MicroRNA Biomarkers of High-Grade Cervical Intraepithelial Neoplasia in Liquid Biopsy

Abstract

New prevention strategies are needed to detect cervical intraepithelial neoplasia (CIN). The microRNA expression analysis has already been reported as molecular biomarkers in the early detection of cervical cancer (CC) through minimally invasive samples, such as liquid biopsy, obtained through collection using liquid-based cytology (LBC). In this study, we aimed to identify molecular signatures of microRNAs in cervical precursor lesions from LBC cervical and the molecular pathways potentially associated with the CC progression. We analyzed 31 LBC cervical samples from women who underwent colposcopy. These samples were divided into two groups: the first group was composed of samples without precursor lesions of CC, considering the control group, referred to as healthy female subjects (HFS; $n=11$ ). The second group corresponded to women diagnosed with cervical interepithelial neoplasia grade 3 (CIN 3; $n=20$ ). We performed microRNA and gene expression profiling using the nCounter® miRNA Expression Assays (NanoString Technology) and PanCancer Pathways (NanoString Technology), respectively. A microRNA target prediction was performed by mirDIP, and molecular pathway interaction was constructed using Cytoscape. Bidirectional in silico analyses and Pearson’s correlation were performed for associated the relation between genes, and miRNAs differentially expressed related cervical cancer progression were performed. We found that the expression of nine microRNAs was significantly higher, two were downregulated (miR-381-3p and miR-4531), and seven miRNAs were upregulated (miR-205-5p, miR-130a-3p, miR-3136-3p, miR-128-2-5p, let-7f-5p, miR-202-3p, and miR-323a-5p) in CIN 3 ( $\text{fold}\text{change}\ge 2$ and $p\le 0.05$ ). The miRNA expression patterns were independent of hr-HPV infection. We identified four miRNAs (miR-205-5p, miR-130a-3p, miR-4531, and miR-381-3p) that could be used as biomarkers for CIN 3 in LBC samples through multiple logistic regression analyses. We found 16 genes differentially expressed between CIN 3 and HSF samples ( $\text{fold}\text{change}\ge 2$ and $p\le 0.05$ ). We found the correlation between miR-130a-3p and CCND1( $R=-0.52$ ; $p=0.0029$ ), miR-205-5p and EGFR ( $R=0.53$ ; $p=0.0021$ ), and miR-4531 and SMAD2 ( $R=-0.54$ ; $p=0.0016$ ). In addition, we demonstrated the most significant pathways of the targets associated with cervical cancer progression (FDR-corrected $p<0.001$ ). This study demonstrated that miRNA biomarkers may distinguish healthy cervix and CIN 3 and regulate important molecular pathways of carcinogenesis.