Periconceptional nutrition with spineless cactus (Opuntia ficus-indica) improves metabolomic profiles and pregnancy outcomes in sheep

We tested whether periconceptional nutrition with cladodes from the cactus, Opuntia ficus-indica, with or without protein-enrichment, improved the metabolomic profile and reproductive outcomes of adult female sheep. Sixty Rambouillet ewes of similar body weight were randomly allocated among three nutritional treatments that were fed during the breeding period (34 days): Control (Control; n = 20), Opuntia (Opuntia; n = 20) and protein-enriched Opuntia (E-Opuntia; n = 20). There were no effects of treatment on body weight but assessment of urine samples indicated that, for 76 metabolites, the Control and Opuntia groups differed completely (P < 0.05), whereas there was overlap between the Control and E-Opuntia groups. It appears that, in Opuntia-fed and Control-fed sheep, different functional groups are activated leading to changes in the metabolism of glucose, tyrosine, methane, and glycerolipids. Fertility and reproductive rate tended to be higher in the Opuntia (70% and 95%) and E-Opuntia (90% and 110%) groups than in the Control (55% and 65%), and an orthogonal contrast revealed the difference between Control and Opuntia to be significant for both reproductive variables (P < 0.05). We conclude that nutritional supplementation with Opuntia cladodes, with or without protein enrichment, increased fertility rate and reproductive rate of female sheep, without any accompanying increases in body weight. Our observations suggest that the reproductive responses to Opuntia do not simply reflect a response to good nutrition, but might be caused by specific metabolites/metabolomic pathways, perhaps by an activation of the metabolism of glucose, methane, tyrosine and glycerolipids. There are few reports relating these metabolomic compounds with the metabolism of the sheep, let alone with reproductive efficiency. The novelty of these discoveries suggests that we need further research into the mechanisms through which nutrition affects the reproductive system.

Small ruminants represent the principal economic output and thus contribute a large share to the income of farmers in arid and semiarid regions. These regions are often characterized by overgrazing, poor-quality soils and low annual rainfall that is also highly variable from year to year. Accordingly, the pattern of forage supply varies from year to year and crop production is unreliable 1 , and this scenario is becoming more difficult with climate change. Animals raised under these conditions generally do not receive nutritional supplements because high-quality pasture and concentrate feeds are too expensive 2 . The productivity of the system therefore depends directly on the forage provided by degraded rangelands 3 . Nevertheless, production from ruminants in these dryland systems is expected to play an increasingly important role in feeding the world, so innovative and versatile options for livestock production are needed to maximize productivity and improve ecosystem health 4 . There is a clear need for abundant, low-cost alternative feed supplements that can improve the production efficiency of animals reared under these difficult conditions.
A variety of alternative supplements have been proposed, such as betacarotene and ball moss (Tillandsia recurvata) 5,6 . For arid and semi-arid areas, there is considerable interest in the Opuntia genus, an abundant plant that can be used as a feed supplement and also seems to confer benefits for reproduction in small ruminants [7][8][9] . As a succulent, Opuntia spp. can withstand water shortage and high temperatures, and they adapt readily to poor soils, so they can be cultivated at low cost 10,11 . Importantly, the distribution of Opuntia ficus-indica encompasses Latin America, South Africa and the Mediterranean area 12 .
The cladodes of Opuntia ficus-indica are highly digestible and can provide water and energy 13,14 , although the protein content is low and varies with plant maturity, ranging from 2.8 to 4.1% 13,14 . The protein content can be improved by fermentation with additives or by adding urea [15][16][17][18] . Supplementation with Opuntia cladodes during specific periods of the production cycle therefore seems to be a real option for arid and semiarid conditions.
Feeding Opuntia cladodes can improve ovulation rate, the response to the male effect, and postpartum ovarian activity 17,19 . The consensus for the effects of improved nutrition on ovarian activity is that circulating glucose and related homeostatic hormones act directly on follicle development 20 . However, most of the studies leading to this perspective were based on experiments using supplementation with lupin grain. There is some information on the metabolic pathways activated by feeding with Opuntia cladodes 21,22 , but responses to protein-enriched cladodes have not been studied, for either metabolic profiles or reproductive outcomes. It is feasible that increasing the protein content of Opuntia cladodes will increase protein synthesis in the ovary and therefore influence follicular development and enhance fertility 20 , especially when the nutritive quality of other feed is low.
For evaluating phytochemical bioactive compounds and determining the quality of diets, metabolomic techniques have emerged as an important tool 23,24 . These techniques are now considered valuable for addressing future needs in agriculture and animal production because they can be used to measure and detect molecules in biofluids and therefore unravel the pathways leading to production of specific biomarkers. Metabolomics might therefore provide further insights into the mechanisms responsible for improvements in production and reproduction after the feeding of supplements such as Opuntia cladodes 25 . Using a suite of metabolomic measurements, we therefore tested the hypothesis that supplementing the diet of adult female sheep with Opuntia cladodes would improve the metabolic profile, increase body weight gain, and enhance reproductive performance. We also tested whether these responses would be improved if the cladodes were protein-enriched.

Results
Experimental measurements were made Day − 14 (the first day of the 2-week adaptation period), Day 0 (the first day of dietary treatment), Day 30 of treatment, and Day 45 (10 days after the end of treatment).

Body live weight and body live weight change.
There were no significant differences in ewe body weight among treatments, in absolute values or in changes over time throughout the experiment (P > 0.05; Table 1).
Fertility and reproductive rate. Fertility rate was 55% (11/20 ewes) in the Control group, 70% (14/20) in the Opuntia group, and 90% (18/20) in the E-Opuntia group. The initial statistical analysis only revealed a tendency for differences among treatments (P = 0.07), but an orthogonal contrast showed that the Control and E-Opuntia treatments differed significantly (P < 0.05). In the ewes that conceived, reproductive rate was 65% (2 pregnant ewes carrying twins out of 11 pregnant ewes) in the Control, 95% (5 pregnant ewes carrying twins out of 14 pregnant ewes) in the Opuntia group, and 110% (4 pregnant ewes carrying twins out of 18 pregnant ewes) in the E-Opuntia group. Again, none of the differences were significant (P > 0.05) except for the orthogonal con- Table 1. Ewe body weight at the start of the experiment, at the start of breeding, and at the end of breeding period, and bodyweight change during the breeding period in mature Rambouillet ewes. The animals were offered only alfalfa (Control), protein-enriched opuntia (E-Opuntia) or Opuntia during the breeding period. a Standard error of the mean from the mixed model output. Urine metabolomic profile. On Day − 14 no difference on metabolomic profile among nutritional treatments was observed (Supplemental Figure 1); afterward, across all urine samples from all treatments, a total of 76 metabolites were consistently identified, of which 13 had indiscrimination index of 0.99 ( Table 2). The PLS-DA score plots based on the urine metabolome database on Day 0 (Fig. 1A), Day 30 ( Fig. 2A)  Time-resolved analysis of metabolomics data. For the time-series metabolomics, an ANOVA-simultaneous component analysis was used to identify the major differences in metabolites between treatments (Control and Opuntia) and major changes in metabolites with time (Table 3 and Fig. 5A). A permutation approach was used to validate the model as verified by significance levels of P < 0.01 for both treatment and time (Fig. 5B). Significant variables were identified on the basis of the leverage and squared prediction errors (SPE) associated with each variable, comparing the Control and Opuntia treatments, and comparing times, as shown in Table 3 and Fig. 5C. Treatment effects were identified in five metabolites (2-furanmethanol, diethyl ether, propanal, E-3octen-2-one, ethanol). As shown in Fig. 5C, several metabolites changed over time in a similar fashion for both the Control and Opuntia treatments, either increasing (Ethanol) or decreasing (2-furanmethanol, 2-octel, methane, bromodichloro-, and perfluronone). Some metabolites remained steady over time: diethyl ether, β-ionone, ethyl butyrate, and heptane in the Opuntia group; propanal and β-pinene in the Control (Fig. 5C). Finally, E-3octen-2-one concentration decreased over time in the Control group but increased over time in the Opuntia group (Fig. 5C).
Metabolites associated with reproduction. Table 4 presents the correlations among urinary metabolomic compounds and fertility or reproductive rate according to nutritional treatment. Only correlations that were P < 1.0 for either fertility or reproductive rate within a nutritional treatment are presented. From further analysis with Mixed Models of the relationships between each urinary metabolomic compound and fertility or reproductive rate, according to nutritional treatment, it is clear that, in the Control treatment, none of the correlations were significant (P > 0.05). By contrast, in the E-Opuntia treatment, the concentration of 2-methyl-2-propanol, 2-methylbutane and β-ionone was positively correlated to fertility rate (P < 0.05). For this treatment, there were no significant correlations between any metabolomic compound and reproductive rate. In the Opuntia treatment, the concentration of demeton-O and β-ionone was positively correlated to fertility rate (P < 0.05); whereas the concentration of butane, demeton-O and trimethylamine was positively correlated to reproductive rate (P < 0.05).

Discussion
Supplementation with Opuntia cladodes, with or without protein enrichment, increased fertility rate and reproductive rate, without any accompanying increases in body weight. The lack of effect on body weight, despite the 48-day duration of treatment, reflects observation in previous studies in two laboratories, with sheep and goats 9,17,19 . The lack of effect on body weight is a critical observation for interpretation of the outcomes for reproduction and fertility because it suggests that the effects of Opuntia on the reproductive system do not simply www.nature.com/scientificreports/ reflect a response to the extra energy provided by supplement and are, in fact, caused by activation of specific metabolomic pathways. Indeed, the metabolomic measurements have revealed pathways that might explain the fertility responses: activation of the metabolism of glucose, methane, tyrosine and glycerolipids. Positive effects on the growth and developmental competence of preovulatory follicles, with a subsequent improvement in ovulation rate were described in the initial studies showing that cactus cladodes can increase reproductive success in sheep in harsh environments 19,26 . These effects were later confirmed in goats 27,28 . High energy diets increase the concentrations of glucose and metabolic hormones, both of which can act directly on the ovary to enhance the responsiveness of the follicles to follicle-stimulating hormone 29,30 . The responses to Opuntia are coherent with this situation because the cladodes are a rich source of soluble sugars 31 and cladode supplements activate the metabolism of glucose. Specifically, we observed changes in the metabolomic footprints of alcohols (ethanol, 2-methyl-2-propanol; 2-furanmethanol) and gases (propyne, 2R, 4R-tetrahydro-4-methyl-2-2-methylprop-1-enyl-2H-pyran) produced as a result of the fermentation of glucose. We also observed changes in (E,Z)-2,4-heptadienal, a medium-chain aldehyde that is produced during the fermentation of glucose by E. coli 32 , perhaps in the rumen 33,34 . Additionally, in grazing lambs, (E,Z)-2,4-heptadienal has been identified in adipose tissue 35 having apparently been formed from the decomposition of 18:3 n-3 fatty acids 36 . The changes in these processes caused by Opuntia supplements suggest that we should look beyond glucose and glucosamine as energy-related factors that affect the ovarian follicles 37,38 .
After correlating each metabolomic compound with either fertility or reproductive rate, we observed that 17 compounds from the E-Opuntia treatment, 17 compounds from the Opuntia treatment and 1 compound from the Control treatment influenced either positive or negative the reproductive variables tested. Yet, a further statistical analysis using mixed models indicated a positive relationship between a few metabolic compounds and fertility rate (2-Methyl-2-propanol, 2-Methylbutane, and β-ionone) in the females from the E-Opuntia treatment. Similarly, within the females from the Opuntia treatment, few metabolic compounds were positively relate to fertility rate (demeton-O, β-ionone) or reproductive rate (demeton-O, butane, trimethylamine). To our knowledge, www.nature.com/scientificreports/ there are few reports relating these metabolomic compounds with the metabolism of the animal, let alone with reproductive efficiency, so further research is needed to elucidate the underlying mechanism driving these effects. A positive effect of cactus cladodes on final fertility rates was only found when they were combined with a high-protein supply such as soybean 26 . A reason for this interdependence has not been elucidated but, keeping in mind that the high sugar content of cactus cladodes improves rumen fermentation 39 , a possible hypothesis would be a synergistic interaction between the sugars and a nitrogen source, whether that be urea, soybean or lupin grain. The importance of dietary protein in reproduction in adult female sheep has long been controversial 40,41 and we suggest that this situation could be resolved by further metabolomic studies.
The sheer diversity of outcomes in metabolomic analysis opens avenues for further research into the relationships between nutrition and reproductive performance in general, not just with Opuntia. For example, the metabolomic analysis showed that intake of Opuntia cladodes also affected the metabolism of tyrosine and glycerolipids. Tyrosine is essential amino for the production of the thyroid hormones and the catecholamines 42 that, in turn, influence the activity of GnRH neurons 43 . Tyrosine and lysine are link by hydrogen bonds 44 and lysine also has a prominent role in cell proliferation 45 and is therefore important in feto-placental growth during pregnancy 46,47 . Glycerolipids, being mainly triacylglycerides, are used for energy storage and also play a prominent role in pregnancy because they are a major source of energy for the developing fetus following through the provision of fatty acids 48 .
In conclusion, feeding ewes with Opuntia, with or without protein-enrichment, can increase the reproductive performance of adult ewes and seems to be a viable option for managing sheep under extensive conditions in arid and semiarid regions. The enhanced reproductive efficiency with Opuntia was not accompanied by increases in body weight despite the 34-day duration of nutritional supplementation. This observation, along with the metabolomic analysis of responses to Opuntia and protein-enriched Opuntia, suggests activation of functional processes that affect the metabolism of glucose, tyrosine, methane, and glycerolipids, all of which could affect the reproductive system through specific processes that do not reflect responses to a simple improvement in www.nature.com/scientificreports/ feed energy supply. The novelty of these discoveries suggests that we need further research into the mechanisms through which nutrition affects the reproductive system.

Material and methods
The study was conducted during the breeding period on a commercial farm in northern Mexico (22°15′ N, 100°52 W). All animal procedures were consistent with international guidelines 49 and with national guidelines 50 for the care and use of laboratory animals, and met the ARRIVE guidelines for reporting animal research 51 . Our institutional committee approved the experimental study with the reference number 10561934075. Fig. 6. In August (the breeding season), 60 mature Rambouillet ewes of proven fertility were randomly allocated among three dietary treatments, ensuring similar average body weights for each group: Opuntia (Opuntia; n = 20), protein-enriched Opuntia (E-Opuntia; n = 20) and a group fed with only alfalfa (Control; n = 20; see details below). The ewes were fed twice daily, at 8-9 a.m. and at 3-4 p.m. Opuntia treatments (Opuntia and E-Opuntia) started two weeks before mating (Day − 14) to allow adaptation to the diet and continued until the last day of the mating period (Day 34). Initially, each ewe received 500 g per day and the amount offered was gradually increased during adaptation period to reach 3 kg per animal per day when the experiment started (Day 0). Food refusals were quantified on a pen basis but, after the two first days of the adaptation period, all of the diet offered was consumed. Therefore, on average, the dry matter intake for the Control treatment was 1.3 kg per day for each animal, and for both Opuntia treatments it was 0.78 kg per day for each animal. Before mating; female sheep were dewormed with a commercially available product containing 1 g ivermectin (BAYMEC; BAYER; MEXICO). Additionally, female sheep had received an intramuscular injection of commercially available product containing 0.005 g of B12 (CATOSAL; BAYER; MEXICO) and 500,000 IU of VitA, 75,000 IU of VitD3 and 50 mg of VitE according to manufacturer's guideline     Experimental diets. Control animals were fed alfalfa hay at 3% of their average body mass and treatment groups were fed with Opuntia (Opuntia) or protein-enriched opuntia (E-Opuntia) at 3 kg per head per day. Opuntia cladodes were harvested daily and cut into small pieces to facilitate consumption. For the E-Opuntia treatment, 600 g urea and 80 g ammonium sulfate were dissolved in 20 L of water, and this solution was sprayed on 100 kg chopped cladodes, after which the material was allowed to stand for 24 h. Diets were provided in a fence-line feeder with sufficient space for all animals to consume all of their respective diet at once. Feed rejection was quantified, but none was rejected after the first few days of the adaptation period. On average, the daily dry matter intake per animal was 1.3 kg for the Control treatment and 0.78 kg for the Opuntia and E-Opuntia treatments. The composition of the diets was analysed by AGROLAB México S.A de C.V (Table 5). All diets provided less than the nutritional requirements for a dry ewe with low physical activity 52 .

Experimental design. The experimental protocol is shown in
Metabolomic assessment. Urine was sampled from a sub-group of ewes from each treatment on four occasions: before the start of treatment (day − 14), during treatment (days 0 and 30) and 10 days after the end of treatment (day 45). For the first sampling, 10 ewes were selected at random from each nutritional treatment and these same 10 ewes were then sampled on every occasion. Urine was collected over ice in sanitized plastic containers with lids and immediately transported to the CIACyT laboratory for storage at − 80 °C until analysis.
Urine samples were analysed by ultrafast gas chromatography with an 'electronic nose' (FGC eNose; HERA-CLES II, ALPHA MOS COMPANY, France) equipped with two columns connected in parallel with different stationary phases, coupled to two ultrasensitive flame ionization detectors (μ-FIDs). Two non-polar MTX 5 columns (10 m × 0.18 mm × 0.4 μm film thickness) and two medium polar MXT-1701 (10 m × 0.18 mm × 0.4 μm film thickness) were used. A subsample (1 mL) of urine was injected in headspace mode under the following conditions: the sample was incubated for 900 s at 40 °C with agitation at 500 rpm; for chromatography, the temperature as initially held at 50 °C for 30 s, after which it was increased isocratically to 280 °C at a rate of 10 °C s −1 . The temperature of the injector was maintained at 200 °C. The temperature of the detector was maintained at 280 °C. Prior to the chromatographic separation, the 1 mL headspace sample was adsorbed on a Carbowax trap maintained at 60 °C for 50 s while the carrier gas flowed through it in order to concentrate the analytes and to remove excess air and moisture. Subsequently, the sample was desorped by increasing the temperature of the trap to 240 °C in 30 s and the sample was injected. The total separation time was 80 s. Ultrapure hydrogen was used as a drag gas. A standard solution of C6-C16 alkanes was used for the determination of Kovats index. The FGC e-nose data were processed by Alpha soft V12. 44 and AroChembase (Toulouse, France) and VOCs were identified according to the 2011 NIST Retention Index database. Statistical analysis. We used the SAS statistical package version 9.3 53 . With linear mixed model procedures (PROC-MIXED) we analyzed the effects of treatment on body weight at the start and end of the breeding period, and on changes in body weight during treatment. Treatment was considered as fixed effect in the model. Body weight at the start and end of the breeding period, and body weight change during the breeding period, were included independently as covariates where appropriate. Body weight values were fitted in a linear regression model (weight on time) for each individual and the regression coefficient was estimated as a measurement of change in weight per unit of time.
Data for fertility rate (percentage of ewes pregnant per 100 ewes mated) was analyzed using the PROC-GLIMMIX with a binomial distribution and logit link function. Treatment was considered as fixed effect. Covariates were body weight at the start and end of the breeding period, and body weight change during the breeding period, and were included independently where appropriate. Data for reproductive rate (number of fetuses in utero per 100 ewes mated) were analyzed using the PROC-GLIMMIX with a multinomial distribution and logit www.nature.com/scientificreports/ link function. The same fixed effect and covariates were used as for the analysis of fertility. The data for fertility and reproductive rate are presented as logit values and back-transformed percentages. The correlations among urinary metabolites, body weight change, fertility and reproductive rate during the breeding period were predicted using PROC GLM with the MANOVA option to allow removal of major fixed effects. The fixed effects included in the model were treatment and sampling date.
All two-way interactions among the fixed effect and covariates were included in each model, and non-significant (P > 0.05) interactions were removed from the analysis. Significant differences among means for treatments for variables measured at different time points during the experiment were analyzed using LSD of PROC GLM.
The multivariate statistical analysis of the urinary metabolite data was performed using the web-based metabolomics data processing tool, MetaboAnalyst 4.0 54 . Briefly, for quality Control, variables containing more than 50% missing values (i.e., values lower than limit of detection) were not considered for the statistical analysis. The metabolite data were transformed using the generalized log transformation and then range-scaled to correct for heteroskedasticity, and to reduce mask effects 55 . Partial least squares discriminant analysis (PLS-DA) and variable importance in projection (VIP) were performed using R 56 to identify the differential metabolites among groups and to rank the metabolites according to their importance in discriminating groups. In addition, PLS-DA was used to differentiate the Control treatment from the Opuntia and E-Opuntia treatments based on the urinary metabolomic data, and to identify the most important metabolites that contributed to the treatment differences. Permutation tests were used to assess the significance of the class discrimination determined by PLS-DA. Classification and cross-validation were performed using the corresponding wrapper function offered by the caret package. In each permutation, a PLS-DA model was built between the data (x) and the permuted class labels (y) using the optimal number of components determined by cross-validation for the model based on the original class assignment. The leverage and the squared prediction error (SPE) were used to evaluate the importance of the metabolite in the model and the fitness of the model for the particular metabolite.
Ethics approval and consent to participate. All animal procedures were consistent with international guidelines 49 and with national guidelines 50 for the care and use of laboratory animals, and met the ARRIVE guidelines for reporting animal research 51 . The experimental study was approved by INIFAP Committee with the reference number 10561934075.

Data availability
The data presented in the manuscript are not deposited in an official repository.