2024 Volume 92 Issue 2 Pages 022009
Cell adhesion to culture substrates plays a crucial role in cellular activities, such as proliferation, differentiation, and apoptosis. Recently, electrochemiluminescence (ECL) imaging has been utilized for analyzing cell adhesion. In addition to ECL imaging, electrochemical impedance spectroscopy (EIS) is widely used as a conventional method for evaluating cell adhesion. However, the relationship between the results obtained from ECL imaging and EIS has not yet been investigated. In this study, the relationship between the results obtained using two methods is explored, and their advantages and disadvantages have been discussed. Endothelial cells were cultured on an extracellular matrix-coated indium tin oxide electrode for 24 h, which was further used for ECL imaging and EIS. To the best of our knowledge, this is the first report of ECL imaging and EIS of the same samples for cell adhesion analysis. This comprehensive analysis of cell adhesion using EIS and ECL imaging could be further used to evaluate drugs that target cell adhesion.
Cell-substrate interactions play a critical role in biological processes such as cell proliferation, cell differentiation, cell migration, and apoptosis.1,2 Particularly, the endothelial and epithelial cells have been widely investigated because they adhere to the extracellular matrix (ECM) and form a cell monolayer, which is important for the functioning of the cellular barrier and the permeability of drugs and chemicals. Therefore, it is important to evaluate these interactions for investigating biological phenomena and successful drug development. Fluorescence imaging is a powerful tool for detecting cell adhesion molecules such as integrins.3 However, it requires labeling of target analytes with fluorescence tags, which is time-consuming and damages the cells.
Electrochemical techniques have been proposed as excellent alternatives to fluorescence imaging techniques because of their high sensitivity, simplicity of equipment, and low cost. For instance, redox currents have been used to analyze cell adhesion. In this strategy, cells adhere to an electrode, and redox species are added to the culture medium. When the cells cover the electrode, the active area of the electrode is reduced owing to cellular insulation, and the resulting redox currents are suppressed.4,5 As the suppression is related to the cell area, the redox currents can be effectively used as an indicator of cell adhesion. Electrochemical impedance spectroscopy (EIS) is another electrochemical approach that is used to evaluate cell-substrate interactions by measuring the impedance.6–9 In EIS, frequency signals (potential or currents) in the range of approximately 0.1–1 × 106 Hz are applied to measure the impedance of the electrode interface and cellular gaps for cell analysis. The approach is advantageous as it applies small amplitude perturbations from the steady state during measurements, resulting in low cell toxicity. The cell adhesion on the electrode affects the impedance, therefore, it can be used as an indicator of cell adhesion. These strategies are useful for investigating cell-substrate interactions; however, they are unsuitable for obtaining spatial information on cell-substrate interactions owing to the utilization of a single working electrode. Although electrode array devices containing many individual electrodes have been reported to address this issue,4,5 their spatial resolution is low because of the limited number of electrodes compared to that of fluorescence imaging. Therefore, the electrode array devices are unsuitable for determining cellular junctions.
Electrochemiluminescence (ECL) devices have garnered considerable attention for the electrochemical imaging of bioassays,10,11 including cell analysis.12–14 ECL is chemiluminescence induced by electrochemical reactions.15,16 Compared to the background signals of fluorescence assays, those of ECL are low because ECL does not require an external source of light. Recently, ECL imaging has been used to visualize cellular adhesion.17–24 Briefly, cell adhesion on an electrode blocks the ECL reactions, and no light emission occurs from the cell areas. However, the gaps between the cells exhibit large emissions because ECL chemicals diffuse from the bulk to the electrode through the gap. Thus, cellular junctions are visualized without fluorescent labeling.
Therefore, EIS and ECL imaging have been used to evaluate cell adhesion. However, a comprehensive analysis and a relationship between the results obtained using these techniques has not been carried out. In this study, ECL and EIS studies were conducted on the same samples to analyze the cell-substrate adhesion (Fig. 1). Further, the relationship between the results obtained using ECL and EIS was discussed, representing a novel aspect in EIS and ECS studies. In this study, the term “cell adhesion” refers to cell-substrate interactions.
Schematic illustration of (A) ECL imaging and (B) EIS for cellular adhesion analysis.
The device fabrication and detection processes are illustrated in Fig. 2. An indium tin oxide (ITO) electrode (Sanyo Vacuum Industries Co., Ltd., Japan) was used as a working electrode (WE). As the ITO electrode is transparent, it was used for the microscopy of cells on the electrode. A polydimethylsiloxane (PDMS: Dow Corning Toray, Japan) framework with 2 mm thickness and 4 mm hole diameter was used as the 1st layer and bonded to the ITO to define the electrode area. Subsequently, it was bonded to a second PDMS layer (thickness: 3 mm) with a hole (diameter: 10 mm), which acted as a reservoir for the culture medium and measurement solutions. An extracellular matrix (ECM) was coated onto the ITO electrode to induce cell adhesion, wherein Matrigel (product #354234, Corning, NY, USA) was used as the ECM and mixed with a culture medium followed by incubation for 1 h. Thereafter, the solution was entirely removed and washed with Dulbecco’s phosphate-buffered saline (PBS; pH 7.4; Nacalai Tesque, Japan). Subsequently, human umbilical vein endothelial cells (HUVECs: Lonza, Switzerland) were seeded in 0, 2.0 × 104, 4.0 × 104, and 8.0 × 104 cells cm−2 cell densities on the ITO electrode and cultured in endothelial cell growth medium 2 (Promo Cell, Germany) containing 1 % penicillin/streptomycin (Gibco, USA) in a humidified incubator at 37 °C with 5 % CO2 for 24 h.
Device fabrication process and detection scheme. (i) ITO electrode. (ii) Bonding of the 1st and 2nd PDMS layers. (iii) Coating Matrigel. (iv) HUVEC culture. (v) EIS measurement after bonding the 3rd and 4th PDMS layers and setting CE and RE. (vi) ECL imaging after removing the 3rd and 4th PDMS layers.
After 24 h, the culture medium was changed to a PBS containing 2 mM [Fe(CN)6]3−/4− (Kanto Chemical Co., Ltd., Japan) and 0.1 M KCl for EIS measurements. Subsequently, it was bonded to 3rd and 4th PDMS layers with 1 mm thickness and hole diameters of 3 and 5 mm, respectively. A Pt wire was inserted into these layers as a counter electrode (CE). The holes in the 3rd and 4th layers were used to connect the Ag/AgCl (sat. KCl) electrode (BAS Inc., Japan) as a reference electrode (RE). These electrodes were further connected to a potentiostat (Compactstat, Ivium Technologies, Netherlands), and EIS was performed at frequencies between 1 MHz and 0.1 Hz with an amplitude of 10 mV under an open circuit potential. A software (IviumSoft, Ivium Technologies) was used for EIS fitting.
ECL imaging was conducted after the EIS measurements. For imaging, the 3rd and 4th PDMS layers were removed, and the cells were fixed using 4 % paraformaldehyde (FUJIFILM Wako Pure Chemical Corporation, Japan) for 15 min followed by washing with PBS twice. After washing, PBS solution containing 1 mM [Ru(bpy)3]2+ (Tokyo Chemical Industry Co., Ltd., Japan) and 50 mM tripropylamine (TPA) (FUJIFILM Wako Pure Chemical Corporation, Japan) was added to the device. The CE and RE were inserted into the solution, and three electrodes were connected to a potentiostat (HA1010mSM8, Hokuto Denko, Japan). Phase-contrast images were captured using an inverted microscope (Olympus, Tokyo, Japan) equipped with a CCD camera (DP71, Olympus). The potential was then stepped from 0 to 1.2 V, and ECL images were captured using the same microscope in a dark room with an exposure time of 10 s following the potential step. ECL images were then analyzed using ImageJ software or a LabVIEW program based on NI Vision (National Instruments, USA). Briefly, the ECL images were binarized by manually selecting an appropriate threshold. The island areas of the cells were automatically obtained using the program. The cell islands were shrunk in the ECL images to remove debris and distinguish the connected cells, and then, the cell islands were counted.
Cyclic voltammetry was conducted using the ITO device filled with PBS containing 1 mM [Ru(bpy)3]2+ and/or 50 mM TPA at a scan rate of 50 mV s−1 to analyze an ECL mechanism.
Figure 3 shows images of the device. The ITO electrode was covered with the 1st PDMS layer to define the active area of the electrode (12.6 mm2), and the 2nd PDMS layer was used as the reservoir for a 0.2 mL solution (Fig. 3A). For high repeatability in EIS, the positions of the CE and RE were fixed by inserting them into the 3rd and 4th PDMS layers (Figs. 3B, 3C) whereas for cyclic voltammetry and ECL imaging, the 3rd and 4th PDMS layers were removed.
Device images. (A–C) Device with the (A) 1st and 2nd PDMS layers, (B) 1st, 2nd, 3rd, 4th PDMS layers, and CE, and (C) 1st, 2nd, 3rd, 4th PDMS layers, CE, and RE. The brightness of the images was adjusted.
Cyclic voltammetry was performed in the ECL solution to investigate the ECL reaction (Fig. 4). When using the solution without ECL chemicals, no peak is observed between 0 and 1.2 V (Fig. 4A). The cyclic voltammogram in Fig. 4B shows the absence of a reduction peak and oxidation of TPA at 0.90 V approximately. The oxidation and reduction peaks of [Ru(bpy)3]2+ are observed at 1.12 and 1.04 V, respectively (Fig. 4C). In contrast, the oxidation peak current when detecting TPA and [Ru(bpy)3]2+ solution is more than twice greater than that using only [Ru(bpy)3]2+, and the reduction peak disappears (Fig. 4D). The oxidation peak at 1.13 V is similar to that obtained using only [Ru(bpy)3]2+. These results indicate that the ECL reaction proceeds via a catalytic route (Fig. 4E). A direct oxidation pathway is observed as well (Fig. 4F). Although three reaction pathways have been previously described in the literature,25 including the direct oxidation, low oxidation potential, and catalytic paths, the primary pathway followed in the present study could not be determined.
Cyclic voltammograms of 1 mM [Ru(bpy)3]2+ and/or 50 mM TPA in PBS. Scan rate: 50 mV s−1. (A) PBS (no ECL chemicals). (B) TPA. (C) [Ru(bpy)3]2+. (D) TPA and [Ru(bpy)3]2+. Presumed ECL reactions via (E) catalytic and (F) direction oxidation paths.
Figures 5A–5C shows phase-contrast and ECL images of HUVECs adhered to the ITO electrode coated with Matrigel (cell seeding number: 2.0 × 104, 4.0 × 104, and 8.0 × 104 cells cm−2, culture time: 24 h). The ECL images are black in the area wherein the cells adhered to the electrode because the ECL reactions are blocked when the cells cover the electrodes. The merged phase-contrast and ECL images reveal that the position of the edges of the individual cells correspond to that of the ECL emission areas, indicating that ECL chemicals diffuse via the cellular gaps from the bulk to the electrodes (Fig. 5D). The small shadow areas might indicate local cell adhesion of cellular membranes. Surprisingly, even though the cells are confluent and cell density is high, many gaps are observed between them (Figs. 5A–5C). Additionally, even though the cell number increases, cell-cell interactions did not occur after the 24-h culture, which indicates that the 24-h culture was insufficient to induce cell-cell interactions and form tight junctions between the cells. However, we used this culture condition because it was easy to prepare the samples using 24 h of culture. Although we assumed that a high cell density could induce cell-cell interactions even if the culture period was short, unfortunately, the interaction was insufficiently induced. We understand the importance of analyzing cell-cell interactions; however, in the present study, we focused on the relationship between ECL imaging results and EIS results using the simple culture condition. In the present study, a thin Matrigel layer was coated, therefore, a thicker Matrigel culture might be suitable for inducing cell-cell interactions. Previously, we reported the ECL imaging of tube formation in HUVECs,17 and there was no gap between the cells during tube formation in the thick Matrigel (5 µL/cm2). Figure 5E shows the magnified ECL image of a single cell. Although no ECL signal is observed in the thick cellular nuclei area, the thin cellular peripheral areas exhibit high ECL signals. These results indicate that the diffusion thickness of [Ru(bpy)3]2+ is greater than that of the cellular peripheral area. Thus, ECL imaging provides information on cellular height. However, to visualize the cellular peripheral area in the ECL images, the thickness of the diffusion of [Ru(bpy)3]2+ should be shortened by adjusting the ratio of the concentrations of [Ru(bpy)3]2+ and TPA.26
ECL imaging of HUVECs. (A–C) Phase-contrast and ECL images of HUVECs at various cell concentrations. Cell seeding number: (A) 2 × 104, (B) 4 × 104, and (C) 8 × 104 cells cm−2. Culture period: 24 h. (D) Magnified images of the area with the high cell density (left: phase-contrast image, middle: ECL image, right: merged image). (E) Magnified image of a single cell. Yellow and white arrows indicate the thick and thin cell areas, respectively. (F) Number of cell islands vs. cell seeding number. The island number was counted in an ECL image (6.2 × 105 µm2). (G) ECL-suppressed area (%) vs. cell seeding number. (H) Cell area (µm2 cell−1) vs. cell seeding number. The error bars indicate the standard deviations. n = 3 or 4.
Next, the ECL images were analyzed to determine the area of cellular adhesion. The number of the islands of the black regions, indicating cell number, increases as the cell seeding number increases (Fig. 5F), which is reasonable. As shown in Fig. 5G, the ECL-suppressed area (%) increases as the cell seeding number increases. However, the value approaches saturation as the cell seed number increases, because the cells are confluent when using 4.0 × 104 cells cm−2 and the space for cell spreading is limited even though the cell seeding number is changed from 4.0 × 104 to 8.0 × 104 cells cm−2. Therefore, the adhesion areas of individual cells decrease owing to the limited cell culture areas when the cell seeding number increases. The cell area (µm2 cell−1) decreases (Fig. 5H) owing to contact inhibition, which is in accordance with the results of the phase-contrast images.
Nyquist plots were obtained using EIS of samples at different cell seeding numbers (0, 2.0 × 104, 4.0 × 104, and 8.0 × 104 cells cm−2) (Fig. 6A). Figure 6B shows the equivalent circuit model, where Rsol is the solution resistance, Rct is the charge-transfer resistance at the electrode, and CPE is the constant phase element representing the capacitance related to the electric double layer. Zw represents the impedance associated with diffusive ion transportation (Warburg impedance). This simple model is widely used for cell analysis,6 although several complex models have been proposed for electrical cell-substrate impedance sensing.27 Although it is necessary to discuss the ECM layer for the precise analysis of EIS, this simple model was used to avoid complex discussions in EIS. Additionally, the first small semicircles in the Nyquist plots were ignored in the analysis. The Rct increases as the cell seeding number increases (Fig. 6C) because the cells adhere to the electrode, which reduces the free electrode area and increases the interface impedance. This trend is well in agreement with the results for the ECL-suppressed areas in ECL imaging. Thus, EIS also provides information on the number of cells. The Rct is less than 400 Ω cm2 when the HUVECs of 8.0 × 104 cells cm−2 are cultured for 24 h. In contrast, Bouafsoun et al. reported an Rct value of 2 kΩ cm2 approximately when using a modified electrode covered with an endothelial cell layer without considerable defects,28 which indicates that the present endothelial cell layer did not form completely, and the culture period is insufficient. These observations were also supported by the ECL results, which revealed many gaps between the cells. In previous reports, experimental conditions (e.g., cell type, culture media, ECM type, electrode type, and device configuration) dramatically affect Rct, and the values range from 2–2 × 103 Ω cm2.28–31 Therefore, further comparison was not carried out in this study. Transepithelial/transendothelial electrical resistance (TEER) values are widely used as impedance indicators of cell barriers. An endothelial cell layer forms on a porous membrane, and the impedance is measured at a certain frequency. TEER values are approximately in the 10–20 Ω cm2 range when using a transwell seeded with HUVECs at near confluence and cultured for approximately 1–2 weeks.32–34 In contrast, since the present EIS measurements differ from TEER, the TEER values were not compared with the present Rct values.
EIS analysis. The cell seeding numbers were 0, 2 × 104, 4 × 104, and 8 × 104 cells cm−2, and the cells were cultured for 24 h. (A) Nyquist plots. (B) Equivalent circuit. (C) Rct vs. cell seeding number. The error bar indicates the standard deviations. n = 3 or 4.
Finally, the relationship between the ECL imaging and EIS results was investigated. As shown in Fig. 7, the Rct is approximately proportional to that of the ECL-suppressed area. However, the lack of a strong correlation indicates that the information obtained from ECL imaging differs from that obtained using EIS, even though the results are derived from cell adhesion. In the ECL imaging analysis, the images were binarized, and some of the information was lost, which could have resulted in the disagreement. In addition, part of the cellular peripheral area was not visualized in the ECL images, which may have caused a mismatch. The results of the samples without any cell-cell interactions are discussed in the present study, and in future studies, the cell-cell interactions will be investigated with a longer culture time, because these interactions also play important biological roles such as forming cellular barriers. Additionally, time-course analyses of longer cultures will be conducted.
Plot of ECL-suppressed area vs. Rct.
An advantage of ECL imaging over EIS is that ECL provides local area information. In addition, cellular height information could be obtained using ECL imaging by adjusting the experimental conditions. However, ECL chemicals may exhibit cellular toxicity and hence may be unsuitable for the analysis of living cells. In contrast, EIS can be used for real-time monitoring of living cells because it is more biocompatible than ECL imaging, which is highly advantageous. If porous membranes are used, cancer migration is observed in real time.35 EIS can in-situ monitor cell damages, such as electroporation.36 For the application of ECL imaging to living cells, the selection of suitable ECL chemicals and optimization of the measurement conditions is required. Recently, bipolar electrochemistry has been used for cell analyses,37,38 which included the evaluation of cell adhesion.39 Since this device comprised a closed-bipolar system, the measurement area was separated from the cell culture areas, indicating that ECL chemicals did not affect the cellular activities. In the future, bipolar electrochemistry would be used for the ECL analysis of cells.
This study presents an analysis of cell adhesion using ECL imaging and EIS of the same samples. In ECL imaging, the cells were imaged as black because they blocked the reactions at the surface of the electrode. Nyquist plots were obtained from EIS measurements, which revealed that Rct increased as the cell seeding number increased because the cells were insulation materials. The relationship between the ECL-suppressed area and the resistance of the surface electrode was approximately linear. The merits and disadvantages of these methods have been discussed in the main text. Although the combination of EIS and phase-contrast imaging for the analysis of cell adhesion has already been reported,7 to the best of our knowledge, this is the first report on the combination of ECL and EIS for cell analysis. These techniques could be utilized for screening drugs that target cell adhesion.
This work was supported by Grant-in-Aid for Scientific Research (A) (No. 20H00619), Grant-in-Aid for Scientific Research (B) (Nos. 21H01957 and 22H02102), and Grant-in-Aid for Challenging Research (Exploratory) (No. 23K17926) from the Japan Society for the Promotion of Science. This work was also supported by Amano Institute of Technology and the Japan Association for Chemical Innovation. This study was partially supported by AMED (Grant Number JP22be1004205).
Kimiharu Oba: Data curation (Lead), Formal analysis (Lead), Investigation (Lead), Writing – original draft (Equal), Writing – review & editing (Equal)
Kosuke Ino: Conceptualization (Lead), Formal analysis (Supporting), Funding acquisition (Lead), Investigation (Lead), Project administration (Lead), Software (Lead), Supervision (Lead), Writing – original draft (Lead), Writing – review & editing (Lead)
Yoshinobu Utagawa: Investigation (Supporting), Writing – original draft (Supporting), Writing – review & editing (Supporting)
Hiroya Abe: Investigation (Supporting), Supervision (Supporting), Writing – original draft (Supporting), Writing – review & editing (Supporting)
Hitoshi Shiku: Conceptualization (Supporting), Funding acquisition (Lead), Investigation (Supporting), Project administration (Lead), Supervision (Lead), Writing – original draft (Supporting), Writing – review & editing (Supporting)
The authors declare no conflict of interest in the manuscript.
Japan Society for the Promotion of Science: Nos. 20H00619, 21H01957, 22H02102, and 23K17926
Amano Institute of Technology and the Japan Association for Chemical Innovation
AMED: JP22be1004205
K. Ino, H. Abe, and H. Shiku: ECSJ Active Members
