In Vitro Antitumor and Neuroprotective Effects and Optimization of

Hadassah UniVersity Hospital, Jerusalem, Israel, and Department of ... Therapeutics, School of Pharmacy, Faculty of Medicine, the Hebrew UniVersity of...
1 downloads 0 Views 211KB Size
Download PDF
Langmuir 2007, 23, 1937-1947

1937

The Antioxidant Tempamine: In Vitro Antitumor and Neuroprotective Effects and Optimization of Liposomal Encapsulation and Release Veronica Wasserman,†,# Pablo Kizelsztein,†,# Olga Garbuzenko,† Ron Kohen,‡ Haim Ovadia,§ Rinat Tabakman,⊥ and Yechezkel Barenholz*,† Laboratory of Biochemistry and Liposome Research, Hebrew UniVersity-Hadassah Medical School, Jerusalem, Israel, Department of Pharmaceutics, School of Pharmacy, the Hebrew UniVersity of Jerusalem, Jerusalem, Israel, Agnes Ginges Center for Human Neurogenetics, Department of Neurology, Hadassah UniVersity Hospital, Jerusalem, Israel, and Department of Pharmacology and Experimental Therapeutics, School of Pharmacy, Faculty of Medicine, the Hebrew UniVersity of Jerusalem, Jerusalem, Israel ReceiVed January 23, 2006. In Final Form: August 27, 2006 The piperidine nitroxide tempamine (TMN) is a cell-permeable, stable radical having antioxidant, anticancer, and proapoptotic and/or pronecrotic activities, as was demonstrated by us in cell cultures. We also demonstrated synergism between TMN and doxorubicin in doxorubicin-sensitive and doxorubicin-resistant cell lines. Treatment of the C26 mouse colon carcinoma model in vivo also demonstrated synergism between TMN and doxorubicin in sterically stabilized liposomes (SSLs) containing TMN (SSL-TMN) and those containing doxorubicin. The above effects of TMN and SSL-TMN motivated us to develop and optimize the SSL-TMN formulation so that it will be able to reach the disease site with a sufficiently high TMN level and a release rate needed to achieve a therapeutic effect. Because TMN is an amphipathic weak base, it was remote loaded by an intraliposome high/extraliposome low transmembrane ammonium sulfate gradient. The kinetics and level of TMN loading were monitored by cyclic voltammetry (CV) and electron paramagnetic resonance (EPR); the latter also indicates TMN precipitation in the intraliposomal aqueous phase. The regeneration of the original CV and EPR signals by the ionophore nigericin indicates that TMN remained fully intact during loading and release. The cardinal role of the transmembrane ammonium ion gradient in the loading process was proven by the use of the selective ionophores nonactin (for NH4+) and nigericin (for H+). The anion of the ammonium salts affects loading stability and the rate of TMN release, both mediated through the TMN state of aggregation in the intraliposomal aqueous phase. The greater the TMN salt precipitation, the slower the TMN release rate. This was supported by measurement of osmolality, which is inversely related to TMN salt precipitate. Precipitation is in the order SO4-2 > Cl-1 > glucuronate-1. Liposome lipid composition, magnitude of the transmembrane ammonium ion gradient, and type of anion of the ammonium salt determine the amount of TMN loaded and its release rate.

1. Introduction Piperidine nitroxides such as Tempol, Tempo, and tempamine (TMN) are cell-permeable, nontoxic, and nonimmunogenic stable cyclic radicals.1,2 Among antioxidants, nitroxides are unusual in their mode of action, being mainly oxidants rather than reductants.3,4 They also possess the ability to be at least partially regenerated.4 Nitroxides exert their antioxidant activity through several mechanisms: SOD-mimicking, oxidation of reduced metal ions, reduction of hypervalent metals, and interruption of radical chain * To whom correspondence should be addressed. E-mail: [email protected]. † Hebrew University-Hadassah Medical School. ‡ Department of Pharmaceutics, the Hebrew University of Jerusalem. § Hadassah University Hospital. ⊥ Department of Pharmacology and Experimental Therapeutics, the Hebrew University of Jerusalem. # Equal contribution authors. (1) Afzal, V.; Brasch, R. C.; Nitecki, D. E.; Wolff, S. Nitroxyl spin label contrast enhancers for magnetic resonance imaging. Studies of acute toxicity and mutagenesis. InVest. Radiol. 1984, 19 (6), 549-52. (2) DeGraff, W. G.; Krishna, M. C.; Kaufman, D.; Mitchell, J. B. Nitroxidemediated protection against X-ray- and neocarzinostatin-induced DNA damage. Free Radical Biol. Med. 1992, 13 (5), 479-87. (3) Mitchell, J. B.; DeGraff, W.; Kaufman, D.; Krishna, M. C.; Samuni, A.; Finkelstein, E.; Ahn, M. S.; Hahn, S. M.; Gamson, J.; Russo, A. Inhibition of oxygen-dependent radiation-induced damage by the nitroxide superoxide dismutase mimic, Tempol. Arch. Biochem. Biophys. 1991, 289 (1), 62-70. (4) Samuni, A.; Mitchell, J. B.; DeGraff, W.; Krishna, C. M.; Samuni, U.; Russo, A. Nitroxide SOD-mimics: modes of action. Free Radical Res. Commun. 1991, 12-13 (1), 187-94.

reactions.3-5 Tempol and Tempo have been shown to possess antineoplastic activity by themselves and to enhance chemotherapy-induced apoptosis.6,7 In this study, we show that another piperidine nitroxide, TMN, possesses antitumor activity in various cell lines. Although its IC50 is lower than that of Tempol, it is still too high to make it a viable anticancer drug by itself. However, much lower TMN concentrations are sufficient to achieve synergistic effects with the anticancer drug doxorubicin (DOX) encapsulated in liposomes (Doxil). We also demonstrate the following potential applications of TMN: (i) protecting agent (antioxidant) against oxidative damage (using PC12 neurons as a model system); (ii) proapoptotic and/ or pronecrotic agent acting synergistically with commonly used chemotherapeutic drugs such as DOX. To increase therapeutic efficacy, one has to address the major issues of short half-life in the circulation and potential toxicity. Encapsulation in liposomes, and especially in ∼100-nm sterically stabilized liposomes (SSLs) may help to overcome these obstacles, as it (5) Krishna, M. C.; Grahame, D. A.; Samuni, A.; Mitchell, J. B.; Russo, A. Oxoammonium cation intermediate in the nitroxide-catalyzed dismutation of superoxide. Proc. Natl. Acad. Sci. U.S.A. 1992, 89 (12), 5537-41. (6) Gariboldi, M. B.; Lucchi, S.; Caserini, C.; Supino, R.; Oliva, C.; Monti, E. Antiproliferative effect of the piperidine nitroxide TEMPOL on neoplastic and nonneoplastic mammalian cell lines. Free Radical Biol. Med. 1998, 24 (6), 91323. (7) Shacter, E.; Williams, J. A.; Hinson, R. M.; Senturker, S.; Lee, Y. J. Oxidative stress interferes with cancer chemotherapy: inhibition of lymphoma cell apoptosis and phagocytosis. Blood 2000, 96 (1), 307-13.

10.1021/la060218k CCC: $37.00 © 2007 American Chemical Society Published on Web 01/09/2007

1938 Langmuir, Vol. 23, No. 4, 2007

should prolong plasma half-life and enable extravasation into tumor and inflammation sites. However, a major obstacle to the use of ∼100-nm SSLs is the small trapped aqueous volume (low drug-to-lipid ratio) obtained when conventional passive drug loading is applied, leading to a low therapeutic value. Thus, we selected the nitroxide TMN, which is an amphipathic weak base and therefore can be actively remote-loaded into the intraliposomal aqueous phase by means of an ammonium sulfate gradient under conditions in which [(NH4)SO4]liposome . [(NH4)2SO4]medium. A major part of this study was dedicated to the optimization of TMN loading in liposomes and, especially, to determine the various factors that affect the level of stable loading and the profile of TMN release from the liposome. The utility of SSLTMN was shown by demonstrating its synergism with a DOXSSL formulation (Doxil) in a mice tumor model. 2. Materials and Methods 2.1. Materials. 2,2,6,6-Tetramethylpiperidine-4-amino-1-oxyl (4amino-Tempo, termed tempamine and abbreviated TMN) free radical, 97%; 2,2,6,6-tetramethylpiperidine-1-oxyl (Tempo); and 4-hydroxy2,2,6,6- tetramethylpiperidine-1-oxyl (Tempol) were purchased from Aldrich (Milwaukee, WI). Egg phosphatidylcholine (EPC I) and hydrogenated soybean phosphatidylcholine (HSPC) were obtained from Lipoid KG (Ludwigshafen, Germany). N-carbamyl-poly(ethylene glycol methyl ether)-1,2-distearoyl-sn-glycero-3-phosphoethanolamine triethylammonium salt (2000PEG-DSPE) (the polyethylene moiety having a molecular mass of 2000 Da) was a gift from ALZA Corp. (Mountain View, CA). Cholesterol was obtained from Sigma (St. Louis, MO). Sephadex G-50 was obtained from Pharmacia (Uppsala, Sweden). tert-Butyl alcohol was purchased from BDH (Poole, UK). Other chemicals, including buffers, were obtained from Sigma and Biological Industries (Beit Haemek, Israel). 2.2. Methods. 2.2.1. Biological Assays. 2.2.1.1. Cell Lines and Culture Conditions. The following cell lines and culture media were used: two DOX-sensitive, MCF-7 (human breast adenocarcinoma) and M-109S (human breast carcinoma), and one DOX-resistant, M-109R (human breast carcinoma). Cells were maintained in RPMI medium supplemented with 10% fetal calf serum (FCS) (Biological Industries). The cell lines were maintained under standard culture conditions at 37 °C in a humidified 5% CO2 atmosphere. Pheochromocytoma PC12 neuronal cells were grown in Dulbecco’s modified Eagle’s medium, supplemented with 7% FCS, 7% horse serum, 100 µg/mL streptomycin, and 100 U/mL penicillin (Biological Industries). The cultures were maintained in an incubator at 37 °C in a humidified atmosphere of 6% CO2. The growth medium was changed twice weekly, and the cultures were split at a 1:6 ratio once a week.8 For cell differentiation, equal quantities of PC12 cells (3.75 × 105 cells) were plated on six-well plates coated with rat tail type I collagen (0.1 mg/mL) (Biological Industries) to promote cell adhesion.9 The differentiation of the cultures was induced by treatment with nerve growth factor (NGF) (50 ng/mL), added every 48 h for a period of 7-8 days. 2.2.1.2. Measurement of Cell Death by Lactate Dehydrogenase (LDH) Leakage Assay. PC12 neuronal cell death was evaluated by measuring the leakage of LDH into the growth medium as previously described.10 Samples of 50 µL of the growth medium were collected from each well and centrifuged at 3500 rpm for 5 min at 25 °C; the supernatant was collected, and LDH release was measured by the leakage of LDH enzyme into the medium, using a Sigma Diagnostics (8) Abu Raya, S.; Trembovler, V.; Shohami, E.; Lazarovici, P. A tissue culture ischemic device to study eicosanoid release by pheochromocytoma PC12 cultures. J. Neurosci. Methods 1993, 50 (2), 197-203. (9) Abu-Raya, S.; Bloch-Shilderman, E.; Lelkes, P. I.; Trembovler, V.; Shohami, E.; Gutman, Y.; Lazarovici, P. Characterization of pardaxin-induced dopamine release from pheochromocytoma cells: role of calcium and eicosanoids. J. Pharmacol. Exp. Ther. 1999, 288 (2), 399-406. (10) Tabakman, R.; Lazarovici, P.; Kohen, R. Neuroprotective effects of carnosine and homocarnosine on pheochromocytoma PC12 cells exposed to ischemia. J. Neurosci. Res. 2002, 68 (4), 463-9.

Wasserman et al. LD-L assay kit. LDH activity was determined using an ELISA reader (TECAN, SPECTRAFluor PLUS, Grodig, Salzburg, Austria) at 340 nm by following the rate of conversion of oxidized nicotinamide adenine dinucleotide (NAD+) to the reduced form, NADH. 1-Methyl4-phenyl pyridinium (MPP+)-induced LDH release was expressed as 100% toxicity compared to the control (untreated) cultures. Each experiment was performed three times in duplicate. 2.2.1.3. MPP+ Toxicity Experiment on PC12 Neurons. On the day of the experiment, NGF-containing medium was replaced with fresh medium. The cultures were divided into three groups: (1) control/untreated cells, (2) cultures exposed to MPP+ insult, and (3) TMN-treated cultures exposed to MPP+ insult. MPP+ was dissolved in growth medium containing NGF and added to each well in a final concentration of 1.5 mM. At the end of the experiment, medium was taken for evaluation of LDH release. During the experiment, all cultures were maintained in an incubator at 37 °C in a humidified atmosphere of 6% CO2. The experiment was ended when the percentage of cell death measured by the release of LDH into the medium was in the range of 30-60%. TMN dissolved in growth medium containing NGF was added to the cultures 1 h prior to the exposure to MPP+. For dose-response assay, TMN was administrated to each well in a final concentration of 0.1, 1, 10, 100, 500, or 1000 µM. Samples of 50 µL of medium were taken after 48 h for assessment of LDH release. 2.2.1.4. Cytotoxicity Assay. The effect of TMN, TMN+DOX, and Tempol on cell proliferation was determined by the methylene blue assay.11 Briefly, cells were seeded onto 96-well plates (MCF-7 cells at a density of 6 × 103 cells per well, and M-109S and M-109R cells at a density of 1.5 × 103 cells per well) and allowed to grow for 24 h prior to treatment with different concentrations of the nitroxides and DOX. After the addition of TMN (5 × 10-5 to 4 × 10-4 M), DOX (10-60 nM), or both, the cells were incubated in RPMI + 10% FCS for 3 days without changing the medium. Ninetysix hours after treatment, cells were fixed with 2.5% glutaraldehyde solution, and the amount of cells was assayed colorimetrically by methylene blue staining. 2.2.1.5. Determination of Phosphatidylserine (PS) Externalization. The percentage of cells showing PS externalization (indication of apoptosis and/or necrosis) was assessed by fluorescence-activated cell-sorting (FACS) analysis of cells with the fluorophore merocyanine-540. 1 × 106 MCF-7 cells were removed from culture, washed with PBS, and stained with merocyanine-540, which, by its binding to the externalized PS from the cell membrane, indicates the early steps of apoptosis.12 Briefly, the cell pellet was resuspended in 500 µL of phosphate-buffered saline (PBS). A 2.5 µL portion of a 1 mg/mL solution of merocyanine-540 was added to the cells and incubated for 2 min at room temperature in the dark. The cells were washed, resuspended in 1 mL of PBS, and run immediately on an FACS flow cytometer (Vantage, Becton Dickinson, Rutherford, NJ). 2.2.1.6. Determination of In Vitro Synergism between TMN and DOX. The effect of the combination of TMN and DOX on cell proliferation was determined by a methylene blue assay. For the assessment of synergy, the combination index (CI) was determined using the following equation: CI ) (D1/Dx1) + (D2/Dx2) + (D1D2/Dx1Dx2), where Dx is the individual drug concentration at its respective IC50, and D is the concentration of the drug in the combination that results in 50% growth inhibition. The subscripts 1 and 2 refer to TMN and DOX, respectively. A CI value less than 0.9 indicates synergism, CI ) 0.9-1.1 indicates additivity, and CI > 1.1 indicates antagonism. 2.2.1.7. Mice Tumor Model: In Vivo Studies. Eight to twelveweek-old BALB/c female mice were obtained from the Animal (11) Horowitz, A. T.; Barenholz, Y.; Gabizon, A. A. In vitro cytotoxicity of liposome-encapsulated doxorubicin: dependence on liposome composition and drug release. Biochim. Biophys. Acta 1992, 1109 (2), 203-9. (12) Reid, S.; Cross, R.; Snow, E. C. Combined Hoechst 33342 and merocyanine 540 staining to examine murine B cell cycle stage, viability and apoptosis. J. Immunol. Methods 1996, 192 (1-2), 43-54.

The Antioxidant Tempamine Facility and housed at the SPF unit of the Hebrew University (Jerusalem, Israel). The experimental procedures were in accordance with the standards required by the Institutional Animal Care and Use Committee of the Hebrew University-Hadassah Medical School. Each mouse was injected with one inoculum of tumor cells (1 × 106 C26 mouse cells) subcutaneously in the left flank. One week after inoculation, the mice were injected once intravenously (iv) through the tail vein with 0.36 mg of TMN in 100 µL buffer (2.1 µmol)/mouse ) 18 mg (105 µmol)/kg body weight, either in free TMN form or as SSL-TMN. The SSL-TMN phospholipid dose was 11 mg (14.7 µmol)/mouse ) 377 mg (514 µmol)/kg body weight. For the groups that received DOX either as free DOX (f-DOX) or as Doxil, the dose of the drug was 8 mg/kg, which was also administrated iv at the same time that TMN or SSL-TMN was administrated. 2.2.2. Physical and Chemical Assays Related to Formulating TMN in Liposomes. 2.2.2.1. n-Octanol/Aqueous Phase Partition Coefficient (Kp) of TMN. The n-octanol/water (oct/aq) partition coefficient Kp is the concentration ratio of a substance (permeant) between two phases in contact at equilibrium. One of the phases is a less polar phase, such as an n-octanol, heptane, or lipid bilayer, and the second phase is a more polar phase, usually an aqueous buffer.13 This parameter represents the lipophilicity (or hydrophilicity) of a substance, being one of the main factors in determining its membrane permeability.13,14 The oct/aq phase distribution of TMN was determined by a method described elsewhere,15 at pHs of 4.0, 7.0, and 10.6 for two different TMN concentrations (2.0 and 20.0 mM) and different concentrations of ammonium sulfate (20-400 mM) to imitate the variable conditions of the intraliposomal aqueous phase. TMN in each of the two phases was quantified by electron paramagnetic resonance (EPR), as described below. 2.2.2.2. Lipid Bilayer/Aqueous Phase Partition Coefficient (Kp) of TMN. The lipid bilayer/aqueous phase (0.15 M NaCl) distribution was determined using equilibrium dialysis.16 2.2.2.3. Liposome Preparation. The approach of lyophilization from tert-butyl alcohol (freezing temperature 22 °C) followed by mechanical hydration (vortexing) and downsizing by extrusion was used.16 All lipids were dissolved and mixed in tert-butyl alcohol, then the mixtures were lyophilized overnight. The dry lipid powder was hydrated with ammonium sulfate solution (150 mM). Hydration was carried out above the Tm of the matrix lipid: for HSPC, 60 °C (Tm ) 52.2 °C), and for EPC, room temperature (Tm ) -5 °C).14 Hydration was performed under continuous shaking, forming multilamellar vesicles (MLVs). The volume of hydration medium was adjusted to obtain a 10% (w/v) lipid concentration. Large unilamellar vesicles (LUVs; ∼100 nm) were prepared by stepwise extrusion using a 100-nm-pore-size polycarbonate filter as the final extrusion step.16 The liposome size distribution was determined by dynamic light scattering using either a Coulter (Model N4 SD) submicron particle analyzer or an ALV-NIBS/HPPS with an ALV5000/EPP multiple digital correlator (ALV-Laser Vertriebsgesellschaft GmbH, Langen, Germany).17 Size distributions of 1200 ( 200 nm (polymodal) and 100 ( 15 nm (unimodal) were obtained for MLVs and LUVs, respectively. Lipid (as raw material) purity, liposome lipid composition, and integrity after liposome preparation and during 2 months storage were determined by quantitative thin-layer chromatography.17

Langmuir, Vol. 23, No. 4, 2007 1939 Phospholipid concentration was determined using the modified Bartlett procedure.18 2.2.2.4. Formation of Transmembrane Intraliposome High/ Extraliposome Low Ammonium Sulfate Gradient. The dialysis approach developed in our laboratory16 was utilized. We used two consecutive dialysis exchanges against 100 vol of 0.15 M NaCl (pH ) 5.2) and a third dialysis exchange overnight against 100 vol of 10 mM histidine buffer, pH 6.7, in either 0.15 M NaCl or in 5% dextrose or 10% sucrose. 2.2.2.5. Determination of Magnitude of Transmembrane Ammonium Gradient Formation by [14C]-Methylamine Distribution between the Intraliposome Aqueous Phase and the Extraliposome Medium. The liposome transmembrane pH gradient was assessed from the distribution of [14C]-methylamine (MA) between the intraliposome aqueous phase and the external medium. This assay is based on the pioneer studies of Padan and Schuldiner,19 later adapted for liposomology by others, as exemplified by Dos Santos et al.20 Briefly, liposomes (10 mM) were incubated for 30 min above the Tm for a liposome-forming lipid (for EPC at 37 °C and for HSPC at 60 °C) with ∼50 × 103 dpm of [14C]-MA (1 × 108 dpm/mole). At the end of incubation, an aliquot of this mixture was passed down a Sephadex G-50 mini spin column equilibrated in 10 mM histidine buffer, pH 6.7, containing 10% sucrose. Liposomes were eluted at the column void volume and separated from the unencapsulated MA. The concentration of liposomes in the original liposomal dispersion and in the void volume fraction was determined from the organic phosphorus (phospholipid) concentration.18 The percentage of [14C]-MA distribution into the liposomes (D) was determined using the following equation: D ) 100 × Z/W

(1)

where Z is the ratio between the [14C]-MA (dpm/mL) and the phospholipid concentration (mM) in the liposome fraction eluted in the Sephadex G-50 mini spin column void volume, and W is the ratio between the [14C]-MA (dpm) and the phospholipid concentration (mM) in the original liposome dispersion. To determine the magnitude of the transmembrane liposome pH gradient, liposomes (100-nm HSPC/Chol 60/40 mol/mol) were prepared in 300 mM citrate buffers with different pH values (3, 4, 5, and 6). All liposomes were dialyzed against 10 mM histidine buffer with 10% sucrose, pH 6.7, to obtain four liposome preparations differing in transmembrane pH gradient (∆pHlip./med.). The distribution of [14C]-MA into each of these liposome preparations was determined as described above. A calibration curve that shows the dependence of the [14C]-MA distribution on the difference between the internal/ external liposome pH (∆pH) was obtained. The internal liposomal pH before and after TMN loading was calculated using the equation derived from this calibration curve: D ) -25.22X + 167.07

(2)

where X is the internal pH, and D is the percentage of [14C]-MA distribution into liposomes (dpm/mL) per millimolar concentration of phospholipid. The magnitude of the transmembrane liposome pH gradient (∆pH) is given by ∆pH ) 6.7 - X

(3)

where 6.7 is the pH of the external medium. (13) Stein, W. D. Transport and Diffusion Across Cell Membranes; Academic Press: Orlando, FL, 1986; Chapter 1. (14) Clerc, S.; Barenholz, Y. A quantitative model for using acridine orange as a transmembrane pH gradient probe. Anal. Biochem. 1998, 259 (1), 104-11. (15) Samuni, A. M.; Barenholz, Y. Stable nitroxide radicals protect lipid acyl chains from radiation damage. Free Radical Biol. Med. 1997, 22 (7), 1165-74. (16) Haran, G.; Cohen, R.; Bar, L. K.; Barenholz, Y. Transmembrane ammonium sulfate gradients in liposomes produce efficient and stable entrapment of amphipathic weak bases. Biochim. Biophys. Acta 1993, 1151 (2), 201-15. (17) Barenholz, Y.; Amselem, S. Quality control assays in the development and clinical use of liposome-based formulations. Liposome Technology, 2nd ed.; Gregoriadis, G., Ed.; Liposome Preparation and Related Techniques; CRC Press: Boca Raton, FL, 1993; Vol. I, pp 527-616.

(18) Shmeeda, H.; Even-Chen, S.; Honen, R.; Cohen, R.; Weintraub, C.; Barenholz, Y. Enzymatic assays for quality control and pharmacokinetics of liposome formulations: comparison with nonenzymatic conventional methodologies. Methods Enzymol. 2003, 367, 272-92. (19) Padan, E.; Schuldiner, S. Energy transduction in the photosynthetic membranes of the cyanobacterium (blue-green alga) Plectonema boryanum. J. Biol. Chem. 1978, 253 (9), 3281-6. (20) Dos Santos, N.; Cox, K. A.; McKenzie, C. A.; van Baarda, F.; Gallagher, R. C.; Karlsson, G.; Edwards, K.; Mayer, L. D.; Allen, C.; Bally, M. B. pH gradient loading of anthracyclines into cholesterol-free liposomes: enhancing drug loading rates through use of ethanol. Biochim. Biophys. Acta 2004, 1661 (1), 47-60.

1940 Langmuir, Vol. 23, No. 4, 2007 2.2.2.6. Determination of the Magnitude of Transmembrane Ammonium Gradient Formation by Acridine Orange (AO) Distribution between the Intraliposome Aqueous Phase and the Extraliposome Medium. The magnitude of the transmembrane ammonium sulfate gradient before TMN loading was determined by the AO distribution method.14,16 Briefly, 10 µL of 0.1 µM AO was added to the cuvette containing 3 mL of magnetically stirred buffer, then 10 µL of the liposomes (0.5-1.0 µmol of phospholipid) was added to the cuvette, and the decrease of fluorescence intensity at 525 nm (excitation 490 nm) due to AO distribution into the liposomes was monitored continuously using a Perkin-Elmer LS50B luminescence spectrofluorimeter.14,16 2.2.2.7. Effect of Ammonium Sulfate Gradient Magnitude on [14C]-MA and AO Distribution into Liposomes. The effect of the ammonium sulfate gradient on [14C]-MA and AO distribution into liposomes was studied by using 100-nm HSPC/Chol 60/40 (mol/ mol) liposomes prepared in 250 mM ammonium sulfate. The liposomes were dialyzed against 10 mM histidine buffer, pH 6.7, with 10% sucrose, containing 250, 25, 2.5, 0.25, and 0 mM ammonium sulfate, to obtain five liposomal preparations differing in ammonium sulfate gradient. All liposome preparations were then assayed for [14C]-MA and AO distribution, as described above. 2.2.2.8. Liposome Remote Loading with TMN. TMN was dissolved in 20% ethanol (in water) to a final concentration of 25 mM. A 0.8 mL portion of this TMN solution was added to 10 mL of a liposome suspension having a phospholipid concentration of 50 mM. Liposomes having a transmembrane ammonium sulfate gradient were used. The final solution contained 2 mM TMN and 1.4% ethanol. Remote loading was performed above the Tm of the liposome-forming PC. The loading process was terminated at the desired time by removal of unencapsulated TMN using the dialysis method, or, alternatively, using the cation exchanger Dowex 50 W X-4.16 Loading efficiency was determined by cyclic voltammetry (CV) and EPR, as described below. Liposome phospholipid concentration post-loading was determined using the modified Bartlett method.18 2.2.2.9. TMN Quantification. 2.2.2.9.1. EPR Measurements. We employed EPR using a JES-RE3X EPR spectrometer (JEOL Co., Japan) to quantify TMN concentration and study its physical state in the liposomes and after its release from the liposomes. Samples were injected by syringe into a gas-permeable Teflon capillary tube of 0.81 mm i.d. and a 0.05-mm wall thickness (Zeus Industrial Products, Raritan, NJ). The capillary tube was inserted into a 2.5mm-i.d. quartz tube open at both ends and placed in the EPR cavity. EPR spectra were recorded with the center field set at 329 mT, a 100 kHz modulation frequency, a 0.1 mT modulation amplitude, and a nonsaturating microwave power. Just before EPR measurements, loaded liposomes were diluted with 0.15 M NaCl to reach the suitable TMN concentration range (0.02-0.1 mM). The experiment was carried out under air, at room temperature. The EPR spectra of TMN loaded into liposomes having a transmembrane ammonium sulfate gradient are modified and do not represent TMN concentration; therefore, to measure the total TMN in liposomes in vitro and in animal plasma and tissues after liposome injection into animals, liposomes have to be solubilized by detergent so that TMN is released and becomes molecularly dispersed. Briefly, samples were homogenized in a Polytron homogenizer (Kinematica, Luzern, Switzerland) in 2% Triton X-100 (1:2, organ lipids/Triton X-100 mol/mol solution), followed by several cooling and heating cycles. Once the liposomes were solubilized, we found that, in plasma and tissues spiked with SSL-TMN, there is complete TMN release according to the EPR measurements. For determination of the total concentration of nitroxide + hydroxylamine, potassium ferricyanide at a final concentration of 2-3 mM (depending on the tested tissue) was added to all the samples (liposome extracts, plasma, and organ homogenates) in order to oxidize all the hydroxylamine to the nitroxide. 2.2.2.9.2. CV Measurements. Cyclic voltammograms were performed between -200 and 1300 mV. Measurements were carried out in PBS, pH 7.4. A three-electrode system was used throughout the study. The working electrode was a glassy carbon disk (BAS MF-2012, Bioanalytical Systems, W. Lafayette, IN), 3.3 mm in

Wasserman et al.

Figure 1. Comparison of the cytotoxicities of TMN and Tempol to MCF-7 human breast carcinoma cells. The cells were exposed to nitroxides for 72 h and then stained with methylene blue. diameter. The auxiliary electrode was a platinum wire, and the reference electrode was Ag/AgCl (BAS). The working electrode was polished before each measurement using a polishing kit (BAS PK-1).21 Just before CV measurements, the samples were diluted with buffer to the optimal TMN concentration range (0.05-0.2 mM). The experiment was carried out under air, at room temperature. 2.2.2.9.3. Effect of Nigericin and Nonactin on TMN Encapsulation by an Ammonium Ion Gradient. For measuring the ionophore effect, 0.6 µmol of liposome phospholipids was diluted in 3 mL (final volume) of 150 mM KCl, and 10 µL of either nigericin (final concn 5 µM) or nonactin (final concn 4 µM) was added. The cuvette was magnetically stirred above the Tm of the liposome-forming phospholipids (23 °C for EPC- and 60 °C for HSPC-based liposomes). To determine the effect of the ionophore on the liposome transmembrane ammonium ion and/or proton gradients before and after TMN loading into the liposomes, we applied the [14C]-MA and AO distribution assays as described above. To get complete TMN release, the liposomes were solubilized by Triton X-100 [final concentration 1% (w/v)]. Under these conditions, the Triton concentration was above the sum of the Triton critical micelle concentration + Triton concentration at 2 times the phospholipid concentration, the level needed for complete liposome solubilization.17 2.2.2.10. Osmolality Measurements. Highly purified water (resistance 18.2 MΩ) was obtained using Water ProPlus (Labconco, Kansas City, MO), referred to here as “pure water” solutions. Osmolality was measured using a 5500 Vapor Pressure-Osmometer (Wescor, Inc., Logan, UT). Calibration curves of three amphipathic weak bases (TMN, AO, and bupivacaine (BUP)) were used to describe the relationship between osmolality (in the range of 10-40 mOsM) and their concentration. On the basis of these calibration curves, we decided to use a concentration of 25 mM for all three amphipathic weak bases, as it gave reliable osmolality measurements.

3. Results 3.1. The “Justification” for Liposome TMN Formulation. 3.1.1. EValuation of TMN Cytotoxicity and PS Externalization in Cells in Culture. Figure 1 shows a comparison of the doseresponse curves of the cytotoxicity of TMN and Tempol to MCF-7 cells in culture after a 72-h exposure to the nitroxides. The IC50 of TMN is 210 µM, and the IC50 of Tempol is 320 µM. The same cell line was used to investigate whether and to what extent PS externalization is involved in the mechanism of cell killing by TMN. Untreated cells and TMN-treated cells (24-h exposure) were stained with merocyanine-540, and analyzed by FACS as described in the Materials and Methods section (subsection 2.2.1.5). As seen from Figure 2, most of the cells (77%) after TMN treatment were fluorescently labeled, compared to only 14% fluorescently labeled cells without TMN treatment. (21) Kohen, R.; Fanberstein, D.; Tirosh, O. Reducing equivalents in the aging process. Arch. Gerontol. Geriatr. 1997, 24 (2), 103-23.

The Antioxidant Tempamine

Langmuir, Vol. 23, No. 4, 2007 1941

Figure 2. TMN-induced apoptosis. Cells were treated with TMN (1 mM) for 24 h, trypsinized, stained with merocyanine-540, and then analyzed by flow cytometry. The M1 area indicates fluorescently labeled apoptotic cells.

Figure 3. TMN protection in PC12 neurons against damage by 1.5 mM MPP+. TMN was added at the concentrations specified in the figure. Cell death was evaluated by measuring the leakage of LDH into the medium. The cytotoxicity of cells growing with MPP+ without TMN are used as the 100% cytotoxicity.

Table 1. Effect of TMN Concentrations on the IC50 of DOX (nM) in DOX-sensitive Cells Following TMN Addition TMN (µM) MCF-7 M-109S

0

50

100

200

487 ( 32 60 ( 4.0

475 ( 38

67 ( 5.8 27 ( 1.7

55 ( 4.1 18 ( 1.1

This supports a previously reported finding6 that piperidine nitroxides kill cancer cells via apoptosis induction. 3.1.2. Synergism between TMN and DOX In Vitro. The ability of antioxidants to regulate the antitumor effect of chemotherapy agents has been recently explored.22 The piperidine nitroxides Tempol and Tempo were shown to enhance chemotherapyinduced apoptosis.6,7 TMN alone possesses anticancer activity in vitro that is at least as good as that of Tempo, but the concentration needed to achieve a cytotoxic effect of TMN is too high for in vivo application. However, TMN may be synergistic with chemotherapy treatment and improve the chemotherapeutic agent by shifting its dose response curve downward. Therefore we studied the degree of synergism between TMN and DOX. For this, the enhancement of DOX cytotoxicity by TMN was evaluated. Low noncytotoxic TMN concentrations in the range of 100-200 µM were added to low DOX concentrations in the range of 10-60 nM and tested on DOX-sensitive (M-109S) and DOX-resistant (M-109R) cell lines. MCF-7 cells, which are more sensitive to TMN (100 µM caused 75% growth inhibition) but are less sensitive to DOX than are M-109S cells, were also studied. In the latter cells, combinations of 50-100 µM TMN with different DOX concentrations were tested. The results are summarized below: MCF-7. The IC50 value of DOX was significantly (by almost 1 order of magnitude) decreased in the presence of 100 µM TMN (Table 1), but did not change in the presence of 50 µM TMN. (22) Lamson, D. W.; Brignall, M. S. Antioxidants in cancer therapy; their actions and interactions with oncologic therapies. Altern. Med. ReV. 1999, 4 (5), 304-29.

Figure 4. Survival curves of mice inoculated with C-26 tumor cells treated with various drug combinations. f-DOX ) free DOX. After 23 days, mice survival was control: 0 out of 5; f-DOX: 2 out of 9; Doxil: 1 out of 9; f-DOX+TMN: 1 out of 9; Doxil+SSL-TMN: 7 out of 9; and SSL-TMN: 0 out of 9.

M-109S. Despite the absence of a cytotoxic effect of low TMN concentrations on this cell line, the addition of 100 and 200 µM TMN decreased to 50% and 25% the observed IC50 of DOX, respectively (Table 1). Synergism was assessed using the equation described in section 2.2.1.6 and by Modrak et al.23 A CI value of 0.83 was obtained, indicating synergism between TMN and DOX. 3.1.3. In ViVo Synergism between TMN (as SSL-TMN) and DOX (as Doxil). To evaluate whether synergism between DOX and TMN also occurs in vivo, we used an animal model of BALB/c mice bearing subcutaneous implants of C26 mouse colon carcinoma cells. Briefly, 1 × 106 mouse colon carcinoma (C26) cells were injected intraperitoneally into mice. The anti-cancer efficacies of SSL-TMN, DOX, and Doxil alone and in various combinations were compared. As shown in Figure 4 and Table 2, combined treatment of Doxil and SSL-TMN increased survival time relative to all other treated groups. On day 21 after carcinoma cell inoculation, in the Doxil+SSL-TMN group, all mice were alive compared with the Doxil alone group, where only five out (23) Modrak, D. E.; Cardillo, T. M.; Newsome, G. A.; Goldenberg, D. M.; Gold, D. V. Synergistic interaction between sphingomyelin and gemcitabine potentiates ceramide-mediated apoptosis in pancreatic cancer. Cancer Res. 2004, 64 (22), 8405-10.

1942 Langmuir, Vol. 23, No. 4, 2007

Wasserman et al.

Table 2. Mean Survival Time and Percentage Increase in Life Span (%ILS) of C-26-Tumor-Inoculated Mice under Different Treatments treatment group

no. of animals

mean survival timea

%ILS

control free DOX free TMN Doxil free DOX+free TMN Doxil+SSL-TMN

5 9 9 9 9 9

17.6 19.1 15.7 21.2 19 26.9

8.5 11.9 20.4 8 52.8

Table 3. Distribution of TMN between the n-Octanol/Aqueous Phase (Kp): Effect of pH and Ammonium Sulfate Concentration ammonium sulfate, mM

TMN, µmol

oct/aq phase Kp, pH 10.6

oct/aq phase Kp, pH 7

oct/aq phase Kp, pH 4

0 20 20 150 150 400 400

20 2 20 2 20 2 20

2.331 2.418 2.386 2.184 3.969 3.005 4.569

0.278 0.048 0.116 0.055 0.038 0.040 0.034

0.111 0.031 0.074 0.024 0.049 0.028 0.046

of nine mice (55%) had survived. The most significant difference was observed on day 23 after tumor inoculation (16 days after beginning of the treatment): in the Doxil+SSL-TMN group, seven mice (78%) had survived compared with one mouse (11%) in the Doxil alone group, and one mouse in the free DOX-TMN group; in the control group, the SSL-TMN alone group, and the free DOX group, all mice were dead at this time point. An increase of 52% in the life span was observed in the Doxil+SSL-TMN combination treatment compared with 20% of Doxil alone (Table 2). These and the in vitro results encouraged us to initiate the development of the SSL-TMN formulation. 3.1.4. TMN ProtectiVe Effects in PC12 Neuronal Cells Exposed to Damage by MPP+. Results shown in Figure 3 demonstrate that concentrations of TMN protect PC12 neurons from oxidative damage by 1.5 mM MPP+ in a bell-shaped manner, with 100 µM being most protective. The bell-shaped curve suggests that, at higher concentrations, TMN is toxic to the cells. However, such high concentrations are irrelevant for in vivo treatment, since with liposome delivery tissue concentration will never reach such high TMN concentrations. 3.2. Development of SSL-TMN Formulation with Focus on Optimization of TMN Loading and Release. 3.2.1. N-Octanol/Aqueous Phase Partition Coefficient (Kp) of TMN. Kp (oct/aq) is a parameter describing the lipophilicity (or hydrophilicity) of a substance. Although not directly related to the lipid bilayer/aqueous phase, it enables one to predict simple diffusion across lipid bilayers.13 The oct/aq phase partition coefficients (Kp) under various conditions of pH and in the absence and presence of ammonium sulfate in the various systems are shown in Table 3. At acidic and neutral pHs, Kp values were below 0.3 and were unaltered by high ammonium sulfate concentrations. At alkaline pH, much more TMN partitioned into the octanol phase, and the Kp increased to >2.2. This effect is further increased with increasing ammonium sulfate and TMN concentrations. This implies that, at acidic and neutral pHs, TMN, being highly charged, forms a salt with the sulfate ion, since otherwise the ammonium sulfate ions would shift the amphipathic molecules into the less polar phase.24 These data also show that, under nonalkaline conditions (which is the situation inside liposomes remote loaded by an (24) Bauer, V.; Bauer, F. Reactive oxygen species as mediators of tissue protection and injury. Gen. Physiol. Biophys. 1999, 18, 7-14.

ammonium sulfate transmembrane gradient), TMN concentrates mainly in the intraliposome aqueous phase and is not associated with the liposome membrane, as expected from the fact that TMN is a weak base and, at acidic and neutral pHs, it is highly charged (protonated). Other piperidine nitroxides, which have a similar chemical structure but lack a charged moiety (Tempo, Tempol), have a much higher Kp (10-100 times higher).15 This high Kp value of the noncharged nitroxides suggests that it will be difficult to form liposomes that stably include Tempo and Tempol inside the intraliposome aqueous phase. This is also supported by the results describing the liposome lipid bilayer aqueous phase Kp below. 3.2.2. Lipid Bilayer/Aqueous Phase Partition Coefficient (Klip/aq). The level of TMN distributed into the liposome membrane was below the detection limit. In this respect, TMN differs from Tempo and Tempol, which distribute at measurable levels to the liposome lipid bilayer.25 These differences are explained by the presence of an ionizable primary amino group in TMN and the lack of charged groups in Tempo and Tempol. Therefore, it is expected that TMN is better retained by the liposomes than Tempo or Tempol. 3.2.3. Determination of Percent Encapsulation Using EPR. TMN is a stable radical, and, therefore, one of the best methods for its characterization and quantification is EPR.15 EPR can also be used to characterize the physical state of the nitroxide and to differentiate between TMN dispersed as separate molecules or as aggregates in aqueous phase, as well as in the lipid bilayer. To determine the percent of encapsulation, we took advantage of the ability to release all TMN from liposomes by nigericindependent collapse of the ammonium sulfate gradient. The determination is based on three steps: (1) Measuring total TMN in post-loading liposome preparation (TMNmix). (2) Measuring TMN in post-loading liposome preparation in the presence of potassium ferricyanide, an EPR broadening agent that eliminates (quenches) the signal of free (nonliposomal) TMN. The remaining signal is of TMN in liposomes and is referred to as TMNliposome(quenched). This spectrum is very broad, as TMN concentration inside the liposomes is very high, leading to quenching of its EPR signal due to spin interaction between the TMN molecules, which are close to one another. (3) Measuring total TMN after its release from liposomes by nigericin (TMNnigericin). This signal is identical to that of total TMN used for loading (TMNnigericin ) TMNtotal) and is completely dequenched. TMNliposome(notquenched) represents the signal of liposomal TMN if the ammonium sulfate gradient would be collapsed and all the TMN becomes free. Calculation of the percent encapsulation and of the quenching factor:

TMNfree ) TMNmix - TMNliposomes(quenched)

(1)

TMNliposomes(not quenched) ) TMNnigericin - TMNfree (2) Percent encapsulation ) 100 × TMNliposome(not quenched)/TMNnigericin (3)

(25) Samuni, A. M.; Barenholz, Y. Site-activity relationship of nitroxide radical’s antioxidative effect. Free Radical Biol. Med. 2003, 34 (2), 177-85.

The Antioxidant Tempamine

Langmuir, Vol. 23, No. 4, 2007 1943

Quenching factor ) TMNliposome(not quenched)/TMNliposome(quenched) (4) This experiment proved that there is no TMN loss during the loading and release procedures. Nigericin releases all the TMN from the liposomes, as indicated by the fact that the TMN signal after the addition of nigericin and the signal of free TMN of the same concentration used during loading are identical. Using eq 4, we calculated the quenching factor to be ∼3-fold. 3.2.4. Determination of Percent Encapsulation Using CV. CV is a direct method to quantify TMN and to evaluate the loading efficiency of TMN, without the need to separate the liposomes from the external medium. In general, this method is used with compounds that possess a stable redox potential, and measures the ability of these compounds to donate or accept electrons to or from the working electrode, producing an electric current.21 The parameters that can be calculated from the cyclic voltammogram are the current produced in the systemsanodic current (AC) or cathodic current (CC)s and the peak oxidation or reduction potential of the compound. This current correlates with the concentration of the reducing/oxidating equivalents in each wave. A decrease in current indicates a decrease in concentration, while an increase in current indicates an increase in concentration. The peak electrical potential indicates the ability of the compound to donate or accept electrons. An analyte has to be in direct contact with the working electrode in order to be determined; this method cannot measure an analyte encapsulated in the intraliposome aqueous phase but only the free unencapsulated analyte. Thus, when the liposomes are intact, only free TMN molecules in the extraliposomal aqueous phase are determined. Thus, a comparison between the current magnitude of TMN before loading, after loading, and of the latter after release of TMN by ionophore or detergent solubilization enables the user to determine the percentage of TMN encapsulation. In addition, because this assay is based on the ability of the analyte to donate or accept electrons, it can be considered as a functional assay that describes the integrity and potency of the analyte throughout various processes such as loading, and release, both upon storage and when present in biological milieux such as plasma or tissues. 3.2.5. Demonstrating that TMN Remote Loading Is DriVen by the Ammonium Ion Gradient. On the basis of our previous experience with Doxil development,16,26 we used a transmembrane intraliposome high/extraliposome low ammonium ion gradient to load the amphipathic weak base TMN into SSL. During the loading process, EPR and CV measurements were made at 5, 10, 30, and 60 min. Already after 5 min of loading, the TMN signal with both the EPR and CV assays changed dramatically (Figures 5 and 6). The CV spectra (Figure 5) suggest loading, while the EPR spectra (Figure 6A) suggest TMN precipitation. No change in CV or EPR signal occurs with liposomes that lack an ammonium sulfate gradient, demonstrating that ammonium sulfate is the driving force for the TMN loading. 3.2.6. InVestigation of the Mechanism of TMN Loading. Two ionophores were used: nigericin, which exchanges K+ for H+ (Figure 8B), and nonactin, which exchanges NH4+ for H+ (Figure 8C). Table 5 demonstrates that nigericin and nonactin in the presence of K+ greatly reduces the uptake of AO into liposomes having a transmembrane ammonium gradient. The results presented above indicate that TMN loading into the liposomes is driven (26) Barenholz, Y. Liposome application: problems and prospects. Curr. Opin. Colloid Interface Sci. 2001, 6, 66-77.

Figure 5. Cyclic voltammogram of TMN before (B) and after (A) its remote loading into liposomes. Anodic and cathodic waves disappeared completely after the loading was completed. Inset: Calibration curve of TMN, based on CC measurements.

Figure 6. EPR signal of 0.5 mM TMN before and after encapsulation into EPC liposomes by means of (A) 150 mM ammonium sulfate gradient and (B) 150 mM ammonium glucuronate gradient.

by an ammonium ion gradient (Figure 8A), as was demonstrated for DOX loading.16 The effect of TMN loading on the transmembrane proton gradient was studied by comparing the distribution of [14C]-MA into liposomes before and after TMN loading. AO distribution cannot be used because of the quenching of AO fluorescence by nitroxides. Table 4 shows the transmembrane ammonium gradient determination by [14C]-MA with two different liposomal formulations before and after TMN remote loading. AO distribution into the liposomes before loading was determined to confirm [14C]-MA distribution. Table 6 demonstrates that only part of the proton gradient was utilized by TMN loading according to the percentage of [14C]-MA encapsulation into liposomes.

1944 Langmuir, Vol. 23, No. 4, 2007

Wasserman et al. Table 4. [14C]-MA and AO Distribution into Liposomes Containing 250 mM (NH4)2SO4 and Having Different (NH4)2SO4 Concentrations in the External Medium

Figure 7. Stability of TMN remote loading into EPC (1200 ( 200 nm) multilamellar liposomes at 37 °C.

% of % of [C14]-MA AO in in liposomes liposomes

liposomes

extra liposome mediuma

1 2 3 4 5 6

250 mM (NH4)2SO4 buffer +25 mM (NH4)2SO4 buffer +2.5 mM (NH4)2SO4 buffer +0.25 mM (NH4)2SO4 buffer + 0 mM (NH4)2SO4 buffer +6.5 mM nigericin + K+

a

8.3 ( 0.5 13.1 ( 0.5 22.2 ( 0.5 82.3 ( 3.5 87.5 ( 4.2 3.0 ( 0.7

0.69 5.8 7.3 49.5 96.4 NH4 glucuronate. The difference between the leakage rates may be attributed to the difference in the extent of precipitation of TMN salts with the different anions of the ammonium salts. That is, the higher the fraction precipitated, the slower the leakage rate. The differences between the anions cannot be explained by the differences between anion permeability coefficients, which are in the order of SO42- ∼glucuronate ,Cl-, which further supports the idea that the level of TMN aggregation controls (at least partially) the TMN leakage rate. The magnitude of the NH4+ gradient in the tested range (100-400 mM) has only a small effect on the stability of TMN loading, as the rate of leakage for each of the three salts was similar for liposomes in which gradients were based on either 100 or 400 mM ammonium salt. 3.2.8. Effect of Ammonium Sulfate on the Aggregation of Amphipathic Weak Bases: Comparison between Three Amphipathic Weak Bases. The ability of three amphipathic weak basess

The Antioxidant Tempamine

Langmuir, Vol. 23, No. 4, 2007 1945 Table 6. Percentage of [C14]-MA and AO Distribution into EPC- and HSPC-Based SSL-TMN Formulation with Transmembrane Ammonium Gradient liposome composition

% of [C14]-MA loading

% of AO loading

EPC/PEG-DSPE/Chol EPC/PEG-DSPE/Chol-TMN HSPC/PEG-DSPE/Chol HSPC/PEG-DSPE/Chol-TMN

98.5 57 97 63

90 ND 98 ND

ND (not done) due to TMN being a quencher of AO fluorescence.

Figure 9. Extent of aggregation of 25 mM TMN in the presence of different ammonium salts.

TMN, AO, and bupivacaine (BUP)sto precipitate in the presence of ammonium sulfate was evaluated. For this we used osmotic pressure measurements. Osmotic pressure is a colligative property, and it is proportional to the actual concentration of particles in the solution, irrespective of particle molecular mass and charge. If the osmotic pressure of a solution that has more than one compound is lower than the sum of the osmotic pressures of each solute separately, it means that the actual concentration of solute particles decreased as a result of association, aggregation, or the formation of microprecipitates.27 Thus, osmotic pressure measurement is a direct and appropriate way to determine molecular aggregation, even under conditions when no precipitate is visible to the eye. The 25 mM solutions had the following osmolalities: TMN ) 21 mOsm, BUP HCl ) 32 mOsm, and AO ) 48 mOsm. When two compounds are mixed and the osmotic activity of the mixture measured is smaller than the sum of the osmotic pressures of the separate components (under conditions in which no chemical reaction occurs between the components), aggregation and/or precipitation is suggested; the lower the percentage of the expected value, the higher the percentage of aggregation. Of the three, only TMN undergoes major aggregation, even at 40 mM ammonium sulfate (Figure 9). 3.2.9. Comparison of the State of Aggregation of Loaded TMN in the Intraliposome Aqueous Phase Containing Various Ammonium Salts: Correlation to TMN Leakage Rate from Liposomes. Above, we demonstrated that loading stability (TMN retention in liposomes) is affected by the anion that formed the ammonium salt used to obtain the transmembrane ammonium gradient. We suggested that this may be related to the level of aggregation of TMN with that anion in the intraliposome aqueous phase. The osmotic pressure measurements of TMN in the presence of the different ammonium salts presented in Figure 9 support this assumption. To directly test the state of aggregation of TMN in the intraliposome aqueous phase, liposomes were prepared in either ammonium sulfate or ammonium glucuronate. The transmembrane ammonium ion gradient was formed, and TMN was loaded into the liposomes. Comparison of the EPR spectra of these two TMN-loaded liposomes showed major broadening only when the gradient of ammonium sulfate was used, and the CV spectrum almost disappeared. When TMN was loaded into ammonium glucuronate liposomes, the CV peak in the spectrum was reduced to the same level as that in the case of TMN loaded into ammonium sulfate liposomes, suggesting similar TMN loading with both ammonium salts. However, no such similarity occurred for the EPR spectrum, as the TMN loaded into ammonium glucuronate (Figure 6B) did (27) Berns, D. S. Protein aggregation in phycocyaninsosmotic pressure studies. Biochem. Biophys. Res. Commun. 1970, 38 (1), 65-73.

not show the broadening that occurs when TMN is remote loaded into liposomes having a transmembrane ammonium sulfate gradient (Figure 6A). This indicates that no precipitation of the loaded TMN occurs in ammonium glucuronate-containing liposomes, in agreement with the above-described osmotic pressure measurements. 3.3. Stability of TMN Loading: The Interplay between Lipid Composition, Storage Time, and Temperature and the Effect of Plasma. As expected, the stability of loading upon storage in 0.15 M NaCl was affected by the liposome-forming PC. In general, loading stability was higher for a liposomeforming PC with a high Tm (such as HSPC, Tm ) 52.0 °C) than for a liposome-forming PC with a low Tm (such as EPC, Tm ) ∼5 °C). After 2 months at 4 °C, the percent retention was reduced to 85% for HSPC/Chol/2000PEG-DSPE (∼100-nm LUV), and to 53% for EPC/Chol/2000PEG-DSPE (∼100-nm LUV). When stability in 0.15 M NaCl was compared at 37 °C after 15 h of incubation, the difference was similar: 85% retention for HSPCbased liposomes versus 44% for EPC-based liposomes. Therefore, it is clear that TMN is released from the liposomes in a temperature-dependent manner. These results demonstrate that the TMN release is highly dependent on the lipids forming the liposomes. Liposomes having the fluid PC EPC as their liposomeforming lipid release TMN much faster than liposomes in which the solid HSPC serves as the liposome-forming lipid (Table 7). EPC-based liposomes released TMN even at 4 °C, while HSPCbased liposomes did not. Although the high resistance to release is important for long shelf-life stability, it may reduce the therapeutic efficacy due to a slower release rate at 37 °C.28 Concerning the effect of plasma on the stability of TMN loading, Table 7 demonstrates that there is a large difference between HSPC-based (71%) and EPC-based (8%) liposomes.

4. Discussion Nitroxides, because of their activity as proapoptotic agents29 and efficacious antioxidants,30 may serve as potent drugs against tumors and inflammation. One of the main requirements to achieve therapeutic efficacy in such diseases is that the drug should be delivered to the disease site in sufficient quantity and be released there in a way that enables achieving a long enough exposure of the diseased cells to the drug. Such a goal can be achieved by the encapsulation of an antioxidant or anti-inflammatory agent in liposomes, and specifically in SSLs less than or equal to 100 nm, which can reach the disease site and release the drug (28) Barenholz, A.; Fishel, F.; Yakir, E.; Gatt, S.; Barenholz, Y.; Bercovier, H. Liposomes enhance bioremediation of oil-contaminated soil. J. Liposome Res. 2003, 13 (2), 173-86. (29) Suy, S.; Mitchell, J. B.; Ehleiter, D.; Haimovitz-Friedman, A.; Kasid, U. Nitroxides tempol and tempo induce divergent signal transduction pathways in MDA-MB 231 breast cancer cells. J. Biol. Chem. 1998, 273 (28), 17871-8. (30) Hahn, S. M.; Tochner, Z.; Krishna, C. M.; Glass, J.; Wilson, L.; Samuni, A.; Sprague, M.; Venzon, D.; Glatstein, E.; Mitchell, J. B. Tempol, a stable free radical, is a novel murine radiation protector. Cancer Res. 1992, 52 (7), 1750-3.

1946 Langmuir, Vol. 23, No. 4, 2007

Wasserman et al.

Table 7. Comparison of TMN Retention in EPC- and HSPC-Based SSL Formulations: Effect of Temperature

liposome composition

media

incubation temperature (°C)

EPC/Chol/2000PEG-DSPE HSPC/Chol/2000PEG-DSPE EPC/Chol/2000PEG-DSPE HSPC/Chol/2000PEG-DSPE EPC/Chol/2000PEG-DSPE HSPC/Chol/2000PEG-DSPE

0.15 M NaCl (pH 5.5) 0.15 M NaCl 0.15 M NaCl 0.15 M NaCl plasma (pH 7.4) plasma

4 4 37 37 37 37

there.31 However, the small size of the liposomes imposes a restriction on the amount of drug that can be passively loaded into the liposomes. This is the reason we selected TMN, a nitroxide that is an amphipathic weak base that can be remote loaded efficiently into the e100-nm SSL via an ammonium ion gradient, so a high drug-to-lipid ratio can be achieved.26,28 Before developing the formulation, we evaluated the biological activity of TMN in a biologically relevant model and compared it to the well-characterized nitroxide Tempol. We proved that TMN is at least equivalent, or, in some cases, superior to Tempol, and therefore we initiated the process of the development and characterization of the SSL-TMN formulation and demonstrated its utility in vivo. We examined the cytotoxic activity of TMN and TMN+DOX on different cancer cell lines. TMN alone possesses somewhat better cytotoxicity toward the MCF-7 cells than the most-studied nitroxide, Tempo. Flow cytometry analysis of TMN-treated cells revealed that TMN induces PS externalization in tumor cells, which suggests the involvement of apoptosis and/or necrosis in the cell death. Although we did not study whether the reason for death is apoptosis and/or necrosis, previous results with the effects of Tempol6 suggest the involvement of apoptosis.6 Our results also show that, while the concentration of TMN alone is not sufficient for therapeutic efficacy, its synergism with DOX is encouraging and worth in-depth evaluation, as relatively low TMN concentrations are needed to increase cell sensitivity to DOX in both DOX-resistant (M-109R) and sensitive (M-109S) cells. Combined treatment of cells with TMN and DOX revealed a significant decrease in the IC50 of DOX when cells were exposed to a low, noncytotoxic concentration (100 µM) of TMN. A 2-fold, but still nontoxic, concentration of TMN partly eliminated M-109R cell resistance to DOX, although M-109R cell sensitivity to DOX was still lower than that of M-109S. We also performed a preliminary study to evaluate whether such synergism is also relevant to animal models. For this, we examined the anti-cancer efficacies of SSL-TMN, DOX, and Doxil, alone and in various combinations, in a mice colon carcinoma (C26) tumor model. These experiments indicate that, while SSL-TMN alone did not improve mice survival, the combined therapeutic activity of Doxil+SSL-TMN was much better than additive. As for protection against oxidative damage, we used PC12 neurons to which oxidative damage by MPP+ was introduced. TMN displayed protectivity in a dose-dependent manner in the range of 0.1-100 µM, while at high doses of 500-1000 µM (irrelevant to the situation of delivery by SSL) it was toxic to the cells. Activity against oxidative damage in vivo was not studied here, but in another, ongoing study (Kizelsztein et al., in (31) Grant, G. J.; Barenholz, Y.; Piskoun, B.; Bansinath, M.; Turndorf, H.; Bolotin, E. M. DRV liposomal bupivacaine: preparation, characterization, and in vivo evaluation in mice. Pharm. Res. 2001, 18 (3), 336-43.

incubation time

% of retention

60 days 60 days 15 h 15 h 15 h 15 h

53 85 44 85 8 71

preparation), we demonstrate that SSL-TMN is efficacious against an animal model of multiple sclerosis, which involves oxidative damage.32 TMN, a chemically stable free radical that can serve as a powerful antioxidant,4,15 due to its primary amino group, is an amphipathic weak base, which meets the requirement for remote loading by a transmembrane ammonium gradient.26 TMN’s low, pH-dependent oct/aq partition coefficient ensures that, at acidic and neutral pHs in the presence of an ammonium gradient, it will be stably encapsulated in the intraliposomal aqueous phase (and not in the membrane). This is also supported by the EPR spectra. Since ammonium sulfate at lower pH further reduces the TMN partition coefficient, it should further reduce drug-membrane interaction and drug release while increasing drug level in the intraliposomal aqueous phase. We used EPR and CV methods to determine the TMN concentration and to quantify the level of its encapsulation, while EPR also enabled us to follow the physical state of TMN in the intraliposomal aqueous phase, as the EPR method enables detection of encapsulated and free TMN separately, while the CV method detects only free TMN available to the electrode. Our results indicate that, during remote loading driven by a transmembrane ammonium sulfate gradient, EPR and CV signals of TMN change significantly (area under the peak and peak shape). These changes are attributed to encapsulation into the liposomes by means of the ammonium sulfate gradient, while the shape of the EPR peak suggests TMN precipitation in the intraliposomal aqueous phase. This is supported by the fact that TMN, which has a low molecular weight, coelutes with the liposomes at the G50-Sephadex or Sepharose 6B void volume, and remains associated with the liposomes after extensive dialysis. It is also supported by osmolality measurements of TMN in the presence of ammonium sulfate and other ammonium salts. This study shows that only for ammonium sulfate most of the TMN precipitates, while, with two other ammonium salts (ammonium chloride and ammonium glucuronate), TMN precipitates to a much smaller extent. The TMN release rate was not well correlated with the anion permeability coefficient, but was correlated with the degree of TMN precipitation, as determined from the osmolality and EPR results. The addition of the ionophores nigericin or nonactin in the presence of K+ collapses the ammonium ion gradient in liposomes, and restores the EPR and CV signals to their original size and shape, proving the direct role of the ammonium ion gradient in TMN loading. It also demonstrates (similarly to what was shown for Doxil26) that TMN aggregation is reversible, and TMN can became fully active once it is released.

5. Conclusions The development of the SSL-TMN formulation is justified by the efficacy of TMN as both a protective and an antitumor agent. (32) Lu, F.; Selak, M.; O’Connor, J.; Croul, S.; Lorenzana, C.; Butunoi, C.; Kalman, B. Oxidative damage to mitochondrial DNA and activity of mitochondrial enzymes in chronic active lesions of multiple sclerosis. J. Neurol. Sci. 2000, 177 (2), 95-103.

The Antioxidant Tempamine

Regarding this formulation, we show that remote loading of TMN into liposomes by a transmembrane ammonium ion gradient is an effective method of loading, achieving a high degree of stable encapsulation. We also demonstrate that the TMN release rate depends on the SSL lipid composition and the type of ammonium salt anion, with ammonium sulfate being the slowest TMN-releasing formulation. On the basis of the release rate measurements, TMN should be biologically available, and therefore SSL-TMN is an attractive candidate to treat major

Langmuir, Vol. 23, No. 4, 2007 1947

pathological conditions in which the intervention of such antioxidants is needed. Acknowledgment. This study was supported in part by the Belfer Foundation and by the Barenholz Fund. Alberto Gabizon is acknowledged for his advice on the animal tumor model studies. The help of S. Geller in editing the manuscript and of B. Levene in helping with its typing is acknowledged with pleasure. LA060218K