Aims Gene promoter hypermethylation is recognised as an essential early step in carcinogenesis, indicating important application areas for DNA methylation analysis in early cancer detection. The current study was set out to assess the performance of CADM1, MAL and miR124-2 methylation analysis in cervical scrapes for detection of cervical and endometrial cancer.
Methods A series of cervical scrapes of women with cervical (n=79) or endometrial (n=21) cancer, cervical intraepithelial neoplasia grade 3 (CIN3) (n=16) or CIN2 (n=32), and women without evidence of CIN2 or worse (n=120) were assessed for methylation of CADM1, MAL and miR124-2. Methylation analysis was done by the PreCursor-M assay, a multiplex quantitative methylation-specific PCR.
Results All samples of women with cervical cancer (79/79, 100%), independent of the histotype, and 76% (16/21; 95% CI 58.0% to 94.4%) of women with endometrial cancer scored positive for DNA methylation for at least one of the three genes. In women without cancer, methylation frequencies increased significantly with severity of disease from 19.2% (23/120; 95% CI 12.1% to 26.2%) in women without CIN2 or worse to 37.5% (12/32; 95% CI 20.7% to 54.3%) and 68.8% (11/16; 95% CI 46.0% to 91.5%) in women with CIN2 and CIN3, respectively. Overall methylation positivity and the number of methylated genes increased proportionally to the lesion severity.
Conclusions DNA methylation analysis of CADM1, MAL and miR124-2 in cervical scrapes consistently detects cervical cancer and the majority of CIN3 lesions, and has the capacity to broaden its use on cervical scrapes through the detection of a substantial subset of endometrial carcinomas.
- CERVICAL CANCER
- diagnostic screening
Statistics from Altmetric.com
Cervical cancer is the third most common cancer in women worldwide, accounting for 9% (529 800) of all new cancer cases and 8% (275 100) of all cancer-related deaths in 2008.1 The introduction of cytology-based screening programmes, either organised or opportunistic, has markedly reduced the incidence and mortality rates of invasive cervical cancer in developed countries. Additionally, different studies have described detection of other gynaecological pathologies such as endometrial carcinoma by the cytology-based cervical screening programme.2–6 However, the effects of cytology-based screening have levelled off. This is mainly due to the suboptimal sensitivity of the screening tool (i.e., cytology) for cervical intraepithelial neoplasia (CIN) grades 2 or 3 (CIN2/3) and cancer (CIN2/3+), as well a substantial number of women not attending cervical screening. Therefore, efforts to improve screening have focused on alternative screening tools to overcome the limitations of cytology, and attracting more women into the screening programme.
Over the last years, the importance of primary screening by detection of human papillomavirus (HPV) DNA in cervical samples in the prevention of cervical cancer has become clear. A persistent infection with a high-risk type of HPV (hrHPV) is necessary for the development of cervical cancer and its high-grade precursors.7 HPV DNA testing appeared substantially more sensitive in detecting CIN2/3+ lesions than cytology, providing 50% greater protection against cervical carcinoma and its high-grade precursor lesions.8 ,9 In addition, HPV DNA testing can also be applied to self-collected (cervico-) vaginal specimens, the latter being important to increase the compliance rate in cervical screening.10–12 Yet, HPV DNA testing shows a reduced specificity compared with cytology, since many infected women harbour transient viral infections that are not associated with clinically meaningful disease. Therefore, it is important to triage hrHPV-positive women for colposcopy to identify those women with the highest risk for CIN2/3+ thereby reducing over-referral and overtreatment.7 ,13 ,14 DNA methylation analysis of cancer-related genes by quantitative methylation-specific PCR (qMSP) has emerged as a promising and objective triage tool for early detection of cervical neoplasia.15–17 Hypermethylation of CpG islands in the promoter regions of tumour suppressor genes leads to gene silencing and is recognised as an essential step in cancer development.18–22 Methylation of cancer-related genes has been described in a variety of gynaecological carcinomas, including cervical cancer15 ,18 ,20 ,23–27 and endometrial cancer.28–31 We previously showed that methylation-mediated silencing of CADM1 (cell adhesion molecule 1), MAL (T-lymphocyte maturation-associated protein) and miR124-2 (micro-RNA124-2) is functionally involved in cervical carcinogenesis,19 ,21 ,22 and a frequent event detectable in tissue biopsies from CIN3 lesions and cervical carcinoma.20 ,32 Methylation analysis of the promoter regions of these genes by qMSP is valuable in colposcopy triage of hrHPV-positive women, both when using cervical scrapes33 and self-collected specimens.34 ,35 The bi-marker panel CADM1/MAL has been validated on hrHPV-positive cervical scrapes collected in a population-based screening setting, and showed to be equally discriminatory for CIN3+ as cytology at the same specificity level.33 In a recent prospective randomised trial among non-responders of the regular cervical screening programme, the bi-marker panel MAL/miR124-2 on hrHPV-positive self-samples was non-inferior to cytology triage via a physician-taken cervical scrape in detecting CIN2+.35 Of note, methylation levels of these three genes have shown to be related to the severity and duration of cervical disease and are exceptionally increased in cervical cancer.20 ,36 As a consequence, methylation analysis could be particularly effective in detecting advanced precursor lesions (with likely high short-term progression risk) and cervical cancers,15 and could serve as complementary tool to cytology triage of hrHPV-positive women to gain a higher reassurance of not missing advanced lesions and cervical cancer.37 These data suggest that cancers are unlikely to remain undetected by DNA methylation analysis of cervical scrapes. Yet, no large series of cervical scrapes from women with cervical cancer have been evaluated so far, as these are not merely encountered in large numbers in a screening setting given the rarity of cervical cancer in the screening population.
In the current study, we evaluated CADM1, MAL and miR124-2 methylation in a large series of cervical scrapes from women with various underlying disease grades, including 79 women with cervical cancer and 21 with endometrial cancer. We assessed the performance of a multiplex qMSP kit (PreCursor-M, Self-screen B.V., The Netherlands) that allows the combined detection of CADM1, MAL and miR124-2 methylation in cervical scrapes.
For this study, cervical scrapes of 268 women who participated in population-based cervical screening or attended a gynaecological outpatient clinic were used. Samples were selected to include a higher than normal proportion of HPV-positives and underlying carcinoma. Details on the number of cervical scrapes in relation to underlying disease category and age of the women are listed in table 1. Cervical scrapes were retrieved from the pathology archive and in case no slides were available, cytological preparations were made from diagnostic left over specimen. In this way, we were able to obtain 235/268 (88%) cytology results. Cytology was reported according to the standard classification in the Netherlands, i.e., CISOE-A classification, that can be easily translated into Bethesda classification38 in which borderline or mild dyskaryosis (BMD) equals ASC-US/ASC-H/LSIL. Presence of hrHPV-DNA was determined by GP5+/6+ PCR.39 For this study, ethical approval was waived since diagnostic leftover specimens were used that had been anonymised according to the Dutch regulations.40
qMSP methylation analysis
DNA from cervical scrapes was isolated using the Nucleo-Spin 96 Tissue kit (Macherey-Nagel) and a Microlab Star robotic system (Hamilton, Germany) according to manufacturers’ protocol or by standard salt-chloroform extraction and isopropanol precipitation.23 ,33 Extracted DNA was subjected to bisulfite treatment using the EZ DNA Methylation Kit (Zymo Research, USA) as described previously.21 ,22 DNA methylation analysis of CADM1, MAL and miR124-2 was performed by a commercial multiplex qMSP (PreCursor-M). PreCursor-M uses primers and probes specific for methylated DNA of CADM1, MAL and miR-124-2, and methylation-independent β-actin as sample quality control. The multiplex format enables simultaneous amplification and detection of the four targets within one reaction.41 Analyses were performed on an ABI 7500 real-time PCR-system (Applied Biosystems, USA). All samples had a quantification cycle (Cq) value for β-actin <29 to assure sample quality. The target methylation result relative to that of a calibrator was calculated as ΔΔCq ratio for each marker gene (i.e., CADM1, MAL or miR124-2), being a measure for hypermethylation using the 2−ΔΔCq method.42 Cervical scrapes were scored positive based on preset ΔΔCq ratio thresholds according to manufacturers’ instructions (i.e., validated thresholds that on a validation set of cervical scrapes of hrHPV-positive women gave rise to a maximum CIN3+ sensitivity at 70% specificity).
Data and statistical analysis
For calculations, a sample was considered methylation-positive for a specific target if the ΔΔCq ratio was above the preset threshold of the respective target and overall methylation-positive when at least one target was above its preset threshold. The threshold for cytology positivity was BMD. The 95% Wald CIs (95% CIs) were determined for the proportions of positive samples. The proportions of overall methylation-positive samples per disease category were compared using χ2 analysis. Calculations were performed in Microsoft Excel (2010) and SPSS (V.20).
Methylation analysis of cervical scrapes
A series of 268 cervical scrapes was evaluated (table 2). DNA methylation of at least one of the three loci (CADM1, MAL and/or miR124-2) was detected in all cervical scrapes of women with cervical cancer, independent of the histotype. This percentage was 76.2% for women with endometrial cancer, 68.8% for CIN3 and 37.5% for CIN2. By comparison, 19.2% of women without evidence of CIN2+ had a methylation-positive cervical scrape, comprising 24.7% (19/77; 95% CI 15.1% to 34.3%) of the hrHPV-positive women and 9.3% (4/43; 95% CI 6.2% to 18.0%) of the hrHPV-negative women. The frequency of methylation positivity increased significantly with the severity of the cervical lesion (figure 1).
Relative contribution of the three targets
The frequency of methylation positivity and the number of markers with a methylation-positive score increased with the severity of the underlying cervical lesion. Among methylation-positive cervical scrapes, those of women without evidence of CIN2+ and women with CIN2 lesions were mostly single marker positive (78.3% (18/23) and 58.3% (7/12), respectively), whereas the majority of women with CIN3 were double marker positive (54.5%; 6/11) and those with cervical carcinoma triple marker positive (54.4%; 43/79). Regarding the two major histotypes of cervical carcinoma, the majority of scrapes of women with squamous cell carcinoma (SCC) were triple marker positive (61.3%; 38/62), whereas adenocarcinoma was predominantly double marker positive (66.7%; 8/12). All double marker positive women with CIN3 and cervical carcinoma were positive for MAL and miR124-2. Among methylation-positive scrapes of women with endometrial adenocarcinoma, both double (43.8%; 7/16) comprising various marker combinations) and single marker (56.3%; 9/16) positivity were seen with overall the largest contribution by miR124-2 (87.5%; 14/16).
In this study, we explored the clinical performance of multiplex methylation analysis of CADM1, MAL and miR124-2 promoter regions in cervical scrapes, and demonstrated a detection rate of 100% for cervical cancer, and 76% (95% CI 58.0% to 94.4%) for endometrial cancer. Our data indicate that methylation analysis of CADM1, MAL and miR124-2 in cervical screening would identify all cervical carcinomas and holds the detection of endometrial carcinomas as important supplement. The standardised assay detects three methylation markers in one single reaction, which requires less sample material over separate reactions as has been mainly used in previous studies.32 ,34 ,35 As demonstrated in the present study, the three methylation markers have additive value with respect to disease detection, in line with previous findings.32–35 The assay furthermore allows high-throughput analysis, which is advantageous in screening programmes.41
To the best of our knowledge, this is the first study evaluating a large series of cervical scrapes from women with cervical or endometrial cancer for DNA methylation of CADM1, MAL and miR124-2. The figures corroborate with previous data describing a methylation positivity rate of 100% for cervical cancer in smaller sample sets using parallel singleplex assays.32 ,33 ,36 In line with previous findings, the number of methylation markers scoring positive, which is inherent to higher methylation levels, increased with disease severity.36 This strongly supports the concept that CIN lesions positive with this assay represent more advanced CIN lesions in need of treatment.15 Based on our data, MAL and miR124-2 are the predominant markers for the detection of high-grade cervical and endometrial lesions, at the thresholds defined for this assay, with CADM1 adding to reach 100% detection rate of cervical cancer. Also HPV-DNA positivity rates in scrapes of women with cervical cancer were high in our study. Positivity rates for HPV-DNA in cervical adenocarcinoma tended to be lower than SCC, as reported in literature.43 The lower HPV detection rates in scrapes of cervical adenocarcinoma versus SCC may be due to technical aspects or misdiagnosis of endometrial carcinoma as cervical adenocarcinoma. Although comparison of the methylation test to cytology is skewed as cancer patients were mainly referred because of abnormal cytology, it was noted that scrapes with normal cytology (n=8) of women with cervical cancer were detected by the methylation assay. The detection of a substantial subset of endometrial carcinomas is an important supplement of methylation analysis of cervical scrapes. The addition of other methylation markers31 or other molecular markers such as tumour-specific mutations44 might improve the overall diagnostic accuracy for gynaecological malignancies in the future.
The promise of methylation analysis for early detection of cervical neoplasia is widely acknowledged.15–18 23–27 33–35 ,45 ,46 At present, methylation analysis is proposed as a triage test for HPV-positive women, with clinical utility in detection of CIN2+/3+ shown for both hrHPV-positive cervical scrapes33 ,37 and self-collected specimens.34 ,35 Current findings also suggest the prospect of the methylation assay as a primary screening tool. Methylation markers with a high sensitivity for cancer may particularly be attractive in middle- and low-income countries, where quality-controlled cytology is absent and implementation of follow-up algorithms for HPV-positive women is complicated.
In conclusion, multiplex methylation analysis of the CADM1, MAL and miR124-2 loci in cervical scrapes consistently identifies cervical cancer and the majority of CIN3 lesions, and has the capacity to broaden its use on cervical scrapes through the detection of a substantial subset of endometrial carcinoma. As such, it is a promising step toward a broadly applicable screening methodology and lays the foundation for a new generation of molecular screening.
Take home messages
Methylation analysis of CADM1, MAL and miR124-2 in cervical scrapes consistently detects cervical cancer and the majority of cervical intraepithelial neoplasia grade 3 lesions.
Overall methylation positivity and the number of methylated genes in cervical scrapes increase proportionally with underlying disease severity.
DNA methylation analysis of CADM1, MAL and miR124-2 in cervical scrapes has the capacity to broaden its use through the detection of a substantial subset of endometrial carcinomas.
We thank HME de Bruin and A van Splunter for expert technical assistance, Dr J Berkhof for statistical advice and Dr L Rozendaal for supportive work.
Abstract in Dutch
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Abstract in Dutch - Online abstract
LMADS and MvZ contributed equally.
Contributors DAMH, RDMS, PJFS and CJLMM designed the study. LMADS, MvZ and DAMH drafted the original manuscript. LMADS, MvZ and ATH were responsible for data analyses. MCGB and GBAW provided clinical samples. All authors reviewed the manuscript and approved the final version.
Funding This work was supported by the Dutch Cancer Society (VU2009-4522) and Eureka/European Community (EUROSTARS E6679).
Competing interests DAMH, RDMS, PJFS and CJLMM are shareholders of Self-screen B.V.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.