RESULTS

The effect of REM and total sleep deprivation on somatostatin mRNA

    REM sleep deprivation for 24 h  increased the number of  somatostatin mRNA expressing cells significantly in the arcuate nucleus when compared to 72 h REM sleep deprivation and control groups (F_{(4, 22)}=6.72, P = 0.001, Duncan P < 0.05) (II) (Fig. 7a). After 72 h of REM sleep deprivation there was no difference in the number of somatostatin mRNA expressing cells in the arcuate nucleus when compared to the corresponding large platform control or home group.  In the periventricular nucleus the number of somatostatin cells was significantly higher after 72 h of REM sleep deprivation compared to large platform controls (F_{(4, 18)}=3.42, P = 0.03, Duncan P < 0.05) but neither treatment group differed from the home control group (II) (Fig. 7b). 24 h REM sleep deprivation did not affect the number of somatostatin mRNA expressing cells in the periventricular nucleus. After a 24 h recovery after 72 h REM sleep deprivation the number of somatostatin mRNA expressing cells in the arcuate and periventricular nuclei did not differ from the control and 72 h deprivation groups (II) (Fig. 7a-b).
 

Fig. 7  The average number of somatostatin (SS) mRNA-expressing cells in the rat arcuate (ARC) (a) and periventricular (PeN) (b) nuclei. H: home control, S: REM sleep deprivation with the platform method (small platforms), L: large platform controls. D24 and D72: deprivation for 24 and 72 hours, D72R24: deprivation for 72 h followed by 24 h of rebound sleep in normal conditions. Vertical bars indicate the standard error of means (SE). * Duncan’s test P < 0.05 after one-way ANOVA at significance level P < 0.05.
 

  Both 6 and 12 h total sleep deprivation increased the amount of somatostatin mRNA in the arcuate nucleus measured by densitometric analysis when compared to the control groups (F_(1,20)= 5.84, P = 0.03) (III) (Fig. 8a). Cell count showed a similar but non-significant trend between the deprivation and control groups (F_(1,20)= 3.04, P = 0.10). In the periventricular nucleus,  total sleep deprivation did not affect the amount of somatostatin mRNA when compared to controls (F_{(1, 20)}= 0.01, P = 0.92)(III) (Fig. 8b).
 

Fig. 8  Mean optical density of somatostatin (SS) mRNA visualized by in situ hybridization in the arcuate  (ARC) (a) and periventricular (PeN) (b) nuclei. Groups: control (C), sleep deprivation by gentle handling (SD) either for 6 h during the light phase (decapitated at 15:00) (6 h) or 12 h during the dark phase (decapitated at 9:00) (12 h). * The amount of somatostatin mRNA increased as a consequence of total sleep deprivation in the arcuate nucleus (two-way ANOVA: P = 0.03 for SD vs. control).

The effect of REM and total sleep deprivation on GHRH mRNA

    REM sleep deprivation for 24 and 72 h on the small platforms decreased the number of GHRH mRNA expressing cells in the paraventricular nucleus (II) (Fig. 9b). 72 h on the large platforms also decreased the number of GHRH mRNA cells when compared to home and 24 h large platform groups (F_{(4, 19)}= 9.01, P = 0.0003, Duncan P < 0.05). In the arcuate nucleus the number of GHRH mRNA cells tended to be lower in all platform treatment groups when compared to the home controls. (F_{(4, 21)}= 2.57, P = 0.07) (II) (Fig. 9a). 24h rebound sleep after 72 h deprivation had no effect on the number of GHRH mRNA expressing cells in the paraventricular or arcuate nuclei when compared to 72 h deprivation or large platform controls (II) (Fig. 9).

Fig. 9  The average number of GHRH mRNA-expressing cells in the rat arcuate (ARC) (a) and paraventricular (PaV)  (b) nuclei. H: home control, S: REM sleep deprivation with the platform method (small platforms), L: large platform controls. D24 and D72: deprivation for 24 and 72 hours, D72R24: deprivation for 72 h followed by 24 h of rebound sleep in normal conditions. Vertical bars indicate the standard error of means (SE). * Duncan’s test P < 0.05 after one-way ANOVA at significance level P < 0.05.
 

    Total sleep deprivation for 6 h during the light phase increased the amount of GHRH mRNA in the paraventricular nucleus measured by densitometric analysis when compared to controls and 12 h deprivation during the dark phase (F_{(3, 20)}= 3.57, P = 0.03, Newman-Keuls: P < 0.05) (III) (Fig. 10a). 12 h total sleep deprivation did not affect the amount of GHRH mRNA when compared to controls. Cell count from the same areas showed a similar but non-significant result (F_{(3, 20)}= 1.71, P = 0.20). In the arcuate nucleus 6 or 12 h sleep deprivation did not affect the amount of GHRH mRNA (F_(1,20)=0.43, P=0.52) (III) (Fig.10b). In the periventromedial area  the amount of GHRH mRNA was significantly higher in the morning (9:00am) than in the afternoon (3:00pm) (F_(1,20)= 5.17, P = 0.03) (III) (Fig.10c). 6 or 12 h sleep deprivation did not have any effect when compared to the control treatments (F_{(1, 20)}= 0.15, P = 0.71). Cell count showed again a similar but non-significant trend in the difference between 9:00am and 3:00pm groups in the periventromedial area (F_{(1, 20)}= 3.55, P = 0.07).
 

Fig. 10  Mean optical density of GHRH mRNA visualized by in situ hybridization in the paraventricular (PaV) (a) and arcuate (ARC) nuclei (b) and in the periventromedial (pVMH) area (c). Groups: control (C), sleep deprivation by gentle handling (SD) either for 6 h during the light phase (decapitated at 15:00) (6 h) or 12 h during the dark phase (decapitated at 9:00) (12 h). * The amount of GHRH mRNA increased after 6 h total sleep deprivation in the arcuate nucleus when compared to all other groups (one-way ANOVA P = 0.03, Newman-Keuls P < 0.05) (a). The amount of GHRH mRNA was higher in the rats decapitated in the morning (9:00) than in the afternoon (15:00) in the periventromedial area (two-way ANOVA: P = 0.03 for 9:00 (12 h) vs. 15:00 (6 h)) (c).
 

The effect of REM and total sleep deprivation on galanin mRNA

    After 24 h REM sleep deprivation on the small platforms the number of galanin mRNA expressing cells increased in the medial preoptic area and the periventricular nucleus when compared to home and large platform controls (I) (Fig. 11). There was no significant difference in the number of galanin mRNA expressing cells between the two control groups (home and large platforms). 6 hour total sleep deprivation during the first half of the light phase and 12 hour  deprivation during the dark phase did not affect the number of galanin mRNA expressing cells  in the medial preoptic area and the periventricular nucleus (Toppila et al. 1996).

Fig. 11  Number of galanin mRNA-expressing cells in the medial preoptic area (MPA) (a) and in the periventricular nucleus (PeN) (b) after  REM sleep deprivation or control conditions. S: 24 h REM sleep deprivation on small platforms, L: large platform controls, H: home controls. * one-way ANOVA P < 0.05, Newman-Keuls: S vs. L and H P < 0.05.

The effect of REM sleep deprivation on plasma GH

    Secretion of GH occurred in typical secretion pulses at approximately 3 h intervals. These pulses were generally much lower in the REM sleep deprived (Fig. 12b) and large platform rats (data not shown) than in the home controls (Fig. 12a). The plasma GH content approximated by the integrated area under the curve (AUC) was significantly lower during REM sleep deprivation (small platforms) and also during large platform treatment when compared to the corresponding home control group (small platforms: t₍₇₎ = 2.49, P = 0.04, large platforms: t₍₈₎= 2.89, P = 0.02) (II).

Fig. 12  The effect of REM sleep deprivation on the rat plasma GH profile. (a) untreated animal, (b): REM sleep deprivation with the platform method. The samples were collected between 24-30 h of REM sleep deprivation. The amplitude of GH pulses is lower in the REM sleep deprivation animal.
 
 

I.C.V. INJECTIONS OF SOMATOSTATIN, SOMATOSTATIN ANTAGONIST,  AND GALANIN

Somatostatin and somatostatin antagonist

    I.c.v. injection of 0.5 and 2 nmol of somatostatin antagonist reduced  the amount of REM sleep during the post injection period from 0.5-2 hours when compared to artificial CSF controls (ANOVA: F_{(2, 8)} = 4.66 P < 0.05, Newman-Keuls: P < 0.05) (IV). (Fig. 13a) There was no difference between the two doses of somatostatin antagonist. Later, during the periods 2-4 and 4-6 hours, there was no significant difference in the amount of REM sleep between the treatments. Somatostatin antagonist did not affect the amount of non-REM sleep during the 6 h post-injection recording period. I.c.v. injected boluses of corresponding doses of somatostatin did not affect the amount of REM (Fig. 13b) or non-REM sleep during any post-injection time period when compared to artificial CSF controls (IV).

Fig. 13  The effect of injected somatostatin (SS) and somatostatin antagonist (SA) on mean proportions of REM sleep during three post-injection time periods.  a: i.c.v. injection of artificial CSF (aCSF) (vehicle control), SA 0.5 or 2 nmol ( n=5), * F(2, 8)= 4.66, P < 0.05, Newman-Keuls: SA 0.5 and SA 2 vs. aCSF P < 0.05. b: i.c.v. injection of aCSF, somatostatin 0.5 or 2 nmol (n=5). c: i.c.v. injection of aCSF or 2 nmol of SA after 24h REM sleep deprivation (n=5). * Paired t-test: t(4)= 4.36, P < 0.05. d: microinjection of aCSF, somatostatin  0.25nmol or SA 0.25nmol (n=8), * ANOVA  F(2, 10) = 8.09, P < 0.01, Bonferroni’s t-test: SA vs.aCSF t(6)=5.10, P < 0.01. Vertical bars indicate the standard error of means (SE).
 

Somatostatin antagonist after REM sleep deprivation

    After 24 h REM sleep deprivation an i.c.v. injection of 2 nmol of somatostatin antagonist reduced REM sleep during the time period 2-4 h post injection when compared to  artificial CSF (t₍₄₎ = 4.36, P < 0.05) (IV) (Fig. 13c). During the periods 0.5-2 and 4-6 h there was no significant difference in the amount of REM sleep. Non-REM sleep was not significantly affected. 0.5 nmol of somatostatin antagonist after REM sleep deprivation did not significantly affect the amount of REM or non-REM sleep during the 6 h recording of recovery sleep when compared to artificial CSF injection (IV).
 
 
 

Galanin

    I.c.v. injection of 0.06, 0.6 or 6 nmol of galanin did not affect the daytime proportions of REM  or  non-REM sleep when compared to controls during the 8 h post-injection recording (I) (Tab. 1). During the first 8 h of the dark phase, when the amount of natural sleep is low in the rat, 0.6 nmol of galanin i.c.v. did neither affect the amount of REM or non-REM sleep when compared to control injection (I) (Tab. 1).
 
 
 

	REM sleep		non-REM sleep
	%	SD	%	SD
Light
Saline	10.0	1.8	51.9	3.7
Galanin 0.06	9.6	2.0	52.9	7.2
Galanin 0.6	12.8	2.6	48.7	5.0
Galanin 6	12.8	2.4	49.7	7.8
Dark
Saline	5.0	0.5	27.7	2.0
Galanin 0.6	5.0	0.9	21.7	1.9

Tab. 1  The effects of i.c.v. injections of 0.06-6 nmol of galanin on sleep compared to saline controls. Sleep polygraphy was recorded 8 h during light (started ar 9:00 am) or dark (started at 8:00 pm) phase (lights 8:00 am - 8:00 pm). SD = standard deviation. n = 5-6 per group.
 
 

Microinjections of somatostatin, somatostatin antagonist and galanin into the locus coeruleus

    The accepted sites of the microinjection were in the locus coeruleus or in the immediate vicinity of the nucleus not penetrating the fourth ventricle (Fig. 6). Microinjection of 0.25 nmol of somatostatin antagonist into the locus coeruleus reduced REM sleep during the post injection period 0.5-2 hours (ANOVA: F_{(2, 10)}= 8.09, P < 0.01, Bonferroni t₍₆₎= 5.10, P < 0.01) (IV) (Fig. 13d). Non-REM sleep was not significantly affected. A corresponding dose of somatostatin did not affect post-injection proportions of sleep phases when compared to artificial CSF (IV). Microinjection of 0.25 nmol of galanin or a combined injection of 0.25 nmol of galanin and 0.25 nmol of somatostatin did not affect REM or non-REM sleep during the 6 post-injection hours when compared to artificial CSF injection (IV).
 

-> DISCUSSION