A.C.O. Evans, 3 C.M. Komar, S-A. Wandji, and J.E. Fortune 2 BIOLOGY OF REPRODUCTION 57, (1997)

BIOLOGY OF REPRODUCTION 57, 394-401 (1997) Changes in Androgen Secretion and Luteinizing Hormone Pulse Amplitude Are Associated with the Recruitment and Growth of Ovarian Follicles during the Luteal Phase of the Bovine Estrous Cycle' A.C.O. Evans, 3 C.M. Komar, S-A. Wandji, and J.E. Fortune 2 Department and Section of Physiology, College of Veterinary Medicine and Division of Biological Sciences, Cornell University, Ithaca, New York 14853 ABSTRACT In cattle the development of large antral follicles occurs in two or three successive waves during the estrous cycle, with one follicle per wave selected for dominance. To test the hypothesis that negative feedback effects of steroids secreted by the dominant follicle are critical to the regulation of follicular waves, we examined temporal relationships among ovarian follicular growth, steroid secretion, and gonadotropin secretion. Follicular growth was monitored by ultrasonography. In the first experiment, blood was collected from 5 Holstein heifers every 8 h between Days 0 and 14 of the estrous cycle from both a jugular vein and the vena cava (to collect ovarian blood). Jugular samples were also collected every 12 min for 8 h during three periods (Days 3 or 4, 7 or 8, and 11, 12, or 13; n = 6) to characterize the pulsatile pattern of LH secretion. Both estradiol and testosterone concentrations in the vena cava increased as prewave elevations in FSH concentrations decreased (p < 0.05) between Days 1 and 4 (first follicular wave) and between Days 9 and 12 (second follicular wave). LH pulse amplitude was greater during the second period of frequent blood collection (Day 7 or 8, end of the growth phase of the first dominant follicle) compared to the other two periods (p < 0.05), suggesting that increased LH pulse amplitude may be important for the later stages of dominant follicle growth. In the second experiment, to determine whether ovarian steroids are secreted primarily by dominant follicles, blood samples were collected from the utero-ovarian veins draining ovaries with (n = 4) and without (n = 4) a dominant follicle during the first follicular wave. Testosterone, androstenedione, and estradiol concentrations in the utero-ovarian veins fluctuated in relation to the pattern of follicular growth (p < 0.05), and secretion was much greater from ovaries with a dominant follicle. In blood collected both from the vena cava and from the utero-ovarian veins, estradiol secretion reached a peak and started to decline before androgen concentrations peaked (p < 0.05), suggesting that the initial decrease in estradiol secretion from the dominant follicle is not due to a lack of androgen precursors. The results suggest that 1) a transient increase in LH pulse amplitude during the earlymidluteal phase may be important for supporting the final stages of dominant follicle growth; 2) ovarian androgens, as well as estradiol, may play a critical role in the control of FSH secretion during waves of follicular development; and 3) the dominant follicle is responsible for fluctuations in circulating estradiol and androgens during follicular waves. INTRODUCTION It is now well established that in cattle the development of large antral follicles occurs in two or three successive Accepted April 1, 1997. Received May 14, 1996. 'This research was supported by a grant from the USDA (90-37240- 5715). 2Correspondence: T9-023 VRT, Cornell University, Ithaca NY 14853. FAX: (607) 253-3476; e-mail: jfl 1 @cornell.edu 'Current address: Department of Animal Science and Production, University College Dublin, Ireland. 394 waves during the estrous cycle (reviewed in [1-3]). However, the mechanisms that allow one follicle from each wave of recruited follicles to become dominant and to suppress, not only the subordinate follicles of the same wave but also any further growth of follicles > 5 mm, are not understood. It has been hypothesized that the dominant follicle suppresses subordinate follicles on both ovaries either directly, by secreting something that inhibits the growth of other follicles, or indirectly, by secreting hormones that exert negative feedback on gonadotropin secretion (reviewed in [3]). There is no convincing evidence for direct inhibition by the dominant follicle (reviewed in [3]). Although there is more evidence for regulation of waves of follicular development via negative feedback by ovarian products, there are important gaps in our knowledge of how negative feedback acts to facilitate continuous and regular waves of follicular growth. Of major importance was the initial observation by Adams et al. [4] that each follicular wave is preceded by a small elevation in FSH. Although the increase in FSH that precedes the first follicular wave of the cycle had long been recognized as the "secondary surge" of FSH [5, 6], the elevations that precede subsequent waves had previously been undetected because the second and third waves do not occur as synchronously among heifers as does the first. Experimentally delaying the secondary surge of FSH delays the first follicular wave and the increase in estradiol that occurs during the first few days of that wave [7]. One might hypothesize, therefore, that the small rises in FSH prior to waves initiate follicular recruitment and that the increase in estradiol that accompanies recruitment then decreases FSH by negative feedback. Estradiol then declines due to lack of FSH, and this is followed by the next rise in FSH and growth of a new follicular wave. However, estradiol appears to peak and decline several days before the next rise in FSH [4, 7], suggesting that other follicular products (inhibin, androgens?) also contribute to negative feedback. In addition, changes in LH pulse amplitude that are not explained by negative feedback of luteal progesterone have been reported [8], but their relationship to follicular dynamics is unknown. Hence a number of factors may play a role in the regulation of follicular development by negative feedback. Since there is a lack of studies that examine changes in all three classes of ovarian steroids and gonadotropins relative to each other, and simultaneously to follicular patterns throughout a follicular wave, we have determined these relationships using blood collected from the vena cava to better detect changes in ovarian hormones. In addition, since plasma estradiol peaks quite early during follicular waves, it is unclear whether circulating factors that may exert negative feedback are secreted primarily by the dominant follicle or by the follicular wave in toto. To answer this im-

GONADOTROPINS, STEROIDS, AND FOLLICLE GROWTH IN CATTLE 395 portant question, blood samples were obtained from both ovarian veins during follicular waves, and steroid secretion by ovaries bearing a dominant follicle was compared to secretion by ovaries without a dominant follicle. MATERIALS AND METHODS Holstein heifers with prior records of regular estrous cycles were used. Animals were checked for estrous behavior once daily (estrus = Day 0 of the cycle). Prior to surgery, anesthesia was induced with thiopental sodium (Sanofi Winthrop, New York, NY) and was maintained with 3-5% halothane (Hallocarbon Laboratories, River Edge, NJ) in oxygen. Two experimental procedures were used to obtain blood samples containing high concentrations of ovarian hormones. In experiment 1, catheters were placed in the vena cava to collect blood from the point where the ovarian venous drainage enters the vena cava. In experiment 2, the utero-ovarian vein draining each ovary was cannulated. These approaches were used to provide accurate and sensitive assessments of the temporal patterns of ovarian steroid secretion and, in experiment 2, to determine differences in secretion between ovaries with and without a dominant follicle. All animal procedures were approved by the Cornell University Animal Care and Use Committee (Protocol #86-214-95). Blood samples were collected into tubes containing sodium citrate. Plasma was separated by centrifugation within 20 min of collection, and plasma samples were stored at -20 C until analyzed. For the duration of each experiment the ovaries of each heifer were examined daily by transrectal ultrasonography using a 7.5-MHz linear array transducer and a B-Mode scanner (Aloka 500V; Corometrics Medical Systems Inc., Wallingford, CT) to monitor changes in ovarian follicular growth [9, 10]. To determine whether surgery and blood collection had effects on follicular dynamics, follicle growth was compared between the experimental animals and a contemporaneous group of untreated heifers in each experiment (n = 4 and n = 3, respectively). Experiment 1 (Cannulation of Vena Cava) Six heifers were injected with prostaglandin F 2 (25 mg Lutalyse, i.m.; Upjohn, Kalamazoo, MI) in order to induce luteolysis and start a new estrous cycle. Between Days 3 and 5 of the cycle, one medial saphenous vein was cannulated in each heifer [11]. Briefly, a sleeve (Silastic tubing, length 55 cm, i.d. 2.3 mm, o.d. 4.1 mm; Cole-Parmer, Niles, IL) with a wider cuff (length 1.5 cm, o.d. 5.9 mm, glued 25 cm from one end) was fed into a small incision in one medial saphenous vein (15-20 cm from the inguinal region) and secured at the depth of the cuff. The free end of the sleeve was exteriorized through an s.c. tunnel 10-15 cm distal to the primary incision. To determine the distance from the end of the saphenous vein sleeve to the position where the utero-ovarian veins drain into the vena cava, a 140-cm catheter (Tygon tubing, i.d. 1.0 mm, o.d. 1.8 mm; Cole-Parmer) was fed through the saphenous vein sleeve during the midluteal phase (range, Day 10-13 of the cycle) to a depth of 130 cm. Blood samples (5 ml) were collected at intervals of 5 cm from 130 to 90 cm as the catheter was withdrawn; a blood sample was also collected from a jugular vein. The correct vena cava catheter length was determined for each heifer as the length at which the blood sample with the highest plasma progesterone concentration was collected [12, 13]. Between Days 13 and 16 of the cycle, heifers were injected with 25 mg prostaglandin F 2 a (i.m.) to start a second cycle. On Day 0 of the second cycle (14.0 0.4 days after placement of the saphenous vein sleeve), a vena cava catheter of predetermined specific length (Tygon tubing; i.d. 1.0 mm, o.d. 1.8 mm) was inserted through the saphenous vein sleeve (with the aid of a wire guide; Cook Inc., Bloomington, IN), and an indwelling catheter (Micro-Renathane tubing, i.d. 1.0 mm, o.d. 2.0 mm; Braintree Scientific, Braintree, MA) was placed in one jugular vein of each heifer. Starting on Day 0, blood samples (10 ml) were collected every 8 h (at 0700 h, 1500 h, and 2300 h) until Day 14 of the cycle from both the jugular vein catheters and the vena cava catheters. To determine LH pulse characteristics, blood samples (10 ml) were collected from the jugular vein catheters every 12 min for 8 h (0700-1500 h) on Days 3 or 4, 7 or 8, and 11, 12, or 13 of the cycle. Ultrasonographic examinations of the ovaries continued until the end of the second cycle. Saphenous vein sleeves, vena cava catheters, and jugular vein catheters were flushed daily, or after the collection of each blood sample, with 5 ml of physiological saline solution containing 50 000 IU/ml penicillin G potassium (Marsam Pharmaceuticals Inc., Cherry Hill, NJ) and 200 IU/ml heparin (Elkins-Sinn Inc., Cherry Hill, NJ). During frequent blood collections, jugular vein catheters were flushed with saline containing 100 IU/ml heparin. If a vena cava catheter became blocked it was replaced with a new catheter of specific length for that heifer via the saphenous vein sleeve. Experiment 2 (Cannulation of Utero-Ovarian Veins) Between Days 2 and 4 (mean day 2.8 ± 0.4) of a natural estrous cycle, the utero-ovarian veins draining each ovary were cannulated (Micro-Renathane tubing, i.d. 0.8 mm, o.d. 1.6 mm, length 130 cm; Braintree Scientific) in five heifers via a midventral incision [14]. Care was taken to ensure that the tip of the catheter did not extend into the vena cava. The cannulas were externalized through a high sublumbar stab incision. Starting the evening after surgery, blood samples (10 ml) were collected at 8-h intervals from utero-ovarian vein catheters. Throughout the blood collection period, catheters were continuously infused with physiological saline containing 100 IU/ml heparin and 10 000 IU/ml penicillin G (3-5 ml/h). Occasionally, blood samples could not be obtained due to temporarily plugged cannulae, but at least two blood samples were obtained per day from each cannula. RIAs Using methods long established in our laboratory, LH [15, 16], FSH [16, 17], estradiol [7], and progesterone [16, 18] were measured by RIA. Procedures and assay sensitivities were as described previously, except that for the LH and FSH assays the standards were made up in assay buffer containing 5% BSA (Sigma Chemical Co., St. Louis, MO). Intra- and interassay coefficients of variation varied between 3% and 9%. Androstenedione [19] and testosterone [19, 20] concentrations were determined in ether extracts of 200-p1l aliquots of plasma. Recovery of tritiated androstenedione or testosterone was 91% and 92%, respectively, and values in ether blanks were negligible. When between 35 and 300 pg of androstenedione or testosterone was added to 200 Ipl of steroid-free plasma, the mean recovery efficiencies were

396 EVANS ET AL. 91% and 94%, respectively. Values were not corrected for recovery. The sensitivity of both the androstenedione and testosterone assays was 30 pg/ml of plasma. The intra- and interassay coefficients of variation were 3% and 5%, respectively, for androstenedione (n = 5 assays) and 3% and 8%, respectively, for testosterone (n = 6 assays), based on the results for four quality control samples (which were low, intermediate, or high relative to the standard curve) assayed in duplicate in each RIA. At_ I u. E 30. e c 0 20 Vena Cava Data Analyses and Statistics Based on ultrasonographic observations, the day of emergence of a follicular wave was defined as the first day on which the dominant follicle of that wave was retrospectively identified at 5 mm in diameter, and the end of the growth phase of a follicle was the day on which its progressive increase in diameter ceased [10, 21]. The pulsatile pattern of LH secretion in samples collected frequently from the jugular vein (every 12 min for 8-h periods) was defined by standard deviation criteria [22] using PC-Pulsar (J. Gitzen and V. Ramirez, University of Illinois, Urbana, IL) and is described in terms of mean concentrations, pulse frequency, pulse amplitude, and basal concentrations (after subtraction of the pulses). The standard deviation criteria (G) were G(1) 13.20, G(2) 7.80, G(3) 5.76, G(4) 4.38, and G(5) 3.39 [22]. In experiment 2, concentrations of steroids in samples collected at 24-h intervals, relative to the peak steroid concentration, were used for statistical analyses; this approach emphasized the major changes over time and between the two ovaries and obviated the problem of missing samples due to temporarily plugged cannulae. All data were examined by repeated measures ANOVA using the General Linear Models procedure of SAS (SAS Institute, Cary, NC). For hormone profile data, each model included the main effects of animal and day. When the days of peak and nadir concentrations were compared, the model also included hormone (experiment 1) and ovary (with or without a dominant follicle, experiment 2). Interactions were also considered but were excluded when not significant (p > 0.05). Post-ANOVA comparisons were made using Tukey's HSD (p < 0.05). Correlation analyses were used to examine the relationships between various hormones using the PROC CORR procedure or the General Linear Models procedure of SAS (including animal in the model statement). Pearson product moment correlation coefficients (r) are given. All values given are the mean + SEM. RESULTS Experiment (Cannulation of Vena Cava) All heifers exhibited normal reproductive function throughout the experimental period. Mean length of the estrous cycle was 21.5 0.2 days (n = 6); four heifers had three waves of follicular development, whereas two heifers had two waves of follicular development. Waves 1 and 2 emerged on Days 1.3 0.2 and 10.7 + 0.5 of the cycle, respectively, and the end of the growth phase of the dominant follicle of the first follicular wave was on Day 7.3 + 0.5 of the cycle. The pattern and characteristics of follicle growth in cannulated heifers (n = 6) were not different from those of uncannulated control heifers (n = 4, data not shown). For illustrative purposes, progesterone concentrations in samples collected every 5 cm from the vena cava were (L o Jugulal O 90 100 110 120 130 Distance from saphenous vein sleeve (cm) FIG. 1. Mean ( SEM) progesterone concentrations in blood samples collected from the jugular vein and the vena cava (at given distances from the end of a catheter sleeve placed in a saphenous vein) in heifers (n = 6) during the midluteal phase of the estrous cycle. Progesterone concentrations in the vena cava are aligned to the peak concentration in each animal at a mean distance of 110.0 2.6 cm (arrow). aligned to the peak concentrations in each profile and plotted to the mean distance at which the highest values were measured (Fig. 1). The mean distance from the end of the saphenous vein sleeve to the point in the vena cava where the greatest progesterone concentrations were measured was 110.0 2.6 cm (Fig. 1), and the range was 105-120 cm; however, catheters of specific length for each animal were used. In one heifer the saphenous vein sleeve came out on Day 4 of the cycle; consequently her vena cava data were omitted from the analyses. Jugular blood samples were collected for the duration of the experiment from all six animals. The growth profiles of the dominant and two largest subordinate follicles were determined for the first and second waves of follicular development. Mean values were normalized to the mean day of emergence of each new follicular wave [10, 21]. In a similar way, jugular concentrations of FSH and both jugular and vena cava concentrations of estradiol, testosterone, and androstenedione were aligned to the mean days of wave emergence. There were significant fluctuations in FSH concentrations in the jugular vein and in estradiol and testosterone concentrations in the vena cava, associated with the emergence of follicular waves (all p < 0.0001; Fig. 2). FSH concentrations were high between Days 1 and 2 and between Days 9 and 10, and were lower (p < 0.05) between Days 3 and 7 and after Day 12 of the cycle (Fig. 2). Estradiol concentrations in blood samples collected from the vena cava were low on Day 1 and between Days 7 and 10 of the cycle and showed a transient increase (p < 0.05) between Days 3 and 5 of the cycle and a nonsignificant increase between Days 11 and 12 of the cycle (Fig. 2). Testosterone concentrations in blood samples collected from the vena cava were lowest on Day 1 of the cycle, increased (p < 0.05) to peak at about Day 6, were at intermediate concentrations between Days 7 and 10 of the cycle, and peaked again on Day 13 of the cycle (Fig. 2). Mean concentrations of androstenedione in blood samples collected from the vena cava did not show consistent changes during the sampling period (p > 0.40; Fig. 2); however, profiles from individual animals showed fluctuations. In samples collected from the jugular vein, estradiol concentrations were greatest on Day 5 of the cycle (p <

GONADOTROPINS, STEROIDS, AND FOLLICLE GROWTH IN CATTLE 397, - ge 16 12.2 U 8- S! 8 u- Is E 16- -IE 2'6 12- U. 8-141.= ' 0 110U cd C i _ f I--......... I 30 0 2 4 6 8 10 12 14 Day of cycle FIG. 3. Days of maximum and minimum jugular FSH and vena cava estradiol and testosterone concentrations (means + SEM, n = 5) in heifers. a,b,c,d,e,f, Means with no common letter are different days of the cycle both within and between hormones (p < 0.05). 0 2 4 6 8 10 12 14 Day of cycle FIG. 2. Mean (+ SEM) diameters of the dominant and two largest subordinate follicles of the first and second follicular waves (n = 6), jugular FSH concentrations (n = 6), and vena cava concentrations of estradiol, testosterone, and androstenedione (n = 5) in cycling heifers. Hormone concentrations were normalized to the mean days of emergence of the first and second follicular waves. Striped boxes indicate the mean period during which frequent blood samples were collected (see Table 1). 0.05; 2.98 0.65 pg/ml); testosterone concentrations increased linearly from Day 1 (43.6 3.3 pg/ml) to Day 6 of the cycle (p < 0.05; 56.4 4.0 pg/ml) and were not significantly different thereafter; and androstenedione concentrations were not different among days (p < 0.20; overall mean 99.4 ± 4.0 pg/ml). For concentrations of FSH in the jugular vein and of estradiol and testosterone in the vena cava, the peak and nadir concentrations and the days on which they occurred were determined from individual profiles, and these days were compared within and between hormones (Fig. 3). The mean day of the cycle on which FSH concentrations were lowest (Day 4.3 ± 0.7) was not different (p > 0.05) from the days on which either estradiol or testosterone concentrations were greatest (Days 4.0 0.5 and 6.4 0.7, respectively); however, the days of peak estradiol and testosterone concentrations were different (p < 0.05; Fig. 3). The days on which estradiol and testosterone concentrations were at nadir values (Days 8.7 0.7 and 8.6 0.7, respectively) were not different (p > 0.05) from the day on which FSH concentrations had reached a second peak (Day 9.0 0.3; Fig. 3). Correlation analyses between hormone concentrations on the days of peak and nadir values (Fig. 3) showed significant negative correlations between FSH and both estradiol (r = -0.66, p < 0.001) and testosterone concentrations (r = -0.51, p < 0.02). The mean days on which intensive blood sampling was conducted (Days 3.7 0.2, 7.7 + 0.2, and 12.0 0.2 of the cycle; Table 1) correspond to the middle and end of the growth phase of the first dominant follicle and the early growth phase of the second dominant follicle. Mean LH concentrations and LH pulse frequency decreased between the first and second sampling periods (p < 0.05), whereas LH pulse amplitude increased from the first to the second sampling period (p < 0.05) but had decreased again by the third sampling period (p < 0.05; Table 1). The mean jugular or vena cava concentrations of steroids during each period of intensive blood sampling were calculated as the mean concentration in plasma samples col- TABLE 1. Pulsatile characteristics of LH secretion and steroid concentrations in vena cava or jugular vein samples associated with each period of intensive blood sampling.* Periods of intensive blood sampling (day of cycle) Characteristic (3.7 + 0.2) (7.7 + 0.2) (12.0 + 0.2) Mean LH (ng/ml) 0.98 + 0.14' 0.70 + 0.1 4 b 0.72 + 0. 15 b Basal LH (ng/ml) 0.75 + 0.07' 0.62 + 0.08' 0.68 + 0.13 a LH pulse frequency (pulses/h) 0.65 + 0.08' 0.36 _+ 0.07 b 0.42 + 0.07"b LH pulse amplitude (ng/ml) 0.67 + 0. 13 b 1.04 + 0.06a 0.68 + 0. 08 b Vena cava estradiol (pg/ml) 6.9 ± 1.5 a 1.9 + 0. 3 b 3.6 + 1.2"b Vena cava testosterone (pg/ml) 58.7 + 3.7a 76.8 + 9.8a 77.8 + 7.6a Vena cava androstenedione (pg/ml) 114.7 + 5.9 a 153.8 + 13.1a 144.9 ± 8.1a Jugular progesterone (ng/ml) 0.43 + 0.11c 1.68 + 0.15 b 3.26 + 0.27' abc Values within rows with no common superscript are different (p < 0.05). * Samples were collected from the jugular vein every 12 min for 8 h on Days 3 or 4 (I), 7 or 8 (11), and 11, 12, or 13 (111) of the cycle to determine the pulsatile characteristics of LH secretion (n = 6); steroid concentrations in the jugular vein (n = 6) or vena cava (n = 5) were the mean concentrations in samples collected at the beginning and end of each 8-h sampling period.

398 EVANS ET AL. TABLE 2. Correlations between the pulsatile characteristics of LH secretion and estradiol concentrations in the vena cava and progesterone concentrations in the jugular vein (n = 5).* Vena cava estradiol Jugular progesterone Correlation Correlation Characteristic coefficient (r) p coefficient (r) p Mean LH +0.90 0.001-0.95 0.001 LH pulse frequency +0.59 0.078 0.75 0.007 LH pulse amplitude -0.63 0.049 +0.46 0.247 * Samples were collected from the jugular vein every 12 min for 8 h on Days 3 or 4, 7 or 8, and 11, 12 or 13 of the cycle to determine the pulsatile characteristics of LH secretion; steroid concentrations were the mean in plasma samples collected at the beginning and end of each 8-h sampling period. lected at the beginning and at the end of each intensive sampling period (Table 1). As expected, jugular progesterone concentrations increased from the first to the second period and from the second to the third period (p < 0.05, Table 1). There were significant negative correlations between jugular progesterone concentrations and LH pulse frequency (r = -0.75, p < 0.01, Table 2) and mean LH concentrations (r = -0.95, p < 0.001, Table 2). Vena cava estradiol concentrations, associated with the periods of intensive blood sampling, decreased between the first and second period (p < 0.05) and showed a nonsignificant increase during the third period (Table 1). Vena cava estradiol concentrations and LH pulse amplitude were negatively correlated (r = -0.63, p < 0.05, Table 2). In contrast, concentrations of estradiol in the vena cava were positively correlated with mean LH concentrations (r = 0.90, p < 0.001, Table 2), and there was a tendency for a positive correlation with LH pulse frequency (r = 0.59, p < 0.08, Table 2). Mean concentrations of androgens in the vena cava did not differ among the three periods of intensive blood sampling; hence changes in circulating androgen concentrations were not associated with the changes observed in mean LH or LH pulse frequency. The observation that testosterone concentrations were the same on Days 3-4 (period I of intensive blood sampling) as on Days 7-8 (period II), whereas estradiol concentrations had decreased 3.5-fold between periods I and II, is consistent with the temporal dissociation between the peaks in vena cava estradiol and testosterone during the first follicular wave shown in Figures 2 and 3. Experiment 2 (Cannulation of Utero-Ovarian Veins) The mean day of emergence of the first follicular wave was Day 1.6 + 0.2 of the cycle, and the end of the growth phase of the dominant follicle of the first follicular wave was Day 7.5 0.3 of the cycle (n = 5). The pattern and characteristics of follicle growth in cannulated heifers (n = 5) were not different from those of uncannulated control heifers (n = 3, data not shown). Based on ultrasonographic recordings, the ovaries in each heifer were described as either containing a developing dominant follicle (dominant ovary) or not containing a dominant follicle (nondominant ovary). From the five animals cannulated, complete steroid hormone profiles were obtained from four ovaries with a dominant follicle and from four ovaries without a dominant follicle. To emphasize the major differences over time, between ovaries, and among steroid hormones, steroid concentrations in samples collected at 24-h intervals, relative to peak concentrations, were aligned and plotted for ovaries with and without a dominant follicle (Fig. 4). ---- ZUUU E 1500 1000. C u E soo uj ---- 1 z -- E I a).o 0 a) 400-0- E 2000- - 1500- c 0 X5 1000- C ( g0 500- C a * Ovary with dominant follicle * Ovary without dominant follicle a b~~~~ab ab b ab z YZ. Z z. _ ' ' ' 4 5 6 7 8 9 b Day of cycle FIG. 4. Mean (+ SEM) concentrations of estradiol, testosterone, and androstenedione in plasma samples collected from the utero-ovarian veins draining ovaries with (n = 4) or without (n = 4) a dominant follicle during the first follicular wave of the estrous cycle. a,b or y,z, Means with no common letter are significantly different (p < 0.05) within a hormone profile. Steroid concentrations in samples collected from the utero-ovarian vein draining ovaries with a dominant follicle were greater than those without a dominant follicle (p < 0.001; Fig. 4); the differences were very dramatic during the early growth phase of the dominant follicle. Each steroid hormone profile showed significant fluctuations with time (p < 0.02) except for estradiol concentrations in samples collected from ovaries without a dominant follicle (p > 0.20; Fig. 4). When the days on which steroid concentrations were at their greatest were compared among steroids and between ovaries (with and without a dominant follicle), there was no difference between ovaries (relative to day of the estrous cycle or day of the follicular wave; p > 0.45). When data for the two ovaries were combined, estradiol concentrations peaked earlier than androstenedione concentrations (p < 0.05, Day 4.9 + 0.2 vs. 5.9 0.3 of the cycle), and testosterone concentrations peaked at an intermediate time (Day 5.6 0.3 of the cycle). When the peak concentrations were considered relative to the day of wave emergence, estradiol concentrations peaked earlier (day 3.1 0.3 of the wave, p < 0.05) than either testosterone (day 3.9 0.4 of the wave) or androstenedione (day 3.9 0.5 of the wave) concentrations. DISCUSSION - The results show that ovarian androgen secretion fluctuated in relation to the pattern of follicle growth and that androgen secretion was inversely related to circulating FSH concentrations. This suggests that androgens (as well as estradiol) exert negative feedback on FSH secretion. In ad- b b

GONADOTROPINS, STEROIDS, AND FOLLICLE GROWTH IN CATTLE 399 dition, the ovary with the dominant follicle was the major source of ovarian steroids during nonovulatory follicular waves. There was a transient increase in LH pulse amplitude after selection of the dominant follicle of the first follicular wave but before the emergence of the second follicular wave, and it appears that this increase is regulated by ovarian estradiol production. Not only were estradiol concentrations inversely related to FSH concentrations, as expected, but testosterone concentrations showed a similar relationship (Fig. 3), and it was the ovary with the dominant follicle that was responsible for the increases in secretion of ovarian estradiol and androgens into the blood (Fig. 4). When androstenedione concentrations in the vena cava were aligned to the pattern of follicle development, no clear pattern was observed (Fig. 2); however, clear changes were observed in samples collected from the utero-ovarian veins (Fig. 4). Nonetheless, the aligned data from the vena cava did show that androstenedione concentrations, like testosterone concentrations, were low about the time of wave emergence (Days 0-2 and 8-10) and higher during the periods of follicle growth and regression (Fig. 2). Therefore, both testosterone and androstenedione were secreted from the ovaries into the peripheral circulation in association with the pattern of follicle development. Negative feedback effects of the elevated concentrations of androgens later during the follicular wave may explain the lag between the earlier estradiol peak and the rise in FSH concentrations that precedes the next follicular wave. We suggest that ovarian androgens as well as estradiol, secreted mainly from the developing dominant follicle, modulate FSH secretion and thereby play a regulatory role in the emergence and growth of subsequent follicular waves. Elevated androgen secretion late in the growth phase of a follicle could also have a functional role at the level of the ovary. In the present study the peak in testosterone secretion occurred about the same time as the end of the growth phase of the dominant follicle of the first follicular wave. Deleterious effects of androgens on follicle growth have been observed in rats [23-25], and androgens can inhibit FSH-induced LH-receptor formation [26, 27] and hcg-stimulated aromatase activity [28]. Atresia of dominant follicles is associated with a decrease in LH-receptor binding [29, 30] and in the expression of mrna for LH receptors [31]. Hence, high androgen secretion toward the end of the growth phase of a dominant follicle may cause the loss of LH receptors and eventually lead to atresia. There was an inverse relationship between ovarian estradiol secretion and LH pulse amplitude (Table 2). To date, this has not been reported in untreated heifers during the estrous cycle. High LH pulse amplitude during the early part of the midluteal phase and in the late luteal phase has been reported recently [8, 32] and was observed toward the end of the growth phase of the dominant follicle in the present experiment (Table 1, Fig. 1). Estrogen-active follicles have greater concentrations of gonadotropin binding sites than estrogen-inactive follicles [30], and mrna for LH receptors in granulosa and theca cells of dominant follicles increases during development of the dominant follicle [31]. It has been suggested that once FSH concentrations reach a minimum, the dominant follicle becomes dependent on LH for its continued growth and development [33]. This hypothesis is supported by experiments in which the suppression of LH secretion impeded the later stages of dominant follicle development or caused atresia of the dominant follicle [34-36]. Hence, in the early luteal phase, when mean LH concentrations and LH pulse frequency are decreasing (Table 1), a short period of high-amplitude LH pulses may be critical for maintaining the later stages of dominant follicle growth. In contrast to this idea, high-amplitude pulses of LH have been shown to antagonize the stimulatory effects of FSH on follicle growth in ewes [37]. In the present study (Fig. 2) and other studies [7, 32, 38, 39], circulating estradiol concentrations reached a peak and declined well before the dominant follicle stopped growing. The factors involved in the decline in estradiol secretion from the dominant follicle are unclear. It has been suggested that the decline is due to a decrease in androgen substrate [1, 39-41] critical for estradiol production [42]. The present results suggest that in the early period after estradiol reaches its peak, androgens are not limiting for aromatase activity, since androgen secretion into the venous drainage continued to increase even after estradiol secretion started to decrease (Figs. 2, 3, and 4). Studies in vitro have shown that in the early stages of atresia in bovine follicles, granulosa cell aromatase activity decreases before a reduction in thecal androgen production occurs [43]. Granulosa cells of dominant follicles are reported to produce a protein that is not secreted from the follicle and has local aromatase-inhibiting actions [44, 45], and it has been suggested that this inhibitor keeps estradiol secretion by the dominant follicle in check [46]. Similar events regulating estradiol secretion have been suggested to occur in ewes [47, 48]. In the present study, estradiol secretion during the second follicular wave was lower than during the first follicular wave (Figs. 2 and 3). Mean LH concentrations and LH pulse frequency are reduced by progesterone (Table 2; [8, 49-51]), and mean LH concentrations and LH pulse frequency have been correlated with circulating estradiol concentrations (Table 2; [12, 51]). Hence, low estradiol secretion from the second wave of follicular development appears to be due indirectly to the effects of high progesterone concentrations that decrease LH secretion. On the basis of the data in the present study and information in the literature, we propose the following model for the pattern of endocrine events associated with the growth of dominant follicles in cattle. In the milieu of low circulating concentrations of estradiol and progesterone on Day 1 of the cycle, an increase in FSH secretion stimulates the emergence of the first wave of follicular development [4, 7, 10, 35, 52]. As the dominant follicle grows, peripheral estrogen and androgen concentrations increase and have negative feedback effects decreasing FSH secretion (Fig. 2; [38]). After estradiol secretion reaches a peak (about Day 4), the dominant follicle continues to grow in an environment of basal FSH (Figs. 2 and 3; [4, 32, 52]) and decreasing mean LH concentrations and LH pulse frequency (due to increasing progesterone secretion from the corpus luteum [8, 51]) but increasing LH pulse amplitude (Table 1). An increase in LH receptors [30, 31] enables the growing follicle to respond to higher LH pulse amplitude (due to low estradiol secretion) for continued growth. FSH secretion is kept low by continued high androgen secretion (Figs. 2 and 3). Under the influence of high androgen production the dominant follicle enters a static phase of growth (around Day 8). As the static phase continues, androgen secretion decreases, allowing FSH secretion to increase (Figs. 2 and 3). This stimulates the emergence of the second wave of follicular development around Day 11 (Fig. 2; [4, 35, 52]). LH pulse amplitude is reduced as estradiol secretion increases (Table 1). Estradiol, together with increased androgen secretion, again depresses circulating FSH con-

400 EVANS ET AL. centrations. Cyclic changes in steroid and gonadotropin concentrations continue to maintain a pattern of follicular waves until luteolysis occurs and a follicular phase is established. Clearly, some parts of the model presented above have strong experimental support, whereas others require additional testing. In addition, it should be emphasized that inhibin secretion by follicles may play an important part in the regulation of FSH secretion and follicular dynamics. However, elucidation of its role during the bovine estrous cycle awaits the development of sensitive and specific assays. In summary, ovarian androgen secretion was inversely related to FSH secretion, suggesting that ovarian androgens, as well as estradiol, closely regulate the fluctuating pattern of FSH secretion and the emergence and growth of follicular waves. Androgen secretion, associated with the growth of the dominant follicle of the first wave, peaked later than estradiol secretion, implying that the initial decrease in estradiol secretion did not occur due to a reduction in androgen substrate. Ovarian estradiol secretion was inversely related to LH pulse amplitude, which was transiently high during the midluteal phase and may be important for the final stages of dominant follicle growth. The dominant follicle (as opposed to the subordinate follicles) was the major source of the fluctuations in circulating steroid concentrations and therefore is primarily responsible for the negative feedback effects of ovarian steroids during waves of follicular development. ACKNOWLEDGMENTS The authors thank R.R. Saatman, T.L. Kimmich, R.O. Gilbert, S.L. Fubini, TE Robinson, K.L. Bigelow, and D. Bianchi for surgical assistance and care of the animals; G.D. Niswender for the LH and testosterone antibodies; L.E. Reichert for purified LH; D.J. Bolt for the FSH antibody and purified FSH; D.T. Armstrong for the androstenedione antibody; and W.R. Butler for reading the manuscript. REFERENCES I. Lucy MC, Savio JD, Badinga L, de la Sota RL, Thatcher WW. Factors that affect ovarian follicular dynamics in cattle. J Anim Sci 1992; 70: 3615-3626. 2. Webb R, Gong JG, Law AS, Rusbridge SM. Control of ovarian function in cattle. J Reprod Fertil Suppl 1992; 45:141-156. 3. Fortune JE. Ovarian follicular growth and development in mammals. Biol Reprod 1994; 50:225-232. 4. Adams GP, Matteri RL, Kastelic JP, Ko JCH, Ginther OJ. Association between surges of follicle-stimulating hormone and the emergence of follicular waves in heifers. J Reprod Fertil 1992; 94:177-188. 5. Dobson H. Plasma gonadotrophins and oestradiol during oestrus in the cow. J Reprod Fertil 1978; 52:51-53. 6. Roche JF, Ireland JJ. The differential effect of progesterone on concentrations of luteinizing hormone and follicle-stimulating hormone in heifers. Endocrinology 1981; 108:568-572. 7. Turzillo AM, Fortune JE. Suppression of the secondary FSH surge with bovine follicular fluid is associated with delayed ovarian follicular development in heifers. J Reprod Fertil 1990; 89:643-653. 8. Cupp AS, Stumpf TT, Kojima FN, Werth LA, Wolfe MW, Roberson MS, Kittok RJ, Kinder JE. Secretion of gonadotrophins change during the luteal phase of the bovine oestrous cycle in the absence of corresponding changes in progesterone or 17j-oestradiol. Anim Reprod Sci 1995; 37:109-119. 9. Sirois J, Fortune JE. Ovarian follicular dynamics during the estrous cycle in heifers monitored by real-time ultrasonography. Biol Reprod 1988; 39:308-317. 10. Evans ACO, Adams GP, Rawlings NC. Endocrine and ovarian follicular changes leading up to the first ovulation in prepubertal heifers. J Reprod Fertil 1994; 100:187-194. 11. Boisclair YR, Bell AW, Bauman DE. Chronic catheterization of external iliac vessels in growing cattle. J Appl Physiol 1993; 74:444-449. 12. Walters DL, Schams D, Schallenberger E. Pulsatile secretion of gonadotrophins, ovarian steroids and ovarian oxytocin during the luteal phase of the oestrous cycle in the cow. J Reprod Fertil 1984; 71:479-491. 13. Benoit AM, Dailey RA. Catheterization of the caudal vena cava via the lateral saphenous vein in the ewe, cow, and gilt: an alternative to utero-ovarian and medial coccygeal vein catheters. J Anim Sci 1991: 69:2971-2979. 14. Ireland JJ, Fogwell RL, Oxender WD, Ames K, Cowley JL. Production of estradiol by each ovary during the estrous cycle of cows. J Anim Sci 1984; 59:764-771. 15. Niswender GD, Reichert LE, Midgley AR, Nalbandov AV. Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 1969; 84:1166-1173. 16. Quirk SM, Fortune JE. Plasma concentrations of gonadotrophins, preovulatory follicular development and luteal function associated with bovine follicular fluid-induced delay of oestrus in heifers. J Reprod Fertil 1986; 76:609-621. 17. Bolt DJ, Rollins R. Development and application of a radioimmunoassay for bovine follicle-stimulating hormone. J Anim Sci 1983; 56: 146-154. 18. Beal WE, Milvae RA, Hansel W. Oestrous cycle length and plasma progesterone concentrations following administration of prostaglandin F-2a early in the bovine oestrous cycle. J Reprod Fertil 1980; 59: 393-396. 19. Fortune JE, Eppig JJ. Effects of gonadotropins on steroid secretion by infantile and juvenile mouse ovaries in vitro. Endocrinology 1979; 105:760-768. 20. Gay VL, Kerlan JT. Serum LH and FSH following passive immunization against circulating testosterone in the intact male rat and in orchidectomized rats bearing subcutaneous silastic implants of testosterone. Arch Androl 1978; 1:257-266. 21. Adams GP, Evans ACO, Rawlings NC. Follicular waves and circulating gonadotrophins in 8-month-old prepubertal heifers. J Reprod Fertil 1994; 100:27-33. 22. Merriam GR, Wachter KW. Algorithms for the study of episodic hormone secretion. Am J Physiol 1982; 243:E310-E318. 23. Payne RW, Runser RH. The influence of estrogen and androgen on the ovarian response of hypophysectomized immature rats to gonadotropins. Endocrinology 1958; 62:313-321. 24. Bagnell CA, Mills TM, Costoff A, Mahesh VB. A model for the study of androgen effects on follicular atresia and ovulation. Biol Reprod 1982; 27:903-914. 25. Louvet J-P, Harman SM, Schreiber JR, Ross GT. Evidence for a role of androgens in follicular maturation. Endocrinology 1975; 97:366-372. 26. Jia X-C, Kessel B, Welsh TH, Hsueh AJW. Androgen inhibition of follicle-stimulating hormone-stimulated luteinizing hormone receptor formation in cultured rat granulosa cells. Endocrinology 1985: 117: 13-22. 27. Farookhi R. Effects of androgen on induction of gonadotropin receptors and gonadotropin-stimulated adenosine 3',5' monophosphate production in rat ovarian granulosa cells. Endocrinology 1980; 106:1216-1223. 28. Shaw HJ, Hillier SG, Hodges JK. Developmental changes in luteinizing hormone/human chorionic gonadotropin steroidogenic responsiveness in marmoset granulosa cells: effects of follicle-stimulating hormone and androgens. Endocrinology 1989; 124:1669-1677. 29. Ireland JJ, Roche JE Development of antral follicles in cattle after prostaglandin-induced luteolysis: changes in serum hormones, steroids in follicular fluid, and gonadotropin receptors. Endocrinology 1982: 111:2077-2086. 30. Ireland JJ, Roche JE Development of nonovulatory antral follicles in heifers: changes in steroids in follicular fluid and receptors for gonadotropins. Endocrinology 1983; 112:150-156. 31. Xu Z, Garverick HA, Smith GW, Smith MF, Hamilton SA, Youngquist RS. Expression of follicle-stimulating hormone and luteinizing hormone receptor messenger ribonucleic acids in bovine follicles during the first follicular wave. Biol Reprod 1995; 53:951-957. 32. Rhodes FM, Fitzpatrick LA, Entwistle KW, Kinder JE. Hormone concentrations in the caudal vena cava during the first ovarian follicular wave of the oestrous cycle in heifers. J Reprod Fertil 1995; 104:33-39. 33. Roche JE Control and regulation of folliculogenesis-a symposium in perspective. Rev Reprod 1996; 1:19-27. 34. Peters KE, Bergfeld EG, Cupp AS, Kojima FN, Mariscal V, Sanchez T, Wehrman ME, Grotjan HE, Hamernik DL, Kinder JE. Suppression

GONADOTROPINS, STEROIDS, AND FOLLICLE GROWTH IN CATTLE 401 of pulsatile secretion of LH diminishes ovarian follicular development in heifers. J Anim Sci 1994; 72(suppl 2):78. 35. Gong JG, Bramley TA, Gutierrez CG, Peters AR, Webb R. Effects of chronic treatment with a gonadotrophin-releasing hormone agonist on peripheral concentrations of FSH and LH, and ovarian function in heifers. J Reprod Fertil 1995; 105:263-270. 36. Manikkam M, Rajamahendran R, Rawlings NC. Effects of bovine follicular fluid and luteinizing-hormone-releasing hormone antagonist on the dominant follicle maintained by a norgestomet ear implant in heifers. Can J Anim Sci 1995; 75:71-78. 37. Picton HM, Tsonis CG, McNeilly AS. The antagonistic effect of exogenous LH pulses on FSH-stimulated preovulatory follicle growth in ewes chronically treated with a gonadotropin-releasing hormone agonist. J Endocrinol 1990; 127:273-283. 38. Kaneko H, Terada T, Taya K, Watanabe G, Sasamoto S, Hasegawa Y, Igarashi M. Ovarian follicular dynamics and concentrations of oestradiol-173, progesterone, luteinizing hormone and follicle stimulating hormone during the periovulatory phase of the oestrous cycle in the cow. Reprod Fertil Dev 1991; 3:529-535. 39. Badinga L, Driancourt MA, Savio JD, Wolfenson D, Drost M, de la Sota RL, Thatcher WW. Endocrine and ovarian responses associated with the first-wave dominant follicle in cattle. Biol Reprod 1992; 47: 871-883. 40. Tian XC, Berndtson AK, Fortune JE. Differentiation of bovine preovulatory follicles during the follicular phase is associated with increases in messenger ribonucleic acid for cytochrome P450 side-chain cleavage, 3-hydroxysteroid dehydrogenase, and P450 17it-hydroxylase, but not P450 aromatase. Endocrinology 1995; 136:5102-5110. 41. Xu Z, Garverick HA, Smith GW, Smith ME Hamilton SA, Youngquist RS. Expression of messenger ribonucleic acid encoding cytochrome P450 side-chain cleavage, cytochrome P450 17cx-hydroxylase, and cytochrome P450 aromatase in bovine follicles during the first follicular wave. Endocrinology 1995; 136:981-989. 42. Fortune JE, Quirk SM. Regulation of steroidogenesis in bovine preovulatory follicles. J Anim Sci 1988; 66(suppl 2):1-8. 43. McNatty KP, Heath DA, Henderson KM, Lun S, Hurst PR, Ellis LM, Montgomery GW, Morrison L, Thurley DC. Some aspects of thecal and granulosa cell function during follicular development in the bovine ovary. J Reprod Fertil 1984; 72:39-53. 44. Shimasaki S, Shimonaka M, Ui M, Inouye S, Shibata E Ling N. Structural characterization of a follicle-stimulating hormone action inhibitor in porcine ovarian follicular fluid. Its identification as the insulin-like growth factor-binding protein. J Biol Chem 1990; 265:2198-2202. 45. Driancourt MA. Lack of between-follicle interactions in the sheep ovary. Reprod Nutr Dev 1994; 34:249-260. 46. Driancourt MA. Local control of follicular growth by ovarian follicles. J Reprod Fertil 1995; Abstract Series 15:2-3. 47. Webb R, England BG. Relationship between LH receptor concentrations in thecal and granulosa cells and in-vivo and in-vitro steroid secretion by ovine follicles during the preovulatory period. J Reprod Fertil 1982; 66:169-180. 48. Campbell BK, Picton HM, Mann GE, McNeilly AS, Baird DT. Effect of steroid- and inhibin-free ovine follicular fluid on ovarian follicles and ovarian hormone secretion. J Reprod Fertil 1991; 93:81-96. 49. Price CA, Webb R. Steroid control of gonadotropin secretion and ovarian function in heifers. Endocrinology 1988; 122:2222-2231. 50. Roberson MS, Wolfe MW, Stumpf TF, Kittok RJ, Kinder JE. Luteinizing hormone secretion and corpus luteum function in cows receiving two levels of progesterone. Biol Reprod 1989; 41:997-1003. 51. Bergfeld EG, Kojima FN, Wehrman ME, Cupp AS, Peters KE, Mariscal V, Sanchez T, Kittok RJ, Garcia-Winder M, Kinder JE. Frequency of luteinizing hormone pulses and circulating 171-oestradiol concentration in cows is related to concentration of progesterone in circulation when the progesterone comes from either an endogenous or exogenous source. Anim Reprod Sci 1995; 37:257-265. 52. Sunderland SJ, Crowe MA, Boland MP, Roche JE Ireland JJ. Selection, dominance and atresia of follicles during the oestrous cycle of heifers. J Reprod Fertil 1994; 101:547-555.