Skip to main content
. Author manuscript; available in PMC: 2022 Jun 1.
Published in final edited form as: Cell Rep. 2022 May 3;39(5):110772. doi: 10.1016/j.celrep.2022.110772

Figure 4. Plasma lipidomics reveals biomarkers of neonatal BCG vaccination.

Figure 4.

(A) Complex lipid panel pipeline schematic diagram. Lipid extracts were subjected to infusion-MS analysis by combining differential mobility spectrometry (DMS) with an extensive library of lipid masses to quantify 14 lipid classes. Lipids were profiled and identified based on known lipid reference standards.

(B) Lipidomics panel identified 963 lipids from the in vivo newborn cohort belonging to 14 different lipid subclasses, with triglycerides being the most abundant. Data represent counts of identified lipids.

(C) Supervised sPLS-DA from Guinea-Bissau newborns distinctly discriminated lipids between early and delayed BCG groups, with PC1 accounting for 52% of variance and PC2 for 15% variance.

(D) Early administration of BCG at birth significantly perturbs plasma lipid concentrations at 4 weeks of age. Significantly altered lipid metabolites in the early versus delayed BCG newborn group were annotated in red (p < 0.05), those that were nearly significant in pink (p < 0.1–p > 0.05), and those that were not significant in gray.

(E) Stimulation of human newborn cord blood in vitro with BCG perturbs phospholipid pathways. Supernatants from blood stimulated in vitro for 18 h with vehicle control, or BCG were subjected to lipidomics and demonstrated much lower lipids production than control. Significantly altered lipid metabolites in the BCG versus vehicle conditions were annotated in red (p < 0.05), those that were nearly significant in pink (p < 0.1–p > 0.05), and those that were not significant in gray.

(F) A majority of significantly perturbed lipids were decreased in vitro following 18 h BCG stimulation and in vivo following early BCG immunization, reflecting a BCG-induced metabolic signature. Comparison of significant lipids between the in vivo Guinea-Bissau and in vitro Boston cohorts reported as log2 fold-change of early BCG/delayed BCG (Guinea-Bissau) and BCG-stimulated/control (Boston). Except for free fatty acid components and LPC 16:0, the directionality of lipid fold-changes was similar in vivo and in vitro.

(G) Concentrations of multiple LPC components were significantly decreased in BCG-stimulated samples in vivo and in vitro. Data are presented as log2 fold-change of early versus delayed (or no) BCG.