Strategies of inorganic and organic trace mineral supplementation in gestating hyperprolific sow diets: effects on the offspring performance and fetal programming

Abstract

The aim of the present study was to evaluate the effect of trace mineral nutrition on sow performance, mineral content, and intestinal gene expression of neonate piglets when inorganic mineral sources (ITM) were partially replaced by their organic mineral (OTM) counterparts. At 35 d postmating, under commercial conditions, a total of 240 hyperprolific multiparous sows were allocated into three experimental diets: 1) ITM: with Zn, Cu, and Mn at 80, 15, and 60 mg/kg, respectively; 2) partial replacement trace mineral source (Replace): with a 30 % replacement of ITM by OTM, resulting in ITM + OTM supplementation of Zn (56 + 24 mg/kg), Cu (10.5 + 4.5 mg/kg), and Mn (42 + 18 mg/kg); and 3) Reduce and replace mineral source (R&R): reducing a 50% of the ITM source of Zn (40 + 24 mg/kg), Cu (7.5 + 4.5 mg/kg), and Mn (30 + 18 mg/kg). At farrowing, 40 piglets were selected, based on birth weight (light: <800 g, and average: >1,200 g), for sampling. Since the present study aimed to reflect results under commercial conditions, it was difficult to get an equal parity number between the experimental diets. Overall, no differences between experimental diets on sow reproductive performance were observed. Light piglets had a lower mineral content (P < 0.05) and a downregulation of several genes (P < 0.10) involved in physiological functions compared with their average littermates. Neonate piglets born from Replace sows had an upregulation of genes involved in functions like immunity and gut barrier, compared with those born from ITM sows (P < 0.10), particularly in light piglets. In conclusion, the partial replacement of ITM by their OTM counterparts represents an alternative to the totally inorganic supplementation with improvements on neonate piglet gene expression, particularly in the smallest piglets of the litter. The lower trace mineral storage together with the greater downregulation of gut health genes exposed the immaturity and vulnerability of small piglets.

Introduction

Although swine genetic selection has successfully increased the litter size (>16 piglets), this has also involved a lower average piglet birth weight, a higher variation of birth weights within the litter, and an increase in the percentage of piglets with low birth weights (even lower than 800 g; Kemp et al., 2018; Oliviero et al., 2019). In general, the presence of small piglets, weighing less than 1 kg, in large litters can range between 15.2% and 25% (De Vos et al., 2014; Feldpausch et al., 2019), who also has a 5.9 times higher risk of dying before weaning compared with their heavier littermates (Feldpausch et al., 2019). Moreover, up to 30% of these small piglets can be exposed to different degrees of intrauterine growth restriction (Foxcroft et al., 2006; Amdi et al., 2013). Birth weight is not only a crucial economic feature in swine production, it is also critical for the piglet development. Indeed, the small vulnerable piglets often show an impaired performance due to a delayed and reduced colostrum intake, and poor thermoregulatory capacity (Kemp et al., 2018), while it may have permanent negative impacts on organ structure and postnatal growth efficiency (Ji et al., 2017) compared with their heavier littermates. Some nutritional interventions during pregnancy, such as the supplementation of diets with functional amino acids (Mateo et al., 2007; Nuntapaitoon et al., 2018) have been proposed to mitigate the incidence of small piglets. Along this line, it is well known that trace minerals (TM) are essential nutrients due to their vital roles in a wide variety of physiological processes, including structural and metabolic functions (Suttle, 2010). For instance, in sows, they are critical for metabolism, and growth of fetuses, as well as for the production of colostrum and milk and uterine involution (Hostetler et al., 2003; Suttle, 2010; NRC, 2012). However, unlike energy and amino acid requirements, very little research is available that focuses on the effects and requirements of TM in highly prolific sows, and definitive information is scarce (NRC, 2012). The National Research council (NRC, 2012) estimated a requirement of Zn (100 mg/kg), Cu (10 mg/kg), and Mn (25 mg/kg) for gestating sows. In the swine industry, TM are supplemented at different levels (Zn: 80 to 125 mg/kg; Cu: 6 to 20 mg/kg; Mn: 40 to 60 mg/kg), and frequently with inorganic sources (ITM), most as sulfates and oxides. Inorganic sources, as sulfates, are characterized by an unstable structure, making them highly soluble in water and acid solutions, while promoting a negative interaction with other feed components like phytic acid (Villagómez-Estrada et al., 2020) and minerals (Walk et al., 2015). In contrast, organic TM (OTM) have a stable chemical structure formed by coordinate covalent bonding between the metal ions and organic molecules (AAFCO, 2021), which make them less prone to feed interactions and therefore to have greater mineral bioavailability (Acda and Chae, 2002). In the present study, the OTM utilized are chelates of protein hydrolysate. Studies conducted in weaned and growing pigs have shown similar (Hill et al., 2014; Liu et al., 2016) or greater efficiency (Veum et al., 2004; Zhao et al., 2014) in promoting growth when ITM were totally or partially replaced by OTM counterparts. However, little is known about the effects of these strategies on the performance of gestating sows (Peters and Mahan, 2008; Ma et al., 2020; Tsai et al., 2020) and, in particular, their likely effects on fetal programming of small piglets. During gestation, after 35 d postmating, when the embryonic implantation has already been established and, consequently, the potential size of the litter, the uterine capacity begins to be the main limiting factor in fetal development, which in terms of nutrition means an increased fetal competition for available maternal nutrients to grow and survive (Langendijk and Plush, 2019). An optimal maternal nutrition, with specific nutrients as TM, may improve the development and survival of fetus within the litter. Consequently, in the present study, we hypothesized that, due to the chemical properties of TM sources, the productive performance of hyperprolific sows fed from 35 d postmating with two combinations and doses of inorganic and organic mineral sources of Zn, Cu, and Mn would be higher than those fed only with a complete dose of inorganic sources. The objectives of the study were to evaluate the effect of a partial substitution of Zn, Cu, and Mn inorganic sources by Zn, Cu, and Mn organic forms, as well as to evaluate the effect of a dietary reduction level on the hyperprolific sow performance, mineral content in colostrum and piglet tissues, as well as intestinal gene expression of neonate piglets. The influence of maternal nutrition on the offspring was evaluated on two neonate littermates from 10 sows per diet, which were selected based on their birth weights (light: <800 g and average: >1,200 g).

Materials and Methods

All animal experimentation procedures were approved by the Ethics Committee of the Universitat Autònoma de Barcelona (CEEAH 3817) in compliance with the European Union guidelines for the care and use of animals in research (European Parliament, 2010).

Animals and housing

The experiment was carried out on a commercial farm under Spanish commercial conditions. In Spain, the sow population of hyperprolific line is about 35% of the total sow population. At the end of early gestation (35 d postmating; confirmed gestation by ultrasound), a total of 240 healthy sows from hyperprolific line (DanBred hybrid line; Landrace × Yorkshire, parity 4.5 ± 1.8), obtained from the same commercial farm, were randomly assigned to three experimental diets (n = 80). For each experimental diet, sows were individually weighted and housed in two gestation pens, one corresponding to light weight sows with parities from second to fourth and another pen for heavy weight sows with parities from fifth to eighth. Given that from day 35 after mating, there is an intensive sibling competition for maternal nutrients, that conditioned the fetal development, the present experiment was designed to assess how gestational TM nutrition can improve litter development and survival. Therefore, experimental diets were fed daily from day 35 postmating until sows were transferred to farrowing crates (day 110). Each gestation barn (6.7 × 14.2 m; 40 sows) had an electronic feeding station (model Intec-Mac, Mannebeck, Schuttorf, Germany) used to control individual sow feed intake through an electronic ear tag. Feed was provided throughout the day. Sows were fed a mean of 2.4 kg/d up to 35 d postmating, and subsequently a mean of 2.3 kg/d until day 110 postmating, with individual adjustments if necessary. During the experiment, two evaluations were performed at 63 d postmating and at 93 d postmating aimed to adjust individual feed intake of sows using a visual body condition score on a scale from 1 (thin) to 5 (overconditioned) performed by a trained technician (Fitzgerald et al., 2009), and also evaluating the sow back-fat thickness using a veterinary ultrasound scanner (model WED-3000 V, Welld, Shenzhen, China). Thus, sows considered as thin (body condition score of 2 and back-fat thickness between 9 and 13 mm) received an additional 30% of feed only for 30 consecutive days, which resulted on 2.6 kg of average provided feed per day during gestation. Although overweight or fat sows (body condition scores up above 4 and back-fat thickness up to 26 mm) were restricted to 1.9 kg/d resulting in an average provided feed of 2.1 kg/d gestation. The general health and welfare status of sows was daily assessed. Water was provided ad libitum through commercial nipple waterers. Light was provided by daylight (via windows) and artificial light (programmable; from 07 00 h to 19 00 h). The temperature inside the buildings was automatically controlled and maintained in a range between 19 °C and 20 °C using force-speed fans linked to temperature sensors with cooling. At day 110 of gestation, sows were individually weighed and moved to farrowing crates (0.7 × 2.2 m) where the respective experimental diet of lactation phase was provided ad libitum. The farrowing facility was environmentally controlled with 32 individual stalls of fully slatted floor for sows and a heated hard plastic floor for piglets, and equipped with a feeding ball system (Rotecna, Agramunt, Lleida, Spain) and two nipple waterers (1 for sow and 1 for piglets) to ensure feed and water ad libitum. The feeders were manually filled twice a day (08:00 and 15:00 h) to ensure ad libitum intake of the experimental lactation diets.

Experimental design and dietary treatments

Gestation basal diet (Table 1) was formulated to meet or exceed nutrient requirements (NRC, 2012). A basal vitamin-premix without Zn, Cu, and Mn was previously prepared. The sources and doses of Zn, Cu, and Mn were subsequently added to this basal vitamin-premix according to the experimental design. Therefore, three vitamin-premix were obtained and consequently added to the basal diet to obtain three different experimental diets. The experimental diets were designed to provide mineral levels close to the swine industry and NRC (2012) requirements, as follows: 1) inorganic trace mineral diet (ITM) was supplemented with Zn, Cu, and Mn at 80, 15, and 60 mg/kg, respectively; 2) partial replacement diet (Replace) replaced a 30% of ITM diet levels by organic minerals, resulting in a ITM + OTM supplementation of Zn (56 + 24 mg/kg), Cu (10.5 + 4.5 mg/kg), and Mn (42 + 18 mg/kg); and 3) reduce and replace diet (R&R) consisted of a 50% reduction of inorganic mineral sources from ITM diet but maintaining the same amount of organic minerals added in Replace diet, thus resulting in a supplementation of Zn (40 + 24 mg/kg), Cu (7.5 + 4.5 mg/kg), and Mn (30 + 18 mg/kg). The inorganic sources of Zn (zinc sulfate monohydrated, 35%), Cu (copper sulfate pentahydrate, 25%), and Mn (manganese oxide, 60%) were purchased from Pintaluba, Reus, Spain. The organic sources of Cu (Optimin Cu, 15%), Zn (Optimin Zn, 15%), and Mn (Optimin Mn, 15%) were provided by Trouw Nutrition, Amersfoort, the Netherlands. Phytase (Ronozyme NP (M), DSM Nutritional Products, Basel, Switzerland) was added at 1,500 FYT per kg of complete feed in all diets. In order to avoid cross contamination with elements from previous productions, feed was manufactured in an appropriate rank order starting with the lower concentrations to be included in the diet. Gestating and lactating diets were offered in pellet form, and the first and last 100 kg of the final diet from each experimental diet were discarded to reduce cross contamination. Experimental diets were prepared in four batches of 4,000 kg each one. From each manufacturing batch, composite samples (1 kg) were collected during the bagging process to be representative of every single experimental diet. Batch samples, from each experimental diet, were mixed and thereafter proportionally split using a riffle splitter into four 250-g samples that were stored for further analysis. Mineral content in feed samples were measured in duplicate.

Table 1.

Ingredient and nutrient composition of the basal sow diet, as fed-basis

Ingredients, %	Gestation
Wheat	35.00
Barley	23.50
Wheat bran	20.00
Sunflower cake	10.00
Maize	7.70
Lard	0.50
Lysine 50	0.44
l-Threonine	0.12
Choline chloride 50	0.04
Salt	0.40
Calcium carbonate	1.45
Di calcium phosphate	0.45
Vit–min premix1	0.40
Calculated composition
NE2, kcal/kg	2,260
CP3, %	13.01
Ether extract, %	3.17
Crude fiber, %	5.87
Ca, %	0.81
Total P, %	0.56
Dig P, %	0.35
Analyzed nutrient composition, %
CP	13.03
Ether extract	3.57
NDF4	13.27
Ash	6.20
Ca	0.89
P	0.61

¹The following amounts were provided per kg diet: 10,000 IU vitamin A (acetate); 2,000 IU 25-hydroxy vitamin D3 (HyD, DSM Nutritional Products, Basel, Switzerland); 70 mg dl-alfatocoferol; 1 mg vitamin B1; 2.7 mg vitamin B2; 1.8 mg vitamin B6; 0.03 mg vitamin B12; 11 mg d-pantothenic acid; 15 mg niacin; 1 mg folic acid; 150 mg iron (FeSO₄); 0.5 mg iodine (Ca(IO₃)₂); 0.4 mg selenium (Na₂SeO₃); 25 mg butylhydroxytoluene.

²NE = net energy.

³CP = crude protein.

⁴NDF = neutral detergent fiber.

Experimental procedures and sampling

Sow body weight (BW) and back-fat thickness were recorded at days 35 and 110 of gestation. Back-fat thickness was recorded using a portable veterinary ultrasound scanner (model WED-3000 V, Welld, Shenzhen, China) measured at P2 (7.5 cm from midline at last rib). Individual feed intake was recorded daily, according to the electronic feeding station, during the gestation phase under study (35 to 110 d of gestation). Since it was not aimed to evaluate nursery sow performance, the feed intake was recorded until the sows were transferred to farrowing crates. Sow reproductive performance included total number of piglets born as live and stillborn. Only litter weight of alive piglets was recorded. A subset of 20 sows per diet from third to fifth parity was selected for sampling. At day 110, blood samples for mineral analysis were collected by jugular puncture into 10-mL vacutainer tubes (BD Vacutainer, Z, BD-Plymouth, UK) free of detectable Zn. Serum was obtained after centrifugation (3,000 × g for 15 min) and immediately frozen at −20 °C. Samples of colostrum (40 mL) were collected from the same sows by hand stimulation of all functional mammary glands within 12 h after farrowing. Colostrum samples were immediately frozen at −20 °C until mineral analysis. At farrowing, two littermates from 10 sows (from the same subset) per experimental diet were selected to take samples. Piglets were selected using their birth BW as a criterion, adjusted from the methodology of Wang et al. (2016) and Bauer et al. (2000), in two categories: light and average littermates. Briefly, a light piglet was defined as having a birth weight between 600 and 800 g and belonging to the lower quartile of litter birth BW, whereas an average littermate had a birth BW within the average birth BW of the litter (1,200 to 1,400 g). Moreover, these categories were adapted from the birth BW distribution of hyper prolific sows previously described (Quiniou et al., 2002; Škorput et al., 2018). During farrowing, when the piglets were born, their individual BW was examined and if they matched into the light or average BW category, they were selected to obtain samples. Selected piglets were removed from the sow before colostrum consumption and euthanized by an overdose of sodium pentobarbital (Dolethal, Vetoquinol, S.A., Madrid, Spain). The entire liver was extracted and rinsed in phosphate-buffered saline solution to eliminate blood residues. Approximately 3 g of liver was collected in cryotubes and immediately frozen at −80 °C until antioxidant analysis. The remaining liver and left leg were collected and immediately frozen at −20 °C until mineral analysis. Samples of jejunum for gene expression analyses were taken only from neonate piglets of ITM and Replace sows. One sample of about 1.5 cm was collected, approximately at the midpoint of jejunum, and then rinsed in PBS solution, and later snap frozen in 1 mL of RNA (Deltalab, Barcelona, Spain), and stored at −80 °C for gene expression analysis.

Chemical analysis

Analytical determinations of diets were performed according to the AOAC International (2005) methods for dry matter (Method 934.01), Dumas method for crude protein (Method 968.06), traditional Soxhlet extraction method for ether extract (Method 920.39), and ash (Method 942.05). Neutral detergent fiber was analyzed using the Ankom nylon bag technique (Ankom 200 fiber Analyzer, Ankom Technology, Macedon, NY).

Mineral analysis

Mineral content was analyzed in feed, liver, bone, colostrum, and serum. Samples were processed as described in Villagómez-Estrada et al. (2021). Briefly, samples of feed were milled at 0.5 mm before mineral analysis. Liver was dried in a forced air oven at 102 °C for 12 h and then milled at 0.5 mm. The left leg was dissected, and tibia was autoclaved to remove all the adjacent muscle, tissue, and the fibula bone (121 °C for 30 min). The tibia bone was then oven-dried for 12 h at 102 °C and soaked in acetone under a chemical hood for 48 h to extract fat. After this period, the bone was again oven-dried for 12 h at 102 °C and then broken in the middle before being ashed in a muffle furnace overnight at 550 °C. Samples of colostrum were freeze dried at −55 °C and 0.06 mbar for 72 h (model Telstar, Fisher Scientific S.L., Madrid, Spain). Prior to mineral analysis, samples of feed, liver, colostrum, and bone ash were digested with concentrated nitric acid (65%) in a microwave oven (model Ultrawave, Milestone Srl, Sorisole, Italy). Serum samples were diluted in 1:20 volume ratio with 5 mL of a solution constituted by 0.05% ethylenediaminetetraacetic acid and 0.5% of ammonia. Colostrum samples were additionally mixed with deionized water before mineral analysis. All samples were analyzed by inductively coupled plasma-mass spectroscopy (model 7500 Agilent, Santa Clara, CA).

Antioxidant and oxidant analysis

For liver antioxidant analysis, approximately 1 g of liver was weighted and homogenized on ice with 4 mL of buffer solution using a Polytron homogenizer (model T18 digital Ultra-turrax, IKA, Staufen, Germany). The buffer solution used for homogenization consisted of 109.5 g/L of 0.32 M of saccharose (Merck, Darmstadt, Germany), 1.21 g/L of 10 mM Tris (Merck), and 0.37 g/L of the phosphate ethylenediaminetetraacetic acid buffer (Merck). Hydrochloric acid (37%; Merck) was used to adjust the solution at pH 7.4. The total homogenate was centrifuged at 4 °C at 1,500 g for 15 min to obtain a supernatant for the antioxidant enzyme analysis. The activity of glutathione peroxidase (GPX) and the malondialdehyde content in supernatant were determined by spectrometry, and following, respectively, the instructions of Ransel kit (Randox, County Antrim, UK) and Cayman kit (Chemical, Ann Arbor, MI). The GPX activity was referred to the total protein content in the supernatant and was determined by the Bradford (1976) method using the Bio-Rad protein assay reagent (Bio-Rad, Santa Clara, CA).

Gene expression analysis

For a complementary evaluation of the effect of maternal diet on fetal programming, a gene expression analysis of jejunum tissue was performed. The expression of 56 genes was studied by using a self-designed Open Array Real-Time PRC Platform (Applied Biosystems, Waltham, MA). Genes involved in multiple physiological functions closely related to intestinal health were selected based on the literature, and grouped based on their main function as follows: 1) barrier function genes such as the family members of claudins, mucins, and occludins (OCLN, ZO1, CLDN1, CLDN4, CLDN15, MUC2, MUC13, and TFF3); 2) immune response genes such as pattern recognition receptors, cytokines, chemokines, and stress proteins (TLR2, TLR4, IL-1β, IL6, IL8, IL10, IL17A, IL22, IFNG, TNF, TGF-β1, CCL20, CXCL2, IFNGR1, HSPB1, HSPA4, REG3G, PPARGC1α, FAXDC2, and GBP1); 3) antioxidant enzyme genes (GPX2, SOD2); 4) digestive enzyme and hormone genes involved in the digestion and metabolism process (ALPI, SI, DAO1, HNMT, APN, IDO1, GCG, CCK, IGF1R, and PYY); 5) nutrient transport coding genes (SLC5A1, SLC16A1, SLC7A8, SLC15A1, SLC13A1, SLC11A2, MT1A, SLC30A1, and SLC39A4); and 6) stress response genes (CRHR1, NR3C1, and HSD11B1); additionally 7) four reference genes were evaluated (ACTB, B2M, GAPDH, and TBP).

The RNA gene expression analysis for jejunum tissue was performed according to the methodology described by Reyes-Camacho et al. (2020). Briefly, RNA was obtained from 50 mg of frozen jejunum tissue using the RiboPure kit (Ambion, Austin, TX) and following the manufacturer’s instruction. The quality and quantity of RNA was assessed with a NanoDrop ND-1000 spectrophotometer (NanoDrop products, Wilmington, DE), whereas the RNA integrity was checked with Agilent Bioanalyzer-2100 equipment (Agilent Technologies, Santa Clara, CA). Reverse transcription of approximately one microgram of total RNA to single-stranded cDNA in a final volume of 20 μL was performed using a High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA) and random hexamer primers. The thermal cycler conditions applied were as follows: 25 °C 10 min; 37 °C 120 min; 85 °C 5 min; 4 °C hold. A total of 25 ng of cDNA sample was pre-amplified using a 2X TaqMan PreAmp Master Mix and a 0.2X Pooled Taqman Gene Expression Custom Assays in a final volume of 10 µL under the following thermal cycling conditions: 10 min at 95 °C; 14 cycles of 15 s at 95 °C, and 4 min at 60 °C, and a final step of 10 min at 99 °C. The pre-amplified cDNA product was stored until use at −20 °C. One replicate per sample was run in a Taqman Open Array gene expression custom plate format for gene expression, with 56 assays of 48 samples per plate (OpenArray plate) in a QuantStudio 12K Flex Real-Time PCR System (Applied Biosystems).

Gene expression data analysis

Data were collected and analyzed using the ThermoFisher Cloud software 1.0 (Applied Biosystems) applying the 2^−ΔΔCt method for relative quantification and using the sample with the lowest expression as a calibrator. The maximum cycle relative threshold allowed was adjusted at 26, amplification score < 1.240, quantification cycle confidence > 0.80, and the maximum SD allowed between duplicates was set at <0.38. Samples that did not meet these criteria or showed an inadequate amplification were removed. All data were subjected to a logarithmic transformation to get closer to the Gaussian distribution. Statistical analyses were performed using R software v.3.4.3 and Bioconductor software (Gentleman et al., 2004).

Statistical analysis

Data were analyzed with ANOVA by using the GLIMMIX procedure of SAS (version 9.4, SAS Institute, Cary, NC). The experimental unit for sow performance parameters was the individual sow. The model included the fixed effects of diet and parity group, as well as the random effect of gestation pen. Initial BW (day 35) was used as covariable for the analysis of BW at day 110. The standardized sow BW was calculated by subtracting the litter BW from the sow BW at day 110. Sow weight losses attributed to placenta tissue or fluids were not considered. Sow reproductive performance included total number of piglets born, born alive, stillborn, as well as the weight of alive litter and the average of alive piglet weight. Reproductive performance at farrowing was related to the sow. Colostrum and sow serum mineral content were analyzed with sow as experimental unit, and parity number as random effect. The experimental unit for mineral content and antioxidant enzymes of newborn tissues was the individual piglet nested within the sampled sow and was analyzed considering the sow diet, BW piglet category, and the interaction between sow diet × BW piglet category as main factors. The normality and homogeneity of the data were examined using the Shapiro–Wilk test and assessing the normal plot before statistical analysis. The normal adjustment was carried out for sow performance data, such as BW, back-fat, feed intake, tissue mineral content, and reproductive performance. Significantly different means were found using Tukey adjust. Significance was declared at a probability P ≤ 0.05, and tendencies were considered when P-value was between >0.05 and <0.10.

Gene expression data were analyzed with ANOVA, using sow diet and BW piglet category as main factors. The Benjamini and Hochberg false discovery rate (FDR) multiple testing correction was also performed (Benjamini and Hochberg, 1995). The variability comparison between the experimental diet and the residual variability within diets is expressed as statistical contrast value. The differences between experimental diets were determined at P-values ≤ 0.10 and at FDR values ≤ 0.20. Heat map visualization was performed using the heatmap.2 function of the gplots package of R (Warnes et al., 2020).

Results

The analyzed TM content in the experimental diets is shown in Table 2. The expected differences in TM content between diets were achieved with all sources.

Table 2.

Analyzed TM content in gestation experimental diets¹

TM element, mg/kg	Experimental diets
	ITM	Replace	R&R
Zn	130	148	113
Cu	30	30	26
Mn	90	99	88

¹Feed samples, from each diet, were collected from four manufacturing batches and mixed into one representative sample. One feed sample from each diet was analyzed by duplicated. The total supplemented levels on ITM and Replace diets of Zn, Cu, and Mn with both mineral sources were 80, 15, and 60 mg/kg, respectively. The Zn, Cu, and Mn total supplemented levels on R&R diet were 64, 12, and 48 mg/kg, respectively.

Reproductive sow performance

Sow productive and reproductive performance are shown in Table 3. The number of sows allotted per experimental diet at farrowing was 78, 78, and 77 for ITM, Replace and R&R, respectively. Difference with the initial number of sows corresponds to sows that showed physiological issues: lameness injuries, sudden dead, and abortion. Since the present study aimed to reflect results under commercial conditions, it was difficult to obtain an equal parity number between the experimental diets. Therefore, this factor was considered within the statistical analysis. No differences (P > 0.10) were observed between experimental diets on sow BW, back-fat thickness, or average daily feed intake. It is necessary to emphasize that the reported feed intake during gestation phase is the real consumption including the individual adjustment, which is carried out as a common commercial practice in the management on gestating sows. Reproductive performance was not affected by experimental diets (Table 3).

Table 3.

Reproductive performance of sows fed diets with total ITM, Replace, and R&R

Item	Experimental diets1
	ITM	Replace	R&R	SEM	P-value
Sows, n	78	78	77	—	—
Parity, n	4.78	3.59	5.03	—	—
Sow BW, kg
Gestation (day 35)	248.83	241.17	244.59	3.394	0.253
Gestation (day 110)2	266.99	263.27	266.20	1.631	0.231
Farrowing standardized3	246.30	242.58	244.80	1.625	0.247
Sow back-fat, mm
Gestation (day 35)	14.54	14.53	14.41	0.423	0.968
Gestation (day 110)	10.41	9.96	9.83	0.222	0.121
Total feed intake, kg	177.13	178.68	176.3	0.923	0.192
Average daily feed intake, kg	2.33	2.35	2.32	0.012	0.191
Reproductive parameters
Total born piglets, n4	19.07	19.14	19.27	0.472	0.951
Piglets born alive, n	16.17	16.38	15.98	0.416	0.790
Born alive litter weight, kg	20.97	20.83	21.43	0.526	0.703
Born alive average BW, kg	1.33	1.30	1.34	0.027	0.608

1ITM, inorganic trace mineral; Replace, partial replacement of trace mineral source; R&R, reduce and replace mineral source.

²Sow BW at day 35 used as covariate.

³Sow BW at day 110 excluding litter birth weight.

⁴Total born includes born alive and stillborn piglets.

^a,bMeans within a row with different superscripts indicate significant differences (P < 0.05).

Mineral contents in sow and newborn piglet tissues and colostrum

Serum TM content of sows was not affected by experimental diets, except on Mn content (Table 4). Sows fed Replace and R&R diets had higher Mn content than sows fed ITM diet (P < 0.001). Colostrum Mn and Fe contents were higher in sows fed Replace diet compared to sows fed R&R (P < 0.05). Likewise, colostrum Cu content tended to be higher in sows fed Replace diet than that from sows receiving the other diets (P = 0.080). No differences in colostrum Zn content were observed (P > 0.10).

Table 4.

TM content of colostrum and serum of sows fed diets with total ITM, Replace, and R&R¹

Item	Experimental diets2
	ITM	Replace	R&R	SEM	P-value
Serum3
Zn, mg/L	0.68	0.77	0.79	0.048	0.187
Cu, mg/L	1.64	1.68	1.58	0.043	0.283
Mn, mg/L	1.50b	2.41a	2.85a	0.185	<0.001
Fe, mg/L	1.29	1.35	1.49	0.122	0.462
Colostrum4
Zn, mg/L	14.62	12.83	13.00	0.715	0.143
Cu, mg/L	4.40	4.54	3.85	0.238	0.080
Mn, µg/L	0.017ab	0.023a	0.016b	0.002	0.017
Fe, mg/L	1.41ab	1.63a	1.19b	0.071	<0.001

¹Data are means of 20 sows sampled per experimental diet.

2ITM, inorganic trace mineral; Replace, partial replacement of trace mineral source; R&R, reduce and replace mineral source.

³Serum at day 110 of gestation.

⁴Basis on fresh matter.

^a,bMeans within a row with different superscripts differ significantly at P < 0.05.

Piglets TM content was not affected by the two-way interaction between sow diet × BW piglet category (P > 0.10); therefore, only the main effects are shown in Table 5. Sow diet has no influence on TM content of piglet tissues, except for a tendency in tibia ash percentage. Piglets from sows fed ITM tended to have higher ash percentage than piglets from sow fed R&R diet (P = 0.079). The effect of BW category was evident on tissue TM content (P < 0.05). Light piglets had lower Mn serum content, and Zn, Cu, and Mn tibia content than average piglets (P < 0.05). Average piglets had lower tibia ash percentage compared to light piglets (P = 0.029).

Table 5.

Tissue TM content and liver antioxidant activity of newborn piglets of sows fed diets with total ITM, Replace, and R&R¹

Item	Experimental diets2				BW piglet category			P-value
	ITM	Replace	R&R	SEM	Light	Average	SEM	Diet	Category	Diet × category
BW, g	1,049	1,043	1,012	34.366	721	1,348	24.996	0.714	<0.001	0.806
Liver, mg/kg DM3
Zn	85.46	101.76	118.17	16.729	94.61	108.98	10.634	0.385	0.145	0.857
Cu	161.60	142.04	170.01	13.612	153.50	162.26	8.928	0.324	0.316	0.225
Mn	5.46	5.59	5.43	0.339	5.15	5.84	0.220	0.933	0.003	0.285
Fe	0.72	0.70	0.72	0.067	0.71	0.71	0.053	0.975	0.986	0.534
Tibia bone4
Zn, mg/tibia	98.40	86.37	84.31	6.387	66.31	113.07	4.461	0.253	<0.001	0.898
Cu, mg/tibia	1.40	1.39	1.26	0.099	0.93	1.76	0.074	0.537	<0.001	0.536
Mn, mg/tibia	5.34	5.02	4.62	0.284	4.03	5.97	0.216	0.220	<0.001	0.955
Ash, %5	60.72	59.42	59.09	0.575	60.42	59.07	0.418	0.079	0.029	0.852
Liver antioxidant
GPX, U/mg protein	0.59ab	0.65a	0.56b	0.027	0.61	0.60	0.021	0.046	0.627	0.956
MDA, µM6	29.52	31.70	30.47	1.729	30.89	30.23	1.170	0.666	0.601	0.411

¹Data are means of 10 newborn piglets for each BW piglet category and per sow experimental diet.

2ITM, inorganic trace mineral; Replace, partial replacement of trace mineral source; R&R, reduce and replace mineral source.

³DM, dry matter.

⁴Total TM expressed as total content of ash bone.

⁵Bone ash as percent of the weight of dried and defatted bone.

⁶MDA, malondialdehyde.

^a,bMeans within a row with different superscripts differ significantly at P < 0.05.

Oxidant and antioxidant activity

Piglets born from sows fed Replace diet had a higher GPX liver activity than piglets from sows fed R&R diets, whereas pigs from ITM sows were intermediate (P = 0.046; Table 5). No interactions were observed between sow diet × BW piglet category on the activity on GPX or MAD content in liver (P > 0.10; Table 5). Likewise, no differences were observed in GPX activity (0.61 vs. 0.60 U/mg protein) and MAD content (30.9 vs. 30.2 µM) between light and average piglets (P > 0.10; Table 5).

Jejunum gene expression

A total of 46 genes were successfully amplified from either ITM or Replace diet. As BW at birth is critical for the piglet development, differences in mRNA gene expression were examined between light and average littermates of ITM and Replace sows (Table 6). The number of genes which show differences (P < 0.10) in mRNA gene expression between light and average littermates was lower in piglets born from Replace sows (five genes) than in those from ITM sows (nine genes). In general, light piglets showed a downregulation of genes involved in barrier, immune, and digestive functions compared with their average littermates.

Table 6.

Relative gene expression differences between light and average BW newborn littermates of sows fed diets with total ITM and Replace¹

Experimental diets2	Function	Response	Gene	Light	Average	Contrast statistic3	P-value	FDR
ITM	Barrier function	Trefoil factor 3	TFF3	0.93	1.33	1.712	0.207	0.650
	Immune response	Fatty acid hydrolase domain containing 2	FAXDC2	1.70	3.12	8.356	0.010	0.138
		Guanylate binding protein 1	GBP1	0.32	0.06	3.944	0.062	0.319
		Heat shock protein 1	HSPB1	1.27	1.81	1.338	0.002	0.083
		Peroxisome proliferative activated receptor gamma, coactivator 1 alpha	PPARGC1α	2.87	4.79	7.799	0.012	0.138
		Interleukin 6	IL6	1.18	1.44	0.437	0.517	0.820
		Tumor necrosis factor alpha	TNFα	2.67	3.64	0.855	0.367	0.650
	Digestive enzyme	Aminopeptidase-N	ANPEP	0.61	0.72	4.528	0.047	0.311
		Indoleamine 2,3-dioxygenase	IDO1	1.40	4.72	4.973	0.040	0.310
		Sucrase-isomaltase	SI	7.27	55.70	9.834	0.006	0.131
	Nutrient transporter	Solute carrier family 39 (Zn transporter) member 4	SLC39A4	1.31	2.20	5.704	0.028	0.258
	Stress enzyme	Hydroxysteroid (11-beta) dehydrogenase 1	HSD11B1	2.88	1.92	3.970	0.062	0.319
Replace	Barrier function	Trefoil factor 3	TFF3	1.80	1.06	6.314	0.023	0.589
	Immune response	Fatty acid hydrolase domain containing 2	FAXDC2	3.18	5.58	2.395	0.139	0.589
		Guanylate binding protein 1	GBP1	0.19	0.51	0.491	0.493	0.839
		Heat shock protein 1	HSPB1	1.96	1.68	1.909	0.184	0.634
		Peroxisome proliferative activated receptor gamma, coactivator 1 alpha	PPARGC1α	4.06	4.98	1.612	0.220	0.634
		Interleukin 6	IL6	1.07	2.64	3.625	0.073	0.589
		Tumor necrosis factor alpha	TNFα	3.08	5.59	3.413	0.081	0.589
	Digestive enzyme	Aminopeptidase-N	ANPEP	0.82	0.79	0.026	0.873	0.990
		Indoleamine 2,3-dioxygenase	IDO1	3.34	8.33	1.735	0.204	0.634
		Sucrase-isomaltase	SI	9.52	50.86	3.977	0.062	0.589
	Nutrient transporter	Solute carrier family 39 (Zn transporter) member 4	SLC39A4	2.18	2.40	0.255	0.619	0.919
	Stress enzyme	Hydroxysteroid (11-beta) dehydrogenase 1	HSD11B1	2.11	2.92	4.033	0.060	0.589

¹Data are means of 10 newborn piglets for each BW piglet category and per sow experimental diet. Only significant gene expression differences are presented in both treatments (P < 0.10). Gene expression values are presented as ratios of cycle relative threshold value for each gene normalized to that of the reference sample.

2ITM, inorganic trace mineral; Replace, partial replacement of trace mineral source; R&R, reduce and replace mineral source.

³Contrast statistic expresses the variability comparison between the experimental diet and the residual variability within diets.

The influence of maternal dietary intervention was shown by the expression of 19 genes in neonate piglets (P < 0.10; Figure 1). Piglets born from sows fed Replace diet had a higher jejunum mRNA expression of genes of Barrier function (CLDN15, MUC2, MUC13, ZO1), Immune response (CXCL2, FAXDC2, HSPA4, HSPB1, IL-1β, IL8, REG3G, TGF-β1), Antioxidant function (GPX2, SOD2), Digestive enzymes (ALPI, ANPEP), Digestive hormones (CCK, IGF1R), and Nutrient transport (SLC16A1) compared with piglets born from ITM sows (P < 0.10; Figure 1).

Figure 1.

Heatmap and hierarchical clustering of gene expression levels of newborn piglets of sows fed diets with total ITM and partial replacement TM source (Replace). Significant differences are indicated using an asterisk symbol (P < 0.10). Data are means of 20 newborn piglets per sow experimental diet.

In order to explore the influence of sow diet on gene expression of each piglet BW category (light and average), the analysis of gene expression is presented separately. Light piglets born from ITM sows showed a lower mRNA expression of 15 different genes involved in different physiological functions such as Barrier function (CLDN15, MUC2, TFF3, ZO1), Immune response (FAXDC2, HSPB1, PPARGC1α, TGF-β1), Antioxidant function (GPX2, SOD2), Digestive enzymes (ALPI, ANPEP), Digestive hormones (CCK, IGF1R), and Nutrient transport (SLC39A4) compared with light piglets born from Replace sows (P < 0.10; Table 7).

Table 7.

Relative gene expression differences between light BW piglets of sows fed diets with total ITM and Replace¹

Function	Response	Gene	ITM	Replace	Contrast statistic2	P-value	FDR
Barrier function	Claudin-15	CLDN15	1.36	2.37	6.312	0.022	0.143
	Mucin 2	MUC2	0.87	1.79	10.848	0.004	0.054
	Trefoil factor 3	TFF3	0.93	1.80	13.598	0.002	0.042
	Zonula occludens 1	ZO1	1.06	1.34	4.020	0.060	0.267
Immune response	Fatty acid hydrolase domain containing 2	FAXDC2	1.70	3.17	10.410	0.005	0.054
	Heat shock protein 1	HSPB1	1.27	1.96	18.122	0.0005	0.022
	Peroxisome proliferative activated receptor gamma, coactivator 1 alpha	PPARGC1α	2.87	4.06	3.179	0.091	0.301
	Transforming growth factor beta 1	TGF-β1	1.42	2.03	9.581	0.006	0.057
Antioxidant enzyme	Glutathione peroxidase 2	GPX2	9.08	22.64	3.034	0.099	0.302
	Superoxide dismutase	SOD2	1.17	1.56	7.290	0.015	0.112
Digestive enzyme	Intestinal alkaline phosphatase	ALPI	1.23	1.76	3.595	0.074	0.284
	Aminopeptidase-N	ANPEP	0.61	0.82	3.252	0.088	0.301
Digestive hormone	Cholecystokinin	CCK	0.85	1.35	4.076	0.059	0.267
	Insulin-like growth factor 1 receptor	IGF1R	1.53	2.22	4.756	0.043	0.246
Nutrient transporter	Solute carrier family 39 (Zn transporter) member 4	SLC39A4	1.31	2.18	3.896	0.064	0.267

¹Data are means of 10 newborn piglets per each sow experimental diet. Only significant gene expression differences are presented (P < 0.10). Gene expression values are presented as ratios of cycle relative threshold value for each gene normalized to that of the reference sample.

²Contrast statistic expresses the variability comparison between the experimental diet and the residual variability within diets.

Likewise, average piglets born from ITM sows showed a lower mRNA expression of genes involved in Barrier function (MUC13, ZO1), Immune response (GBP1, HSPA4, IL-1β, IL8, IL6, TGF-β1), and Stress enzyme (HSD11B1) compared with average piglets from Replace sows (P < 0.10; Table 8).

Table 8.

Relative gene expression differences between average BW piglets of sows fed diets with total ITM and Replace¹

Function	Response	Gene	ITM	Replace	Contrast statistic2	P-value	FDR
Barrier function	Mucin 13	MUC13	0.75	0.95	4.597	0.046	0.424
	Zonula occludens 1	ZO1	1.14	1.39	3.368	0.083	0.424
Immune response	Guanylate binding protein 1	GBP1	0.06	0.51	4.567	0.047	0.424
	Heat shock protein 4	HSPA4	2.02	2.70	4.475	0.049	0.424
	Interleukin 1 beta	IL-1β	2.39	18.23	4.140	0.058	0.424
	Interleukin 8	IL8	1.74	3.59	3.463	0.079	0.424
	Interleukin 6	IL6	1.43	2.64	3.406	0.082	0.424
	Transforming growth factor beta 1	TGF-β1	1.58	2.43	4.554	0.047	0.424
Stress enzyme	Hydroxysteroid (11-beta) dehydrogenase 1	HSD11B1	1.92	2.92	6.058	0.024	0.424

¹Data are means of 10 newborn piglets per sow experimental diet. Only significant gene expression differences are presented (P < 0.10). Gene expression values are presented as ratios of cycle relative threshold value for each gene normalized to that of the reference sample.

²Contrast statistic expresses the variability comparison between the experimental diet and the residual variability within diets.

Discussion

Effects of feeding strategies on sow reproductive performance

Since, in the present study, TM diets were offered from day 35 postmating, when the embryonic implantation and consequently litter size had already been established, it was expected that maternal nutrition had effects on the development and survival (born alive) of piglets rather than in the total number of piglets born. Despite that the reproductive sow performance in the present study was not affected by experimental diets, a numerical decrease on live born piglets, or an increase in those stillborn, were observed on R&R sows. Coincidentally, these sows showed a moderate increase in parity number than Replace sows, which may have influenced this parameter besides receiving the reduced TM supplementation. However, the incidence of stillborn piglets should be considered as a multifactorial challenge with different noninfectious causes such as genetic lines, maternal, piglet, and environmental factors (Vanderhaeghe et al., 2013). Among these different noninfectious causes, the farrowing length, sow age, and proper nutrition may be of interest to the present study. An increase in litter size has been associated with a prolonged farrowing, which, in turn, is directly related to an increased risk of hypoxia in piglets (Vanderhaeghe et al., 2013). This prolonged farrowing may be attributed on the one hand to sow parity: as parity is increased, there is a greater risk of sow having stillborn piglets (Vanderhaeghe et al., 2013), presumably attributed to a loss on uterine muscle tone due to poor mineral reserves or oxytocin secretion (Langendijk and Plush, 2019). Although some studies did not confirm this relationship (van Dijk et al., 2005; Oliviero et al., 2009), others suggested that the duration of farrowing may have a greater impact on stillbirths than parity itself (Vanderhaeghe et al., 2013; Udomchanya et al., 2019). On the other hand, in terms of nutrition, for both young and old sows, a proper supply of macro and micro-minerals is considered important for the effective contractions of muscles surrounding the uterus, improving the transport of fetuses through the birth canal (NRC, 2012; Theil, 2015). Although TM requirements for sows are not well established, another possible reason for this decreased alive birth rate may be the low dosage of TM supplied in the R&R sow diet. In this sense, Vallet et al. (2014) showed that the dietary supplementation of Zn (700 mg/kg), at higher levels than those recommended, decreased stillbirth rate from late gestation (day 80), probably by decreasing birth intervals. Regarding the influence of TM source, early studies reported that a complete supplementation of OTM in sow diets increased the total number of piglets, as well as live born piglets at farrowing, compared with the ITM (Peters and Mahan, 2008; Caine et al., 2009). Whereas Holen et al. (2020) reported no differences in the number of stillborn, born alive, or total born piglets when diets were supplemented with three Zn levels (125, 365, and 595 mg/kg) by combining two Zn sources (sulfate and chelated). Similarly, strategies such as replacing 50% of ITM by OTM (Tsai et al., 2020) or reducing the total dose of ITM to 80% by a complete supplementation with OTM (Ma et al., 2020) did not show differences on litter size parameters compared with ITM.

As mentioned before, one of the major challenges associated with large litters are the greatest presence of low-birth-weight piglets and the variability within the litter. In the present study, no influence of TM strategies was observed on litter weight nor average live piglet weight. However, recent studies shown equivocal outcomes. Supplementing Zn at higher levels (365 mg/kg) than in the present study (64 to 80 mg/kg), during late gestation, appears to enhance piglet birth weight (Holen et al., 2020). The authors showed that a mixture of Zn sulfate with Zn chelate at 365 mg/kg resulted in heavier piglets (1.42 kg) and less incidence of low-birth-weight piglets (11.6%) compared with those sows fed a mixture of Zn sulfate and chelate at 125 (1.38 kg; 15.3 %) or 595 (1.40 kg; 15.1 %) mg/kg. On the contrary, Ma et al. (2020) supplementing sows, from breeding until 21 d postpartum, with a reduced OTM diet of Zn (72 mg/kg), Cu (12 mg/kg), Mn (20 mg/kg), and Fe (72 mg/kg) did not observe differences in the weight at birth of the pigs, except for a greater number of piglets with a birth weight greater than 1 kg (10.47 vs. 9.83) compared with sows fed a complete ITM diet (Zn: 90, Cu: 15; Mn: 25; Fe: 90 mg/kg). Likewise, Tsai et al. (2020) supplementing sows the entire gestation period shown no effect of Zn (120 mg/kg), Cu (30 mg/kg), and Mn (50 mg/kg) as ITM or partially replaced by OTM (50%) on litter birth weight nor the number of piglets weighing less than 0.91 kg. Therefore, further studies should be conducted to investigate the optimal levels or combinations of TM to enhance neonate weight and survival.

In the present study, an improvement of some Mn, Fe, and Cu levels in serum and colostrum were observed in OTM replacement sows. During the perinatal phase, the nutrition of piglets relies entirely on the nutrition of the sow through placental and colostrum transfer. In hyperprolific situations, this perinatal transfer of nutrients might not be optimal (Mahan et al., 2009; Matte and Audet, 2020), and according to Matte and Audet (2020), there is an active placental transfer of Zn but limited for other TM, such as Fe, Cu, and Se. In the present study, no differences were observed between maternal diets in neonatal piglet TM content. In the same way, Peters et al. (2010) indicated that the transfer of minerals to the fetus in utero was not greatly affected by the dietary source (ITM and OTM), or level of TM provided to sow. On the other hand, the higher mineral content in colostrum and serum from OTM replacement sows may suggest an improved postnatal transfer of nutrients for neonatal development compared to those fed the R&R diet. Similarly, Peters et al. (2010) reported that the OTM supplementation in sow diets resulted in the increase of Cu, Se, P, and Mg colostrum content compared with the ITM diets. Interestingly, although the dietary Fe source and level were kept constant in the three experimental diets, the colostrum Fe level was greater in replacement sows. Although, each TM has a specific pathway of intestinal absorption, a competition for DMT1 transporter may occur between Fe, Zn, Cu, and Mn (Goff, 2018). Our finding may suggest a greater competition for transporters between ITM than with their chelated counterparts.

Since the present study was focused only on gestating sows, the performance of the sow and litter during lactation phase was not recorded. Further complementary studies that also consider the lactation phase could provide more information on the midterm effects of these dietary strategies on sow and piglet performance.

Differences between light and average BW piglets

It was observed that the tissue mineral content and the jejunum gene expression differed according to the BW of the neonate piglets, with light piglets being unable to store TM and express genes as effectively as their heavier littermates. According to Mahan et al. (2009), as litter size increases, large quantities of minerals would be expected to be transferred from the sow to litter, but this transfer may differ according to the neonatal BW. Indeed, it is widely accepted that intrauterine embryo crowding in large litters implies that the first embryos to implant can limit the physiological development of the embryos that adhere later, with this embryonic competition dramatically increasing with each successful embryo attachment (van der Waaij et al., 2010). The impaired physiological development of light piglets was also observed through the deficient expression of several genes implicated in metabolic and immune responses. A compromised gut can not only reduce the absorption of nutrients, but it also impairs the mucosal immune system, which is necessary to defend and restore the body’s homeostasis (Weström et al., 2020). In fact, among the frequent complications of low-birth-weight piglets are, among others, the poor development of several organs and reduction in the wall thickness of gastrointestinal tract (Edwards et al., 2019). In humans, intrauterine growth restricted placentas have showed a reduced content of some TM such as Cu (23%) and Zn (37%) compared to normal neonate placentas (Zadrozna et al., 2009). Likewise, in pigs, the intrauterine growth restricted placentas have shown a greater oxidative damage and impaired angiogenesis compared with those heavier littermates (Hu et al., 2020). Overall, the outcome from the biomarkers evaluated in this study emphasizes the important maternal–fetal relationship, and in which the imbalance induces digestive and immunological disorders, particularly in light birth weight neonates.

Effects of maternal sow diet on piglet gene expression

Maternal nutrition not only plays not only a critical role in fetal growth and development but also is the major factor that alters expression of the fetal genome and hence may have lifelong consequences (Wu et al., 2004). The partial substitution of ITM by OTM in sow diets resulted in an upregulation of 15 genes implicated in intestinal barrier, immune, and antioxidant responses, particularly in light newborn piglets. Since Zn, Cu, and Mn are essential micronutrients involved in all major metabolic pathways (Suttle, 2010), it is possible that an improved maternal nutrition may influence the function of these TM on the progeny. Previous studies feeding sows with several types of nutrients (De Vos et al., 2014), as well as with different levels and sources of TM (Holen et al., 2020; Ma et al., 2020), have described positive effects on the birth weight of the piglets. However, none have described the positive effects that these maternal dietary interventions have on the fetal imprinting on the smallest piglets of the litter. The influence of maternal dietary intervention on gut health genes will be discussed in the following sections, according to their main physiological function.

An upregulation of genes involved in barrier functions (MUC2, MUC13, ZO1, TFF3, CLDN15) was observed in light and average neonate piglets born from sows fed Replace diet. An adequate intestinal barrier function involves multiple intestinal mechanisms, to ensure the defense against pathogenic microorganisms, such as immune cells, antimicrobial peptides, tight junctions, and mucins (Wijtten et al., 2011). Mucins and linker proteins like ZO-1, in particular, play an important role in the protective barrier within the gastrointestinal tract, safeguarding the intestine and its structural integrity (Zhang and Guo, 2009; Liu et al., 2014; Pelaseyed and Hansson, 2020). Among all TM, perhaps Zn is the most widely investigated with consistent evidence of its vital role in the regulation of intestinal integrity and immune function in live beings (Liu et al., 2014; Chasapis et al., 2020). For instance, in female broiler breeders, the Zn supplementation in chelated form increased the MUC2 expression in progeny (Li et al., 2015).

In general, as the intestinal mucosa possesses a complex function, it does not only play a role in nutrient absorption and epithelial barrier but is also part of a well-organized immune system (i.e., gut-associated lymphoid tissue; Ramiro-Puig et al., 2008; Weström et al., 2020). GBP1 gene is involved in the control of inflammation and host immune response (Honkala et al., 2020), particularly against several bacterial, parasite (Selleck et al., 2013; Fisch et al., 2019), and viral pathogens (Khatun et al., 2020). Since the intestinal GBP1 gene expression levels were similar in light piglets from both ITM and OTM diets, it seems that GPB1 gene expression is especially required under critical circumstances. Moreover, when comparing the expression of GBP1 gene between light and average littermates under ITM conditions it was also higher in light piglets. Whereas OTM diet seems to be able to induce an upregulation of this defensive molecule even in those average piglets, suggesting a strengthening immune effect until later in life. This immunoregulatory effect by the OTM diet is also observed for TGF-β1 gene expression, a key mediator molecule in the mucosal immune system (Konkel and Chen, 2011), which was upregulated in both light and average piglets compared with those in the ITM group.

Moreover, the partial feeding of OTM in sows not only resulted in an upregulation of the above anti-inflammatory genes (GBP1, TGF-β1), but also of the proinflammatory ones (IL-1β, IL-6) in average piglets. A balance between the expression and function of anti-inflammatory and proinflammatory cytokines is necessary to rapidly detect invaders (proinflammatory) and to initiate an appropriate immune response (balanced proinflammatory and anti-inflammatory responses; Cicchese et al., 2018). For instance, in weaned piglets, the pharmacological doses of Zn downregulated the proinflammatory cytokine IL-1β and IFN-γ, while increasing the expression of TGF-β (Zhu et al., 2017).

Although the upregulation of anti-inflammatory and proinflammatory genes was observed in average piglets of both diets compared with their light littermates, the average OTM piglets seems to upregulate more anti-inflammatory genes than proinflammatory genes compared with the ITM piglets. Complementary studies focused on the specific influence of the dose and chemical form of TM in the stimulation or suppression of the expression of immune cells could help to decide the precise dose and form of TM to strengthen the immune system.

Interestingly, changes in the genes encoding antioxidant enzymes were observed only in light piglets, being upregulated in those born from Replace sows. Three key enzymes (superoxide dismutase, catalase, and GPX) constitute the first line of defense in neutralizing any free radicals (Ighodaro and Akinloye, 2018). TM such Zn, Cu, and Mn are extensively recognized as cofactors and constituents of these antioxidant enzymes (Suttle, 2010). In the present study, the increased activity of the GPX enzyme in the liver, along with upregulation of GPX2 and SOD2 mRNA levels in the intestine, may suggest an improved antioxidant capacity in piglets born from sows fed OTM, particularly in light piglets, compared with those born from ITM sows.

In the same way, changes in mRNA levels of digestive response genes were only detected in light newborn piglets. An upregulation of several digestive response genes (ALPI, ANPEP, CCK, IGF1R, SLC39A4) were observed on light piglets born from Replace sows, compared with those from ITM sows. In particular, IGF-1 is known as the main mediator of the effects of the growth hormone, but also for its effects on the improved absorption of nutrients and electrolytes, and stimulation of the recovery of the intestinal epithelium (Li et al., 2016). Whereas the SLC39A4 (ZIP4) gene is considered as the crucial factor for Zn dietary uptake from the gut lumen and in the control of the systemic Zn homeostasis (Lichten and Cousins, 2009; Hashimoto et al., 2016). The increased expression of this Zn transporter gene in light piglets born from the Replace OTM may be in line with the highest mRNA levels of all the genes mentioned above that need Zn for their expression.

In the light of the present results, it is concluded that the partial substitution of inorganic TM by their organic counterparts represents an alternative to the totally inorganic supplementation, with positive effects on neonate piglet gene expression. Although the lower TM storage, together with the greater downregulation of physiological genes, exposes the immaturity and vulnerability of the smallest neonates of the litter to cope with the challenges after birth, the partial supplementation of maternal diets with OTM may denote a nutritional strategy to mitigate the negative effects of the intrauterine fetal growth competition.

Glossary

Abbreviations

ANPEP

aminopeptidase-N

BW

body weight

CP

crude protein

FDR

false discovery rate

MDA

malondialdehyde

NDF

neutral detergent fiber

NE

net energy

OTM

organic trace minerals

PCR

polymerase chain reaction

TM

trace minerals

Conflict of interest statement

The authors declare no real or perceived conflicts of interest.