NEW YORK (GenomeWeb) – A Stanford University School of Medicine-led team has identified a set of 136 genes whose expression seems to coordinate activity across various brain regions while it is at rest.
Imaging studies, the researchers noted, have indicated that distinct brain regions work together to form networks that synchronize resting brain activity, but the molecular mechanisms behind this functional connectivity have been unclear.
As the Stanford-led researchers reported in Science today, they combined such imaging data with microarray data from the Allen Institute for Brain Science and genome-wide SNP data from the IMAGEN consortium to explore the molecular mechanisms of these networks.
"There's been some skepticism regarding the validity of resting-state network activity," senior author Michael Greicius, an associate professor of neurology and neurological sciences at Stanford, said in a statement. "We wanted to dig deeper and get to the molecular underpinnings of these imaging results, which indicated that the brain maintains its exquisite functional-network architecture even at rest."
Based on resting state functional MRI data from 15 healthy people, Greicius and his colleagues computed 14 known and reproducible functional networks using an independent component analysis. They mapped microarray data from six subjects — encompassing more than 3,700 brain samples — from the Allen dataset to those networks, though focused their analyses on four large, well-characterized, non-overlapping networks: the dorsal default-mode, salience, sensorimotor, and visuospatial networks.
To define linked gene expression networks, the researchers relied on the transcriptional similarities of the expression profiles between the various tissue samples. They defined the nodes of the network by the brain tissue samples and the edges were weighted by similarities between vectors of gene expression values at each sample, the researchers said.
They examined the expression similarity of nearly 17,000 genes.
Using an algorithm, they found that the gene expression correlations in the functionally grouped regions were higher than expected by chance.
By calculating the strength of each gene's influence on the four networks, they developed a consensus list of 136 genes that seem to underlie these functional networks.
To validate this list, the researchers turned to the genome-wide SNP and resting state fMRIs of 259 people from the IMAGEN database. Based on this, they also concluded that polymorphisms in the consensus genes were related to the strength of the functional networks.
Additionally, the researchers reported that 57 mouse orthologs for those 136 genes also exhibited a higher correlation between transcriptional similarity and axonal connectivity than expected by chance.
This work, the researchers noted, will inform future studies of gene sets linked to functional connectivity, including how neurodegeneration spreads through a network.
Greicius and his colleagues noted that this list of 136 genes was enriched for gene ontology terms linked to ion transport, especially sodium channels, and were linked to nine neuropsychiatric disorders, including Alzheimer's and schizophrenia. They further validated these associations using the IMAGEN dataset.
"Our work holds potential implications for a number of neuropsychiatric disorders," Jonas Richiardi, a former postdoc in the Greicius lab who is now at the University of Geneva, said in a statement.
In particular, the researchers noted that Alzheimer's spreads from one brain region to the next through the default-mode network, and focusing on genes in this network could inform studies of the disease.