Variations in the Composition of Human Milk Oligosaccharides Correlates with Effects on Both the Intestinal Epithelial Barrier and Host Inflammation: A Pilot Study

Abstract

Background: Human milk oligosaccharides are complex, non-digestible carbohydrates that directly interact with intestinal epithelial cells to alter barrier function and host inflammation. Oligosaccharide composition varies widely between individual mothers, but it is unclear if this inter-individual variation has any impact on intestinal epithelial barrier function and gut inflammation.

Methods: Human milk oligosaccharides were extracted from the mature human milk of four individual donors. Using an in vitro model of intestinal injury, the effects of the oligosaccharides on the intestinal epithelial barrier and select innate and adaptive immune functions were assessed.

Results: Individual oligosaccharide compositions shared comparable effects on increasing transepithelial electrical resistance and reducing the macromolecular permeability of polarized (Caco-2Bbe1) monolayers but exerted distinct effects on the localization of the intercellular tight junction protein zona occludins-1 in response to injury induced by a human enteric bacterial pathogen Escherichia coli, serotype O157:H7. Immunoblots showed the differential effects of oligosaccharide compositions in reducing host chemokine interleukin 8 expression and inhibiting of p38 MAP kinase activation.

Conclusions: These results provide evidence of both shared and distinct effects on the host intestinal epithelial function that are attributable to inter-individual differences in the composition of human milk oligosaccharides.

Keywords: human milk oligosaccharides; inflammation; intestinal epithelial barrier.

Figure 1

Composition of HMOs prepared from four individuals. (A) Percentage composition of HMOs extracted from four mothers (#1, 2, 3 and 4) determined by using HPAEC–PAD. (B–C) HMO composition of mothers #2 and #3 versus mothers #1 and #4 differ in the abundance of both LNT and 2′-FL. (D) HMO composition across four maternal donors clustered by STAMP using the unweighted pair group method with arithmetic mean. 2FL, 2′-fucosyllactose; 3FL, 3′-fucosyllactose; LDFT, lacto-difucotetraose; LNT, lacto-N-tetraose; LNnT, lacto-N-neotetraose; LNFPI, lacto-N-fucopentaose I; LNFPII, lacto-N-fucopentaose II; LNFPIII, lacto-N-fucopentaose III; LNDHI, lacto-N-difucohexaose I; LNDHII, lacto-N-difucohexaose II; 3SL, 3′-sialyllactose; 6SL, 6′-sialyllactose; LSTa, sialyl-lacto-N-tetraose a; LSTb, sialyl-lacto-N-tetraose b; LSTc, sialyl-lacto-N-neotetraose c; DSLNT, disialyl-lacto-N-tetraose.

Figure 2

Functional effects of HMOs from four individuals on Caco-2Bbe1 epithelial barrier integrity. (A) En face images of immunofluorescence microscopy showing the intercellular tight junction protein ZO-1 in Caco-2Bbe1 cells either in the absence of HMOs (control) or pre-treated (20 mg/mL; 24 h) with four individual HMO preparations alone (-EHEC) or followed by EHEC O157:H7 challenge (4 h, MOI 100:1). Green denotes ZO-1 and blue DAPI nuclear staining. Images are representative of 3 separate experiments. (B) Transepithelial electrical resistance (TER), expressed as percentage of PBS treatment used as control, was measured in polarized Caco-2Bbe1 epithelial cells grown on semi-permeable polyester Transwell filters incubated with HMOs prepared from each of four individuals, followed by 4 h EHEC O157:H7 challenge (MOI: 100 to 1; n = 4–5 per group). (C) Caco-2Bbe1 cells in Transwells were then washed with PBS and exposed to 10 kDa dextran in the apical compartment (5 h) and flux of the fluoresceine-labelled macromolecule then measured in the basolateral compartment (n = 4–5/group). Bars represent means, ±SEM; * denotes p < 0.05 (ANOVA with Bonferroni post hoc analysis). M denotes medium; E denotes EHEC; numbers 1 through 4 denote HMO composition preparations from individuals 1 to 4.

Figure 3

Effects of individual HMO blends on the expression of innate defense and intercellular tight junction proteins in Caco-2Bbe1 cells. Epithelial cells grown in 24-well plates were treated with each of the HMOs prepared from four individuals followed by EHEC O157:H7 challenge (4 h, MOI 100:1). mRNA was extracted to measure, using RT-PCR, levels of: (A) Muc1, (B) Muc2, (C) Zo-1 and (D) Cldn-1 (n = 3–4 per group). Western blotting shows protein levels of (E) ZO-1 and (F) Claudin-1 after 4 h of EHEC challenge with corresponding densitometry of immunoblots shown below each representative immunoblot (n = 4 per group). Densitometry ratios were measured using GAPDH as a loading control. The dotted line indicates spliced sites from a single original immunoblot. Values are expressed as means, ±SEM; * denotes p < 0.05 (ANOVA with Bonferroni post hoc analysis). M denotes medium; E denotes EHEC; numbers 1 through 4 denote HMO composition preparations from individuals 1 to 4.

Figure 4

Effects of individual HMO blends on immune signaling and chemokine/cytokine expression in Caco-2Bbe1 cells. Epithelial cells grown in 24-well plates were pre-treated with each of the four HMO preparations followed by EHEC O157:H7 challenge (3 h, MOI 100:1) and then assessed for (A) phosphorylation of ERK1/2 (n = 4–5 per group), (B) phosphorylation of P38 MAP kinase (n = 4–5 per group) and (C) NF-κB activation (n = 4 per group). Densitometry ratios were measured using corresponding non-phosphorylated MAPK or GAPDH as loading controls. Dotted lines indicate spliced sites from a single, original Western blot. At 4 h post-EHEC challenge, mRNA was extracted to measure levels of (D) Il8, (E) Il18, (F) Tgfβ and (G) Il10 (n = 3–4 per group). Data are expressed as means, ±SEM; * denotes p < 0.05 (ANOVA with Bonferroni post hoc analysis). M denotes medium; E denotes EHEC; numbers 1 through 4 denote HMO composition preparations from individuals 1 to 4.