Calcif Tissue Int (1991) 49:373-377

Calcified Tissue International 9 1991 Springer-Verlag New York Inc.

Diurnal Rhythm in Serum Osteocalcin: Relation with Sleep, Growth Hormone, and PTH(1-84) Henning K. Nielsen, 1'3 K. Brixen, 1 M. Kassem, 1 S. Engkjaer Christensen, 2 and L. Mosekilde 1 IUniversity Department of Endocrinology and Metabolism, Aarhus Amtssygehus, DK 8000 Aarhus C, Denmark; 2Second University Clinic of Medicine, Aarhus Kommunehospital, Denmark; and 3Department of Clinical Chemistry, Randers Centralsygehus, Denmark Received January 4, 1991, and in revised form May 13, 1991

Summary. We examined the role of sleep, growth hormone (GH), and parathyroid hormone [PTH(1-84)] as regulators of the diurnal rhythm of the osteoblastic bone marker, serum osteocalcin (OC). Nine normal subjects were followed with hourly blood sampling during one 24-hour period with normal sleep pattern, and one 24-hour period with absolute sleep deprivation. We found that the rhythm in serum OC did not exhibit significant changes (P > 0.50). Serum OC (mean -+ SE) was 30.9 -+ 2.5 ~g/liter during sleep (2330-0730 hours) versus 29.9 -+ 4.9 ixg/liter during sleep deprivation (not significantly different). The serum GH rhythm was significantly different on the two occasions (P < 0.01). A maximum GH peak (mean -+ SE) of 10.3 • 2.4 txg/liter occurred at 0136 hours - 6 minutes during sleep compared with a maximal peak of 7.6 -+ 1.2 txg/liter (P < 0.01) at 0245 hours - 20 minutes (P < 0.01) during sleep deprivation. During sleep (2330-0730 hours), mean serum GH was 3.61 --- 0.60 p~g/liter compared with 2.39 --- 0.40 ixg/liter during sleep deprivation (P < 0.005). Small insignificant changes occurred in serum PTH(1-84) and serum ionized calcium during the two occasions. We conclude that sleep and GH are not acute controlling factors of the diurnal rhythm in serum OC and the role of serum PTH(1-84) remains unsettled. Key words: Circadian rhythm - Osteocalcin - Sleep deprivation.

Circadian rhythmicity of bone cell activity has been reported in animal studies, applying microchemical techniques directly on bone tissue [1]. In humans, the existence of bone cell periodicity is supported by the findings of diurnal rhythms in biochemical bone markers. Serum osteocalcin (OC) [bone 7-carboxyglutamic acid-containing protein, (BGP)], a specific marker of osteoblastic activity [2-5], varies diurnally with high nocturnal levels [6--9] in a rhythm that is temporally related with the rhythm in bone isoenzyme alkaline phosphatase [10], another marker of osteoblastic activity [ 11]. The environmental and hormonal factors controlling diurnal changes in serum OC are unknown. The sleep-wake cycle has a synchronizing effect on many biological rhythms [12, 13] and it is possible that sleep directly or indirectly influences serum OC, as serum OC increases just around bedtime and decreases in the morning. Among the possible

Offprint requests to: H. K. Nielsen

hormonal factors are several of the major calciotropic hormones. Serum cortisol [14], parathyroid hormone (PTH) [15, 16], and growth hormone (GH) [17] show characteristic diu r n a l v a r i a t i o n s , w h e r e a s s e r u m l e v e l s o f 1,25dihydroxyvitamin D 3 (1,25 (OH)2D 3) have no great daily fluctuations [18]. A possible role of cortisol and PTH as controlling factors is further indicated by two studies showing close temporal negative correlation between serum cortisol and serum OC [15, 19], and another study suggesting a temporal positive association between serum levels of intact parathyroid hormone [PTH(1-84)] and serum OC [15]. GH, however, may also be a factor mediating the diurnal rhythm in serum OC. In children with GH deficiency, administration of GH normalizes the low serum OC and abnormal nocturnal profile in serum OC [20]. Moreover, GH normalizes the abnormal diurnal rhythm in serum OC seen in patients of idiopathic short stature [21]. Furthermore, GH secretion has a characteristic 24-hour profile, where the major event is a spike occurring just before or shortly after bedtime [11, 2226], the same time as the increase in serum OC. The diurnal rhythm in serum cortisol is relatively resistant to acute changes in the sleep-wake rhythm, but is changed after several days reversal of the sleep-wake rhythm [12, 13]. The diurnal secretion of GH is considered strongly associated with the sleep-wake cycle and is generally related to the deep sleep stages III and IV [17, 25, 26]. A sudden shift in the timing of sleep should, therefore, immediately change the pattern of GH secretion [12] and leave the pattern of serum cortisol unchanged. The relation between serum PTH(1-84)/serum Ca 2§ and sleep has not been previously addressed. In the present study, we examined the role of sleep, GH, and PTH as acute regulators of the diurnal rhythm of serum OC during study conditions in which the diurnal rhythm of serum OC was unlikely to be influenced by changes in serum cortisol. The diurnal rhythms of serum OC, serum GH, and PTH(1-84) during a normal day were, therefore, compared to rhythms measured during a day of total sleep deprivation in 9 normal, young subjects.

Subjects and Methods Subjects

Five women and 4 men, aged 21-36 (mean 26 years), participated in the study. They were carefully screened by physical examination and routing biochemistry. None had history of endocrine, renal, or metabolic diseases or was taking any medications at the time of the study. The protocol was reviewed by the local ethical committee. Each subject gave informed consent before the study.

374 All subjects were studied on two occasions in March with an interval of 1 week. On each occasion, volunteers were admitted to the hospital at 1600 hours. Blood samples were collected every 60 minutes through an indwelling venous catheter until 1800 hours the following day. Normal meals were given at 1730-1800 hours, 08000830 hours, and 1200-1230 hours. Minor meals were offered at 21302200 hours and 1500-1530 hours. Coffee/tea and smoking were unrestricted during the study period. Physical activities were restricted, and the subjects spent most of the daytime sitting in a chair in an upright position. Gentle walking was allowed in the corridor. Daytime naps were not allowed. After randomization, 5 of the subjects were not allowed to sleep at the first occasion, whereas the other 4 had a normal sleep-wake cycle, and vice-versa on the second occasion. On a normal day, subjects slept (and fasted) during blood collection from 2330 hours until 0730 hours. During sleep deprivation, subjects spent the night sitting up and were kept awake under close observation. They were fasting, but fluid intake (coffee and tea) was unrestricted, and smoking was allowed. In order to keep the conditions on the 2 study days as constant as possible during daytime, each subject kept a complete diary of activities, smoking habits, and food and fluid intake on the first day, and this diary was then followed in detail on the second study day.

H. K. Nielsen et al.: Osteocalcin, Sleep, GH, and PTH(1-84) Rhythms 45 ~"

40

~ Z ~

35 30

< o r~ O

't

20 1 5

i

,

16:30

'

'

i

'

20:30

'

'

i

,

00:30

,

,

i

,

04:30

CLOCK

'

'

i

,

08:30

,

,

i

,

12:30

'

'

i

,

16:30

TIME

Fig. 1. The diurnal rhythm in mean serum osteocalcin in 8 normal subjects during a normal sleep-wake cycle ( O . . . O ) and during sleep-deprivation ( O - - - Q ) .

Biochemistry Results The following analyses were done on all samples from each individual except for serum PTH(1-84), which was only measured in samples from 2330 hours until 1130 hours. Serum OC (~.g/liter) was analyzed by a radioimmunoassay (RIA) modified from that described by Price and Nishimoto [27]. Purified bovine OC, generously supplied by Dr. J. Poser (Procter and Gamble Co., Cincinnati, OH), was used for standard and tracer. To each assay tube was added 100 ixliter rabbit antibovine OC serum (final dilution, 1:5,000). After 1 hour, 100 ixliter 1~5I-OC (10,000 cpm in assay buffer) was added. After an additional 48-hour incubation at 4~ the assay was terminated by precipitation of 125I-OC bound to antibody by adding 500 ixliter 25% (wt/vol) polyethylene glycol to each tube. The tubes were centrifuged for 10 minutes at 2,000 g, followed by a second precipitation with 500 txliter 12.5% (wt/vol) polyethylene glycol. The supernatants were decanted, and the tubes were counted in a LKB rack minigamma counter; the sensitivity was 1.4 ng/ml. The intra- and interassay variations were 5% and 10%, respectively. All samples from an individual subject were analyzed in duplicate in the same assay. Serum GH (~mol/liter) was measured in duplicate by RIA using Wick-chromatography [28]. Precision within-assay in the relevant concentration range was better than 0.1 ~g/liter. The interassay variation was 9%. All samples from an individual subject were analyzed in duplicate in the same assay. Plasma-glucose (mmol/liter) was measured in peripheral blood using a standard laboratory method. Serum PTH(1-84)(pg/ml) was determined by using a two-site immunoradiometric assay (Allegro, Nichols Institute, San Juan Capistrano, USA) [29]. The assay detects only intact human PTH. Intraassay variation was 5%. Serum Ca 2+ (mmol/liter) was determined by a calcium ion selective electrode (Radiometer, Copenhagen) using a method with an incorporated pH correction [30].

Statistical Analysis The mean values of each parameter (raw data) for each study day were plotted against time. The effect of time and sleep on the time courses were analyzed by repeated measures analysis of variance (ANOVA) using SPSS (Statistical Package for Social Sciences). Areas under the curves (AUC) were calculated by trapezoidal integration and transformed to mean serum concentrations by division with appropriate time interval and compared by unpaired t-test. P values less than 0.5 were considered significant. Results are given as mean +- SE.

One w o m e n left the study at 0800 hours after sleep deprivation (the first occasion) because she felt discomfort, but she returned a w e e k later to the second occasion, which she completed. The following data represent the results from the 8 subjects w h o c o m p l e t e d the w h o l e study; h o w e v e r , similar results w e r e obtained if available data f r o m all 9 subjects were included in the calculations. On both occasions, serum O C varied in a diurnal rhythm (P < 0.01) characterized by an increase j u s t after bedtime, peak value at 0130 hours, and a decline towards initial levels in the morning. On the control day, serum O C was a little h i g h e r - - f r o m 0330 to 0930 h o u r s - - c o m p a r e d with the sleep deprivation day (Fig. 1). Still, the two rhythms were statistically indistinguishable (P > 0.50). The m e a n serum OC levels, determined by analyses of the A U C s during the time intervals from 1630 to 2330 hours, 2330 to 0730 hours, and 0730 to 1630 hours, showed no significant differences between the 2 days (Table 1). S e r u m G H exhibited significant (P < 0.001) diurnal variation under normal day r h y t h m as well as during acute sleep deprivation (Fig. 2). H o w e v e r , the time pattern on the 2 days differed significantly (P < 0.01). F o r each subject, we recorded the m a x i m u m G H concentration in s e r u m (Cm~) and the corresponding time (Tmax). M e a n Tma x w a s 0136 hours 6 minutes during sleep c o m p a r e d with 0245 hours + 20 minutes (P < 0.01) during sleep deprivation. M e a n Cma• was 10.3 -+ 2.4 izg/liter during sleep c o m p a r e d with a significantly l o w e r maximal level of 7.6 --- 1.2 txg/liter during sleep deprivation (P < 0.01). In Figure 2, a second m a j o r p e a k at 1330 hours is o b s e r v e d after sleep deprivation, but this was due to an unexplained peak (40 ixg/liter) in 1 subject. The mean G H level during 2330-0730 hours was significantly higher during sleep than during sleep deprivation (3.61 -+ 0.6 ixg/liter versus 2.4 --- 0.4 txg/liter (P < 0.004). M e a n serum G H was higher during the interval 0730-1630 hours after sleep deprivation (Table 1), but the difference was not significant (P < 0.30). Plasma glucose showed peaks in relation to meals, but the time pattern and m e a n levels during acute sleep deprivation w e r e not different f r o m the pattern during normal sleep (P > 0.50) (Fig. 2). The nocturnal serum PTH(1-84) profiles (Fig. 3) varied

375

H. K. Nielsen et al.: Osteocalcin, Sleep, GH, and PTH(1-84) Rhythms Table 1. Serum osteocalcin, growth hormone, immunoreactive intact parathyroid hormone PTH(1-84), ionized calcium (Ca2+), and plasma glucose during specific intervals of the day. a

Osteocalcin ~g/liter GH p,g/liter Glucose mmol/liter PTH(I-84) pg/ml Ca2+ mmol/liter

Interval

Sleep

Nonsleep

P

I II III I II III I II III Ib II IIIb Ib II IIIb

23.2 +-1.7 30.9 • 2.5 25.8 • 2.8 1.9 • 0.6 3.6 - 0.6 2.0 • 0.5 5.5 - 0.2 5.1 • 0.1 4.9 • 0.2 29.5 • 4.7 39.1 • 5.5 28.3 +- 3.8 1.20 • 0.01 1.22 • 0.01 1.23 • 0.01

26.8 • 3.1 29.9 - 4.9 24.9 - 3.8 2.1 • 0.6 2.4 • 0.4 1.8 • 0.6 5.4 - 0.2 5.0 • 0.1 5.1 • 0.1 36.2 - 5.9 37.9 • 4.6 25.7 -+ 4.3 1.21 • 0.01 1.20 • 0.01 1.22 • 0.02

ns ns ns ns