The project was approved by the Ethics Committee of China-Japan Friendship Hospital (approval number: 2020–28-K20). One hundred thirty-four histological specimens, including hysterectomy tissues and loop/cone biopsied samples, were retrieved from the files of the Department of Pathology of China-Japan Friendship Hospital. The specimens consisted of normal cervix (n = 20), CIN1 (n = 36), CIN2/3 (n = 32), and invasive SCC (n = 46) tissues.

Four-micrometer-thick formalin-fixed, paraffin-embedded tissue sections were deparaffinized with xylene and rehydrated through a graded series of ethanol. Antigen retrieval was performed with 0.01 M citrate buffer (pH 6.0) in a microwave oven for 10 min. After cooling to room temperature, the slides were incubated with 3% H₂O₂ for 10 min to inactivate endogenous peroxidase. The slices were incubated overnight at 4°C in a humidified chamber with the following primary antibodies: 1:500 diluted anti-Wnt3a (ab28472), 1:500 diluted anti-β-catenin (ab32572), 1:1000 diluted anti-CBP (ab50702), 1:100 diluted anti-cytokeratin 17 (CK 17, ab53707), 1:1000 diluted anti-involucrin (ab53112), 1:200 diluted anti-CDKN2A/p16INK4α (p16, ab54210), and 1:200 diluted anti-Ki67 (ab16667). All of the primary antibodies were purchased from Abcam, United States. The binding of the primary antibodies was visualized using the ChemMate detection kit (PV-9000, ZSBIO, China). The slices were lightly counterstained with Mayer’s hematoxylin for 30 s.

Immunoreactivity was semiquantitatively evaluated. The degree of positive staining for all antibodies was evaluated by score on a scale of 0–3 for strength of intensity of staining and on a scale of 0–4 for percentage of positive cells. The final total score was generated by the intensity score × proportion score. The intensity score was defined as follows: 0, no staining of epithelial cells; 1, weak staining; 2, moderate staining; or 3, strong staining. The proportion score was defined as follows: 0, no staining of the cervical epithelial cells in any field; 1, <10% of the epithelium stain-positive; 2, 11–50% stain-positive; 3, 51–80% stain-positive; or 4, >80% stain-positive. The total score ranged from 0 to 12. Two different pathologists evaluated all the specimens in a blinded manner. Notably, the distribution of staining in the intraepithelial lesions (CIN1 and CIN2/3) was compared with that of the normal cervical squamous epithelium. CIN tissues were categorized as having the thickness of the samples being from two-thirds to fully stained.

To investigate the expression pattern of genes associated with the activation of Wnt/β-catenin signaling in the progression of cervical lesions, the data set GSE64217 (containing 2 normal cervix, 2 CIN2/3, and 2 cervical squamous cell cancer samples), GSE51993 (containing 7 HPV- normal cervix samples, 9 HPV + CIN I samples and 8 HPV + CIN III samples), GSE67522 (containing 11 HPV16- normal cervix samples, 11 HPV16+ non-malignants and 20 HPV16 + cervical cancer tissues) were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The platforms for these data sets are all based on GPL10558 and an Illumina HumanHT-12 V4.0 Expression BeadChip. The downloaded files were processed using the R package and then calibrated, standardized, and converted to log2 data, or normalized with cubic spline method. The limma package was used to standardize the gene chip and analyze the differential expression. The quantile normalized values for canonical Wnt signaling pathway associated genes were found out via VLOOKUP from the gene expression description tables. Boxplot was used to show the expression difference among groups of normal cervix, CIN and SCC.

The data are presented as the means ± SD. Statistical analyses were performed with SPSS version 25 (SPSS Inc., Chicago, IL, United States). The significance of gene expression between groups in microarray data sets were analyzed with one-way analysis of variance. The comparison of immunoreactivity score between normal cervix and cervical lesions were performed with two-tailed Chi-square test. Pearson test was used to analyze the correlation between the expression of the biomarkers and the cervical lesion degree. A p value <0.05 was considered significant.

To identify the differentially expressed genes (DEGs), the data set GSE64217, GSE51993 and GSE67522 were analyzed by the GEO2R method on the GEO website. The boxplot of the distribution of the normalized expression values for selected samples demonstrates that the data were suitable for the analysis and “limma” (Figure 1A), which was used to perform a top table computation to extract a list of the top-ranked genes. Within the top 250 differentially expressed genes, the canonical Wnt signaling genes Wnt3a, CTNNB1 (β-catenin), and CREBBP (CBP) were significantly upregulated in the samples of precancerous and cancer of the cervix compared with those of the normal cervix, thereby contributing to the distinct upregulation of their downstream transcription factors KLF4, MYC, and SOX2, except KLF4 slightly downregulated in CIN (Figure 1B). As previously reported, CBP/β-catenin signaling is important for the maintenance of stem cell pluripotency [6]. KRT17 (CK17), TP63 and CDK2A (P16), markers of HPV target cells/cervical cancer stem cells (CSCs), were apparently upregulated in CIN and SCC. In contrast, markers of squamous cell differentiation, LORICRIN, INVOLUCRIN and KRT4 (CK4), were significantly downregulated in CIN and SCC (Figure 1B). It was somewhat unusual that KRT14 (CK14) and KLF4, showed in these GSE gene sets, were moderately downregulated in CIN tissues compared with normal cervix, nevertheless they were all significantly upregulated in SCC tissues (Figure 1B).

In normal cervical squamous epithelium, Wnt3a staining was confined to the basal cells, whereas in CIN1 diffuse immunoreactivity was limited to the lower one-third of the epithelial layer, and in CIN2/3, the staining extended through the full thickness of the epithelial layer (Figure 2A). Strong to moderate Wnt3a immunoreactivity was observed in 45 (97.8%) of 46 invasive SCCs and 26 (81.3%) of 32 CIN2/3, while 44.4% (16/36) CIN1 showed light-to-moderate expression, as occasional scattered staining was observed within basal cells in normal cervix that was considered negative (Table 1).

A gradual increase in the β-catenin score and its positive translocation from the membrane to the cytoplasm and finally to the nucleus was noted in cervical lesions at higher stages (Figure 2A). Only 5% (1/20) of the normal cervix samples were membrane positive for β-catenin in the parabasal cells, while one-half (18/36) of the CIN1 samples were membrane positive for β-catenin. Approximately 71.9% (23/32) of the CIN2/3 and 84.8% (39/46) of the SCC samples had cytoplasmic and even nuclear positive staining. The pattern of CBP expression in each stage of cervical lesion was similar to that of β-catenin with only nuclear immunoreactivity (Table 1). Overall, the total scores of Wnt3a, β-catenin and CBP immunoreactivity were significantly higher in CIN2/3 and cervical cancer tissues compared to the normal cervix and CIN1 tissues, respectively (Figure 2B, p < 0.01).

Markers useful in the assessment of lesion grade were further observed. P16, which is valuable in the diagnosis of high-grade CIN, was positively stained in the upper two-thirds and full thickness of the epithelium in most cases of CIN2/3 and in all of cervical cancer tissues, while minimal or negative P16 was observed in cervical squamous epithelium of the normal cervix or CIN1 samples (Figure 3A). As the Ki67 antigen is expressed in all active phases of the cell cycle, diffuse immunoreactivity was distributed throughout SCC and the full thickness of the CIN2/3 samples, while scattered positive staining was confined to only the lower one-third of the CIN1 and the basal cells in the normal cervix samples (Figure 3A). Statistical analysis showed that the IHC scores of P16 and Ki67 were significantly higher in both the CIN2/3 and SCC tissues compared to the CIN1 and normal cervix (Figure 3B, p < 0.01).

In this work, the characterization of β-catenin and coactivator CBP expression implied that the canonical Wnt signaling pathway was activated in high-grade lesions. Consistent with previous studies, CBP-Wnt activity was associated with tumorigenesis and stem cell maintenance because it modulates the levels of intrinsic pluripotency factors, such as Oct4, Nanog, and Sox2 [6]. Here, CK17, a marker of reserve cells and TICs [23, 24], was expressed in the basal compartment of the normal cervix and CIN1 samples, while it displayed diffuse immunoreactivity in the SCC and CIN2/3 samples (Figure 4). On the other hand, the differentiation marker Involucrin was extensively expressed in the upper two-thirds of the squamous epithelium in the normal cervix and CIN1 samples, while minimal immunoreactivity was observed in the CIN2/3 or SCC samples (Figure 4).

The heatmap, which was made with the total IHC scores of various biomarkers and the cervical lesion grade, apparently displayed the association between them (Figure 5A). Spearman test further verified that the expression of Wnt3a, β-catenin, CBP, P16, Ki67, and CK17 were positively correlated with the cervical lesion grade (r = 0.679, 0.577, 0.688, 0.898, 0.892, 0.867, respectively, all p value <0.01), while the expression of Involucrin was negatively correlated with it (r = −0.557, p < 0.01) (Table 2). Taken together, these data suggest that Wnt/β-catenin/CBP is activated in CIN2/3 lesions and SCC, where it maintains TIC pluripotency and promotes cell proliferation, thereby contributing to neoplastic development.