Synthesis and Evaluation of Partly Fluorinated ... - ACS Publications

374

Biomacromolecules 2009, 10, 374–381

Synthesis and Evaluation of Partly Fluorinated Block Copolymers as MRI Imaging Agents Hui Peng,†,‡ Idriss Blakey,†,‡ Bronwin Dargaville,†,‡ Firas Rasoul,†,‡ Stephen Rose,† and Andrew K. Whittaker*,†,‡ Centre for Magnetic Resonance and Australian Institute for Bioengineering and Nanotechnology, The University of Queensland, St. Lucia 4072, Australia Received October 9, 2008; Revised Manuscript Received November 25, 2008

A series of well-defined diblock copolymers of acrylic acid with partially fluorinated acrylate and methacrylate monomers were synthesized using ATRP as potential 19F MRI imaging agents. The diblock copolymers could undergo spontaneous self-assembly in mixed and aqueous solvents to form stable micelles with a diameter from approximately 20-45 nm, having a fluorine-rich core that provides a strong signal for MRI examinations. The observed MRI image intensities were related to the NMR longitudinal and transverse relaxation times, and were found to depend on polymer structure and method of micellization. Two distinct T2 relaxation times were measured; on comparison of expected MRI image intensities with those observed experimentally, it was found that methacrylate polymers show systematically lower signal intensity than acrylate polymers. This is related to the presence of a population of nuclear spins having very short T2 relaxation times that cannot be detected under high-resolution NMR and MRI conditions.

1. Introduction Recently there has been intense interest in the development of fluorinated molecules to allow tracking of therapeutic particles and cells in vivo. The motivation for this is the very high selectivity of the 19F imaging experiment because, unlike in 1H NMR imaging, the body does not contain a confounding endogenous fluorine background signal. In principal, therefore, if doubly tuned MRI coils are available, highly selective 19F images can be superimposed on high-resolution anatomical 1H images, thus allowing tracking of suitably labeled biomarker molecules. With this aim in mind, a number of groups have proposed strategies for incorporation of fluorine into cells or for potential attachment to targetting molecules. The approaches range from use of self-assembled perfluorinated surfactants to partly fluorinated small and large molecules. In this paper we present the synthesis and properties of a range of partly fluorinated amphiphilic block copolymers with potential as 19F MRI imaging agents. The field of 19F medical imaging has its origins in the work of Holland and co-workers who reported in 1977 the images of sodium fluoride solutions and perfluorotributylamine within phantoms.1 These authors suggested but did not report direct imaging of fluorinated molecules in vivo. Subsequently, McFarland et al.2 reported abdominal images of rats injected with large doses of fluorocarbons; in that paper the authors discuss the effects of the large chemical shift dispersion of fluorine on the image quality. The focus of the research then shifted to the known dependence of the 19F longitudinal relaxation time on the oxygen tension in the blood stream.3 Later, Eidelberg4 reported images of the cat brain after infusion with a fluorocarbon to a level corresponding to 30 vol % of the vascular fluid (approximately 2.4 M); maps of cerebral oxygen tension were reported. At the same time, Longmaid et al.5 reported * To whom correspondence should be addressed. Tel.: +61-7-33463885. Fax: +61-7-33463973. E-mail: [email protected]. † Centre for Magnetic Resonance. ‡ Australian Institute for Bioengineering and Nanotechnology.

images of the liver of rats repeatedly dosed with emulsions of perfluortributylamine. More recently, a number of workers, notably Ahrens et al.6,7 and Partlow and co-workers,8 have reported the tracking of immunotherapeutic and stem/progenitor cells in vivo. Both of these groups have loaded cells in culture with emulsions of perfluoropolyethers or perfluorooctylbromide. Most recently, the group of Ahrens has reported the development of fluorescently labeled linear perfluoropolyethers to allow fluorescence microscopic analysis of the cells and cell sorting.9 Higuchi et al.10 have reported the imaging of amyloidophilic compounds containing fluorine as potential markers for the early stages of Alzheimer’s disease; Flaherty and colleagues11 have proposed a modified structure with higher fluorine content to address potential problems with low signal-to-noise ratio. Most recently, Du and co-workers have reported the synthesis and testing of partially fluorinated hyperbranched copolymers.12 A series of hyperbranched molecules were prepared by selfcondensed vinyl copolymerization of 4-chloromethyl styrene and lauryl acrylate from a trifunctional ATRP initiator, followed by ATRP statistical copolymerisation of trifluoroethyl methacrylate and t-butyl acrylate. The t-butyl acrylate groups were hydrolyzed to acrylic acid to form a water-soluble block copolymer. 19F T1 and T2 values of 500 and 50 ms, respectively, were reported, and images were obtained over an extensive imaging period. In summary, the potential use of fluorinated markers for noninvasive studies of physiology is abundantly clear and has been reviewed by Yu and others.13 The lack of clinical MRI scanners capable of detection of fluorine has meant that the early work in vivo remained of largely scientific interest; however, in recent years, commercial scanners have become available with dedicated single- or double-tuned resonators. In response to this, there has been renewed interest in the field. However, a number of fundamental challenges identified in the early studies remain. These include (1) the low sensitivity of the method due to the low fluorine

10.1021/bm801136m CCC: $40.75  2009 American Chemical Society Published on Web 01/07/2009

Block Copolymers as MRI Imaging Agents

content of many potential marker molecules, (2) the large chemical shift dispersion of the 19F nucleus, and (3) the short T2 relaxation times observed for potential structures arising from the large shift anisotropy and strong dipole-dipole couplings to near-neighbor fluorine and proton nuclei. It is the aim of this paper to present the first use of partially fluorinated amphiphilic block copolymers to address each of these issues in turn. The inherent low sensitivity of current biomarkers is addressed through synthesis of structures having fluorine-rich structures. Simple chemical structures have been chosen to eliminate problems due to multiple resonances in the 19F spectra. The effects of polymer and micellar structure on the spin-spin and spin-lattice relaxation times and, hence, ultimately on image intensity, are discussed.

2. Experimental Section 2.1. Materials. tert-Butyl acrylate(tBA), 2,2,2-trifluoroethyl acrylate (TFEA), 2,2,2-trifluoroethyl methacrylate (TFEMA), 1,1,1,3,3,3hexafluoroisopropyl acrylate (HFPA), 1,1,1,3,3,3-hexafluoroisopropyl methacrylate (HFPMA), and n-butyl acrylate (BA) were purchased from Aldrich and dried over CaH2 and then distilled under reduced pressure before use. CuBr was obtained from Aldrich and purified by washing with glacial acetic acid, followed by absolute ethanol and ethyl ether, and then dried under vacuum. Ethyl 2-bromopropionate (EBrP), N,N,N′,N′′,N′′-pentamethyl diethylene triamine (PMDETA), and trifluoroacetic acid (TFA) were purchased from Aldrich and used as received. Acetone, toluene, dichloromethane (DCM), and dimethylformamide (DMF) were refluxed over CaH2 and distilled under nitrogen before use. Poly(tert-butyl acrylate) (ptBA) was synthesized and purified according to the method reported by Matyjaszewski et al.14 The expected structure and molecular weight of PtBA was confirmed by 1H NMR and GPC. The purified yield was 84.9% after polymerization for 6.5 h. Mn ) 6803, Mw/Mn ) 1.09. 2.2. Characterization Techniques. 1H NMR spectra of the polymer intermediates were measured on a Bruker Avance 500 MHz spectrometer with a TXI probe at 298 K. A 90° pulse of 10 µs and a repetition delay (D1) of 10 s were used in all cases. The spectrum width was 8 kHz, and 32 k data points were collected. Gel permeation chromatography (GPC) measurements were performed using a Waters Alliance 2690 Separations Module equipped with an autosampler, column heater, differential refractive index detector, and a photodiode array (PDA) connected in series. HPLCgrade tetrahydrofuran was used as eluent at a flow rate of 1 mL/min. The columns consisted of three 7.8 × 300 mm Waters Styragel GPC columns connected in series, comprising two linear UltraStyragel and one Styragel HR3 columns. Poly(styrene) standards ranging from 517 to 2 × 106 g mol-1 were used for calibration. Differential scanning calorimetry (DSC) analyses were performed using a Mettler Toledo DSC1 Star System calorimeter with a subambient temperature attachment. The glass transition temperature (Tg) was determined from the midpoint of the change in heat capacity, measured at a heating rate of 10 °C/min. Dynamic light scattering (DLS) measurement was performed on a Nanoseries (Malvern, UK) zetasizer. The scattering angle used was 90° and the temperature was fixed at 298 K. 19 F NMR spectroscopy and imaging experiments were performed on an AMX300 spectrometer interfaced to a 7 T vertical superwide bore magnet. The system is equipped with a Bruker microimaging gradient set and the probe used was a Bruker 5 mm 19F single-tuned bird-cage resonator probe tuned to 282.404 MHz for fluorine detection. The 90° pulse time was 5.5 µs. Samples of the fluorinated block copolymers were prepared as described above and either maintained in mixed DMF-water solvent or examined after dialysis in water. The concentration of the solutions prepared in 5 mm NMR tubes was around 2.5 wt % of polymer. All measurements were performed at 310 K.

Biomacromolecules, Vol. 10, No. 2, 2009

375

19 F NMR spectra of the micelle solutions were recorded using a 90° pulse of 5.5 µs and a repetition delay of 10 s. The spectrum width was 50 kHz, and 4 k data points were collected. Typically, 16 acquisitions were coadded to improve the signal-to-noise ratio. 19 F T2 relaxation times were measured using the CPMG pulse sequence, with from 2 to 256 180° pulses in the echo train. A total of 12 data points were collected. Two separate experiments were performed to allow the initial short T2 decay and the very long T2 decay to be delineated. The echo times used for these two experiments were 100 µs and 10 ms. The data collected using the shorter echo time was best described by the sum of two exponential relaxation decays, while that collected using the long echo time was adequately described by a single exponential function, as the rapidly relaxing component had disappeared during the shortest echo time used. Spin-lattice relaxation times were measured using the standard inversion-recovery pulse sequence, with 16 value of inversion time used. A single T1 relaxation time was observed for all samples. 19 F images were collected on the same spectrometer using the 3D spin-echo pulse sequence. Typical experimental details are as follows. The field of view was 20 × 20 mm2, and the slice thickness was 2.5 mm in a 128 × 128 × 8 matrix. The echo time was 4.8 ms and the repetition time was equal to 3 s. The images presented here were all acquired by coadding two acquisitions, resulting in a constant imaging time of 1 h and 20 min.

2.3. Synthesis of Poly(tert-butyl acrylate)-b-poly[n-butyl acrylate)co-)fluoroacrylate] (p(tBA)-b-p(nBA-co-FA)). Monobromo-terminated p(tBA) (Mn ) 6803, Mw/Mn ) 1.09) and CuBr were added to a 50 mL round-bottom flask, which was sealed with a rubber septum and then degassed with argon. Deoxygenated toluene was added, after which nBA and the fluoroacrylate (FA, here FA refers to TFEA, TFEMA, HFPA, and HFPMA monomer) were added via syringes, and then PMDETA was introduced. The molar ratio of p(tBA)/CuBr/PMDETA/ nBA/FA ) 1:1:1:30:100. The solution was stirred until the Cu complex had formed. After complex formation, the flask was placed in an oil bath at 90 °C. After certain times, the polymer was dissolved in THF and passed through an alumina column to remove the copper catalyst. The THF was removed by evaporation and the polymer was then dried under vacuum at 40 °C. For further purification, the polymer was dissolved in THF, precipitated into a 10-fold excess of hexane, and then washed with hexane twice. The polymer was dried under vacuum at 40 °C for 2 days. 2.4. Hydrolysis: Preparation of Amphiphilic Polymers. A clean, 100 mL, round-bottom flask fitted with a stirrer bar was charged with p(tBA)-b-p(nBA-co-FA), followed by dried dichloromethane. The mixture was stirred for 10 min to dissolve the polymer. Trifluoroacetic acid (TFA, 5.0 equiv to the tert-butyl ester) was then added. After the mixture was allowed to stir at room temperature for 10 h, the dichloromethane and excess TFA were removed at room temperature with dry nitrogen gently flowing through the flask. The resulting lightbrown polymer solid was dried under vacuum for 3 days under room temperature. For further purification, the polymer was redissolved in THF, precipitated into a 10-fold excess of water, and then washed with water twice. The polymer was dried under vacuum at room temperature for 3 days to give to a white product pAA-b-p(nBA-co-FA). 2.5. Preparation of Polymeric Micelles. Spherical micelles were obtained by the dissolution of the purified amphiphilic copolymer pAAb-p(nBA-co-FA) (20 mg/mL) in a good solvent DMF for both blocks, followed by the gradual addition (10.0 mL/h) of an equal volume of nonsolvent for the hydrophobic polyfluoroacrylate to induce micellization. The micelles were allowed to stir for 24 h before being transferred to presoaked and rinsed dialysis bags (molecular weight cut off, MW ) 3500) and dialyzed against ultrapure water (18.2 mΩ · cm) for three days for removal of the remaining organic solvent. The particle sizes of the micelles were determined by DLS.

376


Peng et al.

Scheme 1. Synthetic Scheme Used for the Preparation of Amphiphilic Copolymers of PAA-b-P(nBA-co-TFE(M)A)

Table 1. Molecular Characteristics of the Copolymers Prepared in this Studya

a

polymer

Mn (GPC)

Mw/Mn

P(tBA)53 P(tBA)34-b-PTFEA44 P(tBA)50-b-P(nBA39-co-TFEA115) P(tBA)53-b-P(nBA28-co-TFEMA126) PtBA53-b-P(nBA4-co-HFPA3) PtBA53-b-P(nBA4-co-HFPMA11) PAA34-b-PTFEA44 PAA50-b-P(nBA39-co-TFEA115) PAA53-b-P(nBA28-co-TFEMA126) PAA53-b-P(nBA4-co-HFPA3) PtBA53-b-P(nBA4-co-HFPMA11)

6800 8670 19970 38350 7690 12540

1.09 1.33 1.47 1.32 1.13 1.27

COOH/nBu/CF3 (1H NMR)

1:1.3 0.44:0.34:1.0 0.42:0.22:1.0 17.67:1.33:1.0 4.82:0.36:1.0

Tg(°C; expt.)

Tg(°C calcd.)

31.8 22.8, -1.7 11.4, -5.2 34.8

43 43, -10 43, -21 43, 45 43,-36 43, 46 106, -10 106, -21 106, 45 106, -36 106, 46

94.4, -1.1 92.8, -9.3 87.6, 23.1 66.2, 23.9 102.4, 45.2

The Tg values in the right-hand column were calculated using the Fox equation as described in the text.

3. Results and Discussion In this study diblock copolymers having hydrophilic and hydrophobic blocks, in which the hydrophobic blocks are partially fluorinated, are prepared as potential 19F imaging agents. The copolymers can be self-assembled in water to form stable block copolymers; the hydrophobic core of the block is the “NMR-active” component of the imaging agent, and the physical and NMR properties of the core are especially important in determining the performance of the materials as imaging agents. The polymer must contain as high as possible a concentration of fluorine nuclei which have favorable NMR properties, as described below. These properties are intimately related to the physical properties of the material, most notably the glass transition temperature. For this reason the fluorinecontaining block was prepared as a statistical copolymer of n-butyl acrylate (n-BA) and the fluoro-monomer. The role of the n-BA is to lower the glass transition temperature of this component of the block copolymer, thus lengthening the spin-spin relaxation time of the fluorine nuclei, as explained below. 3.1. Synthesis of Poly(acrylic acid)-b-poly(n-butyl acrylate -co-2,2,2-trifluoroethyl(meth)acrylate) (PAA-b-P(nBA-co-TFE(M)A)) Amphiphilic Copolymers. The synthetic route used to prepare the amphiphilic copolymer PAA-b-P(nBA-co-TFE(M)A) is outlined in Scheme 1. The same procedures were followed for the synthesis of the block copolymers of HFP(M)A. The initial block of poly(t-butyl acrylate) was prepared using the ATRP method and resulted in polymers with controlled molecular weight and narrow polydispersity index. The polymeric macroinitiator was extended in the second step through the statistical copolymerization in toluene of trifluoroethyl

Figure 1. 1H NMR spectra of P(tBA)50-b-P(nBA39-co-TFEA115) (in CDCl3) (A) and PAA50-b-P(nBA39-co-TFEA115) (in DMSO) (B). Solvent peaks are marked with an asterisk.

(meth)acrylate and n-butyl acrylate. It can be seen from the properties of the polymers listed in Table 1 that the molecular weight distribution became slightly broader during the second step. Following this the tertiary-butyl groups were removed by hydrolysis in DCM solution; the solution 1H NMR spectra shown in Figure 1 show the complete disappearance of the peak at 1.38 ppm due to t-butyl protons and appearance of the peak at 12.2 ppm due to carboxylic acid protons, evidence of hydrolysis of the tertiary-butyl ester to carboxylic acid. The second blocks formed in the copolymerization are statistical copolymers of n-butyl acrylate and trifluoroethyl (meth)acrylate. Reactivity ratios for the copolymerization of TFEA and TFEMA (monomer 1) with n-butyl acrylate have been estimated using published Q-e values15,16 to be r1 ) 1.860, r2 ) 0.497 and r1 ) 2.618, r2 ) 0.376, respectively. Therefore, we expect these to be statistical polymers with faster initial incorporation of the fluorinated monomer and some tapering of the sequence distribution to longer blocks of n-butyl acrylate



377

Table 2. Physical and NMR Properties of the Copolymer Micelles Before and after Dialysis with Water polymer

fluorine content (wt %)

concentration (wt %)

particle size (nm)

T1 (ms; 37 °C)

19

T2 (ms; 37 °C)

F NMR line width (Hz)

image SNR

Before Dialysis PAA50-b-P(nBA39-co-TFEA115)

24.9

2.22

25.2

PAA53-b-P(nBA28-co-TFEMA126)

25.1

2.36

42.8

511 52% 676 26%

48% 20.14 74% 342.43

0.96

231

18.9

0.37

280

5.2

476 2241

8.8

After Dialysis PAA50-b-P(nBA39-co-TFEA115) PAA53-b-P(nBA28-co-TFEMA126)

24.9 25.1

2.68 2.5

at higher conversions. In our experiments, conversion was limited to around 40%. The thermal properties of the block copolymers were measured using DSC and the values of Tg listed in Table 1. For the diblock copolymers two resolved glass transition temperatures were observed, indicating effective separation of the two blocks in the solid state. The exception was the diblock copolymer of tBA with TFEMA-co-nBuA for which the expected glass transition temperatures of the two blocks closely coincide. The reported glass transition temperatures for the pure homopolymers of the seven monomers used are 109, 43, -54, -10, 72, -22, and 78 °C for polymers of acrylic acid, t-butyl acrylate, n-butyl acrylate, TFEA, TFEMA, HFPA, and HFPMA, respectively. 15,17,18 The glass transition temperatures expected for the respective phases of the copolymers calculated using the Fox equation and the copolymer compositions determined by 1H NMR spectroscopy are listed below in Table 1. The agreement between experimental and predicted is reasonable given the assumptions inherent in the method of calculation. 3.2. Formation of Micelles and Spin-Spin Relaxation Time Studies. The amphiphilic diblock copolymers were dissolved in DMF, and micelles formed by dropwise addition of water into the organic solution until an equivolume solution of water and DMF was obtained. Stable micelles of the block copolymers were achieved. The particles sizes were measured using dynamic light scattering and are listed below in Table 2. Particles of the order of 20-45 nm were formed and were stable over the duration of the subsequent experiments (several weeks). As the measured NMR parameters were not observed to depend on particle size, at least across the narrow size range we have investigated, this aspect was not further investigated. Spin-spin relaxation times of the 19F nuclei within the micelles of the block copolymers were measured using the CPMG pulse sequence. In all samples prior to dialysis two wellseparated T2 relaxation times was observed. After extensive dialysis with distilled water a single, short relaxation time was measured. The T2 relaxation times were found to increase from room temperature to 37 °C, consistent with the observed decrease in 19F line width at the higher temperatures. All subsequent measurements were made at physiological temperature. For all samples a single peak only was observed in the 19 F spectra. The observation of a single 19F resonance is important since this allows relatively simple MRI pulse sequences to be utilized for imaging. In the case of materials having multiple peaks in the 19F spectrum chemical shiftselective sequences would have to be employed to prevent ghosting in the MRI images. The spin-lattice (T1) and spin-spin (T2) relaxation times of the various block copolymer micelles in the mixed solvent at 37 °C are listed in Table 2. It can be seen that for all block copolymer micelles two T2 relaxation times are observed. The

21.8 39.8

526 509

1.75 0.37

proportion of the short T2 relaxation time is close to 50% for the acrylate copolymer but significantly higher for the methacrylate polymer (Table 2). Other work in progress has shown that the proportions and absolute values of the relaxation times depend on the method of manufacture of the micelles and the identity of the solvent used. The short relaxation times are the result of incomplete averaging of the dipolar couplings and a contribution from the large chemical shift anisotropy of the fluorine nuclei.19 It is well-known that the presence of solvents in polymeric materials leads to higher segmental and side chain mobility, and hence more effective averaging of these dipolar and chemical shift interactions. We can conclude that the short relaxation time arises from fluorine nuclei in chain segments experiencing relatively restricted motion, while the longer T2 relaxation time is due to chain segments undergoing largeamplitude molecular motion. The 19F T2 relaxation times of the two block copolymers in the good solvent DMF (i.e., prior to micellization) are 331 (TFEA) and 249 ms (TFEMA), respectively. The micelles in the presence of the mixed solvent therefore appear to consist of a central core having either high entanglement density and/or lower content of organic solvent, surrounded by a shell of hydrophobic polymer extensively plasticized by solvent and less entangled. The micelles are stabilized by an outer PAA shell as schematically represented in Scheme 1. A single T1 relaxation time of around 500-600 ms was observed in all cases, and indicates that the spectral density of motion of the fluorine nuclei in the mid-MHz range is not appreciably affected by differences in chain packing. This is not unexpected for methyl groups which will be undergoing rapid rotation around the C3 axis. T2 relaxation times, which are sensitive to lower-frequency motions, are expected to be much more sensitive to changes in packing density proposed here. It is instructive to consider the differences in behavior of the diblock copolymers of either fluorinated acrylate or methacrylate units. The hydrophobic block of the former has a glass transition temperature in the bulk state well below the measurement temperature (-9.3 °C, see Table 1), while the methacrylate polymer has a Tg closer to the measurement temperature (23.1 °C). As a result, one would expect significantly larger-scale segmental motion of the acrylate backbone compare with the methacrylate. When designing these polymers it was therefore expected that the spin-lattice relaxation times of the fluorine nuclei attached the acrylate backbones would be substantially longer, as a result of the anticipated more-efficient averaging of the dipole-dipole and chemical shift interactions. What is observed and reported in Table 2 is a small population of spins with a longer spin-spin relaxation time for the methacrylate. It is apparent that greater free volume is available in the methacrylate diblock polymers for side chain motion despite the higher glass transition temperatures in these polymers. This

378


Peng et al.

Figure 2. 3D spin echo 19F MRI images of solutions of block copolymer micelles in mixed solvent (A, B, D, E, F) and in pure water after dialysis (C).

could be ascribed to a higher content of plasticizing solvent in the methacrylate polymer compared with the acrylate, increasing the effective free volume around the fluoroethyl side chains. Alternatively, there may be a proportion of the chains in the methacrylate micelles, presumably close to the hydrated shell, with greater free volume. In fact, as will be demonstrated below, the behavior of these polymers is more complex again, with a significant proportion of the fluorinated components not being detected under the high-resolution NMR conditions employed here. The spin-lattice (T1) relaxation times are also listed in Table 2. A shorter spin-lattice relaxation time is observed for the acrylate systems. The rate of T1 relaxation depends on spectral density of motion of dipoles in the MHz frequency range; the relatively short values, especially for the acrylate system shows that the molecules are close to a minimum in T1, and the longer value for the methacrylate consistent with the polymers being on the high-frequency side of the minimum in T1. Only a single value of T1 was observed for each sample, indicating that the processes responsible for the two T2 relaxation processes (variations in chain packing and extent of plasticization) do not strongly affect the high frequency motions of the side chains. 3.3. Properties of Micelles after Dialysis with Water. The micelles formed from block copolymers containing TFEA were found to be persistent and stable after dialysis for three days with water. The particles sizes listed in Table 2 show a small decrease in particle size which can be ascribed to removal of the swelling solvent from the particles during dialysis. Also listed in Table 2 are the T1 and T2 relaxation times of the dialysed particles. It is clear that remove of the residual plasticizing solvent results in a dramatic decrease in the spin-spin (T2) relaxation times, as a result of the decreased segmental mobility of the hydrophobic core. The spin-lattice relaxation times are

not changed significantly indicating only a small change in the spectra density of motions in the high MHz frequency range. 3.4. NMR Imaging of Polymeric Micelles. Dilute solutions of the partly fluorinated micelles were imaged at 7 T using a standard three-dimensional spin echo pulse sequence with an echo time of 4.8 ms, and a repetition time of three seconds. The imaging time was restricted to 80 min, and comparisons are made of the signal-to-noise ratio of the images under identical conditions. For a spin-echo imaging sequence the image intensity is a function of the fluorine content, the spin-spin and spin-lattice relaxation times, and the imaging echo and repetition times, as described in equation 1.20 Typical images of the block copolymer micelles, in the mixed solvent system and after dialysis are shown in Figure 2 below. The most intense images were obtained for the TFEA copolymer micelles despite the T2 relaxation times being significantly shorter than for the methacrylate micelles; the presence of a much larger proportion of fluorine nuclei having shorter T2 relaxation times leads to the poorer imaging behavior of the methacrylatecontaining micelles. After dialysis with water, the block copolymers containing TFEA remained in solution while the methacrylate micelles were not stable. Despite the decrease in spin-spin relaxation time, excellent images were obtained from the dialysed samples.

[

I ) N(F) 1 - 2 exp

(

) ( )] ( )

-(TR - TE ⁄ 2) + T1 -TR exp T1

exp

-TE T2

(1)

In eq 1, I is the image intensity, N(F) is a measure of the number of fluorine nuclei in the sensitive volume of the imager, TR and TE are the pulse sequence repetition and echo times,


Figure 3. Comparison of NMR and MRI intensities of block copolymer micelles in mixed solvent (filled: DMF/H2O ) 1:1; unfilled: DMSO/ H2O ) 1:1).

respectively, and T1 and T2 are spin-lattice and spin-spin relaxation times, respectively. It is possible to predict the imaging performance of the block copolymer micelles prepared in this study using eq 1, taking into account the structure of the materials (specifically the fluorine content) and the measured relaxation properties. The predicted relative image intensities are listed in Table 2, along with the experimentally determined signal-to-noise ratios from the MRI images. The relationship between the observed and predicted signal intensities is rather poor as seen in Figure 3. However, a number of conclusions can be drawn from this data. Most notably, the image intensities for the methacrylate polymers are systematically much lower than those for the acrylate polymers. Compare the results for TFEA (circles) and TFEMA (triangles), and HFPA (squares) and HFPMA (diamonds). After observation of the differences in predicted and observed relative image intensities, we examined carefully the highresolution 19F spectra of TFEA and TFEMA in a series of mixed solvents from pure deuterated-DMF to pure D2O. The 19F spectra of PAA50-b-P(nBA39-co-TFEA115) are shown in Figure 4, and demonstrate that the line width of the observed NMR signal increases as the solvent becomes more polar. On passing from DMF to D2O the block copolymers are forced to assemble into micelles with a corona of polar poly(acrylic acid). Evidently the mobility of the fluorine-containing side-chains becomes progressively more restricted on dilution with D2O. More importantly for the imaging experiments, the integrated intensities of the 19F signal of the trifluoroethyl side-chains decreases as solvent quality changes. The signal intensities are plotted in Figure 5, and while there is an approximately 30% decrease in intensity for the TFEA-containing block copolymers in D2O, the TFEMA-containing materials experience a decrease in intensity of 55% in 1:1 DMF/D2O and approximately 90% in pure D2O. This observation accounts for the lower MRI imaging intensity observed for the methacrylate polymers; in addition to the observed two T2 populations described above, there must exist a third population of very short T2 relaxation times, of the order of several microseconds, which is not visible in the highresolution NMR experiment (Figure 4) and in the MRI experiment. The receiver dead time used to acquire the high-resolution spectra was 15 µs, and thus nuclear spins with relaxation times shorter than say 3-5 µs will not be visible in the NMR spectrum. This population of spins corresponds to fluorine nuclei


379

Figure 4. Stack plot of the 19F NMR spectra of PAA50-b-P(nBA39-coTFEA115) in the DMF-d7-D2O with different volume fraction of D2O. The vertical scales have been adjusted to equal height.

Figure 5. Plot of the integrated intensities of 19F NMR spectra of the fluoro(meth)acrylate block copolymers in solutions ranging from pure DMF, through mixed solvents, and to D2O. Table 3. Solubility parameters, δ, of Monomer Fragments and Solvents Reported Previously or Calculated Following van Krevelen monomer or solvent

solubility parameter, δ (MPa1/2)

TFEA TFEMA nBuA AA DMF DMSO H2O

17.4621 17.1023 17.9515 24.615 24.722 26.422 4822

in effectively rigid, solid polymer. The model for the packing of the chains in the hydrophobic core is therefore very complex, with at least three environments identifiable by these NMR experiments. The behavior of the block copolymers in solution can be simply rationalized by considering the relative solubility parameters of the monomer fragments and the solvents found in the literature15,21,22 or calculated from the tables list by van Krevelen.23 The relevant solubility parameters are listed in Table

380


Peng et al.

Table 4. Physical and NMR Properties of the Copolymer Micelles Containing Hexafluoroisopropyl (Meth)acrylatea polymer

fluorine content (wt %)

concentration (wt %)

6.8

2.54

18.8

2.51

PAA53-b-P(nBA4-co-HFPA3) PAA53-b-P(nBA2-co-HFPMA11) a

particle size (nm) 334

T1 (ms; 37 °C) 2078 42% 512 24%

58.8

19

T2 (ms; 37 °C) 58% 334.8 76% 193.6


image SNR

0.32

805

11.9

0.27

464

3.7

In 1:1 DMF/H2O solution.

Table 5. Physical and NMR Properties of the Copolymer Micelles in Equimolar Mixed DMSO/H2O Solvent concentration (wt %)

T1 (ms; 37 °C)

PAA50-b-P(nBA39-co-TFEA115)

2.35

PAA53-b-P(nBA28-co-TFEMA126)

2.38

PAA53-b-P(nBA4-co-HFPA3)

2.40

PAA53-b-P(nBA4-co-HFPMA11)

2.45

515 45% 627 36% 620 39% 451 25%

polymers

3. Recall that a low polymer-solvent interaction parameter (good solvent) requires that the solubility parameters of the components be similar. Examination of Table 3 shows that all of the monomers comprising the core are strongly hydrophobic, and that on passing from the solvent DMF to water the only component to be effectively solubilized will be the acrylic acid corona of the particle. The values of δ also indicate that the methacrylate monomer is more polar than TFEA, and this property may help explain the lower stability of the micelles containing methacrylate segments. As mentioned above, Du and co-workers have prepared a series of partially fluorinated hyperbranched copolymers as potential imaging agents.12 The approach of these workers differs from that of the current study in a number of ways. First, in this study we have prepared diblock copolymers capable of self-assembly into particles 20-40 nm in diameter, as opposed to Du who prepared hyperbranched block copolymers through a two-stage ATRP polymerization. Second, we have utilized acrylate monomers, as opposed to the fluorinated methacrylates used by Du, and have demonstrated in this study that the acrylates provide superior imaging performance in the diblock micelles. The same design rules would be expected to apply to more complex architectures. Most importantly, we have identified a population of “MRI-invisible” spins that must be ascribed to immobile polymer segments and that adversely affect the imaging performance. Du et al.12 reported images obtained only after long acquisition times. This was ascribed to the relatively low fluorine contents, but may also be partly due to a population of spins not detected in the imaging experiment, as was found here. A series of hyperbranched molecules were prepared by selfcondensed vinyl copolymerization of 4-chloromethyl styrene and lauryl acrylate from a trifunctional ATRP initiator, followed by ATRP statistical copolymerisation of trifluoroethyl methacrylate and t-butyl acrylate. The t-butyl acrylate groups were hydrolyzed to acrylic acid to form a water-soluble block copolymer. 19F T1 and T2 values of 500 and 50 ms, respectively, were reported, and images were obtained over an extensive imaging period. 3.5. Block Copolymers Containing Hexafluoroisopropyl Side Chains. It is clear from an examination of equation 1 and a comparison of the results of measurements on micelles of block copolymer of acrylate and methacrylate monomers that the structure of monomer unit not only affects the NMR

T2 (ms; 37 °C) 55% 13.88 64% 147.8 61% 210.8 75% 71.93

19


image SNR

0.40

793

7.2

0.38

806

4.2

0.33

928

8.8

0.23

537

2.9

relaxation properties, but also dramatically affects the NMR image intensities. As discussed above it is believed that the most effective imaging agents will have long spin-spin relaxation times, short spin-lattice relaxation times, and high fluorine content. To address this last issue a series of diblock copolymers of acrylic acid and n-butyl acrylate-co-hexafluoroisopropyl (meth)acrylate were prepared. These materials have the potential to have significantly higher fluorine contents than the fluoroethyl analogues. Unfortunately, it was found that the polymerization under ATRP conditions of the statistical copolymer of n-butyl acrylate and hexafluoroisopropyl (meth)acrylate proceeded very slowly, most likely due to the strong electron-withdrawing nature of the hexafluoroisopropyl group. None-the-less, block copolymers with short hydrophobic blocks were obtained and their NMR and imaging properties examined. The results of the measurements of NMR relaxation times are listed in Table 4. Again, two spin-spin relaxation times are observed, and for these polymers it is difficult to postulate chain entanglement of the hydrophobic core considering the very low degree of polymerization of the n-butyl acrylate-co-hexafluoroisopropyl (meth)acrylate block. If chain entanglement is responsible for the observed short T2 relaxation time in these block copolymers, then the hydrophilic acrylic acid block must be involved. The relationships between polymer structure and NMR and relaxation properties are being further investigated and will be reported in a subsequent publication. Spin echo images of the micelles containing hexafluoroisopropyl (meth)acrylate are shown in Figure 2D,E. The intensity of the hexafluoroisopropyl methacrylate-containing material is especially low as a result of the high proportion of spin having a short T2 relaxation time. However, an excellent image is seen for the hexafluoroisopropyl acrylate-based micelles, despite the low fluorine content and the relatively long T1 relaxation time. 3.6. Effect of Preparation Conditions on NMR Relaxation and Imaging. The effect of method of sample preparation on the NMR properties was also investigated. In a first set of experiments, the influence of sample concentration on the relaxation times was examined. It was found that varying the polymer concentration in the DMF solution prior to micellization did not have an appreciable effect on the spin-spin relaxation times of the fluorine nuclei. In a second set of experiments micelle samples were prepared from DMSO solution. The NMR properties of the stable micelles


observed in the mixed solvent are listed in Table 5. DMSO appears to be slightly less effective at plasticizing the core of the micelles, and as in all cases, the spin-spin relaxation times of the micelles in DMSO/H2O are lower than in the DMF/H2O mixed solvent. This may be due to the higher viscosity of DMSO compared with DMF or to a lower tendency of the DMSO to associate with the micellar core. As seen in Table 3, the solubility parameter of DMSO is larger than DMF and, hence, we expect less association of DMSO with the hydrophobic core.

4. Conclusions This paper reports the first report of the synthesis and NMR/ MRI properties of partially fluorinated linear block copolymers as potential MRI imaging agents. A range of block copolymers consisting of hydrophilic blocks of poly(acrylic acid) and hydrophobic blocks of n-butyl acrylate copolymerized with a partially fluorinated acrylate or methacrylate. All of the materials prepared could form stable micelles in mixed water-organic solvents, and a number were stable in pure water. The NMR and MRI properties of the micelles in the mixed solvents were particularly favorable, and excellent MRI images were obtained. In general, the use of acrylate backbones confers greater chain flexibility and, hence, better imaging properties than the corresponding methacrylate; however, other considerations, such as the effect of plasticizing solvent, have to be taken into account when designing materials. Examination of the NMR relaxation transverse relation time behavior and signal intensities lead to the understanding that the core of the micelles must consist of regions of varying packing density. This leads to at least three populations of fluorinated side chains. The acrylate polymers also gave excellent images after full dialysis in water. The work highlights the importance of careful consideration of the three essential parameters for a successful imaging agent, namely, high fluorine content, long T2 and short T1 relaxation times. The results presented here allow further rational design and development of more effective polymer 19F imaging agents. Acknowledgment. The authors gratefully acknowledge funding for the International Biomaterials Research Alliance under the Queensland State Government National and International Research Alliance Program. Funding from the University of Queensland is also gratefully acknowledged. The authors would also like to thank Prof. Karen Wooley of Washington University in St. Louis for valuable discussions.


381

References and Notes (1) Holland, G. N.; Bottomley, P. A.; Hinshaw, W. S. J. Magn. Reson. 1977, 28, 133–136. (2) McFarland, E.; Koutcher, J. A.; Rosen, B. R.; Teicher, B.; Brady, T. J. J. Comput. Assist. Tomo. 1985, 9 (1), 8–15. (3) Parhami, P.; Fung, B. M. J. Phys. Chem. 1983, 87 (11), 1928–31. (4) Eidelberg, D.; Johnson, G.; Barnes, D.; Tofts, P. S.; Delpy, D.; Plummer, D.; McDonald, W. I. Magn. Reson. Med. 1988, 6 (3), 344– 52. (5) Longmaid, H. E., 3rd; Adams, D. F.; Neirinckx, R. D.; Harrison, C. G.; Brunner, P.; Seltzer, S. E.; Davis, M. A.; Neuringer, L.; Geyer, R. P. InVest. Radiol. 1985, 20 (2), 141–5. (6) Srinivas, M.; Morel, P. A.; Ernst, L. A.; Laidlaw, D. H.; Ahrens, E. T. Magn. Reson. Med. 2007, 58 (4), 725–734. (7) Ahrens, E. T.; Flores, R.; Xu, H.; Morel, P. A. Nat. Biotechnol. 2005, 23 (8), 983–987. (8) Partlow, K. C.; Chen, J.; Brant, J. A.; Neubauer, A. M.; Meyerrose, T. E.; Creer, M. H.; Nolta, J. A.; Caruthers, S. D.; Lanza, G. M.; Wickline, S. A. FASEB J. 2007, 21 (8), 1647–1654. (9) Janjic, J. M.; Srinivas, M.; Kadayakkara, D. K. K.; Ahrens, E. T. J. Am. Chem. Soc. 2008, 130 (9), 2832–2841. (10) Higuchi, M.; Iwata, N.; Matsuba, Y.; Sato, K.; Sasamoto, K.; Saido, T. C. Nat. Neurosci. 2005, 8 (4), 527–533. (11) Flaherty, D. P.; Walsh, S. M.; Kiyota, T.; Dong, Y.; Ikezu, T.; Vennerstrom, J. L. J. Med. Chem. 2007, 50 (20), 4986–4992. (12) Du, W.; Nystrom, A. M.; Zhang, L.; Powell, K. T.; Li, Y.; Cheng, C.; Wickline, S. A.; Wooley, K. L. Biomacromolecules 2008, 9 (10), 2826–2833. (13) Yu, J.-x.; Kodibagkar, V. D.; Cui, W.; Mason, R. P. Curr. Med. Chem. 2005, 12 (7), 819–848. (14) Davis, K. A.; Matyjaszewski, K. Macromolecules 2000, 33 (11), 4039– 4047. (15) Brandrup, J. K.; Immergut, E. H., Polymer Handbook, 3rd ed.; WileyIntersciences: New York, 1989. (16) Narita, T.; Hagiwara, T.; Hamana, H. Makromol. Chem., Rapid Commun. 1985, 6 (1), 5–7. (17) Zhou, D.; Teng, H.; Koike, K.; Koike, Y.; Okamoto, Y. J. Polym. Sci., Part A: Polym. Chem. 2008, 46 (14), 4748–4755. (18) Gong, B. M.; Chien, J. C. W. J. Polym. Sci., Part A: Polym. Chem. 1989, 27 (4), 1149–60. (19) Ulrich, A. S. Prog. Nucl. Magn. Reson. Spectrosc. 2005, 46 (1), 1– 21. (20) Hendrick, R. E. Magn. Reson. Imaging 1987, 5, 31–37. (21) Granville, A. M.; Boyes, S. G.; Akgun, B.; Foster, M. D.; Brittain, W. J. Macromolecules 2004, 37 (8), 2790–2796. (22) Barton, A. F. M. CRC Handbook of Solubility Parameters and Other Cohesion Parameters; CRC Press: Boca Raton, FL, 1983. (23) Van Krevelen, D. W.; Hoftyzer, P. J., Properties of Polymers: Correlations with Chemical Structure; Elsevier Science: Amsterdam, 1972; pp 136-143.

BM801136M

Synthesis and Evaluation of Partly Fluorinated ... - ACS Publications

Recommend Documents