REVIEW OF THE LITERATURE

The French astronomer Jean Jacques d'Ortous de Mairan discovered in 1729 that when a Mimosa plant was placed in the constant darkness of a closet, the opening and closing of the leaves still occurred on the basis of an intrinsic rhythm (de Mairan 1729, referred by Meijer & Rietveld 1989). Thus, he pointed out for the first time the independence of biological rhythms of the light-dark cycle.

All organisms generally have the capacity for endogenous temporal organization of biological processes over the course of day. The cellular, neural and humoral machinery that generates this ability is collectively known as the biological clock. This timing device functions as a pacemaker of a complex oscillation network. The endogenous period of the clock is not exactly 24 hours. Thus, the clock generally requires daily resetting by external time cues. This process is called entrainment.

The output of the pacemaker is called circadian rhythms. These rhythms have two principal properties: they are normally entrained to external time cues, but in the absence of synchronizing signals, they free-run with a period slightly different from 24 hours (Aschoff 1965a). The light-dark (LD) cycle of a 24-hour solar day is the main environmental cue entraining the clock and the rhythms driven by it.

The existence of a circadian or daily rhythm is apparent in physiological, pharmacological and pathological events and parameters. These include physical activity, body temperature, plasma levels of hormones (e.g. cortisol, renin, angiotensin, aldosterone, noradrenaline, insulin, prolactin, growth hormone, thyrotropin, atrial natriuretic peptide, vasopressin), response to a glucose-tolerance test, performance variables (e.g. reaction time, reading error, subjective alertness, working memory speed, self-chosen work-rate), birth and death rate, pharmacokinetics and effects of drugs, and risk for cardio- and cerebrovascular attacks (Wever 1979, Reilly et al. 1997, Schwartz 1997, Lemmer 1999). Moreover, various functions of the heart and circulation (e.g. cardiac output, heart rate, blood pressure), of the respiratory system (e.g. minute volume, oxygen consumption, carbon dioxide production) and of the kidneys (e.g. glomerular filtration rate, urine flow rate, electrolyte excretion) can vary with the time of day (Wever 1979, Reilly et al. 1997).

In addition to the action of light on the period of the circadian rhythms, light exposures are able to affect the phase of the rhythms. There are three possible effects: no shift, delay, or advance. The magnitude and direction of the phase shift depend on the phase of the rhythm at the time of stimulus application. This relationship can be illustrated by a phase-response curve (Aschoff 1965b).

Temporal regulation of behavioral and internal homeostatic events is a fundamental feature of mammalian adaptation. It helps to optimize the economy of biological systems, and prepares an organism to foresee and cope with the 24-hour solar day resulting from the rotation of our planet. Although normally the biological clock and the rhythms driven by it are entrained by environmental time cues, in certain situations the rhythms are not synchronized with the time cues (external desynchronization) or with each other (internal desynchronization). For example, time zone transitions, night work or blindness may cause deficient entrainment and desynchronized circadian rhythms, namely circadian disorders.

In vertebrates, the master biological clock system is located in the hypothalamic suprachiasmatic nucleus (SCN). The paired nucleus consists of thousands of neurons, for instance in rats of about 8000 neurons occupying 0.07 mm³ and in humans of about 10 000 neurons occupying 0.25 mm³ (van den Pol 1980, Swaab et al. 1993, Hastings 1997). In humans, the SCN is situated on either side of the brain midline on the top of the optic chiasma and about 3 cm behind the eyes in the basal part of the anterior hypothalamus. The SCN is involved in the generation and expression of physiological functions with circadian properties. These include e.g. water and food intake, motor activity, sleep-wake rhythm, corticosterone release, activity of pineal N-acetyltransferase enzyme, and body temperature (reviewed by Rusak & Zucker 1979). In addition, lesions of the SCN abolish the nocturnal increase of melatonin synthesis (Reppert et al. 1981, Lehman et al. 1984, Kalsbeek et al. 1996).

The circadian timing system consists of three essential components: entrainment pathways (inputs), oscillation mechanisms generating the rhytmicity (pacemaker) and efferent pathways mediating the expression of circadian rhythms (output).

The best-described entrainment pathways to the SCN originate in retinal ganglion cells. The retinohypothalamic tract (RHT) is the main route responsible for photoentrainment (Hendricson et al. 1972, Moore & Lenn 1972). The RHT originates from a distinct and homogenous group of ganglion cells (Moore et al. 1995). It leaves the optic nerve mainly in the anterior part of the optic chiasm and terminates predominantly within the ventrolateral part of the SCN (Levine et al. 1991, Speh & Moore 1993, Reuss 1996). It has been shown that the RHT has several components; the largest projects to the SCN, but there are additional projections to the lateral and anterior hypothalamic area, and to the retrochiasmatic area which is a complex zone immediately caudal to the SCN (Johnson et al. 1988b). Glutamate appears to be the main transmitter of the RHT (de Vries et al. 1993, Shirakawa & Moore 1994, Mintz et al. 1999).

The retinal pathway responsible for the projection of the circadian system to the SCN branches off the projection to the visual centers of the brain (Klein & Moore 1979, Moore et al. 1995). In fact, the section of visual pathways beyond the RHT does not affect stable entrainment (Klein & Moore 1979), whereas ablation of the RHT (Johnson et al. 1988a) or blinding (Wurtman et al. 1964, Klein & Weller 1970, Klein & Weller 1972, Nelson & Zucker 1981, Lucas & Foster 1999) results in the loss of the ability of light to entrain circadian rhythmicity.

In addition to the RHT, the SCN also receives input from the intergeniculate leaflet (IGL), raphe nuclei, the paraventricular thalamus and the limbic telencephalon (reviewed by Moga & Moore 1997). The pathway via the IGL (Pickard 1985) of the lateral geniculate complex of the thalamus through the geniculohypothalamic tract (Rusak et al. 1993, Moore & Card 1994) to the SCN is suggested to modulate photic and nonphotic information to the circadian system (Johnson et al. 1989, Moore & Card 1994, Moga & Moore 1997). In this tract, gamma-aminobutyric acid, neuropeptide Y and enkephalin are likely candidates for being the chemical transmitters (Moore & Speh 1993, Moga & Moore 1997). Midbrain raphe nuclei also project to the SCN (Rusak et al. 1993, Miller et al. 1996) and serotonergic input from these nuclei is suggested to play a role in the circadian system, but not to be primarily involved in the entrainment to the LD cycle (Meijer & Rietveld 1989, Meyer-Bernstein & Morin 1999). The spesific functions of the thalamus, the limbic region and other unexplored tracts in the circadian timing machine await clarifying.

The late Curt Richter deserves the credit for pinpointing the hypothalamus as the home of the biological clock, even though he did not specifically identify the SCN. Richter (1967) investigated the function of the hypothalamus by making lesions, and he discovered that lesions in the region of the ventral median nucleus eliminated eating and drinking rhythms in rats.

The SCN seems not to be just a simple hourglass timer directly driven by external signals. In addition to the SCN being the major site for the RHT input, there are five lines of evidence indicating that the SCN acts as a circadian pacemaker. First, ablation of the SCN eliminates circadian functioning as shown originally by two teams of investigators working independently. Moore and Eichler (1972) found that SCN lesions destroyed circadian adrenal corticosterone rhythms in the rat, and that in the absence of the SCN not even the preservation of the optic tracts was sufficient to maintain synchronization. In the same year, Stephan and Zucker (1972) demonstrated that electrolytic lesions in the SCN permanently eliminated circadian rhythms of drinking behavior and locomotor activity of the rats.

The second line of evidence suggesting a pacemaker function for the SCN comes from the studies in which circadian functioning was found to be maintained in isolated SCN, both in vivo (Inouye & Kawamura 1979) and in vitro (Shibata & Moore 1988). In addition, Green and Gillette (1982) showed that single cells, removed from the SCN and put in culture, could persist in their rhythms for up to 60 hours.

Third, electrical stimulation of the SCN in hamsters and rats resulted in phase-dependent shifts of the free-running activity cycle (Rusak & Groos 1982). Phase-shifting effects are also observed after local stimulation of the SCN with a cholinergic agonist, glutamate, or a glutamate agonist (Zatz & Herkenham 1981, Meijer et al. 1988, Mintz & Albers 1997).

The fourth line of evidence for the role of the SCN as a pacemaker is provided by the findings on the circadian rhythm of metabolic activity in the SCN. Indeed, in the absence of periodic environmental light cues, glucose utilization of the nucleus was found to be high during the subjective day and low during the subjective night (Schwartz et al. 1980, Reppert & Schwartz 1984).

Fifth, transplantation of the fetal anterior hypothalamus containing the SCN into the third ventricle of arrythmic, SCN-lesioned animals was demonstrated to restore rhythmicity (Lehman et al. 1987); the period of the restored rhythm being determined by the graft (Ralph et al. 1990).

Mechanisms of the pacemaker in the suprachiasmatic nucleus

There is no point having a watch unless it keeps time. The timekeeping mechanisms underlying the oscillating action of the SCN have been the focus of many recent studies. To date, molecular clock models of similar structure have been proposed for cyanobacteria, fungi and animals, but the key proteins for these models do not share sequence similarity (Dunlap 1999). Thus, it appears likely that the clock systems have developed independently in different organisms, and by alternating environment natural selection has apparently led to the evolvement of clock systems which have only some physiological properties in common (reviewed by Kondo & Ishiura 1999).

The fruit fly Drosophila has autonomous circadian clocks throughout the body (Plautz et al. 1997), suggesting that individual cells are capable of supporting their own independent clocks. In mammals, however, experimental findings suggest that SCN neurons are born as circadian oscillators coupled within the nucleus to form a complex network pacemaker (Moore & Bernstein 1989). Support for this hypothesis is provided by the observations that the SCN contains many autonomous, single-cell circadian oscillators (Welsh et al. 1995), that isolated glands and pieces of tissue can continue to show circadian oscillations in culture (reviewed by Takahashi & Zatz 1982) and that various rhythms of an organism sometimes free-run with different circadian periods during constant environmental conditions (Wever 1979, Turek et al. 1982). In addition, this hypothesis of a multioscillator system is supported by the report of Moore-Ede (1983) as well as by the study of Illnerová and Vanecek (1982), indicating a two-oscillator pacemaking system with an evening oscillator coupled to dusk and a morning oscillator coupled to dawn.

Indeed, the light-induced immediate increase in the multiple unit activity of SCN neurons was found to be independent of the timing of the exposure (Inouye 1984), whereas the shift of the activity of the neurons was found to be depend on the timing of the exposure (Inouye & Kawamura 1982). Thus, while the direct effect of a light stimulus at the input side of the SCN shows no circadian variation, the output of the SCN shows circadian variation, indicating that the state and internal process of the pacemaker dictate the outcome of the circadian event.

In addition, researchers are trying to assign molecular clock components by investigating different genes involved in or capable of affecting the operation of the biological clock. Until 1997, clock researchers had only three clock components to work on: two proteins from a fruit fly (Drosophila) and one from a bread mold (Neurospora) (reviewed by Dunlap 1999). Takahashi and coworkers (Antoch et al. 1997, King et al. 1997) demonstrated the existence of the first mammalian clock gene, Clock, from mice. In addition, common elements of oscillation are likely to be present up and down the evolutionary tree. For instance, in the mouse there are three different gene relatives (Per1, Per2 and Per3) that are related to the Drosophila per gene (reviewed by Dunlap 1999). Transcription of clock genes and synthesis of the proteins they encode form the basis of the timekeeping function. Interactions among these proteins result in feedback inhibition of gene transcription. With degradation of the protein products, gene transcription is again initiated to reestablish the cycle (Dunlap 1999). The major elements of the cycle have been identified, but exact functional roles of the molecular components await revealers.

The function of the different projections of the pacemaker and transmitters communicating circadian information to the rest of the brain is not well understood. In addition, the morphological organization of SCN projections is rather difficult to study because of technical difficulties combined with the very small size of the SCN. The human SCN can be apportioned into five chemoarchitectonic subdivisions of different kinds of neurons (Mai et al. 1991); vasopressin and vasoactive intestinal polypeptide containing neurons are the common ones in the human SCN (Hofman et al. 1996, Dai et al. 1997).

Efferent projections leave the SCN via ventrocaudal, lateral, dorsal and rostral pathways (Stephan et al. 1981) and many of these projections terminate in other hypothalamic nuclei as well as in the thalamus and midbrain (Berk & Finkelstein 1981) in rats. On the basis of immunological studies, there is general agreement that the SCN project to a number sites in the basomedial hypothalamus and the midline thalamus in rodents (Watts et al. 1987, Watts & Swanson 1987, Kalsbeek et al. 1993). In humans, the efferent projections of the SCN appear to be comparable to those described in rats and hamsters (Dai et al. 1998). The human SCN was found to be connected with the nuclei in the hypothalamus that are involved in hormone secretion, cardiovascular regulation and integration of autonomic information. It is suggested that through these projections the SCN may influence brain areas that regulate e.g. thirst, food intake, metabolism, sleep, sexual behavior, and body temperature (reviewed by Dai et al. 1998).

The best-described efferent pathway of the SCN emerges from the pineal gland. This route from the SCN runs via the paraventricular nucleus (PVN) of the hypothalamus, the intermediolateral cell column of the spinal cord, the superior cervical ganglion, and sympathetic efferents to the pineal gland (Moore 1996).

The visual and circadian light detection systems have been shown to differ from each other (Chase et al. 1969, Frost et al. 1979). Eyes provide the primary source of circadian photoreception because eye loss in both humans and other mammals abolishes photoentrainment (Wurtman et al. 1964, Klein & Weller 1970, Klein & Weller 1972, Nelson & Zucker 1981, Lucas & Foster 1999). However, the nature of the photoreceptors that provide input from the retina to the circadian timekeeping system remains elusive.

Originally, orthodox photoreceptor cells, i.e. retinal rods and cones capturing light for vision, were thought to represent also the photoreceptive elements in light-to-circadian-clock transmission. The light sensitivity of the mammalian pineal gland has been demonstrated with single cell recordings to be contributed by both rods and cones (Thiele & Meissl 1987). On the other hand, it was suggested that a rhodopsin-like pigment mediate circadian vision, because both the spectral sensitivity of the phase-shifting effect and the suppression of pineal N-acetyltransferase enzyme by light resemble the absorption spectrum of the pigment (Takahashi et al. 1984, Bronstein et al. 1987). These studies are in line with the study on the wavelength sensitivity of melatonin suppression in humans (Brainard et al. 1985) reporting that rods arbitrate the effects of light in the mammalian circadian timing system.

More recently, studies in strains of mice with hereditary retinal degeneration suggest that the hypothalamic effects of light might be mediated rather by the remaining fragments of cone photoreceptors than by rods (Provencio et al. 1994, Lucas & Foster 1999). In a study by Rutkowska and colleagues (1998), the rhythm of core temperature in color-deficient subjects was found to be phase-delayed as compared with normal sighted subjects, suggesting a role of retinal cones in the mediation information between environment and the circadian clock. On the other hand, in aged mice, circadian responses to light do not correlate with the number of surviving cones (Provencio et al. 1994), and at least in humans, a normal trichromatic visual system is not necessary for light-mediated melatonin regulation (Ruberg et al. 1996).

It has been suggested that there may be some unidentified retinal light-sensitive cells that send information to the lower brain centers (Foster et al. 1993, Provencio et al. 1994, Huerta et al. 1999, Lucas & Foster 1999). In a strain of retinally degenerate mice with remaining but high-threshold circadian responses to light, the spectral sensitivity of the phase-shifting effect was different from the sensitivity of sighted animals, indicating that various photopigments may be involved in the response (Yoshimura & Ebihara 1996). More recently, it has been shown that neither rods nor cones are required for light-induced melatonin suppression (Lucas et al. 1999) or for photoentrainment of wheel-running activity (Freedman et al. 1999) in mice. In addition, Provencio and coworkers (2000) have identified a novel human opsin which is expressed in cells of the inner retina, but not in retinal photoreceptor cells involved in image formation. The question remains whether there are still uncharacterized photoreceptors in the retina for the transduction of light stimulus to the circadian system.

Besides the studies conducted on opsin/retinal-based photopigments, the possible involvement of cryptochrome blue-light photoreceptors in circadian photic responses has recently become a topic of interest. Cryptochrome proteins are light-sensitive, putative vitamin-B2 based pigments, and cryptochromes 1 and 2 are expressed in the mouse retina and SCN (Miyamoto & Sancar 1998). The observation that the expression of cryptochrome 1 gene exhibits circadian oscillations in the mice SCN (Miyamoto & Sancar 1998) proposes that cryptochromes have a role in circadian photoreception in mammals. Recently, it was shown that the mammalian cryptochrome 1 and 2 are essential for the circadian clockwork (van der Horst et al. 1999). However, the cryptochrome 1 and 2 genes are not essential for the light-induced phase shifting of the clock (Thresher et al. 1998, Okamura et al. 1999).

Extraocular photoreceptors are capable of providing sufficiently light to the circadian timing system in nonmammalian vertebrates (Underwood & Groos 1982, Yoshikawa & Oishi 1998), and some researchers have also demonstrated nonocular photoreception in mammals. For instance, the neonatal rat pineal has been shown to be photosensitive (Blackshaw & Snyder 1997) and extraretinal mechanisms are reported to mediate light-induced changes in the regulation of pineal N-acetyltransferase enzyme in newborn rats blinded by bilateral orbital enucleation (Torres & Lytle 1989). In addition, other investigators have obtained evidence suggesting that light can directly affect hypothalamic neurons in enucleate adult rats (Lisk & Kannwischer 1964).

In 1998, a surprising finding about nonocular phototransduction in humans was reported. The results of this study suggested that a bright light exposure to the back of the knee could phase shift human body temperature and melatonin secretion rhythms without any transmission of light through the eyes (Campbell & Murphy 1998). A photosensitive property of the skin has also been proposed by the finding that the LD alterations synchronize melatonin levels in genetically mutant anophthalmic rats lacking a complete visual system (Jagota et al. 1999).

One possible mechanism for extraocular photoreception has been attributed to chronobiological photoreceptors in blood (Oren & Terman 1998). In fact, Oren (1996, 1997) has hypothesized that heme moieties and bile pigments contained by blood could serve as photoreceptors. This humoral phototransduction model postulates that tetrapyrrole-based pigments, e.g. primary light-sensitive plant pigments of chlorophyll and phytochrome, and mammalian hemoglobin and bilirubin, mediate light-induced circadian effects.

The skin is an interesting candidate for being a photoreceptive element. Two research reports of Iyengar (1994, 1998) have suggested that the melanocyte network in cultured human skin senses light, indicating that cells in the skin might be able to take part in chronobiological events.

In contrast to the above studies on nonocular phototransduction, an earlier study by Nelson and Zucker (1981) demonstrated that the activity rhythms of blinded diurnal ground squirrels and nocturnal grasshopper mice failed to entrain to the LD cycle. In line with this, the study in bilaterally anophthalmic rats showed no evidence of extraocular photoreception (Ibuka 1987). Recent studies by Meijer and coworkers (1999) and Yamazaki and associates (1999) report that illumination of the skin of blinded and shaved hamsters did not result in phase shifting effects on activity rhythms. In addition, light responsiveness of metabolic activity and gene expression of the SCN have been found to be mediated only through the eyes in preterm infant baboons (Hao & Rivkees 1999). Indeed, the absence of nonocular photic regulation of the circadian rhythms in humans has been suggested by our recent study indicating that a bright light exposure on the skin of the abdomen and chest does not induce phase shifting of melatonin, cortisol and thyrotropin rhythms (Lindblom et al. 2000).

Although a phase shifting effect of extraocular light exposure on the melatonin rhythm has been suggested (Campbell & Murphy 1998), there is no evidence of suppression of melatonin levels by skin illumination either in humans (Lockley et al. 1998, Hébert et al. 1999, Lindblom et al. in press) or hamsters (Yamazaki et al. 1999).

It seems that there are still discrepancies in the findings on the photoreceptive elements of circadian events. Several possible mechanisms and pathways could be involved in the effect that light has on the melatonin synthesis and other circadian rhythms. In conclusion, more research is needed before the nature of retinal photoreceptors and the role of extraocular phototransduction in the function of the circadian timing system are solved.

The visible portion of the electromagnetic spectrum covers the wavelength range from 380 to 760 nm, and the eye discriminates between different wavelengths within this range by sensation of color. Light is used to generate a visual image of the environment and to provide time-of-day information. In addition to the timing effects of light on the circadian system, light has also some direct neural effects. For instance, light exposures are able to increase body temperature (Strassman et al. 1991), enhance alertness (Campbell et al. 1995) and suppress melatonin (Brainard et al. 1997).

A temporal structure is needed for circadian timekeeping. The internal circadian clock generally requires daily resetting by external time cues (Wever 1979, Czeisler et al. 1980). Several environmental and behavioral stimuli have been shown to act as circadian synchronizers. These include the timing of food availability, social interaction and physical activity (Stephan 1981, Mrosovsky & Salmon 1987, Mrosovsky et al. 1989, Van Reeth & Turek 1989, Edgar & Dement 1991, Marchant & Mistlberger 1996).

However, in animals the dominant synchronizing signal for circadian rhythmicity is provided by environmental LD cycles. In many species of animals, light plays an important role in the regulation of circadian rhythms, e.g. motor activity, hormone secretion and temperature (McGuire et al. 1973, Wurtman 1975, Elliott 1976). Phase response curves (PRCs) are constructed by exposing the organism to an external signal or by administering a compound at different phases of the endogenous rhythm and measuring the resulting effect on the phase of the cycle (Aschoff 1965b). The first PRC to light was demonstrated by DeCoursey (1960) 40 years ago in flying squirrels. In 1978, Honma and associates (1978) reported the PRC of the locomotor activity rhythm to light pulses in rats which showed that phase delays occurred in the early subjective night followed by phase advances.

The influences of light and darkness on circadian rhythms can also be demonstrated by studies conducted in constant environmental lighting conditions. In constant darkness the rhythms free-run in rats (Redman et al. 1983, Thomas & Armstrong 1988). Continuous light treatment induces suppression of melatonin biosynthesis (Klein & Weller 1970, Laakso et al. 1994a) and the circadian rhythmicity of locomotor activity is lost (Honma & Hiroshige 1978, Chesworth et al. 1987) in rats. Several other circadian rhythms in rats (e.g. behavioral, temperature and some humoral rhythms) may persist for several weeks depending on the intensity of light (Honma & Hiroshige 1978, Eastman & Rechtshaffen 1983, Deprés-Brummer et al. 1995).

Based on current knowledge, the circadian rhythms of humans are also sensitive to light, although earlier findings of temporal isolation experiments proposed that social contacts are more effective than light in the entrainment of human circadian rhythms (Wever 1979). Indeed, as early as in 1960, support for the capacity of light and darkness to synchronize the human circadian system was provided by Sharp (1960) who reported a phase delay in the plasma levels of leucocytes in response to extension of darkness following normal wake time. A similar kind of finding was reported by Orth and Island (1969) and by Osterman (1974), who demonstrated that the circadian rhythm of plasma corticosteroids could be phase shifted by prolongating the dark period of the day.

In addition, more recent studies have shown that the effects of a single light pulse on the phase of human circadian rhythm markers can be observed after only a single day (Honma et al. 1987, Burešová et al. 1991, Minors et al. 1991, Van Cauter 1994) and in primates effective entrainment can be induced by a 1-s light pulse (Sulzman et al. 1981). In fact, the importance of the LD cycle in human circadian resetting was shown in the study of Czeisler and associates (1981) and Middleton and coworkers (1996a) underlining the lesser role of social cues and knowledge of clock time.

In humans, light has been found to be a stronger synchronizer of circadian rhythms than the sleep-wake rhythm. For instance, light-induced phase shifts of body temperature, cortisol, melatonin and sleep propensity rhythms could be seen even when the timing of the sleep-wake cycle was held constant (Czeisler et al. 1986, Drennan et al. 1989, Dijk et al. 1987, Lewy et al. 1987, Dijk et al. 1989). Clodoré and associates (1990) and Foret and coworkers (1993) have reported that the rhythms of cortisol, alertness and performance can be phase advanced by repeated morning exposure to bright light.

The theoretical basis of the PRC to light in humans was postulated by Lewy and collaborators (1983). The first contribution toward the findings of light-induced phase advances and delays were provided by Honma and colleagues (1987) who showed that phase shifts of both sleep-wake and temperature rhythms could be induced by a bright light pulse. In 1989, the PRC to a non-24-hour LD cycle for body temperature was demonstrated (Wever 1989). More recently, Minors and coworkers (1991), Dawson and colleagues (1993) and Van Cauter and associates (1994) established the human PRCs to single bright light pulses for body temperature, melatonin and thyrotropin rhythm (Figure 1). In addition, a strong resetting of the human circadian timing system by multiple light pulses has been demonstrated (Czeisler et al. 1989, Shanahan et al. 1999).

Together, circadian PRCs produced in response to light have been found to share the following time-dependent properties: light stimuli early in the subjective night induce phase delay shifts, light stimuli late in the subjective night induce phase advance shifts, and light stimuli during the subjective day induce no or minimal phase shifts (Figure 1). Dose-response relationships have also been established between light intensity and phase shifting of the human temperature rhythm (Boivin et al. 1996).

Constant environmental lighting conditions also influence the circadian rhythms in nonhuman primates and humans. For instance, in many blind people the circadian rhythms free-run (Miles et al. 1977, Orth et al. 1979, Smith et al. 1981, Lewy & Newsome 1983, Nakagawa et al. 1992, Sack et al. 1992b, Klein et al. 1993, Skene et al. 1999). In the sighted, the period of human melatonin, body temperature and cortisol rhythms has been demonstrated to average 24.18 hours in controlled conditions of low light levels (Czeisler et al. 1999). Continuous light treatment induces suppression of melatonin rhythms in primates (Perlow et al. 1980, Perlow et al. 1981, Tetsuo et al. 1982). On the other hand, constant light conditions have variable disturbing effects on other circadian rhythms. For instance, the daily rhythmicity of cortisol and urinary potassium excretion in humans (Krieger et al. 1969) and cortisol secretion in primates are not altered (Perlow et al. 1981) in constant illumination.

Figure 1. SCHEMATIC HUMAN PHASE-RESPONSE CURVES TO LIGHT AND MELATONIN.

The y-axis of the phase-response curve (PRC) shows the direction and relative magnitude of the phase shift of the body temperature rhythm induced by light exposure (--) and the phase shift of the melatonin rhythm produced by melatonin administration (…) at various times shown on the x-axis as local or circadian times. The circadian time 18 corresponds roughly to the minimum of the body temperature. The black bar indicates a typical time for sleep relative to the minimum of the body temperature when the circadian system is entrained to the 24-hour day. For orientation purposes, sunrise and sunset are drawn in for the time of year when the night and day are of equal length. The PRC to light is about 12 hours out of phase with the PRC to melatonin (modified from Minors et al. 1991, Lewy et al. 1998a, Eastman & Martin 1999).

Over two millennia ago, Herophilos (325-280 B.C.), who was a famous anatomist at the University of Alexandria in Egypt, was probably the first reseacher to discover the pineal organ in humans. Nothing is left of his writings, which have been cited by a Greek physician Galen (±130-300 A.D.). Galen termed the pineal konareion (Latin, conarium) after its pineconelike shape, and described its possible glandular function (reviewed by Kappers 1979).

In the 17th century, the French philosopher and mathematician René Descartes (1596-1650) regarded the pineal organ as the soul's right hand, meaning that the soul exercises its functions to a great degree in the unpaired pineal organ. He likened pineal functions to a valve that regulates the direction and volume of the animal spirits flowing to and from the brain. Thus, he suggested that the pineal initiates the motor stimulus by sending the spirits into the tubular motor nerves (reviewed by Kappers 1979).

In 1917, McCord and Allen (1917) demonstrated that extracts of the pineal gland cause the skin of tadpoles to lighten. A major breakthrough in pineal studies occurred in 1958 when the dermatologist Aaron Lerner and coworkers (1958) extracted bovine pineal glands and were able to isolate a skin lightening compound, and determine its structure (1959) after 4 years of work and the use of more than 250 000 pineal glands (Lerner 1999). The compound was called melatonin (Greek, melas=black and tosos=labor). Axelrod and associates (Axelrod & Weissbach 1960, Weissbach et al. 1960, Weissbach et al. 1961) demonstrated that melatonin could be synthesized within the pineal gland. In later studies, in which it was shown that pinealectomy abolished the circadian rhythm of melatonin, it was confirmed that the pineal gland is the main source of melatonin in rats (Ozaki & Lynch 1976, Lewy et al. 1980a), rhesus monkeys (Tetsuo et al. 1982) and humans (Neuwelt & Lewy 1983, Petterborg et al. 1991).

The pineal gland (Latin, epiphysis cerebri, glandula pinealis) is a part of the brain derived from the caudal portion of the embryonic dorsal diencephalic cell column, the epithalamus. The morphology of the pineal changes dramatically with phylogenetic development. In mammals, the pineal has lost virtually all of its photoreceptor elements and instead contains parenchymal cells whose appearance suggests an entirely secretory function (Korf & Oksche 1986). According to current knowledge, the mammalian pineal (weight, 100 mg in humans, 1 mg in rats; Axelrod 1974) is considered a neuroendocrine organ. In humans, it is located at the posterior wall of the third ventricle near the geometric center of the brain. In addition, different indoleamines, including melatonin, and also several peptides are found in the pineal (Vaughan 1984, Ebels & Balemans 1986).

Previously, a widely held view was that the human pineal is a calcified vestigial organ which provides only a useful landmark for neuroradiologists. According to present knowledge, the physiological significance of the pineal gland in humans is mainly related to the function of melatonin. Melatonin is regulated by the circadian timing system, and LD cycles control the timing of melatonin secretion. Indeed, elevated melatonin levels are associated with nighttime, and melatonin has become to be known as the chemical expression of darkness (Reiter 1991a). In addition to the function of melatonin as a marker of environmental lighting conditions, melatonin may have some other physiological effects e.g. on reproduction and immune system.

Figure 2. THE REGULATION PATHWAY OF THE PINEAL MELATONIN.

Sagittal view of human brain showing the pineal and its innervation. Retinohypothalamic fibers synapse in the suprachiasmatic nuclei (SCN), and there are connections from the SCN to the intermediolateral gray column in the spinal cord. Preganglionic neurons pass from the spinal cord to the superior cervical ganglion (SCG), and the postganglionic neurons project from this ganglion to the pineal in the nervi conarii (modified from Ganong 1997).

Synthesis of melatonin

Melatonin is a small (molecular weight 232.3) indoleamine secreted rhythmically with increased synthesis during the dark period of day. Alterations in environmental lighting are signaled as multisynaptic neural inputs by the central nervous system via periferal nerves into a hormonal output of the pineal gland (Figure 2). In mammals, circadian photoreceptors of the retina convert light and darkness into signals that are sent directly to the SCN through the main pathway, the RHT (Hendrickson et al. 1972, Moore & Lenn 1972).

From the SCN, neuronal projections make synaptic connections in the PVN of the hypothalamus descending onward through the medial forebrain bundle to the intermediolateral cell column of the spinal cord from where preganglionic fibers reach the superior cervical ganglia (Swanson & Cowan 1975, Saper et al. 1976, Rando et al. 1981, Moore 1996). Sympathetic postganglionic noradrenergic fibers from the superior cervical ganglia innervate the pineal gland through the nervi conarii (Kappers 1960). Interruption of this regulation pathway by a lesion of the SCN or PVN, or by superior cervical ganglionectomy abolishes the pineal gland synthesis (Wurtman et al. 1964, Hastings & Herbert 1986, Bittman et al. 1989). Thus, production of pineal melatonin occurs in response to noradrenergic stimulation which produces a cascade of biochemical events within the pinealocyte.

The N-acetyltransferase (NAT) activity represents a key regulatory step in melatonin synthesis. Noradrenaline release from the sympathetic nerves that innervate the pineal gland is normally high at night and low during the day (Brownstein & Axelrod 1974). In most species, noradrenaline interacts with both beta₁- and alpha₁-adrenergic receptors present in the pineal gland (Vanecek et al. 1985). In the rat pinealocyte, stimulation of adrenergic receptors induces a rise in adenylate cyclase, and the cyclic adenosine 3´,5´-monophosphate (cAMP) signaling pathway activates the NAT enzyme that catalyses the rate limiting step of melatonin synthesis (Deguchi & Axelrod 1972, Axelrod 1974, Sugden 1989). Simultaneous activation of alpha₁-receptors potentiates the effects mediated through the beta₁-receptors (Klein et al. 1983). Locally in the pineal gland, the rhythmic melatonin synthesis is ensured by the oscillating cAMP-dependent transcriptional control mechanism (Stehle et al. 1993, Foulkes et al. 1996).

Melatonin's biosynthetical pathway involves tryptophan, one of the 9 essential aminoacids in humans, as a precursor. Tryptophan is hydroxylated and decarboxylated to serotonin, and then serotonin is acetylated by the rate-limiting enzyme NAT and further methylated by hydroxyindole-O-methyltransferase to melatonin (5-methoxy-N-acetyltryptamine) in the pineal gland (Reiter 1991b). After the synthesis, melatonin is secreted into the blood and cerebrospinal fluid (Smith et al. 1976a, Smith et al. 1976b, Arendt et al. 1977). Although melatonin is found to be a very lipophilic compound which readily crosses the blood-brain barrier (Pardridge & Mietus 1980, Le Bars et al. 1991), there is also evidence of high hydrophilicity of the molecule (Shida et al. 1994).

Irrespective of whether a person is asleep or awake in dim light, melatonin is usually secreted between 2100 and 1000 h, with peak levels occurring between 0200 and 0600 h (Laakso et al. 1990, Laakso et al. 1994b). The circulating daytime serum levels of melatonin in healthy adults do not normally exceed 20 ng/l, while the range of the nighttime values may be about 20-170 ng/l (Laakso et al. 1990, Brzezinski 1997).

In fullterm newborn infants, the rhythm of a melatonin urinary metabolite is not apparent but develops between the ninth and twelfth week after birth (Kennaway et al. 1992). The highest nocturnal levels are reached at the age of one to three years (Waldhauser et al. 1993). These peak levels of nighttime blood melatonin levels and the urinary excretion of melatonin metabolites decrease during puberty (Waldhauser et al. 1993, Cavallo & Dolan 1996). After the attainment of sexual maturity, there seems to be a moderate decline of melatonin synthesis until old age (Hartmann et al. 1982, Sack et al. 1986), although in recent studies the difference in the amplitude of melatonin levels between old and young subjects was not significant (Luboshitzky et al. 1998, Zeitzer et al. 1999).

Metabolism and pharmacokinetics of melatonin

Melatonin in the circulation is quickly metabolized in the liver primarily to a watersoluble hydroxy derivative followed by conjugation with sulfate and glucuronic acid (Kopin et al. 1961). The remaining melatonin is excreted into urine (Lynch et al. 1975, Ozaki & Lynch 1976) or converted into some other metabolites (Hirata et al. 1974).

Waldhauser and associates (1984) examined the pharmacokinetic profile of exogenously administered melatonin in healthy humans. After oral administration of melatonin (80 mg) in gelatin capsules during daytime, the absorption half-life was 0.40 h with an elimination half-life of 0.80 h, and the melatonin levels ranged between 350 and 10 000 times those occurring endogenously at nighttime. Those findings are in line with the experiments in which the mean half-time of the elimination phase after oral administration was 47 min (Di et al. 1997). In both studies a great interindividual variability was observed in the peak melatonin concentrations. In the study by Di and coworkers (1997) it was suggested that the variable bioavailability of oral melatonin (intervariability from 10 to 56 %) is a consequence of variation in hepatic first-pass extraction.

Light and dark alterations constitute the principal timing signal of melatonin secretion from the pineal gland. Light influences melatonin synthesis in three ways in humans. First, light exposure acutely suppresses elevated melatonin levels. Second, light is able to phase shift the melatonin rhythm. Third, changes in the photoperiod can alter the melatonin secretion.

Suppressing effects of light

The first piece of evidence for the suppressing effect of light on the enzymatic capacity of the pineal gland to synthesize melatonin came from the study by Wurtman and coworkers (1963b) in rats. The early studies in human were unsuccessful in their attempts to suppress melatonin synthesis by light (Vaughan et al. 1976, Jimerson et al. 1977, Vaughan et al. 1979, Åkerstedt et al. 1979), probably because neither the exposure conditions nor the light stimulus were optimized.

However, in 1980, Lewy and associates (1980b) showed that also in humans this suppressing effect takes place if an exposure to bright light occurs during the abundant synthesis of melatonin. The discovery that bright light (³2500 lux) could suppress melatonin secretion in humans (Lewy et al. 1980b) led to the assumption that only high-intensity (>500 lux) light is able to affect on human circadian rhythms. Although human melatonin synthesis is less sensitive to light than that of nocturnal rodents (Minneman et al. 1974, Illnerová & Vanecek 1979), an interruption or a significant decrease in the synthesis has been produced by intensities as low as 300 lux (Bojkowski et al. 1987), 350-400 lux (McIntyre et al. 1989), 500 lux (Laakso et al. 1994c, Hashimoto et al. 1996) or 650 lux (Laakso et al. 1991). In addition, it has been shown that melatonin is suppressed by light in an intensity- and duration-dependent manner (Brainard et al. 1988, McIntyre et al. 1989, Aoki et al. 1998).

A study of Brainard and coworkers (1993) suggests that the peak sensitivity for melatonin suppression is in the blue-green range (509 nm). In fact, in specially controlled conditions when the pupils have been dilated, the volunteers' heads kept motionless and the light beam directed uniformly on the retina the mean threshold illuminance of monochromatic light of 509 nm for producing a statistically significant melatonin suppression is between 6 and 17 lux (photopic) or 28 and 86 lux (scotopic) in normal volunteers (Brainard et al. 1988); a level of illuminance equal to civil twilight and well below typical indoor light. This means that under optimal conditions much lower intensities of light can suppress melatonin than was originally believed.

Phase delaying and advancing effects of light

Studies in humans (Hashimoto et al. 1996) and in golden hamsters (Nelson & Takahashi 1991) suggest that the threshold intensity of light needed to phase shift the circadian system is greater than that needed for suppression of the melatonin level. After the suppressing effects of light on human melatonin secretion (Lewy et al. 1980b) and the theoretical prediction of light PRC (Lewy et al. 1983) were reported, it was shown that the onset of nocturnal melatonin production could be phase shifted depending on the timing of the bright light (~2500 lux) exposure (Lewy et al. 1987). Furthermore, the PRC to single light pulses of high intensity (~5000 lux, 3 hours) for human melatonin rhythm was constructed by Van Cauter and coworkers (1994). This study demonstrated delays or advances of the melatonin rhythm, indicating that the response to light is phase dependent.

In addition, the magnitude of the phase shifts of melatonin rhythms vary with the number and timing of light pulses. For example, in a study of Burešová and colleagues (1991), one day after a single exposure to bright light (~3000 lux, 6 hours) in the morning, melatonin onset and offset phase advanced by 0.6-2.6 hours. Dijk and coworkers (1989) reported that an exposure to bright light during three consecutive mornings (2000 lux, 3 hours) advanced melatonin rhythms by about an hour. Deacon and Arendt (1994) found melatonin phase delays of about 2-3 hours when exposing the subjects to light (1200 lux, 6 hours) during three consecutive nights.

Recently, light exposures as weak as the common indoor intensity (150-500 lux) have been shown to shift the endogenous circadian rhythm of plasma melatonin (Zeitzer et al. 1997, Boivin & Czeisler 1998) and body temperature (Waterhouse et al. 1998) in humans.

Effect of photoperiod

In nonhuman mammals, the photoperiod participates in the regulation of annual rhythms, e.g. breeding, renewal of fur and body weight rhythms (Martinet & Allain 1985, Zucker et al. 1991). Humans are not generally considered to be as photoperiodic as other species. A study by Van Dongen and coworkers (1998) demonstrated the absence of seasonal variation in the phase of circadian rhythms. However, data from Northern temperate (Beck-Friis et al. 1984, Kivelä et al. 1988) and Polar regions (Broadway 1987, Makkison & Arendt 1991), and from simulated long and short photoperiod experiments (Wehr 1991, Burešová et al. 1992, Van Dongen et al. 1997, Vondrašová et al. 1997) indicate that we retain a number of photoperiodic responses as a function of daylength. For instance, changes in melatonin levels, in duration of nocturnal melatonin secretion, and in the phase of the melatonin rhythm have been induced by seasonal variation of natural lighting conditions (Illnerová et al. 1985, Martikainen et al. 1985, Kauppila et al. 1987, Kivelä et al. 1988, Laakso et al. 1994b, Stokkan & Reiter 1994, Luboshitzky et al. 1998). These findings suggest that the regulation of melatonin is influenced by the photoperiod in humans as in other mammals.

Melatonin is evolutionarily well preserved and present in most organisms, from unicell algae to humans (Pelham et al. 1973, Smith et al. 1976a, Hardeland & Fuhrberg 1996). Melatonin's role as a phylogenetically conservative signal might be associated with its involvement in the adaptation of early living cells to the demands of climate. According to a hypothesis of Paietta (1982), the increased level of free oxygen in the early eukaryote evolution could have pushed the organisms to minimize the deleterious effects of the diurnal photooxidative exposures by developing circadian rhythmicity of metabolic activities. In addition, melatonin can detoxify highly reactive oxygen radicals (Poeggeler et al. 1993, Reiter et al. 1993, Tan et al. 1993). Therefore, melatonin as a natural oxidant may represent a property which during evolution has made this molecule a suitable indicator of the dark period (Hardeland 1993).

The effects of pinealectomy and melatonin administration on rhythms in animals

Pinealectomy, in which the lack of clear rhythmical production of melatonin is induced surgically, results in the elimination of the normal circadian rhythm of locomotor activity in birds (Gaston and Menaker 1968), whereas in rats pinealectomy has little effect on free-running circadian activity patterns (Richter 1964, Richter 1967, Quay 1968, Karppanen et al. 1973). However, pinealectomized rats do re-entrain to the phase-shifted photoperiods more rapidly than their sham-operated controls (Quay 1970).

A major step was taken in the studies of the role of melatonin in mammalian circadian organization in 1983 when Redman, Armstrong and Ng (1983) reported that daily injections of melatonin entrained the circadian locomotor rhythms of rats in constant dim red light. Somewhat later, it was shown that exogenous melatonin was able to phase shift activity rhythms (Redman & Armstrong 1988). In constant light, however, melatonin injections are not always able to repair a light-induced disruption of the rhythms (e.g. wheel-running or locomotor activity, drinking, and body temperature), while a clear synchronization occurs in some individuals (Chesworth et al. 1987, Thomas & Armstrong 1988, Marumoto et al. 1996, Deprés-Brummer et al. 1998, Witte et al. 1998). Thus, the mammalian pineal seems not to be essential for the generation of the circadian rhythm of locomotor activity, and the action of melatonin as a circadian synchronizer may be somewhat limited in strength (reviewed by Redman 1997).

The relationship between intrinsic rhythms and external signals is classically defined by phase response curves. The first PRC to melatonin in a vertebrate was conducted by Underwood (1986). Exogenous melatonin injected into free-running lizards induced phase delays (injections were administered late in the subjective night or in the first half of the subjective day) or phase advances (injections were administered between midsubjective day or early subjective night) of circadian activity rhythms. Afterwards, the PRC to melatonin was also constructed in other animals, e.g. in rats (Armstrong 1989).

The effects of exogenous melatonin on rhythms in humans

The first finding on a substance which had melanin aggregating and chromatographic properties similar to those of melatonin and a cyclic pattern was published in 1973 by Pelham and coworkers (1973) in humans. After that study, several studies have yielded evidence about melatonin's role as a rhythm modulator. For example, Arendt and colleagues (1985) first suggested that exogenous melatonin administration might be able to phase advance the endogenous melatonin profile. After it had been shown that daily melatonin administration induced a phase advance of the endogenous melatonin rhythm in blind subjects (Sack et al. 1991), Lewy and coworkers (1992) provided the first piece of evidence for a human PRC to melatonin (Figure 1). Melatonin administration in the morning elicited phase delays, and administration in the afternoon or early evening induced phase advances of the endogenous melatonin onset. In addition to the phase shifting effects of exogenous melatonin, Sack and associates (1991) have demonstrated a normally entrained endogenous melatonin rhythm in one blind person after about a year of melatonin treatment.

Compared to the human phase-response curve to light (Minors et al. 1991), PRC to exogenous melatonin (Lewy et al. 1992, Zaidan et al. 1994, Lewy et al. 1998a) has been described to be nearly opposite in phase (Fugure 1). In practice, the optimal time to produce a phase delay by exogenous melatonin is near the offset of endogenous secretion (at about 0700 h) and the optimal time of day to administer melatonin to produce a phase advance is 4-8 h before the onset of endogenous melatonin (at about 1700 for people on a conventional schedule) (Lewy & Sack 1997, Lewy et al. 1998a). Other circadian hormonal rhythms (e.g. cortisol, prolactin) and temperature rhythms can also be affected by melatonin administration (Arendt et al. 1987, Mallo et al. 1988, Kräuchi et al. 1997b).

Melatonin binding sites and feedback to the pacemaker

Although an intracellular function and a direct gene regulatory action of melatonin have been suggested (Carlberg & Wiesenberg 1995, Steinhilber & Carlberg 1999), many of the established effects of physiological concentrations of melatonin have been shown to be mediated via high-affinity cell membrane receptors belonging to the superfamily of G-protein-coupled receptors (reviewed by Kokkola & Laitinen 1998).

Based on sequence dissimilarities, melatonin receptors are classified into three subtypes (Mel_1a, Mel_1b and Mel_1c), and two of the subtypes (Mel_1a and Mel_1b) have been found in mammals (Kokkola & Laitinen 1998). The regions which show specific melatonin-binding in most mammals studied are the SCN and the hypophyseal pars tuberalis (Morgan et al. 1994). In humans, melatonin-binding sites are found in various regions of the brain, including the SCN (Weaver et al. 1993, Weaver & Reppert 1996), the temporal cortex (Fauteck et al. 1995), the cerebellum (Fauteck et al. 1994) and possibly in the hypophyseal pars tuberalis (Weaver et al. 1993), and also in peripheral sites such as the kidney (Song et al. 1995), granulosa cells (Yie et al. 1995) and prostate (Laudon et al. 1996).

There is also evidence of circadian oscillators within the SCN which are sensitive to the pineal hormone melatonin, raising the possibility that the circadian organization may be modulated by pineal feedback via melatonin. First, specific melatonin receptors in the SCN (Vanecek et al. 1987, Reppert et al. 1988, Morgan et al. 1994), also in humans (Weaver et al. 1993), provide a critical link for a functional feedback loop whereby circulating melatonin can influence the circadian pacemaker. Second, pinealectomy, i.e. suppressed melatonin production, increases the density of these binding sites (Gauer et al. 1992), and can alter the firing-rate rhythm in the SCN (Rusak & Yu 1993).

Third, the pinealectomy-induced increase in the density of binding sites can be inversed by a single melatonin injection (Gauer et al. 1993). In addition, exogenous melatonin inhibits the spontaneous electrical activity of the SCN of the rat when melatonin is administered near the transition to subjective night (Mason & Brooks 1988, Shibata et al. 1989, Stehle et al. 1989). In fact, the circadian rhythm of the discharge rate of the SCN can be reset by melatonin within two windows of sensitivity corresponding to dusk and dawn (McArthur et al. 1991, McArthur et al. 1997).

Fourth, melatonin alters the metabolic activity of the rat SCN (Cassone et al. 1988). Administration of melatonin to rats induced expression of Fos, the protein product of the c-fos proto-oncogene, in the SCN (Kilduff et al. 1992), and in vivo a single melatonin administration phase advanced the evening rise in the light-induced SCN c-fos expression (Sumová & Illnerová 1996). In conclusion, it seems that melatonin may modulate circadian overt rhythms probably by coupling individual oscillators to form an internal synchronization and affecting the photic sensitivity of the circadian timing system.

Biological calendar of reproduction

Previous classical studies on melatonin have mainly focused on its action on the hypothalamic-hypophyseal-gonadal axis, whereby it controls seasonal reproduction (Reiter 1980, Tamarkin et al. 1985). Because melatonin can act as a signal of photoperiod to the body, a long duration of melatonin secretion at night may cause gonadal atrophy in spring and summer breeders (e.g. hamsters), and gonadal growth in species which breed in the autumn and winter (e.g. sheep) (Bartness et al. 1993). In addition, melatonin was found to have antigonadal effects on rodent species that are not seasonal breeders (e.g. rats, mice) (Wurtman et al. 1963a, Glass & Lynch 1981).

In the late 19th century, Otto Heubner (1898) observed precocious puberty in a child with a pineal tumour. Since then, there have been several reports about the relationship between sexual maturation disorders and abnormal melatonin levels (Berga et al. 1988, Karasek et al. 1990, Puig-Domingo et al. 1992, Luboshitzky et al. 1995). Although correlations between melatonin and reproductive hormones have been observed in humans, the functional relationship remains to be determined (reviewed by Luboshitzky & Lavie 1999).

Melatonin as a protective agent

Melatonin has been found to be a potent intracellular scavenger of hydroxyl and peroxyl free radicals when administered at pharmacological doses both in vivo and vitro (Reiter 1995, Reiter 1998), suggesting that it has a protective role against oxidative damage. For instance, melatonin treatment may be beneficial in some neurodegenerative diseases, e.g. Alzheimer's and Parkinson's diseases (Reiter 1998, Reiter et al. 1998).

Melatonin may also function as an immunomodulator, and the immunoenhancing action seems to be mediated by certain T-lymphocytes and lymphokines (Maestroni 1993, Neri et al. 1998). In addition, some antitumoral effects of melatonin have also been reported (Buswell 1975, Wilson et al. 1992, Lissoni et al. 1995, Lissoni et al. 1996). In contrast to these findings, melatonin has been shown to be able to promote melanoma growth in hamsters (Stanberry et al. 1983) and induce suppression of human lymphocyte natural killer cell activity in vitro (Lewinski et al. 1989).

Furthermore, it has been suggested that melatonin has a role in the modulation of brain excitability. An anticonvulsant potential of exogenous melatonin has been proposed on the basis of in vivo studies in humans (Anton-Tay 1974, Brueske et al. 1981, Molina-Carballo et al. 1997, Jan et al. 1999) and an in vitro experiment in which epileptiform activity in human temporal slices was reduced by melatonin (Fauteck et al. 1995). On the other hand, Sheldon (1998) reported that melatonin induced an increased seizure frequency in 4 of 6 neurological patients. Taken together, it seems that the effects of melatonin may not be completely protective.

The pineal gland is not the only source of melatonin hormone. The suggestion that extrapineal melatonin exists in humans was made by Raikhlin and coworkers (1975) when they blanched the frog's skin with extracts of human intestinal mucosa. In fact, the gastrointestinal tract represents the most abundant extrapineal source of melatonin, for it contains over 80 times more melatonin than the pineals of chicks and rats (Huether et al. 1992). Arrhythmic melatonin levels have been observed in alligators (Alligator mississippiensis) lacking pineal bodies (Roth et al. 1980). Melatonin is also found in the blood of surgically pinealectomized rats (Ozaki & Lynch 1976, Pang et al. 1977). In vertebrate retinas, melatonin is thought to act locally to regulate various dark-adaptive functions (Cahill & Besharse 1995). Melatonin is also synthesized in small amounts in the human retina (Leino 1984), but the role of extrapineal melatonin in circadian functions in humans is not known.

The entry of the amount of light needed for the resetting of the rhythms to the circadian timing system may be prevented by blindness. Thus, blind people may fail to maintain the circadian rhythms of sleep and wake, body temperature, melatonin and cortisol synchronized with the LD cycles (Miles et al. 1977, Orth et al. 1979, Smith et al. 1981, Lewy & Newsome 1983, Okawa et al. 1987, Nakagawa et al. 1992, Sack et al. 1992b, Klein et al. 1993, Palm et al. 1997, Klerman et al. 1998, Skene et al. 1999, Lockley et al. 2000). They may have sleep-wake disruption and may develop either free-running circadian rhythms or disorganized sleep-wake cycles (Arendt et al. 1988, Lockley et al. 1997a, Lockley et al. 1997b, Tabandeh et al. 1998, Lockley et al. 1999). In addition, periodic insomnia and daytime sleepiness (Sack et al. 1992b, Tzischinsky et al. 1992) may occur when the internal clock becomes desynchronized from the solar and social 24-hour cycle.

Although in a study by Czeisler and coworkers (1995), a 3-hour, 10000-lux facial light exposure in patients with different origins of blindness (e.g. retinopathy or optic neuropathy) did not suppress melatonin synthesis in most of the 11 blind subjects with no conscious perception of light, a clear decrease of melatonin was found in 3 subjects who had totally lost the pupillary reflexes. The light suppressing effect was only found when the patients' eyes were uncovered. This work indicates that in a subgroup of patients without conscious light perception the photic pathway through the eyes used by the circadian system may be functional.

Classification of NCL diseases

Neuronal ceroid lipofuscinoses (NCLs) are the most common group of progressive neurodegenerative diseases of childhood in the Western world. The disorders were first described by Stengel, Batten and others in the 19th and early 20th centuries (reviewed by Goebel et al. 1999), and termed neuronal ceroid lipofuscinoses by Zeman and Dyken (1969) because the neural and extraneural accumulation of storage material resembled ceroid and lipofuscin. Based on recent findings on molecular genetics, NCLs can be classified as atypical lysosomal storage diseases (Mole 1998, Mole et al. 1999).

Classically, NCLs have been devided into four main forms: the infantile (INCL, Santavuori-Haltia disease), late infantile (LINCL, Jansky-Bielschowsky or early onset Batten disease), juvenile NCL (JNCL, Spielmeyer-Vogt-Sjögren or late onset Batten disease) and adult onset NCL (Kufs' disease and Parry disease). In addition, a congenital NCL and Northern Epilepsy Syndrome (Progressive Epilepsy with Mental Retardation) have been reported (Hirvasniemi et al. 1994, Goebel et al. 1999).

To date, depending on the age of onset, clinical course of the disease, characterization of the storage material, presence of vacuolated lymphocytes, and chromosomal location of the disease gene, at least 11 childhood and two adult types are recognized (Mole 1999). All childhood and most adult types are inherited in an autosomal recessive fashion. According to present knowledge, eight genes (symbols from CLN1 to CLN8) underlie NCLs (Goebel et al. 1999). Five of the genes have been identified: CLN1 (INCL), CLN2 (classical LINCL), CLN3 (JNCL), CLN5 (Finnish variant LINCL) and CLN8 (Northern Epilepsy Syndrome) (Vesa et al. 1995, Sleat et al. 1997, The International Batten Disease Consortium 1995, Savukoski et al. 1998 and Ranta et al. 1999, respectively). In addition, 86 different disease-causing mutations have been reported (http://www.ucl.ac.uk/ncl, accessed on Mar 23, 2000).

Incidence and common symptoms of NCL diseases in Finland

The most common NCL types in Finland are INCL and JNCL, whereas LINCL is rarer (Santavuori 1999). One adult form of NCL has been diagnosed in Finland (Haltia, personal communication). At present, the incidence of INCL is 1:20000, approximating three new INCL patients per year (Santavuori 1999, Santavuori personal communication). Altogether 158 INCL patients have been diagnosed. The number of diagnosed LINCL patients is 31, consisting of five classic LINCL (gene symbol CLN2) and 26 variant LINCL (CLN5) patients. 185 JNCL patients have been diagnosed, with an incidence of 1:21000.

NCL children are healthy at birth with normal development until the onset of the disease. In the most common Finnish NCL types, including INCL (gene symbol CLN1) and JNCL (CLN3), the clinical course is characterized by loss of vision, epilepsy, progressive cognitive impairment, and psychomotor disturbances leading to premature death.

Clinical picture of INCL

In INCL patients the development is normal until the age of 0.5-1 year, and in some children until 1.5 years (Santavuori 1988). The onset of epilepsy occurs at the mean age of 30 months (Vanhanen 1996). INCL patients suffer from anxiety and hyperexcitability (Santavuori 1988) causing sleep problems, which are among the most disturbing factors in the everyday life of the patients and their families. In our earlier study a fragmented activity rhythm was observed in 4 of 5 INCL patients, but disturbances in the daily melatonin and cortisol rhythms occurred only in the minority of patients, and only at an advanced stage of the disease (Heikkilä et al. 1995).

In neurophysiological examination of INCL patients, electroencephalogram (EEG) shows posterior low activity to eye opening and closing starting at a mean age of 1.8 (range 1.4-2.3) years, and the activity in the EEG disappears after the age of 2.7 years (Vanhanen et al. 1997). Neuroradiological findings by magnetic resonance imaging (MRI) may show hypointensity of the thalamus as compared with the white matter and basal ganglia, and cerebral and cerebellar atrophy (Autti et al. 1997,Vanhanen et al. 1994).

The condition of the patient deteriorates rapidly and before the age of 3 years the child has lost all cognitive and active motor skills (Santavuori et al. 1974). Death usually takes place at the age of 8-13 years (Santavuori et al. 1999).

The characteristic neuropathological finding in INCL patients is extreme atrophy of the brain, whereas the brain stem and particularly the spinal cord are less affected (Haltia et al. 1973, Santavuori et al. 1974). The consistency of the brain is tough and rubberlike, and all cerebral gyri are narrowed and sulci widened (Haltia et al. 1973).

Pathology of visual system in INCL

Blindness is one of the main symptoms in INCL patients, and visual disturbance is usually observed between 12 and 24 months of age (Santavuori et al. 1974). Perception of light can be lost as early as at the age of 18 months, but most often a few months later (Kohlschütter & Goebel 1997). Visual failure leads rapidly to blindness. INCL children are usually practically blind at the age of two (Raitta & Santavuori 1973). Pupillary reactions are slow or absent after the age of two years (Santavuori et al. 1974) but, surprisingly, may reestablish during a later stage of the disease (Santavuori et al. 1999).

The earliest ophthalmoscopical findings are hypopigmentation of the fundus without distinct macular changes (Raitta & Santavuori 1981). Macular changes appear by the age of 12-18 months as depigmentation and mottling of the pigmentlayer. In fluoresceinangiography the changes can be distinguished by the age of 18-24 months and are parallel to the general dystrophy of the retina. Retinal vessels become extremely narrow. They are hardly visible by the age of 3 years when the optic disc appears atrophic (Raitta & Santavuori 1981). In addition to retinal degeneration, clearly visible choroidal vessels are seen (Santavuori et al. 1974). Pigment aggregation of the fundus periphery is not usually found in INCL patients (Raitta & Santavuori 1973, Kohlschütter & Goebel 1997).

In neurophysiological examination, visual evoked potential (VEP) and electroretinogram (ERG) abnormalities appear between 22 and 25 months of age (Vanhanen et al. 1997). The mean age of abolition of ERG is 3.1 (range 2.3-4.1) years and that of VEP 3.8 (2.2-5.4) years.

Neuropathological autopsy of 5 INCL patients showed that the retinas were completely destructed (Tarkkanen et al. 1977). The retina was severely atrophic with complete loss of photoreceptors, bipolar and ganglion cells which were replaced by marked glial proliferation. Loss of pigment in the retinal pigment epithelium had taken place to some extent. The optic nerve was atrophic and gliosed with a complete loss of myelin sheets. In addition, there was accumulation of granular material in the nonpigmented ciliary epithelium, the retina, the retinal pigment epithelium, and the optic nerve.

Clinical picture of JNCL

In JNCL patients the disease usually becomes manifest at early school age. The first symptom is usually visual failure detected between 4 and 7 years of age (Santavuori 1988). At first, mental impairment is slight and apparent only in school performance.

Some time after the onset of visual loss, a personality change may be noted which can be accompanied by various types of psychiatric disturbances (Boustany 1992, Wisniewski et al. 1992, Hofman et al. 1999). The greatest decline in motor functions and intelligence usually takes place between 11 and 15 years of age (Järvelä et al. 1997). The onset of epilepsy occurs, on the average, at the age of 11 years (Järvelä et al. 1997).

Sleep problems are reported in more than half of JNCL patients (Santavuori et al. 1993). The most common problems include frequent awakenings, difficulties in falling asleep, nightmares and night terrors. Despite the disturbances observed in the sleep-wake rhythm in 7 of 8 JNCL patients, most of the patients had normal melatonin, cortisol and temperature rhythms (Heikkilä et al. 1995).

In neurophysiological recordings, EEG is usually found normal at the preclinical stage, and normal or slightly abnormal by the age of 5-6 years in most patients (Santavuori et al. 1988). In MRI, abnormal periventricular white matter may be found in all age groups, and the thalamus, caudate nucleus and putamen usually give low signal in patients above the age of ten years (Autti et al. 1996). During the progression of the disease, movement difficulties develop (Boustany 1992, Wisniewski et al. 1992, Hofman et al. 1999). The disease leads to death at the mean age of 24 years (Järvelä et al. 1997).

In neuropathological examination of JNCL patients, the macroscopic appearance of the brain may be nearby normal (Zeman & Dyken 1969, Zeman et al. 1970). The cerebral cortex is slightly atrophic, whereas the number of neurones is usually not markedly reduced. The most marked changes in the white matter are seen in periventricular areas (Vanhanen et al. 1995).

Pathology of the visual system in JNCL

In JNCL patients the leading initial symptom is visual failure. The visual problems are noticed at a mean age of 5.8 (range 4-10) years, and the patients become practically blind between 6 and 20 years of age (Järvelä et al. 1997, Kohlschütter & Goebel 1997). The characteristic ophthalmological findings are retinal degeneration, macular dystrophy, optic atrophy and thinning of vessels (Spalton et al. 1980, Traboulsi et al. 1987, Santavuori 1988, Seeliger et al. 1997, Neppert & Kemper 1998). In addition, the typical pigment aggregation is usually seen in the peripheral retina (Raitta & Santavuori 1981, Kohlschütter & Goebel 1997).

In neurophysiological examination, ERG is usually abolished around 6 years of age (Santavuori et al. 1988, Kohlschütter & Goebel 1997, Seeliger et al. 1997, Weleber 1998). VEP is pathological when clinical signs are apparent and usually become abolished between 13 and 16 years. In some patients VEP was still recordable, although abnormal, between the age of 18 and 21 years (Santavuori et al. 1988). The abnormality of the ERG is usually seen as a severe loss of rod and cone responses (Horiguchi & Miyake 1992, Seeliger et al. 1997, Weleber 1998).

Pathological examination of one JNCL patient revealed a near-complete loss of photoreceptors (Traboulsi et al. 1987). The autopsy of 3 JNCL patients showed that two of them had lost all retinal photoreceptor cells, whereas in one patient with the shortest clinical course of the disease there were still photoreceptors present in the periphery of the retina (Goebel et al. 1974). Changes were observed in the optic nerves, ganglion cell layers and the pigment epithelium (Goebel et al. 1974, Traboulsi et al. 1987) but the choroid and sclera were normal (Traboulsi et al. 1987).

Treatment of NCL diseases

At present, the treatment of NCLs is merely palliative, including medication (e.g. antiepileptics and antiparkinsonian drugs, analgetics, antidepressants, antipsychotic therapy, antioxidants and hormonal treatment) combined with family guidance, psychological support, social help, artificial feeding, physiotherapy, speech and occupational therapy (Hofman et al. 1999, Santavuori et al. 1999). Thus, there is no cure yet for this devastating group of diseases. However, new forms of treatment will hopefully emerge as the understanding of the pathological processes increases.

The current knowledge about the effects of light and melatonin on human physiology is partly based on unsystematic observations. The indications for phototherapy and melatonin administration are not yet conclusively established, neither are the long-term benefits and risks thoroughly known.

Patients treated with bright light have reported some temporary side effects including headache, eyestrain, irritability, nausea and insomnia (Levitt et al. 1993). There is also evidence that symptoms of mania may emerge as a consequence of phototherapy (Schwitzer et al. 1990). However, short- or long-term daily treatment with 10 000 lux light has not been observed to result in ocular changes in patients suffering from seasonal affective disorder (Gallin et al. 1995).

Melatonin is available as a nutritional supplement in the US without a premarket approval from the FDA (Food and Drug Administration), and it is widely sold in health-food and drug stores. In Canada, Finland, Great Britain and some other European countries melatonin is classified as a medicine and available only on prescription. However, the standard clinical trial methodologies for judging the safety of a drug have not been applied to melatonin. Indeed, very little is known about the long-term adverse effects of melatonin in humans. The results of experiments in rodents indicate low toxicity (Sudgen 1983, Jahnke et al. 1999) and no mutagenicity (Neville et al. 1989). This is consistent with the human studies in which doses ranging from 0.1 mg to several grams have not been found to result in medical catastrophes (Lerner & Nordlund 1978, Wirtz-Justice & Armstrong 1996, Sack et al. 1997, Zhdanova et al. 1997).

The potential side effects of melatonin are drowsiness (Arendt et al. 1984, Dollins et al. 1993, Dollins et al. 1994) and phase shifting of circadian rhythms (Dawson & Armstrong 1996). In addition, sleep could be disrupted if melatonin is not properly timed (Middleton et al. 1996b). The possible adverse effects of melatonin also include cutaneous flushing, gastrointestinal disorders, headache and vascular constriction (Papavasiliou et al. 1972, Guardiola-Lemaître 1997, Mahle et al. 1997).

The most common group of diseases or symptoms treated with light or exogenous melatonin comprises circadian disorders, which are linked to poor resetting of the circadian timing system and usually associated with free-running or abnormal overt rhythms. Disturbances in the circadian control of various body functions have been found to be related to working during unusual hours (Sack et al. 1992a, Weibel et al. 1997) and jet lag (Désir et al. 1981, Fevre-Montange et al. 1981, Désir et al. 1982, Goldstein et al. 1983, Härmä et al. 1993, Suvanto et al. 1993). In addition, abnormal circadian rhythms have been suggested to underly the sleep difficulties of blind people (Arendt et al. 1988, Nakagawa et al. 1992, Lockley et al. 1997a, Lockley et al. 1997b, Tabandeh et al. 1998, Lockley et al. 1999), of some neurologically disabled children (Laakso et al. 1993, Palm et al. 1997, McArthur & Budden 1998, O'Callaghan et al. 1999, Zhdanova et al. 1999) and of the elderly (Haimov et al. 1994). Circadian pathology has been suggested to be related to some affective disorders (Wehr & Goodwin 1983).

The sleep-wake cycle is one of the most obvious rhythms regulated by the circadian clock. The disorders that are related to the timing of sleep within the 24-hour day are named circadian rhythm sleep disorders (American Sleep Disorders Association 1997). In most of the disorders, the underlying problem is that the person cannot sleep or be awake when desired, needed, or expected.

The phase shifts of circadian rhythms are the probable mechanism by which light improves sleep (Dijk et al. 1995). The alerting and activating property of bright light exposures has also been proposed to be one possible way in which light affects arousal and sleep onset (Campbell et al. 1995). Administration of melatonin might improve sleep by producing corrective circadian phase shifts (Zhdanova et al. 1997), thereby improving the alignment of the endogenous sleep propensity rhythm with the desired sleep schedule (Sack & Lewy 1997). In addition, melatonin may increase sleepiness by a soporific effect (Wirz-Justice & Armstrong 1996).

After rapid air travel across several time zones, the endogenous circadian timing system is desynchronized and begins to adapt to new environmental time cues (American Sleep Disorders Association 1997). Because the adjustment process of the circadian system is slow, jet lag symptoms, e.g. disturbed sleep, daytime sleepiness and impaired performance, can last for several days after the flight. Scheduled exposure to bright light can alleviate the symptoms (Daan & Lewy 1984, Boulos et al. 1995). During the lasts ten years, evidence has accumulated about the ability of melatonin administration to decrease jet lag symptoms (Arendt & Marks 1983, Arendt et al. 1986, Arendt et al. 1987, Petrie et al. 1989, Claustrat et al. 1992, Petrie et al. 1993), although in a recent extensive study by Spitzer and coworkers (1999) melatonin did not show any beneficial effects as compared with placebo.

Shift work is a common cause of circadian rhythm sleep disorders. The symptoms, e.g. difficulties in sleep maintaining and reduced alertness and performance, can be explained by a mismatch between the work-sleep schedule and the internal circadian rhythms (American Sleep Disorders Association 1997). There is evidence indicating that maladaptation to shift work can be treated with properly timed exposures to light (Dawson & Campbell 1991, Bjorvatn et al. 1999). Also, the use of artificial nocturnal bright light combined with enforced daytime dark periods can phase shift circadian rhythms effectively despite an exposure to conflicting 24-h time cues (Czeisler et al. 1990, Eastman et al. 1994, Koller et al. 1994, Eastman & Martin 1999). Application of exogenous melatonin has been tested as therapy in sleep disorders of night workers (Folkard et al. 1993).

The delayed sleep phase syndrome is a disorder in which the major sleep episode is delayed in relation to the desired clock time, resulting in symptoms of sleep-onset insomnia or difficulty in awakening at a desired time (American Sleep Disorders Association 1997). Unlike the advanced sleep phase syndrome, which is rather rare, the delayed sleep phase syndrome is quite a common sleep schedule disorder. Lewy and coworkers (1985) were the first ones to recommend the use of a morning light exposure to treat the delayed sleep phase syndrome. After that study, more evidence has been obtained about the beneficial implications of light (Rosenthal et al. 1990) and melatonin (Dahlitz et al. 1991, Tzischinsky et al. 1993, Oldani et al. 1994, Dagan et al. 1998) in the therapy of the delayed sleep phase syndrome.

Blind people may have a non-24-hour sleep cycle that reflects a nonentrained intrinsic rhythm of the circadian pacemaker (American Sleep Disorders Association 1997), resulting in recurrent insomnia and daytime sleepiness (Sack et al. 1992b, Tabandeh et al. 1998). Several reports indicate beneficial effects of melatonin administration in these patients (Arendt et al. 1988, Folkard et al. 1990, Sarrafzadeh et al. 1990, Sack et al. 1991, Tzischinsky et al. 1992, Lockley et al. 2000).

Melatonin has also been found to promote sleep in patients with neurological disabilities, especially in those with circadian sleep disorders (Palm et al. 1991, Jan et al. 1994, Lapierre & Dumont 1995, Palm et al. 1997, McArthur & Budden 1998, O'Callaghan et al. 1999, Zhdanova et al. 1999). On the other hand, controversial effects have also been observed in children with mental retardation and fragmented sleep (Camfield et al. 1996), and melatonin was also ineffective in our previous study in neuronal ceroid lipofuscinosis patients with fragmented or normal motor activity rhythms (Hätönen et al. 1999). In addition, bright light therapy has been demonstrated to have some positive effects in neurologically disabled children with sleep-wake rhythm problems (Guilleminault et al. 1993).

The study of Haimov and associates (1994) has documented lower melatonin levels in elderly people with poor sleep compared with the elderly with no sleep complaints. It is suggested that exogenously administered melatonin may be beneficial in elderly insomniacs (Garfinkel et al. 1995, Haimov et al. 1995), and in sleep problems that are related to melatonin deficiency (Etzioni et al. 1996, Lehmann et al. 1996). On the other hand, a study of Youngstedt and coworkers (1998) showed no association between low melatonin and insomnia. Furthermore, bright light treatment has been shown to have beneficial effects on rest-activity rhythm disturbances and sleep maintaining in old demented patients and elderly subjects (Murphy & Campbell 1996, Van Someren et al. 1997).

Seasonal affective disorder (SAD) was originally defined as a syndrome in which depression developed during the autumn or winter and remitted in the spring or summer (Marx 1946, Rosenthal et al. 1984). At present, it is classified as a form of recurrent depressive or bipolar disorder, and it is characterized by episodes that vary in severity (Partonen & Lönnqvist 1998).

Bright light exposure treatment has been used successfully in SAD (Lewy et al. 1987, Eastman et al. 1992, Partonen 1994, Eastman et al. 1998, Lewy et al. 1998c, Terman et al. 1998). However, the mechanism by which light reduces depressive symptoms is unknown, and even nonphotic high-density negative air ionization has been shown to act as an antidepressant in patients with SAD (Terman et al. 1998). This rhythmic blue disorder was also proposed to be linked to melatonin secretion (Lewy et al. 1987), but later findings showed that winter depression patients do not have abnormal melatonin levels (Checkley et al. 1993), and melatonin suppression does not appear to be causally involved in the antidepressant effects of bright light therapy on the affective illness (Dietzel et al. 1985, Wehr et al. 1986). Exogenous melatonin has been demonstrated to decrease (Lewy et al. 1998b) or have no effect on (Witz-Justice et al. 1990) the depressive symptoms of SAD patients. However, because alternative treatment is available for SAD, such as nondrug therapy by bright light, melatonin is not considered a first-choice treatment at the present time.

Light may well affect mood (Lewy et al. 1987, Eastman et al. 1992, Partonen 1994, Eastman et al. 1998, Lewy et al. 1998c, Terman et al. 1998). In fact, there is now general agreement that the treatment of choice in patients suffering from seasonal affective disorder is bright light exposures (Partonen & Lönnqvist 1998). In addition, light therapy can be useful in the treatment of delayed and advanced sleep phase syndromes (Chesson et al. 1999). However, the benefits of light treatment are less clear in jet lag, shift work, and in non-24-hour sleep-wake syndrome of blind people. Above all, further applied research in everyday conditions is required, because it is important to find out how alterations in indoor and outdoor lighting conditions and avoidance of bright light exposures affect the entrainment of circadian timing system.

Although melatonin promotes sleep in humans (Zhdanova et al. 1997), it is clear that melatonin is not primarily a sleep hormone, because in nocturnal species melatonin is associated with wake and activity (Mendelson et al. 1980, Huber et al. 1998). The majority of published data indicates that melatonin has a therapeutic potential in circadian rhythm-related sleep disorders (Arendt et al. 1997, Arendt & Deacon 1997). As yet, however, its mechanism of action remains unclear, the appropriate dose, timing and method of delivery with respect to any given condition and individual are uncertain, the contraindications remain to be defined, and there are virtually no data on the long-term safety or about the use with concomitant medication or organic disease.

Although melatonin functions in mammalian physiology as an important photoperiodic messenger molecule (Wehr 1997), very little information is available concerning its function as a neuroendocrine phototransducer in humans. Indeed, much further research is needed on melatonin's interaction with light and LD alterations.