Data-derived reference profiles with corepresentation of progesterone, estradiol, LH, and FSH dynamics during the bovine estrous cycle

Theriogenology 79 (2013) 331–343

Contents lists available at SciVerse ScienceDirect

Theriogenology journal homepage: www.theriojournal.com

Data-derived reference profiles with corepresentation of progesterone, estradiol, LH, and FSH dynamics during the bovine estrous cycle O. Martin a, b, *, N.C. Friggens a, b, J. Dupont c, P. Salvetti d, S. Freret c, C. Rame c, S. Elis c, J. Gatien d, C. Disenhaus e, f, F. Blanc g, h a

INRA, UMR0791 Modélisation Systémique Appliquée aux Ruminants, Paris, France AgroParisTech, UMR Modélisation Systémique Appliquée aux Ruminants, Paris, France INRA, UMR085 Physiologie de la Reproduction et des Comportements, Nouzilly, France d UNCEIA, Department R&D, Maisons-Alfort, France e INRA, UMR 348 Physiologie, Environnement et Génétique pour l’Animal et les Systèmes d’Elevage, Saint-Gilles, France f AGROCAMPUS OUEST, UMR1348 Physiologie, Environnement et Génétique pour l’Animal et les Systèmes d’Elevage, Saint-Gilles, France g Clermont Université, VetAgroSup, UMR Herbivores, Clermond-Ferrand, France h INRA,UMR1213 Herbivores, Saint-Gènes-Champanelle, France b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 4 May 2012 Received in revised form 21 September 2012 Accepted 29 September 2012

Subfertility in cows is often associated with alterations in the hormonal patterns involved in the regulation of the estrous cycle. Reference profiles are needed to ground modeling projects aimed at describing these alterations and to develop tools for detecting abnormal dynamics. Various schematic views of LH, FSH, progesterone (P4) and estradiol (E2) patterns have been published but with no clear indication of the extent to which they are derived from real data. The objective of this study was to generate standard profiles for the main reproductive hormones that can be proposed as reliable references to represent the normal dynamics of these hormones over the estrous cycle. A database of hormonal profiles was compiled with 40, 23, 33, and 34 profiles for LH, FSH, E2, and P4, respectively, derived from publications in which changes over time of at least three of these four hormones, including LH, were reported. These profiles were digitalized and standardized over the time throughout the estrous cycle, considering the interval between two successive LH surges to be 21 days. After this standardization on the x-axis, a transformation on the y-axis was performed to center the profiles around their common dynamics. For each hormone, the reference profile was then considered to be the median of the adjusted profiles. Quartiles were reported to account for the time evolution of the variability around each reference profile. The reference profiles obtained showed that the procedure used was satisfactory for extracting the overall changes over time of LH, P4, and E2. Results were less satisfactory for FSH, because of a higher variability observed between the original profiles in our database. The corepresentation of the reference profiles, i.e., when depicted together on the same scale, emphasizes the interplay between these hormones more precisely than most of the schematic views available in literature. These data-derived profiles can be considered to be generic and useful for benchmarking the normal dynamics of gonadotrophins and steroid hormones over the estrous cycle in cow. Ó 2013 Elsevier Inc. All rights reserved.

Keywords: Cattle Estrous LH FSH Progesterone Estradiol

1. Introduction * Corresponding author. Tel.: þ33 (0) 1 44 08 18 12; fax: þ33 (0) 1 44 08 18 53 E-mail address: [email protected] (O. Martin). 0093-691X/$ – see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.theriogenology.2012.09.025

Many studies have reported a decrease in the fertility of dairy cows in the past few decades [1–3] and pointed

332

O. Martin et al. / Theriogenology 79 (2013) 331–343

out that subfertility might involve various components of the reproductive process such as failure to recover postpartum ovarian activity, poor estrous expression, defective oogenesis, decreased embryo survival, and altered endometrium functions. All of these problems are associated with alterations of the hormonal patterns involved in the regulation of the estrous cycle [4,5]. In this context, the question arises as to what is a normal hormonal profile in cows? Numerous studies considered this question mainly with respect to progesterone (P4) patterns that were reported to be indicators of the reproductive status of the cow [6–10]. Data concerning E2, LH, and FSH are much less frequently measured on a sufficient number of cows compared with progesterone data, and they are often studied only during short periods in the cycle [11]. The exceptions are studies dealing with the development of synchronization protocols [12] but, in this case, the profiles have been artificially modified. So far, there have been no data-derived attempts to characterize reference profiles throughout the estrous cycle of reproductive hormones such as E2, LH, and FSH. Using P4 profiles, studies have suggested rules to classify estrous cycles as normal versus altered [8,13]. Usually, these rules rely on the definition of threshold values that concern the length of luteal phase and are often determined by visual assessment [10]. Thresholds retained to classify profiles often differ between studies partly because of biological specificities (e.g., suckling vs. dairy cows) but also because of differences between the hormone assays or the hormone monitoring methods used (milk vs. blood samples, number of measurements per week, chemical analyses, etc.) [14–16]. Further, despite the interdependence between reproductive hormones, equivalent rule sets do not exist for the other reproductive hormones. Thus, it is difficult to compare profiles from different experiments and difficult to fully describe distortions in reproductive hormone profiles. Such a difficulty is emphasized if we consider that variation over time of P4 concentrations influences embryo survival and might explain fertility problems [17]. Consequently, characterizing estrous profiles based only on the length of luteal phases might be not precise enough. Meier et al. [18] and Gorzecka et al. [19] have suggested defining parameters that account for the variation over time (dynamics) of the hormonal concentrations. Consequently, in order to develop benchmarking reference profiles it will be necessary to consider the dynamics of reproductive hormones, i.e., to capture the parameters which describe the variability in the shape of hormones profiles. Thanks to mathematical models, some progress has been made to quantify the rise and decline in P4 concentrations over the luteal phase [20,21] and this offers potential reference parameters for P4 dynamics in milk. However, this has not been extended to the other major reproductive hormones, mainly because of the difficulty in measuring plasma and/ or milk concentrations of gonadotrophins and E2 over the whole cycle and over a large number of animals. Little quantitative information is available for the bovine species specifically addressing the simultaneous time course for changes in gonadotrophins and ovarian hormone

concentrations [22]. To date, only schematic views have been reported to describe the overlapping patterns of the main reproductive hormones [23–25]. However such schematic representations cannot be said to provide valid quantitative reference profiles, especially because they provide no representation of the possible variability in the shape of profiles. This is a major limitation for the development of models that propose to interpret and predict the dynamics of the reproductive cycle based on a representation of the physiological mechanisms involved [26,27]. Calibration of these dynamic models relies on input curves that can be considered as references for nonaltered hormonal dynamics [28]. A representation of the biological variability observed around reference dynamics is needed for prediction. The objective of the present work was to generate dataderived standard profiles for the main hormones involved in the ovarian cycle (LH, FSH, P4, and E2) that can be considered as a reference to represent their codynamic and to evaluate their variability over the estrous cycle. 2. Materials and methods 2.1. Data collection A database was compiled with 22 studies on hormones of the estrous cycle in dairy and beef cows [29–50] published between 1973 and 2010. The articles selected to be included in the database each contained a description through the complete or partial estrous cycle of the time changes of plasma concentrations of LH (showing a preovulatory surge) associated with at least two of the three other main hormones involved in the estrous cycle: FSH, E2, and P4. In the present study, a profile denotes a time series of one hormone over the estrous cycle and a data set denotes a series of at least three hormonal profiles out of LH, FSH, E2, and P4 for the same individual or group mean. When articles were on the influence of a specific hormonal treatment on hormonal profiles [34,35], only the control group was considered. Data were extracted from tables [38,42] or digitized from figures reported in articles using Windig [51]. Some data sets could not be included, for example, when the LH profile did not show a distinct surge [49] or when profiles belonged to short or long cycles [47]. This led to a collection of 40 data sets (from one to four data sets per publication) including 40, 23, 33, and 34 profiles for LH, FSH, E2, and P4 respectively. 2.2. Data standardization Hormonal plasma concentration units were homogenized to ng/mL for LH, FSH, and P4, and to pg/mL for E2 using, when necessary, the following molar weights : LH (C1014H1609N297O294S27), 23530.59 g/mol; FSH (C975H1513N267O304S26), 22673.11 g/mol; P4 (C21H30O2), 314.47 g/mol; and E2 (C18H24O2), 272.39 g/mol. Even though the average cycle duration varies from 21 days to approximately 23 days for Holstein cows [9,52,53], data were standardized to provide information on the dynamic hormonal pattern within a standard estrous cycle

O. Martin et al. / Theriogenology 79 (2013) 331–343

333

Table 1 References included in the database. Reference

Animals

Time period (days) LH peak at t ¼ 0

Breed

Parity

British Friesian Holstein Holstein Gobra Crossbred beef Dairy Crossbred beef Composite beef Holstein Dutch-Friesian Friesan Dairy

P P, M P ? M P, M P M P, M P, M P ?

5.8 9.2 3.0 18.3 1.7 15.0 1.6 4.0 6.7 4.3 0.9 0.5

1 (35)

Holstein

M

1 (?)

Dairy cattle

2 2 2 2 2

Data set Bleach et al., 2001 [48] Bloch et al., 2006 [47] Blödow et al., 1990 [49] Bousquet, 1989 [46] Bryner et al., 1990 [45] Chenault et al., 1975 [44] Cooke et al., 1997 [43] Cupp et al., 1995 [42] Desaulniers et al., 1995 [41] Dieleman et al., 1986 [50] Dobson, 1978 [39] Dobson and Kamonpatana, 1986 [40] Echternkamp and Hansel, 1973 [38] Eilts and Paccamonti, 2004 [37] Evans et al., 1994 [36] Gong et al., 1996 [35] Kaneko et al., 1992 [34] Malhi et al., 2005 [33] Ramirez-Godinez et al., 1982 [32] Schams et al., 1978 [31] Van Eerdenburg, 2010 [30] Wolfenson et al., 2004 [29]

a

2 1 4 1 2 1 2 1 3 3 1 1

(2) (35) (1) (?) (15) (6) (6) (18) (6) (3 to 8) (6) (?)

(3) (7) (5) (6 or 7) (5)

2 (2) 1 (?) 2 (10)

Hormone concentration unit FSH

LH

P4

E2

Estrous cycle

ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL -

ng/mL ng/mL mg/L ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL mg/L ng/mL ng/mL

ng/mL ng/mL nmol/L ng/mL ng/mL ng/mL ng/mL ng/mL ng/mL nmol/L ng/mL

pg/mL pg/mL pmol/L pg/mL pg/mL pg/mL pg/mL pg/mL pg/mL pmol/L pg/mL pg/mL

S S S S N N N, S S S N N N

6.0 to 29.0

-

ng/mL

ng/mL

pg/mL

N

?

7.0 to 28.0

ng/mL

ng/mL

ng/mL

pg/mL

N

Hereford Hereford x Friesan Japanese brown Crossbred Hereford Polled Hereford

P P M P, M M

2.0 1.0 4.6 1.5 4.0

23.1 43.8 7.0 21.0 10.2

ng/mL ng/mL ng/mL ng/mL ng/mL

ng/mL ng/mL ng/mL ng/mL ng/mL

ng/mL ng/mL ng/mL ng/mL ng/mL

pg/mL pg/mL pg/mL -

N S S S N

Braunvieh Dairy Holstein

M ? P, M

15.4 to 41.7 2.8 to 25.3 1.8 to 24.9

ng/mL -

ng/mL mg/L ng/mL

ng/mL nmol/L ng/mL

pmol/mL pg/mL

N N S

to to to to to to to to to to to to

to to to to to

10.6 20.0 21.1 21.3 16.4 23.3 12.7 19.0 8.7 28.9 1.2 25.2

Abbreviations: E2, estradiol; M, multiparous; N, natural estrous cycle; P, primiparous; P4, progesterone; S, estrous cycle following a synchronization treatment (induced luteal regression). a Number of data sets per reference and number of animals per profile in parenthesis; data set is a pool of hormonal profiles for the same individual or group in the study.

of 21 days to be consistent between breeds. The original time t, ranging from 15 days to 44 days, with t ¼ 0 corresponding to the first reported LH peak, was standardized to the standard time t* using the following adjustment at each time point: t* ¼ t21 þ D, where t21 ¼ t , 21/L and L is the estrous cycle length given in the article, derived from the interval between LH peaks in the figures, or otherwise assumed to be equal to 21 days, and D ¼ þ21 if t21 < 0.5, D ¼ 21 if t21  21.5, D ¼ 42 if t21  42.5, and D ¼ 0 otherwise. Segments of the original profile curves were thus adjusted to fall within the range of time t* in the range 0.5 to 21.5. To derive the reference profile of each hormone from the database of profiles that showed variability in both the absolute concentrations and in the amplitude of the profiles, original data Yi (denoting LH, FSH, P4, or E2 for data set i) were transformed to standardized values Yi* using: Yi* ¼ bi þ si , Yi, where bi and si are basal and scaling transformation coefficients, respectively. The couple (bi, si) was estimated for each profile of each data set i and corresponded to the transformation that minimized the deviation of each profile Yi* from the profile YM, defined as the median of the original profiles. This procedure was performed with Scilab [54] and the minimization criteria for each profile was the sum of the squares of the differences between Yi* and YM over the estrous cycle. Because of the shifts between times at which original data were available, the median YM was computed on linearly interpolated original profiles.

2.3. Derivation of reference profiles Using the standardized profiles Yi*, derived from the procedure described above, the reference profile was calculated as the median YM* of Yi*, and the variability of this reference described by the first and third quartiles of Yi*. To represent these reference profiles for each hormone on the same graph, the median YM* of the standardized profiles Yi* was scaled to lie in the range 0.1 to 1 (using: 1– 0.9 , [YM*  M] , [M  m] where M and m are respectively the maximum and minimum value of YM* over the estrous cycle). This set of reference profiles was graphically compared with classic schematic views of the bovine estrous cycle gathered from the literature (see Supplementary Material). These profiles were digitized from figures, standardized to the standard time t* and scaled to lie in the range 0.1 to 1 as described above. Additionally, six data-derived profiles of P4 described in Friggens et al. [7] and Meier et al. [18] were graphically compared with the reference profile of P4 proposed in this work. In these reports, the time t ¼ 0 was the onset of estrus, which approximately corresponded to 24 hours after the time of LH surge defining t* ¼ 0. These profiles were thus shifted by 1 day to be comparable on a consistent time scale. The three profiles of P4 from Friggens et al. [7] correspond to the 25%, 50%, and 75% quartiles of the distribution of milk P4 data representing

334

O. Martin et al. / Theriogenology 79 (2013) 331–343

Fig. 1. Original profiles collected from literature during one estrus cycle of (A) lutenizing hormone (LH, N ¼ 22 profiles), (B) follicle stimulating hormone (FSH, N ¼ 13), (C) progesterone (P4, N ¼ 20), and (D) estradiol (E2, N ¼ 18).

121 lactations from 380 dairy cows. The three profiles of P4 from Meier et al. [18] correspond to the peak-like (profile reaching indistinguishable peak), flat top-like

(profile reaching a plateau), and structured-like profiles (exhibiting a wave-like pattern) identified by clustering of plasma data from 27 Holstein dairy cows.

O. Martin et al. / Theriogenology 79 (2013) 331–343

3. Results and discussion 3.1. Description of the literature database The purpose of this study was to derive reference profiles for the dynamics of P4, E2, LH, and FSH during a normal estrous cycle in cows. Although there exists a number of schematic representations of such profiles [23–25], we wanted our reference profiles to be based on real data, permitting us to obtain at the same time-relevant estimates of the normal variability of such profiles. It was also important that these reference profiles were derived from a reasonably wide range of sources to confer a degree of generalizability to the profiles. The 40 data sets included in the database (Table 1) contained 18 natural cycles (i.e., cycles not preceded by estrous synchronization) and 22 cycles after an estrous that had been synchronized. The data sets represented group mean (N ¼ 36, two to 35 animals) or individual cow (N ¼ 4) hormone profiles. Animals were from dairy (N ¼ 26) or beef (N ¼ 14) breeds, and either primiparous (N ¼ 17), multiparous (N ¼ 14), both (N ¼ 5), or of unknown parity (N ¼ 4). All of the profiles of LH, FSH, P4, and E2 from the original publications are shown in Figure 1 and summary statistics are given in Table 2. The length of the estrous cycle, defined as the interval between two successive LH peaks, ranged from 20.2 to 21.7 in the five publications in which it was reported and was otherwise considered equal to 21 days. As shown in Figure 1, there was a substantial variation in the absolute concentration and amplitude of the profiles. This is not surprising because these data come from different studies and thus different laboratories with all the potential differences in methods and calibration that this implies (see 1. Introduction). This would clearly be a disadvantage if this study was of a comparative nature (e.g., to quantify differences between breeds in reproductive hormone profiles). However, the purpose of this study

Table 2 Quantitative description of the untransformed data and transformation coefficients. LH (ng/mL) Data Number of references 22 Number of data sets 40 Quartiles 25% 0.90 50% 1.47 75% 2.97 a Transformation coefficients Basal (bi) coefficient quartiles 25% 0.06 50% 0.82 75% 1.39 Scaling (si) coefficient quartiles 25% 0.26 50% 0.60 75% 1.69

FSH (ng/mL)

P4 (ng/mL)

E2 (pg/mL)

13 23

20 34

18 33

0.76 12.14 40.78

0.81 2.73 5.81

3.02 5.20 9.27

0.79 0.87 0.91

0.08 0.31 0.53

2.42 3.28 3.89

0.08 0.14 0.22

0.68 0.88 1.24

0.36 0.48 0.65

Abbreviations: E2, estradiol; P4, progesterone. a In the case of FSH, transformation was performed on data expressed in percent of the median of each profile.

335

is to derive reference profiles of these hormones that are broadly representative of the normal cow population. In this context, using data from a wide range of sources is an advantage, contributing to the generalizability of the profiles obtained. 3.2. Data standardization To remove source-specific effects assumed to be largely because of methodological issues, the hormonal concentrations were recalibrated using a procedure to adjust the concentrations for a given profile according to the deviation of that profile from the median profile of the original profiles (Fig. 1; Table 1). The median was used rather than the mean because the median is more robust to outliers and thus was assumed to better represent the profile of the normal population. The resultant standardized profiles are shown in Figure 2 over the standard 21-day estrous cycle. In the case of FSH, original data range from 0.39 to 127.53 ng/mL and are thus shown on a logarithmic scale in Figure 1. Three different subpopulations of original profiles can be seen that differ in orders of magnitude of 1, 10, and 100. These differences of level in FSH plasma concentration are known to be dependent on the assay type chosen and the standard used [55]. Considering this, the method failed to extract a common profile, mainly because the median used to tighten profiles was unrealistic. It was thus necessary to apply a preliminary rescaling of the original FSH data. To keep the approach as simple as possible, original data for each profile were divided by the median of each profile. Afterward, the transformation of FSH was applied on data sets sharing the same median, arbitrarily rescaled to one. The procedure used involved adjustment to both the time and the concentration scales to avoid deformations in the resulting reference profiles that would have occurred by simply averaging the profiles. The comparison of the original profiles with first the standardized profiles and then with the median profiles indicates that our approach was successfully achieved for each hormone (Figs. 1–3). It is important to state that the reference profiles presented in the present study are representative of the overall dynamic of these hormones during one estrous cycle. However, the standardization procedure used does not allow the capture of the temporal variability that is observed within individual profiles of FSH, and to a lesser extent E2 (Fig. 2). It is well established that FSH concentrations are strongly associated with follicular waves [56], whose number per estrous cycle might vary between cows [57]. The variability in the number of follicular waves is clearly apparent in some of the individual profiles of FSH (Fig. 1). Capturing these features would require a far more sophisticated time alignment procedure than the inter-LH peak interval adjustment used in this study. Ideally, it would require data in which not only were the 4 hormones measured but also follicular growth (ultrasound of the ovaries) to pin point the start and end of follicular waves. Alternatively, it would be possible to identify these points from a time-series analysis of the wave patterns observed

336

O. Martin et al. / Theriogenology 79 (2013) 331–343

Fig. 2. Standardized profiles during one estrus cycle of (A) lutenizing hormone (LH, N ¼ 22 profiles), (B) follicle stimulating hormone (FSH, N ¼ 13), (C) progesterone (P4, N ¼ 20), and (D) estradiol (E2, N ¼ 18). Original profiles are adjusted to lie within (0.5–21.5) days from LH peak, standardized for a 21-day interval between LH peaks and standardized around the original median profile (see text for standardization procedure).

in the FSH data. Our preliminary investigation considering such a data analysis method revealed that this was not feasible with the available data.

The 25% and 75% quartiles of the distributions of the basal (bi) and scaling (si) transformation coefficients estimates are given in Table 2. The basal transformation

O. Martin et al. / Theriogenology 79 (2013) 331–343

337

Fig. 3. Reference profiles, for a 21-day interval between LH peaks, duplicated over two successive estrus cycles of (A) lutenizing hormone (LH), (B) follicle stimulating hormone (FSH), (C) progesterone (P4), and (D) estradiol (E2), defined as the median (in bold) of the standardized profiles shown with the 25% and 75% quartiles of the standardized profiles (in gray) (see text for standardization procedure).

coefficients were used to correct for differences in the overall level of original profiles and corresponded to a vertical shift, i.e., an additive value in the unit of the

hormone concentration. The basal transformation effect was very small for FSH, LH, and P4 and mainly concerned E2. The scaling transformation coefficients were used to

338

O. Martin et al. / Theriogenology 79 (2013) 331–343

Fig. 4. Scaled reference profiles for a 21-day interval between LH peaks, duplicated over two successive estrus cycles of lutenizing hormone (LH, thin black curve), follicle stimulating hormone (FSH, bold light gray curve), progesterone (P4, thin gray curve), and estradiol (E2, bold dark gray curve), defined as the median of the standardized profiles scaled to a 0.1 to 1 y-axis (see text for standardization procedure).

correct for differences in the amplitude of variation of original profiles and corresponded to a vertical flattening (si < 1) or stretching (si > 1). Results showed that the transformation coefficients for LH, P4, FSH, and E2 are quite distinct. Values observed for E2 and FSH are notably lower than 1 (0.48 and 0.14 respectively) and evoked a moderate and an important flattening of original profiles. This result is related to the heterogeneity in the dynamics of the original profiles for these two hormones, and is consistent with the diversity of the number of follicular waves. In the case of LH and P4, estimates of the scaling transformation coefficient are distributed around 1, meaning that solutions were found to tighten the original profiles around a common dynamic, either by flattening or stretching the original profiles. These results allow us to qualify the variability in the hormone dynamics that appeared to relate mainly to: the amplitude for P4 and LH, the shape itself for FSH, and both amplitude and shape for E2. This is consistent with the functional lifespan of the associated reproductive structures giving rise to these dynamics. The shape of FSH is indirectly linked to the follicular dynamics, via the paracrine feedback of E2 and inhibin on the anterior pituitary gland. In turn, FSH modulates the rhythm of follicular waves. The FSH dynamic is thus a highly distortable signal. The LH dynamic is a quasi-stationary signal interrupted by tremendous peaks, that shape most of the pattern. The amplitude and shape for E2 are directly linked to the number and stages of follicles and the number of follicular waves, making the E2 dynamics

rather irregular. Though the P4 dynamic is directly linked to the regular development of the corpus luteum, its amplitude is directly linked to the systemic clearance which can vary considerably from one individual to another. In this work, the unit considered for the adjustment procedure was the profile and not the data set or the study. All profiles were thus considered as being members of a common collection of curves, regardless of their source. Because the focus of this study was to capture a common dynamic pattern for each hormone, our procedure did not preserve the variability between profiles. Thus, the intrastudy variability that we assumed to be mainly because of biological factors, and the interstudy variability that we assumed to be largely because of differences in experimental designs and measurement methods, were voluntary removed from our reference profiles to preserve only the variability pertaining to the dynamics. Despite a sizeable body of literature that describes the effects of factors such as energy balance [58–60], milk yield [9,61,62], metritis [63], or endometritis [64–66] on reproductive performance, few studies reported their effects on the reproductive hormone profiles. With respect to P4, metritis was shown to modify the shape of the profile, with a tendency toward lower P4, and flatter luteal phase profiles [19]. However, the effect of metritis only accounted for 10% of the observed variability in profile shapes [19]. In that study, breed and parity had no significant effect on the shape of the P4 profile during estrous cycles although the length of luteal phases was shown to be affected by breed [9,19].

O. Martin et al. / Theriogenology 79 (2013) 331–343

339

Fig. 5. Reference profiles of progesterone (P4) on a 0.1 to 1 scale, for a 21-day interval between LH peaks, duplicated over two successive estrus cycles (bold line with 25%–75% interquartile area in gray) shown together with the three P4 profiles (thin gray lines) redrawn from (A) Friggens et al. (2008) [7] and (B) Meier et al. (2009) [18] (see text for procedure of reference derivation and literature references).

3.3. Reference profiles The median and quartiles of the standardized profiles are presented in Figure 3, in which the 21-day dynamics are duplicated over two successive cycles to highlight the transition around the LH surge. These profiles show that the standardization procedure was satisfactory for synthesizing the overall dynamic of these four hormones during one estrous cycle. Several key time points (in days from LH peak) can be extracted from these reference profiles. The nadir of LH occurs on day 12 (t* ¼ 12.5) and the LH surge (arbitrary defined here as LH > 0.2; scale, 0.1–1) occurs during days 20 to 22 (t* in the range 20.1–21.6). The FSH surge globally occurs during days 20 to 23 and is characterized by a first peak on day 21 (t* ¼ 21) with an exact synchronization with the LH peak, followed by a second peak on day 22 (t* ¼ 21.7). A second peak of FSH is not always reported in the literature, probably because of infrequent sampling. It has been well established that the synchronous surges of LH

and FSH are involved in ovulation and development of the corpus luteum [67], but understanding is still lacking about the function of the second peak of FSH. In ewes and cows it was described to occur 4 to 24 hours after the LH peak, i.e., around the time of ovulation [68–70], but the cause of this peak has not been determined. The nadir of P4 level and the onset of the rise in P4 occur on Day 1 (t* ¼ 0.8). The maximal value of P4 is reached on Day 14 (t* ¼ 14.0) and P4 is higher than 90% of the maximum value from Day 11 to 16 (t* in the range 11.4–16.4). The decrease in P4 pertaining to luteal regression starts after Day 14 and is established on Day 17. The profile of E2 clearly shows a small surge between Days 2 to 7 with a peak on Day 5 (t* ¼ 5.1). The nadir of E2 is thereafter reached on Day 11 (t* ¼ 10.7) and the preovulatory E2 surge occurs globally from Day 17 to 22 with a peak value on Day 21 (t* ¼ 20.7). The interval between peaks of E2 and LH thus equals 8 to 9 hours, which is consistent with the result reported by Walters et al. [68] who observed the highest E2 concentrations 6 to 8 hours before the LH surge.

340

O. Martin et al. / Theriogenology 79 (2013) 331–343

Fig. 6. Reference profiles, for a 21-day interval between LH peaks, duplicated over two successive estrus cycles of (A) lutenizing hormone (LH), (B) follicle stimulating hormone (FSH), (C) progesterone (P4), and (D) estradiol (E2) shown together with literature schematic profiles on a 0.1 to 1 y-axis (see text for procedure of reference derivation and literature references).

These are, to our knowledge, the first reference profiles for these hormones in cattle that have been calculated from a broad range of observed data in which each data source

provided profiles of at least three of the four hormones in question. The fact that each study measured at least three of the hormones allows these reference profiles to be used

O. Martin et al. / Theriogenology 79 (2013) 331–343

not only independently but also to reference the correlated changes in these four hormones. The medians of the standardized profiles scaled to 0.1 to 1 for LH, FSH, P4, and E2 are presented together in Figure 4 (duplication of the 21day dynamics over two successive cycles). It is interesting to note that the pronounced increase in E2 occurs during the decline from peak of P4 and that this is followed by the increase in FSH and LH. Plasma E2 concentrations are elevated before the dual surge of LH and FSH, but the concentration of this steroid has declined before the second surge of FSH. There is no doubt today that the preovulatory LH surge is related to previous E2 secretion in domestic animals. An increase in plasma E2 precedes LH surge in natural conditions. Accepting that these profiles can be considered as reference profiles, at least until larger data sets that have measured these four reproductive hormones become available, provides a useful framework against which to compare particular profiles. In this context, it is clear that there is a need to be able to quantify the effects of potential perturbing factors on reproductive physiology, and this requires a reference for normal estrous cycles. For example, Royal et al. [1,71,72] examined the effects of genetic selection for milk production, and the associated performance changes, on fertility as indexed by changes in the types of P4 profile relative to a base population measured in the 1970s. Because of recent technological advances, P4 profiles can now be routinely measured on commercial farms [73], and thus offer the opportunity to carry out herd level diagnosis of reproductive performance and management. In order to benchmark the profiles of a particular herd there is a need for a reference profile. This study provides a first basis for developing such benchmarks. However, it should be noted that the present profiles reflect the particular conditions of the studies from which they came and as such should not be assumed to represent all conditions and all populations. Even though the references profiles we obtained from the standardization procedure appear quite relevant from an expert point of view, the question arises of validation. This is not an easy task because the number of articles reporting multiple profiles is limited. The consequence is that we could not reserve a subset of data to provide independent validation. Although we could have carried out cross-validation, this is only a limited form of validation especially because of the inherent scale differences between individual studies. It is, however, possible to compare single hormonal profiles, and in particular P4 profiles. A number of studies have analyzed P4 profiles from measurements in milk, on relatively large numbers of animals [1,7,74]. In the study of Friggens et al. [7], the average profile of 120 estrous cycles that preceded successful insemination was presented. Figure 5 shows this profile, derived from milk P4 measurements, plotted on the same scale as the median adjusted P4 profile obtained from the present study (based on plasma measures). A very close agreement between the two can be observed. In Figure 5, the peak-like, flat top-like, and structured-like P4 profiles identified by clustering of plasma data from 27 HolsteinFriesian dairy cows by Meier et al. [18] are also presented. The variability between these shape profiles is close to the

341

interquartile variability around our reference profile. This provides validation, for P4, that the median profile in this study can be considered as a useful reference curve, for instance to calibrate or validate models. With respect to the other reproductive hormones, no reliable, independent, data-derived profiles were found in the literature we have analyzed, so that the only comparison we made is with classic schematic views of the bovine estrous cycle gathered from the literature (see Supplementary Material). These schematic profiles (Fig. 6) appear to be quite different when they are plotted on a same figure with similar level (y-axis) and time scales. Such variability reveals that when drawing a representation of reproductive hormone profiles, authors might have different views to account for the dynamics and simplify them. The statistical comparison between such schematic views and our reference profiles is not possible because there would be no way of ascertaining whether the difference was significant. To use reference profiles to illustrate the dynamics of reproductive hormones over the bovine estrous cycle, we consider that one would be forced to conclude that the data-driven profile is the most valid. There is today a need to be able to predict how the reference profile will be deformed or altered according to the known factors that can perturb reproductive performance [75]. This is necessary if the farm advisor wishes to interpret any discrepancies between an observed and a benchmark profile. Being able to predict quantitative changes in the dynamics of reproductive hormones according to various perturbing factors implies developing simulation models of reproductive physiology. This need has been first recognized for humans [76–78] and now for farm animals [27,79]. Initiatives to do this are now under way [26,28]. However, validating and extending such models requires robust reference profiles. For example, Selgrade et al. [78] clearly showed how the data sets used for estimating the parameters of their model for hormonal control of the menstrual cycle might influence the ability of the model to simulate normal and altered cycles in women. With this study, we believe we have taken a first step toward providing such reference profiles for cows. 4. Conclusion The data-derived reference profiles presented in this study give a coherent representation of the main reproductive hormone dynamics when compared with most of the schematic views available in literature and to P4 median profiles or models resulting from the analysis of large data sets. These data-derived profiles can then be considered to be generic and useful for the development of benchmarking procedures of the normal dynamics of gonadotrophins and steroids hormones over the estrous cycle in cow. Furthermore, they provide reference profiles to be used in modeling approaches. Acknowledgments The authors thank the division of Animal Physiology and Livestock Systems of INRA for supporting this study.

342

O. Martin et al. / Theriogenology 79 (2013) 331–343

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. theriogenology.2012.09.025. References [1] Royal MD, Darwash AO, Flint APF, Webb R, Woolliams JA, Lamming GE. Declining fertility in dairy cattle: changes in traditional and endocrine parameters of fertility. Anim Sci 2000;70:487–501. [2] Lucy MC. Reproductive loss in high-producing dairy cattle: where will it end? J Dairy Sci 2001;84:1277–93. [3] Dobson H, Walker SL, Morris MJ, Routly JE, Smith RF. Why is it getting more difficult to successfully artificially inseminate dairy cows? Animal 2008;2:1104–11. [4] Kesler DJ, Garverick HA, Bierschwal CJ, Elmore RG, Youngquist RS. Reproductive hormones associated with normal and abnormal changes in ovarian follicles in postpartum dairy cows. J Dairy Sci 1979;62:1290–6. [5] Båge R, Gustafsson H, Larsson B, Forsberg M, Rodríguez-Martínez H. Repeat breeding in dairy heifers: follicular dynamics and estrous cycle characteristics in relation to sexual hormone patterns. Theriogenology 2002;57:2257–69. [6] Lamming GE, Darwash AO. The use of milk progesterone profiles to characterise components of subfertility in milked dairy cows. Anim Reprod Sci 1998;52:175–90. [7] Friggens NC, Bjerring M, Ridder C, Højsgaard S, Larsen T. Improved detection of reproductive status in dairy cows using milk progesterone measurements. Reprod Domest Anim 2008;43(Suppl 2): 113–21. [8] Cutullic E, Delaby L, Gallard Y, Disenhaus C. Dairy cows’ reproductive response to feeding level differs according to the reproductive stage and the breed. Animal 2011;5:731–40. [9] Disenhaus C, Cutullic E, Blanc F, Agabriel J. Breed comparison of post partum ovarian activity in cows. J Dairy Sci 2009;92(Suppl 1):498. [10] Mann GE, Keatinge R, Hunter M, Hedley BA, Lamming GE. The use of milk progesterone to monitor reproductive function in beef suckler cows. Anim Reprod Sci 2005;88:169–77. [11] Ginther O, Bergfelt D, Kulick L, Kot K. Pulsatility of systemic FSH and LH concentrations during follicular-wave development in cattle. Theriogenology 1998;50:507–19. [12] Stevenson JS. Progesterone, follicular, and estrual responses to progesterone-based estrus and ovulation synchronization protocols at five stages of the estrous cycle. J Dairy Sci 2008;91:4640–50. [13] Bulman DC, Lamming GE. Milk progesterone levels in relation to conception, repeat breeding and factors influencing acyclicity in dairy cows. J Reprod Fertil 1978;54:447–58. [14] Nebel RL, Whittier WD, Cassell BG, Britt JH. Comparison of on-farm laboratory milk progesterone assays for identifying errors in detection of estrus and diagnosis of pregnancy. J Dairy Sci 1987;70: 1471–6. [15] Colazo MG, Ambrose DJ, Kastelic JP, Small JA. Comparison of 2 enzyme immunoassays and a radioimmunoassay for measurement of progesterone concentrations in bovine plasma, skim milk, and whole milk. Can J Vet Res 2008;72:32–6. [16] Posthuma-Trumpie GA, van Amerongen A, Korf J, van Berkel WJH. Perspectives for on-site monitoring of progesterone. Trends Biotechnol 2009;27:652–60. [17] Mann GE, Payne JH, Lamming GE. Hormonal regulation of oxytocininduced prostaglandin F2alpha secretion by the bovine and ovine uterus in vivo. Domest Anim Endocrin 2001;21:127–41. [18] Meier S, Roche JR, Kolver ES, Boston RC. A compartmental model describing changes in progesterone concentrations during the oestrous cycle. J Dairy Res 2009;76:249–56. [19] Gorzecka J, Callesen H, Pedersen KM, Friggens NC. The relationship between postpartum vaginal discharge symptoms and progesterone profile characteristics in lactating dairy cows in Denmark. Theriogenology 2011;75:1016–28. [20] Friggens NC, Chagunda MGG. Prediction of the reproductive status of cattle on the basis of milk progesterone measures: model description. Theriogenology 2005;64:155–90. [21] Meier S, Roche JR, Kolver ES, Verkerk GA, Boston RC. Comparing subpopulations of plasma progesterone using cluster analyses. J Dairy Sci 2009;92:1460–8. [22] Bergfeld EG, Kojima FN, Cupp AS, Wehrman ME, Peters KE, Mariscal V, et al. Changing dose of progesterone results in sudden

[23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39] [40]

[41]

[42]

[43]

changes in frequency of luteinizing hormone pulses and secretion of 17 beta-estradiol in bovine females. Biol Reprod 1996;54:546–53. Peters AR, Ball PJH. Reproduction in cattle. Second Edition. Oxford, Blackwell Science; 1995. Forde N, Beltman ME, Lonergan P, Diskin M, Roche JF, Crowe MA. Oestrous cycles in Bos taurus cattle. Anim Reprod Sci 2011;124: 163–9. Perry G. The bovine oestrus cycle. South Dakota State University, Cooperative Extension Service, USDA. 2004. Available at: http:// pubstorage.sdstate.edu/AgBio_Publications/articles/FS921A.pdf. Boer HMT, Stötzel C, Röblitz S, Deuflhard P, Veerkamp RF, Woelders H. A simple mathematical model of the bovine estrous cycle: follicle development and endocrine interactions. J Theor Biol 2011;278:20–31. Vetharaniam I, Peterson AJ, McNatty KP, Soboleva TK. Modelling female reproductive function in farmed animals. Anim Reprod Sci 2010;122:164–73. Boer HMT, Röblitz S, Stötzel C, Veerkamp RF, Kemp B, Woelders H. Mechanisms regulating follicle wave patterns in the bovine estrous cycle investigated with a mathematical model. J Dairy Sci 2011;94: 5987–6000. Wolfenson D, Inbar G, Roth Z, Kaim M, Bloch A, Braw-Tal R. Follicular dynamics and concentrations of steroids and gonadotropins in lactating cows and nulliparous heifers. Theriogenology 2004;62: 1042–55. Van Eerdenburg F. Oestrous detection and fertility management. In: Valdmann A, editor. Reproductive challenges facing today’s dairy industry. Estonia, Tartu: Estonian University of Life Sciences Summer School, Estonian Institute of Veterinary Medicine and Animal Science; May 2010. p. 24–8. Schams D, Schallenberger E, Menzer C, Stangl J, Zottmeier K, Hoffmann B, et al. Profiles of LH, FSH and progesterone in postpartum dairy cows and their relationship to the commencement of cyclic functions. Theriogenology 1978;10:453–68. Ramirez-Godinez JA, Kiracofe GH, Schalles RR, Niswender GD. Endocrine patterns in the postpartum beef cow associated with weaning: a comparison of the short and subsequent normal cycles. J Anim Sci 1982;55:153–8. Malhi PS, Adams GP, Singh J. Bovine model for the study of reproductive aging in women: follicular, luteal, and endocrine characteristics. Biol Reprod 2005;73:45–53. Kaneko H, Watanabe G, Taya K, Sasamoto S. Changes in peripheral levels of bioactive and immunoreactive inhibin, estradiol-17 beta, progesterone, luteinizing hormone, and follicle-stimulating hormone associated with follicular development in cows induced to superovulate with equine chorionic gonad. Biol Reprod 1992;47:76–82. Gong JG, Campbell BK, Bramley TA, Gutierrez CG, Peters AR, Webb R. Suppression in the secretion of follicle-stimulating hormone and luteinizing hormone, and ovarian follicle development in heifers continuously infused with a gonadotropin-releasing hormone agonist. Biol Reprod 1996;55:68–74. Evans AC, Adams GP, Rawlings NC. Endocrine and ovarian follicular changes leading up to the first ovulation in prepubertal heifers. J Reprod Fertil 1994;100:187–94. Eilts B, Paccamonti D. The bovine estrous cycle. Comparative Theriogenology, Louisiana State University, Department of Veterinary Clinical Sciences, VMED 5361. 2004. Available at: http://www. vetmed.lsu.edu/eiltslotus/theriogenology-5361/the_bovine_ estrous_cycle.htm. Echternkamp SE, Hansel W. Concurrent changes in bovine plasma hormone levels prior to and during the first postpartum estrous cycle. J Anim Sci 1973;37:1362–70. Dobson H. Plasma gonadotrophins and oestradiol during oestrus in the cow. Reproduction 1978;52:51–3. Dobson H, Kamonpatana M. A review of female cattle reproduction with special reference to a comparison between buffaloes, cows and zebu. J Reprod Fertil 1986;77:1–36. Desaulniers DM, Lussier JG, Goff AK, Bousquet D, Guilbault LA. Follicular development and reproductive endocrinology during a synchronized estrous cycle in heifers and mature cows displaying contrasting superovulatory responses. Domest Anim Endocrin 1995;12:117–31. Cupp AS, Stumpf TT, Kojima FN, Werth LA, Wolfe MW, Roberson MS, et al. Secretion of gonadotrophins change during the luteal phase of the bovine oestrous cycle in the absence of corresponding changes in progesterone or 17ß-oestradiol. Anim Reprod Sci 1995;37:109–19. Cooke DJ, Crowe MA, Roche JF. Circulating FSH isoform patterns during recurrent increases in FSH throughout the oestrous cycle of heifers. J Reprod Fertil 1997;110:339–45.

O. Martin et al. / Theriogenology 79 (2013) 331–343 [44] Chenault JR, Thatcher WW, Kalra PS, Abrams RM, Wilcox CJ. Transitory changes in plasma progestins, estradiol, and luteinizing hormone approaching ovulation in the bovine. J Dairy Sci 1975;58: 709–17. [45] Bryner RW, Garcia-Winder M, Lewis PE, Inskeep EK, Butcher RL. Changes in hormonal profiles during the estrous cycle in old lactating beef cows. Domest Anim Endocrin 1990;7:181–9. [46] Bousquet D. Aspect hormonal du cycle chez la vache. In: Mieux maîtriser la reproduction des espèces domestiques par le transfert d’embryons. Senegal, Dakar: Sommet de la Francophonie, Journées Scientifiques; May 2–11, 1989. [47] Bloch A, Folman Y, Kaim M, Roth Z, Braw-Tal R, Wolfenson D. Endocrine alterations associated with extended time interval between estrus and ovulation in high-yield dairy cows. J Dairy Sci 2006;89:4694–702. [48] Bleach EC, Glencross RG, Feist SA, Groome NP, Knight PG. Plasma inhibin A in heifers: relationship with follicle dynamics, gonadotropins, and steroids during the estrous cycle and after treatment with bovine follicular fluid. Biol Reprod 2001;64:743–52. [49] Blödow G, Rommel P, Kanitz E, Kitzig M, Rodewald S, Schneider F. Investigations on the blood plasma hormone profile in heifers following synchronization of the cycle with PGF2a. Arch Exp Veterinarmed 1990;44:599–610. [50] Dieleman SJ, Bevers MM, Van tol H TM, Willemse AH. Peripheral plasma concentrations of oestradiol, progesterone, cortisol, LH and prolactin during the oestrous cycle in the cow, with emphasis on the peri-oestrous period. Anim Reprod Sci 1986;10:275–92. [51] Lovy D. Windig data digitizer. 1996. Available at: http://www.unige. ch/sciences/chifi/cpb/windig.html. [52] Sartori R, Haughian JM, Shaver RD, Rosa GJM, Wiltbank MC. Comparison of ovarian function and circulating steroids in estrous cycles of Holstein heifers and lactating cows. J Dairy Sci 2004;87:905–20. [53] Sakaguchi M, Sasamoto Y, Suzuki T, Takahashi Y, Yamada Y. Postpartum ovarian follicular dynamics and estrous activity in lactating dairy cows. J Dairy Sci 2004;87:2114–21. [54] Consortium Scilab. Digiteo. Scilab: Logiciel open source gratuit de calcul numérique, (OS, Version 5.3.3). 2011. Available at: http:// www.scilab.org. [55] Rose MP, Gaines Das RE, Balen AH. Definition and measurement of follicle stimulating hormone. Endocr Rev 2000;21:5–22. [56] Schams D, Schallenberger E, Hoffmann B, Karg H. The oestrous cycle of the cow: hormonal parameters and time relationships concerning oestrus, ovulation, and electrical resistance of the vaginal mucus. Acta Endocrinol 1977;86:180–92. [57] Adams GP, Jaiswal R, Singh J, Malhi P. Progress in understanding ovarian follicular dynamics in cattle. Theriogenology 2008;69: 72–80. [58] Beever DE. The impact of controlled nutrition during the dry period on dairy cow health, fertility and performance. Anim Reprod Sci 2006;96:212–26. [59] Garnsworthy PC, Lock A, Mann GE, Sinclair KD, Webb R. Nutrition, metabolism, and fertility in dairy cows: 1. Dietary energy source and ovarian function. J Dairy Sci 2008;91:3814–23. [60] Roche JF. The effect of nutritional management of the dairy cow on reproductive efficiency. Anim Reprod Sci 2006;96:282–96. [61] Buckley F, O’Sullivan K, Mee JF, Evans RD, Dillon P. Relationships among milk yield, body condition, cow weight, and reproduction in spring-calved Holstein-Friesians. J Dairy Sci 2003;86:2308–19.

343

[62] Wathes DC, Fenwick M, Cheng Z, Bourne N, Llewellyn S, Morris DG, et al. Influence of negative energy balance on cyclicity and fertility in the high producing dairy cow. Theriogenology 2007;68(Suppl 1): S232–41. [63] Williams EJ, Sibley K, Miller AN, Lane EA, Fishwick J, Nash DM, et al. The effect of Escherichia coli lipopolysaccharide and tumour necrosis factor alpha on ovarian function. Am J Reprod Immunol 2008;60:462–73. [64] LeBlanc SJ, Duffield TF, Leslie KE, Bateman KG, Keefe GP, Walton JS, et al. Defining and diagnosing postpartum clinical endometritis and its impact on reproductive performance in dairy cows. J Dairy Sci 2002;85:2223–36. [65] McDougall S, Hussein H, Aberdein D, Buckle K, Roche J, Burke C, et al. Relationships between cytology, bacteriology and vaginal discharge scores and reproductive performance in dairy cattle. Theriogenology 2011;76:229–40. [66] McDougall S, Macaulay R, Compton C. Association between endometritis diagnosis using a novel intravaginal device and reproductive performance in dairy cattle. Anim Reprod Sci 2007; 99:9–23. [67] Schwartz NB, McCormack CE. Reproduction: gonadal function and its regulation. Annu Rev Physiol 1972;34:425–72. [68] Walters DL, Schams D, Schallenberger E. Pulsatile secretion of gonadotrophins, ovarian steroids and ovarian oxytocin during the luteal phase of the oestrous cycle in the cow. J Reprod Fertil 1984; 71:479–91. [69] Schams D, Karg H. Radioimmunologic determination of LH in bovine serum, with special attention to the estrous cycle. Acta Endocrinol 1969;61:96–103. [70] Kiddy CA, Odell WA. Radioimmunoassay of blood LH at estrus and ovulation in cattle [Abstract]. J Anim Sci 1969;29:192. [71] Royal MD, Flint APF, Woolliams JA. Genetic and phenotypic relationships among endocrine and traditional fertility traits and production traits in Holstein-Friesian dairy cows. J Dairy Sci 2002; 85:958–67. [72] Royal MD, Pryce JE, Woolliams JA, Flint APF. The genetic relationship between commencement of luteal activity and calving interval, body condition score, production, and linear type traits in HolsteinFriesian dairy cattle. J Dairy Sci 2002;85:3071–80. [73] Friggens NC, Løvendahl P. The potential of on-farm fertility profiles: in-line progesterone and activity measurements. In: Royal MD, Friggens NC, Smith RF, editors. Fertilityindairycows:bridgingthegaps.BSAS,Cambridge,UK:CambridgeUniversity Press; 2008. p. 72–8. [74] Petersson KJ, Berglund B, Strandberg E, Gustafsson H, Flint APF, Woolliams JA, et al. Genetic analysis of postpartum measures of luteal activity in dairy cows. J Dairy Sci 2007;90:427–34. [75] Friggens NC, Disenhaus C, Petit HV. Nutritional sub-fertility in the dairy cow: towards improved reproductive management through a better biological understanding. Animal 2010;4:1197–213. [76] Reinecke I, Deuflhard P. A complex mathematical model of the human menstrual cycle. J Theor Biol 2007;247:303–30. [77] Selgrade JF, Schlosser PM. A model for the production of ovarian hormones during the menstrual cycle. Fields Inst Commun 1999;21: 429–46. [78] Selgrade JF, Harris LA, Pasteur RD. A model for hormonal control of the menstrual cycle: structural consistency but sensitivity with regard to data. J Theor Biol 2009;260:572–80. [79] Blanc F, Martin GB, Bocquier F. Modelling reproduction in farm animals: a review. Reprod Fertil Dev 2001;13:337–53.

343.e1

O. Martin et al. / Theriogenology 79 (2013) 331–343

Online Supplementary Material Schematic views of reproductive hormones profiles (n ¼ 11) redrawn from references cited below Reference of original schematic view Advanced Animal Technology Ltd., Hamilton, New Zealand http://www.aat.co.nz/images/graph123.gif

Ennuyer M 2000. Follicular growth in cows: practical aspects for management of breeding. Point Vétérinaire 31, 9-15.

Digitized and rescaled schematic view

O. Martin et al. / Theriogenology 79 (2013) 331–343

343.e2

(continued ) Reference of original schematic view

Digitized and rescaled schematic view

Forde, N., Beltman, M.E., Lonergan, P., Diskin, M., Roche, J.F. & Crowe, M. a. (2011) Oestrous cycles in Bos taurus cattle. Animal reproduction science, 124, 163-9.

Learning Reproduction in Farm Animals - Cow and Mare Estrous Cycles. Oklahoma State University. Adapted from Ginther, O.J. 1992. 2nd Edition Reproductive Biology of the Mare. p288 http://animalsciences.missouri.edu/reprod/Notes/estrous/estrous.htm

Ownby, 2001 Oklahoma State University College of Veterinary Medicine http://instruction.cvhs.okstate.edu/histology/fr/HiFRp15.htm

Chapter 5 The Estrous Cycle - by Choong-Saeng Park Gyeongsang National University (South Korea) http://nongae.gsnu.ac.kr/wcspark/teaching/chap5.html

(continued on next page)

343.e3

O. Martin et al. / Theriogenology 79 (2013) 331–343

(continued ) Reference of original schematic view New Bolton Center Field Service Department, University of Pennsylvania School of Veterinary Medicine’s Computer Aided Learning program http://cal.vet.upenn.edu/projects/fieldservice/Dairy/REPRO/estfolwv.htm

THE BOVINE ESTROUS CYCLE, 2004 - George Perry, Extension Beef Reproduction Management Specialist, South Dakota State University, Cooperative Extension Service, USDA http://agbiopubs.sdstate.edu/articles/FS921A.pdf

Peters A and Lamming E 1983. Hormone patterns and reproduction in cattle. In Practice 5, 153-158.

Ponsart, 2003. Le cycle oestral. BTIA 110, 20-22. D’après GAMEPI, CD Rom AFC

Digitized and rescaled schematic view

O. Martin et al. / Theriogenology 79 (2013) 331–343 (continued ) Reference of original schematic view Lee Rinehart, 2009: Dairy Production on Pasture: An Introduction to Grass-Based and Seasonal Dairying, Publication of ATTRA, National Sustainable Agriculture Information Service, Graph courtesy of University of Missouri Extension http://attra.ncat.org/attra-pub/PDF/grassbaseddairy.pdf

Digitized and rescaled schematic view

343.e4