Role of glial-like type II cells as paracrine modulators of carotid body chemoreception

Abstract

Mammalian carotid bodies (CB) are chemosensory organs that mediate compensatory cardiorespiratory reflexes in response to low blood PO₂ (hypoxemia) and elevated CO₂/H⁺ (acid hypercapnia). The chemoreceptors are glomus or type I cells that occur in clusters enveloped by neighboring glial-like type II cells. During chemoexcitation type I cells depolarize, leading to Ca²⁺-dependent release of several neurotransmitters, some excitatory and others inhibitory, that help shape the afferent carotid sinus nerve (CSN) discharge. Among the predominantly excitatory neurotransmitters are the purines ATP and adenosine, whereas dopamine (DA) is inhibitory in most species. There is a consensus that ATP and adenosine, acting via postsynaptic ionotropic P2X2/3 receptors and pre- and/or postsynaptic A2 receptors respectively, are major contributors to the increased CSN discharge during chemoexcitation. However, it has been proposed that the CB sensory output is also tuned by paracrine signaling pathways, involving glial-like type II cells. Indeed, type II cells express functional receptors for several excitatory neurochemicals released by type I cells including ATP, 5-HT, ACh, angiotensin II, and endothelin-1. Stimulation of the corresponding G protein-coupled receptors increases intracellular Ca²⁺, leading to the further release of ATP through pannexin-1 channels. Recent evidence suggests that other CB neurochemicals, e.g., histamine and DA, may actually inhibit Ca²⁺ signaling in subpopulations of type II cells. Here, we review evidence supporting neurotransmitter-mediated crosstalk between type I and type II cells of the rat CB. We also consider the potential contribution of paracrine signaling and purinergic catabolic pathways to the integrated sensory output of the CB during chemotransduction.

INTRODUCTION

The peripheral control of breathing is dependent on sensory input to the central pattern generator that regulates the respiratory rhythm at the level of the brain stem. In mammals, this sensory input originates in arterial chemoreceptor organs, i.e., the bilaterally paired carotid bodies and to a lesser extent, the diffusely distributed aortic bodies (16, 25). The carotid bodies (CB) are richly vascularized and are stimulated by changes in the chemical composition of arterial blood, particularly low oxygen (hypoxemia) and elevated CO₂/H⁺ (acid hypercapnia). It is now recognized that CB are polymodal receptors that respond to other sensory modalities including low blood glucose (37, 62). The afferent discharge from the CB is carried by the carotid sinus nerve (CSN) whose cell bodies are located in the petrosal ganglia. Increases in CSN discharge during chemoexcitation trigger a compensatory increase in respiration, as well as an increase in sympathetic drive that influences the cardiovascular system (25, 59). Interest in the mechanisms that regulate CSN discharge has grown in recent years because exaggerated CB output is associated with several clinical disorders including sleep apnea, obstructive pulmonary disease, hypertension, and heart failure (25, 54, 59).

The magnitude of the CSN discharge during chemoexcitation is determined by a combination of pre- and postsynaptic factors. At the presynaptic level, sensory transduction is initiated at clusters of glomus or type I cells that depolarize and release neurotransmitters in response to chemostimuli such as acute hypoxia (2, 25, 38, 45). The mechanisms underlying hypoxia chemotransduction in type I cells remain controversial though significant progress has been made recently; the reader is referred to the following articles for an update on the competing theories (2, 5, 38, 56). On the postsynaptic side, the released excitatory neurotransmitter ATP and its breakdown product adenosine play key roles in increasing the sensory discharge by activating purinergic P2X2/3 and adenosine A2 receptors, respectively (7, 17, 44, 45, 73). In addition, release of other neuroactive agents, including the inhibitory neurotransmitters dopamine and GABA, may help fine-tune the afferent CSN discharge via autocrine-paracrine signaling pathways (16, 21, 46).

Despite significant progress in understanding the main synaptic interactions between type I cells and sensory nerve endings, there is increasing evidence that the final shaping of the CSN discharge may be more complex. For example, a role for glial-like type II cells, found in close association and sometimes displaying “synapse-like” contacts with type I cells (53), has largely been ignored during general discussions of CB physiology. Yet, the ability of glial cells to regulate synaptic strength by releasing “gliotransmitters” has attracted much attention in recent years, especially in studies of the central nervous system [for example see (39)]. Interest in the possibility that type II cells may be involved in paracrine signaling in the CB was first suggested by the observation that exogenously applied ATP can elicit robust intracellular Ca²⁺ responses in type II cells (67). More recent studies suggest this pathway may have downstream consequences in boosting ATP release via gap junction-like pannexin-1 channels (72). In this review, we consider how ATP and other paracrine signals released from type I cells might lead to the activation of type II cells and subsequent regulation of synaptic transmission at the CB chemosensory synapse.

TYPE II CELLS EXPRESS G PROTEIN-COUPLED RECEPTORS FOR SEVERAL CAROTID BODY NEUROTRANSMITTERS

Studies using ratiometric Ca²⁺ imaging indicate that several CB neurotransmitters evoke intracellular Ca²⁺ responses in isolated CB type II cells. Though a small voltage-activated outward current was seen in some cases, these cells generally lack significant voltage-activated inward currents (12, 67, 72), and, as explained below, the source of the Ca²⁺ rise is predominantly intracellular in all cases examined so far.

ATP stimulates type II cells via P2Y2 receptors.

Application of ATP induced robust intracellular Ca²⁺ elevations in isolated rat type II cells (65, 67, 72). Pharmacological tests using P2 receptor agonists indicated the order of potency was UTP ≥ ATP > 2MeSATP (52, 67, 72), consistent with the expression P2Y2 receptors and confirmed by immunolocalization studies (67). The ATP-induced Ca²⁺ response persisted in the presence of Ca²⁺-free extracellular solutions, suggesting release of Ca²⁺ from internal stores (67). Because ATP is released from CB type I cells during exposure to chemostimuli such as hypoxia and acid hypercapnia (3, 8, 47), these findings raise the possibility of paracrine stimulation of type II cells by ATP during chemotransduction (47, 65). Confirmation of this prediction was obtained in studies on monolayer cultures containing rat type I cell clusters intermixed with type II cells. Here, type I cell depolarizations induced by chemostimuli such as hypoxia or isohydric hypercapnia, or by application of high K⁺ solutions, led to delayed intracellular Ca²⁺ responses in nearby type II cells. Given that only isolated type I, but not type II, cells respond directly to these three stimuli (49, 65, 72), the observed delay is expected if the type II cell response is the result of cross talk from adjacent type I cells and involves G protein-coupled signaling pathways. Importantly, the delayed type II cell responses were sensitive to both the P2Y2 receptor blocker suramin and the nucleoside hydrolyase apyrase, suggesting they arose from the endogenous or paracrine release of ATP from nearby type I cells (40).

ACh stimulates type II cells via muscarinic receptors.

Though controversial, ACh is thought to be an excitatory CB neurotransmitter and a variety of nicotinic and muscarinic (mainly M1 and M2) ACh receptors (AChR) has been localized to type I cells and/or petrosal afferent terminals in various species (15, 20, 44, 45, 60). Applications of exogenous ACh and muscarinic AChR agonists, which typically signal via release of Ca²⁺ from internal stores, elicit a rise in intracellular Ca²⁺ in type II cells (43, 65). These data suggest that type II cells express muscarinic AChR (mAChR), stimulation of which leads to the generation of intracellular Ca²⁺ signals as previously reported for type I cells (11). It remains to be determined whether ACh release from type I cells during chemotransduction is sufficient to activate nearby type II cells via paracrine signaling pathways.

Angiotensin II stimulates type II cells via AT₁ receptors.

The enzymatic machinery for synthesis of the vasoactive peptide angiotensin II (ANG II) has been localized to CB type I cells. For example, both the precursor of ANG II, i.e., angiotensinogen, and the angiotensin-converting enzyme are expressed in type I cells (27, 29). The possibility that ANG II may display autocrine-paracrine functions in the CB was suggested by the observation that ANG II can evoke a rise in intracellular Ca²⁺ in type I cells which express losartan-sensitive AT₁ receptors (1, 14, 31). As exemplified in Fig. 1A, the more recent demonstration that ANG II can also elicit a rise in intracellular Ca²⁺ in type II cells also supports a paracrine role in the CB (41, 65). The latter ANG II-induced Ca²⁺ response had an EC₅₀ of ~8 nM and was inhibited by losartan, consistent with the presence of AT₁ receptors (41). The response also persisted in nominally Ca²⁺-free medium (Fig. 1A) and was markedly inhibited following store depletion by cyclopiazonic acid, suggesting Ca²⁺ release from internal stores (41). It remains to be determined whether this paracrine pathway is activated by ANG II released from type I cells during chemoexcitation. However, the type II cell response to exogenous ANG II persisted in the presence of P2Y2R blockade, suggesting it did not arise indirectly, i.e., from ATP released by costimulation of nearby type I cells (41).

Fig. 1.

Angiotensin II and 5-HT elicit a rise in intracellular Ca²⁺ in type II cells. Application of 100 nM angiotensin II (A) and 10 µM 5-HT (B) to type II cells elicited a rise in intracellular calcium concentration ([Ca²⁺]_i) that persisted in Ca²⁺-free medium, suggesting release from intracellular stores. Note type II cells are identified by the presence of an intracellular Ca²⁺ response to the P2Y2R agonist UTP (100 µM); also, they do not respond to the depolarizing stimulus high K⁺ (30 mM) when isolated from type I cells.

5-HT stimulates type II cells via 5-HT₂ receptors.

5-HT acts as a neuromodulator in the CB and is synthesized and stored by type I cells (32, 44, 48, 69). There is strong evidence that 5-HT plays a critical role in promoting long-term facilitation of the CB sensory discharge following exposure to chronic intermittent hypoxia (CIH), as occurs in sleep apnea patients (50, 55). The actions of 5-HT have been attributed mainly to the presence of 5-HT_2A receptors on type I cells, and possibly CSN nerve terminals as well (22, 23, 70, 71). As illustrated in Fig. 1B, recent studies suggest that type II cells also express 5-HT receptors stimulation of which leads to a rise in intracellular Ca²⁺ that persists in Ca²⁺-free medium and is inhibited following store depletion (42). In the latter study, the 5-HT-evoked Ca²⁺ response had an EC₅₀ of ~180 nM and was reversibly inhibited by the 5-HT₂ receptor antagonist ketanserin (1–10 μM). Thus it is plausible that stimulation of type II cells by the paracrine action of 5-HT released from type I cells might play a greater role in synapse integration during conditions of CIH.

Endothelin-1 stimulates intracellular Ca²⁺ signaling in type II cells.

In addition to the above, type II cells may also respond to other CB neurochemicals via an increase in intracellular Ca²⁺ signaling. Endothelin-1 (ET-1) immunoreactivity has been localized to type I cells, and it has been proposed that ET-1, acting via type I cell autoreceptors, is involved in the potentiation of the CB sensory discharge following exposures to both chronic hypoxia and intermittent hypoxia (6, 57). Moreover, recent studies suggest that the adult rat CB contains stem cells of glial-like type II cell origin that express ET-1 receptors, especially the ET_B subtype (53). In the latter study it was shown that stimulation of these ET_B receptors by ET-1 released from type I cells plays a key role in stem cell proliferation and CB growth during chronic hypoxia. In cultured CB cells, exogenous ET-1 was shown to stimulate a rise in intracellular Ca²⁺ in a significant population of type II cells (43), though the source of the Ca²⁺ rise and the ET receptor subtype was not determined.

Do any neurotransmitters inhibit intracellular Ca²⁺ signaling in type II cells?

All the neurotransmitters discussed above lead to a rise in intracellular Ca²⁺ in type II cells and, as discussed below, tend to produce excitatory effects. However, sensory output emanating from the CB is best described as a balance between excitatory and inhibitory inputs, which provide the substrate for CB plasticity (44, 45). This begs the question whether the action of some neurotransmitters might inhibit the function of type II cells by suppressing the rise in intracellular Ca²⁺. In recent attempts to test this hypothesis histamine (10 μM), a neurotransmitter that is highly expressed in type I cells (24), was found to inhibit the UTP-induced rise in intracellular Ca²⁺ in a small subpopulation (<25%) of type II cells (Fig. 2, A–D). The histamine receptor subtype mediating this inhibitory response remains to be determined though, notably, inhibition of mAChR-mediated intracellular Ca²⁺ signaling by activation of histamine H3 receptors has been reported in type I cells (61). Interestingly, exogenous dopamine (10 μM) was also found to inhibit the UTP-induced Ca²⁺ responses in a much larger proportion (~75%) of type II cells (28). Further studies are required to elucidate the mechanisms by which these inhibitory neuromodulators suppress intracellular Ca²⁺ signaling in type II cells.

Fig. 2.

Histamine attenuates purinergic signaling in a subpopulation of type II cells. A and C are representative traces demonstrating the reduction of the UTP-evoked intracellular Ca²⁺ response ([Ca²⁺]_i) in type II cells when histamine (His; 10 µM) was present. Summary data in B and D compare the mean ± SE ∆[Ca²⁺]_i (nM) responses during exposure to UTP with and without histamine (n = 5 dishes; 10–25 cells sampled per dish). In B, 18 of 82 cells and D 17 of 97 cells showed a reduction in the UTP (100 µM)-evoked Ca²⁺ response in the presence of histamine; the remaining cells did not respond to histamine. Lowercase letters denote significant differences between groups; data analyzed by 1-way ANOVA followed by Tukey’s post hoc test; P < 0.05.

NEUROTRANSMITTER-INDUCED RISE IN INTRACELLULAR Ca²⁺ LEADS TO ACTIVATION OF PANNEXIN-1 CURRENTS IN TYPE II CELLS

As discussed above, stimulation of a variety of G protein-coupled receptors in type II leads to a rise in intracellular Ca²⁺ derived mainly from internal stores. What is the downstream consequence of activating this Ca²⁺ signaling pathway? In voltage clamp experiments, several of the ligands that caused a rise in intracellular Ca²⁺ in type II cells were also shown to activate of an inward current at negative membrane potentials (Fig. 3, A–C). The current appeared to have the same characteristics and origin regardless of whether the ligand was ATP/UTP, ACh, ANG II, or 5-HT acting at P2Y2, mAChR, AT₁, or 5-HT₂ receptors, respectively (41–43, 72). Consistent with the opening of nonselective ion channels the current reversed direction near 0 mV and, as exemplified in Fig. 3D, was reversibly inhibited by low concentrations of carbenoxolone (5 μM), a blocker of pannexin-1 (Panx-1) channels (41, 42, 72). Given that carbenoxolone may also block gap junctional hemichannels, though typically at much higher concentrations (36), the purported involvement of Panx-1 channels was confirmed by the demonstration that ¹⁰Panx peptide (100 μM), a more specific Panx-1 mimetic peptide blocker, reversibly blocked the UTP-activated inward current in type II cells (42). Moreover, GFAP-immunoreactive type II cells in CB tissue sections in situ, and in dissociated CB cultures in vitro, also stained positively when treated with Panx-1-specific antibodies (72).

Fig. 3.

Activation of an inward current in type II cells by several carotid body neurochemicals. Under voltage clamp, rapid application of the neurochemicals ATP (50 µM), angiotensin II (0.5 µM), and 5-HT (10 µM) to 3 different type II cells elicited an inward current in A, B, and C, respectively; holding potential = −60 mV. In D, the inward current evoked by 100 nM angiotensin was blocked by the pannexin-1 channel inhibitor carbenoxolone (5 µM CBX).

The observation that the same ligands that activated the Panx-1 current also induced a rise in intracellular Ca²⁺ raised the question whether or not the two events were linked. Indeed, this appeared to be the case since addition of the membrane-permeable Ca²⁺ chelator BAPTA-AM (1–10 μM) to the bathing solution reversibly inhibited the inward current activated by either ATP, ANG II, or 5-HT (41, 42). There is precedence for the idea that Panx-1 channels can be activated by P2Y2 receptor (P2X2R) stimulation through an elevation of intracellular Ca²⁺, even though they lack known Ca²⁺ binding sites and appear to be Ca²⁺-independent in some cell types, e.g., hippocampal neurons (63). In particular, when Panx-1 channels and P2Y2R were coexpressed in oocytes, stimulation with ATP led to the activation of an inward Panx-1 current at negative holding potentials (35); moreover, when the cytoplasmic face of excised inside-out patches from these oocytes was exposed to 0 Ca²⁺, Panx-1 channels remained closed at negative potentials but became strongly activated when the Ca²⁺ concentration was elevated to micromolar levels. In addition, dye uptake through Panx-1 channels in neuroblastoma cells is Ca²⁺ dependent (26), and Panx-1 channel opening is regulated by intracellular Ca²⁺ in other glial cell types such as microglia and astrocytes (10). Thus, it is possible that the regulation of Panx-1 channel opening by intracellular Ca²⁺ may differ between excitable and nonexcitable cells.

ROLE OF PANX-1 CHANNELS AS CONDUITS FOR ATP RELEASE FROM TYPE II CELLS

Panx-1 channels, which have a similar topology and structure as gap junction channels formed from connexin subunits, typically assemble as hexamers at the cell surface where they function as ion channels but without the need for an adjoining cell (10, 51). The pores of Panx-1 channels are large enough to allow release of large signaling molecules (<1.5 kDa) such as ATP from a variety of cell types including red blood cells, brain astrocytes, taste cells, and central neurons (10, 18, 19, 34, 35, 64). Given the role of ATP as a key excitatory CB neurotransmitter (44, 45), it was of interest to determine whether the P2Y2R-mediated and Ca²⁺-dependent activation of Panx-1 channels in type II cells provided an auxiliary pathway for boosting synaptic ATP levels. In one model designed to test this hypothesis petrosal neurons, which express ligand-gated purinergic P2X2/3 receptors (P2X2/3R), were cocultured with type I cell clusters containing contiguous type II cells (72). In that study, selective stimulation of type II cells using the P2Y2 receptor agonist UTP led in some cases to membrane depolarization and/or increased excitability in adjacent petrosal neurons. These responses were inhibited by blockers of either P2X2/3R (10 μM PPADS) or Panx-1 channels (5 μM carbenoxolone) suggesting that Panx-1 channels in type II cells also function as ATP release channels. Taken together, these data point to a potential physiological role of type II cells as the vehicle for “ATP-induced ATP release” following activation of paracrine signaling pathways during chemotransduction.

FATE OF ATP RELEASED FROM TYPE II CELLS AT THE CHEMOSENSORY SYNAPSE: ROLE OF ECTONUCLEOTIDASES

As discussed above, one potential fate of ATP released via Panx-1 channels in type II cells is to bind postsynaptic P2X2/3R and contribute to the afferent sensory discharge. This process could be facilitated by the physical coupling between P2X2/3R and Panx-1 channels as demonstrated in coimmunoprecipitation studies (30) and the proximity between type I cell processes and the afferent terminals (53). In addition, opening of Panx-1 channels could aid in the propagation of “Ca²⁺ waves” within chemoreceptor clusters via the mechanism of ATP-induced ATP release coupled to P2Y2R stimulation. At least three different mechanisms could potentially help prevent or limit overexcitation induced by ATP including: 1) desensitization of postsynaptic P2X2/3 receptors on afferent terminals (74); 2) negative feedback inhibition of type I cell function due to stimulation of purinergic P2Y1R and/or P2Y12R leading to membrane hyperpolarization and decreased voltage-gated Ca²⁺ entry (4, 65, 68); and 3) inhibition of Panx-1 channels in type II cells by high ATP concentrations (10).

ATP released through Panx-1 channels could also contribute to chemoexcitation via alternative pathways (40). For example, ATP can be broken down to other physiologically active nucleotides, e.g., ADP, AMP, and adenosine, by surface-located ectonucleotidases. Recent studies using quantitative PCR techniques revealed that surface-located members of the ectonucleotidase family, i.e., NTPDase1,2,3, and E5′Nt/CD73, were expressed in the rat CB, and positive immunoreactivity against NTPDase2,3 and E5′Nt/CD73 was localized to the periphery of type I cell clusters (58). Of these, NTPDase2 can efficiently hydrolyze extracellular ATP to ADP, whereas NTPDase3 can efficiently hydrolyze ATP to ADP and ADP to AMP (13, 76). In the final rate-limiting step, E5′Nt/CD73 hydrolyses AMP to adenosine, which can further facilitate CB neurotransmission by acting on excitatory A2a receptors on type I cells and petrosal afferent terminals (7, 8, 15, 17, 40, 47, 66, 73). If adenosine levels become too high, activation of low-affinity A2b receptors may lead to an enhancement of DA secretion from type I cells (9, 33), and blunting of the sensory discharge by acting on inhibitory pre- and postsynaptic D2 receptors (7, 8, 73). Equilibrative nucleoside transporter(s) on type I cells can aid in the regulation of extracellular adenosine levels (7). In summary, ATP released from type II cells through Panx-1 channels may contribute to purinergic signaling in the CB via multiple paracrine signaling pathways involving the afferent nerve terminal, as well as neighboring type I and type II cells.

CONCLUSIONS

In this review we summarized the evidence supporting a role for glial-like type II cells in CB chemoreception as a result of paracrine cell-cell interactions. These cells respond to several neurotransmitters synthesized and secreted from chemoreceptor type I cells, with which they form an intimate association. Most of these neurotransmitters, including the excitatory neurotransmitter ATP, stimulate intracellular Ca²⁺ signaling cascades in type II cells leading to opening of Panx-1 channels that can promote the further release of ATP. Catabolism of ATP by ectonucleotidases generates other physiologically active ligands such as ADP and adenosine, which can interact with presynaptic or postsynaptic purinergic receptors at the sensory synapse. A summary diagram highlighting the effects of some of these neurotransmitters is shown in Fig. 4. This model was derived from in vitro experiments on dissociated rat CB cells grown in monolayer cultures. Further studies using more intact in vitro and in vivo preparations are required to confirm and extend the main conclusions. For example, notwithstanding the possibility for compensation, it would be of interest to know whether transgenic Panx-1 knockout mice elicit a normal hypoxic or hypercapnic ventilatory response. Also, is direct stimulation of type II cells alone in the intact CB, e.g., by optogenetic activation of Panx-1 channels (75), sufficient to trigger an increase in CSN discharge or ventilation in live animals? Finally, given the plastic changes in the neurochemistry of the CB sensory synapse during exposures to chronic and intermittent hypoxia (25), the role of type II cells in paracrine signaling needs to be examined in those conditions.

Fig. 4.

A schematic showing the proposed signaling pathways in type II cells activated by neurochemicals at the carotid body sensory synapse. During acute hypoxia (or acid hypercapnia) type I cells depolarize and release several neurochemicals, e.g., ATP, ACh, angiotensin II, endothelin-1, 5-HT, dopamine, and histamine. It is possible that some of these are more relevant after exposures to chronic or intermittent hypoxia. Paracrine activation of several of the corresponding G protein-coupled receptors shown on the elongated type II cell causes a rise in intracellular Ca²⁺, presumably mediated via the PLC-IP3 signaling pathway. However, in some cases, e.g., dopamine receptor (not shown) and histamine receptor activation, the response may be an inhibition of Ca²⁺ signaling. The rise in intracellular Ca²⁺ in turn leads to opening of pannexin-1 (Panx-1) channels, which act as conduits for the further release of ATP. Extracellular ATP may act on excitatory P2X receptors on the nerve terminal or may be broken down to ADP, AMP, and adenosine by a series of surface-located nucleotidases, i.e., NTPDase2,3 and ecto-5′-nucleotidase (E5′Nt). Adenosine can further contribute to purinergic signaling by acting on excitatory A2 receptors on type I cells and the sensory nerve terminal. For clarity, the model does not show the normally close relationship between type I and type II cells. The labels M1/2, H1/2/3, P2Y2, 5-HT2a, ETA/B, and AT1 refer to metabotropic receptors for the ligands ACh, histamine, ATP, 5-HT, endothelin-1, and angiotensin II, respectively.

GRANTS

Grant support to C. A. Nurse was provided by the Canadian Institutes of Health Research (MOP 12037, MOP 142469) and the Natural Sciences and Engineering Research Council of Canada.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.A.N., E.M.L., and S.S. conceived and designed research; C.A.N., E.M.L., and S.S. analyzed data; C.A.N., E.M.L., and S.S. interpreted results of experiments; C.A.N., E.M.L., and S.S. prepared figures; C.A.N. drafted manuscript; C.A.N., E.M.L., and S.S. edited and revised manuscript; C.A.N., E.M.L., and S.S. approved final version of manuscript; E.M.L. and S.S. performed experiments.