(B-56) Metabolic modeling of sex-specific liver tissue suggests mechanisms of differences in toxicological responses

Introduction:: Male subjects in animal and human studies are disproportionately used for toxicological testing^1,2. This discrepancy is evidenced in clinical medicine where females are more likely than males to experience liver-related adverse events in response to xenobiotics. While previous work has shown gene expression differences between the sexes³, there has been a lack of systems-level approaches to understand the mechanisms underlying these observations. Here, we integrate gene expression data with metabolic network models to characterize the impact of transcriptional changes of metabolic genes in the context of sex differences and drug treatment.


Materials and Methods:: We used AERSMine⁴ to evaluate the difference in adverse event reporting in the United States between the first quarter of 2004 to the third quarter of 2021. Only liver-related reports with age and sex information were used, and patients between the ages of 15 to 65 were considered. Once compiled, the percentage of total reports was calculated for each quarter.

Differential expression between male and female samples from data sets were found in the Gene Expression Omnibus (Liver: GSE130991, Kidney: GSE36059, Brain: GSE5281). Only samples considered “control” were used. Drug treatment samples were sourced from ToxicoDB⁵. Differentially expressed genes were found using the limma package⁶.

To find the presence of sex-biased metabolic subsystems, tasks inferred from differential expression (TIDEs)⁷ and Human1⁸ were used. TIDEs identifies all genes related to each KEGG ortholog subsystem and assigns each reaction a weight based on the log2 fold difference in those genes’ expression. Each fold difference is averaged for a given task to calculate its task score. This score is compared to 1000 randomized task scores, calculated using randomly chosen log fold difference weights from other tasks.

Sex-specific liver models were created using Human1, gene expression data (GSE130991), and Reaction Inclusion by Parsimony and Transcript Distribution (RIPTiDe)⁹. RIPTiDe uses sex-specific gene expression data to prune the model of reactions that do not have significant transcriptional evidence. This technique was performed with male and female liver data, with 582 (457 female/125 male) individual models and averaged models for male and female created and sampled.


Results, Conclusions, and Discussions:: Our analysis of the Adverse Event Reporting System indicates that women report liver-related side effects more frequently than their male counterparts. However, we find many drugs are more frequently reported by male patients, suggesting sex may be an important factor in toxicity. To explore the metabolic differences between male and female livers, we use TIDEs to compare which metabolic subsystems are sex-biased in the liver, kidney, and brain using sex-specific gene expression data. We find several subsystems to be altered by sex exclusively in the liver including fatty acid metabolism, glycolysis, pentose and glucuronate interconversions, and xenobiotics metabolism. We then use TIDEs to compare hepatocytes treated with several known toxicants to those that were untreated and find fatty acid metabolism broadly affected by treatment. We then create individual and averaged male and female liver models using RIPTiDe and find reactions in glycolysis, the TCA cycle, the pentose phosphate pathway, and nucleotide metabolism to be significantly different between the sexes. This collection of subsystems suggests there may be a sexually dimorphic response to oxidative stress that causes some drugs to exhibit sex-specific toxicity.


Acknowledgements (Optional): :

References (Optional): : [1] Feldman, Sergey, et al. "Quantifying sex bias in clinical studies at scale with automated data extraction." JAMA network open 2.7 (2019): e196700-e196700.

[2] Zucker, Irving, and Annaliese K. Beery. "Males still dominate animal studies." Nature 465.7299 (2010): 690-690.

[3] Yang, Xia, et al. "Tissue-specific expression and regulation of sexually dimorphic genes in mice." Genome research 16.8 (2006): 995-1004.

[4] Sarangdhar, M., Tabar, S., Schmidt, C., Kushwaha, A., Shah, K., Dahlquist, J. E., ... & Aronow, B. J. (2016). Data mining differential clinical outcomes associated with drug regimens using adverse event reporting data. Nature biotechnology, 34(7), 697-700.

[5] Nair, S. K., Eeles, C., Ho, C., Beri, G., Yoo, E., Tkachuk, D., ... & Haibe-Kains, B. (2020). ToxicoDB: an integrated database to mine and visualize large-scale toxicogenomic datasets. Nucleic acids research, 48(W1), W455-W462.

[6] Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK (2015). “limma powers differential expression analyses for RNA-sequencing and microarray studies.” Nucleic Acids Research, 43(7), e47. doi: 10.1093/nar/gkv007.

[7] Dougherty, B. V., Rawls, K. D., Kolling, G. L., Vinnakota, K. C., Wallqvist, A., & Papin, J. A. (2021). Identifying functional metabolic shifts in heart failure with the integration of omics data and a heart-specific, genome-scale model. Cell Reports, 34(10), 108836.

[8] Robinson, J. L., Kocabaş, P., Wang, H., Cholley, P. E., Cook, D., Nilsson, A., ... & Nielsen, J. (2020). An atlas of human metabolism. Science signaling, 13(624), eaaz1482.

[9] Jenior, M. L., Moutinho Jr, T. J., Dougherty, B. V., & Papin, J. A. (2020). Transcriptome-guided parsimonious flux analysis improves predictions with metabolic networks in complex environments. PLoS computational biology, 16(4), e1007099.