Clinical Implications of the Melatonin Phase Response Curve

In this issue of JCEM, the Burgess et al. report (1) on the melatonin phase response curve (PRC) has many clinical implications worthy of further discussion. A PRC is a mathematical tool established by convention because of its heuristic value for describing the direction and magnitude of phase shifts in the endogenous circadian pacemaker and its driven rhythms that result from applying a stimulus, such as melatonin or bright light. One of these driven rhythms results in a melatonin curve or profile; melatonin production by the pineal gland occurs almost exclusively at night in both diurnal and nocturnal species, resulting in a profile shaped like a square wave (2). Thus, the timing of melatonin production is governed by, and feeds back time-of-day information to, the pacemaker—which is primarily entrained to the light/dark cycle. [However, there is no negative feedback inhibition limiting the amount of melatonin produced (3).] Endocrinologists, as well as other clinicians, are primarily interested in: 1) the function of melatonin in humans as a neurohormone; 2) pathologies related to melatonin production; and 3) therapies resulting from melatonin supplementation. The jury is out as to whether or not there are significant melatonin disorders that are simply a matter of hyper- or hyposecretion. The medical implications of melatonin are best understood in a circadian context. Circadian disorders are numerous and important, and their treatment is informed by the melatonin (and light) PRCs.

A PRC is characterized by two opposing regions and a nonresponsive “dead zone” and thus has the appearance of a partially flattened sine wave. The dead zone of the PRC to light is during the day. The dead zone of the melatonin PRC, however, occurs during the “biological night,” that is, the time when endogenous levels of melatonin are usually high. Responses to melatonin are greatest when it is given exogenously at times when endogenous levels are not normally present, that is, during the day; when given in the morning, melatonin causes phase delays (shifts to a later time), and when given in the afternoon/evening it causes phase advances (shifts to an earlier time). Bright light causes phase shifts opposite to those caused by melatonin; that is, light exposure in the morning causes phase advances, and in the evening causes phase delays (responses are greatest during the night). The Burgess et al. paper (1) is the first to describe the melatonin PRC to two different doses (0.5 vs. 3 mg). It is of interest that they are not much different in the maximal phase shifts they cause. However, their PRCs indicate different times when each dose produces its maximal phase shifts, and Burgess et al. (1) discuss the idea of achieving overlap between the exogenous and endogenous levels for maximizing phase-shifting magnitude (4). Another such pharmacokinetic principle is that even when exogenous melatonin is administered on the correct zone, care must be taken to avoid unnecessarily high doses that would cause trailing levels that spill over onto the wrong zone of the melatonin PRC (5). These pharmacokinetic parameters are not as applicable to bright light, which may account for its more robust phase-shifting effects (5,6); however, taking melatonin is more convenient and may therefore be the preferred treatment modality.

Both bright light and melatonin can be used to treat circadian phase disorders. In some cases, the therapeutic goal is to move the sleep bout to a desired time, such as in advanced and delayed sleep phase syndromes (one common variant of the latter is Monday morning blues, which is caused by circadian rhythms that have delayed over the weekend, making it difficult to fall asleep Sunday night). In other cases, realigning the internal phase relationship between the endogenous circadian pacemaker and the sleep/wake cycle needs to be corrected. Disorders of circadian misalignment include shift work maladaptation and jet lag, as well as adjusting to and from the transition to Daylight Saving Time. More serious disorders that may be related to circadian misalignment are seasonal and nonseasonal depression.

In disorders of circadian misalignment, the best biomarker for the timing (phase) of the endogenous circadian pacemaker is the dim light melatonin onset (DLMO) (7). The DLMO can be obtained in the clinical laboratory, in the sleep lab, or at home; very dim ambient light conditions are required to avoid suppression of melatonin by light (8), although lenses that filter blue light, to which melanopsin [the retinal photoreceptor most important for melatonin suppression (9)] is most sensitive, may allow more permissive conditions. Saliva samples are collected every 30–120 min beginning at 1800 h to determine when melatonin levels rise above a certain threshold, usually 3 pg/ml [which was the average value in the Burgess et al. paper (1)].

A person’s DLMO marks the time of the beginning of their biological night. Other circadian rhythms can also be used as phase markers, such as cortisol or wake time; therefore, it is important to be able to relate the phase markers of these different rhythms to each other. The conventional way to do this is by using circadian time (CT). The 3 pg/ml salivary DLMO occurs on average about 14 h after wake time (7); habitual wake time is CT 0, the DLMO is CT 14, and either of these phase markers can be used to estimate the time points of an individual’s PRCs. The melatonin PRC is divided into two zones (7): the advance zone and the delay zone. The advance zone extends from CT 6 to CT 18 [bedtime is usually CT 16, so taking melatonin at this time will cause a (small) phase advance, although the optimal time to take melatonin to cause a phase advance is CT 8–12]. The delay zone extends from CT 18 to CT 6 (CT 24 = CT 0); wake time, or CT 0, is the most convenient time to take melatonin to cause maximal phase delays. In the Burgess et al. paper (1), CT is implied but is not explicitly used, as when CT 8–12 (the time when melatonin causes maximal phase advances) can be referred to as 6 to 2 h before the DLMO (1,7). (For a person who habitually wakes up at 0600 h, CT 0, CT 6, CT 12, and CT 18 convert on average to clock times of 0600, 1200, 1800, and 2400 h, respectively.)

The DLMO is also useful for assessing circadian misalignment and for phase typing. A DLMO that is relatively delayed with respect to the sleep/wake cycle indicates phase-delayed circadian misalignment, and therefore bright light and melatonin should be used to cause a corrective phase advance; a DLMO that is relatively advanced with respect to the sleep/wake cycle indicates phase-advanced circadian misalignment requiring a corrective delay. According to Burgess et al. (1), additive effects can be expected, at least with respect to causing phase advances. In healthy controls, the DLMO occurs about 6 h earlier than the middle of the sleep bout; therefore, this phase angle difference or time interval has been used for phase typing, at least for patients with seasonal affective disorder [a greater than 6-h interval indicates a phase-advanced type of seasonal affective disorder (SAD) patient; an interval of 6 h or less indicates the more prototypical phase-delayed type (10)].

SAD, or winter depression, was the first disorder in which the antidepressant effects of melatonin were first demonstrated (10). More recently, melatonin has been found to be antidepressant in delayed sleep phase syndrome with comorbid depression (11). In nonseasonal depression, circadian misalignment seems to be of the phase-delay type, although the jury is still out on whether or not a phase-advance type of nonseasonal affective disorder exists, particularly in bipolar depression or, more likely, mania. In severe depression, patients have more phase-delayed circadian misalignment than controls (12); perhaps more important is that symptom severity correlates with the degree of circadian misalignment (13). Thus, circadian misalignment may be one of the first biomarkers in psychiatry, particularly because it appears to be at least partly causal of depression, as shown in the case of SAD in which melatonin given at the incorrect time increases circadian misalignment and is not antidepressant (compared with inactive placebo) and makes some patients more depressed (10).

Much of the above could not have been accomplished without a great many animal studies applied to human melatonin physiology and then pathophysiology (an example of translating basic research to clinical applications). However, there have been examples of “reverse translation” as well, in which clinical investigations have preceded basic studies. Giving melatonin over 3–4 d (7), undertaken to show small phase shifts in humans when melatonin is given near the dead zone, was later adapted to demonstrate the melatonin PRC in rodents (14). Previously, an unusually shaped PRC had been proposed to explain why a single dose of melatonin only appeared to cause phase shifts (advances) during a small window in the late subjective day (15) to explain the first demonstration of entraining a free-running mammal (rodent) to melatonin, a seminal paper (16) that inspired the use of melatonin to entrain free-running totally blind people (17) that is to date the most recognized use of melatonin as a chronobiotic in the treatment of a clinically significant disorder. Another example of reverse translation occurred when it was shown that rodents living in the wild were less sensitive to laboratory light than those who were maintained in indoor-intensity light (18), which was based on the results and explanation offered as to why humans appeared to be less sensitive to laboratory light than other animals in the melatonin suppression test (8). The Burgess et al. study (1) of the melatonin PRC under highly controlled laboratory light and sleep conditions is yet another example of reverse translation, in that it validates the first study of the melatonin PRC undertaken in humans living at home (7).

Returning to that which is especially important to endocrinologists, one possible function of melatonin is that it augments the phase-shifting effects of light using the suppressant effect of light unique to melatonin production; this would explain why humans have retained the suppressant effect of light when our lack of reproductive and other seasonal rhythms obviates the need to measure the length of the biological night by the shortened duration of melatonin production in the summer. A more important function ends when we are 3 months old, at which time entrainment to the light/dark cycle becomes functional (see Ref. 19 for additional references). For entrainment before that age (19), the third trimester fetus and “fourth trimester” suckling infant may have to rely on the mother’s melatonin signal, through placental transfer or breast milk, respectively; although entrainment may or may not be directly important for development, a mother whose sleep is not disturbed by a baby on a different circadian rhythm will improve her caregiving. It should also not be ruled out that entrainment may be important in the course of early and neonatal development (19); indeed, low melatonin production in the mothers of some autistic children (20) may have been a necessary but not sufficient cause that might have been ameliorated by giving the mother melatonin at least during the third and fourth trimesters when the neonate’s melatonin receptors begin to function (19).