CRC cells HCT116, HT29, SW620, HCT-8 were collected in our laboratory, and a 5-FU-resistant cell line HCT-8/5-FU and its parental HCT-8 cell line were ordered from Hunan Fenghui Biotechnology Co., Ltd. HCT116, HT29 and SW620 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM). HCT-8 and HCT-8/5-FU cells were cultured in RPMI-1640 supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin and 100 μg/mL streptomycin. The HCT-8/5-FU cells ordered from Hunan Fenghui Biotechnology Co., Ltd., possess a certain initial tolerance to 5-FU, with a resistance index (RI) of around 10. Subsequently, we further induced the cells with 5-FU to enhance the resistance to 5-FU, gradually increasing the concentration from 200 μM, 400 μM, 800 μM–1,600 μM over a period of approximately 5 months. The HCT-8/5-FU cells were passaged no more than 20 times, and 50 μM 5-FU was added to the culture medium on a daily basis to maintain their resistance to 5-FU.

Cell lysates used for Western blot analysis was prepared as our published methods (Chen et al., 2021; Sheng et al., 2023). Specific primary antibodies: rabbit anti-SUMO1 (ET1606-53, HuaBio), rabbit anti-SUMO2/3 (ET1701-17, HuaBio), rabbit anti-UBC9 (ET1610-21, HuaBio), rabbit anti-SAE1 (ET7108-22, HuaBio), rabbit anti-SAE2 (ET1705-73, HuaBio), mouse anti-Flag (F1804, Sigma-Aldrich). Mouse anti-β-Tubulin (TA-10, Zsbio) and mouse anti-β-Actin (TA-09, Zsbio) were used for internal control. The primary antibodies were diluted at 1:1,000, and second antibodies were diluted at 1:10000. For global proteins SUMOylation detection, a final concentration of 20 mM N-ethylmaleimide (NEM) (HY-D0843, sigma) was added to RIPA lysis buffer.

Cell proliferation and IC₅₀ detection was measured by cell counting kit-8 (CCK-8) assay referred as our previous publications (Yang et al., 2019). 2 × 10³ cells/well of HCT-8 or HCT-8/5-FU were seeded into a 96-well plate for indicated time. According to the manufacturer’s instruction, 10 µL CCK-8 reagent was added into 100 µL DMEM medium in each well for 2 h incubation at 37°C, then absorbance was measured at 450 nm. For IC₅₀ detection, cells were seeded in 96-well plate and cultured with different concentration of drugs for 48 h. As for colony formation assay, 1,000 cells were seeded in 6-well plate with or without treatment of drugs, and cultured for 12–16 days.

EdU assay was performed in accordance with manufacturer’s instruction (KTA 2031, Abbkine). Briefly, EdU agent was supplied to cell culture with a final concentration at 50 µM for 6 h. After fixation, click-reaction and DAPI staining, images were taken with a fluorescence microscope.

Flow cytometry was conducted using a Novoexpress system. For cell cycle assays, we first seeded 1 × 10⁵ cells in 6-well plate and FBS starvation was treated for 24 h. Then a cell cycle detection kit (KGA511, KeyGen BioTECH) was used to prepare the cells for cell cycle according to the manufacturer’s instructions.

The promoter sequence of UBC9 gene (Gene ID:7329) was cloned into firefly luciferase reporter plasmid pGL3-Basic (Chen et al., 2014) at MluⅠ and XhoⅠ sties (pGL3-UBC9), and the sequence was confirmed by DNA sequencing. The Renilla luciferase was used as an internal control. The Ras-responsive element binding protein 1 (RREB1) binding sequence in promoter of UBC9 and SAE1 was predicted in online website (https://jaspar.genereg.net/). Briefly, 2 kb sequence upstream of transcription start site was downloaded from NCBI in a FASTA-formatted style and was imported into JASPAR website to scan RREB1 binding motif (ID: MA0073.1). Relative profile score threshold was set at 80%.

pFlag-RREB1 and pFlag-RREB1^3KR plasmids were generated in our laboratory. In detail, RREB1 coding sequence (GI: 1519315172) was cloned into pCMV6-entry-Flag plasmid between Sgf Ⅰ and Xho Ⅰ (pFlag-RREB1). On the basis of pFlag-RREB1 plasmid, we mutated three lysine (K) amino acids into arginine (R) amino acids at sites 551, 885 and 913 to generate a pFlag-RREB1^3KR plasmid, expressing a mutant RREB1 protein with a lack of SUMOylation modification.

Statistical analysis was performed as required. Two-tailed unpaired Student’s t-test was applied to compare two groups of data. Two-way ANOVA was used to analyze the statistical of cell cycle distribution. Data was represented as mean ± SD (standard deviation). p-value <0.05 was considered as a statistically significant difference.

We first detected SUMOylation level of global proteins under chemotherapy drug exposure in CRC cell lines, including HCT116, HT29 and SW620 and normal colon epithelial NCM460 cells. As results, levels of SUMO1 and SUMO2/3 modifications of global proteins in CRC cell lines were significantly higher than those in normal colon epithelial cells NCM460 (Figure 1A), suggesting that SUMOylation modification is linked with occurrence and development of CRC.

To investigate whether SUMOylation modification is involved in the response of chemotherapy drug treatment, HT29 cells were treated with different concentrations of 5-FU, irinotecan and oxaliplatin for 24 h, and then changes of global protein SUMOylation profiling were detected by Western blot. The results showed that irinotecan and oxaliplatin had little effect on the SUMOylation level of global proteins, while 5-FU obviously reduced the SUMOylation modification of total proteins in a concentration-dependent manner. The downregulation of global protein modified by SUMO2/3 was more obvious than that modified by SUMO1. 10 μM 5-FU treatment downregulated the level of SUMO2/3-modified global proteins to 0.45-fold compared with no 5-FU exposure. We also detected the enzymes involved in SUMOylation process, among which the expression level of SAE2 did not show obvious changes under chemotherapy drug exposure, while the expression of SAE1 and UBC9 level was respectively decreased to 0.45-fold and 0.38-fold under 10 µM 5-FU treatment (Figure 1B), suggesting that global proteins SUMOylation may play a role in 5-FU response.

We further treated HCT116 and HT29 cells with 4 µM oxaliplatin or 10 µM 5-FU to validate the effect of chemotherapy agents on the decrease of SUMOylation in CRC. The results showed that 5-FU treatment reduced the level of SUMO2/3-modified global proteins in HCT116, while oxaliplatin treatment cause a little downregulation of the level of SUMO2/3-modified global proteins (Figure 1C), demonstrating that the SUMOylation modification of global proteins is sensitive to 5-FU exposure, but not to irinotecan and oxaliplatin. These results suggest that global proteins SUMOylation modification, especially SUMO2/3-modified global proteins, may play an important role in 5-FU therapy in CRC cells.

A 5-FU-resistant cell line HCT-8/5-FU was established by inducing partial 5-FU-tolerant HCT-8/5-FU cells continuously with 5-FU induction. Compared with the parental HCT-8 cells, HCT-8/5-FU cells showed a significant increase in tolerance to 5-FU, resistance index (RI) is 56. After treatment with 20 µM 5-FU, parental HCT-8 cells showed obvious apoptotic morphology such as cell shrinkage and floating, while the morphology of 5-FU-resistant HCT-8/5-FU cells did not change significantly and still maintained adherent growth (Figure 2A). We further compared cell response characterization under 0–40 µM 5-FU exposure between HCT-8 cells and HCT-8/5-FU cells, and found that 5-FU treatment gradually decreased cell viability of HCT-8 cells, with the OD value (measured at 450 nm) dropping from 1.7 to 0.6. However, 40 µM 5-FU exposure had a little influence on cell viability of HCT-8/5-FU cells, indicating that our established HCT-8/5-FU cells harbored a 5-FU-resistance ability (Figure 2B).

Next, we measured the growth characteristics of HCT-8 and 5-FU-resistant HCT-8/5-FU cells by CCK-8, EdU and colony formation assay. HCT-8/5-FU cells showed weaker cell viability compared with HCT-8 cells from 48 to 96 h. At 96 h, the OD (measured at 450 nm) of HCT-8 reached 1.770 ± 0.301, while the OD of HCT-8/5-FU was significantly lower (0.867 ± 0.053, p < 0.01) (Figure 2C). We next performed an EdU assay in HCT-8 and HCT-8/5-FU cells to monitor the cell proliferation. The results showed that the EdU-positive cells in HCT-8 cells reached 20% (32/160), which is obviously higher than that in HCT-8/5-FU cells (almost none EdU-positive cells in HCT-8/5-FU). In addition, 5-FU treatment induces the cell apoptosis in HCT-8 cells, thereby dramatically reducing the proportion of EdU-positive cells in HCT-8 (Figure 2D).

The colony formation number of HCT-8/5-FU cells was significantly lower than that of HCT-8 cells (19.6 ± 4.72 vs. 42.67 ± 6.8, p < 0.01) (Figure 2E). These results demonstrated that HCT-8/5-FU cell line decreased its cell proliferation and colony formation ability. Furthermore, flow cytometry analysis revealed that the cell cycle of HCT-8/5-FU was reset, with a significant increased proportion in G2/M phase (p = 0.02) (Figure 2F). Together, our results showed that HCT-8/5-FU slows down cell proliferation rate by reprogramming cell cycle, thereby increasing tolerance to 5-FU.

We investigated the total SUMOylation modification profiling of cell global proteins in HCT-8/5-FU and HCT-8 cells. To exclude cell culture condition difference between HCT-8 and HCT-8/5-FU, HCT-8/5-FU cells were cultured in RPMI-1640 for 1 week without 5-FU supplement, proteins were collected for Western blot analysis. The level of whole SUMO1-modified proteins in HCT-8/5-FU was lower than that in parental HCT-8 cells, while the level of SUMO2/3-modified proteins was elevated in HCT-8/5-FU (Figure 3A). The decrease in the level of SUMO1-modified proteins was consistent with the expression downregulation of total proteins caused by 5-FU exposure in HCT116 and HT29 cells, indicating that the level of some SUMO1- modified proteins in HCT-8/5-FU continued to decrease during continuous induction of 5-FU. In contrast, the level of global SUMO2/3-modified proteins was upregulated under continuous 5-FU induction in HCT-8/5-FU cells.

Similarly, HCT-8/5-FU and HCT-8 cells were re-treated with different concentrations of 5-FU for 24 h to validate the SUMOylation level. In parental HCT-8 cells, 20 µM 5-FU treatment downregulated the level of global SUMO1-modified proteins and SUMO2/3-modified proteins. However, re-treatment with 5-FU in HCT-8/5-FU did not downregulate the level of SUMO1-modified proteins, but instead, it increased the level of SUMO2/3-modified proteins in a concentration-dependent manner. Moreover, the protein expression levels of UBC9 and SAE1 in HCT-8/5-FU were gradually increased in a concentration-dependent manner under 5-FU treatment.

On the contrary, 20 µM 5-FU treatment in HCT-8 cells reduced the expression of UBC9 and SAE1 (Figure 3B). In addition, we measured the mRNA expression of UBC9, SAE1 and SAE2 in HCT-8 and HCT-8/5-FU cells with or without 5-FU exposure. In HCT-8 cells, mRNA expression of UBC9, SAE1 and SAE2 showed no changes under 5-FU treatment (Figure 3C). While, in 5-FU-resistant HCT-8/5-FU cells, 5-FU treatment significantly improved the expression of UBC9 with 3-fold increase and expression of SAE1 with 2-fold enhancement, but showed little influence on the mRNA expression of SAE2 (Figure 3D). Our results indicated that increase of the SUMO2/3-modified proteins under 5-FU treatment is helpful for acquisition of 5-FU resistance in HCT-8/5-FU cells.

To uncover the role of SUMOylation in 5-FU resistance, we applied an inhibitor of SUMOylation, ML-792, to inhibit the global proteins SUMOylation modification and to observe the changes of cell tolerance to 5-FU. Firstly, we confirmed the inhibitory effect of ML-792 on CRC cells, and found that 0.1 µM ML-792 began to inhibit the conjunction of SUMO1 and SUMO2/3 on substrates (Figure 4A). 0.1 µM ML-792 greatly downregulated the IC₅₀ of 5-FU in HCT-8 cells (IC₅₀ = 0.848 µM) (Figure 4B). In addition, we applied a combination of 100 µM 5-FU with 0 μM, 1 µM or 5 µM ML-792 to compare colony formation influence of HCT-8/5-FU cells. ML-792 in combination with 5-FU treatment significantly reduced survival ability of HCT-8/5-FU cells (Figure 4C). These results confirmed that SUMOylation plays an important positive role in 5-FU resistance and that inhibition of global proteins SUMOylation can reverse 5-FU resistance.

Upon with 5-FU exposure, the expression level of UBC9 and SAE1 in HCT-8/5-FU cells has been enhanced (as shown in Figure 3), and we further investigate the potential role of these proteins in the acquisition of resistance to 5-FU in CRC. We transfected pHA-UBC9 or pFlag-SAE1 in HCT116 cells to increase the global proteins SUMOylation and found that overexpression of UBC9 and SAE1 elevated more SUMO2/3-modified proteins than SUMO1-modified proteins (Figure 5A). Overexpression of HA-UBC9 increased the IC₅₀ of 5-FU from 8.7 µM to 14.5 µM in HCT-8 cells (Figure 5B). Moreover, the overexpression of Flag-SAE1 in HCT-8 also improved the cell viability (IC₅₀ = 11.02 µM) under 5-FU treatment compared with Flag-Control group (IC₅₀ = 5.9 µM) (Figure 5C). Next, we measured the effect of UBC9 and SAE1 on colony formation under 5-FU treatment in HCT116 cells. The results showed that overexpression of UBC9 promoted the survival number of HCT116 cells under 0.5 µM 5-FU treatment (164.2 ± 30.5 vs. 105.3 ± 24.4, p < 0.05). Similarly, overexpression of SAE1 also promoted the colony number of HCT116 cells (178.3 ± 23.5 vs. 97.4 ± 18.6, p < 0.05) (Figures 5D, E). Together, these results demonstrate that UBC9 and SAE1 are crucial for 5-FU resistance acquisition in CRC.

Analysis of the UBC9 and SAE1 promoter sequence revealed that UBC9, but not SAE1 (data not shown), contained a motif recognized by RREB1 (Figure 6A), suggesting that RREB1 may directly bind to the UBC9 promoter sequence. RREB1 is overexpressed in CRC and has been identified as a transcription factor to regulate the expression of multiple targets in CRC, including miR-143/145 (Kent et al., 2013) and ITGA7 (Li et al., 2018). The UBC9 promoter sequence was cloned into a dual luciferase reporter gene vector to construct the pGL3-UBC9 vector. pFlag-RREB1 plasmid or control plasmid pFlag-control together with pGL3-UBC9 were co-transfected into HEK293T cells to detect the expression of the reporter gene. The results showed that RREB1 promoted the expression of UBC9 (Figure 6B).

Our team previously reported the oncogenic role of RREB1 in CRC development and migration (Deng et al., 2020; Chen et al., 2021). pFlag-RREB1 plasmid and mutated SUMOylation sites plasmid pFlag-RREB1^3KR were transiently transfected into HCT116 and HEK293T cells. Our results showed that RREB1 increase the expression level of SUMO1-modifed proteins and SUMO2/3-modified proteins, while the pFlag-RREB1^3KR plasmid with mutated SUMOylation sites abolished this effect. RREB1 also promoted the expression of the enzyme UBC9 but showed little influence on SAE1 expression (Figures 6C, D), which is consistent with the result of promoter prediction in UBC9 and SAE1. We have now discovered that RREB1 also regulates the level of global protein SUMOylation.

In colony formation assay, we found that RREB1 but not RREB1^3KR significantly increased the number of clones of HCT116 under 1 µM 5-FU treatment, indicating that RREB1 could increase tolerance to 5-FU. In the presence of ML-792, RREB1 did not significantly increase the number of clones of HCT116 under 1 µM 5-FU treatment (Figure 6E). Our results demonstrate that the elevation of global protein SUMOylation modification mediated by RREB1 promotes the tolerance of CRC cells to 5-FU.