Folic acid inhibits the Wnt/β-catenin pathway by upregulating DKK3 to exert anti-tumor effects in cervical squamous cell carcinoma

Abstract

Cervical cancer poses a significant threat to women’s health. Although folic acid (FA) has been recognized as a protective factor in cervical carcinogenesis, its precise molecular mechanisms remain incompletely understood. Here, we identify Dickkopf Wnt signaling pathway inhibitor 3 (DKK3), an inhibitor of the Wnt signaling pathway, as a potential target of FA. This study provides systematic evidence that DKK3 expression decreases during cervical squamous epithelial carcinogenesis, showing progressive downregulation from squamous intraepithelial lesions to squamous cell carcinoma, which correlates with advanced FIGO stage and lymphovascular space invasion. Functional assays confirmed DKK3’s tumor-suppressive role and therapeutic potential. DKK3 overexpression downregulated β-catenin protein levels and inhibited malignant behavior in SiHa cells, whereas its knockdown produced opposite effects. Notably, this study demonstrated for the first time that FA intervention upregulated DKK3 expression, suppressed Wnt/β-catenin signaling, leading to a reduction in β-catenin protein abundance, and exerted potent anti-tumor effects—suppressing proliferation, migration, and invasion while promoting apoptosis. Even under DKK3-knockdown conditions, FA intervention partially reversed β-catenin accumulation by enhancing residual DKK3 expression. Overall, this study establishes the prognostic and interventional value of DKK3 in cervical squamous epithelial carcinogenesis. FA intervention may serve as an effective strategy to restore DKK3 expression to inhibit the Wnt/β-catenin pathway, thereby exerting antitumor activity. Future cervical cancer prevention and treatment strategies may benefit from dynamic monitoring of DKK3 expression to identify potential beneficiaries and provide targeted FA intervention.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-32762-9.

Introduction

Cervical cancer remains a major global health burden for women, with a staggering estimated annual 660,000 new cases and 350,000 fatalities¹. Although persistent human papillomavirus (HPV) infection is established as the primary pathogenic factor in cervical carcinogenesis, only a subset of infected individuals ultimately progresses to cervical cancer². This clinical phenomenon suggests the important role of other synergistic factors in the process of carcinogenesis, including dysregulation of key signaling molecules and nutritional deficiencies^3,4. Current clinical management requires comprehensive consideration of disease stage, pathological type, patient age, and fertility preservation requirements⁵. While surgery, radiotherapy, and chemotherapy remain therapeutic cornerstones, advances in precision medicine are propelling a paradigm shift toward targeted molecular interventions, specifically through developing effective methods to regulate key molecular targets and exploring more economical, personalized early-intervention strategies^2,5.

Dickkopf Wnt signaling pathway inhibitor 3 (DKK3) is recognized as a tumor suppressor through its inhibition of the Wnt/β-catenin pathway (also known as canonical Wnt signaling), which is indispensable for tumorigenesis^6,7. A major hallmark in multiple cancer types is the excessive presence of β-catenin protein, driven by aberrant Wnt/β-catenin signaling^7,8. As a key antagonist of canonical Wnt signaling, DKK3 promotes β-catenin degradation via ubiquitin-mediated proteolysis, suppressing nuclear translocation^6,9. Multiple studies have reported reduced or absent DKK3 expression accompanied by abnormal β-catenin accumulation in numerous tumor tissues^10,11. In cervical cancer, low DKK3 expression has similarly been demonstrated in cancerous tissues^12–14. Reversing the negative effects caused by abnormal β-catenin accumulation through regulating DKK3 expression represents a potential therapeutic approach for cervical cancer.

Folic acid (FA), an essential nutrient for maintaining genomic stability, is increasingly studied for its potential role in cervical carcinogenesis^15,16. Multiple large-scale population-based studies have provided compelling evidence suggesting an inverse association between the prevalence of squamous intraepithelial lesions (including cervical intraepithelial neoplasia [CIN] grade 1 and CIN2/3) or cervical cancer and folate levels^17–20. Furthermore, epidemiological studies in countries implementing FA supplement fortification programs indicate that supraphysiological serum folate levels not only reduce the risk of CIN2 + but may even promote the regression of CIN1^21,22. These findings from large-scale population studies suggest that folate deficiency is an important synergistic factor in cervical carcinogenesis, while FA within specific concentration ranges may exert anti-cancer effects^17–22. However, it is essential to recognize that the molecular mechanisms connecting folate deficiency to cancer promotion and FA intervention to cervical cancer suppression remain unclear. The anti-tumor effects of FA intervention and their underlying molecular mechanisms require further exploration.

Notably, although evidence for FA intervention in cervical cancer remains limited, research on other disease models strongly suggests that FA may significantly influence the activity of the Wnt/β-catenin pathway to modulate β-catenin protein abundance, thereby affecting disease progression^23–28. This notion has been extensively validated in studies of disease models such as neural tube defects²⁶, congenital heart defects^27,28, and colorectal tumors^23,24. Although the specific direction of FA’s effects reported in these existing studies is not entirely consistent, the Wnt/β-catenin pathway is widely recognized as the central signaling hub of FA action. More direct evidence comes from a recent cervical cancer study, which observed that folate deficiency drives abnormal β-catenin accumulation along with accelerated proliferation²⁹. Collectively, this evidence reveals that FA may participate in cervical cancer progression by targeting canonical Wnt signaling to modulate β-catenin expression. However, whether DKK3 serves as a key molecular mediator directly involved in FA’s regulatory mechanism remains unexplored.

With the available scientific knowledge background about DKK3 and the role of FA in the canonical Wnt signaling pathway, we started this study with a hypothesis that FA may upregulate DKK3 to suppress this oncogenic pathway, thereby promoting β-catenin degradation and imparting anti-tumor effects in cervical cancer. Herein, we evaluated DKK3 expression in cervical squamous epithelial tissues with different lesion levels and established in vitro SiHa cell models featuring altered DKK3 expression and FA intervention to investigate whether FA exerts its anti-tumor effects by regulating DKK3 expression and subsequently inhibiting Wnt/β-catenin signaling activation. This study aims to reveal the DKK3-mediated molecular mechanism underlying FA, thereby providing theoretical support for FA-based targeted prevention and treatment of cervical cancer.

Materials and methods

Clinical samples

Clinical samples were collected from patients undergoing colposcopy, cervical biopsy, and surgical procedures at the Second Hospital of Shanxi Medical University from August 2020 to August 2022. The inclusion criteria for the control group comprised women with normal cervical cytological results and HPV genotype testing results who underwent total hysterectomy for uterine myomas, all of whom had normal postoperative cervical histopathological results (n = 90). The research groups consisted of patients diagnosed with CIN1 (n = 102), CIN2/3 (n = 103), and cervical squamous cell carcinoma (SCC) (n = 69), based on cervical histopathological examination. Slides underwent independent blinded assessment by two gynecologic pathologists from our hospital. Slides with discordant diagnoses were reviewed by a senior pathologist to reach a consensus diagnosis (two of three agreements). Exclusion criteria included: (1) pregnancy; (2) history of hysterectomy; (3) previous treatment for cervical or vaginal lesions; (4) other concomitant malignancies; (5) hematological or digestive system diseases. This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Second Hospital of Shanxi Medical University (approval number: 2019 YX No.309, Taiyuan, China). All participants involved in this study were fully informed about the study’s purposes and procedures, and provided written informed consent prior to participation.

Immunohistochemistry

For histopathological examination, 4 μm sections were prepared from paraffin-embedded cervical tissues and underwent standard hematoxylin and eosin staining. For immunohistochemistry, sections were routinely dewaxed and hydrated via high-pressure thermal repair for 10 min for antigen retrieval. Endogenous peroxidase activity was blocked through a 15-min incubation of the sections in 0.3% hydrogen peroxide. Next, anti-DKK3 primary antibody (1:100 dilution; cat # 66758-1-lg, Proteintech, Wuhan, China) was applied to the sections for overnight incubation at 4 °C after a 30-min blocking step. Following this, sections were exposed for 1 h at room temperature to a ready-to-use immunohistochemistry secondary antibody conjugated to horseradish peroxidase (HRP) (cat. No. PV-6000; ZSGB-Bio, Beijing, China). Then the sections were developed with diaminobenzidine (DAB) for 2 min, counterstained with hematoxylin, and subsequently dehydrated, dried, and coverslipped in a routine manner. The assessment of immunostaining was conducted in a blinded manner by two seasoned pathologists, evaluating both the intensity and extent of staining. The final immunoreactivity score was calculated as the product of staining intensity (0 for no staining, 1 for weak, 2 for moderate, and 3 for strong staining) and the percentage of stained area (0 for < 10%, 1 for 10%-24%, 2 for 25%-49%, and 3 for ≥ 50%). Based on this composite score, samples were finally stratified into two groups: DKK3 high expression (score ≥ 4) and DKK3 low expression (score < 4).

Cell culture

SiHa cell line, an HPV16-positive cervical squamous cell carcinoma cell line, was obtained from Pricella Co., Ltd. (Wuhan, China). It was selected as the experimental model for this study because it best represents the biological characteristics of cervical squamous cell carcinoma. DMEM supplemented with 10% fetal bovine serum (FBS) (Cellmax, Lanzhou, China), 100 µg/ml streptomycin, and 100 U/ml penicillin was used for cell cultivation at 37 °C in a 5% CO₂ humidified incubator. For FA intervention, SiHa cells were exposed to various concentrations of FA (MCE, Beijing, China) for 48 h. During the intervention period, the FA-containing medium was replaced with freshly prepared medium every 24 h.

Lentiviral transfection

Lentiviral vectors for DKK3 overexpression (DKK3-OE), DKK3 knockdown (DKK3-KD), and their respective negative controls (NC-OE, NC-KD), each containing green fluorescent protein and puromycin resistance genes, were procured from Genechem Co., Ltd. (Shanghai, China) to modulate DKK3 expression in SiHa cells. The negative controls contained non-targeting scrambled sequences, using the same viral backbone but lacking a target-specific insert, to ensure that any observed effects were due to specific DKK3 modulation rather than non-specific factors. Cells were transfected following the manufacturer’s instructions. After a 72-h transfection, the successfully transfected cells were selected with 3 µg/ml puromycin. Then, fluorescence microscopy was used to examine cells and to determine transfection efficiency. Finally, the transfection effect was further verified through reverse transcription-quantitative polymerase chain reaction (RT-qPCR) for gene expression and Western blotting for protein detection.

RT-qPCR

Total RNA was extracted from SiHa cells using a commercial RNA extraction kit (cat # MF-036; Mei5bio, Beijing, China). Subsequently, RNA was converted into cDNA utilizing the reverse transcription kit (cat # MF-166; Mei5bio) and then amplified using a real-time PCR Mix reagent (cat # MF-797; Mei5bio). The PCR cycling initiated with 45 s denaturation at 95 °C, and then it proceeded through 40 cycles of 15 s at 95 °C, 15 s at 55 °C, and 30 s at 72 °C. Finally, DKK3 mRNA expression was quantified using the 2^−ΔΔCt method (normalized to GAPDH). DKK3 and GAPDH primer sequences were designed as follows:

DKK3 FP: 5′ AGCTATCACAATGAGACCAACA 3′;

DKK3 RP: 5′ CATTTGTCCAGTCTGGTTGTTG 3′;

GAPDH FP: 5′ CAGGAGGCATTGCTGATGAT 3′;

GAPDH RP: 5′ GAAGGCTGGGGCTCATTT 3′.

Western blotting

The total protein was isolated using RIPA cell lysis buffer with 1 mM PMSF (Boster, Wuhan, China), and concentrations were quantified by bicinchoninic acid assay. Protein separation was achieved through sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the proteins were then transferred to polyvinylidene fluoride membranes via wet electroblotting. Membranes were then incubated for 2 h at room temperature in 5% skim milk. Following blocking, membranes were incubated overnight at 4 ℃ with primary antibodies against DKK3 (1:2000 dilution; cat # ab186409, Abcam, Boston, MA, USA), β-catenin (1:8000 dilution; cat # ab32572, Abcam, Boston, MA, USA), β-actin (1:8000 dilution; cat # ab0035, Abways, Shanghai, China), and were then exposed for 60 min at room temperature to HRP-conjugated secondary antibody(1:8,000 dilution; cat # BA1054; Boster, Wuhan, China). Ultimately, protein signals were visualized using an enhanced chemiluminescence solution (Meilunbio, Dalian, China) with the gel imaging system.

Cell counting kit-8 (CCK-8) assay

To determine the appropriate FA concentration, SiHa cells (4000 cells/well) were seeded into 96-well plates. After 24 h, cells were intervened with an FA gradient (200, 600, 1000, 1400, 1800 µg/ml). Cell viability was assessed 48 h post-treatment by adding 10 µl CCK-8 solution (Boster) per well. Then the optical density at 450 nm (OD₄₅₀) was acquired using a microplate spectrophotometer after 1.5 h incubation.

For proliferation analysis, SiHa cell suspensions from different groups were inoculated into 96-well plates (2000 cells/well). At designated time points, 10 µl of CCK-8 solution was added to each well, and the OD₄₅₀ was quantified after 3 h incubation.

Colony formation assay

SiHa cells from different groups were inoculated in six-well plates (1000 cells/well). Following two weeks of incubation, methanol was employed to fix the cells for 30 min and 0.1% crystal violet was used to stain them for 30 min. Ultimately, we counted and photographed clones that had a minimum of 50 cells each.

Wound healing assay

Different groups of SiHa cells were seeded at 1 × 10⁶/well in six-well plates and cultured to confluence. Then linear wounds were generated using a sterile pipette tip (10 µl), followed by replacement with serum-free medium with or without FA. The wound area was photographically recorded at 0 h and 48 h and was measured utilizing Image J software.

Transwell assay

Matrigel (Mogengel-Bio, Xiamen, China) was diluted with a serum-free medium at 1:8. Before been used in the invasion assay, the polycarbonate membrane of the transwell chamber (Corning, New York, USA) was pre-coated with 60 µl diluted matrigel. The chambers were subsequently transferred to a 24-well plate and incubated in the culture medium for three hours. In the upper chamber, 200 µl of serum-free SiHa cell suspension was plated, whereas the lower chamber contained 600 µl of 10% FBS-supplemented medium serving as chemoattractant. Each transwell chamber was seeded with 100,000 cells and cultured for 24 h to evaluate DKK3 manipulation effects. For FA intervention studies, transwell chambers for the migration and invasion assays were seeded with 30,000 cells and 50,000 cells, respectively, followed by 48 h culture. After adherent cells were gently removed from the inner surface, the chambers were exposed to methanol fixation for 30 min, followed by 30 min of staining with 0.1% crystal violet solution. Following this, the cell count was performed by randomly selecting five visual fields using a microscope (×100).

Flow cytometry

Apoptosis in SiHa cells exposed to varying concentrations of FA was detected using flow cytometry with the Annexin V-FITC/PI kit (Solarbio, Beijing, China). After trypsinization without ethylenediaminetetraacetic acid, cells from each group were harvested, rinsed with pre-cooled phosphate-buffered saline, and resuspended in binding buffer. Subsequently, cells were stained with 5 µl Annexin V-FITC for 5 min in darkness, and then 5 µl propidium iodide was added to identify apoptotic cells.

Statistical analysis

Statistical analyses were executed in SPSS software (v. 27.0). The data were displayed as arithmetic mean ± standard deviation (conforming to a normal distribution) and were obtained from at least three independent experiments. To compare multiple groups, one-way ANOVA coupled with Tukey’s test was conducted (conforming to the homogeneity of variance). Chi-squared and Fisher’s exact tests were used to evaluate associations between DKK3 expression and clinicopathological characteristics. Statistical significance was denoted by P < 0.05. The Bonferroni-corrected z-tests were used to assess the differential expression of DKK3 across various lesion stages.

Results

Reduced DKK3 expression significantly correlates with cervical squamous epithelial carcinogenesis and adverse clinicopathological characteristics

To systematically evaluate the changes in DKK3 expression during cervical squamous epithelial carcinogenesis, we compared its protein levels by immunohistochemistry and observed a progressive downregulation from normal epithelium to SCC tissues (Fig. 1; Table 1). Specifically, DKK3 expression was significantly downregulated in CIN tissues (including CIN1 and CIN2/3) compared to normal cervical epithelium (Fig. 1b), despite no significant difference between CIN1 and CIN2/3 themselves (Fig. 1a). This decreasing trend was most pronounced in SCC tissues, which exhibited a striking reduction in DKK3 expression compared to both normal and CIN tissues (Fig. 1b). The specific high-expression rates for all tissue types are detailed in Table 1.

Fig. 1.

DKK3 expression in different grades of cervical epithelial lesions. (a) Representative immunostaining images of DKK3 in different grades of cervical lesions (scale bar: 50 μm; 10 μm). (b) Summary analysis of DKK3 expression in different grades of cervical lesions (low expression -; high expression +).

Table 1.

DKK3 protein expression in different lesion grades.

Group	Cases (n = 364)	Expression of DKK3		P-value
		High (n = 193)	Low (n = 171)
Lesion grade				< 0.001
Normal	90	74(82.2%) a	16(17.8%)
CIN1	102	62(60.8%) b	40(39.2%)
CIN2/3	103	44(42.7%) b	59(57.3%)
SCC	69	13(18.8%) c	56(81.2%)

The superscripts ^abc denoted the variations among groups. If the identification remained consistent, the inter-group differences were statistically insignificant. Conversely, different identifications indicated statistically significant inter-group differences.

Furthermore, we explored the relationships linking DKK3 expression to clinicopathological characteristics in SCC. As documented in Table 2, lower DKK3 expression in SCC tissues was significantly associated with higher FIGO stage and the presence of lymphovascular space invasion. However, no associations were observed with DKK3 expression across different age groups, lymph node metastasis status, or nerve invasion patterns in this study.

Table 2.

Association between DKK3 expression and clinicopathological characteristics of cervical squamous cell carcinoma patients.

Characteristics	Cases (n = 69)	Expression of DKK3		P-value
		Low (n = 56)	High (n = 13)
Age				1.000
< 45 years	14	12(85.7%)	2(14.3%)
≥ 45 years	55	44(80.0%)	11(20.0%)
FIGO stage				0.023*
I	44	32(72.7%)	12(27.3%)
II-III	25	24(96.0%)	1(4.0%)
Lymphovascular space invasion				0.027*
Negative	34	24(70.6%)	10(29.4%)
Positive	35	32(91.4%)	3(8.6%)
Lymphnode metastasis				0.676
Negative	58	46(79.3%)	12(20.7%)
Positive	11	10(90.9%)	1(9.1%)
Nerve invasion				0.674
Negative	59	47(79.7%)	12(20.3%)
Positive	10	9(90.0%)	1(10.0%)

FIGO, International Federation of Gynecology and Obstetrics. The Fisher’s exact test was conducted on datasets where over 20% of the sets had expected counts less than 5, while the Chi-square test was employed for the remaining datasets. *P < 0.05.

DKK3 regulates the Wnt/β-catenin pathway and suppresses malignant behaviors in SiHa cells

Next, we established cell models with stable DKK3 overexpression (DKK3-OE) and knockdown (DKK3-KD). After puromycin selection, transfection efficiency was confirmed by fluorescence microscopy (Fig. 2a,f). Subsequently, DKK3 mRNA and protein levels were quantified in SiHa wild-type (CTRL), negative control (NC), and DKK3-manipulated cells. In the overexpression model, DKK3 expression was robustly elevated in the DKK3-OE group, showing an approximately 25-fold increase at the mRNA level and a 35-fold increase at the protein level compared to the CTRL group, whereas CTRL and NC-OE groups were comparable (Fig. 2b-d). Conversely, in the knockdown model, DKK3 expression was markedly downregulated in the DKK3-KD group, with levels reduced to approximately 0.44-fold (mRNA) and 0.58-fold (protein) relative to the CTRL group, and no difference was observed between the CTRL and NC-KD groups (Fig. 2g-i). These results confirm the successful modulation of DKK3 expression. Notably, as DKK3 is an inhibitor of canonical Wnt signaling, alterations in its expression significantly impacted the protein abundance of its downstream effector β-catenin^6,12. Corresponding to DKK3 expression changes, β-catenin protein levels were significantly reduced to approximately 0.20-fold in the DKK3-OE group (Fig. 2c,e), but were correspondingly upregulated to approximately 1.85-fold in the DKK3-KD group (Fig. 2h,j). These findings demonstrate that DKK3 critically regulates Wnt/β-catenin signaling activity in SiHa cells by controlling the protein abundance of β-catenin.

Fig. 2.

Establishment of DKK3-manipulated SiHa cell models with altered Wnt/β-catenin signaling activity. (a, f) Brightfield and fluorescence microscopy images after transfection with different lentivirus in SiHa cells cultured in medium (scale bar: 100 μm). (b, g) DKK3 mRNA expression in different treated cells (n = 3). (c-e, h-j) Western blotting and corresponding quantitative expression analysis of DKK3 and β-catenin (n = 3). Original blots are presented in Supplementary Fig. 1–2. *P < 0.05, **P < 0.01, and ***P < 0.001 represent significant difference.

Using these established models, we further assessed the impact of DKK3 on the malignant behaviors of SiHa cells. CCK-8 assays revealed that DKK3-OE cells exhibited a reduced proliferation rate as early as 24 h, and this suppressive effect sustained at 48 and 72 h, whereas proliferation rates were comparable between CTRL and NC-OE cells across all time points (Fig. 3a). Conversely, at the same time points, the proliferation rate of the DKK3-KD cells was accelerated compared to both CTRL and NC-KD cells (Fig. 3i). Colony formation assays further corroborated DKK3’s effect on proliferation: DKK3-OE cells yielded fewer colonies (Fig. 3b-c), whereas DKK3-KD cells showed a slight increase in colony formation (Fig. 3j-k).

Fig. 3.

Cell proliferation, migration and invasion are affected by DKK3 expression in vitro. (a, i) CCK-8 assay detected the ability of cell proliferation (n = 3). (b-c, j-k) Colony formation assay and count of different groups of SiHa cells(n = 3). (d-e, l-m) Representative wound healing assay images and analysis of different groups of SiHa cells (n = 3). (f-h, n-p) Representative cell migration and invasion images and analysis of SiHa cells from different groups using transwell assay (n = 3). *P < 0.05, **P < 0.01, and ***P < 0.001 represent significant difference.

Additionally, wound healing assays revealed slower wound closure over 48 h in DKK3-OE cells (approximately 79.5% of the CTRL group), whereas no significant difference was observed between the NC-OE and CTRL groups (Fig. 3d-e). Further validating these findings, transwell assays revealed that fewer DKK3-OE cells traversed the chambers compared to CTRL and NC-OE groups. Compared to the CTRL group, the number of traversing DKK3-OE cells was reduced to approximately 71.1% in uncoated (migration) chambers and 72.6% in Matrigel-coated (invasion) chambers. In both conditions, no significant difference was observed between the NC-OE and CTRL groups under either setting (Fig. 3f-h). Conversely, in the knockdown model, DKK3-KD cells exhibited enhanced migratory and invasive capacities compared to CTRL and NC-KD groups. This enhancement was evidenced by a higher wound closure ratio (approximately 1.30-fold compared to CTRL group) at 48 h (Fig. 3l-m) and more cells traversing the chambers in transwell migration (approximately 1.25-fold) and invasion assays (approximately 1.31-fold) (Fig. 3n-p). These results collectively demonstrate that DKK3 significantly modulates migratory and invasive capacities in SiHa cells.

FA suppresses malignant behaviors in SiHa cells by upregulating DKK3 to inhibit Wnt/β-catenin signaling activity

As schematized in Fig. 4a, we hypothesized that FA suppresses cervical cancer cells through the upregulation of DKK3, thereby inhibiting the canonical Wnt signaling and reducing β-catenin levels. To elucidate the anti-tumor effects of FA and their correlation with DKK3 expression, we established a FA-intervention cell model to conduct a series of experiments. First, the impact of FA at graded concentrations on SiHa cell viability was evaluated via CCK-8 assay. Relative to untreated cells, lower FA concentrations (200 µg/ml, 600 µg/ml) did not cause significant suppression of proliferative capacity, while higher concentrations (1000 µg/ml, 1400 µg/ml, 1800 µg/ml) inhibited cell proliferation, reducing cell viability to 87.56%, 85.23%, and 78.07%, respectively (Fig. 4b). After determining effective FA concentrations, we further assessed the effect of different intervention durations on proliferation. As shown in Fig. 4c, FA intervention for 6 and 12 h caused no significant changes in proliferation compared to untreated cells. After 24 h of intervention, an inhibitory difference first appeared in the 1800 µg/ml group. Extending intervention to 48 h resulted in inhibitory effects in all FA intervention groups (1000 µg/ml, 1400 µg/ml, 1800 µg/ml). Consistent with the suppression of cell viability, flow cytometric analysis detected an increase in apoptotic cells with FA intervention. Relative to approximately 5.05% in untreated cells, a modest increase was observed at 1000 and 1400 µg/ml FA intervention (approximately 5.99% and 6.48%, respectively), while a more pronounced increase to 10.62% was seen at 1800 µg/ml (Fig. 4d-e). Collectively, these results establish FA’s capacity to potently inhibit SiHa cell viability and promote apoptosis.

Fig. 4.

FA inhibits cell viability, migration and invasion by upregulating DKK3 and reducing β-catenin protein abundance. (a) schematic illustration of the DKK3 regulation after cells treated with FA. (b) Cell viability of SiHa cells treated with different concentrations of FA (n = 3). (c )Cell viability of SiHa cells treated with different time course of FA (n = 3). (d) Flow cytometry analysis for cell apoptosis of SiHa cells with different treatments (n = 3). (e) Representative images for flow cytometry of SiHa cells with different treatments. (f-g) Representative wound healing assay images and analysis of FA treatment in SiHa cells (n = 3). (h) Representative images of cell migration and invasion in SiHa cells treated with FA, using transwell assay. (i-j) Corresponding quantitative analysis of migration and invasion (n = 3). (k) DKK3 mRNA levels in SiHa cells treated with different concentrations of FA (n = 3). (l–n) Protein expression analysis of DKK3 and β‑catenin in SiHa cells under different FA treatments: representative western blot images (l) and corresponding quantification of DKK3 (m) and β‑catenin (n) (n = 3). (o) Rescue experiments: DKK3 mRNA levels in DKK3‑knockdown SiHa cells treated with different concentrations of FA (n = 3). (p–r) Rescue experiments: Protein expression analysis in DKK3‑knockdown SiHa cells under FA treatment: representative western blot images (p) and corresponding quantification of DKK3 (q) and β‑catenin (r) (n = 3). Original blots are presented in Supplementary Fig. 3–4.*P < 0.05, **P < 0.01, and ***P < 0.001 represent significant difference.

Next, we investigated the impact of FA intervention on SiHa cellular migratory and invasive properties. Wound healing assays indicated that FA intervention inhibited the wound closure of SiHa cells, with inhibitory effects observed in the 1000 µg/ml (approximately 59.19%), 1400 µg/ml (approximately 50.00%), and 1800 µg/ml FA groups (approximately 38.56%) compared to untreated cells (Fig. 4f-g). Analogously, transwell assays showed that FA intervention reduced the number of migrating cells compared to untreated cells in all FA intervention groups, with the number decreasing to approximately 46.03%, 30.70%, and 24.75% of the control in 1000, 1400, and 1800 µg/ml FA groups, respectively (Fig. 4h-i). This was also confirmed in invasion assays using matrigel coating, where the number of invading cells was reduced to approximately 64.87%, 49.33%, and 29.57% of the control in the respective FA intervention groups (Fig. 4h,j). All these findings consistently demonstrate the anti-tumor effect of FA intervention in SiHa cells.

Following this, we conducted RT-qPCR and Western blotting to explore the potential link between the anti-tumor effects of FA and DKK3 expression. As depicted in Fig. 4k, DKK3 mRNA expression showed no significant upregulation in the 1000 µg/ml FA group compared to untreated SiHa cells, but was elevated to approximately 2.17-fold and 3.47-fold in the 1400 µg/ml and 1800 µg/ml FA groups, respectively. Western blot analysis further indicated that DKK3 protein expression was elevated to about 1.57-fold in the 1000 µg/ml FA group, and continued to rise to 1.84-fold and 2.04-fold in the 1400 µg/ml and 1800 µg/ml FA groups (Fig. 4l–m). Notably, concomitant with the FA-induced upregulation of DKK3 protein, β-catenin protein expression was reduced to approximately 0.80-fold, 0.78-fold, and 0.63-fold in the 1000, 1400, and 1800 µg/ml FA intervention groups, respectively (Fig. 4l,n). Rescue experiments under partial DKK3-knockdown (DKK3-KD) conditions further demonstrated that FA intervention restored DKK3 expression. In these cells, DKK3 mRNA levels increased to about 1.27-fold, 1.68-fold, and 1.70-fold, while DKK3 protein expression rose to approximately 1.34-fold, 1.63-fold, and 2.18-fold in the 1000, 1400, and 1800 µg/ml FA-treated DKK3-KD cells, respectively (Fig. 4o–q). Correspondingly, the accumulation of β-catenin protein was reversed to about 0.85-fold, 0.74-fold, and 0.56-fold in these groups (Fig. 4p,r). In summary, these results indicate that FA exerts its anti-tumor effects predominantly through DKK3 upregulation, which suppresses Wnt/β-catenin signaling and reduces β-catenin protein abundance in cervical cancer.

Discussion

The prevention and treatment of cervical cancer are highly focused topics in the fields of medical research and public health^1,30. Investigating the mechanisms of carcinogenesis, identifying key molecular targets with warning and therapeutic value during malignant transformation, and exploring cost-effective targeted interventions against these targets are expected to become key components of a future precision medicine strategy in cervical cancer^2,5. DKK3 serves as a pivotal antagonist that inhibits canonical Wnt signaling^6,12. This study reveals the expression changes of DKK3 during cervical squamous epithelial carcinogenesis and establishes its value as a therapeutic target. Crucially, we demonstrate that FA intervention exerts anti-tumor effects in cervical cancer by specifically upregulating DKK3 expression, thereby suppressing Wnt/β-catenin signaling activation.

Reduced DKK3 expression is considered a key contributor to aberrant β-catenin protein accumulation in cervical cancer tissues^12,14. Previous studies on DKK3 in cervical cancer have focused on its expression characteristics after the development of invasive carcinoma and its association with adverse clinicopathological characteristics^13,14. However, its expression and significance during the precancerous stages have not yet received sufficient attention. In fact, the expression and function of molecules may vary with the stage of disease and histological type^31–33. Given that cervical cancers predominantly arise from malignant transformation of basal squamous epithelial cells, we systematically examined DKK3 expression changes across different stages of cervical epithelial lesions. Consistent with previous reports^13,14, we also observed decreased DKK3 expression in cancerous tissues and its association with higher FIGO stage and lymphovascular space invasion. Importantly, this study provides the first evidence of reduced DKK3 protein expression in both CIN1 and precancerous CIN2/3 tissues, with a further decline observed in invasive carcinoma tissues. These findings indicate that DKK3 downregulation is not only linked to late-stage events post-carcinogenesis but may also play a role in driving cervical malignant transformation during the early disease phase. Subsequently, we validated the effect of targeted regulation of DKK3 in in vitro experiments. Our data demonstrate that DKK3 overexpression in SiHa cells significantly reduces β-catenin protein abundance while effectively suppressing proliferation, migration, and invasion capabilities. Conversely, DKK3 knockdown elevated β-catenin levels and enhanced malignant phenotypes. Taken together, our findings establish DKK3 as a critical therapeutic target that suppresses the Wnt/β-catenin signaling and reduces β-catenin protein abundance, exerting tumor-suppressive effects in cervical cancer. Targeted restoration of DKK3 expression may therefore constitute a viable interventional approach for preventing and treating cervical cancer.

As an essential micronutrient, FA exerts profound influence on human health maintenance and disease prevention through diverse regulatory mechanisms^16,34,35. Currently, over 80 countries worldwide have implemented FA supplement fortification programs to reduce the risk of various diseases and have achieved positive outcomes³⁴. As research advances, accumulating evidence reveals significant disease-context dependency in FA’s biological effects^36,37. While most studies indicate FA supplement fortification either lowers disease risk or has neutral effects across diverse disease models, supraphysiological folate levels may exert adverse outcomes in specific contexts, which raises concerns about systemic overexposure risks^36–39. Notably in cervical cancer research, multidisciplinary studies have provided highly consistent evidence confirming the protective role of FA in cervical carcinogenesis^15,17–22. Therefore, we established in vitro FA intervention models to verify its effects and investigate underlying molecular mechanisms in cervical cancer. Our results confirm that FA intervention exerts anti-tumor effects in cervical cancer, significantly inhibiting SiHa cell proliferation, migration, and invasion, along with increasing apoptosis, within a specific concentration range. These findings suggest that appropriate FA intervention holds potential as a low-toxicity, cost-effective adjuvant strategy for cervical cancer prevention and treatment.

Evidence from multiple disease models suggests that the canonical Wnt signaling pathway may act as a key mediator of FA’s regulation of disease progression, particularly through β-catenin protein stability control^23,26,29,40. Nevertheless, the response of the Wnt/β-catenin signaling to FA is highly dependent on the disease context, which potentially explains its divergent effects across pathologies. The current understanding primarily stems from embryonic development research^26–28. Studies on congenital heart defects have reported that FA intervention can suppress teratogen-induced overactivation of the Wnt/β-catenin pathway, thereby preventing birth defects^27,28. This protective role has been corroborated in neural tube defects research, where folate deficiency activated canonical Wnt signaling, leading to impaired neural tube closure²⁶. However, relevant research evidence remains scarce and inconsistent in oncological research. A clinical trial on colorectal adenomas demonstrated that FA reduced β-catenin nuclear translocation and slowed tumor progression²⁴, whereas a recent animal study yielded conflicting results, suggesting that FA might promote β-catenin activation and colitis-associated colorectal cancer under specific conditions of chronic colonic inflammation²³. This striking discrepancy underscores the necessity of in-depth investigation into the FA-Wnt pathway regulatory relationship within specific disease contexts.

In the context of cervical cancer, a recent study provided a crucial clue by demonstrating that folate deficiency increased β-catenin protein levels and promoted malignant behaviors in cervical cancer cells²⁹, which is consistent with our findings. Our study provides functional insights into the protective mechanism of FA in cervical cancer from a novel perspective. We demonstrated that FA intervention suppressed Wnt/β-catenin pathway activity and reduced β-catenin abundance. Crucially, we identified and validated DKK3 as the pivotal molecular mediator linking FA action to Wnt signaling regulation. We demonstrated that FA intervention dose-dependently upregulated both DKK3 mRNA and protein expression in SiHa cells, with a concurrent decrease in β-catenin protein abundance. Even more compelling evidence emerged from rescue experiments, where under conditions of partial DKK3 knockdown, FA intervention still markedly enhanced the remaining DKK3 expression and partially counteracted the β-catenin accumulation resulting from DKK3 deficiency. All these findings demonstrate that the suppression of the Wnt/β-catenin pathway and the consequent anti-tumor effects of FA intervention are largely dependent on its ability to upregulate DKK3 expression. Therefore, monitoring DKK3 changes during cervical epithelial carcinogenesis may help identify potential beneficiaries of FA intervention. Based on this, implementing appropriate FA intervention tailored to individual profiles at optimal timing could potentially prevent further malignant progression or serve as an effective adjuvant therapy.

Overall, this study offers the first systematic elucidation of DKK3 expression dynamics throughout the process from the occurrence of cervical intraepithelial lesions to cervical squamous epithelial carcinoma. Furthermore, it identifies FA intervention as an effective means to restore DKK3 expression, inhibit the Wnt/β-catenin signaling, and thereby exert anti-tumor effects in cervical cancer. These findings offer significant support for the future development of FA-based targeted intervention strategies for cervical cancer. However, while this study has preliminarily confirmed the efficacy of FA intervention and its molecular mechanism in SiHa cells, the detailed mechanisms governing FA-mediated DKK3 regulation (e.g., epigenetic regulation, transcriptional control) and the general applicability of this regulatory axis across different subtypes and pathological stages of cervical cancer remains to be fully elucidated. The optimal modes of FA intervention and its therapeutic efficacy also require more rigorous validation in vivo. Unlike deeper-seated organs like those within the gastrointestinal tract, the cervix’s uniquely superficial anatomical location enhances the feasibility of developing localized intervention strategies to minimize systemic exposure risks. Future research should focus on deeper mechanistic insights into FA-mediated DKK3 regulation and critically evaluate the translational potential of localized FA intervention for the prevention and treatment of cervical carcinogenesis.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Q.L. and Z.W. conceived and designed the study. Q.L., J.L., X.G., and B.W. collected the clinical samples. Q.L. performed the experiments and analyzed the data. J.L. and X.G. assisted in data analysis. Q.L. wrote the initial manuscript draft. Z.W. critically revised the manuscript. All authors reviewed, edited, and approved the final manuscript.

Funding

This research received funding from the International Cooperation Project of key R & D Program in Shanxi Province (201903D421065) and the Natural Science Research Program of Shanxi basic Research Program (20210302123271).

Data availability

Data is provided within the manuscript or supplementary information files. For more detailed information, reasonable requests can be made to the corresponding author via email.

Declarations

Conflict of interest

The authors disclose no competing interests.

Ethics statement

This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Second Hospital of Shanxi Medical University (approval number: 2019 YX No.309, Taiyuan, China). All participants involved in this study were fully informed about the study’s purposes and procedures, and provided written informed consent prior to participation.