Self-Associating Poly(ethylene oxide)-b-poly(Î±-cholesteryl carboxylate

Biomacromolecules 2009, 10, 471–478

471

Articles Self-Associating Poly(ethylene oxide)-b-poly(r-cholesteryl carboxylate-ε-caprolactone) Block Copolymer for the Solubilization of STAT-3 Inhibitor Cucurbitacin I Abdullah Mahmud,† Sarthak Patel,‡ Ommoleila Molavi,† Phillip Choi,‡ John Samuel,† and Afsaneh Lavasanifar*,† Faculty of Pharmacy and Pharmaceutical Sciences and Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta, Canada Received July 29, 2008; Revised Manuscript Received December 10, 2008

An increase in the degree of chemical compatibility between drug and polymeric structure in the core has been shown to raise the encapsulation efficiency and lower the rate of drug release from polymeric micelles. In this study, to achieve an optimized polymeric micellar delivery system for the solubilization and controlled delivery of cucurbitacin I (CuI), the Flory-Huggins interaction parameter (χsc) between CuI and poly(ε-caprolactone) (PCL), poly(R-benzylcarboxylate-ε-caprolactone) (PBCL) and poly(R-cholesteryl carboxylate-ε-caprolactone) (PChCL) structures was calculated by group contribution method (GCM) as an indication for the degree of chemical compatibility between different micellar core structures and CuI. The results pointed to a better compatibility between CuI and PChCL core rationalizing the synthesis of self-associating methoxy poly(ethylene oxide)-bpoly(R-cholesteryl carboxylate-ε-caprolactone) block copolymer (MePEO-b-PChCL). Novel block copolymer of MePEO-b-PChCL was synthesized through, first, preparation of substituted monomer, that is, R-cholesteryl carboxylate-ε-caprolactone, and further ring opening polymerization of this monomer by methoxy PEO (5000 g mol-1) using stannous octoate as catalyst. Synthesized block copolymers were characterized for their molecular weight and polydispersity by 1H NMR and gel permeation chromatography. Self-assembled MePEO-b-PChCL micelles were characterized for their size, morphology, critical micellar concentration (CMC), capacity for the physical encapsulation of CuI, and mode of CuI release in comparison to MePEO-b-PCL and MePEO-b-PBCL micelles. Overall, the experimental order for the level of CuI encapsulation in different polymeric micellar formulations was consistent with what was predicted by the Flory-Huggins interaction parameter. Although MePEO-b-PChCL micelles exhibited the highest level of CuI loading, this structure did not show any significant superiority over MePEO-b-PCL in controlling CuI release. The most efficient control over the rate of CuI release was achieved by MePEO-b-PBCL micelles that had more viscous cores than that of MePEO-b-PChCL, instead. The results point to a potential for MePEO-b-PChCL micelles for the solubilization of cholesterol compatible drugs. It also highlights the inadequacy of the Flory-Huggins interaction parameter calculated by GCM in predicting the order of drug release from different polymeric micellar structures.

Introduction Compatibility between the solubilizate and micellar core, commonly characterized by low values of Flory-Huggins interaction parameter (χsc), is known to be one of the key factors in determining the effectiveness of polymeric micellar delivery systems in drug solubilization and controlled release. A polymer is a good solubilizing agent for a drug when there are favorable interactions between polymer/drug pairs. On the other hand, a strong interaction between micellar core and drug is expected to slow down the rate of drug release from polymeric micelles. Predictive molecular thermodynamic theories for the solubilization in polymeric micelles has been formulated and experimentally tested for the solubilization of hydrophobic molecules in tri- and di block copolymer micelles by several groups.1-3 * To whom correspondence should be addressed. Phone: 780-492-2742. Fax: 780-492-1217. E-mail: [email protected]. † Faculty of Pharmacy and Pharmaceutical Sciences. ‡ Department of Chemical and Materials Engineering.

In polymeric micelles, enhanced degree of compatibility between drug and micellar core can be achieved through hydrophobic interaction, ionic complex formation and hydrogen bonding between polymer and drugs.4-8 In the case of poly(ethylene oxide)-block-poly(L-amino acid) (PEO-b-PLAA) based micelles, a number of drug compatible groups such as benzyl, doxorubicin (DOX), carboxyl, or fatty acid esters have been attached to the free functional groups of PLAA core to facilitate the interaction between drug and micellar core. This way, improvements in the stability, drug loading and release properties have been achieved for the delivery of DOX,6,9 amphotericin B,4 cisplatin,8 and KRN-5500.10 However, in the case of PEO-b-poly(ester) based micelles, chemical modifications in the micellar core structure have been restricted until recently due to the lack of proper functional groups on the poly(ester) backbone. Our research group has recently reported on the development of a new family of methoxy poly(ethylene oxide)-b-poly(ε-caprolactone) (MePEO-b-PCL) block copolymers bearing pendent carboxyl and benzylcarboxylate groups on the poly-

10.1021/bm800846a CCC: $40.75  2009 American Chemical Society Published on Web 01/28/2009

472

Biomacromolecules, Vol. 10, No. 3, 2009

Scheme 1. Synthetic Scheme for the Preparation of R-Cholesteryl carboxylate ε-Caprolactone

Mahmud et al. Scheme 2. Synthetic Scheme for the Preparation of MePEO-b-PChCL Block Copolymers

Figure 1. Chemical structure of cucurbitacin I.

(ester) block.11 In further studies, the pendent carboxyl group has been used to chemically attach DOX to the core of MePEOb-PCL micelles providing polymeric micellar drug conjugates with hydrolyzable core structures.12 The synthesis of core functionalized MePEO-b-PCL block copolymers was made possible through functionalization of the ε-caprolactone monomer. The first objective of this study was to assess the feasibility of this synthetic strategy for the substitution of different moieties on the PCL block. In this paper, the attachment of pendent cholesteryl group to the PCL backbone was accomplished using a similar synthetic method (Schemes 1 and 2). Polymeric nanocarriers with cholesteryl modified core structures, for example, micelles from MePEO-b-poly(R-cholesteryl carboxylate ε-caprolactone) (MePEO-b-PChCL), are expected to increase the solubilization and improve the release profile of several drugs that are chemically compatible and/or strongly interact with cholesterol such as cucurbitacin I (CuI; Figure 1). Cucurbitacins are of great interest due to their selective inhibitory activity on signal transducer and activator of transcription 3 (STAT3) pathway and strong antiproliferative function against a number of human carcinoma cell lines.13-15 The IC50 values of cucurbitacins against several cancer cells are comparable with DOX, a widely used anticancer drug.14 Moreover, selective STAT3 inhibitory activity of cucurbitacins makes them excellent drug candidates for delivery to tumor microenvironment to overcome tumor-induced immunosuppression, which may eventually lead to potent antitumor immune responses through inhibition of STAT3.16 Despite great efficacy, clinical development of this drug and its analogues has mostly been limited by their poor water solubility and nonspecific toxicity. A polymeric nanocarrier that can encapsulate cucur-

bitacins efficiently, control the premature rate of drug release in systemic circulation, restrict their distribution to nontarget sites in the body while directing the drug to tumor site may overcome CuI limitations.

Materials and Methods Materials. MePEO (average molecular weight of 5000 g mol-1), diisopropylamine (99%), chlolesteryl chloroformate (tech. 95%), sodium (in kerosin), benzophenone, butyl lithium (Bu-Li) in hexane (2.5 M solution), and pyrene were purchased from Sigma, St. Louis, MO. MePEO was dried in vacuum oven at 50 °C temperature for 48 h prior to use. Diisopropylamine was dried over calcium hydride at room temperature and freshly distilled before use. ε-Caprolactone was purchased from Lancaster Synthesis, U.K., dried over calcium hydride for 48 h at room temperature, and freshly distilled before polymerization. Tetrahydrofuran (THF) was refluxed over sodium and benzophenone for several hours and distilled immediately before use. Stannous octoate was purchased from MP Biomedicals Inc., Germany, and used without further purification. Fluorescent probe 1,3-(1,1′dipyrenyl)propane was purchased from Molecular Probes, U.S.A. All other chemicals were reagent grade and were used as received. Methods. Synthesis of R-Cholesteryl carboxylate ε-Caprolactone Monomer. Cholesteryl bearing monomer, that is, R-cholesteryl carboxylate ε-caprolactone, was synthesized according to the method reported by our group in a previous publication (Scheme 1).11 Briefly, Bu-Li (24 mL) in hexane was slowly added to dry diisopropylamine (8.4 mL) in 60 mL of dry THF in a three-neck round-bottomed flask at -30 °C under vigorous stirring with continuous argon supply for the in situ preparation of nonnucleophilic base lithium diisopropylamine (LDA). The solution was cooled to -78 °C. ε-Caprolactone (3.42 g), dissolved in dry THF, was added to the above-mentioned mixture slowly and stirred for 30 min. Finally, R-cholesteryl chloroformate (13.47 g),

Solubilization of STAT-3 Inhibitor Cucurbitacin I


473

Table 1. Characteristics of Synthesized Block Copolymers block copolymera MePEO114-b-PCL42 MePEO114-b-PBCL19 MePEO114-b-PChCL14

[M]/[I]b 44 21 18

theoretical mol. wt. (g mol-1)

Mn (g mol-1)c

Mn (g mol-1)d

f

10000 10200 14500

f

9800 9700f 12400

11500 9200f 11800

PDIe 1.04f 1.74f 1.53

a The number showed as subscript indicates the polymerization degree of each block determined from 1H NMR spectroscopy. b Monomer/initiator molar ratio. c Number average molecular weight of block copolymer measured by 1H NMR. d Number average molecular weight of block copolymer measured by GPC. e Polydispersity index (Mw/Mn) measured by GPC. f Data are reproduced from our previous publication for comparison.11

Table 2. Characteristics of Empty Block Copolymer Nanocarriers (n ) 3)a block copolymerb MePEO114-b-PCL42 MePEO114-b-PBCL16 MePEO114-b-PChCL14

avg micellar size ( SD (nm)c 45.0 ( 2.0 65.5 ( 3.6h 93.0 ( 7.5

size of the secondary peaks (nm)

220 (∼40%)g

PDId

CMCe ( SD (µM)

Ie/Im ( SDf

0.20 0.31 0.58

-2

0.055 ( 0.007I 0.028 ( 0.002I,h 0.23 ( 0.030

18.2 × 10 ( 0.01 9.8 × 10-2 ( 0.01I,h 7.5 × 10-2 ( 0.01 I

a

Values are the average of three different measurements. b The number showed as subscript indicates the polymerization degree of each block. Intensity mean estimated by DLS. d Polydispersity index (PDI) of size distribution estimated by DLS technique. e Measured from the onset of a rise in the intensity ratio of peaks at 339 nm to peaks at 334 nm in the fluorescence excitation spectra of pyrene plotted versus logarithm of polymer concentration. f Intensity ratio (excimer/monomer) from emission spectrum of 1,3-(1,1′ dipyrenyl) propane in presence of polymeric micelles. g Numbers in the parenthesis indicate the frequency of secondary peak in micellar population in percentage. h Significantly different from MePEO-b-PCL and MePEO-b-PChCL (P < 0.05, one way ANOVA). I The data are reproduced from our previous publication for comparison.11 c

dissolved in dry THF, was added to the reaction mixture slowly and stirred for additional 1 h. The temperature was allowed to rise to 0 °C, and the reaction was quenched with 5 mL of saturated ammonium chloride solution. The reaction mixture was diluted with water and extracted with ethyl acetate. The combined extracts were dried over Na2SO4 and evaporated. The yellowish semisolid crude mixture was purified twice over a silica gel column using hexane:ethyl acetate at 3:1 ratio as eluent to get the pure compound as white solid powder. The yield of the monomer synthesis was found to be 9.0 g (65%). The purity of the compound was confirmed with thin-layer chromatography (TLC). The chemical structure was analyzed by 1H NMR, IR, and mass spectroscopy. In 300 MHz 1H NMR spectroscopy in CDCl3, corresponding proton peaks were observed at δ (ppm): 0.681 (s, 3H), 0.86-1.7 (m, 34H), 1.8-2.1 (m, 10H), 2.35 (m, 2H), 3.68 (dd, 1H), 4.25 (m, 2H), 4.73 (m, 1H), 5.38 (m, 1H). Mass spectroscopy resulted in the formation of molecular ion peak (M+) at m/z 526.49; M+ + Na peak at m/z 549.15; M+ + K at m/z 565.09. Benzyl group containing monomer, that is, R-benzylcarboxylate ε-caprolactone, was also synthesized by the similar method described in previous publication.11 Synthesis of MePEO-b-PChCL Block Copolymer. Cholesteryl group bearing block copolymer, that is, MePEO-b-PChCL, was synthesized by ring-opening polymerization of R-cholesteryl carboxylate ε-caprolactone using MePEO as initiator and stannous octoate as catalyst (Scheme 2).11 Briefly, dried MePEO (MW: 5000 g mol-1; 1.5 g), R-cholesteryl carboxylate ε-caprolactone (3.0 g), and stannous octoate (0.002 equiv of monomer) were added to a 10 mL previously flamed ampule, nitrogen purged and sealed under vacuum. The polymerization reaction was allowed to proceed for 4 h at 160 °C in oven. The reaction was terminated by cooling the product to room temperature. 1 H NMR spectrum of MePEO-b-PChCL in deuterated chloroform (CDCl3) was obtained by a 300 MHz, Bruker Unity-300 NMR spectrometer at room temperature and used to assess the conversion of R-cholesteryl carboxylate ε-caprolactone monomer to PChCL by comparing the peak intensities of methylene protons (-O-CH2-, δ 4.28 ppm) of R-cholesteryl carboxylate ε-caprolactone monomer to the intensity of the same proton of PChCL (-O-CH2, δ 4.10 ppm). In 1H NMR spectroscopy in CDCl3, corresponding proton peaks were observed at δ (ppm) 0.68 (s, 3H), 0.84-1.7 (m, 34 H), 1.73-2.0 (m, 10 H), 2.30 (m, 1H), 3.28 (m, 1H), 3.65 (s, 4H), 4.63 (m, 1H), and 5.35 (m, 1H). The number of protons in the parentheses represents the corresponding number of protons in one ethylene oxide versus one cholesteryl caprolactone unimer in MePEO-b-PChCL block copolymer. Characterization of MePEO-b-PChCL Block Copolymer. The number average molecular weight (Mn) of MePEO-b-PChCL was determined from 1H NMR spectrum by comparing the peak intensity of

Table 3. Calculated Values of Solubility Parameters of Different Polymers and CuI

polymer/ drug PCL PBCL PChCL CuI

partial solubility parametersa (S.I.) δd 17.46 18.88 21.42 18.72

δp

δh

4.92 8.38 3.64 8.52 2.56 6.05 4.77 15.69

molar volume total solubility (cm3/mol) parametersb (S.I.) 19.98 21.03 22.40 24.88

99.66 192.75 382.84 349.35

a δd, δp, and δh are the partial solubility parameters for CuI and different polymers calculated by GCM as outlined in eqs 3, 4, and 5. b Total solubility parameters were calculated from partial solubility parameters as outlined in eq 2. All solubility parameters were calculated in standard international unit (S.I.; mega Pascal0.5).

MePEO (-CH2CH2O-, δ 3.65 ppm) to that of PChCL (-O-CH2-, δ 4.10 ppm), considering a 5000 g mol-1 molecular weight for MePEO. The Mn and weight average molecular weight (Mw) as well as polydispersity (Mw/Mn) of the prepared block copolymers were further assessed by gel permeation chromatography (GPC). Briefly, samples (20 µL from 10 mg/mL polymer stock solutions in THF) were injected into a 4.6 × 300 mm Waters Styragel HT4 column (Waters Inc., Milford, MA). The elution pattern was detected at 35 °C by refractive index (PD2000, Percision Detectors, Inc.)/light scattering detectors (Model 410, Waters Inc.). THF was used as eluent at a flow rate of 1.0 mL/min. The column was calibrated with a series of standard polystyrenes of varying molecular weights (Mw: from 4750 g mol-1 to 13700 g mol-1). Assembly of Block Copolymers and Characterization of Self-Assembled Structures. Micellization was achieved by dissolving prepared block copolymers (30 mg) in THF (0.5 mL) and dropwise addition (1 drop/15 s) of polymer solutions to doubly distilled water (3 mL) under moderate stirring at 25 °C, followed by the evaporation of THF under vacuum. Average diameter (intensity mean) and size distribution of prepared micelles were estimated by dynamic light scattering (DLS) using a Malvern Zetasizer 3000 at a polymer concentration of 2.5 mg/mL in water at 25 °C after centrifuging the micellar solution at 11600 × g for 5 min. Morphology of the selfassembled structures was investigated by transmission electron microscopy (TEM). An aqueous droplet of micellar solution (20 µL) with a polymer concentration of 1-1.5 mg/mL was placed on a copper coated grid. The grid was held horizontally for 20 s to allow the colloidal aggregates to settle. A drop of 2% solution of phosphotungstic acid (PTA) in PBS (pH ) 7.0) was then added to provide the negative stain. After 1 min, the excess fluid was removed by filter paper. The samples were then air-dried and loaded into a Hitachi H 700 transmission electron microscope. Images were obtained at a magnifica-

474


Mahmud et al.

tion of ×22000 at 75 KV. The diameter of individual particles (n ) 50) from micrographs was measured manually to obtain the average size of micelles in a dry state. A change in the fluorescence excitation spectra of pyrene in the presence of varied concentrations of MePEO-b-PChCL block copolymer was used to measure its critical micellar concentration (CMC).17 The viscosity of prepared micellar core was estimated by measuring excimer to monomer intensity ratio (Ie/Im) from the emission spectra of 1,3(1,1′-dipyrenyl) propane at 373 and 480 nm, respectively.18 Calculation of the Flory-Huggins Interaction Parameter (χsc) between Micellar Core and CuI. The compatibility between CuI and micellar core was calculated by the Flory-Huggins interaction parameter (χsc) as outlined in eq 1.

χsc ) (δs - δc)2VS ⁄ RT

(1)

where (δs - δc) is the difference in solubility parameters between drug (s) and micellar core (c). Vs is the molar volume of the solubilizate, R is the gas constant, and T is the Kelvin temperature. The solubility parameter (δ) was obtained by Hansen’s approach,19 which uses partial solubility parameters to calculate the total solubility parameter as outlined in eq 2.

δ ) (δd2 + δp2 + δh2)1⁄2

(2)

where, δd, δp, and δh are the partial solubility parameters indicating contributions from Van der Waals dispersion forces between atoms, electrostatic interactions between molecules, and propensity of hydrogen bonding between molecules, respectively. The partial solubility parameters for the drug (CuI) and polymers (Table 3) were calculated by group contribution method (GCM) using the following three equations:

∑F ⁄V δ ) [∑ F ] ⁄ V δ ) [∑ E ] ⁄ V δd )

di

(3)

p

2 1⁄2 pi

(4)

h

hi

1⁄2

by a nonlinear gradient to a final ratio of 60:40 (v/v) over 8 min at a constant flow rate of 0.2 mL/min. A calibration curve was constructed over the quantification range of 50-5000 ng/mL of CuI solution in methanol. The ratios of CuI to I.S. peak areas were calculated and plotted versus CuI concentration. CuI loading and encapsulation efficiency were calculated by the following equations: CuI loading (CuI ⁄ monomer;%) ) CuI loading (CuI ⁄ polymer;%) )

Encapsulation efficiency (%) )

moles of loaded CuI × 100 (6) moles of monomer

moles of loaded CuI × 100 (7) moles of copolymer

amount of loaded CuI (mg) × 100 amount of CuI added (mg) (8)

Release of CuI from Polymeric Micelles. Release study was performed using dialysis method.12 Briefly, CuI loaded micellar solutions (3 mL) were prepared at 2.5 mg/mL polymer concentration from MePEO-b-PCL, MePEO-b-PBCL, and MePEO-b-PChCL block copolymers. As a control, free CuI solution in water was prepared at a concentration of 500 µg/mL with the aid of methanol (2% v/v). The micellar solutions were transferred into a dialysis bag (MW cutoff: 3500 Da, supplied by Spectrum Laboratories, U.S.A.). The dialysis bags were placed into 500 mL of doubly distilled water in a beaker. Release study was performed at 37 °C in a Julabo SW 22 shaking water bath (Germany). At selected time intervals, 20 µL of micellar solution was withdrawn from inside the dialysis bag to measure the drug concentration by LC-MS. Three parallel measurements were performed for each time point for all formulations under current study. Statistical Analysis. Data are reported as mean ( standard deviation (S.D.). Differences among the mean of formulation characteristics for polymeric micelles were compared by either one-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls post hoc test for multiple comparisons using Sigma stat software or Student’s unpaired t test assuming unequal variance.

(5)

where Fdi, Fpi, and Ehi refer to the specific functional group contributions: Van der Waals dispersion forces (Fdi), electrostatic interactions (Fpi), and hydrogen bonding (Ehi). The total molar volume (V) of CuI and core forming blocks of different polymer repeat units were obtained by the Fedors method.20 Fdi, Fpi, and Ehi were obtained by the Hoftyzer Van Krevelen’s method.21 We divided the molecules into small chemical groups and used their Fdi, Fpi, and Ehi values to calculate the partial and total solubility parameters of CuI and different core forming blocks of MePEO-b-PCL based block copolymers. Encapsulation of CuI in Polymeric Micelles. Encapsulation of CuI in MePEO-b-PCL, MePEO-b-poly(R-benzyl carboxylate ε-caprolactone) (MePEO-b-PBCL), and MePEO-b-PChCL micelles was achieved by a cosolvent evaporation method. Briefly, 15 mg of copolymer and 1.5 mg of CuI were dissolved in 0.5 mL of THF. This solution was added to 3 mL of doubly distilled water in a dropwise manner. After 4 h of stirring at room temperature, the remaining THF was removed by applying vacuum. The aqueous solution of the micellar formulation was then centrifuged at 11600 × g at room temperature for 10 min to remove free cucurbitacin precipitates. The hydrodynamic diameter of CuI loaded micelles were measured by light scattering as described above. CuI loading level and encapsulation efficiency were determined by liquid chromatography mass spectrometry (LC-MS).22 An aliquot of the micellar solution after centrifugation (20 µL) was diluted with 980 µL of methanol to disrupt the micellar structure and release incorporated drug. Diluted micellar solution (200 µL) was added to 200 µL of internal standard (I.S.; 4-hydroxy benzophenone solution in methanol, 2 µg/ mL). This solution (10 µL) was injected to Waters, Micromass ZQ4000 LC-Mass spectrophotometer. For chromatographic separation a mobile phase consisting of a mixture of acetonitrile:water (40:60) containing 0.2% ammonium hydroxide was employed for 3 min. This was followed

Results Preparation of MePEO-b-PCL Block Copolymers with Cholesteryl Side Groups on the PCL block. Attachment of cholesteryl side groups to MePEO-b-PCL block copolymer was carried out through conjugation of chlolesteryl chloroformate with ε-caprolactone monomer producing R-cholesteryl carboxylate ε-caprolactone (Scheme 1) and further ring-opening polymerization of cholesteryl-substituted monomer (Scheme 2). The structure of monomer, R-cholesteryl carboxylate ε-caprolactone was confirmed by combined analysis of 1H NMR, IR, and mass spectroscopy. In the 1H NMR spectrum, the peak at δ 3.68 ppm for R-cholesteryl carboxylate ε-caprolactone, which corresponds to a single proton instead of two protons of ε-caprolactone monomer, indicates the successful substitution of the cholesteryl carboxylate on ε-caprolactone monomer at the R-position (Figure 2). The presence of two sharp carbonyl peaks in the IR spectrum at 1725 and 1700 cm-1 corresponds to the carbonyl groups in lactone and cholesteryl carboxylate, respectively (data not shown). Finally, the molecular ion peak obtained from mass spectrum is in agreement with the calculated molecular weight of the synthesized compound (526.76 g mol-1). The composition of MePEO-b-PChCL block copolymer was confirmed from 1H NMR spectrum in CDCl3 (Figure 3). The percent of conversion of MePEO-b-PChCL block copolymer from R-cholesteryl carboxylate ε-caprolactone was found to be 78%. The presence of characteristic peaks for cholesteryl moiety at δ 5.35 (-CH)C< proton), δ 4.63 (>CH-O- proton) and δ 0.84-1.7 ppm (other cholesteryl protons) confirm the presence


Figure 2. 1H NMR of R-cholesteryl carboxylate ε-caprolactone (substituted monomer) in CDCl3 and peak assignments with corresponding number of protons.

Figure 3. 1H NMR spectrum and peak assignment of MePEO-bPChCL block copolymer.

of pendant cholesteryl groups in the structure of the block copolymer. Furthermore, the characteristic downfield shift of -OCH2- protons (from δ 4.25 to 4.10 ppm) and O)C-CHproton (from δ 3.75 to 3.28 ppm) of caprolactone backbone in the 1H NMR spectra (Figures 2 and 3) strongly indicates the ring opening polymerization of the monomer and formation of block copolymers. The calculated molecular weight of PChCL block, determined by comparing the peak intensity of MePEO (-CH2-CH2-) at δ 3.65 ppm to that of PChCL (-CH2-O-) at δ 4.10 ppm (Figure 3) was found to be 7400 g mol-1. Therefore, total molecular weight of the synthesized MePEO-b-PChCL block copolymer obtained from 1H NMR (i.e., 12400 g mol-1) was close to the molecular weight determined by GPC (Mn ) 11800 g mol-1) (Table 1). However, the calculated molecular weight from 1H NMR indicates lower degree of polymerization (DP ) 14) of R-cholesteryl carboxyl ε-caprolactone compared to the theoretical molecular weight according to the feed ratio (M/I ) 18; Table 1), which is consistent with the previously reported result for MePEO-b-PBCL11 and may indicate the lower reactivity of substituted monomer compared to the original ε-caprolactone. The resulting copolymer showed a broad polydispersity (Mw/ Mn ) 1.53) compared to the unfunctionalized MePEO-b-PCL


475

Figure 4. TEM picture of micelles prepared from MePEO-b-PChCL block copolymer (magnification 22000). The bar on the images represents 500 nm.

block copolymer (Mw/Mn ) 1.04), which may be due to the presence of trace amount of PChCL homopolymer. Assembly of MePEO-b-PChCL Block Copolymers and Characterization of Self-Assembled Structures. The average diameter for MePEO-b-PChCL micelles determined by the DLS technique was 93.0 ( 7.5 nm. A second peak averaging at 220 ( 15 was also observed which was associated with the secondary aggregation of MePEO-b-PChCL micelles. On the other hand, micelles formed from MePEO-b-PCL and MePEOb-PBCL were found to be much smaller showing average diameters of 45.0 and 65.5 nm, respectively (Table 2). Secondary aggregates were not seen for MePEO-b-PCL and MePEOb-PBCL micelles. Intermicellar aggregation in amphiphilic block copolymer systems has been reported for other systems, as well.23,24 Despite the presence of secondary aggregation the micellar solution was still clear. The size of polymeric micelles is controlled by various factors, among which the length and nature of the core and corona-forming chains are predominant.25,26 The TEM picture of MePEO-b-PChCL micelles shows the formation of true spherical carriers having a clear boundary, and the average diameter was 59 ( 5.4 nm (Figure 4), which is much smaller than the size obtained from DLS measurement. A tendency for micellar aggregation is also evident from the TEM results (Figure 4). The CMC of synthesized block copolymer was found to decrease upon attachment of cholesteryl group. Indeed, the CMC of MePEO-b-PChCL copolymer with a degree of polymerization (DP) of 14 in hydrophobic block was 7.5 × 10-2 µM, which was 2.5 and 1.3 times lower than that of MePEO-b-PCL (ε-caprolactone DP)42) and MePEO-b-PBCL (R-benzylcaroboxylate-ε-caprolactone DP)19) micelles, respectively (P < 0.05, one way ANOVA) (Table 2). The lower CMC value for MePEO-b-PChCL clearly shows that the introduction of more hydrophobic cholesteryl carboxylate makes the self-association of block copolymers thermodynamically more favorable. The synthesized MePEO-b-PChCL block copolymer micelles possessed viscous cores, as evidenced by the low Ie/Im ratios (0.23) compared to that of surfactant micelles.18,27 However, the Ie/Im value of synthesized block copolymer was higher than the Ie/Im values of MePEO-b-PBCL and MePEO-b-PCL block copolymer, which indicates lower viscosity of cholesteryl containing core compared to benzyl containing (PBCL) or unfunctionalized PCL core.

476


Mahmud et al.

Table 4. Characteristics of CuI-Loaded Block Copolymer Nanocarriers (n ) 3)a CuI loading content (%) ( SD block copolymer nanocarriersb

CuI/monomer (molar)

CuI/polymer (molar)

χsc (polymer/CuI)c

encap. efficiency (%) ( SD

avg diameterd (nm)

MePEO114-b-PCL42 MePEO114-b-PBCL19 MePEO114-b-PChCL14

3.0 ( 0.3 8.3 ( 0.2f,g 15.4 ( 0.5f

125.8 ( 5.0 154.7 ( 3.7f,g 203.3 ( 7.0f

3.39 2.08 0.86

64.7 ( 3.0 81.6 ( 1.9f 87.2 ( 3.0f

50.4 ( 5.0 64.5 ( 4.0f,g 98.0 ( 7.5f

a

size of the secondary peaks (nm)

PDIe

230 (∼40%)

0.21 0.29 0.49

b

Values are the average of three different measurements. The number showed as subscript indicates the polymerization degree of each block. Calculated Flory-Huggins interaction parameter between CuI and different polymers. d Estimated from the intensity mean by dynamic light scattering technique. e Polydispersity index (PDI) of size distribution estimated by DLS technique. f Significantly different from MePEO-b-PCL (P < 0.05, one way ANOVA). g Significantly different from MePEO-b-PChCL (P < 0.05, one way ANOVA).

c

Calculation of the Compatibility between Micellar Core and CuI. Flory-Huggins interaction parameter as outlined in eq 1 was used to predict the compatibility of CuI with different polymeric micellar cores. The interaction parameter (χsc) was calculated from the solubility parameters (δ) of CuI (s) and different micellar cores (c). The solubility parameter (δ) was calculated by Hansen’s approach from partial solubility parameters of CuI and different core forming blocks as outlined in eq 2 (Table 3). The partial solubility parameters were obtained by GCM using different interaction forces and total molar volume (V) of CuI and different core forming polymers (Table 3). The lower the positive value of χsc, the greater the compatibility between the solubilizate and the core-forming block reflecting a favorable interaction between the hydrophobic block of copolymer and the encapsulated drug. This approach has successfully been used to predict drug-polymer miscibility and solubility.2,28-30 The predicted compatibility between the drug and different core structures under current study according to the interaction parameter (χsc) were in the order of PChCL > PBCL > PCL as χsc for PCL/CuI, PBCL/CuI, and PChCL/CuI was calculated at 3.39, 2.08, and 0.86, respectively (Table 4). Encapsulation of CuI in Polymeric Micelles. The calculated loading levels based on the drug to polymer (CuI/polymer) molar percentage for PChCL, PBCL and PCL core were 203.3, 154.7, and 125.8%, respectively (Table 4). To account for differences in the polymerization degree of hydrophobic block between block copolymers under study and demonstrate the contribution of each monomer to the drug loading efficiency in the micellar core, the mole % of loaded CuI to monomer (CuI/monomer) was calculated. The molar loading ratio of drug to monomer in MePEO-b-PCL micelles was 3.0%, where the core was unmodified. Compared to PCL core, the molar loading content was increased 2.8- and 5-fold in MePEO-b-PBCL and MePEO-bPChCL nanocarriers, respectively. The size of the CuI-loaded MePEO-b-PCL, MePEO-b-PBCL micelles measured by DLS technique was 50.4 and 64.5 nm, respectively (Table 4). In contrast, the size of MePEO-b-PChCL nanocarriers was found to be 98.0 nm using the same technique. No significant differences between the average diameters of unloaded and loaded carriers were observed for all three block copolymers under study (P > 0.05, unpaired student’s t test). In Vitro Release of CuI from Different Block Copolymer Micelles. The release profile of free CuI (solution in 2% methanol) and polymeric nanocarriers is shown in Figure 5. The transfer of free CuI from methanolic solution through the dialysis bag was found to be relatively rapid (>60 and 90% transfer in 1 and 4 h, respectively). On the other hand, polymeric nanocarriers were able to reduce the rate of drug transfer from the dialysis membrane to outside medium in the in vitro release experiment (Figure 5). MePEO-b-PCL micelles exhibited a burst release of 42% in the first hour followed by 80% drug release within 8 h (Figure 5). MePEO-b-PBCL micelles, having benzyl

Figure 5. In vitro CuI release profile from polymeric nanocarriers at 37 °C, (---b---) free CuI; (-0-) CuI loaded MePEO-b-PCL; (---O---) CuI loaded MePEO-b-PChCL; (---9---) CuI loaded MePEO-b-PBCL. The average drug loading levels based on drug to polymer molar percentage for three core forming blocks were 125.8, 203.3, and 154.7%, respectively. Each point represents mean ( SD (n ) 3). *Shows significant difference in release profile from MePEO-b-PCL and MePEO-b-PChCL (P < 0.05, one way ANOVA). †Shows significant difference from MePEO-b-PCL at 1 h (P < 0.05, one way ANOVA).

group in the core, significantly reduced the burst release of CuI and resulted in a significant decrease in accumulative drug release at the further time points (P < 0.01, one way ANOVA). This system showed an initial release of 28% within 1 h followed by an accumulative release of 71% within 8 h. On the other hand, nanoassemblies of MePEO-b-PChCL, having cholesteryl group in the core, were able to reduce the burst release at initial time points (35% drug release at 1 h) compared to MePEO-b-PCL micelles (P < 0.05, one way ANOVA), but failed to show any superiority in reducing the rate of drug release at later time points (Figure 5). In fact, the release pattern of CuI from MePEO-b-PChCL nanocarriers was almost identical with MePEO-b-PCL micelles after 2 h.

Discussion The long-term objective of this study is to develop an engineered polymeric micellar delivery system for the solubilization and controlled delivery of STAT-3 inhibitor, CuI. Delivery vehicles, specifically those constructed from polymers cannot serve as a universal carrier for the solubilization and controlled delivery of all drugs. The compatibility or degree of interactions between loaded drug and polymer governed by the unique physicochemical characteristics of individual drugs affects their solubilization and release kinetics from a polymer based delivery system, including polymeric micelles.2,31 In polymeric micelles, it has been demonstrated that an increase in the degree of compatibility between the core-forming block of the copolymer and the drug to be delivered results in an improvement in both drug loading and drug retention within


the delivery system.10,32-34 In the present study, the optimization of polymeric micellar formulation has been carried out through chemical modification of the core of MePEO-b-PCL micelles with cholesteryl carboxylate groups in order to achieve high encapsulation and controlled release properties for CuI. The selection of the cholesteryl moiety was based on our calculations for Flory-Huggins interaction parameter which has predicted a higher degree of compatibility and probable interactions between CuI and PChCL in comparison to PBCL and PCL. According to the Flory-Huggins theory, the critical χ value above which a polymer and a low molecular weight compound (e.g., drug) become immiscible is 0.5. In other words, if χ for a drug/block copolymer pair is lower than 0.5, they are soluble into each other. However, as the theory was developed mainly for solvent and polymer that interact mainly through non polar dispersion forces, the use of 0.5 for our systems, which involve both Coulombic and hydrogen bonding interactions, is not suitable. Therefore, we judge whether compatibility prediction is successful or not by comparing the trend of the computed interaction parameters from the GCM with the observed experimental solubility trend. Similar approach has been used successfully to predict the solubility of different drug molecules in polymeric micellar system in previous studies.2,35,36 Ring opening polymerization of cholesteryl bearing monomer, R-cholesteryl carboxylate ε-caprolactone was used to prepare MePEO-b-PChCL block copolymers (Scheme 1). Unlike MePEO-b-PCL and MePEO-b-PBCL, a reaction temperature of 140 °C resulted in incomplete polymerization of R-cholesteryl carboxylate ε-caprolactone. However, sufficient conversion (78%) of R-cholesteryl carboxylate ε-caprolactone to PChCL occurred at a reaction temperature of 160 °C and reaction time of 4 h. Potential degradation of polymer has been reported at higher temperature and longer reaction period due to transesterification or backbiting side reaction.37,38 The reactivity of R-cholesteryl carboxylate ε-caprolactone was found to be lower than benzyl group bearing monomer, R-benzylcarboxylate ε-caprolactone where 91% of the of R-benzyl carboxylate ε-caprolactone was converted to the polymer during the ring opening polymerization.11 Overall, attachment of bulky groups appears to decrease the ring opening polymerization capacity of lactones and this effect seems to be stronger for the larger substituent. Self-assembly of MePEO-b-PChCL led to the formation of micelles with an average diameter of 93 nm at hydrated state. However, a significant tendency for the formation of secondary aggregates was noted for this structure. A decrease in the CMC of MePEO-b-PChCL block copolymer compared to MePEOb-PCL and MePEO-b-PBCL indicated higher thermodynamic stability for the cholesterol containing block copolymer, although the core forming block of MePEO-b-PChCL had a shorter hydrophobic chain length than that of MePEO-b-PCL and MePEO-b-PBCL under study. High thermodynamic stability of polymeric micelles is an important factor determining the in vivo stability, pharmacokinetics, and chance of tumor accumulation for polymeric micellar delivery systems.39,40 Although if MePEO-b-PChCL micelles show a tendency for aggregation in biological fluids, as they did in water, it may limit their success in acting as long circulating carriers in biological systems. On the other hand, measurement of the rigidity of the micellar core may provide means for predicting the relative kinetic stability of polymeric micellar delivery systems after extreme dilution in blood circulation. Unlike pendent carboxyl and benzyl carboxylate groups which have shown to increase the viscosity of the MePEO-b-PCL micelles,18,41 the presence


477

of cholesteryl substituent on the PCL backbone reduced the viscosity of the micellar core. This observation may be attributed to the larger size of cholesteryl moieties in MePEO-b-PChCL which may induce steric hindrance restricting the intra micellar interaction of PChCL chains leading to lower core viscosities. Among the different core forming blocks, PChCL showed better efficiency for the solubilization of CuI (Table 4) confirming the prediction made by the Flory-Huggins interaction parameter on the superiority of PChCL core for the solubilization of CuI. The in vitro release studies, however, clearly demonstrated a better control over the rate of CuI release by MePEOb-PBCL micelles (Figure 5). Surprisingly, CuI release from MePEO-b-PChCL, was similar to that from MePEO-b-PCL micelles. The lower core viscosity and lower degree of polymerization of the PChCL core are assumed to have contributed to this observation. In this study, the prediction of compatibility between CuI and different polymers was assessed by calculating Flory-Huggins interaction parameter based on the GCM which provides reasonable compatibility predictions for compounds with small chemical structures since this method is solely based on the theory that the total cohesive energy density (CED) for a molecule is the linear sum of the contributions from individual functional groups within the molecules. However, the FloryHuggins interaction parameter calculated by the GCM has failed to predict the solubility of complex drug molecules and copolymers in a few instances,21,35 mostly because it does not take into account the three-dimensional arrangements and conformation of different groups in the structure of drug and polymer pairs. Therefore, the shortcoming of this parameter as a sole predictor of drug release from polymeric micelles, which is a far more complex phenomenon specially for polymeric micelles and may be affected by several factors (such as micellar core viscosity and size) is not surprising. Finally, the argument made in the manuscript on the comparative solubilization and release profile of CuI from different polymeric micelles in the current study, is based on the assumption that the majority of CuI is located in the core of polymeric micelles. Nonetheless, the possibility for the incorporation of drug in the core/shell interface and the effect of interface thickness, which itself is dependent on the interfacial tension42 cannot be ruled out at this point and needs more investigations.

Conclusions Consistent with predictions made by the Flory-Huggins interaction parameters calculated from GCM for polymer/CuI pairs, chemical tailoring of the MePEO-b-PCL micellar core through substitution of cholesteryl moieties enhanced the solubilization of poorly water soluble and cholesterol-compatible drug CuI in polymeric micelles. The in vitro rate of CuI release from polymeric nanocarriers was not affected by cholesteryl substitution, however. The most control over the rate of CuI release was achieved by polymeric micelles bearing benzyl groups in their core structure. The results also highlight the need for the development of more appropriate models that can predict the release of encapsulated drugs from polymeric nanocarriers more accurately. Acknowledgment. This study was supported by the National Science and Engineering Research Council (NSERC) Grant Number G121210926. A.M. and O.M. were supported by Rx and D HRF/CIHR graduate student research scholarship.

References and Notes (1) Nagarajan, R. Polym. AdV. Technol. 2001, 12, 23–43. (2) Liu, J.; Xiao, Y.; Allen, C. J. Pharm. Sci. 2004, 93, 132–143.

478


(3) Patel, S.; Lavasanifar, A.; Choi, P. Biomacromolecules 2008, 9, 3014– 3023. (4) Lavasanifar, A.; Samuel, J.; Sattari, S.; Kwon, G. S. Pharm. Res. 2002, 19, 418–422. (5) Nakanishi, T.; Fukushima, S.; Okamoto, K.; Suzuki, M.; Matsumura, Y.; Yokoyama, M.; Okano, T.; Sakurai, Y.; Kataoka, K. J. Controlled Release 2001, 74, 295–302. (6) Yokoyama, M.; Fukushima, S.; Uehara, R.; Okamoto, K.; Kataoka, K.; Sakurai, Y.; Okano, T. J. Controlled Release 1998, 50, 79–92. (7) Li, Y.; Kwon, G. S. Pharm. Res. 2000, 17, 607–611. (8) Nishiyama, N.; Kato, Y.; Sugiyama, Y.; Kataoka, K. Pharm. Res. 2001, 18, 1035–1041. (9) Kataoka, K.; Matsumoto, T.; Yokoyama, M.; Okano, T.; Sakurai, Y.; Fukushima, S.; Okamoto, K.; Kwon, G. S. J. Controlled Release 2000, 64, 143–153. (10) Yokoyama, M.; Satoh, A.; Sakurai, Y.; Okano, T.; Matsumura, Y.; Kakizoe, T.; Kataoka, K. J. Controlled Release 1998, 55, 219–229. (11) Mahmud, A.; Xiong, X. B.; Lavasanifar, A. Macromolecules 2006, 39, 9419–9428. (12) Mahmud, A.; Xiong, X. B.; Lavasanifar, A. Eur. J. Pharm. Biopharm. 2008, 69, 923–934. (13) Blaskovich, M. A.; Sun, J. Z.; Cantor, A.; Turkson, J.; Jove, R.; Sebti, S. M. Cancer Res. 2003, 63, 1270–1279. (14) Jayaprakasam, B.; Seeram, N. P.; Nair, M. G. Cancer Lett. 2003, 189, 11–16. (15) Sun, J. Z.; Blaskovich, M. A.; Jove, R.; Livingston, S. K.; Coppola, D.; Sebti, S. D. M. Oncogene 2005, 24, 3236–3245. (16) Yu, H.; Kortylewski, M.; Pardoll, D. Nat. ReV. Immunol. 2007, 7, 41–51. (17) Zhao, C. L.; Winnik, M. A.; Riess, G.; Croucher, M. D. Langmuir 1990, 6, 514–516. (18) Lavasanifar, A.; Samuel, J.; Kwon, G. S. Colloids Surf., B 2001, 22, 115–126. (19) Hansen, C. M. J. Paint Technol. 1967, 39, 104–117. (20) Fedors, R. F. Polym. Eng. Sci. 1974, 14, 147–154. (21) 3rd ed.; Krevelen, D. V., Krevelen, D. V., Eds.; Elsevier Scientific Pub. Co.: New York, 1990; pp 189-224. (22) Molavi, O.; Shayeganpour, A.; Somayaji, V.; Hamdy, S.; Brocks, D. R.; Lavasanifar, A.; Kwon, G. S.; Samuel, J. J. Pharm. Pharm. Sci. 2006, 9, 158–164.

Mahmud et al. (23) Chu, B. Langmuir 1995, 11, 414–421. (24) Xu, R. L.; Winnik, M. A.; Hallett, F. R.; Riess, G.; Croucher, M. D. Macromolecules 1991, 24, 87–93. (25) Jain, S.; Bates, F. S. Science 2003, 300, 460–464. (26) Zhang, L. F.; Eisenberg, A. J. Am. Chem. Soc. 1996, 118, 3168–3181. (27) Aliabadi, H. M.; Brocks, D. R.; Lavasanifar, A. Biomaterials 2005, 26, 7251–7259. (28) Forster, A.; Hempenstall, J.; Tucker; Rades, T. Drug DeV. Ind. Pharm. 2001, 27, 549–560. (29) Greenhalgh, D. J.; Williams, A. C.; Timmins, P.; York, P. J. Pharm. Sci. 1999, 88, 1182–1190. (30) Marsac, P. J.; Shamblin, S. L.; Taylor, L. S. Pharm. Res. 2006, 23, 2417–2426. (31) Allen, C.; Maysinger, D.; Eisenberg, A. Colloids Surf., B 1999, 16, 3–27. (32) Adams, M. L.; Andes, D. R.; Kwon, G. S. Biomacromolecules 2003, 4, 750–757. (33) Lee, J.; Cho, E. C.; Cho, K. J. Controlled Release 2004, 94, 323– 335. (34) Liu, J.; Zeng, F.; Allen, C. J. Controlled Release 2005, 103, 481– 497. (35) Forrest, M. L.; Zhao, A.; Won, C. Y.; Malick, A. W.; Kwon, G. S. J. Controlled Release 2006, 116, 139–149. (36) Nagarajan, R.; Barry, M.; Ruckenstein, E. Langmuir 1986, 2, 210– 215. (37) Beletsi, A.; Leontiadis, L.; Klepetsanis, P.; Ithakissios, D. S.; Avgoustakis, K. Int. J. Pharm. 1999, 182, 187–197. (38) Liggins, R. T.; Burt, H. M. AdV. Drug DeliVery ReV. 2002, 54, 191– 202. (39) Kwon, G. S. Crit. ReV. Ther. Drug 2003, 20, 357–403. (40) Nishiyama, N.; Kataoka, K. AdV. Polym. Sci. 2006, 193, 67–101. (41) Kwon, G.; Naito, M.; Yokoyama, M.; Okano, T.; Sakurai, Y.; Kataoka, K. Langmuir 1993, 9, 945–949. (42) Linse, P. In Amphiphilic block copolymers: Self-Assembly and Applications; Alexandridis, P., Lindman, B., Eds.; Elsevier: New York, 2000; pp 13-40.

BM800846A

Self-Associating Poly(ethylene oxide)-b-poly(Î±-cholesteryl carboxylate

Recommend Documents