Prazosin modulates rapid eye movement sleep deprivation-induced changes in body temperature in rats


Birendra N. Mallick, School of Life Sciences, Jawaharlal Nehru University, New Delhi-110067, India. Tel.: +91 11 26704522; fax: +91 11 26742558; e-mail:


Prolonged rapid eye movement sleep deprivation (REMSD) causes hypothermia and death; however, the effect of deprivation within 24 h and its mechanism(s) of action were unknown. Based on existing reports we argued that REMSD should, at least initially, induce hyperthermia and the death upon prolonged deprivation could be due to persistent hypothermia. We proposed that noradrenaline (NA), which modulates body temperature and is increased upon REMSD, may be involved in REMSD- associated body temperature changes. Adult male Wistar rats were REM sleep deprived for 6–9 days by the classical flower pot method; suitable free moving, large platform and recovery controls were carried out. The rectal temperature (Trec) was recorded every minute for 1 h, or once daily, or before and after i.p. injection of prazosin, an alpha-1 adrenergic antagonist. The Trec was indeed elevated within 24 h of REMSD which decreased steadily, despite continuation of deprivation. Prazosin injection into the deprived rats reduced the Trec within 30 min, and the duration of effect was comparable to its pharmacological half life. The findings have been explained on the basis of REMSD-induced elevated NA level, which has opposite actions on the peripheral and the central nervous systems. We propose that REMSD-associated immediate increase in Trec is due to increased Na-K ATPase as well as metabolic activities and peripheral vasoconstriction. However, upon prolonged deprivation, probably the persistent effect of NA on the central thermoregulatory sites induced sustained hypothermia, which if remained uncontrolled, results in death. Thus, our findings suggest that peripheral prazosin injection in REMSD would not bring the body temperature to normal, rather might become counterproductive.


Rapid eye movement (REM) sleep loss adversely affects several body functions including food intake, energy metabolism, body weight, hormonal levels, immune functions and several behaviors including memory consolidation, concentration, aggressiveness as well as brain excitability (Bergmann et al., 1989; Everson and Crowley, 2004; Everson and Toth, 2000; Mallick et al., 2005; Rechtschaffen et al., 1989, 2002), and prolonged REM sleep deprivation (REMSD) is reported to be fatal (Kushida et al., 1989). REMSD increases Na-K ATPase activity in the brain (Adya and Mallick, 2000; Gulyani and Mallick, 1993, 1995) that would increase brain temperature. As a consequence for maintenance of thermal homeostasis in the body, there is likely to be a mechanism for reduction of body temperature after REMSD.

Direct and indirect evidences suggest that REMSD elevates the level of noradrenaline (NA) in the blood (Bergmann et al., 1989) and in the brain (Brock et al., 1994; Majumdar and Mallick, 2003; Porkka-Heiskanen et al., 1995; Shouse et al., 2000; Stern et al., 1971), and that elevated NA stimulated the Na-K ATPase activity in the brain (Gulyani and Mallick, 1995; Mallick et al., 2002b). On one hand, NA is known to stimulate the metabolic rate (Matthews et al., 1990; Ratheiser et al., 1998), the released energy is likely to elevate body temperature, which in turn should trigger heat dissipation processes for thermoregulation. On the other hand, it is known that NA acting on alpha-1 adrenoceptors in the medial preoptic area, the main thermoregulatory centre (Alam et al., 1995; McGinty and Szymusiak, 1990), induces hypothermia (Mallick and Alam, 1992; Poole and Stephenson, 1979), and the presence of alpha-1 adrenoceptors on the thermosensitive neurons has been confirmed (Mallick et al., 2002a). In addition, NA action in the periphery is reported to have effect opposite to that of the central effects (Handley and Spencer, 1972). Thus, the findings from these isolated and independent studies suggest that the brain temperature is likely to be elevated as a direct consequence of REMSD; however, to maintain normal physiological processes there should be triggering of heat dissipation processes possibly as a compensatory mechanism; further, REMSD-induced elevated NA by acting on alpha-1 adrenoceptor may play a significant role in such thermoregulation. There are conflicting reports on thermoregulation after REM sleep loss; some showed hypothermia (Bergmann et al., 1989; Landis et al., 1992; Patchev et al., 1991) while others have reported hyperthermia (Hoshino, 1996; Salin-Pascual et al., 1997). A common lacuna in all these studies was that none attempted to study the effect of REMSD within 24 h, especially on body temperature. Those that reported hypothermia have recorded the temperature after 7 or more days of deprivation, whereas those that reported hyperthermia recorded the temperature on or around 4 days of deprivation. Apparently these findings indicate reversal of body temperature upon prolongation of REMSD, which needed in depth systematic investigation for confirmation and possible explanation. Further, in almost all the studies where the body temperature was recorded on or around 4 days of REMSD, the most important control, the large platform control (LPC), was lacking; also, the role of NA in the process was not investigated. In this study, we ran adequate controls and studied the effects of varying period of REMSD (24 h–9 days) on rectal temperature (Trec), and also investigated the role of alpha-1 adrenoceptor antagonist, prazosin, in modulating such effects in freely moving normally behaving rats.



Healthy inbred male Wistar rats (260–300 g, n = 42) maintained under 12 : 12 h light : dark cycle (lights on at 07:30 hours) and at 25 ± 1 °C were used in this study. The rats were individually housed with ad libitum access to food and water. To minimize possible handling stress during experiments, prior to the start of the experiments, the rats were acclimatized to the experimental environment as well as to the experimenter by gentle handling acclimatized to about 10 min daily twice a day for at least 3 days; the rats were also acclimatized to Trec recording using a thermocouple-based rectal probe.


Rapid eye movement sleep deprivation was performed using the classical flower pot method (Gulyani and Mallick, 1993; Mendelson et al., 1974) where the rats were maintained on a small platform (diameter 6.5 cm), 2–3 cm elevated over surrounding water. At the onset of REM sleep, because of atonia in the antigravity muscles, the rats on the small platform tended to fall into or touch the surrounding water resulting in loss of REM sleep. The LPC group of rats was also kept individually over a platform surrounded by water just like the REMSD group except that the platform size was larger (diameter 13 cm) so that the rats could relax and enjoy all stages of sleep including REM sleep. In addition, in another intrinsic control experimental rats were allowed 4 days recovery of REM sleep after 6 days of REMSD. Another rat was maintained in the same room in a normal dry rat cage, the free moving control (FMC). To rule out non-specific environmental effects all the experimental and control sets of experiments were conducted simultaneously and in the same room. The experiments were conducted following NIH guidelines and were approved by the Institutional Animal Ethics Committee.

Temperature recording and injection

The Trec was recorded with the help of a thermocouple-based rectal probe. To facilitate introduction of the thermocouple and to reduce discomfort due to its entry through the anal opening, lidocaine gel was applied around the anal opening and on the tip of the probe before it was gently inserted ∼6 cm through the anal sphincter, held there undisturbed for about 30 s until the Trec stabilized before recording it. The Trec was recorded every minute for 1 h, or once daily, or before and after injections, depending on the experimental plan as described below. The probe was gently withdrawn after recording the Trec and the rats were returned on their respective platforms or dry home cages as the case may be.

Prazosin hydrochloride and N, N-dimethylacetamide (N, N-DA) were obtained from Sigma-Aldrich, USA. Because prazosin was sparingly soluble in normal saline, 20%N, N-DA in pyrogen-free normal saline was used as a vehicle for dissolving prazosin hydrochloride. Each rat from every group received intraperitoneal (i.p.) prazosin (4 mg kg−1, 0.5 mL) or vehicle (0.5 mL) injections on different days keeping all other parameters of the experiment similar. The dose of the prazosin was decided based on our earlier study (Gulyani and Mallick, 1995). No rat was used for another set of experiments to avoid repeated deprivation and possible acclimatization.

Experimental design

Effect of REMSD on Trec

Five sets of experiments were conducted to evaluate the effect of REMSD on the Trec. Each set contained one REM sleep-deprived rat, one LPC rat and one FMC rat. The rats were REM sleep deprived for 9 days; the LPC group of rats was maintained on a large platform over the water, and the FMC group of rats was maintained in individual home cages for an equivalent period. The Trec was recorded daily between 10:00 and 11:30 hours from experimental REM sleep deprived as well as the control FMC and LPC rats.

Effect of prazosin on REMSD-induced changes in Trec

The rats were divided into two major groups, the prazosin injection group (n = 15, five sets) and vehicle injection group (n = 12, four sets) each having rats from REM sleep deprived, LPC and FMC as mentioned above. The rats were REM sleep deprived for 6 days as before (Majumdar and Mallick, 2003); LPC as well as FMC rats were also maintained for an equivalent period. Prazosin was i.p. injected daily between 10:00 and 11:30 hours first after ∼24 h postREMSD and continued every 24 h for five consecutive days to all the experimental and control rats. Similarly, in the vehicle injection group, 20%N, N-DA was i.p. injected to all the rats. The Trec was recorded at the start of the experiment, and every day before prazosin or vehicle injection, after 30 min and 4 h postprazosin and vehicle injections. After 6 days of REMSD, the rats were allowed 4 days of recovery, during which no injection was given; however, the Trec was recorded daily between 10:00 and 11:30 hours during recovery.

Additionally, 11 temperature recordings were done on randomly selected days of deprivation (out of those 5 days of prazosin or vehicle injection) after the rectal probe was temporarily inserted into the rectum and gently fixed with adhesive surgical tape to the base of the tail and the rats were maintained on the small platform after the injections. The Trec was recorded every minute for 60 min to evaluate the changes in the Trec profile within an hour of prazosin or vehicle injection (i.p.) in the REM sleep-deprived rats. Five recordings were done from four prazosin-injected rats, while six recordings from three vehicle-injected rats. One rat from both the prazosin- and vehicle-injected groups was extra sensitive and, hence, recording could not be done after the injections.

Statistical analysis

The mean (±SEM) Trec within each of the experimental and control groups was tested for normality using Sigmastat software (Jandel Scientific, San Rafael, CA, USA). As the data passed the normality test, one-way repeated measure analysis of variance (RM-anova) followed by Tukey test was applied using Sigmastat to compare the mean (±SEM) of Trec within the same group of rats at different times. Further, the mean (±SEM) Trec between different groups of rats was statistically compared using one-way anova followed by Tukey test; at least P < 0.05 was taken as significantly different.


Effect of REMSD on Trec

The baseline Trec (mean ± SEM) was 38.3 ± 0.1 °C, which sharply increased by 1.3 ± 0.1 °C [F(1, 9) = 232.97; P < 0.001, compared to baseline; F(2, 14) = 14.80; P < 0.001, compared to either control groups of rats] and reached a peak of 39.5 ± 0.1 °C after 2 days of REMSD. Subsequently, although the Trec continued to fall consistently, it remained significantly elevated for up to 6 days of REMSD compared to the preREMSD, LPC and FMC. The Trec continued to fall, and by the 9th day Trec was comparable (statistically non-significant) to the baseline (Fig. 1). As one rat died on the 10th day of deprivation, we did not continue further.

Figure 1.

 Rectal temperature (Trec) (mean ± SEM) on different days in REM sleep deprived, free moving control (FMC) and large platform control (LPC) rats have been shown. Trec significantly increased even after the first day postREMSD and remained elevated compared to baseline and control (LPC and FMC) rats till 8 days after which Trec was comparable to baseline and controls. ‘*’– significance compared to preREMSD; ‘$’– significance compared to LPC; ‘&’– significance compared to FMC. *$&P < 0.05; **$$&&P < 0.01; ***$$$&&&P < 0.001.

Effect of recovery on REMSD-induced Trec

The Trec was recorded during REMSD and during recovery period. The mean (±SEM) Trec rose from 38.0 ± 0.1 °C to 39.2 ± 0.1 °C [F(1, 17) = 132.18; P < 0.001, compared to preREMSD; and F (2, 26) = 19.98; P < 0.001, compared to FMC and LPC group of rats] at the end of 6 days of REMSD. Thereafter, the rats were transferred to dry home cages to recover sleep including REM sleep while the Trec was monitored every day as before. The Trec (±SEM) returned to comparable preREMSD level of 38.2 ± 0.1 °C within 24 h of recovery [F(1, 17) = 42.18, P < 0.001, compared to Trec after REMSD] and it was maintained within 0.2 °C of baseline value throughout the following 4 days of recovery period that was studied (Fig. 2).

Figure 2.

 Rectal temperature (mean ± SEM) records of rats during REM sleep deprivation (REMSD) till 4 days of recovery are shown. After the 6th day of REMSD rats were allowed 4 days of recovery. ‘*’– significance compared to preREMSD (baseline) Trec; ‘$’– significance compared to large platform control; ‘&’– significance compared to free moving control. ***$$$&&&P < 0.001. FMC, free moving control; LPC, large platform control.

Effect on REMSD-induced changes in Trec within 60 min of prazosin injection

Alpha-1 adrenergic influence on the REMSD-induced changes in Trec was studied by comparing the effect of prazosin injection with that of vehicle injection. After 24 h REMSD the mean Trec (±SEM) dropped from 39.1 ± 0.2 °C to 38.5 ± 0.1 °C within 15 min of prazosin injection [F(1, 10) = 28.06; P < 0.01, as compared to preinjection; F(1, 10) = 10.20, P < 0.05, as compared to vehicle injection group]. The mean (±SEM) Trec further reduced to 37.7 ± 0.1 °C [F(1, 10) = 79.02; P < 0.001, as compared to vehicle injection group] after 30 min of prazosin injection (Fig. 3).

Figure 3.

 Every minute graphical plots of mean (±SEM) rectal temperature for 60 min upon prazosin and vehicle injection (i.p.) into REM sleep-deprived rats till 60 min. Arrow shows the time of intraperitoneal injection.

Effect on REMSD-induced changes in Trec within 4–24 h after prazosin injection

The Trec was significantly high after 24-h REMSD. After prazosin injection the Trec significantly reduced in REM sleep-deprived rats (as mentioned above); however, subsequently the Trec increased when recorded up to 4 h although it remained significantly lower compared to 4 h postvehicle injection [F(1, 8) = 35.08; P < 0.001]. Further, on continuation of REMSD when the Trec was recorded the following day, it showed a peak Trec comparable to that of post 24-h REMSD without prazosin injection [F(1, 9) = 53.72; P < 0.01, compared to preREMSD; F(2, 14) = 15.75; P < 0.001, compared to either of the control groups LPC and FMC]; also, it was comparable to postvehicle injection at 24 h after REMSD [F(1, 8) = 0.50, P = 0.501]. The second prazosin injection on completion of 48 h REMSD showed a comparable Trec as that of prazosin injection after 24 h. A similar trend was observed in the Trec recorded on the 3rd, 4th and 5th consecutive days before and after prazosin injection after every completed 24 h of REMSD (Fig. 4).

Figure 4.

 Five days of records of mean (±SEM) rectal temperature of REM sleep-deprived rats which received daily i.p. injections of prazosin (n = 15, five in each group viz REMSD, LPC and FMC) or vehicle (n = 12, four in each group). ‘*’– significant compared to vehicle injection group; ‘$’– significant compared to preinjection. **$$P < 0.01; ***$$$P < 0.001. FMC, freely moving control; LPC, large platform control; REMSD, REM sleep deprivation.


The major findings of this study are that the Trec sharply increased in rats within 24 h of REMSD by the flower pot method. Although the Trec gradually decreased with progression of deprivation, it remained significantly elevated and reached predeprivation level only on the 9th day of deprivation. Upon continuation of REMSD the Trec continued to fall apparently, resembling a reversal of Trec, i.e. initial hyperthermia followed by hypothermia. Further, i.p. injection of prazosin induced a significant decrease in Trec within 30 min of injection into REM sleep-deprived rats.

The role of adrenoceptor in modulating the Trec in REM sleep-deprived rats was unknown. The dose of prazosin was used based on our earlier REMSD studies (Gulyani and Mallick, 1995), hence, dose response was not carried out in this study as a compromise to sacrifice a minimum number of rats. As the objective of this study was to investigate the role of REMSD-induced NA in mediating the changes in Trec, we used the adrenoceptor antagonist. The role of NA in thermoregulation was already reported (Alam et al., 1995; McGinty and Szymusiak, 1990). Further, the following observations supported receptor-mediated effects. (i) The effect of prazosin lasted about 4 h which correlates well with the pharmacological half-life of prazosin (Chau et al., 1980; Hamilton et al., 1985; Jaillon, 1980) and (ii) vehicle injection did not affect the Trec. A fall in body temperature after prolonged REMSD was reported (Bergmann et al., 1989); however, the rise in temperature within 24 h after deprivation was unknown. The rise in Trec in this study was due to REMSD and not due to non-specific effects because it was not seen in the LPC rats, and also Trec was comparable to predeprivation levels within 24 h of recovery after REMSD. Although the mechanism is yet unknown, the findings support our hypothesis (see Introduction) that REMSD is likely to elevate the Trec, which should fall upon continuation of deprivation. One may argue that the observed changes in temperature in this study could be due to the flower pot method followed, unlike the disk over water method followed by others (Bergmann et al., 1989;Landis et al., 1992). However, it may be pointed out that in the disk over water method the control rats are also awakened the same number of times as that of the experimental rats, and earlier studies (Bergmann et al., 1989; Landis et al., 1992) did not report temperature changes within 24 h of REMSD.

The fall in Trec after i.p. prazosin injection was apparently contradictory to an earlier report that prazosin into medial preoptico-anterior hypothalamus raises body temperature (Mallick and Alam, 1992), which needs discussion. Body temperature homeostasis is maintained as a balance between heat loss and heat gain by modulating either or both, which may have central and/or peripheral regulation. The NA plays a significant role in thermoregulation, which, however, has opposite effects depending on its action on the central nervous system or the periphery (Handley and Spencer, 1972). Prazosin, an alpha-1 adrenoceptor antagonist, in the preoptic area increased Trec in normal rats suggesting NA acting on the central nervous system induces hypothermia. However, in the periphery, alpha-1 adrenoceptor was shown to induce NA-mediated vasoconstriction causing hyperthermia (Borbujo et al., 1989; Freedman et al., 1992; Gomez et al., 1991; Lindblad and Ekenvall, 1986; Nielsen et al., 1990). Direct and indirect studies have reported that upon REMSD circulating levels of NA or its synthetic enzyme are elevated in the blood (Bergmann et al., 1989) and in the brain (Brock et al., 1994; Majumdar and Mallick, 2003; Porkka-Heiskanen et al., 1995; Shouse et al., 2000; Stern et al., 1971). Because NA is known to be involved in thermoregulation, it was reasonable to assume that it might play an important role in REMSD- induced change in Trec, however, it was unknown if the elevated NA was the cause of hyperthermia or hypothermia. Further, it was also not known if the elevated NA in the brain and the periphery plays a differential role in REMSD-induced alteration in body temperature.

In this study, we observed that within 24 h of REMSD there was elevation of Trec which, however, showed a gradual fall upon continuation of deprivation. This may be explained by the fact that soon after REMSD there was increased metabolism, which is likely to be responsible for the elevation of Trec. This may further be supported by the fact that an increased metabolism (Bergmann et al., 1989; Hipolide et al., 2006; Koban and Swinson, 2005) as well as increased Na-K ATPase activity (Gulyani and Mallick, 1993) has been reported after REMSD. An earlier report from this laboratory showed that glucose 6-phosphatase activity was reduced in the brain within 24 h of REMSD and it was maintained even after continuation of deprivation for 4 days, suggesting that REMSD favored increased glucose metabolism (Thakkar and Mallick, 1993), supporting the present observation.

Elevated body temperature is injurious to brain and body, hence, there should be a compensatory regulation. It is known that NA levels increase in the brain after REMSD and central application of NA has been reported to induce hypothermia (Lin and Pivorun, 1986; Quan et al., 1992; Veale and Whishaw, 1976) by acting on alpha-1 adrenoceptors (Mallick and Alam, 1992), and the presence of such adrenoceptors has been shown on the hypothalamic thermosensitive neurons (Mallick et al., 2002a). Hence, we speculate that elevation of NA in the brain after REMSD is a compensatory thermoregulatory physiological response for the benefit of the body. Our contention may be supported by the fact that in spite of continuation of REMSD, the Trec did not rise significantly beyond the temperature that was reached within about 24 h of REMSD. Conversely, the Trec started declining upon prolongation of REMSD, which may be either due to increased heat dissipation or decreased heat generation or both. As the fall in temperature was slow and not abrupt, it is extremely unlikely that heat dissipation and heat generation were simultaneously affected in opposite directions. Pending confirmation, we hypothesize that upon continuation of REMSD, in the process to maintain body temperature, the rate of heat dissipation exceeds that of heat generation resulting in a slow fall in Trec. This view may be supported by the fact that there is increased NA in the hypothalamus after REMSD and that on prolongation of REMSD, when those mechanisms tend to fail, there is a rapid fall in temperature resulting in hypothermia as reported by others (Bergmann et al., 1989; Landis et al., 1992; Patchev et al., 1991) and also observed in this study (Fig. 1).

Apparent contradiction between earlier reports and the present observation that the Trec increased after REMSD that decreased following i.p. injection of prazosin may be explained as follows. REMSD is reported to increase NA in the blood (Bergmann et al., 1989) and in the brain (Brock et al., 1994; Porkka-Heiskanen et al., 1995; Shouse et al., 2000). Peripheral applications of NA and prazosin have been reported to induce vasoconstriction resulting in hyperthermia and vasodilatation resulting in hypothermia, respectively (Morris, 1994; Stephens et al., 2001; Struijker-Boudier et al., 1996); on the other hand central application of NA and prazosin has been reported to induce opposite effects, i.e. hypothermia (Lin and Pivorun, 1986; Quan et al., 1992; Veale and Whishaw, 1976) and hyperthermia (Mallick and Alam, 1992), respectively. Although our studies do not allow differentiating between central or peripheral effects of NA or prazosin, based on the findings of this study, subject to confirmation, we propose that REMSD-induced elevated NA initially increased Na-K ATPase activity (Mallick et al., 2002b), metabolism (Bergmann et al., 1989; Hipolide et al., 2006; Koban and Swinson, 2005) and food intake (Rechtschaffen et al., 2002) as well as peripheral vasoconstriction, resulting in hyperthermia as evidenced from increased Trec. Possibly, the elevated NA also acted on the hypothalamic thermosensitive neurons and facilitated heat dissipation as a compensatory mechanism and the body temperature starts reducing, resulting in hypothermia (Mallick and Alam, 1992). However, if the REMSD is prolonged, sustained action of NA on the preoptic area neurons caused persistent hypothermia, which if prolonged beyond limit causes thermoregulatory failure and becomes life threatening as reported earlier (Kushida et al., 1989) and also as observed by us (one animal).

Notwithstanding the above arguments, as a caution we must add that the entire effect of REMSD may not be mediated exclusively by NA acting on alpha-1 adrenergic receptors; the effects secondary to other changes in the body can not be ruled out. One school of thought against REMSD by platform method is if the method causes stress which may induce the effects. Arguably it is an important but debatable issue, and stress would depend on several factors including experimental protocol, life style, accommodation, adaptation, species, psychic and mental conditions and so on. Thus, although it is extremely difficult to reach a definite conclusion on such an open-ended debatable issue, we argue with reasonable confidence that our observations are unlikely to be due to the classical concept of stress effect because if the effects were due to stress the hyperthermia observed in this study would not have been reversed within 24 h. Our views may be supported by the fact that other studies which supported the questionable stress argument showed that REMSD-induced increased ACTH and corticosterone levels (Suchecki et al., 1998) did not reduce after 4 days of recovery (Hipolide et al., 2006), and it has been shown that REMSD by this method indeed is unlikely to be responsible for changes in serum corticosterone levels (Porkka-Heiskanen et al., 1995). Further, to rule out non-specific effects we have carried out LPC and recovery controls where the effects were not expressed.

In conclusion, it may be said that REMSD-associated rise in NA-induced heat production is due to increased Na-K ATPase activity, metabolism as well as peripheral vasoconstriction, leading to hyperthermia. However, upon prolongation of REMSD, sustained elevation of NA in the brain probably acts on the central thermoregulatory center and induces hypothermia. Subject to confirmation by designing appropriate experiments, based on our findings we propose that upon REMSD initially the body tries to maintain a balance between the heat production and dissipation. However, with prolongation of REMSD sustained effect of elevated NA on the preoptic area neurons results in persistent fall in body temperature, which if not controlled, becomes fatal (Kushida et al., 1989). These findings help us understand the pathophysiology of thermoregulation during REM sleep loss, particularly cardiovascular-respiratory and metabolic dysfunctions especially in relation to sleep disorders.


Funding from University Grant Commission, India, under UPOE scheme and Department of Biotechnology, India, to BNM and fellowship to MKJ from Council of Scientific and Industrial Research, India.

Conflict of interest

Authors report no conflict of interest.