Enhancement of TIRF Imaging of 3D-Cultured Spheroids via Hydrostatic Compression Using a Balloon Actuator

Abstract

Three-dimensional (3D) cultured cells can mimic the in vivo tumor microenvironment more accurately than conventional monolayer cultures. Therefore, they are essential in cancer research and drug discovery. However, high-sensitivity fluorescence imaging of 3D spheroids remains challenging owing to their limited contact with the observation surface and the low penetration depth of total internal reflection fluorescence microscopy (TIRFM). In this study, we developed a microfluidic device equipped with a water-driven balloon actuator that enables the hydrostatic compression of 3D-cultured spheroids. This system gently presses spheroids against a glass surface, significantly enhancing the contact area and improving TIRFM and epifluorescence imaging quality, with more evident improvement observed in TIRFM. Our results show that hydrostatic compression markedly enhances optical accessibility in spheroids while preserving cell viability and structural integrity. The method is designed to complement volumetric imaging techniques, including confocal and light-sheet microscopy, by enabling high-contrast visualization of cell–surface molecular dynamics. Although the current system focuses on surface accessibility, future studies will incorporate rotational mechanisms and automated pressure control to facilitate multi-angle, high-throughput imaging. This platform offers a promising strategy for the dynamic observation of cell–surface interactions in living 3D systems.

1. Introduction

Three-dimensional (3D) cultured cells more closely resemble the in vivo cellular environment than monolayer-cultured cells and are therefore being actively investigated [1,2,3,4,5,6,7]. Compared with monolayer cultures, 3D-cultured cells exhibit different responses to anticancer drugs [8,9] and distinct gene expression profiles [10]. They are also widely used for organ simulation [11] and regenerative medicine research [12,13]. Precision medicine relies on the binding of drugs to specific proteins on the cell surface to reduce side effects [14]. Consequently, studies of precision medicines require observation of living 3D-cultured cells rather than formalin-fixed or frozen sections. For live-cell investigations, fluorescence microscopy is commonly employed to detect proteins through the attachment of fluorescence-labeled antibodies [15,16]. However, owing to the inherent limitations of conventional microscopy for imaging 3D cultures, confocal microscopy is often used to visualize their fluorescence.

Confocal microscopy is based on the principle of a pinhole aperture [17]. It enables visualization of three-dimensional structures, such as distribution patterns on the cell surface [18,19]. However, its axial (z-direction) resolution of approximately 300 nm limits the detection of weak fluorescence signals [20,21]. Therefore, protein distribution on the cell surface that constitutes spheroids has not yet been investigated in detail. Clarifying differences in the distribution and behavior of surface molecules between monolayer- and 3D cultured cells will facilitate the integration of cell biology with tissue engineering [22,23,24,25]. Such advances will not only promote drug discovery but also support the development of new disease treatments. Overall, high-sensitivity fluorescence methods for observing 3D-cultured cells are necessary.

Although confocal and light-sheet microscopy enable volumetric imaging, they often lack the surface sensitivity required for observations at the plasma membrane [26,27,28]. One approach to achieving such high sensitivity is total internal reflection fluorescence microscopy (TIRFM). In this method, proteins on the cell surface are fluorescently labeled and detected with high sensitivity [21]. Evanescent field, which can penetrate only up to 200 nm beyond the total reflection surface, serves as a highly localized excitation source; it also enables a clear observation of the binding sites of fluorescently labeled molecules [29,30,31].

However, 3D cultured cells suspended in a medium can rarely establish stable contact with the glass interface required for TIRFM (Figure 1), and compressing samples between a slide and coverslip causes lysis due to atmospheric pressure loading. Therefore, a method is required to increase the glass contact area of intact spheroids within a closed microfluidic environment.

Although 3D-cultured cells can be manipulated using various techniques [32], such as micromanipulators [33,34], acoustic waves [35,36], and magnetic fields [37,38], these techniques require expensive equipment and cannot be applied in closed environments, such as within microfluidic devices. This study aims to overcome these technical challenges and facilitate the observation of 3D-cultured cells under evanescent illumination. To this end, we developed a device incorporating a hydraulic actuator that enables 3D-cultured cells to be pressed against a glass surface. A fabrication method for this device is described herein.

2. Materials and Methods

2.1. Device Design

The device must satisfy two requirements: (1) it must apply sufficient force to press spheroids against the glass surface without causing cellular damage, and (2) it must maintain spheroids within the microscope observation range. Microdevices for cell culture typically incorporate water-hydraulic or pneumatic, electromagnetic, or electrostatic actuators. Among these, water-hydraulic actuators are easily sterilized by autoclaving, allow device miniaturization, and do not interfere with bright-field observation using transmitted light. Additionally, the required peripheral equipment can be readily accommodated in a cleanroom environment. Therefore, the proposed device was designed to incorporate a water-hydraulic actuator, which was used to control the distance between 3D-cultured cells and the glass surface. To ensure that spheroids remained within the observation range, a chamber slightly larger than the spheroid diameter was fabricated to house the cells.

A conceptual diagram of the proposed device is shown in Figure 2. The device consists of a chamber for cell storage and a hydraulic actuator for cell compression. The chamber dimensions (L × W × H) are 500 × 500 × 316 µm. The chamber bottom is formed by a glass slide to enable observation under evanescent illumination. A surrounding space is provided to store culture medium around the cell containment area. The hydraulic actuator has a diameter of 4 mm, and its membrane is 40 μm thick. The overall device dimensions are shown in Figure S1.

2.2. Device Fabrication

Fabrication of the proposed microfluidic device involved two main steps: (1) fabrication of molds for the chamber and actuator using a 3D printer, and (2) casting and assembly of the chamber and actuator components. A schematic of the fabrication process for the microfluidic device components is shown in Figure S2.

2.2.1. Mold Fabrication Procedure

The molds were casted using a 3D printer (PartPro150xP, XYZ printing, Taipei, Taiwan) and a heat-resistant resin (ProTemp SL, XYZ Printing Japan, Japan). The process is as follows:

(1)

The device design, created using the CAD software Fusion (Version 2605.1.1,8 Autodesk, San Francisco, CA, USA), was printed using a 3D printer.

(2)

Then, printed molds were immersed in ethanol and cleaned in an ultrasonic cleaner (SWS510, Citizen Systems Co., Ltd., Tokyo, Japan) for 5 min.

(3)

After washing, the molds underwent UV irradiation for 10 min in a UV curing chamber (EeezCure180, XYZprinting, Taipei, Taiwan).

(4)

Then, the supports were removed, and the surfaces were sanded.

2.2.2. Chamber Fabrication Procedure

Polydimethylsiloxane (PDMS, SILPOTTM 184W/C, The Dow Chemical Company, Midland, MI, USA) was used to fabricate the device chamber. The process is as follows:

(1)

The PDMS base and hardener were mixed in a 10:1 ratio, depressurized to −0.08 MPa using a vacuum pump (VP-215N, Aitcool, Taizhou, China), and defoamed for 30 min.

(2)

The solution was poured into the molds and allowed to defoam for 30 min.

(3)

After defoaming, the PDMS was cured by baking the molds at 100 °C for 30 min on a hot plate (Ninos ND-1A, AS ONE Corporation, Osaka, Japan).

(4)

The baked PDMS parts were removed from the mold using tweezers and ethanol.

2.2.3. Balloon Actuator Fabrication Procedure

A balloon actuator was fabricated using the following procedure:

(1)

A thin film of PDMS was dropped onto the substrate and spin-coated at 2000 rpm for 30 s.

(2)

Then, the thin film was cured by baking on the hot plate at 100 °C for 1 min.

(3)

Both the balloon and thin film were irradiated 60 s with excimer lamps (Min-Excimer, Ushio, Tokyo, Japan).

(4)

Then, the balloon and thin film were bonded and baked on the hot plate at 120 °C for 1 h.

2.2.4. Fabricated Device

The fabricated chamber and balloon are shown in Figure 3a. A microgram of the chamber is shown in Figure 3b.

2.3. Cell Preparation

The N87 cell line was used to prepare 3D cultured cells. D-MEM (high glucose) with L-glutamine, phenol red, and sodium pyruvate (FUJIFILM Wako, Osaka, Japan) was used as the culture medium. N87 cells at a density of 2 × 10⁴ cells/mL were seeded into 96-well plates (100 μL per well) and cultured in a CO₂ incubator at 37 °C for 3 d to form spheroids. The resulting N87 spheroids had diameters of approximately 100–200 μm.

2.4. Experimental Setup

In this section, the experimental setup and procedures are described. The balloon inflation was measured. Cellular imposition using the soft actuator was measured. Fluorescently stained spheroids were observed using a TIRF microscope.

2.4.1. Method for Measurement of Balloon Expansion

A schematic of the measurement setup is shown in Figure 4. A 2.5 mL syringe filled with deionized water was connected to the balloon actuator, and a syringe pump (KDS210, KD Scientific, Holliston, MA, USA) was used to inject water into the balloon. The actuator was prefilled exclusively with deionized water, and no gas was introduced. Deionized water (130 μL) was injected at a rate of 10 µL/min in increments of 10 µL. The balloon actuator was placed on an automatic contact angle meter (SImage AUTO 100, Excimer Inc.,Yokohama, Japan) and pressurized with deionized water. Images of the inflated actuator were captured using a camera. The maximum apex height was determined using ImageJ (Version 1.54, NIH, Bethesda, MD, USA) by referencing the baseline (unpressurized membrane) and the apex of the inflated membrane. Three measurements were obtained for each volume, and the values were averaged.

2.4.2. Method for Cellular Imposition Using the Soft Actuator

A schematic of the experimental setup is shown in Figure 5. A syringe pump (KDS210) and a 2.5 mL syringe were used to inject deionized water into the balloon actuator to pressurize the cells. The 3D-cultured cells were placed in the chamber of the device, which was then tilted to guide the cells toward the inlet hole. After the cells entered the hole, they were observed under the microscope (Figure 5). The balloon actuator was positioned over the device, and up to 90 μL of deionized water was injected at a rate of 10 µL/min in increments of 10 µL.

The ratio A(V)/A(0) was calculated, where A(V) is the projected contact area of the spheroid on the glass surface at an injected volume V, and A(0) is the projected area at 0 μL (baseline). Areas were determined from thresholded binary masks in ImageJ (units: μm²), using the same threshold for all images in the sequence.

2.4.3. Method for Observation of Spheroids by TIRFM Under Compression

Fluorescence observations were performed using an objective-based total internal reflection fluorescence (TIRF) microscope while spheroids were mechanically compressed by the balloon actuator. Cultured spheroids were stained with a plasma membrane-specific fluorescent probe (Cell Navigator™ Cell Plasma Membrane Staining Kit, Green Fluorescence; AAT Bioquest, Pleasanton, CA, USA) and placed on glass coverslips with a thickness of 0.13–0.17 mm (C050701, Matsunami Glass, Osaka, Japan) integrated into the microfluidic device.

TIRF and epifluorescence imaging were performed using an inverted fluorescence microscope (IX71, Olympus, Japan) equipped with an oil-immersion objective lens (UApo N 100×, NA 1.49; Olympus, Tokyo, Japan). For TIRF imaging, excitation was provided by a 488 nm laser, and emission was collected through a dichroic mirror optimized for 488 nm excitation (488/561 nm) and a band-pass emission filter (505–540 nm). The laser power at the source was set to 13.5 mW, and fluorescence images were acquired using a CCD camera (iXon Ultra 897, Oxford Instruments, High Wycombe, England). with an exposure time of 100 ms. For epifluorescence imaging, a mercury lamp was used as the excitation light source in combination with a standard NIBA filter set. During imaging, 80 μL of deionized water was injected into the balloon actuator at a flow rate of 10 μL/min to apply controlled compression to the spheroids.

3. Results and Discussions

3.1. Measurement of Balloon Expansion

The measured height of the balloon actuator increased with the injected volume of deionized water (Figure 6). Balloon expansion experiments were repeated three times under identical conditions. Representative images of balloons at different injected water volumes are shown in Figure S3. Expansion exhibited an approximately linear trend up to 130 μL, after which the rate of increase became slightly steeper. At 90 μL, which was the condition used in compression experiments, the balloon reached a height of approximately 0.91 mm and continued to expand to approximately 1.2 mm at 130 μL. These results confirm that the balloon actuator can achieve vertical displacements exceeding 500 μm under moderate loading conditions. This displacement was sufficient to establish contact between the balloon and spheroids positioned in a chamber approximately 480 μm below the actuator mounting surface.

The balloon shape remained stable after water injection because the syringe pump piston was fixed in position. In principle, this configuration allows the balloon to maintain its shape without strict time limitations. However, considering practical factors such as gradual water evaporation through the PDMS membrane, the shape was typically stable for approximately 30–60 min, which was sufficient for the TIRFM observations performed in this study.

To assess the applied pressure, the correlation between measured pressure and balloon expansion height was characterized (Figure S4). For actuator operation, injected volume was used as the primary control parameter. This approach is more straightforward and reproducible in real time than direct pressure control.

3.2. Cellular Imposition Using the Soft Actuator

The effect of gradual compression on spheroid structural clarity was investigated by acquiring a series of images with an inverted microscope while injecting deionized water into the balloon actuator. Water was added in 10 μL increments at a flow rate of 10 μL/min, yielding 10 sequential image frames corresponding to injection volumes from 0 to 90 µL.

The relationship between injected water volume and the spheroid area magnification factor was analyzed using ImageJ. Quantification of the projected spheroid area under increasing compression is shown in Figure 7, and representative images at different compression levels are presented in Figure S5. The ratio of the projected area at each injection volume to that at the initial state (0 µL) was plotted on the vertical axis. For all spheroids, the projected area increased gradually with increasing injection volume, indicating progressive flattening and spreading on the glass surface under applied pressure.

At higher injection volumes, the limited number of data points precluded a definitive conclusion. However, the observed trend suggested a plateau or slight decrease. Additional measurements in the 60–80 μL range will be performed in future work to clarify this behavior. This behavior can be attributed to the mechanical resistance of the spheroid. Once a certain degree of deformation is reached, further expansion is constrained by cell–cell adhesion and membrane elasticity. Moreover, excessive compression may partially disrupt the three-dimensional structure, leading to an apparent reduction in projected area. These results suggest that an injection volume of approximately 60–80 µL provides an optimal pressure range to effectively flatten spheroids while preserving structural integrity.

Importantly, spheroids retrieved after compression and subsequently cultured exhibited no significant difference in growth rate compared with uncompressed controls. Thus, the compression protocol did not adversely affect cell viability or proliferation (Figure S6).

3.3. Observation of Spheroids by TIRFM Under Compression

We evaluated the effect of balloon actuator-mediated compression for the visualization of 3D-cultured spheroids under evanescent illumination and epifluorescence (widefield) imaging. For this, TIRFM and epifluorescence observations were conducted both with and without the application of pressure. Representative phase-contrast, TIRF, and epifluorescence images are shown in Figure 8. Although different spheroids were used for the compressed and uncompressed conditions, all imaging parameters were kept constant. Low-magnification phase-contrast image of each spheroid, showing the overall morphology and the region subjected to TIRF illumination is shown in Figure 8(id,iid’).

Without compression (Figure 8i), fluorescence signals at the interface region appeared diffuse and weak. By contrast, balloon compression increased the contact area between the spheroid and the glass coverslip, revealing clearer spatial variations in fluorescence intensity, with the effect being more pronounced in TIRF images than in epifluorescence images (Figure 8ii). Quantitative analysis based on the standard deviation of fluorescence intensity within the regions of interest demonstrated a significant increase under balloon compression (Figure S7, p = 0.027), indicating enhanced spatial heterogeneity of membrane-associated fluorescence signals. Representative line profile analyses further showed steeper intensity fluctuations and clearer local contrast in compressed spheroids compared with non-compressed ones (Figure 9).

Although epifluorescence imaging with a high-numerical aperture objective provides a relatively shallow depth of field and can visualize the contact region, it excites fluorophores throughout the spheroid volume and therefore suffers from higher background signals. In contrast, TIRF imaging selectively excites fluorophores within approximately 200 nm of the glass surface, making the improvement in image quality produced by balloon compression more evident.

These results indicate that balloon-based mechanical compression improves the effective optical accessibility for fluorescence imaging of 3D-cultured spheroids, particularly for membrane-associated signals observed by TIRF microscopy.

3.4. Limitation and Future Prospective of This Research

Although the proposed system successfully improved fluorescence imaging of 3D-cultured spheroids at the spheroid–glass interface, several limitations remain. First, the applied hydrostatic pressure may not be perfectly uniform across the spheroid, potentially introducing local variations in deformation. Quantitative pressure measurements and calibration curves have been included in the supplementary information to improve reproducibility. Nevertheless, further optimization of pressure distribution is required. Second, the balloon actuator is fabricated from PDMS, which is known to be gas permeable. In our experiments, the actuator was filled exclusively with deionized water to ensure that no gas was introduced, thereby minimizing the risk of bubble formation. However, long-term stability may still be affected by gradual water evaporation through the PDMS membrane. The balloon shape is typically maintained for 30–60 min, which is sufficient for TIRFM observations but may limit extended imaging sessions.

From an imaging perspective, TIRFM inherently restricts observations of the near-surface region (<200 nm), limiting access to intracellular structures. Although high-numerical-aperture epifluorescence imaging can also visualize the contact region, it excites fluorophores throughout the spheroid volume and, therefore, produces higher background signals. By contrast, TIRF selectively excites fluorophores near the glass surface. Accordingly, our system is designed to enhance surface accessibility for high-contrast visualization of membrane-associated fluorescence signals rather than to replace volumetric imaging techniques. Several complementary imaging modalities, including spinning-disk confocal microscopy, lattice light-sheet microscopy, and oblique plane microscopy, provide deeper penetration and volumetric imaging capabilities and should, therefore, be considered complementary to our approach.

Additionally, although cell viability was maintained after compression, the long-term effects of repeated or prolonged compression were not evaluated. The present study was conducted using N87 gastric carcinoma spheroids. Therefore, further validation using different cell types and spheroid sizes is required to confirm the general applicability of this approach.

From a technical perspective, the current system relies on manual or semi-automated control of water injection, which may limit high-throughput applications. Future work will focus on integrating pressure sensors for real-time feedback control, improving automation, and combining the compression system with rotational mechanisms to enable multi-angle imaging.

Finally, the applicability of the proposed device to different spheroid sizes should be noted. In this study, spheroids with diameters of 100–200 μm were used. However, the chamber dimensions allow accommodation of spheroids up to approximately 300 μm. Therefore, the device is applicable to spheroids in the size range of 50–300 μm. Future work will test larger spheroids and additional cell types to further confirm the versatility and generalizability of this approach.

4. Conclusions

We developed a microfluidic system that improves fluorescence imaging at the spheroid–glass interface. By employing a water-driven balloon actuator, we successfully established a method to gently press 3D spheroids against a glass surface, thereby enhancing the effective optical accessibility of membrane-associated fluorescence signals under total internal reflection fluorescence (TIRF) imaging conditions and epifluorescence, with more evident improvement observed in TIRF images. This approach offers a simple and adaptable platform for studying cell–substrate interactions and membrane-associated molecular dynamics in three-dimensional cell culture models and can be combined with complementary imaging techniques in future applications. In the future, we will develop a technique to rotate the spheroids using fluid flow, allowing for multi-angle observation of the cell surface.