Advisory Committee Chair
Sadeep Shrestha
Advisory Committee Members
Ronald Alvarez
Elizabeth E Brown
Michael G Conner
Xiangqin Cui
Laura K Vaughan
Document Type
Dissertation
Date of Award
2014
Degree Name by School
Doctor of Philosophy (PhD) School of Public Health
Abstract
Coexisting lesions can reduce genomic heterogeneity in precancer progression analysis by reducing variation and bias that can mask gene effects. Our goal was to use novel methodologies to depict the neoplastic stage effect (Normal v. LSIL v. HSIL) in cervical precancer. We analyzed the neoplastic stage effect via in silico and in vivo methodologies. For in silico analyses, we calculated differential expression (DE) estimates from a systematic review of DNA methylation and gene expression literature. Significant genes (FC≤2.0 or p-value≥0.05) were grouped by histology for pathway analysis. For in vivo analyses, we performed RNA-seq on microdissected FFPE coexisting cervical tissue from 6 women, each sample consisting of LSIL, HSIL, and adjacent "normal" cervical epithelium. The Tuxedo package was used to calculate DE for each individual (6) and by tissue group. Pathway annotation and enrichment analyzed top genes (FDR≤0.05) in at least 2/6 participants. Finally, HPV16 transcripts from RNA-seq coexisting cervical lesion data was assessed. We extracted 151 genes from 13/264 gene expression and 7/66 DNA methylation papers. There were 80/151 genes differentially expressed (FC≥2.0 or p-value≤0.05). Characteristic cellular dysregulation was identified from pathway analysis of DE genes including the cell cycle:G2/mDNA damage checkpoint pathway (p-value:1.5E-04). RNA sequencing generated 1.2 billion reads, and after alignment, assembly, and differential analysis there were 63 genes in 2/6 participants, of which 53/63 pairs (Normal v. LSIL) were upregulated. HPV16 transcripts were detected in HSIL samples of the three individuals who shared similar expression profiles and 6 DE genes (FDR≤0.05: GALANT, S100A7, SPRR2E, KLK6, and MUC1) and significant over expression in epithelial cell and keratinocyte differentiation pathways. We used novel methods to generate DE in gene signatures that grouped into hallmarks of cervical precancer progression: inflammation, cellular differentiation, and cell cycle dysregulation.
Recommended Citation
Royse, Kathryn Elizabeth, "Molecular Profiling in Cervical Carcinogenesis and Progression" (2014). All ETDs from UAB. 2868.
https://digitalcommons.library.uab.edu/etd-collection/2868