Noninvasive Stimulation of Neurotypic Cells Using Persistent Photoconductivity of Gallium Nitride.

The persistent photoconductivity (PPC) of the n-type Ga-polar GaN was used to stimulate PC12 cells noninvasively. Analysis of the III-V semiconductor material by atomic force microscopy, Kelvin probe force microscopy, photoconductivity, and X-ray photoelectron spectroscopy quantified bulk and surface charge, as well as chemical composition before and after exposure to UV light and cell culture media. The semiconductor surface was made photoconductive by illumination with UV light and experienced PPC, which was utilized to stimulate PC12 cells in vitro. Stimulation was confirmed by measuring the changes in intracellular calcium concentration. Control experiments with gallium salt verified the stimulation of neurotypic cells. Inductively coupled plasma mass spectrometry data confirmed the lack of gallium leaching and toxic effects during the stimulation.

Introduction

Electrical stimulation of cells is of interest to understand the fundamental role of bioelectrical signals with respect to cell, tissue, and organ function.[1] In particular, mapping how electrical stimulation affects the neural cells has garnered additional interests because of the possibility to develop devices with clinical relevance.[2] As a result of all of this work, researchers have shown that electrical stimulation can enhance as well as guide the directionality of neurites. The published in vitro studies have utilized a number of materials’ approaches to facilitate electrical stimulation.[3] Both two-dimensional and three-dimensional scaffolds composed of both organic and inorganic materials have been studied.[4] A variety of fabrication strategies have been published including electrospinning, surface functionalization, and nanoparticle–polymeric composites.[5] Biomaterials’ interfaces in conjunction with electrical stimulation have induced the following cellular responses: morphology, migration, differentiation, proliferation, and adhesion. However, the application of electrical stimulation during in vitro studies varies widely, and many methodologies have been described as powerful but invasive, and with the disadvantage of being coupled with specialized equipment.

The need for a noninvasive stimulation has resulted in a number of seminal studies associated with interfaces for light-dependent cell stimulation. Pulsed IR light from a laser has been employed in vivo,[6] but is often times cited as potentially damaging to tissues due to high local thermal changes. The field of optogenetics,[7] where one uses light-sensitive proteins to trigger a response, relies on the light-emitting diodes and is still invasive due to the way light is delivered. The drive to reduce the invasive nature of the light stimulation has led to the development of photoactive and photoconductive surfaces. Many researchers have focused on adapting conductive polymers or quantum dots for this purpose.[8−10] These entities are attractive because one can fabricate thin films with them using flexible and low-cost fabrication strategies such as the layer-by-layer approach. Photoconductive silicon has been utilized in an inverted microscope setup where the invasive nature of the excitation is reduced but the need for laser light during excitation still remains.[11] Adapting materials with persistent photoconductivity (PPC) can eliminate the need for a light source during the stimulation.

In this article, we report on the ability of gallium nitride (GaN) to stimulate neurotypic PC12 cells due to its PPC properties, Scheme . This report builds on our recent efforts where we showed that PPC, variable topography, and surface bound functional groups can guide PC12 cells on specific surface locations.[12] We now describe the interfacial changes, along with surface and bulk charge properties that contribute to our ability to use the PPC of GaN to noninvasively stimulate the PC12 cells in vitro. We demonstrate that the stimulation of the PC12 cells is possible from a direct control of the surface charges. We report on the semiconductor characterization in the context of the in vitro studies using X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), Kelvin probe force microscopy (KPFM), and photovoltage measurements. We monitor the photoconductive stimulation via changes in intracellular calcium concentration that are governed by voltage-dependent channels.

Scheme 1

Representation of the Approach Used to Noninvasively Stimulate PC12 Cells in Vitro

Results and Discussion

Before the in vitro studies, two types of Ga-polar samples have been characterized. We evaluate both p- and n-type material to determine their suitability for stimulation experiments. The carrier concentrations are 4 × 1016 and 1017 cm–3 for the n- and p-type, respectively. We measure the photocurrent as a function of time before and after exposure to UV light. The relative intensity of photocurrent is significantly higher for the n-type sample compared to the dark photocurrent, Figure A. In addition to the magnitude of the photocurrent, we record different lifetime characteristics of the PPC for the n- and p-type samples. The time needed to drop to 10% of the dark photocurrent for the n-type sample is 100 s. The small amount of photocurrent generated by the p-type sample decays very slowly. The PPC for the p-type sample can be detected hours after illumination. Earlier work has extensively investigated the role of types and amounts of extrinsic dopants with respect to PPC in III–V semiconductor materials.[13] A combination of spectroscopy techniques has shown that extrinsic dopants are not responsible for the PPC characteristics of the materials. The PPC is thought to be determined by large lifetimes induced by point defects such as DX centers or the charge-separating polarization fields in the samples.[14] The data in Figure A determine the bulk conductivity of the material before and after UV light illumination. Literature reports have suggested that the nanostructured thin film materials have a significant number of charge carriers near the surface and are responsible for the specific PPC behavior.[15] Both N-polar and Ga-polar GaN thin films have a distinct nanostructured morphology recorded by a number of researchers in earlier work.[16] To understand the role of surface charge in the observed PPC characteristics, Kelvin probe force microscopy (KPFM) is performed on both types of materials.

Figure 1

(A) Photoresponse of the p-type and n-type Ga-polar GaN before (zone 1), during (zone 2), and after (zone 3), exposure to UV light. (B) Changes in surface potential collected by KPFM of p- and n-type Ga-polar GaN before and during exposure to UV light. (C) Changes in surface potential as a function of time, and before and after illumination with UV light. (D) Time dependence of the recorded photocurrent and surface potential voltage (SPV) for the p-type Ga-polar GaN. The change in the SPV for the p-type Ga-polar GaN is obvious, but it is not so for the n-type Ga-polar GaN due to the time scale.

Figure B summarizes the KPFM results for both p- and n-type Ga-polar GaN samples before and after UV illumination. When the KPFM is performed, one sets up the experiment to record any difference in the potential between the probe and the sample.[17,18] Any time one applies an alternating current voltage to the cantilever, it will cause it to experience electrostatic force and vibrate. For this reason, during the experiment, a direct current (DC) voltage is applied to eliminate tip vibrations.[19] The amount of applied DC voltage corresponds to the surface potential. In all of the cases when KPFM is done, a topography image is acquired simultaneously with the surface potential map. The two types of samples we studied have distinct differences in roughness. The rms values are 1.01 ± 1.01 and 13.7 ± 16.2 nm for the n- and p-type, respectively. Taken in sum, the KPFM data indicate that the magnitude of the surface potential is lower before exposure to light for both samples. The surface potential is positive for n-type and negative for p-type, as expected for the surface work function. The increase in the magnitude of the surface potential under illumination is a consequence of the induced surface photovoltage accompanied by an accumulation of minority carriers (i.e., increased positive charge and negative charge for n-type and p-type materials, respectively) at the surface, Figure C. The decay in the conductivity and associated surface potential indicates an associated gradual decrease in the near-surface accumulated charge, Figure D. Furthermore, the n-type sample demonstrates a larger change in conductivity and surface potential under illumination in comparison to the p-type GaN. Consequently, the n-type Ga-polar GaN with a large accumulation of charge at the surface is a more suitable material for cell stimulation studies.

Further spectroscopic characterization is performed on the n-type Ga-polar GaN samples to understand changes in the surface composition during solution exposure. More specifically, we carried out X-ray photoelectron spectroscopy (XPS) characterization with the intent of quantifying oxide formation[20] in air and in solution before and after UV exposure. The solution used in these studies is Dulbecco’s modified Eagle’s medium (DMEM), which is suitable for the growth of PC12 cells. All of the UV exposure is done outside of the solution for the same period of time. The surface compositions and all types samples’ conditions are summarized in the Supporting Information. In the absence of subsequent solution treatment, UV light does not result in a significant change in oxygen concentration on the surface. However, exposure to UV light in air leads to a lower amount of carbon present on the surface. We observed a statistically significant increase in the oxygen content on the surface after UV exposure and soaking in DMEM media. The amount of sodium detected under these conditions also increases proportionally, Figure A,B. These results support the notion that electrostatic interactions facilitate the adsorption of ions from the solution onto the surface. Our data are in agreement with a recent report that charge on the surface governs the assembly of charged diamond particles from solution onto both Ga-polar and N-polar GaN.[21] We quantified the change in surface roughness after incubation in DMEM solution before and after exposure to UV light. The amounts of adsorbates from the media present on the surface are substantial after soaking for 24 h leading to a statistically significant change in roughness. We observe a larger change when the samples are exposed to UV light before soaking in the DMEM solution. The observed change in protein adsorption from cell media upon exposure to UV light is consistent with earlier reports with photoresponsive GaN nanowires.[22] The roughness data are in agreement with the XPS observations and support the conclusion that an increase in oxides and oxyhydroxides on the surface can result in more adsorption onto the semiconductor interface.

Figure 2

(A) Changes in % O on the surface under different conditions extracted from the high-resolution O 1s XPS data. (B) Changes in % Na on the surface under different conditions extracted from the high-resolution Na 1s XPS data. (C) Changes in surface roughness of n-type Ga-polar GaN under different conditions. All of the rms values were extracted from the AFM images collected by AFM in air.

After the material characterization, we perform in vitro assays with the PC12 cells to test the hypothesis that n-type Ga-polar GaN can be used to stimulate neurotypic cells noninvasively. We validate the stimulation of the PC12 cells using an established assay to measure changes in the intracellular calcium concentration.[23] The exact mechanism and details of each step of stimulation of neuronal and neurotypic cells are still under investigation. However, an overwhelming amount of data in the literature have established that an increase in the intracellular calcium concentration is one of the steps of electrical stimulation of neurotypic cells.[24] In all of our experiments, the cells are never exposed to UV light. The n-type Ga-polar GaN is exposed to UV light before any introduction into the cell culture dishes. Based on the protocols developed by others, the PC12 cells are loaded with Fluo-4, a well-studied calcium indicator dye.[25] The data shown in Figure A indicate that the samples that experience PPC can indeed trigger an increase in intracellular calcium. The same amounts of cells and same types of semiconductor samples that are not exposed to UV light show an absence of elevated intracellular calcium concentration. The data in Figure A also include control experiments with no cells and no semiconductor samples that solidify the above conclusion.

Figure 3

(A) Calcium response of PC12 cells in the presence or absence of n-type Ga-polar GaN before and after its exposure to UV light, nd indicates the control condition which no Fluo-4 dye was added. (B) Calcium response of PC12 cells in the presence or absence of different concentrations of Ga(NO3)3 salt.

An increase in intracellular calcium can also be due to a cytotoxic response as has been established with the nanostructured materials such as silver.[26] In our earlier article, we report that UV light exposure of GaN samples causes only a small increase in the reactive oxygen species (ROS) production.[12] The ROS production can be triggered by the dissolved metal ions in solution. GaN is often quoted as an extremely stable material in harsh conditions[27] and a very promising interface for many biosensing applications.[28] Despite the fact that our samples are in solution for only a few minutes to perform the calcium assay, we verify the amount of gallium leached in a solution from the samples during the assay and after one day of incubation in the solution (see the Supporting Information for all of the data and statistical analysis). The data collected by inductively coupled plasma mass spectrometry (ICP-MS) indicate that at the time the calcium assay is performed in the cell culture wells with and without GaN samples, the amount of gallium is around 250 ppt, and there is no statistically significant difference in the results for all of the samples. After one day of incubation in a solution, the amount of gallium leached increases in the cell culture wells with GaN, but remains bellow 750 ppt. These results are in agreement with a number of ICP-MS analysis we have published on GaN with different polarities and compositions.[29−31] The ICP-MS data we collected for the present work confirms that leaching of gallium ions during the initial stimulation of the cells does not occur and will therefore not obscure the interpretation of results.

Additional experiments are performed to quantify the effect of gallium ions on the changes in intracellular calcium concentration, Figure B. A water-soluble Ga(NO3)3 is used for these experiments. Statistical comparisons among the data collected with three different gallium salt concentrations introduced in the PC12 cell culture show that an increase in gallium ion concentration leads to a decrease in the amount of intracellular calcium concentration. Control experiments are performed in the absence of gallium salt and the absence of PC12 cells. Earlier work with PC12 cells has concluded that a decrease in intracellular calcium concentration is a respiratory response and can be triggered by an excess of ions in the extracellular solution.[32] The ICP-MS and the additional assay data with the gallium salt support the conclusion that the calcium responses we observe with the semiconductor samples that experience PPC are due to stimulation of the PC12 rather than a response to the degradation of the semiconductor material.

Conclusions

In summary, we demonstrate that PPC of the n-type Ga-polar GaN can be used to noninvasively stimulate neurotypic PC12 cells. The type of semiconductor material chosen determines the amount of surface charge used for cell stimulation. The method to “charge” the semiconductor surface requires no special setup and only an affordable UV lamp. The surface characterization detailed the role of surface charge in the amount and persistency of the observed photoconductivity. The recorded change in intracellular calcium concentration in the presence of GaN with PPC demonstrates that one can noninvasively stimulate PC12 cells in vitro. The reported results are significant, as they open the possibility to utilize the electronic properties of III–V semiconductors to electrically stimulate cells without the use of electrodes or externally applied current. In future work, one can vary the electronic properties of the semiconductor substrate to a achieve dose-dependent stimulation.

Experimental Section

Semiconductor Growth

The n-type and p-type Ga-polar GaN were grown on sapphire wafers by metalorganic chemical vapor deposition using typical precursors: triethylgallium, ammonia, silane (n-type), and cyclopentadienyl magnesium (p-type). The film thickness was around 1 μm. Details on the growth procedures are described elsewhere.

Photocurrent Measurements

A Keithley 4200SCS characterization system was used to measure and record the photocurrent from n-type and p-type GaN semiconductor samples. The samples were left in the dark for 24 h before taking the measurement and UV illumination was done using a 4 W UVP, Longwave UV Lamp (Analytik Jena) with a 365 nm wavelength. UV illumination on the samples was ∼75 s.

AFM and KPFM

A minimum of three random 5 × 5 μm2 scans containing height, phase, and topography information was obtained using an Asylum Cypher S AFM. These scans were conducted on GaN wafers in various conditions using tapping mode with a scan frequency of 1 Hz and rectangular tipped cantilevers obtained from Asylum Research (f = 75 kHz, k = 3 N/m). We repeated the experiment three times for each condition. The the root mean square roughness values were calculated through Igor Pro (version 13.01.68), was measured in three random locations on each replicate.

Kelvin probe force microscopy was used to measure the topography, phase, amplitude, and surface potential of a minimum of three random 5 × 5 μm2 scans conducted on each of the condition’s three replicates. Conductive probes coated with Ti/Ir obtained from Asylum Research (f = 285 kHz, k = 42 N/m) were used for electrical measurements. The high conductivity of the samples caused significant (∼2–3 V) drift when left ungrounded. Thus, to prevent this drift, metallized contacts were deposited on the surface of the semiconductor samples, and a flexible wire was placed connecting the sample to a ground lead on the AFM. This sample grounding decreased the standard deviations from ∼2–3 V to ∼200 mV. The average surface potentials of each scan were statistically analyzed.

Surface charge was monitored with KPFM before, during, and after exposure to UV light. This was done by measuring the surface potential of a 5 × 5 nm2 area with a scan time of 1 min 5 s. UV illumination was conducted by placing the UV lamp in the AFM before scanning, and turning it on (UV+) and off (UV−) to minimize surface potential drift within each scan. The average surface potential of each area was plotted against time, and the standard deviation of each area was used as error bars.

XPS

A Kratos Axis Ultra DLD X-ray photoelectron spectrometer was used to analyze the surface chemistry of GaN surfaces under a variety of conditions. The experimental design was a fully factorial design measuring the change in surface chemistry from the effect of soaking in Dulbecco’s modified Eagle medium (ThermoFisher) for 24 h and UV illumination for 1 h, before the DMEM soaking. Three areas of each replicate surface were individually scanned, and the results were analyzed using a one-way analysis of variance (ANOVA). In addition to a survey scan (pass energy = 160 eV), high-resolution regional scans (pass energy = 20 eV) were also performed for Ga 2p, Ga 3d, N 1s, C 1s, O 1s, and Na 1s. Atomic concentrations were calculated from the survey scans using Casa XPS (version 2.3.12.8), and the averages from each replicate condition were analyzed after conducting a Shirley background subtraction and C 1s calibration to 284.8 eV.

GaN Wafers—Calcium Assay

Fourth passage PC12 cells (ATCC) were grown to 80% confluency in collagen-coated petri dishes. A 0.5 mL of Trypsin-EDTA was added to the cultures for cellular detachment, and the cultures were then incubated at 37° at a 5% CO2 atmosphere. Sterilized PC12 cell media consisting of DMEM containing 12.5% horse serum, 2.5% bovine serum, and 1% penicillin/streptomycin was prepared and used throughout all of the cell growth and studies. The cell suspensions were then centrifuged at 200 g for 8 min, and the medium was aspirated and replaced with 12 mL of the Fluo-4 Direct calcium assay reagent-calcium assay buffer (F10472, Invitrogen). The included probenecid was not added to decrease the potential cytotoxic effects. The cell suspension was then vortexed to break up the cell pellet. Meanwhile, a collagen-coated 24-well plate was prepared with the semiconductor samples. Two test conditions and three control conditions were used: (1) semiconductor samples were exposed to UV light for 1 h before cell seeding; (2) semiconductors were left unexposed with UV light; (3) no semiconductor was placed in the well; (4) neither samples nor cells were placed in the wells; and (5) semiconductors that were not exposed to UV were placed in the well with no dye present. Each condition was replicated in three wells.

For each replicate, 0.5 mL of the the Fluo-4 Direct calcium assay reagent/Ca assay buffer suspension was deposited in each well with a cellular concentration of ∼210 000 cells/well. Warmed cell medium (0.5 mL) was also deposited into each well for a 1.0 mL total volume that contains a 50/50 mix of a cell suspension of dye (Fluo-4 Direct calcium assay buffer and reagent) and cell media. The well plate was then incubated at 37° in 5% CO2 for 1 h. After incubation, the well plate was inserted in a Tecan Microplate reader with excitation at 494 nm and emission at 516 nm. The fluorescence signal for each well was measured three times, sequentially. The data were analyzed in terms of relative fluorescence units (RFUs), whereby the , where all of the conditions were compared against the control condition where no semiconductor was present with the cells. Soon after the fluorescence measurements were conducted, a 200 μL aliquot was obtained from each well (replicate). The well plate was incubated for 24 h, and an additional 200 μL aliquot was taken from each well. These aliquots were stored at ∼2 °C until they were analyzed by ICP-MS.

Ga Salt—Calcium Assay

Ga(III) nitrate hydrate (Ga(NO3)3·xH2O) obtained from Sigma-Aldrich was diluted into sterilized phosphate buffer solution (PBS) at concentrations of 4, 2, and 1 μg/L. In addition to the salt conditions, three controls were prepared: 0 μg/L salt with PBS, 0 μg/L salt in PBS with no cells present, and 0 μg/L salt in prepared cell media. In total, 0.5 mL of each of the salt/PBS or salt/media conditions was deposited into a prepared collagen-coated 24-well plate. Each condition was replicated in three different wells. Meanwhile, 7th passage PC12 cells were harvested by introducing 0.25% Trypsin-2.21 mM EDTA into cell culture dishes. After centrifugation and media aspiration, the cells were resuspended in a volume containing Fluo-4 Direct calcium assay reagent combined with 100 mL of the Fluo-4 Direct calcium assay buffer. Dye-cell suspension (0.5 mL) was then passaged into wells of the 24-well plate at a cell density of 230 000 cells/well and then incubated at 37° in 5% CO2 in the dark for 1 h. Thus, the salt concentrations were diluted by 2× to the final concentrations of 2, 1, and 0.5 μg/L. After incubation, the well plate was transferred to a microplate reader, tested using the Fluo-4 Direct calcium assay methods previously described.

ICP-MS

All data was collected with a Thermo Element XR instrument. 200 μL aliquots were taken from each replicate at day 0 and at day 1 from the same wells that the Ca-assay on GaN wafers was conducted. These aliquots were prepared for inductively coupled plasma mass spectrometry (ICP-MS) by filtering and acid digestion in water. The Ga concentrations were measured for each replicate and analyzed using statistical analyses.

Statistical analysis

All of the statistical analyses were conducted using OriginPro 2016 (b9.3.1.273). One-way and two-way ANOVAs were used to find differences between the experimental conditions with a significance level of 0.05.