Metabolomic Approaches to Explore Chemical Diversity of Human Breast-Milk, Formula Milk and Bovine Milk

Abstract

Although many studies have been conducted on the components present in human breast milk (HM), research on the differences of chemical metabolites between HM, bovine milk (BM) and formula milk (FM) is limited. This study was to explore the chemical diversity of HM, BM and FM by metabolomic approaches. GC-TOFMS and UPLC-QTOFMS were applied to investigate the metabolic compositions in 30 HM samples, 20 FM samples and 20 BM samples. Metabolite profiling identified that most of the non-esterified fatty acids, which reflected the hydrolysis of triglycerides, were much more abundant in HM than those in FM and BM, except for palmitic acid and stearic acid. The levels of tricarboxylic acid (TCA) intermediates were much higher in FM and BM than those in HM. Each type of milk also showed its unique composition of free amino acids and free carbohydrates. In conclusion, higher levels of non-esterified saturated fatty acids with aliphatic tails <16 carbons, monounsaturated fatty acids and polyunsaturated fatty acids and lower levels of TCA intermediates are characteristic of HM, as compared with FM and BM. The content of non-esterified fatty acids may reflect the hydrolysis of triglycerides in different milk types.

Keywords: GC-TOFMS; UPLC-QTOFMS; amino acids; carbohydrate; fatty acids; metabolomics; tricarboxylic acid intermediates.

Figure 1

The scores plots of PLS-DA model constructed with annotated metabolites. (A) The scores plot of PLS-DA model discriminating HM, LFM, PFM and BM; (B) the scores plot of PLS-DA model discriminating HM and FM; (C) the scores plot of PLS-DA model discriminating HM and BM. Each component axis is denoted as t[x] on the score plots where x is the component number. HM, human breast milk; LFM, liquid formula milk; PFM, powdered formula milk; FM, formula milk; BM, bovine milk.

Figure 2

Heatmaps of hierarchical cluster analysis comparing the metabolite changes among HM, FM and BM. The heatmaps were generated using the differential metabolites (variable importance in the projection (VIP) >1, p < 0.05) for the comparison of FM vs. HM (A); and BM vs. HM (B). The heatmap graphically represents individual changes of ion intensity. The color red represents the relatively higher abundance, and the color blue represents the relatively lower abundance of each metabolite. The fold of change was normalized by log2 transformation. Hierarchical clustering agglomeration was performed on the identified metabolites using the Pearson correlation as the distance metric. HM, human breast milk; FM, formula milk; BM, bovine milk.

Figure 3

Comparison of the relative abundances of saturated fatty acids (A), monounsaturated fatty acids (B) and polyunsaturated fatty acids (C) among HM, FM and BM. The Y-axis represents the normalized peak area. Boxes contain the 25%–75% measurements for each group, and whiskers cover the 5%–95% measurements. The horizontal lines within the box represent the median values. * p < 0.05 vs. FM; † p < 0.05 vs. BM. Kruskal–Wallis H test followed by Bonferroni-corrected post-hoc test. HM, human breast milk; FM, formula milk; BM, bovine milk.

Figure 3

Comparison of the relative abundances of saturated fatty acids (A), monounsaturated fatty acids (B) and polyunsaturated fatty acids (C) among HM, FM and BM. The Y-axis represents the normalized peak area. Boxes contain the 25%–75% measurements for each group, and whiskers cover the 5%–95% measurements. The horizontal lines within the box represent the median values. * p < 0.05 vs. FM; † p < 0.05 vs. BM. Kruskal–Wallis H test followed by Bonferroni-corrected post-hoc test. HM, human breast milk; FM, formula milk; BM, bovine milk.

Figure 4

Bar plot of the amino acid changes for the comparison of FM vs. HM (A); and BM vs. HM (B). A fold change (FC) value for each amino acid (X) was calculated as follows: X_{fold change} = (X_{level in FM or BM}/X_{level in HM}). Each bar representing an FC value was colored to indicate its corresponding p-value and thereby to specify the statistical significance of the difference (see the color scale). The FC value was normalized by log2 transformation, and the p-value was normalized by log10 transformation. HM, human breast milk; FM, formula milk; BM, bovine milk.

Figure 5

Bar plot of tricarboxylic acid (TCA) intermediates that significantly differed among HM, FM and BM. Blue labeling represents no significant difference among milk types. Red labeling represents a significant difference among milk types. The Y-axis represents the normalized peak area. Boxes contain the 25%–75% measurements for each group, and whiskers cover the 5%–95% measurements. The horizontal lines within the box represent the median values. * p < 0.05 vs. FM; † p < 0.05 vs. BM. Kruskal–Wallis H test followed by Bonferroni-corrected post-hoc test. HM, human breast milk; FM, formula milk; BM, bovine milk.

Figure 6

Bar plot of carbohydrates that significantly differed among HM, FM and BM. The Y-axis represents the normalized peak area. Boxes contain the 25%–75% measurements for each group, and whiskers cover the 5%–95% measurements. The horizontal lines within the box represent the median values. * p < 0.05 vs. FM; † p < 0.05 vs. BM. Kruskal–Wallis H test followed by Bonferroni-corrected post-hoc test. HM, human breast milk; FM, formula milk; BM, bovine milk.