e 117 (2007) 28–39www.elsevier.com/locate/jconrel

Journal of Controlled Releas

Investigating the mechanism of enhanced cytotoxicity of HPMAcopolymer–Dox–AGM in breast cancer cells

Francesca Greco a,b,1, María J. Vicent a,2, Siobhan Gee a, Arwyn T. Jones a, Julia Gee b,Robert I. Nicholson b, Ruth Duncan a,⁎

a Centre for Polymer Therapeutics, Welsh School of Pharmacy, Redwood Building, Cardiff University, King Edward VII Ave., Cardiff CF10 3XF, UKb Tenovus Centre for Cancer Research, Welsh School of Pharmacy, Redwood Building, Cardiff University, King Edward VII Ave., Cardiff CF10 3XF, UK

Received 12 July 2006; accepted 5 October 2006Available online 14 October 2006

Abstract

Recently we have described an HPMA copolymer conjugate carrying both the aromatase inhibitor aminoglutethimide (AGM) and doxorubicin(Dox) as combination therapy. This showed markedly enhanced in vitro cytotoxicity compared to the HPMA copolymer–Dox (FCE28068), aconjugate that demonstrated activity in chemotherapy refractory breast cancer patients during early clinical trials. To better understand the superioractivity of HPMA copolymer–Dox–AGM, here experiments were undertaken using MCF-7 and MCF-7ca (aromatase-transfected) breast cancercell lines to: further probe the synergistic cytotoxic effects of AGM and Dox in free and conjugated form; to compare the endocytic properties ofHPMA copolymer–Dox–AGM and HPMA copolymer–Dox (binding, rate and mechanism of cellular uptake); the rate of drug liberation bylysosomal thiol-dependant proteases (i.e. conjugate activation), and also, using immunocytochemistry, to compare their molecular mechanism ofaction. It was clearly shown that attachment of both drugs to the same polymer backbone was a requirement for enhanced cytotoxicity. FACSstudies indicated both conjugates have a similar pattern of cell binding and endocytic uptake (at least partially via a cholesterol-dependentpathway), however, the pattern of enzyme-mediated drug liberation was distinctly different. Dox release from PK1 was linear with time, whereasthe release of both Dox and AGM from HPMA copolymer–Dox–AGM was not, and the initial rate of AGM release was much faster than thatseen for the anthracycline. Immunocytochemistry showed that both conjugates decreased the expression of ki67. However, this effect was moremarked for HPMA copolymer–Dox–AGM and, moreover, only this conjugate decreased the expression of the anti-apoptotic protein bcl-2. Inconclusion, the superior in vitro activity of HPMA copolymer–Dox–AGM cannot be attributed to differences in endocytic uptake, and it seemslikely that the synergistic effect of Dox and AGM is due to the kinetics of intracellular drug liberation which leads to enhanced activity.© 2006 Elsevier B.V. All rights reserved.

Keywords: Polymer–drug conjugates; Polymer therapeutics; Tumour targeting; Endocrine therapy; Chemotherapy; Breast cancer

1. Introduction

A growing number of polymer–drug conjugates have enteredclinical trial as antitumour agents (reviewed in [1–4]) and a poly(glutamic acid)–paclitaxel conjugate (XYOTAX™) is currentlyshowing particular promise for the treatment of non smallcell lung cancer (NSCLC) in women [5]. Due to their changed

⁎ Corresponding author. Tel.: +44 2920874180; fax: +44 2920874536.E-mail address: DuncanR@cf.ac.uk (R. Duncan).

1 Current address: School of Pharmacy, University of Reading, Whiteknights,PO Box 226, Reading, Berkshire, RG6 6AP, UK.2 Current address: Centro de Investigación Príncipe Felipe, Polymer

Therapeutics Laboratory, Medicinal Chemistry Unit, Av. Autopista del Saler16, E-46013, Valencia, Spain.

0168-3659/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.jconrel.2006.10.012

pharmacokinetics (at the cellular and whole organism level)compared to low molecular weight agents, polymer–drugconjugates have recognised advantages compared to conven-tional chemotherapy. These include passive tumour targetingdue to the enhanced permeability and retention (EPR) effect, aphenomenon that arises due to the hyperpermeability ofangiogenic tumour vasculature [6], the ability to reduce toxicityof bound drug [7], and following cellular uptake by theendocytic route, the potential to bypass mechanisms of drugresistance [7,8]. Our (RD) previous studies have given rise toseveral N-(2-hydroxypropyl)methacrylamide (HPMA) copoly-mer conjugates bearing doxorubicin (Dox), paclitaxel, camp-tothecin and platinates that have progressed into clinical trial(reviewed in [9,10]). Given the proven lack of clinical toxicity of

Fig. 1. Structure and characteristics of drugs tested.

29F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

this polymer, more recently conjugates containing experimentaldrugs (reviewed in [9,11]), and drugs directed towards newtherapeutic targets (e.g. anti-angiogenic HPMA copolymer–TNP470 [12]) have emerged.

This project is developing the first polymer-based combina-tion therapy for breast cancer. Breast cancer is one of the mostcommonmalignancies worldwide and it affects∼1 in 10womenin the UK [13]. Although recent statistics show improvedprognosis due to early diagnosis and surgery, often followed byendocrine and/or chemotherapy [15], the outlook is still poor forthose patients with metastatic or resistant disease (∼20%survival at 5 years). Whereas the selective oestrogen receptormodulator (SERM) tamoxifen was a major advance (28%reduction in mortality of breast cancer patients at 5 years [14]),recent clinical trials have shown that an aromatase inhibitor,anastrozole, is superior to tamoxifen in terms of efficacy, time torecurrence and has less side-effects [15]. Nevertheless, inherentand acquired resistance to both chemotherapy and endocrinetherapy still present an unresolved problem [16].

The observations that HPMA copolymer–Dox (FCE28068)showed activity in chemotherapy refractory breast cancer patientsin Phase I clinical trials [7], and that aromatase inhibitors can actsynergistically with chemotherapy [17], led us to synthesise anHPMA copolymer conjugate containing the aromatase inhibitoraminoglutethimide (AGM) and doxorubicin (Dox) attached to thesame polymeric carrier (Fig. 1) [18]. This combination conjugateshowed markedly enhanced cytotoxicity against MCF-7 cells invitro compared to HPMA copolymer–Dox. Furthermore, experi-ments studying a library of HPMA copolymer conjugates con-taining AGM alone (the first conjugates to contain endocrinetherapy), confirmed aromatase inhibition in vitro, and that AGMliberation was a requirement for activity [18,19].

The aim of this study was to further investigate the mechanismof enhanced cytotoxicity of HPMA copolymer–Dox–AGM,particularly in comparison with HPMA copolymer–Dox. First,further control experiments were conducted to confirm theenhanced cytotoxicity of the polymer combination. It washypothesised that HPMA copolymer–Dox–AGM and HPMAcopolymer–Dox might display (i) different mechanisms or ratesof endocytic uptake, (ii) differences in the rate of release ofbioactive drug(s) – this has been seen previously with otherHPMA copolymer conjugates due to differences in conjugateconformation [20], and/or (iii) differences in the molecularmechanisms of cell killing. The cytotoxicity of conjugates, freeDox and AGM and their mixtures was determined using theMTTassay in MCF-7 and MCF-7ca (aromatase-transfected) cell lines.flow cytometry and live-cell imaging were used to evaluate cellbinding (4 °C) and endocytic uptake (37 °C). In addition, studiesin the presence of methyl-β-cyclodextrin (MβCD) (inhibitsclathrin-mediated and clathrin- and caveolin-independent endo-cytosis), chlorpromazine (inhibits clathrin-mediated endocytosis)and cytochalasin B (inhibits macropinocytosis) were undertakento probe the mechanism of endocytic internalisation. The rateof Dox and AGM liberation from the conjugates was measuredin vitro in the presence of rat liver lysosomal enzymes (tritosomes)as previously described [21,22]. Finally, immunocytochem-istry was used to assess the effect of both conjugates on the

proliferation marker ki67 and the anti-apoptotic protein bcl-2[23–25].

2. Materials and methods

2.1. Materials and cells

HPMA copolymer precursors carrying –GG–ONp or–GFLG–ONp side chains (either 5 or 10 mol%) ofMw∼20.000–25.000 g/mol and Mw/Mn=1.3–1.5 were fromPolymer Laboratories Ltd. (Shropshire, UK). The ONp contentwas calculated using ε274 nm in DMSO=9500 L/mol cm, and theHPMA copolymer–Dox±AGM and HPMA copolymer–AGMconjugates shown in Fig. 1 were synthesised and characterised aspreviously described [18]. Anhydrous DMF and optical gradeDMSO were from Sigma-Aldrich Company Ltd. (Dorset, UK)and all HPLC grade solvents by Fischer Scientific (GreaterManchester, UK). Medical grade O2, N2 and CO2 (all 95% v/v)and liquid nitrogenwere all fromBOCGases (Surrey, UK). AGMwas from Aldrich (Dorset, UK) and Dox·HCl was fromPharmacia (Lombardia, Italy). All other reagents were of generallaboratory grade and were from Aldrich (Dorset, UK) unlessotherwise stated.

The human oestrogen-dependant, breast carcinoma cell linesMCF-7 and MCF-7ca (human aromatase-transfected) werefrom the Tenovus Centre for Cancer Research at CardiffUniversity. Cells were cultured in RPMI 1640 with L-glutaminemedium supplemented with 5% of foetal bovine serum (FBS) asstandard tissue culture conditions. RPMI and FBS were from

30 F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

GIBCO BRL Life Technologies (Paisley, UK). In order tomaintain the transfected strain, the culture medium of MCF-7cawas always (for routine tissue culture and for all the experi-ments) further supplemented with 0.75 mg/mL of geneticin(supplied by Fluka, Dorset, UK). Tissue culture grade DMSO,3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide(MTT), Trypan blue solution (0.4%) (cell culture grade) andoestradiol were from Sigma (Dorset, UK). Steroid-deprivedFCS (SFCS) was prepared as described previously [19].

For immunocytochemistry, the solvents used for fixation(methanol and acetone) and formaldehyde solution (40%) werefrom Fisher Scientific (Leicestershire, UK). The monoclonalmouse primary antibodies against Ki67 antigen (clone MIB-1),and bcl-2 oncoprotein (clone 124), polyclonal goat anti-mouse(GAM) immunoglobulins, peroxidase anti-peroxidase (PAP),and diaminobenzidine (DAB) were all from DakoCytomation(CA, USA). The supersensitive concentrated detection kit (i.e.Link (biotinylated anti-mouse immunoglobulins) and theconcentrated label (streptavidin peroxidase) were from Bio-genex (CA, USA). The DPX mountant used for histology,Tween and the methyl green were all from Sigma-Aldrich(Dorset, UK). Glycerol and sucrose were from Fisher Scientific(Leicestershire, UK).

2.2. Cytotoxicity of Dox, AGM and HPMA copolymer conjugates

The cytotoxicity of free drugs (AGM and Dox) and theirconjugates was evaluated using the MTT cell viability assay(72 h of incubation) with MCF-7 and MCF-7ca cells. Both celllines were seeded into sterile 96-well microtitre plates(4×104 cells/mL) in WRPMI 1640 with 5.0 mM L-glutamineand 5% (v/v) of SFCS±10−9 M oestradiol. They were incu-bated for 5 days before the test compounds (0.2 μm filter-sterilised) were added to give a final concentration in the rangeof 0–2 mg/mL drug-equiv. After a further 67-h incubation,MTT (20 μL of a 5 mg/mL solution in PBS) was added to eachwell, and the cells were incubated for 5 h. After removal of themedium, the precipitated formazan crystals were dissolved inoptical grade DMSO (100 μL), and the plates were read spec-trophotometrically at 550 nm after 30 min using a microtitreplate reader. Cell viability was expressed as a percentage ofthe viability of untreated control cells. The IC50 values wereexpressed as concentration (μg/mL) of AGM or Dox equiv.

2.3. Binding and uptake of HPMA copolymer–Dox±AGMconjugates by MCF-7 and MCF-7 cells

MCF-7 cells were seeded in 6-well plates at a density of1×106 cells/mL (1 mL cell suspension per well) and allowed toadhere for 24 h. In the binding experiments conducted at 4 °C,the cells were pre-incubated at this temperature for 30 min priorto addition of conjugate. HPMA copolymer–Dox conjugate(5) and HPMA copolymer–Dox–AGM conjugate (4) (both6.3 μg/mL Dox-equiv.) were added and the cells incubated for 0to 60 min either at 37 °C or 4 °C. Note that for 0 min incubation,the conjugate was added to the cells and immediately removed.At each sample time, cells were placed on ice, then washed

three times with ice-chilled PBS (5 mL), and PBS (1 mL) wasadded before the cells were scraped from the plate with a rubberpoliceman and collected in falcon tubes. The cell suspensionwas centrifuged at 4 °C (600×g for 10 min). The cell pellet wasre-suspended in ice-chilled PBS (200 μL). Cell-associatedfluorescence was then analysed using a Becton DickinsonFACSCalibur cytometer (CA, USA) equipped with an argonlaser (488 nm) and emission filter for 550 nm. Data collectioninvolved 25,000 counts per sample, and the data were analysedusing CELLQuest™ version 3.3 software. Data are expressed byplotting the shift in geometric mean of the entire population (i.e.geometric mean after incubation with the polymer−geometricmean in absence of polymer). Cells incubated without polymerconjugate were used to account for the background fluorescence.

2.4. Evaluation of the cytotoxicity of endocytosis inhibitors

Before using the endocytosis inhibitors it was important todefine their general toxicity. MCF-7 and MCF-7ca were seededin 24 well plates (3.2×105 cells/mL; 500 μL per well) inWRPMI+5% SFCS. The seeding density was adjusted to obtainthe same concentration (cells/area) used in the flow cytometryexperiments. After 24 h, cells were washed with warm (37 °C)PBS (500 μL) to remove dead cells and residual serum. Theinhibitors were added in medium (500 μL) using the followingconcentrations: a combination of MβCD (0–15 mM) andlovastatin (1 μM); chlorpromazine (0–50 μM); or cytochalasinB (0–25 μM). After 2 h, themediumwas removed, the cells werewashed with warm (37 °C) PBS (3×500 μL), and then, 150 μLof a Trypan blue solution (0.2% Trypan blue in PBS) was addedto each well. The total number of stained cells in the visual fieldwas counted (SCvis). The total number of cells in the visual field(TCvis, i.e. stained and unstained) was determined applying agrid that allowed the counting of 25% of the cells in the visualfield. The data were expressed as % of dead cells as follows:

percentage dead cells ¼ ðSCvis=TCÞ � 100:

2.5. Evaluation of cellular uptake in presence of inhibitors ofendocytic pathways

MCF-7 and MCF-7ca were seeded as described above in 6-well plates (1×106 cells/mL). After 24 h, the cells were washedwith warm (37 °C) PBS and fresh medium (900 μL) was addedthat contained the endocytosis inhibitor; MβCD+lovastatin,chlorpromazine or cytochalasin B. The solutions were preparedstarting from stock solutions of the inhibitor (50 mg/mL and10 mg/mL, 10 mg/mL or 2 mg/mL, respectively). The finalconcentrations used were 10 mM+1 μM forMβCD+lovastatin,25 μM for cytochalasin B and 15 μM for chlorpromazine,respectively. In each case it was ensured that solvents were alsoused at non-toxic concentrations.

After a 1-h pre-incubation with the inhibitors at 37 °C, HPMAcopolymer–Dox (5) or HPMA copolymer–Dox–AGM (4)(100 μL in medium) was added (final concentration 6.3 μg/mLDox-equiv.). After 1 h the cells were placed on ice, and preparedfor FACS as described above.

31F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

2.6. Confocal fluorescent microscopy: live-cell imaging

For live-cell imaging, MCF-7 and MCF-7ca cells(1×106 cells) were seeded in glass bottom culture dishes andleft to adhere to the coverslips for N12 h. To assess the subcellularlocalisation of the polymers, cells were incubated for either15 min or 1 h at 37 °C in complete medium (1 mL) containingHPMA copolymer–Dox±AGM conjugates (6.3 μg/mL Dox-equiv.). Before visualisation, medium containing the polymerswas removed from the cells by aspiration, and the cells werewashed three times with fresh medium (37 °C). Finally, clearcomplete media was added and cells were subsequently viewedfor a maximum of 30 min. Images were captured with an invertedDM IRE2 microscope equipped with a λ-blue 60× oil immersionobjective and handled with a TCS SP2 system, equipped with anAcoustic Optical Beam Splitter (AOBS). Excitation was with anargon laser (548 nm, 476 nm, 488 nm, 496 nm and 514 nm) andblue diode (405 nm). Images were captured at an 8-bit grey scaleand processed with LCS software (version 2.5.1347a, LeicaGermany) containing multicolour, macro and 3D components.

2.7. Release of AGM and Dox from HPMA copolymerconjugates during incubation with isolated rat liver lysosomalenzymes

Tritosomes were prepared in sucrose according to themethod of Trouet [21]. Protein content was determined usingbicinchoninic acid protein assay (BCA) and proteolytic activitywas determined by measuring the release of p-nitroanilide(NAp) from N-benzoyl-Phe-Val-Arg–NAp. In this case, thetritosomes used had a protein content of 2.3 mg/mL and anactivity of 62.4 nm/min/mg protein.

HPMA copolymer–Dox±AGM conjugates (50 μg/mLdrug-equiv.) were incubated (37 °C) in citrate phosphate buffer(pH 5.5, 1 mL total volume) with Triton X-100 (0.2% w/v),EDTA (100 μL, 10 mM) and GSH (100 μL, 50 mM) [22]. Tobegin the degradation study tritosomes were added (300 μL)and the Eppendorf tubes thoroughly mixed. Aliquots (100 μL)were taken at times up to 24 h, immediately frozen in liquidnitrogen, and stored frozen in the dark until assayed by HPLC.In control experiments conjugates were incubated in bufferalone (without addition of tritosomes) to assess non-enzymatichydrolytic cleavage. In addition, free drug (1 or 2) (50 μg/mL)was also incubated under same conditions and later used as thereference control in the HPLC assay.

2.7.1. Determination of AGM or Dox release by HPLCSamples were defrosted and added to polypropylene tubes

and made up to 1 mL with water. The pH of the samples wasadjusted to 8.5 with ammonium formate buffer (100 μL, 1 M,pH 8.5), and then a mixture of chloroform/2-propanol at a ratioof 4:1 (5 mL) was added. Dnm was used as internal referencestandard; 100 μL of a 1 μg/mL stock aqueous solution wasadded to each sample. Samples were then thoroughly extractedby vortexing (3×10 s). The upper aqueous layer was carefullyremoved and the solvent was evaporated under N2. The dryresidue was dissolved in 100 μL of HPLC grade methanol. In

parallel, the same procedure was carried out for the parentcompounds 1 and 2 (using 100 μL of a 1 mg/mL stock aqueoussolution) to construct a standard curve. Addition of 1 mL ofmethanol to redissolve the product gave a 100 μg/mL stockfrom which a range of concentrations were prepared (2 to 60 μg/mL). The amount of drug released from the conjugates wasdetermined by HPLC using a μBondapak C18 column(150×3.9 mm), with a flow rate of 1 mL/min and using agradient elution [solvent A: 2-propanol/H2O 12:88 (v/v),solvent B: 2-propanol/H2O 29:71 (v/v)] adjusted to pH 3.2with o-phosphoric acid. Total run time was 20 min and thegradient profile was: t=5 min A 100%, t=9 min A 0%,t=14 min A 0%, t=16 min A 50%, t=18 min A 100% andt=20 min A 100%). To monitor AGM (1) an UV detector(Spectroflow 783 Kratos analytical) with a fixed-wavelengthfilter (254 nm) was used. To monitor Dox (2) and the Dnmstandard a Fluoromonitor III fitted with interference filters at485 nm for excitation and 560 nm for emission was used. Theretention time was 4.67 min for AGM and 12.48 min for Dox.

2.8. Investigating the ability of AGM to inhibit tritosomes andcathepsin B

The activity of tritosomes and cathepsin B against the N-benzoyl-Phe-Val-Arg–NAp substrate were determined by UVas described above. To study the effect of AGM on enzymeactivity, AGM (25–200 μg/mL in DMSO, 10 μL) was added tothe incubation mixture. Addition of DMSO alone (10 μL) wasused as a control. A known inhibitor of the thiol-dependentproteases, leupeptin (50 μg/mL) was used as a positive control.As above, the tritosomes or cathepsin B were warmed to 37 °C(30 μL) and added last to start the reaction. Release of NAp at410 nm was measured by UV over 20 min.

2.9. Immunocytochemistry

The expression of the proliferation protein ki67 and of theanti-apoptotic protein bcl-2 was evaluated by immunocyto-chemistry in MCF-7 and MCF-7ca cells in the presence andabsence of drugs and conjugates. The experimental protocolwas designed to resemble, as closely as possible, the ex-perimental conditions used for the determination of conjugatecytotoxicity as described above. Both cell lines were seeded insterile 12-well plates that contained one sterile tespa-coatedround coverslip (16 mm diameter) per well in WRPMIsupplemented with 5% of SFCS. The seeding density was1.66×105 cells/mL (0.5 mL/well) and the cells were allowed toadhere for 24 h. Then, the medium was replaced with freshmedium that contained 5% SFCS±10−9 M oestradiol. After5 days the medium was replaced with fresh medium thatcontained either HPMA copolymer–Dox–AGM or HPMAcopolymer–Dox (0.2 μm filter sterilized) at a Dox con-centration that corresponded to the IC50 values determined forthe HPMA copolymer–Dox–AGM (IC50 in MCF-7 75 or21 μg/mL Dox-equiv. with or without oestradiol, respectively;in MCF-7ca 8.2 or 12 μg/mL Dox-equiv. with or withoutoestradiol, respectively).

Fig. 2. Comparison of the cytotoxicity HPMA copolymer–Dox–AGM andconjugate mixtures in MCF-7 and MCF-7ca cells. Panel (a) shows cytotoxicityin MCF-7; panel (b) shows cytotoxicity in MCF-7ca; panel (c) the aromataseexpression in both cell lines. n, Dox; □, HPMA copolymer–Dox–AGM; ●,HPMA copolymer–Dox+HPMA copolymer–AGM; ○, HPMA copolymer–Dox. Data are represented as mean±S.E.M.; n=3.

32 F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

On day 8 (after 72 h of incubation with the polymer con-jugates), the cells were fixed as follows. The coverslips wereplaced in formaldehyde solution (4% in PBS) for 15 min atroom temperature, then into PBS for 5 min room temperature.Subsequently, they were transferred to methanol (5 min) andacetone (3 min) which were kept on dry ice to ensure that thetemperature of the solvents was maintained between −30 °Cand −10 °C. Finally there was a PBS wash (5 min at roomtemperature) and the coverslips were stored in a sucrose storagesolution (42.8 g sucrose, 0.33 g magnesium chloride anhydrous,250 mL PBS and 250 mL of glycerol) at −20 °C.

2.9.1. Detection of Ki67The storage mediumwas removed by thorough washings with

PBS followed by a wash with PBS–Tween (0.02%). Then, thesolution containing the primary antibody (against Ki67 antigen,clone MIB-1) (1:75 in PBS) was added (30 μL per coverslip)before incubation at room temperature for 60 min. Then, theprimary antibody was removed by repeated washing with PBS–Tween and the secondary antibody added (Biogenex mouse link1:50 in BSA–PBS (1%); 50 μL per coverslip). After 30 min atroom temperature, the coverslip was washed with PBS–Tweenand the tertiary antibody was added (Biogenex ConcentratedLabel dilution 1:50 in BSA–PBS (1%); 50 μL per coverslip)followed by a further 30 min incubation at room temperature.Finally, the coverslips were washed with PBS–Tween andincubated with the peroxidase substrate and DAB for 10 min.After thorough washing with dH2O and the cells were counter-stained with a solution of methyl green (0.5% in dH2O) for 30 s atroom temperature, washed again with dH2O and allowed to airdry before mounting onto a glass slide using DPX mountant.

2.9.2. Detection of bcl-2The storage medium was removed by thorough washings

with PBS followed by a wash with PBS–Tween (0.02%). Then,a solution of normal goat serum (5%) and normal human serum(5%) in PBS was added (50 μL/coverslip) and after 10 min theexcess was then removed. The primary antibody (against bcl-2oncoprotein, clone 124) solution was prepared pre-absorbing1 μL of primary antibody with 2 μL of normal human serum for90 min and then adding 747 μL of 0.1% BSA in PBS. Theprimary antibodywas added (30 μL/coverslip) and left overnightbefore repeated washing with PBS–Tween. Then the secondaryantibody, goat–anti-mouse (1:35 in PBS; 50 μL/coverslip) wasadded and left for 30min before another wash with PBS–Tween,addition of the tertiary antibody (peroxidase anti–peroxidase;1:250 in PBS; 50 μL/coverslip), and left for another 30-minincubation. Then the coverslips were washed with PBS–Tween,and a solution containing the peroxidase substrate and DAB wasadded and left for 10 min. Finally, the coverslips were washedwith dH2O, and the cells were counterstained with a solution ofmethyl green (0.5% in dH2O) for 30 s at room temperature. Aftera final wash with dH2O they were allowed to air dry before beingmounted onto a glass slide.

In each case the immunostaining was assessed by two in-dependent observers. Data are expressed as a percentage of cellswith positive staining.

2.10. Statistical analysis

When only two groups were compared, the Student's t testfor small sample size was used to estimate statistical sig-nificance. If more than two groups were compared evaluation ofsignificance was performed using one-way analysis of variance(ANOVA) followed by Bonferroni's post hoc test. In all cases,statistical significance was set at pb0.05.

3. Results

The synthesis of the HPMA copolymer conjugates (3, 4, 5)has been described elsewhere [18,19], but the characteristics ofthe specific samples prepared for these studies are given inFig. 1. Earlier studies using MCF-7 and MCF-7ca cells clearlyshowed an enhancement of cytotoxicity in vitro when AGMand Dox were conjugated within the same polymeric backbonecompared to the effects seen when HPMA copolymer–Dox orHPMA copolymer–AGM (or a simple mixture of both were

Fig. 3. Cytotoxicity of Dox, AGM and HPMA copolymer conjugates and theirmixtures in MCF-7 and MCF-7ca cells incubated in the presence and absenceof oestradiol. n, MCF-7 plus oestradiol; □, MCF-7 without oestradiol; ●,MCF-7ca plus oestradiol; ○, MCF-7ca without oestradiol. Data are representedas mean±S.E.M.; n=3.

Table 1Cytotoxicity of Dox andAGMandHPMAcopolymer conjugates againstMCF-7and MCF-7ca human breast cancer cell lines a

Entry Compound(s) Dox (IC50)b AGM (IC50)

c

MCF-7 MCF-7ca MCF-7 MCF-7ca

1 AGM ( 1) NA NA 1223±274 1043±4412 Dox ( 2) 0.4±0.2 22±8.5 NA NA3 Dox ( 2)+AGM ( 1) 0.4±0.1 16±7.6 7.4±1.7 308±1514 HPMA copolymer–

AGM ( 3)NA NA N180 d N180 d

5 HPMA copolymer–Dox ( 5)

N126 d N126 d NA NA

6 HPMA copolymer–Dox ( 5)+AGM ( 1)

62±12 60±8.9 988±177 967±145

7 HPMA copolymer–Dox ( 5)+HPMAcopolymer–AGM ( 3)

N157 d N157 d N151 d N151 d

8 HPMA copolymer–Dox–AGM ( 4)

75±45 8.2±3.1 77±45 8.8±3.5

a Cell viability 72 h MTT assay, seeding density 4×104 cells/mL for theexperiments carried out in the presence of 10−9 M oestradiol in medium.Published as supplementary information in Vicent et al. [18].b Data expressed in μg/mL of Dox-equiv. mean±S.E.M.; n=3.c Data expressed in μg/mL of AGM-equiv. mean±S.E.M.; n=3.d Maximum concentration tested.

33F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

added (preliminary supplementary information in [18]). Furtherexperiments reported here confirm the uniqueness of this effect(Figs. 2 and 3, and Table 1). The HPMA copolymer–Dox–AGM conjugate (4) was more toxic in both MCF-7 and MCF-7ca cells than HPMA copolymer–Dox (5) and a mixture ofHPMA copolymer–Dox (5)+HPMA copolymer–AGM (3)(Fig. 2), and activity was similar to free Dox in MCF-7cacells, an effect which correlates with the higher aromatase levelsin these cells.

The cytotoxicity of all free drugs, HPMA copolymer con-jugates and their mixtures against MCF-7 and MCF-7ca cellscultured in the presence and absence of oestradiol (to see if thecytotoxicity of these drugs was affected by the presence of thishormone) is shown in Fig. 3 and Table 1). Free AGM andHPMA copolymer–AGM showed little cytotoxicity in eithercell line under the experimental conditions used. As expected,Dox was cytotoxic, being ∼55-fold more active in MCF-7 thanMCF-7ca cells. Addition of Dox+AGM did not change the IC50

values obtained when expressed as Dox-equiv. indicating nosynergistic effect. Neither addition of AGM (1), nor HPMAcopolymer–AGM conjugate (3) to HPMA copolymer–Dox (5)caused an increase in cytotoxicity compared to the profile seenfor the individual compounds (Fig. 3). For example, in MCF-7ca cells the IC50 value of 1+5 was 967 μg/mL (AGM-equiv.)compared to 1043 μg/mL for (1) alone. HPMA copolymer–Dox–AGM (4) displayed greater activity than HPMA copol-ymer–Dox (5), and this enhanced activity was most evident inMCF-7ca cells where the IC50 value seen (Dox-equiv.) wassimilar to that seen for free Dox (Table 1).

3.1. Comparison of the endocytic properties of HPMAcopolymer–Dox–AGM and HPMA copolymer–Dox

Prior to flow cytometry studies, the fluorescence character-istics of HPMA copolymer–Dox (5) and HPMA copolymer–Dox–AGM (4) were determined to ensure that correctcomparisons would be made. Although the fluorescence ofneither conjugate was affected by pH, HPMA copolymer–Doxhad a higher fluorescence output (∼1.3- to 1.4-fold) than con-jugate (4) across the Dox concentration range studied (Fig. 4a).

Fig. 4. Fluorescence characteristics of HPMA copolymer conjugates and theircell binding and endocytic uptake. Panel (a) shows the pH and concentrationdependence of Dox fluorescence in conjugates 4 and 5. Panel (b) showsrepresentative data in flow cytometry experiments collected after incubation ofMCF-7ca cells with HPMA copolymer–Dox. Panels (c) and (d) show time-dependent binding (4 °C) and uptake (37 °C) of HPMA copolymer–Dox andHPMA copolymer–Dox–AGM by MCF-7 and MCF-7ca cells, respectively.□,HPMA copolymer–Dox at 37 °C; ○, HPMA copolymer–Dox at 4 °C; n,HPMA copolymer–Dox–AGM at 37 °C; ●, HPMA copolymer–Dox–AGMat 4 °C. Inserts show magnification of the data obtained with the combinationconjugate. Data are represented as mean±S.E.M.; n≥3. ⁎pb0.05. Fig. 5. Live-cell confocal microscopy images of MCF-7 and MCF-7ca cells

incubated (37 °C) with conjugates (4) and (5). In each case a conjugateconcentration of 6.3 μg/mL Dox-equiv. was used. The white arrows indicatemembrane binding, blue arrows indicate vesicular localisation. Scalebar=20 μm. (For interpretation of the references to colour in this figure legend,the reader is referred to the web version of this article.)

34 F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

Flow cytometry and confocal microscopy experiments con-ducted at equi-concentrations of Dox-equiv. should always beinterpreted with this in mind.

Addition of HPMA copolymer conjugates to both MCF-7and MCF-7ca cells led to an immediate fluorescence shift thatprogressed further during the 1 h incubation (Fig. 4b). Bothconjugates (3 and 4), in both cell lines, showed higher cell-associated fluorescence at 37 °C than at 4 °C (Fig. 4c,d).Although the initial increase in cell association was rapid (withinminutes), there was a much lower time-dependent increasethereafter (10–60 min) (Fig. 4c,d). Incubation with HPMAcopolymer–Dox (5) resulted in higher cell-associated fluores-cence than was observed for the combination conjugate (4) atboth temperatures and in both cell lines, but this have beenpartially attributed to the higher fluorescence output of theconjugate. Live-cell imaging revealed most fluorescenceassociated to the plasma membrane at 15 min (data notshown), and in agreement with the flow cytometry studies, thiswas more marked for conjugate (5) than (4). Even at 15 min, itwas possible to see HPMA copolymer–Dox–AGM (4) invesicles withinMCF-7 cells, and these were more numerous andprominent than could be seen for (5). After 1 h, vesicularlabelling was much clearer for both conjugates, and similarobservations could be seen with the MCF-7ca cell line (Fig. 5).

A number of inhibitors of endocytosis were selected toinvestigate whether conjugates (4) and (5) would demonstratedistinct mechanisms of uptake. It is necessary to take care thatsuch inhibitors are always used at non-toxic concentrations. InMCF-7 cells, all the compounds showed a dose-dependenttoxicity (Fig. 6a,b). MβCD caused less than 20% cell death atconcentrations ≤15 mM. Cytochalasin B toxicity never causedN10% cell death, even at the maximum concentration tested(25 μM), and chlorpromazine caused ∼100% cell death at

35F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

50 μM. Results obtained in MCF-7ca cells showed a similarpattern of toxicity, and again chlorpromazine was the most toxic(∼30% cell killing at 50 μM). As the inhibitor concentrationscommonly used in the literature are 10 mM for MβCD, 15 μMfor chlorpromazine and 25 μM for cytochalasin B (all caused≤10% cell killing) these concentrations were adopted forfurther studies.

After a 1 h incubation, MβCD decreased the uptake of bothHPMA copolymer–Dox (5) and HPMA copolymer–Dox–AGM (4) in MCF-7 by ∼60% and ∼50%, respectively. Asimilar trend (although at a lower extent) was seen in MCF-7cawere the uptake of both conjugates was decreased by ∼25%.Incubation with chlorpromazine or cytochalasin B decreasedonly the uptake of HPMA copolymer–Dox (5) in MCF-7(∼50% and ∼20%, respectively) but did not have any effect onthe uptake of the other conjugate or in MCF-7ca. Overall, asimilar response to the presence of these inhibitors was seenfor both cell lines and for both conjugates. Chlorpromazineinhibited HPMA copolymer–Dox (∼60%) uptake in MCF-7cells. However, the inhibitors used (at these concentrations) alsohad a similar effect on both conjugates (4) and (5).

Fig. 6. Cytotoxicity of inhibitors of endocytosis and their effect (at non-toxicconcentrations) on the uptake of HPMA copolymer conjugates (4) and (5) byMCF-7 and MCF-7ca cells. Panels (a) and (b) show inhibitor concentration-dependent toxicity. Panels (c) to (f) show the effect of inhibitors on the uptakeof conjugate (6.3 μg/mL Dox-equiv.) when cells were incubated for 1 h. n,Control;□, MβCD; , chlorpromazine; , cathepsin B. Data are represented asmean±S.E.M., n=6. ⁎pb0.05.

Fig. 7. Drug release from HPMA copolymer conjugates on incubation with ratliver lysosomal enzymes (tritosomes). Panel (a) shows release of Dox fromconjugate (5) in the presence and absence of conjugate (3); panel (b) shows therelease of AGM from conjugate (3) in the presence and absence of conjugate (5),and panel (c) shows both Dox and AGM release from the HPMA copolymer–GFLG–Dox–AGM conjugate (4). In all cases, drug release is expressed as apercentage of the total drug bound, and data are represented as mean±S.E.M.(n=4). The release of Dox from HPMA copolymer–Dox (panel a) and therelease profiles reported in panel (c) have been published previously in Vicentet al. [18].

3.2. Drug release from HPMA copolymer conjugates

As anticipated, tritosomes released AGM and Dox from allthe HPMA copolymer conjugates containing the Gly-Phe-Leu-Gly polymer–drug linker. No degradation occurred in theabsence of enzymes (e.g. Fig. 7b). The drug-release profiledisplayed marked differences depending on conjugate compo-sition. The combination conjugate HPMA copolymer–Dox–AGM released between 15–20% AGM and ∼20% of Dox over5 h, and unexpectedly release was not linear with time (Fig. 7c).Release of both Dox and AGM had a lag phase with littlerelease over the first 30 min, but after this time AGM initialrelease (to a plateau) was much faster than Dox release whichcaught up over 5 h.

To understand the reason for these unusual kinetics, therelease of Dox from conjugate (5) was measured in the presenceand absence of HPMA copolymer–AGM (3) (Fig. 7a). In bothcases, Dox release began immediately after addition of enzymeand was linear with time. Although the presence of (3) reducedthe extent of Dox release this was not surprising due to increaseof effective substrate concentration. The release of AGM fromconjugate (3) was also measured in the presence of conjugate(5). In both cases, AGM was initially released quickly, but after∼200 min started to plateau. The reason for this behaviouris not clear as active enzyme is available over the whole

Table 2AGM inhibition of tritosomes and cathepsin B

AGM(μg/mL)

Tritosomes Cathepsin B

Activity a

(nM/min/mg)Inhibition a (%) Activity b

(nM/min/mg)Inhibition b

(%)

0 37 0 91±12 0±025 24 34 75±5 8±750 21 42 78±8 4±3200 25 32 78±10 4±5Leupeptin(50 μg/mL)

0 100 0±0 100±0

a Data expressed as mean of n=2 experiments.b Data expressed as mean±S.E.M. (n=3).

36 F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

incubation period. One explanation could be that the AGMreleased causes enzyme inhibition.

To test this hypothesis, cathepsin B and tritosomes wereincubated with Bz–ValArgPhe–NAp as substrate in the presenceand absence of AGM, using the inhibitor leupeptin as positivecontrol. Studies on the kinetics of degradation by tritosomesgave a Vmax=0.107 Abs/min and Km=0.3 μM. Addition ofAGM (a concentration 50 μg/mL corresponds to the maximumrelease of AGM in Fig. 8b,c) caused ∼30–40% inhibition oftritosome activity, but only 4–8% inhibition of cathepsin Bactivity (Table 2). AGM inhibition was not dependent onconcentration, and lower than that observed for leupeptin.

3.3. Immunocytochemistry

To investigate whether HPMA copolymer–Dox–AGM (4)and HPMA copolymer–Dox (5) had different effects on cellproliferation, immunocytochemistry was used to study ki67 and

Fig. 8. Immunocytochemistry showing the effect of HPMA copolymerconjugates (4) and (5) on the expression of ki67 in MCF-7 and MCF-7cacells. Cells were incubated in the presence of oestradiol as shown. The blackarrows indicate nuclear staining, red arrows nucleoli staining and white arrowsunstained cells. Scale bar=10 μm. (For interpretation of the references to colourin this figure legend, the reader is referred to the web version of this article.)

bcl-2 expression in MCF-7 and MCF-7ca cells followingincubation with each conjugate. In MCF-7ca cells, HPMAcopolymer–Dox caused a decreased in expression of ki67(Fig. 8b), and HPMA copolymer–Dox–AGM decreasedexpression even further (Figs. 8c and 9). Similar trends wereseen in MCF-7 cells (Figs. 8d–f and 9), and also in absence ofoestradiol (Fig. 9).

Levels of bcl-2 were used to assess the effect of both twoconjugates (4 and 5) on apoptosis. There was no difference inexpression in MCF-7 cells incubated with or without HPMAcopolymer–Dox (Fig. 10). However, MCF-7 cells incubatedwith HPMA copolymer–Dox–AGM showed decreased expres-sion of bcl-2 (∼30% reduction) (Fig. 10c,d). A similar trendwas found in MCF-7 cells in absence of oestradiol (Fig. 10d).Assessment of bcl-2 in MCF-7ca led to poor staining (b10%) in

Fig. 9. Expression of ki67 inMCF-7 andMCF-7ca cells.□, Control;n, control+oestradiol; , HPMA copolymer–Dox; , HPMA copolymer–Dox–AGM. Notethat the set of results on the left was performed in absence of oestradiol, while theset of results on the right was performed in presence of oestradiol (10−9 M). Dataare represented as mean±S.E.M.; n≥8. ⁎pb0.05.

Fig. 10. Effect of HPMA copolymer conjugates (4) and (5) on the expression ofbcl-2 in MCF-7 cells. Cells were incubated in the presence or absence ofoestradiol as shown. The key for panel (d) is as shown in Fig. 9. The values inpanel (d) are represented as mean±S.E.M.; n≥6. ⁎pb0.05.

Fig. 11. Proposed mechanism of action of the HPMA copolymer–Dox–AGM.

37F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

all cases (data not shown), therefore investigation of the impactof the conjugates on this marker was impossible for this cellline.

4. Discussion

The last decade has seen significant advances in the manage-ment of breast cancer that are attributable to earlier diagnosis, theemergence of endocrine therapy (e.g. tamoxifen and aromataseinhibitors), and most recently the introduction of Herceptin®.However, inherent and acquired resistance is still a significantproblem. It is increasingly clear that monotherapy is unlikelyto provide a long-term solution for the treatment of metastaticand/or resistant disease [26]. Conventional chemotherapy isroutinely used as combination therapy and recent studies areshowing the potential of polymer–protein [27] and polymer–drug conjugates when combined with another chemotherapeuticagent [28,29]. Furthermore, HPMA copolymer conjugatescarrying both chemotherapy (Dox) and phototherapy havebeen described that have increased antitumour activity againstOVCAR-3 xenografts in mice [30].

A polymeric carrier provides an ideal platform for the si-multaneous delivery of a cocktail of drugs, and we recentlyreported the first endocrine–chemotherapy combination in theform of the model compound HPMA copolymer–Dox–AGM(4) [18]. Although this conjugate displayed markedly enhancedcytotoxicity compared to HPMAcopolymer–Dox (5), its precisemechanism of enhanced activity remains unclear. AGM releasedfrom the polymer conjugate can inhibit aromatase in MCF-7cells [19], and HPMA copolymer–Dox can exhibit clinical

antitumour activity in heavily pretreated (even anthracycline-resistant) breast cancer patients [7]. The inability of AGM+Dox,and all other permutations of non covalent mixtures of drugs andconjugates studied here to show synergistic effects in theMCF-7cell lines underlines the uniqueness of the combination con-jugate. Hypothetical opportunities for differences in themechanism of action of HPMA copolymer–Dox and HPMAcopolymer–Dox–AGM conjugate are summarised schematical-ly in Fig. 11.

Although direct comparison of binding and uptake of con-jugates 4 and 5 proved more challenging than expected dueto their differences in fluorescence yield (Fig. 4a), it is clearthat both conjugates displayed the same pattern of cell asso-ciation. If anything, HPMA copolymer–Dox–AGM was cap-tured more slowly (Fig. 4c,d) and certainly not fast enough toaccount for the marked increase in cytotoxicity compared toHPMA copolymer–Dox (Fig. 2). Surprisingly in the MCF-7cell lines both conjugates showed very high binding. Althoughlive-cell confocal microscopy confirmed vesicular uptake after15 min, the internalisation measured by FACS was relativelyslow thereafter. Inhibitors of endocytic entry pathways wereused to see whether the conjugates were using distinctly dif-ferent entry portals. Reduced uptake of both conjugates 4 and 5seen in presence of MβCD (Fig. 6) suggested, at least in part,internalisation via a cholesterol-dependent pathway [31].Clathrin- and caveolin-dependent uptake require cholesterol[32–34], but as wild type MCF-7 cells do not express a detect-able amount of caveolin [35] this uptake may be via coatedvesicle pathway. Further studies with chlorpromazine (thiscompound is claimed to inhibit only clathrin-dependent uptake[31,36,37]) resulted in some inhibition of uptake (but less than

38 F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

seen for MβCD). Cytochalasin B only caused significantinhibition of HPMA copolymer–Dox in MCF-7 cells but theextent of inhibition was relatively small. All that can be con-cluded from these studies is that conjugates 4 and 5 did notshow marked differences in uptake mechanism.

Both 4 and 5 were designed using a Gly-Phe-Leu-Gly linkerto ensure lysosomotropic Dox delivery [38]. It has been shownpreviously that linker degradation by thiol-dependent proteases(particularly cathepsin B) are a prerequisite for HPMA co-polymer conjugate activity (reviewed in [9,19]). The lysosomalenzyme-mediated drug release profiles for conjugates 3, 4 and 5were strikingly different (Fig. 7). Failure of either conjugate(3 and 5) to modify the drug release profile of the otherwhen simply added to the mixture (Fig. 7a,b) underlinesthe uniqueness of the kinetics of release observed for thecombination conjugate 4. Differences in structure (or solutiondynamics) of the complex unimolecular micelles that are 3, 4and 5 could theoretically be responsible for the release patternobserved. The conjugates certainly have different Rg as definedby SANS (Fig. 1) [18] and the structural geometry of each mayguide enzyme access, to the monomers bearing drug (eitherAGM or Dox). Another physicochemical factor that couldtheoretically influence enzyme accessibility would be “blocky-ness” in structure, with polymer domains containing a highdensity of pendant drug molecules. Theory suggests this shouldnot be the case as the HPMA copolymer precursor is a randomcopolymerisation of HPMA monomer and monomer containingthe Gly-Phe-Leu-Gly–ONp side chain. Subsequent drugconjugation should also give a statistical distribution alongthe polymer chain. However, this cannot be guaranteed inpractice and further studies using block copolymers especiallyprepared with an AGM-containing block and a Dox-containingblock would help to elucidate further this question. The ratherstrange kinetics of AGM release in both 3 and 4 (drug liberationreaching a plateau) might theoretically be explained by AGM–enzyme inhibition. However, AGM showed little inhibition ofcathepsin B, and although it did inhibit tritosome activity thiswas not concentration-dependent (Table 2).

Although HPMA copolymer–Dox was synthesised morethan 20 years ago [39] and its activity is well established both inanimal models [40] and in clinical setting (reviewed in [9,10]),its precise mechanism of action at cellular level is still unclear[4,10]. This is not unusual for many anticancer agents. Thepreliminary immunocytochemical studies (Figs. 8 and 9)showed decreased expression of ki67 following incubation ofMCF-7 and MCF-7ca cells with conjugates 4 and 5. Thiswas most marked for the combination polymer. As ki67 is awell described marker for the proliferating fraction of a cellpopulation [23] this is consistent with the higher cytotoxicityof 4. Bcl-2 expression following incubation of various cell typeswith HPMA copolymer–Dox has been investigated before,but the results obtained are conflicting. Some studies suggestdecrease in bcl-2 expression [41,42] while others did not seeany effect on the level of this protein [43]. The MCF-7 resultsreported here agree with the latter study as 5 had no effect onbcl-2. Decreased bcl-2 was however seen following incubationof MCF-7 with conjugate 4 suggesting that combined AGM and

Dox leads to a synergistic effect that induces apoptosis, hencethe increased activity of the combination polymer.

5. Conclusions

Complex cellular mechanisms seem to be responsible for theincreased antitumour activity of HPMA copolymer–Dox–AGM has in vitro. Differences in the drug release profile wereevident and this conjugate caused a significant change in theexpression of the anti-apoptotic protein bcl-2. Further studiesare needed to investigate these effects further, but it is likely thatin vivo testing will be required to better define both therapeuticpotential of HPMA copolymer–Dox–AGM conjugate and theexact mechanism of action.

Acknowledgements

F.G. and M.V. thank the Centre for Polymer Therapeutics,Tenovus Centre for Cancer Research, and Marie Curie (HPMF-CT-2002-01555) for supporting their work. The authors wouldalso like to thank Pauline Finlay, Sue Kyme, Iain Hutcheson andJan Knowlden for advice regarding immunocytochemistry andWestern blotting techniques.

References

[1] R. Duncan, The dawning era of polymer therapeutics, Nat. Rev. DrugDiscov. 2 (2003) 347–360.

[2] R. Satchi-Fainaro, R. Duncan, C.M. Barnes, Polymer therapeutics forcancer: current status and future challenges, Adv. Polym. Sci. 193 (2006)1–65.

[3] M.J. Vicent, R. Duncan, Polymer-based nanomedicines: novel treatmentsfor cancer, Trends Biotechnol. 24 (2006) 39–47.

[4] R. Duncan, Polymer–drug conjugates as anticancer nanomedicines, Nat.Rev. Cancer. 6 (2006) 688–701.

[5] J.W. Singer, S. Schaffer, B. Baker, A. Bernareggi, S. Stromatt, D. Niensted,M. Besman, Paclitaxel poliglumex (XYOTAX; CT-2103) [XYOTAX™]:an intracellularly targeted taxane, Anti-Cancer Drug 16 (2005) 243–254.

[6] Y. Matsumura, H. Maeda, A new concept for macromolecular therapies incancer chemotherapy: mechanism of tumouritropic accumulation ofproteins and the antitumour agent SMANCS, Cancer Res. 6 (1986)6387–6392.

[7] P.A. Vasey, S.B. Kaye, R. Morrison, C. Twelves, P. Wilson, R. Duncan,A.H. Thomson, L.S. Murray, T.E. Hilditch, T. Murray, S. Burtles, D.Fraier, E. Frigerio, J. Cassidy, Phase I clinical and pharmacokinetic studyof PK1 [N-(2-hydroxypropyl)methacrylamide copolymer doxorubicin]:first member of a new class of chemotherapeutic agents–drug–polymerconjugates, Clin. Cancer Res. 5 (1999) 83–94.

[8] T. Minko, P. Kopeckova, V. Pozharov, J. Kopecek, HPMA copolymerbound adriamycin overcomes MDR1 gene encoded resistance in a humanovarian carcinoma cell line, J. Control. Release 54 (1998) 223–233.

[9] R. Duncan, N-(2-Hydroxypropyl) methacrylamide copolymer conjugates,in: G.S. Kwon (Ed.), Polymeric Drug Delivery System, Marcel Dekker,Inc., New York, 2005, pp. 1–92.

[10] R. Duncan, Polymer–drug conjugates, in: D. Budman, H. Calvert, E.Rowinsky (Eds.), Handbook of Anticancer Drug Development, LippincottWilliams and Wilkins, Philadelphia, 2003, pp. 239–260.

[11] M.J. Vicent, S. Manzanaro, J.A. de la Fuente, C. Pérez, R. Duncan, HPMAcopolymer–1,5-diazaanthraquinone conjugates as promising anticancertherapeutics, J. Drug Target. 12 (2004) 503–515.

[12] R. Satchi-Fainaro, M. Puder, J.W. Davies, H.T. Tran, D.A. Sampson, A.Greene, G. Corfas, J. Folkman, Targeting angiogenesis with a conjugate ofHPMA copolymer and TNP-470, Nat. Med. 10 (2004) 225–261.

39F. Greco et al. / Journal of Controlled Release 117 (2007) 28–39

[13] D.M. Parkin, F. Bray, J. Ferlay, P. Pisani, Global cancer statistics, 2002, CACancer J. Clin. 55 (2005) 74–108.

[14] V.C. Jordan, Tamoxifen: a most unlikely pioneering medicine, Nat. Rev.Drug Discov. 2 (2003) 205–213.

[15] ATAC Trialists Group, Results of the ATAC (Arimidex, Tamoxifen Aloneor in Combination) trial after completion of 5 years adjuvant treatment forbreast cancer, Lancet 365 (2005) 60–62.

[16] F.J. Cummings, Evolving uses of hormonal agents for breast cancertherapy, Clin. Ther. 24 (2002) C3–C25.

[17] S.R.D. Johnston, M. Dowsett, Aromatase inhibitors for breast cancer:lessons from the laboratory, Nat. Rev., Cancer 3 (2003) 821–831.

[18] M.J. Vicent, F. Greco, R.I. Nicholson, A. Paul, P.C. Griffiths, R. Duncan,Polymer therapeutics designed for a combination therapy of hormone-dependent cancer, Angew. Chem., Int. Ed. 44 (2005) 2–6.

[19] F. Greco, M.J. Vicent, N.A. Penning, R.I. Nicholson, R. Duncan, HPMAcopolymer–aminoglutethimide conjugates inhibit aromatase in MCF-7cell lines, J. Drug Target. 13 (2005) 459–470.

[20] F. Searle, S. Gac-Breton, R. Keane, S. Dimitrijevic, S. Brocchini, R.Duncan, N-(2-Hydroxypropyl)methacrylamide copolymer–6-(3-amino-propyl)–ellipticine conjugates, synthesis, characterisation and preliminaryin vitro and in vivo studies, Bioconjug. Chem. 12 (2001) 711–718.

[21] A. Trouet, Isolation of modified liver lysosomes, Methods Enzymol. 31(1974) 323–329.

[22] R. Duncan, H.C. Cable, J.B. Lloyd, P. Rejmanova, J. Kopecek, Polymerscontaining enzymatically degradable bonds: 7. Design of oligopeptideside-chains in poly[N-(2-hydroxypropyl)methacrylamide] co-polymers topromote efficient degradation by lysosomal-enzymes, Makromol. Chem.184 (1983) 1997–2008.

[23] T. Scholzen, J. Gerdes, The Ki-67 protein: from the known to theunknown, J. Cell. Physiol. 182 (2000) 311–322.

[24] A. Burlacu, Regulation of apoptosis by bcl-2 family proteins, J. Cell. Mol.Med. 7 (2003) 249–257.

[25] F.S. Kenny, P.C. Willsher, J.M.W. Gee, R.I. Nicholson, S.E. Pinder, I.O.Ellis, J.F.R. Robertson, Change in the expression of ER, bcl-2 and MIB1on primary tamoxifen and relation to response in ER positive breast cancer,Breast Cancer Res. Treat. 65 (2001) 135–144.

[26] J.M. Gee, W.A. Howell, W.J. Gullick, C.C. Benz, R.L. Sutherland, R.J.Santen, L.-A. Martin, F. Ciardiello, W.R. Miller, M. Dowsett, P. Barrett-Lee, J.F.R. Robertson, S.R. Johnston, H.E. Jones, A.E. Wakeling, R.Duncan, R.I. Nicholson, Workshop on therapeutic resistance in breastcancer: impact of growth factor signaling pathways and implications forfuture treatment: Consensus Statement, Endocr.-Relat. Cancer 12 (2005)S1–S7.

[27] G. Molineux, The design and development of Pegfilgrastim (PEG-rmetHuG-CSF, Neulasta®), Curr. Pharm. Des. 10 (2004) 1235–1244.

[28] C.J. Langer, M.A. Socinski, H. Ross, K.J. O'Byrne, Paclitaxel poliglumex(PPX)/carboplatin vs paclitaxel/carboplatin for the treatment of PS2patients with chemotherapy-naïve advanced non-small cell lung cancer(NSCLC): a phase III study, Proc. Am. Soc. Clin. Oncol. (2005) 7011.

[29] T. Herzog, R.J. Barret, R. Edwards, F.B. Oldham, Phase II study ofpaclitaxel poliglumex (PPX)/carboplatin (C) for 1st line induction and

maintenance therapy of Stage III/IV ovarian or primary peritonealcarcinoma, Proc. Am. Soc. Clin. Oncol. (2005) 5012.

[30] C.M. Peterson, J.M. Lu, Y. Sun, C.A. Peterson, J.G. Shiah, R.C. Straight,J. Kopecek, Combination chemotherapy and photodynamic therapy withN-(2-hydroxypropyl)methacrylamide copolymer bound anticancer drugsinhibit human ovarian carcinoma heterotransplanted in nude mice, CancerRes. 56 (1996) 3980–3985.

[31] J. Rejman, V. Oberle, I.S. Zuhorn, D. Hoekstra, Size-dependentinternalization of particles via the pathways of clathrin and caveolae-mediated endocytosis, Biochem. J. 377 (2004) 159–169.

[32] S.D. Conner, S.L. Schmid, Regulated portals of entry into the cells, Nature422 (2003) 37–44.

[33] D. Hailstones, L.S. Sleer, R.G. Parton, K.K. Stanley, Regulation ofcaveolin and caveolae by cholesterol in MDCK cells, J. Lipid Res. 39(1998) 369–379.

[34] S.K. Rodal, G. Skretting, O. Garred, F. Vilhardt, B. van Deurs, K. Sandvig,Extraction of cholesterol with methyl-β-cyclodextrin perturbs formation ofclathrin-coated endocytic vesicles, Mol. Biol. Cell 10 (1999) 961–974.

[35] G. Fiucci, D. Ravid, R. Reich, M. Liscvitch, Caveolin-1 inhibitsanchorage-independent growth, anoikis and invasiveness in MCF-7human breast cancer cells, Oncogene 21 (2002) 2365–2375.

[36] S.H. Kee, E.-J. Cho, J.-W. Song, K.S. Park, L.J. Baek, K.-J. Song, Effectsof endocytosis inhibitory drugs on rubella virus entry into VeroE6 cells,Microbiol. Immunol. 48 (2004) 823–829.

[37] X. Sun, V.K. Yau, B.J. Briggs, G.R. Whittaker, Role of clathrin-mediatedendocytosis during vesicular stomatitis virus entry into host cells, Virology338 (2005) 53–60.

[38] C. De Duve, T. De Barsy, B. Poole, A. Trouet, P. Tulkens, F. Van Hoof,Lysosomotropic agents, Biochem. Pharmacol. 23 (1974) 2495–2531.

[39] J. Kopecek, P. Rejmanova, J. Strohalm, K. Ulbrich, B. Rihova, V. Chytry,R. Duncan, J.B. Lloyd, Synthetic polymeric drugs, EU Patent Application85309560.2, December 31st, 1985.

[40] R. Duncan, L.W. Seymour, K.B. O’Hare, P.A. Flanagan, S. Wedge, I.C.Hume, K. Ulbrich, J. Strohalm, V. Subr, F. Spreafico, M. Grandi, M.Ripamonti, M. Farao, A. Suarato, Preclinical evaluation of polymer-bounddoxorubicin, J. Control. Release 19 (1992) 331–346.

[41] T. Minko, P. Kopeckova, J. Kopecek, Preliminary evaluation of caspases-dependent apoptosis signalling pathways of free and HPMA copolymer-bound doxorubicin in human ovarian carcinoma cells, J. Control. Release71 (2001) 227–237.

[42] K. Kunath, P. Kopeckova, T. Minko, J. Kopecek, HPMA copolymer-anticancer drug–OV–TL16 antibody conjugates: 3. The effect of free andpolymer-bound adriamycin on the expression of some genes in theOVCAR-3 human ovarian carcinoma cell line, Eur. J. Pharm. Biopharm.49 (2000) 11–15.

[43] L. Kovar, K. Ulbrich, M. Kovar, J. Strohalm, T. Stastny, T. Etrych, B.Rihova, The effect of HPMA–copolymer-bound doxorubicin conjugateson the expression of genes involved in apoptosis signalling, J. Control.Release 91 (2003) 247–248.