SAN FRANCISCO (GenomeWeb) – Understanding brain circuitry is critical for understanding what goes wrong in disorders such as autism and schizophrenia, but current techniques are either too laborious and costly to do at large scale or do not provide resolution at the individual neuron level.
Now, under a three-year, $2.7 million grant from the National Institutes of Health, researchers from Cold Spring Harbor Laboratory in New York are developing a technique that enables them to sequence and map individual neurons in the brain. The team, led by Anthony Zador, a professor of neuroscience at CSHL, plans to both scale up the technique and develop a modified version to enable in situ analysis of the neuronal transmissions.
Zador said the method could have implications for understanding basic biology and connectivity of the brain as well as how neural transmissions may be disrupted in disorders such as autism and schizophrenia.
In a proof-of-principle study published last year in Neuron, Zador's team demonstrated that the method, called multiplexed analysis of projections by sequencing (MAP-seq), could track single neuron activity in the mouse brain. MAP-seq relies on being able to create genetic barcodes. In that study the researchers made more than 1018 unique barcodes. A viral delivery system is then used to label the cells.
"The main thing that we're going to be doing is to scale it up so that we can apply it simultaneously to an entire mouse brain in a single experiment," Zador said. That would essentially enable them to figure out where specific neurons are and where they send their long-range connections.
As part of the Allen Brain Atlas, a 10-year project launched in 2012 to better understand the neural code, the Allen Institute created several programs, including the Allen Mouse Brain Connectivity Atlas, a high-resolution map of neural connections in the mouse brain. That atlas was generated using many different transgenic mice. Imaging techniques are used to look at individual neurons or specific cell types, one at a time. But Zador said that the goal with MAP-seq is to be able to study neural projections in a more high-throughput manner using barcodes that enable many different neurons to be analyzed simultaneously in one mouse.
To scale up MAP-seq, the researchers will first need to create more unique barcodes. Currently, Zador said, they can make between 10 million and 20 million unique 30-nucleotide barcodes. But, to map neuron projections across the entire cortex, they will need around 1 billion unique barcodes. "That's doable, it just takes elbow grease," Zador said.
The researchers will also be optimizing each step, figuring out the best balance between efficiency and accuracy.
Another key piece will be to ensure that cells aren't double labeled with different barcodes. Overcoming this is primarily a matter of generating many more barcodes than needed.
However, Zador said that the main limitation of the MAP-seq method is its spatial resolution. Currently, spatial resolution is determined by the precision with which they can dissect the brain, currently around 300-micron slices about 1 millimeter across.
To increase spatial resolution, his team has collaborated with researchers from George Church's lab at Harvard University to combine the MAP-seq protocol with an in situ sequencing technique Church's lab developed.
The group described the new protocol, dubbed BaristaSeq, in a preprint posted to the BioRxiv server last month. Zador said the goal was to take advantage of in situ sequencing's ability to preserve cellular location and combine it with his lab's highly multiplexed barcode strategy. However, in situ sequencing has traditionally not been good at distinguishing barcodes, or at least the billions of barcodes that the Cold Spring Harbor team aims to use.
Several in situ RNA sequencing methods exist, including one developed by researchers from George Church's lab that is now being commercialized by the spinout ReadCoor. Another approach, previously described in Nature Methods, was developed by researchers from the Science for Life Laboratory based at Stockholm University and Uppsala University. That method involves hybridizing padlock probes to the targeted sequence, circularizing the DNA, amplifying it with rolling circle amplification, and then using sequencing-by-ligation to analyze four-base barcoded tags with fluorescently labeled nucleotides.
In the BioRxiv paper, the researchers describe a version of this padlock probe-based approach, but include modifications to increase the sensitivity and enable the sequencing of longer barcodes. The researchers made use of a different polymerase to increase sensitivity and made modifications to enable Illumina sequencing chemistry to be used, rather than sequencing-by-ligation.
BaristaSeq combines the advantages of Zador's team barcoded, high-throughput sequencing method with the spatial resolution of in situ sequencing. The method "allows us to achieve cellular and subcellular spatial resolution," Zador said, "so we can see where each neuron is that we've labeled." That can be combined with the MAP-seq approach so that the neurons' long-range projections can also be identified.
The ultimate goal, he said, is to use these techniques to better understand the neuromechanisms underlying the sensory processes and decision-making. "I'd like to know where every neuron sends its projections and what the targets are for each neuron," he said. These techniques should help enable that, he said. "Knowing the underlying circuitry will make it easier to design and interpret experiments about function," he said. "It's hard to understand function without enough information about structure."