Molecular characterization of well-differentiated human thyroid carcinomas by cDNA arrays

Abstract

Well-differentiated papillary and follicular thyroid carcinomas are the two most common types of thyroid cancer. Although cancerous cells in both types phenotypically resemble the epithelial follicular cell, the tumors display different histological characteristics and clinical outcomes. Molecular defects contributing to the separate development pathways remain largely unclear. We evaluated gene expression profiles to generate a detailed molecular characterization of the two tumor types, attempting to detect novel diagnostic and clinical markers. Gene expression profiling of 46 thyroid samples (16 papillary carcinomas, 13 follicular carcinomas and 17 normal thyroid specimens) was performed by using high-density human UniGene cDNA arrays. The identification of differentially expressed genes was based on a comparison of signal intensity ratios of tumor versus normal tissues. A cross-validation of individual filter-array hybridizations and real-time PCR analysis of selected genes were carried out to confirm data reproducibility and reliability. The majority of genes with altered expression were found in both papillary and follicular carcinomas, reflecting a close relationship between the two tumor types. However, 123 genes consisting of 45 known and 78 unknown genes were shown to be differentially expressed between papillary and follicular carcinomas. Follicular variants of papillary carcinoma, clustered together with classical papillary carcinoma, could be differentiated from follicular carcinoma. Our study revealed a set of genes differentiating follicular carcinoma from classical papillary carcinoma and follicular variant. The data generated in this study could serve as a useful source for further investigation of pathways of papillary and follicular differentiation of thyroid cancer.