Long-Distance Signaling in Neurons

A sympathetic neuron collects its survival signal at the tip of an axon that can be a thousand times longer than its cell body is wide, and the genes that must answer are back in the nucleus. You met the signal, nerve growth factor (NGF), in the lesson on finding a signal, and its receptor TrkA in the lesson on receptor tyrosine kinases. Here the question is how a message received at one end of a cell reaches the other end, and how you would know which way it went.

A Problem: the signal and the genes are far apart

Sympathetic neurons are the nerve cells that control organs such as the heart and blood vessels. The cell body of one, where the nucleus is, measures a few hundredths of a millimeter across. Its axon, the single long fiber that carries its output, runs out to a target organ, which makes NGF, and the axon’s terminals collect it there. Those terminals can be centimeters from the cell body.

What does the cell body need to learn? That its axon has reached a target that makes NGF. It answers with gene expression. The neuron stays alive, and it switches on genes that include the one for TrkA itself and the ones for the tubulin and actin a growing axon is built from.

How does this happen? Rule out the easy answer first. A molecule that spreads by diffusion is slow over long distances, because the time grows with the square of the distance.

time to diffuse a distance x ≈ x² / 2D

Take D as about 10 µm² per second, a typical figure for a protein in cytoplasm. For x = 1 cm, which is 10,000 µm, the time is 5 million seconds, or about 58 days. The survival signal arrives far sooner than that. Something has to be carried, or passed along.

The evidence: NGF at the far end is enough

In an ordinary culture dish one fluid bathes the whole neuron, so you cannot ask which part of the cell needs NGF. In 1977 Robert Campenot published a dish that could ask. A barrier sealed to the floor of the dish divides it into chambers. The cell bodies sit in one chamber, their axons grow under the barrier into the next, and the fluid in each chamber can be changed separately.

Two runs made the case. When NGF was removed from the far chamber, the axons there stopped growing and later appeared to degenerate, even though the cell bodies behind them were in NGF the whole time. When NGF was removed from the cell bodies instead, many neurons survived, as long as their axons reached a chamber that still had it.

So the part of the neuron that needs NGF is the part at the target. And whatever kept those cell bodies alive had to reach them through the axon, whether carried or relayed, because the barrier kept the fluids apart. This is retrograde signaling: signaling from the axon terminal back to the cell body.

Where does the signal end up? In David Ginty’s laboratory, Antonella Riccio and colleagues applied NGF only to the distal axon, its far end. That activated CREB in the nucleus, the transcription factor you met in the lesson on signaling to the nucleus.

Be exact about what this shows. It shows where the signal ends. On its own it does not show how the signal traveled. In the same paper Riccio and colleagues reported that taking in NGF and TrkA and carrying them to the cell body were required for the signal.

The model: the endosome as a parcel

The leading model is the signaling endosome. NGF binds TrkA at the axon tip, and the pair is taken into the cell by endocytosis, the subject of the previous lesson. Look at how the receptor now sits. NGF is inside the vesicle, still held by the part of TrkA that used to face outward. The kinase part of TrkA is on the other side of the vesicle’s membrane, in the cytoplasm, where it can go on phosphorylating its partners. The vesicle is a patch of active membrane that can be shipped.

Endosomes need motors and roads. The roads are microtubules, long protein filaments whose two ends differ and are called plus and minus. In an axon they all lie the same way round, with their plus ends toward the tip. The kinesins that haul axonal cargo walk toward plus ends, so they carry it out, away from the cell body. Cytoplasmic dynein walks toward minus ends, so it carries cargo back, with the help of a partner complex called dynactin.

A motor-driven vesicle moves at roughly a micrometer per second. At that pace, the centimeter that would take diffusion 58 days takes a motor under three hours.

In this model the parcel reaches the cell body still signaling, and kinases there pass the message on to transcription factors such as CREB.

Not everything is sent home. In Campenot’s dish, axons grew only where NGF touched them, whatever the cell bodies were sitting in. Local mediators? Some of NGF’s work is done where it binds, by signaling proteins beside the receptor at the tip. The survival decision is the part that needs the nucleus.

Carried, or relayed? Why blocking the motor does not settle it

There is a second way to move a message down an axon. Think of a courier against a bucket brigade. The courier carries one parcel the whole way. In the brigade nobody walks far: each stretch of the axon hands the message to the next stretch, and nothing physical makes the whole journey.

The popular test is to block dynein. In 2004 Heerssen, Pazyra and Segal found that dynein motors transport activated Trk receptors, and that this transport is needed for the survival of neurons that depend on their target. That is real evidence for a courier.

A brigade would fail the same test, though. Suppose each stretch of the axon needs dynein-driven transport for its own local step. Then blocking dynein stops the brigade as completely as it stops the courier.

The relay is a published alternative. In 2002 MacInnis and Campenot reported in Science that retrograde signals can support neuron survival without retrograde transport of NGF itself. In 2003 Ye, Kuruvilla, Zweifel and Ginty titled their paper “Evidence in support of signaling endosome-based retrograde survival of sympathetic neurons”. The title claims support. It does not claim proof.

Here is the score. Settled: NGF at the terminal is enough to support the neuron, a signal from the terminal reaches the nucleus, and dynein carries activated receptors that survival depends on. Not settled: whether the carried receptor is the only route, and whether one kind of parcel does every job NGF asks for.

Design problem

Two models explain how a survival signal gets from the axon tip to the nucleus. In the first, the receptor itself travels the whole way inside a vesicle. In the second, the receptor stays at the tip and a chain of kinases passes the message along. Design an experiment that could tell them apart. Say first why a drug that stops the dynein motor cannot do it.

Two more answers to distance, and a road that fails

See where the endosome ends up

A sympathetic neuron is also a listener. Neurons in the spinal cord make synapses, the contacts where one neuron signals to the next, on its dendrites, the short branches that collect input. In 2010 Sharma, Deppmann, Harrington and colleagues asked whether NGF from the distant target controls those synapses too. They studied mice that lack NGF and also lack BAX, the death protein from the lesson on apoptosis, so that the neurons live without their survival signal and their synapses can be examined. NGF turned out to be needed for those synapses to form.

How does NGF-TrkA control this? The group grew neurons in a microfluidic chamber, a small relative of Campenot’s dish developed by Noo Li Jeon, using TrkA that carries a tag called Flag. After NGF was given to the axons, tagged TrkA appeared as dots in the dendrites over the next six hours, and the group compared that with the clustering there of postsynaptic proteins, the ones on the receiving side of a synapse. A long-distance relationship? The message that the axon has found its target reaches the side of the cell that receives input.

Make it where you need it

A dendrite has its own distance problem. A synapse can sit a long way from the nucleus, and some changes at a synapse need new protein soon after the signal arrives. Erin Schuman found that the neuron does not ship everything. With Hyejin Kang she applied the growth factor BDNF to slices of hippocampus, a region of the brain, and synapses strengthened. Then they cut the slices so that the synapses were separated from the cell bodies of their own neurons. BDNF still strengthened them, and the effect still depended on protein synthesis. They published that in 1996. In 2001 her laboratory watched a fluorescent reporter protein appear in dendrites. In 2012 the laboratory counted 2,550 different mRNAs in the dendrites and axons of the neuropil, the mesh of fibers where the synapses lie.

When the road fails

A cell built this way depends on its motors. Mutations in DCTN1, the gene for a subunit of dynactin, are found in people whose motor neurons, the neurons that drive muscle, degenerate. Two mouse mutants, Loa and Cra1, carry point mutations in the dynein heavy chain, and their motor neurons degenerate. Transport defects have been reported in ALS, a disease of motor neurons, and in Huntington’s, Parkinson’s and Alzheimer’s diseases, although a defect that accompanies a disease is not proof that it causes it. One case names the cargo. In a mouse model of Down’s syndrome that carries an extra copy of App, the gene for the amyloid precursor protein, Salehi and colleagues found that the raised expression of App disrupts the transport of NGF.