Graphical abstract graphic file with name fx1.jpg [39]Open in a new tab Highlights * • ILA-producing capacity distinguishes bifidobacteria with antidepressant effects * • Aldh is essential for bifidobacteria’s ILA synthesis and antidepressant efficacy * • ILA’s antidepressant effects linked to AhR-mediated neuroinflammation relief __________________________________________________________________ Qian et al. reveal that psychobiotic Bifidobacterium breve reverses the abnormal reduction of hippocampal indole-3-lactic acid (ILA) in depressed mice. The study highlights the role of Aldh gene in ILA production by bifidobacteria and its antidepressant effects through AhR signaling, advancing the understanding of psychobiotics in mood disorder therapies. Introduction Over the past few decades, the prevalence of depression has steadily risen, surpassing cardiovascular disease and cancer, making it the foremost global cause of disability.[40]^1^,[41]^2 Social and psychological stressors play pivotal roles in contributing to depression, a significance accentuated during the COVID-19 pandemic, which have resulted in an upsurge in depression rates, attributed to stress from unemployment, prolonged sedentary time, and social isolation.[42]^3^,[43]^4 Post-COVID-19 depression is also considered to be connected to peripheral immune inflammation triggered by viral infection, gliosis, and neuroinflammation.[44]^5 However, due to the complexity and heterogeneity of depression’s pathogenesis,[45]^6 conventional antidepressants and psychotherapy currently used in clinical practice demonstrate only limited efficacy. Evidence suggests that over one-third of patients show an insufficient response to first-line treatments.[46]^7 Classical antidepressants, rooted in the “monoamine deficiency hypothesis,” inadequately account for delayed efficacy and varied treatment responses.[47]^7^,[48]^8 The deduction from the serendipitous discovery of monoamine oxidase inhibitors’ efficacy to the postulation that serotonin (5-hydroxytryptamine, 5-HT) deficiency causes depressive symptoms remains highly contentious.[49]^9^,[50]^10^,[51]^11 Research has established that altering brain 5-HT levels, either by reducing them in healthy volunteers or increasing them in depressed patients, is insufficient to induce or alleviate clinical depressive symptoms.[52]^10^,[53]^12 Therefore, there is a pressing need to develop new antidepressant strategies that address the multifaceted nature of depression’s pathogenesis. In recent years, attention has shifted toward the bidirectional signaling between the gut microbiota and central nervous systems, known as the microbiota-gut-brain axis.[54]^13 Some studies not only associate changes in the gut microbiota with depression but also establish causal relationships, indicating a direct impact of microbiota alterations on depressive conditions.[55]^14^,[56]^15 For instance, two individuals with severe/moderate depression experienced a notable reduction in symptom severity, transitioning to a state of mild depression following a 4-week regimen of fecal microbiota transplantation (FMT) as an adjunctive therapeutic intervention; this sustained efficacy was observed in one patient for an extended period of up to 8 weeks.[57]^16 Similar effects were also observed in patients with irritable bowel syndrome, functional diarrhea, or functional constipation who underwent FMT.[58]^17 Despite the standardized management of donor screening for FMT and potential safety risks limiting the widespread application of FMT in depression, these findings underscore the intricate interplay between the gut microbiome composition and mental health outcomes.[59]^18 Alternatively, probiotics emerge as a safer and more widely applicable strategy.[60]^19^,[61]^20 Coined as “psychobiotics” in 2013, these live organisms, when ingested in adequate amounts, produce health benefits in patients with psychiatric illnesses.[62]^21 Over the past decade, psychobiotics have gained academic recognition for their proven antidepressant efficacy in both animal and clinical trials.[63]^22^,[64]^23^,[65]^24^,[66]^25 This alternative is particularly valuable for patients intolerant or unresponsive to traditional medications.[67]^25 Psychobiotics can also complement conventional drug therapies, potentially enhancing overall efficacy.[68]^20^,[69]^26 Mechanisms underlying psychobiotics’ mood improvement involved increased brain-derived neurotrophic factor/serotonin levels, reduced microbiota-mediated inflammation, and vagus nerve activation.[70]^27^,[71]^28^,[72]^29 However, understanding the pharmacological mechanisms of psychobiotics remains challenging compared to traditional antidepressants, given the diverse strain-specific effects and limited knowledge about their common microbiological and molecular genetic features.[73]^30^,[74]^31 In this study, we observed a significant decrease in indole-3-lactic acid (ILA) levels in the gut and brain of chronically stressed depressive mice, while supplementation with the psychobiotic Bifidobacterium breve CCFM1025 (Bre1025) can restore these levels back to normal. To underline the critical role of ILA production for the antidepressant efficacy of Bre1025, we conducted experiments involving the insertion mutant of the aromatic lactate dehydrogenase (Aldh) gene responsible for ILA synthesis. Additionally, we investigated the psychobiotic potential of various bifidobacteria species; we confirmed that those harboring the Aldh gene exhibit psychobiotic potential. Furthermore, we suggest that antidepressant efficacy achieved by bifidobacteria-derived ILA may be attributed to the activation of the aryl hydrocarbon receptor (AhR) signaling pathway, leading to the alleviation of neuroinflammation. These findings offer insights into the therapeutic potential of psychobiotics in stress-related mood disorders. Results The antidepressant effects of psychobiotic Bifidobacterium breve linked to the regulation of gut ILA Initially, we noted significant strain-level variations in the emotional regulation capabilities of Bifidobacterium breve. Specifically, strain Bre1025 demonstrated the ability to alleviate anxiety-like and depressive-like behaviors induced by chronic unpredictable mild stress (CUMS), in contrast to the control strain, Bre3M5 ([75]Figures 1A–1D). Notably, despite comparable colonization levels in the mouse gut ([76]Figure 1E), Bre1025 intervention significantly altered the composition of gut microbiota metabolites ([77]Figure 1F). Compared with the CUMS group (p < 0.05), Bre1025 intervention increased the content of ILA, indole-3-acrylic acid, and genistein, while resulting in a decrease in the content of metabolites such as 5-hydroxyindole-3-acetic acid, ornithine, and acrylic acid ([78]Data S1). Based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic pathway enrichment analysis, we found that Bre1025, compared to the control strain, reversed the abnormal decrease in the metabolite ILA in the Tryptophan metabolism pathway induced by CUMS ([79]Figure 1G). This change was further validated through targeted metabolomic detection of indole derivatives ([80]Figure 1H). Meanwhile, we confirmed that Bre1025 exhibited significantly higher ILA production in vitro than the control strain Bre3M5 ([81]Figure S1A). Subsequent debiased sparse partial correlation analysis highlighted ILA as the pivotal gut metabolite, demonstrating a significant negative correlation with the immobility time in the forced swimming test (FST) and tail suspension test (TST) ([82]Figure 1I). Importantly, Bre1025 counteracted the abnormal decrease of ILA levels in the serum and hippocampus induced by CUMS ([83]Figures 1J and 1K). Bre1025 intervention also elevated the level of the downstream indole derivative indole-3-propionic acid (IPA) production ([84]Figure S1B), with no significant impact on other indole derivatives. Figure 1. [85]Figure 1 [86]Open in a new tab Indole-3-lactic acid associated with Bre1025-mediated antidepressant effect (A) Schematic representation of the animal experimental design. In this animal experiment, 20 mice were randomly divided into four groups (n = 5/treatment): control (blank control group), CUMS (CUMS model group), Bre1025 (B. breve Bre1025 intervention group, 1 × 10^∧9 CFU/d), and Bre3M5 (B. breve Bre3M5 intervention group, 1 × 10^∧9 CFU/d); OFT, open field test; TST, tail suspension test; FST, force swim test. (B–D) Behavioral outcomes of mice assessed in different tests. (E) Quantification of B. breve cell counts in mouse feces. (F) Principal component analysis illustrating the metabolome of mouse gut contents. (G) Analysis of differential metabolites in mouse gut. KEGG pathway enrichment analysis of gut metabolites are shown on the left (p < 0.05), and volcano plots of differential metabolites are shown on the right (p < 0.05 and fold change ≥1.5). 5-HIAA, 5-hydroxyindole acetic acid. (H) Targeted and quantitative detection of indole-3-lactic acid (ILA) in mouse gut contents. (I) Debiased sparse partial correlation analysis of gut metabolites and gut microbiota. (J and K) Targeted quantification of ILA in mouse serum and hippocampus. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, determined by one-way analysis of variance (ANOVA) followed by Sidak post hoc test in (B–E), (H), (J), and (K). The potential antidepressant effects of gut microbiota-derived ILA are also supported by clinical and population studies. Previous research has found that patients with major depressive disorder (MDD) have significantly lower serum levels of indole derivatives, including indole-3-acetamide, ILA, and indole-3-aldehyde (IAld), compared to healthy controls.[87]^32 Here, we further conducted a validation experiment of serum and fecal indole derivatives in 40 subjects (patients with depression: healthy individuals = 1:1). The results revealed that patients with depression had significantly lower serum levels of ILA, indole-3-acetic acid, and IAld compared to healthy individuals, while only fecal ILA levels were significantly reduced ([88]Figures S1C and S1D). Additionally, the results from the aforementioned animal experiments are consistent with previous data from our clinical intervention study with depressed patients using this strain, specifically showing that Bre1025 can increase ILA levels in the gut of patients with depression.[89]^20 Based on all the evidence presented, we hypothesize that ILA plays a crucial role in mediating the emotional regulation facilitated by Bre1025. Elevated gut and serum ILA levels were attributed to the supplementation of Bre1025 To determine whether the gut ILA content is contributed by the intake of Bre1025 rather than the indigenous microbiota, we designed strain-specific primers for Bre1025 to investigate its colonization level and assess its contribution to the gut ILA level. Healthy volunteers, upon continuous intake of Bre1025, exhibited a rapid accumulation in the gut ([90]Figures 2A and 2B), leading to an elevation in gut ILA, as well as downstream indole acrylic acid (IA) and IPA levels ([91]Figure 2C). In contrast, the levels of other indole derivatives from the gut microbiota significantly decreased ([92]Figure S2A). Notably, even after discontinuation of intake for 14 days, Bre1025 was still detectable in the feces of nearly half of the volunteers (7/15) ([93]Figure 2B), here termed as “colonizers.” Colonizers maintained stable ILA and downstream indole derivative levels after discontinuation, whereas in individuals where Bre1025 was washed out of the gut, ILA levels significantly decreased ([94]Figures 2C and [95]S2B). Furthermore, in individuals with the presence of Bre1025, their gut ILA and downstream IPA levels were positively correlated with the abundance of Bre1025 ([96]Figures 2D and 2E). We subsequently analyzed the host gut microbiota structure before and after Bre1025 intake ([97]Figures S1E–S1G). The results showed no significant changes in the β-diversity and α-diversity of the gut microbiota of healthy volunteers after 2 weeks of Bre1025 intake. However, Bre1025 intake promoted the abundance of the genera Lachnospira and Allobaculum ([98]Figure S1H), which have been reported to be involved in the conversion of ILA to the downstream indole derivative IPA.[99]^33^,[100]^34 This increase in these genera potentially explains the elevated IPA levels observed. In the mouse model, we further demonstrated that the ILA present in serum is primarily contributed by ILA produced from gut microbes. We established mice with varying doses of Bre1025 in the gut ([101]Figure 2F), revealing a significant microbial dose-dependent ILA/IPA level trend in both the gut and serum ([102]Figures 2G–2J). Especially noteworthy is the positive correlation between fecal and serum ILA/IPA levels observed in these mice ([103]Figures 2K–2L). Additionally, Bre1025 ingestion did not exert significant effects on other tryptophan metabolites in either the gut or blood of mice ([104]Figures S2C–S2J). Figure 2. [105]Figure 2 [106]Open in a new tab Ingestion of Bre1025 improved host gut and serum ILA levels (A) Schematic representation of the clinical trial experimental strategy. In this clinical trial, 15 healthy subjects were recruited for testing the colonization amount and ILA production capability of Bre1025. (B) Quantification of Bre1025 cell counts in human feces. (C) Content changes in indole derivatives in human feces. (D and E) Correlation between Bre1025 biomass and ILA/IPA content in human feces. (F) Quantification of Bre1025 cell counts in mouse feces. In this animal experiment, 24 mice were randomly divided into four groups for the quantitative analysis of intestinal and serum indole derivatives (n = 6/treatment): control (blank control group), 1025-low (low-dose Bre1025 intervention group, 1 × 10^∧7 CFU/d), 1025-med (medium-dose Bre1025 intervention group, 1 × 10^∧8 CFU/d), and 1025-high (high-dose Bre1025 intervention group, 1 × 10^∧9 CFU/d). (G and H) Quantification of ILA/IPA concentrations in mouse feces. (I and J) Quantification of ILA/IPA concentrations in mouse serum. (K) Correlation between fecal ILA content and serum ILA content in mouse feces. (L) Correlation between fecal IPA content and serum IPA content in mouse feces. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, determined by two-tailed paired Student’s t test in (B) and (C) and one-way ANOVA followed by Sidak post hoc test in (F–J). Aldh is the key gene for tryptophan-dependent ILA production in Bre1025 Microbial indole and its derivatives are directly or stepwise biotransformed from tryptophan (Trp).[107]^35 We assessed the ILA production ability of Bre1025 in a Bifidobacterium-defined medium (BDM, a fully defined medium) supplemented with varying concentrations of Trp. The substrate quantity did not affect the biomass accumulation of Bre1025 ([108]Figure S3A), but it did impact ILA production, with ILA levels increasing with higher Trp concentrations ([109]Figure 3A). Figure 3. [110]Figure 3 [111]Open in a new tab Bre1025 metabolizes tryptophan to ILA through the Aldh gene (A) ILA production by Bre1025 using tryptophan (Trp) as a substrate in a Bifidobacterium-defined medium (BDM) (n = 3/treatment). (B) Expression of genes related to ILA production in Bre1025 with different amounts of Trp substrate. The upper pathway diagram depicts a schematic representation of how microorganisms metabolize Trp into indole derivatives. The bottom line chart displays the expression of genes in Bre1025 that are potentially involved in the microbial metabolism of Trp into ILA. IAM, indole-3-acetamide; IAA, indole-3-acetic acid; IAld, indole-3-aldehyde; TA, tryptamine; IAAld, indole-3-acetaldehyde; IPYA, indole-3-pyruvate; IA, indole acrylic acid; IPA, indole-3-propionic acid. (C) In vitro growth curve of the Bre1025 mutant strain (n = 3/treatment). Wild-type: Bre1025 wild-type strain; Δtarget-gene: mutant strain of the Bre1025 target gene; ND, not detected. (D) In vitro ILA production of the Bre1025 mutant strain. (E) Principal component analysis illustrating the metabolome of the Bre1025 wild-type and mutant strains (n = 5/treatment). Control: mMRS medium supernatant. (F) Metabolite heatmaps of Bre1025 wild-type and mutant strains (n = 3/treatment). (G–L) Metabolism of Bre1025 in different substrates (n = 3/treatment). (M) Proposed Trp-indole metabolism pathway of Bre1025. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, determined by two-tailed Student’s t test in (E–J) and one-way ANOVA followed by Sidak post hoc test in (A) and (D). Next, we examined whether genes previously reported in microorganisms to be involved in the conversion of Trp to ILA are present in the genome of Bre1025, including aromatic amino acid aminotransferase (Arat),[112]^36 phenyllactate dehydrogenase (Fldh),[113]^37 and Aldh[114]^38 ([115]Figure 3B). However, the expression patterns of these three genes in response to changes in Trp substrate concentration were nearly identical ([116]Figure 3B), making it challenging to distinguish their contributions to Bre1025’s ILA production. Therefore, we individually constructed insertion mutants for these three genes ([117]Figure S3B). While the growth of the mutants in vitro was unaffected ([118]Figure 3C), there was a significant reduction in the extracellular ILA production of Bre1025 to varying degrees ([119]Figure 3D). Notably, only the Aldh mutant completely abolished Bre1025’s ILA production, regardless of whether it was in mMRS medium, BDM medium, or physiological saline supplemented with the substrate Trp ([120]Figure 3D). Following that, we analyzed the metabolic composition of the Bre1025 wild-type strain and the Aldh gene mutant strain in the classic mMRS medium. The results showed no differences in metabolites between the two strains, except for ILA ([121]Figures 3E and 3F). Subsequently, we conducted an analysis of the Trp metabolic pathway based on both the wild-type and Aldh mutant strains. When Trp served as the substrate, the mutant failed to produce ILA, and neither the wild type nor the mutant produced indole-3-pyruvate (IPYA) ([122]Figure 3G). This suggests that Bre1025 may either lack the enzyme necessary for converting Trp to IPYA, or the enzyme may be inactive under nutrient-deficient conditions. Additionally, when IPYA was used as the substrate, both the wild type and the mutant produced Trp, but with a significant difference in production ratios ([123]Figure 3H). This indicates that there is an enzyme-a converting IPYA to Trp, and that Aldh is responsible for converting IPYA to ILA. Meanwhile, neither strain demonstrated the ability to reverse ILA into IPYA or Trp, nor to further convert ILA into downstream indole metabolites ([124]Figure 3I). Furthermore, Bre1025 did not participate in another Trp-indole pathway widely reported in the intestinal microbiota ([125]Figures 3J–3M). Based on these results, we speculate that Bre1025 has two pathways for converting Trp to ILA: one where Arat converts Trp to IPYA, followed by Aldh converting IPYA to ILA,[126]^35 and another potential pathway bypassing Arat (or another enzyme-b substituting Arat), where Aldh directly converts Trp to ILA, and lacks the ability to convert Trp and ILA into other indole metabolites ([127]Figure 3M). ILA production capacity is essential in enabling the antidepressant effect of Bre1025 To determine the role of ILA production in Bre1025’s emotional regulation capacity, we investigated potential variations in ILA metabolism and behavioral outcomes in CUMS mice subjected to treatment with the wild-type and the Aldh mutant strain ([128]Figure 4A). Before intervention, mice were pre-treated with antibiotics to remove the effects of ILA produced by the native gut microbiota. The findings revealed that the absence of Aldh function did not influence the accumulation of Bre1025 in the mouse intestine ([129]Figure 4B). Nevertheless, in line with the oral intake of ILA (35 mg/kg⋅bw/day), the administration of Bre1025 counteracted the CUMS-induced reduction in serum and brain (hippocampus or prefrontal cortex) ILA levels ([130]Figures 4C, 4D, and [131]S4A). Notably, the loss of Aldh gene hindered the strain’s ability to regulate ILA content in the gut-blood-brain axis of mice. Figure 4. [132]Figure 4 [133]Open in a new tab The production of ILA determines the antidepressant capability of Bre1025 (A) Schematic representation of the animal experimental design. In this animal experiment, 60 mice were randomly divided into five groups to evaluate the mood-regulating function of wild-type and genetically modified Bre1025 strains (n = 12/treatment): control (blank control group), CUMS (CUMS model group), ILA (oral ILA intervention group, 35 mg/kg/d), wild-type (Bre1025 intervention group, 1 × 10^∧9 CFU/d), and ΔAldh (Bre1025-ΔAldh intervention group, 1 × 10^∧9 CFU/d). (B) Quantification of Bre1025 cell counts in mouse feces. (C–E) ILA concentrations in mouse serum, hippocampus, and colon content. (F) IPA concentration in mouse colon content. (G–I) Behavioral outcomes of mice assessed in different tests. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, determined by two-tailed Student’s t test in (B) and one-way ANOVA followed by Sidak post hoc test in (C–I). It is noteworthy that orally administered ILA did not augment colonic ILA levels ([134]Figure 4E); rather, it may be absorbed into the bloodstream before reaching the colon. Furthermore, both oral ILA and Bre1025 influenced the levels of IPA and other indole derivatives in the gut and serum ([135]Figures 4F and [136]S4B–S4J). Concurrently, both ILA and Bre1025 interventions significantly reduced immobility time in the FST and TST in mice, while increasing exploration time in the center zone of the open field test ([137]Figures 4G–4I). Conversely, the Aldh mutant strain exhibited no such effects. These results underscore the essential role of ILA production in enabling Bre1025 to modulate host ILA metabolism and enhance emotional well-being. ILA-mediated antidepressant effect extends across diverse Bifidobacterium species Considering that the mechanisms mentioned earlier were initially observed in only two bacterial strains, our subsequent aim was to explore their universality at the Bifidobacterium species level. Initially, we assessed the in vitro ILA production capabilities of nine different Bifidobacterium species, totaling 75 strains (using BDM). Among them, strains belonging to Bifidobacterium bifidum, B. breve, B. longum subsp. longum, and B. longum subsp. infantis exhibited ILA production, while strains from B. adolescentis, B. pseudocatenulatum, B. dentium, B. animalis subsp. lactis, and B. animalis subsp. animalis did not ([138]Figure 5A). There were no significant differences in Trp consumption among the various Bifidobacterium species ([139]Figure S5A). Following this, we scrutinized the genomic features related to Trp metabolism in these strains and identified that strains capable of ILA synthesis all harbor the Aldh gene. The presence of the Fldh gene may contribute to higher ILA synthesis in B. bifidum compared to other strains ([140]Figure 5B), but further substantiation is still required. Figure 5. [141]Figure 5 [142]Open in a new tab The Aldh gene distinguishes the antidepressant effects among different Bifidobacterium species (A) Quantification of in vitro ILA production in diverse Bifidobacterium species. (B) Identification of Trp-indole pathway genes in the genome of diverse Bifidobacterium species. (C–E) Quantification of ILA concentrations in mouse colon content, serum, and hippocampus. In this animal experiment, 48 mice were randomly divided into six groups (n = 8/treatment): control (blank control group), CUMS (CUMS model group), and four intervention groups with Aldh^+B. breve (1 × 10^∧9 CFU/d): group A (bif-1, bre-2, inf-2), group B (lon-1, bif-2, bre-3), group C (ado-1, pse-2, lac-2), and group D (lac-1, den-2, ani-3). (F–H) Behavioral outcomes of mice assessed in different tests. (I) Principal component analysis of mouse gut microbiota. (J) Debiased sparse partial correlation analysis of indole derivatives and gut microbiota. (K) Abundance of gut bacteria related to the transformation of indole derivatives. Different letters in (A), ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 in (C–H) and (K) represent significant differences between groups, as determined by one-way ANOVA followed by Sidak post hoc test. Subsequently, we selected two groups of high-ILA-producing strains (Aldh^+, group A and B) and two groups of non-ILA-producing strains (Aldh^−, group C and D) from the pool of Bifidobacterium strains ([143]Figure 5A) for intervention in the CUMS mouse model. Bifidobacteria with the Aldh gene significantly elevated ILA levels in mouse colonic contents, serum, and hippocampus ([144]Figures 5C–5E). Moreover, these interventions alleviated anxiety- and depression-like behaviors induced by CUMS in mice ([145]Figures 5F–5H). Bifidobacteria without the Aldh gene intervention failed to effectively regulate indole metabolism in CUMS mice ([146]Figures 5C–5E and [147]S5B–S5I). Moreover, an examination of the gut microbiota in all intervention groups revealed a significant alteration in the gut microbial structure in CUMS mice ([148]Figure S5J). Although none of the four bifidobacteria intervention groups fully restored the microbial structure to that of the control group ([149]Figures S5J–S5L), the microbial characteristics of the group with Aldh^+ bifidobacteria intervention were notably distinct from those of CUMS mice ([150]Figure 5I). The debiased sparse partial correlation analysis between microbial abundance ([151]Data S2) and indole derivative levels indicated a positive correlation between certain genera and the levels of IA and IPA ([152]Figures S5B and S5C), including Clostridium, Mucispirillum, and Lachnospiraceae UCG-006 ([153]Figure 5J). Their abundance increased following the Aldh^+ bifidobacteria intervention ([154]Figure 5K). Considering that IA and IPA are downstream metabolites of ILA, these changes in microbial ecosystem structure may be a response to the elevated ILA levels. Microbial-derived ILA alleviates neurobehavioral abnormalities by reducing neuroinflammation In light of Bre1025’s ability to metabolize Trp, we investigated its effect on brain serotonin (5-HT) content, given its pivotal role as a neurotransmitter implicated in mood regulation and as one of the products of Trp metabolism.[155]^39^,[156]^40 Although we observed that Bre1025 supplementation led to a restoration of brain 5-HT levels ([157]Figure S6A), it did not significantly impact serum and hippocampal 5-HT precursor 5-hydroxytryptophan levels ([158]Figures S6B and S6C). This suggests that the observed changes in brain 5-HT levels were not influenced by alterations in precursor supply within the peripheral system but may represent concurrent alterations subsequent to the mitigation of mood disorders by Bre1025 in mice. Numerous studies have highlighted that stress can trigger an upsurge in corticosterone release in mice, and prolonged high corticosterone level is associated with the inhibition of neurogenesis in neural stem cells and the promotion of neuroinflammation in the hippocampus.[159]^41^,[160]^42 Compared to the Aldh^- strain, Aldh^+ bifidobacteria intervention effectively mitigated the aberrant rise in serum corticosterone induced by CUMS ([161]Figures 6A and [162]S5M). Aldh^+ bifidobacteria intervention also demonstrated inhibition of the gene overexpression of pro-inflammatory cytokines interleukin-1β (IL-1β) and IL-6 in the hippocampus ([163]Figures 6B, 6C, [164]S5N, and S5O). Simultaneously, it enhanced the expression of the anti-inflammatory cytokine IL-10 ([165]Figure 6D). However, the gene expression of tumor necrosis factor alpha (TNF-α) remained unaltered ([166]Figure S6D). Immunofluorescence detection on brain slices revealed that Bre1025 intervention suppressed the abnormal activation of microglial cells in the hippocampal CA3 region and dentate gyrus, while the Aldh mutant strain’s intervention was ineffective ([167]Figures 6E and [168]S6E). Specifically, CUMS causes hippocampal microglia to become hyper-ramified, and intervention with ILA and its producing bacteria (Bre1025) can inhibit this abnormal change ([169]Figure 6E). Given that Bre1025 intervention significantly increased ILA levels in the hippocampus of CUMS mice ([170]Figure 4D), we hypothesized that the neurophysiological mechanism underlying the strain’s antidepressant effect might involve the ILA-AhR signaling pathway, known for its role in modulating inflammatory levels.[171]^43^,[172]^44 As expected, the gene expression of AhR induced by ILA as its ligand, along with its downstream targets CYP1A1, CYP1A2, and CYP1B1, was upregulated under the strain intervention ([173]Figures 6F, S5P, and S6F–I), consequently promoting the gene expression of the downstream anti-inflammatory cytokine IL-22 ([174]Figures 6G and [175]S5Q). Figure 6. [176]Figure 6 [177]Open in a new tab Microbial-derived ILA alleviates neuroinflammation through the aryl hydrocarbon receptor (A) Serum corticosterone concentrations in mice (n = 8/treatment). (B–D) Expression analysis of genes associated with inflammatory factors in mouse hippocampus (n = 8/treatment). (E) Immunofluorescence and immunohistochemical detection of Iba1 in the mouse brain (n = 3/treatment). Immunofluorescence analysis measured the percentage of Iba1 signal area relative to the total CA3 area in hippocampal. Sholl analysis was conducted on immunohistochemical slices to assess branching of labeled microglial cells, recording the intersection number every 4 pixels. (F) Western blot analysis of aryl hydrocarbon receptor (AhR) and CYP1A1 in the hippocampus of mice (n = 4/treatment). (G) Expression of interleukin-22 gene (Il22) in mouse hippocampus (n = 8/treatment). (H) Protective effect of ILA intervention on the viability of NE-4C and BV-2 cells (n = 3/treatment). (I and J) Expression of AhR (Ahr) and IL-22 (Il22) in NE-4C and BV-2 cells (n = 3/treatment). (K) Expression of TNF-α (Tnf) in NE-4C and BV-2 cells (n = 3/treatment). ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 indicate significant differences between groups, as determined by one-way ANOVA followed by Sidak post hoc test. P = numbers indicate significant differences between groups, as determined by two-tailed Student’s t test. To further substantiate this hypothesis, we employed corticosterone intervention to establish an in vitro neuroinflammation model using neural stem cells (NE-4C) and microglial cells (BV-2). Corticosterone intervention significantly inhibited the viability of NE-4C cells ([178]Figure S6J), while pre-treatment with ILA demonstrated a dose-dependent mitigation of corticosterone-induced neurotoxicity ([179]Figures 6H and [180]S6K). Furthermore, ILA intervention activated the AhR signaling pathway in a dose-dependent manner, which led to the promotion of gene expression of the anti-inflammatory cytokine IL-22 and inhibition of the expression of pro-inflammatory cytokines TNF-α and IL-1β ([181]Figures 6I–6K and [182]S6M). The use of the AhR inhibitor CH-223191 (2-methyl-2H-pyrazole-3-carboxylic acid) inhibited the activation of the AhR pathway by ILA in NE-4C cells ([183]Figures 6I, 6J, and [184]S6L), consequently nullifying its protective and anti-inflammatory effects on NE-4C viability ([185]Figure 6H). While pre-treatment with ILA did not alleviate the viability damage caused by corticosterone in microglial BV-2 cells ([186]Figure 6H), its activation of the AhR signaling pathway and regulation of inflammatory factors were consistent with the results observed in NE-4C cells ([187]Figures 6I–6K). Therefore, we speculate that Bre1025, by elevating brain ILA levels, activates the AhR signaling pathway to maintain neuroimmune homeostasis and ultimately suppress CUMS-induced neurobehavioral abnormalities. Discussion Previously, several independently gathered pieces of evidence suggest that microbial ILA from the gut may play a role in regulating brain function. This is primarily evident in (1) endogenous indole derivatives in the human body that are primarily synthesized from dietary tryptophan by the gut microbiota and distributed throughout the body via the bloodstream.[188]^45 (2) Patients with MDD exhibit significantly lower serum levels of ILA compared to healthy controls. Additionally, supplemental ILA at healthy physiological levels appears to adversely affect neurobehavioral outcomes.[189]^32 (3) Indole derivatives have been shown to cross the blood-brain barrier,[190]^43^,[191]^46 and their primary receptor, AhR, is expressed throughout the brain, including regions associated with emotional regulation.[192]^47 (4) Oral administration of exogenous indole compounds can activate the AhR signaling pathway in the central nervous system, influencing behavior and modulating neuroinflammation.[193]^45 Overall, direct evidence regarding the involvement of gut microbiota-derived ILA in the regulation of brain function is currently very scarce and remains in the stage of data accumulation. Here, we have identified that the level of gut ILA serves as a core metabolic marker distinguishing normal and depressed mice ([194]Figures 1G and 1I), and supplementation with specific bifidobacteria strains successfully reversed the reduced levels of ILA in the gut and hippocampus of depressed mice ([195]Figures 5C–5E). Our exploration of the molecular pathway for ILA synthesis in bifidobacteria identified Aldh as the key gene responsible for ILA synthesis ([196]Figure 3). This has been confirmed in several bifidobacteria strains, including nine different strains across three species: B. longum subsp. longum, B. longum subsp. infantis, and B. breve.[197]^48 Our study further validates that the Aldh gene is essential for ILA production across all nine commonly found human gut bifidobacteria species, encompassing a total of 75 strains. Given genetically manipulated strains to germ-free animals, we demonstrated that the mutation of Aldh in Bre1025 directly resulted in the loss of its antidepressant effects in mice ([198]Figures 4G–4I). These results provide compelling evidence for the direct involvement of gut microbiota-derived ILA in emotional regulation and establish ILA-producing capacity as an indicator for the antidepressant effects of psychobiotic strains. While the majority of research within the “microbiota-gut-brain axis” framework tends to focus on macroscopic effects on gut microbiota balance,[199]^49^,[200]^50 we advocate for the crucial recognition of the distinct role of individual psychobiotic strains in their interactions with the host. Firstly, clarifying the pharmacological targets and mechanisms of psychobiotics can contribute to their more scientific use, including dosage selection and personalized choices, thus facilitating the translational potential of psychobiotics into clinical applications.[201]^51^,[202]^52 Secondly, confirming the common molecular and genetic characteristics of strains can guide more efficient psychobiotic screening.[203]^53 As demonstrated by our widespread verification results within the Bifidobacterium genus, species with the Aldh gene and ILA synthesis ability exhibit better antidepressant potential compared to those without ([204]Figure 5). This shift in perspective has the potential to optimize psychobiotic screening, moving away from the constraints of labor-intensive and somewhat random phenotype-based methods, such as animal behavioral experiments and psychological assessments after human consumption trials.[205]^54 Notably, Bifidobacterium species capable of ILA synthesis, including B. longum, B. breve, and B. bifidum, exhibit proficiency in utilizing human milk oligosaccharides (HMOs) and establishing dominance in the early-life gut environment.[206]^55 Since our study was conducted in adult mice, we do not propose a direct association between bifidobacteria ILA synthesis and the utilization of HMOs. Previous research has emphasized the role of these species in producing significant amounts of aromatic lactic acids, including ILA, during early life and their regulation of gut immune development through the AhR signaling pathway.[207]^38 In line with this but extending further, our results demonstrate that the immunomodulatory effects of ILA produced by gut microbiota extend to the brain. Specifically, we confirmed that ILA is directly synthesized from Trp through Aldh and is substrate dependent ([208]Figure 3A). This discovery partially explains earlier observations, such as the occurrence of depressive symptoms under Trp-deprived diets and the overactivation of microglia in the brains of germ-free mice,[209]^45^,[210]^56 which often exhibit depressive-like behavior. However, the precise interconnections between these factors still require further confirmation. In summary, our research establishes that gut microbiota-derived ILA plays a direct role in regulating brain inflammation through the AhR signaling pathway, thereby influencing neurobehavioral outcomes. Bifidobacterium strains with the ability to synthesize ILA emerge as promising psychobiotics. These findings contribute to the expansion of our understanding of the mechanisms underlying depression and the microbiota-gut-brain axis theory. Additionally, they provide valuable insights at the molecular and genetic levels that can guide the selection and development of psychobiotic strains, as well as their future clinical applications in addressing mental disorders, including but not limited to depression. Limitations of the study Our study has some limitations that need to be addressed in future work. Firstly, the presence of the Aldh gene strongly distinguishes antidepressant effects among bifidobacteria primarily at the species level, rather than the strain level, since the Aldh gene is also annotated in Bre3M5, despite significant differences in ILA production between the two strains ([211]Figure S1A). This suggests that relying solely on genomic features for psychobiotics discrimination is relatively inefficient; a combined approach considering both genomic and metabolic characteristics is essential. Nonetheless, this emerging approach proves much more efficient compared to traditional screening methods based on in vivo assessments. Additionally, we observed that B. bifidum exhibits higher ILA production capabilities compared to other species. Whether this is related to the additional presence of the Fldh gene and the superior potential in antidepressant effects warrants further evaluation. Lastly, due to constraints on animal welfare, a limited number of mice were used to explore potential differences in ILA-mediated neurobehavioral regulation between male and female individuals, despite the recognized necessity for such investigations in behavioral neuroscience studies. Future studies need to be done to address these issues. Resource availability Lead contact For additional information and resource requests, please contact and they will be addressed by the lead contact, Peijun Tian (pjtian@jiangnan.edu.cn). Materials availability Requests for access to the materials generated in this study can be directed to the [212]lead contact. It should be noted that this study did not produce any new unique reagents. Data and code availability * • Genome sequencing data have been deposited at SRA and are publicly available as of the date of publication. Accession numbers are listed in the [213]key resources table. Original western blot images have been deposited at Mendeley Data and are publicly available as of the date of publication. The DOI is listed in the [214]key resources table. Microscopy data reported in this paper will be shared by the [215]lead contact (Peijun Tian, pjtian@jiangnan.edu.cn) upon request. * • This study does not include original code. * • Any additional information required to reanalyze the data reported in this paper is available from the [216]lead contact (Peijun Tian, pjtian@jiangnan.edu.cn) upon request. Acknowledgments