cationic liposome complexes for gene delivery and vaccine therapy

Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Microfluidic devices for continuous production of pDNA/cationic liposome complexes for gene delivery and vaccine therapy Tiago Albertini Balbino a , Adriano Rodrigues Azzoni b , Lucimara Gaziola de la Torre a,∗ a b

University of Campinas, UNICAMP, School of Chemical Engineering, PO BOX 6066 13083-970, Campinas, SP, Brazil University of São Paulo, Chemical Engineering Department, Escola Politécnica, Av. Prof. Luciano Gualberto, Travessa 3, CEP 05508-010 São Paulo, SP, Brazil

a r t i c l e

i n f o

Article history: Received 7 November 2012 Received in revised form 3 April 2013 Accepted 3 April 2013 Available online 15 April 2013 Keywords: Microfluidics Cationic liposomes DNA Micromixer Gene delivery

a b s t r a c t To evaluate the process parameters for the production of plasmid DNA/cationic liposome (pDNA/CL) complexes in microfluidic systems, we studied two microfluidic devices: one with simple straight hydrodynamic flow focusing (SMD) and a second one with barriers in the mixing microchannel (patterned walls, PMD). A conventional bulk mixing method was used as a comparison to microfluidic mixing. The CL and the pDNA were combined at a molar positive/negative charge ratio of 6. The results showed that incorporating pDNA into the liposomal structures was different for the two microfluidic devices and that the temperature influenced the average size of complexes produced by the simple microfluidic device, while it did not influence the average complex size in the patterned wall device. Differences were also observed in pDNA probe accessibility in the complexes. The SMD yielded a similar quantity of non-electrostatic bound pDNA as that provided by the bulk mixing method. The complexes produced by the PMD had their pDNA probe accessibility decreased in 40% and achieved lower in vitro transfection levels in HeLa cells than the bulk mixing and simple microfluidic complexation methods. These differences are most likely due to different degrees of association between pDNA and CL, as controlled by the microfluidic devices. This study contributes to the development of rational strategies for controlling the formation of pDNA/CL complexes for further applications in gene and vaccine therapy. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Microfluidics is a multidisciplinary science and technology where small amounts of fluids are processed. The technology is primarily dedicated to miniaturized plumbing and fluidic manipulation [1,2]. Microfluidics uses hydrodynamic characteristics to control concentrations of different molecules in space and time [3–7]. Microfluidic systems have laminar flows and vastly increased surface-to-volume ratios compared to other systems [8], leading to unique transport properties. Unlike macroscale systems, where mixing is generally accomplished by turbulent flow, mixing in microfluidic devices occurs by molecular diffusion, as influenced by convection, which is an inherently slow process. Fast mixing times (i.e., in the order of ms) in microfluidic devices can be achieved because of their small dimensions and thus small diffusion length between the components [9,10]. Micromixers with different geometries have been designed to enhance mixing efficiency by increasing the contact area between the mixing species as they flow along the

∗ Corresponding author. Tel.: +55 19 35210397; fax: +55 19 35213910. E-mail addresses: [email protected], [email protected] (L.G. de la Torre). 0927-7765/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfb.2013.04.003

mixing channel [11–16]. For passive mixing, there is a variety of specially-designed microchannel configurations, such as Y-type and multi-laminating flows, and three-dimensional serpentine channel that has enhanced mixing efficiency [15,16]. One prospective application of microfluidic processes is in the field of gene delivery and vaccine therapy [17]. Gene delivery is a potential method for treating different diseases, including gene-related disorders, infectious diseases, AIDS, and cancer [18]. Gene delivery is based on the process of introducing into a nucleus an engineered pDNA that encodes a functional, therapeutic gene that helps modulate cellular functions and responses [17,19]. However, efficient delivery requires the pDNA to be protected; pDNA/cationic liposome (CL) complexation is a promising strategy for nonviral gene therapy [20]. Liposomes are vesicles of colloidal dimension that are formed when phospholipids self-assemble in aqueous media with an interior core, mimicking cell membranes. The use of cationic lipids allows electrostatic complexation with pDNA. CL effectively protect pDNA from the extracellular matrix and shuttles those plasmids into cells [17,20]. The production of liposomes using microfluidic hydrodynamic focusing devices has been extensively studied [21–25]. Our research group recently investigated the production of CL in microfluidic devices using a lipid composition previously studied for tuberculosis gene vaccination and treatment [26–29]. However,

204

T.A. Balbino et al. / Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

the final step for gene vaccine or therapy requires electrostatic complexation with the therapeutic molecule, the pDNA. The spontaneous complexation process between pDNA and CL relies on pDNA/CL packing. Depending on the lipid composition, the molar charge ratio between positive (from cationic lipids) and negative charges (from the pDNA phosphate group) (R+/− ), the temperature of complexation, the solution concentrations, the lipid and pDNA mixture ratio and the order of mixing the solutions, complexes with different morphological and physicochemical characteristics can be obtained [20,30–32]. These differences generate different biological results [33]. Currently, the conventional method of preparing nonviral complexes is the bulk mixing (BM) process in which the two solutions are followed by hand shaking or brief vortexing [32,34,35]. Due to the several factors that influence pDNA/CL complex formation, the conventional BM method for preparing pDNA/CL complexes may yield inconsistent and poorly reproducible transfection efficiencies [35,36], most likely due to uncontrolled complexation. Consequently, the development of different strategies that can produce pDNA/CL complexes with a narrow particle size distribution and polydispersity is essential for the reproducibility of gene transfection results, following the requirements for applications in gene therapy and DNA vaccination. Another important fact is the future demand for higher quantities of pDNA/CL complexes for clinical trials. Considering the difficulties in controlling the complexation between pDNA and LC using bulk methods at higher volumes, microfluidics emerges as a promising strategy that can be explored to increase the productivity and consequent development of innovative processes. The use of microfluidic devices for forming nonviral carriers and pDNA complexes has been previously demonstrated. One approach of this is employing the hydrodynamic flow focusing technique with a central nucleic acid stream and two adjacent streams with nonviral carriers [37–40] or also employing staggered herringbone micromixers [41,42]. A second technology use employs picolitre incubators, where lipid and pDNA solutions are confined into droplets formed by emulsion water in oil, with a centrifugation post-processing step for recovering the organic solvent [36,43]. However, as far as we know, there has been no systematic study on the operational parameters and production of pDNA/CL complexes for microfluidic flow focusing devices. In this research, we investigated the parameters involved in the continuous production of pDNA/CL complexes by employing two microfluidic devices: one with a simple hydrodynamic flow focusing microfluidic device (SMD) (Fig. 1A) and a second one with barriers installed in the mixing channel (patterned microfluidic device, PMD) (Fig. 1B). We used complexes produced by the BM method as a control. We were thus able to develop a continuous flow microfluidic technology that was able to produce pDNA/CL complexes with desirable physicochemical and gene delivery properties. 2. Materials and methods 2.1. Materials The lipids egg phosphatidylcholine (EPC), 1,2-dioleoyl-snglycero-3-phosphoethanolamine (DOPE) and 1,2-dioleoyl-3trimethylammonium-propane (DOTAP) (50/25/25% molar) were obtained from Lipoid and used with no purification. Absolute ethanol was purchase from Sigma. 2.2. Microfluidic devices fabrication The polydimethylsiloxane (PDMS)/glass microfluidic devices fabrication was carried out according to Moreira et al. [44],

employing the conventional UV photolithographic and softlithography methods. The designs of the devices’ geometries were projected using AutoCAD (Autodesk). Basically, the mask layouts were photo-plotted with 8000 dpi resolution, and the UV exposures were made in a MJB-3 UV300 contact mask aligner (Karl-Suss, Garching, Germany). Sylgard 184 Silicone Elastomer Kit (Dow Corning, Midland, MI, USA) was used as material precursor of PDMS layers. PDMS channels and glass were irreversibly sealed by oxidizing the surfaces using an oxygen plasma cleaner (Plasma Technology PLAB SE80 plasma cleaner, Wrington, England). All channels had a rectangular cross section with a depth of 100 ␮m and a width of 140 ␮m, as depicted in Fig. 1. Fig. 1A shows a simple microfluidic device (SMD) with a 30-mmlong single straight channel. Fig. 1B shows the second geometry, named patterned microfluidic device (PMD). The device includes uniform barriers and the total channel is 130 mm long. The barriers are used to form swirling sections that accelerate mixing. 2.3. Plasmid DNA purification We used pVAX-Luc as our plasmid DNA vector model with the luciferase gene as the reporter gene and the human cytomegalovirus (CMV) as our immediate-early promoter [45]. The pDNA was amplified in Escherichia coli bacteria and purified using PureLinkTM HiPure Plasmid pDNA Purification Kit-Maxiprep K2100-07 (Invitrogen). The pDNA quality and quantity were spectrophotometrically measured with an ND-1000 NanoDrop UV-vis spectrophotometer (PeqLab, Erlangen, Germany). 2.4. Cationic liposome production CL production was carried out in a hydrodynamic flow focusing microfluidic device (Fig. 1A), according to Balbino et al. [46]. The lipids EPC, DOTAP and DOPE (50:25:25% molar) were dispersed in ethanol at a lipid concentration of 25 mM. The lipid mixture was loaded into a 1 mL glass syringe, and deionized water was loaded into two 3 mL glass syringes, all at room temperature. The syringes were mounted on syringe pump (Kd Scientific, model 200, USA) that controlled the flow of each solution into the inlet channels. The lipid dispersion in ethanol flowed through the center inlet channel, and the deionized water flowed through the two side inlet channels that hydrodynamically compressed the central stream. In the present study for producing CL, we used a Flow rate ratio (FRR) of 10, a lipid concentration in the central stream ethanol dispersion of 25 mM, an average fluid flow velocity (Vf ) of 143 mm/s and room temperature. 2.5. Preparation of pDNA/CL complexes We carried out complex preparation with both microfluidic devices and the bulk mixing method. The pDNA/CL complexes obtained by all methods were prepared at a cationic lipid/pDNA molar charge rate (R+/− ) of 6. This R+/− was found to be best for in vitro transfection in HeLa cells [29]. We compared two microfluidic devices (Fig. 1) to verify the effect of geometry on the physicochemical and biological properties of the complexes. The pDNA and liposomal solutions were loaded into sterile 1 and 3 mL glass syringes, respectively. The syringes containing the pDNA and the CL were placed in two separate programmable infusion syringe pumps (Kd Scientific, model 200, USA) that were adjusted to the desired volumetric FRR between the pDNA and CL. The pDNA solution was injected in the central stream and hydrodynamically compressed by two liposomal streams, as illustrated in the inset of Fig. 1. The FRR of the liposomal solution stream to the pDNA solution stream in all experiments was 5. A Peltier thermoelectric system (Watronix, Inc.) was used to maintain the temperature at approximately 4 ◦ C as the cooling platform was placed under the

T.A. Balbino et al. / Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

205

Fig. 1. Schematic diagram of (A) A simple microfluidic hydrodynamic focusing device with inset illustrating the hydrodynamic focusing (SMD). (B) A patterned microfluidic device and the inset illustrating the geometry and dimensions of the barriers (PMD). The microchannels are 140 ␮m wide and 100 ␮m deep.

microfluidic devices. In the bulk mixing method, the complexes were prepared by adding the appropriate amount of pDNA to CL with vortex mixing [27,28]. 2.6. Physicochemical characterization 2.6.1. Average hydrodynamic diameter and polydispersity We measured polydispersity index (PdI) and the average hydrodynamic diameter of the complexes with a Dynamic light scattering (DLS) technique. We used a Zetasizer Nano ZS (Malvern) with a backscattering configuration and detection at a scattering angle of 173◦ , with a He/Ne laser emitting at 633 nm and a 4.0 mW power source. 2.6.2. Zeta potential () measurement The  was obtained by measuring the velocity of the electrophoretic mobility of the particles using laser Doppler anemometry. The measurements were performed in triplicate for each sample at 25 ◦ C, in water, using Malvern Zetasizer 3000 (Malvern) equipped with a standard 633 nm laser. 2.6.3. Morphology The morphology of the pDNA/CL complexes obtained with the microfluidic devices were observed by means of Transmission electron microscopy (TEM) and the negative staining method. Approximately 5 ␮L of each sample were put on a copper grid coated with a carbon film. Excess liquid was removed with blotting paper. Negative staining was performed with a 2% solution of ammonium molybdate for ∼10 s and then blotted dry. The electron microscopic model was LEO 906E (Zeiss, Germany) at 80 kV, the camera was a MEGA VIEW III/Olympus and the software was ITEM E 23082007/Olympus Soft Imaging Solutions GmbH. 2.6.4. Gel retardation assay We used an agarose gel electrophoresis technique to verify whether the pDNA/CL complexes prepared by bulk and microfluidic methods were able to compact the pDNA into the liposomal structures. The pDNA/CL complexes prepared at a molar charge ratio of 6 (containing 1 ␮g of pDNA) were electrophoresed through a 0.8% agarose gel in 40 mM TRIS-acetate buffer solution and 1 mM EDTA (TAE 1X) at 60 V for ∼3 h and stained for 30 min in a 0.5 ␮g/mL ethidium bromide solution. The location of the pDNA was then visualized and photographed using an ultraviolet image acquisition system.

2.6.5. Plasmid DNA accessibility We evaluated the pDNA accessibility to the fluorescence probe using a double-stranded pDNA quantification kit (Quant-iT PicoGreen dspDNA assay, Invitrogen) in accordance with manufacturer’s instructions. Briefly, a working solution was prepared by diluting the Pico Green stock solution 200 times in TE buffer (10 mM Tris–HCl/1 mM EDTA, pH 7.5). One hundred microliters of the working solution was added to the same volume of complexes and then incubated for 3 min. The intensity of the fluorescence was measured using a plate fluorimeter (Gemini XS, Molecular Device) using excitation and emission wavelengths of 485 and 525 nm, respectively. The fluorescence intensity was expressed as the relative fluorescence intensity, using free pDNA as the positive control and taken as the 100% fluorescence level. 2.7. Culture and transfection of mammalian cells The biological evaluation of the complexes obtained by the different methods was assessed by in vitro transfection in human epithelial carcinoma (HeLa) cells. Cells were cultivated in F-12 (Ham) nutrient mixture (Gibco, UK), containing 10% (v/v) fetal bovine serum (FBS) (Gibco, UK) and supplemented with nonessential amino acids (Gibco, UK), gentamicin (Gibco, UK), sodium pyruvate (Gibco, UK), and antibiotic–antimycotic (Gibco, UK). Cultures were incubated at 37 ◦ C under 5% CO2 for 24 h in a humid atmosphere, allowed to grow near confluency and harvested with trypsin. After reaching confluence, cells were seeded into 24-well culture plates (5 × 104 cells per well). The cells were then incubated for 48 h and transfected with 0.8 ␮g of pVAX-Luc per well. The medium was replaced 6 h after the transfection to remove plasmids not internalized by cells. Twenty-four hours post transfection, the medium was removed, the cells were washed using PBS, lysed, and luciferase activity was determined according to the Promega Luciferase Assay protocol. The Relative light units (RLUs) were measured with a chemiluminometer (Lumat LB9507, EG&G Berthold, Germany). The total protein was measured with to a BCA protein assay kit (Pierce), and luciferase activity was expressed as RLU/mg of protein. 2.8. Statistical analysis The data were expressed as the mean of triplicates ± standard deviation (SD). When indicated, the standard deviation was obtained based on the Propagation Uncertainties Theory [47].

206

T.A. Balbino et al. / Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

Statistical significance was determined using the two-tailed Student’s t-test at a confidence level p < 0.05. 3. Results and discussion The primary motivation for our work was to investigate a continuous process using microfluidic devices that produce pDNA/CL complexes for gene delivery and vaccine therapy. We used two different microfluidic devices: one with a simple straight channel with hydrodynamic focusing, as shown in Fig. 1A (SMD), and another with a patterned channel (Fig. 1B, PMD), which has contraction and expansion regions through regular blocks in the microchannel walls. Those contractions and expansion regions generate turbulence or eddies and mixing occurs through chaotic advection [48,49]. As a comparison, we formed the complexes with a bulk mixing method. In our comparative study, we investigated the pDNA/CL bulk complexation at the same R+/− of 6. This R+/− was previous identified as the best range for in vitro HeLa transfection of EPC/DOTAP/DOPE liposomes complexed with pDNA using the bulk method [29]. Because the pDNA/CL molar charge proportion was previously established, the FRR between the pDNA and CL streams were constant. The investigated parameters were thus the complexation temperature and the total fluid flow velocity. We introduced the pDNA stream at the central inlet in both microfluidic devices (Fig. 1). This central inlet of pDNA stream was previously evaluated by Jellema et al. [40] and by Otten et al. [37]. 3.1. Effect of temperature control on complexation process between CLs and pDNA in microfluidic devices Due to the spontaneous, fast and irreversible electrostatic interactions between positively charged lipids and negatively charged pDNA, we hypothesized that the processing temperature in the microchannel could influence complexation and affect the formation of the complexes. We assessed the effect of temperature (at 25 ◦ C and 4 ◦ C) on particle size and polydispersity of the complexes (Fig. 2). As observed in Fig. 2, the temperature control was significant only for the SMD for both investigated parameters (particle size and PdI). No significant difference was observed when we changed the process temperature for production of pDNA/CL complexes using the BM and the PMD, where the latter has a longer microchannel so that complexes have a longer mixing time compared to SMD. We separated the complexation methods into two groups: (i) group I with a high average diameter and polydispersity (BM at 4/25 ◦ C and SMD at 25 ◦ C), and (ii) group II with a low average diameter and polydispersity (SMD at 4 ◦ C and PMD 4/25 ◦ C). The influence of the temperature during the complexation process between pDNA and CL was reported by Wasan et al. [31] using a repeated slow mixing pipetting of equivolumetric suspensions at a typical total volume of less than 1 mL. These authors showed that the sizes of formed complexes could be controlled by equilibrating the CL and pDNA solutions prior to mixing at temperatures between 2 ◦ C and 7 ◦ C. Because pipette mixing is inefficient and usually poorly reproducible, the complexation temperature likely exerts a great influence on the size of the aggregates, as it most likely decreases the velocity of the complexation reaction. As with pipette mixing, vortex system mixing is generally disadvantageous because the fluid moves circularly in streamlines, so there is little mixing between fluids at different tank heights [50]. As a consequence, the average diameter and PdI of complexes obtained at 4 and 25 ◦ C were high and classified in group I. In addition to the high average diameter, we could not identify differences in the process temperature for BM, most likely due to the low lipid

Fig. 2. Effect of temperature on (A) particle size and (B) polydispersity index of the complexation process between CLs and pDNA for different methods. Average fluid flow velocity of the microfluidic devices of 140 mm/s. The flow rate ratio was maintained at 5. Error bars correspond to SD of three independent experiments. *Significantly different (p < 0.05) when compared between the pairs. **Significantly different (p < 0.05) when compared to the bulk mixing method at 4 or 25 ◦ C.

and pDNA concentrations. In addition, a vortex mixer is a procedure that cannot be used in case of increasing the volumetric production. Similar temperature behaviors to those described by Wasan et al. [31] were identified for the SMD. Because the mixing in hydrodynamic focusing depends basically on diffusive gradients and convective flow, the temperature has to be decreased for proper size control. When the pDNA/CL complexation was carried out in the PMD, there was an enhancement in the mixing performance as previously described in simulation studies, based on Computational Fluid Dynamics using numerical methods and algorithms to investigate the behavior of the fluid flow throughout similar microchannels [48,49]. In this latter case, the channel geometry with obstacles allows the formation of swirling vortexes or eddies, increasing the contact area between the two fluids and reducing the diffusion distance to achieve a uniform concentration. Thus, with higher mixing quality, the temperature exerts no significant influence, as identified in Fig. 2. These results show that the complexation between CL and pDNA depends on a balance between the mixing pattern and temperature. Cases of poor mixing with microfluidic devices were associated with decreases in the process temperature in the same manner as previously described by Wasan et al. [31], as identified using the SMD. However, when the PMD was used, the mixing was enhanced and there was no need for temperature control. As a consequence, this PMD process was more robust than the SMD process. In previous work [27], we found that the structural conformation of the pDNA/CL complexes formed by the conventional bulk method at a molar charge ratio of R+/− 6, with the same lipid composition and pDNA, was composed of two juxtaposed cationic vesicles bounded by the negative pDNA. We believe that this conformation remains the same when the complexes are formed using microfluidic methods. Unlike, the results in a previous study [35], the vesicles became multilamellar after complexation with pDNA in the

T.A. Balbino et al. / Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

207

the transition range of the characteristic flow, i.e., in all studied ranges of Vf the fluid flow was laminar. The walls in microfluidic devices have a strong influence on particle motion [56] and thus on the wall shear rate (S). The wall shear rate between parallel plates is given by the following equation: S=

QT 40 W d2

(2)

where QT is the total volumetric flow rate in terms of mL/min, W is the microchannel width in terms of cm, and d is the halfmicrochannel depth in terms of cm [53]. The shear rate is presented in Table 1. Considering the PMD had two distinct regions (one with 140 ␮m and a second one with 70 ␮m), we calculated the maximum shear rate. We can observe that the PMD had regions with twice the shear rate of SMD. This additional information confirms the observations of better mixing and no influence on temperature for pDNA/CL complex production with the PMD. 3.3. Physicochemical properties and morphology of pDNA/CL complexes

Fig. 3. Particle size and polydispersity index of the complexes formed in simple and patterned microfluidic devices as function of mean fluid flow velocity in the center outlet channel (the combined flow velocities of all inlets) at 4 ◦ C. The flow rate ratio was maintained at 5. The error bars represent the standard error of independent triplicates.

outlet microchannel. These authors studied a binary liposomal formulation composed of DOTAP and DOPC (dioleoylphosphatidylcholine), which may have produced different results because it is well known that the use of a third lipid (e.g., EPC, as employed in the present work) can preserve the liposomal structure even in the presence of pDNA [24,27]. 3.2. Effect of average fluid flow velocity (Vf ) on particle size of complexes We investigated the effects of Vf , or total volumetric flow rate, on complex size and polydispersity over five different conditions. In this case, all experiments were carried out at 4 ◦ C. As observed in Fig. 3, the particle size and the polydispersity remain approximately unaffected by the fluid flow velocity in the studied range. Previous reports also indicated that the Vf has little impact on particle formation [21–23,46,51,52]. This suggests that the fluid residence time in the microchannel for forming the complexes was sufficient even for higher Vf . The explored ranges of Vf had residence times between 0.3 and 0.06 s, based on the minimum and maximum Vf , respectively. The fluid flow velocity is a microfluidic parameter directly proportional to the Reynolds number (Re). Re (Eq. (1)) represents the ratio of inertial forces to viscous forces and consequently quantifies the influence of these forces for a given flow conditions [53,54]. Re =

Vf Dh 

(1)

where Dh is the hydraulic diameter,  is the fluid density and  is the dynamic viscosity. The Re values are presented in Table 1. Despite having distinct dimensions when compared to macroscale equipment, the transition from laminar to turbulent flow in microchannels is similar. The onset of transition from laminar to turbulent flow on the macroscale has a minimum lower critical Re between 1800 and 2300 and such is the case for channels with diameters between 50 and 247 ␮m [55]. Thus, as summarized in Table 1, in our microfluidic system, the variation in Re did not reach

To characterize the differences among pDNA/CL complexes produced by SMD and PMD, we selected the process condition of Vf 140 mm/s, an FRR of 5 and process temperature 4 ◦ C. The pDNA/CL complexes were obtained and characterized for particle size, PdI and  (Table 1). The bulk method was evaluated for comparison. The pDNA/CL complexes produced by both microfluidic devices as well as the bulk method presented positive  in water, as summarized in Table 1. This indicated that the complexes produced by both devices and by the bulk method were cationic and considered suitable for in vitro transfection experiments. The difference in  was most likely due to the modifications of physicochemical properties when shear was applied during the complexation process between pDNA and CL [57–59]. Thus, the difference in  is most likely due to the different shear rates applied; as observed by Balbino et al. [46], higher applied shear rates result in higher  values for producing CL. The morphological characterization of the complexes formed by the microfluidic complexation method, performed by TEM, is presented in Fig. 4. Complexes formed by both microfluidic devices had relatively small and homogeneous particle sizes, which is in good agreement with the DLS results (Fig. 2 and Table 1). The complexation process between pDNA and CL for the different methods was evaluated using a gel retardation assay (Fig. 5). It is possible to see in lane 1 that the free pDNA migrated out of the well toward the cathode and presented two characteristic bands. As no running band of pDNA was observed in lanes 2–4, we can conclude that all complexation methods were able to complex completely pDNA into the CL, indicating that all processes produced stable and intact structures. To better understand the influence of the microfluidic device on the formation of pDNA/CL complexes, we evaluated the pDNA accessibility to the fluorescent probe (Fig. 6). The fluorescent probe specifically binds to double-strand pDNA, and its fluorescence increases proportionally to the quantity of non-electrostatically bound pDNA in the liposomal structures [60]. We can observe that pDNA/CL complexes produced by the bulk SMD mixing methods presented similar relative fluorescence values of 25.2 ± 1.1 and 24.5 ± 5.0%, respectively. However, when we used the PMD, the relative fluorescence decreased by 44% (Fig. 6). Considering that we processed the pDNA/CL at the same conditions and all pDNA was incorporated into the CL (Fig. 5), the lower relative fluorescence for the PMD indicates a lower quantity of non-electrostatically bound pDNA in the liposomal structures. This shows that the PMD can increase the association between the CL and electrostatically bound pDNA, modifying the structural organization of the nucleic acid

208

T.A. Balbino et al. / Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

Table 1 Physicochemical and flow properties of pDNA/CL complexes obtained by the bulk mixing and microfluidic methods with Vf of 140 mm/s and at 4 ◦ C. Complexation method Bulk mixing SMD PMD a b c d e

Reynolds number c

na 16.3 23.5

Maximum wall shear rate (104 s−1 ) c

na 0.9 ± 0.1d 1.7 ± 0.3d

Particle sizea (nm)

PdI

 b (mV)

141.4 ± 7.2 119.8 ± 1.6e 118.8 ± 7.9e

0.29 ± 0.03 0.22 ± 0.02e 0.23 ± 0.03

52.7 ± 0.7 53.6 ± 1.7 64.5 ± 1.4e

Intensity-weighted averaged hydrodynamic diameter. Zeta potential measured in water at 25 ◦ C. Not applied. Standard deviation based on the Propagation Uncertainties Theory [43]. Significantly different (p < 0.05) when compared to the bulk mixing method.

into the liposomal structures. The higher shear and intense mixing applied in the PMD are most likely the reasons for different structural pDNA organization in the liposomes. Differences in relative fluorescence were previously reported using DOTAP/DOPE and dehydrated–hydrated vesicles (DRVEPC/DOTAP/DOPE). DOTAP/DOPE liposomes presented lower relative fluorescence than DRV-EPC/DOTAP/DOPE [26]. In this reported case, the lipid composition and most likely the production method influenced the quantity of non-electrostatically bound pDNA into the liposomal structures. pDNA-DOTAP/DOPE complexes presented 10% of the relative fluorescence. In our case, the PMD presented a relative fluorescence of 13.8 ± 1.1%, similar to pDNA/DOTAP/DOPE complexes. To our knowledge, there have

Fig. 5. Electrophoretic mobility retardation assay of pDNA complexed with cationic liposomes with different methods. Lanes: (1) free pDNA; complexation processes: (2) bulk mixing; (3) simple microfluidic device; and (4) patterned microfluidic device. Average fluid flow velocity of the microfluidic devices of 140 mm/s, at 4 ◦ C and an FRR of 5.

been no related studies regarding differences in association with pDNA and CL using different microfluidic devices. 3.4. In vitro HeLa transfection The ability of pDNA/CL complexes obtained by BM and microfluidic methods to transfect HeLa cells in vitro was assessed (Fig. 7) based on the previous physicochemical characterization and the established process conditions.

Fig. 4. TEM images of pDNA/cationic liposomes complexes produced by (A) simple microfluidic device and (B) patterned microfluidic device at flow rate ratio of 5, at 4 ◦ C and Vf of 140 mm/s. Scale bars indicate 500 nm.

Fig. 6. Relative fluorescence obtained by the probe in ultra-pure water. The results with free pDNA used as control for complexes pDNA/cationic liposomes formed by three different methods: bulk and in the simple and patterned microfluidic devices. The operational conditions of the microfluidic process were average fluid flow velocity of 140 mm/s at 4 ◦ C and an FRR of 5. Error bars correspond to the SD of three independent experiments (n = 3). * p < 0.05 was considered significantly different when compared to other samples.

T.A. Balbino et al. / Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

209

References

Fig. 7. In vitro efficacy on HeLa cells transfection of pDNA/CL complexes formed by the bulk mixing and microfluidic methods. Activity was expressed in RLU/mg protein. The pDNA/CL complexes were obtained applying average fluid flow velocity of the microfluidic devices of 140 mm/s at 4 ◦ C and an FRR of 5. Each data represents SD of three experiments, * p < 0.05 was considered significantly different. NS: not significant (p < 0.05).

As observed in Fig. 7, the biological effects of complexes obtained by both microfluidic devices proved in vitro in HeLa cells. It is possible to see that free pDNA did not transfect cells (or had a very small effect) and the complexes prepared by the PMD presented a lower transfection efficacy for the period of time during which the transfection was conducted. This may be due to the limited accessibility that the probe had to the pDNA for these complexes in the fluorescence study. The high quantity of electrostatically bound pDNA in the liposomal structures in the PMD (Fig. 6) most likely influenced the pDNA release to cell, impairing the transfection process for the evaluated transfection time. The complexes formed in the simple microfluidic device yielded similar transfection levels as the complexes formed in the bulk mixing methods, suggesting the same pDNA/lipid organization, in accordance with the same level of pDNA accessibility to the probe in the fluorescence study. 4. Conclusion In conclusion, we have demonstrated the feasibility of employing two different microfluidic devices to produce pDNA/CL complexes on a nanometric scale. It was demonstrated that the microfluidic complexation methods provided better control of particle size than the bulk mixing complexation method. The process temperature is a factor that must be investigated based on the type of microfluidic device. Additionally, the use of different microfluidic devices can control the mode of association between pDNA and CL, which is reflected in different biological results. Moreover, with a higher shear rate and better mixing, the association between pDNA and CLs is higher. Acknowledgements The authors gratefully acknowledge the financial support of the Fundac¸ão de Amparo à Pesquisa do Estado de São Paulo – FAPESP (São Paulo, Brazil) and Universidade Estadual de Campinas – UNICAMP (Campinas, Brazil). The TEM analyses were performed at Instituto Butantan (São Paulo, Brazil), and the microfluidic devices were constructed at Laboratório de Microfabricac¸ão (LMF) of Laboratório Nacional de Luz Síncrotron (LNLS).

[1] G.M. Whitesides, The origins and the future of microfluidics, Nature 442 (7101) (2006) 368–373. [2] T.M. Squires, S.R. Quake, Microfluidics: fluid physics at the nanoliter scale, Rev. Mod. Phys. 77 (3) (2005) 977–1026. [3] L. Draghiciu, et al., Manipulation of nanoparticles within a microfluidic system based on SU-8 polymer for bio-applications, Mater. Sci. Eng. B: Adv. Funct. Solid-State Mater. 169 (1–3) (2010) 186–192. [4] J. Baek, et al., Investigation of indium phosphide nanocrystal synthesis using a high-temperature and high-pressure continuous flow microreactor, Angew. Chem. Int. Ed. 50 (3) (2011) 627–630. [5] T. Yamashita, et al., Cultivation and recovery of vascular endothelial cells in microchannels of a separable micro-chemical chip, Biomaterials 32 (10) (2011) 2459–2465. [6] J.E. Reiner, et al., Accurate optical analysis of single-molecule entrapment in nanoscale vesicles, Anal. Chem. 82 (1) (2010) 180–188. [7] H. Kurita, et al., Fluorescence observation and manipulation of individual DNA molecules in a microfluidic channel, in: International Symposium on MicroNanoMechatronics and Human Science (MHS), 2010. [8] A. Terray, J. Oakey, D.W.M. Marr, Microfluidic control using colloidal devices, Science 296 (5574) (2002) 1841–1844. [9] C.Y. Lee, et al., Microfluidic mixing: a review, Int. J. Mol. Sci. 12 (5) (2011) 3263–3287. [10] V. Mengeaud, J. Josserand, H.H. Girault, Mixing processes in a zigzag microchannel: finite element simulations and optical study, Anal. Chem. 74 (16) (2002) 4279–4286. [11] F. Parvizian, M. Rahimi, M. Faryadi, Macro- and micromixing in a novel sonochemical reactor using high frequency ultrasound, Chem. Eng. Process. 50 (8) (2011) 732–740. [12] C.Y. Lim, Y.C. Lam, Analysis on micro-mixing enhancement through a constriction under time periodic electroosmotic flow, Microfluid. Nanofluid. 12 (1–4) (2012) 127–141. [13] J.S. Kuo, D.T. Chiu, Controlling mass transport in microfluidic devices, Ann. Rev. Anal. Chem. 4 (4) (2011) 275–296. [14] Z. Fan, et al., Evaluation of the performance of a constructal mixer with the iodide-iodate reaction system, Chem. Eng. Process. 49 (6) (2010) 628–632. [15] A.D. Stroock, et al., Chaotic mixer for microchannels, Science 295 (5555) (2002) 647–651. [16] R.H. Liu, et al., Passive mixing in a three-dimensional serpentine microchannel, J. Microelectromech. Syst. 9 (2) (2000) 190–197. [17] T.-i. Kim, S.W. Kim, Bioreducible polymers for gene delivery, React. Funct. Polym. 71 (3) (2011) 344–349. [18] S. Ozbas-Turan, et al., Co-encapsulation of two plasmids in chitosan microspheres as a non-viral gene delivery vehicle, J. Pharm. Pharm. Sci. 6 (1) (2003) 27–32. [19] K. Corsi, et al., Mesenchymal stem cells: MG63 and HEK293 transfection using chitosan-DNA nanoparticles, Biomaterials 24 (7) (2003) 1255–1264. [20] D.D. Lasic, Liposomes: from Physics to Applications, Elsevier, Amsterdam, New York, 1993. [21] A. Jahn, et al., Microfluidic mixing and the formation of nanoscale lipid vesicles, ACS Nano 4 (4) (2010) 2077–2087. [22] A. Jahn, et al., Microfluidic directed formation of liposomes of controlled size, Langmuir 23 (11) (2007) 6289–6293. [23] A. Jahn, et al., Controlled vesicle self-assembly in microfluidic channels with hydrodynamic focusing, J. Am. Chem. Soc. 126 (9) (2004) 2674–2675. [24] P. Pradhan, et al., A facile microfluidic method for production of liposomes, Anticancer Res. 28 (2A) (2008) 943–947. [25] J.M. Zook, W.N. Vreeland, Effects of temperature, acyl chain length, and flowrate ratio on liposome formation and size in a microfluidic hydrodynamic focusing device, Soft Matter 6 (6) (2010) 1352–1360. [26] L.G. de la Torre, et al., The synergy between structural stability and DNAbinding controls the antibody production in EPC/DOTAP/DOPE liposomes and DOTAP/DOPE lipoplexes, Colloids Surf. B: Biointerfaces 73 (2) (2009) 175–184. [27] R.S. Rosada, et al., Protection against tuberculosis by a single intranasal administration of DNA-hsp65 vaccine complexed with cationic liposomes, BMC Immunol. 9 (2008). [28] R.S. Rosada, et al., Effectiveness, against tuberculosis, of pseudo-ternary complexes: peptide-DNA-cationic liposome, J. Colloid Interface Sci. 373 (2012) 102–109. [29] T.A. Balbino, et al., Correlation between physicochemical and structural properties with in vitro transfection of pDNA/cationic liposomes complexes, Langmuir 28 (2012) 11535–11545. [30] D. Pozzi, et al., How lipid hydration and temperature affect the structure of DC-Chol-DOPE/DNA lipoplexes, Chem. Phys. Lett. 422 (4–6) (2006) 439–445. [31] E.K. Wasan, et al., A multi-step lipid mixing assay to model structural changes in cationic lipoplexes used for in vitro transfection, Biochim. Biophys. Acta: Biomembr. 1461 (1) (1999) 27–46. [32] S.C. De Smedt, J. Demeester, W.E. Hennink, Cationic polymer based gene delivery systems, Pharm. Res. 17 (2) (2000) 113–126. [33] J. Rejman, et al., Size-dependent internalization of particles via the pathways of clathrin- and caveolae-mediated endocytosis, Biochem. J. 377 (2004) 159–169. [34] B. Ma, et al., Lipoplex morphologies and their influences on transfection efficiency in gene delivery, J. Control. Release 123 (3) (2007) 184–194.

210

T.A. Balbino et al. / Colloids and Surfaces B: Biointerfaces 111 (2013) 203–210

[35] M. Morille, et al., Progress in developing cationic vectors for non-viral systemic gene therapy against cancer, Biomaterials 29 (24/25) (2008) 3477–3496. [36] A.T.-H. Hsieh, et al., Nonviral gene vector formation in monodispersed picolitre incubator for consistent gene delivery, Lab Chip 9 (18) (2009) 2638–2643. [37] A. Otten, et al., Microfluidics of soft matter investigated by small-angle X-ray scattering, J. Synchrotron Radiat. 12 (2005) 745–750. [38] C.G. Koh, et al., Delivery of polyethylenimine/DNA complexes assembled in a microfluidics device, Mol. Pharm. 6 (5) (2009) 1333–1342. [39] C.G. Koh, et al., Delivery of antisense oligodeoxyribonucleotide lipopolyplex nanoparticles assembled by microfluidic hydrodynamic focusing, J. Control. Release 141 (1) (2010) 62–69. [40] R.K. Jellema, et al., Transfection efficiency of lipoplexes for site-directed delivery, J. Liposome Res. 20 (3) (2010) 258–267. [41] D. Chen, et al., Rapid discovery of potent siRNA-containing lipid nanoparticles enabled by controlled microfluidic formulation, J. Am. Chem. Soc. 134 (16) (2012). [42] N.M. Belliveau, et al., Microfluidic synthesis of highly potent limit-size lipid nanoparticles for in vivo delivery of siRNA, Mol. Ther. Nucleic Acids 1 (2012) e37. [43] Y.-P. Ho, et al., Tuning physical properties of nanocomplexes through microfluidics-assisted confinement, Nano Lett. 11 (5) (2011) 2178–2182. [44] N.H. Moreira, et al., Fabrication of a multichannel PDMS/glass analytical microsystem with integrated electrodes for amperometric detection, Lab Chip 9 (1) (2009) 115–121. [45] M.A.S. Toledo, et al., Development of a recombinant fusion protein based on the dynein light chain LC8 for non-viral gene delivery, J. Control. Release 159 (2) (2012) 222–231. [46] T.A. Balbino, et al., Continuous flow production of cationic liposomes at high lipid concentration in microfluidic devices for gene delivery applications, Chem. Eng. J. (2013), http://dx.doi.org/10.1016/j.cej.2013.04.053.

[47] J.R. Taylor, An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements, 2nd ed., 1997. [48] S.H. Wong, et al., Investigation of mixing in a cross-shaped micromixer with static mixing elements for reaction kinetics studies, Sens. Actuators B: Chem. 95 (1–3) (2003) 414–424. [49] S. Naher, et al., Effect of micro-channel geometry on fluid flow and mixing, Simulat. Model. Pract. Theor. 19 (4) (2011) 1088–1095. [50] P.M. Doran, Bioprocess Engineering Principles, Academic Press, London, 1995. [51] X.M. Huang, et al., Ultrasound-enhanced microfluidic synthesis of liposomes, Anticancer Res. 30 (2) (2010) 463–466. [52] C.T. Lo, et al., Controlled self-assembly of monodisperse niosomes by microfluidic hydrodynamic focusing, Langmuir 26 (11) (2010) 8559–8566. [53] R.B. Bird, W.E. Stewart, E.N. Lightfoot, Transport Phenomena, 2nd ed., John Wiley & Sons, New York, NY, 2007. [54] M.C. Ruzicka, On dimensionless numbers, Chem. Eng. Res. Des. 86 (8A) (2008) 835–868. [55] K.V. Sharp, R.J. Adrian, Transition from laminar to turbulent flow in liquid filled microtubes, Exp. Fluids 36 (5) (2004) 741–747. [56] G. D‘Avino, et al., Effects of confinement on the motion of a single sphere in a sheared viscoelastic liquid, J. Non-Newton. Fluid Mech. 157 (1/2) (2009) 101–107. [57] O. Diat, D. Roux, F. Nallet, Effect of shear on a lyotropic lamellar phase, J. Phys. II 3 (9) (1993) 1427–1452. [58] L. Soubiran, et al., Effects of shear on the lamellar phase of a dialkyl cationic surfactant, Langmuir 17 (26) (2001) 7988–7994. [59] M.A. Alvarez, et al., Influence of the electrical interface properties on the rheological behavior of sonicated soy lecithin dispersions, J. Colloid Interface Sci. 309 (2) (2007) 279–282. [60] M.E. Ferrari, et al., Trends in lipoplex physical properties dependent on cationic lipid structure: vehicle and complexation procedure do not correlate with biological activity, Nucleic Acids Res. 29 (7) (2001) 1539–1548.