Tag Archives: Pericyte

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”.


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.”


The pericyte controls local cerebral blood flow

Actively firing neurons get all the blood flow they need. More in fact. And this is the entire basis of functional magnetic resonance imaging (fMRI). At long, long last we may be close to understanding exactly how this happens.

Almost 100 years ago Wilder Penfield operating on unanesthetized patients with epilepsy to find the epileptic focus and remove it, noted that when a patient had a seizure on the table, veins became red, because so much blood flowed to the active area that it couldn’t absorb all the oxygen contained in the hemoglobin of the red cells, so they stayed red. Penfield was not a sadist, the brain contains no pain fibers, and so the skull could be opened using just local anesthetics. 

Exactly the same thing happens locally when neurons become active firing lots of action potentials. The functional MRI signal is due to the difference in magnetic susceptibility of the iron atom in hemoglobin when it is binding oxygen and when it isn’t.

So how does a firing neuron tell blood vessels it needs more flow?  A superb paper [ Proc. Natl. Acad. Sci. vol. 117 pp. 27022 – 27033 ’20 ]–https://www.pnas.org/content/pnas/117/43/27022.full.pdf probably explains exactly how this happens.  

The pericyte is a cell which is found outside cerebral capillaries and very small arteries.  It isn’t like a rubber band around the vessel (that’s for smooth muscle).  It’s like our bony spine with ribs coming from it, so the spine lies on the long axis of the vessel with the ribs coming down and wrapping (partially) around the vessel.

Pericytes in the brain and the retina are found primarily where two capillaries join each other according to the paper (which provides a convincing picture).

Neurons firing impulses release potassium into the extracellular space.  The endothelial cells of brain capillaries sense this and open up the inwardly rectifying potassium channel KIR2.1, exposing the outside to the resting potential of potassium which is quite negative (e. g the endothelial cell hyperpolarizes in response to neuronal activity.  The signal propagates upstream THROUGH the endothelial cells (because they are coupled together by gap junctions). 

Enter the pericytes which are electrically coupled to the underlying capillary endothelium by gap junctions, so they can receive the endothelial hyperpolarizing signal directly.  This causes the pericyte process receiving the signal to relax opening up the capillary or small artery increasing blood flow.  The authors followed this by watching intracellular calcium changes in pericytes, and noted that individual processes (ribs in the analogy above) could respond individually.  This is how a pericyte straddling the junction of two capillaries will open just the one which is hyperpolarized by neural activity.  

An incredibly elegant mechanism.  Of course with something so dramatic the work needs to be repeated. 

It is a pleasure to write something not involving the pandemic virus and our response to it.