- Email: [email protected]
Chloride channel block phase advances the single-unit activity rhythm in the SCN
BrainResearchBulletin, Vol. 34,No.l,pp.69-72,1994 Coovrieht 0 1994 Elsevier Science Ltd P&e~ in the USA. Allrights reserved 0361-9230/94 $6.00+ .OO
Pergamon 0361-9230(93)EOO36-L
BRIEF COMMUNICATION
Chloride Channel Block Phase Advances the Single-Unit Activity Rhythm in the SCN RADHA
RANGARAJAN,*
H. CRAIG
HELLER,*f
AND
JOSEPH
D. MILLER**’
*Department of Biological Sciences, fstanford Center for Sleep and Circadian Neurobiology, and #Department of Psychiatry, Stanford University, Stanford, CA 94305 Received
23 April 1993; Accepted
30 November
1993
RANGARAJAN, R., H. C. HELLER AND J. D. MILLER. Chloride channel blockphase advances the single-unit activity rhythm in the SCN. BRAIN RES BULL 34(l) 69-72, 1994.-The mammalian suprachiasmatic nuclei (SCN) contain a circadian pacemaker that exhibits a 24 h rhythm in single-unit activity in vivo and in vitro. Chloride channel block by a saturating concentration of picrotoxin at either CT6 or CT15 produces large phase advances in the SCN single-unit activity rhythm in vitro. These phase advances are not affected by simultaneous blockade of voltage-sensitive sodium and calcium channels by ‘lTX and magnesium. Thus, the effects of picrotoxin appear to be mediated by direct blockade of the chloride channel, rather than subsequent membrane depolarization. GABA-A receptor-mediated chloride flux may be part of the mechanism of circadian timekeeping. Suprachiasmatic
nucleus
Circadian
rhythm
Chloride
Picrotoxin
to circadian timekeeping, chloride channel blockade should strongly perturb the circadian oscillator. For that reason the effects on the SCN slice of the chloride channel blocker, picrotoxin, were studied at two circadian times. Because picrotoxin is a strong depolarizing agent, its effects were also studied under joint fast sodium and calcium channel blockade to determine whether observed phase shifting effects were secondary to membrane depolarization and subsequent intracellular cation flux.
THE suprachiasmatic nuclei (SCN) of the hypothalamus are necessary and sufficient for the maintenance of circadian rhythmicity in mammals [for a recent review, see (12)]. Circadian rhythms in single-unit activity, neurotransmitter release, and cellular metabolism are maintained in vitro in the hypothalamic slice preparation (4,516). While the cellular substrate of the circadian pacemaker within the SCN has not been determined, virtually all the neurons of the SCN are GABAergic in nature (18). Mathematical models of the circadian pacemaker have suggested an important role for inhibitory coupling, possibly of a GABAergic nature, among component oscillators in a neural network in the SCN (12,14,19). GABAergic agonists profoundly reduce cellular metabolism in the SCN when administered to the SCN slice during the extrapolated light portion of the 1ight:dark cycle, but not in the extrapolated dark portion (13,27). In contrast, most neurons of the SCN are inhibited by iontophoretic or bath application of GABA, independent of time of day (7-9,17,23). The GABAergic agonist, muscimol, produces phase advances of the circadian rhythm of locomotor activity when administered in vivo in subjective day, but phase delays when administered in subjective night (24). The GABA-A receptor subtype is present in the SCN, but may have a somewhat variant subunit composition [most likely (Ye& y2; (1,2,31)], perhaps explaining the difficulties that some investigators have had in unequivocally demonstrating its presence (11). This GABA receptor subtype ultimately gates a chloride channel. If a local GABAergic intemeuronal network is essential 1 Requests
for reprints
should be addressed
GABA
METHOD The methods used in these experiments have been described in detail earlier (20). Briefly, male Wistar rats were housed on a 12:12 1ight:dark cycle. They were sacrificed during the light portion of their cycles and hypothalamic brain slices, 500 pm thick, were prepared. Slices containing the SCN were maintained in a Hatton-style brain slice chamber (6) at a temperature of 37°C. The slices were superfused at a rate of 20 ml/h with Earle’s balanced salt solution (Sigma), aerated with a 95% O#% CO2 mixture, and supplemented with 24.6 mM glucose and 26.2 mM sodium bicarbonate, pH 7.4. The first set of experiments tested the capacity of picrotoxin to alter the normal time of peak in the single unit activity rhythm. Slices were treated with a chloride channel saturating concentration of picrotoxin at CT6 and CT15. Superfusion was interrupted for an hour at the appropriate time (CT6 or CT15) and the control medium in the chamber was replaced with test medium contain-
to Dr. Joseph Miller, Department
69
of Biological
Sciences,
Stanford
University,
Stanford,
CA 94305.
70
RANGARAJAN,
(A) a
2
I , .1
3
6
9
12
15
Circadian
18
21
Time
24
27
30
(hr)
33
36
(B) 10
HELLER
AND MILLER
ing picrotoxin at a concentration of 10 PM. The test medium was adjusted for pH, temperature, and oxygen concentration before exchanging it with the control medium. After an hour the test medium was removed, replaced with control medium, and superfusion was reinstated. Control experiments to confirm the time of peak activity in normal slices in the absence of any drugs were performed in an identical fashion, but in the absence of picrotoxin. A second set of experiments examined the effects of 10 ,uM picrotoxin in the presence of 1 PM tetrodotoxin (to block fast sodium channels and subsequent action potentials) and 10 mM magnesium chloride (to block voltage-sensitive calcium channels), using the same method as described above. Control experiments were performed with 1 PM tetrodotoxin and 10 mM magnesium chloride to determine whether these cation channel blockers had any independent effect on the rhythm in single-unit activity. The effect of drug treatment was determined by monitoring the rhythm of spontaneous neuronal activity during the first circadian cycle after preparation and treatment of slices. The single unit activity rhythm was determined via extracellular recording from single neurons, using glass micropipettes filled with 3 M NaCl. Action potentials from single cells, encountered sequentially in electrode tracks run through the SCN, were recorded. The firing rate of each cell was averaged over a 5 min period, and cells were recorded sequentially for 8 to 10 h per day. Data were collected and analyzed using a BrainWave system. The firing rates of individual cells were averaged over successive 2 h intervals lagged by 1 h. The time of peak activity was then determined. Phase shifts were determined by comparing the time of peak activity in the presence of the pharmacological manipulation with the time of peak activity in control slices.
RESULTS
2
0
, . 3
9
12
15
18
Circadian (C)
21
Time
24
27
30
33
36
(hr)
lo
2 3
6
9
12
15
Circadian
18
21
Time
24
(hr)
27
30
33
36
The first set of control experiments confirmed that the time of peak activity was at CT5 t 0.56 h (n = 3) (Fig. lA), in good agreement with previous work (e.g., 21). One hour treatment with picrotoxin (10 PM) at CT6 caused a 4-5 h advance in time of peak activity the next day so that the new peak was seen at CT1 + 0.49 h (n = 3) (Fig. 1B). The same treatment at CT 15 caused the new peak to appear at CT23 ? 0.67 (n = 4 ) (Fig. 1C). Two control experiments with 1 PM tetrodotoxin and 10 mM magnesium chloride joint treatment at CT6 showed a normal time of peak activity the next day at CT6 ? 0.78 (n = 2) (Fig. 2A). Action potentials were not detected during the 1 h treatment. One hour treatment with 10 PM picrotoxin at CT6, in the presence of 1 PM tetrodotoxin and 10 mM magnesium chloride, induced a 5-6 h advance in the time of peak activity on the following day. The peak activity was seen at CT1 ? 0.46 (n = 3) (Fig. 2B).
FIG. 1. (A) Average (n = 3 experiments) single-unit activity rhythm recorded from the SCN in vitro. Plotted are the grand averages -C SE across all untreated control experiments at each CT of the 2 h means of firing rates of single SCN cells. Slices are prepared on day 1. The time of peak (dotted line) occurs on day 2 at CT5 -t 0.56. Stippled horizontal bar = lights off in animal colony. (B) Effect of picrotoxin (10 PM), administered for 1 h at cT6, on average (n = 3) single-unit activity rhythm in the SCN recorded the following day. Vertical bar indicates time of drug administration. The time of peak (dotted line) is now advanced to CT1 f 0.49. (C) Effect of picrotoxin (10 PM), administered for 1 h at CT15, on average (n = 4) single-unit activity rhythm in the SCN recorded the following day. The time of peak (dotted line) is now advanced to CT23 + 0.67.
PICROTOXIN
71
PHASE ADVANCES RHYTHMS
(A) 1412-
-i
-I--.
0
3
6
9
I
12
15
18
21
24
7
1
27
30
33
36
27
30
33
36
Circadian Time (hr)
03
I
5
3 I_ 0
-. 3
6
I
9
12
15
’
I
18
’
I
21
-
.
24
Circadian lime (hr) FIG. 2. (A) Effect of ‘lTX (1 /IM) and Mg” (10 PM) coadministered for 1 h at CT6. Vertical bar indicates time of drug administration. The time of peak (CT6 2 0.78) on day 2 is not significantly different from that of the untreated slicesin Fig. iA. Format & in Fig.. 1. (B) Effect of nicrotoxin 110 uhf\, administered for 1 h at CBS, in the uresence of lTX il PM) and‘Mg+ (i0 PM) on average (n = 3) single-unit activity rhythm in the SCN recorded the following day. The time of peak (dotted line) is now advanced to CT1 5 0.46, indistinguishable from the time of peak in Fig. 1B.
Previous work has shown that GABAergic agonists (muscimol, baclofen, triazolam) can phase shift the circadian clock in vivo (24,25,29). Furthermore, GABA-A receptor blockade by bicuculline or chloride channel blockade by picrotoxin prevents the phase-shifting effect of muscimol, while benzodiazepine antagonists [e.g., RO 15-1788; (30)], selectively block the phase-shifting effects of triazolam. In addition, light-induced phase delays are blocked by bicuculline or baclofen and light-induced phase advances are blocked by diazepam or baclofen (22). However,
picrotoxin administered alone systemically or directly into the SCN at CT6, CI’13.5, or CT18 does not phase shift the locomotor rhythm in vivo nor does it block light-induced phase shifts. In contrast, chloride channel blockade in vitro with a relatively high [and efficacious; (15)] concentration of picrotoxin induces large phase advances in the single-unit activity rhythm of the SCN when administered in vitro at either CT6 or CT15 Inhibition of chloride channels in the retinal pacemaker neurons of Bulla produces phase advances, similar to those described here, in the circadian rhythm of firing rate (10). However, such apparent phase advances actually result from a shortening of the intrinsic period. We cannot exclude a similar period shortening effect of picrotoxin in this study (although such an effect has not been observed in the mamm~i~ SCN in viva). In fact, the observed phase-independent phase advance is consistent with such an explanation. Period measurement in the slice is limited by slice viability; slice recordings are limited to about 3 days in vitro, an insufficient survival time for accurate period, as opposed to phase, determination. Whether the apparent phase shift in vitro is truly a phase shift or, rather, period shortening, the discrepancy remains with the in vivo observations. In light of the ubiquitous distribution of GABA receptors in the CNS and the at best moderate binding of muscimol and benzodiazepines to GABA receptors in the SCN (ll), we suspect that the phase shifting effects of such agents in vivo, and the blockade of such effects by picrotoxin, are probably mediated by neuropil remote to the SCN [as has been repeatedly suggested by other authors; for review, (ll)]. Any direct effect of picrotoxin on the clock in vivo might be masked by such effects, but revealed in the reduced neuropil of the SCN slice preparation. The blockade of chloride channels can produce profound membrane depolarization. Certain depolarizing stimuli [potassium, (4); nicotine, (28)] can also induce large phase advances in the peak of single-unit activity rhythms or neuro~~smitter release in the SCN in vitro. For these reasons the effects of picrotoxin were also studied in the presence of agents that prevent the inward cation flux (primarily sodium and calcium) that is associated with membrane depolarization. This and previous studies (21) have shown that these agents (administered singly, or together as in this study) do not themselves alter the peak in the SCN single-unit activity rhythm recorded the next day, when administered at CT6 or CT15 on the previous day. However, when applied at the concentrations used in this study, these agents prevent the generation of action potentials and the release of neurotransmitters, that depend, respectively, on sodium and calcium fluxes, for at least the duration of drug administration (21,26). Most impo~antly, they did not alter the phase advances produced by administration of picrotoxin at either CT6 or CT1.5. A strong inference is that the phase-advancing effects of picrotoxin are mediated by the direct blockade of chloride flux across the cell membrane, rather than subsequent opening of voltagesensitive sodium and calcium channels. This is particularly surprising because the GABA agonist, muscimol, thought to open chloride channels, also produces phase advances in the circadian rhythm of locomotor activity when administered in subjective day in vivo (24). As suggested above, the muscimol effect may be mediated by remote neuropil [but presumably within the diffusion volume of an intra-SCN drug administration; (24)]. Alternatively, very different mechanisms may underlie the effects of muscimol (phase advance) as opposed to picrotoxin (period sho~ening). Finally, because the SCN neuropil is an almost entirely GABAergic network (18), probably involving complex patterns of inhibition and disinhibition, it is possible that any perturbation of GABA transmission away from some optimal level
72
RANGARAJAN,
could result in a similar alteration of clock parameters (e.g., phase or period). Such possibilities underscore the importance of future study of the effects of GABA agonists in the SCN slice for as long as is technically feasible. Phase shifts are thought to represent either the direct effect of a perturbing stimulus on the circadian oscillator or an indirect effect via systems afferent to that oscillator. Because picrotoxininduced phase shifts persist in the presence of agents that block action potential propagation and neurotransmitter release, it is possible that a chloride channel in the clock cell membrane is
HELLER
AND MILLER
to the phase shift. However, intercellular communication in the SCN does not necessarily require synaptic transmission (3,12). In any event, it appears that the chloride channel, and by implication, the GABA-A receptor that gates the chloride channel, may be intimately involved in the generation and/or modulation of circadian rhythmicity. necessary
ACKNOWLEDGEMENTS This work was supported
by a grant from the Upjohn Company
and
by NIA grant AG 11084-01.
REFERENCES 1. Araki, T.; Kiyama, H.; Tohyama, M. The GABA, receptor y, subunit is expressed by distinct neuronal populations. Mol. Brain Res. 15121-132; 1992. 2. Araki, T.; Tohyama, M. Region-specific expression of GABA.., receptor a3 and oq subunits mRNAs in the rat brain. Mol. Brain Res. 12:293-314; 1992. 3. Bousklla, Y.; Dudek, F. E. Synchronous neuronal activity in the suprachiasmatic nucleus (SCN) independent of chemical synaptic transmission. Sot. Neurosci. Abstr. 18:1416; 1992. 4. Earnest, D. J.; Sladek, C. S. Circadian vasopressin release from perifused rat suprachiasmatic explants in vitro: Effects of acute stimulation. Brain Res. 422:398-402; 1987. 5. Green, D. J.; Gillette, R. Circadian rhythm of firing rate recorded from single cells in the rat suprachiasmatic brain slice. Brain Res. 245:198-200; 1982. 6. Hatton, G. I.; Doran, A. D.; Salm, A. K.; Tweedle, C. D. Brain slice preparation: Hypothalamus. Brain Res. Bull. 5:405-414; 1980. 7. Liou, S. Y.; Albers, H. E. Single unit response of neurons within the hamster suprachiasmatic nucleus to GABA and low chloride perfusate during the day and night. Brain Res. Bull. 25:93-98; 1990. 8. Liou, S. Y.; Shibata, S.; Albers, H. E.; Ueki, S. Effects of GABA and anxiolytics on the single unit discharge of suprachiasmatic neurons in rat hypothalamic slices. Brain Res. Bull. 25:103-107; 1990. 9. Mason, R.; Biello, S. M.; Harrington, M. E. The effects of GABA and benxodiazepines on neurones in the suprachiasmatic nucleus (SCN) of Svrian hamsters. Brain Res. 55253-57: 1991. S. B. S.; Block, G. D. Phase shifting of the 10. Michel, S.;-Khalsa, circadian rhythm in the eye of Bullu by inhibition of chloride conductance. Neurosci. Lett. 146:219-222; 1992. 11. Michels, K. M.; Morin, L. P.; Moore, R. Y. GABA,&enxodiazepine receptor localization in the circadian timing system. Brain Res. 531:16-24; 1990. 12. Miller, J. D. On the nature of the circadian clock in mammals. Am. J. Physiol. 264:R821-R832; 1993. 13. Miller, J. D.; Cao, V. H.; Heller, H. C.; Kilduff, T. S. GABA-ergic modulation of 2-deoxyglucose uptake in the SCN in vitro. Sot. Res. Biol. Rhythms Abstr. 215; 1990. 14. Miller, J. D.; Fuller, C. A. Isoperiodic neuronal activity in the suprachiasmatic nucleus of the rat. Am. J. Physiol. 263:R51-R58; 1992. 15. Moss, S. J.; Smart, T. G.; Porter, N. M.; Nayeem, N.; Devine, J.; Stephenson, F. A.; Macdonald, R. L.; Barnard, E. A. Cloned GABA receptors are maintained in a stable cell line: Allosteric and channel properties. Eur. J. Pharmacol. 189:77-88; 1990. nucleus 16. Newman, G. C.; Hospod, F. E. Rhythm of suprachiasmatic 2-deoxyglucose uptake in vitro. Brain Res. 381:345-350; 1986. 17. Nishino, H.; Koizumi, K. Responses of neurons in the suprachiasmatic nuclei of the hypothalamus to putative transmitters. Brain Res. 120:167-172; 1977.
18. Okamura, H.; Berod, A.; Julien, I. F.; Geffard, M.; Kltahama, K.; Mallet, J.; Bobillier, P. Demonstration of GABAergic cell bodies in the suprachiasmatic nucleus: In situ hybridization of glutamic acid decarboxylase (GAD) mRNA and immunocytochemistry of GAD and GABA. Neurosci. Lett. 102:131-136; 1989. 19. Pavlidis, T. Populations of interacting oscillators and circadian rhythms. J. Theor. Biol. 22:418-436; 1969. 20. Prosser, R. A.; Gillette, M. U. The mammalian circadian clock in the suprachiasmatic nuclei is reset in vitro by CAMP. J. Neurosci. 9:1073-1081; 1989. 21. Prosser, R. A.; Miller, J. D.; Heller, H. C. Serotonergic phase shifts of the mammalian circadian clock: Effects of tetrodotoxin and high Mg*+. Brain Res. 573:336-340; 1992. 22. Ralph, M. R.; Menaker, M. GABA regulation of circadian responses to light. I. Involvement of GABAa-benzodiaxepine and GABA, receptors. J. Neurosci. 9:2858-2865; 1989. 23. Shibata, S.; Liou, S. Y.; Ueki, S. Effect of amino acids and monoamines on the neuronal activity of suprachiasmatic nucleus in hypothalamic slice preparations. Jpn. J. Pharmacol. 33:1225-1231; 1983. 24. Smith, R. D.; Inouye, S. T.; Turek, F. W. Central administration of muscimol phase-shifts the mammalian circadian clock. J. Comp. Physiol. 164:805-814; 1989. 25. Smith, R. D.; Turek, F. W.; Slater, N. T. Bicuculline and picrotoxin block phase advances induced by GABA agonists in the circadian rhythm of locomotor activity in the golden hamster by a phaclofeninsensitive mechanism. Brain Res. 530:275-282; 1990. 26. Sperling, E.; Cao, V. H.; Miller, J. D. Circadian variation in 5-HT release in vitro. Sot. Res. Biol. Rhythms Abstr. 3:102; 1992. 27. Tominaga, K.; Shibata, S.; Ueki, S.; Watanabe, S. Effects of inhibitory and excitatory drugs on the metabolic rhythm of the hamster suprachiasmatic nucleus in vitro. Eur. J. Pharmacol. 217~79-84; 1992. 28. Trachsel, L.; Heller, H. C.; Dement, W. C.; Miller, J. D. Effect of nicotine and carbachol on the circadian rhythm of the suprachiasmatic nuclei in vitro. Sot. Neurosci. Abstr. 17:672; 1991. S. H. A benzodiazepine used in the 29. Turek, F. W.; Losee-Olson, treatment of insomnia phase-shifts the mammalian circadian clock. Nature 321:167-168; 1986. J. J.; Turek, F. W. A benxodiazepine 30. Van Reeth, 0.; Vanderhaeghen, antagonist, RO 15-1788 c&r block the phase-shifting effects of triazolam on the mammalian circadian clock. Brain Res. 444:333-339: 1988. 31. Wisden, W.; Laurie, D. J.; Monyer, H.; Seeburg, P. H. The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon. J. Neurosci. 12: 1040-1062; 1992.
Pergamon 0361-9230(93)EOO36-L
BRIEF COMMUNICATION
Chloride Channel Block Phase Advances the Single-Unit Activity Rhythm in the SCN RADHA
RANGARAJAN,*
H. CRAIG
HELLER,*f
AND
JOSEPH
D. MILLER**’
*Department of Biological Sciences, fstanford Center for Sleep and Circadian Neurobiology, and #Department of Psychiatry, Stanford University, Stanford, CA 94305 Received
23 April 1993; Accepted
30 November
1993
RANGARAJAN, R., H. C. HELLER AND J. D. MILLER. Chloride channel blockphase advances the single-unit activity rhythm in the SCN. BRAIN RES BULL 34(l) 69-72, 1994.-The mammalian suprachiasmatic nuclei (SCN) contain a circadian pacemaker that exhibits a 24 h rhythm in single-unit activity in vivo and in vitro. Chloride channel block by a saturating concentration of picrotoxin at either CT6 or CT15 produces large phase advances in the SCN single-unit activity rhythm in vitro. These phase advances are not affected by simultaneous blockade of voltage-sensitive sodium and calcium channels by ‘lTX and magnesium. Thus, the effects of picrotoxin appear to be mediated by direct blockade of the chloride channel, rather than subsequent membrane depolarization. GABA-A receptor-mediated chloride flux may be part of the mechanism of circadian timekeeping. Suprachiasmatic
nucleus
Circadian
rhythm
Chloride
Picrotoxin
to circadian timekeeping, chloride channel blockade should strongly perturb the circadian oscillator. For that reason the effects on the SCN slice of the chloride channel blocker, picrotoxin, were studied at two circadian times. Because picrotoxin is a strong depolarizing agent, its effects were also studied under joint fast sodium and calcium channel blockade to determine whether observed phase shifting effects were secondary to membrane depolarization and subsequent intracellular cation flux.
THE suprachiasmatic nuclei (SCN) of the hypothalamus are necessary and sufficient for the maintenance of circadian rhythmicity in mammals [for a recent review, see (12)]. Circadian rhythms in single-unit activity, neurotransmitter release, and cellular metabolism are maintained in vitro in the hypothalamic slice preparation (4,516). While the cellular substrate of the circadian pacemaker within the SCN has not been determined, virtually all the neurons of the SCN are GABAergic in nature (18). Mathematical models of the circadian pacemaker have suggested an important role for inhibitory coupling, possibly of a GABAergic nature, among component oscillators in a neural network in the SCN (12,14,19). GABAergic agonists profoundly reduce cellular metabolism in the SCN when administered to the SCN slice during the extrapolated light portion of the 1ight:dark cycle, but not in the extrapolated dark portion (13,27). In contrast, most neurons of the SCN are inhibited by iontophoretic or bath application of GABA, independent of time of day (7-9,17,23). The GABAergic agonist, muscimol, produces phase advances of the circadian rhythm of locomotor activity when administered in vivo in subjective day, but phase delays when administered in subjective night (24). The GABA-A receptor subtype is present in the SCN, but may have a somewhat variant subunit composition [most likely (Ye& y2; (1,2,31)], perhaps explaining the difficulties that some investigators have had in unequivocally demonstrating its presence (11). This GABA receptor subtype ultimately gates a chloride channel. If a local GABAergic intemeuronal network is essential 1 Requests
for reprints
should be addressed
GABA
METHOD The methods used in these experiments have been described in detail earlier (20). Briefly, male Wistar rats were housed on a 12:12 1ight:dark cycle. They were sacrificed during the light portion of their cycles and hypothalamic brain slices, 500 pm thick, were prepared. Slices containing the SCN were maintained in a Hatton-style brain slice chamber (6) at a temperature of 37°C. The slices were superfused at a rate of 20 ml/h with Earle’s balanced salt solution (Sigma), aerated with a 95% O#% CO2 mixture, and supplemented with 24.6 mM glucose and 26.2 mM sodium bicarbonate, pH 7.4. The first set of experiments tested the capacity of picrotoxin to alter the normal time of peak in the single unit activity rhythm. Slices were treated with a chloride channel saturating concentration of picrotoxin at CT6 and CT15. Superfusion was interrupted for an hour at the appropriate time (CT6 or CT15) and the control medium in the chamber was replaced with test medium contain-
to Dr. Joseph Miller, Department
69
of Biological
Sciences,
Stanford
University,
Stanford,
CA 94305.
70
RANGARAJAN,
(A) a
2
I , .1
3
6
9
12
15
Circadian
18
21
Time
24
27
30
(hr)
33
36
(B) 10
HELLER
AND MILLER
ing picrotoxin at a concentration of 10 PM. The test medium was adjusted for pH, temperature, and oxygen concentration before exchanging it with the control medium. After an hour the test medium was removed, replaced with control medium, and superfusion was reinstated. Control experiments to confirm the time of peak activity in normal slices in the absence of any drugs were performed in an identical fashion, but in the absence of picrotoxin. A second set of experiments examined the effects of 10 ,uM picrotoxin in the presence of 1 PM tetrodotoxin (to block fast sodium channels and subsequent action potentials) and 10 mM magnesium chloride (to block voltage-sensitive calcium channels), using the same method as described above. Control experiments were performed with 1 PM tetrodotoxin and 10 mM magnesium chloride to determine whether these cation channel blockers had any independent effect on the rhythm in single-unit activity. The effect of drug treatment was determined by monitoring the rhythm of spontaneous neuronal activity during the first circadian cycle after preparation and treatment of slices. The single unit activity rhythm was determined via extracellular recording from single neurons, using glass micropipettes filled with 3 M NaCl. Action potentials from single cells, encountered sequentially in electrode tracks run through the SCN, were recorded. The firing rate of each cell was averaged over a 5 min period, and cells were recorded sequentially for 8 to 10 h per day. Data were collected and analyzed using a BrainWave system. The firing rates of individual cells were averaged over successive 2 h intervals lagged by 1 h. The time of peak activity was then determined. Phase shifts were determined by comparing the time of peak activity in the presence of the pharmacological manipulation with the time of peak activity in control slices.
RESULTS
2
0
, . 3
9
12
15
18
Circadian (C)
21
Time
24
27
30
33
36
(hr)
lo
2 3
6
9
12
15
Circadian
18
21
Time
24
(hr)
27
30
33
36
The first set of control experiments confirmed that the time of peak activity was at CT5 t 0.56 h (n = 3) (Fig. lA), in good agreement with previous work (e.g., 21). One hour treatment with picrotoxin (10 PM) at CT6 caused a 4-5 h advance in time of peak activity the next day so that the new peak was seen at CT1 + 0.49 h (n = 3) (Fig. 1B). The same treatment at CT 15 caused the new peak to appear at CT23 ? 0.67 (n = 4 ) (Fig. 1C). Two control experiments with 1 PM tetrodotoxin and 10 mM magnesium chloride joint treatment at CT6 showed a normal time of peak activity the next day at CT6 ? 0.78 (n = 2) (Fig. 2A). Action potentials were not detected during the 1 h treatment. One hour treatment with 10 PM picrotoxin at CT6, in the presence of 1 PM tetrodotoxin and 10 mM magnesium chloride, induced a 5-6 h advance in the time of peak activity on the following day. The peak activity was seen at CT1 ? 0.46 (n = 3) (Fig. 2B).
FIG. 1. (A) Average (n = 3 experiments) single-unit activity rhythm recorded from the SCN in vitro. Plotted are the grand averages -C SE across all untreated control experiments at each CT of the 2 h means of firing rates of single SCN cells. Slices are prepared on day 1. The time of peak (dotted line) occurs on day 2 at CT5 -t 0.56. Stippled horizontal bar = lights off in animal colony. (B) Effect of picrotoxin (10 PM), administered for 1 h at cT6, on average (n = 3) single-unit activity rhythm in the SCN recorded the following day. Vertical bar indicates time of drug administration. The time of peak (dotted line) is now advanced to CT1 f 0.49. (C) Effect of picrotoxin (10 PM), administered for 1 h at CT15, on average (n = 4) single-unit activity rhythm in the SCN recorded the following day. The time of peak (dotted line) is now advanced to CT23 + 0.67.
PICROTOXIN
71
PHASE ADVANCES RHYTHMS
(A) 1412-
-i
-I--.
0
3
6
9
I
12
15
18
21
24
7
1
27
30
33
36
27
30
33
36
Circadian Time (hr)
03
I
5
3 I_ 0
-. 3
6
I
9
12
15
’
I
18
’
I
21
-
.
24
Circadian lime (hr) FIG. 2. (A) Effect of ‘lTX (1 /IM) and Mg” (10 PM) coadministered for 1 h at CT6. Vertical bar indicates time of drug administration. The time of peak (CT6 2 0.78) on day 2 is not significantly different from that of the untreated slicesin Fig. iA. Format & in Fig.. 1. (B) Effect of nicrotoxin 110 uhf\, administered for 1 h at CBS, in the uresence of lTX il PM) and‘Mg+ (i0 PM) on average (n = 3) single-unit activity rhythm in the SCN recorded the following day. The time of peak (dotted line) is now advanced to CT1 5 0.46, indistinguishable from the time of peak in Fig. 1B.
Previous work has shown that GABAergic agonists (muscimol, baclofen, triazolam) can phase shift the circadian clock in vivo (24,25,29). Furthermore, GABA-A receptor blockade by bicuculline or chloride channel blockade by picrotoxin prevents the phase-shifting effect of muscimol, while benzodiazepine antagonists [e.g., RO 15-1788; (30)], selectively block the phase-shifting effects of triazolam. In addition, light-induced phase delays are blocked by bicuculline or baclofen and light-induced phase advances are blocked by diazepam or baclofen (22). However,
picrotoxin administered alone systemically or directly into the SCN at CT6, CI’13.5, or CT18 does not phase shift the locomotor rhythm in vivo nor does it block light-induced phase shifts. In contrast, chloride channel blockade in vitro with a relatively high [and efficacious; (15)] concentration of picrotoxin induces large phase advances in the single-unit activity rhythm of the SCN when administered in vitro at either CT6 or CT15 Inhibition of chloride channels in the retinal pacemaker neurons of Bulla produces phase advances, similar to those described here, in the circadian rhythm of firing rate (10). However, such apparent phase advances actually result from a shortening of the intrinsic period. We cannot exclude a similar period shortening effect of picrotoxin in this study (although such an effect has not been observed in the mamm~i~ SCN in viva). In fact, the observed phase-independent phase advance is consistent with such an explanation. Period measurement in the slice is limited by slice viability; slice recordings are limited to about 3 days in vitro, an insufficient survival time for accurate period, as opposed to phase, determination. Whether the apparent phase shift in vitro is truly a phase shift or, rather, period shortening, the discrepancy remains with the in vivo observations. In light of the ubiquitous distribution of GABA receptors in the CNS and the at best moderate binding of muscimol and benzodiazepines to GABA receptors in the SCN (ll), we suspect that the phase shifting effects of such agents in vivo, and the blockade of such effects by picrotoxin, are probably mediated by neuropil remote to the SCN [as has been repeatedly suggested by other authors; for review, (ll)]. Any direct effect of picrotoxin on the clock in vivo might be masked by such effects, but revealed in the reduced neuropil of the SCN slice preparation. The blockade of chloride channels can produce profound membrane depolarization. Certain depolarizing stimuli [potassium, (4); nicotine, (28)] can also induce large phase advances in the peak of single-unit activity rhythms or neuro~~smitter release in the SCN in vitro. For these reasons the effects of picrotoxin were also studied in the presence of agents that prevent the inward cation flux (primarily sodium and calcium) that is associated with membrane depolarization. This and previous studies (21) have shown that these agents (administered singly, or together as in this study) do not themselves alter the peak in the SCN single-unit activity rhythm recorded the next day, when administered at CT6 or CT15 on the previous day. However, when applied at the concentrations used in this study, these agents prevent the generation of action potentials and the release of neurotransmitters, that depend, respectively, on sodium and calcium fluxes, for at least the duration of drug administration (21,26). Most impo~antly, they did not alter the phase advances produced by administration of picrotoxin at either CT6 or CT1.5. A strong inference is that the phase-advancing effects of picrotoxin are mediated by the direct blockade of chloride flux across the cell membrane, rather than subsequent opening of voltagesensitive sodium and calcium channels. This is particularly surprising because the GABA agonist, muscimol, thought to open chloride channels, also produces phase advances in the circadian rhythm of locomotor activity when administered in subjective day in vivo (24). As suggested above, the muscimol effect may be mediated by remote neuropil [but presumably within the diffusion volume of an intra-SCN drug administration; (24)]. Alternatively, very different mechanisms may underlie the effects of muscimol (phase advance) as opposed to picrotoxin (period sho~ening). Finally, because the SCN neuropil is an almost entirely GABAergic network (18), probably involving complex patterns of inhibition and disinhibition, it is possible that any perturbation of GABA transmission away from some optimal level
72
RANGARAJAN,
could result in a similar alteration of clock parameters (e.g., phase or period). Such possibilities underscore the importance of future study of the effects of GABA agonists in the SCN slice for as long as is technically feasible. Phase shifts are thought to represent either the direct effect of a perturbing stimulus on the circadian oscillator or an indirect effect via systems afferent to that oscillator. Because picrotoxininduced phase shifts persist in the presence of agents that block action potential propagation and neurotransmitter release, it is possible that a chloride channel in the clock cell membrane is
HELLER
AND MILLER
to the phase shift. However, intercellular communication in the SCN does not necessarily require synaptic transmission (3,12). In any event, it appears that the chloride channel, and by implication, the GABA-A receptor that gates the chloride channel, may be intimately involved in the generation and/or modulation of circadian rhythmicity. necessary
ACKNOWLEDGEMENTS This work was supported
by a grant from the Upjohn Company
and
by NIA grant AG 11084-01.
REFERENCES 1. Araki, T.; Kiyama, H.; Tohyama, M. The GABA, receptor y, subunit is expressed by distinct neuronal populations. Mol. Brain Res. 15121-132; 1992. 2. Araki, T.; Tohyama, M. Region-specific expression of GABA.., receptor a3 and oq subunits mRNAs in the rat brain. Mol. Brain Res. 12:293-314; 1992. 3. Bousklla, Y.; Dudek, F. E. Synchronous neuronal activity in the suprachiasmatic nucleus (SCN) independent of chemical synaptic transmission. Sot. Neurosci. Abstr. 18:1416; 1992. 4. Earnest, D. J.; Sladek, C. S. Circadian vasopressin release from perifused rat suprachiasmatic explants in vitro: Effects of acute stimulation. Brain Res. 422:398-402; 1987. 5. Green, D. J.; Gillette, R. Circadian rhythm of firing rate recorded from single cells in the rat suprachiasmatic brain slice. Brain Res. 245:198-200; 1982. 6. Hatton, G. I.; Doran, A. D.; Salm, A. K.; Tweedle, C. D. Brain slice preparation: Hypothalamus. Brain Res. Bull. 5:405-414; 1980. 7. Liou, S. Y.; Albers, H. E. Single unit response of neurons within the hamster suprachiasmatic nucleus to GABA and low chloride perfusate during the day and night. Brain Res. Bull. 25:93-98; 1990. 8. Liou, S. Y.; Shibata, S.; Albers, H. E.; Ueki, S. Effects of GABA and anxiolytics on the single unit discharge of suprachiasmatic neurons in rat hypothalamic slices. Brain Res. Bull. 25:103-107; 1990. 9. Mason, R.; Biello, S. M.; Harrington, M. E. The effects of GABA and benxodiazepines on neurones in the suprachiasmatic nucleus (SCN) of Svrian hamsters. Brain Res. 55253-57: 1991. S. B. S.; Block, G. D. Phase shifting of the 10. Michel, S.;-Khalsa, circadian rhythm in the eye of Bullu by inhibition of chloride conductance. Neurosci. Lett. 146:219-222; 1992. 11. Michels, K. M.; Morin, L. P.; Moore, R. Y. GABA,&enxodiazepine receptor localization in the circadian timing system. Brain Res. 531:16-24; 1990. 12. Miller, J. D. On the nature of the circadian clock in mammals. Am. J. Physiol. 264:R821-R832; 1993. 13. Miller, J. D.; Cao, V. H.; Heller, H. C.; Kilduff, T. S. GABA-ergic modulation of 2-deoxyglucose uptake in the SCN in vitro. Sot. Res. Biol. Rhythms Abstr. 215; 1990. 14. Miller, J. D.; Fuller, C. A. Isoperiodic neuronal activity in the suprachiasmatic nucleus of the rat. Am. J. Physiol. 263:R51-R58; 1992. 15. Moss, S. J.; Smart, T. G.; Porter, N. M.; Nayeem, N.; Devine, J.; Stephenson, F. A.; Macdonald, R. L.; Barnard, E. A. Cloned GABA receptors are maintained in a stable cell line: Allosteric and channel properties. Eur. J. Pharmacol. 189:77-88; 1990. nucleus 16. Newman, G. C.; Hospod, F. E. Rhythm of suprachiasmatic 2-deoxyglucose uptake in vitro. Brain Res. 381:345-350; 1986. 17. Nishino, H.; Koizumi, K. Responses of neurons in the suprachiasmatic nuclei of the hypothalamus to putative transmitters. Brain Res. 120:167-172; 1977.
18. Okamura, H.; Berod, A.; Julien, I. F.; Geffard, M.; Kltahama, K.; Mallet, J.; Bobillier, P. Demonstration of GABAergic cell bodies in the suprachiasmatic nucleus: In situ hybridization of glutamic acid decarboxylase (GAD) mRNA and immunocytochemistry of GAD and GABA. Neurosci. Lett. 102:131-136; 1989. 19. Pavlidis, T. Populations of interacting oscillators and circadian rhythms. J. Theor. Biol. 22:418-436; 1969. 20. Prosser, R. A.; Gillette, M. U. The mammalian circadian clock in the suprachiasmatic nuclei is reset in vitro by CAMP. J. Neurosci. 9:1073-1081; 1989. 21. Prosser, R. A.; Miller, J. D.; Heller, H. C. Serotonergic phase shifts of the mammalian circadian clock: Effects of tetrodotoxin and high Mg*+. Brain Res. 573:336-340; 1992. 22. Ralph, M. R.; Menaker, M. GABA regulation of circadian responses to light. I. Involvement of GABAa-benzodiaxepine and GABA, receptors. J. Neurosci. 9:2858-2865; 1989. 23. Shibata, S.; Liou, S. Y.; Ueki, S. Effect of amino acids and monoamines on the neuronal activity of suprachiasmatic nucleus in hypothalamic slice preparations. Jpn. J. Pharmacol. 33:1225-1231; 1983. 24. Smith, R. D.; Inouye, S. T.; Turek, F. W. Central administration of muscimol phase-shifts the mammalian circadian clock. J. Comp. Physiol. 164:805-814; 1989. 25. Smith, R. D.; Turek, F. W.; Slater, N. T. Bicuculline and picrotoxin block phase advances induced by GABA agonists in the circadian rhythm of locomotor activity in the golden hamster by a phaclofeninsensitive mechanism. Brain Res. 530:275-282; 1990. 26. Sperling, E.; Cao, V. H.; Miller, J. D. Circadian variation in 5-HT release in vitro. Sot. Res. Biol. Rhythms Abstr. 3:102; 1992. 27. Tominaga, K.; Shibata, S.; Ueki, S.; Watanabe, S. Effects of inhibitory and excitatory drugs on the metabolic rhythm of the hamster suprachiasmatic nucleus in vitro. Eur. J. Pharmacol. 217~79-84; 1992. 28. Trachsel, L.; Heller, H. C.; Dement, W. C.; Miller, J. D. Effect of nicotine and carbachol on the circadian rhythm of the suprachiasmatic nuclei in vitro. Sot. Neurosci. Abstr. 17:672; 1991. S. H. A benzodiazepine used in the 29. Turek, F. W.; Losee-Olson, treatment of insomnia phase-shifts the mammalian circadian clock. Nature 321:167-168; 1986. J. J.; Turek, F. W. A benxodiazepine 30. Van Reeth, 0.; Vanderhaeghen, antagonist, RO 15-1788 c&r block the phase-shifting effects of triazolam on the mammalian circadian clock. Brain Res. 444:333-339: 1988. 31. Wisden, W.; Laurie, D. J.; Monyer, H.; Seeburg, P. H. The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon. J. Neurosci. 12: 1040-1062; 1992.