A new guidebook to the brain

Anyone navigating a new city needs a guidebook. What’s inside the enormous brick building on First and Main? Ask the guidebook. Where’s the natural history museum? The guidebook will get you there.

A guidebook is also invaluable in navigating the brain. Intrepid brain explorers need to know how the different areas connect, as well as functions of the different neurons within each area. But the Lonely Planet: Brain Edition has long been incomplete. Neuroscientists used to have to choose between mapping, or tracing, neuronal circuits, or peering at the molecules inside single cells removed from their circuits. It was like they were traversing a new city with a roadmap in one hand and a list of landmarks in the other.

Now neuroscientists can combine this information to create the guidebooks they need. In work published recently in the Proceedings of the National Academy of Sciences, researchers at Fred Hutchinson Cancer Research Center describe a new technique they dubbed Connect-seq, which makes it possible to map neural connections while also gathering information about the key signals sent by individual neurons within a circuit. The approach integrates two technologies previously used by researchers, each of which provides incomplete information about the brain’s circuitry.

Using Connect-seq in mice, the team traced neurons directly upstream of those that orchestrate the body’s physiological stress response. The work gives insight into individual neuron function and, potentially, the crosstalk between the stress neurons and other biological functions, including reproduction and metabolism.

“The big question that we were interested in was, what are the neural circuits that control the brain’s different functions? And can we define neurons that are interconnected based on the genes that they express [turn on]?” said senior author Dr. Linda Buck, who received the 2004 Nobel Prize in physiology or medicine for her work on the olfactory system.

The answers to these questions will help scientists better understand not only which individual neurons control a specific bodily function, but also how. In the case of the stress response, insights like these could be a step toward developing medical interventions that reduce the negative effects of stressful situations, the researchers said.

“This method is the first of its kind to build that entire molecular map,” said Dr. Naresh Hanchate, the postdoctoral fellow in Buck’s lab who spearheaded the work with fellow postdoc Dr. Eun Jeong Lee. “Each brain area can have many different neuron types. If we identify molecular markers of the neurons involved in stress, then we have a way to map their brain locations. We tested just a handful in this paper.”

Missing information

“The mammalian brain is highly, highly complex,” Hanchate said. “It has millions to billions of neurons that are highly interconnected with each other, to control specific functions. … A handful of them with known functions have been genetically defined, but they are connected to many different neurons in different brain areas, so they are receiving different kinds of signals.”

The molecules within a neuron dictate its function: which signals it sends to other neurons and which signals it’s able to receive. To better understand how the brain works, neuroscientists need to know both what molecules a neuron contains and the neuron’s partners within a circuit.

“To some extent, we know that certain brain areas are involved in certain functions,” Hanchate said, pointing to the hippocampus’s role in memory as an example. “If we know the brain areas, then we can start to dig into their functional roles. But then in each brain area there are many, many different neuronal types.”

Information travels in one direction through a neural circuit, and each neural cell has two ends: one end that receives information, the other that transmits it. The space between the upstream transmitting neuron and the downstream receiving neuron is known as the synapse. Neurons communicate across this synapse via a variety of signaling molecules. Fast communication occurs primarily via two neurotransmitters, glutamate and GABA, which have opposing functions. Glutamate quickly excites downstream neurons. GABA quickly acts to inhibit downstream neuron activity.

A different and more varied class of signaling molecules, called neuropeptides, add a little zest to the two flavors of neurotransmitter. Neuropeptides work more slowly to tune the properties and responsiveness of downstream neurons.

It was information about these kinds of signaling molecules that was missing from previous anatomical maps, Hanchate said.

Buck’s lab focuses, in part, on brain responses to stress, including the stress experienced by prey animals who smell a predator. But stress comes in many forms: there’s emotional stress, as felt by a socially isolated person. There’s physical stress, experienced by someone with a painful broken leg. Whatever the stressor, stress signals get processed by a cluster of neurons in the hypothalamus, a region deep inside the brain that links the nervous system with the endocrine system.

This cluster of neurons are known as CRHNs, after the corticotropin-releasing hormone, or CRH, that they release. CRH is a key mediator of the physical stress response. The team developed Connect-seq to construct a molecular map of the neurons directly upstream of the CRHNs.

Connect-seq: molecular circuitry

To create an anatomical map that includes molecular information about the component neurons, Hanchate combined two technologies that scientists already use.

The first technology is based off engineered viruses that infect neurons and then hop their way up the neural circuit (opposite the direction that information flows from neuron to neuron). Using these viruses, scientists can trace which neurons are connected upstream — and are therefore sending signals — to the originally infected neuron, like a tree branching from its trunk. This builds the neural road map.

The second technology, called single-cell RNA sequencing, or RNA-seq, allows researchers to determine which genes each neuron turns on, or expresses. This is the list of landmarks.

Researchers, including those working at Seattle's Allen Institute for Brain Science, have applied single-cell RNA-seq to collect a census of the individual neuronal types in the brain, Buck said.

Each technology provides just a partial picture to a neuroscientist trying to navigate the brain’s complexities. Virus tracing doesn’t tell you what’s going on inside each neuron within a specific circuit, though you may know the general role of the brain area in which the neurons are located. And single-cell sequencing doesn’t tell you to which circuits an individual neuron connect.

Previous work in Buck’s lab had shown that 31 different brain areas connect to CRHNs.

“What kinds of stressor information are those neurons sending to CRHNs? We knew where they were, but we didn't have any way to identify them,” Buck said. “If you could find genes expressed in those neurons, then you would have a handle that would allow you to figure out what those neurons are doing and what kinds of responses they're controlling.”

So Buck and Hanchate developed Connect-seq to create molecular maps that contained information about neurotransmitters and neuropeptides in each neuronal branch.

Hanchate and Lee did this by using viral tracing to identify neurons upstream of CRHNs and then following it up with single-cell RNA-seq to add information about individual neurons. First, they infected CRHN stress neurons with viruses that made them glow. After the viruses hopped across the synapse to upstream neurons, these neurons also glowed, which allowed the researchers to isolate them. They then used single-cell RNA-seq to see the signaling molecules within individual upstream neurons.

Signaling molecules on an anatomical map

Connect-seq made it possible for Hanchate and Lee to build an anatomical map complete with information about individual neurons’ complements of signaling molecules. And there were some surprises lurking in the data.

For one thing, Hanchate and Lee found that individual neurons produce many more different signaling molecules than they and Buck had expected.

“Traditionally, it was thought one neuron, one signaling molecule,” Hanchate said. “It was thought that a neuron has either glutamate or GABA, and maybe one or two neuropeptides. But more and more studies now provide evidence suggesting that a neuron can release two or more chemical messengers.”

The Connect-seq data confirmed these reports and showed that while many neurons produce either glutamate or GABA, a sizable minority produce both. And Hanchate also found that, rather than add just one or two neuropeptides in addition to a neurotransmitter, neurons upstream of CHRNs can also produce an unexpected variety of neuropeptides: 43 in total.

While some were only found alongside glutamate, and others only with GABA, some neuropeptides accompanied either or both neurotransmitters — another surprise. Buck’s team also reported for the first time that a given neuron can produce as many as eight neuropeptides. Hanchate confirmed these results in a publicly available single-cell RNA-seq dataset released by others. And, working with Andria Ellis, a graduate student in the lab of Dr. Cole Trapnell at the University of Washington, the team found that within different neurons, neuropeptides can be expressed in various combinations and at different levels, adding another layer of complexity to the results.

The different neuropeptides a neuron produces, as well as its location, give hints to its function, Hanchate said. He and Lee were able to trace CRHN-connected neurons that produce specific neuropeptides to several different areas of the hypothalamus. For example, they found one neuropeptide only within neurons in an area of the hypothalamus that regulates the feeling of satiety, or having eaten enough.

“When these neurons are activated, they suppress food intake and increase metabolism,” Hanchate said. “That is intriguing, because what are neurons linked to appetite to do with stress?”

Another type of neuron that Hanchate found communicates with CRHNs controls reproduction. It will take future studies to determine why neurons that control sexual behavior or appetite may be activating or inhibiting stress neurons, he said.

Putting the guidebook to work

Though the researchers focused on just one function of the brain, scientists could use the approach to build guidebooks for any brain function, Hanchate said. And there’s yet more information they could use to fill out the picture. The current analysis focused on the molecules that neurons use to send signals — but not on those that they use to receive them. Hanchate plans to examine these molecules next.

Connect-seq will also help Buck, Hanchate, and Lee better understand how the brain processes stress. Now that they better know some of the brain areas that communicate with stress neurons, they can ask which neuronal types in each area respond to specific stressors, he said. The team is working to develop methods to detect which neurons are activated in response to a given stress signal, as well as take another step upstream and study the neurons two synapses away from CRHNs.

Far off in the future is the potential application of these findings to human health. A deeper understanding of how a specific molecule mediates responses to a stressor could be a step toward developing drug interventions that help blunt the negative effects of stress, Hanchate and Buck said.

“Many of the brain areas and neuronal types are conserved from mouse to humans. So the hope is that if we can learn about neural circuits that control basic functions — not higher cognitive functions — in the mouse, that data can be applied to humans, for example, to identify drug targets,” she said.