OC 11-02GENOME-WIDE METHYLOME ANALYSIS UNCOVERS NEW HYPERMETHYLATION BIOMARKERS FOR BOTH ADENO- AND SQUAMOUS CELL CERVICAL CARCINOMA.

13. Molecular markers
R. Van Leeuwen 1, R. Wang 2, A. Boers 1, T. De Meyer 3, R. Steenbergen 4, W. Van Criekinge 3, A. Van Der Zee 1, E. Schuuring 5, B. Wisman 1.
1Department of Gynaecologic Oncology, Cancer Research Centre Groningen, University of Groningen, University Medical Centre Groningen, Groningen (Netherlands), 2Department of Laboratory Medicine, Tianjin Medical University, Tianjin (China), 3Department of Mathematical Modelling, Statistics and Bio-informatics, University of Ghent, Ghent (Belgium), 4Department of Pathology, VU University Medical Centre, Amsterdam (Netherlands), 5Department of Pathology, Cancer Research Centre Groningen, University of Groningen, University Medical Centre Groningen, Groningen (Netherlands)

Background / Objectives

The sensitivity of cytology-based screening methods for adenocarcinoma of the cervix (ADC) and its precursors is low compared to the already suboptimal sensitivity for squamous-cell cervical neoplasia for a variety of reasons. Analysis of DNA methylation markers has been reported promising to detect squamous-cell carcinoma (SCC) in earlier stages, but few methylation markers have been described for the early detection of ADC. The aim of this study was to identify methylation markers for the early detection of both ADC and SCC.


Methods

Global methylation profiles were generated for frozen tissues of 20 normal cervices, 6 ADC and 6 SCC by capturing methylated DNA and performing next-generation sequencing (MethylCap-seq). Differentially methylated regions were appraised by means of a step-wise validation approach using bisulfite pyrosequencing and methylation-specific PCR (MSP) on various cohorts of patient material (i.e. formalin-fixed paraffin-embedded (FFPE) tissues of 17 normal cervices and 6 ADC and 7 SCC; 225 cervical scrapings from women either with or without cervical cancer; and 229 scrapings of women who were referred for colposcopy).


Results

MethylCap-seq revealed 53 regions that were methylated in both ADC and SCC and not in normal cervices. Of the 15 most significant regions, 5 markers exhibited significantly different methylation rates between normal and cancer FFPE tissues. Quantitative MSP (QMSP) on cervical scrapings from an independent cohort of 89 women with a normal cervix and 68 patients with SCC and 57 with ADC, revealed a sensitivity of 79% to 88% to detect cervical cancer, while 94% to 99% of normal scrapings tested negative. The QMSPs for 4 markers (SOX1, SOX14, SLC6A5 and TBX20) detected ADC and SCC with a similar sensitivity. Finally, the analysis of scrapings from a cohort of women referred with an abnormal smear with known histological diagnosis (n=229 including 14 adenocarcinoma in situ (AdCIS), 12 ADC and 29 SCC), revealed that the individual QMSPs were able to detect cervical intraepithelial neoplasia grade 2 or worse (CIN2+) with a sensitivity ranging from 28% to 59%, of which the AdCIS showed a sensitivity of 71-93%, and with a high specificity (88-98%). Compared to hrHPV testing, the combination of SOX1 or SOX14 methylation showed a similar sensitivity for CIN3+ (73% vs. 80% for hrHPV, p > .2), whereas the specificity was significantly better (88% vs. 42% for hrHPV, p < 10^-5).


Conclusion

We identified 4 new methylation markers with a high sensitivity for both ADC as well as SCC in cervical scrapings. Additionally, the combined methylation of SOX1 or SOX14 showed a better specificity compared to HPV analysis.


References