Human milk oligosaccharides reduce necrotizing enterocolitis-induced neuroinflammation and cognitive impairment in mice

Abstract

Necrotizing enterocolitis (NEC) is the leading cause of morbidity and mortality in premature infants. One of the most devastating complications of NEC is the development of NEC-induced brain injury, which manifests as impaired cognition that persists beyond infancy and which represents a proinflammatory activation of the gut-brain axis. Given that oral administration of the human milk oligosaccharides (HMOs) 2'-fucosyllactose (2'-FL) and 6'-sialyslactose (6'-SL) significantly reduced intestinal inflammation in mice, we hypothesized that oral administration of these HMOs would reduce NEC-induced brain injury and sought to determine the mechanisms involved. We now show that the administration of either 2'-FL or 6'-SL significantly attenuated NEC-induced brain injury, reversed myelin loss in the corpus callosum and midbrain of newborn mice, and prevented the impaired cognition observed in mice with NEC-induced brain injury. In seeking to define the mechanisms involved, 2'-FL or 6'-SL administration resulted in a restoration of the blood-brain barrier in newborn mice and also had a direct anti-inflammatory effect on the brain as revealed through the study of brain organoids. Metabolites of 2'-FL were detected in the infant mouse brain by nuclear magnetic resonance (NMR), whereas intact 2'-FL was not. Strikingly, the beneficial effects of 2'-FL or 6'-SL against NEC-induced brain injury required the release of the neurotrophic factor brain-derived neurotrophic factor (BDNF), as mice lacking BDNF were not protected by these HMOs from the development of NEC-induced brain injury. Taken in aggregate, these findings reveal that the HMOs 2'-FL and 6'-SL interrupt the gut-brain inflammatory axis and reduce the risk of NEC-induced brain injury.NEW & NOTEWORTHY This study reveals that the administration of human milk oligosaccharides, which are present in human breast milk, can interfere with the proinflammatory gut-brain axis and prevent neuroinflammation in the setting of necrotizing enterocolitis, a major intestinal disorder seen in premature infants.

Keywords: gut-brain axis; human milk oligosaccharides; necrotizing enterocolitis; neonate; neuroinflammation.

Graphical abstract

Figure 1.

Human milk oligosaccharides (HMOs) reduce necrotizing enterocolitis (NEC)-induced brain injury. A–D: quantitative real-time polymerase chain reaction (qRT-PCR) showing the relative mRNA expression of proinflammatory cytokine Tnf in ileum (A) and brain (B) in mice and ileum (C) and brain (D) in piglets. E–X: representative confocal images of the sagittal sections of mice brain immunostained with 3′-nitrotyrosine (3′-NT) and dihydroethidium (DHE), markers of reactive oxygen species (ROS)-induced oxidative injury, in control mice (E–I) and NEC mice without (J–N) or with 2′-fucosyllactose (2′-FL; O–S) or 6′-sialyslactose (6′-SL; T–X) HMO fortification. Merged colors show 3′-NT (white) and DHE (red) staining (E, J, O, T). White dashed insets are regions of interest (ROIs) of hippocampus showing staining for 3′-NT (F, K, P, U) and DHE (H, M, R, W) and midbrain showing staining for 3′-NT (G, L, Q, V) and DHE (I, N, S, X). Scale bars, 200 μm. Y and Z: each dot in graphs represents data from an individual mouse or piglet. The mRNA levels are represented as relative to housekeeping gene Rplp0 expression. Mouse and piglet data collected from either sex; mice n = 8/group, piglets n = 4 or 5/group. Statistical significance was determined by 1-way ANOVA, followed by Tukey’s multiple comparisons tests using GraphPad Prism 9 software. *P < 0·05, **P < 0.01, *** or ****P < 0.001.

Figure 2.

Human milk oligosaccharides (HMOs) reduce proinflammatory cytokine levels in mice with active necrotizing enterocolitis (NEC). A: schematic of the HMO treatment strategy. p, Postnatal day; sac, euthanasia. B and C: quantitative real-time polymerase chain reaction (qRT-PCR) expression of proinflammatory cytokines Il6 and Tnf in the ileum (B) and brain (C) of control or NEC mice without or with 2′-fucosyllactose (2′-FL) or 6′-sialyslactose (6′-SL) HMO treatment. The mRNA levels are represented as relative to housekeeping gene Rplp0 expression. Each dot in dot graphs represents data from an individual mouse ileum or brain. Statistical significance was determined by 1-way ANOVA, followed by Tukey’s multiple comparisons tests using GraphPad Prism 9 software. *P < 0.05, **P < 0.01, *** or ****P < 0.001. Mouse data collected from either sex; mice n ≥ 7/group.

Figure 3.

Fortification with human milk oligosaccharides (HMOs) reduces necrotizing enterocolitis (NEC)-induced microglial activation, and restores myelination, in the newborn mouse brain. A–T: representative confocal images of the sagittal sections of mouse brain, immunostained for myelination marker myelin basic protein (Mbp) and microglial activation marker allograft inflammatory factor 1 (Aif1) in control (A–E) and NEC mice without (F–J) or with 2′-fucosyllactose 2′-FL; K–O) or 6′-sialyslactose (6′-SL; P–T) HMO fortification. Merged colors show staining for Mbp (white) and Aif1 (red) (A, F, K, P). White dashed insets show region of interest (ROI) of corpus callosum showing staining for Mbp (B, G, L, Q) and Aif1 (D, I, N, S) and midbrains showing staining for Mbp (C, H, M, R) and Aif1 (E, J, O, T). Further zoom-in insets show more detailed morphology of microglia. Scale bars, 200 μm or 20 μm (insets, D and E, I and J, N and O, S and T). U and V: dot graphs showing the quantification of fluorescent intensity of ROI with ImageJ software. W: qRT-PCR showing mRNA expression of Aif1. Each dot in dot graphs represents data from an individual mouse brain. The mRNA levels are represented as relative to housekeeping gene Rplp0 expression. Statistical significance was determined by one-way ANOVA, followed by Tukey’s multiple comparisons tests using GraphPad Prism 9 software. *P < 0.05, **P < 0.01, *** or ****P < 0.001. Mouse data collected from either sex; mice n = 6/group.

Figure 4.

2′-Fucosyllactose (2′-FL) and 6′-sialyslactose (6′-SL) prevent cognitive impairment in mice with necrotizing enterocolitis (NEC). A: schematic of behavioral testing strategy. P, postnatal day. B: schematic of the Morris water maze training trial task. C: dot plot showing the latency to platform for mice in training trial 11. D: schematic of the Morris water maze reversal trial task in which the hidden platform is moved to an opposite quadrant from the original trial. E: dot plot showing the latency to platform for mice in reversal trial 11. *P < 0.05, **P< 0.01, *** or ****P < 0.001. Mouse data collected from either sex; mice n ≥ 4/group.

Figure 5.

Administration of human milk oligosaccharides (HMOs) limits the degree of inflammation in neonatal brain in a necrotizing enterocolitis (NEC)-in-a-dish model. A: schematic of the strategy for generation of 3-dimensional (3-D) mouse brain organoids and NEC-in-dish model. E, embryonic day; IHC, immunohistochemistry. B: representative confocal images of the brain organoids immunostained with myelination marker myelin basic protein (Mbp, red) and mature neuronal marker neurofilament peptide, heavy (NF-H, yellow). C: quantitative real-time polymerase chain reaction (qRT-PCR) expression of most common brain-specific genes in cultured brain organoids after 3 wk in culture. D–G: qRT-PCR expression of Tlr4 (D), Il6 (E), Il1b (F), and Tnf (G) in the brain organoids without or with 2′-fucosyllactose (2′-FL) and 6′-sialyslactose (6′-SL) HMO treatments. H–K, M–P, and R–U: representative confocal images showing oxygen sensor dihydroethidium (DHE, red; H–K) staining and oxidative injury marker 3′-nitrotyrosine (3′-NT, green; M–P) immunofluorescence and apoptosis marker terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL, green; R–U) staining in brain organoids untreated (Ctrl; H, M, R), NEC-in-dish treated (I, N, S), NEC-in-dish + 2′-FL treated (J, O, T), and NEC-in-dish + 6′-SL treated (K, P, U). L, Q, and V: dot graphs showing the quantification with ImageJ software of fluorescent intensity of regions of interest (ROIs). The mRNA levels are represented as relative to housekeeping gene Rplp0 expression. Each dot in dot graphs represents data from an individual brain organoid in culture. Statistical significance was determined by 1-way ANOVA, followed by Tukey’s multiple comparisons tests using GraphPad Prism 9 software. *P < 0.05, **P < 0.01, ***or ****P < 0.001. Scale bars, 25 μm.

Figure 6.

NMR analysis to determine ¹³C-labeled human milk oligosaccharide (HMO) relative abundance in control and necrotizing enterocolitis (NEC) brains 3 h after oral administration. A: schematic of experiment to determine the presence of 2′-fucosyllactose (2′-FL) or its metabolites in the brains of control and NEC mice subjected to an experimental model of NEC for 4 days and fed 2 mg/mouse ¹³C-labeled 2′-FL 3 h before euthanasia. p, postnatal day. B: Lewis structure of ¹³C-labeled 2′-FL HMO showing the position of the ¹³C-labeled carbon in fucose. C: 1-dimensional (1-D) slice from a 2-dimensional (2-D) heteronuclear single-quantum coherence spectroscopy (HSQC) showing the peaks associated with 2′-FL HMO. D: 2-D NMR spectrum showing the HSQC peaks associated with 2′-FL HMO. E: 2-D NMR plot showing the HSQC peaks associated with control mice (black) and 2′-FL HMO (green). F: 2-D NMR plot showing the HSQC peaks associated with NEC mice (red) and 2′-FL HMO (green). G: quantification of peaks from E and F. Statistical significance was determined by unpaired t test using GraphPad Prism 9 software. ****P < 0.001.

Figure 7.

Human milk oligosaccharides (HMOs) maintain blood-brain barrier integrity and function in neonatal mice. A–T: representative confocal images of the sagittal sections of mouse brain immunostained for tight junction protein Zona occludin 1 (Zo-1) in control (A–E) and necrotizing enterocolitis (NEC) mice without (F–J) or with 2′-fucosyllactose (2′-FL; K–O) or 6′-sialyslactose (6′-SL; P–T) HMO fortification. White dashed insets show region of interest (ROI) of lateral ventricle (B, G, L, Q), 3rd ventricle (C, H, M, R), 4th ventricle (D, I, N, S), and blood vessels (E, J, O, T). Scale bars, 200 μm. White arrows point toward choroid plexus (chp), ependymal cells (ep), and blood vessels (bv). U: dot graphs showing the quantification using ImageJ software of fluorescent intensity of ROI. V: blood-brain barrier permeability measured by the amount (relative fluorescent units, RFU) of fluorescein isothiocyanate-dextran (FITC-dextran) in the whole brain lysates of control and NEC mice without or with 2′-FL or 6′-SL HMO fortification. Statistical significance was determined by 1-way ANOVA, followed by Tukey’s multiple comparisons tests using GraphPad Prism 9 software. *P < 0.05, **P < 0.01, *** or ****P < 0.001. Mouse data collected from either sex; mice n ≥ 6/group.

Figure 8.

Human milk oligosaccharides (HMOs) reduce necrotizing enterocolitis (NEC)-induced brain injury in mice via Bdnf signaling. A: quantitative real-time polymerase chain reaction (qRT-PCR) showing relative mRNA expression of Bdnf in C57BL/6 wild-type control (Ctrl), NEC, and NEC mice fortified with 2′-fucosyllactose (2′-FL) and 6′-sialyslactose (6′-SL) HMOs. B: relative mRNA expression of proinflammatory cytokine Il6 in C57BL/6 wild-type and Bdnf-mutant NEC mice fortified with 2′-FL and 6′-SL HMOs. C: relative mRNA expression of Bdnf on brain organoids derived from C57BL/6 mice and exposed to NEC-in-dish model with or without treatment with 2′-FL and 6′-SL HMOs for 6 h in culture. D: relative mRNA of proinflammatory cytokine Il6 on brain organoids derived from wild-type and Bdnf mutant mice and exposed to NEC-in-dish model with or without treatment with 2′-FL and 6′-SL HMOs for 6 h in culture. E–H, J–M, O–R, and T–W: representative confocal images of brain organoids immunostained with oxidative injury marker dihydroethidium (DHE; E–H, J–M) and 3′-nitrotyrosine (3′-NT; O–R, T–W) exposed in culture to NEC-in-a-dish (F, K, P, U), NEC-in-dish + 2′-FL (G, L, Q, V), and NEC-in-a-dish + 6′-SL HMO (H, M, R, W) for 6 h. Scale bars, 25 μm. I, N, S, and X: dot graphs showing the quantification using ImageJ software of fluorescent intensity of regions of interest (ROIs). The mRNA levels are represented as relative to housekeeping gene Rplp0 expression. Each dot in dot graphs represents data from an individual brain organoid in culture. Statistical significance was determined by 1-way ANOVA, followed by Tukey’s multiple comparisons tests using GraphPad Prism 9 software. *P < 0.05, **P < 0.01, *** or ****P < 0.001. Mouse data collected from either sex; mice n ≥ 4/group.