Identification of immunohistochemical biomarkers for papillary thyroid carcinoma using gene expression profiling

We report the successful validation of a combined gene expression profiling and tissue microarray approach to papillary thyroid carcinoma (PTC) biomarker identification. Our ultimate goal is the identification of protein biomarkers that can be effectively used in immunocytochemical assays applied to thyroid fine needle aspiration biopsy (FNAB) samples. To that end, we designed our approach to prioritize molecules that were minimally expressed in normal thyroid and highly expressed in PTC. We first generated gene expression profiles from 11 normal thyroid tissue samples and 9 samples of classic PTC. The results were segregated to rank most highly those molecules not expressed in normal thyroid and up-regulated at least 6-fold in PTC. From this list, we chose 2 molecules (P-cadherin and Bax) for immunohistochemical analysis for which commercial antibodies were available. These were compared with 2 other molecules that have been previously studied in thyroid cancer (cytokeratin-19 and galectin-3). For immunohistochemistry, a tissue microarray was generated that contained the following tissues: classic PTC (n = 20), follicular variant of PTC (n = 9), normal thyroid (n = 19), Hashimoto thyroiditis (n = 11), follicular adenoma (n = 15), and follicular carcinoma (n = 14). Immunohistochemical staining was scored and compared with the gene expression profiling. As anticipated, cytokeratin-19 and galectin-3 were highly expressed in PTC and less expressed in other tissues. Bax and P-cadherin were also expressed in PTC, but to a lower level than cytokeratin-19 and galectin-3; however, Bax and P-cadherin demonstrated virtually no staining of normal thyroid, unlike cytokeratin-19 and galectin-3. These results validate our approach for PTC biomarker discovery and identify several candidate biomarkers for further development.