repetitive GnRH stimulation on pituitary responsiveness in vivo

Transcription

1 Pititary elf-priming Actions of Gonadotropin-releasing ormone Kinetics of stradiol's Potentiating ffects on Gonadotropin-releasing ormone-facilitated ormone and Follicle-stimlating ormone Release in ealthy Postmenopasal Women ILteinizing Johannes D. Veldhis, William. vans, Alan D. Rogol, Lisa Kolp, Michael 0. Thomer, and Pal tmpf* Departments ofinternal Medicine, Pharmacology, Pediatrics, and Obstetrics and Gynecology, University of Virginia chool ofmedicine, Charlottesville, Virginia 22908; and *Department ofobstetrics and Gynecology, Jersey hore Medical Center, Neptne, New Jersey Abstract We examined the kinetically distinct characteristics of estradiol's effects pon pititary lteinizing hormone (L) and folliclestimlating hormone (F) release in response to plses of exogenos gonadotropin-releasing hormone (GnR) in healthy postmenopasal individals. The ptative self-priming actions of GnR on L and F release were tested by intravenos injections of eqal paired doses of GnR (10 jig) before and after 1, 5, 10, and 30 d of pre estradiol-17,6 delivery via an intravaginal silastic ring. elf-priming actions of GnR, as defined by heightened gonadotropin release in response to the second plse of GnR compared with the first, were completely absent in the hypoestrogenemic state. owever, estradiol administration nmasked GnR self-priming in a time-dependent fashion, with maximal expression after 5 and 10 d of steroid replacement, followed by attenation by 30 d. ince estradiol's modlation of GnR action was expressed differentially on L and F release, we sggest that sch facilitation of GnRstimlated pititary L and F release may provide an additional mechanism for dissociated secretion of gonadotropic hormones in health or disease. Introdction Under certain physiological conditions, repetitive stimlation of the anterior pititary gland by serial plses of gonadotropinreleasing hormone (GnR)' reslts in marked potentiation of gonadotropin release (1-5). This ability of repetitive GnR stimlation to facilitate pititary responsiveness has been referred to as the "self-priming" action ofthis hypothalamic decapeptide (1). tdies in a variety of experimental animals have sggested that sch amplifying effects of serial GnR stimlation may be critically important to the genesis of the preovlatory srge-like release of gonadotropic hormones dring the final stages of folliclar development (1-8). Investigations in ovariectomized rodents sbjected to varios regimens of sex-steroid hormone replacement have implicated estradiol as one critical determinant of the facilitative effects of Address correspondence to Dr. Veldhis. Receivedfor pblication 14 November 1985 and in revisedform 21 Janary Abbreviations sed in this paper: F, follicle-stimlating hormone; GnR, gonadotropin-releasing hormone; L, lteinizing hormone. J. Clin. Invest. The American ociety for Clinical Investigation, Inc /86/06/1849/08 $ 1.00 Volme 77, Jne 1986, repetitive GnR stimlation on pititary responsiveness in vivo and in vitro (8-16). imilarly, short-term administration of estrogen to postmenopasal women is accompanied by altered pititary responsiveness to exogenosly infsed GnR (17, 18). ch alterations inclde either inhibition or facilitation ofgnr actions (17, 18). tdies in the folliclar phase of the hman menstral cycle frther sggest that injected estradiol and/or one or more events associated with folliclar matration can reslt in either diminished or enhanced pititary responses to exogenos GnR stimli (19-24). owever, the ability to relate sch alterations in pititary responsiveness explicitly to estradiol is limited in gonadally intact individals. In addition, available data have not yet delineated: (a) the kinetics of estrogen's elicitation of self-priming actions ofgnr; (b) the extent to which acte responses to single GnR injections and GnR selfpriming represent temporally distinct events; and (c) the differential self-priming effects of GnR on L and F release. In the present stdy, we have sed a model of physiological estradiol replacement and paired exogenos GnR plses to test for precise temporal correlations between circlating estradiol levels, basal gonadotropin concentrations, and the GnR-facilitated release of L and F. To obviate the confonding inflences ofnstable serm estradiol concentrations that reslt after oral or intramsclar estrogen dosing, we have sed an estradiol-impregnated silastic ring placed intravaginally. The latter mode of steroid-hormone delivery reslts in the rapid attainment of steady state serm estradiol concentrations commensrate with those of the mid-to-late folliclar phase of the normal menstral cycle. This paradigm has permitted s to elcidate kinetically distinct characteristics of estradiol's effects on pititary lteinizing hormone (L) and follicle-stimlating hormone (F) release in response to exogenos GnR plses in previosly hypoestrogenemic postmenopasal women. Methods Vaginal rings. stradiol-containing silastic rings were prepared exactly as described earlier, with a dose of 400 mg of estradiol-17,7 impregnated in each ring (25). bjects. ealthy spontaneosly postmenopasal women were stdied after provision ofwritten informed consent, approved by the man Investigation Committee ofthe University ofvirginia chool ofmedicine. The sbjects who participated in this stdy ranged in age from 55 to 63 yr (mean, 58±2 yr, n = eight women) and were 3-9 yr postmenopasal. ach volnteer nderwent a detailed history and physical examination, with the docmentation ofnormal hepatic, renal, and hematologic fnction, biochemical ethyroidism, and postmenopasal concentrations of gonadotropic hormones. At least 5 wk before stdy, women were withdrawn from any drgs, inclding estrogen or sex-steroid hormone treatments. Blood sampling protocols. ampling was condcted in the Clinical Research Center of the University of Virginia by withdrawing 2.5 ml stradiol Modlates Gonadotropin-releasing ormone elf-priming 1849

2 blood at 15-min intervals for 6 h beginning at 0800 h. After 2 h of basal sampling, a dose of 10 g GnR was administered intravenosly by bols injection. After two more hors, a second dose of 10 g GnR was given similarly. The first dose was sed to appraise acte pititary responsiveness. The paired plses of GnR were sed to test for the emergence and/or disappearance of self-priming actions of GnR. elfpriming was defined as an increase in pititary responsiveness to the second dose ofgnr compared with the first (see below: data analysis). ampling was performed basally (no estrogen treatment), and on days 1, 5, 10, and 30 after initiation of estradiol replacement. Assays. erm concentrations of L and F were assayed in dplicate with a dal-label radioimmnoassay (RIA) kit (Clinetics Corporation, Tsten, CA). The sensitivities for L and F were 1.8 miu/ ml and 1.4 miu/ml, respectively. amples were dilted 1:2 or 1:4 to fall within the least variable region of the displacement crve, where the intraassay coefficients of variation averaged %, and interassay variability was 9-14% (L) and 5-13% (F). The cross-reactions of L and F with alpha sbnit in this assay were <10%. erm estadiol levels were qantitated by specific RIA after celite chromatography (26). Data analyses. Data are expressed as means±m for the grop of eight volnteers. ignificant overall treatment effects were soght by analysis of variance with the Newman-Kel's procedre to test for individally significant effects (27). Where indicated, specific a priori comparisons of mean, incremental, or absolte peak (maximal) gonadotropin concentrations were made by paired two-tailed tdent's t testing with Bonferonni's correction (27). Fractional (percentage) increases in gonadotropin concentrations for the first and second GnR-stimlated gonadotropin peaks were compared by the Wilcoxon signed ranks test (27). To assess the possible contribtion of mltiple parameters to the magnitde of the gonadotropin response to the second plse of GnR, mltivariate analysis was employed. Mltiple linear regression was performed sing the mean peak 2 gonadotropin level as the dependent variable, and the following individal independent variables: (a) the mean basal serm gonadotropin concentration over the 2 h preceding GnR administration; (b) the mean gonadotropin response to the first GnR plse; and (c) the simltaneos serm estradiol concentration. In this statistical model, individally significant correlations as well as partial and mltiple linear correlations were soght. This permitted s to evalate the most significant contribtors to the amplitde of the gonadotropin response to the second GnR plse at varios times after (or before) estradiol administration. Reslts erm estradiol concentrations erm estradiol concentrations in the eight women varied significantly over time (P < 0.001) in the manner depicted in Fig. 1. There was an approximately 25-fold increase in mean serm estradiol concentrations within 24 h of intravaginal placement of the estradiol-impregnated silastic ring. On day 5, this vale declined to a significant degree (P < 0.01), and then remained stable from day 5 throgh 30 (serm estradiol concentrations on days 5, 10, and 30 did not differ significantly). Time-dependent inflences ofestradiol on basal and GnR-stimlated L and F concentrations The temporal profiles of mean serm L and F concentrations derived from eight women sampled at 15-min intervals for 6 h before estradiol treatment (basal day 0) and on days 1, 5, 10, and 30 ofestradiol administration are shown in Fig. 2 A and B. In each panel, the crve depicting mean serm L and F concentrations on day 0 is reprodced for comparison with levels observed on the varios days of estradiol administration. To assess time-dependent changes in baseline (pre-gnr) i J ;U DURATION OF TRADIOL RPLACMNT (DAY) Figre 1. erm estradiol concentrations attained after intravaginal placement of a silastic ring containing 400 mg pre crystalline estradiol- 1 7#. erm concentrations of estradiol were measred basally (day 0) and on day 1, 5, 10, and 30 of estradiol replacement. Data are means±m (n = 8 women). Different sperscripts denote significantly different serm estradiol concentrations. gonadotropin concentrations in relation to estradiol replacement, means for the 2-h intervals preceding GnR injections were analyzed over the five stdy sessions. As shown in Fig. 3 (top), baseline serm L concentrations declined within 24 h of estradiol replacement (P = treatment effect). Mean serm L concentrations after the first dose of GnR ( , peak 1) and after the second dose of GnR ( , peak 2) are also smmarized in Fig. 3 (middle and bottom). Mean vales for peak 1 and peak 2 exhibited significant biphasic changes in response to estradiol replacement (P = for peak 1, and P = for peak 2). Mean 2-h baseline serm F concentrations also decreased significantly within 24 h of estradiol replacement (P < treatment effect over time): Fig. 3. Mean F concentrations in peak 1 and peak 2 (bottom left) similarly declined progressively dring the time corse of estradiol replacement (P < 0.001). Fig. 3 also contrasts the patterns observed for mean L and F concentrations over time. Baseline L and F concentrations differed significantly from each other, with F levels exceeding those of L at all times except on day 10. On the other hand, mean GnR-stimlated gonadotropin peak 1 vales were similar for L and F at all time points, except on day 30 when F peak 1 vales exceeded corresponding L peak 1 levels significantly. In addition, mean GnR-stimlated L peak 2 vales significantly exceeded those of F peak 2 on days 5 and 10 (bt not on days 1 or 30) of estradiol administration, exemplifying the prominent self-priming actions ofgnr on L release (discssed frther below). Inflence ofestradiol on the self-priming action ofgnr: comparison ofpeak 2 and peak I properties To evalate the self-priming actions of GnR, the properties of peak 2 were compared with those of peak 1. elf-priming by GnR was defined as a significantly greater gonadotropin response to the second plse of GnR (peak 2) compared to the first (peak 1). We have compared peak 2 and peak 1 in relation to the following characteristics: (a) mean (2-h) gonadotropin concentrations; (b) absolte maximal gonadotropin concentrations attained within the peak (miu/ml); (c) incremental (miu/ ml) increases; and (d) percentage increases. These separate analyses of the relationship of peak 2 to peak 1 have permitted s 1850 Veldhis, vans, Rogol, Kolp, Thorner, and tmpf

3 b3sa A ' 49 ii ' II Clod. Time &Bsal verss Dag 1 Basal verss Day '1'0 *11* Clocd Tam 14.' 129 r 1t L 6 4. A~~~~~.t4~~ ~ 9.13.ti Clodc Time l ves D 19 Basal verss Day r 19 L 6C 4.4* 4. -, , a,. '.., i Clode Tim 1t 129 1it 9 L 69 I' 4, I'.4 28 B Clock Tame B Basal IL.g 129 Cr 199, of a 199 t 60 Bsal verss Dao I.*4IJ,.' '*+t,t t1 ' 41 8e 9 1i t1 12 Clock Taim 8 9 ie 1 ' "l, 14 clock r Figre 2. Individal time-corses for serm L (A) and F (B) concentrations in postmenopasal women sampled at 15-min in- tervals for 6 h before estrogen replacement and at varios intervals after intravaginal BI vms D9 5 placement of an estradiol-containing silastic 129 t t t ffi ring. The mean serm immnoactive L /NA t tt healthy postmenopasal women were deter- 99 t mined from blood samples drawn at 15- min intervals over 6 h. The first 2 h of sam- 199 > Y ; g: and F concentrations (miu/ml) for eight 9, ; pling represented basal conditions ( t-t IOM1000 h). Thereafter (1000 h), 10 Ag of 4C! GnR &" were administered by intravenos Clocd Timw bols injection, which was repeated 2 h later (1200 h). This schedle of sampling 12a v and GnR administration was repeated on five occasions: a basal day (before the administration of estradiol), and on days 1, 5, 9.' 10, and 30 after intravaginal placement of a silastic ring containing 400 mg of pre es- 69 : X t t l l:tradiol- 1 7,. In each figre, the basal (pre-estradiol) time-corse for L or F is tt 14 t4 shown for comparison (solid circles and ? open triangles). Data are presented as Clol Tame means±m (n = 8 sbjects). stradiol Modlates Gonadotropin-releasing ormone elf-priming 1851

4 0 U) c 0cx 0 co 0 I c 0 c a ' I,,j N 40- \ _ *- L BAAL *--* F PAK 1 N N PAK NI \ 100- '''''''' \ N Dration of stradiol (Days) Figre 3. Comparison of time-dependent changes in mean serm immnoactive F and L concentrations before and dring estradiol administration. 2-h mean serm concentrations of F and L (miu/ml) are depicted for baseline (pre-gnr) conditions (top) and for GnR-stimlated gonadotropin peak 1 (middle) and peak 2 (bottom). Data were obtained before estradiol replacement (day 0), and on days 1, 5, 10, and 30 of estradiol administration via an intravaginal silastic ring. *P < 0.05 for the comparison of L and F (N, not significant). * Peak 1 X Peak _150- P<0 _1 P<0 00 P< 001o P< o RN O <T I R ( ~P<0,, 000I 1ilO ~~~~~~P< DURATION OF TRADIOL RPLACMNT (DAY) A to search for significant time-dependent effects of estradiol on one or more specific properties of the GnR-stimlated gonadotropin peak. Mean and maximal amplitdes ofgonadotropin peak I and peak 2. Mean L concentrations in peak 1 compared with peak 2 are presented in Fig. 4 A. Peak 1 and peak 2 are compared nder estrogen-deficient conditions (day 0), and on days 1, 5, 10, and 30 of estradiol replacement. As shown in the left panel of Fig. 4 A, in the absence of estradiol, peak 1 and peak 2 mean amplitdes were statistically indistingishable. In particlar, the mean (±M) L concentration for peak 1 was 93.7±10.4 miu/ ml vs. 99.9±8.7 miu/ml for peak 2. Ths, we cold demonstrate no self-priming action of GnR in the absence of estradiol replacement. In contrast, after 24 h of estradiol administration, mean L peak 2 concentrations significantly exceeded those of peak 1 (P < 0.001). This self-priming pattern was also observed on days 5, 10, and 30 of estradiol replacement (P < 0.001). Moreover, since samples were withdrawn at eqally spaced intervals (every 15 min), the mean vales discssed above are directly proportional to the effective area nder the GnR-stimlated L peaks. This implies that estradiol inflences integrated gonadotropin responses to the first and second GnR plses in different ways. Maximal (absolte peak) L concentrations attained within peak 1 and peak 2 also did not differ significantly nder basal conditions (before estradiol replacement) (Fig. 4 A, right). owever, within 24 h of estradiol administration, the absolte peak vales of L achieved in response to the second GnR stimls significantly exceeded those elicited by the first stimls (P < 0.001). This pattern was sstained to a significant degree on days 5, 10, and 30 (P < to P < 0.001). In the case of F, mean and maximal (absolte peak) F concentrations were also compared for peak 1 and peak 2 before and at varios times after estradiol replacement (Fig. 4 B). Like L, the magnitde of F peak 2 and peak 1 did not differ significantly in the absence of estradiol. owever, within 24 h of estradiol replacement, the mean and maximal amplitdes of DURATION OF TRADIOL RPLACMNT (DAY) Figre 4. Impact of estradiol replacement on serm gonadotropin concentrations associated with GnR-stimlated gonadotropin peak 1 or peak 2. Peak 1 and peak 2 were defined as L (A) and F (B) concentrations after the first or second injections of 0,g GnR. The left panel gives the mean hormone concentration (averaged over 2 h), and the right panel gives the maximal (absolte peak) gonadotropin concentration. P vales denote significant differences between peak 1 and peak 2 amplitdes. F peak 2 significantly exceeded those of peak 1 (P < 0.001). ch differences were sstained for days 5, 10, and 30 of estradiol replacement (P = 0.05 to P < 0.001). Incremental andjfractional amplitdes ofgonadotropin peak I and peak 2. The amplitdes of the first and second GnRstimlated gonadotropin peaks were compared in relation to incremental (miu/ml difference) and percentage (fractional) increases, which were defined as follows. The incremental increase for peak 1 was taken as the mean vale of peak 1 mins the mean vale of the corresponding baseline ( ) hormone concentration. The incremental vale for peak 2 was defined as the mean vale of peak 2 mins the nadir for peak 1 (measred at 1200, jst before injection of the second dose of GnR). The percentage increases for peak 1 and peak 2 were defined similarly; viz., as the ratio of mean peak 1 to mean baseline concentrations, and the ratio of mean peak 2 to peak 1 nadir concentrations. As smmarized in Fig. 5 A, before estradiol replacement (day 0), the increment of L peak 1 above baseline significantly exceeded that of L peak 2 above the peak 1 nadir (P < 0.003). owever, within 24 h of treatment with estradiol, the increase of L peak 2 over L peak 1 nadir was agmented, while the increase of peak 1 over basal declined. Moreover, by day 5 of estradiol replacement, the increment of peak 2 over peak 1 nadir was significantly greater than the increment of peak 1 over basal (P < 0.025). This agmentation ofthe peak 2 amplitde occrred to a lesser and insignificant degree on days 10 and 30 of estradiol 1852 Veldhis, vans, Rogol, Kolp, Thorner, and tmpf

5 * Peak 1 over baseline X Peak 2 over peak 1 nadir A Table I. Regressions ofmean GnR-stimlated Gonadotropin Peak 2 on Peak I at Varios Times after stradiol Replacement a c I -J T 160 P.0.05 N' N N 003 PO0O05OT ]P N N h ~~~ar 40 co DURATION OF TRADIOL RPLACMNT (DAY) Dration of estradiol replacement L* Ft days ± ± ± ± ± ± ± ± ± ± * Peak 1 over baseline v Peak 2 over peak 1 nadir B * ach linear regression coefficient was significant at P < t ach linear regression coefficient was significant at P < lopes of the linear regression of mean peak 2 on mean peak 1. Data are means±m for the slope of the linear regression (n = eight women stdied at each time) ' 10 co U DURATION OF TRADIOL RPLACMNT (DAY) Figre 5. ffects of estradiol pon GnR-stimlated gonadotropin peak 1 and peak 2 amplitdes expressed as incremental or percentage increases. Data are presented as in Fig. 4, except that the amplitdes of peak 1 and peak 2 are defined either as incremental (miu/ml difference) or as percentage increases. For peak 1, the incremental and percentage increases were measred with respect to mean basal concentrations. For peak 2, the incremental and percentage increases were measred relative to the nadir of peak 1 (taken as 1,200 h, jst before the second dose of GnR). Reslts are shown for L (A) and F (B) in relation to the dration (days) of estradiol replacement. Data are means+m (n = 8 sbjects). replacement. A generally similar pattern relating peak 2 and peak 1 was observed in relation to estradiol replacement when peak amplitdes were expressed as percentage increases (right panel of Fig. 5 A). An analogos temporal pattern was observed for incremental or percentage increases when F peak 1 and F peak 2 were compared (Fig. 5 B). Mltivariate analysis ofprincipal contribtors to the amplitde ofthe gonadotropin peak 2 The correlations between mean peak 2 and peak 1 gonadotropin concentrations were appraised for each of the five individal stdy sessions (day 0, and days 1, 5, 10, and 30 of estradiol treatment). In each stdy session, mean peak 2 and peak 1 concentrations were significantly correlated for L as well as for F (P < 0.001). Ths, five individal regression coefficients related mean peak 2 to mean peak I vales for each gonadotropic hormone (Table I). Within 24 h of estradiol treatment, the regression coefficient (i.e., slope of the line traversing the origin that related mean L peak 2 to mean L peak 1 concentrations) increased significantly, and contined to increase to a maximal vale on day 10 of estradiol treatment. The overall profile for F was similar, bt at each time after estrogen treatment the slope of the regression of F peak 2 on F peak 1 was less than that for L. Ths, the self-priming action ofgnr (defined here by the relationship of peak 2 to peak 1) depended significantly pon the dration of estradiol treatment. Moreover, GnR self-priming was more prominent for L than F. The mltivariate relationship between mean L peak 2 concentrations and the following other parameters was also assessed: baseline L, mean L peak 1, and serm estradiol concentrations. This permitted s to test the relative contribtions of these parameters to the magnitde of L peak 2 at varios times before and after estradiol replacement. As shown by the correlation coefficients given in Table II (top), the mean amplitde of L peak 2 on day 0 (no estradiol) was correlated significantly to three factors: basal L, mean L peak 1, and estradiol concentrations (P = 0.013). The mean amplitde of L peak 2 on day 0 was associated most significantly with two parameters: baseline and peak 1 concentrations (P = 0.002). In Table II. Relation ofgonadotropic Peak 2 to Basal, Peak I and/or erm stradiol Concentrations Partial and mltiple correlation Basal, peak I Basal and Basal and Peak I and Condition and estradiol peak I estradiol estradiol day * L * 5 L * 10 L N * N 30 L N * F N * N F * F * F N * N F N * N, P > 0.05 (not significant): P < 0.05 for all other correlations. * Greatest F ratio for that day. stradiol Modlates Gonadotropin-releasing ormone elf-priming 1853

6 contrast, after 1 and 5 d of estradiol administration, the amplitde of L peak 2 was correlated most significantly to mean L peak 1 and serm estradiol concentrations (P < 0.001). On days 10 and 30 ofestradiol administration, peak 2 was described best by its relationship to baseline and L peak 1 concentrations (P = 0.003). Ths, the statistical correlates ofthe mean amplitde of L peak 2 varied in a distinctive manner over time, with peak 1 amplitde and serm estradiol concentrations being most inflential on days 1 and 5, while baseline and peak 1 concentrations were most contribtory on days 0, 10, and 30 ofestrogen administration. The preceding overall pattern for L differed from that observed in the case of F (bottom, Table II), since basal F and mean F peak 1 concentrations provided the best statistical correlates of the amplitde of F peak 2 for days 0, 1, 5, and 10 after estradiol (P = to P < 0.001). By day 30 of estradiol, baseline F and the serm estradiol concentrations represented the most significant correlates ofmean F peak 2 levels. These analyses indicate that the correlates of L and F responsiveness to paired GnR plses are distinctively inflenced by the relative and time-dependent contribtions of baseline gonadotropin concentrations, mean peak 1 gonadotropin concentrations, and concrrent estradiol levels. Inflence ofestradiol on total incremental gonadotropin release in response to paired exogenos GnR plses The time-dependent effects ofestradiol on total GnR-promoted gonadotropin release were estimated by determining the sm of peak 1 and peak 2 increments over basal before and at varios times after estradiol administration. ch estimates of total incremental L and F release are given in Table III for days 0, 1, 5, 10, and 30 of stdy. For both L and F, total incremental gonadotropin release was relatively redced on day 30. In contrast, on day 10 of estradiol treatment, the total incremental vale for GnR-stimlated L release was significantly greater than that on day 0 (pre-estrogen) or day 30. imilarly, for F, the sm of the peak 1 and peak 2 increments on days 5 and 10 significantly exceeded that on day 30. Ths, estimated total L and F release in response to exogenos paired GnR plses varies significantly in relation to the dration of estradiol replacement. Table III. Time-dependent Inflence ofstradiol on Total Incremental Gonadotropin Release in Response to Paired Plses ofxogenos GnRh Dration of estradiol treatment Total L increments Total F increments days miu/mi* miu/mi* 0 109±14 64± ±16 58± ±14 72± ±10t 68± ±15 52±5.4 * Data are means±m (n = eight sbjects) for the total incremental amplitdes (miu/ml) of GnR-stimlated gonadotropin peaks 1 and 2 at the indicated times before and after estradiol administration. * P = vs. day 0 or day 30. P = vs. day 30. Discssion Or reslts clearly show that estradiol differentially reglates the responsiveness of anterior pititary L and F release to exogenos GnR plses, and that these actions of estradiol are critically time-dependent. Ths, in the estrogen-deprived state, L reponses to the first and second of paired GnR plses did not differ significantly. owever, within 24 h of estradiol administration, the mean and maximal vales of L elicited by the second GnR stimls significantly exceeded those evoked by the first stimls. This pattern of increased L release in response to the second, compared with the first, plse of releasing factor exemplifies the self-priming action of GnR (1-5). Or appraisal of the detailed time-corse of estradiol's inflence on this self-potentiating action of GnR indicates the dynamic natre of estrogen action, which is characterized by the emergence of maximal GnR self-priming within 5-10 d of increased circlating estradiol concentrations, followed by an attenation of GnR self-priming after 30 d of continos estradiol exposre. Moreover, increased GnR-stimlated L release after the second plse of releasing factor was accompanied by parallel changes in total gonadotropin release in response to both GnR plses, indicating that self-priming did not simply reflect redistribtion of L release from a diminishing peak 1 to an expanding peak 2. Of additional interest, the biphasic pattern of emergence and sbseqent attenation of GnR selfpriming occrred despite nchanging serm estradiol concentrations over days 5-30 of estrogen administration. The intravaginal rote of estradiol delivery via polysiloxane vehicle reslts in niform and selective elevation of serm estradiol concentrations into the normal physiological range characteristic of the mid- to late folliclar phase, with a lesser rise in serm estrone concentrations (approximately twofold increase) (28). In response to this estrogenic milie, serm free testosterone concentrations decline significantly within 5 d and retrn to baseline by day 10 (28). These steroid hormone changes sggest that the self-priming actions of GnR observed on day 5 of estradiol replacement might reflect the combined impact of an increase in circlating estradiol concentrations as well as a dimintion in androgen negative-feedback effects associated with declining serm free testosterone concentration. Althogh the present data do not permit s to distingish between these two possibilities nambigosly, the inflence of decreased serm free testosterone levels appears to be relatively minimal, since prominent self-priming actions of GnR were also observed on day 10 of estradiol administration, when plasma concentrations of free testosterone, total testosterone, dehydroepiandrosterone slfate, and androstenedione are no different from basal (before estradiol administration) (28). Ths, the dominant steroidal correlate of GnR self-priming is estradiol per se. As sch, these reslts may be pertinent to nderstanding estradiol's reglation of gonadotropin secretory patterns. owever, the present model may not necessarily apply in all its details to yonger, normally menstrating women. The ability of estradiol to amplify pititary responsiveness to paired exogenos plses of GnR in a time-dependent manner cold be observed whether GnR-stimlated gonadotropin peaks were appraised as mean, maximal, incremental, fractional, or total increases above baseline. Althogh estrogen does not seem to inflence the metabolic clearance of GnR per se (17), the effects of estradiol on GnR-stimlated L and F release cold be modified by estrogen-associated alterations in rates of 1854 Veldhis, vans, Rogol, Kolp, Thomer, and tmpf

7 gonadotropin metabolic clearance (29, 30). owever, estrogen's sppression of mean plasma gonadotropin concentrations wold actally tend to accelerate L clearance, since the metabolic clearance of L increases at lower serm hormone concentrations (31). This increase in the rate of gonadotropin removal from the circlation wold actally render the detection of GnR self-priming more difficlt. Moreover, or analysis of incremental and fractional increases in peak 2 relative to the nadir of peak 1, which provides the least favorable conditions for detecting self-priming effects of GnR, still reveals preferential agmentation of GnR-stimlated gonadotropin peak 2 over peak 1. ch self-priming was maximal on day 5 ofsstained estrogen administration. Ths, these different bt complementary analyses sggest that estradiol amplifies pititary responsiveness to paired exogenos plses of GnR, and that sch amplifying actions of estradiol emerge in a distinct time-dependent fashion in previosly hypoestrogenemic postmenopasal women. stradiol administration potentiated both GnR-stimlated L and F release. In comparing the individal time corses of the self-priming actions ofgnr on L and F secretion, we observed a consistently more prominent facilitative effect of estradiol on GnR-stimlated L than F release independently of how the data were expressed. owever, the overall temporal profile ofestradiol's potentiation of GnR action was analogos for L and F, with maximal GnR self-priming of gonadotropin release observed after 5 and 10 d of estradiol replacement. The statistical correlates of GnR self-priming varied over time and were also distingishable for L and F. In relation to maximal GnR self-priming of L release (day 5), the amplitde of GnR-stimlated peak 2 cold be acconted for predominantly by the amplitde of corresponding L peak 1 and the simltaneos serm estradiol concentration. In contrast, for F, the predominant predictors of peak 2 amplitde were basal and peak 1 F concentrations. ch mltivariate analyses indicated that the individal correlates of GnR-stimlated L and F release in response to estradiol administration were temporally distingishable for the two gonadotropic hormones, at least in postmenopasal individals. This may sggest that different pititary mechanisms operate to reglate GnR selfpriming of L and F. owever, independently ofthe precise mechanisms proposed, or observations docment significant differences in the relative self-priming actions of GnR on L and F release. We sggest that sch differences may provide an additional mechanism for dissociated release of gonadotropic hormones nder conditions of health or disease. Acknowledgments We thank Chris McNett for her skillfl preparation of the manscript; Pala P. Azimi for the artwork; National ormone and Pititary Program for the provision of prified hman L; the Clinetics Corporation for its grant-in-aid in spport of the gonadotropin assay kits;. lizabeth Taylor and Rebecca Weaver for technical spport; andra Jackson and the expert nrsing staff at the Clinical Research Center at the University of Virginia; and Mr. David Boyd for valed assistance with Clinfo. This work was spported in part by National Instittes of ealth grant No. RR to the Clinical Research Center of the University of Virginia, by RCDA 1 K04 D (JDV), D (MOT), D 0439 (W), Diabetes and Research Training Center grant No. 5 P60 AM , and National Instittes of ealth-spported Clinfo Data Redction ystems. References 1. Aiyer, M..,. A. Chiappa, and G. Fink A priming effect of lteinizing hormone releasing factor on the anterior pititary gland in the female rat. J. ndocrinol. 62: Castro-Vaszqez, A., and. M. McCann Cyclic variations in the increased responsiveness of the pititary to lteinizing hormonereleasing hormone (LR) indced by LR. ndocrinology. 97: Fink, G.,. A. Chiappa, and M.. Aiyer Priming effect of lteinizing hormone releasing factor elicited by preoptic stimlation and by intravenos infsion and mltiple injections of the synthetic decapeptide. J. ndocrinol. 69: Pickering, A. J. M. C., and G. Fink Priming effect of lteinizing hormone releasing factor in vitro and in vivo evidence consistent with its dependence pon protein and RNA synthesis. J. ndocrinol. 69: Pickering, A. J. M. C., and G. Fink Priming effect of lteinizing hormone releasing factor with respect to release of follicle-stimlating hormone in vitro and in vivo. J. ndocrinol. 75: Waring, D. W., and J. L. Trgeon Lteinizing hormonereleasing hormone-indced lteinizing hormone secretion in vitro: Cyclic changes in responsiveness and self-priming. ndocrinology. 106: Gordon, J.., and. Reichlin Changes in pititary responsiveness to lteinizing hormone-releasing factor dring the rat estros cycle. ndocrinology. 94: peight, A., R. Popkin, A. G. Watts, and G. Fink Oestradiol- 17,B increases pititary responsiveness by a mechanism that involves the release and the priming effect oflteinizing hormone releasing factor. J. ndocrinol. 88: Vilchez-Martinez, J. A., A. Arimra, L. Debeljk, and A. V. chally Biphasic effect of estradiol benzoate on the pititary responsiveness of LR. ndocrinology. 94: Libertn, C., R. Orias, and. M. McCann. Biphasic effect of estrogen on the sensitivity of the pititary to L-R (LRF). ndocrinology. 94: Droin, J., L. Lagace, and F. Labrie stradiol-indced increase of the L responsiveness to L releasing hormone (LR) in rat anterior pititary cells in cltre. ndocrinology. 99: Goodman, R. L., and. Knobil The sites of action of ovarian steroids in the reglation of L secretion. Neroendocrinology. 32: Clayton, R. N., A. olano, A. Garcia-Vela, M. L. Dfa, and K. J. Catt Reglation of pititary receptors for gonadotropinreleasing hormone dring the rat estros cycle. ndocrinology. 107: igchi, T., and M. Kawakami Lteinizing hormone responses to repeated injections oflteinizing hormone releasing hormone in the rat dring the oestros cycle and after ovariectomy with or withot oestrogen treatment. J. ndocrinol. 93: Padmanabhan, V., K. Leng, and. M. Convey Ovarian steroids modlate the self-priming effect of lteinizing hormone-releasing hormone on bovine pititary cells in vitro. ndocrinology. 110: vans, W.., D. R. Uskavitch, D. L. Kaiser, P. ellmann, J. L. Borges, and M. 0. Thorner The self-priming effect ofgonadotropinreleasing hormone on lteinizing hormone release: Observations sing rat anterior pititary fragments and dispersed cells continosly perifed in parallel. ndocrinology. 114: Keye, W. R., and R. B. Jaffe Modlation of pititary gonadotropin response to gonadotropin-releasing hormone by estradiol. J. Clin. ndocrinol. Metab. 38: Lasley, B. L., C. F. Wang, and.. C. Yen The effects of estrogen and progesterone on the fnctional capacity ofthe gonadotrophs. J. Clin. ndocrinol. Metab. 41: Yen,.. C., G. Vandenberg, R. Rebar, and Y. hara Variation in pititary responsiveness to synthetic LRF dring different phases of the menstral cycle. J. Clin. ndocrinol. Metab. 35: stradiol Modlates Gonadotropin-releasing ormone elf-priming 1855

8 20. haw, R. W., W. R. Btt, D. R. London, and J. C. Marshall Variation in response to synthetic lteinizing hormone-releasing hormone (LR) at different phases of the same menstral cycle in normal women. J. Obstet. Gynaecol. Br. Commonw. 81: Jaffe, R. B., and W. R. Keye, Jr strdiol agmentation ofpititary responsiveness to gonadotropin-releasing hormone in women. J. Clin. ndocrinol. Metab. 39: Keye, W. R., Jr., and R. B. Jaffe trength-dration characteristics of estrogen effects on gonadotropin response to gonadotropin releasing hormone in women. I. ffects of varying dration of estradiol administration. J. Clin. ndocrinol. Metab. 41: Yong, J. R., and R. B. Jaffe trength-dration characteristics ofestrogen effects on gonadotropin-releasing hormone in women. II. ffects of varying concentrations of estradiol. J. Clin. ndocrinol. Metab. 42: haw, R. W., W. R. Btt, and D. R. London The effect ofoestrogen pretreatment on sbseqent response to lteinizing hormone releasing hormone in normal women. Clin. ndocrinol. 4: tmpf, P. G., J. Marca, R. J. anten, and L. M. Demers Development of a vaginal ring for achieving physiologic levels of 17Bestradiol in hypoestrogenic women. J. Clin. ndocrinol. Metab. 54: amojlik,., J. D. Veldhis,. A. Wells, and R. J. anten Preservation of androgen secretion dring estrogen sppression with aminogltethimide in the treatment of metastatic breast carcinoma. J. Clin. Invest. 65: Winer, B. J tatistical Principles in xperimental Design. McGraw-ill, Inc., New York Veldhis, J. D.,. amojlik, W.. vans, A. D. Rogol, C.. Ridgeway, W. F. Crowley, L. Kolp,. Checinska, M. A. Kirschner, M. 0. Thorner, and P. tmpf ndocrine impact ofpre estradiol replacement in postmenopasal women: alterations in anterior pititary hormone release and circlating sex steroid hormone concentrations. Am. J. Obstet. Gynecol. In press. 29. Kohler, P. O., G. T. Ross, and W. D. Odell Metabolic clearance and prodction rates of hman lteinizing hormone in preand post-menopasal women. J. Clin. Invest. 47: Pepperell, R. J., D. M. de Kretser, and. G. Brger tdies on the metabolic clearance rate and prodction rate of hman lteinizing hormone and on the initial half-time of its sbnits in man. J. Clin. Invest. 56: Veldhis, J. D., F. Fraioli, A. D. Rogol, and M. L. Dfa Metabolic clearance of biologically active lteinizing hormone in man. J. Clin. Invest. 77: Veldhis, vans, Rogol, Kolp, Thorner, and tmpf