The experimental site is located in northeast Mexico (25° 32’ N, 103° 40’ W) at 1150 m above sea level. Mean annual precipitation is 225 mm, most of which falls as high-intensity thunderstorms from June to October. The mean annual temperature is 22.3° C. The overstory was predominantly Prosopis spp., Larrea tridentata, and Atriplex canescens. Other important shrubs present were Agave lechuguilla, and Opuntia rastrera. The most abundant forbs are Sphaeralcea angustifolia, Solanum elaeagnifolium, Salsola kali, and Lepidium virginicum. Grasses constitute only a small part of the vegetation and grow mainly beneath shrubs. The principal species are Munroa pulchella, Setaria macrostachya, and Muhlenbergia porteri.

A total of sixteen fruit for each treatment group, in four replicates of four apples each, were randomly selected (two apples per tree). Tissue samples consisted of 2 g of whole fruit (skin plus pulp) for each treatment. Samples were then frozen with liquid nitrogen, pulverized and homogenized in a mortar, and then extracted according to the method described by Coseteng & Lee (1987) with some modifications. The tissue was extracted twice (for 10 and 5 min) at 100 °C, with an 80% ethanol solution (ethanol:water 80:20, v/v), and then filtered. Samples were volumetrically diluted to 10 mL with 80% ethanol and kept at -20 °C until use.

Animal procedures were agreed upon and performed following the Institutional Animal Care and Use Committee of the Agrarian Autonomous University Antonio Narro (Protocol # 03001-2258) and carried out following FASS (2010). The study was conducted from May to October 2021 in a large goat herd in a microphyll desert scrub ecosystem of northern Mexico. A total of 89 mixed-breed (Central Europe dairy breeds × criollo) goats were used in the present study. Goats did not present any physical defects, had not given birth in the previous five months, were not pregnant, and had BCS ranging from poor to good (BCS 1–3 on a scale of 1–5). Goats grazed on open degraded Chihuahuan desert rangeland in plain terrain, year-round, driven by a herdsman for 6 h per day (from 1100 to 1700 h). Goats were confined after returning from grazing in an unshaded pen, without access to feed supplementation and water. No salt mineral mix was provided to the goats throughout the year; goats had access to water from a pond only once a day. Goats were not treated against gastrointestinal and external parasites or vaccinated against endemic diseases.

Four adult (64.5 ± 7.3 kg BW) mixed-breed bucks with adequate BCS (3.5; scale 1-5) with previous mating experience and an account of satisfactory kidding percentages were joined to dry goats for 30 days in May 2021. Approximately thirty days after the end of the mating season, pregnancy was detected using ultrasonography. Pregnancy detection was again assessed 120 days after mating. At kidding, kidding rate and litter size were recorded.

At mating, trans-rectal ovarian ultrasonographic evaluations were performed using an ultrasound scanner (Aloka 500V, Corometrics Med. Syst. Inc., Wallingford, CT, USA) equipped with a linear array transrectal probe (7.5 MHz transducer) by a single experienced operator. Follicles were counted, and their height was recorded on a frozen image. Follicles were classified according to follicular height as small (<3 mm in size), growing mid-sized follicles (3 to 5 mm), and large follicles (>5 mm) (Rateb et al., 2019). The ovulation rate was determined by identifying the corpus luteum and the disappearance of the previously observed mature follicle(s) 10 days after mating.

Blood was sampled from the selected goats by jugular venipuncture in non-anticoagulation gel separator vacuum tubes (Vacuntainer®). Immediately after collection, the samples were centrifuged, and the serum was placed in Eppendorf tubes and stored at −20 °C until it was assayed. Serum glucose, total protein (TP), cholesterol, blood urea nitrogen (BUN), creatinine, and phosphorus (P) concentrations were determined using colorimetric methods following protocols supplied by the kits´ manufacturers (Sigma Diagnostics Inc., Livonia, MI, USA). Non-esterified fatty acids (NEFA) were determined using a commercial kit (WAKO; Mountain View, CA, USA). In addition, serum minerals were determined by atomic absorption spectrophotometry.

Continuous variables were tested for normality and group homogeneity using the UNIVARIATE procedure of SAS (SAS Inst. Inc., Cary, NC, USA, vers. 9.4); continuous data were normally distributed. The variables were compared between groups (pregnant vs. non-pregnant; pregnancy loss vs. non-pregnancy loss, single vs. twin pregnancy) using the GLM procedure of SAS. Variables were described as mean values ± standard deviation, differences between groups with 95% confidence intervals were computed (TTEST procedure of SAS), and significance was declared when α was 5%.

These variables were dichotomized for the effect of serum metabolites and minerals on ovarian structures, using their mean as a cutting point for classifying concentrations below or above the mean. Then, the GLM procedure of SAS was used to detect differences between levels of metabolites and minerals on the number of total follicles and ovulatory follicles, mean ovulatory follicles height, number of corpus luteum, and average corpus luteum height. Parity was included in the model as a covariate.

Canonical correlations for exploring the relationships between two multivariate sets of variables were applied to select follicular traits to identify groups of metabolites with a related biological role in ovarian structures using Statgraphics Centurion 19 (Statgraphics Technol. Inc., The Plains, VA, USA). Also, principal component analyses were carried out to understand the sources of variation of data for pregnancy of goats and see distances between important serum variables affecting pregnancy rate. Significance was declared at p<0.05.

In the present study, of the 89 experimental goats, 64 were confirmed as pregnant (72% pregnancy rate) on day 30 post-service using ultrasonography. Upon reexamination on day 120 of mating, ultrasonography indicated eight pregnancy losses, and consequently kidding rate was 63%, with a mean litter size of 1.5 ± 0.59 (± SD). Principal components derived from serum metabolites, ovulatory follicle height, minerals, and ovarian variables showed a clear separation between pregnant and non-pregnant goats (Fig. 1). The first two principal components explained 58.8% of the variation in the data.

Data for BCS, serum metabolites, and minerals for pregnant and non-pregnant goats are shown in Table 1. The univariate general linear model analyses showed that BCS was significantly higher (p<0.05) in pregnant than non-pregnant goats. No significant effect of the confounding variable age on pregnancy outcome was observed. Mean serum glucose concentration was 13 mg/kg higher (p<0.05) for goats that became pregnant compared with non-pregnant goats. There was a significant (p<0.01) effect of pregnancy on serum BUN concentrations, which indicated that high blood BUN levels were associated with pregnancy outcomes. All other serum metabolites indicative of body energy reserves and minerals were not different for non-pregnant compared with pregnant goats. No effect was observed for BCS, serum metabolites, and minerals concentration on litter size of goats (p>0.05).

Mean serum glucose concentrations at mating were much higher (p<0.01) in goats that did not lose their pregnancy than in goats that experienced a pregnancy loss before 120 days post-mating (Table 2). In addition, serum NEFA concentration at mating was 2.4 times higher (p<0.05) in goats that did not lose their pregnancy than in goats that lost their pregnancy. All other serum metabolites and minerals were not different at mating for goats with pregnancy loss and animals with no pregnancy loss. Regarding ovarian structure characteristics at mating, the mean ovulatory follicular height was significantly higher (p<0.01) in goats that became pregnant than in non-fecundated goats (Table 3).

High BCS, serum glucose, and BUN concentrations did not significantly affect the number or size of ovulatory follicles and corpus luteum. The total number of ovulatory follicles at mating was significantly higher (p<0.01) in goats with high serum TP than in goats with lower blood TP levels (Table 4). Likewise, higher serum TP concentrations led to more prominent follicles (p<0.05). Goats with higher serum P concentration presented significantly bigger (p<0.05) ovulatory follicles than goats with lower serum P levels. None of the variables studied affected the number of corpus luteum. The average height of the corpus luteum was larger (p<0.05) in goats with lower serum creatinine levels than in goats with high serum creatinine concentrations. Lower serum magnesium concentration was associated with a shorter corpus luteum (p<0.05). Additionally, the canonical correlation showed that follicle traits, BCS, and some blood metabolites were moderately correlated (r = 0.45; p<0.05), which indicates that serum glucose, creatinine, and NEFA had a significant but moderate association with the number of follicles and size.

Optimization of reproductive efficiency of goat herds on arid rangelands is a continuous challenge to goat producers. In the present study kidding rate of unsupplemented goats bred during the non-breeding season was 63%, a figure close to 72% reported in the same environment and nutritional conditions (De Santiago-Miramontes et al., 2011; Mellado et al., 2020). The suboptimal reproductive performance observed in this study is attributable to the arid conditions under which the goats are reared.

The present study evaluated the variations in BCS, serum metabolites, and mineral concentrations in pregnant and non-pregnant goats on rangeland fecundated during the non-breeding season (May). The results indicated that BCS at mating was higher in pregnant than in non-pregnant goats. These findings align with previous works showing that adequate body energy reserves, mainly represented by body fat and muscle content in goats, are required for maximum estrus response (Rivas-Muñoz et al., 2010) and pregnancy rate (Serin et al., 2010). In addition, BCS in goats is associated with blood glucose concentration (Milosevic-Stankovic et al., 2020; Sitaresmi et al., 2020), BUN (Sitaresmi et al., 2020) and NEFA (Lunesu et al., 2021), which means that BCS vary with the change of energy balance. In goats, reduced BCS leads to ovarian dysfunction (inactive or acyclic; Widiyono et al., 2020). Thus, goats with higher body energy reserves had more energy for reproductive function, which was reflected in higher odds of getting pregnant.

A comparison of the mean serum glucose concentrations between pregnant and non-pregnant goats also indicated that the higher circulating glucose concentrations occurred in goats that become pregnant. Glucose is the primary source of energy for the body’s cells (Milosevic-Stankovic et al., 2020) and the primary indicator of energy status in goats (Khan & Ludri, 2002). Improved nutrition directly affects animal metabolism by providing substrates for metabolism and cellular processes (Scaramuzzi et al., 2006). Well-fed goats show higher serum glucose concentration than underfed animals (Mellado et al., 2020).

Glucose availability for ovarian follicles can be used for energy production (Sutton-McDowall et al., 2010), and circulating glucose, insulin, and glucagon levels affect folliculogenesis and the intrafollicular environment (Ying et al., 2011; Al-Hamedawi et al., 2017). Additionally, glycemia is critical in regulating ovarian follicle responsiveness to gonadotropins (Selvaraju et al., 2003). In this context, the present study confirmed that high serum glucose availability at mating in grazing goats increases the odds of pregnancy. Furthermore, short-term nutritional supplementation rises the number of ovulatory follicles, and the ovulation rate is associated with blood glucose levels in goats (Zabuli et al., 2010). This explains the higher proportion of pregnant goats with higher serum glucose concentration when joined to bucks.

These results showed that low serum BUN in goats resulted in a reduction in pregnancy rate. BUN reflects a higher nitrogen intake, as a positive correlation exists between protein intake and BUN concentration in goats (Rondina et al., 2005; Senosy et al., 2017). Pregnancy was more likely in goats that had a mean BUN value of 12.6 mg/dL, suggesting that high BUN in goats grazing arid rangeland is favorable for conception as the increased serum BUN in goats on rangeland indicates a positive effect of diet on rumen ammonia-nitrogen concentration (Zhu et al., 2020).

Serum cholesterol concentrations were not significantly different among pregnant, non-pregnant, and goats that experienced pregnancy loss or those whose pregnancy was carried to term. Also, this metabolite did not affect the number or size of ovarian structures. This response could be explained by the fact that serum cholesterol concentrations do not differ among BCS grades (Moeini et al., 2014; Sitaresmi et al., 2020) and, therefore, under the current nutritional conditions, may not have influenced the reproductive outcome.

Blood metabolites and minerals were not different at mating between single and twin-bearing goats. These results align with observations of Cepeda-Palacios et al. (2018), who documented that the number of developing fetuses did not affect any measured hematochemical parameters. However, other studies have found differences between goats carrying singles or twins (Cappai et al., 2019; Sarıbay et al., 2020). However, in these previous studies, blood metabolite concentrations were not determined at the time of fecundation.

Regarding pregnancy losses, low serum glucose concentration at mating led to higher spontaneous loss of a pregnancy. Of the several possible mechanisms bringing non-infectious gestational failure in goats, hypoglycemia in the mother and subsequently in the fetus trigger the premature eviction of fetuses (Mellado et al., 2004, 2020). Energy-deficient diets during gestation, especially in young goats or bearing double fetuses, are important factors triggering fetal losses (Waideland & Loken, 1991; Cronjé, 1998). The few aborted fetuses observed in these grazing goats did not show signs of decomposed fetuses or placentitis, which suggests that malnutrition at mating and subsequent weeks of pregnancy caused pregnancy loss. Goats that maintain their gestation to term in this ecosystem select diets higher in nutrients than goats suffering pregnancy loss (Mellado M et al., 2014); therefore, goats’ genotypes unable to ingest nutrient-rich diets would present low serum blood glucose, which eventually leads to fetal expulsion.

Another blood metabolite at mating linked to pregnancy loss was NEFA. Goats with high serum NEFA concentrations were more likely to lose their pregnancy than goats with low circulating NEFA. This finding is in line with Hussain et al. (1996) who observed higher blood NEFA concentrations in goats with nonviable pregnancy than in goats with no pregnancy loss. This suggests increased lipolysis in goats with pregnancy loss, which presented elevated blood NEFA concentrations, indicative of acute energy restriction and mobilization of body reserves when the glucose level decreased (Veerkamp et al., 2003). Perhaps, the higher serum NEFA concentration was related to deficient nutrition for non-adapted goats to lower forage quality or a high fiber diet on rangeland, and they had lower energy status to sustain pregnancy than their better-adapted counterparts (Mellado M et al., 2014).

Ovulatory follicle size affected fertility when ovulation occurred after the buck stimulus. Contreras-Solís et al. (2021) observed that large follicles from prepubertal ewes had higher estradiol and progesterone concentrations, more competent oocytes, and blastocyst produced in vitro than less developed follicles. Also, it has been reported that GnRH-induced ovulation of follicles ≤11 mm results in lower pregnancy rates and augmented late embryonic/fetal mortality, associated with lower circulating estradiol concentrations and decreased circulating progesterone concentrations (Perry et al., 2005). Therefore, ovulatory follicle size is a robust indicator of fertility (Perry et al., 2007) and follicular growth stimulation with equine chorionic gonadotropin leads to greater pregnancy rate in goats (Hameed et al., 2020).

Lower serum creatinine concentrations were associated with larger corpus luteum. High blood creatinine concentrations are a reliable biomarker of body protein breakdown and muscle mass change (Patel et al., 2013), and therefore, creatinine concentrations change in response to body protein mobilization. In goats, creatinine increased linearly with the decreasing crude protein concentration in the diet (Zhu et al., 2020). Therefore, researchers have used this metabolite to monitor nutrient status and muscle mass (Turner et al., 2005).

Higher serum TP concentrations were associated with a greater number of ovulatory follicles and their size. This metabolite accommodated both the variations in the albumin and globulins, showing a clear separation for the undernourished and well-fed sheep (Caldeira et al., 2007). Thus, this study reaffirms the effects of energy level during the antral phase and subsequent follicular development on follicle recruitment and size. This response has been observed in goats, where dietary energy levels positively influence oocyte follicular development and meiotic competence (Kabir et al., 2022).

The fewer ovulatory follicles in goats with elevated serum NEFA concentrations clearly show the effect of the well-known decrease in basal metabolic rate in animals in a state of undernutrition. Energy supplementation in sheep has increased the number of follicles and the amount of double ovulation (Habibizad et al., 2015), which are connected to a rise in the number and size of preovulatory follicles (Cuadro et al., 2018). Also, greater serum P concentration resulted in larger preovulatory follicles, a singularity of this mineral that improves reproductive performance of anestrus sheep restoring their ovarian activity, increasing the number and size of ovarian follicles and size of corpora lutea (Senosy et al., 2018).

In summary, this study showed that high serum BUN and glucose, greater follicular development, and better BCS at mating are sensitive biochemical and physical markers to detect grazing goats capable to become pregnant. Further, high serum glucose and low NEFA concentration at mating are predictive of goats maintaining pregnancy on rangeland.