Circadian clocks and periodic anticipated fasting prevent fasting-associated hepatic steatosis in calorie restriction

Calorie restriction (CR) improves health and longevity. CR induces a periodic fasting cycle in mammals; our study compares CR with unanticipated fasting (F), when the food is unexpectedly withheld. F induces hepatic steatosis, while CR reduces it; surprisingly, the difference is not due to hepatic ß-oxidation. Liver transcriptome analysis identifies fatty acid transporters (Slc27a1 and Slc27a2), TAG synthesis (Gpat4), and lipid storage (Plin2 and Cidec) genes to be upregulated only in F in agreement with hepatic steatosis. The circadian clock and anticipated fasting contribute to preventing fasting-associate hepatic steatosis in CR. Mechanistically, Slc27a1, Plin2 and Cidec genes are upregulated, and liver TAGs accumulate in circadian clock mutant mice on CR or if wild type CR mice miss their anticipated meal. The results highlight similarities and differences between unanticipated fasting and CR, suggesting that circadian clock dependent gating of transcriptional response to fasting controls lipid homeostasis and prevents hepatic steatosis. Overall design: Total RNA was isolated from 30 mg of female mouse liver tissue using mini spin column QIAGEN RNeasy Mini Kit according to the manufacturer's protocol. Sample purity was checked on Nano Drop (Thermo Fisher Scientific, Waltham, MA, USA) with 260/280 and 260/230 ratios = 2.0. The RNA sequencing was performed by Novogene Corporation Inc (Sacramento, CA, USA). After sample quality control, non-directional – polyA enrichment library was prepared using NEBNext Ultra™ II RNA Library Prep Kit for Illumina (New England BioLabs), and 50M reads (150bp, paired end) per sample were generated on Illumina NovaSeq 6000 platform. RNA-seq reads were mapped to the mouse protein coding genes (Ensembl: Mus_musculus; GRCm38) using Bowtie allowing up to 2-mismatches. The gene expected read counts and Transcripts Per Million (TPM) were estimated by RSEM (v1.2.3). The TPMs were further normalized by EBSeq R package to correct potential batch effect.