Fasting plasma metabolomics reveal specific dietary patterns in sow‐fed neonatal piglets compared to soy‐ or dairy‐based formula feeding.

The beneficial effects of breastfeeding on infant health is well‐known; however, the mechanisms underlying these effects have not been fully elucidated. To characterize the impact of formula‐vs. sow‐feeding on metabolic pathways, we used untargeted metabolomics on fasting sera samples from neonatal piglets to compare effects of sow‐fed and infant formula‐based diets. A total of 36 piglets were weaned exclusively on ad‐libitum sow‐feeding for 2 days after birth and then equally split into 3 exclusive dietary groups for 21 consecutive days: continuation of ad‐libitum sow‐feeding, dairy‐, or soy‐based infant formula fed diets. Total energy intake in dairy‐and soy‐based infant formula diets were given to match weight gain in ad‐libitum sow‐fed piglets (i.e., piglets were matched by digestible energy intake). Fasting serum samples were collected at the end of the feeding period and analyzed using LC‐MS. Partial least squares‐discriminant analysis (PLS‐DA) models were used to identify ions that discriminate dietary group differences. Data were split into model training and test datasets to determine PLS‐DA model accuracy and only training data was used for model development and feature selection. Optimal cross‐validation accuracy was reached with 21 metabolites included in the model, and permutations of final model revealed a significant difference between nested cross‐validated model accuracy relative to nested permuted accuracy ( P < 0.01). Classification of data held out in test set was correctly predicted using 2 latent variables. Amino acids and amino acid derivatives, including phenylalanine, arginine, kynurenine, dimethylglycine, urea, and pipecolic acid, were selected in the final model and found to be significantly different among dietary groups (all P < 0.05; 1‐way ANOVA; Benjamini and Hochberg adjusted). Phosphoric acid, citric acid, and sphingosin‐1‐phosphate were also featured in final PLS‐DA model. Other phospholipid species were identified in the final model and validation of these species are ongoing. Our data suggest that amino acid and lipid metabolism is significantly altered by infant formula‐based diets matched for digestible energy intake in a neonatal piglet model, likely due to the disparate protein and lipid composition of the three diets. Future work will be to assess and integrate the metabolomic profile in liver, ileum, and urine to provide a comprehensive overview of metabolic changes associated with infant formula feeding. Support or Funding Information USDA‐ARS Project 6026‐5100‐010‐05S