Physiology of the neuromuscular junction and its receptors

This chapter is most relevant to Section L1(ii) from the 2017 CICM Primary Syllabus, which expects the exam candidates to "describe the physiology of the neuromuscular junction and its receptors". This is probably a fairly reasonable expectation, as this knowledge is a necessary precondition for understanding how neuromuscular junction blockers work, and those drugs occupy a tremendous amount of examiner attention in past papers. For the neuromuscular junction on its own, there has only been one specific SAQ - Question 17 from the first paper of 2019, which asked us to "explain the physiology of neuromuscular transmission". So:

Understandably, this is a topic well supplied with academic interest from both ICU and anaesthesia, and everywhere you turn you find an excellent review, short study-oriented summary or a deep dive into the smallest detail. It is not clear what the following entry could possibly be expected to add to this overabundance of resources, other than to focus the discussion in a way that answers the CICM examiners' impression of an ideal answer to Question 17. For the reader interested in this topic for some reason other than the Australasian College of Intensive Care Medicine primary exam and unwilling to read any more self-indulgent garbage from Deranged Physiology, the best quick reference would probably be this easily digestible paper by King & Hunter (2002). It is, unfortunately, almost entirely without academic references, and is therefore no better than the relevant chapter from any undergrad textbook. A better-referenced alternative would be Ruff (2003), and it is still mercifully brief, at under seven pages. Nobody could possibly need more than this, but if they did, they would probably find it in Engel (2007) from the 91st volume of Handbook of Clinical Neurology. Whenever some factoid is dropped without a reference in the text below, it will have come from the Engel paper.

Axonal transmission down to the terminals

Each motor axon is a thick heavily myelinated structure, leveraging the benefits of saltatory conduction to increase the speed of signal propagation to some kind of natural maximum, at least for vertebrates. Figures of 48-120 m/s are quoted (eg. Brown, 2001). This speed is maintained all the way down to the ends of the terminals, where the motor axon branches extensively. How many branches sprout from each axon depends on the size of the motor unit, eg. for a large muscle expected of gross imprecise movements such as the gastrocnemius each axon might spread into about a thousand terminals.

Each muscle fibre usually receives only one such connection, usually near their centre. There are exceptions receiving distributed innervation, where there are several neuromuscular junctions on the same fibre spaced out at regular intervals, but this is unusual among mammals (for example, in humans, this sort of connectivity is seen only in some of the muscles supplied by cranial nerves, such as laryngeal and extraocular muscles). In vertebrates who are occasionally derogatorily referred to as "lower", such as lampreys, there is also the option of "myoseptal" innervation, where the motor axon ends in a weird basket-like terminal near the insertion of the muscle.

At reading this, the reader may wonder "where, then, are my neuromuscular junctions?" To answer this question, Van Campenhout et al (2011) were actually able to map them using a cholinesterase-specific stain. In the images below, which were shamelessly stolen from their paper, the spots identify the motor endplates they had found, and the lines give representation to the direction of the fibres. 

To maintain speed until the very end, the motor axon loses myelination only at its very tips. Here, a representative electron microscope image from Engel (2008) has been colourised to change (improve?) its explanatory power:

The myelin ends at a node of Ranvier, leaving the action potential to walk the rest of the way - but it should not be far, as the magnification of the synapse depicted above is × 30,600 and everything is happening on a tiny scale. For Coërs (1967), who mapped these things back when they were called the "myoneural junctions", the innervated area of each motor nerve terminal ranged from 10 to 88 μm, with an average of around 33 μm.

Structure of the neuromuscular junction

The unusually prescriptive answer to Question 17 from the first paper of 2019 specified that "well-constructed answers ... elucidated the structure of the neuromuscular junction". The term "elucidate" of course means "to make clear", deriving from "lucid", and it is ironic that the examiners chose to use this specific word to obscure what they wanted from the people reading their comments, to some of whom the English language may not be native. Their other point, that this is "best done with a detailed diagram", is still however valid. Words fail when trying to describe something as complicated as this succinctly, and if you had to sum up the neuromuscular junction in a sentence, you'd probably ramble for several. The best short statement to describe this thing is probably the opening line from Engel (2007):

"The neuromuscular junction (NMJ) is a chemical synapse that is anatomically and functionally differentiated for the transmission of a signal from the motor nerve terminal to a circumscribed postsynaptic region on the muscle fiber."

It feels like anything shorter would result in an unacceptable loss of information. Thankfully there are a million possible ways to draw this structure, and it does not require even the slightest artistic talent. The most important thing is that the essential features are labelled. These would probably have to be:

• Presynaptic membrane
• Acetylcholine vesicles
• Postsynaptic membrane
• Nicotinic receptors
• Acetylcholine (perhaps an action shot, pouring out of the vesicle)
• Acetylcholinesterase

Optionally, also:

• The myelin sheath of the axon and the last node of Ranvier
• The Schwann cell enveloping the terminal
• A sheath of endoneurium that surrounds them all

One occasionally sees amateur NMJ artwork embellished with mitochondria and endoplasmic reticulum, but drawing these in an exam would waste precious seconds, and would not contribute to the explanatory power of the illustration. However, one feels the need to point out that these nerve terminals are definitely packed with subcellular structures of all sorts - for example mitochondria account for approximately 15% of the nerve terminal volume. 

The neuromuscular presynaptic membrane

The presynaptic membrane of the nerve terminal and the immediately underlying cytosol are home to teeming swarms of neurotransmitter vesicles that contain ready acetylcholine. Below, an image so ubiquitous it is impossible to track its origin demonstrates these multitudes, queueing to empty into the gap.

Along the membrane itself, groups of vesicles are "docked" to the membrane, forming "active zones" where the neurotransmitter is readied for release. To give some sense of their numbers, in the mouse each nerve terminal has perhaps 700 active zones, spaced about 500nm apart, and each active zone has 2-3 docked vesicles. There should be about  50–70 synaptic vesicles per μm² of the nerve terminal area, and each terminal releases about 160 vesicles with each action potential.

The neuromuscular junction gap

Otherwise referred to as the synaptic space, this is the extracellular fluid compartment that divides the presynaptic and postsynaptic membranes. It is usually about 70nm wide, and - though the diagrams in textbooks depict it as an empty void for acetylcholine molecules to traverse - it is filled with all kinds of stuff, most of which is broadly referred to as the "synaptic basal lamina" because this proteinaceous gunk shares a lot of its composition with the sort of normal basal lamina you find elsewhere. Well, perhaps "proteinaceous gunk" is an unfairly libellous description of this material, as it makes the place sound like a crack between bathroom tiles. The complex and busy content of the synaptic gap deserves a more respectful treatment, such as this gorgeous candy-coloured molecule orgy from Goodsell (2009):

The most distinctive inhabitant of this space is the acetylcholinesterase enzyme, appearing with a density of 2000 to 3000 molecules per every μm² of synaptic area. The other members include agrin, perlecan, collagen, laminin, entactin and nidogen - some probably sounding unfamiliar to the ICU trainee, but all of them are rather ubiquitous structural macromolecules (Kefalides & Borel, 2005). 

This space extends in all directions from the synapse, and has no specific boundary other than the loose and fuzzy margins of the basal lamina proteins, which means it it functionally continuous with the extracellular fluid. Theoretically, this means there could probably be some spillover of acetylcholine from one synapse to some neighbouring synapse if the conditions were right (eg. where the acetylcholinesterase was inhibited or deficient). That would lead to the unexpected activation of nearby muscle fibres, or even sensory fibres for that matter. Nervo et al (2019) demonstrated that this is probably responsible for some of the toxicological phenomena seen with acetylcholinesterase inhibitor toxicity, such as fasciculations and respiratory depression.

The motor endplate

Though occasionally used interchangeably with "neuromuscular junction", the term "motor endplate" seems to specifically mean "the post-synaptic membrane" for some authors, and because taxonomists recoil in horror from having two names for the same thing, we will side with the latter group for this chapter. They call it an "endplate" which implies flatness, but in fact this thing is furrowed with deep fjords, and the "plate" thing probably comes from the overall flatness of the nerve terminal structure, which in mammals typically looks a bit like the splayed suction cups at the tips of a frog's toes:

Trying to track the sources for Deranged Physiology is sometimes rather difficult. This image, for example, comes from an excellent online histology resource from the University of Delaware, which appears to be a time capsule of the early Web 1.0, written in raw HTML. Their server somehow allows access to the staff directories (files last modified in 2002), where the electron microscopy collection is listed under "Wags". We do not know who Wags is, but wags.gif in the same directory appears to represent an early stage in the life cycle of Dr Roger C. Wagner, Professor Emeritus of Biological Sciences, and so to him (we assume) we owe the excellent shots of the motor endplate above. 

The post-junctional membrane of the myocyte

To lift one of those flat frog toes reveals its footprint on the myocyte surface immediately below the synapse, a flattened plateau surrounded by deeply grooved valleys occasionally referred to as the "synaptic gutter". The best representation of this is again seen in Wagner (2002?), as well as Engel (2008):

These deep folds in the postjunctional myocyte membrane (unimaginatively referred to as "junctional folds" throughout the literature) are unique to the neuromuscular junction, and do not appear in any other synapse. They are deep enough to increase both the post-synaptic surface area and the volume of the synaptic space by a substantial amount. According to Engel, the postsynaptic membrane area is about ten times larger than the area of the presynaptic synapse.

Apparently the point of this extensive folding and expanded surface is to increase the reliability of signal transmission. Wouldn't want to miss even one of those precious acetylcholine molecules, it seems. Probably for the same reason the surface of the post-synaptic membrane is inundated with nicotinic acetylcholine receptors, with some authors quoting a density of 10,000 per μm². 

The sarcoplasm below the junction is also a busy place. This is where the acetylcholine receptors and synthesised and degraded, and it is also where acetylcholinesterase comes from. Of the many nuclei of a skeletal muscle cell, those that find themselves immediately below the synapse are typically specialised to transcribe mRNA for nicotinic receptor components and other kinds of membrane ion channels. None of this, of course, is of any interest or importance to the CICM trainees, who only need to know about the NMJ so they can learn to block it, but it was probably worth mentioning before we segue awkwardly from structure to function.

Presynaptic synthesis and storage of acetylcholine

Acetylcholine is made locally at the nerve terminal, from choline and acetate which are obtained from the extracellular and synaptic fluid. Choline availability is the rate-limiting step here, acetate being abundant. Choline acetyltransferase is the enzyme that performs this step. The flowchart of this synthesis pathway can be borrowed from the autonomic nervous system section, as it is common to all nervous system sites that deal in acetylcholine.

Free acetylcholine in the cytosol is then rapidly sequestered into vesicles by the actions of the vesicular acetylcholine transporter (VAChT) and trapped there by the low intravesicular pH (~ 6.7) maintained by a proton-pumping ATPase on the vesicular surface.

The vesicles then remain in storage, where about 10-20% of them are docked at active zones (with 1-2% ready for immediate release), and 80-90% of them are hovering in the cytosol nearby as a sort of reserve pool. The reason for this abundant redundancy is probably the theoretical lag built into the cyclic process of acetylcholine release degradation and re-synthesis. These steps take time. We mammals, being impatient beasts fond of sustained and vigorous action, cannot afford to wait for the synthesis of new molecules - we must have our acetylcholine right now,  which means a large number of vesicles must be held in reserve if the synapse is expected to still be available after multiple broadsides of exocytosis.

Yes, it is possible to deplete those stores, but it requires a lot of effort. For some examples of what it takes, we can look at experiments by Heuser & Reese (1973) who depleted vesicular stores of some frog sartorius muscles by continuous electrical stimulation at 10 Hz.  After fifteen minutes of electrotorture, the appearance of the synapses was examined using electron microscopy, revealing that the vesicles were depleted by about 60% when compared to controls. Sustained electrocution is not often seen in the ICU, but another more clinically plausible example comes from Clark et al (1972) who stimulated an avalanche of neurotransmitter release using the venom of a black widow spider. This substance (a-latrotoxin) forcibly increases the permeability of the presynaptic membrane to calcium, stimulating massive acetylcholine exocytosis. Which brings us to:

Calcium-mediated exocytosis of acetylcholine

Acetylcholine release is mediated by the entry of calcium into the presynaptic terminal. Neuromuscular neurotransmission is actually the best-studied version of your neurotransmitter release systems, mainly because the neuromuscular junction is so large and so easy to study, and so our understanding of the steps of neurotransmitter release is mainly generalised from this model. It therefore feels normal and fair to mindlessly cut-and-paste the relevant section from the neurotransmission chapter here to list the steps involved in calcium-mediated acetylcholine release. For actual original content, a diversion that combines detailed information with some delightful historical perspectives, the interested reader is redirected to an excellent paper by Dittrich et al (2018). 

In short:

• 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. They are the targets of the peptides found in the venom of Conus magnus, the giant predatory marine snail. 
• With the arrival of the action potential, the membrane depolarises, causing these N-type calcium channels to open. Apparently only about 20% of the available channels will open for any single action potential.
• 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.

To very briefly digress on the last step of this process, and only because it is sensitive to botulinum toxin and therefore potentially examinable:

Calcium-sensitive proteins that mediate NMJ vesicle fusion

Exocytosis machinery at the neuromuscular junction is actually just a mass-produced off-the-shelf solution mediated by the ubiquitous SNARE system.  SNARE somehow stands for "Soluble N-ethylmaleimide-sensitive factor Attachment protein Receptors", and they are a common choice of exocytosis mediator, ubiquitous across the plant and animal kingdoms. At the NMJ this complex links the membrane-docked vesicle to the outer cell membrane in preparation for release, and is capable of joining the two membranes with a latency of less than 100 microseconds (Sabatini & Regehr, 1996), which is close enough to being magically instantaneous.  It just so happens that critical care trainees learn about these mechanisms while reading about the neuromuscular junction, but in fact various SNAREs mediate all kinds of vesicle fusion all over the place, including all kinds of neurotransmitter release, the fusion of phagolysosomes and control of Golgi vacuole traffic. If you're a eukaryotic cell and you need to fuse two membranes, chances are you'll be using a SNARE of some sort.

So, the neuromuscular junction SNARE complex is not special. It just happens to be the one that comes up in ICU exams, and gets mentioned mainly in the context of botulinum toxicity, as the neuromuscular SNARE complex is the target of botulinum toxin (whereas the central nervous system SNAREs get attacked by tetanus). The paper by Han et al (2017) contains a lot more detail for the SNARE enthusiast, which is ultimately going to be unnecessary for most normal CICM people, as the only thing you really need to know about the NMJ exocytosis machinery is that it is activated by intracellular calcium. The relationship between the magnitude of calcium flux and magnitude of neurotransmitter release is unimaginatively called the CRR (calcium release ratio) and - at least in the frog - is extremely nonlinear, obeying a 4th power law. In other words, a doubling of the presynaptic intracellular calcium ion concentration results in a sixteen-fold increase in neurotransmitter release (Dittrich et al, 2018).

Quantum of neuromuscular neurotransmission

Very little calcium is actually required to release a single vesicle - theoretically the calcium-sensitive domains of the SNARE complex could be triggered by as few as six or eight atoms - and in fact there is constantly some calcium accidentally making it way through the presynaptic membrane, resulting in little random blips of acetylcholine release which Katz termed "miniature end plate potentials" (MEPPs). They are MEPPs or mEPPs, as the "M" is variably capitalised; in their original paper Fatt & Katz (1952) didn't feel the need to use that abbreviation at all. These tiny potential fluctuations always produced muscle motor endplate depolarisation in fixed increments of around 0.5 mV, and probably represent the absolute minimum value for acetylcholine release. One of these by itself is not enough to trigger muscle contraction, but several such MEPPs in summation definitely can, and in fact when de Castillo and Katz (1954) looked at the shape of the end plate potential during artificially weakened neurotransmission, they were able to determine that each potential was actually created by the summation of small identical potentials, or "quanta". This was most obvious when, by cooling the synapse down to around 2.5° C to slow the process of exocytosis,  Katz & Miledi (1965) observed that the voltage-over-time waveform of a single motor endplate potential was built up of many little step-like increments, more apparent because they were dispersed over a longer timeframe by the cold. The term "quantum" is occasionally used in reference to these concepts,  to describe the minimum amount of neurotransmitter that can be released from the presynaptic terminal, which is the "dose" of acetylcholine required to produce a single MEPP.  Estimates of the number of ACh molecules released during a quantum at the neuromuscular junction range from 8000 to 10000 (Van der Kloot & Molgó, 1994), probably representing the release of 2-4 vesicles. 

Synaptic reliability through prolific acetylcholine excess

One would likely not be satisfied with excessively economic neurotransmission, for example if each time only some minimum amount of neurotransmitter was released.  After all, only about 50% of the released acetylcholine actually makes it to the neuromuscular junction, as it needs to traverse a hostile space filled with high-affinity acetylcholinesterase enzymes. Ergo, if you really want that synapse to synapse, you would want to release a decent amount of neurotransmitter to ensure their reliability - probably several times more than the amount required to trigger a muscle contraction. This practice is usually referred to as the "safety factor", referring to the multiplier applied; for example in mammals this is usually 3-5, with 10-100 quanta released. This value also seems to vary depending on which muscle you look at, as can be seen from this table (Wood & Slater, 2001):

Mammals have a much lower safety factor as compared to other animals (eg. frogs) because of the increase in the area of the postsynaptic membrane. Where in frogs the postsynaptic membrane is only twice the area of the presynaptic membrane, the only option is to flood the synaptic space with neurotransmitter molecules. In contrast, in humans the difference in surface area is a factor of eight or ten because of extensive junctional folding which brings a lot more acetylcholine receptors into the same small volume, increasing the chances of receptor-ligand interaction and making the junction more reliable. 

The nicotinic acetylcholine receptor

The acetylcholine receptor at the neuromuscular junction is a member of a fairly large family of ligand-gated cation channels that open in response to acetylcholine. They are massively overrepresented in the literature, considering the importance and popularity of the NMJ as a model of a synapse, and everywhere you turn you find several excellent review articles at different levels of detail (for one example, this one by Bouzat & Mukhtasimova, 2018). For the CICM exam, a deep understanding of receptor structure and function is not essential. It is hard to know exactly how much detail is expected, but judging from past papers, the examiners are not especially fixated on the minutiae. The clinically relevant points which are the most likely to appear in exams are noted below:

• We call them "cation channels" because the pore is sufficiently unselective that you couldn't call it a strict sodium channel, even though that ion represents most of the traffic through it. Other cations can also slip through. Notably, potassium can escape from the cell just as easily as sodium can enter it, and in fact the permeability to potassium is actually slightly greater. The delicate reader may blush to learn just how whorishly promiscuous the nicotinic receptor channel really is. Apart from the normal monovalent sodium and potassium ions, it will gladly scandalise divalent cations such as calcium or magnesium, and is open to even more unnatural suitors (barium, caesium, rubidium) according to these data from Nutter & Adams (1995):
  
  For record, the P_x/P_Na here is the value of interest, and represents the permeability ratio of cation x to the permeability of sodium. As you can see the permeability of the nicotinic receptor to ammonium caesium and potassium is higher than its permeability to sodium. Why even mention this? Because suxamethonium, which fixes these channels in the open position, causes hyperkalemia by this exact mechanism (Martyn et al, 2006). The escape of potassium via neuromuscular junctions is sufficiently great during the effect of suxamethonium that the whole of the extracellular fluid is affected.
• This has greatest relevance in the scenario of the chronically immobile patient, or those who suffer from stroke or other illness that results in a large-scale denervation of muscle tissue. The compensatory proliferation of acetylcholine receptors causes them to experience a widespread and massive flux of intracellular potassium into the extracellular fluid, resulting in potentially life-threatening hyperkalemia. 

Unlike the G-protein coupled muscarinic receptors, nicotinic receptors are made from five subunits assembled into a barrel-shaped structure with a central water-filled pore. All nicotinic receptors are some combination of different subunits, which give them their different properties, among which (for example) is their affinity for nicotine. CNS receptors and autonomic receptors have a much higher affinity for it than those found at the neuromuscular junction. 

Which brings us to the next, most important question:

How much nicotine would I need to consume to achieve NMJ blockade?

An insane amount. The affinity of NMJ nicotinic receptors for nicotine is truly minimal, and huge concentrations of nicotine would be required to use nicotine as a depolarising muscle relaxant.  As an example, Paton & Savini (1968) found they needed to inject about 250-1000 μg directly into the cat iliac artery in order to achieve any sort of appreciable neuromuscular block in the downstream limb. The human literature for paralysing nicotine toxicity is thankfully rather sparse, but there are a few case reports of this occurring under various unusual circumstances.  For one example, Manoguerra & Freeman (1982) reported a case of a 76 year old male who picked and ate "a wild green plant that was growing in a vacant lot in Southeast San Diego", which turned out to be Nicotiana glauca - a tobacco-like plant that contains the nicotine-like alkaloid anabasine. He developed features consistent with nondepolarising neuromuscular junction blockade, including fasciculations, and required intubation overnight for respiratory failure. 

In general, nicotine toxicity is objectively fascinating. The casual reader will surely not begrudge, and the returning reader will downright expect, a long and rambling digression on this topic to emerge in the middle of this otherwise task-focused chapter. The literature on nicotine toxicity has historically been rather sparse because the normal way of consuming nicotine (smoking, chewing) has depended on the availability of nicotine in plant matter, where it is routinely derived from, and this does not seem to be a large enough amount to do you harm.  The leaves of the tobacco plant, Nicotiana tabacum, usually only contain 0.3-3% nicotine by dry weight, although some "heavy bodied" tobaccos can contain as much as 7% according to a 1998 review by Blakeley & Bates. That gives an average concentration of 3-30mg per gram. A commercial cigarette will usually contain about 0.5-0.9g of tobacco, and perhaps up to 16-20mg of nicotine. 

But is that a large dose? What represents a dangerous dose? Published answers to these questions are surprisingly obscure. Even toxicology literature tends to uncritically quote the  LD₅₀ from Goldfranks' Manual of Toxicological Emergencies, which is 0.8mg/g, and which seems preposterous, as this would be something like 60mg per person, or about four or five cigarettes. Many people would have had a grandfather for whom that would have been a standard morning routine over coffee, which makes it difficult to take this dose threshold seriously. Bernd Mayer (2014), in what is perhaps the best review of nicotine toxicity literature currently available, excavated decades of references and eventually determined that this 60 mg dose actually comes from some experiments by a German pharmacologist named Rudolf Kobert (1906), which he probably performed on himself, and which caused "dryness of the throat, coldness of the limbs, ructus [belch], flatulence, nausea, vomiting and rectal tenesmus". More modern literature suggests the dose required to cause harm is much higher. To summarise Mayer:

• A "strong" 16mg cigarette delivers perhaps 2mg of nicotine (12.5% bioavailability), because not all of its content will be inhaled, and not all of the inhaled content will be absorbed. Interestingly, because nicotine has an oral bioavailability of about 20%, it appears you'd get more of a hit by eating the cigarette. 
• Either way this produces a blood concentration of 0.03 mg/L  of nicotine.
• Postmortem studies suggest 2mg/L is probably the lower limit of lethality.
• This means one would have to consume about 0.5-1g of nicotine orally (equating to from 30g to 300g of dry tobacco plant matter), or smoke about 62 cigarettes simultaneously, to risk severe toxicity.

This is reassuring, in the sense that normal people would not accidentally do this, and abnormal people would find it so difficult to do this intentionally that they would hopefully give up half-way. However, the modern world seems to insist on presenting us with increasingly sophisticated upgrades of the Last Worst Thing, and highly concentrated sources of nicotine are paradoxically even more available today then they were during times of totally unregulated tobacco markets. Specifically, vape fluid often contains spectacularly large amounts of nicotine (up to 100mg/ml), and reports it with spectacularly inaccurate labelling (Miller et al, 2021, reports that the labels are basically never correct). Most of the lethal and near-lethal overdose literature seems to be related to people accidentally or intentionally ingesting this highly concentrated nicotine product, and other similar products (for example pesticides that use nicotine as the main ingredient).

What is the pathophysiology of nicotine overdose? Applying basic principles, the reader would readily conclude that it should cause ganglionic effects, cholinergic CNS effects, and neuromuscular effects, in that order of appearance with escalating doses. First it hits the peripheral nervous system where there are high affinity receptors and no diffusion barrier, then it goes after the central nervous system (when the concentration is high enough), and finally if the dose it truly humongous the concentration increases to the point where even the neuromuscular junction receptors bind it, even though their affinity is extremely low.  Borrowing from Brčić Karačonji (2005) and Paik et al (2018), the following clinical picture develops:

• Low dose toxicity: mainly autonomic (ganglionic) effects, which are therefore a mixture of sympathetic and parasympathetic:
  • Parasympathetic (cholinergic):
    • Abdominal pain (writhing peristalsis)
    • Salivation 
    • Diarrhoea
    • Bronchorrhoea and bronchospasm
    • Miosis
    • Nausea, vomiting (by directly triggering the chemoreceptor trigger zone, which means at least theoretically anticholinergic antiemetics would be effective against it)
  • Sympathetic (noradrenergic, looking like a sympathomimetic toxidrome)
    • Hypertension
    • Headaches 
    • Diaphoresis
    • Pallor (cutaneous vasoconstriction)
    • Tachycardia
    • Increased respiratory drive
    • Agitation and mania
  • The sympathetic effects seem to be dominant in low dose toxicity, but if the dose taken was reasonably large, the clinical picture is biphasic, and the next stage is completely opposite:
• Moderate dose toxicity: CNS depressant and cholinergic effects begin to dominate
  • Confusion, lethargy, somnolence
  • Seizures
  • Coma
  • Decreased respiratory drive
  • Paradoxic inhibition of ganglionic nicotinic receptors occurs, which is described by many authors (eg. Schep et al, 2009, and Paik et al, 2018), but not very well explained by any of them. To get to the bottom of this, one needs to delve into Volle's chapter from Pharmacology of ganglionic transmission ​​​(1980). In short, the initial depolarisation of ganglionic postsynaptic nerve terminals is replaced with hyperpolarisation, which makes them unresponsive to further acetylcholine. The mechanism appears to be at least partly due to the inactivation of sodium conductance by various  subsynaptic and extrajunctional nicotinic receptors. It persists longer than the initial direct agonist effect, and gives rise to these "paradoxic" ganglionic blocker effects, which look a lot like the effects of ganglion-blocking antihypertensive agents:
    • Inhibition of adrenal medullary catecholamine release
    • Bradycardia, including heart blocks
    • Hypotension (vasodilation)
    • Mydriasis
    • Bladder atony
• Severe toxicity: nicotinic musculoskeletal effects
  • Weakness
  • Fasciculations
  • Death from respiratory muscle paralysis  (or, intubation)
  • Rhabdomyolysis 

In addition to this, the modern intensivist needs to consider that the nicotine overdose will - these days - rarely present in the form of a person who consumed pure nicotine. Likely, they drank some kind of vape fluid, which will inevitably contain a bunch of toxic non-nicotine ingredients, of which Kim & Baum list but a few (ethylene glycol, formaldehyde and other aldehydes, acetone, acrolein, vegetable glycerin and various others). Some of these are toxic in their own right, and others are metabolised into toxic byproducts. All of them should theoretically cause a widened osmolar gap, and many would cause a high anion gap.  Another factor to worry about is the fact that nicotine interacts with a lot of things. Famously, it inhibits the metabolism of suxamethonium, which would be relevant if anybody still used it. 

How high can you go? What are the limits of human survival for nicotine overdose? It appears that they are truly massive. The largest reported nicotine overdose appears in a case report by Schmidt from 1931, where the patient apparently ingested 4g of Nicotinum purissmum by Merck (even in the 1930s you couldn't just buy this stuff off the shelf; the patient was working in a scientific laboratory and had access to 100% pure nicotine). It appears he vomited most of the ingested toxin, as he survived (albeit with seizures and coma) in spite of being in Wrocław during the 1930s, 20-30 years before the establishment of intensive care as a speciality. 

Endplate potential due to the opening of nicotinic receptors

Here's an excellent image of a typical endplate potential recording, from a paper that describes in some detail the process of experimentally recording such potentials in the course of undergraduate labwork (Zanetti et al, 2018). This one even has a couple of bonus MEPPs thrown in.

The resting membrane potential for the myocyte is usually said to be something like -70-90 mV, i.e. approximately the same as the motor neuron axon. It is also maintained by the normal interplay of electrical and chemical gradients, but unlike in the neuron, for myocytes the chloride concentration gradient plays a larger role. The reason for this is the need to maintain polarity in the face of repeated depolarisations. The myocyte leaks potassium each time the nicotinic channels open, and with enough of these events the potassium gradient would diminish, reducing the potential difference across the membrane and fixing it in a depolarised state (similar to the effect of suxamethonium). A chloride current assists the repolarisation of this membrane, and it is mediated by the CLCN1 channel, which - if mutated into poor function - produces myotonia congenita, a disease characterised by weakness, muscle hypertrophy and depressed reflexes.

After it is brought to a threshold by the opening of sodium-conductive acetylcholine-gated channels, the endplate depolarises suddenly and massively through the opening of voltage-gated sodium channels, also in a manner reminiscent of a neuron. 

Just like with all the other excitable tissues the potential overshoots zero by 30-45 mV, or to whatever the Nernst potential for the membrane is. A rapid recovery follows, led by rectifying potassium channels, and the potassium current quickly restores the resting membrane potential. 

The events that lead to excitation-contraction coupling

This tidbit is left here because "the events that lead to excitation-contraction coupling in skeletal muscle" were noted as a part of an "ideal answer" in the examiner comments for Question 17 from the first paper of 2019. Not excitation-contraction coupling itself, mind you; but "the events that lead to" excitation-contraction coupling. The task of unpacking this comment is frustrated further by the fact that the rest of the examiner comments list a series of physiological phenomena which all technically lead to excitation-contraction coupling. The last of that list is the endplate potential, suggesting that these "events" are something that happens between the generation of an endplate potential and the beginning of the calcium-triggered contraction.

If this interpretation is correct, then these events are easy to identify: they are the activation of the voltage-gated sodium channels in the skeletal muscle sarcolemma, the propagation of the action potential, and its transmission to the contractile apparatus via the T tubules. These events are not in any way special, and will be familiar to anybody who has any knowledge whatsoever of excitable tissues. In brief, the transient depolarisation at the site of the nicotinic receptor activates nearby voltage-gated sodium channels and the action potential propagates away from the neuromuscular junction much in the way it normally does in other excitable tissues. The main difference here is that it cannot stay on the surface of the myocyte, as all the contractile machinery is deep inside the muscle fibre, and so it dives along the T tubules to reach each and every sarcomere.

Degradation of acetylcholine

It would obviously do no good to have acetylcholine hanging around in the synapse where it could keep opening nicotinic receptors, and this is why we have high-affinity acetylcholinesterase enzymes present in a large concentration within the synaptic basal lamina. Human acetylcholinesterase is a monstrously rapacious 67 kDa hydrolase enzyme, with each molecule capable of shredding 25,000 acetylcholine molecules every second, a rate so fast that substrate availability by diffusion becomes the rate-limiting step of the reaction (Quinn, 1987). "Less than 1 millisecond" is the normal lifespan of an acetylcholine molecule in the synapse, a timeframe given by resources so reputable that they do not feel the need to offer a reference for this value.

Why mention the degradation rate of acetylcholine? Well: clinical relevance does exist wherever this rate is affected by drugs or disease, as excessive or insufficient persistence of acetylcholine has all sorts of interesting implications:

• Decreased rate of acetylcholine degradation can be harmful, eg. where depolarising NMJ paralysis results from organophosphate poisoning
• Decreased rate of acetylcholine degradation can be beneficial, eg. where acetylcholinesterase inhibitors such as neostigmine are used to increase the availability of acetylcholine to overcome competitive nicotinic receptor blockade (eg. at the end of an anaesthetic case)
• Increased rate of acetylcholine degradation is also not a picnic, and is hypothesised to be one of the aetiological factors behind memory loss in Alzheimer disease, which is how centrally acting acetylcholinesterase inhibitors such as donepezil are expected to improve cognitive function.