NEW YORK (GenomeWeb) – Understanding neuronal connections in the brain has relied on either imaging methods that look at neuron transmissions in bulk or labor-intensive techniques that look at one single neuron at a time. But now, researchers from Cold Spring Harbor Laboratory in New York have developed a sequencing-based method that lets them track in a high-throughput manner the flow of information from single neurons throughout the brain.
The researchers described the method, which they call MAP-seq, for multiplexed analysis of projections by sequencing, last week in a study published in the journal Neuron.
The technique could have applications in understanding the structures and functions of the brain, as well as in understanding diseases such as autism and schizophrenia, and in helping to screen for drug targets, according to lead author Justus Kebschull, a graduate student in Anthony Zador's laboratory at Cold Spring Harbor.
There are lots of methods to look at brain activity, but "we are lagging behind in figuring out the wiring of the brain," Kebschull told GenomeWeb. The "millions of individual cells are all different with different properties, and they send information to different corners of the brain," he said.
So far, the methods to map that flow of information have been very general. For example, he said, imaging studies can enable researchers to see, for example, that one brain region sends information to three other regions, but, what's missing are the details. For instance, whether that region consists of one population of cells sending information to all three downstream regions, or whether separate populations each send information to one of the three regions.
Kebschull said the MAP-seq method is able to get down to that detail. Instead of relying on imaging, MAP-seq make use of genetic barcodes to label individual cells. With 30-nucleotide barcodes, more than 1018 unique barcodes can be generated, allowing individual cells to each have their own barcode.
Then, the researchers use a viral delivery system to label the cells. The barcodes are packaged into deactivated virus particles, which are incorporated into the cells. The barcode includes an mRNA transcript to make sure the virus is expressed inside the cell, as well as an engineered protein that ensures the barcode is transmitted along with the other information that travels across the synapse. "So we have this link between the RNA barcode and the protein that's moving through the synapse," Kebschull said. The barcode and engineered protein don't affect the normal action of the cell.
One key to the method is making sure that one cell is infected with one virus particle and that each virus particle is labeled with a unique barcode. If more than one cell has the same barcode, that is the "same situation of not being able to distinguish individual cells," Kebschull said. "You get the aggregate activity rather than single-cell activity."
To reduce this probability, Kebschull said the researchers generate many more unique barcodes than cells that will be labeled. This reduces the probability that two different cells will have the same barcode.
Next, the amount of virus that is injected is diluted to reduce the probability that more than one virus particle will infect more than one cell. As a result, many cells will not get labeled at all and a few cells will still get two virus particles, Kebschull said, but that is less of a problem than having multiple cells with the same barcode.
Finally, RNA sequencing is performed. The barcodes identified by sequencing can then be traced back to the cell of origin.
In the study, the researchers demonstrated in a proof of principle that the method could track single neuron activity of the locus coeruleus (LC) region in the mouse brain. The team first validated that the barcoded viral particles would indeed infect the neurons and that the labels would not impact neuron activity, using in situ RNA hybridization.
Next, as a proof-of-principle demonstration, they used MAP-seq to determine the patterns of 995 barcodes labeled in four mice. The barcoded virus library was injected into the LC of four different mice. Two days later, the mice's brains were dissected and the researchers performed RNA sequencing on tissue extracted from the olfactory bulb and from 22 different slices from the cortex.
Looking at the patterns showed that "neurons projected in diverse and idiosyncratic ways to specific targets, innervating some areas hundreds of times more strongly than others," the authors wrote. The finding was "in contrast to the simplest prediction from conventional bulk tracing." For instance, the researchers found that some neurons projected only to the olfactory bulb, some only to the cortex, and some to both.
Kebschull said that the researchers are now making some tweaks to the method to improve it and make it more high-throughput, but it is essentially ready to be used. Already, he said, the group is collaborating with other research teams on specific projects, for example, to study different disease models. His lab is applying the method to study autism, since it is thought that brain connections could play a role in that disease.
The technique could also be used to study drugs, he said. By using MAP-seq to barcode potential drug targets and then perturbing those targets, "you can see the effects of your perturbation on the individual targets" and "see the effects on the wiring on thousands of targets at a time," he said.
In addition, he said, the lab is working on setting up a facility at Cold Spring Harbor to offer MAP-seq as a service. Neuroscience labs may not have the sequencing and bioinformatics capabilities, he said, but they likely do have expertise in the viral injection and brain dissection techniques. The goal is to set up a facility where labs can send samples to be processed using MAP-seq, he said. Researchers would pay a fee for the service, but the fees would just cover the costs of the service, and it wouldn't be a profit-generating enterprise, Kebschull said.