Abstract Studies show that disrupted circadian clocks are associated with breast and other forms of cancer. However, the pathways and cell processes linking broken circadian clocks to breast cancer are still mostly unknown. In this study, we conducted a reanalysis of mRNA expression in breast cell lines, specifically MCF10a (a fibrocystic cell line) and MCF7 (a metastatic adenocarcinoma cell line). To achieve this, we utilized microarray data obtained from Gutierrez et al. in 2016, specifically the datasets GSE76370, GSE76368, and GSE76369. First, Metacycle, an R statistical package employing a cosine fitting algorithm, was used to identify circadian-oscillations in mRNA expression from both cell lines. MetaCycle (Wu et al., 2016) analysis produced an amplitude, phase, and period value for each circadian gene. Following the analysis, we conducted principal component analysis on the Metacycle output to identify any outlier genes. Subsequently, hierarchical clustering analysis was employed to group these outlier genes based on their oscillating expression levels, with the purpose of identifying clusters that exhibited similar period and phase patterns. We hypothesized that genes found in the same cluster would share common transcription factors that would drive and synchronize the expression of those genes. We tested this hypothesis and found that genes with the same circadian pattern did indeed share common transcription factor binding sites, using Ciiider 3.0 (Gearing et al, 2019), an internet-based tool for predicting and analyzing transcription factor binding sites. We found 96 genes grouped into 6 clusters for MCF10a, and 92 genes grouped into 6 for MCF7 cell lines. Arid3A binding sites were highly enriched across MCF7 and MCF10a. In MCF10a, where circadian oscillations were more pronounced, Sox5 binding sites were highly enriched for cluster 1, Meis1 was enriched for cluster 2, and Arid3A was enriched in clusters 3-6, in combination with other transcription factors. In MCF7, a different set of genes showed circadian-like oscillating expression patterns. However, these different genes showed enrichment of Arid3A across all clusters. Since Arid3A binding sites were found in genes involved in all of the MCF7 clusters, and found in four of the six clusters in MCF10a, further work is needed in understanding the additional regulatory mechanisms that shift the phase and period of the oscillations of these groups of genes. In addition, more work needs to be done to study the role of these transcription factors in cancer progression and basic cell processes. Most of these transcription factors are known to be related to breast or other forms of cancer. Though this list is not complete, some of the more enriched transcription factor binding sites include Arid3A, Meis1, Znf354c, in MCF10a and in MCF7 they include Arid3A, Znf354c, Nfatc2, and Ahr:Arnt. The most enriched transcription factor amongst these circadian-like genes is Arid3A. Arid3A is an H3K9me3 demethylase and may impact regulation of other genes through demethylation. Suppression of Arid3A is associated with leukemia. Meis1 is also implicated in leukemia and tumor growth, as well as sleep disorders. Arnt splice variants are also associated with breast cancer. Future research is needed to determine if these transcription factors are important for circadian-like control. Our in silico results suggest that genes found in the same clusters share their respective transcription factors that regulate the expression of these circadian-like genes. Citation Format: Yocanxóchitl Perfecto-Avalos, Janice Yoo, Cinthya Saravia Heyman, Princess Jemeala Deranamie, John Tran, Sean-Patrick Scott. Predicted Circadian-related Transcription Factors in Human Breast Cell Lines–A key to understanding breast cancer pathology? [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-07-01.
Read full abstract