Biomacromolecules 2005, 6, 2875-2880
2875
Preparation and Evaluation of Polyvinyl alcohol-co-oleylvinyl ether Derivatives as Tumor-Specific Cytotoxic Systems I. Orienti,*,† G. Zuccari,† R. Carosio,‡ and P. G. Montaldo‡ Department of Pharmaceutical Sciences, University of Bologna, Italy, and Laboratory of Oncology, G. Gaslini Institute, Genoa, Italy Received June 22, 2005
A series of poly(vinyl alcohol) amphiphilic derivatives have been prepared to obtain polymeric aggregates in aqueous phase holding thermodynamic instability. The aim was to evaluate their ability to interact with tumor cells eliciting selective cytotoxicity. The poly(vinyl alcohol) derivatives were prepared by partial substitution of poly(vinyl alcohol) (MW 10 kDa) with both oleyl chains and poly(ethylene glycol) monoethyl ethers (PEGMEE) of different molecular weights. The substitution degree was 1.5% for the oleyl chains and 1% for the PEGMEE chains (moles of substituent per 100 mol of hydroxyvinyl monomer). The polyvinyl derivatives obtained easily dissolved in water. Dynamic and static light scattering measurements on the polymer aqueous solutions indicated the formation of polymeric aggregates characterized by low polydispersity (0.232-0.299) and mean size (218-382 nm) in the range suitable for intravenous administration. Moreover, they were characterized by different packing densities and thermodynamic instabilities driving the polymers to interact with hydrophobic membranes. Among the analyzed polymers, the poly(vinyl alcohol)-co-oleylvinyl ether substituted with triethylene glycol monoethyl ether (P10(4)) provided in solution the highest affinity for hydrophobic membranes. P10(4), moreover, was the most cytotoxic toward the tumor cell lines analyzed (neuroblastoma: SH-SY5Y, IMR-32, HTLA-230. melanoma: MZ2-MEL, RPMI7932.), while it did not appreciably alter the viability of the normal resting lymphocytes. The peculiar behavior of the P10(4) aggregates has been correlated to their high thermodynamic instability in solution due to the high packing density that triggers the polymeric aggregates to interact with hydrophobic membranes such as the tumor cell membranes, thus eliciting cytotoxicity. Introduction As is well-known, tumor cell membranes are characterized by abnormalities with respect to the normal cells. The most recurrent abnormalities are conformational changes, enhanced lateral diffusion of membrane molecules, alteration of membrane-cytoskeleton attachments, alteration of absolute transmembrane potential,1 and increase in membrane fluidity.2-4 These abnormalities drive the peculiar tumor cell-environment interactions which contribute to the development and maintenance of the malignant phenotype.1 Among the studies aimed at exploiting these abnormalities as possible targets for selective cytotoxicity, no one points to the increased fluidity of the tumor cell membrane. The fluidity increase could be an interesting target, as raising the free energy of the membrane5-7 may promote cell interactions with thermodynamically unstable environmental structures, such as amphiphilic aggregates, on the basis of maximal decrease in free energy-maximal affinity for ligands.8,9 The amphiphilic aggregates, both formed by low molecular weight surfactants or amphiphilic polymers, are * Corresponding author: Prof. Isabella Orienti, Department of Pharmaceutical Sciences, University of Bologna, Via San Donato 19/2, 40127 Bologna (Italy), Tel-FAX +39 051 253664. E-mail: orienti@ biocfarm.unibo.it. † University of Bologna. ‡ G. Gaslini Institute.
Figure 1. Poly(vinyl alcohol) substituted with oleyl chains and PEGMEE chains of different molecular weights; n values range from 1 to 7.
characterized by inherent thermodynamic instability;10-12 thus, they may be considered possible candidates to selectively interact with tumor cells. The aim of this work was to evaluate if a series of amphiphilic polymers, self-assembling in aqueous environment, could selectively interact with tumor cells providing cytotoxicity. To this purpose, we substituted poly(vinyl alcohol) with both hydrophobic oleyl chains and hydrophilic PEGMEE chains of different molecular weights (Figure 1). In previous works, we prepared amphiphilic polymers by poly(vinyl alcohol) partial substitution with hydrophobic acyl chains through ester bonds13 or carbamate bonds.14 These polymers formed micelles in aqueous environment due to aggregation of the grafted acyl chains but did not display cytotoxicity either toward tumor cells or toward normal cells, indicating that simple polymer aggregation does not provide the thermodynamic instability sufficient to promote cytotoxicity through polymer-cell interaction. In this work, we linked the poly(vinyl alcohol) backbone to the hydrophobic oleyl substituent through an ether rather
10.1021/bm0504328 CCC: $30.25 © 2005 American Chemical Society Published on Web 08/09/2005
2876
Biomacromolecules, Vol. 6, No. 5, 2005
than an ester or carbamate bond to decrease the steric hindrance between the hydrophobic chains. This should favor their aggregation in more densely packed structures.15 At the same time, we substituted the polyvinyl chain with different molecular weight poly(ethylene glycol) monoethyl ethers to improve the hydration ability of the hydrophilic portions and thus further improve the packing density of the polymeric aggregates, according to the headgroup area-hydrocarbon chain volume balance determining the packing structure in micelles, vesicles, and liposomes.16,17 The aim was to obtain polymeric aggregates in solution endowed with high thermodynamic instability16,17 and evaluate if a selective cytotoxicity toward tumor cells could be achieved. Materials and Methods Preparation of the Substituted PVA. Poly(vinyl alcohol) (PVA, MW 10 kDa, 80% hydrolyzed) was a commercial sample from Aldrich Chemical Gmbh (Steinheim, Germany). Oleyl bromide was purchased from Sigma Chemical Co. (St. Louis, MO); tetraethylene glycol monoethyl ether and heptaethylene glycol monoethyl ether were from Kyowa Hakko Europe Gmbh (Dusseldorf, Germany); ethylene glycol monoethyl ether, diethylene glycol monoethyl ether, triethylene glycol monoethyl ether, 1,8-diazabicyclo[5.4.0]undec7-ene (DBU), and all the other reagents and solvents employed were from Fluka Chemie GmbH (Buchs, Suisse). Self-assembling polymers were prepared by partial substitution of PVA with both oleyl chains and poly(ethylene glycol) monoethyl ethers. The synthesis was carried out by dissolving PVA (3.93 g of polymer corresponding to 75 mmol of hydroxyvinyl monomer) in 75 mL of N-methylpyrrolidone (NMP) in the presence of DBU (75 mmol). The solution was stirred at room temperature for 24 h, and subsequently, ethyl ether was added to induce precipitation of the PVADBU complex. The yellow rubbery solid obtained was dissolved in 75 mL of NMP, and oleyl bromide was subsequently added in a molar amount (15 mmol) corresponding to 20% of the hydroxyvinyl monomers present in solution. The solution was stirred at room temperature for 24 h. Poly(ethylene glycol) monoethyl ether monochlorides were then added in molar amounts corresponding to 50% of the hydroxyvinyl monomers present in solution (37.5 mmol). The obtained solutions were stirred at room temperature for 24 h, and subsequently, diethyl ether was added to induce precipitation of the substituted polymer. The solid obtained was reprecipitated twice from NMP, dried under vacuum to constant weight, subsequently dissolved in water, and dialyzed against water for 15 days. After dialysis, the aqueous solution of the polymer was lyophilized. The final solid product was stored in a desiccator. Poly(ethylene glycol) monoethyl ether monochlorides were prepared following a method for halogenation of alcohols.18 Briefly, poly(ethylene glycol) monoethyl ethers were dissolved in toluene in the presence of stoichiometric amounts of thionyl chloride, and the mixtures were stirred at room temperature for 24 h. The solvent was then evaporated under vacuum, and the residue obtained was stored in a desiccator until use. The final substituted PVAs were named according to the number of
Orienti et al.
ethylene oxide (EO) units present in the poly(ethylene glycol) monoethyl ethers linked to the polymeric backbone: P10(2) when the substituent was ethylene glycol monoethyl ether (2EO), P10(3) diethylene glycol monoethyl ether (3EO), P10(4) triethylene glycol monoethy lether (4EO), P10(5) tetraethylene glycol monoethyl ether (5EO), and P10(8) heptaethylene glycol monoethyl ether (8EO). P inidicated the PVA backbone, and 10 its molecular weight in kilodaltons. In addition, PVA substituted with the oleyl chains without PEGMEE chains was also prepared and named P10(0). The degrees of substitution of the final products were determined by elemental analysis performed using a PerkinElmer elemental analyzer (model 240 B) and by 1H NMR using a Gemini 600 spectrometer and recording the spectrum in (CD3)2SO. Solubilization Studies. Solubilization studies were carried out by dispersing in water amounts of polymers varying from 0.05 to 10 mg/mL. The solutions obtained were observed microscopically to evaluate the appearance of solid, undissolved material indicative of the achievement of saturation with respect to the polymer. Dynamic Light Scattering (DLS) Measurements. DLS measurements were performed on the polymeric aqueous solutions at different concentrations and temperatures with the aim to evaluate the mean size of the polymeric aggregates and their sensitivity to temperature and concentration changes. Measurements were performed by a Brookhaven 90-PLUS instrument equipped with a 50 mW He-Ne laser (532 nm). The scattering angle was fixed at 90°. Results were the combination of three 10-min runs for a total accumulation correlation function (ACF) time of 30 min. The mean size of the polymeric aggregates in solution was provided by their average hydrodynamic radius; results were volume weighted. The concentrations analyzed ranged from 0.05 to 1 mg/mL, as at higher concentrations, a rapid increase in the scattered intensity indicated the establishment of multiple scattering.19,20 The temperature of the polymer solutions ranged from 10 to 37 °C. Static Light Scattering (SLS) Measurements. SLS measurements were performed over the same concentration and temperature ranges used for the DLS measurements on a Brookhaven BI-200SM to obtain the molecular weight of aggregates by the Zimm plot.21,22 Differential Scanning Calorimetry (DSC). DSC measurements on the polymer aggregates were performed using a VP-DSC apparatus (Microcal Inc., North Hampton, MA). Prior to analysis, the polymers were dissolved in water at a concentration of 5 mg/mL. Water was used for baseline scans. The solutions were allowed to equilibrate at 20 °C for 15 min, and temperatures from 20 to 50 °C were scanned at a rate of 1.5 °C/min. DSC curves were analyzed with MicroCal Origin software. Affinity for Hydrophobic Phases. Affinity studies toward hydrophobic phases were carried out to simulate the polymer’s ability to interact with cell membranes in aqueous environment. The polymer aqueous solutions (1 mg/mL) were introduced in glass tubes containing at the bottom a polyethylene membrane (Celgard 2700, Hoechst) fixed by a stopping glass-rubbery disk. The surface of the membrane
PVA Derivatives as Cytotoxic Systems
in contact with the aqueous phase was 1.54 cm2, and the volume of the aqueous phase in the glass tube was 2 mL. The solutions were allowed to equilibrate with the membrane for 6 h at 37 °C and then analyzed by HPLC to evaluate the decrease in polymer concentration due to its interaction with the polyethylene membrane. The solution-membrane equilibration was fully established after 6 h, as at longer time periods, no further variation in the polymer concentration had been detected in all the systems analyzed. The decrease of the polymer concentration in solution was considered representative of the polymer affinity for hydrophobic phases. The polyethylene membrane was selected as a solid hydrophobic phase to avoid the use of liquids which could emulsify into the polymeric aggregates, thus hampering any determination of the polymer concentration decrease in solution. High-performance liquid chromatography (HPLC) assays were carried out by a size exclusion silica diol column (Chromegapore MSE Diol 5 µm, 1000 A, Superchrom, Milan, Italy) using water at 0.2 mL/min as the mobile phase and an IR detector (Shimadzu RID-10A). The system was thermostated at 37 °C. Biological Studies. Antitumor activity of the polymers has been evaluated on the following neuroblastoma cell lines, SH-SY5Y, IMR-32, and HTLA-230, and melanoma cell lines, MZ2-MEL and RPMI7932. The polymers active toward tumor cells have also been tested on normal resting lymphocytes. Cell lines were maintained in Dulbecco’s Modified Eagle Medium (DMEM) growth medium containing 10% fetal bovine serum and 100 ng/mL each penicillin and streptomycin (all from Sigma) at 37 °C in 5% CO2. Experiments were performed during the logarithmic phase of cell growth. Cells were seeded in six-welled plates (Corning Incorporated, NY) (105 cells/well) as triplicates. After 48 h, cells were treated with growth medium containing varying amounts of the different polymers. The effects of the polymers on cell growth and death were determined by cell count and the trypan blue exclusion method. Cells maintained in medium alone and treated with PVA were used as controls. Statistical Analysis. All values in the figures and text are expressed as mean plus/minus standard deviation (SD) of N experiments (with N g 3). Statistical data analysis was performed using the student t-test. Data sets were examined by analysis of variance (ANOVA). P values less than 0.05 were considered statistically significant. Results and Discussion Characterization of the Modified Polymers. Elemental analysis revealed that the substitution degree (moles of substituent per 100 mol of hydroxyvinyl monomer) in the final polymer corresponded to 1.48% for the oleyl chains and 0.99% for PEGMEE chains. By 1H NMR analysis, the substitution degree of oleyl chains was obtained by comparing the integral of the peak at 5.32 δ assigned to the protons (CHdCH) of the oleyl chain with the integral of the peak at 1.95δ assigned to the protons (COCH3) of the acetyl moiety present at 20% in the PVA backbone. The substitution degree of the PEGMEE chain
Biomacromolecules, Vol. 6, No. 5, 2005 2877
was obtained by comparing the integral of the peak at 1.17 δ assigned to the protons (CH3) of PEGMEE with the integral of the peak at 1.95 δ assigned to the protons (COCH3) of the acetyl moiety. The substitution degrees obtained by 1H NMR analysis resulted at 1.50% for the oleyl chains and 1.02% for the PEGMEE chains (moles of substituent per 100 mol of hydroxyvinyl monomer). Solubilization Studies. The polymers analyzed were characterized by high water solubility, as no saturation was observed up to 5 mg/mL. Beyond this concentration, all the solutions, except those containing P10(4), displayed solid formations indicative of saturation. The P10(4) solution did not display solid formations up to the maximum concentration analyzed (10 mg/mL). DLS and SLS Measurements. DLS measurements of polymer aqueous solutions revealed the presence of aggregates characterized by low polydispersity (minimum 0.232; maximum 0.299) indicating that polymers aggregate in nearly monodisperse systems. The mean size of aggregates slightly decreased with increasing polymer concentration in solution and was not significantly affected by temperature variations (Table 1). At each concentration and temperature analyzed, the mean size of aggregates increased in the sequence P10(4) < P10(3) < P10(2) < P10(0) < P10(5) < P10(8). SLS measurements at different polymer concentrations provided, by the Zimm plot, the molecular weight of the polymeric aggregates which allowed us to obtain, from the molecular weight of the polymers, the aggregation number, i.e., the number of polymer molecules in each aggregate.21,22 The same measurements at different temperatures provided the aggregation number variation with temperature. The polymer aggregation number increased with temperature increase and followed the sequence P10(8) < P10(5) < P10(0) < P10(2) < P10(3)
