The digital expression profile of BMP7, CDKN2C, HIST1H3G, and PKMYT1 genes improves high-grade cervical lesion detection in liquid-based cytology.

Abstract

Some studies reported that differential gene expression could be used as a biomarker for high-grade cervical lesion identification. The aim was to evaluate the gene expression profile of cervical intraepithelial neoplasia (CIN) to identify a gene expression signature of CIN2+ in liquid-based cytology (LBC) samples. LBC samples (n=85) obtained from women who underwent colposcopy were included with benign (n=13), CIN1 (n=26), CIN2 (n=16), and CIN3 (n=30) diagnoses. After RNA isolation, gene expression profiling was performed using the nCounter PanCancer Pathways, which consists of 730 cancer-related genes. The genes identified were in silico expression evaluated using the UALCAN database. An accurate prediction model to discriminate CIN2+ from <CIN2 lesions was determined. Immunohistochemistry was performed to assess the expression of p16 and Ki67 proteins. This study identified a gene expression profile that significantly differentiates CIN2+ cases from <CIN2. The gene signature comprised 18 genes, two genes downregulated and 16 upregulated. In silico analysis corroborated the differential expression of 11 of those genes. Further observed was that high expression of BMP7 (odds ratio [OR],4.202), CDKN2C (OR,5.326), HIST1H3G (OR,3.522), PKMYT1 (OR,4.247), and menarche age (OR,1.608) were age-adjusted and associated with CIN2+. This model demonstrates a probabilityof43% leading to an area under the curve (of.979; sensitivity of 94.9%, and specificityof91.2% for CIN2+ prediction. It was observed that p16 expression was significantly associated with CDKN2A mRNA overexpression (p=.0015). A gene expression profile that may be helpful in the identification of patients with CIN2+ was identified. This approach could be used together with currently used LBC in a clinical setting, allowing the identification of patients with high risk of CIN2+.