the nature of these objects may not be the same as those produced by ultrasonication. Therefore, it is impossible to speculate about the biological activity of these dimers.
4. Conclusion This study has shown that some physical treatments used in the formulations of liposomes are not extremely detrimental to rIL-2 and that a high percentage association of the cytokine with small liposomes can be obtained. We can conclude that liposome-associated rIL-2 is less active than free rIL-2 because the access of the protein to its receptor is sterically hindered by the lipid. rIL-2 release from the liposomes in the 48-h culture system is probably too slow to prevent a large proportion of the cells from dying by apoptosis. However, this sort of formulation may be useful as a slow release vehicle for interleukin 2 in vivo. A study of the release of rIL-2 from liposomes under cell culture conditions is now under way. Another issue that will be addressed is the formulation of liposomes which remain in the circulation for an extended period, for example by including poly(ethylene glycol)-grafted lipids in the bilayer.
References [1]
Advant SJ, Braswell EH, Kumar CV, Kalonia DS. The effect of pH and temperature on the self-association of recombinant human interleukin-2 as studied by equilibrium sedimentation. Pharm Res 1995;12:637–41.
Y. Pellequer et al. / Biomedicine & Pharmacotherapy 58 (2004) 162–167 [2]
Anderson PM, Katsanis E, Sencer SF, Hasz D, Ochoa AC, Bostrom B. Depot characteristics and biodistribution of interleukin-2 liposomes: importance of route of administration. J Immunother 1992;12:19.
[3]
Bergers JJ, Den Otter W, Dullens HF, Kerkvliet CT, Crommelin DJ. Interleukin-2-containing liposomes: interaction of interleukin-2 with liposomal bilayers and preliminary studies on application in cancer vaccines. Pharm Res 1993;10:1715.
[4]
Bergmann CA, Landmeier BJ, Kaplan DR. Phase separation analysis of recombinant interleukin 2. Mol Immunol 1991;28:99–105.
[5]
Castignolles N, Betbeder D, Ioualalen K, Merten O, Leclerc C, Samain D, et al. Stabilization and enhancement of interleukin-2 in vitro bioactivity by new carriers: supramolecular biovectors. Vaccine 1994;12:1413.
[6]
Fleischmann JD, Wentworth D, Valencic F, Imbembo AL, Koehler KA. Interleukin-2 self-association. Biochem Biophys Res Commun 1988;152:879–85.
[7]
Fleury L. Paris Sud: Thesis University; 1993.
[8]
Fleury L, Ollivon M, Dubois JL, Puisieux F, Barratt G. Preparation and characterization of dipalmitoylphosphatidylcholine liposomes containing interleukin-2. Braz J Med Biol Res 1995;28:519–29.
[9]
Kaplan D, Smith D, Huang R, Yildirim Z. Self-association of interleukin 2 bound to its receptor. Faseb J 1995;9:1096–102.
[10] Kaplan DR. Delivery of interleukin 2 for immunotherapy. J Chromatogr B 1994;662:315–23. [11] Kedar E, Gur H, Babai I, Samira S, Even-Chen S, Barenholz Y. Delivery of cytokines by liposomes: hematopoietic and immunomodulatory activity of interleukin-2 encapsulated in conventional liposomes and long-circulating liposomes. J Immunother 2000;23: 131.
167
[12] Khanna C, Anderson PM, Hasz DE, Katsanis E, Neville M, Klausner JS. Interleukin-2 liposome inhalation therapy is safe and effective for dogs with spontaneous pulmonary metastases. Cancer 1997;79: 1409. [13] Koppenhagen FJ, Balemans LTM, Steerenberg PA, Jagmont TM, Den Otter W, Storm G. The design of a pharmaceutically acceptable liposomal formulation of recombinant interleukin-2 (rIL-2) for locoregional anticancer immunotherapy. J Liposome Res 1999;9:313. [14] Koppenhagen FJ, Visser AJ, Herron JN, Storm G, Crommelin DJ. Interaction of recombinant interleukin-2 with liposomal bilayers. J Pharm Sci 1998;87:707–14. [15] Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680. [16] Liu LS, Liu SQ, Ng S, Heller J, Froix M. The potential application of alginate/chitosan porous microsphere loaded with interleukin-2 in tumour immunotherapy. Proceed Intern Symp Control Rel Bioact Mater 1995;22:542. [17] Matsuoka J, Sakagami K, Shiozaki S, Uchida S, Fujiwara T, Gohchi A, et al. Development of an interleukin-2 slow delivery system. ASAIO Trans 1988;34:729. [18] Moya G, Gonzalez LJ, Huerta V, Garcia Y, Morera V, Perez D, et al. Isolation and characterization of modified species of a mutated (Cys125-Ala) recombinant human interleukin-2. J Chromatogr A 2002;971:129–42. [19] Pellequer Y, Ollivon M, Barratt G. Methodology for assaying recombinant interleukin-2 associated with liposomes by combined gel exclusion chromatography and fluorescence. J Chromatogr B 2003; 783:151–62. [20] Vaage J, Mayhew E. Immunotherapy of a mouse mammary carcinoma by sustained peritumor release of IL-2. Int J Cancer 1991;47:582–5. [21] Wray W, Boulikas T, Wray VP, Hancok R. Silver staining of proteins in polyacrylamide gels. Anal Biochem 1981;118:197.