This article is mentioned in:

Introduction

Colorectal cancer (CRC) is the third-most common cancer, accounting for 10.2% of new diagnoses as well as the second-leading cause of cancer-related mortality, accounting for 9.2% worldwide (1,2). Although adjuvant and neo-adjuvant chemotherapy are standard front-line approaches in support of surgery, 5-fluorouracil (5-FU), one of the original fluoropyrimidines, has been considered a mainstay of chemotherapy for CRC (3,4). A recent review (5) summarized the various mechanisms against 5-FU including the alterations in drug transport, changes in the cell cycle, epithelial-to-mesenchymal transition (EMT), and cancer stem cell (CSC) involvement. Certain resistance mechanisms that are 5-FU-specific have also been ascertained in 5-FU-resistant SNU-C5 (SNU-C5/5-FUR) CRC cells as compared with wild type SNU-C5 cells to include the upregulation of cyclooxygenase-2 derived prostaglandin E2 (6) and over-activation of protein kinase B (Akt) (7,8). The differential activation of extracellular signal-regulated protein kinase (ERK) between wild type and 5-FU-resistant CRC cells after yeast extract treatment has been previously suggested by the authors (8). It has also been reported that SNU-C5/5-FUR cells are more susceptible to an aqueous extract of Orostachys japonica A. Berger than wild type SNU-C5 cells via the activation of mitogen-activated protein kinase signaling pathways including ERK and p38 (9).

Although the CSC markers between cancers are not identical (10), CSCs exhibit common characteristics regarding the maintenance of CSC pool, tumorigenesis, metastasis, and treatment resistance and recurrence (11–13). As p21 attenuates Ras- and c-Myc-dependent EMT and CSC-like gene expression in vivo (14), it suppresses the cell cycle and is also used as a biomarker for CSCs (15). Similarly, the Wnt/β-catenin signaling pathway is critical for the regulation of cell proliferation, differentiation and apoptosis during regeneration (16). CSCs are typically determined by the cell surface proteins [cluster of differentiation (CD) 44, and nuclear transcription factors, octamer binding transcription factor-4 (Oct-4), and sex determining region Y-box-2 (Sox-2)] (13,17) as well as ATP-binding cassette super-family G member 2 (ABCG2) transporter (18). Moreover, non-CSCs can acquire a CSC-like phenotype during EMT (19), and EMT-induced cells can form spheres in an anchorage-independent growth environment (20).

The three-dimensional (3D) tumor spheroid formation model has been suggested to be an essential tool for confirming CSC-like features in vitro (21), because it resembles in vivo solid tumors rather than the conventional, two-dimensional (2D) monolayer culture (22). Although various methods and conditions have been proposed, spheroids consist of an external proliferating zone, an internal quiescent zone, and a necrotic core (22). During the spheroid formation process in any environment, it is known that cells initially aggregate and then form compact spheroids via E-cadherin (22). E-cadherin, an epithelial cell to cell adhesion molecule, is inversely correlated with EMT (23). Although E-cadherin is also known to be frequently downregulated with tumorigenesis, cell adherence between cancer cells are disturbed by addition of the soluble fragment (80 kDa) of E-cadherin, leading to malignancy in cancers (24,25).

Growth factors (GFs) have been used in spheroid formation culture methods to maintain the integrity of CSCs (17,22), which is also applied to CRC cells in order to effectively obtain CSCs to date (26,27). Although fetal bovine serum (FBS) supplementation has recently been suggested as an adaptable, efficient, and cost-effective tool to maintain pluripotency in a hepatocellular carcinoma cell (28), it has not been investigated in acquired drug-resistant CRC cells. Therefore, it was aimed to investigate whether spheroid formation culture depend on supplementations is appropriate on acquired 5-FU-resistant SNU-C5/5-FUR CRC cells as compared with wild type SNU-C5 cells. Spheroid formation culture methods with different culture environments supplemented with FBS and GFs, respectively, were used. Accordingly, the feasibility of appropriate spheroid formation culture methods was examined in each of different CRC cell, and the differences between wild type and acquired 5-FU-resistance were revealed.

Materials and methods

Antibodies

The antibodies specific for β-catenin (1:1,000; cat. no. sc-7199), c-Myc (9E10; 1:1,000; cat. no. sc-40), E-cadherin (H-108; 1:1,000; cat. no. sc-7870), epithelial cell adhesion molecule (EpCAM) (c-10; 1:1,000; cat. no. sc-25308), glyceraldehyde 3-phophage dehydrogenase (GAPDH; 1:2,000; cat. no. sc-47724), glycogen synthase kinase-3 beta (GSK-3β; 1:1,000; cat. no. sc-9166), Oct3/4 (C-10; 1:1,000; cat. no. sc-5279), Pan-Ras (C-4; 1:1,000; cat. no. sc-166691), Sox2 (E-4; 1:1,000; cat. no. sc-365823) and Wilms tumor protein [WT1 (6F-H17); 1:1,000; cat. no. sc-81619] were obtained from Santa Cruz Biotechnology, Inc. cAMP response element-binding protein (CREB) [phosphor S133 (E113); 1:1,000; cat. no. ab32096], CD24 [M1/69] (1:1,000; cat. no. ab64064), and CD44 (1:1,000; cat. no. ab157107) were obtained from Abcam. ABCG2 (BCRP1, clone 5D3; cat. no. MAB4155; 1:500) and α-smooth muscle actin (αSMA; 1:2,000; cat. no. A2547) from were obtained MilliporeSigma. CREB (1:1,000; cat. no. CSB-PA005947HA01HU; Cusabio Technology, LLC), fibronectin (1:2,000; cat. no. CL54951AP; Cedarlane Laboratories), p21 (1:1,000; cat. no. 60214-1; Proteintech Group, Inc.) and p90RSK (Ab348; 1:1,000; cat. no. 79-554; Prosci, Inc.) were obtained from the corresponding listed company.

Cell culture

SNU-C5 (Korean Cell Line Bank; Seoul, South Korea) and SNU-C5/5-FUR (Research Center for Resistant Cells; Chosun University, Gwangju, South Korea) cells were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated FBS, 100 U/ml penicillin and 100 mg/ml streptomycin at 37°C (Welgene, Inc.) in a humidified atmosphere with 5% CO2 as previously described (8).

Spheroid formation

96-well plates were covered with poly-2-hydroxyethylmethacylate (cat. no. P3932; Sigma-Aldrich; Merck KGaA) to create an anchorage-independent environment. Cells were cultured with 1% B27 supplement (cat. no. 17504-044), 20 ng/ml epidermal GF (cat. no. PHG0311) and 20 ng/ml basic fibroblast GF (cat. no. 13256029; all from Thermo Fisher Scientific, Inc.) in DMEM/F12 medium (sphere/GF group) (17,21), or in same culture media with FBS (sphere/FBS group) as previously reported (28). Spheroid formation was checked for morphometry on days 7, 14 and 28.

Cell viability assay

The effect of 5-FU on cell viability was evaluated in terms of the reduction of MTT (Amresco, LLC) using a VERSAmax microplate reader (Molecular Devices, LLC) as previously described (10,19). Dissociated cells from 2D monolayer and spheroid-formation cultures with trypsin-EDTA (Welgene, Inc.) were seeded in triplicate wells in 96-well plates (2×103 cells/well), and treated with 5-FU at various concentrations as previously described (22). The number of viable cells was estimated for 3 days after incubation in 2D culture, and for 4 days after incubation in spheroid formation culture methods. The effect of the drug was calculated and compared with untreated (DMSO-treated only) cells using Microsoft Excel (MS Office 2016).

Western blotting

Cells were incubated for 3 days in monolayer culture, and incubated in spheroid formation culture for 4 weeks. Cells were harvested in M-PER mammalian protein extraction reagent (Thermo Fisher Scientific, Inc.) including 1% protease inhibitor cocktail set III (EMD Millipore), 0.5% phosphatase inhibitor cocktail 2 and 0.5% phosphatase inhibitor cocktail 3 (both from Sigma-Aldrich; Merck KGaA). Protein concentration was assessed using BCA protein assay (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions.

The electrophoresis of protein in cell lysates on an TGX Stain-Free FastCast™ Acrylamide Starter kit (Bio-Rad Laboratories, Inc.) using tris/glycine buffer systems (Bio-Rad Laboratories, Inc.) onto PVDF membranes was performed as previously described (8,17). The membranes were first blocked at room temperature with 5% skim milk for 1 h and then incubated with primary antibodies overnight at 4°C. After washing, peroxidase anti-mouse or anti-rabbit IgG antibodies (1:3,000; cat. no. PI-2,000 and PI-1,000, Vector Laboratories, Inc.) were applied for 1 h at room temperature. Next, western lighting chemiluminescence reagent (PerkinElmer, Inc.) was used to detect proteins. The anti-GAPDH antibody was used as a loading control on the stripped membranes. The bands were captured using Azure™ c300 (Azure Biosystems, Inc.) and quantified using the AzureSpot analysis software (version 14.2; Azure Biosystems, Inc.).

Enzyme-linked immunosorbent assay (ELISA)

ELISA was performed according to the manufacturer's instruction (Human E-cadherin SimpleStep ELISA Kit, cat. no. ab233611; Abcam). Briefly, standard and spheroid formation culture media samples were added into each well to bind E-cadherin antibody cocktail followed by incubation of 1 h at room temperature. After being washed with washing buffer, TMB1 substrate reagent was added into each well and incubated for 10 min. At this point, the stop solution was added and optical density was measured at 450 nm using a VERSAmax microplate reader.

Immunocytochemistry

Four weeks after spheroid formation e culture, the formed spheres were fixed for 24 h at 4°C in 4% paraformaldehyde, and 4 µm-thick-sections were prepared for immunocytochemistry. The sections were blocked with 10% normal horse serum (cat. no. MP-7401; Vector) for 1 h at room temperature. Incubation with the anti-E-cadherin antibody (1:100) was performed for overnight at 4°C. The binding was visualized using an anti-rabbit secondary antibody (1:200; cat. no. MP-7401; Vector), and the nuclei were counterstained with hematoxylin (cat. no. H-3404, Vector) for 1 min at room temperature.

Statistical analysis

All data were compiled from a minimum of three replicate experiments. Data are expressed as the mean values ± SD. P<0.05 was considered to indicate a statistically significant difference as determined using the Student's paired t-test or one-way ANOVA followed by a Bonferroni post-hoc test. MS Excel 2016 was used for statistical analysis.

Results

Spheroid formation in different environments supplemented with FBS or GF

To evaluate the tumorigenic capacities of SNU-C5 and SNU-C5/5-FUR CRC cells, the cells were cultured in an anchorage-independent condition for 4 weeks. Both cell lines successfully formed spheres in the FBS-supplemented environment. SNU-C5/5-FUR cells only formed spheres in the GF-supplemented environment. The mean size of each cell was significantly increased with the passage of time (P<0.001) (Fig. 1 and Table I).

Figure 1. Spheroid formation of CRC cells in different environments. The wild type SNU-C5 and 5-FU-resistant SNU-C5 (SNU-C5/5-FUR) CRC cells were cultured in an anchorage-independent condition supplemented with FBS or GF. The number and size (diameter) of spheres were estimated at 7, 14 and 28 days after seeding. Data are presented as the mean ± SD (n=3). Scale bar, 100 µm. *P<0.05, **P<0.01 and ***P<0.001 vs. GF; ###P<0.001 vs. p7. CRC, colorectal cancer; FBS, fetal bovine serum; GF, growth factor.

Table I. Number and size (µm) of spheres in GF- and FBS-supplemented environments of SNU-C5 and SNU-C5/5-FUR cells.

Table I.

Number and size (µm) of spheres in GF- and FBS-supplemented environments of SNU-C5 and SNU-C5/5-FUR cells.

	SNU-C5			SNU-C5/5-FUR
Variables	Sphere/GF	Sphere/FBS	P-value	Sphere/GF	Sphere/FBS	P-value
No.
D7	11.08±2.57	64.67±9.35	<0.001	10.25±1.27	19.25±3.72	<0.0309
D14	7.46±0.77	67.46±5.77	<0.001	6.58±1.36	26.88±5.74	<0.0069
D28	12.39±1.21	74.78±4.00	<0.001	5.28±0.35	23.17±0.52	<0.001
Size (µm)
D7	71.25±5.12	93.09±3.16	-	90.30±3.13	89.00±8.33	-
D14	82.59±4.72	245.36±15.08a	<0.001	197.04±10.24a	245.06±22.03a	<0.001
D28	85.50±5.04	312.34±41.14a	<0.001	467.19±59.78a	371.57±55.88a	<0.001

{ label (or @symbol) needed for fn[@id='tfn1-ol-25-5-13793'] } The numbers of spheres were compared between environments.

a The sizes of the spheres were compared with the previous time point (P<0.05). D, days of incubation; FBS, fetal bovine serum; GF, growth factor.

Cell viability and proliferation of spheroid formation in different environments supplemented with FBS or GF

To investigate the proliferation and acquisition of drug resistance in spheroid formation, the effect of an anticancer drug (5-FU) on cell viability was assessed using the MTT assay. Spheroid-formed cells from SNU-C5 cells revealed higher cell viability to 5-FU at 10 (P=0.0174 in sphere/GF; P=0.0421 in sphere/FBS) and 100 µM (P=0.0030 in sphere/GF; P=0.0162 in sphere/FBS) compared with the cells from monolayer culture. The difference in cell viability between environments (sphere/GF vs. sphere/FBS) was statistically significant (P=0.0421/10 µM; P=0.0455/100 µM). Resistance-acquired SNU-C5/5-FUR cells did not show any change to 5-FU with spheroid formation (Fig. 2A). Sphere-formed cells demonstrated significantly slower proliferation than those in monolayer, wherein sphere-formed cells reached the same level at 4 days of incubation when original cells reached proliferation level at 3 days of incubation (8). In addition, sphere-formed cells in GF-supplemented environment showed relatively slower proliferation than those in FBS-supplemented environment (Fig. 2B).

Figure 2. Cell viability changes after spheroid formation of colorectal cancer cells in different environments. (A) The SNU-C5 and SNU-C5/5-FUR cells were mock-treated with DMSO or treated with indicated doses of 5-FU for 3 days. The extent of cell viability was determined by MTT assay. Data are presented as the mean ± SD (n=3). *P<0.05, **P<0.01 and ***P<0.001 vs. DMSO; #P<0.05 and ##P<0.01 vs. sphere/GF. (B) The SNU-C5 and SNU-C5/5-FUR cells were mock-treated with DMSO or treated with indicated doses of 5-FU for indicated days. The extent of cell viability was determined by MTT assay using the aforementioned procedures. Data are presented as the mean ± SD (n=3). (C) Expression levels of cell cycle-related proteins in monolayer and spheroid formation cultures in SNU-C5 and SNU-C5/5-FUR cells were detected by immunoblotting. Immunoblotting analysis was performed for p21, β-catenin and GSK3β, while GAPDH was used for a loading control. Band density was analyzed by AzureSpot analysis software, and results are expressed as the mean ± SD (n=3). **P<0.01 and ***P<0.001 vs. monolayer; ##P<0.01 vs. sphere/FBS. 5-FU, 5-fluorouracil; FBS, fetal bovine serum; GF, growth factor.

To delineate the characteristics of spheroid-formed cells, the levels of p21, β-catenin and GSK3β were first measured to confirm the feasible mechanisms of slower proliferation (Fig. 2C and Table II). Compared with monolayer culture, p21 was increased in both CRC cells. Whereas β-catenin was found to be decreased in both CRC cells, GSK3β was decreased in FBS-supplemented environments and sustained in GF-supplemented environment (P=0.0046 between environments).

Table II. Densitometric results of western blotting on SNU-C5 and SNU-C5/5-FUR cells.

Table II.

Densitometric results of western blotting on SNU-C5 and SNU-C5/5-FUR cells.

	SNU-C5			SNU-C5/5-FUR
Variables	Sphere/FBS	P-value	Sphere/FBS	P-value	Sphere/GF	P-value
Proliferation
p21	1.19±0.13	0.1248	3.22±0.57	0.0015	4.73±0.67	<0.001
β-catenin	0.65±0.09	0.0013	0.54±0.14	0.0043	0.64±0.07	0.0014
GSK3β	0.80±0.06	0.0032	0.48±0.06	<0.001	1.06±0.16	0.3696
Stemness
pan-Ras	1.38±0.04	0.0065	0.71±0.07	0.0011	1.24±0.06	0.0013
Oct3/4	0.90±0.04	0.0948	0.61±0.07	<0.001	1.09±0.08	0.1597
Sox2	0.64±0.08	0.0019	0.49±0.03	<0.001	0.61±0.05	<0.001
WT1	0.88±0.05	0.0158	0.50±0.05	<0.001	1.35±0.08	0.0005
Drug resistance
ABCG2	0.64±0.05	<0.001	0.85±0.10	0.0982	0.72±0.12	0.0284
P90RSK	0.80±0.08	0.0370	0.97±0.21	0.4514	1.39±0.05	<0.001
pCREB	0.96±0.17	0.4190	1.33±0.16	0.0473	0.92±0.10	0.2358
CREB	0.82±0.05	0.0073	0.67±0.08	<0.001	0.81±0.14	0.0800
Surface marker
CD44	0.21±0.01	<0.001	0.16±0.03	<0.001	0.60±0.07	<0.001
CD24	0.69±0.03	<0.001	0.73±0.07	0.0075	0.94±0.08	0.2360
EMT
Fibronectin	4.90±0.73	<0.001	11.57±2.15	<0.001	0.67±0.10	0.0094
EpCAM	0.33±0.08	<0.001	0.11±0.03	<0.001	0.20±0.08	<0.001
αSMA	0.71±0.03	0.0095	0.51±0.06	<0.001	1.01±0.15	0.4869
E-cadherin
135-120 kDa	0.88±0.12	0.1727	0.55±0.13	0.0073	0.83±0.10	0.0873
80 kDa	3.80±0.70	0.0013	4.52±0.96	0.0054	0.70±0.10	0.0550

[i] Each result was first normalized by GAPDH and then compared with the results obtained from 2D monolayer culture as a standard. Results are shown in terms of the relative fold with p value, and GF- and FBS-supplemented environments were also compared in SNU-C5/5-FUR cells. EMT, epithelial-to-mesenchymal transition; FBS, fetal bovine serum; GF, growth factor; EpCAM, epithelial cell adhesion molecule.

Cellular and molecular markers of spheroid formation in different environment supplemented with FBS or GF

CSC markers were assessed in the spheroid-formed cells to reveal the stemness (Fig. 3A and Table II). Compared with monolayer culture, pan-Ras was significantly increased while other markers were decreased in SNU-C5 cells. In the case of SNU-C5/5-FUR cells, pan-Ras and WT1 were increased in sphere/GF environment but decreased in sphere/FBS environment, and considerable differences were observed between supplementations (P<0.001/each). Oct3/4 and Sox2 were decreased or sustained.

Figure 3. Cellular and molecular markers for CSCs and drug resistance on spheroid formation of colorectal cancer cells in different environments. (A) Expression levels of CSC markers in monolayer and spheroid formation cultures in SNU-C5 and SNU-C5/5-FUR cells were detected by immunoblotting. Immunoblotting analysis was performed for pan-Ras, Oct3/4, Sox2 and WT1, while GAPDH was used for a loading control. Band density was analyzed by AzureSpot analysis software, and results are expressed as the mean ± SD (n=3). (B) Expression levels of drug efflux markers in monolayer and spheroid formation cultures in SNU-C5 and SNU-C5/5-FUR cells were detected by immunoblotting. Immunoblotting analysis was performed for ABCG2, p90RSK, pCREB, and CREB, while GAPDH was used for a loading control. Band density was analyzed by AzureSpot analysis software, and results are expressed as the mean ± SD (n=3). (C) Expression levels of cell surface markers in monolayer and spheroid formation cultures in SNU-C5 and SNU-C5/5-FUR cells were detected by immunoblotting. Immunoblotting analysis was performed for CD44 and CD24, while GAPDH was used for a loading control. Band density was analyzed by AzureSpot analysis software, and results are expressed as the mean ± SD (n=3). *P<0.05, **P<0.01 and ***P<0.001 vs. monolayer; #P<0.05, ##P<0.01 and ###P<0.001 vs. sphere/FBS. CSCs, cancer stem cells; FBS, fetal bovine serum; GF, growth factor; p-, phosphorylated.

Next, drug resistance-related markers were assessed in the spheroid-formed cells (Fig. 3B and Table II). Compared with monolayer culture, ABCG2, p90RSK, phosphorylated (p)-CREB and CREB were all decreased or unchanged in SNU-C5 cells. Regarding SNU-C5/5-FUR cells, p90RSK was increased in GF-supplemented environment and pCREB was increased in FBS-supplemented environment. The other markers were decreased or unchanged in sphere-forming cells.

Surface markers related to stemness and/or EMT were also assessed in the spheroid-formed cells (Fig. 3C and Table II). Compared with monolayer culture, epithelial CD24 and mesenchymal CD44 were decreased in spheres in both CRC cells. The ratio between CD44/CD24, which is useful data for searching CSCs, was significantly decreased in sphere/FBS (0.30±0.02-fold in SNU-C5 cells, 0.24±0.06-fold in SNU-C5/5-FUR cells; P<0.001/each) and in sphere/GF (0.65±0.08 fold; P=0.0048). There was a significant difference between supplementations (P=0.0013) in SNU-C5/5-FUR cells.

Differential expression of EMT markers of spheroid formation in different environment supplemented with FBS or GF

As the spheroid formation culture method inhibits the adhesion of cells to the base of culture dish, the EMT markers were assessed (Fig. 4A and Table II). Compared with the monolayer culture, fibronectin was significantly increased in FBS-supplemented environment in both CRC cells but decreased in GF-supplemented environment in SNU-C5/5-FUR cells (P<0.001 between environments). EpCAM and αSMA were not changed in the spheroid formation culture. Compared with monolayer culture, 135–120 kDa E-cadherin was decreased in all spheroid formation conditions that were examined in this experiment. Soluble E-cadherin (80 kDa) was significantly increased in the FBS-supplemented environment in both CRC cells, but sustained in the GF-supplemented environment of SNU-C5/5-FUR cells (Fig. 4B and Table II). The ratio of soluble to 135–120 kDa E-cadherin was significantly increased in FBS-supplemented environment in SNU-C5 (2.50±0.49-fold; P=0.0062) and in SNU-C5/5-FUR (4.64±0.42-fold; P<0.001) cells. The ratio was sustained in sphere/GF (1.06±0.08-fold), while there was a significant difference between supplementations (P<0.001) in SNU-C5/5-FUR cells.

Figure 4. Markers for EMT on spheroid formation of colorectal cancer cells in different environments. (A) Expression levels of EMT markers in monolayer and spheroid formation cultures in SNU-C5 and SNU-C5/5-FUR cells were detected by immunoblotting. Immunoblotting analysis was performed for fibronectin, EpCAM and αSMA, while GAPDH was used as a loading control. Band density was analyzed by AzureSpot analysis software, and results are expressed as the mean ± SD (n=3). (B) Expression of E-cadherin in monolayer and spheroid formation cultures in SNU-C5 and SNU-C5/5-FUR cells was detected by immunoblotting. Immunoblotting analysis was performed for E-cadherin, which was subdivided into 135 −120 kDa and soluble 80 kDa fractions, while GAPDH was used as a loading control. Band density was analyzed by AzureSpot analysis software, and results are expressed as the mean ± SD (n=3). (C) Enzyme-linked immunosorbent assay of E-cadherin on culture media in SNU-C5 and SNU-C5/5-FUR cells. Data are presented as the mean ± SD (n=3). (D) Immunocytochemical staining of E-cadherin on spheroid formation in SNU-C5 and SNU-C5/5-FUR cells. Asterisks indicate areas that were neither immunostained with E-cadherin, nor stained with hematoxylin (Scale bar, 100 µm). **P<0.01 and ***P<0.001 vs. monolayer; #P<0.05, ##P<0.01 and ###P<0.001 vs. sphere/FBS. EMT, epithelial-to-mesenchymal transition; FBS, fetal bovine serum; GF, growth factor; EpCAM, epithelial cell adhesion molecule.

The concentration of E-cadherin in cultured media of SNU-C5 and SNU-C5/5-FUR cells in 2D monolayer, sphere/FBS, and sphere/GF group were estimated by ELISA (Fig. 4C). The ratio of soluble E-cadherin was found to be significantly increased in 28 days of incubation in sphere/FBS compared with 2D monolayer (3 days of incubation) and with 7 days of incubation in sphere/FBS environment in SNU-C5 cells (25.47±1.98-fold vs. monolayer, P<0.001), and in SNU-C5/5-FUR cells (4.48±0.50-fold vs monolayer, P<0.001), respectively. However, the ratio of E-cadherin was not changed until 28 days after incubation in GF-supplemented environments in both CRC cells.

The morphological features were evaluated by the expression of E-cadherin (Fig. 4D). E-cadherin was immunostained on the cell membrane of spheroid-formed cells in both CRC cells with GF and FBS supplementations. Some central area was neither immunostained with E-cadherin, nor stained with hematoxylin for nuclei.

Discussion

The efficiency of spheroid formation is known to differ between cell lines, even within the same tumor type (22), which was also observed in a recent study by the authors (17). It was found that the spheroid formations of both CRC cell lines were induced in an improved manner when supplemented with FBS compared with GFs, while SNU-C5/5-FUR cells only formed spheres supplemented with GFs. Sphere-formed cells showed 5-FU resistance in SNU-C5 cells irrespective of the supplementations used as previously suggested (22). However, SNU-C5/5-FUR cells did not show any further changes against 5-FU in spheroid formation culture. Sphere-formed cells showed slower cell proliferation than cells from monolayer culture, which coincided with an increased level of p21 and a decreased level of β-catenin as previously reported (19,21,22). However, cellular and molecular markers for CSCs, drug resistance, and EMT were not significantly changed between 2D and 3D culture conditions although the cells could acquire CSC-like phenotype via EMT to form spheres (19,20). As a result, spheroid formation culture methods are not appropriate to study CSCs or drug resistance in acquired 5-FU-resistant CRC cells, at least in long-term maintenance condition. Notably, spheroid formation supplemented with FBS environment showed significantly increased level of soluble E-cadherin.

GFs were used in spheroid formation culture methods (17,22,26,27). The suitability of spheroid formation culture methods has been investigated, and the results of such studies have revealed that cancer cells showed differential efficiency of spheroid formation, where the efficiency differed depending on cell lines even for the same tumor type (22). SNU-C5/5-FUR cells showed variable morphology of spheroids in different environments, wherein complete spheroids were formed in GF-supplemented environment with smaller numbers than those formed in FBS-supplemented environment. Although the spheroids in GF-supplemented environment did not acquire further drug resistance, they exhibited increased CSC markers compared with monolayer or FBS-supplemented spheroids. Accordingly, drug-resistance-acquired cells were not suitable for spheroid formation culture methods to investigate CSCs depending on drug resistance, at least in SNU-C5/5-FUR cells. As FBS supplementation was suggested for the in vitro cultivation of CSCs (28), spheroid formation was easily induced and maintained for up to 4 weeks in FBS-supplemented environment in both CRC cells. The spheres showed acquired drug resistance, slow proliferation, and increased level of pan-Ras. However, further changes on cellular and molecular markers of CSCs, drug resistance, surface protein, and EMT were not observed. Therefore, FBS-supplemented environment was not considered to be such an effective tool for CSCs in CRC cells, at least in SNU-C5/5-FUR cells and in long-term maintenance.

During the spheroid formation process in any environment, it is known that cells initially aggregate and then form compact spheroids via a high level of E-cadherin (22). However, expression of E-cadherin was reported as ‘decrease’ in CSC, as induced by spheroid formation culture methods, of the colon (29,30), breast (31,32), liver (33), pancreas (34), and oral and prostate (35) cancers with variable vendors in short-term maintenance. Previous studies (28,29,31–33) reported fractions with 120 kDa, not soluble (80 kDa), E-cadherin. The decrease in E-cadherin in CSC was also supported by the findings of confocal imaging (36) and immunohistochemistry (31,32). However, there have also been controversial studies regarding the expression of E-cadherin in spheroid formation. Transforming GF-beta-induced EMT decreases the expression of E-cadherin in spheroid formation (33,34). E-cadherin enhances CSC-like properties and induces mesenchymal features in colon cancer (37). The overexpression of E-cadherin compromises the EMT-like properties of spheroid formation (32). Spheroid formation in breast cancers depends on the expression of E-cadherin (21). Moreover, Morata-Tarifa et al (38) suggested that trypsin-resistant subpopulation showed increased expression of E-cadherin, but trypsin-sensitive subpopulation (EMT-like) represented CSC-like colon cancer cells.

Previous studies have indicated that decreased expression of E-cadherin in cancers may be re-interpreted as the decrease of 120 kDa E-cadherin irrespective of vendors. In the present study, 120 kDa E-cadherin was significantly decreased only in the sphere/FBS of SNU-C5/5-FUR cells. In contrast to expectations, the soluble E-cadherin was significantly increased in the sphere/FBS of both CRC cells as revealed using western blotting and ELISA, which means the essential component of spheroid formation culture was the soluble E-cadherin in FBS-supplemented environment. As the increase in soluble E-cadherin was observed in 28 days after incubation in both CRC cells, it may promote growth and proliferation of spheroids rather than act as an CSCs marker or drug resistance. These findings are confirmed by previous studies that E-cadherin could be a target for spheroid formation or the formation of CSC-like cells in breast cancer (39), and soluble E-cadherin could be an oncogene like EGF when it induces EMT in 3D (24,25,40) as well as in 2D (24) culture environments. Nevertheless, there has been controversy about the mechanism of soluble E-cadherin. The soluble E-cadherin influences invasion via activation of matrix metalloproteinase (MMP), and thus contributes to skin carcinogenesis (24). Exogenous soluble E-cadherin promotes migration and invasion of non-small-cell lung cancer cells, but silencing MMP9 suppresses soluble E-cadherin expression (41). Although ABCG2 could also modulate the expression of E-cadherin in lung cancer (42), ABCG2 did not significantly change in the present study.

Nonetheless, the present study did not reveal the feasible mechanisms of soluble E-cadherin on spheroid formation. As the increase in soluble E-cadherin was observed in long-term maintenance of spheroids, the application of exogenous E-cadherin in spheroid formation culture should be performed to reveal whether soluble E-cadherin is a result or a cause of spheroid formation. If soluble E-cadherin act as an oncogene as it was hypothesized, the possibility of replacement to EGF would be explored based on serial dilution as performed in a recent study (26). Based on these experiments, following signaling pathways of soluble E-cadherin should be further investigated under the more detailed strategy.

In conclusion, spheroid formation culture methods are not appropriate to study CSCs or drug resistance in acquired 5-FU-resistant CRC cells as compared with wild type cells, at least in long-term maintenance condition. Spheroids were easily formed in FBS-supplemented environment via adaptation to anchorage-independent condition, which are related to soluble E-cadherin. These results suggested that soluble E-cadherin could act like an oncogene to grow the spheroids supplemented with FBS, at least in the long-term maintenance condition.

Acknowledgements

Not applicable.

Funding

The present study was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT; grant no. 2021R1F1A1063023).

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

IYC and SPY conceived and designed the present study, performed the experiments for data acquisition and analysis and interpreted the experimental results. IYC and SPY confirm the authenticity of all the raw data. IYC wrote the original manuscript. SPY revised the manuscript. Both authors read and approved the final manuscript and agree to be accountable for all aspects of the research in ensuring that the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References