Sleep Deprivation in the Rat: XI. The Effect of Guanethidine-Induced Sympathetic Blockade
on the Sleep Deprivation Syndrome
June J. Pilcher, Bernard M. Bergmann, *Victor S. Fang, *tSamuel Refetoff, and Allan Rechtschaffen
Sleep Research Laboratory, Departments of Psychiatry and Psychology, and Departments of *Medicine and tPediatrics, University of Chicago, Chicago, Illinois, U.S.A.
Summary: In earlier studies, rats totally deprived of sleep by a disk-over-water apparatus (TSD rats) had shown an increase in energy expenditure (EE) that could not be explained by increased motor activity or the metabolic expense of wakefulness. Excessive activation of a calorigenic mediator was a possibility, and norepinephrine-mediated sympathetic activation was the most likely candidate, because plasma norepinephrine (NE) levels had risen sharply in TSD rats. To determine whether this activation was necessary for increased EE in sleep deprived rats, the peripheral sympathetic blocking agent guanethidine (OU) was administered to six sleep-deprived (OD) rats and their yoked control (GC) rats. GU attenuated the increase in NE previously seen in TSD rats, but the increase in EE was not attenuated. Apparently, NE-mediated sympathetic activation was not critical for increased EE in sleep-deprived rats. On the other hand, plasma epinephrine (EPI) levels were significantly increased in GD (but not in GC) rats above those previously seen in TSD rats, suggesting the substitution of one calorigenic mediator for another in response to an abnormally elevated need for EE. Temperature data suggest that increased need for EE could arise from an elevated temperature setpoint and an inability to retain body heat. GD (but not GC) rats also showed other effects previously seen in TSD rats, including debilitated appearance; severe ulcerative and hyperkeratotic lesions on the tails and plantar surfaces; initially increased and later decreased body temperature; decreased plasma thyroxine; increased triiodothyronine-thyroxine ratio; and eventual death. Evidently, NE-mediated sympathetic activation was not critical to any of these effects, although a role for catecholamines cannot be ruled out. Key Words: Sleep deprivation-Energy expenditure-Sympathetic activation-Ouanethidine-Catecholamines.
A recent study by Everson and co-workers (1) showed that chronic total sleep deprivation in the rat produced progressive increases in energy expenditure (EE) to about twice baseline values. The increase in EE was more than could be accounted for by the metabolic expense of wakefulness or gross motor activity, thereby indicating an in-
Accepted for publication November 1989. Address correspondence and reprint requests to Dr. Allan Rechtschaffen, Sleep Research Laboratory,
5743 S. Drexel Avenue, Chicago, Illinois, 60637, U.S.A.
218
SLEEP DEPRIVATION AND SYMPATHETIC BLOCKADE 219
crease in resting metabolic rate (2). Of interest relative to the function of sleep is the issue of whether the increased EE resulted from impaired control of calorigenesis or from an appropriate calorigenic response to an abnormally high need for EE, such as might result from abnormal thermoregulatory mechanisms (3).
The present study evaluates whether abnormally high norepinephrine (NE)-mediated sympathetic activation is necessary for the increased EE in sleep-deprived rats by determining whether the increase still occurs when peripheral postganglionic sympathetic activity is blocked by guanethidine mono sulfate (GU). NE-mediated sympathetic activation was hypothesized as a putative mediator of the high EE in the totally sleep deprived (TSD) rats in the earlier studies on the basis of their hormonal profiles (2). Levels of plasma NE, which is released from terminals of stimulated sympathetic neurons, increased progressively to about 300% over baseline levels in TSD rats, compared to increases ofless than 100% in yoked control (TSC) rats. By contrast, percentage increases were smaller in plasma levels of epinephrine (EPI), adrenocorticotropic hormone (ACTH), and corticosteroids, and did not parallel the progressive increase in EE as well as NE did. Plasma levels of thyroid hormones decreased (2).
Although high EE was the major focus of this study, the approach also permitted an evaluation of the possible role of NE-mediated sympathetic activation on other effects of chronic sleep deprivation, including debilitated appearance, skin lesions on the tail and plantar surfaces, increased heart rate, initially increased and later decreased body temperature, decreased plasma thyroxine, increased tri-iodothyronine-thyroxine ratio, and eventual death.
GU was selected to block sympathetic activation because it is highly selective in its action, with no discernible effect on NE in the brain or the adrenal medulla of the adult rat. It inhibits both uptake and release of NE from the noradrenergic terminals of the sympathetic system, where it is selectively accumulated (4,5). Chronic administration of GU, orally or intraperitoneally, has resulted in dose-dependent, progressive decreases in NE and morphological changes in sympathetic ganglia with no notable side effects (6-11).
Sleep deprivation was maintained by the same disk-over-water method used previously in this series of studies (12). To control for the effects of GU per se, GU-treated, sleep-deprived (GD) rats were individually yoked to GU-treated control (GC) rats. As a supplementary analysis to evaluate how certain sleep deprivation effects may have been modified by sympathetic blockade, the GD rats as a group were compared with the TSD rats of the earlier study (1).
METHOD
Except for GU administration and schedules, all aspects of methodology have been described in detail previously (12) .
. " A GD and a yoked GC rat were housed in separate plastic cages that had as a common floor a 46-cm-diameter horizontal disk suspended over water 2 to 3 cm deep. Whenever the GD rat started to sleep, as identified by electrophysiological recordings, the disk automatically began to rotate, forcing both rats to walk opposite to the direction of the disk rotation to avoid falling into the water. Thus, both rats received the same mild physical stimulation, but the GD rat was forced to remain mostly awake, whereas the GC rat could sleep whenever the GD rat was spontaneously awake.
Sleep, Vol. 13, No, 3, 1990
220 1. 1. PILCHER ET AL.
Subjects
Six pairs of GD and GC rats were studied. Rats were Sprague-Dawley males, 92 to 138 days old (x = 116.5 ± 16.8 SD), and they weighed 413 to 645 g (X = 500.9 ± 67.9) at the time of surgery. From their arrival at the laboratory, they were maintained in constant light to flatten their circadian rhythms (13) to avoid confounding the effects of sleep loss with shifts in phase, period, or amplitude of circadian rhythms.
Surgical procedures
Under deep anesthesia maintained by ketamine hydrochloride and xylazine, rats were implanted with the following electrodes: two miniature screws placed through the skull in a lateral position to record the electroencephalogram (EEG); two screws in a more medial position to maximize e activity in cortical recordings; three silver plates under the temporalis muscle to record electromyogram (EMG) activity; two stainless steel wires sutured bilaterally to the cutaneus maximus muscles to record heart rate. Electrode wires were soldered to a miniature plug fastened to the skull. A transmitter was implanted in the peritoneal cavity to record body temperature (Tb). To permit daily blood sampling, a catheter was implanted through the external jugular vein into the right atrium of the heart.
Experimental schedule and procedures
After surgery, rats resided in the experimental apparatus for at least 8 days of adaptation, during which they rested on a solid floor that was installed over the apparatus disk and covered the entire bottom of the cage. This was followed by a 5-day normal baseline period during which the "adaptation" floor was removed, and the rats thereafter resided on the apparatus disk (over water). Then, during an extended GU baseline of 14 days and the deprivation period which followed, 50 mg of GU mixed with approximately 1 g of chocolate was administered orally each day to both the GD and the GC rats. Similar repeated oral doses have previously been shown to produce extensive ganglial pathology and progressive NE decline in rats (11). For both the untreated and GU baselines, the disk was rotated once an hour for 6 s to wipe the disk surface clean and to familiarize the rats with disk rotation. Following baseline, the deprivation-yoked control procedure was initiated and maintained until the GD rat died or was killed when death seemed imminent. During both baseline and experimental periods, food and water were available ad libitum, cage temperatures were regulated thermostatically between 28°C and 29°C, and electrophysiological parameters were recorded continuously.
Recording procedures and variables measured
Electrophysiological signals were passed from the rat sequentially to a commutator, a polygraph for continuous ink recording, and a Synertek "SYN-l" microcomputer. The microcomputer was programmed to initiate disk rotation in a randomly chosen direction at a rate of 3.33 rotations/min whenever EEG, EMG, and e values indicated sleep onset in the GD rat. Procedures for establishing criteria for sleep onset were described previously (12). Disk rotation was stopped when the rat had been awake for about 6 s.
Each 30 s, data from the preceding epoch, including heart rate and the integrated rectified EEG, EMG, and e values, were passed from the SYN-l to a PDP 11-23
Sleep, Vol. 13, No.3, 1990
. '
SLEEP DEPRIVATION AND SYMPATHETIC BLOCKADE 221
computer (Digital Equipment Corporation) for storage. Every 24 h, the data were computer scored for sleep stages by the Parametric Animal State Scoring system (14). Stages scored were as follows: Waking (W)-low-amplitude EEG and e, high-amplitude EMG: low-amplitude non-rapid-eye-movement (NREM) sleep (LS)-low-amplitudeEEG, e, and EMG; high-amplitude NREM sleep (HS)-high-amplitude EEG and e, lowamplitude EMG; paradoxical sleep (PS)-low-amplitude EEG and EMG, moderately high amplitude e. Because LS values were similar during all baselines in all rat groups (about 4% of total time) and were similarly reduced in GD and TSD rats during deprivation, LS and HS were combined as NREM sleep for data presentation. When daily comparisons between computer scores and the polygraph record revealed a poor match, the data were rescored either by the computer with revised criteria or manually. Approximately two-thirds of all epochs were compared and 5% of scores altered manually.
Body weight, food intake, and water intake were measured daily. Water intake was underestimated by an unknown amount because rats occasionally drank the pan water even though it was laced with 130 mg of quinine sulphate to discourage such drinking . Daily EE was estimated by the formula EE = (caloric value offood taken from hopper x fraction of food energy metabolized) + caloric value of weight loss. Conversion factors were provided previously (12).
Plasma levels of NE, EPI, thyroxine (T4), and tri-iodothyronine (T3) were measured as described previously (12) at intervals of 3 to 5 days from blood withdrawn daily. Tb was measured at intervals of 5 to 13 h at least three times daily. Slide photographs, taken every 2 to 3 days, were used to rate the appearance of rats on a six-point scale (1 = completely healthy, 6 = extremely debilitated). Photographs of the GU-treated rats were intermixed with photographs from the original TSD study, randomized with respect to run, experimental condition, and time in experiment and given to two independent judges for blind ratings. The mean of the two ratings was used as the appearance score. Inter-rater reliability was 0.865 for all deprived rats and 0.748 for the GU-treated rats.
Survival time was defined as the number of days from the start of sleep deprivation to either death or the killing of the rat when death seemed imminent. Impending death was judged by the following morbid signs seen in rats who died spontaneously in earlier studies (1,15): (a) severely debilitated appearance; (b) marked decline from maximal food intake; (c) severe ataxia or weakness, sometimes indicated by difficulty in negotiating the moving disk; (d) precipitous decline in body temperature; (e) marked edema of the paws; (0 loss of EEG amplitude. When a GD rat was killed, its yoked GC was also killed within the half-hour, and the internal organs of both rats were examined.
Data presentation Similar pathologies developed in all GD rats, but at different rates. Therefore, to
emphasize the similarities in the course of this development, each rat's survival time was divided into quarters, and data were averaged for each quarterly bin, first within rats and then across rats. To assess which deprivation pathologies developed, the GD and GC rats were compared by an analysis of covariance with time in quarters as the covariant. Only the group by time interactions will be reported, as this evaluates the critical issue of whether there were progressive changes in GD rats compared to GC rats. For some specific comparisons, t tests were used; when comparisons were between yoked experimental and control rats or between pretreatment and posttreatment
Sleep. Vol. 13, No.3, 1990
222 1.1. PILCHER ET AL.
values from the same rats, the t tests were for paired comparisons. "NS" indicates not significant at the p < 0.05 level by two-tailed test; ± indicates standard deviation.
To evaluate the effectiveness of sympathetic blockade on both measured catecholamines, statistical comparisons were made between GD and TSD rats on NE and EPI. There were two pre-experimental differences between GD and TSD rats that conceivably could have produced differences in NE and EPI, but this seems unlikely for the following reasons: (a) The GD rats had a total of 19 days of combined untreated and G U baseline compared to 14 days of untreated baseline for most TSD rats. It does not seem likely that a 5-day difference in baseline would substantially affect response to subsequent deprivation. (b) GD rats were significantly younger at surgery than the TSD rats (116.5 ± 17.9 days versus 169.6 ± 35.7 days; t = 3.64, p < 0.01). However, TSD rats ranged in age from 113 to 219 days (which overlapped the 92- to 138-day range for GD rats), and they all developed similar deprivation effects. Also, correlation coefficients between age and NE or EPI in either the GD or TSD groups were all small.
For NE and EPI, an analysis of variance on the GU baseline and four quarters of deprivation in the GD rats and the normal baseline and four quarters of deprivation in TSD rats gave a pooled estimate of the variance which was used in Bonferroni corrected post hoc tests to compare corresponding time bins in the two groups for significant differences. Similar tests across five time bins were compared on the previous TSD and TSC data and across six time bins on the GD and GC data for post hoc deprived versus control comparisons.
Some values reported for TSD and TSC rats are slightly different from those published because, to enhance comparability across studies, only the data from the eight (of 10) TSD-TSC pairs that were cannulated for hormone studies were used. Graphs of TSD and TSC data by quarters also differ slightly in shape from the original presentation. The data were originally presented as mean percentage changes from individual baseline, whereas here, to facilitate cross-study comparison, absolute values are given. Therefore, each mean of percentage change is slightly different from what the corresponding percentage change in mean absolute values would be.
RESULTS
Norepinephrine
GU successfully blunted NE during GU baseline and during the experimental period in both the GD and GC rats (Fig. 1). NE decreased from normal baseline to GU baseline in GD rats (t = 7.58; p < 0.01) and in GC rats (t = 3.13; p < 0.05). Also, GU baseline levels of NE were lower for GD rats than baseline levels of NE for TSD rats (p < 0.05; Bonferroni corrected post hoc test). As in the previously studied TSD rats, the GD rats showed significant increases over their respective control rats in NE (group x time F 1.79 = 3.66; P < 0.01), with significantly higher levels in quarters 3 and 4 of deprivation (p < 0.01; Bonferroni corrected post hoc test). However, the NE values of the GD rats never approached the levels seen in normal TSD rats. In every quarter of sleep deprivation, mean NE levels in TSD rats were more than double those in GD rats. The differences in NE level between GD and TSD rats increased over time (group x time F'.92 = 10.1; p < 0.001). Quarters 1 and 2 of deprivation were not significantly different between the GD and TSD animals, but quarters 3 and 4 were (p < 0.01; Bonferroni corrected post hoc test). The relatively small increase in NE in the GD rats could not
Sleep. Vol. 13, No.3, 1990
E ---OJ B-ID c ".: £: 0. ID c "u ID >-0 Z
SLEEP DEPRIVATION AND SYMPATHETIC BLOCKADE 223
1250
1000
750
500
250
0 Base
line GU
Base-line
2 3 4
Quarters of Oepriva tion
FIG. 1. Mean plasma norepinephrine (NE) levels in TSC, GC, TSD, and GD rats. Vertical lines above bars indicate standard errors. GU, guanethidine; TSD, totally sleep deprived (D); TSC, yoked controls of TSD rats (Ii!il); rats; GD, guanethidine-treated sleep-deprived rats (_); GC, yoked controls of GD rats (12).
account for the sleep deprivation syndrome seen in the GD rats, as their absolute mean levels of NE never rose above those of TSC rats, which did not develop the syndrome.
Sleep data GD and GC rats showed small increases in NREM sleep and PS during GU baseline
compared to normal baseline, but these increases could have resulted from extended adaptation rather than from GU. GU-treated rats had only 5 days of normal baseline compared to 14 days for untreated rats. It is likely that GU-treated rats acclimated to the apparatus to the same extent as untreated rats only after additional time during GU baseline. Thus, their pre-experimental sleep quotas were more comparable to those of untreated rats during GU baseline than during normal baseline (Table n.
During deprivation, GD rats showed percentage reductions of 86.2,95.0, and 87.2%
TABLE 1. Mean percentages and standard deviations of percent time spent in sleep stages
from GU baseline in NREM, PS, and total sleep respectively, compared to declines of 91.4, 94.8, and 91.8% from untreated baseline in TSD rats. The GD rats did have significantly more NREM (t = 4.41, p < 0.01) and total sleep (t = 4.28, p < 0.01) than TSD rats during the deprivation period, but it is doubtful that the small absolute differences, 2.5 and 2.6% respectively, could have accounted for major differences in outcome variables. Total sleep percentage in TSD rats varied from 3.2 to 6.0% (which overlapped the 5.2-8.4% range in GD rats), and all TSD rats developed the same syndrome. As a result of sleep interruption by some disk rotations, GC rats suffered partial reductions in sleep quotas-31.4, 56.7, and 34.2% in NREM, PS, and total sleep respectively. Comparable reductions had been 24.5, 43.6, and 26.6% in TSC rats. Disk rotation averaged 22.9 ± 4.3% of time in GU rats compared to 20.4 ± 2.9% in the untreated rats.
Survival
Two GD rats died after 8 and 21 days of deprivation. Four GD rats were killed after 15, 17, 17, and 21 days of deprivation when death seemed imminent. Mean survival of GD rats was 16.5 ± 4.8 days, which was not significantly different from mean survival of 22.1 ± 5.2 days in TSD rats (t = 2.02, p < 0.075). No GC rat ever appeared near death.
Necropsy evaluation
The internal organs of the GC rats appeared completely normal. As in the earlier study (1), sleep-deprived rats showed a profound absence of body fat, but no observable anatomical abnormality that could account for their actual or imminent deaths. One kidney and both adrenals were weighed. Kidney weights were nearly identical in GD and GC rats. Average adrenal weights were slightly heavier in GD rats (0.043 ± 0.010 g) than in GC rats (0.035 ± 0.005 g), but the difference was not significant (t =
1.17, p = 0.31). Kidneys and adrenals had been significantly heavier in TSD than in TSC rats.
Appearance
As in the previous study 0), the physical appearance of the GD rats declined progressively during deprivation, whereas the appearance of GC rats remained near normal (Fig. 2). By the end of deprivation, all GD rats looked worse than all GC rats. The group x time interaction in photograph ratings (F1,49 = 39.81) was significant at p < 0.001. Although GD rats did not appear as debilitated as TSD rats had appeared during the last quarter of deprivation, the changes from their respective baselines were similar (Fig, 2). GD rats may have had better appearance scores during baseline because they were somewhat younger than TSD rats.
The fur of the GD rats became yellowish, looked disheveled, and stuck together in clumps as deprivation progressed. Ulcerative and hyperkeratotic lesions developed on the paws and tails of all animals, but much more extensively and severely in the GD than in the GC rats. The lesions of the GD and GC rats were similar to those of the TSD and TSC rats (16), respectively, except that in the GU-treated animals, the lesions had a slight build-up of pus around the edges (indicative of infection), which gave them a yellowish appearance. It is unlikely that the debilitated appearance resulted from a failure to groom, because grooming was shown not to decrease in the TSD rats (1).
Sleep, Vol. 13, No.3, 1990
SLEEP DEPRIVATION AND SYMPATHETIC BLOCKADE 225
Ol ~ 6 +-' cd
IT: 5
QJ U c 4 cd (0 QJ 3 Q Q
« 2 c cd QJ 2
0 Base- 2 3 4
line Qua rt e r S 0 fOe p r i vat ion FIG. 2. Mean appearance ratings in TSC (fZ!l), GC (r./), TSD (0), and GD (_) rats. Vertical lines above bars indicate standard errors. TSD and TSC baseline is compared to GD and GC GU-baseline. Abbreviations as in legend to Fig. I.
Body temperature The small decreases in T b during GU baseline were not statistically significant (Fig.
3). Over the entire experimental period, the group x time interaction (F 1,129 = 6.89) between GD and GC rats was significant at p < 0.01. The GD rats showed an initial rise in T b during quarter 1, followed by a progressive decline as deprivation progressed. Mean Tb during quarter 4 was approximately 1°C lower than during GU baseline. The GC rats maintained T b near baseline levels. During quarter 4, all GD rats had lower mean Tb than their respective yoked control rats. Inspection of Fig. 3 reveals a very similar pattern of Tb change in GD and TSD rats.
Heart rate As would be expected from sympathetic blockade, heart rate decreased significantly
from normal to GU baseline (for GD rats, t = 6.33, p < 0.01; for GC rats, t = 3.14,
38.50
0 'L. 38
Q) lo-::J +-' 37.50 co lo-Q) Q E 37 QJ f-
» \J 36.50 0 m
36 Base- GU 2 3 4
line Base- Quarters of line Oepriva tion
FIG. 3. Mean body temperatures (Tb) in TSC (~), GC (r.I), TSD (0), and GD (_) rats. Vertical lines above bars indicate standard errors. Abbreviations as in legend to Fig. 1.
Sleep, Vol. ]3, No.3, 1990
226 1. 1. PILCHER ET AL.
p < 0.05). Heart rate increased markedly during the course of deprivation in both GD and GC rats (Fig. 4); the group x time interaction (F 1,80 = 1.42) was not significant, unlike the significantly greater rise in TSD than in TSC rats reported previously (1).
Food intake, body weight, and energy expenditure
The GD and GC rats lost body weight as deprivation progressed despite increased food intake. Mean daily food intake of both the GD and GC rats increased progressively from 25 g during GU baseline to 38 g during quarter 4 of deprivation (group x time
. interaction; F 1.76 = 1.12, NS). However, the GD rats lost weight more rapidly than the GC rats (group x time interaction; F 1,76 = 24.29; P < 0.001). The weight of GD rats declined progressively from a mean of 463.4 ± 77.7 g during GU baseline to 379.6 ± 56.3 g during quarter 4 of deprivation, whereas the weight of GC rats declined from 473.6 ± 58.7 to 405.2 ± 17.5 g during the same period.
EE of the GD rats showed a much greater increase than EE of GC rats (group x time F1,75 = 11.34, P < 0.001). During quarter 4, all GD rats had higher mean EE than their respective yoked control rats. Fig. 5 shows that EE differences between GD and GC generally corresponded well to TSD-TSC differences. Clearly, GU did not blunt the EE response to sleep deprivation.
Water intake
From untreated baseline to GU baseline, mean daily water intake increased from 39.6 ± 12.7 g to 51.4 ± 13.2 g in GD rats (t = 3.70, p < 0.05) and from 37.5 ± 7.1 g to 45.4 ± 11.1 g in GC rats (NS). Water intake was generally erratic during the deprivation period, but it tended to decrease in both groups, decreasing more in the GD rats than in the GC rats (group x time F 1,76 = 19.15, P < 0.001). By the fourth quarter of deprivation, mean daily water intake had declined to 21.6 ± 16.8 g in GD rats and to 39.4 ± 24.8 g in GC rats. TSD rats had shown a small increase in water intake during deprivation, whereas TSC rats had shown little change.
+-' L
lU OJ I
450
420
390
360
330
300 L.LLJIU_
Baseline
GU Base
line
2 3 4
Quarters of Deprivation
FIG. 4. Mean waking heart rates in TSC (f2J), GC (~), TSO (D), and GO (_) rats. Vertical lines above bars indicate standard errors. Abbreviations as in legend to Fig. 1.
Sleep, Vol. 13, No.3, 1990
:;, (\j
LJ
---ro u ~ (]) "--::J +-' iJ c (]) a.. X W >. ()) "--(]) c W
SLEEP DEPRIVATION AND SYMPATHETIC BLOCKADE 227
250
200
150
100
50
0 Base
line GU
Baseline
2 3 4
Quarters of Oepriva tion
FIG. 5. Mean energy expenditure in TSC (~), GC (~), TSD (0), and GD (_) rats. Vertical lines above bars indicate standard errors. Abbreviations as in legend to Fig. l.
Thyroid hormones In GD rats, T4levels declined from a GU baseline mean of 4.0 ± 0.30 /J-g/dl to a fourth
quarter mean of 1.9 ± 0.32 /J-g/dl; for GC rats, mean T4 concentration was 3.1 ± 0.39 /J-g/dl during GU baseline and remained near that level throughout deprivation (group x time Pl,18 = 4.82, P < 0.05). During quarter 4, all GD rats had lower mean T4 levels than their respective yoked control rats. In GD rats, mean T3levels decreased from 45.9 ± 7.78 ng/dl l during GU baseline to 35.0 ± 7.21 ng/dl during quarter 3, and then remained at that level during quarter 4. In GC rats, T3levels increased from a baseline mean of 37.9 ± 10.11 to a fourth quarter mean of 46.3 ± 14.98. The group X time interaction for T 3 was not significant.
In GD rats, the T3/T4 ratio increased progressively from a GU baseline mean of 0.0116 ± 0.0020 to 0.0191 ± 0.0050 during quarter 4. In GC rats, the T31T4 ratio increased from a GU baseline mean of 0.0122 ± 0.0022 to a quarter 3 mean of 0.0169 ± 0.0036; it then decreased during quarter 4 to 0.0119 ± 0.0091, near the baseline average (group x time interaction F1,l8 = 5.69, P < 0.05).
Thus, the thyroid hormone results paralleled those of TSD-TSC rats (2) inasmuch as T 4 level decreased more in deprived rats; the level of T 3 tended to decrease more (but not significantly); and the T31T4 ratio increased more.
Epinephrine There was a surprisingly large increase in EPI levels (Fig. 6) during deprivation in GD
rats compared to a relatively modest rise in GC rats (group x time Fl 76 = 8.58; P < 0.001). By quarter 4 of deprivation, EPI levels in GD rats were about seven times greater than during normal baseline and about 14 times greater than during GU baseline. The increase in EPI was significantly greater in GD rats than it had been in TSD rats (Fl.92 = 3.75, P < 0.01). Bonferroni corrected post hoc tests indicated that the EPI levels of GD rats were significantly higher (p < 0.01) than those of the GC rats during
I Units for T3 were incorrectly given as pg/ml in reference 2.
Sleep, Vol. 13, No.3, 1990
228
E -----OJ 2-(l)
.~ "--
J::: Q (l) c Q W
500
400
300
200
100
0 Base
line
1. J. PILCHER ET AL.
2 3 4 GU Base
line Quarters of Oepriva ti on
FIG. 6. Mean plasma epinephrine (EPI) levels in TSC (123), GC (t'.Il), TSD (0), and GD (_) rats. Vertical lines above bars indicate standard errors. Abbreviations as in legend to Fig. 1.
quarters 3 and 4 of deprivation and those of TSD rats during quarter 4. By the fourth quarter of deprivation, EPI levels were 76.0% higher in the GD rats than in the TSD rats.
DISCUSSION
Sympathetic blockade
That GU produced a blockade of NE-mediated sympathetic activity in GD and GC rats was indicated by several results.
a. NE was reduced during G U baseline and did not increase as much during deprivation as it had in TSD rats. The degree of NE-mediated sympathetic blockade is probably underestimated by the fall in NE levels, because NE release from the adrenal medulla also contributes to circulating NE. Because the decline of sympathetic NE is progressive with chronic GU administration (II), it can reasonably be assumed that the increase in plasma NE levels during the deprivation period resulted primarily from the adrenal medulla, which is innervated by preganglionic sympathetic fibers and is thus presumably immune to attack (reviewed references 4 and 5). Similarly, other organs with sympathetic innervation by cholinergic fibers, such as sweat glands and pilomotor muscles, would also be unaffected.
b. Tb decreased from normal to GU baseline. Other investigators have shown that NE-depleting substances such as ex-methyl metatyrosine (17) and reserpine (18,19) cause sustained decreases in T b' Somerville and Whittle (19) have proposed that the effect of reserpine on T b is mediated primarily by peripheral organs. This hypothesis supports the probability that G U, which is selectively active in the sympathetic system, could cause a decrease in T b'
c. Heart rate was reduced during GU baseline and remained low relative to rates in TSD and TSC rats during the early portion of sleep deprivation. Eventually, heart rate reached levels comparable to those seen in untreated rats. This later increase in heart rate probably resulted primarily from deprivation-induced release of EPI and NE from
Sleep, Vol. 13, No.3, 1990
SLEEP DEPRIVATION AND SYMPATHETIC BLOCKADE 229
the adrenal medula, but compensatory adrenergic receptor hypersensitivity may have also a played role.
Replication of the sleep deprivation syndrome
The partial blockade of sympathetic activation did not prevent the development of the most salient features of the sleep deprivation syndrome. The following effects still appeared: eventual death, increased EE, debilitated appearance, lesions on the tail and plantar surfaces, initially increased and subsequently decreased Tb , decreased T4 levels, increased TiT4 ratio, and increased EPI. Although plasma NE level increased, the increase per se could not have caused the aforementioned syndrome in GD rats, because their NE levels never reached those seen in TSC rats, which did not develop the syndrome to the same degree.
The group x time interaction for heart rate was not significant, as it had been in the earlier study 0), but Fig. 4 reveals that overall heart rate profiles were not entirely dissimilar in the two studies. In both studies, deprived rats showed higher heart rates than control rats. In the last quarter, heart rates for both deprived groups fell slightly from third quarter levels, perhaps in response to declining Tb , because heart rate responds sensitively to a lowering of T b (20). Heart rate decreased between quarters 3 and 4 in TSC rats but continued to increase in GC rats-perhaps because T b decreased during the fourth quarter in the former group, but increased in the latter. Thus, deprived rats showed similar heart rate profiles in the two studies, but the control rats did not.
Differences in GD and TSD effects
Although EE increased in GD rats as it had in TSD rats, GD rats used their energy sources somewhat differently. TSD rats had lost more weight and eaten more food than TSC rats. GD rats lost more weight than GC rats, but their increases in food intake were similar. Apparently, there was a differentially greater use of stored energy in GD rats, perhaps because they weighed more than TSD rats at the start of deprivation, or because sympathetic activity normally plays a role in processing large amounts of food (21).
Although water intake would be expected to rise concurrently with food intake, it decreased during deprivation in GD rats. Water intake was above baseline levels in TSD and TSC rats. The reason for the decrease in GD rats is unexplained. It is unlikely that they drank enough pan water to account for the difference between water intake and the expected water intake. The decreased water consumption might have led to dehydration, which could have contributed to weight loss; however, this seems unlikely, as organs and tissues did not appear dehydrated at necropsy. Conceivably, the GD rats drank less than expected because of an increased production of metabolic water or because of increased GU-related water retention (22).
Catecholaminergic mediation of EE
This study was prompted by the issue of whether the increase in EE in sleep-deprived rats resulted from abnormal activation of a calorigenic mechanism or from an appropriate response of calorigenic mechanisms to an abnormally high need for EE. Our results rule out abnormal activation of what seemed to be the most likely mechanism; blockade of NE-mediated sympathetic activation did not prevent an increase in EE in
Sleep, Vol. ]3, No.3, 1990
230 1. 1. PILCHER ET AL.
sleep-deprived rats. Furthermore, the increase in EPIlevels in GD rats may be interpreted as a substitution of adrenergic for noradrenergic stimulation of calorigenic mechanisms. The apparent shift from noradrenergic mediation in TSD rats to adrenergic mediation in GD rats was substantial. Only one GD rat had higher fourth quarter NE levels than a TSD rat. One GD rat had lower fourth quarter EPI levels than two TSD rats. Otherwise, there was no overlap between the groups in fourth quarter catecholamine levels. There are experimental precedents for shifts in calorigenic mediation. In fasting, pregnant rats (23) and in fasting, cold-exposed rats (24), calorigenic mediation was achieved by EPI, because sympathetic outflow was decreased.
Since the sympathetic input to the adrenal medulla was presumably undamaged, the sympathetic system may still have played a role in the increase in catecholaminergic activity in the GU-treated rats. However, given the much greater increase of EPI in GD rats than in TSD rats, this role would necessarily have been secondary to some functional demand placed on the sympathetic system.
The maintenance of high EE combined with the apparent substitution of EPI for NE as a calorigenic mediator in GD rats suggests that the primary pathology underlying the EE rise was not abnormal calorigenic mediation, but a need for increased EE by whatever mechanisms are available. It has been postulated that two specific thermoregulatory deficits could be responsible for an increased need for EE in sleep-deprived rats (3). One postulated deficit is an increase in preferred T b or setpoint, which requires increased EE or lower heat loss to attain the new setpoint. This deficit would explain the rise in T b during the first half of deprivation. The second postulated deficit is excessive heat loss, which requires a rise in EE to maintain T b at near normal levels. When heat production could no longer match heat loss, Tb would decline, as it did late in deprivation. These two hypotheses suggest that sleep is necessary for effective thermoregulation; without it, the need for EE is increased.
Some confirmation for these ideas is found in a recent study from our laboratory (25). Given a choice of ambient temperature, TSD rats chose progressively higher temperatures as deprivation advanced, even when Tb was elevated, thus confirming an elevated setpoint. These data also support the hypothesis of excessive heat loss, as the rats were unable to achieve the elevated setpoint despite greatly increased EE.
Following a thermoregulatory model, we might expect an additional need for EE in both GD and GC rats, given that both might have suffered additional heat loss because EPI is a relatively poor mediator of peripheral vasoconstriction compared to NE. In fact, EE was higher in GD rats than in TSD rats and higher in GC rats than in TSC rats during every quarter of deprivation.
Could a switch from NE- to EPI-mediated calorigenesis be induced by the GU sympathectomy to counter thermoregulatory deficits caused by sleep deprivation? In the previously mentioned concomitant fasting and cold exposure of rats, sympathetic activity was decreased by fasting, whereas cold exposure increased the demand for EE, which was met by the secretion of EPI (24). Similarly, GU-treated rats survived cold exposure if they had an intact adrenal medulla (26). However, demedullated, GUtreated, cold-exposed rats failed to increase their NE or EPI and died fairly rapidly (27). Presumably, EPI was necessary for survival in the cold when sympathetic activity was blocked. These switches to EPI in response to a thermoregulatory challenge are congruent with the results of the present study, except, of course, that the thermoregulatory challenge and consequent need for EE in GD rats presumably resulted from in-
Sleep, Vol. 13, No.3, 1990
SLEEP DEPRIVATION AND SYMPATHETIC BLOCKADE 231
ternally induced heat loss caused by sleep deprivation rather than from externally induced heat loss caused by cold exposure.
Acknowledgment: This research was supported by NIH granst MH4151, MH18428, DK26678, and DK17050. We gratefully acknowledge the use of the TSD and TSC data provided by Carol A. Everson.
REFERENCES
1. Everson CA, Bergmann BM, Rechtschaffen A. Sleep deprivation in the rat: III. Total sleep deprivation. Sleep 1989;12:13-21.
2. Bergmann BM, Everson CA, Kushida CA, et al. Sleep deprivation in the rat: V. Energy use and mediation. Sleep 1989;12:31-41.
3. Rechtschaffen A, Bergmann BM, Everson CA, Kushida CA, Gilliland MA. Sleep deprivation in the rat: X. Integration and discussion of the findings. Sleep 1989;12:68--87.
4. Boura ALA, Green AF. Adrenergic neurone blocking agents. Annu Rev of PharmacoI1965;5:183-212. 5. Furst CI. The biochemistry of guanethidine. Adv Drug Res 1967;4:133-61. 6. Jensen-Holm J, Juul P. Changes ofthe rat superior cervical ganglion induced by guanethidine (histology
and cholinesterase histochemistry). Br J Pharmacol 1968;34:211P-12P. 7. Jensen-Holm J, Juul P. The effects of guanethidine, pre- and postganglionic nerve division on the rat
superior cervical ganglion: Cholinesterases and catecholamines (histochemistry), and histology. Acta Pharmacol Toxieol 1970;28:283-98.
8. Bumstock G, Evans B, Gannon BJ, Heath HW, James V. A new method of destroying adrenergic nerves in adult animals using guanethidine. Br J Pharmacal 1971 ;43:295-301.
9. Jensen-Holm J, Juul P. Ultrastructural changes in the rat superior cervical ganglion following prolonged guanethidine administration. Acta Pharmacol Toxicol 1971 ;30:308-20.
10. Coleoni AH. Effects of the administration of catecholamine-depleting drugs on the thyroid function of the rat. Pharmacology 1972;8:300-10.
II. Bittiger H, Maitre L, Krinke G, Sachnider K, Hess R. A study of long-term effects of guanethidine on peripheral noradrcnergic neurones of the rat. Toxicolugy 1977:8:63-78.
12. Bergmann BM, Kushida CA, Everson CA, Gilliland MA, Obermeyer W, Rechtschaffen A. Sleep deprivation in the rat: II. Methodology. Sleep 1989;12:5-12.
13. Eastman C, Rechtschaffen A. Circadian temperature and wake rhythms of rats exposed to prolonged continuous illumination. Physiol Behav 1983 ;31 :417-27.
14. Rechtschaffen A, Gilliland MA, Bergmann BM, Winter JB. Physiological correlates of prolonged sleep deprivation in rats. Science 1983;221:182-84.
15. Bergmann BM, Winter JB, Rosenberg RS, Rechtschaffen A. NREM sleep with low voltage EEG in the rat. Sleep 1978;10:1-11.
16. KushidaCA, Everson CA, Suthipinittharm P, et al. Sleep deprivation in the rat: VI. Skin changes. Sleep 1989;12:42-6.
17. Shellenberger MK, Elder JT. Alterations in rabbit core temperature produced by alpha-methyl metatyrosine. Eur J Pharmacol 1968;3: 196-202.
18. Shellenberger MK, Elder JT. Changes in rabbit core temperature accompanying alterations in brainstem monoamine concentrations. J Pharmacol Exp Ther 1967;158:219-26.
19. Somerville AR, Whittle BA. The interrelation of hypothermia and depletion of noradrenaline, dopamine and 5-hydroxytryptamine from brain by reserpine, p-chlorophenylalanine and alpha-methyl metatyrosine. Br J Pharmacol Chemother 1967;31:120-31.
20. Popovic V, Popovic P. Hypothermia in biology and medicine. New York: Grune and Stratton, 1974. 21. Landsberg L, Young JB. Diet-induced changes in sympathoadrenal activity: Implications for thermo
genesis. Life Sci 1981 ;28: 1801-19. 22. Weiner N. Drugs that inhibit adrenergic nerves and block adrenergic receptors. In: Gilman AG, Good
man LS, Rail TW, Murad F, eds. Goodman and Gilman's the pharmacological basis of therapeutics. New York: Macmillan, 1985:181-214.
23. Young JB, Landsberg L. Sympathoadrenal activity in fasting pregnant rats. J Clin Invest 1979;64: 109-16. 24. Young JB, Landsberg L. Effect of concomitant fasting and cold exposure on sympathoadrenal activity
in rats. Am J Physiol1981 ;240:E314-9. 25. Prete FR, Bergmann BM, Rechtschaffen A. The effect of total sleep deprivation on behavioral ther
moregulation in the rat. Sleep Res 1988;17:324. 26. Pouliot M, Leblanc J. Effects of reserpine, guanethidine, and a1phamethyl-DOP A on resistance to cold.
Arch Int PhysiaI1963;71:73-82. 27. Pouliot M. Catecholamine excretion in adreno-demedullated rats exposed to cold after chronic guaneth
