Molecular Mechanisms and Tumor Biological Aspects of 5-Fluorouracil Resistance in HCT116 Human Colorectal Cancer Cells

5-Fluorouracil (5-FU) is a cornerstone drug used in the treatment of colorectal cancer (CRC). However, the development of resistance to 5-FU and its analogs remain an unsolved problem in CRC treatment. In this study, we investigated the molecular mechanisms and tumor biological aspects of 5-FU resistance in CRC HCT116 cells. We established an acquired 5-FU-resistant cell line, HCT116RF10. HCT116RF10 cells were cross-resistant to the 5-FU analog, fluorodeoxyuridine. In contrast, HCT116RF10 cells were collaterally sensitive to SN-38 and CDDP compared with the parental HCT16 cells. Whole-exome sequencing revealed that a cluster of genes associated with the 5-FU metabolic pathway were not significantly mutated in HCT116 or HCT116RF10 cells. Interestingly, HCT116RF10 cells were regulated by the function of thymidylate synthase (TS), a 5-FU active metabolite 5-fluorodeoxyuridine monophosphate (FdUMP) inhibiting enzyme. Half of the TS was in an active form, whereas the other half was in an inactive form. This finding indicates that 5-FU-resistant cells exhibited increased TS expression, and the TS enzyme is used to trap FdUMP, resulting in resistance to 5-FU and its analogs.

In this study, we established a 5-FU-resistant HCT116 CRC cell line (HCT116R F3 and HCT116R F10 ) and analyzed its biological features. HCT116R F10 cells, which are crossresistant to the 5-FU analog fluorodeoxyuridine (FUdR), were collaterally sensitive to SN-38 and CDDP compared with the parental HCT16 cells. In addition, HCT116R F10 cells exhibited a lower ability to form tumor spheres compared with parental HCT116 cells. Notably, HCT116R F10 cells maintained the tumor sphere formation ability compared with HCT116 cells under 5-FU exposure conditions. Furthermore, a gene cluster associated with 5-FU metabolic pathway was not significantly mutated in HCT116 and HCT116R F10 cells as determined by whole-exome sequencing. We found that HCT116R F10 cells regulate intracellular TS states in which half of the TS enzyme is in a functional form and the other half exists as an FdUMP-covalent complex (inactive form). These findings provide a better understanding of resistance to anticancer 5-FU and its analogs.

Establishment of the 5-Fluorouracil-Resistant HCT116 Cells
To elucidate the mechanisms underlying resistance to 5-FU, we generated a variant of the HCT116 human colorectal cancer cell line that was resistant to 5-FU, an important anticancer drug used for CRC treatment [2,3]. We established 5-FU-resistant HCT116R F3 or HCT116R F10 cells by repeated exposure of parental HCT116 cells to stepwise increasing concentrations of 5-FU over a period of approximately 12 weeks at 3 µM and 14 weeks at 10 µM, respectively (Figure 1a). The EC 50 of 5-FU in HCT116R F3 (intermediate variant) and HCT116R F10 cells were determined by a WST-8 assay after continuous exposure for 72 h. As shown in Table 1 and Figure 1b, the EC 50 value of the 5-FU-resistant HCT11 6 cells was higher (1.5 × 10 −5 M in HCT116R F3 and 2.9 × 10 −5 M in HCT116R F10 cells) than that of sensitive, parental HCT116 cells (5.1 × 10 −6 M). The RI was approximately 2.9 for HCT116R F3 cells and 5.7 for HCT116R F10 cells (Table 1). In addition, similar results were obtained by colony formation assay (Figure 1c,d). The EC 50 value of 5-FU-resistant HCT116 cells was significantly higher (1.6 × 10 −5 M in HCT116R F3 and 3.8 × 10 −5 M in HCT116R F10 cells) than that of the parental HCT116 cells (5.5 × 10 −6 M) ( Table 1). The RI of HCT116R F3 and HCT116R F10 cells was approximately 2.9 and 6.9, respectively (Table 1). Furthermore, parental HCT116, HCT116R F3 , and HCT116R F10 cells exhibited nearly similar morphological features (Figure 1e).

Biological Features of the 5-FU-Resistant HCT116 Cells
We analyzed the tumor sphere formation ability of HCT116R F10 cells and parental HCT116 cells in three-dimensional cell culture experiments ( Figure 4). HCT116R F10 cells exhibited a lower ability to form tumor spheres compared with parental HCT116 cells under untreated conditions (Figure 4a left panel and b). Interestingly, HCT116R F10 cells maintained a tumor sphere formation ability compared with parent HCT116 cells during 5-FU treatment conditions (Figure 4a,c). We next examined the sensitivity of parental HCT116 and HCT116R F10 tumor sphere cells to 5-FU. As shown in Figure 4d, HCT116R F10 cells were 18.7-fold (EC 50 = 2.8 × 10 −5 M) more resistant to 5-FU than parental HCT116 cells (EC 50 = 1.5 × 10 −6 M). These data indicate that 5-FU-resistant HCT116R F10 cells are less prone to tumorigenesis than sensitive, parental HCT116 cells, but formed tumor spheres that retained a higher 5-FU resistance.

Exome Sequencing Analysis of HCT116 Parent Cells and 5-FU-Resistant HCT116R F10 Cells
We analyzed variants of 5-FU metabolic pathway-related enzyme genes, including TYMS, which encodes for TS, and DPYD, which encodes for DPD in HCT116 and HCT116R F10 cells. TS is a major intracellular target of 5-FU, whereas DPD catalyzes the rate-limiting step in the catabolism of 5-FU [2,3,11]. The pathways involved in the metabolism of 5-FU and its analog FUdR are shown in Figure 5. The genetic alteration status of nearly all of the 5-FU metabolic pathway-related genes was of similar status in both cells. Importantly, the variants of TYMS and DPYD in HCT116 and HCT116R F10 cells contained heterozygous mutations or intron variants. We identified two TYMS intron variants, 454+197_454+202delTTTTTT and 454+199_454+202delTTTT, in HCT116R F10 cells.

Regulation of TS and DPD in HCT116 Parent Cell and 5-FU-Resistant HCT116R F10 Cells
To elucidate the association of TS and DPD expression with 5-FU resistance, we analyzed TS and DPD expression levels in parental HCT116 and 5-FU-resistant HCT116R F10 cells by Western blot analysis (Figure 6a). Interestingly, as shown in Figure 6a

Regulation of TS and DPD in HCT116 Parent Cell and 5-FU-Resistant HCT116R F10 Cells
To elucidate the association of TS and DPD expression with 5-FU resistance, we analyzed TS and DPD expression levels in parental HCT116 and 5-FU-resistant HCT116R F10 cells by Western blot analysis (Figure 6a). Interestingly, as shown in Figure 6a (top panel) and 6b, free-TS protein levels were almost identical in HCT116R F10 and HCT116 cells. Conversely, the FdUMP-TS covalent complex was 1.8-fold higher in HCT116R F10 cells than in HCT116 cells (Figure 6a top panel and 6c). Importantly, it should be noted that total TS, the free form, the FdUMP-covalent form, and total TS was overexpressed in HCT116R F10 cells rather than in HCT116 cells (Figure 6a top panel and Figure 6d). The upper band of TS, indicated FdUMP-covalent form, which represents TS in ternary complexes and is correlated with the intracellular concentration of FdUMP [12][13][14]. In addition, DPD protein levels were slightly decreased in HCT116R F10 cells than in parental HCT116 cells (Figure 6a second panel and Figure 6e). GAPDH and beta-actin were used as an internal controls (Figure 6a third and bottom panels). In parental HCT116 cells and HCT116R F10 cells, both internal control proteins, GAPDH and beta-actin, had similar levels. After treatment with 1 × 10 −4 M 5-FU for 24 h, the protein levels of free TS, FdUMP-TS covalent complex, and total TS were individually about 1.5-fold higher in HCT116R F10 cells than in parental HCT116 cells (Figure 7a-d). Intriguingly, these data indicated that the proportion of active free TS in the intracellular total TS was highly regulated in the 5-FU resistant HCT116R F10 cells. These findings suggested that the regulation of TS status, which includes the balance of active free TS or the inactive FdUMP-TS covalent complex, may confer resistance to 5-FU.    [2,3]. Studies to date have shown that cancer cells develop resistance to 5-FU through complex mechanisms [2,3]. Of note, the TS enzyme and other enzymes involved in 5-FU anabolism or catabolism are often altered in expression or function to promote 5-FU resistance [2,3]. In addition, altered cell death and autophagy, expression/functional changes in drug transporters, epigenetic changes, and non-coding RNA (i.e., microRNA and long non-coding RNA) dysfunction represent putative 5-FU-resistant mechanisms [2,3]. It has been widely believed that TS is the main molecular mechanism that influences 5-FU sensitivity and targeting TS is a major strategy for reversing 5-FU resistance. Importantly, there are currently no specific therapies to overcome 5-FU resistance.

5-FU and its derivatives are widely used in anticancer chemotherapy
We established a 5-FU-resistant cell line, HCT116R F10 , and analyzed its characteristics. Importantly, HCT116R F10 cells were cross-resistant to the 5-FU analog, FUdR ( Figures  2 and 3). In contrast, HCT116R F10 cells did not exhibit cross-resistance to the anticancer drugs, SN-38 and CDDP (Figures 2 and 3). Similarly, Boyer et al. also reported that 5-FU-  [2,3]. Studies to date have shown that cancer cells develop resistance to 5-FU through complex mechanisms [2,3]. Of note, the TS enzyme and other enzymes involved in 5-FU anabolism or catabolism are often altered in expression or function to promote 5-FU resistance [2,3]. In addition, altered cell death and autophagy, expression/functional changes in drug transporters, epigenetic changes, and non-coding RNA (i.e., microRNA and long non-coding RNA) dysfunction represent putative 5-FU-resistant mechanisms [2,3]. It has been widely believed that TS is the main molecular mechanism that influences 5-FU sensitivity and targeting TS is a major strategy for reversing 5-FU resistance. Importantly, there are currently no specific therapies to overcome 5-FU resistance.

5-FU and its derivatives are widely used in anticancer chemotherapy
We established a 5-FU-resistant cell line, HCT116R F10 , and analyzed its characteristics. Importantly, HCT116R F10 cells were cross-resistant to the 5-FU analog, FUdR (Figures 2 and 3). In contrast, HCT116R F10 cells did not exhibit cross-resistance to the anticancer drugs, SN-38 and CDDP (Figures 2 and 3). Similarly, Boyer et al. also reported that 5-FU-resistant HCT116 cells were not cross-resistant to oxaliplatin or irinotecan [15]. In addition, the sensitivities to 5-FU and FUdR were similar to the levels observed individually in parental HCT116 cells. Of note, previous reports indicated that FUdR is more potent than 5-FU and that the inhibition of cell proliferation was approximately 10-to 100-fold higher than that of 5-FU in multiple cancer cell lines [16][17][18]. These findings suggest that the common target or mechanism of action of 5-FU and FUdR is the key to 5-FU resistance in this resistant cell model. Furthermore, our results revealed that HCT116R F10 cells are resistant to 5-FU and its derivatives, but are not multidrug resistant.
To elucidate the underlying cause of 5-FU resistance, we investigated 5-FU metabolismrelated genes, including TYMS and DPYD, in HCT116R F10 and parent HCT116 cells by using whole-exome sequencing. The results revealed that the genetic alteration of almost all of the 5-FU metabolic pathway-related genes was similar in status, intron variants, and heterozygous mutation in both cells (Table 3). Interestingly, we found that the one functional DPD mutation, Asp1000Val, is present in HCT116R F10 cells. However, the effects of DPYD missense mutation on 5-FU resistance are not well understood.
Next, to evaluate TS and DPD in HCT116R F10 and parent HCT116 cells, we analyzed the expression of these genes by Western blot analysis ( Figure 6). 5-FU and FUdR are converted to FdUMP, and it has been shown to form a covalent complex with TS in the presence of CH 2 -THF [2,3,5]. Our results indicated that the free-TS protein (active form) levels were similar in HCT116R F10 and HCT116 cells. Interestingly, the FdUMP-TS covalent complex (inactive form) was higher in HCT116R F10 cells than in HCT116 cells. Notably, this result indicates that TS is not overexpressed, but rather there are two types of TS in HCT116R F10 cells: free TS and FdUMP-coupled TS. We observed that 5-FU-resistant HCT116R F10 cells exhibit upregulated TYMS expression and use a fraction of TS to trap FdUMP, resulting in resistance to 5-FU and its analogs. In addition, our data suggest that the regulation of the TS complex, which refers to the balance of the active free-TS form and the inactive FdUMP-TS covalent complex, may confer to 5-FU resistance.
Numerous studies have shown that TYMS gene amplification, leading to mRNA and enzyme overproduction, is a major mechanism of resistance to fluoropyrimidines 5-FU and FUdR and their derivatives [19]. Also, free TS binds to its own mRNA, resulting in translational repression, that is, translational autoregulation [12,[20][21][22][23]. Indeed, TS ligands, including 5-FU, disrupt the interaction of the TS enzyme with TS mRNA, leading to translational derepression and enzyme upregulation [12,22,23]. Additionally, to translational derepression, enzyme stabilization has been indicated as the primary mechanism of TS induction by fluoropyrimidines in human colon and ovarian cancer cell lines [24][25][26]. Furthermore, it is proposed that fluoropyrimidine-mediated increases in TS levels occur through an effect on enzyme stability with no effect on its mRNA [25,27]. It is also suggested that TS stabilization could be the result of conformational changes that may occur upon the formation of a ternary complex, reducing the susceptibility of the TS enzyme to proteolysis [28]. These findings indicated that understanding translational derepression and enzyme stabilization as the process of TS induction has significance for elucidating the mechanism of resistance acquisition. Further investigation is needed on the functions of the FdUMP-TS covalent complex and free TS in both translational regulation and enzyme stabilization for fluoropyrimidine resistance mechanisms using 5-FU-resistance and 5-FU-sensitive parental HCT116 cell lines. Collectively, our findings provide a better understanding of the anticancer drugs, 5-FU and its fluoropyrimidine derivatives, with respect to resistance mechanisms and anticancer treatment strategies.

Cell Culture
The human colon cancer cell line HCT116 was obtained from the American Type Culture Collection. Parental and 5-FU-resistant HCT116 cell lines were cultured in DMEM medium containing 10% heat-inactivated fetal bovine serum, 100 units/mL penicillin, and 100 µg/mL streptomycin in a 37 • C incubator under an atmosphere containing 5% CO 2 and 100% relative humidity.

Cell Viability by WST-8 Assay
Cell viability assays were performed as previously described [29]. Cell viability was determined using the WST-8 (Cell Counting Kit-8) cell proliferation assay (Dojindo, Tokyo, Japan). Briefly, cells were seeded into 96-well plates (1000 cells per well) in triplicate and then treated with various concentrations of anticancer drugs or DMSO and water (as a negative control). Following incubation for 72 h, WST-8 reagent was added to each well and the plate was placed in a 5% CO 2 incubator at 37 • C for an additional 1 h. Optical density was measured at 450 nm on a Tecan microplate reader (Mannedorf, Switzerland). The EC 50 value was defined as the concentration of drug producing 50% inhibition of cell proliferation. The resistance index (RI) was defined as the ratio of EC 50 values between the resistant and parental cell lines. Experiments were repeated at least three times.

Colony Formation Assay
Colony formation assay was performed as previously described [29][30][31][32]. HCT116 and HCT116R F10 cells were dissociated with Accutase, suspended in medium, inoculated into 6-well plates (200 cells per well) in triplicate, and then incubated overnight. The cells were treated with various concentrations of drugs or with solvent (DMSO or water) as a negative control. After incubation for 10 days, cells were fixed with 4% formaldehyde solution and stained with 0.1% (w/v) crystal violet, and the number of colonies in each well was counted.

Tumor Sphere Assay
HCT116 and HCT116R F10 cells were seeded into 96-well PrimeSurface ® plate 96U (Sumitomo Bakelite Co., Ltd., Tokyo, Japan) plates (1000 cells per well) in triplicate and then treated with various concentrations of 5-FU or DMSO (as a negative control). Following incubation for 14 days, tumor sphere size was monitored once every 3-4 days. Tumor sphere volume (V) was calculated using the following formula: V = ab 2 /2 (a and b are the long and short diameters of the tumor sphere, respectively).

Exome Sequencing Analysis
DNA extraction was performed as previously described [29]. Genomic DNA was extracted from cells (5 × 10 6 cells) by using a DNeasy Tissue Kit (QIAGEN, Venlo, Netherlands), according to the manufacturer's instructions. Exome sequencing of parental HCT116 and HCT116R F10 cells was performed by APRO Life Science Institute, Inc. (Tokushima, Japan) and Macrogen Global Headquarters (Seoul, Korea).

Statistical Analysis
The data are presented as means ± standard deviation. The significance of differences among groups was evaluated using a Student's t-test; p < 0.05 was considered statistically significant.