Mouse is the most widely used preclinical animal model yet the translational rate is dismal. To understand the molecular underpinnings of the differential pathophysiology between human and mouse, we analyzed similar or discordant gene expression networks regulated by a variety of metabolic conditions under multiple complementary experimental settings, and tested these regulations in humanized mice where the confounding factors are minimized. We first analyzed human and mouse liver transcriptomes derived from conditions ranging from differentially expressed genes in fatty liver disease to genes regulated by transcriptional or posttranscriptional factors. In addition to genes that are exclusively regulated in human or mouse, we identified that a large fraction of commonly regulated genes (up to 50%) exhibit completely opposite regulation. To determine the intrinsic similarity and differences between human and mouse liver biology, we further analyze gene expression in a humanized mouse model in which human and mouse hepatocytes co-exist in the humanized liver and are exposed to identical in vivo metabolic milieu. Strikingly, divergent gene expressions in human and mouse hepatocytes are recapitulated in the humanized mice. By integrating the data of humanized mice, we are able to reveal the intrinsic consistently and oppositely regulated genes and pathways between human and mouse in each condition we analyzed. Although the regulations of these genes are largely condition dependent, the most constant consistently or oppositely regulated pathways can be identified. Furthermore, our analysis suggests that human and mouse lncRNAs might play a critical role in the differential gene regulations in human and mouse. Our work provides a foundation to understand the molecular basis for divergent metabolic responses in human and mouse liver, highlighting the importance of a human or humanized system for translational studies.
C. Jiang: None. H. Cao: None.