0021-972X/92/7503-0661$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright C 1992 by The Endocrine Society
A Circadian Rhythm Hormone in Women* J. F. MORTOLAt,
G. A. LAUGHLIN,
Vol. 75, No. 3 Printed in U.S.A.
of Serum AND
S.
S.
Department of Reproductiue Medicine (0802) and the General California-San Diego, La Jolla, California 92093-0802
YENJ:
C.
Clinical
ABSTRACT While a nocturnal decline in serum LH levels in the early follicular phase of the menstrual cycle has been well established, a diurnal variation in serum FSH levels in women has not been demonstrated. We addressed this issue by determining serum LH and FSH levels at 15-min intervals for 24 h in the early follicular phase (EFP; n = 16) and late follicular phase (LFP; n = 10) of the menstrual cycle and in postmenopausal women (PMW; n = 10). Serum estradiol was simultaneously measured at hourly intervals. As expected, EFP, but not LFP and PMW, women had a 15% nocturnal decline (P < 0.01) in transverse mean LH levels compared to values in the daytime hours. In contrast, nocturnal FSH transverse mean values were significantly lower than daytime values in all groups studied, demonstrating an 18% decline in EFP (P < O.OOl), a 17% decline in LFP (P < O.OOOOl), and a 4.3% decline in PMW (P < 0.01).
I
N 1973, Kapen et al. (1) reported a nocturnal decline of serum LH restricted to the early follicular phase of the menstrual cycle. This observation was subsequently confirmed and demonstrated to be a result of sleep-dependent slowing of the LH pulse frequency (Z-6). Attempts to discern a diurnal rhythm in circulating FSH levels in humans have yielded conflicting results (7-13). For example, in men using frequent sampling, diurnal variations of both LH and FSH were found in one study (12), but not in another (13). To date, circadian variations in serum FSH levels in women have not been defined. We report here the finding of a remarkable circadian rhythm of serum FSH measuredin 24h samples at 15-min intervals in women during the early follicular phase (EFP) and the late follicular phase (LFP) of the menstrual cycle and in postmenopausal women (PMW). Subjects
Follicle-Stimulating
and Methods
Twenty-six normal cycling women between the ages of 20-41 yr and 10 PMW between the ages of 51-61 yr completed the study. PMW had been amenorrheic for at least 1 yr and were free from hormonal Received September 23, 1991. Address all correspondence and requests for reprints to: Dr. Samuel 5. C. Yen, Department of Reproductive Medicine (OSOZ), University of California-San Diego, La Jolla, California 92093-0802. * This work was supported by NIH NICHD Center Grant HD-1230313, a grant from the General Clinical Research Branch, Division of Research Resources (NIH MOl-RR-00827), and in part by the Mellon Foundation. This research was conducted in part by the Clayton Foundation for Research. t Current address: Beth Israel Hospital, Boston, Massachusetts 02215. $ Clayton Foundation investigator.
Research Center, University
of
Cosinor analysis revealed a circadian rhythm for FSH, with acrophases in the afternoon and nadirs at night in all three groups. The circadian amplitudes were 1.43 + 0.22, 1.02 + 0.16, and 8.42 + 1.31 IU/ L for EFP, LFP, and PMW, respectively. The EFP nocturnal decline in LH did not conform to a cosine rhythm. A diurnal variation in estradiol was not present in any of the groups of women. These data constitute the first demonstration of a robust circadian rhythm of serum FSH in women. The comparable timing of the acrophase in all groups of subjects and its presence in the postmenopausal years suggest a central, rather than peripheral, feedback mechanism(s) for the circadian rhythmicity. This observation provides strong evidence for a dissociation in the hypothalamic regulation of pituitary LH and FSH secretion in women. The circadian peak and nadir of circulating FSH may prove to be determining for appropriate follicular development. (J Clin Endocrinol Metab 75: 861-864, 1992)
replacements. Before enrollment, all premenopausal women reported regular menstrual cycles, ranging from 26-32 days in length. Subjects were without a history of medical illness, and physical examination was normal before inclusion in the study. This protocol was approved by the Committee of Investigations Involving Human Subjects at the University of California-San Diego. Studies were conducted in the Clinical Research Center (CRC). Sixteen EFP women (defined as cycle days 3-5) and 10 LFP women (defined as cycle days 10-12) participated in this investigation. Phases of the cycle were later confirmed by serum estradiol (E,) measurements. PMW were studied on an arbitrary day. Subjects were admitted to the CRC 2 h before the start of the investigation, during which time an iv catheter was placed in the forearm of the nondominant hand. Blood samples were obtained at 15-min intervals for LH and FSH determinations and at hourly intervals for El measurements. A sleep room that allowed blood drawing through a catheter extending to an adjacent room was used, so as not to disturb the subject during sleep. A standardized activity schedule was maintained for the 24-h study period, with meal times at 0800, 1200, and 1800 h. Ambulation and daytime napping were avoided. Lights were off between 2300-0700 h, and the sleep pattern of each subject was monitored by an experienced nurse through a video camera. LH and FSH values were determined in duplicate by RIA (14, 15). The sensitivities of the assays were 0.3 and 1.0 II-I/L for LH and FSH, respectively. The intraassay coefficients of variation (CVs) were 6% and 7%, and the interassay CVs were 10% and 11% for LH and FSH, respectively. El was determined by a previously described RIA (16). The intra- and interassay CVs for this steroid were 8% and 12%, respectively, and the sensitivity of the assay was 33 pmol/L. All samples from the same subject were included in the same assay run. Transverse mean levels of LH and FSH for each subject were determined for daytime hours (0700-2300 h) and nighttime hours (23000700 h) by calculating the arithmetic mean of all values obtained during the respective time intervals. The magnitude of the nocturnal changes in LH and FSH were defined as the difference between the calculated daytime and nighttime transverse mean levels for each subject. Significant differences between these values within each group were compared
861
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
862
MORTOLA,
LAUGHLIN.
by paired t tests. Cosine analysis of the concentration series of LH, FSH, and E2 was performed on each of the 15-min (60 min for E,) interval data sets for each individual by fitting the general cosine function, CS(t) = M + A cos(w + I#J), where CS(t) is the hormone concentration at time t, M is the mesor (midline value), A is the absolute amplitude, and w and 4 are the angular frequency and acrophase, respectively (17). The acrophase and nadir are defined as the times of occurrence of the maxima and minima in the cosine curve, respectively. Because individuals may display a significant cosine rhythm without a significant rhythm for the group as a whole (usually due to a disparity of acrophase times), individual cosinor analysis was followed by a population mean cosinor determination for each group (EFP, LFP, and PMW) (17). Cosine analysis was also performed on mean percent change from the transverse mean data for each group for the purpose of displaying the group cosine rhythm (see Fig. 2). Parameters of individual cosinor analysis were compared between groups by analysis of variance with Newman-Keuls post-hoc testing. For all statistical analyses, P < 0.05 was considered significant.
Results In EFP women, a nighttime decline in LH was observed in 13 of the 16 subjects studied. Mean nocturnal LH levels fell by 15% (P < 0.01) from a daytime level of 12.7 f 1.2 (+ SE) to 10.9 + 1.1 lU/L (Fig. 1). However, group cosinor analysis did not reveal a significant circadian rhythm for LH in EFP women. In LFP and PMW women, nighttime and daytime transverse mean LH levels were similar (12.8 + 1.5 VS. 12.9 f 1.6 lU/L and 81.7 f 10.5 us. 77.0 + 9.0 lU/L), and no significant group cosine rhythm was detected (Fig. 1). Transverse mean serum FSH levels in EFP declined by iS% (P C 0.001) compared to daytime values (11.0 + 0.7 VS. 9.2 + 0.4 lU/L). A similar 17% decline in FSH occurred in LFP (P < 0.00001) where values were reduced from a dayEARLY
30r
AND
YEN
JCE & M. 1992 Vol75.No3
time transverse mean of 7.2 + 0.8 lU/L to a nighttime value of 5.9 + 0.7 lU/L. Significantly lower transverse mean FSH levels were also seen at night compared to those during the day in PMW (P C 0.01). Although the absolute decline in FSH levels at night was larger in PMW (155.8 f 14.0 to 148.9 + 14.4 lU/L) than in cycling women, the relative change was reduced to 4.3%. Significant cosine rhythms for FSH were detected in 15 of 16 EFP women, 9 of 10 LFP women, and 9 of 10 PMW. These contributed to significant (P c 0.01) group circadian rhythms for all groups tested (Fig. 2). Arithmetic mean circadian rhythm results for FSH for each group are presented in Table 1. A similar time course of FSH concentrations was observed, with early afternoon acrophases (Table 1) and nadirs at night (EFP, 0304 h f 53 min; LFP, 0242 h k 41 min; PMW, 0145 h + 37 min; P = NS). When expressedin serum concentration units, the amplitudes of the circadian FSH variation were similar in both stages of the follicular phase (EFP, 1.43 + 0.22 lU/L; LFP, 1.02 + 0.16 lU/L) and elevated in PMW [8.42 f 1.31 lU/L; P < 0.001 us. EFP and EARLY
20 r
FOLLICULAR
I
10 0
f Y
-10
z 9
-20
I t i
2
20
f E
10 Ib I
LATE
FOLLICUL
4
5 E %
-10
1
-20
L
FOLLICULAR f 3
II
20 5 w
?v)
LATE FOLLICULAR
3o r
POST-MENOPAUSAL
.,-.
-20
1 0900
1700 CLOCK
3 t
TABLE women’s in EFP,
1. Mean (&SE) results of cosinor analysis of individual FSH values obtained on serum samples at 15-min intervals LFP, and PMW
-0900
1700 CLOCK
0100
0900
FIG. 2. Mean (&SE) percent change from 24-h mean in serum FSH values obtained at 15-min intervals for 24 h in EFP (n = 16), LFP (n = lo), and PMW (n = 10). The significant circadian variation of the hormone in each group of women is represented by the best-tit cosine curve.
POST-MENOPAUSAL
3o 1
0100 HOURS
0900
HOURS
FIG. 1. Mean (&SE) percent change from 24-h mean in serum LH levels obtained at 15-min intervals for 24 h in EFP (n = 16), LFP (n = lo), and PMW (n = lo), demonstrating a nocturnal decline in EFP, not present in LFP or PMW.
EFP (n = 16) LFP (n = 10) PMW (n = 10) “P < 0.001 * P < 0.001
Absolute amplitude W/L)
% amplitude m)
W/L)
Acropbase (clock hours)
1.43 + 0.22 1.02 f 0.16 8.42 f 1.31*
13.7 + 1.8 15.5 + 1.6 5.3 + 0.7*
10.31 f 0.58 6.48 + 0.79” 157.1 + 13.0*
15:04 + 0:53 14:42 + 0:41 13:45 + 0:37
us. EFP us. EFP
MSSCX
and PMW. and LFP.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
CIRCADIAN
RHYTHM
LFP). However, when expressed as a percentage of the 24-h mean, the relative amplitude of the circadian rhythm in PMW was markedly attenuated compared to that in follicular phase women (EFP, 13.7 + 1.8%; LFP, 15.5 f 1.6%; PMW, 5.3 f 0.7%; P < 0.001 us. EFP and LFP; Table 1 and Fig. 2). Hourly serum EZ levels were relatively stable throughout the 24-h sampling period within each group studied (Fig. 3). The transverse mean values were 127 + 8, 355 + 54, and 46 + 3 pmol/L for EFP, LFP, and PMW, respectively. There were no differences in daytime and nighttime values, and no significant group circadian rhythms were detected. Discussion This study constitutes the first demonstration of a circadian rhythm of circulating FSH in women. This circadian periodicity of FSH was observed in women of reproductive age during both the EFP and LFP of the menstrual cycle and in PMW. The results of our study reveal a marked disparity between the 24h profiles of serum FSH and LH levels. While the magnitudes of the EFP nocturnal declines in FSH and LH were similar (18% and 15%, respectively), the diurnal variation in LH was limited to this cycle phase and was without discernible circadian rhythmicity, as previously reported (4, 5). As demonstrated by sleep reversal (12) and naloxone infusion (6) studies, the nocturnal LH decline in EFP women is a sleep-entrained and opioidergic-mediated event, In contrast, the 24-h serum FSH profile appears to reflect a circadian rhythmicity, which is evident throughout the follicular phase of the cycle and in PMW. The lower relative (percentage), but not absolute (concentration), of the FSH circadian amplitude observed in PMW may be related to the hypersecretion of FSH in PMW and the long half-life of FSH (15), which could combine to mask the circadian amplitude. Alternatively, it may reflect attenuation of the central mechanism(s) controlling the circadian timekeeping system due to aging (18). The neuroendocrine mechanism(s) to account for the presence of a circadian rhythm of FSH, but not LH, is unclear. While it is well known that E2 exerts a preferential inhibition of FSH secretion (19, 20), the magnitude of the circadian excursion of FSH around the mean was similar in EFP and LFP despite a 3-fold increase in Ez levels. Moreover, since none of the groups of women in this study demonstrated significant diurnal variations in E2, the circadian rhythm of
i
400 300
OF SERUM
Acknowledgments We are grateful to Ms. Shannon Petze and Mr. Jeff Wong for their technical assistance, and to Ms. Del Alsobrook for preparation of the manuscript. Our thanks to Dr. Allen Lein for helpful advice in data analysis, and the CRC nursing staff for providing excellent care.
References 1. Kapen
2.
3. LFP
w” 5. 1700 CLOCK
0100 HOURS
0900
FIG. 3. Mean (SE) serum EZ levels determined at hourly the EFP (n = 16), LFP (n = lo), and PMW (n = 10).
intervals
in
863
FSH was not coupled with E2 fluctuations. The presence of a circadian rhythm for FSH in PMW suggests that ovarian factors, including inhibin, activin, and follistatin (22-24), may not play a significant role in the expression of the circadian rhythmicity of FSH secretion. Although unlikely, intrapituitary inhibin, activin and follistatin, acting as autocrine regulators of FSH gene expression and secretion (2528), may function independent of GnRH in generating the circadian rhythm of FSH. The similar timings of the acrophase of this rhythm in younger cycling women and older postmenopausal women suggest that the rhythm is mediated by a central mechanism(s). Neuroendocrine factors, such as sleep and a hypothalamic FSH-releasing factor distinct from GnRH (29) that is linked to the circadian pacemaker, have to be considered. The physiological significance of the FSH circadian rhythm described here is unclear. FSH should now be added to the number of pituitary hormones with circadian rhythmicity. However, the nocturnal nadir of FSH is opposite the nocturnal rise of ACTH, TSH, GH, and PRL (13, 18). In terms of target cell response, the presence of a circadian peak and nadir of FSH may be an important determining factor in the regulation of folliculogenesis and maturation; this is a speculation worthy of investigation. In summary, we have observed a robust circadian rhythm for circulating FSH levels in women with or without ovarian function. This finding provides additional evidence for the dissociation in the neuroendocrine regulation of pituitary release of FSH and LH. Further studies are needed to define the neuroendocrine mechanism(s) governing the circadian rhythmicity of pituitary FSH secretion and its aberrations in disorders of folliculogenesis.
4.
0900
FSH
6.
S, Boyar R, Perlow M, Hellman L, Weitzman ED. 1973 Luteinizing hormone: changes in secretory patterns during sleep in adult women. Life Sci. 13:693-g. Kapen S, Boyar R, Hellman L, Weitzman ED. 1976 The relationship of luteinizing hormone to sleep in women during the early follicular phase: effects of sleep reversal and a prolonged 3 hour sleep-wake schedule. J Clin Endocrinol Metab. 43:1031-7. Soules M, Steiner R, Cohen N, Bremner W, Clifton D. 1985 Nocturnal slowing of pulsatile luteinizing hormone secretion in women during the follicular phase of the menstrual cycle. J Clin Endocrinol Metab. 61:43-9. Filicori M, Santoto N, Merrium G, Crowley WJ. 1986 Characterization of the physiologic pattern of episodic gonadotropin secretion throughout the menstrual cycle. J Clin Endocrinol Metab. 62:113644. Rossmanith WC, Liu CH, Laughlin GA, Mortola JF, Suh BY, Yen SSC. 1990 Relative changes in LH pulsatility during the menstrual cycle: using hypogonadal women as a reference. Clin Endocrinol (Oxf). 32:647-60. Rossmanith WG, Yen SSC. 1987 Sleep associated decrease in luteinizing hormone pulse frequency during the early follicular
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
864
7. 8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
MORTOLA.
LAUGHLIN,
phase: evidence for an opioidergic mechanism. J Clin Endocrinol Metab. 65:715-g. Faiman C, Ryan RJ. 1967 Diurnal cycle in serum concentrations of follicle-stimulating hormone in men. Nature. 215:857. Saxena BB, Demura H, Gandy HM, Peterson RE. 1968 Radioimmunoassay of human follicle stimulating and luteinizing hormones in plasma. J Clin Endocrinol Metab. 28:519-34. Peterson Jr NT, Midgley AR, Jaffe RB. 1968 Regulation of gonadotropins. III. Luteinizing and follicle stimulating hormone in sera from adult males. J Clin Endocrinol Metab. 28:1473-g. Krieger DT, Ossowski R, Fogel M, Allen W. 1972 Lack of circadian periodicity of human serum FSH and LH levels. J Clin Endocrinol Metab. 35:619-23. Alford FP, Baker HWG, Pate1 YC, et al. 1973 Temporal patterns of circulating hormones as assessed by continuous blood sampling. L J Clin Endo&inol Metab. 36:108-12. . Veldhuis ID. Kine IC. Urban RI. et al. 1987 Oueratinz ” characteristics of the male hypothalamo-pituitary-gonadal axis: pulsatile release of testosterone and follicle stimulating hormone and their temporal coupling with luteinizing hormone. J Clin Endocrinol Metab. 65:929-41. Veldhuis JD, Iranmanesh A, Johnson ML, Lizarralde G. 1990 Twenty-four hour rhythms in plasma concentrations of adrenohypophyseal hormones are generated by distinct amplitude and/or frequency modulation of underlying pituitary secretory bursts. J Clin Endocrinol Metab. 71:1616-23. Yen SSC, Llerena 0, Little B, Pearson OH. 1968 Disappearance rates of endogenous luteinizing hormone and chorionic gonadotropin in man. J Clin Endocrinol Metab. 28:1763-7. Yen SSC, Llerena LA, Pearson OH, Littell AS. 1970 Disappearance rates of endogenous follicle-stimulating hormone in serum following surgical hypophysectomy in man. J Clin Endocrinol Metab. 30:3259. De Vane GW, Czekala NM, Judd HL, Yen SSC. 1975 Circulating gonadotropins, androgens, and estrogens in polycystic ovarian disease. Am J Obstet Gynecol. 121:496-501. Nelson W, Tong YL, Lee JK, Halberg F. 1979 Methods for cosinor ,.
“,.
,.
I
AND
YEN
JCE & M. 1992 Vol75.No3
rhythmometry. Chronogiologia. 6:305-9. 18. Van Coevorden A, Mockel J, Laurent E, et al. 1991 Neuroendocrine rhythms and sleep in aging men. Am J Physiol. 260:E851-61. 19. Yen SSC, Tsai CC, Vandenberg G, Rebar R. 1972 Gonadotropin dynamics in patients with gonadal dysgenesis: a model for the study of gonadotropin regulation. J Clin Endocrinol Metab. 35:897-904. 20. Yen SSC, Lasley BL, Wang CF, Leblanc H, Siler TM. 1975 The operating characteristics of the hypothalamic-pituitary system during the menstrual cycle and observations of biological action of somatostatin. Recent Prog Horm Res. 31:321-63. 21. McLachlan RI, Robertson DM, Healy DL, Burger HG, de Kretser DM. 1987 Circulating immunoreactive inhibin levels during the normal menstrual cycle. J Clin Endocrinol Metab. 65:954-61. 22. McLachlan RI, Robertson DM, De Kretser DM, Burger HG. 1988 Advances in the physiology of inhibin and inhibin-related peptides. Clin Endocrinol (Oxf). 29:77-114. 23. Ueno N, Ling N, Ying S-Y, Esch F, Shimasaki S, Guillemin R. 1987 Isolation and partial characterization of follistatin: a novel MW 35,000 monomeric protein that inhibits the release of follicle stimulating hormone. Proc Nat1 Acad Sci USA. 84:8282-6. 24. Roseff SJ, Bangah ML, Kettel LM, et al. 1989 Dynamic changes in circulating inhibin levels during the luteal-follicular transition of the human menstrual cycle. J Clin Endocrinol Metab. 69:1033-9. 25. Roberts V Meunier H. Vaughan 1. et al. 1988 Production and regulation of inhibin subunitrin pituitary gonadotropes. Endocrinology. 124:552-4. 26. Corrigan AZ, Bilezikjian LM, Carroll RS, et al. 1991 Evidence for an autocrine role of activin B within rat anterior pituitary cultures, Endocrinology. 128:1682-4. 27. Carroll RS, Corrigan AZ, Vale W, Chin WW. 1991 Activin stabilizes follicle-stimulating hormone-beta messenger ribonucleic acid levels. Endocrinology. 129:1721-6. 28. Kogawa K, Nakamura T, Sugino K, Takio K, Titani K, Sugino H. 1991 Activin-binding protein% present in pituitary. Endoc&ology. 128:1434-40. 29. Mizunuma H, Samson WK, Lumpkin MD, Moltz HJ, Fawcett CP, McCann SM. 1983 Purification of a bioactive FSH-releasine ” factor (FSHRF). Brain Res Bull, 10:623-9.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
A Circadian Rhythm Hormone in Women* J. F. MORTOLAt,
G. A. LAUGHLIN,
Vol. 75, No. 3 Printed in U.S.A.
of Serum AND
S.
S.
Department of Reproductiue Medicine (0802) and the General California-San Diego, La Jolla, California 92093-0802
YENJ:
C.
Clinical
ABSTRACT While a nocturnal decline in serum LH levels in the early follicular phase of the menstrual cycle has been well established, a diurnal variation in serum FSH levels in women has not been demonstrated. We addressed this issue by determining serum LH and FSH levels at 15-min intervals for 24 h in the early follicular phase (EFP; n = 16) and late follicular phase (LFP; n = 10) of the menstrual cycle and in postmenopausal women (PMW; n = 10). Serum estradiol was simultaneously measured at hourly intervals. As expected, EFP, but not LFP and PMW, women had a 15% nocturnal decline (P < 0.01) in transverse mean LH levels compared to values in the daytime hours. In contrast, nocturnal FSH transverse mean values were significantly lower than daytime values in all groups studied, demonstrating an 18% decline in EFP (P < O.OOl), a 17% decline in LFP (P < O.OOOOl), and a 4.3% decline in PMW (P < 0.01).
I
N 1973, Kapen et al. (1) reported a nocturnal decline of serum LH restricted to the early follicular phase of the menstrual cycle. This observation was subsequently confirmed and demonstrated to be a result of sleep-dependent slowing of the LH pulse frequency (Z-6). Attempts to discern a diurnal rhythm in circulating FSH levels in humans have yielded conflicting results (7-13). For example, in men using frequent sampling, diurnal variations of both LH and FSH were found in one study (12), but not in another (13). To date, circadian variations in serum FSH levels in women have not been defined. We report here the finding of a remarkable circadian rhythm of serum FSH measuredin 24h samples at 15-min intervals in women during the early follicular phase (EFP) and the late follicular phase (LFP) of the menstrual cycle and in postmenopausal women (PMW). Subjects
Follicle-Stimulating
and Methods
Twenty-six normal cycling women between the ages of 20-41 yr and 10 PMW between the ages of 51-61 yr completed the study. PMW had been amenorrheic for at least 1 yr and were free from hormonal Received September 23, 1991. Address all correspondence and requests for reprints to: Dr. Samuel 5. C. Yen, Department of Reproductive Medicine (OSOZ), University of California-San Diego, La Jolla, California 92093-0802. * This work was supported by NIH NICHD Center Grant HD-1230313, a grant from the General Clinical Research Branch, Division of Research Resources (NIH MOl-RR-00827), and in part by the Mellon Foundation. This research was conducted in part by the Clayton Foundation for Research. t Current address: Beth Israel Hospital, Boston, Massachusetts 02215. $ Clayton Foundation investigator.
Research Center, University
of
Cosinor analysis revealed a circadian rhythm for FSH, with acrophases in the afternoon and nadirs at night in all three groups. The circadian amplitudes were 1.43 + 0.22, 1.02 + 0.16, and 8.42 + 1.31 IU/ L for EFP, LFP, and PMW, respectively. The EFP nocturnal decline in LH did not conform to a cosine rhythm. A diurnal variation in estradiol was not present in any of the groups of women. These data constitute the first demonstration of a robust circadian rhythm of serum FSH in women. The comparable timing of the acrophase in all groups of subjects and its presence in the postmenopausal years suggest a central, rather than peripheral, feedback mechanism(s) for the circadian rhythmicity. This observation provides strong evidence for a dissociation in the hypothalamic regulation of pituitary LH and FSH secretion in women. The circadian peak and nadir of circulating FSH may prove to be determining for appropriate follicular development. (J Clin Endocrinol Metab 75: 861-864, 1992)
replacements. Before enrollment, all premenopausal women reported regular menstrual cycles, ranging from 26-32 days in length. Subjects were without a history of medical illness, and physical examination was normal before inclusion in the study. This protocol was approved by the Committee of Investigations Involving Human Subjects at the University of California-San Diego. Studies were conducted in the Clinical Research Center (CRC). Sixteen EFP women (defined as cycle days 3-5) and 10 LFP women (defined as cycle days 10-12) participated in this investigation. Phases of the cycle were later confirmed by serum estradiol (E,) measurements. PMW were studied on an arbitrary day. Subjects were admitted to the CRC 2 h before the start of the investigation, during which time an iv catheter was placed in the forearm of the nondominant hand. Blood samples were obtained at 15-min intervals for LH and FSH determinations and at hourly intervals for El measurements. A sleep room that allowed blood drawing through a catheter extending to an adjacent room was used, so as not to disturb the subject during sleep. A standardized activity schedule was maintained for the 24-h study period, with meal times at 0800, 1200, and 1800 h. Ambulation and daytime napping were avoided. Lights were off between 2300-0700 h, and the sleep pattern of each subject was monitored by an experienced nurse through a video camera. LH and FSH values were determined in duplicate by RIA (14, 15). The sensitivities of the assays were 0.3 and 1.0 II-I/L for LH and FSH, respectively. The intraassay coefficients of variation (CVs) were 6% and 7%, and the interassay CVs were 10% and 11% for LH and FSH, respectively. El was determined by a previously described RIA (16). The intra- and interassay CVs for this steroid were 8% and 12%, respectively, and the sensitivity of the assay was 33 pmol/L. All samples from the same subject were included in the same assay run. Transverse mean levels of LH and FSH for each subject were determined for daytime hours (0700-2300 h) and nighttime hours (23000700 h) by calculating the arithmetic mean of all values obtained during the respective time intervals. The magnitude of the nocturnal changes in LH and FSH were defined as the difference between the calculated daytime and nighttime transverse mean levels for each subject. Significant differences between these values within each group were compared
861
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
862
MORTOLA,
LAUGHLIN.
by paired t tests. Cosine analysis of the concentration series of LH, FSH, and E2 was performed on each of the 15-min (60 min for E,) interval data sets for each individual by fitting the general cosine function, CS(t) = M + A cos(w + I#J), where CS(t) is the hormone concentration at time t, M is the mesor (midline value), A is the absolute amplitude, and w and 4 are the angular frequency and acrophase, respectively (17). The acrophase and nadir are defined as the times of occurrence of the maxima and minima in the cosine curve, respectively. Because individuals may display a significant cosine rhythm without a significant rhythm for the group as a whole (usually due to a disparity of acrophase times), individual cosinor analysis was followed by a population mean cosinor determination for each group (EFP, LFP, and PMW) (17). Cosine analysis was also performed on mean percent change from the transverse mean data for each group for the purpose of displaying the group cosine rhythm (see Fig. 2). Parameters of individual cosinor analysis were compared between groups by analysis of variance with Newman-Keuls post-hoc testing. For all statistical analyses, P < 0.05 was considered significant.
Results In EFP women, a nighttime decline in LH was observed in 13 of the 16 subjects studied. Mean nocturnal LH levels fell by 15% (P < 0.01) from a daytime level of 12.7 f 1.2 (+ SE) to 10.9 + 1.1 lU/L (Fig. 1). However, group cosinor analysis did not reveal a significant circadian rhythm for LH in EFP women. In LFP and PMW women, nighttime and daytime transverse mean LH levels were similar (12.8 + 1.5 VS. 12.9 f 1.6 lU/L and 81.7 f 10.5 us. 77.0 + 9.0 lU/L), and no significant group cosine rhythm was detected (Fig. 1). Transverse mean serum FSH levels in EFP declined by iS% (P C 0.001) compared to daytime values (11.0 + 0.7 VS. 9.2 + 0.4 lU/L). A similar 17% decline in FSH occurred in LFP (P < 0.00001) where values were reduced from a dayEARLY
30r
AND
YEN
JCE & M. 1992 Vol75.No3
time transverse mean of 7.2 + 0.8 lU/L to a nighttime value of 5.9 + 0.7 lU/L. Significantly lower transverse mean FSH levels were also seen at night compared to those during the day in PMW (P C 0.01). Although the absolute decline in FSH levels at night was larger in PMW (155.8 f 14.0 to 148.9 + 14.4 lU/L) than in cycling women, the relative change was reduced to 4.3%. Significant cosine rhythms for FSH were detected in 15 of 16 EFP women, 9 of 10 LFP women, and 9 of 10 PMW. These contributed to significant (P c 0.01) group circadian rhythms for all groups tested (Fig. 2). Arithmetic mean circadian rhythm results for FSH for each group are presented in Table 1. A similar time course of FSH concentrations was observed, with early afternoon acrophases (Table 1) and nadirs at night (EFP, 0304 h f 53 min; LFP, 0242 h k 41 min; PMW, 0145 h + 37 min; P = NS). When expressedin serum concentration units, the amplitudes of the circadian FSH variation were similar in both stages of the follicular phase (EFP, 1.43 + 0.22 lU/L; LFP, 1.02 + 0.16 lU/L) and elevated in PMW [8.42 f 1.31 lU/L; P < 0.001 us. EFP and EARLY
20 r
FOLLICULAR
I
10 0
f Y
-10
z 9
-20
I t i
2
20
f E
10 Ib I
LATE
FOLLICUL
4
5 E %
-10
1
-20
L
FOLLICULAR f 3
II
20 5 w
?v)
LATE FOLLICULAR
3o r
POST-MENOPAUSAL
.,-.
-20
1 0900
1700 CLOCK
3 t
TABLE women’s in EFP,
1. Mean (&SE) results of cosinor analysis of individual FSH values obtained on serum samples at 15-min intervals LFP, and PMW
-0900
1700 CLOCK
0100
0900
FIG. 2. Mean (&SE) percent change from 24-h mean in serum FSH values obtained at 15-min intervals for 24 h in EFP (n = 16), LFP (n = lo), and PMW (n = 10). The significant circadian variation of the hormone in each group of women is represented by the best-tit cosine curve.
POST-MENOPAUSAL
3o 1
0100 HOURS
0900
HOURS
FIG. 1. Mean (&SE) percent change from 24-h mean in serum LH levels obtained at 15-min intervals for 24 h in EFP (n = 16), LFP (n = lo), and PMW (n = lo), demonstrating a nocturnal decline in EFP, not present in LFP or PMW.
EFP (n = 16) LFP (n = 10) PMW (n = 10) “P < 0.001 * P < 0.001
Absolute amplitude W/L)
% amplitude m)
W/L)
Acropbase (clock hours)
1.43 + 0.22 1.02 f 0.16 8.42 f 1.31*
13.7 + 1.8 15.5 + 1.6 5.3 + 0.7*
10.31 f 0.58 6.48 + 0.79” 157.1 + 13.0*
15:04 + 0:53 14:42 + 0:41 13:45 + 0:37
us. EFP us. EFP
MSSCX
and PMW. and LFP.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
CIRCADIAN
RHYTHM
LFP). However, when expressed as a percentage of the 24-h mean, the relative amplitude of the circadian rhythm in PMW was markedly attenuated compared to that in follicular phase women (EFP, 13.7 + 1.8%; LFP, 15.5 f 1.6%; PMW, 5.3 f 0.7%; P < 0.001 us. EFP and LFP; Table 1 and Fig. 2). Hourly serum EZ levels were relatively stable throughout the 24-h sampling period within each group studied (Fig. 3). The transverse mean values were 127 + 8, 355 + 54, and 46 + 3 pmol/L for EFP, LFP, and PMW, respectively. There were no differences in daytime and nighttime values, and no significant group circadian rhythms were detected. Discussion This study constitutes the first demonstration of a circadian rhythm of circulating FSH in women. This circadian periodicity of FSH was observed in women of reproductive age during both the EFP and LFP of the menstrual cycle and in PMW. The results of our study reveal a marked disparity between the 24h profiles of serum FSH and LH levels. While the magnitudes of the EFP nocturnal declines in FSH and LH were similar (18% and 15%, respectively), the diurnal variation in LH was limited to this cycle phase and was without discernible circadian rhythmicity, as previously reported (4, 5). As demonstrated by sleep reversal (12) and naloxone infusion (6) studies, the nocturnal LH decline in EFP women is a sleep-entrained and opioidergic-mediated event, In contrast, the 24-h serum FSH profile appears to reflect a circadian rhythmicity, which is evident throughout the follicular phase of the cycle and in PMW. The lower relative (percentage), but not absolute (concentration), of the FSH circadian amplitude observed in PMW may be related to the hypersecretion of FSH in PMW and the long half-life of FSH (15), which could combine to mask the circadian amplitude. Alternatively, it may reflect attenuation of the central mechanism(s) controlling the circadian timekeeping system due to aging (18). The neuroendocrine mechanism(s) to account for the presence of a circadian rhythm of FSH, but not LH, is unclear. While it is well known that E2 exerts a preferential inhibition of FSH secretion (19, 20), the magnitude of the circadian excursion of FSH around the mean was similar in EFP and LFP despite a 3-fold increase in Ez levels. Moreover, since none of the groups of women in this study demonstrated significant diurnal variations in E2, the circadian rhythm of
i
400 300
OF SERUM
Acknowledgments We are grateful to Ms. Shannon Petze and Mr. Jeff Wong for their technical assistance, and to Ms. Del Alsobrook for preparation of the manuscript. Our thanks to Dr. Allen Lein for helpful advice in data analysis, and the CRC nursing staff for providing excellent care.
References 1. Kapen
2.
3. LFP
w” 5. 1700 CLOCK
0100 HOURS
0900
FIG. 3. Mean (SE) serum EZ levels determined at hourly the EFP (n = 16), LFP (n = lo), and PMW (n = 10).
intervals
in
863
FSH was not coupled with E2 fluctuations. The presence of a circadian rhythm for FSH in PMW suggests that ovarian factors, including inhibin, activin, and follistatin (22-24), may not play a significant role in the expression of the circadian rhythmicity of FSH secretion. Although unlikely, intrapituitary inhibin, activin and follistatin, acting as autocrine regulators of FSH gene expression and secretion (2528), may function independent of GnRH in generating the circadian rhythm of FSH. The similar timings of the acrophase of this rhythm in younger cycling women and older postmenopausal women suggest that the rhythm is mediated by a central mechanism(s). Neuroendocrine factors, such as sleep and a hypothalamic FSH-releasing factor distinct from GnRH (29) that is linked to the circadian pacemaker, have to be considered. The physiological significance of the FSH circadian rhythm described here is unclear. FSH should now be added to the number of pituitary hormones with circadian rhythmicity. However, the nocturnal nadir of FSH is opposite the nocturnal rise of ACTH, TSH, GH, and PRL (13, 18). In terms of target cell response, the presence of a circadian peak and nadir of FSH may be an important determining factor in the regulation of folliculogenesis and maturation; this is a speculation worthy of investigation. In summary, we have observed a robust circadian rhythm for circulating FSH levels in women with or without ovarian function. This finding provides additional evidence for the dissociation in the neuroendocrine regulation of pituitary release of FSH and LH. Further studies are needed to define the neuroendocrine mechanism(s) governing the circadian rhythmicity of pituitary FSH secretion and its aberrations in disorders of folliculogenesis.
4.
0900
FSH
6.
S, Boyar R, Perlow M, Hellman L, Weitzman ED. 1973 Luteinizing hormone: changes in secretory patterns during sleep in adult women. Life Sci. 13:693-g. Kapen S, Boyar R, Hellman L, Weitzman ED. 1976 The relationship of luteinizing hormone to sleep in women during the early follicular phase: effects of sleep reversal and a prolonged 3 hour sleep-wake schedule. J Clin Endocrinol Metab. 43:1031-7. Soules M, Steiner R, Cohen N, Bremner W, Clifton D. 1985 Nocturnal slowing of pulsatile luteinizing hormone secretion in women during the follicular phase of the menstrual cycle. J Clin Endocrinol Metab. 61:43-9. Filicori M, Santoto N, Merrium G, Crowley WJ. 1986 Characterization of the physiologic pattern of episodic gonadotropin secretion throughout the menstrual cycle. J Clin Endocrinol Metab. 62:113644. Rossmanith WC, Liu CH, Laughlin GA, Mortola JF, Suh BY, Yen SSC. 1990 Relative changes in LH pulsatility during the menstrual cycle: using hypogonadal women as a reference. Clin Endocrinol (Oxf). 32:647-60. Rossmanith WG, Yen SSC. 1987 Sleep associated decrease in luteinizing hormone pulse frequency during the early follicular
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
864
7. 8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
MORTOLA.
LAUGHLIN,
phase: evidence for an opioidergic mechanism. J Clin Endocrinol Metab. 65:715-g. Faiman C, Ryan RJ. 1967 Diurnal cycle in serum concentrations of follicle-stimulating hormone in men. Nature. 215:857. Saxena BB, Demura H, Gandy HM, Peterson RE. 1968 Radioimmunoassay of human follicle stimulating and luteinizing hormones in plasma. J Clin Endocrinol Metab. 28:519-34. Peterson Jr NT, Midgley AR, Jaffe RB. 1968 Regulation of gonadotropins. III. Luteinizing and follicle stimulating hormone in sera from adult males. J Clin Endocrinol Metab. 28:1473-g. Krieger DT, Ossowski R, Fogel M, Allen W. 1972 Lack of circadian periodicity of human serum FSH and LH levels. J Clin Endocrinol Metab. 35:619-23. Alford FP, Baker HWG, Pate1 YC, et al. 1973 Temporal patterns of circulating hormones as assessed by continuous blood sampling. L J Clin Endo&inol Metab. 36:108-12. . Veldhuis ID. Kine IC. Urban RI. et al. 1987 Oueratinz ” characteristics of the male hypothalamo-pituitary-gonadal axis: pulsatile release of testosterone and follicle stimulating hormone and their temporal coupling with luteinizing hormone. J Clin Endocrinol Metab. 65:929-41. Veldhuis JD, Iranmanesh A, Johnson ML, Lizarralde G. 1990 Twenty-four hour rhythms in plasma concentrations of adrenohypophyseal hormones are generated by distinct amplitude and/or frequency modulation of underlying pituitary secretory bursts. J Clin Endocrinol Metab. 71:1616-23. Yen SSC, Llerena 0, Little B, Pearson OH. 1968 Disappearance rates of endogenous luteinizing hormone and chorionic gonadotropin in man. J Clin Endocrinol Metab. 28:1763-7. Yen SSC, Llerena LA, Pearson OH, Littell AS. 1970 Disappearance rates of endogenous follicle-stimulating hormone in serum following surgical hypophysectomy in man. J Clin Endocrinol Metab. 30:3259. De Vane GW, Czekala NM, Judd HL, Yen SSC. 1975 Circulating gonadotropins, androgens, and estrogens in polycystic ovarian disease. Am J Obstet Gynecol. 121:496-501. Nelson W, Tong YL, Lee JK, Halberg F. 1979 Methods for cosinor ,.
“,.
,.
I
AND
YEN
JCE & M. 1992 Vol75.No3
rhythmometry. Chronogiologia. 6:305-9. 18. Van Coevorden A, Mockel J, Laurent E, et al. 1991 Neuroendocrine rhythms and sleep in aging men. Am J Physiol. 260:E851-61. 19. Yen SSC, Tsai CC, Vandenberg G, Rebar R. 1972 Gonadotropin dynamics in patients with gonadal dysgenesis: a model for the study of gonadotropin regulation. J Clin Endocrinol Metab. 35:897-904. 20. Yen SSC, Lasley BL, Wang CF, Leblanc H, Siler TM. 1975 The operating characteristics of the hypothalamic-pituitary system during the menstrual cycle and observations of biological action of somatostatin. Recent Prog Horm Res. 31:321-63. 21. McLachlan RI, Robertson DM, Healy DL, Burger HG, de Kretser DM. 1987 Circulating immunoreactive inhibin levels during the normal menstrual cycle. J Clin Endocrinol Metab. 65:954-61. 22. McLachlan RI, Robertson DM, De Kretser DM, Burger HG. 1988 Advances in the physiology of inhibin and inhibin-related peptides. Clin Endocrinol (Oxf). 29:77-114. 23. Ueno N, Ling N, Ying S-Y, Esch F, Shimasaki S, Guillemin R. 1987 Isolation and partial characterization of follistatin: a novel MW 35,000 monomeric protein that inhibits the release of follicle stimulating hormone. Proc Nat1 Acad Sci USA. 84:8282-6. 24. Roseff SJ, Bangah ML, Kettel LM, et al. 1989 Dynamic changes in circulating inhibin levels during the luteal-follicular transition of the human menstrual cycle. J Clin Endocrinol Metab. 69:1033-9. 25. Roberts V Meunier H. Vaughan 1. et al. 1988 Production and regulation of inhibin subunitrin pituitary gonadotropes. Endocrinology. 124:552-4. 26. Corrigan AZ, Bilezikjian LM, Carroll RS, et al. 1991 Evidence for an autocrine role of activin B within rat anterior pituitary cultures, Endocrinology. 128:1682-4. 27. Carroll RS, Corrigan AZ, Vale W, Chin WW. 1991 Activin stabilizes follicle-stimulating hormone-beta messenger ribonucleic acid levels. Endocrinology. 129:1721-6. 28. Kogawa K, Nakamura T, Sugino K, Takio K, Titani K, Sugino H. 1991 Activin-binding protein% present in pituitary. Endoc&ology. 128:1434-40. 29. Mizunuma H, Samson WK, Lumpkin MD, Moltz HJ, Fawcett CP, McCann SM. 1983 Purification of a bioactive FSH-releasine ” factor (FSHRF). Brain Res Bull, 10:623-9.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 17:31 For personal use only. No other uses without permission. . All rights reserved.
