In Vitro and in Vivo Characterization of Doxorubicin and Vincristine

728 Vol. 10, 728–738, January 15, 2004 Clinical Cancer Research

In Vitro and in Vivo Characterization of Doxorubicin and Vincristine Coencapsulated within Liposomes through Use of Transition Metal Ion Complexation and pH Gradient Loading

Sheela A. Abraham,1,2 Cheryl McKenzie,1 studies evaluating doxorubicin and vincristine combinations Dana Masin,1 Rebecca Ng,1 Troy O. Harasym,4 analyzed using the Chou and Talalay median effect princi- Lawrence D. Mayer,1,3,4 and Marcel B. Bally1,2,4 ple. These data clearly indicated that simultaneous addition 1 of vincristine and doxorubicin resulted in pronounced an- Division of Medical Oncology, Department of Advanced tagonism. Therapeutics, BC Cancer Agency; 2Department of Pathology and Laboratory Medicine, Faculty of Medicine, and 3Faculty of Conclusion: These results emphasize that in vitro drug Pharmaceutical Sciences, University of British Columbia; and combination screens can be used to predict whether a co- 4Celator Technologies Incorporated, Vancouver, British Columbia, formulated drug combination will act in an antagonistic or Canada synergistic manner.

ABSTRACT INTRODUCTION Purpose: There is an opportunity to augment the ther- Combination treatment regimes must take into consider- apeutic potential of drug combinations through use of drug ation drug resistance mechanisms (1) and tumor heterogeneity delivery technology. This report summarizes data obtained (2). In practice, however, drug combinations usually take ad- using a novel liposomal formulation with coencapsulated vantage of nonoverlapping toxicities and unique mechanisms of doxorubicin and vincristine. The rationale for selecting these action. We believe that there is an opportunity to capture the drugs is due in part to the fact that liposomal formulations therapeutic potential of drug combinations through the use of of doxorubicin and vincristine are being separately evalu- drug delivery technology. This hypothesis is supported by three ated as components of drug combinations. lines of evidence: (a) appropriately designed drug delivery Experimental Design: Doxorubicin and vincristine were systems can improve drug delivery and efficacy for single coencapsulated into liposomes using two distinct methods of agents (Refs. 3–5; b) polymer, protein, and lipid-based drug drug loading. A manganese-based drug loading procedure, carriers have the capacity to coformulate two or more therapeu- which relies on drug complexation with a transition metal, tic agents (6–8); and (c) the pharmacokinetic behavior of the was used to encapsulate doxorubicin. Subsequently the iono- coformulated drugs will be dictated by the pharmacokinetic phore A23187 was added to induce formation of a pH gra- behavior of the drug carrier system used, thus offering the dient, which promoted vincristine encapsulation. potential to coordinate the plasma elimination and tissue distri- Results: Plasma elimination studies in mice indicated bution of the combined agents. that the drug:drug ratio before injection [4:1 doxorubicin: There are a multitude of approved chemotherapy combi- vincristine (wt:wt ratio)] changed to 20:1 at the 24-h time nations that could be used to test this principle, and this report point, indicative of more rapid release of vincristine from summarizes data obtained using a novel liposomal formulation the liposomes than doxorubicin. Efficacy studies completed with coencapsulated doxorubicin and vincristine. The rationale in MDA MB-435/LCC6 tumor-bearing mice suggested that for selecting these drugs is due in part to the fact that liposomal at the maximum tolerated dose, the coencapsulated formu- formulations of the individual agents have already proven to be lation was therapeutically no better than liposomal vincris- of clinical value in the treatment of patients with HIV-associated tine. This result was explained in part by in vitro cytotoxicity Kaposi sarcoma (9), ovarian cancer (10), breast cancer (11), and various hematological malignancies (12, 13), including relapsed non-Hodgkin’s lymphoma. Furthermore, these liposomal formulations of doxorubicin and vincristine are now being evalu- Received 7/31/03; revised 9/24/03; accepted 9/29/03. ated as components of drug combinations used routinely in the Grant support: Canadian Institutes for Health Research and Celator management of patients with cancer. Liposomal formulations of Technologies Incorporated. S. A. A. is a recipient of a Science Council doxorubicin, for example, have been evaluated as part of the of British Columbia GREAT scholarship and the William Armstrong vincristine, doxorubicin, and dexamethasone regimens used to Memorial Award. The costs of publication of this article were defrayed in part by the treat patients with multiple myeloma (14), and in the cyclophos- payment of page charges. This article must therefore be hereby marked phamide, doxorubicin, vincristine, and prednisone regimen for advertisement in accordance with 18 U.S.C. Section 1734 solely to treatment of aggressive non-Hodgkin’s lymphoma (15). Alter- indicate this fact. natively, this regimen has been modified by incorporation of a Requests for reprints: Marcel B. Bally, Department of Advanced Therapeutics, BC Cancer Agency, 601 West 10th Avenue, Vancouver, liposomal vincristine formulation for use in the treatment of a British Columbia V5Z 1L3, Canada. Phone: (604) 877-6098, extension similar patient population (16). A liposomal formulation of 3191; Fax: (604) 877-6011; E-mail: [email protected]. doxorubicin has been used in combination with cyclophospha-

mide for treatment of patients with metastatic breast cancer (17), Technologies, Vancouver, British Columbia, Canada) supple- and with cyclophosphamide and fluorouracil for first-line treat- mented with 10% fetal bovine serum from Hyclone Laboratories ment of patients with metastatic breast cancer (18). More re- (Logan, UT), 1% L-glutamine, and 1% penicillin and strepto- cently, liposomal formulations of doxorubicin have been tested mycin solution (Stem Cell Technologies, Vancouver, British in combination with novel agents targeting the multidrug resist- Columbia, Canada). ance protein Pgp as well as with trastuzumab (Herceptin) for All of the animal studies were done according to proce- treatment of breast cancer patients with tumors overexpressing dures approved by the University of British Columbia Animal Her-2/neu (19, 20). Care Committee. These studies meet the requirements outlined These examples highlight two points. First, liposomal for- in the current guidelines for animal use established by the mulations of anticancer drugs are gaining wider acceptance by Canadian Council of Animal Care. the clinical oncology community, which is evaluating how these Preparation of Liposomes. Lipids (DSPC/Chol; 55/45; 3 14 formulations function when used as part of combination regi- mol%) were dissolved in chloroform, and [ H]CHE or [ C]- mens. Second, vincristine and doxorubicin (or other anthracy- cholesterol hexadecyl ether was added to achieve ϳ5 ␮Ci/100 clines) are still in use together as part of many combination mg lipid. The chloroform was removed under a gentle stream of regimens. This is likely a consequence of many decades of nitrogen gas, and subsequently the lipid samples were placed research providing data highlighting the different modes of under a high vacuum for at least3htoremove residual solvent. action and nonoverlapping toxicities of these agents. Dried lipid films were hydrated with 300 mM MnSO4 (adjusted Previous studies from our laboratory have attempted to to pH 3.5 with dilute HCl; 0.1 N) to achieve a final lipid coformulate these two drugs into a single liposomal formulation concentration of 100 mg/ml. After hydration the multilamellar by an established transmembrane pH gradient; however, this vesicles were subjected to 5 freeze-thaw cycles (freezing in effort proved unsuccessful because of the instability of the liquid nitrogen and thawing at 60°C; Ref. 22). Samples were resulting formulation (21). More recently, a novel doxorubicin extruded 10 times through stacked polycarbonate filters with 0.1 encapsulation method, which relies on drug binding to an en- and 0.08 ␮m pore size at 60°C using a water-jacketed Extruder capsulated transition metal, manganese, has been characterized. (Northern Lipids Inc., Vancouver, British Columbia, Canada). It is proposed here that a manganese metal gradient across The mean size distribution of the resulting liposome prepara- liposomes could be used to encapsulate doxorubicin and, in a tions ranged between 100 and 120 nm as determined by a second step, to form a pH gradient across the doxorubicin- NICOMP Submicron Particle Sizer Model 270 (Pacific Sci- loaded liposomes by addition of the electroneutral ionophore, entific, Santa Barbara, CA) with an argon laser operating at A23187 (divalent cation/proton exchanger). It is demonstrated 632.8 nm. here that the resulting pH gradient can facilitate the rapid In Vitro Drug Loading. Before the addition of drug to accumulation of vincristine without loss of encapsulated doxo- liposomes, a pH gradient (ϳ3 units) and a transition metal rubicin. The coencapsulated formulation exhibited plasma elim- gradient was established across liposome bilayers. To generate ination and drug release rates comparable with those observed the gradients, DSPC/Chol liposomes in MnSO4 buffer (pH 3.5) with liposomes containing the individual active agents. Surpris- were eluted on Sephadex G-50 columns equilibrated with a 300 ingly, the coformulated combination of doxorubicin and vincris- mM Sucrose/20 mM HEPES/15 mM EDTA (SHE) buffer at pH tine exhibited antitumor activity that was no better than that 7.5. To stabilize the pH gradient across the liposomes, the observed using liposomal vincristine alone. divalent cation ionophore A23187 was added to some of the indicated liposome preparations (23, 24). A23187 is an electroneutral ionophore capable of translocating a divalent cation for MATERIALS AND METHODS the exchange of two hydrogen ions (25). A23187 shuttles pro- Materials. Doxorubicin hydrochloride for injection and tons to the vesicle interior in exchange for Mn2ϩ ions, which are vincristine sulfate for injection (Faulding; Kirkland, Quebec, subsequently chelated by the EDTA contained in the SHE Canada) were purchased from the BC Cancer Agency. 1,2 buffer. As described elsewhere (24), doxorubicin was added to Distearoyl-sn-glycero-3-phosphocholine (DSPC) was obtained liposomes (with or without A23187) at 60°C, to achieve a final from Avanti Polar Lipids (Alabaster, AL). Cholesterol (Chol), drug:lipid ratio (wt:wt) of 0.2:1.0. Unless otherwise stated, the A23187, Sephadex G-50, and all of the other chemicals were final lipid concentration was 5 mg/ml, and the final doxorubicin obtained from the Sigma Chemical Company (St. Louis, MO). concentration was 1 mg/ml. For the vincristine-loaded samples, [3H]Cholesteryl hexadecyl ether and [14C]-cholesterol hexade- vincristine in solution was added to liposomes (with A23187 cyl ether were obtained from NEN Life Science Products (Bos- ionophore), and all of the components were preincubated at the ton, MA). [3H]Vincristine sulfate was obtained from Amersham specified incubation temperature (20°C, 40°C, 50°C, or 60°C) to Pharmacia Biotech. (Oakville, Ontario, Canada). Pico-Fluor15 achieve a final drug:lipid ratio (wt:wt) of 0.05:1.0. The final and 40 scintillation fluid was purchased from Canberra-Packard lipid concentration was 5 mg/ml, and the final vincristine con- (Meriden, CT). BALB/c mice and SCID-RAG-2M mice (8–10 centration was 0.25 mg/ml. If A23187 was used, then it was weeks of age) were obtained from the BC Cancer Agency Joint added to the liposomes before addition of drug (after the estab- Animal Facility breeding colony and were housed in microiso- lishment of the pH and metal gradients), at an ionophore:lipid lator units according to established operating procedures. The ratio of 0.2:1.0 (␮g:␮mol). MDA435/LCC6 human breast cancer cell line was a gift from For experiments requiring coencapsulation of doxorubicin Dr. Robert Clarke (Georgetown University, Washington, DC). and vincristine, doxorubicin was encapsulated without A23187. Cell culture medium was composed of DMEM (Stem Cell Doxorubicin was added to liposomes at 60°C, to achieve a final

doxorubicin:lipid ratio (wt:wt) of 0.2:1.0. The doxorubicin- mg/ml, the final vincristine concentration was 0.6 mg/ml, and loaded liposomes were cooled to room temperature, and A23187 the final lipid concentration was 12 mg/ml. was added to achieve an ionophore:lipid ratio of 0.2:1.0 (␮g: All of the drug loaded samples were fractionated on a ␮mol). This mixture was incubated at 50°C for 5 min before Sephadex G-50 column equilibrated with 25 mM HEPES/150 vincristine addition. Vincristine in solution at 50°C was added mM NaCl at pH 7.5 to remove the SHE buffer and/or the to achieve a final vincristine:lipid ratio (wt:wt) of 0.05:1.0 and A23187 (23). The drug-loaded liposomes were diluted with 25 incubated for 60 min. Unless otherwise indicated, the final mM HEPES/150 mM NaCl to a lipid concentration required to doxorubicin concentration was 1 mg/ml, the final vincristine administer a 10 mg/kg doxorubicin dose or a 2.5 mg/kg vincris- concentration was 0.25 mg/ml, and the final lipid concentration tine dose in an injection volume of 200 ␮l. The liposomal lipid was 5 mg/ml for the coencapsulated liposomes. dose was ϳ50 mg/kg. Formulations were injected via the lateral tail Accumulation of the drugs into liposomes was determined vein into 20–22-g female BALB/c mice. At 1, 4, and 24 h after at the indicated time points by removing 100-␮l aliquots and injection the animals (4 mice/group) were killed by CO2 asphyxi- separating unencapsulated drug from encapsulated drug on 1 ml ation, and blood was collected by cardiac puncture and placed into Sephadex G-50 (medium) spin columns equilibrated with SHE EDTA coated microtainers. Plasma was prepared by centrifuging ϳ ϫ buffer. Sample volume was adjusted to 100 ␮l with SHE, to the blood samples at 500 g for 10 min. The liposomal lipid ␮ concentrations in plasma were measured on the basis of incorpo- which 900 l of 1% Triton X-100 was added, and the liposomes 14 were disrupted with excess 1% Triton X-100. Before assessing rated [ C]-cholesterol hexadecyl ether, which is a nonexchange- absorbance at 480 nm, the samples were placed in a Ͼ90°C able and nonmetabolizable marker (26). Plasma vincristine concentrations were measured via the [3 waterbath until the cloud point of the detergent was observed. H]vincristine marker. Doxorubicin was extracted by mixing the plasma with 10% The concentration of doxorubicin in the excluded fraction was SDS and 10 mM H SO (1:1:1; vol:vol), diluting with H Oto1 determined by measuring absorbance (at 480 nm on a Hewlett 2 4 2 ml followed by organic extraction with 1:1 (vol:vol) isopropa- Packard 8453 Spectrophotometer). [3H]Vincristine and lipo- nol-chloroform (2:1; vol:vol with sample). To precipitate plasma some lipid (as measured using [14C]-cholesterol hexadecyl proteins, the samples were frozen at Ϫ80°C ϫ 48 h and then ether) concentrations were determined by scintillation counting thawed at room temperature. The doxorubicin-containing or- (Packard 1900TR Liquid Scintillation Analyzer) using the Sin- ganic phase was separated by centrifugation at 3000 ϫ g for 10 gle and Full Spectrum Dual label DPM analysis counting pro- min at room temperature. Fluorescence in the organic phase was tocols. determined using a Perkin-Elmer LS50B luminescence spec- Doxorubicin-Manganese Binding. To assess doxorubi- trometer using an excitation wavelength of 470 nm (slit width ϭ cin binding to and dissociation from manganese, a spectropho- 2.5) and an emission wavelength of 550 nm (slit width ϭ 10). ⌬ tometric assay was developed. Changes in the A480 ( A480; free The fluorescence readings were compared with a standard curve ⌬ 2ϩ doxorubicin) and A550 (doxorubicin complexed to Mn ) were of doxorubicin that was extracted into the organic phase using determined after addition of A23187. Doxorubicin-loaded lipo- the procedure described above. somes were prepared as described above (without A23187) at a Human Breast Cancer Xenograft Model. Tumors were final lipid concentration of 1 mg/ml. After doxorubicin loading, established in SCID-RAG-2M mice by a single s.c. injection of the sample was diluted to 0.6 mg lipid/ml, A23187 was added, 2 ϫ 106 MDA435/LCC6 cells with 4 mice total/group. Control 2ϩ and the A400-A600 measured. The Mn complexed doxorubicin mice were treated with 0.9% saline. Tumor growth was noted appeared purple in color on visual inspection. As the pH gradi- within 12–14 days after cell injection, and within 18 days, ent formed, resulting in dissociation of the metal-drug complex, measurable tumors (ϳ0.1 g) were observed. Animal weights the color changed to an orange-red, and the A␭ maximum and tumor weights were measured daily until the tumor mass shifted from 550 to 480 nm. exceeded 10% of the original body weight of the animals or In Vivo Drug Release. For in vivo studies assessing until the tumors showed any sign of ulceration. Tumor weight doxorubicin and vincristine release, drug-loaded liposomes were was calculated as weight (g) ϭ {[width (mm)]2 x [length prepared such that doxorubicin was loaded at a 0.2:1.0 drug: (mm)]}/2. When the tumors reached Ϸ0.1 g, mice were treated lipid ratio (wt:wt) and vincristine was loaded at a 0.05:1.0 with doxorubicin and/or vincristine (free or liposomal form) via drug:lipid ratio (wt:wt). Coencapsulated doxorubicin and vin- the lateral tail vein. Liposomal drugs were prepared as described cristine were loaded at these same ratios. For the doxorubicin above and exchanged into 25 mM HEPES/150 mM NaCl using loaded samples with or without A23187, doxorubicin in solution column chromatography before administration at the specified was added to liposomes and incubated at 60°C for ϳ15 min. The doses. All of the treatments were given in 200-␮l injections; final lipid concentration was typically 12 mg/ml, and the final therefore, when two drugs were combined the mice received two doxorubicin concentration was 2.4 mg/ml. For vincristine load- 200-␮l injections separated by approximately 4–6 h with doxo- ing, vincristine in solution was added to liposomes (with rubicin (free or encapsulated) injected first. Coformulated lipo- A23187 ionophore:lipid ratio of 0.2:1.0 ␮g:␮mol) to achieve a somal doxorubicin and vincristine was administered in one final drug:lipid ratio (wt:wt) of 0.05:1.0 and incubated at 50°C. injection of 200 ␮l. The control group received 200 ␮l of sterile The final lipid concentration was 12 mg/ml, and the final vin- saline injection. cristine concentration was 0.6 mg/ml. 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium For the in vivo studies requiring coencapsulated doxorubi- Bromide (MTT) Cytotoxicity Assays and Drug Combination cin and vincristine, liposomes were prepared as described for the Studies. Logarithmically growing MDA435/LCC6 human in vitro studies, and the final doxorubicin concentration was 2.4 breast cancer cells were counted and plated onto 96-well mi-

crotiter Falcon plates at a density of 2.0 ϫ 103 cells/well in 0.1 trapped transition metal. The complexation reaction is readily ml of DMEM supplemented with 10% fetal bovine serum, 1% observed as the solution turns purple when doxorubicin com- L-glutamine, and 1% penicillin and streptomycin solution (Stem plexes with manganese. Cell Technologies, Vancouver, British Columbia, Canada). The The entrapment properties of this loading reaction are perimeter wells of the 96-well plates were not used and con- presented in Fig. 1A, where doxorubicin has been added to the tained 0.2 ml of sterile water. After 24 h at 37°C in humidified liposomes at a drug:lipid ratio of 0.2:1.0 (wt:wt). Upon incuba- Ͼ air with 5% CO2, the media were replaced with 0.2 ml of fresh tion at 60°C, 98% of the added drug is sequestered within the media containing a range of concentrations of doxorubicin or interior of the liposomes within 10 min. Doxorubicin will also vincristine. Drug combination studies where doxorubicin and load into liposomes exhibiting a transmembrane pH gradient. vincristine were added simultaneously at fixed ratios of 4:1, 7:1, Thus, if A23187 is added to the MnSO4 loaded liposomes to or 20:1 doxorubicin to vincristine (mol:mol) were also evalu- generate a pH gradient, the resulting pH gradient is sufficient to ated. Control cells received 0.2 ml of medium. After 72 h cell load doxorubicin (Fig. 1B). The time course indicated that for viability was assessed using a conventional MTT dye reduction both metal complexation and pH gradient loading, Ͼ98% load- assay. Fifty ␮l of 1.25 mg/ml MTT reagent in complete medium ing efficiencies are obtained within 10 min at 60°C for doxo- was transferred to each well, and the plates were incubated for rubicin. 3.5hat37°C. The colored formazan product was then dissolved In contrast with doxorubicin, vincristine does not interact ␮ with Mn2ϩ in a manner that promotes drug accumulation; Ͻ5% using 200 l of DMSO. Plates were read (A570) using a micro- titer plate reader (Dynex Technologies Inc., Chantilly, VA). loading was observed in the absence of the ionophore (Fig. 1C). The percentage of cell survival after treatment was normal- However, after addition of A23187, the pH gradient generated ized with untreated controls. All of the assays were performed at promotes vincristine loading (Fig. 1D). Upon incubation with the ionophore at 60°C, Ͼ98% of the added drug was encapsu- least three times in triplicate to determine the IC90. CalcuSyn software (Biosoft, Ferguson, MO) was used to analyze data lated into liposomes within 10 min. from the MTT assays. The program provides a measure of It is known that the drug accumulation process is temper- whether the combined agents act in an additive, synergistic, or ature- and time-dependent for both doxorubicin and vincristine. antagonistic manner. The combination index (CI) equation in Efficient loading of doxorubicin into DSPC/Chol (55/45 mol%) CalcuSyn is based on the multiple drug-effect equation of Chou liposomes requires an incubation temperature of 60°C (29). The and Talalay (27), and defines synergism as a more-than- influence of temperature on vincristine loading into liposomes expected activity effect and antagonism as a less-than-expected using A23187 to generate a pH gradient is shown in Fig. 2. The additive effect. Chou and Talalay defined a parameter, CI, results indicate that efficient vincristine loading occurs only which can be used to assess synergism (CI Ͻ1), additivity (CI ϭ when the incubation temperature is held between 50°C and 1), or antagonism (CI Ͼ1). 60°C. When the incubation temperature was 40°C, Ͻ20% of the Statistical Analysis. A post hoc comparison of means added drug was loaded into the liposomes at the end of the 1-h (Scheffe´ test) was performed with the Statistica software pack- time course. It is possible that reductions in vincristine loading age (StatSoft Inc., Tulsa, OK) on plasma elimination and tumor efficiencies may be a consequence of reduced vincristine per- growth analyses involving the administration of both free and meation across the lipid bilayer at these temperatures (30, 31); liposomal formulations of doxorubicin. Differences were con- however, the temperature effects could also be attributed to sidered significant at P Ͻ 0.05. reduced activity of the ionophore (32). Considering the objective of loading both doxorubicin and vincristine into the same liposome population using two loading RESULTS methods (transition metal complexation and pH gradient) and Coencapsulation of Doxorubicin and Vincristine into knowing that the complexation reaction between doxorubicin Manganese-Sulfate Containing DSPC/Chol Liposomes. and manganese is dependent on pH (greatest at pH Ͼ6.5; Ref. The primary objective of this study was to coencapsulate doxo- 33), it was important to define conditions where vincristine rubicin and vincristine within one liposome population by tak- loading occurred before dissociation of the doxorubicin-manga- ing advantage of two distinct loading methods. The first step nese complex. To determine the rate of this dissociation reaction relies on complexation of doxorubicin to the transition metal at different temperatures, MnSO4-loaded DSPC/Chol (55/45 manganese (24, 28). The second step uses an ionophore-gener- mol%) liposomes were prepared and incubated with doxorubi- ated transmembrane pH gradient (23) driving the encapsulation cin at a drug:lipid ratio of 0.2:1.0 (wt:wt). On encapsulation of of vincristine into the manganese-doxorubicin liposomes. Lipo- doxorubicin, we have measured the pH gradient across liposome somes were prepared in 300 mM MnSO4 (pH 3.5) followed by bilayers using methylamine as a pH-sensitive probe (24, 34) and the exchange of outside buffer to 300 mM sucrose/20 mM have estimated the interior liposomal pH to be ϳ7.3 (data not HEPES/15 mM EDTA (SHE) at pH 7.5, thereby establishing shown). After drug loading, A23187 was added and the conver- both a pH and a metal ion gradient. In this system the neutral sion of the manganese-doxorubicin complex to uncomplexed species of doxorubicin crosses the lipid bilayer, and subse- doxorubicin was followed spectrophotometrically. We have also quently doxorubicin is protonated within the interior of lipo- measured existing pH gradients across the doxorubicin-loaded somes. This causes the interior pH of the liposomes to increase liposomes with added A23187 and have estimated that the because there is no entrapped solute that can buffer the change interior liposomal pH is ϳ3.9. in proton concentration. When the internal pH is Ͼ6.5, the The results shown in Fig. 3 were obtained by measuring the encapsulated doxorubicin can form a complex with the en- decrease in A550 as a function of time after addition of A23187

Fig. 1 Doxorubicin or vincristine encapsulation in 1,2 distearoyl-sn-glycero-3- phosphocholine/cholesterol (55/45 mol%) liposomes using (A and C) the manganese- sulfate loading procedure alone or (B and D) with A23187. Liposomes were prepared as described in “Materials and Methods,” and the outer buffers were exchanged using column chromatography to create a metal gradient. For B and D A23187 was added and incubated 5 min before the addition of drug. Doxorubicin was added to the liposomes to achieve a 0.2:1.0 (wt:wt) drug:lipid ratio, and vincristine was added to the liposomes to achieve a 0.05:1.0 (wt:wt) drug:lipid ratio, and all preparations were incubated at 60°C. Doxorubicin was quantitated by the

A480 of a detergent-solubilized sample, and vincristine was quantitated using a [3H]vincristine-radiolabeled marker as described in “Materials and Methods.” Data points represent mean drug:lipid ratios; bars, Ϯ SD (n ϭ 3).

to the doxorubicin-loaded liposomes. A23187 exchanges pro- no decrease in A550 was observed over the 40-min time course, tons in and Mn2ϩ out of the liposomes reducing the interior indicating that the manganese-doxorubicin complex was stable liposomal pH. A decrease in absorbance at 550 nm indicates the under these conditions. As noted in Fig. 2, vincristine does not dissociation of the doxorubicin-metal chelate and reprotonation load into these liposomes at this temperature, and it is likely that of the chromophore at the reduced pH. At 40°C (filled squares), the ionophore is not functioning optimally at this temperature. When the incubation temperature was increased to 50°C (filled

circle) and 60°C (filled triangle), a decrease in A550 indicated dissociation of the manganese-doxorubicin complex. Destabili- zation of this complex was much slower at 50°C, with the

maximal changes in A550 observed in 30 min after A23187 addition. The visible color change from purple to orange was not seen in the sample incubated at 40°C. When considered together, the results in Figs. 2 and 3 suggest that, at an incubation temperature of 50°C: (a) vincristine loading would occur within 40 min after A23187 addition to

the MnSO4 containing DSPC/Chol (55/45 mol%) liposomes; and (b) over the same time period and under the same conditions the manganese-doxorubicin complex dissociates. It can be suggested that vincristine encapsulation could be achieved in doxorubicin-loaded liposomes before or during dissociation of the manganese:doxorubicin complex. This loading reaction is shown in Fig. 4. Doxorubicin-loaded DSPC/Chol liposomes Fig. 2 Vincristine encapsulation in 1,2 distearoyl-sn-glycero-3-phos- (0.2:1.0 drug:lipid ratio; wt:wt) were mixed with A23187 at phocholine/cholesterol (55/45 mol%) liposomes using the manganese- sulfate loading procedure with the A23187 ionophore at 20°C(f), 40°C 50°C just before the addition of vincristine (added at a 0.05 (F), 50°C(Œ), and 60°C(). Liposomes were prepared as described in drug:lipid weight ratio; wt:wt). Vincristine loading was com- “Materials and Methods,” and the outer buffers were exchanged using plete within 40 min with no loss of doxorubicin from the column chromatography to create a metal gradient. A23187 was added liposomes during this time frame, indicating that the encapsu- and incubated 5 min before the addition of drug. Vincristine was added to the liposomes to achieve a 0.05:1.0 (wt:wt) drug:lipid ratio. Data lation efficiency approached 100% for both drugs using the points represent mean drug:lipid ratios; bars, Ϯ SD (n ϭ 3). coencapsulation procedure.

were prepared in the absence of the ionophore (filled circle). This is most evident at the 24-h time point, where the plasma doxorubicin levels are almost 10-fold lower when compared with those achieved after administration of the doxorubicin- loaded liposomes prepared with A23187 (filled square). Plasma vincristine levels (Fig. 5C) were comparable with those observed for DSPC/Chol liposomal vincristine prepared using citrate based loading methods (4). When both drugs were coencapsulated within liposomes (Fig. 5D) the plasma levels of doxorubicin and vincristine were comparable (P Ͼ 0.1181) to that observed for liposomal formulations containing only one of the agents. Drug release from liposomes in the plasma compartment can be estimated from the data shown in Fig. 5 by calculating the drug:lipid ratios at the indicated time points. Results for the doxorubicin-loaded liposomes (single agent), shown in Fig. 6A, clearly indicate that doxorubicin release from the DSPC/Chol 2ϩ ⌬ liposomes is faster when loading is achieved through Mn - Fig. 3 Changes in A550 ( A550) of doxorubicin loaded 1,2 distearoyl- sn-glycero-3-phosphocholine/cholesterol liposomes after the addition of doxorubicin complexation. The formulation prepared using the ionophore A23187 at 40°C(f), 50°C(F), and 60°C(Œ). Liposomes A23187 in the procedure exhibited no measurable change in were prepared as described in “Materials and Methods,” and the outer drug:lipid ratio over the 24 h time course, consistent with buffers were exchanged using column chromatography to create a metal gradient. Doxorubicin was added to the liposomes to achieve a 0.2:1.0 previous data for doxorubicin loaded into DSPC/Chol liposomes (wt:wt) drug:lipid ratio. A23187 was added in each experiment at t ϭ 0 using the pH gradient procedure [encapsulated 300 mM Citrate and incubated at the indicated temperatures; a decrease in absorbance at (pH 4.0); Ref. 36]. In contrast, vincristine is readily released 550 nm is indicative of the conversion of the manganese complex to from DSPC/Chol liposomes over the 24-h time course (Fig. 6B), uncomplexed doxorubicin in a low pH environment. Data points are where the vincristine:lipid ratio at 24 h is 3.3-fold lower than representative of at least six replicate experiments. that of the injected formulation. When the injected liposomes contain doxorubicin and vincristine (Fig. 6C), the drug release profiles are superimposable to those measured for liposomes In Vivo Plasma Elimination of Liposomes with Coen- containing the single agents encapsulated using A23187. capsulated Doxorubicin and Vincristine. The procedure described above results in efficient encapsulation of both doxorubicin and vincristine into DSPC/Chol (55/45 mol%) liposomes. As noted elsewhere, it is possible that the presence of two encapsulated drugs may alter drug release from liposomes in the plasma compartment (21) and perhaps even the plasma elimination rate of the liposomes themselves. The plasma elimination of liposomal lipid and associated drugs was evaluated after i.v. administration in mice. These data have been summarized in Fig. 5, where liposomal lipid and drug levels at 1, 4, and 24 h after i.v. injection are shown. Four DSPC/Chol liposomal formulations were evaluated in these studies: (a) doxorubicin encapsulated without the ionophore; (b) doxorubicin encapsulated with the ionophore; (c) vincristine encapsulated with the ionophore; and (d) doxorubicin and vincristine coencapsulated as describe above. BALB/c mice were injected with lipid doses of ϳ50 mg/kg, and doxorubicin doses of 10 mg/kg and/or vincristine doses of 2.5 mg/kg. As judged by the plasma levels of liposomal lipid (Fig. 5A), all of the liposomes were eliminated Fig. 4 Doxorubicin and vincristine encapsulation within 1,2 dis- from the plasma compartment at comparable rates, although tearoyl-sn-glycero-3-phosphocholine/cholesterol (55/45 mol%) lipo- there are significantly higher levels (P Ͻ 0.016) of liposomal somes. Liposomes were prepared as described in “Materials and Meth- lipid at the 24-h time point when liposomes contained both ods,” and the outer buffers were exchanged using column chromatography to create a metal gradient. Doxorubicin (F) was added vincristine and doxorubicin compared with either alone. It is to the liposomes to achieve a 0.2:1.0 (wt:wt) drug:lipid ratio and known that doxorubicin or vincristine, when loaded into a incubated at 60°C. A23187 was added to the doxorubicin-loaded lipo- liposomal carrier, can alter the plasma elimination of the carrier somes and incubated 5 min before the addition of vincristine (f), which (30, 35); therefore, this result could suggest that the combination was added to achieve a final drug:lipid ratio of 0.05:1.0 (wt:wt). Vin- cristine loading was completed using an incubation temperature of of drugs actually enhances this effect. Plasma doxorubicin levels 50°C. Doxorubicin was quantitated by the A480 of a detergent-solubi- (Fig. 5B) are consistent with previous studies where doxorubicin lized sample as described in “Materials and Methods.” Data points was eliminated more rapidly when the drug-loaded liposomes represent mean drug:lipid ratios; bars, Ϯ SD (n ϭ 3).

Fig. 6 In vivo release of doxorubicin and vincristine from 1,2 distearoyl-sn-glycero-3-phosphocholine/cholesterol (55/45 mol%) liposomes. Panels represent drug:lipid ratios calculated from Fig. 5. Lipo- somes containing (A): doxorubicin encapsulated using the manganese- sulfate procedure either without A23187 (F); or with A23187 (f); B, vincristine encapsulated using the manganese-sulfate procedure with Fig. 5 The lipid and drug plasma elimination profiles of drug-loaded A23187 (ࡗ); C, doxorubicin (Œ) and vincristine () coencapsulated 1,2 distearoyl-sn-glycero-3-phosphocholine/cholesterol (55/45 mol%) using the manganese-sulfate procedure as described in Fig. 4 after i.v. liposomes after i.v. administration into female BALB/c mice. Lipo- administration to female BALB/c mice. Drug-loaded liposomes were somes were loaded with doxorubicin, vincristine, or both. Loaded lipo- adjusted to a concentration such that a dose of 10 mg/kg doxorubicin somes were adjusted to a concentration such that 10 mg/kg doxorubicin and/or 2.5 mg/kg vincristine could be administered in an injection or 2.5 mg/kg vincristine could be administered in an injection volume of volume of 200 ␮l. Drug levels in the plasma were assessed via meas- 200 ␮l. A, plasma lipid levels in mice administrated either doxorubicin- urements of doxorubicin fluorescent equivalents or the [3H]vincristine loaded liposomes using the manganese sulfate loading procedure (F), sulfate marker. Data points represent mean drug:lipid ratios; bars, Ϯ SD doxorubicin-loaded liposomes using the manganese sulfate loading pro- (n ϭ 6). cedure with A23187 ionophore (f), vincristine-loaded liposomes using the manganese sulfate loading procedure with A23187 ionophore (Œ), or simultaneously loaded doxorubicin and vincristine liposomes (). Lipid levels were determined using ; [3H]-cholesteryl hexadecyl ether or [14C]-cholesteryl hexadecyl ether as a liposomal lipid marker. B, the Efficacy of the Coencapsulated Liposomal Doxorubicin/ level of doxorubicin fluorescent equivalents in plasma from mice ad- Vincristine Formulation Against a Human Breast Cancer ministered doxorubicin-loaded liposomes using the manganese sulfate loading procedure with (f) or without (F) the A23187 ionophore. C, Xenograft Model. Vaage et al. (37) have evaluated previ- vincristine levels in plasma were determined from mice that received ously the therapeutic activity of vincristine and doxorubicin vincristine-loaded liposomes prepared using the manganese sulfate loading encapsulated in sterically stabilized liposomes. In these studies procedure with A23187 (ࡗ). D, doxorubicin (Œ) and vincristine () levels the drugs were not coencapsulated, and the drugs were used were also measured in plasma of mice receiving coencapsulated liposomes. Data points represent mean drug:lipid ratios; bars, Ϯ SD (n ϭ 6). A alone and in combination to treat murine MC2 tumors. The one-way ANOVA analysis was performed to assess differences among the liposomal drugs were more active than free agents; however, treatment groups. A P Ͻ 0.05 was considered significant. when they were used in combination the liposomal vincristine

Table 1 Treatment of SCID/RAG 2M mice bearing MDA MB435/LCC6 tumors Time taken for tumor Percentage to reach 0.2 gramsa,b Growth delayc growth Treatment group (days Ϯ SE) (days Ϯ SE) delayd Saline 39 Ϯ 40 Free vincristine (2.5 mg/kg; maximum tolerated dose, MTD) 48 Ϯ 3 8.5 Ϯ 2.9 22 Liposomal vincristine (2.5 mg/kg; MTD) 55 Ϯ 316Ϯ 2.6 41 Free doxorubicin (10 mg/kg) 42 Ϯ 3 3.3 Ϯ 2.9 8 Liposomal doxorubicin (10 mg/kg) 46 Ϯ 3 7.5 Ϯ 2.3 19 Free vincristine/doxorubicin (2.5/10 mg/kg; MTD) 47 Ϯ 2 8.5 Ϯ 1.8 22 Coencapsulated vincristine and doxorubicin (2.5/10 mg/kg; MTD) 52 Ϯ 2 12.5 Ϯ 2.1 32 Liposomal vincristine and liposomal doxorubicin (2.5/10 mg/kg; MTD) 50 Ϯ 311Ϯ 2.4 28 a Tumors were established in SCID-RAG-2M mice by a single s.c. injection of 2 ϫ 106 MDA435/LCC6 cells with 4 mice total per group. Tumor growth was noted within 12–14 days after cell injection and within 18 days measurable tumors (ϳ0.1 g) were noted. b Data represents days required for the tumor mass to reach 0.2 g. Results were averaged from at least 4 mice, and a one-way ANOVA analysis was performed to assess differences among the treatment groups. A P of Ͻ0.05 was considered significant. c Growth delay is determined by subtracting the mean time taken for the control tumors to reach 0.2 g from the mean time taken for the treated tumors to reach 0.2 g. d Percentage growth delay is growth delay expressed as a percentage of mean time taken for the control tumors to reach 0.2 g.

and liposomal doxorubicin did not provide improved therapy 16-day delay in tumor growth, which was ϳ2-fold greater than when simultaneously injected (37). Their results suggested that that achieved with free vincristine at the same dose. Liposomal liposomal vincristine inhibited the activity of liposomal doxo- doxorubicin used as a single agent resulted in a 7-day delay in rubicin. These authors proposed that reduced activity could have tumor growth, and this was better activity than free doxorubicin been a consequence of the effects of vincristine on the cell cycle, by Ͼ2-fold. When combined with liposomal vincristine or ad- which prevented/delayed progression into S phase in which the ministered coencapsulated with vincristine, delays in tumor impact of doxorubicin could be greatest (37). Alternatively, they growth of 11 and 12 days were observed, respectively. argued that inhibition of activity could have been a consequence In Vitro Cytotoxicity Analysis of Doxorubicin and Vin- of dosing schedules and liposome-mediated drug delivery, cristine Alone and in Combination. Results presented here where liposomal vincristine administration may have interfered as well as those of Vaage et al. (37) suggest that coadministra- with the delivery of liposomal doxorubicin to the tumor. This tion of liposomal doxorubicin and liposomal vincristine results later concern would not arise if the drugs were coencapsulated; in less than additive activity or antagonism. In an effort to gain therefore, a therapeutic assessment of this liposomal formulation a better understanding about the cytotoxic effects of vincristine is described here. and doxorubicin when used in combination, in vitro cytotoxicity Efficacy was determined in SCID mice bearing solid tu- studies were completed using MDA435/LCC6 cells as the target mors derived from the s.c. injection of MDA435/LCC6 human cell population. Cytotoxicity effects were determined using the breast cancer cells. This is an aggressive breast tumor model that MTT assay as described in “Materials and Methods.” The ef- has been described elsewhere (38). The results of these studies fects of the combined free drugs were analyzed using the mul- are summarized in Table 1. Complete dose titrations were com- tiple drug-equation developed by Chou and Talalay (27). This pleted; however, the activity observed at the maximum tolerated method is most readily applied to dose titration data collected dose of drug (2.5 mg/kg vincristine ϩ10 mg/kg doxorubicin) for individual agents alone, and the combination of drugs added resulted in only a 32% delay in tumor growth; thus, only the at fixed ratios and over a broad range of effective doses. results obtained at the maximum tolerated dose are presented In this study the vincristine:doxorubicin ratios were se- here. Tumor growth delay values calculated as the time in days lected to reflect the drug ratios measured in the plasma after i.v. for the tumors to reach 200 mg in treated animals minus the time administration. Three fixed vincristine:doxorubicin ratios of 1:4, to reach the same size in saline-treated control animals are 1:7, and 1:20 (mol:mol) were evaluated against the MDA435/ shown in Table 1 for animals treated with: (a) free vincristine LCC6 cell line. The effects of the drugs alone and in combina- and DSPC/Chol (55/45 mol%) liposomal vincristine (2.5 mg/ tion were evaluated after a 72-h drug exposure. The MTT results kg); (b) free doxorubicin and DSPC/Chol (55/45 mol%) liposo- (Fig. 7) were completed in triplicate, three separate times. The mal doxorubicin (10 mg/kg); (c) a combination of free vincris- resulting nine data points at each concentration were then ana- tine and free doxorubicin (2.5 and 10 mg/kg, respectively); (d) lyzed using the CaluSyn Software program. This software pack- a combination of liposomal vincristine and liposomal doxoru- age takes the nonlinear dose response curves and applies an bicin (2.5 and 10 mg/kg, respectively); and (e) the coencapsu- unweighted linear regression analysis. Our analysis, which as- lated liposomal vincristine and doxorubicin formulation (2.5 sumed mutually exclusive interactions, provided a correlation mg/kg vincristine:10 mg/kg doxorubicin). The results demon- coefficient of Ͼ0.9 for all of the data sets. The regression strate that optimal tumor growth inhibition is achieved when the analysis provided estimations of the drug concentrations re- tumor-bearing animals are treated with liposomal vincristine. quired to inhibit cell growth/proliferation at a variety of effect Ͻ ϳ This improvement in therapy was statistically significant (P levels. The IC90 of vincristine alone was 3.7 nM, whereas 0.05). At the 2.5 mg/kg dose, liposomal vincristine caused a doxorubicin was 10 ␮M (Fig. 7). These data demonstrate that the

cussion will consider the potential for encapsulating multiple drugs into liposomes using distinct loading methods as well as providing a rationale as to why the coencapsulated vincristine/ doxorubicin formulation exhibited antagonistic effects. Our initial efforts to encapsulate two drugs into a single liposome population using pH gradient-mediated approaches was hindered in part because the stability of these formulations was poor both in vitro and in vivo (21). More recently our group (24) has defined a unique drug loading method that relies on formation of a manganese-doxorubicin drug complex rather than a transmembrane pH gradient. Applying this new method to the problem of coencapsulating two anticancer drugs, it is demonstrated here that a stable formulation at drug:lipid ratios shown to be therapeutically relevant (4, 29) can be easily prepared. An established manganese metal gradient across lipo- Fig. 7 IC90 values determined from in vitro 3-(4,5-dimethylthiazol-2- somes permitted the efficient loading of doxorubicin (Fig. 1A) yl)-2,5-diphenyltetrazolium bromide (MTT) assays assessing the activ- due to formation of a drug-metal complex. Sufficient levels of ity of doxorubicin and vincristine combinations. MDA435/LCC6 human 2ϩ breast cancer cells were plated and allowed to adhere for 24 h. Drug was uncomplexed Mn remained to support the establishment of a diluted and added at the indicated ratios. Control wells received 0.2 ml transmembrane pH gradient after addition of A23187, an elec- medium without drug. After 72 h, the cell viability was assessed using troneutral divalent metal/proton pump. The pH gradient formed a conventional MTT dye reduction assay (see “Materials and Meth- supported the efficient loading of vincristine. The methods ods”). The MTT results were completed in triplicate, three separate times. Data were analyzed using the Calcusyn program, which incorpo- described here are generally applicable to many drugs that have rates the median dose effect method developed by Chou and Talalay. chemical groups capable of complexing transition metals like 2ϩ This analysis was used to estimate drug concentrations required for IC90. Mn [for example anthracyclines (40, 41), camptothecins (42, These cells were exposed to doxorubicin (f) and vincristine (u) sepa- 43), and anticancer antibiotics such as bleomycin (44, 45)] and rately (1) and combined in three different ratios (2) 4:1 doxorubicin second agents that can be encapsulated using pH gradients (for (DOX):vincristine (VINC) ratio (mol:mol), (3) 7:1 DOX:VINC ratio (mol:mol), and (4) 20:1 DOX:VINC ratio (mol:mol). example mitoxanthrone, camptothecins, and vincalkaloids; Refs. 5, 23, 46, 47). Plasma elimination data suggest that the release rates of coencapsulated drugs observed are consistent with what would be expected for the drugs encapsulated into therapeutic activity of the combination would be dominated by DSPC/Chol liposomes using pH gradient loading methods (see the activity of vincristine. The estimated drug combination Figs. 5 and 6). The doxorubicin and vincristine drug release concentrations required to achieve 90% inhibition of cell profiles from individually loaded DSPC/Chol liposomes (Fig. 6, growth/proliferation at the 1:4 vincristine:doxorubicin ratio A and B) are not significantly different from the drug release (mol:mol) indicate that the amount of vincristine required was rates observed in the coencapsulated formulation (Fig. 6C). This comparable with that needed when vincristine was used alone suggests that the drugs are not interacting in a manner that ϳ ( 3nM). As the ratio changed to 1:20 vincristine:doxorubicin affects their release from the liposomes. (mol:mol), the estimated amount of the drugs required to The rationale for selecting vincristine and doxorubicin for achieve 90% cell kill increased substantially for both vincristine coencapsulation was based on the fact that these two drugs are ␮ and doxorubicin. Specifically, the IC90 observed was 4.3 M for commonly used in combination chemotherapy. The drugs ex- ␮ vincristine and 86 M for doxorubicin. hibit different dose-limiting toxicities and different mechanisms These data reveal the presence of strong antagonism for of activity. Because liposomal forms of vincristine and doxoru- this drug combination, and this was reflected in the CI values bicin are more active than the respective free drugs, it was defined by the CalcuSyn Software (39). It should be noted that anticipated that combining the two liposomal drugs would be a CI value of 0.9–1.1 is indicative of additive activity, whereas beneficial. One significant disadvantage of administrating indi- Ͻ Ͼ CI values 0.9 indicate drug synergy and values 1.1 indicate vidually encapsulated doxorubicin and vincristine is the poten- antagonism. When analyzed in this manner, the combination of tial interference each liposome population may exert on the vincristine and doxorubicin, at all of the ratios studied, exhibited pharmacokinetic profile of the other, thereby altering drug de- Ͼ CI values 1.1 over the entire range of effective concentrations. livery to tumor cells. We attempted to address this problem with At the 1:20 vincristine:doxorubicin (mol:mol) the CI value was coencapsulation, which would make this an unlikely issue. Ͼ 20, indicative of strong antagonism. However, it should be noted that a previous study demonstrated that simultaneous administration of separately encapsulated li- DISCUSSION posomal doxorubicin and liposomal vincristine exhibited activ- The results summarized in this report can sustain two ity no better than that achieved for liposomal doxorubicin alone avenues of discussion. Of practical significance, a procedure to (37). An explanation as to why codelivery of vincristine and coencapsulate two anticancer drugs into a single liposomal doxorubicin may result in less than additive activity can be formulation has been delineated. However, data are provided developed when considering the mechanisms of action of the that clearly demonstrate that the coencapsulated vincristine and free drugs. It is established that free doxorubicin has some doxorubicin formulation exhibits antagonistic effects. This dis- activity against cells in various phases of the cell cycle, but its

maximal effects occur in early S phase. This action causes a 3. Rahman, A., Kessler, A., More, N., Sikic, B., Rowden, G., Woolley, delay in the progression through all phases of the cell cycle P., and Schein, P. S. Liposomal protection of adriamycin-induced car- 3 diotoxicity in mice. Cancer Res., 40: 1532–1537, 1980. except M G1 (48, 49). Vincristine causes cells to accumulate in mitosis (50–52). On the basis of this information, it is 4. Mayer, L. D., Bally, M. B., Loughrey, H., Masin, D., and Cullis, P. R. Liposomal vincristine preparations which exhibit decreased drug reasonable to suggest that these two drugs could interfere with toxicity and increased activity against murine L1210 and P388 tumors. each other’s action. The Chou and Talalay analysis of vincris- Cancer Res., 50: 575–579, 1990. tine and doxorubicin combinations at fixed ratios (where the 5. Burke, T. G., and Gao, X. Stabilization of topotecan in low pH drugs are added simultaneously) clearly indicated that this drug liposomes composed of distearoylphosphatidylcholine. J. Pharm. Sci., combination is antagonistic (Fig. 7). These findings (Fig. 7; 83: 967–969, 1994. Table 1) indicate that drug combinations selected on the basis of 6. Rihova, B. Immunomodulating activities of soluble synthetic polymer-bound drugs. Adv. Drug Deliv. Rev., 54: 653–674, 2002. nonoverlapping toxicity and unique mechanisms of action can 7. Alakhov, V., Klinski, E., Lemieux, P., Pietrzynski, G., and Kabanov, provide less than optimal therapeutic effects, even when the A. Block copolymeric biotransport carriers as versatile vehicles for drug pharmacokinetics and therapeutic properties of the individual delivery. Exp. Opin. Biol. Ther., 1: 583–602, 2001. drugs are improved through use of carefully designed drug 8. Gariepy, J., and Kawamura, K. Vectorial delivery of macromolecules carriers. into cells using peptide-based vehicles. Trends Biotechnol., 19: 21–28, Accepting the concerns identified above for coencapsu- 2001. lated vincristine and doxorubicin, there is also a tremendous 9. Newell, M., Milliken, S., Goldstein, D., Lewis, C., Boyle, M., Dolan, G., Ryan, S., and Cooper, D. A. A phase II study of liposomal doxo- opportunity to define strategies where the drugs encapsulated rubicin in the treatment of HIV- related Kaposi’s sarcoma. Aust. N. Z. J. within the liposomal formulation are selected using arguments Med., 28: 777–783, 1998. that consider the potential for drug combinations to exhibit 10. Tejada-Berges, T., Granai, C. O., Gordinier, M., and Gajewski, W. supra-additive activity or synergy. On the basis of evidence Caelyx/Doxil for the treatment of metastatic ovarian and breast cancer. supporting the mechanism of selected agents on tumor cells, it Exp. Rev. Anticancer Ther., 2: 143–150, 2002. should be possible to better rationalize the design of a formu- 11. Batist, G., Barton, J., Chaikin, P., Swenson, C., and Welles, L. lation with an effective drug combination. The use of multiple Myocet (liposome-encapsulated doxorubicin citrate): a new approach in breast cancer therapy. Exp. Opin. Pharmacother., 3: 1739–1751, 2002. agents has to date been developed based on pragmatic princi- 12. Sharpe, M., Easthope, S. E., Keating, G. M., and Lamb, H. M. ples, addressing issues of overlapping toxicities, drug resistance, Polyethylene glycol-liposomal doxorubicin: a review of its use in the and tumor cell heterogeneity. Yet there are now a variety of in management of solid and haematological malignancies and AIDS-re- vitro assays, like the Chou and Talalay analysis used here, that lated Kaposi’s sarcoma. Drugs, 62: 2089–2126, 2002. can establish whether two drugs act synergistically or antago- 13. Christou, L., Hatzimichael, E., Chaidos, A., Tsiara, S., and Bou- nistically when used in combination (53). Although these in rantas, K. L. Treatment of plasma cell leukemia with vincristine, liposomal doxorubicin and dexamethasone. Eur. J. Haematol., 67: 51–53, vitro assays can provide information on drug sequencing effects, 2001. typically the drugs are added simultaneously to the cells in 14. Hussein, M. A., Wood, L., Hsi, E., Srkalovic, G., Karam, M., Elson, culture. Given the fact that we have established a method P., and Bukowski, R. M. A Phase II trial of pegylated liposomal whereby two agents can be efficiently loaded into a liposomal doxorubicin, vincristine, and reduced-dose dexamethasone combination formulation, which, when given i.v., will dictate the plasma therapy in newly diagnosed multiple myeloma patients. Cancer (Phila.), elimination and biodistribution of both agents, it is now reason- 95: 2160–2168, 2002. able to consider combining data obtained from in vitro synergy 15. Tsavaris, N., Kosmas, C., Vadiaka, M., Giannouli, S., Siakantaris, M. P., Vassilakopoulos, T., and Pangalis, G. A. Pegylated liposomal assays with liposomal drug formulation methods to define car- doxorubicin in the CHOP regimen for older patients with aggressive rier systems capable of delivering coencapsulated drugs in a (stages III/IV) non-Hodgkin’s lymphoma. Anticancer Res., 22: 1845– synergistic manner. Importantly, it is also reasonable to consider 1848, 2002. the potential of formulated drug combinations, where the drugs 16. Sarris, A. H., Hagemeister, F., Romaguera, J., Rodriguez, M. A., selected are encapsulated in different liposomes, each optimized McLaughlin, P., Tsimberidou, A. M., Medeiros, L. J., Samuels, B., Pate, O., Oholendt, M., Kantarjian, H., Burge, C., and Cabanillas, F. Liposo- for the individual agents. These could then be combined and mal vincristine in relapsed non-Hodgkin’s lymphomas: early results of administered as a single drug product. If this strategy was an ongoing phase II trial. Ann. Oncol., 11: 69–72, 2000. pursued it would be important to demonstrate that the pharma- 17. Swenson, C. E., Bolcsak, L. E., Batist, G., Guthrie, T. H., Jr., cokinetic and biodistribution attributes of the different liposo- Tkaczuk, K. H., Boxenbaum, H., Welles, L., Chow, S. C., Bhamra, R., mal carriers are comparable. and Chaikin, P. Pharmacokinetics of doxorubicin administered i. v. as Myocet (TLC D-99; liposome-encapsulated doxorubicin citrate) compared with conventional doxorubicin when given in combination with cyclophosphamide in patients with metastatic breast cancer. Anticancer ACKNOWLEDGMENTS Drugs, 14: 239–246, 2003. We thank the staff of the Joint Animal Care Facility 18. Valero, V., Buzdar, A. U., Theriault, R. L., Azarnia, N., Fonseca, G. A., Willey, J., Ewer, M., Walters, R. S., Mackay, B., Podoloff, D., Booser, D., Lee, L. W., and Hortobagyi, G. N. Phase II trial of liposome- REFERENCES encapsulated doxorubicin, cyclophosphamide, and fluorouracil as first- 1. Goldie, J. H. Drug resistance in cancer: a perspective. Cancer Metas- line therapy in patients with metastatic breast cancer. J. Clin. Oncol., 17: tisis Rev., 20: 63–68, 2001. 1425–1434, 1999. 2. Hobbs, S. K., Monsky, W. L., Yuan, F., Roberts, W. G., Griffith, L., 19. Gianni, L. The future of targeted therapy: combining novel agents. Torchilin, V. P., and Jain, R. K. Regulation of transport pathways in Oncology, 63: 47–56, 2002. tumor vessels: role of tumor type and microenvironment. Proc. Natl. 20. Winer, E. P., and Burstein, H. J. New combinations with Herceptin Acad. Sci. USA, 95: 4607–4612, 1998. in metastatic breast cancer. Oncology, 61: 50–57, 2001.

21. Saxon, D., Mayer, L., and Bally, M. Liposomal anticancer drugs as bicin encapsulated in sterically stabilized liposomes. Int. J. Cancer, 54: agents to be used in combination with other anticancer agents: studies on 959–964, 1993. a liposomal formulation with two encapsulated drugs. J. Liposome Res., 38. Leonessa, F., Green, D., Licht, T., Wright, A., Wingate-Legette, K., 9: 507–522, 1999. Lippman, J., Gottesman, M. M., and Clarke, R. MDA435/LCC6 and 22. Mayer, L. D., Hope, M. J., Cullis, P. R., and Janoff, A. S. Solute MDA435/LCC6MDR1: ascites models of human breast cancer. Br. J. distributions and trapping efficiencies observed in freeze- thawed mul- Cancer, 73: 154–161, 1996. tilamellar vesicles. Biochim. Biophys. Acta, 817: 193–196, 1985. 39. Chou, T. C., and Hayball, M. P. CalcuSyn. Windows Sorfware for 23. Fenske, D. B., Wong, K. F., Maurer, E., Maurer, N., Leenhouts, Dose Effect Analysis, pp. 1–56. Cambridge, UK: Biosoft, 1996. J. M., Boman, N., Amankwa, L., and Cullis, P. R. Ionophore-mediated uptake of ciprofloxacin and vincristine into large unilamellar vesicles 40. Beraldo, H., Garnier-Suillerot, A., Tosi, L., and Lavelle, F. Iro- exhibiting transmembrane ion gradients. Biochim. Biophys. Acta, 1414: n(III)-adriamycin and Iron(III)-daunorubicin complexes: physicochem- 188–204, 1998. ical characteristics, interaction with DNA, and antitumor activity. Bio- 24. Abraham, S. A., Edwards, K., Karlsson, G., MacIntosh, S., Mayer, chemistry, 24: 284–289, 1985. L. D., McKenzie, C., and Bally, M. Formation of transition metal- 41. Fiallo, M. M., Garnier-Suillerot, A., Matzanke, B., and Kozlowski, doxorubicin complexes inside liposomes. Biochim. Biophys. Acta, H. How Fe3ϩ binds anthracycline antitumour compounds. The myth 1565: 41–54, 2002. and the reality of a chemical sphinx. J. Inorg. Biochem., 75: 105–115, 25. Wang, E., Taylor, R. W., and Pfeiffer, D. R. Mechanism and 1999. specificity of lanthanide series cation transport by ionophores A23187, 42. Kuwahara, J., Suzuki, T., and Sugiura, Y. Studies on antitumor 4-BrA23187 and ionomycin. Biophysical J., 75: 1244–1254, 1998. drugs targeting DNA: photosensitive DNA cleavage of copper-campto- 26. Pool, G. L., French, M. E., Edwards, R. A., Huang, L., and Lumb, thecin. Nucleic Acids Symp. Ser., 201–204, 1985. R. H. Use of radiolabeled hexadecyl cholesteryl ether as a liposome 43. Kuwahara, J., Suzuki, T., Funakoshi, K., and Sugiura, Y. Photo- marker. Lipids, 17: 448–452, 1982. sensitive DNA cleavage and phage inactivation by copper(II)-campto- 27. Chou, T. C., and Talalay, P. Quantitative analysis of dose-effect thecin. Biochemistry, 25: 1216–1221, 1986. relationships: the combined effects of multiple drugs or enzyme inhib- 44. Dabrowiak, J. C., Greenaway, F. T., and Grulich, R. Transition- itors. Adv. Enzyme Regul., 22: 27–55, 1984. metal binding site of bleomycin A2. A carbon-13 nuclear magnetic 28. Cheung, B. C., Sun, T. H., Leenhouts, J. M., and Cullis, P. R. resonance study of the zinc(II) and copper(II) derivatives. Biochemistry, Loading of doxorubicin into liposomes by forming Mn2ϩ-drug com- 17: 4090–4096, 1978. plexes. Biochim. Biophys. Acta, 1414: 205–216, 1998. 45. Suzuki, T., Kuwahara, J., and Sugiura, Y. DNA cleavages of 29. Mayer, L. D., Tai, L. C., Ko, D. S., Masin, D., Ginsberg, R. S., bleomycin-transition metal complexes induced by reductant, hydrogen Cullis, P. R., and Bally, M. B. Influence of vesicle size, lipid compo- peroxide, and ultraviolet light: characteristics and biological implica- sition, and drug-to-lipid ratio on the biological activity of liposomal tion. Nucleic Acids Symp. Ser., 161–164, 1984. doxorubicin in mice. Cancer Res., 49: 5922–5930, 1989. 46. Lim, H. J., Masin, D., Madden, T. D., and Bally, M. B. Influence of 30. Mayer, L. D., Nayar, R., Thies, R. L., Boman, N. L., Cullis, P. R., drug release characteristics on the therapeutic activity of liposomal and Bally, M. B. Identification of vesicle properties that enhance the mitoxantrone. J. Pharmacol. Exp. Ther., 281: 566–573, 1997. antitumour activity of liposomal vincristine against murine L1210 leukemia. Cancer Chemother. Pharmacol., 33: 17–24, 1993. 47. Madden, T. D., Harrigan, P. R., Tai, L. C., Bally, M. B., Mayer, L. D., Redelmeier, T. E., Loughrey, H. C., Tilcock, C. P., Reinish, 31. Boman, N. L., Mayer, L. D., and Cullis, P. R. Optimization of the L. W., and Cullis, P. R. The accumulation of drugs within large retention properties of vincristine in liposomal systems. Biochim. Bio- unilamellar vesicles exhibiting a proton gradient: a survey. Chem. Phys. phys. Acta, 1152: 253–258, 1993. Lipids, 53: 37–46, 1990. 32. Deleers, M., and Malaisse, W. J. Ionophore-mediated calcium ex- 48. Barranco, S. C., Gerner, E. W., Burk, K. H., and Humphrey, R. M. change diffusion in liposomes. Biochem. Biophys. Res. Commun., 95: 650–657, 1980. Survival and cell kinetics effects of adriamycin on mammalian cells. Cancer Res., 33: 11–16, 1973. 33. Bouma, J., Beijnen, J. H., Bult, A., and Underberg, W. J. Anthra- cycline antitumour agents. A review of physicochemical, analytical and 49. Kim, S. H., and Kim, J. H. Lethal effect of adriamycin on the stability properties. Pharm. Weekbl. Sci., 8: 109–133, 1986. division cycle of HeLa cells. Cancer Res., 32: 323–325, 1972. 34. Harrigan, P. R., Hope, M. J., Redelmeier, T. E., and Cullis, P. R. 50. Hill, B. T., and Whelan, R. D. Comparative cell killing and kinetic Determination of transmembrane pH gradients and membrane potentials effects of vincristine or vindesine in mammalian cell lines. J. Natl. in liposomes. Biophys. J., 63: 1336–1345, 1992. Cancer Inst., 67: 437–443, 1981. 35. Bally, M. B., Nayar, R., Masin, D., Hope, M. J., Cullis, P. R., and 51. Hill, B. T., and Whelan, R. D. Comparative effects of vincristine Mayer, L. D. Liposomes with entrapped doxorubicin exhibit extended and vindesine on cell cycle kinetics in vitro. Cancer Treat Rev., 7(Suppl. blood residence times. Biochim. Biophys. Acta, 1023: 133–139, 1990. 1): 5–15, 1980. 36. Mayer, L. D., Tai, L. C., Bally, M. B., Mitilenes, G. N., Ginsberg, 52. Friedland, M. L. Combinational Chemotherapy. In: M. C. Perry R. S., and Cullis, P. R. Characterization of liposomal systems containing (ed.), The Chemotherapy Source Book. Baltimore, Maryland: Williams doxorubicin entrapped in response to pH gradients. Biochim. Biophys. and Wilkins, 1992. Acta, 1025: 143–151, 1990. 53. Greco, F. A., and Hainsworth, J. D. Etoposide phosphate or etopo- 37. Vaage, J., Donovan, D., Mayhew, E., Uster, P., and Woodle, M. side with cisplatin in the treatment of small cell lung cancer: randomized Therapy of mouse mammary carcinomas with vincristine and doxoru- phase II trial. Lung Cancer, 12(Suppl. 3): S85–S95, 1995.

Downloaded from clincancerres.aacrjournals.org on September 28, 2021. © 2004 American Association for Cancer Research. In Vitro and in Vivo Characterization of Doxorubicin and Vincristine Coencapsulated within Liposomes through Use of Transition Metal Ion Complexation and pH Gradient Loading

Sheela A. Abraham, Cheryl McKenzie, Dana Masin, et al.

Clin Cancer Res 2004;10:728-738.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/10/2/728

Cited articles This article cites 41 articles, 8 of which you can access for free at: http://clincancerres.aacrjournals.org/content/10/2/728.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/10/2/728.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/10/2/728. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.