Alternative splicing is an FXRa loss-of-function mechanism and impacts energy metabolism in hepatocarcinoma cells.

Farnesoid X receptor a (FXRa) is a bile acid-activated nuclear receptor that sets the expression of glycolytic and lipogenic target genes by interacting with the 9-cis-retinoic acid receptor a (RXRa). Jointly with recruited cofactors, the current FXRa proteins reported so far in human and rodent have been described to regulate both isoform (a1-4)- and tissue-specific gene expression profiles in order to integrate energy balance and adapt metabolism. Here, we studied the biological functions of a FXRa naturally occurring spliced exon 5 isoform (FXRase5) lacking the second zinc-binding module of the DNA Binding Domain (DBD). We show that the FXRase5 is expressed in all FXRa-expressing tissues and cells from human and mouse, and that it is not able to either bind to its response element or activate FXRa dependent transcription. In parallel, this spliced variant displayed differential interaction capacities with its obligate heterodimer partner RXRa that may account for silencing of this permissive dimer for signal transduction. Finally, deletion of exon 5 by gene edition in HepG2 cells leads to FXRa loss-of-function, increased expression for LRH1 metabolic sensor and CD36 fatty acid transporter in conjunction with changes in glucose and triglycerides homeostasis. All together, these findings highlighted a novel mechanism by which alternative splicing may regulate FXRa gene function to fine-tune adaptive and/or metabolic responses. Taking this into account could deepen our understanding on how splicing events might hinder FXRa activity to regulate specific transcriptional programs and contribute to modify energy metabolism in normal tissues and metabolic diseases. Overall design: To verify whether and to what extent the FXRase5 isoform may modulate FXRa transcriptional network through specific binding to other FXRE DNA sequence motifs, we used RNA sequencing to explore the transcriptome of HepG2 overexpressing either FXRa1 or FXRase5. By comparing genes regulated by these two FXRa isoforms with the empty control vector, we generated the heatmap and the Venn diagram of the RNAseq datasets