Applying quantitative metabolomics based on chemical isotope labeling LC-MS for detecting potential milk adulterant in human milk

Abstract

There is an increasing demand for donor human milk to feed infants for various reasons including that a mother may be unable to provide sufficient amounts of milk for their child or the milk is considered unsafe for the baby. Selling and buying human milk via the Internet has gained popularity. However, there is a risk of human milk sold containing other adulterants such as animal or plant milk. Analytical tools for rapid detection of adulterants in human milk are needed. We report a quantitative metabolomics method for detecting potential milk adulterants (soy, almond, cow, goat and infant formula milk) in human milk. It is based on the use of a high-performance chemical isotope labeling (CIL) LC-MS platform to profile the metabolome of an unknown milk sample, followed by multivariate or univariate comparison of the resultant metabolomic profile with that of human milk to determine the differences. Using dansylation LC-MS to profile the amine/phenol submetabolome, we could detect an average of 4129 ± 297 (n = 9) soy metabolites, 3080 ± 470 (n = 9) almond metabolites, 4256 ± 136 (n = 18) cow metabolites, 4318 ± 198 (n = 9) goat metabolites, 4444 ± 563 (n = 9) infant formula metabolites, and 4020 ± 375 (n = 30) human metabolites. This high level of coverage allowed us to readily differentiate the six different types of samples. From the analysis of binary mixtures of human milk containing 5, 10, 25, 50 and 75% other type of milk, we demonstrated that this method could be used to detect the presence of as low as 5% adulterant in human milk. We envisage that this method could be applied to detect contaminant or adulterant in other types of food or drinks.