This week, I presented this article published in the Journal of Neuroscience in 2010 to our sleep and circadian rhythms journal club. The experiments were undertaken by a group of Harvard neurobiologists who are experts on the temporal relationship between central levels of adenosine and time spent awake and asleep. In this study, they extended their interest in extracellular adenosine release to changes in adenosine triphosphate (ATP; the “currency” of the cell as we learned in our seventh grade biology class) during the normal course of sleep across a 24 hr period and pharmacologically-(adenosine) induced sleep in mice. The paper received a large amount of criticism from the basic sleep research community of which was communicated in several letters to the editor in another (unaffiliated) clinical and basic research journal; SLEEP. Instead of summarizing the methodological, analytical, and interpretational concerns raised by readers and addressed by the authors (some are linked here), I am focusing on the experimental questions and results (post-erratum).

It is well known that ATP is critical for keeping us alive; it directs cellular transport, signaling, division, and respiration, and muscle and cytoskeleton motility, to name a few biological processes. It is also the traffic controller of anabolic (biologially building) and catabolic (biologically destructive) processes. Given that sleep is necessary for biological restoration, the authors hypothesized that ATP levels would rise during sleep to replenish energy stores for subsequent energy use during waking. Indeed, ATP levels precipitously rose (and fell) during the first three hours of sleep which predominantly consists of deep, NREM sleep. More beautifully, the timing and extent of the surge in ATP could be modified by, for example, 3 hrs of sleep restriction followed by recovery sleep (there was a delay). Temporal changes in ATP levels during sleep, sleep restriction (SD), and recovery sleep (RS) are detailed below.

The authors’ hypothesis was also supported by reductions in metabolites such as phosphocreatine, which is utilized to make ATP, and elevations in P-AMPK during sleep restriction. AMPK serves as the (metabolic) switch operator to activate catabolic processes (it’s phosphorylated [P-AMPK]) or continue anabolic processes.

The minor concern of the study, however, was that the underlying biochemical dynamics of this surge in ATP was not discussed…or investigated. Hence, does the surge manifest from an increase in ATP synthesis, a decrease in ATP degradation, or a decrease in ATP usage? And, most notably, the calculated levels of ATP in tissue were GROSSLY OVERCALCULATED [by 100-fold] because the authors made a basic chemical error; they confused molarity with molality. Please let this be a reminder to all of us scientists who routinely undertake simple, but yet large amounts of calculations, to consult and review numbers the next time a manuscript is submitted for peer-review. I imagine a mistake of this sorts, although it did not change the general “trend” in results here, is a career killer, nevertheless.

Dworak M, McCarley RW, Kim T, Kalinchuk AV, & Basheer R (2010). Sleep and brain energy levels: ATP changes during sleep. The Journal of neuroscience : the official journal of the Society for Neuroscience, 30 (26), 9007-16 PMID: 20592221