Joumal of Sensory, Comparative and .oo.~,, Physiology A .Physiology oho~o,o,

J Comp Physiol A (1984) 154:435-440

9 Springer-Verlag 1984

Multiple redundant circadian oscillators within the isolated avian pineal gland Joseph S. Takahashi* and Michael Menaker Institute of Neuroscience, University of Oregon, Eugene, Oregon 97403, USA Accepted September 2, 1983

Summary. The avian pineal gland contains a circadian pacemaker that oscillates in vitro. Using a flow-through culture system it is possible to measure melatonin production from very small subsections of an individual gland. We have used this technique to attempt to localize the oscillators in the pineal. Progressive tissue reduction did not affect the rhythmicity of cultured pineals. Multiple pieces (up to eight) from a single pineal all were capable of circadian oscillation - establishing directly that a pineal gland contains at least eight oscillators. All pineal pieces were responsive to light, and single light pulses shifted the phase of the melatonin rhythm. Because pieces equivalent to less than one per cent of the whole gland were rhythmic and because the capacity for oscillation was distributed throughout the gland, an individual pineal appears to be composed of a population of circadian oscillators.

tration of melatonin is approximately 400 times that detectable with the radioimmunoassay we employ (Takahashi et al. 1980). The capacity for melatonin synthesis is distributed throughout the gland and it is therefore possible to measure melatonin output from very small pieces of pineal. We have used this technique to ask whether there is more than one circadian oscillator in the pineal, and if so whether the oscillators responsible for the melatonin rhythm are localized in a discrete region of the gland, or are distributed throughout. Additionally, we have asked whether there are any correlations between the size of pineal fragments and circadian properties of the oscillation. Here we report direct evidence that an individual avian pineal is composed of a population of circadian oscillators and that its circadian rhythm results from the combined output of these oscillators. Materials and methods

Introduction

The physiological systems responsible for generating circadian rhythms in multicellular organisms are generally thought to be composed of multiple oscillators (for review see Takahashi and Menaker 1983; Takahashi and Zatz 1982). Although the evidence that supports this view is compelling, it is as yet indirect. The avian pineal gland contains a circadian pacemaker that oscillates in vitro (Deguchi 1979; Kasal etal. 1979; Takahashi etal. 1980). When entire pineal glands are cultured in a flow-through system, the release of melatonin into the perfusate is circadian and the peak concen* Present address : Department of Neurobiology and Physiology, Northwestern University, Hogan Hall, Evanston, Illinois 60201, USA

Animals. Newly hatched chicks (Gallus domesticus, white leghorn, Babcock strain) were purchased from Babcock Western (Lakeview CA) and raised under LD 12:12 lighting conditions. Food (Purina Chick Starter) and water were continuously available. Pineal culture. A previously described flow-through superfusion culture system was used (Takahashi et al. 1980) with several modifications. A peristaltic pump (Harvard Apparatus, Model 1203) was used to infuse culture medium and up to 16 individual culture chambers could be run simultaneously. The input and output tubing for the culture medium was Silastic (Dow Corning; input: 0.508 mm I.D.; output: 1.02 mm I.D.) and no gassing chamber for the culture medium input was used. Because Silastic tubing is extremely permeable to both oxygen and carbon dioxide gas, the bicarbonate buffer system used previously was inadequate and required the addition of HEPES buffer. The culture medium formulation consisted of Medium 199 with Hank's salts (GIBCO), 25 mmol/1 HEPES (Research Organics), 50 mg/ml Gentamicin antibiotic (Schering), and 95% 02/5% CO 2 gas. The flow rate was 0.5 ml/h. Melatonin was assayed by radioimmunoassay (Rollag and Niswender 1976) as de-

436

J.S. Takahashi and M. Menaker: Population of oscillators in the isolated pineal

scribed in Takahashi et al. (1980). All pineal pieces were exposed to one cycle of LD 12:12, followed by two days of constant darkness. Pineal dissection. Pineal glands were cut into pieces with a scalpel blade or microscissors with the aid of a dissecting microscope in a laminar flow hood. Pineals were cut midsagittally to produce left and right halves. Quarter pineals were made by cutting the pineal across the long axis to produce a top quarter (dorsal portion of pineal in the intact bird), a second quarter, a third quarter and a 'tail' quarter (the distal portion of the pineal stalk). Eighth pineals were made by first cutting the pineal midsagittally to produce two halves, and then cutting each half into four pieces across the longitudinal axis as with quartering. 'Lobules' were pieces of pineal that were approximately 1/16 to 1/32 of the giand and were dissected from portions of the gland that were distinctly partitioned by septa into lobular regions.

80 60 40

= 2o e-

z z o

10

8

_..1 u.I

N

6

Results

Effects of tissue reduction on the circadian oscillation All regions of the pineal were capable of circadian oscillation. Initially, two whole pineals, three pairs of left and right half pineals, and two sets of quarter pineals were compared. In this experiment, the whole pineals were rhythmic in LD 12: 12, but were not convincingly rhythmic in constant darkness. The half and quarter pineals were strongly rhythmic in L D 12:12 and expressed a circadian oscillation in constant darkness (Fig. 1). Halves and quarters appeared to oscillate more consistently than whole glands. The smaller pieces of tissue may survive better as a result of a greater surface-to-volume ratio (preliminary electron micrographs of previously cultured whole pineals indicate that there is considerable necrosis in the center of the gland). The melatonin production of the tissue was approximately proportional to the size of the pieces. Whole pineals normally produce 112 + 24 ng/ml ( x _ SE, N = 8). In this experiment, the half pineals produced peak values averaging 65__ 9.2 ng/ml (N= 6); and quarter pineals produced 14.8_+2.9ng/ml (N=8). These values are within the range expected, considering the variation found in whole glands, if melatonin production is proportional to tissue volume. In further experiments, smaller pieces of pineal were examined for rhythmicity. Reducing pineals into eighths did not affect the melatonin rhythm (Fig. 2). Furthermore, all eight pieces from an individual gland were rhythmic. The melatonin rhythms obtained from 'lobules' were similar to those observed in larger pieces (Fig. 2). Although the lobules appeared to be about 1/16 to 1/32 of the whole gland, their peak melatonin production

12

24 36 48 60 72 84 TIME (hours) Fig. 1 A, B. Circadian rhythms of melatonin release from subsections of chicken pineal glands maintained in a flow-through culture system. Panel A shows the melatonin release from a half pineal, and panel B shows results from a quarter pineal. The in vitro light conditions are indicated by the shading pattern (white = light, gray = dark). Samples of perfusate were collected every 90 rain. Each point represents a single radioimmunoassay determination and is plotted at the end of the collection interval

was much lower than expected and averaged 0.49 _+0.14 ng/ml (+SE, N=4). This melatonin level is less than one percent of the average melatonin production of whole glands. Because the dissections to produce 'lobules' involved cutting the tissue on all sides, the possibility of damage to cells from the cuts was high. The low melatonin levels may indicate that only a small portion of the original tissue was viable. If we assume that melatonin production is proportional to tissue size, this would suggest that the viable cells in the lobules were equal to less than one percent of the gland. Tissue reduction did not appear to affect the period length of the rhythm. Although it is difficult to quantify the period length on the basis of short free runs, large changes in period length could have been detected by visual inspection of the raw data and were not found.

Effects of light on the melatonin rhythm from pineal pieces In all the pieces of pineal tissue that were examined, light appeared both to inhibit the production of melatonin during the 12-h photoperiod, and to

J.S. Takahashi and M. Menaker: Population of oscillators in the isolated pineal 6

437

48

40 32 24

2 E

I

16

r-

z z o b< ,,-fi, 0.5

E

8

tv

z z

o

48

"'

40

b-

0.4 0.3 0.2

32

0.1

24

12

24 36 48 60 72 84 TIME (hours) Fig. 2A, B. Circadian rhythms of melatonin release from an eighth pineal (A) and from a 'lobule' (B). All eight pieces from this gland were rhythmic. Note that the rising phase of the melatonin peak on the first day in culture was delayed and that the light cyc!e appeared to truncate the peak by inhibiting melatonin. Collection interval and plotting conventions are the same as Fig. I

maintain the amplitude of the first melatonin peak that occurred after the light-to-dark transition (at hour 36 in Figs. 1, 2). These effects of light demonstrate that the photoreceptors as well as the oscillators contained within the gland are distributed throughout the organ. In preliminary experiments using quarter pineals, we observed clear cases in which light shifted the phase of the free-running melatonin rhythm. Single six-hour pulses of light cause changes in the phase and sometimes the amplitude of the subsequent melatonin oscillation. Two examples of these effects are shown in Fig. 3. In each panel, the two records are from quarter pineals obtained from one gland. In the first panel (Fig. 3A), one piece of the gland was maintained in constant darkness, while the other piece was exposed to a six-hour light pulse beginning early in the first night at hour 13.5. The light pulse delayed the rise of melatonin about four hours relative to that seen in constant darkness. The spontaneous decline in melatonin,

16

48 60 72 84 96 108 TIME (hours) Fig. 3A, B. The effect of single 6-h light pulses on the phase of the melatonin rhythm. Each panel shows melatonin release from two quarter pineals from the same gland. A One pineal piece (closed circles) was exposed to a 6-h pulse (1500 lux) beginning at hour 13.5; while the other piece (open circles) was maintained in constant darkness beginning at hour 12. B One pineal piece (closed circles) was exposed to a 6-h pulse beginning at hour 18; while the other piece (open circles) was maintained in constant darkness 12

24

36

however, occurred at the same time in both pieces; and the phases of the second and third melatonin peaks were similar (there is a slight delay in the light exposed pineal piece which is difficult to quantify). Although a light pulse at this phase did not cause a significant phase shift, there was a clear increase in the amplitude of oscillation. Six-hour light extensions during the first night of culture have similar effects on the waveform and amplitude of the rhythm. The second panel (Fig. 3 B), illustrates the effects of a light pulse given later in the first night, from hour 18 to hour 24. After the onset of darkness at hour 12 melatonin levels from both pieces increased. In the experimental

438

J.S. Takahashi and M. Menaker: Population of oscillators in the isolated pineal

piece the light pulse caused a rapid inhibition of melatonin and levels remained low for the duration of the pulse. After the pulse, melatonin levels increased and peaked about six hours later. The subsequent decline of melatonin was delayed about eight hours relative to that seen in constant darkness. Furthermore, the subsequent phase of the melatonin oscillation was delayed about eight hours relative to the dark control, and the amplitude of the rhythm was increased. These experiments show that single pulses of light can induce shifts in the phase of the freerunning oscillation. We do not yet know whether these phase shifts vary in magnitude or direction as a function of the phase of light exposure. However, if we use the peak of the melatonin rhythm as a phase reference point for the middle of the 'subjective night' (CT 18), then the preliminary results obtained (slight delays at CT 8, and large delays at CT 14) are consistent with phase response curves to light described in other systems, in which delays occur during the early subjective night ( ~ CT 12 to ~ CT 18) and advances occur during the late subjective night ( ~ C T 18 to ~ C T 24). These experiments, although preliminary, provide the only available evidence that the circadian oscillators in the pineal can be phase shifted by light pulses in vitro. Discussion

Although it is generally accepted that circadian organization in multicellular organisms is based on multioscillator systems (Pittendrigh 1974; Block and Page 1978; Takahashi and Menaker 1982, 1983), there are few examples in which this proposition has been directly demonstrated. The only unequivocal cases in which multiple oscillators have been demonstrated in an individual are found among the marine mollusks 1. In Aplysia, Navanax, Bursatella, and Bulla, each of the two eyes contains a circadian pacemaker that oscillates in vitro (Jacklet 1969; Hudson and Lickey 1980; Eskin and Harcombe 1977; Block and Roberts 1981; Block and Wallace 1982). Therefore each of these organisms must contain at least two circadian oscillators. In Aplysia, it has been suggested that an individual eye is composed of a population of interacting oscillators (Jacklet and Geronimo 1971); however, this proposition is not well supported (Strum1 In a number of organisms, there is indirect evidence for the existence of multiple oscillators (Aschoff and Wever 1981 ; Moore-Ede and Sulzman 1981; Underwood 1977; Koehler and Fleissner 1978; Page 1982); however, in none of these cases has it been possible to isolate these oscillators directly

wasser 1974; Eskin 1979). Although it is clear that only a small fraction of the cells of the eye (50 to 100 of about 1000 in the Bulla eye) are necessary for generating circadian rhythms in molluscan eyes (Strumwasser 1974; Block and Wallace 1982), in no case has it been possible to record circadian oscillations from multiple pieces of an individual eye. Using chick pineals, Kasal and Perez-Polo (1980) have provided evidence that half pineals can oscillate in culture. However, because the rhythm assayed required a population of half pineals, it was not possible to demonstrate that any single piece could oscillate. Deguchi (1979) has shown that dispersed cell cultures of chicken pineal glands express circadian rhythms of N-acetyltransferase (NAT) activity. Although the cells in his cultures may be in physical and/or chemical contact with each other, his evidence suggests that both the circadian oscillation and a response to light (inhibition of NAT activity) might be properties of individual pineal cells. Our experiments demonstrate directly that the capacities for circadian oscillation and photoreception are distributed throughout the chicken pineal gland and that tissue reduction has no detectable effects on the period length of the rhythm. In one case eight pieces, cultured separately but originating from the same pineal were all capable of circadian oscillation. This establishes directly that there are at minimum eight oscillators in the gland. The persistence of rhythmicity in' lobules' which may be equivalent to less than 1% of the gland, suggests that there are in fact many more than eight oscillators. These experiments confirm the impression conveyed by the ultrastructure of the gland that it is quite homogeneous both in structure and in function. The structure of the chicken pineal gland has been relatively well described (Beattie and Glennie 1966; Boya and Calvo 1978; Wight and MacKenzie 1971; Menaker and Oksche 1974). The gland is encapsulated within a connective tissue sheath. Septa of collagen penetrate inward from the capsule and divide the pineal into many lobules. Within these lobules are numerous cell clusters or follicles that are supported by connective tissue partitions arising from proliferations of the lobule septa. A network of nerve fibers (sympathetic) and an extensive capillary system are interlaced within the connective tissue partitions. Ultrastructural studies indicate that the follicles contain three identifiable cell types : ependymal, secretory, and photoreceptor-like cells (Bischoff 1969). In adult chickens, the ependymal cells have apical cilia with a ' 9 + 2' arrangement of microtubules; the secretory cells have numerous granules

J.S. Takahashi and M. Menaker: Population of oscillators in the isolated pineal Input

Light

Photoreceptor

Circadian Pacemaker

Melatonin Biosynthesis

~~--~~ i i S ~ i i , i

throughout the cytoplasm; and the photoreceptorlike cells have synaptic contacts along their basal border and have modified '9 + 0' cilia which occasionally connect the apical portion of the cell to a membranous lamellar whorl located in the lumen of the follicle (Bischoff 1969). The photoreceptor cells resemble degenerate versions of the cone-like photoreceptor cells found in lower vertebrate pineals (Eakin 1973; Dodt 1973). In two- to five-dayold chicks, there appear to be more photoreceptorlike cells (relative to secretory cells) than in the adult (Boya and Zamarano 1975). The photoreceptor-like cells of young chicks do not appear to have membranous whorls attached but almost always have 9 + 0 cilia ending in a cytoplasmic bulb. The ultrastructural studies indicate that the pineal of the chicken contains cells with structural features corresponding to both secretory and photoreceptive functions. At present no clear correlation at the cellular level can be made between structure and function in the avian pineal. The avian pineal gland can be described functionally as a system containing at least three major components: (1) an input pathway that is photoreceptive; (2) a circadian oscillator or pacemaker that generates the rhythm; and (3) an output pathway that results in the synthesis of melatonin (Fig. 4). For entrainment to occur, the photoreceptor must be coupled to the oscillator; and to generate rhythmic melatonin output, the oscillator must be coupled to the melatonin synthetic pathway. We know from the tissue reduction experiments that this three component system is redundant. Our inability to separate the three components, suggests that they are a closely associated functional unit, how closely we do not yet know. The experiment of Deguchi (1979) suggests that we can rephrase the question more succinctly: are all three functions properties of a single cell-type, or alternatively, are they distributed in more than one celltype ? The redundancy of the system raises a number of organizational questions. Do the oscillators interact? Are they coupled together? If they are coupled, what pathways mediate the coupling? The strong damping of the melatonin rhythm in constant darkness suggests that if there is any coupling, it is weak. In whole glands, the damping

Output

tolIin ' / X /

439 Fig. 4. A schematicdiagram of the components of a circadian systemin the isolated avian pineal gland. Abbreviations: N A T N-acetyltransferase;HIO M T hydroxyindole-O-methyltransferase;5 - H T serotonin; N A S N-acetylserotonin

is rapid and the waveform is irregular. However, in small pieces of pineal, the damping of the amplitude of the rhythm is correlated with a clear broadening of the peak which suggests that it may be due to progressive asynchrony in the population of circadian oscillators. Additional evidence for this view comes from the effects of single light pulses on the amplitude of the rhythm. Light pulses which cause the subsequent peak of melatonin to become narrow also cause the subsequent oscillation to have a larger amplitude, suggesting that these pulses increase coherence among the oscillators. In contrast, pulses given during the middle of the subjective night frequently obscure the melatonin peak, and this effect appears to decrease the amplitude of the rhythm. Light falling at this phase, which should be close to the breakpoint between delays and advances of the phase response curve (Eskin 1971; Binkley et al. 1981), would be expected to decrease the coherence of a population of oscillators. In the presence of a light-dark cycle, the gland could express a coherent rhythm of melatonin output even if there were no coupling among its constituent oscillators because the entraining cycle imposes phase control on the entire population (if each oscillator has photoreceptive input). In culture, light cycles with long photoperiods are necessary for entrainment to occur and to prevent the damping of the oscillation in constant darkness (Takahashi 1981). In the intact chicken, however, the pineal maintains a low amplitude but persistent circadian rhythm of melatonin for at least two weeks in constant darkness (Ralph et al. 1974). The long-term persistence of a pineal rhythm in constant conditions suggests that the oscillators in the pineal remain in synchrony in the intact chicken. Whether this synchrony is maintained by mutual coupling within the gland, or is imposed upon the pineal oscillators by inputs from elsewhere in the organism (e.g., the sympathetic input from the superior cervical ganglion (Cassone and Menaker 1983)), remains to be determined. Acknowledgements. We thank SherryWisner for technicalassistance, Dr. M. Rollag for providing the melatonin antibody. Research was supported by NIH grant AM26972 to M. Menaker and predoctoraltraining grant GM07257 to J. Takahashi.

440

J.S. Takahashi and M. Menaker: Population of oscillators in the isolated pineal

References Aschoff J, Wever R (1981) The circadian system of man. In: Aschoff J (ed) Handbook of behavioral neurobiology, vol 4, Biological rhythms. Plenum Press, New York, pp 311-331 Beattie CW, Glennie FH (1966) Some aspects of the vascularization and chemical histology of the pineal gland in Gallus. Anat Anz 118 : 396M04 Binkley S, Muller G, Hernandez T (1981) Circadian rhythm in pineal N-acetyltransferase activity: phase shifting by light pulses (I). J Neurochem 37:798-800 Bischoff MB (1969) Photoreceptoral and secretory structures in the avian pineal organ. J Ultrastruct Res 28:16-26 Block GD, Page TL (1978) Circadian pacemakers in the nervous system. Annu Rev Neurosci 1 : 19-34 Block GD, Roberts MH (1981) Circadian pacemaker in the Bursatella eye: properties of the rhythm and its effects on locomotor behavior. J Comp Physiol 142:403-410 Block GD, Wallace SF (1982) Localization of a circadian pacemaker in the eye of a mollusk, Bulla. Science 217:155-157 Boya J, Calvo J (1978) Post-hatching evolution of the pineal gland of the chicken. Acta Anat 101 : 1-9 Boya J, Zamarano L (1975) Ultrastructure study of the pineal glands of the chicken (Gallus gallus). Acta Anat 92: 20~226 Cassone VM, Menaker M (1983) Sympathetic regulation of chicken pineal rhythms. Brain Res 272:311-317 Deguchi T (1979) A circadian oscillator in cultured cells of chicken pineal gland. Nature 282:94-96 Dodt E (1973) Parietal eye (pineal and parietal organs) of lower vertebrates. In: Jung R (ed) Handbook of sensory physiology, vol VII/3B. Springer, Berlin Heidelberg New York, pp 113-140 Eakin RM (1973) The third eye. University of California Press, Berkeley Eskin A (1971) Some properties of the system controlling the circadian activity rhythm of sparrows. In: Menaker M (ed) Biochronometry. Natl Acad Sci, Washington DC, pp 55 78 Eskin A (1979) Circadian system of the Aplysia eye: Properties of the pacemaker and mechanisms of its entrainment. Fed Proc 38:2573-2579 Eskin A, Harc0mbe E (1977) Eye of Navanax: optic activity, circadian rhythm and morphology. Comp Biochem Physiol 57 A: 443-449 Hudson DJ, Lickey M (1980) Internal desynchronization between the identified circadian oscillators in Aplysia. Brain Res 183:481-485 Jacklet JW (1969) Circadian rhythm of optic nerve impulses recorded in darkness from isolated eye of ApIysia. Science 164:56~563 Jacklet JW, Geronimo J (1971) Circadian rhythm: population of interacting neurons. Science 174:299-302

Kasal C, Perez-Polo R (1980) In vitro evidence of photoreception in the chick pineal gland and its interaction with the circadian clock controlling N-acetyltransferase (NAT). J Neurosci Res 5 : 579-585 Kasal C, Menaker M, Perez-Polo R (1979) Circadian clock in culture: N-acetyltransferase activity of chick pineal glands oscillates in vitro. Science 203 : 656-658 Koehler WK, Fleissner G (1978) Internal desynchronization of bilaterally organized circadian oscillators in the visual system of insects. Nature 274:708-710 Menaker M, Oksche A (1974) Avian pineal gland. In: Farner DS, King JR (eds) Avian biology, vol IV. Academic Press, New York, pp 79-118 Moore-Ede MC, Sulzman FM (1981) Internal temporal order. In: Aschoff J (ed) Handbook of behavioral neurobiology, vol. 4. Biological rhythms. Plenum Press, New York, pp 215-241 Page T (1982) Transplantation of the cockroach circadian pacemaker. Science 216:73-75 Pittendrigh CS (1974) Circadian oscillations in cells and the circadian organization of multicellular systems. In: Schmitt FO, Worden FG (eds) The Neurosciences Third Study Program. MIT Press, Cambridge, pp 437-458 Ralph CL, Pelham RW, MacBride SE, Reilly DP (1974) Persistent rhythms of pineal and serum melatonin in cockerels in continuous darkness. J Endocrinol 63:319-324 Rollag MD, Niswender GD (1976) Radioimmunoassay of serum concentration of melatonin in sheep exposed to different lighting regimes. Endocrinology 98:482-489 Strumwasser F (1974) Neuronal principles organizing periodic behaviors. In: Schmitt FO, Worden FG (eds) The Neurosciences Third Study Program. MIT Press, Cambridge, pp 459-478 Takahashi JS (1981) Neural and endocrine regulation of avian circadian systems. PhD dissertation, Dept Biology and Inst Neuroscience, Univ Oregon, Eugene Takahashi JS, Menaker M (1982) Entrainment of the circadian system of the house sparrow: a population of oscillators in pinealectomized birds. J Comp Physiol 146:245-253 Takahashi JS, Menaker M (1984) Circadian rhythmicity: regulation in the time domain. In: Yamamoto KR, Goldberger RF (eds) Biological regulation and development, vol 3 B. Plenum Press, New York, (in press) Takahashi JS, Zatz M (1982) Regulation of circadian rhythmicity. Science 217:1104-1111 Takahashi JS, Harem H, Menaker M (1980) Circadian rhythms of melatonin release from individuai superfused chicken pineal glands in vitro. Proc Natl Acad Sci USA 77:2319-2322 Underwood H (1977) Circadian organization in lizards: the role of the pineal organ. Science 195:587-589 Wight PAL, MacKenzie GM (1971) The histochemistry of the pineal gland of the domestic fowl. J Anat 108:261-273