The carotid body is a peripheral sensory organ that can transduce modest falls in the arterial PO2 (partial pressure of oxygen) into a neural signal that provides the afferent limb of a set of stereotypic cardiorespiratory reflexes that are graded according to the intensity of the stimulus. The stimulus sensed is tissue PO2 and this can be estimated to be around 50 mmHg during arterial normoxia, falling to between 10–40 mmHg during hypoxia. The chemoafferent hypoxia stimulus-response curve is exponential, rising in discharge frequency with falling PO2, and with no absolute threshold apparent in hyperoxia. Although the oxygen sensor has not been definitely identified, it is believed to reside within type I cells of the carotid body, and presently two major hypotheses have been put forward to account for the sensing mechanism. The first relies upon alterations in the cell energy status that is sensed by the cytosolic enzyme AMPK (AMP-activated protein kinase) subsequent to hypoxia-induced increases in the cellular AMP/ATP ratio during hypoxia. AMPK is localized close to the plasma membrane and its activation can inhibit both large conductance, calcium-activated potassium (BK) and background, TASK-like potassium channels, inducing membrane depolarization, voltage-gated calcium entry and neurosecretion of a range of transmitter and modulator substances, including catecholamines, ATP and acetylcholine. The alternative hypothesis considers a role for haemoxygenase-2, which uses oxygen as a substrate and may act to gate an associated BK channel through the action of its products, carbon monoxide and possibly haem. It is likely however, that these and other hypotheses of oxygen transduction are not mutually exclusive and that each plays a role, via its own particular sensitivity, in shaping the full response of this organ between hyperoxia and anoxia.
Type I cells of the mammalian carotid body sense hypoxia; that is, they have the ability to detect and convert, within seconds, physiological decreases in the PO2 (partial pressure of oxygen) of arterial blood into calcium-dependent neurosecretion that results in coded neural discharge in the post-synaptically located carotid sinus nerve afferents. This discharge is appropriately graded for the initiation of remedial cardiorespiratory reflexes that act to reduce the impact of systemic hypoxia. Carotid body hypoxia chemotransduction thus differs from oxygen sensing in many other tissues and cells of the body as it demonstrates a rapidly responding, low affinity sensing system, i.e. acting without alteration in protein expression levels and at degrees of PO2 that should be without compromise to cellular metabolism, for the initiation of reflex responses to hypoxaemia. Although hypoxia is believed to be the key excitatory stimulus of the carotid body, it is important to remember that what is being sensed is a lack of oxygen. Put another way, hypoxia must remove or modify some component of cellular physiology that normally occurs in the presence of adequate levels of molecular oxygen. Although a consensus is emerging that places the inhibition of specific potassium channels located within the plasmalemmal membrane of type I cells of the carotid body as pivotal to the sensing mechanism, no single sensor for a lack of oxygen has yet been convincingly defined and it may be that a number of sensors, each with different oxygen affinities and downstream targets, co-exist to provide the graded response that is known to occur throughout the wide range of PO2 between hyperoxia and anoxia. This chapter aims to provide an introduction and general overview of the main features of carotid body chemoreception, from the nature of the stimulus, to the generation of action potentials in response to acute hypoxia and through to the molecular mechanism involved in stimulus transduction while attempting to highlight where consensus has yet to be established and possible reasons for this.
Tissue PO2 is the adequate stimulus of the carotid body
The carotid bodies are small (<100 μg wet weight in the rat) sensory organs located bilaterally in the vicinity of the bifurcation of the common carotid arteries where they branch into the internal and external carotid arteries in the neck of mammals (Figure 1). They receive their sensory, afferent innervation from the carotid sinus nerve, a branch of the glossopharyngeal (IXth cranial) nerve that has its cell bodies located in the petrosal ganglion. These afferent fibres project from the NTS (nucleus tractus solitarius) of the medulla where they mediate cardiorespiratory and autonomic reflex responses. The carotid body also receives a sympathetic, efferent innervation from the closely located superior cervical ganglion and an additional efferent innervation from parasympathetic fibres originating from the vagus (Xth cranial) nerve. The organ is protected by a connective tissue capsule and is highly vascular, receiving a blood flow, relative to its size, of 10–15 times that received by the brain. In addition to a capillary supply consisting of both typical (approx. 7–8 μm) diameter vessels and wider (maximum 18 μm) diameter penetrating sinusoid capillaries that make more intimate contact with the type I cells, there is also anatomical evidence that the carotid body possesses arterio-venous shunts, thus providing various routes for blood to flow through the organ. Its efferent innervation may act principally to shape the afferent response of the organ by modulation of this blood flow and hence the stimulus intensity in the vicinity of the sensory cells. An extensive anatomical description has been described elsewhere . The high blood flow (approx. 2000 ml per 100 g−1·min−1, at normal blood gases and blood pressure) ensures a very low arterio-venous oxygen difference in normoxia and provides sufficient oxygen for metabolic demands during periods of increased activity. Technical difficulties confound accurate measurements of oxygen consumption, but a value of between 1–2 ml per 100 g−1·min−1 during normoxia is a best estimate. In accordance with its sensory function, on lowering PO2, the carotid body markedly increases its glucose consumption and oxidation and an increase in the rate of oxygen disappearance has been positively correlated to sinus nerve afferent discharge . Thus oxygen is needed to provide the energy necessary to sustain the chemosensory responses during hypoxia but, at some point, with increasing hypoxia, failure to sustain aerobic metabolism must occur.
Each carotid body consists of thousands of 8–12 μm diameter, spherical type I cells arranged in clustered islets of 3–5 cells (glomeruli) within a lobular structure. Each cluster is closely associated with a single, elongated type II cell, blood vessels and an afferent nerve supply. Numerous studies have identified the type I cell as the oxygen-sensing element with the type II cell allocated a supportive, glial-like role. Type II cells do not appear to make synaptic contact with afferent nerve fibres but type I cells do. In addition, type I cells are connected to each other via chemical and electrical gap junctions and the consequences of losing these communication pathways during cell isolation procedures may account for some of the differences noted between different preparations. The question of a heterogeneous population of type I cells has been raised, but this is not really considered in the present hypotheses of chemotransduction, and there is no evidence to suggest otherwise than all afferent fibres that show a basal discharge are able also to convey neural information related to hypoxia intensity. Most, but not all, type I cells receive afferent innervation from one or more fibres of the carotid sinus nerve and both calyx and bouton synaptic configurations have been described. Each single fibre can oppose more than one type I cell and from more than one glomerulus. The fibre diameter ranges from 1–10 μm and the conduction velocity between 5–50 ms−1. The complex innervation of the carotid body could lead to a form of elementary signal processing via post-synaptic summation or action potential collision, but there have been very few studies directed at this question and chemoreceptor discharge is most often reported simply as a mean level of frequency, although time-related information does exist, as reported in vivo .
Historically, carotid body chemoreceptor responses, recorded in vivo, have been correlated against arterial PO2 levels and responses in vitro against the superfusate PO2. However, non-endothelial cells sense tissue PO2 which is lower than arterial and dependent upon the inspired PO2, the oxygen transport capacity of the blood and the particular tissue vascular structure and blood flow, as well as, ultimately, the local oxygen consumption and diffusion conditions (Figure 2). A microvascular PO2 of around 50 mmHg (7.5 mmHg=approx. 1 kPa) has been reported using a non-invasive phosphorescence quenching technique , a value that broadly agrees with earlier determinations made using sharp electrodes . Thus the blood flow of the carotid body appears well matched to its metabolism and it does not appear to have a particularly low tissue PO2 that might account for its distinct sensitivity to arterial PO2. Instead, this suggests the existence of some inherent cellular mechanism with an apparent low affinity for oxygen. It should be remembered, however, that there exists a rich sympathetic innervation of the carotid body vasculature that might play a role in determining how this tissue PO2 varies during hypoxia. Thus the precise horizontal position of any PO2 stimulus-response determination will depend principally upon the point at which the PO2 was measured. Measurements of single, type I cell acute responses to hypoxia in vitro would not therefore be expected to occur until the PO2 of the medium falls below approx. 40 mmHg, with a lower value of less than approx. 10 mmHg forcing these cells into anaerobic metabolism  and perhaps potentially triggering at that point another set of non-physiological responses. Stimulus-response curves measured in vivo, with arterial blood PO2 taken as the stimulus, would thus be right-shifted relative to single cell, in vitro responses (Figure 3).
In addition to reductions in the arterial PO2 (i.e. hypoxic hypoxia), stagnant or ischaemic hypoxia and histotoxic hypoxia are also powerful stimuli of the carotid body. Thus sustained and significant haemorrhage leading to hypotension and metabolic poisons including cyanide all increase chemoafferent discharge. In contrast, the arterial content of oxygen does not appear to be sensed by the carotid body of all species and anaemic hypoxia, the reduction of blood oxygen content, without alteration in arterial PO2, is not a principal stimulus of the carotid body and therefore does not stimulate ventilation. Accordingly, carbon monoxide at non-lethal arterial tensions also does not stimulate ventilation and carbon monoxide applied to the carotid body may even, through elevating arterial PO2, decrease chemoafferent discharge . The degree of binding of carbon monoxide to haemoglobin (termed HbCO) needs to be >20% before an increase in discharge is noted but even up to 50% HbCO, the effect is small and an explanation for this lack of excitation by HbCO may lie with the relatively high blood flow of the carotid body and thus its intrinsic inability to sense oxygen delivery itself. The rodent carotid body may differ in this respect, with one report [5a] suggesting that the rat carotid body might detect the arterial content of oxygen; although, to date, this has only been reported at the level of ventilation. The mechanism underlying this species difference is not clear.
Chemoafferent and reflex responses to hypoxia
Chemoafferent discharge provides the final common output of the carotid body and could be considered therefore as a definitive measure of its response to hypoxia. A tonic activity of around 1–2 Hz is observed in carotid sinus nerve single-fibre chemoafferents during hyperoxia in vivo with no absolute physiological threshold to arterial PO2. The discharge is aperiodic with interspike interval distributions fitting a Poisson distribution. The response to hypoxic hypoxia is rapid with a latency of <1 s and a time to peak of 1–5 s, after subtraction of circulatory transit delays. This response does not adapt and remains relatively sustained for the duration of the hypoxia episode, provided the intensity is not too severe. Steady-state discharge rises increasingly with mounting hypoxia, at a constant arterial PCO2 (partial pressure of carbon dioxide), and may be best described by single exponential function with an offset to account for the lack of an oxygen threshold. Chemoreceptor discharge begins to rise gradually from between 400–140 mmHg arterial PO2 until approx. 100 mmHg when the slope increases and then rises more abruptly at an arterial PO2 of approx. 70–75 mmHg, reaching a maximum activity at an arterial PO2 of approx. 20–30 mmHg. A reduction in discharge frequency at an arterial PO2 of below approx. 20 mmHg has also been observed and thus the full arterial PO2-response curve may therefore best be described by a sigmoidal function, which is also positioned appropriately as an inverse function of the haemoglobin oxygen-dissociation curve.
Single-fibre afferent frequency is elevated by 10–20 times basal at its peak across most species studies, with the in vivo absolute level appearing species-dependent and difficult to quantify precisely due to various degrees of anaesthesia/paralytic agents. Different fibres appear to have different thresholds and peaks and the inflection in the curves of different chemoreceptor fibres can occur over the arterial PO2 range of 60–110 mmHg , reflecting perhaps variability in the affinity range of an oxygen sensor(s) or to gradients in tissue PO2 within the organ. At present, there is no compelling evidence to suggest that a single oxygen sensor is responsible for the entire stimulus-response characteristics of the carotid body and a compelling alternative is that a range of sensors with various affinities and thresholds might better explain the response  or, indeed, provide an important failsafe redundancy. Recordings from multi-fibre rather than single-fibre preparations are therefore prone to errors if used for determining sensitivity at a single-cell level and should only be interpreted with caution to infer transduction processes.
In the adult, the reflex response to hypoxia includes hyperventilation, i.e. an increase in alveolar ventilation above metabolic demand, which acts to decrease alveolar PCO2 levels and thus increase alveolar (and hence arterial) PO2 levels towards that of inspired. In addition to respiratory reflexes, an elevated chemoafferent discharge can also induce direct cardiac and vascular reflex responses that include bradycardia, decreased cardiac output and peripheral vasoconstriction as well as augmented adrenomedullary output, but these effects are often masked by secondary feedback responses, via thoracic afferents and baroreceptors, subsequent to the hyperventilation . These complex reflexes and interactions hint at how carotid body dysfunction might be implicated in the aetiology of a variety of cardiorespiratory disorders, including sleep apnoea, hypertension and cardiac arrhythmia although a definitive cause-effect has not been established.
Hypoxia induces intracellular calcium elevations in isolated type I cells and neurosecretion from these cells is calcium-dependent. Although a role for intracellular calcium stores is not entirely without experimental support, the elevated intracellular calcium during hypoxia arises predominantly subsequent to calcium entry via membrane channels . Thus, in the absence of external calcium, chemoafferent discharge responses to hypoxia are greatly attenuated. Attenuation in discharge  and in catecholamine secretion  is also observed, and in a dose-dependent manner, with increasing application of a variety of calcium channel blockers, including nifedipine, suggesting an important role for entry via L-type VGCCs (voltage-gated calcium channels). This common finding implicates cell depolarization as an essential requirement of hypoxia sensing. Additionally, and crucially, mitochondrial membrane depolarization during hypoxia occurs after cell membrane depolarization . Finally, confirming a role for VGCCs, when the type I cell membrane potential is clamped at its resting membrane potential, the elevation in calcium during hypoxia/anoxia is attenuated and slowed . The calcium response is graded with hypoxia intensity but it is worth noting that the relationship between calcium and neurosecretion need not be linear. In addition to L-type, the presence of N-, P-/Q- and R-type calcium channels have also been demonstrated in chemoreceptor cells with, in the rabbit, perhaps only L- and P-/Q- activation being coupled to neurosecretion . More recently, TRP (transient receptor potential) channels have been localized to the carotid body  and a voltage-independent entry pathway for calcium during hypoxia may therefore also exist. More work is required to establish the role of these channels in oxygen sensing.
Carotid body type I cells contain a variety of both excitatory and inhibitory neurotransmitter substances. These range from amines including catecholamines and acetylcholine to neuropeptides, including substance P, the purines ATP and adenosine, amino acids, including GABA (γ-aminobutyric acid) and even the gas signalling molecules, nitric oxide and carbon monoxide. A detailed description of the effects of these various substances on carotid body function is not possible here and can be found elsewhere , but it is becoming apparent that a single transmitter cannot account for the hypoxia response of the carotid body . In addition, a considerable species differences is apparent in the specific transmitter used to signal hypoxia and different species may even use the same transmitter in dissimilar ways. Thus dopamine and acetylcholine have been shown to have different effects upon excitation in cats and rabbits. The consequences of these differences, if there are any, are not known. The presence of pre-synaptic (i.e. type I cell) autoreceptors, in addition to post-synaptic receptors, reveals a complexity of control that is not yet fully understood. For example, adenosine, released from type I cells during hypoxia, can act on type I cells to inhibit a 4-aminopyridine-sensitive outward potassium current . The role of this in shaping the response to hypoxia is not known, but it presumably acts to augment the response.
Of the various transmitters postulated to be central to the transduction process, attention has focussed in recent years upon the co-transmitter substance, ATP. This follows the finding that ATP-activated purinoceptors, specifically the ligand-gated P2X ion channel subtype, exists on chemoafferent nerve terminals and type I cells , that blocking the purinergic receptor attenuates (and inhibits when combined with blockage of the cholinergic receptor) the sensory response to hypoxia  and that modified mice lacking the P2×2, but not P2×3 receptors, have a blunted hypoxia sensitivity at both the ventilatory and chemoafferrent level . P2X receptors desensitize rapidly and although this may be an important information-containing process in central neurones, it does not appear consistent with the sustained chemoafferent discharge observed during a period of hypoxia. Thus interplay between both excitatory and inhibitory transmitters might exist to prolong the excitation and Prabhakar  has forwarded what he calls a ‘push-pull’ hypothesis to account for this phenomenon. This interesting hypothesis has yet to be tested. Additionally, the existence of various transmitter substances might be an essential requirement for the polymodal receptor nature of the carotid body type I cell  as well as for enabling plasticity in response to changes during development or in its tissue environment due, for example, to chronic diseases predisposing to alterations in blood-gas chemistry. The precise role of each of the chemical substances in the carotid body in either determining post-synaptic action potential frequency or in shaping the response is not yet known with certainty but should prove a demanding challenge in the future if we are to understand fully how the carotid body functions.
Potassium channel inhibition and depolarization
The absolute requirement for VGCCs implicates membrane depolarization as a key upstream event in the hypoxia transduction process and this is believed to be achieved by closure of potassium channels located in the plasma membrane of type I cells. These channels must therefore have the property of being either directly, or indirectly, inactivated by hypoxia and must have a finite open conductance at the Em (resting membrane potential; RMP), which, in the rat type I cell, is between −45 and −55 mV. The first description of a fast inactivating, delayed-rectifier potassium current in rabbit type I cells that was decreased, reversibly, by hypoxia was made almost 20 years ago . This effect appeared to be membrane delimited, suggesting an intrinsic sensor mechanism that might therefore be associated with either the α-pore forming subunits of the channel or with associated, auxiliary β-subunits. Later attempts at recreating oxygen sensitivity using recombinant α- and β-subunits in cell expression systems did not, however, prove conclusive and, together with studies in other species demonstrating potassium channel oxygen sensitivity only in patch clamp configurations that would retain cytosolic or closely-associated factors, have returned the focus to identifying an alternative sensor mechanism that couples to potassium channels.
Although a number of potassium channel subtypes are now known to be sensitive to hypoxia, there appears to be a significant species difference, as well as discrepancies within a species, and a definitive identity of the channel(s) responsible for initiating depolarization in type I cells has yet to be made. There is, however, no absolute requirement for just one class of potassium channel to be involved in oxygen sensing and all types may play a role within a single cell to both induce and maintain the depolarization. Thus, in addition to the originally-described Kv (voltage-dependent potassium channels; most likely the Kv4 subtype in rabbit and the Kv3 subtype in mice ), non-inactivating, calcium-dependent potassium channels (principally the large conductance, BK subtype ), hERG potassium channels  and a heteromeric combination of TASK1 and 3 subunits of the voltage-independent, acid-sensitive tandem pore domain potassium channel family  have also been forwarded as candidate oxygen-dependent ion channels. Of these, PO2-response curves for channel inactivation that correlate with the cellular response to tissue hypoxia have been demonstrated convincingly only with the TASK channels that have a K1/2 of 12–13 mmHg PO2. These channels have a low (sub 20 pS) conductance and are activated at potentials greater than approx. −90 mV, making them suitable for the initiation of hypoxia-mediated depolarization. Selective pharmacological approaches to studying these channels are limited at present, with barium, quinidine and both local and gaseous anaesthetics being the prime agents and, more recently, the cannabinoid, anandamide. The gating of these channels by other aspects of cellular metabolism, including pH and lipids, add complexity to the interpretation, but it is noteworthy that barium can induce carotid body afferent discharge . Conversely, the selective blockers of BK channels, namely iberiotoxin and charybdotoxin, are without effect upon resting carotid body chemoafferent discharge or membrane potential  as would have been expected were they contributing to the RMP. Additionally, the relatively non-selective, but TASK-independent, potassium channel blockers TEA (tetraethylammonium) and 4-AP (4-aminopyridine), do not appear to induce afferent discharge or prevent hypoxia-induced discharge and it is therefore not easy to reconcile these findings with others that show a depolarization in type I cells during normoxic, current clamp with charybdotoxin . Some of the discrepancy might exist due to different preparations and this explanation appears founded as a depolarizing and neurosecretory effect of TEA in a novel slice preparation of the carotid body in which cell–cell interactions would be more likely preserved than in the previous isolated cell experiments has been described . It has also been reported that, although charybdotoxin was without effect upon resting chemoafferent discharge, in the presence of hypoxia, this blocker of BK channels right-shifted the PO2-response curve, implying a greater effect with increasing membrane depolarization . Thus were the large conductance BK channels not inhibited by hypoxia, the small, depolarizing currents of approx. 7 pA induced by TASK channel closures, would be rapidly reversed, potentially preventing a sustained calcium influx and perhaps leading to an adapting hypoxia response, which is not observed in vivo. On the other hand, the process of neurosecretion does appear to initiate a limiting control on TASK channels as it has been shown that GABAB autoreceptors can activate TASK-like channels on type I cells thus potentially reversing hypoxia-induced depolarization . It is perhaps worth noting that up to 10% of type I cells might not respond to hypoxia and in those that do the magnitude of depolarization in response to hypoxia is generally quite small, at around 10 mV. The reason, or function, of this heterologous response is not apparent, nor why, within and across species, variations exist in oxygen sensitive potassium channels and the case for an essential requirement for potassium channel inhibition in hypoxia transduction, whilst compelling and widely accepted, is not yet fully proven.
The nature of the oxygen sensor
The sigmoidal shape of the PO2 afferent nerve response curve and its similarity to the haemoglobin-oxygen dissociation curve suggests that the sensor of the lack of oxygen might be a haem protein, an idea borne out by the finding that most hypotheses of hypoxia transduction invoke the participation of such proteins, from either intra- or extra-mitochondrial locations. The relative importance of these various proteins to the full functional response of the carotid body would ultimately depend upon their oxygen affinity and the nature of their coupling to the transduction process and it is important to re-iterate that the role of any putative sensors need not be mutually exclusive.
The oxygen affinity of mitochondrial proteins of the electron transport chain in general appears too high for a role in sensing systemic levels of hypoxia, yet a role for these complexes has long been suggested based, principally, upon the finding that blockers of the electron transport chain or uncouplers of oxidative phosphorylation are potent chemostimuli, increasing afferent chemodischarge and intracellular calcium and inhibiting TASK channels of type I cells . Interestingly, the inhibition of TASK channels by hypoxia appears cell specific  and thus type I cell mitochondria may possess some, as yet undefined, specialization, as suggested by the particularly exquisite sensitivity to hypoxia of their membrane potential and NAD(P)H autofluorescence  and/or the activity of cytochrome c oxidase in complex IV of the electron transport chain . Importantly for this hypothesis, hypoxia responsiveness is greatly reduced or abolished in the presence of the various blockers and uncouplers, which in itself, suggests some commonality of mechanism, although care needs to be exercised when interpreting some of these data, especially when the complex I blocker, rotenone, has been used as this appears to have undefined extra-mitochondrial effects upon oxygen sensing .
If the mitochondria are acting as oxygen sensors, the question arises as to how they couple their reduced activity during hypoxia to the plasma membrane. Current mechanisms proposed include variations in the cellular redox potential, consequent to variations in the concentration of ROS (reactive oxygen species), including the superoxide anion, O2−, formed by incomplete electron transfer in the electron transport chain (principally in complexes I and III) and its diffusible, dismutased product, hydrogen peroxide, H2O2. Although ROS may affect potassium channel function, their role in oxygen sensing is contentious  with global cellular redox apparently unchanged during hypoxia. Another source of ROS localized more closely to plasmalemmal potassium channels is the extra-mitochondrial, membrane-bound NADPH oxidase complex, which produces O2− and H2O2 in proportion to PO2 . Although proposed as a sensor mechanism, an action of NADPH oxidase-derived ROS on type I cell potassium channel inactivation is not proven. Thus, whilst subunits of NADPH oxidase are located within type I cells, mice with a selective (gp91phox) gene deletion of NADPH oxidase activity retained normal type I cell and carotid body function in response to hypoxia [38,39]. Further dispelling a role for NADPH oxidase in sensing hypoxia, is the more recent finding that targeted deletion of the p47phox subunit of NADPH oxidase led to an enhanced inactivation of type I cell BK channels  and this suggests that ROS may be involved in cell repolarization during the recovery from hypoxia, rather than depolarization. Although an action on other potassium channels in the carotid body has not been confirmed to date, a role for ROS in mediating hypoxia chemotransduction is not compelling.
An alternative signalling pathway to couple mitochondrial function with potassium channel inhibition may be via changes in the cytosolic concentration of ATP or factor(s) related to ATP synthesis. This is a long established hypothesis and is based upon a predicted final common consequence of hypoxia and electron transport chain blocking and uncoupling drugs and includes evidence demonstrating an excitatory action upon chemodischarge of the mitochondrial ATPase inhibitor, oligomycin,  and more recently and specifically, inhibitory actions of carbon monoxide, rotenone, myxothiazol, cyanide, DNP (2,4-dinitrophenol) and FCCP (carbonyl-p-trifluoromethoxyphenylhydrazone) upon TASK channel activation in type I cells [31,42]. The common feature of all of these agents would be a predicted fall in cytosolic ATP concentration. These findings, additionally, argue against an important role for ROS as these levels would be affected differently by these various agents. A role for mitochondria in oxygen sensing is, however, not without criticism and it has been reported that, other than rotenone, many inhibitors of oxidative phosphorylation have non-specific actions upon potassium channel activity that are independent of the effects of hypoxia . In addition, partial deletion of sdhd, a component of mitochondrial complex II, in mice was without effect upon hypoxia sensitivity as assayed by catecholamine release from type I cells . Reconciling these differences will require further experimentation and it remains to be seen whether type I cell mitochondria are specialized for oxygen sensing in the carotid body.
Sensing systemic hypoxia is an energy consuming process and, not surprisingly, cellular ATP levels do not fall until the cell is severely compromised. Thus, although a direct action of ATP upon TASK channels has been described in isolated membrane patches , it seems likely that this occurs indirectly in the whole cell. One such mechanism is consequent to the activity of adenylate kinase which, by converting two molecules of ADP into ATP and AMP, increases the cellular AMP:ATP ratio, raising AMP levels sufficiently to stimulate de novo synthesis of AMPK (AMP-activated protein kinase). AMPK is ubiquitously expressed in eukaryotic cells and plays a vital role in cellular metabolic regulation. Cellular stressors, including hypoxia, which decrease the energy status of the cell  activate AMPK which functions to augment ATP-generating processes while inhibiting non-essential energy consuming processes. Recently, a novel downstream target for this kinase, namely potassium channels of the carotid body type I cell, was identified and a hypothesis that places AMPK activation at the heart of oxygen sensing has been proposed . Thus, AMPK, located within approx. 1 μm of the plasma membrane in type I cells, can, subsequent to its [AMP]-independent activation by the drug, AICA (5-aminoimidazole-4-carboxamide) riboside, reversibly cause both BK and TASK channel inactivation, VGCC-dependent elevations in intracellular calcium and increased afferent neural discharge . These effects could also all be reversed or inhibited by the newly synthesized, AMPK antagonist, compound C. This hypothesis is particularly attractive as it unites the mitochondrial and potassium channel responses to hypoxia without a need to necessarily invoke mitochondrial specialization or decreases in [ATP], neither of which has compelling evidence for its support. An implication, however, if one assumes that the type I cell has a sensitivity to hypoxia not observed in other non-oxygen sensing tissues, must be that the carotid body either possesses unique isoform complexes of AMPK or that signalling by AMPK is somehow modified for a specific function. Further experiments will be required to ascertain the basis for this organ’s unique oxygen-sensing ability.
In contrast to the hypotheses requiring alterations in the cellular redox or energy status, another proposal has been forwarded that assumes a key role for gas molecule signalling in the inactivation of BK channels. Thus, Kemp and co-workers  were able to demonstrate that the oxygen-dependent enzyme, HO-2 (haemoxygenase-2) was closely associated in a complex with membrane-bound potassium channels and that at least one of its products, carbon monoxide, appeared essential for BK channel activation. According to this model, in the presence of oxygen and its co-substrates, haem and NADPH, carbon monoxide increases BK channel open probability, but when oxygen levels are compromised the reduction in carbon monoxide production would lead to BK inactivation and membrane depolarization. Carbon monoxide is a remarkably inert molecule, interacting almost exclusively with haem groups, of which the haem-binding domain of the a-subunit of BK appears the most likely site of its action. These data support earlier reports of an inhibitory function of HO-generated carbon monoxide at the carotid body [49,50] and further experiments are keenly anticipated.
The critical importance of matching oxygen delivery to metabolic rate is dependent upon a controlled reflex response to hypoxia and an understanding of this process may go some way towards a more full appreciation of the mechanisms and consequences of cardiorespiratory disease states. A full understanding of carotid body chemotransduction processes would be an important contribution to this appreciation and appears tantalisingly close. Consensus exists on the importance of calcium-dependent neurosecretion with the elevated intracellular calcium arising via VGCCs activated by membrane depolarization, induced by potassium channel closure subsequent to the detection of reduced oxygen by an intracellular protein sensor. The major focus of interest presently is in defining the oxygen sensor and several mitochondrial and non-mitochondrial candidate proteins have been proposed of which, presently, the cytosolic AMPK and the plasma membrane-bound HO-2 are prime candidates (Figure 4). These may all be important in shaping the full response to various degrees of hypoxia and it is hoped that future studies might attempt to begin a reconciliation of the various hypotheses. One way to aid this would be in the generation of full response curves for each putative sensor, from which oxygen affinities could be better determined.
• The carotid body is a peripheral oxygen sensor located in the bifurcation of the internal and external carotid arteries that senses arterial hypoxia translating this into afferent neural discharge that initiates corrective cardiorespiratory reflexes.
• Afferent discharge occurs subsequent to calcium dependent neurosecretion from type I cells on to closely apposed free nerve endings.
• The elevation in intracellular calcium is dependent upon calcium entry into type I cells through VGCCs and this is consequent to cell depolarization following the hypoxia-inactivation of specific, species-dependent classes of potassium channels, including TASK and BK.
• The sensitivity of the carotid body to hypoxia appears high and suggests the presence of a low-affinity oxygen sensor system(s).
• The nature of the oxygen sensor is not known for certain. The generation or reduction of ROS during hypoxia appears not to be a vital part of the transduction process but the proteins AMPK and HO-2 both have been put forward as essential components of this pathway and both have been associated with potassium channel inactivation during hypoxia.
I thank the Lister Institute, the Wellcome Trust and the British Heart Foundation for their generous support of my research and my colleagues for their collaboration over the years. Special thanks also go to Dr Joanne Wilton, University of Cambridge, for providing Figure 1.
- © The Authors Journal compilation © 2007 Biochemical Society