Competition: How a Target Picks Its Neurons

If every sympathetic neuron needs NGF, and the target makes too little for all of them, which neurons get to live?

Earlier in this unit you met the neurotrophic hypothesis in the form Chris Deppmann teaches it: “Neurons compete for a limiting amount of neurotrophic factor for survival.” Targets such as the eye, the salivary gland and the heart release nerve growth factor (NGF). Sympathetic neurons, which live in small clusters of nerve cells called ganglia, send axons to those targets. Neurons that get enough NGF live; those that do not die through the pro-death gene Bax. See that lesson for the evidence.

The hypothesis says there is a contest. It does not say how the contest is decided. This lesson follows one answer, worked out in the sympathetic neurons of the mouse superior cervical ganglion, a ganglion in the neck that supplies the head. Keep that scope in mind: this is one cell type in one species.

The Competition Model: a limited supply means a contest

The target-size evidence came first, from the chick. In 1976 Margaret Hollyday and Viktor Hamburger grafted an extra limb onto chick embryos, and fewer of the motor neurons supplying that side died. More target, more survivors, which fits competition for a limited resource, though the authors could not tell whether the resource was a trophic factor or synaptic sites.

A limiting supply alone does not pick winners. Suppose NGF were shared evenly. Every neuron would get the same fraction, and they would all live or all die together. Suppose instead the first axon to arrive took everything. The target keeps releasing NGF throughout the death period, and nothing in a first-come rule explains how a neuron already receiving some NGF could fall behind. Something has to turn a small, chance difference into a large one.

That is the job of positive feedback. In the Cell Signaling course you saw that a loop in which a signal raises its own production can make a switch with two stable states (see that lesson). Here is a caricature of the idea:

change in TrkA = basal synthesis + k × (NGF signal) − decay × TrkA, where NGF signal rises with both NGF and TrkA

If TrkA, the receptor tyrosine kinase that binds NGF (see that lesson), rises with its own signal, a neuron with a little more TrkA collects more NGF, makes more TrkA, and pulls further ahead.

How is the developing circuit tuned? NGF raises its own sensitivity

To find what NGF does to a neuron in a living animal, Deppmann, Ginty and colleagues needed ganglia that lack NGF yet still have neurons. Without NGF the neurons die, so they used mice lacking Bax. In a Bax-null mouse, sympathetic neurons survive with or without NGF. Comparing ganglia from Bax-null mice with ganglia from mice lacking both Bax and NGF gave a list of “NGF Dependent Genes”: transcription factors, synaptic genes, genes for metabolism and cell size, and genes for axon growth.

One gene on the list was the receptor itself. “NGF is Necessary and Sufficient for TrkA Expression”: in newborn ganglia and in dissociated cultures, TrkA fell without NGF and rose with it. Neurons deprived of NGF also lowered TrkA over 12 to 48 hours and signalled less efficiently, measured by phosphorylated MAPK and AKT, two kinases downstream of TrkA.

The lab called this “NGF Transcriptional Regulation of NGF Sensitivity” and described four stages:

1. Naive. A young neuron responds poorly to NGF.
2. Sensitization. A neuron that receives enough NGF crosses a threshold.
3. Competitive Advantage. It raises TrkA and the proteins downstream of it.
4. Selection. It wins the supply, and neighbours that stayed behind lose.

NGF also controls how long its own signal lasts. With the computational neuroscientists Stefan Mihalas and Ernst Niebur, the group simulated neurons competing for fixed NGF under three rules: no feedback, feedback on TrkA level, and feedback on signal duration. Feedback separated the neurons into a winning and a losing group.

How does NGF lengthen its signal? Much NGF signaling travels from axon tip to cell body inside endosomes (the subject of that lesson). In 2014 Deppmann’s laboratory, with Dong Suo as first author, reported that NGF induces Coronin-1, a protein that associates with these signaling endosomes, prolongs their signaling, and is required for normal competition in the mouse. It is one mechanism; there may be others.

Punishment Signaling: winners push losers

Feedback explains how a winner pulls ahead. Does it need help? In his lectures Deppmann lists the qualities a punishment signal should have. It should come from strongly supported neurons, kill weak ones, spare strong ones, and start only after TrkA has risen.

The candidates were already on the list of NGF-dependent genes. Strongly supported sympathetic neurons make BDNF and NT-4, two other neurotrophins. Sympathetic neurons carry little of their receptor TrkB, but they do carry p75, a member of the tumour necrosis factor receptor family (see that lesson and the course’s lesson on the receptors). In this setting, p75 can promote death. A neuron rich in TrkA is protected; a neuron poor in TrkA is not.

The lab tested three predictions in mouse superior cervical ganglia:

1. Competing neurons differ in TrkA level (tested in Bax-null ganglia).
2. Neurons with low TrkA show high p75 death signaling, read out by phosphorylated c-Jun.
3. Removing punishment delays but does not prevent the loss (tested in p75-null mice).

All three held, and the summary slide reads “A Model for Neuronal Competition.” An earlier study from the laboratory of Freda Miller had also reported delayed sympathetic neuron death in p75-null mice.

“What is the benefit of punishment to the system?” In the simulations, punishment made the split faster: it hastened the killing of weak neurons more than ten-fold. Without it, weak neurons lingered.

Design problem

If punishment through p75 is part of competition, what happens to sympathetic neuron numbers over the death period in a p75-null mouse? Sketch neuron number against age for wild-type and p75-null ganglia. How could you test this?

Activity Modulates Trophic Support

Now a result that seemed to point the other way. In 1979 Robert Pittman and Ronald Oppenheim treated chick embryos with curare, a drug that blocks the acetylcholine receptors of muscle, during the period of motor neuron death. The muscles stopped responding to their nerves, and fewer motor neurons died.

Why was that a contradiction? If activity helps an axon win, as it does in many later forms of refinement, then silencing the connection should increase death. Instead it saved neurons.

The usual explanations are interpretations, so hold them loosely. Paralysed muscle may allow more nerve branching and more synapses, giving each motor neuron more access to trophic support. Inactive muscle may also make more of the factors motor neurons need. Note that chick motor neurons do not depend on NGF. They draw on several factors at once, which Oppenheim called “an embarrassment of riches.”

In other systems, activity helps a winner. Active synapses can release more BDNF, which layers activity-dependent refinement on top of trophic competition. In mouse sympathetic neurons, axon pruning uses the same BDNF and p75 punishment logic you met above (see that lesson). The two kinds of result need not conflict: activity in the muscle and activity in the competing neurons are different variables.

Systems matching: NGF also sets how much a target gets

“How is the developing circuit tuned?” Survival is one answer. Deppmann calls the broader task “Systems matching”: each target gets a fitting number of neurons, amount of innervation, synapses and dendrite shapes. NGF acts on several of these. In mice lacking both Bax and NGF, neurons survive, but sympathetic axons fail to fill some targets at birth, such as the salivary glands and, partly, the heart: “NGF Controls Final Target Innervation.” In his teaching he puts it informally: losing axons “get more aggressive as they find a pot of NGF.” So the same factor that picks which neurons live also shapes how densely the survivors cover their target.