Abstract
Objective
Cervical intraepithelial neoplasia grade 3 (CIN 3)-like squamous cell carcinoma (SCC) is a recently identified subtype of cervical cancer with a deceptive growth pattern. It mimics the phenotype of CIN 3, involving endocervical crypts, presenting significant challenges and potential for diagnostic confusion. This study aims to clarify the differences in epigenetic profiles and HPV viral load(VL) among CIN 3, CIN 3-like SCC, early invasive SCC, and invasive SCC.
Methods
A retrospective analysis was conducted on 138 cases of cervical high-grade lesions and SCC diagnosed at Peking University People's Hospital from March 2022 to June 2024. The cases included 65 high-grade squamous intraepithelial lesions (HSIL) [23 CIN 2 and 42 CIN 3] and 73 SCC [27 CIN 3-like SCC, 21 early invasive SCC, and 25 invasive SCC]. Histological samples were analyzed for methylation levels of the epigenetic factors PAX1 and JAM3, as well as HPV VL.
Results
Methylation levels of PAX1 and JAM3, as well as total HPV VL, were significantly higher in cervical cancer than in HSIL(< 0.001). The PAX1 methylation level of CIN3 was significantly lower than that of invasive SCC, but there was no significant difference compared with CIN3 and CIN 3-like SCC; For the PAX1/JAM3 methylation levels in CIN 3-like SCC, no statistically significant differences were observed when compared with those in early invasive SCC and invasive SCC. The HPV VL in CIN 3-like SCC was significantly higher than that in CIN 3 and early invasive SCC, but no significant difference was found compared with invasive SCC. In the same stage of FIGO, although PAX1/JAM3 methylation and HPV VL (especially HPV16/18 VL) showed a tendency to increase in CIN3-like SCC compared with early invasive SCC, the difference was not significant, further expansion of sample size is needed to confirm this.
Conclusion
PAX1 and JAM3 methylation levels are significantly higher in SCC than HSIL, but no difference in methylation levels was observed between CIN 3-like SCC, early invasive SCC, and invasive SCC. Total HPV VL (especially HPV16/18 VL) in CIN3-like SCC is significantly higher than that in CIN3 and early invasive SCC, suggesting potential shared and distinct features in epigenetics and HPV VL between CIN3-like SCC and early invasive SCC, which need more molecular biology research to explore the difference.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s13148-026-02201-1.
Keywords: Cervical intraepithelial neoplasia (CIN), CIN3-like squamous cell carcinoma, PAX1, JAM3, Methylation, HPV viral load
Introduction
Cervical cancer remains one of the most common female cancers globally, particularly in countries lacking effective cervical cytology screening or human papillomavirus (HPV) vaccination programs [1]. According to the 2020WHO classification [2], cervical high-grade squamous intraepithelial lesions (HSIL) include cervical intraepithelial neoplasia(CIN) 2 and CIN 3. CIN 3 is considered a true precancerous lesion due to its higher risk of progression and high risk HPV(hrHPV) genotype distribution in the population being more similar to cervical cancer [3]. Its most common malignant transformation is squamous cell carcinoma (SCC), defined by diagnostic criteria such as infiltrating the basement membrane.
A recently recognized special type of SCC, termed CIN 3-like SCC, is a rare variant of cervical squamous cell carcinoma often misdiagnosed as CIN 3 with glandular involvement. The most remarkable feature of CIN3-like SCC is the presence of large cavities at the center of cell nests, accompanied by necrosis and dyskeratosis. Histopathologically, its growth pattern differs from traditional invasive patterns, exhibiting a pushing growth pattern that can infiltrate the superficial or deep muscular layers of the cervical wall. Due to the lack of recognition and emphasis on this unique invasive pattern, underestimation of the tumor's invasive scope and depth is inevitable, especially in small biopsy tissues. This underdiagnosis may lead to delayed surgery and disease progression [4]. Studies have recently suggested that CIN 3-like tumors may demonstrate "collective cellular invasion," characterized by retention of an epithelial phenotype [5]. Our team’s prior analysis of histomorphology and immunohistology in 23 cases of CIN 3-like SCC revealed slightly differences in staining intensity and distribution of immunophenotypic markers (E-cadherin, p16, cyclin D1, and p53) compared to invasive SCC, but no differential statistical analysis was performed [6].In molecular targeted testing, methylation assays and HPV viral load (VL) have emerged as actively studied biomarkers in cervical cancer screening. The paired box gene 1 (PAX1) is recognized as a tumor suppressor gene in multiple cancers [7]. Studies [8, 9] have shown that PAX1 methylation increases with the severity of cervical lesions, especially in cervical cancer. Another marker, junctional adhesion molecule 3 (JAM3), has demonstrated utility in stratified management of hrHPV-positive cases with cytological abnormalities and as a marker for lesion progression and prognosis [10]. Research indicates that JAM3 methylation may serve as a novel biomarker guiding grading and treatment for hrHPV-positive patients [11]. Our team previously found a significant correlation between PAX1 and JAM3 methylation (p < 0.001), suggesting their use as cumulative evidence of prolonged HPV infection prior to precancerous lesions [12].
HPV VL testing quantifies viral infection severity to some extent. 2019 U.S. FDA document established a correlation between hrHPV viral load and cervical precancerous lesions, particularly for HPV 16-related A9 genotype family members (HPV 16/31/33/52/58) [13]. Our group also found that PAX1 hypermethylation is associated with high HPV VL, especially for HPV 16/18, and both show superior specificity in detecting CIN 2 + [9].
The present study was undertaken to investigate whether differences exist in methylation status and HPV viral load among CIN 3-like SCC, CIN 3, early invasive SCC, and invasive SCC.
Materials and methods
Case collection and definition
This study was approved by the Ethics Committee of Peking University People's Hospital. Patients were informed before surgery that resection specimens would be preserved by the Department of Pathology at Peking University People's Hospital. A retrospective analysis was performed on 138 cases treated at Peking University People's Hospital from March 2022 to June 2024. All pathological slides were independently reviewed by two experienced gynecological pathologists (Zhang and Shen) according to the 5th edition of the WHO Classification of Tumors of the Female Reproductive System [2], and any cases with controversial or unclear diagnoses were excluded. The clinical and pathological staging of cervical cancer follows the International Federation of Gynecology and Obstetrics (FIGO) 2018 classification [14].
Definition
CIN 3-like SCC was defined as tumors with a pushing growth pattern (in contrast to conventional invasive patterns), surrounded by lymphocyte/plasma cell infiltration and mild stromal reaction. On high-power microscopy, cell nests resembled HSIL, with central cavities containing parakeratotic or necrotic material, mimicking CIN 3 involving endocervical glands [4]. Early invasive SCC was defined according to the WHO classification as microscopic foci of invasion with depth < 5 mm and no lympho-vascular space invasion(LVSI) [2]. Invasive SCC was defined as invasion with depth > 5 mm, with or without LVSI [2]. Figure 1 shows the key points for differentiation of the three types: CIN3 with glandular involvement, early invasive SCC, and CIN3-like SCC. When both types of lesions are present, we use the depth of invasion as the grouping criterion—the lesion with the greater depth of invasion determines the group assignment.
Fig. 1.
HE stains of CIN3 with glandular involvement, early invasive SCC, and CIN3-like SCC. A HSIL/CIN3, with glandular involved(HE × 200); B CIN 3-like squamous cell carcinoma. The cell nests have clear boundaries, with a cavity visible in the center, and dyskeratotic cells are present within the cavity (HE × 100); C Early invasive squamous cell carcinoma, with small, irregular, and scattered infiltration foci (HE × 200)
Methylation and HPV viral load
Retrospective paraffin-embedded tissue blocks were utilized to extract genomic DNA for methylation detection. Briefly, each paraffin block was sectioned into six to eight thin slices with a thickness of 4 µm. Subsequently, xylene was employed to dissolve the paraffin within the tissue sections, followed by re-extraction using 100% ethanol to completely remove the dissolved paraffin. This xylene-ethanol extraction process was repeated to ensure the purity of the resulting DNA. Genomic DNA (gDNA) was extracted from the tissue samples using the JH-DNA Isolation and Purifying Kit (OriginPoly Bio-Tec Co., Ltd., Beijing, China) according to the manufacturer’s instructions. DNA concentration was quantified using a NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Subsequently, 200–1000 ng of gDNA underwent bisulfite conversion using the JH-DNAMethylation-Lightning MagPrep Kit (OriginPoly Bio-Tec Co., Ltd., Beijing, China) following the provided protocol. PAX1 methylation (PAX1m) and JAM3 methylation (JAM3m) levels were measured using the "Human PAX1 and JAM3 Gene Methylation Detection Kit (Real-Time PCR)" for cervical cancer [Class III medical device approved by the National Medical Products Administration (No. 20233400253)], with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an internal control (OriginPoly Bio-Tec Co., Ltd., Beijing, China). Analysis was conducted on an SLAN-96S automatic medical PCR system (Shanghai Hongshi Med Tech Co., Ltd., Shanghai, China) according to the manufacturer’s guidelines. Hypermethylation of PAX1 or JAM3 was determined by the ΔCt value (ΔCtPAX1 = CtPAX1 − CtGAPDH; ΔCtJAM3 = CtJAM3 − CtGAPDH). Positive PAX1/JAM3 hypermethylation was defined as ΔCtPAX1 ≤ 6.6 or ΔCtJAM3 ≤ 10.0. A lower ΔCt value corresponds to a higher abundance of methylated PAX1 and JAM3. e.
HPV genotyping and viral load assessments were conducted using the BioPerfectus Multiplex Real-Time PCR (BMRT), a method validated to be as sensitive as the Cobas 4800 for primary cervical cancer screening [33]. Nucleic acid extraction used a starting sample volume of 300 µL, with an elution volume of 85 µL. The BMRT employs a multi-fluorescence HPV DNA genotyping kit (Bioperfectus Ltd., Taizhou, China) that targets the HPV L1 gene for 21 prevalent HPV genotypes, including 13 high-risk genotypes (HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68), 5 medium-risk genotypes (HPV 26, 53, 66, 73, 82), and 3 low-risk genotypes (HPV 6, 11, 81). In this study, a 14-genotype high-risk BMRT assay was utilized. A single-copy human TOP3 gene was co-amplified to ensure DNA quality control and to calculate relative cell copy numbers. HPV type-specific viral loads were normalized using the formula: VL = log10[(CnHPV/CnTOP3) × 10,000] copies/10,000 cells, where CnHPV represents the quantity of HPV DNA and CnTOP3 denotes the human cell count. The experimental protocol was consistent with previous reports [9, 12].
Statistical methods
Statistical analyses were performed using R programming software (version 4.4.3). The data analysis and visualization in this study were performed using R (version 4.5.1). The core R packages utilized include: tidyverse (version 2.0.0) for data manipulation and integration, ggplot2 (version 4.0.0) for graphic construction, and ggpubr (version 0.6.1) for plot composition and aesthetic enhancement. Samples were stratified into two groups based on the duration of HPV infection. Descriptive statistics were employed to characterize the samples, with categorical variables presented as frequencies and percentages (e.g., high-risk HPV genotypes). Quantitative data were expressed as mean ± standard deviation (e.g., age), while non-normally distributed continuous variables—including all HPV viral loads [log(all-hrHPV)], HPV16/18 viral loads [log (HPV16/18)], viral loads of the other 12 high-risk HPV types [log(other-hrHPV)], PAX1 methylation (PAX1m), and JAM3 methylation (JAM3m)—were described using median and interquartile range (IQR). The Wilcoxon test and Kruskal–Wallis test were used to compare continuous variables between two independent groups. The Fisher’s exact test and Pearson’s Chi-squared test were used to compare categorical variables. Univariate and multivariate logistic regression analyses were conducted to calculate the risk of each variable for the lesions and their corresponding odds ratios (ORs). Pearson's correlation test was used to examine the correlations between the ΔCt values of methylated gene markers and other variables. All statistical tests were two-sided, and a p-value < 0.05 was considered statistically significant.
Results
Clinicopathological characteristics
Among 138 cases, there were 65 cases of high-grade squamous intraepithelial lesions (23 cases of CIN2 and 42 cases of CIN3), and 73 cases of cervical SCC (including 21 cases of early invasive SCC, 27 cases of CIN3-like SCC, and 25 cases of invasive SCC). The median age of onset for HSIL lesions was 49 years (CIN2 47 years, and CIN3 45 years), while that for cervical squamous cell carcinoma was 52 years (early invasive:52 years, CIN3-like SCC:50 years, invasive cancer: 54 years), with a statistically significant difference between the two groups (p = 0.009). The total positive rate of HPV is 94%(130/138), that of HSIL is 90.8% (59/65), and the incidence rate of cervical cancer is 97.3% (71/73). The positive rate of HPV16/18 is 70.3% (97/138, HPV16: 91 cases, HPV18: 10 cases, among which 4 cases were positive for both HPV16 and HPV18), that of HSIL is 60% (39/65), and the incidence rate of cervical cancer is 79.5% (58/73), there was no statistically significant difference in the positive rate of HPV and the rate of HPV16/18 among different pathological types (Table 1).
Table 1.
Characteristics with all pathological phenotypes
| Pathology | ||||||||
|---|---|---|---|---|---|---|---|---|
| Characteristic | N | Overall N = 1381 |
CIN2 N = 231 |
CIN3 N = 421 |
Early invasive N = 211 |
CIN3-like-SCC N = 271 |
Invasive N = 251 |
p value |
| Age, mean (SD) | 138 | 49 (12) | 47 (8) | 45 (12) | 52 (9) | 50 (11) | 54 (14) | 0.0092 |
| PAX1, mean (SD) | 138 | 5.90 (4.44) | 9.83 (4.16) | 5.52 (3.51) | 6.52 (5.23) | 4.54 (3.07) | 3.86 (4.57) | < 0.0012 |
| JAM3, mean (SD) | 138 | 10.01 (4.07) | 12.77 (2.47) | 10.79 (3.82) | 9.93 (4.37) | 8.50 (3.43) | 7.83 (4.44) | < 0.0012 |
| HPV, n(%) | 138 | 0.153 | ||||||
| Negative | 8(5.8%) | 3(13%) | 3(7.1%) | 2(9.5%) | 0(0%) | 0(0%) | ||
| Positive | 130(94.2%) | 20(87%) | 39(92.9%) | 19(90.5%) | 27(100%) | 25(100%) | ||
| HPV 16/18, n(%) | 138 | 0.0394 | ||||||
| Negative | 41(29.7%) | 12(52.2%) | 14(33.3%) | 5(23.8%) | 7(25.9%) | 3(12%) | ||
| Positive | 97(70.3%) | 11(47.8%) | 28(66.7%) | 16(76.2%) | 20(74.1%) | 22(88%) | ||
| Non-HPV 16/18, n(%) | 138 | 0.194 | ||||||
| Negative | 105(76.1%) | 14(60.9%) | 31(73.8%) | 18(85.7%) | 20(74.1%) | 22(88%) | ||
| Positive | 33(23.9%) | 9(39.1%) | 11(26.2%) | 3(14.3%) | 7(25.9%) | 3(12%) | ||
| HPV 16/18 loading (log transformed), mean (SD) | 138 | 6.57 (5.10) | 4.23 (4.89) | 6.59 (5.13) | 6.55 (4.68) | 6.85 (5.16) | 8.41 (5.07) | 0.0402 |
| Non-HPV 16/18 loading (log transformed), mean (SD) | 138 | 5.06 (3.47) | 6.03 (3.32) | 4.53 (2.93) | 4.17 (3.21) | 6.21 (4.35) | 4.57 (3.34) | 0.122 |
| HPV loading (log transformed), mean (SD) | 138 | 9.36 (3.05) | 7.98 (3.12) | 8.89 (3.05) | 8.62 (3.18) | 10.95 (1.68) | 10.32 (3.22) | < 0.0012 |
HPV, human papillomavirus; CIN, cervical intraepithelial neoplasia; SCC, squamous cell carcinoma
1Median (IQR), Mean (SD), or Frequency (%)
2Kruskal-Wallis rank sum test
3Fisher's exact test
4Pearson's Chi-squared test
Totally, 77 cases were cervical conization specimens, and 61 cases were total hysterectomy specimens with bilateral adnexa and pelvic lymph node resection. The proportions of SCC cases in stages IA1 (38.4%), IA2 (6.8%), IB1 (16.4%), IB2 (21.9%), IB3 (6.8%), IIA2 (1.4%), IIB (1.4%), and IIIC1 (6.8%) are as follows in Table 2. In the Early invasive SCC, the IA1 stage accounts for 85.7% (18/21). In the CIN3-like SCC, all cases are within IB3; the IA stage accounts for 44.4% (12/27). Both LVSI and lymph node metastasis are negative in early invasive SCC and CIN3-like SCC (Table 2).
Table 2.
Clinicopathologic characteristics of all types of squamous cell carcinoma
| Pathology | ||||
|---|---|---|---|---|
| Characteristic |
Early invasive N = 21 |
CIN3-like-SCC N = 27 |
Invasive N = 25 |
p value |
| FIGO stage, n/N (%) | ||||
| IA1 | 18/21 (85.7%) | 10/27 (37%) | – | |
| IA2 | 3/21 (14.3%) | 2/27 (7.4%) | – | |
| IB1 | / | 4/27 (14.8%) | 8/25 (32%) | |
| IB2 | – | 10/27 (37.0%) | 6/25 (24%) | |
| IB3 | – | 1/27 (3.7%) | 4/25 (16%) | |
| IIA2 | – | – | 1/25 (4%) | |
| IIB | – | – | 1/25 (4%) | |
| IIIC1 | – | – | 5/25 (20%) | |
| LVSI, n/N (%) | < 0.0011 | |||
| Negative | 20/21 (95.2%) | 27/27 (100%) | 12/25 (48%) | |
| Postive | 1/21 (4.8%) | – | 13/25 (52%) | |
| Lymph node metastatsis, n/N (%) | 0.0051 | |||
| Negative | 21/21 (100%) | 27/27 (100%) | 20/25 (80%) | |
| Positive | – | – | 5/25 (20%) | |
FIGO, International federation of gynecology and obstetrics; LVSI, lymphovascular space invasion; CIN, cervical intraepithelial neoplasia; SCC, squamous cell carcinoma
1Fisher’s exact test
Among 27 cases of CIN3-like SCC, 16 cases were detected after conization, 10 cases were found in total hysterectomy specimens, and 1 case was discovered in residual cervix removed. An interesting finding was that among the 6 available preoperative colposcopy images, 4 cases had a history of previous conization (1–2 procedures) performed at other hospitals for HSIL, see Fig. 2. The pathological results indicate that among 27 cases of CIN3-like SCC, 2 cases were combined with invasive cancer, and one case was combined with adenocarcinoma in situ (AIS).
Fig. 2.
Four cases had a history of cervical conization, with an average interval of 6.5 years after the first conization
Methylation levels of PAX1 and JAM3 in HSIL and squamous cell carcinoma
Methylation levels of PAX1 and JAM3 were significantly higher in cervical cancer than in HSIL(< 0.001). In HSIL, PAX1/JAM3 methylation levels were significantly lower than those in CIN 3-like SCC and invasive SCC, but no statistically significant difference was observed between HSIL and early invasive SCC. If HSIL is divided into CIN2 and CIN3: The PAX1 methylation levels in CIN2 were lower than those in CIN3 and all SCC groups. The PAX1 methylation level of CIN3 was significantly lower than that of invasive carcinoma, but there was no significant difference compared with CIN3 and CIN 3-like SCC; The JAM3 methylation level in CIN3 was lower than that in CIN3-like SCC and invasive SCC, but showed no statistically significant difference compared with early invasive SCC. For the PAX1/JAM3 methylation levels in CIN 3-like SCC, no statistically significant differences were observed when compared with those in early invasive SCC and invasive SCC (Fig. 3A, B).
Fig. 3.
PAX1/JAM3 methylation in different cervical lesion/cancer types. Box plots with scatter points showing the ΔCt values of PAX1 (A) and JAM3 (B) methylation across different grades of cervical lesion/cancer types.Statistical significance is denoted by asterisks: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. ΔCt values reflect the relative methylation levels of PAX1 or JAM3, the smaller ΔCt values reflect the higher level of methylation
HPV viral load in HSIL and squamous cell carcinoma
Total HPV VL and HPV 16/18 VL were both weakly positively correlated with PAX1/JAM3 methylation levels (Fig. 4). In HSIL, total HPV VL were significantly lower than those in CIN 3-like SCC(p < 0.001) and invasive SCC(p < 0.05), but no statistically significant difference was observed between HSIL and early invasive SCC. The HPV VL in CIN 3-like SCC was significantly higher than that in CIN 3 and early invasive SCC, but no significant difference was found compared with invasive SCC. The VL of HPV16/18 was higher in invasive SCC than in HSIL, and this difference was primarily observed between CIN2 and invasive SCC, while other comparisons showed no statistical significance. Similarly, For non-HPV16/18 VL, only CIN2 and early invasive SCC showed statistically significant differences (Fig. 5A–C).
Fig. 4.
Correlation between PAX1/JAM3 methylation and HPV viral load. Scatter plots showing the correlation between PAX1/JAM3 methylation levels (ΔCt values) and HPV viral load (log-transformed) in cervical samples, with blue regression lines and gray shaded areas representing 95% confidence intervals. A Correlation between PAX1 methylation (ΔCt) and overall HPV viral load (log-transformed), with a correlation coefficient R = − 0.42 and p = 2e−07. B Correlation between JAM3 methylation (ΔCt) and overall HPV viral load (log-transformed), with R = − 0.47 and p = 4.8e−09. C Correlation between PAX1 methylation (ΔCt) and HPV 16/18 viral load (log-transformed), with R = − 0.27 and p = 0.0014. D Correlation between JAM3 methylation (ΔCt) and HPV 16/18 viral load (log-transformed), with R = − 0.22 and p = 0.0086. E Correlation between PAX1 methylation (ΔCt) and non-HPV 16/18 viral load (log-transformed), with R = − 0.003 and p = 0.97. F Correlation between JAM3 methylation (ΔCt) and non-HPV 16/18 viral load (log-transformed), with R = − 0.15 and p = 0.08. ΔCt values reflect the relative methylation levels of PAX1 or JAM3. R represents the correlation coefficient, and p values indicate statistical significance.
Fig. 5.
HPV viral load levels in different cervical lesion/cervical cancer types. Box plots with scatter points showing log-scaled HPV viral load across different grades of cervical lesions, stratified by HPV types [overall HPV VL (A), HPV 16/18 VL (B), and non-HPV 16/18 VL (C)]. Statistical significance is denoted by asterisks: *p < 0.05, **p < 0.01, ***p < 0.001
The relationship between methylation levels, HPV viral load, and high-risk clinical-pathological factors of squamous cell carcinoma
With the increase of FIGO stage of squamous cell carcinoma, the overall PAX1/JAM3 methylation trend gradually increased. However, due to the limited number of cases, only a few groups showed statistical differences (Supplementary Fig. 1A, D). The HPV VL also showed an upward trend with the increase of FIGO stage, especially at the HPV16/18 load level before IB3 stage. However, due to the limited number of cases, the overall difference was not statistically significant (Supplementary Fig. 2A, D). Both PAX1/JAM3 methylation and HPV viral loads lack consistent correlations with LVSI or lymph node metastasis (Supplementary Figs. 1, 2).
Considering the possible influence of the different FIGO staging grade on the comparison of methylation levels and HPV VL levels of CIN 3-like SCC, early invasive SCC, and invasive SCC, we compared CIN3-like SCC with early invasive SCC and invasive SCC at the same FIGO stage respectively and found that, the PAX1/JAM3 methylation level and HPV VL (especially HPV 16/18 VL) of CIN3-like SCC showed an increasing trend comparing with early invasive SCC(especially IA2 stage), but the difference was not statistically significant. Except for PAX1 methylation at the IB1 level, which was a statistically significant difference between CIN3-like SCC and invasive SCC. None of the others were statistically significant (Fig. 6).
Fig. 6.
Comparisons of methylation levels and HPV VL of CIN 3-like SCC in the same FIGO stage with those of early invasive carcinoma and invasive carcinoma. A Comparison of PAX1 methylation between CIN3-like SCC and early invasive SCC in Stage IA; B Comparison of PAX1 methylation between CIN3-like SCC and invasive SCC in Stage IB; C Comparison of JAM3 methylation between CIN3-like SCC and early invasive SCC in Stage IA; D Comparison of JAM3 methylation between CIN3-like SCC and invasive SCC in Stage IB; E Comparison of HPV VL between CIN3-like SCC and early invasive SCC in Stage IA; F Comparison of HPV VL between CIN3-like SCC and invasive SCC in Stage IB; G Comparison of HPV16/18 VL between CIN3-like SCC and early invasive SCC in Stage IA; H Comparison of HPV16/18 VL between CIN3-like SCC and invasive SCC in Stage IB; I Comparison of none HPV16/18 VL between CIN3-like SCC and early invasive SCC in Stage IA; J Comparison of none HPV16/18 VL between CIN3-like SCC and invasive SCC in Stage IB Statistical significance is denoted by asterisks: *p < 0.05, **p < 0.01.
Discussion
CIN3-like SCC is a rare type of SCC with a specific growth pattern described by Al-Nafussi and Monaghan. According to a study by Xing Z et al. [6], its invasive depth ranges from 1.8 to 9.7 mm, so according to FIGO staging, it can be early invasive SCC (stage IA1) or late-stage invasive SCC (above stage IB). In this study, CIN3-like SCC mainly concentrated in IA1-IB3. Cases of CIN3-like SCC with retroperitoneal lymph node and psoas muscle metastases have been reported, showing the same appearance as the primary lesion [15]. In this study, none of the CIN3-like SCC cases showed positive LVSI or lymph node metastasis, suggesting a greater tendency toward an early stage, similar to early invasive SCC. In addition, we unexpectedly discovered from the available colposcopy images that 4 out of 6 patients did not show obvious abnormalities (Fig. 2); Instead, they presented with post-conization status, with the average interval since the previous conization being 6.5 years. Although the cases may be coincidental, it suggests that CIN3-like SCC may occur in the reserve cells of the residual cervical canal after cervical conization, which is a worthy area for further exploration.
Detection of methylated genes in cervical cancer can identify high-grade cervical lesions at an early stage and guide clinical interventions, emerging as a novel approach for future cervical cancer prevention and control [16]. PAX1, belonging to the pPAX gene family, is a methylation-silenced gene identified in cervical cancer and a potential methylation biomarker. As a highly conserved transcription factor in vertebrates and invertebrates, it plays a critical role in multiple biological processes and can serve as an auxiliary biomarker for clinical cervical cancer screening to improve the effectiveness of screening [8, 17]. Liu et al. [18] found that PAX1 gene methylation is common in cervical exfoliated cells of patients with CIN3 and invasive cervical cancer. Xu et al. [19] reported no PAX1 methylation in normal cervical exfoliated cells, with a methylation rate of 9% in CIN1, 44% in CIN2/3, and 100% in invasive cervical cancer. Our team’s previous study also showed that high methylation levels of the PAX1 gene increase with the severity of cervical lesions, especially in CIN3 and cervical cancer, but there was no significant difference between CIN1 and CIN2 [9].
JAM3 are a unique type of transmembrane protein family within the immunoglobulin superfamily, located at the tight junctions of polar epithelial and endothelial cells. According to the literature, JAM3 is involved in processes such as leukocyte migration, angiogenesis, and tumor metastasis [20]. Studies have identified JAM3 methylation as a potential novel biomarker for guiding hierarchical diagnosis and treatment in hrHPV-positive patients. Boers et al. [11, 21] used quantitative methods to detect JAM3 gene methylation levels, demonstrating a sensitivity of 68% and specificity of 94% for diagnosing CIN2 and above in hrHPV-positive women, with sensitivity and specificity of 80% and 76% for diagnosing CIN3 + lesions, respectively. Reports also show that the JAM3 gene exhibits significant sensitivity and specificity in detecting cervical adenocarcinoma, which is particularly valuable for identifying HPV-negative cervical adenocarcinomas [22].
HPV 16/18 VL appears to play a critical role in HPV persistent infection [23]. Our team's previous study [9] found that HPV VL significantly differed between CIN1 and CIN2 but not between CIN3 and cervical cancer. In general, PAX1 methylation positively correlated with all hrHPV VL, especially for HPV16/18 VL (p < 0.001), but showed no association with the VL of the other 12 hrHPV types [9]. However, it has also been found that lower HPV VL and viral genome integration predict the persistence of HPV 16 infection but cannot predict the progression of HPV 16 persistent infection to CIN3 + in women with normal cytology. As viral integration suppresses the replication and release of new virions, it exhibits a negative correlation with circulating HPV VL.
In this study, DNA methylation of PAX1 and JAM3 showed significant changes during the CIN2 → CIN3 and HSIL → cervical cancer transitions. This discovery histologically confirms the previously observed methylation patterns in the progression from cervical precancerous lesions to cancer (at the level of exfoliated cytology [9]). However, no significant differences were observed in methylation levels between CIN3-like SCC, early invasive SCC, and invasive SCC. Similarly, methylation levels showed no significant association with high-risk factors such as LVSI and lymph node metastasis as previously reported [24]. When we analyzed all squamous cell carcinoma according to FIGO staging, although there was no statistically significant difference in the methylation levels of PAX1/JAM3 across different FIGO stages, a trend of increasing levels with advancing stage was still observed even with the limited sample size. Furthermore, when comparing within the same FIGO stages, PAX1/JAM3 methylation in CIN3-like SCC showed an upward trend compared to early invasive SCC(especially in IA2 stage), although the difference was not statistically significant.
On the other hand, the total viral load in CIN3-like SCC was higher than that in CIN3 and early invasive SCC, but showed no significant difference from that in invasive cancer. Previous study also found that E6/E7 mRNA expression in CIN3-like SCC was higher than that in endocervical crypt involvement by CIN3, yet lower than that in invasive cancer [6]. Considering that this situation may be related to the later FIGO staging of CIN3-like SCC compared to early invasive cancer, when comparing with the same FIGO stages, CIN3-like SCC shows a tendency towards higher HPV VL (especially HPV16/18 VL) than early invasive cancer (mainly in Stage IA2), although the difference is not statistically significant. The above findings suggest potential differences in epigenetics and HPV viral load between CIN3-like SCC and early invasive cancer, which need to be confirmed with a larger sample size.
Limitations
This study has several limitations that warrant consideration. Firstly, the sample size was relatively small, which may have limited the statistical power to detect subtle differences in methylation levels or viral load across certain subgroups, particularly for the rare CIN3-like SCC subtype. Larger, multi-center studies with expanded cohorts would enhance the generalizability of the findings and improve the reliability of statistical analyses. Secondly, tissue sampling relied on conventional histological sections rather than microdissection techniques, which might have introduced cellular heterogeneity (e.g., admixture of normal stromal cells with target epithelial lesions). Microdissection could provide more precise isolation of specific cell populations, minimizing confounding factors from non-target tissues and enhancing the accuracy of methylation and viral load measurements. Additionally, the study utilized a retrospective design based on archived tissue samples, which may be subject to selection bias or variability in pathological interpretation across different institutions. Prospective studies incorporating non-invasive exfoliated cell-based testing (e.g., cervical cytology or liquid-based cytology) would offer a more clinically applicable approach, enabling early detection of methylation markers in vivo and longitudinal monitoring of cervical lesion progression. Moreover, More CIN3-like SCC colposcopy photos need to be collected further to determine whether the findings of this study are coincidental. Finally, this study did not perform immunohistochemical staining such as PAX1 and JAM3, as previous studies have reported the loss of expression of both in cervical cancer immunohistochemistry [25, 26], Meanwhile, p16 staining has been detected in previous study but no differential statistical analysis was performed [6]. One research suggests that SOX1 immunohistochemistry may be useful in distinguishing CIN3-like SCC from CIN3 with glandular involvement [27], and all of them are worth exploring in the future.
Conclusion
PAX1 and JAM3 methylation levels are significantly higher in SCC than HSIL, but no difference in methylation levels was observed between CIN 3-like SCC, early invasive SCC, and invasive SCC. Total HPV VL (especially HPV16/18 VL) in CIN3-like SCC is significantly higher than that in CIN3 and early invasive SCC, suggesting potential shared and distinct features in epigenetics and HPV VL between CIN3-like SCC and early invasive SCC, which need more molecular biology research to explore the difference.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
Thanks to Wang Yishan, Guo Yueyi, Wang Linhai, Pei Liu, and Su Yinyue from the Department of Medical Laboratory, Beijing Origin-Poly Bio-Tec Co., Ltd. for methylation detection and data management; Thanks to Song Lili from Bioperfectus Ltd. for performing HPV viral load tests on samples.
Author contributions
Data collection, study conception, and study design were performed by M.L.and X.Z.. Personnel organization and data collection were performed by X.Z..and D.S.. Sample collection was performed by Y.Z., J.L. and C.Z. Histopathological diagnosis were performed by Xiao.Z. Xin.Z.and J.W, respectively. Supervision of the research program and manuscript review and guidance during the study was provided by L.W. and D.S.. The first draft of the manuscript was written by M.L. and all authors commented on previous versions of the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by The National Key Research and Development Program of China (2025YFC2708404) and the Peking University People’s Hospital Horizontal Scientific Research Topics (2025-h1124).
Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Peking University People’s Hospital (IRB number:2025PHB437-001, date of IRB approval: 29 September 2025.
Informed consent
Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patients to publish this paper.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Mingzhu Li and Xiaobo Zhang have contributed equally to this work.
Contributor Information
Danhua Shen, Email: danhuashen@126.com.
Lihui Wei, Email: weilh@bjmu.edu.cn.
References
- 1.Arbyn M, Castellsague X, de Sanjose S, Bruni L, Saraiya M, Bray F, et al. Worldwide burden of cervical cancer in 2008. Ann Oncol. 2008;22:2675–86. [DOI] [PubMed] [Google Scholar]
- 2.World Health Organization. Female genital tumours. IARC press. WHO classification of tumours. 5th Edition Volume 4[EB/ OL]. (2020-09-09)[2021-01-25]. https: //tumourclassification. iarc.who.int/9789283245049.
- 3.Li M, Li J, Li X, et al. Guidelines for cervical cancer screening in China part II. Gynecol Obstet Clin Med. 2025;5:e000205. [Google Scholar]
- 4.Al-Nafussi AI, Monaghan H. Squamous carcinoma of the uterine cervix with CIN 3-like growth pattern: an under-diagnosed lesion. Int J Gynecol Cancer. 2000;10:95–9. [DOI] [PubMed] [Google Scholar]
- 5.Stewart CJR, Crook ML. Podoplanin and SOX2 expression in CIN 3-like squamous cell carcinoma of the cervix. Int J Gynecol Pathol. 2018;37:59–67. [DOI] [PubMed] [Google Scholar]
- 6.Xing Z, Danhua S, Xiaobo Z. Diagnostic validity of p16, E-cadherin, cyclin D1, p53, and HPV E6/E7 mRNA in CIN 3-like squamous cell carcinoma of the cervix. Front Oncol. 2024;14:1354838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Fang C, Wang S, Liou Y, Chen M, Ouyang W, Duan K. The promising role of PAX1 (aliases: HUP48, OFC2) gene methylation in cancer screening. Mol Genet Genom Med. 2019;7:e506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Liang H, Li GL, Liu J, Fu M, Huang H, Zhao K, et al. The application value of PAX1 and ZNF582 gene methylation in high grade intraepithelial lesion and cervical cancer. Clin Transl Oncol. 2021;23:283–8. [DOI] [PubMed] [Google Scholar]
- 9.Li M, Zhao C, Zhao Y, Li J, Zhang X, Zhang W, et al. Association and effectiveness of PAX1 methylation and HPV viral load for the detection of cervical high-grade squamous intraepithelial lesion. Pathogens. 2022. 10.3390/pathogens12010063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kong L, Wang L, Wang Z, Xiao X, You Y, Wu H, et al. DNA methylation for cervical cancer screening: a training set in China. Clin Epigenetics. 2020;12:91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Boers A, Bosgraaf RP, van Leeuwen RW, Schuuring E, Heideman DAM, Massuger LFAG, et al. DNA methylation analysis in self-sampled brush material as a triage test in hrHPV-positive women. Br J Cancer. 2014;111:1095–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li M, Zhao C, Zhang X, Li J, Zhao Y, Zhang W, et al. PAX1/JAM3 methylation and HPV viral load in women with persistent HPV infection. Cancers. 2024. 10.3390/cancers16071430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Food and Drug Administration. New Approaches in the evaluation for high-risk human papillomavirus nucleic acid detection devices[EB/OL]. [2019–03–08].https://www.fda.gov/media/122799/download..
- 14.Saleh M, Virarkar M, Javadi S, Elsherif SB, de Castro Faria S, Bhosale P. Cervical cancer: 2018 revised International Federation of Gynecology and Obstetrics staging system and the role of imaging. AJR Am J Roentgenol. 2018;214:1182–95. [DOI] [PubMed] [Google Scholar]
- 15.Wilkens J, Beattie G, Al-Nafussi A. Metastatic CIN3-like squamous carcinoma. Diagn Histopathol. 2011;17:140–3. [Google Scholar]
- 16.Guzel C, van Sten-Van’T HJ, de Kok IMCM, Govorukhina NI, Boychenko A, Luider TM, et al. Molecular markers for cervical cancer screening. Expert Rev Proteomic. 2021;18:675–91. [DOI] [PubMed] [Google Scholar]
- 17.Li B, Guo R, Lai T, Qiao L, Fu H. The application of PAX1 methylation detection and HPV E6/E7 mRNA detection in cervical cancer screening. J Obstet Gynaecol Res. 2021;47:2720–8. [DOI] [PubMed] [Google Scholar]
- 18.Liu H, Meng X, Wang J. Real time quantitative methylation detection of PAX1 gene in cervical cancer screening. Int J Gynecol Cancer. 2020;30:1488–92. [DOI] [PubMed] [Google Scholar]
- 19.Xu J, Xu L, Yang B, Wang L, Lin X, Tu H. Assessing methylation status of PAX1 in cervical scrapings, as a novel diagnostic and predictive biomarker, was closely related to screen cervical cancer. Int J Clin Exp Patho. 2015;8:1674–81. [PMC free article] [PubMed] [Google Scholar]
- 20.Yin A, Zhang Q, Kong X, Jia L, Yang Z, Meng L, et al. JAM3 methylation status as a biomarker for diagnosis of preneoplastic and neoplastic lesions of the cervix. Oncotarget. 2015;6:44373–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Boers A, Wang R, van Leeuwen RW, Klip HG, de Bock GH, Hollema H, et al. Discovery of new methylation markers to improve screening for cervical intraepithelial neoplasia grade 2/3. Clin Epigenetics. 2016;8:29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kong L, Wang L, Wang Z, Xiao X, You Y, Wu H, et al. Cytological DNA methylation for cervical cancer screening: a validation set. Front Oncol. 2023;13:1181982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Dong L, Wang MZ, Zhao X, Feng R, Hu S, Zhang Q, et al. Human papillomavirus viral load as a useful triage tool for non-16/18 high-risk human papillomavirus positive women: A prospective screening cohort study. Gynecol Oncol. 2018;148:103–10. [DOI] [PubMed] [Google Scholar]
- 24.Li X, Zhou X, Zeng M, Zhou Y, Zhang Y, Liou Y, et al. Methylation of PAX1 gene promoter in the prediction of concurrent chemo-radiotherapy efficacy in cervical cancer. J Cancer. 2021;12:5136–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Su P, Lai H, Huang R, Chen L, Wang Y, Wu T, et al. Paired Box-1 (PAX1) Activates Multiple Phosphatases and Inhibits Kinase Cascades in Cervical Cancer. Sci Rep-UK. 2019;9:9195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Gu Y, Chu C, Yuan B, Wang J, Pan S. Expression and hypermethylation of JAM and EPB41L3 in cervical squamous cell carcinoma: Clinical significance and applications. Histol Histopathol. 2024;39:1043–51. [DOI] [PubMed] [Google Scholar]
- 27.Nagi CS, Schlosshauer PW. Endocervical glandular involvement is associated with high-grade SIL. Gynecol Oncol. 2006;102:240–3. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.