Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Liposome formulations developed in an effort to enhance the therapeutic properties of anticancer drugs traditionally focused on lipid compositions that exhibit decreased plasma elimination rates and slow drug release rates (Mayer et al., 1989, 1993; Boman et al., 1994; Gabizon et al., 1996). This strategy was pursued based on a putative biological mechanism relying on the inherent ability of liposomes to be preferentially taken up in disease sites such as tumors (Proffitt et al., 1983; Mayer et al., 1990; Allen et al., 1991; Bally et al., 1994). This uptake occurs as a consequence of increases in tumor blood vessel permeability, typically seen in newly formed blood vessels that arise in response to factors such as vascular endothelial growth factor/vascular permeability factor (VEGF/VPF) (Folkman, 1985; Dvorak et al., 1988, 1991). Emphasis on maximizing liposome-mediated drug delivery to tumors has resulted in the selection of formulations designed to retain drug well after administration; however, it is important to establish a balance between liposome-mediated drug delivery and free drug access to tumor cells. The latter can only be attained through approaches that affect drug release rates and/or cell specific delivery of the liposome and its entrapped drug.

With liposomal formulations of mitoxantrone differing in their in vivo drug-retention characteristics, it was demonstrated that drug release was required for optimal therapeutic activity when the tumor growth was restricted primarily to the liver and spleen (Lim et al., 1997). This study addressed a hypothesis that drug release is the dominating factor controlling biological activity of liposomal drugs in tissues where the rate of liposome accumulation was rapid. The experiments summarized herein addressed the question of whether drug release or liposome-mediated drug delivery becomes the dominant factor controlling therapeutic activity under conditions where the rate of liposome accumulation is slow and tumor development is outside the liver.

We developed these mechanistic studies with liposomal encapsulated mitoxantrone for several reasons. Mitoxantrone biodistribution and elimination parameters are dictated solely by attributes of the liposomal carrier (Lim et al., 1997). This is in contrast to liposomal vincristine and doxorubicin, where entrapped drug changes the pharmacokinetic behavior of the liposome. This effect has been attributed to a direct toxicity of the encapsulated drug on phagocytic cells that play an important role in effecting liposome elimination from the plasma (Bally et al., 1990; Daemen et al., 1995).

Of course, it is also important to note that mitoxantrone is a clinically relevant drug and that its use is indicated for patients with: 1) breast cancer (locally advanced and metastatic), 2) relapsed adult leukemia, and 3) lymphoma or hepatoma (Smith et al., 1983; Durr, 1984). In combination with other drugs, mitoxantrone is used in the treatment of prostate cancer and nonlymphocytic leukemia (Bloomfield et al., 1998; Smith, 1999). It is less cardiotoxic than doxorubicin (Dukart et al., 1985; Neidhart et al., 1986; Bennett et al., 1988; Weiss, 1989) and has proved to be a suitable substitute for doxorubicin in clinical settings where alopecia and/or cardiotoxicity are concerns (Neidhart et al., 1986; Bennett et al., 1988; Weiss, 1989).

Two liposomal formulations of mitoxantrone [1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC)/cholesterol (Chol) and 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)/Chol] were evaluated in animals bearing tumors established after s.c. injection of human LS180 and A431 cell lines. These cell lines were selected on the basis of empirical observations that indicated more rapid liposome uptake in LS180 tumors compared with A431 tumors (Lim, 1999). The results suggest that delays in tumor growth induced by liposomal mitoxantrone were achieved with a liposomal formulation that was selected on the basis of drug release attributes, even when the liposome accumulation rate in the site of tumor growth was slow.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Novantrone (mitoxantrone hydrochloride) was obtained from the British Columbia Cancer Agency and is a product of Wyeth Ayerst, Canada (Montreal, Canada). DSPC and DMPC were purchased from Avanti Polar Lipids (Alabaster, AL). HEPES, citric acid, cholesterol, and Sephadex G50 (medium) were purchased from Sigma Chemical Company (St. Louis, MO). Dibasic sodium phosphate was obtained from Fisher Scientific (Fairlawn, NJ). Solvable was obtained from NEN Research Products (DuPont Canada, Mississauga, Canada). [¹⁴C]mitoxantrone, used as a tracer, was generously donated by Wyeth Ayerst, Canada. The specific activity of the mitoxantrone at the time of synthesis was 6.3 Ci/mmol. [³H]cholesteryl hexadecyl ether (40-60 Ci/mmol), a lipid marker that is not exchanged or metabolized in vivo (Stein et al., 1980), was purchased from Amersham (Oakville, Canada). Aquacide II was purchased from Terochem Laboratories Ltd. (Edmonton, Canada). LS180 and A431 tumor cells were purchased from the American Type Culture Collection (ATCC; Manassas, VA) and maintained in culture. Female SCID/RAG 2 mice were bred at the British Columbia Cancer Agency Animal Breeding Facility (Vancouver, Canada).

Preparation of Liposomes. DSPC/Chol (55:45, mol/mol) and DMPC/Chol (55:45, mol/mol) liposomes were prepared via extrusion technology (Hope et al., 1985). The indicated phospholipid and cholesterol mole ratios were dissolved in chloroform, and [³H]cholesteryl hexadecyl ether (CHE) was added (1-5 µCi/100 mg of lipid) before drying down the mixture to a homogenous lipid film under a stream of nitrogen gas. This lipid film was dried under vacuum for 3 h to remove any residual chloroform. Subsequently, the lipid film was hydrated in a 300 mM citric acid buffer (pH 4.0) to a final lipid concentration of 100 mg/ml. The resulting multilamellar vesicle mixture was frozen and thawed five times (Mayer et al., 1986) and extruded through three stacked 100-nm polycarbonate filters (Nuclepore, Pleasanton, CA) with an extrusion device (Lipex Biomembranes Inc., Vancouver, Canada). The resulting large unilamellar vesicles were sized by quasielastic light scattering with a Nicomp 270-submicron particle sizer (Pacific Scientific, Santa Barbara, CA) operating at 632.8 nm. The mean diameter of these liposomes was 100 to 120 nm.

Transmembrane pH Gradient Loading of Mitoxantrone. Mitoxantrone was encapsulated as described previously (Lim et al., 1997) with a transmembrane pH gradient-driven loading procedure (Mayer et al., 1985; Madden et al., 1990). The procedure used was analogous to that used for vincristine (Boman et al., 1993) and consisted of adding sufficient mitoxantrone to achieve a final drug-to-lipid weight ratio of 0.1. The pH of this mixture was then increased from 4.0 to 7.2 by the addition of 0.5 M Na₂HPO₄ buffer to the drug-liposome mixture. The resulting mixture was incubated at 65°C for an additional 15 min, which resulted in >98% encapsulation of the drug. For in vivo studies that included radiolabeled mitoxantrone as a tracer, between 0.2 and 0.5 µCi [¹⁴C]mitoxantrone was added per milligram of drug before drug loading.

Tumor Accumulation and Plasma Elimination Studies of Liposomal Mitoxantrone. Female SCID/RAG 2 mice (18-20 g, four per group) mice were inoculated with 2 × 10⁶ A431 cells (a human squamous cell carcinoma cell line obtained from ATCC) or 1 × 10⁶ LS180 cells (a human colon cell carcinoma cell line obtained from ATCC) s.c. on the hind regions of the back. Each animal was injected bilaterally with the tumor cell lines, but only one cell line was given to each animal. Once the tumors reached an estimated mass of 0.5 g, mice were injected with a 10-mg/kg dose of free mitoxantrone, DSPC/Chol mitoxantrone, or DMPC/Chol mitoxantrone via the lateral tail vein. At 1, 4, 24, 48, and 96 h, animals were terminated by CO₂ asphyxiation, and whole blood was collected via cardiac puncture and placed into EDTA-coated tubes (Microtainers; Becton Dickinson, Lincoln Park, NJ). Plasma was isolated after centrifugation of whole blood at 500g for 10 min. As indicated elsewhere (Bally et al., 1993), 100% of the liposomes were retained in the plasma compartment, and no interaction with blood cells could be measured. An assessment of drug binding to serum proteins was not made. Aliquoted plasma samples (100 µl) were mixed with 5 ml Pico Fluor 40 (Packard, Meriden, CT), and [³H] and [¹⁴C] was measured with a Packard 1900 scintillation counter.

Liposomal Mitoxantrone Antitumor Efficacy with the Human A431 and LS180 Solid-Tumor Model. SCID/RAG 2 mice were inoculated with 2 × 10⁶ A431 or LS180 cells 14 days before initiation of drug treatment. Each animal was given only one tumor cell line. Tumor-bearing animals (tumor size >0.050 but <0.200 cm³) were given a single i.v. injection of free mitoxantrone, DSPC/Chol liposomal mitoxantrone, or DMPC/Chol liposomal mitoxantrone. Control mice were treated with 0.9% saline. Based on drug dose titrations from 5 to 20 mg/kg (data not shown), the maximum therapeutic dose of drug when given as a single i.v. injection was defined as 5 and 10 mg/kg for the free and liposomal drugs, respectively. Animal weights and tumor volumes were measured daily until the tumor mass exceeded 10% of the animals' original body weight or until the tumors showed any sign of ulceration. Tumor volume was determined by measuring tumor dimensions (Tomayko and Reynolds, 1989) and calculating volume with the equation (/6) × length × width².

Treatment of LS180 and A431 Tumors Before Formation of Measurable Tumors. To establish optimal conditions for treating SCID/RAG 2 mice inoculated with LS180 and A431 cells, studies evaluated treatment of animals 2 days after tumor cell inoculation was completed. There are several reasons for completing these studies, including a need to evaluate treatment when the tumor burden is small and blood vessel formation in response to tumor factors has probably not begun. Treatment was based on single (5 mg/kg of free drug and 10 mg/kg of liposomal drug) and multiple (1.5 mg/kg of free drug and 3.5 of mg/kg liposomal drug) doses. The latter consisted of an i.v. injection on days 2, 3, and 4. Other dose schedules were evaluated (e.g., days 2, 4, and 6; days 2, 6, and 10), but under the conditions used, the most significant therapeutic responses were obtained with the days 2, 3, and 4 schedule.

Modified Microculture Tetrazolium Assay. The modified microculture tetrazolium assay was used to assay the IC₅₀ (drug concentration that caused 50% inhibition of cell proliferation/cytotoxicity) of mitoxantrone, doxorubicin, and vincristine on LS180 and A431 cells. Briefly, cells harvested from exponential phase cultures were counted by trypan blue exclusion (only cell preparations demonstrating viability >90% were used) and dispensed within 96-well flat-bottomed Costar (Cambridge, MA) culture plates (2000 cells/100 µl per well for a 3-day incubation). These cells were exposed to serial concentrations of the indicated drug (12.5-0.001 µM in 100 µl of culture medium, RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum), over a 3-day incubation at 37°C, 5% CO₂, and 100% relative humidity. After incubation, 50 µl of modified microculture tetrazolium {5 mg [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide]/ml PBS, filtered through 0.45-mm filter units, and stored at 4°C for 1 month} was added and further incubated for 4 h at 37°C. Medium was aspirated slowly through a blunt 18-gauge needle leaving ~20 µl of the supernatant, and the reaction product thoroughly solubilized with 150 µl of dimethyl sulfoxide. The plates were read spectrophotometrically at 570 nm in a Titertek multiscan multiplate reader (Flow Laboratories, Mississauga, Canada). Cytotoxicity was expressed in terms of percentage of control absorbance (mean ± S.D.) after subtraction of background absorbance. The IC₅₀ was determined from a linear regression curve (plot of percent control absorbance versus log drug concentration) of the data from studies completed in triplicate.

Statistical Analysis. ANOVA was performed, with the Statistica software package (StatSoft Inc., Tulsa, OK), on the results obtained after administration of the two liposomal formulations and free mitoxantrone. Common time points were compared with the post hoc comparison of means, Scheffé test. Differences were considered significant at P < .05. Area under the curve (AUC) analysis was performed with trapezoidal integration from 0 to 96 h.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Lipid and Drug Accumulation in Solid LS180 and A431 Tumors. Lipid and drug levels were measured in established (0.05-cm³) A431 and LS180 solid tumors, grown in SCID/RAG 2 mice, over a 96-h period after a single i.v. injection of free mitoxantrone (10 mg/kg), DSPC/Chol liposomal mitoxantrone (10 mg drug/kg, 100 mg total lipid/kg), and DMPC/Chol liposomal mitoxantrone (10 mg drug/kg, 100 mg total lipid/kg), and the results are summarized in Fig. 1. The level (µg lipid/g tumor) of liposomal lipid in the LS180 and A431 tumors is shown in Fig. 1, A and B, respectively, and the tissue concentration (µg drug/g tumor) of mitoxantrone in the LS180 and A431 tumors is shown in Fig. 1, C and D, respectively.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Lipid and mitoxantrone accumulation in A431 and LS180 tumors in SCID/RAG 2 mice over a 96-h period. SCID/RAG 2 mice were injected bilaterally with 2 × 106 A431 cells or 1 × 106 LS180 cells s.c. Once the tumors reached a size of ~0.05 to 0.2 cm3, mice were injected with 10 mg/kg of free mitoxantrone (), DSPC/Chol mitoxantrone (), or DMPC/Chol mitoxantrone () via the lateral tail vein. Mice were terminated by CO2 asphyxiation, and tumors were removed and processed as described in Experimental Procedures. After addition of Pico-Fluor 40 to a defined volume of the tumor homogenate, radioactivity ([3H]CHE and [14C]mitoxantrone tracer) was determined with the scintillation counter. All tissue drug and lipid levels were corrected for drug and lipid in the plasma compartment by use of published plasma volume correction factors. A and B, lipid accumulation in LS180 and A431 tumors, respectively. C and D, drug accumulation in LS180 and A431 tumors, respectively. Data points represent the average and S.E.M. of at least four animals.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Percentage of initial drug-to-lipid ratio of DSPC/Chol mitoxantrone and DMPC/Chol mitoxantrone after 48 h in plasma. SCID/RAG 2 mice were injected bilaterally with 2 × 106 A431 cells or 1 × 106 LS180 cells s.c. Once the tumors reached a size of ~0.05 cm3, mice were injected with 10 mg/kg of DSPC/Chol mitoxantrone (shaded columns) or DMPC/Chol mitoxantrone (open columns) via the lateral tail vein. Plasma and tumors were collected and processed as outlined in Experimental Procedures. A, drug-to-lipid ratio in plasma. B, drug-to-lipid ratio in the tumor. Data points represent the average and S.E.M. of the data collected from at least four animals.

1

1

1

1

1

1

1

1

Efficacy of Single-Dose Administration of Liposomal and Free Mitoxantrone in Established A431 and LS180 Human Solid Tumors. In previous studies, we demonstrated that treatment of mice bearing L1210 and P388 liver tumors with DMPC/Chol liposomal mitoxantrone resulted in 100% long-term survivors (Lim et al., 1997). Although the DSPC/Chol liposomal mitoxantrone formulation delivered more mitoxantrone than the DMPC/Chol formulation to the tumor site, treatment with this formulation proved to be less effective. It is important to determine whether these carrier-associated differences in mitoxantrone efficacy extend to solid tumors. The A431 and LS180 tumors provided suitably different liposome uptake characteristics so that comparisons between the liposomal formulations could be made. However, note that the selected tumor cells exhibit different growth characteristics and drug sensitivity (Table 2). Particularly, the LS180 tumors grow approximately two times faster than the A431 tumors. In contrast to the A431 tumors, LS180 tumors are highly vascularized, and the LS180 cells are about five times more sensitive to free mitoxantrone than A431 cells. Gross observations indicate that the LS180 tumors are less cohesive than the A431 tumor and that the LS180 tumors ulcerate more rapidly than A431 tumors.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Efficacy of DSPC/Chol mitoxantrone, DMPC/Chol mitoxantrone, and free mitoxantrone in established LS180 and A431 solid tumors in SCID/RAG 2 mice. Mice were injected with 1 × 106 LS180 cells (A) or 2 × 106 A431 cells (B) s.c. Fourteen days after tumor cell inoculation (tumor size >0.05 cm3), mice were injected with 5 mg/kg of free mitoxantrone (), 10 mg/kg of DSPC/Chol mitoxantrone (), or 10 mg/kg of DMPC/Chol mitoxantrone () via the lateral tail vein. Control mice were injected with saline (). Tumor width and length were measured with calipers, and volume was calculated as outlined in Experimental Procedures. Points represent average data and the S.E.M. from at least six tumors.

Efficacy of Multiple-Dose Administration of Liposomal and Free Mitoxantrone in A431 and LS180 Human Solid Tumors. The studies summarized in Fig. 3 were obtained when mice with well-established tumors were treated with the different mitoxantrone formulations. It can be argued that optimal therapy should be observed when treating tumors at a time point before formation of a measurable tumor and through use of repeated injections of the drug. To address this, mice were treated with single and multiple doses of free and DMPC/Chol liposomal mitoxantrone 2 days after tumor cell inoculation. The results of these studies are summarized in Table 3. For simplicity, the table reports results as the "day of initiation of tumor growth", a parameter determined by taking a linear least-squares analysis of tumor volumes during the rapid-growth phase and extrapolating to a tumor volume of zero. The effect of mitoxantrone treatment can then be determined as a delay in initiation of tumor growth. This analysis relies on the assumption that treatment does not alter the growth rate of the tumor once it is established (i.e., tumor volume in excess of 0.05 cm³ is attained).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

We demonstrated in a previous study with liposomal formulations that exhibit comparable plasma elimination rates that slow drug release results in reduced antitumor activity compared with liposomes that release drug rapidly (Lim et al., 1997). This conclusion was reached by comparing the antitumor activity of mitoxantrone encapsulated in DSPC/Chol and DMPC/Chol liposomes after i.v. administration to mice bearing tumors residing primarily in the liver and spleen. The studies reported here were initiated because of concerns that this conclusion was only applicable to disease localized to the liver, a site where significant and rapid accumulation of liposomes is observed after i.v. administration. To address this concern, the therapeutic activity of DSPC/Chol and DMPC/Chol liposomal mitoxantrone was measured with two human xenograft models grown as ectopic (s.c.) tumors. The results are considered by focusing this discussion on three important points, all critical if the central hypothesis is to be sustained, including: 1) the role of liposome delivery and liposome tumor/host cell interactions, 2) differences in drug-targeting efficiencies between free and liposomal drug, and 3) the importance of considering capillary endothelium permeability to circulating macromolecules as well as capillary density within a tumor.

The 3-fold increase in drug exposure achieved with liposomal formulations of mitoxantrone (Table 2) resulted in improvements in antitumor effects (see Fig. 3). However, the results presented in this study do not support the notion that the greatest therapeutic activity would be obtained with liposome formulations that facilitate the greatest increase in tumor drug AUC. The AUC_D values obtained after administration of DMPC/Chol mitoxantrone were 25 to 40% lower then those obtained for DSPC/Chol for the A431 and LS180 tumors, respectively. Despite reduced drug delivery, the DMPC/Chol liposomal mitoxantrone formulation was therapeutically better than the DSPC/Chol formulation when treating LS180 tumors (Fig. 3A). Treatment of A431 tumors suggested that the DMPC/Chol is as active as the DSPC/Chol formulation (Fig. 3B).

Drug AUC values in solid tumors are dependent on the dose of lipid, the liposome plasma elimination rate, and the drug-retention characteristics of the liposome. The latter are illustrated in Fig. 1, where it is demonstrated that comparable liposomal lipid accumulation does not result in comparable drug uptake levels. In this example, reduction in mitoxantrone uptake is partially a consequence of drug release from the DMPC/Chol liposomes. This explanation, however, is a simplistic analysis that does not account for the accumulation of drug released from liposomes in the plasma compartment or from other tissues that are accumulating and metabolizing liposomes. It has been proposed, for example, that the liver is capable of acting as a drug reservoir, where macrophage processing of drug-loaded liposomes can result in drug release back into the circulation (Storm et al., 1988). Indications of free (released)-drug accumulation in tumors after i.v. administration of a liposomal drug have been based on comparisons between the estimated drug-to-lipid ratio in the plasma compartment versus the tumor. As shown in Fig. 2, the ratio of tumor drug-to-lipid ratio to plasma drug-to-lipid ratio at 48 h after administration of the DSPC/Chol mitoxantrone is ~0.92 for both tumors. A similar analysis for the DMPC/Chol mitoxantrone formulation results in a ratio of 2.5 for A431 tumors and 3.8 for LS180 tumors. A ratio of >1 suggests that more drug is present in the tissue than would be predicted on the basis of liposome accumulation from the plasma.

The higher ratios observed in tumors after administration of DMPC/Chol mitoxantrone are most likely a consequence of released drug accumulation. This can be suggested on the basis of T_e, a value that is determined by dividing the AUC_D in the tumor by the AUC_D in the plasma compartment (see Table 1). The T_e for free mitoxantrone is at least 8-fold greater than that measured for the liposomal formulations and is a consequence of differences in size between free drug and the liposomal drug. The free drug is small and readily distributes after i.v. administration; hence, the T_e for free drug is large. Because drug is released from DMPC/Chol liposomes while in the plasma compartment, it is reasonable to assume that this drug could be efficiently taken into the tumor.

It was demonstrated in this study that the rate and extent of liposome accumulation in tumors is also dependent on the type of tumor, which is probably a function of the tumor's specific attributes such as capillary density and structure. In the LS180 tumor model, extravasation occurs rapidly, reaching the C_max within 4 h after administration (Fig. 2A). In contrast, in the A431 tumor model, the C_max is achieved 48 h after i.v. administration. Almost twice the amount of liposomal lipid accumulates within the LS180 tumor (AUCs of 10,167.32 and 9925.82 µg lipid · g¹ tumor · h¹ for the DSPC/Chol and DMPC/Chol formulations, respectively) compared with the A431 tumors (AUCs of 5728.22 and 5149.66 µg lipid · g¹ tumor · h¹ for the DSPC/Chol and DMPC/Chol formulations, respectively). Gross inspection of the tumors suggests that the LS180 tumor is better vascularized than the A431 tumor (Table 2), which may account for differences in the rate of accumulation. Liposome extravasation will depend on tumor microvascular density as well as capillary endothelium permeability. The increased microvascular density would lead to greater delivery of liposomes to the site of tumor growth. In addition, the extravasation of liposomes is dependent on the permeability of the blood vessel. An increase in the permeability (due to secreted factors such as VEGF) could also result in increased liposome accumulation; however, release of VEGF from the A431 tumor cells was not sufficient to generate a highly vascular tumor in the SCID mice.

A discussion relating microvascular density and endothelium permeability invites consideration of whether liposome extravasation is a relevant parameter when studying tumors before establishment of a significant tumor burden. For the LS180 and A431 tumors studied, measurable tumors were obtained 12 to 15 days after tumor cell inoculation. It would be unexpected to see significant vascularization of the tumors shortly after cell inoculation, although no direct measurement of tumor vascularization was made in these studies. Clearly, it is important to develop methodologies that can measure liposomal lipid and drug levels in areas where tumor growth is initiating, particularly considering the fact that most studies evaluating liposome extravasation use large tumors that may have the greatest microvascular density and the most permeable blood vessels.

A fundamental element of the central hypothesis is that drug encapsulated inside the liposome is not bioavailable. Furthermore, the liposome-encapsulated drug is not therapeutically active unless a feature promoting tumor cell delivery is incorporated. This may involve use of targeting ligands that are known to be internalized, for example, the folate acid receptor (Lee and Low, 1993, 1994; Wang et al., 1995). In addition, noninternalized targets have also been used in an effort to specifically deliver the drug to tumor and release drug in the vicinity of the tumor cells (Longman et al., 1995; Scherphof et al., 1997). Alternatively, the liposomes can be designed to nonspecifically bind and fuse with cells after extravasation into a site of tumor growth. An elegant example of this approach, resulting in a lipid-based delivery system referred to as programmable fusogenic liposomes, has recently been described (Adlakha-Hutcheon et al., 1999). In the absence of cell delivery, cell fusion, and/or intracellular processing by phagocytic cells in the site of extravasation, the encapsulated drug must, however, be released from the liposomes to maximize drug bioavailability and therapeutic activity.

In conclusion, to maximize the benefits of using liposomal carriers, a balance between delivery and drug release must be achieved. It has been argued that the primary source of drug within the tumor is from liposomes that have extravasated into the site (Mayer et al., 1994)an argument that links the rate and extent of liposome accumulation and the rate of drug release to therapeutic activity. However, the possibility that drug release from sites that are distinct from the tumor may contribute to the therapeutic activity cannot be excluded. This is perhaps most important when the tumor burden is small and vascularization is low. The results suggest that a conventional (nontargeted, nonfusogenic) formulation of mitoxantrone prepared with DMPC/Chol liposomes is active in the treatment of ectopic tumors and tumors progressing primarily in the liver and spleen (Lim et al., 1997). This activity is believed to be a consequence of the rate at which mitoxantrone is released from DMPC/Chol liposomes. The DMPC/Chol formulation of mitoxantrone is particularly well suited for treatment of tumors (or sites of tumor growth) where liposome accumulation is rapid. Future studies will focus on the antitumor effects of DMPC/Chol mitoxantrone when used to treat cancer within the liver.