Interferon was one of the first (1983) cytokines to be implicated in sleep [
20] but has been less investigated than others, probably because its species specificity limits generalizations. The following year IL-1, previously known as endogenous pyrogen (the link being unexpectedly made because both were identical to serum amyloid A-inducer and lymphocyte-activating factor (LAF) [
21]) was first associated with sleep [
22]. TNF, first described in 1975 for its
in vivo capacity to kill tumor cells [
23], was, six years later, shown to kill malaria parasites
in vivo, and proposed, along with IL-1 (then known as LAF), to cause the disease complexities of malaria and sepsis [
24,
25]. While sleep aberrations are part of these conditions, they were not singled out as a particular outcome of the presence of these cytokines. Soon after becoming available in recombinant form, these cytokines were confirmed to be linked to physiological sleep in 1987 [
26], and an awareness developed of the metabolic and disease relevance of this association [
27]. As reviewed in 1995 [
28], this group and others had, by then, done considerable work on these effects being amplified by the increased cytokines generated by microbial infections, and also the implications of their functional redundancy. Moreover, just before the normal time of sleep onset for rats, TNF levels in brain tissue were shown to be 10-fold higher than their daily minimum [
29]. Diurnal variations of the soluble TNF receptors (two forms exist, induced by increases in TNF) in plasma from healthy human volunteers are consistent with this model [
30]. Key steps in establishing the importance of TNF in sleep were its suppression by an anti-TNF antibody [
31] and both spontaneous and influenza-induced sleep being variously altered in double TNF receptor-deficient mice [
32]. In brief, when uninfected, these mice had less non-rapid eye movement sleep (NREMS) than wild-type mice at night-time and more rapid eye movement sleep (REMS) than control mice during the day, whereas challenge with mouse-adapted influenza X-31 enhanced NREMS and decreased REMS in both strains to roughly the same extent. In addition, the strain lacking TNF receptors had higher levels of orexin mRNA. As recently summarized [
33], wakefulness enhances TNF protein levels and expression in brain, and the highest normal brain levels, at least in the rat, occur at the time of usual sleep onset. Sleep deprivation elevates levels even further, the effects of which we experience in jetlag.
The nocturnal surge of melatonin that arises in the pineal gland, and determines the synchronization of pineal function with the diurnal cycle, has been studied extensively in normal physiology. Melatonin is, however, relatively absent from the literature on sleep variation in disease, with the exception of a recent valuable contribution [
34]. In brief, therefore, we note that melatonin is well-recognized as an inhibitor of TNF [
35‐
37], and that TNF, in turn, transiently inhibits its production [
38]. Inferences regarding the previous paragraph can be drawn from these observations.