Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The ability of liposomes to decrease the toxicity of doxorubicin while maintaining or increasing its antitumor potency has been well documented for a variety of liposomal systems in preclinical models (Gabizon et al., 1985; Mayer et al., 1989, 1990a; Huang et al., 1992b) as well as Phase I and Phase II clinical trials (Cowens et al., 1993; Uziely et al., 1995; O'Day et al., 1994). It is generally accepted that the buffering of toxicity observed with liposomal doxorubicin formulations is related to the fact that lipid carriers reduce 1) free drug peak plasma concentration and 2) exposure to susceptible tissues such as the heart (Gabizon et al., 1982, 1989; Mayer et al., 1990b). The latter observation reflects the low uptake of small (100 nm) PC/Chol-based liposomes by these tissues. The mechanism(s) whereby liposomes improve the antitumor activity of doxorubicin, however, are less well understood. This may be caused by the fact that few studies have focused on the processes whereby liposomes gain access to tumor sites and, perhaps more importantly, what factors control exposure of malignant cells to encapsulated drug.

Several potential mechanisms have been postulated for the improved antitumor activity of liposomal doxorubicin compared with unencapsulated drug (Gabizon, 1994). First, because the most efficacious liposomal formulations display increased circulation lifetimes, these systems could act as slow release or systemic infusion delivery vehicles in the blood compartment. Although doxorubicin released from liposomes in the circulation will certainly contribute to therapeutic activity, the fact that liposomes displaying remarkably different plasma drug leakage rates provide similar levels of therapy (Mayer et al., 1994) suggests that release in the blood compartment may not be the most critical factor governing antitumor activity. Second, liposomes may deliver doxorubicin directly to tumors through a passive accumulation process. Subsequent exposure of the tumor cells to the encapsulated drug would therefore rely on drug release from regionally localized liposomes because it seems that nontargeted liposomes are incapable of directly delivering their contents to tumor cells (Working et al., 1994). Third, circulating macrophages may mediate delivery of liposomal drugs and, finally, TAMs may promote liposome delivery to the tumor.

Recent investigations, including elegant imaging studies by Jain and co-workers (Yuan et al., 1994), have demonstrated that intact liposomes are capable of extravasating the tumor vasculature and accumulating in the interstitial space within tumors and areas of tumor growth (Yuan et al., 1994; Mayer et al., 1990a; Huang et al., 1992a; Wu et al., 1993). The processes whereby liposomes cross vascular endothelial barriers as well as the fate of the lipid carriers and their entrapped contents after extravasating into the tumor interstitium, however, are not well understood. In particular, the mechanisms that lead to the release of entrapped anticancer drugs from liposomes, which is presumably required for action on the intracellular tumor target, have not been elucidated. Previous studies have suggested that engulfment and subsequent intracellular degradation of liposomal carriers by macrophages can render encapsulated agents such as doxorubicin bioavailable to surrounding tissue or the central blood compartment (Storm et al., 1988). Given that many human malignancies exhibit significant levels of infiltrated macrophages, studies with two murine fibrosarcoma lines differing primarily in their TAM level were undertaken here to determine whether the presence of phagocytic cells control enhanced tumor drug accumulation observed after i.v. administration of liposomal formulations of doxorubicin.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[poly(ethylene glycol) 2000] (PEG₂₀₀₀-DSPE) were purchased from Avanti Polar Lipids (Alabaster, AL). Chol was obtained from Sigma Chemical Co. (St. Louis, MO). Radiolabeled ³H-CHDE was purchased from Amersham (Oakville, ON) and DOX from Adria Laboratories (Mississauga, ON). FITC-labeled rat antimouse CD11b (MAC-1) antibody was purchased from Pharmingen (San Diego, CA) and Tissue-Tek O.C.T. compound from Miles (Etobicoke, ON). Dulbecco's Modified Eagle's Medium and Hanks' Balanced Salt Solution were purchased from StemCell Technologies, Inc. (Vancouver, Canada). Tissue culture grade fetal bovine serum, trypsin and glucose were obtained from ICN Biomedicals (Costa Nesa, CA). Solvable was obtained from NEN (Mississauga, ON), Pico-fluor 40 scintillation cocktail from Packard and microtainer EDTA-coated tubes from Becton Dickinson (Mississauga, ON). Female C3H/HeJ mice were obtained from the Joint Animal Care Facility breeding colony at the BC Cancer Research Centre. Parental stock were obtained from Jackson Laboratories (Bar Harbor, ME).

Preparation of liposomes. Large unilamellar vesicles (100 nm) were prepared as described previously (Mayer et al., 1986). Lipid mixtures consisting of DSPC, Chol and PEG₂₀₀₀-DSPE (50:45:5, mole %) were prepared in chloroform and subsequently concentrated to a homogeneous lipid film under a stream of nitrogen gas. The lipid film was then placed under high vacuum for at least 4 h before hydration at 65°C with 300 mM citrate, pH 4.0. The resulting multilamellar vesicle preparation was frozen and thawed five times (Mayer et al., 1985) before extruding the sample 10 times through stacked 100-nm polycarbonate filters (Poretics) with an extrusion device (Lipex Biomembranes, Inc., Vancouver, Canada) at 65°C. The resulting liposomes were sized by QELS with a Nicomp 270 submicron particle sizer operating at 632.8 nm.

DOX encapsulation. DOX was encapsulated in liposome preparations with the transmembrane pH gradients as described previously (Mayer et al., 1989). The liposome preparation (prepared at pH 4.0) was heated to 65°C for 10 min before addition to a preheated (65°C for 10 min) solution of DOX (5-6 mM in saline). A final drug-to-lipid ratio of 0.2 (w/w) was typically employed. This mixture was incubated with periodic mixing for 10 min at 65°C. Unencapsulated DOX was removed by passing the sample through a Sephadex G-50 column and the DOX-to-lipid ratio was measured by fluorescence detection at 550 nm as described previously (Bally et al., 1994).

Murine tumor models. The cell line designated as FSa-R is a highly immunogenic, methylcholanthrene-induced fibrosarcoma and the FSa-N cell line is a "spontaneous" fibrosarcoma arising in a nonirradiated part of an irradiated mouse (Howie et al., 1982). Cell lines were maintained in culture with Dulbecco's Modified Eagle's Medium supplemented with 4500 mg/l glucose and 10% fetal bovine serum. Cells were passaged (1:10) every 3 to 4 days after removal of adherent cells with a 2.5% trypsin solution. For tumor initiation, harvested cells were centrifuged at 250 × g for 5 min and resuspended in Hanks' Balanced Salt Solution at a concentration of 2 × 10⁶ cells/ml. Female C3H/HeJ mice weighing 18 to 22 g were then injected in both the left and right flank with 1 × 10⁶ cells (50 µl). When tumors grew to approximately 200 mg (12 and 18 days for FSa-R and FSa-N, respectively) tumor-loading studies were initiated. Large unilamellar vesicles were injected via the lateral tail vein at a lipid and drug dose of 100 and 20 mg/kg, respectively (0.2 drug/lipid). For tumor efficacy studies, tumor weights were determined by converting caliper measurements to grams with the formula; (length × width²)/2 (Mayer et al., 1990a) which was confirmed by comparing the actual weights of excised tumors with those estimated from caliper measurements. Treatment was given i.v. when tumors reached a size of 200 to 500 mg. Statistical significance in tumor growth was determined by analysis of variance analysis with the statistical software program Statistica.

Preparation of plasma and tissues. Mice were anesthetized with an intraperitoneal injection (100 µl) of a ketamine/xylazine mixture (160 and 20 mg/kg, respectively). Whole blood was collected via cardiac puncture and immediately placed in EDTA-coated Microtainer tubes. Plasma was subsequently prepared by centrifugation at 1500 × g for 10 min. One tumor from each mouse was prepared for histology (see below), the second tumor as well as lung, liver, spleen, kidney and heart were collected, weighed and homogenized to a 10% homogenate (w/v in H₂0) with a Polytron tissue homogenizer.

Quantitation of liposomal lipid and doxorubicin. Liposomal lipid was quantified by use of the nonexchangeable and nonmetabolizable lipid marker ³H-CHDE (Bally et al., 1994). Solvable (0.5 ml) was added to 200 µl of the 10% tissue homogenates. Samples were then incubated in a water bath at 50°C for 3 h. After cooling to room temperature, 50 µl of EDTA (200 mM), 200 µl H₂O₂ (30%) and 25 µl HCl (10 N) were added. After 1 h at room temperature 5 ml of Pico-fluor 40 scintillation cocktail was added and samples were analyzed in a Packard Tri-carb 1900 TR liquid scintillation analyzer. For plasma samples, 100 µl of plasma was added directly to the Pico-fluor 40 and analyzed.

Tumor fixation and staining. Isolated tumors (approximately 200 mg) were collected in PBS at 4°C. Tumors were subsequently fixed with a 3% paraformaldehyde solution in PBS at 4°C for 30 min (sufficient to fix through the tumor), washed with PBS and immersed in increasing sucrose gradients for 20 min each: 10% sucrose-PBS, 15% sucrose-PBS and 15% sucrose-PBS containing O.C.T. compound (1:1, v/v). The processed tissue was then embedded in O.C.T. compound and frozen in liquid nitrogen. Five-micrometer sections were then prepared on a Frigocut 2800E microtome from Leica.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Characterization of TAM levels in FSa-R and FSa-N tumors. The murine fibrosarcoma lines used here for characterizing the tumor-accumulating properties of liposomes were selected on the basis of the different levels of infiltrating macrophages displayed in vivo (Howie et al., 1982). Cellular isolates from FSa-N and Fsa-R tumor digests have been characterized for macrophage content by immunohistochemistry (G. Dougherty, unpublished observations). These studies revealed that the percent macrophage content in cells obtained from cytospins of FSa-N homogenates were 2.3-fold higher than observed for cytospins from FSA-R tumors. The higher TAM level in FSa-N tumors occur throughout tumor progression when growing as subcutaneous solid tumors in C3H/HeJ mice (Howie et al., 1982). Macrophage content comparisons were also determined here with frozen thin sections of FSa-R and FSa-N tumors that were stained with FITC-labeled anti-CD 11b antibody and visualized for positive cells by fluorescence microscopy. Comparison of three fields in three different tumors of each cell type revealed that the FSa-N tumor sections contained 28.8 ± 4.4 CD 11b positive cells per 400× magnification field, whereas FSa-R tumor sections contained 7.3 ± 2.4 CD 11b positive cells per field. This reflected a 3.9-fold increase in macrophage content in the FSa-N tumors relative to the FSa-R line. It should be noted that other studies (not presented here) demonstrated that both FSa tumor types had similar vascular plasma volumes (8 µl/g tumor). In vitro cytotoxicity studies demonstrated that the FSa-N line was 3-fold less sensitive to DOX than the FSa-R line (24 h continuous exposure DOX IC₅₀ values of 1 µM and 3 µM, respectively, data not shown).

Plasma elimination characteristics of DSPC/Chol and PEG-PE/DSPC/Chol liposomal systems. Plasma levels of DOX and liposomal lipid were determined after i.v. injection of 100 nm drug-loaded DSPC/Chol and PEG-PE/DSPC/Chol liposomes to C3H/HeJ mice. Comparisons were also made between tumor-bearing and tumor-free mice to fully characterize the pharmacokinetic and tumor accumulation properties of the liposomal systems. Figure 1A presents the plasma liposomal lipid levels in tumor-free mice after injection of 100 mg lipid/kg (equivalent to 20 mg/kg DOX dose). As indicated elsewhere (Parr et al., 1993) liposomes containing DOX (entrapped by the pH gradient loading technique) exhibited extended circulation lifetimes compared with empty liposome counterparts (results not shown). DOX-loaded liposomes containing PEG-PE exhibited circulation half-lives that were increased significantly over those observed for DSPC/Chol-formulated liposomal doxorubicin in the C3H/HeJ mice (18 and 3.5 h, respectively). Plasma concentrations of ³H-CHDE indicated that 13% and 37% of the injected dose of DSPC/Chol and PEG-PE/DSPC/Chol liposomes, respectively, were present in the circulation 24 h after i.v. injection (fig. 1A).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Plasma clearance of DOX-loaded DSPC/Chol () and PEG-PE/DSPC/Chol () 100 nm liposomes in normal (A) and Fsa-N tumor-bearing (B) C3H/HeJ mice (three per time point). DOX was encapsulated in the liposomes at a drug-to-lipid ratio of 0.2:1 (w/w). Liposomes were injected i.v. at a dose of 100 mg lipid/kg and DOX at a dose of 20 mg/kg.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Plasma clearance of DOX in C3H/HeJ mice (three per time point) bearing FSa-N tumors after i.v. injection of free and liposomal DOX at a drug dose of 20 mg/kg (100 mg lipid/kg for liposomal formulations). Mice were administered free drug () as well as DOX encapsulated in 100 nm DSPC/Chol () or PEG-PE/DSPC/Chol () liposomes.

Tumor lipid and drug accumulation properties. DOX and liposomal lipid levels in FSa-R and macrophage-enriched FSa-N tumors after i.v. injection were determined for DSPC/Chol and PEG-PE/DSPC/Chol formulations as well as free (unencapsulated) drug. The rapid distribution of free DOX from plasma is accompanied by maximum FSa-N tumor drug uptake levels of 12.8 µg/g 1 h after administration (fig. 3A). Tumor drug levels subsequently decreased gradually from this value to 11.5 µg/g and 7.6 µg/g at 4 h and 24 h, respectively. In contrast, DOX accumulates in FSa-N tumors at a slower rate but for a prolonged period of time when administered in liposome-encapsulated form. FSa-N tumor-associated DOX levels after injection of drug entrapped in DSCP/Chol or PEG-PE/DSPC/Chol liposomes were approximately 2.5-fold lower than those obtained for free DOX at 1 h, equivalent to free drug levels at 4 h and roughly 4-fold higher at 24 h (fig. 3A). This pattern of drug accumulation was similar to that observed for liposomal lipid accumulation in FSa-N tumors for both liposomal doxorubicin formulations (fig. 3B). Tumor-associated lipid levels of approximately 16 µg/g were observed at 1 h and increased to 45 to 60 µg/g and 130 to 140 µg/g at 4 h and 24 h post-injection, respectively. It is important to note that FSa-N tumor levels of liposomal lipid and DOX were not statistically (P < 0.05) different between DSPC/Chol and PEG-PE/DSPC/Chol systems throughout the entire time course.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   DOX (A) and liposomal lipid (B) accumulation in s.c. FSa-N tumors after i.v. injection of free DOX () or DOX encapsulated in 100 nm DSPC/Chol () or PEG-PE/DSPC/Chol () liposomes. Mice (four per time point) were injected with a DOX dose of 20 mg/kg and a lipid dose of 100 mg/kg.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Tumor-to-plasma DOX ratios observed in FSa-R and FSa-N tumors 24 h after i.v. injection of 100 nm DSPC/Chol (open bars) and PEG-PE/DSPC/Chol (solid bars) liposomes containing DOX at a drug-to-lipid ratio of 0.2:1 (w/w). Drug and liposomal lipid doses were 20 mg/kg and 100 mg/kg, respectively. Tumor-to-plasma ratios were calculated as the tumor DOX content (µg/g tumor) divided by the plasma concentration of DOX (µg/ml).

Studies on the mechanism of tumor drug and liposome accumulation. Given that TAM-enriched FSa-N tumors take up significantly more drug and lipid than FSa-R tumors and that PEG-PE-containing liposomes are less efficient in accumulating in these tumors, studies were undertaken to determine whether TAMs mediate tumor delivery of liposomal DOX, particularly for PEG-free liposomes. Figure 5 shows fluorescence micrographs of FSa-N tumor frozen sections of tissues derived 24 h after i.v. injection of DOX (20 mg/kg) encapsulated in DSPC/Chol (fig. 5, A and C) and PEG-PE/DSPC/Chol (fig. 5, B and D) liposomes. DOX fluorescence was monitored directly (fig. 5, A and B) and TAMs were visualized after staining with FITC-labeled anti-CD 11b antibody (fig. 5, C and D). The fluorescence micrographs presented reflect representative fields from tumor sections exhibiting DOX fluorescence; however, it should be noted that areas within the tumors could be found that exhibited negligible DOX fluorescence. Such drug-poor areas were not associated with any specific tumor ultrastructure and could be detected near the tumor periphery or the tumor interior. The pattern and intensity of DOX fluorescence was similar in drug-containing sections for DSPC/Chol- and PEG-PE/DSPC/Chol-based liposomal DOX formulations. Evidence of diffuse nuclear-associated DOX fluorescence as well as concentrated foci of drug were present (fig. 5, A and B). Direct coordinate comparisons (by caliper measurements) of DOX fluorescence and CD 11b positive cells revealed that no intense foci of drug fluorescence occurred in these cells identified as macrophages (fig. 5). In fact, CD 11b positive cells were often observed in positions devoid of fluorescence when visualized with optical filters specific for DOX fluorescence. Similar observations were made with frozen sections of FSa-R tumors (data not shown). It should be noted that the amount and distribution of DOX fluorescence was unaffected by the immunohistochemistry processing steps; DOX fluorescence appeared similar for sections visualized immediately after sectioning as well as after antibody-staining procedures. Further, liposomes containing DOX remained intact and fluorescent when processed through the cryofixation procedure, and isolated peritoneal macrophages that had phagocytosed liposomal DOX exhibited strong fluorescent images when exposed to the cryofixation conditions (data not shown). These results in conjunction with previous studies demonstrating fluorescent imaging of liposomal DOX accumulated in peritoneal macrophages (Bally et al., 1994) indicate that liposomal DOX is not actively taken up by macrophages residing in the fibrosarcoma tumors studied here.

View larger version (K): [in this window] [in a new window]   Fig. 5.   Fluorescence micrographs of FSa-N tumor frozen sections obtained from C3H/HeJ mice 24 h after i.v. injection of DOX (20 mg/kg) encapsulated in 100 nm DSPC/Chol (A and C) or PEG-PE/DSPC/Chol (B and D) liposomes. Frozen sections were stained with FITC-labeled anti-CD 11b antibody and visualized by optical filters specific for DOX (A and B) or FITC (C and D) fluorescence with a 40× objective lens.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Plasma and tumor drug-to-lipid ratios in C3H/HeJ mice (four per experimental point) bearing FSa-R (solid bars) and FSa-N (open bars) tumors 1 h after i.v. injection of DOX (20 mg/kg) encapsulated in 100 nm DSPC/Chol and PEG-PE/DSPC/Chol liposomes. Quantitation of DOX and liposomal lipid was accomplished as described under "Materials and Methods."

Therapeutic activity of free and liposomal DOX formulations against FSa-R and FSa-N solid tumors. The data shown to this point suggest that DOX delivery to both FSa-R and FSa-N tumors is enhanced when the drug is given in liposomal form. The amount of drug obtained in the tumor is similar for conventional liposomes (DSPC/Chol) and sterically stabilized liposomes (PEG/DSPC/Chol) that exhibit enhanced circulation lifetimes. In addition, the macrophage-enriched FSa-N tumors accumulate approximately 2-fold more drug than the FSa-R tumor. Clearly, it is important to determine whether formulation and/or tumor-induced differences in drug delivery result in a therapeutic advantage. Therefore, the efficacy of free DOX as well as DOX encapsulated in DSPC/Chol and PEG-PE/DSPC/Chol liposomes against FSa-R and FSa-N solid tumors growing s.c. in C3H/HeJ mice was evaluated. Mice bearing FSa-R and FSa-N tumors (50-250 mg) were treated with a single i.v. dose of 20 mg DOX/kg (100 mg lipid/kg for liposomal formulations) or saline for control animals. FSa-R tumors in control mice grew rapidly and reached sizes requiring termination of the animals within 12 days after initiation of therapy. Administration of free or liposomal DOX resulted in a significant delay in tumor growth (fig. 7A). The antitumor effects induced by free DOX and DOX entrapped in DSPC/Chol liposomes were similar throughout the course of the study, and the decreased tumor growth rates observed for the PEG-PE/DSPC/Chol liposomal DOX preparation are statistically significant (P < 0.05) compared with free and DSPC/Chol formulations only on days 14 and 16 post-treatment.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Antitumor activity of DOX (20 mg/kg) administered as a single i.v. dose (day 0) to C3H/HeJ mice (five per group) bearing bilateral 50- to 250-mg s.c. FSa-R (A) or FSa-N (B) tumors. DOX was administered in free form () as well as encapsulated in 100 nm DSPC/Chol () or PEG-PE/DSPC/Chol () liposomes. Control mice () were administered sterile saline. Tumor growth was monitored by caliper measurements and conversion to tumor weights as described under "Materials and Methods." Error bars represent the standard error of the mean resulting from determinations on 10 tumors per group.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The studies described here were conducted to determine whether TAMs play a role in the accumulation and antitumor activity of DOX administered in liposomal form. The rationale for comparing the pharmacokinetic, tumor accumulation and therapeutic properties of free and liposomal DOX formulations in fibrosarcoma tumors differing in TAM content was based on three primary considerations. First, the use of liposomes exhibiting extended circulation lifetimes and minimized plasma drug leakage properties will require processing events at the tumor site to render the DOX bioavailable to tumor cells. Second, previous studies have demonstrated the ability of macrophages to engulf, process and re-release free DOX after phagocytosis of liposome-entrapped drug (Storm et al., 1988). Third, the use of RES avoiding liposomes containing steric stabilizing lipids may compromise the beneficial effects of potential TAM-based liposome-processing phenomena. The results from these studies suggest that TAMs do not have a direct role in liposomal drug accumulation in tumors. Specifically, there was no histological correlation observed between DOX and macrophages within the tumors studied. The results, however, do suggest that TAMs may play a secondary role in the accumulation of liposomal drugs. The discussion provided here focuses on how this information can be used to enhance the antitumor potency of liposomal DOX as well as other liposome-based anticancer drug formulations.

Pharmacokinetic studies with liposomal DOX in the FSa-R and FSa-N tumor models suggest that, in tumor-bearing mice, a faster plasma liposome distribution phase is observed. This can account for approximately 20% of the injected drug and liposome dose within 1 h after administration and could not be attributed to accumulation in the tumor or to tumor-induced increases in blood volume. This effect, therefore, maybe caused by tumor-induced nonspecific activation of alternative particulate clearance mechanisms. Regardless, the results obtained here are consistent with previous reports (Gabizon et al., 1989; Huang et al., 1992b) which indicate that incorporation of PEG-PE into liposomal DOX formulations results in significantly enhanced circulation lifetimes. In accordance with our original hypothesis, the TAM-enriched FSa-N tumors exhibited significantly elevated levels of drug and liposome uptake for encapsulated DOX compared with FSa-R tumors. This phenomenon was observed at all time points studied and peaked at approximately 4 h after drug administration. The direct role of TAMs in mediating this increased uptake, however, was not consistent with several experimental observations: 1) increases in drug accumulation in the FSa-N tumor were observed for free drug as well as both liposomal formulations, and 2) time-dependent increased uptake by FSa-N tumors did not correlate with plasma pharmacokinetics or tumor-specific changes in kinetics of accumulation compared with FSa-R tumors. This behavior was unexpected in view of the dramatic differences in plasma drug levels observed between free and liposomal DOX (fig. 2) as well as the fact that peak free drug accumulation in tumors occurs shortly after injection compared with the extended accumulation kinetics obtained with both liposomal preparations (fig 3A).

Although drug and liposome accumulation is increased for TAM-enriched FSa-N tumors, the lack of specificity of this effect for liposome-encapsulated DOX formulations compared with free drug suggests that direct phagocytosis of the lipid carriers is not intricately involved. This interpretation is supported by histological studies which demonstrate that tumor DOX fluorescence did not correlate with TAM location and often was completely absent in CD IIb positive cells after injection of liposomal DOX. Because the role of TAMs within tumors will be partly to remove cellular debris and dead cells, it may be expected that these mature phagocytic cells will already be saturated in terms of phagocytic capacity and unable to accumulate additional foreign particulates, such as liposomes. Further, unlike the condition in ascitic tumor models (Bally et al., 1994), the contact between extravasated liposomes and macrophages within the tumor may be restricted because of the tumor matrix itself.

Although the data discussed above do not support mechanisms for enhanced tumor accumulation of liposomal DOX based on direct uptake by TAMs, macrophages may indirectly increase drug and liposome tumor levels. Factors secreted by TAMs may alter tumor properties such as vascular permeability or intratumor pH, both of which could contribute to increased tumor drug accumulation. Such phenomena would be consistent with the observation that the drug-to-lipid ratios for both liposomal DOX formulations are significantly elevated compared with plasma drug-to-lipid ratios in FSa-N tumors, which indicates tumor site-specific increases in drug release from the liposomes (see fig. 6). The fact that tumor drug accumulation is increased in FSa-N tumors for both free and liposomal DOX is also consistent with an indirect TAM effect on characteristics such as vascular permeability.

Two interesting observations in the studies presented here concern the efficiency of liposome accumulation in tumors as well as the relative efficacy of free and liposomal DOX formulations against FSa-R and FSa-N tumors. Empty and DOX-containing liposomes composed of PEG-PE/DSPC/Chol exhibited significantly extended circulation lifetimes compared with DSPC/Chol systems. However, this increase in plasma area under the curve liposome exposure did not result in a substantial improvement in the uptake of liposomal DOX by FSa-R and FSa-N tumors. These results are in contrast to some previous reports with similar systems (Huang et al., 1992b; Wu et al., 1993; Gabizon and Papahadjopoulos, 1988) and indicate that optimization of liposomal drug carriers based solely on extended plasma pharmacokinetic properties may not necessarily reflect improvements in tumor drug delivery. We have observed similar tumor uptake efficiency relationships for DSPC/Chol and PEG-PE/DSPC/Chol liposomes in various other solid tumor models including murine Lewis lung and human xenograft A431 squamous cell carcinoma (L. Mayer and M. Bally, unpublished observations). This suggests that particular attention should be given to the nature of the tumor model used for the evaluation of liposome-based drug formulations. It is important to note that tumor accumulation characteristics of liposomes containing the steric stabilizing lipid PEG-PE were not adversely affected by the presence of elevated TAM levels. This suggests that the RES-avoiding properties of these systems do not compromise their biological activity in macrophage-enriched tumors and is consistent with our conclusion that TAMs do not play a significant direct role in tumor accumulation.

The therapeutic studies performed here indicate that the 3.4-fold increases in tumor DOX concentrations achieved after 24 h with the liposomal formulations (compared with free DOX) provide only marginal improvements in antitumor activity against FSa-R tumors and are actually less effective than free drug in treating the TAM-enriched FSa-N tumors (see fig. 7). We would suggest that these surprising results are a consequence of poor DOX bioavailability in tumors in which liposomal drug has accumulated compared with free doxorubicin. These results also argue against the ability of macrophages to engulf liposomes and release DOX after intracellular processing because this should result in increased bioavailability and improved therapeutic activity for DOX administered in liposomal form compared with free drug. This indicates that the full therapeutic potential of liposomal anticancer drugs, such as DOX, will require increased sophistication in controlling tumor site-specific release of encapsulated agents to enhance activity. The advantage of such an approach has been demonstrated recently with thermosensitive liposomal DOX formulations (Huang et al., 1994).