Systems that utilize microstructures, such as microfluidic channels, are being increasingly applied in biotechnology research. Cell culture inserts, which are used for the co-culture of cells, have utilized microstructures for many years. The trademarked system “Transwell®” is a well-known co-culture vessel that comprises two containers that are stacked vertically. The upper container is composed of a filter at the bottom through which the substances in the two containers can interact. This structure is essentially the same as a microfluidic machine because micro-materials that are secreted by the cells pass through a filter or microfluidic channel to affect distant cells. Thus, co-culture research is one of the most relevant fields in which microstructures have been applied.
Recently, several co-culture methods have emerged, and they can be broadly classified into two types [1
]: (1) direct co-culture, in which cells are mixed in a single well without separation, and (2) indirect co-culture, in which two different cell types are cultured simultaneously and isolated from each other using a separation mechanism. Although direct co-culture is more commonly employed, the use of this approach to analyze individual cells is complicated. In contrast, indirect co-culture involves placing the cells in different environments, and cell–cell interactions are mediated via humoral factors, meaning that it is easier to observe the cells because they are not mixed. Indirect co-culture systems are mainly classified into three types: (1) separation by filters, (2) separation by a semi-solid material such as a gel, with humoral factors exchanged via the gel, and (3) the formation of separate cell colonies or layers by culturing, in which complete isolation of cell types is not possible [1
]). A typical example of filter separation is a cell culture insert, which is characterized by a bucket-in-a-bucket structure that is constructed vertically. This type of co-culture plate is therefore referred to as a vertical-type co-culture plate (VTCP).
The most popular type of VTCP is the Boyden chamber, which was developed by Stephan Boyden in 1962 [2
]. It was initially designed to study cell invasion by counting the number of cells that passed through a filter [3
]. In the Boyden chamber co-culture method, the culture fluids at the top (on the insert) and the bottom (on the bottom dish) of a system are shared for cell–cell interaction studies. This type of VTCP is generally referred to as a “Transwell” style [4
] or “cell culture insert” type [7
VTCPs have long been used as the standard for co-culture, and this technique has not changed fundamentally because of the difficulty in making novel vessels. When using this design, it is expected that cells will grow on an insert, and that the liquid factors will be exchanged with the cells in the lower wells through the filter. However, this is not always achieved, mainly because of the fundamental problems associated with high cell densities. There are three main issues when using VTCPs:
The top container is invisible: Neither cell line could be simultaneously observed under a microscope. The cells in the culture vessel are usually observed from the bottom of the vessel because the culture fluid hinders observation from the top of the vessel. The upper container cannot be observed because of the short focal length of the microscope. Therefore, cells were only observed in the lower parts of the VTCP-type vessels (Figure S1a
Different materials could impact the results: The cells are not cultured on the same type of material surface: one group of cells is cultured on the filter material and the other on a plastic material; this difference could significantly impact the experimental results (Figure S1b
High cell densities are not considered: A high density of cells may prevent co-culture due to blockage of the filter. As the Boyden chamber was initially invented to evaluate cell invasion [2
], scenarios in which cells were present at high densities were not considered (Figure S1c
To overcome these problems and facilitate studies on cell–cell communication, we developed a novel horizontal-type co-culture plate (HTCP), as shown in Figure 1
. This design is based on recent advances in three-dimensional computer-aided design technology, which offers novel possibilities for creating different culture vessels. We hypothesized that a horizontal system would solve the problems associated with VTCPs; it would enable simultaneous observation of cells in both containers using an inverted microscope, through which a variety of cellular interactions and cell morphologies during the maturation of each cell line can be observed.
In addition, researchers can freely choose from a variety of commercially available filters that are 13 mm in diameter to control the sharing factors of the liquid media when using HTCP, meaning that a wide variety of filter options can be used with the system.
Intercellular communication has traditionally been considered to involve the interaction of cells with their neighbors through direct interactions that are mediated by soluble molecules, such as cell-secreted cytokines. However, recent studies have revealed that Extracellular vesicles (EVs) play an important role in cell–cell communication [9
]. Johnstone et al. [11
] first discovered vesicles secreted by sheep reticulocytes [12
] in 1987, which were subsequently named exosomes. Exosomes are a type of EVs that are secreted by many cells and are present in body fluids, including blood, urine, and saliva. Although the detailed features of exosomes remain unknown, the discovery of microRNA and RNA in EVs in 2006–2007 demonstrated the possibility that informative genetic transmission between cells may occur via exosomes [14
]. Since then, many researchers have turned their attention to EVs, and exosomes have since become a hot topic of study in various fields. EVs also include slightly larger plasma membrane-derived microvesicles and apoptotic bodies that are released during cell apoptosis [16
]. Although the lack of a clear definition has led to confusion among researchers, the description of an exosome as an EV has been driven mainly by the International Society for Extracellular Vesicles [17
]. EVs are currently divided into three main groups: (1) exosomes that are 50–150 nm in diameter and are bound by lipid bilayer membranes, (2) microvesicles that are 100–1000 nm in size and are directly secreted by cells, and (3) apoptotic bodies that arise from cell apoptosis [16
]. Since the precise definition of exosomes remains elusive because of the lack of standard protocols for identifying and distinguishing exosomes and microvesicles, exosomes have recently begun to be referred to as EVs. Although exosomes are released via exocytosis and microvesicles are released via plasma membrane shedding, EVs are an appropriate term to describe both because the different types of vesicles cannot be distinguished by ultracentrifugation.
As it became clear that extracellular vesicles play an essential role in communication between cells, it became necessary to analyze them separately. The EVs can be separated from cells using three main methods [18
]: (1) the ultracentrifugation method that uses weight differences, (2) using the properties of the surface proteins on extracellular vesicles, and (3) using the difference in the sizes of the EVs. The size-separation method is the most suitable method for use with microfluidic devices and requires either a filter or the fabrication of tiny pores in the device to enable size separation. The problem with this method is that extracellular vesicles are extremely small, with exosomes reaching only 50 to 150 nm in size. Track-etching technology is ideally suited for producing filters with uniform pores. Track-etched membrane filters were fabricated by irradiating a raw polymer film with high-energy heavy ions, creating damaged linear tracks across the film, which were later converted into pores by selective wet chemical etching (Figure S2a
). Etching conditions can be adjusted to obtain uniform pores with a precise diameter adapted to separate the extracellular vesicles (Figure S2b
). In addition, for the efficient filter, it is necessary to prevent substances from leaking across the pores in the filter. VTCPs are made by molding the culture vessel and filtering together, so that the filter cannot be replaced. In the HTCP, an O-ring is used to sandwich the filters, meaning that it is possible to replace the filter. We are the first group to our knowledge, to commercialize containers with multiple horizontal connections. HTCP has been reported in several studies [19
]. However, the migration characteristics of substances secreted by HTCP cells have not yet been reported. The novel HTCP vessel is expected to help advance cell co-culture research, including shedding new light on the role of EVs as messengers that can deliver signals between cells. Here, we describe the design of a HTCP system as a new type of co-culture vessel. We also describe the results of experiments conducted to evaluate the feasibility of exchanging liquid factors using a track-etched membrane in comparison to that of using a VTCP. The functional ability of the track-etched membrane to separate EVs was examined to elucidate the influence of EVs on co-cultures.
2. Materials and Methods
2.1. Device Design and Fabrication
The design of the HTCP, its assembly, size information, and how to use it are depicted in Figure 2
and Figures S3–S5
. All components comprising the HTCP were designed using Rhinoceros 6. Available online: https://www.rhino3d.co.jp/
(accessed on 14 November 2021). The products were prepared via injection molding by Shinko Chemical Co., Ltd. The design aims to achieve an easy transition from a monoculture to co-cultures, with or without an optional filter, and the option to culture separately with the possibility of transitioning to the co-culture mode at any time.
This product (HTCP) is already commercially available through AR Brown Co., Ltd. (Tokyo, Japan), Fujifilm Wako Pure Chemicals Corporation (Tokyo, Japan), Blast Inc. (Kawasaki, Japan), AdvantigenBioScience L.L.C. (Madison, WI, USA), Bulldog Bio. Inc. (Portsmouth, NH, USA), Xceltis GmbH (Mannheim, Germany), and 1st PhileKorea Inc. (Seoul, Korea). The product is named “ICCP”, “NICO-1”, “UniWells”, and “Pair-N-Share Tandem Co-Culture Wells & Plates”, but the product itself is the same.
2.2. Cell Lines
Human pancreatic cancer PANC-1 and AsPC-1 cells were obtained from RIKEN BRC (Tsukuba City, Japan). The cell lines were passaged for fewer than six months after resuscitation, cultured in complete Dulbecco’s modified Eagle medium (DMEM)-high glucose with 1% L-glutamine, 1% sodium pyruvate, 10% fetal bovine serum, and 1% penicillin-streptomycin (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), and maintained at 37 °C with 5% CO2
in DMEM. To establish stable pancreatic cancer cell lines expressing CD63-GFP (gPANC1) and CD63-RFP (rPANC1), cells were transfected with the lentiviral vectors pCT-CD63-GFP and pCT-CD63-RFP (pCMsV, exosome/secretory, CD63 tetraspanin tag, System Biosciences, CYTO120-PA-1, CYTO120R-PA-1), respectively, as reported previously [21
]. Supernatants containing lentiviral particles obtained from the gPANC1 cell line and transduced with pCT-CD63-GFP were filtered through a 0.6-μm filter and then applied for EV isolation.
Microscopic images were acquired using the EVOS system (Thermo Fisher Inc., Waltham, MA, USA) to confirm cell conditions and obtain fluorescent images. Excitation wavelengths of 488 nm and 561 nm were used to detect GFP fluorescence (gPANC1) and RFP fluorescence (rPANC1), respectively.
2.4. Cell Viability
According to the manufacturer’s instructions, cell viability was analyzed using the Superoxide Dismutase (SOD) Assay Kit-WST (FUJIFILM WAKO Pure Chemical Corporation, Osaka, Japan. S311). In brief, WST reagent (10 µL) was added to each well, and the sample was incubated for 45 min at 37 °C. Absorbance was measured at 450 nm (A450) and 620 nm (A620). The value obtained by subtracting A620 from A450 was used as the viable cell number.
2.5. Nanoparticle Tracking Analysis
Nanoparticle tracking analysis was used to visualize and measure the size of small particles (10–1000 nm) in the suspension based on Brownian motion analysis from a video sequence. We used the NanoSight system (NS300, Malvern Instruments Ltd., Malvern, UK) to detect EVs. The measurements were performed at least thrice; the mean value and standard error were calculated and are represented in the graph.
2.6. EV Isolation
EVs were isolated using the MagCapture™ Exosome Isolation Kit PS 293-77601 (Fujifilm Wako Pure Chemical Corporation, Osaka, Japan) as described previously [22
2.7. Whole-Cell Protein Extracts
Cells were lysed with radioimmunoprecipitation (RIPA) buffer (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) composed of 50 mM Tris-HCl (pH 7.4), 1% NP-40, 0.5% sodium deoxycholate (SDC), 0.1% sodium dodecyl sulfate (SDS), 150 mM NaCl, and a protease inhibitor cocktail (FUJIFILM WAKO Pure Chemical Corporation, Osaka, Japan). Supernatants obtained after centrifugation at 15,000× g for 10 min at 4 °C were used as whole-cell proteins.
2.8. Measurement of Glucose, Lactate, and Ammonium Ion Concentrations
The culture supernatant was filtered at each growth phase through a 0.22 μm membrane and directly analyzed with a BioProfile FLEX2 analyzer following the manufacturer’s instructions (Nova Biomedical, Inc., Waltham, MA, USA). The measurements were performed at least thrice; the mean value and standard error were calculated and are represented in the graph.
2.9. Protein Assay and Western Blotting
The protein was resuspended in 5× RIPA buffer (125 mM Tris-HCl, 750 nM NaCl, 5% NP-40, 5% SDC, and 0.5% SDS), after which the suspension was sonicated for 5 min and incubated for 15 min on ice. Protein extract concentrations were measured using Coomassie Protein Assay Reagents (Pierce, Thermo Fisher Scientific, Waltham, MA, USA). A 20 μg aliquot of protein was separated by SDS-polyacrylamide gel electrophoresis and analyzed by Western blotting to detect the proteins with anti-flotillin-1 (Cell Signaling Technology 13174), anti-CD-9 (CosmoBio Japan SHI-EXO-M01), anti-CD-81 (CosmoBio Japan, SHI-EXO-M03), and anti-CD-63 (BD-BioScience, Franklin Lakes, NJ, USA, 556019), followed by incubation with the secondary antibodies goat anti-mouse IgG (Thermo Fisher Inc., #31430), and anti-rabbit IgG (Thermo Fisher Inc., #31460). Electroblotted membranes (Amersham) were blocked with 5% bovine serum albumin (BSA). The expression of β-actin was used as a reference for the amount of protein in each cell sample. The signals were measured using enhanced chemiluminescence.
2.10. Silver Staining
The volume of proteins that passed through the filter was determined using a silver staining kit (FUJIFILM Wako Pure Chemical Inc., Osaka, Japan) according to the manufacturer’s instructions.
Originally, our aim was to investigate the interactions that occur between cancer cells and normal cells. We believe that the results of this paper will be useful not only for biological researchers who use HTCPs but also for future engineering researchers who develop microfluidic devices.
Microfluidic and co-culture systems have a high affinity. The co-culture method has been classically used for studying cell–cell interactions, either with a mixture of cells [23
] or by adding the supernatant of one cell type to the culture of another cell type [25
]. A co-culture vessel with upper and lower chambers is frequently used to prevent cells from mixing [27
] while allowing examination of the impact that groups of cells have on each other.
However, various new cell culture methods have been proposed in recent years to overcome the limitations of these traditional approaches, which are summarized in Figure S9
(modified and permitted from Shimasaki et al.) [1
Co-culture methods can be classified according to the presence or absence of separation material and structural characteristics. The traditional and conventional methods use a filter and upper and lower chambers, forming a bucket-in-a-bucket structure, resulting in a vertical configuration (VTCP). Even today, VTCPs are widely used for co-culture. However, as no other effective co-culture vessels have been developed, no comparison has been made in terms of their performance.
As summarized in Figure S10
(modified from Shimasaki et al. [1
] with permission), the major advantages of the HTCP include the ability to confirm the conditions of various cell types simultaneously, use the same material on both sides of the filter, use high-density co-cultures with no detrimental effects, and use multiple types of filters.
In this study, our experiments revealed that HTCP has a high co-culture efficiency. One reason for the excellent co-culture effect is the ability to control the volume ratio of the containers and the positional relationship between the vessels and the filter.
The structure of VTCP in which the “buckets” are stacked means that the volume ratio of the culture fluid does not become 1:1, which results in a large dilution effect. Achieving the same volume ratio is desirable for the effective co-culture of each cell type (Figure 9
a). The VTCP systems also allow placing cells directly onto the filter (Figure 9
b), meaning that high cell densities lead to pore blockage, preventing the exchange of material through the pores.
Therefore, the decrease in the co-culture effect that can occur in cell cultures over long periods is not visually apparent. Some researchers have argued that in VTCPs, substances that are secreted by the upper cells pass through the filter to affect the lower cells. However, this could not be tested because there was no container with which the efficiency could be compared. The development of the HTCP has allowed such verification to be performed. The better performance of the HTCP is not surprising because the problems associated with filter blockage at high cell densities are bound to have deleterious effects on the results obtained with VTCPs. As the substance moves to an equilibrium concentration, a 1:1 volume ratio between containers allows each substance to influence the other effectively. In Brownian motion, liquid factors such as EVs also pass through the filter away from the cell, meaning that it does not matter if the filter and cells are at a distance.
The track-etched membrane filter, which can be regarded as a microfluidic channel, is an important component. Both the VTCP and HTCP connect two vessels through track-etched membrane filters, which is the same mechanism as that used in microfluidic channels. Although the vessel is not a microstructure in terms of size, the mass transfer characteristics obtained in this study will be useful for other microstructures. We selected a 0.03 μm filter for restricting the passage of exosomes and a 0.6 μm filter for allowing the passage of exosomes. The experimental results showed that the size of the filter used could control the passage of exosomes. It was also found that small molecules such as glucose, lactic acid, and NH4+
can traverse the filer rapidly, even when pore sizes of 0.03 μm were used. The ammonia in the medium was also observed to migrate rapidly in the multi-coupled experiments, but the concentration gradient was maintained when ammonia was continuously secreted by the cells. This gradient means that the degree to which a cell affects other cells can be controlled by considering the location of the wells in a multiconnected experiment. In addition, as shown in the results for the 0.03 µm filter for KLM in Figure S7
, the fact that it took 72 h to equilibrate substances of extremely small size such as NH4+
and that the results were the same in multiple experiments indicates that the filter sandwiching mechanism that uses the O-ring in HTCP works well and is made with precision. This can be assumed because any leakage from a source other than the nano-sized filter pores will allow micro materials such as NH4+
to reach equilibrium in a very short time.
As shown in Figure 10
, when the cells were placed in the left container (Q), the concentration of the substance from the cells increased in Q. When the cells were placed in the middle container (U), the concentrations in both containers (T and V) were the same. When the cells were placed in the second container from the left (X), W = Y did not occur because Y was diluted by the influence of Z, resulting in W > Y > Z.
In the relationship between concentration and time, the concentration of each container was represented by an approximate curve, with a correlation coefficient of approximately 1. These results indicate that metabolites and exosomes were not secreted randomly. They also show that the secreted material undergoes Brownian motion and movement due to osmotic disparity, moving precisely through the pores of the filter to the neighboring vessel. The degree of influence can change depending on the position of the cells in the connected plates. These results may also be applicable to sensor technology via filters and microfluidic channels. It was very interesting that the concentration of extracellular vesicles also changed along the approximate curve. This means that extracellular vesicles migrate according to the theoretical concentration curve even when the vessels are separated and filtered. In other words, the combination of horizontal co-culture vessels and the track-etched membrane made it possible to control the co-culture effect.
Co-culture is a powerful tool to observe the intrinsic cell–cell interaction of extracellular vesicles. To understand the reason for this, we need to know about exosome extraction techniques and research methods.
Several techniques have been proposed for the extraction of EVs [18
] and the addition of collected EVs to other cells to confirm their effects. However, such artificial administration of EVs cannot be performed in vivo. Moreover, these extraction methods can only recover specific EV populations because of the nature of the approach. Techniques such as polymer-based precipitation [30
], immunoaffinity capture using surface markers [31
], purification by flow cytometry, and size-exclusion chromatography [32
] have all been used to recover specific EVs. Among these methods, ultracentrifugation is the standard approach [33
] because it can recover a particular population of EVs. However, the details of EVs have not yet been described, which means that experimental methods that can assess the precise nature and characteristics of exosomes are required. Indeed, information on the characteristics of exosomes remains largely unknown, although proteins and nucleic acids appear to be involved. Exosomes include features such as lipid bilayer membranes, tetraspanins such as CD9 and CD63, membrane transport proteins, and adhesion molecules such as integrins.
Similarly, when extracting exosomes, only the effects of a particular population can be examined. The addition of large numbers of EVs may be an artificial condition that would not otherwise occur, making it difficult to reproduce the natural phenomena that occur in cells. Although experiments in which EVs are extracted and administered are useful, this is also a one-sided experiment. Some phenomena occur only when the cells communicate with each other. Because a living cell has unique mechanisms that are akin to a personal identification number, only experiments that allow such interaction can decrypt a biological identification number. In this regard, the co-culture technique is useful in that it can reproduce natural phenomena without requiring the extraction of EVs. The use of co-culturing technologies is necessary to examine the essential interactions of EVs.
To date, many effects of EVs produced by cancer cells on normal cells, or vice versa, have been reported [26
]. However, the simultaneous analysis of these effects has been limited owing to the lack of suitable techniques.
Using the HTCP experimental system, it is possible to conduct experiments with different degrees of influence from other cells in other wells in the same experiment. For example, to confirm the effect of cancer cells on normal cells, we can examine the difference between Y and Z in the experimental system, as shown in Figure S11a
. To compare the degree of influence exerted by the two different types of cells, the experiment shown in Figure S11b
is recommended. The difference in substance levels between ‘X’ and ‘Y’ occurred due to the difference in the degree of influence of ‘W’ and ‘Z.’
Thus, horizontal co-culture systems offer a useful tool to unravel the characteristics of cells. We believe that our newly designed HTCP system can help promote EV research methodologies and co-culture techniques. As shown in Figure 9
and Figure S11
, HTCP can control the effects of co-culture by selecting the location of the cells.