Synthesis and Characterization of Carboxylic Acid Conjugated

Dec 3, 2010 - Departments of Pathology and Lab Medicine and Department of Chemistry, ... University of British Columbia, Vancouver, British Columbia, ...
0 downloads 0 Views 1MB Size
Biomacromolecules 2011, 12, 145–155

145

Synthesis and Characterization of Carboxylic Acid Conjugated, Hydrophobically Derivatized, Hyperbranched Polyglycerols as Nanoparticulate Drug Carriers for Cisplatin Lucy Ye,†,‡ Kevin Letchford,†,‡ Markus Heller,§ Richard Liggins,§ Dechi Guan,§ Jayachandran N. Kizhakkedathu,⊥ Donald E. Brooks,⊥,| John K. Jackson,‡ and Helen M. Burt*,‡ Faculty of Pharmaceutical Sciences, University of British Columbia, 2146 East Mall, Vancouver, British Columbia, Canada V6T 1Z3, Centre for Drug Research and Development, 2259 Lower Mall, University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z4, Centre for Blood Research and Departments of Pathology and Lab Medicine and Department of Chemistry, 2350 Health Science Mall, University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z3 Received September 10, 2010; Revised Manuscript Received November 10, 2010

Hyperbranched polyglycerols (HPGs) with hydrophobic cores and derivatized with methoxy poly(ethylene glycol) were synthesized and further functionalized with carboxylate groups to bind and deliver cisplatin. Low and high levels of carboxylate were conjugated to HPGs (HPG-C8/10-MePEG6.5-COOH113 and HPG-C8/10-MePEG6.5-COOH348) and their structures were confirmed through NMR and FTIR spectroscopy and potentiometric titration. The hydrodynamic diameter of the HPGs ranged from 5-10 nm and the addition of COOH groups decreased the zeta potential of the polymers. HPG-C8/10-MePEG6.5-COOH113 bound up to 10% w/w cisplatin, whereas HPG-C8/10MePEG6.5-COOH348 bound up to 20% w/w drug with 100% efficiency. Drug was released from HPG-C8/10MePEG6.5-COOH113 over 7 days at the same rate, regardless of the pH. Cisplatin release from HPG-C8/10-MePEG6.5COOH348 was significantly slower than HPG-C8/10-MePEG6.5-COOH113 at pH 6 and 7.4, but similar at pH 4.5. Release of cisplatin into artificial urine was considerably faster than into buffer. Carboxylated HPGs demonstrated good biocompatibility, and drug-loaded HPGs effectively inhibited proliferation of KU-7-luc bladder cancer cells.

Introduction Hyperbranched polyglycerols (HPGs) are highly and randomly branched polymers forming unimolecular nanoparticles. HPGs, synthesized via anionic ring-opening multibranching polymerization of glycidol using trimethylolpropane (TMP) as an initiator were first reported by Sunder et al.1 Attractive features of this synthesis are the relatively simple “one-pot” synthesis, good control over molecular weight, narrow polydispersity, and the ability to modify the HPGs with various functional groups such as alkyl (C8/10) chains into the core and methoxy poly(ethylene glycol) (MePEG) onto the surface.2,3 HPGs have been investigated for a number of biomedical applications including human serum albumin substitutes4 and MRI contrast agents.5 Furthermore, a number of researchers have modified the HPG structure to enable the loading of drugs by either conjugation6,7 or physical encapsulation for drug delivery purposes.8-10 Hydrophobically modified HPGs, originally designed as a substitute for human serum albumin, possess low intrinsic viscosity, excellent biocompatibility and the ability to bind fatty acids.4 Comprehensive studies of these HPGs, possessing a range of molecular weights and different extents of MePEG substitution, showed excellent blood compatibility, no evidence * To whom correspondence should be addressed. Tel.: (604) 822-2440. E-mail: [email protected]. † Made equal contributions. ‡ Faculty of Pharmaceutical Sciences. § Centre for Drug Research and Development. ⊥ Center for Blood Research and Department of Pathology & Lab Medicine. | Department of Chemistry.

of animal toxicity, and long circulation lifetimes.11 Hydrophobically modified HPGs have been shown to encapsulate the hydrophobic drug, paclitaxel, and release the drug in a controlled manner over 16 days.10 When administered via a urethral catheter into the bladders (intravesical instillation) of mice bearing orthotropic bladder tumors, paclitaxel loaded HPGs showed significantly greater inhibition of tumor growth compared to Taxol instillation.10 These studies confirmed the biocompatibility of these polymers and established the rationale for the use of HPGs as intravesical chemotherapeutic delivery systems. The anticancer drug cis-diamminedichloroplatinum (cisplatin) is one of the agents used in combined chemotherapeutic regimens routinely administered for the systemic treatment of muscle invasive bladder cancer.12 The drug is usually combined with methotrexate, vinblastine and doxorubicin (so-called MVAC) and is effective at controlling tumor growth and metastatic progression of advanced bladder cancer.13,14 However, systemic administration of cisplatin is associated with unpredictable and life threatening toxicities whereby comorbid conditions may actually prevent older patients from receiving the drug.13,15 At the time of diagnosis, the majority of bladder cancer cases are nonmuscle invasive or superficial bladder cancer and are often treated by the intravesical administration of chemotherapeutic agents. Intravesical administration of cisplatin for the treatment of superficial bladder cancer (>70% of cancer cases) in humans was associated with mild but unpredictable toxicities including dysuria, bladder irritation, hematuria, and some nausea with only minor evidence of any chemotherapeutic benefit.16,17 It was suggested that the limited efficacy associated with the intravesical route of administration might arise from the poor uptake of this hydrophilic drug into bladder tissues and binding of the

10.1021/bm101080p  2011 American Chemical Society Published on Web 12/03/2010

146

Biomacromolecules, Vol. 12, No. 1, 2011

drug to urine components.16 On the other hand, effective intravesical chemotherapy with cisplatin has been described whereby tailored dosing and combination therapy with immunomodulators prevented tumor recurrence with negligible toxicity except cystitis.18 Our group has shown inhibition of the growth of orthotopic bladder tumors in mice following the intravesical administration of cisplatin at 1-2 mg/mL.19 Although the use of these high concentration of cisplatin allowed for some penetration of the drug into the bladder tissue (approximately 0.2% of the total dose), there was also significant systemic uptake of the drug with serum concentrations reaching as high as 20 µg/mL. Clearly, the effective intravesical use of cisplatin is compromised by poor tissue uptake and the need to use high concentrations of the drug with the potential for both local and systemic toxicity. The constant dilution of the drug by urine further complicates any cisplatin-based regimen in the bladder, especially as urine components may bind and sequester the drug.16 Recently, Hwang et al. encapsulated cisplatin in a nanoemulsion controlled release system and following intravesical administration to rats achieved a significant increase in bladder tissue uptake and retention of the drug.20 It was suggested that the improved uptake of the drug might arise from the association of hydrophobic components of the nanoemulsions with the bladder tissue (better drug partitioning into the tissue) and by endocytosis of the drug-loaded nanodroplets. Several researchers have conjugated cisplatin to a variety of polymers to selectively accumulate the drug at tumor sites and limit its distribution to healthy tissue, thereby reducing systemic toxicities after intravenous injection.21-23 Kataoka and coworkers have made considerable contributions to this field, developing micellar systems based on poly(ethylene glycol)block-poly(amino acid) copolymers and currently have a formulation, termed NC-6004, in clinical trials.24,25 This system, composed of poly(ethylene glycol)-block-poly(glutamic acid), relies on the coordination of cisplatin to the polymer to provide the driving force for micellization and, hence, formation of nanoparticles. Similar to the majority of cisplatin polymeric nanoparticulate delivery systems, NC-6004 is intended for intravenous administration with the main benefit being decreased nephrotoxicity. We have previously described the effective complexation of cisplatin onto a commercially available HPG (PG2) based on conversion of the hydroxyl-terminated PG2 to carboxylic acid groups.26 The system allowed for drug loadings greater than 25% w/w and controlled release over a period of a few days, but the commercially sourced polymer could not be readily purified or well characterized. Our goal is to produce a controlled release delivery system, which may decrease the toxicity of cisplatin, based on literature reports of reduced toxicity of platinum compounds formulated as polymer complexes or nanoparticles.27 The objective of this work was to functionalize hydrophobically derivatized HPGs, originally designed for intravesical taxane delivery,10 with varying densities of carboxylic acid functional groups for the encapsulation and intravesical delivery of cisplatin for the treatment of superficial bladder cancer.

Experimental Section Materials. Trimethylolpropane (TMP), glycidol, octyl/decyl glycidyl ether (O/DGE or C8/10), potassium methylate, potassium hydride, dimethylaminopyridine, succinic anhydride, cisplatin, and 1,2 phenylenediamine (OPDA) were purchased from Sigma-Aldrich-Fluka (Oakville, ON, Canada) and used as received, except as noted. Glycidol (96%) and O/DGE (technical grade) were purified by vacuum distil-

Ye et al. lation and stored at 4 °C. All solvents were HPLC grade and were purchased from Fisher Scientific. MePEG 350 epoxide was synthesize in-house as described previously.2 Synthetic urine (Surine) was purchased from Dyna-Tek Industries (Lenexa, KS). Synthesis of HPG-C8/10-MePEG. The polymerization of O/DGE core modified HPGs was carried out according to protocols described in our previous report.11 Briefly, 1.2 g (8.9 mmol) of the initiator (TMP) was mixed with 1.5 mL (5.3 mmol) of 25% w/v potassium methylate solution in methanol and added to a three-neck round-bottom flask under argon atmosphere. The mixture was stirred at 105 °C for 60 min, after which methanol was removed under vacuum. A mixture of 13 mL (0.2 mol) of glycidol and 9 mL (13.4 mmol) of O/DGE was injected into the reaction vessel using a syringe pump at a rate of 1.4 mL/h and stirred at 68 rpm using an overhead stirring system. For the synthesis of HPG-C8/10-OH (without the MePEG), the reaction was continued for at least 1 h prior to termination. For the preparation of polymers capped with MePEG 350, after the injection of the mixture of glycidol and O/DGE, the reaction was continued for approximately 6 h followed by the addition of 0.1 mL (0.35 mmol) of potassium hydride. The mixture was stirred for 60 min, after which 10 mL of MePEG 350 epoxide was added using a syringe pump at a rate of 1.4 mL/h. In the synthesis, the “target” amount of MePEG 350 epoxide was added to the reaction mixture. For instance, 10 mL of MePEG 350 epoxide was used for the target 6.5 mol of MePEG per mole of HPG, based on a 6.5:1 mol ratio of MePEG 350 epoxide/TMP. Accordingly, the resulting HPG was denoted as HPG-C8/10-MePEG6.5. The stirring rate was then increased to 90 rpm and the reaction was continued at 105 °C for a minimum of 1 h. The final products (both with and without MePEG) were semitransparent gels. Unreacted O/DGE was removed by extraction with hexane, and the residual methanol was removed under vacuum. The final products were then dissolved in deionized water and dialyzed for 3 days against deionized water using cellulose acetate dialysis tubing (MWCO 10000 g/mol, Snakeskin, Thermo Fisher Scientific, Rockford, IL) with three water changes per day. Polymers were obtained by freeze-drying. Synthesis of HPG-C8/10-MePEG-COOH. The functionalization of C8/10 core-modified HPGs with carboxylic acid groups was carried out according to protocols described earlier by our group.26 For a typical reaction, 5.0 g of the HPG-C8/10-OH or HPG-C8/10-MePEG6.5 was dissolved in 100 mL of pyridine, and the solution was kept under a nitrogen atmosphere, followed by the addition of dimethylaminopyridine and succinic anhydride, which were adjusted according to the target amount of carboxylic acid groups on HPGs. For the synthesis of HPG with the highest amount of COOH groups, all available free hydroxyl groups were targeted for modification to carboxylates; therefore, an excess amount of dimethylaminopyridine (0.075 g, 0.61 mmol) and succinic anhydride (4.5 g, 45 mmol) were added to the reaction mixture. Through calculation of the theoretical moles of free hydroxyl groups, it was determined that there were 348 mols of free hydroxyl groups per mole of HPG and, thus, theoretically the same number of carboxyl groups per mole of HPG. Therefore, the resulting HPG was denoted as HPG-C8/10-MePEG6.5-COOH348. The use of lower amounts of dimethylaminopyridine (0.015 g, 0.12 mmol) and of succinic anhydride (0.9 g, 9 mmol) produced HPGs in which not all the free hydroxyls were targeted for modification. The theoretical number of carboxylate groups added to the HPG was determined through the calculation of the number of moles succinic anhydride added to the reaction mixture. Therefore, this low carboxylate containing HPG was denoted as HPGC8/10-MePEG6.5-COOH113. After addition of the dimethylaminopyridine and succinic anhydride, the solution was stirred using a magnetic stir bar overnight at room temperature. Deionized water (100 mL) was added to the flask and the mixture was kept stirring for 30 min. Solvents were removed by rotary evaporation with the periodic addition of water to enable better evaporation of pyridine by azeotropic distillation. The final products were dissolved in methanol and dialyzed against a mixture of 80:20 methanol/deionized water for 3 days using cellulose acetate dialysis tubing (MWCO 10000 g/mol, Spectrum Laboratories Inc.,

Polyglycerols as Nanoparticulate Drug Carriers Rancho Domunguez, CA). The dialysis medium was changed every 8 h, each time with a lower methanol concentration until during the final three stages, the dialysis medium was 100% water. Polymers were obtained by freeze-drying. 13 C NMR of HPG-C8/10-MePEG-COOH (400 MHz, methanol-d4) δC: 0 (tetramethylsilane, internal reference), 14.73 (CH3, alkyl on O/DGE), 23.92-33.24 (C(O)CH2CH2COOH), 48.51-49.86 (solvent, methanol-d4), 59.29 (CH3O-MePEG), 64.19-65.36 (-CH2OH, unreacted primary alcohol groups in polymer), 69.98-73.74 (-CH2-O-, -CH-Oin polymer), 78.93-80.14 (CH in polymer), 173.84-174.16 (C(O)CH2CH2COOH), 175.92 (C(O)CH2CH2COOH). Polymer Characterization. NMR spectra of HPG polymers were acquired using a 400 MHz Bruker Avance II+ spectrometer (Bruker Corporation, Milton, ON). Polymers were dissolved in DMSO-d6 or methanol-d4 (Cambridge Isotope Laboratories, Andover, MA). Onedimensional proton and carbon spectra were obtained, as well as twodimensional, multiplicity-edited heteronuclear single quantum coherence (HSQC), heteronuclear multiple-bond correlation (HMBC), and HSQCTOCSY (total correlation spectroscopy) NMR experiments. Chemical shifts were referenced to the residual solvent peak. Two-dimensional spectra were analyzed using Sparky (T. D. Goddard and D. G. Kneller, Sparky 3, University of California, San Francisco). The mole fractions of COOH on HPGs were estimated from HSQC data as follows: For each of the modifications, the peak corresponding to the four methylene protons was integrated and its integral corrected for the number of protons. This value was divided by the integral of the TMP methyl group (corrected for proton multiplicity) to yield the mole fraction of COOH. FT-IR spectra for HPGs were obtained using a Perkin-Elmer FTIR spectrometer (Perkin-Elmer, Woodbridge, ON) with a universal ATR sampling accessory. The scanning range was 4000-650 cm-1 with a resolution of 4 cm-1. Weight average molecular weights (Mw) and polydispersities (PDI) of the HPGs were determined by gel permeation chromatography (GPC) equipped with a DAWN-EOS multiangle laser light scattering (MALLS) detector (GPC-MALLS) and Optilab RI detector (Wyatt Technology Inc., Santa Barbara, CA). Aqueous 0.1 N sodium nitrate solution was used as the mobile phase at a flow rate of 0.8 mL/min. The details have been described in a previous report.11,28 The dn/dc values for various HPGs were determined to be 0.146, 0.165, and 0.138 for HPGC8/10-MePEG6.5, HPG-C8/10-MePEG6.5-COOH348, and HPG-C8/10-MePEG6.5-COOH113, respectively, in aqueous 0.1 N NaNO3 solutions and were used for the calculation of molecular weight of polymers. The data were processed using Astra software provided by Wyatt Technology Corp. Number average molecular weights of the polymers were calculated by dividing Mw by PDI. Potentiometric/pH titrations, to quantify the total concentration of HPGs surface-grafted with COOH, were performed on a T-50 M titrator (Mettler Toledo, Mississauga, ON). HPG-C8/10-MePEG6.5-COOH348 and HPG-C8/10-MePEG6.5-COOH113 samples were dissolved at 0.2 mg/mL in 10 mL of 10 mM NaOH. The pH of each solution was manually increased up to approximately 11 by the addition of 0.1 M NaOH. Samples were then titrated with 0.01 M HCl. Injections were set up in a dynamic range of 10-50 µL and a time interval of 30-60 s between injections was to ensure equilibration was established. Titrations were terminated once the pH reached 3.0. Titration end points were determined using the standard extrapolation/intersection method. The reported COOH titration values represent the mean of three measurements. The solubility characteristics of HPG polymers were assessed by dissolving known weights of the polymer in various aqueous buffers or distilled water. The samples were gently vortexed to speed dissolution. The absorbance of polymer solutions at 550 nm was regularly measured for signs of turbidity for several days to assess whether the polymer remained in solution. For some of the carboxylic acid-derivatized HPG polymers, the pH of the solution was adjusted to facilitate dissolution.

Biomacromolecules, Vol. 12, No. 1, 2011

147

Particle size and zeta-potential analysis were conducted using a Malvern NanoZS Particle Size analyzer (Malvern Instruments Ltd., Malvern, U.K.) using disposable sizing cuvettes. Polymer solutions at a concentration of 15 mg/mL were prepared in 1 mM NaCl at pH 6.0 and filtered with a 0.22 µm syringe filter (Pall Life Sciences, Ann Arbor, MI) prior to measurement. Cisplatin Binding to HPGs. The binding of cisplatin to carboxylate modified HPGs was assessed by preparing 10 mg/mL solutions of the polymers in 0.01 M NaOH. To these solutions, cisplatin was added so that the final concentration of drug ranged from 0.5 to 4 mg/mL. The pH of each solution was adjusted to 6.0 with small volumes of 5 M NaOH. The solutions were incubated overnight at 37 °C with shaking at 50 rpm. Solutions were transferred to Nanosep 3K Omega centrifugal filtration devices (Pall Life Sciences, Ann Arbor, MI) and centrifuged at 5000 rpm for 10 min. A small volume of the filtrate (10-40 µL) was diluted to 400 µL with 0.01 M NaOH, and the concentration of unbound cisplatin in the filtrate was assayed by a previously described o-phenylenediamine (OPDA) colorimetric assay.26 The concentration of cisplatin bound to the HPG was determined by subtracting the concentration of unbound cisplatin found in the filtrate from the initial concentration of drug added to the HPG. In Vitro Cisplatin Release. Cisplatin was bound to the carboxylate modified HPGs as described above with final polymer and cisplatin concentrations of 10 and 1 mg/mL, respectively. Into 7000 MWCO Slide-A-Lyzer mini dialysis units (Thermo Scientific, Rockford, IL), 20 µL of cisplatin bound polymer solution, or a 1 mg/mL solution of free cisplatin, were added and the samples were dialyzed at 37 °C with stirring against 4 L of 1 mM PBS adjusted to pHs of 4.5, 6.0, and 7.4 or synthetic urine at pH 7.0. At predetermined time points, three dialysis units were removed from the release media, and the entire contents were removed with three washings of the dialysis unit followed by dilution to 1 mL with fresh release media. The cisplatin concentration of contents of the dialysis units was determined by OPDA colorimetric assay. The cumulative percent drug released was calculated by subtracting the amount of drug remaining from the initial amount of drug in the dialysis bag at the beginning of the experiment. The data were expressed as cumulative percentage of drug released as a function of time. Cytotoxicity Evaluation. Cytotoxicity studies were performed using the MTS cell proliferation assay (Promega, Madison, WI). This assay does not measure immediate cytolytic effects of agents but measures the effect of the polymer on cellular proliferation over long time periods. In this study, KU-7-luc bladder cancer cells, kindly provided by Dr. M. Tachibana (Keio University, Tokyo, Japan), were chosen for cell cytotoxicity studies because this cell line has been previously used by our group to evaluate the effect of paclitaxel-loaded HPGs.10 The cells were plated at 5000 cells/well into 96-well plates in 180 µL of Dulbecco’s modified Eagle (DMEM) medium (Invitrogen Canada, Inc., Burlington, ON) supplemented with 10% fetal bovine serum (FBS) (Invitrogen Canada, Inc., Burlington, ON), 1% penicillin-streptomycin, and 1% L-glutamine and allowed to grow for 24 h at 37 °C in 5% CO2 to reach approximately 80% confluence for cytotoxicity assays. Cells were then incubated for 2 or 72 h with HPGs alone, ranging from 0.01-100 mg/mL, or free cisplatin or cisplatin-loaded HPGs with drug concentrations ranging from 0.01-100 µg/mL. After treatment, the cells were washed twice with Hank’s balanced salt solution (HBSS) and 180 µL of fresh culture media was added into each well and cells were allowed to grow for 72 h. Proliferation of these cells was measured using a CellTiter 96 aqueous non-radioactive cell proliferation assay (Promega, Madison, WI) as described previously.10 Briefly, 180 µL of a 10% v/v solution of 3-(4,5-dimethythiazol-2-yl)-5-(3-carboxylmethonyphenol)-2-(4-sulfophenyl)-2H-tetrazolium in HBSS was added to each well and the cells were incubated for 2 h. The absorbance was measured at 490 nm with a reference of 620 nm using a microplate reader.

148

Biomacromolecules, Vol. 12, No. 1, 2011

Ye et al.

Table 1. Properties of a Series of Surface-Modified C8/10 Alkyl Derivatized Hyperbranched Polyglycerols titration datab

molecular weightc

polymer compositiona

COOH (mol/mol HPG)

Mw (g/mol)

PDI (Mw/Mn)

particle sized (nm)

HPG-C8/10-OH HPG-C8/10-MePEG6.5 HPG-C8/10-COOH HPG-C8/10-MePEG6.5-COOH348 HPG-C8/10-MePEG6.5-COOH113

N/A N/A N/D 318 87

N/D 7.6 × 104 N/D 1.3 × 105 9.1 × 104

N/D 1.2 N/D 1.4 1.3

9.2 ( 3.5 8.7 ( 3.8 5.3 ( 1.8 5.9 ( 2.1 7.4 ( 3.0

a Nomenclature is designated as follows. HPG-C8/10-OH is the “base polymer” and all others were surface modified with MePEG and COOH, expressed as the theoretical number of moles of surface group added in the reaction per mole of HPG. b Moles of COOH groups per mole of HPG as determined by pH titration. c Weight average molecular weight and polydispersity index determined by GPC. Number average molecular weight is calculated by Mw/PDI. d Particle size (diameter), as determined by dynamic light scattering.

Results and Discussion Intravesical therapy by the instillation of anticancer drugs directly into the bladder is frequently used in the chemotherapeutic treatment of superficial bladder cancer. However, hydrophilic drugs such as doxorubicin, cisplatin and mitomycin C partition poorly into bladder tissues compromising the efficacy of these drugs via this route of administration.19,29 In addition, urine dilution of drugs within the bladder and voiding after 2 h further reduces the effectiveness of intravesical chemotherapy for any drug. In this work, we have prepared a series of HPGs functionalized with carboxylic acid groups that complex with cisplatin and act as a nanocarrier drug delivery system for this drug. We have previously shown that hydrophobically modified HPGs readily solubilize paclitaxel and this formulation effectively reduced the tumor growth in an orthotopic tumor model of superficial bladder cancer.10 This hydrophobically modified HPG was chosen for further modification for the binding of cisplatin to potentially develop a system that would allow for the codelivery of cisplatin and paclitaxel in future studies. Well-defined hydrophobically modified HPGs with narrow polydispersities may be synthesized from glycidol and comonomers via a controlled anionic ring-opening multibranching polymerization process using trimethylolpropane as an initiator and a slow monomer addition method. This synthetic procedure provides good control over the molecular weight of the resulting HPGs.1,30,31 In the preparation of all functionalized HPGs, target amounts of MePEG and COOH groups were added to reaction mixtures. The target amounts of MePEG and COOH of various functionalized HPGs are summarized in Table 1. The reaction yields for the high and low carboxylate functionalized HPGs were 84 and 74%, respectively. HPG polymers are described by the following nomenclature: HPG-C8/10-MePEGA-COOHB wherein HPG-C8/10 represents the alkyl substituted HPG, A is the target content of MePEG conjugated to the polymer, based on the stoichiometry of reagents (moles of MePEG/mol of TMP initiator), and B is the expected molar content of COOH per mole of HPG polymer, based on the calculated molecular weight of the polymer from GPC data. NMR Analysis. After purification, all of the HPGs were characterized by NMR analysis. From Figure 1 it can be seen that all of the peaks of functionalized HPG-C8/10-MePEG6.5COOH polymers were assigned to the structural components of the HPGs and were consistent with previous reports.2,4,26 While the 1H NMR and 13C NMR spectra were useful for confirming HPG structures, there were significant overlapping peaks in the 1 D NMR spectra (such as the CH3 group from TMP and -CH- and -CH2- from the HPG polymer) making it difficult to fully characterize these polymers (Figure 1). Therefore, we used HSQC, HMBC, and HSQC-TOCSY to

estimate the fractions of the substituents, C8/10 alkyl chain, MePEG, and COOH on HPGs using integrated peak volumes. Degree of Branching and Degree of Polymerization. Hyperbranched polymers are typically characterized by the degree of branching (DB) and degree of polymerization (DPn) using the following equations32

DB )

2D 2D + L13 + L14

where DB is the degree of branching, D, L13, and L14 represent the fractions of dendritic, linear 1-3, and linear 1-4 units, respectively. The structures of the dendritic and linear repeat units of glycidol that are present in the hyperbranched structure are summarized in Figure 2. Furthermore, the degree of polymerization (DPn) for these polymers is calculated as follows:1

DPn )

T + L13 + L14 + D · fc T-D

where D, L13, and L14 are defined as above, T represents the fraction of terminal units, and fc is the functionalization of the core molecule (which is 3 for TMP). D is given by the sum of primary and secondary units, Dp and Ds (see Figure 2), and L13, L14, and T are defined in an analogous manner. Unmodified hyperbranched polymers have been characterized using these equations before; see for example.33-35 We investigated if this routine characterization would be applicable to the HPG polymers we have synthesized, modified by the addition of MePEG, alkyl C8/10 groups and COOH groups. The alkyl and MePEG modifications are introduced via the same mechanism as the polymerization of glycidol, namely, anionic ring-opening polymerization of glycidol derivatives. Because the expected chemical shifts for modified structural units are virtually identical to those expected for unmodified structural units, an underestimation of T units and an overestimation of linear and dendritic units are expected. Modifications can be introduced at any hydroxyl group. For example, a MePEG unit can be added to the secondary hydroxyl group of a T unit. This newly formed unit is still a terminal unit, since it does not contribute to polymer branching, but in the NMR spectrum, its proton and carbon chemical shifts are indistinguishable from a L13 unit. In an analogous manner, T units can be misinterpreted not only as L13, but also as L14 or D units, whereas L13 and L14 units can be misinterpreted as D units. This leads to the expected underestimation of T and overestimation of linear and dendritic units.

Polyglycerols as Nanoparticulate Drug Carriers

Figure 1. (A) 1H NMR spectrum and (B)

Biomacromolecules, Vol. 12, No. 1, 2011

149

13

C NMR spectrum of HPG-C8/10-MePEG-COOH in methanol-d4.

When a combination of 2D HMBC and HSQC-TOCSY experiments is used, a number of peaks corresponding to primary and secondary L13, L14, T, and D units were assigned for an unmodified HPG polymer (which we synthesized as a reference material, containing no C8/10 alkyl component and no MePEG addition or carboxyl modification, data not shown), and peak volumes from a multiplicity-edited HSQC were used to calculate DB and DPn. The results obtained for our unmodified HPG were DB ) 0.51 and DPn ) 14.83, and the relative abundances for structural units are 39% for linear units, 20% for dendritic units, and 41% for terminal units. These values are in good agreement with literature values.1,32 When comparing the HSQC spectrum of HPG modified with C8/10 alkyl chains (HPG-C8/10-OH), to the HSQC spectrum of unmodified HPG, two new peaks are visible in the spectral region of the polymer core of the former (Figure 3). One peak was assigned to the R-methylene group of the aliphatic chain, whereas the second peak could not be assigned unambiguously.

Based on chemical shifts, this peak may correspond to a T unit with one alkyl chain attached to the secondary hydroxyl group; however, this speculation could not be confirmed. A similar situation was observed for HPG-MePEG6.5. The peak from the R-methylene group of the MePEG could be assigned, but the additional, unknown peak could not be assigned unambiguously. Similar to HPG-C8/10-OH, the chemical shifts of the new peak are similar to an L14-like unit. In summary, DB and DPn could not be calculated from NMR data due to lack of unambiguous signal assignment. Currently, it is, to our knowledge, not possible to distinguish between true and pseudo linear and dendritic units in a way that would be suitable for routine determination of DB and DPn. Considering the complexity of the NMR spectra of HPG-type polymers, this is not unexpected. While the degree of branching and the degree of polymerization could not be obtained, NMR data allows for a straightforward characterization of free hydroxyl groups through observation of linear or terminal units and confirmation

150

Biomacromolecules, Vol. 12, No. 1, 2011

Ye et al.

Figure 2. Structural units in HPG polymers. Each dendritic, terminal, and linear unit exists as primary and secondary unit. For unmodified polymers, R ) HPG; for modified polymers as described in the present work, R ) HPG, C8/10, MePEG, or COOH. The numbering scheme is indicated for the Dp unit.

that all expected branching patterns and modifications (alkyl, MePEG, and carboxyl) are present. Figure 3 illustrates the various assigned peaks in the NMR spectra, showing the presence of the expected branching pattern, and of MePEG, alkyl chains, and COOH groups. Mole Fractions of COOH. For all HPG polymers modified with COOH, the mole fractions of COOH were estimated from HSQC NMR spectra. By this method, the number of COOH in the HPG polymer is not an absolute number, because it is expressed as relative to the TMP methyl groups present in the sample. We assume that each HPG molecule contains only one TMP; however, we have not independently quantified the amount of TMP per mole of HPG in the various batches of polymer. Therefore, these numbers serve as a qualitative indicator of how many hydroxyl groups were capped with COOH. Furthermore, because the HPG-C8/10-MePEG6.5-COOH polymers were both synthesized from the same batch of HPGC8/10-MePEG6.5, the TMP content is expected to be identical and the NMR spectra can be compared to determine the relative amount of COOH in the two HPG-C8/10-MePEG6.5-COOH polymers. The molar ratios indicate that there is a 2.8-fold higher COOH content in HPG-C8/10-MePEG6.5-COOH348 compared with the HPG-C8/10-MePEG6.5-COOH113. This is in good agreement with the 3.1-fold ratio of target COOH content in the two polymers. For HPG-C8/10-COOH and the high-carboxyl density HPGC8/10-MePEG6.5-COOH348 polymers, no peaks corresponding to linear or terminal groups were observed, indicating that no hydroxyl groups are present in this polymer (see Figure 4 for a representative NMR spectrum). For the lower density COOH polymer, HPG-C8/10-MePEG6.5-COOH113, peaks of linear and terminal groups were observed in addition to the new peaks,

Figure 3. Representative multiplicity-edited HSQC spectra of (A) HPG-C8/10-OH, (B) HPG-C8/10-COOH (high COOH), and (C) HPGC8/10-MePEG6.5. Positive contour levels corresponding to CH and CH3 groups are shown in black, while negative contour levels corresponding to CH2 groups are shown in red. Representative assignments are indicated in the spectra.

indicating only a partial saturation of hydroxyl groups with carboxylic acids (data not shown). FT-IR. FT-IR spectra of HPG-C8/10-MePEG6.5, HPG-C8/10MePEG6.5-COOH348 and HPG-C8/10-MePEG6.5-COOH113 are shown in Figure 5. The peak at 2800-3000 cm-1 is consistent with C-H vibrations and occurs in all HPGs. The peaks at 1680-1780 cm-1 arose from CdO bands, indicating the presence of COOH groups in the HPG-C8/10-MePEG-COOH polymers. The peaks at 1200-1400 and 1000-1180 cm-1 arise from a C-H bend and C-O vibration, respectively, and therefore, they can be found in all of these polymers. By comparing the FT-IR spectra of the HPG-C8/10-MePEG6.5, the HPG-C8/10-MePEG6.5-COOH113, and the HPG-C8/10-MePEG6.5-COOH348, it can be seen that the OH peak (3300 - 3500 cm-1) decreases and the CdO peak (1680-1780 cm-1) in the

Polyglycerols as Nanoparticulate Drug Carriers

Biomacromolecules, Vol. 12, No. 1, 2011

151

Figure 4. Expansions of regions of the HSQC spectra of (A) HPG-C8/10-OH and (B) HPG-C8/10-COOH. Representative assignments are given. New peaks resulting from the addition of COOH are indicated, but no specific assignment was obtained. Note that all peaks corresponding to linear or terminal units (indicated by gray rectangles) are absent in the HPG-C8/10-COOH spectrum (panel B), indicating that no free hydroxyl groups are present in this polymer. This observation is in accordance with the FT-IR data obtained in this work.

Figure 5. FT-IR spectra of HPG-C8/10-MePEG6.5 (top), HPG-C8/10-MePEG6.5-COOH113 (center), and HPG-C8/10-MePEG6.5-COOH348 (bottom).

HPG-C8/10-MePEG-COOH appears, indicating the OH groups have been consumed and converted to COOH. The spectrum of the HPG-C8/10-MePEG6.5-COOH348 showed the near elimination of the OH peak, indicating that the OH groups were largely consumed and converted to COOH, whereas the OH peak for HPG-C8/10-MePEG6.5-COOH113 was decreased but still evident, in agreement with the NMR results. Furthermore, the latter HPG also shows a smaller CdO peak, indicating a lower mole ratio of COOH compared to HPG-C8/10-MePEG6.5-COOH348. The FT-IR data showed good evidence to support the changes in the functionalization of the HPGs and also confirmed that unreacted reagents were removed by the purification procedures. Molecular Weight. The molecular weights and polydispersities of these HPGs are shown in Table 1. The functionalized HPGs (HPG-C8/10-MePEG6.5-COOH) showed increases in molecular weight compared to HPG-C8/10-MePEG6.5. Furthermore, it was found that after the surface functionalization, the polydispersities of the polymers were not altered greatly, indicating a relatively uniform surface modification. Molecular weight values were similar to those of previously reported HPG-C8/10-MePEG.2,4,10,11

Titration of COOH Groups. The mole ratios of COOH groups conjugated to the HPGs were measured by potentiometric/pH titration (Table 1) and showed good agreement with target mole ratios and the measured molecular weights. For instance, the molecular weight of HPG-C8/10-MePEG6.5COOH113 can also be calculated by the addition of the number average molecular weight of the HPG-C8/10-MePEG6.5 (6.3 × 104) with the molecular weight ascribed to COOH groups, which equals the number of carboxylate per HPG molecule (87 from titration data) multiplied by the carboxylate molecular weight (101 g/mol). Based on this calculation, the number average molecular weight of HPG-C8/10-MePEG6.5-COOH113 is 7.2 × 104 g/mol, in good agreement with the measured value (7.0 × 104 g/mol). Solubility. As potential drug nanocarriers, the solubility characteristics in aqueous media of these polymers are critical. It was found that the HPG-C8/10-MePEG polymer had good water solubility (greater than 100 mg/mL) in distilled water, PBS buffer (pH of 7.4), and synthetic urine. HPG-C8/10-COOH was found to be practically insoluble in aqueous media or PBS

152

Biomacromolecules, Vol. 12, No. 1, 2011

Ye et al.

Figure 7. Binding of cisplatin to (4) HPG-C8/10-MePEG6.5-COOH113 or (9) HPG-C8/10-MePEG6.5-COOH348 in distilled water adjusted to pH 6.0. Each point is the average of three samples ( SD.

Figure 6. Representative structure of HPG-C8/10-MePEG-COOH bound to cisplatin.

(pH 7.4) and only soluble in alkaline solutions such as 0.1 M NaOH, due to decreased ionization at neutral pH. The hydrophobic (alkyl chains) components of the HPG core likely dominated the solubility characteristics. As expected, carboxylate-derivatized HPGs also conjugated with MePEG groups showed increased water solubility. Accordingly, it was found that HPG-C8/10-MePEG6.5-COOH113 could be completely dissolved in 10 mM PBS at a concentration of 100 mg/mL without heating, although the pH of the solution dropped from 7.4 to 4.5. HPG with a higher amount of carboxylate (HPG-C8/10-MePEG6.5-COOH348) was poorly soluble in water or PBS buffer and exceeded the buffering capacity, resulting in acidification of PBS buffer and dropping the pH from 7.4 to approximately 3.8. The solution exhibited significant turbidity as measured by absorbance at 550 nm, demonstrating an insoluble residual fraction of polymer (data not shown). The addition of sodium hydroxide was required to achieve a concentration of 100 mg/mL and a clear solution at pH 4.25. This type of pH-dependent solubility is common for polymers possessing ionizable groups36 and has been previously described for chitosan with positively charged amine groups in acidic conditions37 and for polythiol functionalized HPGs.38 Particle Size and Zeta Potential. Carboxyl-terminated HPG polymers had particle sizes in the 5-10 nm range (Table 1). These values are in agreement with previous studies showing that the HPG polymers containing C10 or C18 alkyl chains and MePEG chains have a hydrodynamic radius of approximately 10 nm.4,10 Zeta potentials of the nanoparticles were strongly negative at -41.2 ( 3.2 and -60.3 ( 2.1 mV for HPG-C8/10MePEG6.5-COOH113 and HPG-C8/10-MePEG6.5-COOH348, respectively. The decrease in zeta potential is attributed to the number of carboxyl groups conjugated to the surface of the HPGs. Cisplatin Binding. Binding of cisplatin to the HPGs was achieved through coordination of the drug to terminal carboxylate groups on the polymer (Figure 6). For HPG-C8/10-MePEG6.5-

Figure 8. In vitro release of free cisplatin (]) or cisplatin bound to (A) HPG-C8/10-MePEG6.5-COOH113 or (B) HPG-C8/10-MePEG6.5COOH348 at a drug concentration of 1 mg/mL and polymer concentration of 10 mg/mL. Release media were 1 mM PBS at pHs of 4.5 (9), 6.0 (4), 7.4 (3), or artificial urine (0) at 37 °C. Each point is the average of three samples ( SD.

COOH113 cisplatin bound to the polymer with nearly 100% efficiency up to a maximum of 1 mg/mL (10% w/w; Figure 7). Above this concentration, free drug was detected in the filtrate, indicating saturation of the carboxylate binding sites and the presence of unbound drug in the media. HPG-C8/10-MePEG6.5COOH348 bound up to 2 mg/mL with 100% efficiency before free drug was detected in the filtrate. This increase in bound drug is attributed to the increased number of carboxylate groups and, thus, number of cisplatin binding sites on HPG-C8/10MePEG6.5-COOH348 as compared to HPG-C8/10-MePEG6.5COOH113. In Vitro Drug Release Studies. Release of free cisplatin in PBS was rapid and 100% complete within 7 h, demonstrating that the membrane did not impede the release of free drug to any great extent (Figure 8). For all cisplatin-bound HPG samples, the drug was found to release in a controlled fashion, considerably slower than the free drug. In PBS, regardless of the pH, cisplatin bound to HPG-C8/10-MePEG6.5-COOH113 was released at nearly the same rate in PBS with approximately 5% released in the first 2 h, 40% release after 1 day, and up to

Polyglycerols as Nanoparticulate Drug Carriers

90% released after 7 days. Cisplatin bound to HPG-C8/10MePEG6.5-COOH348 at pHs of 6.0 and 7.4 released the drug in PBS at similar rates, in a nearly linear manner, with approximately 3% of bound cisplatin released in 2 h, 20% released in 1 day, and up to 70% over 7 days. This decreased drug release rate is likely due to the higher number of carboxylate groups, thus, providing more opportunities for stronger bidentate binding of cisplatin or providing free carboxylate sites for rebinding of released drug. Interestingly, the release rate for cisplatin bound to HPG-C8/10-MePEG6.5-COOH348 at pH 4.5 was faster than its higher pH counterparts, with a release profile similar to those of HPGC8/10-MePEG6.5-COOH113. The release of cisplatin from carboxylate ligands is largely mediated by the chloride concentration in the release media;27 however, in these experiments, regardless of the pH of the media, the chloride concentration in the PBS was kept constant. The faster release rate may have arisen from the protonation of free carboxyl groups at the low pH, leading to inhibition of the rebinding of released cisplatin. In an in vivo situation during intravesical administration, the production of urine will reduce the concentration of drug present in the bladder but may also influence the release of cisplatin from the HPG; therefore, the release rate of cisplatin was investigated using artificial urine as the release media. Interestingly, the release rate of cisplatin was considerably faster in the presence of urine, with just over 10% of the dose released in 2 h and complete drug release by 2 days for HPG-C8/10MePEG6.5-COOH113 and 3 days for HPG-C8/10-MePEG6.5COOH348. Similar to release in PBS, the difference in cisplatin release between the two HPGs may be attributed to the increased number of carboxylate groups present on HPG-C8/10-MePEG6.5COOH348. As urine is a complex mixture made up of several components, it is uncertain which compounds are responsible for the increased release of the cisplatin from the HPGs; however, this increased release rate may be advantageous, providing a mechanism by which the drug release increases upon dilution with urine. Upon displacement of cisplatin from the HPG it is possible that nitrogen containing compounds in urine, such as urea, uric acid and creatinine, may bind and inactivate cisplatin. Although cisplatin has been shown to complex with these compounds to some degree, it has been determined that the majority of cisplatin present in urine after IV administration is in the originally administered form and the highly active monoaqua hydrolysis product.39 In light of this finding, it is likely that the majority of the cisplatin released in from the HPGs in urine is in a pharmacologically active form. Cell Proliferation Studies. The inhibition effects of nondrug-loaded HPGs and cisplatin-loaded HPGs on KU-7-luc bladder cancer cells were investigated for incubation times of 2 and 72 h (Figure 9). These incubation times were chosen to allow us to imitate the typical intravesical instillation period as well as to compare against previously determined inhibitory concentrations for cisplatin. Inhibitory concentrations at 50% (IC50) for 2 h incubations were determined to be 1.3, 45.7, 47.0, and 63.0 mg/mL for HPG-C8/10-OH, HPG-C8/10-MePEG6.5, HPGC8/10-MePEG6.5-COOH348, and HPG-C8/10-MePEG6.5-COOH113, respectively. When a 72 h incubation was used, the polymers inhibited cell proliferation to a greater degree than those found with a 2 h incubation. The HPG-C8/10-OH and HPG-C8/10MePEG6.5 had IC50s of 0.1 and 0.2 mg/mL, respectively. The IC50 for the carboxylate-modified HPGs decreased approximately 10-fold; however, these polymers still exhibited a high degree of cellular compatibility with IC50 values of approximately 5 mg/mL. The overall excellent biocompatibility of the HPG-C8/10-MePEG and HPG-C8/10-MePEG-COOH prob-

Biomacromolecules, Vol. 12, No. 1, 2011

153

Figure 9. Cell viability of KU-7-luc cells after (A) 2 and (B) 72 h of incubation with HPG-C8/10-OH (9), HPG-C8/10-MePEG6.5 (4), HPGC8/10-MePEG6.5-COOH113 (b), and HPG-C8/10-MePEG6.5-COOH348 (]). Each point is the average of six samples ( SD.

ably arises from the known cellular compatibility of MePEG surfaces ensuring little interaction with the plasma membrane of the cells. The added benefit of carboxylation may arise from the net negative charge of this moiety at a pH of 7.4, establishing a slight repulsive force with the negatively charged cell surface. Following a 72 h incubation, free cisplatin inhibited KU7-luc cell proliferation with an IC50 of 1 µg/mL (Figure 10A), consistent with previous reports for this drug and cell combination.19 With a 2 h incubation, this IC50 value increased to approximately 10 µg/mL (Figure 10B). When bound to HPG-C8/10-MePEG6.5-COOH polymers, the complexed form of cisplatin also inhibited KU-7-luc proliferation with higher IC50 values observed for the 2 h incubation (approximately 50 µg/mL) as compared to the 72 incubation values (approximately 5 µg/mL). Clearly, for both 2 and 72 h incubations, the complexed form of cisplatin inhibited cell proliferation less than the free drug by a factor of almost 5. This increase in the IC50 for the drug complexed to the HPGs is likely due to the slow release rate of the drug from the polymer. This increase in IC50 has been reported by other researchers developing conjugated cisplatin systems and has been found to be upward of a 100-fold decrease in potency.40-42 The efficacy of intravesically administered cisplatin is limited by the short residence time of the drug due to voiding of the bladder in addition to toxicities such as cystitis from the instillation of high doses of the drug. A drug delivery system that could promote uptake into the urothelium and, thus, form a controlled release depot would be of great therapeutic benefit. Often, intravesical administration of chemotherapeutics occurs immediately after transure-

154

Biomacromolecules, Vol. 12, No. 1, 2011

Ye et al.

Acknowledgment. This work was supported by an operating grant from the Canadian Institutes of Health Research (CIHR; H.M.B.), the Natural Sciences and Engineering Research Council of Canada (H.M.B and D.E.B), Michael Smith Foundation for Health Research (D.E.B), and Canadian Blood Services (D.E.B.). Research performed with the assistance of the LMB Macromolecular Hub at the Centre for Blood Research was supported in part by grants from the Canada Foundation for Innovation and the Michael Smith Foundation for Health Research. J.N.K. is a recipient of CIHR/CBS new investigator award in Transfusion Science. This work was also supported with support from the Centre for Drug Research and Development and a grant from the Canada Foundation for Innovation (CFI).

References and Notes

Figure 10. Viability of KU-7-luc cells after (A) 2 and (B) 72 h incubation with free cisplatin (b), cisplatin-loaded HPG-C8/10-MePEG6.5-COOH113 (4), and HPG-C8/10-MePEG6.5-COOH348 (9). Each point is the average of six samples ( SD.

thral resection of the tumor for cases involving low grade noninvasive bladder cancer; however, tumor reimplatation and subsequent tumor regrowth typically occurs at these traumatized sites.43,44 Recent and ongoing studies in our laboratory have demonstrated that paclitaxel-loaded HPGs effectively increased the drug partitioning into bladder tissue in an ex vivo model. Furthermore, it has been shown that anionic, carboxylate bearing polyamidoamine dendrimers are readily taken up by endocytosis.45 Therefore, in future studies we intend to investigate the partitioning of HPG-bound cisplatin into bladder tissue using our ex vivo bladder tissue model as well as explore the ability of this system to undergo endocytosis.

Conclusions We have demonstrated the synthesis of several well-defined carboxylic acid, MePEG and C8/10 alkyl core modified HPGs, with structures that could be confirmed using a combination of one- and two-dimensional NMR, FTIR, and titration studies. The nanoparticles all possess average diameters less than 10 nm in water. The carboxylate functionalized HPGs effectively bound cisplatin, released the drug in a controlled manner and effectively inhibited proliferation of KU-7-luc bladder cancer cells. These HPGs may be promising nanocarriers for the delivery of cisplatin to tumors.

(1) Sunder, A.; Hanselmann, R.; Frey, H.; Mu¨lhaupt, R. Macromolecules 1999, 32, 4240–4246. (2) Kainthan, R. K.; Mugabe, C.; Burt, H. M.; Brooks, D. E. Biomacromolecules 2008, 9, 886–895. (3) Kainthan, R. K.; Gnanamani, M.; Ganguli, M.; Ghosh, T.; Brooks, D. E.; Maiti, S.; Kizhakkedathu, J. N. Biomaterials 2006, 27, 5377– 5390. (4) Kainthan, R. K.; Janzen, J.; Kizhakkedathu, J. N.; Devine, D. V.; Brooks, D. E. Biomaterials 2008, 29, 1693–1704. (5) Jaszberenyi, Z.; Moriggi, L.; Schmidt, P.; Weidensteiner, C.; Kneuer, R.; Merbach, A. E.; Helm, L.; Toth, E. J. Biol. Inorg. Chem. 2007, 12, 406–420. (6) Kolhe, P.; Khandare, J.; Pillai, O.; Kannan, S.; Lieh-Lai, M.; Kannan, R. Pharm. Res. 2004, 21, 2185–2195. (7) Calderon, M.; Graeser, R.; Kratz, F.; Haag, R. Bioorg. Med. Chem. Lett. 2009, 19, 3725–3728. (8) Xu, S.; Luo, Y.; Graeser, R.; Warnecke, A.; Kratz, F.; Hauff, P.; Licha, K.; Haag, R. Bioorg. Med. Chem. Lett. 2009, 19, 1030–1034. (9) Tziveleka, L.-A.; Kontoyianni, C.; Sideratou, Z.; Tsiourvas, D.; Paleos, C. M. Macromol. Biosci. 2006, 6, 161–169. (10) Mugabe, C.; Hadaschik, B. A.; Kainthan, R. K.; Brooks, D. E.; So, A. I.; Gleave, M. E.; Burt, H. M. BJU Int. 2009, 103, 978–986. (11) Kainthan, R. K.; Brooks, D. E. Bioconjugate Chem. 2008, 19, 2231– 2238. (12) Sonpavde, G.; Sternberg, C. N. BJU Int. 2010, 106, 6–22. (13) Gallagher, D. J.; Milowsky, M. I. Curr. Treat. Options Oncol. 2009, 10, 205–215. (14) Font, A.; Taron, M.; Gago, J. L.; Costa, C.; Sanchez, J. J.; Carrato, C.; Mora, M.; Celiz, P.; Perez, L.; Rodriguez, D.; Gimenez-Capitan, A.; Quiroga, V.; Benlloch, S.; Ibarz, L.; Rosell, R. Ann. Oncol. 2010, DOI: 10.1093/annonc/mdq333. (15) Takata, R.; Katagiri, T.; Kanehira, M.; Tsunoda, T.; Shuin, T.; Miki, T.; Namiki, M.; Kohri, K.; Matsushita, Y.; Fujioka, T.; Nakamura, Y. Clin. Cancer Res. 2005, 11, 2625–2636. (16) Blumenreich, M. S.; Needles, B.; Yagoda, A.; Sogani, P.; Grabstald, H.; Whitmore, W. F., Jr. Cancer 1982, 50, 863–865. (17) Mobley, W. C.; Loening, S. A.; Narayana, A. S.; Culp, D. A. Urology 1986, 27, 335–339. (18) Saxena, S.; Agrawal, U.; Agarwal, A.; Murthy, N. S.; Mohanty, N. K. BJU Int. 2006, 98, 1012–1017. (19) Hadaschik, B. A.; ter Borg, M. G.; Jackson, J.; Sowery, R. D.; So, A. I.; Burt, H. M.; Gleave, M. E. BJU Int. 2008, 101, 1347–1355. (20) Hwang, T. L.; Fang, C. L.; Chen, C. H.; Fang, J. Y. Pharm. Res. 2009, 26, 2314–2323. (21) Aryal, S.; Hu, C. M.; Zhang, L. ACS Nano 2010, 4, 251–258. (22) Verma, A. K.; Sachin, K. Curr. Drug DeliVery 2008, 5, 120–126. (23) Ye, H.; Jin, L.; Hu, R.; Yi, Z.; Li, J.; Wu, Y.; Xi, X.; Wu, Z. Biomaterials 2006, 27, 5958–5965. (24) Nishiyama, N.; Okazaki, S.; Cabral, H.; Miyamoto, M.; Kato, Y.; Sugiyama, Y.; Nishio, K.; Matsumura, Y.; Kataoka, K. Cancer Res. 2003, 63, 8977–8983. (25) Matsumura, Y. AdV. Drug DeliVery ReV. 2008, 60, 899–914. (26) Haxton, K. J.; Burt, H. M. Dalton Trans. 2008, 5872–5875. (27) Haxton, K. J.; Burt, H. M. J. Pharm. Sci. 2009, 98, 2299–2316. (28) Kumar, K. R.; Kizhakkedathu, J. N.; Brooks, D. E. Macromol. Chem. Phys. 2004, 205, 567–573. (29) Wientjes, M. G.; Badalament, R. A.; Au, J. L. Cancer Chemother. Pharmacol. 1996, 37, 539–546.

Polyglycerols as Nanoparticulate Drug Carriers (30) Hanselmann, R.; Holter, D.; Frey, H. Macromolecules 1998, 31, 3790– 3801. (31) Kainthan, R. K.; Muliawan, E. B.; Hatzikiriakos, S. G.; Brooks, D. E. Macromolecules 2006, 39, 7708–7717. (32) Ho¨lter, D.; Burgath, A.; Frey, H. Acta Polym. 1997, 48, 30–35. (33) Cheng, H.; Wang, S.; Yang, J.; Zhou, Y. J. Colloid Interface Sci. 2009, 337, 278–284. (34) Li, M.; Yang, X.; Liu, Y.; Wang, X. J. Appl. Polym. Sci. 2006, 101, 317–322. (35) Sun, X.; Yang, X.; Liu, Y.; Wang, X. J. Polym. Sci., Part A 2004, 42, 2356–2364. (36) Brittain, H. G. Biotech. Pharm. Aspect. 2007, 6, 29–51. (37) Kumar, M. N.; Muzzarelli, R. A.; Muzzarelli, C.; Sashiwa, H.; Domb, A. J. Chem. ReV. 2004, 104, 6017–6084. (38) Wan, D.; Li, Z.; Huang, J. J. Polym. Sci., Part A 2005, 43, 5458– 5464.

Biomacromolecules, Vol. 12, No. 1, 2011

155

(39) Tang, X.; Hayes Ii, J. W.; Schroder, L.; Cacini, W.; Dorsey, J.; Elder, R. C.; Tepperman, K. Met. Based Drugs 1997, 4, 97–109. (40) Malik, N.; Evagorou, E. G.; Duncan, R. Anti-Cancer Drugs 1999, 10, 767–776. (41) Uchino, H.; Matsumura, Y.; Negishi, T.; Koizumi, F.; Hayashi, T.; Honda, T.; Nishiyama, N.; Kataoka, K.; Naito, S.; Kakizoe, T. Br. J. Cancer 2005, 93, 678–687. (42) Zhou, P.; Li, Z.; Chau, Y. Eur. J. Pharm. Sci. 2010, 41, 464–472. (43) Sexton, W. J.; Wiegand, L. R.; Correa, J. J.; Politis, C.; Dickinson, S. I.; Kang, L. C. Cancer Control 2010, 17, 256–268. (44) Gunther, J. H.; Jurczok, A.; Wulf, T.; Brandau, S.; Deinert, I.; Jocham, D.; Bohle, A. Cancer Res. 1999, 59, 2834–2837. (45) Perumal, O. P.; Inapagolla, R.; Kannan, S.; Kannan, R. M. Biomaterials 2008, 29, 3469–3476.

BM101080P