Enhancing social behavior in an autism spectrum disorder mouse model: investigating the underlying mechanisms of Lactiplantibacillus plantarum intervention

Figures & data

Figure 1. PS128 rescues behavioral abnormalities of the VPA mice.

(a) Schematic of the experimental design. WT mice were the offspring of PBS-injected dams and received pure water (H₂O) as a control group. VPA mice were the offspring of VPA-injected dams, and were randomly divided into two groups to receive either pure water or water mixed with PS128. (b–d) Three-chamber social test protocol (b). Sociability (c) was quantified using the social preference index, which reflects each individual animal’s preference to interact with a mouse (M1) versus an empty cup (Empty). Social novelty (d) reflects each individual animal’s preference to interact with an unfamiliar mouse (M2) versus a familiar mouse (M1). N = 10-13 per group; one-way ANOVA with Fisher’s LSD post hoc test. *: p < .05, **: p < .01, ***: p < .001. (e–h) Reciprocal social interaction test setup (e). The amount of interaction time of the test mouse spent in sniffing (f) and following (g) an unfamiliar reference mouse or self-grooming time (h) during the test period. N = 10-12 per group, one-way ANOVA with Tukey post hoc test. *: p < .05, **: p < .01, ***: p < .001. (i–j) Anxiety, quantified as the Anxiety index (j), is defined as the percentage of time the test mouse spent in the closed arms in the elevated plus maze (i). N = 10 per group, one-way ANOVA with Tukey post hoc test. *: p < .05, **: p < .01, ***: p < .001. (k–l) Spatial memory was accessed using the Y maze test protocol (k) to measure the preference of the animal’s preference to the novel arm. N = 10 per group, one sample t-test for each group with Bonferroni correction; dashed line indicates equal time spent exploring novel and familiar arms. n.s.: not significant, *: p < .05/3, **: p < .01/3, ***: p < .001/3. Plots show mean ± SEM.

Figure 2. PS128 restores dendritic morphology of VPA mice.

(a–c) Representative images and 3D reconstructions of dendrites of GFP-expressing neurons in hippocampal CA1 (a), CA2 (b) or prefrontal cortex (c). Basal, apical dendrites and soma are marked in black, blue and green, respectively. (d–f) Dendritic architecture and complexity, quantified by Sholl analyses, are impaired in VPA mice. PS128 restores dendritic structure abnormalities. N = 5 mice per group and 5 neurons per mouse, one-way ANOVA with Tukey post hoc test; # indicates p < .05 between WT H₂O vs. VPA H₂O and VPA H₂O vs. VPA PS128. (g) Total dendritic length is reduced in VPA mice, but is rescued by PS128. (h) Total number of dendritic branch points are decreased in VPA mice, but restored by PS128. N = 5 mice per group and 5 neurons per mouse; one-way ANOVA with Tukey post hoc test. *: p < .05, **: p < .01, ***: p < .001. Plots show mean ± SEM.

Figure 3. PS128 rescues spine loss in VPA mice.

(a, c, e) Representative images of dendritic spines of pyramidal neurons in hippocampal CA1, CA2 and prefrontal cortex. (b, d, f) Quantification of spine density. Spine density is significantly reduced in both apical and basal dendrites of pyramidal neurons in VPA mice. PS128 rescues spine loss in VPA mice. N = 5 mice per group and 5 neurons per mouse, one-way ANOVA with Tukey post hoc test. (g) Schematic for spine type classification. (h–j) Spine type distributions are comparable among all groups. N = 5 mice per group and 5 neurons per mouse, two-way ANOVA with Tukey post hoc test. L: length, D: diameter, Dh: head diameter, Dn: neck diameter. *: p < .05, **: p < .01, ***: p < .001. Plots show mean ± SEM.

Figure 4. PS128 restores reduced kinase activities and synaptic glutamate receptors in the hippocampus of VPA mice.

(a) Phosphorylated Erk1/2, relative to total Erk1/2, is significantly reduced in VPA mice, but is restored by PS128. (b) Phosphorylated CaMKIIα, relative to total CaMKIIα, is attenuated in VPA mice, but is rescued by PS128. (c) Phosphorylated PKA-C, relative to total PKA-C, is normal in all groups. (d) GluN1 is reduced in the homogenate of VPA mice, but is restored by PS128. Immunoblots were stripped and reprobed with GAPDH for normalization. (e–g) Phosphorylated ERK 1/2 (e), CaMKIIα (f), and PKA-C (g), relative to their own total protein levels, are decreased in the hippocampal PSD fraction of VPA mice. PS128 restores their activities in the PSD. (h) Glutamate receptor subunits GluN1 and GluA2 are reduced in the PSD fraction of VPA mice, but are restored by PS128. N = 6-7 mice per group; Protein levels were analyzed by one-way ANOVA with Tukey post hoc test. *: p < .05, **: p < .01, ***: p < .001. Plots show mean ± SEM.

Figure 5. PS128 restores oxytocin expression in the PVN of VPA mice.

(a) Representative images of PVN stained with Oxytocin and NeuN antibodies and DAPI counterstaining. (b) Quantification of oxytocin immunofluorescence signals in the PVN. (c) Quantification of oxytocin positive cells in the PVN. N = 5 mice per group, one-way ANOVA with Tukey post hoc test. *: p < .05, **: p < .01, ***: p < .001. Plots show mean ± SEM.

Figure 6. PS128 loses ability to rescue VPA-induced behaviors in mice lacking oxytocin receptors.

(a) qPCR confirmation of complete oxytocin receptor loss in the hippocampus of Oxtr-KO mice. N = 6 mice per group, Mann-Whitney test; *: p < .05, **: p < .01, ***: p < .001. (b and c) Sociability and social novelty are impaired in VPA mice and Oxtr-KO VPA mice. PS128 only rescues VPA mice but not Oxtr-KO VPA mice. N = 10 per group; one-way ANOVA with Fisher’s LSD test; *: p < .05, **: p < .01, ***: p < .001. (d and e) Reciprocal social interaction in VPA mice and Oxtr-KO VPA mice are decreased in sniffing (d) and following (e) the unfamiliar mice. PS128 only restores reciprocal social behaviors in VPA mice but not Oxtr-KO VPA mice. N = 10-11 per group, one-way ANOVA with Tukey post hoc test; *: p < .05, **: p < .01, ***: p < .001. (f) Anxiety is higher in VPA mice and Oxtr-KO VPA mice. PS128 only rescues VPA mice but not Oxtr-KO VPA mice. N = 10-11 per group, one-way ANOVA with Tukey post hoc test; *: p < .05, **: p < .01, ***: p < .001. (g) Spatial memory is impaired in the VPA and Oxtr-KO VPA mice. PS128 only rescues VPA mice but not Oxtr-KO VPA mice. N = 10-11 per group, one sample t-test for each group with Bonferroni correction; dashed line indicates equal time spent exploring novel and familiar arms; *: p < .05/4, **: p < .01/4, ***: p < .001/4. Plots show mean ± SEM.

Figure 7. PS128 increases Bifidobacterium abundance in the gut, which correlates with improvements in ASD-associated behaviors in VPA mice.

(a) Alpha diversity- The Chao1 index, a species richness index that estimates the expected number of species based on the abundance profile, is comparable across all groups. (b) Alpha diversity- The Shannon index, a diversity index that incorporates species richness and evenness, is lower in the VPA PS128 group. For (a) and (b), N = 10-11 per group, one-way ANOVA with Tukey post hoc test; boxes represent the interquartile range; lines indicate medians, and whiskers indicate the range. *: p < .05, **: p < .01, ***: p < .001; Plots show mean ± SEM. (c) Beta diversity- Principal Coordinates Analysis (PCoA) derived from weighted UniFrac distances between all samples in the six groups. Colored dots represent all samples from six groups. N = 10-11 per group, permutational multivariate ANOVA. (d) The LEfSe analysis for differentially abundance (Linear Discriminant Analysis (LDA) scores > 2) shows significant differences in bacterial abundance among all groups. (e and f) The relative abundance of L. plantarum (e) and Bifidobacterium (f) is significantly higher in the PS128 treated groups. N = 10-11 per group. Mann-Whitney test for comparisons between the WT H₂O vs. WT PS128 groups, VPA H₂O vs. VPA PS128 groups, and Oxtr-KO, VPA H₂O vs. Oxtr-KO, VPA PS128 groups. The Bonferroni method was applied to correct for Type I errors. *: p < .05/3, **: p < .01/3, ***: p < .001/3). (g–j) Correlation analysis was conducted between the relative abundance of Bifidobacterium in the VPA H₂O and VPA PS128 groups combined (N = 20-21) and their corresponding behavioral performance. Notably, a significant positive correlation was observed in the sociability test (g) and the elevated plus maze task (j). r = .4686, p = .037 and r = .5683, p = .0072, respectively. No significant correlations were found for reciprocal social test (h) and the Y maze test (i). *: p < .05, **: p < .01, ***: p < .001. (k and l) Tax4Fun2 prediction based on metagenomic 16S rRNA sequencing shows that the groups supplemented with PS128 have higher abundance in amino acid (k) and secondary metabolite (l) biosynthesis. N = 10-11 per group, Mann-Whitney test between WT H₂O vs. WT PS128 groups; VPA H₂O vs. VPA PS128; and Oxtr-KO, VPA H₂O vs. Oxtr-KO, VPA PS128 groups, respectively. The Bonferroni correction was used to control the Type I error for multiple comparisons. *: p < .05/3, **: p < .01/3, ***: p < .001/3).

Data availability statement

The data supporting the findings of this study are available within the article and its supplementary materials. The raw sequencing data for all 61 samples are deposited on Figshare https://figshare.com/s/6e1b71b946d8f74a7505.