Researchers at Cold Spring Harbor Laboratory are proposing to use high-throughput sequencing to draw a wiring diagram of all neurons in the brain at approximately the cost of sequencing a human genome.
In an article published in PLoS Biology last week, the team, led by Tony Zador, a professor of neurobiology at CSLH, outlined their approach, called BOINC for barcoding of individual neuronal connections, in which sequencing serves as the final readout step.
"The appeal of using sequencing is that its scale — sequencing billions of nucleotides
per day is now routine — is a natural match to the complexity of neural circuits," the scientists write. "An inexpensive high-throughput technique for establishing circuit connectivity at single neuron resolution could transform neuroscience research."
While they still need to overcome a number of hurdles to apply their method to an entire brain, the researchers have already conducted proof-of-principle experiments in cultured mouse neurons, which they plan to submit for publication in a peer-reviewed journal in the near future.
According to Zador, BOINC was inspired by the so-called Brainbow technique, a light microscopy approach that visualizes axons and dendrites of individual nerve cells. Each neuron expresses a random collection of fluorescent dyes in different ratios, giving it a distinct color. The combination of fluorophores is determined by an enzymatic recombination system that shuffles them around in the genome of each neuron. Brainbow is limited, though, by the number of colors that can be created and resolved, on the order of several hundred.
Zador said that after hearing a talk about Brainbow, he thought about how it could be expanded to larger numbers of neurons and concluded that "rather than coming up with more complicated ways of coloring neurons, we could just skip the colors and read out the sequence."
The first step of BOINC generates a unique DNA barcode in each neuron. This could be achieved through a transgenic mouse that carries a cassette of different DNA sequences. A recombinase could scramble these in each individual nerve cell, resulting in a unique barcode, similar to the way that T cells obtain their unique receptor through somatic recombination.
In the second step, a neuron's barcode is moved to all its neighbors that directly connect with it through synapses. Currently, the scientists achieve this by using an engineered pseudorabies virus, a herpes virus that has evolved to pass from one neuron to the next across synapses. Following this step, each neuron is a "bag of barcodes," carrying both its own and copies of its neighbors' barcodes.
The third step joins a cell's own barcode pairwise with all the barcodes that came from the outside, using an integrase, such as phiC31.
Finally, the barcode DNA is amplified and turned into a sequencing library, and from the pattern of joint barcodes, the scientists can reconstruct a map of the neuron connections.
To a limited extent, it will also be possible to characterize these neurons, though unlike microscopy-based approaches, nothing would be known about their morphology. For example, barcode DNA from different regions of the brain could be sequenced separately, thus distinguishing neurons from different brain areas. Also, the experiment could be conducted in transgenic mice that have a tag for a specific subclass of neurons, for example inhibitory or excitatory ones. In addition, Zador and his colleagues are working on a way to barcode the transcriptomes of individual neurons, he said.
Zador said they can currently perform all three of the steps involved in BOINC — barcoding the neurons, moving the barcodes across synapses, and joining the barcodes — in neuronal cultures and get some sequencing results, which they are about to write up for publication. He said they can also obtain results from mice, though they haven’t made a transgenic mouse yet so they inject the animals with marked viruses instead.
One challenge is that they cannot currently determine the number of false positive and false negative connections they detect, although Zador believes the number of false positives will be small. An ideal control would be C. elegans, he said, because the complete wiring diagram for the worm's 302 neurons is known and is identical between animals. However, the pseudorabies virus does not work in invertebrates, so Zador and his team are working on a "slightly different technology" to associate barcodes from synaptically connected neurons in worms that he declined to describe in more detail.
The efficiency of BOINC is also not yet known, but "there are questions that can be addressed even if the efficiency is very low," he said. "Of course our final goal is the complete connectivity, the complete wiring diagram" of the brain, though as with the human genome, the first version might only be 90 percent complete.
Another challenge will be to generate a transgenic mouse that genetically encodes all the required pieces, which can be tricky. "Sometimes, it works really well from the beginning, and sometimes it turns out that for mysterious reasons, it doesn't work for the first five or ten times you make it," he said.
In their article, the scientists estimate that sequencing the brain "connectome" of a mouse might cost on the order of $50,000 "and could easily drop several orders of magnitude in a few years," and the cost would be $1 for a Drosophila brain and even less for a C. elegans brain.
Low costs like these, Zador said, might convince those who believe there is not much to be learned from a single-cell wiring diagram to try it anyway. "If and when we figure out how to do it, the cost of actually getting the circuit will be very small, of the order of the cost of a human genome, which is cheap and getting cheaper," he said.
Zador thinks neuronal circuits could help researchers better understand diseases such as autism and schizophrenia, which he said are believed to result in part from a disruption of neural circuitry. While several mouse models of autism are available, for example, it is difficult to phenotype them, and being able to look at the underlying circuit of neurons "would be a lot more compelling."
But until a wiring map of the brain is available, it will be impossible to say how useful it will be. "On the one extreme, it could be that once we see the wiring diagram, we will slap our heads and say, 'Oh, that's how the brain works!'," Zador said. "At the other extreme, we might get the wiring diagram and we shrug and say, 'Uh?' I think it will be somewhere in between."