Influence of melatonin and oestradiol on the opioidergic regulation of LH and prolactin release in pony mares

Influence of melatonin and oestradiol on the opioidergic regulation of LH and prolactin release in pony mares C Aurich, J Lange, H-O Hoppen and J E Aurich 'Institut für Reproduktionsmedizin, 2Klinik für Geburtshilfe und Gynäkologie des Rindes and 3 Abteilung für Endokrinologie, Tierärztliche Hochschule, Bünteweg 15, D-30559 Hannover, Germany (Requests for offprints should be addressed to C Aurich, Institut für Reproduktionsmedizin, Tierärztliche Hochschule, Bünteweg 15, D-30559 Hannover, Germany} Abstract The aim of this study was to investigate the influence of oestradiol, melatonin and season on the opioid regulation oflh and prolactin release. Effects of the opioid antagonist naloxone (0\m=.\5mg/kg) on LH and prolactin secretion were determined in ovariectomized pony mares. In experiment 1, mares in January (n=6) were pretreated with oestradiol benzoate (5 \g=m\g/kg)for 20 days. In experiment 2, beginning in May, mares (n=7) received melatonin (15 mg) for 15 days and subsequently a combination of melatonin plus oestradiol for 20 days. In experiment 3, beginning in May, mares (n=6) were pretreated with oestradiol for 30 days, left untreated for 12 days and then given melatonin for 35 days. In all experiments the animals were injected with the opioid antagonist naloxone and saline on 2 consecutive days prior to treatment. In experiment 1, animals received naloxone and saline on days 10 and 11 and 20 and 21 following oestradiol treatment. In experiment 2, naloxone and saline were administered on days 15 and 16 following melatonin treatment and on days 10 and 11 and 20 and 21 of melatonin plus oestradiol treatment. In experiment 3, the animals received naloxone and saline on days 10 and 11, 20 and 21 and 30 and 31 of oestradiol treatment, prior to melatonin treatment and on days 15 and 16, 25 and 26 and 35 and 36 following melatonin. In January (experiment 1), naloxone evoked a significant (P<0\m=.\05)LH release at all times, however the LH increment in response to naloxone increased during oestradiol pretreatment (P<0\m=.\05)During the breeding season (experiments 2 and 3), naloxone induced a significant (P<0\m=.\05)increase in plasma pretreated LH concentrations when mares had not been with oestradiol or melatonin and after oestradiol pretreatment. Basal LH concentrations and the LH increment in response to naloxone increased significantly (P<0\m=.\05)during the 30-day oestradiol pretreatment. Melatonin decreased the naloxone-induced LH release and the LH release in response to naloxone and saline no longer differed after 25 and 35 days of melatonin pretreatment. When melatonin was given together with oestradiol for 20 days, again a significant (P<0\m=.\05)LH release in response to naloxone occurred. Prolactin release was significantly (P<0\m=.\05)increased by naloxone when mares had been pretreated with only melatonin. The opioid antagonist did not affect prolactin release in mares that had not been pretreated or received oestradiol either alone or in combination with melatonin. In conclusion, in long\x=req-\ term ovariectomized mares, opioids inhibit LH secretion inhibition of independent from ovarian factors. This opioid LH secretion is enhanced by oestradiol and reduced by melatonin. Although short-term melatonin treatment in\x=req-\ activates the opioid regulation of LH release, a prolonged influence of melatonin as occurs in winter does not prevent activation of the opioid system. This indicates that effects of melatonin on the opioid regulation of LH release change with time. An opioid inhibition of prolactin secretion is activated by melatonin given for 15\p=n-\35days but is lost under the prolonged melatonin signal in winter. Journal of Endocrinology (1997) 154, 241\p=n-\248 influence of a short-day Introduction Reproductive activity in the horse changes throughout the year and the physiological breeding season lasts from approximately March to October. The reduction in ovar ian function during the winter months is primarily caused by a decrease in gonadotrophic stimulation. In female and male horses, endogenous opioids participate in the regu lation of luteinizing hormone (LH) release as inhibitory neuromodulators (Behrens et al. 1993, Aurich et al. 1994!, > 1995). In mares, the opioid antagonist naloxone causes an increase in plasma LH concentrations during the luteal but not the follicular phase of the cycle (Behrens et al. 1993) and during the anovulatory season in winter (Aurich et al. 1994a). It has therefore been suggested that opioidergic mechanisms inhibit LH release during seasonal anoestrus and that transition into the breeding season coincides with a switch from a continuous opioidergic

inhibition to an intermittent, cycle-dependent blockade of LH release (Aurich et al. 1994a). In cyclic mares, the opioidergic inhibition of LH secretion is most likely activated by a sequence of oestrogen and progesteronedominated steroid environments, characteristic for the oestrous cycle but, in ovariectomized mares, could also be stimulated by oestradiol alone (Aurich et al. 1995). In contrast to the breeding season, the opioidergic systems regulating LH release in seasonally anovulatory mares do not depend on progesterone. In these animals, the opioid ergic inhibition of LH secretion might be activated by continuous exposure to low oestrogen concentrations or independent from ovarian factors (Aurich et al. 1994 ). Besides regulating LH release, opioids inhibit prolactin secretion in stallions (Aurich et al. 1994fr) and in ovariec tomized, steroid-treated mares (Aurich et al. 1995). Plasma prolactin concentrations in the horse are low in winter and increase during the breeding season (Worthy et at. 1987). Circannual changes in prolactin release, as in other species, are largely regulated by photoperiod (Johnson 1987). Photoperiodic stimuli are transmitted into changes in melatonin release from the pineal gland and melatonin directly or indirectly inhibits gonadotrophin-releasing hormone (GnRH) synthesis (Strauss et al. 1979). Effects of melatonin on plasma prolactin concentrations have not yet been investigated in the equine species. We have therefore investigated effects of oestradiol, melatonin and season on the opioidergic regulation of LH and prolactin release in ovariectomized mares. Materials and Methods Animals The animals used in this study (n=7) were non-lactating Shetland pony mares, aged between 4 and 12 years (7-6 ± 2-7 years, ± s.d.) and weighing 161-4 ±8-0 kg ( ± s.d.). Ponies were kept in outdoor paddocks and fed hay and mineral supplements. Water was freely available. The mares had been ovariectomized by laparotomy using standard surgical procedures between 4 and 6 months before the first experiment. Experimental design The animals were used for three different experiments carried out over a 19-month period. Experiment 1 was performed in January and experiments 2 and 3 between May and July during 2 consecutive years. In experiment 1, mares (n=6) were first treated i.v. with naloxone (0-5 mg/kg body weight) and saline on 2 con secutive days to determine the effect of naloxone on LH and prolactin secretion in non steroid-treated, ovari ectomized mares outside the breeding season. Half of the mares were given naloxone on day 1 and saline on day 2 and the other half were treated in opposite order. Starting on day 2 after naloxone or saline experiments, mares received oestradiol benzoate (5 µg/kg body weight, s.c.) for 20 days once daily. On days 10 and 11 and days 20 and 21, the animals were again injected with naloxone and saline respectively. Oestradiol was always given in the afternoon and thus after naloxone or saline experiments. In experiment 2, mares (n=7) were treated with naloxone and saline on days 1 and 2 as described for experiment 1. After the naloxone/saline experiment on day 2, mares received melatonin (15 mg) for 15 days. Melatonin was given orally at 1530 h approximately 4 h before dusk. This treatment has been shown to advance the daily rise in plasma melatonin concentrations and thus imitates a short-day photoperiodic signal (Guillaume & Palmer 1991). Naloxone and saline experiments were repeated after 15 and 16 days. The mares were then given melatonin (15 mg orally) plus oestradiol benzoate (5 µg/kg body weight, s.c.) for 20 days and on days 10 and 11 and days 20 and 21 received naloxone and saline. One animal, for reasons not related to the experiment, had to be excluded from the study after day 15 of the combined melatonin plus oestradiol treatment. In experiment 3, mares (n=6) received naloxone and saline on 2 consecutive days to determine opioidergic effects on LH and prolactin secretion in non-pretreated, ovariectomized mares during the breeding season. Starting on day 2, mares received oestradiol benzoate (5 µg/kg body weight, s.c.) for 30 days once daily. Oestradiol was always injected after naloxone or saline experiments. On days 10 and 11, days 20 and 21 and days 30 and 31 after the start of oestradiol pretreatment, the animals received naloxone and saline as described for days 1 and 2. They were then left untreated for 12 days and subsequently pretreated with melatonin (15 nig orally) for 35 days. The LH and prolactin response to naloxone and saline was determined on the 2 days immediately before melatonin treatment, on days 15 and 16, days 25 and 26 and days 35 and 36 after the start of melatonin pretreatment with half of the mares receiving naloxone on the first day and saline on the second day and the other half treated in opposite order. Experimental procedures For each naloxone or saline injection an indwelling cath eter was placed in one jugular vein 30 min before the first blood sample was taken. Blood samples for determination of LH and prolactin were then collected at 15-min intervals for 3 h. After 60 min sampling, naloxone or 10 ml saline were injected via the catheter into the jugular vein. Blood was collected into heparinized polystyrol tubes. Samples were centrifuged immediately for 20 min at 1000 g and plasma was frozen at 20 C until hormone analysis. Naloxone (Sigma, Deisenhofen, Germany) was freshly dissolved in 10 ml saline and filter sterilized before

injection. Oestradiol benzoate (Sigma) was dissolved in corn oil at a concentration of 1 g/1 and administered s.c. at 24-h intervals. Melatonin (Sigma) was dissolved in ethanol at a concentration of 30-0 g/1 and 0-5 ml solution were given on 5 g sugar cubes. After evaporation of ethanol at room temperature, the cubes were stored at 20 C. The ponies were fed one melatonin-containing sugar cube per day. Hormone analysis Plasma concentrations of LH and prolactin were measured by RIA as described previously. The assay for LH utilized equine LH (UCB, Braine L'Alleud, Belgium) as standard and for iodination and an antibody raised in rabbits against equine LH (A 543, UCB). Intra- and interassay coef ficients of variation (CV) were 5-8 and 18-1% respectively; the minimal detectable concentration was 0-5 µg/l and cross-reactivity of the antibody with follicle-stimulating hormone was <2-8% (Behrens et al. 1993). The RIA for prolactin was performed with an antiserum raised in rabbits against equine prolactin (AFP 361687; Dr A F Parlow, Harbor-UCLA Medical Center, Torrance, CA, USA). Intra- and interassay CV values were 4-1 and 10-3% respectively; the minimal detectable concentration was 0-25 µg/l and cross-reactivity of the antibody with equine growth hormone (AFP 7424C, Dr A F Parlow) was <0-l% (Aurich et al. 1995). Statistical analysis Prolactin and LH release in individual mares was calcu lated as the area under the curve (AUC, µ /1/15 min) for the time period from 0 to 120 min after injection of naloxone or saline. The unit takes into account the 15-min sampling interval. The pretreatment baseline was sub tracted from the value for each post-treatment sample. For LH, the baseline was calculated as the mean of five values before naloxone or saline injections and for prolactin as the mean of the last three values before naloxone or saline injections. The first two samples were excluded for calculating basal prolactin concentrations in order to mini mize effects of a possible stress-induced prolactin release in association with i.v. catheterization. Statistical comparisons were made with the SPSS/PC+ statistics package (Norusis 1988). Because no assumption was made about the distri bution of data, non parametric tests were used. Naloxone and respective saline experiments were compared by Wilcoxon matched pairs ranked sum test. Basal hormone concentrations in the same animals at different times and naloxone-induced LH and prolactin release (AUC) after different oestradiol and melatonin pretreatments were compared by Friedman's two-way ANOVA, taking into account the sequential nature of data. For comparisons of basal hormone concentrations, values from control experiments were used. Data given are means ± s.e.m. Results Oestradiol As can be expected from the experimental design, plasma oestradiol concentrations were affected by the oestradiol pretreatment. Oestradiol concentrations were significantly with oestradiol increased when mares had been pretreated benzoate, whereas melatonin treatment did not cause any changes (see Table 1). LH In January (experiment 1), basal LH was not significantly affected by the oestradiol pretreatment (see Table 1). Naloxone induced a significant ( <0 5) release of LH both before and following 10 and 20 days of pretreatment with oestradiol, when the AUC was compared with the respective saline experiment (see Fig. la). The release of LH in response to naloxone (AUC), was significantly ( <0 5) higher when mares had been given oestradiol for 10 and 20 days than without steroid pretreatment. After 10 days of oestradiol pretreatment, maximal LH concen trations in response to naloxone were reached within 15 min (naloxone 32-4 ±11-7 vs saline 22-1 ± 10-1 µg/l) and alter 2(1 days of oestradiol pretreatment within 30 min (naloxone 50-5 ± 28-0 vs saline 32-5 ± 17-3 ng/1). In May (experiment 2), basal plasma LH concentrations were not affected by melatonin pretreatment for 15 days but increased significantly (P<001) following the com bined oestradiol plus melatonin pretreatment (see Table 1). Naloxone induced a significant (P<0-05) LH release, when mares had neither been pretreated with melatonin nor with oestradiol. After oral melatonin pretreatment for 15 days and after subsequent combined melatonin and oestradiol pretreatment for 10 days, naloxone did not significantly increase LH release. However, when the combined treatment was continued for an additional 10 days, naloxone injection was again followed by a signifi cant (P<0-05) LH release. The LH increment in response to naloxone (AUC) was significantly higher when mares had been pretreated with melatonin for 15 days followed by melatonin plus oestradiol benzoate for 20 days than in mares not pretreated ( <0 01; see Fig. 2d). Maximal LH concentrations in mares not pretreated with melatonin and oestradiol were 19-0 ± 3-3 and 14-9 ± 3-1 µg/l respect ively, 30 min after injection of naloxone and saline. After the combined 20-day melatonin plus oestradiol pre treatment maximal LH concentrations 15 min after naloxone injection were 47-4 ± 4 µg/l, compared with 37-4 ± 4-4 µg/l in saline experiments. In experiment 3, during the physiological breeding season of the horse, basal LH concentrations increased continuously during the 30-day oestradiol pretreatment (P<0-001). In contrast, melatonin pretreatment decreased basal LH concentrations when compared with the

Table 1 Basal LH, prolactin and oestradiol concentrations n plasma of mares before injections of naloxone and saline. Significant differences n LH, prolactin and oestradiol concentrations respectively, within Individual experiments are indicated in parentheses (ANOVA). ''^Significant differences between experiments when LH, prolactin and oestradiol concentrations respectively, before any melatonin or oestradiol treatment was initiated, were compared among experiments (ANOVA) Experiment No pretreatment Oestradiol for 10 days Oestradiol for 20 days No pretreatment Melatonin for 1 5 days Melatonin for 25 days + oestradiol for 10 days Melatonin for 35 days + oestradiol for 20 days No pretreatment Oestradiol for 10 days Oestradiol for 20 days Oestradiol for 30 days No pretreatment Melatonin for 1 5 days Melatonin for 25 days Melatonin for 35 days LH (µ8/ ) 9-9 ± 1-3a 20-1 ±8-3 30-4 ± 16 1 14-4±3-2b 1 5-7 ± 1-6 35-1 ±4-8 38-2 ±3-3 (P<0-01) 15-0±2-4b 39-1 ± 100 75-7 ± 13-7 93-9 ± 1 5 8 (P<0-001) 30-3 ± 6-0b 23-5 ±4-1 20-4 ±2-9 21-5±3-l (P<0-01) Prolactin (µß/ ) 2-8 ± 1-2 1-4 ±0-2 30 ±0-6 1-7±0-3a 1-0 ±0-4 1-2 ±0-4 0-4 ±0-1 (P<0-05) 5-0 ± 0-9b 6-6 ±1-1 7-6 ± 1-1 6-5 ± 1-2 6-9 ±1-6 2-3 ±0-3 2-5 ±0-6 2-4 ±0-7 (P<0-05) Oestradiol (pmol/l) 29-0 ±4-4 42-6 ±8-1 49-6 ± 4-0 (P<0-01) 36-3 ± 11-4 33-0 ±9-2 57-6 ±7-3 62-8 ±9-5 (P<0-05) 27-2 ±8-1 91-0 ± 19-8 59-8 ±7-7 75-3 ± 7-3 (P<0-05) 13-6 ±4-4 22-0 ±4-4 9-9 ±2-9 17-6 ±5-1 Values within columns with different superscripts are significantly (P<0-05) different. 300-, (a) 50 40- < 3 " 1 20- ra o ri 10-0 /» 10dE2 20dE2-10dE2 20dE2 Figure 1 (a) LH and (b) prolactin release calculated as area under the curve (AUC; ug/l/1 5 min) for the time period from 0 to 120 min after administration of naloxone (0-5 mg/kg; solid bars) or saline (open bars) in ovariectomized mares (ovx, n=6-7) during the nonbreeding season ± oestradiol benzoate (E2/ 5 µ /kg s.c.) pretreatment for 10 or 20 days. Values are means ± s.e.m. 'Significant difference between naloxone and respective saline experiments (P<0-05). pre-melatonin period ( <0 01; see Table 1). Injection of naloxone on day 1 and 2 (no pretreatment) resulted in a significant (P<0-05) release of LH, calculated as AUC. Mean maximal plasma LH concentrations 15 min after of naloxone and saline were 13-7 ±2-1 and injection 15-0 ±2-1 pg/1 respectively. Oestradiol pretreatment for 10, 20 and 30 days increased the naloxone-induced LH increment (P<0-05; see Fig. 2b). When mares had been pretreated with oestradiol for 30 days, LH concentrations 15 min after injection of naloxone and saline were 129-9 ± 21-9 and 95-3 ± 15-7 µg/l respectively. In con trast, melatonin pretreatment markedly reduced the LH response to naloxone. A significant difference between naloxone and saline experiments was found after 15 days but was no longer present after 25 and 35 days of melatonin pretreatment (see Fig. 2c).

75 (a) 50 (d) 50- i* * I. 40 < ^ 1 20 J! o a. 10- o- ^à^acm -25- ovx 15dM 25dM+E2 35d M+E2-10- 15dM 25dM+E2 35d M+E2 300 (b) 50- ir! 200- I" 100- I * 40 O Ü 30 c V 20 jo o rx 10- aill -100- ovx 10dE2 20dE2 30d E2-10 10dE2 20dE2 30d E2 75- (c) 50 (0 8< 50-25- Jcâie 40 3 30H - 20 Q. 10- III -25- ovx 15dM 25dM 35d M -10 15dM 25dM 35d M Figure 2 (a-c) LH and (d-f) prolactin release calculated as area under the curve (AUC; µg/l/15 min) for the time period from 0 to 120 min after administration of naloxone (0-5 mg/kg; solid bars) or saline (open bars) in ovariectomized mares (ovx, n=6-7) during the breeding season ± pretreatment over different time periods with (b, e) oestradiol benzoate (E2, 5 µg/kg s.c), (c, f) melatonin (M; 15 mg) or (a, d) E2+M. Values are means ± s.e.m. 'Significant difference between naloxone and respective saline experiments (P<005). Prolactin Basal prolactin concentrations were not affected by oestra diol pretreatment (experiments 1 and 3) but were signifi cantly (P<0-05) reduced by either melatonin (experiment 2) or melatonin combined with oestradiol (experiment 1; see Table 1). In experiment 1 (January), naloxone did not cause any changes in plasma prolactin concentrations, irrespective of any oestradiol pretreatment (see Fig. lb). In experiment 2,

during the breeding season, naloxone failed to induce prolactin release in mares not pretreated with melatonin or oestradiol. After oral melatonin pretreatment for 15 days, naloxone caused a significant ( <0 5) prolactin release, compared with AUC values from respective saline exper iments. Prolactin concentrations 30 min after naloxone or saline injections were 2-1 ±0-6 and 1-2 ±0-4 pg/1 respectively. No naloxone-inducible prolactin secretion was found after the combined melatonin and oestradiol pretreatment for 10 and 20 days (see Fig. 2d). In exper iment 3, injection of naloxone on days 1 and 2 (no pretreatment) again did not significantly change prolactin release. Oestradiol pretreatment for 10, 20 and 30 days had no effect on the prolactin response to naloxone (see Fig. 2e). After melatonin pretreatment for 15, 25 and 35 days, naloxone evoked a significant prolactin secretion. Although the naloxone-induced increment in prolactin release tended to increase during the melatonin pretreat ment, no significant difference was found. Prolactin con centrations 15 min after naloxone and saline injections were 3-9 ± 0-6 and 2-7 ± 0-4 pg/1 after 15 days, 5-5 ± 1-4 and 2-2 ± 0-5 pg/1 after 25 days and 5-7 ±1-3 and 23 ± 0-6 pg/1 respectively, after 35 days of melatonin pretreatment (see Fig. 2/). Discussion In this study, effects of oestradiol and melatonin on the opioidergic regulation of LH and prolactin release were investigated during the breeding and non-breeding season of the horse. In May, oestradiol pretreatment caused an increase in basal LH concentrations, reflecting a stimula tory effect of oestradiol on LH secretion previously described (Garcia & Ginther 1978, Baldwin et al. 1991, Thompson et al. 1991). This stimulation is exerted primar ily at the pituitary level (Baldwin et al. 1991). Melatonin pretreatment for up to 35 days beginning in May did not affect LH but decreased prolactin concentrations. In the horse, as in sheep (Kennaway et al. 1982, Lincoln & Ebling 1985) melatonin therefore is able to inhibit prolactin secretion. Before oestradiol pretreatment, LH concentrations in January were significantly lower than corresponding values in May, which is in agreement with the seasonal variation in LH release described previously (Turner et al. 1979). Animals in our study had been ovariectomized several months earlier. As has been suggested for sheep (Goodman et al. 1982, Robinson et al. 1985), seasonal changes in LH release in the horse seem to be regulated at least in part independently from ovarian factors. In experiment 3, LH concentrations before melatonin pretreatment tended to be higher than before oestradiol pretreatment. Although plasma oestradiol concentrations were not markedly different, the 12-day interval between oestradiol and melatonin pretreatments might have been too short, and residual, LH-stimulatory oestrogen effects before the melatonin treatment cannot be ruled out. Basal prolactin concentrations in experiment 3 were lower than in exper iment 2 although the time of year and experimental procedures were identical. Because experiment 3 was performed 12 months later than experiment 2, the differ ence may be linked to long-term changes in prolactin release after ovariectomy. In ovariectomized mares not pretreated with oestradiol or melatonin, naloxone caused a significant LH release, irrespective of the time of the year. This indicates, that in long-term ovariectomized pony mares, opioids inhibit LH secretion and is in contrast to results from short-term ovariectomized mares (Aurich et al. 1995). Therefore, an inhibitory opioidergic tone on LH release exists in ovaryintact mares (Behrens et al. 1993), is lost after ovariectomy (Aurich et al. 1995) but can be restored after prolonged absence of the ovaries. This is in agreement with results from ewes (Schillo et al. 1985) and female rabbits (Younglai et al. 1988), where a steroid-independent opioidergic inhibition of LH release exists under certain conditions. Melatonin pretreatment during the markedly reduced the opioidergic breeding season inhibition of LH secre tion. While LH concentrations still tended to increase in response to naloxone after 15 days of melatonin pretreat ment, no naloxone-reversible, i.e. opioidergic, inhibition was found after 25 and 35 days of melatonin. It is therefore unlikely that a prolonged period of daily melatonin secre tion at the end of the breeding season in autumn activates the opioidergic inhibition of LH release found in seasonally anovulatory mares (Aurich et al. 199Ad). In contrast, in non oestradiol-pretreated mares in winter, exposed to a natural short-day melatonin signal for several -weeks, an opioid ergic inhibition of LH release was present. Although short-term, i.e. 15- to 35-day, melatonin pretreatment inactivates the opioidergic regulation of LH release, the natural, prolonged influence of a short-day melatonin signal does not prevent reactivation of the opioidergic systems. The opioidergic regulation of LH release there fore seems to be influenced not only by short-term melatonin effects but also by the photoperiodic history of the animal. The exogenous melatonin signal given during the breeding season may also not be totally equivalent the endogenous profile to observed under natural winter days. The melatonin treatment used in our study has been shown to elevate plasma melatonin concentrations for several hours until after the onset of endogenous melatonin release at dusk (Guillaume & Palmer 1991). Because melatonin was always given at the same hour, gradual changes in the period of daily melatonin release as seen in winter were not imitated. Differences in the of LH release between melatonin- opioidergic regulation pretreated mares in summer and non-pretreated mares in winter may reflect in part a different pattern of melatonin signals.

In winter as well as during the breeding season, an opioidergic inhibition of LH release was present under the influence of exogenous oestradiol and the LH increment in response to naloxone increased with the duration of oestradiol treatment. This indicates that the opioidergic inhibition of GnRH and thus LH secretion can be enhanced by oestrogens. An oestradiol-dependent acti vation of LH-inhibitory opioidergic systems does exist also in short-term ovariectomized mares, pretreated for 8 days with oestradiol (Aurich et al. 1995). However, in that study, mares were investigated only in the breeding season and plasma oestradiol concentrations resembled values found in ovary-intact mares during the follicular phase of the cycle. Oestradiol concentrations in the present study were lower and comparable to those physiologically seen in mares with reduced ovarian activity outside the breed ing season (Aurich et al. 1994a). The opioidergic inhibition oflh release in seasonally anovulatory mares (Aurich et al. 1994a) therefore is in part steroid-independent enhanced by low oestradiol concentrations. Oestrogens but can be may be part of a negative feed-back loop acting on GnRH and thus LH secretion in non-breeding season mares as has been found also in sheep (Karsch et al. 1993). In the horse, these feedback mechanisms seem to involve opioidergic neuronal pathways. When melatonin and oestradiol were given combined, the inactivating effect of melatonin mechanisms was alone on the LH regulatory opioidergic reversed by oestradiol and an inhibitory opioidergic tone reactivated. An oestrogen-induced, opioidergic inhibition of GnRH secretion does not exclude stimulatory effects of oestradiol on LH release at the pituitary gland, explaining increased pre-naloxone LH concentrations in plasma of mares pretreated with oestradiol. Prolactin release is in part under opioidergic control but its regulation is different from the mechanisms regulating LH release. No naloxone-induced changes in prolactin secretion were found in non-pretreated, ovariectomized mares, irrespective of the season. Melatonin pretreatment during the breeding season induced an opioidergic inhi bition of prolactin secretion and the amount of prolactin secreted in response to the opioid antagonist increased with the duration of melatonin pretreatment. Although melatonin given for 15-35 days activated an opioidergic inhibition of prolactin release, the secretion of this hor mone was not under opioidergic control in mares in winter, exposed for a prolonged time period to an endog enous short-day melatonin signal. This suggests that al though a short-day melatonin signal initially activates an opioidergic inhibition of prolactin release, activity of the opioidergic systems changes with time. Opioids therefore could participate in a reduction in prolactin secretion in autumn but do not inhibit prolactin release throughout the non-breeding season. As in sheep, a prolonged inhibitory effect of short daylength on prolactin secretion could also be mediated through non-opioidergic mechanisms directly at the pituitary gland (Lincoln & Clarke 1994, 1995, Houghton et al. 1995) and such mechanisms would mask an opioidergic inhibition at the hypothalamic level. In conclusion, opioidergic neuronal systems, involved in the regulation of LH and prolactin secretion in the mare, interact with both melatonin and oestradiol. However, although seasonal variations in prolactin and LH release in the horse occur in parallel (Turner et al. 1979, Evans et al. 1991), the opioidergic regulation of these hormones is activated differentially. Given during the breeding melatonin inactivates the opioidergic inhibition season, of LH secretion but a prolonged influence of melatonin as in winter does not prevent reactivation of opioidergic systems. Irrespective of the season, the naloxone-reversible inhibition of LH release is enhanced by oestradiol, suggesting that the opioidergic regulation of LH release in seasonally anovulatory mares might in part be activated by low oestradiol concentrations. Oestradiol is also able to counteract the effects of melatonin on the LH regulatory opioidergic pathways. An opioidergic inhibition of prolactin release can be activated by melatonin but is lost under the prolonged influence of an endogenous short-day melatonin signal. Acknowledgements The authors are grateful to the Mehl-Miilhens Foundation for financial support, the Dorothea Erxleben Program of the State of Niedersachsen for a research fellowship to Dr C Aurich and to Dr A F Parlow, Harbor-UCLA Medical Center, Torrance, CA, USA, for reagents used in the prolactin RIA. References Aurich C, Schlote S, Hoppen H-O, Klug E, Hoppe H & Aurich JH 1994(3 Effects of the opioid antagonist naloxone on release of luteinizing hormone in mares during the anovulatory season. Journal of Endocrinology 142 139 144. Aurich C, Sieme H, Hoppe H & Schlote S 1994b Involvement of of LH and testosterone release endogenous opioids in the regulation in the male horse. Journal of Reproduction and Fertility 102 327 336. Aurich C, Daels PF, Ball BA & Aurich JE 1995 Effects of gonadal steroids on the opioid regulation of LH and prolactin release in ovariectomized pony mares. Journal of Endocrinology 147 195 202. Baldwin DM, Roser JF, Muyan M, Lasley & Dybdal 1991 Direct effects of free and conjugated steroids on GnRH stimulated LH release in cultured equine anterior pituitary cells. Journal of Reproduction and Fertility Suppl 44 327 332. Behrens C, Aurich JE, Klug E, Naumann & Hoppen H-O 1993 Inhibition of gonadotrophin release in mares during the luteal phase of the oestrous cycle by endogenous opioids. Journal of Reproduction and Fertility 98 509-514. Evans MJ, Alexander SL, Irvine CHG, Livesay JH & Donald RA 1991 In vitro and in vivo studies of equine prolactin secretion throughout the year. Journal of Reproduction and Fertility Suppl 44 27-35. and season on Garcia MC & Ginther OJ 1978 Effects of ovariectomy plasma luteinizing hormone in mares. Endocrinology 98 958-962.

Goodman RL, Bittman EL, Foster DL & Karsch FJ 1982 Alterations in the control of luteinizing hormone pulse frequency underlie the seasonal variation in estradiol negative feedback in the ewe. Biology of Reproduction 27 580-589. Guillaume D & Palmer E 1991 Effect of oral melatonin on the date of the first ovulation after ovarian inactivity in mares under artificial photoperiod. Journal of Reproduction and Fertility Suppl 44 249 257. Houghton DC, Young IR & McMillen IC 1995 Response of prolactin to different photoperiods after surgical disconnection of the hypothalamus and pituitary in sheep fetuses. Journal of Reproduction and Fertility 104 199-206. Johnson AL 1987 Seasonal and photoperiod-induced changes in serum prolactin and pituitary responsiveness to thyrotropin releasing hormone in the mare. Proceedings of the Society for Experimental Biology and Medicine 184 118-122. Karsch FJ, Dahl GE, Evans NP, Manning JM, Mayfield KP, Moenter SM & Foster DL 1993 Seasonal changes in gonadotrophin-releasing hormone secretion in the ewe: alteration in response to the negative feedback action of estradiol. Biology of Reproduction 49 1377 1383. Kennaway DJ, Gilmore TA & Seamark RF 1982 Effect of melatonin implants on the circadian rhythm of plasma melatonin and prolactin in sheep. Endocrinology 110 2187 2188. Lincoln GA & Ebling FJP 1985 Effect of constant-release implants of melatonin on seasonal cycles in reproduction, prolactin secretion and moulting in rams. Journal of Reproduction and Fertility 73 241 253. Lincoln GA & Clarke IJ 1994 Photoperiodically induced cycles in the secretion of prolactin in hypothalamo-pituitary disconnected rams: evidence for translation of the melatonin signal in the pituitary gland. Journal of Neuroendocrinology 6 251-260. Lincoln GA & Clarke IJ 1995 Evidence that melatonin acts at the mechanism to pituitary gland through a dopamine-independent mediate effects of daylength on the secretion of prolactin in the ram. Journal of Neuroendocrinology 7 637 643. Norusis MJ 1988 SPSS/PC+ for the IBM PC/XT/AT and PS2. SPSS Ine, Chicago. Robinson JE, Radford HM & Karsch FJ 1985 Seasonal changes in pulsatile luteinizing hormone (LH) secretion in the ewe, relationship of frequency of LH pulses to day length and response to estradiol negative feedback. Biology of Reproduction 33 324 334. Schillo KK, Kuehl D & Jackson GL 1985 Do endogenous opioid peptides mediate the effects of photoperiod on release of LH and prolactin in ovariectomized ewes? Biology of Reproduction 32 779-787. Strauss SS, Chen CL, Kalra SP & Sharp DC 1979 Localization of gonadotropin releasing honnone in the hypothalamus of ovariectomized pony mares. Journal of Reproduction and Fertility Suppl 27 123-129. Thompson DL, Johnson L, Garza F, St George L, Rabb MH, Barry BE Sc French DD 1991 Relationships among LH, FSH and prolactin secretion, storage and response to secretagogue and hypothalamic GnRH content in ovariectomized pony mares administered testosterone, dihydrotestosterone, estradiol, progesterone, dexamethasone or follicular fluid. Domestic Animal Endocrinology 8 189-199. Turner DD, Garcia MC & Ginther OJ 1979 Follicular and gonadotropic changes throughout the year in pony mares. American Journal of Veterinary Research 40 1694 1700. Worthy K, Colquhoun K, Escreet R, Dunlop M, Renton JP & Douglas TA 1987 Plasma prolactin concentrations in non-pregnant mares at different times of the year and in relation to events in the cycle. Journal of Reproduction and Fertility Suppl 35 269 276. Younglai EV, Wilkinson M & Thompson 1988 Effects of ovariectomy and estradiol replacement on naloxone induced LH secretion in the female rabbit. Journal of Steroid Biochemist^ 29 347-351. Received 4 December 1995 Revised manuscript received 25 November 1996 Accepted 27 February 1997