Atopic dermatitis (AD) is a chronic inflammatory skin disease of mainly infants and children, which is generally referred to atopic eczema and is often accompanied by xeroderma, pruritus, and inflammation [1]. Exposure of our body to allergens causes imbalance in cytokines produced by Th1 and Th2 cells through excessive production of cytokines by Th2 cells, which stimulates B cells and subsequent secretion of IgE. The increased IgE migrates to the mast cells in the skin to combine with basophils, which are leukocytes. If a person is repeatedly exposed to the same allergen, the exocytosis of the compounds, such as histamine and leukotriene, stored in the mast cells will occur. Such products of exocytosis cause allergic symptoms and induce or aggravate AD that is accompanied by red spots on the skin, edema, or pruritus [2].

While it is well established that AD is caused by the overexpression of Th2 cytokine, recently thymic stromal lymphopoietin (TSLP) has been suggested as an important factor involved in the onset of AD. TSLP was found to be overexpressed in both acute and chronic AD patients, and mouse AD models, and increased mRNA expression of IL-4, IL-13, and TNF-α [3, 4]. The increased expression of these cytokines in turn triggers Th2-type reactions from dendritic cells [5], leading to decreased expression of filaggrin mRNA and subsequent damage to the skin barrier for the development of AD [6].

The anti-histamines and steroids currently used to treat AD are not suitable for long-term treatment or effective cure, because of their frequent adverse toxic effects [7]. Therefore, safe and effective methods for treatment of AD are in great demand and thus, studies searching for natural substances to treat AD effectively without side effects have been conducted actively [8]. Recently, various alternatives to existing AD treatment methods have been sought, including probiotics and bioactive compounds. Probiotics (i.e., living organisms), referring to microorganisms that exert beneficial effects in the body when taken [9], have few adverse effects, and are relatively safe as most of them are in vivo-derived bacteria [10]. Prebiotics (i.e., nondigestible fiber) are non-digestible food ingredients that have beneficial effects on host health by promoting the growth and activity of probiotics. Typical prebiotics include inulin, fructo-oligosaccharide and galacto-oligosaccharide [11]. Synbiotics (i.e., combination of pro- and prebiotics) is a mixture of microbial probiotics and nutrients prebiotics in an appropriate proportion, and synergistic effects occur because each physiological activity is simultaneously expressed by feeding probiotics and prebiotics together [12].

β-glucan is a group of polysaccharide that exists in the cell walls of yeasts, mushrooms, cereals, yeasts, algae, and bacteria, and has been explored as a bioactive substance [13, 14]. Particularly it is present at high levels in medical mushrooms and plays an important role in immune functions of normal cellular tissues by stimulating B and T cells through the production of various cytokines [15]. Although α-glucan is the most widely known among those pullulans that are produced by Aureobasidium pullulans strain [16], this strain also produces β-1,3/1,6 glucan that has an excellent antitumor activity against allogeneic sarcoma-180 [17]. L. plantarum, which belongs to lactic acid bacteria, is a noninvasive, nonpathogenic, and health-promoting Gram-positive commensal microorganism [18]. Various strains of lactic acid bacteria have been reported to reduce allergic symptoms in mouse models and humans [19, 20]. Previously, we have demonstrated that oral administration of β-1,3/1,6-glucan from Aureobasidium pullulans SM-2001 effectively attenuates AD-like phenotypes in animal models [21]. However, studies on the potential variation in the bioactivity of β-1,3/1,6-glucans from different sources and the synergistic immunomodulatory effects of β-1,3/1,6-glucan and L. plantarum on AD are still lacking.

Therefore, in the present study, we expanded upon our previous study to investigate the immunomodulatory effects of β-1,3/1,6-glucan and L. plantarum LM1004 on AD via oral administration of these substances separately or in combination to AD models of rat and mouse that exhibit AD-like symptoms of histamine-induced vasodilation (rat), allergic pruritus and contact dermatitis (mouse). Additionally, the changes in mRNA expression levels of the transcription factors associated with Th1, Th2, Th17, and Treg cells, cytokines, galectin, filaggrin, and TSLP in response to the same treatments were analyzed using the rat model of histamine-induced vasodilation to understand the immuno-modulatory mechanisms underlying the alleviation of AD. The fecal microbiomes were also analyzed to gain insights on the potential role of gut microbiome in the immuno-modulatory effects of β-1,3/1,6-glucan and/or L. plantarum LM1004. The results of the current study provide unique insights on how oral administration of β-1,3/1,6-glucan and/or L. plantarum LM1004 can alleviates AD-like symptoms via mechanisms that might involve modulation of host immune systems and gut microbiota.

Design for the Animal Experiments

The overall design for all four animal experiments are summarized in Table 1. In this study, we used three different AD models using either Sprague-Dawley rats or ddY mice to examine the immunomodulatory effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on different symptoms of AD after 7 days of daily oral administration. The AD models included rat model of histamine-induced vasodilation (Exp. 1), mouse model of pruritus (Exp. 2), and mouse model of contact dermatitis (Exp. 3). More details on each AD model are provided in the following sections. For each of the animal experiments, there were 6 groups, including 2 Control and 4 Treatment groups (Table 1): 1). Positive control group (C) with no disease induction and no treatment, 2) Negative control group (N) with disease induction and no treatment, 3) Wellmune (T1), 4) β-1,3/1,6-glucan (T2), 5) L. plantarum LM1004 (T3), and 6) β-1,3/1,6-glucan and L. plantarum LM1004 (T4).

Table 1 . Summary of the animal experiments..

Groups (10 animals/ group)		Rats	Mice		Rats
		Exp. 1. Histamine-induced vasodilation	Exp. 2. Pruritus	Exp. 3. Contact dermatitis	Exp. 4. No disease induction
Controls	C (+ control)	No disease induction; no treatment
	N (- control)	Disease induction; no treatment			N.A.
Daily oral administration 0.2 ml (0.01 g/kg initial body weight) for 7 days			0.4% (in diet) for 28 days
Treatments	T1	β-1,3/1,6-glucan extracted from baker’s yeast Saccharomyces cerevisiae (Wellmune)
	T2	β-1,3/1,6-glucan secreted from black yeast Aureobasidium pullulans SM-2001
	T3	L. plantarum LM1004
	T4	β-1,3/1,6-glucan from Aureobasidium pullulans SM-2001 and L. plantarum LM1004 (the same dose for each)
Measurements (organ or tissue used)		. Blue-dye spots (dorsa dermis)	. Incidences of scratching behavior	. Ear thickness	. Average daily feed intake
				. Incidences of scratching behavior
		. Leaked dye (dorsa dermis)	. Histamine (serum)		. Body weight gain
					. Feed conversion ratio
		. Histamine (serum)			. 16S microbiome profiling (feces)
		. IgE (serum)
		. Gene expression (mesenteric lymph nodes)

The β-1,3/1,6-glucan product used for T1 group (Wellmune; Wellmune Corp., Korea) was extracted from baker’s yeast Saccharomyces cerevisiae, whereas β-1,3/1,6-glucan used for T2 group (Glucan Corp., Korea) was the β-1,3/1,6-glucan with an average molecular weight of 2.6 × 10⁵ Da, secreted from black yeast Aureobasidium pullulans SM-2001 to the outside of the cells [22]. The purity of the β-1,3/1,6-glucan used in the present experiments was 15%. L. plantarum LM1004 (KCCM 43246; Lactomason Corp., Korea) was originally isolated from Kimchi (a Korean traditional fermentation food), and the lysophilized form of L. plantarum LM1004 at the concentration of 2 × 10¹² cells/g was used in this study. Each animal received daily dose of 0.2 ml (0.01 g/kg initial body weight at 6 weeks old) of the respective treatment via oral administration.

For Exp. 4, rats with no AD induction were raised on the feeds supplemented with β-1,3/1,6-glucan and/or L. plantarum LM1004 for 28 days and used to investigate the treatment effects on growth parameters, feed consumption, and fecal microbiome. The treatment groups for Exp. 4 were the same as those in Exp. 1-3 except that the treatments were supplemented to feeds instead of oral administration, and Negative control group (N; disease induction and no treatment) was not included (Table 1).

For these animal experiments, Sprague-Dawley (SD) rats (Samtako, Korea) and ddY mice (Central lab, Animal Inc., Korea) at the age of six weeks male were used. All procedures for handling animals were approved by the Institutional Animal Care Board of Gyeongnam National University of Science and Technology (Approval No. 2015-2). During the acclimation and experimental periods, the room temperature of 22 ± 1°C and humidity of 60 ± 10% were maintained, and the light/darkness cycles were adjusted to 12-h cycles. Solid pellet feed AIN-76 (Korea) and water were provided ad libitum unless described otherwise. After the acclimation period for one week, the experimental animals were divided into 6 (Exp. 1-3) or 5 groups (Exp. 4) of 10 animals per group for each animal experiment (Table 1).

Exp. 1. Rat Model of Histamine-Induced Vasodilation

AD was induced in SD rats using a histamine secretion promoter compound 48/80 (COM, Sigma Aldrich, USA) according to the method described earlier [23]. The SD rats in each treatment group received respective oral administration daily for 7 days. On day 7, 50 μl of COM (10 μg/ml) was intra-dermally injected into the shaved dorsal dermis, followed by administration of 200 μl of 0.5% Evans blue (Sigma Aldrich) into the tail vein after 30 min.

Measurement of blue dye spots and leaked dye. After 30 min post injection of Evans blue, the SD rats were sacrificed and the dorsal dermis was cut to measure the diameter (mm) of the region of extravasation of the blue dye due to vasodilation using a digital caliper (Bluebird, NA500-150S). The amount of leaked blue dye was measured as previously described [24]. Briefly, the dorsal dermis was cut out into 10 mm diameter-sized pieces and immersed into 1.0 N KOH (1.0 ml) solution, followed by incubation at 37°C for 48 h. Thereafter, 0.6 N H₃PO₄ (2.5 ml) and acetone (6.5 ml) were added to the solution and the mixture was homogenized and centrifuged (900 ×g, 10 min). The supernatant was spectrophotometrically analyzed at 620 nm, and the amount of leaked blue dye was measured and expressed in mg/site.

Measurement of histamine and IgE levels. After the rats were sacrificed at 30 minutes post injection of Evans blue, blood was collected from the vena cava and centrifuged for 15 minutes at 1,500×g to separate the serum. The quantities of IgE and histamine in the separated serum were measured using the Hitachi Optigen Allergen-specific IgE Assay System (Hitachi Chemical Diagnostics, USA) and Histamine EIA Kit (LDN, Germany), respectively.

Gene expression analysis. For analysis of gene expression, the vasodilated rats were sacrificed at the end of the 7 days of treatment, and the mesenteric lymph nodes were separated. The mesenteric lymph nodes were put into Trizol Reagent (Ambion, USA), and homogenized using Silent Crusher M (Heidolph, Germany). RNA was isolated according to the method described earlier [25] and stored at 20°C until analysis. cDNA synthesis was performed for 30 min at 50°C using a reverse transcription polymerase chain reaction (RT-PCR) kit (TaKaRa, Japan). The amplification cycle of qPCR was: initial denaturation (10 min at 95°C) followed by 35 cycles of denaturation (30 sec at 95°C), annealing (30 sec at 55°C), and extension (1 min at 72°C). Thereafter, final extension was performed for 5 min at 72°C. The primers used in the experiment are shown in Additional file 1: Table S1 [26-30]. The mRNA expression of the transcription factors (T-bet, GATA-3, RORγT and Foxp3) and cytokines (IFN-γ, IL-4, IL-17 and TGF-β) of Th1, Th2, Th17, and Treg cells, respectively, was analyzed using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a reference gene. Additionally, the gene expression of galectin-9, filaggrin, and TSLP was also measured.

Exp. 2. Mouse Model of Pruritus

Allergic pruritus was induced in ddY mice using a histamine secretion promoter (COM) as described earlier [23]. Each treatment group received respective oral administration daily for 7 days. On day 7, 10 ml/kg (3 mg/kg) of COM was intra-dermally injected into the shaved dorsal dermis.

Monitoring of scratching behavior. At 30 min post injection of COM, the number of times of pruritus appearance (the incidence of scratching behavior) on the entire body was observed for 20 min according to the method described earlier [31].

Measurement of histamine level. After the mice were sacrificed at 30 min post injection of COM, blood was collected from the vena cava and centrifuged for 15 min at 1,500 ×g to separate the serum. The quantity of histamine in the separated serum was measured using a Histamine EIA Kit (LDN, Germany).

Exp. 3. Mouse Model of Contact Dermatitis

Contact dermatitis was induced in ddY mice as previously described [32]. Briefly, 0.2 ml of a saline solution containing 10 μg of dinitrophenyl-derivatized ovalbumin (DNP-OVA, Alpha Diagnostic Intl. Inc., USA) and 1 mg of aluminum hydroxide gel (Sigma Aldrich) was intraperitoneally injected to sensitize the ddY mice before the initiation of the dietary treatments for 7 days.

Measurement of ear thickness and incidence of scratching behavior. On day 7 after the respective treatments by oral administration for 7 days, 10 μl of 0.1% 2,4-dinitrofluorobenzene (DNFB, Sigma) dissolved in ethanol was applied to the ear and paw regions at 1 h after the oral administration. Ear thickness was measured using digital calipers (Bluebird, NA500-150S) immediately before (0 h), one hour after (1 h), and 24 h (24 h) after the application of DNFB. Additionally, immediately after DNFB application, the incidence of scratching behavior on the entire body was observed for one hour according to the method described previously [33].

Exp. 4. Rat Study for Growth Responses and Gut Microbiome Analysis

To examine the effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on the intestinal microbiota, SD rats were provided with the feeds supplemented with respective treatments for 28 days. We reasoned that the longer period of treatment to the rats with no AD induction would allow more accurate assessment of the changes in growth parameters and gut microbiota in response to β-1,3/1,6-glucan and/or L. plantarum LM1004. The body weights and feed intakes of the rats were measured once per week. At the end of the experiment, fecal samples (4 samples per group) were collected for the analysis of fecal microbiome. The genomic DNA in the fecal samples was extracted using ZR Fecal DNA MiniPrep (Zymo Research Corporation, USA). The DNA samples were processed at Chunlab Inc. to amplify V3-4 region of 16S rRNA genes, and the resulting PCR products were combined in equal amount [34]. The composite sample was sent for sequencing analysis using Illumina Miseq at Chunlab Inc. and bioinformatics analysis was conducted using the pipelines described earlier [35].

Statistical Analyses

The data obtained from experimental replications were analyzed by using SPSS 12.0 (SPSS Inc., USA). Analysis of variance (ANOVA) followed by Duncan's multiple range test was used to identify the significant differences among various groups, where the level of significance was considered at p <0.05 [36].

Effects of β-1,3/1,6-Glucan and/or L. plantarum LM1004 on Vasodilation Changes in the Diameter of Blue Dye Spots

To analyze the effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on AD due to vasodilation induced by COM, the rats were sacrificed and the diameters of leaked Evans blue dye spots were measured as shown in Fig. 1A. The rats in the control (C) and the AD-induced negative control (N) group showed shortest and longest diameters of the leaked blue dye spots, respectively, as expected. All treatment groups (T1, T2, T3, and T4) showed significantly lower values than the AD-induced negative control group (N), while there was no significant difference among the treatment groups, suggesting the inhibitory effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on COM-induced vasodilation. There was no significant synergistic effect of β-1,3/1,6-glucan and L. plantarum LM1004 combined (T4) as compared to either β-1,3/1,6-glucan (T2) or L. plantarum LM1004 (T3).

Figure 1. The effects of oral administration of L. plantarum LM1004 and/or β-1,3/1,6-glucan on COM-induced vasodilation in rats as measured by the diameter (mm) of blue dye spots produced from extravasation of blue dye due to COM-induced vasodilation (A), the amount of leaked blue dye (B), the serum histamine level (C), and the serum IgE level (D). For details of the control (C and N) and treatment groups (T1-T4), see Table 1. Means that do not share the same superscript a to c are significantly different (p < 0.05). Data represent means ± SD of 10 replicates.

Changes in amounts of leaked dye

The amounts of leaked Evans blue in the control and treatment groups are shown in Fig. 1B. The control (C) and the AD-induced negative control (N) showed the lowest and the highest value, respectively. All treatment groups (T1, T2, T3, and T4) showed significantly lower values than the AD-induced negative control (N), thereby reconfirming the inhibitory effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on COM-induced vasodilation.

Changes in Serum Histamine Level

The serum histamine levels of the experimental groups are as shown in Fig. 1C. The AD-induced negative control (N) showed the highest level of serum histamine, and all treatment groups (T1, T2, T3, and T4) showed significantly lower levels of serum histamine than the AD-induced negative control (N), thereby suggesting AD- alleviating effects of β-1,3/1,6-glucan and/or L. plantarum LM1004.

Changes in Serum IgE Level

The serum IgE levels in the experimental groups are shown in Fig. 1D. The AD-induced negative control (N) showed the highest level of serum IgE content, and all treatment groups (T1, T2, T3, and T4) were significantly lower than the AD-induced control (N). There was no significant differences between the control (C) and all treatment groups (T1, T2, T3, and T4), suggesting the effective inhibition of serum IgE by β-1,3/1,6-glucan and/or L. plantarum LM1004.

Changes in the Gene Expression of Immune-Related Genes

To investigate the mechanism(s) by which β-1,3/1,6-glucan and/or L. plantarum LM1004 alleviate AD, the expressions of T-bet, GATA-3, RORγT, and Foxp3, which are transcription factors of Th1, Th2, Th17, and Treg cells, respectively, and cytokines IFN-γ, IL-4, IL-17, and TGF-β associated with respective T cells were analyzed. The mRNA expression levels of these genes normalized to the reference gene GAPDH are shown in Fig. 2. The mRNA expression of the transcription factors of Th1 (T-bet; Fig. 2A) and Treg cells (Foxp3; Fig. 2B) were significantly increased, whereas the mRNA expression of the transcription factors of Th2 (GATA-3; Fig. 2C) and Th17 (RORγT; Fig. 2D) were significantly reduced, in all treatment groups as compared to the AD-induced negative control (N). In accordance to the expression of the transcription factors of their respective T cells, the mRNA expression of the cytokines IFN-γ (Fig. 2E) and TGF-β (Fig. 2F) were significantly increased, whereas the mRNA expression of the cytokines IL-4 (Fig. 2G) and IL-17 (Fig. 2H) were significantly decreased, in all treatment groups as compared to the AD-induced negative control (N). These results indicate that β-1,3/1,6-glucan and L. plantarum LM1004 can induce a balance among Th1, Th2, Th17, and Treg cells to alleviate AD that occurs mainly due to imbalance in immune reactions among T cells. In addition, the significantly increased Th1/Th2 and Treg/Th2 ratios in the groups treated with L. plantarum LM1004 and/or β-1,3/1,6-glucan (T1, T2, T3, and T4) suggest that L. plantarum LM1004 and β-1,3/1,6-glucan can induce Th1 and Treg mediated immune reactions.

Figure 2. The effects of oral administration of L. plantarum LM1004 and/or β-1,3/1,6-glucan on mRNA expression of the transcription factors and cytokines in mice model. T-cell polarization in mesenteric lymph nodes (MLN) was evaluated by analyzing the mRNA expression of T-bet (A), GATA-3 (B), RORγT (C), Foxp3 (D), IFN-γ (E), IL-4 (F), IL-17 (G), and TGF-β (H). For details of the control (C and N) and treatment groups (T1-T4), see Table 1. Means that do not share the same superscript a to d are significantly different (p < 0.05). Data represent means ± SD of 4 replicates.

Changes in Gene Expression of Galectin-9, Filaggrin and TSLP

The results showing the mRNA expressions of Galectin-9, filaggrin, and TSLP are shown in Fig. 3. Galectin-9 has been reported to reduce IgE and induce Treg-mediated immune reactions [37, 38]. All treatment groups (T1, T2, T3, and T4) showed significantly higher expression levels of Galectin-9 as compared to both control groups (C and N). Within treatment groups, there were no significant difference observed except that T1 showed significantly higher expression level of Galectin-9 as compared to T3 and T4. The AD-induced negative control (N) showed the lowest level of filaggrin expression, which is a skin barrier modulator. All treatment groups (T1, T2, T3, and T4) showed the same levels as the control (C) except T1, which was significantly higher as compared to the AD-induced negative control (N). In addition, the expression of TSLP, which is used as a major biological marker of AD, was significantly reduced in all treatment groups (T1, T2, T3, and T4) as compared to the AD-induced negative control (N). The expression level of TSLP was not significantly different within the treatment groups.

Figure 3. The effects of oral administration of L. plantarum LM1004 and β-1,3/1,6-glucan on mRNA expression of the galectin-9, filaggrin, and TSLP in mice model. For details of the control (C and N) and treatment groups (T1-T4), see Table 1. Means that do not share the same superscript a to c are significantly different (p < 0.05). Data represent means ± SD of 4 replicates.

Effects of β-1,3/1,6-Glucan and/or L. plantarum LM1004 on Allergic Pruritus Changes in the Incidence of Scratching Behaviors

The results of the observation of acute pruritus (itchy skin as measured by scratching behavior) induced by COM are shown in Fig. 4A. The AD-induced negative control (N) showed the most frequent scratching behaviors, while the treatment groups (T1, T2, T3, and T4) showed significantly less frequent scratching behaviors as compared to the AD-induced negative control (N).

Figure 4. The effect of oral administration of L. plantarum LM1004 and β-1,3/1,6-glucan on COM induced acute pruritus in mice as measured by: the incidence of scratching behavior (A) and the serum histamine level (B). For details of the control (C and N) and treatment groups (T1-T4), see Table 1. Means that do not share the same superscript a to c are significantly different (p < 0.05). Data represent means ± SD of 10 replicates.

Changes in Serum Histamine Level

Fig. 4B shows the levels of serum histamine in all experimental groups. The AD-induced negative control (N) showed the highest serum histamine level, while the treatment groups (T1, T2, T3, and T4) showed significantly lower serum histamine levels than the AD-induced negative control (N). These results showed that the treatments were effective in suppressing histamine levels. This result also shows that pruritus, an index of AD, has a positive correlation with serum histamine levels.

Effects of β-1,3/1,6-Glucan and/or L. plantarum LM1004 on Contact Dermatitis Changes in Ear Swelling

To investigate the treatment effects on edema due to contact dermatitis induced by DNP-OVA and DNFB, the ear thickness (swelling) of mice was measured at 0, 1, and 24 h after treatment with DNFB (Fig. 5A). When DNFB was applied, the levels of ear thickness due to edema increased significantly at 1h and 24 h as compared to 0 h, demonstrating the effectiveness of edema induction by DNFB. At 1h and 24 h after DNFB application, the treatment groups (T1, T2, T3, and T4) showed significantly lower ear thickness as compared to their respective AD-induced negative controls (N), except for T1 at 1 h. This result showed that the oral administration of β-Glucan and/or L. plantarum LM1004 effectively alleviated edema due to contact dermatitis.

Figure 5. The effect of oral administration of L. plantarum LM1004 and β-1,3/1,6-glucan on contact dermatitis induced by DNP-OVA and DNFB in mice as measured by: the ear thickness (A) and the incidence of scratching behavior (B). For details of the control (C and N) and treatment groups (T1-T4), see Table 1. Means that do not share the same superscript a to c are significantly different (p < 0.05). Data represent means ± SD of 6 replicates.

Changes in the Incidences of Scratching Behaviors

To investigate scratching behaviors due to dermatitis induced by DNP-OVA and DNFB, the experimental animals were observed for an hour starting from one hour post-application with DNFB and the result is shown in Fig. 5B. The AD-induced negative control (N) showed the most frequent scratching behaviors, while all treatments (T1, T2, T3, and T4) significantly reduced scratching behaviors as compared to the AD-induced negative control (N), to the level similar to the control group (C).

Effects of β-1,3/1,6-Glucan and/or L. plantarum LM1004 on the Composition of Gut Microbiota in Rats

To study the changes in gut microbiota due to the treatments, fecal samples were subjected to 16S microbiome profiling. The data analyzed at the phylum level to determine the relative abundances of Firmicutes and Bacteroidetes are shown in Fig. 6A. Although the relative abundance of Firmicutes was not significantly different among all experimental groups, the relative abundance of Bacteroidetes was significantly higher in T1, T3, and T4 as compared to the control (C). In addition, the ratio of Firmicutes to Bacteroidetes (F/B ratio) was significantly lower in all treatment groups (T1, T2, T3, and T4) as compared to the control (C) (Fig. 6B).

Figure 6. The effects of L. plantarum LM1004 and β-1,3/1,6-glucan administration on microbiome composition in rats. Total genomic DNA was isolated from fecal samples of each group and the abundances of bacterial population at the phylum levelwere analyzed by next-generation sequencing. The relative abundance of Firmicutes and Bacteroidetes (A) and the ratio of Firmicutes/Bacteroidetes were calculated (B). In addition, the relative abundances of genera associated with obesity (C) and families/genus associated with butyrate-production were analyzed. For details of the control (C and N) and treatment groups (T1-T4), see Table 1. Means that do not share the same superscript a to c are significantly different (p < 0.05). Data represent means ± SD of 4 replicates.

The relative abundance of bacterial genera, such as Bacteroides, Alistipes, and Prevotella, which are often associated with reduced obesity, were also analyzed as shown in Fig. 6C. The combined relative abundance of the genera Bacteroides, Alistipes and Prevotella was significantly higher in all treatment groups (T1, T2, T3, and T4) as compared to the control (C).

Similarly, the relative abundance of butyrate-producing taxa (families Lachnospiraceae and Ruminococcaceae; genus Roseburia) are shown in Fig. 6D. For each of these taxa, the relative abundance was significantly higher in all treatment groups (T1, T2, T3, and T4) as compared to the respective control group (C), with the exception of T1 and T2 for the family Ruminococcaceae.

Effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on feed intake, daily weight gain, and feed conversion ratio

The results showing the effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on the average daily feed intake, body weight gain, and feed conversion ratios of the rats that were used for microbiome analysis are shown in Table 2. Interestingly, both finished body weight and average daily gain were significantly decreased in all treatment groups (T1, T2, T3, and T4) as compared to the control group (C).

Table 2 . The effects of β-1,3/1,6-glucan and L. plantarum LM1004 administration on body weight and feed consumption in among the experimental animal groups..

	Treatments
	C	T1	T2	T3	T4
Initial body weight (g)	224.17±21.67	224.23±20.72	224.66±21.05	225.10±21.62	225.08±21.19
Finished body weight (g)	370.78±19.64d	330.03±28.38ab	293.68±15.21a	342.44±22.88c	322.54±19.21b
Average daily gain (g)	5.24±1.00c	3.78±1.40b	2.46±0.86a	4.19±1.34b	3.48±1.36b
Average daily feed intake (g)	22.85±1.29c	18.86±0.95b	15.78±0.27a	19.01±1.34b	18.15±0.87b
Feed conversion ratio	0.23±0.05b	0.20±0.08ab	0.16±0.05a	0.22±0.06b	0.19±0.08ab

C: Control, T1: 0.4% diet of Wellmune, T2: 0.4% diet of β-1,3/1,6-glucan, T3: 0.4% diet of L. plantarum LM1004, T4: 0.4% diet for each of β-1,3/1,6-glucan and L. plantarum.

LM1004. Means that do not share the same superscript a to c are significantly different (p < 0.05). Data represent means ± SD of 10 replicates..

The present study was aimed to study the effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 on AD-like symptoms in AD models of rat and mouse. In an experiment conducted using a model of Type I dermatitis due to passive cutaneous anaphylaxis (PCA), the extravasation of blue dye due to vasodilation was noticed remarkably at 30 minutes post-injection of antigens [39]. The leakage of the blue dye caused by Evans blue extravasation was due to vasodilation induced by COM, which is a histamine secretion promoter [23]. Reduction in blue dye diameters and the amounts of leaked Evans blue by any treatments indicates that they are effective against AD-like symptoms through alleviation of vasodilation [23, 40, 41]. Therefore, the significant reductions in the diameter and amount of leaked blue dye in the present study (Figs. 1A and 1B) can be considered as an evidence showing the positive effects of β-1,3/1,6-glucan and/or L. plantarum LM1004 in alleviating AD-like symptoms through inhibition of vasodilation.

In the case of AD, the mast cells sensitized by IgE and histamine can lead to the occurrence of symptoms such as edema, erythema, and pruritus [42, 43]. Administration of COM resulted in edema through secretion of histamine from mast cells [44]. In general, when AD occurred, the IgE existing in the dermis recognizes allergy-inducing substances first. In acute disease of AD patients, the level of IFN-γ, which is Th1 cytokine, decreases to induce IgE overproduction and Th2 immune reactions [45]. Therefore, to alleviate AD, serum IgE contents should be regulated. Thus, the significant decrease in serum IgE contents shown in the present study (Fig. 1D) by oral administration of β-1,3/1,6-glucan and/or L. plantarum LM1004 can be considered as the positive evidence for the alleviation of AD.

COM is known to promote histamine secretion from mast cells [46]. Therefore, by promoting histamine secretion, COM induces scratching behaviors in mice [47]. Decrease in scratching behaviors indicates alleviation of allergic pruritus [48]. In the present experiment, oral administration of β-1,3/1,6-glucan and/or L. plantarum LM1004 decreased the scratching behaviors significantly as compared to the negative control (N) (Fig. 4A), suggesting effective inhibition of COM-induced pruritus.

When DNFB was repeatedly applied to the ears of mice, the level of serum IgE and ear thicknesses was reported to increase [49]. In addition, the application of DNFB induced the accumulation of the epidermis, the formation of dermatolysis, and the infiltration of inflammatory cells [50]. The contact dermatitis induced by DNFB was used as an evaluation model for Type I allergic dermatitis [33]. When DNFB is applied to the ears of mice after sensitizing the mice with DNP-OVA, two types of edema are induced. The edema occurring one hour after DNFB application is the immediate phase response (IPR), and the edema occurring 24 h after DNFB application is the late phase response (LPR) [51]. After DNFB application, scratching behaviors are observed during the IPR. The IPR is an inflammatory reaction occurred when chemical mediators are secreted from mast cells, whereas LPR is a cytokine-induced reaction. In the present study, oral administration of β-1,3/1,6-glucan and/or L. plantarum LM1004 inhibited ear thickness in both IPR and LPR (Fig. 5A), and significantly inhibited the scratching behavior accompanying IPR (Fig. 5B). Therefore, β-1,3/1,6-glucan and L. plantarum LM1004 can be considered to have antagonistic effects on allergic dermatitis and pruritus.

Th1 enhances the functions of macrophages by secreting cytokines such as IFN-γ, TNF-α, and IL-2, thereby improving cellular immune reactions. Th2 secretes cytokines such as IL-4, IL-5, and IL-13 to increase the production of antibodies by B cells, thereby activating humoral immune reactions [52, 53]. Th1/Th2 balance is maintained by the antagonism of those cytokines that are generated by Th1 and Th2, which are two major T cells that inhibit the activity of each other [52, 54]. Th2 cells secrete cytokines to promote the inflows and activation of allergy mediating cells such as B-cells, mast cells, and eosinophils and cause damage to tissues through fibrosis [55]. Activated immune cells release IgE, histamine, and cytotoxic substances to aggravate inflammatory responses, and induce clinical signs [56]. Inhibiting Th2 cell reactions, which are central to allergic reactions, is essential for treatment of allergic conditions. Th17 cells are known to characteristically secrete IL-17A, IL-17F, and IL-22, which can play an important role in inflammatory responses and autoimmune diseases [57]. IL-17 induces the secretion of inflammatory cytokines to accelerate inflammation and induce the differentiation of inflammatory cells [58]. Treg cells secrete immune inhibitory cytokines such as IL-10 and TGF-β, and therefore they not only inhibit excessive inflammatory responses, but also induce immune tolerance for harmless antigens or autoantigens [59].

Galectin-9 is produced by activated immune cells which enhances Th1 immune responses from CD4+ T cells to promote the secretion of IFN-γ [60], inhibits Th17 immune responses, and increases Treg cell induction [38]. There is a report indicating that when synbiotics were taken after the induction of AD, the secretion of galectin-9 from intestinal epithelial cells increased to induce Th1 and Treg responses [61]. The expression of filaggrin was up-regulated through the specific activation of aryl hydrocarbon receptors (AHR)[62]. AHRs are core molecules that establish intestinal microbiota communities by affecting the balance between Th17 and Treg cells [63]. Decrease in the expression of filaggrin increases Th2 cytokine and are associated with the rise of IgE and the functional disorder of skin barriers [64]. Thus, L. plantarum LM1004 and β-1,3/1,6-glucan can provide beneficial effects to suppress AD-like symptoms by increasing the expression of filaggrin through AHRs. TSLP was found to express in keratinocytes of the dermis, epithelial cells, smooth muscle cells, and fibroblasts of the lung and stimulates mast cells and NK-T cells to cause excessive Th2 and Th17 immune reactions [65]. In the negative control (N), mRNA expression of both TSLP (Fig. 3C) and cytokines of Th2 and Th17 (Figs. 2G and 2H) were significantly higher as compared to all treatment groups, suggesting the another possible mechanism by which β-1,3/1,6-glucan and/or L. plantarum LM1004 can ease AD-like symptoms.

One of the known action mechanisms of β-glucan is mediated by pathogen-associated molecular patterns (PAMPs) such as dectin-1, Toll-like receptor (TLR)-2, TLR-4, TLR-6, and complement receptor (CR) 3 to stimulate the secretion of cytokines, thereby exhibiting immunomodulatory effects [66, 67]. Oral administration of Paramylon (β-1,3-D-glucan) after the induction of AD in NC/Nga mice inhibited Th2 cell responses as reported earlier [68]. Probiotics such as Lactobacilli can induce Th1 responses not only in the intestinal immune systems including Peyer's patch and mesenteric lymph nodes, but also in the entire immune system including internal blood [69, 70]. Most probiotics induce Th1 type cytokines such as IFN-γ and IL-12, which play an important role for the inhibition of IgE production in human B-cells [71].

At least 90% of the intestinal microbiota is composed of Firmicutes and Bacteroidetes. The relative abundance levels of Firmicutes and Bacteroidetes were found to be associated with obesity, since it has been observed that the increase in body weight was associated with the increase in the ratio of Firmicutes to Bacteroidetes [72]. High-fiber diets and prebiotics such as galacto-oligosaccharides (GOS) were shown to decrease the ratio of Firmicutes to Bacteroidetes by increasing the relative abundance of Bacteroidetes [73, 74]. In the present study, when the feed of experimental rats were supplemented with β-1,3/1,6-glucan and/or L. plantarum LM1004, the relative abundance of Bacteroidetes increased (Fig. 6A) and the average daily gain significantly decreased (Table 2) as compared to the control (C). Therefore, β-1,3/ 1,6-glucan and/or L. plantarum LM1004 might be considered as an effective dietary supplements for weight loss.

At the genus level, Bacteroides, Alistipes, and Prevotella have been reported as anti-obesity microorganisms [72, 75]. In the present study, when experimental rats were treated with β-1,3/1,6-glucan and/or L. plantarum LM1004, these anti-obesity microorganisms were increased significantly as compared to the control (C) (Fig. 6C).

Lachnospiraceae and Ruminococcaceae at the family level and Roseburia at the genus level have been reported as butyrate-generating microorganisms [76, 77]. Butyrate is known to stimulate Treg cells to increase IL-10 production [78], and has intestinal anti-inflammatory effects and immunoregulatory functions to regulate T cell differentiation and proliferation. Given that the relative abundance of butyrate-generating microorganisms increased when the experimental rats were treated with β-1,3/1,6-glucan and/or L. plantarum LM1004 in the present study (Fig. 6D), the increased production of butyrate via modulation of gut microbiota might be one of the mechanisms by which β-1,3/1,6-glucan and/or L. plantarum LM1004 alleviate AD-like symptoms.

In the present study, β-1,3/1,6-glucan and/or L. plantarum LM1004 inhibited Th2 cell responses, which are central to allergic reactions, while activated the responses of Treg cell and galectin-9, which have immunoregulatory functions. In addition, oral administration of β-1,3/1,6-glucan and/or L. plantarum LM1004 increased and decreased the expression of filaggrin and TSLP, respectively. Moreover, β-1,3/1,6-glucan and/or L. plantarum LM1004 increased the relative abundance of bacterial taxa associated with anti-obesity and butyrate productions. Therefore, the alleviating effects of L. plantarum LM1004 and β-1,3/1,6-glucan on AD-like symptoms as reported here may be achieved through regulation of immune responses of Th1 and Treg cells as induced by increased secretion of galectin in the dendritic and intestinal epithelial cells of the AD animal models. These regulation of immune responses may be possible through a mechanism(s) that involves modulation of gut microbiota.

However, the interpretation of the results presented in this study should be done with cautions due to the inherent limitations of the animal models used in this study. There are many new AD-models that have been developed in recent years [79, 80], and further evaluation of the β-1,3/ 1,6-glucan and/or L. plantarum using these new models would expand our knowledge on these promising compounds with the great potential for alleviating AD.

JMB-29-1693_Supple.pdf

This research was conducted with the aid of the Industry Core Technology Development Project (Nos. 10049026 and 10063302), Ministry of Trade, Industry, and Energy, Korea.

The authors have no financial conflicts of interest to declare.