An Efficient Intestinal Organoid System of Direct Sorting to Evaluate Stem Cell Competition in Vitro

Stem cell competition could shed light on the tissue-based quality control mechanism that prevents carcinogenesis. To quantitatively evaluate stem cell competition in vitro, we developed a two-color intestinal organoid forming system. First, we improved a protocol of culturing organoids from intestinal leucine-rich-repeat containing G-protein-coupled receptor 5 (Lgr5)- enhanced green fluorescent protein (EGFP)high stem cells directly sorted on Matrigel without embedding. The organoid-forming potential (OFP) was 25% of Lgr5-EGFPhigh cells sorted at one cell per well. Using this culture protocol with lineage tracing, we established a two-color organoid culture system by mixing stem cells expressing different fluorescent colors. To analyze stem cell competition, two-color organoids were formed by mixing X-ray-irradiated and non-irradiated intestinal stem cells. In the two-color organoids, irradiated stem cells exhibited a growth disadvantage, although the OFP of irradiated cells alone did not decrease significantly from that of non-irradiated cells. These results suggest that stem cell competition can be evaluated quantitively in vitro using our new system.


Results
Improvement of organoid-forming efficiency (OFE) by optimizing the culturing medium contents. We established a high-efficiency organoid culture protocol that could generate an organoid from a single Lgr5-EGFP high stem cell by direct sorting into medium containing Matrigel in flat-bottomed plates (Fig. 1A). In this method, it was not necessary to concentrate intestinal stem cells by centrifugation after sorting, then mixing and embedding them in Matrigel. To evaluate the organoid forming capacity of stem cells, the OFE was calculated as a percentage of the number of organoids per number of plated stem cells (Fig. 1B).
The OFE differed greatly depending on the type of medium and supplements provided (Fig. 2). The highest OFE was obtained using IntestiCult without conditioned media containing Wnt3a and/or R-spondin1, which are factors secreted by Paneth cells, although Noggin and epidermal growth factor were added to IntestiCult as described in Methods. The organoids showed fine shapes with budding. In 20% Wnt3a conditioned medium,  Definition of organoid-forming potential (OFP) for evaluating the potential of single stem cells to form organoids. The OFE decreased with an increasing number of plated stem cells per well (Fig. 3A) because many organoid-initiating stem cells contacted each other and created only a single organoid under high density conditions. Additionally, too many cells in a well inhibited cell proliferation; therefore, organoid areas and budding rates also decreased with an increasing concentration of cells (Fig. 3B,C). These results suggest that this method cannot assess the OFE and growth rate accurately when numerous stem cells are sorted into 10% Matrigel-containing medium with flat-bottomed 96-well plates because they are greatly affected by the concentration of cells. Thus, we constructed a "one cell/well direct sorting" method with 1% Matrigel using plates with V-shaped wells. For the one cell/well direct sorting method, organoid medium (50 μL) was added into each well of a 96-well plate (Fig. 4A,B). Then, Lgr5-EGFP high cells were plated into the wells (one cell/well). To evaluate the organoid forming capacity of stem cells, the OFE (%) was calculated as the number of organoids per number of plated stem cells when a large number of stem cells were sorted into a well. We newly defined the OFP (%) as a percentage of the number of organoids per number of plated stem cells when one cell was sorted into a well to avoid confusion between the two organoid formation methods (Fig. 1B). The OFPs of Lgr5-EGFP high , Lgr5-EGFP low , and Lgr5-EGFP neg were 25, 9, and 3%, respectively (Fig. 4C,D).
Quantitative analysis for mixed organoids derived from two different stem-cell populations. The use of one organoid per well simplifies observations of organoid growth by eliminating the potential effects of other organoids. Under this condition, it is important to control the number of stem cells per well accurately by a direct sorting method to assess stem cell competition because one organoid can be formed from multiple stem cells in a well (Fig. 5A). The growth of organoids accelerated as an increasing number of stem cells were plated under the one organoid/well condition (Fig. 5B,C). Organoids were formed in about 20-30% wells www.nature.com/scientificreports www.nature.com/scientificreports/ for one stem cell/well sorting; in other words, the OFP was 20-30% (Figs. 4D and 6B). To confirm that few stem cells can grow in a mixed organoid derived from two different stem cell populations, we evaluated OFP'(10) as the number of organoids per well with 10 stem cells/well sorting (Fig. 1B). In this condition, the OFP'(10) was 80% (Fig. 6C). This method was applicable to evaluate organoid formation and cell competition between different types of cells. For example, stem and Paneth cells, duodenum and colon stem cells, irradiated and non-irradiated stem cells, and color-traced and uncolored stem cells.
We crossed Lgr5-EGFP-IRES-Cre ERT2 knock-in and ROSA26-tdTomato (LRT) mice, and treated them with or without 4-hydroxytamoxifen (4-OHT), to determine the cell origin and whether or not they were stem cells (Figs. 5D and 6A). This enabled us to form two-color organoids derived from different colors of stem cells isolated from different mice. Following stem cell proliferation, a disappearing population of one-color cells from the organoid was observed during the process of duodenal organoid growth (Fig. 5E). Two-color cells were adjacent and moved for growth (Movie S1). These organoids were isolated without dispase/collagenase treatment to avoid dissipation of fluorescence by cellular damage, and analyzed by fluorescence-activated cell sorting. The occupancy ratio of tdTomato + (tdTom + ) cells in the organoid was calculated with following equation: Results of the analyses of two-color organoids derived from LRT mice injected with 4-OHT or sunflower seed oil (vehicle) showed that non-irradiated duodenal stem cells occupied the organoid in almost the same manner (Fig. 6E). Therefore, we assessed stem cell competition in vitro by forming two-color organoids derived from tdTom + non-irradiated stem cells and tdTom − stem cells immediately after exposure to 1 Gy of X-rays. Although the irradiated stem cells grew roughly equal to non-irradiated stem cells in monoculture ( Fig. 6B-D), irradiated stem cells did not grow well in mixed culture with non-irradiated stem cells (Fig. 6E). The occupancy ratio of tdTom + at plating was considered to be the expected value when the occupancies were equal. Thus, the occupancy ratio of tdTom + at plating was adjusted by the difference of the mean number of Lgr5-EGFP high cells in monocultured organoids (Fig. 6D) as shown in Fig. 6E. www.nature.com/scientificreports www.nature.com/scientificreports/ In another experiment, tdTom − 1 Gy-irradiated and tdTom + 1 Gy-irradiated stem cells were mixed. The result was similar to two-color non-irradiated stem cells, but varied more widely than the non-irradiated group. In the case of a mixed culture of tdTom − non-irradiated and tdTom + 1 Gy-irradiated stem cells, the irradiated cells did not grow well (Fig. S1). Because stronger stem cells tended to be selected by the toxicity of 4-OHT injection (Figs. 6B,D and S2), we also evaluated stem cell competition using LRT mice treated with 4-OHT, and Lgr5-EGFP-IRES-Cre ERT2 knock-in × ROSA26-LSL-LacZ (LRZ) mice treated with 4-OHT (Fig. S3A). Similar to LRT mice with 4-OHT or oil injections, the growth of irradiated stem cells was suppressed by mixed culture with non-irradiated stem cells between LRT and LRZ mice with 4-OHT injection (Fig. S3B,C). In LRZ mice without any injection, irradiated stem cells grew equally, or more than, non-irradiated stem cells in monoculture (Fig. S4).

Discussion
Composition and substrates in the media used for stem cell cultures. Wnt3a is an essential protein for stem cell maintenance, self-renewal, and differentiation, and is produced by Paneth cells. R-spondin1 functions as a Wnt modulator. During organoid formation, R-spondin1 can substitute for Wnt3a 12 . Wnt3a and R-spondin1 conditioned media, however, did not contribute to the highest OFE in our results. Rather, CHIR-99021, a WNT activator, was necessary for highly efficient organoid formation (Fig. 2). Nevertheless, Wnt3a conditioned medium improved the OFE slightly because it contains many unknown factors beyond Wnt3a, and the OFE did not increase by adding R-spondin1 conditioned medium 13 . In another report, Lgr5-EGFP + cells could not form organoids without Paneth cells even though Wnt3a conditioned medium was added to the basal medium 14 . Murine fetal enterospheres responded to Wnt stimulation at establishment, but once established could grow without R-spondin1 in the presence of the natural Wnt antagonist, DKK1, and a porcupine inhibitor that prevents Wnt secretion 15 . Thus, stem cells appear to require the Wnt activator rather than Wnt itself, and R-spondin1 is effective at only the early phase.
Our results demonstrate that Thiazovivin, Y-27632, and CHIR-99021 are more important for organoid formation than Wnt3a and R-spondin1 at the time of stem cell isolation and culture (Fig. 2). This is consistent with the report that medium containing Thiazovivin and CHIR99021 increased the OFE more than medium with Wnt3a and R-spondin1 13 . Moreover, Thiazovivin and Y-27632 are Rho-associated coiled-coil forming kinase (ROCK) inhibitors that prevent anoikis (i.e., apoptosis induced by the lack of cell and matrix attachments). Thus, single stem cells are sensitive to death induced by damage at the time of isolation from crypts.
When intestinal stem cells are cultured in flat-bottomed cell culture plates, organoids cannot form in medium with 1% Matrigel. Thus, cells are generally embedded into 10% of Matrigel. Furthermore, treatment with dispase/ collagenase is necessary to extract the organoids from this medium. In this study, we found that organoids can be OFP. Improvement of the efficiency of stem cell cultures and organoid formation enabled us to assess the OFP by one cell/well direct sorting. This procedure was not reported previously, although one cell-derived organoid formation by a dilution method was reported in 2009 10 . When a large number of cells were sorted into a well, the OFE of Lgr5-EGFP high cells isolated enzymatically from organoids was higher than that of cells isolated from www.nature.com/scientificreports www.nature.com/scientificreports/ tissues. Therefore, the OFP of Lgr5-EGFP high cells can be improved if cell damage caused by tissue isolation or culture stress is further reduced. Other advantages of our new method of one stem cell/well sorting include: • The OFP can be evaluated for a stem cell using information from fluorescence-activated cell sorting about forward scatter (information of cell size), side scatter (information of the cell's internal structure), fixable viability dye, and EGFP. • Single organoids can be followed and observed easily for a long time.
• Organoids are available for mutation analyses by whole genome or exome sequencing because a single stem cell grows into an organoid via clonal proliferation. • Many organoids can be formed by using a small number of mice; this can improve animal welfare.
Recently, functional human gut organoids, including neural cells with peristalsis, were reported 16,17 . However, mouse organoids derived from adult stem cells remain essential for comparing data sets obtained from in vivo and in vitro experiments. For example, in discussing the health effects of ionizing radiation based on in vitro experimental results, extensive data from in vivo animal experiments stored for many years can be used, whereas available data from healthy humans are limited.

Cell competition in organoids as an intestinal stem cell pool. Normal intestinal stem cells compete
stochastically by self-renewal in vivo 4,[18][19][20][21][22] . Our in vitro results (Figs. 6E and S1C) are consistent with these in vivo data. In addition, one characteristic of our method is an Lgr5 tracing system using the stem cell marker, Lgr5-cre, rather than Ahcre ERt , which is used in many studies as a marker of intestinal epithelium cells, including both stem and differentiated cells.
In the case of non-neutral competition between normal and transformed cells having Kras and APC mutations, they have differing levels of fitness to occupy a niche in the crypt 22 . Cells with mutations in ß-catenin are easier to eliminate by normal cells in normal intestinal crypts compared to APC mutated cells 23 . In the intestine of Rosa26-LacZ/Ah-Cre/Ascl2 floxed/floxed mice, all cells are visually recombined except for long-lived Paneth cells at five days post injection. However, all crypt cells are replaced by wild-type cells at 20 days post injection 24 . The rapid reappearance of wild-type epithelium suggests a strong selective pressure favoring the few remaining Ascl2 + epithelial Lgr5 stem cells. In addition, wild-type crypts that escape conditional deletion of Sox9 and contain abundant Paneth cells rapidly replace mutant crypts by crypt fission 25 . Cell competition is observed not only in in vivo studies, but also in in vitro and ex vivo studies. Transformed cells are eliminated from intestinal organoids when they are surrounded by normal cells ex vivo 7,8 . Overall, we considered that cell competition is constantly occurring to maintain homeostasis in the intestine.
High doses of ionizing radiation are well-known as mutagenic. However, competition between irradiated and non-irradiated stem cells has not been well-investigated. Activation of tumor suppressor p53 signaling occurs in mouse intestine after irradiation 26 . In the current study, we established a high efficiency two-color organoid culture system by mixing stem cells expressing different fluorescent colors (Figs. 5 and 6). Using this system, we assessed stem cell competition between X-ray-irradiated and non-irradiated intestinal stem cells plated just after irradiation. The results showed that growth of irradiated stem cells was partially suppressed in the organoid in the presence of non-irradiated stem cells. This suggests that X-ray-irradiated stem cells were outcompeted by non-irradiated cells, at least partly. Moreover, our results are important to consider when evaluating the health risk of low-dose and low-dose rate radiation exposures where irradiated and non-irradiated stem cells co-exist in the stem cell pool. Stem cell competition may act as a tissue-based quality control mechanism to prevent radiation-induced carcinogenesis under low-dose and low-dose rate irradiation conditions.
In vitro culture of single crypt cells. Isolated single stem cells were cultured in organoid medium (Advanced Dulbecco's modified Eagle's/F12 medium with 100 × GlutaMax and 1 M HEPES (ThermoFisher Scientific) or IntestiCult Organoid Growth Medium (mouse) (Stemcell Technologies) containing 1 × penicillin/streptomycin, 1 × N2, 1 × B27, 50 ng/mL murine epidermal growth factor (Life Technologies), 1 mM N-acetylcysteine (Sigma-Aldrich), 100 ng/mL murine Noggin (Peprotech)). IntestiCult Organoid Growth Medium (mouse) described above was used for all experiments except for Fig. 2 where it also contained R-spondin1 and Wnt3a conditioned media. Both YTC, and or 10% Matrigel, were added to the medium immediately before dispensing it into each well of the cell culture plate. For OFE evaluation, stem cells were sorted into 10% Matrigel-containing medium using plates with flat-shaped wells. For "one cell/well direct sorting" or forming mixed organoids, stem cells were sorted into 1% Matrigel-containing medium using plates with V-shaped wells. Thereafter, freshly prepared organoid medium was added every 3-4 days. Cells were cultured in a humidified CO 2 incubator at 37 °C. The number of organoids was counted 5-9 days after plating by visual observation under a phase-contrast microscope with a 4 × objective lens(Olympus IX71/ DP71).Organoid images were taken using an Olympus IX71/ DP71 and Olympus IX71/ Sony α NEX-5R. Sectioning and time-lapse imaging were performed with a BZ-X700 fluorescence microscope (Keyence). Images of single organoids per well were captured by an Olympus IX71/ DP71 microscope. The captured images were analyzed by Image J software (National Institutes of Health).
Irradiation. Immediately after sorting, Lgr5-EGFP high cells were irradiated by an MBR-1520R4 X-ray generator (Hitachi Power Solutions) operated at 150 kV with 20 mA and a 0.5 mm Al plus 0.3 mm Cu filter at a dose rate of 0.6 Gy/min.

Isolation of cells from organoids for analyzing organoid passage. Organoids were dissociated with
TrypLE Express + YTC for 10 min at 37 °C to prepare single cell suspensions.
Statistical analyses. Data are presented as means ± standard deviation of the mean. Differences between two groups were determined by Student's t test using Microsoft Excel and GraphPad Prism 8.2, The Tukey-Kramer test was used for comparison of multiple groups with GraphPad Prism 8.2. A P value less than 0.05 was considered significant.