Nature of gonadotropin-releasing hormone self-priming of luteinizing hormone secretion during the normal menstrual cycle Michael J. Sollenberger, MD, Elisabeth C. Carlsen, MD, Robert A. Booth, Jr., MS, Michael L. Johnson, PhD, Johannes D. Veldhuis, MD, and William S. Evans, MD Charlottesville, Virginia To investigate further the nature of the gonadotropin-releasing hormone self-priming effect on luteinizing hormone release, we administered two submaximal doses of gonadotropin-releasing hormone 2 hours apart to women at three stages of the menstrual cycle and analyzed the resultant luteinizing hormone secretory episodes with deconvolution analysis. When the characteristics of the secretory episodes associated with the second gonadotropin-releasing hormone challenge were compared with those associated with the first, both an enhanced maximal secretory rate and mass of luteinizing hormone secreted was demonstrable at each phase of the cycle. No differences in the luteinizing hormone secretory event half-duration were detected when the responses to the first and second gonadotropin-releasing hormone doses were compared. These data confirm the gonadal hormone milieu-associated self-priming effect of gonadotropin-releasing hormone on luteinizing hormone release and indicate that it is the rate with which luteinizing hormone molecules are discharged from the pituitary gland, rather than the duration of the secretory episode itself, that provides for the self-priming effect. (AM J OBSTET GVNECOL 1990;163:1529-34.)
Key words: GnRH self-priming, deconvolution analysis
The ambient gonadal hormone environment modulates gonadotropin-releasing hormone (GnRH)stimulated luteinizing hormone (LH) secretion both in vivo l -5 and in vitro.6 Moreover, under appropriate conditions in humans:' 7 in animal models," 5. 8-11 and in vitro,3. 6.12-14 repetitive administration of GnRH results in significant potentiation of LH secretion by the anterior pituitary gland. This phenomenon of enhanced pituitary responsiveness associated with multiple GnRH challenges has been referred to as "selfpriming" and may represent a primary mechanism subserving the preovulatory LH surge. 3. 6. 8-10.13 From the Departments of Medicine and Pharmacology, Health Sciences Center, and the Biodynamics Institute, University of Virginia. Supported by National Institutes of Health RCDA grant No. 1K04 HD00711, American Diabetes Association Feasibility Grant, and University of Virginia Computer Services grant to W.S.E.; National Institutes of Health Research Career Development Award 1K04 HD00634 and University of Virginia Computer Services grant to J.D. V.; Nationallnstitutes of Health grant R01 AG05977 to W.S.E. and J.D. V.; United States Public Health Service General Clinical Research grant RR-847; and Diabetes and Endocrinology Research Center grant No. DK 38942. Presented in part at the Seventieth Annual Meeting of the Endocrine Society, New Orleans, Louisiana, 1988. Received for publication April 26, 1990; revised July 16, 1990; accepted July 20,1990. Reprint requests: William S. Evans, MD, Box 511, Department of Medicine, University of Virginia Health Sciences Center, Charlottesville, VA 22908. 611 /24038
Despite intensive investigation, the mechanisms that subserve GnRH self-priming remain to be clarified. To date, studies in the human have focused entirely on changes in the circulating concentration of LH after the administration of serial pulses of GnRH. Although of obvious importance with regard to defining the event, these observations have been primarily descriptive in nature and have not provided insight into the determinants of the GnRH-stimulated LH secretory burst. Recently a novel analytic approach known as multiple-parameter deconvolution has been successfully applied to a variety of hormonal secretory profiles. 15 This technique allows for the simultaneous resolution of specific, quantitative features of hormone secretion together with subject-specific metabolic clearance rates. We have used this model in this study to assess LH secretion in response to two submaximal bolus infusions of GnRH administered 2 hours apart in 12 women at three stages of the menstrual cycle. With this design, both the effect of the gonadal hormone milieu on GnRH-stimulated LH secretion (in response to the first bolus of GnRH) and on GnRH selfpriming (LH secretion in response to the second versus the first GnRH bolus) could be assessed. Specifically, we tested the hypotheses that the self-priming phenomenon defined previously by changes in serum concentrations of LH may reflect either (1) alterations in one or more attributes encoding for the luteinizing hor1529
1530 Sollenberger et al.
November 1990 Am J Obstet Gynecol
Table I. Mean (± SEM) serum concentrations of gonadal steroid hormones in normal women at three phases of menstrual cycle* Menstrual cycle phase
17 f3-Estradiol
Early follicular Late follicular Midluteal
24 ± 7a 119 ± 13b
(pglml)
128 ± 21b
Progesterone (ngldl) 34 ± 8a
27 ± 4a 1503 ± 284 b
Entries within each column identified by different superscripts (a and b) differ significantly (Duncan's multiple-range test, p < 0.05). *n = 12 women for each study phase.
mone secretory events stimulated by GnRH or (2) alterations in the half-life of secreted hormone. Material and methods
Subjects and experimental procedures. Twentynine young women (mean age, 28 years; range, 18 to 40 years) were studied on one or more occasions at the University of Virginia Clinical Research Center. The study was approved by the Human Investigation Committee of the University of Virginia and each subject provided written consent. All volunteers had a normal history and physical examination and normal serum concentrations of LH, follicle-stimulating hormone, prolactin, thyroxine, thyrotropin, and total testosterone. During the month of study the phase of the menstrual cycle was documented by daily blood samples for measurement of 17~-estradiol, progesterone, and LH. Twelve women were studied in each of three stages of the menstrual cycle including the early follicular phase (days 2 to 4 after beginning of the menses), the late follicular phase (l to 4 days before the LH surge), and the midluteal phase (5 to 9 days after the LH surge). On the day of study blood samples were obtained from an indwelling heparin cannula at lO-minute intervals from 8 AM to 12 noon. GnRH (lO ILg intravenously) was given at 8 AM and lO AM. To be included in the analysis, the LH concentration response to the GnRH bolus had to be sufficient to fulfill the following criteria: (1) a peak increase of at least 2 mIU/ml above baseline, (2) a peak response occurring within 20 minutes of GnRH injection, (3) a nonzero decay slope, and (4) 95% of samples available for analysis. Assays. Serum concentrations of prolactin, thyroxine, thyrotropin, testosterone, estradiol, and progesterone were determined by radioimmunoassay. Progesterone radioimmunoassay was performed with reagents supplied by Diagnostic Products (Los Angeles). Serum LH and follicle-stimulating hormone concentrations were measured with reagents from Clinetics Corp. (Tusten, Calif.). Median intrassay coefficients of
variation (calculated from each subject'S 145 samples collected over the preceding 24 hours) ranged from 4.6% to 6.2%. All samples from an individual woman were analyzed in duplicate in the same assay. Deconvolution analysis. Serum hormone concentrations result from the combination of the underlying pituitary secretory events and endogenous (subjectspecific) metabolic clearance. Deconvolution analysis allows for the simultaneous determination of the quantitative properties of underlying secretory bursts (characterized by a finite mass, amplitude, and duration) and endogenous clearance kinetics. 15 Endogenous clearance kinetics are related to half-lives by the relationship MCR = (In21t~2)/Vd' where MCR is the metabolic clearance rate and Vd represents the volume of distribution. An LH secretory event is represented algebraically by a gaussian distribution of instantaneous molecular secretory rates that are centered around a particular point in time and dispersed with a finite standard deviation. Multiple-parameter deconvolution procedures were applied separately to the LH data obtained after the administration of each GnRH bolus (8 AM to lO AM and lO AM to 12 noon). Statistical analysis. One-way analysis of variance with Duncan's multiple-range test was used to compare interphase differences in serum concentrations of gonadal steroids. Similar analyses of interphase secretory burst properties and LH half-lives were performed after logarithmic transformation of the data because of departures from normality. Comparisons of these parameters for each GnRH bolus were made by nonparametric methods (Wilcoxon signed-rank test). Correlations between two measures were sought by linear regression analysis of un transformed data. Multiple linear regression using the R2 maximal improvement criterion was applied to define relationships among three or more variables. Results
Circulating concentrations of gonadal hormones. The mean serum concentrations of gonadal hormones are summarized in Table I. Serum 17~-estradiol concentrations were similar in the late follicular and midluteal phases and significantly higher than in the early follicular phase. Circulating progesterone concentrations were increased in the midluteal phase, but was indistinguishable in the two follicular phases studied. Analysis of luteinizing hormone secretion in response to the initial GnRH bolus (GnRH 1). Fig. I shows GnRH-stimulated LH concentration time series and the corresponding deconvolution-resolved LH secretory impulses obtained in a representative woman studied at three phases of the menstrual cycle. Visual inspection of these data suggested the presence of cycledependent differences in LH secretory event amplitude
GnRH self-priming in normal women
Volume 163 Number 5, Part 1
1531
Table II. Mean (± SEM) LH secretory burst characteristics during early follicular, late follicular, and midluteal phases of menstrual cycle LH secretory burst Maximal secretory rate (mIU I mll min) Phase
GnRH 1
Early follicular Late follicular Midluteal
1.5 ± 0.3' 3.8 ± lA' 6.0 ± 1.3b
I
HalFduration (min)
GnRH2
GnRH 1
2.5 ± 0.5* 9.2 ± 2.7t 12.0 ± 1.9t
8.5 ± 1.5' Il.8 ± LOb 8.5 ± 1.3'
I
Mass (mIUlml)
GnRH2
GnRH 1
6.8 ± 0.8 13.9 ± 1.5 604 ± 0.6
13 ± 4' 39 ± 9b 48 ± 9b
T GnRH2 16 ± 3t III ± 26* 78 ± II *
Half-life (min) GnRH 1
125 ± 14' Il7 ± 12' 64 ± 4b
1
GnRH2
117 ± 10 82 ± 8* 61 ± 4
GnRH (10 f.1g intravenously) was given at 0 (GnRH I) and 2 (GnRH 2) hours. Duration (in minutes) of the computer-resolved LH secretory burst at half-maximal amplitude. The volume term (milliliters) represents unit distribution volume. Amplitude is the maximal rate of LH secretion attained in the deconvolution-resolved LH secretory burst. Entries within each GnRH I column identified by different superscripts (a and b) differ significantly (Duncan's multiple-range test of log transformed values, p < 0.05). *Comparisons of secretory burst properties and LH half-lives (GnRH I vs GnRH 2) were made by non parametric options (Wilcoxon signed-rank test): p < 0.01. tComparisons of secretory burst properties and LH half-lives (GnRH I vs GnRH 2) were made by non parametric options (Wilcoxon signed-rank test): 0.01 < P < 0.05.
(maximal secretory rate) and half-duration in response to the first and second challenges with GnRH. Analysis of the deconvolution-resolved secretory event characteristics confirmed this visual impression. Table II summarizes the LH secretory event characteristics inferred in response to the first bolus of GnRH as a function of the phase of the menstrual cycle. Although there was a trend toward a higher maximal LH secretory rate in response to GnRH during the late follicular versus the early follicular phase, a significant difference was not realized (p = 0.076). In contrast, the maximal secretory rate during the midluteal phase was significantly greater than that observed in the early follicular (p = 0.0002) or late follicular (p = 0.021) phases. However, during the late follicular phase the duration of the GnRH-associated LH secretory event was substantially longer than that documented during the early follicular (p = 0.022) and midluteal (p = 0.041 ) phases. The mass of LH released in response to GnRH was greater in both the late follicular and midluteal phases of the cycle compared with the early follicular phase (p = 0.002 and 0.0001, respectively), but the former two were indistinguishable (p = 0.376). Whereas the half-life estimates of LH released during the early follicular and late follicular phases were similar (p = 0.754), that of the material secreted in the midluteal phase was significantly shorter (p = 0.0001). Comparison of luteinizing hormone secretory response to GnRH boluses (GnRH 1 vs GnRH 2). Table II also shows the LH secretory event characteristics in response to a second challenge with a submaximal dose of GnRH administered 2 hours after the initial bolus. During the early follicular phase both the maximal rate of LH secretion and the mass of LH released were
greater (p < 0.01 and p < 0.05, respectively) in response to the second versus the first GnRH challenge. However, neither the secretory event duration nor the half-life of hormone secreted differed. The same pattern was evident during the midluteal phase [i.e., a higher maximal secretory rate (p < 0.05) and mass (p < 0.01)], but indistinguishable duration and half-life of LH were found when the response to the second GnRH bolus was compared with the response to the first. During the late follicular phase, both the maximal secretory rate and mass of hormone secreted were increased in response to the second challenge with GnRH (p < 0.05 and p < 0.01, respectively). However, during the late follicular phase, and in contrast to the results obtained in the early follicular and midluteal phases, the half-life of secreted LH in response to the second injection of GnRH was significantly (p < 0.01) shorter than that released in response to the first dose. No difference in the secretory burst duration was detected. Correlations among GnRH-stimulated secretory event characteristics and the circulating concentrations of the gonadal hormones. When all women were considered as a group across various stages of the menstrual cycle, multiple stepwise linear regression analysis demonstrated that serum estradiol concentrations were significantly predictive of the mass of LH secreted after the first GnRH injection (p < 0.0001 and R2 = 0.41 for mass). Serum estradiol levels also predicted the maximal rate of LH release after the first GnRH pulse (p = 0.004). Serum progesterone concentrations predicted the LH half-life (p < 0.004) and the halfduration (p < 0.02) of the first LH secretory burst. When individual phases of the menstrual cycle were considered alone, the following findings emerged: (1)
1532 Sollenberger et al.
November 1990 Am J Obstet Gynecol
LATE FOLLICULAR GnRH 2
GnRH 1
l
30r---------------~
o
~ 120 TIME (min)
TIME (min)
MID-LUTEAL
EARLY FOLLICULAR GnRH 1
GnRH 1
30,---------______~
120 TIME (min)
80r---------------~
fJ\
120
120
TIME (min)
TIME (min)
j\ 120 TIME (min)
Fig. 1. Illustrative profiles of the deconvolved LH secretory impulses in response to two GnRH challenges (GnRH 1 and GnRH 2) in a representative normal woman. For each phase of the menstrual cycle the upper panel depicts measured serum LH concentrations (mean ± SD) and the fitted reconvolution curve (continuous line), which is based on the calculated secretion and clearance parameters. The resolved LH secretory impulses are shown in the lower panel.
in the early follicular phase, serum estradiol levels were predictive of both the second luteinizing hormone secretory burst's mass (or its maximal rate) and its halfduration (p < 0.05, R2 ~ 0.48); (2) in the early follicular (but not late follicular) phase, serum progesterone concentrations predicted LH secretory burst mass after the first pulse of GnRH (p = 0.03, R2 = 0.42); (3) in the late follicular and midluteal phases, serum estradiol values correlated with the mass and the maximal rate of luteinizing hormone secreted after the first pulse of GnRH (p < 0.03, R2 ~ 0.48); (4) in the midluteal phase, serum 17J3-estradiol predicted both the mass of LH secreted and the half-duration ofthe second LH release episode (p < 0.03, R2 > 0.63); and (5) in the midluteal phase, serum progesterone concentrations were signif-
icantly predictive of the mass of LH secreted after the second GnRH pulse (p = 0.008, R2 = 0.56). Thus, in each of the different phases of the menstrual cyde, serum estradiol concentrations correlated positively and significantly with the mass and maximal rate of LH secretion attained in response to exogenous pulses of GnRH. Comment
Although the importance of the preovulatory LH surge within the context of normal reproductive function has long been recognized, the cellular events that signal the dramatically increased serum concentrations of LH at midcyde remain incompletely defined. Data obtained in animal models suggest that the secretion of
Volume 163 Number 5, Part I
GnRH into the hypothalamic-hypophysial portal circulation is markedly enhanced coincident with the preovulatory LH surge,I6 However, it is generally accepted that increased GnRH secretion does not solely account for the midcycle LH surge; rather it is thought that the surge reflects a situation in which increased concentrations of GnRH impinge on an anterior pituitary gland that has been rendered considerably more responsive to the releasing hormone just at midcycle. Accordingly, significant investigative interest has focused on the mechanisms subserving this alteration in gonadotroph responsivity. That GnRH-stimulated LH release is dependent on the gonadal hormone milieu has been documented in a variety of in vivo and in vitro studies. I-6 Of particular interest are the reports suggesting that, under appropriate conditions, enhanced LH release is observed in response to GnRH after an initial challenge with a "priming" dose of the hypothalamic peptide.'-6 7-'1 Although this phenomenon appears to be dependent on the gonadal hormone environment/' lI-14 the mechanisms involved remain poorly understood. In this study we coupled a moderately intensive sampling paradigm with analysis of the resultant LH concentration time series by use of a multiple-parameter deconvolution procedure to appraise in some detail the determinants of the GnRH secretory-priming event at several stages of the menstrual cycle. Our current results are consistent with previous studies in women that have documented that the stage of the menstrual cycle," presumably reflecting the gonadal hormone environment,' influences the serum concentrations of luteinizing hormone achieved in response to a single bolus of GnRH. The use of deconvolution precedures to "unravel" the GnRH-associated secretory event now provides an enhanced understanding of how this menstrual state-specific modulation occurs. Thus the maximal rate of LH secretion (as distinct from serum LH concentrations per se) achieved in response to GnRH in the midluteal phase is significantly greater than that observed in the early or late follicular phases. In contrast, the duration of the secretory episode in the late follicular phase is prolonged in comparison with the early follicular and midluteal phases. It should be noted that the duration of secretion is considerably shorter than that of the serum concentration peak. Thus the enhanced maximal secretory rate in the midluteal phase and the longer secretory event duration in the late follicular phase are responsible for the greater mass of hormone secreted in these two phases when compared with that secreted in the early follicular phase. Such information about secretion requires a technique such as deconvolution analysis, and cannot be inferred from simple inspection of the serum LH concentration peak. '5 Clearly, these studies do not address the question of which particular intracellular pro-
GnRH self-priming in normal women
1533
cesses subserve the menstrual stage-associated alterations in maximal LH secretory rate and secretory event duration. However, in consideration of this experimental design in which the hypothalamic peptide was administered exogenously, these data do demonstrate that short-term variations in the amount and/or duration of the GnRH stimulus per se are not required to effect variations in the characteristics of the LH secretory event. More prolonged variations in the characteristics of the endogenous GnRH pulse signal throughout the menstrual cycle may act to influence pituitary responsiveness. Our data also support earlier studies that demonstrated GnRH self-priming in the late follicularmidluteaF and midfollicular-midluteaF phases of the menstrual cycle. However, our results also provide evidence for a previously undescribed potentiating effect of GnRH on certain secretory event characteristics during the early follicular phase. Specifically, the maximal rate of LH secretion achieved in response to the second bolus of GnRH compared with the first was also enhanced in the early follicular phase and the midluteal phase. However, the duration of the secretory events in response to the two consecutive GnRH challenges did not differ. These results imply that it is the rate with which LH molecules are discharged from the pituitary gland within a given amount of time, rather than the duration of release itself, that results in the observed augmentation of LH secretion with repetitive GnRH administration. Serum estradiol concentrations were significantly predictive of the mass (as well as maximal rate) of LH secreted after the first GnRH injection in the late follicular and midluteal phases. This finding is consistent with an earlier inference by Lasley et al. 1 that estrogen participates in increasing intra pituitary LH stores that are available for release in response to secretagogue action. 1 Of additional interest is the finding that serum estrogen levels also correlated with the mass of LH released after the second pulse of GnRH in the early follicular phase. This observation suggests that both estrogen and prior exogenous GnRH exposure are required in the early follicular phase to demonstrate an increased availability of LH for release. Serum progesterone concentrations were negatively correlated with LH half-life and secretory burst duration, possibly indicating changes in the posttranslational carbohydrate modification of LH molecules with associated changes in the metabolic clearance rate of LHI7 and alterations in the kinetics of GnRH action and/or gonadotroph secretory responses, which result in a decreased secretory burst duration. In the midluteal phase, progesterone concentrations were related to the mass of LH released after the second GnRH pulse. Thus progesterone also may influence the self-priming actions of GnRH.
1534 Sollenberger et al.
The half-lives reported for endogenous LH correspond to half-lives of LH (calculated from stable metabolic clearance rates) of 123 minutes,!s 136 minutes,19 and 112 minutes,"" assuming a 3.5 L distribution volume for LH. Thus our LH half-life values of 60 to 120 minutes agree with independently estimated values obtained by steady-state infusion of various exogenous LH preparations. It should be noted that the deconvolution model used in this study provides estimates of LH half-life under steady-state conditions (or approximately its slower second component rate constant in non equilibrium studies) and does not reflect acute distribution coefficients (initial equilibration half-times after bolus injection), which could be calculated with even more rapid sampling. This observation presumably reflects the fact that LH distribution is virtually complete between consecutive (lO-minute) sample observations. Moreover, our results do not disagree with those of Wehmann et a!., IS who showed a moderate increase in metabolic clearance rate (or decrease in half-life) in the luteal phase. We thank Sandra Jackson and the nursing staff at the Clinical Research Center for their expert patient care and Catherine Kern and Ginger Bauer for technical assistance with the assays. REFERENCES 1. Lasley BL, Wang CF, Yen SSC. The effects of estrogen and progesterone on the functional capacity of the gonadotrophs. ] Clin Endocrinol metab 1975;41 :820-6. 2. Wang CF, Lasley BL, Lein A, Yen SSC. The functional changes of the pituitary gonadotrophs during the menstrual cycle.] Clin Endocrinol Metab 1976;42:718-28. 3. Pickering AJMC, Fink G. Priming effect of luteinizing hormone r~leasing factor: in vitro and in vivo evidence consistent with its dependence upon protein and RNA synthesis.J Endocrinol 1976;69:373-9. 4. Hoff JD, Lasley CL, Wang CF, Yen SSC. The two pools of pituitary gonadotropin: regulation during the menstrual cycle. J Clin Endocrinol Metab 1977;44:302-12. 5. Higuchi T, Kawakami M. Luteinizing hormone responses to repeated injections of luteinizing hormone-releasing hormone in the rate during the oestrous cycle and after ovariectomy with and without oestrogen treatment.] Endocrinol 1982;93: 161-8. 6. Waring DW, Turgeon]L. Luteinizing hormone-releasing hormone-induced luteinizing hormone secretion in vitro: cyclic changes in responsiveness and self-priming. Endocrinology 1980; 106: 1430-6.
November 1990 Am J Obstet Gynecol
7. Rommler A. Short-term regulation of LH and FSH secretion in cyclic women 1. Altered pituitary response to a second of two LH-RH injections at short intervals. Acta Endocrinol [CopenhJ 1978;87:248-58. 8. Aiyer MS, Chiappa SA, Fink G. A priming effect of luteinizing hormone-releasing factor on the anterior pituitary gland in the female rat.] Endocrinol 1974;62:57388. 9. Castro-Vaszquez A, McCann SM. Cyclic variations in the increased responsiveness of the pituitary to luteinizing hormone-releasing hormone (LHRH) induced by LHRH. Endocrinology 1975; 97: 13-8. 10. Fink G, Chiappa SA, Aujer MS. Priming effect of luteinizing hormone-releasing factor elicited by preoptic stimulation and by intravenous infusion and multiple injections of the synthetic decapeptide. ] Endocrinol 1976; 69:359-72. 11. Nazian SJ. Androgenic and estrogenic control of the selfpriming effect of LHRH in the castrated male rat. Neuroendocrinology 1986;42: 112-9. 12. Beck LV, Bay M, Smith AF, King D, Long R. Steroid priming of the luteinizing hormone response to luteinizing hormone-releasing hormone. ] Endocrinol 1978; 77:293-9. 13. Speight A, Fink G. Comparison of steroid and LH-RH effects on the responsiveness of hemipituitary glands and dispersed pituitary cells. Mol Cell Endocrinol 1981;24: 267-81. 14. Padmanabhan V, Leung K, Convey EM. Ovarian steroids modulate the self-priming effect of luteinizing hormonereleasing hormone on bovine pituitary cells in vitro. Endocrinology 1982;110:717-21. 15. Veldhuis JD, Carlson ML, Johnson ML. The pituitary gland secretes in bursts: appraising the nature of glandular secretory impulses by simultaneous multipleparameter deconvolution of plasma hormone concentrations. Proc Nat! Acad Sci USA 1987;84:7686-90. 16. Sarkar DK, Chiappa SA, Fink G, Sherwood N. Gonadotropin-releasing hormone surge in pro-oestrous rats. Nature 1976;264:461-3. 17. Dufau ML, Veldhuis ]D. Pathophysiological relationships between the biological and immunological activities of luteinizing hormone. In: Burger HG, ed. Balliere's clinical endocrinology and metabolism. Philadelphia: WB Saunders, 1987 vol 1:153-76. 18. Wehmann RE, Blackman MR, Harman SM. Metabolic clearance rates of luteinizing hormone in women during different phases of the menstrual cycle and while taking an oral contraceptive. J Clin Endocrinol Metab 1982; 55:654-9. 19. Marshall]C, Anderson DC, Fraser TR, Harsoulis P. Human luteinizing hormone in man: studies of metabolism and biological action.] EndocrinoI1973;56:431-9. 20. Pepperell~, DeKretser DM, Burger HG. Studies on the metabolic clearance rate and production rate of human luteinizing hormone and on the initial half-time of its subunits in man.] Clin Invest 1975;56:118-26.
A complete list of references is available from the authors on request.
Key words: GnRH self-priming, deconvolution analysis
The ambient gonadal hormone environment modulates gonadotropin-releasing hormone (GnRH)stimulated luteinizing hormone (LH) secretion both in vivo l -5 and in vitro.6 Moreover, under appropriate conditions in humans:' 7 in animal models," 5. 8-11 and in vitro,3. 6.12-14 repetitive administration of GnRH results in significant potentiation of LH secretion by the anterior pituitary gland. This phenomenon of enhanced pituitary responsiveness associated with multiple GnRH challenges has been referred to as "selfpriming" and may represent a primary mechanism subserving the preovulatory LH surge. 3. 6. 8-10.13 From the Departments of Medicine and Pharmacology, Health Sciences Center, and the Biodynamics Institute, University of Virginia. Supported by National Institutes of Health RCDA grant No. 1K04 HD00711, American Diabetes Association Feasibility Grant, and University of Virginia Computer Services grant to W.S.E.; National Institutes of Health Research Career Development Award 1K04 HD00634 and University of Virginia Computer Services grant to J.D. V.; Nationallnstitutes of Health grant R01 AG05977 to W.S.E. and J.D. V.; United States Public Health Service General Clinical Research grant RR-847; and Diabetes and Endocrinology Research Center grant No. DK 38942. Presented in part at the Seventieth Annual Meeting of the Endocrine Society, New Orleans, Louisiana, 1988. Received for publication April 26, 1990; revised July 16, 1990; accepted July 20,1990. Reprint requests: William S. Evans, MD, Box 511, Department of Medicine, University of Virginia Health Sciences Center, Charlottesville, VA 22908. 611 /24038
Despite intensive investigation, the mechanisms that subserve GnRH self-priming remain to be clarified. To date, studies in the human have focused entirely on changes in the circulating concentration of LH after the administration of serial pulses of GnRH. Although of obvious importance with regard to defining the event, these observations have been primarily descriptive in nature and have not provided insight into the determinants of the GnRH-stimulated LH secretory burst. Recently a novel analytic approach known as multiple-parameter deconvolution has been successfully applied to a variety of hormonal secretory profiles. 15 This technique allows for the simultaneous resolution of specific, quantitative features of hormone secretion together with subject-specific metabolic clearance rates. We have used this model in this study to assess LH secretion in response to two submaximal bolus infusions of GnRH administered 2 hours apart in 12 women at three stages of the menstrual cycle. With this design, both the effect of the gonadal hormone milieu on GnRH-stimulated LH secretion (in response to the first bolus of GnRH) and on GnRH selfpriming (LH secretion in response to the second versus the first GnRH bolus) could be assessed. Specifically, we tested the hypotheses that the self-priming phenomenon defined previously by changes in serum concentrations of LH may reflect either (1) alterations in one or more attributes encoding for the luteinizing hor1529
1530 Sollenberger et al.
November 1990 Am J Obstet Gynecol
Table I. Mean (± SEM) serum concentrations of gonadal steroid hormones in normal women at three phases of menstrual cycle* Menstrual cycle phase
17 f3-Estradiol
Early follicular Late follicular Midluteal
24 ± 7a 119 ± 13b
(pglml)
128 ± 21b
Progesterone (ngldl) 34 ± 8a
27 ± 4a 1503 ± 284 b
Entries within each column identified by different superscripts (a and b) differ significantly (Duncan's multiple-range test, p < 0.05). *n = 12 women for each study phase.
mone secretory events stimulated by GnRH or (2) alterations in the half-life of secreted hormone. Material and methods
Subjects and experimental procedures. Twentynine young women (mean age, 28 years; range, 18 to 40 years) were studied on one or more occasions at the University of Virginia Clinical Research Center. The study was approved by the Human Investigation Committee of the University of Virginia and each subject provided written consent. All volunteers had a normal history and physical examination and normal serum concentrations of LH, follicle-stimulating hormone, prolactin, thyroxine, thyrotropin, and total testosterone. During the month of study the phase of the menstrual cycle was documented by daily blood samples for measurement of 17~-estradiol, progesterone, and LH. Twelve women were studied in each of three stages of the menstrual cycle including the early follicular phase (days 2 to 4 after beginning of the menses), the late follicular phase (l to 4 days before the LH surge), and the midluteal phase (5 to 9 days after the LH surge). On the day of study blood samples were obtained from an indwelling heparin cannula at lO-minute intervals from 8 AM to 12 noon. GnRH (lO ILg intravenously) was given at 8 AM and lO AM. To be included in the analysis, the LH concentration response to the GnRH bolus had to be sufficient to fulfill the following criteria: (1) a peak increase of at least 2 mIU/ml above baseline, (2) a peak response occurring within 20 minutes of GnRH injection, (3) a nonzero decay slope, and (4) 95% of samples available for analysis. Assays. Serum concentrations of prolactin, thyroxine, thyrotropin, testosterone, estradiol, and progesterone were determined by radioimmunoassay. Progesterone radioimmunoassay was performed with reagents supplied by Diagnostic Products (Los Angeles). Serum LH and follicle-stimulating hormone concentrations were measured with reagents from Clinetics Corp. (Tusten, Calif.). Median intrassay coefficients of
variation (calculated from each subject'S 145 samples collected over the preceding 24 hours) ranged from 4.6% to 6.2%. All samples from an individual woman were analyzed in duplicate in the same assay. Deconvolution analysis. Serum hormone concentrations result from the combination of the underlying pituitary secretory events and endogenous (subjectspecific) metabolic clearance. Deconvolution analysis allows for the simultaneous determination of the quantitative properties of underlying secretory bursts (characterized by a finite mass, amplitude, and duration) and endogenous clearance kinetics. 15 Endogenous clearance kinetics are related to half-lives by the relationship MCR = (In21t~2)/Vd' where MCR is the metabolic clearance rate and Vd represents the volume of distribution. An LH secretory event is represented algebraically by a gaussian distribution of instantaneous molecular secretory rates that are centered around a particular point in time and dispersed with a finite standard deviation. Multiple-parameter deconvolution procedures were applied separately to the LH data obtained after the administration of each GnRH bolus (8 AM to lO AM and lO AM to 12 noon). Statistical analysis. One-way analysis of variance with Duncan's multiple-range test was used to compare interphase differences in serum concentrations of gonadal steroids. Similar analyses of interphase secretory burst properties and LH half-lives were performed after logarithmic transformation of the data because of departures from normality. Comparisons of these parameters for each GnRH bolus were made by nonparametric methods (Wilcoxon signed-rank test). Correlations between two measures were sought by linear regression analysis of un transformed data. Multiple linear regression using the R2 maximal improvement criterion was applied to define relationships among three or more variables. Results
Circulating concentrations of gonadal hormones. The mean serum concentrations of gonadal hormones are summarized in Table I. Serum 17~-estradiol concentrations were similar in the late follicular and midluteal phases and significantly higher than in the early follicular phase. Circulating progesterone concentrations were increased in the midluteal phase, but was indistinguishable in the two follicular phases studied. Analysis of luteinizing hormone secretion in response to the initial GnRH bolus (GnRH 1). Fig. I shows GnRH-stimulated LH concentration time series and the corresponding deconvolution-resolved LH secretory impulses obtained in a representative woman studied at three phases of the menstrual cycle. Visual inspection of these data suggested the presence of cycledependent differences in LH secretory event amplitude
GnRH self-priming in normal women
Volume 163 Number 5, Part 1
1531
Table II. Mean (± SEM) LH secretory burst characteristics during early follicular, late follicular, and midluteal phases of menstrual cycle LH secretory burst Maximal secretory rate (mIU I mll min) Phase
GnRH 1
Early follicular Late follicular Midluteal
1.5 ± 0.3' 3.8 ± lA' 6.0 ± 1.3b
I
HalFduration (min)
GnRH2
GnRH 1
2.5 ± 0.5* 9.2 ± 2.7t 12.0 ± 1.9t
8.5 ± 1.5' Il.8 ± LOb 8.5 ± 1.3'
I
Mass (mIUlml)
GnRH2
GnRH 1
6.8 ± 0.8 13.9 ± 1.5 604 ± 0.6
13 ± 4' 39 ± 9b 48 ± 9b
T GnRH2 16 ± 3t III ± 26* 78 ± II *
Half-life (min) GnRH 1
125 ± 14' Il7 ± 12' 64 ± 4b
1
GnRH2
117 ± 10 82 ± 8* 61 ± 4
GnRH (10 f.1g intravenously) was given at 0 (GnRH I) and 2 (GnRH 2) hours. Duration (in minutes) of the computer-resolved LH secretory burst at half-maximal amplitude. The volume term (milliliters) represents unit distribution volume. Amplitude is the maximal rate of LH secretion attained in the deconvolution-resolved LH secretory burst. Entries within each GnRH I column identified by different superscripts (a and b) differ significantly (Duncan's multiple-range test of log transformed values, p < 0.05). *Comparisons of secretory burst properties and LH half-lives (GnRH I vs GnRH 2) were made by non parametric options (Wilcoxon signed-rank test): p < 0.01. tComparisons of secretory burst properties and LH half-lives (GnRH I vs GnRH 2) were made by non parametric options (Wilcoxon signed-rank test): 0.01 < P < 0.05.
(maximal secretory rate) and half-duration in response to the first and second challenges with GnRH. Analysis of the deconvolution-resolved secretory event characteristics confirmed this visual impression. Table II summarizes the LH secretory event characteristics inferred in response to the first bolus of GnRH as a function of the phase of the menstrual cycle. Although there was a trend toward a higher maximal LH secretory rate in response to GnRH during the late follicular versus the early follicular phase, a significant difference was not realized (p = 0.076). In contrast, the maximal secretory rate during the midluteal phase was significantly greater than that observed in the early follicular (p = 0.0002) or late follicular (p = 0.021) phases. However, during the late follicular phase the duration of the GnRH-associated LH secretory event was substantially longer than that documented during the early follicular (p = 0.022) and midluteal (p = 0.041 ) phases. The mass of LH released in response to GnRH was greater in both the late follicular and midluteal phases of the cycle compared with the early follicular phase (p = 0.002 and 0.0001, respectively), but the former two were indistinguishable (p = 0.376). Whereas the half-life estimates of LH released during the early follicular and late follicular phases were similar (p = 0.754), that of the material secreted in the midluteal phase was significantly shorter (p = 0.0001). Comparison of luteinizing hormone secretory response to GnRH boluses (GnRH 1 vs GnRH 2). Table II also shows the LH secretory event characteristics in response to a second challenge with a submaximal dose of GnRH administered 2 hours after the initial bolus. During the early follicular phase both the maximal rate of LH secretion and the mass of LH released were
greater (p < 0.01 and p < 0.05, respectively) in response to the second versus the first GnRH challenge. However, neither the secretory event duration nor the half-life of hormone secreted differed. The same pattern was evident during the midluteal phase [i.e., a higher maximal secretory rate (p < 0.05) and mass (p < 0.01)], but indistinguishable duration and half-life of LH were found when the response to the second GnRH bolus was compared with the response to the first. During the late follicular phase, both the maximal secretory rate and mass of hormone secreted were increased in response to the second challenge with GnRH (p < 0.05 and p < 0.01, respectively). However, during the late follicular phase, and in contrast to the results obtained in the early follicular and midluteal phases, the half-life of secreted LH in response to the second injection of GnRH was significantly (p < 0.01) shorter than that released in response to the first dose. No difference in the secretory burst duration was detected. Correlations among GnRH-stimulated secretory event characteristics and the circulating concentrations of the gonadal hormones. When all women were considered as a group across various stages of the menstrual cycle, multiple stepwise linear regression analysis demonstrated that serum estradiol concentrations were significantly predictive of the mass of LH secreted after the first GnRH injection (p < 0.0001 and R2 = 0.41 for mass). Serum estradiol levels also predicted the maximal rate of LH release after the first GnRH pulse (p = 0.004). Serum progesterone concentrations predicted the LH half-life (p < 0.004) and the halfduration (p < 0.02) of the first LH secretory burst. When individual phases of the menstrual cycle were considered alone, the following findings emerged: (1)
1532 Sollenberger et al.
November 1990 Am J Obstet Gynecol
LATE FOLLICULAR GnRH 2
GnRH 1
l
30r---------------~
o
~ 120 TIME (min)
TIME (min)
MID-LUTEAL
EARLY FOLLICULAR GnRH 1
GnRH 1
30,---------______~
120 TIME (min)
80r---------------~
fJ\
120
120
TIME (min)
TIME (min)
j\ 120 TIME (min)
Fig. 1. Illustrative profiles of the deconvolved LH secretory impulses in response to two GnRH challenges (GnRH 1 and GnRH 2) in a representative normal woman. For each phase of the menstrual cycle the upper panel depicts measured serum LH concentrations (mean ± SD) and the fitted reconvolution curve (continuous line), which is based on the calculated secretion and clearance parameters. The resolved LH secretory impulses are shown in the lower panel.
in the early follicular phase, serum estradiol levels were predictive of both the second luteinizing hormone secretory burst's mass (or its maximal rate) and its halfduration (p < 0.05, R2 ~ 0.48); (2) in the early follicular (but not late follicular) phase, serum progesterone concentrations predicted LH secretory burst mass after the first pulse of GnRH (p = 0.03, R2 = 0.42); (3) in the late follicular and midluteal phases, serum estradiol values correlated with the mass and the maximal rate of luteinizing hormone secreted after the first pulse of GnRH (p < 0.03, R2 ~ 0.48); (4) in the midluteal phase, serum 17J3-estradiol predicted both the mass of LH secreted and the half-duration ofthe second LH release episode (p < 0.03, R2 > 0.63); and (5) in the midluteal phase, serum progesterone concentrations were signif-
icantly predictive of the mass of LH secreted after the second GnRH pulse (p = 0.008, R2 = 0.56). Thus, in each of the different phases of the menstrual cyde, serum estradiol concentrations correlated positively and significantly with the mass and maximal rate of LH secretion attained in response to exogenous pulses of GnRH. Comment
Although the importance of the preovulatory LH surge within the context of normal reproductive function has long been recognized, the cellular events that signal the dramatically increased serum concentrations of LH at midcyde remain incompletely defined. Data obtained in animal models suggest that the secretion of
Volume 163 Number 5, Part I
GnRH into the hypothalamic-hypophysial portal circulation is markedly enhanced coincident with the preovulatory LH surge,I6 However, it is generally accepted that increased GnRH secretion does not solely account for the midcycle LH surge; rather it is thought that the surge reflects a situation in which increased concentrations of GnRH impinge on an anterior pituitary gland that has been rendered considerably more responsive to the releasing hormone just at midcycle. Accordingly, significant investigative interest has focused on the mechanisms subserving this alteration in gonadotroph responsivity. That GnRH-stimulated LH release is dependent on the gonadal hormone milieu has been documented in a variety of in vivo and in vitro studies. I-6 Of particular interest are the reports suggesting that, under appropriate conditions, enhanced LH release is observed in response to GnRH after an initial challenge with a "priming" dose of the hypothalamic peptide.'-6 7-'1 Although this phenomenon appears to be dependent on the gonadal hormone environment/' lI-14 the mechanisms involved remain poorly understood. In this study we coupled a moderately intensive sampling paradigm with analysis of the resultant LH concentration time series by use of a multiple-parameter deconvolution procedure to appraise in some detail the determinants of the GnRH secretory-priming event at several stages of the menstrual cycle. Our current results are consistent with previous studies in women that have documented that the stage of the menstrual cycle," presumably reflecting the gonadal hormone environment,' influences the serum concentrations of luteinizing hormone achieved in response to a single bolus of GnRH. The use of deconvolution precedures to "unravel" the GnRH-associated secretory event now provides an enhanced understanding of how this menstrual state-specific modulation occurs. Thus the maximal rate of LH secretion (as distinct from serum LH concentrations per se) achieved in response to GnRH in the midluteal phase is significantly greater than that observed in the early or late follicular phases. In contrast, the duration of the secretory episode in the late follicular phase is prolonged in comparison with the early follicular and midluteal phases. It should be noted that the duration of secretion is considerably shorter than that of the serum concentration peak. Thus the enhanced maximal secretory rate in the midluteal phase and the longer secretory event duration in the late follicular phase are responsible for the greater mass of hormone secreted in these two phases when compared with that secreted in the early follicular phase. Such information about secretion requires a technique such as deconvolution analysis, and cannot be inferred from simple inspection of the serum LH concentration peak. '5 Clearly, these studies do not address the question of which particular intracellular pro-
GnRH self-priming in normal women
1533
cesses subserve the menstrual stage-associated alterations in maximal LH secretory rate and secretory event duration. However, in consideration of this experimental design in which the hypothalamic peptide was administered exogenously, these data do demonstrate that short-term variations in the amount and/or duration of the GnRH stimulus per se are not required to effect variations in the characteristics of the LH secretory event. More prolonged variations in the characteristics of the endogenous GnRH pulse signal throughout the menstrual cycle may act to influence pituitary responsiveness. Our data also support earlier studies that demonstrated GnRH self-priming in the late follicularmidluteaF and midfollicular-midluteaF phases of the menstrual cycle. However, our results also provide evidence for a previously undescribed potentiating effect of GnRH on certain secretory event characteristics during the early follicular phase. Specifically, the maximal rate of LH secretion achieved in response to the second bolus of GnRH compared with the first was also enhanced in the early follicular phase and the midluteal phase. However, the duration of the secretory events in response to the two consecutive GnRH challenges did not differ. These results imply that it is the rate with which LH molecules are discharged from the pituitary gland within a given amount of time, rather than the duration of release itself, that results in the observed augmentation of LH secretion with repetitive GnRH administration. Serum estradiol concentrations were significantly predictive of the mass (as well as maximal rate) of LH secreted after the first GnRH injection in the late follicular and midluteal phases. This finding is consistent with an earlier inference by Lasley et al. 1 that estrogen participates in increasing intra pituitary LH stores that are available for release in response to secretagogue action. 1 Of additional interest is the finding that serum estrogen levels also correlated with the mass of LH released after the second pulse of GnRH in the early follicular phase. This observation suggests that both estrogen and prior exogenous GnRH exposure are required in the early follicular phase to demonstrate an increased availability of LH for release. Serum progesterone concentrations were negatively correlated with LH half-life and secretory burst duration, possibly indicating changes in the posttranslational carbohydrate modification of LH molecules with associated changes in the metabolic clearance rate of LHI7 and alterations in the kinetics of GnRH action and/or gonadotroph secretory responses, which result in a decreased secretory burst duration. In the midluteal phase, progesterone concentrations were related to the mass of LH released after the second GnRH pulse. Thus progesterone also may influence the self-priming actions of GnRH.
1534 Sollenberger et al.
The half-lives reported for endogenous LH correspond to half-lives of LH (calculated from stable metabolic clearance rates) of 123 minutes,!s 136 minutes,19 and 112 minutes,"" assuming a 3.5 L distribution volume for LH. Thus our LH half-life values of 60 to 120 minutes agree with independently estimated values obtained by steady-state infusion of various exogenous LH preparations. It should be noted that the deconvolution model used in this study provides estimates of LH half-life under steady-state conditions (or approximately its slower second component rate constant in non equilibrium studies) and does not reflect acute distribution coefficients (initial equilibration half-times after bolus injection), which could be calculated with even more rapid sampling. This observation presumably reflects the fact that LH distribution is virtually complete between consecutive (lO-minute) sample observations. Moreover, our results do not disagree with those of Wehmann et a!., IS who showed a moderate increase in metabolic clearance rate (or decrease in half-life) in the luteal phase. We thank Sandra Jackson and the nursing staff at the Clinical Research Center for their expert patient care and Catherine Kern and Ginger Bauer for technical assistance with the assays. REFERENCES 1. Lasley BL, Wang CF, Yen SSC. The effects of estrogen and progesterone on the functional capacity of the gonadotrophs. ] Clin Endocrinol metab 1975;41 :820-6. 2. Wang CF, Lasley BL, Lein A, Yen SSC. The functional changes of the pituitary gonadotrophs during the menstrual cycle.] Clin Endocrinol Metab 1976;42:718-28. 3. Pickering AJMC, Fink G. Priming effect of luteinizing hormone r~leasing factor: in vitro and in vivo evidence consistent with its dependence upon protein and RNA synthesis.J Endocrinol 1976;69:373-9. 4. Hoff JD, Lasley CL, Wang CF, Yen SSC. The two pools of pituitary gonadotropin: regulation during the menstrual cycle. J Clin Endocrinol Metab 1977;44:302-12. 5. Higuchi T, Kawakami M. Luteinizing hormone responses to repeated injections of luteinizing hormone-releasing hormone in the rate during the oestrous cycle and after ovariectomy with and without oestrogen treatment.] Endocrinol 1982;93: 161-8. 6. Waring DW, Turgeon]L. Luteinizing hormone-releasing hormone-induced luteinizing hormone secretion in vitro: cyclic changes in responsiveness and self-priming. Endocrinology 1980; 106: 1430-6.
November 1990 Am J Obstet Gynecol
7. Rommler A. Short-term regulation of LH and FSH secretion in cyclic women 1. Altered pituitary response to a second of two LH-RH injections at short intervals. Acta Endocrinol [CopenhJ 1978;87:248-58. 8. Aiyer MS, Chiappa SA, Fink G. A priming effect of luteinizing hormone-releasing factor on the anterior pituitary gland in the female rat.] Endocrinol 1974;62:57388. 9. Castro-Vaszquez A, McCann SM. Cyclic variations in the increased responsiveness of the pituitary to luteinizing hormone-releasing hormone (LHRH) induced by LHRH. Endocrinology 1975; 97: 13-8. 10. Fink G, Chiappa SA, Aujer MS. Priming effect of luteinizing hormone-releasing factor elicited by preoptic stimulation and by intravenous infusion and multiple injections of the synthetic decapeptide. ] Endocrinol 1976; 69:359-72. 11. Nazian SJ. Androgenic and estrogenic control of the selfpriming effect of LHRH in the castrated male rat. Neuroendocrinology 1986;42: 112-9. 12. Beck LV, Bay M, Smith AF, King D, Long R. Steroid priming of the luteinizing hormone response to luteinizing hormone-releasing hormone. ] Endocrinol 1978; 77:293-9. 13. Speight A, Fink G. Comparison of steroid and LH-RH effects on the responsiveness of hemipituitary glands and dispersed pituitary cells. Mol Cell Endocrinol 1981;24: 267-81. 14. Padmanabhan V, Leung K, Convey EM. Ovarian steroids modulate the self-priming effect of luteinizing hormonereleasing hormone on bovine pituitary cells in vitro. Endocrinology 1982;110:717-21. 15. Veldhuis JD, Carlson ML, Johnson ML. The pituitary gland secretes in bursts: appraising the nature of glandular secretory impulses by simultaneous multipleparameter deconvolution of plasma hormone concentrations. Proc Nat! Acad Sci USA 1987;84:7686-90. 16. Sarkar DK, Chiappa SA, Fink G, Sherwood N. Gonadotropin-releasing hormone surge in pro-oestrous rats. Nature 1976;264:461-3. 17. Dufau ML, Veldhuis ]D. Pathophysiological relationships between the biological and immunological activities of luteinizing hormone. In: Burger HG, ed. Balliere's clinical endocrinology and metabolism. Philadelphia: WB Saunders, 1987 vol 1:153-76. 18. Wehmann RE, Blackman MR, Harman SM. Metabolic clearance rates of luteinizing hormone in women during different phases of the menstrual cycle and while taking an oral contraceptive. J Clin Endocrinol Metab 1982; 55:654-9. 19. Marshall]C, Anderson DC, Fraser TR, Harsoulis P. Human luteinizing hormone in man: studies of metabolism and biological action.] EndocrinoI1973;56:431-9. 20. Pepperell~, DeKretser DM, Burger HG. Studies on the metabolic clearance rate and production rate of human luteinizing hormone and on the initial half-time of its subunits in man.] Clin Invest 1975;56:118-26.
A complete list of references is available from the authors on request.
