Equine embryo transfer: the effect of semen processing and donor mare management on recovery rates

1Departamento de Medicina Veterinária/Centro de Investigação Vasco da Gama (DMV/CIVG), Escola Universitária Vasco da Gama. Coimbra. Portugal. 2Instituto Nacional de Investigação Agrária e Veterinária. Vale de Santarém. Portugal. 3Centro Interdisciplinar de Investigação em Sanidade Animal (CIISA), Universidade de Lisboa. Lisboa. Portugal. 4Central Equina de Reprodução, Boituva, Brasil. 5LAQV – REQUIMTE, Bromatology, Pharmacognosy and Analytical Sciences Group Pharmacy Faculty, Coimbra University. Coimbra. Portugal.


SUMMARY
Embryo transfer (ET) is a biotechnology that allows to get more than one foal from a single mare during a breading season and also to prevent the removal of mares from their competition careers. Nevertheless, to achieve a successful outcome, the association of many factors must be considered, such as the management of donor and recipient mares, the stallions and the veterinarian´s experience. In this context, this study aimed to evaluate (1) if some of characteristics related to the donor mares (e.g. breed and the reproductive status, may influence the embryo recovery; (2) if different donor mares location may have impact in breeding programs; (3) the effect of some reproductive techniques and embryo transfer methodologies. For such purpose, 156 uterine flushes and 88 embryo transfers were performed. Higher embryo recovery rates at day 8 (61.5%) and an overall pregnancy rate of 62.2% were obtained. An influence of the month of the year (P=0.04) and of the type of semen (fresh or frozen; P=0.03) on the number of positive embryo collections were identified. The use of fresh semen for donor mares' insemination and recovery at day 8 increased the number of embryos that were collected during this period, although lower results were obtained in July-August. No other significant effects upon the remaining determinants under testing were observed, although stallions were responsible for inducing 11% of the variability within embryo recovery rates. These results should be considered in mare embryo transfer centres. Transferência de embriões equinos: o efeito do processamento de sêmen e do manejo da égua do doador nas taxas de recuperação resUmo A transferência de embriões (ET) é uma biotecnologia que permite obter mais de um potro de uma única égua durante uma temporada de empanar e também impedir a remoção de éguas de suas carreiras na competição. No entanto, para alcançar um resultado bem-sucedido, deve-se considerar a associação de muitos fatores, como a gestão de éguas doadoras e receptoras, os garanhões e a experiência do veterinário. Nesse contexto, este estudo teve como objetivo avaliar (1) se algumas das características relacionadas às éguas doadoras (por exemplo, raça e estado reprodutivo, podem influenciar a recuperação do embrião; (2) se diferentes locais de éguas doadoras podem ter impacto em programas de reprodução; (3) o efeito de algumas técnicas reprodutivas e metodologias de transferência de embriões. Para isso, foram realizadas 156 descargas uterinas e 88 transferências de embriões. Maiores taxas de recuperação de embriões no dia 8 (61,5%) e uma taxa global de gravidez de 62,2% foram obtidas. Influência do mês do ano (P=0,04) e do tipo de sêmen (fresco ou congelado; P=0,03) no número de coletas positivas de embriões foram identificadas. O uso de sêmen fresco para inseminação e recuperação de éguas doadoras no dia 8 aumentou o número de embriões coletados nesse período, embora resultados mais baixos tenham sido obtidos em julho-agosto. Não foram observados outros efeitos significativos sobre os demais determinantes em teste, embora os garanhões tenham sido responsáveis por induzir 11% da variabilidade dentro das taxas de recuperação de embriões. Esses resultados devem ser considerados em centros de transferência de embriões de égua.

INTRODUCTION
In breeding programs, assisted reproductive biotechnologies such as artificial insemination (AI) and these biotechnologies support owners and mares, to prevent the interruption of their training plans (Campbell 2014, pp. 322-327) allowing to obtain genetically improved animals. Brazil is one of the leading countries in equine embryo collection and transfer, presenting improved results year after year (Squires et al. 2003, pp. 151-170). Nevertheless there are still a few difficulties in equine reproduction such as the response to hormonal treatments, in particular to the superovulation protocols. In fact, these protocols are not routinely used due to the much lower ovarian response of mares compared to other species (e.g. bovine), but also because a reliable hormonal treatment (i.e. products, dosage regimen among others) is not commercially available. Moreover, the cryosurvival of horse embryos is lower when compared to other species and in vitro embryo production is quite difficult. The availability of multiple embryos from each donor mare allied to a successful cryopreservation could further expand this technology in a near future (Squires et al. 2003, pp. 151-170;Squires 2007, pp. 1-8).
To improve the pregnancy rates during the breeding season, three important aspects must be considered: management and status of donor and recipient mares, semen quality of the stallions (Samper 2011, pp. 219-228) and the experience of the veterinarian (Munroe 2011, pp.242-380). In fact and according to Jacob et al. (2012Jacob et al. ( , pp. 1159Jacob et al. ( -1166 the time of the uterine flushing in relation to the day of ovulation is one of the most important factors in determining the success of obtaining an embryo from a donor mare. Nonetheless other factors such as the mare´s age, breed, reproductive category, number of ovulations as well as the location of semen deposition during AI are also essential (Panzani et al. 2014, pp.807-814). Therefore is not surprising that the accomplishment of embryo recovery and transfer in the mare is reported as varying significantly (Jacob et al. 2012(Jacob et al. , pp. 1159(Jacob et al. -1166Aurich et al. 2011, pp.419-422).
The objectives of this study were to determine (1) if some of characteristics related to the donor mares such as the breed and the reproductive status, may influence the embryo recovery; (2) if different donor mares location may have impact in breeding programs; (3) the influence of some reproductive techniques and embryo transfer methodologies such as the day of uterine flushing, the type of semen, the stallion and the moment of insemination and transfer as potential effectors of success.

Embryo donor and recipient mares
This study was conducted in the Equine Reproduction Centre (CER, from Portuguese Central Equina de Reprodução), a commercial equine reproduction centre, accredited by the European Union located in Boituva, São Paulo, Brazil.
From July to December of 2015, eighty two mares of different breeds [Thoroughbred Arabian (n=48), Thoroughbred Lusitano (n=9) and Quarter Horse (n=25)] were included as donors in the embryo transfer pro-gram. Their age ranged from four to twenty years old. Most of them were housed in the CER (n=65) along with the crossbred recipient mares (n=170). The remaining seventeen donor mares were kept in a subsidiary centre (Raphaela Haras) located in Porto Feliz, São Paulo, Brazil. Regardless of the donor mare location, in case of positive embryo recover, recipient mares of CER were used.
Donor mares were kept in stalls with water ad libitum and fed with hay and commercial concentrated dry feed. Recipient mares were provided with water ad libitum but were fed with forage grass. The later ones received concentrated dry feed only when taken to stocks for reproductive status evaluation.

Monitoring ovarian activity
Ovarian activity of the mares was monitored using transrectal palpation and ultrasonography (6.5 MHz rectal transducer, C40 vet™, Landwind Medical, China). The donor mares required a daily detailed evaluation, while recipient mares were submitted to the same control but only three days a week (Monday, Wednesday and Friday). When new mares with no reproductive history arrived at CER, they were examined by rectal palpation and ultrasound to check for their cycle stage. During monitoring, the follicular size and firmness, uterus (edema and tone), cervical tone and the possible presence of reproductive pathologies were assessed. This monitoring was crucial to estimate the moment of ovulation and breeding (Munroe 2011, pp.242-380).
In each mare, to carry out the insemination, either with fresh or frozen semen, compliance with all the following criteria was required: follicular growth ≥ 35 mm; level 3 edema (i.e. heat edema); open and relaxed cervix; and soft uterine tone. Uterine edema was graded on a scale of 0 (no edema) to 4 (maximal) as described by William (2004, pp. 119-120).
The ovulation was induced with human chorionic gonadotropin (Vetecor™, Hertape Calier; Brazil; 0.5 mL; IV) and deslorelin (Sincrorrelin, Ourofino; Brazil; 1 mL; IM). These inducing agents helped to reduce the uterine edema with the approach of the ovulation time and to minimize the number of inseminations per cycle (Samper 2001, pp. 219-228).

StallionS, SeMen proceSSing and artificial inSeMination
AIs were performed with fresh or frozen semen of forty eight stallions always from the same breed of the donor mare. Their age ranged from two and a half to seventeen years old. Some of stallions were used more than others (Figure 1). Semen from all stallions was evaluated at the beginning of each breeding season for progressive motility, vigor, concentration and anomalies. The collection of semen from stallions that were resident at CER or at near locations was done by using an artificial vagina. During collection, the ejaculate was divided into the sperm-rich and the gel fractions, before evaluation of the color, volume, concentration and progressive motility. Semen extender (BotuSemen™; Botupharma, Brazil) was prepared before the semen collection and kept at 37ºC until dilution to obtain a final concentration of at least 500 x 10 6 total viable sper- matozoa but normally 800 x 10 6 to 1 x 10 9 total viable spermatozoa were used. The insemination volume was 20 mL for fresh semen.
For cryopreservation purposes, the collected sperm rich fraction of the ejaculate was immediately diluted with the BotuSemen™ extender (1:1, V:V), evaluated and centrifuged (600 g for 10 minutes). Then the pellet was extended with Botu-Crio® (Botupharma, Brazil) in a final concentration of 200 x 10 6 viable spermatozoa per mL. Semen was packaged in 0.5 mL straws, equilibrated for 20 min at 5ºC, and placed in liquid nitrogen vapors (-30ºC) for 20 minutes. Then the semen straws were immersed and stored in liquid nitrogen.
After the detection of ovulation, some mares (n=20) were inseminated with frozen/thawed semen in the uterotubal junction/uterine horn using a standard AI catheter up to six hours post-ovulation. This increased the number of spermatozoa in oviduct and thus the pregnancy rates (Samper 2001, pp. 219-228). Frozen semen straws (100 x 10 6 viable spermatozoa each, 0.5 mL), were thawed in water at 46ºC during twenty seconds, dried and then spermatozoa quality was subjectively evaluated using a microscope (Tim -107; Opton Microscopes, Brazil).
The number of straws per AI was dependent on the post-thaw progressive motility. Usually, the insemination dose for frozen semen was 4 mL (8 x 0.5 mL straws containing minimum total 100 x 10 6 each, with 40% motility). On the morning after AI, if the uterus had over 15 mm of fluid it was flushed with sterile lactated Ringer´s solution (normally two liters, JPFarma, Brazil) until the recovered flush was clear. Oxytocin (Ocitocina Forte, UCB; IV) twice a day during three days was also administered.
In case of fresh semen, the ovulation occurrence was confirmed by ultrasonography after the AI procedure. If the presence of uterine fluid was detected, the previously mentioned post-breeding therapies were also carried out. Inseminations were performed approximately 24h after induction of ovulation. If ovulation did not occur after hCG administration, mares were inseminated again 48h after the first AI.

eMbryo recovery and tranSfer
The day of embryo collection was chosen according to the experimental protocol of the study and varied between day 7 and day 10 after ovulation (day 0 = day of ovulation). Usually only one attempt for embryo recovery was performed per cycle. In this period, donor mares were led to a stock, where the rectum was evacuated of feces, the tail was wrapped and held to one side, and the perineum was washed with povidone iodine, rinsed and accurately dried (Panzani et al. 2014, pp. 807-814).
A total of 156 flushing procedures were made via a Bivona foley catheter inserted into the uterus through the cervix. The cuff was inflated with 60 mL of air and a warmed sterile ringer lactate solution (3-6L per embryo recovery attempt) was infused. The fluid was recovered into a sterile filter cup and then transferred to a Petri dish (Squires et al. 2003, pp.151-170); Aurich 2015, pp.299-303). Then the embryonic stage was identified under a stereo microscope (Tim -107; Opton Microscopes, Brazil) being the majority of the recovered embryos at the blastocyst or expanded blastocyst stages. These embryos were washed 5 times with BotuEmbryo™ (Botupharma; Brazil) to remove cellular debris. Embryos were transferred non-surgically into synchronized recipient mares. The selection criteria for embryo transfer were based on their developmental stage, quality score, and size. As most of the embryos were of intermediate size, 0.5 mL straws were used to pack and transfer, along with a sterilized pipette (side opening) and cannula (Youngquist & Threlfael 2007).
All embryos were transferred in fresh and placed in the tip of the uterine horns ipsi-or contra-lateral to the ovary with the corpus luteum of the recipient mare. The selection of the recipients was based on the number of the days after ovulation, uterine and cervical tone (excellent, good plus or good) and absence of liquid in the uterine horn (Lopes et al. 2011, pp.261-267). Once the procedure for embryo transfer was completed, a dose of 6 mL (IM) of progesterone (P4) was administrated to each recipient mare.
Immediately after embryo recovery, 2 mL (IM) of prostaglandin (PGF2α) (Pfizer, USA) were administered to each donor mare. This administration of PGF2α could be delayed (e.g. to the end of the afternoon or to the next morning) if gastrointestinal disorders (e.g. colic) occurred to avoid the PGF2α side effects including transient decrease of body temperature and sweating. Less often increased respiratory and heart rates, ataxia, abdominal pain and lying down occurred (Equimed®, USA).

pregnancy diagnoSiS
After embryo transfer was performed, the recipient mare returned to the group of the remaining mares to minimize any situation of stress. Five or six days later, pregnancy was confirmed by detecting an embryonic vesicle using rectal palpation and ultrasonography. Cases of dubious diagnosis included uterine tone examined by rectal palpation compatible with a positive pregnancy but with failure to identify embryonic vesicle through ultrasonography. In such cases the ultraso-nography examination was repeated a couple of days later to confirm the pregnancy or possible occurrence of an early embryonic death.
Prior to pregnancy confirmation through a positive heart beat (fourteen days after), the mares stayed with the rest of the group and afterwards they were placed in the pregnant lot. Progesterone (6 mL, IM) was administered weekly until accomplishment of one hundred and twenty days of gestation.

StatiStical analySiS
The breeding records of the mares were analyzed from July up to December of 2015. The available information was statistically analyzed with the objective of evaluating the effects of environmental conditions; the location and breed of donor mare's; stallion´s breed; type of semen (fresh vs. frozen); the month of embryo collection; interval between AI and ovulation; day of uterine flushing; day of embryo transfer; reproductive stage of recipient mare and day of pregnancy diagnosis. Initially the frequency of the studied factors was estimated by Microsoft Excel 2010 (Microsoft Office 2010™; Microsoft Corporation, USA) and the PROC FREQ of programme Statistical Analysis System Institute (SAS International™, Heidelberg, Germany). The results of the pregnancy diagnosis and uterine flushing were analyzed by a logistic regression using the PROC LOGISTIC with a model including the effects that were mentioned before. A P-value ≤ 0.05 was considered significant.

RESULTS
One hundred and fifty six uterine flushing procedures were performed from July to December of 2015 which enabled the recovery of 88 embryos (overall embryo recovery rate of 56.7%). The number of recovered embryos per breed was: Thoroughbred Arabian, n=37; Thoroughbred Lusitano, n=10 and Quarter Horse n=41. No differences (P>0.05) were observed between the results of the headquarters (CER) and the subsidiary center (Haras Raphaela). The effect of the day in which the uterine flush was performed on embryo recovery rate (ERR) is depicted in Table I.
The majority of the embryos (98.7%) was recovered at the day 8 or 9 post ovulation. Only one embryo was recovered at the day 7. Moreover, increased embryo recovery rates were obtained at day 8 (P=0.008). Significant differences between the months in which embryo recovery was carried out were identified (P=0.04, Table I  No differences were identified for the effect of the interval between ovulation and AI on ERR (P>0.05) as the majority of the mares ovulated at the same day or within an interval of twenty-four hours.
The distribution of embryo recoveries according to the breed of the donor mares was represented in Figure 4. The Quarter horse, breed presented better results (ERR=70.7%, P=0.05) but also had the highest percentage of fresh semen AI (98.3%) while Thoroughbred Arabian and Lusitano presented only 46.7% and 52.6% positive recoveries, respectively. So although differences were identified in ERR among breeds, those differences could be related to the type of semen used. performed mostly at day 8. Lower results of embryo recoveries performed at day 9 were obtained. Indeed the range of 8-9 days of age, generally adopted in the present program, can be justified by the fact that at that age, the embryos are easily visualized in the filter system, which speeds up the process of collection, identification, evaluation and transfer (Fleury & Alvarenga 1999, p.261). Thus it minimizes the risk for the people involved, the stress and the risk for the donor by reducing the time and intensity of transrectal manipulation of the uterus as well as the cost of material and time of exposure of the embryo to unfavorable conditions (Lopes et al. 2011, pp. 261-267). Although no flushings for embryo recovery were performed at day 6, a previous study (Jacob et al. 2012(Jacob et al. , pp.1159(Jacob et al. -1166 reported a lower EER compared to days 7, 8, 9, and 10. Similar results were reported by Lopes et al. (2011, pp. 261-267) during six breeding seasons, however the achieved ERR (72.8%) was higher compared to our results. Their ERR was also higher compared to the 64% reported by Aurich et al. (2011, pp. 419-422) and the 64.2% obtained by Youngquist & Threlfael (2007, p. 233) that are near the range achieved herein. On the other hand, Carnevale et al. (2000, pp. 965-979) mentioned that day 7 embryos produced higher pregnancy rates compared to 6, 8 or 9 days old. Similar results were reported by Lopes et al. (2012Lopes et al. ( , pp. 1159Lopes et al. ( -1166. These authors further demonstrated that embryo age should be related to the post-ovulation day of the recipient mare being the pregnancy rates after ET with a degree of synchrony until -4 or -5 days very similar as also showed herein. Camargo et al. (2013, pp. 924-929) also obtained improved results for a synchronization of 0 to 4 days. The above mentioned data demonstrated that the degree of synchrony between donor and recipient mares does not need to be as restricted as previously reported.
An increased age of donors which can affect the embryonic development and recovery or the use of a large number of stallions with different fertilities, which is a very common practice in large scale commercial programs, have been reported to explain the above differences (Fleury et al. 1989, pp. 73-74;Marinone et al. 2015, pp. 53-59). Thus, the use of stallions and donors with good fertility in horse breeding farms and the conduction of programs by the same experienced operators most likely contribute to the similarity among embryo recovery results (Lopes et al. 2011, pp. 261-267). Moreover, Squires et al. (1999, pp. 91-104) argued that the embryonic development and its transport in the oviduct of an old mare may be delayed. Therefore embryonic recovery on days 8 and 9 post-ovulation may be more appropriate for older mares. Their worst results may also be related to the reduction in the quality of oocytes and reproductive parameters that occurs after 12-13 years of age (Vanderwall 2008, pp. 691-702).
The AI with fresh semen resulted in a significantly higher embryo recovery per ovulation compared to frozen semen (61.6% vs 30.0%, respectively). Although a meticulous management of the mare was performed to guarantee good results, in general pregnancy rates decreased when using frozen semen due to the inferior viability of the spermatozoa (Miller 2008, pp. 463-468). This could also be related to the "capacitation-like" In addition, the stallion had no significant effect on the EER despite inducing great variability (11% of the variability can be attributed to the stallion). All the remaining studied factors (stallion breed, moment of AI, number of flushing/mare, oxytocin treatment) did not have a significant influence in the embryo recovery programme (P>0.05).
The majority of recipient mares was classified as "good" or "good plus" corresponding to 76.8% and 21.4%, respectively. No effect of the post ovulation day of the recipient mares transfer on the pregnancy rates was identified (P>0.05). The degree of synchrony of ovulation between the donor and recipient mares ranged from -4 to 0 and did not influence the pregnancy rate (P>0.05).
Positive pregnancies of 64.4% and 68.7% were obtained five or six days after ET respectively, identified by the presence of an embryonic vesicle through ultrasonography. These results were confirmed by rectal palpation in 92% of the cases. Only four cases were considered doubtful. The overall pregnancy rate was 62.2% (both fresh and frozen semen). This value includes the total pregnancies that were confirmed by rectal palpation and at the same time by ultrasonography. The type of semen had no effect on the pregnancy rate. No differences were identified for donor mare location, although CER had thirty one recipient mares pregnant as compared to fifteen of Haras Raphaela, thus displaying the higher number of embryos that were collected at CER. Likewise no differences were identified among breeds or month of embryo transfer on pregnancy rates (P>0.05). Quarter horse breed had twenty four positive pregnancies (63.2%) followed by Thoroughbred Arabian (18 -60.0%) and Thoroughbred Lusitano (4 -66.7%).

DISCUSSION
In this business, uterine flushes in mares are generally performed between the 7 th and 8 th day post ovulation (Squires 1985, pp.92-95), although a broad range of days may be successfully practiced. Our results were already foreseeable as the procedure was deliberately changes inflicted on the spermatozoa during the process of freezing and thawing responsible for the shorter longevity of cryopreserved sperm (Samper 2001, pp. 219-228). Recently, Govaere et al. (2014, pp. 487-491) proved the superiority of deep horn insemination over uterine body insemination that was reflected by the better pregnancy rates obtained after the former insemination using the same low doses of frozen thawed semen (30.6%). Therefore, the insemination technique is one of the factors that must be considered as crucial to achieve a successful pregnancy or embryo collection results. Nevertheless, although a deep horn insemination was performed in the present study, this technique could not overcome the negative effect of using frozen semen. Therefore it is no surprising that, in the current study, the majority of the mares were inseminated with fresh semen. Previously Squires et al. (Squires 2003, pp. 151-170) also referred that mares inseminated with fresh semen are more likely to produce an embryo compared to those inseminated with either cooled or frozen-thawed semen. Later on, Panzani et al. (2014, pp. 807-814) reported similar results for AI performed with fresh, cooled or even the association of cooled plus frozen semen resulting in significantly more positive flushes compared to the AI with frozen semen. Stallions fertility, insemination doses, insemination method and mares management are some of the factors that must be considered to explain the variability in the reported results. Govaere et al. (2014, pp. 487-491) showed that the dose of frozen/thawed semen and AI protocol have a great impact on pregnancy rates and outcomes. So to achieve good results, if only a low-dose of semen is available, a deep uterine horn insemination can significantly improve the pregnancy results. However, no significant differences were observed in mares that were inseminated with high doses of frozen/thawed semen in the uterine body. Moreover, an inadequate semen manipulation (especially when using low doses of frozen semen) will lead to adverse effects and limitations on semen quality and thus to suboptimal pregnancy rates per cycle. Conversely higher doses due to lower semen quality may be related to the presence of free uterine fluid.
In the mare, a non-pathological response to spermatozoa is normally associated to the AI procedure, which can lead to persistent endometritis, unless it disappears within 24-36 hours (Troedsson 2016, pp. 8-12). Uterine flushing and oxytocin administration are typically used to prevent persistent endometritis. Recently, new therapies emerged to modulate the inflammatory response. Ryan et al. (2014, pp. 1-7) mentioned the use of mesenchymal stem cells and autologous conditioned serum. Both strategies reduced the presence of uterine inflammation and the number of neutrophils at six hours post-breeding by over 50%. The authors reported that mesenchymal stem cells treatment had a better ability to modulate the immune response. However, it was already mentioned (Samper 2001, pp. 219-228) that common therapies should only be implemented in mares really showing signs of uterine inflammation or fluid accumulation.
Nowadays, there is still controversy on whether mares should be bred just prior to or after ovulation with frozen semen. Nevertheless, in either case it is evident that the timing of ovulation is highly desirable in order to reduce the interval between breeding and ovulation (Samper 2001, pp. 219-228). Recently, Avanzi et al. (2015Avanzi et al. ( , pp. 1389Avanzi et al. ( -1393 compare two protocols for equine frozen semen AI using either post-ovulation (the same as in the present study) or fixed-time insemination, and evaluated both pregnancy rates and intrauterine fluid accumulation. The pregnancy results were not influenced by the technique of insemination (41.4% vs 51.7%) but the presence of intrauterine fluid accumulation was higher in the post-ovulation protocol (58.6% vs 34.0%). Although the number of mares inseminated was higher in the study of Loomis & Squires (2005, pp. 480-491), similar rates between mares inseminated twice or once (48.1% vs 47.3%) were observed. They also reported uterine fluid presence in 23% of the mares but still lower than mentioned before. In the present study, seventeen mares had this problem (20.5%), only six were submitted more than once at this procedure and four (23.5%) were inseminated with frozen semen. These differences may be related to the fact that an uniform definition of the "free fluid" terminology does not exist.
The impact of breed on ERR was not often reported in the literature. In this work, apparently the Quarter Horse breed had the best results compared to the other two breeds (Lusitano and Arabian). However, the higher number of Quarter Horse mares included in this ET program that were bred with fresh semen could have contributed to these differences. Panzani et al. (2014, pp.807-814) recognized that the Standard breed and the Quarter Horse mares had significantly higher positive flushes and embryo recovery per cycle and per ovulation rates compared to Show Jumping mares. Nevertheless, such reported differences were also related to the different management of the semen market in Italy. Likewise, it is very common in these ET programs, that mares can be used as embryo donors without neglecting their competitive careers. Recently, Pessoa et al. (2011, pp. 703-705) compared embryo recovery and pregnancy rates between athletic and non-athletic breeding Quarter Horse mares in a tropical warm climate (similar weather conditions as in our research). They mentioned that embryo production from donor mares under well-conditioned, appropriate training, were similar to non-athletic as no differences were observed in EER (76% vs 71%) as well as in the pregnancy rates on days 15 (78% vs 79%) and 40 (69% vs 70%). Some authors (Lopes et al. 2011, pp. 261-267;Aurich et al. 2011, pp. 419-422) mentioned that in mare embryo collections, no differences were verified within breeding seasons but they did not exclude the fact that if a higher number of embryos collections could be analyzed these differences could become significant. Although the studied period did not cover the entire breeding season, in the present study it was observed that on July-August a lower number of embryos were obtained when compared to the rest of the months. This difference might be explained, at least in part, by the environmental circumstances such as the temperature and rainfall conditions along with the number of mares in the program. Mares are seasonal polyestrous breeders. So they have multiple estrous cycles during the breeding season being their ovulatory activity related to the long days (Aurich 2011, pp. 220-228). In the south hemisphere, on July-August the temperature and the length of days begin to increase and in consequence the mares gradually start to cycle while emerging from the anestrus period. From September onwards better results were achieved as their cycles become more regular and thus the chances of obtaining a greater number of embryos per cycle was higher.
In our research, all donor mares were submitted to a PGF2α intramuscular administration after embryo recovery so that four to five days later an estrous cycle could again be monitored. In fact, as mentioned by Goretti et al. (2011Goretti et al. ( , pp. 1170Goretti et al. ( -1174, giving PGF2α to the donor mares 48h after embryo collection helps to reduce the average interovulatory interval by approximately 2.5 days, thus increasing the number of embryos that could be collected during the breeding season. This administration had no deleterious effects on the embryo quality, the interval from PGF2α to ovulation or the pregnancy rate of recipient mares. On the other hand, the study of Cuervo-Arango et al. (2015, pp. 1272-1278 concluded that the interval from PGF -induced luteolysis to ovulation had a significant effect on the posterior pregnancy and EER in the mare. Fertility was reduced as this interval became shorter. Differences in sample size, breed or management of the mares could explain these discrepancies.
In conclusion, the day of embryo collection, the type of semen and month of the year were the most important factors affecting embryo recovery rates and thus the embryo transfer program. Moreover, the specific conditions of this study, working simultaneously on different breeds, with different types of semen and donors management in two different breeding stations, allowed several technical approaches that may be relevant to consider and implement in mare embryo transfer centres.