Published December 12, 2014

Transcription

1 Published December 12, 2014 Effect of inorganic or organic selenium supplementation on reproductive performance and tissue trace mineral concentrations in gravid first-parity gilts, fetuses, and nursing piglets1, 2, 3 Y. L. Ma,* 4 M. D. Lindemann, *5 J. L. Pierce, *6 J. M. Unrine, and G. L. Cromwell* *Department of Animal and Food Sciences, University of Kentucky, Lexington 40546; and Department of Plant and Soil Sciences, University of Kentucky, Lexington ABSTRACT: The objective of this experiment was to evaluate 2 supplemental forms of Se on reproductive performance and tissue trace mineral concentration in fetus and first-parity gilts during pregnancy and their progeny. Crossbred gilts (n = 100) were selected at 183 ± 2.7 d and 137 ± 10 kg BW and fed a common diet. After 1 mo, 8 gilts were sacrificed to establish baseline liver Se concentration and the remaining 92 gilts allotted to receive Se (0.3 mg/kg diet) as inorganic Se (Na 2 SeO 3 ) or a Se supplement that contains organoselenium compounds (Sel-Plex; Alltech Inc., Nicholasville, KY). At 267 ± 5.7 d (171 ± 11 kg), gilts were estrus-synchronized and bred. Gilts were then slaughtered at defined time points throughout gestation (d 0, 43, 58, 73, 91, 101, or 108 of gestation; n = 6 to 12 gilts/time point). A week before the expected farrowing day, 10 pregnant gilts (5 from each treatment) were moved to farrowing crates and monitored. Two pigs from each litter were randomly selected and euthanized at d 0 (within 2 h after birth; nursing deprived), 7, 14, and 21 from each litter. During the gestation phase, maternal liver, and fetal body and liver were collected for determination of trace mineral concentration by inductively coupled plasma mass spectrometry. Total number of fetus, crown-rump length, and corpora lutea of gilts were recorded as well. During the lactation phase, pigs (without liver and gastrointestinal tract) and associated liver were analyzed for Se concentration. The results demonstrated that the source of Se generally did not affect the maternal reproductive traits and fetal characteristics. Also, the source of Se supplemented to the maternal diet did not, in general, affect Cu, Fe, Mn, or Zn concentrations in the tissues evaluated other than the observation of a greater maternal liver Mn content (P < 0.01) in gilts fed Sel-Plex and a greater amount of Fe accumulated in the entire litter (P < 0.01) in gilts fed Sel-Plex. However, with regard to Se concentrations, Se in fetal body, fetal liver, and maternal liver were greater (P < 0.01) when Sel-Plex was fed. Postnatal pigs from gilts fed Sel-Plex had greater (P < 0.05) Se retention in body and liver with similar growth performance during the 21-d period. The results demonstrate Se form differences wherein Sel-Plex is associated with greater Se accumulation in both maternal and fetal tissues. Key Words: fetus, gestation, postnatal pig, reproduction, selenium 2014 American Society of Animal Science. All rights reserved. J. Anim. Sci : doi: /jas This manuscript is based on research supported in part by the Kentucky Agricultural Experiment Station (Lexington) and it is published by the Kentucky Agricultural Experiment Station as paper Appreciation is expressed to Alltech Inc. (Nicholasville, KY) who provided partial support for the project. 3 Appreciation is expressed to H. J. Monegue, B. G. Kim, A. D. Quant, J. S. Monegue, I. F. Hung, and N. Inocencio (University of Kentucky, Lexington) for assistance in the care of pigs and laboratory analysis; D. Higginbotham (University of Kentucky, Lexington) for help in diet preparation; and Akey Inc. (Lewisburg, OH), APC Inc. (Ankeny, IA), and DSM Nutritional Products Inc. (Parsippany, NJ) for ingredients used in the diets. 4 Current address: Novus International, Inc., 20 Research Park Dr., St. Charles, MO Corresponding author: [email protected]. 6 Current address: Nutriad Inc., 401C Airport Road, Elgin, IL Received January 13, Accepted October 12, INTRODUCTION Vitamin E-selenium deficiency symptoms are still reported in swine, particularly in the progeny of highproducing sows. Mahan and Kim (1996) stated that Se deficiency seems to be more prevalent from those regions where sodium selenite serves as the major contributor of total dietary Se, which implies that selenite may not be as biologically effective as the indigenous Se in grains or grain byproducts, which is primarily selenomethionine. The benefits of feeding of Se-enriched yeast have been demonstrated with relatively consistent results in

2 Selenium source for sows 5541 reproducing females. Sows fed organic Se have a lower number of stillborn pigs, higher Se content in colostrum and milk, and greater Se concentrations in neonatal pigs and liver (Mahan and Kim, 1996; Mahan, 2000; Mahan and Peters, 2004; Yoon and McMillan, 2006; Peters et al., 2010) compared with sows fed sodium selenite; however, there have been no reports that evaluate the effects of organic Se on fetal characteristics, sow reproductive performance, and Se retention in fetal and maternal tissue during gestation. Therefore, the objective of the current experiment was to evaluate an organic form of Se on reproductive performance during the gestation and lactation phases and fetal and maternal tissue Se, Cu, Fe, Mn, and Zn concentrations when that Se was supplied at the maximum allowable level of 0.3 ppm supplemental Se. MATERIALS AND METHODS The experimental use of animals and procedures followed for their management, the collection of tissues, and slaughter procedures were conducted under protocols approved by the Institutional Animal Care and Use Committee of the University of Kentucky. Pretrial Conditions This experiment utilized a total of 100 crossbred gilts (Yorkshire Duroc; [Yorkshire Landrace] Duroc; [Yorkshire Duroc] Chester White; [Yorkshire Landrace Duroc] Chester White) at about 6 mo of age and approximately 130 kg. These gilts were examined for structural soundness and general health. Before being fed the experimental diets, gilts were fed a common finishing diet containing 0.3 ppm supplemental Se from sodium selenite for about 1 mo. Eight gilts were sacrificed and liver samples were collected for Se analysis to establish the initial Se status in the body. The remaining 92 gilts were then randomly allotted by body weight and ancestry to 2 treatment diets. Experimental Treatments and Animal Allotment The 2 dietary treatments were a basal corn soybean meal diet supplemented with 0.3 mg/kg Se from inorganic source (Na 2 SeO 3 ) or a Se supplement that contains organoselenium compounds (Sel-Plex; Alltech Inc., Nicholasville KY), a selenium-enriched yeast product containing various organic seleno amino acid analogs, the principal one being selenomethionine (Kelly and Power, 1995). All standard trace minerals were supplied as organic minerals (Bioplex Cu, Fe, Zn, Mn; Alltech Inc., Nicholasville KY) at 150% of NRC (1998) requirements with the exception of I. For the macrominerals, Table 1. Ingredient composition of the experimental breeding diet (%, as-fed basis) Ingredient % Corn, ground Dehulled soybean meal, 48% CP Mono-dicalcium phosphate 1.60 Limestone, ground 0.85 Salt 0.50 Choline mix (60% choline chloride) 0.18 Vitamin premix Trace mineral premix Antibiotic Calculated composition ME, kcal/kg 3,272 CP, % True ileal-digestible protein, % Lys, % 0.62 True ileal-digestible lysine, % Met, % 0.24 True ileal-digestible methionine, % TSAA, % 0.51 True ileal-digestibile TSAA, % Ca, % 0.75 P, % 0.62 Available P, % Supplied per kilogram of diet: vitamin A (acetate), 6600 IU; vitamin D 3 (cholecalciferol), 880 IU; vitamin E (DL-a tocopheryl acetate), 44 IU; vitamin K (as menadione sodium bisulfite complex), 6.4 mg; thiamin, 4.0 mg; riboflavin, 8.8 mg; pyridoxine, 4.4 mg; vitamin B 12, 33 μg; folic acid, 1.3 mg; niacin, 44 mg; pantothenic acid, 22 mg; and D-biotin, 0.22 mg. 2 Trace minerals were provided at 150% of NRC (1998) requirement estimates for Fe, Zn, Cu, and Mn. Supplied per kilogram of diet from Bioplex: Cu, 7.5 mg; Fe, 120 mg; Mn, 30 mg; Zn, 75 mg; whereas I (calcium iodate), 1.0 mg, and Se (sodium selenite or Sel-Plex), 0.3 mg. 3 Aureomycin50 granular (Alpharma Inc, Bridgewater, NJ) supplied 441 mg chlortetracycline per kilogram of diet. 4 The true ileal protein digestibilities of corn and soybean meal were adapted from Stein et al. (2001) and the ileal lysine, methionine, as well as TSAA digestibilities of corn and soybean meal were adapted from NRC (1998). salt was included in the diet at 0.50% and dietary Ca and P formulated to 0.75 and 0.62%, respectively. Amino acids and energy were at or above NRC (1998) requirement estimates (Table 1). The experimental diets contained 3.27 Mcal/kg of ME and 0.62% total lysine. The basal diet without Se premix was mixed homogenously, subdivided, and then Se premix was added to the basal diet to obtain the 2 dietary treatments. Prebreeding, gestation, and lactation were the same diet formulation with the exception of the antibiotic; the diet contained chlortetracycline at 55 mg/kg diet at all times other than the period of 3 wk prebreeding to 1 wk prebreeding (no antibiotic) and 1 wk prebreeding to 2 wk postbreeding (441 mg/kg diet). Experimental diets were sampled at 6 time points across the experiment; the mean trace mineral concentrations are presented in Table 2.

3 5542 Ma et al. Table 2. Trace mineral requirements (NRC, 1998) and analysis of those trace minerals in the experimental diets (mg/kg, as-fed basis) 1 Treatment Mineral NRC (1998) Supplemented Inorganic Organic Fe ± ± 36.9 Zn ± ± 5.9 Cu ± ± 0.5 Mn ± ± 2.3 Se ± ± Each value is mean ± standard deviation of 6 samples from 6 different time points during the experiment. The total of 92 gilts, at 6 mo of age (183 ± 2.7 d) with an initial BW of 137 ± 10 kg, were housed individually in gestation crates and had free access to water throughout the gestation period. Gilts were fed a single meal of 2.73 kg/d (as-fed basis) of the gestation diet, which provided 14.5 and 329 g/d of true ileal-digestible lysine and protein, respectively. At 9 mo of age (267 ± 5.7 d; 171 ± 11 kg), gilts were estrus-synchronized with altrenogest (the product Matrix, concentration of altrenogest was 0.22%, Intervet/ Schering-Plough Animal Health, Millsboro, DE) that was topdressed on their daily feed at a rate of 6.8 ml for 14 d. The following day, the gilts were checked by boars to determine onset of estrus and, on observation of estrus, bred by artificial insemination with Duroc semen (Swine Genetics International, Cambridge, IA) a minimum of 2 times on consecutive days. Bred animals continued on their respective dietary treatment and were slaughtered at defined time points throughout gestation. Lactation Phase A week before the expected farrowing day, pregnant gilts (n = 10; 5 from each treatment) were moved to farrowing crates and monitored. Gilts were provided a single meal from their respective gestation diets until farrowing. After farrowing, they were given ad libitum access to feed and water until weaning. Two pigs from each litter were randomly selected and euthanized at given time points from birth to weaning. Slaughter Time Points Three gilts from each treatment were slaughtered at mating to provide prebreeding baseline tissue composition values. Bred females that conceived were assigned randomly to be slaughtered at d 43, 58, 73, 91, 101, or 108 of gestation from each dietary treatment with a postassignment assessment to assure that unequal ancestry balance had not occurred. Four gilts per diet were assigned to be slaughtered at d 43 and 6 gilts per diet at other slaughter dates; both number of gilts and exact slaughter date was allowed to vary by ±1 to accommodate scheduling through the abattoir in relation to the breeding dates of the females. Females that were bred but did not conceive (open gilts) continued on their respective dietary treatment and were assigned to a final slaughter date (d 120) to provide data on the difference in maternal tissue nutrient content simply due to pregnancy compared with gilts slaughtered at d 101 and 108. During lactation, 2 pigs from each litter were randomly selected and euthanized at d 0 (from those delivered within 2 h after the initiation of birth; pigs were nursing deprived), 7, 14, and 21 of age. As with gestating gilts, slaughter age of the pigs was allowed to vary by ± 1d (except for d 0) to accommodate scheduling of labor needs relative to the exact age of the pigs. Slaughter and Sample Collection Upon reaching their designated harvest date (±1 d), gilts were transported to the University of Kentucky Meat Laboratory in Lexington at 0600 in the morning and then slaughtered at 0800, in compliance with standard UK Meats Laboratory procedures. Gilts were weighed, stunned by electric shock, and killed by exsanguination. After they were killed, livers were collected and weighed during the slaughter process. The ovaries were collected and the number of corpora lutea, as an indicator of ovulation rate, was visually counted and recorded by a minimum of 2 individuals. Reproductive tracts were also collected. The uterus was removed intact and placed on a table, allowing for both horns to be placed in the original orientation within the animal body to orient the fetus position from the right (R) side or the left (L) side and numbered from the anterior tip of each horn (i.e., as R1, R2, L1, L2, etc.). Placenta units (a placenta associated with membranes, fluid, and an individual fetus) were separated from the uterine wall and fetuses were then removed from the placenta units by cutting the top of the amniotic sac of the placenta and the base of the umbilicus. Fetal numbers, position, gender, and weight were recorded. Crown rump length was measured from the crown of the head to the base of the tail (rump). After the measurement of fetal length, an incision was made to expose internal organs of fetuses (other than the d 43 fetuses). The liver and the gastrointestinal tract (GIT) from the fetal bodies were collected via blunt dissection to ensure proper organ separation. The weights of fetal tissues (i.e., liver and GIT) were recorded and fetuses from each litter were collectively inserted into a plastic bag. The whole empty placenta and uterus were weighed; 3 placenta were randomly chosen and composited for a placenta sample for each litter and a portion of the midsection of each uterine horn of about

4 Selenium source for sows g was excised and the 2 samples composited for the uterus subsample. During the lactation phase, pigs were euthanized with sodium pentobarbital and liver and GIT were removed from the body. All samples were placed on ice and frozen (-20 C) until they were later processed. Tissue and Diet Sample Analysis All tissue samples and the whole fetal body were ground and thoroughly mixed (TorRey model M22-R-2, TorRey USA, Houston, TX) using an 82.6-mm kidney plate (TorRey model TOR 223KP) followed by a 15.9-mm plate (TorRey model TOR 12P 5/8) and then a 4.8-mm plate (TorRey model TOR 12P 3/16). Subsamples were taken and put into plastic containers and placed into the freeze dryer (Botanique model 18 DX48SA, Botanique Preservation Equipment, Inc., Peoria, AZ) for 12 to 14 d to a constant weight by lyophilization. All samples were removed from the dryer and weighed to determine DM content. After lyophilization, a small food processor (HC 3000, Black & Decker Inc., Shelton, CT) and a coffee grinder (Proctor Silex E160B, Hamilton Beach Brands Inc., Washington, NC) were used to finely grind the samples for Se and other trace mineral analysis. After fine grinding, 3 random individual fetal bodies were pooled from the same litter to create 1 fetal body sample; the same pooling procedure was used for an associated fetal liver sample. Pooled fetal body, fetal liver, and sow liver samples were subsequently analyzed for Se, Cu, Fe, Mn, and Zn (postnatal pigs and liver were only analyzed for Se concentration). Before analysis, samples were microwave digested with trace-metal grade nitric acid in sealed Teflon bombs (MDS- 2000, CEM Corporation, Matthews, NC) and appropriately diluted according to USEPA method 3052 (EPA, 1996). Analysis was conducted using an inductively coupled plasma mass spectrometer (ICP MS; Agilent 7500cx, Santa Clara, CA) using National Institute of Standard and Technology (NIST) traceable standards (Inorganic Ventures, Christiansburg, VA) and standard reference tissue (bovine liver; NIST Standard Reference Material 1577c). The ICP MS was calibrated using the method of standard additions to account for matrix interferences. Quality control and analysis procedures generally followed USEPA method 6020a (EPA, 1998). The procedure for diet Se analysis was the same as tissue Se analysis (Stadlober et al., 2001) except for lyophilization. Following the mineral analysis, a mean whole-fetal-body mineral value was calculated using the fetal body and liver values based on the relative proportion by dry weight of liver to whole body. Statistical Analysis The data were analyzed as a completely randomized design. Analyses were performed using the GLM procedure of SAS (SAS Inst., Inc., Cary, NC). Gilt and associated litter was the experimental unit. The model included day of gestation and source of Se as the main effects to separate the means between treatments. Least squares means, the LSD procedure, and pooled standard errors were used to evaluate the differences between means. Data were also analyzed using PROC REG of SAS, with the forward option to describe the quantitative (linear, quadratic, or cubic) changes of weight or length of each tissue as day of gestation progressed. The model was: y x x x e 2 3 = b0 + b1 + b2 + b3 + ; where y is a variable, x is a day of gestation, b 0 is an intercept, b 1 is a coefficient for linear regression, b 2 is a coefficient for quadratic regression, b 3 is a coefficient for cubic regression, and e is an experimental error. With regard to maternal tissue responses over time, data from d 0 unbred gilts was excluded because of qualitative differences between bred and unbred gilts that may have affected the response. Contrasts in nutrients between open gilts and d-101 and d-108 pregnant gilts provided the difference due to pregnancy alone. Variables for which variation could be explained by quadratic or cubic regressions were further analyzed to find the breakpoint (day of gestation) where the rate of accretion changed at a = The SAS NLIN 2-slope, straight broken-line procedure was used to obtain those multiphasic regressions and breakpoints (Robbins et al., 2006). RESULTS Gilts and Litters Final numbers of gilts in each time point varied between 6 and 12 (total 88 gilts) due to pregnancy failure in some gilts (Table 3). Generally, the number of gilts from each slaughter group was divided equally into 2 groups from the 2 dietary treatments: organic Se (Sel- Plex) and inorganic Se (Na selenite). Reproductive Characteristics The effects of Se source, when that Se was supplied at the maximum allowable level of 0.3 ppm supplemental Se, on reproductive traits, maternal tissues, and fetal characteristics are shown in Tables 4 and 5. Generally, there were no differences (P > 0.10) due to the source of Se in the diet on reproductive traits of gilts, maternal tissue changes, and fetal characteristics except for a tendency related to weight gain (Sel-Plex vs. inorganic, vs kg; P = 0.06), ratio of placenta weight to uterine weight (Sel-Plex vs. inorganic, 0.69 vs.0.80; P < 0.01), and placental efficiency (a ratio of fetal weight

5 5544 Ma et al. Table 3. The number of gilts allotted per day of gestation and the number of fetuses collected in organic and inorganic Se treatments Day of gestation 1 Lactation Variable Open 3 Farrowed No. of gilts Sel-Plex Inorganic No. of fetuses Sel-Plex Inorganic No. of fetuses/gilt Sel-Plex Inorganic SE Sel-Plex Inorganic Days of gestation indicated above were the average of defined slaughter period; individual gestation day of gilt = mean ± 1 d. 2 Gilts in the d 0 unbred group were slaughtered at the average of 7 d after first gilt was bred. 3 Gilts, bred but not pregnant, were slaughtered at the average of 120 d after the first gilt was bred. to placenta weight; Sel-Plex vs. inorganic, 2.19 vs. 1.95; P < 0.01). Specific weight comparisons across the designed slaughter points is shown in Table 5. In general, gestational weight gains from inorganic Se at different slaughter points were greater than those from Sel-Plex; however, none were significant. Regarding the ratios of placenta to uterine weight, inorganic Se values were also greater than those from the Sel-Plex gilts, especially in d 101 and 108 (P < 0.05). On the contrary, placental efficiency values from Sel-Plex were greater than those of inorganic Se in late gestation, especially for d 101 (P < 0.01). As would be expected, day of gestation caused a dramatic change in maternal reproductive traits (Table 4). Trace Mineral Content in Maternal and Fetal Livers Tables 6 to 11 provide various gestation mineral responses. These 6 tables are 3 sets of 2 tables in which the first table provides the overall gestation response. For all response measures that had P < 0.10, the second table evaluates the response across the gestational ages. It is understood that dividing the total gestational sample number into 6 gestational ages reduces the statistical power of each individual gestational age, which will increase the P-value for many of the individual ages. This is done, however, to evaluate potential patterns of mineral accumulation across gestation. The effects of Se source on trace mineral concentrations during gestation in maternal and fetal livers are shown in Tables 6 and 7. Maternal liver Se concentrations in gilts fed Sel-Plex Table 4. Effect of Se source on reproductive traits, maternal tissue, and fetal characteristics during gestation 1 Diet P-value 3 Variable Sel-Plex Inorganic Se rmse 2 Diet Day No. of total gilts No. of culled gilts No. of bred gilts No. of conceived gilts Conception rate, % Diet Day Breeding age, d < Breeding BW, kg Slaughter BW, kg < Weight gain, kg < Liver, kg < Kidney, g Ovaries, g Uterus, kg < Total placenta, kg < Placenta/uterus <0.01 < Corpora lutea, n Total fetuses, n Live fetuses, n Total survival, % Live survival, % Live/total, % Individual Fetus, g < Fetal litter, kg < Fetal length, cm < Fetal liver, g < Fetal GIT, g < Fetal placenta, g < Placental efficiency <0.01 < All the data were pooled across gestation age from the same diet treatments (the number of maternal tissue: organic vs. inorganic, 40 vs. 38, including d 0 and open gilts; the number of fetus and placenta: organic vs. inorganic, 33 vs. 32, excluding d 0 and open gilts). 2 rmse = Root mean square error; when divided by the square root of the number of observations, provides the standard error associated with each mean. 3 Diet effect = contrast between organic and inorganic treatment; day = contrast among different days of gestation; diet day = interaction between the source of Se and day of gestation. In this table, the main comparison was the effect from the different sources of Se. 4 Culled gilts = gilts were removed from the trial due to death or disease. 5 Kidney was collected only from left side of gilts. 6 Fetal survival percentages were calculated using both the total number and the number of presumably live fetuses compared with the total number of corpora lutea counted. 7 Live/total was the percentage of presumably live fetuses compared with the total number of fetuses. 8 Fetal liver and gastrointestinal tract (GIT) were collected after d 58 of gestation. 9 Fetal placenta weights were the total weight of placenta divided by the total number of fetuses. 10 Placental efficiency was calculated by the ratio of average fetal weight to average placental weight in a litter.

6 Selenium source for sows 5545 Table 5. Effect of Se source on weight gain, ratio of placenta to uterine weight, and placental efficiency at different days of gestation Open 3 Variable Day of gestation 1 Gilts Sel-Plex Inorganic Weight gain during gestation, kg Sel-Plex Inorganic SE P-value Ratio of placenta to uterine weight Sel-Plex Inorganic SE P-value Placental efficiency 4 Sel-Plex Inorganic SE P-value < Days of gestation indicated above were the average of defined slaughter period; individual gestation day of gilt = mean ± 1 d and regression analysis limited to 43 to 108 d of gestation in maternal tissue. 2 Gilts in the d-0 unbred group were slaughtered at the average of 7 d after first gilt was bred. 3 Gilts, bred but not pregnant, were slaughtered at the average of 120 d after the first gilt was bred. 4 Placental efficiency was calculated by the ratio of average fetal weight to average placental weight in a litter. 5 Linear response, P < Quadratic response, P < diet were greater (P < 0.01) than those in gilts fed inorganic Se (2.60 vs mg/kg); this difference was likewise reflected in the Se concentration in fetal liver (2.13 vs.1.43 mg/kg; P < 0.01). Se source did not affect other trace mineral concentrations in maternal and fetal livers with an exception of Mn in maternal liver, which showed that more Mn was accumulated from the Sel-Plex treatment than from the inorganic Se treatment (6.18 vs mg/kg; P < 0.01). Table 7 shows the effect of Se source on Se and Mn concentrations in maternal and fetal livers at different days of gestation. Overall, liver Se concentrations in gilts fed Sel-Plex were greater than those in gilts fed inorganic Se across gestation age, even though at some specific gestation points there were no statistical significances (P > 0.10) except for d 43 and open gilts. Surprisingly, liver Se concentrations in open gilts fed inorganic Se at 120 d were even lower than those in pregnant gilts from the same treatment (P < 0.10 for all the contrasts); however, this anomaly was not manifest in response to the Sel-Plex treatment. That is, the liver Se in open gilts fed Sel-Plex Table 6. Effect of Se source on trace mineral concentrations (mg/kg, DM basis) in maternal and fetal livers 1 Diet P-value 2 Variable Sel-Plex Inorganic Se SE Diet Day Diet Day No. of gilts Maternal liver Cu Fe Mn < Zn Se < Fetal liver Cu < Fe < Mn < Zn Se <0.01 < Maternal liver = each value represents the mean of 30 gilts across gestation age from the same treatment, including d 0 and open gilts; fetal liver = each value represents the mean of 20 litters from d 58 to 108 of gestation with 1 liver value per litter from 3 pooled fetal livers in that litter. 2 Diet effect = contrast between organic vs. inorganic Se treatment; day = contrast among different days of gestation; diet day = interaction between the source of Se and days of gestation. accumulated to the highest level after 120 d compared with that of pregnant gilts from the same treatment. In general, fetal liver Se concentrations associated with gilts fed Sel-Plex from d 58 to 108 of gestation were much greater (P < 0.05 at all times) than those fed inorganic Se, which indicates that Se in an organic form (e.g., seleno-methionine) may more easily pass the placental barrier than an inorganic form (e.g., selenite) for accumulation in fetal liver. An interaction (P = 0.07) between Se source and gestation age was observed in fetal liver concentration wherein the numerical difference that was observed in liver Se concentration at d 58 of gestation was reduced almost 50% by d 108 of gestation. Selenium concentrations in fetal liver from the 2 treatments decreased quadratically (P < 0.01; R 2 = 0.80 and 0.74 from Sel-Plex and inorganic treatments, respectively) with gestation age. Trace Mineral Concentrations in Fetus Effects of Se source on trace mineral concentrations in the fetuses are shown in Table 8. Generally, Se concentration in the fetuses from gilts fed Sel-Plex were greater (P < 0.01) than those fed the inorganic diet (0.95 vs mg/kg). Interestingly, fetal Fe concentration tended to be elevated (P = 0.08) in response to the Sel- Plex diet (239.3 vs mg/kg). Table 9 shows the effect of Se source on Se and Fe concentrations in the fetus at different days of gestation. Fetal Se concentrations generally were elevated in

7 5546 Ma et al. Table 7. Effect of Se source on Se and Mn concentrations (mg/kg, DM basis) in maternal and fetal liver at different days of gestation Open 3 Variable Day of gestation 1 Gilts Sel-Plex Inorganic Se concentration in maternal liver, mg/kg Sel-Plex Inorganic SE P-value Mn concentration in maternal liver, mg/kg Sel-Plex Inorganic SE P-value Se concentration in fetal liver, 4 mg/kg Sel-Plex Inorganic SE P-value < <0.01 < Days of gestation indicated above were the average of defined slaughter period, individual gestation day of gilt = mean ± 1day and regression analysis limited to 43 to 108 d of gestation in maternal liver and 58 to 108 d in fetal liver. 2 Gilts in the d-0 unbred group were slaughtered at the average of 7 d after first gilt was bred. 3 Gilts, bred but not pregnant, were slaughtered at the average of 120 d after the first gilt was bred. 4 Each individual value for fetal liver was a composite of 3 fetal livers within a litter. 5 Quadratic response, P < response to the Sel-Plex (P < 0.01), except for a tendency at d 43 (P = 0.09) and no statistical difference at d 73 (P = 0.22). As for Fe, although the total gestational response presented in Table 8 demonstrated a tendency (P = 0.08) for elevated concentrations in response to Sel-Plex, examination of specific gestation time points revealed that differences were not significant except for Table 8. Effect of Se source on trace mineral concentration (mg/kg, DM basis) in the whole fetal body 1 Diet P-value 2 Variable Sel-Plex Inorganic Se SE Diet Day Diet Day No. of litters Cu < Fe < Mn < Zn < Se <0.01 < Each value represents the mean of 24 litters from d 43 to 108 of gestation with 1 fetal value per litter from 3 pooled fetuses in that litter. 2 Diet effect = contrast between organic vs. inorganic Se; day = contrast among different days of gestation; diet day = interaction between the source of Se and days of gestation. Table 9. Effect of Se source on Se and Fe concentrations (mg/kg, DM basis) in the whole fetal body at different days of gestation 1 Day of gestation Variable Litters Sel-Plex Inorganic Se concentrations Sel-Plex Inorganic SE P-value 0.09 < <0.01 <0.01 <0.01 Fe concentrations Sel-Plex Inorganic SE P-value Days of gestation indicated above were the average of defined slaughter period, individual gestation day of gilt = mean ± 1 d and regression analysis limited to 43 to 108 d of gestation and each value of fetus was pooled from 3 fetuses in 1 litter. 2 Quadratic response, P < Linear response, P < Cubic response, P < d 91. Regression analysis showed fetal Se concentrations from Sel-Plex treatment decreased quadratically (P < 0.01; R 2 = 0.90) and in those fed the inorganic Se diet they decreased linearly (P < 0.01; R 2 = 0.80) as gestation progressed. Iron concentration decreased cubically in response to the Sel-Plex treatment (P < 0.01; R 2 = Table 10. Effect of Se source on trace mineral contents (mg) in the individual whole fetal body and whole litter 1 Diet P-value 2 Variable Sel-Plex Inorganic Se SE Diet Day Diet Day No. of litters Fetus, mg/fetus Cu < Fe < Mn < Zn < Se <0.01 <0.01 <0.01 Whole litter, mg 3 Cu < Fe <0.01 < Mn < Zn < Se <0.01 <0.01 < Each value represents the mean of 24 litters from d 43 to 108 of gestation with 1 fetal value per litter from 3 pooled fetuses in that litter. 2 Diet effect = contrast between organic vs. inorganic Se; day = contrast among different days of gestation; diet day = interaction between the source of Se and days of gestation. 3 Trace mineral content in whole litter = the average of trace mineral content in fetus litter size of gilts.

8 Selenium source for sows 5547 Table 11. Effect of Se source on Se and Fe content (mg) in individual fetus and whole litter at different days of gestation 1 Day of gestation Variable Litters Sel-Plex Inorganic Se content in fetus, mg/fetus Sel-Plex Inorganic SE P-value 0.14 < <0.01 Fe content in fetus, mg/fetus Sel-Plex Inorganic SE P-value Se content in whole litter, mg Sel-Plex Inorganic SE P-value < Fe content in whole litter, mg Sel-Plex Inorganic SE P-value Days of gestation indicated above were the average of defined slaughter period, individual gestation day of gilt = mean ± 1 d and regression analysis limited to 43 to 108 d of gestation and each value of fetus was pooled from 3 fetuses in 1 litter. 2 Quadratic response, P < ) and decreased linearly in response to the inorganic treatment (P < 0.01; R 2 = 0.67). Total Trace Mineral Content in Individual Fetus and Whole Litter Effect of Se source on trace mineral concentrations in individual fetus and whole litter is shown in Table 10. As expected, Se concentration in the fetus was elevated greatly in response to the Sel-Plex (individual fetus: vs mg in fetus from Sel-Plex vs. inorganic, P < 0.01; whole litter: 0.98 vs mg, P < 0.01). Surprisingly, Sel-Plex tended to increase Fe concentration in the individual fetus (fetus: vs mg, P = 0.07; whole litter: vs mg, P < 0.01). Also, an interaction (P < 0.01) between Se source and gestation age was observed for Se content for both individual fetus and whole litter. Table 11 provides further information showing that Se content in the fetus from Sel-Plex was greater than from inorganic Se; the ratio of Se content in fetus from Sel-Plex vs. inorganic ranged from 1.2:1 at d 43 of gestation to 1.50:1 at d 108 of gestation; however, these differences varied in their level of significance. Selenium content increased quadratically (P < 0.01) from d 43 to 108 of gestation in response to both treatments. The interaction between Se source and days of gestation might be due to the fact that the Sel-Plex has a faster rate of accumulation than the inorganic Se (i.e., selenite). When examining the whole-gestation observations of Table 10 by individual gestation day in Table 11, whereas the Fe contents observed for the Sel-Plex treatment were generally greater numerically than from the inorganic treatment in both the individual fetus and whole litter at different days of gestation, only at d 91 for the entire litter was significance observed. The ranges of the relative differences between treatments were smaller than those observed for Se and the Fe was accumulated quadratically (P < 0.01) from both treatments. The SAS NLN 2-slope, straight broken-line analysis showed the break points for accumulation of Se in individual fetuses from the different treatment. The break point of Se accumulation in individual fetus from Sel- Plex treatment occurred at d 86.8 (P < 0.01; R 2 = 0.95; see Fig. 1) and from inorganic Se treatment occurred at d 69.9 (P < 0.01; R 2 = 0.92; see Fig. 2); the ratios of the slopes (which are the accretion rates) for Sel-Plex vs. inorganic treatments were from 1.94:1 for the first segment (early gestation) to 2.34:1 for the second segment (late gestation). The slope ratio was 1:5.6 (first slope vs. second slope) for the Sel-Plex treatment and was 1:4.7 (first slope vs. second slope) for the inorganic treatment, indicating that Se from Se-enriched yeast is retained more than Se from selenite, especially during late gestation. Lactation Performance Effect of Se source on full-term sow reproductive performance and tissue Se concentration is shown in Tables 12 and 13. In general, the source of Se did not affect the litter size, litter weight, birth weight, liver weight, or growth performance during the 21-d period. Examining the overall effect of Se source in pig body (P < 0.01) and liver (P = 0.03) shows a clear response of tissue Se to Se source (Table 13). The Se concentrations in pig body and liver from sows fed Sel-Plex during gestation were numerically greater at birth, d 7, 14, and 21 compared with those of pigs from sows fed the inorganic Se treatment. However, due to the reduced statistical power because of reduced observation numbers at the individual times, the ability to declare individual days different varied. Selenium concentrations in pig body and liver from both treatments were decreased with the growth of pigs from birth to d 21 of age (pig body: 0.53 to 0.28 mg/kg from Sel-Plex treatment and 0.30 to 0.20 mg/kg from inorganic treatment; pig liver: 0.32 to 0.23 from Sel-Plex and 0.31 to 0.15 from inorganic treatment).

9 5548 Ma et al. Figure 1. Fetal Se content broken-line analysis from d 43 to 108 of gestation (n = 24 litters; organic Se treatment). Breakpoint of Se content in fetus (mg) from Sel-Plex treatment occurred at d 86.8 of gestation (R 2 = 0.95, P < 0.01), showing that fetal Se accumulation accretion mainly occurred after d 86.8 of gestation; the regression equation before d 86.8 was: (d 86.8) After d 86.8, the equation was: (d 86.8) , where d is day of gestation. The 95% confidence interval was 79.9 to 93.6 d. DISCUSSION In this experiment, gilts supplemented with organic Se (Sel-Plex; Se-enriched yeast) at 0.3 ppm had reproductive performance (total and live litter size, fetal survival rate, and fetal weight and length) similar to that of gilts supplemented at the same level of inorganic Se (Na selenite). In agreement, Mahan and Kim (1996) found no effect on reproductive performance with 0.1 and 0.3 ppm Se supplementation as Sel-Plex or selenite in first parity gilts, although reproductive responses to supplemental 0.1 ppm Se as selenite have been noted during later reproductive cycles and lower litter sizes have been attributed to Se inadequacy (Mahan et al., 1974; Chavez and Patton, 1986). Figure 2. Fetal Se content broken-line analysis from d 43 to 108 of gestation (n = 24 litters; inorganic Se treatment). Breakpoint of Se content in fetus (mg) from selenite treatment occurred at d 69.6 of gestation (R 2 = 0.92, P < 0.01), showing that fetal Se accumulation accretion mainly occurred after d 69.6 of gestation; the regression equation before d 69.6 was: (d 69.6) After d 69.6 the equation was: (d 69.6) , where d is day of gestation. The 95% confidence interval was 52.7 to 86.6 d.

10 Selenium source for sows 5549 Table 12. Effect of Se source on sow reproductive performance and piglet growth 1 Variable Sel-Plex Inorganic Se SE P-value No. of litters 5 5 No. of piglets Litter data Birth Total no No. of stillbirths No. born live Litter wt, kg Pig wt, kg Liver wt, g d ADG, g (0 to 7 d) Pig wt, kg Liver wt, g d ADG, g (7 to 14 d) Pig wt, kg Liver wt, g d ADG, g (14 to 21 d) Pig wt, kg Liver wt, g Growth performance data from all piglets but specific pig body and liver weights only from those piglets sacrificed at different periods for the data presented in Table 13. Two piglets from 3 to 5 litters were randomly selected at each time point and the analytical values pooled for a single observation per litter. In the present experiment, growth performance of postnatal pigs was not affected by the different sources of Se supplementation, which is in agreement with the results of Mahan and Kim (1996). However, Zhan et al. (2010) fed either selenomethionine or sodium selenite at 0.3 ppm Se to sows (32 and 28 d for gestation and lactation periods, respectively) and demonstrated that the daily weight gain of piglets after birth from sows that were fed selenomethionine increased about 12% compared with those fed sodium selenite. These differences might be due to different parities of sow used in different experiments; that is, the first parity gilts were used both in the present experiment and in research by Mahan and Kim (1996), whereas third-parity sows were used in research by Zhan et al. (2010). In addition, the serial slaughter of piglets for tissue sampling over the course of the present experiment (resulting in a smaller litter size nursing the female in the latter parts of lactation) might also have contributed to the difference between studies. Placental Se transfer, as reflected by the fetal body and liver tissue Se contents, were greater when the Seenriched yeast was fed to gestating gilts, which demonstrates that neonatal tissue Se content can be influenced by the source of Se provided to the gestating female. In the present experiment, the Se concentration from fetal Table 13. Effect of Se source on Se concentration (mg/ kg DM) in pig body and associated liver from birth to d 21 of age Day of age 1 Variable Overall Pigs Sel-Plex Inorganic Se in pig body Sel-Plex Inorganic SE P-value < <0.01 Se in pig liver Sel-Plex Inorganic SE P-value Days of age indicated above were the average with individual ages = mean ± 1 d. body and liver were greater from gilts fed Sel-Plex than from gilts fed sodium selenite, which may indicate that greater placental Se transfer from dam to offspring occurs by feeding Sel-Plex compared with sodium selenite. The ratio of Se content in the fetus from Sel-Plex vs. inorganic Se ranged from 1.2:1 at d 43 of gestation to 1.50:1 at d 108 of gestation. Greater amounts of Se seem to be supplied from Sel-Plex, especially when the Se demand is greater (e.g., in late gestation). The interaction between Se source and days of gestation is due to the accumulated daily difference as gestation proceeds. In agreement, Mahan and Peters (2004) suggested sows fed the Sel-Plex source had a greater transfer of Se into the neonate, colostrum, milk, weaned pigs, and various sow tissues than did sows fed inorganic Se. In a similar study, Yoon and McMillan (2006) found that piglet serum at birth from sows fed Se yeast had greater Se concentration than selenite. Even though those experiments did not measure the Se concentration in the fetus during gestation directly, the 0-d-old piglet might be the indicator of Se status before birth. In another study, Peters et al. (2010) fed organic trace minerals (Bioplex Cu, Fe, Mn, Zn; Se as Sel-Plex) to sows over 6 parities and found that only the Sel-Plex increased sow body and liver Se concentration, a finding consistent with the present study. Unexpectedly, in the present study, Sel-Plex tended to increase Mn content of the maternal liver and Fe accumulation in the fetus, which was not observed by Peters et al. (2010). In the present experiment, piglets from sows fed Sel- Plex had a greater Se content at birth in both body and liver. These Se advantages (about 40%) persisted through weaning. In agreement, Mahan and Kim (1996) showed that pig liver Se content at weaning was approximately 31% higher

11 5550 Ma et al. when piglets nursed sows that were fed 0.3 ppm Se from the Se-enriched yeast than when dams were fed diets with added selenite at 0.3 ppm Se. Liver is considered the major labile body storage site for Se and is perhaps the organ that best reflects the Se status of the animal. Milk is the major source of nutrients from birth to weaning, which indicates that either milk Se content was elevated when organic Se was fed in the present study or that the pigs obtained Se from another source (e.g., sow feed or feces) to have an increased Se content in the weaned pigs. This finding was consistent with those of previous investigators (Mahan and Kim, 1996; Mahan and Peters, 2004). LITERATURE CITED Chavez, E. R., and K. L. Patton Response to injectable Se and vitamin E on reproductive performance of sows receiving a standard commercial diet. Can. J. Anim. Sci. 66: EPA Method 3052: Microwave assisted acid digestion of siliceous and organically based matrices. United States Environmental Protection Agency, Washington, DC. EPA Method 6020a: Inductively coupled plasma-mass spectrometry United States Environmental Protection Agency, Washington, DC. Kelly, M. P., and R. F. Power Fractionation and identification of the major selenium compounds in selenized yeast. J. Dairy Sci. 78(Suppl. 1):237 (Abstr.). Mahan, D. C Effect of organic or inorganic selenium sources and levels on sow colostrum and milk selenium content. J. Anim. Sci. 78: Mahan, D. C., and Y. Y. Kim Effect of inorganic or organic selenium at two dietary levels on reproductive performance and tissue selenium concentrations in first-parity gilts and their progeny. J. Anim. Sci. 74: Mahan, D. C., L. H. Penhale, J. H. Cline, A. L. Moxon, A. W. Fetter, and J. T. Yarrington Efficacy of supplemental selenium in reproductive diets on sow and progeny performance. J. Anim. Sci. 39: Mahan, D. C., and J. C. Peters Long-term effects of dietary organic or inorganic selenium sources and levels on reproducing sows and their progeny. J. Anim. Sci. 82: NRC Nutrient requirements of swine. 10th ed. Natl. Acad. Press, Washington, DC. Peters, J. C., D. C. Mahan, T. G. Wiseman, and N. D. Fastinger Effect of dietary organic and inorganic micromineral source and level on sow body, liver, colostrum, mature milk, and progeny mineral compositions over six parities. J. Anim. Sci. 88: Robbins, K. R., A. M. Saxton, and L. L. Southern Estimation of nutrient requirements using broken-line regression analysis. J. Anim. Sci. 84:E155 E165. Stadlober, M., M. Sager, and K.J. Irgolic Effects of selenate supplemented fertilisation on the selenium level of cerealsidentification and quantification of selenium compounds by HPLC-ICP-MS. Food Chem. 73: Stein, H. H., S. W. Kim, T. T. Nielsen, and R. A. Easter Standardized ileal protein and amino acid digestibility by growing pigs and sows. J. Anim. Sci. 79: Yoon, I., and E. McMillan Comparative effects of organic and inorganic selenium on selenium transfer from sows to nursing pigs. J. Anim. Sci. 84: Zhan, X., Y. Z. Qie, M. Wang, X, Li, and R. Q. Zhao Selenomethionine: An effective source for sow to improve Se distribution, antioxidant status, and growth performance of pig offspring. Biol. Trace Elem. Res. 142: