Fabrication of tissue‐engineered cell sheets by automated cell culture equipment

Abstract Most cells for regenerative medicine are currently cultured manually. In order to promote the widespread use of regenerative medicine, it will be necessary to develop automated culture techniques so that cells can be produced in greater quantities at lower cost and with more stable quality. In the field of regenerative medicine technology, cell sheet therapy is an effective tissue engineering technique whereby cells can be grafted by attaching them to a target site. We have developed automated cell culture equipment to promote the use of this cell sheet regenerative treatment. This equipment features a fully closed culture vessel and circuit system that avoids contamination with bacteria and the like from the external environment, and it was designed to allow 10 cell sheets to be simultaneously cultured in parallel. We used this equipment to fabricate 50 sheets of human oral mucosal epithelial cells in five automated culture tests in this trial. By analyzing these sheets, we confirmed that 49 of the 50 sheets satisfied the quality standards of clinical research. To compare the characteristics of automatically fabricated cell sheets with those of manually fabricated cell sheets, we performed histological analyses using immunostaining and transmission electron microscopy. The results confirmed that cell sheets fabricated with the automated cell culture are differentiated in the same way as cultures fabricated manually.

maintaining a high level of cleanliness to avoid bacterial contamination. To make regenerative medicine more available, it is essential to facilitate mass production, reduce costs, and stabilize the quality of cell and tissue production. It is expected that automated cell culture techniques will be developed by combining advanced technologies.
There are two types of automated cell culture system: one is a sealed-chamber culture system using a robotic arm; the other is a sealed-vessel culture system (Kino-oka & Taya, 2009). The sealedchamber culture system requires sterilization of the entire interior of the device during batch changes. On the other hand, the sealed-vessel culture system only has a small volume that requires sterilization. For this reason, we chose a sealed-vessel culture system and developed closed-system automated cell culture equipment that has excellent sterility qualities. The main advantage of this equipment is its closed culture vessel/circuit system that completely avoids contamination by bacteria and the like from the external environment (T. Kobayashi, Kan, Nishida, Yamato, & Okano, 2013;Matsumoto et al., 2019;Nakajima et al., 2015;Shu et al., 2016).
After cells have been expanded on a culture dish, they are usually treated with an enzyme such as trypsin or dispase so they can be detached from the culture dish. However, since enzymes can destroy adhesive proteins and important cell membrane proteins, this process could cause drastic changes to the structure and function of the cells.
In contrast, a cell sheet cultured on a dish with a temperatureresponsive surface can be harvested from the culture dish without using any enzymes and without impairing its structure and function simply by reducing the temperature from 37°C to 20°C (Okano, Yamada, Sakai, & Sakurai, 1993;Yamada et al., 1990). Since this means the cell sheet can be harvested intact together with the intra-cellular binding protein and the adhesive protein, it can be transplanted very efficiently, and has excellent therapeutic effects. This technique has a very wide range of applications, and can be used to fabricate cell sheets from many types of cell, including corneal epithelial cells, oral mucosal epithelial cells, fibroblasts, and nasal mucosal cells Kanzaki et al., 2008;Nishida et al., 2004;Yamamoto et al., 2017). Myoblast cell sheets have been put to practical use, and clinical trials have also been conducted on other cell sheets including chondrocytes (Sato, Yamato, Hamahashi, Okano, & Mochida, 2014;Sawa et al., 2015). Clinical studies in Japan and Sweden have confirmed that cell sheets of the oral mucosal epithelial cells targeted in this study are effective at preventing stenosis following the excision of early-stage esophageal cancer; clinical trials are currently under way in Japan (S. Kobayashi et al., 2014;Ohki et al., 2012).
Using this automated cell culture equipment, we fabricated 50 sheets of oral mucosal epithelial cells in five automated culture tests in this study, and we evaluated the quality of these sheets according to the cell sheet quality standards for clinical research that apply to the use of cell sheets fabricated manually. As a result, we confirmed that 49 out of the 50 cell sheets satisfied seven standard criteria including cell morphology, structure, cell number, cell viability, cell purity, and cleanliness. To see if the automatically fabricated cell sheets are equivalent to manually fabricated ones, we performed histological observations of oral mucosal epithelial cell sheets by means of immunostaining and transmission electron microscopy, from which we confirmed that the automatically fabricated cell sheets had differentiated into basal, spinous, and granule layers in the same way as manually fabricated cell sheets.

| Cell culture
For the automated culture test, we used human oral mucosal epithelial cells manufactured by CELLnTEC (Bern, Switzerland), Lot numbers ES1305033 and MC1507246. These cells were passaged and expanded to prepare the number of cells required for testing. As the culture medium for growth, we used a special culture medium provided with the cell sample for Lot No. ES1305033, while we used an epithelial cell culture medium produced by CELLnTEC for Lot No. MC1507246. A 1% mixture of antibiotics (Nacalai Tesque, Kyoto, Japan) was added to each culture medium. For passaging, we used 0.25% trypsin, 1 mM ethylenediaminetetraacetic acid (EDTA) solution (Nacalai Tesque), and a trypsin inhibitor solution (derived from 0.1% soy beans) (DS Pharma Biomedical, Osaka, Japan). The cells were seeded at 0.4 × 10 4 cells/cm 2 , cultured for 5 days, passaged, and then cultured for a further 5 days.
For the cell sheet culture, we used a keratinocyte medium (KCM).
KCM is a culture medium made from a 3:1 mixture of Dulbecco's modified Eagle medium (DMEM; Merck, Darmstadt, Germany) and F12 (Merck), to which 5% fetal bovine serum (Biowest, Nuaillé, France), 1% mixture of antibiotics (Nacalai Tesque), 1 nM cholera toxin (Wako Pure Chemicals, Osaka, Japan), 2 nM triiodothyronine (Wako Pure Chemicals), 5 μg/ml Insulin-Transferrin-Selenium solution (Thermo Fisher Scientific, Waltham, MA, USA), 5 μg/ml transferrin (Thermo Fisher Scientific), 10 ng/ml epidermal growth factor (Proteinexpress, Chiba, Japan), and 0.4 μg/ml hydrocortisone (Wako Pure Chemicals) are added. In order to fabricate the cell sheets, we used a temperature-responsive insert manufactured by CellSeed (CellSeed, Tokyo, Japan). After detaching the cultured cells, a cell suspension was prepared in KCM culture medium with a cell concentration of 2.2-2.8 × 10 5 /ml. In the automated cell culture, a suspension of cells was dispensed into seeding bottles with 8.0-10.0 × 10 4 cells/cm 2 per 1.5 ml, and 3.4-4.2 × 10 5 cells per insert, after which the automated cell culture was started. For the manual culture used as a control, we set the inserts into a 6-well plate, and seeded the inserts with a pipette from a cell suspension. The culture medium was replaced on the 3rd, 5th, 7th, 8th, 9th, 10th, and 11th days. On the 12th day, the culture was terminated and the cell sheets were harvested.

| Harvest of cell sheets and measurement of cell numbers/viability
When the culture had been completed, the culture vessels were treated in a 5% CO 2 incubator at 20°C for 30 min. After that, the cells were washed three times with Hank's balanced salt solution (HBSS; Merck), and the cell sheets were harvested under a stereoscopic microscope. The cell sheets were inspected visually. Then, they were placed in a tube containing a solution of 0.25% trypsin and 1 mM ethylenediaminetetraacetic acid (Nacalai Tesque), cut finely with scissors, and treated at 37°C for 15 min. After dispersing the cells by pipetting, KCM was added and the mixture was passed through a 40 μm cell strainer (BD Biosciences, #352340). The tube was then centrifuged for 5 min at 1,000 rpm at room temperature, resuspended, and stained with trypan blue so that the number of cells and the cell viability could be ascertained.

| Measurement of cell purity
The cell suspensions were centrifuged and washed once with 3 ml of a flow cytometer washing solution (BD Bioscience, Franklin Lakes, NJ, USA). After centrifuging at 1,500 rpm for 5 min at 4°C, the samples were incubated for 20 min at 4°C in a fixative (BD Bioscience). After fixation, they were centrifuged at 3,500 rpm for 10 min at 4°C. After washing, one sample was reacted for 1 hr at room temperature with fluorescein isothiocyanate (FITC)-labeled anti-pan-cytokeratin (CK) antibodies (Progen, Heidelberg, Germany), and another was reacted for 1 hr at room temperature with fluorescein isothiocyanate-labeled anti-mouse immunoglobulin G2a (IgG2a) antibodies (SC-2856 Santa Cruz Biotechnology, Santa Cruz, CA, USA). After these reactions, the cells were washed twice in a washing solution and resuspended to measure the pan-CK positive ratio by flow cytometry.

| Transmission electron microscopy (TEM)
The cell sheets were fixed with 10% neutral buffered formalin solution (Wako Pure Chemicals). Prefixing was performed with 2% glutaraldehyde/0.1 M phosphate buffer at 4°C, followed by postfixing with 2% aqueous osmium tetroxide at 4°C for 3 hr and dehydration processing with 50%-100% ethanol. This was followed by thermal polymerization embedding using EPN812 → preparation of ultra-thin 90 nm sections using an ultra-microtome → staining with a two-part dye consisting of lead/uranyl acetate. Observations were performed using a transmission electron microscope (TEM) with an acceleration voltage of 80 kV.

| Device configuration
An overview of the automated cell culture equipment (ACE3) is shown in Figure 1a,b. The equipment is relatively compact (1,700 mm × 795 mm × 1,580 mm W × D × H), and can be installed in a cell processing facility. It is easy to culture 10 cell sheets in parallel by installing 10 closed culture vessels inside this equipment. Each of the 10 closed culture vessels has independent circuits to maintain the accuracy of the air and liquid feed rates. The CO 2 concentration in the circuit is maintained at 4.8-4.9%, and the liquid feed accuracy is ±3.2%. The equipment consists mainly of an incubator, a refrigerator, a control box, an uninterruptible power supply, a controller personal computer (PC), and a router PC. The controller PC can be operated remotely via the router PC. The incubator is equipped with a weight sensor, circuit module, and phase contrast microscope. The difference between an open culture system, which uses the conventional culture method, and our closed culture system is shown in Figure 1c. The open culture system requires sterilization of the entire interior of the device during batch changes. On the other hand, the closed culture system requires very little space for cleanliness, and only the closed culture vessel and circuit modules which are sterilized by gamma ray irradiation in advance have to be replaced in this system, making the risk of contamination lower. During culture, the interiors of the closed culture vessel and circuit modules of the ACE3 are filled with a high (95% or more) humidity environment with 5% CO 2 , while the other parts of the equipment interior are filled with the same atmosphere as the outside air. By reducing the humidity inside the main body of the equipment to provide a dry environment, we can reduce the risk of contamination by bacteria and the like, resulting in a lower risk of equipment failure. To keep the interior of the equipment sterile, it also uses an externally driven pinch valve and a peristaltic pump.
The main automated functions of ACE3 are cell seeding, cell culture, gas exchange, medium exchange, and microscopic observation.
The microscopic observations can automatically take photographs of each closed culture vessel from multiple fixed points at a frequency set by the user. The user can manually observe and photograph from arbitrary positions in the closed culture vessels or can remotely control these operations.
Here, we should note that there are two major differences from our previous equipment (T. Kobayashi et al., 2013). First, our previous equipment was a semi-closed system, because the culture vessels were connected to the circuits only at the times of cell seeding and medium exchange. In contrast, it is a completely closed system in ACE3, because the culture vessels were always connected and never disconnected throughout the automated culture. Second, we increased the number of cell sheets that can be cultured at the same time from four to 10, because the transplantation target was expanded from the cornea to the esophagus after endoscopic submucosal dissection. Ten cell sheets are available for inspection the day before transplantation and as spares. Multiple sheets are used when the affected area is large (Ohki et al., 2012).

| Closed culture vessel and circuit modules
The closed culture vessel and circuit modules consist of culture vessels, circuits (tubes), bottles and bags. These are provided as disposable assemblies that can be detached from the equipment and replaced for each culture. This makes it possible to avoid cross contamination between lots. The modules, except for temperatureresponsive cell-culture inserts, are sterilized by gamma irradiation.
The temperature-responsive cell-culture inserts are set in the culture vessels of the sterilized modules on a clean bench before they are set on the equipment. Sterile adapters are used to attach them to the equipment. Figure 2 shows a closed culture vessel and circuit module.  (Figure 2a-2). For cell sheets cultured on the cell culture insert, 5% CO 2 gas cylinder is supplied from the port for the 35 mm dish in the lower layer in order to avoid shaking the medium.
In the fully closed circuit, bottles and bags for supplying and discharging the cell suspension, culture medium, and 5% CO 2 are Difference between open culture system and closed culture system. The incubator area is at the top, and the control area and storage cabinet (including the refrigerator) are at the bottom. The white dotted outline indicates the incubator area where cell culturing is performed at 37°C, but humidified 5% CO 2 is only supplied in the closed culture vessel and circuit surrounded by the white solid line. The inside of the device, including the shaking mechanism, valves, and pumps, is kept dry. The refrigerator at the bottom is where the culture medium and recovered supernatant can be stored. The equipment also includes a sterile welding-cutting device for aseptically removing culture vessels and culture supernatant and control device. The open culture system is the conventional culture method and it means a manual culture and sealed-chamber culture system using robotic arm. Light gray parts indicate spaces that need to be sterilized. The closed culture system that we use requires very little space for cleanliness connected to the culture vessel with tubes, and various sensors for monitoring are attached. The details and an image of the circuit are shown in Figures 2b and S1. The cell suspension, culture medium, and 5% CO 2 are supplied and discharged using a rotary valve mechanism, solenoid valve, and peristaltic pump (Matsumoto et al., 2019).
The culture medium is kept in a refrigerator. When exchanging the medium, the required amount is introduced into a preheating bottle in the incubator and preheated to 37°C before being used to exchange the culture medium. To allow the culture supernatant to be inspected for bacterial contamination, it can be collected in a special bag at any time. The closed culture vessels and culture medium supernatant bags can be aseptically removed by using the installed sterile weldingcutting device. This equipment has a case composed of heat storing materials especially for transferring culture vessels. Once the culture vessels are in the case, they can be carried to the operating room while maintaining sterility and a temperature of 37°C.

| Automated culture tests
To verify the performance of the ACE3, we performed automated culture tests using commercially available human oral mucosal epithelial cells. For these tests, we selected the best lot in which the cells form a cell sheet in a manual culture in advance. We set up an automated cell culture protocol corresponding to the manual culture protocol and seeded the cells automatically in 10 closed culture vessels. After   (Shu et al., 2016). (a-2) Cross-section through a closed culture vessel. The culture dish and temperature-responsive cell culture insert are set in the vessel, and the culture medium and gas are exchanged via the port. (b) Circuit diagram of the closed system circuit. By switching valves, it is possible to pump the culture medium and gas to any location be at least 1 × 10 5 cells per sheet, (4) the cell viability should be at least 70%, (5) the epithelial cell purity (pan-CK positive ratio) should be at least 70%, and the cells must give negative results in (6) sterilization tests and (7) mycoplasma inspections. Table 1 (3) the number of cells, (4) cell viability, and (5) epithelial cell purity were calculated as the mean and standard variation over five automatically fabricated sheets and three manually fabricated sheets as a control. In the 49 sheets that were available for testing, we found that the automatically fabricated cell sheets met all the above criteria. Furthermore, they all tested negative for bacteria and mycoplasma. No tests for bacteria or mycoplasma were performed on the manually fabricated cell sheets.
Phase contrast microscope images acquired during the automated cell culture are shown in Figure 3. Epithelial stem cells and progenitor cells derived from the basal layer that are used as the source cells grow after adhering to the surface of the culture vessel. They become confluent, resulting in a cobblestone appearance, and then form overlapping layers of cells that are differentiated from the epithelial stem cells and progenitor cells. This process results in the formation of a cell sheet . In both the automated and manual cell culture methods, a cobblestone-like morphology was confirmed on the 3rd day of culture and the cells formed layers on the 7th day. At the end of the 12th day of automated cell culture, the culture vessel was aseptically removed from the equipment, after which it was incubated at 20°C for 30 min in a 5% CO 2 before harvesting the cell sheet.
The cell sheet is shown in Figure 4. In both the automated and manual cell culture techniques, it was possible to recover a circular cell sheet without any damage.

| Histological analysis of cell sheets
For the histological analysis, we performed HE staining, immunostaining, and transmission electron microscope observations. The results of HE staining are shown in Figure 5a. In HE staining, the cell nuclei are stained purple with hematoxylin, and the cytoplasm is stained red with eosin. We confirmed that epithelial stem cells and progenitor cells derived from human oral mucosal epithelium had differentiated and formed roughly 3-5 layers in cell sheets formed by both automated and manual cultures. The results for item 1 (multilayer) were obtained from five sheets per test with the automated cell culture, and two sheets per test with the manual culture. For items 3-5, the results show the mean and standard deviation over 5 sheets with the automated cell culture and over 3 sheets with the manual cell culture. For items 6 and 7, we did not perform tests on the manual culture.
b Seeding was performed with 3.4-4.2 × 10 5 cells per insert in the automated cell culture and with 3.4 × 10 5 cells per insert in the manual culture control.
Manual culturing was performed with the inserts set in a 6-well plate. In each test, 10 sheets were fabricated with the automated cell culture, and five sheets were fabricated manually.
The results of immunostaining with keratinocyte 3 (CK3) and p63 antibodies are shown in Figure 5b. CK3 is a marker for mucosal epithelial differentiation, and p63 is a marker that localizes in the nuclei of epithelial stem cells and progenitor cells Nishida et al., 2004;Oie et al., 2010). In both the automated and manual cell cultures, CK3 stained the whole cell sheet dark reddishbrown, thus confirming that CK3 is generally expressed as previously reported. With p63, the nuclei in the basal layer of the cell sheet were stained dark reddish-brown in both the automated and manual cell cultures, and as previously reported, we confirmed that the basal layer  and (4) keratinocytes (Schoop, Mirancea, & Fusenig, 1999;Tobin, 2006  As a result of automated culture tests using human oral mucosal cells, we succeeded in fabricating epithelial cell sheets that satisfy the criteria for cell sheets fabricated manually, and we confirmed that cell sheets fabricated by automated cell culture are histologically equivalent to those fabricated manually. We also succeeded in culturing cell sheets five times in succession and were able to show that the automated cell culture equipment is capable of stable production of cell sheets. Automated cell culture has the advantage that it is As a result, it causes smaller changes in temperature and CO 2 concentration than manual culture, and provides a better culture environment.
This report compared results for an automated culture and those for a manual culture. When the manual culture did not go well, the automated culture did not go well either. The cell sheet fabricated by automated cell culture sometimes did not reach the quality of the manually cultured cell sheet. The automated equipment seeds, supplies gas, and feeds in the medium via dedicated tubes for each closed culture vessel. It is conceivable that the environment of the closed system may cause physical damage to the cells. By performing cell seeding, gas supplying, and medium feeding via tubes in a closed system to minimize physical damage to cells, it should be possible to perform culturing more stably than with manual techniques.
We are also studying the use of this automated culture equipment for the fabrication of cell sheets with different cell types. If we can fabricate cell sheets of various types, then it should be possible to increase the use of cell sheets in regenerative medicine.
Expertise in manual cultures is currently needed for making regenerative medicine products; this is an obstacle to the popularization of regenerative medicine. If automated cell culture techniques can be used to mass produce cell sheets at lower cost with stable quality, more patients could have access to regenerative medicine therapies.