Mechanisms of action of melatonin within the central nervous system

AnimalReproduction Science, 30 (1992) 45-65

45

Elsevier Science Publishers B.V., Amsterdam

Mechanisms of action of melatonin within the central nervous system David J. Kennaway ~ and Helmut M. Hugel b aDepartment of Obstetrics and Gynaecology, The University of Adelaide, North Terrace,Adelaide, S.A. 5000, Australia bDepartment of Applied Chemistry, Phillip Institute of Technology, Bundoora, Vie. 3083, Australia

ABSTRACT Kennaway, D.J. and Hugel, H.M., 1992. Mechanisms of action of melatonin within the central nervous system.Anita. Reprod. Sci., 30: 45-65. Melatonin plays a key role in the transfer of photoperiodic information to the neuroendocrine axis. Less clear are the biochemical mechanisms mediating these effects.This review critically examines selectedliterature published during recent years pertaining to putative sites of melatonin action within peripheral and brain tissues. Interactions between melatonin, neurotransmitters and prostaglandins are discussed. Finally, we discuss recent studies on putative melatonin receptors, second messenger systems and the possiblerole of melatonin metabolitesin the mediation of pineal action. INTRODUCTION There is unequivocal evidence that melatonin, through its impact on the timing o f p u b e r t y and adult breeding activity, plays a key role in the control o f reproductive events in a wide range o f species. Its action is apparently unique, as the same signal can both stimulate and inhibit reproductive activity, depending on the species. There is currently no consensus about what aspect o f the melatonin rhythm is o f importance in transferring p h o t o p e r i o d information to the endocrine glands. The duration o f melatonin secretion is important since (a) most species studied produce melatonin for proportionately longer periods during the short days o f winter c o m p a r e d with s u m m e r (Rollag et al., 1978; Illnerova and Vanecek, 1980), and ( b ) seasonality o f reproductive function in pinealectomised animals can be restored b y prog r a m m e d infusions o f melatonin which vary only in duration and not amplitude (Carter and G o l d m a n , 1983; Bittman et al., 1983 ). There are, however, anomalies which are difficult to reconcile with simple duration, for example, the observation that 13 h o f darkness is either stimulatory or inhibitory in Djungarian hamsters ( H o f f m a n et al., 1986) and sheep ( R o b i n s o n and Karsch, 1987 ), depending on whether the previous p h o t o p e r i o d was o f 10 or 16 h duration. There is n o w a better understanding o f the control of melaCorrespondence to: D.J. Kennaway, Department of Obstetrics and Gynaecology, The University of Adelaide, North Terrace, Adelaide, S.A. 5000, Australia. © 1992 Elsevier Science Publishers B.V. All rights reserved 0378-4320/92/$05.00

46

D.J. KENNAWAYAND H.M. HUGEL

tonin production by the pineal gland, and the consequences of pineal ablation and administration of melatonin in many animals, but the biochemical actions of melatonin remain poorly understood. This review provides an interpretation of recent studies on the possible mechanisms of action of melatonin in relation to its effect upon the reproductive organs of various species. SITES OF ACTION Melatonin clearly affects the activity of the reproductive axis in many diverse species, but whether melatonin acts directly on gonadal tissue to either activate or inhibit function, or at the level of the hypothalamus, is still not clear, due to the employment of pharmacological dosages of melatonin and the lack of critical assessment of specificity. Evidence suggesting that melatonin can suppress steroidogenesis at the level of the ovary and testis has appeared sporadically over many years, (e.g. MacPhee et al., 1975; Ng and Lo, 1988). Dose-response studies (D.J. Kennaway, P. Blake and H.M. Hugel, unpublished results, 1988), (Fig. 1 ) suggest that the phenomenon may be pharmacological, since concentrations in the order of 10-7_ 10 -6 M are required to elicit the in vitro response, whereas circulating levels of melatonin are usually in the order of 10- 1 0 10-- 9 M during the dark phase. Nevertheless, much of our knowledge about the biological effects of melatonin is derived from experiments in which doses of 0.1-1.0 mg kg -1 melatonin were employed. The blood melatonin concentrations in such experiments commonly exceed 10 -6 M during the distribution and early elimination phases (Lang et al., 1983; Cassone et al., 1986; Vaughan et al., 1986). The possibility should not be ignored that many studies intended to demonstrate the physiological effects of melatonin instead generated results due to transient pharmacological suppression of gonadal steroidogenesis. Specificity studies in vitro with indoles related to melatonin have provided interesting results; in vitro assays measuring LH-stimulated testosterone production by Leydig cells indicated the importance of the 5-methoxy group and the relative unimportance of Nacylation (Ng and Lo, 1988) (Fig. 1 ) (D.J. Kennaway et al., unpublished results, 1988 ). This is in direct contrast to results obtained with the putative brain receptors for melatonin (Fig. 2 ). Direct actions on gonadal tissue clearly cannot be ignored, but it is generally thought that brain sites are a more likely target of endogenous melatonin. In addressing the question of site of action of melatonin, it is worth considering the fate of the secreted hormone after it leaves the pineal. For many years, it was believed that melatonin was secreted into the blood and that cerebrospinal fluid (CSF) levels were either equal to or lower than blood. Recent re-evaluation of this question by Shaw et al. ( 1989 ) and Kanematsu et al. ( 1989 ) has shown unequivocally that in the sheep and goat, melatonin is found in ten-fold higher concentration in the lateral ventricles than in the

MODE OF ACTION OF MELATONIN IN CENTRAL NERVOUS SYSTEM

~

80_L_

I.o

600no_ 1 4 0 _ ILl Z 0 t~ 2 0 uJ p-

0 cO wpi, 0

\ \ \ \ \ \ ~tMT

j

",,j ",,,j

47

\ \ \

,,j ",,,j \1

aB,

\ \ \ \ \

\ \ \ aCT

aFT

DH.aMT

~80-

I-~ 60u.I ~ 40tZ ~ 20w aMT

pMT

bMT

aFoMK

aMK

Fig. 1. The effect of melatonin analogues and metabolites on LH-stimulated testosterone production by rat Leydig cells in vitro. Leydig cells ( ~ 1.5 × 105cells) were incubated for 3 h in the presence of 10 ng ovine LH and test compounds at concentrations ranging from 10 - 6 t o 10 - 4 M. At the completion of the incubation period, the ceils were killed and testosterone in the medium determined by RIA. The results are presented as mean + SEM percentage inhibition of testosterone for four independent experiments. Rq, 10-6 M; D, 10-5 M; [] 10-4 M. Abbreviations: aBT, 5-bromo-N-acetyltryptamine; aCT, 5-chloro-N-acetyltryptamine; aFoMK, Nacetyl-N 1-formyl-5-methoxykynurenamine; aFT, 5-fluoro-N-acetyltryptamine; aMK, N-acetyl5-methoxykynurenamine; aMT, melatonin; bMT, 5-methoxy-N-butyryltryptamine; DH.aMT, 2.3-dihydromelatonin; pMT, 5-methoxy-N-propionyltryptamine. cisterna magna or jugular blood. It would follow that concentrations in the third ventricle would be o f equal or greater magnitude and as a consequence, circumventricular organs would be exposed to these high levels. Although it is not known h o w the high C S F / b l o o d gradient is maintained, this finding alone points to the brain in general and areas in close contact with CSF (e.g. suprachiasmatic nuclei ( S C N ) ) in particular as likely sites o f action. Since the discovery o f melatonin, there have been many attempts to identify the organs which concentrate radiolabelled melatonin. Areas o f high uptake have been presumed to represent the site o f action or metabolism. The early studies used very low specific activity [I4C]- or [3H ]melatonin (Anton-Tay and Wurtman, 1969; Cardinali et al., 1973 ); nevertheless the brain and the hypothalamus in particular were found to accumulate melatonin. Un-

48

D.J. K E N N A W A Y A N D H . M . H U G E L

LEYDIG CELL TESTOSTERONE Z 100O m -1_z tu 0

125 I - MELATONIN BINDING

3H- DIAZEPAM BINDING

... 2,3DH.aMT

2 , 3 D H . a M T /"~~

80"

N'~./- aFo MK aMT.. .~ j ' ' "

2,°..aii12,

60-

aMT ....

40-

aMK ......

\

w a.

aMT ...... "

20-

b

a

(~

l

-12

I

I

-10

I

i

-8

I

I

-6

I

I

-4

I

I

~-12

I

I

-10 LOG

I

I

-8

I

I

-6

I

I

-4

I

I

%12

I

I

-10

I

-8

I

I

-6

I

-4

CONCENTRATION (M)

Fig. 2. The effect, of melatonin, melatonin metabolites and analogues on the inhibition of testosterone production in rat Leydig cells, inhibition of [3H ] diazepam binding to rat cortical membranes and inhibition of [ '25I] iodomelatonin binding to chicken brain membranes. (a) increasing concentrations of test compounds were added to freshly isolated rat Leydig cells ( 105 cells per 200/~1). The cells were incubated for 3 h in the presence of 10 ng LH and the testosterone content of the medium measured by radioimmunoassay. The data shown are mean percentage inhibition of testosterone production from four independent experiments. (b) increasing concentrations of test compounds were added to crude rat cortical membrane preparations as per Marangos et al. (1980). The membranes were incubated for 30 rain at 4°C in the presence of 1.5 nM [3H]diazepam and specific binding assessed following filtration through glass fibre filters. The data shown are the mean percentage inhibition of the specific binding from four independent experiments. (c) Increasing concentrations of test compounds were added to crude adult chicken brain membrane preparations. The membranes were incubated for 120 rain at 26 °C in the presence of 20 pM [ 125I] iodomelatonin and specific binding assessed following centrifugation. The data shown are the mean percentage inhibition of specific binding from four independent experiments. Abbreviations as for Fig. 1 and aMT.21, 2-iodomelatonin. f o r t u n a t e l y , this t y p e o f e x p e r i m e n t is a f f e c t e d b y the e x t r e m e l y r a p i d h e p a t i c m e t a b o l i s m o f the p e r i p h e r a l l y i n j e c t e d h o r m o n e . T h o s e studies r e p o r t i n g int r a c r a n i a l a d m i n i s t r a t i o n o f m e l a t o n i n s h o w e d clearer e v i d e n c e o f h y p o t h a l a m i c u p t a k e (e.g. C a r d i n a l i et al., 1973 ). M o r e o v e r , local b r a i n a d m i n i s t r a t i o n results in the synthesis o f m e l a t o n i n m e t a b o l i t e s b y b r a i n e n z y m e s ( H i r a t a et al., 1 9 7 4 ) . A m o r e r e c e n t a p p r o a c h to the q u e s t i o n o f specific sites o f a c t i o n o f transm i t t e r s a n d h o r m o n e s w i t h i n the b r a i n has i n v o l v e d the use o f in v i t r o autor a d i o g r a p h y . W h e n u s e d w i t h [ 125I ] 2 - i o d o m e l a t o n i n the t e c h n i q u e has ident i f i e d the S C N a n d p a r s t u b e r a l i s ( P T ) as the m a j o r sites o f a c c u m u l a t i o n o f r a d i o a c t i v i t y . T h e r e v i e w o f S t a n k o v a n d R e i t e r ( 1 9 9 0 ) is a g o o d s o u r c e o f r e f e r e n c e s for the o t h e r b r a i n areas in w h i c h 'specific' labelling o f tissue occurs. It is clear t h a t t h e r e are species d i f f e r e n c e s a n d i n t e r l a b o r a t o r y i n c o n sistencies; n e v e r t h e l e s s , labelling o f t h e S C N a n d P T is the m o s t c o n s i s t e n t finding. It has p r o v e n v e r y difficult to d e t e r m i n e sites o f m e l a t o n i n a c t i o n using in

MODE OF ACTION OF MELATONIN IN CENTRAL NERVOUS SYSTEM

49

vivo experiments, mainly because the physiological response to photoperiod change or to melatonin administration can take from 3 to 12 weeks to become evident. The two main approaches have been ( 1 ) localised administration of melatonin using injections, implants or infusions, and (2) ablation of brain areas thought to be melatonin targets. Glass and co-workers conducted an elegant series of experiments in the white-footed mouse (Peromyscus leucopus) which provide compelling evidence that anterior hypothalamic (AH) and SCN are melatonin targets. In the initial study, melatonin/beeswax pellets implanted in the AH and SCN decreased the weight of the female reproductive tract in animals maintained on a photostimulatory 16L/8D lighting regime. Subsequent studies using timed daily injections (Glass and Lynch, 1982) or infusions (Dowell and Lynch, 1987) in the region of the SCN confirmed the site as a possible target for melatonin, at least in this species. Despite these results, the SCN have tended to be discounted as being essential for melatonin action in hamsters on the basis of a study by Bittman et al. ( 1979 ). Destruction of the SCN prevented testicular regression in hamsters injected once daily with melatonin, but not when melatonin was injected three times daily. Bittman et al. (1979) however stated that their findings "do not eliminate the possibility that under physiological conditions melatonin acts on the SCN to facilitate the process of testicular regression". A pharmacological effect of melatonin is also possible in this study since the dose used (25 #g) at 3-hourly intervals would be expected to produce melatonin blood levels 100-1000 times the normal night-time values and within the range of direct inhibition of testosterone production. In contrast, melatonin-filled capsules implanted into hamsters retarded testicular recrudescence in intact and pinealectomised animals, but not in animals with SCN lesions (Rusak, 1980). One of the explanations for this result was that the antigonadal action ofmelatonin may depend on its uptake in the suprachiasmatic region. More recently, Maywood et al. (1990) also tested the hypothesis that the SCN are an essential component of the neural circuits that measure the melatonin signal. Their experimental design overcame problems associated with the loss of normal pineal melatonin secretion in SCN-lesioned animals by using programmed infusions of melatonin into pinealectomised male Syrian hamsters bearing SCN lesions. Infusion of 250 ng melatonin every 10 h for 35 days induced gonadal atrophy in sham-lesioned and SCN-lesioned hamsters after 5 weeks. In contrast, SCN lesions prevented the induction of shortday responses by 10-h melatonin infusions in Djungarian hamsters (Bartness et al., 1991 ). There may be methodological reasons for the apparent species differences (i.e. the extent of lesion damage to other hypothalamic structures in the study by Bartness et al., 1990). Nevertheless, despite the elegance of the methodology, the issue remains to be resolved. Mess and his colleagues (Mess and Ruzsas, 1987) have used a different

50

D.J. KENNAWAYAND H.M. HUGEL

experimental model which provides interesting if circumstantial evidence of melatonin interactions with the SCN. Female rats maintained in constant light for up to 120 days develop a constant oestrous/anovulatory (CEA) state. This develops presumably because of the loss of a photoperiodic circadian signal. An obvious explanation may be the obliteration of the melatonin rhythm by constant light and this was addressed by Mess and Ruzsas ( ( 1987 ). Single daily injections of melatonin were observed to either prevent or reverse the CEA state induced by constant light. It cannot be stated definitely that melatonin acts in this instance directly upon the SCN in preference to other brain centres projecting to the SCN, but such a relationship is worth further investigation. Additional evidence supporting interactions between melatonin and the SCN is available from experiments exploring the effect on the circadian rhythm of wheel running in rats. Daily injections of melatonin entrained running activity in rats kept in constant darkness but only when the time of injection coincided with activity onset in free-running animals (Redman et al., 1983 ). Moreover, melatonin administration has been shown to accelerate the re-entrainment of running activity (Redman and Armstrong, 1988 ), the melatonin rhythm (Kennaway et al., 1989), and pineal NAT rhythm (Illnerova et al., 1989 ) following a phase shift of the light/dark cycle. These actions of melatonin clearly involve interactions with the SCN. Whilst the monitoring of wheel running has provided important insight into circadian rhythmicity, it has not generally been studied in association with cyclic reproductive phenomena. Puchalski and Lynch ( 1988 ), however, recently reported results of experiments addressing the effect of daily late afternoon melatonin injections into hamsters who had access to running wheels. Analysis of activity rhythms after 10 weeks of treatment indicated two distinct responses to melatonin administration. The majority of hamsters had a 1.5 to 2-h phase advance of their activity rhythm, whilst the remaining hamsters responded to melatonin injections by free-running their rhythms (i.e. they failed to stay entrained to the light/dark cycle). Hamsters with phase advanced but entrained activity rhythms were found to have responded physiologically to melatonin by undergoing gonadal collapse. By contrast, those animals free-running in response to melatonin treatment were also apparently resistant to the antigonadal effects of melatonin and maintained large sexual and accessory organ weights. While these findings are yet to be confirmed in other species, the results provide compelling evidence for melatonin/SCN/gonadal interactions. In keeping with the proposed role of the SCN as a major rhythm generating centre, there is now extensive evidence that electrophysiological (Green and Gillette, 1982) and biochemical events (Schwartz et al., 1983) within the nuclei change predictably during the day. The SCN are the only part of the brain capable of sustaining a circadian rhythm of electrical activity when isolated from the rest of the brain, and evidence is accumulating which indicates

MODE OF ACTION OF MELATONIN IN CENTRAL NERVOUS SYSTEM

51

that melatonin administration in vivo and in vitro can alter intrinsic SCN functions. Cassone et al. (1988) injected melatonin into rats 2 h prior to darkness, followed 15 min later with an injection of 2-deoxy- [ 1-~4C] glucose. This resulted in a reduction of glucose uptake to a level normally observed 2 h after darkness, i.e. at the onset of melatonin secretion. Subsequent experiments testing the dose-response effects of melatonin indicated that doses as low as 1.5/tg kg- 1were effective in decreasing glucose utilisation (Cassone et al., 1988 ), compared with the dose of 5/tg kg- 1 day- 1 melatonin required to entrain circadian running activity in the rat (Cassone et al., 1986). A similar circadian time-dependent response to melatonin has been observed with spontaneous electrical activity in the SCN of the rat and Syrian hamster in vitro. Typically, the neuronal firing rate is highest 6 h prior to darkness and lowest during mid-darkness with an amplitude of two to four impulses p e r second. When melatonin was added to the bathing solution (Shibata et al., 1989) or iontophoretically applied (Stehle et al., 1989), it consistently inhibited neuronal firing rate in a dose-dependent manner 0-3 h prior to darkness. Mason and Rusak (1990) have recently extended the studies to test whether hamsters exhibiting photorefractoriness to short days exhibit altered neuronal sensitivity to melatonin. Suprachiasmatic nuclei obtained from hamsters maintained on long days (LD14:10) had a higher proportion of cells suppressed by melatonin than those obtained from animals kept for 150-190 days in short day length (LD10:14). Melatonin-responsive cells were found primarily during the late projected day and early projected night in both long-day and short-day animals. A feature of SCN rhythmicity is the persistence of the circadian rhythm of neuronal firing rate (Green and Gillette, 1982 ) and vasopressin rhythmicity (Earnest and Sladek, 1986) during prolonged culture in vitro. McArthur et al. (1989) found that a 1 h exposure of isolated SCN to 10 -9 M melatonin, 2 h prior to subjective darkness induced a significant advance (3.8 _+0.1 h) in the time of peak neuronal firing the following day. In contrast, melatonin applied 10 h before subjective darkness failed to alter the rhythm of neuronal firing. In a further examination of this effect, McArthur and Gillette (1990) showed that prior exposure to pertussis toxin blocked the melatonin-induced phase advance. Other sites

The median eminence region of the brain has been implicated as a site of action of melatonin for many years. In vitro autoradiography studies confirmed that the region was a significant area of [ 125I ] 2-iodomelatonin concentration (Vanecek et al., 1987 ), although careful histological investigations have identified the cells involved as PT cells. Most of the physiological evidence that the median eminence ( M E ) / P T is a site of action for melatonin

52

D.J. KENNAWAY AND H.M. H U G E L

has come from in vitro studies of the inhibition of prostaglandin production (Franchi et al., 1987 ). The PT is a particularly interesting, but poorly studied region which has cells with testosterone stimulating hormone (TSH)-like immunoreactivity that change morphologically with alterations in photoperiod (Wittowski et al., 1984, 1988 ). There are also cells in this region which contain luteinising hormone (LH) and follicle stimulating hormone (FSH) and changes in cell content are evidently influenced by medial preoptic area inputs and vary during the oestrous cycle (Aguado et al., 1982). It is possible that the PT cells are in some way involved in thyroid hormone interactions with the reproductive axis, since seasonal changes in the cell morphology have been observed. At this time, there is no clear concept of how the PT (if it is an important melatonin target) does alter the gonadal axis. A pituitary site of action was suggested by Martin and Klein (1976) following studies showing that melatonin inhibited luteinising hormone stimulating hormone (LHRH)-stimulated LH release from 5-day-old rat pituitaries in vivo and in vitro at nanomolar concentrations. The response was apparently absent in adult rats and neonatal hamsters and the significance of this response is difficult to incorporate into overall understanding of melatonin/ pineal function. Nevertheless, interest in this area has been stimulated recently by identification of binding sites for melatonin in neonatal pituitaries (Vanecek and Jansky, 1989 ) and the observation that nanomolar concentrations of melatonin inhibit cAMP production in the glands (Vanecek and Vollrath, 1989 ). In summary, there is considerable evidence that a primary site ofmelatonin action is located in the brain and in the SCN region in particular. The potential importance of melatonin/SCN interactions is obvious considering the extensive efferent connections with other major hypothalamic centres involved in the control of reproductive function, such as the preoptic area and arcuate nucleus. Since the SCN is also involved in the generation of the melatonin rhythm, light and melatonin may interact within the same nucleus to maintain an appropriate duration of melatonin production, and consequently the appropriate reproductive state. MECHANISMSOF ACTION Transmitter systems

One of the earliest proposed mechanisms of action of melatonin involved the modulation of gamma amino butyric acid (GABA) nerves. Anton-Tay (1974) first reported that melatonin injections increased synthesis of GABA in the brain as a result of increased pyridoxal phosphokinase activity. Little progress was made in this area until Rosenstein and Cardinali (1986) confirmed that acute administration of 25-300 #g kg-1 melatonin to rats increased GABA accumulation in the hypothalamus, cerebellum, cortex and pineal gland. The same group found that in vivo and in vitro melatonin in-

MODE OF ACTION OF MELATONIN IN CENTRAL NERVOUS SYSTEM

53

creased the numbers of both the high- and low-affinity GABA receptor sites in rat cerebral cortex (Acuna Castroviejo et al., 1986). There is unfortunately no information about the mechanisms involved in these changes, or whether they are related. A special feature of the GABA receptor is the presence of an additional binding site for benzodiazepines; occupation of these sites allosterically alters the GABA receptor complex and potentiates the effects of GABA occupancy of its receptor, i.e. chloride ion flux is increased. Marangos et al. ( 1981 ) made the observation that diazepam binding to the benzodiazepine binding site on the GABA receptor was inhibited by melatonin. Subsequently, injections of melatonin were found to reverse the decrease in benzodiazepine binding site density following pinealectomy (Lowenstein et al., 1985 ). These interactions with GABAergic neurons may explain some of the pharmacological actions of melatonin, e.g. its anti-convulsive, sedative and hypnotic activities, but they also suggest a physiological role via hypothalamic GABA neurons which play a major role in the control of reproductive function. For example, GABAergic neurons are implicated in the steroid-dependent modulation of LH pulse frequency, in the control of prolactin and in SCN function (particularly the entrainment of rhythms). There has been considerable interest recently in melatonin/dopamine relationships, in particular the ability of melatonin to inhibit dopamine release from brain tissues in vitro. In these studies, brain slices were preloaded with labelled dopamine and the release of dopamine following field stimulation monitored. The greatest inhibition of dopamine release by melatonin occurred in the preoptic region of the hypothalamus (which contains the suprachiasmatic nucleus) followed by the ventral and dorsal hippocampus (Zisapel et al., 1982 ). The medulla, ports, cerebellum, cerebral cortex and striatum slices were unaffected by melatonin. Inhibition of dopamine release from the hypothalamus is not constant during the day; at a dose of 10- 5 M, melatonin inhibited dopamine release by 30% 9 h before darkness, but was ineffective 1 h after onset of darkness (Zisapel et al., 1985 ). The discrepancy between the time of maximal dopamine sensitivity and the time of maximal inhibition of SCN neuronal firing rate and metabolic activity cannot be explained at this stage. Little is known about the cellular mechanisms involved in the inhibition of dopamine release.

Prostaglandins Studies on the interactions between melatonin and prostaglandins have concentrated on the relationship between the prostanoids and L H R H release from LHRH-secreting neurons. Intraventricular PGE2 clearly stimulates LH release, whereas inhibitors of prostaglandin synthesis (e.g. indomethacin) decrease LH levels. Indomethacin and melatonin are both 5-methoxyindoles and several groups have shown that melatonin can inhibit PG synthesis in

54

D.J. KENNAWAY AND H.M. HUGEL

vitro. In the early experiments, very high ( 10- 3 M ) levels of melatonin were required to demonstrate inhibition of bioassayable prostaglandin from uteri or hypothalami. Subsequently inhibition of in vitro PGE release by rat medial basal hypothalami was observed with 10 -8 M melatonin concentrations (Franchi et al., 1987). There have been few published attempts to demonstrate in vivo interactions between melatonin and prostanoids. Leach et al. (1982) administered melatonin intracisternally to rabbits and blocked the cervical stimulation-induced rise in CSF PGE levels and prevented ovulation. The PGE released following cervical stimulation was presumed to reflect production in the medial basal hypothalamus-median eminence. The specificity of the melatonin action on prostaglandin synthesis has not been extensively studied, but 5-methoxytryptamine is more potent than melatonin and its 6-halogenated analogues whereas serotonin and 6-hydroxymelatonin are ineffective (Franchi et al., 1987 ). The high potency of 5-methoxytryptamine must cast some doubt upon the physiological significance of the phenomenon because in vivo the N-acetyl group distinguishing melatonin from 5-methoxytryptamine is essential for most of melatonin's reproductive effects. For example, puberty onset in male rats was not delayed by the melatonin analogues N-propionyl-5-methoxytryptamine and N-butyryl-5-methoxytryptamine (Kennaway et al., 1988 ). Moreover, binding studies utilising [ 125I] 2-iodomelatonin consistently indicate poor binding of 5-methoxytryptamine to the binding sites. CELLULARFACTORSMEDIATINGIN THE ACTION OF MELATONIN

Receptors It has been presumed that melatonin binds to a receptor and in some way alters the physiological functions of the target cells. Over 10 years ago, reports appeared describing specific binding of [3H]melatonin to brain membrane (Cardinali et al., 1979 ) and ovarian cytosol preparations (Cohen et al., 1978 ), but these original observations have proven difficult to replicate, even for the original authors. The reasons why it has proven so difficult to repeat the work are not known, although instability of the labelled ligand has often been cited. This 'instability' is not observed in laboratories performing melatonin radioimmunoassays (RIA) using various highly specific and sensitive antisera. If there was a problem with the ligand for RIA, special procedures would be required for handling and storing the ligand. In our laboratory, [ 3H ]methoxymelatonin (notoriously bad in receptor studies according to Niles (1987) ) has been stored for over 12 months at - 10 °C without loss of binding activity or increase in non-specific binding to a highly specific antiserum (D.J. Kennaway, unpublished results, 1991 ) or to chicken brain membranes (Fig. 3).

MODE OF ACTION OF MELATONININ CENTRAL NERVOUS SYSTEM

IODINATED MELATONIN 6000.

yTotal

4500.

/~pf~cifi,

55

TRITIATED MELATONIN

o t a l

-1000 Specific

//

600

o ooo 11

1 5 0 0 ~ N S

215 50

E

16-

"~ LU rr LL 7, Z O ..J iii _L

1284-

1130 ,pM

~

B

200

KD

0 c Z 0

NSB 200

goo lobo pM 20'00

20.8 pM 10 fmoVg

150 300 J 6;0 I-MEL BOUND (frnol/gm)

0.8

I I I' I 150 300 460 600 H-MEL BOUND(fmol/gm)

Fig. 3. Binding isotherms and Scatchard plots for chicken brain membrane preparations incubated with either [ 1=51]iodomelatonin or [ 3H] melatonin. Chicken brains were homogenised in Tris-HC1 buffer and washed three times and tubes containing either 2 mg or 10 mg wet weight brain tissue incubated for 120 rain with 5-150 pM [125I]iodomelatonin or 62-1500 pM [3H]melatonin, respectively. Specific binding was determined by incubating the membranes with 1/tM melatonin. The breakthrough in melatonin receptor research came with the synthesis o f [ 125I ] 2-iodomelatonin by Vakkuri et al. (1984) initially for use in RIA. It was subsequently recognised that 2-iodomelatonin was a potent melatonin agonist. Since the initial studies o f Laudon and Zisapel (1986) and Duncan et al. ( 1988 ), many groups have reported binding o f the radioligand to brain, retinal and pituitary membrane preparations in a wide variety o f species. The binding o f iodinated melatonin to some tissue preparations (rat, Syrian and Djungarian hamster whole brain) is characterised by low-affinity (nanomolar), rapid association/dissociation rates and at least in the hamsters, with specificities which are not consistent with physiological knowledge (Duncan et al., 1988, 1989). For example, N-acetyl-5-hydroxytryptamine (N-acetyl serotonin) has a Ki less than melatonin in hamster whole-brain preparations, although there is no evidence that it has biological activities similar to melatonin. In contrast, preparations o f M E / P T tissue from sheep and hamsters

56

D.J. KENNAWAY AND H.M. H U G E L

(Morgan et al., 1989c; Williams et al., 1989) have extremely high affinity (KD - - 20 pM), slow association and dissociation times and specificity of binding consistent with the known pharmacology of melatonin analogues. Because melatonin levels in blood and CSF usually range from 30 to 1000 pM, it has been presumed that the high-affinity binding sites represent the functional melatonin receptors. The restricted distribution of the putative melatonin receptors unfortunately makes them extremely difficult to study; for example, the rat SCN weighs approximately 300/lg, rat ME/PT 1-2 mg and sheep ME/PT about 20 mg. It is not surprising then, that little is currently known about the physical characteristics of the receptor protein in animals where clearly defined melatonin effects are observed. Studies by Rivkees et al. (1989) on a melatonin-binding protein from the brain ofAnolis carolinensis have nevertheless provided some preliminary information. These authors presented data consistent with agonist-promoted guanine nucleotide-binding protein coupling and estimated the apparent molecular size of the putative receptor uncoupled from G protein to be about 110 000. Second messenger systems

Binding studies provide valuable but restricted information regarding mechanisms of action, with specificity and affinity of the interactions being the most useful. If the binding site is a true receptor rather than an acceptor, the interaction between the ligand and receptor protein must initiate some type of intracellular response. This may involve interactions with ion channels or the stimulation or inhibition of the synthesis of specific protein(s) as occurs in the binding of many other hormones to their receptors. Occupation of melatonin-binding sites has not been shown to be associated with any specific intracellular response. There is however some evidence of interactions between the melatonin-bindingproteins and guanine nucleotide-bindingproteins (G proteins) which are often involved in transducing signals to second messenger systems to effect intracellular changes. The guanine nucleotide-induced changes in [ x25I] 2-iodomelatoninbinding have nevertheless been rather modest and often inconsistent. Rivkees et al. (1989) showed that GTP altered the Bma x of lizard brain membranes without altering the affinity constant. Dubocovich et al. ( 1989 ) failed to alter binding ofiodinated melatonin to chicken brain membranes with GTP, whereas Sugden and Chong ( 1991 ) reported inhibition of binding. In ovine PT tissue GTP analogues altered receptor density but not affinity (Morgan et al., 1989a). Incubation of rat area postrema and suprachiasmatic nuclei with GTP has indicated that G-proteins may be involved in melatonin binding in these tissues also (Laitenen and Saavedra, 1990; Laitenen et al., 1990). The concept of G-protein involvement in melatonin binding is clearly predicated on high-affinity binding sites

MODE OF ACTION OF MELATONIN IN CENTRAL NERVOUS SYSTEM

57

being converted to 'inactive' low-affinity sites, but these lower affinity sites are not universally found. Even when low-affinity sites are demonstrated, generally on the basis of upward curved Scatchard plots, it is worth considering that explanations other than hypothetical second sites are possible (Kermode, 1989). If receptor occupation by melatonin does trigger cellular responses, then, based upon evidence with other receptors, second messenger systems are likely to be activated. To date the ovine PT is the best studied tissue, because it can be prepared in relatively pure form in high yields. Morgan et al. (1989b) found that addition of forskolin (1/zM) to PT cells stimulated cyclic AMP accumulation 12-fold, and that melatonin inhibited this response in a dosedependent manner with 50% inhibition occurring at a dose of 6 pM. In other species, ME explants which contain the PT have been used with qualitatively similar results obtained, although the doses of melatonin required to elicit a 50% decrease in cAMP have generally been approximately 1000-fold higher (Vanecek and Vollrath, 1989; Carlson et al., 1989). Forskolin increases cAMP accumulation in tissue homogenates of ovine PT and in hamster hypothalamus, however melatonin either failed to alter cAMP accumulation (Morgan et al., 1989b ) or had a very minor inhibitory effect over a wide range of doses (Niles and Hashemi, 1990). There is currently no clear explanation why melatonin is ineffective in homogenates.

Physiological changes in receptors There have been few studies reporting changes in high-affinity binding sites under different physiological conditions. Laitenen et al. ( 1989 ) found a variation in labelling density over a 24 h period in the rat SCN with the lowest number of apparent binding sites prior to darkness and the highest just prior to light onset. This is not the pattern one would necessarily expect, given the well-documented sensitivity of animals to melatonin 0-3 h before darkness and insensitivity outside these times. Whether endogenous melatonin occupying the binding sites affected these results is unclear; the well-documented slow dissociation rate from SCN and the lack of extensive prewashing of the slides, however, would suggest that the number of binding sites during the night were underestimated. The positive physiological response of animals to late afternoon melatonin injections as opposed to ineffective morning injections suggested that there might be a rhythm in melatonin receptor density. In addition, morning melatonin injections block the effects of afternoon injections, and this phenomenon had previously been suggested to involve receptor down-regulation. Anis and Zisapel (1989) tested the effect of chronic morning and late-afternoon melatonin injections on melatonin binding to various hamster brain regions. Daily injections (morning or afternoon) failed to alter binding in the hypo-

58

D.J. KENNAWAYAND H.M. HUGEL

thalamus whereas binding in the mid-brain decreased regardless of the time of injection. Binding of melatonin to hippocampal membranes was decreased at all times recorded following morning injections, whereas afternoon injections decreased binding only the following morning. The lack of changes within the hypothalamus and the very low KD of the binding sites poses doubts regarding the significance of these results. Several groups have begun to investigate the relationship between day length and melatonin-binding sites. Vanecek and Jansky (1989) found that the number of binding sites 1-2 h before darkness in Syrian hamster M E / P T decreased by over 50% 18 weeks after a change from 14L/10D to 8L/16D. In contrast, Weaver et al. ( 1990, 1991 ) failed to detect any alterations in melatonin binding to M E / P T of Siberian hamsters housed in 16L/8D or 10L/ 14D. The animals in 10L/14D were photorefractory, i.e. gonadal recrudescence had occurred despite the previously inhibitory short day length. The authors also found that melatonin inhibited forskolin-induced cAMP accumulation in the photorefractory state. Clearly this is an area which still requires more investigation. METABOLISM

The cellular effects of occupancy of the putative melatonin receptors have been difficult to determine. Could the difficulty in studying these effects be an indication that melatonin, like androgens, estrogens and thyroxine is metabolised intracellularly to a more potent hormone? This hypothesis has been developed on the basis of three separate observations. In 1974, Hirata et al. (1974) showed that the rat brain contained an enzyme, indoleamine-2,3-dioxygenase (IDO) which oxidised melatonin to form N-acetyl-N2-formyl-5-methoxykynurenamine (aFoMK). Over 35% of radiolabelled melatonin injected into the cisterna magna of the rat was recovered in the brain or urine as a F o M K or as the product of further metabolism by formamidase (N-acetyl-5-methoxykynurenamine ( a M K ) ) . Subsequently Marangos et al. (1981) found that aMK was one of the most potent naturally occurring compounds to interact with the brain bendodiazepine receptor. The benzodiazepine receptor forms part of the GABA receptor complex and aMK inhibits both GABA-stimulated benzodiazepine binding and unstimulated binding. The Ki for aMK interactions with [ 3H]diazepam binding to brain membranes is in the order of 5 × 10 -5 M (Fig. 2). Whilst this concentration would appear to be high, there are no reports of tissue or blood aMK or aFoMK levels. The final study was performed in our laboratory by Kelly et al. (1984), who observed that aMK was a very potent inhibitor of prostaglandin synthesis in ovine epididymal microsomes. Kelly et al. (1984) highlighted the close structural similarities between aMK and the

MODE OF ACTION OF MELATONIN IN CENTRAL NERVOUS SYSTEM

59

fenemate series of prostaglandin synthesis inhibitors. When aMK was compared with both aspirin and melatonin, aMK was more potent. To establish the validity of the hypothesis, it is clear that either aFoMK or aMK should be shown to precipitate physiological changes in animals similar to those observed with melatonin administration. The first such experiments were reported incidentally by Iwasaki et al. ( 1978 ) who found that injection of aFoMK but not aMK into the lateral ventricles of rats stimulated prolactin release into blood. Subsequently, Kennaway et al. ( 1988 ) reported that daily subcutaneous injections of a F o M K ( 100/tg d a y - 1) and aMK ( 1 mg d a y - i ) resulted in retarded growth of the testes, seminal vesicles and ventral prostate of prepubertal rats. The doses required to elicit this response were two to 20 times higher than the doses of melatonin required to retard sexual organ growth to the same extent. This is consistent with the putative intracellular site of action of the kynurenamines in the brain. Kennaway et al. (1989) used a different paradigm to confirm the melatonin-like actions of aFoMK in vivo. A single subcutaneous injection of aFoMK ( 1 mg) mimicked the ability of the melatonin agonist, 6-chloromelatonin, to accelerate the re-entrainment of the endogenous melatonin rhythm to a phase-advanced photoperiod. This phenomenon was not observed when 2,3-dihydro-6-chloromelatonin (a synthetic analogue of melatonin unable to be oxidised by IDO) was injected under similar conditions. Besides establishing the specificity of the response, this experiment clearly demonstrated an important role of melatonin in controlling its own circadian rhythm, presumably via interactions directly on the SCN or brain centres affecting the SCN. As discussed in the previous section, many groups have pursued investigations looking for melatonin receptors. The above prohormone hypothesis would thus reduce these 'receptors' to 'transport' proteins to pump the melatonin into the cytoplasm of target cells, where it becomes a substrate for indoleamine-2,3-dioxygenase. Although it has not been generally acknowledged, many of the receptor reports are consistent with such a transport role of the melatonin binding proteins. The marked temperature dependence of binding and the slow, incomplete dissociation of ligand from binding sites suggest an energy-dependent reaction, as opposed to an equilibrium system. The issue is discussed more extensively elsewhere (Kennaway and Hugel, 1992). MELATONIN AND CLOCK GENES

There is compelling evidence that melatonin interacts with the circadian timing system (s) within the brain to alter reproductive and other physiological activities. Whilst there is considerable evidence pointing to the SCN as the particular centre involved, some of the experiments discussed in earlier sections of this review do warn against dogmatism. Nevertheless, it is worth

60

D.J. KENNAWAY AND H.M. HUGEL

considering how melatonin/clock interactions might occur. Relatively little is known about the underlying biochemical events which lead to rhythmic neuronal and metabolic activities in vertebrate clocks. In contrast to vertebrates, there is considerable understanding of biochemical rhythmicity in invertebrates such as Drosophila and Neurospora. It has been shown that alterations in the structure or abundance of the per (period) gene product can lengthen, shorten or abolish periodicity of circadian and ultradian behavioural rhythms (Hall and Rosbash, 1988 ). The per gene product (a proteoglycan) alters the electrical coupling between cells, and Bargiello et al. (1987 ) proposed that per product may play a central role in a neural system based on interactions between electrical (gap junctions ) and chemical synapses in cells containing an oscillator, i.e. a biological clock. The challenge is to identify similar genes in vertebrate rhythm centres and to assess any interactions between their protein products and light/melatonin. Preliminary evidence has been published indicating (a) the existence of a period Gly-Thr region (per repeat) in rat SCN and (2) a rhythm in expression of the gene in the SCN (Ishida et al., 1991 ). In situ hybridisation studies showed that per repeat was highest during the day and lowest during the night and was located in neuronal cells and not glial cells. It will clearly be of interest to investigate the possible interactions between melatonin and per products within the SCN. Molecular biological techniques have also been used recently to investigate the occurrence and light regulation of the transcriptional regulatory protein Fos in the SCN of rats. Fos immunoreactivity in the SCN was highest during the light period and lowest in the dark (Aronin et al., 1990). Moreover, Rusak et al. (1990), reported increased expression of the C-fos gene and increased Fos immunoreactivity in the SCN following light pulses given at a time when they can shift circadian rhythms. Again this is a fertile area for further research not only into clock mechanisms and factors altering circadian rhythms, but also into possible melatonin interactions with the Fos genes. THEFUTURE

It is remarkable that 30 years after the discovery ofmelatonin and in an era notable for gigantic leaps forward in our understanding of reproductive hormones, the site (s) and mechanism (s) of action of melatonin are not better understood. We can hope that the answers to those important questions will be available in the next few years. The answers will come only if the right questions are asked, and as we have indicated in this review, melatonin may not act quite the same as other hormones. Nevertheless, the rewards will be substantial once the biochemistry of melatonin action is elaborated.

MODEOF ACTIONOF MELATONININ CENTRALNERVOUSSYSTEM

61

ACKNOWLEDGEMENTS

Preparation of this review and original research cited herein were supported by a grant from the Australian National Health and Medical Research Council to D.J.K.

REFERENCES Acuna Castroviejo, D., Rosenstein, R.E., Romeo, H.E. and Cardinali, D.P., 1986. Changes in gamma-aminobutyric acid high affinity binding to cerebral cortex membranes after pinealectomy or melatonin administration. Neuroendocrinology, 43:24-31. Aguado, L.I., Hancke, J.L., Rodriguez, S. and Rodriguez, E.M., 1982. Changes in the luteinizing hormone content of the rat Pars tuberalis during the estrous cycle and after lesions in the preoptic area. Neuroendocrinology, 35:178-185. Anis, Y., Nir, I. and Zisapel, N., 1989. Diurnal variations in melatonin binding sites in the hamster brain: impact of melatonin. Mol. Cell. Endocrinol., 67:121-129. Anton-Tay, F., 1974. Melatonin: effects on brain function. Adv. Biochem. Psychopharmacol., 11: 315-324. Anton-Tay, F. and Wurtman, R.J., 1969. Regional uptake of H3-melatonin from blood or spinal fluid by rat brain. Nature, 221: 474-475. Aronin, N., Sagar, S.M., Sharp, F.R. and Schwartz, N.J., 1990. Light regulates expression of a Fos-related protein in rat suprachiasmatic nuclei. Proc. Natl. Acad. Sci. USA, 87: 59595962. Bargiello, T.A., Saez, L., Baylies, M.K., Gasic, G., Young, M.W. and Spray, D.C., 1987. The Drosophila clock gene per affects intercellular junctional communication. Nature, 328:686691. Bartness, T.J., Goldman, B.D. and Bittman, E.L., 1991. SCN lesions block responses to systemic melatonin infusions in Siberian hamsters. Am. J. Physiol., 260: R 102-R 112. Bittman, E.L., Goldman, B.D. and Zucker, I., 1979. Testicular responses to melatonin are altered by lesions of the suprachiasmatic nuclei in golden hamsters. Biol. Reprod., 21: 647656. Bittman, E.L., Dempsey, R.J. and Karsch, F.J., 1983. Pineal melatonin secretion drives the reproductive response to daylength in the ewe. Endocrinology, 113: 2276-2283. Cardinali, D.P., Hyyppa, M.T. and Wurtman, R.J., 1973. Fate ofintra cisternally injected melatonin in the rat brain. Neuroendocrinology, 12: 30-40. Cardinali, D.P., Vacas, M.I. and Boyer, E.E., 1979. Specific binding of melatonin in bovine brain. Endocrinology, 105:437-441. Carlson, L.L., Weaver, D.R. and Reppert, S.M., 1989. Melatonin signal transduction in hamster brain: inhibition ofAdenyl Cyclase by a pertussis toxin-sensitive G protein. Endocrinology, 125: 2670-2676. Carter, D.S. and Goldman, B.D., 1983. Antigonadal effects of timed melatonin infusion in pinealectomised male Djungarian hamsters (Phodopus sungorus sungorus): Duration is the critical parameter. Endocrinology, 113:1261-1267. Cassone, V.M., Chesworth, M.J. and Armstrong, S.M., 1986. Dose dependent entrainment of rat circadian rhythms by daily injection ofmelatonin. J. Biol. Rhythms, 1: 219-229. Cassone, V.M., Roberts, M.H. and Moore, R.Y., 1988. Effects ofmelatonin on 2-deoxy- [ 1-14C] glucose uptake within rat suprachiasmatic nucleus. Am. J. Physiol., 225: R332-R337.

62

D.J. KENNAWAY AND H.M. H U G E L

Cohen, M., Roselle, D., Chabner, B., Schmidt, T.J. and Lippman, M., 1978. Evidence for a cytoplasmic melatonin receptor. Nature, 274: 894-895. Dowell, S.F. and Lynch, G.R., 1987. Duration of the melatonin pulse in the hypothalamus controis testicular function in pinealectomised mice (Peromyscus leucopus). Biol. Reprod., 36: 1095-1101. Dubocovich, M.L., Shankar, G. and Mickel, M., 1989.2- [ 1251]-Iodomelatonin labels sites with identical pharmacological characteristics in chicken brain and chicken retina. Eur. J. Pharmacol., 162: 289-299. Duncan, M.J., Takahashi, J.S. and Dubocovich, M.L., 1988.2- [ 12sI]-Iodomelatonin binding sites in hamster brain membranes: pharmacological characteristics and regional distribution. Endocrinology, 122: 1825-1833. Duncan, M.J., Takahashi, J.S. and Dubocovich, M.L., 1989. Characteristics and autoradiographic localisation of 2-[ 1251]-Iodomelatonin binding sites in Djungarian hamster brain. Endocrinology, 125:1011-1018. Earnest, D.J. and Sladek, C.D., 1986. Circadian rhythms of vasopressin release from individual rat suprachiasmatic explants in vitro. Brain Res., 382:129-133. Franchi, A.M., Gimeno, M.F., Cardinali, D.P. and Vacas, M.I., 1987. Melatonin, 5-methoxytryptamine and some of their analogs as cyclo-oxygenase inhibitors in rat medial basal hypothalamus. Brain Res., 405: 384-388. Glass, J.D. and Lynch, G.R., 1982. Diurnal rhythm of response to chronic intrahypothalamic melatonin injections in the white-footed mouse, Peromyscus leucopus. Neuroendocrinology, 35: 117-122. Green, D.J. and Gillette, R., 1982. Circadian rhythm of firing rate recorded from single cells in the rat suprachiasmatic brain slice. Brain Res., 245: 198-200. Hall, J.C. and Rosbash, M., 1988. Mutations and molecules influencing biological rhythms. Annu. Rev. Neurosci., 11: 373-393. Hirata, F., Hayaishi, O., Tokuyama, T. and Senoh, S., 1974. In vitro and in vivo formation of two new metabolites of melatonin. J. Biol. Chem., 249:1311-1313. Hoffman, K., Illnerova, H. and Vanecek, J., 1986. Change in duration of the night time melatonin peak may be a signal driving photoperiodic responses in the Djungarian hamster (Phodopus sungorus). Neurosci. Lett., 67: 68-72. Illnerova, H. and Vanecek, J., 1980. Pineal rhythm in N-acetyltransferase activity in rats under different artificial photoperiods and in natural daylight in the course of a year. Neuroendocrinology, 31: 321-326. Illnerova, H., Trentini, G.P. and Maslova, L., 1989. Melatonin accelerates re-entrainment of the circadian rhythm of its own production after an 8h advance of the light/dark cycle. J. Comp. Physiol. A., 166: 97-102. Ishida, N., Noji, S., Ono, K., Koyama, E., Nohno, T., Taniguchi, S., Tokunaga, A., Fuji, T. and Mitsui, Y., 1991. Diurnal regulation of per repeat mRNA in the suprachiasmatic nucleus in rat brain. Neurosci. Lett., 122:113-116. Iwasaki, Y., Kato, Y., Ohgo, S., Abe, H., Imura, H., Hirata, F., Senoh, S., Tokuyama, T. and Hayaishi, O., 1978. Effects of indoleamines and their newly identified metabolites on prolactin release in rats. Endocrinology, 103: 254-258. Kanematsu, N., Mori, Y., Hayashi, S. and Hoshino, K., 1989. Presence of a distinct 24-hour melatonin rhythm in the ventricular cerebrospinal fluid of the goat. J. Pineal Res., 7:143152. Kelly, R.W., Amato, F. and Seamark, R.F., 1984. N-acetyl-5-methoxykynurenamine, abrain metabolite of melatonin is a potent inhibitor of prostaglandin biosynthesis. Biochem. Biophys. Res. Commun., 121: 372-379. Kennaway, D.J. and Hugel, H.M., 1992. Melatonin, melatonin metabolites and the suprachiasmatic nucleus. Adv. Pineal Res., 6: 169-178. Kennaway, D.J., Hugel, H., Johnson, D.W., Royles, P., Webb, H.A. and Carbone, F., 1988.

MODEOFACTIONOFMELATONININCENTRALNERVOUSSYSTEM

63

Structure-activity studies of melatonin analogues in underfed prepubertal male rats. Aust. J. Biol. Sci., 41: 393-400. Kennaway, D.J., Blake, P. and Webb, H.A., 1989. A melatonin agonist and N-acetyl-N 1-formyl5-methoxykynurenamine accelerate the re-entrainment of the melatonin rhythm following a phase advance of the light-dark cycle. Brain Res., 495: 349-354. Kermode, J.C., 1989. The curvilinear scatchard plot. Experimental artifact or receptor heterogeneity? Biochem. Pharmacol., 38: 2053-2060. Laitenen, J.T. and Saavedra, J.M., 1990. Characterisation of melatonin receptors in the rat suprachiasmatic nuclei: Modulation of affinity with cations and guanine nucleotides. Endocrinology, 126:2110-2115. Laitenen, J.T., Castren, E., Vakkuri, O. and Saavedra, J.M., 1989. Diurnal rhythm of melatonin binding in the rat suprachiasmatic nucleus. Endocrinology, 124:1585-1587. Laitenen, J.T., Flugge, G. and Saavedra, J.M., 1990. Characterisation of melatonin receptors in the rat area postrema: modulation of affinity with cations and guanine nucleotides. Neuroendocrinology, 51: 619-624. Lang, U., Aubert, M.L., Conne, B.S., Bradtke, J.C. and Sizonenko, P.C., 1983. Influence of exogenous melatonin on melatonin secretion and the neuroendocrine reproductive axis of intact male rats during sexual maturation. Endocrinology, 112:1578-1584. Laudon, M. and Zisapel, N., 1986. Characterisation of central melatonin receptors using [ ~25I]melatonin. FEBS Lett., 197: 9-12. Leach, C.M., Reynoldson, J.A. and Thorburn, G.D., 1982. Release of E Prostaglandins in the cerebrospinal fluid and its inhibition by melatonin after cervical stimulation in the rabbit. Endocrinology, 110: 1320-1324. Lowenstein, P.R., Rosenstein, R. and Cardinali, D.P., 1985. Melatonin reverses pinealectomyinduced decrease of benzodiazepine binding in rat cerebral cortex. Neurochem. Int., 7:675681. MacPhee, A.A., Cole, F.E. and Rice, B.F., 1975. The effect of melatonin on steroidogenesis by the human ovary in vitro. J. Clin. Endocrinol. Metab., 40: 688-696. Marangos, P.J., Patel, J., Hirata, F., Sonhein, D., Paul, S.M., Skolnick, P. and Goodwin, F., 1981. Inhibition of diazepam binding by tryptophan derivatives including melatonin and its brain metabolite N-acetyl-5-methoxykynurenamine. Life Sci., 29: 259-267. Martin, J.E. and Klein, D.C., 1976. Melatonin inhibition of the neonatal pituitary response to luteinizing hormone-releasing factor. Science, 191: 301-302. Mason, R. and Rusak, B., 1990. Neurophysiological responses to melatonin in the SCN of short day sensitive and refractory hamsters. Brain Res., 533:15-19. Maywood, E.S., Buttery, R.C., Vance, G.H.S., Herbert, J. and Hastings, M.H., 1990. Gonadal responses of the male Syrian hamster to programmed infusions of melatonin are sensitive to signal duration and frequency but not to signal phase nor to lesions of the suprachiasmatic nucleus. Biol. Reprod., 43: 174-182. McArthur, A.J. and Gillette, M.U., 1990. Pertussis toxin blocks melatonin's ability to reset the suprachiasmatic circadian clock in vitro. Neurosci. Abstr., 16:317.7. McArthur. A.J., Prosser, R.A, and Gillette, M.U., 1989. Melatonin resets the suprachiasmatic circadian clock in vitro. Soc. Neurosci. Abstr., 15:420.14. Mess, B. and Ruzsas, C., 1987. Central nervous regulation of pineal function. Pineal Res. Rev., 5: 191-216. Morgan, P.J., Lawson, W., Davidson, G. and Howell, H.E., 1989a. Guanine nucleotides regulate the affinity of melatonin receptors on the ovine pars tuberalis. Neuroendocrinology, 50: 359-362. Morgan, P.J., Lawson, W., Davidson, G. and Howell, H.E., 1989b. Melatonin inhibits cyclic AMP production in cultured ovine pars tuberalis cells. J. Mol. Endocrinol., 3: R5-R8. Morgan, P.J., Williams, L.M,, Davidson, G., Lawson, W. and Howell, H.E., 1989c. Melatonin

64

D.J. KENNAWAY AND H.M. HUGEL

receptors on the ovine pars tuberalis. Characterization and autoradiographical localisation. J. Neuroendocrinol., 1: 1-4. Ng, T.B. and Lo, L.H., 1988. Inhibitory actions of pineal indoles on steroidogenesis in isolated rat Leydig cells. J. Pineal Res., 5: 229-243. Niles, L.P., 1987. [ 3H ]-melatonin binding in membrane and cytosol fractions from rat and calf brain. J. Pineal Res., 4: 89-98. Niles, L.P. and Hashemi, F., 1990. Picomolar-affinity binding and inhibition of adenylate cyclase activity by melatonin in Syrian hamster hypothalamus. Cell. Mol. Neurobiol., 10: 553558. Puchalski, W. and Lynch, G.R., 1988. Daily melatonin injections affect the expression of circadian rhythmicity in Djungarian hamsters kept under a long day photoperiod. Neuroendocrinology, 48: 280-286. Redman, J. and Armstrong, S.M., 1988. Re-entrainment of rat circadian activity rhythms: effects of melatonin. J. Pineal Res., 5:203-213. Redman, J., Armstrong, S.M. and Ng, M.T., 1983. Free running activity rhythms in the rat: entrainment by melatonin. Science, 219:1089-1091. Rivkees, S.A., Carlson, L.L. and Reppert, S.M., 1989. Guanine nucleotide-binding protein regulation of melatonin receptors in lizard brain. Proc. Natl. Acad. Sci. USA, 86: 3882-3886. Robinson, J.E. and Karsch, F.J., 1987. Photoperiodic history and a changing melatonin pattern determine the neuroendocrine response of the ewe to daylength. J. Reprod. Fertil., 80:159165. Rollag, M.D., O'Callaghan, P.L. and Niswender, G.D., 1978. Serum melatonin concentrations during different stages of the annual reproductive cycle in ewes. Biol. Reprod., 18: 279-285. Rosenstein, R.E. and Cardinali, D.P., 1986. Melatonin increases in vivo GABA accumulation in rat hypothalamus, cerebellum, cerebral cortex and pineal gland. Brain Res., 398: 403-406. Rusak, B., 1980. Suprachiasmatic lesions prevent an antigonadal effect of melatonin. Biol. Reprod., 22: 148-154. Rusak, B., Robertson, H.A., Wisden, W. and Hunt, S.P., 1990. Light pulses that shift rhythms induce gene expression in the suprachiasmatic nucleus. Science, 248: 1237-1240. Schwartz, W.J., Reppert, S.M., Eagan, S.M. and Moore-Ede, M.C., 1983. In vivo metabolic activity of the suprachiasmatic nuclei: a comparative study. Brain Res., 274:184-187. Shaw, P.F., Kennaway, D.J. and Seamark, R.F., 1989. Evidence of high concentrations ofmelatonin in lateral ventricular cerebral spinal fluid of sheep. J. Pineal Res., 6:201-208. Shibata, S., Cassone, V.M. and Moore, R.Y., 1989. Effects ofmelatonin on neuronal activity in the rat suprachiasmatic nucleus in vitro. Neurosci. Lett., 97: 140-144. Stankov, B. and Reiter, R.J., 1990. Melatonin receptors: current status, facts and hypotheses. Life Sci., 46: 971-982. Stehle, J., Vanecek, J. and Vollrath, L., 1989. Effects of melatonin on spontaneous electrical activity of neurons in rat suprachiasmatic nuclei: an in vitro iontophoretic study. J. Neural Transm. (Gen. Sect), 78: 173-177. Sugden, D. and Chong, N.S.W., 1991. Pharmacological identity of 2- [ 125I] Iodomelatonin binding sites in chicken brain and sheep pars tub eralis. Brain Res., 539:151-154. Vakkuri, O., Lasma, E., Rahkamaa, E., Ruotsalainen, H. and Leppaluoto, J., 1984. Iodinated melatonin: preparation and characterisation of molecular structure by mass and 3H-NMR spectrometry. Anal. Biochem., 142: 284-289. Vanecek, J. and Jansky, L., 1989. Short days induce changes in specific melatonin binding in hamster median eminence and anterior pituitary. Brain Res., 477: 387-390. Vanecek, J. and Vollrath, L., 1989. Melatonin inhibits cyclic AMP and cyclic GMP accumulation in the rat pituitary. Brain Res., 505:157-159. Vanecek, J., Pavlik, A. and Illnerova, H., 1987. Hypothalamic melatonin receptor sites revealed by autoradiography. Brain Res., 435: 359-362.

MODE OF ACTION OF MELATONIN IN CENTRAL NERVOUS SYSTEM

65

Vaughan, G.M., Mason, A.D. and Reiter, R.J., 1986. Serum melatonin after a single aqueous subcutaneous injection in Syrian hamsters. Neuroendocrinology, 42: 124-127. Weaver, D.R., Carlson, L.L. and Reppert, S.M., 1990. Melatonin receptors and signal transduction in melatonin-sensitive and melatonin insensitive populations of white footed mice (Peromyscus leucopus). Brain Res., 506: 353-357. Weaver, D.R., Provencio, 1., Carlson, L.L. and Reppert, S.M., 1991. Melatonin receptors and signal transduction in photorefractory Siberian hamsters (Phodopus sungorus). Endocrinology, 128: 1086-1092. Williams, L.M., Morgan, P.J., Hastings, M.H., Lawson, W., Davidson, G. and Howell, H.E., 1989. Melatonin receptor sites in the Syrian hamster brain and pituitary. Localisation and characterisation using [ 125I] Iodomelatonin. J. Neuroendocrinol., 1: 315-320. Wittowski, W., Hewing, M., Hoffmann, K., Bergmann, M. and Fechner, J., 1984. Influence of photoperiod on the ultrastructure of the hypophysial pars tuberalis of the Djungarian hamster. Phodopus sungorus. Cell Tissue Res., 238:213-216. Wittowski, W., Bergmann, M., Hoffmann, K. and Pera, F., 1988. Photoperiodic-dependent changes in TSH-like immunoreactivity of cells in the hypophysial pars tuberalis of the Djungarian hamster Phodopus sungorus. Cell Tissue Res., 251: 183-187. Zisapel, N., Egozi, Y. and Laudon, M., 1982. Inhibition of dopamine release by melatonin: regional distribution in the rat brain. Brain Res., 246: 161-164. Zisapel, N., Egozi, Y. and Laudon, M., 1985. Circadian variations in the inhibition of dopamine release from adult and newborn rat hypothalamus by rnelatonin. Neuroendocrinology, 40: 102-108.