Functional Analyses of Live-cell Membrane Proteins Using Ion-sensitive Field-effect Transistor

Abstract

Functional analyses of the membrane proteins on live cells using ion-sensitive field effect transistors (ISFETs) are described in this review. Expressions of human epidermal growth factor receptor (HER2) and epidermal growth factor receptor (EGFR) on live cancer cells have been detected using cell-based field effect transistors (FETs) in combination with enzymatic signal amplification. A good correlation could be obtained between the pH values measured with the cell-based FETs and the fluorescence intensities measured using the fluorescence-activated cell sorting (FACS), with a correlation coefficient of 0.976. The interactions between membrane proteins/transporters and ligands at cell membranes using a cell-based FET with an oocyte were monitored non-invasively. Xenopus laevis oocytes were injected with the capped human organic anion transporting peptide C (hOATP-C) cRNA. Estrone-3-sulfate (E3S) was used as a substrate for hOATP-C during the uptake measurements. The transporting kinetics of the substrate when mediated by the wild-type and the mutant-type transporters could be distinguished using the cell-based FETs. It was found that the signal generation mechanism of the cell-based transistor could be explained by direct or indirect proton transport via the transporters. Measurements of expression levels of membrane proteins is important to analyze their signaling pathways and cellular outcomes. Moreover, membrane proteins and transporters constitute one of the most extensively studied classes of drug targets. Therefore, a system based on cell-based FETs would be suitable for rapid and cost-effective identification of biomarkers and high throughput analysis of drug candidates.

1. Introduction

Since the ion-sensitive field-effect transistor (ISFET) was first reported in 1970,¹ various types of solid-state biosensors have been reported that used ISFETs in combination with biomolecular recognition reactions.² Enzymatic reactions on the gate insulator surfaces of ISFETs were first designed to detect penicillin,³ urea,⁴ and glucose.⁵ In addition to biomolecules, cells were placed directly on the surfaces of these transistors. For example, a neuron was captured on the surface of an open gate transistor to realize multiple recordings in neurons and neural networks.⁶ The electronic signals obtained matched the shape of the action potential and the magnitudes of the voltage signals ranged up to 25 % of the intracellular voltage changes. The voltage signal was considered to be generated by capacitive coupling that occurred at the junction between the cell plasma membrane and the transistor gate oxide. Several studies using high density arrays of transistors were subsequently reported to have achieved higher spatiotemporal resolutions and higher signal-to-noise ratios. A capacitor array with 400 stimulation sites was fabricated to apply electrical stimuli at different positions in the same retinal circuit, including a retinal ganglion cell.⁷ The area of this capacitor was 50 × 50 µm² and the gap between adjacent capacitors was 0.5 µm. Ex vivo rabbit retinas were interfaced and activated in either epiretinal or subretinal configurations using low current densities. The stimulation sites were then miniaturized further and integrated to achieve high spatial resolution. High-density microelectrode arrays featuring 26400 bidirectional electrodes at a pitch of 17.5 µm over an electrode area of 5 × 9 µm² were developed to realize single-neuron stimulation.⁸ Because of the high spatial resolution of these microelectrodes and their close proximity to their respective neurons, it was possible to simulate neurons with signal amplitudes of 70 mV or 100 nA and generate action potential initiation and propagation. The miniaturized electrodes and their densely packed configuration were beneficial for stimulation of neurons and could improve the stimulation accuracy while also enabling reduced power consumption.

In addition to the stimulation of neurons and the detection of action potentials, ISFETs have also been used to monitor the metabolic activity of living cells. Living cells are sensitive to their environments and show their responses using specific metabolic pathways. Environmental change would induce alterations in the respiratory activity of these cells. Because an increase in carbon dioxide due to cell respiration causes a pH change in the solution, the extracellular acidification rate of cell cultures can be used as an indicator of cellular metabolism and respiration activity.⁹ For example, biosensors using living yeast cells have been reported for use in the detection of toxic agents including cationic and anionic surfactants, Hg²⁺, and Cd²⁺ ions.¹⁰ The yeast cells were immobilized in agarose gel, which was in close contact with the gate surface of the FET. The detection limits for the cationic surfactants, anionic surfactants, Hg²⁺ ions, and Cd²⁺ ions were 0.02, 0.05, 0.02, and 0.01 mM, respectively. A microbial biosensor was also fabricated using Escherichia coli bacteria and an ISFET.¹¹ The bacteria were immobilized in agarose gels that were formed on the ISFET surface. The extracellular acidification rate could then be analyzed with or without the addition of a culture medium or an antibiotic solution. The initial acidification and subsequent alkalization could be observed because of the bacterial metabolism of carbohydrates and peptones. When the concentration of the streptomycin antibiotic increased, the responses of the microbial biosensor decreased because of the reduction in metabolic activity. However, the bacterial activity could be detected for several days. An ISFET was also fabricated using complementary metal-oxide-semiconductor (CMOS) technology to monitor cellular acidification within tumor cells.¹² Colon adenocarcinoma cells (LS 174T) were cultured on the ISFET surface. The pH in the microenvironment that was defined by the space between the ISFET and the cell membrane was reduced by approximately 0.52 ± 0.06 pH units when compared with the pH of the measurement medium. This acidification of the microenvironment would be caused by the enhanced acid extrusion of the tumor cells. The reduced microenvironmental pH close to the surface and the acidification rate were mainly caused by glycolysis. Chondrocytes were also cultured on the gate surface for periods as long as several weeks, and the extracellular matrix (ECM) production was then measured to estimate the metabolic activities.¹³ The local pH change that occurred near the cultured chondrocyte, which served as a measure of the energy metabolism of the living cells, was evaluated for three weeks after the addition of a biologically active substance. The electrical signal change increased gradually over the three weeks with the addition of ascorbic acid phosphate magnesium salt n-hydrate (APM), which activated the chondrocyte’s metabolic activities. The amounts of the main components of the ECM that were quantified via colorimetry were found to have a linear relationship with the electrical signal changes in the ISFET. The respiration activity of embryos was also evaluated using an ISFET.¹⁴ This noninvasive process for detection of embryo activity could be applied to embryo diagnosis for in vitro fertilization of mammals.

On the other hand, there are a number of mechanisms by which cells are able to receive information from their external environment and integrate and transmit intracellular responses. One of the major mechanisms of signal transduction is based on the activity of the superfamily of transmembrane receptor tyrosine kinases, which mediate numerous cell activities.¹⁵ Receptor tyrosine kinases are not only important regulators of normal cellular functions, but also play important roles in the development and progression of many types of cancer.¹⁶ Mutations in these receptor tyrosine kinases lead to the activation of a series of signal transduction cascades, which have many effects on protein expression. Therefore, quantification of the expression levels of the specific membrane proteins is highly important for cancer diagnosis and treatment. Other types of transmembrane proteins function as gateways that allow transport of specific substances across membranes. Typical examples of the transmembrane proteins in this category include transporters and ion channels. These proteins often undergo major structural changes to move substances across membranes. In drug discovery, there is a need for a technology that can perform rapid analysis of the interactions between transporters on the surfaces of cell membranes and drug candidate compounds (substrates).

In this review, we focus on functional analyses of the membrane proteins on live cells in combination with ISFETs. Detection of cancer biomarkers was first described using a breast cancer cell line and ISFET. Epidermal growth factor receptor (EGFR) and human epidermal growth factor receptor (HER2) were used as model proteins for cancer biomarkers, and detected based on enzymatic chemical amplification using glucose oxidase and urease, respectively. Interactions between transporters and their substrates were detected through proton transport using a cell-based FET as the second example of functional analyses of the membrane proteins. Transporters were expressed at cell membrane of an oocyte and uptake experiments were performed on the surface of ISFET. The advantages of use of ISFETs include large-scale integration of multiple sensors and sensing area miniaturization because advanced semiconductor technologies can be used for sensor fabrication and production. A high-density ISFET array used in combination with live cell assay would be suitable for use in rapid and cost-effective identification of biomarkers and high throughput analysis of drug candidates.

2. Live Cell Assay of Membrane Proteins Using Enzymatic Signal Amplification

In a cell-based field-effect transistor (FET) such as an ISFET, cells are placed and cultured on the transistor’s gate surface. Figure 1 shows the conceptual structure of the cell-based FET. Because hydrogen ions (H⁺ ions) can be detected using the ISFET, molecular recognition reactions that involve the target membrane protein should be designed on the surface of the cell membrane to ensure that local pH change is induced near the gate surface as a result of molecular recognition. Metal oxides such as Ta₂O₅ or silicon nitrides such as Si₃N₄ are usually used as the pH-sensitive membrane in an ISFET. To enhance the adhesion of the cells and to maintain their activity on the surface of this pH-sensitive membrane, the surface of the pH sensitive membrane is sometimes modified using functional molecules, including fibronectin, poly-L-lysine, RGD peptides, and extracellular matrix gels.

Figure 1.

Conceptual structure of cell-based field effect transistor.

The first example of live cell assay of membrane proteins using an ISFET was the detection of cancer biomarkers to understand the heterogeneous nature of cancer cells and to elucidate their metastasis mechanisms. Recently, the liquid biopsy technique has been recognized as a minimally invasive method and is now frequently embedded in the design of clinical trials. Circulating tumor cells (CTCs)¹⁷ in the blood are one of the main detection targets in liquid biopsy. For CTC detection, capture and enrichment of the CTCs are required because CTCs are rare in blood. The CELLSEARCH^® circulating tumor cell kit (Menarini Silicon Biosystems, USA), which has been approved by the United States Food and Drug Administration, is currently used to detect the epithelial cell adhesion molecule (EpCAM) and cytokeratin (CK)-positive CTCs based on the semi-automatic fluorescence detection method.^18,19 When circulating in the bloodstream, the membrane proteins of CTCs may change throughout tumor progression. The CTCs undergo an epithelial–mesenchymal transition (EMT) and as a result, the expression of epithelial markers such as EpCAM decreases, whereas the expression of mesenchymal markers such as vimentin increases. Although the CELLSEARCH^® circulating tumor cell kit can be applied to analysis of metastatic breast, prostate, and colorectal cancers using patient blood samples, epithelial markers, including EpCAM, are downregulated during EMT, which then limits the capture and detection of the epithelial marker-negative CTCs.

To overcome this problem, use of cell-based FETs in combination with enzymatic signal amplification has been proposed for detection of human epidermal growth factor receptors (ErbB2, HER2) as a model membrane protein.²⁰ A breast cancer cell line was used as a model for the CTCs. It is known that the expression of HER2 is higher on the membranes of breast cancer cells than on those of healthy cells.^21,22 Overexpression of HER2 is also related to malignancy and to a poor prognosis in breast cancer. The scheme for detection of HER2 expression on live cancer cells is shown in Fig. 2. The cells are captured on the gate surface of an ISFET. An anti-HER2 antibody is bound to HER2 on the cell membrane and a secondary antibody labeled with the enzyme urease is introduced for binding to the anti-HER2 antibody. The HER2 expression on the cell membrane is then detected by introducing urea into the ISFET sensing area for an enzymatic reaction with urease that will induce a local pH change. The urease enzyme catalyzes hydrolysis of urea according to the reaction shown in Eq. 1.

  

\begin{equation} \begin{array}{ccccccccc} \text{Urea} & & & & & \text{Urease} & & & \\ \text{(NH$_{2}$)$_{2}$CO} & + & \text{2H$_{2}$O} & + & \text{H$^{+}$} &\to & \text{2NH$_{4}{}^{+}$} & + & \text{HCO$_{3}{}^{-}$} \end{array} \end{equation} (1)

A proton is consumed based on this reaction and the local pH near the cell membrane changes toward the alkaline direction. The enzymatic reaction rate is dependent on the number of urease enzymes that are bound to the cell membrane, which corresponds to the number of HER2 receptors at the cell surface. Therefore, the rate of the pH change reflects the level of HER2 expression under a constant substrate concentration. The pH change produced during the enzymatic reaction is known to be suppressed by the buffer capacity. The effect of the buffer capacity on the responses of enzyme-coupled ISFETs have been well studied in the literature,²³ and diluted buffer solutions are usually used.

Figure 2.

Detection scheme for HER2 expression on live cancer cells. Reprinted from Chem. Commun., 2022, 58, 7368–7371.

The human breast cancer cell line BT474 was used to demonstrate the operational principle of this method based on the detection scheme shown in Fig. 2. Expression of HER2 on BT474 cells was confirmed using fluorescein isothiocyanate (FITC)-conjugated antibodies and fluorescence observation via confocal microscopy, as shown in Fig. 3. The green color indicated fluorescence from FITC, which was bound to HER2, and the blue color indicated fluorescence from 4′,6-diamidino-2-phenylindole (DAPI), with which the DNA in the nucleus was stained. As shown in Fig. 3b, HER2 was confirmed to be expressed at the surfaces of the BT474 cells. The surface of the Ta₂O₅ pH-sensitive layer was modified with a positively charged polymer such as poly-L-lysine for efficient cell capture. The cells were cultured on the sensor surface in a confluent manner. The effects of both the poly-L-lysine coating and the cell culture on the ISFET’s pH sensitivity were confirmed to be negligible. The pH changes in the cell-based FETs were monitored when urea solutions were introduced to the gate surface. Figure 4a shows the time course of the responses of the cell-based FETs to the urea solution.²⁰ An ISFET without BT474 cells was used in a control experiment. The output signal of this ISFET without cells changed slightly with respect to the basic direction and reached a steady state immediately. In contrast, the pH values of the cell-based FETs shifted gradually in the basic direction as a response to the urea. These gradual pH changes reached the steady state in 10–15 min after the introduction of the urea solution. The pH changes were generated by the enzymatic reaction of the urease that was bound to the HER2 expressed on the cell membrane surface. The average pH change was 0.49 ± 0.02 in the cell-based FET, while that in the ISFET without cells was 0.23 + 0.03, as shown in Fig. 4b. Because the amount of urease that is bound to the cell surface reflects the expression level of HER2, the pH change obtained corresponds directly to the expression level of HER2 on the cell membrane.

Figure 3.

Immunofluorescence image of HER2 on BT474 cells. HER2 receptor: Green (FITC), Nucleus: Blue (DAPI). Reprinted from Chem. Commun., 2022, 58, 7368–7371.

Figure 4.

Response of the cell-based FETs to urea solutions. Reprinted from Chem. Commun., 2022, 58, 7368–7371.

The principle of the cell-based FET described above was further confirmed by quantitative detection of epidermal growth factor receptor (EGFR) (e.g., ErbB1, HER1) expression on live cancer cells. EGFR is a 170-kDa transmembrane glycoprotein with an intracellular tyrosine kinase domain, and overexpression of EGFR is associated with a factor of poor prognosis and more aggressive clinical progression.²⁴ Four breast cancer cell lines with different levels of EGFR expression were used to evaluate detection of the expression levels using the cell-based FETs in comparison with the results of flow cytometry with fluorescence detection.²⁵ The expression levels of EGFR on the membranes of the breast cancer cell lines, which comprised MM231, MM453, MM468, and BT474 cells, were analyzed using the fluorescence-activated cell sorting (FACS) method. Figures 5a and 5b show the FACS data and the histograms of the fluorescence intensity, respectively, for each cell type. The fluorescence intensity of MM468 was the highest among the four cell lines. The order of the EGFR expression levels was MM453 < BT474 < MM231 < MM468. ECM gel was used to capture breast cancer cells on the Ta₂O₅ gate surface of the ISFET. In this study, glucose oxidase (GOx) was used as the enzyme to detect EGFR expression on the cell membrane rather than urease. After the cells were captured on the pre-treated ISFET surface with the ECM gel, an anti-human EGFR antibody and a secondary antibody with GOx were subsequently introduced into the incubation chamber of the cell-based FET. When a glucose solution was then added to the surface of the cell-based FET, the local pH near the surface of the cells changed depending on the expression level of EGFR that resulted from a GOx-glucose reaction according to Eqs. 2 and 3.

  

\begin{equation} \text{C$_{6}$H$_{12}$O$_{6}$} + \text{O$_{2}$} + \text{H$_{2}$O}\xrightarrow{\text{GOx}}\text{C$_{6}$H$_{12}$O$_{7}$} + \text{H$_{2}$O$_{2}$} \end{equation} (2)

  

\begin{equation} \text{C$_{6}$H$_{12}$O$_{7}$} \to \text{C$_{6}$H$_{11}$O$_{7}{}^{-}$} + \text{H$^{+}$} \end{equation} (3)

The hydrogen ion is produced via the dissociation of the carboxyl groups on gluconic acid. Some of the hydrogen ions that are produced then diffuse to the surface of the pH-sensitive gate of the ISFET, while the others diffuse into the bulk of the solution.

Figure 5.

Immunofluorescence detection of EGFR expressed on four cell lines. Reprinted with permission from J. Am. Chem. Soc., 144, (2022) 36, 16545–16552. Copyright 2022 American Chemical Society.

A 10 mM glucose solution was then added to the reaction chamber of each cell-based FET and the changes in potential in response to the glucose–GOx enzymatic reaction were monitored. A comparison of the potential shifts among the four cell lines in the 10 min period after the addition of the glucose is shown in Fig. 6a.²⁵ The magnitude of the potential shift (ΔV) was greatest in order from MM468 to MM231, BT474, and MM453, which reflected the order of their higher expression levels of EGFR. This tendency was consistent with the results of the EGFR expression level analysis with FACS, as shown in Fig. 5. A comparison of the relative cell densities among the four cell lines is shown in Fig. 6b. No significant differences were observed in the relative cell densities before and after the enzymatic reaction, which indicated that the cell detachments were negligible during the measurements. In addition, the variation in the relative cell densities among the four cell lines after these measurements was 9.5 %. Therefore, the measured potential changes recorded using the different cell-based FETs in Fig. 6a could be attributed to differences in the EGFR expression levels among the four cell lines within an error range of 9.5 %. Figure 7 shows a comparison of the EGFR expression levels that were measured using both the ISFET method and the FACS method. A good correlation can be observed between the pH values measured with the cell-based FETs and the fluorescence intensities measured using the FACS, with a correlation coefficient of 0.976. Both pH values and fluorescence intensities reflected the EGFR expression levels on the cell membranes, the pH changes were not as big as the changes in fluorescence intensities. This is because only some of the generated hydrogen ions diffuse to the surface of the gate and contribute to the signal generation of the ISFET, while the pH values were suppressed by the buffer capacity to some extent. Although the kinetics of the responses of the cell-based FETs need to be analyzed in more detail, the expression levels of EGFR on the cell membranes could be detected successfully.

Figure 6.

Detection of EGFR with the cell-based FETs in combination with enzymatic chemical signal amplification. Reprinted with permission from J. Am. Chem. Soc., 144, (2022) 36, 16545–16552. Copyright 2022 American Chemical Society.

Figure 7.

Correlation between fluorescence intensity and potential change (n = 3). Reprinted with permission from J. Am. Chem. Soc., 144, (2022) 36, 16545–16552. Copyright 2022 American Chemical Society.

The types of enzymes that can be used in cell-based FETs are those that induce pH changes as a result of a reaction with their specific substrates, and they include GOx, urease, penicillinase, and acetylcholine esterase. If these enzymes are conjugated with antibodies that recognize epithelial antigens, mesenchymal antigens, and CD45 leukocyte markers, it would then be possible to detect all the surface antigens above simultaneously by sequential addition of their respective substrates; it would also be possible to analyze the heterogeneity of the cells when the cells are captured on the surfaces of the ECM gel-coated gates of the ISFETs. Because the sizes of these cells are of the order of a few tens of micrometers, a single cell can be placed on the surface of a single gate of a FET. Selective capture of the cells could then be realized based on hydrophobic/hydrophilic control of the gate surface.^26,27 In this case, variations in the cell density on the sensing areas among the different FETs will not present a problem. Use of single cell-based FETs would make it possible to enumerate the number of cells that show higher expression of specific antigens on their cell membranes.

3. Real-time Monitoring of Transporter-substrate Interactions at the Cell Membrane

The pharmacological action of a drug is determined by the concentration of the drug at the site of action and is also highly dependent on the pharmacokinetics of the drug, which comprise absorption, distribution, metabolism, and excretion. Therefore, because absorption, distribution, metabolism, and excretion all involve cell membrane permeation processes, transporters, which are responsible for the drug transport, play important roles in determining the efficacy and the side effects of drugs. The patch clamp technique has been used in studies of ion channel proteins. A fine glass pipette tip must be pressed carefully against the cell membrane and a tiny area of the cell membrane is then sucked into it. High levels of skill are required to achieve good contact between the cell membrane and the pipette tip, and to measure the ionic currents that result from opening of the ion channels without any leakage. Xenopus laevis oocytes express their membrane-bound transporters efficiently and can be used as a convenient model system in pharmaceutical lead discovery²⁸ to study the functions of membrane proteins such as transporters, ion channels, and ion pumps. To measure the uptake of substrates when mediated by transporters, radioisotope (RI)-labeled compounds are generally used to perform fundamental studies of properties such as substrate selectivity and uptake rates. In this method, the oocytes must be solved and fractured before detection of the specific radioactivity.

Non-invasive monitoring of the interactions between membrane proteins/transporters and ligands at cell membranes using a cell-based FET with an oocyte has also been reported.²⁹ In this case, the inflow and outflow of ions associated with molecular recognition of the membrane proteins was measured using the ISFET in real time. Xenopus laevis oocytes were injected with the capped human organic anion transporting peptide C (hOATP-C) cRNA. Estrone-3-sulfate (E3S) was used as a substrate for hOATP-C during the uptake measurements. As illustrated in Fig. 8a, when E3S is introduced into the reaction chamber of the cell-based FET with the hOATP-C expressed oocyte, E3S is then taken into the oocyte via hOATP-C. In contrast, E3S uptake did not occur in the case of a control oocyte on which hOATP-C is not expressed. The uptake experiments above were performed on the surface of the ISFET, as illustrated in Fig. 8b. Figure 9a shows the uptake results of the control experiments, where E3S was labeled with the radioisotope tritium [³H]. The intensity of the radioactivity increased over 2 h for the oocyte with expression of hOATP-C because of the E3S uptake, whereas it did not increase in the case of the control oocyte without hOATP-C expression. The output signals from the cell-based FETs were monitored when E3S was added to the reaction chambers of the FETs, as shown in Fig. 9b. The cell-based FET with the control oocyte and an ISFET without an oocyte were also evaluated in addition to the cell-based FET with the hOATP-C expressed oocyte. Only the output signal from the cell-based FET with the hOATP-C expressed oocyte increased dramatically, reaching a steady state in approximately 30 min after the introduction of E3S. The increase in the output signal from the cell-based FET with the hOATP-C expressed oocyte was based on the efflux of some ionic species that was associated with the E3S uptake. The constant flux of charged species continued during uptake of the E3S, which resulted in a steady-state output signal from the cell-based FET with hOATP-C expression.

Figure 8.

Detection scheme for transporter-substrate interaction on ISFET.

Figure 9.

Detection of substrate uptake into oocyte. Reprinted with permission from Anal. Chem. 2008, 80, 5, 1493–1496. Copyright 2008 American Chemical Society

It is known that there are several genotypes for hOATP-C and that these genotypes have differing transporting abilities. The mutant-type hOATP-C*15 was expressed in an oocyte and its transporting ability was compared with that of the wild-type hOATP-C using the common estradiol 17β-d-glucuronide (E₂17βG) substrate. The uptake signals obtained from the radioisotope method and the cell-based FET method are shown in Figs. 10a and 10b, respectively.²⁹ In the control experiment, the transporting ability of the mutant-type transporter was shown to be approximately half that of the wild type, as shown in Fig. 10a. In the case of the cell-based FET method, the time course data of the mutant-type transporter also showed approximately half the magnitude of the wild type. Therefore, the transporting kinetics of the substrate when mediated by the wild-type and the mutant-type transporters could be distinguished using the cell-based FETs.

Figure 10.

Comparison of transporting ability between the wild type and the mutant type. Reprinted with permission from Anal. Chem. 2008, 80, 5, 1493–1496. Copyright 2008 American Chemical Society

The signal generation mechanism of the cell-based FET was also investigated using various combinations of transporters and substrates.³⁰ The responses of the cell-based FETs with the different transporters are summarized in Fig. 11. The first example on the upper left shows the response of the cell-based FET with a proton-driven amino acid transporter to proline. In this case, the proline and the proton are transported from the exterior to the interior of the oocyte by a ratio of 1 to 1. The local proton concentration near the cell surface was reduced based on this transportation process. Therefore, the signal from the cell-based transistor decreased, as shown in the upper left figure. In contrast, when the proton-driven amino acid transporter was not expressed on the cell membrane, the signal from the cell-based transistor remained unchanged, as shown in the upper right figure. The second example is the case with the sodium-coupled phosphate cotransporter, which mediates the transport of sodium ions and hydrogen phosphate ions from the exterior to the interior of the oocyte by a ratio of 3 to 1. As a result, the hydrogen phosphate ion concentration on the exterior of the cell decreases, and the proton concentration increases locally near the cell surface based on the following equilibrium reaction.

  

\begin{equation} \text{HPO$_{4}{}^{2-}$} + \text{H$^{+}$} \leftrightarrow \text{H$_{2}$PO$_{4}{}^{-}$} \end{equation} (4)

The signal from the cell-based transistor increased as a result of increase in the proton concentration near the cell. When the sodium-coupled phosphate cotransporter was not expressed on the cell membrane, the signal from the cell-based transistor remained unchanged. The third example is the case with the sodium-coupled phosphate cotransporter, which mediates the transport of sodium ions and dihydrogen phosphate ions from the exterior to the interior of the oocyte by a ratio of 2 to 1, as illustrated on the bottom left of Fig. 11. As a result of the reduction of the dihydrogen phosphate ion concentration on the exterior of the cell, the proton concentration decreases locally near the cell surface because the following equilibrium relation shifts toward the right side.

  

\begin{equation} \text{HPO$_{4}{}^{2-}$} + \text{H$^{+}$} \leftrightarrow \text{H$_{2}$PO$_{4}{}^{-}$} \end{equation} (5)

The final examples are the cases with the γ-aminobutyric acid (GABA) transporter and the epithelial sodium channel. The GABA transporter mediates the transport of sodium ions, chloride ions, and GABA by a ratio of 2 to 1 to 1, and the epithelial sodium channel mediates sodium ion transport from the exterior to the interior of the oocyte. In both cases, protons are not involved in the transporting species, and in these cases, the signals from the cell-based transistors were not obtained, as shown in the bottom right figures in Fig. 11. From the results above, it was found that the signal generation mechanism of the cell-based transistor could be explained by direct or indirect proton transport via the transporters, and that proton transport should be involved in the transporting events to generate signals from a cell-based transistor using an oocyte with transporter expression. Figure 12 shows a lateral proton diffusion model at the oocyte-ISFET interface.³⁰ The vitelline layer of the defolliculated oocyte comprises a network of fibrous filaments that surround the cytoplasmic membrane. The total thickness of the vitelline layer is estimated to be 2–3 µm for the oocytes used in the experiment. It is known that the bulk diffusion rate of protons in water is extremely fast and that the pH values in the cell membrane are different from those in the bulk solution due to the change in the local association/dissociation rate constants that results in lateral diffusion rates different from the bulk diffusion rate. Once the system is in equilibrium, the proton concentration at the detection site [H⁺]_D is equal to the proton concentration at the membrane surface [H⁺]_S, assuming ideal coupling between the sensor surface and the membrane surface. The surface potential of the ISFET then reflects the surface pH for steady state membrane transport.

Figure 11.

The responses of the cell-based FETs with different transporters. Reprinted from Schaffhauser DF, et al., PLoS ONE 7(7): e39238.

Figure 12.

Lateral proton diffusion model at the cell-ISFET interface. [S]: Substrate concentration. [H⁺]_D, [H⁺]_S, [H⁺]_B: Proton concentrations at the detection site, membrane surface and in buffer, respectively. Reprinted from Schaffhauser DF, et al., PLoS ONE 7(7): e39238.

In addition to Xenopus laevis oocytes, mammalian cells have also been used to analyze membrane proteins using an ISFET.³¹ A sodium–proton exchanger (Na⁺/H⁺ exchanger; NHE), a membrane protein that transported Na⁺ ions into the cell and H⁺ ions out of the cell, was used in combination with three cell types: Chinese hamster ovary (CHO) cells, NHE3-deficient mouse skin fibroblasts (MSFs) (MSF NHE3−/−), and NHE3-expressed MSFs. The NHE was activated using an inward transmembrane sodium gradient as the energy source and was inhibited by the amiloride derivative 5-(N-ethyl-N-isopropyl)amiloride (EIPA). A continuous superfusion system was developed to realize rapid changes in the pH at the cell surface through alternate loading and unloading of ammonium ions. The ammonium ions are in equilibrium with the dissolved ammonia gas in an aqueous solution, as shown in Eq. 6.

  

\begin{equation} \text{H$_{3}$O$^{+}$} + \text{NH$_{3}$} \leftrightarrow \text{H$_{2}$O} + \text{NH$_{4}{}^{+}$} \end{equation} (6)

Ammonia gas can diffuse through the cell membrane but ammonium ions cannot penetrate the cell membrane, which results in a shift in the equilibrium in Eq. 6 that leaves protons on the outside of the cell. The typical output signal from the cell-based FET in response to repeated addition/withdrawal of NH₄Cl is shown in Fig. 13. The point designated ① in the signal indicated that the interior of the cell became basic because of the exposure to NH₄⁺ and the diffusion of NH₃ into the cell, while the pH on the exterior of the cell became acidic. In the case of point ②, the ammonia gas diffused back from the interior to the exterior of the cell because of NH₄⁺ being flushed away by the reflux system. As a result, the pH in the interior of the cell became acidic and the H⁺ stored within the cell was released simultaneously through NHE. After the inhibitor EIPA was applied to the cell as indicated at point ③, proton release mediated by NHE was inhibited, and the extracellular pH then became higher than the case in point ②. The difference in the potential values with and without EIPA could be used to analyze the interaction between NHE and EIPA. The NHE3 expressed MSF showed potential behavior similar to that of the CHO cells, and as expected, no significant potential change was observed in MSF NHE3−/−. Further investigation will be required to analyze the potential behavior of the cell-based FETs by considering the exchange transport processes of NH₄⁺ and Na⁺ via NHE.

Figure 13.

Typical output signal of the cell-based FET in response to repeated addition/withdrawal of NH₄Cl. Reprinted from Biosensors 2016, 6, 11.

This methodology was applied to assessment of the integrity of cell membranes.^32–34 The membrane integrity of live cells is evaluated routinely for cytotoxicity induced by either chemical or physical stimuli. HepG2 cells were cultured on a gate insulator of an ISFET. When these cells were subject to a membrane-toxic reagent, irreversible attenuation of the pH transient was then observed. The attenuated pH signal was correlated with the degree of hemolysis produced by the model reagents. The pH assay was sensitive to formation of molecular-sized pores that were otherwise not measurable via detection of the hemoglobin leakage because the H⁺ and NH₄⁺ ions could be used as indicators in the ISFET assay. Based on this technique, assays for tight junction integrity were developed by detecting the proton leaks in the cell gaps.^35,36 Furthermore, the translocation mechanisms of specific nanomaterials through cell membranes were analyzed using cultured cells on the ISFET. A water-soluble nanoaggregate that comprised an amphiphilic random copolymer of 2-methacryloyloxyethyl phosphorylcholine (MPC) and n-butyl methacrylate (BMA) was shown to pass through live cell membranes in an endocytosis-free manner.³⁷ Additionally, minor differences in the monomer structures and the functional groups of a series of zwitterionic poly(methacrylamide) and poly(methacrylate) species had dramatic effects on their interactions with plasma membranes during translocation.³⁸ The cell-penetrating polysulfobetaines had limited or zero detrimental effects on cell proliferation.

4. Conclusion and Future Perspectives

Various types of solid-state biosensors have been developed using ISFETs in combination with specific molecular recognition for simple, rapid, inexpensive, and multiplexed analysis of biomarkers for point-of-care testing and drug screening. In this review, two types of signal transduction schemes for detecting membrane protein functions were described as recent advances in cell-based FETs. In both schemes, local changes in proton concentration and their diffusion to the gate surface played a crucial role in signal generation of the cell-based FETs. Although fundamental principles of operation for both of the cell-based FETs were confirmed, further investigation will be necessary for practical use in clinical testing and drug screening. In the case of biomarker analyses on cancer cells, simultaneous detections of both epithelial and mesenchymal markers would be required using a single cell in order to apply the cell-based FETs to CTC testing. Recently, extracellular vesicles (EVs) have attracted considerable attention as potential biomarkers for use in clinical diagnosis. Among several types of EVs, micro vesicles (MVs), which have sizes that range from 100 nm to 1 µm, are generated from the plasma membrane of the host cells. Tumor-derived EVs express tumor-related proteins in the membrane that can be used for either cancer disease diagnosis or progress monitoring. EVs are contained in bodily fluids such as urine and saliva, and it is thus possible to collect EV samples in non-invasive and minimally invasive manners. The cell-based FET can be applied to the detection of surface antigens on the membranes of EVs. Although the size of an EV is approximately two orders of magnitude smaller than that of a cell, it is technically possible to fabricate the gate of a FET with a size that is comparable to that of an EV. Therefore, such an EV-based FET would be suitable for rapid, cost-effective, and noninvasive identification of biomarkers at various points in time during a disease’s course in future liquid biopsy applications.

In the case of monitoring transporter-substrate interaction, transporting species other than proton should be investigated to apply the cell-based FET to wider variety of transporters. Ionic species such as sodium, potassium, calcium, and chloride ions can be detected by depositing respective ion-selective membrane on the surface of ISFET. It is also preferable to use mammalian cells instead of Xenopus laevis oocytes. In research fields such as drug discovery, medical care, and cosmetics, it is necessary to screen for highly active compounds from vast numbers of candidate substances, ranging from natural compound libraries to newly synthesized compounds. Various toxicity tests and pharmacokinetic tests are generally performed in vivo using mammals to evaluate the efficacy of these substances and determine the appropriate dosage. However, recently, from an animal welfare viewpoint, various alternatives to animal experiments have been proposed, and analysis techniques using cultured cells are expected to become mainstream methods. By taking advantage of the potential for large-scale integration of multiple sensing areas, the cell-based FET will contribute greatly to drug discovery in the future.

CRediT Authorship Contribution Statement

Miyuki Tabata: Conceptualization (Equal), Writing – original draft (Equal), Writing – review & editing (Equal)

Yuji Miyahara: Conceptualization (Equal), Writing – original draft (Equal), Writing – review & editing (Equal)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Footnotes

M. Tabata and Y. Miyahara: ECSJ Active Members

References

• 1)  P. Bergveld, IEEE Trans. Biomed. Eng., BME-17, 70 (1970).
• 2)  M. Tabata and Y. Miyahara, Sens. Actuators B, 352, 131033 (2022).
• 3)  S. Caras and J. Janata, Anal. Chem., 52, 1935 (1980).
• 4)  Y. Miyahara, T. Moriizumi, H. Matsuoka, I. Karube, and S. Suzuki, J. Chem. Soc. Jpn., 1983, 823 (1983).
• 5)  Y. Hanazato and S. Shiono, Proc. Int. Meeting Chem. Sensors (Fukuoka, 1983), Kodansha Press, Tokyo, p. 513 (1983).
• 6)  P. Fromherz, A. Offenhäusser, T. Vetter, and J. Weis, Science, 252, 1290 (1991).
• 7)  M. Eickenscheidt, M. Jenkner, R. Thewes, P. Fromherz, and G. Zeck, J. Neurophysiol., 107, 2742 (2012).
• 8)  S. Ronchi, M. Fiscella, C. Marchetti, V. Viswam, J. Müller, U. Frey, and A. Hierlemann, Front. Neurosci., 13, 208 (2019).
• 9)  M. Lehmann, W. Baumann, M. Brischwein, H.-J. Gahle, I. Freund, R. Ehret, S. Drechsler, H. Palzer, M. Kleintges, U. Sieben, and B. Wolf, Biosens. Bioelectron., 16, 195 (2001).
• 10)  L. Campanella, G. Favero, D. Mastrofini, and M. Tomassetti, Sens. Actuators B, 44, 279 (1997).
• 11)  F. Bettaieb, L. Ponsonnet, P. Lejeune, H. B. Ouada, C. Martelet, A. Bakhrouf, N. Jaffrézic-Renault, and A. Othmane, Bioelectrochemistry, 71, 118 (2007).
• 12)  M. Lehmann, W. Baumann, M. Brischwein, R. Ehret, M. Kraus, A. Schwinde, M. Bitzenhofer, I. Freund, and B. Wolf, Biosens. Bioelectron., 15, 117 (2000).
• 13)  H. Satake and T. Sakata, Anal. Chem., 91, 16017 (2019).
• 14)  T. Sakata, I. Makino, S. Kita, and Y. Miyahara, Microelectron. Eng., 85, 1337 (2008).
• 15)  M. A. Lemmon and J. Schlessinger, Cell, 141, 1117 (2010).
• 16)  E. Zwick, J. Bange, and A. Ullrich, Endocr. Relat. Cancer, 8, 161 (2001).
• 17)  L. Keller and K. Pantel, Nat. Rev. Cancer, 19, 553 (2019).
• 18)  F.-C. Bidard, D. J. Peeters, T. Fehm, F. Nolé, R. Gisbert-Criado, D. Mavroudis, S. Grisanti, D. Generali, J. A. Garcia-Saenz, J. Stebbing, C. Caldas, P. Gazzaniga, L. Manso, R. Zamarchi, A. F. de Lascoiti, L. De Mattos-Arruda, M. Ignatiadis, R. Lebofsky, S. J. van Laere, F. Meier-Stiegen, M.-T. Sandri, J. Vidal-Martinez, E. Politaki, F. Consoli, A. Bottini, E. Diaz-Rubio, J. Krell, S.-J. Dawson, C. Raimondi, A. Rutten, W. Janni, E. Munzone, V. Carañana, S. Agelaki, C. Almici, L. Dirix, E.-F. Solomayer, L. Zorzino, H. Johannes, J. S. Reis-Filho, K. Pantel, J.-Y. Pierga, and S. Michiels, Lancet Oncol., 15, 406 (2014).
• 19)  B. Hu, X.-R. Yang, Y. Xu, Y.-F. Sun, C. Sun, W. Guo, X. Zhang, W.-M. Wang, S.-J. Qiu, J. Zhou, and J. Fan, Clin. Cancer Res., 20, 6212 (2014).
• 20)  M. Tabata, C. Khamhanglit, S. Kotaki, and Y. Miyahara, Chem. Commun., 58, 7368 (2022).
• 21)  M. Yan, M. Schwaederle, D. Arguello, S. Z. Millis, Z. Gatalica, and R. Kurzrock, Cancer Metastasis Rev., 34, 157 (2015).
• 22)  S. Pernas and S. M. Tolaney, Ther. Adv. Med. Oncol., 11, 1 (2019).
• 23)  S. Caras and J. Janata, Anal. Chem., 52, 1935 (1980).
• 24)  C. R. Chong and P. A. Janne, Nat. Med., 19, 1389 (2013).
• 25)  M. Tabata, X. Liu, C. Khamhanglit, S. Kotaki, and Y. Miyahara, J. Am. Chem. Soc., 144, 16545 (2022).
• 26)  H. Otsuka, Molecules, 15, 5525 (2010).
• 27)  H. Otsuka, M. Nagamura, A. Kaneko, K. Kutsuzawa, T. Sakata, and Y. Miyahara, Sci. Technol. Adv. Mater., 13, 064217 (2012).
• 28)  T. Abe, M. Kakyo, T. Tokui, R. Nakagomi, T. Nishio, D. Nakai, H. Nomura, M. Unno, M. Suzuki, T. Naitoh, S. Matsuno, and H. Yawo, J. Biol. Chem., 274, 17159 (1999).
• 29)  T. Sakata and Y. Miyahara, Anal. Chem., 80, 1493 (2008).
• 30)  D. F. Schaffhauser, M. Patti, T. Goda, Y. Miyahara, I. C. Forster, and P. S. Dittrich, PLoS One, 7, e39238 (2012).
• 31)  D. Schaffhauser, M. Fine, M. Tabata, T. Goda, and Y. Miyahara, Biosensors, 6, 11 (2016).
• 32)  Y. Imaizumi, T. Goda, Y. Toya, A. Matsumoto, and Y. Miyahara, Sci. Technol. Adv. Mater., 17, 337 (2016).
• 33)  Y. Imaizumi, T. Goda, D. F. Schaffhauser, J. Okada, A. Matsumoto, and Y. Miyahara, Acta Biomater., 50, 502 (2017).
• 34)  Y. Imaizumi, T. Goda, A. Matsumoto, and Y. Miyahara, Analyst, 142, 3451 (2017).
• 35)  H. Hatano, T. Goda, A. Matsumoto, and Y. Miyahara, Anal. Chem., 91, 3525 (2019).
• 36)  H. Hatano, T. Goda, A. Matsumoto, and Y. Miyahara, ACS Sens., 4, 3195 (2019).
• 37)  T. Goda, Y. Imaizumi, H. Hatano, A. Matsumoto, K. Ishihara, and Y. Miyahara, Langmuir, 35, 8167 (2019).
• 38)  T. Goda, H. Hatano, M. Yamamoto, Y. Miyahara, and N. Morimoto, Langmuir, 36, 9977 (2020).

Biographies

Miyuki Tabata (Senior Assistant Professor, Graduate School of Bio-Applications and Systems Engineering Tokyo University of Agriculture and Technology (TUAT))

Miyuki Tabata received her PhD from the Graduate School of Pure and Applied Sciences, University of Tsukuba. She then began work at the Tokyo Medical and Dental University in the laboratory of Bioelectronics (Prof. Yuji Miyahara). She has been a PI at the Tokyo University of Agriculture and Technology since 2023. Her research focuses on developing electrical/electrochemical biosensing devices and their medical applications.

Yuji Miyahara (Specially Appointed Professor, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University)

Yuji Miyahara was born in Kanagawa Prefecture in 1956. He graduated from the School of Engineering, Tokyo Institute of Technology in 1980 and received his PhD from Tokyo Institute of Technology in 1985. He joined the Hitachi Central Research Laboratory in 1985. He was a visiting researcher at the Swiss Federal Institute of Technology from 1988 to 1989. He became the group leader of the Bioelectronics group and a director of the Biomaterials Center at the National Institute for Materials Science in 2002 and 2008, respectively. He became a professor at the Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University in 2010, and a director of the Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University in 2014. He has been a specially appointed professor at Tokyo Medical and Dental University since 2022. He has been working in the fields of electronics and biotechnology for more than 40 years. He has been studying bio-transistors for DNA analyses, micro total analysis systems (µTAS), and solid-state biosensors.