1. Introduction
Early life nutrition is crucial for life-long health, and strongly influences the development of the innate immune system and the composition of the intestinal microbiota [
1,
2]. The World Health Organization recommends exclusive breastfeeding for infants up to an age of 6 months [
3]. Human milk oligosaccharides (HMOs) are an important component of mother milk which are considered to be prebiotics that guide immune development and gut barrier maturation [
4,
5]. However, for many reasons, in Europe, only approximately 25% of infants receive exclusive breastfeeding up to an age of 6 months [
6]. For infants for which breast milk is not an option, different types of infant formula are commercially available, which are often supplemented with non-digestible carbohydrates (NDCs) to mimic the effects of HMOs. Commonly used NDCs in infant formula include fructo-oligosaccharides (FOS) and galacto-oligosaccharides (GOS), which are added either individually or in combination [
7,
8]. The beneficial effects of FOS and GOS have been studied extensively [
7,
8,
9]. However, there are many other potential NDCs with beneficial effects on the microbiota composition and the immune system, presumably by simulating the conformational structure of HMOs rather than the constituent monosaccharides [
10]. An example of such NDC that could be of interest as a supplement in infant formulas is β-glucan [
11].
β-glucans are polymers of D-glucose building blocks linked by β(1→3), β(1→6) or β(1→4) linkages [
12], and can be found in many food and feed components including cereals, yeast and mushrooms [
13]. β-glucans from cereals, like oat and barley, are linear β(1→3) and β(1→4) linked glucose polysaccharides, whereas β-glucans from yeast and fungi consist of a β(1→3) linked glucan backbone with β(1→6) linked glucose residues as sidechains [
12]. Even though β-glucans are comprised of different building blocks than HMOs, e.g., β-linked glucose moieties rather than β-linked glucose, galactose and N-acetylglucosamine that form the backbone of HMOs, β-glucans have been shown to partially resemble the immune modulating functions of HMOs. These effects are exerted via binding to different pattern recognition receptors (PRRs) including C-type lectin Dectin-1, complement receptor 3 and Toll-like receptors (TLRs) [
14,
15,
16]. For oat β-glucans, it has also been shown that microbiota-derived enzymes may enhance their immune activity upon digestion by increasing Dectin-1 stimulation. This was demonstrated by the ability of oat β-glucans to stimulate Dectin-1 receptors, which increased after enzyme treatment with endoglucanase [
17].
Next to their direct effects, β-glucans might also influence immunity by changing the composition of the gut microbiota and the formation of short chain fatty acids (SCFAs) [
18]. Previous studies demonstrated that the fermentability of β-glucans is dependent on their structure. Enzymatic pretreatment of barley β-glucans enhanced their fermentability by adult fecal microbiota in an in vitro setup [
18]. Depolymerization of barley β-glucans by acid hydrolysis into fractions with MWs ranging from 6–104 kDa has also been shown to increase glucan fermentability with the inoculum of 9–15 month-old infants [
19]. However, currently, only limited knowledge is available about the effects of fermentation of β-glucans on their immune-modulating properties, and whether an increase in fermentability, due to depolymerization, affects these processes.
In the present study, native oat β-glucan was treated with an enzyme preparation containing predominantly endo-1,3(4)-β-glucanase to decrease the molecular weight. Both native and enzyme-treated oat β-glucan were fermented in an in vitro set-up using the infant fecal inoculum of 2- and 8-week old infants. We decided to use the fecal inoculum of 2- and 8-week-old infants because microbiota compositions are rapidly developing in the first weeks of life [
20] and different sets of enzymes for the degradation of NDCs will become available, depending on the bacteria present in the infant gut. Therefore, it is likely that both age classes will have different fermentation capabilities. The glucan degradation kinetics, impact on microbiota composition and SCFA production during fermentation were studied. In addition, the effects of the fermentation digesta on Dectin-1 receptor binding and immature dendritic cells cytokine production were studied.
4. Discussion
Non-digestible carbohydrates are often added to cow milk-based infant formula to substitute HMO functions. Commonly used NDCs are FOS and GOS; however, β-glucans might also have potential as possible addition to infant formula, as these molecules also stimulate the growth of beneficial microbes [
18,
19] and have immune modulating effects [
15,
16]. However, at present, only limited knowledge is available about the effects of fermentation of oat β-glucans on their immune-modulating properties. In this study, the degradation of native and enzyme-treated oat β-glucan by infant fecal microbiota was studied, as well as the effect of the fermentation digesta on cytokine production by dendritic cells and dectin-1 receptor activation, i.e., the receptor for β-glucans on immune cells [
37].
Although it has been demonstrated in vitro that adult fecal microbiota can utilize native oat and barley β-glucans in the size range of 130–243 kDa [
18], native oat β-glucan was not degraded by the fecal microbiota of either 2- or 8-week-old infants in the present study. This finding, however, is in line with the differences observed in GI microbial functionality in people of different ages [
38]. Bacterial species belonging to the
Clostridium histolyticum group were mainly involved in the described utilization in adults, which was corroborated by the presence of genes encoding endo-β-glucanase for several members of this group, e.g.,
C. longisporum [
39] and
C. acetobutylicum [
40]. The stimulation of
Clostridium was also observed upon intake of barley β-glucans in an in vivo study with healthy adult volunteers [
41]. In our study, using infant microbiota, bacteria belonging to the genus
Clostridium sensu stricto 1 comprising many different
Clostridium species [
31] were detected. Nevertheless, the absence of fermentation of native oat β-glucan by infant fecal microbiota could possibly be ascribed to differences at the species level, since the presence and abundance of specific
Clostridium species gradually changes upon aging [
42].
Instead, a clear increase in
Enterococcus was observed upon fermentation of media supplemented with native oat β-glucan by the fecal microbiota of 2- and 8-week-old infants, which was more pronounced than in the control fermentations without added native oat β-glucan. As such, it is suggested that the presence of the unfermentable native oat β-glucan stimulated the growth of Enterococcus on SIEM medium components. As enterococci are known to be among the first colonizers of the infant gut [
43], the increase in enterococci might be relevant in infants for creating a new environment that allows colonization of the gut by strict anaerobes [
2,
20]. A similar stimulation of
Enterococcus was shown in an in vitro fermentation study of barley β-glucans using the fecal microbiota of 9–15-month old infants which received both human milk and solid foods [
19]. With these age groups, substantial production of SCFAs was also observed, which indicates that the fecal microbiota of 9–15-month-old infants are capable of degrading native oat β-glucan. It is therefore likely that differences in
Enterococcus on species level upon aging [
44,
45] result in the availability of different sets of enzymes for the degradation of β-glucans for infants of different age groups. Our findings suggest that the bacteria and their enzymes which are responsible for the degradation of β-glucans are not yet present the first weeks after birth, but are introduced over time.
In contrast to native oat β-glucan that was not fermented by infant fecal microbiota, endo-1,3(4)-β-glucanase-treated oat β-glucan was degraded by the fecal microbiota of both 2- and 8-week old infants. This result corroborates the findings of Lam et al. [
19], who observed an increase in SCFA production when lowering the molecular weight of barley β-glucans in an in vitro fermentation study using infant fecal inocula. We also found that decreasing the size of oat β-glucans by enzyme treatment resulted in a stronger stimulation of
Enterococcus and increased production of lactic acid, which were more pronounced in the fecal inoculum of 2-week-old infants. Our data suggest that the fermentation of enzyme-treated oat β-glucan resulted in a selective increase of
E. faecium and
E. faecalis using the fecal inoculum of 2- and 8-week-old infants, respectively. Notably, for both inocula, there was a clear preference for the β(1→4)-linked oligomers present in the enzyme-treated oat β-glucan. Not much is known about specific carbohydrate-degrading enzymes and their structural preferences expressed by
Enterococcus species. However, a previous in vitro study using adult fecal inoculum also concluded that lower molecular weight barley β-glucans results in the stimulation of
Enterococcus and
Lactobacillus [
18].
The fermentation digesta of both native and enzyme-treated oat β-glucan could attenuate multiple pro-inflammatory cytokine responses in immature dendritic cells. The digesta of the control fermentation with the inoculum of either 2- or 8-week-old infants induced high levels of the chemokines and cytokines MCP-1/CLL2, MIP-1α/CCL3, IL-1β, IL-6 and TNFα, as well as the anti-inflammatory cytokine, IL-10. This effect was only observed for the
t = 14 and
t = 26 digesta, and therefore, was likely caused by products formed during fermentation, like metabolic products of protein fermentation and other bacterial metabolites such as ATP, lipoteichoic acid, polysaccharide A, peptidoglycan, RNA/DNA sequences and exopolysaccharide [
46]. The most pronounced attenuating effect by the digesta of medium supplemented with native and enzyme-treated oat β-glucan was observed for IL-6 production. The strongest attenuation of IL-6 was observed for the digesta of fermentations with the fecal inoculum of 2-week-old infants. Incubation with
t = 14 and
t = 26 digesta resulted in a significantly lower induction of IL-6 compared to the controls. In addition, the
t = 26 digesta of enzyme-treated oat β-glucan resulted in a significantly lower IL-6 production compared to the
t = 26 digesta of native oat β-glucan. The observed difference in DC stimulating capacity could possibly be explained by the formation of lactate which was only observed during the fermentation of enzyme-treated oat β-glucan. Lactate is the main end-product of the fermentation of carbohydrates by Enterococcus [
47], and has been reported to modulate cytokine responses, resulting predominantly in reduced inflammation in PBMCs and monocytes [
48].
Despite the absence of glucan degradation and lactate formation, the digesta of medium supplemented with native oat β-glucan displayed significant immune attenuating effects. As with the fermentation of both medium supplemented with native and enzyme-treated oat β-glucan, a clear increase in
Enterococcus was observed; it is suggested that the microbiota composition plays an important role in the attenuation of pro-inflammatory cytokine responses. The unfermentable native oat β-glucan resulted in a lower abundance of Enterococcus after 26 h of fermentation than the fermentable enzyme-treated oat β-glucan. Next to the SCFAs, differences in the relative abundance of
Enterococcus could thus also possibly explain the difference in DC stimulation capacity between native and enzyme-treated oat β-glucan digesta. Although the bacterial content was removed from the digesta before incubation with DCs in our experiments, the presence of soluble immune-active bacterial fragments in the digesta might be possible, as exemplified for the supernatant of a
Lactobacillus casei cell wall extract [
49].
Enterococcus have been associated with immunomodulatory functions and gut health in several studies. For example, a cohort study by Bjorksten et al. found a negative correlation between the colonization of
Enterococcus in the first year of life and the development of allergies [
50]. Other in vivo studies observed a correlation between
E. faecium and the reduction of infections in the gut, as reviewed by Franz et al. [
51]. There are also multiple studies demonstrating the immunomodulatory properties of
E. faecalis on different cell types. For example,
E. faecalis could upregulate IL-10 through PPARγ1 activation in colonic cell lines [
52]. A study by Wang et al. showed that several strains of
E. faecalis isolated from newborn infants downregulated IL-8 secretion in Caco-2, HT-29 and HCT116 intestinal cell lines [
53]. Interestingly, the immune attenuating effects were virtually gone when the carbohydrate moieties on the cell surface of
E. faecalis were oxidized. This indicates the carbohydrates present on the cell surface of
E. faecalis are the major effector molecules for regulating IL-8 secretion [
53]. An additional study by Wang et al. demonstrated that IL-8 secretion in intestinal cell lines is attenuated by
E. faecalis through the inhibition of MAPK signaling pathways. It was suggested that this effect was also caused by factors on the exterior cell wall of the bacteria [
54]. As such, it can be hypothesized that the enzyme treatment induced the immune-attenuating effect of native oat ß-glucan due to the increased stimulation of
Enterococcus and consequent production of lactate during fermentation.
As Dectin-1 receptors are important β-glucan binding receptors present on DCs [
55], we tested the immune stimulating capacity of our digesta on Dectin-1 reporter cell lines. The
t = 0 digesta of medium supplemented with the native and enzyme-treated oat β-glucan did not stimulate the Dectin-1 receptors. This is in line with an earlier study showing that particulate β-glucans induce stronger immune responses than soluble β-glucan [
15], as, in our study, we used soluble β-glucan, and any insoluble material was removed from the digesta by centrifugation and filtration prior to incubation with the receptors. The absence of any insoluble material could possibly also explain why we did not observe an increase in Dectin-1 stimulation after endoglucanase treatment, as observed before [
17].
Both the Dectin-1a and Dectin-1b receptors were strongly activated by the
t = 14 and
t = 26 digesta of both native and enzyme-treated oat β-glucan, fermented with the inoculum of either 2- or 8-week-old infants. It is unlikely that conformational changes of oat β-glucan played a role in the receptor activation, as reported elsewhere [
56], as the stimulatory effects of the
t = 14 and
t = 26 digesta of the unfermentable native oat β-glucan and the fermented enzyme-treated oat β-glucan were very similar. This finding also indicates that the degradation of the enzyme-treated oat β-glucans and the consequent lactate production were not responsible for the observed effects. It is tempting to speculate that the observed Dectin-1 activation is an (in)direct consequence of the increase in
Enterococcus. Although it is commonly accepted that fungal β-glucan with β(1→3) glycosidic bonds is an agonist of Dectin-1, there are reports of Dectin-1 activation by β(1→1) trehalose [
57], mannoprotein [
57], lipopeptide [
58] and bacterial cell walls [
49,
59]. As such, it could be hypothesized that the complex cell wall of
Enterococcus containing, amongst others, lipoteichoic acid, surface proteins and capsular polysaccharides [
60], plays a role in the activation of Dectin-1 receptors, as observed in our study. The activation of Dectin-1 receptors has been reported to have immune effects. For example, in macrophages, Dectin-1 regulates IL-10 and induces regulatory macrophage markers [
61]. The ability of the
t = 14 and
t = 26 digesta to stimulate Dectin-1 activation, therefore, might contribute to the immune attenuating effects we observed in DCs.