Chips-on-a-plate device for monitoring cellular migration in a microchannel-based intestinal follicle-associated epithelium model

Abstract

This paper describes a chips-on-a-plate (COP) device for monitoring the migration of Raji cells in the Caco-2/Raji coculture. To generate a model of the human intestinal follicle-associated epithelium (FAE), the coculture method using a conventional Transwell cell culture insert was established. Due to the structural limitations of the Transwell insert, live-cell tracking studies have not been performed previously using the existing FAE model. In this study, we designed a COP device to conduct long-term live-cell tracking of Raji cell migration using a microchannel-based FAE model. The COP device incorporates microfluidic chips integrated on a standard well plate, consistent humidity control to allow live-cell microscopy for 2 days, and microchannels connecting the two cell culture chambers of the COP device, which serve as a monitoring area for cellular migration. Using the COP device, we provide the first analysis of various migratory characteristics of Raji cells, including their chemotactic index in the microchannel-based FAE model. We showed that the migration of Raji cells could be controlled by modulating the geometry of the connecting microchannels. Cellular treatments with cytokines revealed that the cytokines increased the permeability of an FAE model with a detachment of Caco-2 cells. Live-cell monitoring of Raji cells treated with a fluorescent reagent also indicated exocytosis as a key agent of the Caco-2/Raji interaction. The COP device allows live-cell tracking analyses of cocultured cells in the microchannel-based FAE model, providing a promising tool for investigating cellular behavior associated with the recruitment of Raji to Caco-2 cells.

INTRODUCTION

The human intestines, containing the largest number of immune cells in the body, are exposed to a variety of antigens, including commensal and pathogenic microorganisms.¹ Gut-associated lymphoid tissue (GALT) consists of follicles of immune cells and associated tissues, which orchestrate the intestinal immune system.² Different from the majority of the intestinal epithelium, the follicle-associated epithelium (FAE), which covers GALT, senses luminal antigens and translocates them into GALT.³ The FAE is considered to be an entrance for pathogens and drugs, thus attracting recent scientific investigation into its involvement in pathogenesis and drug delivery.

Kernéis et al. established the first in vitro FAE model, consisting of two immortalized cell lines, Caco-2 and Raji cells, cocultured in a Transwell insert.⁴ Caco-2 cells, a human intestinal epithelial cell line derived from colorectal adenocarcinoma, were selected to model intestinal epithelial cells; Raji cells, a human B lymphocyte cell line derived from Burkitt's lymphoma, were selected to model intestinal lymphocytes of GALT. The Transwell in vitro model simplifies the complexity of the FAE tissue, facilitating studies focused on various cellular mechanisms. For example, Transwell-based studies were used to characterize genetic similarities between the in vitro model and the human FAE.⁵ In addition, various Transwell-based models were used to investigate hypotheses deduced from in vivo studies, elucidating surface markers and antigen translocation mechanisms of the FAE.^6–8 Nevertheless, the Transwell apparatus has limitations, which have prevented the analysis of Raji cell migration; instead, studies have focused on Caco-2 cells for the FAE model. As in vivo studies have illuminated lymphocytic migration and contact with intestinal epithelial cells to differentiate the FAE,³ the migratory behavior of Raji cells in Caco-2/Raji coculture could be informative about the formation of the FAE-like phenotype in Caco-2 cells. Additionally, the translocation of Raji cells through 3-μm pores of the Transwell membrane would signify active migration toward Caco-2 cells.⁹ Live-cell monitoring is required to investigate the migration of and cellular interaction between cocultured Caco-2 and Raji cells. However, Transwell assays allow only indirect visualization. Accordingly, end-point visualization techniques such as immunohistochemical imaging, confocal microscopy, and transmission electron microscopy have been used in Transwell-based models.^10–12

To investigate live monitoring of cellular migration, basic principles of microfabrication were applied to develop cell culture devices.^13–15 Heralded by the invention of soft lithography,¹⁶ advancements in microfluidics and microfabrication technologies have resulted in the development of numerous cell culture devices made of poly(dimethylsiloxane) (PDMS) for live-cell monitoring.^17–21 Previously, PDMS-based microfluidic cell culture devices were used to measure the migratory rate and directionality against soluble stimulants of Raji cells.^22,23 However, a comparable analysis of Raji cells in Caco-2/Raji coculture has not been reported. The Caco-2/Raji coculture requires a long culture period in a hydrostatic environment.⁹ Although existing microfluidic cell culture devices enabled long-term culture, customized perfusion pumps are required to provide a hydrostatic environment.^24–27 Recently, an open-well based microfluidic device was used successfully in long-term culture; however, its dimensions were specific for the coculture of mammalian and bacterial cells.²⁸

Here, we report a microfluidic cell culture chips-on-a-plate (COP) device for live-cell monitoring of the Caco-2/Raji coculture. The device consists of microfluidic chips for cell culture, integrated on a 6-well plate for live-cell microscopy. The open-well design of the COP device is compatible with a common culture plate and ensures consistent prevention of media evaporation, allowing maintenance of long-term coculture and live-cell monitoring. The small molecule diffusion properties and the apparent permeability of Caco-2 cells cultured in the COP device were analyzed to validate the establishment of the in vitro microchannel-based FAE model consisting of the Caco-2/Raji coculture. This study demonstrated visual tracking and analyses of the migratory behavior of Raji cells cocultured with Caco-2 cells in the COP device. This device will facilitate investigations of key characteristics of the microchannel-based FAE model.

MATERIALS AND METHODS

Design and fabrication of a COP device

A COP device was fabricated using soft lithography of PDMS. To prepare the mold, the negative photoresist SU-8 (MicroChem Corp., Newton, MA, USA) was spin-coated on a bare silicon wafer and developed into microstructures using photolithography. The PDMS elastomer was mixed with a curing agent at a ratio of 10:1 (w/w). The PDMS mixture was cast onto the mold and cured on a hot plate at 120 °C for 2 h to complete the cross-linking. Cured chips were sterilized with 70% ethanol and autoclaved. The COP devices and the surfaces of a glass-bottomed confocal 6-well plate (SPL Life Sciences Co., Ltd., Pocheon, Korea) were plasma-activated and bound together. The bonding was completed by placing the assembly in a dry oven at 85 °C for 30 min. The COP device was filled with media prior to use.

Caco-2/Raji coculture

Caco-2 and Raji cells were obtained from the American Type Culture Collection (Manassas, VA, USA) and maintained in Dulbecco's modified Eagle's medium supplemented with 4.5 g/l glucose, 50 U/ml penicillin, 50 U/ml streptomycin, and 10% fetal bovine serum (media components purchased from Corning Inc., Corning, NY, USA). The cells were cultured to confluence in T25 flasks. The Caco-2 cells were collected by trypsinization while the Raji cells were collected by centrifugation, and the cell numbers were counted using an automated hemocytometer (Logos Biosystems, Inc., Anyang, Korea). Following a seeding method described previously,²⁰ Caco-2 cells were introduced into the Caco-2 chamber of the COP device at a concentration of 200 000 cells/ml. The Caco-2 cells were cultured up to 14 days (∼7 days after confluence) to promote their epithelial phenotype.²⁹ Subsequently, Raji cells were seeded at a concentration of 200 000 cells/ml into the Raji cell chamber. The COP device was placed upright (e.g., perpendicular to the floor) in a humidified 37 °C incubator for 15 min to allow settling of the Raji cells at the entrances of the connecting microchannels between the two cell culture chambers.

Characterization of diffusive transport

The connecting microchannels were reconstituted on a 2D plane and analyzed with computational fluid dynamics (ESI Group SA, Paris, France). The COP device was filled with distilled water, and one drop of 100 μM 10 kDa fluorescein isothiocyanate–dextran (FD-10) was introduced into the Caco-2 chamber. The diffusive transport of FD-10 through the connecting microchannels was recorded using a fluorescent microscope (AX10; Carl Zeiss AG, Jena, Germany). The acquired images were analyzed with ImageJ (https://imagej.nih.gov/ij/) to measure the gradient of fluorescent intensity over the middle section of the connecting microchannels at 10 min. The fluorescent intensity was compared to a mathematical prediction expressed by the equations listed below. The diffusion coefficient of FD-10 in water at 37 °C was calculated using the Stokes–Einstein equation [Eq. (1)], where D_FD-10 is the diffusion coefficient of FD-10, k_B is Boltzmann's constant, T is the temperature in Kelvin, η_water is the dynamic viscosity of water at 37 °C, and r_FD-10 is the radius of FD-10 reported as 2.9 nm,³⁰

$$D_{\mathrm{FD}\text{-}10}={\displaystyle \frac{k_{\mathrm{B}}T}{6\pi {\eta }_{\mathrm{water}}\times r_{\mathrm{FD}\text{-}10}}=1.131548\times {10}^{-10}{\mathrm{m}}^2{\mathrm{s}}^{-1}.}$$ (1)

For mathematical prediction, the measured middle section of the connecting microchannels was simplified as a straight line, and the mathematical prediction of the concentration of FD-10 was modeled as diffusion in a semi-infinite medium [Eq. (2)],³¹ where C(x,t) is the concentration at point x at time t, C₀ is the initial concentration, and D is the diffusion coefficient from Eq. (1),

$$C(x,t)=C_{0}erfc\left({\displaystyle \frac{x}{2\sqrt{(Dt)}}}\right).$$ (2)

Measurement of media evaporation

The COP device was filled with 220 μl of 100 μM erioglaucine aqueous solution and humidified using integrated humidity control or conventional wet tissue paper humidification. The COP device was placed in a conventional humidified incubator at 37 °C, and samples were taken every 12 h. A microplate spectrophotometer (Molecular Devices, LLC, San Jose, CA, USA) was used to measure the concentration of erioglaucine from which the volume loss was calculated. The differences were considered significant at p < 0.05.

FD-10 permeability assay

FD-10 (100 μM) in Hank's balanced salt solution (Corning Inc.) was added into the Caco-2 chamber of the COP device with Caco-2 cells cultured for 14 days. The opposite chamber contained Raji cells cocultured for 3 days, and the control condition contained a single culture of Raji cells without coculture. The cells were incubated at 37 °C for 2 h. Samples were taken from the Raji chamber, and the fluorescent intensity of FD-10 was measured using a multilabel plate reader (PerkinElmer, Inc., Waltham, MA, USA). The apparent permeability coefficient (P_app) was calculated, where dQ/dt is the flux for 2 h, A is the total area of the entrances of the connecting microchannels, and C₀ is the initial concentration of FD-10 [Eq. (3)].³² Differences were considered significant at p < 0.05,

$$P_{\mathrm{app}}={\displaystyle \frac{dQ}{dt}\times {\displaystyle \frac{1}{A\times C_0}.}}$$ (3)

Live-cell tracking analysis

Caco-2 cells were cultured for 14 days in the Caco-2 chamber, and Raji cells were introduced into the Raji chamber as described in the Caco-2/Raji coculture section. The COP device was mounted on a microscope stage-top incubator (Live Cell Instrument, Seoul, Korea) for live-cell tracking. Images were acquired every 10 min for 48 h, and the observed migration of Raji cells was tracked as coordinates on an xy plane using the MTrackJ plugin of ImageJ. The migration speed was calculated by dividing the total migratory distance over time. The chemotactic index was calculated by dividing the absolute distance between the initial and final position of the Raji cells over their total migratory distance.²³ Differences were considered significant at p < 0.05.

Chemical treatment of Caco-2 and Raji cells in a COP device

Tumor necrosis factor-α (TNF-α) and lipopolysaccharide (LPS) from E. coli O26:B6 (Sigma-Aldrich, St. Louis, MO, USA) were used for the proinflammatory cytokine treatments of Caco-2 cells. Caco-2 cells cultured to confluence in a COP device were exposed to LPS (10 μg/ml) or TNF-α (100 ng/ml) solubilized in Hank's balanced salt solution (Corning Inc.) for 1 day. The cytokine solution was replaced with the growth medium for subsequent Raji seeding. For visualization in fluorescent microscopy, Raji cells were stained with 5-chloromethylfluorescein diacetate (CMFDA; Thermo Fisher Scientific, Inc., Waltham, MA, USA).

RESULTS AND DISCUSSION

The COP device for monitoring cellular migration in a microchannel-based FAE model

The COP device was designed to study cellular migration in the Caco-2/Raji coculture, representing in vitro microchannel-based FAE. Using the device, Caco-2 cells were cultured to confluence, and then Raji cells were seeded into the opposite chamber. The connecting microchannels linked the two cell culture chambers and served as passages for migration of Raji cells to interact with Caco-2 cells [Fig. 1(a)]. A PDMS disk with micropatterned structures and punched inlets was bound on a glass-bottomed 6-well plate to form the COP device [Fig. 1(b)]. In the complete COP device, diffusive transport of FD-10 through the connecting microchannels was characterized, as it was assumed that chemokines secreted by Caco-2 cells regulate Raji migration.^33–35 The computational simulation of the fluorescent tracer diffusing through the connecting microchannels was comparable to the fluorescent images of actual diffusion [Fig. S1(a) in the supplementary material], and the measured fluorescent intensity was comparable to the mathematical prediction of diffusion [Fig. S1(b) in the supplementary material].

FIG. 1.

Design principle of a chips-on-a-plate (COP) device. (a) Illustration of Caco-2/Raji cellular migration analysis. Caco-2 cells were cultured to confluence in one chamber, and then Raji cells were introduced to the opposite chamber. Raji cells interacted with Caco-2 cells via the connecting microchannels, and their migration was tracked using live-cell microscopy. (b) Design and dimension of the COP device. An exploded view displays the four regions composing the COP device. The Caco-2 and Raji chambers were juxtaposed between the connecting microchannels. The COP device was generated by bonding PDMS chips on the glass-bottomed wells. Blue-dyed water was injected into the chambers to visualize the inner structures of the device. A microscope image shows the dimension of the connecting microchannels (3 μm height × 50 μm width × 500 μm length). (c) Photograph of a complete device. The peripheral area within a well was used as an internal water bath maintaining optimal humidity for several days of live-cell microscopy. (d) Experimental workflow. Schematics representing side views of the COP device along the a–a′ cut in panel (c). Caco-2 cells were seeded and cultured to confluence prior to loading Raji cells. Live-cell microscopy was performed immediately following Raji cell seeding.

Figure 1(c) shows a photograph of six COP devices integrated on a conventional 6-well plate. The peripheral area in each well was used as an internal water bath to maintain optimal humidity during several days of live-cell microscopy. Furthermore, the remaining area in each well acted as a humidity controller that prevented rapid media evaporation, which is problematic when maintaining a microvolumetric cell culture. The internal chambers and the humidity controlling area were filled with medium to prime the device prior to seeding Caco-2. The medium was changed periodically up to 14 days until the Caco-2 cells formed a confluent cell layer. Live-cell microscopy commenced as Raji cells were introduced into the Raji chamber, focusing on the connecting microchannels to monitor their migratory behavior [Fig. 1(d)]. Integration into commercial cultureware allowed improved reliability and ease-of-use of the COP device as a microfluidic cell culture device.^28,36–38 In addition, a microfluidic device integrated into commercial cultureware was easily mounted (without modification) on a commercial microscope stage-top incubator for live-cell monitoring. It is expected that the analytical throughput could be increased by the use of a fully motorized microscope, which allows simultaneous image acquisition on multiple areas of interest.³⁷

Prevention of media evaporation in the COP device

As the Caco-2/Raji coculture requires long-term microvolume culture in the COP device, evaporation of the media is a concern. Microvolumetric cell culture is susceptible to media evaporation, causing inaccurate readings of cell numbers and drug efficacy.³⁹ In addition, media evaporation in a PDMS-based microfluidic cell-culture device could introduce a critical failure causing a dramatic increase in media osmolality, which affects cellular behavior and viability.⁴⁰ However, a few studies on evaporation have been performed aimed at increasing the duration of monitoring.^40–42

In this study, the peripheral area of the 6-well plate attached to the COP device served as an integrated humidity controller with consistent geometry. Caco-2 and Raji cells were easily maintained in the open-well design, while maintaining the internal humidity with the closed lid of the 6-well plate. The preventive efficacy of the COP device was compared to the use of wet tissue paper surrounding the COP device. We compared the three humidity configurations (no source of external humidity, the integrated humidity controller, and wet tissue paper) [Fig. 2(a)] in a 37 °C humidified incubator for 48 h. Whereas the lack of proximal humidity resulted in complete evaporation, the media volume was maintained within 3% of its original volume over 48 h (a period between media exchanges) when the integrated humidity controller was active. Without this function, the media volume was reduced to 80% in 12 h [Fig. 2(b)]. Therefore, basic cell culture could not be maintained without the proximal source of humidity. The surrounding wet tissue paper also prevented media evaporation, but with a deviation of 10%. Such inconsistent results might be due to manual handling of the wet tissue paper when surrounding the apparatus. Our results emphasize the need for a proximal humidity source for microvolumetric cell culture, requiring a reliable media evaporation prevention system such as the integrated humidity controller of the COP device.

FIG. 2.

Prevention of media evaporation in the COP device. (a) Images of the COP device operated with or without the integrated humidity controller and wet tissue paper. Scale bar: 1 cm. (b) Changes during the evaporation of the medium over 48 h (n = 5). *p < 0.05.

Characteristics of a microchannel-based FAE model in the COP device

As the COP device differs structurally from the Transwell, the microchannel-based FAE model required verification using the established coculture protocol. Measurement of transepithelial electrical resistance (TEER) is a candidate assay for measuring changes in epithelial permeability of the Caco-2/Raji coculture; however, the TEER assay is highly sensitive to the placement of measuring electrodes, which are often designed for optimal use with a Transwell device. As summarized in a review of various studies investigating the development of new TEER equipment for non-Transwell microfluidic cell culture devices,⁴³ a different electrical measurement setup is required to conduct a TEER assay on a COP device. For the coculture of Caco-2 and Raji cells in a Transwell insert, the establishment of the FAE model was confirmed by the apparent permeability (P_app) of Caco-2 cells and the migration of Raji cells into the Caco-2 cell layer.^9,11 In this respect, P_app was measured instead of TEER, representing the amount of a Raji cell invasion in a COP device.

P_app is generally used to evaluate the permeation of the tracer molecules across the Caco-2 cell culture.^10,11,44 Here, 10 kDa FD-10 was chosen as the tracer, as it represented the molecular mass of chemokines.⁴⁵ As shown in Figs. 3(a) and 3(b), permeation of FD-10 occurred from the Caco-2 cell chamber into the Raji cell chamber at 2 h. The Caco-2 cells cultured in the COP device behaved as a barrier against permeation, and thus the amount of FD-10 in the Raji chamber could not be measured from images acquired using a fluorescent microscope. Therefore, a multilabel plate reader was used to measure the low concentration of FD-10 in the Raji chamber to calculate P_app. The P_app values were measured for FD-10 across the Caco-2 monolayer with or without the Raji cell coculture. The Caco-2 monolayer cocultured with Raji cells for 3 days showed a 1.76-fold increase in the P_app value [Fig. 3(c)]. Moreover, the invasion of Raji cells into the Caco-2 culture was observed [Fig. 3(d)]. The increased permeability in the coculture was similar to that described in previous reports.^10,44 The increased P_app and Raji cell invasion suggest that Caco-2 cells acquired an FAE-like phenotype.¹⁰ Thus, the application of the coculture protocol in the COP device displayed the two characteristics of in vitro FAE. Therefore, Raji cell migration toward Caco-2 cells could be visually tracked using the COP device.

FIG. 3.

Apparent permeability values (P_app) measured for FD-10 across a Caco-2 monolayer with or without Raji cell coculture. (a) Schematic of the FD-10 permeability assay. The amount of FD-10 that permeated the Raji chamber over 2 h was measured. (b) Microscope images of FD-10 diffusing through the connecting microchannels in a blank COP device (left) and with Caco-2 cells cultured for 14 days (right). Scale bar: 100 μm. (c) P_app for the Caco-2 single culture and the Caco-2/Raji coculture (n = 3). *p < 0.05. (d) The invasion of a single Raji cell into the Caco-2 layer was monitored. A Raji cell body is emphasized with white borderlines. Scale bar: 10 μm.

Live-cell tracking analysis of Raji cells

The migratory characteristics of Raji cells in the coculture were studied using live-cell microscopy. A single culture of Raji cells in the Raji chamber of the COP device was set as a negative control [Fig. 4(a), Multimedia view] and compared to the Raji cells of the microchannel-based FAE model [Fig. 4(b), Multimedia view]. Time-lapse imaging was used to monitor Raji cells migrating inside the connecting microchannels of the COP device [Figs. 4(c) and 4(d)]. Raji cells introduced into the Raji chamber showed a tendency to congregate in clusters from which the migrating cells moved out. The majority of Raji cells in the control condition returned to the clusters [Fig. 4(e)]. On the other hand, Raji cells in the microchannel-based FAE model displayed migration toward the Caco-2 layer [Fig. 4(f)], infiltrating the Caco-2 culture. Such behavior was comparable to that of the Raji cells found within the membrane or the Caco-2 layer of the Transwell-based FAE model.^10,11

FIG. 4.

Live-cell tracking of Raji cells in a COP device. (a) An image showing a single culture of Raji cells monitored for 48 h in the Raji chamber of a COP device and (b) an image showing a coculture of Caco-2 and Raji cells monitored for 45 h in a COP device. Both images were taken at the 30th hour of live monitoring experiments. Scale bars: 100 μm. In the dashed rectangular areas, images (c) and (d) show the migration of Raji cells inside a connecting microchannel for 8 h. The triangle symbols indicate the time-dependent position of a selected Raji cell. The migration tracks of Raji cells in (e) the control and (f) cocultures, normalized to a common origin (0,0) on an xy plane. Multimedia views: https://doi.org/10.1063/1.5128640.1Download video file ^{(5.4MB, mp4)} DOI: 10.1063/1.5128640.1; https://doi.org/10.1063/1.5128640.2Download video file ^{(5.1MB, mp4)} DOI: 10.1063/1.5128640.2

Migratory characteristics were analyzed and compared between the control (a single culture of Raji cells) and the microchannel-based FAE model in the COP device with 3- or 8-μm high connecting microchannels. The change in height of the connecting microchannels revealed the importance of structural hindrance for migrating cells. At the onset of live-cell monitoring, Raji cells were observed inside the 8-μm high connecting microchannels [Fig. S2 in the supplementary material]. As the measured average diameter (10.83 ± 1.64 μm) of Raji cells is similar to the 8-μm height, a considerable number of Raji cells randomly entered the connecting microchannels without geometric constraints. The migration rates of the Raji cells were recorded in the range of 0.75–1.25 μm/min under the three conditions [a single culture of Raji cells (control) and in coculture in the COP device with 3- or 8-μm high microchannels] [Fig. 5(a)]. Visual tracking was initiated when a Raji cell entered the connecting microchannels and terminated when the observed cell returned to a cluster of Raji cells or infiltrated the Caco-2 culture. The tracking time of the Raji cells of the microchannel-based FAE model in the COP device with 3-μm high connecting microchannels showed greater consistency than that under the other two conditions [Fig. 5(b)]. The low chemotactic index observed in the control condition was consistent with the reported value obtained from a microfluidic chemotaxis assay.²³ Indeed, the chemotaxis value obtained in the COP device increased fourfold in the coculture [Fig. 5(c)]. Consistent and directed migration of Raji cells toward the Caco-2 layer was observed in the 3-μm high connecting microchannels. This report is the first demonstration of live-cell monitoring; however, our results showed that the migratory tendency of Raji cells toward the Caco-2 layer was not dominant. The majority of Raji cells seeded inside the connecting microchannels migrated toward the seeding inlet. Removing the Raji cells clustered in the inlet will allow visualization of Raji migratory behaviors in relation to the connecting microchannels.

FIG. 5.

Comparison of the migratory behavior of Raji cells under three conditions: single culture of Raji cells (control, n = 6), coculture in the COP devices with 3-μm high connecting microchannels (FAE 3 μm, n = 10), and coculture in the COP devices with 8-μm high connecting microchannels (FAE 8 μm, n = 10). (a) Migration speed of Raji cells. (b) Tracking time of Raji cells, measured until incorporation into a cluster of Raji cells or infiltration into the Caco-2 monolayer. (c) Chemotactic index of the Raji cells. *p < 0.05.

Effects of proinflammatory cytokines and a fluorescent dye on the in vitro FAE cellular behaviors

Caco-2 cells of the microchannel-based FAE were treated with LPS or TNF-α to study the effect of proinflammatory cytokines on Raji cell migration. At the beginning of the time-lapse imaging, Caco-2 cells detached from the entrances of the connecting microchannels were observed in both treatments [Fig. 6(a), Multimedia views]. Reattachment of Caco-2 cells to the entrances indicated cellular viability maintained with the cytokine treatments. No Raji cells entering the connecting microchannel were observed, and the motility of the Raji cells residing in their culture chamber decreased with the cytokine treatments [Fig. 6(b)]. Stimulating Caco-2 with TNF-α was reported to increase the secretion of an in vivo FAE chemokine CCL20.⁴⁶ However, Raji does not express CCR6, which is the receptor of CCL20.^34,47 The reduced Raji motility measured in this study agreed with the inability of Raji to recognize CCL20 and disputed its role as an FAE protagonist in the Caco-2/Raji coculture. With no mean for live-cell monitoring, changes in cellular behavior of a Transwell-based FAE were inferred from the measurements of Caco-2 permeability. Increased permeability of the Caco-2 culture by TNF-α treatment implied a stronger FAE-like function established in a Transwell-based FAE.⁴⁸ On the other hand, the COP device provided direct observational data from both Caco-2 and Raji cells in the microchannel-based FAE coculture. The result of studying cellular behavior under the influence of proinflammatory cytokines suggested that the increased Caco-2 permeability was due to the detachment from the interconnecting passage rather than the strengthened FAE-like function.

FIG. 6.

Monitoring of Caco-2 and Raji cells in a COP device after cellular treatments. (a) The detachment of Caco-2 cells from the entrances of the connecting microchannels after treatments with LPS or TNF-α. Black arrows indicate the detachments. Scale bars: 100 μm. (b) Migration speed of Raji cells in the Raji chamber of a COP device with Caco-2 cells treated with LPS or TNF-α. (c) Brightfield and fluorescent images of CMFDA-treated Raji cells in the FAE model. Scale bars: 100 μm. White triangular symbols indicate granular bodies from Raji cells streaming through the connecting microchannels. Multimedia views: https://doi.org/10.1063/1.5128640.3Download video file ^{(1.9MB, mp4)} DOI: 10.1063/1.5128640.3; https://doi.org/10.1063/1.5128640.4Download video file ^{(5.7MB, mp4)} DOI: 10.1063/1.5128640.4; https://doi.org/10.1063/1.5128640.5Download video file ^{(2.8MB, mp4)} DOI: 10.1063/1.5128640.5

Live-cell monitoring of nonlabeled cells is considered as a gold standard assay because labeling reagents often interfere with cellular behavior. Nevertheless, fluorescent stains help locate cells that are hardly seen with brightfield microscopy. To enhance cell visibility, the FAE coculture in a COP device was constructed with Raji cells infused with a fluorescent tracking reagent, CMFDA. Fluorescent live-cell monitoring also revealed streams of granular bodies originating from Raji cells and migrating toward Caco-2 [Fig. 6(c), Multimedia view]. Using a COP device, the reported exocytosis of the fluorescent label from Raji cells was observed.⁴⁹ The observed streams of granular bodies suggested a bigger role of exocytosis than cellular migration in the Caco-2/Raji interaction of the FAE model.

CONCLUSIONS

Raji cell migration toward Caco-2 has been reported previously, however, the migratory characteristics were unclear due to the structural limit of the Transwell insert used. Microfluidic cell culture devices for live-cell analysis have often been optimized for assays shorter than 1 day. However, Raji cell migration in the cocultured FAE model requires at least 3 days to verify the end-point position. In this study, we developed a COP device to provide live-cell tracking for an extended period. During continuous live-cell tracking over 2 days, Raji cell migration was observed from beginning to end. This is the first report of Raji cell migration using a microchannel-based FAE model. The migration rate, actual tracking time, and chemotactic index were obtained from the live-cell tracking conducted with the COP device. Changing the height of the connecting microchannels elucidated the role of structural hindrance in reducing the random migration of Raji cells inside the COP device. The dominant self-clustering behavior of Raji cells observed in this study suggests that the COP device structure provides an effective microenvironment for promoting the interaction between Caco-2 and Raji cells. When the microchannel-based FAE model was influenced by proinflammatory cytokines, Caco-2 cells were detached from the entrances of the connecting microchannels, and Raji cells displayed reduced motility. This result suggested that the cytokines increased the permeability of an FAE model with morphological changes in Caco-2 culture, rather than the induction on FAE-like function. The involvement of exocytosis in the Caco-2/Raji interaction was inferred from the observation of massive streams of granular bodies emanating from Raji cells. The COP device for live-cell analyses of the microchannel-based FAE model is expected to help elucidate the cellular migration of Raji cells as an immunological response against stimulants such as microorganisms and cytokines.

SUPPLEMENTARY MATERIAL

See the supplementary material for a detailed description of the diffusive transport of the fluorescent tracer through the connecting microchannels (Fig. S1) and an image of Raji cells located inside the 8-μm high connecting microchannels (Fig. S2).