Neuronal excitability

Topics

• What makes neurons excitable?
• How do neurons integrate information.
• How do neurons transmit information.

Topic Slide

Charles Sherrington won the Nobel Prize in 1906 for his work on neuron communication. He discovered (and named) the synapse (Greek for clasp, which is the location of information transmission between neurons). His monograph, The Integrative Action of the Nervous System, was based on his series of ten Silliman Lectures given at Yale University in 1904.

Signaling

Types of signaling

Cells need to communicate with other cells. A key idea is that a molecule released by one cell affects the activity of another cell. The released signaling molecule interacts with a receptor in the target cell. There are different names for signaling molecules depending in which biological system the signaling takes place. For example, we generally talk about hormones in the endocrine system, cytokines in the immune system, and neurotransmitters and neuromodulators in the nervous system. Notably, the very same signaling molecules can have roles in different biological systems. For example, cortisol is a steroid hormone that has widespread effects throughout the body but also has an important role in the nervous system. We will discuss cortisol in more detail in a later lecture on stress, as cortisol is released by the adrenal gland in response to stressors.

In many instances, the receptor for a signaling molecule is located within the membrane that forms the border of the target cell. In other instances, the receptor may be located within the cytosol of the target cell. Some molecules have both membrane-bound and cytosol receptors. Again, cortisol provides an example. Cortisol can bind with specialized membrane receptors to activate internal biochemical cascades within the neuron, rapidly affecting neural information processing. Cortisol can also diffuse through the membrane of the neuron (as the membrane is permeable to cortisol — we will discuss permeability later in these notes) and bind with receptor molecules in the cytosol of the neuron. The resulting complex can then pass into the genome and affect gene expression.

Different forms of cell-to-cell communication have different names.

• Endocrine (hormonal)
  • Over a distance using the circulatory system as a medium.
• Autocrine
  • Self-signaling
• Paracrine
  • Cell-to-cell signaling over a short distance
    • Paracrine signaling plays an important role in development where growth factors influence nearby cells.
    • Paracrine signaling is also involved in inflammation and the response of cells to cytokines.
    • Synaptic signaling is a form of paracrine signaling.
• Direct
  • Cells signal through direct contact.
  • Gap junctions between cells provide a basis for direct signaling.

The Khan Academy has a nice graphical presentation of these forms of signaling.

Brief digression: Neural toolkits

In Lecture 01, I stated that a brain coordinates adaptive movement that improves fitness in that movements can be used to acquire food and mates. However, adaptive movements can occur in organisms that do not possess a brain per se. So we can make inferences about the evolution of nervous systems by observing what components of a neural toolkit an organism possesses, what behaviors are enabled by those components, and what components have been preserved in evolution.

One interesting example is the marine sponge (Porifera), which subsists on bacteria obtained through filtration (filter feeders). The cellular sponges do not have a nervous system, neurons, or electrical transmission. However, there are elements of a neural toolkit in the sponge.

The sponge has a series of contractions to move water through its filtration system, and this requires coordination. Water with impurities must be rejected, and so sponges have a ‘sneeze’ mechanism. It appears that the sneeze depends upon the release of glutamate, a small molecule that is used in excitatory neurotransmission in the human and all other brains. The ‘sneeze’ can be inhibited by the application of GABA, which is used in inhibitory neurotransmission in the human brain. Thus, some components of the neural toolkit found in sponges have been preserved by evolution.

Membranes, channels, and pumps

The challenges facing plant and animal cells

The membrane demarcates a cell from its environment. Without a membrane, how would you define the difference between an amoeba and the pond water it inhabits?

However, once you have a membrane, you have problems. Chemical substances will attempt to achieve equilibrium and thus diffuse (move) along their concentration gradients inside or outside of the membrane. Their success in diffusion will depend upon concentration gradients and on the permeability of the membrane to their diffusion.

As a simple example, think of osmosis – the movement of water (diffusion) into and out of cells through pores called aquaporins. Water will diffuse in the direction from the lower concentration of solutes to higher concentration of solutes. So, if a cell filled with a high concentration of solutes enters a region of pure water, water will move into the cell. This can cause the cell to swell and perhaps burst its membrane. Think of beans soaking overnight in a clean water solution and how swollen they appear the next morning. If the same cell enters a region of very high salinity, then water may diffuse out from the cell, causing it to shrink.

One-way cells can solve this osmotic pressure problem is to separate the inside and outside of the cell with very thick walls. This is what plants have done. However, with such thick walls, plant cells cannot move around. Thus, sessile plant cells cannot move about the environment, searching for food. They must make their own food (through photosynthesis).

Animal cells need to move about and thus cannot be burdened by thick cell walls. Thus, they have thin cell membranes that permit movement. However, the membranes must now be able to regulate dynamically what is on the inside and outside.

Diffusion

Think about adding a drop of ink to a glass of water. Over time, the ink molecules will spread out with the water through a process called diffusion, whereby substances (such as ink molecules) move from an area of high concentration to an area of low concentration.

Diffusion applies to all substances in all mediums. However, if the diffusing substance has an electrical charge (such as an ion), the situation is somewhat more complicated. We refer to the diffusion of electrically charged substances as electrochemical diffusion.

Electrochemical diffusion is governed by two gradients:

• Concentration gradient: a chemical or molecule will move from a region of higher concentration to a region of lower concentration.
• Electrical gradient: a charged molecule (i.e., an ion) will be attracted to its opposite charge (opposites attract). For example, a negative ion (anion) (e.g., Cl-) will move towards a region of positive charge (anode). A positive ion (Na+) (cation) will move toward a region of negative charge (cathode). Ions with particular relevance to neurons are sodium (Na+), calcium (Ca++), potassium (K+), and chloride (Cl-).

The diffusion of a particular molecule is governed by both gradients operating simultaneously. An electrochemical equilibrium equation can be written that takes both gradients into account for each molecule. The result is an electrochemical driving force for each ion.

The Nernst Equation

The Nernst Equation provides the electrochemical equilibrium potential for a single ion species. At the equilibrium potential, an ion will have no driving force in to, or out of, the membrane (i.e., equal numbers of ions will move across in each direction). Although upon initial inspection, the Nernst Equation may look daunting, it is just the (log) ratio of the concentration of the ion on the inside and outside of the membrane. The other terms are some physical constants, the valence of the ion species (i.e., whether it is negative or positive), and the temperature.

You can read about the Nernst Equation on page 41 of your PN6 textbook (page 33 of PN5). For those of you who like Khan Academy graphics, they have a nice explanation of the membrane potential that can be found here.

The Goldman-Hodgkin-Katz Equation

If the only ion species that was distributed across the membrane was K+ (potassium), then a measurement of the voltage across the membrane (i.e., the voltage measured by one electrode inside the cell membrane, and a second electrode outside the cell membrane) would be equal to the Nernst equilibrium potential for K+. However, there are other ion species present – such as Cl- (chloride), Na+ (sodium), and Ca++ (calcium). Each of these are also unequally distributed between the inside and the outside of the cell. Each has a different Nernst equilibrium potential when taken in isolation.

To determine the membrane potential for all of these ion species simultaneously, we must use the Goldman-Hodgkin-Katz equation (page 43 of PN6, page 34 of PN5). The G-H-K equation looks similar to the Nernst equation expanded to include a term for each ion species. However, it also incorporates the relative permeability of the membrane to ion movements from the inside to the outside, and from the outside to the inside, for that ion species.

When we solve for the resting membrane potential using all of these ion species, we find that it is ~~not equal to ~~any of the ion species measured in isolation. Thus, the neuronal membrane (at rest) is actually in a dynamic state where each ion species has an electrochemical force pushing it to enter (influx) or exit (efflux) the cell.

To maintain this ionic disequilibrium, energy must be expended by the neuron to operate an ionic pump that continually moves ions ~~against their equilibrium potential. This is an important point. While ions can move effortlessly (i.e., without additional energy being expended) along their combined concentration and electrical gradients, it takes the expenditure of additional energy to move against those gradients.

Think of a water tower. The water in the water tower will move (with considerable pressure) down the pipes along the gravity gradient without additional energy. However, to get water up into the tower, an energy consuming pump is required.

Membranes and channels

Cell membranes consist of a lipid bilayer, interrupted by openings (pores, or channels). An important property of a membrane is its permeability.

Membranes have to have some permeability to substances in the environment, otherwise cellular respiration would be impossible. The neuronal membrane, for example, allows diffusion of oxygen and carbon dioxide. Some lipophillic steroid hormones like glucocortioids (e.g., cortisol) can diffuse across neuronal membranes and activate receptors within the cytosol of the neuron. Other molecules, however, cannot diffuse across the neural membrane and can only be admitted through special channels (pores or holes) through the membrane.

• Permeability: membranes have pores, or channels, through which some molecules can diffuse.
• Some channels are nonselective, though which any molecule can diffuse along its concentration or electrical gradient.
• Some channels are selective, meaning that their size or conformation will only permit particular molecules to diffuse, but not others.
• Some channels are gated and are normally closed. These gated channels require a molecular key to open them to permit diffusion.
• Some gated channels are normally closed, but open when the membrane potential reaches a threshold voltage.
• Some ~~membranes are selectively permeable. This means that they will contain disproportionate numbers of selective channels for different ion species. For example, neurons have 30x the number of K+ channels as Na+ channels. Thus, the neuron’s membrane is selectively permeable to K+ but much less so for Na+.

The movement of a molecule along its concentration gradient, or its electrical gradient, does not require additional energy. It is like water running down its gravity gradient. In a water tower, the water does not require energy to run down the pipes below the tower.

Pumps

There are molecular pumps that transport some ions across the membrane against their concentration gradients. For example, the sodium-potassium ATPase pump transports three Na+ ions out of the cell while transporting two K+ ions in to the cell. This pump requires energy (in the form of one ATP molecule per turn of the pump) to operate. Using our water tower analogy, water runs down from a water tower without the need for additional energy. But it requires energy to pump the water up into the tower.

Why are some molecules unequally distributed?

• Some large molecules, such as large proteins and nucleic acids, are too big to cross through a membrane pore. The trapped anion nucleic acids make the inside of a cell negative in electrical potential relative to the outside of the cell. This is one basis for a resting potential difference – i.e., the inside of the cell is negative relative to the outside of the cell.
• There are differences in permeability to different ions. For example, the membrane of a neuron is 30x more permeable to potassium (K+) than sodium (Na+). Some believe that life evolved in a potassium rich environment, and thus the cellular mechanisms of living organisms utilize potassium.
• The molecular pumps transport some ions across the membrane against their concentration gradients and thus maintain an unequal concentration that requires energy to maintain. In the case of the sodium-potassium ATPase pump, this results in a large concentration of Na+ outside of the cell, and a larger concentration of K+ inside of the cell. ~
  ~ Thus Na+ has a strong electrochemical driving force to enter the cell – as its concentration gradient AND its electrical gradient both point towards the inside of the neuron. Potassium has a weaker electrochemical driving force to leave the neuron, as its concentration gradient points outward, but its electrical gradient is inward.

Three fun facts:

• All animal cells have this the sodium-potassium pump including sponges, jellyfish, insects, and humans.
• The sodium-potassium pump can be poisoned by Ouabain, which is a substance used to make poison arrow heads.
• The sodium channel can be poisoned by Tetrodotoxin, which is found in the puffer fish used to make the fugu delicacy.

Excitability and Signaling

Excitability

At rest, the neuron has a resting transmembrane potential of ~–70mV where the inside of the neuron is negatively charged related to the outside. Due to the differential permeability of the neuron’s membrane to different ions, and the action of the sodium-potassium pump, there is more sodium outside the neuron than inside, and more potassium inside the neuron than outside.

Note that all cells in the body have a membrane potential, and use it to bring materials in and out of the cell. But neurons have something special that makes them excitable: gated channels.

We already discussed the fact that membranes have pores or channels that permit diffusion of some species of molecules. However, a special kind of gated channel is normally closed, but opens to permit diffusion when a molecular key is applied to unlock it. Thus, the selective permeability of the neuron’s membrane can be altered by the presence of other molecules that open these gated channels. There are different kinds of gated channels:

• Ionotropic: ligand gated channels opens when ligand attaches (analogy to a locked gate opened by a key). These gates work very quickly (milliseconds) and also close quickly when the ligand detaches. There are relatively few types of ionotropic gated channels – most use glutamate, GABA, acetylcholine (particularly in the neuromuscular junction), and serotonin.
• Metabotropic: ligand initiates intracellular cascade (usually involving g-proteins) that opens gates (analogy to an intercom, you call, something happens inside house, and then gate opens). Metabotropic receptors operate more slowly than ionotropic receptors (from 10s to 100s of milliseconds). There are many dozens of metabotropic receptors.
• Voltage-gated: opens when the membrane around it reaches a particular threshold voltage (analogy to dominos).
• Mechanical: opens when physically stretched. (We will encounter such channels in special sense organs in the skin).

Because of the unequal distribution of ions caused by selective permeability and pumps, opening a gate that is highly permeable to sodium, would cause an influx of positive sodium ions into the neuron.

When we say ‘ligand’, we are usually referring to neurotransmitters released by another cell. Thus neurotransmitters released by other neurons can open gated channels in target neurons.

There is a good article in the Wikipedia that covers these topics concerning neuronal membranes and excitability which can be found here.

Post-synaptic potentials (PSPs)

Let’s consider a prototypical signaling event. In this case we have end of an axon from neuron A opposed to a spine of the dendrite of neuron B. Their membranes are separated by a small gap through which ions and other molecules can diffuse. We will refer to the the axon membrane of neuron A as the pre-synaptic side, and the membrane of the dendritic spine of neuron B as the post-synaptic side. The gap will be called the synaptic cleft. The entire assembly will be referred to as the synapse (so-named by Sherrington). Because chemicals will move through the synaptic cleft, we will refer to this as a chemical synapse.

For the sake of completeness, I mentioned that there is another type of synapse called the electrical synapse where the pre- and post-synaptic membranes are physically joined (there is no cleft) and channels through these membranes directly connect the pre- and post-synaptic neurons. This harkens back to Golgi’s view of continuous connections among neurons. While less common than chemical synapses, electrical synapses are abundant in some brain regions, such as the superior olivary nuclei of the phons that provides input to the cerebellum. Electrical synapses also exist among interneurons in cerebral cortex.

If the presynaptic neuron is activated (we will discuss the particulars of its activity when we discuss action potentials further below), it will be stimulated to release chemicals called neurotransmitters into the synaptic cleft. The neurotransmitters will diffuse quickly across the synaptic cleft and bind to molecules on the surface of the post-synaptic membrane that open ligand gated channels (the neurotransmitter is the ligand). The gated channel will open, and will permit ions to pass through the post-synaptic membrane, altering chemical and electrical gradients. This is the basis of signaling.

Excitatory glutamate synapse

Let’s first consider a prototypical excitatory synapse in which glutamate is release by the pre-synaptic process onto to the post-synaptic membrane at the synapse.

When the pre-synaptic membrane is activated, it releases glutamate, which binds to glutamate receptors on the post-synaptic membrane that open gated channels. Think of a lock (receptor channel) and key (glutamate) arrangement. Glutamate is an amino acid that acts as a neurotransmitter, and it is released at 90% of synapses in the brain

You should be aware that there are different kinds of glutamate receptors with different names and properties. Common ionotropic glutamate receptors are the AMPA, Kainate, and NMDA type (although the NMDA receptor also has a voltage-gated characteristic discussed below). There are also metabotropic glutamate receptors, in which glutamate initiates activity within biochemical pathways inside the post-synaptic cell and nearby glia (astrocytes). The ionotropic glutamate (iGluR) receptors work very quickly while the the metabotropic glutamate receptors (mGluR) open and close much more slowly. This has implications for their function that we will discuss further.

At the excitatory synapse, glutamate opens ionotropic gated channels that allow sodium ions (Na+) to enter the cell. Recall that Na+ has a considerable electrochemical drive to enter the post-synaptic cell. The in-rush of positive sodium ions causes the local transmembrane potential in the area of the synapse to change in a positive direction (i.e., from –70 mV towards zero). This is called depolarization. The change in membrane potential due to this excitation is called a post-synaptic potential and, as it is a depolarizing change, we called it an excitatory post-synaptic potential, or EPSP. As we will learn below, depolarization is a critical step in activating the post-synaptic neuron, so that it sends signals to other neurons through its own axonal processes.

The fast iGluR channels produce EPSPs with a very short latency (within milliseconds, or ms) and very short duration (~5 ms). The slow mGluR produce EPSPs with a longer latency and with a much longer duration (~100–200 ms). Thus, the resulting EPSP at a glutamate receptor will depend upon the number and type of its glutamate receptors.

Astrocytes (a type of glial cell) participate in regulating the synapse, and thus the combination of the pre-synaptic, post-synaptic, and nearby astrocyte process is referred to as the tripartite synapse. The astrocyte process takes up excess K+ from the synapse, and takes up excess glutamate. Glutamate in excess is toxic to neurons, and so this important process of astrocytes helps avoid excitotoxicity. The astrocyte recycles glutamate to a non-toxic form called glutamine, which is transported back from the astrocyte to the pre-synaptic neuron to be recycled into glutamate.

The glutamate released by the pre-synaptic membrane also stimulates metabotropic receptors on the astrocyte process that admit calcium ions (Ca++). By processes beyond our present concern, the Ca++ influx creates calcium waves through a network of glia (using direct signaling through gap junctions) that influence blood local blood flow. In this way, glia respond to neuronal activity to increase local blood flow to support signaling. Please see Box 4.1 at the end of these notes for a summary of recent evidence concerned with the role of astrocytes in the tripartite synapse.

Inhibitory GABA synapse

Another amino acid, GABA (gamma-aminobutyric acid), is the most common inhibitory neurotransmitter. When released by the pre-synaptic neuron onto the post-synaptic membrane, is opens a ionotropic gated channel for chloride (Cl-). This is called a GABA-A receptor. There is a little chloride present inside the neuron, so there is a concentration gradient for Cl- to move into the neuron. However, Cl- is a negative ion, and so the total electrochemical drive for Cl- to enter the cell is less than the drive for Na+ to enter the cell in response to glutamate at an excitatory synapse. Indeed, the Nernst equation for Cl- is very close to the typical membrane potential of a neuron, and so the electrochemical driving force for chloride is low. Depending upon what the actual membrane potential is at a synapse, Cl- can move either inwardly or outwardly when a GABA-A channel is open.

GABA release also opens a metabotropic receptor (the GABA-B receptor). The post-synaptic GABA-B receptor opens K+ channels and, thus, K+ flows out of the neuron. Both K+ efflux from GABA-B channels, and Cl- influx from GABA-A channels influences the local transmembrane potential, but in a negative direction. That is, it goes from –70 mV to an even more negative number. The neuron thus becomes hyperpolarized and more difficult to activate. We call the hyperpolarizing potential change initiated by GABA an inhibitory post-synaptic potential, or IPSP.

Passive spread of PSPs and the Law of Conservation of Charge

Post-Synaptic potentials (both IPSPs and EPSPs) are primarily local events, that alter the transmembrane potential maximally near the site of the synapse itself. However,

However, in a closed system, like a neuron, the charge entering the depolarized region of the cell must be matched by charge leaving the cell. This is the Law of Conservation of Charge.

Let’s imagine the following example. Imagine that you were to attend a party in a closed room with a single entrance door, and a number of windows along the sides of the room. The room is completely packed such that another person (or charged ion) cannot enter the room. Nevertheless, there is a strong attraction to the party and a lot of people outside want to enter. As each person pushes in to enter, somebody falls out of the window. The person who falls out of the window wants to re-enter, and so s/he attempts to enter again at the doorway. This creates a loop, where charge entering is matched by charge leaving.

This inflow and outflow of current creates a loop, which is strongest close to the point of depolarization (where the neuron membrane is depolarized by influx of sodium), or hyperpolarization (where the neuron membrane is hyperpolarized by the influx of chloride) and which falls off with distance along the neuron.

Let’s consider the excitatory synapse, as sodium rushes into the cell (its strong electrochemical driving force due to both a concentration gradient and an electrical gradient) , it begins to diffuse inside the post-synaptic membrane in both directions. As sodium in positive, it creates a positive current flowing inside the neuron. This positive flow causes other positive ions (like potassium, K+) to flee (like charges repel). The positive K+ ions diffuse through their leakage channels in post-synaptic membrane into the extracellular space. As mentioned above, some of the excess potassium is taken up by astrocytes. The remainder is eventually returned to the post-synaptic neuron when the sodium-potassium pump operates.

Eventually, the gated channels opened by glutamate close and the influx of sodium ends. The glutamate is recycled by the astrocytes, and the sodium-potassium pump returns the post-synaptic membrane back to its original state. So what was that all about? If the system returns to its original state, what information has been transferred?

Summation of PSPs

Although these passive depolarizations and hyperpolarizations are graded (i.e., fall off with distance), they still influence the membrane potential at the axon hillock, which is the part of the neuron’s soma at the point where the axon begins. An individual neuron might have hundreds of synapses on its extensive dendritic branches and on its soma, and many will be generating PSPs at the same time. Thus, the axon hillock computes a ‘sum’ of all transmembrane currents. It is this summation constitutes information processing in the neuron.

Imagine that you have a long garden hose attached at one end to a water spigot. Further imagine that the hose has holes along its length, but the distant end of the hose is capped. If you turn on the water pressure for a few moments (our EPSP), the holes closest to the water spigot will shoot out water with the most pressure, and the holes near the distant capped end will have little water pressure. However, it will not be zero. No matter how long the hose, every hole will dribble a little water.

Now imagine that you have many hoses of the type described above, and the distant ends are all located on the same patch of grass. If you give a momentary pulse of water to all of the individual hoses, that distant patch of grass will receive a tiny bit of water from every hose. The total water at that patch will be the sum of all the tiny bits of water. With enough hoses, that could be large amount. Thus, the axon hillock will experience the sum of all of the PSPs that all of the synapses are experiencing.

The ‘summation’ at the axon hillock is influenced by three factors:

• In this ‘summation’, the IPSPs and EPSPs can be thought of as negative and positive numbers, that is, they move the membrane potential at the axon hillock in different directions.
• The spatial distribution of depolarizations and hyperpolarizations. PSPs initiated closer to the axon hillock will be more consequential than those distant to the axon hillock.
• The temporal interval between successive depolarizations and hyperpolarizations matter. PSPs have a duration, and successive PSPs will summate over time. However, if the interval between PSPs is too great, they will not summate.

In class, I gave examples of how the spatial arrangement of the same number of depolarizations and hyperpolarizations could influence activity at the axon hillock very differently. For the computer logic minded, I also used the idea of the neuron as a logic gate (AND gate) that would fire an action potential when two synapses are simultaneously depolarized, but not when either one of the them was depolarized alone. However, while the example of a two-input AND gate is appealingly simple, I showed an example of a neuron in the spinal cord in which there were many dozens of synapses visible.

To summarize, the axon hillock computes a complex sum based upon spatial geometry and temporal synchronization of dozens to thousands of synapses. This is how neurons integrate information. These synapses primarily occur on dendrites but also on the soma of neuron. The PSPs spread passively along the membrane (a process called electrotonic conduction), but diminish quickly with distance from the synapse. The PSPs summate over time and space (in this sense, the neuron is a form of analog computer). The axon hillock at the base of the axon experiences the ‘sum’ of the PSPs. If a particular threshold in the transmembrane current of the axon hillock is reached, then the neuron initiates an action potential.

The Action Potential.

Transmitting information to other neurons

At this point in our understanding, the membrane at the axon hillock has experienced a net depolarization by the sum of near-simultaneous EPSPs (minus IPSPs) impinging upon synapses on the dendrites and soma of the neuron. Well, so what? What is special about the depolarization of the axon hillock? Won’t that depolarization just spread passively down the axon, rapidly diminishing with distance? Since the EPSP will die off with distance, how can one neuron communicate at a distance with another neuron? After all, some axons are more than a meter long!

Voltage-gated channels

To transmit a signal over a long distance, that signal must be actively regenerated, not passively conducted.

Here we need to reintroduce another type of gated channel, the voltage-gated channel. The voltage-gated channel does not require a ligand, but rather opens when the surrounding membrane potential reaches a threshold. If you imagine a long line of voltage-gated channels, then the opening of the first will depolarize the membrane surrounding the second, it will depolarize and then depolarize the membrane surrounding the third, and so forth. Thus, voltage-gated channels are like dominoes that fall in order. As each opens, it regenerates the signal from the prior. Thus the Action Potential moving down an axon of voltage-gated channels is self-propagating.

Details of the Action Potential (AP)

• The Action Potential depends upon sodium (Na+) voltage-gated channel that opens when the transmembrane potential (the resting potential) exceeds (or becomes closer to zero than) a threshold of about –50mV (from –70mV). The Na+ voltate-gated opens and Na+ flows into the neuron, depolarizing it (i.e., making it less negative inside relative to outside). So much Na+ flows inside, that the the transmembrane current actually reverses and becomes positive inside relative to the outside. The sodium voltage-gated channel inactivates (closes) when the transmembrane current reaches +30mV.
• At +30mV, the potassium (K+) voltage-gated channel opens, and allows K+ to rapidly flow out of the neuron. It wants to escape the high concentration of K+ inside the cell, and also wants to escape the now positive charged region inside of the neuron’s membrane.
• The potassium channel stays opens until the neuron reaches about –80mV – that is, until the neuron is hyperpolarized.
• When the neuron is hyperpolarized, it cannot fire another action potential. It is in an absolute refractory phase. This helps prevent the action potential from flowing backward down the axon towards the cell body (soma)
• After the action potential, there is an excess of Na+ inside the axon, and an excess of K+ outside the neuron. The sodium-potassium pump (sodium ATPase) then pumps out Na+ and bring in K+, thus restoring the membrane to its initial composition. This pump is the single largest energy user in the entire brain.

Myelination and the Action Potential

We discussed a type of glial cell called the oligodendrocyte which wrapped the axons of some neurons with a myelin sheath. I mentioned that this speeded nerve conduction, but I didn’t say how.

• The axon in many neurons is wrapped in a myelin sheath produced by oligodendrocytes, a type of glial cell. The myelin is interrupted by open areas called the nodes of Ranvier.
• Myelinated axons conduct action potentials much fast than unmyelinated axons (and larger axons conduct faster than thinner axons).
• The AP jumps from node to node because the membrane between nodes is insulated and there is no way for the current to leak out through the myelin insulation. This intensifies the current flow inside the axon which then depolarizes the axon at the nodes of Ranvier where Na+ gated channels reside.
• Some diseases demyelinate axons and cause motor and sensory disturbances in affected individuals. One such disease is multiple sclerosis.

Revisiting the neuron doctrine.

The canonical neuron doctrine emphasizes one-way transmission – i.e., the dendrites and soma receive excitation and then transmit action potentials through action potentials. This is an ‘information processing’ view of neurons.

There are deviations from this canonical view:

• There is evidence for propagated dendritic potentials (‘dendritic spikes’) that use voltage-gated channels to propagate depolarization to the axon hillock. This would provide an advantage in influencing the axon hillock – e.g., a distant synapse could have more influence on the axon hillock if it could propagate depolarization rather than depend upon passive conduction.
• There is evidence for retrograde action potentials – i.e., action potentials that sweep backwards from the axon hillock through the soma and dendrites. This would influence the excitability of the soma and dendrites. As we will see later, it may also serve as a mechanism for synaptic plasticity.
• There are autoreceptors. Presynaptic receptors that respond to the neurotransmitters released by the very same presynaptic process.
• There are axo-axonic synapses, where an axonal process is presynaptic (i.e., releases neurotransmitter) onto a postsynaptic axon (i.e., another axon that has receptors for the neurotransmitter released).
• There is retrograde signaling in which the ‘post-synaptic’ membrane releases a neurotransmitter (endogenous cannibinoids and/or nitric oxide) that stimulate receptors on the ‘pre-synaptic’ membrane and influence neurotransmitter release from the pre-synaptic membrane.

These non-canonical mechanisms enumerated above may play an important role in neuronal plasticity – i.e., they are mechanisms by which the post-synaptic neuron can modify itself in response to stimulation, and by which the post-synaptic neuron can provide feedback to the pre-synaptic neuron.

Neurotransmitters, neuromodulators, and neurohormones

Confusion can result from terms with overlapping meanings, such as neurotransmitter, neuromodulator, and neurohormone. In every case, a molecule is released by one cell (usually neuron) that influences another neuron. The distinction among the terms comes from the proximity of release to effect, the requirement for a specialized synapse, and the duration of action. But these distinctions become muddied in that some molecules can satisfy the definitions for each term.

The following was taken from Oxford Scholarship Online and presents a good summary.

A neurotransmitter is a messenger released from a neuron at an anatomically specialised junction and that diffuses across a narrow cleft to affect one or sometimes two postsynaptic neurons, a muscle cell, or other effector cell. Typically, a neurotransmitter acts directly on a postsynaptic neuron to cause a change in its membrane potential, although it may sometimes act through second messengers.

A neuromodulator is a messenger released from a neuron in the central nervous system, or in the periphery, that affects groups of neurons, or effector cells that have the appropriate receptors. It may not be released at synaptic sites, often acts through second messengers and can produce long-lasting effects. The release may be local so that only nearby neurons or effectors are influenced, or may be more widespread, which means that the distinction with a neurohormone can become very blurred. The act of neuromodulation, unlike that of neurotransmission, does not necessarily carry excitation of inhibition from one neuron to another, but instead alters either the cellular or synaptic properties of certain neurons so that neurotransmission between them is changed.

A neurohormone is a messenger that is released by neurons into the haemolymph and which may therefore exert its effects on distant peripheral targets. It may differ only in degree from a neuromodulator in the extent of its action. Those with restricted actions may sometimes be called local neurohormones to emphasise that their effects are localised.

These terms do not define rigid categories but rather the peaks in a continuum of effects. Some substances, such as nitric oxide (NO), can act as a transmitter but by virtue of their diffusibility might act on many cells, while substances such as octopamine can fulfill the requirements of all three definitions. A restricted but … workable definition of neuromodulation (Kaczmarek and Levitan, 1987) suggests that it is ‘the ability of neurons to alter their electrical properties in response to intracellular biochemical changes resulting from synaptic or hormonal stimulation’. On this basis, neuromodulation can result from the actions of a substance defined in any of the three categories.

Box 4.1 Gliotransmission

We briefly introduced glia in our last lecture. The functional role of glia — particularly astrocytes — in the brain has been unfolding for the past several decades. Once thought to play only a passive and supportive role for neurons, astrocytes are now understood to be active participants in the synapse and to play an important role in regulating blood flow, and thus energy, to active neurons.

Astrocyte processes participate in a tripartite synapse with a pre and post-synaptic neuron. Astrocytes buffer extracellular potassium and take up glutamate released by presynaptic neurons into the synapse and recycle it to non-active glutamine. This prevents excitoxicity by excess glutamate in the synapse. Astrocytes can also respond to some forms of synaptic activity with increases in intracellular calcium concentrations. This can result in a ‘calcium wave’ which passes through the astrocyte. By virtue of gap junctions between adjacent astroctyes, these calcium-mediated signals can pass from astroctye to astrocyte in a syncytium.

But what is the purpose of this calcium signaling? Is this all in the service of blood flow regulation? Do glia take part in information processing? This issue has been hotly debated in the recent literature.

One intriguing and yet controversial feature of astrocytes is that of gliotransmission. It has been established that astrocytes in immature brains have mGluR5 receptors expressed in the tripartite synapse. These are metabotropic glutamate receptors that respond to glutmate released presynaptically. Activation of mGluR5 receptors in astrocytes increase intracellular calcium within the astrocyte which can then spread from astrocyte to astrocyte. Furthermore, studies have indicated that increases in calcium within the astroctyte can initiate the release of transmitters from the astrocytes (gliotransmitters) including glutamate, GABA, ATP, and D-serine. D-serine is particularly interesting, as it is a co-factor in the activation of the NMDA receptor on the post-synaptic membrane. Thus, an astrocyte could be stimulated by the presynaptic release of glutamate and release D-serine into the synapse that would facilitate activate of post-synaptic NMDA receptors and subsequent plasticity.

This process has been extensively studied by Park and colleagues (2022) in the Drosophila (fruit fly) brain where thirst signals cause gene expression changes in astrocytes which lead them to release D-serine into the tripartite synaptic space and affect post-synaptic neuronal NMDA receptors that cause behavioral changes related to the alleviation of thirst. These data strongly suggest that astrocytes are participating in neural circuites assocaited with sensing and responding to thirst.

In addition to gliotransmission, activation of the astrocyte syncytium would be transmitted to astrocytic endfeed that are wrapped around capillaires and would thus provide the astrocytes with an important role in channeling energy to the active neurons. It is notable that a single astrocyte can participate in 100,000 synapses. Thus, the astroctyte is in a position to detect local changes in neuronal activity.

The hypothesis relating calcium activation in astrocytes to gliotransmission and local blood flow regulation flourished in studies published in the 1990s and early 2000s, but it ran into difficulties in the mid-to-late 2000s. There were a number of criticisms raised that are beyond this brief summary. Chief among them were the fact that astrocytic mGluR5 receptors, though abundant in the immature brain, diminished in adult animals, and anyway were also represented in neurons and not just astrocytes. and that the Ca+ signal responded too slowly to meaningfully affect local blood flow regulation.

More recent work has emphasized that there are different astrocytic compartments for calcium increases. For example, in fine astrocyte processes that extend into the tripartite synapse, exogenous calcium can enter the astrocyte and increase concentrations, while release of endogenous calcium can occur in the soma. Also, calcium increases can occur from stimulation of mGluR2 and mGluR3 metabotropic glutamate receptors, as well as norepinehprine and acetycholine receptors in the astrocyte. Thus, there has been an increase of understanding the complexity of astrocyte signaling.

These different processes involve in calcium signaling for astrocytes were summarized by Barzagani and Attwell in their 2016 Nature Neuroscience paper and represented in their figure below, which I slightly modified.

Figure legend modified from Barzagani and Attwell, 2016. Note that [Ca2+]i refers to increase in intracellular calcium concentration in astrocyte. [Na+]i refers to increase in intracellular sodium concentration, and [K+]o refers to extracellular potassium concentration. I added the 9th bullet point and process number. Astrocyte morphology has been distorted to define the locations of signaling processes.

1. [Ca2+] i transients in the processes of astrocytes (1a) differ from those in the soma (1b) in terms of frequency, kinetics and spatial spread.

2. [Ca2+]i transients in the processes of astrocytes depend roughly equally on Ca 2+ entry (2a) from the extracellular space through ion channels (40%) and on Ca2+ release from intracellular stores (60%), while those in the soma (2b) depend largely (90%) on Ca2+ release from the intracellular stores.

3. [Ca2i transients can be generated by Ca2+ entry through spontaneously opening TRPA1 channels or neurotransmitter-gated channels (3a), by mGluR2 or mGluR3 (3b), and by neurotransmitter uptake raising [Na+]i and reversing Na+ /Ca2+ exchange (3c).

4. [Ca2+]i rises TRPA1 may release transmitters via ion channels such as P2X7 Best–1, as well as via exocytosis.

5. [Ca2+]i rises NMDA alter the surface expression of neurotransmitter transporters.

6. Activation of Na+ /Ca2+ exchange by a [Ca2+]i rise can raise [Na+]i and activate the sodium pump, lowering [K+]o and hyperpolarizing nearby neurons. This increases the release probability (prelease ) for action potential–driven vesicle release and thus decreases synaptic failure rate.

7. ATP released by a [Ca2+]i rise may act on P2X or P2Y receptors to raise [Ca2+]i farther along the cell, propagating a Ca2+ wave along the cell (7a), or be converted to adenosine, which acts on presynaptic receptors to increase (A2A ) or decrease (A1 ) transmitter release (7b).

8. Noradrenaline (NA) released from locus coeruleus neurons and acetylcholine (ACh) released from nucleus basalis neurons produce large [Ca2+]i rises in astrocytes.

9. Elevation in astrocytes. Ca2+ waves can spread through the astrocyte’s processes to its vascular endfeet, where vasoactive messengers are released. AA, arachidonic acid; PGs, prostaglandins; 20-HETE, 20-hydroxyeicosatetraenoic acid; EETs, epoxyeicosatrienoic acids; PLC, phospholipase C; Ado, adenosine; A1 , adenosine’s A1 receptor; P2Y1 , ATP’s P2Y1 receptor.

Bazargani, N., & Attwell, D. (2016). Astrocyte calcium signaling: the third wave. Nat Neurosci, 19(2), 182–189.

Park, A., Croset, V., Otto, N., Agarwal, D., Treiber, C. D., Meschi, E. et al. (2022). Gliotransmission of D-serine promotes thirst-directed behaviors in Drosophila. Curr Biol, S0960–9822(22)01175.

Savtchouk, I., & Volterra, A. (2018). Gliotransmission: Beyond Black-and-White. J Neurosci, 38(1), 14–25.