Biophysical Characterization of Complexation of DNA with Block

To 25 mL of solutions of fixed DNA (0.06%) and EB (0.03%) concentration in pH 5 ... for Pluronic L92 (open circles) and L92-pDMAEMA (open squares) sol...
0 downloads 0 Views 141KB Size
5142

Langmuir 2005, 21, 5142-5148

Biophysical Characterization of Complexation of DNA with Block Copolymers of Poly(2-dimethylaminoethyl) Methacrylate, Poly(ethylene oxide), and Poly(propylene oxide) Carmen Alvarez-Lorenzo,* Rafael Barreiro-Iglesias, and Angel Concheiro Departamento de Farmacia y Tecnologı´a Farmace´ utica, Facultad de Farmacia, Universidad de Santiago de Compostela, 15782-Santiago de Compostela, Spain

Ludmila Iourtchenko and Valery Alakhov Supratek Pharma, Incorporated, 531 Boulevard des Priries, Building 18, Laval, Que´ bec, Canada H7V 1B7

Lev Bromberg, Marina Temchenko, Smeet Deshmukh, and T. Alan Hatton Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received January 21, 2005. In Final Form: March 21, 2005 The interactions of DNA (salmon testes) with two new cationic block copolymers made of poly(2dimethylaminoethyl) methacrylate and poly(ethylene oxide), PEO-pDMAEMA, or poly(ethylene oxide)poly(propylene oxide)-poly(ethylene oxide), L92-pDMAEMA, were studied with the aim to understand their different in vitro transfection efficiencies when used as nonviral delivery vectors. PEO-pDMAEMA does not show surface activity while L92-pDMAEMA is as surface active as its parent Pluronic L92. Surface tension, titration microcalorimetry, ethidium bromide displacement, and ζ-potential measurements were carried out in phosphate buffers at pH 5 and 7. The association of L92-pDMAEMA with DNA was strongly exothermic at both pHs; the critical aggregation concentration (CAC) corresponded to a N/P ratio of 0.3, the maximum energy evolved was reached for N/P ratios of 0.82 and 1.27 at pH 5 and pH 7, respectively, and the saturation occurred for N/P ratios close to 2. The presence of L92 in the structure of this new block copolymer apparently did not modify the thermodynamic parameters of the interaction with DNA. In contrast, the interaction with PEO-pDMAEMA was significantly less exothermic, and CAC and saturation occurred for N/Ps equal to 0.43 and 1.37, respectively. The strong affinity of L92-pDMAEMA for DNA was reflected in its capacity to displace ethidium bromide and in the jump in the values of the ζ potential when N/P is near 1. Above the N/P ratio at which electroneutral polyplexes are formed, only at pH 5 an excess of L92-pDMAEMA is incorporated in the complexes, resulting in positively charged complexes. The profile of the ζ-potential values obtained for mixtures of L92-pDMAEMA with Pluronic P123 showed a shift to a lower N/P ratio, owing to an easier interaction of L92-pDMAEMA molecules with DNA in the presence of P123. Additionally, a visual inspection of the systems indicates that P123 contributes to stabilize/solubilize the DNA/cationic polymer aggregates, by avoiding the typical phase separation near the charge neutralization point. The information obtained can be particularly useful to optimize the conditions to form efficient polyplexes for gene delivery systems.

* To whom correspondence should be addressed. E-mail: [email protected]. Fax: 34-981547148.

formation reduces the exposure of DNA to enzymes, notably increasing its stability and the possibility of reaching the nucleus without degradation.6,7 Despite the high yield of complexation, some drawbacks still prevent the use of the DNA-polymer complexes (polyplexes). As a result of the interaction and the neutralization of charges, the complexes become hydrophobic, form large aggregates, and precipitate. Additionally, cationic polymers are usually highly toxic,8 and the differences between in vitro and in vivo transfection efficiency are still unclear.2 Although positively charged polyplexes are shown to be efficient in vitro, electroneutral complexes may be preferred in vivo to avoid clinging to or repulsion by cellular

(1) De Smedt, S. C.; Demeester, J.; Hennink, W. E. Pharm. Res. 2000, 17, 113-126. (2) Kircheis, R.; Wagner, E. Gene Ther. Regul. 2000, 1, 95-114. (3) Gebhart, C. L.; Kabanov, A. V. J. Controlled Release 2001, 73, 401-416. (4) Wagner, E. Pharm. Res. 2004, 21, 8-14. (5) Andersson, T.; Aseyev, V.; Tenhu, H. Biomacromolecules 2004, 5, 1853-1861.

(6) Van de Wetering, P.; Schuurmans-Nieuwenbroek, N. W. E.; Hennink, W. E.; Storm, G. J. Gene Med. 1999, 1, 156-165. (7) Nisha, C. K.; Manorama, S. V.; Ganguli, M.; Maiti, S.; Kizhakkedathu, J. N. Langmuir 2004, 20, 2386-2396. (8) Jones, R. A.; Poniris, M. H.; Wilson, M. R. J. Controlled Release 2004, 96, 379-391.

Introduction Cationic polymers are being widely studied as promising nonviral vectors to facilitate the delivery and expression of DNA into cells.1-4 Formation of polyelectrolyte complexes of DNA with polycations spontaneously occurs in an aqueous medium as a result of electrostatic attraction between the phosphate groups of DNA with the positively charged groups of the polymers. The process is entropically favorable because the counterions are released to the solution, increasing the entropy of the system.5 Complex

10.1021/la050170v CCC: $30.25 © 2005 American Chemical Society Published on Web 04/15/2005

Complexation of DNA with Block Copolymers

membranes.9 To address the challenging application of polyplexes in gene therapy, detailed biophysical characterization of the complexes is required to understand what makes the transfection efficient.1,4 Small changes in the nature, architecture, and charge density of the cationic polymers can dramatically alter the complexation process, the lifetime, the in vivo distribution, and the transfection efficacy of the complexes.2,7,10-12 Among the cationic polymers, those partially protonated at physiological pH are, at the moment, the most promising candidates.3 They can establish ionic interactions with DNA while keeping the ability to protect DNA by buffering, to some extent, the adverse acid medium of endosomal/ lysosomal vesicles,6 altering the morphology and function of these vesicles,8 or forming nuclease-resistant complexes.7 The complexation of DNA with polyethyleneimine (PEI) and poly(dimethylaminoethyl) methacrylate (pDMAEMA) has been characterized by several techniques, and transfection studies have been undertaken.6,13 Conjugation of the cationic polymer to a hydrophilic block, such as poly(ethylene glycol) (PEG) or a carbohydrate derivative, markedly improves the solubility and stability of the polyplexes, cell survival, and biodistribution.1,5,14 The hydrophilic shell at the outer surface of the complexes prevents the aggregation due to steric repulsion and reduces nonspecific interactions with other biomolecules and blood components, extending systemic circulation.7,15,16 However, PEGylation of cationic polymers can decrease transfection efficiency through undesirable interactions of the PEG segment and the cationic groups or by preventing the DNA phosphate groups to come close enough to the cationic moiety.17 The PEG proportion and the length of its chains as well as the strategy used to introduce the hydrophilic segments in the polymer or in the polyplex strongly condition its performance.16 In the particular case of block copolymers of pDMAEMA-PEG, transfection activity is low.16,18 One possible way to overcome this problem is to attach the cationic segment to the amphiphilic block PEO-PPOPEO (Pluronic).3,19 Pluronic block copolymers are recognized as pharmaceutical multipurpose excipients able to increase the solubility, stability, and circulation time or to control the release of drugs.20-24 Additionally, they can (9) Prokop, A.; Kozlov, E.; Moore, W.; Davidson, J. M. J. Pharm. Sci. 2002, 91, 67-76. (10) Ochietti, B.; Lemieux, P.; Kabanov, A. V.; Vinogradov, S.; StPierre, Y.; Alakhov, V. Gene Ther. 2002, 9, 939-945. (11) Ochietti, B.; Guerin, N.; Vinogradov, S. V.; St-Pierre, Y.; Lemieux, P.; Kabanov, A. V.; Alakhov, V. Y. J. Drug Targeting 2002, 10, 113121. (12) Funhoff, A. M.; van Nostrum, C. F.; Koning, G. A.; SchuurmansNieuwenbroek, N. M. E.; Crommelin, D. J. A.; Hennink, W. E. Biomacromolecules 2004, 5, 32-39. (13) Takeda, N.; Nakamura, E.; Yokoyama, M.; Okano, T. J. Controlled Release 2004, 95, 343-355. (14) Tseng, W.-C.; Jong, C. M. Biomacromolecules 2003, 4, 12771284. (15) Oupicki, D.; Ogris, M.; Howard, K. A.; Dash, P. R.; Ulbrich, K.; Seymour, L. W. Mol. Ther. 2002, 5, 463-472. (16) Pirotton, S.; Muller, C.; Pantoustier, N.; Botteman, F.; Collinet, S.; Grandfils, C.; Dandrifosse, G.; Dege´e, P.; Dubois, P.; Raes, M. Pharm. Res. 2004, 21, 1471-1479. (17) Rungsardthong, U.; Deshpande, M.; Bailey, L.; Vamvakaki, M.; Armes, S. P.; Garnett, M. C.; Stolnik, S. J. Controlled Release 2001, 73, 359-380. (18) Bromberg, L.; Deshmukh, S.; Temchenko, M.; Iourtchenko, L.; Alakhov, V.; Alvarez-Lorenzo, C.; Barreiro-Iglesias, R.; Concheiro, A.; Hatton, T. A. Bioconjugate Chem. 2005, in press. (19) Gebhart, C. L.; Sriadibhatla, S.; Vinogradov, S.; Lemieux, P.; Alakhov, V.; Kabanov, A. V. Bioconjugate Chem. 2002, 13, 937-944. (20) The United States Pharmacopeial Convention, Inc. United States Pharmacopoeia 25 -National Formulary 20; Rockville, MD, 2002. (21) Rowe, R. C.; Sheskey, P. J.; Weller, P. J. Handbook of Pharmaceutical Excipients, 4th ed.; Pharmaceutical Press and American Pharmaceutical Association: Bath, U.K., 2003; pp 447-450.

Langmuir, Vol. 21, No. 11, 2005 5143 Table 1. Structural Characteristics of the Copolymers copolymer

MW

EO/PO units

%N

pKa

L92a

3720 7100 6950 5750

17/52 17/52 76/0 39/69

4.15 4.75

7.1 7.4

Pluronic L92-pDMAEMAb PEO-pDMAEMAb Pluronic P123a a

Data from ref 22. b Data from ref 18.

themselves stimulate gene expression of naked DNA or that complexed with PEI.25,26 Polypexes based on Pluronic P123-g-PEI and free Pluronic P123 are physically stable, and their transfection effectiveness is similar to that of PEI.3,19 Other structurally similar block polymers, Tetronic, composed of poly(propylene oxide) (PPO), PEO, and ternary amino moieties, behave as efficient gene delivery carriers into subcutaneous tissues.9 Pluronic-based polyplexes are expected to be formed by a core of cationic units that interact with the DNA, probably stabilized by hydrophobic interactions with PPO, and a shell of PEO that prevents interaction with serum components and enzymes.19 In a previous paper, Pluronic L92-pDMAEMA, prepared by a new synthetic one-step route, was demonstrated to possess a strong transfection efficiency, comparable to that of the best available polyplexes and lipid-based systems.18 The aim of this work was to carry out a biophysical characterization of DNA complexation with this new copolymer of Pluronic L92-pDMAEMA and to compare its behavior with that of Pluronic L92 alone or PEO-pDMAEMA. The results obtained should help the mechanism of the transfecting activity of the new copolymer to be understood and provide useful information to optimize its structure with a view to using it as a nonviral gene delivery system. Materials and Methods Materials. Deoxyribonucleic acid, sodium salt (DNA) from Salmon testes, was from Sigma-Aldrich Co. (St. Louis, MO). Ethidium bromide (EB) solution (1% in water) was supplied by Fluka (Switzerland). Pluronic L92 (PEO8PPO52PEO8) and P123 (PEO19PPO69PEO19) were received from BASF Corp. (Mount Olive, NJ). L92-pDMAEMA and POE-pDMAEMA were synthesized as previously described.18 The most relevant structural characteristics are summarized in Table 1. Purified water by reverse osmosis (Milli-Q, Millipore, Spain) with a resistivity above 18.2 MΩ‚cm was used. All other chemicals were of analytical grade and used as received. Characterization of Copolymer Solutions. Solutions of L92, L92-pDMAEMA, or PEO-pDMAEMA in water or in phosphate buffers at pH 5 or 7 (prepared as indicated in USP25, with a final concentration in monobasic potassium phosphate of 0.05 M)20 were obtained by dispersing adequate amounts of copolymer powder in the medium under stirring and left to rest for 24 h before measurements at 298.0 K. The following properties were characterized: (a) pH, using a model GLP22 Crison pH meter (Barcelona, Spain), equipped with a silver/silver chloride no. 52-21 electrode; (b) apparent viscosity, using Cannon-Fenske capillary viscometers (Afora, Spain); (c) surface tension, by the platinum ring method using a Lauda Tensiometer TD1 (LaudaKo¨nigshofen, Germany); and (d) conductivity, using a conductivity meter (model CDM2e, Radiometer, Copenhagen, Denmark) equipped with a Crison platinum sensor (Barcelona, Spain). (22) Kabanov, A. V.; Batrakova, E. V.; Alakhov, V. J. Controlled Release 2002, 82, 189-212. (23) Barreiro-Iglesias, R.; Alvarez-Lorenzo, C.; Concheiro, A. Int. J. Pharm. 2003, 258, 165-177. (24) Barreiro-Iglesias, R.; Bromberg, L.; Temchenko, M.; Hatton, T. A.; Alvarez-Lorenzo, C.; Concheiro, A. J. Controlled Release 2004, 97, 537-549. (25) Lemieux, P.; Guerin, N.; Paradis, G.; Proulx, R.; Chistyakova, L.; Kabanov, A.; Alakhov, V. Gene Ther. 2000, 7, 986-991. (26) Kuo, J. H. S. Biotechnol. Appl. Biochem. 2003, 37, 267-271.

5144

Langmuir, Vol. 21, No. 11, 2005

Characterization of Copolymer/DNA Interactions. Surface Tension. Solutions of L92, L92-pDMAEMA, or PEOpDMAEMA in phosphate buffers at pH 7 with 0.01% DNA were prepared by mixing adequate volumes of copolymer solution (8%, w/v) and DNA (0.02%, w/v) solution. Surface tension was estimated by the platinum ring method using a Lauda Tensiometer TD1 (Lauda-Ko¨nigshofen, Germany). Titration Microcalorimetry. Calorimetric experiments were performed in duplicate using a Tronac-450 isoperibol microcalorimeter and Tronac FS101 calorimetry software (Tronac, Inc., Orem, UT). In each experiment, a 47.5 mL DNA solution with a concentration of 0.03% (w/w) was placed in a Dewar reaction vessel, and a relatively concentrated copolymer solution (4%) was loaded into a 2 mL calibrated buret. The entire assembly was then immersed into a constant temperature water bath (298.0 K). After thermal equilibration, the copolymer solution was delivered at a constant rate of 0.3332 mL/min into the reaction vessel, in which a stirrer mixed the two solutions rapidly. The rise or decrease in the temperature of the system was monitored using a thermistor and was later reproduced using a heating coil in the reaction vessel. The apparent enthalpy was calculated from the applied current and voltage and the heating time. As a blank, a buffer solution was used instead of DNA dispersion. Calibration of the system was assured by titration of tris(hydroxymethyl)aminomethane with HCl. The integral binding heat for the polymer/DNA interaction (Qint) process was estimated by subtracting from the measured heat produced by addition of the copolymer to the DNA dispersion (Qp) and the heat effects due to the dilution/demicellization of the copolymer in the buffer solutions used as a blank (Qd).23

Qint ) Qp - Qd The enthalpy of DNA dilution (final concentration 3% lower) was negligible. Similar experiments were carried out with L92pDMAEMA/Pluronic P123 (1:1 and 1:5, w/w) mixtures (in the buret), keeping the concentration of L92-pDMAEMA constant at 4%. EB Displacement Assay. To 25 mL of solutions of fixed DNA (0.06%) and EB (0.03%) concentration in pH 5 or pH 7 phosphate buffer were added adequate volumes of cationic polymer solution to cover a wide range of N/P ratios (expressed via the ratio of equivalents of DMAEMA units to the number of nucleotides in DNA). The systems were diluted with buffer medium up to 50 mL. The final concentrations of cationic polymer ranged from 0.0009 to 0.1%. The samples were stirred overnight, protected from light, and stored at 20 °C. EB emission spectra were recorded in a Perkin-Elmer LS50B spectrofluorometer (Buckinghamshire, U.K.). The excitation and emission wavelengths were 545 and 596 nm. As a blank, DNA (0.03%) and EB (0.015%) solutions in each medium were used. The DNA/EB ratio was chosen to have approximately two EB molecules per five base pairs.27 An average molecular weight per nucleotide of 324.5 g/mol PO4 was used for the calculations of N/P ratios.28 ζ Potential. Samples were prepared by mixing DNA solutions with L92-pDMAEMA, its mixtures with Pluronic P123, or POEpDMAEMA solutions in phosphate buffers (pH 5 and 7), which were previously filtered through 0.2 µm nylon filters (Millipore). The DNA concentration in the samples was held constant at 0.01% whereas the concentration of cationic copolymer was varied to cover a wide range of N/P ratios. The experiments were carried out in a Zetamaster 5002 particle electrophoresis analyzer (Malvern Instruments, Ltd., U.K.) equipped with a He-Ne laser (633 nm, 5 mW) and using a standard electrophoresis cell of quartz (5 mm × 2 mm).

Results and Discussion This section begins with a general characterization of the fundamental properties of the cationic copolymers to (27) Bronich, T. K.; Nguyen, H. K.; Eisenberg, A.; Kabanov, A. V. J. Am. Chem. Soc. 2000, 122, 8339-8343. (28) Lobo, B. A.; Davis, A.; Koe, G.; Smith, J. G.; Middaugh, C. R. Arch. Biochem. Biophys. 2001, 386, 95-105.

Alvarez-Lorenzo et al.

Figure 1. Concentration dependencies of surface tension and conductivity of Pluronic L92 in water (open triangles) and L92pDMAEMA solutions in water (closed triangles) and in phosphate buffer at pH 7 (closed squares). Surface tension plots also show the values obtained in the presence of DNA (0.01%) for Pluronic L92 (open circles) and L92-pDMAEMA (open squares) solutions in phosphate buffer at pH 7. In the conductivity plot, the Y scale on the left corresponds to the L92-pDMAEMA solutions and on the right to the Pluronic L92 solutions.

obtain information useful to explain the interactions with DNA that are discussed in the second subsection. Characterization of the Copolymer Solutions. Similar to its parent Pluronic L92, the copolymer L92pDMAEMA is a surface-active compound in aqueous solutions. In contrast, PEO-pDMAEMA, which has a significantly greater hydrophilic character, had little effect on the surface tension values of the medium. Figure 1 depicts concentration dependencies of the equilibrium surface tension and conductivity of the copolymer solutions in water (∼pH 4) and in phosphate buffer (pH 7). A change in slope was observed in the surface tension curve at a specific copolymer concentration, after which the values decreased more gradually. The critical micelle concentration (CMC) value observed for L92-pDMAEMA around 4 × 10-3 g/mL was similar to that found for Pluronic L92 (2 × 10-3 g/mL) at the same temperature (298.0 K). The obtained CMC of the Pluronic L92 corresponded well with the literature data.22,29 In the case of PEO-pDMAEMA solutions, only those at pH 7 showed an apparent CMC at around 0.01 g/mL.18 Owing to the ionizable character of the amino groups of the pDMAEMA copolymers, the higher the ionization of the tertiary amino groups, the lower the surface activity. This is due to the ionization leading to a higher solubility and lower concentration at the hydrophobic water-air interface. Despite the greater conductivity of the L92-pDMAEMA solutions compared to those of L92 (owing to the cationic character of the former), the plots were superimposable, which confirms the similarity of their CMC values. Within the range of concentrations studied, the apparent viscosity of the solutions was between 1 and 8 mPa‚s, which is (29) Kozlov, M. Y.; Melik-Nubarov, N. S.; Batrakova, E. V.; Kabanov, A. V. Macromolecules 2000, 33, 3305-3313.

Complexation of DNA with Block Copolymers

small enough to ensure a good syringeability if they were parenterally administered.30 Characterization of Copolymer/DNA Interactions. Surface Tension. The presence of DNA did not modify the surface tension profiles of Pluronic L92 or PEO-pDMAEMA solutions (the values were around 60 mN/m in the whole range) but slightly decreased the surface tension values of L92-pDMAEMA solutions (Figure 1). Vinogradov et al.31 found that PEI-PEG and its complex with DNA behaves as a surfactant with a low CMC and related this fact to the greater stability of DNA against nuclease degradation in the presence of PEI-PEG. The fact that the self-assembly of L92-pDMAEMA is promoted in the presence of DNA makes these systems particularly attractive as nonviral vectors. Titration Microcalorimetry. Figure 2 shows the apparent enthalpies associated with the dilution and/or demicellization processes of Pluronic L92 (a), L92-pDMAEMA (b, c), and PEO-pDMAEMA (d) in phosphate buffer and those of the interaction processes with DNA. The concentration of the polymers (0.04 g/mL) in the buret, before being added to the Dewar, was sufficient to be above their CMC, if any. Therefore, when the polymer solution was slowly added to an aqueous medium without DNA, the micelles of Pluronic L92 and L92-pDMAEMA broke up until the concentration in the Dewar reached the CMC. For the Pluronic L92, the demicellization process was exothermic (negative enthalpy change), while for the L92pDMAEMA an endothermic process was observed. The exothermic dilution process of Pluronic solutions has been attributed to the hydrogen bond formation between the triblock copolymer and water molecules after breakage of water-water and surfactant-surfactant hydrogen bonds.32 The apparent enthalpy change observed for the demicellization of Pluronic L92 (∆Hd ) -198 J/mmol) was significantly lower than that obtained for other more hydrophilic Pluronics, such as F127 (∆Hd ) -410 J/mmol).23 The chemical linking of pDMAEMA to L92 significantly changes the thermodynamic parameters of the demicellization, the process being endothermic but entropically favorable, as for many other typical ionic surfactants.33 No significant interaction with DNA was observed for Pluronic L92 (Table 2), for which the demicellization process was not altered by the presence of DNA. In contrast, the association of L92-pDMAEMA with DNA was extraordinarily intense at both pH 5 and 7 (Figure 2b,c). When the L92-pDMAEMA solution was added to the DNA solution, an initial endothermic demicellization process was observed. However, this process stopped when the L92-pDMAEMA concentration in the Dewar reached 1.1 × 10-4 g/mL, and a new exothermic process began. This concentration can be identified as the critical associative concentration (CAC) and corresponds to a N/P ratio of 0.3, being similar for both pHs evaluated. The peak in enthalpy was observed when the L92-pDMAEMA concentration reached 2.4 × 10-4 g/mL at pH 5 or 3.7 × 10-4 g/mL at pH 7, which correspond to N/P ratios of 0.82 and 1.27, respectively. Therefore, the maximum in heat evolved is achieved at a cationic/anionic molar ratio close to 1. The effect of pH is related to the lower ionization (30) Allahham, A.; Stewart, P.; Marriott, J.; Mainwaring, D. E. Int. J. Pharm. 2004, 270, 139-148. (31) Vinogradov, S. V.; Bronich, T. K.; Kabanov, A. V. Bioconjugate Chem. 1998, 9, 805-812. (32) Irwin, J. J.; Beezer, A. E.; Mitchell, J. C.; Buckton, M. G.; Chowdhry, B. Z.; Eagland, D.; Crowther, N. J. J. Phys. Chem. 1993, 97, 2034. (33) Barreiro-Iglesias, R.; Alvarez-Lorenzo, C.; Concheiro, A. Int. J. Pharm. 2003, 258, 179-191.

Langmuir, Vol. 21, No. 11, 2005 5145

Figure 2. Calorimetric titration curves observed during the addition of 4% copolymer solutions into a dewar containing just phosphate buffer (open symbols) or a 0.03% DNA solution in phosphate buffer (closed symbols). The polymer/DNA interaction heat, Qint, was estimated as the difference between the heat evolved in the presence and that in the absence of DNA (continuous lines). (a) Pluronic L92, pH 5; (b) L92pDMAEMA, pH 5; (c) L92-pDMAEMA, pH 7; (d) PEOpDMAEMA, pH 5.

degree of pDMAEMA groups at pH 7. The process of interaction was exothermic up to 6 × 10-4 g/mL of cationic polymer in the Dewar. This finding suggests that the saturation of the binding was reached at a N/P ratio close to 2; that is, nonstoichiometric polyplexes could be formed. In previous studies with other cationic molecules, formation of polyplexes and lipoplexes with a N/P ratio above 1 was observed.7,28 In the case of PEO-pDMAEMA, the dilution process in pH 5 phosphate buffer (Figure 2d) was exothermic but did not show the characteristic profile of a demicellization. The interaction with DNA was significantly more exothermic, although the energy evolved in the process was notably lower than for L92-pDMAEMA. The CAC can be

5146

Langmuir, Vol. 21, No. 11, 2005

Alvarez-Lorenzo et al.

Table 2. Thermodynamic Parameters of the Interaction of the Copolymers with DNA at 298 Ka copolymer Pluronic L92 L92-pDMAEMA PEO-pDMAEMA P123 P123/L92-pDMAEMA (1:1) P123/L92-pDMAEMA (5:1) a Values within (5%. mixture.

b

pH

∆Hint (J/mmol)

5 7 5 7 5 5 5 5

-3.4 +6.4 -165.4 -205.5 -31.1 +45.0 -40.7 (-131.8)b -35.1 (-237.3)b

J/mmol of L92-pDMAEMA in the

established at about 1.13 × 10-4 g/mL, while the interaction process was completed when the polymer concentration was 3.63 × 10-4 g/mL. Considering that the polymer contains a 4.75% N, these values correspond to N/P ratios of 0.43 and 1.37, respectively. The greater CAC, lower saturation concentration, and lower interaction energy observed with POE-pDMAEMA, compared to those of L92pDMAEMA, clearly indicate a lower affinity for DNA. The contents of the Dewar after the experiments were cloudy in the case of L92-pDMAEMA titrations. The enthalpies of interaction (Table 2) were calculated considering the average energy involved in the process up to theoretical neutralization of charges. Taking into account that each chain of copolymer contains approximately 23 units of DMAEMA,18 the enthalpies of interaction (Table 2) expressed per mol of cationic unit can be estimated as -7.2 kJ/mol and -8.9 kJ/mol for L92pDMAEMA at pH 5 and at pH 7, respectively, and -1.3 kJ/mol for PEO-pDMAEMA at pH 5. Therefore, the microenvironment of the cationic DMAEMA units, provided by the nature and structure of polymer to which they are attached, plays an important role in the binding process. Although L92 itself shows a low affinity for DNA, both molecules can act as both hydrogen-bond donor and acceptor and, also, can interact hydrophobically.9,34 Neutral amphiphilic polymers, such as poly(vinylpyrrolidone) (PVP) or Tetronics (tetrafunctional block copolymers of PEO-PPO bound to ethylenediamine), can form polyplexes with DNA via hydrogen bonding as well as hydrophobic interactions.34 This kind of association is shown as an endothermic process in which the entropic contribution occurs via release of water and/or counterions from both PVP and DNA upon their mixing.34 Binding of DNA (both plasmid and total DNA from salmon testes) to cationic lipids was found to be endothermic or exothermic depending on the composition of the lipids and pH of the medium.35,28 Rungsardthong et al.17 observed that the interaction of DNA with pDMAEMA was exothermic and that the enthalpy increased from pH 4 to pH 7.4. This effect is attributed to the proton abstraction from the buffer by the nonionized amino groups of the polymer to interact with the anionic phosphate groups of DNA (i.e., there is a proton transfer from the hydrogen phosphate of the buffer to the amino groups of pDMAEMA). Protonation changes have been also observed for PEI during complex formation with DNA.36 Similarly, L92-pDMAEMA showed a greater binding enthalpy at pH 7 than at pH 5. The observed interaction enthalpy at pH 7 is coincident with (34) Mumper, R. J.; Duguid, J. G.; Anwer, K.; Barron, M. K.; Nitta, H.; Rolland, A. P. Pharm. Res. 1996, 13, 701-709. (35) Barreleiro, P. C. A.; Olofsson, G.; Alexandridis, P. J. Phys. Chem. B 2000, 104, 7795-7802. (36) Choosakoonkriang, S.; Lobo, B. A.; Koe, G. S.; Koe, J. G.; Middaugh, C. R. J. Pharm. Sci. 2003, 92, 1710-1722.

Figure 3. Schematic drawing of the association of DNA with L92-pDMAEMA. L92-pDMAEMA keeps its surface activity even after DNA binding. The amphiphilic character prompts the formation of polyplexes with a core (PPO and neutralized complexes) and shell (PEO and non-neutralized pDMAEMA groups) structure.

the value reported by Rungsardthong et al.37 for pure pDMAEMA at pH 6.6 in phosphate buffer (-8.9 kJ/mol). Therefore, it appears that the bonding of L92 did not modify the thermodynamic parameters of the copolymer interaction with DNA significantly. Nevertheless, because the data reported for pDMAEMA and those obtained with L92-pDMAEMA were not recorded under exactly the same pH (6.6 vs 7.0), the possibility that L92 can strengthen the association through hydrogen bonding and hydrophobic interactions cannot be disregarded. In contrast, the process was significantly less exothermic for PEOpDMAEMA. In this sense, Nisha et al.7 have recently observed that the binding enthalpy of DNA to copolymers made of methoxy[poly(ethylene glycol)] monomethacrylate and [3-(methacryloylamino)propyl] trimethylammonium chloride decreased proportionally to the content in PEG in these copolymers. The enthalpy per cationic unit was measured to be about -2.5 kJ/mol for the most hydrophilic copolymer (we obtained an enthalpy even lower for PEOpDMAEMA, -1.3 kJ/mol). Compared to the L92-pDMAEMA (pKa 7.1), the PEO-pDMAEMA (pKa 7.4) is more hydrophilic and has a greater ionization degree at pH 5. Thus, contribution of the ionic component to the interactions of the PEO-pDMAEMA with DNA (exothermic) should be considerably greater than the other weak noncovalent interactions such as hydrogen bonding and hydrophobic (endothermic). However, the extended PEO chains may impede the ionic groups of the two macromolecules from approaching each other, as close as in the case of L92-pDMAEMA.7 Figure 3 is a schematic drawing of the changes in conformation of L92-pDMAEMA molecules (below the CMC) when DNA is added to the solution. The amphiphilic character of L92-pDMAEMA facilitates the formation of polyplexes with a core-shell structure in which PPO comprises the cores and neutralized DNApDMAEMA complexes and the shells are composed of extended PEO chains and some ionized pDMAEMA groups. In the case of PEO-pDMAEMA, the interaction with DNA is less energetically favorable owing to the more hydrophilic character of the copolymer. Polyplexes with compact, neutralized cores and extended PEO chains forming a shell have been reported.38 Because the transfection efficiency of cationic copolymers of Pluronics has been shown to be improved by Pluronic P123,3,18,19 the next step was to evaluate how the presence of Pluronic P123 could modify the energetic aspects of the interaction with DNA. Previously, we have shown that the addition of P123 to the L92-pDMAEMA and DNA complexes diminished the size of the complex aggregates and masked the charges.18 We, thus, assayed mixtures of P123/L92-pDMAEMA at weight ratios of 1:1 (37) Rungsardthong, U.; Ehtezazi, T.; Bailey, L.; Armes, S. P.; Garnett, M. C.; Stolnik, S. Biomacromolecules 2003, 4, 683-690. (38) Guo, Y.; Sun, Y.; Li, G.; Xu, Y. Mol. Pharm. 2004, 1, 477-482.

Complexation of DNA with Block Copolymers

Langmuir, Vol. 21, No. 11, 2005 5147

Figure 5. Relative fluorescence of DNA-EB solutions with increasing proportions of Pluronic L92 (open circles), PEOpDMAEMA at pH 5 (down triangles), and L92-pDMAEMA at pH 5 (solid squares) and pH 7 (open squares).

Figure 4. (a) Demicellization energy recorded for L92pDMAEMA (4%) + P123 (4%) solution (open circles) and L92pDMAEMA (4%) + P123 (20%) solution (open squares) compared to the values theoretically predicted (continuous lines, Table 2). Calorimetric titration curves observed during the addition of (b) a L92-pDMAEMA (4%) + P123 (4%) solution or (c) a L92-pDMAEMA (4%) + P123 (20%) solution into a Dewar containing just pH 5 phosphate buffer (open symbols) or a 0.03% DNA solution in pH 5 phosphate buffer (closed symbols). The polymer/DNA interaction heat, Qint, was estimated as the difference between the heat evolved in the presence and absence of DNA (continuous lines). The enthalpy values are referred to the molar amount of L92-pDMAEMA added to the Dewar.

and 5:1. The demicellization processes of Pluronic P123 and its mixtures with L92-pDMAEMA were also characterized. Figure 4 shows the energy recorded in the calorimeter during the addition of concentrated P123 and P123/L92-pDMAEMA solutions to phosphate buffer at pH 5. The demicellization of P123 alone was strongly exothermic (∆Hd ) -150 J/mmol), similarly to the analogous process in L92 solutions. The demicellization of the mixtures P123/L92-pDMAEMA was even more exothermic, being above the energy expected assuming an additive contribution of the demicellization heat of both components (continuous line). This observation suggests formation of mixed micelles of P123 and L92-pDMAEMA.39 Compared to the demicellization process, when P123 solution (0.04 or 0.20 g/mL) was added to the DNA solution, an endothermic event was observed (∆Hint ∼ +45.0 J/mmol; Table 2), which suggests an entropy-driven association owing to the gain in entropy of water released to the bulk as the hydrophobic part of P123 interacts with (39) Liu, T.; Nace, V. N.; Chu, B. Langmuir 1999, 15, 3109-3117.

hydrophobic regions of DNA, displacing the water bound to the P123. This phenomenon may be more intense than in the case of L92 because of the greater molecular weight and the more even content of both PPO and PEO groups (i.e., hydrogen-bond donors and acceptors).9,34 The calorimetric profiles of the interaction of the P123/ L92-pDMAEMA with DNA (Figure 4) were similar to those recorded for the L92-pDMAEMA solutions without P123 added. However, the initial endothermic process was not observed, which indicates that the interaction with DNA began at slightly lower cationic polymer concentrations, especially in the P123/L92-pDMAEMA (1:1) system. The energy of interaction with DNA (Qint) was lower for the 1:1 copolymer mixture (-0.12 J) than for the 5:1 one (-0.36 J). When referring the energy to the moles of L92pDMAEMA in the mixture, the enthalpy values were in the range of those obtained with the L92-pDMAEMA alone system (Table 2). The maximum in heat evolved was achieved for approximately a N/P ratio equal to 1. EB Displacement and ζ Potential. These two techniques were chosen to elucidate the practical consequences of the different affinities for DNA of the copolymers observed by calorimetry. The ability of copolymers to effectively bind to DNA was monitored by recording the fluorescence of DNA-EB complexes.13,40 Intercalation of EB into the DNA is accompanied with an increase in the fluorescence of this probe.27 The addition of a polymer able to displace EB from the DNA-EB complex results in quenching of the fluorescence. As can be observed in Figure 5, the presence of L92 almost did not modify the fluorescence of EB; the results were similar for both pHs. In contrast, L92-pDMAEMA caused the fluorescence to decrease. The degree of quenching depended on the N/P ratio. Above 1 × 10-4 g/mL of cationic copolymer, the fluorescence decreased markedly when more copolymer was added. This result is in agreement with the CAC observed by calorimetric titration (N/P ratio of about 0.3). L92pDMAEMA quenched the EB fluorescence completely when the N/P ratio was around 1. However, as observed in the calorimetric experiments, a slightly greater concentration of L92-pDMAEMA is required at pH 7 than at pH 5 for a similar displacement of EB, which is a consequence of the lower ionization degree of the copolymer at pH 7.37 The lower affinity of POE-pDMAEMA for DNA was reflected in the copolymer’s weak displacement of EB; despite being almost completely ionized at pH 5, the profile of fluorescence intensity was similar to that of L92. These observations are in accordance with the ζ-potential values. To carry out these experiments, the samples (40) Veron, L.; Gane´e, A.; Charreyre, M. T.; Pichot, C.; Delair, T. Macromol. Biosci. 2004, 4, 431-444.

5148

Langmuir, Vol. 21, No. 11, 2005

Alvarez-Lorenzo et al.

the polyplexes have a smaller particle size when P123 is present, which also provides a more hydrophilic shell that prevents aggregation.18 It is possible that P123 can stabilize polyplexes by depletion flocculation. As expected from the calorimetric and EB studies, no significant changes in the ζ potential of DNA were observed after adding quantities of POE-pDMAEMA up to an N/P ratio of 11 (Figure 6, down triangles), which clearly indicates a low binding affinity. The solutions were clear at any N/P ratio, as reported for other POE-grafted cationic polymers.5,27 Figure 6. Variation of the ζ potential as a function of the charge ratio in the polyplexes, for L92-pDMAEMA at pH 5 (solid squares) and pH 7 (open squares), P123/L92-pDMAEMA (1:1) at pH 5 (up triangles), and PEO-pDMAEMA at pH 5 (down triangles).

were prepared as indicated above, but the concentration of DNA and copolymer was 3 times lower to make possible a precise measurement. As can be seen in the Figure 6, there is a jump in the values of the ζ potential of L92pDMAEMA polyplexes when N/P is near 1. The curve is shifted to slightly greater N/P at pH 7 compared to pH 5. Because the titration calorimetry data shows a strong binding affinity, it is expected that above a N/P ratio of 1, the proportion of free DNA in the samples containing L92-pDMAEMA may be small enough to avoid any interference with the measurement of the ζ potential. It is also interesting to note that, above the N/P ratio at which electroneutral polyplexes are formed, only at pH 5 an excess of L92-pDMAEMA is incorporated in the complexes, resulting in positively charged complexes. The ζ-potential values obtained for the P123/L92-pDMAEMA mixtures (Figure 6, up triangles) were similar to those recorded for L92-pDMAEMA. Only a shift to lower polymer concentrations (lower N/P) was observed, which may be related to an easier interaction of L92-pDMAEMA molecules with DNA in the presence of P123. Additionally, a visual inspection of the systems indicates that the presence of P123 contributes to stabilize/solubilize the DNA/cationic polymer aggregates, by avoiding the typical phase separation near the charge neutralization point. This ability of P123/L92-pDMAEMA mixtures to provide more homogeneous systems can be related to the fact that

Conclusions The new block copolymer L92-pDMAEMA combines the surface activity of Pluronic L92 and the strong affinity for DNA of pDMAEMA. The self-assembly of L92-pDMAEMA is promoted in the presence of DNA. Judging by the lower CAC and greater saturation concentration and interaction energy (e.g., -7.2 kJ/mol vs -1.3 kJ/mol at pH 5), the L92-pDMAEMA copolymer is more prone to forming polyplexes with DNA than its more hydrophilic counterpart, PEO-pDMAEMA copolymer. Therefore, the microenvironment of the cationic DMAEMA units, provided by the nature and structure of the polymer to which they are attached, plays an important role in the DNA binding process. Because the ionization of L92-pDMAEMA is pHdependent, the copolymer affinity for DNA can be modulated in the physiological range, the interaction being stronger at pH 5 than at pH 7. The presence of P123 promotes the interaction process and enhances the solubility of the polyplexes. In the polyplexes, the neutral pDMAEMA-DNA complexes and PPO segments may form a core, the PEO and protonized pDMAEMA segments being oriented toward the aqueous interface. These results explain the high in vitro efficiency of L92-pDMAEMAbased polyplexes, compared to the PEO-pDMAEMA ones, as nonviral gene vectors. Acknowledgment. This work was financed by the Xunta de Galicia (PGIDT 02BTF20302PR) and Ministerio de Ciencia y Tecnologı´a, Spain (RYC2001-8). L.B. and T.A.H. were supported, in part, by the Singapore-MIT alliance and by the Cambridge-MIT Institute. LA050170V