Melatonin and the pineal gland: Influence on mammalian seasonal and circadian physiology

Abstract

The pineal hormone melatonin is secreted with a marked circadian rhythm. Normally, maximum production occurs during the dark phase of the day and the duration of secretion reflects the duration of the night. The changing profile of secretion as a function of daylength conveys photoperiodic information for the organization of seasonal rhythms in mammals. The role of melatonin in mammalian circadian physiology is less clear. However, exogenous melatonin can phase shift, and in some cases entrain, circadian rhythms in rodents and humans. It can also lower body temperature and induce transient sleepiness. These properties indicate that melatonin can be used therapeutically in circadian rhythm disorder. Successful outcomes have been reported, for example in jet lag and shift work, and with cyclic sleep disorder of some blind subjects. Melatonin receptors of several subtypes are found in the brain, the retina, the pituitary and elsewhere. They are currently under intense investigation. Melatonin agonists and antagonists are under development.

Full text

© 1998 Journals of Reproduction and Fertility
1359-6004/98 $12.50
Reviews of Reproduction (1998) 3, 13–22
The pineal gland appears to serve the same function in all
mammals studied to date. The pattern of secretion of its major
hormone, melatonin, conveys information concerning light–dark
cycles to the body physiology for the organization of seasonal
and circadian rhythms (Arendt, 1995). Recently, much progress
has been made in identifying, localizing and characterizing
melatonin receptors. This review provides a brief summary of
the important features of the production and effects of mela-
tonin that impinge upon its circadian physiology and seasonal
functions and that have proved of use in the general study of
biological rhythms. Wherever possible, broad reference is made
to reviews rather than the original literature. Where no citation
is given to publications before 1994, references can be found in
Arendt (1995) and other cited reviews.
The pineal gland is part of the visual system and mammalian
pinealocytes are derived evolutionarily from the pineal photo-
receptors of lower vertebrates. The influence of the pineal gland
on the circadian system appears to be more important in lower
vertebrates than in mammals. In some reptiles and birds, the
pineal appears to act as a central circadian rhythm generator. In
house sparrows (Passer domesticus), pinealectomy leads to ar-
rhythmicity which can be restored by transplanting a pineal from
another bird (Menaker et al., 1981). The circadian phase of the
donor bird is conveyed to the host with the transplant. It is possi-
ble to culture pineal explants and dispersed pineal cells from
fish, lizards and birds, and these preparations retain their cir-
cadian melatonin production in vitro. In contrast, the mammalian
pineal does not retain endogenous rhythmicity in culture.
The retinae of lower vertebrates also generate melatonin
rhythms in culture and the hamster retina (maintained at low
temperature; Tosini and Menaker, 1996) can show the same
phenomenon, which suggests that there may be a circadian
pacemaker in the mammalian eye.
In contrast to reptiles and birds, pinealectomy in mammals
has rather more subtle effects on the circadian system. For
example, the rate of resynchronization of rats to a phase shift
of the light–dark cycle is faster in pinealectomized than in in-
tact animals, and pinealectomy leads to disrupted circadian
rhythms in rats kept in constant light (Armstrong, 1989; Cassone,
1992). The pineal, and indeed melatonin secretion, cannot be
regarded as essential in the adult mammalian circadian
system given that, in their absence, virtually normal function
is maintained.
In mammals, the ability to respond to changing artificial
daylength in terms of seasonal functions is abolished by
pinealectomy or denervation of the gland. In long-lived
species, such as sheep and ferrets, pinealectomy leads to
desynchronization of seasonal rhythms in reproductive func-
tion from the annual periodicity, which suggests that the pineal
essentially synchronizes the endogenous cycle to a yearly
periodicity.
Production of melatonin
Melatonin is synthesized within the pineal gland itself, in
the retina and possibly in some other sites. However, in
mammals, most if not all of the hormone reaching peripheral
sites is derived from the pineal, and pinealectomy leads to a
great reduction, in most cases to undetectable concentrations,
of circulating melatonin. Melatonin is synthesized from
tryptophan via 5-hydroxylation by tryptophan-5-hydroxylase
to 5-hydroxytryptophan, decarboxylation by aromatic amino-
acid decarboxylase to 5-hydroxytryptamine (serotonin), N-
acetylation by N-acetyl transferase to N-acetylserotonin (NAT)
and O-methylation by hydroxyindole-O-methyltransferase
(HIOMT) to melatonin (N-acetyl-5-methoxytryptamine) (Fig. 1).
The cardinal feature of this synthetic pathway is its rhyth-
micity. The activity of the enzyme NAT in particular increases
from 30-fold to 70-fold at night and, in most circumstances, is
rate-limiting in melatonin synthesis (Klein et al., 1997).
Melatonin and the pineal gland: influence on mammalian seasonal and
circadian physiology
Josephine Arendt
School of Biological Sciences, University of Surrey, Guildford, Surrey GU2 5XH, UK
The pineal hormone melatonin is secreted with a marked circadian rhythm. Normally, maximum
production occurs during the dark phase of the day and the duration of secretion reflects the
duration of the night. The changing profile of secretion as a function of daylength conveys
photoperiodic information for the organization of seasonal rhythms in mammals. The role of
melatonin in mammalian circadian physiology is less clear. However, exogenous melatonin
can phase shift, and in some cases entrain, circadian rhythms in rodents and humans. It can
also lower body temperature and induce transient sleepiness. These properties indicate that
melatonin can be used therapeutically in circadian rhythm disorder. Successful outcomes have
been reported, for example in jet lag and shift work, and with cyclic sleep disorder of some
blind subjects. Melatonin receptors of several subtypes are found in the brain, the retina, the
pituitary and elsewhere. They are currently under intense investigation. Melatonin agonists
and antagonists are under development.

The pineal is innervated primarily from a peripheral sym-
pathetic tract arising in the superior cervical ganglion, and
there is some evidence for direct central innervation. Sectioning
the sympathetic nerve supply abolishes melatonin rhythmicity
and mimics the effects of pinealectomy. The endogenous cir-
cadian rhythm of melatonin, like most other circadian rhythms,
is generated in the suprachiasmatic nuclei (SCN) and entrained
principally by the light–dark cycle acting via the retino–
hypothalamic tract, probably with a contribution from the
lateral geniculate nucleus (Fig. 1).
Sympathetic pineal input terminates in adrenoceptors, char-
acterized as beta 1 and alpha 1 in rodents and humans, with
a recent report of alpha-2D adrenoceptors being present in
rodents and bovines. There is good evidence that the control of
melatonin secretion in humans is similar to that of rats. In rats,
it is clear that beta-adrenoceptor stimulation of pineal activity
is potentiated by concomitant α-adrenoceptor stimulation. The
transmitter is noradrenaline. The events surrounding pineal
adrenoceptor stimulation have been used extensively to charac-
terize β- and α-adrenoceptor function.
Human melatonin rhythms and reproductive function
In humans, the melatonin rhythm in either plasma or saliva
is arguably the best marker of the phase of the endogenous
‘biological clock’, principally because it is demonstrably con-
trolled by the SCN and because very few ‘masking’ factors
influence its production. Only light suppression grossly
masks the rhythm (Fig. 2). Furthermore, the melatonin
rhythm is coupled tightly to the core temperature rhythm,
known to be a good circadian rhythm marker, with the peak
of melatonin secretion corresponding closely to the nadir of
temperature (Fig. 2). In the same individual, it is highly repro-
ducible in amplitude, details of profile and in timing in a
synchronized environment, almost like a hormonal finger-
print (Fig. 2). This means that small differences can be highly
14 J. Arendt
Light
RH tract
Entrainment
Hypothalamus
Pinealocyte
Tryptophan
5-Hydroxytryptophan
5-Hydroxytryptamine
N
-Acetylserotonin
Protein
synthesis
NAT
HIOMT
cAMP
ATPβ-Adrenergic
receptor
α-Adrenergic
receptor
C-Kinase
SCN
SCG
Preganglionic
fibres
Postganglionic
fibres
NA
NA
PVN

Eye Hindbrain
Spinal cord
Melatonin
Fig. 1. Diagram of the major controlling mechanisms in melatonin synthesis. The rhythm is generated in the suprachiasmatic nucleus (SCN),
entrained by light via the retino–hypothalamic pathway (RH tract). The signal passes via the paraventricular nucleus (PVN, albeit with some
controversy), hindbrain, spinal cord, superior cervical ganglion (SCG) to pineal noradrenergic (NA) receptors. Serotonin N-acetyltransferase
(NAT), in most circumstances the rate-limiting enzyme in melatonin synthesis, increases 30–70-fold during the dark phase through nor-
adrenergic stimulation. HIOMT, hydroxyindole-O-methyltransferase.

significant. Between individuals there are large differences in
amplitude.
The disadvantage of plasma melatonin is that only ‘snap-
shots’ can be taken of evolving circadian status in order to
avoid unacceptable blood loss. In humans, the use of saliva
avoids this, but brings with it the problem that it can only be
collected during the night by waking the subject. In view of the
non-invasive nature of sampling, the major urinary melatonin
metabolite 6-sulfatoxymelatonin (aMT6s) is especially useful
for long term and field studies. aMT6s reflects accurately both
the quantitative and qualitative aspects of melatonin secretion,
albeit with some loss of detail. Abnormal melatonin profiles
have rarely been reported in clinical studies. One problem with
many clinical observations is the lack of control of environ-
mental and postural variables, and much work needs reassess-
ment with this in mind.
The onset of a circadian melatonin rhythm with peak values
at night is established by 9 months of age in humans. Secretion
reaches a lifetime peak between 3 and 5 years, subsequently
declining to adult amounts by 15–18 years. Amplitude remains
relatively stable until old age, when a marked decline is re-
ported in most studies. Low amplitude in old age may be
related to general lack of robustness of the circadian system.
Whether or not the declining plasma concentrations during
puberty have any causal role in pubertal development remains
to be proven. However, evidence is accumulating for a role
of melatonin in human reproduction. In precocious puberty,
amplitude is high for age, and in delayed puberty, amplitude
is low for age. Moreover, hypothalamic amenorrhoea is associ-
ated with high melatonin concentrations and studies in Finland
have shown high daytime melatonin associated with anovu-
latory cycles (references in Arendt, 1995). Very large doses of
melatonin (80–300 mg) have been shown variously to suppress
LH and, in combination with norethistrone (an oral contra-
ceptive minipill), to suppress ovulation when given at night,
to increase the amplitude of LH pulses when given in the
morning and to potentiate testosterone-induced suppression of
LH when given in the late afternoon (references in Arendt,
1995).
Light suppression and entrainment of
melatonin secretion
For a given period of darkness, melatonin is produced during
the dark phase (Fig. 3) and the duration of its secretion is dic-
tated by the duration of darkness up to a defined duration of
secretion which is species dependent. In sheep, the duration
of melatonin secretion extends to approximately 16 h in 8 h:16 h
light:dark (but does not expand further in dark:dark) and re-
tracts to 8 h in 16 h:8 h light:dark, whereas in hamsters, while
the duration of secretion positively reflects the duration of dark-
ness, it occupies rather less of the dark phase than in sheep
(Arendt, 1986).
Light of sufficient duration and intensity suppresses noctur-
nal melatonin production. The amount of light required is de-
pendent on both species and photoperiodic environment. For
example, in some laboratory-raised animals less light is needed
than in the same species raised in the wild. Humans require
2500 lux for complete suppression, although partial sup-
pression can be observed with as little as 100–300 lux (Fig. 2c).
Melatonin 15
18 22 02 06 10 18 22 02
Clock time (h)
Clock time (h)
06 10 18 22 02 06 10
60
50
40
30
20
10
0
50
40
30
20
10
0
30
(a)
(b)
(c)
20
10
0
014 16 18 20 24 02 04 06 08 10 12
10
20
30
40
50
60
70
36.9
37.0
37.1
37.2
37.3
37.4
37.5
37.6
Melatonin (pg ml–1)Melatonin (pg ml–1)
Core body temperature (°C)
Clock time (h)
014 16 18 20 22 24 02 04 06 08 10 12
10
20
30
40
50
60
70
Melatonin (pg ml–1)
Fig. 2. Diagrammatic representation of some examples of human
melatonin secretion profiles ( ). (a) two major peaks (top), a
peak with a shoulder (middle), a low amplitude unimodal pro-
file (bottom). Note that individual profiles are very reproducible
but that there are major individual differences in both amplitude
and pattern of secretion. (b) The melatonin and core temperature
rhythms ( ) are closely coupled, the nadir of core temperature
occurs within approximately 1 h of the peak of melatonin. (c) Bright
light is required to suppress melatonin completely at night. The
effect on plasma melatonin of three increasing light intensities
( , , ) given from 24:00 h to 02:00 h is shown. Full suppression
is usually found with 2500 lux white light, partial suppression has
been observed with as little as 100 lux, with substantial individual
variations in sensitivity.

There are very large individual variations. Elucidating this
human requirement for bright light has proved to be of funda-
mental importance in our understanding of human circadian
and photoperiodic physiology.
The phase of the free-running melatonin rhythm can be reset
by a single light pulse, with the magnitude and direction of
shift being dependent on the circadian time at which light is
applied. Phase delays are induced by light exposure in the late
subjective day and early subjective night whereas phase ad-
vances follow light in the middle-to-late subjective night and
early subjective day. The direction and magnitude of shift in
response to light is summarized as the phase response curve
(PRC). Such light-induced phase shifts are being exploited to
aid adaptation to shift work, to time zone change and in
pathology (Eastman et al., 1995). It is possible that they depend
partly on suppression of melatonin production for their
efficiency but this remains to be proven.
The change in duration of melatonin secretion as a function
of duration of night is found in the vast majority of species
studied, with the possible exception of domestic pigs. Usually,
the only seasonal change in human melatonin is in phase, with
delayed phase being characteristic of winter in normal healthy
individuals. Two daily light pulses of 2500 lux given as a skel-
eton spring photoperiod in the dim light conditions of winter
in the Antarctic will phase-advance the delayed melatonin to
a summer phase position (Broadway et al., 1987). Thus, the
elements of a photoperiodic response remain in human physi-
ology. If humans are kept for 2 months in 8 h:16 h light: dark,
where the 16 h of darkness are completely dark, it is possible to
show an increase in the duration of melatonin secretion com-
pared with the same individuals kept for 2 months in 14 h:10 h
light: dark (Wehr, 1991).
The clearest abnormalities in melatonin secretion are ob-
served in blind subjects with no light perception (NLP) and in
circadian dysrhythmia during adaptation to phase shift.
Several groups have described both free running rhythms and
entrained but abnormally phased rhythms of melatonin and
other variables in the blind (see references in Lockley et al.,
1997). Czeisler et al. (1995) have shown that, in some patients
with no light perception, circadian entrainment is possible and
that this is linked to the ability of light to suppress melatonin in
these individuals. This phenomenon of ‘hypothalamic light
perception’ is known in animals in which the retinohypothal-
amic projection is intact but the primary and accessory optic
tracts have been sectioned. The nature of the photoreceptors
concerned with circadian responses is as yet unknown but
is under intense investigation. Free-running blind people at-
tempting to live on a 24 h day suffer periodic sleep disturbance
and this occurs mostly when the melatonin rhythm, and indeed
other endogenous circadian rhythms, are 180° out of phase
with the normal cycle.
Role of melatonin in seasonal cycles
The duration of melatonin secretion is the critical parameter
signalling daylength for the organization of seasonal rhythms.
Infusion of appropriate duration profiles of physiological
concentrations of melatonin to pinealectomized hamsters and
sheep has shown the critical role of melatonin duration in sig-
nalling daylength. In the short-day breeding sheep, the daily
infusion of ‘long-duration’ melatonin to simulate long nights
(short days) is inductive to reproductive activity. In contrast, in
the long-day breeding hamster, such infusions are inhibitory
and short-duration melatonin is inductive. The frequency of
melatonin infusions does not appear to be critical over period-
icities ranging from 20 to 26 h but, for an infusion to be read
efficiently, it appears that it must be perceived as a single block
(although there is some recent controversy on this point). For
example, two daily infusions of 4 h and 4 h separated by a 2 h
gap in hamsters are read as 4 h not 8 h whereas, if the gap is
sufficiently small, the two periods are perceived as 8 h. In ad-
dition, photoperiodic history is important in the interpretation
of any melatonin duration signal. For example, transferring
hamsters from 8 h:16 h light:dark to 12 h:12 h light:dark leads
to perception of 12 h:12 h light:dark as a long day, whereas
transferring from 16 h:8 h light:dark to 12 h:12 h light:dark
leads to perception of 12 h:12 h light:dark as a short day
(Hastings et al., 1989).
Karsch’s group have provided evidence in pinealectomized
sheep that the seasonal breeding cycle is an endogenous annual
16 J. Arendt
(a)
Summer
Long days
Short melatonin profile
Winter
Short days
Long melatonin profile
(b)
Summer
Long days
Sleep
Photoperiod
Artificial light
Winter
Short days
Sleep
Photoperiod
Artificial light
Fig. 3. Duration of melatonin secretion response to photoperiod
change in sheep (a) and humans (b). , darkness; , sleep; ,
artificial light (humans). Note that humans rarely show duration
changes unless living in artificial photoperiods with total darkness
for long periods. A small phase delay is usually seen in winter.

rhythm that desynchronizes from 365 days in the absence of the
pineal. It can be resynchronized by a single block of 70 consecu-
tive days of long-day melatonin infusions (Woodfill et al., 1994;
Fig. 4). Thus, the long days of spring are presumably both
necessary and sufficient to cue the entire annual cycle in the
absence of other seasonal time cues. Long days are similarly
essential for the appropriate timing of pubertal development
in sheep.
In ruminants, it is possible to create a winter duration mela-
tonin profile in animals on a natural or artificial summer photo-
period by feeding 3–5 mg melatonin adsorbed onto a food pellet
5–6 h before onset of darkness (Arendt, 1986). If animals are fed
daily from midsummer in this way, they respond as if exposed
to winter photoperiod: that is, by early onset of seasonal repro-
ductive function, winter coat growth and the suppression of
prolactin secretion as found in short winter days (Arendt,
Melatonin 17
Jul JanJul JanJul JanJul
(a) Jan
Jul JanJul JanJul JanJul
(b) Jan
PX
Jul JanJul JanJul
MMM
JanJul
(c) Jan
PX
Fig. 4. Synchronization of the breeding cycle in pinealectomized sheep by a 70 day block of daily 8 h infusions (long-day signal) in spring
from data published by Woodfill et al. (1994), kindly supplied by F. Karsch, University of Michigan. Three groups of sheep are shown:
(a) pineal intact controls; (b) pinealectomized animals (note the desynchronization of the breeding season from the annual cycle); (c) pineal-
ectomized and infused daily for 70 days with a long-day melatonin profile (note the synchronization of reproductive activity). , Increase
of LH, representative of the normal breeding season. This long-day signal is sufficient to set the whole annual cycle in the absence of other
photoperiodic information from melatonin secretion. PX, pinealectomy.

18 J. Arendt
1986). Melatonin mimics the winter photoperiod. The use of
continuous release subcutaneous implants or oral soluble glass
bolus preparations (retained in the rumen) leads to increased
melatonin concentrations throughout the light–dark cycle.
During summer, this is read by body physiology as a ‘super
short’ day and seasonal functions are modified appropriately.
For agricultural purposes, this is a practical approach to in-
duction of early breeding: for early lamb production or ma-
nipulation of other commercially important animal products
such as winter pelage and milk production. However, it is
necessary for the animal to experience a period of long days
before exogenous melatonin will induce a short-day effect. This
phenomenon is known as refractoriness and is common to
photoperiodic responses in many species. Refractoriness is
broken either by the natural long days of spring or by treat-
ment with artificial long days early in the year.
Commercial preparations of melatonin for use in agriculture
for the manipulation of seasonal breeding in sheep have been
developed, and at least one is registered and available in a
number of countries including Australia and the UK. The long-
term results of its use have yet to be fully evaluated.
Effects of melatonin on circadian rhythms
The relationship of the pineal to the circadian system in mam-
mals has already been referred to as modulatory. It is difficult
to make a case for a major role of the pineal or melatonin in
mammalian circadian control. Nevertheless, there are effects of
melatonin on mammalian circadian rhythms that are proving
of considerable therapeutic use.
In rats free running in dark:dark, daily melatonin injections
(5 µg–1 mg) can resynchronize activity–rest cycles to 24 h when
the injection time approaches the time of onset of free-running
activity (Armstrong, 1989). In rats exhibiting a persistent phase
delay of activity–rest cycles in entrained conditions, melatonin
will phase advance the rhythm and, after a forced 8 h advance
phase shift, given at the original dark onset, it can dictate the
direction of re-entrainment. The PRC of activity–rest to mela-
tonin injections in rats shows only phase advances. Similar
effects have been reported in hamsters, although vehicle in-
jections can entrain by ‘arousal’. However, there is clear evi-
dence for entrainment of hamster pups by prenatal melatonin
treatment.
In humans, both pharmacological and physiological doses of
melatonin can induce phase shifts. Early work suggested that
2 mg melatonin given daily in the late afternoon (17:00 h) for a
period of one month was sufficient to induce a phase advance
of the endogenous melatonin rhythm of 1–3 h (when the en-
dogenous and exogenous components could be distinguished)
(Arendt et al., 1985). This was accompanied by an earlier onset
of evening fatigue or sleep and also an earlier timing of the pro-
lactin rhythm. There were no effects on other anterior pituitary
hormones. These observations suggested that melatonin had
‘chronobiotic’ effects on the human circadian system. An acute
sleep-inducing effect and modification of the EEG by pharma-
cological doses of melatonin has been known since the early
1970s. There is now evidence that melatonin shifts sleep time in
delayed sleep phase insomnia, that low dose (5 mg) melatonin
increases sleep propensity in a time-dependent manner, with
an increase in subjective fatigue after acute administration of
physiological amounts of melatonin (< 1 mg) and that there are
acute, dose-dependent advances in sleep timing over the dose
range 0.05–5.0 mg (for a review see Arendt, 1997).
Melatonin will also induce phase delays in humans when
given in circadian early morning, and several human PRCs
have been described (Fig. 5). The melatonin PRC approxi-
mately mirrors the light PRC, lending further credence to the
concept of melatonin as a darkness hormone with respect to
the circadian system as well as in photoperiodism. Given the
PRCs to light and melatonin, and the general tendency of
human circadian rhythms to delay, it is possible that early
morning light resets the endogenous clock daily by phase ad-
vance, and the evening melatonin increase may reinforce this
effect. Thus, it is possible to propose a physiological role for the
pineal and melatonin in circadian organization in mammals.
Whether or not melatonin can entrain all free-running
rhythms remains doubtful (Middleton et al., 1997). In the case
of some blind subjects, melatonin clearly stabilizes sleep–wake
cycles to 24 h; however, simultaneous entrainment of other
rhythms has only been reported in one case. Nevertheless,
when melatonin is used in concert with other time cues to
hasten adaptation to phase shift, it clearly affects endogenous
variables, such as core temperature, and cortisol and endog-
enous melatonin, as well as the sleep–wake cycle.
In humans, the nadir of core temperature occurs within 1 h of
the peak of melatonin together with the peak of fatigue and the
trough in ability to perform certain tasks (reflected in high night-
time accident rates in night shift workers), and this association
may in part be causal. Melatonin (0.5–5.0 mg) acutely sup-
presses core body temperature and Cagnacci et al. (1992) have
calculated that possibly 40% of the amplitude of the rhythm
in core body temperature may be due to night-time melatonin
secretion. The acute suppression induced by a single dose of
exogenous melatonin is dose-dependent, is accompanied by
decreased alertness, and correlates closely with the acute phase
advance induced when given at 17:00 h (Deacon and Arendt,
1995). Importantly, bright light increases body temperature and
alertness when given at night, and this can be reversed by
melatonin. Whether the changes in core body temperature are
relevant to phase-shifting mechanisms remains doubtful.
Target sites of melatonin for the control of
seasonal rhythms
The target sites of melatonin are currently under intense in-
vestigation. Lesions of the mediobasal hypothalamic area can
block photoperiodic effects of melatonin on gonadotrophins
(Lincoln, 1992; Maywood and Hastings, 1995). Implants of
Fig. 5. Two individual subjects showing phase delay (a) and phase advance (b) of the sleep–wake cycle after administration of melatonin
(5 mg day–1) at 20:00 h from day 16. Placebo (identical capsule with lactose filler) was given at the same time on days 1–15. (Redrawn from
Middleton et al., 1997, by permission.)

Melatonin 19
1 + 2
2 + 3
3 + 4
4 + 5
5 + 6
6 + 7
7 + 8
8 + 9
9 + 10
10 + 11
11 + 12
12 + 13
13 + 14
14 + 15
15 + 16
16 + 17
17 + 18
18 + 19
19 + 20
20 + 21
21 + 22
22 + 23
23 + 24
24 + 25
25 + 26
26 + 27
27 + 28
28 + 29
29 + 30
30 + 31
1 + 2
(b)
(a) 2 + 3
3 + 4
4 + 5
5 + 6
6 + 7
7 + 8
8 + 9
9 + 10
10 + 11
11 + 12
12 + 13
13 + 14
14 + 15
15 + 16
16 + 17
17 + 18
18 + 19
19 + 20
20 + 21
21 + 22
22 + 23
23 + 24
24 + 25
25 + 26
26 + 27
27 + 28
28 + 29
29 + 30
30 + 31
12 16 20 24 04 08 12 16 20 24 04 08 12
12
Key: Wake Sleep Activity
acrophase Demasked temperature
acrophase aMT6s
acrophase Melatonin
administration
16 20 24 04 08 12
Time (h)
Time (h) 16 20 24 04 08 12
Trial dayTrial day

melatonin in the hypothalamus suppress LH release in rats.
Implants or infusion of melatonin in the mediobasal hypothala-
mus mimic or block photoperiodic responses in several species
(for example, Lincoln, 1992; Maywood and Hastings, 1995). In
prepubertal rats, melatonin inhibits LHRH-induced LH release
in pituitary cultures at concentrations comparable with those
circulating in the blood. There is also evidence that melatonin
influences LHRH secretion from the hypothalamus in co-
cultures of median eminence and pars tuberalis.
Lincoln and Clarke (1994) have investigated seasonal
rhythms in hypothalamic–pituitary-disconnected Soay rams,
animals in which the pars tuberalis and pituitary are intact but
have no input from hypothalamic neurohormones and trans-
mitters. In such conditions, the tonic inhibitory and presum-
ably dopaminergic control of prolactin secretion is abolished,
as is the circadian rhythm. Nevertheless, the well-characterized
seasonal variations in prolactin concentrations persist and re-
spond to both changing artificial daylength and melatonin
administration. The inference is that melatonin acts at the pitu-
itary for this particular seasonal variation via pars tuberalis
melatonin receptors. However, the gonadotrophic hormones
do not respond and clearly there is scope for seasonal effects at
the hypothalamus as well as at the pituitary.
The development of 2-[
125
I]iodomelatonin as a high specific
activity ligand has permitted the identification of high affinity
(K
d
value 25–175 pmol l
–1
), saturable, specific and reversible
melatonin binding to cell membranes in the central nervous
system, initially in the SCN and the pars tuberalis of the pitu-
itary (for review see Morgan and Williams, 1989). Both periph-
eral and central high-affinity binding sites for melatonin are
found using 2-iodomelatonin as ligand. The most consistent
binding site is within the pars tuberalis of the pituitary gland,
but with frequent binding in the mediobasal hypothalamus, the
retina, the SCN and many other brain sites. There are changes
in detectable binding with age; for example, in fetal rats, the
first appearance of binding is in the pars distalis and pars
tuberalis of the pituitary, with SCN labelling appearing in later
gestation. Pars distalis binding is absent in adult rats but per-
sists after birth in the neonate. This suggests that such binding
may indeed underlie function since melatonin will inhibit
GnRH-induced pituitary LH release in prepuberty but not in
adulthood.
Target sites of melatonin influencing
circadian rhythms
The most obvious target tissue for the actions of melatonin on
the circadian system is the SCN and, in view of its effects
on the mechanisms of light transduction (Iuvone and Gan,
1995), receptors in the retina may also be of importance in
circadian light perception. SCN-lesioned rats do not respond
to melatonin by restoration of activity–rest cycles, although
since the SCN itself is essential for generation and expression
of activity–rest rhythms this experiment is susceptible to dif-
ferent interpretations.
The SCN shows clear melatonin binding in human post-
mortem tissue. The most convincing evidence that the SCN is a
target site for circadian effects of melatonin has derived from
SCN-containing hypothalamic slice preparations in vitro. It is
possible to show that the 24 h rhythm of electrical activity
persists in vitro for several cycles and is phased in relation to
the donor’s previous light–dark cycle. Physiological concen-
trations of melatonin added to the cultures at different cir-
cadian times induce substantial phase advances according to a
phase–response curve resembling that seen in intact animals
using activity–rest as a circadian marker rhythm (Gillette and
McArthur, 1996).
The amphibian melatonin receptor was cloned in 1994, and
cloning of the sheep and human receptors was reported shortly
thereafter (Reppert et al., 1994) with high structural similarity
(80%) between sheep and human clones. Three subtypes are
reported: Mel 1a, 1b and 1c. In situ hybridization studies in
several mammals have shown signals in both the pars tuberalis
and the SCN, and the pharmacology of the recombinant recep-
tors is identical to the endogenous G-protein-linked receptor.
The Mel 1b receptor does not appear to be necessary for photo-
periodic responses in hamsters (Reppert, 1996).
These receptors are members of a new receptor group that is
distinct from other G-protein-linked groups. This work opens
large new perspectives and approaches not only for the study
of the mechanism of action of melatonin but also for the de-
velopment of new molecules for therapeutic use. Initial studies
using ‘knockout’ mice genetically engineered to be devoid of
Mel 1a receptors suggest that phase shifts of the SCN to mela-
tonin in vitro are retained but that the ability of melatonin to
suppress SCN neural activity is abolished (Liu et al., 1997).
Thus, Mel 1a is implicated in inhibition of SCN activity with
the possibility that Mel 1b (present in barely detectable
amounts in these animals) is the phase shifting receptor.
Therapeutic uses of melatonin
Circadian rhythm disorder affects a very large number of
people, including many shift workers and individuals suffering
from jet lag, blindness, insomnia and other problems of old age,
some psychiatric disorders and conditions where natural zeit-
gebers are very weak (for example dim light in high latitude
winters). Circadian rhythm disturbance may also accompany
many other conditions as a secondary effect.
The problems of shift work, where individuals are required
to work during the low point of performance and high point
of fatigue, and to sleep at an inappropriate circadian time
are of real economic importance. Moreover, the health prob-
lems of shift workers include sleep disorder, gastrointestinal
disturbance and increased susceptibility to cardiovascular
disease. All of these may be related to rhythm disruption. In
some respects shift work and jet lag are similar, in others
they are not. For example night shift workers operate through-
out their night shift counter to the natural zeitgebers whereas
time zone travellers adapt with the help of the environment.
The use of increased light intensity at night (> 1000 lux) and
specified phase shifting light regimens is proving to be success-
ful as an aid to adaptation and improved efficiency. However,
not all circumstances allow the use of bright light: it is ex-
pensive to install and maintain and its potential deleterious
effects on the eyes remains fully to be assessed. The use
of melatonin as a phase shifting mechanism offers a con-
venient alternative. The combined use of timed melatonin
and bright light is likely to provide optimum phase shifting
conditions.
20 J. Arendt

Early work in a placebo controlled study indicated that, over
an advance phase shift of eight time zones, suitably timed
melatonin (5 mg) significantly improved night sleep latency
and quality, daytime alertness and hastened the resynchron-
ization of endogenous melatonin and cortisol rhythms. Most
subsequent work has confirmed the behavioural effects in field
studies and the circadian re-entrainment in simulation studies.
The timing of melatonin treatment for westward time zone
transitions poses problems over fewer than eight time zones as,
according to the PRC, it should be taken in the middle of the
night. Moreover, preflight early morning treatment to initiate a
phase delay can lead to loss of alertness at a very inappropriate
time of day. In the author’s experiments, melatonin is not given
preflight but is taken at bedtime after westward flight for
4 days. To date, with 474 people taking melatonin and 112
taking placebo, the overall reduction in perceived jetlag (visual
analogue scale, 100 = very bad jetlag, 0 = negligible jetlag) over
all time zones and in both directions is 50%. Included in these
figures are both placebo controlled and uncontrolled studies,
between which no difference was observed. Side effects reported
more than once are (melatonin–placebo): sleepiness (8.3–1.8%),
headache (1.7–2.7%), nausea (0.8–0.9%), giddiness (0.6–0%),
and light-headedness (0.8–0%).
Melatonin has been used in two small controlled studies of
night shift workers, compared with both placebo and no treat-
ment, and timed to phase delay by administration after the
night shift and before daytime sleep (Folkard et al., 1993; Sack
et al., 1994). It was successful in improving night shift alertness,
sleep duration and quality in one study and improved circadian
adaptation assessed by endogenous melatonin production in
the other: clearly, more data are required, particularly about
the effect of melatonin on work-related performance.
We have developed a means of simulating time-zone tran-
sition or shift work without environmental isolation using
a combination of timed moderately bright light (1200 lux, 9 h
followed by 8 h imposed darkness or sleep) delaying or ad-
vancing by 3 h each day followed by 2 days stable treatment at
the new phase. In this way, it is possible to induce 9 h phase
shifts while maintaining internal synchronization. Subsequently,
subjects are abruptly returned to the local time cues, simulating
time zone transition or rotating shift systems. After an abrupt
phase advance of 9 h, melatonin (5 mg), timed to phase ad-
vance, immediately improves sleep quality and duration to-
gether with alertness and the ability to perform low and high
memory load cancellation tests (Deacon and Arendt, 1996)
(Fig. 6). This improvement is evident before any major phase
shift has occurred in endogenous markers such as core body
temperature. Although resynchronization of endogenous vari-
ables is hastened, the acute effects of melatonin appear to be as
important as any induced phase shift. In the author’s opinion,
melatonin reinforces acutely and chronically physiological
phenomena connected with darkness, in particular sleep.
Melatonin antagonists and agonists
In view of the very large potential market for melatonin-like
effects in both occupational health and pathology, a number
of pharmaceutical companies have initiated development of
specific formulations of melatonin and of novel analogues. The
most interesting napthalenic agonists have similar effects to
Melatonin 21
5
Rapid 9 h advance phase-shift
6 7 8 9 10 11 Mean
baseline
130 Sleep quality
110
90
70
50
30 Alertness
110
100
90
80
Performance efficiency
aMT6s
110
100
90
80
06:00
08:00
10:00
12:00
14:00
16:00
04:00
Acrophase time (h) Daily mean (% of baseline) Percentage of baseline
Melatonin
treatment Day of study
Fig. 6. Melatonin ( ) hastens behavioural adaptation to phase shift
compared with placebo ( ), when taken for 3 days at 23:00 h after
an abrupt simulated 9 h phase advance (equivalent of travelling
from Los Angeles to Paris); subjective sleep quality (and poly-
somnographically recorded sleep, not shown), alertness and per-
formance are improved, before and during an increase in the
rate of adaptation of the endogenous melatonin rhythm, as
assessed by 6-sulfatoxymelatonin (aMT6s) measurement in urine.
Note that beneficial effects are seen long before circadian ad-
aptation is complete, indicating a major contribution of acute
effects of melatonin. (Redrawn from Arendt and Deacon, 1997, with
permission).

melatonin on SCN electrical activity, on in vivo rhythm physi-
ology in rodents (for example, Ying et al., 1996) and on human
circadian rhythms (Cajochem et al., 1995). Besides these ana-
logues, many other potentially useful molecules are under
investigation. The detailed pharmacology of these new antag-
onists is eagerly awaited.
Melatonin has provoked the development of a whole new
pharmacology of so-called chronobiotic drugs. It is also claimed
to be a powerful antioxidant, immunostimulant, anti-ageing
factor and general cure-all. The publicity surrounding these
claims has been detrimental to serious investigation of its
proven therapeutic uses. It is to be hoped that suitable clinical
trials will lead to a more rigorous definition of its uses and
limitations and, most particularly, to its long-term safety.
References
Key references are identified by asterisks.
Arendt J (1986) Role of the pineal gland in seasonal reproductive function
Oxford Reviews of Reproductive Physiology 8 266–320
*Arendt J (1995) Melatonin and the Mammalian Pineal Gland Chapman and
Hall, London
Arendt J (1997) Sleep science: integrating basic research and clinical practice.
Melatonin Karger Monographs in Clinical Neuroscience Vol. 15 pp 196–228
Series Ed. M Fisher, Vol. Ed. W Schwartz. Karger, Basel
Arendt J and Deacon S (1997) Treatment of circadian rhythm disorders –
melatonin Chronobiology International 14 185–204
Arendt J, Bojkowski C, Folkard S, Franey C, Minors DS, Waterhous JM,
Wever RA, Wildgruber C and Wright J (1985) Some effects of melatonin
and the control of its secretion in man. In Ciba Foundation Symposium 117:
Photoperiodism, Melatonin and the Pineal pp 266–283 Eds D Evered and
S Clark. Pitman, London
Armstrong SM (1989) Melatonin and circadian control in mammals
Experientia 45 932–939
Broadway J, Arendt J and Folkard S (1987) Bright light phase shifts the
human melatonin rhythm during the Antarctic winter Neuroscience Letters
79 185–189
Cagnacci A, Elliott JA and Yen SSC (1992) Melatonin: a major regulator of
the circadian rhythm of core temperature in humans Journal of Clinical
Endocrinology and Metabolism 75 447–452
Cajochen C, Kraeuchi K, Moeri D, Hetsch C and Wirz-Justice A (1995) A
single administration of melatonin or the melatonin agonist S-20098
lenthens the first REM sleep episode Sleep Research 24 40
Cassone V (1992) The pineal gland influences rat circadian activity rhythms
in constant light Journal of Biological Rhythms 7 27–40
Czeisler CA, Shanaghan TL, Klerman E, Martens, H, Brotman DJ, Emens JS,
Klein T and Rizzo J (1995) Suppression of melatonin secretion in some
blind patients by exposure to bright light New England Journal of Medicine
332 6–11
Deacon S and Arendt J (1995) Melatonin-induced temperature suppression
and its acute phase-shifting effects correlate in a dose-dependent manner
in humans Brain Research 688 77–85
Deacon S and Arendt J (1996) Adapting to phase-shifts. II. Effects of mela-
tonin and conflicting light treatment Physiology and Behaviour 59 675–682
Eastman CI, Boulos Z, Terman M, Campbell SS, Dijk D-J, and Lewy AJ
(1995) Light treatment for sleep disorders: consensus report. VI. Shift
work Journal of Biological Rhythms 10 157–165
Folkard S, Arendt J and Clark M (1993) Can melatonin improve shift workers’
tolerance of the night shift? Some preliminary findings Chronobiology
International 10 315–320
*Gillette MU and McArthur AJ (1996) Circadian actions of melatonin at the
suprachiasmatic nucleus Behavioural Brain Research 73 135–139
Hastings MH, Walker AP and Powers JB (1989) Differential effects of photo-
peroidic history on the responses of gonadotrophins and prolactin to in-
termediate daylengths in the male Syrian hamster Journal of Biological
Rhythms 4335–350
Iuvone PM and Gan JW (1995) Functional interaction of melatonin receptors
and D1 dopamine receptors in cultured chick retinal neurons Journal of
Neuroscience 15 2179–2183
Klein DC, Coon SL, Roseboom PH, Weller JL, Bernard M, Gastel JA, Zatz M,
Iuvone PM, Rodriguez IR, Begay V, Falcon J, Cahill G, Cassone VM
and Baler R (1997) The melatonin rhythms generating enzyme: molecular
regulation of serotonin N-acetyltransferase in the pineal gland Recent
Progress in Hormone Research 52 307–357
Lincoln GA (1992) Administration of melatonin into the mediobasal hypo-
thalamus as a continuous or intermittent signal affects the secretion of
follicle-stimulating hormone and prolactin in the ram Journal of Pineal
Research 12 135–144
*Lincoln GA and Clarke IJ (1994) Photoperiodically-induced cycles in the
secretion of prolactin in hypothalamo–pituitary disconnected rams: evi-
dence for translation of the melatonin signal in the pituitary gland Journal
of Neuroendocrinology 6 251–260
Liu C, Weaver DR, Jin X and Reppert SM (1997) Molecular dissection of
two distinct actions of melatonin on the suprachiasmatic circadian clock
Neuron 19 1991–2101
Lockley SW, Skene DJ, Tabandeh H, Bird AC, Defrance R and Arendt J
(1997) Relationship between napping and melatonin in the blind Journal of
Biological Rhythms 12 16–25
*Maywood ES and Hastings MH (1995) Lesions of the iodomelatonin-binding
sites of the mediobasal hypothalamus spare the lactotropic, but block the
gonadotropic response of male syrian hamsters to short photoperiod and
to melatonin Endocrinology 136 144–151
Menaker M, Hudson DJ and Takahashi JS (1981) Neural and endocrine
components of circadian clocks in birds. In Biological Clocks in Seasonal
Reproductive Cycles pp 171–183 Eds BK Follett and DE Follett. Wright, Bristol
Middleton B, Stone B and Arendt J (1997) Complex effects of melatonin on
human circadian rhythms in constant dim light Journal of Biological
Rhythms 12 467–477
Morgan PJ and Williams LM (1989) Central melatonin receptors; impli-
cations for a mode of action Experientia 45 955–965
Reppert SM (1996) Nature’s knockout: the Mel1b receptor is not necessary
for reproductive and circadian responses to melatonin in Siberian ham-
sters Molecular Endocrinology 10 1478–1487
Reppert SM, Weaver DR and Ebisawa T (1994) Cloning and characterisation
of a mammalian melatonin receptor that mediates reproductive and cir-
cadian responses Neuron 13 1177–1185
Sack RL, Blood ML and Lewy AJ (1994) Melatonin administration promotes
circadian adaptation to shift work Sleep Research 23 509
Tosini G and Menaker M (1996) Circadian rhythms in cultured mammalian
retina Science 272 419–421
Wehr TA (1991) The durations of human melatonin secretion and sleep re-
spond to changes in daylength Journal of Clinical Endocrinology and
Metabolism 73 1276–1280
*Woodfill CJI, Wayne N, Moenter SM and Karsch F (1994) Photoperiodic syn-
chronisation of a circannual reproductive rhythm in sheep: identification
of season-specific time cues Biology of Reproduction 50 965–976
Ying SW, Rusak B, Delagrange P, Mocaer E, Renard P and Guardiola-
Lemaitre B (1996) Melatonin analogues as agonists and antagonists in the
circadian system and other brain areas European Journal of Pharmacology
296 33–42
22 J. Arendt