Theriogenology 61 (2004) 10611076

Inuence of oral melatonin on natural and gonadotropin-induced ovarian function in the domestic cat
L.H. Grahama,*, W.F. Swansonb, D.E. Wildta, J.L. Browna
a

Smithsonian National Zoological Park, Conservation & Research Center, Front Royal, VA, USA b Cincinnati Zoo and Botanical Garden, Center for Research of Endangered Wildlife, Cincinnati, OH, USA Received 5 December 2002; accepted 15 May 2003

Abstract Ovarian hyperstimulation after exogenous gonadotropin stimulation is believed to be a cause of poor success after articial insemination (AI) in felids. The objectives of this study were to assess the effect of oral melatonin on endogenous ovarian activity in the domestic cat and subsequent eCG/hCG-induced ovarian activity. Serum melatonin concentrations peaked $1 h after a single oral dose of 30 mg melatonin and remained elevated above endogenous day-time concentrations for >8 h. The calculated circulating half-life (mean S:E:M:) of oral melatonin was 45:4 3:5 min, and the elimination rate constant (k10) was 55:2 4:2 min1. Oral melatonin (30 mg per day) administered 3 h before lights-off effectively and reversibly suppressed estrous elevations in fecal estrogens after 25 days of treatment. There was a progressive decrease in baseline estrogen concentrations from inter-estrous concentrations after 25 days of treatment to below inter-estrous concentrations after 35 days of treatment. Oral melatonin treatment (30 mg per day for 30 days) prior to eCG/hCG administration only marginally reduced ancillary follicle development and had no signicant effect on the quantity or quality of embryos produced by AI. Thus, oral melatonin effectively inhibited endogenous ovarian activity and had no adverse impact on embryo quality after AI in the domestic cat; however, this treatment was only marginally effective in minimizing eCG/hCG-induced ovarian hyperstimulation. # 2003 Elsevier Inc. All rights reserved.
Keywords: Melatonin; Domestic cat; Articial insemination; Ovary

* Corresponding author. Present address: Toronto Zoo, Reproductive Physiology, 361A Old Finch Avenue, Scarborough, Ont., Canada M1B 5K7. Tel.: 1-416-392-5980; fax: 1-416-392-4979. E-mail address: endolaura@yahoo.com (L.H. Graham).

0093-691X/$ see front matter # 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.theriogenology.2003.05.004

1062

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1. Introduction Articial insemination (AI) has signicant potential for improving the ex situ management and conservation of wild felid populations [1]. However, AI does not consistently produce pregnancies in most felid species, possibly because eCG/hCGinduced ovarian hyperstimulation disrupts oviductal embryo transport [2,3]. Suppressing endogenous ovarian activity before AI or IVF/ET improved pregnancy rates in many species. In cows, hormonal control of the estrous cycle to inhibit follicular development and induce premature luteolysis can increase AI conception rates [4]. Analogues of GnRH are used in humans before AI or IVF/ET to induce temporary hypopituitaryhypogonadism that suppresses endogenous follicular activity [5,6]. Perhaps inhibiting endogenous ovarian activity before eCG/hCG treatment could reduce ovarian hyperstimulation, thus producing a maternal endocrine environment more conductive to establishing pregnancy. One approach for controlling follicular activity in the cat is to mimic seasonal ovarian inactivity. Annually recurring transitions between reproductive activity and ovarian quiescence in seasonal breeders constitute a natural process of gonadal activation and deactivation [712]. The female domestic cat is a seasonal breeder when exposed to natural photoperiod, with ovarian activity ceasing under decreasing photoperiod and resuming with increasing photoperiod (i.e. a long-day breeder). Melatonin secretion is controlled by the prevailing photoperiod (as in other mammals), with higher concentrations during the dark phase [1315]. Exogenous melatonin administered intravenously suppressed ovarian activity in the domestic cat maintained under a 24-h light photoperiod [15]. These preliminary studies suggested that melatonin is the signal by which the female domestic cat measures photoperiod and that exogenously administered melatonin may mimic the effects of decreasing photoperiod. The objectives of the present study were to: (1) determine the inuence of exogenous melatonin (oral) on ovarian dynamics in the domestic cat using non-invasive fecal hormone monitoring; and (2) assuming ovarian quiescence was induced, evaluate the gonadal response to eCG/hCG therapy and embryo characteristics after AI.

2. Materials and methods 2.1. General methods 2.1.1. Experimental animals Domestic short-hair queens (n 12) purchased commercially (Liberty Research Inc., Waverly, NY, USA) were 818 months of age at project onset. Queens were housed in pairs in stainless steel cages (1 m 1:5 m 1:5 m) and provided a commercial cat food diet (Purina Cat Chow; Ralston-Purina Co., St. Louis, MO, USA) and water ad libitum. Three intact male cats (age, 49 years) used as semen donors were housed individually in an adjacent room and fed the same diet. Cats were maintained in a controlled ambient environment under articial uorescent illumination (12 h light:12 h dark, with lights-off at 19:00 h). Litter boxes were checked daily, and all fecal samples were collected and

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1063

stored at 20 8C until processed. To facilitate identifying the origin of fecal samples from paired females, one cat per pair was fed 15 ml canned cat food daily containing 4 ml of green commercial food dye. 2.1.2. Fecal sample processing Frozen fecal samples were dried (Labconco Lyophilizer, Kansas City, MO, USA), pulverized and steroid metabolites extracted as described previously [3,16]. In brief, 0.2 g of fecal powder was extracted twice with 5 ml of 90% ethanol:distilled water, the supernatants combined, dried completely and re-dissolved in 1 ml methanol. Samples were diluted in phosphate buffered saline (0.01 M PO4, 0.14 M NaCl, 0.01% sodium azide) before analysis by RIA. All fecal data are expressed on a dry-weight basis. 2.1.3. Radioimmunoassays For quantication of melatonin, serum samples (0.5 ml) were extracted with 2.5 ml of dichloromethane. Extracts were taken to dryness under air and re-suspended in 0.5 ml of tricine-buffered saline (0.1 M, pH 5.5) with gelatin (0.1%). To monitor procedural losses, 3 H-labeled melatonin (400 dpm; Du Pont de Nemours Co., Boston, MA, USA) was added to each serum sample before extraction. Extraction efciency was 79:9 0:9%. Melatonin concentrations were measured in duplicate 100 ml aliquots of the extracted serum by RIA, using an antibody made in sheep to N-acetyl-5 methoxy tryptophan/bovine thryoglobulin (#YGU-N010; Accurate Chemical and Scientic Co., Westbury, NY, USA), 3 H-labeled melatonin tracer (Du Pont de Nemours) and melatonin standards (Sigma Chemical Co., St. Louis, MO, USA). The antibody cross-reacted 100% with melatonin and <1% with N-acetyl tryptamine, 6-hydroxy melatonin, N-acetyl tryptophan and other related compounds (product literature). The antisera and labeled melatonin were combined with extracted serum and standards (4500 pg per tube) and incubated overnight at 4 8C. After adding 0.5 ml charcoaldextran (2% charcoal, 0.02% mg dextran in tricine buffer) and a 15-min incubation (4 8C), tubes were centrifuged (500 g, 15 min) and the supernatant counted for 2 min in a Beckman LS5801 counter (GMI Inc., Albertville, MN, USA) after adding 3 ml of scintillation uid (Ultima Gold, Packard Instruments, Meriden, CT, USA). Final hormone concentrations were corrected for individual procedural losses and expressed as ng/ml of serum. The melatonin RIA was validated for cat serum by demonstrating: (1) parallelism between serial dilutions of serum extracts and the melatonin standards; and (2) recovery of exogenous melatonin added to serum extracts (y 1:04x 0:89, r 0:99). The intraand inter-assay coefcients of variation for cat serum controls were 8.7 and 10.0%, respectively. Gonadal steroid metabolite concentrations in the feces were quantied as described previously [3]. In brief, fecal estrogen concentrations were analyzed by RIA using an antibody raised in rabbits against estrogen-17b 6-o-carboxy-methyloxime:BSA (provided by Dr. Samuel Wasser, Center for Wildlife Conservation, Seattle, WA, USA). Fecal progestagens were measured by a RIA that used a monoclonal antibody produced against 4-pregnen-11-ol-3, 20-dione hemisuccinate:BSA (Quidel clone #425, provided by Dr. Jan Roser, University of California, Davis, CA, USA). Assay sensitivities were 2 and 5 pg/ml

1064

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

for the estrogen and progestagen assays, respectively, and intra- and inter-assay coefcients of variation were <10%. 2.1.4. Gonadotropin administration and laparoscopic evaluation Gonadotropin-induced ovarian stimulation and laparoscopy followed previously described procedures in our laboratory [17,18]. Queens were given 100 IU eCG i.m. (Sigma) to stimulate follicular development, followed by 75 IU hCG i.m. (Sigma) 80 h later to induce nal oocyte maturation and ovulation. Thirty-six to 38 h after hCG injection, queens were induced into a surgical plane of anesthesia with ketamine hydrochloride (Vetelar; Parke-Davis, Morris Plains, NJ, USA) plus acepromazine maleate (Ayerst Laboratories, Rouses Point, NY, USA) (10.0 mg/kg and 0.2 ng/kg body weight, respectively, i.m.). Anesthesia was maintained with isourane (Aerane; Anaquest, Madison, WI, USA) gas (12%) via face mask. The ovaries of anaesthetized queens were examined laparoscopically to determine number of mature vesicular follicles (!2 mm in diameter) and fresh corpora lutea (CL). 2.1.5. Articial insemination Ovulating queens were inseminated as follows. Sperm donors were anaesthetized with a tiletaminezolazepam combination (Telazol, Fort Dodge Laboratories, Fort Dodge, IA, USA; 5 mg/kg) and semen was collected via a standardized electroejaculation protocol [19,20]. Semen was diluted with an equal volume of 37 8C Hams F10 medium (HF10) supplemented with 5% fetal calf serum (FCS, Irvine Scientic, Irvine, CA, USA) [21] and centrifuged (150 g, 8 min). After decanting the supernatant, the sperm pellet was resuspended in 210 ml of fresh medium, maintained at room temperature and protected from light for 13 h until insemination. Queens were inseminated with washed sperm (100 ml per horn) deposited into the uterine lumen by inserting a 20-g indwelling catheter (Sherwood Medical, Tullamore, Ireland) percutaneously into the cranial portion of each uterine cornua [17]. Average number of motile spermatozoa inseminated was 9:8 106 (range, 4.1 to 14:8 106 ). 2.1.6. Embryo recovery Reproductive organs were excised via laparotomy and immediately processed by severing the oviducts from the uterine cornua at the uterotubal junctions. Oviducts and uterine cornuae were ushed separately and repeatedly into plastic Petri dishes with 15 ml of warm (37 8C) HF10 (Sigma) supplemented with 5% FCS. Number of embryos and oocytes were counted and embryos assessed for quality and developmental stage at 12 magnication [22,23]. Embryos were stained with Hoechst 33342 (Sigma) and the nuclei counted [24]. Embryos with !16 cells were considered morulae, and those with a developing blastocoelic cavity were classied as blastocysts. Criteria for three quality grades were as follows. Grade 1 embryos (G-1) had symmetrical blastomeres that were dark brown or black in color with no fragmentation or vacuolation. Grade 2 embryos (G-2) had slightly asymmetrical blastomeres, some with minimal fragmentation or vacuolation. Grade 3 (or degenerate) embryos (G-3) had blastomeres of variable size and were light or variable in color, most having pronounced fragmentation or vacuolations. Number and size of ovarian structures (follicles and CL) and ovarian mass was recorded.

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1065

2.2. Series of experiments 2.2.1. Experiment 1 The rst study was designed to determine the oral dose of melatonin required to maintain circulating melatonin at or above endogenous night-time concentrations for a minimum of 3 h. Blood samples were collected from eight female cats via jugular catheters as previously described [18]. In brief, 2 days before oral melatonin administration queens were anaesthetized, and a polymer resin catheter (Intra-cath Intravenous Catheter Placement Unit; Deseret Medical Inc., Parke, Davis and Company, Sandy, UT, USA; 19-gauge, 20.3-cm long) was inserted percutaneously into an external jugular vein. The catheter was sutured in place with the external portion extended laterally to the dorsal aspect of the neck and then thoroughly secured using bandage material (Vetrap Bandaging Tape; Animal Care Products, 3M Company, Saint Paul, MN, USA) and tape (Elastikon; Johnson and Johnson Medical Inc., Arlington, TX, USA). Forty-eight hours after catheterization, blood was collected from each animal at 09:00 h, and then again at 01:00 h under a light with a red lter (Premiere Multipurpose Darkroom Safelight, Penn Camera, Fairfax, VA, USA). Data from these samples were used to establish endogenous day- and night-time concentrations of serum melatonin, respectively. Seventy-two hours post-catheterization at 09:00 h, queens (n 4 cats per treatment) were dosed orally with gelatin capsules (A.J. Buck and Son Inc., Owings Mills, MD, USA) containing 30 or 60 mg melatonin (Sigma). Blood samples were collected from the catheters at 0, 0.5, 1, 2, 3, 4, 6, and 8 h after melatonin administration, centrifuged and the serum stored at 20 8C until the RIA was performed. 2.2.2. Experiment 2 The effect of daily oral melatonin administration (30 mg per day) on endogenous ovarian activity was examined in domestic cats (n 6) using noninvasive fecal hormone monitoring. Oral melatonin was administered daily at 16:00 h (3 h before lights-off) for 35 days. Fecal samples were collected daily for 8 weeks before melatonin treatment, during treatment and for 8 weeks after treatment cessation (n 821, $135 samples per cat). These fecal samples were processed and analyzed for hormone metabolites to assess ovarian activity as described above. 2.2.3. Experiment 3 The inuence of oral melatonin pre-treatment on ovarian hormonal and morphological responses to eCG/hCG administration was assessed in queens assigned randomly to: (1) ovarian stimulation by eCG/hCG only (n 5); or (2) ovarian stimulation initiated after 30 days of pre-treatment with oral melatonin (30 mg per day given 3 h before lights-off) with eCG administration 2 days after pre-treatment cessation (n 5). All queens were subjected to laparoscopic ovarian evaluation 36 h after hCG administration. Fecal samples were collected daily for 12 weeks before and after the onset of gonadotropin stimulation and stored before being used to assess endocrine patterns as described above (n 1428, $140 samples per cat). 2.2.4. Experiment 4 Effect of oral melatonin pre-treatment on embryo characteristics after AI was examined in queens assigned randomly to ovarian stimulation: (1) by eCG/hCG only (n 6); or (2)

1066

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

initiated after 30 days of pre-treatment with oral melatonin (30 mg per day 3 h before lights-off) with eCG administration (as in Experiment 3) 2 days after pre-treatment cessation (n 6). All queens were subjected to laparoscopic evaluation of their ovaries 36 h following hCG administration and ovulating queens were inseminated with fresh semen. All inseminated queens were anaesthetized and subjected to a midventral laparotomy and ovariohysterectomy [23] 160 h after hCG administration (5 days post-AI). Fecal samples were collected daily from 8 weeks before eCG/hCG stimulation to 5 days after ovariohysterectomy (n 722 samples, $60 samples per cat) and stored and processed for endocrine evaluation as described above. 2.3. Statistical analyses In Experiment 1, endogenous day-time (09:00 h) versus night-time (01:00 h) melatonin concentrations were compared using a Students t-test. To calculate the half-life and elimination rate of orally administered melatonin, data were analyzed as a one-compartment open model, assuming rst-order absorption and rst-order excretion from the system (PCNONLIN Version 4.2, Statistical Consultants, Inc., Lexington, KY, USA). Serum melatonin concentrations at different time points after exogenous administration within a treatment group were compared to endogenous day- and night-time concentrations using a Students t-test assuming unequal variances (Welchs t-test) and 95% Bonferroni interval comparisons. Serum melatonin concentrations were compared between treatment groups at the same time points using a Students t-test. In Experiment 2, fecal estrogen concentrations during estrus and inter-estrous periods were calculated for each individual by an iterative process [3]. Estrous cycle length was calculated as the number of days between fecal estrogen elevations lasting at least 3 days with no subsequent increase in fecal progestagen concentrations. Inter-estrus concentrations were dened as the fecal estrogen concentrations between estrus elevations in cycling queens. Because not all queens defecated every day, fecal estrogen and progestagen concentrations were averaged over 2-day intervals for each queen to compare temporal endocrine patterns after gonadotropin stimulation in Studies 3 and 4. Hormone data were compared between treatment groups using a Welchs t-test. The day of hCG injection was considered the day of the ovulatory stimulus (comparable to the rst copulation in natural estrual cats [23]) and designated as Day 0. Numbers of ovarian structures (follicles !2 mm and CL) at laparoscopy or after ovariohysterectomy were compared between groups using a Welchs t-test. In Experiment 4, ancillary follicle development (difference between the number of total structures, i.e. follicles !2 mm plus CL observed at laparoscopy and at ovariohysterectomy), the number of G-1/G-2 embryos, the number of G-1/G-2 morulae/blastocysts, and the proportion of embryos within the uterus also were compared using the Welchs t-test. Differences in ovulation rates (number of CL/number of follicles plus CL) and embryo recovery rates between treatment groups were compared using Chi-square analysis. Correlation coefcients were calculated between the location of recovered embryos (number of embryos recovered in the uterus/number of embryos recovered in the uterus plus oviducts) and fecal estrogen hormone concentrations.

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1067

3. Results 3.1. Experiment 1 Mean (S.E.M.) serum melatonin concentrations were higher (P < 0:05) at 01:00 h (8:94 2:6 ng/ml) than at 09:00 h (0:53 0:1 ng/Ml). Circulating melatonin also was elevated (P < 0:05) above day-time baseline concentrations in the rst sample after oral melatonin (0.5 h) and remained elevated (P < 0:05) for at least 4 h (Fig. 1). Serum melatonin peaked approximately 1 h after the 30 mg oral dose and 23 h after administering 60 mg, and was higher (P < 0:05) at 2, 3 and 4 h in the queens given 60 mg versus 30 mg melatonin. Concentrations decreased to endogenous night-time concentrations by 6 h in the 30 mg group and by 8 h in the 60 mg group. Serum melatonin concentrations failed to reach day-time concentrations during the 8-h sampling period in either group. Calculated circulating half-life of orally administered melatonin was 45:4 3:5 min, with an elimination rate constant (k10) of 55:2 4:2 min1. Since 30 mg melatonin maintained serum concentrations at or above endogenous night-time concentrations for at least 3 h, this dose was used in all subsequent studies. 3.2. Experiment 2 Fecal hormone monitoring revealed that each of the six test cats exhibited regular estrous cycles (923 days in duration; mean, 17:1 0:7 days) before melatonin treatment.
10000
* * *

1000

Serum melatonin (ng/mL)

100

10

0.1 0 2 4 6 Interval from melatonin administration (h) 8

Fig. 1. Serum melatonin concentrations (mean S:E:M:) in domestic cats (n 4 per treatment) after oral ingestion of 30 mg (*) or 60 mg (^) melatonin (P < 0:05).

1068

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

500 400 300 200 100 0 500 400 300 200 100 0 500 400 300 200 100 0 -40 -20 0

Melatonin treatment

Fecal estrogens (ng/g)

Melatonin treatment

Melatonin treatment

20 40 Day

60

80

120

Fig. 2. Fecal estrogen concentrations in three female cats before, during and after oral melatonin treatment (30 mg per day for 35 days).

Inter-estrual fecal estrogen concentrations averaged 127:1 6:9 ng/g and increased during estrus to 277:9 3:5 ng/g. Three of six cats experienced elevated fecal estrogen concentrations (indicative of follicular activity) during the melatonin treatment interval on Days 815, 1218 and 2024, respectively (Fig. 2; Day 1 first day of melatonin treatment). However, concentrations were at inter-estrous concentrations by Day 25 in all cats and remained at inter-estrous concentrations or below thereafter. During the last week of melatonin treatment (Days 2935), and for the next 3 weeks, fecal estrogen concentrations were lower (P < 0:05) than pre-treatment inter-estrous concentrations (Fig. 3). Queens returned to estrus 33:0 2:8 days (range 2140 days) after treatment cessation, and there was no synchronization of post-treatment estrous cycles. Fecal progestagens remained at baseline (9:1 1:1 mg/g) throughout the collection period in all queens. Based on this study, 30 days of melatonin treatment was sufcient to suppress estrual ovarian activity in all cats. Therefore, this treatment duration was used in subsequent studies. 3.3. Experiment 3 Although queens pre-treated with 30 days of oral melatonin tended to have a shorter elevation of fecal estrogens in response to eCG/hCG than queens without pre-treatment, the

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1069

Fig. 3. Suppression of fecal estrogen concentrations (mean S:E:M:) by daily oral melatonin (30 mg per day for 35 days) (, Different from inter-estrus pre-treatment mean; P < 0:05; n 6 queens).

average difference was not signicant (P 0:11; Fig. 4). Luteal fecal progestagen concentrations also tended (P 0:050.07) to be lower initially following eCG/hCG administration in melatonin pre-treated queens. Numbers of follicles and CL were similar between queens without and with melatonin pre-treatment (follicles: 2:2 0:8 versus 3:0 0:7, P 0:48; CL: 11:2 3:1 versus 11:6 1:0, P 0:91, respectively). 3.4. Experiment 4 Average fecal estrogen and progestagen concentrations did not differ between eCG/hCG treated cats that were or were not pre-treated with melatonin. The number of unovulated follicles and CL observed at laparoscopic AI or after ovariohysterectomy also did not differ (P > 0:05; Table 1). Queens pre-treated with oral melatonin tended (P 0:08) to have fewer ancillary follicles (follicles that developed after AI) than queens without pre-treatment (4:5 1:78 versus 11:2 3:0, respectively). Ancillary follicle development was observed in 10 of 12 (83%) eCG/hCG stimulated cats, regardless of presence or absence of pre-treatment. Mean number of embryos recovered, number of good quality embryos (G-1/G-2), stage of development and location of recovery (oviducts, uterus) were similar (P > 0:05) with or without melatonin pre-treatment (Table 2). Three of six eCG/hCG-treated cats with no pre-treatment and two of six cats pre-treated with melatonin produced good quality morulae or blastocysts. There was a correlation between the timing of the eCG/hCG-induced fecal estrogen peak and the location of the embryos in queens with or without melatonin pre-treatment (P < 0:05; r 0:74 and 0.68, respectively; Fig. 5). As the time from gonadotropin administration to the fecal estrogen peak increased, proportionately fewer embryos reached the uterus by the time of ovariohysterectomy.

1070

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1200 Fecal estrogens (ng/g) 1000 800 600 400 200 0 -2

(a)

0 2 4 6 8 10 12 14 Days from hCG administration

300 Fecal progestagens (g/g)

200

100

0 0
(b)

7 14 21 28 35 Days from hCG administration

42

Fig. 4. Temporal proles (mean S:E:M:) of fecal estrogen (a) and progestagens (b) in gonadotropin stimulated queens with (^) and without (*) melatonin pre-treatment.

Table 1 Mean (S.E.M.) numbers of ovarian structures observed at laparoscopy (Day 2) or at ovariohysterectomy (Day 7) in queens given eCG and hCG Pre-treatment Day 2 Follicles (>2 mm) Control Melatonin NB* P-value** 6.3 2.7 2.8 0.8 0.25 Corpora lutea 9.7 2.9 6.3 2.3 0.39 Day 7 Follicles (>2 mm) 8.3 2.4 5.0 2.1 2.6 1.6 0.31 Corpora lutea 18.8 6.4 8.7 3.3 6.5 1.1 0.19 Ancillary follicles 11.2 3.0 4.5 1.8 0.08

Day 0: day of hCG administration; control: no melatonin pre-treatment; melatonin: melatonin pre-treatment of 30 mg per day for 30 days. * For comparative purposes, historical data adapted from a previous report (from our laboratory on naturally estrual, mated queens ovariohysterectomized 148 h after breeding) are included [23]; data not included in statistical analyses. ** Differences between pre-treatment groups.

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1071

Table 2 Mean (S.E.M.) embryo number, development and location within the reproductive tract at the time of ovariohysterectomy (Day 7) in queens administered eCG and hCG and subjected to AI Pre-treatment Control Melatonin NB** P-value*** Recovery (%)* 52.5 12.8 60.9 10.1 56.4 0.62 G-1/G-2 embryos recovered per cat 2.5 1.1 2.5 1.2 3.0 0.9 1.00 G-1/G-2 embryos developed to morula or blastocyst per cat 1.3 0.6 1.2 0.7 2.5 0.87 Embryos in uterus (%) 58.0 17.4 57.8 20.1 94.4 0.99

Day 0: day of hCG administration; control: no melatonin pre-treatment; melatonin: melatonin pre-treatment of 30 mg per day for 30 days. * Number of recovered embryos/number of corpora lutea observed at the time of ovariohysterectomy. ** For comparative purposes, historical data adapted from a previous report (from our laboratory on naturally estrual, mated queens ovariohysterectomized 148 h after ovulatory stimulus) are included [23]; data not included in statistical analyses. *** Differences between pre-treatment groups.

100

Embryos recovered from uterus (%)

80

60

40

20

10

Interval from hCG to estrogen peak (d )

Fig. 5. Relationship between the day of peak estrogens and the anatomic location of recovered embryos in gonadotropin stimulated queens with (^, solid line, r 0:73, P < 0:05) or without (*, dashed line, r 0:68, P < 0:05) melatonin pre-treatment.

4. Discussion This investigation describes the inuence of oral melatonin administration on endogenous and gonadotropin-induced ovarian activity and embryo characteristics after AI in

1072

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

the domestic cat. There were three major ndings: (1) orally administered melatonin (30 mg per day 3 h prior to lights-off) effectively and reversibly inhibited endogenous ovarian activity; (2) a 30-day oral melatonin pre-treatment to suppress ovarian activity was only marginally effective in reducing eCG/hCG-induced ovarian hyperstimulation; and (3) a 30-day oral melatonin pre-treatment before eCG/hCG administration and AI had no detrimental impact on embryo quantity or quality. Endogenous night-time serum melatonin concentrations reportedly range from 20 pg/ml (ferrets) to 500 pg/ml (mink) [12,2527]. Thus, night-time melatonin concentrations in the domestic cat (39 ng/ml in this study; 49 ng/ml [13]) were considerably higher than those measured in other mammals. Although the pineal gland of the cat is known to have several unique morphological features [28], there have been no adaptive or functional relationships established between pineal gland morphology, melatonin concentrations and ecological niche. Regardless, serum melatonin concentrations in the domestic cat were approximately 15-fold higher during the dark phase than the light phase, similar to ndings in other mammals [8,9,11,12,2931]. Serum melatonin concentrations increased rapidly in domestic cats after oral administration, indicating that it was readily absorbed. Furthermore, 30 mg melatonin was sufcient to maintain serum concentrations at or above endogenous night-time concentrations for at least 3 h. The half-life of intravenously injected melatonin is only $4 min [15]. In contrast, our study revealed that the half-life of orally administered melatonin is considerably longer (45:4 3:5 min) and more in agreement with ranges reported in other species; e.g. 20 min in the human [31] and hamster [32] and several hours in the sheep [33]. Finally, serum melatonin concentrations in the two dosage groups (30 mg versus 60 mg) overlapped during the 2-h post-administration period, suggesting that containment within gelatin capsules may have limited the rate at which it dissolved in the gut. In the domestic cat, the duration of endogenous melatonin elevation corresponds to the duration of the dark cycle of the prevailing photoperiod [13]. The goal of Experiment 2 was to extend the daily duration of melatonin elevation from the photoperiod-induced 12-h elevation to a 15-h elevation with the administration of oral melatonin 3 h prior to the onset of darkness. Although 50% of the queens exhibited an elevation in fecal estrogens suggestive of follicular growth during the rst 24 days of oral melatonin treatment, all cats were at baseline inter-estrous concentrations after 25 days of treatment and remained there or lower thereafter. Similarly, intravenously administered melatonin (5 mg melatonin every other day) suppressed ovarian activity in cats maintained on 24-h light regimen, although 50% of cats exhibited ovarian activity during the rst 5 days of treatment [15]. In contrast, an increase from 10 to 16 h of dark caused an immediate shutdown in ovarian activity in cats [14]. These variable responses to a decrease in actual or perceived photoperiod could be due to a host of combined/integrated factors (e.g. duration of melatonin elevation, photoperiodic threshold and photoperiodic history) all of which could inuence the magnitude and temporal prole of reproductive responses [8]. For example, the Syrian hamster expresses a graded response to changes in photoperiod, with the rate of reproductive arrest correlated with the duration of melatonin elevation [34]. The cat may express a similar graded response, which would explain why extending the duration of melatonin elevation by 6 h immediately suppressed follicular activity [14],

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1073

whereas extending it 3 h or less resulted in delayed suppression in the present study and a previous study [15]. A graded response to changes in melatonin elevation could also explain the decrease in fecal estrogen concentrations from baseline inter-estrous concentrations after 25 days of treatment to below inter-estrus concentrations after 35 days of treatment. In our investigation, cats maintained under 12 h of darkness with night-time melatonin extended 3 h by oral melatonin returned to estrus, on average, 33 days after the end of treatment. This seemed longer than the mean of 16 days observed in cats switched from an inhibitory (16 h of darkness) to a stimulatory (10 h of darkness) photoperiod [14]. As discussed above, the different durations of melatonin elevation and photoperiodic history in the two studies may account for the disparity in results. Both studies reported considerable variation in resumption of estrous activity post-treatment (1144 days, in our study; 1226 days [14]) with no estrous synchrony among queens. Administering melatonin i.v. to domestic cats already exposed to an inhibitory photoperiod (14 h of darkness) resulted in ovarian follicular growth after 70 days of treatment [15]. Thus, the cat appears to become desensitized to elevated melatonin over time (i.e. photorefractory) [8]. Although the ovarian response to prolonged oral melatonin treatment (>35 days) has not been formally assessed in the cat, we conducted a preliminary study where two oral melatonin treatments (30 day per treatment) were conducted within 50 days of one another. The rst treatment (30 mg per day) inhibited ovarian activity, but the second (60 mg per day) had no effect, and cats in the latter group continued to exhibit normal estrous cycles throughout the treatment period (Graham and Brown, unpublished results). This observation suggests that the domestic cat can become refractory to repeated changes in melatonin exposure. In the Syrian hamster, photorefractoriness is associated with a loss of gonadal responsiveness to melatonin, and subsequent exposure to long days is required to restore sensitivity to its suppressive effects [35]. Our preliminary results suggest that in the cat, more than 50 days of exposure to long days is required to restore gonadal responsiveness to melatonin and its suppressive effects. It was hypothesized that pharmacologically mimicking a non-breeding season photoperiod and inactivating the ovaries would prevent ovarian hyperstimulation and ancillary follicular development in response to eCG/hCG administration. Melatonin pre-treated queens tended to have lower fecal hormone concentrations and fewer ancillary follicles. Even in melatonin-treated cats, the mean number of unovulated follicles at ovariohysterectomy remained twice that reported for naturally estrual, mated queens [23]. Furthermore, many queens had a delayed peak in fecal estrogens (presumably from ancillary follicle development) and delayed oviductal embryo transport, as reported previously in eCG/hCG treated cats [3]. In the present study, only 60% of embryos were recovered from the uterus at 160 h post-hCG. In contrast, >90% of embryos are found in the uterus by 148 h after rst copulation in naturally estrual cats [23]. There was some concern that a 30-day melatonin pre-treatment might adversely inuence embryo development, but we observed no reduction in the quality or quantity of embryos in melatonin-treated cats after AI. This contrasts with the results of a recent study in domestic cats that showed oocyte nuclear maturation, fertilization and developmental rates were depressed in oocytes recovered from August to October, compared to oocytes collected from May to July in cats exposed to natural photoperiod [36].

1074

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

These contrasting results could be explained by the observed progressive ovarian inactivation under perceived shortened day-length. Our results from Experiment 2 suggested that ovarian inactivation induced by increasing the daily duration of melatonin elevation was greater after 35 days of treatment compared to 30 days of treatment. In the previous study [36], cats would have been exposed to increasing duration of melatonin elevation for more than 35 days from the summer solstice (June 21) to early August. Thus, the shorter duration (30 days) of melatonin treatment in our AI study may have prevented any negative effects of the perceived shortened day-length on embryo quality. A prolonged (>30 days) melatonin treatment might result in ovarian inactivation more similar to that observed under natural photoperiod during early August, with a corresponding decrease in oocyte quality. These studies provide interesting insights about the ovarian responses to changes in the perceived photoperiod in domestic cats. Orally administered melatonin was effective in inhibiting ovarian activity but was not completely effective in preventing the ovarian hyperstimulation that results from eCG/hCG administration. Perhaps a longer duration of melatonin pre-treatment would provide a greater inhibition of ovarian activity and reduce the ovarian hyperactivity response to eCG/hCG; however, this could result in decreased oocyte quality. Perhaps the variable ovarian hyperactivity in felids treated with eCG/hCG is not only due to differences in the stage of the estrous cycle at the time of gonadotropin administration, but is also related to differences in individuals sensitivity to eCG and hCG. Protocols using steroids or GnRH analogs to suppress endogenous ovarian activity as well as changes in the type and timing of the exogenous gonadotropin treatment are alternatives that could be investigated to improve AI efciency in felids.

Acknowledgements This research was supported by grants from the Smithsonian Institution Scholarly Studies Program, the Smithsonian Institution Pre-Doctoral Fellowship Program, the Womens Committee of the Smithsonian Institution and the Friends of the National Zoo. We thank Michelle Sommers; Reagan Hardy and Lisa Ware for animal care assistance. We especially thank Dr. Blaire Osborn for providing the pharmacokinetic analysis of the melatonin data. We also thank the Purina C.A.R.E.S. Fund for donation of cat food. References
[1] Wildt DE, Monfort SL, Donoghue AM, Johnston LA, Howard JG. Embryogenesis in conservation biologyor, how to make an endangered species embryo. Theriogenology 1992;37:16184. [2] Graham LH, Byers AP, Wildt DE, Armstrong DL, Brown JL. Natural versus chorionic gonadotropininduced ovarian responses in the tiger assessed by fecal steroids. Biol Reprod 1996;54:114 [abstract]. [3] Graham LH, Swanson WF, Brown JL. Chorionic gonadotropin administration in domestic cats causes an abnormal endocrine environment that disrupts oviductal embryo transport. Theriogenology 2000;54: 111731. [4] Macmillan KL, Burke CR. Effects of estrous cycle control on reproductive efciency. Anim Reprod Sci 1996;42:30720.

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

1075

[5] Lindner C, Braendle W, Lichtenberg V, Kohler S, Bettendor G. Ovarian response and pregnancy rates in in vitro fertilization, gamete intrafallopian transfer and in vivo fertilization therapies after combined gonadotropin-releasing hormone agonist/human menopausal gonadotropin stimulation. Gyn Obst Invest 1990;29:1404. [6] Ron-El R, Herman A, Golan A, Nachum H, Soffer Y, Caspi E. Gonadotropins and combined gonadotropin-releasing hormone agonist-gonadotropins protocols in a randomized prospective study. Fertil Steril 1991;55:5748. [7] Karsch FJ, Woodll CJI. Neuroendocrinology of seasonal breeding: mode of action of melatonin. In: Yen SSC, Vale WW, editors. Neuroendocrine regulation of reproduction. Norwell, MA: Serono Symposia; 1990. p. 917. [8] Karsch FJ, Woodll CJI, Malpaux B, Robinson JE, Wayne NL. Melatonin and mammalian photoperiodism: synchronization of annual reproductive cycles. In: SSC Yen, Vale WW, editors. Suprachiasmatic nucleus: the minds clock. Norwell, MA: Serono Symposia; 1991. p. 21731. [9] Bartness TJ, Powers JB, Hastings MH, Bittman EL, Goldman BD. The timed infusion paradigm for melatonin delivery: what has it taught us about the melatonin signal, its reception and the photoperiodic control of seasonal responses? J Pineal Res 1993;15:16190. [10] Matthew CD, Guerin MV, Deed JR. Melatonin and photoperiodic time measurement: seasonal breeding in the sheep. J Pineal Res 1993;14:10516. [11] Paterson AM, Foldes A. Melatonin and farm animals: endogenous rhythms and exogenous applications. J Pineal Res 1994;16:16777. [12] Arendt J. Melatonin and the pineal gland: inuence on mammalian seasonal and circadian physiology. Rev Reprod 1998;3:1322. [13] Leyva H, Addiego L, Stabenfeldt G. The effect of different photoperiods on plasma concentrations of melatonin, prolactin and cortisol in the domestic cat. Endocrinology 1984;115:172936. [14] Leyva H, Madley T, Stabenfeldt GH. Effect of light manipulation on ovarian activity and melatonin and prolactin secretion in the domestic cat. J Reprod Fertil Suppl 1989a;39:12533. [15] Leyva H, Madley T, Stabenfeldt GH. Effect of melatonin on photoperiod responses, ovarian secretion of estrogen and coital responses in the domestic cat. J Reprod Fertil Suppl 1989;39:13542. [16] Brown JL, Wasser SK, Wildt DE, Graham LH. Comparative aspects of steroid hormone metabolism and ovarian activity in felids, measured noninvasively in feces. Biol Reprod 1994;51:77686. [17] Howard JG, Barone MA, Donoghue AM, Wildt DE. The effect of pre-ovulatory anesthesia on ovulation in laparoscopically inseminated domestic cats. J Reprod Fertil 1992;96:17586. [18] Swanson WF, Wolfe BA, Brown JL, Martin-Jimenez T, Riviere JE, Roth TL, et al. Pharmacokinetics and ovarian-stimulatory effects of equine and human chorionic gonadotropins administered singly and in combination in the domestic cat. Biol Reprod 1997;57:295302. [19] Wildt DE, Bush M, Howard JG, Oien SJ, Meltzer D, van Dyk A, et al. Unique seminal quality in the South African cheetah and a comparative evaluation in the domestic cat. Biol Reprod 1983;29:101925. [20] Howard JG, Bush M, Wildt DE. Semen collection, analysis and cryopreservation in nondomestic mammals. In: Morrow DA, editor. Current therapy in theriogenology. Philadelphia: WB Saunders Co; 1986. p. 104753. [21] Johnston LA, Donoghue AM, OBrien SJ, Wildt DE. Inuence of temperature and gas atmosphere on in vitro fertilization and embryo development in domestic cats. J Reprod Fertil 1991;92:37782. [22] Goodrowe KL, Howard JG, Wildt DE. Comparison of embryo recovery, embryo quality, oestradiol-17b and progesterone proles in domestic cats (Felis catus) at natural or induced estrus. J Reprod Fertil 1988a;82:55361. [23] Swanson WF, Roth TL, Wildt DE. In vivo embryogenesis, embryo migration and embryonic mortality in the domestic cat. Biol Reprod 1994;51:45264. [24] Goodrowe KL, Wall RJ, OBrien SJ, Schmidt PM, Wildt DE. Developmental competence of domestic cat follicular oocytes after fertilization in vitro. Biol Reprod 1988;39:35572. [25] Kennaway DJ, Frith RG, Phillipou G, Matthews CD, Seamark RF. A specic radioimmunoassay for melatonin in biological tissue and uids and its validation by gas chromatography-mass spectrometry. Endocrinology 1977;101:11927. [26] Baum MJ, Lynch HJ, Gallagher CA, Deng MH. Plasma and pineal melatonin levels in female ferrets housed under long or short photoperiods. Biol Reprod 1986;34:96100.

1076

L.H. Graham et al. / Theriogenology 61 (2004) 10611076

[27] NToumi F, Martinet L, Mondain-Monval M. Effects of melatonin treatment on the gonadotropin-releasing hormone neuronal system and on gonadotropin secretion in male mink, Mustela vison, in the presence or absence of testosterone feedback. J Pineal Res 1994;16:1825. [28] Boya J, Calvo JL, Rancano D. Structure of the pineal gland in the adult cat. J Pineal Res 1995;18:1128. [29] Ebels I, Balemans MGM. Physiological aspects of pineal functions in mammals. Physiol Rev 1986;66: 581605. [30] Reiter RJ. Mechanisms of reproductive regulation by the pineal gland. In: Yen SCC, Vale WW, editors. Neuroendocrine regulation of reproduction. Norwell, MA: Serono Symposia; 1990. p. 10512. [31] Waldhauser F, Waldhauser M, Lieberman HR, Deng MH, Lynch HJ, Wurtman RJ. Bioavailability of oral melatonin in humans. Neuroendocrinology 1984;39:30713. [32] Brown GM, Seggie J, Grota LJ. Serum melatonin response to melatonin administration in the Syrian hamster. Neuroendocrinology 1985;41:315. [33] Kennaway DJ, Seamark RF. Circulating levels of melatonin following its oral administration or subcutaneous injection in sheep and goats. Aust J Biol 1980;33:34953. [34] Powers JB, Jetton AE, Mangels RA, Bittman EL. Effects of photoperiod duration and melatonin signal characteristics on the reproductive system of male Syrian hamsters. J Neuroendocrinol 1997;9:45166. [35] Bittman EL, Zucker I. Photoperiodic termination of hamster refractoriness: participation of the pineal gland. Biol Reprod 1981;24:56872. [36] Spindler RE, Wildt DE. Circannual variations in intraovarian oocyte but not epididymal sperm quality in the domestic cat. Biol Reprod 1999;61:18894.