Gene Expression Profiling of Women with Varying Degrees of Cervical Intraepithelial Neoplasia

Abstract

A pilot study to determine genomic microarray differences in patients with normal cervical tissue and cervical intraepithelial neoplasia (CIN) 3. After institutional review board approval, patients referred to the University of Alabama at Birmingham Colposcopy Clinic for a loop electrosurgical excisional procedure were identified. Tissue biopsies of both normal tissue and CIN 3 from fresh loop electrosurgical excisional procedure specimens were obtained and sent to pathology for histological confirmation. Procurement of these 2 types of tissue from the same patient controls for different types of human papillomavirus infection, smoking and nutritional status, age, immunocompetency, and other microenvironment factors (i.e., the patient serves as their own control). Standard RNA extraction techniques (Qiagen, Inc., Valencia, CA) were used to prepare the tissue specimens for microarray analysis with the Affymetrix GeneChip U133A expression array (Affymetrix, Inc., Santa Clara, CA). Paired samples were eligible for microarray analysis only when both normal tissue and CIN 3 were confirmed by pathology. The data were then subsequently subjected to a log-like transformation and analyzed with a t test. For this study, p<.001 was determined to be statistically significant. Twenty-one pairs of both normal and CIN tissues were obtained and underwent histological evaluation followed by RNA extraction. Within the normal group, 86% were confirmed as true normals and 14% as CIN. Within the CIN group, 28.5% were found to have CIN 2, 62% were found to have CIN 3, and 9.5% had no evidence of CIN. Mean RNA content of normal samples was 2.0 microg, whereas the mean RNA content of CIN samples was significantly higher at 7.4 microg (p=006). Ultimately, 5 pairs of normal and neoplastic tissues were subjected to microarray analysis. Using a cutoff of p<.001, 24 candidate genes were identified from more than 18,000 genes. In the CIN 3 group, 14 genes were overexpressed and 10 genes were underexpressed. Of the 14 overexpressed genes, 9 were noted to have identities listed in the National Center for Biotechnology Information public domain. Five (56%) of these 9 genes were directly related to immunity-related pathways, and 3 (33%) of the 9 genes were found to be involved in cell cycle function/control. One overexpressed gene was identified as p53. The presence of CIN is marked by increased transcriptional activity, evident by an almost 4-fold increase in mean RNA content obtained from our CIN samples versus normal cervical tissue. Furthermore, a number of statistically significant overexpressed genes of interest related to immune function/response and cell cycle control were identified in our pilot microarray study. This data have the ability to direct future research endeavors in cervical neoplasia. Future endeavors include the use of laser capture microdissection to evaluate genomic changes strictly at the epithelial level and, as such, exclude stromal response contributions.