The staggering implications of one axon synapsing on another

It isn’t often that a single paper can change the way we think the brain works.  But such is the case for the paper described in the previous post (full copy below *** ) if the implications I draw from it are correct.

Unfortunately this post requires a deep dive into neuroanatomy, neurophysiology, neuropharmacology and cellular molecular biology.  I hope to put in enough background to make some of it comprehensible, but it is really written for the cognoscenti in these fields.

I’m pretty sure that some of these thoughts are both original and unique

Briefly, the paper provided excellent evidence for one axon causing another to fire an impulse (an action potential).   The fireror was from a neuron using acetyl choline as a neurotransmitter, and the fireree was a dopamine axon going to the striatum.

Dopamine axons are special.  They go all over the brain. The cell body of the parent neuron of the axon to be synapsed on uses dopamine as a neurotransmitter.  It sits in the pars compacta of the substantia nigra a fair piece away from the target they studied (the striatum). “Individual neurons of the pars compacta are calculated to give rise to 4.5 meters of axons once all the branches are summed”  — [ Neuron vol. 96 p. 651 ’17 ].”  These axons release dopamine all over the brain.  There aren’t many dopamine neurons to begin with just 80,000 which is 1 millionth of the current (probably unreliable) estimate of the number of neurons in the brain 80,000,000,000.

Now synapses between neurons are easy to spot using electron microscopy.  The presynaptic terminal contains a bunch of small vesicles and is closely apposed (300 Angstroms — way below anything the our eyes can see) to the post synaptic neuron which also looks different, usually having a density just under the membrane (called, logically enough, post-synaptic density).  Embedded in the postsynaptic membrane are proteins which conduct ions such as Na+, K+, Cl- into the postsynaptic neuron triggering an action potential.

But the dopamine axons going all over the brain have a lot of presynaptic specialization, but in many of the cases the post-synaptic neuron and its postsynaptic density is nowhere to be found (or the receptors for dopamine aren’t near the presynaptic specialization).  This is called volume neurotransmission.

However, in the nuclei studied (the striatum) dopamine synapses on dendrites of the major cell type (the medium spiny neuron) are well described and the 5 receptors for dopamine (called G Protein Coupled Receptors — GPCRs) are found there.  None of the GPCRs conduct ions or trigger action potentials (immediately anyway).  Instead, they produce their effects much more slowly and change the metabolism of the interior of the cell.  This is true for all GPCRs, regardless of the ligand activating them — and humans have 826 GPCR genes.

Note also that volume neurotransmission means that dopamine reaches nonNeuronal tissue — and there is good evidence that dopamine receptors are present on glial cells, pericytes and blood vessels.

The story doesn’t end with dopamine.  There are 3 other similar systems of small numbers of neurons collected into nuclei, using different neurotransmitters, but whose axons branch and branch so they go all over the brain.

These are the locus coeruleus which uses norepinephrine as a neurotransmitter, the dorsal raphe nucleus which uses serotonin and the basal nucleus of Meynert which uses acetyl choline.  There is excellent evidence that the first two (norepinephrine and serotonin) use volume neurotransmission. I’m not sure about those of the basal nucleus of Meynert.

What is so remarkable about the paper, that it allows the receiving neurons to (partially) control what dopamine input it gets.

All norepinephrine receptors are GPCRs, while only one of the 16 or so serotonin receptors conducts ions, the rest being GPCRs.

Acetyl choline does have one class of receptors (nicotinic) which conducts ions, and which the paper shows is what is triggering the axon on axon synapse.  The other class (muscarinic) of acetyl choline receptor is a GPCR.

Addendum 29 September — it goes without saying (although I didn’t say it) that any molecule released by volume neurotransmission doesn’t confine itself to finding targets on neurons.  Especially with norepinephrine, it could bind to receptors for it on the vasculature causing circulatory effects.  They could also bind to GPCRs on pericytes and glia.

Now the paper tested axon to axon firing in one of the four systems (dopamine) in one of the places its axons goes (the striatum).  There is no question that the axons of all 4 systems ramify widely.

Suppose axon to axon firing is general, so a given region can control in someway how much dopamine/serotonin/norepinephrine/acetyl choline it is getting.

Does this remind you of any system you are familiar with?  Perhaps because my wife went to architecture school, it reminds me of an old apartment building, with separate systems to distribute electricity, plumbing, steam heat and water to each apartment, which controls how much of each it gets.

Perhaps these four systems are basically neurological utilities, necessary for  the function of the brain, but possibly irrelevant to the computations it is carrying out, like a mother heating a bottle for her baby in water on a gas stove on a cold winter night.  The nature of steam heat, electricity, water and gas tell you very little about what is going on in her apartment.

The paper is so new (the Neuron issue of 21 September) that more implications are sure to present themselves.

Quibbles are sure to arise.  One is that fact that the gray matter of our brain doesn’t contain much in the way of neurons using acetyl choline as a neurotransmitter.  What it does have is lots of neurons using GABA which we know can act on axons, inhibiting axon potential generation.  This has been well worked out with synapses where the axon emerges from the neuron cell body (the initial segment).  However the different ionic composition of axons in the developing brain results in GABA having an excitatory effect.  Perhaps ionic composition varies in different parts of the neuron.

The work was done in living animals, so the paper contains no electron micrographs.  Such work is sure to be done.  No classical presynaptic apparatus may be present, just two naked axons touching each other and interacting by ephaptic transmission (the term does not appear in the paper).

So a lot of work should be done, the first of which should be replication. As the late Carl Sagan said “extraordinary claims require extraordinary evidence”.

Finally:

As Mark Twain said ” There is something fascinating about science. One gets such wholesale returns of conjecture out of such a trifling investment of fact.”

 

How the brain really works (maybe) – 2

I sent the previous post to a very intelligent friend — a PhD Electrical Engineer who responded as follows

“Correct me if I’m wrong, but it sounds like you are proposing that in addition to direct communication in the nervous system via electrical and chemical synapses, you are proposing that there could also be communication coupling in nerve fibers via local electric fields. But isn’t this a known phenomenon, ephaptic coupling? See
https://en.wikipedia.org/wiki/Ephaptic_coupling”

I didn’t think EE’s knew about such things (but I told you he’s very smart). Here are a few extra points of mine concerning his response and the article in general.

Excellent point. Thanks. What I propose could certainly be called ephaptic transmission. It has been well described between two axons in peripheral nerves (this was the initial description). Ephaptic transmission is fairly well established in muscle (which also has action potentials spreading along the muscle fiber allowing it to contract). Investigation in the brain has primarily been between adjacent neurons or adjacent axons. Questions have arisen as to whether it could be a mechanism of seizure generation.

As far as I can tell, the following ideas are actually original.
(1) Ephaptic transmission could normally occur between dendrites in the cerebral cortex.
(2) The brain and cerebral cortex is built the way it is to allow dendritic ephaptic transmission to occur.
(3) This is the way serious computations are carried out by the cerebral cortex.

Why now? Probably because there was no way of measuring dendritic electric potential changes directly before this paper (prior to this calcium levels in dendrites were used as a surrogate). Another example of new technology driving the science.

I didn’t put it in the original post, but actual paper notes that the potential flucutations across the dendritic membrane were much larger than the fluctuations recorded at the cell body.

People have wondered for years how various electrical activities in the brain could be synchronized over large areas (every electrical wave seen in the electroencephalogram is the activity of hundreds of thousands to millions of neurons). This may be an explanation — previously people had figured that it was coming from neurons lower in the brain (particularly the thalamus) sending axons all over the place stimulating neurons simultaneously. Even this doesn’t really work, because various areas of the brain are separted from each other, axonal speed is thought to be constant, and the impulses have different distances to travel.

One disturbing aspect to the picture in the previous post — If you regard that neuron as embedded in a cube 50 x 50 x 50 microns on a side, you’d get about 8,000 neurons per cubic millimeter (1,000 x 1,000 x 1,000 cubic microns). The literature says over twice that at 20,000 neurons/cubic millimeter.

I doubt that the above constitutes all the implications of these ideas. Any comments? I am quite interested to hear them.