Paeoniflorin inhibits colorectal cancer cell stemness through the miR‐3194‐5p/catenin beta‐interacting protein 1 axis

Abstract

Paeoniflorin (PF) is a natural plant ingredient with remarkable antitumor effects. Herein, we investigated the biological effects and mechanism of PF in colorectal cancer (CRC) cell stemness. The messenger RNA (mRNA) and protein expressions were assessed using quantitative real‐time polymerase chain reaction and western blot. The viability, proliferation, and migration and invasion of CRC cells were evaluated using cell counting kit‐8, clone‐formation, and transwell migration and invasion assays, respectively. The sphere‐formation capacity was determined using the sphere‐formation assay. A dual‐luciferase reporter gene assay was employed to analyze the interaction between miR‐3194‐5p and catenin beta‐interacting protein 1 (CTNNBIP1). The viability, migration, invasion, epithelial–mesenchymal transition, and stemness of CRC cells were repressed by PF. MiR‐3194‐5p was upregulated in CRC tissues and cells. MiR‐3194‐5p knockdown suppressed CRC cell stemness, while miR‐3194‐5p overexpression had the opposite effect. In addition, the inhibition of CRC cell stemness caused by PF was eliminated by miR‐3194‐5p overexpression. CTNNBIP1 functioned as the target of miR‐3194‐5p, whose knockdown abrogated the repression of CRC cell stemness and Wnt/β‐catenin signaling activation by PF.PF regulated the miR‐3194‐5p/CTNNBIP1/Wnt/β‐catenin axis to repress CRC cell stemness.

Abbreviations

ANOVA

analysis of variance

Axin2

axis inhibition protein 2

CCK‐8

cell counting kit‐8

CRC

colorectal cancer

COAD

colon adenocarcinoma

CSCs

cancer stem cells

CTNNBIP1

catenin beta‐interacting protein 1

EMT

epithelial–mesenchymal transition

miRNAs

microRNAs

MMP2

matrix metalloproteinase‐2

MMP9

metalloproteinase‐9

OCT4

octamer binding factor 4

PF

paeoniflorin; PVDF, polyvinylidene fluoride

qRT‐PCR

Quantitative real‐time polymerase chain reaction

Sox2

SRY‐box transcription factor 2

1. INTRODUCTION

Colorectal cancer (CRC), characterized by multiple distal metastases, is a common human malignancy that ranks third in incidence and second in mortality among all cancers worldwide. ¹ Surgery combined with chemotherapy is the main clinical treatment for CRC. ² Although treatment can prolong CRC patient survival, the prognosis of patients with metastatic CRC remains poor. Cancer stem cells (CSCs) are a small subpopulation of tumor cells with the abilities of unlimited self‐renewal, unlimited proliferation, heterogeneous tumor cell production, and enhanced migration and invasion. ³ As reported, CSCs are regulated by the overexpression of octamer‐binding factor 4 (OCT4), Nanog, and SRY‐box transcription factor 2 (Sox2), among others, which are considered CSC markers. ⁴ Research has shown that the proportion of CSCs in metastatic CRC is significantly higher than that in other CRC subtypes, and evidence has revealed that CRC stem cells are considered the main factor for the poor prognosis of metastatic CRC. ⁵ Therefore, finding a drug that can effectively inhibit CRC cell stemness is significant for CRC treatment.

Paeoniflorin (PF) is the main active ingredient of Paeonia lactiflora with antitumor effects. ⁶ , ⁷ For example, Zhang et al. demonstrated that PF treatment markedly suppressed CRC cell migration and invasion abilities and epithelial–mesenchymal transition (EMT). ⁸ Previous studies have shown the antitumor effects of PF on CRC; however, the function of PF in regulating CRC cell stemness remains unclear and deserves further research.

MicroRNAs (miRNAs) refer to noncoding RNA molecules approximately 20 nucleotides in length. MiRNA dysregulation can effectively induce CRC pathogenesis and development. For example, miR‐875‐3p was significantly upregulated in CRC, and its downregulation repressed malignant behaviors of CRC cells. ⁹ In addition, miR‐590‐5p expression was markedly reduced in CRC, and its overexpression suppressed the malignant behaviors of cancer cells. ¹⁰ A previous study showed that miR‐3194‐5p knockdown suppressed the proliferation, migration, and tube formation of hypoxia‐treated vascular endothelial cells and increased cell apoptosis. ¹¹ Young et al. demonstrated that the estradiol‐mediated inhibition of SP1 decreased miR‐3194‐5p expression, promoting lung cancer progression. ¹² In addition, Hufbauer et al. revealed that miR‐3194‐5p was highly expressed in primary human papillomavirus type 16 oropharyngeal squamous‐cell‐carcinoma and matched metastasis. ¹³ However, the function of miR‐3194‐5p in CRC has not been elucidated. Notably, previous research has reported that PF inhibited the proliferation and migration of psoriasis cells by downregulating miR‐3194‐5p. ¹⁴ Thus, considering the anti‐CRC effect of PF and the inhibitory effect of PF on miR‐3194‐5p, we reasonably hypothesized that PF inhibits CRC progression by acting on miR‐3194‐5p.

Catenin beta‐interacting protein 1 (CTNNBIP1), which encodes the protein ICAT, is a classic tumor suppressor gene. ¹⁵ Previous studies reported that patients with high CTNNBIP1 expression had improved prognoses and that CTNNBIP1 overexpression suppressed CRC cell proliferation. ¹⁵ CTNNBIP1 inhibits β‐catenin‐mediated transcription and is a target in Wnt‐dependent cancer therapy. ¹⁶ Wnt/β‐catenin pathway dysregulation facilitates cancer initiation, progression, differentiation, and metastasis. ¹⁷ Kang et al. demonstrated that CTNNBIP1 upregulation inhibited CRC growth by deactivating the Wnt/β‐catenin signaling pathway. ¹⁸ Herein, we predicted that miR‐3194‐5p had a potential binding site to CTNNBIP1 using the starBase database. We believe that miR‐3194‐5p activates the Wnt/β‐catenin pathway by downregulating CTNNBIP1 to accelerate CRC development.

Based on the above evidence, we speculate that PF inhibits CRC cell stemness by acting on the miR‐3194‐5p/CTNNBIP1/Wnt/β‐catenin pathway. In this study, we propose a potential drug for CRC treatment and elucidate its mechanism of action.

2. MATERIALS AND METHODS

2.1. Cell culture and treatment

Normal human colon epithelial cells and human CRC cells (HCT116 and SW480) were obtained from ATCC (VA, USA) and cultured in DMEM (Gibco, CA, USA) containing 10% fetal bovine serum (Gibco) at 37°C with 5% CO₂. For PF treatment, CRC cells were treated with PF (0, 5, 10, 30, and 60 μM) (S2410, Selleckchem, TX, USA) for 48 h as previously described. ¹⁹

2.2. Cell transfection

The short hairpin RNA of CTNNBIP1 (sh‐CTNNBIP1), mimic/inhibitor of miR‐3194‐5p, and their negative controls were purchased from GenePharma (Shanghai, China) and transfected into cells using Lipofectamine™ 3000 (Invitrogen, CA, USA). The sequences were listed as follows: sh‐CTNNBIP1 (5′‐AGTCCGGAGGAGATGTACATT‐3′), sh‐NC (5′‐GTTCTCCGAACGTGTCACGT‐3′), miR‐3194‐5p mimics (sense: 5′‐GGCCAGCCACCAGGAGGGCUG‐3′, antisense: 5′‐CAGCCCUCCUGGUGGCUGGCC‐3′), mimics NC (sense: 5′‐UCACAACCUCCUAGAAAGAGUAGA‐3′, antisense: 5′‐UCACAACCUCCUAGAAAGAGUAGA‐3′), miR‐3194‐5p inhibitor (5′‐CAGCCCUCCUGGUGGCUGGCC‐3′), and inhibitor NC (5′‐UCUACUCUUUCUAGGAGGUUGUGA‐3′).

2.3. Cell counting kit‐8 assay

CRC cells were seeded in a 96‐well plate with 5 × 10³ cells. After treatment, the cells were incubated with a 10‐μL cell counting kit‐8 (CCK‐8) solution (Beyotime, Shanghai, China) for 3 h at 37°C. Subsequently, the absorbance at 450 nm was analyzed with a microplate spectrophotometer.

2.4. Clone‐formation assay

CRC cells were inoculated into 6‐well plates (1 × 10³ cells per well) and incubated at 37°C for 15 days. The culture was terminated when visible colonies were formed. Afterward, the cells were washed in phosphate‐buffered saline (Beyotime, Shanghai, China) and fixed with 4% paraformaldehyde (Beyotime), followed by 0.1% crystalized violet (Sigma‐Aldrich, USA) staining. The cell colonies were observed and photographed under a microscope (Olympus, Tokyo, Japan).

2.5. Transwell migration and invasion assays

The migration and invasion abilities of CRC cells were assessed using transwell and Matrigel‐coated transwell assays, respectively. For the migration assay, 500 μL of DMEM containing 1 × 10⁴ cells was moved to the upper transwell chamber (BD, NJ, USA), and 1000 μL of complete DMEM was added to the bottom chamber. The transwell invasion assay was performed using an upper chamber coated with Matrigel (200 μg/mL, Corning, NY, USA), as previously reported. ²⁰ After being fostered for 12 h at 37°C, the cells were fixed in ice methanol and stained with crystalized violet for 30 min. The filter removed the upper part of the cells, which was observed and photographed under a microscope.

2.6. Sphere‐formation assay

CRC cells were cultured in ultralow adhesion plates (2 × 10³ per well) with DMEM‐F12 (Gibco) containing 10 μg/L fibroblast and epidermal growth factors and 2% B27 for 14 days. Subsequently, spheres larger than 100 μm were counted using an inverted microscope (Nikon, Tokyo, Japan).

2.7. Dual‐luciferase reporter gene assay

CTNNBIP1 3′UTR fragments containing binding sites with miR‐3194‐5p were amplified using a polymerase chain reaction. Site‐directed mutagenesis was performed using a site‐directed mutagenesis Kit (Stratagene, CA, USA). Wild‐type (wt, 5′‐CAGCUACUCAGUGGGGCUGGCA‐3′) and mutant (mut, 5′‐CACGGAGAGAGACCCCGACCGA‐3′) CTNNBIP1 3′UTR sequences were cloned into the pmiRGLO vector (Promega, WI, USA). HCT116 and SW480 cells were cotransfected with CTNNBIP1‐wt/CTNNBIP1‐mut and the miR‐3194‐5p mimics/inhibitor. Subsequently, luciferase activity was examined.

2.8. Quantitative real‐time polymerase chain reaction

Total RNA was extracted with TRIzol (ThermoFisher Scientific, MA, USA). The cDNA for messenger RNA (mRNA) and miRNA was synthesized using the HiFiScript cDNA synthesis kit (CWBIO, Beijing, China) and the first‐strand cDNA synthesis kit (ThermoFisher Scientific), respectively, and subjected to quantitative real‐time polymerase chain reaction (qRT‐PCR) assay with SYBR (ThermoFisher Scientific). GAPDH and U6 were used as the reference genes of mRNA and miRNA, respectively, and the data were analyzed using the 2^−ΔΔCT method. The primers were listed as follows (5′–3′):

miR‐3194‐5p (F): ACACTCCAGCTGGGGGCCAGCCACCAGGA.

miR‐3194‐5p (R): TGGTGTCGTGGAGTCG. ¹¹

CTNNBIP1 (F): GGAAGAGTCCGGAGGAGATGTACATTC.

CTNNBIP1 (R): CTACTGCCTCCGGTCTTCCGTCT. ²¹

GAPDH (F): GAAGGTGAAGGTCGGAGTC.

GAPDH (R): GAAGATGGTGATGGGATTTC. ²²

U6 (F): CTCGCTTCCGCAGCAAT.

U6 (R): AACGCTTCAGTAATTCGCGT. ¹¹

2.9. RNA pull‐down assay

miR‐3194‐5p‐sense was transcribed, labeled with biotin, and purified. Approximately 2 × 10⁷ cells were dissolved in soft lysis buffer plus 80‐U/mL RNasin (Promega, WI, USA). The cell extract was incubated with biotinylated RNA for 1 h. Washed streptavidin‐coupled agarose beads (Invitrogen) were added to each binding reaction and incubated for 1 h. The beads were further washed six times in the lysis buffer. The retrieved RNA was assessed using qRT‐PCR.

2.10. Western blot

Total proteins were extracted using radio immunoprecipitation assay (Thermo Fisher Scientific), and a bicinchoninic acid kit (Beyotime) was employed to quantify the concentration. Proteins were separated using 10% sodium dodecyl sulfate‐page gel and transferred into a polyvinylidene luoride (PVDF) membrane (Millipore, MA, USA). Subsequently, the PVDF membranes were incubated overnight with antibodies against CTNNBIP1 (Abcam, VA, USA; 1: 1000, ab129011), OCT4 (Abcam; 1: 1000, ab181557), Nanog (Abcam; 1: 5000, ab109250), Sox2 (Abcam; 1: 1000, ab92494), β‐catenin (Abcam; 1: 5000, ab32572); p‐β‐catenin (Abcam; 1: 1000, ab75777); Axis inhibition protein 2 (Axin2, Abcam; 1: 2000, ab109307), C‐myc (Abcam; 1: 1000, ab32072), matrix metalloproteinase‐2 (MMP2, Abcam; 1: 1000, ab92536), metalloproteinase‐9 (MMP9, Abcam; 1: 1000, ab283575), E‐cadherin (CST, MA, USA; 1: 1000, #3195), Vimentin (Abcam; 1: 1000, ab92547), and the GAPDH antibody (Abcam; 1: 10000, ab9485). A corresponding second antibody (Abcam; 1: 10000, ab7090) was then applied for 1 h at 37°C. Protein bands were analyzed using an ECL detection kit (Beyotime).

2.11. Statistical analysis

All our data were obtained from three independent experiments. Statistical data were analyzed using SPSS 19.0 (IBM, Armonk, NY) and expressed as means ± SD. Between‐group differences and multi‐group comparisons were determined using Student's t‐test and one‐way analysis of variance, respectively. P values <0.05 were considered significant.

3. RESULTS

3.1. PF suppressed CRC cell viability, migration, invasion, stemness, and EMT by deactivating Wnt/β‐catenin signaling

To preliminarily explore the effect of PF on the cellular behaviors of CRC, we treated HCT116 and SW480 cells with PF (0, 5, 10, 30, and 60 μM). The CCK‐8 assay results showed that PF suppressed CRC cell viability in a dose‐dependent manner (Figure 1A). Subsequently, the clone‐formation assay demonstrated that PF inhibited CRC cell proliferation in a dose‐dependent manner (Figure S1). Moreover, CRC cell migration (Figure 1B) and invasion (Figure 1C) were markedly reduced in a dose‐dependent manner following PF treatment. Wnt/β‐catenin signaling activation is involved in facilitating CRC cell stemness. ²³ In this study, we observed that the protein levels of Wnt/β‐catenin signaling molecules (p‐β‐catenin, Axin2, and C‐myc) in CRC cells were significantly decreased in a dose‐dependent manner following PF treatment (Figure 1D). In addition, PF treatment reduced the levels of stemness markers (OCT4, Nanog, and Sox2) in CRC cells in a dose‐dependent manner (Figure 1E). We examined the effects of different PF doses on the cell viability of normal human colonic epithelial cells and found that cytotoxic effects were absent until the PF concentration was upregulated to 90 μM (Figure S2). Considering that the inhibitory effect of 30 μM PF on CRC cell viability was approximately 50% and did not affect normal cell viability, it was selected for subsequent experiments. Consistently, we observed that PF (30 μM) treatment resulted in reduced sphere‐formation capacity (Figure 1F). PF treatment also decreased the protein levels of MMP2, MMP9, and vimentin and increased the protein level of E‐cadherin in CRC cells (Figure 1G), suggesting that PF inhibited EMT in CRC cells. These results suggested that PF could dramatically inhibit CRC cell malignant phenotypes by inhibiting Wnt/β‐catenin signaling.

FIGURE 1.

PF suppressed CRC cell viability, migration, invasion, stemness, and EMT by deactivating Wnt/β‐catenin signaling. HCT116 and SW480 cells were treated with PF (0, 5, 10, 30, and 60 μM). (A) Cell viability was measured using CCK‐8 assay. (B–C) Cell migration and invasion were measured using transwell assays. (D–E) β‐Catenin, p‐β‐catenin, Axin2, C‐myc, OCT4, Nanog, and Sox2 levels were analyzed using western blot assay. (F) Sphere‐formation capacity was measured using the sphere‐formation assay. (G) Protein levels of MMP2, MMP9, E‐cadherin, and vimentin in CRC cells were determined using western blot assay. The measurement data were presented as mean ± SD. All data were obtained from at least three replicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001. Axin2, axis inhibition protein 2; CCK‐8, cell counting kit‐8; CRC, colorectal cancer; EMT, epithelial–mesenchymal transition; MMP2, matrix metalloproteinase‐2; MMP9, metalloproteinase‐9; PF, paeoniflorin; OCT4, octamer binding factor 4; Sox2, SRY‐box transcription factor 2.

3.2. MiR‐3194‐5p enhanced CRC cell viability, migration, invasion, EMT, and stemness

We analyzed miR‐3194‐5p expression in CRC and found that miR‐3194‐5p was notably upregulated in CRC cells (Figure 2A). What's more, starBase database found that miR‐3194‐5p was upregualated in colon adenocarcinoma (COAD) tissues (Figure S3A). The survival rate of COAD patients with high expression of miR‐3194‐5p showed a decreasing trend (Figure S3B). To further investigate the role of miR‐3194‐5p in CRC, miR‐3194‐5p knockdown and overexpression were induced in HCT116 and SW480 cells (Figure 2B). Functional experiments demonstrated that miR‐3194‐5p knockdown decreased CRC cell viability (Figure 2C), migration (Figure 2D), and invasion (Figure 2E), while miR‐3194‐5p upregulation had the opposite effect. In addition, miR‐3194‐5p inhibitor transfection resulted in reduced MMP2, MMP9, and vimentin levels and an increased E‐cadherin level in CRC cells, while the transfection of miR‐3194‐5p mimics had the opposite effect (Figure 2F), indicating that miR‐3194‐5 promoted CRC cell EMT. OCT4, Nanog, and Sox2 levels in CRC cells were significantly reduced by miR‐3194‐5p downregulation and increased by miR‐3194‐5p overexpression (Figure 2G). Moreover, the sphere‐formation capacity of CRC cells was inhibited by miR‐3194‐5p knockdown and promoted by miR‐3194‐5p overexpression (Figure 2H). Collectively, miR‐3194‐5p knockdown suppressed CRC cell viability, migration, invasion, EMT, and stemness, while miR‐3194‐5p overexpression had the opposite effect.

FIGURE 2.

MiR‐3194‐5p enhanced CRC cell viability, migration, invasion, EMT, and stemness. (A) qRT‐PCR was employed to detect miR‐3194‐5p expression in CRC cells. We induced miR‐3194‐5p knockdown/overexpression in CRC cells by transfecting miR‐3194‐5p inhibitor/mimics into the cells. (B) miR‐3194‐5p expression in cells was examined using qRT‐PCR. (C) CCK‐8 assay was conducted to examine cell viability. (D–E) Cell migration and invasion were measured using transwell assays. (F) Protein levels of MMP2, MMP9, E‐cadherin, and vimentin in CRC cells were measured using western blot assay. (G) OCT4, Nanog, and Sox2 protein levels in CRC cells were measured using western blot assay. (H) Sphere‐formation capacity was determined using the sphere‐formation assay. The measurement data were presented as mean ± SD. All data were obtained from at least three replicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001. CCK‐8, cell counting kit‐8; CRC, colorectal cancer; EMT, epithelial–mesenchymal transition; MMP2, matrix metalloproteinase‐2; MMP9, metalloproteinase‐9; OCT4, octamer binding factor 4; qRT‐PCR, quantitative real‐time polymerase chain reaction; Sox2, SRY‐box transcription factor 2.

3.3. MiR‐3194‐5p upregulation eliminated the suppression of CRC cell stemness by PF

To investigate whether miR‐3194‐5p was modulated by PF in CRC, we induced miR‐3194‐5p overexpression in PF‐treated CRC cells. We found that miR‐3194‐5p expression in CRC cells was markedly decreased after PF treatment, while this effect was abrogated following the transfection of miR‐3194‐5p mimics (Figure 3A). The CCK‐8 and transwell migration and invasion assays revealed that miR‐3194‐5p upregulation abolished the repression of CRC cell viability, migration, and invasion by PF (Figure 3B–D). MiR‐3194‐5p upregulation also abolished the repression of MMP2, MMP9, and Vimentin levels and the promotion of the E‐cadherin level by PF in CRC cells (Figure 3E). Additionally, the inhibition of the OCT4, Nanog, and Sox2 levels by PF in CRC cells was eliminated by miR‐3194‐5p overexpression (Figure 3F). Meanwhile, miR‐3194‐5p overexpression abrogated the repression of the sphere‐formation capacity of CRC cells by PF (Figure 3G). In summary, PF inhibited CRC cell stemness by reducing miR‐3194‐5p expression.

FIGURE 3.

MiR‐3194‐5p upregulation eliminated the suppression of CRC cell stemness by PF. PF‐treated CRC cells were transfected with mimics NC or miR‐3194‐5p mimics. (A) qRT‐PCR was performed to determine miR‐3194‐5p expression in cells. (B) Cell viability was measured using CCK‐8 assay. (C–D) Cell migration and invasion were measured using transwell assays. (E) Protein levels of MMP2, MMP9, E‐cadherin, and vimentin in CRC cells were assessed using western blot assay. (F) Western blot assay was performed to assess OCT4, Nanog, and Sox2 protein levels in cells. (G) Sphere‐formation assay was conducted to detect the sphere‐formation capacity of cells. The measurement data were presented as mean ± SD. All data were obtained from at least three replicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001. CCK‐8, cell counting kit‐8; CRC, colorectal cancer; MMP2, matrix metalloproteinase‐2; MMP9, metalloproteinase‐9; PF, paeoniflorin; OCT4, octamer binding factor 4; Sox2, SRY‐box transcription factor 2.

3.4. MiR‐3194‐5p downregulated CTNNBIP1 by directly binding to CTNNBIP1 in CRC cells

CTNNBIP1 is a tumor suppressor gene in various human malignant tumors, including CRC. ¹⁵ In this study, we predicted tmiR‐3194‐5p to have a potential binding site to CTNNBIP1 using the starBase database (Figure 4A). The luciferase activity of the CTNNBIP1‐wt group was suppressed by the transfection of miR‐3194‐5p mimics but enhanced by the transfection of the miR‐3194‐5p inhibitor (Figure 4B). However, no such changes occurred in the CTNNBIP1‐mut group, suggesting that miR‐3194‐5p is directly bound with CTNNBIP1. To further verify the binding relationship between miR‐3194‐5p and CTNNBIP1, RNA pull‐down assay was performed. The results showed that biotin‐labeled miR‐3194‐5p could enrich CTNNBIP1 mRNA in CRC cells (Figure 4C). In addition, CTNNBIP1 expression in CRC cells was decreased after miR‐3194‐5p overexpression, while CTNNBIP1 expression was increased by miR‐3194‐5p inhibition (Figure 4D,E). Overall, miR‐3194‐5p negatively regulated CTNNBIP1 expression in CRC cells by targeting CTNNBIP1.

FIGURE 4.

MiR‐3194‐5p downregulated CTNNBIP1 by directly binding to CTNNBIP1 in CRC cells. (A) starBase was employed to predict the potential binding site between miR‐3194‐5p and CTNNBIP1. (B–C) The interaction between miR‐3194‐5p and CTNNBIP1 was analyzed using dual‐luciferase reporter gene and RNA pull‐down assays. (D–E) CTNNBIP1 expression in CRC cells after miR‐3194‐5p overexpression/knockdown was determined using qRT‐PCR and western blot. The measurement data were presented as mean ± SD. All data were obtained from at least three replicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001. CRC, colorectal cancer; CTNNBIP1, catenin beta‐interacting protein 1; qRT‐PCR, quantitative real‐time polymerase chain reaction.

3.5. CTNNBIP1 knockdown eliminated the inhibition of CRC cell stemness and Wnt/β‐catenin signaling pathway by PF

We examined CTNNBIP1 expression in CRC and found that CTNNBIP1 was markedly downregulated in CRC cells compared with normal human colonic epithelial cells (Figure 5A). Subsequently, we transfected sh‐NC and sh‐CTNNBIP1 into CRC cells to knockdown CTNNBIP1. CTNNBIP1 expression in CRC cells was markedly decreased following sh‐CTNNBIP1 transfection (Figure 5B), indicating a successful transfection. Furthermore, we induced CTNNBIP1 knockdown in PF‐treated CRC cells to investigate the role of CTNNBIP1 in PF‐mediated cell functions. PF treatment markedly increased the CTNNBIP1 mRNA level in CRC cells, while this effect was abolished by sh‐CTNNBIP1 transfection (Figure 5C). Functional experiments subsequently revealed that CTNNBIP1 knockdown eliminated the repression of CRC cell viability, migration, and invasion by cPF (Figure 5D–F). Furthermore, CTNNBIP1 silencing reversed the repression of MMP2, MMP9, and vimentin levels and promotion of the E‐cadherin level by PF in CRC cells (Figure 5G). The inhibition of Wnt/β‐catenin signaling molecules (p‐β‐catenin, Axin2, and C‐myc) by PF in CRC cells was eliminated by CTNNBIP1 downregulation (Figure 5H). Moreover, CTNNBIP1 silencing abrogated the inhibitory effect of PF on OCT4, Nanog, and Sox2 levels and the sphere‐formation capacity in CRC cells (Figure 5I,J). To sum up, PF upregulated CTNNBIP1 to repress CRC cell stemness by inhibiting Wnt/β‐catenin signaling activation.

FIGURE 5.

CTNNBIP1 knockdown eliminated the inhibition of CRC cell stemness and the Wnt/β‐catenin pathway by PF. (A) qRT‐PCR was employed to measure CTNNBIP1 expression in cells. (B) CTNNBIP1 expression in CRC cells after sh‐NC and sh‐CTNNBIP1 transfection was detected using qRT‐PCR. CTNNBIP1 knockdown was induced in PF‐treated CRC cells. (C) CTNNBIP1 expression in cells was examined using qRT‐PCR. (D) CCK‐8 assay was conducted to examine cell viability. (E–F) Transwell assays were employed to detect cell migration and invasion. (G) Protein levels of MMP2, MMP9, E‐cadherin, and vimentin in CRC cells were examined using western blot assay. (H–I) β‐Catenin, p‐β‐catenin, Axin2, C‐myc, OCT4, Nanog, and Sox2 levels were determined using western blot assay. (J) Sphere‐formation assay was conducted to detect the sphere‐formation capacity of CRC cells. The measurement data were presented as mean ± SD. All data were obtained from at least three replicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001. Axin2, axis inhibition protein 2; CCK‐8, cell counting kit‐8; CRC, colorectal cancer; CTNNBIP1, catenin beta‐interacting protein 1; EMT, epithelial–mesenchymal transition; MMP2, Matrix metalloproteinase‐2; MMP9, metalloproteinase‐9; PF, paeoniflorin; OCT4, octamer binding factor 4; qRT‐PCR, quantitative real‐time polymerase chain reaction; Sox2, SRY‐box transcription factor 2.

4. DISCUSSION

CSCs are cells that have unlimited self‐renewal ability and can cause multilineage differentiation of malignant tumor cells. CSCs may explain various cancer phenomena, including chemical radiation resistance, metastasis, and drug resistance, thereby contributing to cancer recurrence. ²⁴ Recent research has shown that stemness is a key biological process driving CRC cell metastasis and invasion. ²³ In this study, we found that PF treatment markedly inhibited CRC cell viability, migration, invasion, and stemness. Mechanistically, PF inhibited Wnt/β‐catenin signaling activation by acting on the miR‐3194‐5p/CTNNBIP1 axis.

PF is the main bioactive component isolated from the root of Paeonia suffruticosa and is one of the formulations of many traditional Chinese medicines with various biological functions. ²⁵ Recent studies have shown that PF exhibits antitumor activity in various human malignancies, including CRC. ¹⁹ , ²⁶ A previous study revealed that PF could inhibit EMT and angiogenesis in human glioblastoma via K63‐Linked C‐Met polyubiquitination‐dependent autophagic degradation. ²⁷ Zhou et al. illustrated that PF prevented hypoxia‐induced EMT in human breast cancer cells. ²⁸ Furthermore, PF can suppress CRC cell growth, induce cell cycle arrest ¹⁹ markedly suppress CRC cell EMT. ⁸ Consistently, our results have shown that PF suppressed CRC cell migration, invasion, EMT, and stemness.

Several annotated miRNAs act as oncogenes or anti‐oncogenes. ²⁹ For instance, miR‐34a upregulation can inhibit EMT‐mediated CRC invasion and metastasis. ³⁰ In addition, miR‐96 upregulation can promote CRC occurrence and progression, ³¹ and miR‐3194‐5p might be the oncogene in oropharyngeal cancer. ¹³ To study the function of miR‐3194‐5p in CRC, we induced miR‐3194‐5p knockdown and overexpression in CRC cells and examined related indicators. Our results revealed that miR‐3194‐5p knockdown remarkably inhibited CRC cell viability, migration, invasion, and stemness. However, miR‐3194‐5p overexpression had the opposite effect, which was reported for the first time. In this study, we detected that miR‐3194‐5p had similar cancer‐promoting effects in two CRC cell lines (HCT116 and SW480). HCT116 has a high metastatic capacity, while SW480 has a low metastatic capacity. ³² To reflect the heterogeneity of miR‐3194‐5p in CRC cell lines, it is necessary to further detect the effects of miR‐3194‐5p in more types of CRC cells in the future. A previous study demonstrated that PF suppressed psoriatic keratinocyte proliferation and migration by reducing miR‐3194‐5p expression. ¹⁴ However, the mechanism and role of miR‐3194‐5p in PF‐mediated anticancer activity for CRC have not yet been elucidated. Our results demonstrated that miR‐3194‐5p overexpression partly abolished the repression of CRC cell malignant behaviors by PF. Thus, we proved that PF suppressed CRC cell stemness by downregulating miR‐3194‐5p.

The Wnt/β‐catenin pathway is involved in tumor‐initiating cells and CSCs. ³³ Wnt/β‐catenin signaling activation facilitates tumor cell migration, invasion, and stemness, ²³ and Wnt/β‐catenin pathway activation has been frequently observed in CRC. ¹⁵ In our study, we found that PF treatment inhibited Wnt/β‐catenin pathway activation in CRC cells in a dose‐dependent manner, suggesting that the pathway played a key role in the PF‐mediated biological effects on CRC. Extensive research has shown that CTNNBIP1 upregulation can inhibit tumor cell stemness by deactivating the Wnt/β‐catenin signaling pathway. ³⁴ Herein, we observed that miR‐3194‐5p downregulated CTNNBIP1 in CRC cells by directly targeting CTNNBIP1. As expected, CTNNBIP1 silencing abrogated the suppression of CRC cell stemness and Wnt/β‐catenin signaling activation by PF. Collectively, PF inhibited CRC cell stemness by repressing the miR‐3194‐5p/CTNNBIP1 axis to inhibit Wnt/β‐catenin signaling activation. Wang et al. showed that PF could inhibit the proliferation and migration of psoriatic keratinocytes through the lncRNA NEAT1/miR‐3194‐5p/galectin‐7 axis. ¹⁴ Notably, galectin‐7 plays an anticancer role in CRC. ³⁵ However, whether miR‐3194‐5p promotes CRC progression by targeting galectin‐7 is worth further study. Among the various miR‐3194‐5p target genes predicted by the database, adenomatous polyposis coli 2, ³⁶ KLF13, ³⁷ and LDLRAD4 ³⁸ play anticancer roles in CRC. However, further study is required to determine whether miR‐3194‐5p promotes the occurrence and development of CRC through targeted inhibition of these genes.

Overall, we report for the first time that PF exerts antitumor effects on CRC cell by mediating cell stemness and deactivating Wnt/β‐catenin signaling, which is achieved through the inhibition of the miR‐3194‐5p/CTNNBIP1 axis. However, clinical trials were lacking in this study. In future mechanism studies, we will further explore the relationship between the expression level of miR‐3194‐5p and CRC prognosis. Our findings provide a novel theoretical basis for deeply exploring metastatic CRC treatment using PF.

CONFLICT OF INTEREST STATEMENT

All authors declare no conflict of interest.

Supporting information

Figure S1. HCT116 and SW480 cells were treated with PF (0, 5, 10, 30, and 60 μM), and CRC cell proliferation was detected using clone‐formation assay. The measurement data were presented as mean ± SD. All data were obtained from at least three replicate experiments. *p < 0.05, **p < 0.01, ***p < 0.001. CRC, colorectal cancer; PF, paeoniflorin.

Figure S2. Normal human colon epithelial cells were treated with PF (0, 5, 10, 30, 60, and 90 μM), and cell viability was measured using CCK‐8 assay. The measurement data were presented as mean ± SD. All data were obtained from at least three replicate experiments. *p < 0.05. PF, Paeoniflorin; CCK‐8, cell counting kit‐8.

Figure S3. (A) starBase database analysis of miR‐3194‐5p expression in COAD and normal samples. (B) starBase database analysis of the relationship between miR‐3194‐5p expression and CRC patient prognosis. The measurement data were presented as mean ± SD. ***p < 0.001. COAD, colon adenocarcinoma.

Su Z, Hu B, Li J, Zeng Z, Chen H, Guo Y, et al. Paeoniflorin inhibits colorectal cancer cell stemness through the miR‐3194‐5p/catenin beta‐interacting protein 1 axis. Kaohsiung J Med Sci. 2023;39(10):1011–1021. 10.1002/kjm2.12736