Upsala J Med Sci 79: 5 5 - 6 2 , 1974 Nocturnal Secretory Patterns of FSH, GH, LH and TSH P. 0. OSTERMAN, G. WALLIN and L. WIDE Departments of Neurology, Neurophysiology and Clinical Chemistry, University Hospital. Uppsala, Sweden ABSTRACT TSH (17) have been reported, whereas in other The nocturnal secretory patterns of foIIicIe-stimulating investigations (8, 18) such a rhythm has not been hormone (FSH), growth hormone (GH), luteinizing observed. An episodic secretion of FSH and LH hormone (LH) and thyroid-stimulating hormone (TSH) has been found in some studies (15, 2 5 ) , whereas were studied in 1 female and 4 male volunteers. Immuno- in other investigations the plasma have been reactive FSH, GH, LH and TSH in serum were assayed 11-hydroxycorticosteroids in plasma were also measured Of Only One or two hypophyseai hormones have in the male volunteers by a fluorimetric method. been determined at the same time. The Dresent by a radioimmunosorbent technique. Non-conjugated rather constant (20). the piasma levels were taken half from lo pem- to 8 a.m. Polygraphic monitoring of sleep stages was per- formed. In all the subjects a marked elevation of serum GH oecurred within 60 minutes affer the onset of investigation was undertaken to ascertain whether there is any correlation between the nocturnal secretory Patterns Of FSH, GH, LH and TSH or sleep and additional, isolated GH peaks were observed in 3 subjects. The serum concentrations of FSH, LH and TSH also fluctuated, but the variations were small between the secretory patterns of these hormones and different stages of sleep. compared with the fluctuations of GH. The serum levels of FSH, LH and TSH respectively were not cor- related with the depth or stages of sleep. The observed fluctuations, in the serum levels of all hormones assayed, are larger than could be expected from intra-assay errors and probably represent changes in the secretion rate of the hormones. A lack of concordance between the secretory patterns of FSH, GH, LH and TSH in- dicates that different mechanisms may be involved in their release. INTRODUCTION During recent years several studies have been made on the secretory patterns of hypophyseal hormones. The occurrence of a circadian rhythm, and an episodic secretion of ACTH and cortisol are well established (7, 12). The secretory pattern of the growth hormone (GH), which is characterized by sporadic peaks and a marked elevation of plasma G H at the onset of deep sleep, is also well known (9, 27). However, the results from studies on the plasma levels of follicle-stimulating hormone (FSH), luteinizing hormone (LH) and thyroid-stimulating hormone (TSH) are more divergent. A circadian rhythm of the secretion of FSH (9, L H (26) and MATERIALS AND METHODS Four healthy male volunteers, in the age range 19-24, and one 24-year-old, healthy female volunteer were studied. None of them took any medicine, tobacco or alcohol during the study. The female (MO) had a normal menstrual cycle and did not take oral contraceptives. Blood samples were obtained from her 9 and 10 days respectively after the onset of the preceding menstrual period. All the subjects reported a normal sleepwaking cycle. All had normal EEG activity during sleep and wakefulness. Three of the subjects (CB, JB and MO) spent 3 nights, and two (HH and BS) spent 4 con- secutive nights in a quiet room in the neurophysio- logical department. Only a weak light was used in the room from 11 p.m. to 7 a.m. The subjects were fasting after 7 p.m. Electroencephalographic, electrooculographic and electromyographic recordings were made con- tinuously from 10 p m . to 7 a.m. on the last 2 nights. Because of a technical defect, sleep-stage scoring for the female volunteer was possible only for the last night of the study. For the female subject blood sampling took place on the last 2 nights of the study and for the males on the last night. A flexible venous cannula was inserted into an antecubital vein and connected to a 2-foot long plastic tube, so as to permit blood drawing without touching the subject’s arm. In 3 cases the system was filled with heparinized isotonic saline between samples, and in 2 cases the cannula was kept patent with a slow drip of isotonic saline. In each case Upsala J Med Sci 79 56 P . 0. Osterman et al. blood samples were taken half hourly from 10 p.m. to 8 a.m. From the female subject a total volume of approximately 200 ml of blood was removed on each of the 2 nights of the study. Plasma ll-hydroxycorti- costeroids were not determined in the female volunteer in order to avoid removal of more blood. A total volume of approximately 300 ml of blood was removed from each male subject. Blood samples for assay of plasma 1 1-hydroxycorticosteroids were run into heparinized tubes, whereas FSH, GH, L H and TSH were determined in serum. After centrifugation plasma and serum were stored at -20°C until assayed. Plasma 1 1-hydroxycorticosteroids were assayed by a fluorimetric method measuring non-conjugated 1 l-hydro- xycorticosteroids (14). During the period of investigation the intra-assay error estimated from duplicate determina- tions was 4 4 % (coefficient of variation). Immunoreactive FSH, GH, L H and TSH in serum were assayed by a radioimmunosorbent technique (3 1, 32, 33). FSH in serum was measured by utilizing human pituitary FSH (23) labelled with lZ5I and guinea- pig anti-human pituitary FSH. The FSH preparation had a biological activity of 14 OOO IU (2nd IRP-HMG) per mg. The results were expressed in ng per ml. One ng of FSH was equivalent to 369 ng of LER-907 in the immunoassay. GH in serum was measured by utilizing human pituitary G H (24) labelled with lz5I and rabbit anti-human pituitary GH. The results were expressed in ng per' ml using the WHO International Reference Preparations of Growth Hormone, 66/217, as a reference. LH in serum was measured by utilizing human pituitary LH (23) labelled with 1251 and rabbit anti- human pituitary LH. The LH preparation had a biolo- gical activity of 14 000 IU (2nd IRP-HMG) per mg. The results were expressed in ng per ml. One ng of L H was equivalent to 83 ng of LER-907 in the immunoassay. TSH in serum was assayed by utilizing human pituitary TSH (National Pituitary Agency) labelled with lZ5I and rabbit anti-human pituitary TSH. The activity was expressed in p U per ml using Medical Research Council HTSH Research Standard A as a reference. In all radioimmunoassays the standards were dis- solved in sera with undetectable or low levels of the hormones to be assayed. The samples were assayed i n duplicate. The results of the immunoassays were cal- culated by the use of a logit transformation as pro- posed by Rodbard e t al. (22). The percentage standard error of the mean of the two determinations was between 4 and 6%. For recordings of EEG, neck muscle EMG and eye movements Ag-AgC1 surface electrodes were glued to the skin. Two symmetrical EEG channels were re- corded with electrodes in position P3-01 and P4-Oz. For recordings of eye movements one electrode was placed a t the lateral corner of one eye and another electrode above the same eye. The EMG electrodes were placed symmetrically over the muscles on the back of the neck. The 4 signals were amplified in a standard EEG apparatus (Elema, Sweden) and from there fed to a 4 channel FM tape recorder (P1-6200, USA), Upsala J Med Sci 79 where they were stored for later analysis. EEG ana- lysis was made visually and 4 sleep stages were separated. Stage 1 : Light sleep with low voltage background, sleep spindles and K-complexes. No rapid eye movements. Stage 2: Intermediary stage with a fair amount of high voltage slow waves and superimposed spindles. Stage 3: Deep sleep dominated by high voltage, slow delta waves. Rapid eye movernenf ( R E M ) stage: Low voltage, fast activity without spindles and with simulta- neously occurring rapid eye movements. Absence of EMG activity in the neck muscles. RESULTS The results are shown in Figs. 1-3. Onset of sleep occurred between 10 and I 1 p.m. in 2 sub- jects and between midnight and 1 a.m. in 3 sub- jects. Each person slept well for several hours during the night, but the blood sampling procedure occasionally awakened the subjects. In 2 men (CB and BS) REM sleep constituted about 10% and in 2 men and one woman (JB, H H and MO) REM sleep constituted about 15% of total sleep. The serum or plasma concentrations of the 5 hormones assayed were, throughout, within the normal range as determined in this laboratory in fertile men and women. In the 4 male volunteers, in whom plasma I l - hydroxycorticosteroids were determined, maximum levels occurred during the early morning hours. The pattern of plasma 1 1-hydroxycorticosteroid levels was in all the subjects Characterized by sharp increases followed by falls in the plasma level. The lowest levels occurred between 2 and 4 a.m. It can be seen on the graphs that in all the subjects the serum concentrations of FSH, G H , L H and TSH fluctuated. Each point on the graphs represents the mean of duplicate determinations. The variation between the points in each curve is caused by the error of the mean of the duplicates (intra-sample variation) and a possible true varia- tion between the samples (inter-sample variation). The variance of the results obtained during the period of investigation was calculated for each curve. The difference between this value and the variance of the mean of the duplicate determinations (intra-sample variation) gave an estimate of the inter-sample variance. The median and range of the inter-sample variation, expressed as percentage standard deviation was 12 (4.2-17.3) for FSH, 22.7 (16.4-30.7) for L H , 15.3 (10.0-22.3) for TSH and 95.5 (71.8-130) for GH. Thus, an Nocturnal secretory patterns of F S H , G H , LH and TSH 57 REM - 1 . 2 - 3 - Awake - REM . 1 . 2 . 3 - c.a. d J. 8. d” 2 0 . 10. 5 . A . 3 . 3 . 2 - 2 - , . L H nglml - 3 1 1.0 0.5 O.S1 LH n g h l - 1StIpUlml-t GH nglml - 30 ’ 20 . 10 - 5 - L - 3 - 2 - 1.0- TSH uUlml JWI GH nglml - 20 10 - 5 - A - 3 - 2 - 58 P . 0. Osterman et al. 2.0 - 1.0 - 0.5 - 0.3 - H.H. d 1.0 - 0.5 - 0.3 - B.S. d 2 - 1.0 - 0.5 - Awake 1 2 3 E H pU/ml YLN GH nglml -lo 0.5 J Awoke REM * 1 - 2 - 3 - TSH pU/ml Y G H nglml - 10 5 I 3 w I 1 - Hydroxy - cwticosterolds 22 24 02 04 06 08 Time of day 22 2.4 02 04 06 08 Time of day Fig. 2. Patterns of serum FSH, LH, TSH, GH and lationship to sleep stages in male subjects HH and BS. plasma 11-hydroxycorticosteroid levels and their re- Blood samples taken half hourly from 10 p.m. to 8 a.m. Vpsala J Med Sci 79 Nocturnal secretory patterns of F S H , G H , L H and FSH 59 U.O. 0 R E M - 1 - 2 - 3 - U.O. Q 1.0. 0.5 J 22 2L 02 OL 06 08 TSH uUlml )-* GH ng'ml -Lo 1 2.0. J 1.0 1 1/22 0.5 2 2 2 4 02 04 06 08 Time of day Fig. 3. Patterns of serum FSH, LH, TSH and GH levels in the female subject. Blood samples taken half hourly two consecutive nights, 9 and 10 days respectively evident inter-sample variation was found for all four pituitary hormones. In all the subjects a significant elevation of serum G H concentration occurred within 60 minutes after the onset of sleep. A peak in serum G H concentration preceded the G H peaks associ- ated with the onset of sleep in 2 subjects (HH and BS). In the female subject a second GH peak at 3 a.m. followed the peak which appeared at the onset of sleep. The variation in the level of L H was greater than those of the FSH and TSH levels. Serum levels of FSH, L H and TSH were not correlated (p>0.05) with the depth or stages of sleep. The secretory patterns of FSH, L H and TSH were not constant ,between nights, in the female subject FSH ng/ml n-x LH ng/ml - 2.0 1 0.5 - 22 2 4 02 OL 06 08 0.5 I 22 21 02 04 06 08 Time of day after the onset of the preceding menstrual period. Sleep stages recorded the last night. studied. Furthermore no relationship was observed between the secretory patterns of FSH, G H , LH and TSH in any of the 5 subjects. DISCUSSION Several studies have shown that cortisol is secreted in a series of brief episodes separated by intervals during which there is either no, or only low, secretion. In men and women with a >>normal>> sleep-waking cycle most of the secretory bursts occur in the early morning hours, with maximum plasma cortisol levels between 4 and 8 a.m. (7, 12, 19, 30). In the present investigation the pattern of plasma 1 1-hydroxycorticosteroid levels was, in all 4 men in accordance with that de- Upsala J Med Sci 79 60 P . 0. Osterman el al. scribed previously for plasma cortisol. Thus, all the men had a .normal. pattern of plasma 11- hydroxycorticosteroid levels in spite of a certain amount of stress caused by frequent blood sampling. Krieger et al. (12) found a very good correlation between the times of peaks of plasma corticosteroid levels and those of plasma ACTH levels. The time at which the ACTH peak occurred coincided in most cases with that of the plasma 11-hydroxy- corticosteroidpeak. However, there was no apparent proportionality between plasma corticosteroid and plasma ACTH levels. In all the subjects a marked elevation of serum G H was observed at the onset of sleep. In 3 sub- jects, 2 male and one female a second G H peak was observed. The peak at 10.30 p.m. in subject H H may be due to the stress of the initial vene- punction, that at midnight in subject BS is pro- bably part of the nocturnal peak. The observed secretory pattern of G H is similar to that previously reported (9, 10, 21, 27). Evidently, the peak of serum G H at the onset of sleep is very constant, whereas additional nocturnal peaks are often missing. Dierschke et al. (4) found, in ovariectomized rhesus monkeys, wide rhythmic oscillations in the plasma L H levels. In intact female monkeys no such fluctuations were observed. Similarly, in castrated male and female rats fairly regularly occurring peaks of LH were found, whereas the serum FSH levels were more stable (6). Yen et al. (34) reported, in women of fertile age, fluc- tuations in the serum level of L H which were rather regular in frequency and magnitude. However, these parameters varied according to the stage of the menstrual cycle. In the present study the fluctuations of FSH and L H occurred irregularly, and were of varying magnitude. This finding is in accordance with observations reported in recent investigations in adult men (2, 13, 16, 25) and in one investigation on women of fertile age (15). The fluctuations found for FSH were smaller than for L H which agrees with previous studies (15, 32, 34). There was no correlation between the levels of FSH and L H , and the depth or stages of sleep. Rubin et al. (25) observed that in adult men the L H values were 14% higher during REM sleep than during other stages of sleep; but release of FSH was not clearly related to the stages of sleep. On the other hand, Boyar et al. (1) did not find, in adult men and prepubertal children, any signifi- cant difference between mean LH concentration, regardless of whether the subjects were asleep or awake; and no relation between LH secretory activity and the stages of sleep was observed. In pubertal children, however, the mean, sleep L H concentration was significantly higher than the mean, awake LH concentration ( I ) . Vanhaelst et al. (28) observed, in 7 of 8 euthy- roid subjects, a peak of TSH between 4 and 6 a.m. Moreover, all the subjects exhibited a rhythm of higher frequency, lasting from 1 to 2 hours. The findings of these authors are at variance with those of Webster et al. (29), who demonstrated, in normal subjects and in patients with myxedema, occasional significant variations in serum TSH concentration, but no convincing pattern of cir- cadian rhythm or other mode of variation of TSH secretion. The present study supports the results of Webster et al. (29). Fluctuations of the serum concentrations of TSH were observed, but there was no evidence of a regular periodicity. The observed fluctuations, in the serum levels of all hormones assayed, are larger than could be expected from intra-assay errors. The fluctuations of the radioimmunoassayed hormones, FSH, GH, L H , TSH, are not synchronized. Therefore it is unlikely that the fluctuations are due to variations of substances in serum, causing nonspecific inter- ference with the assay technique. Kohler et al. ( l l ) , and Coble et al. (3) have shown that the metabolic clearance rates of L H and FSH are rather constant, and are not influenced by gonadal function. It is likely that the observed fluctuations in the serum levels of FSH, G H , L H and TSH probably represent changes in the secretion rate of the hormones. Takahashi et al. (27) did not find any correlation between the secretory patterns of cortisol and GH. Rubin et al. (25) observed no relation between the peaks of FSH and L H in adult men; nor did Krieger et al. (13) detect any relation between the patterns of FSH, G H and L H in adult men. Yen et al. (34) failed t o observe any relation between the secretory patterns of FSH and L H in women of fertile age, but in postmenopausal women the fluctuations of FSH and L H usually were coincidental with minor asynchrony. In the study by Midgley & Jaffe (15), in women of fertile age, the changes in the FSH level often paralleled those of L H . In this study, no correlation was found between Upsala J Med Sci 79 Nocturnal secretory p a t t e r n s of F S H , G H , L H and F S H 61 the nocturnal secretory patterns of FSH, G H , L H a d TSH. The mechanisms, regulating the secretion of the different hypophyseal hormones, are still un- known. The lack of concordance between the secretory patterns of FSH, G H , L H and TSH indicates that different mechanisms may be in- volved in their release. It is not known whether the observed fluctuations in the trophic hormone levels are necessary for the maintenance of a normal function of the target organs. Further similar studies on patients with hypothalamo-hypo- physeal diseases may prove valuable for under- standing the importance of the observed secretory patterns. ACKNOWLEDGEMENTS We are indebted to Dr Paul Roos at the Institute of Biochemistry, Uppsala for supplying the highly purified LH, FSH and G H preparations, and to Mrs Anna-Lena Barmark, Mr Christer Bengtsson and Miss Vera Star- felt for skilful technical assistance. This study was supported by the Swedish Medical Research Council (grant No. 13X-3145) and by a grant from the Medical Faculty of the University of Uppsala. REFERENCES 1. Boyar, R., Finkelstein, J., Roffwarg, H., Kapen, S., Weitzman, E . & Hellman, L.: Synchronization of augmented luteinizing hormone secretion with sleep during puberty. New Eng J Med 287: 582, 1972. 2. Boyar, R., Perlow, M., Hellman, L., Kapen, S. & Weitzman, E.: Twenty-four hour pattern of luteinizing hormone secretion in normal men with sleep stage recording. J Clin Endocr35: 73, 1972. 3. Coble, Y . D., Kohler, P. O., Cargille, C. M. & Ross, G. T.: Production rates and metabolic clearance rates of human follicle-stimulating hormone in premenopausal and postmenopausal women. J Clin Invest 48: 359, 1969. 4. Dierschke, D. J., Bhattacharya, A. N., Atkinson, L. E. & Knobil, E.: Circhoral oscillations of plasma LH levels in the ovariectomized rhesus monkey. Endocrinology 87: 850, 1970. 5. Faiman, C. & Winter, J. S. D.: Diurnal cycles in plasma FSH, testosterone and cortisol in men. J Clin Endocr 33: 186, 197 1. 6 . Gay, V . L. & Sheth, N. A.: Evidence for a periodic release of L H in castrated male and female rats. Endocrinology 90: 158, 1972. 7. Hellman, L., Nakada, F., Curti, J., Weitzman, E. D., Kream, J., Roffwarg, H., Ellman, S., Fukushima, D. K. & Gallagher, T. F.: Cortisol is secreted episodically by normal man. J Clin Endocr 30: 411, 1970. 8. Hershman, J . M. & Pittman, J . A.: Utility of the radioimmunoassay of serum thyrotrophin in man. Ann Intern Med 74: 481, 1971. 9. Honda, Y., Takahashi, K., Takahashi, S., Azumi, K., h i e , M., Sakuma, M., Tsushima, T. & Shizume, K.: Growth hormone secretion during nocturnal sleep in normal subjects. J Clin Endocr 29: 20, 1969. 0. Hunter, W . M., Rigal, W. M. & Sukkar, M. Y.: Plasma growth hormone during fasting. In: Proc. Int. Symp. Growth Hormone, Milan, Italy, p. 408. Excerpta Medica, Amsterdam, 1968. 1. Kohler, P. O., Ross, G. T. & Odell, W. D.: Meta- bolic clearance and production rates of human luteinizing hormone in pre- and postmenopausal women. J Clin Invest 47: 38, 1968. 12. Krieger, D. T., Allen, W., Rizzo, F. & Krieger, H. P.: Characterization of the normal temporal pattern of plasma corticosteroid levels. J Clin Endocr 32: 266, 1971. 13. Krieger, D. T . , Ossowski, R., Fogel, M. & Allen, W.: Lack of circadian periodicity of human serum FSH and L H levels. J Clin Endocr 35: 619, 1972. 14. Mattingly, D.: A simple fluorimetric method for the estimation of free 1 I-hydroxycorticoids in human plasma. J Clin Path 15: 374, 1962. 15. Midgley, A. R. & Jaffe, R. B.: Regulation of human gonadotropins: X. Episodic fluctuation of L H during the menstrual cycle. J Clin Endocr 33: 962, 1971. 16. Nankin, H. R. & Troen, P.: Repetitive luteinizing hormone elevations in serum of normal men. J Clin Endocr33: 558, 1971. 17. Nicoloff, J . T., Fisher, D. A. & Appleman, M. D.: The role of glucocorticoids in the regulation of thyroid function in man. J Clin Invest 49: 1922, 1970. 18. Nillius, S. J. & Wide, L.: Effects of oestrogen on serum levels of LH and FSH. Acta Endocr (Koben- havn)65: 583, 1970. 19. Otth, D. N., Island, D. P. & Liddle, G. W.: Experi- mental alteration of the circadian rhythm in plasma cortisol (17-OHCS) concentration in man. J Clin Endocr 27: 549, 1967. 20. Peterson, N. T . , Midgley, A. R. & Jaffe, R. B.: Regulation of human gonadotropins. 111. Luteinizing hormone and follicle stimulating hormone in sera from adult males. J Clin Endocr 28: 1473, 1%8. 21. Quabbe, H-J., Schilling, E. & Helge, H.: Pattern of growth hormone secretion during a 24-hour fast in normal adults. J Clin Endocr 26: 1173, 1966. 22. Rodbard, D., Rayford, P. L., Cooper, J. A. & Ross, G. T.: Statistical quality control of radioimmuno- assays. J Clin Endocr 28: 1412, 1968. 23. Roos, P.: Human follicle-stimulating hormone. Acta Endocr (Kobenhavn) Suppl. 131: 1, 1968. 24. Roos, P., Fevold, H. R. & Gemzell, C. A.: Pre- paration of human growth hormone by gel filtration. Biochim Biophys Acta 74: 525, 1963. 25. Rubin, R. T., Kales, A., Adler, R., Fagan, T. & Odell, W.: Gonadotropin secretion during sleep in normal adult men. Science 175: 196, 1972. 26. Saxena, B. B., Leyendecker, G., Chen, W., Gandy, Vpsulu J Med Sci 79 62 P. 0. Osterman et al. H. M. & Peterson, R. E.: Radioimmunoassay of follicle-stimulating (FSH) and luteinizing (LH) hor- mones by chromatoelectrophoresis. Acta Endocr (Kobenhavn) Suppl. 142: 185, 1969. 27. Takahashi, Y., Kipnis, D. M. & Daughaday, W. H.: Growth hormone secretion during sleep. J Clin Invest 47: 2079, 1968. 28. Vanhaelst, L., Van Cauter, E., Degaute, J. P. & Golstein, J.: Circadian variations of serum thyro- tropin levels in man. J Clin Endocr 35: 479, 1972. 29. Webster, B . R . , Guansing, A. R. & Paice, J . C.: Absence of diurnal variation of serum TSH. J Clin Endocr 34: 899, 1972. 30. Weitzman, E. D . , Fukushima, D., Nogeire, C., Roffwarg, H., Gallagher, T. F. & Hellman, L.: Twenty-four hour pattern of the episodic sercretion of cortisol in normal subjects. J Clin Endocr 33: 14, 1971. 31. Wide, L.: Radioimmunoassays employing immuno- sorbents. Acta Endocr (Kobenhavn) Suppl. 142: 207, 1%9. 32. Wide, L., Nillius, S. J . , Gemzell, C. & Roos, P.: Radioimmunosorbent assay of follicle-stimulating hormone and luteinizing hormone in serum and urine from men and women. Acta Endocr (Koben- havn) Suppl. 174: 1 , 1973. 33. Wide, L. & Porath, J.: Radioimmunoassay of pro- teins with the use of sephadex-coupled antibodies. Biochim Biophys Acta 130: 257, 1966. 34. Yen, S. S. C., Tsai, C. C., Naftolin, F., Vanden- berg, G. & Ajabor, L.: Pulsatile patterns of gonado- tropin release in subjects with and without ovarian function. J Clin Endocr34: 671, 1972. Received June 8,1973 Address for reprints: Per Olof Osterman, M.D. Department of Neurology University Hospital S-750 14 Uppsala 14 Sweden Upsala J Med Sci 79