Synthesis and Characterization of Variable-Architecture

Uniquely, variable-temperature aqueous-phase AFM was able to show that changes in the structures of these ... View: PDF | PDF w/ Links | Full Text HTM...
0 downloads 0 Views 587KB Size
Langmuir 2007, 23, 41-49

41

Synthesis and Characterization of Variable-Architecture Thermosensitive Polymers for Complexation with DNA† Sivanand S. Pennadam, James S. Ellis, Matthieu D. Lavigne,‡ Dariusz C. Go´recki,‡ Martyn C. Davies, and Cameron Alexander* The School of Pharmacy, Boots Science Building, UniVersity of Nottingham, UniVersity Park, Nottingham NG7 2RD, U.K., and School of Pharmacy and Biomedical Sciences, and Institute of Biomedical and Biomolecular Science, UniVersity of Portsmouth, St. Michael’s Building, White Swan Road, Portsmouth PO1 2DT, U.K. ReceiVed July 10, 2006. In Final Form: September 21, 2006 Copolymers of N-isopropylacrylamide with a fluorescent probe monomer were grafted to branched poly(ethyleneimine) to generate polycations that exhibited lower critical solution temperature (LCST) behavior. The structures of these polymers were confirmed by spectroscopy, and their phase transitions before and after complexation with DNA were followed using ultraviolet and fluorescence spectroscopy and light scattering. Interactions with DNA were investigated by ethidium bromide displacement assays, while temperature-induced changes in structure of both polymers and polymer-DNA complexes were evaluated by fluorescence spectroscopy, dynamic light scattering, laser Doppler anemometry, and atomic force microscopy (AFM) in water and buffer solutions. The results showed that changes in polymer architecture were mirrored by variations in the architectures of the complexes and that the overall effect of the temperature-mediated changes was dependent on the graft polymer architecture and content, as well as the solvent medium, concentrations, and stoichiometries of the complexes. Furthermore, AFM indicated subtle changes in polymerDNA complexes at the microstructural level that could not be detected by light scattering techniques. Uniquely, variable-temperature aqueous-phase AFM was able to show that changes in the structures of these complexes were not uniform across a population of polymer-DNA condensates, with isolated complexes compacting above LCST even though the sample as a whole showed a tendency for aggregation of complexes above LCST over time. These results indicate that sample heterogeneities can be accentuated in responsive polymer-DNA complexes through LCST-mediated changessa factor that is likely to be important in cellular uptake and nucleic acid transport.

Introduction The field of gene therapy requires vector systems that can protect an exogenous nucleic acid during transport across cellular barriers but then release the gene payload as and when required.1-6 Viruses are able to package and release genetic material very effectively and are able to transfect many cell types with high efficiency.7-9 However, their use in the clinic remains problematic, owing to concerns over integration of viral nucleic acids in the host genome and repeated use risks instigation of the host immune response.10,11 As a consequence, nonviral vectors are widely regarded as the most likely materials for gene therapy in the clinic. For DNA therapies, the synthetic vectors must transport the biopolymer ‘payload’ to the nucleus followed by release for † Part of the Stimuli-Responsive Materials: Polymers, Colloids, and Multicomponent Systems special issue. * To whom correspondence should be addressed. Tel: +44 (0) 846 7678. Fax: +44 (0) 115 951 5102. E-mail: [email protected]. ‡ University of Portsmouth.

(1) Hillery, A. M.; Lloyd, A. W.; Swarbrick, J., Eds. Drug DeliVery and Targeting; Taylor and Francis: London, 2001. (2) Lechardeur, D.; Verkman, A. S.; Lukacs, G. L. AdV. Drug DeliVery ReV. 2005, 57, 755-767. (3) Mastrobattista, E.; van der Aa, M.; Hennink, W. E.; Crommelin, D. J. A. Nat. ReV. Drug DiscoVery 2006, 5, 115-121. (4) Eliyahu, H.; Barenholz, Y.; Domb, A. J. Molecules 2005, 10, 34-64. (5) Gardlik, R.; Palffy, R.; Hodosy, J.; Lukacs, J.; Turna, J.; Celec, P. Med. Sci. Monit. 2005, 11, RA110-RA121. (6) Emerich, D. F.; Thanos, C. G. Curr. Nanosci. 2005, 1, 177-188. (7) Go´recki, D. Emerging Drugs 1999, 4, 247-261. (8) Green, N. K.; Seymour, L. W. Cancer Gene Ther. 2002, 9, 1036-1042. (9) Kochanek, S.; Schiedner, G.; Volpers, C. Curr. Opin. Mol. Ther. 2001, 3, 454-463. (10) Shoji, Y.; Nakashima, H. Curr. Pharm. Des. 2004, 10, 785-796. (11) Kabanov, A. V.; Batrakoval, E. V.; Sherman, S.; Alakhov, V. Y. In Polymer Therapeutics II: Polymers As Drugs, Conjugates And Gene DeliVery Systems; Springer: Berlin, 2006; Vol. 193, pp 173-198.

transcription and ultimately translation, whereas for RNA, delivery into the cytosol may be sufficient for regulation of the gene message12,13 However, irrespective of the nucleic acid, the conflicting demands of protection and release are still formidable14-16sand thus responsive polymers that are able to switch states, for example, between potentially ‘binding’ and ‘nonbinding’, have emerged as promising candidates to address the problem of controlled DNA binding and release.17-24 The structures of these responsive polymer-DNA complexes are likely to be critical in controlling their gene transfection efficacy, but the factors affecting these structures at temperatures above and below their phase transitions remain rather less studied in comparison with nonresponsive vectors. In previous papers, we showed that poly(ethyleneimine) (PEI) polymers with attached (12) Jensen, K. D.; Nori, A.; Tijerina, M.; Kopeckova, P.; Kopecek, J. J. Controlled Release 2003, 87, 89-105. (13) Fisher, K. D.; Hermiston, T.; Seymour, L. W. Curr. Opin. Mol. Ther. 2002, 4, 289. (14) Belting, M.; Sandgren, S.; Wittrup, A. AdV. Drug DeliVery ReV. 2005, 57, 505-527. (15) Ferrari, S.; Griesenbach, U.; Geddes, D. M.; Alton, E. Clin. Exp. Immunol. 2003, 132, 1-8. (16) Wiethoff, C. M.; Middaugh, C. R. J. Pharm. Sci. 2003, 92, 203-217. (17) Hinrichs, W. L. J.; Schuurmans-Nieuwenbroek, N. M. E.; Van De Wetering, P.; Hennink, W. E. J. Controlled Release 1999, 60, 249-259. (18) Kurisawa, M.; Yokoyama, M.; Okano, T. J. Controlled Release 2000, 69, 127-137. (19) Kurisawa, M.; Yokoyama, M.; Okano, T. J. Controlled Release 2000, 68, 1-8. (20) Takeda, N.; Nakamura, E.; Yokoyama, M.; Okano, T. J. Controlled Release 2004, 95, 343-355. (21) Turk, M.; Dincer, S.; Yulug, I. S.; Piskin, E. J. Controlled Release 2004, 96, 325-340. (22) Piskin, E.; Dincer, S.; Turk, M. Gene Ther. 2004, 11, 51. (23) Oupicky, D.; Reschel, T.; Konak, C.; Oupicka, L. Macromolecules 2003, 36, 6863-6872. (24) Oupicky, D.; Diwadkar, V. Curr. Opin. Mol. Ther. 2003, 5, 345-350.

10.1021/la061992a CCC: $37.00 © 2007 American Chemical Society Published on Web 11/17/2006

42 Langmuir, Vol. 23, No. 1, 2007

Pennadam et al. Table 1. Key Properties of Polymers Used in This Study

a

polymer

Mw (kDa)

% PNIPAm

λmax change/°C

cloud point (PBS pH 7.4)

P1 (PNIPAm-co-DANSAPP 1) P2 (PNIPAm-co-DANSAPP 2) P3 PEI-PNIPAm P4 PEI-PNIPAm

34a 17a 86b 82b

99.2 99.4 71 70

29 31 33 41

31 36 34 44

Polymer molar masses were determined by GPC and NMR. b Polymer molar masses were calculated from NMR integral ratios and amine content.

poly(N-isopropylacrylamide) (PNIPAm) were able to transfect cells,25,26 and here we report the results of further investigations into the behavior of thermoresponsive PEI-g-PNIPAm copolymers and the effects of phase transitions on the properties of the complexes as determined by physical chemistry techniques, including light scattering, fluorescence spectroscopy, and atomic force microscopy (AFM). Experimental Section Materials and Methods. Reagents, monomers, and solvents for chemical synthesis were purchased from Aldrich, Acros, or Fisher Scientific (UK) in the highest purities available and used as received. Branched poly(ethyleneimine) (PEI, fw ≈ 25 kDa, Aldrich),was dialyzed (3 kDa cutoff) against deionized water (5 × 1000 mL) and lyophilized prior to use. Inhibitors were removed from N-isopropylacrylamide (NIPAm) by recrystallization from hexane. Synthesis of Monomers. The fluorescent monomer N-[2-[[[5(N,N-dimethylamino)-1-naphthalenyl]sulfonyl]amino]propyl]-2-propenamide (DANSAPP) was prepared as described by Twaites et al.25 Synthesis of Polymers. Copolymers (P1, P2) of NIPAm were prepared with the fluorescent marker monomer (DANSAPP). Conjugation of these polymers to PEI via heterobifunctional crosslinker chemistry generated branch-graft polymers (P3, P4) (Table 1). PNIPAm Copolymers P1, P2. A Schlenk vessel was charged with NIPAm (1.75 g, 15.46 mmol), DANSAPP (87.5 mg, 0.24 mmol), 2,2′-azobisisobutyronitrile (AIBN, 35 mg, 0.213 mmol), and 2-aminoethanethiol hydrochloride (35 mg, 0.309 mmol) in propanol (20 mL). The resulting solution was degassed by three freeze-thaw cycles under vacuum. The vessel was then placed in a thermostated oil bath at 65 °C for 24 h. Once cooled to room temperature, the solution was concentrated under reduced pressure and the residue added to diethyl ether (500 mL) to precipitate the polymer. The precipitated material was redissolved in THF and reprecipitated into light petroleum ether (40-60 °C) three times to afford the purified copolymer as a colorless precipitate, which was then dried in vacuo overnight. Polymer P2 was made in an analogous way using monomers in the appropriate molar ratios: molecular weight was controlled by the addition of initiator and chain transfer agent, respectively. PEI/PNIPAm P3, P4. The amine terminus group of polymer P1 or P2 was activated using the heterobifunctional cross-linker N-(maleimidocaproyloxy)succinimide ester (EMCS, Pierce). To a solution of polymers P1 or P2 (500 mg in 1 mL PBS) EMCS (324 µg, 50 equiv) in DMSO (32 µL) was added. The reaction was stirred for 30 min, PBS (2 mL) was added, and the polymer was separated from unreacted EMCS and concentrated using a Vivaspin 6 membrane concentrator (50 kDa cutoff) against PBS (3 × 3 mL). The activated polymer was then added to bPEI (excess) in water, and the pH was adjusted to 9.0 prior to reaction overnight at room temperature with stirring. The polymer solution was dialyzed to remove both unreacted PEI and PNIPAm-g-EMCS for 48 h against double distilled deionized water (50 kDa cutoff), changing the dialysis solution hourly for the first 10 h and then at 12 h intervals or until no dansyl-labeled polymer (25) Twaites, B. R.; De las Heras Alarco´n, C.; Lavigne, M.; Saulnier, A.; Pennadam, S. S.; Cunliffe, D.; Go´recki, D. C.; Alexander, C. J. Controlled Release 2005, 108, 472-483. (26) Twaites, B. R.; de las Heras Alarcon, C.; Cunliffe, D.; Lavigne, M.; Pennadam, S.; Smith, J. R.; Gorecki, D. C.; Alexander, C. J. Controlled Release 2004, 97, 551-566.

could be detected in the dialysate. The resulting polymers were lyophilised to yield an oily off-white solid (P3 or P4). Preparation of Polymer-DNA Complexes. Initial screening experiments employed a stock solution of plasmid DNA (pX61 or pCS2) diluted into a total volume of 3 mL in double-distilled DNAsefree water or HEPES buffer (10 mM) at pH 7.4 or PBS pH 7.4 to give a solution of 40 µg/mL concentration. Polymers P3, P4 and PEI stock solutions were diluted into a volume of 3 mL in double distilled DNAse-free water or HEPES buffer at pH 7.4 or PBS ph 7.4 such that the resulting amine/phosphate (N/P) ratios of polymer to plasmid DNA were in the ranges 2:1, 2.5:1, 2.8:1, 3:1, and 4:1. For AFM analysis a stock solution of plasmid DNA (pX61 or pCS2) was diluted into a total volume of 450 µL in double-distilled DNAse-free water or PBS pH 7.4 to give a solution of 40 µg‚mL-1. Polymers P3, P4 and PEI stock solutions were diluted into a volume of 450 µL in double distilled DNAse free water or PBS pH 7.4. Complexes were formed as described below and then for solutions initially prepared in water the pH was adjusted using 100 µL of 10 mM HEPES buffer at pH 7.4 such that the resulting amine/phosphate (N/P) ratios of polymer to plasmid DNA were in the ranges 2:1, 2.5:1, 2.8:1, 3:1, and 4:1. In all cases, complexes were formed by rapid addition of polymer solutions (3 mL) to DNA solutions (3 mL) in a sterile tube with gentle shake of the mixed solutions. The mixed suspensions were allowed to stand for 30 min at room temperature prior to use. Instrumentation. Infrared Spectroscopy. Infrared spectra were obtained with a Perkin-Elmer Paragon 1000 FT-IR spectrophotometer, in transmittance mode, at a resolution of 4 cm-1. NMR Spectroscopy. 1H and 13C NMR spectra were recorded on JEOL EX-270 and Eclipse 400 spectrometers at 270 and 399.8 MHz (1H) and 67.6 and 100.5 MHz (13C), respectively, in D2O and CDCl3 solutions. All chemical shifts are reported in ppm relative to TMS or DSS. Gel Permeation Chromatography. The solvent system was tetrahydrofuran (THF) at 1.0 mL/min (nominal). A single solution of each sample was prepared by adding the polymer (10 mg) to solvent (THF, 1 mL) and with shaking to dissolve. The solutions were then filtered (0.4 µm cutoff), and part of each filtered solution transferred to glass sample vials. The vials were then placed in the instrument sample compartment, and after an initial delay of 15 min to allow the samples to equilibrate to room temperature, injection of part of the contents of each vial was carried out automatically. Molecular weights were calculated relative to PMMA or poly(ethylene oxide)/poly(ethylene glycol) calibrants. A Waters 150CV GPC instrument fitted with PLgel 2 × mixed bed-B or PLAquagel guard plus 2 × mixed-OH, 30 cm, 10 µm columns and refractive index detection was used. Data capture and subsequent data handling were carried out using Viscotek ‘Trisec’ 3.0 software. Ethidium Bromide Displacement Assay. Plasmid DNA (10 µL, 80 µg/mL, 20 nM) was added to PBS (75 µL), and emission intensity at 580 nm was recorded after each addition: for variable temperature experiments, fluorescence intensities for polymer-DNA complexes were corrected for the intrinsic changes in fluorescence, owing to increased nonradiative decay. Ethidium bromide displacement was recorded against the percentage free phosphate available for binding. Fluorescence Spectroscopy. Fluorescence spectra (excitation 330 nm, emission spectra 350-650 nm) were recorded in PBS using a Varian spectrophotometer, Cary Eclipse software). Polymer-DNA complexes were prepared by rapid addition of polymer to DNA in

Variable-Architecture ThermosensitiVe Polymers an Eppendorf tube at the appropriate N/P ratio followed by inversion of the tube. Measurements were carried out after 30 min maturation time. Dynamic Light Scattering. Hydrodynamic radii of the polymerDNA complexes were measured via scattered light recorded at 90° angle to incident radiation in a Viscotek 802 dynamic light scattering (DLS) instrument equipped with a 50 mW internal laser operating at a wavelength of 830 nm. From standard autocorrelation functions, measured diffusion coefficients were related to particle hydrodynamic radius via the Stokes-Einstein equation. In addition, it was assumed that particles were spherical and noninteracting. RH ) kT/6πηD Where RH is the hydrodynamic radius, k is the Boltzmann constant, T is the temperature, and η is the viscosity of the solvent. Measurements quoted are the averages of triplicate samples with at least 10 readings of particle size recorded at each temperature. Radii quoted are averages for samples where >90% of the scattered light in terms of numbers of particles was from complexes within the size range quoted. Laser Doppler Anemometry. Zeta potentials (ζ) of the polymerDNA complexes were measured on a Malvern Zetasizer 2000 equipped with a 10 mW He-Ne laser operating at a wavelength of 633 nm. Freshly prepared polymer-DNA complex suspensions at 80 µg/mL DNA were injected into the instrument: the total volume of sample was ∼3.5 mL. Each sample at the appropriate temperature was measured six times to check for repeatability. ζ (mV) was derived from the measured electrophoretic mobilities ((µm/cm‚V)‚s) using the Smoluchowski approximation. The statistical significance of the assessed temperature-response changes of zeta potentials of polymerDNA complexes was confirmed by paired Student’s t-test (P < 0.05). Atomic Force Microscopy. A Multimode AFM (Veeco Instruments, Santa Barbara, CA) operating with a NanoScope IIIa controller with a size E scanner was used throughout this study. For imaging in liquid, silicon nitride, oxide-sharpened, triangular cantilevers (Veeco Instruments) were selected, operating at resonant frequencies of ∼8-10 kHz. All imaging was conducted in tapping mode at 512 × 512 pixel resolution, with scan speeds of 3.05 Hz. We chose optimal set-points to reduce sample distortion. Post-capture analysis was conducted using either Nanoscope IIIa software (version 5.12 v b2) or SPIP version 3.3.60 (Image Metrology, Lyngby). Images were flattened using a first order polynominal function to adjust for the z offset and sample tilt. Images were then extracted as tagged image file (TIF) formats. The freeware WSxM version 4 development 8.4 (www.nanotec.es) was also used, as all the software automatically embeds a scale bar onto the image file. Following the flattening of each image file using Nanoscope Program, widths and diameters were measured using the line cross section of SPIP. This function provides a profile of the area under investigation. Using the cursor functions, dimension readout can be achieved providing a measurement reading. For imaging of polymer-DNA complexes, suspensions of the polymer-biopolymer condensates were incubated on mica for 10 min to allow maximal immobilization of complexes prior to analysis. This methodology was used so as to obtain as representative random samples of the population of complexes as possible. Grain size analysis was chosen to determine the mean diameters of condensates on the surface. In this instance, grain size identified the diameter of the core of the condensates, appearing as the white center in an AFM image. To test the reliability of grain size analytical methods, the diameters of a variety of random different DNA condensates were measured manually. The diameter was then obtained for each grain size using grain size analysis software in SPIP; these were compared against each manual measurement by means of regression analysis. The grain-size analysis was conducted using SPIP; each file was flattened to remove background noise. The detection level was adjusted such that both the background and unwanted data were removed from the sample. The segment image was cross checked

Langmuir, Vol. 23, No. 1, 2007 43 with the original AFM file to ensure the required condensates were highlighted. For each sample, a minimum diameter of 20 nm was chosen; in addition, any condensates or objects at the border of the image were ignored. We suggest the grain diameter is more representative of particle size than the grain surface area.27

Results and Discussion Synthesis of Monomers and Polymers. Two thermoresponsive polymers were prepared from NIPAm and the fluorescent monomer DANSAPP using 2-aminoethanethiol as chain transfer agent to control molecular weight and provide amine functionality for subsequent derivitisation. Spectroscopic analysis was consistent with the desired product, and 1H NMR integrals of diagnostic peaks (N(CH3)2 2.85 ppm for dansyl label compared to CH(CH3)2 3.90 ppm for PNIPAm) suggested 0.8 mol % incorporation of the fluorescent monomer in P1 and 0.6 mol % incorporation in P2. As expected from a conventional free-radical polymerization, copolymer composition did not exactly match monomer feed, and the molecular weight distributions were broad (Mw/Mn ) 2.3, 1.9 for P1 and P2, respectively). Determination of lower critical solution temperatures (LCSTs) was via measurement of cloud-point temperatures from buffered aqueous solutions of polymer. Conventional cloud-point determinations involve following changes in solution turbidity until no further increase is observedsthese are concentration dependent, as they measure overall extent of aggregation. We defined the LCST for this work as the onset of a sharp increase in absorption at 500 nm in order to have a concentrationindependent value for biological studies where the actual polymer concentration in complexes with DNA is not easily measurable. The onset of the phase transition was further confirmed using fluorescence spectroscopy, monitoring the increase in fluorescence emission intensity as the dansyl label experienced a more hydrophobic environment. The values for LCST of polymers P1 and P2 determined in this way were 29 and 31 °C, respectively. This corresponded with an expected decrease in LCST compared to pure PNIPAm through incorporation of the hydrophobic fluorescent monomer. Polymers P1 and P2 were then reacted with an excess of the heterobifunctional linker EMCS at pH 7.4 to generate the maleimide-tipped intermediates P1* and P2*, which were coupled to branched PEI (Figure 1). Structures of the resulting polymers P3 and P4 were confirmed by spectroscopy and extent of PNIPAm incorporation determined by integration of diagnostic signals at 1.14-1.17 and 3.94-3.98 ppm for the methyl and isopropyl protons of N-isopropyl side chains compared to the PEI methylene protons centered at 2.65 and 2.72 ppm. Dansyl aromatic protons at 8.2 and 8.4 ppm were detectable from the labeled PNIPAm side-chains, but the dimethylamino protons were masked by the PEI backbone methylene protons. However, fluorescence spectra and 13C resonances at 150 and 120 ppm left no doubt of the incorporation of the labeled side-chains. Integral ratios indicated that ∼1.9 PNIPAm chains of 34 kDa molecular mass were attached to the PEI backbone in P3 and ∼2.8 17 kDa PNIPAm chains were attached in P4. Further quantification of substitution was carried out by titration of primary amines on PEI and the substituted PEI-PNIPAm copolymers with 2,4,6-trinitrobenzenesulfonic acid (TNBS). The extent of substitution was found to be 1.15% of amine groups in P3 and 2.5% in P4, corresponding to the attachment of ∼1.8 and 3.5 PNIPAm chains to polymers P3 and (27) Wittmar, M.; Ellis, J. S.; Morell, F.; Unger, F.; Schumacher, J. C.; Roberts, C. J.; Tendler, S. J. B.; Davies, M. C.; Kissel, T. Bioconjugate Chem. 2005, 16, 1390-1398.

44 Langmuir, Vol. 23, No. 1, 2007

Pennadam et al.

Figure 1. Schematic and synthesis of copolymers P1-P4. Table 2. Particle Sizes of P3- and P4-Plasmid DNA Complexesa Polymer P3 polymer/DNA ratio

temperature (°C)

RH (nm)

w/w ) 0.8 N/P ) 2:1 RH (nm)

30 50

74 ( 25 (68%) 255 ( 31 (40%)b

75 ( 11 (73%) 75 ( 14 (67%)

P3 (no DNA)

w/w ) 1.0 N/P ) 2/5:1 RH (nm)

w/w ) 1/2 N/P ) 3:1 RH (nm)

w/w ) 1.6 N/P ) 4:1 RH (nm)

102 ( 31 (93%) 106 ( 16 (82%)

111 ( 8 (90%) 86 ( 14 (81%)

95 ( 25 (57%) 92 ( 15 (41%)b

Polymer P4 polymer/DNA ratio (N/P)

temperature (°C)

RH (nm)

w/w ) 0.8 N/P ) 2:1 RH (nm)

30 50

56 ( 10 (73%) 331 ( 30 (23%)b

94 ( 24 (96%) 80 ( 18 (97%)

P4 (no DNA)

w/w ) 1.0 N/P ) 2.5:1 RH (nm)

w/w ) 1.2 N/P ) 3:1 RH (nm)

w/w ) 1.6 N/P ) 4:1 RH (nm)

91 ( 47 (86%) 89 ( 40 (81%)

101 ( 34 (79%) 79 ( 36 (71%)

88 ( 37 (51%)b 267 ( 12 (85%)

a The percentage of the total population of the complex (intensity distribution) is given in parentheses for each particle size. b Remainder of the particle size distribution due to >1 µm aggregates.

P4. We therefore calculated final molecular masses for the polymers assuming two P1 chains were attached to PEI in P3 and three P2 chains were grafted to PEI in P4, corresponding to mass ratios of 71% PNIPAm in P3 and 70% PNIPAm in P4, respectively. Cloud-point measurements for P3 and P4 showed a slight increase in LCST of the branch graft copolymers compared to the homopolymer precursors. The measured onsets of the phase transition in isotonic buffer (phosphate buffered saline, PBS) at pH 7.4 were 33 (P3) and 41 °C (P4). Fluorescence spectra of the copolymers confirmed the onsets of the phase transitions, with changes in both fluorescence intensity and emission wavelength maxima between 33 and 37 °C for P3 and 41-44 °C for P4. Ethidium bromide displacement assays indicated these polymers were able to displace the intercalating dye from DNA at slightly lower N/P ratios than branched PEI alone, with complete displacement at N/P > 1.5:1. Complexes of these polymers were subsequently prepared with plasmid DNA (pDNA) at N/P ratios varying from 2:1 to 10:1. We initially prepared these complexes in double-distilled DNAse-free water and 10mM HEPES buffer at pH 7.4 in order to monitor intrinsic phase transition effects on complexes without the competing effects of high salt concentrations. Characterization of Polymer-DNA Complexes by Dynamic Light Scattering. Variable-temperature light scattering experiments of complexes at DNA concentrations of 500 nm) took place over a 2-3 h period. In all cases, DLS indicated that polymer-DNA complexes maintained above LCST for prolonged periods tended to associate into larger particles, although this effect was more noticeable for the higher N/P ratios. The stability to aggregation of these particles was likely to have been low, despite positive zeta potentials, owing to the presence of the hydrophobic PNIPAm domains above LCST. Similar effects on

heating of PEI-PNIPAm complexes were observed by Bisht et al, who reported expansion from 160 nm diameter particles to >250 nm above LCST for N/P ratios of 5:1 and 10:1. The variations apparent in the light scattering data suggested heterogeneous polymer-DNA complexes in the samples, and this was confirmed by AFM. Considerable differences in polymer-DNA complex size and geometry were apparent across the samples imaged by AFM for both P3 and P4. In particular,

48 Langmuir, Vol. 23, No. 1, 2007

some particles were poorly condensed and contained a core of ∼100 nm with a corona of DNA extending the overall diameter of the complexes to >500 nm in some cases. Close examination of representative poorly condensed complexes (P3, Figure 3gi) showed small reductions (∼10%) in the compacted core with temperature but no change in the outlying corona. The ‘corecorona’ structures most likely account for the discrepancies in the values obtained for radii of the complexes from light scattering data compared to AFM across the LCST. Judging by the particle structures and sizes from our AFM images, DLS was consistently reporting smaller complexessmost likely due to the more intense scattering that would have been observed from the highly condensed core compared to freely diffusing DNA strands. The indication from AFM that more subtle microstructural changes were occurring in these polymer-DNA complexes than could be detected by DLS led us to examine further the responses of the pendant PNIPAm chains as reported by the fluorescent dansyl label. Fluorescence spectra showed that architecture changes in P3 and P4 due to the phase transitions of their constituent P1 and P2 labeled PNIPAms gave rise to different environments for the fluorophore. In P1 and P2, the blue-shift in emission maxima above LCST could be obviously correlated with the reduction in hydration of the dansyl label as the PNIPAm chains collapsed. However, for P3 and P4, wherein the labels were contained on PNIPAm chains grafted to PEI, the appearance of the 420 nm maxima below LCST was strongly suggestive of micelle formation, with the PNIPAm chains at the core and the dansyl labels in low-mobility nonpolar environments, removed from the PEI. Above LCST, the 420 nm peak was reduced and the emission maxima for both polymers centered at ∼500 nm, as was observed for P1 and P2. This indicated that the dansyl labels were in a PNIPAm-rich domain that was more hydrophobic than for isolated PNIPAm domains below LCST, but at least partially hydrated. When complexed with DNA, P3 exhibited very similar fluorescence spectra to P1, implying that no micellar domains were present and that the PNIPAm chains were still able to exhibit phase transitions when the copolymer was bound to DNA. P4-DNA complexes displayed the 420 nm dansyl emission below LCST but showed similar spectral features to P2 above LCST, suggesting that there were contributions from both ‘micellar’ PEI-PNIPAm species and polymer-DNA complexes below LCST but more condensed complexes above LCST. For both P3 and P4 complexes with DNA, phase transitions were shown to take place, demonstrating that the PNIPAm chains were not strongly interacting with DNA. The changes in fluorescence maxima of P3-DNA and P4DNA complexes were also concentration dependent, with the greatest changes in the intensity of the absorption maxima at 500 nm at higher concentrations of polymer. We attribute this to association/aggregation of complexes above LCST, with the PNIPAm chains at higher total content able to associate and the dansyl labels retained in the hydrophobic PNIPAm regions. Further AFM analysis in PBS showed clearly the association and aggregation of P3 and P4 complexes above LCST, as predicted by the fluorescence spectroscopy experiments. The presence of large particles, most likely resulting from hydrophobic association and/or ‘salting-out’ due to the higher ionic strength,36 was more marked in P3 than P4 complexes. Once again, AFM indicated that polymer responses with temperature resulted in greater heterogeneities in the samples, with disproportionation of particle size taking place. Isolated polymer-DNA complexes compacted above LCST, whereas closely neighboring complexes showed (36) Kabanov, A. V.; Kabanov, V. A. AdV. Drug DeliVery ReV. 1998, 30, 49-60.

Pennadam et al.

a tendency to aggregate, and this was most apparent for P3 complexes at 3:1 compared to P4 at N/P 4:1. The total amount of PNIPAm was the same in these experiments for P3- and P4-DNA complexes (vide infra), thus a ‘diluting’ effect of the PEI segments, which were more abundant in P4 may have helped to reduce the chances of hydrophobic regions in neighboring complexes associating and thus causing aggregation. Significantly, as determined from grain size analysis, P3 complexes overall increased in size above LCST, presumably driven by hydrophobic self-association whereas P4 complexes increased in size up to polymer LCST (∼40 °C in PBS) before contracting. This behavior demonstrated that these two polymers, although very similar in chemical functionality, exhibited subtly different biophysical properties, owing to their different graft architectures. It also suggests that such polymers might display markedly different biological properties in terms of serum stability and/or cell entry/trafficking on account of their different conformational or architectural changes around body temperature. Indeed, the recent paper of Bisht et al.28 noted differences in the uptake of PEI-PNIPAm-DNA complexes above LCST, as well as enhanced transfection at the higher temperaturessresults which were attributed to increased surface charge density and reduced colloidal stability. For the P3 and P4 polymers in this study, it is expected that transfection experiments carried out around 37 °C, i.e., around the LCST of P3 but below the LCST of P4, would result in differences in transgene expression as a consequence of P3-DNA complexes being chain-collapsed and prone to association at hydrophobic sites, whereas P4 would be chain-extended, more hydrophilic, and less membrane-associative. Experiments to test this hypothesis are in progress.

Conclusions Cationic gene delivery vectors with responsive copolymer grafts containing an environment-sensitive fluorophore were synthesized and characterized. The combination of light scattering, fluorescence spectroscopy, and atomic force microscopy confirmed that for both the PEI-g-PNIPAm copolymers P3 and P4, the coil-to-globule transitions of the pendant PNIPAm chains resulted in a change in the structure of the polymers alone and when complexed to DNA. The fluorescent label proved to be a useful probe of polymer architecture, enabling sensitive detection of side-chain phase changes and variations in environment of the responsive chains as they passed through lower critical temperature transitions, both within and without complexes with plasmid DNA. AFM provided important further microstructural data demonstrating that isolated polymer-DNA complexes were compacted above polymer LCST, but that the increase in hydrophobicity led to aggregation of complexes at higher polymer/ DNA ratios most likely due to self-association of hydrophobic surfaces. Importantly, however, the results demonstrate that different architecture changes occur in P3 and P4 complexes at temperatures close to those in vivo as a consequence of their individual polymer components and phase transition temperatures and suggest that these may lead to predictable temperature variations between polymers P3 and P4 in terms of their transfection efficacies. If these architecture changes can be directed by biological stimuli such that reversibly DNA-binding structures are generated in a controlled manner, more efficient transport and release of DNA should be achieved, and ultimately, effective gene therapies will result. Acknowledgment. We thank the Biotechnology and Biological Sciences Research Council (BBSRC, Grant BB/C515855/

Variable-Architecture ThermosensitiVe Polymers

1) and Engineering and Physical Sciences Research Council (EPSRC, Grant GR/N AF/001572)) for financial support. We also thank Dr Steve Holding (RAPRA) for gel permeation chromatography and Pacawat Hongthong for repeat DLS experiments. We additionally thank reviewers of this manuscript for detailed and helpful comments.

Langmuir, Vol. 23, No. 1, 2007 49

Supporting Information Available: Turbidity temperature data, figure of agarose gel electrophoresis, and variable-temperature fluorescence emission spectra. This material is available free of charge via the Internet at http://pubs.acs.org. LA061992A