Polymer Terminal Group Effects on Properties of Thermoresponsive

Biomacromolecules 2005, 6, 2320-2327

2320

Polymer Terminal Group Effects on Properties of Thermoresponsive Polymeric Micelles with Controlled Outer-Shell Chain Lengths Masamichi Nakayama and Teruo Okano* Institute of Advanced Biomedical Engineering and Science, Tokyo Women’s Medical University, 8-1 Kawada-cho, Shinjuku-ku, Tokyo 162-8666, Japan Received March 31, 2005; Revised Manuscript Received May 12, 2005

Well-defined amphiphilic diblock copolymers comprising thermoresponsive polymer segments of poly(Nisopropylacrylamide-co-N,N-dimethylacrylamide) (PID) and hydrophobic polymer segments, poly(benzyl methacrylate) (PBzMA), were synthesized by controlled living radical polymerization. Terminal derivatization of PID segments to either hydroxyl or phenyl groups was achieved through reactions of coupling agents with thiol groups exposed by cleavage of terminal dithiobenzoate groups. Diblock copolymers formed coreshell type polymeric micelles with thermoresponsive outer shells. Hydrodynamic micellar diameters ranged from 12 to 31 nm, controlled by varying PID chain lengths. Differences in PID terminal groups did not affect the critical micelle concentration or micellar diameters. However, these groups demonstrated a significant influence on the micellar thermoresponses. Hydroxylated PID/PBzMA micelles exhibited a phase transition of approximately 40 °C, independent of PID molecular weights. Even though molecular weights and compositions of PID chains were equivalent except for terminal groups, micelles having the outermost surface phenyl groups exhibited drastically lower phase transition temperature shifts, especially for micelles with low molecular weight PID chains. Introduction Amphiphilic block or graft copolymers form spherical multimolecular assemblies called polymeric micelles with reliable structural stability and nano-order sizes.1,2 Kataoka and Okano reported a concept of new therapeutic systems using anti-cancer drug-incorporated polymeric micelles composed of poly(ethylene glycol)-b-poly(L-aspartate) (PEG/ PBLA).3 These polymeric micelles escape from reticuloendotherium systems (RES) due to their nano-order sizes and accumulate preferentially in solid tumor tissue through the enhanced permeability and retention (EPR) effect.4,5 This system is now under clinical trials6 and is extremely attractive for targeted drug delivery applications.7,8 Recently, we and other researchers independently designed polymeric micelles possessing stimuli-responsive outer shells (corona) and/or inner aggregated cores as next generation polymeric micelles for application in effective cancer chemotherapy systems.9-11 These micelle systems can achieve ON-OFF release of incorporated drugs and cellular uptake prompted by physicochemical stimuli such as heat9 and pH variations.10,11 In our previous studies, several polymeric micelles with thermoresponsive outer shells comprising poly(N-isopropylacrylamide) (PIPAAm) or its copolymers were designed, and successful controlled release of anti-cancer drugs and subsequent expression of in vitro cytotoxicity were both demonstrated with applied temperature changes.12,13 * Corresponding author. Phone: +81-3-3353-8111, ext. 30233; fax: +81-3359-6046; e-mail: [email protected].

PIPAAm is well-known to exhibit a reversible phase transition across its lower critical solution temperature (LCST) in aqueous medium.14 This polymer is water-soluble and hydrophilic, existing in an extended conformation, below its LCST but undergoes a phase transition to insoluble, hydrophobic aggregates above 32 °C. We have previously demonstrated the preparation of PIPAAm-grafted interfaces for various applications including novel aqueous chromatography systems to separate bioactive compounds15 and thermally regulated cell adhesion and detachment controlled solely by temperature changes.16 We have also reported contributions of both molecular weight and functional end groups on the LCST of various PIPAAm derivatives.17 Our previous reports thus suggest that molecular design of thermoresponsive polymers, in terms of their molecular weights and functional end groups, is important for the production of well-defined intelligent biomaterials and biointerfaces with appropriate properties. Here, we focus on characterization of thermoresponsive polymeric micelles with varying thermoresponsive polymer chain lengths and effects of terminal functional groups, especially surface polarity and hydrophobicity, on thermoresponses of the corresponding polymeric micelles. Terminal functional groups on the micelle’s hydrophilic corona chains can be localized on the micellar outermost surface via densely packed hydrophilic polymer chains. Thus, novel functional nanoparticles can be created by controlling their thermoresponsive phase transition through the modulation of surface polarity and hydrophobicity. For this purpose, well-defined diblock copolymers of thermoresponsive poly(N-isopropyl-

10.1021/bm050232w CCC: $30.25 © 2005 American Chemical Society Published on Web 06/23/2005

Properties of Thermoresponsive Polymeric Micelles

Biomacromolecules, Vol. 6, No. 4, 2005 2321

Figure 1. (a) Synthetic scheme of PBzMA-b-PID diblock copolymers by RAFT polymerization and (b) chemical structure of diblock copolymers and formation of surface derivatized thermoresponsive polymeric micelles.

acrylamide-co-N,N-dimethylacrylamide) (PID) and hydrophobic polymer segments, poly(benzyl methacrylate) (PBzMA), possessing different thermoresponsive PID chain lengths were synthesized by reversible addition-fragmentation chain transfer (RAFT) polymerization.18,19 RAFT polymerization allows for the control of the molecular weight and its distribution during the polymerization process, applicable to a wide range of monomers under various experimental conditions. Another advantage in RAFT polymerization is that thiol-terminated polymers can be easily prepared via base cleavage of terminal polymer dithioester groups.20,21 Thus, terminal modification of PID/PBzMA diblock copolymers was performed by reaction of select coupling agents with exposed polymer thiol groups to obtain two types of surface functionalized thermoresponsive polymeric micelles possessing either phenyl (apolar) or hydroxyl (polar) groups on their corona surfaces. We investigated the effects of PID thermoresponsive block chain lengths on characteristics of the formed polymeric micelles such as their critical micelle concentration (CMC) in water and hydrodynamic diameters. Furthermore, we analyze the influence of PID thermoresponsive chain lengths and outermost surface functionalities on thermoresponsive behavior of polymeric micelles. Experimental Procedures Materials. N-Isopropylacrylamide (IPAAm) was kindly provided by Kohjin (Tokyo, Japan) and recrystallized twice from n-hexane. Benzyl methacrylate (BzMA) and N,Ndimethylacrylamide (DMAAm) were purchased from Wako

Pure Chemicals Industries Ltd. (Osaka, Japan) and distilled under reduced pressure. 2,2′-Azobisisobutyronitrile (AIBN, Wako Pure Chemicals) was recrystallized from methanol. Benzene (Wako Pure Chemicals), N,N-dimethylformamide (DMF, Wako Pure Chemicals), N,N-dimethylacetamide (DMAc, Wako Pure Chemicals), diethyl ether (Wako Pure Chemicals), 2-ethanolamine (Kanto Chemical Co. Inc., Tokyo, Japan), iodoethanol (Aldrich, Milwaukee), pyrene (Wako Pure Chemicals), and N-(1-pyrenyl)maleimide (Aldrich) were used as received. A RAFT agent, 2-cyanopropyl dithiobenzoate (CPDB), was prepared by modification of previous methods.22,23 General RAFT Polymerization Procedure. PBzMA was prepared by RAFT polymerization using CPDB as a chain transfer agent (CTA). BzMA (0.03 mol), CPDB (3.60 mmol), and AIBN (0.72 mmol) as initiators were dissolved in 60 mL of benzene. The solution was degassed under reduced pressure by three freeze-pump-thaw cycles, and polymerization was carried out at 70 °C for 7 h. After polymerization, polymers were precipitated in an excess of methanol and purified by repeated precipitations, followed by thorough drying under vacuum. The obtained BzMA homopolymers were used as a macro-CTA. Diblock copolymers with thermoresponsive segments, poly(IPAAm-co-DMAAm) (PID), were synthesized using characterized PBzMA as the macro-CTA (Figure 1a). IPAAm, DMAAm, AIBN, and PBzMA were dissolved in 5 mL of benzene, and solutions were degassed by freezepump-thaw cycles three times. Polymerization was then carried out at 70 °C for 22 h. After polymerization, polymer products were obtained by precipitation twice into an excess

2322

Biomacromolecules, Vol. 6, No. 4, 2005

Nakayama and Okano

Table 1. Characterization of PBzMA and Diblock Copolymers code

[M]/[CTA]/[I] (mM)a

yield (%)

Mnb

PDIc

monomer unitsd IPAAm/DMAAm/BzMA

PBzMA b-1 b-2 b-3 b-4

500/60/12 2000/5/1 2000/10/2 2000/20/4 2000/25/5

53 56 45 44 34

2900 38800 17300 11200 9400

1.10 1.19 1.15 1.15 1.17

-/-/16 189/116/16 72/51/16 44/26/16 33/23/16

feed

a Concentration of monomer [M], RAFT agent [CTA], and AIBN [I]. b Number-averaged molecular weight, M , estimated by 1H NMR. c Polydispersity, n PDI, determined by GPC in DMF with 10 mM LiCl. d Estimated by 1H NMR.

of diethyl ether and then dried in vacuo. IPAAm homopolymers with various chain lengths were also prepared by RAFT polymerization in the same manner as for PBzMA. Number-averaged molecular weights (Mn) of obtained polymers were estimated by a 1H NMR spectrometer (400 MHz, Varian Inc.) using methyl sulfoxide-d6 (Aldrich), and their polydispersities (PDI) were determined by gel permeation chromatography (GPC; Tosoh, SC-8020, Tokyo, Japan, calibrated with polystyrene standards, elution rate: 1.0 mL/ min) at 45 °C using DMF containing 10 mM LiCl as eluent. Elution profiles were obtained with a refractive index meter (Tosoh, RI-8022). Preparation and characterization of each polymer are summarized in Table 1. Aminolysis and Conversion of Polymer Terminal Groups. The obtained polymers (300 mg) were dissolved in 2 mL of DMF deoxidized by N2 gas bubbling for 1 h. 2-Ethanolamine (20 mol equivalents vs terminal dithiobenzoate groups) in 1 mL of DMF was then added to polymer solutions. After 5 min, 40 mol equivalents (vs terminal groups) of iodoethanol in 1 mL of DMF were added to the reaction solutions, followed by reaction at 40 °C for 15 h in a nitrogen atmosphere. Reaction solutions were then dialyzed against distilled water at 5 °C for 48 h (MWCO 1000, Spectra/Por 6, Spectrum Medical Industries), and polymers were recovered by freeze-drying. Aminolyzed diblock copolymer b-2 was then reacted with N-(1-pyrenyl)maleimide (20 mol equivalents vs terminal dithiobenzoate groups) under dark conditions to provide a terminal pyrene modification of this polymer. After this coupling reaction, the polymer was purified by precipitation in diethyl ether 5 times and then dried in vacuo. Cleavage of terminal dithiobenzoate moieties was then confirmed by the disappearance of polymer maximum absorbance at 500 nm in 1 w/v% ethanol solution using a UV-vis spectrometer (V-530, JASCO Co., Tokyo, Japan). Preparation of Polymeric Micelles. The chemical structures of PBzMA-b-PID diblock copolymers with various terminal groups and core-shell type micellar formation are shown in Figure 1b. Formation of micelles was simultaneously performed using a dialysis method.13 Amphiphilic diblock copolymers were dissolved in DMAc, and solutions were then dialyzed against distilled water at 5 °C for 24 h using a dialyzer with a Spectra/Por 6 membrane (MWCO: 1000), followed by freeze-drying. The obtained surfacederivatized PID/PBzMA micelles are coded by the components of the diblock copolymers and their PID terminal groups (Table 2). (Phe) and (OH) indicate these functionalities on the micellar surface with phenyl and hydroxyl groups, respectively. For example, M-1(Phe) represents

Table 2. Characterization of Surface-Derivatized Thermoresponsive Polymeric Micelles codea

CMC (mg/L)b

particle size (nm)c

LCST (°C)d

M-1(Phe) M-2(Phe) M-3(Phe) M-4(Phe) M-1(OH) M-2(OH) M-3(OH) M-4(OH)

1.1 3.3 4.5 5.0 1.2 3.3 4.6 4.9

30.0 ( 5.0 22.0 ( 3.1 13.5 ( 2.9 12.0 ( 2.0 30.6 ( 5.6 20.1 ( 3.5 13.5 ( 2.2 12.2 ( 2.0

38.5 34.1 22.7 14.7 39.8 39.6 40.7 39.3

a Coded by component of diblock copolymer and PID terminal groups. (Phe) and (OH) indicate respective micellar phenyl and hydroxyl groups, in the corona region. b Determined from pyrene spectra. c Determined by DLS. Data are number-averaged diameters and expressed as the mean with standard error (n ) 3). d Determined by transmittance changes in PBS.

polymeric micelles having phenyl groups on their surfaces and comprising diblock copolymer b-1. Dynamic Light Scattering Measurements. Hydrodynamic mean diameters and size distributions of polymeric micelles were determined by dynamic light scattering (DLS) using a DLS-7000 instrument (Otsuka Electronics Co., Tokyo, Japan) equipped with a He-Ne laser (633 nm) at a scattering angle of 90°. The concentration of all polymer samples was 1 mg/mL in distilled water. Measurements were performed at 20 °C except for M-4(Phe) micelles measured instead at 15 °C. Fluorescence Measurements. Fluorescence spectra were obtained using an FP-6500 spectrofluorimeter (JASCO) with pyrene as a hydrophobic fluorescent probe. Pyrene dissolved in acetone (480 µM, 5 µL) was added to 4 mL of each polymer solution at various concentrations. Sample solutions were kept at 20 °C for 2 days before measurements were conducted to allow for complete evaporation of acetone. Fluorescence excitation was carried out at 370 nm, and excitation spectra were recorded from 300 to 360 nm. The ratio of the intensities at 337 and 334 nm (I337/I334) were then analyzed to determine critical micelle concentrations (CMC). Emission spectra were recorded from 350 to 400 nm with an excitation wavelength of 340 nm. From pyrene emission spectra, the intensity ratio (I1/I3) of the first and third bands was used to investigate the polarity of the pyrene environment. The emission spectrum of pyrenyl-labeled PID/ PBzMA micelles, M-2(Py), comprising pyrene-modified diblock copolymer b-2-Py was further evaluated to investigate the interactions between terminal pyrene hydrophobic groups on PID chains and the hydrophobic micellar inner cores. The excitation and emission bandwidths were 10 and 3 nm, respectively. Measurements were performed at 20 °C except for M-4(Phe) micelles performed at 15 °C instead.



Figure 2. GPC curves of PBzMA and PBzMA-b-PID diblock copolymers with various chain lengths, determined using DMF containing 10 mM LiCl as eluent at 45 °C with a flow rate of 1 mL/min.

Determination of LCST Using Optical Transmittance Measurements. Optical transmittance of IPAAm homopolymers in water (10 mg/mL) and polymeric micelles in phosphate buffered saline (PBS) (10 mg/mL) at various temperatures was measured at 600 nm using a UV-vis spectrometer with a heating rate of 0.1 °C/min. A sample cell was thermostated with a Peltier-effect cell holder (EHC477T, JASCO). The LCST of the polymer solutions was defined as the temperature inducing a 50% decrease in optical transmittance. Results and Discussion RAFT Polymerization. RAFT polymerization is a novel controlled living radical polymerization method that proceeds through a reversible addition-fragmentation mechanism. Polymerization is conducted in the presence of a dithiocarbonyl compound, resulting in the formation of terminally functional polymers. With this approach, we synthesized PBzMA in benzene using CPDB and AIBN as the RAFT agent and the initiator, respectively. Polymerization proceeded effectively with 53% monomer conversion after 7 h at 70 °C, and unreacted PBzMA homopolymers were removed in a purification process of diblock copolymers using an excess amount of diethyl ether that dissolves low molecular weights of PBzMA. The resultant polymer had a low PDI of 1.10 as measured by GPC. The 1H NMR spectrum showed three peaks derived from the terminal dithiobenzoate group at 7.4, 7.6, and 7.9 ppm, corresponding to resonances of meta-, para-, and ortho-phenyl, respectively (data not shown). The number-averaged molecular weight, Mn, of PBzMA, determined from the number of monomer units in PBzMA estimated from the areas of ortho-phenyl and methylene peaks corresponding to BzMA side chains (4.9 ppm), was 2900. In the next step, AB-type diblock copolymers with thermoresponsive polymer segments (PBzMA-b-PID) were prepared by RAFT polymerization using PBzMA as a macro-CTA (Figure 1a). PID segments with various chain lengths were obtained by varying the molar ratios of total monomer [M], macro-CTA [S], and initiator [I] during the syntheses of the diblock copolymers. The formulation of the diblock copolymers was then confirmed by GPC. In Figure 2, GPC elution curves for diblock copolymers were compared with that of PBzMA macro-CTA: a significant increase in molecular weight was

Figure 3. 1H NMR spectra of (a) diblock copolymer b-3-Phe in DMSO-d6 and (b) PID/PBzMA micelles M-3(Phe) composed of b-3Phe in D2O at 20 °C.

observed in each diblock polymerization step. GPC elution profiles were also unimodal without any high molecular weight species indicative of uncontrolled polymerization. In the RAFT polymerization process, polymerization kinetics change significantly due to differences in the chemical structures of monomers.24 Therefore, to control monomer compositions of polymers close to the initial feed, PID segments were prepared by copolymerization of IPAAm with the same acrylamide derivative, DMAAm, in a random fashion. Monomer compositions of IPAAm, DMAAm, and BzMA in each of the diblock copolymers were then analyzed by the peak areas of methine (4.0 ppm), methyl (2.9 ppm), and methylene (4.9 ppm) protons, respectively (Figure 3a). Results suggest that DMAAm units were introduced to the thermoresponsive PIPAAm chains at approximately 39 mol % for each PID segment of the diblock copolymers and that this resulting composition could be controlled by using a monomer ratio of 35 mol % DMAAm in the initial feed. Aminolysis of Dithiobenzoate Ester and Derivatization of Polymer End Groups. For polymers synthesized by the RAFT technique, the terminal functional dithioester groups are easily removed to yield the corresponding thiol groups in the presence of nucleophiles such as hydroxide ions and primary and secondary amines.20,21 Attachment of fluorescent dyes or biomolecules including saccharides and peptides onto micellar surfaces can be successfully achieved by such substitution of terminal groups located at the hydrophilic

2324


Figure 4. Plot of log C vs intensity ratio (I337/I334) obtained by pyrene excitation spectra for diblock copolymer b-3 in aqueous solutions at 20 °C. [pyrene] ) 6.0 × 10-7 M.

chains forming the outer micellar shell. In this study, we attempted substitutions of the thermoresponsive corona chain ends with two different terminal groups to investigate their effects on the characteristics of the thermoresponsive micelles. We confirmed reliable aminolysis of the terminal dithiobenzoate groups by analyzing UV spectra of the polymer solutions. Polymers obtained by the RAFT technique were pink in color, with a n-π* absorption band around 500 nm corresponding to the dithiobenzoate moieties. After cleavage of terminal groups, dithiobenzoate absorption at 500 nm was not detected, and polymers became a white powder. Alternatively, aminolysis was confirmed by the disappearance of 1H NMR resonances near 7.4-7.9 ppm, corresponding to the terminal dithiobenzoate groups (data not shown). For polymers treated only with base, PDI increased in the GPC profiles, indicating that some fraction of aminolyzed polymers became coupled through disulfide bond formation. Therefore, in this work, we adopted a one-pot reaction for the modification of the polymer end groups by adding the coupling agent just after visual confirmation of terminal aminolysis by observing solution color changes from pink to pale yellow. Using this procedure, GPC elution curves and the PDI values were comparable before and after the coupling reaction. Thus, hydroxylation of polymer termini was possible by reacting terminal thiol groups with iodoethanol in DMF. Reaction efficiency was determined to be greater than 90% for the terminal modification by estimating the new proton resonance at 4.2 ppm attributed to -S-CH2. In addition, we successfully prepared pyrene-labeled PID/ PBzMA block copolymers by reaction with pyrenyl maleimide in DMF under dark conditions. Formation and Characterization of Polymeric Micelles. Core-shell type micellar formation through self-association of amphiphilic PID/PBzMA diblock copolymers was achieved using dialysis of the polymer solutions in DMAc against water at a temperature below the polymer LCST (e.g., at 5 °C). PBzMA-b-PIPAAm diblock copolymers with PIPAAm segments having less than 120 monomer units did not form stable micellar structures and resulted in precipitation of the diblock copolymers in water (data not shown). In this work, successful structural stabilization of thermoresponsive micelles was achieved by introduction of a highly hydrophilic comonomer, DMAAm, into PIPAAm main chains, even with

Nakayama and Okano

a significantly shorter thermoresponsive chain length (e.g., 56 monomer units) through the improved water solubilities of micellar outer shells. All obtained micelles dispersed completely in water to become highly transparent solutions. In DMSO-d6, where micellar formation is not expected, all NMR resonances attributed to IPAAm, DMAAm, and BzMA units were detected as shown in Figure 3a. However, the NMR spectrum in D2O showed a complete loss of BzMA resonances (0.6, 0.8, 4.9, and 7.3 ppm, respectively) due plausibly to suppressed molecular motion of the aggregated hydrophobic chains (Figure 3b). 1H NMR study in aqueous solution showed mainly hydrophilic PID signals but not hydrophobic PBzMA signals, strongly indicative of a stable core-shell micellar formation with a highly viscous inner core.25 Hydrophobic microenvironments within diblock copolymer aggregates and micelles in aqueous medium were then characterized by fluorescence spectroscopy using pyrene. Pyrene is a widely used hydrophobic fluorescent probe because of its high sensitivity to the local polarity of the medium.26,27 CMC values for the amphiphilic diblock copolymers were evaluated from fluorescent excitation spectra of pyrene. At low concentrations of the diblock copolymer, changes in fluorescence intensity and shift of the excitation band at 334 nm are negligible. As the polymer concentration is increased, an increase in fluorescence intensity and a red shift of the excitation band to 337 nm can be clearly detected.26 This peak shift indicates that pyrene molecules are partitioned into a less polar environment with increasing polymer concentration. Figure 4 shows the intensity ratio I337/I334 plot of the excitation spectrum against the logarithm of the polymer concentration. CMC values were defined as the intersection of the lines drawn through the points of flat regions at low concentrations and the drastically increasing regions at high concentrations. CMC values of the diblock copolymers used in this study were estimated to be in the range of 1.1-5.0 mg/L (Table 2). Such lower CMC values than those of conventional low molecular weight amphiphiles (i.e., CMC >10 mg/L) because of strong aggregation and entanglement are considered as a very important factor for the structural stability necessary for longcirculating polymeric drug carriers in the body.28 We further investigated the influence of polymer terminal functionalities (hydrophilic/hydrophobic properties) on CMC values. CMC values increased with increasing diblock copolymer molecular weight; a lower CMC value for M-1(Phe) was demonstrated as compared with M-4(Phe), even though the hydrophobic segments comprised PBzMA with the same molecular weight. These results are likely due to an increase in chain lengths of the hydrophilic polymer segments. Additionally, for diblock copolymers with the same PID chain lengths, the outermost surface derivatives (phenyl vs hydroxyl) have little effect on the CMC values. Furthermore, as the ratio between the intensities of the first and third vibrational band, I1/I3, for pyrene is sensitive to the polarity of the local environment,27 we were able to analyze the microenvironment of the hydrophobic inner cores. The values of I1/I3 for pyrene under various conditions are summarized in Table 3. At polymer concentrations below the CMC, the



Table 3. Intensity Ratio I1/I3 of Pyrene Placed into Various Environmentsa location of pyrene

I1/I3

waterb micellar inner corec PID chain terminusd

1.81 1.18 1.83

a Emission spectra were recorded over 350-400 nm with excitation wavelength at 340 nm. [pyrene] ) 6.0 × 10-7 M, detected at 25 °C. b Dissolved in water. c Localized in micellar inner core. d Hydrophilic chain end group forming micellar outer shell.

I1/I3 values for pyrene were identical to those for water (ratio ) 1.81). As the concentration of polymers increased above the CMC, the I1/I3 values showed a drastic decrease to 1.18, a polarity for pyrene consistent with that for ethanol.29 These pyrene partitioning results strongly suggest that PID/PBzMA micelles created here can successfully incorporate hydrophobic drugs into the micellar inner cores via hydrophobic partitioning interactions. To investigate the interactions between phenyl terminal groups on the hydrophilic shell-forming chains and the hydrophobic micellar inner cores, we analyzed the microenvironment of PID terminal hydrophobic groups using pyrenyl-derivatized PID/PBzMA diblock copolymers. At a polymer concentration above the CMC (0.5 g/L), the I1/I3 value for pyrene located on the hydrophilic chain end was 1.83, a value close to that of pyrene in an aqueous environment. In addition, proton signals for dithiobenzoate groups were detected despite the lack of 1H NMR resonances corresponding to hydrophobic PBzMA chains in the PID/ PBzMA micelles, meaning that hydrophobic end groups on the PID segments are mobile (Figure 3b). These results suggest that hydrophobic end groups that did not interact with or partition into micellar inner cores are exposed to the surrounding aqueous environment and that hydrophilic PID chains have an extended corona conformation. Hydrophobic moieties such as cholesteryl or long alkyl chain groups positioned at a hydrophilic polymer terminus are known to interact with fluid liposomal bilayer membranes via hydrophobic insertion interactions.30 However, in these polymeric micelles, surface polymer chains forming the outer corona are in a highly extended conformation so that free-end, mobile hydrophobic end groups are structurally separated and cannot interact with the highly aggregated and rigid micellar inner cores. From a DLS study, a series of PID/PBzMA micelles appears monodisperse with a narrow distribution and numberaveraged diameters less than 31 nm (Table 2). Hydrodynamic diameter ranges of dispersed polymeric micelles are known to be between 10 and 100 nm.2 For passive targeting using polymeric particles, nano-ordered particle sizes (5-200 nm) are a very important factor for long circulation in the blood stream, avoiding RES uptake5 and allowing for selective tumor targeting due to the EPR effect of solid tumors.4 The prepared PID/PBzMA micelles have appropriate particle sizes with highly dispersed, nonaggregating properties, indicating their utility as targeted carrier systems. Micellar diameters also increased with increasing molecular weights of the diblock copolymers, with a mean diameter of M-1(Phe) micelles (composed of diblock copolymers b-1 with the longest PID segment) determined to be 30 nm, while

Figure 5. LCST changes of semitelechelic PIPAAms with various molecular weights, determined by transmittance changes at 600 nm in deionized water at a heating rate of 0.1 °C/min and [polymer] ) 10 mg/mL.

M-4(Phe) micelles (with the shortest PID chains) had the smallest particle size of 12 nm. Therefore, micellar sizes can be regulated by varying the outer shell-forming hydrophilic polymer chain lengths. We also investigated effects of two surface derivatives on the hydrodynamic diameters; results are summarized in Table 2. Surface derivatives did not affect micelle hydrodynamic diameters made from same diblock copolymers but bearing different terminal groups. This suggests that the terminal chemistry on the hydrophilic chains does not affect their diameters: hydrophilic polymer segments that form the outer shells are in a highly extended polymer conformation. Thermoresponsive Phase Transition Behavior. The effect of semitelechelic terminal groups on the LCSTs of IPAAm homopolymers with various chain lengths was investigated in water by a turbidity method with results shown in Figure 5. PIPAAm with hydroxyl terminal groups exhibited LCSTs of nearly 32 °C, similar to PIPAAm prepared by conventional radical polymerization regardless of their molecular weights. In contrast, PIPAAm with terminal phenyl groups had LCSTs shifted toward lower temperatures as compared to the corresponding hydroxyl terminated PIPAAms. Among synthesized PIPAAms with various chain lengths, the shortest polymer containing 60 monomer units demonstrated the largest LCST shift of approximately 5 °C. In our previous reports,17 contributions of hydrophobic groups to the phase transition temperature of PIPAAm were particularly notable when such groups were located on freely mobile chain termini. The LCSTs of alkylterminated PIPAAms were shifted to lower temperatures as compared with PIPAAm.17 Hydrophobic groups promote dehydration of proximal IPAAm units and also can disrupt polymer hydration, typically shifting the LCST to lower values. Therefore, the effect of hydrophobic terminal groups is enhanced by reducing the PIPAAm’s molecular weight. We further investigated the thermoresponsive behavior of PID/PBzMA micelles, paying special attention to surface derivatives, and LCSTs are listed in Table 2. As shown in Figure 6a, all micelles having hydroxyl-terminated PID chains (hydroxyl PID/PBzMA micelles) exhibited thermoresponsive phase transitions around 40 °C, above typical

2326


Figure 6. LCST profiles for dispersions of micelles having (a) PID hydroxyl terminal groups (hydroxyl PID/PBzMA micelles) and (b) PID phenyl terminal groups (PID/PBzMA micelles) on the micellar surfaces. Micelles comprising diblock copolymers with various molecular weights of thermoresponsive PID segments: closed circle, b-1; open circle, b-2; closed diamond, b-3; and open diamond, b-4. LCST profiles were determined by transmittance changes at 600 nm in PBS at a heating rate of 0.1 °C/min and [polymer] ) 10 mg/mL.

human body temperature, within narrow temperature ranges, indicating that chain lengths of the thermoresponsive polymer segments have little effect on LCSTs of hydroxyl PID/ PBzMA micelles. LCST shifts to higher temperatures than those for IPAAm homopolymers are due to the introduction of a hydrophilic comonomer unit, DMAAm, into the PIPAAm chains. We have previously reported that thermoresponsive polymeric micelles comprising diblock copolymers of PIPAAm and poly(n-butyl methacrylate) (PBMA) showed an LCST of 32 °C, identical to that of the IPAAm homopolymer, irrespective of hydrophobic PBMA cointroduction. Retention of the PIPAAm LCST in this terpolymer was rationalized to be due to a clearly phase-separated micellar structure having a hydrophilic outer shell and hydrophobic inner core.12 Considering that hydroxylterminated PIPAAm demonstrated similar temperature responses to PIPAAm, regardless of the molecular weight differences (see Figure 5), micellar phase transitions are assumed to be determined by the temperature response of hydroxyl-terminated PID chains located at the micellar outer corona. Of interest, micelles with hydrophobic phenyl groups on their surfaces (phenyl PID/PBzMA micelles) showed significant LCST shifts to lower temperatures than hydroxyl PID/PBzMA micelles composed of diblock copolymers of the same composition (Figure 6b). In addition, the molecular weights of thermoresponsive chains have significant effects on the phase transition of nanoparticles. Micelles with the shortest thermoresponsive chains (56 monomer units) showed a dramatic LCST shift from 39.3 °C for hydroxyl PID/ PBzMA micelles to 14.7 °C for phenyl PID/PBzMA micelles. These LCST shifts were attributed to the influence of hydrophobic phenyl groups located at the PID chain ends (outermost surface of polymeric micelles), promoting the dehydration of thermoresponsive polymer segments and thus the alteration of micellar solubilities and aggregation states. These results strongly suggest that micellar temperature response can be regulated by surface hydrophobicity without any variation of the CMC or particle size. To accomplish consistent, large magnitude nanoparticle phase transition changes by modulating surface chemistry, the thickness of the thermoresponsive outer shell is a very important factor.

Nakayama and Okano

Molecular weight contributions to these effects in micelles are similar to those in phase transitions of semitelechelic PIPAAms as mentioned previously. However, magnitudes of LCST shifts for micelles are much larger than that for PIPAAm molecules with similar chain lengths. Previously, we have reported that PEG/PBLA block copolymers formed a core-shell micelle structure, and densely packed PEG polymer chains of the micelle outer shells showed particular characteristics such as a drastic decrease of interaction with serum proteins.31 In the present thermoresponsive micellar systems, terminal hydrophobic groups could accumulate to the outermost micelle surfaces through multi-assemblies of diblock copolymers. These hydrophobic groups concentrated at the thermoresponsive polymer brush surfaces could have greater contributions to reducing the LCST than PIPAAm. Solvated and independent thermoresponsive polymer chains in aqueous media were affected by only terminally attached hydrophobic groups. On the other hand, for the polymer chains forming densely packed micelle outer shells, accumulation of hydrophobic groups surrounding diblock copolymer micelles increased the overall hydrophobicity of the micelle systems. Thus, the total enhancement of the hydrophobic effect strongly influenced the micelle thermoresponses and brought about drastic micelle LCST shifts. These results indicated that phase transition behavior of PID/ PBzMA micelles was significantly influenced by the outermost surface chemistries (hydrophilic and/or hydrophobic groups) due to a particular feature of the micelle structure. Conclusions In this paper, we successfully designed thermoresponsive polymeric micelles comprising diblock copolymers with thermoresponsive segments of varying chain lengths. Diblock copolymers with controlled molecular weights and narrow polydispersities were synthesized by RAFT polymerization. All polymeric micelles were formed at a very low polymer concentration (1.1-5.0 mg/L), with nano-ordered hydrodynamic sizes regardless of differences in surface terminal chemistry between phenyl and hydroxyl groups. We clearly demonstrated a significant influence of micelle surface chemistry on micellar thermoresponse, which is dependent on PID chain lengths. Drastic LCST shifts to lower temperatures occur only with micelles having PID hydrophobic phenyl groups. The magnitude of these LCST shifts increases with decreasing molecular weight of the PID thermoresponsive chains. Furthermore, we demonstrated distinct effects of hydrophobic groups present on the surface of the polymeric micelles on their thermal phase transition behavior as compared to single, soluble thermoresponsive PIPAAm polymer chains in aqueous media. By modulating surface hydrophobicities with applied stimuli such as light, pH, and biomolecule interactions, novel drug or reactor carrier systems and/or supramolecular sensors can be designed using these stimuli-responsive nanoparticle systems. Acknowledgment. Funds from the Center of Excellence Program for the 21st Century; the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan; the Grant-in-Aid for Young Scientists (B) (16700368) from


the Japan Society for Promotion of Science (JSPS); and the Core Research for Evolution Science and Technology (CREST), Japan Science and Technology Agency are acknowledged. The authors are also grateful to Prof. D. W. Grainger (Colorado State University), Dr. A. Kikuchi (Tokyo Women’s Medical University), and Mr. J. Yang (Tokyo Women’s Medical University) for their valuable discussion and comments on this research. References and Notes (1) Tuzar, Z.; Kratochvil, P. AdV. Colloid Interface Sci. 1976, 6, 201232. (2) Kwon, G. S.; Kataoka, K. AdV. Drug DeliV. ReV. 1995, 16 (2-3), 295-309. (3) (a) Kataoka, K.; Kwon, G. S.; Yokoyama, M.; Okano, T.; Sakurai, Y. J. Controlled Release 1993, 24, 119-132. (b) Kwon, G. S.; Suwa, S.; Yokoyama, M.; Okano, T.; Sakurai, Y.; Kataoka, K. J. Controlled Release 1994, 29, 17-23. (c) Yokoyama, M.; Okano, T.; Sakurai, Y.; Fukushima, S.; Okamoto, K.; Kataoka, K. J. Drug Targeting 1999, 7, 171-186. (4) (a) Matsumura, Y.; Maeda, H. Cancer Res. 1986, 46, 6387-6392. (b) Maeda, H.; Seymour, L. W.; Miyamoto, Y. Bioconjugate Chem. 1992, 3, 351-362. (5) Ishida, O.; Maruyama, K.; Sasaki, K.; Iwatsuru, M. Int. J. Pharm. 1999, 190, 49-56. (6) Matsumura, M.; Hamaguchi, T.; Ura, T.; Muro, K.; Yamada, Y.; Shimada, Y.; Shirao, K.; Okusaka, T.; Ueno, H.; Ikeda, M.; Watanabe, N. Brit. J. Cancer 2004, 91, 1775-1781. (7) Duncan, R. Nat. ReV. Drug DiscoV. 2003, 2, 347-360. (8) Okano, T. In Biorelated Polymers and Gels; Okano, T., Ed.; Academic Press: San Diego, 1998. (9) Cammas, S.; Suzuki, K.; Sone, C.; Sakurai, Y.; Kataoka, K.; Okano, T. J. Controlled Release 1997, 48, 157-164. (10) Na, K.; Lee, E. S.; Bae, Y. H. J. Controlled Release 2003, 87, 3-13. (11) Lee, E. S.; Na, K.; Bae, Y. H. J. Controlled Release 2005, 103, 405418. (12) Chung, J. E.; Yokoyama, M.; Okano, T. J. Controlled Release 2000, 65, 93-103. (13) Kohori, F.; Sakai, K.; Aoyagi, T.; Yokoyama, M.; Yamato, M.; Sakurai, Y.; Okano, T. Colloids Surf., B 1999, 16, 195-205. (14) Heskins, M.; Guillet, J. E. J. Macromol. Sci. Chem. 1968, A2, 14411455. (15) (a) Kanazawa, H.; Sunamoto, T.; Matsushima, Y.; Kikuchi, A.; Okano, T. Anal. Chem. 2000, 72, 5961-5966. (b) Kobayashi, J.; Kikuchi, A.; Sakai, K.; Okano, T. Anal. Chem. 2001, 73, 20272033.

Biomacromolecules, Vol. 6, No. 4, 2005 2327 (16) (a) Okano, T.; Yamada, N.; Sakai, H.; Sakurai, Y. J. Biomed. Mater. Res. 1993, 27, 1243-1251. (b) Kushida, A.; Yamato, M.; Konno, C.; Kikuchi, A.; Sakurai, Y.; Okano, T. J. Biomed. Mater. Res. 2000, 51, 216-223. (c) Yamato, M.; Okano, T. Mater. Today 2004, 7 (5), 42-47. (17) (a) Chung, J. E.; Yokoyama, M.; Suzuki, K.; Aoyagi, T.; Sakurai, Y.; Okano, T. Colloids Surf., B 1997, 9, 37-48. (b) Chung, J. E.; Yokoyama, M.; Aoyagi, T.; Sakurai, Y.; Okano, T. J. Controlled Release 1997, 53, 119-130. (18) Mayadunne, R. T. A.; Rizzardo, E.; Chiefari, J.; Chong, Y. K.; Moad, G.; Thang, S. H. Macromolecules 1999, 32, 6977-6980. (19) Yusa, S.; Shimada, Y.; Mitsukami, Y.; Yamamoto, T.; Morishima, Y. Macromolecules 2003, 36, 4208-4215. (20) (a) Mayadunne, R. T. A.; Rizzardo, E.; Chiefari, J.; Krstina, J.; Moad, G.; Postma, A.; Thang, S. H. Macromolecules 2000, 33, 243-245. (b) Schilli, C.; Lanzendorfer, M. G.; Mu¨ller, A. H. E. Macromolecules 2002, 35, 6819-6827. (21) (a) Lowe, A. B.; Sumerlin, B. S.; Donovan, M. S.; McCormick, C. L. J. Am. Chem. Soc. 2002, 124, 11562-11563. (b) Sumerlin, B. S.; Lowe, A. B.; Stroud, P. A.; Zhang, P.; Urban, M. W.; McCormick, C. L. Langmuir 2003, 19, 5559-5562. (22) Thang, S. H.; Chong, B. Y. K.; Mayadunne, R. T. A.; Moad, G.; Rizzardo, E. Tetrahedron Lett. 1999, 40, 2435-2438. (23) Mitsukami, Y.; Donovan, M. S.; Lowe, A. B.; McCormick, C. L. Macromolecules 2001, 34, 2248-2256. (24) (a) Chong, B. Y. K.; Krstina, J.; Le, T. P. T.; Moad, G.; Postma, A.; Rizzardo, E.; Thang, S. H. Macromolecules 2003, 36, 2256-2272. (b) Chiefari, J.; Mayadunne, R. T. A.; Moad, C. L.; Moad, G.; Rizzardo, E.; Postma, A.; Skidmore, M. A.; Thang, S. H. Macromolecules 2003, 36, 2273-2283. (25) Hagan, S. A.; Coombes, G. A.; Garnett, M. C.; Du, S. E.; Davies, M. C.; Illum, L.; Davis, S. S.; Harding, S. E.; Purkiss, S.; Gellert, P. R. Langmuir 1996, 12, 2153-2161. (26) Zhao, C.; Wang, Y.; Winnik, M. A.; Riess, G.; Croucher, M. D. Langmuir 1990, 6, 514-516. (27) Kalyanasundaram, K.; Thomas, J. K. J. Am. Chem. Soc. 1977, 99, 2039-2044. (28) Opanasopit, P.; Yokoyama, M.; Watanabe, M.; Kawano, K.; Maitani, Y.; Okano, T. Pharm. Res. 2004, 21, 2001-2008. (29) Dong, D. C.; Winnik, M. A. Can. J. Chem. 1984, 62, 2560-2565. (30) Ringsdorf, H.; Konings W. N. Biochim. Biophys. Acta 1992, 1106, 23-30. (31) Cammas, S.; Matsumoto, T.; Okano, T.; Sakurai, Y.; Kataoka, K. Mater. Sci. Eng., C 1997, 4, 241-247.

BM050232W

Polymer Terminal Group Effects on Properties of Thermoresponsive

Recommend Documents