NAMPT-dependent NAD+ biosynthesis controls circadian metabolism in a tissue-specific manner

Abstract

Molecular clocks in the periphery coordinate tissue-specific daily biorhythms by integrating input from the hypothalamic master clock and intracellular metabolic signals. One such key metabolic signal is the cellular concentration of NAD⁺, which oscillates along with its biosynthetic enzyme, nicotinamide phosphoribosyltransferase (NAMPT). NAD⁺ levels feed back into the clock to influence rhythmicity of biological functions, yet whether this metabolic fine-tuning occurs ubiquitously across cell types and is a core clock feature is unknown. Here, we show that NAMPT-dependent control over the molecular clock varies substantially between tissues. Brown adipose tissue (BAT) requires NAMPT to sustain the amplitude of the core clock, whereas rhythmicity in white adipose tissue (WAT) is only moderately dependent on NAD⁺ biosynthesis, and the skeletal muscle clock is completely refractory to loss of NAMPT. In BAT and WAT, NAMPT differentially orchestrates oscillation of clock-controlled gene networks and the diurnality of metabolite levels. NAMPT coordinates the rhythmicity of TCA cycle intermediates in BAT, but not in WAT, and loss of NAD⁺ abolishes these oscillations similarly to high-fat diet-induced circadian disruption. Moreover, adipose NAMPT depletion improved the ability of animals to defend body temperature during cold stress but in a time-of-day-independent manner. Thus, our findings reveal that peripheral molecular clocks and metabolic biorhythms are shaped in a highly tissue-specific manner by NAMPT-dependent NAD⁺ synthesis.

Keywords: Brown adipose tissue; Circadian metabolism; Clock rhythm; NAD; Skeletal muscle.

Fig. 1.

NAMPT-dependent NAD⁺ levels regulate the molecular clock in a tissue-specific manner. (A) Mouse models used to study the tissue-specific roles of Nampt-dependent NAD⁺ levels for the molecular clock: fat-specific Nampt knockout (FANKO) and inducible skeletal muscle-specific Nampt knockout mice (iSMNKO). Gene expression and NAD⁺ levels in tissues from 2 to 4-mo-old WT, and tissue-specific Nampt KO (FANKO, iSMNKO) male mice. Tissues were harvested at 4-h intervals over the course of 24 h, with ZT 0 denoting the start of the light phase. Nampt expression in (B) BAT and (C) eWAT (n = 4 to 5). NAD⁺ levels in (D) BAT and (E) eWAT (n = 4 to 5). Arntl, Cry1 and Nr1d1 expression in (F–H) BAT, and (I–K) eWAT (n = 4 to 5). (L) Nampt expression and (M) NAD⁺ levels in gastrocnemius muscle (n = 5 to 8). (N–P) Arntl, Cry1 and Nr1d1 expression in gastrocnemius (n = 5 to 8). Significant differences found by 2-way ANOVA or JTK rhythmicity are noted. Data from RNA sequencing performed on BAT and eWAT at ZT 6, 10, 18, and 22 (n = 4). (Q) Heatmap of amplitude of core clock genes, * denotes significantly different circadian rhythmicity between WT and FANKO. (R) Nicotinamide riboside (NR) intravenous infusion setup. NAD⁺ levels in (S) BAT and (T) eWAT after NR infusions in WT and FANKO mice (n = 5 to 6). Cry1 expression in (U) BAT and (V) eWAT after NR infusions in WT and FANKO mice (n = 5 to 6). (W) Cry2 expression in BAT after NR infusions in WT and FANKO mice (n = 5 to 6). (X) Nr1d1 expression in eWAT after NR infusions in WT and FANKO mice (n = 5 to 6). Significant differences were found by 2-way ANOVA. * denotes significant difference between WT and FANKO. # denotes significant difference with treatment. $ denotes significant difference from WT vehicle treated.

Fig. 2.

Nampt controls global adipose transcriptional rhythmicity through changes in circadian amplitude. Data from RNA sequencing performed on BAT and eWAT at ZT 6, 10, 18, and 22 from Fig. 1. (n = 4). (A) Euler diagram of oscillating transcripts in WT BAT and eWAT. (B) Euler diagram of oscillating transcripts in WT and FANKO BAT. (C) Euler diagram of oscillating transcripts in WT and FANKO eWAT. (D) Heatmap of amplitude and (E) peak time of rhythmic transcripts in BAT and eWAT of WT and FANKO mice. (F) Gene ontology analysis of circadian genes in all four groups. (G) UpSet plot showing selected overlaps of circadian rhythmicity of genes in the four groups. All analyses were performed with FDR cutoff < 0.01.

Fig. 3.

Nampt ablation affects metabolite rhythmicity in brown adipose tissue. Metabolomics analysis of BAT of WT and FANKO mice harvested at ZT 6 and ZT 18 (n = 4 to 6). (A) Euler diagram of metabolic features in BAT significantly affected by genotype, time-of-day, or an interaction between these two factors. (B) List of KEGG pathways with changed diurnal pattern in response to Nampt deletion in BAT. (C) Schematic overview of pyruvate metabolism and the TCA cycle, and plots displaying the content of metabolic intermediates with changed diurnal rhythm in BAT in response to Nampt deletion. (D) Levels of the arginine biosynthesis intermediate, N-acetylornithine, in WT and FANKO BAT. Measurements of (E) blood glucose and (F) blood lactate levels in WT and FANKO mice at ZT 6 and ZT 18 (n = 7). (G) Liver glycogen levels in WT and FANKO mice at 4-h intervals over the course of 24 h. (H) List of selected KEGG pathways regulated at the transcriptional and metabolic level. M: metabolite, G: gene, and T: total number of metabolites and genes in pathway. Significant differences found by 2-way ANOVA or JTK rhythmicity are noted. All metabolomics analyses were performed with a FDR cutoff < 0.05. * denotes significant difference between WT and FANKO. £ denotes significantly difference with time-of-day. $ denotes significant difference with time-of-day in WT.

Fig. 4.

The effect of Nampt ablation on metabolite rhythmicity is tissue and diet dependent. Metabolomics analysis of eWAT from WT and FANKO mice harvested at ZT 6 and ZT 18 (n = 4 to 6). (A) Euler diagram of metabolic features in eWAT significantly affected by genotype, time-of-day, or an interaction between these two factors. (B) Metabolites affected by both genotype and time-of-day in eWAT. Metabolomics analysis of BAT of chow- and high-fat diet (HFD)-fed WT and FANKO mice harvested at ZT 6 and ZT 18 (n = 2 to 4). (C) Euler diagram of metabolic features significantly affected by genotype, time-of-day, and HFD as determined by three-way ANOVA. Day and night levels of (D) cis-aconitate, (E) fumarate, and (F) malate in BAT of WT and FANKO mice fed chow or HFD. (G) Euler diagram of metabolic features in BAT of chow-fed animals significantly affected by genotype, time-of-day, or an interaction between these two factors. (H) Euler diagram of metabolic features in BAT of HFD-fed animals significantly affected by genotype, time-of-day or an interaction between these two factors. (I) Schematic of acute cold tolerance test. (J) Light and (K) dark phase cold tolerance test. (L) Body temperature drop for the two cold tolerance tests. All metabolomics analyses were performed with FDR cutoff < 0.05. * denotes significantly difference between WT and FANKO. £ denotes significantly difference with time-of-day. $ denotes significant difference with time-of-day in WT at the same diet.