The human colon cancer DLD-1 cell line was obtained from the American Type Culture Collection. DMEM/F12 and RPMI 1640 media were purchased from HyClone (Cytiva). EGF and bFGF were purchased from PeproTech, Inc. Fetal bovine serum (FBS) and bovine serum albumin (BSA) were purchased from Amresco, LLC. Insulin was purchased from MilliporeSigma. SFM supplement factor B27 was purchased from Invitrogen (Thermo Fisher Scientific, Inc.). CD133-phycoerythrin (PE; cat. no. 130-098-826), CD44-FITC (130-095-195) were purchased from Miltenyi Biotec GmbH. Corresponding isotype control antibodies Mouse IgG1κ Isotype Control PE (12-4714-41) and FITC (cat. no. 11-4714-41) were purchased from eBioscience. PrimeScript^™ RT Master Mix kits were purchased from Takara Biotechnology Co., Ltd. Phosphate-buffered saline (PBS), crystal violet dye and trypsin were purchased from Wuhan Boster Biological Technology, Ltd. PCR primers were synthesized by Invitrogen (Thermo Fisher Scientific, Inc.). The BD FACSCanto II Flow Cytometry Instrument was purchased from ACEA Bioscience, Inc. (Agilent Technologies, Inc.), and the PCR thermocycler was purchased from Roche Diagnostics. Sevoflurane was purchased from MilliporeSigma. BALB/C-Nu/Nu mice were purchased from Beijing HFK Bioscience Co., Ltd.

SFM cultures were prepared by combining DMEM/F12 with 10% BSA, 5 mg/ml insulin, B27 cell culture additive (1:50), 100 U/ml streptomycin, 0.1 mg/ml penicillin and variable concentrations of growth factors. A total of nine experimental groups were prepared using different combinations of EGF and bFGF: Group 1 (G1), 5 ng/ml EGF; G2, 5 ng/ml bFGF; G3, 5 ng/ml EGF + 5 ng/ml bFGF; G4, 10 ng/ml EGF; G5, 10 ng/ml bFGF; G6, 10 ng/ml EGF + 10 ng/ml bFGF; G7, 20 ng/ml EGF; G8, 20 ng/ml bFGF; and G9, 20 ng/ml EGF + 20 ng/ml bFGF.

After thawing, DLD-1 cells adhered in a monolayer in RPMI 1640 medium containing 10% FBS. The medium was changed every 2 days. When cells reached the logarithmic growth phase, they were digested with EDTA-containing trypsin for 4 min at 37°C, followed by the addition of 4 ml RPMI-1640 and transfer to a centrifuge tube. After centrifugation at 111.8 × g for 5 min with 4°C, the supernatant was discarded, and cells were resuspended in new medium. Finally, cells were cultivated in a new culture dish according to the number of cells. All cell cultures were maintained in a humidified incubator at 37°C and 5% CO₂.

Adherent DLD-1 cells were cultured to the logarithmic growth phase, and then cultured in DMEM/F12 for 24 h at 37°C. Cultures underwent trypsin digestion for 4 min to create single-cell suspensions in DMEM/F12, followed by cell counting. A volume of 20 ml of the G1-G9 culture media preparations were added to super slip plastic culture flasks, followed by the addition of 3×10⁴ cells to each experimental group flask. Cultures were then incubated with the addition of 2 ml of the corresponding group medium daily at 37°C and 5% CO₂. Follow-up tests were performed at days 10, 20 and 30.

Suspended cell spheroids were cultured to days 10, 20 and 30, and then were subjected trypsin digestion and quantification. Single-cell suspensions were obtained in PBS. Aliquots of 5×10⁵ cells were rinsed twice with PBS, followed by the addition of staining buffer blocking solution (pH 7.2, containing 2% FBS and 0.4% BSA) and storage at 4°C for 1 h. The suspension was centrifuged (251.55 × g for 5 min with 4°C) and the supernatant was discarded. Subsequently, the cells were treated with the addition of 25 µl staining buffer and 1.25 µl of either anti-CD133 antibodies, anti-CD44 antibodies, or a mixture of both anti-CD133 and anti-CD44 antibodies (all 1:50), and protected from light at 4°C for 30 min. Finally, the supernatant was discarded after centrifugation (251.55 × g for 5 min with 4°C). CSCs were then suspended in 500 µl PBS and detected via flow cytometry. The data were analyzed using FlowJo software version 10.0 (BD Bioscience, ACEA Bioscience, Inc.).

RNA was extracted from spheroids with chloroform and precipitated with absolute ethanol at days 10, 20 and 30. mRNA was then reverse-transcribed into cDNA (37°C for 15 min, 85°C for 5 min, 4°C for 5 min). GAPDH was used as an internal reference. PrimeScript^™ RT Master Mix kit and TB Green^® Premix Ex Taq^™ (Tli RNaseH Plus; cat. no. RR420A; both from Takara Biotechnology Co., Ltd.) were used according to the manufacturer's instructions to detect the mRNA expression level in each sample using the 2^−ΔΔCq method (31). PCR was executed according to the follows: 1 cycle of pre-incubation at 95°C for 10 min, followed by 40 cycles of amplification for 10 sec at 95°C, 10 sec at 60°C, 10 sec at 72°C, then 1 cycle of melting at 95°C for 10 sec, at 65°C for 60 sec, at 97°C for 1 sec, and cooling at 50°C for 30 sec. The primer sequences used are presented in Table I.

A total of 100 spheroid cells were collected from each group and cultured in 24-well plates with 1 ml of the corresponding group medium for 14 days at 37°C. The number of spheroids in each group was then counted in three fields of view under a light microscope at ×100 magnification for statistical analysis. Count three times for each sample.

Spheroid cells were trypsinized, suspended and counted. A total of 500 spheroid cells were collected from each group and seeded in 6-well plates with 5 ml RPMI-1640 medium containing 10% FBS for 7 days at 37°C. Cells were fixed with 75% ethanol solution at 37°C for 15 min and stained crystal violet alcohol solution (Shangbao Biological Company) at 37°C for 30 min. Colonies (>3 cells) counted manually under a light microscope at 40× magnification for statistical analysis. The experiments was performed three times.

All animal experiments were approved by the Ethical Committee of Huazhong University of Science and Technology (approval no. S255). A total of 81 female nude mice (age, 4–6 weeks; weight, 15–21 g) were randomly divided into 27 groups (n=3 mice/group), and housed in a specific pathogen-free environment (18–26°C; humidity, 40–50%, 12/12-h light/dark cycle and free access to food and water) for 1 week to facilitate adaption to the environment. Spheroid cells were trypsinized, collected and counted. Suspensions of 1×10⁵ cells in 200 µl physiological saline were prepared for later use. According to the grouping of nude mice, the corresponding cells were injected subcutaneously using a 1-ml syringe, and the injection tracks were clamped for 20 sec to prevent leakage. Nude mice were observed daily to evaluate their general condition and changes in body weight. After 30 days, sevoflurane (5%) was used to anesthetize the nude mice via breathing inhalation, then nude mice were sacrificed via cervical dislocation; after confirming the onset of rigor mortis, the tumors were harvested and measured, and tumor volumes were calculated using the following formula: Volume=length × width² × 0.5.

SPSS 25.0 (IBM Corp) was used for statistical analysis, and GraphPad Prism 7 (GraphPad Software, Inc.) was used for graphing of results. The normal distributions of data were presented as the mean ± SD, and one-way analysis of variance and Tukey's post hoc tests were used to compare statistical differences between each group. P<0.05 was considered to indicate a statistically significant difference. All experiments were performed ≥3 times.

Adherent cells were arranged individually and grew close to the bottom of the culture flasks. Cells in each group began to assume a spheroid morphology on day 3, and tended to be suspended in the culture medium as a single oval unit at day 7, after which the volume of cell spheroids gradually increased. There were no intergroup differences in spheroid volume (Fig. 1). The suspension cultures started to produce stable spheroids after the seventh day of culture.

Flow cytometry revealed the percentages of CD133⁺, CD44⁺ and CD133⁺CD44⁺ double-positive spheroid cells of the nine groups at days 10, 20 and 30 of culture (Table II). The percentages of CD133⁺ and CD44⁺ cells were highest in G9 at 30 days, and were significantly elevated compared with the other combinations (F=123.554 and 99.528, respectively, P<0.001). The percentage of CD133⁺ cells in the G9 group at day 30 was 2.558±0.085%, which was similar to the results for G7 at day 10, 20 and 30, and G8 at day 10 and 30 (P>0.999, P=0.686, P>0.999, P=0.282 and 0.818, respectively); however, the percentage of CD44⁺ cells was 72.990±2.188%, which was higher compared with in any other group. The percentage of CD133⁺CD44⁺ double-positive spheroid cells in G9 at day 30 was second only to G3 at day 10 and G6 at day 20 and G9 at day 30, and G9 at day 30 was significantly higher compared with other combinations (except G3 at day 10 and G6 at day 20) (F=57.897, P<0.001). The difference between G3 at day 10 and G6 at day 20 was not statistically significant (P=0.574). These results suggested that the G9 suspension-cultured spheroids at 30 days of culture exhibited more prominent CSC characteristics than those grown under the other experimental conditions.

RT-qPCR detected the mRNA expression of stemness genes [Krüppel-like factor 4 (KLF4), leucine-rich repeat-containing G protein-coupled receptor 5 (Lgr5), CD44, CD133], epithelial-mesenchymal transition (EMT) genes (E-cadherin, Vimentin) and Wnt/β-catenin pathway genes (Wnt-3a; Fig. 2). The mRNA expression levels of KLF4 (Fig. 2A), Lgr5 (Fig. 2B), CD44 (Fig. 2C), CD133 (Fig. 2D), Vimentin (Fig. 2E) and Wnt-3a (Fig. 2F) were significantly altered between the culture groups (F=22.682, 25.401, 3.272, 7.852, 13.331 and 17.445, respectively, P<0.001). with the highest expression values most commonly observed in G9 at day 30. Comparing the spheroids of each group with G9 at day 30, CD44 expression in G2 at day 20, G3 at days 20 and 30, G5 at day 20, G7 at day 30, and G9 at days 10 and 20 was similar (P=0.218, 0.830, 0.069, 0.162, 0.123, 0.060 and 0.133, respectively; Fig. 2C). There were no significant differences in CD133 expression between G9 at day 30 and G3 at day 20, G8 at days 20, G8 at days 30 (P=0.876, 0.304 and 0.664, respectively; Fig. 2D). There were no significant differences in Vimentin expression between G3 at day 30 and G9 at day 30 (P=0.350; Fig. 2E). Wnt-3a mRNA expression was similar between G9 at day 30 and G2, 3, 4 and 6 at day 20 (P=0.339, 0.235, 0.09 and 0.198; Fig. 2F). E-cadherin expression was lowest in the G9 group at day 30 (F=10.851; P<0.001), but was not significantly different compared with 15 other culture conditions (Fig. 2G). Increased expression of Wnt-3a indicated activation of the Wnt/β-catenin signaling pathway, whereas high expression of Vimentin and low expression of E-cadherin indicated that EMT had occurred, and was accompanied with substantially increased expression of stemness genes. In summary, these results indicated that cell spheroids exhibited the characteristics of CSCs. Comprehensive analysis of the data showed that G9 at day 30 exhibited the most upregulated expression of stemness genes and the lowest expression of E-cadherin, indicative of the most characteristic gene expression of CSCs (32,33).

Self-renewal ability is a landmark feature of CSCs, which is closely associated with cancer occurrence, development, recurrence and treatment resistance (34). The purpose of the sphere-forming assay is to test the self-renewal ability of spheroid cells. The sphere count was significantly different between the different conditions (F=19.147, P<0.001; Fig. 2H), with the highest yield of spheroids in the G9 condition from day 30. Of note, the difference between G6 and G9 from day 30 was not statistically significant (P=0.116). These results suggested that G9 spheroid cells cultured for 30 days exhibited the most efficient self-renewal capacity.

To test in vitro tumorigenicity, colony formation assays were performed for each experimental group (Fig. S1). Spheroid cells can redifferentiate into adherent cells. The number of colonies was significantly different between the different conditions (F=60.767, P<0.001), with the largest number of colonies observed for G9 cells cultured for 30 days. However, there was no statistically significant difference between G9 from day 30 compared with G6 from day 30 or G7 from day 20 (P=0.119 and 0.131; Fig. 3). Therefore, the results suggested that G9 spheroid cells cultured for 30 days exhibited the most prominent in vitro tumorigenic capacity.

In vivo tumorigenesis was analyzed by using the tumor cell xenograft formation assay (Fig. S2). G9 at day 30 promoted the largest tumor volume, which was significantly increased compared with other combinations (F=12.539, P<0.001; Fig. 4). However, there were no statistically significant differences between G9 from day 30 compared with G6 from days 20 and 30, G7 from day 30 or G9 from day 20 (P=0.892, 0.940. 0.053 and 0.997, respectively). It is therefore proposed that G9 spheroid cells cultured for 30 days exhibited the most potent in vivo tumorigenicity.