The metabolic role of vitamin D in children's neurodevelopment: a network study

Abstract

Neurodevelopmental disorders are rapidly increasing in prevalence and have been linked to various environmental risk factors. Mounting evidence suggests a potential role of vitamin D in child neurodevelopment, though the causal mechanisms remain largely unknown. Here, we investigate how vitamin D deficiency affects children's communication development, particularly in relation to Autism Spectrum Disorder (ASD). We do so by developing an integrative network approach that combines metabolomic profiles, clinical traits, and neurodevelopmental data from a pediatric cohort. Our results show that low levels of vitamin D are associated with changes in the metabolic networks of tryptophan, linoleic, and fatty acid metabolism. These changes correlate with distinct ASD-related phenotypes, including delayed communication skills and respiratory dysfunctions. Additionally, our analysis suggests the kynurenine and serotonin sub-pathways may mediate the effect of vitamin D on early life communication development. Altogether, our findings provide metabolome-wide insights into the potential of vitamin D as a therapeutic option for ASD and other communication disorders.

Figure 1

Overall scheme of the analysis. (a) The metabolomics matrix containing the metabolite levels of 381 VDAART children was used as the input to the LIONESS algorithm. (b) Pearson correlation coefficients between each pair of metabolites were calculated to build the aggregate correlation network associated with the metabolomics matrix. Only edges with high correlation coefficients are shown for visualization purposes. (c) LIONESS reconstructed the sample-specific metabolic networks based on Eq. (1). (d) The LIONESS edge weights for each pair of metabolites were regressed against the children’s phenotypic traits. (e) We selected edges that are statistically associated with the interaction term between children’s ASQ-comm scores and either maternal or offspring vitamin D levels. These edges constitute the VDI network.

Figure 2

Enrichment analysis of the VDI network. (a) Normalized Enrichment Score of KEGG pathways enriched in the VDI network based on pre-ranked MSEA. Pathways are ranked based on their p-values. (b) VDI network degree of the MSEA leading metabolites of the top three enriched KEGG metabolic pathways, which include Tryptophan metabolism, Linoleic acid metabolism, and Biosynthesis of unsaturated fatty acids. (c) First neighbors’ network of the leading metabolites for 1) Tryptophan metabolism (top), 2) Linoleic acid metabolism (center), and 3) Steroid hormone biosynthesis (bottom). Node sizes are proportional to the node’s degree. Only edges included in the VDI network are shown. For visualization purposes, we highlighted nodes that were mentioned in the main text.

Figure 3

Children’s phenotypic traits in the clusters extracted from the VDI network. (a) Vitamin D levels (ng/mol) of children in each cluster at year 3, their mothers at 32 to 38 gestation weeks and one year after delivery, and in the cord blood. Colored boxes represent clusters with lower or higher Vitamin D levels compared to the individuals in all other clusters (p-value < 0.15, see “Materials and methods”). (b) Stacked histogram of the communication scores for children in each cluster. Cluster 4 (referred as the low-communication cluster) exhibited the lowest distribution of communication scores. (c) Stacked histogram of asthma incidence in children and mothers within each cluster. Cluster 2 (referred as the asthma cluster) exhibited the highest proportion of children and mothers with asthma.

Figure 4

Marker metabolic edges identified for the low-communication and asthma clusters. Top marker metabolic edges identified for the (a) low-communication and (b) asthma clusters. Node sizes are proportional to the node’s degree in the marker edges’ subnetwork. Edges are colored based on the average LIONESS weight across the networks in the cluster. For visualization purposes, we highlighted nodes and edges that were mentioned in the main text.

Figure 5

Shortest path connecting 5HIAA and L-kynurenine in the KEGG reaction network. The edge connecting 5HIAA and L-kynurenine was statistically associated with the interaction term between vitamin D and the ASQ communication score. One shortest path connected these two metabolites in the KEGG reaction network of tryptophan metabolism. This path involved two main subprocesses: (1) the biosynthesis of serotonin from l-tryptophan (orange nodes) and (2) the degradation of l-tryptophan via the kynurenine pathway (blue nodes). The nodes in this network represent chemical compounds listed in the KEGG tryptophan metabolism. The edges represent the chemical reactions between a substrate and a product. Only the largest connected component is shown.