Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Jul 11;16(1):6422.
doi: 10.1038/s41467-025-61544-0.

Gut microbiota and brain-resident CD4+ T cells shape behavioral outcomes in autism spectrum disorder

Affiliations

Gut microbiota and brain-resident CD4+ T cells shape behavioral outcomes in autism spectrum disorder

John Chulhoon Park et al. Nat Commun. .

Abstract

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by alterations in social, repetitive, and anxiety-like behaviors. While emerging evidence suggest a gut-brain etiology in ASD, the underlying mechanisms remain unclear. To dissect this axis, we developed a germ-free BTBR mouse model for ASD. The absence of gut microbiota in male mice ameliorates ASD-associated behaviors and reduces populations of inflammatory brain-resident T cells. Additionally, CD4+ T cell depletion mitigates neuroinflammation and ASD behaviors, suggesting a gut-immune-brain axis. We identify several microbial and metabolic regulators of ASD, particularly those relevant to the glutamate/GABA ratio and 3-hydroxyglutaric acid. Using an in silico metabolite prediction model, we propose Limosilactobacillus reuteri IMB015 (IMB015) to be a probiotic candidate. Administration of IMB015 reduces the glutamate/GABA ratio and neuroinflammation, resulting in improved behaviors. Here we report a gut-immune-brain axis in which the gut microbiota and its metabolites can modulate brain-resident immune cells and ASD-associated behaviors.

PubMed Disclaimer

Conflict of interest statement

Competing interests: J.L., S.L., and S.Y.C. are employees of ImmunoBiome. S.-H.I. is the founder and major shareholder of ImmunoBiome. The other authors declare no competing interests.

Figures

Fig. 1
Fig. 1. GF BTBR do not manifest ASD-associated phenotypes.
A Representative heat maps and time spent in either the social or empty chambers during the three-chamber sociability test. SPF BTBR n = 18, GF BTBR n = 22. P value from two-way ANOVA, F1,76 = 1.948. B Representative heat maps and time spent in either the familiar or novel chambers during the three-chamber social novelty test. SPF BTBR n = 18, GF BTBR n = 22. P value from two-way ANOVA, F1,76 = 5.817. C Representative images and percent of marbles buried during the marble burying test. SPF BTBR n = 12, GF BTBR n = 18. P = 0.000002 from two-tailed Welch’s t-test, t(11.97) = 6.540. D Representative traces, time spent in the center zone, and frequency to the center zone during the open field maze. SPF BTBR n = 10, GF BTBR n = 12. P value from two-tailed Welch’s t-test, t(12.76) = 3.785 and t(18.95) = 0.7207, respectively. E Mean velocity and distance traveled during the open field maze. SPF BTBR n = 10, GF BTBR n = 12. P value from two-tailed Welch’s t-test, t(18.95) = 0.7207 and t(18.94) = 0.7229. FH Representative immunofluorescence staining and number of c-FOS+ neurons per area in the amygdala (two-tailed Student’s t-test, t(5) = 3.562. SPF BTBR n = 5 sections, 3 mice, GF BTBR n = 7 sections, 4 mice) (G) and dentate gyrus of the hippocampus (t(10) = 3.769. SPF BTBR n = 6 sections, 3 mice, GF BTBR n = 6 sections, 3 mice) (H). All data represented as mean ± SEM. Source data are provided as source data file.
Fig. 2
Fig. 2. Microbiota regulates brain-resident T cells in BTBR mice.
A Representative flow cytometry plots and frequency of CD4+ or CD8+ T cells gated from TCRβ+ CD45+ live cells within the whole brain parenchyma of SPF or GF BTBR mice. SPF BTBR n = 5 mice, GF BTBR n = 4 mice. P values from two-tailed Student’s t-test, t(7) = 4.385 and t(7) = 2.958, respectively. B Representative flow cytometry plots and frequency of Foxp3+ Tregs gated from CD4+ T cells in the brain. SPF BTBR n = 5 mice, GF BTBR n = 4 mice. P value from two-tailed Student’s t-test, t(7) = 6.887. C, D Representative flow cytometry plots and frequency of IFN-γ+ Foxp3- CD4+ T cells or IFN-γ+ CD8+ T cells in the brain parenchyma. SPF BTBR n = 5 mice, GF BTBR n = 4 mice. P values from two-tailed Student’s t-test, t(7) = 7.625 and t(7) = 4.912, respectively. E Representative flow cytometry plots and frequency of CD69+ T cells in the brain gated from TCRβ+ CD45+ cells. SPF BTBR n = 5 mice, GF BTBR n = 4 mice. P values from two-tailed Student’s t-test, t(7) = 3.028. F Representative flow cytometry plots and frequency of Ki67+ T cells in the brain gated from CD45+ cells. SPF and GF BTBR n = 8 mice, P value from two-tailed Student’s t-test, t(14) = 3.185. Flow cytometric analyses are representative results from at least 2 independent experiments with minimum of 4 mice. All data represented as mean ± SEM. Source data are available in source data file.
Fig. 3
Fig. 3. Depletion of CD4+ T cells abrogate ASD-associated behaviors and neuroinflammation.
A SPF BTBR pups were injected with 25 μg/g of anti-CD4 by i.p. on P5, P13, and P36. Mice were analyzed at 8 weeks of age. B, C BTBR αCD4 mice had reduced frequency (B) and cell number (C) of brain-resident CD4+ T cells. CD8+ T cell frequency was increased (B), but not its absolute number (C). BTBR Ctrl n = 3 mice, BTBR αCD4 n = 5 mice. P value from two-tailed Student’s t-test, t(6) = 5.437 and t(6) = 4.701 (B) and t(6) = 4.719 and t(6) = 0.9473 (C). D, E Time spent in chambers during the three-chamber sociability (D) and social novelty (E) tests. BTBR αCD4 sociability P = 0.00009, P value from two-way ANOVA, F1,40 = 2.590 and F1,40 = 12.42, respectively. BTBR Ctrl n = 9, BTBR αCD4 n = 13. F Percent of marbles buried during the marble burying test. P value from two-tailed Welch’s t-test, t(28.30) = 3.929. BTBR Ctrl n = 16, BTBR αCD4 n = 15. G, H Time spent in the center zone (G) and mean velocity (H) during the open field maze. BTBR Ctrl n = 14, BTBR αCD4 n = 13. P value from two-tailed Welch’s t-test, t(23.52) = 3.077 and t(24.98) = 2.402, respectively. I, J Representative flow cytometry plots, frequencies (I) and absolute cell numbers (J) of microglia and macrophages from live cells of the brain parenchyma. BTBR Ctrl n = 3 mice, BTBR αCD4 n = 5 mice. P values from two-tailed Student’s t-test, t(6) = 6.006 and t(6) = 1.733 (I) and t(6) = 3.458 and t(6) = 0.6370 (J). K–N Representative flow cytometry plots and frequencies of IL-10+ (K), IL-6+ (L), TNF-α+ (M), and IL-1β+ (N) microglia. BTBR Ctrl n = 3 mice, BTBR αCD4 n = 5 mice. P values from two-tailed Student’s t-test, t(6) = 2.752 (K), t(6) = 4.726 (L), t(6) = 3.538 (M), t(6) = 0.5021 (N). Flow cytometric analyses are representative results from at least 2 independent experiments with minimum of 3 mice. All data represented as mean ± SEM. Source data are available in source data file.
Fig. 4
Fig. 4. Identification of gut microbiota and metabolites that drive ASD in BTBR mice.
A Schematic for antibiotic treatment. B Representative images and percent of marbles buried. SPF and GF BTBR n = 5, AVNM n = 9, AVN n = 22, ANM n = 25, AVM n = 17 VNM n = 20. SPF BTBR vs AVN P = 0.00008, P values from one-way ANOVA with Bonferroni’s correction, F6,96 = 10.42. C Linear discriminant analysis Effect Size (LEfSe) for genera enriched in AVNM or VNM groups. Genera with FDR-adjusted P value cutoff of 0.01 and Log LDA > 2.0 shown. D Volcano plot of fecal metabolites enriched in AVNM or VNM mice. Fold change threshold of 2.0 with two-tailed raw P value threshold of 0.05 from unpaired analysis with equal group variances. Significant metabolites labeled. EG Relative glutamate (E), glutamate to GABA ratio (F), and 3-hydroxyglutaric acid (OH-GlutAcid) (G) levels in the feces normalized to AVNM. n = 6 mice. P = 0.000002 (E) and P = 0.000008 (G). P values from two-tailed Student’s t-test, t(10) = 9.627 (E), t(10) = 3.611 (F), t(10) = 8.325 (G). H Schematic of L. murinus isolation and monoassociation. I Representative images and percent of marbles buried. SPF BTBR n = 9, GF BTBR n = 18, GF BTBR + L. murinus n = 10. SPF BTBR vs GF BTBR P = 0.00003. P values from one-way ANOVA with Bonferroni’s correction, F2,34 = 15.97. JL Relative glutamate (J), glutamate to GABA ratio (K), and 3-hydroxyglutaric acid (L) levels in feces after L. murinus monoassociation. GF BTBR n = 6 mice, L. murinus n = 5 mice. P = 0.00000005 (J), P = 0.00008 (K). P values from two-tailed Student’s t-test, t(9) = 16.37 (J), t(9) = 6.768 (K), t(9) = 2.924 (L). M, N Representative flow cytometry plots and frequencies of IFN-γ+ Foxp3- CD4+ (M) and IFN-γ+ CD8+ (N) T cells in the brain after L. murinus monoassociation. SPF and GF BTBR n = 4 mice, L. murinus n = 5 mice. P values from one-way ANOVA with Bonferroni’s correction, F2,10 = 11.02 (M), F2,10 = 5.215 (N). All data represented as mean ± SEM. Source data are available in source data file.
Fig. 5
Fig. 5. Healthy B6 microbiota confers neurotypical behavior and neuroimmune modulation in BTBR mice.
A 3-weeks-old GF BTBR pups were cohoused with either SPF B6 (BTBR-B6) or SPF BTBR (BTBR-BTBR) mice. B, C Duration in chambers during the sociability (B) and social novelty (C) tests. BTBR-BTBR n = 14, BTBR-B6 n = 24, P values from two-way ANOVA, F1,72 = 8.186 (B), F1,72 = 10.61 (C). D Percent of marbles buried. BTBR-BTBR n = 24, BTBR-B6 n = 22, P value from two-tailed Welch’s t-test, t(40.98) = 3.213. E–G Representative immunofluorescence staining and number of c-FOS+ neurons per area in the amygdala (F) and dentate gyrus (G). P value from two-tailed Student’s t-test, t(26) = 4.445, BTBR-BTBR n = 15 sections, 5 mice, BTBR-B6 n = 13 sections, 5 mice (F). t(22) = 4.180, BTBR-BTBR n = 12 sections, 5 mice, BTBR-B6 n = 12 sections, 5 mice (G). H, I Representative flow cytometry plots and frequencies of IFN-γ+ Foxp3- CD4+ (H) or IFN-γ+ CD8+ (I) T cells in the brain. BTBR-BTBR and BTBR-B6 n = 4 mice, P value from two-tailed Student’s t-test, t(6) = 2.711 (H), t(6) = 2.507 (I). J Representative flow cytometry plots and frequencies of CD11b+ CD45int microglia in the brain. BTBR-BTBR and BTBR-B6 n = 5 mice, P values from two-tailed Student’s t-test, t(8) = 2.404. KN Representative flow cytometry plots and frequencies of IL-10+ (K), IL-6+ (L), TNF-α+ (M), or IL-1β+ (N) microglia in the brain. BTBR-BTBR n = 6 mice, BTBR-B6 n = 5 mice. P values from two-tailed Student’s t-test, t(9) = 2.289 (K), t(9) = 2.988 (L), t(9) = 1.982 (M), t(9) = 3.125 (N). Flow cytometry data are representative of at least 2 independent experiments with minimum 3 mice. All data represented as mean ± SEM. Source data are available in source data file.
Fig. 6
Fig. 6. Healthy gut microbiota reduces ASD-associated metabolites in BTBR mice.
A Beta diversity from Principal coordinates analysis (PCoA) based on Bray-Curtis dissimilarity index demonstrating a unique microbiota composition following colonization with healthy gut microbiota from B6 mice (Pairwise PERMANOVA, F = 5.61, R2 = 0.38, P = 0.004). B LEfSe analysis of genera enriched in SPF BTBR or B6 microbiota recipient BTBR (BTBR-B6) mice. Genera with FDR-adjusted P value cutoff of 0.1 and Log LDA > 2.0 shown. C Volcano plot of fecal metabolites enriched in SPF BTBR or BTBR-B6 mice. Fold change threshold of 1.5 with two-tailed raw P value threshold of 0.05 from unpaired analysis with equal group variances. All significant metabolites labeled. DF Relative glutamate (D), glutamate to GABA ratio (E), and 3-hydroxyglutaric acid (F) levels in feces after colonization with healthy gut microbiota. SPF BTBR n = 5, BTBR-B6 n = 6. P values from two-tailed Student’s t-test, t(9) = 2.299 (D), t(9) = 2.000 (E), t(9) = 2.466 (F). All data represented as mean ± SEM. Source data are available in source data file.
Fig. 7
Fig. 7. L. reuteri IMB015 therapeutically alleviates ASD-associated phenotypes.
A Schematic diagram of in silico metabolic flux prediction. B Predicted GABA and L-glutamate production or uptake in IMB015. C Schematic diagram of IMB015 treatment. D, E Time spent in chambers during the sociability (D) and social novelty (E) tests. BTBR Vehicle n = 24, BTBR IMB015 n = 20, P values from two-way ANOVA, F1,84 = 0.096673 (D), F1,84 = 21.67 (E). F Representative images and percentage of marbles buried. BTBR Vehicle n = 21, BTBR IMB015 n = 13. P = 0.000000005. P value from two-tailed Welch’s t-test, t(22.62) = 9.059. G, H Time spent in the center zone (G) and mean velocity (H) during the open field maze. BTBR Vehicle n = 14, BTBR IMB015 n = 19. P value from two-tailed Welch’s t-test, t(15.43) = 2.494 (G), t(28.60) = 0.190 (H). IK Representative immunofluorescence staining and number of c-FOS+ neurons per area in the amygdala (two-tailed Student’s t-test, t(16) = 3.593. BTBR Vehicle n = 10 sections, 3 mice, BTBR IMB015 n = 8 sections, 4 mice) and dentate gyrus (t(29) = 3.504. BTBR Vehicle n = 16 sections, 3 mice, BTBR IMB015 n = 15 sections, 3 mice). LN Relative glutamate (L), glutamate/GABA ratio (M), and Oh-GlutAcid (N) levels in feces after IMB015 treatment. BTBR Vehicle n = 8, BTBR IMB015 n = 5. P values from two-tailed Student’s t-test, t(11) = 2.382 (L), t(11) = 3.252 (M), t(11) = 2.108 (N). O Absolute cell count of Tregs in the brain from flow cytometry. BTBR Vehicle n = 14, BTBR IMB015 n = 7. P value from two-tailed Student’s t-test, t(19) = 2.891. P, Q Representative flow cytometry plots and frequencies of IFN-γ+ Foxp3- CD4+ (P) or IFN-γ+ CD8+ (Q) T cells in the brain. BTBR Vehicle n = 5 mice, BTBR IMB015 n = 4 mice. P value from two-tailed Student’s t-test, t(7) = 3.741 (P), t(7) = 3.010 (Q). Flow cytometry data from 2 independent experiments with minimum 3 mice. Data represented as mean ± SEM. Source data are available in source data file.

References

    1. Park, J. C. & Im, S. H. The gut-immune-brain axis in neurodevelopment and neurological disorders. Microbiome Res. Rep.1, 23 (2022). - PMC - PubMed
    1. Sharon, G. et al. Human gut microbiota from autism spectrum disorder promote behavioral symptoms in mice. Cell177, 1600 (2019). - PMC - PubMed
    1. Lewandowska-Pietruszka, Z., Figlerowicz, M. & Mazur-Melewska, K. Microbiota in autism spectrum disorder: a systematic review. Int. J. Mol. Sci.24, 10.3390/ijms242316660 (2023). - PMC - PubMed
    1. Buffington, S. A. et al. Dissecting the contribution of host genetics and the microbiome in complex behaviors. Cell184, 1740–1756.e1716 (2021). - PMC - PubMed
    1. Sgritta, M. et al. Mechanisms underlying microbial-mediated changes in social behavior in mouse models of autism spectrum disorder. Neuron101, 246 (2019). - PMC - PubMed

MeSH terms