Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Applications of liposome technology in modulating p-glycoprotein mediated multidrug resistance Krishna, Rajesh 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1999-389200.pdf [ 14.61MB ]
Metadata
JSON: 831-1.0099449.json
JSON-LD: 831-1.0099449-ld.json
RDF/XML (Pretty): 831-1.0099449-rdf.xml
RDF/JSON: 831-1.0099449-rdf.json
Turtle: 831-1.0099449-turtle.txt
N-Triples: 831-1.0099449-rdf-ntriples.txt
Original Record: 831-1.0099449-source.json
Full Text
831-1.0099449-fulltext.txt
Citation
831-1.0099449.ris

Full Text

APPLICATIONS OF LIPOSOME TECHNOLOGY IN MODULATING P-GLYCOPROTEEV MEDIATED MULTIDRUG RESISTANCE by Rajesh Krishna B. Pharm., University of Kerala, India, 1990 M. Pharm., Banaras Hindu University, India, 1992 M. Sc. (Phar), University of British Columbia, 1995 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY (Pharmaceutical Sciences) in THE FACULTY OF GRADUATE STUDIES Faculty of Pharmaceutical Sciences We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA January, 1999 © Rajesh Krishna, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of P rffirnpt-^nPt ceUT/cAt, SClGrtc&S The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT This thesis tested the hypothesis that the inadequate tumor delivery of anticancer agents and selectivity of PGP modulation are primarily responsible for the attenuated therapy of extravascular MDR solid tumors overexpressing PGP. PSC 833 was selected as the MDR modulator for in vivo applications, based on potency, extent of MDR reversal and latent modulating activity. Whereas co-administration of PSC 833 with non-encapsulated DOX required a 3-fold reduction in anticancer drug dose due to increased toxicity, the toxicity of DOX encapsulated in 100 nm diameter DSPC/Chol vesicles was comparable in the presence and absence of PSC 833. Efficacy studies in P388/ADR solid tumor bearing mice revealed that co-administration of non-encapsulated DOX and PSC 833 at the maximum tolerated dose (MTD) resulted in modest MDR modulation and antitumor activity. However, significant tumor growth suppression was observed when DSPC/Chol DOX was co-administered with PSC 833. Pharmacokinetic studies indicated a PSC 833-induced increase in terminal elimination phase and a concomitant 10-fold increase in plasma DOX AUC following treatment with non-encapsulated DOX whereas minimal pharmacokinetic alterations were observed with DOX encapsulated in DSPC/Chol liposomes. The relative toxicities and degree of MDR reversal were also examined using leaky and non-leaky liposomal DOX in the presence and absence of PSC 833. While PSC 833 necessitated a 3-fold dose reduction for the leaky EPC/Chol DOX in the multiple dosing schedule, no dose reductions were required for the non-leaky PEG-DSPE/DSPC/Chol DOX. In human breast carcinoma MDR xenograft tumor bearing mice, co-administration of EPC/Chol DOX and PSC 833 resulted in modest modulation and antitumor activity. n However, co-administration of PSC 833 with non-leaky liposomal DOX resulted in significant tumor growth suppression comparable to that obtained using drug sensitive tumors. PSC 833 caused a 2.6-fold increase in plasma DOX AUC when co-administered with EPC/Chol DOX, whereas PSC 833 did not alter the pharmacokinetics of DOX encapsulated in sterically stabilized liposomes. The renal (CLr) and biliary (CLb) clearances of non-encapsulated DOX were found to be significantly inhibited by PSC 833. While the CLr and CLb of the EPC/Chol and PEG-DSPE/DSPC/Chol DOX were considerably lower than for non-encapsulated drug, PSC 833 co-administration resulted in minor inhibitions of CLr and CLb for the EPC/Chol DOX formulation. PEG-DSPE/DSPC/Chol DOX clearances were unaltered by the MDR modulator. These results confirmed the ability of liposome encapsulation to modify the distribution and metabolism of DOX in a manner that alleviates the effects of PSC 833 on DOX pharmacokinetics observed using non-encapsulated drug. iii TABLE OF CONTENTS Abstract ii Table of Contents iv List of Figures viii List of Tables xv List of Schemes xvii Abbreviations xviii Acknowledgments xxiv Dedication xxv Chapter 1 Introduction 1 1.1. Hypothesis and goal 1 1.2. Background information: Biology of Multidrug resistance (MDR) 3 1.2.1. Non-cellular resistance mechanisms 4 1.2.2. Cellular-based resistance mechanisms 5 1.2.2.1. Non-classical MDR phenotypes 5 1.2.2.1.1. GlutathioneS-transferase 6 1.2.2.1.2. Topoisomerase activity 7 1.2.2.1.3. Altered apoptosis regulation 9 1.2.2.2. Transport-based classical mechanisms 11 1.2.2.2.1. MDR associated protein (MRP) 12 1.2.2.2.2. P-glycoprotein (PGP) 14 1.3. Clinical relevance of multidrug resistance 19 1.4. Therapeutic strategies to overcome PGP-mediated MDR 20 1.4.1. First generation modulators 22 1.4.2. Second generation modulators 23 1.4.3. Third generation modulators 25 1.5. Complications with chemosensitizers 25 1.6. Drug delivery approaches to improve chemotherapy of MDR tumors 30 1.6.1. Liposome components 33 1.6.1.1. Phospholipids 33 1.6.1.2. Cholesterol 38 1.6.2. Preparation of liposomes 40 1.6.2.1. Multilammelar vesicles 40 1.6.2.2. Unilammelar vesicles 41 1.6.3. Drug entrapment within liposomes 44 1.6.3.1. Passive entrapment 45 1.6.3.2. Active entrapment 46 1.6.4. Biological fate of systemically administered liposomes 49 1.6.4.1. Liposome-protein interactions 49 1.6.4.2. Drug release from liposomes 51 1.6.4.3. Factors affecting liposome circulation lifetimes 53 1.6.4.4. Strategies to improve circulation longevity 5 5 1.6.5. Liposomes in cancer chemotherapy 60 1.6.5.1. Normal vasculature 61 1.6.5.2. Tumor vasculature 62 1.6.6. MDR 65 1.6.6.1. Applications of liposomal anticancer drugs for MDR 65 1.6.6.2. Use of liposomes for the delivery of MDR modulators 68 1.7. Thesis rationale and specific aims 70 1.7.1. Rationale 70 1.7.2. Specific aims 71 Chapter 2 Comparative in vitro evaluation of DOX cytotoxicity, cellular uptake and retention, and intracellular distribution in the presence and absence of MDR modulators: Selection of MDR modulator with optimum reversal properties for in vivo applications... 74 2.1. Introduction 74 2.2. Materials and methods 80 2.2.1. Chemicals, animals, cells and supplies 80 2.2.2. General equipment 81 2.2.3. Drug solutions 82 2.2.4. Cell lines and culture 82 2.2.5. Cytotoxicity experiments 84 2.2.6. Cellular DOX uptake and retention studies.... 86 2.2.6.1. Flow cytometric studies 86 2.2.6.2. Cellular DOX concentrations 87 2.2.7. Intracellular drug distribution 88 2.3. Results 89 2.3.1. Cytotoxicity experiments 89 2.3.1.1. Continuous exposure 89 2.3.1.2. Pulsed exposure 92 2.3.2. DOX uptake and release kinetics 95 2.3.3. Intracellular DOX distribution 104 2.4. Discussion 106 Chapter 3 Toxicity, antitumor activity, and pharmacokinetics of non-encapsulated and DSPC/Chol liposome-encapsulated DOX, in the presence and absence of PSC 833, using the murine lymphocytic P388/ADR solid tumor model I l l 3.1. Introduction I l l 3.2. Materials and methods 116 3.2.1. Materials... 116 3.2.2. Liposome preparation 117 3.2.3. Drug encapsulation 120 3.2.4. PSC 833 formulation and use 120 3.2.5. Toxicity evaluation studies 122 3.2.6. Cell preparation for efficacy studies 123 3.2.7. In vivo antitumor activity 124 3.2.8. Pharmacokinetics and tissue distribution 125 3.3. Results 127 3.3.1. Toxicity 127 3.3.2. Efficacy 131 3.3.3. Pharmacokinetics and biodistribution 133 3.4. Discussion 142 Chapter 4 Influence of liposomal drug retention and tumor accumulation properties on the toxicity, antitumor activity, and pharmacokinetics of DOX, in the presence and absence of PSC 833 using the MDA435LCC6 MDR1 human xenograft solid tumor model 147 4.1. Introduction 147 4.2. Materials and methods 151 4.2.1. Materials 151 4.2.2. Liposomes and drug preparation 152 4.2.3. Liposome characterization 153 4.2.4. Toxicity evaluation studies 154 4.2.5. In vivo antitumor activity 155 vi 4.2.6. Pharmacokinetics and tissue distribution 157 4.2.7. Confocal microscopy and imaging studies 162 4.3. Results 164 4.3.1. Toxicity 164 4.3.1.1. Single injection 166 4.3.1.2. Multiple injection 167 4.3.2. Efficacy 168 4.3.2.1. Efficacy experiment #1 168 4.3.2.2. Efficacy experiment #2 174 4.3.3. Pharmacokinetics and tissue distribution 181 4.3.3.1. Plasma drug kinetics 182 4.3.3.2. Tumor drug kinetics 189 4.3.3.3. Liver drug kinetics 194 4.3.3.4. Kidney drug kinetics 200 4.3.4. Confocal imaging 201 4.4. Discussion 209 Chapter 5 Influence of liposomal encapsulation on the renal and hepatobiliary disposition of DOX in the presence and absence of PSC 833 using an instrumented rat model 215 5.1. Introduction 215 5.2. Materials and methods 220 5.2.1. Materials 220 5.2.2. Liposomes and drug preparation 221 5.2.3. Liposome characterization 222 5.2.4. Analytical methods 222 5.2.4.1. HPLC assay for DOX, DOXol 222 5.2.4.2. Lipid radioactivity 223 5.2.5. Animal experiments 223 5.3. Results 227 5.3.1. Plasma pharmacokinetics 227 5.3.2. Urinary excretion 232 5.3.3. Biliary excretion 237 5.4. Discussion 239 Chapter 6 Summarizing discussion 245 References 252 vii LIST OF FIGURES Page Figure l . l .A Structure of the PGP molecule. In the structure of PGP, the 15 circled areas indicate ATP-binding sites, curly lines represent N-linked carbohydrates and filled circles represent amino acids that differ between mouse and human mdrl (Adapted from Kane et al., 1990). Figure 1.1.B Function of the PGP pump. The model illustrates a protein 15 which uses ATP energy to actively efflux drug substrate across the plasma membrane. The MDR modulator (chemosensitizer) may act as a competitive inhibitor by occupying drug binding sites or as a non-competitive inhibitor at chemosensitizer binding sites (Modified from Ford, 1995). Figure 1.2. (A) An illustration of the structure of a liposome bilayer. 32 Lipids upon hydration with aqueous buffer adopt a bilayer configuration. (B) Illustration of three types of liposomes, multilamellar vesicles, large unilamellar vesicles, and small unilamellar vesicles. Figure 1.3. Diagrammatic representation of phospholipid chemical 34 structures. Illustrated are the headgroup, glycerol backbone, and the acyl chains. The figure also shows various headgroup and acyl chain fatty acid compositions for commonly occurring phospholipids. Figure 1.4. Chemical structure of cholesterol molecule (A), diagrammatic 39 representation of the orientation of cholesterol with PC (B), and role of cholesterol when incorporated in the liposome bilayer (C.) Figure 1.5. Diagrammatic representation of an active loading process. A 48 pH gradient is established when the liposomes (with an interior pH 4.0) are titrated with 0.5 M sodium carbonate to create an exterior pH of 7.8. The equilibrium redistribution of a weak base in response to this pH gradient across the liposome membrane is shown. The neutral form of the molecule is membrane permeable. Figure 1.6. In vivo drug-to-lipid ratio determined in plasma of mice 53 treated with 120 nm EPC/Chol, DSPC/Chol, and PEG-DSPE/DSPC/Chol liposomal DOX. Liposomal lipid and DOX were quantified using the methods described in Chapters 3 and 4 (mean ± SD). viii Figure 1.7. Chemical structure of the PEG-DSPE molecule and an 57 illustration of a PEG-coated liposome when this polymerized lipid is incorporated in the initial lipid mixture. Note the polymer coating on interior and exterior surfaces of the liposome. Figure 2.1. Representative plots of counts vs fluorescence for DOX 96 uptake in P388/ADR cells (Panel A), P388AVT cells (Panel B), and PSC 833 sensitized P388/ADR cells (Panel C), as evaluated by flow cytometry. The resulting peaks are single, sharp, and homogeneously constructed. Figure 2.2. A,B Flow cytometric evaluation of DOX uptake in 98 chemosensitized P388/ADR cells. The uptake of DOX was evaluated in P388/ADR cells pretreated with PSC 833 (1 uM), DNG (5 uM), Rol 1-2933 (10 uM), Ro44-5912 (10 uM), tamoxifen (10 uM), quinidine (10 uM), dexVRP (10 pM), prochlorperazine (10 uM), trans-E-flupenthixol (10 uM), and chlorpromazine (10 u.M) as described under Materials and Methods. Data are presented as mean ± standard deviation (n = 4). Figure 2.3. Flow cytometric evaluation of DOX uptake in 99 chemosensitized MCF7/ADR cells. The uptake of DOX was evaluated in MCF7/ADR cells pretreated with Rol 1-2933 (10 uM, •) , DNG (5 uM, A ) , PSC 833 (1 uM, T ) , and VRP (10 uM, •) as described under Materials and Methods in relation to DOX controls in the resistant (MCF7/ADR control (•)) and sensitive lines (MCF7AVT control (+)). Data are presented as mean ± standard deviation (n = 4). Figure 2.4. A,B DOX release profile in MDR P388/ADR cells. Panel A: 101 Effect of PSC 833 ( A , 1 uM), DNG (•, 5 uM) and VRP (•, 10 uM) on DOX retention properties. Cells were incubated with the modulator for 30 min after which DOX was added and further incubated for 90 min. The cells were then washed free of the modulator and DOX release was studied over 70 min using flow cytometry. Release curves are normalized to the P388/ADR resistant modulator-free DOX controls. Data are presented as mean ± standard deviation (n = 4). Panel B: Effect of extended pulse exposure times on DOX retention properties. Cells were exposed for 15 hours to MDR modulators in the presence (PSC 833 + DOX [•] or VRP + DOX [•]) or the absence (PSC 833 [A] or VRP [+]) of ix DOX. These cells were washed free of the modulator and DOX release was studied over a 60 min period using flow cytometry. Release curves are normalized to the P388/ADR resistant modulator-free DOX controls. Data are presented as mean ± standard deviation (n = 4). Figure 2.5. Intracellular DOX levels in P388/ADR, P388AVT, and 103 chemosensitized cells. DOX (10 pM) was incubated in the presence and absence of MDR modulator (PSC 833, PSC, 1 uM or VRP, 10 uM) for 260 min, following which cell associated DOX was quantified using spectrofluorimetry as described under Materials and Methods. Data are presented as mean ± standard deviation (n = 6). Figure 2.6. Representative DOX fluorescence micrographs of DOX 105 controls (P388/WT and P388/ADR) and modulator treated (VRP, PSC 833) P388/ADR cells Figure 3.1. Structures of cyclosporin A and its non-immunosuppressive 112 analog, PSC 833. Note the substitutions at positions Rl and R2. Figure 3.2. Representative NICOMP gaussian analysis of vesicle size 118 distribution as determined using quasi-elastic light scattering technique. DSPC/Chol liposomes, prepared using the extrusion technique, were diluted in normal saline and analyzed for vesicle size distribution at a count rate of 341 kHz. The figure represents the size distribution histogram where liposome diameter (mean ± standard deviation) was determined as 117.8 ± 28.8 nm (Chi2 = 0.22). Figure 3.3. Representative mass spectrum of PSC 833 using electrospray 121 MS system Figure 3.4. Liposomal encapsulation circumvents PSC 833-mediated 129 increased toxicity of i.v. DOX. Intravenous DOX was administered in the free (7.5 mg/kg and 20 mg/kg) and liposomal (50 mg/kg and 70 mg/kg) forms, alone or in combination with oral PSC 833 (100 mg/kg) in normal BDF1 mice and the change in body weight monitored over a 15-day period. Legends: liposomal DOX (70 mg/kg) [•]; liposomal DOX (50 mg/kg) with PSC 833 [•]; non-encapsulated DOX (20 mg/kg) [A] ; non-encapsulated DOX (20 mg/kg) with PSC 833 [A]; non-encapsulated DOX (7.5 mg/kg) with PSC x 833 [+]. Data are presented as group mean value (n = 5/group). Figure 3.5. Antitumor efficacy of free and liposomal DOX against 132 P388/ADR solid tumors in the absence (Panel A) and presence (Panel B) of co-administered PSC 833. Panel A (PSC 833 free) groups: P388/ADR untreated control tumors (•), P388/ADR tumors treated with non-encapsulated DOX 7.5 mg/kg (•) or liposomal DOX 10 mg/kg (A) , and P388/WT sensitive tumors treated with non-encapsulated DOX 7.5 mg/kg (T) . Panel B (incorporating PSC 833) groups: P388/ADR control untreated tumors (•), P388/ADR tumors . treated with p.o. PSC 833 100 mg/kg (•), P388/ADR tumors treated with non-encapsulated DOX 2.0 mg/kg in conjunction with 100 mg/kg p.o. PSC 833 (T) , and P388/ADR tumors treated with liposomal DOX 10 mg/kg in conjunction with 100 mg/kg p.o. PSC 833 (A) . Data are expressed as mean ± standard error of the mean (n = 10/group). Figure 3.6. Plasma (A),tumor (B), liver (C), spleen (D), kidney (E), and 136, heart (F) DOX concentration-time profiles following 137 administration of non-encapsulated i.v. DOX 7.5 mg/kg in the absence (•) or presence of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (•); non-encapsulated i.v. DOX 2.0 mg/kg in the presence of PSC 833 100 mg/kg p.o. (A) ; and i.v. liposomal DOX (55:45, DSPC/Chol, 100 nm) 10 mg/kg in the absence (T ) or presence of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (•). Data are expressed as mean ± standard error of the mean (n = 3 mice/time point; rx-21 mice/group). Figure 4.1. Representative chromatogram of DOX and DOX metabolite 160 standards (DOXol, DOXone, and 7-deoxyDOXone) spiked in a 10% liver homogenate Figure 4.2. Antitumor efficacy of free (F) (Panel A), EPC/Chol DOX 170, (Panel B), DSPC/Chol DOX (Panel C), and PEG- 171 DSPE/DSPC/Chol DOX (Panel D) against MDA435LCC6 WT or MDR1 human xenograft solid tumors in the absence and presence of co-administered PSC 833. Untreated WT and MDR controls as well as free (F) DOX treated WT and MDR tumors are presented in Panel A. Data are expressed as mean ± standard error of the mean (n = 8/group). For legends, see individual panels. Figure 4.3 Figure 4.4 Figure 4.5 Antitumor efficacy of free (F) DOX (Panel A), EPC/Chol 176 DOX (Panel B), DSPC/Chol DOX (Panel C), and PEG-DSPE/DSPC/Chol DOX (Panel D) against MDA435LCC6 WT or MDR1 tumors + PSC 833. MDA435LCC6 tumors were grown on mfp of female SCJD/Rag2 mice. Oral PSC 833 (100 mg/kg) and i.v. DOX treatments were initiated once tumors were established (20-100 mg) and were given on days 1, 5, and 9 at the indicated doses of non-encapsulated and liposomal DOX. PSC 833 was administered 4 h prior to DOX injection. Data are expressed as mean + s.e.m. (n = 8/group; for EPC/Chol DOX + PSC 833, n = 16/group). For legends, see panels. DOX (Panel A) and liposomal lipid (Panel B) concentration- 185 time profiles in plasma following administration of free DOX 5 mg/kg in the absence (open triangles) or presence of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled triangles); EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence of PSC 833 (administered 4 hours prior 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 of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled squares). Data are expressed as mean ± standard error of the mean (n = 3 mice/time point; n = 24/group). Tumor DOX (Panel A) and liposomal lipid (Panel B); Liver 190 DOX (Panel C), and kidney DOX (Panel D) concentration- (A,B), time profiles following administration of free DOX 5 mg/kg in 193 the absence (open triangles) or presence of PSC 833 (C,D) (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled triangles); EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence of PSC 833 (administered 4 hours prior 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 of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled squares). Data are expressed as mean ± standard error of the mean (n = 3/time point; n = 24/group). xii Figure 4.6. Liver (Panel A) and kidney (Panel B) DOXone concentration- 199 time profiles following administration of EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence of PSC 833 (administered 4 hours prior 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 of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled squares). Data are expressed as mean ± standard error of the mean (n = 3 mice/time point; n = 24/group). Figure 4.7.A A comparison of two sample treatments on DOX visualization 203, using CFM. Panel A represents the image after PEG- 204 DSPE/DSPC/Chol -liposomal DOX is injected into a defined mass of freshly isolated muscle tissue, at a concentration of 10 ug/g, (from SCID/Rag2 mice) and visualized fresh. Panel B is a representation of PEG-DSPE/DSPC/Chol -liposomal DOX at identical concentration in muscle, but subjected to a standard cyrofixation protocol before visualization under the confocal microscope. Note the distinct DOX fluorescence signal consistent with nuclear localization following the latter protocol. Figure 4.7.B An illustration of DOX fluorescence quenching properties 203, when encapsulated within liposomes. Freshly isolated muscle 205 tissue from SCJX)/Rag2 mice were injected with known amounts of PEG-DSPE/DSPC/Chol -liposomal DOX (10 ug/g; Panel A), empty PEG-DSPE/DSPC/Chol -liposomes labeled with Dil (Panel B), and free DOX (10 ug/g; Panel C). Note the distinct strong DOX fluorescence in Panel C, but not in Panel B, indicating substantial quenching with PEG-DSPE/DSPC/Chol -liposomal DOX. The visualization of Dil label confirms the presence of empty PEG-DSPE/DSPC/Chol liposomes. Figure 4.8. Tumor confocal images of free DOX (5 mg/kg), EPC/Chol 207, DOX (5 mg/kg) or PEG-DSPE/DSPC/Chol -liposomal DOX 208 (5 mg/kg), in the presence and absence of PSC 833. SCID/Rag2 mice were treated with various combinations at a DOX dose of 5 mg/kg, injected i.v., alone or in the presence of PSC 833 (p.o., 100 mg/kg; 4 h before DOX). Following 1 and 4 h post-DOX administration, tumor samples were aseptically dissected and placed in PBS containing tubes, maintained on ice. These samples were directly imaged fresh by placing them on concave slides under the CFM (see xiii Materials and Methods). Identical settings on the CFM and identical processing times facilitated a comparison of DOX fluorescence intensity in the presence and absence of PSC 833. Figure 5.1. Plasma DOX concentration-time profdes of free DOX (Panel 229 A), EPC/Chol DOX (Panel B), and PEG-DSPE/DSPC/Chol DOX (Panel C) at a DOX dose of 5 mg/kg i.v. in the absence and presence of PSC 833 (50 mg/kg, p.o., administered 4 h prior to DOX). Data are represented as mean ± standard deviation (n = 3/group). Note the different x and y axis scale for Panels A, B and C. Figure 5.2. Urinary (Panel A) and biliary (Panel B) DOX excretion 234 profiles following administration of free, EPC/Chol, and PEG-DSPE/DSPC/Chol DOX at a DOX i.v. dose of 5 mg/kg, in the absence and presence of PSC 833 (50 mg/kg, p.o., 4 h prior to DOX). Data are represented as mean ± standard deviation (n = 3/group). xiv LIST OF TABLES Page Table 1.1. Temperature (Tc) of the gel to liquid-crystalline phase 37 transition and charge of various phospholipids Table 2.1. Summary of fold reversal and modulator concentrations in 91 P388/ADR and MCF7/ADR cells in the 72 h continuous exposure Table 2.2. Cytotoxicity profdes in continuous and pulsed exposures for 93 PSC 833, VRP and Rol 1-2933 inP388/ADR cells Table 2.3. Cytotoxicity profiles in continuous and pulsed exposures for 94 PSC 833, VRP, DNG and Rol 1-2933 in MCF7/ADR cells Table 3.1. Plasma and tissue distribution of DOX administered with or 135 without PSC 833 in free or liposomal encapsulated forms Table 4.1. Toxicity as a function of DOX dose for liposomal 165 formulations in single and multiple injection schedules Table 4.2. Summary of relative tumor weights and percent T/C of free 180 and liposomal DOX treatments in the presence and absence of PSC 833 Table 4.3.A Effect of PSC 833 on plasma DOX pharmacokinetic 186 parameters when DOX is administered in liposome encapsulated forms Table 4.3.B Effect of PSC 833 on plasma liposomal lipid pharmacokinetic 186 parameters when DOX is administered in liposome encapsulated forms Table 4.4. Evaluation of free DOX equivalents in mouse plasma using 188 Microcon-30 filters Table 4.5.A Tumor DOX distribution of free and liposomal (EPC/Chol and 191 PEG-DSPE/DSPC/Chol) DOX in the presence and absence of PSC 833 Table 4.5.B Liver DOX and DOXone distribution characteristics following 195 administration of EPC/Chol and PEG-DSPE/DSPC/Chol DOX in the presence and absence of PSC 833 xv Table 5.1. Summary of plasma pharmacokinetic parameters for free, 228 EPC/Chol, and PEG-DSPE/DSPC/Chol liposomal DOX in the presence and absence of PSC 833 Table 5.2. Summary of urine and bile DOX pharmacokinetic parameters 235 for free, EPC/Chol and PEG-DSPE/DSPC/Chol DOX in the presence and absence of PSC 833 xvi LIST OF SCHEMES Page Scheme 2.1. Chemical structures of neutral, anionic and cationic species of DOX 77 Scheme 4.1. Metabolism of DOX to DOXol, DOXone, and 7-deoxyDOXone 196 xvii LIST OF ABBREVIATIONS ABC ATPase-Binding Cassette ADR Adriamycin AML Acute myeloid leukemia ANOVA Analysis of Variance ApoA-1 Apolipoprotein A- l Ara-C Cytosine arabanoside ATP Adenosine triphosphate AUC Area under the curve AUMC Area under the first moment curve bcl2 B cell lymphoma 2 BD Bile duct BSO Buthionine Sulfoximine C26 Colon carcinoma CAE Cumulative amount excreted CEM Human acute lymphoblastic leukemia CFM Confocal microscopy CFTR Cystic fibrosis transmembrane regulator protein CFJDE (3H) Cholesteryl hexadecyl ether, tritiated CHO Chinese hamster ovary Choi Cholesterol CL b Biliary clearance xviii CLL Chronic lymphocytic leukemia CL P Plasma clearance CL, Renal clearance Cmax Peak concentration achieved cMOAT Cannalicular multispecific organic anion transporter co2 Carbon dioxide CsA Cyclosporin A CV Coefficient of variation CYP3A Cytochrome P450 isoform 3 A DHFR Dihydrofolate reductase DMSO Dimethylsulphoxide DNA Deoxyribonucleic acid DNG Dexniguldipine DOX Doxorubicin DOXol Doxorubicinol DOXone Doxorubicinone 7-deoxyDOXone 7-deoxydoxorubicinone DPM Disintegration's per minute DSPC l,2-Distearoyl-sn-glycero-3-phosphocholine DSPE Distearoylphosphoethanolamine DMPC l,2-Dimyristoyl-sn-glycero-3-phosphocholine EDTA Ethylene diamine tetracetic acid xix EPC Egg phosphatidylcholine FBS Fetal bovine serum FITC Fluorescein isothiocyanate FR Fold reversal GSH Reduced glutathione GST Glutathione transferase H 2 S0 4 Sulfuric acid FfBSS Hank's buffered salt solution HCT-15 Human colon (colorectal) adenocarcinoma HPLC High performance (or pressure) liquid chromatography i.m. Intramuscular i.p. Intraperitoneal i.v. Intravenous IC50 Concentration required to achieve 50% cell kill IgG Immunoglobulin G JVC Jugular vein catheter K562 Human chronic myelogenous leukemia kDa Kilodalton KLN-205 Squamous cell carcinoma, mouse, lung L1210 Mouse lymphocytic leukemia LC/MS-MS Liquid chromatography - Mass spectrometry - Mass Spectrometry LD10 Lethal dose, 10% population xx LD50 Lethal dose, 50% population LOD Limit of detection LOQ Limit of quantitation LSC Liquid scintillation counter LUV Large unilammelar vesicles MCF7/ADR Human breast carcinoma, selected for adriamycin resistance MCF7AVT Human breast carcinoma, sensitive (wild-type) MDA435LCC6/mdr Human breast carcinoma (MDR, transfected with mdrl) MD A43 5LCC6AVT Human breast carcinoma (sensitive, wild-type) MDR Multidrug resistance mdrl Multidrug resistance gene (PGP) MLV Multilammelar vesicles MPS Mononuclear phagocyte system mRNA Messenger ribonucleic acid MRP Multidrug resistance-associated protein MRT Mean residence time MTD Maximum tolerated dose MTT 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide MW Molecular weight Na 2C0 3 Sodium carbonate NMR Nuclear magnetic resonance o2 Oxygen XXI p.o. per os, oral route P388/ADR Murine lymphocytic leukemia, selected for adriamycin resistance P388AVT Murine lymphocytic leukemia, sensitive (wild-type) p53 Tumor suppresser gene PBS Phosphate buffered saline PC Phosphatidylcholine PE Phosphoethanolamine PEG Polyethylene glycol PEG-DSPE Polyethylene glycol polymerized to DSPE PGP P-glycoprotein PS Phosphatidylserine PSC 833 SDZPSC 833 [p'-keto-Bm^MVal^-cyclosporin PSCM9 active metabolite of PSC 833 QELS Quasi-elastic light scattering Rag2 Recombinant activation gene 2 RBC Red blood cell RRF Residual resistance factor R-VRP, dexVRP R enantiomer (dextro) of Verapamil SCID Severe Combined Immunodeficient Mice SCLC Small cell lung cancer SD Standard deviation SDS Sodium dodecyl sulfate xxii SEM Standard error of the mean SUV Small unilammelar vesicles Tin Half life Tc Phase transition temperature Vd Volume of distribution VEGF Vascular endothelial growth factor VPF Vascular permeability factor VRP Verapamil Vss Volume of distribution at steady state (for i.v. bolus) WT Wild type xxiii ACKNOWLEDGEMENTS First and foremost, I am thankful to my advisor, Dr. Lawrence Mayer for his support, intellectual input, and for providing an excellent amiable working environment. My sincere thanks also to Dr. Marcel Bally who showed me that science can be fun. Both Lawrence and Marcel have played crucial roles in my maturing process over the years in the Department of Advanced Therapeutics, BC Cancer Agency. They should be complemented for providing state-of-the-art research opportunities as well as for creating a free thinking and amiable work atmosphere. Thanks also for all those pizzas and several parties! I must especially thank two individuals in the Faculty of Pharmaceutical Sciences, Dr. Wayne Riggs and Dean Frank Abbott, for their support over my years at UBC. Several people have helped me in my thesis work and I know that this work wouldn't have taken shape otherwise. My sincere thanks to: Howie and Troy for teaching me liposomology and for writing chapters for the website; Dana, Natashia, and Rebecca for help with mice (and rats) experiments; Daria for tissue culture; Gary for flow cytometry; Norma and Jean for helping and putting up to my (mis)adventures with the HPLC; Maryse for the confocal imaging and Emma for teaching me in vitro drug metabolism techniques. Thanks to Howie, Frances, Ellen, Gigi, Jennifer, Pierrot, Kelly, Hafiza, Spencer, Philip, Zaihui, Daniel and all others for keeping me buoyant and making my experiences in the lab a memorable one. Special thanks to my wife, Bhuvana for being there and constantly inspiring me through the final phase of the thesis. She has been a great source of support, encouragement and help. Last, but by no means the least, I acknowledge all sources of funding for the research project as well as for the BC Medical Services Foundation fellowship and travel awards. The research was funded by the National Cancer Institute of Canada. I must acknowledge additional sources of travel support from UBC (2), Roche, Procter and Gamble, Green College, NCIC, BCCF, and NATO ASI. xxiv DEDICATION To my mom, who will be the happiest person today, and my dad ..mom you have been a great source of support and encouragement and I couldn't have done this without your constant inspiration and motivation this one is for you! Chapter 1 INTRODUCTION 1.1. Hypothesis and Goal Many attempts to circumvent P-glycoprotein (PGP)-based multidrug resistance (MDR) in cancer chemotherapy have utilized PGP blocking agents (also referred to as MDR modulators), which are co-administered with the anticancer drug. This approach is based on the premise that inhibiting PGP function will result in increased accumulation of many anticancer drugs in the tumor cells and restore full antitumor activity. However, co-administration of MDR modulators with anticancer drugs has often resulted in exacerbated toxicity of the anticancer drugs and relatively poor chemosensitization of MDR tumors. These problems appear to be related to MDR modulator blockade of PGP excretory functions in healthy tissues, such as liver and kidney, which markedly reduces anticancer drug clearance properties. Two consequences of this pharmacokinetic interaction are: 1. Increased toxicity due to modulator-induced changes in biodistribution properties of the anticancer drug. 2. Problems interpreting preclinical and clinical data with respect to: a) Are therapeutic improvements due to altered pharmacokinetics or PGP modulation within the tumor cells? And, b) Does decreasing the anticancer drug dose to that which is equitoxic in the absence of the MDR modulator potentially compromise tumor therapy due to decreased anticancer drug levels in the tumor tissue? 1 Although many of the difficulties associated with co-administration of MDR modulators and anticancer drugs are manifested by toxicity effects, it is ultimately the ability to obtain effective antitumor activity against resistant tumors that will determine the utility of chemosensitization approaches. Liposomes appear to be well suited to solve many of the problems noted above that are associated with conventional anticancer drugs and MDR modulators. In view of these considerations, the hypothesis of this thesis is that inadequate tumor delivery of anticancer agents and selectivity of PGP modulation are primarily responsible for the attenuated therapy of extravascular MDR solid tumors overexpressing PGP. Liposomal carriers have been utilized to provide tumor selective delivery of anticancer agents as well as to circumvent many toxicities associated with these agents by altering the pharmacodistribution properties of encapsulated drugs (Gabizon and Papahadjopolous, 1988; Gabizon, 1992; Mayer et al., 1989; Mayer et al., 1990a,b; Mayer et al., 1995a; Forssen et al., 1996). Given the pharmacokinetic changes induced by the MDR modulator on free (non-encapsulated; conventional) doxorubicin (DOX), this thesis postulates that liposomes may limit these effects by virtue of their ability to reduce the exposure of encapsulated DOX to the kidneys and alter clearance of DOX in the liver (Van Hossel et al., 1984; Mayer et al., 1989). These tissues appear to be key factors involved in modulator-induced DOX pharmacokinetic changes (Colombo et al., 1996). In conjunction with these toxicity buffering effects, the effect of PGP blockade on the cellular uptake of DOX in the tumor may be able to be selectively increased using liposomal carriers. This is based on the ability of small liposomes to passively extravasate in leaky tumors (Gabizon, 1992; Gabizon and Papahadjopolous, 1988; Mayer et al., 1990a,b; Webb et al., 1995b) as well as their 2 inability to accumulate in healthy susceptible tissues. We predict that two factors will contribute to the increased tumor selectivity of PGP blockade effects when liposomal anticancer drugs are employed: 1) due to the high DOX localization in the tumor via liposomal delivery, PGP blockade caused by the co-administered drug efflux blockade component leads to increased cellular DOX uptake in the malignant cells, and 2) due to the reduced DOX accumulation in other PGP expressing healthy tissues following liposomal delivery, DOX is not significantly taken up in these susceptible tissues even under conditions of PGP blockade. The following sections of this thesis will review various areas of the biology of MDR as well as the physical, chemical and biological properties of liposomes that are relevant in this thesis. This will include an overview of the mechanisms of drug resistance, clinical relevance of MDR, strategies and shortcomings of overcoming MDR using MDR modulators with conventional anticancer agents, as well as a review on the applications of liposome technology that appear to be well suited for MDR reversal. 1.2. BackgroundTnformation: The Biology of Multidrug Resistance There are over 1,200,000 estimated new patients with cancer every year in North America (Landis et al., 1998). Less than a half of these are cancers which can be cured through surgery or radiation therapy. In the remainder, some cancers are also curable by chemotherapy; however, a large number of cancers are inherently untreatable or initially respond to treatment only to relapse with refractory disease (Gottesman, 1993). This is often due to the fact that these cancers are resistant to a wide variety of anticancer drugs or their combinations. The term, multi-drug resistance (MDR) has been utilized to describe this phenomenon and is characterized by the ability of such tumors to exhibit 3 simultaneous resistance to numerous structurally and functionally unrelated drugs from several drug classes. In some tumor types (e.g., renal cell, colorectal and pancreatic cancers), MDR is intrinsically presented prior to any treatment. In others, tumors which initially respond well to chemotherapy acquire MDR properties, rendering them refractory to most, if not all, chemotherapy combinations. Although the focus of this thesis is PGP-mediated MDR, a general overview of the underlying mechanisms and types of drug resistance is presented in the following sections in order to place the studies of this thesis in the context of the overall MDR phenomenon. A number of mechanisms have been described to explain the phenomenon of MDR in mammalian cells. They have been broadly classified into cellular and non-cellular mechanisms (Fan et al., 1994), as described below. 1.2.1. Non-cellular resistance mechanisms Non-cellular drug resistance can arise as a consequence of in vivo tumor growth. These phenomena are typically associated with solid tumors which exhibit unique physiological properties compared to circulating tumors such as hematological malignancies. Solid tumors are composed of a vasculature that is characterized by a higher geometric resistance (Sevick and Jain, 1989) and is heterogeneous, where the tumor blood vessels are dilated, tortuous, and saccular (Jain, 1987). In particular, the branching patterns in tumor tissue are much different from normal vessels, with vascular shunts and sprouts being present (Jain, 1988). The extracellular environment associated with solid tumors is characterized by increased interstitial fluid pressure compared to normal tissues. This is due to two contributing factors, namely, higher vascular permeability and absence of a functional lymphatic system (Jain, 1987). Consequently, poor tumor vascularization 4 can result in reduced drug access to regions within solid tumors and thus protect tumor cells from cytotoxicity. The physiological properties of solid tumors also result in tumor regions that are deficient in nutrients and oxygen. Such properties can induce additional resistance mechanisms that arise from extracellular influences. An example of this type of resistance is the increased presence of non-cycling tumor cells in poorly vascularized sections of solid tumors. These cells are often viable, but non-dividing, and consequently are resistant to drugs dependent on cell proliferation. The acidic environment in tumors, due to lactic acid generation by hypoxic tumor cells, has also been suggested to confer a resistance mechanism for weak bases, where cellular uptake is dependent on the pH gradient across membranes (Demant et al., 1990). 1.2.2. Cellular based resistance mechanisms Cellular mechanisms are categorized in terms of alterations in the biochemistry of malignant cells. Such mechanisms can be further classified into two major categories: 1) non-classical MDR phenotypes and 2) transport-based classical MDR phenotypes (see Section 1.2.2.2. for description). 1.2.2.1. Non-classical MDR Phenotypes The term non-classical MDR is used to describe non-transport based mechanisms that affect multiple drug classes. This type of resistance can be caused by altered activity of specific enzyme systems (such as glutathione S-transferase, GST and topoisomerase), which can decrease the cytotoxic activity of drugs in a manner independent of intracellular drug concentrations, which remain unaltered. In addition, changes in the balance of proteins that control apoptosis can also reduce chemosensitivity since most anticancer drugs are believed to exert their cytotoxic effects via apoptotic processes (Hickman, 5 1998). This section outlines some of these MDR mechanisms and their role in the overall MDR phenomenon. 1.2.2.1.1. Glutathione S-Transferease (GST) GST is an enzyme system involved in drug and xenobiotic detoxification. Specifically, biotransformation processes catalyzed by GST conjugate organic molecules with glutathione (GSH), resulting in excretable polar molecules. There are two intracellular pools of GST, one residing in the cytosol and the other in the microsomal compartment. Cytosolic GSTs are composed of 23-29 kDa subunits which may be homo or hetero-dimers (Mannervik and Jensson, 1982). In contrast, microsomal GSTs are trimeric and composed of identical 17 kDa subunits (Morgenstern et al., 1985). Among its various activities, GST plays an important role in protecting cells from reactive epoxides (Kuzmich and Tew, 1991). This is believed to occur via the catalytic addition of GSH to the epoxide moiety, such as that observed for the metabolism of aflatoxin B l to an 8,9-epoxide which is detoxified by GST (Ramsdell and Eaton, 1990). The GST enzyme also protects the cell from damage due to free radicals (Ketterer et al., 1990). The GSTs are extensively involved in the metabolic biotransformation of many anticancer drugs. Among these are nitrogen mustards, such as BCNU, and cyclophosphamides. Several resistant cell lines have been shown to overexpress GST (Lewis et al., 1989; Hao et al., 1994). The GST-pi isoenzyme has been shown to be overexpressed in MCF7/ADR cells (Batist et al., 1986). These MCF7/ADR cells, which also express elevated levels of PGP, exhibit increased peroxidase activity due to the enhanced levels of GST-pi. Similar increases in GST-pi levels have also been reported in other MDR cell lines (Cole et al., 1989; Chao et al., 1991; Raghu et al., 1993). However, 6 evidence suggests that the GST-pi activity is not a resistance mechanism for DOX. Specifically, when the gene expressing GST-pi was transfected into MCF7 cells, it resulted in over 15-fold increase in GST-pi activity compared to WT cells (Moscow et al., 1989), however, the transfected line was not DOX resistant suggesting that GST-pi did not contribute to DOX resistance. This is in contrast to observations with other GST isoforms where an 8-fold increase in resistance to drugs such as chlorambucil was observed when GST-alpha was transfected into yeast cells (Black et al., 1990), while for nitrogen mustards, a 2-fold increase in GST activity resulted in drug resistance (Schisselbauer et al, 1990). Agents such as ethacrynic acid and prostaglandin analogs block GST activity and can be used to increase the sensitivity of chlorambucil (Tew et al., 1988) or mephalan (Clapper et al., 1990). In addition to GST, the cellular regulation of the thiol tripeptide, glutathione (GSH) also appears to play a key role in detoxification and cellular repair following the damaging effects of DOX and alkylating agents. Increases in GSH levels have been observed in many alkylating agent-resistant cell lines (Calcutt and Connors, 1963; Ball et al., 1966; Meister, 1991; Meister, 1994). This suggests that reductions in intracellular GSH will result in chemosensitization of drug resistant cells. In support of this, use of such agents as buthionine sulfoximine (BSO) has led to the modulation of GSH-mediated drug resistance by reducing GSH levels (Batist et al., 1986). 1.2.2.1.2. Topoisomerase activity Two types of topoisomerase have been shown to be present in all eukaryotic cells (Wang, 1985; Osheroff, 1989). Type I topoisomerase, a monomeric 100 kDa protein (Liu and Miller, 1981), serves to alter DNA topology via single strand break (Wang, 1985), 7 while topoisomerase type II alters DNA topology by causing transient double strand breaks. Both enzyme classes are intrinsically involved in the processes of DNA replication (Wang, 1987). There are two subclasses of type II - one, a 170 kDa alpha and two, a 180 kDa beta (Drake et al., 1987; Chung et al., 1989). Consequently, these enzymes constitute therapeutic targets in rapidly dividing tumor cells for anticancer drugs. For example, DOX and etoposide specifically target topoisomerase II (Tewey et al., 1984a,b; Ross et al., 1984), while campothecin analogs target topoisomerase I (Kunimoto et al., 1987; Johnson et al., 1989). Anthracyclines stabilize the transient covalent complex between DNA and topoisomerase II alpha. Rejoining of DNA strands is blocked, resulting in the formation of stable enzyme-DNA complex with double-strand DNA breaks (Liu, 1989), resulting in tumor cell death. These DNA strand breaks appear to trigger apoptosis (Roy et al., 1992; Onishi et al., 1993), which leads to tumor cell death. Cells become resistant to topoisomerase II inhibitors such as DOX and etoposide (Drake et al., 1987) due either to the under-expression of topoisomerase II or topoisomerase II gene mutations. Resistance may occur alone or concurrent to PGP overexpression (Kunikane et al., 1990; Kamath et al., 1992). Reduced activity of topoisomerase II activity (Matsuo et al., 1990) as well as reductions in topoisomerase II mRNA levels have been shown to explain cellular resistance to topoisomerase II inhibitors. Topoisomerase II gene mutations can also occur, where reduced enzyme synthesis occurs following topoisomerase II gene transcription (Deffie et al., 1989; Tan et al., 1989). The exact mechanism of down regulation is poorly understood, but believed to be due to hypermethylation of the topoisomerase II gene. 8 A compensatory overexpression of topoisomerase I has been observed in cells resistant to topoisomerase II inhibitors, where cells have reduced expression of topoisomerase II (Tan et al, 1989). One approach to circumvent topoisomerase II-mediated MDR is to target both enzyme classes at the same time. For instance, when topoisomerase I inhibitor CPT-11 was pretreated in nude mice bearing human xenografts, enhanced activity of DOX, a topoisomerase II inhibitor, was observed (Kim et al, 1992), presumably due to overexpression of topoisomerase II activity mediated by CPT-11 pretreatment. 1.2.2.1.3. Altered apoptosis regulation Anticancer drugs typically induce programmed cell death or apoptosis. This form of cell death is characterized by nuclear condensation within cells, leading to DNA fragmentation caused by endonucleolytic cleavage of genomic DNA. The decision, whether a cell continues through cell cycle or undergoes apoptosis, is dependent upon a complex interplay of a team of genes and proteins that exert a regulatory role in cellular events. It has been suggested that the tumor suppressor gene, p53, not only plays a key role in inducing cell cycle arrest in Gi and apoptosis following DNA damage caused by anticancer drugs, but also in the regulation of expression of downstream proteins, bcl-2 and bax (Lowe et al, 1993a; Myashita et al, 1994). The ability of cells to undergo apoptosis has been linked to the formation of hetero- and homo-dimers generated via bcl-2-bax interactions. For example, bax-bcl-2 heterodimers as well as bax homodimers promote apoptosis, whereas apoptosis is inhibited when bcl-2 forms homodimers (Reed, 1995). Resistance may, therefore, develop with loss of genes required for cell death such 9 as p53 or overexpression of genes that block cell death. B-cell lymphoma-2 (bcl-2) is a gene that plays a key role in the regulation of cell death pathways. In lymphomas, particularly follicular lymphoma, evidence suggests that chromosomal translocations cause the movement of the bcl-2 gene from chromosome 18 to chromosome 14, due to errors in normal DNA recombination mechanisms (Tsuijimoto et al., 1988; Weiss et al., 1987; Zelenetz et al., 1991). Bcl-2 serves to protect the cell from stimuli causing cell death, such as uv or gamma radiation, tumor necrosis factor, or drugs that induce free-radical production (Reed, 1995). Bcl-2 has also been reported to be overexpressed in patients with AML and CLL (Campos et al., 1993; Hanada et al., 1993), whereas in carcinomas and other undifferentiated cancers, Bcl-2 expression is altered (Bronner et al., 1995; Castle et al., 1993; Colombel et al, 1993; Reed et al., 1991). Bcl-2 can confer cellular resistance to the cytotoxic effects of a number of anticancer agents including DOX, taxol, etoposide, camptothecin, mitoxantrone and cisplatin (Reed, 1995). Unlike PGP mediated MDR, overexpression of bcl-2 does not prevent drug influx into tumor cells (Fisher et al., 1993; Walton et al., 1993). When bcl-2 is overexpressed and contributes as a resistance mechanism, it has been shown that the anticancer drugs promote cell cycle arrest; however, their effects are cytostatic rather than cytotoxic (Fisher et al., 1993). Down regulation of bcl-2 can occur with loss of p53, the tumor suppressor gene. Stimuli such as radiation that induce DNA strand breaks are associated with increases in p53 protein as well as increases in p53 transcriptional activity (Zhan et al., 1993). Further, in vitro results have demonstrated that cell lines lacking p53 exhibit increased resistance to induction of apoptosis by anticancer drugs or radiation (Lowe et al., 1993b). 10 1.2.2.2. Transport-based classical MDR mechanisms The ATP-Binding Cassette (ABC) family of membrane transport ATPases are of considerable clinical importance. The family is phylogenetically ancient and its normal functions in eukaryotic cells remain to be fully characterized. The first of these proteins were found in bacteria and yeast (van Veen and Konings, 1998). These transport ATPases are generally composed of four structural domains, two that span the membrane (each containing several transmembrane segments) and two that remain in the cytoplasm. The last two units, called nucleotide binding domains, play a role in cleaving ATP (hydrolysis) to derive energy necessary for transporting cell nutrients, such as sugars, amino acids, ions and small peptides across membranes. As an example, one of the proteins responsible for resistance to the antimalarial agent, chloroquine, is a member of the ABC family, in which resistant Plasmodium falciparum (a malaria causing parasite) exhibit amplified levels of an ABC transporter that pumps chloroquine out of the cell (Wilson et al., 1989; Foote et al., 1989). The cystic fibrosis transmembrane regulator protein (CFTR) is also a member of the ABC family. Cystic fibrosis is a disease caused by a mutation in the gene encoding an ABC transporter that plays a role as a CI" channel in epithelial cell plasma membranes. It has been suggested, therefore, that these ABC transporters may play a dual role as ion channels as well as carrier pumps for influx/efflux of nutrients and hydrophobic drugs. Among the ABC transporters involved in MDR are P-glycoprotein (PGP) and multidrug resistance associated protein (MRP), which can be overexpressed in malignant cells and serve to pump anticancer drugs out of the cell, resulting in lack of intracellular levels of the drug necessary for effective therapy. These two drug transporter pumps 11 constitute well characterized mechanisms of MDR and are discussed in detail in the following sections. 1.2.2.2.1. Multidrug resistance associated protein (MRP) MRP, a member of the ABC family, has been described as a GS-X pump capable of transporting organic anion drug conjugates as well as intact anticancer drugs (Cole et al., 1992; Grant et al., 1994; Barrand et al., 1994; Krishnamachary et al., 1993; Borst et al., 1997). MRP, which was isolated as a transmembrane glycoprotein in non-PGP expressing small cell lung cancer DOX resistant cell lines, is an asymmetrical molecule with eight subunits and four membrane-spanning domains (Almquist et al., 1995). The clinical cancers exhibiting MRP expression include hematological (Burger et al., 1994; Broxterman et al., 1994), lung (Rubio et al., 1994; Savaraj et al., 1994), acute lymphoblastic leukemia relapses and chronic myeloid leukemia (Beck et al., 1994). The classes of anticancer drugs that are substrates of MRP include anthracyclines such as DOX, vinca alkaloids, and etoposide. By pumping these agents out of the tumor cells, MRP causes reduced intracellular accumulation of drugs, leading to resistance. Although MRP is remarkably similar to PGP in terms of substrate specificity and action, only a 15% amino acid homology exists between the two. Several isoforms of MRP have recently been identified, namely MRP1, MRP2 (or cannalicular Multispecific Organic Anion Transporter or cMOAT), and MRP3-6. MRP 1 and 2 have been identified as organic anion transporters (Borst et al., 1997). Although associated with drug resistance properties, MRP1 also plays a normal physiological role in the ATP-dependent unidirectional membrane transport of glutathione conjugates, such as leukotriene C4 and S-(2,4-dinitrophenyl)glutathione (Jedlitschky et al., 1994; Leier et al., 1994b). Inhibition of 12 this transport has been demonstrated with leukotriene receptor antagonists (Jedlitschky et al., 1994; Leier et al., 1994a; Leier et al., 1994b). Effective modulation of MRP-mediated MDR has been observed with the leukotriene LTD 4 receptor antagonist MK 571 (Gekeler et al., 1995b). It has also been shown that MRP1 transports drugs conjugated to glucuronic acid and glutathione in addition to oxyanion complexes, while MRP2 transports agents such as vinblastine (Kool et al., 1998). This would suggest that MRP-mediated MDR requires drug conjugation or modification prior to efflux across the plasma membrane. MRP is also expressed in normal human tissues, such as muscle, lung, spleen, bladder, adrenal gland and gall bladder (Zaman et al., 1993). Reduced glutathione (GSH) has been suggested as an important component of MRP-mediated MDR and drug transport (Olsen et al., 1998). An isoform of MRP was first shown to be expressed in the liver which functions in the excretion of glutathione and glucuronate conjugates across the cannalicular membrane into bile (Mayer et al., 1995b). Excretion of glutathione, glucuronate and sulfate conjugates across the hepatocyte canalicular membrane was functionally characterized (Akerboom et al., 1991; Fernandez-Checa et al., 1992; Ishikawa et al, 1990; Kitamura et al, 1990; Keppler et al, 1996) as being due to an ATP-dependent mechanism/transporter, which has been described under various names: as a leukotriene export pump (Keppler, 1992), a non-bile acid organic anion transporter (Arias et al, 1993), MOAT (Oude Elferink and Jansen, 1994; Oude Elferink et al, 1995), and as a glutathione-S-conjugate export pump (Ishikawa, 1992). In addition to being expressed in the cannalicular membrane, MRP2/cMOAT is also expressed in the human kidney proximal tubule epithelia on the apical side. Specifically, it is localized to the brush 13 border membranes of proximal tubule segments SI to S3 (Schaub et al., 1998). MRP2, therefore, is implicated to play a role in the renal excretion of endogenous substances and xenobiotics, in normal conditions. 1.2.2.2.2. P-glycoprotein (PGP) PGP (Figure 1.1 A), a member of a superfamily of ATP-dependent membrane transport proteins, is a plasma membrane protein that was first characterized in multidrug resistant Chinese hamster ovary cells by Ling and co-workers (Ling and Thompson, 1974; Juliano and Ling, 1976; Riordan and Ling, 1979; Kartner et al., 1983). It has been shown to pump substrates out of tumor cells through an ATP-dependent mechanism in a unidirectional fashion. In tumor cells expressing PGP, this results in reduced intracellular drug concentrations which decreases the cytotoxicity of a broad spectrum of antitumor drugs including anthracyclines (e.g., DOX), vinca alkaloids (e.g., vincristine), podophyllotoxins (e.g., etoposide), and taxanes (e.g., taxol). PGP is present in human (two human genes), mouse (three rodent genes), and hamster cells (Roninson et al., 1986; Van der Bliek et al., 1987; Croop et al., 1989; de Bruijn et al., 1986; Gros et al., 1988; Ng et al., 1989). Gene sequence analysis for mammalian PGP has revealed the presence of two similar halves, each containing 6 putative transmembrane segments, and an ATP-binding consensus motif (Figure 1.1). The human protein is comprised of 1280 amino acids with 12 transmembrane domains and 43% sequence homology between the two halves. Three glycosylation sites on the first extracytoplasmic domain are present (Chen et al., 1986; Gros et al., 1986a). 14 Figure 1.1 A: Structure of the PGP molecule. In the structure of PGP, the circled areas indicate ATP-binding sites, curly lines represent N-linked carbohydrates and filled circles represent amino acids that differ between mouse and human mdrl (Adapted from Kane et al., 1990). (OUT) • Drug H Chemosensitizer I Chemosensitizer binding site ® Drug binding site Figure 1. LB: Function of the PGP pump. The model illustrates a protein which uses ATP energy to actively efflux drug substrate across the plasma membrane. The MDR modulator (chemosensitizer) may act as a competitive inhibitor by occupying drug binding sites or as a non-competitive inhibitor at chemosensitizer binding sites (Modified from Ford, 1995). 15 There are three known isoforms of PGP, namely, Class I, II and III. Rodent cells have all three PGP genes, encoding classes I, II, and III PGP, whereas human cells have two, encoding class I and III PGP (Lee et al., 1993). Classes I and II PGP genes confer MDR when transfected into sensitive wild type (WT) cells, whereas the class III PGP gene is not shown to be associated with drug resistance. All three classes of PGP are inherently expressed in several normal tissues. Specifically, in mammalian tissues, class I PGP is present in intestinal lining epithelium, endothelial cells, bone marrow progenitor cells, peripheral blood lymphocytes and natural killer cells, whereas class II is present in the adrenal cortex. The class III PGP is localized in hepatocytes, cardiac and striated muscle (Thiebaut et al., 1987; Croop et al., 1989; Chaudhary et al., 1992). Human class I PGP is closely related to rodent class I and II PGPs, whereas human class III PGP corresponds to rodent class III PGP. In the liver, class III PGP is localized to the canalicular membrane of the hepatocytes. Mice lacking class III PGP expression are unable to secrete phospholipid into bile and this would suggest that class III PGP may act as a phospholipid transporter in hepatocyte membranes. These localization studies suggest that PGP has functional roles in normal tissues. PGP activity in normal tissues suggests an important role in transepithelial transport. PGP is expressed in such tissues as transporting epithelia of the liver, kidney, colon, small intestine, pancreas, placenta, uterus and in specialized capillary endothelial cells in the brain and testis (biliary face of hepatocytes, brush border of kidney proximal tubule cells, see Fojo et al., 1987; Thiebaut et al., 1987; Sugawara et al., 1988a; Thiebaut et al., 1989; Cordon-Cardo et al., 1989; Arceci et al., 1988). That PGP is expressed in the brain suggests a role in the blood-brain barrier, thus preventing the permeation and persistence 16 of hydrophobic agents in the central nervous system. These observations indicate a role of PGP in the transepithelial secretion of substrates into bile, urine, or gastrointestinal tract lumen. PGP may also confer a protective role to mediate xenobiotic efflux in tissues such as the brain, testis, and placenta. Whereas PGP fulfills critical functions in transport processes involved in normal physiology, overexpression of this protein in tumor cells results in reduced intracellular accumulation of anticancer agents due to increased drug efflux. Direct evidence of PGP-mediated drug transport came from initial studies employing partially purified membrane vesicles (Horio et al., 1988). It was shown that vesicles prepared from resistant cells were more capable of binding and transporting radiolabeled vinblastine than those prepared from sensitive cells (Horio et al., 1988). Vesicle transport experiments have also been performed in biliary canalicular vesicles expressing PGP, where ATP-dependent transport of daunomycin was shown in inside-out membrane vesicles obtained from apical membranes of hepatocytes (Kamimoto et al., 1989). Transepithelial transport of vinca alkaloids, and anthracyclines mediated by PGP was also demonstrated in epithelia prepared from Madin-Darby canine kidney cells (Horio et al., 1989), where a retrovirus carrying a human mdrl cDNA was transfected resulting in epithelia with PGP polarized on the apical surface. Further, transport of radiolabeled colchicine was demonstrated by partially purified and reconstituted PGP (Sharom et al., 1993). Several mechanisms have been put forward to explain this transport function of PGP. The model of Higgins and Gottesman, 1992, postulates that PGP encounters xenobiotics in the inner leaflet of the plasma membrane and flips the agents to the outer leaflet, where they diffuse into the extracellular region (Figure 1.1.B). PGP has also been 17 postulated to increase intracellular pH (Roepe, 1992), depolarizing plasma membrane electrical potential of the cell by acting as a proton pump, or a chloride channel, thus reducing intracellular accumulation of weak bases or reducing pH-dependent binding of agents to their intracellular targets (Roepe et al., 1993). In the hydrophobic vacuum cleaner model proposed by Gottesman and Pastan, 1993, PGP interacts directly with substrates in the plasma membrane and pumps them out of the cell. Evidence of competitive binding of drugs on the PGP molecule has been demonstrated with radiolabeled photoaffinity probes, such as N-(p-azido-[3-125I]-salicyl)-N'-p-amino vindesine and [3H]azidopine. This technique is based on energy transfer from a chromophoric substrate for PGP to a photoactive radiolabeled probe. Utilizing the photoactive analog of vinblastine, N-(p-azido-[3-125I]-salicyl)-N'-B-amino vindesine, as a probe, the presence of vinblastine-binding sites on the PGP molecule was shown in membrane vesicles prepared from MDR Chinese hamster lung cells and human carcinoma cells (Safa et al., 1986; Cornwell et al., 1986). This labeling was inhibited by vincristine and daunorubicin, indicating that the binding was competitive (Cornwell et al., 1986). Since colchicine and actinomycin D did not compete for vinblastine-binding sites, the presence of multiple drug binding domains was suggested (Akiyama et al., 1988; Cornwell et al., 1986). These observations were corroborated by the fact that binding of the photoactive calcium channel blocker, [3H]azidopine to PGP was blocked by colchicine and actinomycin D among others (Safa et al., 1987; Yang et al., 1988). Based on recent methods of purifying and reconstituting class I hamster PGP into liposomes with greater retention of ATPase activity (Shapiro and Ling, 1994), transport of the PGP substrate, 18 Hoechst 33342 was demonstrated (Shapiro and Ling, 1995), lending further support of the hypothesis that PGP is itself capable of direct substrate transport. 1.3. Clinical relevance of MDR A number of cancers are treatable either by surgery, radiation therapy or chemotherapy. However, a considerable number of cancers either are intrinsically resistant or exhibit treatment induced acquired resistance, which complicates efforts to successfully cause long-term regression or cure. As described in appropriate sections above, a number of mechanisms have been shown to exist under the generalized MDR group. Among these mechanisms, the role of PGP in MDR is perhaps the best understood and constitutes a well characterized system for studying MDR. PGP occurrence in clinical tumors has been extensively characterized and PGP overexpression has been shown to occur both during diagnosis as well as during relapse. Examples of human tumors where PGP was detected at the time of diagnosis include: acute lymphoblastic leukemia (Goasguen et al., 1993); acute non-lymphoblastic leukemia (Campos et al., 1992); lymphoma (Niehans et al, 1992); adult sarcoma (Gerlach et al, 1987); colon carcinoma (Goldstein et al, 1989); childhood sarcoma (Chan et al, 1990), and neuroblastoma (Chan et al, 1991). In many human tumors PGP detection correlates well with poor response to anticancer drug therapy and subsequent shorter survival (Bradley and Ling, 1994). The most convincing correlation has been in childhood solid tumors such as soft tissue sarcoma and neuroblastoma (Chan et al, 1990; Chan et al, 1991). A large study of a series of patients with acute non-lymphoblastic leukemia revealed a similar correlation between PGP detection in leukemic cells at diagnosis and lower rate of complete regression after intensive chemotherapy, in addition to shorter patient survival (Campos et al, 1992). 19 However, despite the compelling number of correlations, many other studies have failed to establish this link between PGP detection and shorter survival times (Haak et al., 1993; Vergier et al., 1993; Finnegan et al., 1995). In addition, other mechanisms, such as altered enzyme activity or apoptotic events, may be present concurrent to PGP (Batist et al., 1986; Giaccone et al., 1992; Lehnert, 1996), thus complicating efforts to develop treatment strategies utilizing PGP as an unilateral target. This thesis, however, supports the concept that even in the presence of non-PGP-mediated MDR mechanisms, the expression of PGP still constitutes a therapeutic barrier in the tumor, where the anticancer drugs will not be able to gain access to the cells and exert their anti-proliferative activity. Consequently, pharmacological improvements aimed at MDR reversal will depend on the ability of the anticancer drugs to be taken up in the malignant cells and consequently require the presence of a drug efflux blockade component if PGP or related proteins are overexpressed even under conditions of resistance mechanisms beyond PGP. 1.4. Therapeutic Strategies to overcome PGP-mediated MDR MDR clearly presents a significant obstacle to the chemotherapy of many cancers. A considerable amount of work has focused on reversal of transport-based MDR mechanisms, particularly PGP, with an aim of increasing intracellular accumulation of anticancer agents. Although strategies have varied from developing newer analogs of anticancer agents that are not substrates of the ABC family of transporters to reducing expression of PGP mRNA using monoclonal anti-PGP antibodies (Tsuruo et al., 1989), a considerable amount of attention has been directed at the process of chemosensitization utilizing agents that impair the function of PGP. This method of overcoming MDR involves the co-administration of a PGP inhibitor (MDR modulator) with an anticancer 20 drug in order to cause enhanced intracellular anticancer drug accumulation. Numerous compounds have been shown to inhibit the drug efflux function of PGP and therefore, reverse (modulate) cellular resistance. As described in Section 1.2.2.2.2., PGP (Figure 1.1B) is characterized by the presence of ATP binding sites, anticancer drug binding sites, and phosphorylation sites. These sites can therefore constitute targets for impairing the function of the drug efflux pump. Consequently, MDR modulators have been suggested to exert their action in several ways, such as blocking drug efflux function competitively or non-competitively. Specifically, they may act as substrates for PGP and may serve as competitive ligands for recognition sites on the PGP molecule (Gottesman and Pastan, 1988; Yusa and Tsuruo, 1989) Other mechanisms include inhibition of PGP activity via interference with ATPase activity or phosphorylation (Ffamada and Tsuruo, 1988) as well as alterations in membrane fluidity and increased membrane permeability caused by positively charged hydrophobic MDR modulators (Drori et al., 1995). PGP directed MDR modulators belong to a number of chemical classes including calcium channel blockers, calmodulin inhibitors, coronary vasodilators, indole alkaloids, quinolines, hormones, cyclosporins, surfactants, and antibodies (Ford and Hait, 1990; Ford and Hait, 1993). In general, they have been classified as those belonging to the first, second or third generation, as described in the following sections. 21 1.4.1. First Generation Modulators Tsuruo and co-workers (1981; 1982a,b; 1983a,b,c) were the first to demonstrate the ability of the calcium channel blocker, verapamil (VRP), to reverse MDR. VRP enhanced intracellular accumulation of many anticancer drugs, including DOX in numerous cell lines (Inaba et al., 1979; Harker et al., 1986; Roepe, 1992; Feng et al., 1992). Subsequent studies revealed that this MDR reversing property is shared by many other calcium channel blockers. Clinically available calcium antagonists were demonstrated to reverse MDR in vitro. These included felodipine (Ffollt et al., 1992), isradipine (Hollt et al., 1992), nicardipine (Tsuruo et al., 1983b; 1983c; Ramu et al., 1984a), nifedipine (Tsuruo et al., 1983b; 1983c; Ramu et al., 1984b), bepridil (Schuurhuis et al., 1987), and diltiazem (Tsuruo et al., 1983b; Klohs et al., 1986). These agents modulated MDR at very high concentrations, ranging from 5-50 uM. At these high concentrations, there was enhanced cytotoxicity observed in normal cells (Lampidis et al., 1986; 1990). In addition, it was later demonstrated that even calmodulin antagonists such as trifluorperazine (Tsuruo et al., 1982a,b; Klohs et al., 1986; Ganapathi and Grabowski, 1993; Ganapathi et al., 1986; Akiyama et al., 1986; Ford et al., 1989; Ganapathi et al., 1984), chlorpromazine (Akiyama et al., 1986; Ford et al., 1989; Ganapathi et al., 1984) and prochlorperazine (Ford et al., 1989; Ganapathi et al., 1984) all reversed MDR significantly at concentrations ranging from 1-10 uM. Other more potent calmodulin antagonists included clopenthixol (Ford and Hait, 1990), trifluopromazine (Ford et al., 1989), and flupenthixol (Ford et al., 1989; Ford and Hait, 1990), all of which were effective at concentrations of approximately 3-5 uM. Indole alkaloids, the anti-malarial quinine and the anti-arrhythmic quinidine, have also been shown to reverse MDR in vitro in experimental cell lines in conjunction with 22 DOX (Solary et al., 1990; Solary et al., 1991; Eliason et al., 1990; Sehested et al., 1989). Cyclosporin A, a commonly used immunosuppressant for organ transplantation, remains one of the most effective first generation MDR modulators studied. In conjunction with several anticancer agents, it effectively reverses MDR in many cell lines, such as P388 (Kukl et al., 1992), CHO (Neumann et al., 1992), K562 and CEM (Chao et al., 1990), L1210 (Dorr and Liddil, 1991), and C26 (Spoelstra et al., 1991). A number of these first generation MDR modulators displaying excellent MDR reversal activities in vitro have been tested in murine tumor models. A common model is the evaluation of MDR efficacy following i.p. inoculation of sensitive and MDR tumor cells, such as P388 or L1210, to generate ascites. Subsequently, the drug and chemosensitizer are administered i.p (Skovsgaard et al., 1984). Several MDR modulators including VRP, nicardipine, quinacrine, and CsA demonstrated significant increases in mean and median life span in this i.p.-i.p. model (Slater et al., 1982; 1986b; Tsuruo et al., 1984; 1983c; 1981). 1.4.2. Second Generation Modulators A unique property shared by most first generation MDR modulators is that they are therapeutic agents and typically reverse MDR at concentrations much higher than those required for their individual therapeutic activity (see Section 1.5.). The search for non-toxic second generation modulators resulted in newer analogs of the first generation agents, which were more potent and considerably less toxic. Structural analogs of VRP, dexverapamil (less cardiotoxic R-enantiomer of VRP), emopamil, gallopamil, and Ro 11-2933 (a tiapamil analog) reversed MDR in vitro to a degree equivalent to VRP, but with marginal toxicity in many animal models (Pirker et al., 23 1990; Nawrath and Raschack, 1987; Pirker et al., 1989). In addition, some of these compounds demonstrated increased potency such as the tiapamil analog Rol 1-2933 which is effective at 1-2 uM concentrations compared to 5-10 uM required for VRP (Kessel and Wilberding, 1985; Alaoui-Jamali et al., 1993). The non-immunosuppressive analog of CsA, PSC 833, has demonstrated superior MDR reversal efficacy in many experimental cell lines in vitro (Boesch et al., 1991b; Jonsson et al., 1992; Kreis et al., 1993; Van der Graaf et al., 1993; Gaveriaux et al., 1989; 1991). PSC 833, has been shown to reverse MDR in conjunction with daunorubicin, DOX, vincristine, vinblastine, taxol, or mitoxantrone in MDR P388 leukemic cell lines at in vitro concentrations of 0.5 - 2 uM (Boesch et al., 1991b; Jonsson et al., 1992; Kreis etal., 1993; Van der Graaf et al., 1993; Duran et al., 1993; Krishna et al., 1997). Considerable improvements in antitumor activity and of anticancer agents in conjunction with PSC 833 have been demonstrated in in vivo ascites models as well as solid tumor MDR models (Boesch et al., 1991a; Colombo et al., 1996; Watanabe et al., 1996a; 1995; Keller et al., 1992b). In a study using the P388 ascites leukemia, murine C26 colon carcinoma, and HCT-15 human xenografts in nude mice, the chemosensitizer PSC 833 along with DOX provided varying degrees of MDR reversal. While providing significant increases in median life span in ascites leukemia and significant tumor regression in C26 carcinoma (which expresses moderate to low levels of PGP), co-administration of PSC 833 and DOX, at non-dose reduced conditions, afforded incomplete tumor regression in the human xenograft model (Watanabe et al., 1996a). 24 1.4.3. Third generation modulators Several MDR modulators have recently been developed using structure-activity relationships and combinatorial chemistry approaches targeted against specific MDR mechanisms. These agents exhibit effective reversing concentrations in the nanomolar range (20 - 100 nM), thus requiring low doses to achieve effective reversing concentrations in vivo. One such example is the cyclopropyldibenzosuberane L Y 335979, which is characterized by a 10-fold increased potency compared to CsA, latent modulating activity and a blockade mechanism specific for PGP (Dantzig et al., 1996). The acridonecarboxamide GF 120918 (Hyafil et al., 1993) exhibits similar characteristics to L Y 335979, with increased potency and a PGP-selective blockade mechanism. VX-170 and VX-853 are examples of bispecific chemosensitizers that block both PGP and MRP (Germann et al., 1996; Peck et al., 1996). Using a combinatorial chemistry approach, XR9051, a novel diketopiperazine derivative, was shown to be specific in modulating PGP-mediated MDR (Dale et al., 1998). Although these agents appear to be well tolerated in combination with anticancer drugs such as DOX, it is yet to be determined whether these compounds can achieve full chemosensitization in more stringent MDR solid tumor models or beyond that, in clinical cancers. 1.5. Complications with chemosensitizers First generation modulators such as VRP and CsA exhibit inherent pharmacological activity (cardiovascular for VRP, immunosuppression for CsA) and have been shown both preclinically and clinically to require high doses to achieve in vivo plasma concentrations sufficient to reverse MDR (Ozols et al., 1987; Fisher et al., 1994a). At these elevated doses, both compounds exhibited severe and sometimes life-threatening toxicities. These 25 dose-limiting toxic effects have precluded their application as effective MDR modulators for in vivo applications, particularly in the clinic. For VRP, this is exemplified by the fact that plasma concentrations required to obtain desirable cardiovascular effects average 0.4 -1.2 uM (Anderson et al., 1982). Clinical manifestations of toxicity (arterio-ventricular block and hypotension) have been observed at concentrations higher than these, particularly at plasma concentrations required for MDR modulation activity (2-6 uM). These complications of first generation modulators stressed the need to develop MDR modulators with lower inherent toxicities. This initiated the search to identify congeners of first generation MDR modulators with reduced inherent toxicity while retaining MDR modulating efficacy. These efforts resulted in the development of second generation MDR modulators which were either stereoisomers of their first generation racemic counterparts, such as dexVRP (R-enantiomer of VRP) and dexniguldipine (R-enantiomer of niguldipine) or structural analogs of first generation agents, such as Rol l -2933 (a tiapamil analog) and PSC 833 (a CsA analog). All of these compounds exhibited decreased inherent toxicity while retaining MDR reversal efficacy compared to their parent compounds. An added feature was that some of these agents exhibited increased potency, and consequently required low doses to achieve effective in vivo plasma concentrations required to modulate MDR. Although these agents alleviated many of the problems experienced with first generation MDR modulators, when these agents were co-administered with anticancer agents for modulating PGP-based MDR they often caused significant alterations in pharmacokinetics and biodistribution properties of the anticancer drugs (reviewed in Lum and Gosland, 1995). Early reports suggested that this was perhaps due to PGP blockade in 26 normal tissues, given the localization of PGP in such tissues as the liver and kidney which exhibit a secretory function (See section 1.2.2.2.2.). Examples of such pharmacokinetic interactions include an 8-fold increase in plasma AUC of daunorubicin observed when the anticancer drug was combined with the PGP inhibitor VRP in rats (Nooter et al., 1987), which resulted in increased toxicity and a need to reduce the anticancer drug dose. This was corroborated by observations that high dose VRP infusion resulted in 8-fold increases in tissue uptake of vincristine in liver, kidney and intestine,! necessitating an 8-fold reduction in drug dose (Horton et al., 1989). A recent report demonstrated that VX-853, a potent third generation modulator caused a 3-fold reduction in plasma clearance of co-administered paclitaxel, necessitating anticancer drug dose reduction (Wang and Chaturvedi, 1998). In the liver, expression of PGP is primarily on the canalicular membranes of the biliary tract and anticancer drug transport in this tissue has been shown to be inhibited by MDR modulators (Gatmaitan and Arias, 1993; Bohme et al., 1993; 1994). Blockade of biliary clearance of several anticancer drugs by MDR modulators has also been demonstrated. This includes reports of colchicine blockade by CsA (Speeg et al., 1992a), reduced vinblastine elimination caused by CsA (Samuels et al., 1993), as well as DOX biliary clearance inhibition by PSC 833 (Speeg and Maldonado, 1994) and GF120918 (Brouwer and Mellon-Kusibab, 1996; Booth et al., 1998). These observations suggest that MDR modulators block PGP expressed in the biliary canaliculi and impede drug transport. In the kidney, PGP is highly expressed on the brush border of proximal renal tubule and colchicine renal clearance was inhibited by CsA (Speeg et al., 1992b), suggesting that MDR modulators may also alter the renal elimination processes of 27 anticancer drugs by blocking PGP in the kidneys. These observations of impaired biliary and renal excretion mediated by first and second generation MDR modulators are consistent with clinical observations with a variety of MDR modulators and anticancer drugs where increases in anticancer drug plasma exposure (i.e., increased elimination half-lives and AUC) and hematological toxicities warranted dose reduction (Kerr et al., 1986; Fedeli et al., 1989; Philip et al., 1992; Lum et al., 1992; Mross et al., 1993a; Tolcher et al., 1994; Bartlett et al., 1994; Fisher et al., 1994a,b; Erlichman et al., 1994; Giaccone et al., 1994; Boote et al., 1996; Sarris et al., 1996). When anticancer drug doses are reduced to accommodate modulator-induced anticancer drug pharmacokinetic alterations and toxicity, it complicates interpreting whether these dose adjustments affect drug exposure to the tumor and antitumor activity compared to anticancer drug administration alone. Consequently, results from a number of disappointing Phase II clinical trials in patients with colerectal, lung, ovarian, breast, renal cell, myeloma, and leukemic cancers cannot be used to resolve the importance of PGP in these tumor types. Chemosensitivity results in these patients have ranged from a few partial (Sonneveld and Marie, 1994) and minor responses (Van Kalken et al., 1991), to no responses (Presant et al., 1986; Miller et al., 1994; Ozols et al., 1987; Hendrick et al., 1991; Mross et al., 1993a; Bissett et al., 1991; Rodenburg et al., 1991; Verweij et al., 1991; Murphy etal., 1994; Wishart et al., 1994; Dalmark et al., 1991). The reasons for this failure to achieve favorable responses in the clinic are unclear, however, several possibilities exist: 1) PGP does not mediate MDR in the clinical tumors tested; 2) the tumors may express other resistance mechanisms in addition to PGP and therefore PGP blockade alone is not useful in such tumors; and 3) the dose reductions necessitated by 28 MDR modulator induced inhibitions of anticancer drug clearance may cause reduced drug exposure to the tumor. Given the lack of selectivity for the tumor tissue PGP, it is still unclear whether such pharmacokinetic alterations induced by the modulator may have a negative impact on therapeutic outcome; however, such relationships are clearly possible and have been investigated in this thesis. The reversal of MDR using the chemosensitization approach presents a complex challenge in targeting PGP in tumor tissue. The lack of specificity for action on tumor PGP is an inherent disadvantage for most, if not all, MDR modulators. Given the fundamental disadvantage of MDR modulation with respect to the inability to differentiate between tumor and non-tumor PGP, alternative methods to improve the selectivity of MDR modulation at the tumor site may offer a significant advantage. Central to the objectives of this thesis is the hypothesis that, by altering anticancer drug delivery to cause increased delivery to the tumor site and reduced delivery to healthy susceptible tissues, the selectivity of MDR modulation at the tumor site can be increased using appropriately designed macromolecular carrier systems for anticancer drug delivery. Two factors contribute to the increased selectivity: 1) due to the high DOX localization in the tumor via liposomal delivery, PGP blockade caused by the co-administered drug efflux blockade component leads to increased cellular DOX uptake in the malignant cells, and 2) due to the reduced DOX accumulation in other PGP expressing healthy tissues following liposomal delivery, DOX is not significantly taken up in these susceptible tissues even under conditions of PGP blockade. 29 1.6. Drug delivery approaches to improve chemotherapy of MDR tumors Various polymer and lipid based drug delivery vehicles have been developed which selectively accumulate in sites of tumor growth (Mayer et al., 1989; Gabizon, 1992; Gabizon and Papahadjopolous, 1988; Douglas et al., 1987). Many properties of liposomal delivery systems appear well suited to address the problems associated with therapeutic strategies utilizing conventional anticancer drugs and MDR modulators. Liposomes may be applied to the treatment of MDR tumors in two basic approaches. These are: 1) the use of liposomes as carriers for anticancer drugs to provide increased dose intensity and tumor delivery, and 2) the inclusion of PGP modulators in the lipid bilayer or entrapped inside liposomes for delivery to MDR tumors. Although delivery of the MDR modulator to the site of tumor progression may perhaps be considered ideal, lack of supportive literature reports in this area coupled with the limited success, in our laboratory, of encapsulating such agents, have precluded pursuing this strategy. This inability to effectively administer the MDR modulator in these drug carrier systems stems from difficulties retaining the modulator within the carrier. This thesis, therefore, has focused on the delivery of the anticancer drug to MDR tumors in appropriately designed liposomes. Liposomes are lipid vesicles composed of phospholipid and cholesterol which have been employed as drug delivery vehicles for drugs and macromolecules. Over the past several years, our laboratory has investigated the application of liposomal anticancer agents in various preclinical and clinical tumor models demonstrating the enhanced selectivity of such drug carriers for the tumor tissue with associated reductions in toxicity and favorable pharmacokinetic behavior. Consequently, the ability of these liposomal systems to selectively localize in sites of tumor growth can be used to evaluate the role of drug 30 delivery in modulating MDR. In this context, the following sections will review the basic chemistry and physics of liposomal lipids, methods used to prepare liposomes, techniques to encapsulate drugs inside liposomes, biological fate of systemically administered liposomes, and applications in cancer chemotherapy. Liposomes, first described by Bangham and co-workers in 1965, are typically composed of phospholipids and up to 50% (molar basis) of cholesterol. They are biodegradable, non-toxic and generally non-immunogenic. When bilayer forming lipids (amphiphilic lipids) are dispersed in excess buffer (aqueous phase), they spontaneously adopt an onion skin arrangement of concentric lamellae structures called multilamellar liposomes (Figure 1.2A). As illustrated in Figure 1.2A the hydrophilic head group and hydrophobic acyl chains are oriented in such a manner that the lipid head group is toward the aqueous environment and the acyl chains are towards each other. In this thesis, three types of liposomes will be described, namely, multilamellar vesicles (MLV), large unilamellar vesicles (LUV), and small unilamellar vesicles (SUV) (See Figure 1.2B for an illustration. 31 A bilayer structure B Lipid Multilamellar Large Unilamellar Small Unilamellar Vesicles Vesicles Vesicles Figure 1.2: (A) An illustration of the structure of a liposome bilayer. Lipids upon hydration with aqueous buffer adopt a bilayer configuration. (B) Illustration of three types of liposomes, multilamellar vesicles, large unilamellar vesicles, and small unilamellar vesicles. Liposomes were initially developed as models of biological membrane to evaluate the structural and functional roles of lipids in biological membranes. It was not until later that their ability to serve as drug carriers was exploited (Sessa and Weissmann, 1968). Since then, these delivery systems have been utilized as effective carriers for the delivery of anticancer drugs (Rahman et al., 1990; Cowens et al., 1993; Mayer etal., 1989; Mayer et al., 1990a; Mayer et al., 1990b; Money-Kyrle et al., 1993), antimicrobial agents (Lopez-Berestein and Juliano, 1987; Vincent et al., 1992; Alving and Swartz, 1984), genes 32 (Feigner and Ringold, 1989; Rose et al., 1991; Hyde et al., 1993; Singhal and Huang, 1994), and antisense oligonucleotides (Bennett et al., 1992; Leserman et al., 1994). An understanding of liposome chemical and physical properties is central to applying these drug delivery systems for use in the treatment of MDR tumors. Therefore, subsequent sections will review the building blocks or components of liposomes, factors influencing selection of lipids, methods of preparation, drug entrapment procedures, and how these characteristics impact on their use in the delivery of anticancer drugs for cancer chemotherapy. 1.6.1. Liposome components 1.6.1.1. Phospholipids One of the primary component of liposomes is the phospholipid. Figure 1.3 illustrates the typical phospholipid structure and examples of some species commonly used for liposome drug delivery applications. As shown in this figure, these lipid structures are characterized by two hydrophobic acyl chains linked via ester bonds to a glycerol backbone and a phosphate containing hydrophilic headgroup (Chapman, 1975; Seelig, 1978). Increased unsaturation of acyl chains has been correlated with increased membrane permeability (Papahadjopolous et al., 1973) and reduction in phase transition temperature (Chapman, 1968). 33 o I oc o I O=P—O" o C H 2 C H C H 2 -I I o Headgroup Glycerol backbone Acyl chain-Neutral phospholipids Choline (Phosphatidylcholine) Ethanolamine (Phosphatidylethanolamine) Negative phospholipids Serine (Phosphatidylserine) _ Glycerol (Phosphatidylglycerol) Structures -CH 2CH 2N+(CH3)3 - C H 2 C H 2 N + H 3 - C H 2 C H N + H 3 COO" -CH 2CH(0H)CH 20H Saturated fatty acids Laurie CH 3 (CH 2 ) 1 0 COOH Myristic CH 3 (CH 2 ) 1 2 COOH Palmitic CH 3 (CH 2 ) 1 4 COOH CH 3 (CH 2 ) 1 6 COOH Stearic Unsaturated fatty acids Palmitoleic Oleic Linoleic Linolenic CH3(CH2)5CH=CH(CH2>7COOH CH 3(CH 2) 7CH=CH(CH 2)7COOH CH 3(CH 2) 5CH=CHCH 2CH=CH(CH 2) 7COOH CH 3CH 2CH=CHCH 2CH=CHCH 2CH=CH(CH 2) 7COOH Figure 1.3: Diagrammatic representation of phospholipid chemical structures. Illustrated are the headgroup, glycerol backbone, and the acyl chains. The figure also shows various headgroup and acyl chain fatty acid compositions for commonly occurring phospholipids. 34 When selecting a lipid for liposome formulation, a factor that must be considered is the gel-liquid crystalline phase transition temperature (Tc). Tc is dependent upon the length of the acyl chains as well as saturation. In general, the longer the acyl chains the higher the Tc. Unsaturation of acyl chains and presence of bulky side groups leads to a reduction in Tc. The headgroup of phospholipids can also influence Tc, as well as calcium and magnesium ions which can increase Tc of acidic phospholipids (Jacobson and Papahadjopolous, 1975). Table 1.1 lists the Tc of commonly used phospholipids. Below the Tc, the motion of the acyl chains is restricted (high order) and the chains are packed in gel phase. Above the Tc, the acyl chains are less ordered and the chains exist in a liquid crystalline phase. Increased unsaturation of acyl chains reduces the order which has been correlated to increased membrane permeability (Demel et al., 1972; Papahadjopolous et al., 1973) and a reduction in Tc (Chapman, 1968; Chapman, 1975). It has also been observed that membranes are more permeable to solvents and solutes at or above the Tc (Bittman and Blau, 1972). In this thesis, three liposomal lipid compositions were utilized, namely, 1) 1,2 distearoyl-sn-glycero-3-phosphocholirie (DSPC)/cholesterol (Choi) (55:45), 2) polyethylene glycol derivatized 1,2-distearoyl-sn-glycero-phosphoethanolamine (PEG-DSPEVDSPC/Chol (5:50:45), and 3) egg phosphatidylcholine (EPC)/Chol (55:45). The former two systems are saturated lipid compositions while the EPC system is a mixture of saturated and unsaturated fatty acids. These systems exhibit varying membrane permeability and drug release properties dictated by the individual liposomal lipid compositions. For example, EPC (Tc = -10 °C)/Chol DOX is a rapid release system characterized by high membrane permeability and drug leakage properties (Bally et al., 35 1990a), whereas both DSPC (Tc = 55 °C)/Chol and PEG-DSPE/DSPC/Chol DOX formulations are long circulating formulations which retain DOX for extended periods of time (Mayer et al., 1994). When phospholipids are hydrated, they can adopt a variety of structures, including bilayer, micellar, or hexagonal phase structures. This ability of lipids to adopt structures other than the bilayer has been described as lipid polymorphism (Cullis et al., 1986). Surfactants (detergents) possess head groups which occupy a large volume compared to their acyl tails, and these may adopt a cone shaped configuration with an increased tendency to form micelles. In PE, the polar group is small relative to acyl hydrocarbon chains. These lipids do not form bilayers or liposomes on their own. Instead, such lipids upon hydration enter, what has been called, the hexagonal phase, where lipid cylinders exist in an inverted rodlike micelle configuration (Cullis and de Kruijff, 1979). Several other phospholipids, including those with a single acyl chain, have also been shown to adopt a non-bilayer structure depending on such factors as temperature, pH and presence of ionic gradients (Cullis and de Kruijff, 1979). However, most lipids employed for liposomal drug delivery fall under the category of bilayer forming phospholipids. These bilayer forming lipids spontaneously adopt a bilayer configuration by preference (see Figure 1.2.B). Further, long saturated acyl chains also tend to form extensive van der Waals interactions with each other in the bilayer, limiting their motion. The most commonly utilized bilayer forming lipids with respect to drug delivery applications are phosphatidylcholines (PCs). PC is a zwitterion composed of a glycerol-phosphate ester with a choline headgroup and two acyl chains esterified to the sn-1 and sn-2 positions (Hamilton et al., 1991). 36 Table 1.1.: Temperature (Tc) of the gel to liquid-crystalline phase transition and charge of various phospholipids. Lipid Tc (°C) Charge Reference DilauroylPCa(12:12)b -1 0 Mabrey and Sturtevant, 1976 DimyristoylPCa(14:14)b 24 0 Ladbrooke and Chapman, 1969 DipalmitoylPCa(16:16)b 41 0 Ladbrooke and Chapman, 1969 DistearoylPCa(18:18)b 55 0 Ladbrooke and Chapman, 1969 DipalmitoylPEa(16:16)b 63 - Van Dijck et al., 1976 EggPCa' c -10 0 Ladbrooke and Chapman, 1969 a: PC, phosphatidylcholine; PE, phosphatidylethanolamine b: Number in parentheses represents the number of carbons per acyl chain. c: Fatty acid content of egg derived PC is as follows: palmitic acid (33.95%), oleic acid (31.06%), linoleic acid (17.72%), and stearic acid (10.52%). The membrane surface charge of a liposome is dependent on the presence of negatively charged phospholipids (refer to Figure 1.3.) (in the absence of exogenous synthetic cationic lipids such as those used in gene therapy). Phosphatidylcholines (PCs) and phosphatidylethanolamines (PEs) are zwitterionic at physiological pH due to the presence of positively charged amine group. However, in contrast to these PCs and PEs, phosphatidylserines (PS) carry a net negative charge. These properties which affect the biological fate of systemically administered liposomes will be discussed further in Section 37 1.6.1.2. Cholesterol Another commonly employed component of liposomes is cholesterol (Figure 1.4. A). Cholesterol is the major sterol found in membranes of most animal tissues and possesses amphipathic characteristics conferred by the polar 3-P-hydroxy group. Although it can exist in bilayer membranes at levels up to 50 mole %, cholesterol is unable to form bilayers in the absence of phospholipid. Cholesterol orients itself in the bilayer in such a manner that the hydophobic steroid nucleus aligns itself with the phospholipid acyl chains and the hydroxyl group is positioned along with the carbonyl ester bond of the phospholipid (Figure 1.4. A, B). The inclusion of cholesterol in the membrane influences the Tc of the phospholipid. Specifically, cholesterol increases the order of acyl chains of phospholipids that are in the liquid-crystalline state and decreases the order of phospholipids in the gel state (see Figure 1.4. C; Demel and de Kruijff, 1976; Hyslop et al., 1990). 38 Gel Liquid-crystalline Figure 1.4: (A) Chemical structure of the cholesterol molecule. (B) Diagrammatic representation of cholesterol orientation with PC in a membrane environment. (C.) Role of cholesterol when incorporated in the liposome bilayer. Note the association of cholesterol with the PC in the bilayer. 39 A reduction in the enthalpy of the gel to liquid-crystalline transition occurs when cholesterol is incorporated into membranes composed of saturated PCs (Ladbroke et al., 1968; Hubbel and McConnel, 1971). The presence of cholesterol, therefore, suppresses phase transition when incorporated at 30 mole % and higher (Chapman, 1975). In this thesis, cholesterol was used in all liposome compositions at 45 mole %. The interaction of cholesterol with the hydrocarbon chains decreases order as it interferes with the association of hydrocarbon chains between phospholipids (Ladbroke et al., 1968; Hubbel and McConnel, 1971). Furthermore, inclusion of cholesterol helps stabilize systemically administered liposomes as a result of reduced lipid exchange with lipoproteins (Kirby et al., 1980a,b). This role of cholesterol is discussed in Section 1.6.4. 1.6.2. Preparation of liposomes Bilayer membrane liposomes can adopt a wide range of physical properties depending on the lipid composition used and the method of preparation. Consequently, efforts have been directed at generating liposomes using reproducible and readily controllable procedures. Features that need to be monitored are: vesicle size and size distribution, trapped volume, as well as physical and chemical stability (Bally et al., 1988a; Barenholz et al., 1993). The following sections will review the methods described in the literature to generate multilamellar (MLV), large (LUV) and small unilamellar vesicles (SUV). 1.6.2.1. Multilamellar vesicles (MLV) MLVs can be prepared by dispersing lipid in an aqueous medium (Bangham et al., 1965). Typically, MLVs are obtained when lipids deposited from organic solvents in a thin film on a round bottom flask or test tube are hydrated in aqueous buffer above the Tc of 40 the highest melting lipid. The method of hydration can influence size distribution and interlamellar spacing of the resulting MLVs. MLV systems are typically heterogenous in size with vesicle diameters ranging from 0.5 - 10 pm (Mayer et al., 1985b; Hope et al., 1986). It has been shown that for egg PC, the trapped volume (internal aqueous volume) is low («0.5 ul/pmol) due to closely packed bilayers of the lamellae (Mayer et al., 1985b). This trapped volume can be increased to as high as 10 pl/pmol lipid by subjecting these MLVs to five repetitive freeze-thaw cycles, leading to substantial increases in interlamellar spacing and formation of intervesicular structures between lamellae (Mayer et al., 1985b). An alternative method of generating MLVs is the reverse phase evaporation method (Szoka and Papahadjopolous, 1978; Gruner et al., 1985b). This method involves hydration of lipids from an organic solvent. As the organic solvent is evaporated in the presence of an aqueous buffer, the lipids moving from the organic to the aqueous phase spontaneously form vesicles. Although this method is limited by difficulties removing any residual organic solvent from the final preparation as well as by lipid solubility (Gruner et al., 1985b), the resulting MLVs are characterized by high trapped volumes (-10 pl/pmol) and increased interlamellar spacing. In general, these systems have limited applicability for systemic use in vivo. However, MLVs are typically used as precursors for the preparation of LUVs, such as those used in this thesis (see Section 1.6.4.2.). 1.6.2.2. Unilamellar Vesicles Unilamellar vesicles (SUVs and LUVs) may range in size from 30 nm to 200 nm, the size of which depends on the lipid components employed and the method of preparation. Optically clear SUV suspensions can be obtained by sonicating a dispersion of phospholipids (Saunders et al., 1962; Abramson et al., 1964) or MLVs (Huang, 1969). 41 Other methods to obtain SUVs involve use of the French press extrusion technique which involves extrusion of lipid dispersions through a specialized French press device at 20,000 lbs/in2 (Barenholz et al., 1979), the ethanol injection method, where lipids, when dissolved in ethanol are rapidly injected into an aqueous buffer, spontaneously form SUVs (Batzri and Korn, 1973), and the high shear homogenization method involving a specialized apparatus such as the Microfluidizer, where cavitational processing leads to vesicles with markedly reduced sizes (Mason, 1989; Talsma et al., 1989; Vemuri et al., 1990). SUVs may also be generated using extrusion methods (Szoka and Papahadjopolous, 1980; Hope et al., 1985; Mayer et al., 1986b) using a specialized extrusion device operating at pressures approaching 800 psi. Precursor MLVs can be extruded through polycarbonate filters of a defined limit SUV size, resulting in a homogenous distribution of SUVs. Because of the high radius of curvature in these SUVs (30 - 50 nm in diameter), there is a greater percent of phospholipid in the outer monolayer than the inner monolayer, which can potentially lead to asymmetric distribution between the bilayers. Further, a major drawback of SUV systems is that they are limited in terms of trapped volume (<0.5 p:l/p,mol lipid) and therefore exhibit low encapsulation efficiency for most drug delivery applications (Huang, 1969; Barenholz et al., 1979). Because of the instability due to small curvature radius, SUVs can spontaneously fuse to form conglomerates (Parente and Lentz, 1984). Given these limitations for SUVs, LUVs appear suited for drug delivery applications and have become a system of choice for most therapeutic applications. All of the liposomes used in this thesis are LUVs that exhibit vesicle size ranging from 100 - 120 nm. 42 LUVs are desirable as drug carriers for several reasons. Some of the most pertinent attributes of LUVs for their applications in this thesis include. 1) increased trapped volumes, 2) increased membrane physical stability, and 3) greater stability of vesicles in the circulation. There are several methods described in the literature for preparing LUVs. LUVs can be prepared using detergent dialysis (Kagawa and Packer, 1971) where dried lipids are dissolved in detergent containing buffer and removal of detergent by controlled dialysis conditions results in phospholipids adopting a bilayer configuration which are in the 100 - 200 nm in size range. However, factors such as the type of detergent used and the rate of detergent removal must be carefully monitored as these influence size and trapped volume of resulting vesicles (Madden, 1986). In many cases, difficulties remain in completely removing detergent, restricting the use of such vesicles as drug carriers. LUVs can also be obtained using a reverse phase evaporation method, where lipids are hydrated in the presence of an organic solvent, followed by removal of organic solvent under reduced pressure, leading to vesicle formation (Szoka and Papahadjopolous, 1978). While these vesicles have a high trapped volume («10 ul/p:mol lipid), complete removal of organic solvent is cumbersome (Gruner et al., 1985b), and the vesicles are heterogenous in size distribution which often require subsequent processing in order to obtain LUVs (Szoka and Papahadjopolous, 1980). Many of the problems associated with methods used to generate LUVs described above were alleviated using the method of extrusion (Szoka and Papahadjopolous, 1980; Hope et al., 1985; Mayer et al., 1986b), which has been modified, optimised, and made amenable for pharmaceutical scale-up (van Winden et al., 1998). Extrusion methods employ specialized devices operating at a pressure up to 800 psi and precursor MLVs are 43 forced through polycarbonate filters of a defined pore size. This is based on the principle that when frozen and thawed MLVs are forced through filters under pressure, they undergo a deformation and blebbing process, generating LUVs. The method involves formation of a thin film by evaporating a lipid mixture in organic solvent under vacuum, hydration of the film with an aqueous buffer to generate MLVs, freezing and thawing 5x these MLVs to improve trapped volume and extruding them repeatedly at temperatures above Tc of the highest phase transition temperature phospholipid (Nayar et al., 1989). The resulting vesicles have combined advantages of high trapped volume, homogenous size distribution, and high encapsulation efficiencies since they can be extruded at lipid concentrations up to 400 mg/ml (Mayer et al., 1986b). Further, no addition of organic solvents or detergents during hydration are necessary. The method is simple, rapid, and amenable for pharmaceutical scale-up (Barenholz and Lasic, 1996; van Winden et al., 1998). The extrusion method, therefore, was used to prepare all the liposomes described in this thesis. 1.6.3. Drug entrapment within liposomes There are two general methods for incorporating drugs into liposomes depending on the chemical property of the drug to be encapsulated as well as the liposome properties required for the specific drug delivery applications. These include 1) passive entrapment and 2) active entrapment. There are two types of passive entrapment, depending on the chemical nature of the drug. Encapsulation of hydrophilic agents depends upon the trapped aqueous volume in the liposome core whereas the association of hydrophobic agents with the liposome depends upon the ability of the liposome bilayer to solubilize the drug while maintaining vesicle integrity. The method of active entrapment involves the 44 encapsulation of hydrophilic drugs into empty liposomes in response to an active transmembrane ion gradient. This technique is dependent on the ability of the agent to re-distribute across the bilayer in response to a ion gradient. These methods are discussed below. 1.6.3.1. Passive entrapment Passive entrapment has been used for hydrophilic and hydrophobic drugs alike. Hydrophobic drugs can be mixed in organic solvent with the lipids during the preparation of lipids for MLV precursor formation. Examples of such drugs include amphotericin B (Janoff, 1992; Lopez-Berestein et al., 1983) and CsA (Ouyang et al., 1995; Choice et al., 1995). Incorporation efficiency of these agents will depend on packing constraints and lipid - drug compatibility. These drugs, however, demonstrate exchangeability into other lipophilic membranes and consequently, may exhibit high drug leakage rates in vivo (Choice et al., 1995; Madden et al., 1990b). In contrast to hydrophobic agents, hydrophilic drugs are dissolved in the hydration buffer and are entrapped within the intravesicle compartments during liposome formation. Consequently, entrapment of hydrophilic agents depends on the aqueous trapped volume of the liposomes and the lipid concentration used during preparation. For SUVs with trapped volumes -0.2 ul/p:mol, trapping efficiencies can be as low as 1% (Szoka and Papahadjopolous, 1980). For frozen and thawed MLV derived LUVs with trapped volumes averaging 1.5 uVumol and lipid concentrations as high as 400 mg/ml, efficiencies up to 80% can be achieved (Mayer et al., 1985b). This method is largely dependent on the type and physico-chemical properties of both drug and lipid used and these properties may also dictate drug retention and release properties. 45 1.6.3.2. Active entrapment Certain classes of drugs can be encapsulated within empty preformed liposomes by generating a transmembrane ion gradient (Bally et al., 1985; Bally et al., 1988b; Mayer et al., 1986a,c; Madden et al., 1990a; Mayer et al., 1993; Haran et al., 1993). These are typically lipophilic cationic drugs containing ionizable amino groups. An example of this method is illustrated in Figure 1.5, where DOX, a weak base, accumulates inside liposomes in response to a proton gradient. This method exploits the fact that the neutral unprotonated drug species permeates across the membrane orders of magnitude more rapidly than the protonated form (Addanki et al., 1968; Rottenberg, 1979). When utilizing a transmembrane pH gradient, this method involves titrating empty liposomes (which were prepared using pH 4.0 hydration buffer, i.e., interior acidic) with 0.5 M sodium carbonate buffer to pH 7.8 (exterior pH 7.8). DOX is then added to the liposome solution at a temperature above the Tc of the lipid. In the neutral form (unionized) the drug is able to permeate across the bilayer and once inside the acidic interior, DOX becomes protonated. Since the charged species displays very low membrane permeablity, it is retained within the liposome interior. This occurs until equilibrium levels corresponding to [BH+]iN/rBH+]ouT = [FT]IN/[H +]OUT are attained, where BFT is the protonated form of the drug. Therefore, for a ApH of 3 units, interior concentrations of drugs 1000-fold higher than liposome exterior concentrations are attainable. For DOX, when encapsulated within 100 nm diameter EPC/Chol liposomes (drug to lipid ratio 0.2:1.0) using this method, trapping efficiences approach 100% (Madden et al., 1990a). Further, the presence of a pH gradient can also reduce the rate of drug leakage by as high as 30-fold (Mayer et al., 1986a). Although this method is not independent of lipid - drug interactions, it is versatile and can 46 be applied to many liposome formulations capable of maintaining a stable transmembrane pH gradient (Mayer et al., 1993). Further, it has been shown that for several PC bilayers, the pH gradients appears to be stable for several days (Harrigan et al., 1992). Also, addition of cholesterol retards the decay of this pH gradient (Madden et al., 1990a). Amphipathic weak acids can also be encapsulated in a similar manner using interior basic pH (Eastman et al., 1991) using a calcium acetate gradient (Clerc and Barenholz, 1995). Another variation of this method involves the use of ammonium sulphate to generate pH gradients (Ffaran et al., 1993; Lasic et al., 1995; Bolotin et al., 1994). Here, the diffusion of ammonia out of the liposomes generates a pH gradient, enabling agents to be entrapped. Further, sulfate ions are postulated to stabilize the gradient as well as enhance anthracycline retention within liposomes via formation of anthracycline sulfate aggregates (Bolotin et al., 1994). Although these methods can be extended to other lipophilic weak bases, it is important to note that the level of accumulation will vary case by case. Loss of the proton gradient may result in poor drug retention and greater leakage. For example, vincristine leakage from the liposome has been related to a collapse of the pH gradient (Boman et al., 1993). For the studies described in this thesis, DOX was encapsulated within preformed liposomes using the transmembrane pH gradient method. 47 Bring exterior liposome pH to 7.8 with 0.5 M Na 2 C0 3 65°C Add DOX pH 7.8 Outside pH 7.8 Inside pH 4.0 BH+ t l B + H+ Figure 1.5.: Diagrammatic representation of an active loading process. A pH gradient is established when the liposomes (with an interior pH 4.0) are titrated with 0.5 M sodium carbonate to create an exterior pH of 7.8. The equilibrium redistribution of a weak base in response to this pH gradient across the liposome membrane is shown. The neutral form of the molecule is membrane permeable. 48 1.6.4. Biological fate of systemically administered liposomes A number of factors dictate the disposition and pharmacodynamics of systemically administered liposomes. Liposome properties such as liposomal lipid composition, size, dose administered and type of encapsulated agent can influence the biodistribution properties of liposomes and directly impact toxicity and efficacy behavior. Although liposomes have been administered via a number of routes (i.p., i.v., p.o., s.c, i.m.), the discussions here will focus on i.v. administration in view of the applications being investigated for the treatment of MDR solid tumors. Following i.v. injection, liposomes readily interact with various circulating blood proteins and cellular components within the plasma compartment including circulating cells, lipoproteins, proteins, carbohydrates, and ions. These interactions may cause liposome destabilization or initiate processes that mediate liposome drug leakage or clearance by the mononuclear phagocyte system (MPS). The fate of these liposomes depends on interactions between the liposome surface and serum proteins. When proteins adsorb to liposomes, two consequences compromising liposome integrity may occur, namely, increased membrane permeability and MPS recognition and subsequent clearance. 1.6.4.1. Liposome - protein interactions Initial studies utilizing PC liposomes demonstrated the transfer of PC from SUVs to high density lipoprotein (HDL) resulting in liposome destabilization (Scherphof et al., 1978; Scherphof et al., 1983b). Liposomes can interact with lipoproteins and this can result in drug leakage due to lipid exchange and dissolution of the liposome carrier (Kirby et al., 1980a,b). Subsequently, it was shown that liposomes that contained sufficient amounts of cholesterol can avoid the lipoprotein-mediated liposome destabilization (Kirby 49 et al., 1980a,b; Sweeny and Jonas, 1985). HDL and its apolipoproteins are primarily responsible for liposome destabilization. Specifically, ApoA-1 apolipoprotein inserts itself into the lipid bilayer (Klausner et al., 1985). Liposomes may also interact with complement proteins leading to activation of complement cascade (Devine et al., 1994), causing increased membrane permeability and drug leakage. Liposome-protein interactions often lead to opsonization (adsorption of proteins on liposomes allowing macrophages to recognize these foreign bodies and phagocytosize them) of the liposomes which are eventually cleared from the circulation, via Kupffer cells residing in the liver and fixed macrophages in the spleen (Kimelberg and Mayhew, 1978; Finkelstein and Weissman, 1978; Scherphof et al., 1980; Gregoriadis and Senior, 1980; Senior and Gregoriadis, 1982a,b; 1984; Coleman, 1986; Moghimi and Patel, 1989). Opsonins include serum albumin, beta-2-glycoprotein 1, fibronectin, IgG, and certain complement proteins such as C3, C3bi (Absolom, 1986; Bonte and Juliano, 1986; Reinish et al., 1988; Loughrey et al., 1990a; Chonn et al., 1992). Studies have shown that the increased opsonization of liposomes following i.v. administration is directly related to increased elimination from the blood (Woodle et al., 1994; Bonte and Juliano, 1986; Chonn et al., 1992). Further, increased protein binding and subsequent MPS clearance has been shown for liposomes containing PS, cardiolipin, and PA. In addition to increased membrane permeability and clearance observed following liposome - protein interactions, protein adsorption may potentially lead to increased collapse of transmembrane gradients caused by increased stresses on the gradients as well as the high levels of entrapped agents. These high concentrations of buffer components and encapsulated drug in the liposome can lead to osmotic gradients being generated in the 50 physiological milieu and consequently enhance drug leakage (Mui et al., 1994). It has also been shown that DSPC/Chol liposomes can better withstand osmotic gradients than EPC/Chol liposomes in the presence of proteins, suggesting that the effect is less pronounced with highly ordered membranes (Mui et al., 1994). This is corroborated by the fact that EPC/Chol liposomal DOX liberates over 50% of the'encapsulated DOX within 1 h post i.v. injection, compared to the DSPC/Chol DOX formulation (Bally et al, 1990a; Mayer et al, 1989). 1.6.4.2. Drug release from liposomes For most non-targeted liposomal drug formulations, the released drug is the therapeutically active component which is bioavailable for cellular uptake and processing. Consequently, drug release from the liposomes must be measured in order to relate drug leakage from liposomal systems with pharmacological and toxicological properties of the drug. In vitro and in vivo analyses can be utilized to identify liposomal formulations whose drug retention properties meet minimum criteria for their intended in vivo application. In vitro drug release studies can be performed in order to evaluate liposomal drug permeability characteristics in the presence of serum compared to protein-free buffer medium. In vivo drug release, on the other hand, can be determined by measuring liposomal lipid (using non-exchangeable non-metabolizable tracer) and drug in plasma and calculating the drug-to-lipid ratio. This drug-to-lipid ratio is based on the assumption that free drug, not associated with the liposomes is rapidly cleared from the circulation. However, as observed from studies performed in our laboratory, in vitro drug release studies cannot always be used to predict in vivo drug leakage properties due to poor in vitro-in vivo correlations (Bally et al, 1998). This is exemplified for liposomal vincristine, 51 mitoxantrone and DOX formulations. Specifically, vincristine release from DSPC/Chol liposomes was significantly increased in vivo compared to data obtained from in vitro experiments (Bally et al., 1998). Interestingly, in vitro release studies with DSPC/Chol and dimyristoyl glycerophosphocholine (DMPC)/Chol liposomal mitoxantrone demonstrated no significantly different drug leakage properties with <2% released over 72 h for both formulations. However, in vivo release data indicated that 73% of mitoxantrone associated with DMPC/Chol liposomes was released within 48 h, whereas <5% of the drug was released from the DSPC/Chol formulation (Lim et al., 1997). Similar results were obtained with liposomal DOX formulations where in vitro release data did not match in vivo release (Mayer et al., 1994). Based on the information presented above, in vivo drug-to-lipid ratio was used to monitor liposomal drug leakage. In this thesis, two long circulating (saturated synthetic lipid-containing DSPC/Chol and PEG-DSPE/DSPC/Chol) and a rapid release (primarily unsaturated naturally occurring EPC/Chol) liposomal DOX systems are described. Figure 1.6 illustrates the drug-to-lipid ratios for these three liposome formulations. The two saturated lipid formulations exhibit constant drug-to-lipid ratio over the time course evaluated, indicating that DOX remains associated with the liposomal carrier. In contrast, for the EPC/Chol DOX formulation, over 50% of the drug is released within 1 h of systemic administration. These results are consistent with earlier reports originating from our laboratory (Bally et al., 1990a; Mayer et al., 1994). 52 0.25' 0.20 }L 0.15 CD - O 0.10 0.05 0.00--0-EPCADhclDOX - • - D S P C / C h o l DOX ^ _ - A - PEG-DSPE/DSPC/Chol DOX I) 1\ 10 15 20 25 30 35 40 45 50 Time (h) Figure 1.6: In vivo drug-to-lipid ratio determined in plasma of mice treated with 120 nm EPC/Chol, DSPC/Chol, and PEG-DSPE/DSPC/Chol liposomal DOX. Liposomal lipid and DOX were quantified using the methods described in Chapters 3 and 4 (mean ± SD). 1.6.4.3. Factors affecting liposome circulation lifetimes As seen from the above sections, interactions of liposomes with proteins can cause premature drug leakage via increased membrane permeability as well as lead to rapid removal of the liposomes from the circulation. It is generally believed that increasing circulation longevity in the plasma compartment will enhance therapeutic efficacy and benefits of liposomal drug delivery (Gabizon, 1992; Gabizon and Papahadjopolous, 1992). This has led to significant efforts to understand and exploit liposome properties that enhance liposome longevity in the central blood compartment. A number of factors influence liposome clearance from the circulation. These include liposome size, lipid composition, charge and lipid dose. Liposomes of sizes greater than 1 um are more rapidly cleared from the circulation compared to smaller liposomes (particularly those <0.2 p:m; see Gregoriadis and Ryman, 1972; Juliano and Stamp, 1975; Abra and Hunt, 1981). Also, a size specific clearance by macrophages has been shown: 53 liver Kupffer cells take up liposomes with diameters exceeding 0.3 um (Freise et al., 1980), while spleen macrophages take up smaller liposomes (<0.1 um, Poste et al., 1984) however, it should be noted that these processes are not well understood. There is also a relationship between lipid dose and clearance where increased liposome circulation lifetimes are observed as the dose of liposomes is increased. This is thought to be due to the saturability of phagocytic cells involved in liposome removal since a reduction in the percent of injected liposomes in the liver has been related to increases in lipid dose (Poste et al., 1984; Hwang, 1987). However, under these conditions an increase of liposome accumulation in spleen is observed and when this organ becomes saturated, liposomes tend to accumulate in other MPS sites such as bone marrow and lung (Poste et al., 1984; 1983). Chronic administration of liposomes at higher doses also results in an impaired ability of MPS to clear liposomes (Allen et al., 1984; Allen and Smucler, 1985). Such large doses of liposomal lipid may also deplete certain opsonins which can result in increased liposome circulation times (Oja et al., 1996). Liposome clearance is also influenced by liposomal lipid composition and surface charge. In general, a negative charge on liposomes increases their rate of clearance and macrophage uptake (Juliano and Stamp, 1975; Senior et al., 1985; Schroit et al., 1986; Allen et al., 1988; Gabizon and Papahadjopolous, 1992). MLVs composed of 30 mole% PS, PG or PI were cleared by mouse peritoneal macrophages 25-fold, 18-fold, and 15-fold faster, respectively, than PC containing vesicles (Schroit et al., 1986). Depending upon acyl chain composition, protein binding and MPS uptake may be altered suggesting a role of lipid composition in clearance (see Section 1.6.4.1.). The liposomes used in this thesis were typically neutral and were designed to exhibit long circulation lifetimes. 54 1.6.4.4. Strategies to improve circulation longevity It has now become apparent that liposomes exhibiting rapid drug leakage in vivo do not confer increased therapeutic benefits. Unless the disease site lies within the MPS (such as liver and spleen), rapid clearance from the plasma compartment will likely compromise their ability to provide substantial increments in therapy. Consequently, strategies to improve liposome circulation longevity will offer a significant advantage for tumor-targeted liposome systems. Attempts to increase liposome circulation longevity have utilized several approaches. One strategy to improve circulation longevity was based on observations that erythrocytes, the basic structure of which more or less resembles a liposome, were able to circulate for several days before being opsonized. This strategy includes incorporating specialized lipids such as the ganglioside GM1 (Allen et al, 1985; Allen and Chonn, 1987; Gabizon and Papahadjopolous, 1988; Allen et al, 1989; Allen and Mehra, 1989) or sphingomyelin instead of PC as the bilayer forming lipid (Mayer et al, 1995a; Parr et al, 1994) in order to reduce liposome-protein interactions. Incorporation of these lipids results in reduced uptake by MPS, resulting to increased circulation longevity. For example, a reduction in MPS uptake by 3-fold was accompanied by a 4-fold increase in circulation half-lives upon addition of 10 mole% of GM1 in 100 nm PC/Choi liposomes (Allen et al, 1989). This increased circulation life time was suggested as due to reduced opsonin binding (Chonn et al, 1991; 1992). These attempts of reducing opsonization resulted in reasonable improvements in liposome circulation longevity. A significant advance in liposome technology over the past decade has been the identification of the use of polyethylene glycol (PEG) derivatized PE 55 for inhibiting liposome-protein interactions (Kilbanov et al., 1990; Papahadjopolous et al., 1991; Woodle et al., 1994; Lasic, 1996; Allen, 1994; Torchilin et al., 1994; 1995; Cabanes et al., 1998). These lipids are postulated to act by providing a steric barrier that limits the exposure of the liposome surface to plasma proteins, and consequently these liposomes circulate longer in the circulation (Woodle and Lasic, 1992; Torchilin et al., 1994). Since PEG containing formulations have been utilized in this thesis, discussions below will focus on PEG chemistry, mechanisms of liposome-protein inhibition, and applications. 56 1 , 2 - D i s t e a r o y l - s n - g l y c e r o - 3 - p h o s p h o e t h a n o l a m i n e - N - [ P o l y ( e t h y l e n e g l yco l )2000 ] Figure 1.7: Chemical structure of the PEG-DSPE molecule and an illustration of a PEG-coated liposome when this polymerized lipid is incorporated in the initial lipid mixture. Note the polymer coating on interior and exterior surfaces of the liposome. PEG is a flexible hydrophilic polymer with repetitive ethylene glycol units (-[O-CH2-CFf2]n-) (Figure 1.7). It is an inert poorly immunogenic neutral polyether which is available in a variety of MWs (PEG2000, PEG5000, etc). Key observations that led to the application of PEG technology to liposomes include the following: 1) when PEG was 57 used for purification and crystal growth it served to drive proteins away from solution (Poison et al., 1964; Chun et al., 1967), 2) PEG, when covalently attached to proteins, formed conjugates which exhibited enhanced serum lifetimes) (Abuchowski et al., 1977), and 3) protein adsorption on surfaces that are covalently coated with PEG was significantly retarded (Mori et al., 1982). PEG can be incorporated in the liposome in different ways. It can be attached to the surface of preformed liposomes resulting in a polymeric coating on the exterior of the liposome surface (Senior et al., 1991b). PEG can also be coupled to the headgroup of the phospholipid PE and this polymerized lipid can then be incorporated in the initial lipid mixture. A feature of this approach is that the polymer coating exists both on the interior and exterior liposome surfaces (Figure 1.7). PEG modified lipid containing liposomes have been termed sterically stabilized or Stealth™ liposomes. Sterically stabilized liposomes have often been shown to provide enhanced therapeutic activity compared to short circulating liposome systems in a number of animal models (Gabizon and Papahadjopoulos, 1988; Gabizon et al., 1989; Allen et al., 1992; Ahmad et al., 1993; Mayhew et al., 1992; Gabizon, 1992; Williams et al., 1993; Allen et al, 1995; Gabizon et al, 1994; Goren et al, 1996; Gabizon and Martin, 1997; Cabanes et al, 1998). These stealth liposomes possess extended circulation life times (Allen et al, 1991b; Allen and Hansen, 1991; Gabizon et al, 1993), exhibit dose-independent pharmacokinetics (Allan and Hansen, 1991), and several studies have been shown to demonstrate increased solid tumor uptake properties (Gabizon and Papahadjopoulos, 1988; Gabizon, 1992; Torchilin et al, 1994; Gabizon et al, 1996; Goren and Gabizon, 1998). A concentration of 5 mol% PEG-DSPE at a polymer 58 molecular weight of 2000 has been shown to provide maximum inhibition of serum interactions and circulating longevity (Allen, 1994). Incorporation of 5 mol% PEG-polymerized DSPE in 100 nm DSPC/Chol vesicles resulted in a 4-fold increase in circulation life time and a 2-fold reduction in liver accumulation (Papahadjopolous et al., 1991). Studies have also confirmed that addition of PEG at a molar ratio of 5% to the liposome does not alter the normal structure of the bilayer interior (Needham et al., 1992). It should be noted that although PEG lipids improve the circulation lifetime of liposomes, this has not always resulted in increased tumor drug delivery compared to conventional liposomes that retain their encapsulated agents (Parr et al., 1994; Mayer et al., 1997). The reason for these differences in behavior are not understood, but could be related to the nature of solid tumor models used or the liposome dose employed. The mechanism of circulation longevity afforded by PEG derivatized lipids has been attributed to ability of surface presence of hydrated PEG molecules to sterically inhibit electrostatic and hydrophobic interactions between liposome surface and plasma proteins (Allen et al., 1989; Lasic et al., 1991). PEG extends ~5 nm above the liposome surface creating a polymer cloud (Torchilin et al., 1994; 1995), leading to a reduced recognition by the phagocytic cells of the MPS. PEG has been hypothesised to assume a "mushroom" conformation at lower concentrations and a "brush" conformation at higher levels. The PEG coating provides an increased unfavorable surface free energy leading to inhibition of protein binding interactions (Arakawa and Timasheff, 1985). Although these modified lipids reduce the rate of liposome binding and clearance, liposomes do not completely evade the MPS and eventually are taken up by the phagocytic cells in these tissues (Parr et al., 1993). 59 1.6.5. Liposomes in cancer chemotherapy One of the primary means by which clinicians attempt to improve cancer chemotherapy is to maximize anticancer dose intensity (De Vita et al., 1993). This is based on the philosophy that tumors cells must be exposed to the highest levels of the anticancer drug for extended periods of time if maximum therapeutic efficacy is to be achieved (Livingston, 1994). The pharmacodynamic alterations offered by liposomal encapsulation can be effectively utilized for such anticancer drug delivery applications. This is because liposomes provide the potential for 1) increased stability of the encapsulated agent, 2) a circulating microreservoir or depot of the encapsulated drug, releasing it as a function of time (sustained release), 3) increased selectivity of drug delivery to the tumor site, and 4) reduced peak free drug levels and decreased exposure to healthy tissues. Following systemic administration, liposomes are confronted with a healthy vascular endothelium which consists of the blood vessel lining composed of endothelial cells, an underlying basement membrane, and smooth muscle, which poses a significant barrier for the circulating liposomes in healthy tissues. However, disease sites such as infection, inflammation, and tumor growth all share a feature that can be exploited by liposome delivery systems, namely, altered vasculature permeability (see Section 1.6.5.2.). Blood vessels associated with these disease sites are characterized by frequent interruptions along the endothelial cell lining, which is permeable to many circulating macromolecules. Macromolecules may leak through these "fenestrae" or gaps (Kohn et al, 1992) or endothelial cell facilitated transcytosis may occur (Huang et al, 1993). The process by which specific barriers are crossed during passage from blood to the interstitial space has been described as extravasation. Liposomes with long circulation life times 60 exhibit extended access to the openings in the endothelial lining which can often lead to increased accumulation in such extravascular sites of disease (Gabizon and Papahadjopolous, 1988; Gabizon et al., 1990; Gabizon, 1992). In order to better understand mechanisms of passive extravasation and how liposomes can be designed to take advantage of this process, the following sections will review normal vasculature as well as the differences that exist in the tumor vasculature. 1.6.5.1. Normal Vasculature There are three types of capillary endothelium (Poste et al., 1984; Jain, 1987), namely continuous, fenestrated, and sinusoidal (or discontinuous), based on endothelium and basement membrane structures. Continuous endothelium is composed of tight junctions («2 nm) and a continuous basement membrane (e.g., muscle, connective tissue, skin, and heart). These continuous endothelial layers demonstrate the lowest permeability characteristics. Fenestrated endothelium is characterized by interendothelial cell junctions (40-60 nm) and typically a continuous basement membrane. This type of endothelium is found in the kidney glomerulus, pancreas, and other glands. Finally, the sinusoidal type of endothelium, which is perhaps the most permeable, is composed of junctions up to 150 nm and a discontinuous basement membrane (e.g., liver, spleen, and bone marrow). The basement membrane is absent in the liver and is fragmented in spleen. Macromolecules extravasate in normal endothelium by transcapillary pinocytosis as well as via interendothelial cell junctions. These blood capillaries are primary sites of nutrient exchange between blood and tissue in normal vasculature. Inasmuch the permeability of normal vasculature is rigidly controlled, Nagy et al. (1989) and Bally et al. (1994) have 61 argued that macromolecules can cross the normal endothelium via a slow extravasation process. 1.6.5.2. Tumor Vasculature Tumors are generally characterized by a discontinuous or poorly defined vasculature which renders the tumor hyperpermeable to circulating macromolecules. This is because of deterioration in the structure of the endothelium (Dvorak et al, 1988; Brown et al., 1989) such as fenestrations, widened interendothelial junctions, or a lack of endothelial cell lining (Kohn et al., 1992). Tumors are vastly heterogenic in terms of defined vascular structure and in many tumors, this structure is very poorly defined. There are therefore considerable differences compared to normal healthy tissue in terms of cellular composition of the capillary wall, basement membrane, and size of interendothelial cell fenestrae, within any one tumor. This is further complicated by a dynamic process of new blood vessel formation (angiogenesis) as well as unpredictable closure of tumor capillaries leading to formation of shunts within a region. Despite the relative abundance of new blood vessel sprouts, tumor blood flow is slower than in normal tissues, due to the presence of tortuous blood vasculature, dead ends, and cut off loops (Jain 1987, 1988). Solid tumors are characterized by a high interstitial pressure due to lack of a lymphatic system (Jain, 1988). In larger tumors, such pressure increases lead to vascular collapse and eventual development of necrotic foci. In general, due to these irregularities and interruptions in the vascular lining of blood vessels within a tumor mass, tumors are hyperpermeable to circulating macromolecules. There are two underlying physiological implications of this hyperpermeability: 1) to facilitate import of fibrin necessary for a functional tumor interstitium (Nagy et al., 1988), and 2) to facilitate influx of 62 macromolecules and nutrients necessary for tumor cell survival through the release of permeability factors such as vascular permeability factor (VPF also called vascular endothelial growth factor, VEGF) by tumor cells (Dvorak et al., 1991). Consequently, large molecules may pass through the large gaps or fenestrae between adjacent endothelial cells (Kohn et al., 1992; Huang et al., 1993). These gaps in the endothelial cell lining may range in size from 30 nm for fenestrated capillaries to between 400 - 500 nm in liver, tumor and many sites of inflammation (Dvorak et al., 1988; Yuan et al., 1995). In view of the information above, it may be expected that LUVs of size ranging from 100-200 nm, therefore, can passively extravasate through the gaps in the endothelial layer and accumulate in sites of tumor growth. However, given the high hydrostatic pressure within the tumor site which will generate a pressure gradient impeding passage of molecules from the blood into the tumor interstitium (Jain, 1993; Buocher and Jain, 1992), additional factors may dictate accumulation of liposomes. The lack of lymphatic drainage in the tumor, for example, may lead to an extravascular trapping effect (Buocher and Jain, 1992). This is due to interstitial diffusion of molecules which, in the absence of lymphatic drainage, leads to egress from the disease site. The diffusion rate of egress is dependent upon the size of the molecule, for example such that may leave the site more rapidly than larger molecules or macromolecules (Buocher and Jain, 1992). Liposome extravasation and accumulation in solid tumors has been well documented by numerous laboratories using a wide variety of tumor types (Gabizon and Papahadjopolous, 1988; Mayer et al., 1989; Allen et al., 1989; Gabizon, 1992; Wu et al., 1993; Yuan et al., 1994). Liposomes have been shown to enter tumors by two paths: 1) via endothelial cell uptake and transcytosis across endothelial cells, and 2) via the gaps 63 present in the tumor vasculature, determined using videomicroscopy studies (Wu et al., 1993; Yuan et al., 1994). This preferential accumulation of liposomes in solid tumors has been used to increase the tumor delivery of entrapped anticancer drugs which result in tumor drug levels that are 5-15 fold higher than can be achieved with free drug (Gabizon and Papahadjopolous, 1992; Mayer et al., 1994; Parr etal., 1997; Webb etal., 1995b). In summary, liposomes with extended circulation times may passively accumulate in sites of tumor progression, given their ability to permeate across gaps in the "leaky" endothelial layer. Also, since normal healthy tissues are characterized by a "tight" vasculature, liposomes tend not to accumulate due to size limitations. Therefore, by providing selective localization in the tumors, liposomes also provide maximal anticancer dose intensity in the absence of toxicity to susceptible tissues. Consequently, many properties of liposomes appear well suited for MDR tumor applications, namely: 1) their utility to provide increased anticancer drug dose intensity to PGP overexpressing tumors, 2) reduced exposure to normal tissues expressing PGP, and 3) favorable modification of pharmacokinetic behavior. In the following sections, applications of liposome technology in MDR will be reviewed. Initial discussions will focus on use of liposome anticancer agents alone for MDR reversal, followed by applications of liposomes to deliver PGP blockers. 64 1.6.6. Multidrug resistance 1.6.6.1. Applications of liposomal anticancer drugs for MDR Liposomes have been used widely to improve the therapeutic activity of encapsulated anticancer agents based on the properties described in the above section. Improvements in liposome technology such as the production of target specific systems and extended circulation life times have led to an increase in the therapeutic index of several anticancer drugs in numerous tumor types (Yuan et al., 1994; Huang et al., 1992; Williams et al., 1993). Further, the ability of small (100 nm vesicles) liposomes to passively extravasate in tumor tissues has resulted in increased selectivity of anticancer drug delivery at the tumor site, while markedly reducing drug accumulation and toxicity in many susceptible healthy tissues (Gabizon, 1992; Gabizon and Papahadjopoulos, 1988; Mayer et al., 1989). This increased delivery of anticancer drugs has the potential of overcoming MDR based solely on mass action, provided that the level of drug resistance is of a magnitude comparable to the increase in tumor drug levels. One of the early indications that liposomes can be beneficial in this manner for MDR reversal was observed in a study that demonstrated enhanced activity of a long circulating liposomal formulation incorporating either the ganglioside GM1 or PEG-polymerized DSPE for doxorubicin and epirubicin against the murine C26 colon carcinoma (Huang et al., 1992; Mayhew et al., 1992; Papahadjopoulos et al., 1991). These C26 colon carcinoma cells are known to express moderate levels of PGP and doses of free doxorubicin up to 10 mg/kg are rendered ineffective (Huang et al., 1992) in vivo. Results indicated that mice receiving stealth formulations of DOX or epirubicin resulted in tumor regression to nonmeasurable sizes, with 90 and 100% long term 120 day survivors in groups that were treated with sterically 65 stabilized liposomal epirubicin and DOX, respectively. In comparison, free drugs did not have any effect on delaying tumor growth (Huang et al., 1992). The authors attributed this enhanced activity in a relatively resistant tumor model to be due to enhanced localization of the drug in tumor tissue, drug release from liposomes into the extravascular spaces and uptake of the released contents by the tumor (Huang et al., 1992). The therapeutic efficacy of free epirubicin and liposomal epirubicin were also evaluated in the C26 colon carcinoma model by Mayhew et al., 1992. As observed with the study of Huang et al., 1992, liposomal epirubicin exhibited enhanced activity in this anthracycline resistant murine colon carcinoma model. This ability of the liposomes to inherently overcome a certain degree of MDR stems from the fact that liposomes can markedly enhance (between 5-15 fold) the amount of drug that can be delivered to tumors compared to free drug. For example, if the tumor exhibits a resistance factor of 5-fold, this could theoretically be overcome by increasing anticancer drug dose intensity through liposomal delivery. The fact that stealth liposomes were able to provide tumor regression in C26 colon carcinoma cells (Huang et al., 1992; Mayhew et al., 1992) by delivering higher amounts of the drug supports this argument. However, for tumors exhibiting higher resistance levels, liposomes by themselves may be unable to circumvent MDR significantly as demonstrated in a rat glioblastoma tumor model (Hu et al., 1995). Antibody-coated liposomes (also called immunoliposomes) have the ability to provide targeting and selective toxicity towards tumor cells and this has been demonstrated in KLN-205 squamous cell carcinoma in vivo (Ahmad et al., 1993). By targeting tumor cells expressing a specific internalizable surface epitope, these immunoliposomes provide a 66 unique approach of providing a intracellular tumor delivery of the drug (Marjan et al., 1996). This is exemplified by a recent in vitro study which reported that DOX resistance was modulated by an immunoliposome targeting transferring receptor in MDR human leukemic cells (Suzuki et al., 1997). However, site-directed targeting may not result in enhanced antitumor activity unless targeting is directed to surface markers that can be endocytosed since most tumor cells do not actively internalize non-targeted liposomes. Further, studies have shown that the PEG on the surface of the sterically stabilized liposome causes a steric hindrance which may interfere with antibody when directly conjugated to the lipid bilayer (Lee and Low, 1995; Storm et al., 1994; Hansen et al., 1995). In addition, given the penetration difficulties experienced for immunoliposomes in larger solid tumors, it is likely that their application to circumvent MDR may be limited to circulating hematopoietic malignancies as has been suggested by Allen and co-workers (Lopes de Menezes et al., 1995). Some of the alternative approaches using liposomal anticancer drugs include the use of thermosensitive liposome encapsulated DOX in conjunction with hyperthermia which resulted in increased activity against MDR MCF7 cells (Merlin et al., 1993), and the use of folate-mediated tumor cell targeting of liposomal DOX in tumor cells that overexpress folate binding proteins such as ovarian carinomas (Lee and Low, 1995). However, these novel approaches have limited applicability in vivo given the fact that the current conditions for hyperthermia are unlikely to be useful for visceral or widespread systemic malignancies. The use of folate-PEG-liposomes may increase the specificity for tumor cells overexpressing the folate receptor, however, this approach relies on effective 67 internalization of the liposomes into the vast majority of tumor cells which may be difficult in view of the problems experienced with tumor penetration for immunoliposomes. 1.6.6.2. Use of liposomes for delivery of MDR modulators Studies with liposomes composed of certain acidic phospholipid based systems such as phosphatidylserine (Fan et al., 1990) or cardiolipin liposomes (Rahman et al., 1992) have shown that these lipids are able to increase the cytotoxicity of encapsulated or complexed anticancer drugs against MDR cells. MDR reversal using such phospholipids have been related to increased intracellular delivery of anticancer agents as well as a direct PGP blocking effect (Thierry et al., 1992). Several in vitro investigations with acidic phospholipids have been described. In vitro work in colon cancer cell lines (Oudard et al., 1991), human ovarian carcinoma cells (Thierry et al, 1992), human leukemia cells, and human breast carcinoma cells (Thierry et al, 1994) suggested that liposomal encapsulation may have beneficial effects in MDR cells. In some systems, liposomes can increase intracellular drug accumulation and work by Thierry et al (1992) suggested that cardiolipin containing liposomes directly alter PGP function as illustrated by the fact that empty liposomes inhibited specific [3H]-vincristine binding to PGP-enriched membranes. Cardiolipin liposomes complexed with DOX have been shown to reverse MDR in human breast carcinoma MCF7/ADR cells in vitro (Thierry et al., 1994). While these results provided encouraging data implicating certain lipids as PGP modulating agents, acidic lipids such as cardiolipin and phosphatidylserine are readily recognized by phagocytic cells of the reticulo-endothelial system and liposomes containing these lipids are cleared within minutes from the circulation offering limited systemic availability in vivo. Whereas cardiolipin based liposomal encapsulation of DOX has been 68 shown to have enhanced efficacy in an in vivo murine ascitic L1210 leukemia model compared to free drug (Gokhale et al., 1996), this model does not reflect conditions for systemic administration of therapeutic agents to treat a distal extravascular disease site. Given the rapid clearance of negatively charged liposomes from the plasma, it is yet to be demonstrated that such liposome systems will be of use in treating solid tumors. Since liposomal formulations offer considerable advantages over free drugs for many biological agents, the use of strategies involving liposomes for the delivery of biopharmaceuticals has been increasing. As an example, antisense oligonucleotides against mdr-1 mRNA exhibits greater therapy when encapsulated in liposomes and enhances PGP synthesis reductions (Thierry et al., 1993). It has been recently demonstrated (Alahari et al., 1996) that phosphorothioate antisense oligonucleotides reduced levels of mdrl message inhibiting the expression of PGP as well as affecting drug uptake. However, effective reduction in mdrl mRNA and protein levels was only obtained when used in conjunction with cationic liposomes (Alahari et al., 1996). Currently, the therapeutic utility of such cationic liposomes are limited to local delivery since these systems are also cleared very rapidly after systemic administration and are unlikely to provide any systemic delivery. A related study reported the liposome-mediated transfer of MDR-1 ribozymes into MDR human pleural mesothelioma cells which restored sensitivity towards anticancer drugs (Kiehntopf et al., 1994). Attempts have been made to encapsulate MDR modulators in liposomes. However, to date the MDR modulators are either not amenable for liposomal encapsulation or exhibit high leakage rates when encapsulated agents are exposed to circulating components after i.v. injection. For example, it has been shown that modulators 69 such as verapamil, prochlorperazine (Webb et al., 1995a), and cyclosporin A (Ouyang et al., 1995; Choice et al., 1995) could be effectively entrapped within liposomes. However, their very high membrane permeabilities resulted in rapid leakage from the liposomes on i.v. adminstration (Choice et al., 1995; Mayer, unpublished observations) which compromises any potential advantages of such liposomes to provide improved selectivity of tumor delivery for these agents through the use of long circulating liposomes. It is speculated that stable liposomal modulator formulations will increase the selectivity of their reversal properties in solid tumors. However, in order for such liposomes to be formulated, the modulator must possess either higher affinity for the liposomal membranes or significantly reduced membrane permeability when entrapped in the liposome interior neither of which has been achieved to date. 1.7. Thesis rationale and specific aims 1.7.1. Rationale Over recent years, MDR has been a subject of numerous investigations aimed at identifying the underlying molecular mechanisms and methods to overcome it. Although a number of these studies describe convincing reversal in several murine and human experimental cell lines in vitro, in vivo studies have yet to clearly demonstrate increased efficacy. This lack of convincing reversal in vivo could be due to poor drug and/or modulator delivery to the target site, resulting in sub-optimal concentrations being achieved in the tumor cells. PGP is a well characterized and perhaps best understood mechanism associated with MDR (preclinically and clinically) and consequently, this thesis has used PGP overexpressing in vitro and in vivo models. Moreover, the selective procedures that 70 induce PGP based MDR in tissue culture can be carefully controlled which allows for precise interpretation of results. This thesis acknowledges that PGP is not the single underlying mechanism of MDR, and that there may be an interplay of alternate multi-factorial resistance mechanisms that may be clinically important. Regardless of the presence of other mechanisms, this thesis is designed to evaluate whether poor drug delivery is a major cause for the limited therapy of well defined MDR tumors, and for this reason, as well as its prevalent overexpression in many clinical drug resistant tumors, PGP based MDR tumor models were selected for study. Based on historical studies in the areas of both MDR and liposomal drug delivery, it is proposed that sufficient evidence is available to warrant testing the hypothesis that inadequate tumor delivery of anticancer agents and selectivity of PGP modulation are primarily responsible for the attenuated therapy of extravascular MDR solid tumors overexpressing PGP. 1.7.2. Specific aims 1. To perform in vitro cellular pharmacological studies in order to select the MDR modulator best suited for in vivo applications. This is described in Chapter 2, where a series of cellular pharmacology studies were initiated to characterize a wide range of PGP inhibitors with respect to the extent of MDR reversal, potency, and latency of activity. Studies were performed to assess the cytotoxicity of DOX against two MDR tumor lines and their drug sensitive parental cell lines in the presence of eleven different MDR modulators. These cytotoxicity studies were correlated with the ability of MDR modulators to increase DOX uptake and decrease DOX efflux from the MDR cells using flow cytometry. Visualization of DOX cellular uptake properties using 71 fluorescence microscopy was correlated with cellular DOX concentrations, in the presence and absence of the MDR modulators. 2. To demonstrate improved toxicological and therapeutic MDR reversal properties of DSPC/Chol liposomal DOX compared to free (non-encapsulated) DOX co-administered with PSC 833 in a P388/ADR solid tumor model. (a) To perform toxicity studies in order to compare PSC 833 induced changes in maximum tolerated doses and toxicity profiles for free and liposome entrapped DOX. (b) To compare the antitumor activity of free and liposome encapsulated DOX in combination with PSC 833 using a murine MDR solid tumor model (c) To elucidate DOX pharmacokinetics and correlate drug distribution properties with toxicity and therapeutic activity. The results of these experiments are presented in Chapter 3, where increased tumor growth suppression was demonstrated with liposomal DOX and PSC 833 compared to free DOX and liposome encapsulation appeared to circumvent adverse DOX - PSC 833 interactions observed with free drug. 3. To determine the influence of liposome physical characteristics on liposomal DOX toxicity, efficacy and pharmacokinetics when co-administered with PSC 833. These studies utilized a human breast carcinoma MDA435LCC6 MDR1 solid tumor model and three different liposome formulations, namely EPC/Chol, DSPC/Chol, and sterically stabilized PEG-DSPE/DSPC/Chol liposomal DOX. These studies were designed to investigate the role of DOX release from the liposomes as well as liposome tumor accumulation properties in determining the toxicity and efficacy characteristics 72 of liposomal DOX in conjunction with PSC 833. In addition, in depth pharmacokinetic and cellular distribution studies were conducted in order to reveal the processes underlying the empirical toxicity and efficacy observations. These results are presented in Chapter 4, where the three types of liposomal DOX formulations are compared, in the presence and absence of PSC 833, using the human xenograft solid tumor model. 4. To investigate the alterations in DOX renal and biliary clearances by PSC 833 in a chronically instrumented rat model and to determine the role of liposomal encapsulation on the DOX elimination pathways. The results of these studies are presented in Chapter 5, where studies were performed to evaluate DOX pharmacokinetic parameters in plasma, bile, and urine. These studies used free (non-encapsulated) DOX as well as two different liposomal DOX formulations in the presence and absence of PSC 833. 73 Chapter 2 COMPARATIVE IN VITRO EVALUATION OF DOX CYTOTOXICITY, CELLULAR UPTAKE & RETENTION, AND INTRACELLULAR DISTRIBUTION IN THE PRESENCE AND ABSENCE OF MDR MODULATORS: SELECTION OF MDR MODULATOR WITH OPTIMIZED REVERSAL PROPERTIES FOR IN VIVO APPLICATIONS 2.1. Introduction As described in Chapter 1, the use of MDR modulators (reversibly referred to as chemosensitizers) to chemosensitize MDR tumors has been the subject of numerous preclinical and clinical investigations (see Fan et al., 1994 for a review). The in vitro studies described in this Chapter were undertaken in order to select the MDR modulator best suited for in vivo applications based on the extent of MDR reversal, potency, and latent modulating activity. A number of first as well as newer second generation modulators, such as Rol 1-2933, a new tiapamil analog (Alaoui-Jamali et al., 1989; Alaoui-Jamali et al., 1993; Bankusli et al., 1989), Ro44-5912 (de Jong et al., 1995), dexniguldipine (DNG, Hofmann et al., 1995; Hofmann et al., 1991), and PSC 833 (Jiang et al., 1995; Keller et al., 1992b), that were available at the time of initiation of the project, were evaluated for MDR reversal properties in two independent PGP overexpressing cell lines, namely, the murine lymphocytic P388/ADR line as well as the human breast carcinoma MCF7/ADR line. The following sections will discuss MDR reversal with existing chemosensitizers and will highlight properties that are likely to impact in vivo MDR therapy. A wide range of chemical entities have been identified as potential modulators of MDR by examining the degree of chemosensitization that occurs in vitro upon exposure of the resistant cells to the modulators in conjunction with anticancer drugs such as DOX. 74 However, convincing results of in vivo MDR reversal both in preclinical models and clinical trials have been limited (see Section 1.5. in Chapter 1). Although various pharmacological and molecular mechanisms have been postulated to be responsible for the sub-optimal therapy seen in vivo, it is still unclear whether inherent MDR modulating properties of compounds can be correlated with their in vivo activity. Two key properties of MDR reversing agents must be met if successful modulation of transport-based MDR in vivo is to be achieved: These are 1) complete reversal of the resistance mechanism targeted by the modulator and 2) synchronous presence of the modulator activity and anticancer drug in resistant tumor cells to enhance influx of the drug in the cell and improve therapy. The failure to accomplish either of these two conditions may lead to ineffective therapy of MDR tumors regardless of the presence of resistance mechanisms beyond PGP. Previous cell culture studies with MDR modulators have focused primarily on the degree of MDR reversal, often referred to as fold-reversal (Alaoui-Jamali et al., 1993; Gekeler et al., 1995a; Fan et al., 1994; Hofmann et al., 1991; Hofmann et al., 1995; Jiang et al., 1995; Keller et al., 1992b; Ramu et al., 1984a,b; Tsuruo et al, 1982a,b; Tsuruo et al, 1983a,b,c; Tsuruo et al, 1984). However, for in vivo applications, the levels of residual resistance remaining after exposure to chemosensitizers is likely to dictate therapeutic activity rather than the fold-reversal in itself. This is due to the fact that even low levels of residual resistance may pose a substantial barrier to chemotherapy where the window between therapeutic and toxic anticancer drug doses is typically very narrow. Unfortunately, little information is available on the residual resistance that remains after 75 treatment with various MDR modulators and whether this feature correlates with the degree of therapeutic activity obtained in vivo. In addition to considerations on the extent of MDR reversal, few studies have addressed the duration of MDR reversal that exists once the extracellular modulator concentrations fall to sub-therapeutic levels. For example, the fact that the modulation efficiency of verapamil (VRP) is lost when the cells were washed free of the chemosensitizer and only extremely high pulses of the modulator provide any extended activity, suggests that the PGP blockade may perhaps be transient depending on chemical properties of the modulator and its binding affinities (Hyafil et al., 1993; Dantzig et al., 1996). Presumably, the clearance of the modulator from the site of action may lead to poor PGP blocking activity (Perez et al., 1993; Boesch and Loor, 1994; Mayer and Hartley, 1998). Again, the lack of understanding in this area significantly limits the ability to resolve the factors that dictate in vivo MDR reversal as well as attempts to improve therapeutic strategies. DOX (see Scheme 2.1) was chosen as the model anticancer drug for the studies described in this thesis for the following reasons: 1) It is a versatile broad spectrum anticancer drug used in a variety of human malignancies. These include acute leukemia, advanced Hodgkin's disease, advanced breast carcinoma, ovarian cancers, and adult soft tissue sarcomas. Although effective as a single agent in many cancer types including advanced breast cancer, it is typically used in conjunction with other anticancer drugs in combinations, such as bleomycin, vinblastine, cyclophosphamide, vincristine, fluorouracil, among several others (De Vita et al., 1993); 2) DOX is a PGP substrate and is therefore pumped out of MDR cells overexpressing PGP. This has been demonstrated in a number 76 of preclinical and clinical MDR models (See section 1.2.2.2.2.). Consequently, it serves as a model anticancer drug for the evaluation of PGP-based MDR modulation using the chemosensitization strategy in conjunction with a PGP blocker; and 3) Finally, liposomal encapsulated formulations of DOX have been well characterized and are readily available. Consequently, this allows testing the hypothesis whether MDR tumor therapy can be improved by presenting DOX in liposomes in the presence of an MDR modulator (see Chapters 3 and 4). Scheme 2.1.: Chemical structures of neutral, anionic, and cationic species of DOX. DOX, a weak base with a pKa of 8.2, is a red crystalline powder. Illustrated in Scheme 2.1 are the various species of neutral, cation, and anion forms. A number of mechanisms have been postulated for the underlying mechanisms of DOX action in tumor cells. DOX has been suggested to cause cytotoxicity via DNA intercalation (Pigram et al., 1972) and topoisomerase II activity (Tewey et al., 1984b), via alterations in membrane 77 binding (Tritton et al., 1978), and via the generation of reactive oxyradicals (Bachur et al., 1977; Doroshow, 1986). DNA intercalation was one of the earliest described mechanism of DOX action, where it was postulated to bind to DNA, affecting both DNA and RNA synthesis which resulted in single and double strand breaks (Pigram et al., 1972). It has been suggested that anthracyclines insert themselves into the DNA double helix in a manner that the aglycone moiety is between the adjacent base pairs (Pigram et al., 1972). Following DOX binding to DNA, inhibition of DNA synthesis, ribosomal RNA and protein synthesis (Daskal et al., 1978), as well as DNA fragmentation (Schwartz, 1975) and inhibition of DNA repair, have been shown to occur (Painter, 1978). Evidence suggests that these breaks were caused by DOX effects on topoisomerase II, an enzyme that promotes DNA strand breakage and resealing (Tewey et al., 1984b). When DOX intercalates with DNA, this causes alterations in DNA conformation, arresting topoisomerase II activity, at the point of DNA cleavage (Myers and Chabner, 1990). Another mechanism of DOX action stems from its ability to bind to cell membranes and alter membrane function at concentrations that inhibit DNA synthesis. Binding to cardiolipin has been suggested as a mechanism of DOX action as well as its cardiotoxic properties, and in support of this, higher levels of DOX in both malignant cells and the cardiac mitochondria (where elevated cardiolipin concentrations exist) were observed (Tritton et al., 1978). Further, the changes associated with the development of chronic cardiomyopathy with DOX has been correlated with altered sodium and calcium ions in cardiac tissue (Olson et al., 1974). Specifically, DOX has been shown to increase the sodium permeability as well as alter the calcium handling in many experimental models (Solie and Yuncker, 1978; Villani et al., 1978). 78 Anthracyclines have also been shown to trigger production of reactive oxygen radicals. Early observations have shown that the microsomal cytochrome P450 reductase catalyses the reduction of DOX to a semiquinone free radical, which in turn readily reduces molecular oxygen to superoxide ion (Sato et al., 1977; Handa and Sato, 1975; Handa and Sato, 1976). Other flavin-dependent oxireductases are also capable of reducing DOX to the semiquinone free radical and subsequently, the oxygen radical (Pan and Bachur, 1979; Thayer, 1977). Free radical mediated membrane damage in cardiac tissue has been postulated as a mechanism of action of DOX cardiomyopathy, given the observation that vitamin E, a free radical scavenger, is able to confer protection to DOX-mediated cardiac muscle damage (Myers et al., 1977). Chemically reduced DOX is also capable of causing single strand breaks in DNA, via superoxide formation (Lown et al., 1977). The MDR models used for in vitro modulator screening studies included 1) P388 murine lymphocytic leukemia (sensitive, WT and PGP positive, ADR), and 2) MCF7 human breast carcinoma (sensitive, WT and PGP positive, ADR). P388 lymphocytic leukemia is a murine cell line that has been extensively used by the National Cancer Institute as an in vitro and in vivo screen. The ADR line is at least >75 fold resistant than the sensitive WT cell line thus enabling the evaluation of MDR reversal as well as the presence of residual resistance. MCF7 is a human breast carcinoma cell line, available in sensitive and MDR lines. The MDR line is at least >50 fold resistant, allowing comparative assessments of MDR reversal to be made between modulators. Like the P388 cells, these adherent type cells have a well denned MDR profile and are easy to maintain in tissue culture for extended periods of time. 79 In this chapter, the cytotoxicity, cellular uptake and retention, as well as intracellular distribution of DOX have been determined in the presence and absence of a wide range of MDR modulators in the PGP positive murine P388/ADR leukemia and human MCF7/ADR breast carcinoma cell lines with an aim to select an MDR modulator with optimum reversal properties (ability to completely reverse MDR, high potency, and latent modulating activity) for in vivo applications. Specifically, studies were designed to 1) assess the cytotoxicity of DOX in continuous and pulsed exposures using a microculture tetrazolium assay, 2) uptake and release kinetics of DOX using flow cytometry, and 3) intracellular drug distribution characteristics using fluorescence microscopy, in the presence and absence of the MDR modulators. Cellular DOX uptake was evaluated using in-line kinetic laser flow cytometry in order to assess the extent and homogeneity of MDR modulation in viable, individual cells. The results of these investigations provide useful information for the selection of MDR modulators for in vivo applications. 2.2. Materials & Methods 2.2.1. Chemicals, Animals, Cells & Supplies Mice (BDF1, i.e., hybrid of DBA/2 and C57BL/6 mouse strains) were purchased from Charles River (Quebec, Canada). P388 (murine lymphocytic leukemia cell line) sensitive (wild-type, WT) and (multidrug resistant, selected for DOX) ADR cells were obtained from the National Cancer Institute (Maryland, USA). MCF7 WT and ADR cells were obtained from Dr. Erasmus Schneider at the National Cancer Institute (Maryland, USA). RPM1-1640 culture medium, supplemented by 5% L-glutamine, was obtained from Stem Cell Technologies (Vancouver, Canada). Penicillin and Streptomycin (10,000 units/ml), heat inactivated fetal bovine serum (FBS), Nunc tissue culture flasks (75, 175 80 cm2), and trypan blue were purchased from GTBCO (New York, USA). Dimethylsulphoxide (DMSO) was purchased from Sigma Chemical Company (Missouri, USA). DOX was supplied by Adria Laboratories (Ontario, Canada). Experimental chemosensitizers used in these studies were gifts from various sources: PSC 833 (Novartis Canada, Inc.), dexniguildipine (DNG, BYK-Gulden, Germany), dexVRP (Knoll Pharmaceuticals, Canada), Rol 1-2933 (Hoffmann La Roche Inc., Basel, Switzerland), trans-E-flupenthixol (Ff. Lundbeck, Copenhagen, Denmark). Others (VRP, chlorpromazine, tamoxifen, prochlorperazine, quinidine) were purchased from Sigma (Missouri, USA). C219 monoclonal antibody (mouse ascites, IgG2a, specific for PGP), to assess PGP positivity in cell lines, was obtained from Signet Laboratories, Inc (Dedham, MA). The negative isotypic antibody was obtained from Pharmingen (San Diego, CA) and the FITC-labeled goat-anti-mouse IgG-antiserum was purchased from Jackson (West Grove, PA). 2.2.2. General Equipment VWR VanlabR vortex mixer (Scientific Industries, Inc., Bohemia, NY); Centronics11 Silent Series S-103NAR centrifuge, IEC Micromax centrifuge, International Equipment Company, Needham Heights, MA, and GLC-4 General laboratory centrifuge, SorvallR Instruments, Dupont for centrifuging 96-well culture plates; Precision All Stainless Steel Water Bath Model 183 (Precision Scientific, Inc., Chicago, IL); Minishaker (plate shaker), Dynatech Industries, Middletown, NY; LABGARD Laminar Flow Biological Safety Cabinet, NuAire, Inc., Minneapolis, MN; Precision Automatic CO2 Incubator, NAPCO Waterjacketed C0 2 Incubator (VWR Scientific); Edwards LABOPORTR Vacuum System Model N820.3 FT. 18, Edwards High Vacuum 81 International, West Sussex, UK; Acumet Model 230 pH meter (Fisher Scientific, Fairlawn, NJ) for measuring pH of buffers, and Acumet 925 pH meter for titrating liposomal solutions (Fisher Scientific, Fairlawn, NJ). 2.2.3. Drug solutions DOX and MDR modulator solutions were prepared freshly prior to experimentation. VRP, dexVRP, Rol 1-2933, quinidine, prochlorperazine, and chlorpromazine were dissolved in Sterile Water For Injection, USP. DNG, PSC 833, and trans-E-flupenthixol were dissolved in "99% ethanol before preparing subsequent dilutions in water and finally in culture media (final ethanol concentration below 0.5%). 2.2.4. Cell Lines & Culture The culture medium requirements for P388AVT (sensitive P388 cells) and P388/ADR (multidrug resistant P388 cells) cells include RPMI 1640 medium containing added antibiotics, 10% FBS and 1% L-glutamine. MCF7 cells were maintained in tissue culture (RPMI 1640 containing 10% FBS and 1% L-glutamine) whereas P388 cells were passaged in the peritoneum of BDF1 mice in the presence (P388-ADR) or absence (P388-WT) of DOX according to NCI protocols. DOX resistant P388 cells were passaged in mice that were treated with 6 mg/kg i.v. of DOX two days after tumor cell incubation. After 7 days, the cells from the peritoneum of the tumor bearing mice were harvested into EDTA coated tubes. About 3 ml of ascites fluid was obtained from each mouse. The P388 cells were then separated from lymphocytes and RBC by Ficoll Paque density gradient centrifugation technique. The interface was collected into the RPMI 1640 medium containing 10% FBS, and cells washed three times with the same medium containing 5% FBS. The cell pellet was then resuspended in 5 ml RPMI 1640 medium and 82 transferred into a T75 flask with 30 ml medium. The cells were incubated for 2 h, to allow the monocyte macrophages to adhere to plastic and the cells in suspension were transferred into a T175 flask containing 100 ml medium giving a cell concentration of approximately 1 x 10s cells/ml. P388 cells (both WT and ADR cells) were used for experiments within 48 h. In order to characterize the phenotypic and functional resistance of the P388 and P388/ADR cells as well as MCF7 and MCF7/ADR cells, PGP expression was routinely determined by treating 1 x 106 cells from each culture with the antibody, P-glyco CHEK219, using flow cytometry. The C219 monoclonal antibody is specific for PGP. A mouse IgG2a isotype-matched negative control antibody was included in each test. Specimens were fixed in 3.7% formaldehyde for 10 min at 4°C. Cells were washed 2-3 times in PBS, centrifuged and resuspended in 0.4 ml PBS containing blocking agent (1% bovine serum albumin). Aliquots of 0.2 ml were placed in two tubes. Ten microlitres of C219 antibody was added to one and 10 pl of negative isotypic antibody (Pharmingen, San Diego, CA) to the other. The mixture was incubated for 60 min at 4°C. Following 2-3 washes in PBS, 20 pl of 1:10 dilution of FITC-labeled goat-anti-mouse IgG-antiserum (Jackson, West Grove, PA, USA) was added and incubated for 30 min at 4°C. After washing cells twice in PBS, cells were centrifuged and resuspended in PBS and analyzed by flow cytometry. P388/ADR cell aliquots exhibiting >90% positivity (with respect to control sensitive cells) of phenotypic PGP-170 MDR were used in these studies. 83 2.2.5. Cytotoxicity Experiments The modified microculture tetrazolium (MTT) assay (Alley et al., 1988) was used in DOX-induced cytotoxicity experiments to screen the MDR modulators. This assay is based on the principle that the active mitochondria, present in living cells, can convert MTT to a colored formazan product which, when solubilized in DMSO, can be spectrophotometrically analyzed at a wavelength of 570 nm. Consequently, the number of viable cells per well is directly proportional to the production of formazan. Cells harvested from cultures in exponential growth phase were first counted by trypan blue exclusion which is based on the principle that the dye is specifically taken up by dead cells, thus allowing % viability of cell preparations to be determined. Specifically, 50 pi of cells was mixed with 50 u.1 of trypan blue. After mixing well, cells were counted using a hemocytometer. Cell preparations demonstrating viability >90% were used for cytotoxicity experiments. The optimal number of cells per culture plate well for a 72 h cytotoxicity study was determined by plating cells ranging from 500 - 15,000 and evaluating DOX cytotoxicity, cell growth (confluence) and microscopic observations of cell morphology. Using this procedure, the number of cells per well was determined as 2000 for P388AVT, P388/ADR, and MCF7AVT cells, and 5000 for MCF7/ADR cells. Cells were then dispensed within 96-well flat-bottomed Costar* (Cambridge, MA, USA) culture plates (cells/100 |il/well for a 3-day incubation). In-plate controls included non-drug treated cell-free 200 pi culture medium (for subtracting background absorbance) as well as non-drug treated cells, of equivalent cell density as test cells, in 200 ul medium (to calculate % viability). Test cells were exposed to serial concentrations of DOX (0.001 -12.5 uM; m 100 ul of culture medium, RPMI 1640 supplemented with 10% heat-84 inactivated FBS) in the presence or absence of modulators either in pulsed (cells are washed free of the modulator) or continuous exposure, over a 3-day incubation at 37°C, 5% CO2 and 100% relative humidity. Pulsed conditions were performed as follows. Modulator pre-treated cells were plated in 96-well culture plates and incubated for specified time intervals. At the end of this time period, the plate was centrifuged at 1800 rpm for 15 min and the upper layer aspirated using a blunt needle leaving cellular mass. Care was taken to keep cell-loss to a minimum. Two hundred microlitres of RPMI 1640 medium was then added, the well contents mixed, and the plate was centrifuged again at 1800 rpm for 15 min. Following aspiration using a blunt needle, 200 ul of drug-free or DOX containing medium was added and the plate incubated for the remainder of the 3-day incubation period at 37°C, 5% CO2 and 100% relative humidity. At the end of the incubation, 50 ul MTT (5 mg {3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl tetrazolium bromide}/ml PBS, filtered through 0.45 um filter units, and stored at 4°C for not longer than 1 month) was added and further incubated for 4 hours at 37°C. Following that, plates were centrifuged for the suspension type P388 cells, contents aspirated slowly through a blunt 18-gauge needle leaving 20 ul of the supernatant, and the reaction product thoroughly solubilised with 150 ul DMSO. The absorbance in each well was measured spectrophotometrically at 570 nm in a Titertek Multiscan multiplate reader (Flow Laboratories, Mississauga, Canada). Cytotoxicity was expressed in terms of percentage of control (drug-free cell cultures) absorbance (mean ± standard deviation) following subtraction of background absorbance. The IC50, defined as that concentration of drug which decreases absorbance to 50% of that in control (drug-free) cultures, was determined by plotting percent control absorbance on y-axis vs log drug concentration on 85 x-axis. Extrapolating the concentration at which 50% of control absorbance was observed, provided the IC50. IC50 values were determined as mean ± standard deviation (n = 9). The fold reversal and residual resistance factors were calculated using these mean IC50 values and statistical significance between groups was determined as described below. Two criteria were used to evaluate cytotoxicity data: [1] Fold Reversal (FR) = I = IC50 (POX - Modulator) P388/ADR or MCF7/ADR IC50 (DOX + Modulator) P388/ADR or MCF7/ADR [2] Residual Resistance Factor (RRF) = = IC50 (DOX + Modulator) P388/ADR or MCF7/ADR IC50 (DOX - Modulator) P388AVT or MCF7AVT Fold-reversal reflects the extent to which the chemosensitizer reverses resistance in the MDR cell line relative to unmodulated tumor cells, while the RRF is relative to the sensitive or wild type parental (WT) control cell line from which the resistant line was derived. Statistical analysis was performed using one way ANOVA (post-hoc comparison of means, Scheffe test) using Statistica for Windows, Release 4.5 (Statsoft, Inc.). A criterion of p<0.05 was used to determine significance. 2.2.6. Cellular DOX Uptake & Retention Studies 2.2.6.1. Flow Cytometry Studies Flow cytometric studies were conducted to dynamically determine DOX uptake by individual cells (confirmation of cytotoxicity observations relating to PGP blockade), in a Coulter EPICS Elite ESP flow cytometer (Coulter Electronics, Hialeah, FL, USA). The Uniphase 10 mw HeNe laser and Cyonics 15 mw argon laser, are mounted on an optical 86 table. An Arnold type SortSense flow cell with 76-pm orifice and 250-pm square channel was used. Detectors used were a diode forward angle light scatter, 90° light scatter PMT, and three fluorescence PMTs. 12 mm x 75 mm sample containing tubes were held by a vortex-capable Q-PREP-type fixture. Data were collected and processed on an Elite workstation. The MCF7 (adherent cells) cells were used for studies following trypsinization of cultures. These cells were vigorously vortexed before and during flow cytometric analysis. Briefly, modulator treated cells (1 x 106 cells/ml; viability >90%) were incubated with 10 uM DOX (t=0 min) and aliquots of the incubated samples at intervals of 10-20 min subjected to flow cytometric studies (flow rate: 500 cells/sec). Forward and side scatter were used to gate out debris, dead cells, and doublets. DOX fluorescence emission was filtered through a 550 nm long pass fdter. The mean fluorescence intensity associated with DOX within cells was plotted with time. For release studies, 1 x 106 cells/ml cells were incubated with 10 pM DOX at 37°C for 90 min in the presence and absence of the modulator. These cells were then washed free of DOX using PBS and evaluated for cellular DOX fluorescence. 2.2.6.2. Cellular Doxorubicin Concentrations The cell-associated DOX concentrations, in the presence and absence of the MDR modulator, were determined for P388AVT and P388/ADR cells in order to directly correlate DOX fluorescence in flow cytometry experiments with intracellular DOX concentrations, thus correcting for potential changes in fluorescence due to intracellular microenvironments. Tumor cells (1.5 x 106 cells/ml) were incubated with DOX (10 pM) and MDR modulators under standard culture conditions. At various times (30, 120, and 87 260 min), 2 ml samples were removed, cells were counted using a Coulter Counter and non-cell associated modulators and cytotoxic agents were removed by centrifuging (1000 rpm, 5 min at 4°C), washing cells once with drug free cold RPMI-1640 medium (no FBS) at 4 °C, and harvesting the cells by centrifugation. The resulting cell pellet was frozen until DOX assay. DOX quantitation was carried out by thawing cell pellets with water, followed by addition of 10% SDS and 10 mM H2SO4. The drug was extracted with a 1:1 mixture of isopropanol-chloroform. Protein aggregation was facilitated by freezing the samples at -80°C (1 h), thawing the samples, and centrifuging at 3000 rpm. The organic layer containing DOX was analyzed for fluorescence at 550 nm in a Perkin-Elmer F150 spectrofluorimeter (Perkin-Elmer), and concentrations were determined using calibration curves generated with similarly treated DOX standards. Statistical analyses were performed using 2-sample t-test and statistical significance was set at p<0.05. 2.2.7. Intracellular Drug Distribution Tumor cell lines exposed to DOX and MDR modulators were evaluated by fluorescence microscopy in order to determine intracellular drug distribution characteristics and compare the properties for sensitive and resistant tumor cells. This included the assessment of nuclear, cytoplasmic or punctated cytoplasmic DOX fluorescence in cells treated with MDR modulators. P388/ADR cells (5 x 106) were incubated with the indicated concentrations of modulators for 2 hours, after which cells were seeded onto glass slides mounted in 50% glycerol in PBS, and assessed for DOX fluorescence using a Leitz Wetzlar Dialux model (Germany) fluorescence microscope. 88 2.3. Results 2.3.1. Cytotoxicity experiments 2.3.1.1. Continuous exposure DOX-induced cytotoxicity against PGP positive P388/ADR and MCF7/ADR cells in the presence of MDR modulators was evaluated using the MTT assay (Table 2.1). The modulators screened included the extensively characterized VRP (10 pM) as a standard for comparison. The concentrations for the various modulators were selected based on preliminary observations using concentration-activity titrations to determine the concentration at which maximal reversal was obtained as well as literature reports identifying the concentration as that required to achieve maximal activity and any further increase in concentration either resulted in no further increase in activity or was associated with enhanced toxicity (Alaoui-Jamali et al, 1993; de Jong et al., 1995; Keller et al., 1992b; Hofmann et al, 1995; Fan et al, 1994). In the 72-h continuous exposure, all the modulators tested reversed MDR to varying degrees. Three compounds provided near complete reversal in MCF7/ADR and P388/ADR cell lines. These were PSC 833 (1 pM), dexniguldipine (DNG, 5 pM) and Rol 1-2933 (10 pM). Specifically, PSC 833 provided a 75-fold reversal in P388/ADR cells and 263-fold reversal in MCF7/ADR cells. Rol 1-2933 induced a 95-fold and 516-fold reversal in P388/ADR and MCF7/ADR cells, respectively. DNG reversed MDR by 1952-fold in MCF7/ADR cells, however, at 5 uM was observed to be cytotoxic to murine P388 cells over the 72 h period resulting in 30% cell viability. At concentrations of 1, 1.25, and 2.5 pM, DNG did not provide significant modulation in the P388/ADR cell line. Over the same exposure time, DNG was non-toxic to MCF7 cells and afforded complete reversal at a concentration of 5 pM. In fact, the MCF7/ADR line was 89 approximately 5-fold more sensitive to DOX than wild type MCF7 cells in the presence of 5 pM DNG (Table 2.1). Both PSC 833 and Rol 1-2933 provided >95% reversal of MDR with a residual resistance factor (RRF, relative to WT cells with a value of 1.0 indicating complete reversal) less than 2.5 in P388 cells and 1.0 in MCF7 cells (difference relative to sensitive WT control not being statistically significant). In comparison, VRP (10 pM) provided approximately 50% MDR reversal with a RRF of 7.22 (p<0.05) in P388 cells and 132.8 in MCF7 cells (Table 2.1). Other chemosensitizers were either comparable to VRP in modulation activity (trans-E-flupenthixol, dexVRP) or exhibited inferior reversal characteristics (prochlorperazine, quinidine, chorpromazine, and tamoxifen). These agents were unable to fully sensitize MDR cells as evidenced by RRF values greater than 5 (Table 2.1). PSC 833 was the most potent MDR modulator with maximum chemosensitization achieved at 1 pM whereas the other modulators required concentrations of 5 pM or more to obtain maximal activity. 90 Table 2.1: Summary of fold reversal, and modulator concentrations in P388/ADR and MCF7/ADR cells in the 72 h continuous exposure 1 Modulator Cone. (uM) P388 ADR MCF7 ADR Fold Reversal Residual Resistance Factor Fold Reversal Residual Resistance Factor VRP 10 23.6 7.2* 1.417 132.8* DNG 5 312.8* 0.582 1952 0.194 2.5 1.741 110.1* ND ND PSC 833 1 75.4 2.259 263.4 0.714 Rol 1-2933 10 95.4 1.787 516.6 0.734 rraws-flupenthixol 10 20.1 12.9* ND ND Prochlorperazine 10 3.156 85.4* 1.221 258.3* Quinidine 10 2.947 91.4* 1.459 216.1* Chlorpromazine 10 1.707 157.8* 1.116 282.5* Tamoxifen 10 1.613 167.1* ND ND DexVRP 10 5.264 7.7* ND ND # 30% viability 1: P388/ADR or MCF7/ADR cells were preincubated for 30 min with the chemosensitizer and plated in 96-well culture plates. Serial DOX concentrations were added and incubated for 72h, after which the MTT assay was performed as described in Materials & Methods. IC5oS, FR and RRF values were calculated using the formulae described in Materials and Methods. NB: FR and RRF values were calculated using mean IC50 values (based on n = 9). In order to determine statistical significance, statistics were performed using IC50 values corresponding to the respective FR or RRF value. * difference (IC50) relative to sensitive (WT) control is significant at p<0.05 91 2.3.1.2. Pulsed exposures Pulsed exposures were performed with modulators that exhibited effective reversal of MDR in cytotoxicity experiments to determine if these agents possess latent activity. The P388/ADR and MCF7/ADR cells were exposed to the modulator in the absence of DOX for 2, 8, or 24 hours after which cells were washed free of the modulator and DOX was added for the remainder of the 72 h incubation period. PSC 833 (1 uM) exposed for 72 h lowered the IC50 of DOX in P388/ADR cells to that of the sensitive P388 cell line (75.4-fold reversal). In comparison, pulsed PSC 833 exposures for 2 and 8 hours reversed MDR by 3.9- and 77-fold, respectively. It is important to note that PSC 833 is removed from the incubation mixture after the pulsed exposure and DOX is retained in the media for the remainder of the 72 h assay. This suggests that PSC 833 can completely sensitize resistant P388 cells to DOX over extended times if therapeutic levels can be maintained for 8 h. In sharp contrast, the modulating activities of VRP and Rol 1-2933 are substantially lost when the modulator is removed from the cells even for 8 h pulsed exposures where maximum reversal ranged from 1- to 5- fold (Table 2.2). Since DNG was itself somewhat cytotoxic to P388 cells at concentrations required for MDR reversal for the 72 h exposure time period, it was not evaluated in this model. In the MCF7/ADR cell line, however, PSC 833 did not provide complete reversal at an incubation time of 8 h (Table 2.3). When the incubation time was extended to 24h, PSC 833 afforded complete reversal with a FR value of 282 and a RRF of 1.0. Similar latency was observed for DNG and Rol 1-2933 in the MCF7/ADR line (Table 2.3). In marked contrast, the activity of VRP was completely lost when pulsed for 24 h. It should be noted that similar results were obtained when the 92 pulsed exposure period for MDR modulators included DOX in the incubation mixture in addition to the 72 h DOX exposure after the modulators had been removed. Table 2.2: Cytotoxicity profiles in continuous and pulsed exposures for PSC 833, VRP, and Rol 1-2933 in P388/ADR cells1. Fold Reversal (Residual Resistance Factor) Condition PSC 833 [1 uMl VRP [10 uMl Rol 1-2933 [10 uMl Continuous 75.4 (2.259) 23.6(7.2*) 95.4(1.781) Pulsed: [1] 2h 3.851 (44.3*) 0.918 (185.6*) 4.181 (40.8*) [2] 8h 77.0 (2.213) 1.719(99.2*) 1.485 (303.4*) Numbers in parentheses represent residual resistance factor (RRF). 1: In the continuous exposure, P388/ADR cells were preincubated for 30 min with the modulator and plated in 96-well culture plates. Serial DOX concentrations were added and incubated for 72 h after which the MTT assay was performed as described in the section on Materials & Methods. In pulsed conditions, P388/ADR cells were incubated with modulator in the absence of DOX for 2 or 8 hours, after which the cells were washed free of the modulator, DOX added and incubated until the remainder of the 72 h incubation period. IC50, FR and RRF values were calculated using the formulae described in Materials and Methods. NB: FR and RRF values were calculated using mean IC50 values (based onn = 9). In order to determine statistical significance, statistics were performed using IC50 values corresponding to the respective FR or RRF value. * difference (IC50) relative to sensitive (WT) control is significant at p<0.05 93 Table 2.3: Cytotoxicity profiles in continuous and pulsed exposure for PSC 833, VRP, DNG, and Rol 1-2933 in MCF7/ADR cells1. Fold Reversal (RRF)a Condition PSC 833 VRP Rol 1-2933 DNG Continuous 263.4 (0.714) 1.417(132.8*) 516.6(0.734) 1952 (0.194) Pulsed: [1] 2h 2.069 (90.9*) 0.801 (234.8*) 1.887 (201.0*) 6.53 (58.1*) [2] 8h 2.484 (75.8*) 0.798 (235.8*) 4.398 (86.2*) 514.4 (0.737) [3] 24h 282.2(1.056) 1.365 (218.4*) 629.5 (0.474) 1539(0.194) a: Numbers in parentheses represent residual resistance factor (RRF). 1: In the continuous exposure mode, MCF7/ADR cells were preincubated for 30 min with the modulator and plated in 96-well culture plates. Serial DOX concentrations were added and incubated for 72 h after which the MTT assay was performed as described in the section on Materials & Methods. In pulsed conditions, MCF7/ADR cells were incubated with modulator in the absence of DOX for 2, 8 or 24 hours, after which the cells were washed free of the modulator, DOX added and incubated until the remainder of the 72 h incubation period. IC50, FR and RRF values were calculated using the formulae described in Materials and Methods. NB: FR and RRF values were calculated using mean IC50 values (based on n = 9). In order to determine statistical significance, statistics were performed using IC50 values corresponding to the respective FR or RRF value. * difference (IC50) relative to sensitive (WT) control is significant at p<0.05 94 2.3.2. DOX Uptake & Release Kinetics In order to correlate the enhanced cell kill observed in cytotoxicity experiments with increased DOX influx as a result of PGP blockade, chemosensitized cell samples were subjected to flow cytometry studies, which provided a dynamic measurement of DOX uptake by individual live cells. The conditions utilized in this technique provide specificity for anthracycline fluorescence, enabling scattered and fluorescent light from individual cells to be quantified at rapid rates thereby allowing dead cells and debris to be gated out. The peaks resulting from the plot of side scatter vs DOX fluorescence were found to be single, sharp and homogeneously constructed under all conditions studied here (Figure 2.1). Consequently, increases in cell associated DOX fluorescence reflected unimodal averages of DOX content in cell populations rather than averages arising from multiple cell populations exhibiting distinctly different DOX content. Multidrug resistance modulators from a wide range of chemical classes assessed in cytotoxicity studies were evaluated for DOX uptake characteristics in comparison to VRP which was employed as an internal standard. DOX uptake in sensitive, resistant, and VRP chemosensitized cells was found to be reproducible on a inter-week basis, allowing reliable and reproducible comparisons to be made. Modulators screened included tamoxifen (10 pM), quinidine (10 uM), Rol 1-2933 (10 pM), prochlorperazine (10 pM), PSC 833 (1 pM), DNG (5 pM), trans-E-flupenthixol (10 pM), dexVRP (10 pM) and chlorpromazine (10 pM). Selection of the concentrations used was based on those required for maximal chemosensitization in cytotoxicity experiments described above. 95 Figure 2.1: Representative plots of counts vs fluorescence for DOX uptake in P388/ADR cells (Panel A), P388AVT cells (Panel B), and PSC 833 sensitized P388/ADR cells (Panel C), as evaluated by flow cytometry. The resulting peaks are single, sharp, and homogeneously constructed. 96 As shown in Figure 2.2 (A and B), DOX uptake is enhanced by varying magnitudes in chemosensitized P388/ADR cells, the extent dependent upon the type of modulator used. DOX uptake in P388AVT cells was 5-fold higher than that in P388/ADR cells (on a relative DOX fluorescence scale). The rate of DOX uptake in P388/ADR cells was one-tenth of that observed in P388AVT cells, the slope being 0.0042 (5-60 min.). In the first hour (5-60 min.), DOX uptake rates in sensitive P388 cells (slope = 0.0545) and P388/ADR cells chemosensitized with Rol 1-2933 (slope=0.0542), PSC 833 (slope=0.0657), and DNG (slope=0.0516) were similar, the coefficient of variation being <10% (slopes were calculated from a plot of relative DOX fluorescence vs time). The uptake of DOX in these chemosensitized cells was not significantly different from that in sensitive P388AVT cells, indicating complete PGP blockade by these agents. VRP (slope=0.0479) and dexVRP treated cells demonstrated modest DOX uptake (approximately 50-60% of WT control), while transflupenthixol, prochlorperazine, chlorpromazine, tamoxifen and quinidine sensitized cells exhibited <25% maximal DOX uptake compared to that for sensitive P388 cells (Figure 2.2 A and B). DOX uptake in P388/ADR cells sensitized by these agents followed the order PSC 833 (greatest) > DNG « Rol 1-2933 > VRP « dexVRP > trans-E-flupenthixol > prochlorperazine « chlorpromazine « tamoxifen « quinidine (least). Similar DOX uptake modulation effects, both in terms of extent of DOX uptake modulation and rates were observed for MCF7/ADR cells in comparison to sensitive wild type MCF7 cells (Figure 2.3). 97 Figure 2.2 (A and B): Flow cytometric evaluation of DOX uptake in chemosensitized P388/ADR cells. The uptake of DOX was evaluated in P388/ADR cells pretreated with PSC 833 (1 pM), DNG (5 pM), Rol 1-2933 (10 pM), Ro44-5912 (10 pM), tamoxifen (10 pM), quinidine (10 pM), dexVRP (10 pM), prochlorperazine (10 uM), trans-E-flupenthixol (10 pM), and chlorpromazine (10 pM) as described under Materials and Methods. Data are presented as mean ± standard deviation (n = 4). Legends for Panel A: P388/ADR control (•); tamoxifen (•); quinidine (A) ; Rol 1-2933 (T) ; VRP (X) ; dexVRP (•); P388AVT control (*). Legends for Panel B: P388/ADR control (•); DNG (*); prochlorperazine (•); PSC 833 (A) ; trans-E-flupenthixol (T); chlorpromazine (•); VRP (X) . 98 0.0- | 1 1 i 1 i 1 1 — — i 1 1 0 20 40 60 80 100 Time (Min) Figure 2.3: Flow cytometric evaluation of DOX uptake in chemosensitized MCF7/ADR cells. The uptake of DOX was evaluated in MCF7/ADR cells pretreated with Rol 1-2933 (10 uM, •) , DNG (5 uM, A), PSC 833 (1 uM, T) , and VRP (10 uM, • ) as described under Materials and Methods in relation to DOX controls in the resistant (MCF7/ADR control (•)) and sensitive lines (MCF7AVT control (+)). Data are presented as mean ± standard deviation (n = 4). 99 Next, the cellular DOX release profile in chemosensitized P388 MDR and MCF7 MDR cells that were pulsed with MDR modulators were determined. The chemosensitizers used for these studies included PSC 833, DNG, and VRP in order to determine whether modulation latency observed in cytotoxicity experiments correlated with the ability to achieve latent DOX retention in MDR cells. In the release studies, cells incubated with the modulator and DOX for 90 minutes were washed free of the modulator, exposed to media containing DOX only and monitored for DOX fluorescence at 10 min intervals by flow cytometry. Release of DOX from cells was reflected by a decrease in the cell associated fluorescence. VRP treated cells exhibited first order DOX elimination kinetics with first order release rate constants (slope of the % release vs time plot) of 2.95 % release/min for P388/ADR (t1/2=10 min) and 2.84 % release/min for MCF7/ADR cells (t 1/2= 10 min) upon removal of modulator. In contrast, PSC 833 and DNG provided extended DOX retention properties in P388/ADR cells, with 50% DOX released at 25 and 40 min (Figure 2.4A), respectively. The time required for 50% DOX release in MCF7/ADR cells for these two modulators was 20 min and >65 min respectively (data not shown). These DOX release kinetics reflected release rate constants of 2.09 % release/min and 1.41 % release/min, respectively in the P388/ADR cells and 3.86 % release/min and 0.0 % release/min in the MCF7/ADR cells over the first 30 min. When the P388 MDR cells were pulsed with PSC 833 for extended times (15 hours) in the presence or absence of DOX, there was significant latency in its ability to maintain DOX retention as shown in Figure 2.4B. 100 0 10 20 30 40 50 60 Time (Min) Figure 2.4: DOX release profile in multidrug resistant P388/ADR cells Panel A: Effect of PSC 833 (•, 1 uM), DNG (•, 5 uM) and VRP (•, 10 uM) on DOX retention properties. Cells were incubated with the modulator for 30 min after which DOX was added and further incubated for 90 min. The cells were then washed free of the modulator and DOX release was studied over 70 min using flow cytometry. Release curves are normalized to the P388/ADR resistant modulator-free DOX controls. Data are presented as mean + standard deviation (n = 4). Panel B: Effect of extended pulse exposure times on DOX retention properties. Cells were exposed for 15 hours to MDR modulators in the presence (PSC 833 + DOX [€>] or VRP + DOX [•]) or the absence (PSC 833 [A] or VRP [+]) of DOX. These cells were washed free of the modulator and DOX release was studied over a 60 min period using flow cytometry. Release curves are normalized to the P388/ADR resistant modulator-free DOX controls. Data are presented as mean ± standard deviation (n = 4). 101 Specifically, the time to achieve 50% drug release was >60 min for cells pulsed with PSC 833 both in the presence and absence of DOX. In contrast, VRP exposed cells exhibited time for 50% release on the order of 5 min in the absence of DOX and 10 min in the presence of the anticancer drug. These DOX retention results correlated well with cytotoxicity data regarding the latency of modulating activity by PSC 833 and DNG. Figure 2.5 presents the quantitative DOX intracellular levels in P388/ADR cells. Cells (controls and modulator treated) were incubated with DOX after which they were extracted and quantitated by fluorescence spectrophotometry in order to ensure that fluorescence intensity measurements made by flow cytometry reflected actual changes in DOX uptake and not changes in DOX fluorescence efficiency caused by altered cellular localization. DOX levels in P388/WT cells (1223.6 ± 15.62 ng/106 cells) were significantly (2.6-fold, p<0.05) higher than P388/ADR cells (478.26 + 4.42 ng/106 cells) in the absence of the MDR modulators. In comparison, DOX levels in PSC 833 sensitized cells were 2 fold higher than VRP sensitized cells (significantly different at p<0.05), and were not significantly different from those in P388/WT cells (Figure 2.5). These quantitative DOX intracellular levels as well as the flow cytometry results compared favorably to the extent of MDR reversal observed in the cytotoxicity experiments. 102 2000 J L 1500 4 "3 o CO < 1000 o O Q C 500-I ti. i 8 co Q. O w CL + cc o 8 <o 0 . Q. > + tc Q < s co CL Figure 2.5: Intracellular DOX levels in P388/ADR, P388/WT, and chemosensitized cells. DOX (10 uM) was incubated in the presence and absence of MDR modulator (PSC 833, PSC, 1 uM or VRP, 10 uM) for 260 min, following which cell associated DOX was quantified using spectrofluorimetry as described under Materials and Methods. Data are presented as mean ± standard deviation (n = 6). 103 2.3.3. Intracellular DOX distribution Intracellular DOX distribution characteristics were assessed to determine whether increased cellular DOX levels induced by MDR modulators exhibited cellular localization properties comparable to those displayed by the drug sensitive parental tumor cell lines. To accomplish this, P388 cells (WT and ADR) were exposed to DOX for 2 h in the presence and absence of modulator and were evaluated by fluorescence microscopy. DOX fluorescence in P388/ADR cells was negligible in the absence of MDR modulator whereas intense cell-associated fluorescence was observed in drug sensitive P388 cells (Figure 2.6). P388/ADR cells exposed to DOX in the presence of 10 pM VRP did exhibit increased intracellular fluorescence compared to P388/ADR cells exposed to DOX alone. However, the intensity of DOX fluorescence was significantly less than that obtained in the sensitive cell line. In comparison, P388/ADR cells exposed to DOX in the presence of 1 pM PSC 833 demonstrated similar fluorescence intensity and distribution properties as sensitive P388AVT cells. Although qualitative, these results corroborate the relationships between total cell DOX uptake and chemosensitization described above. 104 Fluoresence Microscopy (1) P388/wt + DOX (2) P388/adr + DOX (3) 1 uM PSC 833 (4) 10 uM Verapamil Figure 2.6: Representative DOX Fluorescence micrographs of P388/ADR Control, P388/WT Control, and modulator (VRP or PSC 833) pre-treated P388/ADR cells. 105 2.4. Discussion The circumvention of MDR using chemosensitizers has received considerable attention over recent years, resulting in a number of compounds being identified that modulate MDR in vitro to varying degrees in a variety of experimental MDR cell lines (Hyafil et al., 1993; Dantzig et al., 1996; Perez et al., 1993; Alaoui-Jamali et al., 1993; Gekeler et al., 1995a; Fan et al., 1994; Hofmann et al., 1991; Hofmann et al., 1995; Jiang et al., 1995; Keller et al., 1992b; Ramu et al., 1984a,b; Tsuruo et al., 1982a,b; Tsuruo et al., 1983a,b,c; Tsuruo et al., 1984). However, whereas significant MDR reversal can be readily achieved in vitro, preclinical and clinical chemosensitization studies have provided far less convincing results (Fan et al., 1994; Goldstein, 1995; Boote et al., 1996; Sarris et al., 1996; Sikic et al., 1994). The reasons for this discrepancy between in vitro and in vivo studies could be several: (1) lack of therapeutic plasma concentrations of these agents being achieved and/or maintained for complete chemosensitization to occur, (2) poor drug and/or modulator delivery to the tumor site of action, or (3) the presence of residual resistance due to incomplete chemosensitization or multifactorial resistance mechanisms. One of the features limiting the resolution of these issues is the poor understanding of the pharmacological properties of chemosensitizers that affect MDR modulation. In addition to evaluations typically associated with the in vitro characterization of MDR modulators (fold reversal and potency), the studies described here addressed additional cell pharmacological elements that are likely to be key to successful in vivo MDR modulation, namely residual resistance and modulation latency. Comparison of these characteristics between various modulators provides additional 106 insight into the factors that dictate chemosensitization which can be used to optimize the modulation of MDR in vivo. In these studies, VRP, a representative first generation MDR modulator, was used as a standard for comparing compounds that modulate MDR. Early studies demonstrated that vincristine resistance was completely reversed when P388/VCR cells were chemosensitized with 6.6 uM VRP (Tsuruo et al., 1981). VRP (6.6 uM) lowered the IC50 of vincristine in P388/VCR cells from 26 nM to 0.31 nM with an 84-fold reversal factor. In comparison, VRP had a fold reversal factor of 13 for DOX in P388/ADR cells (Tsuruo et al., 1982a,b). In this study, VRP afforded a moderate 23.6-fold reversal factor for DOX in P388/ADR cells for continuous 72 h exposure. This is consistent with the results of Tsuruo et al., 1982 (Tsuruo et al., 1982a,b). However, reversal of MDR by VRP in P388/ADR cells was shown here to be associated with a residual resistance factor (RRF) of 7.2, indicating incomplete chemosensitization. The RRF value has significant implications on in vivo MDR therapy since dose-limiting toxicities associated with anticancer drug-chemosensitizer co-administration typically do not allow for increasing the dose of the chemotherapeutic agent to compensate for RRF values in this range. In view of the 7.2 RRF it can be predicted that VRP would be relatively ineffective as an MDR modulator in vivo. This is true also for other MDR modulators that exhibit high RRF values in this study including dexVRP, tamoxifen, quinidine, chlorpromazine, and prochlorperazine. Reports of a lack of substantial therapy with VRP and these other modulators in preclinical and clinical studies support this argument (Fan et al., 1994; Goldstein, 1995; Sikic et al., 1994). In contrast, modulators such as PSC 833, Rol l -2933, and DNG which exhibit small RRF values in the murine and human MDR cell lines 107 utilized here have demonstrated significant MDR reversal in vivo in preclinical tumor models (Alaoui-Jamali et al., 1993; Van de Vrie et al., 1996; Keller et al., 1992b; Watanabe et al., 1995; Watanabe et al., 1996a). Of these three compounds, PSC 833 was observed to be the most potent chemosensitizer affording complete reversal of MDR at concentrations 5- to 10-fold lower than required for other reversing agents. However, the small residual resistance factors observed for these more effective compounds in the P388/ADR line may explain why their in vivo therapy, although improved over VRP and other common modulators, often does not achieve full chemosensitization equivalent to the WT sensitive cells. Latent MDR modulating activity was observed to vary extensively among modulators examined here. In this study it was observed that PSC 833 and DNG, but neither VRP nor the tiapamil derivative Rol 1-2933 were capable of inducing MDR reversal long after they are removed from the extracellular medium. It is important to note that the extent of MDR modulation did not correlate with the degree of latency. In particular, Rol 1-2933 was able to provide full reversal of MDR under continuous exposure conditions, however it exhibited negligible latent activity in the P388/ADR tumor line. These results are consistent with preliminary indications of variability in the duration of activity between MDR modulators (Boesch and Loor, 1994; Archinal-Mattheis et al., 1995). Further, the studies here indicate that latent MDR modulators such as PSC 833 and DNG can provide full chemosensitization by pre-exposure in the absence of anticancer drug (Table 2.2). Such properties may be useful to increase the selectivity of modulator activity in tumor cells versus healthy tissue for in vivo applications. While both PSC 833 and DNG are very hydrophobic, large molecular weight compounds, VRP and the tiapamil 108 derivative are amphipathic with significant hydrophilic nature. It can be speculated that PSC 833 and DNG, via hydrophobic interactions confer sustained inhibition of the PGP pump due to their increased affinity for membranes. Interestingly Rol 1-2933, a hydrophilic compound, exhibited latent modulating activity in MCF7/ADR cells but not P388/ADR cells. This information as well as recent evidence (Mayer and Hartley, 1998) that high concentrations of VRP (>50 uM) create residual cell associated VRP pools for extended periods, suggest that further investigations into these biochemical and biophysical phenomena may help to identify pharmacological properties that can be used to a therapeutic advantage in vivo. In conclusion, the extent of MDR reversal (using fold reversal and residual resistance factor), potency, and latent modulating activity of MDR modulators in two independent MDR in vitro tumor models have been characterized. The empirical cytotoxicity results correlated well with DOX uptake and release characteristics in the presence and absence of MDR modulator indicating that the primary underlying mechanism for the increased DOX cytotoxicity brought about by MDR modulators in both continuous and latent settings is presumably due to inhibition of PGP drug efflux. Evaluating such parameters in a screening program will help to identify new modulators with a greater potential for successful in vivo applications. In particular, properties such as residual resistance and modulation latency combined with characteristics traditionally associated with MDR reversing agent activity profiles (i.e., potency, fold-reversal and specificity) may distinguish between modulators that are limited to ex vivo applications from those with the versatility required for systemic in vivo use. In this context, it was observed in this study that PSC 833 at a concentration of 1 uM is a potent, non-toxic modulator of MDR which 109 chemosensitizes PGP-based MDR tumors to levels approaching parental, drug sensitive cell lines. However, further therapeutic improvements may be possible with new MDR modulators in view of the small residual resistance that remains after treatment with PSC 833. An interesting feature of PSC 833 modulation is its latent modulating activity which makes it a particularly attractive candidate for in vivo studies by widening the therapeutic window necessary for chemosensitization dose and timing relationships. In view of the results described here, PSC 833 was selected as the MDR modulator for application in the in vivo studies described in the following Chapters. 110 Chapter 3 TOXICITY, ANTI-TUMOR ACTIVITY, AND PHARMACOKINETICS OF NON-ENCAPSULATED AND DSPC/CHOL LIPOSOME ENCAPSULATED DOX, IN THE PRESENCE AND ABSENCE OF PSC 833 USING THE MURINE P388/ADR SOLID TUMOR MODEL 3.1. Introduction Efforts to overcome MDR using chemosensitization have provided promising results in cell culture systems but have met with limited success in preclinical and clinical studies. First generation modulators (VRP, CsA) exhibit inherent pharmacological activity and were often very toxic at doses required to modulate MDR (Ozols et al., 1987; Fan et al., 1994). Recent investigations have led to more potent and less toxic modulators of MDR such as PSC 833, a non-immunosuppressive analog of cyclosporin, which, based on results of the studies presented in Chapter 2, has been selected for studies presented here. The following sections will discuss the cellular pharmacological properties, mechanism of action, toxicity, drug-drug interactions and pharmacokinetics of the MDR modulator, PSC 833. PSC 833 is a non-immunosuppressive analog of Cyclosporin A (see structures of PSC 833 and CsA in Figure 3.1.) which has been shown to exhibit 10-fold greater potency than CsA. At concentrations of 1 uM, it has been shown to modulate MDR nearly completely (Keller et al., 1992b; Watanabe et al., 1995; Watanabe et al., 1996a; Jachezet al., .1993; Krishna et al., 1997). Using a photoactive [3H]-azidopine and a photoactive cyclosporin analog, it was shown that PSC 833 more effectively blocked the binding of the photoactive compound to PGP compared with CsA (Archinal-Mattheis et al., 1995) and this resulted in prolonged inhibitory effects. This suggests hydrophobic binding interaction 111 between PSC 833 and PGP, a finding consistent with observations of latent modulating activity shown in Chapter 2 (Krishna et al., 1997; Boesch and Loor, 1994). In addition, functional studies have confirmed PGP blockade in various experimental cell lines, where increased intracellular drug accumulation was demonstrated (Watanabe et al., 1995; Boesch et al., 1991b; Michieli et al., 1994; Jiang et al., 1995; Krishna et al., 1997). The following sections will review the physico-chemical properties, toxicity, drug-drug interactions, and pharmacokinetic characteristics of PSC 833. R 2 — C H 2 C H , — C H -C H 3 C H 3 Figure 3.1.: Structures of cyclosporin A and its non-immunosuppressive analog, PSC 833. Note the substitutions at positions Rl and R2. 112 PSC 833 is generally free from any immunosuppressive, anti-inflammatory, or cytotoxic effects (Covelli, 1998). However, PSC 833 may adversely interact with co-administered anticancer drugs, via impacting anticancer drug biliary, urinary, and metabolic processes (Speeg and Maldonado, 1994). The former two pathways are mediated via PGP blockade in excretory organs while metabolic interactions may be via interference with the CYP3 A pathway. These interactions are particularly important since CYP3A is implied in the metabolism of PSC 833 and many anticancer agents. This combined with the fact that many drugs are substrates for both CYP3A and PGP, would indicate that PSC 833 may affect the elimination of anticancer drugs whose elimination processes are governed by the CYP3 A/PGP pathways. Due to the lack of an analytical chromatographic assay for PSC 833 until recently (Scott et al., 1997), little is known about its pharmacokinetics. Limited studies within Novartis (Covelli, 1998) have utilized radiolabeled PSC 833 to determine its disposition. In general, absorption of PSC 833 is slow with variable bioavailabilities in rats and dogs. Following oral administration in rats and dogs, mean terminal elimination half lives of PSC 833 are in the order of 20 h and 142 h respectively. In humans, PSC 833 is highly protein bound (98%) with the major protein component being lipoproteins. This hydrophobic molecule is also highly distributed into tissues. In both rats and dogs, the Vd is high (9 -10 L/kg) confirming extravascular distribution. It is extensively metabolized by CYP3A, the major metabolite being PSC M9 (hydroxylated at y position of L-methyl-leucine). This metabolite, although active, exhibits a 5-40-fold reduced MDR potency and is devoid of any other toxicities. In both rats and dogs, PSC 833 and metabolites are excreted slowly and mainly via bile into the feces (Covelli, 1998). 113 The above cellular pharmacological properties of PSC 833 coupled with results of studies presented in Chapter 2 makes its role as a model MDR modulator well suited for in vivo studies outlined in this Chapter. In spite of the developments of newer, more potent MDR modulators, significant difficulties remain in modulation strategies employing chemosensitization approaches (Colombo et al., 1996). A problem more confounding perhaps than intrinsic toxicity is the modulator-induced alterations in pharmacokinetics of co-administered conventional anticancer drugs. Such modulator-induced pharmacokinetic alterations significantly reduce the elimination of many anticancer drugs from the circulation resulting in a need to reduce the dose of the anticancer agents due to exacerbated toxicity (Erlichman et al., 1993; Linn et al., 1994; Colombo et al., 1994; Gonzalez et al., 1995; Sarris et al., 1996; Boote et al., 1996). It has been unclear whether such alterations may have a negative impact on therapeutic outcome, however such relationships are clearly possible. Liposomal carriers have been demonstrated to provide tumor specific delivery of anticancer agents as well as to circumvent many toxicities associated with these agents by altering the pharmacodistribution properties of encapsulated drugs (Gabizon and Papahadjopolous, 1988; Mayer et al., 1990a,b; Gabizon, 1992; Mayer et al., 1995a; Forssen et al., 1996). Given the pharmacokinetic alterations induced by PSC 833 on non-encapsulated (conventional) DOX, this thesis examined whether liposomes limit these effects by virtue of their ability to reduce the exposure of encapsulated DOX to the kidneys (Van Hossel et al., 1984) and alter clearance of DOX in the liver (Mayer et al., 1989; Parker et al., 1982b). These tissues appear to be key factors involved in PSC 833-induced DOX pharmacokinetic changes (Colombo et al., 1996, also see Sections 1.2.2.2.2. and 1.5 114 in Chapter 1). Small (100 nm) liposomes also preferentially accumulate in sites of tumor progression where they are able to extravasate through the more leaky tumor neovasculature (Gabizon and Papahadjopolous, 1988; Gabizon, 1992; Wu et al., 1993; Yuan et al., 1994). This feature, in conjunction with the potential avoidance of PSC 833-induced DOX pharmacokinetic alterations when liposomes are employed, may lead to substantially increased selectivity of PSC 833-mediated increased cellular DOX uptake for tumor cells and improve the anti-tumor activity of resistant tumors. In general, two types of in vivo tumor MDR models have been described in the literature, namely, ascites and solid tumors. Ascites (liquid tumor) models are those where PGP expressing tumor cells are injected i.p. (the drug is usually also administered intraperitoneally) to allow proliferation of tumor cell growth in the mouse peritoneum. Ascites models lack a physiologic barrier, and therefore the situation may be considered to be no different from in vitro cell culture systems. Solid tumor MDR models are those where PGP overexpressing cells are inoculated in extraperitoneal sites, such as s.c, in mice, generating solid tumors. These models are experimentally more stringent since both MDR modulators and anticancer drugs must gain access to the extravascular tumor site after systemic administration and therefore may be considered more relevant to humans than the ascites model. This thesis has therefore focused on the use of solid tumor MDR models in order to test the potential utility of liposomal drug delivery systems to improve the therapy of MDR tumors. P388/ADR cells, when inoculated s.c. in the flank of BDF1 mice generate well vascularized solid tumors, which exhibit reproducible growth rates in vivo, enabling tumor growth suppression assessments to be made. Although this model is not an epithelial solid 115 tumor model, the biological features exhibited closely simulate solid tumor physiology. Also, it must be mentioned here that the P388/ADR cell line is derived from the parental sensitive WT line through sequential exposure of DOX and is therefore selected for DOX resistance. At the time of initiation of these studies, independent sensitive and transfected MDR lines (where the cDNA of mdrl gene is transfected into the sensitive cells) were not readily available. In this chapter, the evaluation of the toxicity, efficacy and pharmacokinetics of DOX encapsulated in 120 nm diameter DSPC/Chol liposomes when co-administered with the MDR modulator PSC 833 is presented. DSPC/Chol liposomal DOX was chosen on the basis of previous reports that this formulation retains DOX for extended periods of time when injected systemically (Mayer et al., 1989). These biological properties of liposomal DOX are compared with DOX administration in the non-encapsulated drug form as well as for DOX entrapped in DSPC/Chol liposomes in the absence of PSC 833. Measurements of PSC 833 levels in plasma and tissues were not undertaken due to unavailability of a suitable analytical procedure as well as limited supplies. 3.2. Materials and Methods 3.2.1. Materials DOX hydrochloride was purchased from David Bull Laboratories (Canada) Inc., Vaudreil, Quebec. Its purity and analytical integrity was affirmed using spectrofluorescence (Perkin-Elmer) and HPLC (Waters Associates, Milford, MA) analyses. PSC 833 was a generous gift from Novartis (Canada) Inc., Dorval, Quebec. The purity and analytical integrity of PSC 833 was affirmed using LC/MS system (Varian Biotech, Altrincham, UK). Distearoylglycerophosphocholine (DSPC, >99% purity) and cholesterol 116 were obtained from Avanti Polar Lipids (Alabaster, Alabama) and Sigma Chemical Company (St. Louis, Missouri) respectively. Cholesteryl hexadecyl ether (3H), a non-exchangeable, non-metabolizable lipid marker was purchased from Amersham Canada (Oakville, Ontario). Female BDF1 mice were obtained from Charles River Laboratories (St. Constant, Quebec). 3.2.2. Liposome Preparation Liposomes composed of DSPC/Chol (55:45; mokmol) were prepared as follows. DSPC (71.4 mg) and Choi (28.6 mg) were weighed and dissolved in 1 ml of chloroform to obtain a concentration of 100 mg/ml. 3H-Cholesterylhexadecyl ether was then added in trace amounts (15 uCi/15 ul for pharmacokinetic studies) which is a non-exchangeable, non-metabolizable lipid marker (Derksen et al., 1987) typically employed to follow the elimination of liposomal lipid. After mixing on a vortex mixer, a 10 ul aliquot was pippeted into scintillation vials, the chloroform evaporated to dryness under a gentle stream of nitrogen gas, and radioactivity determined using liquid scintillation counting (LSC, see section 3.2.8.) following addition of 5 ml of Pico-Fluor scintillation fluid. This provided a specific activity value which was used for calculations determining amount of lipid following extrusion. The original chloroform solution was then evaporated to a viscous consistency under a gentle stream of nitrogen, followed by vacuum drying for 4 h. The dried lipid film was then hydrated using a 300 mM citric acid pH 4.0 buffer. The resulting multilammelar vesicles (MLVs) were subjected to five freeze (in liquid nitrogen) -thaw (in 65°C water bath) cycles. This was followed by a 10 cycle extrusion through two stacked 100 nm polycarbonate filters (Nuclepore, Pleasanton, California) using a Lipex Extruder (Lipex Biomembranes Inc., Vancouver, British Columbia) (Mayer et al., 1986b), 117 at 65 °C. Another 10 pl aliquot of the extruded sample was analyzed for radioactivity using LSC upon addition of the scintillation fluid to determine amount of lipid in the extruded sample. The resulting large unilammelar vesicles (LUVs) exhibited a mean diameter ranging between 100 - 130 nm as determined using a Nicomp 270 submicron particle sizer, operating at a wavelength of 632.8 nm (Particle Sizing Systems, Santa Barbara, CA). Figure 3.2. is a representative Nicomp vesicle size analysis profile of liposomes prepared in this manner. Diameter (nanometers) 24.2 27.9 32.3 37.2 -43.0 Z 49.7 Z 57.4 Z 66.2 " 76.5 H ^ M H . — . 88.3 102.0 ^ ^ ^ ^ S ^ H ^ H 117.8 ^ ^ ^ ^ ^ ^ ^ Z ^ S 136.0 157.1 ^^^^^^^^S 181.4 209.5 241.9 279.3 322.6 372.5 430.2 Figure 3.2. Representative NICOMP gaussian analysis of vesicle size distribution as determined using quasi-elastic light scattering technique. DSPC/Chol liposomes, prepared using the extrusion technique, were diluted in normal saline and analyzed for vesicle size distribution at a count rate of 341 kHz. The figure represents the size distribution histogram where liposome diameter (mean ± standard deviation) was determined as 117.8 ± 28.8 nm (Chi2 = 0.22). 118 Hydrated and extruded empty liposomes prepared at pH 4.0 are stable for several months when stored at 4 °C (Harrigan et al., 1992), however for the studies described here liposomes were prepared freshly for each experiment. Prior to encapsulating DOX, liposomes were characterized with respect to lipid concentration and vesicle size distribution as a quality control measure following extrusion or storage. Vesicle size distribution was determined using quasi-elastic light scattering method, based on the principle of time-dependent intensity fluctuations of scattered light due to Brownian motion of vesicles in solution or suspension. Vesicles exhibiting a mean vesicle diameter of in the range of 100 - 130 nm (with standard deviation not greater than 30% and a Chi-square value of less than 0.4) were rendered acceptable for studies. Lipid concentration was determined using radioactivity determinations in aliquoted samples before and after extrusion, from which the amount of lipid in the sample and loss, if any, was determined. Following drug encapsulation, the encapsulation efficiency and liposome size distribution were determined. Liposomal DOX was applied on 3-ml Sephadex G-50 spin columns (and centrifuged at 4 °C for 5 min) to remove any untrapped drug prior to dilution. A trapping efficiency of >95% was considered acceptable for in vivo use (see Section 3.2.3.). Liposomal DOX formulations were analyzed for DOX as well as liposomal lipid and the drug-to-lipid ratio (weight:weight) was determined before and after removal of unentrapped DOX. It should be noted that changes in the drug-to-lipid ratio of liposomal DOX over time can be used to assess drug leakage from the liposomes and this parameter can be monitored both in vitro and in plasma after in vivo administration of liposomal DOX. 119 3.2.3. Drug Encapsulation DOX was encapsulated in the liposomes using the transmembrane pH gradient loading procedure (interior acidic) employing sodium carbonate as the alkalinizing agent and a drug to lipid weight ratio of 0.2:1.0 (Mayer et al., 1986a,c). Specifically, preformed liposome solutions were titrated with 0.5 M Na 2C0 3 to pH 7.8, creating a pH gradient across the vesicles. Powdered DOX was then reconstituted with sterile saline. Liposomes and DOX solution were then heated in a 65 °C water bath for 2 min. The solution containing DOX was then added to the liposomes, at 65 °C, at a drug to lipid ratio of 0.2:1.0 and vigorously mixed. The liposomal DOX was heated further for 10 min at 65 °C with intermittent mixing. Vesicle entrapped DOX was then determined using column chromatography methods as described above. Liposomal DOX preparations were prepared freshly before use and were diluted with saline as necessary just prior to in vivo administration. Free DOX was administered in sterile saline. DOX was found to be stable for many months as a lyophilized powder, in reconstituted form at 4 °C for at least 10 days, and for at least four freeze-thaw cycles. Further, the stability of DOX inside liposomes was confirmed using HPLC. 3.2.4. PSC 833 formulations and use PSC 833 (MW 1214) is sparingly soluble in water and consequently, was initially dissolved in a 10:1 mixture of ethanol (95%):polyoxyethylene sorbitan monooleate (tween 80). For in vivo studies, this solution was mixed with corn oil and administered by oral gavage in a 200 pl volume (Keller et al., 1992b). PSC 833 formulations were always prepared freshly (-30 min) before experimentation. Due to the lack of an analytical assay 120 for PSC 833, stability studies were performed using electrospray-mass spectrometric analysis in a LC-MS/MS system. Stability conditions were evaluated by comparing the mass spectra of pure substance (obtained from Novartis, Inc) with the spectra obtained from PSC 833 formulations stored at room temperature at 4, 24 h, and 48 h. At each of the conditions evaluated, PSC 833's mass spectra were identical to those obtained at various test groups, confirming the stability of PSC 833. The representative mass spectrum of PSC 833 is illustrated below in Figure 3.3. Note the molecular ion peak observed at m/z 1215 (MFT peak). 100-, 1214.9 1215.9 1232.01236.9 1237.9 1268.8 /V. 1220 ' 1240 1160 1180 1200 1260 1280 1300 1320 1340 Figure 3.3.: Representative mass spectrum of PSC 833 using electrospray MS system. 121 3.2.5. Toxicity Evaluation Studies Toxicity of the indicated DOX and PSC 833 dose regimens were evaluated in dose range finding studies using normal (non-tumor bearing) female BDF1 mice. Formal studies for determining LD10 and LD50 are not sanctioned by the Canadian Council on Animal Care, therefore, toxic dose range finding studies in tumor free female mice were performed using 5 mice/group according to the following protocol. Mice were administered increasing doses of intravenous DOX (non-encapsulated or liposomal) via the tail vein and oral PSC 833 at a fixed dose of 100 mg/kg (4 h before DOX), in single (day 1) or multiple dosing schedules (days 1, 5, 9). Liposomal DOX doses were changed by increments of 10 mg/kg for single injections and by 5 mg/kg in the day 1, 5, 9 multiple injections. Non-encapsulated DOX doses were changed in increments of 5 mg/kg and 2.5 mg/kg above and below, respectively, a DOX dose of 10 mg/kg. DOX dose escalation was stopped when weight loss exceeded 30% or toxicity related mortality was observed. Survival and the percent change in body weight was monitored over a 15 day period. Animals were monitored for physical manifestations of toxicity such as scruffy coat, lethargy, ataxia, or labored breathing. Animals which demonstrated significant physical manifestations of distress /toxicity or exhibited a body weight loss in excess of 30% were terminated. At the end of the 15-day study period, mice were terminated by carbon dioxide asphyxiation. Necropsies were performed to identify abnormalities in the major organs. The dose at which the body weight loss (group mean value) was < 15% and all mice survived for the duration of study was established as the maximum tolerated dose (MTD). The results were collated from at least two independent experiments. Select groups were repeated as a quality control measure to ensure reproducibility. 122 3.2.6. Cell Preparation for Efficacy Studies Stock cultures of P388AVT and P388/ADR cells were purchased from the NCI. RPMI-1640 culture medium containing 10% FBS was used as the culture medium and 10% DMSO was used for cyropreservation. The cell aliquots and ascites were frozen in a nitrogen tank. Before experimentation, a vial of frozen ascites was removed from the tank and thawed at 37°C. About 0.3 ml of these cells were injected i.p. in mice. P388 cells were passaged in vivo in the mouse peritoneum because in vitro tissue culture does not result in cell preparations that achieve acceptable tumor take rates when inoculated in mice. This method involves the removal of adhering monocytes and RBCs by exposing to plastic culture ware and Ficoll-Paque density centrifugation, respectively, such that the cells are viable, with reproducible in vivo growth characteristics. These transfer female BDF1 mice were euthanised by CO2 overdose and soaked in 70% ethanol. Standard surgical procedures were followed, wherein the skin just above the genitalia is excised, facilitating loosening of the skin from peritoneal wall using a blunt tipped scissor under the skin and pushing towards the breast bone. Care was taken to keep the peritoneal wall intact. Using a 1 cc syringe with a 20G needle, about 500-1000 ul of the peritoneal fluid is withdrawn from each mouse and all aliquots are collected in a 15 ml conical sterile tube containing 5 ml of Hanks' Balanced Salt Solution (HBSS) without calcium and magnesium salts and mixed well. A 0.5 ml aliquot is transferred into another 15 ml conical sterile tube containing 5 ml of HBSS. After mixing, a 0.1 ml aliquot is removed for counting. The sample for counting is diluted 1:1 with trypan blue (2%) stain, mixed well, and counted using a Haemocytometer. Greater than 90% cell viability was rendered acceptable for experimentation. A 2 million cells/ml solution in HBSS (without Ca and Mg 123 salts) was then prepared. Two BDF1 female mice (approx. 7-12 weeks old) were then injected with 1 million cells in 0.5 ml of the P388 suspension intraperitoneally. The above procedure was repeated every 6-8 days for a maximum of 20 passages. These peritoneal cells can be used for animal experiments from 3rd to 20th passage. For efficacy experiments, the above protocol was followed, and care was taken to inject the cells into experimental animals within one hour of harvesting. Resistant P388/ADR cells were also maintained in the mouse peritoneum. Specifically, transfer mice containing these cells in their ascites were injected with free DOX at a dose of 6 mg/kg i.p. After two days, 1 million cells were transferred to another set of mice and DOX administered. DOX was exposed to these transfer mice once a week, 7 days after injection. 3.2.7. In vivo Antitumor Activity Efficacy experiments were conducted in BDF1 mice bearing P388 (resistant, ADR and sensitive, WT) solid tumors. Briefly, 1 x 106 P388 (ADR or WT) cells obtained as described above were inoculated subcutaneously in the flank of female BDF1 mice. After one week when the tumors (n=10/group) are established (20-100 mg weight), treatment was initiated with dosage regimens incorporating i.v. non-encapsulated or liposomal DOX with or without p.o. PSC 833 (given 4 h before DOX) on days 1, 5, and 9. Caliper measurements of the tumors were performed daily, and the tumor weights calculated according to the formula (Mayer et al., 1990a): Tumor weight (g) = length (cm) x [width (cm)]2 2 124 This conversion formula was verified by comparing the calculation derived tumor weights to excised and weighed tumors. Animal weights and mortality were monitored daily. Animals bearing ulcerated tumors or where tumor weights exceeded 10% of the animal's body weight were terminated. The results were collated from two independent experiments where all groups were repeated at least two times. Data was expressed as mean tumor weights ± standard error of the mean. Statistical tests were performed using ANOVA and statistical significance was set at p<0.05. 3.2.8. Pharmacokinetics & Tissue Distribution Pharmacokinetic experiments were performed on non-fasted female BDF1 mice bearing 100-200 mg P388/ADR solid tumors. Mice received i.v. via the tail vein a single bolus dose of liposomal DOX or non-encapsulated DOX. In groups incorporating PSC 833, the modulator was administered orally by gavage (100 mg/kg) and DOX administration was carried out 4 h later using the maximum therapeutic doses observed in efficacy studies. After DOX dosing, groups of 3 mice/time point were anesthetized with 100 ul i.p. of ketamine/xylazine at 15 min, 30 min, 1 h, 2 h, 4 h, 8 h, 12 h, 24 h, and 48 h. Blood samples were collected by cardiac puncture and placed into EDTA coated tubes (Microtainer, Becton Dickinson). Plasma was obtained by centrifuging whole blood samples at 500 g for 10 min. After blood collection animals were terminated and heart, kidneys, liver, and tumors were removed from each animal, rinsed in PBS, homogenized and analyzed for liposomal lipid and DOX by scintillation counting and fluorescence, respectively as described below. Measurements in DOX concentrations in plasma and tissues were carried out by thawing tissue homogenates or plasma samples (0.2 ml), followed by addition of 10% 125 sodium dodecyl sulphate (0.1 ml) and 10 mM sulfuric acid (0.1 ml), and finally making up the aqueous phase to 1 ml with distilled water (Mayer et al., 1989). The drug was extracted by the addition of 2 ml of isopropanol-chloroform mixture (1:1, vol:vol) and vigorously vortexing for 2 min. Protein aggregation was facilitated by freezing the samples at -80°C for 1 h, thawing the samples, and centrifuging at 3000 rpm for 10 min. The lower organic layer was pipetted into another tube and analyzed spectrofluorimetrically in a Perkin-Elmer F150 Spectrofluorimeter using an excitation wavelength of 500 nm and an emission wavelength of 550 nm. DOX concentrations were determined using calibration curves generated with DOX concentrations ranging from 10 - 2000 ng/ml, spiked in plasma and tissue homogenates. DOX extraction recoveries (extracted versus unextracted), at 1 pg/ml, averaged 90.8 ± 8.2% (plasma), 86.6 ± 9.3% (liver homogenate), 88.0 ± 6.8% (kidney homogenate), 85.7 ± 3.9% (spleen homogenate), 82.9 + 9.8% (heart homogenate), and 85.2 ± 9.9% (tumor homogenate). HPLC analysis of selected biological samples obtained from mice receiving liposomal DOX has indicated that >98% of the fluorescence detected in the fluorescence assay described above is due to non-metabolized DOX (data not shown, see also Bally et al., 1990a). Due to these findings, a simple fluorescence assay was used to monitor DOX in these initial pharmacokinetic studies. It should be noted that later studies utilize a more rigorous HPLC assay to fully characterize DOX and DOX metabolites in tissue and plasma samples. Aliquots of tissue homogenates (0.2 ml) to be analyzed for liposomal lipid were incubated with Solvable, a tissue solubilizer (0.5 ml) at 50°C for 3 h. Following incubation, samples were brought to room temperature for 1 h to control foaming. 200 126 mM EDTA (50 ul), 30% hydrogen peroxide (200 ul), and ION hydrochloric acid (25 ul), were then added and the mixture incubated at room temperature for 1 h. Five ml of the scintillation fluid, Pico-Fluor was then added and 3 H dpm's were detected using a Liquid Scintillation Counter (LSC). Plasma samples were directly counted using the LSC following addition of the scintillation fluid. All LSC analyses were performed using a TRI-CARB Model 1900 Liquid Scintillation Analyzer, Packard Instrumentation Company (Meriden, CT). Tissue DOX and lipid levels were corrected for contributions from plasma residing in the vasculature of the tissues. For tumor, this correction was performed by multiplying the tumor plasma volume of 12 pl/g by the concentration of DOX or liposomal lipid and subtracting this value from the total tumor content. Two pharmacokinetic parameters were calculated from the concentration-time curves, namely, i.e., Cmax which is defined as the maximum concentration achieved, and AUC(o-t) which is defined as the area under the concentration - time curve from the time of drug administration until the last time point with detectable drug levels. The trapezoidal rule was used to calculate the AUCs in plasma and tissue concentration-time profiles employing a computer software AUC (K. Wayne Riggs, personal communication). Statistical analyses were performed using 2-sample t-test and statistical significance was set at p<0.05. 3.3. Results 3.3.1. Toxicity Studies Drug-modulator dose response toxicity studies were conducted for different dosage regimens comparing free and liposome encapsulated DOX in order to elucidate the effect 127 of liposome encapsulation on DOX-PSC 833 interactions and to identify the maximum tolerated dose (MTD) for each drug form in the presence and absence of the MDR modulator. Intravenous DOX was administered in non-encapsulated and liposomal forms at escalating doses as described in Materials and Methods, alone or in combination with oral PSC 833 (100 mg/kg) and the change in body weight was monitored over a 15 day period. The increase in non-encapsulated DOX toxicity induced by PSC 833 is illustrated in Figure 3.4. When non-encapsulated DOX was administered as a single injection on day 1 at a dose of 20 mg/kg, the nadir in weight change on day 6 was -12.8% which recovered to a weight loss of 8% by the end of the 15 day study period with 100% survival. Using the criteria of <15% weight loss and no mortality, this dose was identified as the maximum tolerated dose (MTD) for this treatment. However, when combined with PSC 833 at 20 mg/kg non-encapsulated DOX, the body weight decreased rapidly 19.5% within 5 days and all mice had to be terminated on day 10 due to a weight loss of 32.2%. The dose of the non-encapsulated DOX in conjunction with PSC 833 (single dose) had to be reduced by 2.7-fold to 7.5 mg/kg in order to obtain 100% survival and a weight loss nadir of < 15.0%. At this dose combination, the nadir weight loss on day 5 day was 14.2% but mice eventually recovered to near pretreatment weight by day 15 with 100% survival (Figure 3.4). These studies demonstrate that PSC-833 reduced the MTD (denned as no deaths and <15% weight loss nadir) of non-encapsulated DOX from 20 mg/kg to 7.5 mg/kg in the single dose regimen. In a similar fashion, PSC 833 co-administration in a multiple dose (days 1, 5 and 9) regimen decreased the DOX MTD from 7.5 mg/kg in the absence of PSC 833 to 2.5 mg/kg (data not shown). 128 Figure 3.4.: Liposomal encapsulation circumvents PSC 833-mediated increased toxicity of i.v. DOX. Intravenous DOX was administered in the free (7.5 mg/kg and 20 mg/kg) and liposomal (50 mg/kg and 70 mg/kg) forms, alone or in combination with oral PSC 833 (100 mg/kg) in normal BDF1 mice and the change in body weight monitored over a 15-day period. Legends: liposomal DOX (70 mg/kg) [•]; liposomal DOX (50 mg/kg) with PSC 833 [•]; non-encapsulated DOX (20 mg/kg) [•]; non-encapsulated DOX (20 mg/kg) with PSC 833 [A]; non-encapsulated DOX (7.5 mg/kg) with PSC 833 [+]. Data are presented as group mean value (n = 5/group). 129 In contrast to the results obtained with free DOX, the MTD of liposomal DOX decreased only marginally in the presence of PSC-833 for a single IV injection. Specifically, when liposomal DOX was administered at a dose of 70 mg/kg, a weight loss of 5.2% was observed over the first 9 days with a study end point of 1.9% weight loss and 100% survival, indicating that liposomal carriers permitted higher amounts of DOX to be delivered with limited toxicity (Figure 3.4). Increasing the dose of liposomal DOX beyond 70 mg/kg resulted in significant increases in weight loss and mortality. More importantly, the MTD of liposomal DOX was reduced only 20% when co-administered with PSC 833. Although administration of liposomal DOX at 70 mg/kg resulted in toxicity related deaths, a dose of 50 mg/kg combined with oral administration of PSC 833 at 100 mg/kg was well tolerated by all mice resulting in a nadir in weight loss of 7.2% on day 8 and at day 15, the weight recovered completely to +0.5% with 100% survival (Figure 3.4). In a similar fashion, the day 1, 5, 9 dosing of liposomal DOX was unaffected by co-administration of PSC 833 and MTD values of 20 mg/kg were obtained in the presence and absence of MDR modulator. It should be noted that the physical presentation of toxicity symptoms as well as their time of onset were comparable for all groups at and above the respective MTDs. 130 3.3.2. Efficacy Figure 3.5 presents the results of anti-tumor activity studies comparing non-encapsulated and liposomal DOX in the P388/ADR in-vivo MDR solid tumor model. In the absence of any treatment, tumors grow to a size of lg within 26-28 days of tumor cell inoculation (12-14 days after formation of a palpable solid tumor, Figure 3.5A). Treatment with non-encapsulated DOX at a dose of 7.5 mg/kg IV on days 1, 5 and 9 (MTD in the absence of PSC 833) as well as oral administration of PSC 833 at 100 mg/kg on days 1, 5 and 9 in the absence of DOX had no effect on tumor growth rates (Figure 3.5 A and B). In contrast, i.v. injection of non-encapsulated DOX to mice bearing wild-type, drug sensitive P388 tumors resulted in a significant reduction (p<0.05) in tumor growth beyond day 8 (Figure 3.5 A). IV administration of liposomal DOX at a drug dose of 10 mg/kg on days 1, 5 and 9 in the absence of PSC 833 provided moderate antitumor activity against the resistant P388/ADR tumors where tumor growth rates were decreased significantly (p<0.05) compared to non-encapsulated DOX treatment between days 9-13. However, tumor growth increased substantially beyond day 12 and mean tumor weights were greater than 0.75g by day 15 (Figure 3.5A). Although non-encapsulated DOX alone had minimal antitumor activity against solid P388/ADR tumors, combining this treatment with PSC 833 did inhibit tumor growth. This is shown in panel B of Figure 3.5 where administration of non-encapsulated DOX at 2.0 mg/kg (MTD in tumor bearing mice) with oral PSC 833 (100 mg/kg) caused tumor growth inhibition, however tumor growth delay was transient and tumors grew rapidly after day 10. 131 Figure 3.5: Antitumor efficacy of free and liposomal DOX against P388/ADR solid tumors in the absence (Panel A) and presence (Panel B) of co-administered PSC 833. P388/ADR and P388AVT solid tumors were grown s.c. in BDF1 mice. Oral PSC 833 (100 mg/kg) and i.v. DOX treatments were initiated once tumors were established (20-100 mg) and were given on days 1, 5, and 9 at the indicated doses of non-encapsulated and liposomal DOX. PSC 833 was administered 4 hours prior to DOX injection. Panel A (PSC 833 free) groups: P388/ADR untreated control tumors (•), P388/ADR tumors treated with non-encapsulated DOX 7.5 mg/kg (•) or liposomal DOX 10 mg/kg (A), and P388/WT sensitive tumors treated with non-encapsulated DOX 7.5 mg/kg (•). Panel B (incorporating PSC 833) groups: P388/ADR control untreated tumors (•), P388/ADR tumors treated with p.o. PSC 833 100 mg/kg (•), P388/ADR tumors treated with non-encapsulated DOX 2.0 mg/kg in conjunction with 100 mg/kg p.o. PSC 833 (T), and P388/ADR tumors treated with liposomal DOX 10 mg/kg in conjunction with 100 mg/kg p.o. PSC 833 (A). Data are expressed as mean ± standard error of the mean (n = 10/group). 132 Administration of DOX in liposomal form at a dose of 10 mg/kg combined with oral PSC 833 at 100 mg/kg provided maximum antitumor activity of the treatments studied. In this group, tumors remained smaller than 150 mg for the duration of the study which was equivalent to that obtained for non-encapsulated DOX in the drug sensitive P388/WT tumor (compare Figure 3.5 A and B). It should be noted that the maximum therapeutic effect for non-encapsulated and liposomal DOX was observed at doses lower than the MTD determined in toxicity studies conducted on tumor-free BDF1 mice. For non-encapsulated and liposomal DOX, animals receiving DOX at the tumor-free MTDs experienced increased toxicities by day 10 which resulted in the need to terminate the animals prematurely. However, prior to termination of these animals, liposomal DOX plus PSC 833 produced complete suppression of P388/ADR tumors in the majority of mice whereas tumors injected with 2.5 mg/kg non-encapsulated DOX were growing at a rate comparable to that observed with non-encapsulated DOX at a dose of 2.0 mg/kg in the presence of PSC 833 (see Figure 3.5B). 3.3.3. Pharmacokinetics and biodistribution DOX pharmacodistribution studies were performed in order to correlate the toxicity and efficacy results obtained for non-encapsulated and liposomal DOX-PSC 833 combinations with plasma and tissue DOX concentrations. These studies were performed in P388/ADR tumor bearing BDF1 mice using single DOX injections corresponding to the maximum therapeutic doses of non-encapsulated and liposomal drug forms in the presence and absence of PSC 833 using the day 1, 5 and 9 dosing regimen employed in efficacy studies. The results summarized in Table 3.1 and Figure 3.6 demonstrate that the DOX 133 elimination profile following administration of non-encapsulated drug at 7.5 mg/kg appeared to be monophasic with a rapid elimination phase occurring within 2 h. DOX concentrations beyond 2 h were undetectable (Figure 3.6A, Limit of quantitation, LOQ = 0.01 pg/ml or 0.01 pg/g). Plasma DOX concentrations peaked at 2.7 ± 0.3 pg/ml with an AUC (0-2h) of 13 pg.h/ml (Table 3.1). In contrast, co-administration of PSC 833 resulted in a biphasic DOX elimination profile characterized by a 2-fold increase in Cmax and a prolonged terminal phase accounting for the 10-fold increase in AUC (Figure 3.6A, Table 3.1; p<0.05). This reduced DOX clearance associated with PSC 833 administration was associated with increased toxicity (see Section 3.3.1.). 134 Heart CJ P A < 2£ 820.4 1229.4 790.6 67.0 121.0 Heart 3 M e M U * 35.1 ± 1.8 38.4 + 2.7 27.9 + 0.9 7.8 + 1.5 6.6 ±0.5 Kidney p -4 1258.1 1614.2 1050.8 1265.6 1386.3 Kidney M « M J 1 98.5 ± 6.4 95.2 ± 3.7 35.6 ± 0.9 30.9 ± 0.6 32.9 ± 1.6 Spleen n -CJ 'u P 4 < a 496.9 1085* 406.1 8923.8 7088.4 Spleen 3 M E en U ^ 13.5 ± 1.4 30.3 + 4.1* 11.6 ± 3.9 259.8 ±7.8 231.3 ± 21.1 Liver CJ 'Oil p -4 874.9 992.2 602.9 2444.5 2535.9 Liver w M £ DO O *• 59.512.4 72.2 ±3.2 24.8 ± 1.4 68.7 ±3.7 71.4 ± 4.9 Tumor CJ P -4 < 2 73.8 115.6 39.1 470.6 490.8 Tumor 3 M £ ^ o * 3.9 + 1.8 3.9 ± 1.6 2.9 ±0.3 13.3 ± 1.9 16.1 ± 4.7 Plasma u 1 P -4 < SL 13.7* o ro o 1095.2 1549.8 Plasma 3 ^ U 1 2.7 + 0.3 4.2 ±0.5 1.6 ±0.3 140.5 ± 5.3 144.5 ± 3.9 Group Free DOX 7.5 mg/kg Free DOX 7.5 mg/kg + PSC 833 100 mg/kg Free DOX 2.0 mg/kg + PSC 833 100 mg/kg Lipo DOX 10 mg/kg Lipo DOX 10 mg/kg + PSC 833 100 mg/kg CO CO 1 1 I g M fa CU CU - O •° s c CU T 3 CU C o c cu o C o CD o O * v ~ Q ^ ,—. " CN co +J 3.! §> CU co "5 co ^ ^ u t—1 co P H « o o 21 g •S S i O H i—i I - " O OH C 3 u o o -fi >• fcl. C30 ^ ~cu g ol O '55 S '3 3 .2 » cu 1 o c o e o II CA S OH , • cu o 6 a 3 -a Figure 3.6: Plasma (A),tumor (B), liver (C), spleen (D), kidney (E), and heart (F) DOX concentration-time profdes following administration of non-encapsulated i.v. DOX 7.5 mg/kg in the absence (•) or presence of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (•); non-encapsulated i.v. DOX 2.0 mg/kg in the presence of PSC 833 100 mg/kg p.o. (A); and i.v. liposomal DOX (55:45, DSPC/Chol, 100 nm) 10 mg/kg in the absence (T) or presence of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (•). Mice bearing 200 - 500 mg solid P388/ADR tumors were sacrificed at the indicated times and plasma and tissue samples were processed and quantified for DOX content (fluorescent DOX equivalents) as described in Materials and Methods. Data are expressed as mean ± standard error of the mean (n = 3 mice/time point; n = 27 mice/group). (Figure 3.6, continued on next page) 136 Figure 3.6 (Continued from previous page) 137 A similar pharmacokinetic profile to non-encapsulated DOX at 7.5 mg/kg in the absence of PSC 833 was observed for non-encapsulated DOX at 2 mg/kg when co-administered with PSC 833 where a CmaX of 1.6 ± 0.3 \xg/ml was obtained and plasma DOX levels fell below the LOQ after 1 h. It should be noted that because DOX concentrations at later time points were below assay detection limits in groups incorporating non-encapsulated DOX (7.5 mg/kg non-encapsulated DOX without PSC 833 and 2 mg/kg DOX with PSC 833) AUC comparisons with groups exhibiting extended DOX plasma life times may be underestimated. However, even if we were to assume that DOX concentrations at undetected time points were assigned an arbitrary value approximating the assay LOQ, the increase in AUC value associated with PSC 833 co-administration (for non-encapsulated DOX 7.5 mg/kg) would be at least 7.4-fold higher than that derived in groups with no PSC 833. Therefore, PSC 833 co-administration results in a marked increase in plasma AUC of non-encapsulated drug. In contrast to the results obtained with non-encapsulated DOX, there is no significant difference in the C m a x values of DOX at a dose of 10 mg/kg administered in liposomal form in the presence or absence of PSC 833 and the increase in AUC induced by PSC 833 was small (not statistically significant) compared to non-encapsulated DOX indicating that the plasma elimination of DOX administered in liposome encapsulated form was largely unaffected by this MDR modulator. Specifically, there is only a 1.03 fold increase in C ^ (140.5 ± 5.3 ug/ml without PSC 833 vs. 144.5 + 3.9 ug/ml with PSC 833) and 1.4 fold increase in AUC (1095.2 ug.h/ml without PSC 833 vs. 1549.8 ug.h/ml with PSC 833) when PSC 833 (100 mg/kg) is combined with liposomal DOX at 10 mg/kg (Table 3.1). The DOX elimination profile was monophasic both in the presence and 138 absence of PSC 833. Additionally, administration of DOX encapsulated in 120 nm DSPC/Chol liposomes resulted in a 600-fold increase in AUC compared to non-encapsulated DOX at comparable doses (Figure 3.6A, p<0.05). Liposomal lipid (using a non-exchangeable, non-metabolizable marker) and DOX elimination were also followed over time. These parameters were used to calculate the drug-to-lipid ratio which provides an assessment of liposome-associated drug in the plasma as a function of time. These assessments are based on the assumption that free drug released from liposomes is rapidly cleared from the circulation and is no longer associated with the liposomes (Thies et al., 1990). Measurement of liposomal lipid and DOX elimination in plasma indicated that >95% of the drug is associated with the liposomes where the drug to lipid ratio remained constant at 0.2:1 (weight:weight) over the entire time course. This suggests that DOX distribution properties are due to drug being delivered in encapsulated form. DOX accumulation in solid tumors was characterized in order to correlate the efficacy data (Figure 3.5) with tumor DOX concentrations. Tumor DOX levels following administration of non-encapsulated DOX at 7.5 mg/kg resulted in relatively stable accumulation of the drug with a C m a x of 3.9 ± 1.8 ug/g and an AUC of 73.8 ug.h/g. In the presence of PSC 833, although the C m a x was similar at 3.9 ±1 .6 ug/g, DOX concentrations were significantly elevated at later time points resulting in a 1.6 fold increase in AUC (115.6 ug.h/g v s - 7 3 8 Hg-h/g). The AUC for tumor levels of DOX following treatment with 2 mg/kg non-encapsulated DOX in the presence of PSC 833 was observed to be 39.1 ug.h/g, indicative of a decrease in tumor drug exposure when non-encapsulated DOX is combined at its MTD with PSC 833 compared to non-encapsulated DOX at the MTD in 139 the absence of PSC 833 (Figure 3.6B, Table 3.1). Tumor DOX concentrations following liposomal delivery were significantly higher than non-encapsulated drug at similar doses, consistent with previous observations that liposomes selectively localize in tumor tissues due to the leaky vasculature which enhances liposome extravasation (Gabizon and Papahadjopolous, 1988; Gabizon, 1992; Mayer et al., 1990a,b; Jain, 1988). Specifically, there was a 5-fold increase in AUC when liposomes were used as a carrier at a DOX dose of 10 mg/kg compared to non-encapsulated DOX administered at 7.5 mg/kg. Furthermore, tumor DOX kinetics after injection of liposomal DOX were unaffected by PSC 833 as observed by minor increases in tumor C^x and AUC values (Table 3.1, Figure 3.6B). In this case, the lack of an effect of the MDR modulator on tumor drug levels strongly suggests that increased delivery of DOX to P388/ADR tumors via liposomes is in itself insufficient to explain the therapeutic improvements provided by combined therapy with PSC 833. In order to assess if liposome clearance in organs of the reticuloendothelial system are affected by PSC 833, liver and spleen DOX concentrations were determined. DOX distribution in the liver following administration of the non-encapsulated drug at 7.5 mg/kg was indicative of a biphasic elimination profile with a C m a x of 59.5 ± 2.4 ug/g (Table 3.1, Figure 3.6C). Co-administration of PSC 833 at this drug dose increased liver DOX uptake slightly resulting in approximately a 1.2-fold increase in C m a x and AUC. Relatively low drug levels were quantified in spleen after administration of non-encapsulated DOX. However, co-administration of PSC 833 resulted in a doubling of the C m a x and a 2.2-fold increase in AUC (Table 3.1), indicating that PSC 833 reduction of plasma DOX clearance leads to increased drug accumulation in the spleen (significant at p<0.05). This increase 140 was evident when non-encapsulated drug at 2 mg/kg was combined with PSC 833 (Table 3.1), where comparable Cmax and AUC values were obtained at a 3.8-fold lower DOX dose. The higher DOX levels in the liver and spleen following liposomal delivery compared to non-encapsulated drug were characteristic of previous observations with this delivery system (Mayer et al., 1995a; Mayer et al., 1989). However, the distribution profile for liposomal DOX was unaltered by PSC 833 as reflected by comparable C m a x and AUC values (Table 3.1, Figure 3.6C). In the spleen, liposome encapsulation of DOX resulted in a dramatic increase in tissue concentrations with a Cnu,x of 259.8 ±7 .8 ug/g and an AUC value of 8923.8 ug.h/g for a DOX dose of 10 mg/kg (Table 3.1, Figure 3.6D). This represents a 13.5-fold increase compared to the non-encapsulated drug administered at 7.5 mg/kg. However, unlike non-encapsulated drug, co-administration of PSC 833 with liposomal DOX did not affect DOX accumulation characteristics in spleen (Figure 3.6D). The effects of PSC 833 co-administration on DOX toxicity in target organs such as the kidney (nephrotoxicity) and the heart (cardiotoxicity) were correlated with DOX distribution kinetics in .these two susceptible tissues were examined. Following administration of non-encapsulated DOX at 7.5 mg/kg, DOX appeared to undergo a biphasic elimination profile (0 - 4 h rapid distribution phase and 8 - 48 h slow terminal elimination phase) with a C m a x of 98.5 ± 6.4 ug/g in the kidney (Figure 3.6E, Table 3.1). Administration of non-encapsulated DOX at 2.0 mg/kg with PSC 833 resulted in kidney DOX levels similar to those observed for non-encapsulated DOX at 7.5 mg/kg in the absence of PSC 833 at later time points (Figure 3.6E). In the heart, DOX levels following non-encapsulated DOX administration at 7.5 mg/kg were characterized by a rapid accumulation phase (0-2 h) followed by a slow elimination phase (Figure 3.6F). When 141 PSC 833 was co-adrninistered, although comparable Cmax values were obtained, increased DOX tissue levels at later time points led to a 1.5-fold increase in AUC for a non-encapsulated DOX dose of 7.5 mg/kg (Table 3.1). Decreasing the DOX dose to 2 mg/kg with co-administration of PSC 833 resulted in AUC values comparable to those obtained following non-encapsulated DOX at 7.5 mg/kg in the absence of PSC 833 (Table 3.1). DOX levels in both the kidney and the heart following liposomal delivery were lower than for non-encapsulated drug administration. In the kidney, this was reflected by a Cmax of 30.9 ± 0.6 pg/g (compared to' 98.5 ± 6.4 pg/g for free DOX) and total DOX exposure was minimally affected by co-administration with PSC 833 (Figure 3.6E, Table 3.1). In the heart, DOX concentrations were significantly (p<0.05) lower than all free DOX treatment groups both with respect to Cmax and AUC and these levels were comparable in the presence or absence of PSC 833 (Figure 3.6F). 3.4. Discussion The cyclosporin derivative PSC 833 has been shown to be a potent chemosensitizing agent for a wide variety of cells over-expressing the drug efflux pump PGP (Fan et al., 1994; Keller et al., 1992b). When utilized in vivo, PSC 833 itself is non-toxic at typical MDR reversing concentrations. However, studies have demonstrated that it markedly increases the toxicity of non-encapsulated DOX in normal mice (Gonzalez et al., 1995; Froidevaux and Loor, 1994; Pourtier-Mazanedo et al., 1995). PSC 833-mediated increases in toxicity have been correlated with alterations in DOX pharmacokinetics and are presumably a consequence of increased DOX bioavailability to susceptible target organs (Gonzalez et al., 1995). This is consistent with clinical observations where PSC 833 and cyclosporin A have been shown to decrease the clearance 142 of non-encapsulated DOX, taxol and etoposide in Phase I and Phase II clinical trials which resulted in increased toxic side effects and a need to decrease the anticancer drug dose approximately 2-fold (Erlichman et al., 1993; Sarris et al., 1996; Boote et al., 1996; Erlichman et al., 1994). The decrease in the MTD of anticancer drugs induced by co-administration with PSC 833 has hampered our ability to resolve whether this or other MDR modulators exhibiting similar properties are able to effectively reverse MDR in vivo. Preclinical studies on the uptake of i.v. administered non-encapsulated DOX into drug sensitive tumors suggest that accumulation of this anticancer agent in solid tumors is dictated primarily by very early distribution and elimination phases after injection (Mayer et al., 1990a,b; Mayer et al., 1997). These data suggest that reducing the dose of DOX co-administered with PSC 833 to a level that is equitoxic to DOX alone may reduce tumor exposure to the therapeutic agent, a hypothesis corroborated in this Chapter. This could in part explain why the in vivo therapy of anticancer drugs used in combination with PSC 833 has typically fallen short of achieving the degree of antitumour activity obtained with drug sensitive parental cell lines from which the MDR tumor cells are derived (Watanabe et al., 1996a; Boesch et al., 1991a; Watanabe et al., 1995). The use of liposome encapsulated DOX in conjunction with PSC 833 here was based on the demonstrated ability of these lipid carriers to reduce the exposure and toxicity of non-encapsulated drug to susceptible tissues such as the heart, kidney and gastro-intestinal tract (Van Hossel et al., 1984; Mayer et al., 1989). It was postulated that this altered biodistribution may circumvent the effects of PSC 833 on non-encapsulated DOX pharmacokinetics and thereby ameliorate the elevated toxicities associated with this 143 combination therapy. Here, a DSPC/Chol 120 nm liposomal DOX formulation was used. This formulation was selected based on its ability to retain encapsulated DOX over periods of days after i.v. administration (Mayer et al., 1989), a feature confirmed here by the stable plasma drug-to-lipid ratio of 0.2:1.0 observed in the presence or absence of PSC 833. Consequently, the influence of PSC 833 on in vivo properties of liposomal DOX did not suffer significant complications arising from DOX released from liposomes in the circulation. The results of this study have supported our prediction of alleviated pharmacokinetic interactions and have demonstrated that this effect not only results in the ability to increase the dose of DOX, but that this increased dose intensity provides significantly improved therapy of solid tumors exhibiting a PGP-based MDR phenotype. In the absence of PSC 833, the reduced toxicity of DOX encapsulated in DSPC/Chol liposomes compared to non-encapsulated drug was comparable to that observed previously for these systems in single and multiple dose regimens (Mayer et al., 1989; Mayer et al., 1990a,b). For single i.v. injections, there was a 3.5-fold increase in DOX dose administered in normal mice for liposomal formulations compared to free DOX. Liposome encapsulation afforded a an 8-fold increase in DOX dose administered compared to the non-encapsulated form with the multiple i.v. injection in the presence of PSC 833 in normal animals and this correlated with reduced DOX exposure to susceptible normal tissues. In tumor bearing animals, there was a 5-fold increase in the MTD for liposomal DOX. These data indicate that liposomal carriers can deliver substantially elevated amounts of DOX systemically while circumventing the adverse pharmacokinetic interactions observed with non-encapsulated (conventional) drug preparations co-administered with MDR modulators. 144 It is evident from our results that non-encapsulated DOX at 7.5 mg/kg (MTD) alone does not confer any therapeutic activity in the MDR solid tumor model. In the presence of PSC 833, however, there is transient inhibition of tumor growth until day 9 after which the growth characteristics are similar to the untreated P388/ADR control tumors, offering marginal therapeutic value (Figure 3.5B). Furthermore, the dose of the non-encapsulated DOX had to be reduced from 7.5 mg/kg to 2 mg/kg when combined with PSC 833. This, as corroborated by the tissue distribution studies, may be inadequate to provide sufficient tumor drug levels that will lead to sustained antitumor activity. In contrast, liposomal DOX in the absence of PSC 833 does provide some degree of therapy against the MDR tumor, most likely by virtue of the increased tumor drug delivery provided by this system (Figure 3.5A). However, the fact that only the combination of PSC 833 with the liposomal DOX formulation provided complete chemosensitization strongly supports the direct involvement of PGP blockade in MDR solid tumor reversal. This direct role of PSC 833 in the therapy of PGP overexpressing tumors is corroborated by the fact that DOX pharmacokinetic properties were not significantly altered by PSC 833 when co-administered with the liposomal formulation. In agreement with previous studies (Colombo et al., 1996; Gonzalez et al., 1995), the toxicity of non-encapsulated DOX with PSC 833 observed here appears to be related to the prolonged terminal drug elimination phase (Figure 3.6A), resulting in a 10-fold increase in AUC and increased accumulation in susceptible tissues. At equitoxic doses, solid tumor non-encapsulated DOX levels are decreased when co-administered with PSC 833, which could account for its inadequate therapeutic properties. Our results indicate that liposomal delivery of DOX in conjunction with PSC 833 provides plasma levels 145 comparable to liposomal DOX administered alone, indicative of a favorable pharmacokinetic behavior of liposomal DOX-PSC 833 co-administeration. Further, tumor DOX levels following liposomal delivery were significantly higher than that observed for the non-encapsulated drug administration while DOX levels in susceptible tissues (heart, kidney, liver, and spleen) were lower than that for non-encapsulated drug. Most importantly, the tumor associated DOX levels were unaffected by PSC 833 when liposomes were used, unlike the observations seen with non-encapsulated DOX. In addition, no apparent increases in DOX concentrations after liposomal DOX injection were observed to be caused by PSC 833 in plasma and analyzed tissues and this correlated with toxicity and efficacy data in mice bearing P388/ADR tumors. Of particular interest is that although liposomes increase DOX delivery to the liver, which has been associated with PSC 833 induced DOX pharmacokinetic changes, no indications of liver toxicity or altered drug clearance properties were observed. This suggests that liposomes alter the metabolic processes involved for non-encapsulated DOX in the liver, an area addressed in Chapters 4 and 5. The results obtained here suggest that the pharmacological properties of MDR modulators and anticancer drugs can be optimized by selectively targeting PGP blockade effects in tumor tissues. Further, the ability of liposome encapsulated DOX in the presence of PSC 833 to completely inhibit tumor growth as well as the delayed tumor growth observed in the absence of MDR modulators with liposomal DOX demonstrate the importance of anticancer dose intensity in order to effectively treat MDR tumors. 146 Chapter 4 INFLUENCE OF LIPOSOMAL DRUG RETENTION AND TUMOR ACCUMULATION PROPERTIES ON THE TOXICITY, ANTITUMOR ACTIVITY, & PHARMACOKINETICS OF DOX, IN THE PRESENCE AND ABSENCE OF PSC 833 USING THE MDA435LCC6 MDR1 HUMAN XENOGRAFT SOLID TUMOR MODEL 4.1. Introduction The development of second generation MDR reversing agents alleviated many of the problems caused by earlier PGP blockers which were pharmacological agents with their own inherent toxicities. However, co-administration of conventional anticancer drugs with many of these newer MDR modulators has been shown to elicit drug - modulator interactions, by virtue of PGP blockade in normal tissues such as the liver, kidney, intestine, and brain (Keller et al., 1992b; Krishna and Mayer, 1997). In the liver, daunorubicin transport across canalicular membranes has been shown to be unidirectional, a process that may be inhibited by MDR modulators (Gatmaitan and Arias, 1993), suggesting a role of PGP in the biliary excretion of anticancer agents. Recently, expression of another transporter, the canalicular multiple organic anion transporter (cMOAT/MRP2), a conjugate export pump, has also been demonstrated in the biliary canalicular membrane (Paulusma et al., 1996) and cyclosporines, particularly Cyclosporin A (CsA), effectively inhibit cMOAT (Sikic et al, 1997). It may not be surprising then that biliary clearance of several drugs has been shown to be inhibited by MDR modulators, namely; colchicine clearance blockade by CsA (Speeg et al, 1992a), colchicine and DOX clearance blockade by PSC 833 (Speeg and Maldonado, 1994) as well as DOX clearance blockade by GW918 (Booth et al., 1998). These observations indicate that drug transporter pump(s) may be 147 preferentially expressed in biliary canaliculi and that the use of MDR modulators can impede drug transport and their subsequent excretion. In the kidney, PGP is highly expressed on the brush border of proximal renal tubule, indicating a role in secretion of xenobiotics into the urine. In support of this, colchicine renal clearance was inhibited by CsA (Speeg et al., 1992b), suggesting that MDR modulators can alter the renal elimination processes of conventional (non-encapsulated) anticancer drugs. The effects of MDR modulators on drug transport proteins that cause alterations in anticancer drug excretion often have lead to increased exposure to healthy tissues and have necessitated anticancer drug dose reduction in many cases. Although doses can be adjusted to equal levels of toxicity, it is unclear how such pharmacokinetic changes may have an impact on therapeutic activity. Evidence of drug-modulator interactions and necessitation of dose reduction has been demonstrated in preclinical (Nooter et al., 1987; Horton et al., 1989; Keller et al., 1992b; Krishna and Mayer, 1997) and clinical studies (Kerr et al., 1986; Tolcher et al., 1994; Giaccone et al., 1994; Boote et al., 1996; Sarris et al., 1996). These interactions have also been postulated to play a role on limiting therapeutic activity in some patient populations (Wishart et al., 1994; Van Kalken et al., 1991; Murphy et al., 1994; Miller et al., 1994; Mross et al., 1993a). While changes in anticancer drug dose and/or schedule may be able to address toxicity or efficacy alterations brought about by MDR modulators, this clearly represents a significant complication in applying PGP blockade strategies to cancer chemotherapy. This is due to the fact that most chemotherapy regimens utilize drug combinations, of which more than one are often PGP substrates. Consequently, an ability to avoid such anticancer drug clearance alterations may be considered to be a significant advantage in MDR modulation strategies. 148 In Chapter 3, it was shown that liposome encapsulation of DOX can significantly circumvent non-encapsulated drug-PSC 833 interactions, resulting in effective growth suppression of multidrug resistant murine solid tumors (Krishna and Mayer, 1997). The liposomal formulation used in Chapter 3 exhibited a mean diameter of 100-130 nm and was composed of DSPC/Chol (55:45 molar ratio). These vesicles are characterized by their ability to retain DOX for extended periods of time (Mayer et al, 1989), and the decreased DOX toxicity observed appeared to be a consequence of the increased protection from PSC 833-mediated pharmacokinetic changes. However, the mechanisms by which this effect is achieved are not fully understood, particularly since significant amounts of DOX are delivered to the liver by the liposomes without notable toxicological consequences. In addition to alleviation of PK alterations, the increased delivery of DOX to MDR solid tumors using liposome delivery systems was associated with increased antitumor activity when co-administered with PSC 833 compared to non-encapsulated drug. These studies were unable, however, to distinguish the degree of DOX bioavailability (liposome entrapped versus released drug) to malignant cells in the solid tumor. Therefore, the relative roles of PGP blockade and tumor drug levels in determining the degree of antitumor activity remain unresolved. In order to address these issues, the studies described in this Chapter compared the toxicity, efficacy, pharmacokinetics, as well as tissue and cellular distribution properties of EPC/Chol DOX and a sterically stabilized PEG20oo-DSPE/DSPC/Chol DOX formulation when combined with the MDR modulator, PSC 833. These studies were designed to test the hypothesis whether a liposomal carrier system exhibiting optimal retention of the encapsulated drug and tumor uptake properties will confer enhanced ability to provide 149 increased tumor growth suppression and MDR reversal, while decreasing modulator-induced toxicity. EPC/Chol liposomal DOX is a system which exhibits increased initial drug leakage rates (over 50% of the drug is released within the first hour) and consequently, would not be expected to confer an optimized protective effect when co-administered with PSC 833. The non-leaky PEG-DSPE/DSPC/Chol system was chosen on the basis of reports that 5 mol% incorporation of PEGzooo-derivatized DSPE in 120 nm DSPC/Chol vesicles results in a 4-fold increase in liposome circulation longevity and a 2-fold reduction in liver accumulation (Papahadjopolous et al., 1991). In addition, increases in tumor accumulation and efficacy have been demonstrated (Gabizon, 1992; Woodle et al., 1994; Lasic, 1996; Torchilin et al., 1994; Torchilin et al., 1995). Such improvements in the pharmacokinetic behavior of these systems has been attributed to the ability of PEG on the liposome surface to cause steric inhibition of electrostatic and hydrophobic interactions between the liposome and plasma proteins (Allen et al., 1989; Lasic et al., 1991). Therefore, the effects of altering the drug release, liver uptake and tumor accumulation properties of liposomal DOX on toxicological and therapeutic activity may be used to reveal the processes underlying the improvements in MDR tumor therapy achieved with liposomal systems in combination with PSC 833. For toxicity and efficacy studies, DSPC/Chol DOX and non-encapsulated DOX, in the presence and absence of PSC 833, were included to bridge the studies described in Chapter 3 with those described here. Also, an orthotopic human breast carcinoma xenograft solid MDR tumor model was utilized in the present studies, where solid tumors were generated by inoculating MDA435LCC6 human breast carcinoma sensitive (WT) or 150 mdr-1 transfected (MDR) cells in the mammary fat pads of SCID/Rag2 mice (Leonessa et al., 1996). This model was selected based on the well characterized MDR properties of the PGP overexpressing cell line, its reliable in vivo tumor growth characteristics and the theoretically increased relevance of an orthotopic human breast carcinoma model, particularly compared to the P388/ADR lymphocytic leukemia line being grown as a solid tumor (Chapter 3). The availability of sensitive and MDR transfected variants of the human breast carcinoma allows comparative assessments of sensitive and MDR cellular responses to anticancer drugs and MDR modulators to be made. 4.2. Materials and Methods 4.2.1. Materials DOX hydrochloride was purchased from David Bull Laboratories (Canada) Inc., Vaudreil, Quebec, and its purity affirmed by HPLC. PSC 833 was a generous gift from Novartis (Canada) Inc., Dorval, Quebec, and its purity affirmed by LC/MS-MS (Varian V G Biotech LC-MS-MS system, Fisons Instruments, Altrincham, UK). DOX metabolite standards were generous gifts from Pharmacia Carlo Erba (Milan, Italy). Polyethylene glycol 2000 coupled with distearoylphosphoethanolamine (PEG2ooo-DSPE, >99% purity), egg phosphocholine (EPC, >99% purity), and l,2-distearoyl-sn-glycero-3-phosphocholine (DSPC, >99% purity) were obtained from Northern Lipids, Inc (Vancouver, BC) and cholesterol was obtained from Sigma Chemical Company (St. Louis, Missouri). Tritiated cholesteryl hexadecyl ether, a non-exchangeable, non-metabolizable lipid marker (Derksen et al., 1987) was purchased from Amersham Canada (Oakville, Ontario). Di-I (1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine percholate) was purchased from 151 Molecular Probes. Tween 80 (polyoxyethylene sorbitan monooleate), sodium carbonate, 30% w/w hydrogen peroxide, were obtained from Sigma Chemical Company (St Louis, MO), and ammonium formate from Fisher Scientific (Fairlawn, NJ). HPLC grade solvents, iso-propyl alcohol, acetone, ethanol, acetonitrile, and chloroform were all obtained from BDH, Inc (Toronto, Ontario), and used without further purification. HPLC grade water was prepared on-site with a Milli-Q system (Millipore Corporation, Mississauga, Ontario). Compressed air (medical grade) and nitrogen, NF were purchased from Praxair, Inc (Mississauga, Ontario). Pyrex 15 mL disposable culture tubes (Corning Glass Works, Corning, NY) and polytetrafluoroethylene (PTFE) lined screw caps (Canlab, Vancouver, BC) were used. Sterile supplies as well as Sterile saline and water for injection, were obtained from the BC Cancer Agency stores (Vancouver, BC). Female BDF1 mice were obtained from Charles River Laboratories (St. Constant, Quebec). Female SCJD/RAG2 mice were bred in-house at the BC Cancer Agency animal facility. The MDA435LCC6 and its transfected MDR-1 line were generous gifts from Dr. Robert Clarke, Georgetown University, Washington DC. 4.2.2. Liposome and Drug Preparation Liposomes composed of EPC/Chol (55:45), PEG2ooo-DSPE/DSPC/Chol (5:50:45), and DSPC/Chol (55:45; mol: mol) were prepared by initially dissolving the lipid mixtures in chloroform (100 mg lipid/ml), vacuum drying to generate the thin film, and hydrating this dried lipid film in a 300 mM citric acid pH 4.00 buffer (see Section 3.2.2. for a detailed description of this procedure). Briefly, the resulting multilammelar vesicles (MLVs) were subjected to five freeze-thaw cycles followed by a 10 cycle extrusion through two stacked 100 nm polycarbonate filters (Nuclepore, Pleasanton, California) using a Lipex Extruder 152 (Lipex Biomembranes Inc., Vancouver, British Columbia) (Mayer et al., 1986b). 3 H -Choleterylhexadecyl ether, used as a non-exchangeable, non-metabolizable lipid marker (Derksen et al., 1987), was added to the chloroform lipid mixtures prior to thin film formation. The resulting large unilammelar vesicles (LUVs) exhibited a mean diameter ranging between 100 - 120 nm as determined using a Nicomp 270 submicron particle sizer (Particle Sizing Systems, Inc, Santa Barbara, CA), operating at a wavelength of 632.8 nm. DOX was encapsulated in the liposomes using the transmembrane pH gradient loading procedure (interior acidic) employing sodium carbonate as the alkalinizing agent and a drug to lipid weight ratio of 0.2:1.0 (Mayer et al., 1986a), as described in Section 3.2.3. Liposomal DOX preparations were diluted with saline as necessary prior to in vivo administration. PSC 833 (for animal studies) was dissolved in a 10:1 mixture of ethanol (95%):(polyoxyethylene sorbitan monooleate (tween 80) and administered in a corn oil vehicle by oral gavage of a 200 pl volume (Keller et al, 1992b; also see Section 3.2.4.). Non-encapsulated DOX was administered in sterile saline. 4.2.3. Liposome characterization Vesicle size distribution, liposome encapsulation efficiency and drug-to-lipid ratio were determined as quality control measures following extrusion or storage of preformed and encapsulated liposomes (see Section 3.2.2. for a detailed description). Vesicles exhibiting a mean vesicle diameter of 100 - 130 nm, as determined using quasi-elastic light scattering method, with a standard deviation not greater than 30% and a Chi-square value of less than 0.4 were rendered acceptable for studies. A drug encapsulation efficiency of >95% was acceptable for studies. Both in vitro as well as following in vivo administration, 153 liposomal DOX samples were analyzed for DOX as well as liposomal lipid in order to determine the drug-to-lipid ratio. A drug-to-lipid ratio of 0.2:1.0 was used for the studies. 4.2.4. Toxicity Evaluation Studies Toxicity of the indicated DOX and PSC 833 dose regimens were evaluated in dose range finding studies using normal (non-tumor bearing) female BDF1 mice. Formal studies for determining LDio and L D 5 0 are not sanctioned by the Canadian Council on Animal Care, therefore, toxic dose range finding studies in tumor free female mice were performed using 3 mice/group with appropriate group repetitions as described below. Briefly, mice were administered increasing (each dose in different groups) doses of intravenous DOX (free or liposomal) via the tail vein and oral PSC 833 at a fixed dose of 100 mg/kg (4 h before DOX), in single (day 1) or multiple dosing schedules (days 1, 5, 9). Liposomal DOX doses were changed by increments of 10 mg/kg for single injections and by 5 mg/kg in the day 1, 5, 9 multiple injections. DOX dose escalation was stopped when weight loss exceeded 30% or toxicity related mortality was observed. Survival and the percent change in body weight 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 demonstrated significant physical manifestations of distress /toxicity or exhibited a body weight loss in excess of 30% were terminated. At the end of the 21-day study period, mice were terminated by carbon dioxide asphyxiation. Necropsies were performed to identify abnormalities in the major organs. The dose at which the body weight loss (group mean value) was < 15% and all mice survived for the duration of study was established as the maximum tolerated dose (MTD). A deviation in MTD weight loss nadir of ± 1.5% (10%) was allowed provided body weight loss was very transient with full 154 recovery of body weight loss and 100% survival. The results were collated from at least two independent experiments. Select groups were repeated as a quality control measure to ensure reproducibility, particularly for dose groups near the MTD. 4.2.5. In vivo Antitumor Activity MDA435LCC6 WT and MDR-1 cells were cultured in Dulbecco's Minimum Essential Medium containing L-amino acids, vitamins, inorganic salts, as well as glucose, phenol red, and sodium pyruvate. This medium was further supplemented with 10% FBS and 1% L-glutamine. The MDR cells were generated by transfecting the parental cell line with a plasmid containing mdrl cDNA (Leonessa et al., 1996). Prior to using these cells for in vivo experimentation, their sensitivity to DOX in WT and MDR cells were determined using cytotoxicity studies. Also assessed was the restoration of DOX sensitivity in MDR cells caused by PSC 833. These cells overexpress PGP in their plasma membrane by approximately 30-fold. Specifically, PSC 833 (1 pM) caused a 30-fold reversal of DOX resistance with a residual resistance factor of 0.67 in cultured MDR cells, a finding which was corroborated by increased uptake of DOX in MDR cells chemosensitized with PSC 833 using flow cytometric analysis. Further, PSC 833 did not sensitize WT cells as reflected by identical IC50S of DOX in the presence and absence of the MDR modulator. These MDA435LCC6 WT and MDR cells were passaged at least three times in medium. Cell aliquots containing 5 million cells in 0.5 ml HBSS are injected i.p. into two female SCID/Rag2 mice. After 20-25 days, ascites (cells) were removed from the mouse via the peritoneal wall using a 20G needle, and placed in sterile 15 ml conical tubes containing 5 ml of HBSS without Ca and Mg salts. The ascites cell suspension was further 155 diluted in 50 ml tube with a total volume of 25-30 ml HBSS and mixed well. A 0.1 ml aliquot was counted after diluting 1:1 with trypan blue stain using a haemocytometer. The cell suspension was then centrifuged at 1000 rpm for 5 min and the supernatant discarded. The cells were then resuspended in HBSS to obtain a concentration of 40 million cells/ml. Using a 27G needle fitted on a 1 cc syringe, 50 ul of the cell suspension was injected into the two mammary fat pads on each mouse (2 million cells/pad). A period of 3 weeks was needed for the tumors to become palpable and measurable for drug treatments to begin. Tumor growth suppression experiments were conducted in SCJD/Rag2 mice bearing orthotopic human breast carcinoma MDA435LCC6 (multidrug resistant, MDR-1 and sensitive, WT) solid tumors. These mice possess a genetic autosomal recessive mutation and show a combined immunodeficiency affecting both B and T lymphocytes, resulting in an inability to reject allogenic grafts including tumors. Briefly, 2 x 106 LCC6 (MDR or WT) cells were inoculated bilaterally in the mammary fat pads of female SCID/RAG2 mice. After three weeks, when the tumors (n=8/group) were established, treatment was initiated with dosage regimens incorporating i.v. non-encapsulated or liposomal DOX with or without p.o. PSC 833 (given 4 h before DOX) on days 1, 5, and 9. Caliper measurements of the tumors were performed daily, and the tumor weights calculated according to the formula (Mayer et al., 1990a): Tumor weight (g) = length (cm) x [width (cm)]2 2 This conversion formula was verified by comparing the calculation derived tumor weights to excised and weighed tumors. Animal weights and mortality were monitored daily. Animals bearing ulcerated tumors or where tumor weights exceeded 10% of the animal's 156 body weight were terminated. Two independent experiments were performed. The weights of the bilateral tumors were averaged for each mouse and mean tumor weights for each treatment group ± standard error of the mean were calculated. In order to compare the results from the two independent experiments, two parameters were used (Watanabe et al, 1996a). Relative weight (RW) = [Weight on day n / Weight on day of treatment] x 100 T/C (%) = [RW of treatment group / RW of respective control] x 100 Statistical tests for these longitudinal data were performed using repeat measures ANOVA employing Statistica for Windows 4.0 (StatSoft, Inc., Tulsa, OK) and statistical significance was set at p<0.05. 4.2.6. Pharmacokinetics & Tissue Distribution Female SCID/Rag2 mice bearing MDA435LCC6/MDR-1 solid tumors received i.v. via the tail vein a single bolus injection of free (non-encapsulated) or liposomal DOX (3H-Cholesterylhexadecyl ether labeled EPC/Chol, or PEG2ooo-DSPE/DSPC/Chol formulated at a drug-to-lipid ratio of 0.2: 1.0). PSC 833 (100 mg/kg) was administered orally (in 0.2 ml) and doxorubicin administered 4 hours later. After DOX dosing, groups of 3 mice/time point were anesthetized with 100 pl i.p. of ketamine/xylazine at 30 min, 1 h, 2 h. 4 h, 16 h, 24 h, 48 h, and 72 h. The treatment groups included: 1) EPC/Chol Lipo DOX (5 mg/kg), 2) EPC/Chol Lipo DOX (5 mg/kg) + PSC 833 (100 mg/kg), 3) PEG2 0oo-DSPE/DSPC/Chol Lipo DOX (5 mg/kg), and 4) PEG20oo-DSPE/DSPC/Chol Lipo DOX (5 mg/kg) + PSC 833 (100 mg/kg). For comparison, free (non-encapsulated) DOX (5 mg/kg) treatments were also included, in the presence and absence of PSC 833 (100 mg/kg, p.o., 4 h before DOX). Blood was collected by cardiac puncture and placed into 157 EDTA coated microtainer tubes. Terminal blood and selected tissues were collected from all animals for lipid and DOX determinations. After blood collection via cardiac puncture, kidneys, tumors and liver were excised from each animal. Tissues were rinsed with PBS, pat dried on absorbent paper and weighed in pre-weighed 16 x 100 mm tubes. Samples were stored at -20°C pending analysis. A 10-30% homogenate of tissue samples in distilled water was prepared using a Polytron homogenizer (Kinematica, Switzerland). Tissue homogenates (0.2 ml) were incubated with Solvable, a tissue solubilizer (0.5 ml) at 50°C for 3 h. Following incubation, samples were brought to room temperature for 1 h to control foaming. 200 mM EDTA (50 ul), 30% hydrogen peroxide (200 ul), and 10N hydrochloric acid (25 ul), were then added and the mixture incubated at room temperature for 1 h. Five ml of the scintillation fluid, Pico-Fluor (Packard Instruments, Meridien, CT) was then added and samples were analyzed for radioactivity. Plasma samples were directly counted following addition of the scintillation fluid. All radioactivity determinations were performed using a TRI-CARB™ Model 1900 TR Liquid Scintillation Analyzer, Packard Instrumentation Company (Meridien, CT). The specimens obtained from the control animals were used as background samples. HPLC analysis of DOX and its metabolites were performed using a Waters model 45 and 6000 A solvent metering systems (Waters Associates, Milford, MA) with a Waters Model 710 WISP autosampler, a Waters Model 470 fluoresence detector (500 nm excitation wavelength, 580 nm emission wavelength), Waters 510 HPLC pump, and a Waters Millenium Version 2.1. software for data processing was used. A NEC Powermate SX Plus Computer (NEC Information Systems, Inc., Boxborough, MA) and a Waters 158 Systems Interface Module (Waters Associates) were used for data handling. A C18 guard column inlet filter (Rheodyne, Inc., Cotati, CA) was placed prior to the Nova-Pak11 C18 3.9 x 150 mm HPLC analytical column (Millipore Corporation, Milford, MA). Durapore 0.22 pm GVWP and 0.45 pm HVHP membrane filters were used to fdter solvents (Millipore Corporation, Bedford, MA). The HPLC assay of Andersen et al (1993) was used with minor modification. Briefly, sample extraction with acetonitrile was followed by isocratic elution from a C18 reverse phase column and quantified by endogenous fluorescence. The mobile phase consisted of a 16 mM ammonium formate buffer (pH 3.5)/acetone/isopropanol mixture (75:20:5) delivered at a rate of 1.0 ml/min. The column was maintained at 40 °C. Using this method, the retention times of DOXol, DOX, DOXone, and 7-deoxyDOXone, were 3.6 min, 5.9 min, 8.0 min, and 16.5 min, respectively, as shown in the representative chromatogram below (Figure 4.1). 159 6.00-9 B o M C o I 4.50 tc k o a o *» • a 3.00 DOXo! DOX DOXone 7-deoxyDOXone 10.00 20.00 Retent ion T ime (Minutes) Figure 4.1.: Representative chromatogram of DOX and DOX metabolite standards (DOXol, DOXone, and 7-deoxyDOXone) spiked in a 10% liver homogenate. Recoveries, using acetonitrile as the extraction solvent, from plasma over a concentration range of 0.05 - 10 ug/ml of DOX, DOXone, 7-deoxyDOXone, and DOXol, were between 80-110%. Specifically, DOX recoveries (extracted versus unextracted), at a concentration of 1 ug/ml, from plasma and tumor, liver, and kidney homogenates were 98.9 ± 3.6%, 99.1 ± 2.8%, 97.8 ± 4.0%, and 97.2 ± 0.6%, respectively. Five point calibration curves generated were characterized by: linearity over the 0.05 - 10 u,g/ml range, acceptable inter- and intra-day variability within 10% coefficient of variation (CV), and 20% CV at the limit of quantitation (50 ng/ml), at a signal to noise ratio >3. To protect DOX, and DOX metabolites from photodegradation, all procedures were shielded from direct exposure to light. In addition, DOX was found to be stable in the mobile-phase solvent mixture for at least 96 h (40 °C), on the FfPLC autosampler tray for at least 160 96 h, in reconstituted form at 4 °C for at least 10 days, and for at least four freeze-thaw cycles. The plasma data were modeled using WinNONLIN Version 1.5 pharmacokinetic software (Pharsight Corporation, CA), to calculate area under the curve (AUC), half-life (T1/2), and plasma clearance (CLP) according to standard equations (Gibaldi and Perrier, 1982). The data was fitted to one, two, and three compartment models. Appropriate models were selected on the basis of goodness of fit for each model. This included a visual examination of the distribution of residuals, rank, and Akaike's Information Criterion (Akaike, 1976). The trapezoidal rule was used to calculate the AUCs in tissue concentration-time profdes employing a computer software AUC, program provided by Dr. Wayne Riggs (Faculty of Pharmaceutical Sciences, University of British Columbia). Tissue DOX levels were corrected for blood volume to account for material residing in the vasculature of the tissues, as previously described in Section 3.2.8. Data are represented as mean ± standard error of the mean. Statistical analyses were performed using 2-sample t-test and statistical significance was set at p<0.05. Selected samples were assayed for free and liposome-associated DOX using Microcon-30 filters according to the equilibrium filtration method of Mayer and St.Onge (1995). Groups of female SCID/Rag2 mice bearing MDA435LCC6 MDR tumors received liposomal DOX (EPC/Chol or PEG-DSPE/DSPC/Chol) via tail vein injection. DOX encapsulated in 3H-labeled liposomes were prepared in saline to achieve a final dose of 5 mg/kg. Each mouse received 0.2 ml of the respective formulation. At 1, 4, and 24 h, mice were terminated, and blood collected by cardiac puncture and placed in a EDTA-coated microtainer. Samples were centrifuged (Centronics Silent Series S-103NAR Centrifuge) at 161 4°C, 15 min, 3000 rpm and aliquots of the plasma obtained were transferred to Microcon-30 reservoirs. Separation of free from liposomal and protein-bound drug was performed using Microcon-30 (0.5 ml capacity, Amicon Canada Limited, Oakville, Canada) ultrafiltration devices with a molecular weight cut off of 30,000 daltons. Microcon-30 samples were centrifuged at 4°C, 8000g for 20 min in a microcentrifuge (IEC Micromax Centrifuge, International Equipment Company, Needham Heights, MA). The ultrafiltrate was the processed for DOX by HPLC. 4.2.7. Confocal Microscopy and Imaging Studies Confocal images were collected on a Optiphot 2 research microscope (Nikon, Japan) attached to a confocal laser scanning microscope (MRC-600, BioRad Laboratories, Hercules, CA) using COMOS software (BioRad Laboratories). The laser line on the krypton/argon laser was 488 nm. Filterblock BHS was used to detect DOX (488 nm excitation, 515 nm emission). The numerical aperture was 0.75 on the x20 air objective and 1.2 on the x60 oil objective. The images were captured such that the xyz dimensions were 0.4 urn cubed (x20) and 0.2 (im pixel (x60). NIH Image version 1.61 was used for image analysis, and all images were based on maximum intensity projection. Projections made in the NTH Image were saved in TIFF format, then imported to Adobe Photoshop version 4.0 where the different fluorophore images were assigned to individual RGB channels and subsequently merged to provide the final image of the single or multiple sections. For confocal imaging studies, SCID/RAG2 mice bearing MDA43 5LCC6/MDR-1 tumors were treated with free DOX, EPC/Chol DOX, or PEG-DSPE/DSPC/Chol DOX (5 mg/kg) in the presence and absence of PSC 833 100 mg/kg (4 h before DOX). These 162 studies were performed in order to assess the bioavailability of DOX administered in liposome encapsulated form as well as the intracellular uptake of DOX for all formulations in the presence and absence of PSC 833. The evaluation of bioavailability relies on the fact that DOX encapsulated inside liposomes exhibiting transmembrane ion gradients is quenched with respect to its inherent fluorescence. At the indicated times following DOX administration, tumor tissues were asceptically dissected, perfused in PBS, and imaged fresh. Before imaging, thin pieces of liver lobes and tumors were placed on concave slides and observed under a 60x oil immersion lens. These were then viewed under the confocal microscope to determine DOX distribution characteristics. As controls, known amounts of non-encapsulated DOX, or PEG-DSPE/DSPC/Chol -DOX liposomes were infused into freshly isolated muscle tissues and viewed for DOX fluorescence. In order to confirm quenching with liposomal DOX preparations, we compared the fluorescence properties in spiked muscle tissue of PEG-DSPE/DSPC/Chol -DOX liposomes to that obtained with empty PEG-DSPE/DSPC/Chol liposomes of identical composition but with a fluorescent lipid label. The lipid label used was Di-I and was incubated with empty PEG-DSPE/DSPC/Chol liposomes at 60°C for 1 h with agitation. 163 4.3. Results 4.3.1. Toxicity In order to identify the MTD for single (day 1) and multiple (days 1, 5, 9) dosing regimens as well as to investigate the mechanisms of PSC 833 mediated increases in DOX toxicity, we conducted a 21-day dose range-finding toxicity studies with EPC/Chol, DSPC/Chol, and PEG-DSPE/DSPC/Chol liposomal formulations of DOX in the presence and absence of p.o. PSC 833 at a fixed dose of 100 mg/kg in female BDF1 non-tumor bearing healthy mice. The results of this study are presented in Table 4.1, where the body weight loss nadir on day 10 is represented, along with the MTDs (as defined in Materials and Methods) of day 1 and days 1, 5, 9 regimens. The three liposomal DOX formulations yielded toxicity characteristics that depended on their respective abilities of retaining DOX. The following subsections present the results for the single and multiple injections. 164 Table 4.1: Toxicity as a function of DOX dose for liposomal DOX formulations in single and multiple (days 1,5,9) injection schedules. Day 10 weight loss Day 10 weight loss (% Survival) (% Survival) Group Dose (mg/kg) Single Injection Dose (mg/kg) Multiple Injection - PSC 833 + PSC 833 - PSC 833 + PSC 833 EPC/Chol DOX 10 16.4 (100) 15 5.5 (33)* 20 20.0(67) 30 7.4 (100) 40 5.6 (100) 45 8.3 (100) 50 31.8 (0) 5 0.9 (100) 10 2.8(100) 23.1 (0) 15 1.7(100) 20 20.4 (0) 24.4 d6 (0) MTD 45 mg/kg 10 mg/kg 15 mg/kg 5 mg/kg DSPC/Chol DOX 40 3.6 (100) 50 10.8 (100) 60 4.9 (100) 35.6 (0) 70 11.7(100) 80 18.7 (100) 15 10.8 (100) 12.9 (100) 20 3.1 (100) 15.0(100) 25 16.0 (100) 13.5 (67) MTD 70 mg/kg 50 mg/kg 25 mg/kg 20 mg/kg PEG-DSPE/ DSPC/Chol DOX 40 15.2 (100) 45 20.7 (100) 50 22.7 (0) 60 5.3 (100) 70 8.5 (100) 80 12.6 (100) 90 18.7 (100) 15 7.3 (100) 9.8 (100) 20 6.5 (100) 14.8 (100) 25 9.3 (100) 14.8 (100) 30 4.4 (67)** 25.3 (0) MTD 80 mg/kg 40 mg/kg 25 mg/kg 25 mg/kg Data are group mean values (n = 3 mice/group) *Body weight loss nadir is on day 14 (-17.6%) with 33% survival. **Body weight loss nadir is on day 19 (-10.2%) with 67% survival. 165 4.3.1.1. Single Injection In the single injection (day 1) and in the absence of PSC 833, the MTD for EPC/Chol DOX was identified as 45 mg/kg, with 100% survival for the study period (21 days). For the formulations based on DSPC as the bulk phospholipid, the conventional (DSPC/Chol) and sterically stabilized (PEG containing) liposomal DOX the MTDs were considerably higher than the rapid drug-releasing EPC/Chol formulation. Specifically, the MTD was 70 mg/kg with 11.7% weight loss on day 10 for DSPC/Chol DOX and 80 mg/kg with 12.6% weight loss on day 10 for the PEG-DSPE/DSPC/Chol formulation. Compared to the more leaky EPC/Chol formulation, this represented 1.6-fold and 1.8-fold increases with DSPC/Chol and PEG-DSPE/DSPC/Chol liposomal DOX, respectively. In the presence of PSC 833, the MTD for the EPC/Chol DOX formulation was 10 mg/kg with a nadir in weight loss of 16.4%. Significant toxicity resulted when the dose was further increased. This reflects a 4.5-fold reduction in MTD induced by PSC 833. In comparison, the MTDs for the two saturated lipid formulations in the presence of PSC 833 were significantly higher than EPC/Chol DOX, i.e., 50 mg/kg for DSPC/Chol DOX and 40 mg/kg for PEG-DSPE/DSPC/Chol DOX. This represented a 5-fold increase in MTD for DSPC/Chol and a 4-fold increase for the PEG formulation, compared to EPC/Chol DOX, suggesting a greater ability of the non-leaky liposomal DOX formulations to circumvent toxicity increases caused by PSC 833 treatment. Specifically, PSC 833-co-administration necessitated a 1.4-fold reduction in MTD with DSPC/Chol DOX, and a 2-fold reduction with PEG-DSPE/DSPC/Chol liposomal DOX compared to the 4.5-fold reduction observed for the EPC/Chol formulation. 166 4.3.1.2. Multiple injection For the day 1, 5, 9 dosing schedule and in the absence of PSC 833, EPC/Chol DOX exhibited an MTD of 15 mg/kg, where the body weight loss was 1.7% with 100% survival (Table 4.1). Increasing the dose of 20 mg/kg was not tolerated by mice and resulted in 0% survival. For the two saturated lipid formulations, DSPC/Chol and PEG-DSPE/DSPC/Chol DOX, the MTD was identical at 25 mg/kg, representing a 1.7-fold increase in MTD compared to the more leaky EPC/Chol formulation. When EPC/Chol DOX was combined with PSC 833, significant toxicity resulted, necessitating a dose reduction by 3-fold, to 5 mg/kg to achieve well tolerated DOX doses. In contrast, PSC 833 caused very modest changes in the toxicity of DOX encapsulated in DSPC/Chol liposomes (MTD of 25 mg/kg and 20 mg/kg in the absence and presence of PSC 833, respectively) and no dose reduction was required for the PEG containing sterically stabilized liposomes (Table 4.1). 167 4.3.2. Efficacy The antitumor activity of the three types of liposomal DOX formulations was evaluated in vivo in the absence and presence of PSC 833 using the MDA435LCC6 human breast carcinoma xenograft solid tumor model. Figures 4.2 and 4.3 represent the antitumor efficacy, in the presence and absence of PSC 833, of non-encapsulated drug (Panel A), EPC/Chol DOX (Panel B), DSPC/Chol DOX (Panel C), and sterically stabilized PEG-DSPE/DSPC/Chol DOX system (Panel D), of two independent efficacy studies performed four months apart. The following sections will review the results obtained from each of these experiments. It may be mentioned here that day 1 represents the day treatments were first given (post-treatment). However, there was a 21 day waiting period to generate palpable solid tumors (post-inoculation), before treatments began, as described in Materials and Methods. The following sections will use post-treatment days for presentation of results. 4.3.2.1. Efficacy experiment # 1 When MDA435LCC6 cells (sensitive, WT or resistant, MDR) are inoculated in the mammary fat pads of SCID/Rag2 mice, solid tumors readily establish (tumor take rates >95%). Figure 4.2 (Panel A) shows that both MDR and WT untreated controls exhibit comparable tumor growth rates. Both groups were terminated on days 17 and 19 post-treatment, respectively, when tumor weights reached 0.59 ± 0.05 g and 0.58 ± 0.03 g for the MDR and WT tumors, respectively. These groups were not significantly different from each other. Treatment of WT tumors with non-encapsulated DOX at 7.5 mg/kg on days 1, 5, and 9 (Figure 4.2, Panel A), resulted in significant (p<0.05) tumor growth suppression until day 20, when tumors weighed 0.083 ± 0.02 g. These tumors, lacking PGP, eventually 168 grew to 0.35 ± 0.07 g by day 30. However, administration of non-encapsulated DOX at 7.5 mg/kg on days 1, 5, 9 in mice bearing MDR tumors did not cause any tumor growth suppression due to the overexpression of the drug efflux pump (Figure 4.2, Panel A). These tumors were also not significantly different from MDR untreated controls. When MDR tumor bearing mice were treated with non-encapsulated DOX at 3 mg/kg combined with PSC 833 (100 mg/kg), there was tumor growth inhibition until day 10, after which tumor growth rates were similar to untreated MDR controls as well as MDR tumors treated with non-encapsulated DOX alone at the MTD of 7.5 mg/kg, until day 19, when mice were terminated (Figure 4.2, Panel A). Specifically, for non-encapsulated DOX plus PSC 833, the tumor weight on day 10 was 0.08 ± 0.01 g, which grew to a weight of 0.45 ± 0.03 g prior to termination. Growth of MDR tumors treated with non-encapsulated DOX plus PSC 833 was significantly (p<0.05) different from both untreated MDR tumors as well as MDR tumors treated with non-encapsulated drug alone. In comparison to WT tumors treated with non-encapsulated DOX 7.5 mg/kg, the MDR tumor growth inhibition caused by non-encapsulated drug and PSC 833 was transient (Figure 4.2, Panel A). 169 Figure 4.2 (following page) Antitumor efficacy of free (F) (Panel A), EPC/Chol DOX (Panel B), DSPC/Chol DOX (Panel C), and PEG-DSPE/DSPC/Chol DOX (Panel D) against MDA435LCC6 WT or MDR1 human xenograft solid tumors in the absence and presence of co-administered PSC 833. Untreated WT and MDR controls as well as free (F) DOX treated WT and MDR tumors are presented in Panel A. MDA435LCC6 tumors were grown on mammary fat pads of female SCID/Rag2 mice. Oral PSC 833 (100 mg/kg) and i.v. DOX treatments were initiated once tumors were established (20-100 mg) and were given on days 1, 5, and 9 at the indicated doses of non-encapsulated and liposomal DOX. PSC 833 was administered 4 hours prior to DOX injection. Data are expressed as mean ± standard error of the mean (n = 8/group). For legends, see individual panels. 170 pod hH W 1 4 * Q a> E I—« I—"^d W W o s. ra O 0) E (B) jqBiaM-ioujni (B) )i|8i3M J o i u n i 171 When MDR tumors were treated with liposomal DOX formulations in the presence and absence of PSC 833 (Panels B, C, and D of Figure 4.2), varying degrees of tumor growth suppression were observed. As seen in Figure 4.2, Panel B, EPC/Chol DOX alone, at a dose of 5 mg/kg, was unable to induce substantial inhibition of MDR tumor growth, with tumors weighing 0.57 ± 0.05 g on day 21. In the absence of PSC 833, EPC/Chol DOX closely resembled non-encapsulated DOX treatment at 7.5 mg/kg, the two groups being not significantly different from one another (compare 0.57 ± 0.05 g for EPC/Chol DOX and 0.59 ± 0.05 g for non-encapsulated DOX on day of termination). In the presence of PSC 833, however, EPC/Chol DOX at 5 mg/kg treatment closely resembled EPC/Chol DOX alone until day 10 in terms of tumor growth. However, after day 10 mice exhibited physical manifestation of toxicity (indicated by body weight loss) and mortality, with 25% survival on day 13 (Figure 4.2, Panel B), indicating a need to reduce the dose of EPC/Chol DOX. Consequently, tumor bearing SCID/Rag2 mice were administered EPC/Chol DOX doses in a stepwise dose escalation manner from 2 mg/kg to 4 mg/kg, in the presence of PSC 833 to determine the MTD of this combination. In the second efficacy experiment (see Section 4.3.2.2.), a dose of 3 mg/kg EPC/Chol DOX in combination with PSC 833 was identified as the MTD. Consequently, this treatment group was repeated with a n=16 tumors/group in the second efficacy experiment (see Section 4.3.2.2. for description). Figure 4.2, Panel C illustrates the tumor growth inhibition of DSPC/Chol liposomal DOX, in the presence and absence of PSC 833. In the absence of PSC 833, DSPC/Chol liposomal DOX caused a modest reduction in tumor growth. The tumor growth inhibition caused by DSPC/Chol DOX was significant (p<0.05) until day 11, after which tumor 172 growth was characterized by a rate similar to MDR untreated controls, although the tumor growth inhibition profiles with both groups (MDR untreated and MDR treated with DSPC/Chol DOX) were significantly (p<0.05) different from one another. Tumor weights for MDR tumors treated with DSPC/Chol DOX alone were 0.49 ± 0.06 g on day 30. In contrast, DSPC/Chol DOX in the presence of PSC 833 afforded significant MDR modulation, where tumor weights approached 0.33 ± 0.04 g on day 30. The tumor growth inhibition caused by a combination of DSPC/Chol DOX and PSC 833 was significantly (p<0.05) different from the following treatment groups: MDR tumors treated with DSPC/Chol DOX alone, MDR untreated tumors, MDR tumors treated with non-encapsulated DOX, and MDR tumors treated with non-encapsulated DOX and PSC 833. Panel D in Figure 4.2 illustrates the tumor growth inhibition caused by PEG-DSPE/DSPC/Chol formulations in the presence and absence of PSC 833. In the absence of PSC 833, PEG-DSPE/DSPC/Chol DOX caused a significant (p<0.05) reduction in tumor growth until day 10, where tumor weights approached 0.11 ± 0.01 g (compared to a tumor weight of 0.03 ± 0.002 g on day 1), after which tumor growth rates increased and compared closely with untreated MDR tumors. Tumor weights for this group were 0.45 ± 0.04 g on day 30. In the presence of PSC 833, PEG-DSPE/DSPC/Chol DOX significantly (p<0.05) inhibited tumor growth and the tumor growth suppression was not significantly different from WT tumors treated with free DOX at its MTD. Specifically, the tumor weights on day 17 for PEG-DSPE/DSPC/Chol DOX in presence of PSC 833 were 0.1 ± 0.01 g compared to 0.08 ± 0.02 g for WT tumors treated with non-encapsulated DOX. PEG-DSPE/DSPC/Chol DOX and PSC 833 provided optimal suppression and delay of MDR tumor growth. This suppression of tumor growth for PEG-DSPE/DSPC/Chol DOX 173 and PSC 833 was significantly (p<0.05) different from the following treatments: MDR untreated controls, MDR tumors treated with non-encapsulated DOX (with and without PSC 833), MDR tumors treated with EPC/Chol DOX, as well as MDR tumors treated with DSPC/Chol DOX. 4.3.2.2. Efficacy experiment # 2 When MDA435LCC6 cells (sensitive, WT or resistant, MDR) were inoculated in the mammary fat pads of SC1D/Rag2 mice in the repeat experiment, solid tumors were readily established as shown in Figure 4.3 where both MDR and WT untreated controls exhibited comparable tumor growth rates, with exponential growth phase occurring between days 5 and 14 (r = 0.99 and slope of linear regression line = 0.067 for WT versus r=0.99 and slope of linear regression line = 0.056 for MDR). Both groups were terminated on day 18, when tumor weights reached 0.84 ± 0.05 g and 0.76 ± 0.05 g for the WT and MDR tumors, respectively. These groups were not significantly different from each other. Treatment of WT tumors with non-encapsulated DOX at 7.5 mg/kg on days 1, 5, and 9 (Figure 4.3, Panel A), resulted in significant (p<0.05) tumor growth suppression until day 20, when tumors weighed 0.088 ± 0.01 g. These tumors, lacking PGP, eventually grew to 0.68 ± 0.09 g by day 40. However, administration of non-encapsulated DOX at 7.5 mg/kg on days 1, 5, 9 in mice bearing MDR tumors did not cause any tumor growth suppression due to the overexpression of the drug efflux pump (Figure 4.3, Panel A). These tumors were also not significantly different from MDR untreated controls. These results demonstrate that MDA435LCC6 WT and MDR tumors exhibit well defined tumor growth as well as clear differentiation of sensitive and resistant tumor types, which can be utilized to compare MDR reversal properties of various liposomal treatment groups. 174 When MDR tumor bearing mice were treated with non-encapsulated DOX at 3 mg/kg (MTD with PSC 833) combined with PSC 833 (100 mg/kg), there was partial tumor growth inhibition until day 11. After this time, tumor growth rates were similar to untreated MDR controls as well as MDR tumors treated with non-encapsulated DOX alone up to day 18, at which time the mice were terminated (Figure 4.3, Panel A). For non-encapsulated DOX plus PSC 833, the tumor weight on day 11 was 0.2 ± 0.05 g, which was approximately 2-fold lower than MDR tumors treated with non-encapsulated DOX alone (0.42 ± 0.01 g) as well as untreated MDR tumors (0.45 ± 0.05 g). Growth of MDR tumors treated with non-encapsulated DOX plus PSC 833 was significantly (p<0.05) different from both untreated MDR tumors as well as MDR tumors treated with non-encapsulated drug alone. However, in comparison to WT tumors treated with non-encapsulated DOX 7.5 mg/kg, the MDR tumor growth inhibition caused by non-encapsulated drug and PSC 833 was transient (Figure 4.3, Panel A). 175 — A - M D R traalad wRh F-Dox 7.S mg/kg — A - M D R treated wMt F-Dox 7.5 mg/kg - • — M D R traalad w t h D S P C / C h o l Dex S mg/kg - • — M D R treated with P E Q - D S P E / D S P C / C h o l Dox B r n g f t g 0 5 10 15 20 25 30 0 5 10 15 20 25 40 45 Time (Days) Time (Days) Figure 4.3: Antitumor efficacy of free (F) DOX (Panel A), EPC/Chol DOX (Panel B), DSPC/Chol DOX (Panel C), and PEG-DSPE/DSPC/Chol DOX (Panel D) against MDA435LCC6 WT or MDR1 tumors ± PSC 833. MDA435LCC6 tumors were grown on mfp of female SCID/Rag2 mice. Oral PSC 833 (100 mg/kg) and i.v. DOX treatments were initiated once tumors were established (20-100 mg) and were given on days 1, 5, and 9 at the indicated doses of non-encapsulated and liposomal DOX. PSC 833 was administered 4 h prior to DOX injection. Data are expressed as mean ± s.e.m. (n = 8/group; for EPC/Chol DOX + PSC 833, n = 16/group). For legends, see panels. 176 When MDR tumors were treated with liposomal DOX formulations in the presence and absence of PSC 833 (Panels B, C, and D of Figure 4.3), antitumor activities similar to those observed in the first experiment were observed. As seen in Figure 4.3, Panel B, EPC/Chol DOX alone, at a dose of 5 mg/kg, was unable to induce substantial inhibition of MDR tumor growth, with tumors weighing 0.86 ± 0.2 g on day 18. In the absence of PSC 833, EPC/Chol DOX closely resembled non-encapsulated DOX treatment at 7.5 mg/kg, the two groups being not significantly different from one another (compare 0.86 ± 0.2 g for EPC/Chol DOX and 0.67 ± 0.1 g for non-encapsulated DOX on day 18). In the presence of PSC 833, however, EPC/Chol DOX at 3 mg/kg treatment (n=16 tumors/group) closely resembled EPC/Chol DOX alone until day 12 in terms of tumor growth, at which time tumor growth inhibition continued to day 20 (Figure 4.3, Panel B), indicating improved but incomplete modulation of MDR. This MDR modulation caused by EPC/Chol DOX and PSC 833 was significantly better than that caused by non-encapsulated DOX and PSC 833, presumably due to better DOX tumor accumulation via EPC/Chol liposome delivery. The tumor growth inhibition caused by EPC/Chol DOX and PSC 833 was also significantly (p<0.05) different from untreated MDR tumors and MDR tumors treated with non-encapsulated DOX. Figure 4.3, Panel C illustrates the tumor growth inhibition of DSPC/Chol liposomal DOX, in the presence and absence of PSC 833. Similar to results previously seen in the P388/ADR solid tumor model and in experiment 1 of this tumor model, DSPC/Chol liposomal DOX caused a modest reduction in tumor growth in the absence of PSC 833. The tumor growth inhibition caused by DSPC/Chol DOX was significant (p<0.05) until day 12, after which tumor growth was characterized by a rate similar to MDR untreated 177 controls, although the tumor growth inhibition profiles with both groups (MDR untreated and MDR treated with DSPC/Chol DOX) were significantly (p<0.05) different from one another. Tumor weights for MDR tumors treated with DSPC/Chol DOX alone were 0.51 ± 0 . 1 g on day 20. In contrast, DSPC/Chol DOX in the presence of PSC 833 afforded significant (p<0.05) MDR modulation, where tumor weights were 0.25 ± 0.05 g on day 20. The improvement of tumor growth inhibition caused by the combination of DSPC/Chol DOX and PSC 833 was significantly (p<0.05) different from MDR tumors treated with DSPC/Chol DOX alone, MDR tumors treated with non-encapsulated DOX, and MDR tumors treated with non-encapsulated DOX and PSC 833. Furthermore, the tumor growth inhibition caused by DSPC/Chol DOX and PSC 833 was statistically significant (p<0.05) compared to EPC/Chol DOX and PSC 833. Panel D in Figure 4.3 illustrates the tumor growth inhibition caused by PEG-DSPE/DSPC/Chol DOX formulations in the presence and absence of PSC 833. In the absence of PSC 833, PEG-DSPE/DSPC/Chol DOX caused a significant (p<0.05) reduction in tumor growth until day 11, where tumor weights were 0.1 ± 0.01 g (compared to a tumor weight of 0.05 ± 0.01 g on day 1), after which tumor growth rates increased and compared closely with untreated MDR tumors. Tumor weights for this group were 0.39 ± 0.02 g on day 20. In the presence of PSC 833, PEG-DSPE/DSPC/Chol DOX significantly (p<0.05) inhibited tumor growth and the tumor growth suppression was not significantly different from WT tumors treated with free DOX at its MTD. Specifically, the tumor weights on day 20 for PEG-DSPE/DSPC/Chol DOX in presence of PSC 833 were 0.1 ± 0.02 g compared to 0.088 ± 0.01 g for WT tumors treated with non-encapsulated DOX. PEG-DSPE/DSPC/Chol DOX and PSC 833 provided optimal 178 suppression and delay of MDR tumor growth. This suppression of tumor growth for PEG-DSPE/DSPC/Chol DOX and PSC 833 was statistically significantly (p<0.05) different from MDR tumors treated with non-encapsulated DOX (with and without PSC 833), MDR tumors treated with EPC/Chol DOX (with and without PSC 833), as well as MDR tumors treated with DSPC/Chol DOX (with and without PSC 833). The results of the second efficacy experiment compared favorably with the results of the first study. The following Table (Table 4.2.) compares the relative tumor weights and percentage T/C, as described in Materials and Methods, of the results from these two experiments. 179 Table 4.2.: Summary of relative tumor weights and percent T/C of free and liposomal DOX treatments in the presence and absence of PSC 833. Treatment group Relative Weight8 T/C (%)b Expt. 1 Expt. 2 Expt. 1 Expt. 2 WT - Control 914 1630 100 100 WT treated with free DOX 7.5 mg/kg 277 275 30 17 MDR - Control 946 1775 100 100 MDR treated with free DOX 7.5 mg/kg 679 1464 72 82 MDR treated with free DOX 3 mg/kg and PSC 833 100 mg/kg 535 1752 56 99 MDR treated with EPC/Chol DOX 5 mg/kg 549 1227 58 69 MDR treated with EPC/Chol DOX 5 mg/kg and PSC 833 100 mg/kg Lethal dose - Lethal dose -MDR treated with EPC/Chol DOX 3 mg/kg and PSC 833 100 mg/kg - 1084 - 62 MDR treated with DSPC/Chol DOX 5 mg/kg 848 1290 89 73 MDR treated with DSPC/Chol DOX 5 mg/kg and PSC 833 100 mg/kg 287 400 30 23 MDR treated with PEG-DSPE/DSPC/Chol DOX 5 mg/kg 629 620 66 35 MDR treated with PEG-DSPE/DSPC/Chol DOX 5 mg/kg and PSC 833 100 mg/kg 244 533 26 30 a: Relative tumor weight (RW) was calculated using the equation: (Mean tumor weight on day 15/mean tumor weight on day 1) x 100 b: Treated-to-control ratio (T/C, %) was calculated using the equation: (RW of treatment group/RW of respective control) x 100 For the calculation of the relative weight, the ratio of the mean tumor weight on day 15 divided by that on day 1 (and multiplied by 100) was used to compare the tumor growth rates of different treatment groups between the two individual experiments. This is based on the consideration that the tumor size on the first day of treatment is 100%, and all changes in tumor size expressed in percentage of control, whereas T/C (%) was defined as the percentage ratio of relative tumor weight of treated groups (T) to that of the respective 180 control group (C). It should be mentioned that for groups which exhibited prolonged tumor growth suppression, the experiment was continued until day 30 for experiment 1 and day 40 for experiment 2. Although there were variations in tumor growth rates (and therefore relative weights) between the two independent studies, the trends in tumor growth suppression were similar for the treatment groups, as seen from comparable T/C, with negligible reductions in T/C for MDR tumors treated with either free DOX alone or in combination with PSC 833, EPC/Chol DOX alone or in combination with PSC 833, or DSPC/Chol DOX alone. The greatest reductions in tumor growth were observed with the two saturated lipid liposomal formulations in the presence of PSC 833. 4.3.3. Pharmacokinetics and Tissue Distribution A comprehensive pharmacokinetic evaluation was performed to correlate toxicity and efficacy data with plasma and tissue (liver, kidney, tumor) DOX and DOX metabolite distribution properties determined using HPLC analysis. Since the comparison of DSPC/Chol DOX and non-encapsulated (free) DOX pharmacokinetics and tissue distribution properties in the presence and absence of PSC 833 has already been extensively characterized (Chapter 3; Krishna and Mayer, 1997) and toxicity and efficacy properties of these formulations were very comparable in the current models, the results presented in this Chapter focused on comparisons between a liposomal system (EPC/Chol DOX) which leaks a significant portion of entrapped drug into the circulation to a sterically stabilized saturated lipid liposome system (PEG-DSPE/DSPC/Chol DOX) which exhibits negligible drug release in plasma compartment and improved accumulation in tumor tissue. For comparison, free DOX 5 mg/kg treatments (+ PSC 833) were performed and data presented for DOX concentrations quantitated in the plasma and the tumor. By comparing 181 pharmacokinetic properties of the least toxic/most efficacious liposomal formulation with one which exhibits increased release of the drug, it was predicted that important DOX distribution and metabolism properties may be revealed which impact on drug pharmacokinetic alterations caused by the MDR modulator. 4.3.3.1. Plasma drug kinetics Figure 4.4 presents the DOX (Panel A) plasma concentrations after i.v. administration of free (non-encapsulated) DOX and the two liposomal DOX formulations at a DOX dose of 5 mg/kg. Following administration of free DOX at 5 mg/kg, DOX is rapidly eliminated from the circulation. Concentrations of DOX beyond 4 h were below assay detection limits. The concentration-time profile was characterized by a Cmax of 1.5 ± 0.1 pig/ml and an AUC of 4.4 u.g.h./ml (Figure 4.4A). However, when PSC 833 was co-administered with free DOX at 5 mg/kg, DOX elimination from plasma was characterized by a prolonged terminal elimination phase (Figure 4.4A). PSC 833 caused significant (p<0.05) increases in Cmax and AUC of DOX. Specifically, a Cmax of 3.9 ± 0.5 ug/ml and an AUC of 48.1 u.g.h./ml were obtained when free (non-encapsulated) DOX was co-administered with PSC 833. This approximately 11-fold increase in DOX AUC caused by PSC 833 is consistent with the 10-fold increase in DOX AUC for the free DOX-PSC 833 combination observed in the P388/ADR solid tumor model described in Chapter 3. As shown in Figure 4.4A, DOX elimination from plasma for both EPC/Chol and PEG-DSPE/DSPC/Chol DOX systems exhibit a monophasic elimination profile characterized by a one-compartment model with first order elimination. While EPC/Chol DOX plasma concentration-time profile illustrates rapid elimination of the drug within 24 h, PEG-DSPE/DSPC/Chol DOX is characterized by a prolonged circulation life-time with 182 over 20% remaining at 24 h. In all cases, parent DOX was the only detectable entity with no indication of any metabolites present. In the presence of PSC 833, although the elimination profile remained largely monophasic, DOX concentrations at earlier time points were significantly (p<0.05) increased for EPC/Chol liposomes whereas no such PSC 833 effect was observed for PEG containing DSPC/Chol liposomes. This is in contrast to observations for non-encapsulated DOX which demonstrated increases in the terminal elimination phase in the presence of PSC 833 (Krishna and Mayer, 1997). Co-administration of PSC 833 and EPC/Chol DOX increases the AUC of DOX by 2.6-fold and Cmax by 2.3-fold, accounting for the 40% reduction in plasma clearance (Table 4.3; significant at p<0.05). In contrast, PSC 833 did not cause dramatic changes in the pharmacokinetics of DOX encapsulated in PEG-DSPE/DSPC/Chol liposomes, with small (36% and 25%) increases in Cmax and AUC, respectively (Table 4.3). PEG-DSPE/DSPC/Chol DOX was able to provide approximately a 15-fold increase in DOX AUC compared to EPC/Chol DOX liposomes, in the absence of PSC 833 and an 8-fold DOX AUC increase in the presence of PSC 833 (significant at p<0.05). Panel B of Figure 4.4. shows the elimination of liposomal lipid from plasma was monitored using the non-exchangeable, non-metabolizable lipid marker, 3 H -cholesterylhexadecyl ether (Derksen et al., 1987). This label can therefore be used to monitor the levels of intact liposomes circulating in plasma. As seen in Figure 4.4B, similar to DOX pharmacokinetics, the liposomal lipid elimination is monophasic characterized by a one-compartment model with first-order elimination. PEG-DSPE/DSPC/Chol liposomes afforded 2.5-fold (significant at p<0.05) higher plasma lipid levels compared to EPC/Chol liposomes, reflecting a longer circulation lifetime. Specifically, the plasma clearance (CLP) 183 for PEG-DSPE/DSPC/Chol liposomes was 2.4-fold lower than EPC/Chol, and the half-life (T1/2) was 2.2-fold higher for PEG-DSPE/DSPC/Chol liposomes compared to EPC/Chol, demonstrating the utility of PEG to significantly increase circulation longevity. PSC 833 did not significantly alter the liposomal lipid elimination kinetics for either liposomal formulation, as demonstrated by the small changes in half-life. The drug-to-lipid ratio can be used to monitor liposome-associated drug over time and consequently, any reductions in this ratio may be used to determine drug leakage. The drug-to-lipid ratio remained constant for PEG-DSPE/DSPC/Chol liposomal DOX in the presence and absence of PSC 833 (identical at 0.2 at 30 min, 0.18 at 24 h, 0.1 at 48 h, and 0.1 at 72 h). Interestingly, the drug-to-lipid ratio for EPC/Chol DOX, in the absence of PSC 833, was 0.08 at 30 min, indicating that more than 50% of the drug was already released from the system and cleared from the circulation at 0.5 h. Furthermore, the drug-to-lipid ratio for the EPC/Chol DOX system followed a biphasic elimination profile, typified by an early sharp decline (0.2 at 0 h to 0.03 at 4h, and <0.01 at 24 h), and a delayed terminal phase at very low drug-to-lipid ratios. This compared favorably with previous pharmacokinetic observations with these systems (Mayer et al., 1989). In the presence of PSC 833, however, EPC/Chol DOX drug-to-lipid ratios were higher (0.2 at 0 h, 0.14 at 30 min, 0.07 at 16 h), which could arise from altered retention within EPC/Chol liposomes or increased levels of non-liposomal drug due to reduced free drug elimination rates. No such PSC 833 related changes were observed with the more drug retentive PEG-DSPE/DSPC/Chol formulation. 184 - O - E P C / C h d D O X S m g / k g — O — E P C / C h o l D O X 6 mg /kg - • - P E G - D S P E / D S P C / C h o l D O X 5 m g * g — Q — P E G - D S P E / D S P C / C h o l D O X 6 m g A f l E P C / C h o l D O X S n t g f t g + P S C 833 - • - E P C / C h o l D O X 6 mg /kg • P S C 833 I • I • I • I • I • I • I • I ' I • I ' I • I • I • I 1 I • I • I 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 f j m e (Hours) Time (Hours) Figure 4.4: DOX (Panel A) and liposomal lipid (Panel B) concentration-time profdes in plasma following administration of free DOX 5 mg/kg in the absence (open triangles) or presence of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled triangles); EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence of PSC 833 (administered 4 hours prior 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 of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled squares). Mice bearing 200 - 500 mg solid MDA435LCC6 tumors were terminated at the indicated times and plasma samples were processed and quantified for DOX and metabolites by HPLC and liposomal lipid by LSC as described in Materials and Methods. Data are expressed as mean ± standard error of the mean (n = 3 mice/time point; n = 24/group). 185 Table 4.3 A : Effect of PSC 833 on plasma DOX pharmacokinetic parameters when DOX is administered in liposome-encapsulated forms. Group AUCo-t pg.h./ml Ti/2 h Cmax Ug/ml C L P ml/h EPC/Chol DOX 5 mg/kg 97.9 1.9 35.8 1.0 EPC/Chol DOX 5 mg/kg + PSC 833 100 mg/kg 256.9* 2.1 83.8* 0.39* PEG2ooo-DSPE/DSPC/Chol DOX 5 mg/kg 1523.6 9.9 107.1 0.07 PEG2ooo-DSPE/DSPC/Chol DOX 5 mg/kg + PSC 833 100 mg/kg 1905.2 9.0 146.8 0.05 * Significantly different from the no PSC 833 group (p<0.05; 2-sample t-test) Table 4.3B: Effect of PSC 833 on plasma liposomal lipid pharmacokinetic parameters when DOX is administered in liposome-encapsulated forms. Group AUCo-, u£.h./ml Tin h Cmax Ug/ml C L P ml/h EPC/Chol DOX 5 mg/kg 5144.8 8.5 421.9 0.12 EPC/Chol DOX 5 mg/kg + PSC 833 100 mg/kg 4540.6 5.9 535.9 0.13 PEG20oo-DSPE/DSPC/Chol DOX 5 mg/kg 12158.4 16.8 501.2 0.05 PEG20oo-DSPE/DSPC/Chol DOX 5 mg/kg + PSC 833 100 mg/kg 12111.5 12.2 691.2 0.05 NB: Data are mean values (n = 3 mice/time point; n = 24/group) Pharmacokinetic abbreviations: area under the concentration-time curve, A U C ; elimination half-life, Ty2; peak concentration achieved, Cmax; and plasma clearance, C L p . 186 In order to determine the amount of free drug in the plasma following i.v. injection of liposomal DOX as well as to determine whether higher levels of free drug are present when EPC/Chol DOX and PSC 833 are combined, plasma from mice treated with EPC/Chol and PEG-DSPE/DSPC/Chol DOX formulations (at a DOX dose of 5 mg/kg), in the presence and absence of PSC 833 (at a dose of 100 mg/kg, p.o.), were processed using Microcon-30 filters as described in Materials and Methods. The data are summarized in Table 4.4. Free drug was below assay detection limits following administration of EPC/Chol and PEG-DSPE/DSPC/Chol DOX formulations in the absence of PSC 833 at 1, 4 or 24 h. Further, no free drug was present (i.e., below assay detection limits) for PEG-DSPE/DSPC/Chol DOX in the presence of PSC 833. This is consistent with the lack of a PSC 833-mediated changes in plasma DOX concentrations and the unchanged drug-to-lipid ratio with the PEG-DSPE/DSPC/Chol system. However, free drug levels were increased for EPC/Chol DOX in the presence of PSC 833, particularly at early time points where a 2.6-fold increase in plasma DOX concentrations were observed when EPC/Chol DOX was co-administered with PSC 833. Given that the equilibrium between free DOX and protein-bound DOX in plasma is approximately 8:2 (Mayer and St-Onge, 1995), the 1 h and 4 h free DOX determinations following EPC/Chol DOX administration could reflect approximately 32% and 12%, respectively, of the total plasma DOX concentration existing in non-encapsulated form. 187 Table 4.4: Evaluation of free DOX equivalents in mouse plasma using Microcon-30" Group Free DOX Equivalents* Microcon (ug/ml) % total in plasma* 1 h 4h 24 h 1 h 4h 24 h EPC/Chol DOX ND ND ND NA NA NA PEG-DSPE/DSPC/Chol DOX ND ND ND NA NA NA EPC/Chol DOX + PSC 833 4.318 0.652 ND 8.1 2.8 NA PEG-DSPE/DSPC/Chol DOX + PSC 833 ND ND ND NA NA NA a: SCID/Rag2 mice were treated with 5 mg/kg liposomal DOX preparation alone or in combination with PSC 833 (100 mg/kg, p.o.). At 1, 4, and 24 h, blood samples were collected by cardiac puncture, centrifuged, and the plasma processed with Microcon-30 filters as described in Materials and Methods. b: EPC/Chol liposomal DOX and PEG2ooo-DSPE/DSPC/Chol liposomal DOX containing 3H-cholesterylhexadecyl ether as a lipid marker were injected i/v at a dose of 5 mg/kg. c: % Total in plasma = (free drug equivalents/amount of drug per ml plasma at corresponding time point) x 100 ND: Not detectable, samples were below HPLC limit of quantitation, i.e., 50 ng/ml; NA: Not applicable 188 4.3.3.2. Tumor drug kinetics Figure 4.5 A illustrates tumor DOX concentrations following administration of free (non-encapsulated) DOX 5 mg/kg in the absence and presence of PSC 833. The tumor DOX accumulation following administration of free DOX 5 mg/kg was characterized by a C ^ of 6.6 ± 1.4 pg/g and a DOX AUC of 103.9 pg.h/g (Table 4.5A). Coadminstration of PSC 833 with free DOX at 5 mg/kg resulted in a significant (p<0.05) 2.5-fold increase in DOX AUC. In MDR tumors (Figure 4.5, Panel A), the more leaky EPC/Chol DOX formulation alone at a dose of 5 mg/kg provided modest DOX accumulation, with a C m a x of 12.3 ± 0.6 pg/g and total drug exposure of 811.4 pg.h/g (Table 4.5A). In contrast, the long circulating PEG-DSPE/DSPC/Chol formulation was able to provide greater accumulation of DOX in the tumor, substantiated by the 2.6-fold increase in Cmax and a 2.5-fold increase in AUC. Compared to free DOX treatment at 5 mg/kg, administration of DOX encapsulated in EPC/Chol and PEG-DSPE/DSPC/Chol liposomes (at identical doses) resulted in significant (p<0.05) 7.8-fold and 19.6-fold increases in DOX AUC, respectively (Table 4.5A), indicating effective DOX delivery to the tumor using liposome carriers. When PSC 833 (100 mg/kg) was co-administered with EPC/Chol DOX at 5 mg/kg, a small increase in tumor accumulation was observed, as reflected by the 1.8-fold increase in Cma X and a 1.5-fold increase in AUC (Figure 4.5; Table 4.5A). An interesting observation was the presence of the DOX aglycone metabolite, DOXone, at early time points when PSC 833 was given in conjunction with EPC/Chol DOX, indicating possible inhibition of DOX and DOXone elimination caused by the MDR modulator. No other metabolites, however, were recovered in any of the other treatment groups. 189 Figure 4.5: Tumor DOX (Panel A) and liposomal lipid (Panel B) concentration-time profiles following administration of free DOX 5 mg/kg in the absence (open triangles) or presence of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled triangles); EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence of PSC 833 (administered 4 hours prior 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 of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled squares). Mice bearing 200 - 500 mg solid MDA435LCC6 tumors were terminated at the indicated times and tumor samples were processed and quantified for DOX and metabolites by HPLC and liposomal lipid by LSC as described in Materials and Methods. Data are expressed as mean ± standard error of the mean (n = 3/time point; n = 24/group). 190 Table 4.5A: Tumor DOX distribution characteristics following administration of free or liposomal (EPC/Chol and PEG-DSPE/DSPC/Chol) DOX formulations in the presence and absence of PSC 833 \ Group T u m o r Cmax (Pg/g) A U C2 (M-g-h/g) Free DOX (5 mg/kg) 6.6+1.4 103.93 Free DOX (5 mg/kg) + PSC 833 14.1+4.4 253.93* EPC/Chol Dox (5 mg/kg) 12.3+0.6 811.4 EPC/Chol Dox (5 mg/kg) + PSC 833 22.6 ±2 .9 1228.9 PEG-DSPE/DSPC/Chol Dox (5 mg/kg) 32.8 ±2.65 2015.7 PEG-DSPE/DSPC/Chol Dox (5 mg/kg) + PSC 833 56.8 ±5.3 3104.2 1: Values represent intact DOX determined using HPLC analysis; data are mean ± standard error of the mean (n = 3 mice/time point; n = 24 mice/group). 2: AUC 0-72 was calculated using the trapezoidal rule. 3: AUC value represents 0 - 24 h time period. * Significantly different from the no PSC 833 group at p<0.05. 191 When DOX encapsulated in PEG-DSPE/DSPC/Chol liposomes is combined with PSC 833, a Cmax of 56.8 ± 5.3 ug/g and an AUC of 3104.2 ug.h/g was obtained. PSC 833 caused a 1.7-fold increase in Cmax and 1.5-fold increase in AUC for DOX administered in PEG-DSPE/DSPC/Chol liposomes. In both the presence and absence of PSC 833, PEG-DSPE/DSPC/Chol liposomes provided approximately 2.5-fold increases in DOX C m a X and AUC compared to EPC/Chol liposomes. Liposomal lipid levels in tumors were also monitored in order to determine whether increased DOX tumor concentrations associated with PSC 833 co-administration were due to elevated accumulation of the liposomes themselves in the tumor or rather increased retention of DOX in the tumor resulting from PGP inhibition by PSC 833. For both EPC/Chol and PEG-DSPE/DSPC/Chol liposomal DOX formulations, liposomal lipid levels (Figure 4.5, Panel B), observed in the tumors were comparable in the presence and absence of PSC 833. Specifically, for PEG-DSPE/DSPC/Chol system, a C ^ of 455.3 + 32.6 ug lipid/g and an AUC of 23788 pig lipid.h/g was obtained in the absence of PSC 833, whereas a Cmax of 516.6 ± 135.6 ug lipid/g and an AUC of 24883 ug lipid.h/g was obtained in the presence of the MDR modulator. For EPC/Chol system, a Cmax of 270.4 ± 12.9 ug lipid/g and an AUC of 16224 pig lipid.h/g was obtained in the absence of PSC 833, whereas a C^x of 349.5 ± 109.2 ug lipid/g and an AUC of 18140 ug lipid.h/g was obtained in the presence of the MDR modulator (Figure 4.5.B). Compared to EPC/Chol liposomes, PEG-DSPE/DSPC/Chol provided a 1.5-fold higher liposomal lipid accumulation in the tumor. Further, since liposomal lipid levels were comparable in the presence and absence of PSC 833 for each of the liposomal DOX formulation, the 192 increased DOX accumulation in the tumor caused by PSC 833 co-administration appeared consistent with greater drug retention at the tumor site following PGP blockade. Figure 4.5: Liver DOX (Panel C), and kidney DOX (Panel D) concentration-time profiles following administration of EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence of PSC 833 (administered 4 hours prior 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 of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled squares). Mice bearing 200 - 500 mg solid MDA435LCC6 tumors were terminated at the indicated times and liver and kidney samples were processed and quantified for DOX and metabolites by FfPLC as described in Materials and Methods. Data are expressed as mean + standard error of the mean (n = 3 mice/time point; n = 24/group). 193 4.3.3.3. Liver drug kinetics Parent (intact) drug Table 4.5B and Figure 4.5 (Panel C) summarizes the DOX C^x and AUC values for various treatment groups in the liver. When EPC/Chol DOX at 5 mg/kg was administered alone the liver DOX C m a x was 43.1 ± 5.4 ug/g and the AUC was 1511.6 ug.h/g. In the presence of PSC 833, while there was a small increase in C m a x of 1.3-fold, a doubling in AUC was observed (Table 4.5B). PEG-DSPE/DSPC/Chol DOX at 5 mg/kg alone resulted in a C m a x of 41.7 ± 6.8 ug/g and an AUC of 1715.7 ug.h/g. Interestingly, the combination of PEG-DSPE/DSPC/Chol DOX and PSC 833 resulted in a 1.5-fold reduction in C^x and a 1.8-fold reduction in AUC compared to PEG-DSPE/DSPC/Chol DOX in the absence of the MDR modulator. In the presence of PSC 833, the liver DOX accumulation with EPC/Chol DOX was 3.2-fold higher than that for the PEG-DSPE/DSPC/Chol DOX. 194 Table 4.5B Liver DOX distribution following administration of EPC/Chol DOX and PEG-DSPE/DSPC/Chol DOX in the presence and absence of PSC 833. Group Liver Cmax (Pg/g) A U C (pg.h/g) EPC/Chol Dox (5 mg/kg) 43.1 + 5.4 1511.6 EPC/Chol Dox (5 mg/kg) + PSC 833 53.6 ±2.3 3034.5* PEG-DSPE/DSPC/Chol Dox (5 mg/kg) 41.7 + 6.8 1715.7 PEG-DSPE/DSPC/Chol Dox (5 mg/kg) + PSC 833 28.5 ± 10.5 950.5* * Significantly different from the no PSC 833 group at p<0.05. AUC 0-72h was calculated using the trapezoidal rule. Data are expressed as mean ± s.e.m. Liver DOXone distribution following administration of free DOX, EPC/Chol DOX and PEG-DSPE/DSPC/Chol DOX in the presence and absence of PSC 833. Group Liver Cmax (pg/g) AUC (pg.h/g) Free DOX 7.5 mg/kg 84.2 ±6 .8 241.1* Free DOX 7.5 mg/kg + PSC 833 307.6 + 34.1** 1034.5*'** EPC/Chol Dox (5 mg/kg) 65.8 +22.6 262.8 EPC/Chol Dox (5 mg/kg) + PSC 833 102.4 + 4.7 2477.9** PEG-DSPE/DSPC/Chol Dox (5 mg/kg) 8.5 ±5 .6 62.5 PEG-DSPE/DSPC/Chol Dox (5 mg/kg) + PSC 833 13.5 ±0.3 699.5** Data are mean ± s.e.m. (n = 3 mice/time point; n = 24/group) AUC 0-72h was calculated using the trapezoidal rule. *Free DOX AUC values represent 0 - 8 h time period **Significantly different from the no PSC 833 group at p<0.05. 195 Metabolites of DOX Free DOX is mainly metabolized in the liver to (a) doxorubicinol - via reduction of side chain carbonyl group to a secondary alcohol by a cytosolic aldo-keto reductase (Felsted et al., 1974), (b) doxorubicinone - via reductive cleavage of the daunosamine moiety, and (c.) 7-deoxy aglycones, followed by further aglycone conjugation and O-methylation (see Scheme 4.1). In general, the metabolites of DOX identified include, doxorubicinol, doxorubicione, doxorubicinolone, 7-deoxydoxorubicinone, 7-deoxydoxorubicinolone, as well as sulfate and glucuronide conjugates (Arcamone et al., 1984; Andrews et al., 1980; Vrignaud et al., 1986; Bachur et al., 1974; Benjamin et al., 1973; Takanashi and Bachur, 1976; Preiss et al., 1989; Bronchud et al., 1990). Doxorubicinol Scheme 4.1.: Metabolism of DOX to DOXol, DOXone, and 7-deoxyDOXone. Following administration of non-encapsulated (free) DOX alone, DOX and all three metabolites were quantitated. The levels of DOXone were considerably higher than either 196 DOXol or the 7-deoxyDOXone metabolite. Specifically, a Cmax of 10.62 ± 1.26 pg/g and an AUCo-4 of 35.6 pg.h/g for DOXol, a C ™ , of 4.78 ± 0.66 pg/g and an AUCo-s of 22.75 pg.h/g for 7-deoxyDOXone, and a Cnax of 84.2 + 6.79 pg/g and an AUC 0 . 8 of 241.1 pg.h/g for DOXone, were obtained. On a relative scale, the respective contributions of DOX and its metabolites to the total AUC, at 1 h, are as follows: DOX (13.1%), DOXol (9.6%), DOXone (73.5%), and 7-deoxyDOXone (3.8%). PSC 833 co-administration caused substantial alterations in DOX metabolite levels. When free DOX was co-administered with PSC 833, a 4.3-fold increase in DOXone AUCo-g was observed, along with a 8.8-fold increase in 7-deoxyDOXone AUC 0 . 8 (p<0.05, ± P S C 833). DOXol levels following PSC 833 co-administration, however, were below assay detection limits. In contrast to the observations with free DOX, liposome encapsulated DOX exhibited different metabolic properties both in absolute drug levels as well as relative distribution between different DOX metabolites. For liposomal DOX, two major metabolites that were identified were DOXone and 7-deoxyDOXone. Table 4.5B and Figure 4.6 (Panel A) summarizes DOXone C^x and AUC for EPC/Chol DOX and PEG-DSPE/DSPC/Chol DOX, in the presence and absence of PSC 833, in the liver. With EPC/Chol DOX alone at 5 mg/kg, a C m a x of 65.8 ± 22.6 pg/g and an AUC of 262.8 pg.h/g was obtained. EPC/Chol is a leaky formulation of DOX that displays features resembling free DOX with respect to metabolic profiles. With the long circulating PEG-DSPE/DSPC/Chol DOX formulation, formation of the DOXone metabolite was not favored, and a Cmax of 8.5 ± 5.6 pg/g and an AUC of 62.5 pg.h/g was observed. Compared to the EPC/Chol DOX, this represents an 8-fold decrease in C m a x and a 4-fold decrease in AUC of DOXone (significant at p<0.05). 197 DOXol levels were below assay detection limits for both liposomal formulations. On a relative scale, the respective contributions of DOX and its various metabolites to the total AUC at 1 h for EPC/Chol DOX (5 mg/kg) 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%). Clearly, the % DOX recovered was highest for the sterically stabilized formulation (43.4%), intermediate for the leaky EPC/Chol formulation (16.6%), and lowest for non-encapsulated DOX (13.1%). The % DOXone recovered also followed the general relationship: greatest for non-encapsulated DOX (73.5%), intermediate for EPC/Chol DOX (50.2%), and lowest for PEG-DSPE/DSPC/Chol DOX (8.2%). When PSC 833 was co-administered with EPC/Chol DOX at 5 mg/kg, a higher DOXone C m a x value was observed (compare 65.8 ± 22.6 ug/g for EPC/Chol DOX and 102.4 ± 4.7 ug/g for EPC/Chol DOX and PSC 833), along with a significant 9.5-fold increase in AUC. These increases in DOXone Cmax and A U C 0 - 7 2 for EPC/Chol DOX following PSC 833 co-administration are similar to those observed for non-encapsulated DOX in the presence of PSC 833. It must be pointed out here that the AUC values for EPC/Chol DOX reported here are over a 0-72 h period and for non-encapsulated DOX, the time course beyond 8 h were not followed due to undetectable levels. When AUC values over the 0 - 4 h time period are considered, DOXone A U C 0 - 4 values of 77.7 ug.h/g and 311.8 ug.h/g, for the EPC/Chol DOX treatment group, are obtained in the absence and presence of PSC 833, respectively. These DOXone AUC values are much lower than for free DOX, where A U C 0 - 4 values of 173.1 ug.h/g and 681.4 ug.h/g were obtained in the absence and presence of PSC 833, respectively. In contrast, PSC 833 co-administration 198 with PEG-DSPE/DSPC/Chol DOX resulted in negligible increase in C^x (compare 8.5 ± 5.6 pg/g for PEG-DSPE/DSPC/Chol DOX and 13.5 ± 0.3 pg/g for PEG-DSPE/DSPC/Chol DOX and PSC 833), however, a 11-fold increase in AUC was observed. Figure 4.6: Liver (Panel A) and kidney (Panel B) DOXone concentration-time profiles following administration of EPC/Chol DOX 5 mg/kg in the absence (open circles) or presence of PSC 833 (administered 4 hours prior 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 of PSC 833 (administered 4 hours prior to DOX) 100 mg/kg p.o. (filled squares). Mice bearing 200 - 500 mg solid MDA435LCC6 tumors were terminated at the indicated times and liver and kidney samples were processed and quantified for DOXone by HPLC as described in Materials and Methods. Data are expressed as mean ± standard error of the mean (n = 3 mice/time point; n = 24/group). 199 4.3.3.4. Kidney drug kinetics Parent (intact) drug DOX accumulation in another PGP-expressing excretory organ, the kidney, was determined for the two liposomal formulations in the presence and absence of PSC 833, the profiles are illustrated in Figure 4.5 (Panel D). Since the kidneys are characterized by a tight vasculature, liposomes tend not to accumulate in this organ (Van Hossel et al., 1984). When EPC/Chol DOX is administered at 5 mg/kg, in the absence of PSC 833, kidney DOX distribution was characterized by a C m a x of 14.9 ± 0 . 1 ug/g and an AUC of 749.4 Ug.h/g. Coadminstration of PSC 833 with EPC/Chol DOX resulted in negligible increases in Cmax and small increases in AUC (compare 749.4 ug.h/g for EPC/Chol DOX and 1010.2 ug.h/g for EPC/Chol DOX and PSC 833). Compared to the EPC/Chol DOX formulation alone, PEG-DSPE/DSPC/Chol DOX (in the absence of PSC 833) accumulation was reduced by as much as 2-fold. These observations are consistent with the fact that while non-encapsulated drug accumulates in kidney (Krishna and Mayer, 1997), long circulating liposomal DOX formulations do not. We see higher accumulation in kidney with EPC/Chol DOX, a formulation that releases DOX, but not with the sterically stabilized liposomal system. Metabolites of DOX Kidney DOXone and 7-deoxyDOXone levels were monitored for the formulations in the presence or absence of PSC 833 and the results are presented in Figure 4.6 (Panel B). In the absence of PSC 833, DOXone kidney distribution properties following administration of EPC/Chol DOX 5 mg/kg were characterized by a Cmax of 6.3 ± 0.6 u.g/g and an AUC of 12.6 ug.h/g. PSC 833 co-administration with EPC/Chol DOX caused a 200 significant (p<0.05) 9-fold increase in Cmax and a 60-fold increase in AUC. When PEG-DSPE/DSPC/Chol DOX was administered alone, DOXone levels in kidney were below assay detection limits (LOD = 0.1 pg/g). When PEG-DSPE/DSPC/Chol DOX was combined with PSC 833, a Cmax of 2.6 ± 0.5 pg/g and an AUC of 118.3 pg.h/g was obtained. In the presence of PSC 833, the DOXone kidney Cmax and AUC for EPC/Chol DOX liposomes were 22-fold and 6-fold higher than that for PEG-DSPE/DSPC/Chol liposomal DOX (significant at p<0.05), respectively, indicating the enhanced ability of PSC 833 to alter DOXone distribution properties when DOX is co-administered in leaky EPC/Chol liposomes compared to PEG-DSPE/DSPC/Chol long circulating liposomes. These observations of DOX and DOXone distribution is similar to those seen in the liver. 4.3.4. Confocal Imaging In order to elucidate whether the subcellular distribution properties of DOX are influenced by PSC 833 in PGP expressing solid tumors, a non-perturbing confocal fluorescence microscopy (CFM) technique with image processing analysis was used. CFM analysis can be performed using thick (2-3 mm) and viable tissues, perfused in buffer, an approach that has been routinely utilized for brain tissues in psychiatric applications. Such sections are over 500 times the thickness typically used in microscopy studies (1-5 pm). The need for a non-perturbing imaging technique is apparent because standard tissue processing methodologies for conventional immunohistochemical or fluorescence microscopy utilize a tissue fixative, a cryofixation step and/or several washing steps (Mayer et al., 1997) that may disrupt the liposomal bilayer and cause drug leakage. This could potentially lead to high intracellular DOX concentrations, given DOX's propensity to bind to DNA. This is illustrated in Figure 4.7A, where, when PEG-DSPE/DSPC/Chol 201 liposomal DOX is spiked in freshly isolated muscle tissue, fluorescence of DOX is not visualized due to the fluorescence quenching (Panel A). However, when cryofixation is performed (Panel B), a significant increase DOX fluorescence was observed, indicating that drug leakage had occurred following disruption of the liposomal bilayer caused by the sample processing steps. This method also takes advantage of the fact that considerable quenching occurs when DOX is within liposomes (Wu et al., 1997), as corroborated in the present study, leading to the selectivity of visualizing fluorescence of released DOX. Consequently, this method can be applied to determine bioavailable DOX with liposomal DOX formulations. The results presented in Figure 4.7B demonstrate the fluorescence quenching observed for liposome-encapsulated DOX as well as the non-perturbing nature of evaluating fresh tissue. Panel A is the confocal image of PEG-DSPE/DSPC/Chol DOX liposomes injected in freshly isolated muscle tissue at a concentration of 10 ug/g. No DOX fluorescence is visualized, however, the presence of liposomes can be confirmed since PEG-DSPE/DSPC/Chol liposomes labeled with the fluorescent lipid label, Di-I (Panel B) were visualized throughout the tissue. Further, when non-encapsulated DOX is injected at identical concentrations, DOX is readily visualized and appears as intense nuclear fluorescence (Panel C). Therefore, the method that we have utilized here can qualitatively assess concentrations of fluorescent DOX that have been released from the liposomes and can also provide indications of the intracellular localization of DOX in the presence and absence of PSC 833. Further, the method preserves cellular architecture, allows 3-D projections of multiple scans to be made and preserves liposome integrity. Despite the greater thickness of the tissue, DOX visualization was possible due to the strong 202 fluorescent signal afforded by non-encapsulated DOX and the ability of confocal microscopy to image finite sections of the tissues. It must be pointed out here that all the scans were performed on one day at identical instrument settings and imaging exposure times. Therefore, any increases in DOX signal is attributable to increased DOX levels, and enhanced intracellular accumulation may indicate PGP blockade by PSC 833. However, it should be pointed out here that the results presented provide qualitative information, which were not amenable to quantitative analysis, thus precluding definitive statistical comparisons to be made. Figures 4.7 A & B (following pages) Figure 4.7A: A comparison of two sample treatments on DOX visualization using CFM. Panel A represents the image after PEG-DSPE/DSPC/Chol -liposomal DOX is injected into a defined mass of freshly isolated muscle tissue, at a concentration of 10 pg/g, (from SCID/Rag2 mice) and visualized fresh. Panel B is a representation of PEG-DSPE/DSPC/Chol -liposomal DOX at identical concentration in muscle, but subjected to a standard cyrofixation protocol before visualization under the confocal microscope. Note the distinct DOX fluorescence signal consistent with nuclear localization following the latter protocol. Figure 4.7B: An illustration of DOX fluorescence quenching properties when encapsulated within liposomes. Freshly isolated muscle tissue from SCID/Rag2 mice were injected with known amounts of PEG-DSPE/DSPC/Chol -liposomal DOX (10 pg/g; Panel A), empty PEG-DSPE/DSPC/Chol -liposomes labeled with Dil (Panel B), and free DOX (10 pg/g; Panel C). Note the distinct strong DOX fluorescence in Panel C, but not in Panel B, indicating substantial quenching with PEG-DSPE/DSPC/Chol -liposomal DOX. The visualization of Dil label confirms the presence of empty PEG-DSPE/DSPC/Chol liposomes. 203 • t 03 204 Figure 4.8 presents the CFM images of MDA435LCC6 MDR1 tumors from mice administered non-encapsulated DOX (Free), EPC/Chol DOX (EPC), and PEG-DSPE/DSPC/Chol DOX (PEG) all at a dose of 5 mg/kg, in the presence and absence of PSC 833. MDR tumors isolated from mice treated with non-encapsulated DOX in the absence of PSC 833 demonstrated no detectable DOX fluorescence either at 1 h or 4 h. In the presence of PSC 833, however, increased DOX fluorescence was visualized at lh which increased by 24 h (Figure 4.8). The pattern of fluorescence images in the presence of PSC 833 were consistent with intracellular accumulation of DOX. For EPC/Chol DOX treated mice, tumor DOX fluorescence was visualized at 4 h and beyond (Figure 4.8). Again, PSC 833 caused an apparent increase in intracellular DOX accumulation as seen from the more intense foci of DOX fluorescence at 1 h and 4 h compared to mice treated with EPC/Chol DOX in the absence of PSC 833. Similar PSC 833 induced increases in the intracellular accumulation of DOX were observed for PEG-DSPE/DSPC/Chol DOX at 1 h and 4 h. Evaluation of intracellular tumor DOX accumulation for mice treated with either liposomal formulation could not readily be assessed at later time points due to increased diffuse (presumably extracellular) fluorescence which precluded even a qualitative evaluation of intracellular DOX uptake in the presence or absence of PSC 833. 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 identical both in the presence and absence of PSC 833, suggesting that (a) increasing anticancer drug dose intensity alone using saturated lipid formulations results in increased intracellular accumulation, and (b) PSC 833 pre-treatment causes increases in drug retention and increased intracellular DOX accumulation. 206 Figure 4.8 (following page) Tumor confocal images of free DOX (5 mg/kg), EPC/Chol DOX (5 mg/kg) or PEG-DSPE/DSPC/Chol -liposomal DOX (5 mg/kg), in the presence and absence of PSC 833. SCJX)/Rag2 mice were treated with various combinations at a DOX dose of 5 mg/kg, injected i.v., alone or in the presence of PSC 833 (p.o., 100 mg/kg; 4 h before DOX). Following 1 and 4 h post-DOX administration, tumor samples were aseptically dissected and placed in PBS containing tubes, maintained on ice. These samples were directly imaged fresh by placing them on concave slides under the CFM (see Materials and Methods). Identical settings on the CFM and identical processing times facilitated a comparison of DOX fluorescence intensity in the presence and absence of PSC 833. 207 4.4. Discussion The results presented in Chapter 3 demonstrated a lack of a PSC 833-induced adverse pharmacokinetic interactions when DOX was presented in DSPC/Chol liposomes. Also observed were increased efficacy in the P388/ADR solid tumor model and decreased toxicity of the liposomal DOX-PSC 833 combination. In order to investigate the influence of liposomal drug retention and tumor uptake properties on the MDR reversal properties of liposomal DOX in the presence and absence of PSC 833, studies were designed to compare the toxicity, antitumor efficacy, tissue and intracellular DOX distribution properties of two liposomal DOX formulations, namely an EPC/Chol rapid DOX releasing system (EPC/Chol) and a long circulating sterically stabilized system (PEG-DSPE/DSPC/Chol). In addition, these two formulations represent clinically used liposomal DOX systems that are either approved or in the final stages of development, and consequently, determining their MDR reversal properties in the presence of MDR modulators would be of considerable clinical importance. The results presented here demonstrate that liposome properties influencing drug retention and tissue distribution properties significantly affect the susceptibility of encapsulated DOX to PGP blockade by PSC 833 in tumors and healthy tissues expressing this drug efflux pump. In particular, for EPC/Chol DOX, a rapid drug releasing system, the dose reduction required was greater than that required for non-encapsulated DOX, in the presence of MDR modulator (Krishna and Mayer, 1997). Specifically, while the non-encapsulated DOX MTD was decreased 2.6-fold in the single injection and 3-fold in multiple injection schedules, use of EPC/Chol DOX resulted in a 4.5-fold reduction in MTD when combined with PSC 833 for the day 1 schedule. The MTDs and dose 209 reductions for EPC/Chol DOX were similar to those for non-encapsulated DOX, in the presence of PSC 833 for multiple injection, although the MTD for EPC/Chol DOX plus PSC 833 was 2-fold higher than that for the non-encapsulated DOX. These toxicity observations correlated with the degree of DOX retention within liposomes as observed by the plasma drug-to-lipid ratios. Further, the increased toxicity of DOX when administered in EPC/Chol liposomes in the presence of PSC 833 was consistent with the increased plasma levels in the circulation, as indicated by a decrease in drug-to-lipid ratio as well as elevated DOX concentrations in Microcon filtrates, due to rapid drug leakage observed with the EPC/Chol system. This increased drug leakage resulted in a 2.6-fold increased exposure of the drug in susceptible organs in the kidney and the liver. Drug-to-lipid ratio of the PEG-DSPE/DSPC/Chol DOX system, however, remained constant over time, indicating considerably less free drug exposure occurring with these liposomes, corroborating the reductions in toxicity. Tumor growth suppression studies indicated modest modulation of EPC/Chol DOX in the presence and absence of PSC 833, whereas significant reductions in tumor growth were observed with the sterically stabilized liposomal system in the presence of PSC 833. Tumor growth suppressions appeared to correlate most closely with increased tumor DOX accumulation. These observations are consistent with the increases in drug delivery to the tumor (increased local anticancer dose intensity) provided by the sterically stabilized liposomes as well as a definitive PSC 833 effect reflected by an increased intracellular accumulation of DOX, as observed using CFM analysis. An interesting observation was the pharmacokinetic profile of EPC/Chol DOX and non-encapsulated DOX, in the presence of PSC 833. While PSC 833 caused a 10-fold 210 increase in non-encapsulated DOX AUC, typified by a prolonged terminal elimination phase (Krishna and Mayer, 1997), PSC 833 caused an overall increase in DOX AUC by 2.6-fold and an increase in the initial distribution phase when co-administered with EPC/Chol DOX. The increase in free drug levels (Table 4.4) when EPC/Chol DOX is combined with PSC 833 was determined using equilibrium ultrafiltration studies. The presence of free drug in the ultrafiltrate represents the free drug fraction not including the fraction (released from the liposome) that is protein bound. The 2.6-fold increase in plasma DOX AUC with EPC/Chol DOX caused by PSC 833 can largely be accounted by the increase in free fraction plus the protein bound fraction such that the non-liposomal DOX at 1 and 4 h after EPC/Chol DOX treatment could reflect approximately 32% and 12%, respectively, of the total plasma DOX. In contrast, PSC 833 did not significantly alter the pharmacokinetics of DOX encapsulated in PEG-DSPE/DSPC/Chol sterically stabilized systems. This is also corroborated by observations of undetectable free drug levels in the ultrafiltrate for PEG-DSPE/DSPC/Chol DOX liposomes, both in the presence and absence of PSC 833. The above results suggest that: (1) when comparing non-encapsulated and various liposomal DOX formulations, increased tumor levels correlate with increased antitumor * activity in the presence of PSC 833 and (2) increased drug leakage from the liposomes is associated with larger PSC 833-induced toxicity exacerbation. Further, a need to better understand cellular distribution is stressed since liposome-encapsulated drug is not bioavailable. DOX levels in the tumors (where tumors are homogenized in water, frozen until analysis, thawed and analyzed) represent "total" DOX content, i.e., released plus encapsulated drug. Further, these total DOX levels in tumor do not differentiate between 211 cell-associated drug (intracellular accumulation) and non-cell associated (extracellular) drug. The latter is particularly relevant for liposomal systems, because of their inability to permeate the cellular membrane per se, and they may be localized at cell-cell junctions, where they may serve as a depot releasing DOX as a function of time. The increased efficacy of PEG-DSPE/DSPC/Chol DOX and PSC 833 can be explained if PSC 833 were to cause an increase in intracellular free drug levels leading to increased intracellular accumulation of DOX, released from PEG-DSPE/DSPC/Chol liposomes. For liposomal systems, the changes in intracellular distribution of DOX can occur irrespective of whether or not an increase in Cmax or AUC is observed, in the presence of PSC 833. We predicted that although there may be slightly higher pools of PEG-DSPE/DSPC/Chol DOX at the tumor site in the presence of PSC 833, there may be a significant nuclear accumulation caused by PSC 833, whereas with PEG-DSPE/DSPC/Chol DOX alone the liposomal drug pools are perhaps localized at cellular junctions or interstitial spaces. The CFM results corroborated this prediction as seen by the increased intracellular accumulation caused by PSC 833, which is consistent with observations of increased drug retention in tumors. Our CFM results suggest that PSC 833 co-administration causes an increase in the nuclear localization of DOX consistent with the ability of this MDR modulator bo block PGP drug efflux in vitro. In Chapter 3, observations with DSPC/Chol liposomes indicated higher levels of drug exposure in the liver in the presence and absence of PSC 833. Surprisingly, no evidence of PSC 833-mediated toxicities or changes in DOX pharmacokinetics were observed, leading to a speculation that liposomes may alter the metabolic processes for the free drug in the liver. This prediction was also corroborated in the present study, where 212 differences in the extent of metabolite formation as well as in metabolite distribution were observed with the two liposomal formulations. While EPC/Chol DOX closely resembled non-encapsulated DOX with respect to metabolite formation, both liposome systems did not favor the formation of DOXol, which is formed when DOX is metabolized by the cytosolic aldo-keto reductase. Further, higher levels of the two aglycones, DOXone and 7-deoxyDOXone, were observed with EPC/Chol system compared to PEG-DSPE/DSPC/Chol DOX. PSC 833 co-administration also caused dramatic increases in DOX and DOXone levels with EPC/Chol DOX, but did not significantly alter the metabolic distribution of PEG-DSPE/DSPC/Chol DOX. It is not known that whether the aglycone metabolites are substrates of PGP and whether they are effluxed in the same manner as DOX, and consequently, the toxicological implications of these PSC 833-mediated increases in aglycone levels remains to be investigated. Similar observations were obtained for the kidney, which also expresses PGP. The improvements observed with liposomal systems may perhaps be due to favorable alterations in DOX renal and biliary excretion. Liposomes may be expected to deliver DOX to the phagocytic Kupffer cells in the liver (Scherphof et al., 1983a) which may lead to DOX metabolites that are not significantly affected by PGP blockade. In summary, toxicity characteristics observed when liposomal DOX formulations are co-administered with PSC 833 appear to be dictated by liposomal drug retention properties where increased drug retention within liposomes reduces toxicity. Tumor growth suppression, on the other hand, appears to be primarily dictated by a combination of liposomal drug tumor accumulation as well as intracellular drug retention caused by the MDR modulator. Consequently, improvements in MDR reversal properties observed with 213 liposomal DOX appears to be mediated by 1) PSC 833 blockade of PGP mediated increased drug retention by tumor cells, and 2) increased dose intensity provided by liposomes. As seen from the results presented here, these factors are maximized with long circulating non-leaky liposomal systems. Further investigations comparing liposomal and non-encapsulated DOX renal and biliary clearances in an instrumented animal model may shed light on the mechanisms by which liposomes confer enhanced protection to MDR modulator-mediated toxicity and pharmacokinetic alterations of co-administered anticancer drugs. Such studies evaluating the hepatobiliary and renal disposition of DOX and effects of the MDR modulator, PSC 833, are presented in Chapter 5. 214 Chapter 5 INFLUENCE OF LIPOSOMAL ENCAPSULATION ON THE RENAL AND HEPATOBILIARY DISPOSITION OF DOX IN THE PRESENCE AND ABSENCE OF PSC 833 USING AN INSTRUMENTED RAT MODEL 5.1 Introduction The hepatobiliary system plays an important role in the elimination of many drugs from the body (Oude Elferink et al., 1995; Meijer et al., 1997). The hepatobiliary uptake of organic molecules has been extensively studied, particularly in rat isolated perfused livers (Meijer et al., 1976; Ballet et al., 1987; Stapf et al., 1994), as well as in isolated hepatocytes (Okudaira et al., 1992; Miyauchi et al., 1993; Nakamura et al, 1994). Multiple elimination mechanisms, depending on size and chemical nature of drugs, have been shown to exist which control the excretion of xenobiotics in bile and urine (Steen et al, 1991; Grundemann et al, 1994; Martel et al, 1996; Meijer et al, 1990; Meijer et al, 1991). It has been shown using vesicles reconstituted from canalicular membranes that PGP is directly involved in the biliary excretion of many amphiphilic anticancer drugs (Kamimoto et al, 1989). This role of PGP in biliary excretion process has also been a subject of recent reviews that suggest a need to address issues of drug-drug interactions that may potentially cause pharmacokinetic alterations when multiple agents are co-administered (Oude Elferink et al, 1995; Meijer et al, 1997). Chemosensitization approaches to circumvent PGP-mediated multidrug resistance typically utilize anticancer drugs co-administered with a PGP inhibitor in order to cause enhanced intracellular tumor accumulation of the drug. However, due to lack of tumor specificity for current PGP blocking agents (MDR modulators), PGP blockade may also occur in normal healthy excretory tissues, where such transport pumps are expressed. The 215 implication of PGP blockade at extratumoral sites is that excretion of anticancer drugs may be impaired when co-administered with MDR modulators that directly block PGP function. In general, MDR modulator-induced pharmacokinetic alterations may occur at the hepatic-, intestinal, renal, or blood brain barrier levels of elimination (Schinkel et al., 1994; ibid., 1995; Mayer et al., 1996; Sparrboom et al., 1997). In support of this, several reports have confirmed inhibition of anticancer drug biliary excretion (Watanabe et al., 1992; Speeg et al., 1992a,b; Thalhammer et al, 1994; Speeg and Maldonado, 1994; Stapf et al, 1994; Booth et al, 1998). Recently, these interactions have been shown to occur at the bile canalicular level during transport and the inhibitory effects of MDR modulator are related to the drug's lipophilicity (Smit et al., 1998). Likewise, PGP and perhaps other transport pumps have been implicated in the secretion of amphiphiltc agents into the bile. Particularly, in mdrl knock-out mice, biliary excretion was inhibited by at least 60% (Smit et al, 1996). This raises possibilities of involvement of another transport mechanism, concurrent with PGP. The multidrug resistance-associated protein or MRP (Cole et al, 1992) isoforms have been implied as being expressed preferentially in the biliary canaliculi. In support of this, a cannalicular multiple organic anion transporter or cMOAT (Paulusma et al, 1996), a conjugate export pump, which has been identified as MRP2, has been cloned and shown to mediate drug transport. An understanding of DOX distribution and pharmacokinetic properties is central to this thesis, and consequently, the following discussions will focus of the biological fate of systemically administered DOX. Following i.v. administration in humans, DOX is rapidly eliminated from plasma, characterized by a multiphasic profile with distribution half-life of 10-30 min and bi-phasic elimination half-lives of 3 h and 30 h. During the rapid 216 distribution phase, there is extensive tissue distribution as observed by a large volume of distribution (Vd) of 25 1 in humans (De Vita et al., 1993). DOX extensively distributes in heart, lungs, spleen, as well as PGP expressing normal tissues such as the liver, kidney, and the intestine. However, it does not cross the blood-brain barrier. Tissue DOX levels have been correlated with DNA content in tissues (Terasaki et al., 1989). Of the DOX present in plasma, about 75-80% of DOX is protein bound (Greene et al., 1983; Mayer and St-Onge, 1995). DOX is readily eliminated in urine and bile, via a PGP dependent process as a result of the expression of this drug efflux pump in the luminal side of the kidney proximal tubule and biliary canaliculi. Pharmacokinetic studies with DOX have demonstrated that it is excreted in bile and urine. In humans, biliary excretion is a major route of drug elimination where DOX appears in bile within 5 min after an intravenous bolus administration (Cusack et al., 1993; Arcamone et al., 1984). Intact DOX is the major chemical entity in bile and urine, with lower levels of DOXol and aglycones (Carlo et al., 1988). DOX is also characterized by a large volume of distribution, indicating extensive tissue distribution (Greene et al., 1983). Biliary excretion represents 41% of injected drug excreted in bile compared to 14% in urine (Riggs et al., 1977). Further, biliary excretion accounts for 90% of eliminated DOX in rabbits (Bachur et al., 1974). Reports on biliary excretion have been relatively sparse in smaller animals such as the rats. Further, in rats, significant differences in biliary excretion have been observed, ranging from 8% of injected drug in bile (Yesair et al., 1972) to 20% (Israel et al., 1978) and 35% (Tavoloni and Guarino, 1980a). Both colchicine and DOX secretion in urine has been thought to be mediated via PGP located in the kidney proximal tubule (Thiebault et al., 1987; Speeg et al., 1992b). Urinary excretion 217 of DOX (% injected dose) has been reported to be 4-8% in anesthetized rats (Tavaloni and Guarino, 1980a), 14% in humans (Riggs et al., 1977), and 7-8% in conscious rats (Yesair et al., 1972). Consequently, it may be expected that co-administration of MDR modulator with DOX may lead to inhibitions of DOX renal and biliary excretion, via PGP blockade in the kidney and biliary canaliculi. Both cyclosporin A (CsA) and its non-immunosuppressive analog, PSC 833 have been shown to block biliary and renal excretion of colchicine and DOX in the rat (Speeg and Maldonado, 1994; Speeg et al., 1992a; Speeg et al., 1992b). A dose-dependent inhibitory effect was observed with CsA where increasing the CsA dose 5-fold caused the colchicine biliary clearance to be reduced from 64% to 81% (Speeg et al., 1992a). PSC 833 reduced biliary clearance of colchicine from 9.05 to 2.41 ml/min/kg and decreased drug bile-to-plasma ratio from 146 to 35 in the rat. For DOX, PSC 833 caused a similar profound reduction in biliary excretion, from 10.5 to 2.48 ml/min/kg and decreased the drug bile-to-plasma ratio from 228 to 48 (Speeg and Maldonado, 1994). Further, PSC 833 inhibited colchicine's renal clearance completely in one animal (Speeg and Maldonado, 1994). This inhibition caused by PSC 833 was immediate and prolonged. In Chapters 3 and 4, it was observed that liposomes, by increasing exposure of the drug in the tumors as well as reducing exposure of the encapsulated drug to susceptible healthy tissues confers enhanced therapeutic activity against MDR tumors while alleviating toxicity complications when co-administered with PSC 833. It was also observed that the profde of DOX and related metabolites in the liver was very different between free and liposomal formulations. This suggested that DOX metabolism and excretion after administration in liposomal formulations may be much less dependent on the elimination 218 processes involving PGP compared to non-encapsulated DOX. In this Chapter, the evaluation of the DOX renal and biliary clearance, following i.v. injection of free (non-encapsulated) and liposome-entrapped DOX in a chronically instrumented rat model is presented. These comparisons are made in the presence and absence of the MDR modulator, PSC 833 in order to elucidate the mechanism(s) whereby liposomal DOX avoids biodistribution alterations caused by co-administration of PSC 833 on non-encapsulated DOX. The instrumented rat model was chosen for these studies given the need to perform serial sampling of blood and bile, a feature not feasible in the mouse under the constraints of the analytical procedure for measuring DOX and its metabolites (volume of biological sample required for drug measurement assay). 219 5.2. Materials and Methods 5.2.1. Materials DOX hydrochloride for injection, U. S. P. was purchased from David Bull Laboratories (Canada) Inc., Vaudreil, Quebec, and its purity affirmed by HPLC (See below). PSC 833 was a generous gift from Novartis (Canada) Inc., Dorval, Quebec, and its purity affirmed by LC/MS-MS (Varian V G Biotech LC-MS-MS system, Fisons Instruments, Altrincham, UK). DOX metabolite standards were generous gifts (Pharmacia Carlo Erba, Milan, Italy). The enzymes, -^glucuronidase and sulfatase were purchased from Sigma Chemical Company (St. Louis, MO). Polyethylene glycol 2000 coupled with distearoylphosphoethanolamine (PEG2000-DSPE, >99% purity), egg phosphocholine (EPC, >99% purity), and distearoylglycerophosphocholine (DSPC, >99% purity) were obtained from Northern Lipids, Inc (Vancouver, BC) and cholesterol was obtained from Sigma Chemical Company (St. Louis, Missouri). Cholesteryl hexadecyl ether (3H), a non-exchangeable, non-metabolizable lipid marker was purchased from Amersham Canada (Oakville, Ontario). Tween 80 (polyoxyethylene sorbitan monooleate), sodium carbonate, 30% w/w hydrogen peroxide, were obtained from Sigma Chemical Company (St Louis, MO), and ammonium formate from Fisher Scientific (Fairlawn, NJ). HPLC grade solvents, iso-propyl alcohol, acetone, ethanol, acetonitrile, and chloroform were all obtained from BDH, Inc (Toronto, Ontario), and used without further purification. HPLC grade water was prepared on-site with the Milli-Q system (Millipore Corporation, Mississauga, Ontario). Compressed air (medical grade), nirtogen, NF were purchased from Praxair, Inc (Mississauga, Ontario). Pyrex 15 mL disposable culture tubes (Corning Glass Works, Corning, NY) and polytetrafluoroethylene (PTFE) lined screw caps (Canlab, Vancouver, 220 BC). Sterile supplies as well as Sterile saline and water for injection, were obtained from the BC Cancer Agency (Vancouver, BC). Isoflurane (99.9%), USP was purchased from Abbott Laboratories, Ltd. (Montreal, Canada). 5.2.2. Liposome and Drug Preparation Liposomes composed of EPC/Chol (55:45), PEG2000-DSPE/DSPC/Chol (5:50:45), and DSPC/Chol (55:45) were prepared by initially dissolving the lipid mixtures in chloroform (100 mg lipid/ml), vacuum drying to generate a thin film and hydrating the dried lipid film in a 300 mM citric acid pH 4.00 buffer (see Section 3.2.2. for a detailed description of this procedure). Briefly, the resulting multilammelar vesicles (MLVs) were subjected to five freeze-thaw cycles followed by a 10 cycle extrusion through two stacked 100 nm polycarbonate filters (Nuclepore, Pleasanton, California) using a Lipex Extruder (Lipex Biomembranes Inc., Vancouver, British Columbia) (Mayer et al., 1986b). 3 H -Cholesterylhexadecyl ether was used as a non-exchangeable, non-metabolizable lipid marker (Derksen et al., 1987). The resulting large unilammelar vesicles (LUVs) exhibited a mean diameter ranging between 100 - 130 nm as determined using a Nicomp 270 submicron particle sizer (Particle Sizing Systems, Inc, Santa Barbara, CA), operating at a wavelength of 632.8 nm. DOX was encapsulated in the liposomes using the transmembrane pH gradient loading procedure (interior acidic) employing sodium carbonate as the alkalinizing agent and a drug to lipid weight ratio of 0.2:1.0 (Mayer et al., 1986a), as described in Section 3.2.3. Liposomal DOX preparations were diluted with saline as necessary prior to in vivo administration. PSC 833 (for animal studies) was dissolved in a 10:1 mixture of ethanol (95%):polyoxyethylene sorbitan monooleate (tween 80) and administered in a corn oil 221 vehicle by oral gavage of a 200 ul volume (Keller et al., 1992b; also see Section 3.2.4.). Non-encapsulated DOX was administered in sterile saline. 5.2.3. Liposome characterization Vesicle size distribution, liposome encapsulation efficiency and drug-to-lipid ratio were determined as a quality control measure following extrusion or storage of preformed and encapsulated liposomes (see Section 3.2.2. for a detailed description). Vesicles exhibiting a mean vesicle diameter of 100 - 130 nm, as determined using quasi-elastic light scattering method, with a standard deviation not greater than 30% and a Chi-square value of less than 0.4 were rendered acceptable for studies. A drug encapsulation efficiency of >95% was acceptable for studies. A drug-to-lipid ratio of 0.2:1.0 was used for these studies, and this ratio was used to monitor in vivo and in vitro. 5.2.4. Analytical Methods 5.2.4.1. HPLC Analysis for DOX and DOXol A Waters model 45 and 6000 A solvent metering systems (Waters Associates, Milford, MA) with a Waters Model 710 WISP autosampler, a Waters Model 470 fluoresence detector (500 nm excitation wavelength, 580 nm emission wavelength), Waters 510 HPLC pump, and a Waters Millenium Version 2.1. software for data processing was used. A NEC Powermate SX Plus Computer (NEC Information Systems, Inc., Boxborough, MA) and a Waters Systems Interface Module (Waters Associates) were used for data handling. A C18 guard column inlet filter (Rheodyne, Inc., Cotati, CA) was placed prior to the Nova-PakR C18 3.9 x 150 mm HPLC analytical column (Millipore Corporation, Milford, MA). Durapore 0.22 um GVWP and 0.45 um HVHP membrane filters were used to filter solvents (Millipore Corporation, Bedford, MA). The HPLC 222 assay of Andersen et al (1993) was used with minor modification. Briefly, sample extraction with acetonitrile was followed by isocratic elution from a C18 reverse phase column and quantified by endogenous fluorescence. The mobile phase consisted of a 16 mM ammonium formate buffer (pH 3.5)/acetone/isopropanol mixture (75:20:5) delivered at a rate of 1.0 ml/min. The column was maintained at 40 °C. The retention times of DOXol and DOX were 3.6 min and 5.9 min, respectively. Recoveries, using acetonitrile as the extraction solvent, from plasma over a concentration range of 0.05 - 10 ug/ml of DOX and DOXol were between 95-102%. Specifically, DOX recoveries (extracted versus unextracted), at a concentration of 1 ug/ml, from rat plasma, urine, and bile were 98.0 ± 3.3%, 95.1 ± 3.8%, and 94.3 ± 4.6%, respectively. 5.2.4.2. Lipid radioactivity To plasma samples from liposomal drug treated animals (0.02 ml) five ml of the scintillation fluid, Pico-Fluor (Packard Instruments, Meridien, CT) was added and 3 H dpm's were counted. All radioactivity determinations were performed using a TRI-CARB™ Model 1900 TR Liquid Scintillation Analyzer, Packard Instrumentation Company (Meridien, CT). 5.2.5. Animal Experiments Cannulated male Sprague Dawley rats (225-275 g) were purchased from Charles River (St. Constant, Quebec). These rats confirmed to existing standards of surgical, anesthetic use, and animal handling. All animals were used for experimentation within 10 days of surgery and were allowed free access to food and water. These rats had polyurethane cannulas placed in the jugular vein (TV) (0.025" JD x 0.040" OD, 13.5 cm long, dead volume of 40 ul) and common bile duct (0.025" ID x 0.040" OD, 28 cm long). 223 The bile duct was cannulated and exteriorized at the base of the neck in such a way that bile flow (from animal's left to right) into the small intestine can be interrupted and restored. The left half of the cannula was used for the sampling and the right half was plugged with a metal plug. On receipt of the animals, the JVC catheter was flushed with saline and locked with 40 ul of a heparinized saline solution. Animals were housed singly in steel metabolic cages throughout the experiment. Samples were collected by gently restraining the animal or briefly anesthetizing the animal with isoflurane-oxygen (controlled via an anesthetic machine; 2% at 2 L/min for 1 min and reduced to 1% at 1 L/min) at the respective time points. Rats received DOX (5 mg/kg) through the tail vein alone or 4h after PSC 833 (p.o.). Jugular venous blood samples (maximum of 0.25 ml) were collected at 5 min, 15 min, 30 min, 1 h, 1.5 h, 2 h, 3 h, and 4 h (Free DOX), 5 min, 15 min, 30 min, 1 h, 1.5 h, 2 h, 4 h, and 8 h (Free DOX + PSC 833), 15 min 30 min, lh, 2h, 4h, 8h, 16h, 24h (EPC/Chol DOX ± PSC 833), and 30 min, lh, 2h, 4h, 8h, 12 h, 24 h, 48 h (PEG DOX ± PSC 833) following DOX administration. An equal amount of saline was infused back along with a 0.04 ml heparinized saline lock. Rats were sampled for bile in tared polyethylene tubes at predetermined intervals immediately after bile flow was interrupted. One ml of lactated ringer solution was infused via the JVC as fluid replacement at every bile collection point. Bile volume was determined assuming bile density as being lg/ml. A rectal probe was used to monitor body temperature. Blood (plasma) and bile samples were analyzed for DOX, DOXol, and other DOX metabolites by HPLC as described above. In order to determine whether bile-flow interruption alters the pharmacokinetics of DOX, 224 groups of rats with intact JV catheters, where bile-duct catheter loop was left uninterrupted, were sampled for serial blood (plasma) and analyzed for DOX. Biliary excretion rate (ng/min) is a function of drug concentration in bile (ng/ml) and biliary flow rate (ml/min). Biliary clearance (CLb) was estimated by plotting the average biliary excretion rate versus the plasma drug concentration at the midpoint of collection interval. The slope of the line provided the biliary clearance in ml/min. The equation used to generate this value is: Biliary clearance (CLb) = Bile flow rate (ml/h) x Concentration of DOX in bile (pg/ml) Average JV plasma concentration (pg/ml) It should be noted that biliary clearance value for animals receiving liposomal DOX likely reflect clearance of free drug that has been released from liposomes in the plasma or tissues since liposomes are not excreted in bile (confirmed by the lack of detectable lipid levels in bile). Bile samples were analyzed for unconjugated DOX and conjugated DOX using two aliquots of the same sample. To one aliquot, a solution of sulfatase (to cleave sulfate conjugates of DOX) in pH 7.1 buffer was added to bile samples before extraction with organic solvent and incubated at 37°C for 3 hours (Parker et al., 1982b). To another aliquot, a solution of glucuronidase (to cleave glucuronide conjugates of DOX) in pH 5.0 formate buffer was added to bile samples and incubated at 37°C for 3 hours. The concentration of conjugated DOX was calculated by subtracting the non-conjugated DOX concentration from the samples measured following hydrolysis. 225 Urine was collected at predetermined intervals and urinary flow rate was determined. Blood (plasma) and urine samples were analyzed for DOX and DOXol. Urine samples were analyzed for unconjugated DOX and conjugated DOX. Two aliquots of the sample were used. To these aliquots, enzymes sulphatase (to cleave sulphate conjugates of DOX) in pH 7.1 buffer and P-glucoronidase (to cleave glucoronide conjugates of DOX) in pH 5.0 buffer were added to urine samples before extraction with organic solvent and incubated at 37°C for 3 hours. The concentration of conjugated DOX was calculated by subtracting the non-conjugated DOX concentration from the samples measured following hydrolysis. Urinary excretion rate (ng/min) is a function of drug concentration in urine (ng/ml) and urinary flow rate (ml/min). Renal clearance (CLr) was estimated by plotting the average urinary excretion rate versus the plasma drug concentration at the midpoint of collection interval. The slope of the line provided the renal clearance in ml/min. The equation used to determine the value is: Renal clearance (CLr) = Urine flow rate (ml/h) x Concentration of DOX in urine (pg/ml) Average JV plasma concentration (pg/ml) As with biliary clearance, renal clearance values for animals receiving liposomal DOX likely reflect clearance of free drug that has been released from liposomes in the plasma or tissues since liposomes are not excreted in urine (confirmed by the lack of detectable lipid levels in urine). The plasma data were modeled using WinNONLIN Version 1.5 pharmacokinetic software (Pharsight Corporation, CA), to calculate area under the curve (AUC), terminal 226 elimination half-life (Tin), plasma clearance (CLP), mean residence time (MRT) according to standard equations (Gibaldi and Perrier, 1982). In order to determine appropriate models to fit the plasma data, the criteria used to evaluate the goodness of fit for each model included a visual assessment of distribution of residuals, rank, and Akaike's Information Criterion (AIC, see Akaike, 1976). Data are presented as mean + standard deviation (n = 3 animals/group). Statistical analyses were performed using ANOVA and statistical significance was set at p<0.05. 5.3. Results 5.3.1. Plasma Pharmacokinetics Figure 5.1 illustrates the plasma DOX concentration-time profile for non-encapsulated DOX (Panel A), EPC/Chol DOX (Panel B), and PEG-DSPE/DSPC/Chol DOX (Panel C) in the presence and absence of PSC 833. The pharmacokinetic parameters, area under the curve (AUC), terminal elimination half-life (Tin), peak concentration achieved (Cmax), and mean residence time (MRT) for non-encapsulated DOX and the two liposome-encapsulated forms (in the presence and absence of PSC 833), are summarized in Table 5.1. 227 TD O C m TO ^—^ - ro O co £ oo ^ U O co PH ^ P J ^ O CU cu <U u c CU cn CU <2 a. ~a o 3 60 cu ro c3 II | O "i= .S T3 CO T3 eo .*-» CA +1 CU £ CA I-l CU +^ CU s eo t-eO O H 60 e in «*H O cu CA o -a cu S3 O O CCJ X o Q X! O Q IS) CO § .a J3 rs O H X > CA CO S3 "3 eO O H tj CU "C cu *^ CA T*1 _ . o ~Q ~ CJ PH CO § PH 00 Q 3 00 in X o Q o •*-> I-l o •d O H 43 d O H SO •3 w a H PH £ + ,—. ro 60 ro ^ 00 £ co in ^ I O I T ) X o Q 00 00 • I O wo X o Q *o 43 o CJ m PH + ,—.. ro 60 ro 00 "eb CJ co PH 60 ft £ o i n 60 X o o s W + / - ^ ro 60 ro ^ 00 g>U £ to I fe y \ ^ X o Q CU CU X o Q CU fe 60 £ i n cu £ CCJ l a CS PH PH vo ro' ON +1 Ov vq ro 00 +1 © 00 Ov ro Ov vo +1 00 CO ro O tN CN O +1 "3-r-m oflj 00 o +1 00 VO CO © +1 i n m i n r j «n ro O +1 i n Ov +1 o * CN cN O +1 00 in cs m o +1 m o * o +1 00 CM O +1 co m +1 i n i n CN c s o 00 ro +1 00 CN ro c s CO 00 +1 CN CN Ov +1 Ov Ov CN * VO i n VO 00 d d +1 +1 CN ro r -OV +1 VO "5b U VO o 0 +1 o © d +1 CO d Ov d +1 CN CN CN d +1 CN CN Ov +1 m CN CN CU > I— CU *J CU X ! cf 'CA 3 -a cu e cu H-> CU T3 CA CO T3 CU C J 00 (N CN o UH CU O H (3 M O c .£ 8 ^ •a CU PH CU > ^ 9 j> .£ w J*H CA m is -o 6 ^ .1 ^ fi £ •2 g CO 1 t5 3 .£ -o o e I-. £ e § 0 0 a 00 y 2 cj o ' co <5 a • • o <s « e 0 in C IS CU. « J - g "S & 60 3 O -53 CO O H 5 .£ ^ < "3 -B cj § —•— Free D O X - • - Free D O X + P S C 833 Time (h) Figure 5.1: Plasma DOX concentration-time profiles of free DOX (Panel A), EPC/Chol DOX (Panel B), and PEG-DSPE/DSPC/Chol DOX (Panel C) at a DOX dose of 5 mg/kg i.v. in the absence and presence of PSC 833 (50 mg/kg, p.o., administered 4 h prior to DOX). Data are represented as mean ± standard deviation (n = 3/group). Note the different x and y axis scale for Panels A, B and C. 229 Prior to experimentation with various treatment groups, the effect of bile-flow interruption on the plasma pharmacokinetics of DOX was investigated. These control rats containing a JV catheter had their bile-duct catheter loop left uninterrupted, enabling only blood collection. There were no significant differences in plasma DOX pharmacokinetics between bile-flow interrupted and uninterrupted groups (see below), following administration of free and liposomal DOX formulations, indicating a lack of changes in pharmacokinetics of DOX caused by surgical instrumentation. The following sections will therefore focus on results generated in instrumented rats containing JV as well as BD catheters. The non-encapsulated DOX plasma concentration-time data, both in the presence and absence of the MDR modulator, were fitted with a 2-compartment model, characterized by a rapid phase drug concentration decrease and a slower terminal elimination phase (Figure 5.1, Panel A). When non-encapsulated DOX was administered i.v. at a dose of 5 mg/kg, a model-derived estimated Cmax value of 17.61 ±1 .79 p,g/ml was obtained. The pharmacokinetics of non-encapsulated DOX was characterized by an AUC of 5.55 ± 0.3 ug.h/ml, a terminal half-life of 1.41 ± 0.2 h, and a plasma clearance (CLp) of 225.7 ±11 .9 ml/h. In the presence of PSC 833, a 3-fold increase in AUC (significant at p<0.05) was observed while Cmax values were minimally affected. This increase in AUC is consistent with earlier observations in mice (Krishna and Mayer, 1997), described in Chapter 3. Further, PSC 833 co-administration caused a 2-fold reduction in elimination Tl/2. When EPC/Chol DOX was administered at a dose of 5 mg/kg, the plasma concentration-time data was best fitted with a one compartment model characterized by a 230 mono-phasic elimination profile (Figure 5.1, Panel B). A Cmax of 129.9 ± 11.9 ug/ml and an AUC of 571.4 ± 102.2 ug.h/ml were obtained. Compared to the free DOX treatment in the absence of PSC 833, this represented a significant (p<0.05) 103-fold increase in AUC, confirming slower DOX removal from the plasma compartment following liposome delivery. An elimination Tm of 3.05 ± 0.52 h and a CL P value of 2.23 ± 0.42 ml/h were estimated. Again, compared to non-encapsulated DOX, liposomal delivery of DOX in EPC/Chol liposomes resulted in a 2-fold increase in Tm and a 100-fold reduction in CLp. When PSC 833 was co-administered with EPC/Chol DOX, a significant (p<0.05) 1.8-fold increase in AUC was observed. This increase in AUC by PSC 833 for EPC/Chol DOX is consistent with observations in mice presented in Chapter 4, and was not as profound as for non-encapsulated DOX. A 1.9-fold increase in elimination Ti/ 2 and a 1.8-fold decrease in CL P were also observed (Table 5.1; significant at p<0.05). PEG-DSPE/DSPC/Chol DOX administered at a dose of 5 mg/kg to rats resulted in a plasma elimination kinetics that were also fitted to a one-compartment system (Figure 5.1, Panel C), with an estimated C m a x of 232.8 ± 38.0 ug/ml. An AUC value of 3981 ± 183.6 ug.h/ml, elimination half life of 12.01 ± 1.44 h, and a CL P value of 0.31 ± 0.01 ml/h. PSC 833 did not significantly alter the pharmacokinetic parameters of DOX when administered in PEG-DSPE/DSPC/Chol liposomes, as indicated by comparable values, in the presence and absence of the MDR modulator, for AUC, T1/2, C m a x , and CL P values (Table 5.1). Consistent with observations in Chapter 3, the plasma clearance of liposomal DOX was considerably reduced compared to that of non-encapsulated drug. Specifically, compared to non-encapsulated drug, administration of DOX in EPC/Chol liposomes and 231 PEG-DSPE/DSPC/Chol liposomes decreased the CLp by 101-fold and 728-fold respectively. These observations are substantiated by concomitant increases in AUC of 103-fold and 724-fold for EPC/Chol and PEG-DSPE/DSPC/Chol liposome formulations, respectively. Furthermore, at the doses utilized here, no indications of toxicity were observed for the increased presence of liposomal drug in the plasma compartment for 48 h (for EPC/Chol DOX) and 72 h (for PEG-DSPE/DSPC/Chol). DOX metabolites were below detection limits in the plasma for most of the treatment groups. While free DOX treatment in the absence of PSC 833 resulted in detectable but non-quantifiable concentrations of DOXol, the 13-hydroxy metabolite of DOX, this metabolite was below assay detection limits in the presence of PSC 833 as well as in either liposomal DOX treatment groups both in the presence and absence of PSC 833. 5.3.2. Urinary Excretion Urine samples from rats treated with non-encapsulated and liposomal DOX in the presence and absence of PSC 833 were analyzed for DOX, it's primary metabolite in rats (DOXol), as well as for glucuronide and sulphate conjugates. The volume of urine collected was determined, which provided the urinary flow rate, from which the drug urinary excretion rate was determined. Urine flow rate averaged 0.5 - 1.0 ml/h, throughout the experimental time course for all study groups. The DOX urinary excretion profile (in the presence and absence of PSC 833) is presented in Figure 5.2A. Neither DOX nor co-administration with PSC 833 significantly altered urinary flow rate at the doses employed here. When non-encapsulated DOX is administered at a dose of 5 mg/kg, intact DOX is readily excreted in urine. Specifically, over 24 h, the percent DOX excreted in urine was 6.8% of the injected dose. This is 232 consistent with previous observations of DOX excretion in rats (Tavaloni and Guarino, 1980a; Yesair et al., 1972). DOXol was the only DOX metabolite present and then, only at low levels. Approximately 0.1% of the injected DOX dose was excreted as DOXol over the 24 h study period (data not shown). The DOX renal clearance (CLr) was estimated at 13.47 ± 9.25 ml/h (range of 5.65 - 23.69 ml/h; Table 5.2). In the presence of PSC 833, a significant (p<0.05) 3.5 fold decrease in CLr was observed. The cumulative amount of DOX and DOXol excreted over 24 h after administration of non-encapsulated DOX was not altered in the presence of PSC 833 (for DOX, compare 6.8% for non-encapsulated DOX and 6.4% for non-encapsulated DOX in the presence of PSC 833; for DOXol, compare 0.12% for non-encapsulated DOX and 0.47% for non-encapsulated DOX plus PSC 833). This correlated with the elevated plasma DOX levels when non-encapsulated DOX is co-administered with PSC 833 which offset MDR modulator induced decreases in DOX renal clearance. 233 -O-Free DOX Free DOX + P S C 8 3 3 -Q-EPC /Cho l DOX - • - EPC/Chol D O X * PSC 833 -A- PEG-DSPE/DSPC/Chol DOX - A - PEG-DSPE/DSPC/Chol D O X * PSC 833 Time (h) Figure 5.2: Urinary (Panel A) and biliary (Panel B ) DOX excretion profiles following administration of free, EPC/Chol, and PEG-DSPE/DSPC/Chol DOX at a DOX i.v. dose of 5 mg/kg, in the absence and presence of PSC 833 (50 mg/kg, p.o., 4 h prior to DOX). Data are represented as mean + standard deviation (n = 3/group). 234 Table 5.2: Summary of the urine and bile DOX pharmacokinetic parameters (mean ± standard deviation, n = 3/group) following administration of free, EPC/Chol DOX, and PEG-DSPE/DSPC/Chol DOX formulations at a DOX dose of i.v. 5 mg/kg, in the and absence and presence of PSC 833 (50 mg/kg, p.o., 4 h prior to DOX). Group Renal Clearance, CLr, ml/h (range) Biliary Clearance, CLb, ml/h (range) Free DOX 5 mg/kg 13.47 ±9.25 (5.65 - 23.69) 55.16 ± 19.2 (34.42 - 72.2) Free DOX 5 mg/kg + PSC 833 50 mg/kg 3.78 ±0.21* (3.5-3.95) 1.476 ±0.05* (1.43 - 1.53) EPC/Chol DOX 5 mg/kg 0.085 ±0.03 (0.054-0.113) 0.043 ± 0.03 (0.018 - 0.087) EPC/Chol DOX 5 mg/kg + PSC 833 50 mg/kg 0.049 ± 0.009 (0.045 - 0.059) 0.022 ± 0.007 (0.014 - 0.027) PEG-DSPE/DSPC/Chol DOX 5 mg/kg 0.004 ± 0.001 (0.003 - 0.005) 0.007 ± 0.002 (0.005 - 0.008) PEG-DSPE/DSPC/Chol DOX 5 mg/kg + PSC 833 50 mg/kg 0.003 ± 0.001 (0.002 - 0.004) 0.002 ±0.001 (0.001 - 0.004) Renal and biliary clearances were calculated using the equations described in Materials and Methods. Numbers in parentheses represent the range of the estimated clearance parameters. * Statistically significant from no PSC 833 group at p<0.05 (ANOVA) 235 DOX encapsulated in EPC/Chol liposomes administered at a dose of 5 mg/kg resulted in decreased excretion in the urine compared to non-encapsulated drug (Figure 5.2A). Again, neither EPC/Chol DOX nor PSC 833 co-administration resulted in any alterations in urinary flow rate. Cumulative DOX excreted in urine was estimated at 4.8% of injected dose after 24 h. As seen for non-encapsulated DOX, PSC 833 did not alter the cumulative amount of DOX excreted after injection of EPC/Chol DOX (compare 4.8% for EPC/Chol DOX and 4.14 % for EPC/Chol DOX plus PSC 833). The CLr for EPC/Chol DOX was estimated as 0.085 ± 0.03 ml/h, which was reduced by 1.7-fold to 0.049 ± 0.009 ml/h in the presence of PSC 833. This reduction in CL r of EPC/Chol DOX by PSC 833 was significantly less than that caused for non-encapsulated DOX. DOXol was below assay detection limits and therefore, could not be quantitated in rats treated with either EPC/Chol DOX alone or in combination with PSC 833. When PEG-DSPE/DSPC/Chol DOX was administered at a dose of 5 mg/kg, DOX excretion in urine was the slowest of the three DOX formulations studied here. No alterations in urinary output were caused by administration of DOX in these liposomes in the presence and absence of PSC 833. As shown in Figure 5.2A, the urinary elimination of DOX when presented in these liposomes (open and filled triangles) was minimal compared to either non-encapsulated or EPC/Chol encapsulated forms. Specifically, urinary clearance of DOX after injection in PEG-DSPE/DSPC/Chol encapsulated form represented 0.75% of the injected dose in urine over 48 h. A CL r value of 0.004 ± 0.001 ml/h (Table 5.2) was obtained, which was not statistically different from zero. PSC 833 co-administration did not cause significant alterations in urinary excretion, as indicated by a 236 comparable CLr value of 0.003 ± 0.001 ml/h, which was also not statistically different from zero. The percent DOX excreted in urine, in the absence and presence of PSC 833 (1.1% of injected dose at 48 h), were also comparable, indicating a lack of a PSC 833 effect on DOX urinary excretion when DOX is encapsulated in PEG-DSPE/DSPC/Chol liposomes. Due to limited urine samples, aliquots of pooled urine were analyzed for the presence of glucuronide and sulfate conjugates of DOX. For non-encapsulated DOX, 1.4% of total amount over 0-6 h was recovered as sulfate conjugated DOX, whereas there were no indication of glucuronide conjugates. For EPC/Chol DOX, 12.1%was recovered as the sulfate conjugate and 13.7% as the glucuronide conjugate over 0 - 6 h. While there was no indication of glucuronide conjugates with PEG-DSPE/DSPC/Chol DOX, 6.5% was recovered as sulfate conjugated DOX over 0-6 h. 5.3.3. Biliary Excretion DOX is rapidly excreted in bile following administration of non-encapsulated form at a dose of 5 mg/kg,. At the end of 24 h, 17% of administered DOX was excreted unchanged in bile. The biliary flow rate, monitored throughout the study, was not altered by either DOX or PSC 833 co-administration. Figure 5.2B represents the biliary excretion profile of DOX after injection of non-encapsulated and liposomal encapsulated DOX. The 13-hydroxy metabolite of DOX, DOXol, is also excreted in bile. Specifically, about 1.46% of injected DOX dose is excreted as unconjugated DOXol at the end of 24 h. A CLb of 55.16 ± 19.2 ml/h was obtained, which constituted 24% of the systemic plasma clearance. This is consistent with the value 22% reported by Speeg and Maldonado, 1994. In the presence of PSC 833, the cumulative DOX excreted at 24 h was significantly (p<0.05) reduced 5.6-fold to 3% (Figure 5.2B, compare open and solid circles). Further, PSC 833 237 caused a 37-fold reduction in CLb of non-encapsulated DOX (compare of 55.16 ± 19.2 ml/h for non-encapsulated DOX and 1.47 ± 0.05 ml/h for non-encapsulated DOX and PSC 833). PSC 833 also inhibits the biliary excretion of DOXol after i.v. injection of non-encapsulated DOX, where the percent DOXol in bile is reduced by 3-fold (data not shown). DOX is slowly excreted in bile when DOX encapsulated in EPC/Chol liposomes was administered i.v. at 5 mg/kg. Specifically, at the end of 24 h, DOX and DOXol excreted in bile accounts for 1.7% and 0.14% of injected dose, respectively. The CLb of DOX after injection of the EPC/Chol formulation was 0.043 ± 0.03 ml/h (Table 5.2). In the presence of PSC 833, the CLb was reduced by 2.0-fold and the percent of injected DOX excreted as intact DOX decreased from 1.7% to 0.9%. PSC 833 did not alter the percent of DOXol excreted in bile (0.15% of injected dose for rats receiving EPC/Chol DOX alone compared to 0.18% for rats injected with the EPC/Chol DOX and PSC 833 combination). When PEG-DSPE/DSPC/Chol DOX is administered i.v. at a dose of 5 mg/kg, DOX is excreted in bile in very low amounts. No effect on biliary flow rate was observed either by this formulation of liposomal DOX alone or in combination with PSC 833. Figure 5.2B illustrates the biliary excretion profile of PEG-DSPE/DSPC/Chol DOX, in the presence and absence of PSC 833. In the absence of PSC 833, a biliary clearance value of 0.007 ± 0.002 ml/h was obtained, which was not statistically different from zero (Table 5.2). This is substantiated by the fact that only 1% of the injected dose appeared in bile over 48 h. PSC 833 co-administration did not alter the CL b of PEG-DSPE/DSPC/Chol 238 DOX, as indicated by a comparable CLb value of 0.002 ± 0.001 ml/h, which was also not statistically different from zero. Aliquots of pooled bile were analyzed for the presence of glucuronide and sulfate conjugates of DOX in a similar fashion as described for urine. For non-encapsulated DOX, no glucuronide and sulfate conjugates were recovered, compared to 16.7% sulfate conjugate (0-6 h) and 1.7% glucuronide conjugate (0-6 h) for DOX administered in the EPC/Chol formulation. No conjugates were recovered in bile of rats injected with PEG-DSPE/DSPC/Chol DOX in the absence or presence of PSC 833. 5.4. Discussion The blockade of PGP by MDR modulators in extratumoral sites, particularly in excretory organs, such as the kidney (proximal tubule) and the liver (bile canaliculi) presents a potential problem associated with impaired renal and biliary clearance of anticancer agents such as DOX. Recent reports of studies in several laboratories have confirmed these drug-modulator pharmacokinetic interactions where significant reductions in renal and biliary clearances caused by the MDR modulator co-administration have been demonstrated. Speeg and Maldonado (1994) demonstrated a reduction in CL b of at least 76% for biliary clearance with DOX using a 2 mg/kg i.v. bolus dose of PSC 833. However, these authors administered DOX as a continuous i.v. infusion at a dose of 20 pg/min while PSC 833 was given as an i.v. bolus dose of 2 mg/kg. Given previous observations of significant differences in pharmacokinetics and toxicity (such as increases in peak concentrations by as much as 17.4-fold when DOX was given as an i.v. bolus) with continuous infusion versus bolus dosing (Cusack et al., 1993), it is possible that the reductions in CLb are likely to be more severe when DOX is administered over short 239 periods of time, as is typically applied in clinical situations. This is corroborated in the present study, where the experimental design of Chapters 3 and 4 were retained, which involved administering PSC 833 (50 mg/kg) as an oral gavage 4 h before an i.v. bolus dose of DOX. The PSC 833 mediated alterations in DOX pharmacokinetics for non-encapsulated and various liposomal formulations observed in the rat model described here were comparable to the alterations obtained in mice as presented in Chapters 3 and 4. The percent DOX excreted unchanged in urine (6.75%) and bile (16.8%) in the present study for non-encapsulated DOX at a dose of 5 mg/kg, correlates well with previous observations in several studies (Israel et al., 1978; Riggs et al., 1977; Tavoloni and Guarino, 1980a; Yesair et al., 1972). Further, the estimations of CLb as well as its contribution to the overall systemic clearance are similar to those reported by Speeg and Maldonado (1994). These results supported the comparisons here between non-encapsulated and liposomal DOX formulations in the presence and absence of PSC 833, given the abundance of literature of non-encapsulated DOX's hepatobiliary and renal disposition in the rat. The PSC 833-mediated reductions observed for biliary clearance of non-encapsulated DOX is more profound than those seen for renal clearance, likely due to greater contribution of hepatobiliary clearance of DOX relative to total systemic clearance. Dose independent urinary excretion (between 0-20 mg/kg, in rats) and plasma elimination of DOX in humans and rats has been a subject of various investigations (Tavaloni and Guarino, 1980a; Bronchud et al., 1990; Preiss et al., 1989; Vora and Boroujerdi, 1996). At the DOX dose used here (5 mg/kg), no alterations in urinary or biliary output were observed. This contrasts the anti-diuretic effect observed at a high DOX dose of 40 mg/kg 240 (Tavaloni and Guarino, 1980a) as well as the fact that 2 out of the 3 rats used in the Speeg and Maldonado study (1994) exhibited no net renal secretion suggesting that reductions in urine output may have been due to the continuous infusion of DOX. Significant differences in the renal and biliary handling of DOX arising from administration of non-encapsulated and liposomal DOX formulations were observed in this study. Utilizing a relatively leaky liposomal DOX formulation (EPC/Chol) and a long circulating PEG-DSPE/DSPC/Chol formulation that exhibits negligible DOX release resulted in varying degrees of PSC 833-mediated inhibition of DOX clearance for the respective liposome systems. The differences in DOX pharmacokinetics with these liposomal lipid compositions are consistent with literature reports (Gabizon et al., 1993; Spanjer et al., 1986; Mayer et al., 1989). The PSC 833-effect was more marked with EPC/Chol DOX liposomes and negligible with PEG-DSPE/DSPC/Chol DOX. It is important to point out here that EPC/Chol DOX is characterized by an ability to liberate over 50% of encapsulated drug within the 1st hour post-injection and consequently exhibits a component of drug distribution properties similar to non-encapsulated DOX. In contrast, the sterically stabilized liposome system is characterized by an ability to retain DOX for extended periods of time as well as decrease accumulation in liver tissue. While non-encapsulated DOX was rapidly eliminated in bile, DOX administered in EPC/Chol liposomes led to reduced drug excretion in urine and bile. PEG-DSPE/DSPC/Chol DOX injected i.v. into rats resulted in negligible DOX excretion in both urine and bile. Further, whereas PSC 833 caused a 3.6-fold decrease in CL r and 37.5-fold decrease in CLb of non-encapsulated DOX, it reduced the CL r and CL b of DOX after injection of the EPC/Chol formulation by only 1.7-fold and 2.0-fold, respectively. In contrast, PSC 833 did not alter 241 renal or biliary clearance of DOX after administration of DOX encapsulated in PEG-DSPE/DSPC/Chol liposomes. Biliary and urinary excretion are predominant pathways of DOX elimination in rats, humans, and many other species. In rats, more than 90% of DOX is excreted as intact DOX in urine and bile, with the remainder as DOXol (approximately 3%), glucuronide and sulphate conjugates. It has also been previously shown (Parker et al, 1982b) that the percent of DOX excreted as conjugates is at least 5-10 fold higher with DOX encapsulated in liposomes compared to free drug. This is corroborated in the present study, where higher levels of conjugated DOX were observed, particularly with EPC/Chol DOX system, with a general relationship: EPC/Chol DOX (highest) > PEG-DSPE/DSPC/Chol DOX (intermediate) > non-encapsulated DOX (least). This suggests an increased conjugation of DOX occurs when it is entrapped in liposomes that accumulate at increased levels in the liver. Reductions in toxicity observed with liposomal DOX systems may therefore be due to this enhanced formation of polar, non-toxic glucuronide and sulphate conjugates, which are readily excreted in urine and bile, compared to the more toxic parent DOX or its aglycone metabolites. Liposomes have been shown to be predominantly taken up by the Kupffer cells in the liver (Parker etal, 1982b; Moghimi and Patel, 1993; Scherphof et al, 1983a). Also, liposomes do not permeate across hepatocyte membranes per se and consequently drug has to be released from the liposomes in order to enter these cells. These observations are also consistent with the results of this study and those in Chapter 4 where indications of altered metabolic processes of DOX occur with liposome encapsulation. Further, previous studies have indicated slower metabolism of liposome entrapped drugs (Hunt et al, 1979; 242 Kimelberg et al., 1976). This is also consistent with the observations here of reduced DOXol levels following liposomal DOX administration (DOX is metabolized to DOXol by cytosolic aldo-keto reductase) in Chapter 4. Further, in Chapter 4, it was observed that PSC 833 co-administration with non-encapsulated DOX resulted in significant increases in DOX, DOXol, and DOXone (aglycone) levels. However, with liposomes, reductions in aglycone levels were observed with PEG-DSPE/DSPC/Chol liposomes in addition to DOXol formation not being favored, with the EPC/Chol liposomal DOX resembling non-encapsulated DOX in that higher levels of aglycone were present. The presence of higher levels of polar non-toxic conjugates with EPC/Chol liposomal DOX may in part explain the reductions in toxicity of liposomal DOX, despite the enhanced delivery of DOX to the liver. Further, it is unlikely that liposomes are directly excreted intact in bile due to size limitations for their passage into the bile canaliculi (Parker et al., 1982b), which was substantiated by negligible levels of tritiated lipid being detected here in bile. Liposomal encapsulation of DOX reduced the extent to which it was excreted in urine. These observations are consistent with reports of reduced drug renal clearance following administration of liposomal formulations of DOX (Parker et al., 1982b; Martin, 1998), cytosine arabinoside (Parker et al., 1982a), and methotrexate (Kimelberg et al., 1976). This has been postulated to be due to the low renal clearance rate of liposomal carriers themselves thus modifying the clearance behavior of the encapsulated agents (Parker et al., 1982b; Zirenberg and Betzing, 1979; Martin, 1998). The renal clearance of liposomal DOX (alone or with PSC 833) was much lower than that for the free DOX-PSC 833 combination. It is likely that co-administration of 243 PSC 833 and free DOX leads to saturation of the pathways for DOX metabolism and excretion due to increased DOX exposure mediated by the MDR modulator, whereas administration of liposomal DOX (alone or with PSC 833) results in lower concentrations of DOX exposed to the excretory organs via slower release of the drug from liposomes. In support of this, alterations in DOX renal excretion have been shown to occur at high doses of DOX (~40 mg/kg), indicating the limited capacity of the kidneys to eliminate DOX when presented at high concentrations (Tavoloni and Guarino, 1980a). Presumably, PSC 83 3-mediated increases in free DOX exposure to the kidneys may result in saturation of the urinary transport processes leading to reduced capacity of the kidneys to eliminate DOX. Taken together, these results suggest that the slower urinary and biliary elimination by liposomes may explain the lower propensity of PSC 833 to exert a profound inhibitory effect on biliary excretion of DOX. The utility of liposomes to reduce the toxicity of DOX and effects of PSC 833 on DOX pharmacokinetics may be related to the ability of DOX delivered to phagocytic Kuppfer cells to be metabolized by different processes such that metabolites are neither toxic nor PGP substrates. Liposomal delivery to the liver may result in much lower DOX and DOX metabolite levels over much greater lengths of time such that even under conditions of PSC 833 mediated inhibition of PGP, the renal and biliary excretion capacity is sufficient to handle the levels of DOX and DOX metabolites exposed to these tissues. 244 Chapter 6 SUMMARIZING DISCUSSION Reports in the literature investigating the modulation of multidrug resistance (MDR) through chemosensitization have led to the identification of numerous chemical entities capable of reversing MDR to varying degrees. However, lack of a thorough understanding of the cellular pharmacology and molecular mechanisms of MDR reversal have limited the ability to draw definitive conclusions from preclinical and clinical studies. Specifically, issues such as the importance of residual resistance as well as the effect of modulator clearance on anticancer drug uptake and cytotoxicity characteristics have not been addressed adequately to resolve key pharmacodynamic relationships that impact MDR reversal in vivo. The results of Chapter 2 suggest that screening studies for potential MDR modulators should evaluate the residual resistance and latency associated with MDR reversal. For example, agents with low residual resistance as well as those with latent modulating activity are more likely to demonstrate improved activity in vivo. Such properties may be critical for modulating PGP mediated MDR in vivo where chemosensitization is complicated by difficulties in synchronizing therapeutic levels of modulator and anticancer drug at the tumor site. The issue of modulator-induced alterations in anticancer drug pharmacokinetics introduces another layer of complexity to the MDR modulation strategies. This is due to the lack of selectivity of current MDR modulators for the tumor tissue presenting a major dilemma in envisioning strategies for reversal. MDR modulators may elicit adverse pharmacokinetic interactions by competing for drug metabolizing enzymes as well as by 245 impairing renal and biliary excretion of anticancer drugs such as DOX, given the presence of PGP in the kidney proximal tubule and biliary canaliculi. Over recent years, in addition to potent second generation agents which are currently in advanced clinical trials (such as PSC 833) several so-called third generation modulators have also been identified (such as LY335979, GW918). While these agents are not only potent and perhaps even selective for different MDR mechanisms (PGP or MRP), they have still been shown to alter the pharmacokinetics of anticancer drugs via competition for metabolism (such as taxol, etoposide) or blockade of the drug efflux pump in excretory organs such as the liver and the kidney (such as DOX, daunorubicin), impairing drug elimination. Although attention has been focused on the rena