Synaptic transmission and neurotransmitter systems

This chapter is relevant to the aims of Section K1(iv) from the 2017 CICM Primary Syllabus, which asks the exam candidate to "describe the major neurotransmitters and their physiological role, with particular reference to GABA, excitatory and inhibitory amino acids, acetylcholine, noradrenaline, dopamine and serotonin and NMDA receptor". They ask you this, but never follow through with actually putting it into the exams. The only time this has become even vaguely exam-relevant has been in Question 1(p.2) from the first paper of 2010, and even this question is only related to neurotransmitters because it asked for the physiological basis of serotonin syndrome (unsurprisingly, the pass rate was 0%). Also, Question 17 from the second paper of 2012 had asked for a selection of drugs which act on serotonin receptors.

At the same time, it would be disingenuous to state that one can simply coast through these exams without having the slightest idea about neurotransmitter systems. Some basic understanding is probably beneficial, considering especially that all the examinable CNS-active drugs exert their actions by interfering with neurotransmitter function. This will be the focus of this chapter.

You could really do no better than Roy Webster's Neurotransmitters, drugs and brain function (2001). Not all of its 550 pages are relevant to CICM exams, and one could safely limit one's reading to just the headline chapter ("Neurotransmitter systems and functions: overview").  If one is really interested, one could get a hold of Kumar & Deb's cowboy-sounding Frontiers in Pharmacology
of Neurotransmitters (2020), which is over 700 pages and is actually the main source for most of the material in this chapter. On the other end of the detail/value spectrum, Ayano (2016) is free to read, but is a highly oversimplified summary, and may not be suitable to the CICM trainee. Lastly, specifically for the topic of synaptic transmission, Holz & Fisher (2012) is excellent.

Neurotransmitters

If you were a molecule, what characteristics or behaviours would let you call yourself a neurotransmitter? The term sure seems to be thrown around a lot. Most textbooks dealing with neuroscience tend to take upon themselves the task of defining this term. For example, the 2nd ed of Neuroscience by Purves et al (2001) focuses on the location where the substance is secreted: 

"Neurotransmitters are chemical signals released from presynaptic nerve terminals into the synaptic cleft."

Others borrow a prefabricated definition from another authority, which also involves synapses. Webster unironically quotes the Oxford dictionary:

"A chemical substance which is released at the end of a nerve fibre by the arrival of a nerve impulse and, by diffusing across the synapse or junction, effects the transfer of the impulse to another nerve fibre, a muscle fibre, or some other structure"

The Mirriam-Webster definition is similar:

"a substance (such as norepinephrine or acetylcholine) that transmits nerve impulses across a synapse"

But then a lot of neurotransmitter substances are released into the circulation, where they have the same effect (eg. noradrenaline), i.e. bridging a synapse is clearly not essential. That's ok: we can compromise. When noradrenaline is released into the bloodstream, it is a hormone, and when it is released from one neuron and binds to another, it is a neurotransmitter. A molecule can be two things. 

To bring a bit of scientific formality back into this discussion, the following criteria seem to be essential for the definition of a neurotransmitter:

• It must be found in the presynaptic neuron and released from a presynaptic terminal
• It must be released in response to the depolarisation of the presynaptic neuron, and this release must be calcium dependent
• The postsynaptic neuron must express receptors which are specific for this molecule
• The effect of the chemical on these receptors should be blocked by antagonists in a dose-dependent manner
• It  must be reabsorbed into the presynaptic neuron or glia, or metabolized into an inactive form by enzymes to terminate the stimulation.

These are a bit of a remix of Purves et al (2001) and Inoue (2009); the two sources are equally authoritative reference texts for neurophysiology, but their criteria for calling something a neurotransmitter are more than slightly different. Moreover, and most frustratingly, neither author explain why they selected these criteria, or where they came from. Why is reabsorption and metabolism essential for the definition? Where did the calcium dependence come from? 

Synaptic neurotransmission

At a basic level:

• A  nerve impulse is conducted to the presynaptic endplate of a neuron
• At this endplate, the neurotransmitter substance is stored in vesicles
• The arrival of an action potential and the depolarisation of the presynaptic membrane causes the release of the neurotransmitter into the "synaptic cleft"
• This release is generally mediated by intracellular calcium entry, i.e. ionised calcium acts as the secondary messenger
• The released neurotransmitters cross the (narrow) cleft and bind to their receptors
• This changes the properties of the membrane in some way, either by affecting the threshold potential or by directly producing depolarisation 
• In this fashion, the action potential of one neuron, though a neurotransmitter signal molecule, facilitates a change in the state of another neuron, which is the barest essence of neurotransmission. 

It would be tempting at this point to pull out Adobe Illustrator and create some nasty diagram full of popping vesicles and hotdog-shaped receptor molecules, but one cannot help but feel that this would be insulting to the reader. By the point they decide to resort to Deranged Physiology, people are typically at some halfway point though their critical care training, and therefore well familiar with abstract diagrams of the synapse. For these people, it's time for the real thing, or at least its osmium-stained shadow.

Ultrastructure of the synapse

A "synapse" is defined as a junctions between nerve cells, or the site where they communicate. The term is usually credited to Charles Sherrington who had first used it in a 1897 physiology textbook:

‘‘So far as our present knowledge goes we are led to think that the tip of a twig of the [axon’s] arborescence is not continuous with but merely in contact with the substance of the dendrite or cell body on which it impinges. Such a special connection of one nerve cell with another might be called a synapsis’’

Apparently Sherringotn and Foster (the eminent professor who published said textbook) were trying to avoid the use of the word "junction", because for the English speaker it would be "redolent of the continuum" and they wanted to be very clear that neurons were separated. According to Tansey (1997), the pair came across "synapsis" as the term they favoured most, because it "strictly means a process of contact—i.e. a proceeding or act of contact, rather than a thing which enables contact i.e. an instrument of contact". It was a concept probably influenced to a considerable degree by the contemporary theory that neurons could move like amoebae - extending and retracting their axons and dendrites as pseudopods. It was therefore tempting for the authors to represent the site of two neurons connecting as a transient touch, rather than a lasting point of attachment.

Synapses form at numerous sites along a neuron. Here's a picture of pyramidal neuron from Hrvoj-Mihic et al (2013), demonstrate the comparatively massive scale of this cell. As you can see, the cell body of the neuron itself is huge (about ten erythrocytes could fit inside). The branching dendrites projecting from the main body span over hundreds of micrometres, and the axon could potentially be a metre long.

Note how these projections are studded by little thickened swellings. These are "dendritic spikes" and "axonal boutons", the sites of synaptic connection between neurons. They can be terminal boutons (i.e. at the very end of an axon), or they can occur along the length of an axon like huge deformed nodes of Ranvier. Here, some excellent art from Nicol & Walmsley (1993) reconstructs such an en passant bouton from a series of electron microscope slices:

These boutons represent the synaptic interface. Two boutons or a bouton and a dendritic spike would typically lay in close apposition. To help visualise this, Burette et al (2012) published some detailed images of synaptic ultrastructure, which represents the next level of zoom. Here is a couple of terminal boutons, with the presynaptic terminal easily identified by the excess of little (~30nm) round vesicles, all full of neurotransmitter molecules.

As you can see, the space is very small - the average distance between the membranes is about 20-30nm. It is also not empty, as you can see from the high electron density of the cleft content. It is filled with a fibrous extracellular protein matrix which acts to keep the two membranes segregated, while acting as a scaffold for various important membrane components (Nakayama et al 2016). The spaces between these protein fibrils are probably filled with a rather pure and wholesome variant of extracellular fluid.  Ordinarily, the interstitium between cells is full of gross debris and random cellular metabolic byproducts, but obviously that would not be suitable to a space designed for rapid molecular traffic. However, the contents of the synapse and the contents of the extracellular fluid compartment do tend to mingle, which is how the drugs acting here (eg. metaraminol) find their way to their site of action.

Now, for the longest time, it was believed that these vesicles all contained the same neurotransmitter, and that in fact throughout the entire neuron the vesicles will all be the same. This was called Dale's Principle, after Sir Henry Dale generalised some of his findings in the autonomic nervous system (Eccles et al, 1954). We now know that this is probably not true, and there are plenty of examples of neurons capable of releasing multiple different neurotransmitter substances into their synapses. However, the terminology of the 1930s has stuck, which means one will still see people referring to "adrenergic" or "dopaminergic" neurons in the nervous system, as if those cells are restricted in their biochemical range.

Anyway: the presynaptic vesicles bind to the synaptic junction and release their contents into the synaptic cleft. Some fantastic images (again from Burette et al, 2012) demonstrate the size and shape of this process, as the authors were actually able to capture these organelles in flagrante. Here, especially in the image labelled E3, the vesicle's union with the presynaptic membrane can clearly be seen. 

This is what happens when an action potential arrives at the synapse. The process by which the electrical weirdness of an action potential is transformed into the movement of vesicles is well described in basically any textbook you might pick up, but to repeat in in brief:

• The presynaptic membrane has voltage-gated calcium channels. These are distinct from the sort of calcium channels you might block with verapamil, and they are usually referred to as N-type calcium channels.
• With the arrival of the action potential, the membrane depolarises, causing these N-type calcium channels to open
• With this, calcium influx into the presynaptic nerve terminal occurs.
• Inside the cell, this calcium acts as a secondary messenger
• The specific targets relevant to neurotransmission here are proteins such as synaptotagmin, synaptobrevin, syntaxin, and several other proteins which are broadly referred to as the SNARE family
• These proteins are responsible for mediating the fusion of vesicles with the presynaptic membrane, and the exocytosis of vesicle contents
• The neurotransmitter is therefore emptied out of the vesicle and, when in the synapse, it does two main things:
  • Binds to the post-synaptic receptors, producing some change in the other neuron
  • Binds to the pre-synaptic receptors on the same neuron which had just released it, and therefore exerting some sort of feedback effect
• After this, the neurotransmitter is usually scrubbed out of the synaptic cleft by the action of (usually) various reuptake pumps, or (more rarely) by the activity of a high-affinity enzyme that destroys the neurotransmitter molecule, like acetylcholinesterase. In either case, the objective is to rapidly and completely clear the synapse of any residual neurotransmitter, so that the slate is scraped clean for the next synaptic transmission event.
• This process (from the action potential arriving to the triggered post-synaptic changes after neurotransmitter binding) is blindingly rapid, faster than a gunshot. In the room-temperature squid giant axon it is about 0.2 msec (Llinás, 1982), and in the warm mammalian central nervous system Sabatini & Regehr (1996) recorded speeds of up to 60 μs (0.06 ms), which is much faster than even the rate of voltage change at the depolarising axonal membrane. The high concentration of neurotransmitter, high receptor affinity and short diffusion distance (20-30nm) all contribute to the incredible speed of this reaction.

An image of catecholamine behaviour at the synapse, and the drugs which routinely affect it, is borrowed here from the autonomic nervous system chapters, in the unlikely case that it might be somehow clearer than the point-form description it follows:

This process of calcium-mediated release-recapture-repeat is common to basically all synaptic neurotransmitter molecules. Their unique effects, on the other hand, are mediated by the functional characteristics of their postsynaptic receptor complexes. This is what makes neurotransmitters "excitatory" or "inhibitory", for example. Which is a handy segue into a discussion of...

Classification of neurotransmitters

Another method might be to divide them into "excitatory" or "inhibitory". The terms refer not to the net effect of the neurotransmitter on the performance of the CNS systems they are involved with, but on the effect on the post-synaptic neuron. Excitatory neurotransmitters increase the likelihood that the receiving neuron will generate an action potential, and inhibitory neurotransmitters decrease it.

• Excitatory neurotransmitters:
  • Glutamate
  • Dopamine
  • Noradrenaline
  • Acetylcholine (nicotinic receptors)
• Inhibitory neurotransmitters:
  • GABA
  • Serotonin
  • Acetylcholine (muscarinic receptors)

However, this is also not satisfactory, particularly because of the way the central nervous system seems to string double negatives together. A GABA neuron might inhibit another GABA neuron, which is the inhibition of inhibition, i.e. excitation. Moreover, some neurotransmitters (eg. acetycholine) may have either an inhibitory or an excitatory effect, depending on the receptor subtype. A subtle variant of this is to separate neurotransmitters into those with "direct" effects and those which are merely "neuromodulatory", but again many of them could fall into both categories.

Another classification system looks at the biochemistry of the neurotransmitter molecule. In this way, you can group them into monoamines, amino acids, peptides, large molecules, etc. This makes a bit of sense. There is clearly something special about the chemistry of neurotransmitter molecules, as they show considerable evolutionary resilience, in the sense that even locusts and echinoderms use the same basic combination of molecules to exchange information in their nervous system. Apart from the occasional use of weird molecules like octopamine, we animals are all very similar from that perspective. From this, it follows that there is probably some merit in describing them according to their chemical properties. Webster, for example, offers this classification:

Gases Nitric oxide Esters Acetylcholine Monoamines Catecholamines Dopamine Noradrenaline Indoles Serotonin Imidazoles Histamine

Amino acids Glutamate γ-aminobutyric acid (GABA) Glycine Peptides Endorphins Cholecyctokinin Substance P Steroids Pregnenalone Dehydroepiandrosterone

This classification system is technically accurate, in the sense that the categories are clear objective and non-overlapping. Unfortunately, these categories really tell you nothing about how these neurotransmitters behave or their influence on nervous system function. 

In any case the whole point of this is to demonstrate that there is no sensible way to classify neurotransmitters into functionally relevant non-overlapping categories, because the functions of neurotransmitters are so considerably overlapping, and to classify them chemically would be pointless because it would have no functional relevance. This is a defence of the totally arbitrary and unscientific classification system used below, which is based on nothing more than the importance of a neurotransmitter to the intensivist,  grouping them into "interesting" and "not". 

• Glutamate
• GABA 
• Acetylcholine
• Noradrenaline
• Dopamine
• Serotonin
• Histamine

Glutamate and the NMDA receptor

This is probably the most important "excitatory" neurotransmitter, well covered by Zhou & Danbolt (2014) and Fonnum (1984). Textbook authors tend to emphasise its importance by describing it as "the most prevalent amino acid in the brain", or quoting some sort of concentration value, which they say is the highest for any neurotransmitter or amino acid. Sure, that is true - Arne Schousboe (1981) is often quoted as saying that the brain contains 5-15 mmol/kg of glutamate, which would be 0.7-2.2g - but it is probably irrelevant to its neurotransmitter function, or to its "importance", however you want to measure that. Certainly, glutamate neurotransmission seems essential for life; for example knockout mice who lacked a critical subunit of the NMDA channel complex all died within about 15 hours of birth, apparently from respiratory failure (Forrest et al, 1994). 

The reason glutamate is present in such massive amounts is probably more to do with its importance to cerebral energy metabolism than to its neurotransmitter role. Together with GABA, glutamate is involved in the GABA shunt, a reaction that allows the neuron to bypass the α-ketoglutarate dehydrogenase step of Krebs cycle. Depending on who you read, this bypass accounts for 10-40% of the total neuronal TCA cycle activity. The main purpose of this shunt is to create GABA from α-ketoglutarate before it has a chance to get burned as a metabolic fuel, and glutamate is an intermediate step in that process. Observe:

Thus, you'd want to have plenty of glutamate and GABA around for this shunt to keep working properly. One other interesting feature of this otherwise boring diagram is the central role of the activated form of Vitamin B6 in the synthesis of GABA. From here, one might expect that any interference with B6 function or availability might give rise to some sort of pathologically hyperexcitable state, and that would be correct - isoniazid, the anti-TB drug, binds to B6 and prevents the synthesis of GABA, and refractory seizures are the result.

Glutamate receptors are numerous. Altevogt et al (2011) list about 20, each of which has multiple subtypes. They can be broadly separated into ionotropic receptors, which act basically like a ligand-gated ion channel, and metabotropic receptors, which are G-protein coupled and which express their effects more slowly (through gene transcription and protein synthesis). One never really hears about the latter, even though they are probably rather useful. Most attention in critical care exams tends to focus on NMDA, AMPA and kainate receptors, so named after the substances which selectively bind to them (eg. N-methyl-D-aspartate only binds to NMDA receptors and to no other glutamate receptor species). Its is probably also worth mentioning that NMDA receptors are somewhat unique in that they require two distinct molecular species to activate their pore opening. Glutamate is one, and glycine is the other.  

NMDA receptors are ionotropic. Specifically, you'd have to classify them as non-selective ligand-gated cation channels. When they bind glutamate, they open, allowing the traffic of sodium, calcium and potassium. Sodium and calcium enter the cell, and potassium exists concurrently; logically this should be expected to have a depolarising effect, except the membrane is usually already depolarised at this point. That is because of an added level of weirdness, often described as a "coincidence detector" function: these channels remain closed to ions even when they have bound their ligand, because their pore is constantly blocked by a magnesium ion. This ion is only dislodged from its position when the membrane of the post-synaptic neuron is depolarised.  Thus, the receptor is both voltage-gated and ligand-gated, with both factors necessary for it to function. Sure, it can serve a synaptic function, insofar as they could potentially allow the propagation of an action potential, but they are not critically important for synaptic transmission (Blanke et al, 2008) and their main role is in various interesting neuromodulatory second messenger effects.

Speaking of which; here are a few clinically relevant details about the role of glutamate neurotransmission, in case this ever comes up for any reason:

Role in pain transmission: in the spinal cord, with sufficient pain stimulus, the magnesium ion can be removed from the pore by various co-transmitters, for example substance P. This pain-related "recruitability" of  NMDA receptors results in a significant amplification of the pain response, and may underlie the analgesic effects of ketamine.

Role in excitotoxicity:  Because they permit an influx of calcium into the cell, the activation of ionotropic glutamate receptors can result in an overabundance of this intracellular messenger, producing mitochondrial membrane depolarisation, caspase cascade activation, the production of free radicals, and all sorts of other intracellular mischief. In short, a sudden unexpected calcium excess can really send a very strong proapoptotic signal to the neurons (Dong et al, 2009). This probably underlies the brain injury seen with uncontrollable status epilepticus

Role in cognition, learning and memory: without going into too much detail (for that we have Peng et al, 2011, and Gécz, 2010), it will suffice to say that glutamate seems essential for cognition, memory, learning, and probably consciousness in a broader sense. Reidel et al (2003) report extensive evidence to support this, most of which amounts to repeated findings that animals whose NMDA receptors are chronically blocked become really bad at exploring mazes.

Role as a differential for unexplained encephalopathy: for some unhappy reason, NMDA receptors are occasionally the target of autoantibodies which give rise to limbic encephalitis. Come to think of it, this really isn't a physiological role.

Pharmacological properties of glutamate are certainly never going to grant anybody a flawless pass mark in an exam, but some group of people may be interested in the possibility of getting high by eating or injecting this neurotransmitter. This community would be very disappointed with the effects. The sodium salt of glutamate, otherwise known as umami, will not appear in the bloodstream unless it is ingested in truly absurd doses (Fernstrom, 2018).This is because glutamate is immediately transformed into alanine in the small intestine, and so its bioavailability is rather poor.

The normal daily intake of an average Westerner eating a relatively normal MSG-flavoured diet is apparently 0.3-1.0g/d, which should not be expected to elevate serum glutamate levels even by one molecule. In fact, even truly humongous doses are eliminated on first pass.  Back in the 1950s when it was still possible to experiment on mental patients, Himwich et al (1955) tested the effect of escalating oral doses and found that "whether the patient is on placebo or 45 grams of glutamic acid per day, there is no observable difference in the glutamic acid remaining in the plasma". Apparently doses up to 150g/day are well tolerated even after fasting, even though the plasma concentration of glutamate does increase tenfold.  In mouse studies, it appears that around 4g/kg of subcutaneous monosodium glutamate (280g for a normal adult) is required to raise the serum concentration of glutamate to a dangerous level. At this dose, neurotoxicity is observed in the hippocampus (Takasaki, 1978): the neurons become oedematous and begin to degenerate, likely due to the development of excitotoxicity. As far as a lazy Google search can reveal, there are no human case reports of large-scale glutamate toxicity.

GABA , γ-aminobutyric acid

GABA is said the be the most important inhibitory neurotransmitter, which is probably accurate irrespective of which way you look at it. GABA is present in high concentrations in the CNS, it seems important to cerebral energy metabolism, its receptors are important drug targets, and the relative lack of GABA gives rise to life-threatening unpleasantness (eg. the relentless seizures mentioned above in the context of isoniazid overdose).

GABA receptors come in two major distinct varieties, the GABA_A and GABA_B. They are well covered in a review by Enna Salvatore (2007). GABA_A receptors are rapidly acting ligand-gated chloride channels, best known for all the sedative drugs which they acts as the target for. GABA_B are slow metabotropic receptors, coupled to G proteins, and their action leads to an overall increase in  membrane potassium conductance. Both receptors lead to a hyperpolarisation of the membrane: both potassium and chloride have a similar (very negative) Nernst potential, usually more negative than the resting membrane potential. Letting those ions equilibrate across the membrane would drag the overall resting membrane potential in the direction of a more negative resting voltage.

On this basis, GABA is often referred to as an inhibitory neurotransmitter. When GABA affects a neuron, the synaptic site becomes hyperpolarised, and therefore more ion current is required to bring its membrane potential back to the firing threshold. This has a net inhibitory effect if that neuron is responsible for some sort of excitatory action. If the neuron is inhibitory, the inhibition of inhibition is disinhibition, i.e. it permits excitation. 

Role of GABA_A receptors for sedation is central. A GABA_A receptor could say, without false modesty, that they are sedation. The majority of drugs that intentionally or unintentionally produce a decreased level of consciousness act by modifying the actions of these receptors. Notable examples for which the mechanism of action is known to include benzodiazepines and barbiturates, and suspected to include propofol, volatile halogenated ethers, and a whole host of party drugs.

Role of GABA_B receptors for spasticity and epilepsy is well established. GABA_B receptor agonists like baclofen have a muscle relaxant effect. Tiagabine and vigabatrin are GABA_B receptor agonists which act as antiepileptic drugs. 

Relevance to horrible diseases is again not a physiological role, but one is still tempted to note that GAD (glutamic acid decarboxylase) receptor antibodies are responsible for stiff person syndrome, which is the consequence of GABA deficiency, and which is managed by supplementing GABA receptor agonists (Levy et al, 1999).

Pharmacological properties of GABA are also not relevant to exam preparation. It is enough to mention that, like glutamate, people had considered the possibility that consuming raw GABA might have some sort of CNS effect. Hepsomali et al (2020) document the rich history of consuming oral GABA in hope that it might act as some sort of sedative or anxiolytic (spoiler: it does not work), and Nurnberger et al (1986)  reports one of the few studies which described the effects of injecting GABA intravenously (it made the subjects depressed).

Acetylcholine

Acetylcholine is the most important neurotransmitter in the peripheral nervous system. Specifically, it is involved in:

• Motor neurotransmission, at the neuromuscular junction (nicotinic receptors)
• Preganglionic autonomic neurotransmission, for both the sympathetic and parasympathetic autonomic nervous system branches, acting on muscarinic receptors
• Postganglionic parasympathetic nervous system, acting on muscarinic receptors
• Postganglionic sympathetic nervous system, acting on muscarinic receptors specifically at sweat glands
• Central nervous system, predominantly among interneurons, where it appears to act as a major modulator of consciousness and awareness (Perry et al, 1999)

Acetylcholine is dealt with in more detail in a series of dedicated chapters from the series on the autonomic nervous system, starting with the one on synthesis and metabolism of acetylcholine. Without revisiting a lot of that information, it is acetylated choline. Shortly after it is released into the synapse, it is broken down by acetylcholinesterase, which is the main mechanism of its removal from the active sites, and a fashionable drug target for CICM exam vivas.

Acetylcholine receptors come in a variety of flavours, of which the majority are G-protein coupled:

• Muscarinic receptors:
  • M1: G_q protein coupled, second messenger is IP₃,  
    main mechanism is to increase intracellular calcium
  • M2: G_i protein coupled, inhibit adenylyl cyclase, decrease cAMP
  • M3: G_q protein coupled, second messenger is IP_3, 
    increase intracellular calcium
  • M4: G_i protein coupled, inhibit adenylyl cyclase, decrease cAM
  • M5: G_q protein coupled, second messenger is IP_3, 
    increase intracellular calcium
• Nicotinic receptors:
  • Ligand-gated cation channels, vaguely selective for sodium

There is no reason whatsoever to remember the muscarinic receptor subtypes and their various secondary messenger molecules. It is probably more important to keep in mind that nicotinic receptors are channels which are supposed to act as sodium channels, but which in reality permit the bi-directional flow of all cations, leading to embarassing potassium leakage (this is the basis of suxamethonium-induced hyperkalemia, another favourite topic for viva questions). Nicotionic neurotransmission is therefore a clearly "excitatory" phenomenon, as the opening of these acetylcholine channels will lead to the depolarisation of the post-synaptic neuron.

For muscarinic receptor subtypes, the effect is more subtle, and more difficult to classify in this simplistic way.  The excellent paper by Michael Caulfield (1993) is a solid resource for those who want to read more about it. In the briefest possible summary, the effect of their second messenger systems will clearly do something to the excitability of that postsynaptic neuron, which could be a decrease or an increase in the excitability, or even a depolarisation, but this occurs by numerous mechanisms which are not easy to nail down and which may be different for each neuron species affected. For example, Lin et al (2004) looked at some M1 receptors in rat nigrostriatal neurons and found that the activation of these receptors had an excitatory effect, depolarising the cells probably by affecting potassium conductance. In contrast, Busse et al (1988) found that rabbit aorta endothelium seemed to become hyperpolarised following exposure to acetylcholine, which seemed to be a calcium-mediated process. Similarly, M2 receptor stimulation hyperpolarised the neurons of the rat nucleus parabrachialis (Egan & North, 1986). In short, the effect of acetylcholine can be either excitatory or inhibitory, and either effect seems to be the result of changes in potassium and calcium conductance.

For the intensive care trainee at the early stages of their training, the relevance of acetylcholine receptors lies in their following roles:

Role in neuromuscular junction: acetylcholine receptors here are competitively inhibited by non-depolarising neuromuscular junction blockers, and propped open (and therefore also blocked, but in a different way) by depolarising neuromuscular junction blockers. Acetylcholinesterase inhibitors act here to increase the presence of acetylcholine, which will reverse the effects of competitive NMJ blockade.

Role in autonomic nervous system: this is discussed elsewhere, but in summary that role is critical, as basically the entire autonomic nervous system runs on cholinergic neurotransmission. From a practical standpoint, anticholinergic side effects are common ward problems that lead to hilarious ICU referrals for things like tachycardia, decreased level of consciousness and anuria (because urinary retention).  Autonomic effects of anticholinergic drugs are a popular toxidrome for exam writers, as are cholinergic toxins (eg. organophosphates). 

Role in consciousness and alertness: Acetylcholine in the CNS does not seem to act in the same way as it acts in the periphery. Instead of mediating instant nerve-to-nerve transmission, instead central cholinergic neurons seem to act as neuromodulators, changing the excitability of other neurotransmitter pathways (Picciotto et al, 2012). Without going into too much detail (not that it's understood particularly well even by experts), one can broadly say that cholinergic neuromodulation has a major role to play in things like attention, affect, hunger, memory deposition and the overall level of alertness. Acetylcholine seems to be one of the more important players in the regulation of the level of consciousness by the brainstem reticular formation. However, it goes without saying that these statements are mostly speculative, as we do not know what consciousness even is, let alone how the interplay of neurotransmitter systems influences it, or which neurotransmitter system plays thedominant regulatory role. Hal Blumenfield's 2012 textbook chapter on the neuroanatomical basis of consciousness discusses the role of all neurotransmitter molecules in the control of arousal, in case the reader is interested. 

Role in delirium and toxicology: Anticholinergic effects are common among CNS-acting agents, particularly old-school antipsychotic drugs. As already mentioned, cholinergic neurotransmission is involved in the regulation of alertness, and therefore anticholinergic influences will interfere with this and produce delirium.

Noradrenaline

Noradrenaline is a catecholamine neurotransmitter known to ICU trainees mainly because of its peripheral cardiovascular effects. A whole messy chapter is dedicated to the properties of noradrenaline as a vasopressor. As a neurotransmitter, noradrenaline has peripheral autonomic (sympathetic) and central neuromodulator functions. Specifically, it is released from postganglionic sympathetic neurons (whereas all the rest of the autonomic nervous system runs on acetylcholine), and from noradrenergic nerve terminals in the CNS, which are involved in numerous activities. Additionally, noradrenaline is released into the circulation from the chromaffin cells in the adrenal medulla, but really you could hardly call that neurotransmission, as it is meant to reach various cardiovascular destinations and not nerve terminals.

Noradrenaline receptors are all metabotropic:

• α₁ receptors: G_q-protein coupled, second messenger is DAG and calcium
• α₂ receptors: G_i protein-coupled, presynaptic; second messenger is cyclic AMP
• β₁ , β₂ and β₃ receptors: G_s-protein coupled, second messenger is cyclic AMP

The downstream effects of activating these catecholamine receptors in the peripheral nervous system are discussed elsewhere. The most important point for this chapter would be to note that none of these receptors have a direct effect on the resting membrane potential of other neurons, and therefore cannot be classified as "excitatory" or "inhibitory". In spite of this, noradrenaline is usually classified as an excitatory neurotransmitter, probably because it is also involved in the regulation of alertness and attention.

What happens when a noradrenergic neuron connects to another neuron, instead of a cardiac myocyte? That probably depends on which neuron and which receptors. For example, when Grzelka et al (2017) looked at the noradrenergic projections from the locus coeruleus to the pyramidal neurons, they found that β1 receptor activation produced an inward sodium current, which had a depolarising influence (it was carried by some of the same cyclic nucleotide-gated channels that you might have expected from the aforementioned cardiac myocyte). But when Madison & Nicoll (1985) looked at noradrenergic neurotransmission in the rat hippocampus, they found that activation of α-receptors resulted in the hyperpolarisation of the postsynaptic membrane, which seemed to be a mechanism designed to increase the signal-to-noise ratio in the neurotransmission through that system. 

The physiological roles of noradrenaline as a neurotransmitter can be oversimplified as follows:

Role in sympathetic nervous system: though comparatively minor in comparison to acetylcholine, noradrenaline has pharmacological importance for the neural autonomic control of heart rate and blood pressure, as well as thermoregulation, bladder relaxation, and several other maintenance functions. The noradrenergic arm of the autonomic nervous system is a convenient lever for the manipulation of those functions.

Role in alertness and attention: Noradrenergic neurons in the locus ceruleus of the rostral pons
and the lateral tegmental area project all over the place, most notably to the cortex and hippocampus, and appear to regulate the level of arousal and sleep-wake cycles. Certainly, CNS-active drugs which promote noradrenergic neurotransmission by interfering with its uptake or by direct receptor action (eg. amphetamines) tend to increase arousal and decrease the demand for sleep, whereas drugs which inhibit noradrenergic neurotransmission (such as clonidine and dexmedetomidine, presynaptic α2-agonist agents) tend to decrease alertness. In case you are interested, CNS receptors involved in this appear to mainly be of the α1 subtype (Blumenfield, 2012). 

Role in pain and analgesia: Clonidine dexmedetomidine and to a lesser degree moxonidine have a co-analgesic effect which is mediated though their presynaptic α2-agonist effects. This effect appears to be originate from the modulation of noradrenergic neurotransmission in the superficial layers of the dorsal horns.

Dopamine

Noradrenaline is also a catecholamine neurotransmitter that will be better known to ICU trainees because it has historical relevance for the treatment of shock, and because it remains a widely available vasopressor/inotrope, still in clinical use worldwide. The synthesis and structure of dopamine, as well as its catecholamine daughter molecules, is discussed better in the chapters about the autonomic nervous system. There are multiple dopamine receptors, all of which are metabotropic:

• D1: G_s-protein coupled, increase cAMP
• D2: G_i protein coupled, inhibit adenylyl cyclase, decrease cAMP
• D3: G_i protein coupled, inhibit adenylyl cyclase, decrease cAMP
• D4: G_i protein coupled, inhibit adenylyl cyclase, decrease cAMP
• D5: G_s-protein coupled, increase cAMP

So, depending on which receptor system is activated, the effects of dopamine could be excitatory or inhibitory. The increase or decrease in intracellular cyclic AMP could also lead to conflicting membrane effects. For example, D1 receptor secondary messenger signalling produces an inhibition of Na⁺/K⁺ ATPase activity, which should have a depolarising effect by making the resting membrane potential less negative (Pirovarov et al, 2019). D2 effects, on the other hand, seem to be mainly "inhibitory", as they produce a hyperpolarisation of the membrane by the modulation of voltage-gated sodium channels (Maurice et al, 2004). Moreover D2 receptors are available in presynaptic and postsynaptic flavours, each with distinct effects (presynaptic ones, for example, inhibit further dopamine release by a negative feedback mechanism, whereas postsynaptic ones can actually enhance downstream dopamine release). In short, you can't call dopamine an excitatory or an inhibitory neurotransmitter. It's just not that simple. 

However, when one thinks of scenarios where dopamine neurotransmission is in excess, one thinks of a patient cavorting around their room in the nude. For the intensivist, the most important physiological roles of dopamine neurotransmission are related to its effects on movement, arousal and behaviour. In summary:

Role of dopamine in movement relates to the regulatory role of dopamine in the nigrostriatal region of the basal ganglia. From here, projections to the rest of the basal ganglia are involved in the control of voluntary movement and coordination. Interference with dopaminergic neurotransmission here can lead to a paucity of movement (Parkinson disease) or its excess (tardive dyskinesia).

Role of dopamine in alertness and consciousness: Dopaminergic neurotransmission is crucially important to all sorts of complex prefrontal functions such as learning, working memory, aggression, emotion, motor control, and food intake. 

Role of dopamine in goal-oriented behaviour: Dopaminergic neurotransmission, particularly projections from the ventral tegmental area to the nucleus accumbens, is implicated in the stimulus-response arcs associated with addiction. Dopaminergic drugs of abuse influence this system in complex ways which are incompletely understood, but which seem to affect limbic pathways involved in complex behaviours associated with reward and punishment. 

Neuroleptic malignant syndrome is not exactly a physiological role, but needs to be mentioned here at least in passing as an interesting toxidrome. In short, this is a syndrome of decreased consciousness with classical motor manifestations, characterised by muscle rigidity and fever. 

Histamine

This is another monoamine, made up of an imidazole ring connected to an amine group by an ethylene group. Its effects in inflammation and gastric acid secretion need to be mentioned, but are completely unrelated to its neurotransmitter action. Both histamine receptors (H1 and H2) are metabotropic:

• H1 receptors: G_q protein coupled, secondary messengers are IP3 and calcium
• H2 receptors: G_s-protein coupled, increase cAMP

The membrane effects of H1 receptors seems to be excitatory in every sense: their activation leads to a depolarisation of the membrane through changes in potassium conductance, and they inhibit hyperpolarisation.  H2 receptors also seem to have some sort of excitatory effect, as they seemed to prevent the long afterdepolarisation in cortical and thalamic neurons, reducing their refractory period (Haas et al, 1989).

Role in wakefulness and sleep: the histaminergic neurons of the hypothalamic tuberomamillary nucleus seem to be heavily involved in the regulation of the sleep/wake cycle. When they are active and firing at their normal rate, they release histamine, which binds to H1 receptors and increases wakefulness (Sherin et al, 1998). H1 receptor antagonists therefore logically promote sleep. 

Regulation of hypothalamic functions: Pretty much everything the pituitary gland does is mediated by histaminergic neurotransmission from the hypothalamus (Knigge & Warburg, 1991). The release of vasopressin, ACTH, LH and prolactin all seem to be controlled by histaminergic neurons. 

Serotonin

Serotonin, another monoamine, is not just a neurotransmitter, and is found in all sorts of places (platelets, etc). Its contribution to inflammation and clotting will not be mentioned here. The neurotransmitter effects of serotonin are mediated by a huge receptor family of seven different types, each of which has further subtypes. There is probably no merit in describing each and every serotonin receptor subtype, as this would lead to depression. Instead of being tortured with extensive discussions of structure and function, the reader will be spared a detailed discussion, and is instead shown this diagrammatic representation of the different serotonin receptor families, so they know what to be grateful for:

These are all metabotropic receptors, with the exception of the 5-HT₃ family which are ligand-gated cation channels and which mainly conduct depolarising sodium and potassium currents. There's a lot more detail about all this in Kroeze et al (2002), in case the reader needs to know exactly which secondary messengers are used where, but it is not essential for any real-life purpose, unless one plans to become a dedicated serotonin specialist. For the ICU trainee, the main physiological roles to know about consist of the following list:

Role in the regulation of mood: Serotonin is famously involved in the mind-bogglingly complex systems which give rise to the overall behavioural trend we describe as our "mood", though what exactly is meant by this has escaped better authors. Mitchell & Phillips (2007) is a good safe reference that focuses on the executive effects of mood and the effects of serotonin on them. It appears the 5-HT_2A receptor subtype is the most important.  From the perspective of the intensivist, the greatest interest for this lies not in the  mood-related neurotransmitter effects per se, but rather in the side effects of the drugs which are used to treat mood disturbances, and in their toxicology.

Role in sensation of nausea:  The sensation of nausea seems to be mediated by 5-HT₃ receptors. The key area appears to be the chemoreceptor trigger zone, a dorsal medullary structure with the unenviable task of constantly sampling the bloodstream for nauseating substances. These receptors are nonselective ion channels, and depolarise the postsynaptic membrane. Their most exciting feature is their sensitivity to antiemetics like ondansetron granisetron and dolasetron, which act as  5-HT₃ receptor antagonists, interrupting the nauseating neurotransmission and decreasing the central perception of nausea.

Role in the control of intestinal motility: The same 5-HT₃ receptors are also involved in all sorts of autonomic (and automatic) functions of the enteric nervous system, particularly involving functions like motility and secretion. Browning (2015) has a good review of this, in case it peaks your interest for whatever weird reason. In short, its relevance is mainly related to the side effect profile of antiserotonergic antiemetics, which are all known to cause constipation.

Role in thermoregulation: Serotonin receptors, particularly 5-HT₇ receptors which are abundant in hypothalamic nuclei, play a role in  the regulation of body temperature. Schwartz et al (1995) designed and conducted a nightmare experiment, somehow convincing sixteen healthy subjects to be blindfolded and have their rectal temperature measured while having regular infusions of the serotonin receptor agonist m-CPP, a highly potent panic-inducing hallucinogen. Rectal temperature increased in association with this abuse.  "It appears that m-CPP activates a mode of metabolic thermogenesis governed by a nocturnally sensitive proportional control mechanism" the investigators giggled. They published in Nature. Taking advantage of the obvious pun, the bottom line here is that serotonergic neurotransmission is involved in the central regulation of body temperature, and drugs which increase the availability of serotonin at these receptors will produce hyperthermia.

Role in the management of migraine: Though there does not seem to be a well-accepted physiological mechanism for this, serotonin is clearly implicated in the pathogenesis of migrain headache, at least insofar as serotonin receptor agonists seem to have an antimigraine effect (Hamel, 2007). Specifically, 5-HT_1B and 5-HT_1D receptor agonists are of interest to the migraineur. 

So, in short, if one were ever asked to "classify the 5HT receptors and give examples of pharmacological agents that affect them", one could regurgitate this table. In case anybody is wondering where any of this information comes from, it was "xPharm: the Comprehensive Pharmacology Reference" (2009), by Ennam & Bylund. In this encyclopaedic work, each receptor subtype has its own short chapter.

Mechanism of serotonin syndrome

Question 1(p.2) from the first paper of 2010 asked for "the physiological basis of the effects seen in the serotonin syndrome", which appears to be a question none of the trainees could answer. This makes perfect sense, as the international scientific community is still unclear on this, and Francescageli et al, writing in 2019, did not report any clearly established molecular mechanisms. The college answer itself is also weirdly lacking in detail, offering irrelevant factoids like "important neurotransmitter, a local hormone in the GIT and involved in platelet reactions" and listing some clinical features but not explaining them other than to say that "in serotonin syndrome these effects are probably CNS mediated". This disappointing performance by both the trainees and the examiners highlights one important point which should be comforting to the trainee preparing for the CICM primaries: neither the candidate nor the faculty knew how to explain serotonin syndrome physiologically, and both groups are now successful intensivists, i.e. this does not seem to have been much of an impediment to any of their careers.

Also, ondansetron is a 5-HT₃ antagonist, and therefore should not be expected to produce serotonin syndrome (Rojas-Fernandez, 2014)

If one were trying to explain serotonin syndrome in ten minutes or less, one would be forced to take some shortcuts. One would certainly not be able to reproduce the spectacular table of serotonin syndrome causative agents from  Francescangeli et al, 2019. Instead a shorter leaner version of the same thing is offered here. Wherever one sees a confident statement to the tune of "...and this receptor subtype is responsible for this effect", the source for the information is Deka et al, 2020. 

• Specific serotonin receptors are involved:
  • 5-HT_1A and 5-HT_2A receptors are the most responsible
• Stimulation of these receptors exceeds safe thresholds when:
  • Serotonin synthesis is increased (eg. dietary tryptophan, soft cheese etc.)
  • Serotonin release is enhanced (pethidine, dextromethorphan)
  • Serotonin reuptake is inhibited (SSRIs, MDMA)
  • Serotonin metabolism is impaired (MAOIs, linezolid, methylene blue)
  • Serotonin receptor agonists are present (eg. sumatriptan, LSD, fentanyl)
  • Or any combination of the above.
• The result of this hyperstimulation is:
  • Altered function of other neurotransmitter systems for which serotonin is a neuromodulator (usually, an inhibitory influence)
  • For the majority of these, serotonin excess leads to the disinhibition of presynaptic release of these mediators
  • They include noradrenaline, acetylcholine, dopamine and glutamate
• Clinically, this manifests as:
  • Confusion (possibly due to the effect on cholinergic neurotransmission)
  • Pupillary dilatation (a 5-HT_1A effect)
  • Hypertension (a 5-HT_2A effect)
  • Hyperthermia (a 5-HT_2A effect)
  • Diarrrhoea (the effect of increased serotonin activity on the GIT)
  • Muscle rigidity and clonus (probably related to dopamine disinhibition)