Human CRC cell line HCT116 and SW620 cells were kind gifts from Prof. Yunlong Lei (Molecular Medicine and Cancer Research Center, Chongqing Medical University). Cells were maintained in DMEM medium (Hyclone) containing 10% fetal calf serum (Hyclone), penicillin and streptomycin at 37°C in a humidified atmosphere with 5% CO₂. The cells were treated without or with H₂O₂ in media, which were changed daily. Cells were pretreated with antioxidant 5 mM or 10 mM N-acetylcysteine (NAC) for 30 min before addition of H₂O₂.

To clone the CXCL14 cDNA, we isolated total RNAs from the HCT116 cell line using Trizol method. The RNA samples were dissolved with TE buffer, and then quantified by measuring OD value. The extracted RNA samples were further reverse transcribed into cDNA, which would be used as PCR template. The primers used for CXCL14 amplification were synthesized (Invitrogen). The primer sequences for amplification were as follows: cxcl14, forward 5′-GAT​CCC​CGC​CAC​CAT​GTC​CCT​GCT​CCC​ACG​C-3′, reverse 5′-TCG​AGC​TAT​TCT​TCG​TAG​ACC​CTG​CGC-3’; gapdh, forward 5′-ACC​TGA​CCT​GCC​GTC​TAG​AA-3′, reverse 5′-TCC​ACC​ACC​CTG​TTG​CTG​TA-3’. For CXCL14 stable expression in HCT116 cells, the recombinant plasmid of EX-CXCL14-LV203 and the control plasmid EX-NEG-LV203 were transfected into 293T cells with lentivirus packaging system, separately. Then, the concentrated virus particles were collected and the HCT116 cells were infected with the viruses in their logarithmic growth phase. The stable transfectants were selected in the presence of puromycin and identified by real-time PCR and western blotting. The primers for real-time PCR were as follows: cxcl14, forward 5′-CTA​CAG​CGA​CGT​GAA​GAA​GC-3′, reverse 5′-TTC​TCG​TTC​CAG​GCG​TTG​TA-3’. The construction of HCT116 cells lacking CXCL14 was also accomplished via lentiviral-mediated transduction with a scrambled shNCs or verified CXCL14-specific shRNA sequence encoded in LVRH1GP. A 21-mer shRNA expressing vector targeting CXCL14 (shCXCL14) and its scrambled sequence-expressing vector as a negative control (shNC) were synthesized (Invitrogen). The sequence for CXCL14 shRNA was forward 5′-GAT​CCG​CAC​CAA​GCG​CTT​CAT​CAA​TTC​AAG​AGA​TTG​ATG​AAG​CGC​TTG​GTG​CTT​TTT​TGG-3′, and reverse 5′-AAT​TCC​AAA​AAA​GCA​CCA​AGC​GCT​TCA​TCA​ATC​TCT​TGA​ATT​GAT​GAA​GCG​CTT​GGT​GCG-3’.

Intracellular ROS were detected by staining cells with 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H₂DCF-DA) (Genmed Scientifics Inc., Burlington, MA, United States). Cells after firm adhesion to the wall were treated with H₂O₂ for 6 h. After trypsin digestion and centrifugation, supernatant was removed. 10 μmol/L DCFH-DA dye diluted with serum-free medium was added to cells for incubation at 37°C for 20 min. The CM-H₂DCF-DA signal was then analyzed with the BD Biosciences FACSCalibur Flow Cytometer (Mountain View, CA).

Cell proliferation was determined by using the Cell Counting Kit-8 (CCK-8) Assay (Boster). The cancer cells (5000 cells per well) were seeded in the 96-well plates and treated with different concentrations of H₂O₂ for 48 h. The cells were incubated with 10 μL of CCK-8 per well and the optical density (OD) value at 450 nm was determined by microplate reader. The experiment was repeated in triplicate. The formula of promotion rate of cell proliferation (P%) is as follows: P % = O D ( s a m p l e ) − O D ( c o n t r o l ) O D ( c o n t r o l ) − O D ( b l a n k ) × 100 % (1)

For colony formation assay, the cancer cells (1000 cells per well) were resuspended in DMEM supplemented with 10% FBS without or with H₂O₂ in 6-well plates. The cultures were maintained for 14 days and stained with 0.1% crystal violet and estimated under the microscope. Each experiment was performed in triplicate.

Confluent HCT116 cells were wounded with a micropipette tip (200 μL) and immediately placed in 1% serum-containing medium supplemented with or without H₂O₂. Bright-field images of wounded monolayers were obtained immediately after wounding (0 h) and at various times thereafter as indicated. The extent of wound closure was quantified by obtaining three wound measurements for each of three random fields (× 100) per wound, and all wound conditions were performed in triplicate.

The cell migration assay was performed in chambers with 8-μm-porosity polycarbonate filter membranes according to the manufacturer’s instructions (Becton Dickinson Labware). Cells (1 × 10⁵ cells per well) treated without or with H₂O₂ in serum-free medium were added to the upper chamber and incubated for 20 h. The bottom chamber was prepared with 20% FBS as a chemoattractant. Nonmigrating cells on the upper surface of the membrane were removed. Cells that invaded to the lower surface were fixed with 4% paraformaldehyde stained with crystal violet, and photographed under a microscope (Eclipse TS100, Nikon) at × 20 magnification.

Both adherent and floating cells were harvested. All cell protein was extracted in RIPA lysis buffer (50 mM Tris, 1.0 mM EDTA, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 1% sodium deoxycholate, 1 mM PMSF). The concentrations of protein were quantified by using the DC Protein Assay Kit (500–0121; Bio-Rad). The cellular lysates were resolved on SDS-PAGE and electrophoretically transferred to polyvinylidene difluoride membranes. The membranes were then blocked with a buffer [composed of 10 mM Tris (Zhu et al., 2018), 150 mM NaCl, 0.1% Tween 20 and 5% bovine serum albumin], and then incubated with the primary antibodies (CXCL14, diluted 1:400, Proteintech, 10468-1-AP; Cyclin A1, diluted 1:1000–2000, Abcam, ab53699; Cyclin B1, diluted 1:1000–2000, Abcam, ab215436; CDK1, diluted 1:1000–2000, Abcam, ab265590; CDK2, diluted 1:1000–2000, Abcam, ab101682; E-cadherin, diluted 1:1000–2000, Abcam, ab238099; N-cadherin, diluted 1:1000–2000, Abcam, ab207608; vimentin, diluted 1:1000–2000, Abcam, ab92547; ERK, diluted 1:1000–2000, Abcam, ab184699; p-ERK, diluted 1:1000–2000, Abcam, ab214036; and GAPDH, diluted 1:1000–2000, Abcam, ab181602) at 4°C for overnight. Then, the membranes were washed and treated with appropriate secondary antibodies conjugated with horseradish peroxidase for (goat anti-rabbit IgG H&L, Abcam, ab7090; goat anti-mouse IgG H&L, Abcam, ab7068) 2 h. The immunoreactivities were determined by using enhanced chemiluminescence reagents (WBKLS0500; Millipore) (Wang et al., 2011). The GAPDH protein was used as an internal control.

Statistical values were defined by using an unpaired Student’s t-test. Differences among multiple groups were assessed by one-way ANOVA analysis. p < 0.05 was considered to be statistically significant.

Recent studies have demonstrated that cancer cells can be characterized by elevated ROS levels when compared with the normal cells (Maurya and Vinayak, 2015; Prasad et al., 2017; Moloney and Cotter, 2018). Aberrant ROS production is associated with tumor initiation and progression (Yang et al., 2016; Ghanbari Movahed et al., 2019). To determine the role of ROS in CRC cell proliferation and migration, we applied exogenous H₂O₂ to induce cellular ROS stress in HCT116 cells and SW620 cells. As illustrated in Figures 1A,B, the cells treated with the exogenous H₂O₂ showed an endogenous increase in the oxidative stress level when compared with the control cells, as measured by flow cytometry.

Cell proliferation was determined by the CCK-8 assay. As shown in Figures 1C,D, treatment with different doses of H₂O₂ induced different biological outcomes. H₂O₂ increased HCT116 cell proliferation at a concentration of <25 μM when compared with the control cells. Adversely, H₂O₂ decreased the cell proliferation at concentration of >25 μM (Figure 1C). We found a similar phenomenon in SW620 cells (Figure 1D). As expected, the cells treated with H₂O₂ formed a greater number of colonies when compared with the control cells (Figures 1E–H). To further characterize the effect of H₂O₂ on the cell cycle progression, we examined the DNA content of HCT116 cells and SW620 cells in various phases of the cell cycle, as measured by flow cytometry (Figures 1I–L). Results showed that H₂O₂ increased the cell percentage in the S phase and G2/M phase and decreased the cell percentage in the G1 phase in HCT116 cells (Figures 1I,J), and H₂O₂ increased the cell percentage in the S phase in SW620 cells (Figures 1K,L).

The migratory capacity of HCT116 cells and SW620 cells treated with or without H₂O₂ was estimated by the wound healing assay. In the wound healing assay, the number of H₂O₂-treated HCT116 and SW620 cells that migrated into the wound area was much greater than the control cells (Figures 2A–D), which suggests the role of ROS in mediating the closure of cell wounds. The cells in the wound area may be a direct reflection of migration, or partially due to cell proliferation. To circumvent this difficulty, we further performed cell migration assays in 24-well transwell chambers. Results showed that ROS significantly increased the migration capacity of the HCT116 cells and SW620 cells (Figures 2E–H). Taken together, these data strongly suggest that ROS is directly involved in mediating the proliferation and migration of CRC cells.

In mammalian cells, the cell-cycle progression is highly controlled by cyclins and cyclin-dependent kinases (CDKs). The activity of cyclin A1/CDK2 complex is essential for the cell-cycle G1/S phase transition, and the cyclin B1/CDK1 complex plays the fundamental role in cell-cycle G2/M transition (Malumbres and Barbacid, 2009; Xie et al., 2020). We noted that H₂O₂ could up-regulate the expression levels of cyclin A1, cyclin B1, CDK1, and CDK2 in a time-dependent manner in HCT116 cells (Figures 3A,B). H₂O₂ could up-regulated the expression of cyclin A1, CDK2 and survivin in SW620 cells, which could be partially suppressed by the antioxidant NAC pretreatment (Figures 3C,D). We also noted that H₂O₂ could up-regulate the expression of proliferating cell nuclear antigen (PCNA) in SW620 cells (Figures 3E,F). These data indicated increased growth capacity for the CRC cells in the presence of H₂O₂ through regulation of the cell cycle progression and cell proliferation.

The epithelial to mesenchymal transition (EMT) endows cancer cells with the properties of invasion and metastasis (Yuan et al., 2020). In solid tumors, EMT occurs at the invasive front and induces migratory cells with downregulated expression of epithelial markers E-cadherin and upregulated expression of the mesenchymal markers vimentin and N-cadherin (Nieto et al., 2016). By immunoblotting, we observed a substantial decrease in the expression level of E-cadherin and an increase in the expression levels of vimentin and N-cadherin in response to exogenous H₂O₂ in HCT116 cells (Figures 3G,H). We also noted that vimentin expression was down-regulated after exposure to H₂O₂, which could be partially restored by the antioxidant NAC in SW620 cells (Figures 3C,D).

Previous studies have shown that ROS can act as signal transduction molecules and regulate the expression of oncogenic chemokines (Wang et al., 2017). Based on previous reports, we performed immunoblotting to examine the expression level of chemokine CXCL14 after exposure to H₂O₂ in CRC cells. We observed that the CXCL14 expression was significantly elevated in HCT116 cells treated with H₂O₂ when compared with the control cells and in a time-dependent manner (Figures 4A,B). We then used the antioxidant NAC to explore whether the CXCL14 expression was affected. Results showed that NAC pretreatment partially suppressed the expression levels of CXCL14, indicating that CXCL14 expression might be regulated by ROS in HCT116 cells (Figures 4A,B). Meanwhile, we found that the expression levels of CXCL14 mRNA and protein were elevated after exposure of H₂O₂ in a time-dependent way in SW620 cells (Figures 4C,D), which could be partially suppressed by the antioxidant NAC pretreatment (Figures 4E,F).

To gain further insights into the role of CXCL14 in ROS-induced oncogenesis, we first constructed a stably CXCL14-expressing HCT116 cell line (HCT116/CXCL14), as shown in Figures 5A–C. The proliferative and migratory capacity of HCT116/CXCL14 cells was determined by using the colony formation assay (Figure 5D), wound healing assay (Figure 5E) and migration assay (Figure 5F). As compared with the control cells (HCT116/Control), HCT116/CXCL14 cells exhibited stronger proliferative and migratory capacities, suggesting that CXCL14 might play an important role in CRC progression, which was consistent with previous data (Zeng et al., 2013a).

We further examined the expression levels of cyclin A1, cyclin B1, CDK1, and CDK2, and found that CXCL14 could up-regulate the expression levels of cell cycle-related protein (Figures 5G,H). Meanwhile, we also noted that the HCT116/CXCL14 cells were characterized with downregulation of epithelial markers E-cadherin and upregulation of mesenchymal markers vimentin and N-cadherin (Figures 5G,I), suggesting that CXCL14 may promote the CRC progression through the regulation the EMT process.

We also constructed a stably CXCL14-expressing SW620 cell line (SW620/CXCL14), as shown in Figures 5J–L. We found that the expression levels of CDK2, cyclin B1, and vimentin could be up-regulated by CXCL14 (Figures 5M,N).

Taken together, these data strongly suggests that chemokine CXCL14 may be directly involved in CRC cell proliferation and migration.

The typical extracellular-regulated kinase (ERK) cascade is a highly conserved signaling pathway that bridges extracellular signal molecules and intracellular diverse executor proteins, modulating various physiological or pathological processes, including tumor cell proliferation and migration (Maik-Rachline et al., 2019). Aberrant ROS can induce the activation of ERK through different signal cascade reactions in different cancers (Su et al., 2019). In the present study, we found that ROS treatment resulted in a greater accumulation of phosphorylated ERK (p-ERK) in HCT116 cells in a time-dependent manner, which could be partially restored by the antioxidant NAC pretreatment (Figures 6A,B). Previous evidence demonstrated that chemokines could induce cancer development through the activation of Erk1/2 (Zhao et al., 2017). To explore whether chemokine CXCL14 could mediate the level of p-ERK in CRC cells, we compared the levels of p-ERK between HCT116/CXCL14 cells and HCT116/Control cells. The results revealed that the level of p-ERK was markedly higher in the HCT116/CXCL14 cells than that in the HCT116/Control cells (Figures 6C,D). Furthermore, CXCL14-deficiency markedly inhibited the phosphorylation of ERK when compared with the control cells (i.e., scrambled shRNA), as shown in Figures 6E,F. More interestingly, H₂O₂ treatment could partially restore the expression levels of CXCL14 and p-ERK in HCT116/shCXCL14 cells (Figures 6E,F).

Taken together, these data suggest that chemokine CXCL14 may stimulate the expression of ROS-induced p-ERK, thereby promoting the CRC progression.