ARTICLE pubs.acs.org/Langmuir
Gel Network Photodisruption: A New Strategy for the Codelivery of Plasmid DNA and Drugs Diana Costa,*,† Artur J. M. Valente,‡ M. Grac-a Miguel,‡ and Jo~ao Queiroz† † ‡
CICS - Centro de Investigac-~ao em Ci^encias da Saude, Universidade da Beira Interior, 6201-001 Covilh~a, Portugal Department of Chemistry, University of Coimbra, Coimbra, Portugal ABSTRACT: In the last 5 years, we have gained further insight on the physical/chemical field of DNA gels. Our expertise on the gel swelling behavior, compaction of DNA by cationic entities, as lipids and surfactants, as well as on the assembly structures of these complexes allow us for the development of novel systems to be used in a variety of biomedical applications. In our previous reports, the physicochemical characterization has been well-established, and now one can evolve to the challenge of using DNA-based carriers in the biological area. Moreover, a new plasmid DNA (pDNA) hydrogel that is porous, is able to swell in the presence of additives, is biocompatible and, thus, is suitable to be used therapeutically was prepared. Here, the dual release of pDNA and solutes with pharmaceutical interest was the main challenge, and thus, we report on the photodisruption of plasmid DNA (pDNA) gels cross-linked with ethylene glycol diglycidyl ether (EGDE) as a strategy for this simultaneous release. The disruption over time, after the irradiation of the gel with ultraviolet light (400 nm), was characterized through the cumulative plasmid DNA release, the evolution in dry weight, the extent of swelling, and also the variations in the gel mesh size. The controlled release of different molecular weight solutes from plasmid DNA gels was investigated, and the influence of both the hydrogel degradation and cross-linker density on the release kinetics were addressed. While the release of lysozyme follows a Fickian process, the release of bovine serum albumin (BSA) and fluoresceinisothiocyanato-dextran (FITC-dextran) is characteristic of a Super Case II release phenomena. In addition, the size of the three solutes partially influences the release behavior; polymer chain mobility and the degree of swelling also play a role. To gain a fundamental understanding of drug release profile from pDNA matrices, in vitro release studies were evaluated using several anti-inflammatory drugs. The quantification of the release mechanism indicates a Super Case II release profile, which can be related with the gel swelling degree. A correlation between the drug release trend and the drug hydrophobicity can be found, with more hydrophobic drugs showing a slower release rate. In brief, this new pDNA gel system is biocompatible, is degradable upon light irradiation, and allows for the controlled and sustained release of plasmid DNA and incorporated solutes. This codelivery of pDNA and drugs would find relevant clinical uses due to the possibility of gene and nongene therapy combination in order to improve the therapeutic efficiency.
’ INTRODUCTION Hydrogels can be used in the biomedical field due to their biocompatibility with tissue and blood.1,2 The presence of water in the gel makes it soft and rubbery, which offers the least frictional irritation and provides a soothing effect when in contact with the physiological system. Nowadays, hydrogels play a crucial role in biomaterials research, and with increasing needs for controlled molecule release, biosensors, cell scaffolds, organs, and tissues, their development will certainly continue to grow in the near future. Novel approaches in hydrogel design have revitalized the medical area. Synthetic hydrogels such as proteinbased networks, hybrid gels containing protein domains,3,4 DNA-based gels,513 and hydrogels for tissue engineering are just a few advances in this field. The tissue engineering field, for example, has developed to meet the tremendous need for organs and tissues.14,15 The future of pharmaceutical and drug delivery has many challenges, among which the delivery of intact drugs into specific cell types in the human body is far from the least. Not only would protein drugs benefit from a targeted delivery to their site of action, but also highly toxic ones such as anticancer r 2011 American Chemical Society
drugs would benefit. Concerning drug diffusion, polymer networks show interesting properties, such as ionization of the gel, the extent of swelling, and specific mesh size. Mesh size is useful in determining the physical properties of the hydrogels including mechanical strength, degradability, and diffusion of the releasing solute.16,17 Mathematical models are important tools in promoting the understanding of material properties, interaction parameters, kinetic events, and transport behavior within complex hydrogel systems, clarifying the key parameters and mechanisms that govern the rate and extent of drug release.1820 In order to design appropriate delivery systems, the mesh size of swollen hydrogels must be tailored, leading to desired rates of solute diffusion or, instead, the degree of swelling or polymer degradation can be tailored to control the release of drugs smaller than the gel mesh size. Researchers have engineered the chemical and physical properties of hydrogels to optimize the efficacy of their Received: July 11, 2011 Revised: September 12, 2011 Published: September 21, 2011 13780
dx.doi.org/10.1021/la2026285 | Langmuir 2011, 27, 13780–13789
Langmuir use in controlled drug delivery applications.21,22 Development of temperature-responsive hydrogels,23,24 pH-responsive networks,25 glucose-responsive hydrogels,26,27 and control of hydrogel swelling behavior28 were conceived to improve the field of sustained/controlled drug delivery. For most biomedical demands, degradable hydrogels are preferred over nondegradable ones because they degrade in clinically relevant time-scales under relatively mild conditions and eliminate the need for additional surgeries to remove the implanted systems. There are several studies in the literature reporting the controlled release of drugs from biodegradable hydrogels, such as the release of recombinant human interleukin-2 from dextran-based hydrogels;29 similar dextran hydrogels were investigated as drug delivery devices for the posterior part of the eye.30 Moreover, the polymer hyaluronic acid (HA) was used to create a combination tissue engineering scaffold and protein delivery vehicle.31 Recently, in situ forming biodegradable poly(ethylene glycol)-based hydrogels were created for the time-controlled release of tethered peptides or proteins,32 and injectable biodegradable hydrogels were studied as vehicles for the release of anticancer drugs.33 Biodegradable hydrogels appear to be a more challenging strategy for DNA release, once one can readily control the release rate by modulating the network structure with adjusting crosslinking density.12 In a previous work, we prepared plasmid DNA gels by a cross-linking reaction with ethylene glycol diglycidyl ether (EGDE).34 The pDNA gels have been investigated with respect to their swelling in aqueous solution containing different additives, such as metal ions, polyamines, polycations, and surfactants. The mesh size was found to be 39 and 32 Å for pDNA gels cross-linked with 0.2% and 0.5% EGDE, respectively. Moreover, cell viability assays suggested that the plasmid DNA gels are nontoxic to cells and this step contributes to the possibility of using them as carriers in real biological systems.34 Since the future of drug and gene delivery holds for the development of new biobased vehicles and strategies, in this work we studied the capacity of plasmid DNA gels (0.2% and 0.5% EGDE) to degrade, due to the photodegradation of the cross-linker molecule, and pDNA released was quantified. In the current report, we also monitored the weight loss of the hydrogels, the degree of swelling, and the increase in the mesh size as a function of disruption time. To demonstrate the ability of these gels to load and release drugs, the controlled release of different molecular weight solutes, such as bovine serum albumin, fluoresceinisothiocyanato-dextran, lysozyme, and some anti-inflammatory drugs, was characterized using appropriate release models. Therefore, in this work, we clarified the fundamental and basic aspects of the solute release mechanism from pDNA hydrogels, and the significance of this information is enormous as a basic tool for the development of pDNA carriers to be used therapeutically. The most relevant aspect of the present report is the possibility of simultaneous controlled release of plasmid DNA and drugs. The acquired knowledge on this dual carrier is crucial as a model for progresses in the design of systems for the codelivery of gene and anticancer drugs, combining chemical and gene therapies in the treatment of cancer. This issue has attracted a lot of attention in recent years, and it is already under investigation by our group.
’ MATERIALS AND METHODS Materials. Ethylene glycol diglycidyl ether (EGDE) was from Aldrich. N,N,N0 ,N0 -tetramethylethylenediamine (TEMED), sodium
ARTICLE
hydroxide (NaOH), and sodium bromide (NaBr) were obtained from Sigma. A bicinchoninic acid (BCA) kit, bovine serum albumin (BSA), fluoresceinisothiocyanato-dextran (FITC-dextran), and lysozyme were obtained from Sigma. Lysozyme has a molecular weight of 14 100 Da and a Stokes radius of ca. 16 Å, BSA has a molecular weight of 65 000 Da and the radius is around 34.8 Å, while FITC-dextran has a molecular weight of 77 000 Da and the radius is around 55.0 Å.35 Hydrocortisone, cortisone, dexamethasone, prednisolone, and prednisone were obtained from Sigma. Plasmid Production and Recovery. The 2.7-kpb plasmid pUC19, from Invitrogen (Carlsband, CA), was hosted and amplified in Escherichia coli (E. coli) DH5α under ampicillin selection. Growth was carried in Terrific Broth medium (20 g/L tryptone, 24 g/L yeast extract, 4 mL/L glycerol, 0.017 M KH2PO4, 0.072 M K2HPO4) at 37 °C, with an orbital agitation of 250 rpm and suspended at late log phase. The cells were then pelleted by centrifugation (5000g, 15 min at 4 °C). Plasmid DNA was purified using the Qiagen (Hilden, Germany) plasmid maxi kit according to the instructions of the manufacturer. The protocol is based on a modified alkaline lysis procedure. Following lysis, binding of pDNA to the Qiagen anion-exchange resin is promoted under appropriate lowsalt and pH conditions. Impurities are removed by a medium-salt wash and plasmid DNA is eluted in a high salt buffer. The plasmid sample is then concentrated and desalted by isopropanol precipitation and resuspended in Tris-HCl buffer (pH 8.0). Linear plasmid DNA was then prepared by digestion with the Sac I restriction enzyme. The ratio of absorbance at 260 and 280 nm was found to be above 1.8, and, also spectrophotometrically, it was confirmed that A320 was negligible, which suggests the absence of proteins. Preparation of Gels. Plasmid DNA was dissolved in water containing 1 mM NaBr to a DNA concentration of 1 wt %. DNA was chemically cross-linked by EGDE at pH 8.5. After adding 0.1 M NaOH and TEMED, the sample was mixed and then transferred to test tubes and incubated for 2 h in a water bath at 50 °C. Gels with 0.2% and 0.5% EGDE were prepared. The concentration of pDNA in the gels was obtained by weighing gels before and after freeze-drying. Some of the gels were cut into thin cylinders and dried (gels with approximately 1 g, 1 cm length, and 0.25 cm in diameter).
Plasmid DNA Gel Degradation by Ultraviolet Light Irradiation. The pDNA gels (gels with approximately 1 g, 1 cm length, and 0.25 cm in diameter) were irradiated with ultraviolet light (400 nm) using a mercury lamp in order to promote the gel photodisruption. Plasmid DNA Release Measurements. To determine the amount of pDNA released, the gels (around 1 g) were suspended in 20 mL of pH 7.6 10 mM Tris HCl buffer. The samples were incubated at 25 °C with gentle shaking (40 rpm). At certain time intervals, the supernatant was collected and gels were resuspended in fresh solution. Plasmid DNA released into the supernatant was quantified by spectrophotometrically measuring the absorbance at 260 nm using a Shimadzu UVvis 1700 spectrophotometer. Weight Loss Experiments. After formation, pDNA gels crosslinked with different amounts of EGDE were dried, by freeze-drying, and weighed (reference state).The mass changes of dry gels were monitored over light exposure time and normalized to their initial values before network disruption. Swelling Experiments. To study the swelling behavior of the pDNA hydrogels, they were weighed after being preswollen in a NaOH 1 mM solution (reference state), and on several occasions during their disruption. The swelling ratio was calculated by dividing the weight of the gels at steady-state swelling by their weight in the reference state. Determination of the Average Gel Mesh Size. The network mesh size represents the distance between consecutive cross-linking points and provides a measure of the porosity of the network. As for other gel systems,36 the mesh size of pDNA hydrogels was determined using the method described by Canal and Peppas37 and it is understood 13781
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir
ARTICLE
as an average value. The root-mean-squared end-to-end distance of the polymer chain in its unperturbed state was calculated using eq 1: 2 1=2 ¼ lCn 1=2 n1=2 ð1Þ r̅ 0 where l is the length of the bond along the backbone chain, Cn is the characteristic ratio of the polymer, and n is the number of bonds in the cross-link, determined by eq 2: n¼
2M ̅ c M0
ð2Þ
where M0 and Mc are the molecular weight of the repeating units making up the polymer chains and the molecular weight of the polymer chains between cross-links, respectively. The degree of cross-linking in the gel (X) can be defined as X = M0/2Mc, which allows for the determination of M c. The mesh size in angstroms, ξ, of the swollen polymer network was then calculated from eq 3: 1=3
ξ¼υ
2 1=2
ðr 0 Þ
ð3Þ
where υ is the polymer volume fraction in the equilibrium swollen gel, which equals the reciprocal of the volume swelling ratio, Q; it is a measure of the amount of fluid that a hydrogel can incorporate into its structure. This parameter is determined using equilibrium swelling experiments. Incorporation of Solutes and Release. Hydrocortisone, cortisone, dexamethasone, prednisolone, and prednisone were initially dissolved in ethanol. Thereafter, the last solutes and lysozyme, BSA, and FITC-dextran have been incorporated into pDNA gels by imbibition. Concentrated solutions of the mentioned solutes (5 wt %) were prepared. The dried gels were saturated with solute by placing two cylinders in 50 mL of each solute solution. The gels were left in the solutions for approximately 48 h. Lysozyme, BSA, FITC-dextran loaded gels were placed in a 500 mL bottle containing 150 mL of the HEPES buffer (pH 7.4, 37 °C), while hydrocortisone, cortisone, dexamethasone, prednisolone, and prednisone containing gels were placed in a 500 mL bottle containing 150 mL of the PBS buffer (pH 7.4, 37 °C). Aliquots of the sample solutions were withdrawn at appropriate time intervals, and the volume of the medium was kept constant by replacement. The lysozyme concentration was measured with the Bio-Rad protein assay following the microassay procedure. The Bio-Rad protein assay is a dye-binding assay in which a differential color change of a dye occurs in response to various concentrations of protein.38 This protein assay, based on the method of Bradford, is a simple and accurate procedure for determining the concentration of solubilized protein. It involves the addition of an acidic dye to the protein solution and subsequent measurement at 595 nm with a spectrophotometer. The absorbance maximum for an acidic solution of Coomassie Brilliant Blue G-250 dye shifts from 465 to 595 nm when binding to protein occurs.39,40 Comparison to a standard curve provides a measurement of protein concentration. BSA was analyzed by determining the total mass of protein using the bicinchoninic acid (BCA) method.41 Bicinchoninic acid forms a complex with Cu+ ions, producing a purple colored solution that can be quantitatively measured at 562 nm. The protein to be analyzed reacts with Cu2+ in an alkaline solution producing Cu+ ions. These ions are then chelated by the BCA which converts the apple-green color of the free BCA to the purple color of the copperBCA complex.41 As the physical state of the released lyzosyme and BSA could be affected by exposure to UV light and cross-linking agents, leading to changes in their structure and stability, the released proteins were analyzed using a wavelength of 214 nm to study possible conformational changes. It was then observed that the preparation of the gels, as well as the conditions of the release study, did not measurably alter the native protein structure.
Considering the wavelength of maximum absorbance for FITC-dextran, 494 nm, its concentration was determined using a UVvis spectrophotometer at this wavelength. Hydrocortisone, cortisone, dexamethasone, prednisolone, and prednisone become fluorescent when dissolved in ethanol. Their concentrations released in the PBS buffer were detected by ultraviolet light as function of time. The wavelengths used for the detection were 238 243 nm. To avoid interference of gel degradation products on the measurement of the solute concentration, the solutions from identical pDNA gels were used as a blank in measuring the solute concentration as a function of degradation. Release Models. A number of equations have been developed to model the release of solutes from polymeric matrices, ranging from those based on purely theoretical models42 to semiempirical approaches.43 Among these, the most classical one is the simple power law Korsemeyer Peppas equation44 Ct =C∞ ¼ kt n
ð4Þ
where Ct and C∞ are cumulative concentrations of the material released at time t and at infinite time, respectively, and k and n are fitting parameters, giving the later useful information on the release mechanism; for the release from cylinders, values of n near 0.45 indicate a diffusion-controlled release (so-called Fickian). Non-Fickian behavior is observed for 0.45 < n < 0.89, with a limit of Case II (zero-order release) transports for n = 0.89. NonFickian behavior and Case II transport are indicative of coupling of diffusional and relaxational mechanisms. Occasionally, values of n > 0.89 have been observed and considered to be Super Case II kinetics.44 This mechanism could result from increased plasticization at the relaxing boundary (gel layer), that is, when the surface resistance becomes more significant relative to the diffusion resistance.45 The validity of eq 4 is restricted to Ct/C∞ < 0.60.42 Mean dissolution time (MDT), used to characterize the drug release rate from a dosage form and to indicate the drug-release-retarding efficiency of the polymer,46 is calculated through the following equation47 n 1 kn MDT ¼ ð5Þ n þ 1 where n and k have the same meaning as in eq 4. An empirical approach, which allows describing the entire set of the release data, is based on the Weibull function h i ð6Þ Ct ¼ C∞ 1 expð k0 tÞd where k0 and d are constants. Papadopoulou et al.48 have demonstrated that eq 6 gives an insight into the diffusional mechanism, since d and k0 are closely related to the mechanism and rate constant of release, respectively. All the fitting parameters have been estimated through a least-squares method, with a confidential degree of 95%, by using Origin software.
’ RESULTS AND DISCUSSION The Gel Photodisruption. In previous works, we report on the production and purification of plasmid DNA49,50 and for the first time on the preparation of linear plasmid DNA gels by crosslinking reaction with 0.2% and 0.5% EGDE densities.34 The reaction mechanism of gel formation seems to involve the guanine nitrogen atom at position seven (N-7), which attacks the more substituted carbon of the epoxide or the least hindered end of the epoxide. The reaction is of nucleophilic substitution type (SN2 reaction). All plasmid DNA gels are clear and transparent (Figure 1A).34 Structural characterization by means 13782
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir
ARTICLE
Figure 1. Photograph (A) and scanning electron micrograph (B) of plasmid DNA gels cross-linked with 0.2% EGDE.
of scanning electron microscopy revealed that the gels are porous with a smooth surface (Figure 1B). Moreover, the gel preparation occurs with the conservation of plasmid DNA secondary structure.34 To demonstrate that the pDNA gels protect the encapsulated pDNA, the gels were exposed to serum nucleases by combining them with Dulbecco's modified Eagle's medium containing 10% fetal bovine serum at 37 °C. Any degradation of the plasmid DNA by enzymes present in this medium would result in small, linear fragments, and short DNA sequences would not be retained in the network and would diffuse out of the gels. However, when the samples were analyzed spectrophotometrically, no appreciable decrease in the concentration of double stranded DNA was detected with the plasmid DNA encapsulated gels, and thus, no degradation of the pDNA was observed. Ethylene glycol ethers do not persist in the environment or bioaccumulate in tissues, and they are “practically nontoxic” to aquatic organisms; these ethers photooxidize in the presence of sunlight.51 The degradation on ultraviolet light exposure (photooxidation) of EGDE leads to the removal of the chemical crosslinks and can allow the release of the constituent network polymer, inducing changes in gel weight, mechanical properties, mesh size, porosity, and the degree of swelling.12 Moreover, preliminary studies on cross-linked DNA gel systems showed that the ultraviolet light that causes the maximum network degradation is λdisruption = 400 nm.36 Thus, in order to demonstrate the plasmid DNA gel disruption, experiments on plasmid DNA release were performed after the gels being irradiated with light (400 nm) and in the dark conditions, for pDNA gels crosslinked with 0.2% and 0.5% EGDE, as illustrated in Figure 2A. After irradiation, both gels suffered disruption leading to the release of plasmid DNA with time. For 0.2% and 0.5% EGDE cross-linked gels, plasmid DNA release behavior presents a narrow time lag in the first 24 h, after which the release gradually increases until a plateau is reached around 400 h of photodegradation. The initial time lag may be related to the number of cross-links that have to be degraded to permit the release of plasmid DNA. After irradiation, and at maximum release, pDNA gels cross-linked with 0.2% EGDE released 87.8% of plasmid DNA while gels prepared from 0.5% EGDE released 74.7% of pDNA, in approximately 18 days. Not all the cross-linked plasmid DNA is released, probably due to the fact that the inhomogeneous distribution of the cross-linker leads to the existence of very concentrated cross-linker/DNA regions, from where the release of DNA is extremely difficult. Additionally, it is noted that the extent of pDNA release is quite dependent on cross-linker density, as found before for similar systems;12 the higher the cross-linker concentration used in the gel preparation, the lower the amount of plasmid DNA released. Completely different
Figure 2. Cumulative release of pDNA from cross-linked pDNA gels with 0.2% (9) and 0.5% (b) (w/v) EGDE as a function of time. Studies were performed after the irradiation of gels with light (400 nm) (A) and in dark conditions (B).
behavior was found for release studies performed in dark conditions (Figure 2B). In the absence of ultraviolet light irradiation, and for both gel types, minimal amounts of plasmid DNA are released, less than 8%. This corresponds, in each gel system, approximately to the free plasmid DNA in the network; that is, the pDNA not cross-linked. At the preparation time, the relation between the pDNA concentration to EGDE concentration allows the gel to maintain some of the chains in a non-crosslinked state. The disruption of the cross-linked plasmid DNA gels was also monitored, as shown in Figure 3, by determining the weight loss of the dried hydrogel after light irradiation, as a function of disruption time. The disruption rate is dependent on cross-linker density; the higher the cross-linking density used to synthesize the gels, the lower the disruption rate. Plasmid DNA gels crosslinked with 0.2% EGDE disrupt within approximately 2 weeks, and those with 0.5% EGDE within 20 days. As can be seen, the 13783
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir
ARTICLE
Figure 3. Weight loss as a function of time of pDNA gels cross-linked with 0.2% (9) and 0.5% (b) (w/v) EGDE after the irradiation of gels with light (400 nm).
Figure 4. Change of the degree of swelling of pDNA gels cross-linked with 0.2% (9) and 0.5% (b) (w/v) EGDE as a function of time after the irradiation of gels with light (400 nm).
weight loss data differs from the cumulative DNA release in Figure 2. In DNA release measurements, we determined spectrophotometrically the amount of DNA released into the bulk solution, while the weight loss measurements take into account every smaller portion of the gel that separates from the main part, portions of gel that have separated into the release medium. To further characterize the disruption behavior, the changes in the degree of swelling after gels being irradiated with light were analyzed for each of the plasmid DNA gels. The results are summarized in Figure 4. The weight of the gels preswollen in the 1 mM NaOH solution was taken as the first point (this swelling ratio given as unity) in Figure 4 and therefore represents the equilibrium degree of swelling of the “intact” covalently crosslinked plasmid DNA gels. Plasmid gels cross-linked with 0.2% and 0.5% EGDE presented an increased extent of swelling with increasing time of disruption. In addition, 0.2% EGDE plasmid DNA gels reach the plateau, indicating maximum swelling,
Figure 5. Gel mesh size of 0.2% (9) and 0.5% (b) (w/v) EGDE crosslinked pDNA gels as a function of time after the irradiation of gels with light (400 nm) (A) and in the dark conditions (B).
after 8 days of gels irradiation with λdisruption, while 0.5% pDNA gels exhibiting maximum swelling degree after approximately 13 days of irradiation with the same wavelength. This reflects the cross-linking density of the gel and its effect on gel degradation. Further evidence of plasmid DNA gel degradation comes from the evolution of mesh size with disruption time, after 0.2% and 0.5% EGDE cross-linked gels are irradiated with light (Figure 5A). In addition, studies were also performed in the dark conditions, as represented in Figure 5B. As mentioned in the experimental section, knowledge of the equilibrium degree of swelling allows for the calculation of matrix structural parameters such as the distance between cross-links and the mesh size of the gel. Following the method of determination due to Canal and Peppas,37 the initial mesh size, before induced gel disruption, is determined to 39 and 32 Å for pDNA gels cross-linked with 0.2% and 0.5% EGDE, respectively. After ultraviolet light irradiation, variations in the mesh size can be found. For both types of plasmid DNA gels, it increases gradually with time of disruption until a maximum value of 122 and 113 Å for 0.2% and 0.5% EGDE pDNA gels, 13784
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir
Figure 6. Cumulative release of lysozyme, BSA, and FITC-dextran from 0.2% EGDE plasmid DNA gels as a function of time and as a function of gel mesh size (inset) when gels were irradiated with light (400 nm). Solid lines correspond to the best fits of experimental data to eq 6.
respectively, after approximately 13 days of network disruption. An opposite situation was verified in the dark conditions, where gel mesh size remains unaltered with time for both 0.2% and 0.5% EGDE cross-linked gels. This evidenced the ability of using the ultraviolet light irradiation in modulating the pDNA network disruption. Moreover, as the gel network degrades, the mesh size increases, which, as we will see, contributes to the release of incorporated solutes. Release of Solutes. In degradable gels, if the release of solutes is governed by polymer degradation, the degradation rate of the hydrogel must be matched to the size of the solute species. Moreover, control of variations with mesh size in time is crucial to design appropriate solute release devices. During hydrogel degradation, the water content tends to increase, increasing both the network mesh size and the volume swelling ratio, and, consequently, the release of gel entrapped solutes is facilitated. To evaluate the dependence of the release rate on the size of an entrapped solute, three different solutes were incorporated into 0.2% and 0.5% EGDE cross-linked pDNA gels. The chosen solutes were lysozyme, BSA, and FITC-dextran with molecular weights ranging from 14 100 to 77 000 Da with and the hydrodynamic radii being 16, 34.8, and 55 Å, respectively.35 The effect of radiation on the release profiles of these molecules has been studied. Figures 6 and 7 show the release kinetics of the different molecules from 0.2% and 0.5% EGDE pDNA matrices, respectively. In the inset of each figure, it is represented the mesh size as a function of time after the irradiation of gels. From the release curves, it is possible to infer that approximately 98.6% lysozyme, 94.9% BSA, and 96.6% FITC-dextran are released in 48, 192, and 274 h, respectively, from 0.2% EGDE pDNA gels. Similarly, approximately 94.9% lysozyme, 93.1% BSA, and 94.1% FITC-dextran are released in 72, 240, and 300 h, respectively, from 0.5% EGDE pDNA gels. In order to have a deep assessment on the release mechanism, the release kinetics has been fitted to eqs 4 and 6. The fitting parameters are summarized in Table 1. Equation 4 is commonly used for an assessment of the release mechanism; for BSA and FITC-dextran, n values are higher than 0.89, indicating a Super Case II release phenomena. This is also confirmed by d (eq 6) values higher than 1. This mechanism
ARTICLE
Figure 7. Cumulative release of lysozyme, BSA, and FITC-dextran from 0.5% EGDE plasmid DNA gels as a function of time and as a function of gel mesh size (inset) when gels were irradiated with light (400 nm). Solid lines correspond to the best fits of experimental data to eq 6.
emerges as the polymer resistance becomes more significant relative to the diffusion resistance; consequently, the release is linear at early times in the process, but at long time the release rate increases markedly (Figures 6 and 7). On the other hand, and despite the limited number of experimental data for the release of lysozyme, the analysis of the exponent d suggests a Fickian release process. It is interesting to note that a similar mechanism has been found for the release of lysozyme and in the case of BSA a change from Fickian to Super Case II release from DNA (from salmon testes, 2.0-kbp) gels with 5% EGDE cross-linker;36 this clearly suggests that another process (e.g., swelling), beside diffusion, affects the transport of BSA in higher molecular weight plasmid DNA gels.52 Concomitantly, for plasmid DNA gels, the release rate (k0 ) for the biggest solutes decreases (k0 values for BSA and FITC-dextran release from salmon testes DNA gels, crosslinked with 0.5% EGDE, are 1.7 105 and 2.5 106 s1, respectively36). This can be justified by an increase in the mobility of the polymeric chains, with an increase of the molecular weight, leading to a further retardant factor for the solute release.53 From the analysis of the release rate (k0 ) of lysozyme, BSA, and FITC-dextran from pDNA gels with different degrees of crosslinker (Table 1), we can also conclude that (a) k0 decreases by increasing the degree of cross-linker and (b) for both gels, k0 decreases from lysozyme to FITC-dextran by changing ca. 2 orders of magnitude. For pDNA gels with different cross-linker concentrations, we verified that lysozyme is efficiently released upon gel disruption, probably due to its small size (16 Å). Furthermore, a time lag (of 37 and 86 h for EGDE 0.2 and 0.5%, respectively) for the release profile of FITC-dextran is found, which can be explained by its larger size compared to the other solutes. In the case of BSA (which has intermediate dimensions), there is no time-lag for EGDE 0.2% gels, while a rather small time-lag is observed by increasing the cross-linker concentration. Such behavior might suggest that the release is partially dependent on the size of the three solutes; however, contrary to what happens for salmon testes DNA gels, the hydrodynamic model54 does not fit to these experimental data; this is in agreement with previous discussion where other factors such as mobility of polymer chain and swelling cannot be neglected for the interpretation of these solutes release. 13785
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir
ARTICLE
Table 1. Fitting Parameters of eqs 4 and 6 to Experimental Release Data of Lysozyme, BSA, and FITC-Dextran Previously Loaded in pDNA Gels to Water eq 4 EDGE lysozyme
BSA
FITC-dextran a d
n
eq 6 2
log k
R
6 1
k0 (10
s )
d
R2
0.2%
97 ((2)
0.91 ((0.05)
0.9988
0.5%
81 ((2)
1.09 ((0.05)
0.9990
0.2%
0.94 ((0.05)
1.83 ((0.07)
0.99934
5.3 ((0.3)
1.1 ((0.1)
0.9821
0.5%
0.98 ((0.03)a,b
2.27 ((0.05)a,b
0.9958
2.50 ((0.06)
2.1 ((0.2)
0.9916
0.2%
1.4 ((0.1)a,c
3.1 ((0.2)a,c
0.9742
1.88 ((0.01)
2.89 ((0.09)
0.9987
0.5%
a,d
5.8 (0.4)a,d
0.9837
1.31 ((0.01)
4.2 ((0.2)
0.9957
2.4 ((0.2)
A modified version of eq 4 has been used, in order to take into account an observed time-lag (θ): Ct/C∞ = k(t θ)n.36,57 b θ = 3.4 h. c θ = 37 h. θ = 86 h.
To gain further insight into these mechanisms, the effect of gel mesh size on solute release was examined. The results, presented in the insets of Figures 6 and 7, suggested that, in general, there is a correlation between the solute size, the gel mesh size, and the global release profile. Although all the incorporated solutes were able to diffuse out from the 0.2% and 0.5% EGDE pDNA gels, when they suffer degradation, their release patterns are completely different due to their different sizes. As referred above, the mesh size is 39 and 32 Å for pDNA gels cross-linked with 0.2% and 0.5% EGDE, respectively. Due to the larger mesh size of 0.2% EGDE cross-linked gels, incorporated lysozyme and BSA can rapidly diffuse out from the network in consequence of their size, 16 and 34.8 Å, respectively. These solutes were almost totally released at an early stage, probably with the high concentration gradient contributing to this situation. In contrast, negligible amounts of FITC-dextran have been released in the first 48 h, which is referred to its larger size compared to the initial gel mesh size. Upon degradation, the network mesh size is continuously increasing, approaching the dextran size, and so the dextran began to diffuses slowly. At a certain time, the gel mesh size was larger than the solute size and the dextran release rate greatly increases until it reaches its maximum release. An identical situation occurs for 0.5% EGDE pDNA gels, however, with some differences concerning two aspects: first, the evidence of a time lag for the BSA release when the gel mesh size is lower than 40 Å and a consequent slow release trend and, second, the major time lag found for FITC-dextran, with the release profile increasing for mesh size values higher than 60 Å. To further demonstrate the sustained release characteristics of the pDNA gel system, release studies were evaluated in vitro using selected anti-inflammatory drugs. Figure 8 presents the results of drug release from 0.2% and 0.5% EGDE pDNA gels. The drugs used were hydrocortisone, cortisone, prednisolone, dexamethasone, and prednisone. All the mentioned drugs can be successfully released from plasmid DNA matrices. Although no burst effect can be observed in the first hours, the cumulative release of hydrocortisone, cortisone, and prednisolone increases gradually with time until a plateau is reached at maximum release between 35 and 40 h. It is worth noting that a pronounced time lag is verified, for the highest cross-linked gel, in the initial hours, for the release of both dexamethasone and prednisone: 4.1 and 10.8 h, respectively. After that, the cumulative release of such drugs significantly increases until the maximum value is reached around 80 h. The quantification of the release mechanism and
rate constant has been done by using eqs 4 and 6, and the corresponding fitting parameters are shown in Table 2. The exponential factors in both equations, n and d, are higher than 0.89 and 1, respectively, indicating a Super Case II release phenomenon. A possible explanation for that can be found in the swelling degree of gels. Assuming that the loading of drugs does not strongly affect the swelling properties of pDNA gels, it can be seen from Figure 4 that in the first 40 h only a slightl swelling increase (2% and 6% to 0.5% and 0.2% EDGE pDNA gels, respectively) is observed; however, after that, an increase of Q up to 60% of the initial volume occurs, decreasing, in this way, the resistance of polymer to drug release, explaining the burst observed in the release profiles and, consequently, the release mechanism. From the analysis of MDT, it is possible to conclude that the release is dependent on the type of anti-inflammatory and, for each drug, a slight effect of cross-linker can be noted. In order to have an assessment of the factors affecting the release kinetics, the effect of hydrophobicity and hydration is evaluated. It is known that the diffusion coefficients of 5 mM aqueous solutions of cortisone, hydrocortisone, prednisone, and prednisolone are 3.83 1010, 3.86 1010, 4.03 1010, and 4.83 1010 m2 s1, respectively;55 comparing the homologous molecules (cortisone and hydrocortisone, and prednisone and prednisolone), it can be found that also in the case of the first two drugs the MDT values are similar; while comparing the later two, prednisolone shows a significantly smaller MDT. This also explains the increase of MDT values with an increase of cross-linker concentration for the fastest species releasing out. These data suggest that the hydrated volume of these drugs has a non-negligible influence on the release from pDNA gels. However, hydration does not justify the comparatively higher MDT value for prednisone (and dexamethasone). Another important factor that might influence the release is the hydrophobicity of these drugs, quantified through the octanolwater partition coefficients (log P).56 The (log P) values for dexamethasone and prednisone are 1.9 and 2.39, respectively, much higher than those reported for hydrocortisone, cortisone and prednisolone: 1.1, 1.49 and 1.3, respectively. It can be found that the MDT values are higher for the most hydrophobic drugs, suggesting the low affinity for water, when compared with the gel matrix. Two other relevant observations arise from the data analysis: (a) For the most hydrophobic drugs, MDT values decrease by increasing the concentration of cross-linker, which can lead to a lower drug 13786
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir
ARTICLE
Figure 8. Release profiles of (A) hydrocortisone, (B) cortisone, (C) prednisolone, (D) dexamethasone, and (E) prednisone from plasmid DNA gels cross-linked with 0.2% (9) and 0.5% (b) (w/v) EGDE when gels were irradiated with light (400 nm). Solid lines correspond to the best fits of experimental data to eq 6.
stability inside gel matrix. (b) By plotting MDT values as a function of rate constants (k0 ), it can be seen that, for MDT lower than 30 h, there is a good correlation between those two parameters; however, for MDT values higher than 30 h, an
increase of MDT is accompanied by a rather constant value of k0 . This behavior gives a clear indication that for dexamethasone and prednisone the rate constant, k0 , is controlled by the long time range release. The release of the studied drugs is dependent on a 13787
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir
ARTICLE
Table 2. Fitting Parameters of eqs 4 and 6 to Experimental Release Data of Anti-Inflammatories Loaded in pDNA Gels to Water (Figure 8) eq 4
cortisone hydrocortisone prednisolone prednisone dexamethasone a
EDGE
n
log k
eq 6 2
R
MDT (h)
a
2 1
k0 (10
s )
d
R2
0.2%
1.53 ((0.08)
7.7 ((0.3)
0.9828
18
4.62 ((0. 07)
2.5 ((0.1)
0.9940
0.5%
2.2 ((0.1)
11 ((0.5)
0.9816
19
4.15 ((0.03)
3.16 ((0.09)
0.9983
0.2% 0.5%
1.09 ((0.05) 1.9 ((0.1)
5.5 ((0.2) 9.5 ((0.4)
0.9861 0.9849
16 18
5.4 ((0.1) 4.94 ((0.06)
2.1 ((0.2) 2.6 ((0.1)
0.9856 0.9955
0.2%
1.38 ((0.04)
6.8 ((0.2)
0.9930
14
6.2 ((0.1)
1.92 ((0.09)
0.9948
0.5%
1.82 ((0.08)
9.0 (0.4)
0.9863
16
5.36 ((0.07)
2.20 ((0.09)
0.9961 0.9988
0.2%
1.69 ((0.05)
9.1 (0.3)
0.9850
42.3
2.16 ((0.01)
2.42 ((0.04)
0.5%b
2.2 ((0.1)
11.7 ((0.4)
0.9791
39.7
1.95 ((0.02)
2.68 ((0.09)
0.9955
0.2%
1.54 ((0.07)
8.2 ((0.3)
0.9767
35.6
2.36 ((0.02)
2.17 ((0.04)
0.9984
0.5%c
1.70 ((0.04)
8.9 ((0.2)
0.9920
30.1
2.15 ((0.02)
2.37 ((0.05)
0.9981
Mean dissolution time (MDT 60%) calculated from eq 5. b θ = 10.8 h. c θ = 4.1 h.
balance between hydrodynamic volume and hydrophobicity, leading the former to a faster release; on the other hand, a correlation between the drug release trend and the drug hydrophobicity can be addressed, with the evidence of a slow release rate for the more hydrophobic drugs. Moreover, we detected that the drug release occurs within a shorter disruption time than the release of pDNA (see previous section), indicating the existence of pDNA and drug release profiles characterized by different time scales. This opens up the possibility to manipulate the release in accordance with needs of pDNA and/or drug leading to the promising combination of gene and drug delivery in real biosystems.
’ CONCLUSIONS We have made use of our knowledge in the field of physical chemistry of DNA networks to develop a photodegradable plasmid DNA hydrogel as a potential candidate for the sustained delivery of both plasmid DNA and drugs. The gel disruption, upon light irradiation, was demonstrated by means of cumulative pDNA release, weight loss, and increases in both the degree of swelling and the mesh size. Lysozyme, BSA, and FITC-dextran have been incorporated in pDNA matrices; the analysis of the release kinetics indicates a Fickian release process for lysozyme, while a Super case II phenomena has been found for the other two solutes. Although there is a correlation between the gel mesh size and the solute size, not only is drug size important but also the mobility of polymer chains and the extent of swelling are factors that should be taken into account. Modeling of the in vitro release studies using anti-inflammatory drugs is consistent with a Super Case II release mechanism, related to increases in the gel swelling degree. Furthermore, the release of the anti-inflammatory drugs depends on the balance between hydrodynamic volume and hydrophobicity, and there is a correlation between drug release and hydrophobicity; a slow release rate was found for the more hydrophobic drugs. The presented biomaterial can be quite useful as a controlled release device for several therapeutic agents at wound sites because it is biocompatible, it can be degraded, and it can release the incorporated solute in a chemically and conformationally stable form. Our findings in this model system are crucial for the development of pDNA-based formulations for the codelivery of drugs and genes which may
improve the therapeutic efficacy in the treatment of severe diseases, such as cancer, due to their synergistic/combined effect.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT We are grateful for financial support from Fundac-~ao para a Ci^encia e a Tecnologia (FCT) (SFRH/BPD/47229/2008) and to PTDC/EBB-Bio/114320/2009. The authors would like to ^ ngela Sousa for the 2.7-kpb plasmid pUC19 thank Fani Sousa e A production and recovery. ’ REFERENCES (1) Wichterle, O.; Lim, D. Nature 1960, 185, 117–118. (2) Lee, K. Y.; Bouhadir, K. H.; Mooney, D. J. Biomaterials 2004, 25, 2461–2466. (3) Wang, C.; Stewart, R. J.; Kopecek, J. Nature 1999, 397, 417–420. (4) Kopecek, J. Eur. J. Pharm. Sci. 2003, 20, 1–16. (5) Amiya, T.; Tanaka, T. Macromolecules 1987, 20, 1162–1165. (6) Horkay, F.; Basser, P. Biomacromolecules 2004, 5, 232–237. (7) Topuz, F.; Okay, O. Macromolecules 2008, 41, 8847–8854. (8) Orakdogen, N.; Karacan, P.; Okay, O. React. Funct. Polym. 2011, 71, 782–790. (9) Costa, D.; Hansson, P.; Schneider, S.; Miguel, M.; Lindman, B. Biomacromolecules 2006, 7, 1090–1095. (10) Costa, D.; Miguel, M.; Lindman, B. J. Phys. Chem B 2007, 111, 8444–8452. (11) Costa, D.; Miguel, M.; Lindman, B. J. Phys. Chem B 2007, 111, 10886–10896. (12) Costa, D.; Valente, A. J. M.; Pais, A. A. C. C.; Miguel, M.; Lindman, B. Colloids Surf., A 2010, 354, 28–33. (13) Costa, D.; Miguel, M.; Lindman, B. Adv. Colloid Interface Sci. 2010, 158, 21–31. (14) Langer, R.; Vacanti, J. P. Science 1993, 260, 920–926. (15) Yang, S.; Leong, K.-F.; Du, Z. Tissue Eng. 2001, 7, 679–689. (16) Amsden, B. Macromolecules 1998, 31, 8382–8395. (17) Mason, M. N.; Metters, A. T.; Bowman, C. N.; Anseth, K. S. Macromolecules 2001, 34, 4680–4635. (18) Grassi, M.; Grassi, G. Curr. Drug Delivery 2005, 2, 97–116. (19) Siepmann, J.; Siepmann, F. Int. J. Pharm. 2008, 364, 328–343. 13788
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
Langmuir (20) Aguzzi, C.; Cerezo, P.; Salcedo, I.; Sanchez, R.; Viseras, C. Adv. Perform. Mater. 2010, 24, 205–211. (21) Peppas, N. Hydrogels in Medicine and Pharmacy; CRC Press: Boca Raton, FL, 1987. (22) Peppas, N. A.; Bures, P.; Leobandung, W.; Ichikawa, H. Eur. J. Pharm. Biopharm. 2000, 50, 27–46. (23) Jeong, B.; Kim, S.; Bae, Y. Adv. Drug Delivery Rev. 2002, 54, 37–51. (24) Sershen, S.; West, J. Adv. Drug Delivery Rev. 2003, 55, 439–445. (25) Torres-Lugo, M.; Garcia, M.; Record, R.; Peppas, N. A. J. Controlled Release 2002, 80, 197–205. (26) Podual, K.; Doyle, F. J.; Peppas, N. A. J. Controlled Release 2000, 67, 9–17. (27) Podual, K.; Doyle, F. J.; Peppas, N. A. Biomaterials 2000, 21, 1439–1450. (28) Peppas, N. A.; Langer, R. AIChE J. 2004, 50, 536–546. (29) Cadee, J. A.; de Groot, C. J.; Jiskoot, W.; den Otter, W.; Hennink, W. E. J. Controlled Release 2002, 78, 1–13. (30) Maia, J.; Ribeiro, M. P.; Ventura, C.; Carvalho, R. A.; Correia, I. J.; Gil, M. H. Acta Biomater. 2009, 5, 1948–1955. (31) Leach, J. B.; Schmidt, C. E. Biomaterials 2005, 26, 125–135. (32) Brandl, F.; Hammer, N.; Blunk, T.; Tessmar, J.; Goepferich, A. Biomacromolecules 2010, 11, 496–504. (33) Nguyen, M. K.; Lee, D. S. Macrommol. Biosci. 2010, 10, 563–579. (34) Costa, D.; Queiroz, J.; Miguel, M. G.; Lindman, B. Submitted. (35) Kim, B.-S; Mooney, D. J. Trends Biotechnol. 1998, 16, 224–228. (36) Costa, D.; Valente, A. J. M.; Miguel, M. G.; Lindman, B. Colloids Surf., A 2011, doi:10.1016/j.colsurfa.2011.03.046. (37) Canal, T.; Peppas, N. A. J. Biomed. Mater. Res. 1989, 23, 1183–1193. (38) Bradford, M. Anal. Biochem. 1976, 72, 248–253. (39) Reisner, A. H.; Nemes, P.; Bucholtz, C. Anal. Biochem. 1975, 64, 509–514. (40) Sedmack, J. J.; Grossberg, S. E. Anal. Biochem. 1977, 79, 544–550. (41) Scopes, R. K. Protein Purification: Principles and Practice; SpringerVerlag: New York, 1994; pp 4050. (42) Crank, J. The Mathematics of Diffusion, 2nd ed.; Oxford University Press Inc.: New York, 1975. (43) Amsden, B. Polymer 2002, 43, 1623–1630. (44) Korsmeyer, R. W.; Gurny, R.; Docler, E.; Buri, P.; Peppas, N. A. Int. J. Pharm. 1983, 15, 25–35. (45) Polishchuk, A. Y.; Zaikov, G. Multicomponent transport in polymer systems for controlled release; Gordon and Breach Sci. Publ.: Amsterdam, 1997. (46) Sriamornsak, P.; Sungthongjeen, S. AAPS PharmSciTech 2007, 8, 51. (47) Mockel, J. E.; Lippold, B. C. Pharm. Res. 1993, 10, 1066Y1070. (48) Papadopoulou, V.; Kosmidis, K.; Vlachou, M.; Macheras, P. J. Pharm. 2006, 309, 44–50. (49) Prazeres, D. M. F.; Diogo, M. M.; Queiroz, J. A. Process for Production and Purification of Plasmid DNA. Patent Number PT 102491 R; WO 02/04027 A1; United States Patent 7,169,917, January 30, 2007. (50) Sousa, F.; Prazeres, D. M. F.; Queiroz, J. A. Trends Biotechnol. 2008, 26, 518–525. (51) Staples, C.; Boatman, R.; Cano, M. Chemosphere 1998, 36, 1585–1613. (52) Erdogan, M.; Hizal, G.; Tunca, U.; Hayrabetyan, D.; Pekcan, O. Polymer 2002, 43, 1925–1931. (53) Erdogan, M.; Hepuzer, Y.; Cianga, I.; Yagci, Y.; Pekcan, O. J. Phys. Chem. A 2003, 107, 8363–8370. (54) Cukier, R. I. Macromolecules 1984, 17, 252–255. (55) Kabasakalian, P.; McGlotten, J. J. Am. Chem. Soc. 1956, 78, 5032–5036. (56) Leo, A.; Hansch, C.; Elkins, D. Chem. Rev. 1971, 71, 525– 616.
ARTICLE
(57) Gouanve, F.; Marais, S.; Bessadok, A.; Langevin, D.; Metayer, M. Eur. Polym. J. 2007, 43, 586–598.
13789
dx.doi.org/10.1021/la2026285 |Langmuir 2011, 27, 13780–13789
