ARTICLE pubs.acs.org/Biomac
Folic Acid Conjugated Amino Acid-Based Star Polymers for Active Targeting of Cancer Cells Adrian Sulistio,† Justin Lowenthal,‡ Anton Blencowe,† Marie N. Bongiovanni,†,§ Lydia Ong,†,§ Sally L. Gras,†,§ Xiaoqing Zhang,^ and Greg G. Qiao*,† †
Department of Chemical and Biomolecular Engineering, University of Melbourne, Parkville, Melbourne, VIC 3010, Australia Department of Biomedical Engineering, Yale University, New Haven, Connecticut 06511, United States § The Bio21 Molecular Science and Biotechnology Institute, The University of Melbourne, Parkville, Melbourne, VIC 3010, Australia ^ CSIRO Material Science and Engineering, Private Bag 33, Clayton South MDC 3169, VIC, Australia ‡
bS Supporting Information ABSTRACT: Amino acid-based core cross-linked star (CCS) polymers (poly(L-lysine)armpoly(L-cystine)core) with peripheral allyl functionalities were synthesized by sequential ring-opening polymerization (ROP) of amino acid N-carboxyanhydrides (NCAs) via the arm-first approach, using N-(trimethylsilyl) allylamine as the initiator. Subsequent functionalization with a poly(ethylene glycol) (PEG)folic acid conjugate via thiol ene click chemistry afforded poly(PEG-b-L-lysine)armpoly(Lcystine)core stars with outer PEG coronas decorated with folic acid targeting moieties. Similarly, a control was prepared without folic acid, using just PEG. A fluorophore was used to track both star polymers incubated with breast cancer cells (MDA-MB-231) in vitro. Confocal microscopy and flow cytometry revealed that the stars could be internalized into the cells, and higher cell internalization was observed when folic acid moieties were present. Cytotoxicity studies indicate that both stars are nontoxic to MDA-MB-231 cells at concentrations of up to 50 μg/mL. These results make this amino acid-based star polymer an attractive candidate in targeted drug delivery applications including chemotherapy.
’ INTRODUCTION The development of polymeric nanocarriers as targeted drug delivery vehicles for the treatment of cancer offers significant therapeutic benefits over other strategies, such as conventional chemotherapy or radiation theraphy.1 For example, polymeric vesicles can improve drug solubility in an aqueous medium, provide increased payload capacity, prolong the circulation time of the drug in the bloodstream, reduce drug toxicity to healthy cells,2 and passively target tumor tissue due to enhanced permeability and retention (EPR) effects as well as actively targeting cancer cells.1,3,4 Such targeted approaches allow for reduced drug dosage while improving the efficacy of the treatment and minimizing the side effects to the patient. Various nanocarriers have been developed, including polymeric micelles (self-assembled from amphiphilic polymers),510 layer-by-layer (LbL) capsules,11 dendrimers,12 and star polymers.13,14 It has been demonstrated that star polymers possessing hydrophobic cores and radiating hydrophilic arms can combine the advantages of polymeric micelles, dendrimers, and LbL capsules by providing enhanced encapsulation capabilities for hydrophobic drugs in the core,15 while avoiding issues associated with the sudden dissociation (under shear or dilution below the critical micelle concentration (cmc)) of polymeric micelles and the time-consuming multistep synthesis of LbL capsules and r 2011 American Chemical Society
dendrimers.1416 Moreover, star polymers are known to possess unique properties such as low solution viscosities17,18 and enhanced functionalities,1923 which makes this class of macromolecules an attractive candidate for drug delivery applications. However, only a limited number of these polymeric nanocarriers are reported to be biocompatible and biodegradable.24 Recently, we reported the successful synthesis of core crosslinked star (CCS) polymers composed exclusively of amino acid building blocks.25 These star polymer were composed of poly(Llysine) arms radiating from a poly(L-cystine) core and could be site specifically functionalized to possess hierarchical functionalities spanning from the core, along the arms, to the periphery. Furthermore, the core building block, L-cystine, has a centrally located disulfide bond that can be cleaved by reducing agents such as glutathione; hence, the resulting star polymers possess bioreducable cores. Subsequently, it was demonstrated that these star polymers are also capable of sequestering hydrophobic drugs, such as pirarubicin, within their interior. In this study, the viability of this CCS polymer as an active targeting drug delivery vehicle for the treatment of cancer was Received: May 3, 2011 Revised: August 18, 2011 Published: August 22, 2011 3469
dx.doi.org/10.1021/bm200604h | Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules explored by functionalizing the star with a targeting moiety and investigating its in vitro cell distribution with a breast cancer cell line (MDA-MB-231). The star polymer was synthesized via a one-pot, arm-first approach using sequential ring-opening polymerization (ROP) of amino acid N-carboxyanhydride (NCA) derivatives initiated with N-(trimethylsilyl)allylamine.26 Z-L-Lysine was chosen as the building block for the arms since the pendent amine functionalities available after deprotection could potentially be used to conjugate chemotherapeutic drugs via acid-labile linkers, such as aryl imines,27 or tracing agents for imaging (vide infra). The incorporation of allyl functionalities at the star periphery allowed for postpolymerization functionalization with thiolated PEG derivatives via “thiolene” click chemistry. Thus, the use of heterodifunctional α-thiol, ω-folic acid PEG provided a facile route to PEGylation and simultaneously introduce targeting moieties. This functionalization served two purposes: (i) PEGylation of the star would shield the peptide components from rapid enzymatic degradation and suppress their interaction with plasma protein and cells, hence reducing their uptake by the liver and increasing their circulation time in the bloodstream;12,16 and (ii) the folic acid (FA) moieties impart active targeting capabilities due to their high binding affinity for folate receptors (FRs) (Kd ∼ 1010 mol L1),2830 which are overexpressed on the surfaces of endothelial cancer cells.31 Targeting these receptors can facilitate cell internalization, presumably via receptor-mediated endocytosis.32 Gel permeation chromatography (GPC), dynamic light scattering (DLS), and 1H NMR spectroscopic analysis were used to characterize the PEGylated CCS polymers and their precursors. The cellular uptake of the star polymers in MDA-MB-231 cells was assessed via confocal microscopy and flow cytometry, and their cytotoxicity at various concentrations was determined using a cell viability assay.
’ EXPERIMENTAL SECTION Materials. Z-L-Lys(Z)-OH (Bachem), L-cystine (Aldrich), N-(trimethylsilyl)allylamine (96%, Aldrich), benzyl chloroformate (95%, Aldrich), sodium bicarbonate (NaHCO3) (99%, Ajax Fine Chemicals), magnesium sulfate (MgSO4) (>98%, anhydrous, Merck), phosphorus pentachloride (PCl5) (>99%, Merck), HO-PEG-NHCO-C2H4-S-Trt (Mw = 3 kDa) (Rapp Polymere), MeO-PEG-OH (Mw = 2 kDa) (Aldrich), folic acid (FA) (g97%, Sigma), N-(3-(dimethylamino)propyl)-N0 -ethylcarbodiimide hydrochloride (EDCI) (>98%, Fluka), 4-(dimethylamino)pyridine (DMAP) (99%, Aldrich), 2,2-dimethoxy-2-phenylacetophenone (DMPA) (99%, Aldrich), aminomethylpyrene hydrochloride (95%, Aldrich), hydrobromic acid (33% in acetic acid, Aldrich), trifluoroacetic acid (TFA) (98%, Aldrich), hydrochloric acid (37%, Scharlau), lithium bromide (99.9%, Aldrich), tetrahydrofuran (THF) (99.9%, HPLC grade, RCI Labscan), 3-mercaptopropanoic acid (g99%, Aldrich), sulfuric acid (H2SO4) (99.8%, BDH), sodium chloride (NaCl) (AR grade, Chem-Supply) and N,Ndimethylformamide (DMF) (99.8%, anhydrous, Aldrich) were used as received. n-Hexane (99.95%, Merck), ethyl acetate (99.5%, Chem-Supply), dimethyl sulfoxide (DMSO) (99.9%, Chem-Supply), chloroform (99.8%, Chem-Supply), and methanol (99.7%, Chem Supply) were used as received. 1,4-Dioxane (>98%, Fluka) and diethyl ether (99%, Chem-Supply) were used to make a binary solvent (2:3 v/v) and dried for 48 h prior to use over 3 Å molecular sieves (Aldrich). Acetonitrile (99.99%, HPLC grade, B&J) was stored over 3 Å molecular sieves. MALDI ToF MS matrices (dithranol (98.5%)) and cationization agents (NaTFA (98%), KI (99%)) were purchased from Fluka and Aldrich, respectively, and were used as received. Dichloromethane (99.5%, Chem-Supply) was distilled from CaH2 under argon. Milli-Q water (18.2 MΩ 3 cm) was obtained from a Millipore Synergy
ARTICLE
Water System. Acetone-d6 (99.9%), DMSO-d6 (99.9%), DMF-d7 (99.5%), and D2O (99.9%) were purchased from Cambridge Isotope Laboratories and used as received. Breast cancer cells (MDA-MB-231, American Type Culture Collection (ATCC)), Dulbecco’s Modified Eagle’s Medium with L-glutamine and without folic acid (DMEM) (Sigma), Dulbecco’s phosphate buffered saline (PBS) (Sigma), penicillinstreptomycin solution (Sigma), fetal bovine serum (Invitrogen), trypsin (Sigma), 6-well and 96well cell culture plates, and T25 cell culture flasks (Corning) were used for cell culture. Alexa Fluor 647 carboxylic acid succinimidyl ester (Invitrogen) was used to stain the polymer. Alexa Fluor 488 wheat germ agglutinin (Invitrogen) and ProLong Gold antifade reagent doped with 40 ,6-diamidino2-phenylindole dihydrochloride (DAPI) (Invitrogen) were used to stain the cells. Tyton X (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)2-(4-sulfophenyl)-2H-tetrazolium (MTS)) (Promega) was used to perform cytotoxicity assays. Z-L-Lysine-NCA (lys NCA), di-Z-L-cystine ((Z-CysOH)2), L-cystine NCA (cys NCA), and aminomethylpyrene (AMP) were synthesized according to previously published literature procedures.25 Instrumentation. GPC analysis was performed on a Shimadzu liquid chromatography system fitted with a Wyatt DAWN HELEOS LS detector (λ = 658 nm), a Shimadzu RID-10 refractometer (λ = 633 nm), and a Shimadzu SPD-20A UVvis detector, using three identical PLgel columns (5 μm, MIXED-C) in series and HPLC grade DMF with 0.05 M LiBr as the mobile phase (1 mL/min). The oven temperature was set to 70 C to maintain an acceptable pressure across the system, and the detectors were temperature controlled to 25 C. Astra software (Wyatt Technology Corp.) was used to determine the molecular weight characteristics using known dn/dc values for poly(Z-L-lysine) (PZLL) in DMF (dn/dcPZLL = 0.101 mL/g (25 C)). 1H NMR spectroscopy was performed using a Varian Unity400 (400 MHz) spectrometer with the deuterated solvent as reference and a sample concentration of ca. 20 mg/ mL. DLS measurements were performed on a Malvern high-performance particle sizer (HPPS) with a HeNe laser (633 nm) at an angle of 173 and a temperature of 25 ( 0.1 C—initial sample concentrations of 10 mg/mL in DMF were used, and then serial dilutions were performed until stable spectra were obtained. All sample solutions were filtered using 0.45 μm filters. MALDI ToF MS was performed on a Bruker Autoflex III mass spectrometer operating in positive/linear mode. The analyte, matrix (dithranol), and cationization agent (KI or NaTFA) were dissolved in THF at concentrations of 10, 10, and 1 mg/ mL, respectively, and then mixed in a ratio of 10:1:1. 0.3 μL of this solution was then spotted onto a ground steel target plate, and the solvent was allowed to evaporate prior to analysis. FlexAnalysis (Bruker) was used to analyses the data. UVvis spectrophotometry was performed on a Shimadzu UV-2101PC spectrometer using quartz cuvettes with a 1 cm path length. A Cyflow Space (Partec GmbH) flow cytometer using an excitation wavelength of 633 nm was used to measure the intensity of the CCS polymers; at least 10 000 particles were analyzed in each experiment. A FLUOstar OPTIMA (BMG Labtech, Germany) plate reader was used to measure the MTS absorbance at λ = 492 nm. Confocal laser scanning microscopy was performed on a Leica TCS SP2 (Leica Microsystems, Heidelberg, Baden-W€urttemberg, Germany) powered by Ar/Kr and He/Ne lasers. The sample was viewed using an oil immersion 63 lens (1.32 numerical aperture), and the pinhole diameter was maintained at 1 airy unit. DAPI was excited at λ = 310 nm, Alexa Fluor 488 at λ = 488 nm, and Alexa Fluor 647 at λ = 633 nm. The emission filters were set at λ = 410450 nm for DAPI, λ = 500580 nm for Alexa Fluor 488, and λ = 650710 nm for Alexa Fluor 647. Images were recorded at a depth of 1020 μm from the surface of the glass coverslip. Leica confocal software was used to acquire images of 512 512 pixels that were the average of eight scans.
Synthesis of Heterodifunctional α-Thiol, ω-Folic Acid PEG (FA-PEG3000-SH). FA (0.071 g, 0.16 mmol) and DCC (0.034 g, 0.17
mmol) were dissolved in DMSO (15 mL) under argon, followed by addition of DMAP (0.019 g, 0.15 mmol) and HO-PEG-NHCO-C2H4-S-Trt 3470
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules
ARTICLE
Scheme 1. Synthesis of PEGFolic Acid Conjugate via DCC-Mediated Coupling
(0.050 g, 0.017 mmol). The reaction mixture was stirred in the dark at room temperature for 24 h and then centrifuged to remove the insoluble byproduct, dicyclohexylurea. The supernatant was concentrated in vacuo to ca. 3.0 mL in volume and precipitated into DCM (50 mL). The insoluble, unreacted FA was removed via centrifugation and the supernatant concentrated in vacuo. The resulting residue was redissolved in DCM (2 mL) and precipitated into diethyl ether (20 mL), and the precipitated product, FA-PEG-S-Trt, was collected via centrifugation and dried in vacuo (35 mg). UVvis spectrophotometry and MALDIToF MS were used to determine the conjugation efficiency (ca. 30%), and the mixture was not separated. MALDI-ToF MS: Mw = 3803, PDI = 1.00. 1H NMR (400 MHz, d6-DMSO) δH 2.10 (t, 2H, J = 7.0 Hz CH2CO), 2.19 (t, 2H, J = 7.0 Hz, CH2S), 3.12 (m, 2H, CH2N), 3.47 (m, 160H, 80CH2O) 3.65 (t, 2H, J = 6.0 Hz, CH2C), 4.51 (t, 2H, J = 6.0 Hz, CH2O), 7.197.32 ppm (m, 15H, 15ArH). (It is expected that the FA resonances are indistinguishable from the baseline since the 1H NMR spectrum only revealed resonances corresponding to PEG-S-Trt.) The Trt protecting group was removed by dissolving the FA-PEG-S-Trt (15 mg) in TFA (2.0 mL) to afford a clear yellow solution that turned colorless when the deprotection was complete by the dropwise addition of triethylsilane. The product FA-PEG3000-SH (containing 70% unfunctionalized HO-PEG3000-SH) was isolated by precipitation into diethyl ether (30 mL), followed by centrifugation and drying in vacuo (10 mg) for 1 h. This was used immediately for subsequent reactions to prevent oxidation of the thiols. Synthesis of Thiol PEG (MeO-PEG2000-SH). Thiolated PEG without folic acid conjugated was synthesized by acid-catalyzed esterification. MeO-PEG2000-OH (3.50 g, 1.75 mmol) was heated to 90 C under argon, and then 3-mercaptopropanoic acid (0.24 g, 2.28 mmol) and H2SO4 (0.034 g, 0.35 mmol) were added. The reaction mixture was stirred for 16 h, cooled to room temperature, and diluted with DCM (80 mL). The solution was washed with 1:1 saturated NaHCO3:H2O (80 mL 5) and saturated NaCl (80 mL), dried (MgSO4), filtered, and concentrated in vacuo to afford MeO-PEG2000-SH, 3.0 g (80%). MALDI-ToF MS: Mw = 2022, PDI = 1.01. 1H NMR (400 MHz, CDCl3) δH 2.66 (t, 2H, J = 6.0 Hz, CH2CO), 2.75 (q, 2H, J = 6.0 Hz, CH2S), 3.37 (s, 3H, CH3O), 3.493.77 (m, 90H, 45CH2O), 4.28 ppm (t, 2H, J = 4.0 Hz, CH2O).
Synthesis of Poly(Z-L-lysine)armpoly(L-cystine)core CCS Polymer 1. The method used was similar to that previously reported in the literature with slight modification.25 Briefly, lys NCA (1.00 g, 3.26 mmol) was dissolved in anhydrous DMF (10 mL) in a flame-dried, argon-purged two-necked flask. N-(Trimethylsilyl)allylamine (M/I = 40, 13.7 μL, 0.0817 mmol) was added, and the mixture was stirred at room temperature for 12 h to afford poly(Z-L-lysine) (PZLL) (Mw = 9.67 kDa, PDI = 1.12). Cys NCA
(MI/CL = 30, 0.716 g, 2.45 mmol) was added, and the mixture was stirred for a further 6 h to afford the crude CCS polymer (Mw = 358 kDa, PDI = 1.73, f = 22). This solution was added to AMP (0.283 g, 2.35 mmol) dissolved in DMF (20 mL), stirred for 3 h, and then precipitatied in methanol (250 mL). The precipitate was isolated via centrifugation, washed with chloroform (100 mL) and THF (100 mL), and dried in vacuo to afford CCS polymer 1, 1.05 g (ca. 70%). GPC-MALLS: Mw = 668.3 kDa, PDI = 1.83, f = 36. 1H NMR (400 MHz, d6-DMSO) δH 1.102.06 (m, CH2), 3.01 (br s, CH2N), 3.724.31 (m, CHN), 5.03 (br s, CH2O), 7.017.50 ppm (m, ArH). Refer to Supporting Information for calculation of f.
Synthesis of PEGylated Poly(PEG-b-Z-L-lysine)armpoly(Lcystine)core CCS Polymers 2PEG and 2PEG-FA. In two 4 mL vials fitted with stirrer bars, CCS polymer 1 (20 mg, 29.9 nmol) was dissolved in DMF (3 mL). To one, MeO-PEG2000-SH (2.5 equiv per arm, 5.26 mg, 2.52 μmol) was added, and to the other, FA-PEG3000-SH (2.5 equiv per arm, 7.57 mg, 2.52 μmol) was added. Photoinitiator DMPA (0.32 mg, 1.26 μmol) was added to each vial, the vials were sealed, and the reaction mixtures were stirred under UV irradiation (λ = 256 nm) for 3 h. The resulting solutions were precipitated into 1:1 ethyl acetate: diethyl ether (50 mL) and the precipitates isolated via centrifugation and dried in vacuo to afford PEGylated CCS polymers 2PEG (18 mg, 81%) and 2PEG-FA (16 mg, 68%), respectively. GPC MALLS: (2PEG) Mw = 710.2 kDa, PDI = 1.85, f = 36; (2PEG-FA) Mw = 811.5 kDa, PDI = 1.75, f = 36. 1H NMR (400 MHz, d6-DMSO, identical spectra were obtained for both 2PEG and 2PEG-FA) δH 1.231.86 (m, CH2), 2.91 (br s, CH2N), 3.46 (br s, CH2O), 3.724.26 (m, CHN), 4.96 (br s, CH2O), 6.997.32 ppm (m, ArH).
Synthesis of Poly(PEG-b-L-lysine)armpoly(L-cystine)core CCS Polymers 3PEG and 3PEG-FA. Separately, CCS polymers 2PEG and
2PEG-FA (10 mg) were dissolved in TFA (50 μL), and 33% HBr in acetic acid was then added (200 μL). After stirring for 1 h at room temperature the mixture was precipitated into diethyl ether (2.5 mL). The residue was isolated via centrifugation, redissolved in water (2 mL), and precipitated into THF (30 mL); this step was repeated three times and the residue dried in vacuo (0.1 mbar) to afford CCS polymers 3PEG (5.2 mg) and 3PEG-FA (3.2 mg) as off-white solids. 1H NMR (400 MHz, D2O, identical spectra were obtained for both 3PEG and 3PEG-FA) δH 1.20 1.64 (m, CH2), 2.86 (br s, CH2N), 3.65 (br s, CH2O), 4.184.20 (m, CHN).
Fluorescent Tagging of CCS Polymers 3PEG and 3PEG-FA with Alexa Fluor 647. The CCS polymers 3PEG and 3PEG-FA were dissolved in water (0.5 mg/mL), and Alexa Fluor 647 carboxylic acid succinimidyl ester dissolved in DMSO (1 mg/mL) was added (2.5 μL/mg of polymer). The mixtures were stirred for 2 h at room temperature and then precipitated into THF (10 times the reaction volume). The 3471
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules resulting solid was isolated via centrifugation and dried in vacuo to afford the fluorescently tagged derivatives. Cell Culture. MDA-MB-231 human breast cancer cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with penicillin (100 units/mL), streptomycin (100 mg/mL), fetal bovine serum (10% v/v), and 2 mM L-glutamine at 37 C, 5% CO2, and 95% relative humidity. Confluent cultures were rinsed in Dulbecco’s phosphate buffered saline (DPBS) and split (1:3) every 3 days using trypsin ethylenediaminetetraacetic acid (EDTA) (1 liquid, 1 mL per T25 flask). Cellular Uptake of CCS Polymers 3PEG and 3PEG-FA. For confocal microscopy experiments, sterilized glass coverslips were placed in a 6-well tissue culture plate and coated with fibronectin from bovine plasma (10 μg/mL in PBS) for 1 h. The excess fibronectin was removed by washing with PBS. MDA-MB-231 cells were detached from culture using trypsinEDTA and seeded onto the coated coverslips at a density of 10 000 cells/well and then allowed to attach at 37 C for 1 h. The cells were then treated with 50 μL of 3PEG and 3PEG-FA polymer solutions (1 mg/mL in PBS), and the polymers were incubated with the cells for 3 h at 37 C and 5% CO2. The coverslips were gently washed with warmed DPBS then fixed in 4% w/v paraformaldehyde in PBS for 20 min. The cell membrane was stained with AlexaFluor 488 conjugated wheat germ agglutinin overnight. Coverslips were washed with PBS and then mounted onto glass slides using a DAPI-doped glycerol antifade medium. For flow cytometry experiments, the cells were seeded directly in a 6-well culture plate and treated with the star polymers as described for the confocal microscopy experiments. Following an incubation period of 3 h, the cells were detached from the plate using trypsinEDTA (100 μL), pelleted at 500g for 5 min, and then resuspended in PBS (1 mL) and left on ice. Samples were analyzed by flow cytometry performed on a Cyflow Space flow cytometer using an excitation wavelength of λ = 633 nm. 10 000 counts were collected for each run. Forward and side scatter data were used to fix a gate around live cell population. This gate was used to identify the population which demonstrated a higher fluorescence intensity compared to the cells alone. Measurement of Cell Cytotoxicity. Breast cancer cells (MDAMD-231) were detached from the culture vessel using trypsinEDTA, and the cell density was adjusted to 60 000 cells/mL with DMEM. A 96well plate was then seeded with the cells in quadruplicate (3000 cells/ well) and left at 37 C for 30 min before addition of the star polymers. For concentration-dependent toxicity testing different concentrations (0, 2.5, 5.0, 12.5, 5.0, and 100 μg/mL) of star polymers 3PEG and 3PEG-FA were added to the cells and incubated for 72 h. For time-course studies cells were incubated with polymers at 5.0 μg/mL for 24, 48, and 72 h, where each time point was conducted on a separate plate. After the indicated time period MTS assays were conducted according to the manufacturers protocol. The cells were incubated with the MTS solution for 3 h followed by measurement of absorbance at λ = 492 nm in a microtiter plate reader. A student t test was performed to determine statistical significance of differences between treatments (p < 0.05). Each experiment was repeated on two separate occasions, and cells that were untreated or lysed with 2 μL of Triton X-100 prior to the addition of MTS solution were used as controls for live and dead cells, respectively.
’ RESULTS AND DISCUSSION Synthesis of Folic AcidPEG Conjugate. Thiolated and folic acid (FA) conjugated PEG (FA-PEG3000-SH) was synthesized via reaction of HO-PEG3000-S-Trt with FA using DCC-mediated coupling, followed by deprotection of the trityl (Trt) group. It has previously been demonstrated that the coupling reaction occurs predominately between the hydroxyl group of the PEG and the γ-carboxylic acid of FA,33,34 which allows the conjugated FA to maintain a high affinity toward folate receptors on
ARTICLE
Figure 1. (A) MALDI-ToF MS spectrum of FA-PEG3000-S-Trt recorded in linear/positive mode using dithranol/KI. The numbers on the spectrum represent the number of PEG repeat units (n, 44.01 m/z). Inset shows the expanded section of the spectrum complete with peak m/z values and m/z differences. The major series, M1, represents the unreacted starting material, and the minor series, M2, represents the desired FA-PEG3000-S-Trt. (B) Normalized UVvis spectra of (i) FAPEG3000-S-Trt, (ii) HO-PEG3000-S-Trt, and (iii) FA in DCM.
endothelial tumor cells (Scheme 1).30 As a result, the PEG-FA conjugate could offer a “stealth” layer for biocompatibility and increased circulation16,3537 as well as facilitate active targeting2830,33,34 once incorporated onto the star polymers. The synthesis of FA-PEG3000-S-Trt was confirmed by MALDIToF MS and UVvis spectrophotometry (Figures 1A and 1B, respectively). The major series in the MALDI-ToF MS spectrum correlates to the unreacted starting material, with the minor series corresponding to the desired FA-PEG3000-S-Trt conjugate (represented as series M1 and M2, respectively, in Figure 1A). 3472
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules
ARTICLE
Scheme 2. Synthesis of Amino Acid-Based CCS Polymers via a One-Pot, Arm-First Strategy
The normalized UVvis spectra of the starting materials (HO-PEG3000-S-Trt and FA) and resulting product (FAPEG3000-S-Trt) in DCM (Figure 1B) revealed an absorbance at λ = 280 nm and a broadening of the peak between λ = 220270 nm for the FA-PEG3000-S-Trt, which matches the FA absorbance profile. Additionally, pure FA is only slightly soluble in DCM and displays a low absorbance at λ = 280 nm corresponding to the aromatic chromophore. Therefore, the more pronounced peak at λ = 280 nm for the product confirms the successful incorporation of FA and synthesis of FA-PEG3000-STrt.9 From both MALDI-ToF MS and UVvis spectrophotometry, the extent of FA conjugation was calculated to be ca. 30%, which was deemed sufficient to provide active targeting once coupled to the star polymer.38,39 Thus, the mixture of FA functionalized and unfunctionalized PEG was not further purified and was used directly for PEGylation of the CCS polymer (vide infra). 1H NMR spectroscopic analysis was also used to characterize the product; however, due to the low ratio of FA end-groups to the main chain repeat units, the FA proton resonances could not be distinguished from the baseline. The Trt protecting group was subsequently removed by reaction of FA-PEG3000-S-Trt with triethylsilane in TFA to afford the desired thiolated derivative, FA-PEG3000-SH (also containing 70% unfunctionalized HO-PEG3000-SH), which was used for the subsequent reaction immediately after isolation to prevent oxidation. In order to confirm the deprotection of the Trt group, 1 H NMR spectroscopy was performed on a separate sample of HO-PEG3000-S-Trt, which revealed the successful removal of the Trt protecting groups when subjected to the same deprotection conditions (Supporting Information, Figure S1).40 For control experiments a thiolated PEG derivative (MeOPEG2000-SH) without conjugated FA was synthesized by reacting
MeO-PEG2000-OH with 3-mercaptopropanoic acid. The absence of the FA targeting moiety on this PEG derivative allowed the difference in the in vitro cellular interaction between star polymers with and without FA to be studied. The synthesis of MeO-PEG2000SH was confirmed via MALDI-ToF MS and 1H NMR spectroscopic analysis (Supporting Information Figures S2A and S2B, respectively), which revealed quantitative conversion. Synthesis of CCS Polymers 1. The synthesis of star polymer 1 was achieved via ROP of amino acid NCAs using an arm-first, one-pot strategy as outlined in Scheme 2. The ROP of lys NCA using N-(trimethylsilyl)allylamine as the initiator yielded living PZLL (Mw = 9.67 kDa, PDI = 1.12), which served as the macroinitiator (MI) for star formation (Figure 2A, PZLL). Subsequent addition of the cross-linker (CL) cys NCA ([CL]/[MI] = 30) afford an intermediate star polymer, having a Mw of 358 kDa, PDI of 1.73, and f (average number of arms) of 22. The GPC differential refractive index (DRI) chromatogram of the intermediate (Figure 2A, intermediate) revealed a slight shoulder to the right of the star peak, with a retention time similar to the precursor PZLL (ca. 22.5 min), indicating the presence of some unincorporated PZLL derivatives. From the authors’ previous studies,25 it was determined that some pendant NCAs remained unreacted in the core after star formation, which allowed for postpolymerization functionalization with aminomethylpyrene (AMP) to afford CCS polymer 1. After isolation in MeOH and washing with chloroform and THF to remove the excess AMP, GPC analysis revealed an increase in Mw to 668.3 kDa (PDI = 1.89, f = 36) (Figure 2A, 1). Determination of the pyrene loading via UVvis spectrophotometry provided a value of 261 mol/mol star, which equates to an increase in molecular weight of ca. 60 kDa. Therefore, the increase in the Mw observed by GPC (ca. 300 kDa) can be 3473
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules
Figure 2. (A) GPC DRI chromatograms and (B) DLS traces in DMF of PZLL macroinitiator, intermediate CCS polymer prior to core functionalization with AMP, CCS polymer 1, and PEGylated CCS polymers 2PEG and 2PEG-FA. Each DLS result is the average of at least three independent measurements.
attributed mainly to the occurrence of starstar coupling during the functionalization and isolation process.25 The disappearance of the shoulder observed in the GPC DRI chromatogram of the intermediate (Figure 2A, intermediate) indicated that the free PZLL derivatives (∼22.5 min) had either been incorporated into the star or removed during the washing procedure with THF, in which PZLL was slightly soluble. DLS analysis was performed at each reaction stage to follow the change in the hydrodynamic radius (dH) of the polymers (Figure 2B). The macroinitiator PZLL was found to have a dH of 8.4 nm, which increased to 86 nm upon formation of the intermediate star. DLS analysis of the intermediate star also confirmed the presence of some unincorporated poly(Z-L-lysineb-L-cystine) (P(ZLL-b-LC)) copolymer, with a dH of 15 nm. After postpolymerization functionalization with AMP the dH of star 1 was determined to be 110 nm, and unincorporated P(ZLL-b-LC) was no longer observed, which correlates well with the GPC results. PEGylation of CCS Polymer 1. The peripheral allyl groups of star 1 derived from the amine initiator used for the ROP of amino acid NCAs and star formation25,26 allowed for facile functionalization of the stars via thiolene click chemistry (Scheme 2). Since star 1 is composed entirely of peptide chains, it would be prone to bloodstream enzymatic degradation, plasma opsonization, and monocyte phagocytic system (MPS) uptake, which is undesirable for targeted drug delivery vehicles. PEGylation is a widely used technique to shield therapeutic agents from the host’s immune system and was employed in this study to protect the stars against rapid degradation into their amino acid constituents and prolong their circulation time in the bloodstream. Two types of PEGylated stars, 2PEG and 2PEG-FA, were synthesized via thiolene click of star 1 with thiolated PEG derivatives
ARTICLE
Figure 3. MALDI-ToF MS spectrum of control degradation experiment using FA-PEG3000-S-Trt in HBr, recorded in linear/positive mode using dithranol/KI. The numbers on the spectrum represent the number of PEG repeat units (n, 44.01 m/z). Inset shows the expanded section of MS spectrum complete with peak m/z values and m/z differences. The major series, M10 and M20 , represent the starting material and PEG-FA conjugate without the Trt protecting groups, respectively, and the minor series M1 represents the unreacted starting material.
MeO-PEG-SH and FA-PEG3000-SH, respectively, in the presence of a photoinitiator, where the former does not possess targeting capabilities and the latter has 30% FA-conjugated for targeted drug delivery. GPC DRI chromatograms of both stars 2PEG and 2PEG-FA revealed retention times higher than those of the precursor star 1 (Figure 2A, 2PEG and 2PEG-FA, respectively). The dn/dc values of densely branched CCS polymers and linear polymers of the same monomeric constitution have been reported to be comparable.41 Therefore, using the dn/dc value calculated for PEGylated linear PZLL (dn/dc = 0.065 mL/g), GPC analysis provided increases in the molecular weight for both stars 2PEG and 2PEG-FA. For 2PEG the Mw was determined to be 710.2 kDa (PDI = 1.85), which correlates to ca. 58% PEGylation of the arms, whereas for 2PEG-FA the Mw of 811.5 kDa (PDI = 1.75) correlated to >99% PEGylation. The slight peak shift to higher retention time might result from changes in the physical properties of the star polymers after PEGylation, which affects their interaction with the GPC columns. The appearance of peaks between 20 and 25 min was attributed to the unreacted excess PEG, and their dimerized derivatives formed as a result of disulfide bond formation. However, the extent of PEGylation calculated from 1H NMR spectroscopic analysis provided values of ca. 30 and 50% for 2PEG and 2PEG-FA, respectively (Supporting Information, Figure S3). The discrepancies between the GPC and 1H NMR spectroscopy calculations are expected to arise from a combination of resonance broadening 3474
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules
ARTICLE
Figure 4. Confocal microscopy images of (A) untreated breast cancer MDA-MB-231 cells and cells incubated with (B) 3PEG and (C,D,E,F) 3PEG-FA. Images D and F show expanded 3D reconstructions of the cell membrane and nucleus from image E with cross sections through the sample in the z-direction. The cell membrane was stained with Alexa Fluor 488 (green), the cell nucleus with DAPI (blue), and stars 3PEG and 3PEG-FA with Alexa Fluor 647 (red) in all images. The scale bars are (AC, E) 10 μm, (D) 5 μm, and (F) 3 μm in length.
and overlap in the NMR measurements and the dn/dc value used in GPC calculations, which assumes that all of the star’s arms are PEGylated. DLS analysis also revealed increases in size after PEGylation, with dH values of 238 and 262 nm for stars 2PEG and 2PEG-FA, respectively (Figure 2B). This marked increase after PEGylation may result from aggregation.42 However, this size regime ensures that the drug delivery vehicles will not be able to enter the healthy/normal vasculature but only to the “leaky” tumor vasculature due to enhanced permeability and retention (EPR) effects, which provides passive targeting abilities in addition to the active targeting provided by the conjugated folic acid moieties. Deprotection of PEGylated CCS Polymers 2. The removal of the pendent Cbz protecting groups along the arms of stars 2PEG and 2PEG-FA was achieved using HBr to afford water-soluble CCS polymers 3PEG and 3PEG-FA, respectively, with pendent amine functionalities available along the arms. 1H NMR spectroscopic analysis revealed ca. 80% removal of the protecting groups for both 3PEG and 3PEG-FA (Supporting Information, Figure S4). To ensure that the conjugated FA present on 3PEG-FA was stable during the acid-mediated deprotection, a control experiment was conducted by exposing FA-PEG3000-S-Trt to HBr for 6 h, which was longer than the time required to deprotect star 2PEG-FA (1 h). MALDI-ToF MS of the control experiment after 6 h showed that the conjugated FA remained intact during the process despite the fact that 85% of the Trt protecting groups were removed (Figure 3). Cellular Uptake of CCS Polymers 3PEG and 3PEG-FA. Confocal microscopy (Figure 4) and flow cytometry (Figure 5) were used to study the in vitro interaction between cells and stars 3PEG and 3PEG-FA tagged with Alexa Fluor 647 carboxylic succinimidyl ester (red) through the reaction with the pendent amine groups along the
Figure 5. Histogram from flow cytometry of breast cancer MDA-MB231 cells incubated with fluorescently labeled stars 3PEG and 3PEG-FA for 3 h and no star polymer for the control.
stars’ arms. MDA-MB-231 cells were chosen as model cancer cells.43 After incubation with MDA-MB-231 for 3 h, confocal microscopy confirmed that the cells incubated with star 3PEG-FA showed a number of red fluorescent stars associated with the cells. This included stars associated with the nucleus (stained with DAPI, blue), the cell membrane (stained with wheat germ agglutinin conjugated Alexa Fluor 488, green), or other parts of 3475
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules
Figure 6. Viability of breast cancer cells (MDA-MB-231) after incubation with stars 3PEG-FA and 3PEG at (A) different concentration for 72 h and (B) at 5.0 μg/mL for 24, 48, and 72 h, as assessed by MTS assays. Absorbance is proportional to MTS reduction by metabolically active cells. The data are normalized relative to live cells receiving no treatment and is the mean ( the standard deviation (n = 4). The results are representative of two separate assays.
the cell suggesting cellular association and uptake of the star polymer 3PEG-FA (Figure 4C,E). The 3D reconstructions in Figure 4D,F show expanded 3D views of the nucleus and area around the nucleus for two cells highlighted in Figure 4E and clearly confirm that some of the 3PEG-FA appear within the cross section of the cell (Figure 4D and Supporting Information Figure S5) as well as the cell nucleus (Figure 4F). In comparison, the cells incubated with 3PEG had considerably fewer fluorescent stars (Figure 4B), and these appear restricted to the membrane within 3D reconstructions of the cell (Supporting Information Figure S6). This result was expected since star 3PEG did not possess a targeting ligand, and therefore interactions with the cells were nonspecific compared to the receptor mediated attachment and internalization of star 3PEG-FA. Flow cytometry was used to confirm and quantify the association MDA-MB-231 cells with Alexa Fluor 647-labeled stars. Figure 5 shows that after 3 h coincubation only 13% of cells incubated with 3PEG were associated with these particles compared to 55% for cells incubated with 3PEG-FA (Figure 5), consistent with observations by confocal microscopy. The histograms show a broad distribution of the Alexa Fluor 647positive cells most likely as a result of uneven distribution of
ARTICLE
the stars across all cells due to aggregation (also evident from the confocal images (Figure 4CF)), which has been previously observed even in PEGylated systems.42 Nevertheless, these results further confirm that star 3PEG-FA showed preferential attachment to the cells due to the presence of periphery conjugated FA, and a greater number of these star polymers appear internalized. Cell Cytotoxicity Study. The viability of both stars 3PEG and 3PEG-FA as nontoxic drug carriers was assessed by incubation with MDA-MB-231 breast cancer cells in vitro. The star polymers were dissolved in sterile PBS at different concentrations (0, 2.5, 5.0, 12.5, 50.0, and 100 μg/mL) and were incubated with MDA-MB-231 cells. Cell viability was assessed after 72 h via a standard MTS assay, which measures the metabolic activity of live cells. Figure 6A shows that cell viability was comparable between nearly all samples and the control, which received no star polymers. However, addition of the star 3PEG-FA at 100 μg/mL induced a significant decrease (p < 0.05) in cell proliferation of up to 15% after 72 h. This can potentially be explained by the internalization of 3PEG-FA star polymers via FA receptor mediated interactions, which may slow cell proliferation. A second MTS assay using star polymers at a lower concentration (5.0 μg/mL) where viability was assessed after coincubation with cells for 24, 48, and 72 h (Figure 6B) indicated that during the first 24 h the cells incubated with 3PEG-FA had significantly reduced viability (p < 0.05) up to 15% compared to the untreated control. After 48 and 72 h, however, cell viability was comparable to the control, suggesting that any reductions in viability at low concentration are temporary. Changes in cell viability may be due to a decrease in the polymer to cell ratio as the cells proliferate or cellular processing of the polymer.
’ CONCLUSION Core cross-linked star (CCS) polymer 1 (poly(L-lysine) armpoly(L-cystine)core) composed entirely of amino acid building blocks have been synthesized via ring-opening polymerization (ROP) of amino acid N-carboxyanhydrides (NCAs) in a onepot, arm-first strategy using N-(trimethylsilyl)allylamine as the initiator. The resulting allyl groups at the stars’ periphery allowed for PEGylation via thiolene click chemistry to yield CCS polymers 2PEG and 2PEG-FA without and with conjugated folic acid (FA), respectively. Deprotection of the Cbz groups along the arms yielded water-soluble CCS polymers 3PEG and 3PEG-FA. Concentrations of up to 50 μg/mL of both PEGylated CCS polymers showed no effect on MDA-MB-231 breast cancer cell viability for up to 72 h in vitro. Confocal microscopy and flow cytometry revealed cell association and internalization of star 3PEG-FA within 3 h of incubation for 55% of cells, whereas a low level of nonspecific attachment was observed for 13% of cells exposed to star 3PEG. These results suggest that the star polymers with FA described here have the potential to be used as targeted drug delivery vehicles. When combined with previous work showing the capability of these polymers to encapsulate chemotherapeutic agents at high rates, these multifunctional stars show great promise as targeted biocompatible vehicles with the potential for targeted delivery, fluorescent tagging, high circulation time, and low toxicity. Studies are now underway to assess the controlled release of drugs from this amino acid-based star polymer. 3476
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477
Biomacromolecules
’ ASSOCIATED CONTENT
bS
Supporting Information. Figures S1S6. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Tel: +61 3 8344 8665. Fax: +61 3 8344 4153.
’ ACKNOWLEDGMENT The authors thank Dr. Angus Johnston and Ms. Sher Leen Ng for their help with confocal microscopy and flow cytometry. A.S. is a recipient of CSIRO Food Future Flagship Scholarship. The MALDI ToF MS used in these studies was supported under the Australian Research Council’s Linkage Infrastructure, Equipment and Facilities (LIEF) funding scheme (LE0882576). ’ REFERENCES (1) Murakami, M.; Cabral, H.; Matsumoto, Y.; Wu, S.; Kano, M. R.; Yamori, T.; Nishiyama, N.; Kataoka, K. Sci. Transl. Med. 2011, 64, 64ra2. (2) Zhou, H.; Yu, W.; Guo, X.; Liu, X.; Li, N.; Zhang, Y.; Ma, X. Biomacromolecules 2010, 11, 3480–3486. (3) Vicent, M. J.; Duncan, R. Trends Biotechnol. 2006, 24, 39–47. (4) Maeda, H.; Greish, K.; Fang, J. Adv. Polym. Sci. 2006, 193, 103–121. (5) Kataoka, K.; Harada, A.; Nagasaki, Y. Adv. Drug Delivery Rev. 2001, 47, 113–131. (6) Sun, H.; Guo, B.; Cheng, R.; Meng, F.; Liu, H.; Zhong, Z. Biomaterials 2009, 30, 6358–6366. (7) Ryu, J.; Roy, R.; Ventura, J.; Thayumanavan, S. Langmuir 2010, 26, 7086–7092. (8) Chan, Y.; Wong, T.; Byrne, F.; Kavallaris, M.; Bulmus, V. Biomacromolecules 2008, 9, 1826–1836. (9) Pan, D.; Turner, J. L.; Wooley, K. L. Chem. Commun. 2003, 2400–2401. (10) Nystroem, A. M.; Xu, Z.; Xu, J.; Taylor, S.; Nittis, T.; Stewart, S. A.; Leonard, J.; Wooley, K. L. Chem. Commun. 2008, 3579–3581. (11) Johnston, A. P. R.; Cortez, C.; Angelatos, A. S.; Caruso, F. Curr. Opin. Colloid Interface Sci. 2006, 11, 203–206. (12) Liu, M.; Kono, K.; Frechet, J. M. J. J. Controlled Release 2000, 65, 121–131. (13) Cho, H. Y.; Gao, H.; Srinivasan, A.; Hong, J.; Bencherif, S. A.; Siegwart, D. J.; Paik, H.; Hollinger, J. O.; Matyjaszewski, K. Biomacromolecules 2010, 11, 2199–2203. (14) Wiltshire, J.; Qiao, G. Aust. J. Chem. 2007, 60, 699–705. (15) Schramm, O. G.; Meier, M. A. R.; Hoogenboom, R.; Erp, H. P.; Gohy, J. F.; Schubert, U. S. Soft Matter 2009, 5, 1662–1667. (16) Kojima, C.; Kono, K.; Maruyama, K.; Takagishi, T. Bioconjugate Chem. 2000, 11, 910–917. (17) Ho, A.; Gurr, P.; Mills, M.; Qiao, G. Polymer 2005, 46, 6727–6735. (18) Goh, T. K.; Coventry, K.; Blencowe, A.; Qiao, G. Polymer 2008, 49, 5095–5104. (19) Blencowe, A.; Goh, T. K.; Best, S.; Qiao, G. Polymer 2008, 49, 825–830. (20) Wiltshire, J.; Qiao, G. Macromolecules 2008, 41, 623–631. (21) Helms, B.; Guillaudeu, S. J.; Xie, Y.; McMurdo, M.; C. J. Hawker, C. J.; Frechet, J. M. J. Angew. Chem., Int. Ed. 2005, 44, 6384–6387. (22) Spiniello, M.; Blencowe, A.; Qiao, G. J. Polym. Sci., Polym. Chem. 2008, 46, 2422–2432. (23) Blencowe, A.; Tan, J. F.; Goh, T. K.; Qiao, G. Polymer 2009, 50, 5–32.
ARTICLE
(24) Liu, Z.; Cheng, R.; Meng, F.; Cui, J.; Ji, S.; Zhong, Z. Angew. Chem., Int. Ed. 2009, 48, 9914–9918. (25) Sulistio, A.; Widjaya, A.; Blencowe, A.; Zhang, X.; Qiao, G. Chem. Commun. 2011, 47, 1151–1153. (26) Lu, H.; Cheng, J. J. Am. Chem. Soc. 2008, 130, 12562–12563. (27) Matesic, L.; Locke, J. M.; Vine, K. L.; Ranson, M.; Bremner, J. B.; Skropeta, D. Bioorg. Med. Chem. 2011, 19, 1771–1778. (28) Lu, Y.; Low, P. S. Adv. Drug Delivery Rev. 2002, 54, 675–693. (29) Ross, J. F.; Chaudhuri, P. K.; Ratnam, M. Cancer 1994, 73, 2432–2443. (30) Sudimack, J.; Lee, R. J. Adv. Drug Delivery Rev. 2000, 41, 147–162. (31) Weitman, S. D.; Lark, R. H.; Coney, L. R.; Fort, D. W.; Frasca, V.; Zurawski, V. R., Jr.; Kamen, B. A. Cancer Res. 1992, 52, 3396–3401. (32) Canal, F.; Vicent, M. J.; Pasut, G.; Schiavon, O. J. Controlled Release 2010, 146, 388–399. (33) Chen, S.; Zhang, X. Z.; Cheng, S. X.; Zhuo, R. X.; Gu, Z. W. Biomacromolecules 2008, 9, 2578–2585. (34) Saul, J. M.; Annapragada, A.; Natarajan, J. V.; Bellamkonda, R. V. J. Controlled Release 2003, 92, 49–67. (35) Greenwald, R. B.; Conover, C. D.; Choe, Y. H. Crit. Rev. Ther. Drug Carrier Syst. 2000, 17, 101–161. (36) Greenwald, R. B.; Choe, Y. H.; McGuire, J.; Conover, C. D. Adv. Drug Delivery Rev. 2003, 55, 217–250. (37) Berna, M.; Dalzoppo, D.; Pasut, G.; Manunta, M.; Izzo, L.; Jones, A. T.; Duncan, R.; Veronese, F. M. Biomacromolecules 2006, 7, 146–153. (38) Bae, Y.; Nishiyama, N.; Lataoka, K. Bioconjugate Chem. 2007, 18, 1131–1139. (39) De, P.; Gondi, S. R.; Sumerlin, B. S. Biomacromolecules 2008, 9, 1064–1070. (40) Badyal, J. P.; Cameron, A. M.; Cameron, N. R.; Coe, D. M.; Cox, R.; Davis, B. G.; Oates, L. J.; Oyea, G.; Steel, P. G. Tetrahedron Lett. 2001, 42, 8531–8533. (41) Gao, H.; Tsarevsky, N. V.; Matyjaszewski, K. Macromolecules 2005, 38, 9018–9027. (42) Wenig, L.; Kilcher, G.; Tirelli, N. Macromol. Biosci. 2007, 7, 987–998. (43) Jhaveri, M. S.; Rait, A. S.; Chung, K. N.; Trepel, J. B.; Chang, E. H. Mol. Cancer Ther. 2004, 12, 1505–1512.
3477
dx.doi.org/10.1021/bm200604h |Biomacromolecules 2011, 12, 3469–3477