Endotoxin Disrupts the Estradiol-Induced Luteinizing Hormone Surge ...

25 downloads 25 Views 425KB Size Report
EDMUND TANHEHCO, CATHERINE VIGUIE´, AND FRED J. KARSCH. Department of Physiology (D.F.B., F.J.K.), Reproductive Sciences Program, University of ...
0013-7227/99/$03.00/0 Endocrinology Copyright © 1999 by The Endocrine Society

Vol. 140, No. 6 Printed in U.S.A.

Endotoxin Disrupts the Estradiol-Induced Luteinizing Hormone Surge: Interference with Estradiol Signal Reading, Not Surge Release* DEBORAH F. BATTAGLIA, ANDREW B. BEAVER, THOMAS G. HARRIS, ´ , AND FRED J. KARSCH EDMUND TANHEHCO, CATHERINE VIGUIE Department of Physiology (D.F.B., F.J.K.), Reproductive Sciences Program, University of Michigan, Ann Arbor, Michigan 48109 ABSTRACT Three experiments were conducted to investigate whether the immune/inflammatory stimulus endotoxin disrupts the estradiolinduced LH surge of the ewe. Ovariectomized sheep were set up in an artificial follicular phase model in which luteolysis is simulated by progesterone withdrawal and the follicular phase estradiol rise is reproduced experimentally. In the first experiment, we tested the hypothesis that endotoxin interferes with the estradiol-induced LH surge. Ewes were either infused with endotoxin (300 ng/kg/h, iv) for 30 h beginning at onset of a 48-h estradiol stimulus or sham infused as a control. Endotoxin significantly delayed the time to the LH surge (P , 0.01), but did not alter surge amplitude, duration, or incidence. The second experiment tested the hypothesis that the delaying effects of endotoxin on the LH surge depend on when endotoxin is introduced relative to the onset of the estradiol signal. Previous work in the ewe has shown that a 14-h estradiol signal is adequate to generate GnRH and LH surges, which begin 6 – 8 h later. Thus, we again infused

I

MMUNE/INFLAMMATORY challenges, such as endotoxin or the cytokine interleukin-1 (IL-1), which mediates many of the effects of endotoxin, can profoundly disrupt the natural follicular phase of the cycle in a number of species, including the cow, rat, sheep, and monkey (1–5). The mechanisms of this disruption, however, are unclear. Studies in gonadectomized animals have shown that pulsatile LH release (6 –11), and in sheep pulsatile GnRH release (6), are markedly inhibited by immune challenge. Thus, the suppression of follicular phase GnRH and LH pulses may contribute to follicular phase disruption. It is also possible, however, that immune challenge interferes with the expression of the preovulatory LH surge independently of an effect on pulsatile secretion. In support of this, IL-1 can inhibit the LH surge induced by ovarian steroids in the ovariectomized rat (12). However, the manner by which immune challenge in-

Received August 27, 1998. Address all correspondence and requests for reprints to: Dr. Fred J. Karsch, Reproductive Sciences Program, University of Michigan, 300 North Ingalls Building, Room 1101 SW, Ann Arbor, Michigan 481090404. E-mail: [email protected]. * A preliminary report has appeared in the abstracts of the 30th Annual Meeting of the Society for the Study of Reproduction (Biol Reprod [Suppl 1] 56:93, 1997). This work was supported by NIH Grants MH-11653 and HD-18337; the Sheep Research, Standards and Reagents, Data Analysis, and Administrative Core Facilities of the P30 Center for the Study of Reproduction (NIH HD-18258); and the Office of the Vice President for Research at the University of Michigan.

endotoxin for 30 h, but began it 14 h after the onset of the estradiol signal. In contrast to the first experiment, endotoxin given later had no effect on any parameter of the LH surge. In the third experiment, we tested the hypothesis that endotoxin acts during the first 14 h to disrupt the initial activating effects of estradiol. Estradiol was delivered for just 14 h, and endotoxin was infused only during this time. Under these conditions, endotoxin blocked the LH surge in five of eight ewes. In a similar follow-up study, endotoxin again blocked the LH surge in six of seven ewes. We conclude that endotoxin can disrupt the estradiol-induced LH surge by interfering with the early activating effects of the estradiol signal during the first 14 h (reading of the signal). In contrast, endotoxin does not disrupt later stages of signal processing (i.e. events during the interval between estradiol signal delivery and surge onset), nor does it prevent actual hormonal surge output. Thus, endotoxin appears to disrupt estrogen action per se rather than the release of GnRH or LH at the time of the surge. (Endocrinology 140: 2471–2479, 1999)

terferes with surge generation remains an open issue. For example, because the GnRH/LH surge in the rat is linked to a circadian mechanism (13, 14), immune challenge may interfere either with the interaction between ovarian steroids and the neural substrates that relay the positive feedback signal to the GnRH neurons or with the circadian signal that times surge secretion. In the present investigation, we examined how endotoxin influences the estradiol-induced LH surge in the sheep. Endotoxin is a well characterized model of immune/inflammatory challenge and induces a cascade of cytokines (e.g. IL-1 and tumor necrosis factor-a) and other chemical messengers that impact neuroendocrine systems (7, 15, 16). In sheep, it has long been recognized that an increase in circulating estradiol induces the LH surge (17, 18) and that the surge system is not linked to a circadian mechanism (19). Further, an effective estradiol signal may be relatively brief in duration and occur well in advance of the LH surge itself (20, 21). To study the influence of endotoxin on surge induction, we used a physiological artificial follicular phase model (22) that is well characterized in terms of both estradiol signal requirements and GnRH and LH surge responses (23–25). Of particular interest to this study, the surge induction process in the model has been temporally fractionated (25) into an early period, when estradiol must be elevated to produce GnRH/LH surges (estradiol signal-reading stage; ;0 –14 h after the onset of the estradiol signal under the conditions

2471

2472

ENDOTOXIN DISRUPTS LH SURGE INDUCTION

used in our laboratory), an intermediate period closer to the start of the surge when elevated estradiol is no longer necessary (late signal-processing stage; ;14 –22 h), and a final period when the GnRH and LH surges themselves occur (surge release stage; ;22–36 h). With this in mind, our goal for the current experiments was twofold. First, we investigated whether endotoxin interferes with the estradiol-induced LH surge in the sheep. Second, having found that it does, we assessed whether the effects of endotoxin depend upon its time of introduction. Namely, we examined whether endotoxin impacts the system early, during the estradiolreading stage when estradiol must be elevated to generate surges, or subsequently during the late signal-processing stage (interval between estradiol signal delivery and surge onset) and surge release stage when elevated estradiol is no longer necessary for successful surge generation.

Endo • 1999 Vol 140 • No 6

perature. All procedures were approved by the Committee for the Use and Care of Animals at the University of Michigan.

Exp 1: does endotoxin disrupt the estradiol-induced LH surge? (Fig. 1a) The artificial follicular phase was created in 14 ovariectomized ewes as described above. Ewes were either infused with endotoxin for 30 h beginning at the onset of an estradiol stimulus that lasted through the end of the experiment (48 h) or sham infused as a control (n 5 7 ewes/group; Fig. 1a). A 30-h period of infusion was chosen so that endotoxin would be present through all stages of the surge induction process and the majority of the LH surge itself (in this model, LH surge begins around 20 –24 h and lasts approximately 12 h; hatched bar in Fig. 1a). Jugular blood was sampled by venipuncture at 1- to 2-h intervals (hourly samples from 16 –30 h after estradiol implant) for measurement of LH, cortisol, and progesterone beginning at the onset of estradiol/

Materials and Methods Animal model Experiments were conducted during the mid to late anestrous season (Exp 1, June 1996; Exp 2, August 1996; Exp 3, May-July 1997 and JulyAugust 1998) on adult, ovariectomized, Suffolk ewes maintained outdoors under standard husbandry conditions at the Sheep Research Facility in Ann Arbor, MI. All ewes were ovariectomized aseptically under general anesthesia at least 9 days before LH surge induction. During experiments, disturbance to the ewes was minimized, and a nonstressful environment was maintained. An artificial estrous cycle was created as follows. The luteal phase was simulated by treating ewes at the time of ovariectomy with a 1-cm estradiol capsule (26) and either two intravaginal progesterone-releasing devices (Exp 1, 2, 3A) or two 4 3 7 sc packets containing progesterone (Exp 3B) (27, 28). {Estradiol capsules and progesterone packets were prepared with SILASTIC brand tubing, Dow Corning Corp., Midland, MI; intravaginal progesterone pessaries [controlled internal drug releasing (CIDR)] were provided by InterAg, Hamilton, New Zeland}. These treatments maintain serum concentrations of estradiol and progesterone equivalent to mid to late luteal phase levels (;1 pg/ml and 2– 4 ng/ml, respectively) (22, 25, 28). After 10 days (approximate duration of progesterone elevation during the natural luteal phase), an artificial follicular phase was created as follows. Progesterone was removed to simulate luteal regression. Sixteen or 24 h later, four 3-cm estradiol implants (presoaked 24 h in water) were inserted sc; these implants raise serum estradiol to presurge values of 5– 8 pg/ml within 1–2 h (20, 23). This steroid treatment induces preovulatory-like GnRH and LH surges in virtually 100% of animals beginning approximately 20 –24 h after estradiol addition (23–25). Endotoxin was infused iv using a backpack pump system regularly employed in our laboratory (29). This system allowed continuous delivery of endotoxin with minimal human interaction with the animals. Backpacks were placed on ewes several days before progesterone withdrawal to allow adjustment to the devices. In ewes designated to receive endotoxin, an iv jugular catheter was placed at the time of progesterone withdrawal; controls were sham cannulated. Endotoxin was dissolved in saline and infused at 300 ng/kg/h (0.16 ml/h). This dose of endotoxin was chosen because it reportedly produced behavioral sickness effects in gonadectomized rams but did not cause severe distress (Daley, C., University of California-Davis, personal communication). Additionally, we verified that this dose of endotoxin consistently generates fever, stimulates adrenal cortisol secretion, and disrupts the follicular phase of the cycle in our animals (4, 30). The time of endotoxin relative to the estradiol stimulus varied according to the experimental objective. In Exp 1, a core body temperature response was characterized in six of the eight endotoxin-treated ewes using battery-operated, temperature-sensitive telemetry transmitters (model CH-3, MiniMitter, Sunriver, OR) tied to the broad ligament of the uterus at the time of ovariectomy (6). In Exp 2 and 3A and in controls of all three experiments, body temperature responses were confirmed by measurement of rectal tem-

FIG. 1. Designs for Exp 1 (a), Exp 2 (b), and Exp 3 (c). Time is depicted as hours relative to the addition of the peak follicular phase estradiol (E) implants. The dashed line designates the profile of serum progesterone (P) concentrations; E designates the profile of serum estradiol concentrations. Dark bars indicate the time of endotoxin infusion. Hatched bars indicate the approximate time of the LH surge.

ENDOTOXIN DISRUPTS LH SURGE INDUCTION endotoxin treatment (defined as 0 h) and continuing for 48 h. Serum estradiol was measured at three time points (0, 18, and 30 h) to confirm that the implants produced the expected circulating steroid levels. Additionally, blood was sampled before progesterone withdrawal to confirm that the CIDRs produced circulating luteal phase progesterone levels.

Exp 2: does the effect of endotoxin depend on when it is introduced? (Fig. 1b) In this experiment we investigated whether the effect of endotoxin on the surge induction process is time dependent. Previous work in the artificial follicular phase model has shown that a 14-h estradiol signal is adequate to generate GnRH and LH surges beginning approximately 8 h later (25). Thus, it appears that the critical role for elevated estradiol in surge induction is early, during the first 14 h of the surge induction process under the conditions used in our laboratory. This initial reading of the signal is followed by periods of late estradiol signal processing and GnRH/LH surge release, which themselves do not require maintenance of elevated circulating estradiol levels (25). Our goal was to determine whether endotoxin interrupts the LH surge after the estradiol signal has been read, i.e. during the late signal-processing and GnRH/LH surge release stages. The overall design was similar to that of Exp 1 and is shown in Fig. 1b. Endotoxin was infused for 30 h (n 5 7) beginning 14 h after initiating the estradiol stimulus, i.e. after the initial reading of the estradiol signal in our animal model (25). Controls were sham infused (n 5 6). Estradiol implants remained in place until the end of the experiment (60 h). Jugular blood was sampled by venipuncture at hourly intervals for measurement of LH, cortisol, and progesterone beginning 12 h after implanting estradiol and continuing through the end of endotoxin infusion (44 h) and at 2- to 4-h intervals until 60 h. Additional blood samples were taken at the time of progesterone withdrawal.

Exp 3: does endotoxin disrupt reading of the estradiol signal? (Fig. 1c) Part A. Our third experiment continued to dissect the time-dependent effects of endotoxin by testing the hypothesis that endotoxin disrupts the initial reading of the estradiol signal, i.e. during the first 14 h of the estradiol signal when elevated circulating estradiol is required (25). If this hypothesis is correct, then endotoxin should disrupt the LH surge if both endotoxin and estradiol are only present for 14 h. As illustrated in Fig. 1c (part A), ewes underwent steroid treatments similar to those in Exp 1 and 2 with the following differences: 1) the estradiol stimulus was only delivered for 14 h, and endotoxin was infused only during the 14 h of estradiol (n 5 8) (controls sham infused; n 5 7); and 2) sc implants were used to deliver progesterone rather than intravaginal CIDRs (pilot data suggested that a new batch of CIDRs did not produce the expected progesterone levels). Jugular blood was sampled by venipuncture at hourly intervals for measurement of LH, cortisol, and progesterone from 8 – 48 h after implanting estradiol and around the time of progesterone withdrawal to confirm the efficacy of the sc progesterone implants. Part B. Part A of this experiment revealed that the LH surge was blocked by endotoxin infused during the 14-h estradiol signal in the majority of ewes (see Results). A similar experiment was conducted during late anestrus of the following year (July-August 1998) to expand our test of the hypothesis that endotoxin disrupts the initial reading of the estradiol signal. The design is illustrated in Fig. 1c (part B). The artificial follicular phase was created using two CIDR devices to simulate luteal phase progesterone levels. Estradiol was delivered for just 12 h (preliminary studies indicated that this was adequate to generate a LH surge). Endotoxin was given as an iv bolus (400 ng/kg) rather than as an infusion at the time estradiol implants were inserted (n 5 7); controls received saline iv (n 5 11). This endotoxin treatment has been extensively characterized in our laboratory, stimulating cortisol, generating fever, and inhibiting pulsatile GnRH and LH secretion (6). Jugular blood was sampled by venipuncture at hourly intervals for LH and cortisol measurement beginning 3 h before estradiol implant/endotoxin injection and continuing for 36 h.

2473

Assays LH was measured in duplicate aliquots of plasma (10 –200 ml) using a modification (19) of a previously described RIA (31, 32) and is expressed in terms of NIH LH-S12. The mean intra- and interassay coefficients of variation were 5.2% and 6.6%, respectively, and assay sensitivity for 200 ml averaged 0.6 ng/ml. Cortisol was measured in duplicate 50-ml aliquots of unextracted plasma using the Coat-A-Count cortisol assay kit (Diagnostic Products, Los Angeles, CA), previously validated in our laboratory for use in the sheep (6). The mean intra- and interassay coefficients of variation were 9.6% and 9.8%, respectively, and assay sensitivity averaged 0.86 ng/ml. Progesterone was determined in duplicate 100-ml aliquots using the Coat-A-Count progesterone assay kit (Diagnostics Products) validated for use in sheep (33). The mean intraand interassay coefficients of variation were 8.6% and 8.2%, respectively, and assay sensitivity averaged 0.04 ng/ml. Estradiol was measured in a single assay in duplicate diethyl ether extracts of 200 ml plasma using the Serono Diagnostics estradiol MAIA assay (Serono Baker, Allentown, PA) adapted for use in sheep (34). Intraassay variation, as determined by the median variance ratio of assay replicates, was 10.7%, and assay sensitivity was 1 pg/ml.

Data analysis For each ewe, the mean of the presurge baseline for LH was calculated. Surge onset was defined as the time LH rose above 2 times the presurge baseline and remained so for at least 4 h. The end of the LH surge was defined as the time LH fell to below twice the presurge baseline for at least four samples. In several ewes, LH did not fall below this value but remained consistently at concentrations 4 –5 times above baseline for more than four samples (surge peak was generally .50 times baseline). In these animals, the end of the surge was defined as the first of these consistent values. If the surge had not ended before the end of sampling as defined above, then surge duration was taken to be the time from surge onset to the end of sample collection (4 of the 48 surges in this study). The LH surge peak was taken as the highest concentration assayed. The surge peak (data log transformed), duration, and time to onset were compared between the control and endotoxin groups in each experiment by Student’s t test. In Exp 3, not all endotoxin-treated ewes expressed the LH surge. Fisher’s exact probability test was used to determine whether the proportion of control ewes exhibiting the LH surge was significantly different from the proportion of experimental ewes exhibiting the LH surge. Mean cortisol, progesterone, and temperature values were calculated for each ewe before vs. during endotoxin infusion (or during the comparable period for controls). Treatment effects within animals were identified by paired t test (before vs. during endotoxin). Differences in estradiol and progesterone concentrations between the control and experimental groups in each experiment were determined by Student’s t test. Cortisol, progesterone, and temperature values were log transformed before statistical analyses. The level of significance was established at P # 0.05.

Results Exp 1: does endotoxin disrupt the estradiol-induced LH surge?

Mean plasma progesterone on the last 2 days of progesterone treatment in all ewes was 2.2 6 0.1 ng/ml, similar to values in the late luteal phase (19, 35). Estradiol implants produced a rise in serum estradiol, from 1.0 6 0.1 pg/ml at 0 h to 4.6 6 0.4 and 6.3 6 0.6 pg/ml at 18 and 30 h, respectively; this estradiol rise was similar to that in the natural follicular phase (19, 35). Values for both steroids were not different between control and experimental groups. Figure 2 depicts the adrenal cortisol and progesterone and body temperature profiles in all endotoxin-treated and control ewes. In controls, cortisol and progesterone remained at or near levels considered to be basal in our ewes (6) throughout sampling (mean from 0 – 48 h: cortisol, 13.8 6 1.3 ng/ml;

2474

ENDOTOXIN DISRUPTS LH SURGE INDUCTION

Endo • 1999 Vol 140 • No 6

in Fig. 3b. Endotoxin, infused for 30 h beginning at the onset of the estradiol stimulus, significantly delayed the LH surge by approximately 10 h (control vs. endotoxin; latency to LH peak, 21.7 6 1.1 vs. 31.1 6 2.0 h; P , 0.01). Endotoxin did not, however, affect the occurrence of the LH surge or its amplitude or duration (control vs. endotoxin: incidence, 7 of 7 vs. 7 of 7 ewes expressed LH surge; amplitude, 89.8 6 16.6 vs. 70.6 6 9.7 ng/ml; duration, 14.9 6 1.4 vs. 14.9 6 1.6 h). Exp 2: does the effect of endotoxin depend on when it is introduced?

In this experiment, endotoxin was infused for 30 h beginning 14 h after the onset of the estradiol stimulus. As in Exp 1, circulating progesterone before its withdrawal was at a luteal phase level (2.5 6 0.3 ng/ml), and endotoxin significantly stimulated cortisol, progesterone, and core body temperature (pre vs. during endotoxin: cortisol, 12.1 6 1.5 vs. 50.9 6 2.5 ng/ml; progesterone, 0.17 6 0.02 vs. 0.45 6 0.05 ng/ml; temperature, 39.1 6 0.1 vs. 40.6 6 0.2 C; P # 0.001; data not illustrated). Figure 4 illustrates the mean LH surge profile, the percentage of ewes expressing the LH surge, and the latency to peak LH in all controls and endotoxin-treated ewes. In contrast to the delay observed in Exp 1, when endotoxin began at the onset of the estradiol stimulus, endotoxin beginning 14 h later had no effect on any aspect of the LH surge response. The time to the LH surge peak was not affected (Fig. 4b; control vs. endotoxin, 24.3 6 1.3 vs. 26.4 6 0.9 h). Further, the incidence, amplitude, and duration of the LH surge were not significantly different between groups (Fig. 4a; control vs. endotoxin: incidence, 7 of 7 vs. 6 of 6 ewes expressed LH surge; amplitude, 90.9 6 26.5 vs. 65.2 6 20.1 ng/ml; duration, 16.0 6 2.0 vs. 15.6 6 1.9 h). Exp 3: does endotoxin disrupt reading of the estradiol signal?

FIG. 2. Mean 6 SEM cortisol, progesterone, and body temperature responses in all control (open circles) and endotoxin-treated (closed circles) ewes in Exp 1 (n 5 7/group). Endotoxin was infused for 30 h beginning at the time of estradiol implant insertion, i.e. 0 –30 h. Controls were sham infused. The estradiol signal was maintained throughout the experiment. Jugular blood samples were taken at 1to 2-h intervals. Body temperature was monitored periodically by MiniMitter in endotoxin ewes and by rectal thermometer in all controls.

progesterone, 0.1 6 0.02 ng/ml). Rectal temperature also remained basal at the time points measured (mean of three readings taken at 0, 4, and 12 h; 39.4 6 0.1 C). In contrast, endotoxin significantly stimulated cortisol, progesterone, and core body temperature (P # 0.001). Interestingly, after reaching peak values within 4 h of the start of endotoxin, all three of these responses declined and approached baseline despite continued endotoxin infusion. Figure 3 illustrates the mean LH surge responses in all control and endotoxin-treated ewes. LH values, normalized to the surge peak, and percentage of ewes expressing the LH surge are depicted in Fig. 3a; latency to the peak is depicted

Part A. In this experiment, both endotoxin and estradiol were delivered only from 0 –14 h, i.e. during the estradiol-reading stage (25). Circulating progesterone before its withdrawal was at a late luteal phase level (3.5 6 0.2 ng/ml), and endotoxin again significantly stimulated cortisol secretion (before vs. during endotoxin, 8.5 6 1.8 vs. 56.1 6 6.6 ng/ml; P # 0.001). [Note, we did not take adequate basal progesterone and temperature measurements to compare before vs. during endotoxin within the same animal.] Figure 5a illustrates the mean LH surge profile and percentage of ewes expressing the LH surge in all control and endotoxin-treated ewes. All seven controls exhibited the LH surge at the expected time, with the peak at 22.9 6 1.5 h. In marked contrast to Exp 1 and 2 in which all endotoxin ewes expressed the LH surge, the surge was completely blocked in five of eight ewes when endotoxin was infused during the 14-h estradiol signal. Using Fisher’s exact probability test to compare proportions of ewes that expressed the LH surge between groups, the incidence of the LH surge in endotoxintreated ewes was significantly less than that in controls (control vs. endotoxin, seven of seven vs. three of eight ewes expressed LH surge; P 5 0.023). The LH responses in the three endotoxin ewes that exhibited the LH surge were similar to those in controls (control vs. three endotoxin ewes that

ENDOTOXIN DISRUPTS LH SURGE INDUCTION

2475

FIG. 3. Mean 6 SEM plasma LH concentration and percentage of ewes expressing the LH surge (a), and latency to peak LH (b) in all endotoxin-treated and control ewes (n 5 7/group) in Exp 1. Controls are depicted with open circles or bars, and the endotoxin group is shown with solid circles or bars. In a, LH is plotted normalized to the peak concentration. Endotoxin was infused for 30 h beginning from the time of estradiol implant insertion (0 –30 h). Controls were sham infused. The estradiol signal was maintained throughout the experiment. Endotoxin significantly delayed the time to the peak of the LH surge (**, P , 0.01).

FIG. 4. The mean 6 SEM plasma LH concentration and the percentage of ewes expressing the LH surge (a), and latency to peak LH (b) in all endotoxintreated (n 5 7) and control ewes (n 5 6) in Exp 2. Controls are depicted with open circles or bars, and endotoxintreated ewes are depicted with solid circles or bars. In a, LH is plotted normalized to the peak concentration. Endotoxin was infused for 30 h beginning 14 h after estradiol implant insertion (14 – 44 h). Controls were sham infused. The estradiol signal was maintained throughout the experiment. Endotoxin had no significant effect on any parameter of the LH surge.

expressed LH surge: time to peak, 22.9 6 1.5 vs. 24.3 6 1.8 h; peak value, 40.4 6 8.8 vs. 82.3 6 26.5; duration, 14.7 6 1.9 vs. 12.0 6 0.6 h; statistical comparisons were not made due to the small number of endotoxin ewes expressing the LH surge). Of interest, the five endotoxin-treated ewes that failed to exhibit the LH surge tended to have greater plasma cortisol and progesterone concentrations from the start of sampling midway through endotoxin treatment to just before the LH surge (8 –21 h) than did the three ewes that did express a surge ( surge blocked vs. surge expressed: cortisol, 64.8 6 9.1 vs. 34.0 6 8.9 ng/ml; progesterone, 0.32 6 0.08 vs. 0.21 6 0.05 ng/ml; statistical comparisons were not made due to the small number of animals in the responders vs. nonresponders). Part B. As in previous studies, endotoxin significantly stimulated cortisol secretion (before vs. during endotoxin; 3.3 6 0.4 vs. 45.3 6 6.2 ng/ml; P , 0.001). All 11 control ewes responded to the 12-h estradiol stimulus with the LH surge at the expected time (LH surge peak at 27 6 1 h). Similar to part A of this experiment, endotoxin at the onset of the shortened estradiol signal blocked the LH surge in six of the seven ewes; the surge in the single positive responder was low in amplitude (Fig. 5b). Statistical analysis using Fisher’s exact probability test to compare proportions of ewes that expressed the LH surge between groups indicated that the LH surge incidence in endotoxin-treated ewes was signifi-

cantly less than that in controls (control vs. endotoxin; 11 of 11 vs. 1 of 7 ewes expressed LH surge; P 5 0.0004). Discussion

Earlier studies have shown that immune challenge can disrupt the natural follicular phase of the cycle in a number of species (1–5). For example, endotoxin prevents the spontaneous LH surge and induces cystic follicles in the cow (1) and interrupts the preovulatory estradiol rise and blocks or delays the LH surge in the sheep and monkey (4, 5). In the rat, endotoxin or the cytokine IL-1 given on proestrus can inhibit fos expression in GnRH neurons (2, 36) and suppress GnRH neuronal activation and gene expression at the expected time of the LH surge (36). In considering potential mechanisms by which endotoxin disrupts preovulatory events in the natural follicular phase, it is important to note that immune challenge potently suppresses GnRH pulsatile release into the hypophyseal portal blood (6) and thus LH pulsatile secretion (6 –11). Interference with pulsatile gonadotropin secretion in itself could disrupt the follicular phase by impairing follicular development and the subsequent preovulatory estradiol rise. The present findings suggest that endotoxin can also disrupt the follicular phase by preventing the ability of the surge-generating mechanism to respond to the preovulatory estradiol rise. In this regard, our data com-

2476

ENDOTOXIN DISRUPTS LH SURGE INDUCTION

FIG. 5. The mean 6 SEM plasma LH concentration and percentage of ewes expressing the LH surge in all endotoxin-treated and control ewes in Exp 3: part A (top) and part B (bottom). Controls are depicted with open circles; endotoxin-treated ewes whose LH surge was blocked are depicted with solid circles; endotoxin-treated ewes that surged are depicted with the dashed line. LH is plotted normalized to the peak concentration. In Exp 3, part A, estradiol was present for only 14 h. Endotoxin (300 ng/kgzh; n 5 8) was infused for 14 h beginning from the time of estradiol implant insertion. In Exp 3, part B, estradiol was present for 12 h, and endotoxin was given as an iv bolus (400 ng/kg; n 5 7) at the time of estradiol implant. Controls were sham infused (n 5 7, Exp 3A; n 5 11, Exp 3B). Endotoxin had a significant effect to block the LH surge in both experiments (*, P 5 0.023 in part A; **, P 5 0.0004 in part B; by Fisher’s exact probability test).

plement studies in the rat showing that IL-1 can suppress a steroid-induced LH surge (12). Our finding that the effects of endotoxin are dependent on time of exposure relative to the estradiol signal provides important insight into how immune challenge interferes with LH surge generation. It is insightful to consider the basis for this time dependency in conjunction with an integrative model (illustrated in Fig. 6) describing the steps by which

Endo • 1999 Vol 140 • No 6

estradiol induces the GnRH/LH surge in the ewe. The model is based on observations that estradiol may not act directly on GnRH neurons (37– 40), historical data that estradiol induces the LH surge with a time delay (17, 18, 20, 21), and evidence that the critical action of estradiol to induce the surge occurs well in advance of GnRH/LH release itself (21, 25). According to this model, there are three stages to surge induction (evidence for the three stages is provided in Ref. 25). Stage I consists of initial reading of the estradiol signal by the estradiol-sensitive neurons, a time when estradiol must be elevated (;0 –14 h in the artificial follicular phase model as employed in our laboratory). Stage II consists of late signal processing (i.e. events occurring during the interval between estradiol signal delivery and surge onset) when the positive feedback signal is relayed to the GnRH neuron, either directly by estradiol-responsive neurons or by one or more interneuron(s) (;14 –22 h). During this stage, estradiol no longer needs to be elevated to induce a surge. Stage III, also not reliant on elevated estradiol, is the time of hormonal release when the discharge of GnRH and LH actually occurs (beyond 22 h). Of interest to this temporal model for surge induction, observations in sheep and rats using c-fos as a marker of neuronal activation indicate that estradiol induces early expression of c-fos in brain stem regions that regulate GnRH well in advance of the LH surge (41, 42). In the present investigation, endotoxin only interfered with induction of the LH surge when present during the period of estradiol dependence (stage I, black bar in Fig. 6). We thus conclude that endotoxin disrupts the initial reading of the estradiol signal, not later signal processing or surge release itself. It will now be of keen interest to determine whether endotoxin blocks the early c-fos induction in neuronal populations associated with estradiol action and surge induction. Two issues are of interest relevant to the conclusion that endotoxin blocks reading of the estradiol signal. The first pertains to when, relative to estradiol signal onset, other disrupters of the LH surge elicit their inhibitory effect in the ewe. Similar to endotoxin, certain disrupters of the estradiolinduced LH surge also appear to act early in the induction process. Ethanol, for example, can delay the LH surge in the artificial follicular phase model, but only when present during the beginning of the estradiol signal (43). Progesterone, another disrupter of the surge, also exerts its blocking effects before the actual release of GnRH/LH in the artificial follicular phase model (44), and it can block the surge if present only during the estradiol reading stage (Harris, T. G., unpublished observations). Unlike endotoxin, however, progesterone can also act during the late signal-processing stage, i.e. after the estradiol signal is read (44). Also differing from endotoxin, studies in the early postpartum ewe indicate that transport stress delays the estradiol-induced LH surge but only when ewes are transported just before or during the expected time of LH surge release (45). Finally, insulininduced hypoglycemia can delay the LH surge at any time before hormonal surge release (46). Second, it is relevant to consider our findings of the timedependent influence of endotoxin on the LH surge in sheep in light of related studies in the rat. Similar to the sheep, the critical actions of estradiol on surge generation in the rat occur well in advance of GnRH and LH surge release (47, 48).

ENDOTOXIN DISRUPTS LH SURGE INDUCTION

2477

FIG. 6. Theoretical model of the GnRH/LH surge induction process and the time-dependent effect of endotoxin. Typical surges of LH (shaded region) and GnRH (solid line) in the artificial follicular phase model are adapted from prior work (23). The model for LH surge induction is based on findings reported by Evans et al. (25). The endotoxin-sensitive and estradiol-dependent time is shown with the black bar.

Based on studies involving timed ovariectomies and insertion/removal of estradiol implants, the GnRH surge system is primed and activated by estradiol some 12–16 h before surge onset (47, 48). Beyond this time, the GnRH neurosecretory system of the rat awaits the circadian timing signal to initiate hormonal surge release on the afternoon of proestrus (13, 14). In contrast to our findings with endotoxin in sheep, IL-1 can block the steroid-induced LH surge in ovariectomized rats when given at the time of surge onset (12). This suggests that, in the rat, immune challenge interferes with surge induction after the estradiol signal has been read, blocking either the neuronal processes immediately preceding the steroid-induced LH surge or actual hormonal surge release itself. Studies in intact rats also suggest that immune challenge can interrupt GnRH neuronal activation when given after the estradiol signal is delivered (2, 36). To our knowledge, it has not been determined whether immune challenge can block estradiol signal reading in the rat as we observed in the sheep. Relative to our conclusion that endotoxin impairs initial reading of the estradiol signal in sheep, it is important to provide an explanation for the delayed LH surge we observed in Exp 1 vs. the blockade of the LH surge in Exp 3. We consider the most parsimonious explanation to be the development of tolerance to endotoxin in conjunction with the differing durations of the estradiol signals (48 h in Exp 1 vs. 12 or 14 h in Exp 3). Tolerance refers to a hyporesponsive state that develops with continued immune challenge and favors survival when animals are reexposed to the immune challenge (49). Suggestive of tolerance in Exp 1, the adrenal steroid (cortisol and progesterone) and temperature responses began to wane despite continued infusion of endotoxin; values peaked within the first 4 –5 h and approached pretreatment levels by the end of the 30-h infusion. In Exp 1, the estradiol implants maintained surge-inducing steroid levels for 48 h, well beyond the time that tolerance developed, as assessed by the adrenal steroid and temperature responses to endotoxin. Accordingly, estradiol could have restarted the surge induction process once this tolerance set in, causing the delay of the LH surge we observed in Exp 1. In contrast, in Exp 3 the estradiol stimulus was not main-

tained because the implants were removed after 12 or 14 h. Thus, we propose that the surge induction process could not restart once the effects of endotoxin wore off, and hence, the LH surge was blocked. Although the ability of endotoxin to block early reading of the estradiol signal was evident in Exp 3, some endotoxintreated ewes did express the LH surge. The question thus arises, why did this split response occur? Of interest, analysis of circulating cortisol and progesterone in Exp 3 revealed a tendency for greater concentrations of both steroids in those ewes that did not surge. Perhaps the profundity of these steroid responses serves as a marker for the overall impact of endotoxin on neuroendocrine circuitries in general, one being the LH surge system. Additionally, the steroids themselves may contribute to blockade of the surge in a dosedependent manner. Although cortisol itself has not been found to block the estradiol-induced LH surge in the ewe (50), progesterone is a potent blocker of the surge mechanism (18). In this regard, circulating progesterone levels similar to peak values we observed after endotoxin (;0.5–1 ng/ml) may be sufficient to prevent the estradiol-induced LH surge in the ewe (51). It will thus be of interest to determine whether endotoxin induces sufficient amounts of these steroids, alone or in combination, for a sufficient period of time to block the surge. It is worth noting that although endotoxin failed to block actual surge release of LH, pulsatile GnRH and LH secretion are potently suppressed by this immune challenge in the ewe (6). Perhaps such differential effects of endotoxin on pulse and surge release reflect different populations of GnRH neurons or different control mechanisms for these two modes of GnRH/LH secretion. We must emphasize caution, however, in drawing definitive conclusions relative to expression of the GnRH surge because we did not measure GnRH secretion directly; we relied on LH as its marker. Importantly, neither the full amplitude (52) nor duration (24) of the GnRH surge is needed for a normal LH surge in the ewe. Thus, the LH surge could appear perfectly normal under conditions in which the GnRH surge is blunted or shortened. Additionally, endotoxin, or the cytokines it induces, could act at the level of the pituitary to modify responsiveness to GnRH (53), such

2478

ENDOTOXIN DISRUPTS LH SURGE INDUCTION

that LH secretion may not reflect GnRH release under the conditions of our experiments. Further studies will thus be required to establish whether the actual surge release of GnRH is truly insensitive to endotoxin. In conclusion, our data suggest that endotoxin can disrupt the ability of the LH surge-generating mechanism of the ewe to respond to the preovulatory estradiol signal. This disruption would be expected to alter the follicular phase of the cycle. Interestingly, endotoxin has a time-dependent effect, interfering with the action of estradiol per se at the time when the estradiol signal is read by the estrogen-sensitive neurons. Endotoxin does not appear to disrupt later processing and relay of the positive feedback signal to GnRH neurons, nor does it impact actual hormonal LH surge release in the ewe. Our data thus suggest that a window of sensitivity exists to the disruptive effects of endotoxin on the surge generation process, a window in advance of actual hormonal surge release. Acknowledgments The authors sincerely thank Doug Doop and Gary McCalla for their expertise with the animal experimentation. We also thank Jennifer Bowen, Martha Brown, Dr. Vasantha Padmanabhan, and Dr. Lori Thrun for their input into the design and interpretation of the results, and Drs. Gordon D. Niswender and Leo E. Reichert, Jr., for supplying the RIA reagents. Finally, Ms. Barbara Glover, who is no longer with us, will be remembered by all for her technical expertise and years of devotion to our laboratory.

References 1. Peter AT, Bosu WTK, DeDecher RJ 1989 Suppression of preovulatory luteinizing hormone surges in heifers after intrauterine infusions of Escherichia coli endotoxin. Am J Vet Res 50:368 –373 2. Rivest S, Rivier C 1993 Interleukin-1b inhibits the endogenous expression of the early gene c-fos located within the nucleus of LH-RH neurons and interferes with hypothalamic LH-RH release during proestrus in the rat. Brain Res 613:132–142 3. Rivest S, Lee S, Attardi B, Rivier C 1993 The chronic intracerebroventricular infusion of interleukin-1b alters the activity of the hypothalamic-pituitarygonadal axis of cycling rats. I. Effect on LHRH and gonadotropin biosynthesis and secretion. Endocrinology 133:2424 –2430 4. Battaglia DF, Krasa HK, Viguie´ C, Karsch FJ, Does immune stress disrupt pulsatile LH secretion during the follicular phase of the estrous cycle? 27th Annual Meeting of the Society for Neuroscience, New Orleans LA, 1997, vol 23:1245 (Abstract 491.15) 5. Xiao E, Xia-Zhang L, Barth A, Zhu J, Perin M 1998 Stress and the menstrual cycle: relevance of quality in the short- and long-term response to a 5-day endotoxin challenge during the follicular phase in the rhesus monkey. J Clin Endocrinol Metab 83:2454 –2460 6. Battaglia DF, Bowen JM, Krasa HK, Thrun LA, Viguie´ C, Karsch FJ 1997 Endotoxin inhibits the reproductive neuroendocrine axis while stimulating adrenal steroids: a simultaneous view from hypophyseal portal and peripheral blood. Endocrinology 138:4273– 4281 7. Coleman ES, Elsasser TH, Kemppainen RJ, Coleman DA, Sartin JL 1993 Effect of endotoxin on pituitary hormone secretion in sheep. Neuroendocrinology 58:111–122 8. Kujjo LL, Bosu WTK, Perez GI 1995 Opioid peptides involvement in endotoxin-induced suppression of LH secretion in ovariectomized Holstein heifers. Reprod Toxicol 1:169 –174 9. Ebisui O, Fukata J, Tominaga T, Murakami N, Kobayashi H, Segawa H, Muro S, Naito Y, Nakai Y, Maui Y, Nishida T, Imura H 1992 Roles of interleukin-1a and -1b in endotoxin-induced suppression of plasma gonadotropin levels in rats. Endocrinology 130:3307–3313 10. Bonavera JJ, Kalra SP, Kalra PS 1993 Mode of action of interleukin-1 suppression of pituitary LH release in castrated male rats. Brain Res 612:1– 8 11. Feng Y, Shalts E, Xia L, Rivier J, Rivier C, Vale W, Ferin M 1991 An inhibitory effect of interleukin-1a on basal gonadotropin release in the ovariectomized rhesus monkey: reversal by a corticotropin-releasing factor antagonist. Endocrinology 128:2077–2082 12. Kalra PS, Sahu A, Kalra SP 1990 Interleukin-1 inhibits the ovarian steroidinduced luteinizing hormone surge and release of hypothalamic luteinizing hormone-releasing hormone in rats. Endocrinology 126:2145–2152

Endo • 1999 Vol 140 • No 6

13. Everett JW, Sawyer CH 1950 A 24-hour perodicity in the ‘LH release apparatus’ of female rats, disclosed by barbiturate sedation. Endocrinology 47:198 –218 14. Kalra SP, Kalra PS 1983 Neural regulation of luteinizing hormone secretion in the rat. Endocr Rev 4:311–351 15. Redl H, Bahrami S, Schlag G, Traber DL 1993 Clinical detection of LPS and animal models of endotoxemia. Immunobiology 187:330 –345 16. Tilders FJH, DeRijk RH, Van Dam A, Vincent VAM, Schotanus K, Persoons JHA 1994 Activation of the hypothalamus-pituitary-adrenal axis by bacterial endotoxins: routes and intermediate signals. Psychoneuroendocrinology 19:209 –232 17. Goding JR, Catt KJ, Brown JM, Kaltenbach CC, Cumming IA, Mole BJ 1969 Radioimmunoassay for ovine luteinizing hormone. Secretion of luteinizing hormone during estrus and following estrogen administration in the sheep. Endocrinology 85:133–142 18. Scaramuzzi RJ, Tillson SA, Thorneycroft IH, Caldwell BV 1971 Action of progesterone and estrogen on behavioral estrus and luteinizing hormone levels in the ovariectomized ewe. Endocrinology 88:1184 –1189 19. Hauger RL, Karsch FJ, Foster DL 1977 A new concept for control of the estrous cycle of the ewe based on the temporal relationships between luteinizing hormone, estradiol and progesterone in peripheral serum and evidence that progesterone inhibits tonic LH secretion. Endocrinology 101:807– 817 20. Jonas HA, Salamonsen LA, Burger HG, Chamley WA, Cumming IA, Findlay JK, Goding JR 1972 Release of FSH after administration of gonadotrophinreleasing hormone or estradiol to the anestrous ewe. Endocrinology 92:862– 865 21. Coppings RJ, Malven PV 1976 Biphasic effect of estradiol on mechanisms regulating LH release in ovariectomized sheep. Neuroendocrinology 21:146 –156 22. Goodman RL, Legan SJ, Ryan KD, Foster DL, Karsch FJ 1981 Importance of variations in behavioural and feedback actions of oestradiol to the control of seasonal breeding in the ewe. J Endocrinol 89:229 –240 23. Moenter SM, Caraty A, Karsch FJ 1990 The estradiol-induced LH surge of gonadotropin-releasing hormone in the ewe. Endocrinology 127:1375–1384 24. Evans NP, Dahl GE, Caraty A, Padmanabhan V, Thrun LA, Karsch FJ 1996 How much of the gonadotropin-releasing hormone (GnRH) surge is required for generation of the luteinizing hormone surge in the ewe? Duration of the endogenous GnRH signal. Endocrinology 137:4730 – 4737 25. Evans NP, Dahl GE, Padmanabhan V, Thrun LA, Karsch FJ 1997 Estradiol requirements for induction and maintenance of the gonadotropin-releasing hormone surge: implications for neuroendocrine processing of the estradiol signal. Endocrinology 138:5408 –5414 26. Karsch FJ, Dierschke DJ, Weick RF, Yamaji T, Hotchkiss J, Knobil E 1973 Positive and negative feedback control by estrogen of luteinizing hormone secretion in the rhesus monkey. Endocrinology 92:799 – 804 27. Karsch FJ, Legan SJ, Hauger RL, Foster DL 1977 Negative feedback action of progesterone on tonic luteinizing hormone secretion in the ewe: dependence on the ovaries. Endocrinology 101:800 – 806 28. Van Cleef J, Karsch FJ, Padmanabhan V 1998 Characterization of endocrine events during the periestrous period in sheep after estrous synchronization with controlled internal drug release (CIDR) device. Dom Anim Endocrinol 15:23–34 29. Wayne NL, Malpaux B, Karsch FJ 1988 How does melatonin code for day length in the ewe: duration of nocturnal melatonin release or coincidence of melatonin with a light-entrained sensitive period? Biol Reprod 39:6 –75 30. Krasa HB, Battaglia DF, Bowen JM, Thrun LA, Viguie´ C, Karsch FJ, Endotoxin disrupts the follicular phase of the estrous cycle in sheep. 26th Annual Meeting of the Society for Neuroscience, Washington DC, 1996, vol 22:1791 (Abstract 704.20) 31. Niswender GD, Midgley Jr AR, Reichert Jr LE 1968 Radioimmunologic studies with murine, ovine, and porcine luteinizing hormone. In: Rosenberg E (ed) Gonadotropins. GERON-X, Los Altos, p 299 32. Niswender GD, Reichert Jr LE, Midgley Jr AR, Nalbandov AV 1969 Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 84:1166 –1173 33. Padmanabhan V, Evans NP, Dahl GE, McFadden KL, Mauger DT, Karsch FJ 1995 Evidence for short or ultrashort loop negative feedback of gonadotropinreleasing hormone secretion. Neuroendocrinology 62:248 –258 34. Evans NP, Dahl GE, Glover BH, Karsch FJ 1994 Central regulation of pulsatile GnRH secretion by estradiol during the period leading up to the preovulatory GnRH surge in the ewe. Endocrinology 134:1896 –1811 35. Karsch FJ, Legan SJ, Ryan KD, Foster DL 1980 Importance of estradiol and progeserone in regulating LH secretion and estrous behavior during the sheep estrous cycle. Biol Reprod 23:404 – 413 36. Nappi RE, Rivest S 1997 Effect of immune and metabolic challenges on the luteinizing hormone-releasing hormone neuronal system in cycling female rats: an evaluation at the transcriptional level. Endocrinology 138:1374 –1384 37. Herbison AE, Robinson JE, Skinner DC 1993 Distribution of estrogen receptor-immunoreactive cells in the preoptic area of the ewe: co-localization with glutamic acid decarboxylase but not luteinizing hormone-releasing hormone. Neuroendocrinology 57:751–759 38. Lehman MN, Karsch FJ 1993 Do gonadotropin-releasing hormone, tyrosine

ENDOTOXIN DISRUPTS LH SURGE INDUCTION

39.

40.

41. 42. 43.

44.

45.

hydroxylase-, and b-endorphin-immunoreactive neurons contain estrogen receptors? A double-label immunocytochemical study in the Suffolk ewe. Endocrinology 133:887– 895 Nappi RE, Laflamme N, Drolet G, Labrie C, Rivest S, Neuropeptidergic characterization of the novel estrogen receptor (ERb) in the rat hypothalamus. 80th Annual Meeting of The Endocrine Society, New Orleans LA, 1998, p 441 (Abstract P3–267) Hrabvoszky E, Kallo I, Shughrue PJ, Merchenthaler I, Liposits Z, Lack of estrogen-receptor-b (ER-b) mRNA co-expression in luteinizing hormone releasing hormone (LHRH) neurons of the rat brain. 80th Annual Meeting of The Endocrine Society, New Orleans LA, 1998, p 442 (Abstract P3–269) Clarke IJ 1995 The preovulatory LH surge: a case of a neuroendocrine switch. Trends Endocrinol Metab 7:241–247 Conde GL, Bicknell RJ, Herbison AE 1995 Changing patterns of fos expression in brainstem catecholaminergic neurons during the rat oestrous cycle. Brain Res 672:68 –76 Goebbert A, Klein J, Rabil J, Advis JP, Acute ethanol ingestion, and expression of synchronized LH preovulatory surges using a follicular model in ewes. 26th Annual Meeting of the Society for Neuroscience, Washington DC, 1996, vol 22:1346 (Abstract 533.18) Harris TG, Dye S, Robinson JE, Skinner DC, Evans NP 1999 Progesterone can block transmission of the estradiol-induced signal for LH surge generation during a specific period of time immediately after activation of the GnRH surge generating system. Endocrinology 140:827– 834 Smart D, Forhead AJ, Smith RF, Dobson H 1994 Transport stress delays the

46.

47. 48.

49. 50.

51.

52.

53.

2479

oestradiol-induced LH surge by a non-opioidergic mechanism in the early postpartum ewe. J Endocrinol 142:447– 451 Medina CL, Nagatani S, Darling TA, Bucholtz DC, Tsukamura H, Maeda K, Foster DL 1998 Glucose availability modulates the timing of the LH surge in the ewe. J Neuroendocrinol 10:785–792 Kalra SP 1975 Observations of facilitation of the preovulatory rise of LH by estrogen. Endocrinology 96:23–28 Krey LC, Parsons B 1981 Characterization of estrogen stimuli sufficient to initiate cyclic luteinizing hormone release in acutely ovariectomized rats. Neuroendocrinology 34:315–322 Greisman SE 1983 Induction of endotoxin tolerance. In: Nowotny A (ed) Beneficial Effects of Endotoxin. Plenum Press, New York, pp 149 –178 Moberg GP, Watson JG, Stoebel DP, Cook R 1981 Effect of cortisol and dexamethasone on the oestrogen-induced release of luteinizing hormone in the anoestrous ewe. J Endocrinol 90:221–225 Pant HC 1977 Effect of oestradiol infusion on plasma gonadotrophins and ovarian activity in progesterone-primed and unprimed anoestrous ewes. J Endocrinol 75:227–233 Bowen JM, Dahl GE, Evans NP, Thrun LA, Wang Y, Brown MB, Karsch FJ 1998 Importance of the gonadotropin-releasing hormone (GnRH) surge for induction of the preovulatory luteinizing hormone surge of the ewe: doseresponse relationship and excess of GnRH. Endocrinology 139:588 –595 Braden TD, Fry C, Sartin JL 1998 Effects of interleukins on secretion of luteinizing hormone from ovine pituitary cells. Am J Vet Res 59:1488 –1493