NEW YORK (GenomeWeb) – Researchers from Harvard University have sequenced the whole genomes of single neurons in a proof-of-concept study to understand the somatic mutation process in brain cells.
Reporting their work today in Science, the researchers sequenced whole genomes of 36 single neurons from the cerebral cortexes of three healthy individuals. They identified thousands of SNVs and were able to piece together the cells' lineage.
The study was a collaboration between Peter Park's computational biology laboratory and Christopher Walsh's neurodevelopmental lab.
In the study, the researchers found that somatic mutations occur in neurons — a question that had previously not been conclusively answered due to the inability of bulk sequencing to identify low-frequency mutations — and that those somatic mutations occur by a different process than mutations in germline or cancer cells. In addition, identifying SNVs at the single-cell level allowed the researchers to trace the cells' lineage, discovering that neurons come from diverse pluripotent founder cells.
Single-cell sequencing provides "valuable information," Park, a senior author of the study and computational genomicist, told GenomeWeb. "In this case, the fact that we could discover somatic mutations at the single-cell level that would not be detected otherwise gave us a lot of new insights about the potential role of these mutations in the brain."
The researchers examined neurons from the brains of three post-mortem individuals aged 15, 17, and 42 years, all of whom had been healthy. They performed multiple displacement amplification and whole-genome sequencing to 40x coverage of 36 total neurons. Next, they used three different algorithms to call SNVs from the data, using only those SNVs found by all three algorithms for further analysis.
On average, the neurons had between 1,458 to 1,560 SNVs, a similar number of variants per cell as other normal human cells, aside from skin cells, which are exposed to UV light.
The researchers found that the profile of the neuronal SNVs was different than either germline or cancer SNVs. Germline and cancer mutations often arise as a result of DNA replication. However, the neuronal SNVs did not correlate with replication, but instead correlated with transcription.
In addition, when the researchers integrated SNV data with gene-expression data from public databases, they found that genes that are highly expressed in the brain were enriched for neuronal SNVs.
"Neural-related gene sets were enriched for somatic SNVs, and single neurons harbored heterozygous coding SNVs in genes that, when a single copy is mutated in the germline, confer a high risk of neurological disease," the authors wrote.
Specifically, they found neurons in one sample with coding mutations in a gene known to cause seizure disorders and one gene associated with schizophrenia.
In addition, the researchers were able to analyze the neurons' lineage by looking at which SNVs were shared among neurons and by how many.
From one sample, the team genotyped shared variants among the sequenced neurons in an additional 210 neurons isolated from the same brain region. About 60 percent of the neurons contained at least one clonal SNV. The data also implied that mutations likely occurred serially throughout development. For instance, one SNV present in 51, or 23 percent, of the neurons suggested it likely occurred early in development, while another SNV was present in three of those 51 neurons and another in two of the three neurons.
In all, by clustering neurons by their shared SNVs, the researchers were able to identify five distinct branches that gave rise to the set of 226 neurons.
"It's like a map of how these neurons have ended up where they are in the brain," Park said. "We can trace the ancestors of these individual neurons, and I think that will give us a better understanding of the brain." For instance, he said, neurons with mutations in disease-related genes like epilepsy or schizophrenia could potentially be traced to see whether they arise from a particular ancestor.
Park said the next step would be to perform these types of single-neuron sequencing studies in larger numbers of individuals and larger numbers of cells. He is involved in projects to study autism in schizophrenia, which will involve both bulk DNA sequencing as well as single-cell sequencing.
Also, sequencing both the DNA and RNA from single cells could help identify more readily the SNVs that are expressed. Park's study found that transcribed genes had a higher rate of mutation, "but that analysis doesn't tell you whether those genes are expressed in the cell," he said.
The researchers were able to infer that the genes that were mutated were the most likely to be expressed, he said, but they couldn't say whether the specific mutated gene was being expressed in that cell — information that would be important for studying relationships between neuronal SNVs and disease.