Direct Incorporation of Lipophilic Nanoparticles into Monodisperse

Sep 4, 2014 - Physical Sciences, Sunnybrook Research Institute,. ‡ ... and treatment of cancer. ...... (CINO) supported by Cancer Care Ontario with ...
0 downloads 0 Views 6MB Size
Article pubs.acs.org/Langmuir

Direct Incorporation of Lipophilic Nanoparticles into Monodisperse Perfluorocarbon Nanodroplets via Solvent Dissolution from Microfluidic-Generated Precursor Microdroplets Minseok Seo† and Naomi Matsuura*,‡ †

Physical Sciences, Sunnybrook Research Institute, ‡Department of Medical Imaging, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada S Supporting Information *

ABSTRACT: Multifunctional medical agents based on imaging or therapy nanoparticles (NPs) incorporated into perfluorocarbon (PFC) droplets are promising new agents for cancer detection and treatment. For the first time, monodisperse PFC nanodroplets labeled with NPs have been produced. Lipophilic, as-synthesized, hydrocarbonstabilized NPs are directly miscibilized into lipophobic PFCs using a removable cosolvent, diethyl ether (DEE), which eliminates the need of the typical time-consuming and expertise-specific NP surface modification steps previously required for NP incorporation into PFCs. This NP-DEE/PFC solution is then used to synthesize monodisperse, micrometer-scale, DEEinfused NP-PFC precursor droplets in water using microfluidics. After precursor microdroplet generation, the DEE cosolvent is removed by dissolution and evaporation, resulting in dramatically smaller, monodisperse, NP-labeled nanodroplets, with final droplet sizes far smaller than the minimum droplet size limit of the microfluidic system, and easily controlled by the amount of DEE mixed in the PFC phase prior to precursor droplet synthesis. Using this technique, unmodified lipophilic quantum dot (QD) NPs were integrated into monodisperse and PFC nanodroplets 165 times smaller in volume than the precursor microdroplets, with dimensions down to 470 nm. The final droplet sizes scaled with the PFC concentrations in the precursor microdroplets, and the QDs remain localized within the droplets after DEE is removed from the system. This method is robust and versatile, and it comprises a platform technology for other unmodified lipophilic NPs and molecules to be incorporated into different types of PFC droplets for the production of new NP-PFC hybrid agents for medical imaging and therapy applications.

1. INTRODUCTION Multifunctional medical agents based on nanoparticle (NP)incorporated perfluorocarbon (PFC) droplets have attracted great interest due to their promise in the improved detection and treatment of cancer. Biocompatible PFC droplets have a long history as in vivo contrast agents in ultrasound, magnetic resonance, and X-ray imaging and as agents in cancer therapy.1−4 Recently, PFC imaging signal detectability and potential therapeutic efficacy for cancer-specific applications were found to be substantially enhanced through the modular incorporation of other, complementary medical agents within the PFC droplet.5 For example, the inclusion of various solid NPs, including quantum dots (QDs),6,7 iron oxide NPs,7,8 Au NPs,7,9 and lead sulfide NPs,10 has expanded the functionality of current PFC agents both in medical imaging (e.g., in optical, photoacoustic, and magnetic resonance imaging)8−14 and in localized therapy.8,10,15 However, a major challenge in the field remains the production of monodisperse and nanoscale NP-incorporated PFC droplets. PFC droplet agents in the nanoscale size range (95 (345)

solvent solubility in waterc

water solubility in solventc

calculated vapor pressured

20 μm at the orifice). Although DEE is known to cause swelling in PDMS,32 monodisperse DEE/PFC microdroplets were successfully generated using PDMS microfluidic devices, after equilibrating the PDMS microfluidic device for at least 10 min before starting the experiments, after which the microfluidic device became stable for the duration of the experiment. The production of smaller initial (Di < 20 μm) microdroplets with narrow size distribution was carried out using a quartz microfluidic device, with an orifice width and a channel height of 17 and 14 μm, respectively. In both microfluidic devices, the flow rates of the continuous phase (Qc) and dispersed phase (Qd) were varied from 0.1 to 2.0 and 0.01 to 0.2 mL/h, respectively. The system was equilibrated for a minimum of 3 min after each change in flow rate. The continuous and dispersed phases were supplied to the channel through polytetrafluoroethylene tubing attached to syringes operated by digitally controlled syringe pumps (PHD Ultra Syringe Pumps, Harvard Apparatus (Holliston, MA)). Hamilton GASTight glass syringe was used to inject the dispersed phase. A small magnetic bar (3 mm diameter with 10 mm length) with stirrer was used to mix the dispersed phase during the experiments. To stabilize DEE/PFC microdroplets inside of the channel and collector, Zonyl-FSP (fluorosurfactant), Pluronic F-68, and glycerol purchased from Sigma-Aldrich (Ont., Canada) were used. These surfactants were selected due to their previous use in producing very stable, microfluidic-generated perfluorocarbon droplets.20 The continuous phase (0.5 wt % FSP and 1 wt % F-68 in 20−60 wt % glycerol mixture) was supplied to the two side channels of the microfluidic device, and the dispersed phase (different ratios of the DEE/PFC mixture) was introduced in the central channel. During all experiments, the ambient temperature was maintained at 21.0 ± 1.0 °C, and deionized water (Millipore Milli-Q grade) with a resistivity of 18.2 MΩ was used. Characterization. The generation of the microdroplets was imaged using a Leica DMIL inverted microscope coupled to a highspeed CCD camera (CoolSNAP HQ2, Photometrics, Tucson, AZ) using exposure times of 1−10 μs. A Zeiss Axiovert 200 M inverted epifluorescence microscope equipped with a Hamamatsu ORCA-ER camera (Carl Zeiss Canada Ltd., Ont., Canada) was used to take fluorescence microscope images. The sizes of the microdroplets were

measured optically using Image Pro Plus (Media Cybernetics, Rockville, MD) software. The coefficient of variation (CV = σ/D × 100, where σ is the standard deviation of droplet diameter) for each population of microdroplets was obtained by measuring the dimensions of more than 500 microdroplets. The size distribution and number percent of droplets in solution after collection was measured using a Coulter Counter (Multisizer 3, Beckman Coulter, Inc., Danvers, MA). The hydrodynamic diameters of nanoscale size droplets were obtained using a Malvern Zetasizer Nano-ZS 3000HS (Worcestershire, U.K.) dynamic light scattering (DLS) instrument. The emission wavelength of QDs was measured using a spectrofluorometer (FluoroMax-4, Edison, NJ).

3. RESULTS AND DISSCUSION 3.1. Direct Integration of Unmodified NPs into DEE/ PFC Precursor Microdroplets. First, we determined that DEE could be used as a cosolvent to miscibilize lipophilic NPs without any surface modification into lipophobic PFCs. Perfluorohexanes were selected as the PFC material because they have been extensively characterized in the past as medical imaging contrast agents.33−35 Hydrophobic CdSe/ZnS core/ shell QDs were selected as the model NPs because their strong fluorescence permits easy optical assessment of their spatial distribution in relation to the PFC droplets at submicrometer resolution. DEE was selected as a suitable cosolvent because it is one of the few organic solvents that is fully miscible with PFCs at relatively low temperatures, permitting room-temperature NP incorporation into PFC droplets. DEE has limited water solubility at room temperature (∼9.0 vol %) compared to other simple organic solvents, which can permit the generation of monodisperse droplets and its slow diffusion from the generated droplets into the continuous aqueous phase. Further, DEE has high vapor pressure so that the cosolvent can be subsequently removed by evaporation from the continuous phase. DEE also has the advantage that due to its fluid properties, mixtures of PFC and DEE can generate smaller droplets at the microfluidic device orifice with the same flow rates and the same device dimensions compared to PFC alone. This occurs because the viscosity of the pure PFC dispersed phase can be decreased through the addition of DEE (i.e., from 0.67 to 0.22 cP for PFC and DEE, respectively, at 25 °C) and the interfacial tension of the pure PFC dispersed phase with water can also be decreased through the addition of DEE (i.e., from 57.2 and 10.7 mN/m for PFC in water and DEE in water, respectively).36,37 The temperature-dependent miscibilization of QDs into PFCs using DEE was observed (Figure 2). Below the miscibility temperature (Tmisc = 14 ± 0.1 °C, Table 1), DEE (here, containing QDs) is phase-separated from the PFCs. Above the miscibility temperature, the DEE phase (with the QDs) becomes fully miscibilized with PFC to form a homogeneous NP-DEE/PFC solution. Since the microfluidic device will generate droplets at ambient temperature, the fully mixed NPDEE/PFC solution is suitable for the production of the NPDEE/PFC precursor microdroplets. The monodispersity and minimum size of the final NPloaded PFC droplet population (after DEE removal) depends on the monodispersity and minimum size of the precursor NPDEE/PFC microdroplets generated at the orifice. Therefore, the precursor microdroplets were carefully characterized for size and monodispersity as a function of different flow rate ratios and different DEE concentrations. In general, for a particular combination of continuous and dispersed phase materials, the droplet size and size distribution depend on the 12467

dx.doi.org/10.1021/la502462n | Langmuir 2014, 30, 12465−12473

Langmuir

Article

Figure 2. Temperature-dependent miscibility of unmodified, lipophilic QDs in lipophobic PFC using DEE as a cosolvent. Fluorescence photographs (taken under 368 nm UV illumination) of DEE containing fluorescent QD NPs (i.e., the QD-DEE mixture) mixed with PFC, at temperatures of 21 °C (left) and 4 °C (right). Above the miscibilization temperature (Tmisc = 14 ± 0.1 °C), the QD-DEE solution mixed homogeneously with the PFC into a single phase, while below Tmisc the QD-DEE mixture phase separates from PFC to form two phases.

design and dimensions of the microfluidic device, the flow rates and properties of the continuous and the droplet phases, and the interaction between the continuous and the droplet phases.31,38 As expected, the dimensions and coefficient of variation (CV) of the microdroplets at generation using the microfluidic device used in this study exhibited a strong dependence on the flow rates and flow rate ratios. Specifically, consistent with previous studies,30,39−41 the size of the microdroplets decreased with increasing flow rates of the continuous phase (Qc) using a fixed dispersed phase flow rate (Qd) (Figures 3 and 4). The microdroplet size at the orifice was found to decrease with increasing concentrations of DEE in the DEE/PFC dispersed phase at constant values of Qc and Qd (Figure 4a). The presence of NPs in the DEE/PFC dispersed phase had no effect on the generation rate, size, or size distribution of DEE/ PFC microdroplets, which was expected due to the very low volume concentration of QDs in the DEE phase. The observed microdroplet size decrease with increasing concentration of DEE in the dispersed phase was consistent with previous findings, and was attributed to the decrease in both the viscosity of the dispersed phase and the decrease in the interfacial tension between the dispersed and continuous phases.31,38,41,42 We determined here that monodisperse precursor NP-DEE/ PFC microdroplets could be successfully generated, even at very high DEE concentrations (i.e., up to 99.6 vol %). Previous studies have shown that although lower viscosity materials can result in smaller microdroplets,41 lower viscosity can also increase the CV of the microdroplets, which will ultimately limit the maximum DEE concentration that can be used to generate monodisperse NP-DEE/PFC precursor microdroplets. Indeed, the CVs of the generated microdroplets slightly increased with increasing DEE in the DEE/PFC mixture, and with increasing Qc (Figure 4b). For DEE/PFC microdroplets generated in the jetting regime (i.e., beyond the flow focusing regime when the Qc ≥ 0.8 mL/h), the CVs significantly increased (>9%) such that the generated microdroplets could no longer be considered to be monodisperse (as indicated by the broken line in Figure 4). For pure PFC, the microdroplet generation was stable in the flow focusing regime at Qc > 0.8

Figure 3. Example of monodisperse and size-controlled DEE/PFC precursor microdroplet generation using different flow rates of the continuous phase (Qc) at a constant dispersed phase flow rate (Qd = 0.10 mL/h). The concentration of PFC in DEE/PFC mixture was 46.3 vol %. The flow rates of the continuous phase were (a) 0.2, (b) 0.4, (c) 0.6, and (d) 0.8 mL/h. The height and width of the orifice of the quartz microfluidic device are 14.0 and 17.0 μm, respectively, and the width of the outer downstream microchannel is 500 μm. The scale bar represents 100 μm.

mL/h, so the CVs of the pure PFC microdroplets remained between 1.0 and 3.0%. To assess the capability of this microfluidic device to produce NP-incorporated PFC droplets in the submicrometer regime after DEE removal, we established the smallest monodisperse precursor DEE/PFC microdroplet that could be produced using this microfluidic device. The highest concentration of DEE in the dispersed phase that could be used to generate monodisperse DEE/PFC precursor microdroplets (for the biggest potential size change from generated DEE/PFC microdroplet compared to the final PFC droplet after DEE removal) was ∼99.4 vol % DEE (i.e., ∼0.6 vol % PFC). At this concentration, the smallest monodisperse DEE/PFC microdroplets that were generated at the orifice at Qc = 0.7 mL/h were 2.6 μm, equivalent to pure PFC droplets after DEE removal of ∼0.47 μm in size. 3.2. DEE Dissolution for NP-Incorporated, Monodisperse PFC Droplet Formation. We next characterized the dissolution of DEE from the monodisperse precursor DEE/ 12468

dx.doi.org/10.1021/la502462n | Langmuir 2014, 30, 12465−12473

Langmuir

Article

dissolution and evaporation of DEE happened within the first 5−10 min after collection. The diffusion rate of the cosolvent was found to be dependent on the volume fraction of DEE/PFC droplets in the aqueous, continuous phase. The dynamic dissolution of DEE from the DEE/PFC microdroplets as a function of droplet concentration in water is shown in Figure 5. In Figure 5a, we show optical microscope images of precursor DEE/PFC microdroplets, originally 13.9 μm in size, as a function of time in a 60/40 wt % glycerol/water solution after their transfer to a glass slide exposed to air (see movie clip in the Supporting Information). Glycerol was used in the continuous phase to increase its viscosity (from 1.0 cP for water to 11.0 cP for the 60/40 wt % glycerol/water solution) to slow down the DEE diffusion rate such that the dynamic size change of the DEE/ PFC microdroplets as a function of time could be captured. The microdroplet size was observed to change too quickly to reliably monitor (