HIBIT2018 11th International Symposium on Health Informatics and Bioinformatics, Antalya, Turkey, 25 - 27 October 2018, pp.83-84
Network models for understanding boron-induced transcriptomics changes within HepG2
Ayşegül Tombuloğlu, Hülya Çöpoğlu, Tülin Güray and Yeşim Aydın Son
Middle East Technical University, Graduate School of Informatics, Health Informatics
Boron has crucial roles in plant growth and survival; also, it is suggested as an essential trace element for
human physiology. Accumulating evidence show beneficial effects of boron for human health. Along with
its benefits to bone and brain health, many findings support the anti-carcinogenic role of dietary boron.
Although biochemical significance of boron is evident, relatively few studies focus on boron-induced
biological processes and mechanisms at the molecular level.
In this work, we aim to reveal the boron-induced molecular mechanisms in detail, and our preliminary
findings of network modelling studies is presented. HepG2 cell line is treated with boric acid (BA) at halfmaximal
inhibitory concentration (IC50) for 24 hours. Differential gene expression profile relative to nontreated
HepG2 cells is investigated with microarray technology. Gene expression changes in HepG2 cells
cultured with different chemicals are also obtained by re-analysis of published data, and integrated with
gene expression changes due to boric acid treatment. Integrated gene expression data is used for
reconstructing a weighted gene co-expression network to investigate the bioactivity of boric acid in a
comparative manner. As a second approach, a regulatory network is build using boric acid induced gene
expression data with motif knowledge and known physical interactions among transcription factors.
At half-maximal inhibitory concentration, boric acid treatment lead to a massive down-regulation of
genes which take part in in cell-cycle progression and various metabolic processes. Few genes involved in
apoptosis and cytokine-cytokine receptor interaction pathway had elevated levels of expression in the
presence of boric acid exposure. Comparative network analysis indicated the induction of many
functional gene groups, which appear to be specifically associated with boric acid treatment. Regulatory
network revealed transcription factor-gene interactions, which will help us to exploit the effected
regulatory mechanisms at transcriptomics level in the presence of highly concentrated boron.
According to our results, a group of genes involved in lipid metabolism might be particularly meaningful
since latest research suggest potential therapeutic activity of boron in lipid dysregulation disorders like
fatty liver disease and obesityThese genes were down-regulated in boric acid treated cells and formed a
boric acid associated subnetwork contrasting with the carcinogenic chemical induced profile. In the
future we plan to validate the key proteins in the regulatory network in cell culture. Moreover, we aim to
recapitulate the microarray experiments and carry out subsequent network modelling at lower
concentrations of boric acid to study the boric acid related network patterns in a concentration