Research Progress

Activities

1. Develop an acoustic wave cell focuser.

We fabricated and tested a bulk acoustic wave cell focuser to verify the cell manipulation ability of acoustic wave device. A fluidic channel with a height of 200µm was fabricated by binding glass slides together. Polystyrene microspheres with diameters of 10±0.2µm (72986 Fluka, Sigma-Aldrich) were loaded and tested under the effect of bulk acoustic wave while flowing through the channel. The acoustic wave was generated by a piezoelectric sheet (piezoceramic sheet PSI-5H4E, Piezo System Inc.) with a sinusoidal voltage generated by a function generator. The trajectory of the beads were captured by a digital camera (digital camera QIC-F-M-12-C, QIMAGING).

2. Develop a cell electrical stimulator.

For this goal, we have developed a device that can quickly apply versatile electrical stimulation (ES) signals to cells suspended in microfluidic channels and measure extracellular field potential simultaneously. Cardiomyocytes together with primary rat cortex neurons (rCNs) were tested with the device.

2.1 Design concept and sensing principle

The design of the microfluidic device for single cell electrical stimulation is shown in Figure 1. The device has 3 main fluidic channels; each has dimensions of 500 µm in width and 50 µm in height. A constriction microchannel was fabricated between the main channels in order to trap and locate the single cell on top of a measurement electrode. The dimensions of the constriction channel were 10 µm in height, 8 µm in width and 15 µm in length. A pair of 200 µm wide stimulation electrodes were fabricated to apply ES signals. Another pair of electrodes with a 50 µm width were located on both sides of the constriction channel and worked as measurement electrodes to record the EFP. 3 reservoirs with 1.5 mm diameter access holes were fabricated along with the channels: an inlet reservoir (Inlet) for loading the floating cells, an outlet reservoir (Outlet2) to collect the cells after ES, and another outlet reservoir (Outlet1) for collecting the fluidic medium.

Cells suspended in medium were introduced into the device via the inlet reservoir at a concentration of 20-50 cells per µL. We used a flow controller (Flow-EZ, Fluigent, France) to apply a constant pressure of 3kPa through the inlet reservoir to control the cell flow. The single cell was trapped and located on top of the left measurement electrode by the constriction microchannel. Single cardiomyocytes and neurons with diameters from 10 µm to 50 µm should be captured on the electrode surface. This was con-firmed by experimental observations as shown in Fig. 1(e). The constriction dimension can be modified to position single cells with diameters out of this size range.

The flow was approximately stopped once a single cell was trapped by the constriction channel and blocked the constriction channel. This can be confirmed by observing a flow rate reduction by the flow controller, and also by optical observation from the microscope, the applied pressure was removed by setting the pressure to zero. With the low cell concentration (about 20 cells/µL), only one cell could be trapped and positioned on the surface of the measurement electrodes. Then we applied the ES signals on cells through a pair of wide electrodes (200µm wide), and measured the extracellular field potential signals from the trapped cell through a pair of measurement electrodes (50µm wide). The extracellular field potentials (EFP) reflect electrical activity of cardiac and nervous cells or tissues based on the transmembrane currents in the extracellular medium. When a cell is electrically stimulated and invokes changes of transmembrane ion pumps and voltage-gated channels, a transient imbalance in ion concentration forms across the cell membrane. The caused transient alteration of the transmembrane currents will induce response EFP signals and could be quantitatively recorded by this device to reflect the cell activity. Note that all cells loaded into the channel traveled through the stimulation region (channel region between the pair of stimulation electrodes) and were electrical stimulated. Because of the use of the current pulses as ES signals, all cells presented in the stimulation region experienced the same electrical stimulations.

Figure 1. Schematic of the microfluidic device for single cell ES and FP recording. (a) Illustration of device design for single cell electrical stimulation and FP measurement. (b) Picture of the device. (c) Microscopic photo of the measurement region, including the measurement electrodes and the constriction micro-channel. (d) Microscopic images of the cardiomyocyte trapped by the constriction channel (Measurement electrodes were intentionally re-moved to take clear images of cell trapping).

The device was designed to detect the EFP signals of the single cell trapped by the construction channel and located on top of a measurement electrode. The FP magnitude is inversely proportional to distances between each cell and measurement electrodes. Hence the FP magnitude from the cells that were away from the measurement electrode was too weak to detect. By controlling the low concentration of the cell (about 20 cells/µL), the chance that multiple cells situating on or very close to measurement electrode is small. From the microscope observations during the tests, we typically saw only one cell was trapped and situated on the measurement electrode. This can be confirmed by a microscope image, shown in Figure 1(d). As a result, only the FP signals from the cell tapped on top the electrode were detected.

While a cell may have experienced some stress and deformation due to trapping, the trapping was unlikely to cause the cell to generate detectable electrical activities. This was confirmed by the following test: we trapped one cell outside the constriction channel and let it situate on the measurement electrode for 1 minute without apply ES. And then we measured the electric signals from the cells. We did not observe any detectable FP signals from the trapped cell without electrical stimulations. Subsequently, we applied the ES stimulation (0.4mA, 1Hz, 0.5 ms pulse width) to the same cell for 1 minute. With the ES, we measured clear FP signal from the cell. 10 different single cells were trapped and measured separately (with and without applying ES). No FP signal was detected when the cells were not electrically stimulated, while we did detect FP signals when they were stimulated. This test proved that the cell trapping did not cause a noticeable change in the cells’ electrical activity.

The entire system was settled in a Faraday cage to reduce the electromagnetic interferences/noises. During the test, we used an upright microscope (PSM 1000, Motic) equipped with a video camera (QICAM 12 bit, QIMAGING) to observe the cell positioning/trapping. To apply the ES signals applied on cells, we used a waveform generator (33600A, KEYSIGHT) and an input circuits to generate the desired signals and exert the signals via the stimulation electrodes. We then recorded and digitized the EFP signals from cells through a data acquisition board (PCI-6133, National Instrument, USA). A custom LabVIEW program was set for the data acquisition at a sampling rate of 1.5 MHz. Here we used current-controlled signals with a magnitude of 0.1 to 0.8mA, which resulted a 0.3 to 15V/cm electrical field. The frequency, pulse width and duration were from 0.5 to 2 Hz, 0.5ms, and 1 min respectively. Both hardware and software method were used to remove noises and stimulus artifacts resulted from the input ES signals from the detected signals. The stimulation signals also induced artifact signals on the measurement electrodes. Because the induced artifacts signals had much larger peak magnitudes than the FP signal from a single cell, and occurred regularly with the same frequency as the stimulation signals, the stimulus artifacts were filtered out from FP signals using filter circuits and MATLAB algorithm.

2.2 Design fabrication

The microfluidic device was fabricated with the standard soft lithography process. A two-layer master mold with the structure of microchannels and reservoirs made of SU8 (5 & 2025, MicroChem) was created using a two-step photolithography method. The fabrication process is as follows: First, a 10 µm thick SU8-5 layer was spin coated on a silicon wafer. A photolithograph step was then utilized to define the pattern of the constriction channel. Next a second layer of SU8-2025 (40 µm thick) was spin coated on top of the first layer and underwent a photolithograph step to obtain the wide channel and reservoir structures. Subsequently polydimethylsiloxane (PDMS) was used to form the microchannel with the mater mold. The microelectrodes were fabricated with the 100 nm /10 nm gold / titanium coating on a glass slide (1"x3", EMF Corporation). Photolithography and wet etching with KI / I2 gold etchant (Sigma-Aldrich) was used to fabricate electrodes. Finally, an oxygen plasma treatment (200 mTorr, 50 W, 50 s) was used to activate the surfaces of PDMS and the glass. After the treatment, the PDMS structure was bonded to the glass slide. A surface profilometer (Dektak 150, Veeco Instrument) was used to measure the dimensions of the constriction microchannel. The dimensions were 14.67 µm ± 1.34 (length), 9.23 ± 0.95 µm (depth), and 6.87 ± 0.74 µm (width).

The electrodes were platinized to reduce their impedance before bonding the glass slide with the PDMS microchannels. Platinization is the process of depositing a layer of platinum black on the gold electrodes. The electrodeposition of platinum black usually results in a rough surface and largely increases the effective area of electrodes. Before the platinization the electrode surface was rinsed by methanol. Electroplating of platinum black on the gold electrodes was conducted in platinizing solution (YSI 3140 Platinizing Solution, water solution of 5% chloroplatinic acid and 0.1% lead acetate). A 5V DC voltage was applied across the two measurement electrodes for 10s. Then a negative DC voltage (-5V) was applied on the electrodes for another 10s. After the Platinization, we scanned the impedance spectrum of the measurement electrodes using a Gamry Reference 600 potentiostat (Gamry Instruments, Warminster, PA, USA). With the myocyte growth medium, we measured the impedance of the electrodes as a function of AC frequency. Results showed that the electrode impedance was largely reduced at lower frequencies (<1kHz) in which the dominant frequencies of the FP signals fall. The electrodeposition was conducted before we bonded the glass slide with the PDMS microchannels. Because of the small thickness of the gold and black Pt (~0.1 µm), the glass slide and the PDMS channel were bonded well; the microchannel remained sealed. No leakage was found in our tests.

2.3 Cell culture

Human cardiac myocytes (hCMs), myocyte growth medium kit, cryoSFM, and detach kit were obtained from PromoCell GmbH (Heidelberg, Germany). Primary rat cortex neurons (rCNs), B-27TM plus neuronal culture system, LIVE/DEAD viabil-ity/cytotoxicity kit, and AmplexTM red glutamic acid/glutamate oxidase assay kit was purchased from ThermoFisher Scientific (Walkersville, MD, USA). Trypsin/EDTA and antibiotic−antimycotic solution were obtained from Gibco (Carlsbad, CA, USA). Penicillin-streptomycin was obtained from Sigma Aldrich (St. Louis, MI). Poly-D-lysine coated 12-well plate was purchased from Greiner Bio-One (Monroe, NC, USA). Cell scraper and cell culture grade 1X phosphate buffered saline solution was obtained from Corning (Manassas, VA, USA). All the materials were used as received from the manufacturers.

hCMs were cultured following the manufacturer’s instruction using myocyte growth medium kit. rCNs were cultured using the B-27TM plus neuronal culture system containing 1% penicillin-streptomycin solution. Both cells were maintained in cell culture incubator at the temperature of 37 °C with 5% CO2. Medium change was performed every other day for the hCMs and every day for the rCNs. The neurons (rCNs) were cultured in vitro for 1 days prior to the testing. Because of the short culture time, we did not observe any formation of cell-cell connection and neurites in the cultured neurons. hCMs at passages 6~9 and rCNs at the pas-sage 1 were used for all the testing.

1.4 Cell viability test

Tests of cell viabilities of primary rCNs before and after electrical stimulation were performed with LIVE / DEAD viability / cytotoxicity kit (Thermo Fisher Scientific, Waltham, MA, USA) using a microplate protocol according to the manufacturer’s recommendation. First we seeded rCNs on poly-D-lysine coated 12-well plate. The cells were incubated in the humidified cell culture incubator (37 °C, 5% CO2) for 48 h. After the incubation, we stained the adherent neurons with calcein AM (1μM) for live cells and ethidium homodimer (4μM) for dead cells followed by harvesting the cells for experiments. The cells were divided into two experimental groups before loaded into the device. We applied ES to one group, while no ES signals were applied to the other group. The ES used for viability test were 2Hz pulses with 0.4mA in magnitude, 0.5ms in width and 60s in duration. Cells before being introduced into the device were tested and used as control groups for each experiment. The fluorescence intensity was measured at excitation/emission of 495nm/645 nm and 495nm/530nm using a Synergy H1 hybrid microplate reader (Bio-Tek Instruments, Winooski, VT, USA). All the cells loaded into the device were collected and tested for cell viability.

2.5 Glutamic acid measurement

The glutamic acid concentration of rCN before and after the electrical stimulation was measured by AmplexTM red glutamic ac-id / glutamate oxidase assay kit (Molecular Probes, Eugene, OR, USA). Briefly, the rCNs were detached from the culture plate, passed through the microfluidic device, and were applied the ES (0.1-0.4 mA; 0.5-2 Hz; 0.5ms pulse width; 60 s period). For the no ES control group, the cells passed through the microchannel and were collected without applying the ES. For both experiment groups, the collected cells were then centrifuged at 100×g for 3 min at 4°C to collect the supernatant for measurement. After staining the samples with AmplexTM red glutamic acid / glutamate oxidase assay kit, the fluorescence intensity was measured at excitation/emission of 545nm/590 nm using a Synergy H1 hybrid microplate reader (Bio-Tek Instruments, Winooski, VT, USA).

The measurement of the glutamic acid concentration change was to prove the electrical stimulation did trigger electrical activities of excitable neuron cells indicated by the secretion of glutamic acid; thus the measured FP signals were generated by the stimulated cells rather than artifacts or crosstalk signals.

Findings

1. Develop an acoustic wave cell focuser.

We tested a bulk acoustic wave device to focus the microparticles (diameter: ~10µm) to the middle of the channel, shown in Fig 7. The particles were focused at the centerline while flowing through the channel upon the effects of acoustic wave. The acoustic wave was generated by a 20MHz, 10V sinusoidal voltage applied on a piezoelectric sheet. This indicate that microparticles/ cells can be manipulated within a continuous flow by acoustic wave device, such that cell manipulation in an electrical stimulator can be achieved and single cells can be measured/stimulated in situ.

Figure 2. Particles focused at the centerline in an acoustic wave device

2. Develop a cell electrical stimulator.

For this goal, we have developed an electrical stimulation and detection setup to apply controlled current signals on target cells and measure the responses of excitable cells.

2.1 The effects of ES conditions on single hCM responses

Human cardiac myocytes (hCMs) were introduced into the microfluidic device and positioned onto the measurement electrodes one by one. Electrical stimulations with different conditions were applied on single hCMs. Extracellular field potential of the cell was recorded at the same time (Figure 3).

Figure 3. Detected field potential signals from human cardiac myocytes electrically stimulated in the device (a) hCM responses un-der different ES amplitudes (0.2/0.4/0.6 mA, 0.5 Hz pulses with a 0.5 ms pulse width (b) Microscopy phase contrast images of single hCMs in suspension (Scale: 50 µm), taken before cells were loaded into the device (c) Magnitude and occurrence of hCM responses under different ES frequencies.

Different amplitudes of ES were tested from 0.2mA to 0.6mA, shown in Figure 3(a). Without stimulation or with stimulation below 0.4 mA, the range of the detected signal was found to be between -5µV to 5µV, which was within the noise band from the measurement circuit. When the ES amplitude was above 0.4 mA, the field potential responses were detected by the device. When a single cell was evoked, the detected extracellular potential underwent a rapid drop forming a sharp negative pulse. The potential then raised above 0V resulting a positive pulse. Finally the potential returned to the baseline of 0V. The FP signals were clearly identified with waveform patterns. The peak value of the pulses ranging from –60 to 50µV. The magnitude and occurrence of field potentials generated by hCMs in responses to ES with different magnitudes were recorded and compared in Figure 3(c). The magnitude of the hCM’s FP signals was measured as peak to peak value of each FP signal. Average magnitude was calculated from all FP pulses during a 20s ES period. The occurrence of the cell responses was defined to represent the average amount of identified FP pulse signals per second in 20 s. The results showed that the hCMs were electrically activated when the ES was stronger than 0.4 mA. Continuing to increase the ES amplitude to 0.6 mA, the FP responses measured from electrodes did not change significantly. The results demonstrate that a magnitude threshold of ES is needed to sufficiently trigger the cell electrical activity. When the excitable cells are triggered for electrical activity, the voltage-gated ion channels embedded in a cell’s plasma membrane will open and let the ion exchange happen, causing the detected FP signals. While applying higher ES magnitude (0.4-0.6mA), the FP signals did not alter much, which was similar to the observation in our previous study. The possible cellular mechanism we suspect is that the voltage-gated ion channels do not have the function to further alter its ability for ion exchange after they are triggered by the electric current above the threshold.

Figure 4. Effect of stimulation frequencies on responses of single hCMs (a) hCM responses under different ES frequencies (0.4 mA, 0.5/1/2 Hz pulses with a 0.5 ms pulse width). (b) Magnitude and occurrence of hCM cluster responses under different ES fre-quencies.20 (c) Magnitude and occurrence of single hCM responses under different ES frequencies.

To confirm the FP responses of hCMs and study the relation between various stimulation signals and single cell electrical activity, single hCMs introduced into the device were stimulated by ES signals (0.4mA, 0.5ms wide pulses with a duration of 60 s) at multiple frequencies (0.5Hz, 1Hz and 2Hz). The FP signals’ magnitude and occurrence were measured and compared in Figure 2. Results demonstrated that changing input ES signals’ frequency from 0.5Hz to 2Hz, the cell FP signals’ occurrence significantly in-creased from 0.27 to 1.06 pulse/s on average. These results were consistent with the results from our previous study of ES frequency effect on hCM clusters, as shown in Figure 4(b), which means the effects of different ES frequencies on cell clusters also apply to single cells. The input ES signals’ frequency played an important role in controlling hCMs’ FP signals. Higher frequency ES seemed to have the ability to further increase the cell electrical activities of single hCMs. The cellular mechanisms behind this phenomenon have not been fully elucidated. One possible mechanism we suspect is that higher frequency of ES increased the occurrence of cell membrane alterations (e.g. open frequency of the voltage-gated ion channels), which may further increase the cell electrical activities. The results showed that the frequency of the ES might have the ability to regulate the level of triggered cells’ electrical activities. This could be used to optimize the future bioreactor design and apply controlled ES on cardiac tissue construct in order for the desired electrical property.

2.2 The effects of ES frequency on single neuron responses

Neuron cell ES and FP measurement is a widely used tool to study the function and connectivity of neuronal circuits. To validate the ability to electrically stimulate neurons and detect FP signals of our device, single primary rat cortex neurons (rCNs) were introduced into the microfluidic device and electrically stimulated by input signals (0.4mA, 0.5ms wide pulses with a duration of 60s). FP signals generated from the stimulated single neurons under ES of different frequencies (0.5Hz, 1Hz and 2Hz) were recorded (Figure 5). Results also showed a trend of improving cell electrical responses under increasing frequencies of single neurons, similar to the hCMs. The occurrence of the rCN FP signals significantly increased from 0.3 to 1.2 pulse/s on average, when the frequency was increased from 0.5Hz to 2Hz. The average magnitude did not change significantly under different frequencies. These results indicate that the frequency of ES is a critical factor in triggering rCNs’ FP signals. Higher frequency ES had the ability to further increase the cell electrical activities of single rCNs, which is consistent with the results from hCMs. These two types of cells showed similar trends of field potential signal change when ES conditions were varied.

Figure 5. Detected field potential signals from single primary rat cortex electrically stimulated in the device (a) rCN responses under different ES frequencies (0.4 mA, 0.5/1/2 Hz pulses with a 0.5 ms pulse width). (b) Microscopy images of single rCNs (The Scale bar is 50 µm). (c) Magnitude and occurrence of rCN responses under different ES frequencies. * indicates the statistically significant difference between the experiment groups with p-value less than 0.05.

2.3 Cell viability and secretion assay

We conducted cell viability tests to investigate if the applied ES conditions could cause a significant amount of cells to die during experiments. Results shown in Figure 6 demonstrated that optimized ES conditions would not cause a significant number of cells to die after the tests. Percentages of viable cells before and after ES were 84.66±2.37% and 82.16±4.70%, respectively. No statistically significant difference was found between the tested groups in student’s t-test analysis. The average time consumed from harvesting the cells to completing the entire ES tests was ~ 2h, showing that this device has the feasibility to apply ES up to 2h without compromising the cell viability. Thus this device has the ability to electrically stimulate single cells while not affecting the viability of the cells.

Figure 6. Cell viability and secretion of glutamic acid by primary rat cortex neurons before and after the electrical stimulation (ES). (a, b) Fluorescence images of rCN stained with calcein AM (green) and ethidium homodimer (red). The Scale bar is 50 µm. (c) Percentage of viable cells before and after the ES. (d) Secretion of glutamic acid concentration by primary rat cortex neurons with and without the ES. (0.4 mA, 0.5 Hz pulses with a 0.5 ms pulse width) (e) Secretion of glutamic acid concentration with different ES magnitude. (0.5Hz) (f) Secretion of glutamic acid concentration with different ES frequency. (0.4mA) * indicates the statistically significant difference between the tested experiment groups with p-value less than 0.05.

Analysis of the secretion of glutamic acid from single rCNs under different ES conditions was investigated to verify the ability to trigger electrical activities of excitable cells with ES within the device, and also to validate the potential biological applications of this device. Glutamate is the most abundant neurotransmitter in brain and central nervous system, which is involved in all major excitatory brain function. Many studies have used glutamate as a highly sensitive functional marker for excited neurons. After passing through the microfluidic device with different ES conditions, the cells were collected, and centrifuged at 100 × g for 3 min at 4°C. Then the samples were stained by red glutamic acid/glutamate oxidase assay kit, and the fluorescence intensity was measured (See Section “Glutamic Acid Measurement”). The glutamic acid concentration in the supernatant collected from each experiment group with different ES conditions Figure 6(d)-(f). The results showed that the electrical stimulation significantly increased the glutamic acid concentration in electrically stimulated cells compared to control group (Figure 6(d)). The concentration of glutamic acid with and without ES were 2.21±0.17µM and 0.56±0.08µM, respectively. Results showed that applying different ES magnitude did affect the glutamic acid concentration, and high enough magnitude is necessary to achieve sufficient cell activity. It was also shown that the glutamic acid concentration of cells stimulated by different ES frequency did not have a significant difference, which might be due to the sufficient glutamic acid release from the cells under the ES. The neurons released majority of the glutamic acid after the ES in the device even under lower frequency, and increasing the frequency would not further increase the glutamic acid concentration in the supernatant. It is widely understood and proved by many studies that changes in ES magnitude would alter the number of activated neurons in the stimulated tissue28. However, the extent to which different stimulation frequencies and pulse durations affect local neuronal responses remains less explored. Our results indicated the dominant role the pulse magnitude plays as the ES parameter for neuron activation.