Torcetrapib-induced blood pressure elevation is independent of CETP inhibition and is accompanied by increased circulating levels of aldosterone.

BACKGROUND AND PURPOSE: Inhibition of cholesteryl ester transfer protein (CETP) with torcetrapib in humans increases plasma high density lipoprotein (HDL) cholesterol levels but is associated with increased blood pressure. In a phase 3 clinical study, evaluating the effects of torcetrapib in atherosclerosis, there was an excess of deaths and adverse cardiovascular events in patients taking torcetrapib. The studies reported herein sought to evaluate off-target effects of torcetrapib. EXPERIMENTAL APPROACH: Cardiovascular effects of the CETP inhibitors torcetrapib and anacetrapib were evaluated in animal models. KEY RESULTS: Torcetrapib evoked an acute increase in blood pressure in all species evaluated whereas no increase was observed with anacetrapib. The pressor effect of torcetrapib was not diminished in the presence of adrenoceptor, angiotensin II or endothelin receptor antagonists. Torcetrapib did not have a contractile effect on vascular smooth muscle suggesting its effects in vivo are via the release of a secondary mediator. Treatment with torcetrapib was associated with an increase in plasma levels of aldosterone and corticosterone and, in vitro, was shown to release aldosterone from adrenocortical cells. Increased adrenal steroid levels were not observed with anacetrapib. Inhibition of adrenal steroid synthesis did not inhibit the pressor response to torcetrapib whereas adrenalectomy prevented the ability of torcetrapib to increase blood pressure in rats. CONCLUSIONS AND IMPLICATIONS: Torcetrapib evoked an acute increase in blood pressure and an acute increase in plasma adrenal steroids. The acute pressor response to torcetrapib was not mediated by adrenal steroids but was dependent on intact adrenal glands.

Specific interaction between a subset of the p53 protein family and heat shock proteins hsp72/hsc73 in a human osteosarcoma cell line.

We report here immunological evidence for the specific association between p53 and hsp72/hsc73 heat shock proteins in a human cell line derived from an osteosarcoma. The same association between p53 and hsp72/hsc73 was observed in HOS-TE85 clone 5 from which the HOS-SL cell line was derived. This association was indicated by the co-immunoprecipitation from HOS-SL of both p53 and hsp72/hsc73 proteins observed with either an anti-p53 monoclonal antibody (PAb421) or an antiserum specifically reacting with hsp72/hsc73 heat shock proteins. Furthermore, Western blot analysis allowed us to show that hsp72/hsc73 proteins did not share an epitope with p53, confirming that the co-immunoprecipitation of p53 and hsp72/hsc73 was attributable to the physical association of the proteins. Data obtained from SDS-PAGE show that the HOS-SL cells expressed two forms of p53 with distinct molecular weights. Both forms contained several species with different isoelectric points ranging between pH 6.0 and 6.5. The data obtained from both 1D and 2D gel analyses consistently show that the p53 proteins involved in the association with hsp72/hsc73 were mainly the species that migrated with the slower mobility in the SDS dimension. The possibility is discussed that the HOS-SL p53 variant forming a complex with hsp72/hsc73 contains an activating mutation for transformation.

Directional selection associated with clonal expansion of p53 mutant cells during neoplastic development of carcinogen-treated rat embryo lung epithelial cells.

In this study, rat embryo lung organ cultures were exposed to benzo[a]pyrene (B[a]P). After carcinogen-treatment the cells were dissociated and an epithelial cell line (BP) was developed from the primary cell culture derived from the carcinogen-treated explants. Investigations were performed on the sequential changes occurring in the course of neoplastic progression of BP cells and in the tumor cells that arose in vivo from implanted BP cells. During the neoplastic progression a mutation was shown to occur in p53 gene at codon 130 (AAG > AGG; Lys > Arg) in a single cell which expanded and gave rise to a predominant subpopulation. This mutational event was already detected at passage 14 but was probably not a direct consequence of a specific alteration caused by the carcinogen in the target cell. This mutation was retained through the subsequent progressional steps first as a heterozygous mutation, then converted to a homozygous state. From passage 18 on, it was possible in BP cell cultures to detect foci of larger morphologically distinct cells emerging on a background of cells maintaining the original morphology. These foci were shown to derive from a single cell carrying the p53 mutation in a homozygous state. During the neoplastic progression the mutant p53 allele frequency steadily increased and this mutant allele eventually came to predominate completely in the late stages of the neoplastic progression, including in the transplantation-induced tumors. The pattern of a directional selection for mutant p53 gene towards fixation is probably applicable to a wide range of human malignancies and may reflect the particular importance of this gene for tumorigenesis.

Identification and characterization of a p53 gene mutation in a human osteosarcoma cell line.

The nuclear phosphoprotein p53 occurs at elevated levels in many transformed cells. Mutant forms of mouse p53 possess enhanced transforming activity compared with wild type p53. Mutant mouse p53 proteins form complexes with the 70 kDa family of heat shock proteins (HSPs). We previously demonstrated an association between p53 and the 70 kDa HSPs in the human osteosarcoma (HOS) derived cell line HOS-SL. We report here the molecular cloning and sequencing of the p53 gene from HOS-SL cells, and demonstrate that it is in fact mutant. Further, analysis of similar HOS-derived cell lines demonstrates that they also encode the same mutant form of p53, whereas the wild type form of p53 appears to be lost in these cells. Stability studies demonstrate an increased half life of the p53 protein in these cells, in keeping with its association with the HSP 70 proteins. A potential role for this p53 mutant in the transformation process is discussed.

p53 mutations in Raji cells: characterization and localization relative to other Burkitt's lymphomas.

The nuclear phosphoprotein p53 is an important regulator of cell proliferation in normal cells. Interestingly, the gene encoding p53 has usually undergone mutations in a wide range of tumor types. Recent studies of the p53 gene in Burkitt's lymphomas have demonstrated that mutations are extremely common, and in fact it is rare that both alleles of the p53 gene in these tumors are not inactivated by mutation or deletion. We present here genetic data regarding the status of the p53 gene in the Burkitt lymphoma cell line, Raji. As is typical for this type of tumor, both alleles have undergone point mutations. Further, statistical analysis of available data from a large number of Burkitt's lymphomas indicates an apparent tumor-specific distribution of p53 mutations. The possibility that specific mutations of the p53 gene may be important for different tumor types is discussed.

Differential solubilization of proteins, phospholipids, free and esterified cholesterol of rat liver cellular membranes by sodium deoxycholate.

1. Smooth microsomes, Golgi-rich fractions, and light and heavy plasmalemmal subfractions from rat liver were isolated and their purity assessed using enzymic, chemical and morphological criteria. 2. Membranes were prepared by Tris-EDTA washing combined with sonication treatment of the different subcellular fractions. 3. Washed membranes were submitted to differential solubilization with 0.26% sodium deoxycholate. When the deoxycholate/phospholipid molar ratio (R) is raised, all the membranes showed a maximum protein solubilization occurring at R approximately equal TO 2. The higher the membrane neutral lipid to phospholipid molar ratio is, the lower the solubilized protein plateau lies. 4. Phospholipids are solubilized in slightly greater amounts than proteins and their solubilization is complete at R equals 14-16. 5. For R smaller than 2, sterols are solubilized in slightly greater amounts than phospholipids. At maximum protein solubilization, cholesterol and cholesterol esters completely differ in their behaviour. The whole membrane cholesterol goes into solution for R equals 14-16 while the solubilization of esterified cholesterol is never complete. The higher the protein plateau is, the lower the cholesterol esters solubilization curve asymptote lies.

Mitochondrial inhibition and oxidative stress: reciprocating players in neurodegeneration.

Although the etiology for many neurodegenerative diseases is unknown, the common findings of mitochondrial defects and oxidative damage posit these events as contributing factors. The temporal conundrum of whether mitochondrial defects lead to enhanced reactive oxygen species generation, or conversely, if oxidative stress is the underlying cause of the mitochondrial defects remains enigmatic. This review focuses on evidence to show that either event can lead to the evolution of the other with subsequent neuronal cell loss. Glutathione is a major antioxidant system used by cells and mitochondria for protection and is altered in a number of neurodegenerative and neuropathological conditions. This review also addresses the multiple roles for glutathione during mitochondrial inhibition or oxidative stress. Protein aggregation and inclusions are hallmarks of a number of neurodegenerative diseases. Recent evidence that links protein aggregation to oxidative stress and mitochondrial dysfunction will also be examined. Lastly, current therapies that target mitochondrial dysfunction or oxidative stress are discussed.

Neuroendocrine processes underlying ultradian sleep regulation in man.

Sleep is not a uniform state but is characterized by the cyclic alternation between rapid eye movement (REM) and non-REM sleep with a periodicity of 90-110 min. This cycle length corresponds to one of the oscillations in electroencephalographic (EEG) activity in the delta frequency band (0.5-3.5 Hz), which reflect the depth of sleep. To demonstrate the intimate link between EEG and neuroendocrine rhythmic activities in man, we adopted a procedure permitting simultaneous analysis of sleep EEG activity in the delta band and of two activating systems: the adrenocorticotropic system and the autonomic nervous system. Adrenocorticotropic activity was evaluated by calculating the cortisol secretory rate in blood samples taken at 10-min intervals. Autonomic activity was estimated by two measures of heart rate variability: 1) by the ratio of low-frequency (LF) to high-frequency (HF) power from spectral analysis of R-R intervals; and 2) by the interbeat autocorrelation coefficient of R-R intervals (rRR intervals between two successive cardiac beats). The results revealed that oscillations in delta wave activity, adrenocorticotropic activity, and autonomic activity are linked in a well-defined manner. Delta wave activity developed when cortisol secretory rates had returned to low levels and sympathetic tone was low or decreasing, as reflected by a low LF/HF ratio and by low levels in rRR. Conversely, the decrease in delta wave activity occurred together with an increase in the LF/HF ratio and in rRR. REM sleep was associated with a decrease in cortisol secretory rates preceding REM sleep onset, whereas the LF/HF ratio and rRR remained high. These results demonstrate a close coupling of adrenocorticotropic, autonomic, and EEG ultradian rhythms during sleep in man. They suggest that low neuroendocrine activity is a prerequisite for the increase in slow wave activity.

Ultradian oscillations in plasma renin activity: their relationships to meals and sleep stages.

The 24-h pattern of PRA was studied in 6 supine normal subjects, and the relationship between sleep stages and PRA oscillations was analyzed using 18 nighttime profiles and the concomitant polygraphic recordings of sleep. Blood was collected at 10-min intervals. The slow trends obtained by adjusting a third degree polynomial to the 24-h data were not reproducible among individuals, and no circadian pattern was detected. Sustained oscillations in PRA occurred throughout the day. Spectral analysis revealed that PRA oscillated at a regular periodicity of about 100 min during the night. This periodicity was modified during the daytime by meal intake, which induced PRA peaks with large interindividual variations in size. A close relationship was found between the nocturnal PRA oscillations and the alternance of rapid eye movement (REM) sleep and non-REM sleep. Non-REM sleep invariably coincided with increasing or peaking PRA levels. REM sleep occurred as PRA was declining or at nadirs. More precisely, increases in PRA marked the transition from REM sleep to stage II, whereas stages III and IV usually occurred when PRA was highest. This relationship between the periodic nocturnal oscillations in PRA and the alternance of the REM-non-REM cycles may translate a similar oscillatory process in the central nervous system or may be linked to hemodynamic changes during sleep that might be partly controlled by the renin-angiotensin system.

Modulation of episodic renin release during sleep in humans.

We previously described a strong concordance between nocturnal oscillations in plasma renin activity and sleep cycles. To examine whether modifying renal renin content or release influences the response to central stimuli linked to sleep stage alternation, plasma renin activity was measured every 10 minutes from 11:00 PM to 8:00 AM in three groups of six subjects. The first group received one 40 mg dose of the diuretic furosemide; the second group underwent the night experiment after 3 days on a low sodium diet; the third group received one 100 mg dose of the beta-blocker atenolol. Each subject underwent a control night when a placebo was given. The nocturnal curves were analyzed with a pulse detection program. For the control nights, 74 of the 83 sleep cycles were associated with significant plasma renin activity oscillations; non-rapid eye movement sleep occurred in the ascending portions and rapid eye movement sleep in the declining portions of the oscillations. These oscillations persisted in the three groups of subjects during the experimental nights and the relation with the sleep stages was not disturbed. Acute stimulation by furosemide amplified the oscillations and led to a general upward trend of the nocturnal profiles. Similarly, a low sodium diet, which led to a slow increase in renal renin content, provoked large oscillations with high initial levels. However, in both cases the mean relative amplitude of the oscillations, expressed as a percentage of the nocturnal means, was similar to that of the control nights and approximated 60%.(ABSTRACT TRUNCATED AT 250 WORDS)

Twenty-four-hour profiles of plasma renin activity in relation to the sleep-wake cycle.

OBJECTIVE: To evaluate the relative contribution of sleep and the endogenous circadian rhythmicity in producing the 24-h variations in the plasma renin activity. METHODS: Ten normal young men were studied, under basal conditions with normal nocturnal sleep from 2300-0700 h and once after a night of total sleep deprivation followed by 8 h daytime sleep from 0700 to 1500 h. Plasma renin activity was measured every 10 min for 24 h and the profiles were analysed using the pulse detection program ULTRA. RESULTS: During the 8 h night-time sleep a significant increase in the mean plasma renin activity levels occurred compared with the subsequent 8-h waking periods. After the shift in the sleep period, a sleep-associated increase was clearly apparent during the daytime hours. The number of the amplitude of the oscillations, linked to the non-rapid eye movement-rapid eye movement sleep cycles, increased during sleep (at whatever time it occurred), and were dependent on the regularity and the length of the sleep cycles. In awake subjects the plasma renin activity generally fluctuated in a more damped and irregular manner, but occasionally the plasma renin activity oscillated at a regular periodicity with two dominant peaks centred around 100 and 50 min. CONCLUSION: These results demonstrate that the 24-h plasma renin activity variations are not circadian in nature but are related to sleep processes, which create the nycthemeral rhythm by increasing both the frequency and the amplitude of the oscillations.

K-complexes and sleep spindles before transient activation during sleep.

Six subjects spent three consecutive nights in the sleep laboratory. Activation phases (PATs), spontaneous K-complexes, and sleep spindles were visually detected in sleep stages 2 and 3 for nights 2 and 3. The K-complex rate was significantly greater in the 10 sec prior to the PATs than at any other time spent in stage 2 or 3. K-complexes associated with sheep spindles occurred significantly less frequently during the epochs just preceding the PATs. In all subjects, there was a sharp increase of sleep spindles associated with K-complexes when PATs did not follow within 10 sec. These results suggest that spontaneous K-complexes and sleep spindles act antagonistically with respect to the occurrence of PATs. These two phasic events are significantly related to regulating the probability of occurrence of PATs in sleep stages 2 and 3; K-complexes may reflect an organismic state leading towards PAT, whereas sleep spindles may inhibit the occurrence of PAT.

Dynamic relation of sleep spindles and K-complexes to spontaneous phasic arousal in sleeping human subjects.

K-complexes unaccompanied by sleep spindles (K0-complexes) and isolated sleep spindles during stage 2, non-rapid eye movement (NREM) sleep were examined before and after transient activation phase (AP) and also pseudo-AP during human sleep to determine the relationship of K0-complexes and sleep spindles to APs. Sixteen sleep records obtained from 16 young adult males were scanned for isolated APs during stage 2 NREM sleep. One hundred APs and 62 pseudo-APs were identified and analyzed. The number of sleep spindles decreased and reached its minimum at the onset of APs, when an increase was observed in the number of K0-complexes. APs occurred when a decrease in sleep spindles was coupled with an increased incidence of K0-complexes, thus forming an antagonistic relation. A similar antagonism was observed between sleep spindles and slow waves. A working hypothesis was formulated to interpret a triad of sleep events: sleep spindles. K0-complexes, and slow wave sleep. Three kinds of sleep--REM, spindles-dominant, and slow-wave-dominant--are suggested as more useful classifications than the Rechtschaffen and Kales categories.

Comparative effect of night and daytime sleep on the 24-hour cortisol secretory profile.

To determine whether cortisol secretion interacts with daytime sleep in a similar manner to that reported for night sleep, 14 healthy young men were studied during two 24-hour cycles. During one cycle they slept during the night, during the other the sleep period was delayed by 8 hours. Secretory rates were calculated by a deconvolution procedure from plasma cortisol, measured at 10-minute intervals. The amount of cortisol secreted during night sleep was lower than during the corresponding period of sleep deprivation (12.7 +/- 1.1 vs. 16.3 +/- 1.6 mg; p < 0.05), but daytime sleep beginning at the habitual time of morning awakening failed to inhibit cortisol secretion significantly. There was no difference between the amount of cortisol secreted from 0700 to 1500 hours in sleeping subjects and in subjects who were awake during the same period of time (24.2 +/- 1.5 vs. 22.5 +/- 1.4 mg). Even if the comparison between sleeping and waking subjects was restricted to the period 0700-1100 hours or 0700-0900 hours, no significant difference was found. Neither secretory pulse amplitude nor frequency differed significantly in either period. However, detailed analysis of the secretory rates in day sleepers demonstrated a transient decrease in cortisol secretion at about the time of sleep onset, which began 10 minutes before and lasted 20 minutes after falling asleep. Spontaneous or provoked awakenings had a determining influence on the secretory profiles. Ten to 20 minutes after awakening from either night or day sleep cortisol secretion increased significantly.(ABSTRACT TRUNCATED AT 250 WORDS)

Slow oscillations of plasma glucose and insulin secretion rate are amplified during sleep in humans under continuous enteral nutrition.

Concomitant oscillations of plasma glucose and insulin secretion rate with a periodicity of about 80 minutes have been identified in normal humans. To determine whether these slow oscillations are influenced by sleep, peripheral levels of glucose and C-peptide were measured at 10-minute intervals over 24 hours in seven subjects, once with a normal nocturnal sleep from 2300 to 0700 hours, and once with a shifted daytime sleep from 0700 to 1500 hours. The subjects received continuous enteral nutrition and remained supine for the 8 hours preceding blood sampling and throughout the whole experiment. Insulin secretion rate was estimated by deconvoluting peripheral C-peptide levels using an open two-compartment model. The amplitude of glucose and insulin secretion rate oscillations increased by 160% during the 8-hour sleep periods, at whatever time they occurred, whereas the influence of the time of day was not significant. Glucose and insulin secretion rate mean levels were also significantly increased during normal nocturnal sleep compared to the remaining 8-hour waking periods, but this effect did not persist when sleep was shifted to the daytime. The number of oscillations was similar in both experimental series and was not affected by sleep. No systematic concordance was found between glucose and insulin secretion rate oscillations and the rapid eye movement-nonrapid eye movement sleep cycles, despite them having similar periodicities. This study demonstrates that increased amplitude of glucose and insulin secretion rate oscillations is related to sleep rather than to the time of day, without any associated frequency variations.(ABSTRACT TRUNCATED AT 250 WORDS)

Comparison of cardiovascular responses to noise during waking and sleeping in humans.

Eighty subjects, 40 men and 40 women, were allocated to one of two groups according to their self-estimated high or low sensitivity to noise. In the first part of the experiment, they were exposed to sequences of common noises during the morning or the afternoon. The heart-rate and finger-pulse responses were measured and recorded in relation to sensitivity, sex of subjects, and time of day. The different types of noise were compared for both responses. The heart-rate response showed differences between sensitivity groups but not between noises. In contrast, no significant differences were obtained between sensitivity groups when using the finger-pulse response, but clear differences were observed between noises. In a second part of the experiment, 10 men and 10 women subjects were selected from the previous two sensitivity groups. These 20 subjects were exposed during sleep to the same noises as during the daytime. Heart-rate and finger-pulse responses during sleep were significantly greater than during waking, and they did not differ significantly with respect to sensitivity to noise or gender. These two autonomic responses showed differences between noises that appeared to be related to their noise-equivalent-level value. Compared with the silent baseline night, the sleep pattern showed no significant modification in the night of noise disturbance, except for the frequency of transient activation phases, which was significantly increased in the latter.

Nocturnal oscillations in plasma renin activity and REM-NREM sleep cycles in humans: a common regulatory mechanism?

To establish the strength of the relationship between the nocturnal oscillations in plasma renin activity (PRA) and the sleep stage patterns, 42 PRA profiles from blood collected at 10-min intervals and the concomitant polygraphic sleep recordings were analyzed. In all cases, PRA curves exactly reflected the pattern of sleep stage distribution. When sleep cycles were complete, PRA levels oscillated at a regular 100-min period, with a strong spectral density. Declining PRA levels always coincided with REM sleep phases and increasing levels with NREM sleep phases. More precisely, peak levels corresponded to the transition from deep sleep stages toward lighter ones. The start of the rises in PRA generally marked the transition from REM sleep to stage 2. For incomplete sleep cycles, PRA curves reflected all disturbances and irregularities in the sleep structure. Spontaneous and provoked awakenings blunted the rise in PRA normally associated with NREM sleep, which indicates that disturbing sleep modifies the renin release from the kidneys. These results suggest that a common mechanism within the central nervous system controls both PRA oscillations and the REM-NREM sleep alternation.

Relative and combined effects of heat and noise exposure on sleep in humans.

In a counter-balanced design, the effects of daytime and/or nighttime exposure to heat and/or traffic noise on night sleep were studied in eight healthy young men. During the day, the subjects were exposed to baseline condition (ambient temperature = 20 degrees C; no noise) or to both heat (35 degrees C) and noise. The duration of the daytime exposure was 8 h ending 5 h before sleep onset. The following nights, the subjects slept either in undisturbed (20 degrees C; no noise) or in noise, heat, or noise plus heat-disturbed environments. During the day, the various types of traffic noise were distributed at a rate of 48/h with peak intensities ranging between 79 and 86 dB(A). The background noise level was at 45 dB(A). At night, the peak intensities were reduced by 15 dB(A), the rate was diminished to 9/h, and the background noise was at 30 dB(A). Electrophysiological measures of sleep and esophageal and mean skin temperatures were continuously recorded. The results showed that both objective and subjective measures of sleep were more disturbed by heat than by noise. The thermal load had a larger impact on sleep quality than on sleep architecture. In the nocturnal hot condition, total sleep time decreased while duration of wakefulness, number of sleep stage changes, stage 1 episodes, number of awakenings, and transitions toward waking increased. An increase in the frequency of transient activation phases was also found in slow-wave sleep and in stage 2. In the nocturnal noise condition, only total number of sleep stage changes, changes to waking, and number of stage 1 episodes increased.(ABSTRACT TRUNCATED AT 250 WORDS)

Prolactin secretion and sleep.

To clarify the relationship between prolactin (PRL) secretion and sleep, three experimental procedures were employed and secretory rates were estimated from plasma levels using a deconvolution procedure. Eight healthy young men participated in two 24-hour studies, one using normal night sleep and one using delayed sleep, to determine the influence of sleep as a whole on the PRL rhythm. Another group of 24 subjects underwent a 1-night study to investigate the relationship between PRL secretion and the internal sleep structure. The influence of sleep quality was studied in two more groups of eight subjects. Secretory rates were calculated by deconvolution from plasma PRL measured at 10-minute intervals. Sleep was recorded polygraphically in all experiments. PRL secretory pulses occurred throughout the 24-hour cycle without significant variation in frequency, but with enhanced pulse amplitude for both night and day sleep periods. Sleep onset was rapidly followed by an increase in secretion, and awakenings coincided with an immediate offset of active secretion. Analyzing the association between secretory pulses and sleep stages demonstrated that PRL secretory rate is low at the time of rapid eye movement sleep onset. Sleep quality appeared not to affect the PRL secretory profile. These results confirmed that PRL secretion is enhanced during the whole sleep period, as inferred from plasma levels. Considering secretory pulses provides a precise determination of the temporal relations between PRL and sleep structure and demonstrates that occasionally poor sleep does not influence PRL secretion in normal humans.

Effect of continuous heat exposure on sleep stages in humans.

Six young men were exposed to a thermoneutral environment of air temperature (Ta) 20 degrees C for 5 days and nights followed by an acclimation period of 5 days and nights at Ta 35 degrees C and 2 recovery days and nights at Ta 20 degrees C. Electrophysiological measures of sleep, esophageal temperature, and mean skin temperature were continuously monitored. The total nocturnal body weight loss was measured by a sensitive platform scale. Compared with the 5 nights of the baseline period at 20 degrees C, sleep patterns showed disturbances at 35 degrees C. Total sleep time was significantly reduced, while the amount of wakefulness increased. The subjects exhibited fragmented sleep patterns. The mean duration of REM episodes was shorter at 35 degrees C than at 20 degrees C of Ta, while the REM cycle length shortened. In the acclimation period, there was no change in sleep pattern from night to night, despite adaptative adjustments of the thermoregulatory response. The protective mechanisms of deep body temperature occurring with heat adaptation did not interact with sleep processes. Upon return to baseline condition, a recovery effect was observed on a number of sleep parameters which were not significantly affected by the preceding exposure to prolonged heat. This would suggest that during exposure to dry heat, the demand for sleep could overcome that of other regulatory functions that are temperature-dependent. Therefore, a complete analysis of the effect of heat on sleep parameters can be assessed only if heat exposure is compared with both baseline and recovery periods.