Enhancing nanoparticle penetration through airway mucus to improve drug delivery efficacy in the lung

Abstract

Introduction:

Nanoparticles (NPs) administered via an inhaled route offer an array of benefits to pulmonary drug delivery. However, mucus gel layer lining the conducting airways serves as a biological and physiological delivery barrier that limits the performance of inhaled NPs. Specifically, conventional NPs, upon deposition on the airway lumen, are readily trapped by the airway mucus and rapidly cleared from the lung via physiological mucus clearance mechanisms. Thus, these NPs cannot distribute through the lung airways, long-reside in the lung and/or reach the airway epithelium. To address this critical challenge, strategies to enhance particle penetration through the airway mucus have been developed and proof-of-concept has been established using relevant mucus model systems.

Areas covered:

In this review, we first overview the biochemical and biophysical characteristics that render the airway mucus a challenging delivery barrier and how the barrier properties are reinforced in muco-obstructive lung diseases. We then introduce strategies to improve particle penetration through the airway mucus. Specifically, we walk through two classes of approaches developed to overcome the barrier, including modification of physicochemical properties of NPs and modulation of barrier properties of airway mucus.

Expert opinion:

State-of-the-art strategies to overcome the airway mucus barrier have been introduced and experimentally validated. However, data should be interpreted with caution in the comprehensive context of therapeutic delivery from the site of administration to the final target destination to identify approaches that would reliably work in clinic. In addition, safety should be carefully monitored, particularly when it comes to mucus-altering strategies that may exert a negative impact on physiological functions of airway mucus.

1. Introduction

Inhalation or direct administration to the respiratory tract constitutes a straightforward means to deliver therapeutics to lung airways for treating muco-obstructive lung diseases, such as cystic fibrosis (CF), chronic obstructive pulmonary disease (COPD) and asthma, and lung cancers [1, 2]. Inhaled therapeutics are readily deployed to the target tissue (i.e. lung airways), thereby reducing drug doses required to fall into the therapeutic window while minimizing off-target delivery and associated side effects [3, 4]. Nano-formulations (nanoparticles or NPs hereafter) designed for inhaled therapeutic delivery confer numerous additional benefits, as well as augment the aforementioned advantages inherent to inhaled administration [5, 6]. First, NPs allow delivery of poorly water soluble drugs to the lung airways by enhancing the drug concentration that can be administered in an aqueous or a powder form [7–11]. Second, NPs protects encapsulated therapeutic payloads from premature degradation or other modes of loss-of-function by enzymatic or metabolic activities [12, 13]. Third, NPs, due to its nanometric dimension, potentially overcome several extracellular and intracellular delivery barriers encountered during the journey from site of airway deposition to target cells or intracellular milieus [14, 15]. Forth, it is relatively feasible to tune the physicochemical properties to modulate delivery and safety profiles (e.g. preferential delivery to target cells). Finally, NPs can be designed to release payloads in a controlled and/or sustained manner over time, which reduces burdens of frequent dosing and dependence on patient compliance [16, 17].

Despite the unique opportunity provided by inhaled NPs for treating lung airway diseases, success of the approach is highly contingent on the ability of NPs to overcome a series of biological and physiological delivery barriers. Specifically, NPs, upon inhalation, encounter airway surface layer (ASL) that protects lung airways from the invasion of deleterious foreign substances, such as pathogens and pollutants [18, 19]. The ASL is maintained by orchestrated architecture and dynamics of two distinct “gel-on-brush” layers, including the luminal mucus gel layer and the periciliary layer (PCL), covering the airway epithelium [20, 21]. Once beyond these extracellular barriers, NPs are required to breach several cellular and/or intracellular barriers depending on the type of therapeutic payloads or specific cellular and subcellular target destinations [22]. These barriers are often shared by other delivery routes and target tissues and thus are beyond the scope of this review.

2. Key extracellular barriers in lung airways

Inhaled foreign matters, including therapeutic NPs, are readily trapped by the mucus gel layer and rapidly cleared from the lung via the mucociliary clearance (MCC), continuous ciliary beating of the PCL, or cough-mediated expectoration [23, 24]. As a result, conventional NPs deposited on and subsequently trapped in the mucus gel layer, cannot efficiently distribute throughout, as well as long-reside in, the lung airways and/or get an access to the underlying airway epithelium [25, 26]. Underlying the mucus gel layer is the PCL that serves as a second line of permeation barrier that limits NP delivery to the airway epithelium [22].

2.1. Mucus gel layer

Mucus gel layer is primarily composed of water (i.e. < 90% in general), but the residual solid-based makeup and its physicochemical arrangements endow the layer with unique rheological properties to serve essential physiological functions, such as MCC [27–29]. Among the solid content, fiber-like gel-forming mucin glycoprotein (mucin hereafter) constitutes a major macromolecular component, which possesses negatively charged O-glycosylated domains or glycans flanked by hydrophobic protein domains (Figure 1a) [30]. Thus, mucins are prime to interact with inhaled foreign matters via non-specific adhesive binding activities, including but not limited to electrostatic and hydrophobic interactions [31, 32]. These adhesive forces, despite being relatively weak and reversible per se, cooperatively immobilize NPs via multivalent interactions [28, 30, 33]. To this end, NPs possessing positive or hydrophobic surfaces cannot efficiently penetrate human airway mucus [28, 34]. In addition, larger NPs with adhesive surfaces are more likely prone to mucoadhesion compared to smaller NPs with comparable adhesivity due to the greater surface area available for interacting with mucus constituents [35]. Due to the heavily glycosylated nature, diffusion of NPs possessing characteristic affinities to sugar moieties abundant in the mucins are also likely interfered by specific adhesive interactions [36]. Lastly, mucin has intrinsic affinity to N-glycan-rich Fc regions of antibodies and thus particulate matters can potentially be immobilized by antibody-mediated multivalent trapping [37, 38], but this may not be of particular concern for synthetic NPs with negligible immunogenicity. Of note, it has been recently suggested that hydrogen bonding between carboxyl and hydroxyl groups of glycans on antibodies and mucus gel may facilitate antibody-mediated trapping or retardation of immunogenic particulate matters [39]. In addition to these adhesive interactions, particle diffusion in the mucus gel can be negatively affected by physical obstruction. Specifically, mucin fibers are chemically cross-linked or physically entangled to form a dense macromolecular network (Figure 1b), thereby sterically hindering particle diffusion particularly when NP diameters reach or surpass the mesh pore sizes of the mucus gel [28, 40]. We have previously estimated the average pore size of human airway mucus to be ~ 200 nm [35]. Overall, the airway mucus gel layer serves virtually as a “sticky net” to inhaled NPs.

Figure 1. Airway mucus is an adhesive and steric barrier to inhaled nanoparticles.

(a) Biochemical structure of mucin, the primary macromolecular composition of airway mucus. (b) Scanning electron micrograph of airway mucus collected from an individual without lung disease. Scale bar = 500 nm. Reproduced with permission from ref. [30].

Barrier properties of the mucus gel layer are generally more pronounced in disease states. The mucin concentration is often elevated in the airway mucus from patients with CF and COPD due to mucus hypersecretion or dehydration [41, 42]. We have recently demonstrated clear inverse correlations between the mucin concentration and mesh pore size of airway mucus samples expectorated by CF [43] and COPD [44] patients. The findings suggest that increase in the mucin concentration in these disease states tightens the mucus mesh spacing, thereby rendering NPs harder to penetrate the airway mucus. We also note that similar inverse correlations between solid content, inclusive of all solid-based makeup in the airway mucus, and mucus mesh pore size were observed with both diseases [43, 44]. Yuan et al. have previously shown that elevation of oxidative stress emanated from chronic inflammation increases disulfide cross-linking density in airway mucus from CF patients [45]. In this study, the observation was primarily associated with macroscopic changes of airway mucus, but the increase in cross-linking density would most likely to tighten the mucus pore size. To this end, we have confirmed that mucus pore size inversely correlates with the concentration of disulfide cross-links in the airway mucus collected from CF patients [43]. In CF airways, genomic DNA [45, 46] and actin microfilaments [47] released from necrotic neutrophils or bacterial cells may enhance negative charge density of airway mucus and/or tighten the mucus mesh. As a result, NPs are likely to experience greater degrees of adhesive trapping and steric hindrance. In support of this hypothesis, we have reported that mucus pore size is inversely correlated with DNA concentration in the airway mucus from CF [43] but not from COPD [44] patients. Likewise, perturbation of pH and salt concentration often observed in the airway mucus gel from patients with muco-obstructive lung diseases [48] may reinforce the barrier properties via alteration of chemical properties (e.g. protonation state) and/or microscopic rearrangement of mucus microstructure (e.g. swelling state) [49]. Of note, it has been recently demonstrated using a covalently-grafted model mucus gel that hydrophobic domains are more exposed with reducing pH and increasing salt concentration, thereby enhancing adhesive interactions with conventional hydrophobic NPs [50].

2.2. Periciliary layer

The PCL as a delivery barrier has been explored to a much lesser extent compared to the luminal mucus gel layer. The reality is presumably due to the previous hypothesis that the PCL is simply a watery layer [51] and thus believed to be relatively permeable to inhaled NPs. In addition, the PCL is generally considered to be stationary, and thus theoretically, inhaled NPs manage to penetrate the mucus gel layer and reach the PCL are expected to long-reside in the airway and to be readily taken up by the immediately underlying airway epithelium if desired [35, 52]. More recently, Button et al. have demonstrated in a pioneering work using an air-liquid interface culture of primary human airway epithelial cells that the layer is densely packed with cell-tethered mucins, thereby excluding apically administered dextran probes possessing diameters of ~ 40 nm [21]. In healthy state, the PCL is retained since osmotic modulus is greater in the layer compared to the mucus gel layer [21]. However, the osmotic modulus in the gel layer is elevated in muco-obstructive lung diseases due to mucus hypersecretion or dehydration, which in turn leads to PCL collapse and impaired MCC [53–55]. Mucus hyper-accumulation and stasis then trigger a vicious cycle of chronic infection and/or inflammation to solidify the PCL collapse. At this point, mesh spacing of the mucin network within the PCL is likely further tightened. Collectively, the PCL should now be regarded as adhesive and steric barriers similar to the mucus gel layer while the barrier properties may vary along the conducting airways due to various cell types forming the airway epithelium. Of note, the mucus stasis may extend airway retention of inhaled NPs [56]; however, the mucus plugs would be ultimately cleared by cough-mediated expectoration.

3. Strategies to enhance penetration of nanoparticles through airway mucus

As introduced earlier, barrier properties that govern diffusion behaviors of NPs in airway mucus is now well established. We and other have demonstrated using various in vitro particle diffusion assays, including particle tracking analysis, fluorescence recovery after photobleaching and diffusion chamber study, that hydrophobic [35, 44, 57–60] or positively charged [61–65] NPs cannot efficiently penetrate human airway mucus collected from healthy individuals or patients with muco-obstructive lung diseases, such as CF and COPD. Likewise, NPs possessing particle diameters greater than mucus pore sizes [35, 59–61] have been shown inability to permeate human airway mucus. To this end, NPs carefully designed to avoid such properties, while conferring muco-inert surfaces, are likely to percolate airway mucus with enhanced diffusion rates. Importantly, particle integrity and physicochemical properties that enable efficient mucus penetration of NPs should be retained in physiologically relevant lung environment. Alternatively, barrier properties of airway mucus can be transiently and/or reversibly modulated by mucus-altering agents to improve particle diffusion. While approaches to promote mucus penetration of NPs can be interchangeably applicable among different mucosal surfaces, we here primarily focus on strategies developed and validated in human airway mucus ex vivo and/or in animal lung airways in vivo.

3.1. Modification of physicochemical properties of nanoparticles

The most widely explored strategy to endow NPs with muco-inert surfaces has been coating NP surfaces with hydrophilic and neutrally charged polyethylene glycol (PEG) polymers (i.e. PEGylation) via chemical conjugation or physical adsorption [33, 66]. Early studies investigating PEGylation were conducted with commercially available polystyrene (PS)-based NPs due to their defined sizes as determined by electron microscopy. We first demonstrated that these hydrophobic PS-based NPs as large as 200 nm in diameters effectively resisted mucoadhesion and were capable of efficiently penetrating the airways mucus collected from individuals without lung diseases [35] and CF patients [58, 59, 67] when particle surfaces were densely passivated with 3.4 - 5 kDa PEG via chemical conjugation. Of note, PEGylation can increase the hydrodynamic diameters of PS-based NPs up to ~10% [35, 44, 59, 68]. In an independent study, Forier et al. also confirmed that surface coatings with 2 – 5 kDa PEG significantly improved diffusion of PS-based NPs possessing particle diameters of 100 – 200 nm in CF airway mucus [69]. More recently, we showed that diffusion of PS-based NPs possessing particle diameters of ≤ 300 nm in airway mucus collected from COPD patients was significantly enhanced by dense surface coatings with 5 kDa PEG [59]. In contrast, similarly PEGylated NPs with particle diameters of ~ 500 nm were unable to efficiently penetrate human airway mucus regardless of sample donors’ disease states [35, 59], suggesting that the particle size was too large to fit in the mesh pores of airway mucus. Importantly, only NPs capable of efficiently penetrating human airway mucus (i.e. densely PEGylated NPs with particle diameters ≤ 300 nm) were those exhibiting widespread distribution and prolonged retention in the lung airways of healthy mice following inhaled administration [68]. Here, we note that “widespread” distribution refers to efficient lateral dispersion of inhaled NPs from the site of deposition on airway lumen, which is distinguished from proximal-to-distal airway coverage achievable by optimizing aerodynamics. We previously reported that high PEG densities that yielded dense brush, rather than mushroom, conformations provided NPs with muco-inert surface coatings although the study was conducted using human cervicovaginal mucus [70]. Specifically, such conformations were achieved when 100 nm NP surfaces were covered with 5 kDa PEG at ≥ 10 PEG chains per 100 nm², and these particles exhibited diffusion rates orders of magnitude greater than those with no PEG or mushroom PEG conformation [70]. However, it should be noted that PEG densities enabling dense brush conformations vary with particle diameter and geometry, as well as PEG molecular weight (MW).

Based on the aforementioned mechanistic underpinnings with model NPs, we and other have implemented the PEGylation strategy to develop drug delivery NPs capable of efficiently penetrating airway mucus. We engineered mucus-penetrating particles (MPPs) using diblock copolymers of PEG and biodegradable polymers, either poly(lactic-co-glycolic acid) (PLGA) [68] or poly(sebacic acid) [58]. Both MPP formulations exhibited at least two order of magnitude greater diffusion rates in CF airway mucus compared to respective non-PEGylated counterpart NPs [58, 68]. Similarly, PEGylated liposomes have been shown to permeate COPD airway mucus to a greater extent compared to non-PEGylated liposomes [71]. We also formulated PLGA-based MPPs by physical adsorption of poloxamers, specifically Pluronic® F127, to the surfaces of PLGA NPs via the central hydrophobic segment and confirmed for efficient airway mucus penetration [68]. Poloxamers are a family of triblock copolymers composed of a central hydrophobic poly(propylene oxide) (PPO) chain interspersed by two poly(ethylene oxide) (PEO or PEG) polymer segments with varying PPO and PEG MW. For example, Pluronic® F127 is composed of two ~4.2 kDa PEG segments adjoined by a central ~3.6 kDa PPO segment. Surface-adsorbed Pluronic® F127 has been also shown to enhance diffusion of compritol-based solid lipid NPs [72] and excipient-free drug nanocrystals [73] in CF airway mucus. Of note, we found that poloxamer-based muco-inert coating was achieved regardless of PEG MW in the range of 0.7 – 5.8 kDa, but minimal PPO MW of ~2.6 kDa was required to ensure stable association of the coating to the surfaces of drug nanocrystals [73]. Congruent with our observation with PS-based model MPP, PLGA- and drug nanocrystal-based MPPs, irrespective of formulation methods (i.e. chemical conjugation and physical adsorption), exhibited widespread distribution and/or prolonged retention in healthy mouse lung airways in vivo following inhaled administration (Figure 2) [68, 73]. Likewise, we have demonstrated that surface modification of human protein nanocages with 2 – 10 kDa PEG significantly enhanced airway distribution compared to their parent non-PEGylated formulation [74]. Synthetic nucleic acid delivery NPs are routinely produced using cationic carrier materials, including polymers, peptides and lipids, resulting in highly positive NPs that readily adhere to airway mucus [61–65]. We thus have employed PEGylation strategies to engineer MPP formulations with numerous cationic polymers, including polyethyleneimine (PEI) [62, 63] and poly(I-amino ester) (PBAE) [65], poly(amido amine) dendrimer [62], peptides, including poly-L-lysine (PLL) [63] and cell-penetrating peptide (CPP) [64], and validated their abilities to retain physiological stability and to efficiently percolate human airway mucus. Additionally, PEI-, PBAE- and CPP-based MPPs were confirmed for widespread and uniform airway distribution and/or prolonged retention, leading to markedly enhanced reporter transgene expression in the lungs of healthy inbred mice compared to their respective non-PEGylated, mucus-impermeable counterparts [63–65]. Similarly, PEGylated PLGA-PEI composite NPs, confirmed for enhanced penetration through an artificial mucus model in vitro, were shown to delay lung clearance in vivo compared to non-PEGylated control NPs [75]. More recently, we demonstrated in vivo that PBAE-based MPPs carrying plasmids encoding a model antigen, due to its ability to penetrate the airway mucus gel layer, were efficiently taken up by pulmonary dendritic cells and mediated robust and durable antigen-specific immune responses in the lung and other mucosal surfaces (Kim et al., accepted). In another study, we also showed that thymulin-encoding plasmids delivered by the identical NPs normalized pro-inflammatory and pro-fibrotic (i.e. remodeling) responses and mechanical defects manifested in the lungs of a mouse model of allergic asthma, entirely in a therapeutic manner (de Silva et al., accepted).

Figure 2. PLGA-based mucus-penetrating particles, prepared by chemical conjugation or physical adsorption, exhibit more widespread airway distribution and/or prolonged lung retention compared to mucus-impermeable nanoparticles in mouse lungs following inhaled administration.

Bronchial distribution of (a) PEGylated PLGA (PLGA-PEG; red) and non-PEGylated PLGA (PLGA; green) nanoparticles (NPs) or (b) PLGA NPs (red) with different poloxamer coatings, including F127 (PLGA/F127) and F68 (PLGA/F68). The white and yellow arrows indicate aggregation of PLGA NPs at the mucus lumen and deep penetration of PLGA-PEG NPs close to airway epithelium, respectively. F127 (PPO MW = ~3.6 kDa) provides muco-inert coatings whereas F68 (PPO MW = ~1.8 kDa) does not. (c) Tracheal distribution of PLGA-PEG (red) and PLGA (green) NPs. The white arrows indicate aggregation of PLGA NPs at the mucus lumen. The yellow dashed box depicts PLGA-PEG NPs distributed throughout a submucosal gland, which is magnified in (d). The yellow arrow indicates deep penetration of PLGA-PEG NPs into the gland. Cell nuclei are stained with DAPI (blue). Retention of PLGA-PEG and PLGA NPs in (e) bronchoalveolar lavage fluid or (f) entire lung at different time points after the administration. Data represent means ± SEM. Reproduced with permission from ref. [68].

A few alternatives to PEG have been explored to improve particle penetration through airway mucus. Ge et al. have demonstrated that fluorination, due to its unique ability to resist mucoadhesion and retain physiological stability, enhances polypeptide-based siRNA delivery NPs through diluted CF airway mucus (Figure 3) [76]. The authors then showed that siRNA against tumor necrosis factor (TNF)-α delivered via the fluorinated NPs reduced TNF-α expression and pro-inflammatory responses in the lungs of acute lung injury mouse model to a greater extent compared to non-fluorinated formulations incapable of penetrating the airway mucus [76]. Self-emulsifying drug delivery system has been developed for plasmid delivery and shown to efficiently penetrate CF airway mucus [77]. The NPs exhibited mucus-penetrating physicochemical properties, specifically small particle diameters (i.e. < 100 nm) and near-neutral surface charges (< −3 mV) [77]. However, it is unclear which constituents primarily contributed to the muco-inert surface property since the system consists of multiple lipid components and non-ionic surfactants. Recently, Leal et al. used combinatorial peptide-presenting phage libraries and next generation sequencing to identify hydrophilic and near-neutrally charged peptides that can endow NP surfaces with muco-inert coatings [78]. Specifically, phages displaying the selected peptides demonstrated ~600-fold greater permeation through CF airway mucus compared to a positively charged control phage [78]. In addition to these alternatives, N-(2-hydroxylpropyl methacrylamide) [79, 80], poly(2-alkyl-2-oxazoline) [81–83], zwitterion [84], polydopamine [85–87], dextran [88, 89], have been investigated to enhance mucus penetration of various NPs. However, we note that these materials have been employed to design MPPs tackling mucosal surfaces other than lung airways, including gastric, intestinal or urothelial mucosa, and thus their performances in airway mucus are yet to be determined.

Figure 3. Fluorination enhances penetration of polypeptide-based siRNA delivery nanoparticles (NPs) through airway mucus.

(a) Apparent permeability coefficient (P_app) of various fluorinated particle formulations across an air-liquid interface culture of a bronchial epithelial cell line. (b) Representative trajectories of fluorinated (P3F16 and P7F7) and non-fluorinated (PG1) NPs in CF airway mucus. (c) Geometric averaged mean square displacement (〈MSD〉) of NPs as a function of the time scale (τ). (d) Distribution of the logarithmic MSD of an individual NPs at τ = 1 second. (e) Fluorescence resonance energy transfer analysis of NPs after incubation with 5% CF airway mucus for 0 or 4 hours, demonstrating enhanced complexation stability of fluorinated NPs. (f) The fluorescence intensity of the aggregates between mucin (0.3% or 0.5%) and NPs following 4-hour incubation, demonstrating enhanced colloidal stability of fluorinated NPs. (g) Distribution of PG1, P3F16 and P7F7 NPs throughout the lung epithelial tissues after intratracheal administration. Scale bar = 75 μm. Reproduced with permission from ref. [76].

3.2. Modulation of barrier properties of airway mucus

Strategies in this category, modulation of barrier properties, hold inherent benefits over NP modification approaches. Specifically, mucus-altering agents are often used in clinic for treating patients with muco-obstructive lung diseases [90], and thus may be readily applicable with negligible regulatory burden and potentially serve dual purposes of enhancing NP-based therapeutic delivery while promoting patients’ lung function by facilitating removal of mucus plugs from the lung airways [91].

As introduced earlier, mucin is the primary macromolecular building block of airway mucus and thus mucus-altering agents or mucolytics directly acting on mucins have been most widely investigated to improve particle diffusion in airway mucus. In particular, many of available mucolytics are reducing agents that cleave disulfide cross-links of mucin network to increase the porosity of airway mucus [91–93]. We have previously reported that N-acetylcysteine (NAC) at the final concentration of 20 mM increases the average pore size of CF airway mucus from 145 ± 50 nm (range: 50 – 300 nm) to 230 ± 50 nm (range: 50 – 1300 nm) [94]. Remarkably, NAC pretreatment resulted in significant increase in the diffusive fraction of MPPs possessing diameters of ~500 nm which were otherwise immobilized within airway mucus due to steric obstruction imposed by its intrinsically small pores [94]. The finding may pose important implications since large particles generally offer enhanced drug encapsulation and/or longer-lasting drug release kinetics [95]. In a follow-up study, NAC pretreatment was shown to significantly enhance diffusion of PLL-based gene delivery NPs in CF airway mucus (Figure 4) and reporter transgene expression in the lungs of a mouse model of mucus hypersecretion [94]. Of note, these NPs are analogous to those tested in a CF gene therapy clinical trial [96]. Likewise, pretreatment with a lysine salt of NAC, Nacystelyn (NAL), was shown to significantly improve gene transfer efficacy of adenoviral vectors complexed with diethylaminoethyl-dextran in the lungs of healthy mice in vivo [97] and p-ethyl-dimyristoylphosphatidyl choline cholesterol- and PEI-based gene delivery NPs in an ex vivo sheep trachea model [98]. However, direct impact of NAL on their diffusion in airway mucus was not determined in these studies. More recently, Yuan et al. synthesized a thiol-carbohydrate structure, methyl 6-thio-6-deoxy-I-D-galactopyranoside, and demonstrated that the compound provided greater reducing capacity compared to NAC and also the effect was faster in CF airway mucus at the final concentration of 10 mM (Figure 5) [45]. Besides these reducing agents, alginate oligosaccharides that include > 85% guluronic acid content have been shown to increase the porosity of CF airway mucus [99]. In this study, Pritchard et al. attributed the microstructural alteration to interactions of alginate oligosaccharides with both peptide backbone and glycan structure of gel-forming mucins (i.e. MUC5AC) via hydrogen bonding, which let to reduced mucin interlinking network [99]. In agreement with this observation, guluronate oligomers were shown to markedly enhance mobility of PS-based NPs possessing diameters 100 or 200 nm in porcine gastric mucus due to reduced steric hindrance resulted presumably from lowering the cross-linking density within the mucin network [100].

Figure 4. N-acetylcysteine (NAC) significantly enhances airway mucus penetration of poly-L-lysine-based gene delivery nanoparticles (NPs), but recombinant human deoxyribonuclease (rhDNase) does not.

(a) Representative trajectories of NPs in CF airway mucus remained untreated or treated with NAC, rhDNase or in combination. (b) Ensemble-averaged geometric mean effective diffusivity (<D_eff>) and (c) distribution of the logarithmic D_eff of individual NPs in CF airway mucus at a time scale (τ) of 1 second. Data represent 5 independent experiments/samples, with an average of n > 100 NPs tracked per experiment. Reproduced with permission from ref. [94].

Figure 5. Methyl 6-thio-6-deoxy-α-D-galactopyranoside (TDG) is a potent and fast-acting mucolytic.

Chemical structure of (a) N-acetylcysteine (NAC) and (b) TDG. (c) The oxidation-reduction potential of TDG is lower than that of NAC and of its parent sugar lacking a thiol group, methyl α-D-galactopyranoside (MDG), indicating a stronger reducing property. Effects of TDG, NAC and MDG on the elastic properties of CF airway mucus at (d, e) high (61 mM) and (f, g) low (10 mM) concentrations over time, demonstrating potent and fasting-acting mucolytic activity of TDG. Reproduced with permission from ref. [45].

A clinically used mucolytic, recombinant human deoxyribonuclease (rhDNase dornase alfa, Pulmozyme®), has been implicated for its potential to enhance particle penetration through airway mucus due to its ability to degrade DNA. It has been previously shown that an addition of 50 μg/ml rhDNase significantly improves the ability of lipid-based gene delivery NPs to mediate transgene expression in the cells covered with CF airway mucus [101]. However, we found that rhDNase at the final concentration of 7 μg/ml was unable to significantly increase diffusion rates of PS-based NPs [102] and the aforementioned PLL-based gene delivery NPs (Figure 4) [94] in CF airway mucus. Similarly, Sander et al. showed that rhDNase at the final concentration of 6 μg/ml only marginally improved penetration of lipid-based gene delivery NPs through CF airway mucus [103]. Alternative to the pretreatment approach, rhDNase has been directly conjugated to the surfaces of NPs. Specifically, tobramycin-loaded alginate/chitosan NPs functionalized with rhDNase, prepared at the final rhDNase concentration of 50 μg/ml, exhibited significantly greater permeation through CF airway mucus compared to non-functionalized NPs [104]. These findings collectively suggest that the effect of rhDNase on particle diffusion in airway mucus may be concentration dependent. However, it should be noted that these studies were conducted using different experimental settings and techniques, and thus comparison should be made with caution.

Osmotic agents can also enhance particle diffusion in airway mucus by increasing the pore size via osmosis-mediated water efflux from the airway epithelium to the ASL and subsequent mucus gel swelling. Specifically, hypertonic saline and mannitol (Bronchitol™) have been widely used to promote airway hydration, both in luminal mucus layer and PCL, thereby enhancing MCC and cough-mediated expectoration in the lungs of CF patients [105, 106]. However, while hypertonic saline and mannitol have been shown to significantly alter bulk rheological properties of airway mucus [107, 108], their impact at a microscopic level relevant to particle diffusion (i.e. change in mucus pore sizes) is yet to be determined. Nevertheless, pretreatment of lung airways with osmotic agent is likely to enhance mucus penetration of subsequently administered NPs, regardless of microscopic mucus perturbation, due to the reduction in the overall mucus burden.

4. Conclusion

Early preclinical and clinical studies investigating the utility of NPs for inhaled therapeutic delivery have been heavily concentrating on enhancing pulmonary bioavailability of poorly water-soluble drugs, tuning lung deposition pattern and/or targeting specific cell types of interest. For example, disappointing outcomes in clinical trials of CF gene therapy have been in large attributed to inability to deploy therapeutic plasmid payloads to target cells in the lung airways [109, 110]. However, research efforts over the past decade have revealed that airway mucus is a highly challenging hurdle to accomplishing clinically relevant inhaled therapy. In the meantime, numerous strategies to overcome the barrier have been introduced and proof-of-concept has been established using human airway mucus and animals. We are now beginning to encounter preclinical studies demonstrating therapeutic benefits enabled by strategies to enhance airway mucus penetration in disease models, laying a steppingstone towards clinical development to ultimately help patients with unmet need.

5. Expert opinion

In this review, we overviewed barrier properties of airway mucus and introduced approaches developed and employed to improve particle penetration through the airway mucus for inhaled therapy of lung diseases. Accumulated evidence suggests that airway mucus is impermeable to conventional NPs, and studies conducted with various mucus model systems insinuate that the reality can be changed by manipulating properties of NPs and/or the barrier. However, data should be interpreted with caution particularly when a mucus model utilized for evaluating mucus-penetrating strategies deviates from “real” mucus gel layer found in the human lung airways. Reconstituted mucin solutions do not precisely emulate chemical features and/or microstructure of physiological airway mucus [49]. In addition, sample dilution essentially swells the mucus hydrogel, thereby leading to artificially enlarged mesh spacing. We also have previously shown that a long-term (≥ 48 hours) storage at 4 °C or a single freeze-and-thaw cycle significantly alters the pore sizes of human airway mucus [28]. Thus, it would be ideal to further validate strategies established on such conditions using freshly collected human airway mucus to secure clinical relevance. We also note that ex vivo mucus models, regardless of origin, are generally static, whereas mucus gel layer found in lung airways in vivo is under a dynamic condition with a continuous and rapid mucus removal and replenish. Thus, enhanced diffusion rates of NPs observed ex vivo may not necessarily translate to desired in vivo performances, including widespread airway distribution, prolonged lung retention and/or efficient epithelial uptake. For example, Murgia et al. previously demonstrated that PS-based NPs confirmed for efficient diffusion in porcine airway mucus ex vivo after mechanical mixing were found trapped at the air-mucus interface in vivo following inhaled administration [111]. To this end, in vivo validation pursued with an animal model recapitulating the lung environment of a target disease would complement human mucus studies and enhance the reliability and relevance of a developed strategy.

As discussed earlier, near-neutral surface is ideal for particle penetration through airway mucus by minimizing multivalent electrostatic interactions with mucus constituents, including mucin and DNA. However, surface hydrophilicity should be simultaneously implemented since uncharged, yet non-polar particle surfaces are prone to hydrophobic interactions with hydrophobic domains of airway mucus. Negatively charged NPs are often considered non-mucoadhesive due to the charge repulsion imposed by the negatively charged glycans of mucins. Indeed, negatively charged NPs administered within airway mucus ex vivo have exhibited significantly greater diffusion rates compared to positively charged NPs [61]. However, inhaled NPs that possess strongly negative surfaces, encountering repulsive forces, may not readily partition into the mucus gel layer upon the deposition at the air-mucus interface. We have previously shown that highly negatively charged PS-based NPs are clumped up at the mucus lumen following inhaled administration and rapidly cleared from the lung, although the mucoadhesion may have been attributed to the hydrophobic cores of the NPs as well [68]. We also note that non-adhesive surface coatings, while desired for efficient mucus penetration, may inhibit bindings with negatively charged cell surfaces and subsequent uptake (i.e. PEG dilemma) [112], which is of particular concern for gene delivery applications. However, we and other showed that dense PEGylation did not entirely block individual cells to engulf small (< 100 nm) NPs in vitro [65, 74, 113, 114]. Further, densely PEGylated MPPs mediated markedly greater transgene expression in mouse lung airways in vivo compared to conventional gene delivery NPs possessing positive surfaces [63–65]. The findings underscore that the ability of PEGylated MPPs to overcome the airway mucus barrier more than offsets their inferior capacity to interact with cell surfaces by enhancing the probability of NPs approaching cell vicinity. Nevertheless, surface ligands can be conjugated to the surfaces of MPPs if further augmented cellular uptake or specific cell targeting is needed, for which ligand type and surface density should be carefully determined not to compromise the mucus-penetrating property. Alternatively, muco-inert coatings can be incorporated to particle surfaces via environmentally sensitive linkers that can be detached after successful mucus penetration is achieved. However, this approach requires an environmental cue that specifically acts on the linkers at a desired location in a timely manner (e.g. pH-, hypoxia- or reactive oxygen species-responsive linkers). Regarding the strategies to temporarily modulate mucus barrier properties, caution should be taken to obviate any deleterious perturbation of physiological mucus functions (e.g. protection against pathogen, MCC, etc.). In addition, pharmacokinetics and pharmacodynamics of mucus-altering agents should be precisely monitored to determine a time window during which inhaled NPs can fully exploit the reduced barrier properties resulted from the action of these agents.

In summary, necessity and strategies to overcome the airway mucus barrier to enhance inhaled NP-mediated therapy of lung diseases have been well established over the past decade. Given the accumulated knowledge and technology, clinical implementation is expected to take place in near future, and careful validation of such strategies in most relevant and complementary experimental settings would facilitate the process.

Article highlights.

• Airway mucus serves as an adhesive and steric barrier to inhaled nanoparticle-based therapeutic delivery.

• The barrier properties of airway mucus is enhanced and/or altered in disease conditions.

• The airway mucus barrier can be overcome by manipulating nanoparticle properties and/or transiently modulating the airway mucus barrier.

• Strategies to overcome the airway mucus barrier should be validated using pathophysiologically relevant and complementary experimental settings and with safety consideration.

• The airway mucus, while critical and challenging, is not the only hurdle and thus other delivery barriers should be taken into account for clinical development of inhaled nanoparticle-based therapies.