Article

Surface Acoustic Waves (SAW)-Based Biosensing for Quantification of Cell Growth in 2D and 3D Cultures Tao Wang 1,2,† , Ryan Green 1,3,† , Rajesh Ramakrishnan Nair 1,4,5,† , Mark Howell 1,3 , Subhra Mohapatra 1,3, *, Rasim Guldiken 1,2, * and Shyam Sundar Mohapatra 1,4, * Received: 11 September 2015; Accepted: 11 December 2015; Published: 19 December 2015 Academic Editors: Zhiping Wang and Charles Chun Yang 1

2 3 4 5

* †

Center for Research and Education in Nanobioengineering, University of South Florida, Tampa, FL 33612, USA; [email protected] (T.W.); [email protected] (R.G.); [email protected] (R.R.N.); [email protected] (M.H.) Microfluidics and Acoustics Laboratory, Department of Mechanical Engineering, College of Engineering, University of South Florida, Tampa, FL 33612, USA Department of Molecular Medicine, University of South Florida, Tampa, FL 33612, USA Departments of Internal Medicine, University of South Florida, Tampa, FL 33612, USA Transgenex Nanobiotech Inc, Tampa, FL 33612, USA Correspondence: [email protected] (S.M.); [email protected] (R.G.); [email protected] (S.S.M.); Tel.: +1-813-974-8568 (S.S.M.); Fax: +1-813-974-8575 (S.S.M.) These authors contributed equally to this work.

Abstract: Detection and quantification of cell viability and growth in two-dimensional (2D) and three-dimensional (3D) cell cultures commonly involve harvesting of cells and therefore requires a parallel set-up of several replicates for time-lapse or dose–response studies. Thus, developing a non-invasive and touch-free detection of cell growth in longitudinal studies of 3D tumor spheroid cultures or of stem cell regeneration remains a major unmet need. Since surface acoustic waves (SAWs) permit mass loading-based biosensing and have been touted due to their many advantages including low cost, small size and ease of assembly, we examined the potential of SAW-biosensing to detect and quantify cell growth. Herein, we demonstrate that a shear horizontal-surface acoustic waves (SH-SAW) device comprising two pairs of resonators consisting of interdigital transducers and reflecting fingers can be used to quantify mass loading by the cells in suspension as well as within a 3D cell culture platform. A 3D COMSOL model was built to simulate the mass loading response of increasing concentrations of cells in suspension in the polydimethylsiloxane (PDMS) well in order to predict the characteristics and optimize the design of the SH-SAW biosensor. The simulated relative frequency shift from the two oscillatory circuit systems (one of which functions as control) were found to be concordant to experimental data generated with RAW264.7 macrophage and A549 cancer cells. In addition, results showed that SAW measurements per se did not affect viability of cells. Further, SH-SAW biosensing was applied to A549 cells cultured on a 3D electrospun nanofiber scaffold that generate tumor spheroids (tumoroids) and the results showed the device's ability to detect changes in tumor spheroid growth over the course of eight days. Taken together, these results demonstrate the use of SH-SAW device for detection and quantification of cell growth changes over time in 2D suspension cultures and in 3D cell culture models, which may have potential applications in both longitudinal 3D cell cultures in cancer biology and in regenerative medicine. Keywords: surface acoustic waves (SAW); shear horizontal—surface acoustic waves (SH-SAW); Biosensor; zinc oxide (ZnO); cancer; 3D cell culture

Sensors 2015, 15, 32045–32055; doi:10.3390/s151229909

www.mdpi.com/journal/sensors

Sensors 2015, 15, 32045–32055

1. Introduction Detection and quantification of cell viability and growth in two-dimensional (2D) and three-dimensional (3D) cell cultures commonly involve harvesting of cells and therefore requires a parallel set-up of several replicates for time-lapse or dose–response studies. Currently, cell growth or proliferation of flat 2D cultures utilize MTT assay, flow cytometry and Ki67 staining. Similarly, measuring cell growth and proliferation in 3D cultures consist of terminal studies that may include trypsinization and staining with trypan blue and quantification. Thus, a non-invasive and touch-free detection of cell growth or proliferation in longitudinal studies, especially for 3D tumor spheroid cultures and stem cell regeneration remains a major unmet research need. Surface acoustic waves (SAWs) have been broadly applied in many areas of micro-sensor technology. Gas sensors [1], biosensors [2,3] and chemical sensors [4] are a few of the leading applications for SAW sensors. Generally, biosensors are widely used in cancer biomarker detection and bio-agent detection. Due to SAWs’ advantages of low cost, small size and ease of assembly, SAW-based biosensor technologies have the potential to transform the cancer and bio-agent detection fields [3,5]. However, the potential application of SAWs for the detection of cell growth has not been reported and remains to be elucidated. SAWs consist of two particle displacement components. One is along the direction of wave propagation and the second one is normal to the surface, such as Rayleigh waves. Rayleigh waves, which generate compressional waves, are affected and damped by the liquid loading and dissipate the wave energy into the liquid. Therefore Rayleigh surface acoustic waves are less sensitive to mass loading changes [6]. Shear horizontal-surface acoustic waves (SH-SAW) with the substrate polarized normal to wave propagation are most commonly used in sensor applications that involve fluidics. Many different wafer types with special cuts are used for shear horizontal wave excitation, such as ST-cut Quartz [2,7,8] and 36˝ Y-cut LiTaO3 [6,9,10]. ST-cut Quartz and 36˝ Y-cut LiTaO3 are very stable substrates for sensor applications. However, the electroacoustic coupling coefficient (K2 ) of ST-cut Quartz is much smaller than that of 36˝ Y-cut LiTaO3 (36˝ Y-cut LiTaO3 is 4.7 [11] and ST-cut Quartz is 0.0016 [12]). Because of its high electroacoustic coupling coefficient, the 36˝ Y-cut LiTaO3 generates more stable signals when the SH-SAWs travel through polydimethylsiloxane (PDMS), which absorbs the majority of the energy generated by the interdigital transducers [13,14]. PDMS has been widely used in biomedical devices due to its biocompatibility and ease of manufacture into fluidic channels. An optimization of the PDMS channel sidewall thickness was demonstrated to reduce the damping effect of the PDMS on the wave propagation [15], thereby increasing the sensitivity of the sensor. Even though 36˝ Y-cut LiTaO3 has a higher electroacoustic coupling coefficient, it also has a higher temperature coefficient compared to the ST-Quartz. Various guide layers can be deposited on the LiTaO3 to change the phase velocity and temperature coefficient of the system. Zinc Oxide (ZnO) is a relatively common material in sensor and SAW fields. The majority of SAW devices coated with ZnO are used as Ph [16] or UV sensors [17]. Coating a ZnO layer on a LiTaO3 substrate reduces the temperature coefficient and increases the mass sensitivity [18–21], hence addressing the shortcoming of the LiTaO3 substrate as opposed to its alternatives. We have been investigating the possibility of non-invasive touch-free monitoring of cell proliferation/growth in long-term 2D and 3D cell cultures. Serendipitously, we discovered that SAW-based biosensing produced a different frequency shift. We reasoned that SH-SAW using 36˝ Y-cut LiTaO3 wafers coated with ZnO might have the potential to measure and quantify cellular mass changes. To test this idea, we utilized a PDMS channel/well and surface acoustic wave transducers coated with a ZnO layer to measure mass changes due to increasing cell numbers in normal murine RAW264.7 macrophages and human A549 lung adenocarcinoma cell lines. Our results indicate that the proposed microfluidic SAW device is capable of monitoring and quantifying cell density of both cell lines in suspension as well as cultured on a 3D-nanofiber scaffold.

32046

Sensors Sensors 2015, 2015, 15, 15, 32045–32055 page–page

2. Working Principle 2. Working Principle The frequency of the SH-SAW shifts when the waves travel through the media inside the The frequency of the SH-SAW shifts waves travel through the media the PDMS PDMS well due to propagation loss. when Whenthethe substrate where the surfaceinside acoustic wave well due to propagation loss. When the substrate where the surface acoustic wave propagates is liquid propagates is liquid loaded, the phase velocity and attenuation of the wave can be related to the loaded, the phase velocity and attenuation the wave be related to the mechanical properties mechanical properties of the media such as of viscosity andcan density, and electrical characteristics such of the media such as viscosity and density, and electrical characteristics such as permittivity and as permittivity and conductivity. The perturbation formula was originally derived from Auld’s conductivity. The perturbation formula was originally derived from Auld’s perturbation theory perturbation theory which applies to gas sensors and then extended to liquid phase applications by which applies gas sensors and then extended to liquid phase applications by J Kondoh et al. [4,22,23]. J Kondoh et al.to[4,22,23]. Change in velocity: Change in velocity: ⁄ )2 ` ∆ +pε ( 1 ´−εREF q)( K2 (pσ {ωq pε1 `+εPIEZO q) ∆V =´ − S 1 “ 2 ⁄ ( ) + ( + V 2 pσ1 {ωq ` pε1 ` εPIEZO q)2 Change in attenuation: Change in attenuation:

(1) (1)

⁄ )( ∆ ( 1 {ωq + εPIEZO)q K2 pσ pεREF ` ∆α “ = S k 2 (pσ1⁄{ωq)2 ` + pε ( 1 `+εPIEZO q)2

(2) (2)

In the the equations, equations, K2 represents representsthe theelectromechanical electromechanical coupling factor substrate, coupling factor of of thethe substrate, εPIEZO is In S is the effective dielectric constant, is the dielectric constant of sample liquid, is the the effective dielectric constant, ε1 is the dielectric constant of sample liquid, εREF is the dielectric dielectricof constant of the reference is the conductivity of theliquid, sampleand liquid, is the constant the reference liquid, σ1liquid, is the conductivity of the sample ω isand the angular angular frequency of the SH-SAW. Based on the equation above, mixtures containing particles of frequency of the SH-SAW. Based on the equation above, mixtures containing particles of different differentcan density easily be distinguished by the device. SH-SAW device. density easilycan be distinguished by the SH-SAW 3. Designand andFabrication Fabricationof of Bio-Sensor Bio-Sensor 3. Design

3.1. Device Device Design Design 3.1. A 3D 3D COMSOL COMSOL model model that that consists consists of of aa two-port two-port resonator resonator was was built built to to characterize characterize changes changes A to the wave propagation characteristics resulting from alterations to mechanical properties inside to the wave propagation characteristics resulting from alterations to mechanical properties inside the the well. A simplified 3D cell model of the Lithium tantalate resonator was built to obtain the well. A simplified 3D cell model of the Lithium tantalate resonator was built to obtain the resonator resonator frequency shifts by the Eigen frequency of the COMSOL The individual frequency shifts by the Eigen frequency module ofmodule the COMSOL software.software. The individual 3D cell 3D cell was set as periodic to simulate the entire SAW sensor with a fairly-simplified was set as periodic to simulate the entire SAW sensor with a fairly-simplified geometry.geometry. Figure 1 Figure 1 illustrates one wavelength cellsimulated of the simulated with interdigital transducers illustrates one wavelength cell of the design design with interdigital transducers (IDTs).(IDTs). Two Two interdigital transducer fingers are illustrated where one of them was connected to the ground. interdigital transducer fingers are illustrated where one of them was connected to the ground. B

A

C

Total displacement (A)

3.00

Total displacement (A)

2.86

400 200 0 100 0

100

200

300 Frequency = 14.0475e6

Frequency = 14.0265e6

Figure 1. A 3D COMSOL model and simulation results based on the 36˝ Y-cut LiTaO3: (A) 3D cell Figuregeometry 1. A 3D COMSOL and simulation results based thea 36°Y-cut LiTaO3: 3D and cell model with mesh;model (B) resonance frequency of the IDTson with 200 nm thick ZnO(A) layer; model geometry with mesh; resonance of the with 200 nm layer; (C) resonance frequency of the(B) IDTs with 12.5frequency K cells media on IDTs the 200 nmathick ZnOthick layerZnO surface. and (C) resonance frequency of the IDTs with 12.5 K cells media on the 200 nm thick ZnO layer surface.

After the model was built and material properties were applied, a mesh was created with total degrees of freedom of 679,924. The mesh consisted of 111,679 domain elements, 28,942 boundary After the model was built and material properties were applied, a mesh was created with total elements and 2892 edge elements. 36˝ Y-cut LiTaO3 with or without a ZnO coating was employed as degrees of freedom of 679,924. The mesh consisted of 111,679 domain elements, 28,942 boundary the choice of substrate to simulate the resonator frequency. The simulation results indicated that the elements and 2892 edge elements. 36°Y-cut LiTaO3 with or without a ZnO coating was employed as 32047 3

Sensors 2015, 15, 32045–32055

operation frequency would be 14.0475 MHz without ZnO while the fabricated operation frequency was experimentally measured as 14.056 MHz. The simulation result of the 3D-cell with ZnO is 14.03296 MHz, while the experimentally measured value is 14.04120 MHz, illustrating the validity of the developed simulation designs. As expected, the shear horizontal wave propagated in the x direction with the substrate polarized in the y direction, as illustrated in Figure 1B,C. Additional device design details are given in Table 1. Table 1. Device Parameters used for the simulation and fabrication of the IDT transducers. PARAMETERS Wavelength (λ) Number of reflecting fingers Finger width Wavelength of reflecting fingers Number of fingers Well diameter SAW velocity ZnO layer thickness Finger height Operation frequency

SETTINGS 297 µm 30 pairs 74.25 µm 297 µm 30 pairs 6.5 mm 4160 m/s 200 nm 100 nm 14.05 MHz

3.2. Device Fabrication The IDTs were fabricated by the traditional micro-lithography methods while the microfluidic well was fabricated by the conventional PDMS micro molding technique. Further details on the fabrication process can be found in our recent reports [24–26]. After the IDTs were fabricated on the lithium tantalate substrate, ZnO sputtering was carried out. A 200 nm thick ZnO film was deposited at 150 ˝ C in 2.5 h. After the ZnO deposition, the PDMS well was bonded to the lithium tantalate substrate after being exposed to 30 s oxygen plasma for increased bonding. The SAW resonator can be used as a propagation delay-line with a pair of IDT transducers that serve to excite and receive the acoustic wave. Therefore a custom-designed oscillatory circuit system was used for quantifying the cell concentrations as shown in the conceptual view in Figure 2. Compared to other detection methods such as network analyzer, an oscillatory circuit was employed as it offers higher stability as well as higher sensitivity [3,27]. In the oscillator circuit detection system, the SAW sensor was employed as the feedback element of the RF amplifier. The relative change of SAW velocity due to mechanical and electrical changes resulted in an oscillation frequency shift. These changes in oscillation frequency were detected with a digital frequency counter. The setup used two variable gain RF amplifiers (Olympus 5073PR and Olympus 5072PR, Olympus NDT Inc., Waltham, MA, USA), a digital frequency counter (Agilent 53220A, Agilent Technologies Inc, Santa Clara, CA, USA), an oscillator (Tektronix TDS2001C, Tektronix Inc., Beaverton, OR, USA), as reported previously [3]. A band pass filter was used on the amplifier to eliminate the frequencies lower than 5 MHz and higher than 20 MHz in the loop. The two oscillation loops were employed to minimize the background noise and relate the frequency shift to the mass loading of different cell concentrations. A constant volume of different cell concentrations media was supplied to the well in the test loop for each experiment. During the experiments, the frequency changes for the test group while the frequency of the control group remained nearly constant. From the perturbation theory, when the surface acoustic waves propagates thru the detection area of the sensor, the phase velocity changes due to mass loading from the cell media. In Equations (3) and (4) presented below, V1 is the surface wave phase velocity of the control group device and V2 is the surface wave phase velocity 1 of the actual tested device.V2 represents the phase velocity of the surface acoustic wave travelling through different cell concentrations. During the experiments, the only real time relative frequency

32048

Sensors 2015, 15, 32045–32055 Sensors 2015, 15, page–page

f2 the experiments, thethe only real time relativeThen frequency wassorted recorded by the©frequency counter. was recorded by frequency counter. the data were by Matlab and plotted out in f1 © Then the data were sorted by Matlab normalized relative frequency shift. and plotted out in normalized relative frequency shift. 1 V ´V ∆V ∆ “ 2− ′2 (3) (3) V = V1 1

f ´ f2 ∆V ∆f ∆ “ ∆ “ 2− ′ f = V= f1 A

(4) (4)

B

Figure Conceptual view of the oscillatory circuit with Figure 2.2.(A)(A) Conceptual view of the oscillatory circuit system. Two system. resonatorsTwo with resonators custom-designed custom-designed oscillatory circuit system were used with one of them as control group; oscillatory circuit system were used with one of them as control group; (B) Fabricated and assembled (B) Fabricated and assembled resonator and fluidic well. resonator and fluidic well.

4. 4. Experimental ExperimentalSetup Setup 4.1. Experimental Protocol for SAW Measurement 4.1. Experimental Protocol for SAW Measurement A549 human lung adenocarcinoma cells were maintained in RPMI media containing 10% fetal A549 human lung adenocarcinoma cells were maintained in RPMI media containing 10% fetal bovine serum (FBS) and 1% penicillin streptomycin. RAW-264.7 murine macrophages (used as an bovine serum (FBS) and 1% penicillin streptomycin. RAW-264.7 murine macrophages (used as an example of a non-cancerous cell) were maintained in Dulbecco’s modified eagle medium (DMEM) example of a non-cancerous cell) were maintained in Dulbecco’s modified eagle medium (DMEM) media containing 10% FBS and 1% penicillin streptomycin. All cells were cultured in a humidified media containing 10% FBS and 1% penicillin streptomycin. All cells were cultured in a humidified incubator at 37 °C in a 5% CO2 atmosphere. Cells were collected via trypsinization and counted incubator at 37 ˝ C in a 5% CO2 atmosphere. Cells were collected via trypsinization and counted using a hemocytometer. For SAW measurement of cells in suspension, cell suspensions of using a hemocytometer. For SAW measurement of cells in suspension, cell suspensions of decreasing decreasing concentration were prepared by serial dilution in phosphate buffered saline (PBS) concentration were prepared by serial dilution in phosphate buffered saline (PBS) containing 1% FBS. containing 1% FBS. Half a minute after the frequency counter started to record, 100 μL of each Half a minute after the frequency counter started to record, 100 µL of each suspension was placed suspension was placed on the chip of test group to record the relative frequency response for a on the chip of test group to record the relative frequency response for a duration of 10 min. After duration of 10 min. After recording each sample, the cell suspension was removed by vacuum and recording each sample, the cell suspension was removed by vacuum and the well was washed with the well was washed with three changes of PBS followed by three changes of water to clean the three changes of PBS followed by three changes of water to clean the sensing area. sensing area. 4.2. Experimental Protocol for Measuring Cell Viability 4.2. Experimental Protocol for Measuring Cell Viability We determined the cell viability using trypan blue staining in combination with a T20TM We determined the cell viability using trypan blue staining in combination with a T20TM automated cell counter (Bio-Rad). Cell suspension was mixed with trypan blue at a ratio of 1:1 automated cell counter (Bio-Rad). Cell suspension was mixed with trypan blue at a ratio of 1:1 respectively and the resulting cell suspension was loaded on to a cassette for measurement in the respectively and the resulting cell suspension was loaded on to a cassette for measurement in the cell counter. The T20TM uses microscopy in conjunction with an algorithm to calculate the total cell TM cell counter. The T20 uses microscopy in conjunction with an algorithm to calculate the total cell count and assesses the cell viability by trypan blue exclusion without any interference from the user. count and assesses the cell viability by trypan blue exclusion without any interference from the user. The advantage of this methodology is that it ensures reproducibility in the cell count independent of The advantage of this methodology is that it ensures reproducibility in the cell count independent the users. Additionally, we validated the instrument for measuring the cell viability by using a cell of the users. Additionally, we validated the instrument for measuring the cell viability by using a suspension sample that was heated for 15 min at 56 ˝ C to induce cell death and show that the cell cell suspension sample that was heated for 15 min at 56 °C to induce cell death and show that the counter was able to detect increase in cell death (Supplementary Material Figure 1). cell counter was able to detect increase in cell death (Supplementary Material Figure 1). 4.3. Experimental Protocol for Measuring Cell Proliferation 4.3. Experimental Protocol for Measuring Cell Proliferation To examine any long term effects on cell proliferation a re-plating experiment was performed in Tocells examine any long and termthen effects on cell a re-plating experiment was performed which were collected seeded ontoproliferation a 96 well culture plate after SAW measurements and in which cells were collected and then seeded onto a 96 well culture plate after SAW measurements allowed to grow for three days. Cell number and morphology were compared to untested control cells and allowed to grow for(Olympus three days. Cell number and the morphology were compared to untested by both light microscopy BX51) and by staining cells with Hoechst 33342 (NucBlue, Life control cells by both light microscopy (Olympus BX51) and by staining the cells with Hoechst 33342 (NucBlue, Life Technologies) and then capturing images using fluorescence microscopy (Olympus 32049 5

BX51). This assay was designed to reveal any changes to the cell proliferation rate as a result of SAW measurement. 4.4. Experimental for Culturing 3D Tumoroids SensorsProtocol 2015, 15, 32045–32055 For growing cell cultures in three-dimensions (3D), we used a fibrous scaffold developed by Technologies) and then capturing images usingof fluorescence microscopywhen (Olympus BX51). This cancer assay cells. our lab (3P scaffold) which promotes the growth 3D “tumoroids” seeded with was designed to reveal any changes to the cell proliferation rate as a result of SAW measurement. 3P scaffold was prepared by electrospinning as described previously [28]. Scaffold was placed into 4.4. and Experimental Protocol for Culturing 3D Tumoroids a 96 well plate 5000 A549 cells were seeded into each well in RPMI media. Cells were allowed to grow on theFor scaffold for cultures eight in days and the culture media wasscaffold changed everyby two growing cell three-dimensions (3D), we used a fibrous developed our days. lab (3P scaffold) which promotes the growth of 3D “tumoroids” when seeded with cancer cells. 3P Successful growth of cells in 3D was confirmed by staining the cells with Hoechst 33342 (NucBlue, scaffold was prepared by electrospinning as described previously [28]. Scaffold was placed into a 96 Life Technologies) and then capturing images using fluorescence microscopy (Olympus BX51). 5. Results

well plate and 5000 A549 cells were seeded into each well in RPMI media. Cells were allowed to grow on the scaffold for eight days and the culture media was changed every two days. Successful growth and Discussion of cells in 3D was confirmed by staining the cells with Hoechst 33342 (NucBlue, Life Technologies) and then capturing images using fluorescence microscopy (Olympus BX51).

5.1. Cell Viability and Cell Proliferation Is Not Affected after SAW Measurements 5. Results and Discussion

One of our first concerns was that the cellular stress due to seeding of the cells in our device Cell Viability and Cell Proliferation Is Not Affected SAW Measurements followed by 5.1. exposure to acoustic waves would haveafter a negative impact on the cells’ viability and One of our first was thatthe theprocedure’s cellular stress due to seedingnature of the cells our device their ability to proliferate. To concerns demonstrate innocuous andinthus its utility in followed by exposure to acoustic waves would have a negative impact on the cells’ viability translational lab setting, we determined cell viability as described in the experimental and setup. We their ability to proliferate. To demonstrate the procedure’s innocuous nature and thus its utility in tested the viability of both normal (RAW 264.7) and cancerous (A549) cells immediately following translational lab setting, we determined cell viability as described in the experimental setup. We SAW measurement trypan bluenormal exclusion. replicates for each concentration of tested theby viability of both (RAW Three 264.7) and cancerouswere (A549)tested cells immediately following SAWA measurement trypanwas blue used exclusion. replicates weresignificant tested for each concentration both cell lines. student’sbyt-test to Three determine any difference in ofviability both cell lines. A student’s t-test was used to determine any significant difference in viability between between each pair of control and SAW tested groups. A p-value ≤ 0.05 was considered to be each pair of control and SAW tested groups. A p-value ď 0.05 was considered to be significant. significant. As seen in Figure 3, out of the 12 groups compared, only one showed any significant As seen in Figure 3, out of the 12 groups compared, only one showed any significant difference difference in viability. in viability.

Figure 3. CellFigure viability following SAW measurement. Immediately following SAW measurement the 3. Cell viability following SAW measurement. Immediately following SAW measurement the viability of (A) A549 and (B) RAW 264.7 cells was determined by trypan blue exclusion. Data is plotted viability of (A) A549 and (B) RAW 264.7 cells was determined by trypan blue exclusion. Data is as mean ˘ standard error of the mean. The data is a representative of a study that was performed in plotted as mean ± standard error of the mean. The data is a representative of a study that was triplicates and performed at least two independent times. A student’s t-test was used to evaluate performed insignificance triplicates(* and performed at least independent times. A student’s t-test was used to p ď 0.05, when compared to two control). evaluate significance (* p ≤ 0.05, when compared to control).

A

Next, we looked at the long-term effect of the SAW measurement on cell proliferation by performing a re-plating assay as described earlier. Briefly, A549 cells were collected and then seeded B C onto a 96-well culture plate after SAW measurements. As seen in Figure 4, A549 cells exposed to SAW (Figure 4A) and control untested cells which were not exposed to the SAW device (Figure 4C) reveals no obvious changes to cell morphology or growth rate after 72 h. The nuclear staining in Figure 4B shows that cells have in-tact nuclei and appear healthy.

32050

Figure 4. Cell proliferation following SAW measurement. Immediately following SAW

Figure 3. Cell viability following SAW measurement. Immediately following SAW measurement the viability of (A) A549 and (B) RAW 264.7 cells was determined by trypan blue exclusion. Data is plotted as mean ± standard error of the mean. The data is a representative of a study that was Sensors 2015, 15, 32045–32055 performed in triplicates and performed at least two independent times. A student’s t-test was used to evaluate significance (* p ≤ 0.05, when compared to control).

A

B

C

Sensors 2015, 15, page–page

Next, we looked at the long-term effect of the SAW measurement on cell proliferation by performing a re-plating assay as described earlier. Briefly, A549 cells were collected and then seeded Figure 4. Cell proliferation following SAW measurement. Immediately following SAW onto a 96-well culture plate after SAW As seen in Figure 4, A549 cells exposed to Figure 4. Cell proliferation following SAWmeasurements. measurement. Immediately following SAW measurement measurement A549 cells were seeded onto a 96-well plate at a density of 5000 cells per well. SAW (Figure 4A) and control untested cells which were not exposed to the SAW device (Figure 4C) A549 cells were seeded ontoofacontrol 96-well plate at groups a density 5000(A) cells percells well. Representative images and tested are of shown: A549 72 Representative h after SAW test; images reveals no obvious changes to cell morphology or growth rate after 72 h. The nuclear staining of control tested groups are cells shown: (A) A549 cells h after test; of in (B) and NucBlue staining of A549 72 h after test; and (C) 72 Control cellsSAW after 72 h. (B) NucBlue staining Figure 4B shows thattest; cellsand have nuclei and appear A549 cells 72 h after (C)in-tact Control cells after 72 h. healthy. 6

Frequency Shift Increases with IncreasingCell CellConcentration Concentrationand andSensitivity Sensitivity Is Is Further 5.2. 5.2. Frequency Shift Increases with Increasing Further Aided Aidedby bythe the Use Use of ZnO of ZnO Once confirmed that our bio-sensingdevice deviceand andmeasurement measurement protocol Once wewe confirmed that our bio-sensing protocol was wasbio-compatible, bio-compatible, we next wanted to determine the sensitivity of our device in accurately measuring cell we next wanted to determine the sensitivity of our device in accurately measuring cell concentrations. concentrations. For this we used two variations of our device, one that was coated with ZnO and the For this we used two variations of our device, one that was coated with ZnO and the second that was second that was kept bare. The SAW measurement protocol for both the devices was kept constant kept bare. The SAW measurement protocol for both the devices was kept constant as described in as described in the experimental setup section. Both the non-cancerous (RAW 264.7) and cancerous the experimental setup section. Both the non-cancerous (RAW 264.7) and cancerous (A549) cells were (A549) cells were examined in the devices at 6250, 12,500, 25,000 and 50,000 cells per 100 μL. The examined in the devices at 6250, 12,500, 25,000 and 50,000 cells per 100 µL. The concentration of concentration of cells were chosen based on cell density that we would encounter when performing cellsactual wereresearch chosen based density that5,we encounter when performing actual research studies.on Ascell seen in Figure we would saw a cell dependent increase in the frequency shift studies. As seen in Figure 5, we saw a cell dependent increase in the frequency shift both in the in both the cell lines tested. Interestingly, the layer of ZnO increased the sensitivity of theindevice cell recording lines tested. Interestingly, the layer of ZnO increased thetested. sensitivity of the comparing device in recording changes in cell numbers in both of the cell lines Specifically, sensors changes in cell numbers in both of the cell lines tested. Specifically, comparing sensors containing the containing the ZnO layer (Figure 5B,D) to those with the bare substrate (Figure 5A,C) revealed that ZnOthe layer (Figure 5B,D) to those with the bare substrate (Figure 5A,C) revealed that the ZnO layer ZnO layer increased the relative frequency response by four times which means it increased the increased the of relative frequency response by four times which means it increased the sensitivity of sensitivity the device. the device.

Figure 5. Relative frequency shift response to the different cell concentrations. (A,C) The bare SAW Figure 5. Relative frequency shift response to the different cell concentrations. (A,C) The bare SAW resonator response to cell with concentrations of 3000, 6250, 12,500, 25,000 and 50,000 in 100 µL media resonator response to cell with concentrations of 3000, 6250, 12,500, 25,000 and 50,000 in 100 μL for each test, A549 (A) and RAW 264.7 (C). (B,D) The SAW resonator coating with ZnO layer response media for each test, A549 (A) and RAW 264.7 (C). (B,D) The SAW resonator coating with ZnO layer to cells with concentration of 3000, 6250, 12,500, 25,000 and 50,000 in 100 µL media for each test, A549 response to cells with concentration of 3000, 6250, 12,500, 25,000 and 50,000 in 100 μL media for each (B) and RAW 264.7 (D). Data are plotted as mean ˘ 95% CI. The data are representative of a study test, A549 (B) and RAW 264.7 (D). Data are plotted as mean ± 95% CI. The data are representative of a that was performed in triplicates and performed at least two independent times. A student’s t-test study that was performed in triplicates and performed at least two independent times. A student’s wast-test usedwas to evaluate significance (p ď 0.05). indicates a significant difference in the shift used to evaluate significance (p ≤ *0.05). * indicates a significant difference infrequency the frequency between the labeled theand adjacent lower concentration. Best fitBest curves were were calculated using shift between the group labeledand group the adjacent lower concentration. fit curves calculated a second order polynomial model in the GraphPad Prism software application. using a second order polynomial model in the GraphPad Prism software application.

5.3. SAW Measurements of Cell Density Match Simulation Results 32051 Having confirmed that our device accurately distinguishes between cell numbers, we further optimized our device for future experimental protocols. For this, we decided to establish a working theoretical model based on the 3D COMSOL model by adding mass loading module that best

Sensors 2015, 15, 32045–32055

5.3. SAW Measurements of Cell Density Match Simulation Results confirmed SensorsHaving 2015, 15, page–page

that our device accurately distinguishes between cell numbers, we further optimized our device for future experimental protocols. For this, we decided to establish a working mimicked bio-sensing device. OneCOMSOL advantage of having such a model is that itmodule will enable to theoreticalour model based on the 3D model by adding mass loading thatus best tweak several our device and run a simulation experiment to spend mimicked ourparameters bio-sensingon device. One advantage of having such a modelwithout is that having it will enable us time andseveral moneyparameters on actual on experiments. Towards this goal, experiment we set up without the simulation studies, to tweak our device and run a simulation having to spend wherein weight of eachexperiments. individual cell was assumed beset 1 pg different concentrations of time andthe money on actual Towards this goal,towe up and the simulation studies, wherein cells were added the ZnO surface (thickness nm) onand 36°Y-cut LiTaO 3 substrate by the weight of eachtoindividual cell was assumed200 to be 1 pg different concentrations ofmodifying cells were the mass the developed The simulated 0, 6250,the 12,500, added to loading the ZnOinsurface (thicknessmodel. 200 nm) oncell 36˝concentrations Y-cut LiTaO3 substrate bywere modifying mass 25,000, 50,000, and 100,000 cells/microwell. After the cell (mass) loading was applied, the relative loading in the developed model. The cell concentrations simulated were 0, 6250, 12,500, 25,000, frequency to the different cell concentrations was simulated 50,000, andresponse 100,000 cells/microwell. After the cell (mass) loading was (Figure applied,2). the relative frequency In order to normalize the frequency shift obtained in response to the cell concentration chance, response to the different cell concentrations was simulated (Figure 2). ⁄ relative frequency response, which is the shift frequency shiftinover the operation (∆ chance, ), was In order to normalize the frequency obtained response to the cellfrequency concentration plotted. With the cell concentration increasing (henceshift the over mass), phase velocity of the relative frequency response, which is the frequency thethe operation frequency ( ∆ substrate f { f q, was decreased which resulted in decreasing frequency. simulated relative frequency was plotted. With the cell concentration increasing (hence The the mass), the phase velocity of theshift substrate found to bewhich aboutresulted one order of magnitude higher The thansimulated the experimental results, without applying decreased in decreasing frequency. relative frequency shift was found any the simulation when presenting thethan results. The mismatchresults, between the raw simulation to betoabout one order of magnitude higher the experimental without applying any todata the presented and the experimental data is expected which may be attributed to the following factors. simulation when presenting the results. The mismatch between the raw simulation data presented First, theexperimental weight of thedata individual cell which was assumed be 1pg in simulations for referencing and the is expected may be to attributed to the the following factors. First, the purposes, whereas in actual experimentation the weight of the cells can differ significantly weight of the individual cell was assumed to be 1pg in the simulations for referencing purposes, depending the experimentation cell type. Second,the theweight cells with media placed on the bottom of the well in whereas in on actual of the cells were can differ significantly depending on the the experiments performed. Onwere the placed other hand massofloading entire experiments sensor chip cell actual type. Second, the cells with media on thethe bottom the wellon in the the actual surface was simulated for the 3Dthe COMSOL model.onNonetheless, there is a very good concordance performed. On the other hand mass loading the entire sensor chip surface was simulated of and experimental data, withthere a very matchconcordance in the trendsofobtained (Figureand 6). forthe thesimulated 3D COMSOL model. Nonetheless, is asimilar very good the simulated One can easily apply a correction factor to the simulation for the specific cell lines or application if experimental data, with a very similar match in the trends obtained (Figure 6). One can easily apply closer frequency shift magnitude is needed. a correction factor to the simulation for the specific cell lines or application if closer frequency shift magnitude is needed.

Figure 6. SAWs experiment data match the tendency of the simulation results on the ZnO coated Figure 6. SAWs experiment data match the tendency of the simulation results on the ZnO coated sensor. Experimental data are plotted as mean ˘ standard error of the mean The data is a sensor. Experimental data as mean ± standard error of the mean The data is a representative of a study thatare wasplotted performed in triplicates. representative of a study that was performed in triplicates.

5.4. SAW Measurements Aid in Monitoring Growth of A549 3D Spheroid Cultures 5.4. SAW Measurements Aid in Monitoring Growth of A549 3D Spheroid Cultures Use of 3D cell culture techniques in cancer research is rapidly expanding due to the limited Use of 3D cell culture techniques in cancer research is rapidly expanding due to the limited ability of traditional 2D culture to accurately model in vivo cell behavior. To test whether the SAW ability of traditional 2D culture to accurately model in vivo cell behavior. To test whether the SAW device is able to measure the cell density of 3D spheroids, (also referred to as tumoroids) growing on device is able to measure the cell density of 3D spheroids, (also referred to as tumoroids) growing a fiber matrix we cultured A549 cells on the matrix. A549 cells were allowed to grow on the scaffold on a fiber matrix we cultured A549 cells on the matrix. A549 cells were allowed to grow on the for eight days and the culture media was changed every two days (Figure 7A). The 3P scaffold alone scaffold for eight days and the culture media was changed every two days (Figure 7A). The 3P was first measured as control and corresponding reading was designated as Day 0. On Days 4, 6, and scaffold alone was first measured as control and corresponding reading was designated as Day 0. On Days 4, 6, and 8 scaffolds were removed from the plate and assayed on the SAW sensor. Data shown in Figure 7B demonstrate that the sensor was able to detect the change in density resulting 32052 from cell proliferation over time in the 3D environment. There was a linear increase in frequency shifts observed in A549 tumoroids with time. This increase was similar to increases in tumoroid size and number reported previously for other cancer cell lines [28].

Sensors 2015, 15, 32045–32055

8 scaffolds were removed from the plate and assayed on the SAW sensor. Data shown in Figure 7B demonstrate that the sensor was able to detect the change in density resulting from cell proliferation over time in the 3D environment. There was a linear increase in frequency shifts observed in A549 tumoroids with time. This increase was similar to increases in tumoroid size and number reported previously cancer cell lines [28]. Sensors 2015,for 15,other page–page A

B

Figure 7. (A) Representative image of A549 cells growing in 3D tumoroid structures on the 3P on scaffold Figure 7. (A) Representative image of A549 cells growing in 3D tumoroid structures the 3P using NucBlue stain on Day 8 ofon culture. cells cultured on the 3P scaffold for3P scaffold usingnuclear NucBlue nuclear stain Day 8(B) of A549 culture. (B)were A549 cells were cultured on the eight days.for Oneight Daysdays. 0 (scaffold with0 no cells), 4, 6, and 8 scaffolds transferred scaffold On Days (scaffold with no cells), 4, 6, were and 8collected scaffoldsand were collected to and thetransferred SAW device with ZnO layer for measurement. Data is plotted as mean ˘ standard error of the to the SAW device with ZnO layer for measurement. Data is plotted as mean ± standard mean. a representative of aa study that was performed in triplicates. * = significant errorThe of data the is mean. The data is representative of a study that was performed in increase triplicates. from Day 0; ** = significant increase from Day 4 (p ď 0.05). * = significant increase from Day 0; ** = significant increase from Day 4 (p ≤ 0.05).

6. 6.Conclusions Conclusions Based onon these results, the acoustic measurement procedure seems toto have nono illill effect onon cell Based these results, the acoustic measurement procedure seems have effect cell viability in either the A549 cancer cell line or the RAW 264.7 macrophages. Cell proliferation was viability in either the A549 cancer cell line or the RAW 264.7 macrophages. Cell proliferation was also unaffected byby SAW measurements in in A549. The device’s ability toto detect changes inin cell density also unaffected SAW measurements A549. The device’s ability detect changes cell density onon the 3D3D scaffold over time along with itsits biocompatibility reveal great potential forfor this device toto the scaffold over time along with biocompatibility reveal great potential this device bebe incorporated into 3D3D in in vitro cancer models. thus created would enable continuous incorporated into vitro cancer models.The Theplatform platform thus created would enable continuous real-time measurement of cell growth in a 3D during bio assays drug screens, real-time measurement of cell growth in environment a 3D environment during bioincluding assays including drug multi cell co-cultures, gene knockdown/knockout, etc. screens, multi cell co-cultures, gene knockdown/knockout, etc. There are There areseveral severalpotential potentialtranslational translationalimplications implicationsfor forthese. these. Acoustic Acoustic biosensing involves a a highly highly sensitive sensitiveand andtunable tunable SAW (37–46), which canperformed be performed without any electrode SAW (37–46), which can be without any electrode touching touching the tumoroids and acoustical response can be acquired independent of existence of a the tumoroids and acoustical response can be acquired independent of existence of a magnetic/electrical magnetic/electrical field andnanoparticles iron oxide/MnO nanoparticles in the flow field. The potential for field and iron oxide/MnO in the flow field. The potential for miniaturization and miniaturization integration of into complex functions into “multi-cell on chip” exists,to integration of and complex functions “multi-cell tumoroids on chip”tumoroids exists, which is expected which is expected to revolutionize tracking of biomarkers diagnostics andin revolutionize real-time tracking of real-time biomarkers and clinical diagnosticsand andclinical prognostics of cancers prognostics of cancers in for a point-of-care setting for personalizing therapy. In addition, monitoring of a point-of-care setting personalizing therapy. In addition, monitoring of physiologic/metabolic physiologic/metabolic tumor bio-sensing markers viaisacoustic bio-sensing expected to of increase resemblance tumor markers via acoustic expected to increaseisresemblance tumoroid cultures to ofintumoroid cultures in vivoa tumors provide a precise, stable, and well-defined vivo tumors and to provide precise, and stable, and well-defined culture environment forculture cellular environment for cellular assays. assays. Acknowledgments:SSM SSMand andSM SMlike liketotoacknowledge acknowledgefinancial financialsupport supportthrough throughNCI NCIRO1CA152005 RO1CA152005and and Acknowledgments: HHSN261201400022C. Ryan Green is supported American Heart Association pre-doctoral fellowship. HHSN261201400022C. Ryan Green is supported byby anan American Heart Association pre-doctoral fellowship. Author Conducting research, research, analysis analysisofofdata data and writing of the manuscript. AuthorContributions: Contributions:Tao Tao Wang: Wang: Conducting and writing of the manuscript. Rasim Rasim Guldiken: Research and writing of the manuscript. RyanConducting Green: Conducting research,ofanalysis Guldiken: Research designdesign and writing of the manuscript. Ryan Green: research, analysis data and of data and writing of the manuscript. Mark Howell: Conducting research. Subhra Mohapatra: Research writing of the manuscript. Mark Howell: Conducting research. Subhra Mohapatra: Research design and writing design and writing of the manuscript. Rajesh R Nair: Conducting research, analysis of data and writing of the manuscript. Rajesh R Nair: Research Conducting research, analysis and writing of the manuscript. Shyam theofmanuscript. Shyam Mohapatra: design and writing of of thedata manuscript. Mohapatra: Research design and writing of the manuscript. Conflicts of Interest: The authors declare no conflict of interest. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3.

Shen, C.; Liou, S. Surface acoustic wave gas monitor for ppm ammonia detection. Sens. Actuators B Chem. 2008, 131, 673–679. 32053 Onen, O.; Ahmad, A.A.; Guldiken, R.; Gallant, N.D. Surface Modification on Acoustic Wave Biosensors for Enhanced Specificity. Sensors 2012, 12, 12317–12328. Onen, O.; Sisman, A.; Gallant, N.D.; Kruk, P.; Guldiken, R. A urinary Bcl-2 surface acoustic wave biosensor

Sensors 2015, 15, 32045–32055

References 1. 2. 3. 4. 5.

6. 7. 8. 9.

10. 11. 12.

13. 14. 15. 16. 17. 18. 19.

20. 21.

22. 23.

Shen, C.; Liou, S. Surface acoustic wave gas monitor for ppm ammonia detection. Sens. Actuators B Chem. 2008, 131, 673–679. [CrossRef] Onen, O.; Ahmad, A.A.; Guldiken, R.; Gallant, N.D. Surface Modification on Acoustic Wave Biosensors for Enhanced Specificity. Sensors 2012, 12, 12317–12328. [CrossRef] [PubMed] Onen, O.; Sisman, A.; Gallant, N.D.; Kruk, P.; Guldiken, R. A urinary Bcl-2 surface acoustic wave biosensor for early ovarian cancer detection. Sensors 2012, 12, 7423–7437. [CrossRef] [PubMed] Vivancos, J.-L.; Rácz, Z.; Cole, M.; Gardner, J.W. Surface acoustic wave based analytical system for the detection of liquid detergents. Sens. Actuators B Chem. 2012, 171–172, 469–477. [CrossRef] Pomowski, A.; Baricham, C.; Rapp, B.E.; Matern, A.; Länge, K. Acoustic Biosensors Coated With Phosphorylcholine Groups for Label-Free Detection of Human C-Reactive Protein in Serum. IEEE Sens. J. 2015, 15, 4388–4392. [CrossRef] Nomura, T.; Saitoh, A.; Horikoshi, Y. Measurement of acoustic properties of liquid using liquid flow SH-SAW sensor system. Sens. Actuators B Chem. 2001, 76, 69–73. [CrossRef] Deobagkar, D.D.; Limaye, V.; Sinha, S.; Yadava, R.D.S. Acoustic wave immunosensing of Escherichia coli in water. Sens. Actuators B. Chem. 2005, 104, 85–89. [CrossRef] Roederer, J.E.; Bastiaans, G.J. Microgravimetric Immunoassay with Piezoelectric Crystals. Anal. Chem. 1983, 55, 2333–2336. [CrossRef] Kondoh, J.; Tabushi, S.; Matsui, Y.; Shiokawa, S. Development of methanol sensor using a shear horizontal surface acoustic wave device for a direct methanol fuel cell. Sens. Actuators B Chem. 2008, 129, 575–580. [CrossRef] Nomura, T.; Saitoh, A.; Miyazaki, T. Liquid sensor probe using reflecting SH-SAW delay line. Sens. Actuators B Chem. 2003, 91, 298–302. [CrossRef] Litao, Y.; Fu, Y.; Li, Z.; Li, Y.; Ma, J.; Placido, F.; Walton, A.J.; Zu, X. Love mode surface acoustic wave ultraviolet sensor using ZnO films deposited on. Sens. Actuators A. Phys. 2013, 193, 87–94. Lam, C.S. A Review of the Recent Development of Temperature Stable Cuts of Quartz for SAW Applications. In Proceedings of the Fourth International Symposium on Acoustic Wave Devices for Future Mobile Communication Systems, Chiba, Japan, 3–5 March 2010. Shilton, R.J.; Travagliati, M.; Beltram, F.; Cecchini, M. Microfluidic pumping through miniaturized channels driven by ultra-high frequency surface acoustic waves. Appl. Phys. Lett. 2014, 105. [CrossRef] Li, F. The Acoustic Wave Sensor and Soft Lithography Technologies for Cell Biological Studies; ProQuest, UMI Dissertation Publishing: Ann Arbor, MI, USA, 2011. Jo, M.C.; Guldiken, R. Effects of polydimethylsiloxane (PDMS) microchannels on surface acoustic wave-based microfluidic devices. Microelectron. Eng. 2014, 113, 98–104. [CrossRef] Oh, H.; Lee, K.J.; Baek, J.; Yang, S.S.; Lee, K. Development of a high sensitive pH sensor based on shear horizontal surface acoustic wave with ZnO nanoparticles. Microelectron. Eng. 2013, 111, 154–159. [CrossRef] Chivukula, V.; Ciplys, D.; Shur, M.; Dutta, P. ZnO nanoparticle surface acoustic wave UV sensor. Appl. Phys. Lett. 2010, 96, 3–6. [CrossRef] Powell, D.A.; Kalantar-zadeh, K.; Wlodarski, W. Numerical calculation of SAW sensitivity: Application to ZnO/LiTaO3 transducers. Sens. Actuators A Phys. 2004, 115, 456–461. [CrossRef] Powell, D.A.; Kalantar-zadeh, K.; Ippolito, S.; Wlodarski, W. A layered SAW device based on ZnO/LiTaO3 for liquid media sensing applications. In Proceedings of the 2002 IEEE International Ultrasonics Symposium, Munich, Germany, 8–11 October 2002; Volume 1, pp. 493–496. Chang, R.-C.; Chu, S.-Y.; Hong, C.-S.; Chuang, Y.-T. A study of Love wave devices in ZnO/Quartz and ZnO/LiTaO3 structures. Thin Solid Films 2006, 498, 146–151. [CrossRef] Fu, Y.Q.; Luo, J.K.; Du, X.Y.; Flewitt, A.J.; Li, Y.; Markx, G.H.; Walton, A.J.; Milne, W.I. Recent developments on ZnO films for acoustic wave based bio-sensing and microfluidic applications: A review. Sens. Actuators B Chem. 2010, 143, 606–619. [CrossRef] Kondoh, J. A liquid-phase sensor using shear horizontal surface acoustic wave devices. Electron. Commun. Jpn. 2013, 96, 41–49. [CrossRef] Onen, O.; Guldiken, R. Investigation of guided surface acoustic wave sensors by analytical modeling and perturbation analysis. Sens. Actuators A Phys. 2014, 205, 38–46. [CrossRef]

32054

Sensors 2015, 15, 32045–32055

24. 25. 26. 27. 28.

Guldiken, R.; Jo, M.C.; Gallant, N.D.; Demirci, U.; Zhe, J. Sheathless size-based acoustic particle separation. Sensors 2012, 12, 905–922. [CrossRef] [PubMed] Jo, M.C.; Guldiken, R. Active density-based separation using standing surface acoustic waves. Sens. Actuators A Phys. 2012, 187, 22–28. [CrossRef] Jo, M.C.; Guldiken, R. Dual surface acoustic wave-based active mixing in a microfluidic channel. Sens. Actuators A Phys. 2013, 196, 1–7. [CrossRef] Rocha-Gaso, M.-I.; March-Iborra, C.; Montoya-Baides, A.; Arnau-Vives, A. Surface generated acoustic wave biosensors for the detection of pathogens: A review. Sensors 2009, 9, 5740–5769. [CrossRef] [PubMed] Girard, Y.K.; Wang, C.; Ravi, S.; Howell, M.C.; Mallela, J.; Alibrahim, M.; Green, R.; Hellermann, G.; Mohapatra, S.S.; Mohapatra, S. A 3D Fibrous Scaffold Inducing Tumoroids: A Platform for Anticancer Drug Development. PLoS ONE 2013, 8. [CrossRef] [PubMed] © 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

32055