Human Milk Oligosaccharide Composition Analysis

Human milk samples for HMO analysis were obtained by manual milk expression by the mother into a clean plastic container. The samples were kept cool until homogenization by the study team. They were then split into 1–2 mL portions and stored at −20°C. All aliquots were stored at −20°C at the study site until shipment on dry ice to the ETH Zurich, Switzerland. For the HMO composition analysis reported here, the HM samples were transported on dry ice to the glycoanalytical laboratory (glyXera GmbH, Magdeburg, Germany). The qualitative and quantitative HMO composition of each individual HM sample was determined with the glyXboxCE™ system (glyXera GmbH, Magdeburg, Germany) based on multiplexed capillary gel electrophoresis with laser-induced fluorescence detection (xCGE-LIF).^{73 (link)} In accordance with the glyXera GmbH kit protocol (KIT-glyX-OS.P-APTS, glyXera GmbH, Magdeburg, Germany), the pure HM samples were diluted 1:100, spiked with an internal standard (IS) (oligosaccharide (OS) quantification standard solution, OS-A5-N-1 mL-01; part of the KIT-glyX-Quant-DP5, all from glyXera GmbH, Magdeburg, Germany) and treated with a denaturation solution. The free OS were labeled with 8-aminopyrene-1,3,6-trisulfonic acid (APTS), purified and determined with the glyXbox™ system. All measurements included the addition of a migration time alignment standard (glyXalign4; STD-glyXalign-4-S, glyXera GmbH) to the sample. Finally, glyXtoolGUI™ software (Beta v0.8.11, glyXera GmbH, Magdeburg, Germany) was used for the processing and analysis of the HMO Fingerprints data (normalized electropherograms). The limit of quantification (LOQ) was determined from the signal-to-noise ratio (SNR) of each HMO Fingerprint calculated as described by Ullsten et al.^{74 (link)} The LOQ was defined as an SNR of 10 and the limit of detection (LOD) was defined as an SNR of 3. The respective noise for each sample was determined after migration time alignment of the unsmoothed data in the late migration time range (approximation range = degree of polymerization (DP) 18< DP<20). Peaks with intensities below the LOQ but above the LOD were picked. All peaks ≥LOQ were considered and their IS-normalized peak areas were calculated (as percentages relative to the peak area of the IS [% IS] (= nPA)). All peaks ≥LOD but 75 (link) All HM samples were assigned to a maternal secretor and Lewis (Se/Le) phenotype (HM groups I–IV) based on the presence or absence of specific α1-2- and/or α1-4-fucosylated HMOs, as previously described.^{73 (link)} The assignment of maternal secretor status was based on the presence of 2’-fucosyllactose (2’-FL), difucosyllactose (DFL), and lacto-N-fucopentaose (LNFP) I, and the determination of Lewis status was based on the presence of LNFP II and lacto-N-difucohexaose (LNDFH) II. Differences in HMO abundance between maternal secretor status and HM types were assessed with Mann-Whitney tests or Kruskal-Wallis tests followed by post-hoc Dunn’s test, respectively, with adjustment for false discovery rate (FDR) by the Benjamini-Hochberg mechanism (FDR<0.05).