Reevaluating gel-forming mucins' roles in cystic fibrosis lung disease

Abstract

The existence of mucus plugs, containing mucins, bacteria, and neutrophils, blocking the lower airways in the lung of cystic fibrosis (CF) patients has raised the possibility that production of “abnormal” mucins is a critical characteristic of this disease. The molecular nature, if any, of this abnormality is unknown. Recent studies suggest that CF lung disease progression is characterized by an early phase in which airway surface liquid (ASL) increased dehydration is accompanied by altered pH and levels of reduced glutathione (GSH). In a later phase, bacterial infection and neutrophil invasion lead to increased ASL of concentrations myeloperoxidase and hypochlorous acid (HOCl). Independent studies indicate that gel-forming mucins, the key components of airway mucus, form disulfide-linked polymers through a pH-dependent, likely self-catalyzed mechanism. In this article, we present the hypothesis that increased mucus concentration (dehydration) and altered pH, and levels of GSH, myeloperoxidase, and/or HOCl result in the extracellular formation of additional interchain bonds among airway mucins. These novel interactions would create an atypical mucin network with abnormal viscoelastic and adhesive properties.

Introduction

The respiratory tract mucosa is covered by a liquid layer known as airway surface liquid. The ASL encounters noxious agents and particles and, propelled by beating cilia, eliminates them from airway surfaces. Antioxidant compounds in the ASL provide a first line of defense against highly reactive oxygen-derived species [1]. The ASL consists of the periciliary liquid layer (PCL), which is an aqueous solution that surrounds the cilia between the cell surface and the mucus layer or mucus, which is positioned between the PCL and the airway lumen. The PCL is believed to lack mucin aggregates and polymers and to behave as a low-viscosity, water solution. The mucus is a gel-like aqueous solution in which soluble compounds are mixed with mucin polymers and aggregates of mucins and other high-molecular-weight glycoproteins. The soluble and insoluble (gel-phase) components of the mucus can be readily separated by centrifugation. The PCL constituents still are undetermined, though it seems reasonable to postulate that the soluble components of the mucus layer are also distributed in the PCL.

In the human respiratory mucosa, two secreted mucins, named MUC5AC and MUC5B, are the major proteins of the mucus [2], [3], [4]. A third, structurally related mucin (MUC19) has been recently reported to be expressed in submucosal glands of the trachea [5], although further studies are required to assess its quantitative and qualitative contributions to the airway mucus. Because of their large size, high carbohydrate content, extended solution structure, and polymeric nature, gel-forming mucins endow airway mucus with high viscosity and the necessary chemical diversity to interact, entrap, and transport inhaled microorganisms, particles, and chemicals from the lung.

A large, centrally located, and highly glycosylated protein domain, named the mucin domain, comprises most of MUC5AC and MUC5B polypeptide chains [6], [7] (Fig. 1). The mucin domain consists of repeats of a unique sequence rich in threonine and serine residues, whose hydroxyl groups are in O-glycosidic linkage with oligosaccharides. The mucin domains are interrupted by several copies of a small Cys-rich domain, known as the Cys subdomain or CS domain. Other protein domains, usually underglycosylated and rich in Cys residues, are also part of the mucin polypeptides. Among these, the D domains and the CK domains at the N and C termini of the mucin polypeptides, respectively, are involved in formation of mucin polymers [3], [8], [9], [10] (see next section).

In addition to secreted mucins, water, and ions, airway mucus contains other components, including proteins/peptides related to the innate immunity system, proteoglycans, glycoproteins, enzymes, lipids, and likely glycosylated domains of membrane/tethered mucins shed from the cell surface [11], [12], [13]. Due to their large sizes and polyanionic nature, proteoglycans and glycosylated subunits of membrane/tethered mucins likely contribute to the viscoelastic properties of the airway mucus.

In the airways, MUC5AC and MUC5B are predominantly synthesized in and secreted from superficial goblet and submucosal mucus cells, respectively [14]. Although these mucins are constitutively secreted, under acute or chronic airway insult several regulated mechanisms mediate increased synthesis and/or secretion of mucins. First, mucin gene promoters are up-regulated by bacterial exoproducts, air pollutants, or inflammation-derived factors [15]. Second, the discharge of large quantities of mucins from mucus granules in goblet/mucus cells is rapidly accomplished in response to a wide variety of stimuli, including irritants, reactive oxygen species, proteases, inflammatory mediators, and nerve activation. Finally, the number of mucin-producing cells increases with chronic airway insult [16], [17], [18].

In parallel with mucin synthesis and secretion, respiratory mucins are constantly eliminated from the lung by ciliary transport, which propels the mucus layer and whatever is bound to it toward the throat [19]. Mucociliary transport and neutrophil-mediated phagocytosis, together with the synthesis/secretion of lysozyme, lactoferrin, and cationic antibacterial peptides, constitute the main mechanisms of lung defense against bacterial colonization [20]. When the ciliary escalator is unable to eliminate the mucus produced, cough clearance is the major mechanism of mucus removal [20].

Gel-forming mucins are synthesized in the endoplasmic reticulum (ER), where they are N-glycosylated [3], [11] and likely C-mannosylated [21] (Fig. 2). In the ER, mucin monomers dimerize through disulfide bonds between their respective C-terminal CK domains^* [4], [8], [9], [22], [23]. Thereafter, dimeric mucins are transported to the Golgi complex, O-glycosylated, and assembled into disulfide-bonded multimers by interdimeric disulfide bonds among the D domains at their N-terminal regions [4], [10], [24]. This assembly step occurs prior to secretion, in regions of the Golgi complex or beyond, in which the vacuolar H⁺-ATPase maintains an acidic pH [10]. Proteolytical processing may occur also in these late secretory compartments [25], [26]. Since in the Golgi complex multimerization is a gradual process, pools of different oligomeric and polymeric species of mucins are likely formed and subsequently secreted [4], [9]. The smaller disulfide-bonded species would comprise dimers linked by their CK domains, with free N-terminal D domains.

The available evidence from the PSM/pMuc19 assembly, the porcine counterpart of MUC19 [5], [27], suggests that the Cys-Gly-Leu-Cys (CGLC) peptide motif at the N-terminal D1 domain in mucins is critical for pH-dependent mucin multimerization [4], [28]. Because this motif resembles the active site of thiol/disulfide oxido/reductases [29], it has been suggested that the mucin D1 domains catalyze mucin multimerization in the trans-Golgi compartments [4], [28]. As for known Cys oxidases, the mucin D1 domains lack a W residue and a charged amino acid residue N and C terminal of the corresponding D1 CGLC motifs, respectively [29]. Moreover, substitution of the D1 CGLC motif for a CGHC sequence, a motif commonly present in the catalytic center of cysteine oxidases, does not alter disulfide interchain bonding of the N-terminal region of PSM/pMuc19.^* A role for the N-terminal D domains in self-catalyzed polymerization has been best demonstrated for the von Willebrand factor (vWF), a blood factor that also has D-, B-, C,- and CK-like domains, although it lacks mucin domains [30]. Thus, incubation of purified vWF at acidic pH is sufficient for the formation of disulfide-linked multimers through the D domains [31]. Because many aspects of the mechanism of multimerization of vWF and mucins are similar, including the involvement of the CGXC motifs at their respective D1 domains [32], it seems reasonable to propose that gel-forming mucins also form disulfide-linked polymers by a pH-driven, self-catalyzed mechanism. Preliminary studies with recombinant PSM/pMuc19 support this view.*

CF is the most common lethal genetic disease in Caucasian populations with an incidence of around 1 in 2500 live births [33]. It is caused by mutations in a gene known as the cystic fibrosis transmembrane regulator [34] encoding a 1480-amino-acid-residue transmembrane protein that is located on the apical surfaces of epithelial cells. CFTR functions both as a cAMP-dependent Cl⁻ channel and as a regulator of other apical channels and transporters, including amiloride-sensitive Na channels and the Cl/HCO₃ exchanger [35]. In addition, low-molecular-weight compounds such as reduced glutathione (GSH), may be transported through CFTR [35]. The opening and closing of the CFTR Cl⁻ channel is regulated through phosphorylation and dephosphorylation by cellular kinases and phosphatases, respectively, and by the level of ATP in the cell [34].

More than 1000 mutations in the CFTR gene have been identified, with those affecting CFTR translation and/or expression at the cell surface resulting in the most severe disease manifestations [36]. Among the latter, deletion of the codon for F508 is by far the most common CF mutation. Other mutations permit apical expression but the regulation, Cl⁻ conductance, or quantity of the CFTR at the apical surface is altered. In general these mutations are found in patients with milder disease manifestations [36].

The major cause of morbidity and mortality among CF patients is associated with the development of chronic lung disease [33]. Lung disease progression can be divided into two overlapping phases with massive bacterial infection and neutrophil invasion in the transition between them. In the first phase, intrinsic defects caused by direct failure of CFTR are predominant while, in the second phase, and largely due to the massive presence of bacteria and neutrophils, new (acquired) defects appear. These acquired defects produce an increasingly complex pathological condition that to some extent is independent of CFTR dysfunction.

The accumulating evidence suggests that early in the life of the CF patient a lack of functional CFTR alters the coupled transport of ions, GSH, and water across airway epithelia and submucosal glands [37], [38] (Fig. 3). This defect leads to a low water/salt content of the mucus, which becomes more concentrated and, hence, viscous [39]. This event, coupled with the depletion of the PCL, is thought to retard mucociliary transport. It appears that the failure to clear mucus is not sensed by mucus secretory cells, so persistent mucin secretion into a water-depleted environment exacerbates the problem of increased mucus concentration/viscosity and mucus stasis. A second key aspect of mucus stasis reflects adhesion of the concentrated mucus layer to the cell surface due to the absence of the PCL. This event is thought to greatly reduce the efficacy of cough clearance. In addition, the high metabolic activity of the ciliated cells, required for sustaining the excessive salt and water absorptive transport that depletes the CF ASL of water and salt, together with the increasing quantities amounts of accumulated mucus likely produces hypoxia in adherent mucus plaques/plugs near the epithelia cell surface [40], [41]. It is suggested that hypoxia is critical for two events in the acquisition of persistent bacterial infection in CF. First, it may select for bacteria that are able to adapt to hypoxic niches. Second, it contributes to formation of antibiotic-resistant, exopolysaccharides-encased microcolonies known as biofilms.

The development of immobilized mucus plaques, coupled with mucus hypoxia, appears to promote the acquisition of chronic bacterial infection. Although the bacterial species are varied in infancy, typically by the age of 3 to 6 years, Pseudomonas aeruginosa becomes the predominant microorganism. Chronic lung disease with extensive neutrophilic accumulation in airway mucus progressively evolves as the CF lung bordered by mucus retention is persistently infected by P. aeruginosa. Inflammatory mediators and bacterial exoproducts contribute to mucus overproduction by up-regulating mucin gene expression and increasing the number of goblet/mucus cells [16], [18], [42], including stimulation of the expression of goblet cells in the bronchioles where they are usually absent. Further, macromolecules such as actin and DNA are released from lysing neutrophils. These large polyanions likely also contribute to the high viscosity of the CF mucus [43]. The excess of mucus, P. aeruginosa biofilms, and neutrophils lead to obstruction of the airways ducts and, finally, to bronchiectasis.

To date, the existence of a single intrinsic abnormality specific for CF airway mucins has not been demonstrated. First, studies suggesting that altered glycosylation/sulfation of CF airway mucins is correlated with absence of CFTR expression [44], [45], [46], [47] have been countered by studies failing to support such modifications [48], [49], [50], [51]. Indeed, the absence of CFTR expression in airway mucus-producing cells [52] argues against a role of CFTR during mucin biosynthesis and assembly. Therefore, the reported differences in glycosylation/sulfation in normal and CF mucins [53] likely originate from CF lung infection and inflammation [54]. That these alterations have an impact on the progression of the disease has been proposed but not yet convincingly demonstrated [55], [56]. Second, structural studies have not found differences between CF and normal airway mucins [57], [58], [59]. However, these and other studies [60] have revealed that presence of fragmented mucins is a common feature of CF mucin preparations, likely the consequence of increased levels of bacterial- and neutrophil-derived proteases in the CF lung lumen (see Effects of mucin degradation for further discussion) and direct hydrolysis by reactive oxygen species. Not surprisingly, the existence of mucin-active mammalian and bacterial proteases, glycosydases, and sulfatases has been recognized for years [61], [62]. Paradoxically, some of these enzymes, especially proteases, also induce mucin gene transcription and secretion [61], [63], [64]. In any case, whatever the effect of these enzymes on mucin polymers, they do not reduce the high viscosity of the CF mucus [65], which suggests that the supply of mucin polymers counters its degradation. In summary, the available data suggest that airway mucins in CF subjects compared to normal subjects are not synthesized with an intrinsic structural or posttranslational alteration.

The question whether mucins are altered in the CF ASL as lung disease progresses from a state where mucus dehydration and PCL depletion are the primary defects arises (Fig. 3). First, do altered CF ASL properties (e.g., pH and levels of GSH; see below) alter mucin multimerization? Second, does the presence of large numbers of neutrophils and bacteria create conditions that chemically alter mucin polymers in a manner that significantly contributes to the pathogenesis of CF lung disease? If so, why have alterations in either context not been detected? There are several potential explanations. First, the large size and highly glycosylated nature of gel-forming mucins make it very difficult to apply classical analytical techniques to reveal their biochemical and structural properties. Second, the majority of studies on CF mucins have been performed with uncharacterized and/or partially purified samples. Third, in those few cases where exhaustive structural analyses were carried out, respiratory mucins were obtained from the sputum of CF patients using strong denaturants such as guanidine and, often, reducing agents, which result in the disappearance of important structural information.