Identifying Prospective Markers for Thyroid Status Using a Differential Gene Expression Analysis Workflow

Peter Chang

Mentor: Drs. Karen Ross and Mark Danielsen, Department of Biochemistry and Molecular & Cellular Biology, Georgetown University Medical Center.

Date/Time: August 25, 2020 at 2pm

Abstract: The standard of care for treating hypothyroidism is to give synthetic T4 until thyroid stimulating hormone (TSH) levels are within the normal range. In pituitary disease where TSH production is absent, this marker of thyroid status cannot be used. For this reason, we performed a differential gene expression analysis to look for new markers for thyroid status. A workflow was created to analyze microarray data taken from monocyte cells of patients with thyroid cancer and autoimmune diseases affecting the thyroid. Samples were collected before and after treatment was applied. CD8 cell samples from the autoimmune disease patients were also analyzed. Any potential markers that come through in blood cells would allow for a noninvasive marker to monitor patients. In our workflow, normalization was performed using the MAS5 method. Probes that were present were mapped to genes. Probes with low intensity, or that did not map to known genes were filtered out to increase significance and power when running statistical analyses. Comparisons between pre-treatment and post-treatment samples, CD8 cells and monocyte cells, and autoimmune disease and thyroid cancer groups were performed to find differential expression between groups. Preliminary PCA analysis showed a clear difference between CD8 cells and monocyte cells, but little difference between pre-treatment and post-treatment groups, or thyroid cancer vs autoimmune disease groups. We did not find any genes with significant adjusted p-values in any comparisons. Several genes of interest were found including THRB, THRSP, TRIP11, TRIP12 and TRIP13. None of them had significant p-values, and had log fold changes between -1 and 1.  Gene enrichment analysis was performed using DAVID on the genes with the lowest p-values to determine if any functions or pathways may be enriched. None of the gene sets were found to be enriched. However, clustering of samples based on genes with high fold change and low p-values did show distinct clustering, clearly separating the pre-treatment and post-treatment thyroid cancer cells, as well as a lesser distinction between pre- and post-treatment autoimmune samples. While this preliminary analysis of microarray data did not yield significant results, more work can be done with RNAseq data to investigate gene expression in these samples.