Aberrant expression of long noncoding RNAs in cervical intraepithelial neoplasia

Long noncoding RNAs (lncRNAs) are a unique class of messenger RNA-like transcripts of at least 200 nucleotides in length with no significant protein-coding capacity. Aberrant lncRNA expression is emerging as a major component of the cancer transcriptome. Here, we sought to determine if differential lncRNA expression is a feature of the human cervical intraepithelial neoplasia (CIN) transcriptome.

Sequence data were derived from 16 long serial analyses of gene expression (L-SAGE) libraries constructed from cervical specimens representing mild (CIN1), moderate (CIN2), and severe (CIN3) histopathologic grades of CIN. A novel lncRNA discovery pipeline was developed to query the expression of lncRNAs within the SAGE data sets.

A total of 2,230,370 sequence tags were delineated from the 16 SAGE libraries, representing the expression of 367,482 unique tags at varying abundance. Using a novel stepwise filtering strategy, we analyzed the cervical SAGE libraries and identified the expression profiles of 1056 lncRNAs in the human cervix. We present the first lncRNA expression profile derived from nonneoplastic cervical tissue and establish that changes in lncRNA expression do occur in cervical intraepithelial lesions. Our analysis also shows statistically significant aberrant expression of lncRNAs in the 3 CIN grades, suggesting that these unique noncoding RNA transcripts may contribute to the development and progression of precursor lesions.

Through the analysis of L-SAGE libraries constructed from cervical specimens, we provide the first lncRNA expression profile of the cervix and demonstrate aberrant expression in early-stage neoplasia.