Serum, milk, saliva and urine progesterone and estradiol profiles in crossbred (Zebu x Holstein Friesian) dairy cattle

Effective breeding management is a crucial tool to enhance the reproductive and productive performance of dairy cattle. To achieve this, regular monitoring of their reproductive status is important. The aims of this study were to monitor the reproductive status of crossbred dairy cattle based on progesterone (P4) concentration in serum, whole milk, saliva and urine, and estradiol in serum using quantitative laboratory ELISA tests, and to document hormonal profiles at different reproductive stages. Three hundred and thirty-six (336) crossbred (Holstein Friesian x Zebu) dairy cattle (232 cows and 104 heifers) from 47 dairy farms owned by smallholder or organized commercial dairy farmers were included in the study. Matched blood, milk and saliva samples were collected twice at 11-day intervals at afternoon milking, while urine was collected just once. Serum was separated within 48 hr of blood collection. All samples were stored at -80oC until assayed. In both cows and heifers, P4 was higher during pregnancy than in other reproductive conditions. Estradiol was higher during estrus than at other reproductive stages. Estradiol levels were significantly higher (p<0.05) in estrus cows compared with heifers. The mean level of P4 in serum strongly correlated with the value in milk (r = 0.6368, P< 0.0001), however there was no correlation in P4 concentration between serum and saliva or between P4 in serum and in urine. Hence, determination of P4 in serum and milk using an ELISA test is a reliable and precise method to monitor reproductive status in crossbred dairy cattle. Correspondence to: Mekonnin AB, The University of Edinburgh, College of Medicine and Veterinary Medicine, MRC Centre for Reproductive Health, QMRI, 47 Little France Crescent, Edinburgh, EH16 4TJ, UK; E-mail: alemselambnh@yahoo.com


Introduction
Effective breeding management is a crucial tool to enhance the reproductive and productive performance of dairy cattle. Failure to detect and correctly interpret signs of estrus can contribute to significant loss in the dairy industry [1,2]. Although, visual observation of animals for signs of estrus is a common method of estrus detection across the globe, in developed countries objective tests are used to confirm estrus [3], including measuring serum/plasma or milk progesterone (P4) levels, and to successfully breed cattle.
Progesterone plays an essential role in various reproductive functions, including regulating the length of the estrous cycle, maintaining pregnancy [4,5] and regulating embryonic growth and development [6,7]. P4 is high during the luteal phase and in pregnancy, however the levels decline if the animal fails to conceive [8,9]. Monitoring P4 in plasma/serum [10,11] or milk [11][12][13] using an Enzyme Linked Immunosorbent Assay (ELISA) or Radioimmunoassay (RIA) has been used to assess the reproductive status of cattle and ensure successful breeding, or inform decisions to cull non-productive cows/heifers. Additionally, it has been used to detect pathology of the reproductive system such as cystic ovaries [14].
Determination of P4 in saliva using ELISA has also been used to determine the reproductive status in farm animals [15][16][17] as well as in women [18]. However, determination of P4 concentration alone is not sufficient to predict ovulation because there is a large variation in the timing of P4 decrease relative to ovulation [19]. Ovarian estradiol plays an important role in establishing the timing of uterine receptivity to the developing embryo [20] and embryo survival [21]. In the bovine, estrus is initiated after a rise in circulating estradiol level [22]. Identifying the estradiol peak can precisely indicate pre-ovulation [13], which may assist prediction of insemination timing yielding high conception rates. Circulating concentrations of estradiol during the preovulatory period can influence the establishment and maintenance of pregnancy by altering the uterine environment [21,23]. Additionally, estradiol is proposed to induce FSH/ LH receptor expression in granulosa cells [24] and increase the stimulatory action of FSH on aromatase activity [25].
In Ethiopia, pregnancy diagnosis in cattle is only carried out using per-rectal palpation by AI technicians or veterinarians [11]. Recently, we showed that qualitative on-farm milk/serum P4 ELISA tests effectively diagnosed pregnancy as early as 18-days post service and confirmed cyclicity in cattle [26]. However, the on-farm qualitative testing does not quantify P4 levels in cattle, and is designed only to determine the relative P4 concentration (high or low) [27]. As dry cows and heifers are not lactating, it is not feasible to use milk to determine P4 levels, so urine or saliva samples are attractive alternatives for monitoring reproductive status [16].
The metabolism of P4 is rapid and occurs mainly in the liver [28][29][30]. The P4 metabolites are released from the liver into the blood circulation, and excreted by the kidneys into the urine [28,30]. Pregnanediol glucuronide is a common urinary metabolite of P4 [31][32][33][34][35]. Volkery et al. [36] reported that plasma concentration of P4 and estrone sulphate correlated with pregnanediol-3-glucuronide and estrone sulphate levels in matched urine in female alpacas. Similarly, estrone conjugates and pregnanediol glucuronide in urine were measured using RIA and/or EIA, and their value paralleled the profile of the parent steroid in serum/plasma in women [32,37].
Changes in total P4 circulation in blood (protein -bound plus free, non-protein bound fractions) have been widely used for accurate monitoring of reproductive status in cows [15,16,[38][39][40]. Over 80% of circulating P4 is bound to plasma proteins [41,42] and only the free fraction is biologically active [43,44]. Sex steroids in saliva can be used as an index of their unbound plasma concentration [43,45,46]. Hence, the quantitative correlation between the two fractions could be used to determine P4 concentration in matched blood and saliva samples [15,47,48].
Saliva and urine have the advantage of being non-invasive and relatively stress-free methods and can easily be collected by farmers [48][49][50]. However, P4 profiles at different reproductive stages of crossbred cattle in Tigray region and other parts of Ethiopia are not well established. The aims of this study were to monitor the reproductive status of crossbred dairy cattle based on P4 concentrations in serum, whole milk, saliva and urine, and estradiol in serum using quantitative laboratory ELISA tests, and to document P4 and estradiol profiles at different reproductive stages.

Study animals
The study was approved by

Selection of study animals
Overtly healthy animals, having a mean BCS of 3.1, range 2.5-4 (using a scale of 1-5) were included in the study [52]. According to breeding history (farmer's observation), animals that had never come into estrus 60 to 90 days postpartum (anestrus), repeaters, pregnant animals (>18 days post AI/mating) and those that were reported inheat at the time of sample collection were included. Breeding history of each animal in the experimental group was taken. Reproductive status was confirmed on the farm by qualitative on-farm P4 assays and animals were grouped accordingly as anestrus, in-heat, diestrus and pregnant (1 st trimester, 2 nd trimester and 3 rd trimester).

Sample collection
Blood collection and serum separation: Blood was collected in 10 ml plain vacutainer tubes (BD Vacutainer®, BD, Plymouth, UK) from the jugular vein of dry cows and heifers twice at eleven-day intervals.
Blood was allowed to clot at room temperature for 24-48 hr after collection and serum was separated. Serum was stored in 5ml capacity sterile plastic transport tubes (Alpha Laboratories, Hampshire, UK) and stored at -80 o C until assayed.
Milk collection: Milk (10-20ml per animal) was collected into plastic centrifuge tubes (Alpha Laboratories) twice at eleven-day intervals from lactating dairy cows. Samples (foremilk) were taken manually from all four quarters after discarding the first four milk drops. Milk was collected only from cows with a clinically healthy udder and teats. Samples were frozen at -80 o C until assayed.
Saliva collection: Saliva (5ml from each animal) was collected from dairy cows and heifers twice at 11-day intervals using a 7ml capacity plastic transfer pipette directly from the buccal commissures by aspiration. In cases when insufficient saliva was collected, animals were provided with some green feed or hay/straw to stimulate salivation, and in some cases saliva dropping out of the mouth was collected directly to the sterile, plastic transport tubes (Alpha Laboratories). Samples were transported from the farms to the laboratory in a cooling box with ice and samples were centrifuged at 4000g for 10 min at 20 o C to remove any contaminating food particles. In some cases, samples were frozen and centrifuged, multiple times until food particles were completely removed. Samples were stored at -80 o C until assayed. Collection procedures were based on previous reports [15,16].
Urine collection: Urine (5ml/cow) was collected when cows urinated normally or following stimulation through massaging the ventral commissure of the vulva. Samples were frozen at -80 o C until assayed.

Experimental design
Steroid hormone extraction: Whole milk, saliva and urine were used without steroid extraction (intact) for P4 measurement. Steroids were extracted from serum for both P4 and estradiol measurements. Extraction procedures were as follows: 120μl (for P4 measurement) or 250μl (for estradiol) of serum, standards or controls were added into 16 x 100mm glass tubes. 2ml (for P4) or 3ml (for estradiol) diethyl ether (Sigma-Aldrich, St. Louis, MO, USA) was then added to each glass tube and steroids were extracted on a muli-vortexer (IKA® Vibrax VXR basic, Staufen, Germany) for 5 min. Extracted samples were frozen in an ethanol (VWR International, Fontenay-sous Bois, France) plus dry ice bath, and the ether layer decanted. Extracted samples were dried overnight in a fume cupboard, or dried down rapidly under a stream of nitrogen in a hot block (Dri-Block® DB-3, Techne Ltd, Cambridge, UK) at 40 o C for a maximum of 1hr. Extracted samples were reconstituted by vortexing with charcoal-stripped serum (volume equal to original volume of sample) and assayed immediately or stored short term at 4 o C until assayed.

P4 determination in serum, whole milk, saliva and urine:
Quantitative in-house laboratory ELISA that was developed by Dr. Forbes Howie (MRC Centre for Reproductive Health, The University of Edinburgh, UK) was adapted to quantify the level of P4 in serum, whole milk, saliva and urine. All samples, standards and controls were first thawed at 4 o C, and then warmed to room temperature. Whole milk and saliva samples were further warmed in a tube warmer (Dri-Block, Bibbi Scientific Ltd., Stone, Staffordshire, UK) at 40-50 o C for 40 min to homogenize the milk and to break down the mucin present in saliva.
Urine P4 ELISA: The same ELISA protocol as used for saliva was used to determine P4 in urine, without extraction. The P4 level in each urine sample was divided by the creatinine level in the same sample to adjust for dilution effects. Creatinine was determined using the creatininase/creatinase specific enzymatic method described by Bömer et al. [53] utilizing a commercial kit (Alpha Laboratories Ltd.) adapted for use on a Cobas Fara centrifugal analyzer (Roche Diagnostics Ltd., Welwyn Garden City, UK). Within run precision was CV < 3% while intra-batch precision was CV < 5%.
Estradiol measurement in serum: Estradiol was measured using Estradiol Sensitive ELISA kit according to the manufacturer's instructions (Demeditec Diagnostics GmbH, Kiel, Germany). Steroids were extracted prior to assaying. Wells were precoated with a polyclonal anti-estradiol antibody; 100μl of each standard (0, 3, 10, 50, 200 pg/ml of estradiol), control and samples were added into appropriate wells. Then, 200μl enzyme conjugate was added into each well, thoroughly mixed for 10 seconds on a plate shaker and plates were incubated for 4 hr at room temperature. Plates were washed 3 times with 400μl per well of wash solution and 200μl of TMB was added to each well, and then plates incubated for 30 min at room temperature. The reaction was stopped by adding 100μl stop solution. Finally, plates were read on plate reader at 450nm. Inter-and intra-assay CV were determined as for P4, and were 15.2 & 18.1% and 13.6 & 13.5% for high and low pools of controls, respectively. Cross-reactivity of antiestradiol antibody was reported as 0.2% for estrone, 0.05% for estriol and <0.001% for a range of other steroids (Demeditec Diagnostics GmbH).

Statistical analysis
Data was analysed using GraphPad Prism 6. Data was checked for normal distribution using D'Agostino-Pearson Test, and analysed using one-way ANOVA. Post hoc Dunn's multiple comparisons test was used to compare P4 level between pregnant cows and pregnant heifers, between 1 st , 2 nd and 3 rd trimester pregnancies within cows and heifers, between in heat and diestrus cows and heifers and between diestrus and pregnant animals.

P4 profiles
Laboratory serum P4 ELISA results were considered as gold standard to classify and confirm reproductive status. The P4 level in milk, saliva and urine was measured once cattle were classified into their respective reproductive status following serum P4 determination. Three different assays were developed for three different sample types, serum, milk and saliva/urine. Extraction of steroids was performed for serum, to avoid any solvent recovery interference; standards and controls went through the extraction processes at the same time as the unknown samples. Extraction was deemed unnecessary for saliva/urine assay. Extraction of steroids for milk proved unsuccessful; therefore, P4 measurements had to be done without extraction.

Serum P4
The mean levels of P4 in serum strongly correlated with the level in milk (r = 0.6368, P< 0.0001) (Figure 2A), however there was no significant correlation between serum and saliva P4 or between serum and urine P4 levels (Figure 2 B, C and D). Serum P4 measurement in crossbred dairy cattle in different reproductive states is indicated in Figures 3, 4 and 5. Cattle with P4 levels of ≥1ng/ml in either of the tests (measured twice at 11-day intervals) were considered to have functional ovaries; hence, animals were regarded as normally cycling. In contrast, cattle with low P4 level (<1ng/ml) in both tests had non-functional ovaries (no CL or follicle detected per rectum) and were considered as anestrus. Cattle that were observed in-heat and had a low P4 level (P4 <1ng/ml) were considered in-heat or follicular phase. All other animals that had a history of insemination and serum P4 ≥1ng/ml at both tests, together with per-rectal palpation findings, were considered pregnant. The P4 levels (Mean±SEM) of three cows that had ovarian follicular cysts was measured in matched serum and milk, and were 0.37 and 0.13 ng/ml respectively.
The mean serum P4 in cows and heifers either in-heat or anestrus was below 1ng/ml (Figure 3). There was similar serum P4 levels in cows and heifers that were either anestrus, in-heat or in their 3 rd trimester pregnancy. Significantly higher (p <0.05) P4 was recorded in heifers than in cows in diestrus and in the 1 st trimester; whereas significantly higher P4 was detected in cows compared with heifers in 2 nd trimester. While most normally cycling cattle had <1ng/ml of P4 in one of the tests (first or second test), 22 cattle (10 cows and 12 heifers) had P4 levels ≥ 1ng/ml in both the tests. In contrast, two pregnant cows (1 st and 2nd trimester, each) had low (< 1ng/ml) serum P4 concentration, whereas none of the pregnant heifers had <1ng serum P4. The overall mean P4 level in pregnant cows was significantly higher (p < 0.05) than in pregnant heifers. Cows that were in 1 st trimester of pregnancy had significantly lower (p < 0.05) serum P4 levels than heifers at the same stage of pregnancy. In contrast, significantly higher (p < 0.05) serum P4 was detected in cows compared with heifers during 2 nd trimester. P4 concentration in serum was similar between pregnant and diestrus heifers (Figure 4), however pregnant heifers had significantly higher (p < 0.05) P4 in saliva than had diestrus heifers. In contrast, there was significantly higher (p < 0.05) serum and milk P4 concentration in pregnant than diestrus cows ( Figure 3).

Milk P4
Milk P4 concentration was determined in 200 lactating cows. Similar to serum, the level of P4 in milk was low in cows that were anestrus and in-heat, 0.16 and 0.14 ng/ml, respectively ( Figure 3). The P4 levels in diestrus cows and heifers were similar (Figure 3). The overall mean P4 in milk was significantly higher (p < 0.05) in pregnant cows compared to those in diestrus (Figure 3). Cows that were in their   3 rd trimester of pregnancy had significantly higher (p < 0.05) milk P4 than cows in other stages of pregnancy. Milk P4 in cows at 2 nd trimester pregnancy was lower than those at 1 st and 3 rd trimester ( Figure 5). Eight pregnant crossbred cows (one in 1 st trimester, six in 2 nd trimester and one in 3 rd trimester) had milk P4 levels < 1ng/ml.

Saliva P4
The P4 level in saliva in anestrous cows and heifers was low, and similar to cattle in-heat. It was also similar to the levels in serum and milk (Figure 3 & 4). Similarly, cows in-heat had similar saliva P4 levels to heifers in-heat. Diestrus cows and heifers had significantly higher (p < 0.05) saliva P4 levels than those in anestrus or in-heat. Conversely, diestrus cows/heifers had significantly lower (p< 0.05) mean P4 concentrations in saliva than pregnant cows. A significantly lower (p < 0.05) P4 level was detected in saliva than in serum and milk in diestrus and pregnant cattle. Pregnant cows had a similar overall mean P4 to pregnant heifers. Pregnant cattle had low saliva P4 level (0.13ng/ ml) in their second trimester of pregnancy. Significantly higher (p < 0.05) saliva P4 was detected in cows in 1 st trimester than the 2 nd and 3 rd trimester ( Figure 5), with similar P4 between 2 nd and 3 rd trimesters. The P4 levels in saliva were significantly lower (p < 0.05) than in serum in both cows and heifers and in milk in cows at all stages of pregnancy ( Figure 5). In heifers, P4 was significantly higher (p<0.05) in 1 st and 3 rd trimester compared with 2 nd trimester, with similar P4 levels between 1 st and 3 rd trimesters ( Figure 5).

Urine P4
The mean urine concentration of P4 in pregnant cows was 119.24 ng/mg of creatinine. A significantly higher (p< 0.05) P4 level was detected in pregnant cows than in cows that were anestrus and in estrus ( Figure 6).

Serum estradiol profile
Serum estradiol concentration was estimated using a laboratory ELISA test (Figure 7A & B). Anestrus and diestrus cattle had significantly lower (p < 0.05) estradiol concentration compared with cattle in other reproductive stages, and the level was similar between cows and heifers. A significantly higher (p < 0.05) estradiol concentration was measured in cows which were in estrus than cows in other reproductive states. Pregnant heifers in term trimester had significantly higher (p<0.05) estradiol levels than heifers in-heat and in other reproductive states. A significantly higher (p<0.05) difference was found in serum concentration of estradiol in cows than heifers that were in estrus. In contrast, a significantly higher (p < 0.05) estradiol level was detected in pregnant heifers than in cows in their first trimester. In both cows and heifers, higher estradiol concentrations were recorded during term pregnancy in contrast to other stages of pregnancy ( Figure 7B).

P4 detection in non-pregnant cycling and anestrous cattle
In the present study, cattle in estrus had low concentrations of P4 (< 1ng/ml), which is in line to previous reports in serum [54][55][56][57], plasma in Zebu cows [58] and milk [58][59][60]. In contrast, Rajamahendran et al. [61] reported that 32 (4.8%) cows that were submitted for breeding had milk P4 ≥1ng/ml, that was in line with other reports [62][63][64][65]. Saliva P4 concentration in cattle in estrus in this study was higher than in previous reports in Friesian/Jersey cows [16]. It has been demonstrated that P4 levels during the follicular phase fluctuates from 0.025 to 0.08ng/ml [15]. Although P4 levels are low when cows/heifers are in estrus, a low concentration of P4 does not necessary mean the animal is in estrus, as anestrous cattle with small ovaries without palpable follicles or CL, also have low levels of P4. However, it has been proposed that high P4 accurately confirms that the animal is not in estrus, even when animals exhibit behavioral symptoms of estrus [27]. This suggests inaccurate estrus detection [66]. The use of on-farm or laboratory P4 assays had played a significant role in identifying cycling (in-heat and those in the luteal phase) cattle. This increases the chance of breeding cattle in-heat and properly detecting when animals in the luteal phase subsequently come into estrus.
The present study has determined the P4 concentration in urine. Urinary creatinine concentration was used as a reference to determine P4 levels [67,68], based on previous reports suggesting urinary creatinine excretion is constant in individual animals [69,70]. However, a recent report has shown that P4 metabolites (such as pregnanediol glucuronide) concentration can be measured for monitoring reproductive status of dairy cows without the need for creatinine adjustment [49]. Yang et al. [49] has shown that the correlation between the urinary pregnanediol glucuronide and serum P4 was higher without adjustment than when adjusted by urinary creatinine concentration, which is in line with our finding whereby the level of P4 in serum and in urine was not correlated. Furthermore, studies on ovarian function in women using urinary steroid assays have shown that the correlation between serum steroid and their metabolite levels in urine was improved when not adjusted for creatinine levels [32]. The P4 profile in urine of cows that were in heat, in the present study, was similar to a previous report in North American Bison cows [71]. The concentration of P4 (steroid) metabolite varies widely among different animal species [49].   The P4 level is measured, in the present study, in cycling cattle in diestrus. However, as the specific date of the previous ovulation or estrus was not available, P4 level of cattle at unknown stage of the estrous cycle, but with the existence of a palpable CL per-rectal palpation was measured. Some animals had high P4 levels (≥1 ng/ ml) in either of the samples collected twice at 11-day intervals with no breeding history postpartum, reflecting the animals were normally cycling. Hence, of the two samples tested per animals, the one with the higher P4 concentration was taken for analysis in diestrus cattle, assuming it is comparatively closer to the level when the animal is in the middle of her estrous cycle. Some cycling cattle had high P4 levels in both samples collected at 11-day interval. The reason for cows/ heifers having high P4 levels for an extended period could be due to a prolonged luteal phase [72]. A prolonged luteal phase is one of the common reproductive problems of dairy cows, occurring mostly from calving to 90 days postpartum [73,74].
Ovarian cyclicity was determined based on the fluctuation of P4 levels in two samples collected 11 days apart with at least one of the samples having a P4 concentration ≥ 1ng/ml. That indicates the presence of a CL. The level of P4 in non-cycling (anestrous) cattle is low due to the absence of CL. In the present study, crossbred cattle that did not have CL during per-rectal palpation and was confirmed by quantitative laboratory ELISA to have serum and milk P4 levels <1 ng/ml, in two samples collected at 11-day intervals, were considered as true anestrus. This correlates with the commonly accepted threshold of 1.0ng/ml for ovarian function [75]. However, several other studies used different P4 thresholds to distinguish luteal activity. A threshold of 0.5 ng/ml was used to determine non-luteal cows [12], while a threshold of 3 ng/ml was used to indicate luteal activity in the analysis of milk P4 profiles [73]. The circulating P4 profile of crossbred cattle during luteal phase (diestrus) in the present study was similar to previous reports in HF cows in Thailand that used a serum RIA [55]. Conversely, serum P4 levels in diestrus cattle in the present study was higher than a previous report in crossbred cows in India [54]. In contrast, lower milk and serum P4 was recorded in diestrus cattle in the present study than reports in HF cows in the Netherlands [19] and Punganur cows in India [56]. Higher serum P4 levels recorded in the present study in diestrus heifers than in diestrus cows was in agreement with a previous report after administration of CIDR [76]. The P4 concentration in diestrus cattle in this study was higher than previous reports in Friesian /Jersey cattle [15,16].

P4 determination in pregnant cattle
Progesterone is an essential hormone for the maintenance of pregnancy, however, the minimum concentrations of P4 required is not clearly understood [77]. The P4 concentration recorded in the present study in crossbred cattle in early stage (first trimester) of pregnancy was similar to previous reports in crossbred cattle as well as local breed (Zebu) cattle in other part of Ethiopia [78] and in Zebu cows in Bangladesh [58]. In contrast, crossbred cows and heifers in the present study had higher serum P4 during their early pregnancy than HF cows in Pakistan [10].
Milk P4 recorded in crossbred cattle in early pregnancy in the present study was similar to previous reports in Zebu cows in Bangladesh [58]. Conversely, comparatively lower milk P4 was recorded during early pregnancy in the present study than in previous reports [79]. P4 is soluble in milk fat, which is the reason for its higher level in milk than in blood [80]. The mean saliva P4 concentration detected in pregnant crossbred cows in this study was >1ng/ml, which was higher than in previous reports [15,16]. In sheep, a mean P4 concentration of 1.8ng/ml has been reported [81], which is similar to the value in pregnant cows in this study. The P4 level in saliva during first trimester of pregnancy was not significantly higher than second and significantly higher than third trimester and this was significantly higher in cows than in heifers. The higher P4 in cow saliva than in heifers may be because cows have comparatively well-developed reproductive physiology associated with previous calving and have larger ovaries with large CL. Previous reports have shown a positive correlation between the size of CL and P4 level [82].
Low milk P4 (< 1ng/mL) detected in pregnant cows in this study is in agreement with previous reports from British Colombia whereby 135 (29%) cows in early pregnancy (50-60 days post-service) showed low milk P4 level [61]. Furthermore, Bulman and Lamming [83] reported that a pregnant cow calved successfully despite its milk P4 level between 8 and 16 days after insemination being basal.
Several previous studies have shown application of analysis of urinary steroid metabolites to determine pregnancy status in various species [36,49,84,85], including human [86]. Both bovine blood and bovine CL tissue can convert P4 to 20β-hydroxypregn-4-en-3one [87] that can be determined in urine. It has been reported that 5β-pregnane-3α,20α-diol detected in the urine of late pregnant cows [88]. Determination of the level of pregnanediol, which is an inactive metabolite of P4, has also been used as an indirect way of measuring P4 to diagnose pregnancy [49]. The higher P4 level detected in the urine of pregnant cows than in anestrus or cows in estrus, in the present study, reflects the formation and persistence of matured CL. The P4 profile in urine in this study was lower than a previous report in North American Bison cows [71], which may be due to breed difference.
Collecting blood samples for serum/plasma steroid hormone measurement requires venipuncture; hence, it is stressful and painful to the subject [18,49]. Furthermore, the challenges and the risk (to experts) collecting blood samples are formidable, particularly when working in farms where proper restraining facilities are not available. Unfortunately, almost all of the farms where the present study was conducted had no cattle crush for restraining the animals. Therefore, determining P4 or estradiol from serum samples for regular follow up of reproductive status of cattle is not safe for both the animals and the experts, particularly when someone approaches aggressive cattle. Though this particular study directly helped farmers involved in the study, as it was conducted at the field level, there were some farmers reluctant to provide the second blood samples, even some refrained from being included in the study. When asked why, they did not want their animals to feel pain and stressed, particularly those they thought were pregnant.
Milk is routinely used for determining reproductive status in lactating dairy cattle [12,19]. Our studies support previous work [89] that shows P4 concentrations in milk closely correlate those in the blood/plasma during the reproductive cycle in cows. Unlike the present findings that agreed with a recent report [48], previous reports demonstrated a positive correlation in the concentration of P4 in blood and saliva [15,16]. However, it has been proposed that the most significant problem with milk P4 measurements is the presence of considerable amount of lipids, which interfere with the assay [90,91]. Sampling has also been proposed to affect P4 concentration in milk [90]. Additionally, it has been reported that storage conditions can affect P4 level in bovine milk whereby frozen samples were reported to have higher P4 concentration compared to fresh samples [13]. Furthermore, the availability of milk is limited to lactating cattle only.
Saliva and urine could be used as alternative samples to accurately determine the reproductive status of dairy cattle. Both saliva and urine can be collected through noninvasive methods, also from dry cows and heifers, hence avoid the animal welfare inconveniences when blood is collected or other inconveniences from using milk samples. Saliva has additional advantages due to its ease of storage, as it can be stored at -20 o C for over 6 months, at 48 o C for 7 days, or at room temperature for 48 hr with no significant change in steroid levels [92]. Furthermore, urine has several advantages over saliva, serum and milk: it can be collected in large volume that enables for multiple assays [93,94,95] and provides integrated hormone measures without the confounding effects of pulsatile secretion [37]. Additionally, the availability of steroid metabolite in higher concentration in urine (by two-to fourfold or more) than that of parent sex steroids in blood give greater advantage to the use of urine [49,85] or in women [96]. Although the P4 level in urine in the present study did not correlate with its level in matched serum, based on previous reports that have shown urine P4 profiles correlated with serum profiles [49,85], it might be more reliable specimen to use for monitoring reproductive status of dairy cattle. Furthermore, in the present study, collecting saliva was equally challenging as collecting blood in most cattle; whereas, urine was easily collected from both cows and heifers by the farmers themselves. This gives better advantage towards the use of urine in the future when simple and rapid on-farm Dipstick method is developed, and can be used by farmers themselves.

Determination of serum estradiol
Estradiol profiles recorded in crossbred heifers in estrus, in the present study, were in agreement with previous reports in Sahiwal heifers in estrus following treatment with CIDR in India [63], in Swedish dairy heifers [97] and in HF heifers following administration of PGF2α in USA [98]. In contrast, the estradiol level recorded in heifers in estrus in the present study was comparatively higher than reports in crossbred beef heifers in USA [99]. Conversely, crossbred heifers that were in estrus in the present study had lower serum concentration of estradiol than previous reports in HF heifers in estrus induced by PGF2α in USA [98] and in Bali heifers in Indonesia [100]. The reason for lower estradiol profiles in the present finding than in previous reports may be due to breed difference. The estradiol level in crossbred dairy cows in estrus, in the present study, was similar with previous reports in crossbred normally cycling or repeat-breeding cows in India that were in estrus [65] and in Sahiwal cows in India that were in estrus induced by CIDR administration [63]. In contrast, the present study recorded higher circulating estradiol in cows during estrus than previous reports in Punganur cows in India [56]. Conversely, the estradiol profile in cows in estrus in the present study was lower than a previous report in Bunaji cattle in Nigeria [101]. Higher estradiol levels were detected, in the present study, in cattle exhibiting estrus compared to anestrous or cycling (diestrus phase) cattle, which agrees with a previous report [102] that that showed a positive correlation between follicle size and peak estradiol concentration, and greater peak estradiol concentration in cows exhibiting estrus compared to not exhibiting estrus.
Serum concentration of estradiol in anestrous cattle or cycling cattle in the diestrus phase of the estrous cycle, in the present study, was in agreement with a previous report in postpartum crossbred beef cows with calves weaned >30 days and reported to have ovulatorysized follicles (>10 mm) but failed to exhibit standing estrus following Co-Synch protocol [102]. Conversely, the estradiol profile in serum in crossbred cattle in our study was lower than previous reports in postpartum (20-40 days) crossbred Angus beef cows with calves allowed to suckle without restriction throughout the experiment in USA [102], in Punganur cows in diestrus (in 12 th day of the estrous cycle) [56] and in anestrous or cycling (diestrus) Bunaji cattle in Nigeria [101]. The higher estradiol level in pregnant heifers than in heifers in estrus or other reproductive conditions (anestrus or diestrus) was in agreement with a previous report in Bunaji cattle in Nigeria [101]. However, the overall estradiol profile in serum in pregnant cattle in the present study was lower than a previous report [101]. The reason behind this could be breed difference.
A quantitative ELISA technique can be applied in the Ethiopian research laboratories and veterinary clinics to identify pregnant dairy cattle and assess ovarian function. However, unlike the on-farm ELISA, application of quantitative ELISA at field (farm) level is not possible due to its high cost and the requirement for skilled farm personnel [25,103].
In conclusion, the present study confirmed that a quantitative ELISA technique could be used to accurately determine the steroid profile and monitor the reproductive status of dairy cattle. This enables the detection of cycling animals and hence to successfully breed them, and perform pregnancy diagnosis as early as 18-days post service. Lastly, the measurement of sex steroids in various sample types should provide a basis for further related studies in crossbred, as well as local breed cattle in the tropics and other regions.

Conflicts of interest
The authors declare no conflict of interests.