Cur was obtained from MedChemExpress (#HY-N0005/CS-1490, purity ≥98%, CAS 458-37-7). Cell culture medium DMEM/HIGH GLUCOSE (4mML-Glutamine, 4500 mg/L Glucose) was purchased from CYTIVA (SH30022.01). CCK-8 was obtained from EnoGene Cell (E1CK-000208).

Human colon cancer cell lines HT-29 and HCT-116 were obtained from the Cell Bank of the Chinese Academy of Science (Shanghai, China). All these cell lines were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin under standard culture conditions (5% CO₂, 37°C).

Cell proliferation was evaluated by the CCK8 assay. HT29 and HCT116 cells were seeded in 96-well plates at 1 × 10⁴ cells/well in a volume, and incubated overnight. Each cell line was then treated with 0, 5, 10, 20, 40 and 80 μM concentrations of CUR (dissolved in DMSO) for 24 h, the control group was treated with DMSO solvent corresponding to the experimental group. After treatment, 10 μL CCK8 reagent was added to a 96-well plate and incubated at 37 °C for 20 min, and then the absorbance was measured at a wavelength of 450 nm using an enzyme marker (M200Pro). IC50 values were obtained by nonlinear regression curve fitting analysis using GraphPad Prism 5.0.

Five thousand HT29 and HCT116 cells were inoculated in 6-well plates, respectively, and divided into control and CUR (30 μM) groups and cultured for 14 days until visible colonies appeared. The control group was treated with DMSO solvent of the same concentration as the experimental group. Then the colonies were stained with crystal violet for 20 min. The colony count was calculated using ImageJ software.

HT29 and HCT116 cells (80 × 10⁴ cells/well) were inoculated in 6-well plates and cultured for 24 h. CUR (30 μM) was added to the experimental group before scratching and a linear scratch was formed by quickly scratching over the monolayer cell surface using a 1 mL pipette. The control group was also replaced with a medium containing the same concentration of DMSO as the experimental group before scratches. Cell migration was observed under the microscope (EVOS XL Core) at 24 h and 48h, and wound healing was observed by comparing micrographs of the experimental and control groups at different times. The gap area was analyzed using ImageJ software.

HCT116 cells were treated with CUR (30 μM) for 24 h. Total RNA was then extracted from cells in the administered and unadministered groups, respectively, using Trizol reagent (Sangon Biotech, B511311-0100) and reverse transcribed with a reverse transcription kit (TaKaRa, Cat# RR047A). Quantitative RT-PCR, using the FastStart Essential DNA Green Master (Roche Diagnostics GmbH, Mannheim, Germany) was performed on Qtower2.2 equipment. Gene expression was calculated using the 2^−ΔΔCT method, with GAPDH as an internal references. Primers were synthesized as follows,

GAPDH forward: 5′-TGA​CTT​CAA​CAG​CGA​CAC​CCA-3′, and.

GAPDH reverse: 5′-CAC​CCT​GTT​GCT​GTA​GCC​AAA-3′; ARHGEF12 forward: 5′-CTA​TCA​CCG​ACA​GAT​AGC​TCC​TCC-3′, and.

ARHGEF12 reverse: 5′-CGC​TGA​ACA​AGA​CCA​TAT​ATC​TCG; APAF1 forward: 5′-CAA​AGG​CTT​GGC​TCA​TGG​TTG​ACA-3′, and.

APAF1 reverse: 5′-ATG​ATG​TAG​GAT​GTC​TTG​ATG​TCC-3′; VHL forward: 5′-CAG​CTA​CCG​AGT​CCT​CAT​GAC​T-3′, and.

VHL reverse: 5′-AGC​AGG​CAG​GTA​AGT​CAA​TTT​C-3′; CEBPA forward: 5′-CCG​GAT​CTC​GAG​GCT​TGC​CCG​A-3′, and.

CEBPA reverse: 5′-TCC​TCG​CAG​GGA​GAA​GCC​ACC​G-3′; CASP8 forward: 5′-GGA​GCA​TCT​GCT​GTC​TGA​GCA​G-3′, and.

CASP8 reverse: 5′-CAT​AAA​GAT​TTC​TGC​TGA​AGT​C-3′.

To investigate the effect of CUR on gene expression in colon cancer cells, HCT116 cells (30 × 10⁴ cells/well) were inoculated in 6-well plates and cultured for 24 h. The experiment was then divided into CUR (30 μM) treated and untreated groups for another 24 h. Total RNA was extracted using Trizol reagent for high-throughput sequencing. The RNA concentration and quality were determined using Nanodrop 2000 (Thermo Scientific) for library preparation. RNA-seq analysis was performed using Tophat2 (http://ccb.jhu.edu/software/tophat) to compare the sequencing reads to the human reference genome hg38. Reads were calculated using featureCounts (http://subread.sourceforge.net). Differentially expressed genes (DEGs) and statistical analyses were performed with DESeq2 (version 3.12) in R (version 4.0) (fold change >1.5, p < 0.05). Heat maps were created with Complex Heatmap (Bioconductor Project).

ARHGEF12–RhoA crystal complexes (PDB code:1X86) was obtained from the Protein Data Bank (http://www. rcsb. org/). Discovery Studio 2021 was used to perform molecular docking analysis. Before docking, CUR was prepared using “Prepare Ligands” module and CHARM force field was used for minimization, generating ten conformations. The protein was prepared using “Protein Preparation” module, allowing for the addition of hydrogen atoms and the deletion of unnecessary water. Subsequently, the proteins were optimized and minimized. Ligand binding sites are defined using the “Receptor-Ligand Pharmacophore Generation” module. The docking results were evaluated using hydrogen bond interactions and binding mode. The interaction 2D diagram of the CUR with residues of receptor was views in the “View Interaction” module. All structural figures were generated using PyMol 3.7.

All experiments were conducted in triplicate (n = 3) and the data are presented as the mean ± SEM. All statistical analyses were processed with GraphPad Prism 5.0. Differences of unpaired comparisons between two groups were analyzed using the ANOVA. p-value <0.05 was considered statistically significant.

Figure 1A shows the chemical structure of CUR. Firstly, we calculated the IC50 values of HCT116 and HT29 cells after CUR treatment by CCK8 assay. As a result, the corresponding IC50 at 24 h was 27.21 μM and 37.76 μM in HCT116 and HT29 cells, respectively (Figure 1B). Further, we investigated the effects of CUR on the viability of colon cancer cells. Cell viability was measured by CCK8 assay after treatment with CUR at concentrations of 5, 10, 20, 40 and 80 μM for 24 h. Figure 1C, D showed that CUR reduced cell proliferation in a dose-dependent manner. As the dose of CUR increased, the number of cells gradually decreased and the morphology shrank (Figure 1E, F).

Next, we used the CCK8 method to detect the effect of CUR on the proliferation of HCT116 and HT29 cells by CCK8 assay. The results in Figure 2A–D show that cell proliferation decreased in a time-dependent manner with increasing time of CUR treatment, and cell morphology was crinkled.Elucidation of the effects of CUR on colon cancer cell proliferation and migration by colony formation and wound healing assays. In the colony formation assay, the observations showed that the colony formation capacity of HCT116 and HT29 cells was reduced after treatment with CUR at 30 μM, indicated by significantly decreased colony numbers (Figure 2E-H). Wound healing assays showed that CUR treatment inhibited cell migration from 0 to 48 h (Figure 2I-L). The migration rates of the CUR treatment group in HCT116 and HT29 cells were 39.9% and 21.79% compared with those of the corresponding control group after CUR treatment for 24 h and 25.89% and 27.64% of the control group after CUR treatment for 48 h (p < 0.0001), respectively.

To investigate the molecular mechanism of CUR inhibition of colon cancer cell proliferation and migration, we treated HCT116 cells with CUR (30 μM) for 24 h, and collected HCT116 cells from the control and CUR groups separately for mRNA high-throughput sequencing analysis to determine the curcumin-related genes. The results are shown in Figure 3A, 3,505 genes were upregulated and 3,564 genes were downregulated. Cluster analysis by the top 100 differentially expressed genes in Figure 3B showed that these differentially expressed genes clustered significantly between the control and CUR group cell samples. KEGG pathway enrichment analysis showed that these differentially expressed genes after being acted upon by CUR were enriched in the cancer pathway in Figure 3C.

TSG is a key gene involved in DNA damage repair, suppression of cell mitosis, induction of apoptosis and prevention of metastasis. Hence, downregulation of TSG will lead to cancer development and progression. Therefore, re-upregulation of TSGs that are downregulated in cancers may prevent cancer progression (Wang et al., 2018). To explore whether CUR could reactivate the TSGs that were downregulated in colon cancer, we intersected the 1,217 tumor suppressor genes that were downregulated in colon cancer from TCGA database (https://bioinfo.uth.edu/TSGene/) with 135 genes that were upregulated in cancer pathways in HCT116 under the treatment of CUR, 37 TSGs that were reactivated under the action of CUR were obtained, including ARHGEF12, APAF1, VHL, CEBPA, CASP8 et al (Figure 4A). The results of KEGG and GO analysis are shown in Figure 4B, C. These 37 reactivated TSGs are mainly involved in the cancer pathways, which is consistent with the results of KEGG analysis of the first 100 upregulated genes in Figure 3C. In addition, GO profiling suggested that these 37 upregulated TSGs negatively regulated the transcription of RNA polymerase II promoter and positively regulated the apoptotic process. (Figure 4C), suggesting activation of these genes can suppress cell growth and proliferation.

To further verify that TSGs were indeed reactivated by CUR, we randomly selected five of the 37 upregulated TSGs: ARHGEF12, APAF1, VHL, CEBPA, and CASP8 for qPCR assay. The results showed that the expression of these five TSGs was upregulated in both HCT116 and HT29 cells after CUR (60 μM) treatment, with significant differences in ARHGEF12, APAF1, VHL three genes in HCT116 and ARHGEF12, APAF1, VHL, CASP8 four genes in HT29 (Figure 5A, B). Among them, ARHGEF12 and APAF1 genes, which were most significantly upregulated by CUR, were preferentially selected as candidate targets for CUR in colon cancer cells.

Considering that CUR can significantly reactivate TSGs ARHGEF12 and APAF1, to further investigate whether upregulated ARHGEF12 and APAF1 expression in colon cancer correlates with patient prognosis, we analyzed colon cancer subtypes using Kaplan-Meier survival curves. We found in Figure 5C, D that high expression of ARHGEF12 and APAF1 was significantly correlated with a good prognosis for patients of colon cancer. Further, we queried the pan-oncogene expression profiles of ARHGEF12 and APAF1 in the TCGA (https://bioinfo.uth.edu/TSGene/) database using the UALCAN database (http://ualcan.path.uab.edu/), and as shown in Figure 5E, F, ARHGEF12 and APAF1 were significantly downregulated in colon cancer and many other types of cancers. In summary, both TSGs ARHGEF12 and APAF1 expression were downregulated in colon cancer and could be reactivated by CUR, and upregulated ARHGEF12 and APAF1 may be associated with favorable prognosis of patients.

Since ARHGEF12 has not been reported as a therapeutic target for colon cancer, we aimed to investigate the molecular mechanism by which CUR inhibits colon cancer by acting on ARHGEF12. Based on the reported researches, ARHGEF12 binds to RhoA to shape a functional complex, thereby enhancing cancer cell migration and invasion. (Lacoste et al., 2012; Ghanem et al., 2022). We speculate that CUR may block the binding of ARHGEF12-RhoA complex to suppress colon cancer cells, providing a new therapeutic strategy for colon cancer treatment.

To verify whether CUR interacts with ARHGEF12–RhoA, we performed molecular docking to analyze its affinity and binding mode. In the structure of ARHGEF12–RhoA complexes, ARG 923 of ARHGEF12 formed multiple salt bridge interactions with ASP 45 and GLU 54 of RhoA, which was important for RhoA to be selected as a substrate (Aggarwal and Sung, 2009). In our study, we found that CUR occupied the cavity at the interface between ARHGEF12 and RhoA, indicating that CUR effectively inhibited interactions between ARHGEF12 and RhoA (Figure 6A, B). In the binding pocket, ARG 923 formed a direct hydrogen bond with the phenolic hydroxyl group of CUR at 2.8 Å. Additionally, ARG 804 of ARHGEF12 and GLU 40 of RhoA both formed stronger hydrogen bonds with CUR at 2.7 Å and 2.9 Å, respectively. ARG 936 of ARHGEF12 and TYR 42 of RhoA formed weaker hydrogen bonds with CUR at 4.0 Å and 3.7 Å. Moreover, ARG 922 of ARHGEF12 was packed in the hydrophobic core of CUR, which contributed to stabilizing the molecule at the binding site (Figure 6C, D). This suggests that CUR has a high affinity for ARHGEF12 and can effectively intervene in the interaction between ARHGEF12 and RhoA. Therefore, we hypothesized that CUR reactivated the ARHGEF12 TSG to block colon cancer cell proliferation and migration via cancer pathways, whereas CUR might exert an inhibitory effect on invasion and migration by blocking the binding of ARHGEF12 to RhoA.