Limosilactobacillus reuteri Attenuates Atopic Dermatitis via Changes in Gut Bacteria and Indole Derivatives from Tryptophan Metabolism

Atopic dermatitis (AD) is a common inflammatory disease in the skin and affects many infants, children, and even adults worldwide [1]. Susceptibility to AD is closely associated with genetic and environmental factors that increase the dysfunction of the epidermal barrier and/or dysregulation of the immune response, and mutations in the filaggrin gene are the strongest genetic predisposing factors to induce dysfunction of the epidermal barrier in AD [2,3]. Furthermore, there are massive amounts of bacteria in the skin and gut, and they affect and regulate the immune and healthy states of the host. Skin microbiome such as Staphylococcus aureus disturbs the epidermal barrier and induces decreases in antimicrobial peptides, an increase in T helper 2 (Th2) type cytokines, and disturbance of skin lipid metabolism [4,5]. In addition, the intestine is another important bacterial habitat, and gut microbial alteration is closely associated with allergic diseases including asthma, food allergy, and AD [6,7]. Gut microbial dysbiosis is an important cause of the onset of AD and this has been demonstrated in clinical studies [8,9]. Clostridium difficile, Clostridia, S. aureus, Escherichia coli, and obligate anaerobe species were prevalent but Lactobacillus, Bifidobacterium, Akkermansia muciniphila, and Ruminococcus gnavus were reduced in the feces of patients with AD [10,11]. It has been reported that the proportion of E. coli was positively correlated to serum immunoglobulin (Ig) E levels [12], but Bacteroides fragilis decreased IL-4 levels produced by CD4+ T cells in germ-free mice [13]. In a cohort of 24 infants, compared to healthy controls, the proportion of bacilli was significantly higher in AD infants. Furthermore, Clostridia, but not bacilli and E. coli, is significantly related to age at AD onset and negatively correlated with the proportion of eosinophils in the blood [14]. However, compared with patients with AD (aged 6–22 years old), the relative abundance of Clostridium was higher, but Blautia and Parabacteroides were lower in healthy controls [15]. There are significantly different gut microbial taxa in infants with AD compared to that in child/adult patients. Therefore, these specific changes influence gut microbial metabolisms and thus affect the balance between T helper 1 (Th1)- and Th2-type as well as Th17 and T regulatory (Treg) immune responses in AD [16].

Tryptophan metabolism involves the development of many diseases. Changes in tryptophan metabolites significantly affect the clinical characteristics of patients with food allergies [17], depression [18], and cancer [19]. Fructooligosaccharides significantly increased kynurenine levels and restored Th17/Treg balance to alleviate the clinical manifestations in mice with ovalbumin-induced food allergies [20]. Fuzhuan brick tea polysaccharide increased indoleacetic acid (IAA) and indole-3-aldehyde (IAld) levels and thus activated the aryl hydrocarbon receptor (AHR) signaling pathway to improve ulcerative colitis via producing interleukin-22 (IL-22) and evaluating the expression of intestinal tight junction proteins [21]. Gut microbiota converts tryptophan into indole and indole derivatives, which exert various physiological features including anti-allergy [22,23,24,25]. In germ-free mice with AD, reduced AHR signaling expression induced an increase in epidermal barrier dysfunction, which was restored using AHR agonist treatment [26]. Indole derivatives such as IAld, as an endogenous ligand for AHR, have been demonstrated to activate AHR signaling to alleviate psoriasis and AD [27,28]. Tryptophan metabolism is reduced in patients with AD, and treatment with IAld-activating AHR to inhibit aberrant Th2-type response improves AD-like symptoms [29,30]. Therefore, targeting activating AHR is an alternative way to ameliorate AD-like clinical symptoms. Limosilactobacillus reuteri, a tryptophan-metabolizing bacterium, produced indole derivatives and protected against Candida albicans colonization and intestinal mucosal inflammation via activating the AHR-IL-22 axis [31]. However, increased IL-22 is positively related to the development of AD via modulating the gene expression of the skin barrier-related molecules including filaggrin, loricrin, and involucrin [32], and its main producers are Th22 cells that do not produce IL-17A [33]. IL-22 production from Th22 and innate lymphoid cells (ILC) such as ILC22 is dependent on AHR [34,35]. AHR-IL-22 axis signaling pathway may not be the main mechanism to alleviate AD and thus the mechanism of action needs to be elucidated. We assumed that L. reuteri strains could alleviate the clinical symptoms via AHR activation. Furthermore, the effects of gut microbiota should not be excluded from the process. Therefore, combined with these alterations, this study aimed to explore: (1) the effects of L. reuteri strains on AD symptoms and (2) the mechanism of L. reuteri strains in alleviating AD via the gut microbiota.

This study explored the effects of L. reuteri strains with the role of regulating gut microbial tryptophan metabolism on AD-like symptoms and the underlying mechanism in mice. In AD-like mice, a decrease in the integrity of the skin barrier and inflammatory infiltration due to aberrant immune responses. Therefore, AD was characterized by skin inflammation and excessive Th2-type responses in the acute stage [36]. L. reuteri strain has the potential to alleviate AD symptoms via modulating immune responses. L. reuteri Fn041, a strain secreting IgA, significantly improved skin swelling and inflammatory cell infiltration via restoring the systemic Th1 and Th2 cytokine ratios and Treg proliferation [37]. Likewise, L. reuteri DYNDL22M62 variably affected AD-like symptoms, which could be shown based on H&E staining and changes in cytokines. Treatment with L. reuteri DYNDL22M62 significantly reduced ear thickness, skin swelling, and inflammatory infiltration in skin lesions suggesting that L. reuteri DYNDL22M62 might have the function of regulating immunity (Figure 1B,C). L. reuteri strains significantly suppressed IgE levels except for GDLZ105, and L. reuteri DYNDL22M62 reduced Th2 cytokines including TSLP, IL-4, and IL-5 in AD-like mice (Figure 2). IgE-mediated hypersensitivity was one of the mechanisms to induce allergic diseases including AD [38]. IL-4 binds to IL-4Rα, which is expressed in T cells, B cells, and macrophages leading to excessive Th2 cell differentiation and IgE class switching [39]. DNFB-induced an increase in IL-4 led to a higher level of IgE and Th2 responses (Figure 2). Treatments with L. reuteri strains except for GDLZ105 reduced skin swelling and inflammation (Figure 1C), and this was consistent with the alteration of Th2 cytokines. TSLP, expressed in epithelial cells, is closely associated with the initiation and maintenance of AD and epidermal barrier integrity [40,41]. L. reuteri DYNDL22M62-induced a reduction of TSLP was conducive to improving skin lesions in AD-like mice. Collectively, these results indicated that L. reuteri DYNDL22M62 alleviated AD-like symptoms via restoring aberrant immune responses.

Indole derivatives are the main metabolites of tryptophan in the gut due to the catabolism of anaerobic bacteria and are associated with the maturation and function of the immune system. L. reuteri strains WU and 100-23 induced intraepithelial T lymphocytes via the release of AHR ligands including ILA and reprogrammed intraepithelial CD4 T cells into immunoregulatory T cells, suggesting the distinct AHR-mediated immunoregulatory mechanism of indole derivatives [42]. Treatments with L. reuteri DYNDL22M62 increased the tryptophan metabolism of gut microbiota and significantly induced the release of ILA and IPA (Figure 3A–E). Furthermore, L reuteri DYNDL22M62 increased the 1.35-fold-change expression of AHR compared to the DNFB group (Figure 3F). The AHR pathway plays a critical role in the treatment of AD. Treatment with coal tar induced epidermal cell differentiation and increased filaggrin expression and skin barrier proteins via AHR activation in primary keratinocytes from AD patients [28]. An increase in IAld, a ligand of AHR, induced by B. longum CCFM1029 or skin bacteria, reduced TSLP levels and further reduced Th2 cytokine levels in AD-like mice [30,43]. Furthermore, in some cohort studies, ILA was the predominant metabolite from B. infantis grown in human milk and regulated immune responses of human CD4+ T cells and monocytes in a dose-dependent manner by activating AHR and hydroxycarboxylic acid receptor 3 [44,45]. This suggests that ILA from microbial metabolism may affect immune function via activating the receptor signaling pathway. IPA has direct anti-inflammatory regulation effects on the immune cells via increasing IL-10 production or decreasing pro-inflammatory tumor necrosis factor expression [46]. As one of the ligands for AHR, type I interferons were produced in the central nervous system in combination with tryptophan metabolites such as IPA and IAld to activate AHR signaling in astrocytes and suppressed central nervous inflammation [47]. Although there is no direct evidence to link ILA and IPA to the development of AD, their immunomodulating effects may be the important action mechanism to alleviate AD-like symptoms. Collectively, these results suggested that L. reuteri DYNDL22M62 might activate AHR signaling via the increase in the expression of AHR ligands, such as ILA and IPA, to suppress aberrant Th2-type immune responses.

Changes in tryptophan metabolism in the intestine were driven by gut microbiota, and gut microbial alterations contributed to the development of AD. In a cohort study, Bacteroidetes, Xanthomonadaceae, and Bacteroidaceae at the family level, and Stenotrophomonas and Bacteroides at the genus level were more abundant in infants (3–4 months old) with AD versus healthy controls [6]. However, children with eczema showed a higher Simpson’s reciprocal diversity index, a reduction of Bacteroidetes, and more abundant Clostridium clusters Ⅳ and ⅩⅣa in gut microbial alterations compared to at this age healthy children [48]. Among the mice in Dermatophagoides farinae extract-induced AD group, the proportion of S24-7_unclassified decreased but Bacteroides increased versus the control group [49]. In this study, the relative abundances of bacteria belonging to Firmicutes including Lactobacillaceae, Erysipelotrichaceae, and Peptostreptococcaceae were reduced in mice with AD but they were restored by L. reuteri DYNDL22M62 treatment (Figure 4C). Compared to the DNFB group, L. reuteri DYNDL22M62 significantly increased the proportions of the Romboutsia and Ruminococcaceae NK4A214 group, which were positively related to IPA and ILA (Figure 5B). However, L. reuteri DYNDL22M62 decreased the proportion of Dubosiella, which was negatively related to IPA but positively related to ILA (Figure 5C). This suggested that there might be a competitive relationship between these bacteria related to tryptophan metabolism. In some cohort studies, compared to the healthy controls, the proportions of Romboutsia and Ruminococcaceae were decreased in patients with allergic rhinitis and food allergy, respectively [50,51]. These results showed that L. reuteri DYNDL22M62 treatment regulated tryptophan metabolism-related gut microbiota that contributed to the AHR activation and further suppressed the excessive Th2-type response in mice.

There are some issues to be solved in the future. For instance, to evaluate ILA and IPA whether produced by L. reuteri DYNDL22M62 and/or gut microbiota, the role of gut microbiota will be evaluated using fecal microbiota transplantation at first and the role of L. reuteri DYNDL22M62 will be assessed using a germ-free mouse or antibiotic-treated mouse model. Furthermore, it is to be elucidated whether ILA/IPA alleviated AD dependent on AHR signaling although we have analyzed the expression of AHR in the skin lesion. In the next experiment, the effects of L. reuteri DYNDL22M62 will be evaluated using AHR antagonists such as CH223191↗ in wild-type mice or AHR-knockdown mice, and signaling molecules related to the AHR pathway need to be characterized based on different molecular biological methods. Finally, the mechanism of L. reuteri DYNDL22M62 to alleviate AD will be demonstrated comprehensively via gut microbial alterations and changes in immune responses.

In conclusion, L. reuteri DYNDL22M62 treatment regulated gut microbial composition and elevated tryptophan metabolism in the intestine to produce the AHR ligands including ILA and IPA, which activated AHR signaling to reduce aberrant Th2-type response in mice. Our findings suggest that treatment with L. reuteri DYNDL22M62 with the role of regulating gut bacteria and their tryptophan metabolism may be an effective alternative to alleviate AD.