Chiaw Ling Chng ^1* Oi Fah Lai ¹ Lay-Leng Seah ² Kai-Ling Yong ² Yvonne Hsi-Wei Chung ² Rochelle Goh ¹ CHE KANG LIM ¹

• ¹ Singapore General Hospital, Singapore, Singapore
• ² Singapore National Eye Center, Singapore, Singapore

Background There are no reliable biomarkers to identify Graves’ disease patients that will develop severe Graves’ Orbitopathy (GO). We hypothesize that integrating various omics platforms can enhance our understanding of disease mechanisms and uncover potential biomarkers. The aims of this study were (1) Elucidate differential gene expression profile of orbital fibroblasts in GO during early adipogenesis to better understand disease mechanisms (2) Compare tear protein profiles from our earlier study and the transcriptome profiles of orbital fibroblasts is to identify possible biomarkers of the disease. Methods Orbital fibroblasts (OF) were grown from orbital adipose tissue obtained from 9 GO patients (3 for discovery and 6 for validation experiments). Total RNA was extracted from OF on day 0 as baseline for each sample and from differentiated OF on days 4 and 8. Protein - protein interactome (PPI) analysis and functional enrichment analysis were also carried out. The differentially expressed genes (DEGs) from the RNA sequencing experiments were then compared to the full tear proteome profile from the author’s previous study which examined the tear protein changes in tears of GO patients based on fold change > 1.6 or < -1.6, FDR < 0.05 applied within all dataset. Further validation of S100A4 downregulation in GO was performed via quantitative real-time PCR (qPCR). Results The whole transcriptomic analysis revealed 9 upregulated genes and 15 downregulated genes in common between discovery and validation experiments. From the PPI network analysis, an interaction network containing six identified DEG (ALDH2, MAP2K6, MT2A, SOCS3, S100A4 and THBD) were observed. Functional enrichment network analysis identified a set of genes related to oxysterol production. S100 calcium binding protein A4 (S100A4) was found to be consistently downregulated in both our transcriptome studies and the full tear proteome profile from the author’s previous study. Conclusion Our study has identified several DEGs and potential gene pathways in GO patients which concurred with the results of other studies. Tear S100A4 may serve as a biomarker for the propensity to develop TED in patients with AITD before clinical manifestation and should be confirmed in future studies.