PURINERGIC RECEPTORS IN AIRWAY HYDRATION

Abstract

Airway epithelial purinergic receptors control key components of the mucociliary clearance (MCC), the dominant component of pulmonary host defense. In healthy airways, the periciliary liquid (PCL) is optimally hydrated, thus acting as an efficient lubricant layer over which the mucus layer moves by ciliary force. When the hydration of the airway surface decreases, the mucus becomes hyperconcentrated, the PCL collapses, and the “thickened” mucus layer adheres to cell surfaces, causing plaque/plug formation. Mucus accumulation is a major contributing factor to the progression of chronic obstructive lung diseases such as cystic fibrosis (CF) and chronic bronchitis (CB). Mucus hydration is regulated by finely tuned mechanisms of luminal Cl⁻ secretion and Na⁺ absorption with concomitant osmotically driven water flow. These activities are regulated by airway surface liquid (ASL) concentrations of adenosine and ATP, acting on airway epithelial A_2B and P2Y₂ receptors, respectively. The goal of this article is to provide an overview of our understanding of the role of purinergic receptors in the regulation of airway epithelial ion/fluid transport and the mechanisms of nucleotide release and metabolic activities that contribute to airway surface hydration in healthy and chronically obstructed airways.

Introduction

The purinergic concept initially proposed by Geoff Burnstock on the basis of smooth muscle responses to autonomic nerve stimulation [1-4] has been expanded to essentially all peripheral tissues, including non-innervated tissues [5].

The significance of nucleotides/nucleosides as extracellular molecules in the airways is underscored by the presence of a subset of airway epithelial purinergic receptors that control key components of mucociliary clearance (MCC), the primary innate defense mechanism in the lung. In healthy airways, the periciliary liquid (PCL) is optimally hydrated, thus acting as an efficient lubricant layer over which the mucus layer moves by ciliary force [6]. In normal airways, the mucus layer is ~ 2% solids (98% water). When the hydration of the airway surface decreases, the mucus becomes hyperconcentrated, the PCL collapses, and the “thickened” mucus layer adheres to cell surfaces, causing plaque/plug formation. Mucus accumulation is a major contributing factor to the progression of cystic fibrosis (CF) [7, 8], and cigarette smoke (CS)-induced chronic bronchitis (CB), a form of chronic obstructive pulmonary disease (COPD) [9]. CF is caused by a genetically deficient expression/function of the CF transmembrane conductance regulator (CFTR), a cyclic AMP-regulator Cl⁻ channel [10], that regulates airway surface hydration [11]. While the pathogenesis of CB is complex and has only been partially described, recent evidence suggests that CS-induced CFTR dysfunction may contribute to the onset of CB [12-14] (Fig. 1A).

Figure 1. Regulation of mucus hydration in airway epithelia.

A, airway surface hydration is maintained by water fluxes mainly driven by active Cl⁻ and Na⁺ transport. When the hydration of the airway surface decreases, e.g., secondary to CFTR dysfunction, Na⁺ absorption increases, the mucus becomes hyperconcentrated, the PCL collapses, and the “thickened” mucus layer adheres to the cell surface, causing plaque/plug formation, as seen in CF and CB. B, ATP and adenosine (Ado) acting on airway epithelial P2Y₂R and A_2BR, respectively, promote Cl⁻ secretion and inhibition of Na⁺ absorption.

Mucus hydration in health is regulated by finely tuned mechanisms of luminal Cl⁻ secretion and Na⁺ absorption with concomitant osmotically driven water flow. These activities are coordinately controlled by airway epithelial purinergic receptors that promote luminal Cl⁻ secretion (via CFTR and CaCC/TMEM16) and inhibit Na⁺ absorption (via ENaC) (Fig. 1B).

Almost three decades have elapsed since nucleotide and nucleoside actions on airway epithelia were first described [15, 16]. Extensive research has subsequently identified the purinergic receptors and signaling pathways that account for the physiologic actions of nucleotides and nucleosides within the airway surface liquid (ASL), and the mechanisms by which nucleotide release and metabolism control airway surface hydration in health and disease.

The aim of this article is to present an overview of our understanding of the contribution of purinergic receptors to ASL homeostasis and the potential therapeutic benefits of targeting purinergic signaling components to promote mucus hydration in CF- and CB-diseased airways.

ASL adenosine promotes CFTR-mediated airway surface hydration

CFTR is a major element mediating liquid balance on normal airway surfaces. Patch clamp studies with Calu-3 airway epithelial cells indicated that the CFTR Cl⁻ channel activity in naïve cells was abolished by including an adenosine receptor antagonist [100 μM 8-SPT (8-(p-sulfophenyl)theophylline] or an inhibitor of AMP hydrolysis (300 μM α,β-methyleneadenosine 5'-diphosphate) in the patch pipette [17]. Studies with primary cultures of human bronchial epithelial (HBE) cells illustrated that naturally occurring adenosine promotes tonic adenosine receptor activity, as judged by the reduction of cyclic AMP formation observed following the addition of adenosine deaminase to culture surfaces [18]. Importantly, confocal microscopy measurements of ASL height (an index of liquid volume production) showed that normal, resting HBE cells exhibited ~ 7 μm ASL height, the height of extended cilia. Addition of adenosine deaminase or 8-SPT reduced ASL height to ~ 4 μ m, the ASL height observed in naïve CF cultures [18]. These observations suggested that, in normal cells, baseline CFTR activity is controlled by the levels of endogenous adenosine generated from the hydrolysis of released nucleotides.

A body of evidence suggest that adenosine actions on airway epithelial cells are mediated by the Gs-coupled A_2B receptor (A_2BR) (Fig. 1B). The A_2BR is the most abundant adenosine receptor expressed in airway epithelia [19, 20]. Adenosine promotes cyclic AMP formation, CFTR-dependent Cl⁻ transport, and ASL volume regulation with agonist potency order consistent with activation of the A_2B receptor [16, 19-22]. Furthermore, adenosine-promoted fluid secretion in primary HBE cells was blocked by the A_2BR antagonist ATL801, but not by A₁R, A_2AR, or A₃R antagonists [20].

ASL adenosine is generated by dephosphorylation of released ATP by ecto-nucleotidases [23, 24]. Adenosine is converted to inosine by adenosine deaminase. Adenosine and inosine are taken into cells via concentrative nucleoside transporters 2 and 3 [25]. Inosine is not an agonist on the A_2BR although is a weak and partial agonist on A₁R, A_2AR and A₃R [26, 27]; anti-inflammatory actions of inosine in the lung via A_2A and A₃ receptors have been described [28, 29].

Direct measurements of adenine nucleotides/sides in luminal solutions bathing airway epithelial cell cultures indicated adenosine concentrations in the 180-350 nM range [18], which is below EC₅₀ values for A_2BR activation (1-2 μM) [19, 20, 30]. However, these adenosine concentration values are based on bulk measurements (i.e., 22 μl HBSS/cm² added to culture surfaces), which likely underestimated the adenosine levels in the physiological “thin film”. As suggested by in vivo nasal ASL adenosine measurements, it is likely that adenosine concentrations at the airway epithelial cell surface reach the micromolar range [20].

ASL ATP decayed with a half-life that was 10 times faster than adenosine half-life [18]. Despite its robust basal rates of release (200-400 fmol/min/cm²), cell surface ATP concentrations in resting cells (assessed in real-time via cell-attached luciferase) were found to be 1-10 nM [31], i.e., suboptimal for P2Y₂R activation [18]. Indeed, baseline levels of ATP fail to promote ASL volume regulation in CF airway epithelia [18, 32], indicating that ATP levels are not in range of promoting P2Y₂R mediated CaCC activation and ENaC inhibition in resting cells.

Collectively, these observations indicated that ATP release and subsequent formation of adenosine provide a mechanism for fluid secretion in normal (but not CF) airway epithelia, via A_2BR-promoted CFTR activity.

Nucleotides promote CFTR-independent ion transport activities

Nucleotide-promoted Cl⁻ secretion in a CFTR-independent manner was first described by Mason and co-workers in 1991. Luminal nucleotides evoked changes in short circuit currents with similar effectiveness in normal and CF airway epithelial cells and exhibited the same potency order: UTP ≥ ATP > ATPγS > 2MeSATP > ADPβS > βγMeATP. Nucleotide-stimulated Cl⁻ secretion was associated with Ca²⁺ mobilization [15] secondary to phospholipase C activation [33]. The receptor mediating these responses was identified as the P_2U receptor [34-36], now known as P2Y₂R [37].

The molecular identity of the Ca²⁺-regulated Cl⁻ channel downstream of P2Y₂R activation in CF cultures was revealed by three independent groups in 2008. Using siRNA approaches to knock down putative plasma membrane proteins of unknown function, Caputo and co-workers identified TMEM16A as a protein exhibiting calcium-dependent Cl⁻ channel (CaCC) activity [38]. Simultaneously, Yang et al. [39] and Schroeder et al. [40] demonstrated that expression of TMEM16A confers receptor-activated calcium-dependent chloride conductance. In addition to CaCC-mediated Cl⁻ secretory responses, the P2Y₂R activates CFTR via PKC [41] and Ca²⁺-activated adenylyl cyclase [42] (Fig. 1B).

ENaC constitutes the rate-limiting step for Na⁺ absorption across airway epithelia, which in turn regulates ASL volume and the efficiency of mucociliary clearance [43]. Importantly, binding of the β subunit of ENaC to phosphatidylinositol 4,5-bisphosphate (PIP₂) is essential for ENaC channel activity [44-46]. Thus, P2Y₂R activation further contributes to airway surface hydration via phospholipase C-catalyzed depletion of PIP₂, leading to inhibition of Na⁺ absorption [47-49] (Fig. 1B).

The P2Y₂R was subsequently shown to regulate ciliary beat via Ca²/PKC (reviewed in [50]). P2Y₂R, via Ca²⁺/PKC activation, was also shown to be a major regulator of airway mucin secretion [50, 51]. Thus, P2Y₂R regulates multiple components of the MCC system.

While screening various cell lines for the presence of functional P2_U receptors, we discovered a uridine nucleotide-selective G protein-linked receptor that activated phospholipase C in C6-2B rat glioma cells [52]. This receptor was potently activated by UDP and with lower potency by UTP, but not by ATP. This pharmacological finding triggered a major interest in the field for identifying genes encoding uridine nucleotide-selective receptors leading to the cloning and characterization of the UDP-selective P2Y₆R and the UTP-selective P2Y₄R [53-57]. It was subsequently shown that the P2Y₆R is functionally expressed on the apical surface of airway epithelial cells, promoting Ca²⁺-regulated Cl⁻ secretion [58] and ciliary beating [50]. The magnitude of UDP/P2Y₆R-promoted responses in airway epithelia was ~ 30% of the UTP/P2Y₂R-evoked responses [50, 58]. In addition to airway epithelia, the P2Y₆R is expressed in lung fibroblasts; P2Y₆R upregulation in the lung was recently associated with pulmonary fibrosis and inflammation [59].

Nucleotide homeostasis in the airways

Nucleotides in ASL are also potent pro-inflammatory mediators which act on purinoceptors expressed on immune/inflammatory cells or lung stromal cells to promote cytokine release [60, 61]. Therefore, a finely tuned mechanism of nucleotide release and ecto-metabolism must exist to maintain effective MCC without promoting airway inflammation [62].

Initial evidence of cellular release of ATP from non-secretory/non-excitatory cells arose from the observation that 1321N1 astrocytoma cells over-expressing the recombinant P2Y₂R exhibited enhanced second messenger signaling, i.e., inositol phosphate formation, which was sensitive to apyrase [63] and was secondary to mechanical stimulation, e.g., a medium change [36]. The concept that mechanical stimuli trigger non-lytic release of ATP was expanded to airway epithelial cells, which were shown to release robust amounts of ATP, acutely, in response to a plate tilting, touching of the monolayer [64], or medium changes [65]. It has since been well-established that epithelial cells release discrete amounts of ATP constitutively and that enhanced ATP release occurs following pharmacological challenge [66] or controlled mechanical stimulation such as hypotonic cell swelling [18, 31], or phasic motion that mimics the shear stress associated with normal tidal breathing [32].

The realization that a subset of broadly distributed GPCRs recognize UTP (P2Y₂R, P2Y₄R) and UDP (P2Y₆R) as potent agonists suggested that uridine nucleotides are released from cells in addition to ATP. By developing highly sensitive methodologies for the quantification of uridine nucleotides, our lab demonstrated that UTP is released in a regulated manner from a variety of cells, including airway epithelial cells [67, 68]. The ratio of UTP to ATP concentrations assessed in the medium bathing primary human nasal epithelial cells and epithelial cell lines (1:5 to 1:10) correlated with the intracellular UTP to ATP ratio. This observation suggested the occurrence of a common mechanism of release, likely a common transport mechanism driven by the relative intracellular concentrations of ATP and UTP [68].

This hypothesis was revised following the discovery that, in addition to ATP and UTP, airway epithelia and other cells release UDP-glucose [69, 70], a potent agonist of the P2Y₁₄R [71] that is expressed in inflammatory cells, including neutrophils [72] and mast cells [73]. UDP-sugars are concentrated up to 20 times in the lumen of the endoplasmic reticulum (ER) and Golgi (relative to the cytosol), serving as sugar donor substrates for glycosylation reactions. Luminal UDP-sugars are converted during this process to UDP, which in turn is hydrolyzed to UMP. UDP-sugar/UMP antiporters translocate cytosolic UDP-sugars to the lumen of the secretory pathway. Therefore, Golgi UDP-sugars, UDP, and UMP are likely to be delivered as cargo during the transport of glycoproteins to the cell surface [74]. The involvement of the secretory pathway in the release of UDP-sugars was confirmed by experiments with airway epithelial cells and yeast mutants demonstrating that overexpression/deletion of ER/Golgi UDP-N-Acetyl-glucosamine transporters contributed to the release of UDP-N-Acetyl-glucosamine [70]. Based on these observations, it was speculated that transport of ATP into ER/Golgi vesicles, e.g., to serve in luminal phosphorylation reactions, provided a mechanism for the vesicular release of adenine nucleotides.

The pathophysiologic relevance of uridine nucleotide and nucleotide sugar release into the airways is emphasized by studies showing abnormally elevated levels of UTP and UDP in BAL from mice exhibiting RSV infection-associated lung edema [75] and bleomycin-induced pulmonary fibrosis [59], respectively. Furthermore, UDP-glucose levels are markedly increased in lung secretions from CF patients and mouse models of CF/CB-like lung disease, the βENaC transgenic mouse [76]. Administration of the P2Y₁₄R antagonist PPTN [4-(piperidin-4-yl)-phenyl)-7-(4-(trifluoromethyl)-phenyl-2-naphtoic acid] attenuated neutrophil infiltration in these mice [76].

The diversity of conditions in which airway epithelial nucleotide release was observed suggested that both exocytotic pathways and a plasma membrane channel were involved in nucleotide release, perhaps in a cell- and stimulus-specific fashion. For example, primary cultures of naïve HBE cells, which are dominated by ciliated cells, exhibited robust release of ATP upon hypotonicity-induced cell swelling, and this release was, largely, not affected by chelation of the intracellular Ca²⁺ or by pharmacological inhibition of the secretory pathway [31, 77, 78], suggesting a conductive pathway. However, HBE cells exposed to inflammatory challenges that promote goblet cell metaplasia exhibited enhanced hypotonicity-induced ATP release, which was blocked by intracellular Ca²⁺ chelation and by reagents that disrupted vesicle trafficking/exocytosis [77, 78]. Furthermore, agents that increased intracellular Ca²⁺ in HBE cells, such as ionomycin and UTP, caused a minor release of ATP in naïve ciliated cell cultures, but ionomycin- and UTP-promoted ATP release increased markedly in goblet cell-metaplasic cultures [31].

Thus, functional data suggested that conducted and vesicular pathways contribute to ATP release from airway epithelial non-mucous and goblet cells, respectively. Major components of these pathways include the plasma membrane ATP channel pannexin 1 and the vesicular nucleotide transporter VNUT, as discussed below. It is important to note that, in the absence of selective molecular approaches (e.g., knockdown, knockout), studies relaying solely in the action of non-selective reagents should be taken with caution. For example, the commonly used pannexin 1 inhibitor probenecid inhibits several additional targets [79-81].

Pannexin 1-mediated airway epithelial ATP release

The discovery by Dahl and co-workers that pannexin 1 acted as a plasma membrane ATP channel when expressed in Xenopus oocytes [82] led to the elucidation of the conductive pathway responsible for airway epithelial ATP release.

In 2009, Ransford et al. reported that ATP release from hypotonically swollen HBE cells was markedly reduced by pannexin 1 knockdown via shRNA as well as by the pannexin 1 channel inhibitors carbenoxolone (10 μM) and probenecid (1 mM) [83]. Simultaneously, we reported that activation of G protein-coupled protease-activated receptors (PAR) in HBE cells resulted in enhanced release of ATP and enhanced uptake of the pannexin 1 probe propidium iodide, and these responses were inhibited by carbenoxolone (10 μM) [66]. Subsequently, we illustrated that ATP release and dye uptake in HBE cells subjected to hypotonic cell swelling were markedly reduced by the pannexin 1-selective blocking peptide ¹⁰Panxl [84]. Further assessment of the contribution of pannexin 1 to airway epithelial ATP release was obtained using pannexin 1^−/− mice. Utilizing a perfusion approach to assess ATP levels in tracheal luminal secretions under controlled flow conditions, we showed that ATP release from WT tracheas increased up to six-fold following a brief exposure to hypotonicity. In contrast, pannexin 1^−/− animals exhibited markedly impaired hypotonicity-evoked ATP release [84].

We also showed that disrupting Rho signaling via RhoA-dominant negative mutants or the Rho kinase inhibitor H1152 (1 μM) markedly decreased propidium iodide uptake and ATP release from airway epithelial cells in response to thrombin or hypotonic challenges [66, 84]. Given the involvement of Rho/Rho kinase in cytoskeleton rearrangements, one speculation is that Rho activation facilitates pannexin 1 translocation to the plasma membrane and/or its interaction with regulatory proteins.

Transient receptor potential vanilloid 4 (TRPV4) is a widely expressed cation channel (relatively selective for Ca²⁺ and, to a lesser extent, Na⁺) that acts as a sensor of various physical stimuli, including osmotic/shear stresses and stretching [85, 86]. TRPV4 knockout mice exhibited decreased ATP release in the kidney and bladder epithelia [87, 88]. TRPV4 regulation of pannexin1 was recently reported in human pulmonary fibroblasts [89]. In airway epithelial cells, TRPV4 shRNA and the highly selective TRPV4 inhibitor HC67047 (10 μM) reduced hypotonic stress-evoked ATP release and propidium iodide uptake. Notably, HC67047 also abolished RhoA activation in hypotonicity-challenged cells, suggesting that TRPV4 channels transduce hypotonic cell swelling into Rho activation upstream of Panx1 channel activation [84]. Baxter et al. [90] reported that exposure of HBE cells to cigarette smoke-bubbled medium (CSM) for 3-6 h resulted in elevation of ATP release in the cell supernatants, but ATP levels returned to control values at 24 h. The transient elevation of CSM ATP was attenuated by blockers of TRPV4 and, to a lesser extent, TRPV1. Pannexin 1 inhibitors (10 μM carbenoxolone and 1 mM probenecid) also reduced the levels of ATP in CSM. Wild type mouse acutely exposed to CS exhibited enhanced ATP levels in bronchoalveolar lavage fluid, which were reduced in Trpv1 KO, Trpv4 KO, and pannexin 1^−/− mice. The authors proposed that acute CS exposure causes TRPV1 and TRPV4 activation, leading to ATP release via pannexin 1 [90].

Airway inflammation causes mucociliary dysfunction. Krick et al. observed that HBE cells exposed to interferon gamma (IFN-γ) exhibited decreased ATP release in response to hypotonic cell swelling [91], INF-γ-inhibited ATP release was secondary to the upregulation of dual oxidase 2 (Duox2), a member of the NADPH oxidase gene family, and was not associated with changes in pannexin 1 expression. Duox2 upregulation led to intracellular acidification. A direct correlation was found between intracellular acidification and a reduction of pannexin 1 open probability in oocytes. Knock down of Duox2 led to an increase in hypotonicity-promoted ATP release in control HBE cells and restoration of ATP release in cells treated with IFN-γ. Both effects were reduced by 1 mM probenecid. It was concluded that IFN-γ-induced Duox2 upregulation and cell acidification lead to reduced pannexin 1-mediated ATP release, thus contributing to mucociliary dysfunction in response to inflammation [91].

Relevant to the alveolar region of the lung, pannexin 1 has been mechanistically tied to the regulation of surfactant secretion, an essential function for lung health. Stretching of alveoli during lung inflation is the main trigger for surfactant secretion from alveolar type II (ATII) cells; surfactant is stored in lamellar bodies within ATII cells. Diem et al. recently reported that stretch-promoted ATP release from hAELVi cells, an established ATI-like cell line, was markedly reduced by the pannexin 1 blocking peptide ¹⁰Panxl (100 μM) and siRNA knockdown of pannexin 1. Stretch-evoked pannexin 1-mediated ATP release from hAELVi cells was sensed by piezo1 (a mechanosensitive ion channel, unrelated to TRP channels) within caveolae [92]. Importantly, co-cultures of hAELVi and primary rat ATII cells subjected to stretch exhibited enhanced surfactant secretion relative to ATII monocultures alone, and this synergism was markedly reduced by inhibiting piezo1 or pannexin 1. It was concluded that stretching ATI cells resulted in Ca²⁺ entry via piezo1, leading to pannexin 1 activation and ATP release. ATP released from ATI cells promoted activation of P2Y receptors in neighboring ATII cells, resulting in surfactant secretion [92]. However, in a study with co-cultured human primary ATI and ATII cells, Tan et al. showed that ATP release from stretched alveolar cells tightly correlated with the number of ATII cells in culture; stretch-promoted ATP release from ATII-dominated cultures was blocked by Ca²⁺ chelators and was insensitive to carbenoxolone (100 μM) or probenecid (2 mM) [93], suggesting a vesicular pathway. It has been also shown that lamellar bodies store ATP and that ATP is released from ATII cell lamellar bodies upon exocytosis [94]. The extent to which observations with cultured cells reflect the in vivo scenario in which ATI cells cover > 95% of the alveolar surface [95] remains to be addressed.

The studies described above emphasize the diversity of mechanisms controlling pannexin 1 activity in epithelial cells. A comprehensive analysis of the functions and mechanisms of activation of pannexin 1 in different cell types and its channel properties has recently been published [96].

Conductive ATP release pathways additional to pannexin 1 have been described. For example, Foskett and co-workers reported that Ca²⁺ homeostasis modulator 1 (CALHM1) functions as an ATP channel that mediates non-synaptic neurotransmitter release [97]. Subsequently, Workman et al. showed that murine nasal epithelial cells from pannexin 1 KO or from Calhm1 KO mice exhibited decreased ATP release in response to air puffs. Based on these and additional observations, the authors suggested a complementary role for CALHM1 and pannexin 1 in ATP release-promoted increased ciliary beat frequency in nasal epithelia [98]. The expression of CALHM1 in lower airway epithelia remains to be investigated. Using a gain of function approach, Dunn et al. recently identified ABCG1 (ABC subfamily G member 1) as modulator of hypotonicity-induced ATP release through volume-regulated anion channels (VRACs) in HEK-293 cells [99]. However, the contribution of this pathway to airway epithelial ATP release is not known.

The vesicular nucleotide transporter VNUT mediates ATP release from goblet cell mucin granules.

Our initial studies with goblet cell-rich epithelial cell cultures [77, 100] established a strong association between ATP release and Ca²⁺-regulated mucin secretion. The identification by Moriyama and co-workers of SLC17A9 [101] as a vesicular nucleotide transporter that transfers cytosolic ATP into secretory granules provided an approach to more conclusively examine this association. VNUT mRNA was detected in Calu-3 cells and strong VNUT immunoreactivity was observed in mucin granules isolated from these cells [102]. Importantly, isolated mucin granules stored nucleotides. While ATP represented 80% of the total adenine nucleotide pool in Calu-3 cell lysates, ADP and AMP accounted for 90% of the purine pool within the mucin granule fraction, suggesting that ATP is largely dephosphorylated upon entry into the mucin granule. In addition, UTP was detected in the mucin granule fraction at levels equivalent to 30% of ATP levels [102].

Calu-3 cells constitute a mixed population of CFTR-rich (non-mucous) and mucin granule-rich (goblet) cells [100]. We have shown that pannexin 1 shRNA reduced ATP release in Calu-3 cells exposed to hypotonicity [84] and that thrombin-stimulated ATP release from Calu-3 cells was reduced by 10 μM carbenoxolone [103]. However, residual ATP release that was associated with Ca²⁺-regulated exocytosis of mucin granules persisted in thrombin-stimulated Calu-3 cells after complete inhibition of pannexin 1 [103], Ca²⁺-regulated ATP release from Calu-3 cells was blunted after treatment with 2 μM bafilomycin A₁ (an inhibitor of the H⁺-ATPase that loads ATP into specialized granules in secretory cells) [100] and by downregulation of VNUT by shRNA [102]. Because ATP within mucin granules is largely (90%) converted to (and released as) ADP/AMP [102, 103], release of predominantly ADP and AMP from mucin granules minimizes autocrine, P2Y₂R-mediated feedback for mucin secretion. Importantly, adenosine formation from AMP provides a paracrine mechanism to promote ion/water secretion from neighboring non-goblet cells to hydrate newly released mucins (Fig. 2). It is worth noting that VNUT immunoreactivity was also observed in vesicles of Calu-3 cells that were devoid of mucin granules [102], suggesting that vesicular compartments other than mucin granules additionally contribute to ATP release.

Figure 2. Purinergic regulation of mucus hydration and clearance.

ATP is released from non-mucous cells via the plasma membrane channel pannexin 1 (PANX1). Goblet cell mucin granules release ATP and, to a greater extent, ADP and AMP. A subset of ecto-enzymes (not shown for clarity) convert released ATP to ADP, AMP, and adenosine (Ado). The A_2BR and the P2Y₂R on non-mucous cells are major regulators of ASL volume production, via activation of the CFTR and CACC, and promote inhibition of Na⁺ absorption. P2Y₂RS expressed on goblet cells promote mucin secretion.

Collectively, our results are consistent with the notion that pannexin 1- and VNUT-dependent pathways contribute separately to the release of ATP from airway epithelial cells (Fig. 2).

Increased nucleotide hydrolysis in muco-obstructed airways

The concentration of ATP released into the ASL and the concentration of ATP metabolites are tightly controlled by cell surface and perhaps shed enzymes (also see below). As mentioned earlier, adenosine concentrations in ASL confer A_2BR-mediated regulation of CFTR in healthy airways. Notably, adenosine is formed and sustained on the airway surface by a robust ATP hydrolysis activity, which exhibits first-order rate-constant values of 0.3-0.5 min⁻¹ that keeps released ATP concentrations in the low nanomolar range [18, 23]. Therefore, the pathway that promotes CaCC-mediated Cl⁻ secretion and ENaC inhibition downstream of P2Y₂R activation is inefficient to compensate for the impaired CFTR activity of CF airways under resting conditions, as previously shown [18]. Furthermore, ecto-ATPase activities are markedly increased in epithelial cell cultures exposed to CF-like inflammatory environments [32, 104]. The notion that airway inflammation accelerates ATP hydrolysis in vivo is supported by our recent measurements in cell-free lung secretions. A nearly 30-fold reduction of ATP concentrations was observed in CF relative to healthy controls. Notably, the total amount of purines (ATP + ADP + AMP + adenosine) in CF samples was 4-fold higher than in controls [23], suggesting that nucleotide release is upregulated in CF airways, e.g., via VNUT upregulation [78] and, likely, release of nucleotides from inflammatory cells. Therefore, the reduced levels of ATP observed in CF secretions reflect increased ATP hydrolysis. Consistent with this notion, a 4-fold increased ATPase activity was measured in CF sputum samples [23]. A potential association between airway ATPase activities and mucus dehydration also emerged from the observation that lung secretions from subjects with COPD/CB exhibited decreased levels of ATP and increased ATP hydrolysis rates, relative to healthy subjects, which was associated with increased mucus concentration [105].

P2Y₂R as therapeutic target for dehydrated CF airways

Restoration of airway surface hydration is a goal of CF therapies but remains difficult to achieve due to the complexity of the regulatory systems that maintain ASL homeostasis. The efficacy of P2Y₂R to promote activation of Cl⁻ secretion and inhibition of Na⁺ absorption in CF epithelial cells identified this receptor as a candidate therapeutic target to improve airway hydration in CF and, potentially, CB. One strategy to achieve this goal was to administrate hydrolysis resistant P2Y₂R agonists via aerosol onto CF airways.

The realization that diadenosine tetraphosphate (A_p4A) promoted robust and potent activation of P2Y₂R [36] and exhibited relatively high stability on airway epithelial surfaces [106] suggested that nucleoside polyphosphate structures could be developed as potent, hydrolysis-resistant P2Y₂R agonists to be aerosolized onto CF airway surfaces. Accordingly, Inspire Pharmaceuticals synthetized INS37217 (dCp4U, denufosol), a P2Y₂R agonist engineered with enhanced metabolic stability and resistance to extracellular nucleotidase-mediated hydrolysis [107]. A Phase 2 clinical trial showed that inhalation of denufosol (up to 60 mg three times daily) improved lung functions in CF patients after 28 days of treatment [108]. However, denufosol (60 mg via inhalation three times daily) did not consistently improve pulmonary function in long-term (48-week) studies [109], likely due to the receptor desensitization produced by delivery of saturating concentrations of this P2Y₂R agonist onto airway surfaces [62].

An alternative therapeutic approach is to raise steady-state ATP levels on airway surfaces by reducing the metabolism of released ATP. Relevant to this strategy, a substantial literature, including our own studies, indicates that airway epithelial ATP release rates are not reduced in CF [18, 32, 64, 65, 78, 110]. However, effective inhibitors of ATP hydrolysis in CF airways have not been identified, due in part to the complex array of ATP metabolizing ectoenzyme activities expressed on airway surfaces [111].

Three major subfamilies of ecto-enzymes hydrolyze extracellular ATP: a) ecto-nucleoside triphospho diphosphohydrolases (ENTPDs), which hydrolyze nucleoside triphosphates and diphosphates, rendering nucleoside monophosphates and inorganic phosphate (Pi) as final hydrolysis products; b) ecto-nucleotide pyrophosphatases/phosphodiesterases (ENPPs), which remove pyrophosphate (PPi) from ATP, producing AMP; and c) tissue nonspecific alkaline phosphatase (NSAP), a glycosylphosphatidylinositol (GPI)-anchored protein that dephosphorylates ATP, ADP, and AMP, generating adenosine [112]. Ecto-nucleotidase activities consistent with ENTPD1, ENTPD3, NSAP, and ENPP3 have been described in airway epithelia [104, 106, 113, 114]. However, their relative roles in the hydrolysis of naturally occurring ASL ATP and their contribution to airway dehydration in CFTR-deficient epithelia are not known. Additional hydrolases relevant to purinergic signaling in the airways include 5’nucleotidase, which hydrolyzes AMP [115], and adenosine deaminase, which converts adenosine to inosine [25].

As an initial approach to identifying the ecto-ATPase subfamilies which control the levels of naturally occurring ASL ATP, we assessed ATP hydrolysis under first-order kinetics, i.e., using trace amounts of [γ³²P]ATP ([ATP] << K_m), which most closely predicts the decay pattern of physiologic concentrations of ATP in ASL. Analysis of [γ³²P]ATP hydrolysis products demonstrated a strict correlation between [γ³²P] ATP decay and the formation of ³²Pi in both CF and non-CF HBE cells. [γ³²P]ATP was fully hydrolyzed after 15 min, and at no point during the incubation was formation of ³²PPi detected. The absence of ³²PPi accumulation was not due to PPi degradation (PPi → 2Pi), as ³²PPi added to cells remained intact after 60 min. Consistent with these data, the non-selective ENPP inhibitor β,γ,-MetATP (100 μM) had no effect on the decay of [γ³²P] ATP on HBE cells [23]. These observations strongly suggest that ENPP activities play little or no role in regulating ASL levels of released ATP. In addition, the NSAP inhibitor levamisol failed to prevent ATP hydrolysis in HBE cells. Collectively, the data suggest that ENTPDs are the major activities metabolizing physiologically occurring levels of ASL ATP [23]. Real-time PCR analysis revealed that ENTPD3 is the predominant ENTPD expressed in HBE cells. ENTPD1 and ENTPD2 were also detected in control and CF HBE cells but with apparently lower abundances than ENTPD3. ENTPD8 was not detected. No differences in ATPase activity or gene expression were observed between control and CF cells [23].

A screening of commercially available ENTPD inhibitors revealed that the nucleotide analogue ARL 67156 (ENTPD1>ENTPD3 [116, 117]), the suramin analogue NF 279 (ENTPD3>ENTPD1=ENTPD2 [118]), and the anthraquinone derivative PSB 06126 (ENTPD3 [116]) delayed but did not prevent the hydrolysis of [γ³²P]ATP by HBE cells. That is, these reagents (100 μM) reduced by 35-50% the hydrolysis of radiotracer ATP at 15 min, but ATP was fully hydrolyzed after 60 min. Interestingly, the polyoxometalate POM-1 (Na₆[H₂W₁₂O₄₀)].21H₂0) also reduced ATP hydrolysis by 50% at 15 min, and ~ 30% ATP was recovered after 1 h [23]. POM-1 exhibits relatively high affinity towards human ENTPD3 (K_i = 0.7 μM), ENTPD1 (K_i =0.8 μM), and ENTPD2 (K_i = 3.3 μM) [119]. Furthermore, POM-5 (K₁₀[Co₄(H₂O)₂(PW₉O₃₄)₂].22H₂O), which was recently developed as the most potent inhibitor of human ENTPD1, ENTPD2, and ENTPD3 (exhibiting sub-nanomolar Ki values; [119]), showed potent and sustained inhibition of ATP hydrolysis by CF HBE cells. In addition, POM-5 abolished ATP hydrolysis in sputum supernatants from CF patients [23]. Importantly, POM-5 added to HBE culture surfaces increased ATP levels in response to prolonged, physiologically relevant phase shear stress stimulation [23]. Therefore, POM-5 was used as tool to assess the extent to which inhibition of ATP hydrolysis might restore ASL volume regulation in CF cells. Using confocal microscopy to measure changes in ASL height as an index of ATP release-mediated fluid secretion, it was found that POM-5 markedly increased ASL volume production in phasic motion-stimulated CF HBE cells. Indeed, ASL height remained elevated for > 2 h in the presence of POM-5 (Fig. 3). These observations provide a proof of concept for the use of ENTPD inhibitors as therapeutic agents to restore ASL volume regulation in CF cells.

Figure 3. Inhibition of ATP hydrolysis improves ASL volume regulation in CF HBE cells.

CF HBE cells were maintained at an air-liquid interface overnight under phasic motion (0.5 dynes/cm2, 14 cycles/min) [62]. Afterwards, PBS or PBS containing 1 mM POM-5 were applied via an Aeroneb Lab Nebulizer (final [POM-5] = 100 μM; volume applied =100 nl) and ASL height was measured by confocal microscopy, as in [62]. A, representative confocal microscopy image taken at 60 min in the absence (PBS) or presence of 100 μM POM-5; bar =10 μm. B, quantification of ASL height in CF HBE cells. The data represent the mean ± SD (4 cultures / condition; *, p < 0.05) and are representative of two independent experiments. Reproduced from van Heusden et al. 2020. Am J Physiol Lung Cell Mol Physiol 318, L356-L365A [23].

The molecular identification of the enzyme(s) responsible for ATP hydrolysis in the airways awaits the development of ENTPD subtype-specific inhibitors and/or the use of molecular approaches to knock down/knock out these genes. The observation that ATP hydrolysis is elevated in lung secretions from CF and CB patients suggests that secreted/shed ATPases enhance ATP degradation in ASL in vivo, contributing to the mucus dehydration, and inflammation that characterizes muco-obstructed airways. Identifying extracellular nucleotide/nucleoside distribution and release/metabolism patterns associated with mucus dehydration will provide novel biomarker criteria to subdivide CF and COPD patients into defined subpopulations most likely to benefit from a given treatment.

Conclusions

Important advances in the understanding of the purinergic signaling mechanisms that control airway homeostasis were achieved over the past decades, thanks to the pioneering ideas of Geoff Burnstock. The purinergic receptors that regulate MCC activities and the major mechanisms of nucleotide release and hydrolysis/interconversion in ASL have been identified. It is now evident that ATP and adenosine are crucial regulators of MCC activities in normal airways and that the P2Y₂R is a candidate therapeutic target to promote CaCC activity and ENaC inhibition in CF lungs, improving the otherwise poor ASL volume production associated with defective CFTR. Recently, it has also become evident that nucleotide metabolism is abnormally accelerated in CF- and CB-diseased airways, suggesting that ATPases secreted/shed by epithelial and/or inflammatory cells contribute to the positive feedback cycle of ATP degradation, airway/mucus dehydration, obstruction, and inflammation that characterizes the CF and CB airways. Reducing the rates of hydrolysis of released ATP is predicted to facilitate mucus hydration. An ancillary conclusion from these studies is that both the nucleotide distribution and ATPase activity in airway secretions could be novel biomarkers of CF and CB disease progression. An interrelated area of active research focuses on the identification of potent ENTPD inhibitors that could be administered to the airways to elevate the levels of released ATP to promote mucus hydration.

Nucleotides are also potent inflammatory molecules, e.g., by acting on purinergic receptors on inflammatory cells, illustrating the need to control the balance and spectrum of extracellular nucleotide/nucleoside levels. The potential role of purinergic agonists in the pathophysiology of inflammatory airway is highlighted by the finding that levels of UDP-glucose, a poorly hydrolysable P2Y₁₄R agonist, are elevated in neutrophil inflamed airways. The P2Y₁₄R is highly expressed in neutrophils and a P2Y₁₄R-selective antagonist reduces the spontaneous neutrophil lung inflammation associated with the CF/CB-like βENaC mouse. The observation that elevated levels of UDP-glucose are associated with lung inflammation in CF supports further investigations to test the potential therapeutic benefits of aerosolizing P2Y₁₄R antagonists onto CF airways to reduce neutrophil infiltration.

Remembrances of Geoff Burnstock

A highlight of all interactions with Geoff in London were the sessions, hosted by Geoff and Nomi, in his beloved London flat. The light, airy, garden-bedecked flat, coupled to spirits and hors d’oeuvres, was the perfect foil for the spirited, often pugnacious, discussions of the role of extracellular nucleotides that spanned evolutionary biology to human disease-oriented therapeutics. Unique sessions led by a unique man.