Managing the dominant follicle in lactating dairy cows

Available online at www.sciencedirect.com Theriogenology 76 (2011) 1568 1582 Advances in Bovine Reproduction and Embryo Technology Managing the dominant follicle in lactating dairy cows M.C. Wiltbank a, *, R. Sartori b, M.M. Herlihy c, J.L.M. Vasconcelos d, A.B. Nascimento a, A.H. Souza a, H. Ayres e, A.P. Cunha a, A. Keskin f, J.N. Guenther a, A. Gumen f a Department of Dairy Science, University of Wisconsin-Madison, Madison, Wisconsin, USA b Department of Animal Science, USP, Piracicaba, São Paulo, Brazil c Animal and Bioscience Research Department, Animal and Grassland Research and Innovation Centre, Teagasc, Moorepark, Ireland d Department of Animal Production, UNESP, Botucatu, São Paulo, Brazil e Department of Animal Reproduction, FMVZ-USP, São Paulo, Brazil f Department of Obstetrics and Gynecology, University of Uludag, Bursa, Turkey Received 25 May 2011; received in revised form 25 July 2011; accepted 9 August 2011 www.theriojournal.com Abstract Reproductive efficiency is not optimal in high-producing dairy cows. Although many aspects of ovarian follicular growth in cows are similar to those observed in heifers, there are numerous specific differences in follicular development that may be linked with changes in reproductive physiology in high-producing lactating dairy cows. These include: 1) reduced circulating estradiol (E2) concentrations near estrus, 2) ovulation of follicles that are larger than the optimal size, 3) increased double ovulation and twinning, and 4) increased incidence of anovulation with a distinctive pattern of follicle growth in anovular dairy cows. The first three changes become more dramatic as milk production increases, although anovulation has not generally been associated with level of milk production. To overcome reproductive inefficiencies in dairy cows, reproductive management programs have been developed to synchronize ovulation and enable the use of timed AI in lactating dairy cows. Effective regulation of the CL, follicles, and hormonal environment during each part of the protocol is critical for optimizing these programs. This review discusses the distinct aspects of follicular development in lactating dairy cows and the methodologies that have been utilized in the past two decades in order to manage the dominant follicle during synchronization of ovulation and timed AI programs. 2011 Published by Elsevier Inc. Keywords: Follicle; Dairy cattle; Ovsynch; Synchronization of ovulation; Timed AI Contents 1. Follicle growth in cattle... 1569 1.1. Hormonal control of follicular waves... 1569 1.2. Four distinctive aspects of follicular growth in lactating dairy cows... 1569 2. Follicle size and fertility in dairy cows... 1571 2.1. Reduced fertility associated with large ovulatory diameters... 1571 2.2. Reduced fertility associated with small ovulatory follicle diameters... 1571 3. Controlling the follicular wave for timed AI programs in dairy cows... 1572 3.1. Synchronization of the follicular wave at the initiation of the synchronization protocol... 1573 3.2. Optimizing the hormonal environment and duration of follicular growth... 1574 3.3. Optimizing the hormonal environment during the preovulatory period... 1575 3.4. Synchronization of ovulation before timed AI... 1576 0093-691X/$ see front matter 2011 Published by Elsevier Inc. doi:10.1016/j.theriogenology.2011.08.012

M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 1569 4. Post-AI treatments that regulate follicle growth... 1577 5. Conclusions... 1578 References... 1578 1. Follicle growth in cattle In monovular species such as cattle, horses, and humans, the growth of ovarian follicles is characterized by periodic waves of follicular development, termed follicular waves [1 4]. The emergence of a follicular wave has been defined as the growth of follicles to more than 4 mm in diameter and is generally characterized retrospectively after the dominant follicle has been selected. When follicles reach approximately 8.5 mm in Bos taurus cattle, a dominant follicle is selected from the cohort of growing follicles, which has been termed follicular deviation [5]. After follicular deviation, there is continued growth of the dominant follicle, while subordinate follicles experience a reduced growth rate and eventually begin to decrease in size. 1.1. Hormonal control of follicular waves * Corresponding author. Tel.: (608) 263-9413; fax: (608) 263-9412. E-mail address: Wiltbank@wisc.edu (M.C. Wiltbank). Recent studies in bovine and horses have provided substantial insight on the role of FSH, LH, and the insulin-like growth factor family in follicle growth and selection of a dominant follicle [2,6 8]. In many species, including rodents [9], sheep [10], pigs [11], dogs [12], and cattle [13], two distinct FSH surges have been described during the periovulatory period. Synthesis and secretion of FSH is regulated by hypothalamic factors, particularly GnRH, and inhibited by ovarian factors, particularly estradiol (E2) and inhibin [14,15]. The E2-induced preovulatory surge of GnRH stimulates the release of both LH and FSH [13]. Preovulatory LH and FSH surges are followed by a secondary FSH surge [14,16] linked with emergence and growth of the first follicular wave of the estrous cycle [1,17,18]. This periovulatory FSH surge is temporally associated with a decline in inhibin concentrations, specifically inhibin A, and a decline in circulating E2 induced by the LH surge [19,20]. Accordingly, treatment with inhibin delays onset of the periovulatory FSH surge in cattle and other species. Around the time of follicular deviation, there is a decrease in circulating FSH concentrations. In addition, follicular deviation has been temporally associated with an increase in the expression of LH receptors in granulosa cells and in LH responsiveness in most [21 23,24], but not all [25] studies. Indeed, treatment of cows with an ovulatory dose of LH will only cause ovulation in follicles that have grown past the point of deviation [24], suggesting that the development of LH responsiveness associated with deviation is required for ovulatory capacity. Furthermore, a reduction in the maximal size of the dominant follicle has been reported in response to reduced circulating LH concentrations by a GnRH-receptor antagonist [26] or chronic treatment with a GnRH-receptor agonist [27]. The reduction in size of the dominant follicle suggests that the selection mechanism was impaired; however, the dynamics of follicular deviation could not be ascertained from these studies, because critical data pertaining to growth patterns and deviation in growth rates between follicles were not reported. Other researchers have analyzed LHCGR splice variants in cattle and found that functional LHCGR mrna variants in granulosa cells only were observed after selection of a dominant follicle [17]. Intriguingly, acquisition of LH receptors in granulosa cells near deviation was found to depend on LH action, as evidenced by a lack of increase in granulosa cell LH receptors in cows given a GnRH antagonist [28]. Thus, most data are consistent with an increasing role for LH in follicle growth after the time of follicular deviation, probably due to acquisition of LH receptors in granulosa cells of the dominant follicle. 1.2. Four distinctive aspects of follicular growth in lactating dairy cows The high-producing lactating dairy cow has a unique physiology related to follicular growth that is relevant to practical reproductive management programs. Follicular waves continue to occur throughout pregnancy, with each wave being preceded by a surge of FSH. The magnitude of the FSH surge increased as pregnancy progressed (0.79 ng/ml in Month 4 vs 1.01 ng/ml in Month 8 of gestation). However, the maximum diameter of the dominant follicle continuously decreased as pregnancy progressed, from 11.1 mm in Month 4 to 8.5 mm in Month 9 of gestation [29]. Emergence of the last follicle detected during pregnancy occurred 21.6 2.4 d before parturition [29], whereas the interval from the last FSH surge to parturition averaged 11.8 3.1 d. An impressive surge of FSH is observed during the first

1570 M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 week post-partum, with maximal FSH concentrations averaging 1.9 ng/ml (more than two-fold greater than the average FSH surge concentrations during pregnancy [0.9 ng/ml]). The interval from parturition to the day of emergence of the first follicular wave postpartum averaged 4.0 0.9 d (range of 2 to7 d; [29]). The interval to first ovulation post-partum in lactating dairy cows varies depending upon which follicular wave will culminate with ovulation. The dominant follicle from the first follicular wave post-partum has three potential outcomes: ovulation, atresia, or becoming a large anovular follicle such as a follicular cyst. For example, Savio et al [30] reported that 74% (14/19) of lactating cows ovulated the dominant follicle of the first follicular wave; whereas 4 of 19 cows (21%) became cystic. In contrast, Butler et al [31] reported that 17 of 55 (31%) cows ovulated the first dominant follicle post-partum, whereas 24 of 55 (44%) cows had atresia of the first dominant follicle associated with low circulating E2 concentrations. The remaining cows developed a large anovular follicle that either became cystic (8/55 15%) or produced high E2 but did not ovulate (6/55 11%). The pulsatile release of LH during the dominance phase of the follicular wave is likely the major driver of growth and E2 production by the dominant follicle. Insufficient LH is probably the key limiting factor that results in insufficient E2 production and subsequent atresia of the dominant follicle from the first follicular wave post-partum [32]. It is still not clear what physiological changes result in large anovular follicles that are capable of producing great amounts of E2 during the first post-partum wave. The lack of a GnRH/LH surge in response to follicular E2 may be the mechanism that underlies the development of these large anovular follicles [33,34]. This lack of E2 positive feedback on the hypothalamus could be caused by the presence of low circulating progesterone (P4) concentrations in early post-partum dairy cows [35]. The loss of body weight during the early post-partum period has been associated with the release of substantial amounts of P4 from fat, which could result in sub-luteal concentrations of P4 and prevent ovulation [36]. Subsequent follicular waves are subjected to the same fates as the first wave post-partum. There are a number of intriguing aspects of reproductive physiology that differ somewhat in lactating dairy cows and may be related to the control of follicular wave dynamics [37]. First, high-producing dairy cows have lower circulating E2 concentrations than would be expected, given the size of their dominant follicles. For example, Lopez et al [38] reported a decrease in the peak E2 concentrations with increasing milk production. Paradoxically, smaller circulating E2 was associated with larger follicular diameters as milk production increased. A decrease in follicular steroidogenesis has not been demonstrated in high-producing cows. Greater E2 metabolism with increasing milk production, however, has been demonstrated [39]. It seems likely that lower circulating E2 in cows with greater follicular volume is primarily related to greater E2 metabolism in high-producing lactating dairy cows [38]. Greater E2 metabolism may be caused by greater blood flow through the liver associated with greater feed intake in lactating cows [40]. There also appears to be decreased expression of estrus in dairy cattle [41], which may reduce the reproductive efficiency of the more productive cows. Timed AI programs allow high- and low-producing dairy cows to be inseminated with similar efficiency. In addition, the time of AI can be optimized in relation to the time of ovulation; whereas, programs based on expression of estrus have more variability in interval from AI to ovulation. A second intriguing aspect of follicular development in lactating dairy cows is that there is an increasing size of the ovulatory follicle with increasing milk production when cows are inseminated to estrus [40]. These differences may be reduced with timed AI programs because GnRH (or another agent) is used to induce ovulation before expression of estrus. When cows are inseminated to estrus there may be a delay in ovulation in the highest-producing cows because of greater E2 metabolism associated with a delay in the attainment of sufficient circulating E2 to induce the preovulatory GnRH/LH surge. Ovulation of larger follicles by highproducing dairy cows may partially explain the reduction in fertility observed in cows inseminated to estrus [42]. A third intriguing aspect of lactating dairy cows is the large incidence of anovulation. Surprisingly, there is normally no relationship detected between level of milk production and percentage of anovular cows. The follicular dynamics and physiology underlying various types of anovulation in cattle have been reviewed [43,44]. The most common type of anovular lactating dairy cow ( 60% of anovular dairy cows) had follicles larger than ovulatory size but smaller than the classically defined cystic size [43,45]. This type of anovular cows probably has hypothalamic resistance to the positive feedback effects of E2 [34,43]. Timed AI programs generally induce ovulation in most anovular cows [44]. Nevertheless, cows that are anovular or have

M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 1571 low P4 at the start of the Ovsynch program (cows either in proestrus, estrus, or metestrus) have lower fertility than ovular cows with elevated P4 at the initiation of the Ovsynch program [45,46]. A fourth intriguing aspect of follicle development in lactating dairy cows is the increased occurrence of double ovulation associated with increased milk production [38]. Cows that select multiple dominant follicles have increased circulating FSH and LH concentrations during the 24 h before follicular deviation compared with cows that select a single dominant follicle [47]. A reduction in circulating E2 does not appear to be the underlying cause of selection of multiple follicles. Circulating E2 is actually greater in cows with two or three dominant follicles than in cows with one dominant follicle before deviation. A reduction in circulating P4 during this period may be at least part of the explanation for the increases in FSH and LH and increased follicular co-dominance. Differences in other aspects of reproductive physiology in lactating dairy cows have also been reported, including number and dynamics of follicular waves, CL volume, circulating P4 concentrations, and reproductive diseases [48]. The reader is referred to the many other reviews on related topics such as follicular management of anovular cows, patterns of follicular waves and fertility, and nutritional effects on follicular development and fertility in dairy cattle [44,49 53]. cows with a device that produced subluteal circulating P4 for 7 d after the regression of the CL. Treated cows ovulated an older and larger follicle than control cows. The ovulation of a persistent follicle did not alter fertilization rate (97%), but it reduced the percentage of embryos that developed to 16-cell or greater (14 vs 86%; [56]). Evaluation of the oocyte from persistent follicles revealed that germinal vesicle breakdown occurred prematurely, before ovulation of the oocyte, and this probably was the cause of reduced fertility [60,61]. Subsequent research has indicated that lactating dairy cows that are bred to estrus may also ovulate larger than normal follicles and that extended duration of follicular dominance is associated with reduced fertility [62]. The increase in size of the ovulatory follicle has been associated with increased milk production [38,40]. 2.2. Reduced fertility associated with small ovulatory follicle diameters Excessively small ovulatory follicles have also been associated with reduced fertility [63 65]. For example, reducing the size of the ovulatory follicle by aspiration of the dominant follicle during the middle of the Ovsynch protocol dramatically reduced fertility of lactating dairy cows [65]. The reduction in fertility was associated with ovulation of a smaller follicle, reduced 2. Follicle size and fertility in dairy cows The effect of the size of the ovulatory follicle on subsequent fertility has been the subject of substantial research in the past two decades, due to accurate measurements of preovulatory diameter of follicles using transrectal ultrasonography. Synchronization programs based on prolonged treatments with progestins caused ovulation of larger follicles and increased circulating E2 concentrations before breeding, which may reduce fertility in cattle (reviewed by [54]). 2.1. Reduced fertility associated with large ovulatory diameters Prolonged growth of follicles under reduced concentrations of P4/progestin results in increased LH pulses, increased E2 concentrations, larger ovulatory diameter, and, subsequently, decreased fertility [55 58]. Researchers have developed models to investigate the reduction in fertility following ovulation of the persistent follicle. For example, Ahmad et al [59] evaluated fertilization and early embryo development in beef Fig. 1. Typical programs designed to synchronize ovulation of the dominant follicle and facilitate success with a timed AI (TAI) protocol. A) Representation of the typical Ovsynch program that begins with GnRH treatment, followed 7 d later by PGF treatment, 56 h later by a second GnRH treatment to synchronize ovulation, and finally TAI at 16 h after the second GnRH. B) Representation of a typical E2/P4-based program which begins with insertion of a P4-releasing device and treatment with E2 (in this case 3 mg E2-benzoate), followed 7 d later by PGF treatment. One day later, cows are given 1 mg of estradiol cypionate (ECP), with TAI 48 h later.

1572 M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 E2 concentrations before AI, reduced P4 concentrations after AI, and increased percentage of cows exhibiting short luteal phases following AI [65]. This is most likely to be a problem during synchronization programs that induce ovulation of a small follicle using GnRH [65,66]. Fertility of cows bred to estrus is generally reduced with increasing ovulatory follicle size, and not with decreased ovulatory size [62,64,67,68]. Thus, optimization of ovulatory follicle size and health has been an important goal for reproductive management programs, particularly those that use synchronization of ovulation. 3. Controlling the follicular wave for timed AI programs in dairy cows Development of effective reproductive management strategies is generally dependent on successful integration of valid information on reproductive physiology with practical management information from commercial dairy farms [69 74]. This has been the basis for the development of timed AI programs that are now an integrated part of reproductive management strategies in many parts of the world. Typical protocols representing the two general types of program for synchronization of ovulation are shown (Fig. 1). Representation of the Ovsynch program [69], using GnRH at the initiation of the program to synchronize a new follicular wave and to assure the presence of a CL during the program, is shown (Fig. 1A). After 7 d, cows are treated with prostaglandin F 2 (PGF) to regress the CL and allow the dominant follicle to proceed toward ovulation. The final hormonal treatment is GnRH, which is given to synchronize the time of ovulation; timed AI is subsequently done in a temporal relationship to this synchronized ovulation [69]. An E2/P4-based protocol is represented in Fig. 1B. In this representative program, cows are treated with E2 benzoate at the start of the program and a P4-releasing device is inserted into the vagina. Controlled-internal drug release (CIDR) inserts are one of the most common P4-releasing devices approved for use in lactating dairy cows. The increase in circulating E2 after E2 benzoate treatment, in the presence of high P4, causes regression of follicles that are present on the ovaries; approximately 3 to 5 d later, there is initiation of a new follicular wave [75]. Cows are treated with PGF to regress any CL that are present and 1 d later, they are given E2 cypionate (ECP) and the CIDR is removed. The slow increase in circulating E2 that occurs after treatment with ECP [76] in the presence of low circulating P4 produces a synchronized ovulation and the cows receive timed AI in a temporal relationship to this synchronized ovulation [77]. A graphical summary of the key aspects of reproductive physiology that are controlled by these two Fig. 2. Schematic representation of five key aspects of reproductive function that synchronized ovulation protocols attempt to optimize. Below each reproductive function are shown typical methods or processes that protocols have attempted to modify during the last decade. See text for details.

M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 1573 types of ovulation synchronization programs is shown (Fig. 2). Furthermore, this figure also depicts key aspects of protocols that have been investigated in research directed at optimizing timed AI programs. Part A(Fig. 2) represents the two most common methods, GnRH and E2, which have been utilized to synchronize a new follicular wave at the start of an ovulation synchronization protocol (discussed in Section 3.1.). Part B represents the period of growth of the follicular wave with the two key research areas, increased P4 and shortened length of protocol, which have been investigated during this period (Section 3.2.). Part C represents the preovulatory period with the decrease in P4 and increase in E2 as key research areas for optimizing protocols (Section 3.3.). Part D represents the two common methods that have been used to synchronize ovulation: E2 treatment to induce estrus, the LH surge, and ovulation; or GnRH treatment to induce the LH surge and synchronized ovulation (Section 3.4.). Part E represents the key research area of optimizing the post-ai environment that is discussed in Section 4 of this manuscript. 3.1. Synchronization of the follicular wave at the initiation of the synchronization protocol The first critical element of timed AI protocols is synchronization of the follicular wave at the initiation of the program. Synchronization of a new follicular wave generally entails removal of the functional dominant follicle, either by physical destruction, hormonally induced ovulation, or hormonal inhibition of gonadotropins so that the follicle regresses [78]. Following aspiration of the dominant follicle, there is a rapid increase in circulating FSH concentrations caused by removal of follicle-derived inhibitors of FSH secretion. This FSH surge is followed by the emergence of a new follicular wave, usually within 1 d. Similarly, induction of ovulation of the dominant follicle either with LH or hcg is followed by an FSH surge and emergence of a new follicular wave. From a practical perspective, removal of the dominant follicle is usually performed by ovulation of the follicle using GnRH treatment [69,79]. Treatment with GnRH induces an LH and FSH surge that peaks within 2 h after treatment, with a return to basal concentrations approximately 4 h after the treatment. If this gonadotropin surge causes ovulation, then there is a surge in FSH that peaks 24 h after the initial GnRH treatment, with coincident emergence of a new follicular wave [78]. Each of these follicular synchronization methods is primarily effective when the dominant follicle is removed, i.e., when cows ovulate in response to GnRH or gonadotropin treatments. Treatment of dairy cows at random stages of the estrous cycle with GnRH results in ovulation in only 50 to 70% of the cows [69,80]. Different stages of the estrous cycle have different ovulatory responses, with the greatest response at days 5 to 9, and a reduced response earlier and later in the estrous cycle [66,68,81]. It is unrealistic, therefore, to expect 100% synchronization of follicular waves by giving GnRH alone at the beginning of a synchronized ovulation program. Although ovulation to the initial GnRH is not necessarily a requirement for conception with these programs, there is generally a reduction in fertility in cows that do not ovulate to the initial GnRH treatment [51,80]. An alternative method for removing the functional dominant follicle is inhibiting the gonadotropin support required for follicle growth. From a practical standpoint, this can be accomplished by treatment with E2 in the presence of high P4 [75,82]. The combination of E2 and P4 decreases LH and FSH; this reduction in gonadotropin support for follicular growth results in follicular atresia by 36 h after E2/P4 treatment [82,83]. For instance, after treating Holstein cows with 2 mg of E2 benzoate (with or without injectable P4) in a CIDR protocol, circulating FSH and LH reach nadir concentrations in approximately 0.5 d, corresponding to the peak of circulating E2, followed by a subsequent FSH surge 3 to 4 d after the E2 treatment [84]. The interval from the hormonal treatment to emergence of the new follicular wave has been shown to be independent of stage of follicular development after E2 treatment [85]. However, it depends on the type [83] and dose [82,86] of E2 ester that is used. The circulating E2 profile in cattle differs according to the esters of E2 that are used [76,83]. Although long half-life estrogens (i.e., ECP) have been successfully used to synchronize the emergence of follicular wave in beef cows and heifers [86], they seem to be less precise when used in lactating dairy cows [87], perhaps because of the high milk production and rapid E2 clearance rates in the liver of high-producing cows. Based on recent data, the interval from treatment to emergence of a new follicular wave is reduced with increasing milk production [77]. Treatment of all cows with 2 mg of E2 benzoate results in follicular wave emergence at 3.8 d in high-producing cows, but not until 4.5 d in low-producing cows. Collectively, E2 P4-based protocols are expected to induce a synchronous follicular wave emergence in 70 to 90% of the cows; most failures in synchronization are related to lack of dominant follicle regression and late wave emergence [77,88]. One of the key benefits of

1574 M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 synchronizing the follicular wave by inhibiting follicle growth rather than by ovulating the dominant follicle is that a new CL is not present during the synchronization protocol, making incomplete luteal regression less likely in an E2 P4 protocol compared with Ovsynchlike protocols [89]. However, a downside of not inducing a new ovulation at the beginning of the synchronization procedures is that circulating P4 could be lower in E2 P4 protocols than needed for optimal oocyte quality in timed AI protocols [46,89,90]. 3.2. Optimizing the hormonal environment and duration of follicular growth Control over the growth of the ovulatory follicle is necessary to achieve optimal ovulatory diameter, thereby resulting in adequate E2 production around AI and the formation of a CL capable of producing sufficient P4 to support embryo development without compromising oocyte maturation. The hormonal environment and duration of growth of this follicle are key components in optimizing fertility in timed AI programs. Hormonal environment and duration of the follicular wave may impact fertility by altering oocyte function, granulosa/theca cell number or function, development of adequate oviductal or uterine environment, and potentially other reproductive functions. Circulating P4 concentrations have a substantial impact on subsequent fertility. A relationship between pre-ai P4 concentration and subsequent fertility was documented almost three decades ago [91]. Using twice-weekly blood samples from 88 Holsteins and 107 Jersey cows, a relationship was found between average P4 concentrations during the 12 d prior to AI and subsequent first service conception rate (approximately 10% increase for every 1 ng/ml increase in average P4). In seasonal dairy herds in New Zealand, cows were synchronized with PGF 13 d apart and supplemented or not with a CIDR for 5 d before the second PGF to increase circulating P4 before AI [92]. The authors found an increase in the percentage of cows that showed estrus after the second PGF (542/605 89.6% vs 474/572 82.9%) and in the proportion of pregnancy per AI (P/AI) in cows inseminated after detected estrus (353/542 65.1% vs 285/474 59.7%) in response to P4 supplementation. The evaluation of the stage of the cycle at the time of the second PGF demonstrated that the CIDR improved fertility in cows in earlier (days 5 9; 52.3 vs 64.8% P/AI) and mid-cycle (days 10 13; 59.3 vs 66.2%), but not later cycle (days 14 19; 71.3 vs 71.4%). Thus, increasing P4 in cows with lower P4 prior to PGF synchronization improves fertility at the subsequent AI. Indeed, many of the presynchronization programs, such as Presynch, may be improving fertility due to the increase in P4 during the pre-ai period. Recent work in our laboratory tested whether P4 concentrations during growth of the preovulatory follicle alter double ovulation rates and fertility in lactating cows (Cunha & Wiltbank, unpublished). Holstein cows (n 624) were presynchronized before the breeding Ovsynch with an Ovsynch72 protocol (GnRH-7d-PGF- 3d-GnRH), but no insemination was performed. Cows then began Ovsynch immediately (the 2 nd GnRH of the Ovsynch72 was the 1 st GnRH of the breeding Ovsynch) after the presynchronized Ovsynch, to result in an ovulatory follicle developing under low progesterone (Low-P4), or cows received the 1 st GnRH of the breeding Ovsynch 7 d later to result in an ovulatory follicle developing under high progesterone (High-P4). Ovarian ultrasonography and blood samples were used to assess ovulation, pregnancy status, and circulating P4 concentrations. As expected, cows in the High-P4 had greater P4 concentrations than cows in the Low-P4 at the first GnRH of the breeding Ovsynch (1.80 vs 0.38 ng/ml) and at the PGF (4.43 vs 2.51 ng/ml). The double ovulation was greater in Low-P4 than High-P4 (20.6 vs 7.0%). A previous study also reported a reduction in double ovulation in cows with increased P4 concentrations prior to AI [90]. Overall P/AI at day 29 was greater in High-P4 than Low-P4 (51.0%, n 292 vs 37.1%, n 272). Surprisingly, pregnancy loss between 29 to 57 d of gestation was also less in High-P4 than Low-P4 (6.8 vs 14.3%). Thus, High-P4 during follicle development dramatically reduces selection of co-dominant follicles and double ovulation. In spite of ovulating fewer follicles and having reduced P4 concentrations after AI, cows treated with the High-P4 protocol had better fertility than those treated with the Low-P4, providing strong evidence for the importance of high P4 during Ovsynch, as demonstrated previously [46]. There are many potential physiological mechanisms that may underlie the effect of high P4 concentrations during follicular growth to reduce double ovulation, increase fertility, and reduce loss of pregnancies in timed AI protocols. For example, increasing circulating P4 decreases LH pulse frequency [93]. In previous studies, low P4 during follicle growth resulted in the development of persistent, low-fertility follicles [56,61,94]. For example, lactating dairy cows had much lower fertility after development of a persistent follicle compared with control dairy cows (44% vs 12%; [56]). In a recent

M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 1575 elegant study [95], embryos were recovered from cows that started the Ovsynch on day 3 or 6 of the estrous cycle. Only 7.1% of cows that started Ovsynch on day 3 ovulated to the first GnRH, whereas 88.6% of cows ovulated to the first GnRH when Ovsynch was started on day 6. Fertilization rate was similar for the two groups (85 vs 86%); however, the percentage of highquality embryos (Grades 1 and 2) were much greater for cows that started Ovsynch on day 6 (83.7%) compared with cows that started on day 3 (47.0%). The period of follicular dominance averaged 8.0 d in cows that started Ovsynch on day 3 compared with 5.8 d for cows that received the protocol beginning on day 6. Thus, a 2 d increase in follicular dominance reduced embryo quality. In our experiment that utilized Double- Ovsynch with high or low P4 concentrations, the age of the follicles were identical in the two groups. However, the preovulatory follicle from the low P4 group would have been exposed to greater number of LH pulses and this could underlie the low fertility. Thus, reduced fertility can occur either in response to the ovulation of persistent follicles or follicles that were overexposed to LH. This idea may be a critical concept for fertility in dairy cows that ovulate spontaneously or in response to synchronized breeding protocols. Reducing the period of follicular dominance has been evaluated as a strategy to improve fertility in cattle. A shortened Ovsynch strategy has been developed to reduce the duration of follicular dominance during the protocol. The interval between GnRH and PGF was reduced from 7 to 5 d along with an increase in the period of proestrous (interval from PGF to the second GnRH) from 48 to 56 h to 72 h improved fertility in beef cows [96]. A similar strategy in dairy cattle produced encouraging results. Santos et al [42] reported an improvement in fertility in response to the 5-d compared with the 7-d timed AI protocol (5-d: 175/462 37.9% vs 7-d: 144/466 30.9%). It is noteworthy that two PGF treatments were required for proper luteal regression in cows assigned to the 5-d protocol. In that particular study, the second GnRH and timed AI were performed at 72 h after the PGF (Cosynch-72) in both the 5- and 7-d programs. This timing of GnRH and AI reduced fertility in the 7-d protocol [97], but does not alter fertility in the 5-d protocol [98] compared with GnRH treatment at 56 h and AI 16 h later. Additional studies are necessary to evaluate differences between these protocols and to optimize the protocols. It seems likely that protocols with reduced duration of follicular dominance, combined with a longer proestrus period, can increase fertility in lactating dairy cows. 3.3. Optimizing the hormonal environment during the preovulatory period The hormonal milieu during the proestrus is critical for reproductive success. First, a lack of complete CL regression associated with slightly elevated P4 concentrations during proestrus reduced fertility in response to timed AI protocols [99]. In our studies, 15% of the cows did not undergo complete CL regression (P4 0.4 ng/ml) following the Double-Ovsynch protocol and had greatly reduced fertility to timed AI [100]. Treatment with a second injection of PGF, administered 24 h after the first injection, improved CL regression. However, the effects on fertility were marginal. A second critical factor is to optimize E2 concentrations before AI. In cows given GnRH followed 7 d later by an injection of PGF and then inseminated following detection of estrus, treatment with 1 mg of ECP at 24 h after PGF, increased E2 concentrations and fertility [101]. To test the effects of E2 concentrations during a timed AI program, cows were treated with 1 mg of E2 at 48 h after PGF of the Ovsynch-56 (GnRH- 7d-PGF-56h-GnRH-16h-timed AI; [99]). As expected, there was an increase in the expression of estrus among cows treated with E2 (80.2 vs 44.4%). However, there was no improvement in P/AI with E2 treatment (Ovsynch: 161/409 39.4% vs Ovsynch E2: 188/443 42.4%). Treatment with E2 improved fertility in low body condition score (BCS 2.5) cows (Ovsynch: 32/114 28.1% vs Ovsynch E2: 50/125 40.0%) such that these cows had similar fertility compared with cows with high BCS (Ovsynch: 125/286 43.7% vs Ovsynch E2: 138/314 43.9%). Therefore, a sufficient E2 surge may be the key rate-limiting step for obtaining high fertility in low BCS cows during timed AI protocols. Treatment with E2 also tended to increase fertility in cows ovulating medium-sized follicles (15 to 19 mm), but not in cows ovulating smaller or larger follicles. This observation is consistent with the concept that increasing circulating E2 may be important for optimizing the preovulatory hormonal environment during timed AI protocols, at least in cows that ovulate follicles at a more desired diameter. Thus, optimizing circulating E2, along with other follicular/luteal functions, may improve fertility in cows inseminated to estrus or in cows that receive timed AI [99,101]. Changes in ovarian steroid concentrations also influence endometrial morphology during the final few days of the Ovsynch protocol [102]. A clear increase in

1576 M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 endometrial thickness was observed at luteal regression. Cows that had delayed luteolysis and, therefore, higher circulating P4 had a thinner endometrium than cows with complete luteolysis. Probably the most interesting finding in this study was the relationship between endometrial thickness and fertility in lactating dairy cows inseminated after a timed AI protocol. An increase in P/AI from less than 20% to more than 40% was observed when endometrial thickness increased from 6 to 10 mm. Fertility remained above 40% for cows above 10 mm in endometrial thickness. Cows with the thinnest endometrium and lowest fertility also ovulated the smallest follicles and had lower circulating E2 concentrations prior to AI. Thus, the reduction in fertility is likely to be due to ovulation of smaller follicles with insufficient circulating E2 near AI, as previously reported in dairy cows [65]. The hypothesis that treatment with E2 during the period before synchronized ovulation of the follicle would increase fertility, particularly in cows with lower endometrial thickness, has also been evaluated. Surprisingly, treatment with E2 did not improve fertility in cows with adequate endometrial thickness, even though there was a dramatic increase in expression of estrus in response to E2 treatment from 47.5% (no E2, endometrial thickness 8 mm) to 82.4% (E2 treated, endometrial thickness 8 mm). However, in cows with reduced endometrial thickness, there was some indication of an increase in fertility after E2, suggesting that some of the lower fertility in this sub-group of cows may be due to inadequate E2 during the preovulatory period [102]. Thus, adequate P4 priming, followed by low circulating P4 concentrations near ovulation, and longer proestrus with sufficient circulating E2 concentrations is probably required to increase endometrial thickness into the optimal range [103]. A more practical method to optimize the development of the preovulatory follicle is to increase the length of the proestrus period. The importance of increasing the proestrus period has been demonstrated in beef cattle treated with the 5-d timed AI protocol [104]. It is likely that changes in the proestrus period may alter fertility by affecting the uterine environment and/or the follicle and oocyte. 3.4. Synchronization of ovulation before timed AI There are many reasons to synchronize ovulation and perform timed AI rather than wait for cows to come in estrus. Firstly, higher-producing lactating dairy cows have reduced estrous behavior than lower-producing cows [41]. Therefore, timed AI programs should result in an increased percentage of high-producing cows receiving AI compared with programs based on detection of estrus. Secondly, in theory, follicle size could be optimized during a synchronized ovulation program and this could improve fertility. Thirdly, timing of AI can be optimized in relation to ovulation (one previously designated time in all cows), which should increase management efficiency and fertility. The first advantage (increased percentage of cows receiving AI) has been reported in direct comparisons of timed AI and programs based solely on detection of estrus. Improvements in fertility, however, have not been a consistent result. A meta-analysis published in 2005 (71 trials in 53 publications with sufficient experimental details for inclusion in the analysis) reported no differences in P/AI between Ovsynch compared with various other reproductive management strategies [105]. There are several methods to synchronize the time of ovulation. Any method that will synchronize estrus will also synchronize the time of ovulation; however, synchronization may not be sufficient to yield good success with timed AI. Most timed AI programs use either GnRH or E2 to increase the synchrony of ovulation during these programs. Treatment with GnRH results in an LH surge that reaches a peak within 2 h and causes ovulation from 24 to 32 h after treatment [69]. Approximately, 10% of cows do not ovulate to the second GnRH treatment of Ovsynch [66,68,80,99,100]. Two primary reasons explain the lack of synchronized ovulation to the Ovsynch protocol. First, cows may come into estrus before the GnRH treatment and therefore ovulate prematurely because of an endogenous GnRH/LH surge [68]. Premature ovulation during the Ovsynch protocol was only found in cows that started Ovsynch in the later estrous cycle ( d 12) and did not ovulate to the first GnRH treatment of Ovsynch [68]. Second, cows may not have a dominant follicle at the time of the second GnRH treatment, due to initiation of a new follicular wave during the Ovsynch protocol. Immature follicular development at the second GnRH of Ovsynch was found in cows that initiated Ovsynch at various stages of the estrous cycle, and in cows that may or may not have ovulated to the first GnRH of Ovsynch [68]. Obviously, cows that do not ovulate to the second GnRH of Ovsynch have little or no chance of becoming pregnant to that timed AI. Cows that ovulate after Ovsynch may not ovulate an optimal-sized follicle. For example, only 358 of 622 (57.6%) cows ovulated a follicle of 14 to 19 mm after Ovsynch [99]; 20% of cows ovulated a follicle that was

M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 1577 too small ( 13 mm) and 22.5% ovulated a follicle that was too large ( 20 mm). Vasconcelos et al [65] used aspiration to synchronize a new follicular wave midway through the Ovsynch protocol so that there was ovulation of smaller ( 11.5 mm) follicles at the second GnRH of Ovsynch compared with non-aspirated controls ( 14.5 mm). It was concluded that cows ovulating smaller follicles had reduced fertility (12 vs 45% for control). The reduction in fertility in cows ovulating small follicles may be due to reduced E2 before AI, ovulation of a less-mature oocyte, and/or reduced P4 after AI (due to ovulation of a smaller follicle with a subsequently smaller CL). Ovulation of larger follicles may, in some cases, result in the ovulation of oocytes with reduced fertility [56,95]. Therefore, improvements in fertility following the Ovsynch protocol are likely to be obtained by increasing the percentage of cows that ovulate an optimal-sized follicle after the final GnRH treatment. Many studies have used various estrogens to synchronize the time of ovulation in dairy cows treated with E2 P4 or GnRH/PGF-based protocols. Estrogens are available for use in synchronization protocols in many countries, but are not available in the USA or Europe. Estrogen products are generally sold at a lower price than GnRH products and therefore can be economically attractive for producers. Estrogens with shorter half-lives (E2 and E2-benzoate) are typically used 24 to 48 h after PGF treatment with AI at about 1 d after E2 treatment. Studies with the Heatsynch protocol [101,106,107] substituted the longer-acting E2 in ECP 24 h after PGF treatment to synchronize the time of ovulation. Although time of ovulation is shorter after GnRH compared with ECP treatment, fertility was similar in Heat-Synch vs Ovsynch protocols. The Heat- Synch protocol improved fertility as compared to cows inseminated to estrus [101]. Interestingly, there was improved fertility in cows that showed estrus during the Heatsynch protocol, although expression of estrus had no effect on fertility in cows receiving AI to Ovsynch [106]. Some protocols have also used ECP at the time of PGF/CIDR removal with similar results as observed with synchronized ovulation after a GnRH treatment [77]. Synchronized ovulation occurs around 70 to 75 h after CIDR removal. Based on data from an earlier study [108] using beef heifers, better fertility is achieved when ECP was given 24 h after CIDR removal rather than at the time of CIDR removal. Thus, there are a variety of different options for synchronizing ovulation with estrogens. Optimal timing of treatments and of AI can vary, based on half-life of the E2, management factors, and the type of protocol used to synchronize follicular/luteal function before the synchronized ovulation. 4. Post-AI treatments that regulate follicle growth Management of follicles after breeding can also be used to regulate reproduction and potentially improve efficiency of reproductive management programs. Programs that include resynchronization of ovulation (Resynch) in cows that did not become pregnant to the first AI can improve reproductive management programs by reducing the interval between successive inseminations. The physiological aspects of Resynch programs are similar to what has been discussed above; however, some management considerations are critical, including an accurate and systematic routine for pregnancy examinations, due to the need to definitively confirm non-pregnancy in cows before treatment with PGF [109 111]. The use of hcg or GnRH after AI has been tested as a follicular/luteal management strategy to improve fertility in lactating dairy cows [112]. Treatment of cattle with these agents at certain times of the cycle can result in ovulation of the dominant follicle. If a follicle is ovulated, there should be increased circulating P4 concentrations due to the presence of an accessory CL. In addition, hcg could increase circulating P4 due to direct effects of hcg to stimulate luteal function; however, this theoretical hcg effect has not been confirmed experimentally in cattle. The increases in circulating P4 after treatment with hcg ond5to7oftheestrous cycle have been shown in many studies. For example, Santos et al [113] treated cows that had been AI to estrus with 3,300 IU of hcg on day 5 after AI. Treatment with hcg increased circulating P4 and increased fertility (Control: 79/203 38.9% vs hcg: 93/203 45.8%). Recently, studies from our group have shown a smaller, but significant effect of treatment with hcg on day 5 after AI on fertility (Control: 566/1519 37.3% vs hcg: 596/1460 40.8%), with significant hcg effects only in first-lactation cows (Nascimento, Souza, Bender, Wiltbank, unpublished). Thus, it appears that hcg treatment on day 5 improves fertility although the physiological reason for this improvement was not defined in these studies. One possibility is that the increase in circulating P4 improves fertility due to P4 effects on the uterus and/or embryo. Several studies have demonstrated increased embryonic development associated with greater concentrations of P4. However, the P4 effect appears to be

1578 M.C. Wiltbank et al. / Theriogenology 76 (2011) 1568 1582 mainly during the early luteal phase (days 5 to 9; [114]). Treatment with hcg on day 5 does not increase circulating P4 until day 8. Perhaps earlier increases in P4 would produce greater improvements in fertility. Another possible effect of hcg treatment is a change in follicular development pattern that could delay luteolysis and thus improve fertility. Ovulation on day 5 results in the emergence of a new follicular wave by day 6 and likely turnover of the dominant follicle of this wave before luteolysis. Thus, the cow is likely to have three rather than two follicular waves. This change in follicular wave patterns may improve fertility [115]. Absence of circulating E2 from a dominant follicle near the time of normal luteolysis will delay luteolysis until there is sufficient E2 to increase PGF secretion from the uterus [116]. Thus, changes in follicular wave patterns and timing of luteolysis may be part of the mechanism for increased fertility after hcg-induced ovulation of the dominant follicle of the first follicular wave. 5. Conclusions Precise regulation of follicular development has been the focus of extensive research during the past few decades. This was possible because of the availability of high-resolution ultrasound technology, allowing evaluation of the dynamic processes of follicular emergence, selection, growth, atresia, and ovulation. Evaluation of the dynamics of follicular development in anovular conditions has provided insights regarding the physiology of reproductive processes [43]. Also, a better understanding of follicular development allows for the determination of how management and nutritional factors may affect fertility [49,51 53,115]. Synchronization protocols have been adjusted to induce the ovulation of an optimized follicle. Initial programs resulted in fertility that was similar to breeding to estrus, but offered the advantage of a timed AI of all cattle on a preselected day [117]. Recent, optimized synchronization strategies may result in better fertility in lactating dairy cows than breeding to estrus. Additional research is needed to evaluate these programs under various commercial and experimental conditions. Perhaps some of these programs may be best for cows in specific physiological circumstances, as exemplified by the better P/AI in primiparous than multiparous cows following the Double-Ovsynch protocol [118]. Clearly, a great deal of progress has been achieved in understanding the physiology of these reproductive programs, with future progress needed to increase consistency, simplicity, and economic value for dairy producers generated by improved reproductive performance with the use of these programs. References [1] Ginther OJ, Wiltbank MC, Fricke PM, Gibbons JR, Kot K. Selection of the dominant follicle in cattle. Biol Reprod 1996; 55:1187 94. [2] Ginther OJ, Beg MA, Bergfelt DR, Donadeu FX, Kot K. Follicle selection in monovular species. Biol Reprod 2001;65: 638 47. [3] Mihm M, Crowe MA, Knight PG, Austin EJ. Follicle wave growth in cattle. Reprod Domest Anim 2002;37:191 200. [4] Fortune JE, Sirois J, Quirk SM. The growth and differentiation of ovarian follicles during the bovine estrous-cycle. Theriogenology 1988;29:95 109. [5] Ginther OJ. Selection of the dominant follicle in cattle and horses. Anim Reprod Sci 2000;60:61 79. [6] Aerts JMJ, Bols PEJ. Ovarian follicular dynamics. A review with emphasis on the bovine species. Part II: Antral development, exogenous influence and future prospects. Reprod Domest Anim 2010;45:180 87. [7] Beg MA, Bergfelt DR, Kot K, Ginther OJ. Follicle selection in cattle: Dynamics of follicular fluid factors during development of follicle dominance. Biol Reprod 2002;66:120 6. [8] Ginther OJ, Bergfelt DR, Kulick LJ, Kot K. Selection of the dominant follicle in cattle: Role of two-way functional coupling between follicle-stimulating hormone and the follicles. Biol Reprod 2000;62:920 7. [9] Ackland JF, Schwartz NB. Changes in serum immunoreactive inhibin and follicle-stimulating-hormone during gonadal development in male and female rats. Biol Reprod 1991;45:295 300. [10] Baird DT, Campbell BK, Mann GE, McNeilly AS. Inhibin and estradiol in the control of FSH secretion in the sheep. J Reprod Fertil 1991:125 38. [11] Kelly P, Duffy P, Roche JF, Boland MP. Superovulation in cattle: Effect of FSH type and method of administration on follicular growth, ovulatory response and endocrine patterns. Anim Reprod Sci 1997;46:1 14. [12] Mondainmonval M, Farstad W, Smith AJ, Roger M, Lahlou N. Relationship between gonadotropins, inhibin and sex steroid secretion during the periovulatory period and the luteal phase in the blue fox (Alopex Lagopus). J Reprod Fertil 1993:47 56. [13] Kulick LJ, Kot K, Wiltbank MC, Ginther OJ. Follicular and hormonal dynamics during the first follicular wave in heifers. Theriogenology 1999;52:913 21. [14] McNeilly AS. The control of FSH secretion. Acta Endocrinol 1988;119:31 40. [15] Bernard DJ, Fortin J, Wang Y, Lamba P. Mechanisms of FSH synthesis: what we know, what we don t, and why you should care. Fertility and Sterility 2010;93:2465 85. [16] Bergfelt DR, Smith CA, Adams GP, Ginther OJ. Surges of FSH during the follicular and early luteal phases of the estrous cycle in heifers. Theriogenology 1997;48:757 68. [17] Turzillo AM, Fortune JE. Suppression of the secondary FSH surge with bovine follicular fluid is associated with delayed ovarian follicular development in heifers. J Reprod Fertil 1990;89:643 53.