Tick-Tock Consider the Clock: The Influence of Circadian and External Cycles on Time of Day Variation in the Human Metabolome-A Review

Abstract

The past decade has seen a large influx of work investigating time of day variation in different human biofluid and tissue metabolomes. The driver of this daily variation can be endogenous circadian rhythms driven by the central and/or peripheral clocks, or exogenous diurnal rhythms driven by behavioural and environmental cycles, which manifest as regular 24 h cycles of metabolite concentrations. This review, of all published studies to date, establishes the extent of daily variation with regard to the number and identity of 'rhythmic' metabolites observed in blood, saliva, urine, breath, and skeletal muscle. The probable sources driving such variation, in addition to what metabolite classes are most susceptible in adhering to or uncoupling from such cycles is described in addition to a compiled list of common rhythmic metabolites. The reviewed studies show that the metabolome undergoes significant time of day variation, primarily observed for amino acids and multiple lipid classes. Such 24 h rhythms, driven by various factors discussed herein, are an additional source of intra/inter-individual variation and are thus highly pertinent to all studies applying untargeted and targeted metabolomics platforms, particularly for the construction of biomarker panels. The potential implications are discussed alongside proposed minimum reporting criteria suggested to acknowledge time of day variation as a potential influence of results and to facilitate improved reproducibility.

Keywords: blood; breath; circadian rhythms; diurnal rhythms; metabolite rhythms; metabolomics; saliva; skeletal muscle; urine.

Figure 1

Mock data representation of two biological rhythms (e.g., metabolite rhythms) X (top) and Y (bottom) under entrained conditions (left) and constant routine (right). Entrained conditions consist of a light/dark cycle, meals at 08:00, 13:00, 19:00 h (shown by dashed vertical lines) and designated sleep time between 23:00 and 07:00 h (shown by grey shading of x-axis). (Top-left): Metabolite X: A rhythm with a regular 24 h period under entrained conditions may or may not be influenced by Zeitgeber (e.g., wake/sleep). (Top-right): Metabolite X: The rhythm observed under entrained conditions persists and maintains its 24 h periodicity under constant routine, the amplitude may or may not change between the conditions, and the rhythm is considered circadian. (Bottom-left): A rhythm with a more complex cycle, but regular 24 h period, and peaks correspond to mealtimes under entrained conditions (08:00, 13:00, 19:00 h), suggesting some effect of feeding/fasting cycles. (Bottom-right): The rhythm is significantly dampened in constant routine (does not persist), with a regular period/amplitude no longer detectable. The rhythm of metabolite Y is not considered circadian in nature, as it did not persist under constant routine conditions, and is classed as a diurnal rhythm, i.e., a rhythm evoked by exogenous cycles such as feeding/fasting and sleep/wake.

Figure 2

Comparative metabolite/hormone profiles under entrained conditions: sleep vs. prolonged wakefulness (A–C), entrained vs. circadian constant routine conditions (D), and inter-individual variation under constant routine conditions (E), reproduced from Honma et al. [50] (A), Davies et al. [51] (B,C), Czeisler & Klerman [52] (D), and Chua et al. [53] (E). A,B,C: Comparative profiles of cortisol, dodecanoylcarnitine (C12), and taurine under entrained light/dark conditions, sleep (highlighted in black) vs. prolonged wakefulness (highlighted in grey), with meals provided at 07:00 h, 13:00 h, 19:00 h, and 22:00 h (snack). No significant difference observed in cortisol, a SCN-driven hormone, between sleep conditions (A), peaks in measured intensity (y-axis) corresponding to mealtimes observable in various lipids and amino acids (B,C), alongside statistically significant perturbations during prolonged wakefulness vs. sleep [51]. D: Growth hormone rhythm observable under entrained conditions (peak during sleep) but dampened under constant routine. E: Individual rhythmic profiles of six participants showing inter-individual variation in lipid profiles (SM18:1/24:1), with two individuals displaying an inversed rhythm (~12 h ahead/delayed) relative to the other four participants.