2. The Functional Organization of the Human Circadian System
The circadian timing system, a hierarchically organized network of structures responsible for generating circadian rhythms, is driven in mammals by a circadian pacemaker located in the suprachiasmatic nuclei (SCN) of the hypothalamus. It allows organisms to adjust their physiology by anticipating daily environmental changes, instead of merely responding to them in a reactive manner. Thus, under natural conditions, endogenous circadian rhythms are entrained to the 24 h light-dark cycle (for a review, see [
9]). In humans, daily rhythms are observed in a variety of molecular, physiological and psychological processes, such as gene expression, body temperature, heart rate and melatonin production, as well as sleep, mood and higher cognitive functions (for a review, see [
10]). The circadian system (CS) consists of [
11,
12] (
Figure 1):
(a) Oscillatory machinery, including a central pacemaker, the SCN [
13], and peripheral oscillators located in most tissues and cells [
14]. Their rhythms are generated by a transcriptional-translational feedback loop between two groups of clock genes (positive and negative elements). Circadian locomotor output cycles kaput (
Clock) and brain and muscle aryl hydrocarbon receptor nuclear translocator-like (
Bmal1), acting as positive elements, are responsible for the synthesis of two transcription factors which, after heterodimerization, induce the expression of negative components of the molecular circadian clock, such as isoforms of Period (
Per 1,
2,
3) and Cryptochrome (
Cry1 and
Cry2) and a Nuclear receptor subfamily 1 (
Rev-Erbα) [
15,
16] (
Figure 2). An unknown clock gene, referred to as Chrono, has been recently added to this list. It seems to function as a transcriptional repressor of the negative feedback loop in the mammalian clock. Chrono binds to the regulatory region of clock genes, and its occupancy oscillates in a circadian manner [
17]. Variants of the core oscillator alternately use not only orthologs (e.g.,
Per 1,
2,
3;
Cry1,
2,
Bmal1,
2), but also paralogs, such as NPAS2 (neuronal PAS domain protein 2), which can replace CLOCK. The variants can exist as oscillators acting in parallel in the same organ [
18]. Moreover, the core oscillator system is associated with numerous, often tissue-specific accessory proteins that also undergo circadian cycling and additionally feed into the core oscillator. Among these, nicotinamide phosphoribosyltransferase (NAMPT) [
19], peroxisome proliferator-activated receptor-γ (PPARγ) [
20,
21], sirtuin 1 (SIRT1) [
22,
23], AMP-activated protein kinase (AMPK) [
24] and protein kinase Cα (PKCα) [
25,
26] are of particular importance, because they connect oscillators with metabolic sensing and mitochondrial function and are also controlled or modulated by melatonin [
27]. As metabolic sensors, these accessory oscillator components are also relevant to health, especially with regard to metabolic syndrome and diabetes type 2, but also in the context of aging [
28]. The connection between circadian oscillators and health maintenance extends to the prevention and suppression of cancer. Some core oscillator components, such as PER1 [
29], PER2 [
30,
31,
32] and BMAL1 [
33,
34], and the oscillatory output factor and modulator of Rev-erbα, deleted in breast cancer 1 (DBC1) [
35], have been shown to act as tumor suppressors. These insights originated from the crucial finding that mice carrying a mutation in the
Per2 gene are cancer-prone [
30]. Meanwhile, this conclusion has been supported by other data, including epigenetic knockdowns of core oscillator genes and oscillator dysfunction in cancer cells (summarized in references [
18,
28]).
(b) Input pathways carry information about the light-dark cycle to the central pacemaker. This pathway starts in a particular type of retinal ganglion cells containing melanopsin (which makes them intrinsically photosensitive). These cells are directly excited by blue light [
36], and send the information to the SCN through the retinohypothalamic tract (RHT). In addition to their intrinsic photosignal, they receive rod and cone inputs [
37,
38]. Other synchronizers, such as feeding cycles, scheduled physical exercise and social activities, are also connected to the central pacemaker and peripheral oscillators, contributing to their synchronization [
39]; however, only the light-dark cycle has been demonstrated to be a necessary and sufficient condition for circadian synchronization.
(c) Output pathways are responsible for the coordination of circadian rhythms between different functions and parts of the organism. These are the result of humoral mediators, such as prokineticin-2, which is able to generate the rhythm of locomotor activity [
40], and neural outputs, such as the rhythmic change in the parasympathetic/sympathetic balance [
41], or the pineal release of melatonin during darkness [
42]. This ubiquitous molecule is present in all biological domains, and it has been adopted during evolution as a “darkness molecule”. Its original antioxidant function, as well as its photosensitivity, caused it to be consumed during the day time (reducing oxidized molecules), thus peaking during the night [
43]. In mammals, melatonin is produced by the pineal gland at night, and its secretion is inhibited most efficiently by light at ~460–480 nm. It is important to mention that these outputs can also act as inputs in a feed-back loop.
Figure 1.
General overview of the functional organization of the circadian system in mammals. Inputs: environmental periodical cues can reset the phase of the central pacemaker so that the period and phase of circadian rhythms coincide with the timing of the external cues; Central pacemakers: the suprachiasmatic nuclei (SCN) is considered to be the major pacemaker of the circadian system, driving circadian rhythmicity in other brain areas and peripheral tissues by sending them neural and humoral signals (such as melatonin, secreted by the pineal gland (P)). The SCN receives light-dark cycle information through the retinohypothalamic tract (RHT). Peripheral oscillators: most peripheral tissues and organs contain circadian oscillators. Usually, they are under the control of the SCN; however, under some circumstances (e.g., restricted feeding, jet lag and shift work), they can desynchronize from the SCN; Outputs: central pacemakers and peripheral oscillators are responsible for the daily rhythmicity observed in most physiological and behavioral functions. Some of these overt rhythms (physical exercise, core temperature, sleep-wake cycle and feeding time), in turn, provide feedback, which can modify the function of the SCN and peripheral oscillators, (redrawn from [
11]).
Figure 1.
General overview of the functional organization of the circadian system in mammals. Inputs: environmental periodical cues can reset the phase of the central pacemaker so that the period and phase of circadian rhythms coincide with the timing of the external cues; Central pacemakers: the suprachiasmatic nuclei (SCN) is considered to be the major pacemaker of the circadian system, driving circadian rhythmicity in other brain areas and peripheral tissues by sending them neural and humoral signals (such as melatonin, secreted by the pineal gland (P)). The SCN receives light-dark cycle information through the retinohypothalamic tract (RHT). Peripheral oscillators: most peripheral tissues and organs contain circadian oscillators. Usually, they are under the control of the SCN; however, under some circumstances (e.g., restricted feeding, jet lag and shift work), they can desynchronize from the SCN; Outputs: central pacemakers and peripheral oscillators are responsible for the daily rhythmicity observed in most physiological and behavioral functions. Some of these overt rhythms (physical exercise, core temperature, sleep-wake cycle and feeding time), in turn, provide feedback, which can modify the function of the SCN and peripheral oscillators, (redrawn from [
11]).

Figure 2.
Molecular clock of mammals. Circadian locomotor output cycles kaput (CLOCK)/brain and muscle aryl hydrocarbon receptor nuclear translocator-like (BMAL1) heterodimers (red and green ovals) bind the DNA of clock target genes at E-boxes or E’-boxes and permit their transcription. The resulting period (PER) and cryptochrome (CRY) proteins (blue and yellow) dimerize in the cytoplasm and translocate to the nucleus where they inhibit CLOCK/BMAL1 proteins from initiating further transcription (redrawn from [
16]).
Figure 2.
Molecular clock of mammals. Circadian locomotor output cycles kaput (CLOCK)/brain and muscle aryl hydrocarbon receptor nuclear translocator-like (BMAL1) heterodimers (red and green ovals) bind the DNA of clock target genes at E-boxes or E’-boxes and permit their transcription. The resulting period (PER) and cryptochrome (CRY) proteins (blue and yellow) dimerize in the cytoplasm and translocate to the nucleus where they inhibit CLOCK/BMAL1 proteins from initiating further transcription (redrawn from [
16]).
Neuroanatomical and functional studies point to the existence of nervous pathways between the SCN and heart, pancreas, liver, thyroid and pineal gland [
41,
44,
45,
46]. Using such means of communication, the SCN csan activate or silence different tissues, depending on their function at different times of the day. Thus, the circadian system functions as an orchestra, in which the SCN acts as the conductor and the peripheral oscillators are the different instrumental groups.
The cycle of sunrise and sunset has provided a reliable time cue for many thousands of years, until recently when modern life and the “24-hour society” intensified exposure to artificial lighting environments, both during the day and at night, as people engage in shift work and leisure time is displaced towards the nighttime hours. Thus, it is important to find a way to illuminate the night that permits the circadian entrainment and respects the melatonin rhythm. Reducing the blue component of nocturnal light could prevent light-induced disruption of the circadian system and provide an attractive means of reducing the health risks induced by LAN.
5. Output Pathways: Melatonin
The SCN steers numerous rhythmic functions via a number of neuronal output pathways to hypothalamic and thalamic nuclei and structures. Among these, the paraventricular nucleus of the hypothalamus (PVN) is the first relay station towards the pineal gland. This neuronal pathway is extended even further, via the intermediolateral column of the upper thoracic cord to the superior cervical ganglion, from which postganglionic sympathetic fibers innervate the pineal and control melatonin synthesis through β- and α
1-adrenergic stimulation. Further details and modulation by other neurotransmitters (PACAP, VIP, NPY and glutamate) have been summarized elsewhere [
199]. However, other efferences from the SCN also exist, e.g., to the preoptic area, lateral septum, bed nucleus of stria terminalis, subventricular zone, arcuate nucleus, paraventricular nucleus of the thalamus and, perhaps, the amygdala, habenula and intergeniculate leaflet [
200]. In functional terms, the importance of the pathway via the PVN to rhythmic functions lies beyond pineal activity, in particular, because it controls both parasympathetic and sympathetic projections and, additionally, influences the glucocorticoid rhythm, which is otherwise generated by an autonomous adrenocortical clock. However, glucocorticoid secretion can, in turn, be modulated by melatonin [
201,
202]. The absence of robust adrenocortical rhythmicity in melatonin-deficient C57BL mice [
203] may be taken as an additional indication of the need for melatonin in the glucocorticoid rhythm, but this would need direct experimental support. Another SCN output of relevance to rhythmic organization and chronodisruption controls the hypothalamic sleep switch, a structure generating on-off responses on the basis of mutual inhibition. It alternately activates either wake-related downstream neuronal pathways that involve the locus coeruleus, dorsal raphe nucleus and tuberomammillary nucleus, or sleep-related pathways via the ventrolateral preoptic nucleus [
204,
205]. Again, this output function is influenced by melatonin via its feedback to the SCN. By suppressing neuronal firing through MT
1-dependent signalling, the wake-related neuronal downstream pathways are inhibited, whereas the sleep-related ones are activated. This represents a major contribution of melatonin to the promotion of sleep initiation, although additional actions of the methoxyindole are also involved (for further details, see [
206]).
The control of melatonin formation and secretion is one of the major output functions of the SCN. In terms of rhythm coordination and, on the negative side, chronodisruption, the transmission of light information via the SCN is important to the pineal gland in two aspects. Firstly, the rhythmicity of the SCN, which is entrained by the optic information received from the retina, is imposed on the pineal gland, and thus generates the melatonin rhythm. Although the oscillation of the basically rate-limiting melatonin-synthesizing enzyme, aralkylamine
N-acetyltransferase (AANAT), is mainly driven by the SCN, rhythmic noradrenergic stimulation finds its counterpart in an endogenous pineal clock, which exhibits cycles in the expression of core oscillator genes [
207,
208]. In mammals, the function of this pineal clock may be a periodic facilitation of responsiveness rather than the autonomous generation of the melatonin rhythm. Secondly, and of importance with regard to chronodisruption, nocturnal melatonin can be suppressed by LAN [
150,
209,
210,
211,
212,
213,
214], an action that occurs in addition to and independent of the phase shifting effects. While resetting depends on the phase response curve and, therefore, varies within the circadian cycle and also in the course of the scotophase, the so-called photic shutoff can take place at any timepoint within the scotophase and depends only on light intensity, duration of light exposure and spectral quality, but not substantially on the circadian phase. This action is particularly pronounced if the spectral composition allows perception by melanopsin,
i.e., in the range of 460–480 nm. The photic shutoff causes a rapid cessation of melatonin biosynthesis, and a similarly rapid drop of pineal melatonin concentration and release. Notably, it is observed in animals in which the melatonin rhythm is generated by different mechanisms, either by rhythmic transcription of the
Aanat gene, as in rodents, or by AANAT phosphorylation and stabilization of pAANAT by a 14-3-3 protein, as in primates and ungulates (details in [
199]).
The actions of melatonin are manifold. From a chronobiological point of view, an important effect is the feedback to the SCN [
215]. This feedback is the basis for the chronobiotic actions of melatonin,
i.e., its ability to reset the circadian clock in the SCN. As with all time cues that reset an oscillator, the extent and direction of the phase shifts depend on stimulus timing, according to a phase response curve (PRC). Although the human PRC for melatonin has been determined [
216,
217], in practice, the precise PRC of an individual is usually unkwown and varies according to the chronotype and previous illumination schedules. The phase position is, therefore, usually assessed by determining the dim-light melatonin onset (DLMO) [
218,
219]. The chronobiotic properties of melatonin are the basis for all treatments using the pineal hormone or synthetic melatonergic agonists aimed at readjusting the circadian phases, e.g., after time shifts (jet lag or maladapted shift work) or because of poor entrainment in circadian rhythm sleep disorders (advanced or delayed sleep phase syndromes), in blind people and in circadian-related mood disorders [
220,
221]. However, it should be emphasized that phase readjustments can be achieved by light exposure. Moreover, combinations of melatonin and light have already been used for this purpose [
222,
223]. In addition to the feedback to the SCN, melatonin may also entrain or modulate some peripheral circadian clocks, an assumption for which several indications exist [
18]. If supported by further data, this will lead to important implications concerning the maintenance of favorable phase relationships between centrally-driven oscillations and autonomous or semi-autonomous peripheral clocks.
In addition to its chronobiotic effects, numerous other actions of melatonin have been described. With regard to its circadian periodicity, they may represent, in respective target cells, the nocturnal up- or downregulation of gene expression, release of humoral factors, neuronal activities and other physiological functions. A detailed description would greatly exceed the scope of this review, but a comprehensive overview can be found in reference [
224]. However, it is important to be aware that the complexity of functions is only partially related to the dynamics of the circulating hormone, with its prominent nocturnal peak. Much higher levels of melatonin are found in the tissues of vegetative organs, in particular the gastrointestinal tract, in which the day/night differences are considerably smaller than in the pineal gland or the blood. Among the various extrapineal sites of melatonin formation, immune cells and bone marrow can be specifically mentioned. For detailed information, see reference [
224].
A quantitatively important area of melatonin research has been that of antioxidative protection. Again, this is only partially a matter of circadian rhythms and has often been studied under conditions that go far beyond chronobiology, e.g., in fighting sepsis or other forms of high-grade inflammation, in the attenuation of oxidotoxicity by chemicals or insults such as ischemia/reperfusion or brain trauma, and, more recently, in the induction of apoptosis in tumor cells. This interesting field of actions cannot be discussed in detail within the scope of this article. However, it should be briefly mentioned that melatonin displays multiple properties that counteract or even prevent oxidative damage. Apart from direct interactions with reactive oxygen species (ROS) and reactive nitrogen species (RNS), melatonin upregulates several antioxidant enzymes and downregulates inducible and neuronal NO synthases (iNOS, nNOS). While direct radical scavenging requires elevated melatonin concentrations present in some melatonin-synthesizing tissues and, perhaps, in organelles accumulating melatonin as reported for mitochondria [
225,
226,
227], the modulation of gene expression as mediated by MT1 and MT2 receptors can be observed at physiological levels. These effects counteract microglia activation and peroxynitrite formation and support mitochondrial electron flux, thereby reducing electron leakage and thus freeing radical generation, often preventing oxidant-induced apoptosis in nontumor cells [
224,
227,
228,
229,
230,
231]. Without discussing the details of antioxidative protection, emphasis should be placed on an important consequence with regard to chronodisruption and especially to the light-induced melatonin shutoff, as occurs in LAN during shift work. As melatonin exerts numerous antioxidant actions, a suppression of melatonin by LAN should be expected to decrease antioxidative mechanisms inasmuch as they depend on physiological nocturnal melatonin levels [
232]. This assumption would be in line with the repeatedly observed increases in lipid peroxidation and decreases in glutathione peroxidase and superoxide dismutase in pinealectomized animals (e.g., [
233,
234,
235,
236]). Such a direct reduction of protective capacity must be distinguished from perturbed rhythms in protective enzymes and mitochondrial activity, which are induced by phase-shifting light signals that occur in inappropriate circadian phases and represent another aspect of chronodisruption. However, in the practice of clinical or epidemiological studies, the two undesired changes by LAN, namely dysphased rhythms and melatonin shutoff, have been rarely analyzed.
The effect of light timing on the phase resetting response is also described by a PRC [
237,
238]. Although comparisons among reported PRCs are difficult to make, due to differences in methodology, they are generally consistent and show that light exposure in the early biological night (from DLMO timing to the minimum core body temperature timing) induces a phase delay of the circadian pacemaker, whereas light exposure in the late biological night/early morning (from minimum core body temperature timing to 8 h later) induces a phase advance. The rest of the day, light exposure does not induce any phase shift, defining the so-called “dead zone”. Most of the PRCs report phase delays of less than 12 h (type 1, or weak,
versus 0 or strong, according to the average slope of the plot of the initial and final circadian phase following a resetting stimulus) [
239,
240,
241,
242]. Other studies [
243,
244] showed type 0 resetting and accompanying amplitude suppression following three consecutive days with light pulses. While most of light PRC studies have been performed with white light, a recent study evaluated the PRC obtained under blue light exposure (480 nm), reporting similar results [
245].
Chronodisruption, including melatonin shutoff, has been assumed to play an important role in a variety of health problems. However, it is necessary to distinguish among the different disorders or diseases and their mechanistic causes, and among the methods used for demonstrating the relationship to chronodisruption. This becomes particularly evident in the numerous epidemiologic studies, in which health problems related to shift work have been evaluated. Epidemiology is frequently confronted with the general problem of heterogeneity, and this is very much the case in shift work [
246]. This not only concerns differences in lifestyle habits and nutrition and the type of work (including exposure to other unhealthy factors), but also the various forms of shift work (e.g., length and direction of rotating shifts), the duration of employment under shift schedules, and periods after the end of shift work. This latter point is relevant insofar as health problems progressively emerge with age and may become evident only after the periods of shift work have already ended.
Despite these methodological difficulties, the association of shift work with health problems has been demonstrated in a number of studies. This is most evident in the complex of metabolic syndrome, cardiovascular diseases (especially coronary heart disease [
247,
248,
249,
250,
251,
252]), and diabetes type 2 [
253,
254,
255,
256]. The mere demonstration of such associations is, however, not entirely sufficient. It is also important to know whether shift work induces health problems or aggravates disorders. This latter aspect has been addressed in a study by Lin
et al. (2009) [
257], in which a pre-existing metabolic syndrome was shown to be aggravated by shift work. Another point concerns the changes in eating habits induced by LAN or work at night. In fact, altered food intake and obesity were shown to be induced by shift work and to be associated with elevated blood pressure [
258,
259,
260,
261]. These changes were even observed under conditions of a fixed shift work schedule [
260]. Therefore, the health problems arising from shift work and LAN might be assumed to be indirectly caused by an altered food intake. Another indirect influence may arise from sleep disturbances. Sleep deficits and interruptions are also known to be associated with changes in eating behavior and obesity [
262,
263,
264,
265]. Of course, an increase in body mass index related to eating at night also entails the aspect of circadian rhythmicity in nutrient uptake and metabolism, but the mechanistic relationships are not that simple, because of the demonstrable association between sleep debt and obesity. Nevertheless, although food intake and metabolic utilization are influenced by sleep loss, a recent study has shed light on the importance of circadian misalignment on insulin resistance. Authors showed that insulin resistance is promoted by circadian perturbance, under conditions of controlled sleep loss [
254]. Although sleep restriction also promoted insulin resistance, the effect was considerably higher under circadian misalignment. Importantly, the change in insulin sensitivity was associated with increases in inflammatory markers, a finding that is also relevant to numerous other diseases and merits further investigation on a broader scale. Moreover, it is indicative of a considerable influence of chronodisruption beyond rather simple relationships to the amount of food and changes in body mass. It would be of interest to elucidate the role of melatonin deficits in this context.
With regard to both melatonin and other hormones or humoral factors that undergo circadian cycling, more information is required concerning LAN-induced metabolic changes. Evidence is currently accumulating that shift work alters the plasma levels of resistin, ghrelin, leptin and adiponectin [
247]. These findings are not only of interest with regard to the regulation of food intake and nutrient metabolism, but also in terms of inflammation and atherosclerosis. In particular, the leukocyte-derived factor resistin has been discussed as a mediator of cardiovascular risk in rotating shift work [
266]. Again, the question remains to what extent circadian misalignment is decisive and the potential contribution of melatonin shutoff by LAN.
It should be briefly mentioned that LAN is not only a matter of shift work, rather it must be considered as a contributing factor in gerontological problems. This has been recently addressed in a study of an elderly population [
267]. Moreover, an association between midlife insomnia and mortality has been reported [
268], which may also be of relevance to aged subjects.
As mentioned above, melatonin is known to be a potent antioxidant. One of the predictable consequences of a nocturnal melatonin shutoff by LAN is, therefore, an increase in oxidative damage to biomolecules. In addition, circadian perturbations by mutations in clock genes or repeated phase shifts have also been shown to increase oxidative damage [
269] and to reduce lifespan in animals [
270,
271]. In light of these relationships, it is surprising to observe that the connection between shift work and oxidants has been rarely studied. Two investigations have demonstrated increases in 8-hydroxydeoxyguanosine in the DNA of shift workers [
272,
273]. In the future, this aspect should be more extensively investigated, and also with regard to its consequences in numerous other diseases. Moreover, it will be necessary to clarify the contribution of inflammatory responses to LAN-induced damage in biomolecules.
Another area in which inflammation and mutations induced by oxidative stress are of particular importance is cancer. The possible association between shift work and cancer has been vividly discussed during the last years, especially after a respective classification by the International Agency for Research on Cancer in 2007 (
cf. [
274]). However, this relationship to cancer is not nearly as clear as initially thought. In various types of cancer that had been suspected to be promoted by shift work, epidemiology failed to support this assumption. The best documented association with shift work concerns breast cancer, but despite a remarkable degree of variability among studies and a final conclusion that this relationship is demonstrable, it fails to represent a major risk factor [
275,
276,
277,
278,
279,
280,
281,
282,
283]. Other cancers with a high likelihood of being convincingly related to shift work are colorectal cancer [
72,
284], ovarian cancer [
285] and non-Hodgkin lymphoma [
286,
287]. Apart from the general problems related to the heterogeneity of epidemiology, a major problem of the respective cancer studies is the lack of a mechanistic explanation. Disturbed or misaligned circadian rhythmicity is frequently mentioned, but the reasons for the promotion of cancer have remained unclear. Moreover, the necessity of distinguishing between perturbed/shifted circadian rhythms and melatonin shutdown has not frequently been seen. On the other hand, both chronobiology and melatonin physiology offer manifold possible nexuses to cancer, such as mutations in clock genes that make an organism cancer-prone, cancer-related alterations of melatonergic signalling (summarized in [
18]), and proinflammatory effects of melatonin [
28,
288]. Additional hints can be obtained from preclinical studies in animal models. For instance, the finding that LAN favors resistance to an anticancer drug, tamoxifen [
289], must have a mechanistic explanation that might be relevant to the development of cancer in humans. Additional animal data from experimentally well-controlled studies showing that chronodisruption promotes tumorigenesis, along with reductions in life span [
290], may be more convincing to researchers of the connections between cancer and disturbed circadian and melatonin physiology than poorly controlled epidemiological studies affected by many confounding factors. Although the relevance in humans must be ultimately demonstrated, the mechanistic elucidation may be easier in animals.
A further future demand has to be the monitoring of chronodisruption in humans. To a certain extent, this may be done by ambulatory circadian monitoring (ACM), on the basis of wrist actigraphy, thermometry or body position measurements, for example [
291,
292]. However, all such data must be interpreted with due caution, since chronodisruption not only leads to changes in the oscillators, it also has negative and positive masking effects, which must be identified and, to the extent possible, removed to yield a purified time pattern. With regard to melatonin, determinations of dim light melatonin onset (DLMO) have been applied in sleep research, especially concerning circadian rhythm sleep disorders [
219], and most recently using a DLMO “hockey-stick” variant [
218], and have been compared to ACM data [
292]. Although this can also provide valuable data on circadian changes, the other aspect of melatonin shutoff is not sufficiently accessible using this method. If the loss of melatonin by LAN should turn out to be more important than circadian perturbations, the only way to monitor it in a meaningful manner would be through the determination of melatonin. This can be most easily done using salivary melatonin, but it requires the exclusion of confounding factors, such as melatonin-containing food and beverages, especially coffee.
7. Circadian Healthy Light
Scheduled bright light exposure is an effective countermeasure for sleepiness and fatigue in shift workers and those suffering from jet lag or delayed and advanced sleep phase syndrome. Moreover, the beneficial effects of light on mood, sleep quality, and/or cognitive performance have been found in quite different pathological conditions, including Parkinson’s and Alzheimer’s diseases.
Although circadian entrainment in humans can be attained with lower light intensities [
342], lighting levels of at least 1000 lux at eye level have been proposed to be necessary [
343,
344]; however, this level of lighting is not available in most offices and industrial areas [
343,
344]. Working indoors during the day implies light intensities of 40–200 times lower than being outdoors. Since the maximal human circadian spectral sensitivity occurs at 460–480 nm, diurnal lighting should not be poor in this part of the spectrum [
116,
345]. Thus, all the evidence indicates that daytime lighting should not have a low percentage of blue wavelengths and should present greater intensities than those that are usually found. However, not only intensity and spectrum are important in order to obtain healthy day lighting; glare and spatial distribution also need to be considered. The use of devices capable of modulating light intensity throughout the day in a way that mimics the Sun has also been recently proposed [
346].
As reviewed earlier, light at night can have negative effects on circadian rhythms and on health in general [
116], especially if it is enriched with wavelengths from 460 to 480 nm [
347]. Thus, the luminous energy associated with light radiation, especially that from the short wavelength part of the visible spectrum (400–500 nm), can cause toxic effects in the eye [
348]. Short wavelength light can penetrate the cells and their organelles, inducing the generation of reactive oxygen species (ROS) in retinal pigment epithelium mitochondria and even apoptosis, potentially caused by ROS-damaged mitochondrial DNA, as reported in
in vitro studies [
349,
350]. A recent study
in vivo has also demonstrated that white LEDs (with high content in blue light) at domestic lighting levels cause retinal injury in a rat model [
351], although it should be noted that the animal model used is nocturnal and albino, and thus any extrapolation to humans would be inaccurate. On the other hand, it also could be argued that natural sunlight contains greater blue light intensity. However, it should be noted that retinal physiology changes between day and night (see
Section 4.1.2), so it could be hypothesized that blue light at night could entail a greater risk to the retina integrity. Thus, nocturnal lighting should avoid these specifically active wavelengths [
352,
353,
354] due to their negative effects on both the circadian system and the eyes themselves. Technically, spectrum modifications can be achieved through two procedures: (a) by filtering the 460–480 band of the spectrum or (b) by modulating the intensity and spectrum, depending on the time of the day, using light generated by independent red, green and blue LEDs sources, for example.
Light has been also used as a therapy to treat several disorders, including seasonal affective disorders (SAD) and disrupted sleep-wake rhythms [
355,
356,
357]. The idea is to increase the
zeitgeber strength by enhancing the light exposure during the day, and thus the contrast between the day and the night. The most critical phases for circadian light effects are found after dusk or before dawn, so if exposure to artificial light of sufficient intensity occurs at these moments, it will cause phase shifts (a delay or advance) [
358], contributing to the entrainment of the circadian pacemaker to the 24-h light-dark cycle. Recently, a variation of this therapy has been developed, consisting of simulating dawn and dusk (Dawn-Dusk Simulation, DDS). Since the early pioneering times of Chronobiology, twilight has been known to influence the entrainment of circadian rhythms [
359], in particular, enhancing the coupling to a weak
zeitgeber [
360]. DDS is based on imitating the outdoor twilight and sunrise transitions. With this treatment, a gradual onset of dusk and dawn is adjusted to the patient’s sleep time. DDS can be regarded as a “natural” light therapy because of its lower and gradual changes in light intensity. DDS has been shown to be a successful therapy in the treatment of some psychiatric disorders [
356,
361] and patients who suffer circadian sleep-wake cycle disorders [
356,
362]. Moreover, it is known that DDS induces small advances in the circadian rest-activity rhythm by triggering an earlier onset of the most restful period of the night [
356].
It is clear that although we can design some strategies to create a “melatonin-friendly light”, there would probably be some cases in which protecting the melatonin rhythm would entail extra problems. It is important to be aware that the melatonin rhythm is also affected by parameters other than cycling light intensity. As already mentioned in the section on aging, impaired function of the SCN or of signal transmission to the pineal gland can reduce nocturnal melatonin secretion during senescence and, to an even greater extent, in patients with Alzheimerʼs disease (AD) and other forms of dementia [
316,
363,
364]. Although the causes of dysfunction are not necessarily or primarily a matter of light perception, light therapies have been tested in AD patients, (for examples, see references [
316,
363,
364,
365,
366,
367,
368]). In general, the improvements reported in these studies were relatively modest. Although changes in the proportion of daytime/nighttime activities were observed, in addition to some behavioral benefits [
365,
369], improvements of sleep and circadian rhythmicity usually remained marginal. The efficacy of bright light therapy on sleep consolidation and effects on the circadian system depended on the progression of the disease [
366,
370]. While improvements were demonstrable in earlier stages, this was decreased in the case of advanced AD. At the melatonin rhythm level, light therapy remained relatively inefficient in late AD. In particular, reductions of daytime melatonin were not achieved, contrary to findings in patients with other psychiatric disorders [
363]. These observations are supported by more recent data, which indicate, however, a relatively early onset of SCN and pineal dysfunctions [
371]. Reductions in pineal melatonin secretion, as well as disrupted clock gene oscillations in the pineal, were already demonstrable in Braak stages I–II [
371]. The idea that a combination of bright light therapy with melatonin administration may be helpful, which has received some clinical support [
372], may be critically viewed with regard to extreme reductions in the expression of melatonin receptor MT1 in the SCN of AD patients [
371]. Nevertheless, interindividual differences may exist, and melatonin may be beneficial in SCN-independent or only partially dependent functions. Therefore, the use of melatonin should not be precociously ruled out, in the absence of additional studies that consider these possibilities.
Reduced melatonin levels have been also observed in various other diseases and disorders (summarized in [
364,
373]). These include various cases in which there is no reason to assume dysfunction of the SCN. In particular, such reductions have been found to accompany several stressful or painful conditions, such as Menièreʼs disease [
374], fibromyalgia and neuralgia [
375], migraines [
376,
377], heart diseases [
378,
379,
380,
381,
382,
383,
384], critical illness [
109,
385,
386] and cases of cancer [
387,
388], in which the contribution of stress and pain has remained unclear, as well as in some metabolic diseases, such as acute intermittent porphyria [
389,
390], and notably, diabetes type 2 [
391,
392]. In some neurological disorders, decreases in melatonin are only observed in subpopulations of affected individuals or in a very limited number of subjects studied [
364]. As damage to the SCN does not seem to be causal to the reduced melatonin levels, one might be inclined to assume that increases in
zeitgeber strength (e.g., higher light intensities or an enhanced proportion of blue light) might be able to correct the circadian deficits affecting the melatonin rhythm. This may be possible in less severely affected patients, in which the melatonin rhythm would return to normal anyway after the end of the stress- or painful conditions. However, this has not been sufficiently studied. In the case of critical illness in which the patient receives intensive care, attempts at correcting the melatonin level by a light-dark cycle have been unsuccessful [
109]. Some skepticism may be also due in advanced diabetes type 2, at least as far as the disease has already led to a neuropathy (
cf. ref. [
391]). With regard to the more recently discovered connection between diabetes type 2 and AD in terms of insulin resistance in both vegetative organs and the brain [
393,
394,
395,
396,
397], it is still unclear the extent to which resistance to insulin may already affect neuronal functions in patients who have an otherwise asymptomatic neurological condition.
With regard to pathological deviations, the usefulness of strategies to support the melatonin rhythm must be judged according to a different barometer. In all disorders in which circadian malfunction or poor entrainment are causal, i.e., in circadian rhythm sleep disorders and circadian-related forms of depression, such as borderline personality disorder (BP) or SAD; enhancements of zeitgeber strength by means of bright light, more intense blue light and, eventually, twilight phases around dawn and dusk are promising. Additional medication with melatonin or synthetic melatonergic drugs in the evening may further enhance success. In other diseases discussed in the previous paragraph, the chances of improvement are either limited, low or have not been tested.
Difficulties in readjusting or restoring the melatonin rhythm are of an entirely different nature in the case of shift work. Although it should be possible to strongly reduce the blue and green fraction of the light spectrum without too great of a reduction in overall light intensity, the question remains to what extent the working capacity of an individual is affected by a light quality that only moderately reduces melatonin [
398]. The maintenance of a high nocturnal melatonin level may well reduce the alertness of a worker and cause undesired safety problems. This has to be weighed against more convenient durations and sequences of shift periods.
As already mentioned, in the developed countries, the usage of smartphones, tablets and other electronic devices has been widespread over the last few years. Some applications have been recently developed to reduce the negative effects of their use at night. In general, they work by adjusting the display color temperature according to the natural light-dark cycle. Thus, they avoid high color temperatures after sunset, while permitting them during the day.