Phd Ifergan Ilan 031810245

The Role of Transporters in Folate Homeostasis and Anticancer Drug Resistance Research Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

Submitted to the Senate of the Technion Israel Institute of Technology University of Haifa - The Graduate School

(Adar 5767 March 2007)

The Role of Transporters in Folate Homeostasis and Anticancer Drug Resistance Research Thesis Submitted in Partial Fulfillment of the Requirements For the Degree of Doctor of Philosophy Ilan Ifergan Submitted to the Senate of the Technion Israel Institute of Technology University of Haifa - The Graduate School (Adar 5767 Haifa March 2007)

II

The Research Thesis Was Carried Out Within The Framework Of The Joint Program In Economics Of The Technion - Israel Institute Of Technology And The University Of Haifa Under The Supervision Of Professor Yehuda G. Assaraf in the Faculty of Biology.

The Generous Financial Help Of The Fred Wyszkowski Cancer Research Fund Is Gratefully Acknowledged.

III

Contents Abstract ………………………………………………………………….……....1 Abbreviations ……………………………………………………………….…. .3 Introduction ………………………………………………………………….…..5 Folic acid and biologically active folates …………………………………...…….5 Intracellular metabolism of folic acid …………………………………………….6 MTX as an antifolate anticancer drug …………………………………………….7 Cellular uptake of folates and MTX …………………………………..……….….8 Cellular retention of folates and MTX ……………………………………….……9 Mechanisms of resistance to MTX …………………………..………………...…11 The superfamily of ABC transporters ……………………………………….……13 BCRP (ABCG2), a novel mediator of anticancer drug resistance ………...….…..16 The BCRP gene ………………………………………………………...…………17 The MRPs ……………………………………….…………………………..……17 MRP1 ……………………………………………………………………….……..18 The role of the ABC transporters in folate homeostasis ………………………….18 Research Aims ……………………………………………………….…..……….20 Description of the research……………………………………………………….24 Materials and Methods ……………………………………………….………….25 A) The reduced folate carrier mediates intracellular folate depletion and consequent cytotoxicity under folate deprivation……………………………………………….25 RFC dependent cellular proliferation in the presence or absence of folates……….25 Folate deprivation.....................................................................................................27. RNA extraction and Quantitative RT-PCR...............................................................28 [3H] MTX initial rate of uptake measurements.........................................................30 Statistical Analysis.....................................................................................................30 B) Folate deprivation results in the loss of breast cancer resistance protein (BCRP/ABCG2) expression: A role for BCRP in cellular folate homeostasis……..30 Materials………………………………………………………...…………………..30 Tissue culture and folic acid deprivation ………………………….….……………..31 Growth inhibition with Mitoxantrone and MTX ………………….………..………31 Extraction of membrane proteins from cultured cells …………………..….………32 Western blot analysis of MRPs and BCRP expression …………………..……..….32

IV

Contents - continuance Immunohistochemistry studies………………………………………….………….33 Online efflux of Hoechst 33342 …………………………………….…..…….……33 [3H]Folic acid accumulation……………………………………………………..…34 Flow cytometric analysis of Mitoxantrone staining ………………………….…….35 FPGS activity assay …………………………………………………………..….…35 Semi-quantitative RT-PCR and DNA sequencing ………………………….…...…35 Scanning densitometry ………………………………………………………..……36 Statistical analysis …………………………………………………………….....…36 C) Cytoplasmic Confinement of Breast Cancer Resistance Protein (BCRP/ABCG2) as a Novel Mechanism of Adaptation to Short-Term Folate Deprivation ……..…..36 Chemicals...................................................................................................................36 Tissue culture. ...........................................................................................................37 Western Blot Analysis of BCRP, MRP1, and Pgp expression .................................37 Immunohistochemistry studies .................................................................................38. Immunofluorescence analysis with viable cells .......................................................39 Confocal and immunofluorescence microscopy studies with fixed cells …………40 Propidium iodide (PI) staining and cell cycle analysis .............................................40 Assay of cellular rhodamine 123 accumulation .......................................................41 Quantitative analysis of the cytoplasmic and plasma membrane fractions of BCRP..41 [3H]Folic acid accumulation ......................................................................................42 Statistical analysis ......................................................................................................42 Scanning densitometry ...............................................................................................42 D) Novel extracellular vesicles mediate an ABCG2-dependent anticancer drug sequestration and resistance……………………………………………………..….43 Chemicals ……………………………………………………………….…….……43 Tissue culture and growth inhibition with mitoxantrone……………………….…..43 Western blot analysis of BCRP……………………………………………………..44 Mitoxantrone accumulation and immunohistochemical localization of BCRP in specific colonies of MCF-7/MR and MCF-7/FLV 1000 cells ………...…..……44 Determination of the number of light-refracting extracellular vesicles……….……44 Inhibition of mitoxantrone accumulation with BCRP transport inhibitors and ATPdepleting agents ……………………………………………………………………45 Estimation of the intravesicular concentration of mitoxantrone ………………...….45 V

Contents - continuance Autofluorescence detection with viable cells………………………………….….46 Confocal microscopy of BCRP confinement to cell-cell attachment zones ……...46 Confocal microscopy studies of the accessibility of the culture medium to the extracellular vesicles …………………………………………………………...…46 Electron microscopy studies …………………………………………………..….47 Statistical analysis ……………………………………………………………..….48 Results……………………………………………………………………….…...49 A) The reduced folate carrier mediates intracellular folate depletion and consequent cytotoxicity under folate deprivation………………………………………………49 Effect of RFC overexpression on cellular proliferation under folate deplete conditions ……………………………………………………………………………………...49 Gene expression status of folate influx and efflux transporters as well as folatedependent enzymes under folate deplete- and replete conditions ............................52 Decreased RFC activity in folate-deprived cells ……………………………...….53 B) Folate deprivation results in the loss of breast cancer resistance protein (BCRP/ABCG2) expression: A role for BCRP in cellular folate homeostasis……..54 Loss of BCRP expression in the LF-adapted cell lines as revealed by Western blot analysis………………………………………………………………………..…….54 Retention of poor MRP2 through MRP5 expression in the LF-adapted cell lines….55 Poor BCRP gene expression in the LF-adapted cell lines as revealed by RTPCR………………………………………………………………………………….56 Loss of BCRP expression in the LF-adapted cell lines as revealed by immunohistochemistry ………………………………………………………...……57 Loss of Hoechst 33342 efflux in the LF-adapted cell lines…………………….……58 Accumulation of Mitoxantrone in the LF-adapted cell lines as revealed by flow cytometry……………………………………………………………………….…….59 Sensitivity of the LF-adapted cell lines to mitoxantrone ……………………...…….60 Loss of MTX-resistance in MCF/MR-LF cells ………………………………..……63 Increased accumulation of [3H]Folic acid in the LF-adapted cell lines …….….……65 Increased FPGS activity in the LF-adapted cell lines ………………………….……66 C) Cytoplasmic Confinement of Breast Cancer Resistance Protein (BCRP/ABCG2) as a Novel Mechanism of Adaptation to Short-Term Folate Deprivation.………….…..68 Establishment of a Short-Term Folate Deprivation Protocol ………………………..68 VI

Contents - continuance Expression and Glycosylation of BCRP in Short-Term Folate-Deprived Cells and Their Control Counterparts…………………………………………..…………….70 Subcellular Localization of BCRP in Short-Term Folate-Deprived Cells and Their Control Counterparts ……………………………………………..……………….72 Retention of Plasma Membrane Localization of Various Membrane Proteins in the Short-Term Folate-Deprived Cells ……………………….………………...……..76 Colocalization of BCRP in the ER Compartment in Folate-Deprived Cells……....78 Functionality of BCRP in the Various Cell Lines ………………………………...79 D) Novel extracellular vesicles mediate a BCRP -dependent anticancer drug sequestration and resistance ………………………………………………….……83 Overexpression and immunolocalization of BCRP to extracellular vesicles in mitoxantrone-resistant MCF-7/MR cells ……………………………..……………83 Intravesicular concentration of mitoxantrone in an ATP- and BCRP -dependent manner………………………………………………………….…………………...87 Intravesicular concentration of an endogenous green fluorescent chromophore ......91 Discussion ……………………………………………………………………….…95 References …………………………………………………………………….…..115

VII

Figures Fig. 1: Chemical structure of folic acid …………………………..………………5 Fig. 2: Chemical structures of reduced folates and MTX ……………..………..6 Fig. 3: Folate metabolism in mammalian cells …………………………..………7 Fig. 4: Membrane topology of the hRFC…………………………………………9 Fig. 5: The reaction of folate polyglutamylation.…………………….………….10 Fig. 6: Predicted topology of the major classes of mammalian ABC transporters ………………………………………………………………………………………14 Fig. 7: Intracellular folate metabolism model under replete (A) and deplete (B) conditions …………………………………………………………………………..50 Fig. 8: Effect of RFC overexpression on cellular proliferation under folate deplete conditions. ………………………………………………………………....51 Fig. 9: Gene expression status of folate influx and efflux transporters as well as folate-dependent enzymes under folate deplete- and replete conditions………..53 Fig. 10: [3H]MTX transport in folate supplemented and deprived sublines .…54 Fig. 11: Western blot analysis of BCRP as well as MRP1 through MRP5 expression in parental cells and their LF-adapted cell lines ……………………56 Fig. 12: Immunohistochemical detection of BCRP expression in parental cells and their LF-adapted cell lines……………………………………………………57 Fig. 13: Online efflux of Hoechst 33342 from monolayers of parental and LFadapted cell lines ……………………………………………………………….….59 Fig. 14: Flow cytometric analysis of Mitoxantrone accumulation in parental cells and the LF- adapted cell lines ………………………………………………….…60 Fig. 15: Cellular growth inhibition with Mitoxantrone ……………………....…62 Fig. 16: Cellular growth inhibition with MTX …………………………...………64 Fig. 17: [3H] Folic acid accumulation in parental and LF-adapted cell lines…..66 Fig. 18: Histogram of FPGS activity in parental cells and their LF-adapted cell lines ………………………………………………………………………………….67 Fig. 19: Schematic presentation of the short-term folate deprivation protocol… 69 Fig. 20: Cell cycle analysis of folate-deprived cells and their control counterparts…………………………………………………………………………70

VIII

Fig. 21: Western blot analysis of BCRP, MRP1, and Pgp in folate-deprived cells and their control counterparts…………………………………………………….72 Fig. 22: Immunohistochemical and immunofluorescence detection of BCRP in parental cells and their folate-deprived cells…………………………………......74 Fig. 23: Histograms comparing the plasma membrane and cytoplasmic fractions of BCRP in folate-deprived cells and their control counterparts………………76 Fig. 24: Immunohistochemistry and immunofluorescence localization of various plasma membrane proteins in folate-deprived cells and their parental counterparts………………………………………………………………………..77 Fig. 25: Colocalization of BCRP in the ER compartment in folate-deprived cells as revealed by confocal microscopy……………………………………………….79 Fig. 26: Histogram of rhodamine 123 accumulation in folate-deprived cells and their control counterparts………………………………………………………….81 Fig. 27: [3H] Folic acid accumulation in parental and short-term folate-deprived cells……………………………………………………………………………….….82 Fig. 28: Histogram comparing the association between the percentage of the cytoplasmic BCRP versus the number of cells in the different colonies of folatedeprived cells (C) and their control counterparts (A and B)……………….……83 Fig. 29: Cellular growth inhibition with mitoxantrone ……………………….…84 Fig. 30: Immunohistochemical localization of BCRP in parental MCF-7 cells and their MCF-7/MR subline ………………………….. ………………………...……85 Fig. 31: Transmission electron microscopy analysis of the extracellular vesicles in monolayers of MCF-7/MR cells……………………………………………………86 Fig.32: Mitoxantrone accumulation in extracellular vesicles and immunohistochemical localization of BCRP in MCF-7/MR and MCF-7/FLV1000 cells …………………………………………..………………………………………89 Fig. 33: Prevention of intravesicular mitoxantrone accumulation by BCRP transport inhibitors and metabolic energy deprivation……………………….….91 Fig. 34: Detection of intravesicular green autofluorescence in viable MCF-7/MR cells…………………………………………………………………………………..93 Fig 35: Novel model of extracellular vesicles that serve as cytotoxic drug disposal chambers shared by multiple neighbor cancer cells ……………………………..94

IX

X

Abstract The role of plasma membrane transporters in drug resistance and homeostasis of folate vitamins is of great physiological and therapeutic importance. The reduced folate carrier (RFC) is the primary high-affinity bi-directional transporter for reduced folate cofactors (B9 vitamin) essential for nucleotide biosynthesis and thus DNA replication. We hence hypothesized that under conditions of folate deprivation, the folate efflux activity of the RFC may result in intracellular folate depletion and consequently decreased cellular proliferation. Here we show that the cellular proliferation rate is significantly decreased in Chinese hamster ovary (CHO) C5 RFC cells with RFC overexpression relative to RFC null C5/RFC cells or C5/folate receptor (FR) cells overexpressing the FRα only under folate-free growth conditions. Moreover, the mRNA levels and activity of RFC were significantly decreased upon 37 days of folate deprivation in several cell lines. We conclude that upon folate deprivation, RFC activity becomes detrimental as RFC extrudes folate monoglutamates out of cells. Hence, we suggest that this cytotoxic folate efflux activity may be abrogated by a novel adaptive down-regulation of RFC. In contrast to RFC, the breast cancer resistance protein (BCRP/ABCG2) is currently the only known transporter that exports both mono-, di-, and triglutamate conjugates of folate and the antifolate methotrexate (MTX)). We therefore explored the relationship between cellular folate status and BCRP expression as well as transport function. Toward this end, MCF-7 breast cancer cells, with low BCRP protein levels, and their MR (i.e. mitoxantrone, an anticancer drug which is a specific BCRP substrate)resistant MCF-7/MR subline, with BCRP overexpression were gradually deprived (three months) of folic acid from 2.3 µM to 3 nM resulting in the sublines MCF-7/LF

1

and MCF-7/MR-LF, respectively. These low folate adapted sublines displayed only residual mRNA, protein levels and activity of BCRP. Additionally, the low folate adapted sublines displayed a ~2-fold increase in the 4h accumulation of [3H]folic acid along with a significant increase in folylpoly-γ-glutamate synthetase activity (FPGS), an enzyme that catalyzes the conversion of folate monoglutamates to polyglutamates that no longer serve as BCRP substrates. Moreover, we found that cellular adaptation of MCF7/MR cells to short-term (two weeks) folate deprivation is associated with a selective confinement of BCRP to the endoplasmic reticulum instead of the plasma membrane. Hence, consistent with the mono- and polyglutamate folate exporter function of BCRP, down-regulation of BCRP, increased FPGS activity and selective confinement of BCRP to the endoplasmic reticulum appear to be crucial components of cellular folate homeostasis. During these BCRP cellular confinement studies, we discovered that this transporter is highly confined to cell-cell attachment zones in the MCF-7 breast cancer sublines MCF-7/MR and MCF-7/FLV1000 in which wild type (R482) BCRP is overexpressed. The cell-cell attachment zones were found to be the membrane of novel extracellular vesicles in which mitoxantrone was rapidly, dramatically and specifically sequestered via BCRP. We suggested that the novel extracellular vesicles serve as cytotoxic drug disposal chambers shared by multiple neighbor cancer cells. This finding constitutes a novel modality of anticancer drug resistance.

2

Abbreviations ABC-ATP binding cassette AICAR - Aminoimidazole carboxamide ribonucleotide transformylase ALL - acute lymphoblastic leukemia BCRP - Breast cancer resistance protein CHO - Chinese hamster ovary DHF- Dihydrofolate DHFR - Dihydrofolate reductase DNA - Deoxyribonucleic acid EDTA- Ethylenediamine tetra acetic acid EGFR - epidermal growth factor receptor FCS- Fetal calf serum FGFR - Fibroblast growth factor receptor FPGS - Folylpoly-γ -glutamate synthetase FR - Folate receptor FTC - fumitremorgin C GARTF- Glycinamide ribonucleotide transformylase GGH - γ-glutamate hydrolase GSH - glutathione HF - high folate HFM - hereditary folate malabsorption hRFC - Human Reduced folate carrier HRP - Horseradish peroxidase IC50 - the drug concentration inhibiting cell growth by 50% Kd - Kilo dalton

3

LCV- Leucovorin LF - low folate M- Molar MDR - Multidrug resistance mM- millimolar MRP - Multidrug Resistance Protein MRP1- Multidrug resistance protein 1 MR- Mitoxantrone MTX- Methotrexate M.W. - Molecular weigh NBDs-Nucleotide-binding domains NF- No folate nM- Nanomolar PABA - para amino benzoic acid PBS- Phosphate-buffered saline PI3K- Phosphatidylinositol-3-kinase RFC - Reduced folate carrier PRPP- Phosphoribosyl-1-pyrophosphate R.T- Room temperature THF- Tetrahydrofolate TMD - Transmembrane domain TRITC - Tetramethylrhodamine isothiocyanate TS - Thymidylate synthase

4

Introduction

Folic acid and biologically active folatesFolic acid and its reduced forms play an essential role as one–carbon donors in several biosynthetic reactions, including the biosynthesis of purine and pyrimidine precursors of nucleic acids, the metabolism of certain amino acids as well as the initiation of macromolecular synthesis in the mitochondria [1, 2]. Folates cannot be synthesized by mammalian cells and therefore must be consumed from exogenous sources; one of the best dietary sources of folates is green-leafy vegetables. Folic acid is composed of three structural components: pteridine ring, p-amino benzoic acid (PABA) and glutamic acid (Fig 1).

Fig 1: Chemical structure of folic acid. Biologically active folates exist predominantly in the reduced form, i.e. the two double bonds on the pteridine ring are enzymatically reduced (Fig 2).

5

Fig 2: Chemical structures of reduced folates and MTX. Intracellular metabolism of folic acid: Folates are absolutely essential for cellular proliferation due to their key role in purines and pyrimidine biosynthesis. Several key enzymes use folates either as cofactors or as substrates in these biosynthetic pathways (Fig 3). Once folic acid enters the cell, it is rapidly reducted to dihydrofolate (DHF) by the enzyme dihydrofolate reductase (DHFR). The latter enzyme also catalyzes the further reduction of DHF to tetrahydrofolate (THF). Thymidylate synthase (TS) catalyzes the conversion of dUMP to dTMP through the transfer of one carbon unit from 5,10methylene-THF to dUMP. Glycinamide ribonucleotide transformylase (GARTF) transfers a formyl group from 10-CHO-THF to the ribonucleotide following another formyl-group transfer reaction that is carried out by aminoimidazole carboxamide ribonucleotide transformylase (AICARTF), the following reactions lead to the formation of the purines AMP and GMP.

6

Fig 3: Folate metabolism in mammalian cells. MTX as an antifolate anticancer drug: The finding that folates are essential vitamins for the growth and proliferation of neoplastic cells has been exploited for the introduction of folate antagonists (antifolates) that block or inhibit the biosynthesis of purines and pyrimidines. Therefore, antifolates are being used in diseases that are characterized by abnormal cellular proliferation as in human malignancies, rheumatoid arthritis and psoriasis [3]. Methotrexate (MTX, Fig. 2) is an antifolate that has become increasingly important in cancer chemotherapy [4]. MTX binds (stoichiometrically) to the enzyme DHFR very tightly (KD =1 pM) with a million fold higher affinity than the natural substrate DHF (Km = 1µM) [5]. Studies have shown that at least 95% of DHFR must be inhibited in order to block cell growth [6]. The cytotoxic activity of MTX and other antifolates is due to inhibition of DNA synthesis as well as misincorporation of dUTP into DNA, thereby resulting in DNA strand breaks and cell death [7]. MTX is

7

currently used for the chemotherapeutic treatment of various human cancers including childhood acute lymphoblastic leukemia (ALL), non-hodgkin’s lymphoma, osteosarcoma, head and neck cancer, choriocarcinoma, small cell lung cancer, and breast cancer [4, 8]. Cellular uptake of folates and MTX: Folic acid, reduced folates and MTX are divalent anions; hence, their uptake into cells relies on specific transport proteins. Several transport systems are known to accommodate the uptake of folates and antifolates through biological membranes: a) The reduced folate carrier (RFC, Fig 4 ) is the major uptake route that functions as a bi-directional anion exchanger [9, 10] with a high affinity (Km= 1 µM) for reduced folates and MTX but low affinity (Km=200-400 µM) for folic acid [10-12]. Human RFC (hRFC) is a plasma membrane protein with 591 amino acids. According to hydropathy plots and membrane topology studies, the transporter contains 12 transmembrane domains (TMDs), has a short cytosolic N-terminus and long cytosolic C-terminus [13] . RFC contains one consensus site for N -linked glycosylation. The core molecular weight of the RFC is 64 kDa, but the extensive glycosylation it undergoes results in a broadly migrating protein with a molecular mass of ~ 85 kDa [14]. b) Folate receptors, glycosylphosphatidylinositol membrane-anchored proteins that mediate the unidirectional uptake of folates, display a high affinity for folic acid and 5-methyltetrahydrofolate (KD=0.1-10 nM) but lower affinity (KD=10-300 nM) for other reduced folates and MTX [15-18]. c) An apparently independent transport system with optimal uptake activity at low pH, which recognizes folic acid, reduced folates and MTX with comparable affinities (Km= 1-5 µM) [19-22]. This transporter has been recently cloned and termed protoncoupled folate transporter (PCFT/HCP-1) [23]. Furthermore, it was recently

8

discovered that inactivating mutations in this folate transporter results in congenital hereditary folate malabsorption

Fig 4: Membrane topology of the hRFC. Cellular retention of folates and MTX: The intracellular retention of folates and MTX occurs through a mechanism of polyglutamylation. The latter is a unique metabolic process in which multiple equivalents of glutamate are added to the γ-carboxyl residue of folates; this reaction is carried out by the enzyme FPGS (Fig .5). Intracellular (anti)folates exist mainly as poly-γ-glutamate derivatives, with the original (anti)folate being elongated by 7-10 glutamyl residues [24, 25]. Polyglutamylated folate derivatives are no longer substrates for multidrug resistance protein 1 (MRP1) which is the major efflux transporter that exports unglutamylated folates and hydrophilic antifolates out of cells [26, 27]. Hence, polyglutamylation mediates intracellular retention of folates. 9

Polyglutamylated folate derivatives have higher affinities for various folate-dependent enzymes than their non-glutamylated forms [28, 29]. Furthermore, an organellar species of FPGS exists in the mitochondrion which is absolutely essential for the biosynthesis of glycine.

Fig 5: The reaction of folate polyglutamylation.

10

Mechanisms of resistance to MTX: Several mechanisms of resistance to MTX have been described over the past 58 years: 1) Impaired transport of MTX into the cell Defective or altered influx of antifolates via alterations in the RFC either through decreased expression of the transporter or via mutations that results in increased Km and/or decreased Vmax have been described. These resistance modalities were described both in vitro as well as in vivo [30-32]. When cultured cells acquire resistance to MTX under usual growth conditions, the requirement for folates is met through the uptake of folic acid, the oxidized folate species in most media. Transport of folic acid can be also mediated by processes that are distinct from RFC [16, 17]. However, survival of tumor cells that develop MTX resistance due to the loss of RFC activity in vivo, where the folate substrate in the blood 5-CH3-THF is transported by the same mechanism, is not well understood. However, at least one report described that acquisition of antifolate resistance is associated with an increased activity of the low pH folate transporter [22]. Hence, it is possible that PCFT is overexpressed in antifolate transport-deficient cells. 2) Increased efflux of MTX – To date, there are six ATP-driven, unidirectional ABC (ATP binding cassette) transporters, including MRP1-5 [33-37] and the breast cancer resistance protein (BCRP) [38] that actively pump MTX out of cells. The overexpression of MRPs 1–5 [33, 37, 39, 40] has been shown to reduce MTX accumulation thereby leading to decreased MTX sensitivity only in short-term (4 hours) exposure assays. The restriction of this antifolate resistance to only a short-term drug exposure has been attributed to the ability of these transporters to export only monoglutamate but not polyglutamate conjugates of MTX; indeed, these transporters failed to transport MTX

11

diglutamates or longer chain polyglutamates [26, 40]. However, even in the absence of MRP overexpression [41] the MCF7/MX cells displayed MTX resistance to longterm (7 days) MTX exposure. It has been shown that this long-term resistance to MTX is mediated by BCRP [38]. The resistance to MTX is mediated by the wild type BCRP (R482) [38, 41]. However, by contrast to the wild-type BCRP, two mutant BCRP forms threonine 482 and glycine 482 completely appear to have lost their ability to transport folates and MTX in a vesicle system [42-44] and therefore displayed a wild type sensitivity to long-term exposure to MTX (44). In contrast, recently we have shown that Gly 482 and Thr 482 BCRP mediates a high level resistance to short-term (4hr) antifolate exposure thereby reinforcing the role of BCRP in a clinically relevant treatment schedule [45]. Increased MRP1 expression has been established as a mechanism of resistance to MTX [33] and other anticancer drugs [34, 46]. Conversely, loss of MRP1 expression and function along with RFC overexpression resulted in significant expansion of cellular folate pools [47]. Consequently, cells became highly resistant to various antifolates. 3) Overexpression of DHFR Amplification of the DHFR gene after treatment with MTX has been documented both in cultured cells [48] and in clinical samples from patients that were treated with the drug. For certain concentrations of MTX, the increase in the intracellular levels of DHFR produces more free enzyme to carry out its biosynthetic reaction. 4) Altered DHFR Mutations in DHFR can result in a dramatically decreased affinity for MTX [49]. An example for altered DHFR that mediates such resistance is mutant 3T6 fibroblasts which were found to display a 270-fold lower affinity to MTX than normal DHFR [49]; this resulted in a high level resistance to MTX.

12

5) Decreased polyglutamylation Decreased FPGS activity leads to substantial reduction in MTX polyglutamylation, hence the unglutamylated MTX is pumped out of the cells via MRP1, as well as MRP2-4 and BCRP. Decreased FPGS expression mediated resistance to MTX both in vitro [50] and in vivo [32, 51]. A point mutation in the human FPGS was recently discovered that markedly decreased FPGS activity and resulted in MTX resistance via decreased polyglutamylation [52]. The superfamily ABC transporters: The term ABC transporters was introduced in 1992 by Chris Higgins in a memorable review [53], the initials ABC were based on the highly conserved ATP-Binding Cassette. The ATP-binding cassette is the most characteristic feature of this superfamily. Several other names are used for this family, including, Traffic ATPases and P-glycoproteins (Pgps). There are 49 known and putative human ABC transporters [54], however, only 24 of them are with a known function and/or involvement in diseases. The ABC transporters are transmembrane proteins that bind and hydrolyze ATP thereby driving the vectorial transport of various substrates across cell membranes [55-57]. The basic structure of the ABC transporters as exemplified by P-glycoprotein (ABCB1) is thought to consist of 12 transmembrane domains (TMDs) and two ATP-binding sites in a protein of about 1,300 amino acids (Fig 6). This basic structure may be assembled from two nearly equal (BCRP) or unequal halves (ABCG5 and ABCG8). Several ABC transporters (e.g., MRP1) have additional domains and are even larger than P-glycoprotein. For example, MRP1 contains an additional N-terminal spanning domain.

13

Figure 6: Predicted topology of the major classes of mammalian ABC transporters [34]. This simplified scheme shows the intracellular nucleotide-binding domains (NBDs) and the transmembrane segments and indicates the N- and Ctermini of the transporter. Note that the predicted topology is often based on minimal data, as in the case of ABCG2 (BCRP1/MXR/ABCP) and MRP5 (ABCC5). TAP, the transporter associated with T-cell antigen presentation, probably has more than six transmembrane segments[58, 59]. The half-size transporter TAP functions as a heterodimer of TAP1 and TAP2, and BCRP presumably functions as a homodimer.

The ABC transporters have been shown to play a key role in the transport of drugs (xenotoxins) and drug conjugates. This role is exemplified by the multidrug resistance transporters P-glycoprotein, MRP1 (ABCC1), or BCRP1 (MXR, ABCP, ABCG2), each of which can cause multidrug resistance (MDR) in cancer cells. These transporters play an important role in preventing the uptake of toxic compounds, including many drugs and food components, from the gut into the body, and in protecting vital organs in the body, such as the brain, the cerebrospinal fluid, testis, and the fetus against various xenobiotics. Indeed, knock-out of the murine ABC

14

transporter genes have shown altered blood–brain barrier function [60], intestinal drug absorption [61, 62], fetal drug exposure [63] and drug-induced damage to testicular tubules [64]. It is known that many drugs are detoxified by conjugation to glutathione, glucuronate, or sulfate, prior to their ATP-dependent extrusion. The acidic charged conjugates cannot diffuse through cell membranes. The various members of the MRP family mediate the export of these conjugates. Two members of the family, MRP4 (ABCC4) and MRP5 (ABCC5), can transport cyclic nucleotides and nucleotide analogs; these transporters might contribute to resistance against base and nucleoside analogs used in the chemotherapy of cancer and viral diseases. Several genetic variations of some ABC transporters have been recently identified [65]. These genetic variations may potentially modulate the drug resistance phenotype in cancer patients and thereby affect their predisposition to toxicity and response to drug treatment. ABC transporters excrete endogenous metabolites from mammalian secretory epithelia, sometimes against a steep concentration gradient [34]. In the liver these compounds include bile salts (transported by BSEP, the bile salt export pump, ABCB11), phosphatidylcholine (MDR3 P-glycoprotein, ABCB4), bilirubin glucuronides (MRP2, ABCC2), and drugs (MDR1 P-glycoprotein, ABCB1)[34]. Other important functions of ABC transporters include hydrophobic peptide transport. The drug-transporting P-glycoproteins (ABCB1) are excellent transporters of hydrophobic peptides, such as gramicidin D or cyclosporin A [66, 67]. Heterodimeric ABC transporters, the transporter associated with antigen processing (TAP) transports peptides for antigen presentation [58], and an ABC transporter related to TAP have been found to export peptides from mitochondria [68]. The transcription of many mammalian ABC transporters is under tight regulation of nuclear receptors. This is especially the case for the lipid transporters but also for MDR1 and MRP2[34].

15

BCRP (ABCG2), a novel mediator of anticancer drug resistance: The discovery of Pgp and MRP1 is a result of an intensive study of cell lines selected for MDR. Although the increased level of either transporter could account for the drug resistance of most of these cell lines, the resistance of a few cell lines remained unexplained. These lines were characterized by high mitoxantrone resistance and lower resistance to anthracyclines and camptothecins [34]. The resistance to mitoxantrone was a result of decreased drug accumulation, which suggested the presence of a new drug efflux transporter. This transporter was finally identified by Doyle et al. [69] as the breast cancer resistance protein (BCRP), a half-size ABC transporter overproduced in MCF-7 breast cancer cells, and by Allikmets et al. [70] as ABCP, a transporter present at high concentration in placenta. BCRP/ABCP/MXR belongs to the ABCG subfamily and has been renamed ABCG2. Presumably, BCRP functions as a homodimer [71]; there are no indications for other intracellular partners. The range of drugs to which BCRP can confer resistance is less broad than found for Pgp. In addition to mitoxantrone, topotecan derivatives, and anthracyclines, BCRP drug substrates also include bisantrene, etoposide, prazosin, and flavopiridol [72-76]. Other typical Pgp substrates, such as Vinca alkaloids and taxanes, are not included in the BCRP resistance spectrum. Like Pgp, BCRP does not require glutathione (GSH) for the efflux of electroneutral amphipathic drugs [77]. It now appears that Pgp, MRP1, and BCRP can explain MDR in all cell lines analyzed thus far. As has been mentioned before, the resistance to MTX is mediated by the wild type BCRP containing an arginine at amino acid residue 482 (R482) [38, 41].

16

The BCRP gene: The BCRP gene comprises 16 exons and 15 introns spanning 66 kb in length and is located on chromosome 4q22 [78]. Initial characterization of the BCRP promoter revealed that it is TATA-less with 5 putative Sp1 sites downstream from a putative CpG island and contains several AP1 sites [78]. Recently, a functional estrogen response element (ERE) was identified in the BCRP promoter [79]. BCRP confers resistance to anthracycline anticancer drugs and is amplified or involved in chromosomal translocations in cell lines selected with topotecan, mitoxantrone, or doxorubicin treatment [80].

The MRPs: Several toxic compounds that enter the body are modified by oxidation (phase I metabolism) and/or made more water-soluble by conjugation to glutathione (GSH), sulfate, or glucuronate (phase II metabolism). The resulting conjugates are too hydrophilic to diffuse out of the cell and require dedicated transporters to mediate their efflux, as first pointed out by Ishikawa [81]. MRP1 was the first GS-X pump to be identified in cells selected for MDR [46]. Vesicular transport experiments established that MRP1 is in fact a GS-X pump. In 1996 the gene cMOAT (currently known as MRP2) was cloned [82, 83]. MRP3–5 soon followed [84] when Allikmets et al. [85] identified 21 potential human ABC transporters. Recent studies have added four more members to this MRP family: MRP6, MRP7 [86], MRP8 and MRP9 [87]. This probably completes the family, as there are no other putative MRP genes among the 49 human ABC transporter genes. The MRPs studied thus far, MRP1–5, are all organic anion pumps, but they differ in substrate specificity, tissue distribution, and intracellular location. MRPs come in two structural types (Figure 6), one with 17

17

transmembrane segments (MRP1, 2, 3, 6), and one with 12 (MRP4, 5, 7, 8). Structural studies on MRPs have just begun [88].

MRP1: MRP1 is a prototype GS-X pump that transports a variety of drugs conjugated to GSH, to sulfate or to glucuronate, as well as anionic drugs and dyes, but also neutral/basic amphipathic drugs and even oxyanions. However, this enormous range of substrates transported, MRP1 is not indiscriminate. Whereas estradiol-17 βglucuronide is a good substrate, the 3-isomer is not [89] .

The role of the ABC transporters in folate homeostasis: Cellular folate pools are controlled by the previously mentioned folate influx systems, by FPGS activity as well as by ATP-dependent efflux transporters of the ABCC subfamily [34] . Recent studies have established an increased energy-dependent folate and MTX transport into inverted membrane vesicles isolated from cell lines with MRP1 (ABCC1) through MRP5 (ABCC5) overexpression [26, 33, 35, 36, 40] [90]. Recently we have shown [47] that human leukemia cells adapted to grow under extremely low concentrations of leucovorin (5-formyl-tetrahydrofolate) had a 95% loss of MRP1 expression and folate efflux function along with a 100-fold overexpression of the RFC, the primary folate influx transporter [47]. Consistently, replenishment of the latter cells with 5 nM leucovorin resulted in a complete restoration of MRP1 expression. In contrast to the restricted ability of MRP1 through MRP4 to export only monoglutamate forms of folates and MTX, BCRP/ABCG2 has been recently found to transport both mono-, di-, and triglutamate conjugates of folic

18

acid and MTX in membrane vesicles isolated from tumor cell lines with BCRP overexpression [42, 43].

19

Research Aims 1) Examination of our hypothesis that under conditions of folate deprivation, the folate efflux activity of RFC may result in intracellular folate depletion and consequently decreased cellular proliferation. Specific aims: A. To determine whether the cellular proliferation rate of stable transfectants overexpressing the hRFC (i.e. C5/RFC cells) is decreased when compared to their parental RFC null C5 cells under folate free growth conditions B. According to our hypothesis we will determine whether a marked contraction in the cellular folate pool and depletion in folate polyglutamates is enhanced in C5/RFC cells relative to C5 cells under folate free growth conditions. 2) To determine whether down-regulation of RFC and GGH may serve as a novel mechanism of adaptation to short-term (3-7 days) folate-free growth conditions. Specifically, we will explore the gene expression status and activity of RFC as well as GGH, an enzyme catalyzing the conversion of folate polyglutamates to folate monoglutamates that serve as RFC substrates, in several cell lines under folate free growth conditions. 3) A major aim of the present research is to explore the possible role of BCRP in the modulation of cellular folate homeostasis. BCRP may play an important role in controlling cellular folate pools due to its ability to export both mono-, di-, and triglutamate conjugates of folic acid. Specific aims: A- In the current research we will use the two cell lines: MCF-7 breast cancer cells with low BCRP levels and moderate MRP1 (MRP1/ABCC1) levels, and their

20

mitoxantrone (MR)-resistant MCF-7/MR subline with BCRP overexpression and low MRP1 levels. We will develop low folate (LF) adapted sublines from the cell lines MCF-7 and MCF-7/MR. These sublines will be developed by gradual derivation of folic acid from 2.3 μM to 3 nM; resulting in the sublines MCF-7/ LF and MCF-7/MRLF, respectively. B- To determine, by Western blot analysis, the expression of BCRP as well as MRP1 through MRP5 in parental cells and their LF-adapted sublines. C- To examine, by semi-quantitative RT-PCR analysis, the status of BCRP gene expression, in the LF-adapted sublines. D- To determine, by immunohistochemistry studies, both the expression and localization of BCRP in the parental cells and their LF-adapted sublines. E- To determine, by an online Hoechst 33342 efflux assay [91] (a specific chromophoric substrate of BCRP), the BCRP export activity, in the LF-adapted sublines relative to their parental cell lines. F- To examine the Mitoxantrone (a specific substrate of BCRP) accumulation ability and BCRP efflux activity of the LF-adapted sublines. G- To examine the sensitivity of the LF-adapted sublines to Mitoxantrone and MTX, relative to their parental cell lines. The sensitivity to these drugs depends upon the activity of BCRP; therefore, this is a functional assay of BCRP.

The depicted specific aims (A-G) should enable us to examine the expression and efflux function of BCRP in the LF-adapted sublines. The examination of the relationship between folate status and BCRP expression and function may reveal for the first time, the possible role of BCRP in folate homeostasis.

21

4) To explore whether cellular misconfinement of BCRP may serve as a possible mechanism of adaptation that limits folate efflux under low folate conditions: Our preliminary results revealed for the first time that BCRP may have an important role in folate homeostasis [92]. Hence, we will examine whether alternative mechanisms of loss of BCRP function exist (other than loss of BCRP expression), including lack of plasma membrane targeting of BCRP under low folate conditions. The first possible mechanism to be examined is retention of BCRP in the cytosolic compartment and lack of plasma membrane targeting of BCRP after a relatively short -term (3 weeks) deprivation of folic acid. Specific aims: A- To examine, by immunohistochemistry studies, the localization of BCRP in cells after a short -term folic acid deprivation. B- If BCRP is retained in the cytosolic compartment, then, we will determine its exact sub-cellular localization by confocal microscopy. C- To determine, by an online efflux assay [91] of Hoechst 33342 ( a specific substrate of BCRP), the BCRP export activity, in the BCRP-mislocalized cells . D- To examine, by Mitoxantrone (a specific substrate of BCRP) accumulation assay, the BCRP activity, in the BCRP-mislocalized cells. 5) Exploration of the relationship between the expression of BCRP and the activity of FPGS as two tightly-related components of folate homeostasis: Our working hypothesis is that under low folate conditions two reciprocal processes may occur; loss of BCRP expression and/or efflux function, along with increased FPGS activity. BCRP is relatively poorly expressed and its pattern of tissue expression is apparently restricted to only a few tissues including placenta, intestine,

22

colon and the bile canaliculus. Interestingly, all these tissues were reported to express high levels of FPGS mRNA [93] and displayed a relatively high activity of FPGS. It is possible that the high activity of FPGS ensures sufficient intracellular retention of long-chain (> 3 glutamate residues) folate polyglutamates. These long polyglutamylated folates are no longer substrates for BCRP that fails to export folate containing more than three glutamate residues. Hence, we will determine if the FPGS activity of the various low- folate adapted sublines is increased relative to their parental cells. 6) Previously we have shown that folate status strongly affects MRP1 expression [47]: low folate conditions result in a dramatic loss of MRP1 expression, whereas, replenishment of µM concentrations of folic acid leads to restoration of MRP1 expression. Hence, in the present research we will determine if MRP1 mRNA levels are decreased in the various low-folate adapted sublines, relative to their parental cell lines. 7) Overexpression of the multidrug efflux transporter BCRP in the plasma membrane of cancer cells confers resistance to various anticancer drugs including mitoxantrone. Our research aim is to reveal the mechanism underlying drug resistance in the MCF-7 breast cancer sublines MCF-7/MR and MCF-7/FLV1000 cells in which wild type (R482) BCRP overexpression is highly confined to cell-cell attachment zones. Specifically, we will test our hypothesis that the cell-cell attachment zones are a part of a novel extracellular vesicles in which mitoxantrone is sequestered into. The structural characteristics of the extracellular vesicles will be examined by confocal and electron microscopy. Moreover, we will examine if this novel sequestration activity is mediated via BCRP in a specific manner.

23

Description of the research In the current research we used several cell lines including breast cancer, leukemia and ovary cells in order to understand the role of plasma membrane transporters in folate homeostasis and drug resistance. These cells were deprived from folates upon short (days), median (weeks) and long (months) term schedule. The molecular mechanisms that enabled the cells to survive the folate deprivation were then identified. The role of plasma membrane transporters in folate homeostasis was specifically investigated and novel models of folate homeostasis involving the plasma membrane transporters RFC and BCRP have been introduced in this research. Moreover, the molecular and cellular mechanisms by which the transporter BCRP mediates resistance against the anticancer drug mitoxantrone in breast cancer cells was revealed and led to the discovery and characterization of a novel extracellular vesicle.

24

Materials and Methods A) The reduced folate carrier mediates intracellular folate depletion and consequent cytotoxicity under folate deprivation: RFC-dependent cellular proliferation rate in the presence or absence of folates: For C5 cellsThe Chinese hamster ovary cell line deficient in RFC activity named C5 was grown as monolayers in RPMI-1640 medium containing 2.3 μM folic acid (Biological Industries, Beth-Haemek, Israel), 10% fetal calf serum (GIBCO) supplemented with 2 mM glutamine and 100 μg/ml penicillin and streptomycin. Following trypsinization of the C5 monolayer cell line, the cells were washed three times with folic acid free growth medium containing 10% dialyzed FCS and antibiotics. Then, cells (6 x 104) were seeded in each of two T25 flasks in

5 ml folic acid free growth

medium(GIBCO) containing 10% dialyzed FCS (GIBCO) and antibiotics; the subline in the first flask was termed C5-NF (i.e. no folate) and was incubated for 6 days in humidified CO2 incubator without medium refreshment (the flasks' cover must remain loosely open during the incubation period). The second T25 flask was supplemented with 2.3 µM folic acid and was therefore termed C5-HF (i.e. high folate);this flask was incubated for 6 days in humidified CO2 incubator without medium refreshment as well (the flasks' cover must remain loosely open during the incubation period). Following these 6 days of incubation, the cells were detached by trypsinization and the number of viable cells was determined by a haemocytometer counting after trypan blue staining.

25

For C5/RFC-3nMLCV cellsG418 (600µg/ml) and LCV (3 nM) were added to the growth medium (folic acid free growth medium containing 10% dialyzed FCS and antibiotics) of RFC transfected C5 cells (i.e. C5/RFC cells). C5/RFC cells were grown under these conditions for two weeks. Following trypsinization of the C5/RFC-3nMLCV monolayer cell line, the cells were washed three times with folic acid free growth medium containing 10% dialyzed FCS and antibiotics. Then, cells (6x104) were seeded in each of two T25 flasks in 5 ml folic acid free growth medium containing 10% dialyzed FCS and antibiotics; the subline in the first flask was termed C5/RFC-NF (i.e. no folate) and was incubated for 6 days in humidified CO2 incubator without medium refreshment (the flasks' cover must remain loosely open during the incubation period). The second T25 flask was supplemented with 3nM leucovorin and was therefore termed C5/RFC-3nMLCV; this flask was incubated for 6 days in humidified CO2 incubator without medium refreshment as well (the flasks' cover must remain loosely open during the incubation period). Following these 6 days of incubation, the cells were detached by trypsinization and the number of viable cells was determined by a haemocytometer counting after trypan blue staining. For C5/ FR-3nMFA cellsG418 (600µg/ml) and folic acid (3 nM) were added to the growth medium (folic acid free growth medium containing 10% dialyzed FCS and antibiotics) of folate receptor transfected C5 cells (i.e. C5/FR cells). C5/ FR-3nMFA cells were grown under these conditions for two weeks. Following trypsinization of the C5/FR-3nMLCV monolayer cell line, the cells were washed three times with folic acid free growth medium containing 10% dialyzed FCS and antibiotics. Then, cells ( 6 x 104) were

26

seeded in each of two T25 flasks in 5 ml folic acid free growth medium containing 10% dialyzed FCS and antibiotics; the subline in the first flask was termed C5/FR-NF and was incubated for 6 days in humidified CO2 incubator without medium refreshment (the flasks' cover must remain loosely open during the incubation period). The second T25 flask was supplemented with 3nM folic acid and was therefore termed C5/FR-3nM FA; This flask was incubated for 6 days in humidified CO2 incubator without medium refreshment as well (the flasks' cover must remain loosely open during the incubation period). Following these 6 days of incubation, the cells were detached by trypsinization and the number of viable cells was determined by a haemocytometer counting after trypan blue staining. Remark- the experiments for C5-HF, C5-NF, C5/RFC-3nMLCV, C5/RFC-NF, C5/FR-3nM FA and C5/FR-NF were done simultaneously. Folate deprivation: These are the folate deprivation protocols that we developed for each cell line: For MCF-7/MR cellsMitoxantrone was added to the growth medium of MCF-7/MR cells at a final concentration of 100 nM. This drug concentration was maintained for three days, after which monolayer cells were washed twice with PBS; then, a fresh drug free growth medium was added to the cells. The MCF-7/MR cells were grown for 4 days without the drug. Following trypsinization of the MCF7/MR monolayer cell line, the cells were washed three times with folic acid free growth medium containing 10% dialyzed FCS and antibiotics. Then, cells ( 2.3 x 105) were seeded in each of two T75 flasks in 15 ml folic acid free growth medium containing 10% dialyzed FCS and antibiotics; the subline in the first flask was termed MCF7/MR-NF (i.e. no folate) and was incubated for 7 days in humidified CO2 incubator without medium refreshment (the

27

flasks' cover must remain loosely open during the incubation period) . The second T75 flask was supplemented with 2.3 µM folic acid and was therefore termed MCF7/MR-HF (i.e. high folate); This flask was incubated for 7 days in humidified CO2 incubator without medium refreshment as well (the flasks' cover must remain loosely open during the incubation period). Following these 7 days of incubation, the cells are ready for the various analyses (i.e. Quantitative RT-PCR and [3H] MTX initial rate uptake experiments). For CEM/7A cellsThe CEM/7A cells were grown in growth medium containing 0.2 nM leucovorin for one week or more. Then, the CEM/7A cells were washed three times with folic acid free growth medium containing 10% dialyzed FCS and antibiotics. Then, cells (107) are seeded in each of two T75 flasks in 50 ml folic acid free growth medium containing 10% dialyzed FCS and antibiotics; the subline in the first flask was termed CEM/7A-NF (i.e. no folate) and was incubated for 3 days in humidified CO2 incubator without medium refreshment (the flasks' cover must remain loosely open during the incubation period). The second T75 flask was supplemented with 0.2 nM leucovorin and was therefore termed CEM/7A-HF; This flask was incubated for 3 days in humidified CO2 incubator without medium refreshment as well (the flasks' cover must remain loosely open during the incubation period). Following these 3 days of incubation, the cells are ready for the various analyses (i.e. Quantitative RTPCR and [3H] MTX initial rate uptake experiments).

RNA extraction and quantitative RT-PCR: The folate deprivation protocols resulted in the following sublines: MCF7/MR-HF, MCF7/MR-NF, CEM/7A-HF and CEM/7A-NF. These sublines were briefly

28

trypsinized and total RNA was extracted using the TriReagent protocol (Sigma). RNA quantity was determined using an ND-1000 Nano DropTM UV spectrophotometer (Nano Drop Tech, Inc.,). cDNA synthesis was carried out using 12 μg of RNA in a 50 μl reaction mixture containing random hexamer primers and MLV reverse transcriptase at 37oC (Promega). The quality and quantity of the resultant cDNA were estimated by 1% agarose gel electrophoresis. In order to evaluate the level of Actin,GAPDH, BCRP,MRP5,MRP1,GGH, FPGS and RFC gene expression, a semiquantitative RT-PCR method was used.. PCR was carried out in a total volume of 30 μl in the presence of 100 ng of cDNA, 0.4 μM of each of the sense and antisense primers and a 1X RedTaqTM ReadyMixTM PCR reaction mix solution (Sigma).The primers that were used for Actin, GAPDH, MRP1,GGH and FPGS were previously described [94]. The primers that were used for BCRP were 5’TGCCCAAGGACTCAATGCAACA-3’ and a downstream primer 5’ACAATTTCAGGTAGGCAATTGTG-3’ as previously described [95] The GGH primers were 5’-GCTTATTAACTGCCACAGATACGTTG-3’ and 5’-GAACATTCTGCTGTGCAATGAC-3’ [96].The MRP5 primers were 5’CCCAGGCAACAGAGTCTAACC -3’ and 5’- CGGTAATTCAATGCCCAAGTC3’. The RFC primers were 5’-CCCCAGGGCTAGTGAAGTGG -3’ and 5’- CACATCAGATGGTCCCGCAC-3’. Following an initial denaturation at 95°C for 10 min, 25 to 35 cycles each of 1 min denaturation at 95°C, 1 min annealing at 53– 62°C and 1 min elongation at 72°C, as well as a final extension period of 10 min at 72°C, were carried out. PCR products were analyzed on 1% agarose gel electrophoresis in order to verify the size and specificity of the products. Representative results of three independent experiments are shown.

29

[3H] MTX initial rate uptake experiments: The folate deprivation protocols resulted in the following sublines: MCF7/MR-HF, MCF7/MR-NF, CEM/7A-HF and CEM/7A-NF. The [3H] MTX initial rate uptake activity of these sublines was measured as has been previously described [97]. Statistical Analysis: We used paired student’s T-test to examine the significance of the difference between two populations for a certain variable. A difference between the averages of two populations was considered significant if the P-value obtained was < 0.05. B) Folate deprivation results in the loss of BCRP expression: A role for BCRP in cellular folate homeostasis: Materials: The following materials were purchased from Sigma Chemical Co: Triton X-100, Freund Adjuvant (FA) and complete FA, Tween 20. MTX was purchased from Teva Pharmaceuticals Ltd. Mitoxantrone hydrochloride was from Cyanamid of Great Britain Ltd (Gosport, Hampshire, England). Ko143 was generously provided by Dr. A.H. Schinkel, The Netherlands Cancer Institute, Amsterdam, The Netherlands. Hoechst 33342 was purchased from Molecular Probes (Eugene, OR). The completeTM mini-mixture of protease inhibitors was from Roche. Nitrocellulose nylon membrane was purchased from Schleicher & Schuell. Horseradish peroxidase (HRP) conjugated goat anti-mouse IgG was purchased from Jackson Immunoresearch Labs, West Grove, PA. RPMI-1640 medium was purchased from Biological Industries, Beth-Haemek, Israel. Fetal calf serum was from GIBCO. MRPr1 monoclonal antibody was kindly provided by Prof. R.J. Scheper, Dept. of Pathology, VU Medical Center, Amsterdam, The Netherlands. β-actin monoclonal antibody was purchased from Chemicon, USA.

30

Tissue culture and folic acid deprivation: The human breast cancer cell line, MCF7, and its Mitoxantrone-resistant MCF-7/MR subline (originally termed MCF7/Mitox, see ref. [98] with BCRP overexpression [76] were grown as monolayers in RPMI-1640 medium containing 2.3 μM folic acid (Biological Industries, BethHaemek, Israel), 10% fetal calf serum (GIBCO) supplemented with 2 mM glutamine and 100 μg/ml penicillin and streptomycin. The growth medium of MCF-7/MR cells also contained 0.1 μM Mitoxantrone. In order to establish cell lines growing under LF conditions, MCF-7 and MCF-7/MR cells were gradually deprived of folic acid from 2.3 μM (the standard concentration in RPMI-1640 medium) to 3 nM resulting in the sublines MCF-7/LF and MCF-7/MR-LF; this was achieved over a period of three and a half months in a folic acid-free RPMI-1640 medium (Biological Industries, BethHaemek, Israel) supplemented with 10% dialyzed fetal calf serum (GIBCO) to which gradually decreasing folic acid concentrations were added. In order to examine the stability of BCRP expression during the omission of Mitoxantrone from the growth medium, MCF-7/MR-HF cells were continuously cultured in Mitoxantrone-free medium containing 2.3 μM folic acid. The human ovarian carcinoma cell line, 2008, and its sublines stably transduced with MRP1, MRP2 and MRP3 cDNAs were kindly provided by Prof. P. Borst (The Netherlands Cancer Institute, Amsterdam, The Netherlands), whereas HEK293/MRP4 and HEK293/MRP5 cells were used as positive controls for MRP4 and MRP5 overexpression, respectively. These cell lines were cultured in RPMI-1640 medium containing 2.3 μM folic acid, 10% fetal calf serum, 2 mM glutamine and antibiotics. Growth inhibition with Mitoxantrone and MTX: For Mitoxantrone growth inhibition, cells (1x104-1.75x104/well) were seeded in 24-well plates in growth medium containing various concentrations of Mitoxantrone for 3-5 days at 37oC. For

31

MTX growth inhibition, cells were allowed to attach for 24 hr at 37oC. Attached cells were then exposed to various concentrations of MTX for 4 hr at 37oC, following which the drug-containing medium was aspirated and three successive washes each of 10 min in RPMI-1640 containing 10% dialyzed fetal bovine serum at 37 oC were performed. Drug-free medium was added (2ml/well) and cultures were incubated for 4 days at 37oC. After incubation with Mitoxantrone or MTX, cells were detached by trypsinization and the number of viable cells was determined using trypan blue exclusion. Extraction of membrane proteins from cultured cells: Exponentially growing cells (at ~106cells/ml) were harvested by centrifugation and washed with PBS. The cells (1-3x107 cells ) were then incubated in a lysis buffer containing 50 mM Tris-HCL pH 7.5, 50 mM 2-mercaptoethanol, 0.5 Triton X-100, 10 µg/ml PMSF, 60 µg/ml aprotinin, 5µg/ml leupeptin, 10µg/ml pepstatin, 1mM EGTA ph 8, and 1mM EDTA pH 8 (120 µl per 107 cells). After 1h incubation on ice, the protein extract was centrifuged and aliquots of the supernatant were stored at -80° C until analysis. Protein content was determined using the Bio-Rad protein assay according to Bradford. Western blot analysis of MRPs and BCRP expression: To examine the expression of various MRPs and BCRP in the different cell lines, non-ionic detergent-soluble proteins were extracted. Proteins (6-60 µg) were resolved by electrophoresis on 7% (for MRPs) or 10% (for BCRP) polyacrylamide gels containing SDS and electroblotted onto a Protran BA83 cellulose nitrate membrane (Schleicher & Schuell, Dassel,Germany). The blots were blocked for 1 hr at room temperature in TBS buffer (150 mM NaCl, 0.5% Tween 20, 10 mM Tris, at pH 8.0) containing 1% skim milk. The blots were then reacted with the following anti-human

32

MRP monoclonal antibodies (kindly provided by Prof. R.J.Scheper, VU Medical Center, Amsterdam, The Netherlands): rat anti-MRP1 (MRP-r1; 1:1,000),-MRP4 (M4I-10; 1:500), -MRP5 (M5I-1; 1:750), and -BCRP (BXP-53; 1:1000) as well as mouse anti-MRP2 (M2III-5; 1:50) and -MRP3 (M3II-21; 1:500). Blots were then rinsed in the same buffer for 10 min at room temperature and reacted with horseradish peroxidase (HRP)-conjugated goat anti-mouse or anti-rat IgG (1: 10,000 dilution, Jackson Immunoresearch Labs,West Grove, PA) for 1 hr at room temperature. Following three10-min washes in TBS at room temperature, enhanced chemiluminescence detection was performed according to the manufacturer’s instructions (Biological Industries, Beth-Haemek, Israel). To normalize for loading differences, the nylon membranes were stripped and reacted with an antibody against β-tubulin (clone 2-28-33 from Sigma; 1: 4,000). Immunohistochemistry studies: Mid-logarithmic monolayers in 24-well plates were washed twice with PBS and fixed with 4% formaldehyde for 10 min. Endogenous peroxidase activity was neutralized by incubation for 20 min in a solution consisting of 4 volumes of methanol and 1 volume of 3% H2O2 in double distilled water. The fixed cells were washed twice with PBS, blocked for 1 hr at room temperature in PBS containing 1% skim milk and reacted with an anti-BCRP monoclonal antibody BXP53 (1:100 dilution). Then, an HRP-conjugated goat anti-rat IgG (1:100 dilution) was added. Color development was performed with the chromogen 3,3’-diaminobenzidine (0.6 mg/ml) in a solution containing 0.02% H2O2. After counterstaining with haematoxylin, cells were examined with an Olympus BH-2 upright light microscope. Online efflux of Hoechst 33342: Efflux of Hoechst 33342 was measured using an online computerized method. Cells were cultured on glass coverslips that fitted to the wall of a 1 x 1 x 4 cm cuvette (width, depth, and height, respectively). Cells were

33

loaded with 10 μM Hoechst 33342 for 2 hr at 37ºC in a phenol red-free RPMI-1640 medium (Gibco) until steady-state was achieved. Cells were then transferred to ice until the efflux was initiated. To follow Hoechst 33342 efflux, the coverslips were washed twice with ice-cold medium and then placed in a cuvette that contained 3 ml of warm (37ºC) medium. The fluorescence in the extracellular medium originating from the extruded chromophore was monitored online with a spectrofluorometer (FluoroMax, SPEX Industries, Edison, NJ). Hoechst 33342 fluorescence was measured every second at an ultraviolet excitation of 318 nm and emission at 460 nm. To correct for fluctuations in the number of cells/slide, cell numbers were determined by adding 0.4 μM of the DNA dye Syto 13 (Molecular Probes, Eugene, OR). The fluorescence of Syto 13 was determined at excitation and emission wavelengths of 485 nm and 520 nm, respectively. Hoechst 33342 fluorescence was then normalized for the individual Syto 13 signals. [3H]Folic acid accumulation: Adherent cells (~1x106) in 6 cm petri dishes (Nunc) were washed three times in folic acid-free RPMI-1640 containing 10% dialyzed fetal calf serum (Gibco). Following an equilibration for 2 hr in the same medium at 37oC, [3H]folic acid (69 Ci/mmol, Moravek Biochemicals, Brea, CA) was added to a final concentration of 1 μM (specific radioactivity 1,200 dpm/pmol) and incubated at 37oC for 4 hr and 24 hr. Transport was stopped by the addition of 10 ml of ice-cold HBS containing: 20 mM Hepes, 140 mM NaCl, 5 mM KCl, 2 mM MgCl2 and 5 mM Dglucose, pH 7.4 with NaOH

[99]. Cells were then washed three times in ice-cold

HBS, detached by trypsinization, counted, centrifuged at 500 x g for 5 min at 4oC and the radioactivity was determined using an Ultima Gold scintillation fluid and a scintillation spectrometer (Packard). Controls for accumulation of radiolabeled folic acid contained a 1,000-fold excess of unlabeled folic acid (1 mM).

34

Flow cytometric analysis of Mitoxantrone staining: Exponentially growing MCF-7, MCF-7/MR and their LF-adapted cell sublines were trypsinized, adjusted to a density of 5x105-1x106/ml and incubated in growth medium containing 20 μM Mitoxantrone for 1 hr at 37oC. Cells were then harvested by centrifugation at 4oC, washed once with ice-cold PBS and analyzed for mean fluorescence intensity per cell by a FACSCalibur flow cytometer (Becton Dickenson). Excitation was at 633 nm and emission at 661 nm. Autofluorescence intensities of unstained cells were recorded and subtracted from those of Mitoxantrone-stained cells. FPGS activity assay: Frozen pellets of 2x107 cells grown in the absence of drugs for 3-4 passages were suspended in 0.5 ml of an extraction buffer containing: 50 mM Tris-HCl, 20 mM KCl, 10 mM MgCl2 and 5 mM dithiotreitol, at pH 7.5. Total cell extracts were obtained by sonication (MSE Soniprep, amplitude 14 micron, 3 x 5 sec with 10 sec intervals, at 4oC) followed by centrifugation at 12, 000 x g for 15 min at 4oC. The FPGS activity assay mixture contained: 200 μg protein, 4 mM [2,3-3H]-Lglutamic acid (specific activity 6.6 mCi/mmol) and 250 μM MTX in a buffer consisting of 100 mM Tris pH 8.5, 10 mM ATP, 20 mM MgCl2, 20 mM KCl, and 10 mM dithiotreitol in a final volume of 250 μl [100]. Following 2 h incubation at 37oC, the reaction was stopped by adding 1 ml of an ice-cold solution containing 5 mM unlabeled L-glutamic acid. Sep-Pak C18 cartridges (Millipore, Waters Associates, Etten-Leur, The Netherlands) were used for the separation of free, unreacted [3H]-Lglutamate from MTX-[3H]-diglutamate. Controls lacking MTX were included in order to correct for polyglutamylation of endogenous folates present in the cell extract. Semi-quantitative RT-PCR and DNA sequencing: Following RNA isolation using the Tri Reagent protocol (Sigma) and cDNA synthesis, a 172 bp-human BCRP fragment was PCR amplified using the upstream primer:

35

5’-TGCCCAAGGACTCAATGCAACA-3’ and a downstream primer 5’ACAATTTCAGGTAGGCAATTGTG-3’ as previously described [95]. The primers for GAPDH and PCR conditions were as recently described [94]. To analyze whether BCRP in parental MCF-7 cells and their MCF-7/MR subline harbored the wild type R482 amino acid, we performed DNA sequencing of this region using an ABI Prism 310 DNA Sequencer (AME Bioscience, USA). To this end, BCRP primers were designed using the LightCycler Probe Design Software version 1.0 (Idaho Technology Inc., USA); the upstream and downstream primers were as follows: 5’-CAGCGGATACTACAGAG-3’ and 5’GCCGTAAATCCATATCGTG-3’, respectively. All cell lines were homozygous for the wild type R482 amino acid. Scanning densitometry: Relative BCRP and MRP1 protein/mRNA levels were determined by scanning densitometry of several linear exposures using the program "TINA" (version 2.10g) divided by the densitometrical value of β-actin/GAPDH respectively. Statistical Analysis: We used a student T-test to examine the significance of the difference between two populations for a certain variable. A difference between the averages of two populations was considered significant if the P-value obtained was < 0.05.

C) Cytoplasmic Confinement of Breast Cancer Resistance Protein (BCRP/ABCG2) as a Novel Mechanism of Adaptation to Short-Term Folate Deprivation: Chemicals: Folic acid, tunicamycin, Triton X-100, Tween 20, rhodamine 123, 3,3'diaminobenzidine tetrahydrochloride, propidium iodide, 6-diamidino-2-phenylindole 36

(DAPI), emetine hydrochloride, and hematoxylin were obtained from Sigma-Aldrich (St. Louis, MO). Mitoxantrone hydrochloride was from Cyanamid of Great Britain Ltd. (Gosport, Hampshire, England). Tissue Culture: The Mitoxantrone-resistant human breast cancer cell line MCF7/MR originally termed MCF-7/Mitox [98] with BCRP overexpression [92]was grown under monolayer conditions in RPMI-1640 medium containing 2.3 µM folic acid (Biological Industries, Beth-Haemek, Israel), 10% fetal calf serum (Invitrogen, Carlsbad, CA), 2 mM glutamine, 100 µg/ml penicillin, and 100 µg/ml streptomycin. Once every 2 weeks, MCF-7/MR cells were cultured in the presence of 0.1 µM Mitoxantrone. The human ovarian carcinoma 2008/MRP1 subline, stably transduced with an MRP1 cDNA (kindly provided by Prof. P. Borst, The Netherlands Cancer Institute, Amsterdam, The Netherlands) was cultured in the above-mentioned RPMI 1640 medium. The EmtR1 Chinese hamster ovary cell line was derived in our laboratory by stepwise selection in increasing concentrations of emetine, resulting in stable Pgp overexpression [67]. This cell line was routinely grown in RPMI 1640 medium supplemented with 1 µM emetine. Western Blot Analysis of BCRP, MRP1, and Pgp Expression: To examine the expression of BCRP, MRP1, and Pgp in the various cell lines, nonionic detergentsoluble membrane proteins were extracted as described previously [92]. Proteins (1020 µg) were resolved by electrophoresis on 10% (for BCRP) or 7% (for MRP1 or Pgp) polyacrylamide gels containing SDS and electroblotted onto Protran BA83 cellulose nitrate membranes (Schleicher & Schuell, Dassel, Germany). The blots were blocked for 1 h at room temperature in TBS buffer (10 mM Tris, pH 8.0, 150 mM NaCl, and 0.5% Tween 20) containing 1% skim milk. The blots were then reacted

37

with the following anti-human BCRP, MRP1, and Pgp monoclonal antibodies (generously provided by Prof. R. J. Scheper and Dr. G. L. Scheffer VU University Medical Center, Amsterdam, The Netherlands): BXP-53, a rat anti-BCRP antibody (at a dilution of 1:1000); MRP-r1, a rat anti-MRP1 antibody (1:1000); and JSB-1, a mouse anti-Pgp antibody (1:100). Blots were then rinsed in the same buffer for 10 min at room temperature and reacted with second antibodies consisting either of horseradish peroxidase (HRP)-conjugated goat anti-mouse or anti-rat IgG (1: 10,000 dilution; Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 h at room temperature. After three washes (each of 10 min) in TBS at room temperature, enhanced chemiluminescence detection was performed according to the manufacturer's instructions (Biological Industries). To normalize for loading differences, the blots were first stripped using the following procedure. Blots were incubated for 10 min in a stripping solution containing 0.5 M NaCl, 0.5 M acetic acid at pH 2.4. The nylon membranes were then washed twice with TBS and reacted with an antibody against β-tubulin, clone 2-28-33 from Sigma-Aldrich (1: 4000). Immunohistochemistry Studies: Cells (5x104) from each cell line were seeded in 25mm tissue culture flasks and incubated for 4 days in 5 ml of growth medium. Monolayer cells were then washed twice with PBS and fixed for 10 min in a solution of 4% formaldehyde in PBS. Endogenous peroxidase activity was neutralized by incubation for 20 min in a solution consisting of 4 volumes of methanol and 1 volume of 3% H2O2 in double distilled water. The fixed cells were washed twice with PBS, blocked for 1 h at room temperature in PBS containing 1% skim milk, and reacted with the following antibodies: rat anti-human BCRP monoclonal antibody BXP-53 (1:100), mouse anti-human epidermal growth factor receptor (EGFR) monoclonal antibody 111.6 (1:100; generously provided by Prof. Y. Yarden, The Weizmann

38

Institute of Science, Rehovot, Israel), rabbit polyclonal antibodies to human fibroblast growth factor receptor (FGFR)1 Flg (H-76), FGFR2 (Bek C-17), and FGFR3 (H-100; all at a 1:100 dilution; Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Then, HRPconjugated goat anti-rat, -mouse, or -rabbit IgG (all at 1:100 dilution; Jackson ImmunoResearch Laboratories) were added followed by two washes with PBS. Color development was performed with the chromogen 3,3'-diaminobenzidine (0.6 mg/ml) in a solution containing 0.02% H2O2 at pH 7.6. After counterstaining of nuclei with hematoxylin, cells were examined with an Olympus BH-2 upright light microscope at random monolayer positions avoiding the edges of the flasks. Immunofluorescence Analysis with Viable Cells: Cells (104) from each cell line were seeded onto 24-well plates (1 ml of medium/well) on sterile glass coverslips and incubated for 4 days at 37°C. Then, the growth medium was removed, and monolayer cells were washed twice with PBS and blocked for 1 h at room temperature in PBS containing 10% fetal calf serum (Invitrogen) and reacted with a mouse anti-MHC class I monoclonal antibody W6/32 (1:100; kindly provided by Prof. Yoram Reiter, Technion, Haifa, Israel) in a blocking solution for 45 min at room temperature. The coverslips were then washed twice with PBS and reacted in blocking solution with an FITC-conjugated goat anti-mouse antibody (1:100; Jackson ImmunoResearch Laboratories). After 40 min of incubation at room temperature, the coverslips containing viable cells were washed twice with PBS, mounted onto glass slides, and examined using a Leica immunofluorescence microscope. We also performed a control experiment in which the mouse anti-MHC W6/32 monoclonal antibody was omitted.

39

Confocal and Immunofluorescence Microscopy Studies with Fixed Cells: Cells (104) from each cell line were seeded onto 24-well plates (1 ml of medium/well) on sterile glass coverslips and incubated for 4 days at 37°C. Then, the growth medium was removed, and monolayer cells were washed twice with PBS and fixed with 4% formaldehyde in PBS for 10 min. The coverslips were washed twice with PBS and then incubated for 20 min in a solution of 80% methanol in double distilled water. The coverslips were washed twice with PBS, blocked for 1 h at room temperature in PBS containing 1% skim milk and reacted with a rat anti-BCRP monoclonal antibody BXP-53 (1:100) for 60 min at room temperature. The coverslips were washed twice with PBS and reacted with a goat anti-rat IgG (1:100) and mouse anti-calnexin antibody (1:100; BD Transduction Laboratories, Lexington, KY) for 60 min at room temperature. After washing twice with PBS, the cells were incubated with the secondary FITC-conjugated rabbit-anti-goat IgG (1:100; Sigma-Aldrich) and Cy3conjugated donkey-anti-mouse IgG (1:100; Jackson ImmunoResearch Laboratories). Cell nuclei were stained with the DNA dye DAPI (Sigma-Aldrich) at a final concentration of 0.5 µg/ml for 60 min at room temperature. After four washes with PBS (each with 2 ml), the coverslips were mounted onto glass slides using Fluoromount-G (Southern Biotechnology Associates, Birmingham, AL). The slides were then examined using a Leica immunofluorescence microscope and a Bio-Rad MRC1024 confocal microscope. Propidium Iodide (PI) Staining and Cell Cycle Analysis: Monolayer cells were detached by trypsinization, adjusted to a density of 106 cells/ml in PBS, fixed with 70% ethanol, and stained with propidium iodide as described previously [101] . PIstained cells were then analyzed by flow cytometry using a 488-nm laser excitation

40

and emission was collected at 585 nm. The percentages of cells at apoptosis or with a >4n DNA content were calculated using a WinMDI software (version 2.8). Assay of Cellular Rhodamine 123 Accumulation: Cells (5x104) from each cell line were seeded onto 25-mm tissue culture flasks for 4 days; the medium was replaced and allowed to equilibrate by 4 h incubation at 37°C in an atmosphere of 5% CO2. Then, rhodamine 123 was added to the growth medium at a final concentration of 750 nM. After 60 min of incubation at 37°C, monolayer cells in the tissue culture flasks were washed seven times with cold (4°C) PBS (each time with 8 ml/flask). To extract cellular rhodamine 123 fluorescence, cells were lysed with a solution of PBS containing 1% Triton X-100 (1.6 ml/flask). Total cellular fluorescence was determined in quartz cuvettes using a fluorescence spectrophotometer (Cary Eclipse; Varian, Inc., Palo Alto, CA). The fluorescence readings were normalized to the relative number of cells present in each culture flask that were determined with duplicate flasks before rhodamine 123 accumulation. The entire rhodamine 123 accumulation study was carried out in the dark. Quantitative Analysis of the Cytoplasmic and Plasma Membrane Fractions of BCRP: First, the immunohistochemical images were processed using Adobe Photoshop (ver. 6.0; Adobe Systems, Mountain View, CA), and only the brown diaminobenzidine color corresponding to the BCRP staining was selected automatically and then copied to a new picture with a white background. All the other accompanying colors, including that of the counterstain hematoxylin were excluded. This new image represented total cellular BCRP staining. To obtain a picture representing the cytoplasmic BCRP staining, the original pictures were opened once again using Adobe Photoshop software followed by manual erasure of the plasma

41

membrane staining. Then, the brown BCRP staining was selected automatically and copied to yield a new picture of the cytoplasmic staining. Total staining and cytoplasmic staining was transformed to a gray scale picture and analyzed separately by scanning densitometry using the program TINA (version 2.10g). The local background levels were subtracted from the original densitometric values resulting in two corrected values for each colony: total cellular staining of BCRP and cytoplasmic BCRP staining. The percentage of the cytoplasmic fraction was obtained by dividing the value of cytoplasmic BCRP staining by that of the total cellular BCRP staining, multiplied by 100. [3H]Folic Acid Accumulation: Adherent cells ( 8 x 106) in T75 tissue culture flasks (NUNC A/S, Roskilde, Denmark) were washed with HBS, detached by trypsinization, and suspended at 107 cells/ml in the same buffer as described previously [92]. [3H]Folic acid (69 Ci/mmol; Moravek Biochemicals, Brea, CA) was added to a final concentration of 2 µM (specific radioactivity 2500 dpm/pmol) and incubated at 37°C for 30 min; 1 µM trimetrexate was included to block folic acid reduction [99]. Transport was terminated and cells were processed for scintillation counting as described previously [92] . Statistical Analysis: We used a one-tailed Student's t test to examine the significance of the difference between two populations for a certain variable. A difference between the averages of two populations was considered significant if the P value obtained was <0.05. Scanning Densitometry: Relative BCRP and MRP1 protein levels were determined by scanning densitometry of several linear exposures, using the program TINA, divided by the densitometrical value of β-tubulin.

42

D) Novel extracellular vesicles mediate a BCRP-dependent anticancer drug sequestration and resistance: Chemicals: Mitoxantrone hydrochloride was from Cyanamid of Great Britain Ltd. (Gosport, Hampshire, UK). Ko143 was generously provided by Dr. A.H. Schinkel, The

Netherlands

Cancer

Institute,

Amsterdam,

The

Netherlands,

whereas

fumitremorgin C (FTC) and flavopiridol were kindly provided by Dr. S.E. Bates, National Cancer Institute, Bethesda, MD. Sodium azide and carbonyl cyanide-ptrifluoromethoxyphenylhydrazone (FCCP) were from Sigma. Tissue culture and growth inhibition with mitoxantrone : MCF-7 human breast cancer cells and its mitoxantrone-resistant MCF-7/MR subline (kindly provided by Dr. S.E. Bates, National Cancer Institute, Bethesda, MD) were grown as monolayers in RPMI-1640 medium as previously described [92, 102]. It should be emphasized that in a previous study we found that MCF-7/MR cells overexpressed the wild type R482 BCRP and thus no BCRP mutations were present in this cell line [92]. Specifically, MCF-7/MR cells were pulsed with 100 nM mitoxantrone every two weeks for 3 days, whereas MCF-7/FLV1000 cells were continuously grown in the presence of 1 µM flavopiridol. All subsequent experiments were initiated after 4 days of incubation with mitoxantrone-free or flavopiridol-free medium. Cells (5x103/well) were seeded in 24-well plates in growth medium (2 ml/well) and incubated for 48 hr at 37°C. Then, the medium of MCF-7 and MCF-7/MR cells was replaced with a fresh one lacking or containing Ko143 (0.4 µM). After 2 hr of incubation, mitoxantrone was added at various concentrations. Then, the cells were exposed to this drug for 5 h at 37°C, following which the drug-containing medium was aspirated and three successive washes (each of 10-min) in RPMI-1640 containing 10% dialyzed fetal bovine serum were performed at 37°C. Drug-free medium was then added (2

43

ml/well), and cultures were incubated for 4 additional days at 37°C. Finally, cells were detached by trypsinization and the number of viable cells was determined microscopically using trypan blue exclusion. Western blot analysis of BCRP: BCRP levels as normalized to β-tubulin were determined by Western blots using monoclonal antibodies BXP-53 and clone 2-28-33, respectively, as previously described [92, 102]. Mitoxantrone accumulation and immunohistochemical localization of BCRP in specific colonies of MCF-7/MR and MCF-7/FLV 1000 cells: Cells (5x104) were seeded in 25-mm tissue culture flasks (5 ml medium) and incubated for 4 days. Then, the growth medium was replaced with a fresh one containing 20 µM mitoxantrone. After 12 hr of incubation at 37°C, monolayer cells were washed three times with medium containing 10% dialyzed fetal bovine serum. Then, 1ml of medium was added to each flask and random colonies were examined using a LEICA microscope at a bright field mode. For immunohistochemical staining of BCRP, monolayer cells were then processed as previously described [92, 102]. The immunolocalization results obtained with BXP-53 were fully corroborated with other monoclonal antibodies to BCRP including BXP-21 and BXP-43. Determination of the number of light-refracting extracellular vesicles: MCF-7 and MCF-7/MR cells (5x104) were seeded in 25-mm tissue culture flasks (5 ml medium/flask) and incubated for 4 days at 37°C, following which the medium was replaced with a fresh one (1ml/flask). Random colonies were examined for visible, light-refracting extracellular vesicles using a LEICA microscope at a bright field mode. Three independent experiments were performed using ~200 cells in each determination for each cell line.

44

Inhibition of mitoxantrone accumulation with BCRP transport inhibitors and ATP- depleting agents: Cells were seeded in 24-well plates (104/well) and incubated for 4 days at 37oC. Then, the medium of a control well was replaced with a fresh one. In contrast, the medium of the three remaining wells was replaced with one containing either 0.4 µM Ko143, 5 µM FTC or the combination of the metabolic energy inhibitors 5 µM FCCP and 5 mM sodium azide. Following 1 hr of incubation at 37°C, mitoxantrone was added to a final concentration of 20 µM. After 6 hr of incubation at 37°C, the growth medium was aspirated and a wash step with medium containing 10% dialyzed fetal bovine serum was performed. Then, fresh medium was added (0.3 ml/well) and random colonies were rapidly examined for their mitoxantrone blue staining using a LEICA microscope at a bright field mode. In order to remove the BCRP transport and metabolic energy inhibitors, the cells were incubated twice (each for 7 min) in fresh growth medium at 37oC, followed by aspiration of the medium. Fresh medium containing 20 µM mitoxantrone was then added; cells previously incubated with FCCP and azide were also supplemented with the ATP-restoring agents sodium pyruvate (1mM; GIBCO BRL) and D-glucose (5 mM; Sigma). After 6 hr of incubation at 37°C, the medium was discarded, and cells were washed once with medium containing 10% dialyzed fetal bovine serum. Then, fresh medium was added (0.3 ml/well) and random colonies were examined microscopically for mitoxantrone accumulation. Estimation of the intravesicular concentration of mitoxantrone: To explore the time-dependence of mitoxantrone accumulation in the intravesicular lumen, we incubated cells with 20 µM mitoxantrone for 3-12 hr at 37oC. Photographs of random colonies were taken using a LEICA microscope at a bright field mode. In order to generate a calibration curve, 10 µl aliquots of standard solutions containing increasing

45

mitoxantrone concentrations were dispensed onto glass slides which were then covered by glass coverslips. Photographs were taken at random locations using a bright field mode. Photographs were then transformed to a gray scale format and analyzed individually by scanning densitometry using the program "TINA" (version 2.10g). The densitometric background levels of the calibration curve (i.e. at a zero mitoxantrone concentration) and the monolayer cell culture with unstained extracellular vesicles (t = 0) were numerically normalized. The experiment was performed three times using ~150 extracellular vesicles in each experiment for each incubation time with mitoxantrone. Autofluorescence detection with viable cells: Cells (4x103) were seeded in 24-well plates (2ml phenol red-free medium/well) for 4 days at 37 oC in the presence or absence of 0.4 µM Ko143. Then, 1.5 ml of the growth medium was removed and random colonies were examined for their autofluorescence using a fluorescence microscope at an FITC-like mode; a bright field mode examination was also used here. Confocal microscopy of BCRP confinement to cell-cell attachment zones: Cells (4x103) were seeded in 24-well plates (1ml medium/well) on sterile glass coverslips and incubated for 4 days at 37oC. Cells were then washed, reacted with the BCRP -specific monoclonal antibody BXP-53 followed by a secondary goat anti-rat IgG and a third FITC-conjugated rabbit-anti-goat IgG as described recently [102]. The fluorescent cells were then examined using a Bio-Rad MRC1024 confocal microscope for cross-section and perpendicular section analyses. Confocal microscopy studies of the accessibility of the culture medium to the extracellular vesicles: MCF-F/MR cells (4x103/well) were seeded in 24-well plates (2ml medium/well) on sterile glass coverslips and incubated for 3 days at 37oC. The

46

growth medium was removed and monolayer cells were washed twice with a fresh medium. Then, a PBS solution containing 10% fetal calf serum, 15% tetramethylrhodamine isothiocyanate (TRITC)-goat anti-rabbit IgG (Sigma, T5268) was placed between the coverslips and the glass slides. The slides were then examined using a Bio-Rad MRC1024 confocal microscope for cross-section and perpendicular section. The FITC-like mode was used to follow the green autofluorescence of the vesicles and the red fluorescence of the culture medium was detected by a Kr/Ar laser (excitation at 568 nm and emission at 585 nm). Electron microscopy studies: The presence of extracellular vesicles and their fine structure were studied by first seeding the MCF-7/MR cells on glass slides mounted on 8-well tissue culture chambers (Lab-Tek, Nunc). The cells were grown for 4 days until confluence was achieved; an examination under a light microscope revealed the presence of numerous large vesicles. Slides containing monolayer cells were fixed using an overnight incubation in 2% glutardialdehyde in phosphate buffer and an additional 30 min incubation in osmium tetroxide/collidine (2:1) buffer for 30 min. Slides were dehydrated using solutions containing increasing concentrations of ethanol (70%-100%). Then, the chambers were discarded and the slides containing monolayer cells were impregnated for 1.5 hr in a solution of Epon/propylene oxide/DMP-30 (1:1:0.02). Cells were embedded by placing on top of them open-end capsules that were filled with embedding fluid (Epon/DMP-30 1:0.015) following which polymerization was allowed overnight at 70oC. After polymerization, the glass slides were removed by snap freezing in liquid nitrogen and thawing thereby resulting in the entrapment of the monolayer cells in the polymerized resin. Then, ultra thin sections (60-70 nm) were cut using a Diatome diamant knife and an LKB Ultrotome III, and collected on support film-coated (1.5% Formvar in dichloroethane) Cu grids.

47

The sections were then counterstained with 2% uranyl acetate for 20 min and lead nitrate/sodium tricitrate for 20 min and then examined with a Jeol 1200EX electron microscope. Photographs were finally printed using a Leitz Focomat IIc. Statistical Analysis: we used a Student’s t test to examine the significance of the difference between two populations for a certain variable. A difference between the averages of two populations was considered significant if the P value obtained was < .0.05

48

Results A) RFC mediates intracellular folate depletion and consequent cytotoxicity under folate deprivation:

Effect of RFC overexpression on cellular proliferation under folate deplete conditions: RFC functions as an anion-exchanger with a high affinity reduced folate efflux activity. Hence, we hypothesized that under conditions of folate deprivation (i.e. when folates are lacking in the extracellular milieu), the high affinity folate efflux activity of RFC may result in intracellular folate depletion and consequent cell death (Fig. 7B).

49

Fig. 7. Model of intracellular folate metabolism under replete (A) and deplete conditions (B)

To explore this hypothesis, RFC null Chinese hamster ovary (CHO) C5 cells and their stable C5/RFC transfectants overexpressing the hRFC [97] as well as the folate receptor α overexpressing cell line C5/FR were examined for cell growth in medium lacking or containing folates. Following six days of incubation in these media, viable

50

cell numbers were determined (Fig. 8). The cellular growth rates of C5, C5/RFC and C5/FR cells in folate-replete growth medium (i.e. C5-HF, C5/RFC-3nMLCV and C5/FR-3nMFA, respectively) were similar. In contrast, in folate-free medium, the cellular proliferation rate of C5/RFC cells (i.e. C5/RFC-NF) but not C5/FR-NF cells was 14.8-fold (P value= 0.025) decreased when compared to the RFC null C5 cells (i.e. C5-NF). These results lend support to our hypothesis that overexpression of the human RFC is detrimental to cellular proliferation under conditions of folate deprivation.

51

Fig 8: Effect of RFC overexpression on cellular proliferation under folate deplete conditions. The Chinese hamster ovary cell line deficient in RFC activity termed C5 as well as their RFC and folate receptor α (i.e. FR) overexpressing transfectants were trypsinized and washed three times with folic acid-free growth medium. Then, cells from each subline (6x104) were seeded in each of two T25 flasks in 5 ml folic acidfree growth medium. The sublines in the first set of flasks were termed C5/NF, C5/FR-NF and C5/RFC-NF (i.e. no folate) and were incubated for 6 days in a humidified CO2 incubator. The sublines in the second set of flasks were supplemented with 2.3 µM, 3nM folic acid or 3 nM LCV resulting in the sublines C5-HF, C5/FR3nMFA and C5/RFC-3nMLCV sublines respectively. These three folate supplemented sublines were also incubated for 6 days in a humidified CO2 incubator. Following these 6 days of incubation, the cells were detached by trypsinization and the number of viable cells was determined by a haemocytometer counting after trypan blue staining (Y axis). Gene expression status of folate influx and efflux transporters as well as folatedependent enzymes under folate deplete- and replete conditions: After demonstrating the cytotoxic effect elicited by RFC overexpression under folate deplete conditions, we explored the gene expression status of RFC as well as folylpoly-γ-glutamate synthetase (FPGS) and γ-glutamate hydrolase (GGH), key folate-dependent enzymes mediating intracellular folate polyglutamylation and hydrolysis, respectively. Moreover, the transcript levels of the ATP-binding cassette (ABC) multidrug resistance efflux transporters MRP1, MRP5 and BCRP, all of which display low affinity, high capacity ATP-dependent folate efflux activity [103], were also determined. Quantitative RT-PCR analyses revealed a 2.4-fold (p.value= 0.01) and 2.6-fold (p.value= 0.005) decrease in RFC and GGH mRNA levels upon 7 days of folate deprivation in MCF-7/MR breast cancer cells (i.e. MCF-7/MR-NF versus MCF-7/MR-HF), whereas no alterations in the gene expression status were observed with FPGS, MRP1, MRP5, BCRP, GAPDH and actin (Fig. 9). Similarly, although CCRF-CEM-7A cells have a stable RFC gene amplification, a 2.5-fold decrease (p.value= 0.009) in RFC mRNA levels was observed in these cells after 3 days of incubation in folate-free medium (i.e. CEM/7A-NF versus CEM7A-HF; Fig.

52

9). In contrast, no changes were observed in the mRNA levels of FPGS, GGH, MRP1, GAPDH and actin.

Fig 9. Gene expression status of folate influx and efflux transporters as well as folate-dependent enzymes under folate deplete- and replete conditions. Total cellular RNA was extracted from the folate-supplemented cell lines MCF7/MR-HF and CEM/7A-HF as well as from their folate-deprived counterparts MCF7/MR-NF and CEM/7A-NF cell lines, respectively. Then, the transcript levels of RFC, GGH, FPGS, MRP1, MRP5, BCRP and GAPDH of the .various sublines were quantified by quantitative RT-PCR analyses Decreased RFC activity in folate-deprived cells: Experiments were set up to corroborate the decrease in the transcript levels at the activity level. In order to assess the transport activity of RFC under folate-deplete and replete conditions, we determined the initial rates of [3H] methotrexate uptake in two cell lines with low and high levels of RFC expression. Consistent with the decreased transcript levels of the

53

hRFC in folate-deprived cells, both MCF-7/MR (expressing low RFC levels) as well as CEM-7A cells (overexpressing high RFC levels) showed 49% (p. value= 0.03) and 44% ( p. value=0.004 ) decrease in the influx of [3H]MTX under folate-deplete conditions, respectively (Fig. 10A-B).

Fig 10. [3H]MTX transport in folate supplemented and deprived sublines. Initial rates of [3H]MTX transport were determined for the folate-supplemented cell lines MCF7/MR-HF and CEM/7A-HF as well as for their folate-deprived counterparts MCF7/MR-NF and CEM/7A-NF sublines, respectively. B) Folate deprivation results in the loss of breast cancer resistance protein (BCRP/ABCG2) expression: a role for BCRP in cellular folate homeostasis: The following results enabled us to examine the possible role of BCRP in folate homeostasis: Loss of BCRP expression in the LF-adapted cell lines as revealed by Western blot analysis: The levels of BCRP as well as MRP1 through MRP5 were determined

54

in MCF-7/LF and MCF-7/MR-LF cells relative to their parental counterparts (Fig 11). Western blot analysis revealed that folate deprivation in MCF-7/LF cells resulted in an 18-fold decrease in BCRP levels, relative to parental MCF-7 cells (Fig 11A); in contrast, no changes were observed in MRP1 levels (Fig 11B). Similarly, whereas Mitoxantrone-resistant MCF-7/MR cells displayed a ~55-fold BCRP overexpression, relative to parental MCF-7 cells, the MCF-7/MR-LF subline expressed only residual BCRP levels (Fig 11A); remarkably, these barely detectable levels of BCRP in MCF7/MR-LF cells were ~ 4-fold lower than those present in parental MCF-7 cells. Furthermore, folate deprivation in MCF-7/MR-LF cells resulted in a simultaneous 5fold decrease in MRP1 levels, relative to their parental MCF-7/MR cells (Fig 11B). It should be noted that when compared to parental MCF-7 cells, MCF-7/MR cells contained 3-fold less MRP1 levels even before gradual folate deprivation was initiated. Importantly, MCF-7/MR-HF cells grown in the absence of Mitoxantrone but in the continuous presence of 2.3 μM folic acid had only a slight decrease in BCRP expression, relative to the near complete loss of BCRP in MCF-7/MR-LF cells (Fig 11A). Retention of poor MRP2 through MRP5 expression in the LF-adapted cell lines: MRP2-5 have the facility to export folates [34, 103]. We therefore determined the levels of these transporters in the LF-adapted cell lines (Fig 11B). MRP2, MRP3 and MRP5 were essentially undetectable in both parental cells and their LF-adapted sublines, whereas MRP4 was expressed at equally low levels in both cells lines (Fig 11B). Reprobing with a β-tubulin monoclonal antibody was used to correct for any differences in the amounts of Triton X-100-soluble proteins that were actually analyzed (Fig 11 A, B).

55

Fig 11: Western blot analysis of BCRP as well as MRP1 through MRP5 expression in parental cells and their LF-adapted cell lines. Aliquots of Triton X100-soluble membrane proteins (6-60 μg) were resolved by electrophoresis on 7.5% polyacrylamide gels containing SDS and electroblotted onto a Protran nylon membrane. Then, the membranes were reacted with monoclonal antibodies against BCRP (A) or MRP1 through MRP5 (B), following which a second peroxidaseconjugated antibody was added and membranes were developed using a standard ECL procedure. To correct for loading differences, the blots were stripped and reprobed with an antibody against β-tubulin. Note that in order to estimate the barely detectable BCRP expression in the LF-adapted cell lines (A), the MCF-7/LF and MCF-7/MR-LF lanes were intentionally loaded with excess protein (60 μg) relative to their parental counterparts (30 μg). The “control” lane in Panel B contained membrane protein extracts (6 μg) from cell lines with overexpression of MRP1 through MRP5 as detailed in Materials and Methods. Semi-quantitative RT-PCR analysis was performed in order to estimate BCRP gene expression in the LF-adapted cell lines as compared to their parental cells (C). A parallel RT-PCR with GAPDH was used in order to normalize for the amounts of total cDNA used in each lane (see Materials and Methods). Note that a ~3-fold lower levels of MCF-7/MR-HF cDNA were analyzed in order to retain comparability of signal intensity (C). Poor BCRP gene expression in the LF-adapted cell lines as revealed by RT-PCR: We next examined by semi-quantitative RT-PCR analysis whether the marked loss of BCRP protein levels in the LF-adapted cell lines was due to decreased BCRP gene expression. Whereas parental MCF-7 cells expressed notable levels of BCRP mRNA, MCF-7/MR-HF cells displayed a prominent overexpression of BCRP mRNA (Fig 56

11C). In contrast, the low folate (LF) adapted sublines MCF-7/LF and MCF-7/MRLF contained only residual levels of BCRP mRNA (Fig 11C, upper panel). An RTPCR of a housekeeping gene (GAPDH) confirmed that comparable levels of cDNA were being analyzed in the various cell lines (Fig 11C, lower panel).

Loss of BCRP expression in the LF-adapted cell lines as revealed by immunohistochemistry: To confirm the loss of BCRP expression in the LF-adapted cell lines, we performed immunohistochemistry with BXP-53, a novel monoclonal antibody to BCRP (Fig 12). MCF-7/MR cells growing in the continuous presence of 0.1 μM Mitoxantrone displayed an intense cellular staining (Fig 12A) and a dominant plasma membrane localization (Fig 12A, arrows). Consistently, MCF-7/MR-HF cells growing in a medium lacking Mitoxantrone but containing 2.3 μM folic acid retained a relatively strong cellular staining (Fig 12B). In contrast, the poor BCRP expression in MCF-7/MR-LF cells (Fig 12C), MCF-7/LF cells (Fig 12E), and the low BCRP levels in parental MCF-7 cells (Fig 12D) were below the level of detection by immunohistochemistry.

Fig 12: Immunohistochemical detection of BCRP expression in parental cells and their LF-adapted cell lines. Exponentially growing MCF-7/MR (A), MCF-7/MR-HF (B), MCF-7/MR-LF (C), MCF-7 (D) and MCF-7/LF (E) cells in 24-well culture

57

plates were fixed with 4% formaldehyde, reacted with an anti-BCRP monoclonal antibody, BXP-53. Then, an HRP-conjugated rabbit anti-rat IgG was added and color development was carried out using the chromogen 3,3’-diaminobenzidine. Finally, cells were counterstained with haematoxylin and examined with a light microscope at a 200x magnification. The arrows in panel A denote the plasma membrane localization of BCRP in MCF-7/MR cells growing in the continuous presence of 0.1 μM Mitoxantrone. Note that MCF-7/MR-HF cells (B) were grown for three and half months in Mitoxantrone -free medium containing 2.3 μM folic acid. Loss of Hoechst 33342 efflux in the LF-adapted cell lines: In order to determine whether the marked loss of BCRP expression in the LF-adapted cell lines was associated with a parallel fall in BCRP drug efflux activity, we used an online efflux assay [91] of the chromophore Hoechst 33342, an established BCRP efflux substrate. In this functional assay, monolayer cells were loaded with 10 μM of Hoechst 33342 [104]. After 2 hr of chromophore loading at 37oC, the culture medium was replaced by a fluorophore-free medium and the extent of Hoechst 33342 efflux was continuously monitored by measuring the increase in the fluorescence in the external medium (Fig. 13). MCF-7/MR-HF cells with BCRP overexpression displayed a marked efflux of Hoechst 33342, whereas parental MCF-7 cells with low BCRP expression had a lower efflux. In contrast, the LF-adapted sublines MCF-7/LF and MCF-7/MR-LF cells which essentially lost BCRP expression had only a background efflux of Hoechst 33342; curve-fitting of these efflux data confirmed the poor efflux of Hoechst 33342 in the LF-adapted cell lines (Fig. 13, Inset). Thus, loss of BCRP expression was accompanied by a parallel fall in the efflux of Hoechst 33342, an established BCRP substrate.

58

10000

Fluorescence (arbitrary units)

MCF-7 8000

Y Data

6000

4000

2000

MCF-7 MCF-7/LF MCF-7/MR-LF X D a t a MCF-7/MR-HF 0

0

200

400

600

Time (sec)

Fig 13: Online efflux of Hoechst 33342 from monolayers of parental and LFadapted cell lines. Following loading of MCF-7/MR-HF (dark blue tracing), MCF7/MR-LF (yellow), MCF-7 (purple), and MCF-7/LF cells (light blue/green) with 10 μM Hoechst 33342, cells were washed, and the fluorescence of the extruded chromophore in the extracellular medium was continuously monitored (every second for up to 5,000 seconds) by an online computerized spectrofluorometer. The fluorescence depicted was obtained after normalization for differences in the number of cells per well; this was achieved by DNA staining with Syto 13 as detailed in Materials and Methods. Inset: Curve-fitting of the Hoechst 33342 efflux data was performed as previously described by Wielinga et al., [105]. Accumulation of Mitoxantrone in the LF-adapted cell lines as revealed by flow cytometry: Using flow cytometry, we further determined whether the loss of BCRP expression and efflux function would consistently result in increased Mitoxantrone accumulation in the LF-adapted cell lines (Fig 14). Indeed, upon a 1 hr incubation with 20 μM Mitoxantrone, MCF-7/MR-LF and MCF-7/LF cells displayed statistically significant (P=0.003 and P=0.018, respectively) increases of 1.9-fold and 2.0-fold in the net accumulation of Mitoxantrone, respectively, relative to their parental cells (Fig 14). 59

Fig 14: Flow cytometric analysis of Mitoxantrone accumulation in parental cells and the LF- adapted cell lines. Exponentially growing cells were detached by trypsinization and incubated in growth medium containing 20 μM Mitoxantrone for 1 hr at 37OC. Cells were then washed with ice-cold PBS and analyzed for mean linear fluorescence per cell using a flow cytometer. Cellular Mitoxantrone fluorescence was obtained after subtraction of the autofluorescence of unstained cells. Results presented are means ± S.D. of three independent experiments performed in duplicates. The asterisks denote statistically significant (Student T-test) changes in MCF-7/LF vs. its parental MCF-7 cells, as well as MCF-7/MR-LF vs. its MCF-7/MR-HF parental counterpart.

Sensitivity of the LF-adapted cell lines to Mitoxantrone: We therefore determined whether the loss of BCRP expression and Mitoxantrone efflux activity was also accompanied by an increase in Mitoxantrone sensitivity in the low folateadapted cell lines. MCF-7/LF cells that essentially lost BCRP expression

60

exhibited a 2.5-fold increased sensitivity to Mitoxantrone, relative to parental MCF-7 cells (Fig 15A). Consistently, MCF-7/MR-HF cells displayed a 34-fold resistance to Mitoxantrone, relative to parental MCF-7 cells (Fig 15B); whereas, MCF-7/MR-LF cells which lost BCRP expression and efflux activity exhibited 84-fold increased sensitivity to Mitoxantrone, relative to their MCF-7/MR-HF counterpart (Fig 15B). Importantly, disruption of BCRP efflux activity in MCF-7 and MCF-7/MR-HF cells by the specific and potent BCRP inhibitor Ko143 [106] rendered these cell lines 2.1- and ~16.4-fold more sensitive to Mitoxantrone, respectively (Fig 15A and Fig 15B, respectively).

61

Cell Growth (% of Cell Growth (%Cell of control) Growth (% of control) Cell Growth (% of control) control)

140

A

120

MCF-7/W.T MCF-7/LF MCF-7/W.T+KO143 + ko143

100 80 60 40 20 0 1 140

10

B

120

100

1000

[M itoxantrone] (nM )

MCF-7/MR-HF MCF-7/MR-LF MCF-7/MR-HF+KO143 + ko143 MCF-7/W.T

100 80 60 40 20 0 1

10

100

1000

10000

[Mitoxantrone] (nM) [Mitoxantrone] (Mitoxantrone)(nM) [nM]

Fig 15: Cellular growth inhibition with Mitoxantrone: Parental MCF-7 and MCF-7-LF (A) as well as parental MCF-7, MCF-7/MR-HF and MCF-7/MR-LF cells (B) growing in monolayers in 24-well plates were exposed to various concentrations of Mitoxantrone in the absence or presence of 0.3 μM Ko143, a potent BCRP inhibitor. Following 72 hr incubation at 37oC, cells were detached by trypsinization and the number of viable cells was determined using trypan blue exclusion. Results depicted are the means ± S.D. of three independent experiments.

62

Loss of MTX-resistance in MCF/MR-LF cells: BCRP was shown to export MTX, thereby conferring resistance to this antifolate, particularly upon short-term drug exposure [38, 41]. When compared to parental MCF-7 cells, MCF-7/MR cells displayed only 2-fold resistance upon a continuous MTX exposure (Fig 16A), but as high as 28-fold resistance to MTX upon 4 hr antifolate exposure (Fig 16B). In contrast, loss of BCRP expression in MCF-7/MR-LF cells resulted in a complete loss of MTX resistance upon 4 hr drug exposure, thereby resulting in an IC50 value (~80 nM) that was identical to that obtained with MCF-7/LF cells (Fig 16C).

63

Fig 16: Cellular growth inhibition with MTX. Cells were seeded in 24-well plates and allowed to attach for 24 hr at 37oC. Attached cells were then exposed continuously (A) or pulsed for 4 hr at 37oC with various concentrations of MTX (B, C). Following 4 hr of exposure to MTX, the drug-containing medium was removed and three successive washes each of ~10 min in drug-free medium were performed at 37oC. Drug-free growth medium was added (2ml/well) and cultures were incubated for 4 days at 37oC. Cells were then detached by trypsinization and the number of viable cells was determined using trypan blue exclusion.

64

Increased accumulation of [3H]Folic acid in the LF-adapted cell lines: We next examined the ability of the LF-adapted cell lines to accumulate [3H]folic acid as compared to their parental cell lines. After 4 hr (Fig 17A) and 24 hr incubation (Fig 17B) with 1 μM [3H]folic acid at 37oC, MCF-7/LF and MCF-7/MR-LF cells displayed a ~2-fold increase in the accumulation of [3H]folic relative to their parental counterparts. This increased accumulation of [3H]folic acid after 4 hr and 24 hr incubation was statistically significant for both MCF-7/LF (P=0.015 and P=0.026, respectively) and MCF-7/MR-LF cells (P=0.031 and P=0.033, respectively), when compared to their parental cells. In contrast, the trivial elevation in the accumulation of [3H]folic after 4 hr and 24 hr in MCF-7/MR-HF cells relative to MCF-7/MR cells was not statistically significant (P=0.549 and P=0.249, respectively).

65

Fig 17: [3H]Folic acid accumulation in parental and LF-adapted cell lines. Monolayer cells were washed with folate-free medium, then incubated for 4hr (A) and 24 hr (B) at 37oC in HBS containing 1 μM [3H]folic acid as detailed in Materials and Methods. Transport was stopped by the addition of 10 ml of ice-cold HBS. Cells were then detached by trypsinization, washed with ice-cold HBS, and the final cell pellet was lysed in 0.2 ml water and the radioactivity released was determined using a liquid scintillation spectrometer. The asterisks denote statistically significant changes in MCF-7/LF vs. parental MCF-7 cells, as well as MCF-7/MR-LF vs. its MCF-7/MR-HF parental counterpart.

Increased FPGS activity in the LF-adapted cell lines: In a previous study we have shown that gradual folate deprivation achieved by a stepwise increase in the antifolate pressure resulted in a substantial increase in FPGS activity [107] in pyrimethamineresistant CHO cells [108]. Here, consistently, folate deprivation resulted in 66

statistically significant increases of 63% (P=0.001) and 20% (P=0.008) in FPGS activity in MCF-7/MR-LF and MCF-7/LF cells, relative to their parental counterparts (Fig 18). In summary, the gradual folate deprivation in MCF-7 and MCF-7/MR cells resulted in the loss of BCRP expression and efflux function as well as in a significant increase in the activity of FPGS, the key enzyme responsible for cellular retention of long chain (> 3 glutamate residues) folate polyglutamates.

(pmole [3H]Glu hr/mg protein)

FPGS Activity

2000 1500

*

*

1000 500 0

F2 F 71 F H L L F R RC -7 / M M CF 7 / -7 /M M F C CF M M

Fig 18: Histogram of FPGS activity in parental cells and their LF-adapted cell lines. The catalytic activity of FPGS in the cytosolic fraction isolated from the various cell lines was determined as described in Materials and Methods. Results presented are the means ± S.D. of three independent experiments. The asterisks denote statistically significant changes in the LF-adapted sublines when compared to their parental cell lines.

67

C) Cytoplasmic Confinement of BCRP as a Novel Mechanism of Adaptation to Short-Term Folate Deprivation: Establishment of a Short-Term Folate Deprivation Protocol. Our above results showed that long-term gradual deprivation of folic acid from the growth medium resulted in the near complete loss of BCRP expression along with a marked decrease in MRP1 expression in Mitoxantrone -resistant MCF-7/MR breast cancer cells [92]. Here, we explored the mode of adaptation of these BCRP-overexpressing cells upon a short-term deprivation of folic acid from the growth medium. Toward this end, we established a short-term folate deprivation protocol (Fig. 19). In brief, MCF-7/MR cells growing in a high folic acid (2.3 µM) medium containing Mitoxantrone were washed with excess PBS and distributed to three groups; one group continued to grow in the above-mentioned medium and was termed MCF-7/MR-HF-MR, whereas the second group was grown in drug-free medium containing high folic acid and was therefore termed MCF-7/MR-HF. The third group, which was termed MCF-7/MRNF-LF, was grown for 2 weeks in folic acid-free medium (i.e., the folate deprivation step) followed by an additional week of adaptation to low folic acid (1 nM). At the end of this 3-week period, cells from the various groups were processed for various analyses. To confirm that folate-deprived MCF-7MR-NF-LF cells retained normal cell cycle kinetics, we performed a flow cytometric analysis with propidium iodide-stained cells. Folate-deprived cells displayed a normal cell cycle distribution in the G1, S, and G2M phases, compared with the control groups growing in high folate medium (Fig. 20). Furthermore, in the group of folate-deprived cells, the apoptotic/dead cell fraction was relatively small (8.1 ± 0.5%) and was similar to that obtained with control cells growing in high folate medium (8.8 ± 0.5%). As would be expected, MCF-7/MR-HFMR cells growing in a medium containing the cytotoxic drug Mitoxantrone displayed

68

a slight increase in the fraction of apoptotic cells (13.2 ± 0.1%). Furthermore, the percentage of cells with >4n DNA content in MCF-7/MR-HF-MR and MCF-7/MRHF cells was comparable at 21.9 ± 1.9 and 20.2 ± 1.7, respectively, whereas the MCF7/MR-NF-LF subline had a lower fraction of cells with a >4n DNA content (11.2 ± 3.2%). These results are consistent with the well established genomic instability and chromosomal aberrations of cultured tumor cell lines.

MCF-7/MR ( BCRP overexpressing cell line )

High folic acid (2.3µM) medium with 100 nM mitoxantrone ( Promotion of plasma me mbrane localization of BCRP) : One week Two washes with PBS

High folic acid medium with 100 nM mitoxantrone :Three weeks

High folic acid medium: Three weeks

Folic acid free medium (Cellular folate pool depletion step): Two weeks

Low folic acid medium (1nM) (Low folic acid adaptation step) : One week

MCF-7/MR-HF-MR

MCF-7/MR-HF

MCF-7/MR-NF-LF

Fig. 19. Schematic presentation of the short-term folate deprivation protocol. For details, see Results.

69

MCF-7/MR-HF-MR

G0/G1 G2/M

Apoptotic cells =13.2% ± 0.1

G0/G1 Events

MCF-7/MR-HF G2/M

Apoptotic cells =8.8% ± 0.5

MCF-7/MR-NF-LF

G0/G1 G2/M

Apoptotic cells =8.1% ± 0.5

DNA Fluorescence

Fig. 20. Cell cycle analysis of folate-deprived cells and their control counterparts. After cell fixation with ethanol and staining with the DNA dye propidium iodide, cell cycle analysis was performed using a flow cytometer. The average fraction of apoptotic/dead cells (±S.D.) represented by a lower than G1 DNA content is depicted for each cell line and has been obtained from three separate experiments. Expression and Glycosylation of BCRP in Short-Term Folate-Deprived Cells and Their Control Counterparts. Because we shown above that long-term folate deprivation results in a dramatic loss of BCRP and MRP1 expression [92], we first determined the status of expression of these transporters in short-term folate-deprived cells. Western blot analysis with monoclonal antibodies to BCRP (Fig. 21, A and B) and MRP1 (Fig. 21C) revealed a 3-fold decrease in their levels in folate-deprived

70

cells, relative to parental cells growing in a high folate medium (Fig. 21, A-C). Because two closely migrating BCRP species were apparent in both MCF-7/MR-NFLF cells and their parental MCF-7/MR counterparts (Fig. 21A), we undertook experiments to rule out the possibility that folate deprivation results in alterations in BCRP glycosylation. Thus, MCF-7/MR-NF-LF cells and their parental MCF-7/MR counterparts were treated with the N-glycosylation inhibitor tunicamycin (10 µg/ml) for 24 h, after which Western blot analysis was performed with a monoclonal antibody to BCRP (Fig. 21B). The completely unglycosylated BCRP in all tunicamycin-treated cell lines migrated as a faint 72-kDa protein, thereby being much lower in its molecular mass than either of the two glycosylated high molecular mass BCRP bands ( 80 and 82 kDa, respectively). Hence, the 80- and 82-kDa BCRP species observed in equal ratios in both MCF-7/MR-NF-LF and MCF-7/MR-HF cells are both glycosylated but to slightly different extents, whereas the unglycosylated BCRP obtained after treatment with tunicamycin has a much lower molecular mass as would be predicted from the calculated core molecular mass of unglycosylated BCRP. Hence, short-term folate deprivation does not alter the extent of glycosylation of BCRP. Reprobing with a β-tubulin antibody confirmed that equal amounts of proteins were analyzed (Fig. 21E).

71

Fig. 21. Western blot analysis of BCRP, MRP1, and Pgp in folate-deprived cells and their control counterparts. Triton X-100-soluble membrane proteins (20 µg) were resolved by electrophoresis on polyacrylamide gels containing SDS, electroblotted onto Protran BA83 cellulose nitrate membranes, and reacted with monoclonal antibodies against BCRP (A and B), MRP1 (C), or Pgp (D). Membrane proteins shown in B were isolated after 24-h treatment of the various cell lines with 10 µg/ml N-glycosylation inhibitor tunicamycin. Blots were then reacted with a second HRP-conjugated antibody, and these nylon membranes were developed using a standard enhanced chemiluminescence procedure. To correct for loading differences, the blots were stripped and reacted with an antibody against β-tubulin (E). The "overexpressor" lane contained protein extracts (6 µg) from cell lines with overexpression of BCRP (MCF-7/MR), MRP1 (2008/MRP1), and Pgp (EmtR1).

Subcellular Localization of BCRP in Short-Term Folate-Deprived Cells and Their Control Counterparts. We have shown above that, relative to their parental MCF-7 cells, MCF-7/MR cells display a 55-fold BCRP overexpression, the large fraction of which is localized in the plasma membrane [92]. Hence, the surprisingly modest decrease in BCRP and MRP1 levels in the short-term folate-deprived cells

72

here apparently could not account for their survival under folate-deficient conditions when taking into consideration the potent folate efflux activity of BCRP in MCF7/MR cells [42, 43, 92]. Therefore, we further explored the expression and subcellular localization of BCRP in these cells by immunohistochemistry. MCF-7/MR-HF-MR and MCF-7/MR-HF cells growing in high folic acid medium displayed an intense plasma membrane staining of BCRP, particularly at zones of cell-cell adhesion (Fig. 22, A and B, top, see arrows). In contrast, in folate-deprived cells, BCRP was highly confined to the cytoplasm (Fig. 22C, dashed arrow). These results were corroborated with immunofluorescence analysis of cells stained with DAPI, a DNA dye with a blue fluorescence. Thus, the green fluorescence of BCRP clearly localized to zones of cellcell attachment in MCF-7/MR-HF-MR and MCF-7/MR-HF cells as aided by the blue fluorescence of the nuclei (Fig. 22, A and B, bottom). In contrast, BCRP was confined to the cytoplasm in folate-deprived cells (Fig. 22C, bottom). Furthermore, detailed time-course experiments revealed that the first significant appearance of the cytoplasmic BCRP localization was observed only after 2 weeks of folate deprivation followed by at least 4 additional days of adaptation in 1 nM folic acid-containing medium.

73

BCRP(DAB) and Nuclei (haematoxylin )

C. MCF-7/MR-NF-LF

Immunohistoche mistry

A. MCF-7/MR-HF-MR B. MCF-7/MR-HF

BCRP (FITC)

Immunofluorescence

DAPI

Merge

Fig. 22. Immunohistochemical and immunofluorescence detection of BCRP in parental cells and their folate-deprived cells. Top, monolayer MCF-7/MR-HF-MR (A), MCF-7/MR-HF (B), and folate-deprived MCF-7/MR-NF-LF cells (C) were fixed with 4% formaldehyde and reacted with an anti-BCRP monoclonal antibody, BXP-53. Then, an HRP-conjugated rabbit anti-rat IgG was added, and color (brown) development was carried out using the chromogen 3,3'-diaminobenzidine. Cells were then counterstained with hematoxylin and examined with a light microscope at a 200x magnification. The arrows in A and B denote the plasma membrane localization of BCRP, particularly at the regions of cell-cell attachment in MCF-7/MR-HF-MR and MCF-7/MR-HF cells, respectively, whereas the dashed arrow represents the cytoplasmic localization of BCRP in folate-deprived MCF-7/MR-NF-LF cells (C). Bottom, immunofluorescence detection with an FITC-conjugated antibody to BCRP (green fluorescence). Nuclei were counterstained with the DNA dye DAPI (blue fluorescence). Note the plasma membrane localization of BCRP in the region of cellcell attachment in MCF-7/MR-HF-MR and MCF-7/MR-HF cells, whereas BCRP is confined to the cytoplasmic compartment in MCF-7/MR-NF-LF cells.

74

To provide a quantitative assessment of this markedly altered subcellular distribution, we devised a computerized whole-cell scanning technique (see Materials and Methods) and thereby determined the percentage of BCRP in the cytoplasm and plasma membrane fractions in the various cell lines after immunohistochemical staining with an anti-BCRP antibody. MCF-7/MR-HF-MR cells contained 62 ± 9.8% of their BCRP in the plasma membrane and only 38 ± 9.8% in cytoplasm, whereas folate-deprived cells contained as high as 86 ± 1.7% of their BCRP in the cytoplasm and only 14 ± 1.7% in the plasma membrane (Fig. 23A); this dramatic increase in the cytoplasmic fraction in folate-deprived cells was statistically significant (P = 0.002). Furthermore, whereas the cytoplasmic/plasma membrane distribution ratio of BCRP in parental MCF-7/MR-HF-MR cells was 0.65 ± 0.28, folate-deprived cells had a statistically significant, 9.3-fold increase in this ratio (6.06 ± 0.84; P = 0.001; Fig. 23B). These results establish that BCRP is highly confined to the cytoplasm in the short-term folate-deprived cells.

75

% of Fraction

60 40 20 0

M

CF

M 7/

H R-

M F-

R

M

/M -7 F C

R-

M

H

F

/M -7 CF

N R-

F

F -L

M

8 BCRP Ratio

A

Membranal Plasma mefracti mbr ane on Cytopl asmic fraction 80 Cytocolic

Cytoplas mic/ Plas ma Membrane

100

B

6 4 2 0

CF

M 7/

-H R

M F-

R

M

/M -7 F C

R

-H

M

F

/M -7 CF

R

-N

F-

LF

Fig. 23. Histograms comparing the plasma membrane and cytoplasmic fractions of BCRP in folate-deprived cells and their control counterparts. After immunohistochemistry with a BCRP-specific antibody, the percentages of the plasma membrane and cytoplasmic BCRP fractions were determined in the various cell lines (A) as detailed under Materials and Methods. The cytoplasmic/plasma membrane BCRP ratio in the various cell lines is also depicted (B); note the large increase in the cytoplasmic/plasma membrane BCRP ratio in the folate-deprived cells compared with the control cells. Results depicted were obtained from three independent experiments in which a total number of 1200 cells from each cell line were processed for the determination of the cytoplasmic and plasma membrane BCRP fractions.

Retention of Plasma Membrane Localization of Various Membrane Proteins in the Short-Term Folate-Deprived Cells. To determine whether short-term folate deprivation results in a selective cytoplasmic localization of BCRP, we performed immunohistochemistry experiments with antibodies directed to various plasma membrane proteins, including EGFR as well as FGFR1, FGFR2, and FGFR3. We also undertook immunofluorescence studies with viable cells using a monoclonal antibody to an external epitope of MHC class I. Albeit these membrane proteins were expressed at variable levels, EGFR (Fig. 24A), FGFR1 (Fig. 24B), FGFR2 (Fig. 24C), FGFR3

76

(Fig. 24D), and MHC class I (Fig. 24E) retained their normal plasma localization in folate-deprived cells as in their parental cells. These results strongly suggest that the cytoplasmic localization of BCRP in the short-term folate-deprived cells is specific to BCRP because various transmembrane proteins retained their normal plasma membrane localization.

MCF-7/MR-HF-MR MCF-7/MR-HF MCF-7/MR-NF-LF

A) EGFR

B) FGFR-1

C) FGFR-2

D) FGFR-3

E) MHC Class I

Fig. 24. Immunohistochemistry and immunofluorescence localization of various plasma membrane proteins in folate-deprived cells and their parental counterparts. For immunohistochemistry studies, monolayer MCF-7/MR-HF-MR (left column), MCF7/MR-HF (middle column), and folate-deprived MCF-7/MR-NF-LF cells (right column) were fixed with 4% formaldehyde and reacted with antibodies to EGFR (A), FGFR1 (B), FGFR2 (C), and FGFR3 (D). Then, an HRP-conjugated goat anti-mouse or anti-rabbit IgG was added, and color development was carried out using the chromogen 3,3'-diaminobenzidine. Cells were then counterstained with hematoxylin and examined with a light microscope at a 200x magnification. For immunofluorescence studies, viable cells were reacted with monoclonal antibodies to

77

an external epitope of MHC class I (E) and a second FITC-conjugated goat antimouse antibody was added, and cells were analyzed with a fluorescence microscope. For experimental details, see Materials and Methods. The arrows denote the plasma membrane localization of the various membrane proteins.

Colocalization of BCRP in the ER Compartment in Folate-Deprived Cells. To better define the cytoplasmic subcellular localization of BCRP in folate-deprived cells, we used confocal microscopy after immunostaining; cells were stained either with anti-BCRP antibodies followed by a FITC-conjugated antibody (i.e., green fluorescence; Fig. 25A), or with an antibody to calnexin, an established endoplasmic reticulum resident [109, 110] followed by a Cy3-conjugated antibody (red fluorescence; Fig. 25 B). Cell nuclei were counterstained with the DNA dye DAPI (blue fluorescence; Fig. 25C). MCF-7/MR-HF-MR and MCF-7/MR-HF cells displayed an intense green fluorescence (i.e., BCRP) at the plasma membrane, particularly at cell-cell contact zones (Fig. 25A). In contrast, folate-deprived MCF7/MR-NF-LF cells had a green cytoplasmic BCRP fluorescence with no detectable plasma membrane staining (Fig. 25A). In all cell lines, the red fluorescence derived from the anti-calnexin antibodies was highly confined to the perinuclear region (Fig. 25B) as would be expected from an ER marker [109, 110]. The DAPI-stained nuclei with blue fluorescence served to localize the perinuclear ER staining (Fig. 25C). It is remarkable that merging the green BCRP fluorescence and the red calnexin fluorescence in the folate-deprived cells revealed a perfect perinuclear colocalization, as evidenced by the resultant yellow fluorescence (Fig. 25D). In contrast, merging the green BCRP fluorescence and the red calnexin fluorescence did not result in any substantial ER colocalization in MCF-7/MR-HF-MR and MCF-7/MR-HF cells. These results establish that BCRP is highly confined to the ER compartment in folatedeprived cells.

78

A.

B.

C.

D.

MCF-7/MR-HF-MR

MCF-7/MR-HF

MCF-7/MR-NF-LF BCRP (FITC)

ER (CY3)

Nuclei (DAPI)

Merge

Fig. 25. Colocalization of BCRP in the ER compartment in folate-deprived cells as revealed by confocal microscopy. Monolayer cells growing on coverslips in 24-well plates were washed, fixed, and reacted with monoclonal antibodies to BCRP (A) and calnexin (B) followed by counterstaining with the DNA dye DAPI (C). Then, FITCconjugated antibodies (A, green fluorescence representing BCRP staining) and Cy3conjugated antibodies (B, red fluorescence representing calnexin staining) were added. The merging of the green BCRP fluorescence with the red calnexin fluorescence is depicted in D. Note that the merging of the green BCRP fluorescence and the red calnexin fluorescence in folate-deprived cells revealed a perinuclear ER colocalization as evidenced by the perinuclear yellow color. In contrast, this merging experiment did not result in any substantial colocalization to the perinuclear zone in MCF-7/MR-HF-MR and MCF-7/MR-HF cells. All analyses of the fluorescent slides were performed by confocal microscopy.

Functionality of BCRP in the Various Cell Lines. To determine whether the loss of BCRP from the plasma membrane of folate-deprived cells was accompanied by a parallel loss of a plasma membrane efflux function, we measured rhodamine 123 accumulation in these cells. Rhodamine 123 was shown to be a moderate transport substrate of R482 BCRP and an excellent substrate of G482 BCRP [111]. Rhodamine 123 was also shown to be an efflux substrate of Pgp [112]; however, as shown in Fig. 21D, MCF-7/MR-HF-MR cells and their sublines were completely devoid of Pgp. 79

Cells were incubated for 1 h in the presence of 0.75 µM rhodamine 123 after which cell-associated dye was extracted and determined spectrofluorometrically. Folatedeprived cells accumulated 3-fold more rhodamine 123 compared with control cells grown in medium containing high folates (Fig. 26); this increased rhodamine 123 accumulation in folate-deprived cells was statistically significant (P = 0.017). Thus, the loss of BCRP from the plasma membrane in folate-deprived cells was accompanied by an increased cellular accumulation of rhodamine 123. Furthermore, MCF-7/MR-NF-LF cells that had a cytoplasm-to-plasma membrane BCRP distribution ratio of 6 (Fig. 23B) consistently displayed a 4.5-fold increase in [3H]folic acid accumulation (Fig. 27), compared with MCF-7/MR cells, which contained most of their BCRP in the plasma membrane (P = 0.0026). These results provide strong evidence that the cytoplasmic confinement of BCRP in the short-term folate-deprived cells serves a functional role of markedly augmenting cellular folate accumulation. Hence, to explore the possibility of whether the confinement of BCRP to the cytoplasm under conditions of folate deprivation is correlated with cell growth or the number of cells in the colony, we plotted the percentages of cytoplasmic BCRP versus the number of cells in the different colonies for each cell line (Fig. 28). In the folatedeprived cells (Fig. 28C), the percentages of cytoplasmic BCRP were significantly higher in the colonies containing high cell numbers (i.e., cell number per colony > the median cell number of the colonies in the population) than colonies containing low cell numbers (cell number per colony < the median cell number of the colonies; P = 0.013). In contrast, the MCF-7/MR-HF-MR (Fig. 28A) and MCF-7/MR-HF (Fig. 28B) cell lines failed to reveal any significant difference in the percentages of cytoplasmic BCRP when comparing the group of colonies containing high cell

80

numbers and the group of colonies containing low cell numbers (P = 0.19 and P =

30 20 10 2

HF

M

CF -7 /M

R-

RCF -7 /M M

HF -M CF -7 /M M

3

NF -L F

1

R

0 R-

(Arbitrary Units)

Rhodamine 123 Accumulation

0.76, respectively).

Fig. 26. Histogram of rhodamine 123 accumulation in folate-deprived cells and their control counterparts. Monolayer cells growing in 25-mm tissue culture flasks were incubated with 750 nM rhodamine123 for 1 h at 37°C. Cells were then washed extensively, lysed, and rhodamine 123 was extracted with PBS containing 1% Triton X-100. The fluorescence determined by a fluorescence spectrophotometer was normalized to the relative number of cells present in each culture flask. The asterisk denotes that the increased accumulation of rhodamine 123 in folate-deprived cells was statistically significant compared with parental MCF-7/MR-HF-MR cells (P = 0.017) or MCF-7/MR-HF cells (P = 0.027).

81

HF -M

CF R -7 M /M CF R-7 HF /M RNF -L F

M

1

2

3

M

CF -7 /M

R-

[3H] Folic Acid Accumulation (pmol/107cells)

5 4 3 2 1 0

Fig. 27. [3H]Folic acid accumulation in parental and short-term folate-deprived cells. Monolayer cells were washed with folate-free medium and incubated for 30 min at 37°C in HBS containing 2 µM[3H]folic acid in the presence of 1 µM trimetrexate. Transport was stopped by the addition of 10 ml of ice-cold HBS. Then, cells were detached by trypsinization, washed with ice-cold transport buffer, and the final cell pellet was suspended in 0.2 ml of water, and the radioactivity released was determined using a liquid scintillation spectrometer. The asterisk denotes statistically significant change in the MCF-7/MR-NF-LF subline compared with its parental MCF-7/MR-HF-MR counterpart (P = 0.0026).

82

100

A

80

% Cytoplasmic BCRP in the Colony

60 40 20 0 1

10

100

100

1000

B

80 60 40 20 0 1

10

100

100

1000

C

80 60 40 20 0 1

10

100

1000

Number of Cells in the Colony

Fig. 28. Histogram comparing the association between the percentage of the cytoplasmic BCRP versus the number of cells in the different colonies of folatedeprived cells (C) and their control counterparts (A and B). Note that only in the folate-deprived cells, the percentages of cytoplasmic BCRP were significantly higher in the colonies containing high cell numbers (i.e., cell number per colony > the median cell number of the colonies in the population) than colonies containing low cell numbers (cell number per colony < the median cell number of the colonies) (P = 0.013). D) Novel extracellular vesicles mediate a BCRP -dependent anticancer drug sequestration and resistance: Overexpression and immunolocalization of BCRP to extracellular vesicles in mitoxantrone-resistant

MCF-7/MR

cells:

MCF-7/MR

breast

cancer

cells

overexpress BCRP [92, 102] and consequently display 20-fold resistance to the

83

established BCRP transport substrate mitoxantrone, relative to their parental MCF-7 cells (Fig 29). Sensitivity to this anticancer drug was restored with Ko143, a potent and specific BCRP transport inhibitor [106]. Above we have shown that BCRP was highly confined to cell-cell attachment zones in monolayers of MCF-7/MR cells [102] .

Figure 29: Cellular growth inhibition with mitoxantrone: Parental MCF-7 and MCF-7/MR cells were exposed to various concentrations of mitoxantrone for 5 hr in the absence or presence of the BCRP inhibitor Ko143 (0.4 μM). After 4 days of incubation in drug-free medium the number of viable cells was determined; results depicted are the means ± S.D. of three independent experiments. Inset: Western blot analysis of BCRP expression as normalized to β-tubulin. Microscopic examination of immunohistochemical staining with anti-BCRP as well as after hematoxylin-staining of monolayers of MCF-7 (Fig 30A-C) and MCF-7/MR cells (Fig 30D-F) revealed numerous extracellular vesicle-like structures that were highly confined to cell-cell attachment zones between multiple neighbor cells. 84

Immunohistochemical analysis of MCF-7/MR cells with a monoclonal antibody to BCRP (BXP-53) revealed that BCRP was highly confined to the vesicular membrane contacting the surrounding cells (Fig 30D,E; see continuous arrows) as well as to cell-cell attachment zones (Fig 30E, see dashed arrow); no staining was observed in the absence of BXP-53 antibodies (Fig 30F). In contrast, parental MCF-7 cells which poorly express BCRP did not show any detectable staining of the vesicular membrane whether the BXP-53 antibody was present (Fig 30A,B) or absent (Fig 30C). These immunolocalization results with the BXP-53 antibody were recapitulated with additional monoclonal antibodies to BCRP including BXP-21 and BXP-34 (data not shown).

Figure 30: Immunohistochemical localization of BCRP in parental MCF-7 cells and their MCF-7/MR subline. MCF-7 (A-C) and MCF-7/MR cells (D-F) were grown in 24-well culture plates were fixed cells with formaldehyde and reacted with (A,B,D,E) or without (C and F ) the anti-BCRP monoclonal antibody BXP-53 followed by the addition of horseradish peroxidase-conjugated rabbit anti-mouse IgG as the second antibody. Color development was carried out using the chromogen 3,3'diaminobenzidine ( brown color). Cells were then counterstained with hematoxylin (violet color) and examined with a light microscope at a x200 magnification. The continuous arrows denote the extracellular vesicles (A-F). Note that BCRP precisely localizes to the membrane surrounding the extracellular vesicles (D and E, see continuous arrows) and to cell-cell attachment zones in MCF-7/MR cells (E, see dashed arrow). These immunolocalization results with BXP-53 were also

85

corroborated with additional monoclonal antibodies to BCRP including BXP-21 and BXP-34 (data not shown). Transmission electron microscopy studies corroborated the presence of multiple extracellular vesicles in MCF-7/MR cells (Fig 31A,B; see continuous arrows), particularly between multiple neighbor cells (Fig 31B; see dashed arrows). High resolution electron microscopy revealed that the vesicular membrane had a typical lipid bilayer structure (Fig 31C; see continuous arrow). Moreover, these extracellular vesicles contained microvilli-like invaginations protruding into the vesicular lumen (Fig 31C; see dashed arrows).

Figure 31: Transmission electron microscopy analysis of the extracellular vesicles in monolayers of MCF-7/MR cells. MCF-7/ MR cells were grown on glass slides until confluence was achieved. Slides containing monolayer cells were then fixed, dehydrated with ethanol, embedded in an Epon/DMP-30 resin, cut with an ultrotome and analyzed with an electron microscope as detailed in Materials and Methods. Continuous arrows denote the membrane of the extracellular vesicles (A: magnification= 4500x, and C: magnification 18,000x). Note that the dashed arrows in Panel B (magnification= 9000x) denote the plasma membrane of neighbour cells surrounding the vesicle. Furthermore, high resolution electron microscopy revealed that these vesicles contained a lipid bilayer membrane (C, see continuous arrow) with multiple microvilli-like invaginations protruding into the intravesicular lumen (C, see dashed arrow).

86

Intravesicular concentration of mitoxantrone in an ATP- and BCRP -dependent manner: We explored the mechanism underlying mitoxantrone resistance in breast cancer MCF-7/MR cells in which BCRP overexpression is highly confined to these extracellular vesicles in cell-cell attachment zones. The intense blue color of mitoxantrone rendered drug accumulation readily discernible by light microscopy. Pulse-exposure of MCF-7/MR breast cancer cells to 20 µM mitoxantrone for 6 hr resulted in a dramatic sequestration of this blue drug in the lumen of these extracellular vesicles that were confined to cell-cell attachment zones (Fig 32A). These extracellular vesicles residing in between neighbor cells refracted light; this characteristic was used to estimate the number of extracellular vesicles. The number of light-refracting extracellular vesicles per 100 cells was estimated to be 23.3 ± 2.5 and 3.2 ± 0.5 in drug-resistant MCF-7/MR and parental MCF-7 cells, respectively. Furthermore, the number of mitoxantrone-concentrating extracellular compartments was 44.1 ± 6.5 per 100 MCF-7/MR cells and none in parental cells. In agreement with the above results, immunohistochemical analysis revealed that BCRP staining formed a circumferential ring in the membrane of the extracellular vesicles (Fig 32B) thereby establishing that BCRP is highly confined to the vesicular membrane contacting the surrounding cells. In contrast, BCRP was barely detectable in the apical and basal membranes of these vesicles (Fig 32B). The intense blue color of the sequestered mitoxantrone in these extracellular vesicles allowed for the quantification of the intravesicular concentration of the drug. Based on a calibration curve of known mitoxantrone concentrations (Fig 32C), the intravesicular concentration of the drug was estimated to be as high as 12.8 ± 3.5 mM after 6 hr of incubation with 20 µM mitoxantrone and further increased in a time-dependent manner to ~20 mM after 12 hr (Fig 32D). Hence, the intravesicular concentration of mitoxantrone after 12 hr of

87

incubation with this drug was ~1,000-fold higher than in the culture medium (P=5.5x10-12). Similarly, the intravesicular concentration of mitoxantrone was also explored in flavopiridol-resistant MCF-7/FLV1000 cells with BCRP overexpression. Consistently, MCF-7/FLV1000 cells which also contained extracellular vesicles, albeit at a lower frequency than MCF-7/MR cells, displayed a robust intravesicular concentration of mitoxantrone (Fig 32E). This latter result suggests that the extracellular vesicles and their ability to sequester mitoxantrone in an BCRP -dependent manner is not limited to mitoxantrone-resistant MCF-7/MR cells.

88

E

Fig 32: Mitoxantrone accumulation in extracellular vesicles and immunohistochemical localization of BCRP in MCF-7/MR and MCF-7/FLV1000 cells: Cells were incubated in growth medium containing 20 μM mitoxantrone for 12 hr at 37°C. Then, cells were washed three times, fresh medium was added and random colonies were examined by light microscopy at a bright field mode; note that mitoxantrone accumulated in extracellular vesicles (A, blue extracellular vesicles). Cells were then processed for immunohistochemical localization of BCRP (B, brown color); colonies were examined by light microscopy after counterstaining of nuclei with haematoxylin (B, violet color). Note that BCRP precisely localizes to the membrane surrounding the extracellular vesicles that have previously accumulated 89

mitoxantrone (A). The intravesicular concentration of mitoxantrone was estimated using a mitoxantrone calibration curve (C) revealing drug concentrations of ~20 mM after 12 hr of incubation (D). Results depicted are the means ± S.D. of three independent experiments using approximately 150 extracellular vesicles in each experiment for each incubation time. The estimated intravesicular concentration of mitoxantrone at 6 and 12 hr of incubation was significantly higher than that at 3 hr (Pvalue 4.9x10-16 and 5.5x10-12, respectively). E: Mitoxantrone accumulation in extracellular vesicles in flavopiridol-resistant MCF-7/FLV1000 cells- Monolayer cells were incubated in growth medium containing 15 μM mitoxantrone for 24 hr at 37°C. Then, cells were washed three times, fresh medium was added and random colonies were examined by light microscopy at a bright field mode; note that mitoxantrone accumulated in extracellular vesicles (blue extracellular vesicles).

The intravesicular concentration of mitoxantrone in MCF-7/MR cells (Fig 33A) was prevented by the specific and potent BCRP drug efflux inhibitors Ko143 (Fig 33B) and FTC (Fig 33C) as well as by energy deprivation achieved by treatment with the respiration inhibitor sodium azide and the uncoupler FCCP (Fig 33D). To confirm that the high concentration of mitoxantrone did not impair the ability of the vesicles to concentrate this drug, we performed an experiment in which the BCRP transport inhibitor and metabolic energy inhibitors were first washed out followed by mitoxantrone re-accumulation. Hence washing out the drug efflux inhibitors followed by further incubation with mitoxantrone restored the intravesicular concentration of this drug (Fig 33F-G) at a level that was comparable to untreated cells (Fig 33E). Likewise, washing out the metabolic energy inhibitors followed by provision of the energy substrates glucose and pyruvate in the presence of mitoxantrone restored the intravesicular sequestration of the drug (Fig 33H).

90

Figure 33: Prevention of intravesicular mitoxantrone accumulation by BCRP transport inhibitors and metabolic energy deprivation: MCF-7/MR cells were incubated for 1 hr at 37°C in medium lacking (A) or containing either 0.4 µM Ko143 (B), 5 µM fumitremorgin C (C), or a combination of the metabolic energy inhibitors FCCP (5 µM) and azide (5 mM) (D). Mitoxantrone was added at 20 µM and cells were incubated for 6 additional hr at 37°C. Random colonies were then rapidly examined for the intravesicular accumulation of mitoxantrone. After ridding off the various BCRP - and metabolic energy inhibitors (F-H), fresh medium containing 20 µM mitoxantrone was added and cells that were previously incubated with FCCP and azide were supplemented with the ATP-restoring substrates sodium pyruvate (1 mM) and D-glucose (5 mM) along with mitoxantrone. After 6 hr of incubation at 37°C, the growth medium was removed, cells were washed once with medium containing 10% dialyzed fetal bovine serum. Random colonies were finally examined for the intravesicular accumulation of mitoxantrone using a light microscope at a bright field mode (E-H). Intravesicular concentration of an endogenous green fluorescent chromophore: Under normal growth in mitoxantrone-free medium, the extracellular vesicles were easily identifiable by an intense endogenous green fluorescence (Fig 34A-C). This autofluorescence was retained in phenol red-free medium, thereby excluding the possibility that this common pH indicator is the endogenous fluorescent compound. However, this intravesicular fluorescence was completely lost upon cellular growth in the presence of Ko143 (Fig 34D-F). This result indicated that BCRP mediated the intravesicular concentration of some endogenous fluorescent compound(s). Hence, BCRP mediates the intravesicular concentration of both mitoxantrone and the endogenous fluorescent compound(s). Taking advantage of this autofluorescence, the 91

structural and functional characteristics of these vesicles were explored. The accessibility of the culture medium to the extracellular vesicles was first examined; a cell-impermeable red-fluorescence TRITC-IgG conjugate was used to label the extracellular milieu of MCF-7/MR monolayers (Fig 34G,J). Whereas the extracellular vesicles were readily discernible by their endogenous green fluorescence (Fig 34H, K), confocal laser microscopy of cross-sections of monolayer MCF-7/MR cells incubated in medium containing TRITC-IgG revealed that this red fluorescence chromophore was inaccessible to the green fluorescent extracellular vesicle from the cytosol (Fig 34I). In contrast, a section perpendicular to the monolayer plane showed that the apical side of the extracellular vesicle was the only surface accessible to the TRITC-IgG-containing culture medium (Fig 34J-L). The confinement of BCRP to the vesicular membrane of this cylindrical extracellular compartment was also corroborated by confocal laser microscopy after staining with a green fluorescent labeled

antibody

to

BCRP

(Fig

34M-O).

Consistent

with

the

above

immunohistochemistry results (Fig 30D,E and Fig 32B), confocal analysis of a crosssection revealed that BCRP staining formed a circumferential ring (Fig 34M). This further confirmed that BCRP was highly confined to the vesicular membrane contacting the surrounding cells but was barely detectable in the intravesicular lumen. Consistent with the cross-section, a section perpendicular to the plane of the monolayer established the confinement of BCRP to the walls lining the extracellular vesicle (Fig 34N,O). In contrast, the apical membrane of the extracellular vesicle that faces the culture medium was devoid of BCRP (Fig 34N). These results are in accord with the immunohistochemistry findings which show that BCRP was barely detectable in the apical membrane of the extracellular vesicle. These analyses (Fig 34) allowed for the estimation of the average volume of the cylindrical extracellular

92

vesicle which was found to be 190 ± 64 fL. These results establish that BCRP which is highly confined to the membrane walls lining the extracellular vesicles mediates the ATP-driven transport of mitoxantrone from the cytosol into the intravesicular lumen of these extracellular compartments (Fig 35).

Figure 34: Detection of intravesicular green autofluorescence in viable MCF7/MR cells: MCF-7/MR monolayer cells were cultured in medium lacking (A-C) or containing 0.4 µM Ko143 (D-F). Then, random colonies were examined with a fluorescence microscope for their green autofluorescence (B and E); microscopic examination using a bright field mode revealed the intercellular localization of the extracellular vesicles (A and D, black arrows). The green autofluorescence merged perfectly with the extracellular vesicles (C), whereas this autofluorescence was absent from the extracellular vesicles in monolayer cells growing in the presence of Ko143 (E,F). The accessibility of the growth medium to the extracellular vesicles was examined by confocal microscopy (G-L). Cells grown on sterile glass coverslips were incubated in a buffer solution containing a TRITC-IgG conjugate (red fluorescence). Confocal analysis of cross-section (G-I) and perpendicular section (J-L) of the green and red fluorescence was performed with viable monolayer cells; the FITC-like mode 93

was used to detect the extracellular vesicles’ green autofluorescence (H and K), whereas the culture medium red fluorescence of the cell-impermeable TRITC-IgG conjugate was detected using a Kr/Ar laser (G and J). Merging the green fluorescence of the extracellular vesicles and the red fluorescence is shown for both the cross- (I) and perpendicular (L) section analyses. Note that the accessibility of the extracellular vesicles (green fluorescence) to the culture medium (red fluorescence) and not to the cytosol is restricted to its apical side (I,L). The confinement of BCRP to the circumferential membrane of the extracellular vesicles was revealed by crosssection (M) and perpendicular section (N) confocal microscopy after immunofluorescent staining with anti-BCRP antibodies. Cells grown on glass coverslips were fixed with methanol and reacted with a monoclonal antibody to BCRP followed by a second FITC-conjugated antibody. The BCRP green fluorescence was confined to the circumferential membrane of the extracellular vesicles upon a cross-section (M) and to the membrane walls lining the extracellular vesicles upon a perpendicular section (N). Merging the cross- and perpendicular sections is shown as well (O).

Fig 35: Novel model of extracellular vesicles that serve as cytotoxic drug disposal chambers shared by multiple neighbor cancer cells: BCRP (yellow) which is highly confined to the membrane walls lining the extracellular vesicles (green) mediates the ATP-driven transport of mitoxantrone (blue) from the cytosol (gray) into the intravesicular lumen of this extracellular compartment.

94

Discussion Influx as well as efflux transporters play a key role in folate cofactor homeostasis due to their intrinsic nature to translocate folates across cellular membranes. In mammalian cells, tetrahydrofolate cofactor transport proceeds primarily via the RFC, a high-affinity influx transporter of naturally occurring reduced folates (Km~1µM). RFC is a non-concentrative, facilitative transporter with the characteristics of a bi-directional anion exchanger that equally displays a high-affinity efflux of reduced folate cofactors. The total intracellular folate pool size of cultured tumor cells is ~11.3 µM and the intracellular concentration of the abundant 10-CHOTHF is ~7 µM [113]. Since the transport Km of the RFC for these reduced folate cofactors is in the 1µM range, this transporter functions well above its transport Km regarding the efflux component of intracellular tetrahydrofolate cofactors. Hence, in the present study we postulated that the high affinity folate cofactor efflux activity of the RFC may be detrimental to cells subjected to folate-deficiency (i.e. folate-free) conditions. The rationale behind this hypothesis was that under conditions of folatedeprivation, RFC should efficiently extrude reduced folate monoglutamates out of cells hence resulting in a significant depletion and contraction of the intracellular folate pool thereby leading to cessation of DNA replication and consequent cell death (Fig 7). We further hypothesized that this RFC-dependent efflux activity would be most prominent and thus detrimental in cells with RFC overexpression. Several lines of experimental evidence lend support to our hypothesis. First, ectopic overexpression of the hRFC after transfection into RFC null cells resulted in a dramatic decline in the proliferation rates in medium lacking folates (Fig 8). In contrast, only a modest decrease in the proliferation rates of RFC null cells was observed under conditions of relatively short-term (6 days) folate deprivation. In contrast to the RFC

95

overexpressing cells that displayed a markedly decreased proliferation rate in medium lacking folates, ectopic overexpression of the human FRα after transfection resulted in only a modest decrease in the proliferation rates under these short-term folate deprivation conditions (Fig 8). These results provide the biochemical basis for the deleterious effect that RFC-dependent folate efflux activity has on cellular proliferation. Owing to the catalytic activity of FPGS, a predominant fraction of the pool of intracellular folates is present as long-chain polyglutamates [103, 114]. In contrast, under conditions of cellular exposure to a folate-free medium, the small fraction of intracellular folate monoglutamates could be efficiently extruded via the high-affinity efflux activity of the RFC. Hence, given the lack of extracellular folates in the extracellular medium under conditions of folate deprivation along with the high affinity RFC-dependent efflux of folate monoglutamates as well as based on the Le Chatelier principle, one could predict a continuous conversion of intracellular folate polyglutamates to monoglutamate congeners via lysosomal GGH activity (Fig 7). This should result in the continuous high affinity efflux of folate monoglutamates via RFC until the intracellular pool of folates becomes minimal and well below the transport Km of the RFC. This would consequently result in a severe depletion of the intracellular folate pool thereby leading to cessation of DNA replication and cell death (Fig 7B). One of the novel findings in the present research relates to the significant down-regulation of both RFC and GGH gene expression as a cellular adaptiveprotective response to folate deficiency conditions. Thus, MCF-7/MR cells with physiological expression of the RFC (i.e. relatively low levels) displayed a marked down-regulation of RFC and GGH, upon exposure to folate-free medium and survived for relatively long periods under such conditions (Fig. 9). The decreased RFC mRNA levels of the folate-deprived cell lines were associated with a consistent

96

significantly decreased RFC activity (Fig. 10). We therefore propose here that under conditions of folate deficiency, RFC and GGH may undergo an adaptive downregulation of gene expression and a consistent decline in their functional activities thereby presumably serving as a protective mechanism aimed at counteracting the otherwise detrimental conversion of folate poly- to monoglutamates and their high affinity efflux via the RFC. The fact that neither of the low affinity yet high capacity ATP-dependent folate exporters including MRP1, MRP5 and BCRP underwent any down-regulation under these folate-deplete conditions suggests a key detrimental role for RFC-dependent folate efflux activity but not for these ATP-driven efflux transporters. Hence, it appears that as long as reduced folates are present in the medium at a minimal yet sufficient level, RFC activity may not be down-regulated (Fig 7A). However, upon the complete lack of folates in the growth medium (Fig 7B), the apparent protective cellular response of the cells is to down-regulate both RFC and GGH. Further studies are necessary to pin-point the putative RFC and GGH promoter elements that may respond to folate deprivation and thereby result in a marked repression of gene expression. One important emerging question from the current study is what physio-pathological conditions and syndromes could match the transient folate-deprivation conditions used in the present paper. The first pathological syndrome pertaining to such folate deprive status includes a recent discovery reported on the molecular identification of the genetic lesion responsible for hereditary folate malabsorption (HFM) [23]. The mutant transporter gene encodes for a proton-coupled folate transporter (PCFT/HCP-1/SLC46A) responsible for the high affinity intestinal influx of naturally occurring folates. In this recent paper it was found that a single nucleotide inactivating mutation in the consensus splice acceptor site of intron 2 (i.e. intron 2/exon 3 boundary) led to exon 3 skipping. This consequently resulted in an in-

97

frame deletion of 28 amino acids thereby leading to intracellular trapping of the truncated protein in the cytoplasm and loss of function due to the lack of plasma membrane localization. The loss of intestinal folate transport resulted in severely low folate levels (≤ 0.2 nM) in the blood and cerebral spinal fluid. Hence, under such pathological conditions of severe folate deprivation, RFC and GGH may possibly undergo a significant down-regulation of gene expression and activity in order to protect cells from further loss of intracellular folates due to folate efflux via RFC. Second, mammals may occasionally experience transient or prolonged states of starvation and/or B-complex vitamin deficiency. These would also lead to a major folate (B9 vitamin) deprivation and presumably the above protective-adaptive response aimed at preserving the precious intracellular folate pool. It is possible that the ability to down-regulate RFC and GGH gene expression under states of severe folate deprivation stems from evolutionary roots originating in unicellular and perhaps metazoic ancestral organisms undergoing transient yet frequent states of starvation and folate deprivation. Whereas

RFC as well as MRP1 through MRP4 export folate and MTX

monoglutamates, only BCRP has the facility to extrude both mono-, di-, and triglutamate conjugates of folic acid and MTX [42, 43] . We hence undertook the above study in order to explore the possible role of BCRP in folate homeostasis in cells with a ubiquitous expression of MRP1. To this end, we first examined the relationship between folate status and BCRP expression. We therefore performed a gradual folic acid deprivation by using two sets of breast cancer cell lines as follows: parental MCF-7 cells with low BCRP expression and moderate MRP1 levels, as well as their Mitoxantrone-resistant MCF-7/MR subline with high levels of BCRP but low levels of MRP1. Several lines of evidence established that folate deprivation resulted

98

in a dramatic down-regulation of BCRP expression and efflux function. (a) Semiquantitative RT-PCR and Western blot analyses revealed only residual BCRP gene expression and protein levels, respectively, in both LF-adapted cell lines (Fig 11). (b) Immunohistochemistry studies demonstrated that MCF-7/MR cells had an intense plasma membrane staining and a marked intracellular staining, whereas their LFadapted subline MCF-7/MR-LF was devoid of plasma membrane and cytoplasmic staining (Fig 12). (c) Although MCF-7/MR-HF and MCF-7 cells displayed a high and low efflux of the BCRP substrate Hoechst 33342, respectively, only residual chromophore efflux was obtained with the LF adapted sublines (Fig 13). (d) Consistently, the LF-adapted cell lines exhibited a significant increase in the net accumulation of Mitoxantrone, relative to their parental cells (Fig 14). (e) Folate deprivation resulted in the loss of resistance to the established BCRP substrates Mitoxantrone and MTX in MCF-7/MR-LF cells as well as in a markedly increased sensitivity of MCF-7/LF cells to these drugs (Fig 15, Fig 16). Furthermore, the extent of the increased sensitivity to Mitoxantrone in the LF-adapted cell lines was comparable with that achieved with MCF-7 and MCF-7/MR cells that were treated with Ko143, a potent and specific BCRP inhibitor [106]. Based on these cumulative data, we conclude that down-regulation of BCRP expression and efflux activity is an essential component of cellular survival under conditions of folate limitation (i.e. not folate-free conditions). These results support the hypothesis that apart from MRP1, BCRP is an important component of folate homeostasis, particularly under conditions of folate deprivation. We note that loss of BCRP expression and function in MCF7/MR-LF cells was not the sole adaptive response to folate deprivation. When compared with parental MCF-7 cells, MCF-7/MR-LF cells showed a 14-fold decrease in MRP1 levels. Moreover, the latter cells also displayed a 63% increase in FPGS

99

activity relative to their parental MCF-7/MR-HF counterpart (Fig 18). Thus, MCF-7 cells growing in an excess of folic acid in the growth medium (i.e. 2.3 µM) initially had a substantial capacity to actively export folates via MRP1 and BCRP but only poorly via MRP4. In contrast, MCF-7/MR-LF cells growing in a ~770-fold less folic acid in the growth medium essentially lost this ability to export folates via MRP1 and BCRP. It should be emphasized that whereas MRP2, MRP3, and MRP4 have the facility to extrude folates, these transporters were essentially undetectable or poorly expressed in both parental cells and the LF-adapted cell lines. The increased activity of FPGS along with the loss of BCRP and the markedly decreased MRP1 levels under conditions of folate limitation were presumably crucial adaptations aimed at augmenting cellular folate retention. These findings are in agreement with several in vivo and in vitro studies that explored the effect of folate deficiency on cellular FPGS activity and the expression of various MRPs: (a) Mice fed a low folate diet displayed a 50% increase in liver FPGS activity [115]. (b) In a recent study [116] we characterized CHO cells that were subjected to gradually increasing concentrations of the lipid-soluble antifolate, pyrimethamine [108], thereby resulting in a gradually increasing folate deprivation [107]. Consequently, these cells displayed a 3–4-fold increase in FPGS activity [107] along with a complete loss of MRP1 expression [116]. Furthermore, these cells were also devoid of BCRP and MRP2 through MRP4 even before pyrimethamine selection took place. Hence, in the absence of an ABC transporter that would mediate folate efflux activity, these cells could grow on extremely low concentrations of folates (e.g. pM concentrations of leucovorin). (c) Recently we have shown [47] that human leukemia cells adapted to grow under extremely low concentrations of leucovorin had a 95% loss of MRP1 expression and folate efflux function along with a 100-fold overexpression of the RFC, the primary

100

folate influx transporter. In conclusion, disruption of folate exporter function via loss of BCRP and/or MRP1 expressions along with a concomitant increase in FPGS activity are apparently essential adaptations to conditions of folate deficiency, thereby resulting in an increased capacity to accumulate and retain cellular folates. We consistently find here that MCF-7/MR-LF and MCF-7/LF cells accumulated significantly higher levels of [3H]folic acid than their parental counterparts (Fig 17). As mentioned above, we have shown recently [47] that gradual deprivation of leucovorin (5-formyltetrahydrofolate) from the growth medium of human CCRFCEM leukemia cells resulted in a 95% loss of MRP1 expression. Consistently, replenishment of the latter cells with 5 nM leucovorin resulted in a complete restoration of MRP1 expression. In a recent paper [117] we reported that under folatefree conditions, MRP1- and MRP3-overexpressing cell lines were impaired in cellular growth upon a short exposure (4 h) to folic acid or leucovorin, when compared with their parental cells. Furthermore, the folic acid growth stimulation capacity in these cells was dramatically decreased during the pulse exposure to folic acid, when metabolism into rapidly polyglutamatable and hence retainable dihydrofolate and tetrahydrofolate was blocked by trimetrexate, a dihydrofolate reductase inhibitor. In another study [116], as mentioned above, we subjected Chinese hamster ovary cells to gradually increasing concentrations of the lipid soluble antifolate pyrimethamine [108]. This gradually increasing folate deprivation [107] resulted in a complete loss of MRP1 expression [116]. Taken together, these findings suggest that down- and upregulation of the ubiquitously expressed MRP1 can readily influence cellular folate homeostasis, particularly when cellular folate retention by polyglutamylation is attenuated. Based on the unique substrate specificity of BCRP that is capable of exporting mono-, di-, and triglutamate conjugates of folates, one could predict that

101

expression of substantial BCRP levels would not be compatible with folate deficiency conditions. Indeed, although MCF-7 cells expressed 5-fold more MRP1 than BCRP levels, the LF-adapted subline MCF-7/LF almost completely lost BCRP expression with no change in MRP1 levels (Fig 11). In contrast, following folate deprivation, MCF- 7/MR cells with 55-fold BCRP overexpression, relative to parental MCF-7 cells, had a near complete loss of BCRP and MRP1 expression (Fig 11). Hence, elimination of the low expression levels of BCRP in MCF-7/LF cells was apparently sufficient to meet their folate growth requirement. In contrast, the dramatic downregulation of the initially very high levels of BCRP in MCF-7/MR cells was crucial but apparently not sufficient as MRP1 expression was down-regulated to barely detectable levels. It is possible that this repression in MRP1 expression was achieved in the following manner: when BCRP was decreased to levels that were comparable with those of MRP1, the latter became significant in its contribution to folate efflux, thereby promoting its down-regulation as well. Support for this hypothesis could derive from the fact that replenishment of MCF-7/ MR-LF cells with 2.3 µM folic acid for 1 month resulted in restoration of MRP1 expression to levels that exceeded those of parental MCF-7 cells (data not shown). In contrast, no restoration of BCRP expression was observed in these cells. Clearly, resumption of substantial levels of BCRP, an exporter that extrudes the precious triglutamate conjugates of folates, was not consistent with the retention of sufficient cellular folate pools to support cell growth. Both MRP1 and BCRP are capable of ATP-driven efflux of folate monoglutamates. However, whereas MRP1 is ubiquitously expressed at substantial levels in various tissues, BCRP is relatively poorly expressed, and its pattern of tissue expression is apparently restricted to only a few tissues including placenta, intestine, colon, and the bile canaliculus. Most interesting, all these tissues were reported to

102

express high levels of FPGS mRNA [93] and displayed a relatively high activity of FPGS. Hence, based on the unique folate polyglutamate exporter function of BCRP as well as on the ubiquitous expression of MRP1, it is tempting to speculate that expression of substantial levels of BCRP should be accompanied by adequate levels of FPGS activity in order to ensure sufficient intracellular retention of long chain (>3 glutamate residues) folate polyglutamates. The present finding of the loss of BCRP expression under conditions of folate deficiency may have potentially important implications

for anticancer

chemotherapy

including

Mitoxantrone-containing

chemotherapy. First, Mitoxantrone is being used in the treatment of metastatic breast cancer and non-lymphocytic leukemia including acute granulocytic leukemia. Breast cancer patients are initially treated with a chemotherapeutic regimen that contains either MTX or doxorubicin. However, one mechanism of MTX resistance in breast cancer cells may be antifolate drug transport [118, 119]. This may be because of the loss of gene expression and function of RFC, the primary transporter for folates and MTX. Consequently, such folate transport-deficient and MTX-resistant cells may suffer from a markedly diminished intracellular folate pool [120].Cell survival on such a shrunken cellular folate pool may be associated with a significant downregulation of MRP1 and/or BCRP [47, 116]. Hence, such MTX-resistant breast cancer cells may be most vulnerable to the cytotoxic action of various chemotherapeutic agents including anthracyclines and Mitoxantrone. This consideration may possibly have therapeutic value in overcoming drug resistance of certain tumors that have down-regulated the expression of BCRP under folate deficiency conditions. Second, we have shown here that folate deprivation results in the loss of MRP1 and BCRP expression (Fig 11). As such, one potential strategy to overcome anticancer drug resistance that is based on MRP1 and/or BCRP overexpression could be a transient

103

exposure of the tumors to a folate-deficient diet. This could lead to a marked downregulation of MRP1 and/or BCRP, thereby resulting in increased sensitivity of the tumors to various chemotherapeutic agents. We have previously shown that long-term gradual deprivation (three months) of folic acid from the growth medium of breast cancer cells with BCRP overexpression results in almost a complete loss of BCRP expression along with a marked decrease in MRP1 levels [92], (Fig 11). Our next aim was to study the impact of short-term (two weeks) folic acid deprivation (Fig. 19) on BCRP expression, subcellular localization and efflux function. The rational behind these experiments was that as BCRP has the facility to export mono-, di-, and triglutamates of folates [42, 43], the localization of an overexpressed BCRP at the plasma membrane should not be retained under conditions of folate deprivation. The following line of evidence confirms that the plasma membrane localization of BCRP has been lost in breast cancer cells subjected to a short-term folate deprivation. First, western blot, immunohistochemistry and immunofluorescence analyses revealed that although high levels of BCRP were retained (Fig. 21, Fig 22), this transporter was confined to the cytoplasm rather than to the plasma membrane (Fig. 22). Second, confocal microscopy after immunofluorescent staining with antibodies to BCRP as well as to calnexin, an established marker of the ER [109, 110], showed that BCRP was largely confined to the ER compartment in folate-deprived cells (Fig. 25). This was inferred from BCRP colocalization with the ER-resident calnexin (Fig. 25). The latter is a lectin chaperone that functions in the quality control system in the ER [109] . Third, folate-deprived cells with a cytoplasm-to-plasma membrane BCRP distribution ratio of 6 (Fig. 23) displayed a consistent increase (4.5-fold) in [3H]folic acid accumulation, compared with their parental cells that contained most of their BCRP in the plasma

104

membrane (Fig. 27). Fourth, loss of BCRP from the plasma membrane was accompanied by a prominent increase in the cellular accumulation of rhodamine 123, a moderate R482 BCRP substrate [111] (Fig. 26). Because MCF-7/MR-HF-MR cells were devoid of Pgp, which also exports rhodamine 123 [112], it was likely that the lack of sorting of BCRP to the plasma membrane would result in increased rhodamine 123 accumulation in these folate-deprived cells. These data suggest that short-term folic acid deprivation presumably selects for the lack of plasma membrane targeting of BCRP, thereby resulting in the cytoplasmic confinement of this ABC transporter. Overexpressed BCRP with a cytoplasmic residence rather than the normal plasma membrane localization [76] is a useful strategy aimed at eliminating the BCRPdependent efflux of intracellular mono and poly glutamates of folates. This would result in the preservation of the precious cellular folate pools that are particularly shrunken under conditions of folate deficiency [47, 52, 120] . Indeed, as mentioned above, short-term folate-deprived cells had a drastic increase in the accumulation of [3H]folic acid relative to their parental counterpart (Fig. 27). Together, our previous study [92] demonstrates that long-term gradual folate deprivation results in the near complete loss of BCRP expression and a marked decrease in MRP1 levels, whereas our current findings show that short-term folate deprivation leads to lack of plasma membrane targeting of BCRP and its cytoplasmic confinement, along with a moderate decrease in BCRP and MRP1 levels (Fig. 21). We conclude that cytoplasmic confinement and decreased expression of BCRP are important components of cellular adaptation to short-term folate deprivation. The present immunohistochemistry and immunofluorescence data suggest that folate deprivation resulted in a cytoplasmic confinement of BCRP that seemed to be selective for this transmembrane protein. This is based upon the finding that various membrane proteins that are expressed at

105

variable levels in breast cancer cells, including EGFR, FGFR1, EGFR2, EGFR3, and MHC class I, retained their dominant plasma membrane localization under conditions of folate deprivation (Fig. 24). As we show here that the cytoplasmic confinement of BCRP is not shared with various plasma membrane proteins, the phenomenon of plasma membrane confinement cannot be regarded as a pleiotropic effect of folate deprivation such that it would encompass various transmembrane proteins. Hence, these results suggest that the selective confinement of BCRP to the cytoplasmic compartment plays a contributing role in the cellular adaptation to conditions of folate deprivation. Several possibilities exist that can provide a potential molecular basis for the novel finding of the cytoplasmic confinement of BCRP upon short-term folate deprivation. The first involves a recent article [104] that reported on the rapid translocation of BCRP from the plasma membrane to the cytoplasmic compartment in hematopoietic stem cells known as side population (SP); these cells are defined by the efflux of Hoechst 33342, an established BCRP substrate. In this study, it was shown that a brief treatment (1.5 h) of freshly derived mouse bone marrow cells with LY294002, an inhibitor of the Akt effector protein phosphatidylinositol-3-kinase (PI3K), resulted in the translocation of BCRP from the plasma membrane to the cytoplasmic compartment. The authors therefore suggested that the PI3K-Akt signaling axis is an important regulator of BCRP expression and subcellular localization as well as of the bone marrow-derived SP stem cell phenotype. Thus, it is possible that the confinement of BCRP to the cytoplasmic compartment in our shortterm folate-deprived breast cancer cells may be a result of the loss of activity of a component in the PI3K-Akt signaling pathway. However, it should be noted that in the current study, 1.5-h treatment of MCF-7/MR cells with LY294002 did not result in a rapid translocation of BCRP from the plasma membrane to the cytoplasmic

106

compartment. Thus, one cannot exclude the possibility of an ongoing effect of this inhibitor on the cytoplasmic confinement of BCRP in breast cancer cells with BCRP overexpression. In support of this hypothesis, it has been shown that the PI3K-Akt signaling pathway controls the cellular localization of a number of proteins, including GLUT4, an insulin-stimulated glucose transporter; this cytoplasmic localization involved the cycling of GLUT4 between the plasma membrane and specialized intracellular vesicles . In this regard, it has been shown recently that lung SP progenitor cells express BCRP on their surface, whereas muscle SP cells express intracellular BCRP and are therefore incapable of Hoechst 33342 efflux [121].Furthermore, phosphoinositol 3,4-biphosphate enhanced the ATP-dependent transport of taurocholate in canalicular membrane vesicles in vitro an in vivo [122]. Hence, it seems that the PI3K-Akt signaling pathway can alter the subcellular localization of membrane transporters, and, through its lipid products, it can directly modulate the transport activity of membrane transporters, including the Mdr1 and Mdr2 gene products. The second possibility involves a recent article in which the impact of BCRP mutations and single amino acid polymorphisms on its localization, ATPase activity, and efflux function was explored [123]. It was found that an Nterminal BCRP mutation (Val12Met) disrupted the apical plasma membrane localization of BCRP in polarized LLC-PK1 cells. The third possibility involves the use of a BCRP that was tagged with a cyan green fluorescent protein and then transiently expressed in HeLa cells [124]. In this study, it was found that BCRP colocalized to a perinuclear compartment that was positive for lysosomal markers including cathepsin D and synaptotagmin VII. The authors therefore suggested that BCRP can display a variable subcellular localization other than the plasma membrane in different cell lines and under different conditions. Consistent with our current

107

findings, these results establish that BCRP has the inherent capability to be localized at certain cytoplasmic compartments, including ER and lysosomes. The cytoplasmic localization of BCRP in folate-deprived cells has potentially important implications for combination chemotherapy. BCRP has been shown to confer resistance to various anticancer drugs, including doxorubicin, mitoxantrone, topotecan, and the antifolate methotrexate [111, 125]. Hence, chemotherapeutic regimens containing some of these BCRP efflux substrates, including the CAF and CMF protocols, which contain cyclophosphamide, Adriamycin (i.e., doxorubicin), 5-fluorouracil, and methotrexate for the treatment of breast cancer, may become limited in their efficacy if BCRP is overexpressed in these malignant cells [34, 56, 125-128] . As such, one potential strategy to overcome BCRP-dependent drug resistance may be the use of a combined treatment of cancer cells with trimetrexate, a lipid-soluble analog of methotrexate that is not recognized by BCRP as an efflux substrate (A. Shafran and Y. G. Assaraf, unpublished data) along with conventional chemotherapeutic drugs, including cyclophosphamide and 5-fluorouracil. This antifolate treatment should result in an intracellular folate depletion, thereby resulting in the possible confinement of BCRP to the cytoplasmic compartment aimed at preserving cellular folates. The resultant breast cancer cells should be vulnerable and could be easily eradicated even with chemotherapeutic drugs that are BCRP substrates (e.g., mitoxantrone and topotecan), because no plasma membrane BCRP efflux would be operable. An alternative approach would be pulse treatment of BCRP-overexpressing cancer cells with a PI3K inhibitor such as LY294002 [104]. This could possibly lead to the confinement of BCRP to the cytoplasmic compartment, thereby achieving reversal of the MDR phenotype as would be obtained with specific BCRP efflux inhibitors such as Ko143 [106]. It is clear that these potential strategies to overcome BCRP-dependent drug

108

resistance must await further studies to explore their feasibility and potential applicability. In the present article, we note that the cytoplasmic confinement of BCRP was much more pronounced (P = 0.013) in large colonies (i.e., colony number greater than the median) than in small colonies (i.e., colony number lower than the median) of folate-deprived cells (Fig 28). Hence, it is very likely that during the shortterm folate deprivation, clonal subpopulations with a dominant cytoplasmic BCRP localization may have gained a significant growth advantage over subpopulations with high plasma membrane fraction but low cytoplasmic confinement. This presumed growth advantage is based upon the fact that the loss of BCRP from the plasma membrane would lead to a parallel loss of efflux of cellular folates as evidenced by the drastic increase in [3H]folic acid accumulation in both short-term (the present study) and long-term folate-deprived cells [92]. Therefore, cells within colonies that display a dominant cytoplasmic BCRP localization could better preserve their intracellular folate pools, thereby leading to a growth advantage as reflected in the increased number of cells per colony. During these BCRP cellular confinement studies, we discovered that this transporter is highly confined to cell-cell attachment zones in the MCF-7/ MX resistant breast cancer subline MCF-7/MR in which wild type (R482) BCRP is overexpressed. The cell-cell attachment zones were found to be a membrane of novel extracellular vesicles in which mitoxantrone was rapidly and dramatically sequestered into. Several lines of evidence establish that the intravesicular concentration of mitoxantrone is mediated by BCRP. First, inhibition of BCRP transport activity by Ko143 and FTC prevented the intravesicular concentration of mitoxantrone (Fig 33). Furthermore, washing out these drug transport inhibitors resulted in restoration of mitoxantrone compartmentalization (Fig 33). These results are in accord with the finding that Ko143 induced a near complete reversal of

109

mitoxantrone resistance in MCF-7/MR cells (Fig 29). Second, depletion of cellular ATP pools by the respiration inhibitor sodium azide [129] and the uncoupler FCCP [130] prevented the intravesicular concentration of mitoxantrone (Fig 33). Consistently, removal of metabolic energy inhibitors followed by restoration of cellular energy resources by provision of glucose and pyruvate in the presence of mitoxantrone resulted in the resumption of intravesicular drug concentration (Fig 33). These findings are in agreement with the tight coupling of BCRP drug transport to ATP hydrolysis and consequent intravesicular drug concentration. The concentration of mitoxantrone in these extracellular vesicles by BCRP was energy-, time-dependent and reached a ~20 mM concentration after 12 hr of incubation with 20 µM drug (Fig 32). This 1,000-fold concentrative ability of BCRP gained further support by the dramatic intravesicular sequestration of a green fluorescent compound(s) (Fig 34AC). Whereas this chromophore(s) was not fluorescently visible in the cytosol of neighbor cells surrounding the extracellular vesicle, it was intensely fluorescent in the lumen of this compartment in MCF-7/MR cells with BCRP overexpression. This intravesicular fluorescence was completely absent after 4 days of cellular growth in the presence of the specific BCRP transport inhibitor Ko143 (Fig 34D-F). Furthermore, the presence of the light-refracting extracellular vesicles in parental MCF-7 cells was less frequent. When present, these extracellular vesicles in parental MCF-7 cells were completely devoid of the green autofluorescence that is characteristic of the vesicles in drug-resistant MCF-7/MR cells which overexpress the wild type R482 BCRP. This lack of intravesicular autofluorescence in drug-sensitive cells that poorly express BCRP is consistent with the finding that the intravesicular concentration of the green fluorescent compound(s) in MCF-7/MR cells is mediated by BCRP. Hence, the energy-, time-dependence and the ~1,000-fold concentrative

110

capacity of mitoxantrone and that of the autofluorescent compound(s) by BCRP is consistent with the highly concentrative transport of various ABC transporters. First, lysosomal and vacuolar membranes contain V-type ATP-driven proton pumps that maintain a >100-fold proton gradient across the acidic lumen of the lysosome (pH ~ 4.5-5) and the neutral cytosol (~ pH 7.0) [131] . Second, since a rise in the concentration of Ca2+ ions in the cytosol of mammalian cells (e.g. erythrocytes) is an important regulatory signal, the plasma membrane P-class Ca2+ ATPase efficiently transports Ca2+ out of the cell; consequently, the extracellular (i.e. blood) concentration of Ca2+ is as high as 3,600-fold (1.8 mM) than in the cytosol of the erythrocyte (0.5 µM) [132]. Similarly, Ca2+ ATPase from the sarcoplasmic reticulum of muscle cells efficiently pumps Ca2+ ions from the cytosol (0.1-1 µM) into the lumen of the sarcoplasmic reticulum (10 mM), thereby resulting in at least 10,000fold concentrative transport [133]. The third example concerns the H+,K+ ATPase present in the plasma membrane of acid-secreting parietal gastric cells. This P-type H+,K+ ATPase maintains an extremely acidic pH in the gastric fluid, whereas the cytosolic pH of these cells is neutral (pH 7.0). Thus, this H+,K+ ATPase concentrates protons by a factor of 100,000 [134]. Upon cross-section confocal microscopy experiments with a cell impermeable TRITC-IgG conjugate there was no accessibility of the culture medium containing this red chromophore to the extracellular vesicles (Fig 34G-I). Whereas, a section that is perpendicular to the plane of the monolayer revealed that the only contact of these vesicles with the fluorochrome-labeled medium was from the apical side of this extracellular compartment (Fig 34J-L). Furthermore, confocal microscopy and immunohistochemistry revealed that BCRP was primarily localized at the circumference of the extracellular vesicle but was absent from its apical side that faces the culture medium; BCRP was therefore localized at the walls

111

lining this vesicle but was absent from its apical side. (Fig 32, Fig 34M-O) Thus, the ATP-binding and the substrate-binding site of BCRP must face the cytoplasm of the cells surrounding this extracellular vesicle (Fig 35). As such, BCRP presumably extracts mitoxantrone from the cytosol of the surrounding cells and highly concentrates it in the lumen of these extracellular vesicles (Fig 35). Fine structure studies corroborated the presence of numerous large extracellular vesicles emerging from cell-cell attachment zones (Fig 31A-B). Furthermore, high resolution electron microscopy revealed that these vesicles contained a lipid bilayer membrane with multiple microvilli-like invaginations protruding into the intravesicular lumen (Fig 31C). Hence these fine structure projections are reminiscent of the microvilli invaginations of both the gastrointestinal mucosa and the placental epithelium. Given the ATP-driven BCRP -dependent trans-vesicular transport of mitoxantrone into the intravesicular lumen, it is likely that these microvilli-like invaginations increase the vesicular membrane surface thereby allowing for a more efficient intravesicular drug concentration. Similarly, the surface of the syncytial trophoblast of the human placenta is covered by a microvillus (i.e. brush) border that is in direct contact with maternal blood; this location is the site of a variety of transport and receptor activities. For example, endocytosis of gold-labeled LDL into primary human placental cells involved uncoated plasmalemmal invaginations which subsequently became clathrin uncoated endosome vesicles [135]. The encouraging results of the current study with anticancer drug-resistant breast cancer cell lines warrant further clinical evaluation of the presence of such drug-sequestering extracellular vesicles in tumor-derived specimens. The possible future finding of such extracellular vesicles which could efficiently concentrate anticancer drugs may have potential implications for cancer chemotherapy. First, inclusion of specific, potent and non-toxic BCRP transport

112

inhibitors such as Ko143 [106] and GF120918 [136] which reverse anticancer drug resistance,

may

potentially

prove

effective

in

the

prevention

of

drug

compartmentalization in tumors, thereby resulting in reversal of drug-resistance. Moreover, various approved cytotoxic drugs were recently found to be efficient inhibitors of BCRP efflux activity including Iressa (ZD1839, Gefitinib), a selective epidermal growth factor receptor tyrosine kinase inhibitor [137, 138] Imatinib mesylate (Gleevec, STI571), a tyrosine kinase inhibitor, selective for Bcr-Abl [139], and CI-1033, a HER tyrosine kinase inhibitor [140]. In addition, phytoestrogens and flavonoids were also shown to efficiently reverse drug resistance mediated by BCRP [141]. Clearly, these BCRP efflux inhibitors may prove effective reversal agents of drug resistance mediated by BCRP overexpression including when the latter is highly confined to the vesicular membrane. Second, compounds which may interfere with the formation of these novel extracellular vesicles and/or with the sorting of BCRP to the vesicular membrane should render cells sensitive to anticancer drugs like mitoxantrone. For example, a recent paper [104] reported on the rapid translocation of BCRP from the plasma membrane to the cytoplasmic compartment (ER-Golgi) in freshly derived hematopoietic stem cells known as side population (SP); in this study it was shown that a brief treatment (1.5 hrs) of freshly derived mouse bone marrow cells with LY294002, an inhibitor of the Akt effector protein phosphatidylinositol-3kinase (PI3K), resulted in the rapid translocation of BCRP from the plasma membrane to the cytoplasmic compartment. The authors therefore suggested that the PI3K-Akt signaling axis is an important regulator of BCRP expression and subcellular localization of the bone marrow-derived SP stem cell phenotype. Another example involves a recent study from our laboratory demonstrating that short-term deprivation of folic acid from the growth medium of BCRP -overexepressing MCF-7/MR cells

113

resulted in the cytoplasmic retention of this MDR efflux transporter BCRP [102]. Hence it is possible that such agents and treatment strategies which block protein sorting of BCRP from the cytoplasmic compartment to the plasma membrane and vesicular membrane may be used to reverse anticancer drug resistance mediated by the BCRP -rich extracellular vesicles.

114

References 1. Blakely, R.D., and Benkovics, S. J., Chemistry and Biochemistry of Pterins, in Folates and Pterins, J.W. Sons, Editor. 1985: New York. 2. Blakely, R.D., and Whitehead, V. M, Nutritional, Pharmacological, and Physiological Aspects, in Folates and Pterins, J. Wiley and Sons, Editors. 1986: New York. 3. Jukes, T.H., Searching for magic bullets: early approaches to chemotherapyantifolates, methotrexate--the Bruce F. Cain memorial award lecture. Cancer Res, 1987. 47(21): p. 5528-36. 4. Bertino, J.R., Karnofsky memorial lecture. Ode to methotrexate. J Clin Oncol, 1993. 11(1): p. 5-14. 5. Ozaki, Y., R.W. King, and P.R. Carey, Methotrexate and folate binding to dihydrofolate reductase. Separate characterization of the pteridine and paminobenzoyl binding sites by resonance Raman spectroscopy. Biochemistry, 1981. 20(11): p. 3219-25. 6. Jackson, R.C. and K.R. Harrap, Studies with a mathematical model of folate metabolism. Arch Biochem Biophys, 1973. 158(2): p. 827-41. 7. Richards, R.G., et al., Drug concentration-dependent DNA lesions are induced by the lipid-soluble antifolate, piritrexim (BW301U). Mol Pharmacol, 1986. 30(6): p. 651-8. 8. Schornagel, J.H. and J.G. McVie, The clinical pharmacology of methotrexate. Cancer Treat Rev, 1983. 10(1): p. 53-75. 9. Henderson, G.B. and E.M. Zevely, Anion exchange mechanism for transport of methotrexate in L1210 cells. Biochem Biophys Res Commun, 1981. 99(1): p. 163-9. 10. Goldman, I.D., The characteristics of the membrane transport of amethopterin and the naturally occurring folates. Ann N Y Acad Sci, 1971. 186: p. 400-22. 11. Sirotnak, F.M., Obligate genetic expression in tumor cells of a fetal membrane property mediating "folate" transport: biological significance and implications for improved therapy of human cancer. Cancer Res, 1985. 45(9): p. 3992-4000. 12. Sirotnak, F.M., Determinants of resistance to antifolates: biochemical phenotypes, their frequency of occurrence and circumvention. NCI Monogr, 1987(5): p. 27-35. 13. Ferguson, P.L. and W.F. Flintoff, Topological and functional analysis of the human reduced folate carrier by hemagglutinin epitope insertion. J Biol Chem, 1999. 274(23): p. 16269-78. 14. Drori, S., et al., Characterization of a human alternatively spliced truncated reduced folate carrier increasing folate accumulation in parental leukemia cells. Eur J Biochem, 2000. 267(3): p. 690-702. 15. Westerhof, G.R., et al., Carrier- and receptor-mediated transport of folate antagonists targeting folate-dependent enzymes: correlates of molecularstructure and biological activity. Mol Pharmacol, 1995. 48(3): p. 459-71. 16. Wang, X., et al., Differential stereospecificities and affinities of folate receptor isoforms for folate compounds and antifolates. Biochem Pharmacol, 1992. 44(9): p. 1898-901.

115

17.

Antony, A.C., The biological chemistry of folate receptors. Blood, 1992. 79(11): p. 2807-20. 18. Brigle, K.E., et al., Increased expression and characterization of two distinct folate binding proteins in murine erythroleukemia cells. Biochem Pharmacol, 1994. 47(2): p. 337-45. 19. Henderson, G.B. and B.P. Strauss, Characteristics of a novel transport system for folate compounds in wild-type and methotrexate-resistant L1210 cells. Cancer Res, 1990. 50(6): p. 1709-14. 20. Sierra, E.E., et al., pH dependence of methotrexate transport by the reduced folate carrier and the folate receptor in L1210 leukemia cells. Further evidence for a third route mediated at low pH. Biochem Pharmacol, 1997. 53(2): p. 223-31. 21. Kumar, C.K., et al., A protein-tyrosine kinase-regulated, pH-dependent, carrier-mediated uptake system for folate in human normal colonic epithelial cell line NCM460. J Biol Chem, 1997. 272(10): p. 6226-31. 22. Assaraf, Y.G., S. Babani, and I.D. Goldman, Increased activity of a novel low pH folate transporter associated with lipophilic antifolate resistance in chinese hamster ovary cells. J Biol Chem, 1998. 273(14): p. 8106-11. 23. Qiu, A., et al., Identification of an intestinal folate transporter and the molecular basis for hereditary folate malabsorption. Cell, 2006. 127(5): p. 917-28. 24. Shane, B., Folylpolyglutamate synthesis and role in the regulation of onecarbon metabolism. Vitam Horm, 1989. 45: p. 263-335. 25. McGuire, J.J., et al., Enzymatic synthesis of folylpolyglutamates. Characterization of the reaction and its products. J Biol Chem, 1980. 255(12): p. 5776-88. 26. Zeng, H., et al., Transport of methotrexate (MTX) and folates by multidrug resistance protein (MRP) 3 and MRP1: effect of polyglutamylation on MTX transport. Cancer Res, 2001. 61(19): p. 7225-32. 27. G, J., Receptor- and carrier-mediated transport systems for folates and antifolates-Exploitation for folate-based chemotherapy and immunotherapy, in Antifolate drugs in cancer therapy, A.L. Jackman, Editor. 1999, Humana Press: Totowa. p. 293-321. 28. Allegra, C.J., et al., Enhanced inhibition of thymidylate synthase by methotrexate polyglutamates. J Biol Chem, 1985. 260(17): p. 9720-6. 29. Allegra, C.J., et al., Evidence for direct inhibition of de novo purine synthesis in human MCF-7 breast cells as a principal mode of metabolic inhibition by methotrexate. J Biol Chem, 1987. 262(28): p. 13520-6. 30. Assaraf, Y.G. and R.T. Schimke, Identification of methotrexate transport deficiency in mammalian cells using fluoresceinated methotrexate and flow cytometry. Proc Natl Acad Sci U S A, 1987. 84(20): p. 7154-8. 31. Drori, S., et al., Clustering of mutations in the first transmembrane domain of the human reduced folate carrier in GW1843U89-resistant leukemia cells with impaired antifolate transport and augmented folate uptake. J Biol Chem, 2000. 275(40): p. 30855-63. 32. Rots, M.G., et al., Role of folylpolyglutamate synthetase and folylpolyglutamate hydrolase in methotrexate accumulation and polyglutamylation in childhood leukemia. Blood, 1999. 93(5): p. 1677-83. 33. Hooijberg, J.H., et al., Antifolate resistance mediated by the multidrug resistance proteins MRP1 and MRP2. Cancer Res, 1999. 59(11): p. 2532-5.

116

34. 35. 36. 37. 38. 39. 40. 41. 42.

43. 44. 45. 46. 47. 48. 49. 50.

Borst, P. and R.O. Elferink, Mammalian ABC transporters in health and disease. Annu Rev Biochem, 2002. 71: p. 537-92. Kool, M., et al., MRP3, an organic anion transporter able to transport anticancer drugs. Proc Natl Acad Sci U S A, 1999. 96(12): p. 6914-9. Zeng, H., et al., Expression of multidrug resistance protein-3 (multispecific organic anion transporter-D) in human embryonic kidney 293 cells confers resistance to anticancer agents. Cancer Res, 1999. 59(23): p. 5964-7. Lee, K., A.J. Klein-Szanto, and G.D. Kruh, Analysis of the MRP4 drug resistance profile in transfected NIH3T3 cells. J Natl Cancer Inst, 2000. 92(23): p. 1934-40. Volk, E.L., et al., Overexpression of wild-type breast cancer resistance protein mediates methotrexate resistance. Cancer Res, 2002. 62(17): p. 503540. Hirohashi, T., H. Suzuki, and Y. Sugiyama, Characterization of the transport properties of cloned rat multidrug resistance-associated protein 3 (MRP3). J Biol Chem, 1999. 274(21): p. 15181-5. Chen, Z.S., et al., Analysis of methotrexate and folate transport by multidrug resistance protein 4 (ABCC4): MRP4 is a component of the methotrexate efflux system. Cancer Res, 2002. 62(11): p. 3144-50. Volk, E.L., et al., Methotrexate cross-resistance in a mitoxantrone-selected multidrug-resistant MCF7 breast cancer cell line is attributable to enhanced energy-dependent drug efflux. Cancer Res, 2000. 60(13): p. 3514-21. Chen, Z.S., et al., Transport of methotrexate, methotrexate polyglutamates, and 17beta-estradiol 17-(beta-D-glucuronide) by ABCG2: effects of acquired mutations at R482 on methotrexate transport. Cancer Res, 2003. 63(14): p. 4048-54. Volk, E.L. and E. Schneider, Wild-type breast cancer resistance protein (BCRP/ABCG2) is a methotrexate polyglutamate transporter. Cancer Res, 2003. 63(17): p. 5538-43. Mitomo, H., et al., A functional study on polymorphism of the ATP-binding cassette transporter ABCG2: critical role of arginine-482 in methotrexate transport. Biochem J, 2003. 373(Pt 3): p. 767-74. Shafran, A., et al., ABCG2 harboring the Gly482 mutation confers high-level resistance to various hydrophilic antifolates. Cancer Res, 2005. 65(18): p. 8414-22. Cole, S.P., et al., Overexpression of a transporter gene in a multidrugresistant human lung cancer cell line. Science, 1992. 258(5088): p. 1650-4. Assaraf, Y.G., et al., Loss of multidrug resistance protein 1 expression and folate efflux activity results in a highly concentrative folate transport in human leukemia cells. J Biol Chem, 2003. 278(9): p. 6680-6. Schimke, R.T., Gene amplification in cultured animal cells. Cell, 1984. 37(3): p. 705-13. Haber, D.A. and R.T. Schimke, Unstable amplification of an altered dihydrofolate reductase gene associated with double-minute chromosomes. Cell, 1981. 26(3 Pt 1): p. 355-62. Tolner, B., K. Roy, and F.M. Sirotnak, Structural analysis of the human RFC1 gene encoding a folate transporter reveals multiple promoters and alternatively spliced transcripts with 5' end heterogeneity. Gene, 1998. 211(2): p. 331-41.

117

51.

Li, W.W., et al., Intrinsic resistance to methotrexate in human soft tissue sarcoma cell lines. Cancer Res, 1992. 52(14): p. 3908-13. 52. Liani, E., et al., Loss of folylpoly-gamma-glutamate synthetase activity is a dominant mechanism of resistance to polyglutamylation-dependent novel antifolates in multiple human leukemia sublines. Int J Cancer, 2003. 103(5): p. 587-99. 53. Higgins, C.F., ABC transporters: from microorganisms to man. Annu Rev Cell Biol, 1992. 8: p. 67-113. 54. Dean, M., A. Rzhetsky, and R. Allikmets, The human ATP-binding cassette (ABC) transporter superfamily. Genome Res, 2001. 11(7): p. 1156-66. 55. Borst, P., et al., A family of drug transporters: the multidrug resistanceassociated proteins. J Natl Cancer Inst, 2000. 92(16): p. 1295-302. 56. Gottesman, M.M., T. Fojo, and S.E. Bates, Multidrug resistance in cancer: role of ATP-dependent transporters. Nat Rev Cancer, 2002. 2(1): p. 48-58. 57. Dean, M., The human ATP-binding cassette (ABC) transporter superfamily. 2002, NCBI Monographs. National Library of Medicine,: Bethesda, MD, USA. 58. Schmitt, L. and R. Tampe, Affinity, specificity, diversity: a challenge for the ABC transporter TAP in cellular immunity. Chembiochem, 2000. 1(1): p. 1635. 59. Vos, J.C., et al., Head-head/tail-tail relative orientation of the pore-forming domains of the heterodimeric ABC transporter TAP. Curr Biol, 2000. 10(1): p. 1-7. 60. Schinkel, A.H., et al., Normal viability and altered pharmacokinetics in mice lacking mdr1-type (drug-transporting) P-glycoproteins. Proc Natl Acad Sci U S A, 1997. 94(8): p. 4028-33. 61. Jonker, J.W., et al., The breast cancer resistance protein protects against a major chlorophyll-derived dietary phototoxin and protoporphyria. Proc Natl Acad Sci U S A, 2002. 99(24): p. 15649-54. 62. Sparreboom, A., et al., Limited oral bioavailability and active epithelial excretion of paclitaxel (Taxol) caused by P-glycoprotein in the intestine. Proc Natl Acad Sci U S A, 1997. 94(5): p. 2031-5. 63. Smit, J.W., et al., Absence or pharmacological blocking of placental Pglycoprotein profoundly increases fetal drug exposure. J Clin Invest, 1999. 104(10): p. 1441-7. 64. Wijnholds, J., et al., Multidrug resistance protein 1 protects the choroid plexus epithelium and contributes to the blood-cerebrospinal fluid barrier. J Clin Invest, 2000. 105(3): p. 279-85. 65. Fromm, M.F., The influence of MDR1 polymorphisms on P-glycoprotein expression and function in humans. Adv Drug Deliv Rev, 2002. 54(10): p. 1295-310. 66. Sarkadi, B., et al., Interaction of bioactive hydrophobic peptides with the human multidrug transporter. Faseb J, 1994. 8(10): p. 766-70. 67. Borgnia, M.J., G.D. Eytan, and Y.G. Assaraf, Competition of hydrophobic peptides, cytotoxic drugs, and chemosensitizers on a common P-glycoprotein pharmacophore as revealed by its ATPase activity. J Biol Chem, 1996. 271(6): p. 3163-71. 68. Young, L., et al., Role of the ABC transporter Mdl1 in peptide export from mitochondria. Science, 2001. 291(5511): p. 2135-8.

118

69. 70. 71. 72.

73. 74. 75. 76. 77.

78. 79. 80. 81. 82. 83. 84. 85.

Doyle, L.A., et al., A multidrug resistance transporter from human MCF-7 breast cancer cells. Proc Natl Acad Sci U S A, 1998. 95(26): p. 15665-70. Allikmets, R., et al., A human placenta-specific ATP-binding cassette gene (ABCP) on chromosome 4q22 that is involved in multidrug resistance. Cancer Res, 1998. 58(23): p. 5337-9. Ozvegy, C., et al., Functional characterization of the human multidrug transporter, ABCG2, expressed in insect cells. Biochem Biophys Res Commun, 2001. 285(1): p. 111-7. Brangi, M., et al., Camptothecin resistance: role of the ATP-binding cassette (ABC), mitoxantrone-resistance half-transporter (MXR), and potential for glucuronidation in MXR-expressing cells. Cancer Res, 1999. 59(23): p. 593846. Allen, J.D., et al., The mouse Bcrp1/Mxr/Abcp gene: amplification and overexpression in cell lines selected for resistance to topotecan, mitoxantrone, or doxorubicin. Cancer Res, 1999. 59(17): p. 4237-41. Robey, R.W., et al., Overexpression of the ATP-binding cassette halftransporter, ABCG2 (Mxr/BCrp/ABCP1), in flavopiridol-resistant human breast cancer cells. Clin Cancer Res, 2001. 7(1): p. 145-52. Maliepaard, M., et al., Overexpression of the BCRP/MXR/ABCP gene in a topotecan-selected ovarian tumor cell line. Cancer Res, 1999. 59(18): p. 455963. Maliepaard, M., et al., Subcellular localization and distribution of the breast cancer resistance protein transporter in normal human tissues. Cancer Res, 2001. 61(8): p. 3458-64. Maliepaard, M., et al., Circumvention of breast cancer resistance protein (BCRP)-mediated resistance to camptothecins in vitro using non-substrate drugs or the BCRP inhibitor GF120918. Clin Cancer Res, 2001. 7(4): p. 93541. Bailey-Dell, K.J., et al., Promoter characterization and genomic organization of the human breast cancer resistance protein (ATP-binding cassette transporter G2) gene. Biochim Biophys Acta, 2001. 1520(3): p. 234-41. Ee, P.L., et al., Identification of a novel estrogen response element in the breast cancer resistance protein (ABCG2) gene. Cancer Res, 2004. 64(4): p. 1247-51. Allen, J.D. and A.H. Schinkel, Multidrug resistance and pharmacological protection mediated by the breast cancer resistance protein (BCRP/ABCG2). Mol Cancer Ther, 2002. 1(6): p. 427-34. Ishikawa, T., The ATP-dependent glutathione S-conjugate export pump. Trends Biochem Sci, 1992. 17(11): p. 463-8. Paulusma, C.C., et al., Congenital jaundice in rats with a mutation in a multidrug resistance-associated protein gene. Science, 1996. 271(5252): p. 1126-8. Konig, P., et al., The crystal structure of the DNA-binding domain of yeast RAP1 in complex with telomeric DNA. Cell, 1996. 85(1): p. 125-36. Kool, M., et al., Analysis of expression of cMOAT (MRP2), MRP3, MRP4, and MRP5, homologues of the multidrug resistance-associated protein gene (MRP1), in human cancer cell lines. Cancer Res, 1997. 57(16): p. 3537-47. Allikmets, R., et al., Characterization of the human ABC superfamily: isolation and mapping of 21 new genes using the expressed sequence tags database. Hum Mol Genet, 1996. 5(10): p. 1649-55.

119

86. 87. 88. 89. 90. 91. 92. 93. 94.

95. 96. 97. 98. 99. 100. 101. 102.

Hopper, E., et al., Analysis of the structure and expression pattern of MRP7 (ABCC10), a new member of the MRP subfamily. Cancer Lett, 2001. 162(2): p. 181-91. Tammur, J., et al., Two new genes from the human ATP-binding cassette transporter superfamily, ABCC11 and ABCC12, tandemly duplicated on chromosome 16q12. Gene, 2001. 273(1): p. 89-96. Rosenberg, M.F., et al., The structure of the multidrug resistance protein 1 (MRP1/ABCC1). crystallization and single-particle analysis. J Biol Chem, 2001. 276(19): p. 16076-82. Loe, D.W., et al., ATP-dependent 17 beta-estradiol 17-(beta-D-glucuronide) transport by multidrug resistance protein (MRP). Inhibition by cholestatic steroids. J Biol Chem, 1996. 271(16): p. 9683-9. Wielinga, P., et al., The human multidrug resistance protein MRP5 transports folates and can mediate cellular resistance against antifolates. Cancer Res, 2005. 65(10): p. 4425-30. Hooijberg, J.H., et al., Online fluorescent method to assess BCRP/ABCG2 activity in suspension cells. Nucleosides Nucleotides Nucleic Acids, 2004. 23(8-9): p. 1451-4. Ifergan, I., et al., Folate deprivation results in the loss of breast cancer resistance protein (BCRP/ABCG2) expression. A role for BCRP in cellular folate homeostasis. J Biol Chem, 2004. 279(24): p. 25527-34. Turner, F.B., et al., Tissue-specific expression of functional isoforms of mouse folypoly-gamma-glutamae synthetase: a basis for targeting folate antimetabolites. Cancer Res, 1999. 59(24): p. 6074-9. Rothem, L., A. Aronheim, and Y.G. Assaraf, Alterations in the expression of transcription factors and the reduced folate carrier as a novel mechanism of antifolate resistance in human leukemia cells. J Biol Chem, 2003. 278(11): p. 8935-41. Kanzaki, A., et al., Expression of multidrug resistance-related transporters in human breast carcinoma. Jpn J Cancer Res, 2001. 92(4): p. 452-8. Cole, P.D., et al., Effects of overexpression of gamma-Glutamyl hydrolase on methotrexate metabolism and resistance. Cancer Res, 2001. 61(11): p. 4599604. Rothem, L., et al., The reduced folate carrier gene is a novel selectable marker for recombinant protein overexpression. Mol Pharmacol, 2005. 68(3): p. 616-24. Taylor, C.W., et al., Different mechanisms of decreased drug accumulation in doxorubicin and mitoxantrone resistant variants of the MCF7 human breast cancer cell line. Br J Cancer, 1991. 63(6): p. 923-9. Assaraf, Y.G. and I.D. Goldman, Loss of folic acid exporter function with markedly augmented folate accumulation in lipophilic antifolate-resistant mammalian cells. J Biol Chem, 1997. 272(28): p. 17460-6. Jansen, G., et al., Measurement of folylpolyglutamate synthetase activity in head and neck squamous carcinoma cell lines and clinical samples using a new rapid separation procedure. Oncol Res, 1992. 4(7): p. 299-305. Darzynkiewicz, Z., X. Li, and J. Gong, Assays of cell viability: discrimination of cells dying by apoptosis. Methods Cell Biol, 1994. 41: p. 15-38. Ifergan, I., G. Jansen, and Y.G. Assaraf, Cytoplasmic confinement of breast cancer resistance protein (BCRP/ABCG2) as a novel mechanism of

120

103. 104. 105. 106. 107.

108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118.

adaptation to short-term folate deprivation. Mol Pharmacol, 2005. 67(4): p. 1349-59. Assaraf, Y.G., The role of multidrug resistance efflux transporters in antifolate resistance and folate homeostasis. Drug Resist Updat, 2006. 9(4-5): p. 227-46. Mogi, M., et al., Akt signaling regulates side population cell phenotype via Bcrp1 translocation. J Biol Chem, 2003. 278(40): p. 39068-75. Wielinga, P.R., et al., A method for studying plasma membrane transport with intact cells using computerized fluorometry. Anal Biochem, 1998. 263(2): p. 221-31. Allen, J.D., et al., Potent and specific inhibition of the breast cancer resistance protein multidrug transporter in vitro and in mouse intestine by a novel analogue of fumitremorgin C. Mol Cancer Ther, 2002. 1(6): p. 417-25. Jansen, G., et al., Multiple mechanisms of resistance to polyglutamatable and lipophilic antifolates in mammalian cells: role of increased folylpolyglutamylation, expanded folate pools, and intralysosomal drug sequestration. Mol Pharmacol, 1999. 55(4): p. 761-9. Assaraf, Y.G. and J.I. Slotky, Characterization of a lipophilic antifolate resistance provoked by treatment of mammalian cells with the antiparasitic agent pyrimethamine. J Biol Chem, 1993. 268(6): p. 4556-66. Kleizen, B. and I. Braakman, Protein folding and quality control in the endoplasmic reticulum. Curr Opin Cell Biol, 2004. 16(4): p. 343-9. Baron, D., et al., Lack of plasma membrane targeting of a G172D mutant thiamine transporter derived from Rogers syndrome family. Mol Med, 2002. 8(8): p. 462-74. Robey, R.W., et al., Mutations at amino-acid 482 in the ABCG2 gene affect substrate and antagonist specificity. Br J Cancer, 2003. 89(10): p. 1971-8. Assaraf, Y.G. and M.J. Borgnia, Probing the interaction of the multidrugresistance phenotype with the polypeptide ionophore gramicidin D via functional channel formation. Eur J Biochem, 1994. 222(3): p. 813-24. Assaraf, Y.G., et al., Computer modelling of antifolate inhibition of folate metabolism using hybrid functional petri nets. J Theor Biol, 2006. 240(4): p. 637-47. Matherly, L.H. and D.I. Goldman, Membrane transport of folates. Vitam Horm, 2003. 66: p. 403-56. Gates, S.B., et al., Dietary folate and folylpolyglutamate synthetase activity in normal and neoplastic murine tissues and human tumor xenografts. Biochem Pharmacol, 1996. 52(9): p. 1477-9. Stark, M., et al., Antifolate resistance associated with loss of MRP1 expression and function in Chinese hamster ovary cells with markedly impaired export of folate and cholate. Mol Pharmacol, 2003. 64(2): p. 220-7. Hooijberg, J.H., et al., The role of multidrug resistance proteins MRP1, MRP2 and MRP3 in cellular folate homeostasis. Biochem Pharmacol, 2003. 65(5): p. 765-71. Rothem, L., et al., Reduced folate carrier gene silencing in multiple antifolateresistant tumor cell lines is due to a simultaneous loss of function of multiple transcription factors but not promoter methylation. J Biol Chem, 2004. 279(1): p. 374-84.

121

119. 120.

121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136.

Worm, J., et al., Methylation-dependent silencing of the reduced folate carrier gene in inherently methotrexate-resistant human breast cancer cells. J Biol Chem, 2001. 276(43): p. 39990-40000. Rothem, L., et al., Resistance to multiple novel antifolates is mediated via defective drug transport resulting from clustered mutations in the reduced folate carrier gene in human leukaemia cell lines. Biochem J, 2002. 367(Pt 3): p. 741-50. Summer, R., et al., Side population cells and Bcrp1 expression in lung. Am J Physiol Lung Cell Mol Physiol, 2003. 285(1): p. L97-104. Misra, S., et al., The role of phosphoinositide 3-kinase in taurocholateinduced trafficking of ATP-dependent canalicular transporters in rat liver. J Biol Chem, 1998. 273(41): p. 26638-44. Mizuarai, S., N. Aozasa, and H. Kotani, Single nucleotide polymorphisms result in impaired membrane localization and reduced atpase activity in multidrug transporter ABCG2. Int J Cancer, 2004. 109(2): p. 238-46. Rajagopal, A. and S.M. Simon, Subcellular localization and activity of multidrug resistance proteins. Mol Biol Cell, 2003. 14(8): p. 3389-99. Sarkadi, B., et al., ABCG2 -- a transporter for all seasons. FEBS Lett, 2004. 567(1): p. 116-20. Doyle, L.A. and D.D. Ross, Multidrug resistance mediated by the breast cancer resistance protein BCRP (ABCG2). Oncogene, 2003. 22(47): p. 734058. Haimeur, A., et al., The MRP-related and BCRP/ABCG2 multidrug resistance proteins: biology, substrate specificity and regulation. Curr Drug Metab, 2004. 5(1): p. 21-53. Leonard, G.D., T. Fojo, and S.E. Bates, The role of ABC transporters in clinical practice. Oncologist, 2003. 8(5): p. 411-24. Mechetner, E.B., et al., P-glycoprotein function involves conformational transitions detectable by differential immunoreactivity. Proc Natl Acad Sci U S A, 1997. 94(24): p. 12908-13. Stuhlsatz-Krouper, S.M., N.E. Bennett, and J.E. Schaffer, Substitution of alanine for serine 250 in the murine fatty acid transport protein inhibits long chain fatty acid transport. J Biol Chem, 1998. 273(44): p. 28642-50. Forgac, M., Structure and function of vacuolar class of ATP-driven proton pumps. Physiol Rev, 1989. 69(3): p. 765-96. James, P., et al., Modulation of erythrocyte Ca2+-ATPase by selective calpain cleavage of the calmodulin-binding domain. J Biol Chem, 1989. 264(14): p. 8289-96. Clarke, D.M., et al., Location of high affinity Ca2+-binding sites within the predicted transmembrane domain of the sarcoplasmic reticulum Ca2+ATPase. Nature, 1989. 339(6224): p. 476-8. Rabon, E.C., T.L. McFall, and G. Sachs, The gastric [H,K]ATPase:H+/ATP stoichiometry. J Biol Chem, 1982. 257(11): p. 6296-9. Malassine, A., et al., Ultrastructural visualization of the internalization of low density lipoprotein by human placental cells. Histochemistry, 1987. 87(5): p. 457-64. de Bruin, M., et al., Reversal of resistance by GF120918 in cell lines expressing the ABC half-transporter, MXR. Cancer Lett, 1999. 146(2): p. 11726.

122

137. 138. 139. 140.

141.

Yanase, K., et al., Gefitinib reverses breast cancer resistance proteinmediated drug resistance. Mol Cancer Ther, 2004. 3(9): p. 1119-25. Elkind, N.B., et al., Multidrug transporter ABCG2 prevents tumor cell death induced by the epidermal growth factor receptor inhibitor Iressa (ZD1839, Gefitinib). Cancer Res, 2005. 65(5): p. 1770-7. Houghton, P.J., et al., Imatinib mesylate is a potent inhibitor of the ABCG2 (BCRP) transporter and reverses resistance to topotecan and SN-38 in vitro. Cancer Res, 2004. 64(7): p. 2333-7. Erlichman, C., et al., The HER tyrosine kinase inhibitor CI1033 enhances cytotoxicity of 7-ethyl-10-hydroxycamptothecin and topotecan by inhibiting breast cancer resistance protein-mediated drug efflux. Cancer Res, 2001. 61(2): p. 739-48. Imai, Y., et al., Phytoestrogens/flavonoids reverse breast cancer resistance protein/ABCG2-mediated multidrug resistance. Cancer Res, 2004. 64(12): p. 4346-52.

123

‫תפקידם של טרנספורטרים בהומאוסטאזיס‬ ‫של פולאטים ועמידות לתרופות אנטיסרטניות‬

‫חיבור על מחקר‬

‫לשם מילוי חלקי של הדרישות לקבלת התואר‬ ‫דוקטור לפילוסופיה‬

‫הוגש לסנט הטכניון ‪ -‬מכון טכנולוגי לישראל‬ ‫ולאוניברסיטת חיפה ‪ -‬הרשות ללימודים מתקדמים‬

‫‪124‬‬

‫אדר תשס"ז מרץ ‪2007‬‬

‫תפקידם של טרנספורטרים בהומאוסטאזיס‬ ‫של פולאטים ועמידות לתרופות אנטיסרטניות‬

‫חיבור על מחקר‬ ‫לשם מילוי חלקי של הדרישות לקבלת התואר‬ ‫דוקטור לפילוסופיה‬

‫אילן איפרגן‬

‫הוגש לסנט הטכניון ‪ -‬מכון טכנולוגי לישראל‬ ‫ולאוניברסיטת חיפה ‪ -‬הרשות ללימודים מתקדמים‬

‫אדר תשס"ז חיפה פברואר ‪2007‬‬

‫‪125‬‬

‫המחקר נעשה בהנחיית פרופסור חבר יהודה אסרף במסגרת התוכנית‬ ‫המשותפת לכלכלה בטכניון ובאוניברסיטת חיפה‪.‬‬ ‫אני מודה לטכניון על התמיכה הכספית הנדיבה בהשתלמותי‪.‬‬

‫‪126‬‬

‫תפקידם של טרנספורטרים בהומאוסטאזיס של פולאטים ועמידות לתרופות אנטיסרטניות‬ ‫אילן איפרגן‬ ‫תקציר‬ ‫תפקידם של טרנספורטרים ממברנליים בעמידות לתרופות אנטיסרטניות ובהומאוסטאזיס של פולאטים‬ ‫הינו בעל חשיבות רפואית ופיזיולוגית רבה‪ .‬הפולאטים הם נגזרות שונות של ויטמין ‪ B9‬החיוניים‬ ‫לביוסינטזת נוקלאוטידים ולכן גם לשיכפול דנ"א וחלוקת תאים‪ .‬כאשר הפולאטים חודרים לתא הם‬ ‫עוברים תהליך של פוליגלוטמילציה ע"י האנזים ‪ FPGS‬אשר מאפשר את שימור הפולאטים בתא‪.‬‬ ‫בתהליך הפוליגלוטמילציה מתווספים עד ‪ 8‬שיירי גלוטמאט לפולאט ובכך קטנה באופן משמעותי יכולת‬ ‫השלכת הפולאטים אל מחוץ לתא ע"י טרנספורטרים הממוקמים בממברנת התא‪ .‬הנשא ‪ RFC‬ממוקם‬ ‫בממברנת התא ופועל כטרנספורטר דו‪-‬כיווני בעל אפיניות גבוהה לפולאטים מחוזרים‪ .‬בספרות ידוע כי‬ ‫ביטוי הטרנספורטר ‪ RFC‬גדל בתאי הלאוקמיה ‪ CEM/7A‬פי ‪ 100‬יחסית לתאי האב ‪ , CEM‬ותכונה זו‬ ‫איפשרה לתאים לגדול בריכוזי פולאטים הקטנים פי ‪ 120‬מאלו המצויים בסרום רגיל‪ .‬במחקר הנוכחי‬ ‫הנחנו כי לטרנספורטר ‪RFC‬אמורה להיות גם פעילות מזיקה לתאים בשל יכולתו להשליך פולאטים‬ ‫המצומדים לשייר גלוטמי יחיד אל מחוץ לתא‪ .‬פעילות מזיקה זו אמורה לבוא לידי ביטוי בתנאי הרעבה‬ ‫לפולאטים שבהם פעילות השלכת הפולאטים ע"י הנשא ‪ RFC‬תגרום להדלדלות מאגר הפולאטים התוך‪-‬‬ ‫תאי ועקב כך להקטנת קצב חלוקות התאים‪ .‬לפי הנחה זאת הקטנת ריכוז הפולאטים המצומדים לשייר‬ ‫גלוטמי יחיד תגרום להטיית שיווי המשקל לכיוון הסרת שיירי גלוטמאט מהפולאטים המצומדים לשרשרת‬ ‫פוליגלוטמית ע"י אנזים שנקרא ‪ GGH‬בכדי למזער את השינוי ) עקרון לה שטליה(‪ .‬דבר זה יגרום‬ ‫להיווצרות עוד פולאטים המצומדים לשייר גלוטמי יחיד ואלו יושלכו מהתא ע"י הנשא ‪ RFC‬ובכך‬ ‫יתאפשר המשך ריקון מאגר הפולאטים התוך תאי ע"י ‪ .RFC‬במחקר זה הוכחנו כי רק בתנאי הרעבה‬ ‫לפולאטים קצב חלוקת תאים שמבטאים ביתר את הנשא ‪ RFC‬קטן פי~ ‪ 15‬יחסית לתאים חסרי פעילות‬ ‫של הנשא וכן יחסית לתאים המבטאים ביתר את הרצפטור לפולאטים המכונה ‪ .FR‬יתר על כן‪ ,‬הבחנו כי‬ ‫פעילות הנשא ‪ RFC‬קטנה באופן משמעותי בקווי תאים שונים שהורעבו לפולאטים למשך ‪ 7-3‬ימים‪.‬‬

‫‪127‬‬

‫באנליזת ‪ RT-PCR‬כמותי התברר כי רמות תעתיקי הרנ"א של הנשא ‪ RFC‬וכן ‪ GGH,‬אנזים המקטלז‬ ‫את את הפיכת הפולאט בעל שיירי גלוטאמט רבים לפולאט בעל שייר גלוטאמט יחיד המהווה סובסטראט‬ ‫לנשא ‪ RFC,‬קטנו פי ~‪ 2.5‬ופי~ ‪ 2.6‬בהתאמה‪ ,‬כאשר רמות תעתיקי הרנ"א של טרנספורטרים תלויי‬ ‫‪ ATP‬כגון ‪ MRP1, MRP5‬ו ‪ BCRP‬נותרו ללא שינוי‪ .‬הסקנו כי בעת הרעבה לפולאטים‪ ,‬פעילות‬ ‫הנשא ‪ RFC‬הינה מזיקה מכיוון שהנשא משליך פולאטים בעלי שייר גלוטמתי יחיד‪ ,‬תהליך המזורז ע"י‬ ‫האנזים ‪ GGH‬שכאמור מקטלז את הפיכת הפולאט בעל שיירי גלוטמט רבים לפולאט בעל שייר גלוטאמט‬ ‫יחיד‪ .‬יתר על כן הצענו מנגנון הסתגלותי חדש בתאים בעת הרעבה לפולאטים אשר מאופיין בירידה‬ ‫בפעילות הנשא ‪ RFC‬והאנזים ‪ GGH‬וזאת באמצעות ירידה בתעתיקי הרנ"א הספציפיים לחלבונים אלו‪.‬‬ ‫בניגוד לנשא ‪ RFC‬ולטרנספורטרים תלויי ‪ ATP‬כגון ‪ , MRP1-4‬אשר משליכים פולאטים המצומדים‬ ‫לשייר גלוטמי יחיד‪ ,‬הטרנספורטר )‪ ABCG2) BCRP‬הינו היחיד הידוע כמשליך פולאטים המצומדים‬ ‫לעד שלושה שיירי גלוטמאט‪ .‬בשל יכולת יוצאת דופן זו של ‪ BCRP‬להשליך פולאטים המצומדים לעד‬ ‫שלושה שיירי גלוטמאט חקרנו את תפקידו האפשרי של הטרנספורטר בהומאוסטאזיס של פולאטים‪.‬‬ ‫במחקר זה היה שימוש בשני קווי סרטן השד בעלי רמות בינוניות וגבוהות של הנשא המכונים ‪ MCF7‬ו ‪,‬‬ ‫‪ MCF7/MR‬בהתאמה‪ .‬קווי תאים אלה עברו אדפטציה שנמשכה כ‪ 3-‬חודשים לגידול בריכוז פולאטים‬ ‫הנמוך פי ‪ 700‬מהריכוז הרגיל‪ .‬מצאנו כי התאים שנוצרו בעקבות האדפטציה לגידול בריכוזי פולאטים‬ ‫נמוכים ביטאו רמות קלושות של תעתיקי רנ"א והחלבון ‪BCRP‬אך שמרו על רמות זהות של תעתיקי‬ ‫רנ"א עבור הטרנספורטרים תלויי ה‪ .ATP MRP1-5, -‬הירידה ברמת החלבון ‪ BCRP‬אופיינה גם ב‪:‬‬ ‫א( ירידה ביכולת השלכת ‪ Hoechst 33342‬המשמש כסובסטראט ספציפי של ‪ .BCRP‬ב( עלייה של‬ ‫כ‪-‬פי ‪ 2‬באגירה התרופה מיטוזנטרון הידועה כסובסטראט ספציפי של ‪ .BCRP‬ג( איבוד רוב העמידות‬ ‫לתרופה מיטוזנטרון‪ .‬ד( עלייה של פי ‪ 2‬במבחן אגירת חומצה פולית רדיואקטיבית‪ .‬יתר על כן מצאנו כי‬ ‫בקווי התאים שעברו אדפטציה לגידול בריכוזי הפולאטים הנמוכים קיימת פעילות יתר של ‪,FPGS‬‬ ‫אנזים אשר מגדיל את השרשרת הגלוטמית של הפולאט עד לכדי ‪ 9‬שיירי גלוטאמאט ובכך מקטין את‬ ‫הפרקציה הפולאטית התוך‪-‬תאית המוכרת ע"י הטרנספורטר ‪ .BCRP‬כמו כן מצאנו כי בהרעבה‬ ‫לפולאטים של תאי סרטן השד למשך שבועיים בלבד מיקום רוב )‪ (86%‬יחידות הטרנספורטר ‪BCRP‬‬ ‫היה ברטיקולום האנדופלסמאטי ולא בממברנת התא כפי שהיה בטרם ההרעבה‪ .‬השינוי במיקום ‪BCRP‬‬

‫‪128‬‬

‫היה מלווה בירידה של פי ‪ 3‬לערך ברמת החלבון של ‪ . BCRP‬בניגוד ל‪ BCRP ,‬מיקומם הממברנלי של‬ ‫החלבונים הבאים נותר ללא שינוי בתנאי ההרעבה לפולאטים‪ EGFR, FGFR1-3 :‬ו‪MHC class -‬‬ ‫‪ .I‬השינוי במיקום ‪ BCRP‬בתאים שהורעבו לפולאטים היה מלווה בעלייה של פי ‪ 3‬לערך במבחן אגירת‬ ‫‪ rhodamine 123‬המשמש כסובסטראט של ‪BCRP‬מבחן אגירה זה הוכיח כי הפעילות התאית של‬ ‫הטרנספורטר ‪ BCRP‬קטנה באופן דרסטי בקווי התאים שהורעבו לפולאטים‪ .‬לפיכך הצענו לראשונה כי‬ ‫הירידה ברמת הטרנספורטר ‪ BCRP‬ושינוי מיקומו התאי מהווים מנגנוני אדפטציה חשובים‬ ‫בהומאוסטאזיס של פולאטים וזאת בשל יכולתו יוצאת הדופן של הטרנספורטר להשליך אל מחוץ לתא‬ ‫פולאטים המצומדים לעד שלושה שיירי גלוטמאט‪ .‬במהלך המחקר על מיקום הטרנספורטר ‪ BCRP‬בתאי‬ ‫סרטן השד ‪ MCF-7/MR‬העמידים לתרופה האנטי‪-‬סרטנית מיטוזנטרון גילינו כי הטרנספורטר ממוקם‬ ‫בעיקר באיזורי המגע של התאים אלו באלו‪ .‬עד מחקר זה ההנחה בספרות הייתה כי הטרנספורטר ‪BCRP‬‬ ‫ממוקם באופן אחיד בממברנת התא כולה‪ .‬נשאלה השאלה כיצד המיקום הבין‪-‬תאי של ‪ BCRP‬מאפשר‬ ‫הקניית עמידות לתרופה מיטוזנטרון‪ .‬בעזרת מיקרוסקופיה אלקטרונית וקונפוקאלית גילינו כי איזורי מגע‬ ‫אלה הם חלק מוזיקולה חוץ‪-‬תאית שלא היה ידוע על קיומה ואשר מרכזת בתוכה את התרופה האנטי‪-‬‬ ‫סרטנית מיטוזנטרון‪ .‬הראינו כי לאחר ‪ 12‬שעות הדגרה של התאים עם התרופה מיטוזנטרון ריכוזה‬ ‫בוזיקולה גדול פי ‪ 1000‬לערך מריכוזה החוץ‪-‬תאי‪ ,‬ריכוז זה של התרופה בוזיקולה נמנע לחלוטין‬ ‫בנוכחות מעכבים ספציפים לפעילות הטרנספורטר ‪ . BCRP‬ריכוז התרופה בוזיקולה נמנע גם בעת‬ ‫ביטול מאגרי ה ‪ ATP‬התוך‪-‬תאיים ; יש לציין כי הטרנספורטר ‪ BCRP‬מבצע את פעולת השלכת‬ ‫הסובסראטים במחיר הידרוליזת מולקולת ‪ ATP‬ולכן בהעדר ‪ ATP‬פעילות הטרנספורטר ‪BCRP‬‬ ‫תושבת‪ .‬מניסויים אלו למדנו כי פעולת השלכת התרופה מיטוזנטרון מתווכת באופן ספציפי ע"י‬ ‫הטרנספורטר ‪ BCRP.‬באמצעות מיקרוסקופיה קונפוקאלית הראינו כי הטרנספורטר ‪ BCRP‬ממוקם‬ ‫באזורי המגע של הוזיקולה והתאים השכנים אך לא בצד האפיקלי של הוזיקולה‪ .‬כמו כן חתכי רוחב‬ ‫קונפוקאלים של הוזיקולה חשפו כי אין מגע בין המדיום החוץ‪-‬תאי לדפנות הצידיים של הוזיקולה‪ .‬חתכים‬ ‫המאונכים למשטח גידול התאים חשפו כי המגע היחידי בן המדיום החוץ‪-‬תאי לוזיקולה מצוי בחלק‬ ‫האפיקלי של הוזיקולה‪ .‬מניסויים אלה למדנו כי הנפח הממוצע של הוזיקולה הוא כ ‪ 194‬פמטוליטר‬ ‫)כמחצית הנפח של לויקוציט(‪ .‬יש לציין כי וזיקולות דומות בעלות יכולת אגירה לתרופה מיטוזנטרון‬

‫‪129‬‬

‫נמצאו גם בקווי תאים נוספים של סרטן השד כגון ‪ MCF-7/FLV1000‬אשר עמיד לתרופה‬ ‫האנטיסרטנית ‪ .flavopiridol‬הצענו שהוזיקולה שהתגלתה במחקר הנוכחי משמשת את תאי סרטן השד‬ ‫כחדר אשפה המשותף למספר תאים שכנים וזאת ע"י השלכת התרופה לוזיקולה באופן ספציפי ואקטיבי‬ ‫ע"י הטרנספורטר ‪ . BCRP‬מחקר זה מידל מחדש את נושא העמידות האנטיסרטנית המתווכת באמצעות‬ ‫הטרנספורטר ‪ BCRP‬בתאי סרטן השד אך אין לשלול את קיומן של וזיקולות דומות בתאי סרטן מרקמות‬ ‫שונות‪.‬‬

‫‪130‬‬