Drug and nanoparticle transport through the densely O-glycosylated mucin polymer mesh of respiratory, GI, and cervicovaginal mucus — from mesh architecture to mucus-penetrating particle design
Gel-forming mucins are the structural backbone of every mucosal barrier. MUC5B and MUC5AC dominate airway mucus, MUC2 dominates the intestinal mucus bilayer, and MUC5B together with cervical mucins governs the cervicovaginal barrier. Each monomer is built around a central set of PTS (Pro/Thr/Ser-rich) tandem-repeat domains that are O-glycosylated at up to 80% of Ser/Thr residues, flanked by cysteine-rich von Willebrand factor-like D-domains that drive end-to-end disulfide polymerization into linear polymers exceeding 40 MDa and several micrometers in contour length.
Mucin biosynthesis begins in goblet and submucosal gland cells: the MUC5B/MUC5AC/MUC2 apomucin is core-glycosylated in the ER, then subjected to massive Golgi-stage O-glycosylation by a cascade of GalNAc-transferases (GALNT family), core 1–4 branching enzymes (C1GALT1, GCNT enzymes), and terminal sialyl-/fucosyltransferases that decorate the PTS domains with 100+ distinct O-glycan structures per monomer — sialylated and sulfated core-2 structures predominate in respiratory mucus, while cervicovaginal mucins carry a distinct, cycle-dependent glycan repertoire.
Polymerization occurs in two steps: N-terminal D1-D2-D'D3 domains form disulfide-linked trimers in the ER, and C-terminal cystine-knot domains mediate end-to-end dimerization in the trans-Golgi and secretory granules, where mucins are packaged at extraordinarily high concentration (>200 mg/mL) in a condensed, Ca²⁺-crosslinked, low-pH state. Granule exocytosis triggers explosive hydration: Ca²⁺ chelation and pH neutralization cause the mucin to expand up to 1000-fold in volume within milliseconds, unfurling into the low-concentration (1–5% w/v) hydrated gel that lines the mucosal surface.
The resulting network is not a regular lattice but a heterogeneous, physically entangled mesh stabilized by a combination of permanent disulfide crosslinks, transient hydrophobic globular-domain associations, and low-affinity glycan-glycan and glycan-Ca²⁺ bridging interactions. Multiple-particle-tracking studies (Lai, Hanes, and colleagues, PNAS 2007) established that this mesh has a broad, approximately log-normal distribution of pore sizes: human cervicovaginal mucus averages ~340 nm at midcycle (range 50 nm to >1800 nm), fresh porcine gastric mucus averages ~200–500 nm, and induced sputum from healthy airway averages 300–500 nm. This heterogeneity — rather than a single fixed mesh size — is central to why some particles find percolating pathways through the gel while others of similar average size do not.
Critically, mucin glycans are not a passive filter: sialic acid and sulfate groups impart a net negative charge (mucus surface charge density on the order of −20 to −40 mV zeta potential-equivalent), and terminal glycan epitopes (Lewis antigens, mucin-type O-glycans) are direct ligands for pathogen adhesins and lectins, meaning the barrier performs simultaneous size-exclusion and affinity-based capture — the two mechanisms that any drug delivery system must simultaneously defeat.
Bulk rheometry measures the macroscopic viscoelasticity of mucus but cannot report on the local, nanoscale aqueous pathways a drug carrier actually experiences. Multiple particle tracking (MPT) microrheology solves this by filming thousands of individually resolved fluorescent nanoparticles diffusing within a mucus sample and reconstructing each particle's mean squared displacement (MSD) trajectory, converting single-particle motion into a map of local pore architecture and viscosity.
The MPT workflow: fluorescent, surface-defined polystyrene or PLGA nanoparticles (typically 100–1000 nm, monodisperse, ζ-potential characterized) are mixed at low volume fraction into a fresh, undiluted mucus sample mounted between a glass slide and coverslip, and imaged for 10–20 seconds at 15–66.7 frames/second using an epifluorescence or spinning-disk confocal microscope. Particle centroids are localized per frame (typically <10 nm localization precision) and linked into trajectories using nearest-neighbor or Kalman-filter tracking algorithms; each trajectory yields a time-averaged MSD(τ) = <Δx(τ)² + Δy(τ)²> from which an effective diffusion coefficient Deff = MSD(τ)/4τ is extracted (2D projection of 3D motion).
The key quantitative result of two decades of MPT studies (Lai et al. 2007–2010; Cone 2009 review; Suk, Xu, Kim, Hanes 2016) is a sharp size-dependent transition: conventional (uncoated, carboxylate- or amine-modified) polystyrene nanoparticles above roughly 100–200 nm show Deff three to four orders of magnitude below their Stokes-Einstein-predicted diffusion coefficient in water, because they are transiently and repeatedly trapped by adhesive interactions with mucin fibers even when geometrically smaller than the average mesh spacing. Below ~50–100 nm, uncoated particles diffuse faster but are still subject to adhesive trapping; only with adhesion-shielding surface chemistry (Stage 3) does diffusion approach the water-like regime up to particle sizes comparable to the mesh pore diameter.
A critical methodological pitfall is ensemble averaging: because pore size is log-normally distributed, a population of identically sized particles shows a bimodal or highly skewed distribution of individual Deff values — a minority of particles find large pores and diffuse rapidly while the majority are geometrically or adhesively trapped. Reporting only the ensemble-average Deff obscures this and can make a poorly performing formulation look moderately mobile; rigorous MPT analysis instead reports the fraction of the population classified as mobile (using a Deff threshold, e.g. Deff > 0.1 µm²/s at τ=1s) alongside the full trajectory-level distribution, which correlates far better with actual tissue penetration and drug delivery outcomes than the mean alone.
The central design principle for a mucus-penetrating particle (MPP), pioneered by Justin Hanes and Richard Cone at Johns Hopkins, is to coat the nanoparticle surface with a sufficiently dense, low-molecular-weight, near-neutral polyethylene glycol (PEG) brush that sterically shields the core from hydrophobic and electrostatic adhesive contacts with mucin glycans — without the brush itself being long or dense enough to sterically obstruct passage through mesh pores.
The mucoadhesive-to-mucus-penetrating transition is governed by a delicate balance rather than a simple "more PEG is better" rule. At low PEG surface density, exposed patches of the underlying hydrophobic polymer core (polystyrene, PLGA, PLA) or charged functional groups remain accessible and mediate hydrophobic and electrostatic adhesion to mucin glycan domains and to the protein backbone itself; these particles are captured within seconds to minutes of contact and cleared with the mucus layer on its normal turnover timescale. As PEG grafting density increases past a percolation threshold — empirically established at roughly 10 PEG(2–5 kDa) chains per 100 nm² of particle surface, achievable via carbodiimide conjugation chemistry or PEGylated block-copolymer (e.g., PLGA-PEG diblock) self-assembly — the brush enters the "mushroom-to-brush" transition and forms a continuous, near-neutral, highly hydrated steric shield that reduces the particle's effective adhesion energy below the thermal energy scale (kT) required for stable capture.
PEG molecular weight matters independently of density: paradoxically, shorter PEG chains (1–5 kDa) at high density outperform longer PEG (10–20 kDa) at equivalent mass coverage, because long, low-density PEG chains adopt a "mushroom" conformation with residual conformational entropy loss upon confinement in mesh pores, generating polymer-brush steric repulsion against the mesh walls themselves — the same phenomenon that makes long PEG-decorated liposomes and PEGylated antibodies (e.g., certolizumab-type Fab-PEG conjugates) surprisingly poor mucus penetrators despite excellent "stealth" behavior in blood. The optimal MPP formulation therefore uses short, dense PEG brushes: this combination reduces adhesive energy while adding minimal effective hydrodynamic radius, allowing particles up to ~200–500 nm — approaching the mean mesh pore diameter itself — to diffuse through human mucus at 10–50% of their unhindered rate in water, versus <0.1% for uncoated controls.
Muco-inert surface chemistry is not limited to PEG: emerging alternatives include zwitterionic polymers (poly(carboxybetaine), poly(sulfobetaine)), which achieve comparable adhesion-shielding via hydration-shell repulsion rather than steric brush repulsion and may resist protein-corona formation more effectively in vivo, and dense low-density PEG combined with a slightly negative zeta potential (−10 to −20 mV) that exploits charge repulsion from the net-negative sialylated/sulfated mucin surface rather than fighting it.
Even a perfectly engineered MPP must contend with a barrier that is never static: healthy mucus is continuously secreted, transported, and shed, while the diseases that most motivate mucosal drug delivery — cystic fibrosis, COPD, chronic rhinosinusitis, ulcerative colitis, and bacterial vaginosis-associated cervicovaginal inflammation — systematically hyperconcentrate and stiffen the gel, shrinking the mean pore size below the threshold at which even well-PEGylated nanoparticles can percolate.
In healthy conducting airway, coordinated ciliary beating (~10–15 Hz beat frequency) propels the entire mucus gel layer cephalad at 4–10 mm/min, giving any deposited particle a residence-time window of only 10–20 minutes before the mucus carrying it is swallowed or expectorated — meaning a drug delivery system must not just penetrate the mesh but do so faster than this clearance clock, or else even a perfectly non-adhesive particle is cleared along with the mucus that momentarily surrounds it. The intestine presents a structurally distinct problem: MUC2-based colonic mucus forms two discrete layers — an inner layer firmly attached to the epithelium and normally impenetrable to bacteria, and an outer, loosely attached layer that is the substrate for the commensal microbiota — so a colonic drug delivery particle must cross the loose outer layer, avoid microbial and immune capture, and separately penetrate the dense inner layer to reach the epithelium.
Disease fundamentally re-engineers the mesh. In cystic fibrosis, loss of CFTR-mediated bicarbonate and chloride secretion causes airway surface liquid dehydration and a drop in periciliary pH, which prevents normal mucin unpacking after granule exocytosis: MUC5B remains hyperconcentrated and abnormally crosslinked, raising mucus solids content from a healthy ~2% to >4–6% w/v and collapsing the mean pore size from ~340 nm to below 100–150 nm — a regime in which even 100 nm dense-PEG MPPs are substantially hindered, motivating combination approaches (mucolytics such as dornase alfa or N-acetylcysteine, or hypertonic saline, administered ahead of or with the nanoparticle dose to transiently loosen the mesh). Similar hyperconcentration and reduced pore size are documented in COPD sputum, chronic rhinosinusitis nasal mucus, and asthmatic airway mucus (with additional eosinophil-derived crosslinking). In the gut, ulcerative colitis and Crohn's disease reduce the protective inner mucus layer thickness and can eliminate the normally sterile inner layer entirely, paradoxically easing physical nanoparticle access to epithelium in inflamed regions while simultaneously increasing systemic exposure risk — a property exploited by "leaky gut"-targeted oral nanoparticle formulations for inflammatory bowel disease. Cervicovaginal mucus varies over an order of magnitude in pore size across the menstrual cycle, from a highly permeable, low-viscosity, estrogen-dominant periovulatory state to a dense, progesterone-dominant luteal/pregnancy state — a consideration central to on-demand topical microbicide and pre-exposure prophylaxis (PrEP) nanoparticle formulation, since a formulation validated in one cycle phase can fail in another.
A 2018 study comparing 200 nm dense-PEG MPPs in freshly expectorated sputum from healthy volunteers versus cystic fibrosis patients found average Deff dropped roughly 25-fold (from ~1.8 µm²/s to ~0.07 µm²/s) despite identical particle surface chemistry — the barrier, not the particle, had changed. This motivated co-formulation of MPPs with low-dose N-acetylcysteine, which restored Deff to ~0.9 µm²/s by disrupting mucin disulfide crosslinks and transiently enlarging the mesh.
In vitro MPT establishes diffusivity, but the ultimate test of a mucus-penetrating formulation is whether it physically reaches, and is retained at, the epithelial surface in real tissue — assessed ex vivo in freshly excised human or animal mucosal explants, in mucus-secreting organoid and air-liquid-interface (ALI) culture models, and in vivo by fluorescence cross-sectioning, intravital microscopy, and conventional pharmacokinetic/biodistribution sampling.
Ex vivo penetration assays place a defined dose of fluorescently labeled nanoparticle formulation onto a freshly excised, intact mucus-covered tissue explant (human cervical biopsy, porcine intestinal segment, excised murine vaginal tissue, or bronchial explant) mounted mucosal-side-up, incubate for a defined contact time (typically 15–60 minutes at 37°C), then cryosection perpendicular to the surface and image particle fluorescence as a function of depth by confocal microscopy. Penetration depth and the fraction of dose reaching within a defined distance of the epithelial surface (e.g., <10 µm) are the primary endpoints; well-designed MPPs typically show penetration depths of tens to >100 µm within 20–30 minutes in healthy tissue, versus depths under a few micrometers — often not distinguishable from the mucus surface itself — for uncoated, mucoadhesive controls.
Organoid and air-liquid-interface (ALI) culture systems provide a renewable, genetically tractable intermediate model: primary human bronchial or nasal epithelial cells, or patient-derived (including CF genotype-specific) intestinal or airway organoids, are differentiated at an air-liquid interface over 21–28 days to generate a pseudostratified, ciliated, mucus-secreting epithelium that recapitulates native mucus rheology, ciliary transport, and — critically for CFTR-targeted programs — the exact ion transport defect of the donor. These systems allow direct testing of nanoparticle penetration under active mucociliary transport (rather than the static mucus samples used in most ex vivo and MPT work), and allow simultaneous assessment of transepithelial permeability, cytotoxicity, and inflammatory cytokine response.
In vivo validation closes the loop with three complementary readouts: (1) fluorescence or radiolabel-based whole-tissue cross-sectioning after topical, inhaled, or oral dosing to map spatial distribution relative to the epithelium, most influentially demonstrated in mouse vaginal and rat lung studies by the Hanes laboratory showing ~90% mucosal surface coverage by dense-PEG MPPs versus highly non-uniform, aggregated deposition of uncoated controls; (2) intravital or two-photon microscopy for real-time, dynamic visualization of particle transport in living tissue; and (3) conventional pharmacokinetic sampling (plasma concentration-time profiles, AUC, Cmax, tissue Cmax at the target mucosa) to confirm that improved physical penetration translates into the pharmacological endpoint that ultimately matters — local drug concentration at, or absorption across, the epithelial target. Programs advancing on this evidence base span inhaled MPP antibiotics and gene therapy vectors for cystic fibrosis, oral MPP formulations for inflammatory bowel disease and H. pylori eradication, and topical cervicovaginal MPP platforms for on-demand and long-acting HIV pre-exposure prophylaxis.
Ensign et al. (Nature Materials 2012, then Sci. Transl. Med.) showed that dense-PEG coated MPPs applied to the mouse vaginal tract achieved uniform distribution across essentially the entire mucosal surface within 10 minutes, including deep within rugae folds inaccessible to conventional mucoadhesive gels — while uncoated particles of the same 200–500 nm core size aggregated into isolated clumps covering under 10% of the surface. This single result reframed vaginal and rectal microbicide/PrEP formulation strategy across the field, shifting emphasis from maximizing mucoadhesion (long presumed desirable for tissue retention) to maximizing mucus penetration.