Fig 1.
Zinc, Ser, and BCAA act in concert to regulate synaptic signaling and the synaptic response.
Zinc is highly concentrated at synaptic vesicles and is co-released with glutamate. At the postsynaptic site, zinc induces condensate formation of the postsynaptic proteins and maintains dendritic spine morphology. Synaptic stimulation via the action of glutamate and D-Ser on ion channels activates downstream signaling and promotes protein synthesis, which is required for the long-term synaptic response and synaptic remodeling. BCAA enhances protein synthesis via activation of the mTOR pathway.
Fig 2.
Cocktail supplementation alters the total proteome of Tbr1+/− mouse brains.
Four groups of total mouse brain lysates—WT_water, Tbr1+/−_water, WT_1/4 cocktail, and Tbr1+/−_1/4 cocktail—were subjected to LC-MS-MS analysis. The results were subjected to (A) principal component analysis (PCA) and (B–D) Python package for weighted correlation network analysis (PyWGCNA) followed by protein network analysis using STRING. (B) Heatmap of module–trait relationships identified by PyWGCNA. Seven modules and respective numbers of proteins (in brackets) are labeled below the heatmap. In addition to the four groups of samples, the same analyses were performed according to genotype and treatment. In each well, the upper numbers indicate the correlation and the lower numbers in brackets indicate the P value of each analysis. The analysis revealed the highest level of correlation between the “Black” module and the treatment. (C) The Eigengene expression for the Black module. Both the Tbr1+/−_water and WT_water groups display reduced expression compared to the Tbr1+/−_cocktail and WT_cocktail groups. (D) The protein networks and GO for the Black module were analyzed with STRING. The GO terms are color-coded. The corresponding false discovery rates and gene counts are also indicated. See also S1 and S2 Data and S1–S4 Figs.
Fig 3.
Immunoblotting validation of proteins listed in the Black module.
Four groups of total mouse brain lysates—WT_water, Tbr1+/−_water, WT_1/4 cocktail, and Tbr1+/−_1/4 cocktail––were prepared for immunoblotting. Each group comprises eight mice for analysis. A total of nine antibodies were applied for immunoblotting, as indicated. Quantification was performed by normalization against the total protein levels revealed by Coomassie Blue stain. (A) Representative immunoblots. (B) Quantification of immunoblots. The statistical analysis was conducted using two-way ANOVA and Bonferroni’s post-hoc correction. # P < 0.05; ## P < 0.01; ### P < 0.001 in two-way ANOVA test. *, P < 0.05; **, P < 0.01 in the post-hoc test. Full scans of immunoblots are available in S1 Raw Images. The data underlying the graphs in (B) can be found in the S3 Data. All statistical analyses and results, including the actual P-values, are summarized in S4 Data.
Fig 4.
Neural ensembles in the BLA are influenced by Tbr1 haploinsufficiency and cocktail supplementation.
(A) Schematic of the experimental design (WT: n = 4; Tbr1+/−: n = 4). (B) Activity of neurons in the basolateral amygdala (BLA) during object exploration (OE, top) and reciprocal social interaction (RSI, bottom) tests. A set of representative results, including individual neuron activity and mean activity for each neuronal ensemble, is shown. (C) The correlation between the mean neuronal ensemble activity and behavior (binary vector) in the OE (top) and RSI (bottom) tests. (D) Correlation of the activity of all recorded neurons and their corresponding ensembles. The cumulative probabilities of the correlation coefficient of WT and Tbr1+/− mice in OE-1, OE-2, RSI-1, and RSI-2 are shown. (E, F) Activity correlations between dual-recorded neurons and their corresponding neural ensembles. The data points represent the individual neurons. (E) Object exploration, n = 27 from four WT mice; n = 12 from four Tbr1+/− mice. (F) RSI, n = 49 from four WT mice; n = 16 from four Tbr1+/− mice. Data in (C) is mean ± SEM. Statistical analysis in (C) by two-way ANOVA followed by Bonferroni post hoc test. The Kolmogorov–Smirnov test was used for cumulative probabilities in (D). * P < 0.05, ** P < 0.01, *** P < 0.001. See also S5–S7 Figs. The data underlying the figure can be found in the S3 Data. All statistical analyses and results, including the actual P-values, are summarized in S4 Data. The figure was created in BioRender. Lin, M. (2025) https://BioRender.com/cyi56kh.
Fig 5.
Populations and responses of sociality-linked neurons in the BLA are influenced by Tbr1 deficiency and cocktail supplementation.
(A) Definition of the behavior-related neurons. (B) Representative in vivo calcium imaging traces of behavior-related positively correlated (top), negatively correlated (middle), and irrelevant (bottom) neurons. Left, neuronal activity during the entire session. Right, comparison between the raw cosine similarity (R) and null distribution of shuffled data. (C) Ratio of object exploration- and sociality-linked neurons (n = 300 from four WT mice; n = 242 from four Tbr1+/− mice). (D) Mean activity of object exploration- and sociality-linked neurons during behavior tests. Only positively correlated neurons were analyzed. (E) Connectrograms to reveal repetitively recorded neurons across the four behavioral sessions. (F) Overlap between the water and cocktail experimental groups in terms of object exploration- and sociality-linked neurons. Data in (D) represents mean ± SEM. * P < 0.05, ** P < 0.01; two-way ANOVA with Bonferroni post hoc test. The data underlying the figure can be found in the S3 Data. All statistical analyses and results, including the actual P-values, are summarized in S4 Data.
Fig 6.
Cocktail supplementation increases the diversity of activation patterns in the sociality-linked neurons of Tbr1+/− mice.
(A) Representative neuronal activity patterns of RSI-positive neurons in WT and Tbr1+/− mice. (B) Functional networks of WT and Tbr1+/− mice in RSI. Links (lines) between the nodes represent significant similarity in activation patterns relative to shuffled data. (C) Degree centrality of neurons in WT and Tbr1+/− mice across the RSI-1 and RSI-2 sessions. (D) Functional networks of WT and Tbr1+/− mice in OE. Links (lines) between the nodes represent significant similarity in activation patterns relative to shuffled data. (E) Degree centrality of neurons in WT and Tbr1+/− mice across the OE-1 and OE-2 sessions. The results of WT mouse #3 and Tbr1+/− mouse #4 are presented as examples. Data in (C) and (E) are presented as means ± SEM. The data points of individual neurons are also shown. Two-way ANOVA with Bonferroni post hoc test was used for statistical analysis. ** P < 0.01; *** P < 0.001; **** P < 0.0001. See also S8 and S9 Figs. The data underlying the figure can be found in the S3 Data. All statistical analyses and results, including the actual P-values, are summarized in S4 Data.
Fig 7.
Mixing low-dose nutrients synergistically improves social behaviors of autism mouse models.
(A, C) Experimental design of dietary supplementation and behavior assays. RSI, reciprocal social interaction. (B) Synergistic effects of the 1/4 cocktail on RSI of WT and Tbr1+/− mice. Individual nutrients, including BCAA (0.45%), serine (1%), and Zn2+ (20 ppm), and the 1/4 cocktail were sequentially applied to mice for 1 week for each supplementation assay. Five RSI tests were conducted as indicated to investigate the effect of supplements on RSI. Left: the response of WT mice in each test; Middle: the response of Tbr1+/− mice in each test; Right: the comparison of WT and Tbr1+/− mice in each test. (D) Synergistic effects of the 1/8 cocktail on Cttnbp2+/M120I mice. Different from (A) and (B), each Cttnbp2+/M120I mouse was only subjected to two RSI tests: one is water control; the other is supplement treatment. Data are presented as data points for individual mice and/or means ± SEM. The numbers (n) of examined mice for each group are indicated. (B) Left and Middle: paired t test was performed to examine the difference between water control and other experimental groups, individually. Right: Two-way ANOVA with Bonferroni post hoc test. (D) Paired t test. * P < 0.05, ** P < 0.01, ns = non-significant. The data underlying the figure can be found in the S3 Data. All statistical analyses and results, including the actual P-values, are summarized in S4 Data.
Fig 8.
Cocktail supplementation improves the social behaviors of three autism mouse models.
(A, H) Experimental design of cocktail supplementation and behavior assays. OF, open field; EPM, elevated plus maze; 3C, three chamber test; RSI, reciprocal social interaction; CFC, cued fear conditioning. (B-G) The effect of long-term 1/4 cocktail treatment on Tbr1+/− mice. (B) Body weight. (C) Open field. (D) Elevated plus maze. (E) Cued fear conditioning. (F) Reciprocal social interaction. (G) Three-chamber test. (I) The social behaviors of NF1+/−, Cttnbp2+/M120I, and Tbr1+/− mice in the RSI assay were improved by 1/2 cocktail treatment. Data are presented as data points for individual mice and/or means ± SEM. The numbers (n) of examined mice for each group are indicated. (B, I) Paired t test. (C–G) Two-way ANOVA with Bonferroni post hoc test. * P < 0.05, ** P < 0.01, ns = non-significant. Note there is no significant difference in (B–D), though no labeling is indicated. The data underlying the the figure can be found in the S3 Data. All statistical analyses and results, including the actual P-values, are summarized in S4 Data.