|Home | About | Journals | Submit | Contact Us | Français|
Use of microfabricated neural devices is on the rise in medicine today, and their failures in long-term performance have been well documented [1–3]. While some neural devices fail for unforeseen reasons, most lose their function in vivo because they are compromised by reactive cell responses [4–7]. Currently, there are several approaches in the quest to improve the biocompatibility of neural devices. One approach is to develop more biocompatible materials [8–11], while another is to chemically modify the surfaces of these devices to control cell interactions [12–14]. The third approach is to use drug delivery systems for local release of treatment agents [15–18]. For the drug delivery option, polymer coatings adapted to device surfaces are still the predominant form. Although polymer coatings are reliable in providing sustained release of biomolecules, they are not always appropriate because their poor conductivity limits the capacity of the neural device to communicate with neighboring neurons.
This study presents an alternative method for coating neural devices that provide several improvements over current methods. Rather than forming a complete coating that might electrically insulate the surface, polymer nanoparticles (NPs) are assembled individually onto the device surface to potentially improve conductivity. In addition, NPs are capable of providing sustained release of treatment agents from the device surface. To demonstrate the concepts and capabilities of our new technique, we utilized a planar silicon oxide material as our model probe, although we anticipate that other materials such as metals and polymers are also compatible with our proposed system.
Poly(lactide-co-glycolide) (PLGA) was selected for use in this study, because it is a FDA-approved, biocompatible polymer [19–21] commonly used for particle fabrication. With the addition of an anionic surfactant in the formulation, negatively charged PLGA NPs were obtained. These particles readily assembled onto silicon probe surfaces that were modified by adsorption of cationic poly(-L-lysine) (PLL). The extent of NP assembly was tunable by altering the ionic strength of the particle in aqueous suspension. The principle of this coating technique is similar to the multilayering method in colloid sciences, in which thin films are built by assembling alternating layers of oppositely charged polymers [22, 23]. This new system can also be expanded to deliver multiple molecules simultaneously if NPs encapsulating different molecules are mixed prior to attachment. Such versatility of the system is especially advantageous for device coatings meant to improve biocompatibility. Transfection agents, for example, can be loaded into NPs and coated on the same surface as DNA NPs, to be locally released for in vivo transfection of neurons around the implant site to produce proteins that enhance neuron survival. Although several papers in recent years have investigated NPs embedded in hydrogel coatings [17, 24], this is the first to study the controlled attachment of drug-loaded degradable particles.
The objectives of this study were the following: (1) to test several surfactants in fabricating charged PLGA NPs for electrostatic attachment to a device surface, (2) to find the optimal conditions for NP assembly on the surface, (3) to demonstrate that more than one molecule can be released simultaneously from the NP coating, and (4) to characterize the NP coating morphologically. Free fluorescent dyes were loaded into NPs to visualize the NP coatings under fluorescence microscopy. Fluorescently labeled dexamethasone (DEX) was also loaded into NPs to demonstrate drug release from the NP coatings. DEX is a well-known anti-inflammatory glucocorticoid. It has been reported to reduce tissue responses after device implantation by decreasing the number of infiltrating immune cells associated with inflammation [25, 26].
PLGA 50:50 MW 40,000–70,000 was purchased from Birmingham Polymers (Birmingham, AL). Poly(ethylene-alt-maleic anhydride) MW 400,000 (PEMA) was purchased from Polysciences (Warrington, PA). Poly(vinyl alcohol) MW 30,000–70,000 (PVA), cetyltrimethylammonium bromide (CTAB), dichloromethane (DCM), dimethyl sulfoxide (DMSO), N-[2-hydroxyethyl]piperazine-N′-ethanesulfonic acid (HEPES), phosphate buffered saline (PBS), sodium chloride (NaCl), PLL MW 70,000–150,000, rhodamine, fluorescein, and all other chemicals were purchased from Sigma (St. Louis, MO). Fluorescein isothiocyanate labeled dexamethasone (FITC-DEX) was purchased from Invitrogen (Carlsbad, CA). All buffer solutions were prepared with deionized water (18 MΩ) obtained from a Millipore Ultrapure Water System (Milford, MA). Silicon oxide probes were used to simulate microfabricated neural devices in all of the coating experiments. The probes were made by depositing a 1 µm layer of silicon dioxide on a silica wafer by plasma enhanced chemical vapor deposition, then dicing the wafer into 1 mm×10 mm probes.
NPs were fabricated by an emulsion technique. One hundred milligrams of PLGA was dissolved in 2 ml of DCM overnight at room temperature. The polymer solution was then added dropwise to 4 ml of 1% surfactant solution while vortexing and sonicated at 38% amplitude for 3×10 s with a TMX 400 sonic disruptor (Tekmar, Cincinnati, OH). The emulsion was poured into 100 ml of 0.3% surfactant bath and stirred for 3 h at room temperature to evaporate off the organic solvent. The resulting particles were washed with deionized water 3 times and collected by centrifugation at 10,000 RPM. After flash freezing in 5 ml of deionized water at −80 °C and lyophilization at room temperature for 3 days, dried NPs were collected and stored at −20 °C for future experiments.
Blank NPs were made with PVA, PEMA, or CTAB as surfactant. PEMA was later used in all subsequent fabrications of NPs used in coating experiments. To encapsulate molecules in NPs, a desired amount of each molecule was added to the polymer solution in 200 µl aqueous form before sonication. NPs loaded with 1% rhodamine, 1% fluorescein, 2.5% rhodamine and 5% FITC-DEX were fabricated for coating experiments. The theoretical % loading was calculated as the total mass of loaded material divided by the total mass of polymer (times 100 to convert fraction to %).
The mean sizes and size distributions of the NPs were determined by scanning electron microscopy (SEM). Samples were mounted onto an aluminum plate and sputter coated with gold by a Cressington Sputter Coater 108 auto (Ted Pella, Redding, CA) at 40 mA for 30 s, before being examined by an XL-30 Environmental SEM (FEI Company, Hillsboro, OR). Images of the NPs were taken at 5 kV acceleration voltage and 20,000× magnification. The diameters of all NPs were measured from the images by using Image J software (NIH, Bethesda, MD). At least 500 particles were counted for each batch of NPs to obtain an accurate size estimate.
NPs were evaluated by a ZetaPALS analyzer (Brookhaven Instruments, Holtsville, NY). Samples were suspended in 10 mM HEPES buffer solutions at varying pH values and salt concentrations: 1 mM NaCl, pH 5.5; 10 mM NaCl, pH 5.5; 1 mM NaCl, pH 7.4; and 100 mM NaCl, pH 7.4. Zeta potential values were calculated from the measured mobility values using the software provided by the manufacturer. All measurements were performed 10 times or more for each sample.
The number of molecules entrapped within the NPs was determined by a Spectra Max M5 microplate reader (Molecular Devices, Sunnyvale, CA). Five milligrams of NPs were dissolved in 1 ml of DMSO at room temperature overnight, diluted in PBS, and quantified by measuring the fluorescence intensities of the molecules at the appropriate wavelengths: 540 nm excitation/625 nm emission for rhodamine and 495 nm excitation/520 nm emission for fluorescein or FITC-DEX. The total mass of extracted molecules encapsulated in 5 mg of NPs was converted to an experimental percent loading. The encapsulation efficiency was calculated as the experimentally measured loading divided by the theoretical loading (times 100 to express as %).
Silicon oxide probes were sequentially washed in acetone, ethanol, then deionized water. A PLL layer was then deposited by incubating the probes in 0.4 mg/ml of PLL dissolved in HEPES buffer (10 mM HEPES, 100 mM NaCl, pH 7.4) for 30 min at room temperature. The probes were then washed for 10 minutes in buffer to remove unattached excess polymer.
To determine the best concentration of NPs for particle attachment, 0.1 mg/ml, 0.5 mg/ml, or 1 mg/ml of blank NPs were suspended in a buffer solution of 10 mM HEPES and 1 mM NaCl, pH 5.5. PLL coated probes were incubated with the NP suspension for another 30 min at room temperature to allow particle attachment, followed by a 10 minute buffer wash step. Coated samples were vacuum dried overnight and prepared for SEM inspection. Images of the coated probes were taken at 5 kV acceleration voltage and 1,200× magnification.
To determine the best buffer conditions for NP attachment, 1% rhodamine NPs were used in the coating experiments to confirm particle presence by fluorescence microscopy. Rhodamine-loaded NPs were suspended in two different buffers of 10 mM HEPES, 1 mM NaCl, pH 5.5 or pH 7.4. These two buffer conditions were selected for testing because the zeta potentials of NPs in these solutions were the most negative. The PLL coated probes were then incubated in the NP suspensions for 30 minutes, followed by a buffer wash step. Samples were vacuum dried overnight for fluorescence microscopy. Coating results were visualized under an Olympus X71 Fluorescence Microscope (Olympus, Center Valley, PA) with a rhodamine filter (540 nm excitation). Images of the coated probes were taken at 300 ms exposure time and 200× magnification.
Particle density was counted from each fluorescence image using the “spots” option in Imaris software (Bitplane, St. Paul, MN). Initially, each spherical object in the image was recognized by the program as a “spot” – a computer generated representation of the particle with a fluorescent center. With an additional user-defined spot size, the program was able to estimate the number of fluorescent centers in a cluster of fluorescence, thereby giving a more accurate count of the spots. The total number of spots was summed up and considered as the total number of particles on the surface (total particle number divided by surface area of the probe).
Multiple NPs were incubated with a PLL coated silicon probe in a buffer solution (10 mM HEPES, 1mM NaCl, pH 7.4) to attach the particles to a single surface. Two versions of NP coating were prepared for in vitro characterization studies. Version A contained a mixture low-loading NPs (1% rhodamine and 1% fluorescein NPs) at 0.25 mg/ml each. Version B included high-loading NPs (2.5% rhodamine and 5% FITC-DEX NPs) at 0.25 mg/ml each. All coated probes were vacuum dried overnight before being visually examined. The coatings were observed under fluorescence microscopy with a rhodamine filter (540 nm excitation) and a FITC filter (470–490 nm excitation). Images were taken separately under each filter, at 300 ms exposure time and 200× magnification to minimize fluorescence overbleed. Particle densities were counted from the images by Imaris analysis as previously described.
Release studies from probes coated with multiple NPs were conducted in 500 µl of PBS at 37 °C under static conditions for a period of two weeks. At each sample retrieval time point, all of the release samples were collected for spectrophotometry analysis, and the empty vials were replaced with fresh buffer solution for the next time point. Molecules were quantified by their fluorescence intensities at their corresponding wavelengths: 540 nm excitation/625 nm emission for rhodamine and 495 nm excitation/520 nm emission for fluorescein or FITC-DEX.
Surfaces of the coated probes were examined before and after release studies by SEM to see if there were any morphological changes. After the two-week incubation period, samples were removed from PBS, washed with deionized water and vacuum dried overnight before being examined. All images were taken at 5 kV acceleration voltage and 2,500× magnification. The total area of NPs on the probe surface was estimated by Image J analysis, by summing up the areas of all spherical objects on the surface. The surface area coverage (%) was calculated as the total surface area of all spherical objects divided by the total surface area of the probe (times 100 to convert to %).
All samples were prepared and tested in triplicates or more. Sample data is presented as mean ± standard deviation of the mean.
Blank NPs were fabricated using PVA, PEMA or CTAB as the surfactant (Figure 1). SEM micrographs show that NPs fabricated from all three surfactants were spherical in shape and similar in size (~200 nm) (Figure 2A–C). However, since NPs fabricated with PEMA yielded the most negative zeta potential compared to the other two surfactants (Table 1), PEMA was selected as the surfactant for all subsequent NP fabrications. To further characterize NPs fabricated with PEMA, zeta potential values were measured from NPs suspended in various HEPES buffer conditions (Table 2). It was found that raising the pH or salt concentration to nearly physiological conditions resulted in a reduction in the zeta potential, since the adjustment of buffer conditions also introduced additional cations, some of which will adsorb to the NPs. This effect turned out to be an important factor for NP assembly, and the results are reported in the next section. Lastly, SEM imaging showed the surface morphologies of NPs encapsulating rhodamine, fluorescein and FITC-DEX to be spherical, smooth and evenly sized (Figure 2D–F). Particle size for loaded NPs was similar to that of blank NPs (Figure 2B). Percent of encapsulation efficiencies of NPs loaded with the two free dye molecules were slightly higher than that for NPs loaded with FITC-DEX, which has a greater molecular weight (Table 3).
The conditions for NP assembly were optimized by examining the results of various NP concentrations and buffer conditions. SEM results from three NP coating concentrations (Figure 3) show that NPs assembled with low density at the lowest concentration (0.1 mg/ml) and showed heavy aggregation at the highest concentration (1mg/ml). The optimal concentration was found to be 0.5 mg/ml of NPs, which produced an even distribution of NPs on the probe surface.
The pH of the 10 mM HEPES, 1 mM NaCl buffer also affected the particle density on probe surfaces, Fluorescence images from the 1% rhodamine NP coatings are presented in Figure 4. Image analysis calculated the particle density for pH 5.5 buffer to be 10,200 ± 400 particles/mm2, while the buffer at pH 7.4 had 27,900 ± 4,900 particles/mm2, nearly a three-fold increase. Recalling that the zeta potential value was lower in magnitude at pH 7.4 (−40 ± 3 mV) than at pH 5.5 (−52 ± 3 mV), NPs seemed to assemble more at less negative zeta potential.
For assembly of multiple types of NPs on a single surface, particle counts by image analysis were consistent with the results reported above for the attachment of just a single type of NPs. Images of 1% rhodamine + 1% fluorescein NP coatings were taken individually under a rhodamine or FITC filter. Merging of the two images demonstrated the distribution of dual sources of fluorescence on the same probe surface (Figure 5). Density counts were 13,400 ± 2,700 particles/mm2 for rhodamine loaded and 10,900 ± 2,000 particles/mm2 for fluorescein loaded NPs, or 24,300 ± 4,700 combined total particles/mm2. The ratio of rhodamine to fluorescein loaded NPs was nearly 1:1, which was the same ratio in the suspension buffer.
Release kinetics from multiple NP coatings is consistent with typical controlled release profiles from NPs, with an initial burst in the first few days and a steady release in the later time points. Specifically, the release profile for NPs of low loadings (Figure 6A, 1% rhodamine + 1% fluorescein) shows that the two dye molecules exhibited distinct release profiles and can be measured separately. On the other hand, release from NPs of high loadings (Figure 6B, 2.5% rhodamine + 5% FITC-DEX), demonstrates a greater initial burst due to the greater mass loaded into the NPs. The successful measurement of drug from FITC-DEX coatings showed that drug molecules could also be released from the NP assembly system. SEM examination of surfaces with NP coatings before and after release showed that particles remained attached to surfaces after 14 days of exposure to buffer solution, with little change in their surface morphologies (Figure 7). The percentage of total surface area covered by NPs was maintained at ~4% after two weeks.
Because NPs at a less negative zeta potential were found to attach at higher coverage, more conditions were tested for optimization of NP assembly. First, since a single layer of PLL-NPs was saturated, multilayering of PLL-NPs was attempted to increase the NP density on surface (Figure 8). However, the layering technique resulted in heavy NP aggregation. Attaching NPs when zeta potential was lower in magnitude than −40 ± 3 mV yielded better results (Figure 9). NPs were suspended in 10 mM HEPES, 100 mM NaCl, pH 7.4 buffer solution, with a corresponding zeta potential value of −15 ± 2 mV. Image analysis estimated the surface area covered by NPs to be ~13%, over a threefold increase from the previous 1 mM NaCl suspension condition.
Zeta potential is a function of a particle’s surface charge. When a charged particle is suspended in aqueous solution, a counter-ion cloud will form around the particle to compensate for the potential difference. This cloud layer is also known as the electrical double layer – an inner region where the ions are strongly bound to the surface (the Stern layer) and an outer region where ions are loosely associated (the diffusive layer). When the particle moves in solution, part of the diffusive layer remains behind in bulk solution. The potential at this boundary within the diffusive layer, or shear plane, is the zeta potential . The magnitude of the zeta potential gives an indication of the stability of the system1; a large magnitude means that particles will tend to repel each other and not aggregate in solution. Therefore, NPs with large zeta potential values were desired for strong attachment of NPs to an oppositely charged surface and for even distribution of NP coatings. However, if the zeta is too negative, then particle-to-particle repulsion will limit surface coverage, so a balance is needed.
The chemistry of the surfactant plays a role in the overall measured zeta potentials of polymer NPs. During the fabrication process, the surfactant molecules are presumably incorporated into the surface of the hydrophobic polymer NPs. If the surfactant has charged groups, then the surface of the NPs is covered by these groups and the NPs become charged (Figure 1). The typical surfactant used in PLGA NP fabrication, PVA, contains hydroxyl groups and produces NPs with weakly charged hydroxyl groups associated with the surface. Thus, the zeta potential for NPs fabricated with PVA is reported to be low in magnitude [28, 29]. In order to attach NPs to a surface by electrostatic interactions, NPs with strong zeta potentials were hypothesized to be the best candidates. Since it was previously reported that PEMA produces particles with carboxyl groups on the surface [30–32], PEMA was selected as a more strongly charged surfactant to test. CTAB was also selected for testing because it was previously demonstrated to produce strong positively charged NPs due to its amine groups [33, 34].
Substituting PVA with PEMA or CTAB, blank NPs were fabricated and characterized accordingly. As anticipated, NPs fabricated with PEMA exhibited the most negative zeta potentials; zeta potentials stayed strong despite the change in suspension buffer solution conditions. These NPs were initially negatively charged at −52 ± 3 mV in 1 mM NaCl buffer solution. That value decreased slightly to −35 ± 4 mV when the salt concentration of the buffer solution was raised to 10 mM NaCl. This reduction in zeta potential was due to the additional sodium ions introduced into the buffer solution during the salt adjustment. More cations were available to adsorb to the surface of the NPs, thereby shifting the shear plane and altering the zeta potential. This pattern was also observed with NPs fabricated with PVA or CTAB, although the results were more difficult to interpret due to the low starting values of these NPs (−6 ± 2 mV and + 5 ± 2 mV respectively). When suspended in 10 mM NaCl solution, these NPs exhibited almost zero zeta potentials and were therefore eliminated from any future NP coating experiments. The disappointing data on CTAB could be explained by the fact that the surfactant was not a polymer chain like PVA or PEMA and thus was not stably incorporated into the surface layer of NPs. Furthermore, in prior literature, CTAB was typically used to fabricate micron-sized particles, which have much larger surface areas available for CTAB surface association.
To further characterize NPs fabricated with PEMA, zeta potentials were measured from NPs suspended in various HEPES solution conditions. In addition to the salt concentration, pH was also adjusted. Similar effects on zeta potentials were observed with pH adjustments, which is mainly attributed to the NaOH used in pH adjustment. NPs used in the coating experiments were fabricated with PEMA and encapsulated free rhodamine dye, fluorescein dye, or fluorescently labeled DEX. Percent of encapsulation efficiencies for all three types of NPs were high and similar in values, given that the molecules loaded were all small in size (only hundreds of Daltons). NPs loaded with molecules were similar in shape and size compared to blank NPs, as exemplified by their SEM images.
For NP assembly on silicon dioxide surfaces, PLL was tested and found necessary as a priming layer for particle attachment (data not presented). The concentration of PLL was set at 0.4 mg/ml, a value taken from our previous work with PLL multilayers [35, 36]. The concentration of NPs in suspension was optimized to 0.5 mg/ml, because aggregation of particles was observed at higher concentrations. This phenomenon was apparent with NPs in suspension without the probe. Most likely the hydrophobic effect of the NPs (which favors aggregation) was dominant over the charge effect of NPs (which resists aggregation) at high particle concentrations. Another variable that greatly affected the particle distribution on the probe surface was the suspension solution condition. From the data presented, NP assembly was more successful at 10 mM HEPES, 1 mM NaCl, pH 7.4 rather than the same buffer at pH 5.5. Looking at the zeta potentials measured at those two conditions, NPs were more densely attached when they were less charged (−40 ± 3 mV versus −52 ± 3 mV). This finding is consistent with other literature, which report that strongly charged particles exert more repellent forces on their neighbors, thus preventing tight packing of particles [37–39].
To demonstrate that NPs can be used as a drug delivery coating for neural devices, a mixture of NPs encapsulated with various molecules were attached to the probe surface and inspected both visually and in release experiments. Fluorescence microscopy showed dual fluorescence from 1% rhodamine + 1% fluorescein NPs on the same probe surface. Release from NPs attached to the surface was measured in PBS for two weeks and demonstrated the typical biphasic release pattern. The anti-inflammatory agent, DEX, was also released from the surface to prove that the system can deliver biomolecules. Other agents can also be encapsulated and released simultaneously, and there are no restrictions as to how many different types of NPs can be assembled on the same surface, although there is potentially a balance between the number of molecule types to deliver and the concentration of each type of NP present on the probe surface. Finally, the surface morphology of coated probes was examined before and after release studies. NPs were still attached to the surface after two weeks of exposure to buffer solution, and there was little change in their morphology. Since “bulk erosion” of PLGA polymer begins only after two to five months 2 of buffer incubation, the limited degradation of the NP coating during the 2-week period was not detectable under SEM. This observation is consistent with other literature references that cited work with pre-degraded PLGA particles [15, 16].
To revisit the idea that less charge leads to more NP attachment, NPs suspended in 10 mM HEPES, 100 mM NaCl, pH 7.4 were tested for assembly improvement. At near physiological buffer condition, NPs measured −15 ± 2 mV in zeta potential, a much lower magnitude compared to the −40 ± 3 mV from the previous 1 mM NaCl condition. Indeed, NPs were more densely assembled on the probe surface at lesser charge, as evidenced by the fluorescence and SEM images of the new condition. We believe that diminished mutual repulsion is at play here, due to both the diminished surface charge and the stronger charge screening via shorter Debye length. Image analysis counts of the particles tripled, as well as the percent of surface area coverage. Multilayering of PLL-NP was also tested because the single layering was already saturated; however, the result was not as ideal because the NPs aggregated with PLL on the second layering. It was also unclear what effects the long term exposure to buffer solutions would have on the sustained release from biodegradable NPs. Therefore, the optimal coating conditions for NP assembly on silicon probe was to suspend NPs in a 10 mM HEPES, 100 mM NaCl, pH 7.4 buffer at 0.5 mg/ml and to incubate them with the probe for 30 min to allow NP attachment. Under these conditions, the NP are sufficiently negatively charged to strongly attach to the surface, yet repel other particles only weakly, thus enabling a high loading.
The method presented in this study is an improvement in NP distribution compared to NPs embedded in hydrogel coatings. Quite frequently, NPs aggregate in the hydrogel solution before they are coated onto the neural device. This non-homogenous morphology leads to less reliable release profiles. Our new technique allows NPs to be assembled evenly on a device surface, with the additional option of adding a hydrogel coating to increase stability, if this is a concern.
To address the issue of whether or not the NP assembly detach from device surface in vivo, the transport properties of the particles were evaluated in order to predict their stability on surface in the presence of blood or other biological fluids. Given the typical shear flow diagram in Figure 10 and solving for laminar flow conditions, equation 1 describes the boundary layer between diffusive and convective regions of particle transport:
Where δ is the boundary layer between the two regions of transport, D is the diffusion coefficient (cm2/s), x is the horizontal distance of particle on surface (nm), and a is the shear rate (1/s−1). For any given particle with a vertical diameter of y > δ, the influence of convection dominates, so the particle will be caught in the convective flow and carried away from the device surface. For particle with a diameter of y < δ, the influence of diffusion dominates. The particle, then, will be in the diffusive region of transport and should remain close to the surface. Solving for δ with our NPs of 200 nm size, we get an approximate value of 2.5 µm, which is an order of magnitude greater than our particle diameter. Therefore, for our NP assembly system, most of the particles are subjected more to diffusion than convection. This calculation suggests the particles to remain well within the convective diffusive boundary layer; they are thus not expected to be significantly influenced by flow conditions. In short, NPs encapsulating bioactive agents can be assembled onto a neural device surface at high density and remain attached to the surface, despite fluid flow, to provide controlled delivery of treatment agents to neurons around the implant site.
This study demonstrated that PLGA NPs can be fabricated by using PEMA as the surfactant instead of the more commonly used PVA surfactant. The particles produced had normal spherical morphology and high encapsulation efficiency for small molecules. They also had a tunable range of zeta potential values, depending on the suspension buffer solution conditions. This property was important in NP assembly on silicon probe surface. NPs attached better at less negative zeta potentials (at near physiological condition), because the repulsion forces between the particles were minimized when zeta potentials are lower in magnitude. The NP coating was further proven capable of releasing multiple types of molecule simultaneously, and that the particles remained successfully attached to the surface even after two weeks of exposure to buffer incubation. Shear flow boundary layer calculations also supported this finding. The versatility of this coating system gives it a unique advantage over other current coating techniques. Multiple types of NPs (i.e. particles encapsulating different drugs, proteins or DNA) can be prepared independently in advance and easily mixed together in suspension prior to attachment. After they are assembled onto the same device surface, these NPs could then provide sustained release of several biomolecules from a single surface. This new coating technique, therefore, should be of interest to others working on neural device coatings.
This work was supported by grant #NS45236 from the National Institutes of Health and the NIH Neuroengineering Pre-Doctoral Training Grant (CTL). The authors thank Dr. Amarilys Sanchez-Santos for her technical advice on NP formulations, and Dr. Keith Neeves and Dr. Jian Tan for providing the silicon oxide probes.