Lactobacillus paracasei KBL382 administration attenuates atopic dermatitis by modulating immune response and gut microbiota

ABSTRACT Administration of probiotics has been linked to immune regulation and changes in gut microbiota composition, with effects on atopic dermatitis (AD). In this study, we investigated amelioration of the symptoms of AD using Lactobacillus paracasei KBL382 isolated from the feces of healthy Koreans. Mice with Dermatophagoides farinae extract (DFE)-induced AD were fed 1 × 109 CFU d−1 of L. paracasei KBL382 for 4 weeks. Oral administration of L. paracasei KBL382 significantly reduced AD-associated skin lesions, epidermal thickening, serum levels of immunoglobulin E, and immune cell infiltration. L. paracasei KBL382-treated mice showed decreased production of T helper (Th)1-, Th2-, and Th17-type cytokines, including thymic stromal lymphopoietin, thymus, and activation-regulated chemokine, and macrophage-derived chemokine, and increased production of the anti-inflammatory cytokine IL-10 and transforming growth factor-β in skin tissue. Intake of L. paracasei KBL382 also increased the proportion of CD4+ CD25+ Foxp3+ regulatory T cells in mesenteric lymph nodes. In addition, administration of L. paracasei KBL382 dramatically changed the composition of gut microbiota in AD mice. Administration of KBL382 significantly ameliorates AD-like symptoms by regulating the immune response and altering the composition of gut microbiota.


Introduction
The prevalence of atopic dermatitis (AD) is increasing worldwide. 1 AD can be caused by numerous etiological factors, 2,3 and typically characterized by the T helper (Th)2-dominated immune response. 4 For example, IL-4, a key cytokine involved in atopic inflammation, is secreted from mast cells and drives the differentiation of T helper cells into Th2 cells. Further, it produces Th2-related cytokines and induces immunoglobulin (Ig)E isotype. 4,5 Infiltration of inflammatory cells such as eosinophils and lymphocytes into skin lesions is regulated by Th2 chemokines 6,7 and induced by IL-4, IL-5, and IL-13, potent eosinophil cytokines released from activated mast cells. 4,8,9 In addition, Th2derived IL-31 and thymic stromal lymphopoietin (TSLP) produced by keratinocytes can promote the sensation of itching, leading to scratching that further exacerbates skin eruption. 2,10,11 In addition to Th2 cells, Th1 and Th17 cells are also associated with AD progression. The acute phase of AD occurs at the onset of Th1-cell activation. 12 Also, Th17 cells play a role in the development of AD, acting as an initial cytokine source for the development of skin lesions. 13 In AD patients, transepidermal water loss (TEWL) increases due to defects of the intercellular lipid bilayers in the stratum corneum. 14 TSLP-stimulated dendritic cells (DCs) induce differentiation of naive T cells into Th2 cells and Th17 cells, leading to allergic inflammation of the skin. 15 In addition, thymus and activation-regulated chemokine (TARC) and macrophage-derived chemokine (MDC), which are induced by keratinocytes, trigger Th2 development. 16,17 Gut microbiota and their repertoire of biochemical reactions contribute to many aspects of host health, including metabolism, immunity, development, and behavior. Microbial dysbiosis, which is an imbalance of the microbial community, can contribute to the development of numerous diseases. 18 Dysbiosis of gut microbiota is a driver of the development of autoimmune diseases such as AD.
Probiotics, also referred to as live biotherapeutic products (LBPs), are live microorganisms that can confer beneficial health effects on the host when administered in adequate amounts. 19 Probiotics are a potential agent for preventing the symptoms of AD. 20 Lactic acid bacteria are the most commonly used probiotics, with emerging applications for the treatment of various diseases, including AD. [21][22][23] Previous studies have shown that oral administration of probiotics prevents the development of AD by inhibiting skin inflammation 4,24,25 or through the generation of CD4+ Foxp3+ regulatory T (Treg) cells, 26 which migrate to the skin from lymph nodes. 27 Administration of certain Lactobacillus strains can prevent the development of skin legions in an AD model through induction of inhibitory cytokines such as IL-10. 24 In addition, some Lactobacillus strains reverse gut dysbiosis-related diseases by regulating intestinal homeostasis. 28,29 We investigated the effects of Lactobacillus strains on AD-like symptoms in terms of immunomodulation, the composition of cecum microbiota, and their metabolites.

KBL382 treatment alleviates Dermatophagoides farinae extract (DFE)-induced AD-like symptoms
The overall experimental procedure is illustrated in Figure 1A. When the severity of dermatitis was analyzed visually from photographs, we found that oral administration of KBL382 ameliorated the development of AD-like lesions compared to DFE +PBS mice, whereas no difference was observed between KBL365-treated and DFE+PBS mice ( Figure 1B). On week 4, oral administration of KBL382 significantly suppressed the development of dermatitis scores compared to DFE+PBS mice. However, no difference was observed between KBL365-treated mice and DFE+PBS mice ( Figure  1C). We also measured the thickness of ear and dorsal skin from 0 to 4 weeks. On week 4, the thickness of KBL365-treated, DFE+PBS mice were similar. By contrast, that of KBL382-treated mice was significantly reduced compared to KBL365treated or DFE+PBS mice ( Figure 1D and E). Furthermore, AD-like symptoms induced by 2, 4-dinitrochlorobenzene (DNCB) were also ameliorated in KBL382-treated mice only (Figure S1A-D).

KBL382 treatment enhances skin function in dorsal skin lesions and induces immune homeostasis including IgE activity in AD
We measured TEWL and corneometer units, which indicated hydration and moisture in skin, as well as observing scratching behavior, the ratio of spleento-body weight, and IgE levels. Although the TEWL was similar between KBL365-treated mice and DFE +PBS mice, that of KBL382-treated mice was significantly lower than that of DFE+PBS mice, in accordance with the symptoms observed in vivo ( Figure 2A). Oral administration of KBL382 significantly increased corneometer unit compared to DFE+PBS mice. On the other hand, oral administration of KBL365 was not related to skin moisture ( Figure 2B).
The scratching behavior of KBL365-treated mice tended to be less severe than that of DFE+PBS mice, but no significant differences were observed. By contrast, oral administration of KBL382 significantly reduced the frequency of scratching behavior ( Figure 2C). In addition, we compared the weight ratio of the spleen among groups to assess the immune response, in light of previous research showing that an enlarged spleen indicates abnormal immune system function in AD. 30 Oral administration of KBL382 led to a significantly lower spleen-tobody weight ratio compared to DFE+PBS mice ( Figure 2D). We also investigated the ameliorative effects of KBL382 on serum levels of IgE. As observed in DFE+PBS mice, AD was accompanied by a marked increase in IgE concentrations ( Figure 2E). Compared to DFE+PBS mice, the IgE concentrations of KBL382treated mice were significantly lower. Meanwhile, KBL365-treated mice were comparable to DFE+PBS mice in terms of IgE ( Figure 2E). In DNCB-induced model, the spleen-to-body ratio and serum levels of IgE in KBL382-treated mice were also significantly lower than PBS-treated mice ( Figure S2A-B).

KBL382 treatment reduces infiltration of eosinophils and mast cells into dorsal skin lesions
DFE+PBS mice showed severe AD-like lesions and hyperkeratosis of the skin. However, administration of KBL382 led to a decrease in epidermal thickness in skin with AD ( Figure 3A). To evaluate skin lesions, eosinophils and mast cells were stained with Congo red and toluidine blue, respectively. The numbers of both eosinophils and mast cells were significantly reduced in KBL382-treated mice compared to DFE+PBS mice, while there were no significant differences between KBL365-treated mice and DFE+PBS mice ( Figure 3B-E). Also, administration of KBL382 led decrease in epidermal thickness of skin lesion in mice with DNCBinduced AD ( Figure S1E).

KBL382 treatment modulates TSLP, pro-inflammatory chemokines, and cytokines in dorsal skin lesions
The levels of TSLP, TARC, and MDC were significantly reduced in KBL382-treated mice compared to DFE+PBS mice ( Figure 4A-C). However, they were not significantly different in KBL365-treated mice compared to DFE+PBS mice ( Figure 4A-C). We further evaluated the immune response related to T cells based on protein and mRNA levels in the skin. The levels of Th1-(IL-2 and IFN-γ), Th2-(IL-4, IL-5, IL-13, and IL-31), and Th17-(IL-17A) were significantly suppressed in KBL382-treated mice, whereas there were no significant differences between KBL365-treated and DFE+PBS mice ( Figure 5A-E, G, and H). By contrast, significantly elevated levels of IL-10, transforming growth factor (TGF)-β and Foxp3 expression were observed in KBL382-treated mice compared to DFE+PBS mice. Oral administration of KBL365 also increased the IL-10 level and TGF-β expression compared to DFE+PBS mice, but this effect was not significant (figure 5F and I-J). IL-4, IL-5, and IL-13 were significantly decreased, and IL-10 and Foxp3 were clearly elevated in the DNCB-induced AD mice with KBL382 treatment ( Figure S2C-E, H, and I).

KBL382 treatment increases the Treg cells population in the mesenteric lymph node (MLN)
We assessed whether oral administration of KBL365 or KBL382 leads to an increase in Treg cells in the MLN. The proportions of CD4+ CD25+ Foxp3 + Tregs cells in KBL382-treated mice were significantly elevated compared to that of DFE+PBS mice. Oral administration of KBL382 promoted the generation of Tregs in MLN to modulate the immune response. Administration of KBL365 also increased the CD4+ CD25+ Foxp3+ Tregs cells compared to DFE+PBS mice, but this effect was not significant ( Figure 6A and B).

KBL365 or KBL382 treatment alters the cecum microbiota in AD
At the genus level, the bacterial microbiota of Control mice was mainly composed of S24-7_unclassified (18.2%) followed by Bacteroides (4.8%), while DFE+PBS mice reduced abundances of S24-7_unclassified (9.7%) and increased Bacteroides (8.5%). Mice administered the two strains showed a microbiota similar to Control mice, as S24-7_unclassified (20.2%) and Bacteroides (5.6%) were common in KBL365-  treated mice and these taxa (20.5% and 8.7%, respectively) were also abundant in KBL382-treated mice ( Figure 7A). Chao1 and phylogenetic diversity (PD), indicating bacterial alpha diversity,  revealed no differences among the four groups ( Figure 7B). Principle coordinate analyses (PCoA) of the weighted UniFrac distances indicated that bacterial composition (beta diversity) in the gut separated distinctly in groups of KBL365 or KBL382 administration compared to the group with DFE application only ( Figure 7C). In further taxonomic analyses, the two Lactobacillusadministration groups had significantly higher levels of S24-7_unclassified compared to DFE +PBS mice. KBL365 significantly decreased Mucispirillum and Bacteroides, for a composition similar to Control mice. KBL382 significantly increased the relative abundance of Akkermansia compared to DFE+PBS mice, with much higher levels than that in control mice ( Figure 7D and E).

KBL365 or KBL382 treatment regulates metabolites in cecum
DFE+PBS mice exhibited increases in environmental information processes, genetic information processes, and cellular processes ( Figure 8A). By contrast, KBL382-treated mice showed involvement of pathways such as metabolism and biosynthesis. Next, we investigated amino acids, nonvolatile acids, and short-chain fatty acid (SCFA) such as propionate in cecum. KBL382treated mice showed significant increases in the concentrations of both nonvolatile acids such as lactate, succinate, and fumarate, and SCFA such as propionate compared to DFE+PBS mice ( Figure 8B). On the other hand, KBL382-treated mice showed significant decreases of aspartic acid, serine, and methionine compared to DFE+PBS mice ( Figure 8C).

Discussion
Our study demonstrated that the administration of specific Lactobacillus strain could significantly alleviate the symptoms of AD via modulation of the immune response and gut microbiota. We used DFE and DNCB, which causes AD-like skin lesions, in an NC/Nga mouse model. External exposure to epicutaneous allergens such as DFE and DNCB can induce localized allergic inflammation in the skin and systemic sensitization to a specific allergen. 27 The symptoms of skin lesions were categorized based on edema, hemorrhage, erosion, dryness, and were significantly alleviated in the AD mouse model through administration of KBL382. Lichenification, which is indicated by ear and dermal thickness, was reduced in KBL382-treated mice compared to DFE +PBS and DNCB+PBS mice ( Figure 1D, E, 3A, and S1D-E). 31 KBL382-treated mice had a significantly lower TEWL and higher corneometer units than DFE+PBS mice, indicating fewer defects in the intercellular lipid bilayers of the stratum corneum ( Figure  2A and B). 14 Furthermore, scratching behavior increased during the progression of AD, suggesting increases in IgE and IL-31 levels and the number of mast cells. 32 Scratching behavior frequency was significantly lower in KBL382-treated mice than in DFE +PBS mice due to the reduced inflammation response ( Figure 2C). KBL382 also decreased the production of IgE, and thus controlled clinical symptoms related to AD ( Figure 2E and S2B). A decrease in serum concentrations of IgE suppresses the activation of mast cells and eosinophils, which express high-affinity receptors for IgE on their surface to induce degranulation of inflammatory mediators such as histamine and tryptase. 23 KBL382 significantly suppressed the infiltration of both eosinophils and mast cells into affected skin lesions ( Figure 3B and C). Exposure to allergens on atopic skin increases the expression of TSLP in epithelial cells, particularly keratinocytes. TSLP is a potent activator of DCs and induces production of the Th2-attracting chemokines TARC and MDC. TSLP-activated DCs prime naïve Th cells to produce the proallergic cytokines IL-4, IL-5, and IL-13, while downregulating IL-10. 33 The expression levels of TSLP, MDC, and TARC were elevated in DFE +PBS mice but reduced in KBL382-treated mice, which may suppress differentiation of T cells into Th2 cells and the secretion of Th2-type cytokines ( Figure 4).
As Th1, Th2, and Th17 responses are involved in the pathogenesis of AD, 8,10,14 oral administration of KBL382 modulated the adaptive immune response, leading to suppression of Th1 (IL-2 and IFN-γ), Th2 (IL-4, IL-5, IL-13, and IL-31), and Th17 (IL-17A) cytokines (Figure 5A-E, G-H, and S2C-E). Ingested probiotics are recognized by pattern recognition receptors, such as Toll-like receptors (TLRs) on DCs, and are essential to immunological homeostasis in the gut. DCs can directly present antigens from probiotics to the MLN and can interact with T and B cells to maintain noninflammatory immune responses. During this process, DCs play an important role in converting CD4+ Foxp3-T cells into CD4+ Foxp3 + T cells in MLNs. 34,35 CD4+ Foxp3+ Treg cells produce IL-10 and TGF-β, which suppress T cell polarization into Th2 cells, leading to the production of IL-4, IL-5, and IL-13 and inhibition of the activation of T cells, monocytes, and DCs. 23,34 Some probiotics increase the abundance of Treg cells, which mediate immune suppression and attenuate skin symptoms of AD. 26 Our results also showed that KBL382 significantly promoted the Treg cells population compared to DFE+PBS mice, downregulating Th2 cells and inhibiting the progression of inflammation ( Figure 5J, 6 and S2I). Gut microbiota interact dynamically with other microorganisms and hosts and it has become increasingly apparent that they are crucial to host health. 36 Interestingly, although alpha diversity was similar among the four groups tested, the beta diversity of gut microbiota differed distinctly ( Figure 7B and C). The proportions of bacteria at the genus level differed depending on DFE treatment and administration of the two strains. In accordance with previous study gut microbiota was altered in the progression of colitis, with a reduced relative abundance of S24-7_unclassified and increased Bacteroides and Mucispirillum. 37 On the other hand, we observed an increased abundance of S24-7_unclassified with administration of KBL365 or KBL382, and decreased abundance of Bacteroides and Mucispirillum with administration of KBL365 ( Figure 7A, D, and E). The proportion of Akkermansia was significantly increased with administration of KBL382 in AD-induced mice ( Figure 7D and E). Previous study has shown that the relative abundance of Akkermansia is significantly reduced in IBD patients. 38 Akkermansia increases anti-inflammatory activity by inducing Treg cells in adipose tissue, leading to attenuation of high-fat-diet-induced metabolic syndrome. 39 In addition, Akkermansia stimulates IL-10 production, playing an immunological role in homeostasis and taking on a barrier function in the gut mucosa. 40 Therefore, the further longitudinal study to elucidate the effects of KBL382 treatment in the abundances of beneficial microorganisms including Akkermansia should be performed. The compositional changes in specific microorganisms in the gut and their functional roles in the phenotypic changes observed in AD should be evaluated.
To investigate the functional roles of the administered Lactobacillus strains, we used function prediction to characterize gut microbiota using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database and measured metabolites in cecum. Gut microbiota is important to produce nonvolatile acids such as lactate, succinate, and fumarate which serve as pivotal intermediates that are degraded into SCFAs. 41 KBL382-treated mice had reduced amino acid levels but elevated levels of nonvolatile acids and propionate compared to DFE+PBS mice. These findings were consistent with KEGG analyses showing that administration of KBL382 resulted in increased levels of genes associated with metabolism. Moreover, KBL382 administration can affect the amino acids and SCFAs production ( Figure 8A-C). Further study to clarify the microbiota-derived SCFAs related to KBL382 administration should be performed.
Previous studies suggested that L. rhamnosus showed the therapeutic effects for AD. [42][43][44] Especially, L. rhamnosus GG is widely reported strain for AD prevention and treatment. 43,44 Therefore, in this study, we used L. rhamnosus KBL365, isolated from healthy Korean feces, to confirm the effects of L. paracasei on AD. Although some L. rhamnosus strains suppress the progression of symptoms 43 and show therapeutic benefits in infants with AD, 44 KBL365 administration in our study did not suppress AD symptom progression due to differences in the adaptive immune response and biological functions of these strains, even within the same species. This is not surprising because strain specificity of particular functions is commonly observed in bacteria. Compared to more complex organisms, the same species of microorganisms can have a lower percentage of common genes, which can cause different functional capabilities of different strains within a genus or even species. 45 The genes and functions of KBL365 and KBL382 should be further identified and compared, along with those of other strains, in the future.
In conclusion, administration of KBL382 alleviates AD by reducing the severity of clinical symptoms by increasing the immunosuppressive response and changing the metabolic functions of gut microbiota. Our results suggest that KBL382 is a promising candidate for therapeutic treatment to reduce AD symptoms.

Isolation and preparation of L. rhamnosus KBL365 and L. paracasei KBL382
L. rhamnosus KBL365 and L. paracasei KBL382 were isolated from the feces of healthy Koreans, as described in a previous study. 46 Briefly, KBL365 and KBL382 strains were cultured in Lactobacilli MRS Agar (BD Difco, Sparks, MD, USA) supplemented with 0.05% L-cysteine-hydrochloride at 37° C under anaerobic conditions maintained with an Anaeropack (Mitsubishi Gas Chemical Company, Inc., Tokyo, Japan) for 24 h. The cells were harvested through centrifugation (1,200 × g) and washed twice with phosphate-buffered saline (PBS) prior to administration to mice. Based on decreases in levels of Th1, Th2, and Th17-type cytokines and an increase in the antiinflammatory cytokine IL-10 in peripheral blood mononuclear cells in vitro, 46 we selected KBL382 for our AD-like experiment.

Experimental atopic dermatitis
Male 5-week-old NC/Nga mice were purchased from Central Lab Animals Incorporated (Seoul, Korea). All animal experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC: SNU-160928-1-1 and SNU-190919-3-1) of Seoul National University, Korea. The DFE and DNCB (Sigma-Aldrich Corp, St. Louis, MO, USA)-induced in vivo mouse AD model was performed as previously described with minor modifications. 5,31 For DFE-induced model experiment, four groups of mice (N = 7-9) were treated with (1) PBS (Control), (2) DFE+PBS, (3) DFE+KBL365, and (4) DFE+KBL382. After removing hair from the ears and dorsal region using electric clippers and hair removal cream, all groups except Control were applied with 150 µL 4% sodium dodecyl sulfate for 3 h to disrupt the skin barrier. Then, they were treated with 100 mg Biostir AD (Biostir, Hiroshima, Japan), an ointment containing DFE extract, once or twice per week for 7 weeks to induce AD. After 3 weeks from the beginning of the experiment, the DFE +KBL365 and DFE+KBL382 groups were administered suspensions of KBL365 and KBL382 (1 × 10 9 CFU each) in 200 μL PBS and Control and DFE +PBS mice were administered PBS daily via oral gavage for 4 weeks ( Figure 1A). For DNCB-induced model experiment, three groups of mice (N = 6-7) were treated with (1) PBS (Control), (2) DNCB +PBS, and (3) DNCB+KBL382. After removing hair from the ears and dorsal region using electric clippers and hair removal cream, all groups except Control were applied with 200 µL of 1% DNCB dissolved in a vehicle (acetone:olive oil = 3:1) and 15 µL to induce skin legion on d 1 and 4. One week later, 0.4% DNCB was treated to challenge the dorsal skin (200 µL) and ears (15 µL) three times a week for 2 weeks. After 1 week from the beginning of the experiment, the DNCB+KBL382 group was administered suspensions of KBL382 (1 × 10 9 CFU each) prepared in the same as the DFEinduced AD model experiment daily via oral gavage for 2 weeks ( Figure S1A). Then, the mice were sacrificed and further analyzed.

Assessment of skin lesions and clinical dermatitis severity score
NC/Nga mice were anesthetized with isoflurane, and pictures of the skin were taken weekly until the end of the experiment. The severity of dermatitis was evaluated once a week by scoring the symptoms visible on the ear, neck, and dorsal skin. The severity of dermatitis was assessed based on the four aspects of erythema/hemorrhage, scarring/dryness, edema, and excoriation/erosion on a scale of 0 (none), 1 (mild), 2 (moderate), and 3 (severe). A total dermatitis score was evaluated as the sum of the scores for each aspect, for a maximum score of 12. 31

Measurement of TEWL and stratum corneum hydration
On the last day of the experiment, TEWL and skin hydration were determined with a Tewameter TM300 and Corneometer CM 825 (Courage and Khazaka, Cologne, Germany), respectively, applied to the dorsal skin under controlled conditions of 21-22°C and 50-55% humidity. 47 The probe was placed on the skin and stabilized for approximately 30 s, and then the measurement was recorded. This measurement procedure was repeated five times in the same area at the same pressure.

Scratching behavior score
Scratching behavior was observed, as described previously. 48 Briefly, the mice were placed individually in acrylic cases composed of three equal cells for at least 1 h for acclimation. Scratching behavior was evaluated by measuring the number of scratches and time that mice spent scratching their nose, ears, and dorsal skin for 10 min on the day before sacrifice. As mice make several rapid scratching movements over a period of about 1 s, a series of such movements was counted as one bout of scratching. 49

Measurement of total IgE
Blood was obtained from the eye blood vessel of each mouse. Serum was separated through centrifugation at 1,200 × g for 15 min at 4°C. Total serum concentrations of IgE were measured using an IgE ELISA kit (Komabiotech, Seoul, Korea) according to the manufacturer's instructions.

Measurement of cytokines and chemokines
To measure cytokine production at the protein level, frozen skin tissues were weighed and homogenized with radioimmunoprecipitation assay buffer and protease inhibitors (Thermo Fisher Scientific, Inc., Waltham, MA, USA). The homogenates were centrifuged at 12,000 × g for 10 min at 4°C, and their supernatants were collected. Levels of IFN-γ, IL-4, IL-5, IL-13, IL-10, and IL-17 were measured in the skin lysates using a customized multiplex magnetic Luminex kit (R&D Systems, Minneapolis, MN, USA). Expression of Foxp3, IL-2, IL-4, IL-5, IL-6, IL-10, IL-31, MDC, TARC, TGFβ, TNF-α, and TSLP in mouse skin tissue at mRNA level was measured using real-time PCR. RNA was isolated from skin using an Easy-spin Total RNA Extraction Kit (Intron, Seoul, Korea) and cDNA was synthesized using the High-Capacity RNA to cDNA Kit (Thermo Fisher Scientific). Real-time PCR was performed in a Rotor-Gene Q (Qiagen, Hilden, Germany) thermal cycler using the QuantiTect SYBR Green PCR kit (Qiagen) with the primers (0.01 mM) listed in Table S1. Each PCR consisted of denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 5 s and 60°C for 10 s. The relative expression level of the target gene was calculated using the 2 -ΔΔCT method, with normalization to the expression level of the reference gene hypoxanthine-guanine phosphoribosyl transferase. 50

Histological examination
Dorsal skin tissues were fixed in 10% formaldehyde. This tissue was stained with hematoxylin and eosin to evaluate neutrophil infiltration and with Congo red and toluidine blue to detect eosinophil and mast cell infiltration, respectively. The numbers of eosinophils and mast cells were counted using Pannoramic Viewer software (3DHISTECH, Ltd., Budapest, Hungary).

Intestinal microbiota analyses
Intestinal microbiota was observed as described previously. 46 Briefly, the V3-V4 region of the 16S rRNA gene in total genomic DNA collected from cecum was amplified using the barcoded primers 341F and 805R. PCR amplicons were purified using the QIAquick PCR Purification Kit (Qiagen) and sequenced on the MiSeq platform (Illumina, Inc., San Diego, CA, USA). 51,52 Sequence data were processed using the Quantitative Insights into Microbial Ecology 1.8.0 (QIIME) pipeline. The sequences were clustered into operational taxonomic units (OTUs) at the 97% identity level using the Greengenes database (ver. 13_5). Alpha diversity indexes such as chao1 and PD were estimated to identify significant differences among groups. Phylogenetic beta diversity was measured using the UniFrac distance between samples and visualized based on weighted PCoAs. Linear discriminant analyses (LDAs) of effect size (LEfSe) were performed (http://huttenhower.org/galaxy) to identify significantly different phylotypes within the experimental groups. In addition, phylogenetic investigation of the communities was performed by reconstructing unobserved data to identify functional genes in the sampled microbial community based on data in the KEGG pathway database (GenomeNet; https://www.genome.jp/kegg/path way.html), as described previously. 53

Measurement of nonvolatile acids and short-chain fatty acid
Cecum was homogenized with distilled water and centrifuged at 13,000 × g for 5 min. Then, the supernatant was extracted and mixed with an internal standard (benzoic acid for nonvolatile acids and 1% 2-methylpentanoic acid for volatile acids). The supernatant was added to the extraction solvent (chloroform for nonvolatile acids and ethyl ether for volatile acids) and centrifuged at 13,000 × g for 5 min. The organic layer was transferred to a vial (Agilent Technologies, Santa Clara, CA, USA) and measured through gas chromatography. Nitrogen was used as the carrier gas. The oven temperature was set to 170°C. The flame ionization detector and injection port were set to 225°C. The retention times and peak areas of the standard mix were used as references for unknown samples. 54

Amino acid analyses using ultra-performance liquid chromatography (UPLC)
Amino acids were analyzed as described previously. 46 Cecum extracts were transferred to an autosampler vial and derivatized. After incubation for 10 min at 55°C, the derivatized samples were analyzed using Acquity UPLC (Waters Corporation) with a SYNAPT G2-Si mass spectrometer (Waters Corporation). The mass acquisition mode Tof-MRM and positive electrospray ionization mode were used. Capillary voltage was set to 1.5 kV. The sampling cone voltage varied from 20 V to 60 V among amino acids. The desolvation gas and cone gas flow rates were set to 600 L/h and 50 L/h, respectively. The desolvation temperature was 250°C. Amino acids were confirmed through alignment to the analytical standard mixture and internal standard. Data acquisition and quantification were carried out using MassLynx software 4.1 (Waters Corporation). 55

Statistical analyses
Data are expressed as mean ± standard error of the mean (SEM). GraphPad Prism 5.04 (GraphPad Software, Inc., La Jolla, CA, USA) was used to visualize and analyze data using the Mann-Whitney U test. When appropriate, false discovery rate (FDR)correction was not applied in data presentation. Statistical significance is denoted as * P< .05, ** P< .01, *** P< .001.

Disclosure of potential conflicts of interest
GK is the founder of KoBioLabs, Inc. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding
This work was supported by National Research Foundation of Korea (NRF) grants funded by the Ministry of Science and ICT (MSIT) (NRF-2015M3C9A4053391, NRF-2018R1A2A1A05078258, and NRF-2019M3C9A6089879).