Baicalin suppresses colorectal cancer cell proliferation, potentially via ARRDC4: Bioinformatics and experimental analysis

2.1 Chemicals and reagents

Baicalin (MW:446.37, Fig. A.1, ≥98% purity) was purchased from Shanghai yuanye Bio-Technology Co., Ltd (catalog no.:N15GB167969). Caspase-Glo® 3/7 Assay was obtained from Promega (Madison, WI, USA, #G8091). Antibodies against ABP1 (#16338–1-AP), DAPP1 (#14722–1-AP), NLRP3 (#19771–1-AP), GSTA4 (17271–1-AP) and LCN2 (26991–1-AP) were purchased from Proteintech (Chicago, USA). The antibody against GGT5 (#ab283267) was purchased from Abcam (USA). The antibody against Claudin-2 (#48120) was purchased from Cell Signaling Technology (Boston, MA, USA). ANKRD63 polyclonal antibody (#PA5-49152) and ARRDC4 polyclonal antibody (# PA5-36518) were purchased from Invitrogen (Carlsbad, USA). Cell counting kit-8 kit (#CK04, Dojindo, Kumamoto, Japan), apoptosis kit (#88–8007, eBioscience, California, USA), BCA protein assay kit (Beyotime, Shanghai, China, #P0010S), radioimmunoprecipitation assay (RIPA) buffer (Beyotime, #P0013B), endoFree midi plasmid kit (#DP118-2, TIANGEN, Beijing, China) and enhanced chemiluminescent (ECL) plus reagents (#M3121/1859022, Thermo Fisher Scientific, Waltham, MA, USA) were used in the study.

2.2 Cell culture and baicalin treament

The two human CRC cell lines HCT-116 and HT-29 were obtained from the Cell Bank of the Chinese Academy of Science Type Culture Collection (Shanghai, China). RKO cells were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). The cell lines were all cultured in RPMI 1640 media (Thermo Fisher Scientific,Waltham, MA, USA), McCoy’s 5A and Eagle’s Minimum Essential Medium (Sigma Aldrich, Germany), respectively. All mediums were supplemented with 10% fetal bovine serum (#VS500T, Ausbian, Australia), 100 μg/mL streptomycin (Gibco, CA, USA), and 100 units/mL penicillin (Gibco, CA, USA). Cells were all incubated in a humidified incubator under a 5% CO₂ atmosphere at 37 °C. The indicated concentrations of baicalin were achieved by adding appropriate amounts of stock baicalin solution to the culture medium. The cells then incubated for a certain time. The dimethyl sulfoxide solution that did not contain baicalin was used the blank reagent.

2.3 Cell viability and drug sensitivity assays

Cells were seeded into 96 wells (4 × 10³ cells/well) and incubated for 24 h at 37 °C for adhesion, then treated with baicalin at different concentrations or physiological saline (control) for 48 h. Then, the CCK-8 assay was performed to evaluate cell viability (%). A non-linear regression model was used to calculate half maximal (50%) inhibitory concentration (IC₅₀) values with a sigmoidal dose–response curve in GraphPad Prism 8. Three independent experiments were performed.

2.4 RNA sequencing and quantitative real-time PCR (qRT-PCR)

Total RNA was extracted using Trizol reagent from baicalin-treated cells and was reversely transcribed into cDNA with a Fast King RT kit. The primers are listed in Table 2. RNA purity was confirmed using a NanoPhotometer® spectrophotometer (IMPLEN, CA, USA). Next, we screened the RNA to identify DEGs with an Agilent 2100 bioanalyzer (Agilent, Santa Clara, CA, USA). Three ARRDC4 knockdown samples and three negative control samples were used for microarray analysis using the human GeneChip PrimeView array. The cDNA library construction and purification as well as transcriptome sequencing were conducted based on the Shanghai Genechem Company’s instructions. The gene or exon expression levels were normalized as the number of reads per kilobase per million mapped reads (RPKM) (Mortazavi et al., 2008). DEGs were defined as having a P value below 0.05 and an expression fold change above 2. The GoTaq® qPCR Master Mix was applied for qRT-PCR. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control.

2.5 Construction of ARRDC4 gene recombinant lentiviral vectors and transfection

The cell lines with stably knocked-down genes were obtained as previously described (Dave et al., 2014). The knockdown efficiency was confirmed by real-time PCR and western blot, and the shRNA sequences were shown in Table 1. The interference sequence-containing single-strand DNA oligonucleotide was then created, and annealing was used to create the double-strand DNA. After digestion by an enzyme, the two ends of the oligonucleotide were then immediately joined to the lentiviral vector. The prepared Escherichia coli cells received the ligated products. The positive recombinants were then located by PCR, and their sequences were used to verify them and remove their plasmids. Lentiviral vector DNA and packaging vectors were transfected in 293 T cells (Chinese Academy of Science Type Culture Collection, Shanghai, China). We purchased the lentivirus with the ARRDC4 gene overexpression vector (Shanghai Genechem Co., Ltd., Shanghai, China). The lentiviral vector system was composed of GV493, a pHelper 1.0 vector and a pHelper 2.0 vector prior to packaging, and the full-length ARRDC4 gene was encoded into the GV493 vector after being labeled with the enhanced green fluorescence protein (eGFP). The empty GV493 lentiviral vector was used as the shRNA control (shCtrl). After 48 h of culture, supernatants containing the lentiviruses, including shARRDC4 and shCtrl, were harvested and purified. Three days later, GFP expression was quantified by fluorescence microscopy (BX51; Olympus, Tokyo, Japan). Cells with stable eGFP expression were harvested for the following analysis.

2.6 High-content screening and cell growth curve analysis

Using HCS, we discovered the effects of 19 putative target genes on the proliferation of HCT-116 cells were discovered (Supplementary Material). In brief, HCT-116 cells were seeded at 2000 cells/well in 96-well plates and transfected with shRNA lentivirus targeting candidate target genes or negative control lentivirus. The expression of GFP was quantified by fluorescence microscopy. Cells were collected for further experiments when they reached 80% confluence. The system achieves high throughput screening through an automatic Celigo cell imaging analyzer (Nexcelom, USA). In the HCS proliferation assay, the detection target was cells expressing GFP after infection with the virus. The Celigo cell imaging analyzer recognizes cells with green fluorescence and records pictures. Then, the images were analyzed and processed by software to calculate the number of cells in various groups in the orifice plate. After 5 days of continuous reading, the cell growth curves were drawn. A fold change of the proliferation ratio of two or more suggested slowing-down cell proliferation, which was enough to examine the effect of RNAi lentivirus infection on cell proliferation (Mandavilli et al., 2018).

2.7 Western blot

After harvesting baicalin-treated cells, cell pellets were washed with ice-cold phosphate-buffered saline (PBS) and lyzed with RIPA buffer. Total proteins were quantified using a bicinchoninic acid (BCA) assay kit. After the total lysates were treated with a sodium dodecyl sulfate (SDS)-loading buffer at 100 °C, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was performed. The proteins were then transferred to polyvinylidene difluoride (PVDF) membranes, which were blocked with 5% dried skimmed milk, followed by overnight incubation with the specific primary antibodies. Finally, the HRP-conjugated secondary antibody were added and protein expression was quantified using the ECL plus reagents.

2.8 Apoptosis assay

Before the apoptosis assay, cells were seeded in a six-well plate and treated with baicalin for 48 h. The cells (attached and floating) were collected by centrifugation. To perform the apoptosis assay, these cells were resuspended in PBS and adjusted to 1 × 106 cells/mL. Next, 10 μL of Annexin V-APC staining solution (eBioscience, California, USA) was added to the 300 μL cell suspension for staining and the samples were incubated for 10–15 min at room temperature in the dark. Finally, the cells were counted and analyzed using a FAC Sort Flow Cytometer (BD, CA, USA).

2.9 Caspase-Glo 3/7 assay

The caspase activity was detected using a Caspase-Glo® 3/7 Assay Kit (Promega, WI, USA) (Li et al., 2021a, 2021b, 2021c). The cells were seeded in 96-well plates, incubated overnight and then for 24 h with the compounds at various concentrations according to the manufacturer’s instructions. The luminescent signal resulting from caspase 3/7 activity was detected with an EnVision® Multilabel Reader (PerkinElmer, Thermo Fisher Scientific, Shanghai, China).

2.10 In vivo tumor growth assay

Female BALB/c nude mice (4 weeks old, 25–30 g) from Shanghai Lingchang Biotechnology Co., Ltd (SCXK(hu)2018–0003), were maintained under a specific pathogen -free environment according to the Regulations for the Care and Use of Laboratory Animals. All experiments were in accordance with the guidelines of China for animal care, which confirm to the internationally accepted principles in the care and use of experimental animals. This study obtained approval from the Ethics Committee of Animal Experiments of Suzhou TCM Hospital Affiliated to Nanjing University of Chinese Medicine. A xenograft model was established by subcutaneously injecting HCT-116 cells and ARRDC4 knockdown HCT-116 cells under logarithmic growth into the right flank of nude mice (1 × 10⁷/mL cells/mouse). Next, we monitored tumor growth. The treatment was initiated when the volume of the tumor reached approximately 100 mm³ (day 1). Mice were randomly divided into four groups of six mice each. The vehicle and control groups were treated with baicalin (50 or 100 mg/kg/day) for 14 consecutive days. The tumors were measured twice a week using microcalipers and the tumor volume (V) was calculated using: V = (Length × Width × Width)/2. All mice were sacrificed at the end of the experiment or when the tumor volume reached 800 mm³ to avoid unnecessary animal suffering, and their tumors were weighed and photographed.

2.11 Hematoxylin and eosin (HE) staining of subcutaneous tumor tissue sections

Subcutaneous tumor tissues were fixed in 4% polyformaldehyde, then washed thoroughly with water, dehydrated with methanol at different concentrations, and transparentized with dimethylbenzene. They were then embedded in paraffin, sectioned, and subjected to HE staining. The pathological changes of subcutaneous tumor tissues were observed by optical microscopy and photographed (×200, ×400).

2.12 Immunohistochemistry (IHC) analysis

We performed IHC on the paraffin sections of tumor tissues using a kit according to the manufacturer’s instructions. Briefly, the tissue sections (4 μm) were dehydrated and subjected to peroxidase blocking. After blocking, the sections were incubated with the Ki67 antibody for overnight incubation at 4 °C. Next, we incubated the slides with the horseradish peroxidase-conjugated secondary antibody and stained them with 3,3′-diaminobenzidine (DAB). Ki67-positive cells appeared brown under a microscope, and images were recorded. Image J image analysis software was used for semi-quantitative analysis.

2.13 Statistical analysis

All quantitative data were expressed as mean ± standard deviation and analyzed using Student’s t-test or analysis of variance (ANOVA). P-values < 0.05 indicated statistically significant differences. All statistical analyses were conducted using SPSS (version 13.0) and visualized with GraphPad Prism 8 (GraphPad Software, La Jolla, CA, USA).

3.1 Cytotoxicity of baicalin in CRC cells

To study the effect of baicalin on CRC cells, we treated three cell lines with various baicalin doses for 48 h. The calculated IC₅₀ values ranged from 0 to 1000 μM (Fig. 1B) and indicated that HCT116 cells were more sensitive to baicalin than the other lines. Baicalin markedly inhibited HCT-116 cell proliferation and reduced cell growth by more than 30% after 48 h (Fig. 1C and D). Cells treated with 90 µM baicalin were more significantly inhibited than untreated cells. Based on the pre-experiment, baicalin dose-dependently inhibited HCT-116 cell proliferation (1–40 μM). Therefore, we used baicalin concentrations of 80 µM for RNA sequencing and 10 µM for the HCS experiment.

Close

Fig. 1

The inhibitory effect of baicalin was investigated based on the selected cell line. (A-C) Using an MTT assay, the effects of baicalin on the viability of human RKO, HT-29 and HCT-116 cells treated with baicalin (0–1000 μM) for 48 h are illustrated.

Export to PPT

3.2 Identification of ARRDC4 as a critical gene promoting CRC proliferation

In order to study the underlying mechanisms of the anti-proliferation effect of baicalin on HCT-116 cells, we conducted a transcriptomic analysis. We treated HCT-116 cells with baicalin (80 μM) or negative control for 48 h, performed RNA sequencing. and identified 9601 DEGs were identified (4817 upregulated and 4784 downregulated) (Fig. 2A and 2B; Tables A1-A3). Among them, 114 had a fold change larger than 2 (Table A4). We then conducted a GO enrichment analysis on these genes to further characterize the mechanism of baicalin’s effect on CRC cell proliferation (Fig. 2C). To identify novel oncogenes, we focused on 19 downregulated genes that have not been extensively investigated for potential association with CRC.

Close

Fig. 2

Agilent mRNA microarray identified expression profiling of baicalin-treated HCT-116 cells for 48 h by using RNA sequencing technology. (A) Heat map showing gene expression profiles. Each row represents a gene and each column represents a sample. Red indicates high expression, whereas green indicates low expression. (B) Volcano plot representing differentially expressed genes between baicalin-treated and baicalin-untreated HCT-116 cells. (C) GO enrichment (BP, CC and MF) of host genes of significantly differentially expressed circRNAs. X-axis represents the enriched GO term ordered by BP, CC and MF. Y -axis indicates the number (Left) of the genes in the corresponding terms.

Export to PPT

We confirmed these DEGs by qPCR in accordance with the RNA-seq sample preparation guidelines. Except for NUDT4B, 19 genes were compatible with the sequencing data (Fig. 3).Next, we performed HCS on cells infected with a mix of three targets of lentivirus carrying 19 genes to identify genes involved in cell proliferation. Silencing each of the 19 candidate genes in HCT-116 cells revealed that silencing ANKRD63, SMIM32, and ARRDC4 strongly inhibited cancer cell growth (Fig. 4A, B and C). These results suggested that ARRDC4 might play a role in CRC development (Fig. 4D). ARRDC4 has been recently extensively studied in prostate cancer (Choueiri et al., 2015), but few studies have documented its role in CRC. Therefore, we assessed its anti-proliferation properties and the underlying mechanisms.

Close

Fig. 3

RNA-seq transcriptome profiling identifies 19 genes (except NUDT4B) as potential targets for colorectal cancer proliferation. Values were expressed as mean ± SD. Compared with the control group, ^##P < 0.01.

Export to PPT

Close

Fig. 4

High-content screening identified ARRDC4 as a critical gene in promoting colorectal cancer proliferation. (A) A total of 19 genes were selected for validation by high-content screening. (B-C) Representative cell proliferation count of high-content screening for ANKRD63, SMIM32 and ARRDC4. (D) Knockdown efficiency of shRNA targeting ANKRD63, SMIM32 and ARRDC4. Cells were infected with lentivirus containing shRNA, RNA was collected and knockdown efficiency was determined by Real-time PCR.

Export to PPT

3.3 Knocking down ARRDC4 strengthened the baicalin-mediated inhibition of HCT-116 cell proliferation

To determine the role of ARRDC4 in CRC cell proliferation, we established a stable ARRDC4 knockdown HCT-116 cell line using a lentiviral delivery system (Fig. 5A), confirmed the downregulation of ARRDC4 protein and mRNA levels in this cell line (Fig. 5B and C). The role of ARRDC4 in baicalin-mediated activities was further clarified, to down-regulation of ARRDC4 and then investigate baicalin’s activity. HCT-116 cells were knocked down ARRDC4 for 72 h, then the cells were treated with 10 μM baicalin, and cell viability was assessed by the CCK-8 assay (Fig. 5D). The inhibition of CRC cell proliferation by baicalin was dramatically strengthened in ARRDC4-low expression cells, as evaluated by Celigo and cell growth curve analysis utilizing a fluorescent imaging equipment.

Close

Fig. 5

Knockdown of ARRDC4 inhibits HCT-116 cell proliferation in vitro. (A) Infection efficiency was determined at 48 h after infection of lentivirus shARRDC4 or shCtrl in HCT-116 cells. Original magnification × 100, ×200. (B) ARRDC4 mRNA in CRC cells was measured by q-PCR, and normalized to GAPDH. (C) ARRDC4 protein expression was analyzed by western blot analysis in HCT-116 infected with shARRDC4 or shCtrl. ^**P < 0.001 shARRDC4 vs shCtrl. (D) Cell growth curve analysis comparing ARRDC4 knockdown (shARRDC4 + baicalin) with negative control (shCtrl + baicalin) HCT-116 cells. (E) Knocked down of ARRDC4 accelerated the effect of baicalin on apoptosis. Cell apoptosis was analyzed was analyzed by flow cytometry. (F) Activity of cleaved caspase-3/7 was determined in HCT-116 cells transfected with shCtrl + baicalin or shARRDC4 + baicalin (^***P < 0.001).

Export to PPT

3.4 Knocking down ARRDC4 accelerated the baicalin-induced apoptosis of HCT-116 cells

To determine whether ARRDC4 regulated apoptosis in the baicalin-mediated CRC cell proliferation inhibition, we performed Annexin V-APC labeling and flow cytometry analysis. We observed 5% of apoptosis in HCT-116 cells treated with 10 μM baicalin, but the ARRDC4-knockdown cells had a significantly higher apoptotic ratio (Fig. 5E). Morever, compared with the shCtrl + baicalin group, the cleaved caspase-3/7 activity was statistically significantly increased in the shARRDC4 + baicalin group (Fig. 5F). These findings indicated that the baicalin-mediated induction of apoptosis in HCT-116 cells was strongly connected to its effect on ARRDC4.

3.5 Baicalin restrained CRC xenograft growth in vivo

After the in vitro exploration of ARRDC4′s role, we assessed whether knocking down ARRDC4 affected the in vivo formation of tumors. First, we established an in vivo tumor xenograft model by subcutaneously injecting sh ARRDC4-HCT-116 (KD groups) or HCT-116 cells (NC groups) into BALB/c nude mice (Fig. 6A). As shown in Fig. 6B-D, the KD + baicalin groups had significantly smaller and lighter tumors than the NC + baicalin (50 mg/kg) group. At the end of the experiment, the tumors of the vehicle group reached around 500 mm³, while those of the KD + baicalin group averaged 300 mm³. Overall, baicalin (50 and 100 mg/kg) inhibited tumor growth on both NC and KD groups, but the effect was weaker on NC groups. HE staining (Fig. 7A) revealed that the tumors in the NC + baicalin groups presented different degrees of degenerative lesions, and some areas showed obvious cell necrosis. These lesions were more obvious in the KD + baicalin groups, and vacuolar degeneration and necrosis could be seen in some cells. Unlike the NC + baicalin groups, the KD + baicalin groups showed large necrotic areas and cells with more nuclear shrinkage, fragmentation, and dissolution. Subsequently, we investigated effects of baicalin on cell proliferation markers in vivo by IHC analysis. The KD + baicalin groups had substantially lower levels of Ki-67, a well-known cell proliferation marker, than the NC + baicalin groups (Fig. 7B and C). Furthermore, the Western blot analysis revealed that knocking down ARRDC4 in groups treated with baicalin dramatically reduced the protein expression of p21 (Fig. 7D). Contrarily, ARRDC4 knockdown with baicalin treatment might upregulate Cyclin D1 protein expression (Fig. 7D). The results demonstrated that knocking down ARRDC4 in groups treated with baicalin might eventually suppress tumor growth.

Close

Fig. 6

Knockdown of ARRDC4 with baicalin inhibited HCT-116 xenograft growth in vivo. (A) Morphology of HCT-116 cells after silencing ARRDC4. (B) HCT116 cells expressing sh Ctrl or sh ARRDC4 were injected into the right flank of nude mice for xenograft model, then after 2 weeks, the tumors of nude mice from the NC + baicalin (50 mg/kg), NC + baicalin (100 mg/kg), KD + baicalin (50 mg/kg) and KD + baicalin (100 mg/kg) groups were observed and tumor samples were separately collected. (C) Tumor volume growth curves and (D) tumor weight change for subcutaneous xenografts.

Export to PPT

Close

Fig. 7

Knockdown of ARRDC4 with baicalin exerts antitumor effect in vivo. (A) Effect of knockdown of ARRDC4 with baicalin on histological changes was visualized using HE staining (×100). (B) Representative images of IHC staining for Ki67 cell in the tumors from the mice in each group (×100). (C) Quantification of immunohistochemistry intensity of Ki67 area in the tumor tissues of mice. The IHC score was calculated and analyzed. (D) Protein expression levels of CyclinD1 and p21 in tumors of nude mice from the NC + baicalin (50 mg/kg), NC + baicalin (100 mg/kg), KD + baicalin (50 mg/kg) and KD + baicalin (100 mg/kg) groups. All data are presented as mean ± SD. *P < 0.05, compared with NC + baicalin (100 mg/kg) group; ^**P < 0.01, compared with KD + baicalin (100 mg/kg) group; ^***P < 0.01, compared with KD + baicalin (50 mg/kg) group.

Export to PPT

3.6 Baicalin inhibited cell proliferation by regulating LCN2, Claudin-2 and GSTA4

To explore the target of ARRDC4, we performed RNA sequencing with shARRDC4-HCT-116 and HCT-116 cells, and we screened 25 genes with consistent trends with the previous sequencing results (Fig. 8). Knocking down ARRDC4 significantly increased LCN2 levels and reduced Claudin-2 and GSTA4 levels (Fig. 9). These findings suggest that ARRDC4 mediates CRC cell proliferation through LCN2, Claudin-2, and GSTA4.

Close

Fig. 8

RNA-seq sequencing was performed with sh ARRDC4-HCT-116 or HCT-116 cells, and we screened 30 genes with consistent trends with the previous sequencing results.

Export to PPT

Close

Fig. 9

Western blot assay was performed to test ABP1, DAPP1, NLRP3, LCN2, GGT5, GSTA4 and Claudin-2 expression in the HCT-116 cells with or without ARRDC4. β-actin was performed as an internal control. Data are presented as the mean ± SD and are analyzed by one-way ANOVA. ^***P < 0.001.

Export to PPT