INCREASED INTRACELLULAR DRUG ACCUMULATION AND COMPLETECHEMOSENSITIZATION ACHIEVED IN MULTIDRUG-RESISTANT SOLIDTUMORS BY CO-ADMINISTERING VALSPODAR (PSC 833)WITH STERICALLY STABILIZED LIPOSOMAL DOXORUBICINRajesh KRISHNA1,2, Maryse ST-LOUIS3 and Lawrence D. MAYER1,2*1Department of Advanced Therapeutics, BC Cancer Agency, Vancouver, Canada2Division of Pharmaceutical Chemistry, Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, Canada3Biotechnology Laboratory, University of British Columbia, Vancouver, Canada

We have previously demonstrated that liposome encapsu-lation of doxorubicin (DOX) can alleviate adverse interac-tions with non-encapsulated DOX and the cyclosporine multi-drug-resistant (MDR) modulator Valspodar. We have nowinvestigated the behavior of different liposomal DOX formu-lations in MDA435LCC6/MDR-1 human breast cancer solidtumor xenograft models to identify liposome characteristicsassociated with enhanced therapeutic activity and the mecha-nism whereby increased chemosensitization is achieved. Tox-icity studies incorporating conventional phosphatidylcholine(PC)/cholesterol (chol) and sterically stabilized (polyethyleneglycol 2000 [PEG]-containing) formulations of DOX indi-cated that whereas PC/Chol DOX was approximately 3-foldmore toxic in the presence of Valspodar, PEG containingdistearoylglycerophosphocholine (DSPC)/Chol DOX wasminimally affected. In mice bearing MDR tumors, co-administration of Valspodar and egg phosphocholine (EPC)/Chol DOX resulted in modest MDR modulation and efficacy,whereas the sterically stabilized formulation induced reduc-tions in tumor growth equivalent to that achieved for drug-sensitive tumors treated with non-encapsulated DOX. Phar-macokinetic studies revealed a 2.5-fold increase in plasmaDOX area under the curve (AUC) upon co-administration ofValspodar with EPC/Chol DOX whereas no such alterationswere observed with the sterically stabilized liposomes. Com-pared to non-encapsulated DOX combined with Valspodar,improvements in efficacy and toxicity correlated with theextent to which liposomal DOX formulations were able tocircumvent pharmacokinetic interactions. Confocal micros-copy demonstrated that Valspodar increased cell-associatedDOX which correlated with the level of anti-tumor efficacy.Int. J. Cancer 85:131–141, 2000.r 2000 Wiley-Liss, Inc.

The development of second generation multidrug-resistant (MDR)reversing agents alleviated many of the problems caused by earlierPGP blockers which were pharmacological agents with their owninherent toxicities. However, co-administration of conventionalanticancer drugs with many of these newer MDR modulators hasbeen shown to elicit drug-modulator interactions by virtue of PGPblockade in normal tissues such as liver, kidney, intestine and brain(Keller et al., 1992a; Gatmaitan and Arias, 1993). It may not besurprising then that clearance of several anticancer drugs isinhibited by MDR modulators such as cyclosporine (CsA) (Speeget al.,1992), verapamil (Nooteret al.,1987), Valspodar (Speeg andMaldonado, 1994, previously named PSC 833) and GW918 (Boothet al.,1998).

The effects of MDR modulators on drug transport proteins thatcause alterations in anticancer drug excretion often lead to in-creased anticancer drug exposure of healthy tissues and havenecessitated dose reduction in many preclinical (Krishna andMayer, 1997; Kelleret al.,1992a; Nooteret al.,1987) and clinical(Booteet al.,1996; Sarriset al.,1996) studies. Although doses canbe adjusted to equal levels of toxicity, it is unclear how suchpharmacokinetic (PK) changes may impact therapeutic activity.These interactions have been postulated to play a role in limitingthe therapeutic outcome in some patients (Wishartet al., 1994;Miller et al.,1994). While changes in anticancer drug dose and/or

schedule may be able to address toxicity or efficacy alterationsbrought about by MDR modulators, this clearly represents asignificant complication in applying PGP blockade strategies tocancer chemotherapy. This is due to the fact that most chemother-apy regimens utilize drug combinations, of which more than oneare often PGP substrates. Consequently, an ability to avoid suchanticancer drug clearance alterations may be considered to be asignificant advantage in MDR modulation strategies.

We earlier reported that liposome encapsulation of doxorubicin(DOX) can reduce non-encapsulated drug-Valspodar interactions,resulting in improved growth suppression of MDR murine solidtumors (Krishna and Mayer, 1997). The liposomal formulationused in these studies was composed of 120 nm diameter distearoyl-glycerophosphocholine (DSPC)/cholesterol (Chol) (55:45 molarratio) that retains DOX for extended periods of time (Mayeret al.,1989). The enhanced antitumor activity observed in this studyappeared to be a consequence of increased protection fromValspodar-mediated PK changes and toxicity exacerbation. How-ever, the mechanisms by which these effects were achieved are notfully understood, particularly since significant amounts of DOX aredelivered to the liver by liposomes without notable toxicologicalconsequences. In addition to alleviation of PK alterations, theincreased delivery of DOX to MDR solid tumors using liposomedelivery systems was associated with increased anti-tumor activitywhen co-administered with Valspodar compared with non-encapsulated drugs. These studies were unable, however, todistinguish the degree of DOX bioavailability (liposome entrappedvs.released drug) in the solid tumor. Therefore, the relative roles ofPGP blockade and tumor drug levels remain unresolved.

In order to address these questions, we compared here thetoxicity, efficacy, pharmacokinetics, cellular distribution propertiesof egg phosphocholine (EPC)/Chol DOX (a system where over50% of the drug is released in the first hour) and a stericallystabilized PEG2000(PEG) distearoylphosphoethanolamine (DSPE)/DSPC/Chol DOX formulation combined with the MDR modulator,Valspodar. The latter system was chosen on the basis of reports thatincorporation of 5 mol% PEG-polymerized lipid in 100 nmDSPC/Chol vesicles results in increased circulation longevity,reduced liver uptake and increased tumor delivery (Papahadjopou-los et al.,1991). Therefore, the effects of altering the drug release,liver uptake and tumor accumulation properties of liposomal DOXon toxicological and therapeutic activity were compared to revealthe processes underlying the improvements in toxicity and efficacyachieved with liposomal systems in a human breast carcinoma

Rajesh Krishna’s present address is: Department of Metabolism andPharmacokinetics, Pharmaceutical Research Institute, Bristol-Myers SquibbCo, P.O. Box 4000, Princeton, NJ 08543, USA.

*Correspondence to: Department of Advanced Therapeutics, BC CancerAgency, 600 West 10th Avenue, Vancouver, BC, V5Z-4E6, Canada.Fax:11 604 877-6011.

Received 10 May 1999; Revised 12 July 1999

Int. J. Cancer:85,131–141 (2000)

r 2000 Wiley-Liss, Inc.

Publication of the International Union Against CancerPublication de l’Union Internationale Contre le Cancer

MDR solid tumor xenograft model. Furthermore, these formula-tions reflect the 2 liposomal DOX products that are either approved(sterically stabilized) or pending approval (EPC/Chol) for wide-spread clinical use. Consequently, determining their pharmacologi-cal properties in the presence of MDR modulators would be ofconsiderable clinical interest.

MATERIAL AND METHODS

Material

DOX hydrochloride for injection (U.S.P.) was purchased fromDavid Bull (Vaudreil, Canada) and its purity affirmed by high-performance liquid chromatography (HPLC; see below). Valspodarwas a generous gift from Novartis (Dorval, Canada) and its purityaffirmed by LC/MS-MS (Varian LC-MS-MS system; Fisons,Altrincham, UK). DOX metabolite standards were generous giftsfrom Pharmacia Carlo Erba (Milan, Italy). PEG2000-DSPE (.99%purity), EPC (.99% purity) and DSPC (.99% purity) wereobtained from Northern Lipids (Vancouver, Canada) and Chol fromSigma (St. Louis, MO). Cholesteryl hexadecyl ether (3H), anon-exchangeable, non-metabolizable lipid marker was purchasedfrom Amersham (Oakville, Canada). HPLC grade solvents wereobtained from BDH (Toronto, Canada) and used without furtherpurification. Female BDF1 mice were obtained from Charles River(St. Constant, Canada). Female SCID/RAG2 mice were bredin-house at the BC Cancer Agency animal facility. The MDA435/LCC6 and its transfected MDR-1 line were generously provided byDr. R. Clarke (Georgetown University, Washington, DC). Thesecells were maintained in tissue culture in Dulbeco’s modified Eaglemedium (StemCell Technologies, Vancouver, Canada).

Liposome and drug preparationLiposomes composed of EPC/Chol (55:45), PEG2000-DSPE/

DSPC/Chol (5:50:45) and DSPC/Chol (55:45; mol:mol) wereprepared by initially dissolving the lipid mixtures in chloroform(100 mg lipid per milliliter) and hydrating the dried lipid film in a300-mM citric acid, pH 4.00, buffer. The resulting multilamellarvesicles (MLVs) were subjected to 5 freeze-thaw cycles followedby a 10-cycle extrusion through 2 stacked 100-nm polycarbonatefilters (Nuclepore, Pleasanton, CA) using a Lipex Extruder (LipexBiomembranes, Vancouver, Canada).3H-Cholesterylhexadecyl etherwas used as a non-exchangeable, non-metabolizable lipid marker(Derksenet al., 1987). The resulting large unilammelar vesiclesexhibited a mean diameter between 110–120 nm as determinedusing a Nicomp 270 submicron particle sizer (Particle SizingSystems, Santa Barbara, CA).

DOX was encapsulated in the liposomes using the transmem-brane pH gradient loading procedure (interior acidic) employingsodium carbonate as the alkalinizing agent and a drug-to-lipidweight ratio of 0.2:1.0 (Mayeret al., 1989). Liposomal DOXpreparations were diluted with saline as necessary prior toin vivoadministration. Valspodar (for animal studies) was dissolved in a10:1 mixture of ethanol (95%):Tween 80 and administered in acorn oil vehicle by oral gavage of a 200-µl volume (Kelleret al.,1992b). Non-encapsulated DOX was administered in sterile saline.

Toxicity evaluation studiesToxicity of the indicated DOX and Valspodar dose regimens was

evaluated in dose range-finding studies using normal (non-tumor-bearing) female BDF1 mice. Toxic dose range-finding studies intumor-free female mice were performed using 3 mice per groupwith appropriate group repetitions as described below. Briefly,mice were administered increasing (each dose in different groups)doses of i.v. DOX (free or liposomal)via the tail vein and oralValspodar at a fixed dose of 100 mg/kg (4 hr before DOX) on days1, 5 and 9. Liposomal DOX doses were changed by increments of 5mg/kg on the day 1, 5, 9 injection schedule. DOX dose escalationwas stopped when weight loss exceeded 30% or toxicity-relatedmortality was observed. Survival and the percent change in bodyweight was monitored over a 21-day period. Animals were

monitored for other toxicity signs such as scruffy coat, dehydration,lethargy, ataxia or labored breathing. Animals which demonstratedsignificant physical manifestations of distress/toxicity or exhibiteda body weight loss in excess of 30% were terminated. At the end ofthe 21-day study period, mice were terminated by carbon dioxideasphyxiation. Necropsies were performed to identify abnormalitiesin the major organs. The dose at which the body weight loss (groupmean value) was#15% and all mice survived for the duration ofstudy was established as the maximum tolerated dose (MTD). Adeviation in MTD of 61.5% was allowed in cases where bodyweight loss between 15–16.5% lasted only 1 day and where 100%recovery of body weight loss was observed, with 100% survival.The results were collated from at least 2 independent experiments.Select groups were repeated as a quality control measure to ensurereproducibility.

In vivo anti-tumor activityMDA435LCC6 cells were cultured in DMEM and passaged at

least 3 times in medium. A cell aliquot containing 53 106 cells in0.5 ml HBSS was injected i.p. into 2 female SCID/RAG2 mice.After 20–25 days, ascites (cells) were removed from the mouseviathe peritoneal wall using a 20g needle and placed in sterile 15-mlconical tubes containing 5 ml of HBSS without Ca and Mg salts.The cell suspension was then centrifuged at 1000 g for 5 min andthe supernatant discarded. Using a 27g needle fitted on a 1-mlsyringe, 50 µl of the cell suspension was injected into 2 mammaryfat pads on each mouse (23 106 cells per pad). A period of 18–21days was needed for the tumors to become palpable and measurablefor drug treatments to begin. The MDR MDA435LCC6-MDR-1cells were maintained and passaged identically to the WT cells.These MDR cells have had the MDR-1 cDNA introduced into theMDA435LCC6 cells (Leonessaet al., 1996). Cell passaging andinoculation were performed identical to the procedures used in theWT sensitive cell line.

Tumor growth suppression experiments were conducted inSCID/RAG2 mice bearing orthotopic human breast carcinomaMDA435LCC6 (multidrug resistant, MDR-1 and sensitive, WT)solid tumors. After approximately 3 weeks, when the tumors(n 5 8 per group) were established, treatment was initiated withdosage regimens incorporating i.v. non-encapsulated or liposomalDOX with or without p.o. Valspodar (given 4 hr before DOX) ondays 1, 5 and 9. Experiments were repeated to ensure reproducibil-ity of tumor growth inhibition properties. Caliper measurements ofthe tumors were performed daily, and the tumor weights calculatedaccording to the formula (Krishna and Mayer, 1997):

Tumor weight (g)5length (cm)3 [width (cm)]2

2

This conversion formula was verified by comparing the calculationderived tumor weights to excised and weighed tumors. Animalweights and mortality were monitored daily. Animals bearingulcerated tumors or where tumor weights exceeded 10% of theanimals’ body weight were terminated. The weights of the bilateraltumors were averaged for each mouse and mean tumor weights foreach treatment group6 standard error of the mean were calculated.Statistical tests for these longitudinal data were performed usingrepeat measures ANOVA employing Statistica for Windows 4.0(StatSoft, Tulsa, OK) and statistical significance was set atp ,0.05.

Pharmacokinetics and tissue distributionFemale MDA435LCC6/MDR-1 solid tumor-bearing SCID/

RAG2 mice received i.v.via the tail vein a single bolus ofliposomal doxorubicin (3H EPC/Chol or PEG2000-DSPE/DSPC/Chol 0.2:1.0). Valspodar (100 mg/kg) was administered p.o. (in 0.2ml) and doxorubicin administered 4 hr later. After DOX dosing,groups of 3 mice per time point were anesthetized with 100 µl i.p.of ketamine/xylazine at 30 min, 1, 2, 4, 16, 24, 48 and 72 hr. Bloodwas collected by cardiac puncture and placed into EDTA-coated

132 KRISHNA ET AL.

microtainer tubes. Terminal blood and selected tissues werecollected from all animals and were processed to determine lipid(using radioactivity) and DOX concentrations (using HPLC). Afterblood collection, the kidneys, livers and tumors were removedfrom each animal. Tissues were rinsed in PBS, pat dried onabsorbent paper and weighed in pre-weighed 163 100-mm tubes.Samples were stored in220°C pending analysis. A 10–30%homogenate in distilled water was prepared using a Polytronhomogenizer (Kinematica, Littau, Switzerland). A 0.2-ml aliquot ofthe homogenate was digested with a tissue solubilizer, decolorizedwith peroxide and analyzed for the lipid label by scintillationcounting. The specimens obtained from the control animals wereused as background samples.

DOX and its metabolites in plasma and tissue extracts weredetermined using HPLC. The HPLC assay of Andersenet al.(1993) was used to analyze DOX and its metabolites with minormodification. Briefly, sample extraction with acetonitrile wasfollowed by isocratic elution from a Nova-Pak C18 3.93 150 mm(Millipore, Bedford, MA) analytical reverse phase column andquantified by endogenous fluorescence (emission wavelength of515 nm). The mobile phase consisted of a 16-mM ammoniumformate buffer (pH 3.5)/acetone/isopropanol mixture (75:20:5)delivered at a rate of 1.0 ml/min. The column was maintained at40°C. A NEC (Boxborough, MA) Powermate SX Plus Computerand a Systems Interface Module (Waters, Milford, MA) were usedfor data handling.

Using this system, the retention times of DOXol, DOX, DOXoneand 7-deoxyDOXone were 3.6, 5.8, 7.5 and 12.6 min, respectively.Recoveries, using acetonitrile as the extraction solvent, fromplasma over a concentration range of 0.05–10 µg/ml of DOX,DOXone, 7-deoxyDOXone and DOXol, were between 80–110%.To protect DOX and its metabolites from photodegradation, allprocedures were shielded from direct exposure to light. In addition,DOX was found to be stable in the mobile-phase solvent mixturefor at least 96 hr (40°C), on the HPLC autosampler tray for at least96 hr, in reconstituted form at 4°C for at least 10 days and for atleast 4 freeze-thaw cycles.

The plasma data were modeled using WinNONLIN Version 1.5PK software (Pharsight, Mountain View, CA) to calculate areaunder the curve (AUC), half-life (T1/2) and plasma clearance (CLp)according to standard equations (Gibaldi and Perrier, 1982). Thetrapezoidal rule was used to calculate the AUCs in tissue concentra-tion-time profiles employing a computer software AUC (programprovided by Dr. W. Riggs, Faculty of Pharmaceutical Sciences,University of British Columbia). Tissue DOX levels were correctedfor blood volume to account for material residing in the vasculatureof the tissues, using previously published values (Ballyet al.,1993).

Select samples were assayed for free and liposome-associatedDOX using Microcon-30 filters (Amicon, Oakville, Canada) usingthe equilibrium filtration method of Mayer and St. Onge (1995).Separation of free from liposomal and protein-bound drug wasperformed using Microcon-30 (0.5-ml capacity, Amicon) ultrafiltra-tion devices with a m.w. cut-off of 30 kDa. Microcon-30 sampleswere centrifuged at 4°C, 8,000 g for 20 min in a microcentrifuge(IEC Micromax Centrifuge; International Equipment, NeedhamHeights, MA). The ultrafiltrate was processed for DOX by HPLC(see above).

Confocal microscopy and imaging studiesFor confocal imaging studies, SCID/RAG2 mice bearing

MDA435LCC6/MDR-1 tumors were treated with non-encapsu-lated DOX, EPC/Chol DOX or PEG-DSPE/DSPC/Chol DOX (5mg/kg) in the presence and absence of Valspodar 100 mg/kg (4 hrbefore DOX). At the indicated times following DOX administra-tion, tissues were aseptically dissected, bathed in PBS and imagedfresh. Before imaging, thin pieces of tumors were placed onconcave slides and observed under a 603 oil immersion lens.These were then viewed under the confocal microscope to deter-

mine DOX distribution characteristics. As controls, known amountsof non-encapsulated DOX or PEG-DSPE/DSPC/Chol DOX lipo-somes were infused into freshly isolated muscle tissues and viewedfor DOX fluorescence.

Confocal images were collected on an Optiphot 2 researchmicroscope (Nikon, Tokyo, Japan) attached to a confocal laserscanning microscope (MRC-600, BioRad, Hercules, CA) usingCOMOS software (BioRad). The laser line on the krypton/argonlaser was 488 nm. Filterblock BHS was used to detect DOX (488nm excitation, 515 nm emission). The numerical aperture was 0.75on the320 air objective and 1.2 on the360 oil objective. Theimages were captured such that the xyz dimensions were 0.4 µmcubed (320) and 0.2 µm pixel (360). NIH Image version 1.61 wasused for image analysis, and all images were based on maximumintensity projection. Projections made in the NIH Image weresaved in TIFF format, then imported to Adobe Photoshop version4.0 where the different fluorophore images were assigned toindividual RGB channels and subsequently merged to provide thefinal image of the single or multiple sections.

RESULTS

Toxicity

In order to identify the MTD for the day 1, 5 and 9 dosageregimen used in therapeutic experiments as well as to investigatethe mechanisms of Valspodar-mediated increases in DOX toxicity,21-day dose range-finding toxicity studies were conducted withEPC/Chol, DSPC/Chol and PEG-DSPE/DSPC/Chol liposomalformulations of DOX as well as non-encapsulated drug in thepresence and absence of p.o. Valspodar in non-tumor-bearinghealthy mice. The results of this study are shown in Table I, wherethe body weight loss is presented, along with the MTDs (as definedin Material and Methods). The 3 liposomal DOX formulationsyielded variable toxicity characteristics that depended on theirrespective abilities to retain DOX.

In the absence of Valspodar, EPC/Chol DOX exhibited an MTDof 15 mg/kg, where the body weight loss was 1.7% with 100%survival (Table I). For the 2 saturated lipid formulations, DSPC/Chol and PEG-DSPE/DSPC/Chol DOX, the MTD was identical at25 mg/kg, representing a 1.7-fold increase in MTD compared withthe more leaky EPC/Chol formulation. When EPC/Chol DOX was

TABLE I – TOXICITY AS A FUNCTION OF DOX DOSE FOR NON-ENCAPSULATEDAND LIPOSOMAL FORMULATIONS1

Group Dose(mg/kg)

Day 10 weight loss (% survival)

2PSC 833 1PSC 833

Non-encapsulated 2.5 0 (100) 10.2 (100)DOX 5 3.5 (100) 18.5 (0)

7.5 10.7 (100) 28.2 (0)10 29.2 (0)

MTD 7.5 mg/kg 2.5 mg/kg

EPC/Chol 5 0.9 (100)DOX 10 2.8 (100) 23.1 (0)

15 1.7 (100)20 20.4 (0) 24.4 (0)

MTD 15 mg/kg 5 mg/kg

DSPC/Chol 15 10.8 (100) 12.9 (100)DOX 20 3.1 (100) 15.0 (100)

25 16.0 (100) 18.5 (67)MTD 25 mg/kg 20 mg/kg

PEG-DSPE/DSPC/Chol 15 7.3 (100) 9.8 (100)DOX 20 6.5 (100) 14.8 (100)

25 9.3 (100) 14.8 (100)30 4.4 (67)2 25.3 (0)

MTD 25 mg/kg 25 mg/kg

1Data are group mean values (n5 3 mice per group treated i.v. ondays 1, 5 and 9).–2Body weight loss nadir is on day 19 (210.2%) with67% survival.

133COMPLETE REVERSAL OF MDR IN SOLID TUMORS

combined with Valspodar, significant toxicity resulted, necessitat-ing a dose reduction by 3-fold, to 5 mg/kg. This is comparable tothe 3-fold decrease in MTD caused by Valspodar for non-encapsulated drug administration (Table I). In contrast, Valspodarnecessitated a small 1.2-fold decrease in DOX encapsulated inDSPC/Chol liposomes and no dose reduction was required for thePEG-containing sterically stabilized liposomes (Table I).

EffıcacyThe anti-tumor activity of the 3 types of liposomal DOX

formulations was evaluatedin vivo in the absence and presence ofValspodar using the MDA435LCC6 and PGP-overexpressing

MDA435LCC6/MDR-1 human breast carcinoma xenograft solidtumor models. Figure 1 presents the tumor growth curves for micetreated with non-encapsulated drug (a), EPC/Chol DOX (b),DSPC/Chol DOX (c) and sterically stabilized PEG-DSPE/DSPC/Chol DOX (d) in the presence and absence of Valspodar.

When MDA435LCC6 cells (sensitive, WT or resistant, MDR)are inoculated in the mammary fat pads of SCID/RAG2 mice, solidtumors readily establish (tumor take rates.95%). Figure 1 showsthat both MDR and WT-untreated controls exhibit comparabletumor growth rates, with the exponential growth phase occurringbetween days 5 and 14 (slope of linear regression line5 0.067 for

FIGURE 1 – Anti-tumor efficacy of free(a),EPC/Chol DOX(b),DSPC/Chol(c) and PEG-DSPE/DSPC/Chol(d) against MDA435LCC6 WT orMDR1 human xenograft solid tumors in the absence and presence of co-administered Valspodar. MDA435LCC6 tumors were grown on mammaryfat pads of female SCID/RAG2 mice. Oral Valspodar (100 mg/kg) and i.v. DOX treatments were initiated once tumors were established (20–100mg) and were given on days 1, 5 and 9 at the indicated doses of free and liposomal DOX. Valspodar was administered 4 hr prior to DOX injection.Data are expressed as mean6 standard error of the mean. For legends, see individual panels.

134 KRISHNA ET AL.

WT vs.0.056 for MDR tumors). Both groups were terminated onday 18, when tumor weights reached 0.846 0.05 and 0.766 0.05 gfor the WT and MDR tumors, respectively. Treatment of WTtumors with non-encapsulated DOX at 7.5 mg/kg on days 1, 5 and 9(Fig. 1a) resulted in significant tumor growth suppression until day20, when tumors weighed 0.0886 0.01 g. These tumors eventuallygrew to 0.686 0.09 g by day 40. However, administration ofnon-encapsulated DOX at 7.5 mg/kg on days 1, 5 and 9 in micebearing MDR tumors did not cause any tumor growth suppression(Fig. 1a). When MDR tumor-bearing mice were treated with theMTD of non-encapsulated DOX combined with Valspodar (DOXdose of 3 mg/kg), there was partial tumor growth inhibition untilday 11. After this time, tumor growth rates were similar tountreated MDR controls until day 18 when mice were terminated(Fig. 1a). Growth of MDR tumors treated with non-encapsulatedDOX plus Valspodar was significantly different from both un-treated MDR tumors as well as MDR tumors treated withnon-encapsulated drug alone between days 7 and 12 post-DOXadministration. However, in comparison to WT tumors treated withnon-encapsulated DOX 7.5 mg/kg, the MDR tumor growth inhibi-tion caused by non-encapsulated drug and Valspodar was transient(Fig. 1a).

When MDR tumors were treated with liposomal DOX formula-tions in the presence and absence of Valspodar (Fig. 1b–d), varyingdegrees of tumor growth suppression were observed. As seen inFigure 1b, EPC/Chol DOX alone (5 mg/kg) was unable to inducesubstantial inhibition of MDR tumor growth, with tumors weighing0.86 6 0.2 g on day 18. In the presence of Valspodar, however,EPC/Chol DOX at 3 mg/kg treatment closely resembled EPC/CholDOX alone until day 12, after which tumor growth was decreasedbetween days 12 and 20 (Fig. 1b). This indicated modest delayedanti-tumor activity. This MDR modulation caused by EPC/CholDOX and Valspodar was significantly better than that caused bynon-encapsulated DOX and Valspodar (p , 0.05).

Figure 1c illustrates the tumor growth inhibition of DSPC/Cholliposomal DOX in the presence and absence of Valspodar. In theabsence of Valspodar, DSPC/Chol liposomal DOX caused a modestreduction in tumor growth. The tumor growth inhibition caused byDSPC/Chol DOX was significantly different from untreated con-trols until day 12, after which the tumor growth rate increased.Tumor weight for MDR tumors treated with DSPC/Chol DOXalone was 0.516 0.1 g on day 20. In comparison, Valspodar causeda significant increase in DSPC/Chol DOX tumor growth suppres-sion of the MDR solid tumors, where tumor weight was 0.2560.05 g on day 20. The tumor growth inhibition resulting fromDSPC/Chol DOX and Valspodar treatment was significantly( p , 0.05) greater than that observed for EPC/Chol DOX andValspodar.

PEG-DSPE/DSPC/Chol DOX formulations displayed superioranti-tumor activity in the presence and absence of Valspodar whencompared with the 2 liposomal DOX formulations described aboveas well as non-encapsulated drug (Fig. 1d). In the absence ofValspodar, PEG-DSPE/DSPC/Chol DOX caused a significantreduction in tumor growth until day 11, when tumor weight was0.1 6 0.01 g (compared with a tumor weight of 0.056 0.01 g onday 1), after which tumor growth rates increased. Tumor weight forthis group was 0.396 0.02 g on day 20. In the presence ofValspodar, PEG-DSPE/DSPC/Chol DOX significantly inhibitedtumor growth and this growth suppression was not significantlydifferent from WT tumors treated with non-encapsulated DOX atits MTD. Specifically, the tumor weights on day 20 for PEG-DSPE/DSPC/Chol DOX in presence of Valspodar were 0.16 0.02 gcompared with 0.0886 0.01 g for WT tumors treated withnon-encapsulated DOX. This suppression of tumor growth forPEG-DSPE/DSPC/Chol DOX and Valspodar was significantlydifferent (p , 0.05) from all other treatment groups for micebearing MDR solid tumors.

Pharmacokinetics and tissue distributionA comprehensive PK evaluation was performed to correlate

toxicity and efficacy data with plasma and tissue (tumor and liver)DOX and DOX metabolite distribution properties determined usingHPLC analysis. The comparison of DSPC/Chol DOX and non-encapsulated (free) DOX pharmacokinetics and tissue distributionproperties in the presence and absence of Valspodar has beenextensively characterized (Krishna and Mayer, 1997). Since thetoxicity and efficacy properties of these formulations were verycomparable in the current models, the results presented herefocused on comparisons between a liposomal system which leaks asignificant portion of entrapped drug into the circulation (EPC/Chol DOX), and a sterically stabilized system that exhibitsnegligible drug release in plasma compartment and increasedcirculation lifetimes (PEG-DSPE/DSPC/Chol DOX). As a refer-ence comparison, non-encapsulated DOX treatments (with orwithout Valspodar) were evaluated for DOX concentrations inplasma and the tumor. The goal of these comparisons was to revealimportant DOX distribution and metabolism properties that dictatetoxicity and efficacy behavior.

Plasma drug kineticsFigure 2a presents the DOX plasma concentrations after i.v.

administration of non-encapsulated DOX and the 2 liposomal DOXformulations at a DOX dose of 5 mg/kg. Following administrationof non-encapsulated drug, DOX is rapidly eliminated from thecirculation. Concentrations of DOX beyond 4 hr were below assaydetection limits. The concentration-time profile was characterizedby a Cmax of 1.56 0.1 µg/ml and an AUC of 4.4 µg.hr/ml (Fig. 2a).However, when Valspodar was co-administered with non-encapsulated DOX at 5 mg/kg, DOX elimination from plasma wascharacterized by a prolonged terminal elimination phase (Fig. 2a).Valspodar caused significant (p , 0.05) increases in Cmax(3.96 0.5µg/ml) and AUC (48.1 µg.hr/ml) of non-encapsulated DOX com-pared to data obtained in the absence of the MDR modulator. Thisapproximately 11-fold increase in DOX AUC caused by Valspodaris consistent with the 10-fold increase in DOX AUC for thenon-encapsulated DOX-Valspodar combination observed in theP388/ADR solid tumor model described earlier (Krishna andMayer, 1997).

As shown in Figure 2a, DOX elimination from plasma for bothEPC/Chol and PEG-DSPE/DSPC/Chol DOX systems exhibits amonophasic elimination profile characterized by a 1-compartmentmodel with first-order elimination. While EPC/Chol DOX plasmaconcentration exhibits rapid elimination of the drug within 24 hr,PEG-DSPE/DSPC/Chol DOX displays a prolonged circulationlife-time with over 20% remaining at 24 hr. In all cases, parentDOX was the only detectable entity with no indication of anymetabolites present. In the presence of Valspodar, the eliminationprofile remained monophasic, however, DOX concentrations atearlier time points were significantly (p , 0.05) increased forEPC/Chol liposomes whereas no such Valspodar effect was ob-served for PEG containing DSPC/Chol liposomes. This is contrastto observations for non-encapsulated DOX which demonstratedincreases in the terminal elimination phase in the presence ofValspodar (Fig. 2a). Co-administration of Valspodar and EPC/CholDOX increased the AUC of DOX by 2.6-fold and Cmax by 2.3-fold,accounting for the 40% reduction in plasma clearance (Table II;significant at p , 0.05). In contrast, Valspodar caused minorchanges in the pharmacokinetics of DOX encapsulated in PEG-DSPE/DSPC/Chol liposomes, with small 36% and 25% increasesin Cmax and AUC, respectively (Table II).

Figure 2b shows the elimination of liposomal lipid from plasma.These data demonstrate that, similar to DOX pharmacokinetics,liposomal lipid elimination is monophasic, characterized by a1-compartment model with first-order elimination. The plasmaclearance (CLp) for PEG-DSPE/DSPC/Chol liposomes was 2.4-fold lower than EPC/Chol, and the half-life (T1/2) was 2.2-foldhigher for PEG-DSPE/DSPC/Chol liposomes compared to EPC/

135COMPLETE REVERSAL OF MDR IN SOLID TUMORS

Chol (Table II). Valspodar did not significantly alter the liposomallipid elimination kinetics for either liposomal formulation, asdemonstrated by the small changes in lipid half-lives (Table III).

Plasma from mice treated with EPC/Chol and PEG-DSPE/DSPC/Chol DOX formulations in the presence and absence of Valspodarwas processed using Microcon-30 filters in order to determine the

amount of non-encapsulated drug in the plasma following i.v.injection of liposomal DOX and to assess whether higher levels ofnon-encapsulated drug are present upon co-administration ofValspodar. Non-encapsulated drug was below assay detectionlimits following administration of EPC/Chol and PEG-DSPE/DSPC/Chol DOX formulations in the absence of Valspodar at 1, 4 or 24 hr(data not shown). Further, no non-encapsulated drug was present(i.e., below assay detection limits) for PEG-DSPE/DSPC/CholDOX in the presence of Valspodar. This is consistent with the lackof Valspodar-mediated changes in plasma DOX concentrations andthe unchanged drug-to-lipid ratio with the PEG-DSPE/DSPC/Cholsystem. However, non-encapsulated drug levels were elevated forEPC/Chol DOX in the presence of Valspodar, particularly at 1 and4 hr when plasma non-encapsulated DOX concentrations were 4.32and 0.65 µg/ml, respectively (data not shown). This was associatedwith a 2-fold decrease in the plasma drug-to-lipid ratio whichindicated approximately 50% release of DOX from these lipo-somes. Given that the equilibrium between non-encapsulated DOXand protein-bound DOX in plasma is approximately 4:1 (Mayerand St-Onge, 1995), the 1 and 4 hr non-encapsulated DOXdeterminations following EPC/Chol DOX administration implicatethat approximately 32 and 12%, respectively, of the total plasmaDOX concentration exists in non-encapsulated form.

TABLE II – EFFECT OF VALSPODAR ON PLASMA DOX PK PARAMETERS WHENDOX IS ADMINISTERED IN LIPOSOME-ENCAPSULATED FORMS

Group AUC0-t(µg.h./ml)

T1/2(hz)

Cmax(µg/ml)

CLp(ml/hr)

EPC/Chol DOX 5 mg/kg 97.9 1.9 35.8 1.0EPC/Chol DOX 5 mg/kg1 Val-

spodar 100 mg/kg256.91 2.1 83.81 0.391

PEG2000-DSPE/DSPC/Chol DOX 5mg/kg

1523.6 9.9 107.1 0.07

PEG2000-DSPE/DSPC/Chol DOX 5mg/kg1 Valspodar 100 mg/kg

1905.2 9.0 146.8 0.05

1Significantly different from the no Valspodar group (p , 0.05;2-samplet-test).

TABLE III – EFFECT OF VALSPODAR ON PLASMA LIPOSOMAL LIPIDPHARMACOKINETIC PARAMETERS WHEN DOX IS ADMINISTERED IN

LIPOSOME-ENCAPSULATED FORMS1

Group AUC0-t(µg.h./ml)

T1/2(hz)

Cmax(µg/ml)

CLp(ml/hr)

EPC/Chol DOX 5 mg/kg 5144.8 8.5 421.9 0.12EPC/Chol DOX 5 mg/kg1 Val-

spodar 100 mg/kg4540.6 5.9 535.9 0.13

PEG2000-DSPE/DSPC/Chol DOX 5mg/kg

12158.4 16.8 501.2 0.05

PEG2000-DSPE/DSPC/Chol DOX 5mg/kg1 Valspodar 100 mg/kg

12111.5 12.2 691.2 0.05

1Data are mean values (n5 3 mice per time point; n5 24 pergroup). PK abbreviations: area under the concentration-time curve,AUC; elimination half-life, T1/2; peak concentration achieved, Cmax;and plasma clearance, CLp.

FIGURE 2 – DOX (a) and liposomal lipid (b) concentration-timeprofiles in plasma following administration of non-encapsulated DOX5 mg/kg in the absence (open triangles) or presence of Valspodar(administered 4 hr prior to DOX) 100 mg/kg p.o. (filled triangles);EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence ofValspodar (administered 4 hr prior to DOX) 100 mg/kg p.o. (filledcircles); and PEG-DSPE/DSPC/Chol DOX 5 mg/kg in the absence(open squares) or presence of Valspodar (administered 4 hr prior toDOX) 100 mg/kg p.o. (filled squares). Mice bearing 200–500 mg solidMDA435LCC6 tumors (3 per group) were terminated at the indicatedtimes and plasma and tissue samples were processed and quantified forDOX and metabolites by HPLC as described in Material and Methods.Data are expressed as mean6 standard error of the mean.

136 KRISHNA ET AL.

Tumor drug kineticsFigure 3a illustrates tumor DOX concentrations following

administration of non-encapsulated DOX at 5 mg/kg in the absenceand presence of Valspodar. The tumor DOX accumulation follow-ing administration of non-encapsulated DOX is characterized by aCmax of 6.6 6 1.4 µg/g and a DOX AUC of 103.9 µg.hr/g (Table

IV). Co-administration of Valspodar with non-encapsulated DOXat 5 mg/kg resulted in a significant (p , 0.05) 2.5-fold increase inDOX AUC.

In these MDR solid tumors, the more leaky EPC/Chol DOXformulation alone provided modest increases in DOX tumoraccumulation (Fig. 3a), with a Cmax of 12.3 6 0.6 µg/g and totaldrug exposure of 811.4 µg.hr/g (Table IV). In contrast, thelong-circulating PEG-DSPE/DSPC/Chol formulation was able toprovide greater accumulation of DOX in the tumor, as indicated bythe 2.6-fold increase in Cmax and a 2.5-fold increase in AUC.Compared to non-encapsulated DOX treatment at 5 mg/kg, admin-istration of DOX encapsulated in EPC/Chol and PEG-DSPE/DSPC/Chol liposomes (at identical doses) resulted in significant (p , 0.05)7.8- and 19.6-fold increases in tumor DOX AUC values, respec-tively (Table IV). When Valspodar was co-administered withEPC/Chol DOX at 5 mg/kg, a small increase in tumor accumula-tion was observed (Fig. 3, Table IV). An interesting observationwas the presence of the DOX aglycone metabolite, DOXone, atearly time points when Valspodar was given in conjunction withEPC/Chol DOX, indicating possible inhibition of DOX andDOXone elimination caused by the MDR modulator. No othermetabolites, however, were recovered in any of the other treatmentgroups.

DOX encapsulated in PEG-DSPE/DSPC/Chol liposomes com-bined with Valspodar resulted in a tumor-associated DOX Cmax of56.86 5.3 µg/g and an AUC of 3,104.2 µg.hr/g. Valspodar caused a1.7-fold increase in tumor Cmax and a 1.5-fold increase in AUC forDOX administered in PEG-DSPE/DSPC/Chol liposomes. In boththe presence and absence of Valspodar, PEG-DSPE/DSPC/Cholliposomes provided approximately 2.5-fold increases in DOXtumor Cmax and AUC values compared with EPC/Chol liposomes.

Liposomal lipid levels in tumors were also monitored in order todetermine whether increased DOX tumor concentrations associ-ated with Valspodar co-administration were due to elevated accumu-lation of the liposomes themselves in the tumor or rather increasedretention of DOX in the tumor resulting from PGP inhibition by

TABLE IV – TUMOR DOX DISTRIBUTION CHARACTERISTICS FOLLOWINGADMINISTRATION OF FREE OR LIPOSOMAL (EPC/Chol AND

PEG-DSPE/DSPC/Chol) DOX FORMULATIONS IN THE PRESENCE AND ABSENCEOF VALSPODAR1

Group Cmax(µg/g)

AUC2

(µg.h/g)

Non-encapsulated DOX (5 mg/kg) 6.66 1.4 103.92Non-encapsulated DOX (5 mg/kg)1

Valspodar14.16 4.4 253.93,4

EPC/Chol Dox (5 mg/kg) 12.36 0.6 811.4EPC/Chol Dox (5 mg/kg)1 Valspodar 22.66 2.9 1228.9PEG-DSPE/DSPC/Chol Dox (5 mg/kg) 32.86 2.6 2015.7PEG-DSPE/DSPC/Chol Dox (5 mg/kg)1

Valspodar56.86 5.3 3104.2

1Values represent intact DOX determined using HPLC analysis; dataare mean6 standard error of the mean (n5 3 mice per time point;n 5 24 mice per group).–2AUC 0–72 was calculated using the trapezoi-dal rule.–3AUC value represents 0–24-hr time period.–4Significantlydifferent from the no Valspodar group atp , 0.05.

FIGURE 3 – Tumor DOX(a) and liposomal lipid(b) concentration-time profiles following administration of EPC/Chol DOX 5 mg/kg inthe absence (open circles) or presence of Valspodar (administered 4 hrprior to DOX) 100 mg/kg p.o. (filled circles); and PEG-DSPE/DSPC/Chol DOX 5 mg/kg in the absence (open squares) or presence ofValspodar (administered 4 hr prior to DOX) 100 mg/kg p.o. (filledsquares). Mice bearing 200–500 mg solid MDA435LCC6 tumors (3per group) were terminated at the indicated times and plasma and tissuesamples were processed and quantified for DOX and metabolites byHPLC as described in Material and Methods. Data are expressed asmean6 standard error of the mean. See legend for explanation oftriangle symbols.

137COMPLETE REVERSAL OF MDR IN SOLID TUMORS

Valspodar. For both EPC/Chol and PEG-DSPE/DSPC/Chol DOXformulations, liposomal lipid levels (Fig. 3b) observed in thetumors were comparable in the presence and absence of Valspodar.Specifically, for the PEG-DSPE/DSPC/Chol system, a Cmax of455.36 32.6 µg lipid/g and an AUC of 23,788 µg lipid.hr/g wereobtained in the absence of Valspodar, whereas a Cmax of 516.66135.6 µg lipid/g and an AUC of 24,883 µg lipid.hr/g were obtainedfor tumors in the presence of the MDR modulator. Similarrelationships were observed for the EPC/Chol formulation (Fig.3b). Compared with EPC/Chol liposomes, PEG-DSPE/DSPC/Cholprovided a 1.5-fold elevation in liposomal lipid tumor accumula-tion. Since liposomal lipid levels were independent of Valspodarfor both of the liposomal DOX formulations, the increased DOXaccumulation in the tumor caused by Valspodar co-administrationappeared consistent with greater drug retention at the tumor sitefollowing PGP blockade.

Drug kinetics in liver tissueDOX concentrations were evaluated in liver tissue for the

various treatment groups over 0–72 hr. Co-administration ofEPC/Chol DOX (5 mg/kg) with Valspodar resulted in a 1.3-foldincrease in DOX Cmaxand a doubling in DOX AUC (Table V). Thiscompared to Valspodar induced increases in Cmax and AUC fornon-encapsulated DOX of 4.3- and 13-fold, respectively. Interest-ingly, the combination of PEG-DSPE/DSPC/Chol DOX and Valspo-dar resulted in a 1.5-fold reduction in Cmax and a 1.8-fold reductionin AUC compared to PEG-DSPE/DSPC/Chol DOX in the absenceof the MDR modulator. In the presence of Valspodar, the liver DOXaccumulation with EPC/Chol DOX was 3.2-fold higher than thatfor the PEG-DSPE/DSPC/Chol DOX and 5.9-fold higher com-pared with non-encapsulated DOX.

Non-encapsulated DOX is metabolized in the liver of miceprimarily to: (1) doxorubicinol,via reduction of a side chaincarbonyl group to a secondary alcohol by a cytosolic aldo-ketoreductase; (2) doxorubicinone,via reductive cleavage of thedaunosamine moiety; (3) 7-deoxy aglycones, followed by furtheraglycone conjugation and O-methylation (Vrignaudet al., 1986).Following administration of non-encapsulated DOX alone at a doseof 5 mg/kg, DOX and all 3 metabolites were quantitated in the liver.On a relative scale, the respective contributions of DOX and itsmetabolites to the total AUC, at 1 hr, were as follows: DOX(13.1%), DOXol (9.6%), DOXone (73.5%) and 7-deoxyDOXone(3.8%). When non-encapsulated DOX was co-administered withValspodar, a 4.3-fold increase in DOXone AUC0–8 was observed,along with an 8.8-fold increase in 7-deoxyDOXone AUC0–8( p , 0.05, with or without Valspodar). DOXol levels followingValspodar co-administration, however, were below assay detectionlimits.

In contrast with the observations with non-encapsulated DOX,liposome-encapsulated DOX exhibited different metabolic proper-ties both in absolute drug levels as well as relative distributionbetween different DOX metabolites. The 2 major metabolitesidentified after injection of liposomal DOX formulations wereDOXone and 7-deoxyDOXone. Table IV summarizes DOXone

Cmax and AUC for EPC/Chol DOX and PEG-DSPE/DSPC/CholDOX, in the presence and absence of Valspodar, in the liver.EPC/Chol is a leaky formulation of DOX that displays some of thefeatures observed for non-encapsulated DOX with respect tometabolic profiles. With the long-circulating PEG-DSPE/DSPC/Chol DOX formulation, formation of the DOXone metabolite wasnot favored as reflected by an 8- and 4-fold reduction in DOXoneCmax and AUC, respectively, compared to EPC/Chol DOX (signifi-cant atp , 0.05).

DOXol levels were below assay detection limits for bothliposomal formulations. On a relative scale, the respective contribu-tions of DOX and its various metabolites to the total exposure at 1hr for EPC/Chol DOX were as follows: DOX (16.6%), DOXol(0%), DOXone (50.2%) and 7-deoxyDOXone (33.1%). For PEG-DSPE/DSPC/Chol DOX, the respective contributions were: DOX(43.4%), DOXol (0%), DOXone (8.2%) and 7-deoxyDOXone(48.4%). When Valspodar was co-administered with EPC/CholDOX at 5 mg/kg, a higher DOXone Cmax value was observed(compare 65.86 22.6 µg/g for EPC/Chol DOX and 102.46 4.7µg/g for EPC/Chol DOX and Valspodar), along with a significant9.5-fold increase in AUC. These increases in DOXone Cmax andAUC0–72 for EPC/Chol DOX following Valspodar co-administra-tion are similar to those observed for non-encapsulated DOX in thepresence of Valspodar. The AUC values for EPC/Chol DOXreported here are over a 0–72 hr period and for non-encapsulatedDOX, the time course beyond 8 hr was not followed due toundetectable levels.

Confocal imagingIn order to elucidate whether the cellular distribution properties

of DOX in tumor tissue are influenced by Valspodar in PGP-expressing solid tumors, a non-perturbing confocal fluorescencemicroscopy (CFM) technique with image processing analysis wasused. The need for a non-perturbing imaging technique is due to thefact that standard tissue processing methodologies for conventionalor fluorescence microscopy utilize procedures that disrupt theliposomal bilayer, causing DOX leakage and artificially highindications of intracellular uptake (data not shown). This complica-tion does not occur when tumors are imaged by confocal micros-copy immediately after excision. The fresh tissue CFM methodtakes advantage of the fact that near complete fluorescencequenching occurs when DOX is entrapped within liposomes (Wuetal., 1997; data not shown), leading to the selectivity of visualizingfluorescence of released DOX. Therefore, this method can qualita-tively assess concentrations of fluorescent DOX that have beenreleased from the liposomes and have localized intracellularly.

Figure 4 presents the CFM images of MDA435LCC6/MDR-1tumors from mice administered non-encapsulated DOX, EPC/CholDOX and PEG-DSPE/DSPC/Chol DOX, all at a drug dose of 5mg/kg, in the presence and absence of Valspodar. MDR tumorsisolated from mice treated with non-encapsulated DOX in theabsence of Valspodar demonstrated no detectable DOX fluores-cence either at 1 or 4 hr. In the presence of Valspodar, however,increased DOX fluorescence was visualized at 1 hr (Fig. 4). The

TABLE V – LIVER DOX AND DOXone DISTRIBUTION FOLLOWING ADMINISTRATION OF EPC/Chol DOX ANDPEG-DSPE/DSPC/Chol DOX IN THE PRESENCE AND ABSENCE OF VALSPODAR1

GroupDOX DOXone

Cmax(µg/g)

AUC(µg.hr/g)

Cmax(µg/g)

AUC(µg.hr/g)

Non-encapsulated DOX 7.5 mg/kg 27.36 9.2 38.6 84.26 6.8 241.12Non-encapsulated DOX 7.5 mg/kg1 Valspodar 115.16 17.63 514.33 307.66 34.13 1034.52,3

EPC/Chol Dox (5 mg/kg) 43.16 5.4 1511.6 65.86 22.6 262.8EPC/Chol Dox (5 mg/kg)1 Valspodar 53.66 2.3 3034.53 102.46 4.7 2477.93PEG-DSPE/DSPC/Chol Dox (5 mg/kg) 41.76 6.8 1715.7 8.56 5.6 62.5PEG-DSPE/DSPC/Chol Dox (5 mg/kg)1 Valspodar 28.56 10.5 950.53 13.56 0.3 699.53

1AUC 0–72 hr was calculated using the trapezoidal rule. Data are expressed as mean6 s.e.m.–2Non-encapsulated DOX AUC values represent 0–8-hr time period.–3Significantly different from the noValspodar group atp , 0.05.

138 KRISHNA ET AL.

pattern of fluorescence images in the presence of Valspodar wasconsistent with intracellular accumulation of DOX based onnormal-phase and autofluorescence images of the same fields.Valspodar also caused an apparent increase in intracellular DOXaccumulation after injection of EPC/Chol DOX and PEG-DSPE/DSPC/Chol DOX as revealed by the more intense DOX fluores-cence at 1 and 4 hr compared with mice treated in the absence ofValspodar (Fig. 4). Evaluation of intracellular tumor DOX accumu-lation for mice treated with either liposomal formulation could notbe assessed at later time points due to increased diffuse (presum-ably extracellular) fluorescence that precluded even qualitativeevaluations of intracellular DOX in the presence or absence ofValspodar.

The tumor levels of bioavailable DOX followed the order:non-encapsulated DOX (least), EPC/Chol DOX (intermediate),PEG-DSPE/DSPC/Chol DOX (highest). These trends were similarboth in the presence and absence of Valspodar, suggesting that: (1)increasing anticancer drug tumor delivery using saturated lipidformulations alone results in increased intracellular accumulation;(2) Valspodar pre-treatment causes increases in tumor drug reten-tion and increased intracellular DOX accumulation in tumor cells.

DISCUSSION

Our results demonstrate that liposome properties influencingdrug retention and tumor accumulation properties significantlyaffect the susceptibility of encapsulated DOX to PGP blockade byValspodar in both malignant and healthy tissues expressing thisdrug efflux pump. For EPC/Chol DOX, a rapid drug-releasingsystem, the dose reduction required when combined with Valspodarwas comparable to that required for non-encapsulated DOX in thepresence of an MDR modulator (Krishna and Mayer, 1997). Incontrast, the toxicity of PEG-DSPE/DSPC/Chol DOX was notaffected by Valspodar administration. These toxicity observationscorrelated with the degree of DOX retention within liposomes. Theincreased toxicity of DOX when administered in EPC/Chol lipo-somes in the presence of Valspodar was consistent with theincreased plasma levels in the circulation, which appeared to berelated to increased non-encapsulated DOX released from theliposomes as revealed by elevated DOX concentrations in Micro-con filtrates. In contrast, the drug-to-lipid ratio of the PEG-DSPE/DSPC/Chol DOX system remained constant over time, indicatingconsiderably less non-encapsulated drug exposure occurring with

FIGURE 4 – Tumor confocal images of non-encapsulated DOX (5 mg/kg), EPC/Chol DOX (5 mg/kg) or PEG-DSPE/DSPC/Chol DOX (5mg/kg), in the presence and absence of Valspodar. Scale bar5 25 µm. SCID/RAG2 mice were treated with the various combinations at a DOXdose of 5 mg/kg, injected i.v., alone or in the presence of Valspodar (p.o., 100 mg/kg; 4 hr before DOX). Following 1 and 4-hr post-DOXadministration, tumor samples were aseptically dissected and placed in PBS-containing tubes, maintained on ice. These samples were directlyimaged fresh by placing them on concave slides under the CFM (see Material and Methods). Identical settings on the CFM and identicalprocessing times facilitated a comparison of DOX fluorescence intensity in the presence and absence of Valspodar.

139COMPLETE REVERSAL OF MDR IN SOLID TUMORS

these liposomes, corroborating the reductions in toxicity for thisformulation.

Our observations with DSPC/Chol liposomes indicated higherlevels of drug exposure in the liver in the presence and absence ofValspodar (Krishna and Mayer, 1997). Surprisingly, no evidence ofValspodar-mediated toxicities or changes in DOX pharmacokinet-ics were observed with DSPC/Chol liposomes. Our results indicatethat this avoidance of PK interactions in the liver is related toaltered metabolic processes compared to non-encapsulated drug.While EPC/Chol DOX closely resembled non-encapsulated DOXwith respect to metabolite formation, the other liposome systemtested here did not favor the formation of DOXol, which is formedwhen DOX is metabolized by the cytosolic aldo-keto reductase.Further, higher levels of the 2 aglycones, DOXone and 7-deoxy-DOXone, were observed with the EPC/Chol system compared toPEG-DSPE/DSPC/Chol DOX. Valspodar co-administration alsocaused marked increases in DOX and DOXone levels withEPC/Chol DOX, but did not significantly alter the metaboliteformation of PEG-DSPE/DSPC/Chol DOX.

It appears that differences in DOX metabolism when DOX isencapsulated within sterically stabilized liposomes may in part beresponsible for the reduced toxicity and lack of drug-modulator PKinteractions observed for this L-DOX-Valspodar combination. Twopossible consequences are: (1) slower or less biotransformation ofDOX due to the non-leaky nature of the sterically stabilizedL-DOX, and/or (2) metabolism of DOX mediated by a differentmechanism dictated by the liposomal carrier, resulting in com-pounds that are not PGP substrates. It is not known whether theaglycone metabolites are substrates of PGP, consequently thetoxicological implications of these Valspodar-mediated increases inaglycone levels remain to be investigated. The improvementsobserved with the stable liposomal systems may be due to reducedDOX renal and biliary excretion since liposomes may be expectedto deliver DOX preferentially to the phagocytic Kupffer cells in theliver (Scherphofet al.,1983).

Our present investigations document complete chemosensitiza-tion in an MDR human xenograft solid tumor model by combiningPGP blockade with liposomal DOX. While tumor growth suppres-sion studies indicated modest anti-tumor activity against the MDR

solid tumors for the leaky EPC/Chol DOX in the presence ofValspodar, significant reductions in tumor growth were observedwith the sterically stabilized stabilized liposomal system in thepresence of Valspodar. This increased efficacy of the stericallystabilized liposomal DOX-Valspodar combination was accompa-nied by the absence of significant toxicity, PK and metabolicchanges. This enhanced therapeutic activity appeared to correlatewith increased tumor accumulation of DOX afforded by thesterically stabilized liposomes and secondly Valspodar-mediatedincreased retention of the drug at the tumor site. These interpreta-tions are based on 2 key observations. Increases in DOX tumoraccumulation in the presence of Valspodar were not associated withincreased drug delivery caused by liposomes as noted by compa-rable liposomal lipid levels in the tumor following administrationin the presence and absence of Valspodar (Fig. 3a,b). Also, ourCFM results indicate an increase in intracellular DOX accumula-tion caused by Valspodar, consistent with the ability of this MDRmodulator to block PGP drug effluxin vivo.

In summary, the toxicity characteristics observed when lipo-somal DOX formulations are co-administered with Valspodarappear to be dictated by liposomal drug retention properties whereincreased drug retention within liposomes reduces toxicity anddrug-modulator interactions. In comparison, tumor growth suppres-sion appears to be primarily dictated by a combination of liposomaldrug tumor accumulation as well as intracellular drug retentioncaused by the MDR modulator. Consequently, improvements inMDR reversal properties observed with liposomal DOX aremediated by: (1) Valspodar blockade of PGP-mediated increaseddrug retention by tumor cells; (2) increased local drug exposureprovided by liposomes. As seen from our results, these factors aremaximized with sterically stabilized liposomal DOX systems.

ACKNOWLEDGEMENTS

This work was supported by the National Cancer Institute ofCanada with funds from the Canadian Cancer Society. RK is a BCMedical Services Foundation Predoctoral fellow. The authorsacknowledge the able technical assistance provided by Ms. N.McIntosh and Ms. D. Masin in the animal experimentation.

REFERENCES

ANDERSEN, A., WARREN, D.J. and STORDAL, L., A sensitive and simplehigh-performance liquid chromatographic method for the determination ofdoxorubicin and its metabolites in plasma.Ther. Drug Monit.,15,455–458(1993).

BALLY, M.B., MAYER, L.D., HOPE, M.J. and NAYAR, R., Pharmacodynamicsof liposomal drug carriers: methodological considerations. In: G. Gregoria-dis (ed.),Liposome technology, Vol. III, Interactions of liposomes withbiological milieu,pp. 27–41, CRC Press, Boca Raton, FL, (1993).

BOOTE, D.J., DENNIS, I.F., TWENTYMAN, P.R., OSBORNE, R.J., LABURTE, C.,HENSEL, S., SMYTH, J.F., BRAMPTON, M.H. and BLEEHEN, N.M., Phase Istudy of etoposide with PSC-833 as a modulator of multidrug resistance inpatients with cancer.J. clin. Oncol.,14,610–618 (1996).

BOOTH, C.L., BROUWER, K.R. and BROUWER, K.L.R., Effect of multidrugresistance modulators on the hepatobiliary disposition of doxorubicin in theisolated perfused rat liver.Cancer Res.,58,3641–3648 (1998).

DERKSEN, J.T.P., MORSELT, H.W.M. and SCHERPHOF, G.L., Processing ofdifferent liposome markers afterin vitro uptake of immunoglobulin-coatedliposomes by rat liver macrophages.Biochim. biophys. Acta,931, 33–40(1987).

GATMAITAN , Z. and ARIAS, I., Structure and function of P-glycoprotein innormal liver and small intestine.Adv. Pharmacol.,24,77–96 (1993).

GIBALDI , M., and PERRIER, D., Pharmacokinetics,Marcel Dekker, NewYork (1982).

KELLER, R.P., ALTERMATT, H.J., DONATSCH, P., ZIHLMANN , H., LAISSUE, J.A.and HIESTAND, P.C., Pharmacologic interactions between the resistance-modifying cyclosporine SDZ PSC 833 and etoposide (VP 16-213) enhancein vivocytostatic activity and toxicity.Int. J. Cancer,51,433–438 (1992a).

KELLER, R.P., ALTERMATT, H.J., NOOTER, K., POSCHMANN, G., LAISSUE, J.A.,

BOLLINGER, P. and HIESTAND, P.C., SDZ PSC 833, a non-immunosuppres-sive cyclosporine: its potency in overcoming P-glycoprotein mediatedmultidrug resistance of murine leukemia.Int. J. Cancer,50, 593–597(1992b).

KRISHNA, R. and MAYER, L.D., Liposomal doxorubicin circumvents toxici-ties associated with non-encapsulated drug-PSC 833 interactions, resultingin effective therapy of multidrug resistant solid tumors.Cancer Res.,57,5246–5253 (1997).

LEONESSA, F., GREEN, D., LICHT, T., WRIGHT, A., WINGATE-LEGETTE, K.,LIPPMAN, J., GOTTESMAN, M.M. and CLARKE, R., MDA435/LCC6 andMDA435/LCC6MDR1: ascites models of human breast cancer.Brit. J.Cancer,73,154–161 (1996).

MAYER, L.D. and ST-ONGE, G., Determination of free and liposome-associated doxorubicin and vincristine levels in plasma under equilibriumconditions employing ultrafiltration technique.Anal. Biochem.,232,149–157 (1995).

MAYER, L.D., TAI, L.C.L., KO, D.S.C., MASIN, D., GINSBERG, R.S., CULLIS,P.R. and BALLY, M.B., Influence of vesicle size, lipid composition, anddrug-to-lipid ratio on the biological activity of liposomal doxorubicin inmice.Cancer Res.,49,5922–5930 (1989).

MILLER, T., DAHLBERG, S. and GROGAN, T., Southwest Oncology Groupclinical trials of chemosensitizers in non-Hodgkin’s lymphomas.Antican-cer Drugs,5, 73 (1994).

NOOTER, K., OOSTRUM, R. and DEURLOO, J., Effects of verapamil on thepharmacokinetics of daunomycin in the rat.Cancer Chemother. Pharma-col.,20,176–178 (1987).

PAPAHADJOPOULOS, D., ALLEN, T.M., GABIZON, A., MAYHEW, E., MATTHAY,K., HUANG, S.K., LEE, K.D., WOODLE, M.C., LASIC, D.D., REDEMANN, C.

140 KRISHNA ET AL.

and MARTIN, F.C., Sterically stabilized liposomes: improvements in pharma-cokinetics and antitumor therapeutic efficacy.Proc. nat. Acad. Sci. (Wash.),88,11460–11464 (1991).SARRIS, A.H. and 13 OTHERS, Cyclosporin A does not reverse clinicalresistance to paclitaxel in patients with relapsed non-Hodgkin’s lymphoma.J. clin. Oncol.,14,233–239 (1996).SCHERPHOF, G.L., ROERDINK, F., DIJKSTRA, J., ELLENS, H., DE ZANGER, R.and WISSE, E., Uptake of liposomes by rat and mouse hepatocytes andKupffer cells.Biol. Cell.,47,47 (1983).SPEEG, K.V. and MALDONADO, A.L., Effect of the nonimmunosuppressivecyclosporin analog SDZ PSC-833 on colchicine and doxorubicin biliarysecretion by the ratin vivo. Cancer Chemother. Pharmacol.,34, 133–136(1994).SPEEG, K.V., MALDONADO, A.L., LIACI, J. and MUIRHEAD, D., Effect of

cyclosporine on colchicine secretion by a liver canalicular transporterstudiedin vivo. Hepatology,15,899–903 (1992).

VRIGNAUD, P., LONDOS-GAGLIARDI , D. and ROBERT, J., Hepatic metabolismof doxorubicin in mice and rats.Europ. J. Drug Metab., Pharmacokinetics,11,101–105 (1986).

WISHART, G.C., BISSETT, D., PAUL, J., JODRELL, D., HARNETT, A., HABE-SHAW, T., KERR, D.J., MACHAM, M.A., SOUKOP, M. and LEONARD, R.C.,Quinidine as a resistance modulator of epirubicin in advanced breastcancer: mature results of a placebo-controlled trial.J. clin. Oncol.,12,1771–1777 (1994).

WU, N.Z., BRAUN, R.D., GABER, M.H., LIN, G.M., ONG, E.T., SHAN, S.,PAPAHADJOPOULOS, D. and DEWHIRST, M.W., Simultaneous measurement ofliposome extravasation and content release in tumors.Microcirculation, 4,83–101 (1997).

141COMPLETE REVERSAL OF MDR IN SOLID TUMORS