- Email: [email protected]
Systemic hemodynamic changes raising brain temperature in REM sleep
Brain Research 940 (2002) 55–60 www.elsevier.com / locate / bres
Research report
Systemic hemodynamic changes raising brain temperature in REM sleep Pier Luigi Parmeggiani*, Adele Azzaroni, Marcella Calasso Dipartimento di Fisiologia Umana e Generale, Universita` di Bologna, Piazza Porta San Donato 2, I-40127 Bologna, Italy Accepted 16 November 2001
Abstract The present research studied the mechanisms underlying the increase in brain temperature during REM sleep in the unrestrained rabbit carrying chronically implanted electrodes, thermistors and common carotid artery occluders. During the ultradian wake–sleep cycle at constant ambient temperature (2562 8C), we recorded: (i) the ear pinna temperature as an indirect indicator of blood flow affecting heat loss from the systemic heat exchangers of the head, (ii) the temperature of the pons and hypothalamus as indirect indicators of the temperature of vertebral artery blood (systemic cooling only) and carotid artery blood (both systemic and selective cooling), respectively, and (iii) the changes induced in these temperatures by short-lasting bilateral common carotid artery occlusion. The results show that during REM sleep both systemic and selective brain cooling are depressed by a spontaneous decrease in the common carotid artery blood flow and the associated autoregulatory increase in the vertebral artery share of the cerebral blood supply. 2002 Elsevier Science B.V. All rights reserved. Theme: Other systems of the CNS Topic: Brain metabolism and blood flow Keywords: Systemic brain cooling; Selective brain cooling; Cerebral blood flow; Hypothalamic temperature; Pontine temperature; Ultradian wake–sleep cycle
1. Introduction In the cat [10], the carotid artery blood supply to the brain is thermally conditioned once more prior to entering the circle of Willis (selective brain cooling) by countercurrent heat exchange between a network of fine vessels (the carotid rete) in contact with the cranial venous plexuses receiving cool venous blood from the systemic heat exchangers of the head. Vertebral artery blood is not thermally conditioned by selective cooling and enters the circle of Willis at the same temperature as aortic arch blood, which has been cooled only through the venous blood returning from the systemic heat exchangers of the whole body to the heart (systemic brain cooling). This explains the existence of a difference between vertebral artery blood temperature (systemic cooling only) and *Corresponding author. Tel.: 139-051-244-499; fax: 139-051-251731. E-mail address: [email protected] (P.L. Parmeggiani).
carotid artery blood temperature (both systemic and selective cooling). Practically, this difference may be indirectly appraised by computing the difference between pontine and hypothalamic temperatures, since they depend primarily on vertebral and carotid artery blood temperature, respectively [3,4,15]. In the cat at ambient temperatures close to thermal neutrality [2], NREM sleep (non-rapid eye movement (nrem) sleep showing low frequency and high amplitude electroencephalogram) is characterized by an increased blood flow in the systemic heat exchangers of the head as shown by the rise in ear pinna temperature [3,4]. This is due to the state-dependent decrease in tonic vasoconstrictor sympathetic outflow and the head-down posture (decrement of negative hydrostatic load and increment of transmural pressure in the vessels of heat exchangers) with respect to quiet wakefulness (QW). As a result of the increased systemic heat loss enhancing also selective heat loss (s. above), both pontine and hypothalamic temperatures are lower in NREM sleep than during QW and their
0006-8993 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 02 )02591-X
56
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
difference is increased. The evidence of a larger temperature change in the hypothalamus than in the pons shows the cumulative influence of systemic and selective arterial blood cooling at hypothalamic level. These effects are the obvious result of mechanisms which, however, do not underlie the pontine and hypothalamic temperature changes observed during REM sleep (rapid eye movement (rem) sleep showing high frequency and low amplitude electroencephalogram). Experimentally induced variations in heat loss from the systemic heat exchangers of the head exert only a weak effect on the spontaneously occurring increase in pontine and hypothalamic temperatures in this sleep state [3]. This is a result of a state-dependent decrease in the common carotid artery blood supply to the heat exchangers of the head and to the carotid rete, which also implies an autoregulatory increase in the supply of vertebral artery blood to the circle of Willis [3]. The more pronounced rise in hypothalamic temperature with respect to pontine temperature in the cat shows the cumulative effect of the depression of both systemic and selective heat loss on hypothalamic temperature during REM sleep [3]. In conclusion, the spontaneous hemodynamic change in the common carotid artery bed during REM sleep, impairing the efficiency of both systemic and selective brain cooling, would explain the increase in hypothalamic temperature characterizing this sleep state in several species, regardless of the exposure to a wide range of ambient temperatures [1,3,9,14,16,17] above and below ambient thermal neutrality [2]. Concerning the contribution of the systemic heat exchangers of the head (the nasal mucosa and ear pinna) to selective brain cooling, studies in sheep have shown that selective cooling depends mainly on heat loss from the nasal mucosa in this species [5,11]. This may be the case for other species with a carotid rete, like the cat. The heat loss from the large surface of the nasal mucosa must surely cool the venous blood flowing in contact with the carotid rete (selective cooling) more than the heat loss from the small surface of the ear pinna cools the venous blood flowing directly to the heart (systemic cooling). This explains why hemodynamic changes in the common carotid artery bed influence the hypothalamic more than the pontine temperature in the cat. On the basis of the previous results showing the dramatic systemic hemodynamic changes raising brain temperature during REM in the cat, the present research investigated comparatively the changes in systemic and selective brain cooling during REM sleep in the rabbit. In this species, selective brain cooling is provided by conductive heat exchange between the basal portion of the brain (including the circle of Willis) and the cranial venous lakes which drain cool venous blood from the systemic heat exchangers of the head [6]. The anatomical differences between the cat, with a carotid rete and a small ear pinna, and the rabbit, lacking a carotid rete and with a large ear pinna, may underlie quantitatively different influences of
hemodynamic changes in the common carotid artery bed on systemic and selective brain cooling. The results of this study indicate that the spontaneously occurring decrease in common carotid artery blood flow during REM sleep affects systemic brain cooling more than selective brain cooling in the rabbit. Moreover, such hemodynamic change primarily underlies an autoregulatory increase in vertebral artery blood flow warming the brain, since this blood is not thermally conditioned by selective cooling.
2. Material and methods Three adult rabbits (New Zealand) were used. Chronic implantation of (i) extradural screw EEG and wire EMG electrodes, (ii) needle thermistors (Yellow Springs) in the anterior hypothalamic region and in the tegmental field of the upper pons to record the respective temperatures (T hy and T p ), and (iii) air-inflatable occluders placed around the common carotid arteries, was carried out under general (clonazepam, 0.5 mg / kg i.m.; sodium pentobarbital, 40 mg / kg i.p.) anesthesia. Ear pinna temperature (T ep ) was recorded bilaterally by means of surface thermistors (Yellow Springs) since it indicates blood flow changes in the systemic heat exchangers of the head. The rationale for the experiment of short bilateral common carotid artery occlusion, a procedure already applied in the cat [3], is that it influences the mechanisms of both systemic and selective brain cooling and that these effects can be examined separately by considering the changes in pontine and hypothalamic temperature, respectively. The experiments have been carried out according to the European Communities Council Directive (86 / 609 / EEC), under the supervision of the Central Veterinary Service of the University of Bologna and the Ministry of Health, Rome.
2.1. Experimental procedures The experimental sessions lasted 5–6 h at 2562 8C, an ambient temperature close to thermal neutrality [2], in a sound-attenuated, thermoregulated chamber. The animal was unrestrained in a plexiglas box. Preliminary sessions allowed the animal to adapt to the experimental condition to display regularly the natural sequence of ultradian wake–sleep cycles. A polygraph (Grass 7D) recorded EEG and EMG activity, and the temperatures under study. Short-lasting occlusions (duration, 40–60s) of both common carotid arteries were carried out during the ultradian wake–sleep cycle by means of the air-inflatable occluders chronically placed around the vessels.
2.2. Histological control At the end of the experimental sessions, the animal was
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
sacrificed by means of a large dose of sodium pentobarbital (i.p.) and histological sections of the brain (fixed in formalin, embedded in celloidin, stained with the Nissl method) were prepared for determination of thermistor locations in the hypothalamus and pons.
2.3. Data analysis T p and T hy were measured in NREM sleep just before the beginning of REM sleep and at the end of their steep rising phase during REM sleep. The coincident values of T ep were also measured. The daily average values of the three variables were calculated for statistical analysis (5 days for each animal). The slopes of the T p and T hy rise in REM sleep and the differences between coincident T p and T hy values (T p 2T hy ), measured in NREM and REM sleep, were statistically compared by means of the two-tailed paired t-test.
57
3. Results An example of polygraphic recordings of spontaneous changes in T ep , T p and T hy during the ultradian sleep cycle is given in Fig. 1. During NREM sleep, such changes were small at constant ambient temperature. At REM sleep onset, a pronounced and constant decrease in T ep occurred which was related to a steep increase in both T p and T hy . The slope of the T hy change was steeper than that of the T p change (Table 1). Nevertheless, T p was always higher than T hy but the difference (T p 2T hy ) decreased during REM sleep (Table 1). At the end of the REM sleep episode the increase in T ep was associated with a decrease in T p and T hy . The effects of short-lasting (40–60 s) occlusion of both common carotid arteries on T ep , T p and T hy during NREM sleep are shown in Fig. 1. The bilateral carotid artery occlusion decreased T ep and increased both T p and T hy also in QW. These experimentally induced changes were simi-
Fig. 1. Changes in T ep , T p and T hy during the ultradian sleep cycle (ambient temperature, 23.6 8C). During NREM sleep, T ep increases slightly, whereas both T p and T hy show slight opposite changes. During REM sleep, T ep markedly decreases, whereas T p and T hy markedly increase. Note that the rise in T hy is steeper than that of T p . The effects of bilateral common carotid artery occlusion on T ep , T p and T hy are shown during NREM sleep. Note that T ep decreases and both T p and T hy increase as during REM sleep. However, the induced temperature changes are smaller with respect to those spontaneously occurring during REM sleep. EEG, electroencephalogram; EMG, electromyogram; T ep , left (l) and right (r) ear pinna temperature; T p , pontine temperature; T hy , hypothalamic temperature; BCO, bilateral common carotid artery occlusion; REMS, REM sleep.
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
58
Table 1 Slopes (DT p , DT hy ) of pontine and hypothalamic temperature increases from NREM to REM sleep (nrem→rem) and differences (T p 2T hy ) between the coincident values of pontine and hypothalamic temperatures just before the onset of (nrem) and during REM sleep (rem) DT p nrem→ rem DT hy nrem→ rem
0.0860.006 8C / min 0.1260.006 8C / min
P,0.001
df 14
(T p -T hy ) nrem (T p -T hy ) rem
0.1460.004 8C 0.0660.005 8C
P,0.001
df 14
lar but always smaller (independently of baseline) than those occurring spontaneously during REM sleep. During REM sleep onset, the bilateral common carotid artery occlusion did not significantly affect the spontaneous steep decrease in T ep and increase in T p and T hy (Fig. 2). In contrast, the bilateral common carotid artery occlusion just after the end of the REM sleep episode stopped both the T ep rise and the T p and T hy decrease (Fig. 2).
4. Discussion The results show that the changes in T ep , T p and T hy during REM sleep in the rabbit are qualitatively comparable with those previously observed in the cat [3]. However, remarkable species-specific features exist from the quantitative viewpoint. The fact that the T p and T hy rises in REM sleep are closer in amplitude in the rabbit than in the cat shows the greater relevance of systemic brain cooling with respect to selective brain cooling in the former species. In spite of a marked decrease in T ep occurring in both species during REM sleep at ambient temperatures slightly below thermal neutrality [2], systemic brain cooling is depressed more in the rabbit than in the cat, as shown by the larger T p increase in the former. It is likely that such a difference depends on the extension of the ear pinna surface, a heat exchanger prevalently affecting systemic brain cooling. On the other hand, selective brain cooling is probably less
Fig. 2. Changes in T ep , T p and T hy during the ultradian sleep cycle (ambient temperature: 25.8 8C). The effects of bilateral common carotid artery occlusion on T ep , T p and T hy are shown at the onset and just after the end of REM sleep. Note that the occlusion is not significantly affecting the spontaneous steep decrease in T ep and increase in T p and T hy . In contrast, the occlusion after the end of the REM sleep episode stops both the T ep rise and the T p and T hy decrease. EEG, electroencephalogram; EMG, electromyogram; T ep , left (l) and right (r) ear pinna temperature; T p , pontine temperature; T hy , hypothalamic temperature; BCO, bilateral common carotid artery occlusion; REMS, REM sleep.
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
efficient in the rabbit than in the cat since the thermal inertia of conductive heat exchange is greater than that of countercurrent heat exchange in the latter species. Nevertheless, the cumulative effect of the depression of both systemic and selective brain cooling is also revealed in the rabbit at REM sleep onset by the steeper slope of the T hy increase with respect to that of the T p increase. Concerning the systemic hemodynamic mechanisms underlying T ep , T p and T hy changes during REM sleep, the results show that they do not differ in the rabbit with respect to the cat [3]. In particular, the marked spontaneous decrease in T ep and increase in T p and T hy , characterizing REM sleep onset also in the rabbit, is practically unaffected by bilateral common carotid artery occlusion as in the cat. A reasonable inference is that during REM sleep common carotid artery blood flow is spontaneously decreased [3]. Consistent with this view is the fact that the return to normal hemodynamic conditions of cerebral blood perfusion, after the end of the REM sleep episode, is clearly affected in the rabbit by bilateral common carotid artery occlusion which stops the T ep rise and both the T p and T hy decrease as in the cat (Ref. [3], Fig. 1B). Moreover, the decrease in carotid artery blood supply to the brain ought to be buffered by an increased amount of vertebral artery blood flowing into the circle of Willis. In other words, REM sleep is characterized by a shift from the carotid artery to the vertebral artery (and probably also to other arterial sources, cf. Ref. [7]) in the amount of blood contributed to the overall cerebral blood flow. The inference of a spontaneous decrease in carotid artery blood flow during REM sleep in both cats [3] and rabbits is consistent with the alteration in cardiovascular regulation characterizing this behavioral state (cf. Ref. [13]). For instance, a sudden expansion of extracerebral common carotid artery beds, as a result of an altered vasoconstrictor control and a posturally (antigravitary muscle atonia) induced rise in transmural pressure, may underlie a cerebral blood ‘steal’ decreasing the amount of carotid artery blood flowing into the circle of Willis [3]. In this respect, it is worth considering that T ep , T p and T hy changes elicited by short bilateral common carotid artery occlusion during QW and NREM sleep are qualitatively similar but much less pronounced than the spontaneous changes observed in REM sleep. The likely explanation is that systemic baroreceptive reflexes provide a more effective buffering of the blood flow drop in the common carotid artery bed during QW and NREM sleep in comparison to REM sleep (cf. Ref. [3], p. 137) which is characterized by a marked alteration in autonomic homeostatic regulation (cf. Ref. [13]). The indirect evidence of a decrease in the carotid artery blood supply to the circle of Willis is apparently inconsistent with the overall increase in cerebral blood flow observed during REM sleep with respect to NREM sleep in several species (cf. Refs. [8,12]), including the cat and the rabbit. According to the present data, however, such an
59
increase may be considered, at least partially, an autoregulatory response necessarily involving the vertebral artery (and probably other arterial blood sources, cf. Ref. [7]). This view is supported by the fact that blood flow during REM sleep increases more in the cerebellum and brain stem than in the cerebrum (cf. Ref. [8]). Also the early steep rise in hypothalamic and pontine temperatures, followed by a temperature plateau, is an indirect proof of the initial warming effect of an increased vertebral artery blood flow, since this blood is not thermally conditioned by selective cooling like the carotid artery blood [3,10]. On the other hand, the temperature plateau shows that the initial temperature rise does not depend on metabolic heat production under depressed systemic and selective brain cooling during REM sleep [3]. In conclusion, the increase in vertebral artery blood flow appears primarily as an autoregulatory response to the drop in carotid artery blood flow during REM sleep. An additional increase in vertebral artery blood flow ought to be considered a response to brain activation in REM sleep with respect to NREM sleep. Of the two responses, the former would be the most variable, since the ‘steal’ of carotid blood is due to autonomic alterations which are intrinsically irregular (cf. Ref. [13]), whereas the latter would be the most stable as the expression of actual flow-metabolism coupling due to a stereotyped brain activation pattern. Eventually, the overall temporal coupling of flow and metabolism in the brain would be less consistent in REM sleep than in both QW and NREM sleep, according to the different time courses of randomly interacting peripheral and central physiological processes during REM sleep (cf. Ref. [13]).
Acknowledgements The work was supported by grants from the Ministero dell’Universita` e della Ricerca Scientifica e Tecnologica and from the Consiglio Nazionale delle Ricerche, Italy. The authors thank G. Mancinelli, V. Meoni and L. Sabattini for careful technical assistance.
References ¨ ´ Jr., Brain and core [1] P. Alfoldi, G. Rubicsek, G. Cserni, F. Obal temperatures and peripheral vasomotion during sleep and wakeful¨ ness at various ambient temperatures in the rat, Pflugers Arch. 417 (1990) 336–341. [2] P.L. Altman, D.S. Dittmer, in: Environmental Biology, F.A.S.E.B, Bethesda, MA, 1966, p. 694. [3] A. Azzaroni, P.L. Parmeggiani, Mechanisms underlying hypothalamic temperature changes during sleep in mammals, Brain. Res. 632 (1993) 136–142. [4] A. Azzaroni, P.L. Parmeggiani, Postural and sympathetic influences on brain cooling during the ultradian wake–sleep cycle, Brain. Res. 671 (1995) 78–82. [5] M.A. Baker, J.N. Hayward, The influence of the nasal mucosa and
60
[6]
[7] [8] [9]
[10]
[11]
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60 the carotid rete upon hypothalamic temperature in sheep, J. Physiol. (London) 198 (1968) 561–579. M. Caputa, W. Kadziela, J. Narebski, Significance of cranial circulation for the brain homeothermia in rabbits. II. The role of the cranial venous lakes in the defence against hyperthermia, Acta Neurobiol. Exp. 36 (1976) 625–638. L. Edvinsson, E.T. MacKenzie, J. McCulloch, in: Cerebral Blood Flow and Metabolism, Raven Press, New York, 1993, p. 683. C. Franzini, Brain metabolism and blood flow during sleep, J. Sleep Res. 1 (1992) 3–16. C. Franzini, T. Cianci, P. Lenzi, P.L. Guidalotti, Neural control of vasomotion in rabbit ear is impaired during desynchronized sleep, Am. J. Physiol. 243 (1982) R142–R146. J.N. Hayward, M.A. Baker, A comparative study of the role of the cerebral arterial blood in the regulation of brain temperature in five mammals, Brain Res. 16 (1969) 417–440. S.K. Maloney, G. Mitchell, Selective brain cooling: role of angularis oculi vein and nasal thermoreception, Am. J. Physiol. 273 (1997) R1108–R1116.
[12] P. Maquet, Functional neuroimaging of normal human sleep by positron emission tomography, J. Sleep Res. 9 (2000) 207–231. [13] P.L. Parmeggiani, The autonomic nervous system in sleep, in: M.H. Kryger, T. Roth, W.C. Dement (Eds.), Principles and Practice of Sleep Medicine, Saunders, Philadelphia, PA, 1994, pp. 194–203, 1067. [14] P.L. Parmeggiani, L.F. Agnati, G. Zamboni, T. Cianci, Hypothalamic temperature during the sleep cycle at different ambient temperatures, Electroencephalogr. Clin. Neurophysiol. 38 (1975) 589–596. [15] P.L. Parmeggiani, A. Azzaroni, M. Calasso, A pontine-hypothalamic temperature difference correlated with cutaneous and respiratory heat loss, Respir. Physiol. 114 (1998) 49–56. [16] P.L. Parmeggiani, G. Zamboni, T. Cianci, M. Calasso, Absence of thermoregulatory vasomotor responses during fast wave sleep in cats, Electroencephalogr. Clin. Neurophysiol. 42 (1977) 372–380. [17] P.L. Parmeggiani, G. Zamboni, E. Perez, P. Lenzi, Hypothalamic temperature during desynchronized sleep, Exp. Brain Res. 54 (1984) 315–320.
Research report
Systemic hemodynamic changes raising brain temperature in REM sleep Pier Luigi Parmeggiani*, Adele Azzaroni, Marcella Calasso Dipartimento di Fisiologia Umana e Generale, Universita` di Bologna, Piazza Porta San Donato 2, I-40127 Bologna, Italy Accepted 16 November 2001
Abstract The present research studied the mechanisms underlying the increase in brain temperature during REM sleep in the unrestrained rabbit carrying chronically implanted electrodes, thermistors and common carotid artery occluders. During the ultradian wake–sleep cycle at constant ambient temperature (2562 8C), we recorded: (i) the ear pinna temperature as an indirect indicator of blood flow affecting heat loss from the systemic heat exchangers of the head, (ii) the temperature of the pons and hypothalamus as indirect indicators of the temperature of vertebral artery blood (systemic cooling only) and carotid artery blood (both systemic and selective cooling), respectively, and (iii) the changes induced in these temperatures by short-lasting bilateral common carotid artery occlusion. The results show that during REM sleep both systemic and selective brain cooling are depressed by a spontaneous decrease in the common carotid artery blood flow and the associated autoregulatory increase in the vertebral artery share of the cerebral blood supply. 2002 Elsevier Science B.V. All rights reserved. Theme: Other systems of the CNS Topic: Brain metabolism and blood flow Keywords: Systemic brain cooling; Selective brain cooling; Cerebral blood flow; Hypothalamic temperature; Pontine temperature; Ultradian wake–sleep cycle
1. Introduction In the cat [10], the carotid artery blood supply to the brain is thermally conditioned once more prior to entering the circle of Willis (selective brain cooling) by countercurrent heat exchange between a network of fine vessels (the carotid rete) in contact with the cranial venous plexuses receiving cool venous blood from the systemic heat exchangers of the head. Vertebral artery blood is not thermally conditioned by selective cooling and enters the circle of Willis at the same temperature as aortic arch blood, which has been cooled only through the venous blood returning from the systemic heat exchangers of the whole body to the heart (systemic brain cooling). This explains the existence of a difference between vertebral artery blood temperature (systemic cooling only) and *Corresponding author. Tel.: 139-051-244-499; fax: 139-051-251731. E-mail address: [email protected] (P.L. Parmeggiani).
carotid artery blood temperature (both systemic and selective cooling). Practically, this difference may be indirectly appraised by computing the difference between pontine and hypothalamic temperatures, since they depend primarily on vertebral and carotid artery blood temperature, respectively [3,4,15]. In the cat at ambient temperatures close to thermal neutrality [2], NREM sleep (non-rapid eye movement (nrem) sleep showing low frequency and high amplitude electroencephalogram) is characterized by an increased blood flow in the systemic heat exchangers of the head as shown by the rise in ear pinna temperature [3,4]. This is due to the state-dependent decrease in tonic vasoconstrictor sympathetic outflow and the head-down posture (decrement of negative hydrostatic load and increment of transmural pressure in the vessels of heat exchangers) with respect to quiet wakefulness (QW). As a result of the increased systemic heat loss enhancing also selective heat loss (s. above), both pontine and hypothalamic temperatures are lower in NREM sleep than during QW and their
0006-8993 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 02 )02591-X
56
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
difference is increased. The evidence of a larger temperature change in the hypothalamus than in the pons shows the cumulative influence of systemic and selective arterial blood cooling at hypothalamic level. These effects are the obvious result of mechanisms which, however, do not underlie the pontine and hypothalamic temperature changes observed during REM sleep (rapid eye movement (rem) sleep showing high frequency and low amplitude electroencephalogram). Experimentally induced variations in heat loss from the systemic heat exchangers of the head exert only a weak effect on the spontaneously occurring increase in pontine and hypothalamic temperatures in this sleep state [3]. This is a result of a state-dependent decrease in the common carotid artery blood supply to the heat exchangers of the head and to the carotid rete, which also implies an autoregulatory increase in the supply of vertebral artery blood to the circle of Willis [3]. The more pronounced rise in hypothalamic temperature with respect to pontine temperature in the cat shows the cumulative effect of the depression of both systemic and selective heat loss on hypothalamic temperature during REM sleep [3]. In conclusion, the spontaneous hemodynamic change in the common carotid artery bed during REM sleep, impairing the efficiency of both systemic and selective brain cooling, would explain the increase in hypothalamic temperature characterizing this sleep state in several species, regardless of the exposure to a wide range of ambient temperatures [1,3,9,14,16,17] above and below ambient thermal neutrality [2]. Concerning the contribution of the systemic heat exchangers of the head (the nasal mucosa and ear pinna) to selective brain cooling, studies in sheep have shown that selective cooling depends mainly on heat loss from the nasal mucosa in this species [5,11]. This may be the case for other species with a carotid rete, like the cat. The heat loss from the large surface of the nasal mucosa must surely cool the venous blood flowing in contact with the carotid rete (selective cooling) more than the heat loss from the small surface of the ear pinna cools the venous blood flowing directly to the heart (systemic cooling). This explains why hemodynamic changes in the common carotid artery bed influence the hypothalamic more than the pontine temperature in the cat. On the basis of the previous results showing the dramatic systemic hemodynamic changes raising brain temperature during REM in the cat, the present research investigated comparatively the changes in systemic and selective brain cooling during REM sleep in the rabbit. In this species, selective brain cooling is provided by conductive heat exchange between the basal portion of the brain (including the circle of Willis) and the cranial venous lakes which drain cool venous blood from the systemic heat exchangers of the head [6]. The anatomical differences between the cat, with a carotid rete and a small ear pinna, and the rabbit, lacking a carotid rete and with a large ear pinna, may underlie quantitatively different influences of
hemodynamic changes in the common carotid artery bed on systemic and selective brain cooling. The results of this study indicate that the spontaneously occurring decrease in common carotid artery blood flow during REM sleep affects systemic brain cooling more than selective brain cooling in the rabbit. Moreover, such hemodynamic change primarily underlies an autoregulatory increase in vertebral artery blood flow warming the brain, since this blood is not thermally conditioned by selective cooling.
2. Material and methods Three adult rabbits (New Zealand) were used. Chronic implantation of (i) extradural screw EEG and wire EMG electrodes, (ii) needle thermistors (Yellow Springs) in the anterior hypothalamic region and in the tegmental field of the upper pons to record the respective temperatures (T hy and T p ), and (iii) air-inflatable occluders placed around the common carotid arteries, was carried out under general (clonazepam, 0.5 mg / kg i.m.; sodium pentobarbital, 40 mg / kg i.p.) anesthesia. Ear pinna temperature (T ep ) was recorded bilaterally by means of surface thermistors (Yellow Springs) since it indicates blood flow changes in the systemic heat exchangers of the head. The rationale for the experiment of short bilateral common carotid artery occlusion, a procedure already applied in the cat [3], is that it influences the mechanisms of both systemic and selective brain cooling and that these effects can be examined separately by considering the changes in pontine and hypothalamic temperature, respectively. The experiments have been carried out according to the European Communities Council Directive (86 / 609 / EEC), under the supervision of the Central Veterinary Service of the University of Bologna and the Ministry of Health, Rome.
2.1. Experimental procedures The experimental sessions lasted 5–6 h at 2562 8C, an ambient temperature close to thermal neutrality [2], in a sound-attenuated, thermoregulated chamber. The animal was unrestrained in a plexiglas box. Preliminary sessions allowed the animal to adapt to the experimental condition to display regularly the natural sequence of ultradian wake–sleep cycles. A polygraph (Grass 7D) recorded EEG and EMG activity, and the temperatures under study. Short-lasting occlusions (duration, 40–60s) of both common carotid arteries were carried out during the ultradian wake–sleep cycle by means of the air-inflatable occluders chronically placed around the vessels.
2.2. Histological control At the end of the experimental sessions, the animal was
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
sacrificed by means of a large dose of sodium pentobarbital (i.p.) and histological sections of the brain (fixed in formalin, embedded in celloidin, stained with the Nissl method) were prepared for determination of thermistor locations in the hypothalamus and pons.
2.3. Data analysis T p and T hy were measured in NREM sleep just before the beginning of REM sleep and at the end of their steep rising phase during REM sleep. The coincident values of T ep were also measured. The daily average values of the three variables were calculated for statistical analysis (5 days for each animal). The slopes of the T p and T hy rise in REM sleep and the differences between coincident T p and T hy values (T p 2T hy ), measured in NREM and REM sleep, were statistically compared by means of the two-tailed paired t-test.
57
3. Results An example of polygraphic recordings of spontaneous changes in T ep , T p and T hy during the ultradian sleep cycle is given in Fig. 1. During NREM sleep, such changes were small at constant ambient temperature. At REM sleep onset, a pronounced and constant decrease in T ep occurred which was related to a steep increase in both T p and T hy . The slope of the T hy change was steeper than that of the T p change (Table 1). Nevertheless, T p was always higher than T hy but the difference (T p 2T hy ) decreased during REM sleep (Table 1). At the end of the REM sleep episode the increase in T ep was associated with a decrease in T p and T hy . The effects of short-lasting (40–60 s) occlusion of both common carotid arteries on T ep , T p and T hy during NREM sleep are shown in Fig. 1. The bilateral carotid artery occlusion decreased T ep and increased both T p and T hy also in QW. These experimentally induced changes were simi-
Fig. 1. Changes in T ep , T p and T hy during the ultradian sleep cycle (ambient temperature, 23.6 8C). During NREM sleep, T ep increases slightly, whereas both T p and T hy show slight opposite changes. During REM sleep, T ep markedly decreases, whereas T p and T hy markedly increase. Note that the rise in T hy is steeper than that of T p . The effects of bilateral common carotid artery occlusion on T ep , T p and T hy are shown during NREM sleep. Note that T ep decreases and both T p and T hy increase as during REM sleep. However, the induced temperature changes are smaller with respect to those spontaneously occurring during REM sleep. EEG, electroencephalogram; EMG, electromyogram; T ep , left (l) and right (r) ear pinna temperature; T p , pontine temperature; T hy , hypothalamic temperature; BCO, bilateral common carotid artery occlusion; REMS, REM sleep.
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
58
Table 1 Slopes (DT p , DT hy ) of pontine and hypothalamic temperature increases from NREM to REM sleep (nrem→rem) and differences (T p 2T hy ) between the coincident values of pontine and hypothalamic temperatures just before the onset of (nrem) and during REM sleep (rem) DT p nrem→ rem DT hy nrem→ rem
0.0860.006 8C / min 0.1260.006 8C / min
P,0.001
df 14
(T p -T hy ) nrem (T p -T hy ) rem
0.1460.004 8C 0.0660.005 8C
P,0.001
df 14
lar but always smaller (independently of baseline) than those occurring spontaneously during REM sleep. During REM sleep onset, the bilateral common carotid artery occlusion did not significantly affect the spontaneous steep decrease in T ep and increase in T p and T hy (Fig. 2). In contrast, the bilateral common carotid artery occlusion just after the end of the REM sleep episode stopped both the T ep rise and the T p and T hy decrease (Fig. 2).
4. Discussion The results show that the changes in T ep , T p and T hy during REM sleep in the rabbit are qualitatively comparable with those previously observed in the cat [3]. However, remarkable species-specific features exist from the quantitative viewpoint. The fact that the T p and T hy rises in REM sleep are closer in amplitude in the rabbit than in the cat shows the greater relevance of systemic brain cooling with respect to selective brain cooling in the former species. In spite of a marked decrease in T ep occurring in both species during REM sleep at ambient temperatures slightly below thermal neutrality [2], systemic brain cooling is depressed more in the rabbit than in the cat, as shown by the larger T p increase in the former. It is likely that such a difference depends on the extension of the ear pinna surface, a heat exchanger prevalently affecting systemic brain cooling. On the other hand, selective brain cooling is probably less
Fig. 2. Changes in T ep , T p and T hy during the ultradian sleep cycle (ambient temperature: 25.8 8C). The effects of bilateral common carotid artery occlusion on T ep , T p and T hy are shown at the onset and just after the end of REM sleep. Note that the occlusion is not significantly affecting the spontaneous steep decrease in T ep and increase in T p and T hy . In contrast, the occlusion after the end of the REM sleep episode stops both the T ep rise and the T p and T hy decrease. EEG, electroencephalogram; EMG, electromyogram; T ep , left (l) and right (r) ear pinna temperature; T p , pontine temperature; T hy , hypothalamic temperature; BCO, bilateral common carotid artery occlusion; REMS, REM sleep.
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60
efficient in the rabbit than in the cat since the thermal inertia of conductive heat exchange is greater than that of countercurrent heat exchange in the latter species. Nevertheless, the cumulative effect of the depression of both systemic and selective brain cooling is also revealed in the rabbit at REM sleep onset by the steeper slope of the T hy increase with respect to that of the T p increase. Concerning the systemic hemodynamic mechanisms underlying T ep , T p and T hy changes during REM sleep, the results show that they do not differ in the rabbit with respect to the cat [3]. In particular, the marked spontaneous decrease in T ep and increase in T p and T hy , characterizing REM sleep onset also in the rabbit, is practically unaffected by bilateral common carotid artery occlusion as in the cat. A reasonable inference is that during REM sleep common carotid artery blood flow is spontaneously decreased [3]. Consistent with this view is the fact that the return to normal hemodynamic conditions of cerebral blood perfusion, after the end of the REM sleep episode, is clearly affected in the rabbit by bilateral common carotid artery occlusion which stops the T ep rise and both the T p and T hy decrease as in the cat (Ref. [3], Fig. 1B). Moreover, the decrease in carotid artery blood supply to the brain ought to be buffered by an increased amount of vertebral artery blood flowing into the circle of Willis. In other words, REM sleep is characterized by a shift from the carotid artery to the vertebral artery (and probably also to other arterial sources, cf. Ref. [7]) in the amount of blood contributed to the overall cerebral blood flow. The inference of a spontaneous decrease in carotid artery blood flow during REM sleep in both cats [3] and rabbits is consistent with the alteration in cardiovascular regulation characterizing this behavioral state (cf. Ref. [13]). For instance, a sudden expansion of extracerebral common carotid artery beds, as a result of an altered vasoconstrictor control and a posturally (antigravitary muscle atonia) induced rise in transmural pressure, may underlie a cerebral blood ‘steal’ decreasing the amount of carotid artery blood flowing into the circle of Willis [3]. In this respect, it is worth considering that T ep , T p and T hy changes elicited by short bilateral common carotid artery occlusion during QW and NREM sleep are qualitatively similar but much less pronounced than the spontaneous changes observed in REM sleep. The likely explanation is that systemic baroreceptive reflexes provide a more effective buffering of the blood flow drop in the common carotid artery bed during QW and NREM sleep in comparison to REM sleep (cf. Ref. [3], p. 137) which is characterized by a marked alteration in autonomic homeostatic regulation (cf. Ref. [13]). The indirect evidence of a decrease in the carotid artery blood supply to the circle of Willis is apparently inconsistent with the overall increase in cerebral blood flow observed during REM sleep with respect to NREM sleep in several species (cf. Refs. [8,12]), including the cat and the rabbit. According to the present data, however, such an
59
increase may be considered, at least partially, an autoregulatory response necessarily involving the vertebral artery (and probably other arterial blood sources, cf. Ref. [7]). This view is supported by the fact that blood flow during REM sleep increases more in the cerebellum and brain stem than in the cerebrum (cf. Ref. [8]). Also the early steep rise in hypothalamic and pontine temperatures, followed by a temperature plateau, is an indirect proof of the initial warming effect of an increased vertebral artery blood flow, since this blood is not thermally conditioned by selective cooling like the carotid artery blood [3,10]. On the other hand, the temperature plateau shows that the initial temperature rise does not depend on metabolic heat production under depressed systemic and selective brain cooling during REM sleep [3]. In conclusion, the increase in vertebral artery blood flow appears primarily as an autoregulatory response to the drop in carotid artery blood flow during REM sleep. An additional increase in vertebral artery blood flow ought to be considered a response to brain activation in REM sleep with respect to NREM sleep. Of the two responses, the former would be the most variable, since the ‘steal’ of carotid blood is due to autonomic alterations which are intrinsically irregular (cf. Ref. [13]), whereas the latter would be the most stable as the expression of actual flow-metabolism coupling due to a stereotyped brain activation pattern. Eventually, the overall temporal coupling of flow and metabolism in the brain would be less consistent in REM sleep than in both QW and NREM sleep, according to the different time courses of randomly interacting peripheral and central physiological processes during REM sleep (cf. Ref. [13]).
Acknowledgements The work was supported by grants from the Ministero dell’Universita` e della Ricerca Scientifica e Tecnologica and from the Consiglio Nazionale delle Ricerche, Italy. The authors thank G. Mancinelli, V. Meoni and L. Sabattini for careful technical assistance.
References ¨ ´ Jr., Brain and core [1] P. Alfoldi, G. Rubicsek, G. Cserni, F. Obal temperatures and peripheral vasomotion during sleep and wakeful¨ ness at various ambient temperatures in the rat, Pflugers Arch. 417 (1990) 336–341. [2] P.L. Altman, D.S. Dittmer, in: Environmental Biology, F.A.S.E.B, Bethesda, MA, 1966, p. 694. [3] A. Azzaroni, P.L. Parmeggiani, Mechanisms underlying hypothalamic temperature changes during sleep in mammals, Brain. Res. 632 (1993) 136–142. [4] A. Azzaroni, P.L. Parmeggiani, Postural and sympathetic influences on brain cooling during the ultradian wake–sleep cycle, Brain. Res. 671 (1995) 78–82. [5] M.A. Baker, J.N. Hayward, The influence of the nasal mucosa and
60
[6]
[7] [8] [9]
[10]
[11]
P.L. Parmeggiani et al. / Brain Research 940 (2002) 55 – 60 the carotid rete upon hypothalamic temperature in sheep, J. Physiol. (London) 198 (1968) 561–579. M. Caputa, W. Kadziela, J. Narebski, Significance of cranial circulation for the brain homeothermia in rabbits. II. The role of the cranial venous lakes in the defence against hyperthermia, Acta Neurobiol. Exp. 36 (1976) 625–638. L. Edvinsson, E.T. MacKenzie, J. McCulloch, in: Cerebral Blood Flow and Metabolism, Raven Press, New York, 1993, p. 683. C. Franzini, Brain metabolism and blood flow during sleep, J. Sleep Res. 1 (1992) 3–16. C. Franzini, T. Cianci, P. Lenzi, P.L. Guidalotti, Neural control of vasomotion in rabbit ear is impaired during desynchronized sleep, Am. J. Physiol. 243 (1982) R142–R146. J.N. Hayward, M.A. Baker, A comparative study of the role of the cerebral arterial blood in the regulation of brain temperature in five mammals, Brain Res. 16 (1969) 417–440. S.K. Maloney, G. Mitchell, Selective brain cooling: role of angularis oculi vein and nasal thermoreception, Am. J. Physiol. 273 (1997) R1108–R1116.
[12] P. Maquet, Functional neuroimaging of normal human sleep by positron emission tomography, J. Sleep Res. 9 (2000) 207–231. [13] P.L. Parmeggiani, The autonomic nervous system in sleep, in: M.H. Kryger, T. Roth, W.C. Dement (Eds.), Principles and Practice of Sleep Medicine, Saunders, Philadelphia, PA, 1994, pp. 194–203, 1067. [14] P.L. Parmeggiani, L.F. Agnati, G. Zamboni, T. Cianci, Hypothalamic temperature during the sleep cycle at different ambient temperatures, Electroencephalogr. Clin. Neurophysiol. 38 (1975) 589–596. [15] P.L. Parmeggiani, A. Azzaroni, M. Calasso, A pontine-hypothalamic temperature difference correlated with cutaneous and respiratory heat loss, Respir. Physiol. 114 (1998) 49–56. [16] P.L. Parmeggiani, G. Zamboni, T. Cianci, M. Calasso, Absence of thermoregulatory vasomotor responses during fast wave sleep in cats, Electroencephalogr. Clin. Neurophysiol. 42 (1977) 372–380. [17] P.L. Parmeggiani, G. Zamboni, E. Perez, P. Lenzi, Hypothalamic temperature during desynchronized sleep, Exp. Brain Res. 54 (1984) 315–320.