FSH immediately after the end of bff treatment at luteolysis, and they remained above control levels for

Transcription

1 Changes in the secretion of LH pulses, FSH and prolactin during the preovulatory phase of the oestrous cycle of the ewe and the influence of treatment with bovine follicular fluid during the luteal phase J. M. Wallace, G. B. Martin and A. S. McNeilly MRC Reproductive Biology Unit, Centre for Reproductive Biology, University of Edinburgh, 37 Chalmers Street, Edinburgh eh3 9ew (J. M. Wallace is now at The Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen ab2 9sb) (G. B. Martin is now at Animal Science Group, School of Agriculture, University of Western Australia, Nedlands, Western Australia 6009) (Requests for offprints should be addressed to A. S. McNeilly) REVISED MANUSCRIPT RECEIVED 3 AugUSt 1987 ABSTRACT It has previously been shown that treatment of ewes with bovine follicular fluid (bff) throughout the luteal phase of the oestrous cycle lowers plasma levels of FSH but increases the frequency and amplitude of the pulses of LH. Under these conditions, ovarian follicles grow to a maximum diameter of 2\m=.\7 mm and have a reduced capacity to release oestradiol. We have examined the nature of the gonadotrophin signals controlling follicular development in the normally cycling ewe and have investigated the effects of previous exposure to bff on these signals and the follicular responses to them. Control ewes (n l) = injected i.v. with were 9 ml bovine serum and treated ewes were injected with 9 ml bff, twice daily from days 1 to 10 of the luteal phase (day 0 oestrus). The = ewes were injected with prostaglandin analogue on day 11 of the cycle to induce luteolysis and the gonadotrophin patterns were studied in blood sampled from these animals every 10 min for up to 72 h during the subsequent follicular phase. Following luteolysis (and the end of bff treatment), LH pulse frequency increased rapidly in both groups and reached 1 pulse/h within 6 h. Thereafter, pulse frequency increased marginally and reached 1 pulse/50 min by the onset of the LH surge. This pattern was not affected by previous treatment with bff. In the control ewes, the amplitude of the LH pulses did not change significantly following luteolysis or at any time during the follicular phase, while the levels of FSH declined slowly until the onset of the surge. In the treated ewes, on the other hand, there was an immediate increase in both LH pulse amplitude and the concentration of FSH immediately after the end of bff treatment at luteolysis, and they remained above control levels for 24 and 16 h respectively. Plasma prolactin levels did not appear to change around the time of luteolysis but showed a marked and significant diurnal rhythm (nadir around noon and peak around midnight) in both groups. The concentrations of prolactin were significantly (P<0\m=.\001)lower and the preovulatory peak was delayed and reduced in the bff-treated ewes relative to controls. The onset of oestrus was also significantly (P<0\m=.\01)delayed by bff treatment, but the ovulation rates did not differ between the groups. Furthermore, comparisons within or between groups revealed no significant relationships between any of the variables of plasma LH secretion during the follicular phase and the subsequent ovulation rate. These observations provide a complete description of gonadotrophin patterns during the follicular phase of the ewe and confirm the suggestion that an increase in LH pulse frequency is the major driving force behind the follicular growth that ultimately leads to ovulation. On the other hand, it appears most unlikely that the pattern of LH secretion during the follicular phase has any influence on ovulation rate. The levels of FSH declined in the period leading up to the preovulatory surge, presumably as a consequence of rising peripheral levels of oestrogen (and/or inhibin). We also expected LH pulse amplitude to decline during the follicular phase because it has been proposed that pulse amplitude is also controlled by oestrogen. The absence of any significant fall in amplitude suggests that hypotheses about the control of LH

2 secretion drawn from studies with ovariectomized ewes require further verification in the intact ewe. The effect of bff on prolactin levels probably reflects the low rates of secretion of oestradiol by the small ovarian follicles in these ewes. J. Endocr. (1988) 116, 123\p=n-\135 INTRODUCTION We have previously reported that, in ewes treated with bovine follicular fluid (bff) throughout the luteal phase, there was a rebound in plasma folliclestimulating hormone (FSH) concentrations after the cessation of treatment, a delay in the onset of oestrous behaviour and an increase in ovulation rate (Wallace & McNeilly, 1985; Wallace, McNeilly & Baird, 1985). Furthermore, no follicle development beyond 2-7 mm in diameter was observed on day 10 of the luteal phase in ewes treated with bff (Wallace, McNeilly & Tsonis, 1985), so it appears that the high ovulation rate of these animals is due entirely to follicular growth during the follicular phase. The present study was undertaken to determine the nature of the gonadotrophin signal to which these small androgenic follicles are exposed after the cessation of follicular fluid treatment at cloprostenol-induced luteolysis. The hypothesis tested was that differences in gonadotro phin secretion during the follicular phase may be responsible for the increase in ovulation rate normally observed in ewes treated with bff. The current model suggests that when the corpus luteum regresses the subsequent increase in luteinizing hormone (LH) pulse frequency promotes the sus tained increase in ostradiol production by the pre ovulatory follicle(s) required to induce oestrous behaviour and the LH surge (Baird & Scaramuzzi, 1976; Hauger, Karsch & Foster, 1977; Karsch, Foster, Legan et al. 1979; Baird, Swanston & McNeilly, 1981). The increase in pulsatile secretion of LH during the follicular phase is considered to be essential for the final stages of follicular development, but the role of FSH at this time is not clear (Baird & McNeilly, 1981). However, to date, there is a lack of data on the detailed changes in FSH and pulsatile LH secretion which normally occur in untreated ewes at this time. Therefore a further aim of the present study was to determine the changes in FSH secretion and pulsatile LH secretion after cloprostenol-induced luteolysis in normal cyclic ewes to establish a firm base for sub sequent studies on the induction of ovulation using exogenous hormone treatments. MATERIALS AND METHODS Experimental animals Fourteen Damline ewes (47% Finnish Landrace, 24% East Freisland, 17% Border Leicester, 12% Dorset Horn) were studied during the breeding season in November The ewes were 3-6 years old and weighed 59-1 ± 1-2 kg at the start of the experiment. The animals were at pasture throughout the period of synchronization of oestrus and housed indoors during the treatment cycle at Dryden Field Station, Roslin, Midlothian. The initial synchrony was achieved by with drawing progestagen-impregnated intravaginal pes saries (Intervet Laboratories Ltd, Cambridge) 12 days after their insertion. All ewes displayed behavioural oestrus within 48 h as determined by a raddled vasectomized ram. Luteolysis was then induced on day 10 of the second cycle with an i.m. injection of a potent analogue of prostaglandin F2a (100 pg cloprostenol; ICI, Macclesfield, Cheshire). Oestrus was detected as before and the ewes were moved inside before the start of the treatment cycle on day 1 (oestrus + 24 h). Experimental design Ewes were weighed, ranked and allocated to one of two treatments so that weights were equivalent between treatment groups before the start of the study. Bovine follicular fluid was collected and extracted with charcoal as described previously (Wallace & McNeilly, 1985). This removed 98% of the steroids leaving 2-13 ng oestradiol/ml and 5-2 ng progesterone/ml. Bovine serum was treated in an identical manner and used as a control fluid. The con centration of inhibin in the bff was estimated using an in-vitro ovine pituitary cell culture bioassay (Tsonis, McNeilly & Baird, 1986). The potency of the bff was 6840 units/ml (95% confidence limits, units/ml; index of precision of the bioassay, = 0-153). The seven ewes in the treatment group received 10 ml bff (i.v.) at and h on day 1 (oestrus + 24 h) to day 10 of the cycle inclusive. The seven remaining ewes received bovine serum and acted as controls. 11 of Luteal regression was induced at h on day the treatment cycle using cloprostenol. The subsequent onset of oestrus was detected with a vasectomized ram at 8-h intervals (09.00, and h) between 32 and 88 h after cloprostenol injection. Ovulation rate was determined by laparoscopy between days 5 and 7 of the subsequent cycle. Blood sampling Blood was sampled twice daily throughout the treat ment cycle immediately before bff or serum injec-

3 tion. On day 6 of the cycle the ewes were given an indwelling jugular cannula and, beginning 2 h later, samples were collected every 10 min between and h. The cannulae were removed after the sample at h on that day and ewes were cannulated again on the afternoon of day 10. On day 11 blood was sampled at 10-min intervals for 1-5 h before cloprostenol injection and thereafter for a maximum of 72 h. Blood samples were then withdrawn at 2-h intervals for up to 24 h to determine the onset of the preovulatory LH surge. In ewes which showed oestrous behaviour during the 72 h period, blood was sampled at 10-min intervals for 12 h after the first positive test for oestrus and thereafter at 2-h intervals for 24 h. During the subsequent oestrous cycle blood samples were taken daily at h until day 12. The ewes were maintained on an 8 h light: 16 h darkness photoperiod throughout the intensive sam pling period. Lights were on between and h and during the dark phase ewes were bled under red light. Lights were on during oestrus detection for 30-^45 min. All blood samples withdrawn during the luteal and follicular phases of the cycle were assayed for LH and prolactin. Samples taken twice daily during the luteal phase and those withdrawn at 30-min intervals on day 6 and throughout the follicular phase were assayed for FSH. Plasma progesterone concentrations were measured in daily samples taken during the treatment and subsequent cycle. In addition, samples taken at 1-h intervals for 10 h from cloprostenol-induced luteolysis were also assayed for progesterone. Hormone assays Prolactin, LH and FSH were measured in specific double-antibody radioimmunoassays as described previously (McNeilly & Andrews, 1974; McNeilly, McNeilly, Walton & Cunningham, 1976; McNeilly, Jonassen & Fraser, 1986). The sensitivities of the assays were 0-2 ng LH (NIH-LH-S18)/ml, 4ng FSH (NIH-FSH-S14)/ml and 0-4 ng prolactin (NIH-p- S13)/ml. The intra- and interassay variations as coef ficients of variation were 4-0 and 11-9%, 4-2 and 8-3% and 51 and 8-9% for LH, FSH and prolactin respect ively. Progesterone concentrations were measured in 250 pi aliquots of plasma by radioimmunoassay as described by Djahanbahkch, Swanston, Corrie & McNeilly (1981). The intra-assay coefficient of vari ation was 14-8%, the detection limit was 30 pmol per tube and the recovery of progesterone from plasma was % (S.E.M.; «480). = Data analysis Onset of oestrus was considered to be the time minus 4 h when a ewe first stood to allow a vasectomized ram to mount her because of the 8-h intervals between detection of oestrous behaviour. The onset of the preovulatory LH surge was considered to be the time when levels exceeded 20 pg/1 for the first time. An LH rise was considered to be a pulse if the value of two consecutive samples was greater than the mean of the two previous samples (basal value) and the value of at least one of the peak samples exceeded the mean basal value by more than four times the intra-assay coef ficient of variation of the assay (based on Backström, McNeilly, Leask & Baird, 1982). The Mann-Whitney U test was used to examine the effect of treatment on ovulation rate and the number of large follicles at laparoscopy. Student's?-test was used to determine the effect of treatment on onset of oestrus and the characteristics of the preovulatory LH surge. The differences in gonadotrophin secretion between the two groups were analysed by analysis of variance. One way analysis of variance with repeated measures was used to analyse the changes in the characteristics of pulsatile LH secretion within groups during the follicular phase in conjunction with Duncan's multiple range test where appropriate. Linear regression analysis was applied to changes in FSH and progesterone secretion as indicated in the text to determine whether the slopes were significantly different (1) from time zero or (2) between treatment groups. Split-plot analysis of variance was used to assess the differences between groups and the change in progesterone concentrations with time. The prolactin data were initially condensed by taking means of the 10-min samples over consecutive 2-h periods for each ewe and these means were then used as the raw data for testing for treatment effects. We transformed the data logarithmically, because the variance and means were proportional, then used a split-plot analysis of variance with bff as main treat ment and sample time as subtreatment. We also tested for a diurnal rhythm by autocorrelation, in which values for the first hypothesized cycle were correlated with those for the second and subsequent cycles (if there were any), using data from individual ewes. There was no effect of bff treatment on this pattern between the groups so the data from all 14 ewes were grouped and only two cycles were tested. RESULTS Oestrus and ovulation rate Twice daily injections of bff for 10 days of the luteal phase before cloprostenol-induced luteolysis resulted in a significant ( <0 1) delay in onset of oestrus compared with that of control animals (Table 1). Ovu lation rate was not significantly different between

4 TABLK 1. Effect of treatment of ewes with bovine follicular fluid on pulsatile secretion of LH on day 6 of the luteal phase and on time to onset of oestrus from cloprostenol-induced luteal regression, characteristics of the preovulatory LH surge and ovulation rate. Values are means + s.e.m. (n = 1) Control ewes Treated ewes Day 6: luteal phase Basal LH (µ /1) ** LHpulses/6h * LH pulse amplitude/6 h * (Mg/1) Mean LH level/6 h * (µß/1) Onset of oestrus Time from cloprostenol (h) ** LH surge Onset-time from *** cloprostenol (h) Oestrus to surge (h) * Maximum peak height ± (µß/1) Ovulation rate (range) (2-5) (3-5) No. of large follicles * (>5mmdiam) (range) (1-4) (2-7) */><0-05, **/><001,***/><0001 compared with control (Student's/-test). groups but there was significantly (P<0-05) more large follicles ( > 5 mm diamter) as assessed at laparoscopy in the group treated with bff. The delay in onset of oestrus was not correlated with subsequent ovulation rate or the number of large follicles. Hormone concentrations during treatment Follicle-stimulating hormone The changes in plasma levels of FSH measured twice daily throughout the luteal phase were consistant with those described previously (Wallace & McNeilly, 1985, 1986). In control ewes receiving bovine serum, FSH levels were initially high at h on day 1 and then fell during day 2 to reach a nadir on days 3 and 4. Thereafter levels began to rise to reach a peak on days 5 and 6 after which they decreased slightly to a plateau on days 7 to 11. The first injection of bff at h on day 1 suppressed the initial plasma FSH levels by approximately 60% by the h blood sample. Throughout the luteal phase there was a twice daily rebound in FSH levels in most treated ewes as described previously (Wallace & McNeilly, 1986) with concentrations being higher at the h sample (16 h from injection) than at the h sample (8 h from injection). During the frequent sampling period on day 6 of the luteal phase the mean FSH level measured in blood sampled at 30-min intervals was significantly higher in control ewes than in ewes treated with bff ( (s.e.m.) versus pg/1; <0-02). Luteinizing hormone Twice daily LH concentrations were marginally higher throughout the luteal phase in ewes treated with bff than in control ewes. Analysis of the shortterm changes in LH secretion on day 6 of the cycle confirm previous observations (Wallace & McNeilly, 1985, 1986) that basal LH concentrations, pulse fre quency and pulse amplitude were all significantly (P<0-05) higher in the treated group than in controls (Table 1). Prolactin Analysis of the samples taken at 10-min intervals on day 6 of the luteal phase showed that prolactin levels were low and did not fluctuate during the 6h. The mean prolactin concentrations were and pg/1 in control and treated ewes respectively. Progesterone The changes in plasma progesterone concentrations measured daily throughout the luteal phase are illus trated in Fig. la. Progesterone concentrations were significantly (P< 0005) higher in treated ewes than in controls between days 3 and 11 of the luteal phase. Similarly, linear regression analysis of the individual slopes snowed that the rate of increase in progesterone concentrations was significantly (P<005) higher in ewes treated with bff (slope , r = versus , r= in controls). Hormone concentrations after luteal regression Progesterone The decline in plasma progesterone concentrations after cloprostenol-induced luteolysis are illustrated in Fig. lb. Progesterone levels decreased after clopros tenol injection in both groups and were significantly (P<0-05) lower relative to time zero by 1 h in the con trol group and by 2 h in the treated group. Lineal regression analysis of the individual slopes showed that progesterone levels fell significantly (P<0-05) with time in all ewes and that the overall rate of decline during the first 10 h after cloprostenol was greater (P<002) in bff-treated ewes than in controls (slope , r 0-652±0-061 = versus ±0-060, r ±0-052 respectively). = Progesterone levels were at or below basal concen trations (6 nmol/1) by 24 h after cloprostenol injection in both groups.

5 as a Day of luteal phase 0 I Time relative to cloprostenol (h) figure 1.(a) Changes in the plasma concentrations of progesterone in blood samples taken before the morning injection of bovine serum ( O ) or bovine follicular fluid ( ) during the luteal phase of the cycle, (b) Changes in the plasma concentrations of progesterone in blood samples withdrawn at 1-h intervals after the injection of cloprostenol (arrow; 100 pg i.m.) on day 11 of the cycle in control ewes (O) and ewes injected with bovine follicular fluid ( ) throughout the luteal phase. Values are means + s.e.m.; = 7 ewes per group. Follicle-stimulating hormone The changes in plasma FSH concentrations after cloprostenol-induced luteal regression are illustrated in Fig. 2a. Concentrations of FSH in ewes treated with bff increased after luteolysis to reach a maxi mum of pg/1 (range pg/1) between 8 and 14h after cloprostenol injection ( h), i.e h after the last injection of bff. This represents a three- to fourfold increase over control group levels which decreased after cloprostenol injec tion to reach a nadir 16 to 18 h later. Linear regression 11 analysis of the individual slopes of the control group ewes showed that there was a significantly negative regression with time between zero to +16 h after injection of cloprostenol in all seven ewes (range, r to 0-919; /><0001). This represents = an over all decrease in FSH concentrations of approximately 56%. Plasma FSH concentrations in treated ewes were significantly (P<0001) greater than in controls throughout the period 0 to +24 h after cloprostenol injection, i.e. 16^40 h after the last injection of bff (control ewes: 32-2 ±3-2 pg/1; bff-treated ewes: 95-4±4-1 pg/1). Thereafter, during the period from 24 to 48 h after injection of cloprostenol, there were no significant differences between treatments. Similarly there were no significant differences between treat ment groups in the FSH levels during the 24 h before the preovulatory FSH surge, nor in the magnitude of the surge (Fig. 2b) which coincided with the preovula tory LH surge in both groups. Although the sampling period did not encompass the secondary FSH surge in all ewes, in both groups the data indicate that there was no significant difference between treatment groups in the ascending limb of the second FSH surge and further suggest that the magnitude of the second ary FSH surge was considerably greater than that of the preovulatory surge in both groups. There was no relationship between FSH concen trations during the follicular phase and ovulation rate within either treatment group or overall for all ani mals irrespective of treatment group. However, there was a significant positive correlation (r = 0-534; < 0-05; = 1) between the mean FSH concentrations during the first 24 h after cloprostenol injection and the number of large follicles observed at laparoscopy 5-6 days after ovulation within the bff-treated group. No such relationship was evident within the control group. Pulsatile LH secretion: data analysis The pattern of LH secretion on day 6 of the luteal phase and during the preovulatory follicular phase of the cycle, after cloprostenol-induced luteal regression, is illustrated for one control (Fig. 3a) and one bfftreated ewe (Fig. 3b). For analysis, the LH pulse frequency data has been presented as changes in interpulse interval which was calculated as the time between successive LH pulses (Table 2). For the pur poses of statistical analysis within treatment groups the number of pulses and mean pulse amplitude were calculated for each 5-h block after cloprostenolinduced luteolysis. Due to the spread in time of onset of behavioural oestrus and hence the LH surge within both groups it was only possible to compare the changes in pulsatile LH secretion until 40 h and 65 h

6 (a) ZI t/1 U (b) " ft Time relative to cloprostenol (h) 40 ~ 'S i t/5 " Time relative to LH surge (h) ~i 24 figure 2. Changes in the concentrations of FSH after the injection of cloprostenol (arrow) on day 11 of the oestrous cycle in control ewes (O) and ewes injected twice daily with bovine follicular fluid ( ) throughout the luteal phase. The results have been grouped around the time of (a) clopros tenol injection, samples withdrawn at 30-min intervals, or (b) the LH peak, samples withdrawn at 2-h intervals. Values are means + s.e.m.; = 1 ewes per group except where indicated otherwise. after cloprostenol injection in control and bff-treated ewes respectively. Similarly the 10-min sampling sched ule did not extend to the onset of the preovulatory LH surge in all ewes in either treatment group. Thus the data could not be analysed relative to the LH surge. One ewe in the control group (1J482) was excluded from the pulsatile LH secretion analysis. Although regular oscillations above basal concentrations were observed in this ewe throughout the follicular phase the amplitude of many of these 'pulses' was too low to determine whether or not they represented significant pulses using the criteria defined previously.

7 ^_ w 20-, Day 6 luteal Follicular 16 phase phase wmmmm f 12 * î îî î*;.* *. ï Time : 0J -.-1-r h : LH surge g X Time : h : 20 -, lft (b) Time relative to cloprostenol (h) Day 6 luteal Follicular phase phase Z_ "S 12 d OkM VJ\J 0 Time: h : LH surge 20 lft -12 _> *** I î **** Time: h : Time relative to cloprostenol (h) figure 3. Changes in the pulsatile secretion of LH during day 6 of the luteal phase and then through out the follicular phase in (a) a representative control ewe (OJ312) and (b) a representative ewe treated with bovine follicular fluid (bff) (OJ404). Blood samples were withdrawn at 10-min inter vals from 2 h before to + 56 h after injection of cloprostenol (arrow) on day 11 of the cycle. Each asterisk represents a significant LH pulse. The shaded horizontal bars represent periods of darkness.

8 table 2. Effect of treatment of ewes with bovine follicular fluid (bff) throughout the previous luteal phase on the changes in basal plasma levels of LH, and the amplitude and frequency (as inter-pulse interval) of the pulsatile secretion of LH after induction of luteal regression by injection of cloprostenol at h on day 11. The last injection of bff was given at h on day 10. Values are means ± s.e.m., n = l ewes per group Basal LH fog/l) Control bff LH pulse amplitude (µ 1) Control bff Inter-pulse interval (min) Control bff Time after cloprostenol (h) Changes in pulse frequency Control ewes The inter-pulse interval, initially 118±6 min in con trol ewes, decreased rapidly to reach approximately 1 pulse/60 min within 5 h after cloprostenol injection in all ewes (Table 2). Thereafter a more gradual decrease in inter-pulse interval with time was apparent until a minimum of approximately 50 min was reached by 40 h after cloprostenol injection. Correspondingly this decrease in pulse interval during the follicular phase was significant (P<0001) when analysed in 5-h blocks from zero to 40 h after cloprostenol (Table 2). In the profiles of the three ewes in which the 10-min sampling extended to the onset of the preovulatory LH surge there was no further detectable change in pulse interval (Fig. 3a). The mean inter-pulse interval on day 6 of the luteal phase ( min) was signifi cantly (_ <0 1) lower than that observed during the follicular phase from 8 h after cloprostenol injection (40-70 min; Table 2). Ewes treated with bff In ewes treated with bff the inter-pulse interval was also initially high ( min) and decreased rapidly to reach a plateau inter-pulse interval of 60 min identical to that of controls within 5 h after clo prostenol injection (Table 2). The inter-pulse interval remained approximately constant for a longer time than in controls before declining in a similar manner to reach a minimum inter-pulse interval of approxi mately 50 min. Pulse interval showed a significant (P< 0-001) decrease with time during the follicular phase when analysed in 5-h blocks between zero and 55 h after injection of cloprostenol. A complete folli cular phase profile, in which the frequent sampling encompassed the onset of the preovulatory LH surge, was obtained for two ewes treated with bff and showed that there was no further detectable change in LH pulse interval. The mean pulse interval during the luteal phase in ewes treated with bff (96 min) was also distinctly lower than that observed during the fol licular phase (60 min). However, the magnitude of the increase in pulse frequency induced by luteolysis was considerably less in treated ewes than in controls. Changes in pulse amplitude Immediately after cloprostenol injection there was a significant increase in LH pulse amplitude with time for the first five pulses measured in samples from control ewes and ewes treated with follicular fluid (r = 0-917; < 001 and r=0-665; < 0-05 respectively, see Fig. 3 and Table 2). There were no significant differ ences between treatment groups in the rate of increase in amplitude of these pulses as determined by analy sis of the slopes of the individual regression lines (B = 0-219±0066 for control ewes and = 0168±0050 for ewes treated with bff). Thereafter pulse ampli tude in the control group decreased rapidly and did not change significantly with time throughout the fol licular phase (Fig. 3a and Table 2). Overall, the mean pulse amplitude in the control group during the folli cular phase was not significantly different from that observed during day 6 of the luteal phase (2-0 ±0-2 versus 2-0 ±0-3 pg/1 respectively, see Table 2). In contrast, pulse amplitude in ewes treated with bff remained high for up to 15 h after cloprostenol

9 injection (Fig. 3b and Table 2). Subsequently, between 15 and 65 h after cloprostenol, there was no signifi cant change in the mean pulse amplitude which was similar to that of controls. In addition, the pulse amplitude in ewes treated with bff was significantly (P<005) higher on day 6 of the luteal phase than between 16 h and 64 h after cloprostenol injection (3-0 ±0-3 versus 1-9 ±0-2 pg/1). Changes in basal LH levels Basal LH levels did not change significantly through out the follicular phase in control ewes (21 ±0-1 pg/1; Table 2). In contrast, basal LH levels were initially high (3-4±0-3 pg/1) between zero and +15 h in the bff-treated group, coincident with the high pulse amplitude observed at this time (Table 2). Thereafter, basal LH levels (2-0 ±0-1 pg/1) were equivalent to those of controls and did not change significantly with time. None of the parameters of pulsatile LH secretion or basal levels of LH measured during the luteal and/or follicular phase of the cycle were correlated with the subsequent ovulation rate either within or between treatment groups. Neither LH pulse frequency nor amplitude appeared to vary with photoperiod during the follicular phase. Prolactin secretion The 2-hourly mean concentrations of prolactin in plasma during the follicular phase are shown for the two experimental groups in Fig. 4. Analysis of vari ance revealed significant major effects of sampling time ( < ) and bff ( < 0-05), and also a signifi cant (P<0-05) interaction between these two factors. The interaction was significant because the concen trations were consistently higher in the control ewes than in the bff-treated ewes except for short periods at the beginning and end of the observation period. The first of these was associated with a peak in prolactin induced in both groups by the injection of cloprostenol. The second was near the time of the preovulatory surge of LH in the treated ewes, but well after the preovulatory LH surge in the control ewes, most of which were observed about h after injection of cloprostenol. Despite the disruptive effects of cloprostenol and the preovulatory rises, a diurnal rhythm was evident in the data, with a nadir between and 15.00h, and a peak between and h. This was sup ported by the autocorrelation analysis which showed that the concentrations in the first 24-h period (beginning 4 h after injection of cloprostenol and ter minating before the first preovulatory rises) were sig nificantly (r 0-67; P<00001) correlated with those = in the following 24-h period when the data from all ewes were combined (data not shown). Interestingly, the preovulatory peaks in the control ewes coincided with the night-time peak of the second day while those in the treated ewes coincided with the night-time peak of the third day. Preovulatory LH surge There were no significant differences between control ewes and ewes treated with bff in basal LH concen trations measured in 2-h samples collected 24 h before figure 4. Mean plasma prolactin concentrations (logarithmically transformed) after cloprostenol-induced luteolysis in ewes treated with bovine serum (open bars, = 7) or bovine follicular fluid (bff, solid bars, = 1). The shaded horizontal bars indicate periods of darkness. The least significant differences were 1-91 for effect of bff and 0-29for effect of time.

10 the preovulatory LH surge (Table 1). The onset of the preovulatory LH surge was significantly (P< 0-001) later relative to cloprostenol injection in the treatment than in the control group. Similarly the interval from oestrus to the LH surge was significantly (P<005) longer in treated ewes than in control ewes. Subsequent oestrous cycle There was no significant effect of bff treatment on plasma concentrations of FSH or LH during the sub sequent cycle. Similarly, plasma progesterone levels were characteristic of a normal cycle in both groups. DISCUSSION Changes in pulsatile LH secretion: control ewes The results of the present study confirm but con siderably extend previous observations that the pulse frequency is significantly higher in the follicular phase than in the luteal phase of the oestrous cycle (Baird, Swanston & Scaramuzzi, 1976; Baird, 1978; Martensz & Scaramuzzi, 1979; Baird et al. 1981). The LH pulse frequency increased rapidly within 5 h of cloprostenolinduced luteolysis, most probably as a result of falling progesterone levels. This agrees with the trend observed in a previous study in which blood was sampled at 1-h intervals after cloprostenol (Baird et al ). Karsch, Foster, Bittman & Goodman (1983) studied ovariectomized steroid-implanted ewes and found that the acute withdrawal of progesterone alone is insufficient to account entirely for the high frequency of LH pulses seen during the follicular phase and have suggested that oestradiol also stimu lates the increase in pulse frequency. It is unlikely that the rapid increase in LH pulse frequency observed over the first 6 h in the present study was due to a direct stimulatory effect of oestradiol as it is known that the secretion of oestradiol does not begin to increase until 6-12 h after cloprostenol injection (Baird et al. 1981). However, it is possible that the increased secretion of oestradiol from the preovula tory follicle(s) is responsible for the subsequent, more gradual increase in pulse frequency. The transient increase in LH pulse amplitude after cloprostenol injection and before oestradiol secretion has increased suggests that progesterone can regu late pulse amplitude, perhaps through an effect on pituitary responsiveness to gonadotrophin-releasing hormone (GnRH) pulse amplitude. Low levels of progesterone (1 ng/ml) have been shown to reduce pituitary responses to GnRH in anoestrous ewes (Wheaton & Mullet, 1982). Alternatively, the initial increase in pulse amplitude may be due to the self-priming effect of GnRH described by Crighton & Foster (1977) who found that, in ewes given two injections of GnRH 1-5 h apart, the response to the second injection was greater than the response to the first. In the present study the increase in pulse amplitude after cloprostenol injec tion may have been due to this self-priming effect, or to the withdrawal of progesterone, or to both. Further work is required to clarify this point. Several studies have shown that the amplitude of pulses during the follicular phase were lower than during the luteal phase (Baird, 1978; Martensz & Scaramuzzi, 1979) and this has been attributed to a direct effect of increasing oestradiol levels from the preovulatory follicle(s) on pituitary responsiveness to GnRH (Goodman & Karsch, 1980). However, more recent studies have shown that pulse amplitudes are similar throughout the cycle (Baird et al ; Karsch et al. 1983) and the data from the present study sup port this view. Furthermore, there was no detectable change in pulse amplitude with time throughout the follicular phase in spite of the high pulse frequency and presumably rising oestradiol levels. The present study shows convincingly that neither the mean nor the basal LH levels change throughout the follicular phase after the initial rapid increase in pulse frequency at luteolysis. This suggests that while the follicle popu lation is stimulated with a higher frequency of LH pulses than that observed during the luteal phase, a gradual increment in pulse frequency and a rise in basal LH levels throughout the entire follicular phase is not required, and does not occur, before the abrupt onset of the preovulatory LH surge. Changes in pulsatile LH secretion: bff-treated ewes From previous studies (Wallace & McNeilly, 1985, 1986) we concluded that the high LH pulse frequency observed during the luteal phase in ewes treated with bff occurred because of the suppression of plasma FSH concentrations resulting in a reduction of nega tive oestradiol feedback. Furthermore, we suggested that oestradiol was a major hormone exerting nega tive feedback on LH pulse frequency as the high fre quencies were observed in the presence of normal or high progesterone levels (Wallace & McNeilly, 1985, 1986). Similarly, in the present study LH pulse fre quency was significantly higher in treated ewes than in controls during the luteal phase. Nevertheless, as was observed in the control ewes, pulse frequency increased further within 6 h after cloprostenol injec tion in ewes treated with bff even though the amount of oestradiol being secreted from the follicles was pre sumably still very low. This supports the suggestion that the increase in pulse frequency during luteolysis is primarily due to the fall in progesterone levels. Following the initial rapid increase in pulse

11 frequency after cloprostenol injection there was an extended period, not observed in the controls, during which pulse frequency did not change. This delay before the gradual increase in pulse frequency began was of similar duration to the delay in the onset of oestrus in treated ewes and suggests that insufficient oestradiol production by the follicles was temporarily preventing the final increase in pulse frequency. The LH pulse amplitude was also significantly higher in treated ewes compared with controls during the luteal phase in the present study confirming our previous observations (Wallace & McNeilly, 1986). However, pulse amplitude increased even further after cloprostenol injection in bff-treated ewes. As oestra diol secretion by the follicles is known to be low at this time in bff-treated ewes it seems most probable that this initial increase in pulse amplitude is due to a selfpriming effect by GnRH. As the follicles grow and respond to the high FSH concentrations and increased LH secretion, oestradiol production presumably rises until it reaches concentrations similar to those in con trol ewes. This, in turn, results in a gradual fall in pulse amplitude. Thereafter, pulse amplitude did not vary throughout the follicular phase and was similar to that of control ewes. It appears, therefore, that a threshold level of oestradiol secretion is required to maintain normal LH pulse amplitude and that an increase in oestradiol production beyond this threshold, as occurs during the follicular phase, cannot suppress pulse amplitude further. Changes in FSH secretion In contrast to LH, the secretion of FSH in control ewes decreased with time during the first 18 h after cloprostenol-induced luteolysis, in agreement with previous observations of a general decline in FSH levels during the follicular phase (Salamonsen, Jones, Burger et al. 1973; Baird et al. 1981; McNeilly, 1984). The suppression of FSH at this time is probably due to the secretion of oestradiol (Baird et al. 1981) and, possibly, inhibin, as the concentrations of both of these hormones increase in parallel within the preovu latory follicle(s) as they develop (Tsonis, Quigg, Lee et al. 1983). The present study has confirmed previous obser vations (Wallace & McNeilly, 1985, 1986) that treat ment with bff resulted in a significant rebound in FSH levels after the cessation of treatment at clopros tenol-induced luteolysis. This rebound may reflect the release of pituitary FSH stores accumulated in the gonadotrophs during treatment or it may equally be a reflection of low endogenous inhibin and steroid pro duction due to insufficient FSH stimulation of the fol licle population during bff treatment. When the bff treatment ends there is a hypersécrétion of FSH as the pituitary releases its stores and follicle development progresses. As the follicles grow and become oestrogenic in response to LH stimulation they secrete suf ficient oestradiol and inhibin into the circulation and the rebound ceases. This is confirmed by the obser vation that the delay in the onset of oestrus which was most probably due to insufficient oestradiol produc tion by the small follicles was similar to the duration of the hypersécrétion of FSH. Gonadotrophin secretion and ovulation rate In previous experiments the hypersécrétion of FSH after the cessation of bff treatment was considered to be the most probable cause of the increase in ovula tion rate. In the present study follicular fluid treat ment did not significantly increase ovulation rate but the mean ovulation rate of control ewes (3-3) was con siderably greater than that normally recorded for Damline ewes ( ). This highlights one of the problems of working with a prolific breed and suggests that the animals used in the present study were already expressing their maximum ovulation rate. However, the number of large follicles ( > 5 mm diameter) observed at laparoscopy 5-6 days after ovulation in the bff-treated group was significantly higher than that of controls and was correlated with the magnitude of the FSH rebound during the initial 24 h of the follicular phase. It is tempting to suggest that the hypersécrétion of FSH at this time may have resulted in either an increase in the number of antral follicles in the population or a decrease in the incidence of atresia within existing follicles. In spite of the failure of bff treatment during the luteal phase to increase ovulation rate in the present experiment, the changes in LH secretion observed were considered to be characteristic of those normally observed after bff treatment. It is unlikely that the follicle population was sufficiently developed to benefit from the high pulse amplitude or frequency observed during the luteal phase or indeed from the high pulse amplitude during the early follicular phase. As the subsequent pulsatile LH secretion during the follicular phase was similar in both groups it seems unlikely that quantitative differences at this stage of the cycle could be associated with the change in ovulation rate observed in previous studies. It is perhaps not surprising, however, that in this and other studies it has proved largely impossible to associate quantitative differences in gonadotrophin secretion with changes in ovulation rate. The sensi tivity of the hypothalamic-pituitary axis to the feed back effects of steroids and inhibin may vary widely between individual ewes at different times of the cycle. Similarly the amount of feedback, especially in a pro lific breed, may vary with ovulation rate but not

12 necessarily in a proportional manner. These problems were highlighted by one control ewe in the present study. During both the luteal and follicular phases of the cycle, pulsatile LH secretion was largely undetectable. This ewe subsequently had four normal corpora lutea. This suggests that the large quantities of oestradiol being produced by the follicles resulted in a large amount of negative feedback and hence low LH secretion. The concentrations of prolactin in plasma were affected by a diurnal rhythm and by bff treatment. The effect of prior treatment with bff were probably mediated through the effects of this treatment on folli cular development. In the treated ewes, the follicles are small and produce little oestrogen, and oestrogen has been shown to stimulate prolactin secretion in the ewe (Fraser, Clarke & McNeilly, 1981). The low levels of prolactin in bff-treated ewes are therefore probably a consequence of low rates of oestrogen secretion, at least during the initial 2 days of the folli cular phase. The diurnal rhythm is strikingly similar to that recently reported for cyclic (Wallace & McNeilly 1986) and anoestrous ewes (Martin, Cognie, Schirar et al. 1985) with a nadir around noon and a peak before midnight. Interestingly, the lack of oestrogen did not inhibit the expression of the diurnal rhythm which, in fact, was more evident in bfftreated ewes than in the controls (Fig. 4). Thus, while the origin and function of this rhythm remain obscure, it appears that it is independent of other rhythms in the hypothalamic-ovarian axis. ACKNOWLEDGEMENTS We thank Dr C. G. Tsonis, Miss N. Anderson, Mrs H. Ebling and Mr B. Ogilvie for help during the experiment, Miss M. Fordyce and the staffai ABRO Dryden Field Station for performing the laparoscopies, Dr G. C. Tsonis for determining the inhibin potency, Mr T. McFetters and Mr E. Pinner for prep aration of the figures and Miss A. Wallace for typing. J. M. W. was in receipt of a postgraduate studentship from the Department of Agriculture and Fisheries for Scotland. REFERENCES Backström, C. T., McNeilly, A. S., Leask, R. M. & Baird, D. T. (1982). Pulsatile secretion of LH, FSH, prolactin, oestradiol and progesterone during the menstrual cycle. Clinical Endocrinology 17, Baird, D. T. ( 1978). Pulsatile secretion of LH and ovarian oestradiol in the follicular phase of the sheep oestrous cycle. Biology of Reproduction 18, Baird, D. T. & McNeilly, A. S. (1981). Gonadotrophic control of follicular development and function during the oestrous cycle of the ewe. Journal ofreproduction and FertilitySuppl. 30, Baird, D. T. & Scaramuzzi, R. J. (1976). Changes in the secretion of ovarian steroids and pituitary luteinizing hormone in the periovulatory period in the ewe: the effect ofprogesterone. Journal of Endocrinology 70, Baird, D. T., Swanston, I. A. & McNeilly, A. S. (1981). Relation ship between LH, FSH and prolactin concentration and the secretion of androgens and estrogens by the pre-ovulatory follicle in the ewe. Biology ofreproduction24, Baird, D. T., Swanston, I. A. & Scaramuzzi, R. J. (1976). Pulsatile release of LH and secretion of ovarian steroids in sheep during the luteal phase of the estrous cycle. Endocrinology 98, Crighton, D. B. & Foster, J. P. (1977). Luteinizing hormone release after two injections of synthetic luteinizing hormone releasing hormone in the ewe. Journal ofendocrinology 72, Djahanbahkch, O., Swanston, I.A., Corrie, J. E. T. & McNeilly, A. S. (1981). Prediction of ovulation by progesterone. Lancet û, Fraser, H. M., Clarke, I. J. & McNeilly, A. S. (1981). Effect of oestrogen and an LHRH agonist on the release of gonadotro phins in ovariectomized ewes deprived of LHRH. Journal of Reproduction and Fertility 62, Goodman, R. L. & Karsch, F. J. (1980). Pulsatile secretion of luteinizing hormone: Differential suppression by ovarian steroids. Endocrinology 107, Hauger, R. L., Karsch, F. J. & Foster, D. L. (1977). A new concept for control of the estrous cycle of the ewe based on the temporal relationships between luteinizing hormone, estradiol and pro gesterone in peripheral serum and evidence that progesterone inhibits tonic LH secretion. Endocrinology 101, Karsch, F. J., Foster, D. L., Bittman, E. L. & Goodman, R. L. (1983). A role for estradiol in enhancing luteinizing hormone pulse frequency during the follicular phase of the estrous cycle of sheep. Endocrinology 113, Karsch, F. J., Foster, D. L., Legan, S. J., Ryan, K. D. & Peter, G. K. (1979). Control of the preovulatory endocrine events in the ewe: interrelationship of estradiol, progesterone and luteinizing hormone. Endocrinology 105, McNeilly, A. S. ( 1984). Changes in FSH and the pulsatile secretion of LH during the delay in oestrus induced by treatment of ewes with bovine follicular fluid. JournalofReproduction andfertility 72, McNeilly, A. S. & Andrews, P. (1974). Purification and charac terization of caprine prolactin. Journal ofendocrinology 60, McNeilly, A. S., Jonassen, J. A. & Fraser, H. M. (1986). Sup pression of follicular development after chronic LHRH immunoneutralization in the ewe. Journal ofreproduction and Fertility 76, McNeilly, J. R McNeilly, A. S., Walton, J. S. & Cunningham, F. J. (1976). Development and application of a heterologous radio immunoassay for ovine follicle-stimulating hormone. Journal of Endocrinology 70, Martensz, N. D. & Scaramuzzi, R. J. (1979). The plasma concen trations of luteinizing hormone, follicle-stimulating hormone and progesterone during the breeding season in ewes immunized against androstenedione or testosterone. Journal ofendocrin ology 81, Martin, G. B., Cognie, Y., Schirar,., Nunes-Ribeiro,., Fabre- Nys, C. & Thiery, J. C. (1985). Diurnal variation in the response of anoestrous ewes to the ram effect. Journal ofreproduction and Fertility 65, Salamonsen, L.., Jones, J.., Burger, H. G.Buckmaster, J. M., Chamley, W.., Cummins,.., Findlay, J. K. & Goding, J. R. (1973). A heterologous radioimmunoassay for folliclestimulating hormone: application to measurement of FSH in the ovine estrous cycle and in several other species including man. Endocrinology 93,