Neuromuscular Junction

The neuromuscular junction (NMJ) is one of the most widely studied examples of chemical synaptic transmission in the nervous system. The overall concept is deceivingly simple: an action potential at the presynaptic neuron releases a neurotransmitter that in turn activates ion channels on the muscle cell membrane, resulting in postsynaptic action potential. Yet, every step in this process is exquisitely controlled and modulated in health and may be disrupted in disease. Our evolving understanding of the NMJ provides a critical window into synapse function and pathophysiology.

The NMJ consists of three distinct anatomic components: the presynaptic nerve terminal, the synaptic cleft containing the basement membrane, and the postsynaptic muscle fiber. The presynaptic nerve terminal originates from the myelinated axon of a motoneuron in the spinal cord. Lower motoneurons in the ventral gray matter of the spinal cord send their myelinated axons through the ventral root towards peripheral muscle targets. As the axon approaches the muscle fibers, it loses its myelin sheath and branches into a fine network of terminals, each approximately 2 μm in diameter. Each branch has several swellings along its course, termed presynaptic boutons, where the nerve makes synaptic contact with the muscle fiber. Presynaptic boutons are covered by terminal Schwann cells, which provide growth factors, recycle neurotransmitters, and may participate in recovery after nerve and muscle injury. Presynaptic boutons are positioned over specialized regions of the muscle cell membrane called the end-plate regions. Furthermore, underlying each bouton is a specialized invagination of the cell membrane that contains a very high concentration of nicotinic acetylcholine (ACh) receptors and voltage-gated Na⁺ channels. The presynaptic bouton and the end-plate are separated by a 100 nm-wide synaptic cleft containing the basement membrane and the extracellular matrix. The basement membrane anchors a number of proteins, including acetylcholinesterase, the enzyme responsible for rapid hydrolysis of ACh in the synaptic cleft.

Several distinct functional steps occur during synaptic transmission at the NMJ. First, the action potential arriving from the motor axon depolarizes the membrane in the presynaptic boutons, causing Ca²⁺ entry via voltage-gated Ca²⁺ channels on the presynaptic membrane. Ca²⁺ entry results in fusion of synaptic vesicles containing ACh with the presynaptic cell membrane and release of ACh into the synaptic cleft. ACh rapidly diffuses towards the postsynaptic membrane and binds to the nicotinic ACh receptor (nACh); two ACh molecules are required to activate the nACh receptor. Upon activation, the nACh receptor opens and allows both Na⁺ and K⁺ to flow through the ion pore. The inward Na⁺ current, however, dominates over the outward K⁺ current, resulting in net depolarization of the muscle membrane at the end-plate. This so-called end-plate potential propagates a short distance before encountering voltage-gated Na⁺ channels. These Na⁺ channels open when membrane potential rises to a critical threshold value, allowing only Na⁺ ions to flow into the cell and generating the all-or-none muscle action potential. Acetylcholinesterases in the synaptic cleft terminate the depolarizing action of ACh at the postsynaptic membrane by rapidly hydrolyzing ACh into acetate and choline.

A detailed understanding of diseases that affect synaptic transmission at the NMJ, as well as familiarity with clinical pharmacology as it applies to the NMJ, are essential in critical care. Specific diseases affecting the NMJ include toxin-mediated botulism and autoimmune disorders such as myasthenia gravis, Lambert-Eaton syndrome, and neuromyotonia. These are discussed in detail in Chapter 64 (see also Tseng-Ong and Mitchell⁸ and Lang and Vincent⁹). Among the pharmacotherapies targeted at the NMJ are some of the most commonly used drugs in the pediatric ICU: neuromuscular blockers (NMB). NMB agents are reviewed in Chapter 122. Furthermore, a number of toxins either decrease (anticholinergic agents) or increase (acetylcholinestrase inhibitors) the amount of ACh available at cholinergic synapses, resulting in corresponding toxidromes (see Chapter 106).

Chemical Synapses in the Central Nervous System

Chemical synapses in the CNS operate on basic principles similar to those governing synaptic transmission at the NMJ, although the cadre of neurotransmitters and postsynaptic receptors is significantly more diverse in the CNS. Importantly, a given neuron in the CNS may synthesize and store more than a single neurotransmitter, but it releases the same set of neurotransmitters at all of its synapses (Dale’s principle). CNS synapses are generally divided into asymmetric (Gray type I) synapses and symmetric (Gray type II) synapses on the basis of their appearance under electron microscopy. Physiologically, these correspond to excitatory and inhibitory synapses, respectively. Each neuron synthesizes its own complement of neurotransmitters, which are delivered to all synaptic contact sites in the axon and packaged into synaptic vesicles. When an action potential reaches the axon, Ca²⁺ currents cause the synaptic vesicles to fuse with the cell membrane, releasing neurotransmitters into the synaptic cleft. Neurotransmitters then act on their corresponding ionotropic and metabotropic receptors on the postsynaptic membrane. Ionotropic receptor activation leads to either a depolarizing, excitatory current or a hyperpolarizing, inhibitory current. These subthreshold currents are called excitatory and inhibitory postsynaptic potentials (EPSPs and IPSPs), respectively. Temporal and spatial summation of EPSPs in the dendritic tree of the postsynaptic neuron occasionally depolarizes the somatic membrane sufficiently to cross threshold and to generate an action potential. One of the distinguishing features of chemical synaptic transmission in the CNS, compared to the NMJ, is its lack of reliability on a single cell level; an action potential in the presynaptic neuron does not necessarily cause a postsynaptic neuron to fire its own action potential. Such lack of determinism likely allows for individual differences in responses to the same stimuli. After interaction with the receptor, the neurotransmitter is cleared from the synapse by diffusion, active reuptake into the terminal, or by enzymatic destruction, similar to ACh hydrolysis at the NMJ.

Acetylcholine

Acetycholine is an amine that functions as a neurotransmitter at the NMJ as well as in the CNS. At the NMJ, its actions are quick and precise, whereas in the CNS it functions as a slow, more global modulator of synaptic activity. Cholinergic CNS neurons include all the motoneurons in the spinal cord, as well as a number of cell nuclei in the brainstem reticular formation and the basal forebrain. Cholinergic neurons synthesize ACh using choline acetyltransferase (ChAT), which transfers an acetyl group from acetyl coenzyme A (CoA) to choline (Figure 57-1, A). Choline concentration in the extracellular fluid is the rate-limiting step in the reaction. Cholinergic neurons also synthesize acetylcholinesterase (AChE), the enzyme that breaks down ACh. AChE is released with ACh into the synaptic cleft, where it rapidly hydrolyzes ACh into acetic acid and choline.

Figure 57–1 Neurotransmitter structure and synthesis. A, Acetycholine. B, Catecholamines. It should be noted that synthesis of norepinephrine and epinephrine requires dopamine as a precursor. C, Serotonin, an amine synthesized from tryptophan. D, Glutamate and GABA, amino acid derivatives.

Low concentrations of choline in the CNS have been associated with neurologic impairment during fetal¹⁰ and postnatal life.¹¹ Thus, choline supplementation represents an attractive therapeutic strategy in neurologic disorders characterized by decreased CNS choline. Interestingly, patients receiving long-term parenteral nutrition (TPN) occasionally develop choline deficiency, which has been associated with TPN-related liver failure¹² and possibly, cognitive dysfunction. Hence, choline supplementation in TPN-dependent patients may ameliorate some neurologic deficits.¹³

Catecholamines

Catecholamine neurotransmitters include dopamine, norepinephrine and epinephrine. For all three, the initial starting point in biochemical synthesis is the amino acid tyrosine (Figure 57-1, B). Tyrosine is converted into an intermediate compound, called dopa, by tyrosine hydroxylase (TH). In dopaminergic neurons, dopa is converted into dopamine by dopa decarboxylase. Noradrenergic neurons, which utilize norepinephrine as a neurotransmitter, further convert dopamine into norepinephrine using dopamine β-hydroxylase. Finally, norepinephrine is converted into epinephrine by an enzyme phentolamine N-methyltransferase, which is found only in adrenergic neurons. Thus all neurons that synthesize catecholamines contain TH and dopa decarboxylase, but only noradrenergic and adrenergic neurons contain the synthetic enzymes required to produce norepinephrine and epinephrine, respectively.

Catecholamine-utilizing neurons reside primarily in the brainstem. Dopaminergic neurons in humans are located in two mesencephalic nuclei: the substantia nigra and its medial neighbor, the ventral tegmental area. Dopaminergic neurons in the substantia nigra project primarily to the basal ganglia, where they are involved with initiation of voluntary movement. Ventral tegmental neurons send dopaminergic fibers to the amygdala and the cerebral cortex and participate in regulation of emotion, reward and addiction. Brainstem noradrenergic neurons are located in the locus ceruleus and in the reticular formation. They project widely to the thalamus and cortex as well as to the spinal cord, and play a significant role in arousal and vigilance. Adrenergic CNS neurons are located in the ventrolateral medulla, and participate in temperature regulation via their projections to the hypothalamus.

Unlike ACh, which is cleared from the synapse by hydrolysis, catecholamines are cleared from the synaptic cleft by reuptake into the axonal terminal. Once inside the cell, catecholamines are either repackaged into vesicles or destroyed by monoamine oxidase (MAO). Pharmacologic manipulation of synaptic catecholamine concentrations plays a therapeutic role in the management of several disorders, such as depression and attention deficit-hyperactivity disorder (ADHD). Furthermore, recreational drugs affecting catecholamine concentrations at the synapse continue to gain popularity and to grow in number. Therapeutic uses include treatment of severe depression with MAO inhibitors, which inhibit catecholamine breakdown, and treatment of ADHD with amphetamines, that interfere with dopamine transport and increase dopamine concentrations. Recreational drugs include amphetamine and its analogues, as well as cocaine, a selective norepinephrine transporter blocker. Excess catecholamine levels at the synapse result in sensations of euphoria, increased energy levels and improved focus, as well as anxiety, paranoia, and jitteriness. Notably, hypertensive crises leading to myocardial infarction and stroke may occur with use of cocaine, amphetamines, and MAO inhibitors.

Serotonin

Serotonin is an amine neurotransmitter synthesized from the amino acid tryptophan in a two-step process (Figure 57-1, C). First, tryptophan is hydroxylated by tryptophan hydroxylase to form 5-hydroxytryptophan (5-HTP). 5-HTP is then decarboxylated by 5-HTP decarboxylase to form serotonin, also known as 5-hydroxytryptamine (5-HT). After 5-HT is released at the synapse, it is cleared by a specific serotonin reuptake transporter. Serotonergic neurons are located in the rostral and caudal raphe nuclei in the brainstem. Rostral raphe neurons innervate the cerebral cortex, including the limbic system, where serotonin levels help regulate mood and attention. Caudal raphe neurons project to the brainstem and the spinal cord, where they are involved in regulation of general arousal and pain perception, respectively. Importantly, dysfunction of the serotonergic pathways originating in the raphe nuclei has been linked with sudden infant death syndrome (SIDS).¹⁴ Additionally, serotonin levels play a key role in depression, giving rise to an entire class of drugs in clinical use called selective serotonin reuptake inhibitors (SSRIs). SSRI abuse or overdose is rare, but may result in patients presenting with the potentially life-threatening “serotonin syndrome,” characterized by hypertension, tachycardia, mental status changes, myoclonus, and severe hyperthermia. The latter may lead to shock, rhabdomyolysis, renal failure, and death. The serotonin syndrome is particularly likely to occur when SSRIs and MAO inhibitors, inadvertently or intentionally, are taken together. Treatment includes serotonin antagonists, blood pressure control with either adrenergic antagonists or agonists as clinically indicated, and temperature control with benzodiazepines and neuromuscular blockade.

Amino Acids

Neurotransmitters derived from common amino acids include glutamate, gamma-aminobutyric acid (GABA), and glycine. These are among the most widely distributed neurotransmitters in the CNS. Glutamate and glycine exist as amino acids in all cells, where they are utilized as protein building blocks. Glutamatergic and glycinergic neurons have the additional capacity to package glutamate and glycine, respectively, into synaptic vesicles and release them at the synapse. GABA must be synthesized from glutamate via an additional reaction catalyzed by an enzyme glutamic acid decarboxylase (GAD) (Figure 57-1, D). Only GABA-ergic neurons contain GAD.

Glutamate is generally an excitatory neurotransmitter, whereas GABA and glycine are inhibitory. Excitatory glutamatergic neurons exert their influence both locally and over long distance, depending on the shape of their axons. Inhibitory neurons, on the other hand, tend to exert local inhibitory control over neuronal circuitry either in the brain (GABA) or in the spinal cord (glycine). A major exception are cerebellar Purkinje cells, which are GABA-ergic but project over long distances to the brainstem, thalamus, and cerebral cortex (see below).

Adenosine, Peptides, and Nitric Oxide

In addition to the “classic” neurotransmitters described above, a number of substances have been documented to mediate or modulate information transfer between neurons. These include adenosine triphosphate (ATP) and adenosine, which at the synapse is a metabolite of ATP released in the synaptic vesicle. ATP modulates neuronal excitability such that energy may be conserved during times of ATP depletion. Adenosine functions as a neurotransmitter in the autonomic nervous system (ANS), in the basal ganglia, and at some cortical synapses. It also modulates the respiratory rate, and adenosine antagonists, such as caffeine, are used to treat apnea and bradycardia of prematurity. In addition to ATP and adenosine, a number of peptides can be released in synaptic vesicles, including substance P, vasoactive intestinal peptide (VIP), endogenous opioids, and endogenous cannabinoids. These peptides are involved in pain sensation and perception (substance P and opioids), modulation of vascular tone (VIP), and as yet uncharacterized processes (cannabinoids).

Finally, several gaseous molecules function as neurotransmitters. These include nitric oxide (NO) and carbon monoxide (CO), and possibly hydrogen sulfide. NO, the most thoroughly studied of the gaseous neurotransmitters, is produced by brainstem neurons in the nucleus tractus solitarius, where it interacts with α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate (AMPA)-type and N-methyl-D-aspartate (NMDA)-type glutamate receptors and regulates cardiovascular function.¹⁵ Hydrogen sulfide is produced from the amino acid cysteine, and may influence cellular redox state and glutamatergic transmission.¹⁶ Intriguingly, hydrogen sulfide induces a suspended animation-like state in animals.¹⁷ and may be protective after resuscitation from cardiac arrest.¹⁸

Nicotinic Acetylcholine Receptors

Nicotinic ACh receptors are ligand-gated ion channels, related structurally and functionally to GABA_A channels, and a subset of serotonin receptors. Five transmembrane subunits comprise the nAChR and form a central pore that allows ionic currents to pass. There are two subunit classes, α and β, with multiple members in each class. nAChR is generally a heteromer, but homomer channels have been described. Each nAChR binds two acetycholine molecules, with affinity for ACh and for nicotine dependent on subunit composition. The receptor exists in three distinct states: closed, open, and desensitized. In the closed position, no ionic current passes through the central core. When an agonist, such as ACh or nicotine, binds the nAChR, the receptor opens, becoming permeable to monovalent and divalent cations. After a short period of time, the receptor spontaneously closes. Upon continued exposure to an agonist, however, the nAChR assumes a permanently closed, or desensitized, conformation.¹⁹ As discussed above, nAChRs mediate neuromuscular coupling at the NMJ. In addition, nAChRs are widely distributed in the CNS, where they perform a variety of functions depending on subunit composition and location.

In the CNS, unlike at the NMJ, nAChRs are permeable to both Na⁺ and Ca²⁺. In neurons, however, nAChR-evoked Ca²⁺ current exceeds the Na⁺ current twofold to tenfold. The greater Ca²⁺ permeability indicates that nAChRs mostly modulate synaptic transmission and release of other neurotransmitters. Indeed, although direct nAChR-dependent responses have been observed in the hippocampus and the developing visual cortex, the overwhelming evidence points to nicotinic receptors playing the role of modulator in the CNS. Their wide distribution in the CNS with locations presynaptically, postsynaptically, and extrasynaptically further supports that role. Activation of presynaptic nAChRs enhances release of ACh, dopamine, glutamate, and GABA. Coupling of enhanced glutamate release with nAChR-dependent increase in intracellular Ca²⁺ suggests that nAChRs participate in synaptic plasticity during learning. Postsynaptic and extrasynaptic nAChRs regulate excitability and signal propagation in neuronal circuits. In the hippocampus, for example, nAChR activation leads to increased release of GABA from inhibitory interneurons, which decreases the excitability of hippocampal pyramidal neurons. Nicotinic receptors also interact with the dopaminergic system in regulating neuronal circuitry in the basal ganglia and the limbic system. Thus, nicotinic receptors have been implicated not only in learning and memory, but also in regulation of addiction and reward. Furthermore, loss of cholinergic neurons represents one of the distinguishing neuropathologic features of Alzheimer disease, and cholinesterase inhibitors are widely used to improve cognition and memory in Alzheimer patients.¹⁹ In pediatrics, a mutation in nAChR is responsible for a specific clinical epilepsy phenotype, called autosomal dominant nocturnal frontal lobe epilepsy.²⁰ Seizure onset usually occurs around 12 years of age in otherwise healthy children. Seizures originate in the frontal lobe, and occur predominantly during non-REM sleep.²¹ The mutant nAChR is more sensitive to ACh than the wild-type receptor, suggesting that cholinergic medications should be avoided in these patients.

Muscarinic Acetylcholine Receptors

Muscarinic ACh receptors (mAChRs) comprise a group of metabotropic receptors that link ACh exposure at the surface with G protein activation inside the cell. There are five distinct subtypes of mAChRs, designated M₁ through M₅, and these subtypes are divided into two broad classes based on the identity of the G protein with which they interact. M₂ and M₄ mAChRs couple with G_i proteins, and inhibit adenylyl cyclase activity and reduce intracellular cAMP levels. M₁, M₃, and M₅ receptor subtypes couple with G_q proteins, and increase intracellular Ca²⁺ levels via activation of phospholipase C.²² In the CNS, the M₁ mAChR is the most abundant subtype, located on neurons in the cortex, thalamus and striatum.²³ mAChRs are also present in the PNS in the sweat glands and organs of lacrimation and salivation, as well as in the heart, where they mediate the parasympathetic control of heart rate and contractility. Muscarinic AChRs thus mediate many of the systemic effects of organophosphate exposure and nerve gas poisoning.²⁴

Glutamate Receptors

Glutamate is the major excitatory neurotransmitter in the central nervous system. In mammals, it depolarizes postsynaptic neurons by binding to three types of ionotropic glutamate receptors (iGluR), each of which is characterized by different affinities for synthetic analogues, by different selectivity to ions, and by different time-course of the current that is permitted to pass through the cell membrane. The three types of iGluR are the AMPA receptor, the NMDA receptor, and the kainate receptor. All three channels types are widely present throughout the CNS, with AMPA and NMDA channels mediating the bulk of the excitatory transmission. At present, the function of the kainate channel remains to be clearly defined.

AMPA and NMDA receptors differ from each other with respect to ion permeability and time-course of ion flow through the channel. Upon binding of glutamate, AMPA receptors open their pores, which are permeable to Na⁺ and K⁺ ions. At the negative resting potential of the neuronal cell membrane, Na⁺, driven by the electrochemical gradient, flows into the cell and causes a large, fast depolarization. AMPA receptors are generally impermeable to Ca²⁺, although more recent findings indicate that Ca²⁺-permeable AMPA receptors do exist and may significantly contribute to pathology observed in ALS and stroke.²⁵ In contrast, NMDA receptors are universally permeable to Ca²⁺, as well as to Na⁺ and K⁺, generating a slow inward depolarizing current. NMDA receptors possess a unique property, however, that allows them to pass current only when the neuronal membrane is already depolarized. This property, termed voltage-dependence, stems from Mg²⁺ ions blocking the entry pore of the NMDA channel at negative membrane potentials even in the presence of glutamate. As the neuron depolarizes further, mostly due to current flow via the AMPA receptor, the Mg²⁺ block is relieved, and Ca²⁺, as well as Na⁺, flow into the cell. Once inside the cell, Ca²⁺ ions mediate a multitude of effects, from modifying protein phosphorylation and gene expression to overt excitotoxicity and cell death. NMDA receptors are thus thought to function as coincidence detectors, linking events at the cell membrane, e.g., AMPA receptor-mediated depolarization, with long-term changes in synaptic strength and gene expression in the neuron.

In addition to directly modulating current flow via the ionotropic channels, glutamate modulates effector molecules within the neuron by binding to a diversity of G protein-linked metabotropic glutamate receptors (mGluRs). Eight mammalian subtypes of mGluR are divided into three categories, based on sequence homology and coupling to secondary effector systems.²⁶ Group I mGluRs (mGlurR1 and mGluR5) are localized at the edge of the postsynaptic neuronal membrane and are positively coupled to phospholipase C (PLC) via the G_q protein. Group II and III mGluRs are located on the edge of the presynaptic neuronal membrane and are negatively coupled to adenylyl cyclase (AC) via the G_i protein. All three groups are activated only when excess glutamate spills out of the synaptic cleft and diffuses towards mGluRs located at the periphery of the synaptic membrane.

Differential secondary messenger coupling and synaptic localization of the three mGluR groups points to their divergent roles in regulating neuronal function. Binding of glutamate to group I mGluRs leads to activation of PLC, which releases two secondary messengers—diacylglycerol (DAG) and inositol triphosphate (IP₃)—from the membrane phospholipid phosphoinositol 1,4,5-bisphosphate (PIP₂). DAG activates protein kinase C in the neuronal membrane, whereas PIP₂ diffuses towards its receptor on the internal cell membrane and triggers a massive Ca²⁺ release into the cytoplasm. Downstream events lead to 1) modulation of K⁺ currents, resulting in increased neuronal excitability and 2) potentiation of glutamate-dependent current at NMDA receptors specific to the synapses at which excess glutamate release has occurred. Group I mGluRs thus participate in activity-dependent strengthening of synaptic connections and play a significant role in learning and memory.

In contrast, group III, and probably group II, mGluRs on presynaptic neurons provide a negative feedback loop by inhibiting glutamate release. When glutamate binds to group III mGluRs, an inhibitory G protein (G_i) is activated, that then functions to decrease AC-mediated production of cAMP. A decrease in cAMP leads to lower Ca²⁺ concentrations at the presynaptic neuronal membrane and decreased synaptic vesicle fusion. The net effect is a decrease in the amount of glutamate released at the synapse and a reduction in synaptic transmission. Recently, mGluRs have emerged as a major therapeutic target due to the multitude of effects they exert on synaptic transmission. Their extrasynaptic location presumably will allow newly developed pharmaceutical agents to maximize therapeutic value and minimize unwanted side effects.²⁷

GABA_A and GABA_B Receptors

GABA is the major inhibitory neurotransmitter in the brain. Like glutamate and ACh, it binds two distinct classes of GABA receptors; GABA_A receptors are ionotropic, and GABA_B receptors are metabotropic. Both receptor classes are involved in regulation of physiologic and pathologic states, and pharmacological manipulation of GABA receptors plays a major role in the management of pediatric critical illness.

GABA_A receptors are chloride channels. At the normal resting membrane potential, opening of the GABA_A receptor allows chloride ions to flow into the cell down their electrochemical gradient. Influx of negatively charged Cl⁻ ions results in hyperpolarization of the cell membrane. In neurons, membrane hyperpolarization decreases the probability that the neuron will reach threshold and fire an action potential. Thus, on an individual cell level, GABA decreases neuronal activity via the GABA_A receptor.

GABA_A receptors are heteropentamers, similar in structure to the nAChRs. At least eight subunit classes exist in mammals, including humans, and each class consists of several members, allowing for a staggering 150,000 possible subunit combinations to create one functional GABA_A receptor. Only 500 combinations are known to exist, and most receptors contain a various complement of the α, β, and γ subunits.²⁸ Most GABA_A receptors cluster at postsynaptic densities, and such clustering appears to depend on presence of the γ subunit. However, a subset of the GABA_A receptors, in particular those containing the δ subunit, localize to extrasynaptic sites, mediate tonic levels of inhibition in the brain, and may underlie the pathophysiology of absence seizures.²⁹

The pharmacology of the GABA_A receptor is of particular relevance in critical care, since the two classes of first-line anticonvulsants and anxiolytics in clinical practice—the benzodiazepines and the barbiturates—allosterically modulate the GABA_A receptor. GABA itself binds the receptor at the junction of the α and β subunits. Benzodiazepines bind the GABA_A receptor at a different site, classically between the α and γ₂ subunits, and, in the presence of GABA, increase the frequency with which the chloride channel opens. Barbiturates, in contrast, bind at yet a different site and increase the duration of the open state in the presence of GABA. Thus both benzodiazepines and barbiturates increase the efficacy of endogenous GABA in hyperpolarizing the cell membrane. A major difference between the two drug classes is that at increasing concentrations, barbiturates, but not benzodiazepines, become direct GABA agonists and can open GABA_A channels independent of endogenous GABA release. Hence, barbiturates have a significantly narrower safety window compared to benzodiazepines.

Additional GABA_A receptor ligands of clinical importance include (1) general inhalational anesthetics, which are thought to modulate tonic inhibition via the δ subunit-containing receptors; (2) alcohol, although its mechanistic action is poorly understood; and (3) flumazenil, a competitive benzodiazepine antagonist used clinically to reverse benzodiazepine overdose.

GABA_A receptors mutations contribute to several known disease states in humans. Two different point mutations on chromosome 5, both affecting the γ₂ subunit, are associated with development of febrile seizures and generalized epilepsy as well as with a link between the two conditions.^30–31 In patients with temporal lobe epilepsy, GABA_A receptor expression is altered in hippocampal neurons,³² and GABA-evoked responses actually depolarize, rather than hyperpolarize, neurons in excised tissue (see also Developmental Processes section later in the chapter).³³ GABA_A receptor dysfunction is also thought to contribute to anxiety, panic disorder, schizophrenia, and sleep disturbances.³⁴ Thus, pharmacologic modulation of GABA_A receptor function represents an active area of research and novel drug development.

GABA_B receptors are heterodimeric proteins that activate G protein-coupled second messenger systems upon interaction with GABA at the membrane.³⁵ The GABA_B1 subunit binds GABA at the cell membrane, while the GABA_B2 subunit interacts with the G protein on the cytosolic surface. GABA_B receptors are widely distributed throughout the CNS but are particularly abundant in the thalamus, cerebellum, and hippocampus. They can be located both presynaptically and postsynaptically. Presynaptic GABA_B receptors allow G protein–coupled Ca²⁺ influx into the cell, and leads to feedback inhibition of transmitter release. Postsynaptic GABA_B receptors function as K⁺ channels by allowing a slow, G protein–dependent K⁺ current to leak out of the cell, and leads to cell hyperpolarization (since positive ions have left the inside of the cell membrane).³⁶

Baclofen is the only pharmacologic GABA_B receptor ligand in current clinical use. It directly activates the GABA_B receptor and, through its action in the spinal cord, leads to reduction in muscle tone. Thus it is used primarily to relieve spasticity after CNS injury. When administered systemically, it has a significant side-effect profile, including hypotension and bradycardia.³⁵ Systemic side effects have been minimized with intrathecal administration via an indwelling catheter and pump.^37,38 Indeed, intrathecal baclofen infusion has emerged as an alternative to the more invasive dorsal rhizotomy in the treatment of spasticity refractory to medical therapy in pediatric patients.³⁹ Although intrathecal baclofen delivery systems are effective, they have a 10% to 20% failure rate over time. When an intrathecal baclofen pump fails, patients can develop baclofen withdrawal symptoms characterized by agitation, increasing spasticity and dystonia, hypertension, tachycardia, hyperthermia, and potentially death.³⁹ Thus baclofen withdrawal should be considered in the differential diagnosis of an agitated child with an indwelling baclofen pump.

Spinal Cord

The spinal cord is organized into segments delineated by the exiting spinal nerves. In humans, there are 31 segments: 8 cervical, 12 thoracic, 5 lumbar, 5 sacral, and 1 coccygeal.⁴⁰ The first seven cervical nerves exit the spinal canal above the corresponding vertebra, that is, the C1 nerve exits between the occiput and the C1 vertebra (the atlas). Since in humans there are only seven cervical vertebra and eight cervical nerves, the last cervical nerve, C8, exits the spinal canal between C7 and T1. From T1 on, each corresponding spinal nerve exits the spinal cord below its corresponding vertebra, that is, the T12 nerve exits between T12 and L1. Early in fetal life, the spinal cord extends throughout the entire length of the spinal canal. Beginning in gestation week 12, the growth rate of the vertebrae exceeds that of the spinal cord, such that by birth in humans, the spinal cord ends at L3. During postnatal development, further differential growth occurs, and in adults, the spinal cord ends more rostrally, between L1 and L2. The nerve roots continue to exit the spinal canal through their corresponding foraminae, such that the caudal spinal roots extend past the end of the spinal cord towards their exit points and form the cauda equina. The end of the spinal cord forms an important landmark during development, since the lumbar puncture must be performed below the spinal cord in order to avoid severe injury. Hence, in infants, the preferred location of the lumbar puncture is between L4 and L5 vertebrae, with an alternate site between L3 and L4. In adults, it is safe to perform the lumbar puncture between L3 and L4, with both the L2/L3 and the L4/L5 intervertebral spaces as alternative sites.

Spinal cord lesions result in two general subsets of neurologic deficits: those caused by interruption of ascending information flow towards the brain, and those caused by interruption of descending brain control of the spinal cord circuitry and the PNS. Thus complete spinal cord transection leads to loss of sensation and muscular paralysis below the level of the lesion due to injury to the ascending sensory pathways and descending motor pathways, respectively. Immediately after the transection, paralysis is flaccid with loss of deep tendon reflexes, characteristic of spinal shock (not to be confused with neurogenic shock, see below). After a period of time, paralysis becomes spastic, with increased muscle tone and hyperactive deep tendon reflexes, due to disrupted inhibitory control of spinal cord circuitry by the brain’s motor centers.

Medulla

The medulla extends rostrally from the spinal cord in the rough shape of an ice cream scoop.⁴⁰ The closed “handle” portion contains an enclosed central canal contiguous with that in the spinal cord. The open “scoop” portion is located rostrally where the central canal opens into the fourth ventricle. The medulla gives origin to four cranial nerves (CN): the glossopharyngeal (CN IX), vagus (CN X), accessory (CN XI), and hypoglossal (CN XII). It also contains decussations (crossings) of two major fiber tracts. The postsynaptic fibers from the nuclei gracilis and cuneatus, that carry tactile information from the lower and upper parts of the body, respectively, cross the midline in the caudal medulla, giving rise to the sensory decussation. The pyramidal tracts, that contain fibers descending from the motor cortex into the spinal cord, decussate slightly more rostrally, giving rise to the pyramidal, or motor decussation. These decussations are responsible for the fact that the left half of the brain controls and senses the right half of the body, and vice versa. Thus in general, damage to brain structures above the decussation gives rise to contralateral symptoms whereas damage below the decussation results in ipsilateral symptoms (except for the cerebellum as discussed below). Finally, the medulla contains the brain’s respiratory control center, which is of paramount importance in determining brain death (see below). In the absence of neuromuscular blockade, complete apnea in response to rising PaCO₂ results only when the medulla has been extensively injured.

Although extensive damage to the medullary structures is often quickly fatal due to ensuing apnea, more localized injury produces a number of recognizable syndromes. Wallenberg, or lateral medullary, syndrome occurs when the territory supplied by posterior inferior cerebellar artery has been compromised, and consists of loss of temperature and pain sensation on the contralateral side of the body and on the ipsilateral side of the face. Additionally, Horner syndrome and varying degrees of vertigo, dysphagia, and dysarthria can be present. The medial medullary syndrome (Dejerine or inferior alternating syndrome) results from injury to the territory supplied by the anterior spinal artery or, occasionally, the vertebral artery. It is characterized by weakness or complete hemiplegia and loss of tactile and vibratory perception on the contralateral side of the body, together with preservation of temperature and pain sensation in the body and full sensation in the face. Medullary injury also occurs as a life-threatening, albeit rare, complication of tonsillectomy and adenoidectomy.⁴¹