Micro/nanoscale bubble-mediated mucolytic conditioning of the lung model of cystic fibrosis enhances nanoparticle transport through the mucus

Abstract

Cystic fibrosis (CF) is a monogenic disease with severe pulmonary manifestations. Nanoparticle carriers of drug and gene therapeutics for CF are being studied, but only few can traverse the highly concentrated, disulfide-crosslinked airway mucus in the CF lung. Here, we developed mucolytic-associated polydisperse micro/nanoscale bubble formulations (MNB) that couple reducing chemistry (tris(2-carboxyethyl)phosphine, TCEP) with an ultrasound-responsive carrier to disrupt mucus. We evaluated the penetration of a model nanocarrier through a CF-mimetic mucus hydrogel following MNB-TCEP treatment. Under the same reducing-equivalent dosing, MNB-TCEP markedly enhanced penetration of 200 nm anionic nanoparticles relative to both PBS and free TCEP. In the more restrictive thick-mucus condition, therapeutic ultrasound increased anionic nanoparticle penetration approximately 3.5-fold in the MNB-TCEP group relative to its no-ultrasound counterpart, without statistically significant effects on PBS or free TCEP controls. MNB-TCEP also retained diagnostic ultrasound visibility in gel phantoms and produced airway-associated echogenicity in ex vivo porcine lungs within fluid-accessible, non-aerated imaging windows. Moreover, MNB-TCEP with or without ultrasound did not reduce acute viability of ex vivo tracheal biopsies under the tested regimens. Introduction of an AAV1 reporter suggested that bubble-ultrasound conditioning may increase localized reporter-positive nuclei in airway epithelium. Together, these data identify mucolytic-associated MNB formulations as a theranostic strategy to enhance nanocarrier access through CF-mimetic mucus, with exploratory tissue-level data supporting further evaluation for gene-delivery applications.

1. Introduction

While nucleic acid therapeutics and gene delivery vectors hold substantial promise for treating the lethal pulmonary patterns of cystic fibrosis (CF), their clinical translation is constrained by key physical barriers, primarily those imposed by airway mucus [1–3]. In the pathological CF context, mucus is a structurally complex hydrogel that restricts transport of endogenous factors and exogenous therapeutic carriers [4–7]. In CF and related muco-obstructive states, thick mucus increases network density and reduces effective mesh spacing [4,8], while chemically targetable linkages, including disulfide bonds within gel-forming mucins (e.g., MUC5B and MUC5AC), increase mucus stiffness [9,10]. Gene vectors and nanoparticulate carriers are impeded by this dense physicochemical barrier, making the mucus network a key upstream constraint to luminal airway delivery [6,11,12]. Although PEGylation can reduce adhesive interactions, densely PEGylated nanoparticles often remain inadequate against CF-like mucus barriers, where highly concentrated crosslinked mesh can physically entrap carriers [13,14]. These observations motivate strategies to actively remodel the mucus network to allow nanotherapeutic transport.

To dismantle the crosslinked mucin network, thiol-based mucolytics that reduce mucin disulfide linkages such as N-acetylcysteine (NAC) have been explored, as adjuvants that improve transport through CF sputum barriers [12]. However, NAC often requires high concentrations and/or extended airway-surface residence times to achieve meaningful disulfide reduction and mucolysis [9,15]. Related mucolytic approaches, including new thiol-containing compounds, have also demonstrated activity in CF-relevant sputum, mucus-mimetic barriers, and muco-obstructive models [14,16,17]. Among these, phosphine-based reductants such as tris(2-carboxyethyl)phosphine (TCEP) can drive disulfide disruption in mucus-rich systems, reversing pathological mucus dysfunction in allergic airway disease models and enhancing nanoparticle permeation through mucus-mimetic barrier gels [10,14]. Nevertheless, translation of freely soluble mucolytics is limited by dilution within airway surface liquid and limited exposure time, as deposited agents are subject to mucociliary clearance and epithelial absorption, weakening sustained local reducing activity needed to sever extensive mucin disulfide linkages [15]. This motivates the development of delivery strategies that locally concentrate reducing equivalents and prolong effective mucolytic exposure at the mucus interface.

Ultrasound-responsive micro- and nanoscale bubble systems are increasingly recognized as versatile tools for triggered delivery and barrier modulation [18–20], in part because acoustically driven bubble activity can generate localized microscale stresses that transiently perturb biological matrices [21–23]. In infection-associated biofilms, ultrasound-driven bubble activity can disrupt the matrix and enhance antimicrobial or immune-mediated clearance [24,25], in pulmonary-relevant settings [26]. However, biofilms and pathologically crosslinked mucus act as barriers with different physicochemical constraints. Accordingly, pulmonary bubble studies have focused on biofilm disruption [26] and contrast enhancement [27,28], rather than chemical remodeling of the disulfide-cross-linked mucin network. Consequently, bubble-enabled systems engineered as chemical-mechanical disruptors of the mucus transport barrier could achieve therapeutic delivery by coupling localized mucolytic presentation with ultrasound-driven actuation to enhance subsequent nanocarrier penetration.

Here we introduce a bubble-based biomaterials platform that integrates biochemical reducing capacity with ultrasound-responsive mechanical activity. To this end, TCEP with cationic polydisperse micro/nanoscale bubble formulations (MNB-TCEP) is used to concentrate reducing equivalents at the mucus network. As summarized schematically in Fig. 1, mucolytic-associated MNB treatment is intended to chemically weaken disulfide-crosslinked mucus, while ultrasound provides an acoustic/mechanical disruption step. Formulations were normalized on a reducing-equivalent basis using Ellman quantification, and barrier conditioning was evaluated by Transwell-based nanoparticle penetration through CF-mimetic mucus. Theranostic potential was evaluated by (i) testing bubble-dependent ultrasound visibility in gel phantoms and in fluid-accessible, non-aerated airway regions of ex vivo lung, (ii) ex vivo tissue viability in response to MNB-TCEP and ultrasound, and (iii) exploratory AAV1 reporter readout after MNB-TCEP and ultrasound conditioning. This chemical-mechanical bubble platform may provide a new airway-conditioning strategy to transiently remodel pathological mucus and improve access of inhaled or luminally delivered gene vectors and other nanotherapeutics to the diseased CF airway.

Figure 1. Proposed mechanism of mucolytic-associated MNB-mediated mucus disruption and nanoparticle penetration.

MNB-TCEP/NAC formulations are depicted as micro/nanoscale bubble formulations associated with TCEP or NAC, whereas fluorescent nanoparticles are depicted separately as transport probes added to assess mucus-barrier penetration. (1) Dense CF-mimetic disulfide-crosslinked mucus restricts nanoparticle transport. (2) Mucolytic-associated MNB-TCEP/NAC treatment chemically weakens the mucus network by reducing disulfide linkages, permitting limited nanoparticle transport in the absence of ultrasound. (3A) Without ultrasound, chemical mucus weakening alone allows some nanoparticle penetration. (3B) With ultrasound, MNB-associated acoustic/mechanical activity provides an additional disruption step, enabling increased nanoparticle transport through the mucus barrier. Created with BioRender.com.

2. Materials and Methods

2.1. Materials and Equipment

Micro/nanoscale bubble lipids included distearoylphosphatidylcholine (DSPC; Cayman Chemical), 1,2-stearoyl-3-trimethylammonium-propane (DSTAP; Nanosoft Polymers), and PEG-40 stearate (Sigma-Aldrich); mucolytics included NAC (Sigma-Aldrich) and TCEP·HCl (GoldBio). Co-solvents used were propylene glycol (MP Biomedicals) and glycerol (Sigma-Aldrich). Reagents for Ellman reaction were 5,5-dithio-bis-(2-nitrobenzoic acid) (DTNB, Ellman’s reagent; Sigma-Aldrich) and Tris base (EMD Millipore). CF-mimetic mucus hydrogels were formulated from porcine gastric mucin Type III (Sigma-Aldrich) and PEG-4SH (10 kDa; Laysan Bio) with buffer salts NaCl (Sigma-Aldrich), CaCl₂ and NaH₂PO₄ (Flinn Scientific); phosphate-buffered saline (PBS, Corning) was used for dilutions and receiver wells.

Fluorescent polystyrene nanoparticles were carboxylated FluoSpheres (Invitrogen; 20–500 nm) and 200 nm aminated/unmodified polystyrene (Magsphere; 2.5% solids), diluted to 0.02% solids in PBS unless noted. Transwell assays used 24-well inserts (Celltreat): 3.0 μm PET membranes for primary studies and 8.0 μm membranes for 500 nm nanoparticle studies. Lactate dehydrogenase kit (LDH-Glo) was obtained from Promega. Pneumacult ALI medium was obtained from STEMCELL Technologies. Hoechst 33342 and propidium iodide were used for nuclear and dead-cell staining, respectively.

Key instrumentation included a Branson Sonifier 450 (Branson Ultrasonics), Zetasizer Nano (Malvern), Synergy H1 plate reader (BioTek), Olympus IX2-UCB microscope, Chattanooga Intelect 240 therapeutic ultrasound generator, Philips EPIQ CVx ultrasound system with L15–7io transducer, and Nikon Ti Eclipse inverted spinning disk confocal microscope. Image analysis used Fiji/ImageJ (v2.16.0) and GraphPad Prism (v8.4.2).

2.2. Micro/nanoscale bubble fabrication

Micro/nanoscale bubbles were prepared by thin-film hydration and probe sonication under perfluoropropane. This approach was adapted from established lipid-coated bubble preparation methods using lipid-film preparation, perfluorocarbon-containing aqueous media, and probe sonication [29]. DSPC (6 mg), PEG-40 stearate (3 mg), and DSTAP (1.2 mg) were dissolved in chloroform (1 mL) and dried under high vacuum (≥12 h) to form lipid films. Films were hydrated with 2.4 mL ultrapure water (MNB-Cation) or 2.4 mL water containing NAC or TCEP (200 mM; MNB-NAC/MNB-TCEP), adjusted to pH 7.4 prior to hydration. Glycerol (0.3 mL) and propylene glycol (0.3 mL) were added, and suspensions were maintained at 60 °C during sonication. Micro/nanoscale bubbles were generated by sonication (Branson Sonifier 450; 40% duty cycle, output control 4; 5 min) while continuously introducing perfluoropropane (C₃F₈) by injection. For mucolytic-associated formulations (MNB-NAC and MNB-TCEP), post-sonication suspensions were concentrated by centrifugal ultrafiltration (100 kDa; 3000 RCF, 30 min, room temperature), and the retentate was collected and used immediately unless otherwise specified. Micro/nanoscale bubble concentration was determined by hemocytometer counting after 1:10 dilution in water.

Micro/nanoscale bubbles were prepared using a DSPC/PEG-40 stearate/DSTAP lipid composition adapted from established cationic lipid-shelled microbubble formulations [30,31].

2.3. Physicochemical characterization

Micro/nanoscale bubble hydrodynamic diameter and ζ-potential were measured using a Zetasizer Nano (Malvern). Immediately before analysis, concentrated post-filtration bubble formulations were adjusted to a final NaCl concentration of 10 mM and analyzed without further dilution at 25 °C. Hydrodynamic diameter was reported as Z-average (DLS). ζ-potential was measured using a folded capillary zeta cell under the same final 10 mM NaCl and temperature conditions. For each formulation, n = 3 independently prepared batches were analyzed, with each biological replicate value calculated from the mean of three instrument runs per sample.

For cryo-TEM imaging, 3 μL of sample solution was applied to freshly glow-discharged lacey carbon grids with an additional thin continuous carbon film (Ted Pella, Prod #01895-F). Grids were treated for 35 s in a Fischione Nanoclean 1070 plasma cleaner at 70% power in 75% argon/25% oxygen. Samples were vitrified using a Vitrobot Mark IV (Thermo Fisher Scientific) at 4 °C and 100% relative humidity with a 5 s wait time, 4 s blot time, and blot force of 4, followed by plunge-freezing in liquid ethane pre-cooled by liquid nitrogen. Grids were stored in liquid nitrogen and imaged using a Gatan 626 cryo-specimen holder at approximately −170 °C. Cryo-TEM imaging was performed in low-dose mode on a Titan Halo TEM operating at 300 kV with a 16M CETA camera (Thermo Fisher Scientific) by staff at the Imaging Facility, Advanced Science Research Center, CUNY Graduate Center.

2.4. Ellman assay (DTNB) for reducing-equivalent quantification of MNB-TCEP and MNB-NAC

Reducing-equivalent content of MNB-TCEP and MNB-NAC was quantified using Ellman’s reagent (DTNB). Immediately after fabrication and ultrafiltration (Section 2.2), formulations were diluted 1:1000 in ultrapure water. DTNB (1 mM) was freshly prepared in 100 mM Tris-HCl (pH 8.0). Diluted samples (50 μL) were mixed 1:1 with DTNB solution (50 μL) in a 96-well plate and incubated for 10 min at room temperature prior to absorbance measurement at 412 nm (Synergy H1, BioTek). Calibration standards (0–220 μM) were prepared in water using the corresponding mucolytic (TCEP standards for MNB-TCEP; NAC standards for MNB-NAC) and run on the same plate. Samples and standards were assayed in duplicate wells, and concentrations were obtained by linear regression of A412 versus standard concentration. Ellman-derived values were used to set equal reducing-equivalent doses for transport experiments.

2.5. CF-mimetic mucus hydrogel preparation

CF-mimetic mucus hydrogels were prepared from porcine gastric mucin and PEG-4SH using a disulfide-crosslinking formulation adapted from prior mucin–PEG thiol hydrogel methods [32]. Porcine gastric mucin (Type III; Sigma-Aldrich) and 4-arm PEG-4SH (10 kDa; Laysan Bio) were each prepared at 8% (w/v) in physiological buffer (154 mM NaCl, 3 mM CaCl₂, 15 mM NaH₂PO₄), with pH adjusted to 7.4. Equal volumes of mucin and PEG-4SH precursor solutions were combined (1:1, v/v) to yield gels containing 4% (w/v) mucin and 4% (w/v) PEG-4SH.

The final 4% mucin/4% PEG-4SH formulation was selected to reproduce the higher-concentration synthetic mucus barrier configuration previously used to model disease-state mucus, while preserving the equal-volume mucin:PEG-4SH crosslinking approach of the original disulfide-linked mucin hydrogel method [14,32]. This PEG-thiol/mucin hydrogel platform was previously shown to produce disulfide-linked mucin networks with mucus-like viscoelastic properties, mucolytic sensitivity, and network pore sizes within ranges reported for human mucus [32]. The mixture was cast into Transwell inserts (60 μL for standard barrier conditions; 120 μL for thick-barrier conditions). To prevent precursor leakage prior to gelation, insert bottoms were temporarily sealed (e.g., Parafilm) during casting and incubation, and plates were sealed/covered to minimize evaporation. Inserts were incubated for 12 h at 37 °C and 5% CO₂ to allow gelation prior to transport experiments.

2.6. Transwell nanoparticle transport assay

Nanoparticle transport through CF-mimetic mucus hydrogels was quantified using a Transwell-based receiver-well fluorescence readout. After overnight gelation (Section 2.5), mucus-filled inserts were placed into 24-well receiver plates containing 600 μL PBS per well. Mucolytic solutions or micro/nanoscale bubble formulations were applied apically in a fixed 100 μL volume; reported reducing-equivalent concentrations refer to the final apical mucus/treatment mixture at nanoparticle dosing. For 60 μL mucus dose-series studies, treatment stocks were prepared at 16, 40, or 80 mM to yield final concentrations of 10, 25, or 50 mM after mixing with mucus. For 120 μL mucus experiments, a 55 mM treatment stock was used to yield a final 25 mM condition. The 120 μL condition was used to increase barrier thickness/path length while keeping hydrogel composition fixed, allowing comparison of treatment effects without simultaneously altering mucin solids concentration, PEG-4SH concentration, or gel chemistry. Micro/nanoscale bubble concentrations were set by Ellman quantification (Section 2.4).

After 10 min of apical exposure, fluorescent polystyrene nanoparticles were added to the apical chamber (40 μL of 0.02% solids in PBS) without removing the apical treatment solution. For no-ultrasound conditions, plates were incubated statically for 24 h at 37 °C and 5% CO₂. For ultrasound-treated Transwell transport, ultrasound was applied after nanoparticle addition as described in Section 2.7, followed by the same 24 h of static incubation. Inserts were then removed and receiver-well fluorescence was measured on a Synergy H1 microplate reader (Gen5). Fluorescence was acquired at 580/610 nm for carboxylate-modified FluoSpheres (20–500 nm) and 525/570 nm for aminated and unmodified 200 nm polystyrene nanoparticles. Receiver fluorescence was converted to percent transport (percent of the input dose recovered in the receiver compartment) using a same-day input-dose calibration prepared from the same nanoparticle suspension used in the experiment. Calibration wells were prepared by spiking 0–40 μL nanoparticle stock into 600 μL PBS (0–100% of the applied dose).

2.7. Therapeutic ultrasound exposure

Therapeutic ultrasound was applied using a Chattanooga Intelect 240 ultrasound generator equipped with a 1.0 MHz applicator (2 cm²). The Transwell plate was placed directly on the gel-coupled transducer head, and ultrasound was delivered from beneath the plate (through the receiver-well side) using pulsed conditions (20% duty cycle; 1.75 W cm⁻² ISATA, device setting) as four 10 s ON exposures separated by 60 s OFF intervals. For ultrasound-treated Transwell transport groups, mucolytic or micro/nanoscale bubble formulations were added apically and incubated for 10 min before fluorescent nanoparticles were added as described in Section 2.6; ultrasound was then applied where indicated. Sham controls were handled identically without sonication. Ultrasound-treated Transwell groups received a 10 min formulation pre-exposure, followed by nanoparticle addition and ultrasound exposure consisting of four 10 s ON pulses separated by 60 s OFF intervals (40 s ON time), before the 24 h transport incubation.

2.8. Ultrasound imaging and image analysis

Ultrasound imaging was performed on a Philips EPIQ CVx system using an L15–7io linear array transducer in 2D B-mode (MI = 0.2; depth = 3.0 cm). For datasets used for grayscale quantification, imaging parameters were held constant within each experiment, including gain and dynamic range (Dyn R 60), and frame rate was constant within each dataset (e.g., 51 Hz). Images were exported as DICOM files without post-acquisition enhancement and analyzed in Fiji/ImageJ. DICOM frames were converted to 8-bit grayscale and mean intensity was measured in fixed, identically sized regions of interest (ROIs). Airway visibility was quantified using a local contrast metric, ΔI, defined as ΔI = I_lumen − I_background, where I_lumen is the mean grayscale intensity in a lumen ROI and I_background is the mean grayscale intensity in adjacent background tissue at comparable depth.

2.9. Agarose gel mold phantom studies

Agarose gel molds were used to assess micro/nano scale bubble-associated echogenicity and signal persistence under controlled imaging conditions. Agarose gel molds containing luminal channels were fabricated as described in Supplementary Methods S2. Lumens were filled with 1 mL normal saline (NS) or 1 mL micro/nanoscale bubble suspension dosed at 1 × 10⁷ bubbles mL⁻¹ (1 × 10⁷ bubbles per lumen). B-mode images were acquired at 0, 5, 10, 30, and 60 min post-loading using the imaging settings described in Section 2.8. ΔI was computed at each timepoint using fixed lumen and background ROIs at comparable depth.

2.10. Ex vivo porcine lung imaging and micro/nanoscale bubble delivery

Ex vivo porcine lungs were explanted within 30 min post-mortem from animals euthanized during unrelated terminal procedures through the Columbia University tissue-sharing program and imaged the same day. Lungs were mechanically ventilated in a prone position. To create a fluid-filled, non-aerated target region while maintaining ventilation of the remaining lung, an Arndt bronchial blocker was positioned in a lobar bronchus and baseline ultrasound images were acquired using the imaging settings described in Section 2.8. The blocker was then exchanged for a bronchoscope for intrabronchial delivery.

MNB-TCEP was administered through the bronchoscope channel by a single continuous instillation (2 mL; 2 × 10⁷ bubbles total) without flush, and the region was re-imaged immediately using identical settings. For each lung, imaging was performed at three locations along the presumed airway path (Area 1: proximal, Area 2: intermediate, Area 3: distal) by translating the probe along the target lobe. In the additional lung experiment, the collapsed/fluid-accessible target lobe was externally marked at three sequential positions along the anticipated ultrasound probe translation path, designated Areas 1–3. During imaging, the ultrasound probe was translated along the marked axis while tracing a visible airway path as consistently as possible using B-mode airway and branching-airway features as anatomical landmarks. Paired pre- and post-MNB-TCEP images were acquired from corresponding marked regions using identical imaging settings. After MNB-TCEP administration and ultrasound scanning, intrabronchial Omnipaque was administered and radiographic imaging was performed to visualize the contrast-filled airway path and the approximate airway course corresponding to the ultrasound-tracked regions, in n=3 lungs.

2.11. Ex vivo tracheal tissue viability and LDH release

Fresh porcine tracheae were obtained post mortem from swine used in unrelated terminal procedures, through the Columbia University tissue-sharing program. Tracheal procurement, biopsy preparation, and air–liquid interface (ALI) culture were done as in our previous studies [33]. 4 mm mucosal punches were maintained at ALI for 2 weeks prior to LDH studies.

For membrane integrity assessment, biopsies were transferred to Transwell inserts (24-well) and apically exposed to PBS, free mucolytics (TCEP and/or NAC, where indicated), or micro/nanoscale bubble formulations (MNB-Cation, MNB-NAC, MNB-TCEP) in a 100 μL apical volume, with 600 μL culture medium in the basolateral compartment. Where indicated, therapeutic ultrasound was applied as described in Section 2.7 (1.0 MHz; 20% duty cycle; 1.75 W cm⁻² ISATA device setting; four 10 s ON exposures separated by 60 s OFF intervals). After exposure, tissues were returned to ALI culture. Basolateral medium was collected at 10 min post-exposure, or 10 min after completion of ultrasound, and at 24 h. Samples were diluted 1:20 in LDH storage buffer, stored at −20 °C, and analyzed using the LDH-Glo Cytotoxicity Assay (Promega) according to the manufacturer’s protocol, with luminescence measured on a Synergy H1 plate reader (BioTek).

2.12. Live/Dead imaging of tracheal biopsies

Porcine tracheal mucosa biopsies were prepared as described above and cultured at ALI for at least 7 days to allow full epithelial recovery from procurement-related injury [33]. On the day of treatment, individual mucosal biopsies were transferred into Transwell inserts in a 24-well plate and apically exposed to PBS, free TCEP, or MNB-TCEP, with or without ultrasound application (20% duty cycle; 1.75 W cm⁻² ISATA device setting; four 10 s ON exposures separated by 60 s OFF intervals). Following the elapsed ultrasound-treatment window, all groups were immediately rinsed three times in PBS to remove treatment solutions and model transient rather than prolonged tissue exposure. Mucosal biopsies were returned to ALI culture overnight.

The following day, tissues were washed in 10 mM dithiothreitol (DTT) for 5 min followed by three washes in PBS to remove the native mucus layer and enable better tissue visualization. Biopsies were then transferred to a staining solution composed of 1.0 μg/mL Hoechst (all nuclei) and 1.0 μg/mL propidium iodide (dead nuclei) in 1:1 PBS and Pneumacult ALI culture medium (STEMCELL Technologies). Tissues were incubated for 1 h at 37 °C and 5% CO₂ before imaging on a Nikon Ti Eclipse inverted spinning disk confocal microscope at 20×. Total and dead nuclear counts were determined from 20X image fields using the LD3 macro for Fiji/ImageJ [33]. Briefly, paired Hoechst/propidium iodide images were split into total-nuclei and dead-nuclei channels, thresholded to generate masks, processed to separate nuclei, and analyzed for channel overlap. Percent live cells was calculated as 1 - C2/(C1 + C2 - C3), where C1 is the Hoechst-positive total nuclei count, C2 is the propidium iodide-positive dead nuclei count, and C3 is the overlapping nuclei count. This overlap-corrected denominator accounts for dead nuclei detected in the propidium iodide channel but not in the Hoechst channel in thick tissue image fields. Four 20X ROIs were captured from each of two tissue biopsies per condition at north, south, east, and west positions across each 4-mm-diameter tissue; ROI-level measurements were used for exploratory field-level viability comparisons to screen for acute toxicity across sampled tissue regions. The Live/Dead quantification workflow is illustrated in Fig. S5.

2.13. Viral gene-delivery proof-of-concept

On the day of viral delivery, tracheal biopsies were pre-treated with MNB-TCEP, ultrasound stimulation, both, or PBS control in the same manner as for viability studies except without any rinse steps following treatment. Tissues were then transferred to a 96-well Transwell plate with 175 μL of complete Pneumacult ALI medium (STEMCELL Technologies) in the basal chamber. Adeno-associated virus serotype 1 (AAV1) carrying a tdTomato reporter conjugated to a nuclear localization signal (NLS-tdTomato) was delivered by pipetting 9.6 × 10⁹ viral genomes directly onto the surface of each tissue biopsy. Tissues were incubated overnight before nuclear counterstaining in 27 μM Hoechst 33342 (Thermo Scientific) and imaging on a Nikon Ti Eclipse inverted spinning disk confocal microscope approximately 20 h following delivery. Percent positive nuclei was determined in Fiji/ImageJ.

2.14. Ethics statement

Post-mortem porcine tissues (unmanipulated lungs and tracheae) were obtained from animals used in unrelated terminal procedures through the Columbia University tissue-sharing program. No live-animal procedures were performed.

2.15. Statistical analysis

Statistical analyses were performed in GraphPad Prism. Data are presented as mean ± SD. For comparisons among ≥3 groups, one-way ANOVA was used, followed by Tukey’s multiple-comparisons test where applicable. For experiments involving two factors, two-way ANOVA was used with the multiple-comparisons test specified in the corresponding figure legend. Statistical significance was defined as p < 0.05. Significance notation was defined as: ns, p > 0.05; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p < 0.0001.

3. Results

3.1. Fabrication and physicochemical characterization of MNB formulations.

We first established whether MNB formulations could be reproducibly fabricated and whether the resulting formulations retained the physicochemical characteristics needed for downstream transport studies. MNB formulations were generated by lipid film hydration and ultrasonic emulsification (Fig. 1; Fig. 2A). Across formulations (MNB-Cation, MNB-NAC, MNB-TCEP), suspensions appeared visually homogeneous, while brightfield microscopy revealed heterogeneous spherical bubble populations that included micron-scale structures (Fig. 2B,D). Cryo-TEM provided orthogonal morphology characterization and showed rounded bubble-like structures within the micro/nanoscale formulations (Fig. 2E).

Figure 2. Fabrication and physicochemical characterization of MNB formulations.

(A) Schematic of micro/nanoscale bubble fabrication by lipid film formation and vacuum drying, hydration with mucolytic or control solution (MNB-Cation, MNB-NAC, or MNB-TCEP), and ultrasonic emulsification to generate MNB formulations (created with BioRender.com). (B) Representative photographs of freshly prepared MNB-Cation, MNB-NAC, and MNB-TCEP suspensions. (C) Representative Ellman (DTNB) colorimetric assay showing DTNB reactivity in mucolytic-associated formulations; quantitative absorbance-derived DTNB-reactive reducing-equivalent concentrations are reported in Table S1. (D) Representative brightfield microscopy images showing heterogeneous bubble populations. (E) Representative cryo-TEM images showing bubble-like structures within the formulations. (F) Hydrodynamic diameter measured by DLS in concentrated formulations adjusted to final 10 mM NaCl without further dilution. (G) ζ-potential measured under the same final 10 mM NaCl condition. DLS, PDI, peak-distribution, and ζ-potential values are reported in Table S2. Data are mean ± SD from n = 3 independently prepared batches; each batch value represents the mean of three instrument runs.

DTNB-reactive reducing-equivalent content in mucolytic-associated formulations was quantified by absorbance-based Ellman’s assay using formulation-matched calibration curves (Fig. 2C; Table S1). Representative DTNB color-change images showed increased DTNB reactivity in mucolytic-associated formulations relative to PBS and MNB-Cation controls, consistent with retention of reducing functionality after fabrication. Because Ellman readouts reflect total accessible reducing equivalents and cannot distinguish free from bubble-associated mucolytic, calibrated Ellman measurements were used operationally to normalize reducing-equivalent dosing for downstream transport studies. In independent post-filtration micro/nanoscale bubble batches used for dosing, absorbance-calibrated DTNB-reactive reducing equivalents were 87.01 ± 9.24 mM for MNB-TCEP and 99.64 ± 14.38 mM for MNB-NAC (Table S1; mean ± SD; n = 4 batches).

Dynamic light scattering of concentrated formulations adjusted to final 10 mM NaCl showed broad, polydisperse micro/nanoscale distributions (Fig. 2F; Table S2). Z-average diameters ranged from 254.5 ± 60.3 nm for MNB-Cation to 2342.3 ± 1480.6 nm for MNB-TCEP, with high PDI values consistent with formulation heterogeneity. Under the same final 10 mM NaCl measurement condition, ζ-potential values remained positive across formulations and were not significantly different among MNB-Cation, MNB-NAC, and MNB-TCEP (Fig. 2G; Table S2).

3.2. Micro/nanoscale bubble-mediated enhancement of anionic nanoparticle penetration through CF-mimetic mucus.

We next investigated whether micro/nanoscale bubble formulations altered nanoparticle transport through CF-mimetic mucus and whether this effect depended on the presence of associated mucolytic reducing capacity. Throughout, “micro/nanoscale bubble” refers to the ultrasound-responsive, mucus-modulating bubbles that are the focus of this work, while “nanoparticle” refers to polystyrene nanospheres that serve as model nano-sized drug or gene carriers whose transport could be improved following mucus treatment with micro/nanoscale bubbles. Qualitative tube-inversion assays further supported TCEP- and MNB-TCEP-mediated weakening of CF-mimetic mucus structure compared with PBS (Fig. S1; Supplementary Methods S1).

3.2.1. Impact of model anionic polystyrene particle diameter.

Baseline penetration through untreated CF-mimetic mucus decreased with increasing diameter for anionic polystyrene nanoparticles (20, 100, and 200 nm) (Fig. 3A–B). Because the goal of subsequent experiments was to determine whether barrier conditioning with mucolytic-associated MNB formulations could rescue nanoparticle transport under a highly restrictive mucus condition, 200 nm anionic nanoparticles were advanced as the most demanding readout for formulation-dependent enhancement.

Figure 3. Micro/nanoscale bubble-mediated enhancement of anionic nanoparticle penetration through CF-mimetic mucus.

(A) General schematic of the 60 μL or 120 μL CF-mimetic mucus Transwell nanoparticle-transport workflow used for the mucus-barrier experiments in Figs. 3 and 4; ultrasound was applied only for Fig. 4 conditions where indicated. (B) Baseline diameter-dependent penetration screen of anionic nanoparticles (20, 100, 200 nm) through PBS-treated control mucus. Quantified penetration of 200 nm anionic nanoparticles comparing (C) individual formulation components and (D-E) free mucolytics (TCEP or NAC) versus corresponding micro/nanoscale bubble formulations (MNB-NAC, MNB-TCEP), with PBS as control. All data shown in Fig. 3B–F were collected without ultrasound exposure. (F) Representative receiver-well fluorescence images following penetration under the indicated no-ultrasound conditions (identical LUT and intensity settings across images). Data are mean ± SD with n = 6 per group unless noted on the plot. Statistical comparisons in (B-E) were performed by one-way ANOVA with Tukey’s multiple comparisons test.

3.2.2. Impact of micro/nanoscale bubbles and free mucolytics.

To determine whether micro/nanoscale bubbles alone or free mucolytics alone enhanced anionic nanoparticle transport, PBS, free TCEP, and MNB-Cation were compared (Fig. 3C). Using this 200 nm anionic readout (no ultrasound), free TCEP increased penetration modestly relative to PBS, whereas MNB-Cation remained comparable to PBS, indicating that micro/nanoscale bubbles without mucolytic chemistry did not enhance transport. Mean penetration was 3.89 ± 0.98% for TCEP versus 1.96 ± 0.07% for PBS and 1.87 ± 0.11% for MNB-Cation (mean ± SD; n = 6 per group), with TCEP higher than PBS (*p < 0.05).

3.2.3. Impact of micro/nanoscale bubble-formulated mucolytics.

Next, we compared TCEP and NAC mucolytics, formulated with and without micro/nanoscale bubbles. For both reductants, micro/nanoscale bubble formulation markedly increased anionic nanoparticle penetration, whereas the corresponding free mucolytics produced little or no improvement over PBS (Fig. 3D–E). Fig. 3A depicts the general Transwell nanoparticle-transport workflow used for the mucus-barrier experiments in Figs. 3 and 4, including the optional ultrasound step used only for Fig. 4 conditions where indicated. All experiments shown in Fig. 3B–F were performed without ultrasound exposure; ultrasound-dependent mucus transport experiments are shown separately in Fig. 4. In the TCEP comparison, free TCEP was not significantly different from PBS, whereas MNB-TCEP increased penetration to 27.08 ± 7.95% and was significantly higher than both PBS and free TCEP. Similarly, MNB-NAC increased penetration to 11.29 ± 3.61% and was higher than both PBS and free NAC (mean ± SD; n = 6 per group). These quantitative trends were visually supported by representative receiver-well fluorescence images, which showed minimal signal in PBS, free mucolytic, and MNB-Cation conditions, but substantially greater fluorescence following MNB-NAC and especially MNB-TCEP treatment (Fig. 3F).

Figure 4. Nanoparticle penetration depends on particle charge, mucolytic formulation/concentration, mucus thickness, and ultrasound.

Penetration of 200 nm cationic (A) and nominally neutral (B) nanoparticles through 60 μL CF-mimetic mucus with apical PBS, free TCEP, or MNB-TCEP. (C) Penetration of 200 nm anionic nanoparticles through 60 μL mucus across corresponding free TCEP and MNB-TCEP dose series (10/25/50 mM; final mucolytic concentration after combining mucus with the apical treatment volume). (D-E) Anionic (200 nm) penetration through 120 μL mucus without ultrasound (D) and with ultrasound (E). Gel phantom ultrasound images pre/post exposure (F) and Δ grayscale intensity quantification (G) for saline (NS) and MNB-TCEP. Data are mean ± SD; n = 6 (A-B); (C) n = 4 (10, 50 mM) and n = 6 (25 mM); n = 4 (D-E); n = 3 (G). Statistics: one-way ANOVA + Tukey (A-B, D); two-way ANOVA + Tukey (C); two-way ANOVA + Sidak (E, G).

PBS baselines differed between Fig. 3C and Fig. 3E because these comparisons were performed on different days using separate CF-mimetic mucus preparations. Consistent with this run-to-run variability, the free TCEP effect was modest and variable, whereas MNB-TCEP produced a substantially larger enhancement within the same experiment.

Additional experiments evaluated nanoparticle size-dependent trends (Fig. S2). Mean penetration increased directionally after mucolytic exposure within each particle-size condition, with MNB-TCEP comparable to or higher than free TCEP within the 20 nm, 100 nm, and 500 nm experiments. Because different membrane pore sizes were used (3 μm for 20/100 nm; 8 μm for 500 nm) to avoid membrane-limited transport, absolute values should not be compared directly across sizes.

3.3. Dependence of nanoparticle penetration on charge, mucus thickness, and ultrasound.

Having established formulation-dependent transport enhancement, we then examined whether this effect depended on nanoparticle charge, mucus barrier thickness, and the application of ultrasound.

3.3.1. Dependence on Charge.

Penetration of 200 nm polystyrene nanoparticles was quantified across cationic, nominally neutral, and anionic nanoparticle classes under defined mucolytic exposure conditions.

In 60 μL CF-mimetic mucus, both free TCEP and MNB-TCEP increased penetration of cationic 200 nm nanoparticles relative to PBS (Fig. 4A). Penetration was negligible after PBS pretreatment but increased to 23.40 ± 9.41% with TCEP and 31.31 ± 14.25% with MNB-TCEP (mean ± SD; n = 6 per group). Both TCEP and MNB-TCEP were significantly higher than PBS, whereas TCEP and MNB-TCEP did not differ from each other (p = 0.3863). A similar pattern was observed for nominally neutral 200 nm nanoparticles (Fig. 4B). PBS responses were variable, whereas both TCEP and MNB-TCEP produced substantial penetration, reaching 77.68 ± 30.67% and 92.85 ± 10.89%, respectively (mean ± SD; n = 6 per group). There was no significant difference between TCEP and MNB-TCEP treatment groups (p = 0.6557).

In contrast, anionic nanoparticle penetration in 60 μL mucus showed a distinct dependence on mucolytic formulation state and concentration (Fig. 4C). Across the free TCEP dose series, penetration remained low and did not increase with concentration, similar to PBS controls. By contrast, MNB-TCEP produced clear concentration-dependent penetration, increasing from 5.90 ± 3.48% at 10 mM to 27.08 ± 7.95% at 25 mM and 36.04 ± 4.77% at 50 mM (mean ± SD; n = 4, 6, and 4, respectively). There were no significant series differences across PBS or free TCEP concentrations, whereas MNB-TCEP-mediated penetration increased significantly from 10 to 25 mM (p < 0.0001) and from 25 to 50 mM (p = 0.0059).

3.3.2. Dependence on Mucus Layer Thickness.

To test a more restrictive barrier, mucus volume was increased to 120 μL (25 mM reducing-equivalent condition) (Fig. 4D). Without ultrasound, penetration of 200 nm anionic nanoparticles remained low when mucus was treated with PBS and free TCEP but increased when treated with MNB-TCEP to 12.10 ± 3.92% (mean ± SD; n = 4 per group). MNB-TCEP was higher than PBS (p = 0.0143) and free TCEP (p = 0.0173), whereas PBS and TCEP did not differ (p = 0.9911).

3.3.3. Dependence on Ultrasound.

In 120 μL mucus, ultrasound effects were evaluated under the same 25 mM condition (Fig. 4E). Ultrasound did not significantly change penetration of 200 nm anionic nanoparticles after treatment with PBS or free TCEP, but markedly increased penetration in the MNB-TCEP condition, from 12.10 ± 3.92% without ultrasound to 42.80 ± 14.11% with ultrasound (mean ± SD; n = 4 per group). There was no significant ultrasound effect for PBS (p = 0.9062) or free TCEP (p = 0.5125), whereas MNB-TCEP increased significantly with ultrasound (p < 0.0001), indicating that ultrasound enhanced penetration selectively in the MNB-TCEP condition under this restrictive barrier configuration.

3.4. Theranostic Properties of Micro/nanoscale bubbles: Burst Responsiveness and Imaging Visibility.

Micro/nanoscale bubbles can function as theranostic agents by providing ultrasound contrast and demonstrating an ultrasound-triggered burst response (destruction) at higher power output. Because our platform is designed to retain both visibility and responsiveness, we first confirmed ultrasound-triggered destruction and the associated loss of echogenic signal in a controlled phantom model by imaging before and after the destructive pulse.

3.4.1. Burst Responsiveness.

Agarose gel lumen models filled with normal saline (NS) or MNB-TCEP were imaged before and after a high mechanical index ultrasound pulse (Fig. 4F–G). Local contrast was quantified as ΔI = I_lumen − I_background. NS showed minimal signal before and after ultrasound, with no change. In contrast, MNB-TCEP exhibited high pre-ultrasound contrast (ΔI = 96.43 ± 2.45) that fell to zero after the destructive pulse, confirming ultrasound-triggered destruction (mean ± SD; n = 3; p < 0.0001).

3.4.2. Ultrasound visualization in gel and ex vivo lung.

We next examined whether MNB-TCEP retained diagnostic ultrasound visibility in an agarose gel mold phantom over time and produced airway-associated echogenicity in an ex vivo porcine lung model using low mechanical index (non-destructive) imaging (Fig. 5A). In gel molds filled with NS or MNB-TCEP (1 × 10⁷ bubbles mL⁻¹), local contrast remained near zero for NS but was high for MNB-TCEP, decreasing over time while remaining elevated for 60 min (Fig. 5B; mean ± SD; n = 3 gels per condition), consistent with persistence as an ultrasound contrast agent. In ex vivo lungs, bronchoscopic delivery of MNB-TCEP produced increased airway-associated echogenic signal relative to corresponding pre-MNB-TCEP images acquired along ultrasound-tracked airway regions (Fig. 5C–E; Fig. S3). ΔI was quantified from three sampled regions per lung and is presented descriptively across three independent lungs without statistical inference. Regional mean ΔI increased from −5.92 ± 5.67 to 59.43 ± 20.96 in Lung 1, from −23.77 ± 6.75 to 42.96 ± 29.69 in Lung 2, and from −6.79 ± 10.46 to 28.94 ± 12.58 in Lung 3 (mean ± SD across three sampled regions per lung). Signal magnitude varied across lungs and airway regions.

Figure 5. Ultrasound visualization of MNB-TCEP in gel phantoms and ex vivo porcine lung airways.

(A) Representative B-mode ultrasound images of agarose gel molds containing normal saline (NS) or MNB-TCEP acquired over 60 min. (B) Longitudinal lumen contrast quantified as Δ grayscale intensity from fixed ROIs in gel molds (mean ± SD, n = 3 independent gels per condition). (C) Descriptive Δ grayscale signal trends from three sampled airway regions in three independent ex vivo lungs before [MNB-TCEP(−)] and after [MNB-TCEP(+)] bronchoscopic administration of MNB-TCEP; no statistical inference was performed. The plotted regional trends show increased grayscale signal after MNB-TCEP administration across the sampled regions in all three lungs. (D) Anatomical registration workflow for the additional third ex vivo lung. (a) Gross image of Lung 3 showing the collapsed/fluid-accessible target lobe with external markings corresponding to Areas 1–3 along the ultrasound probe translation path. (b) X-ray image acquired with the bronchial blocker in place, showing the collapsed target lung section; numbered arrows indicate Areas 1–3, and the zoomed view highlights the target section and airway course used for ultrasound tracking. (c) Post-scan X-ray image after MNB-TCEP administration, ultrasound imaging, and intrabronchial Omnipaque administration, showing the contrast-filled airway path used to corroborate the approximate airway course corresponding to the ultrasound-tracked regions. (E) Representative paired B-mode images from the additional third lung acquired from corresponding marked regions along the ultrasound-tracked airway path before [MNB-TCEP(−)] and after [MNB-TCEP(+)] MNB-TCEP administration. Colored arrows indicate the tracked primary airway and branching-airway landmarks used to support anatomical comparison across conditions. Additional representative ex vivo lung ultrasound images from Lung 1 and Lung 2 are provided in Fig. S3. Scale bars: gross photographs and radiographs, 5 cm; ultrasound images, 0.5 cm.

3.5. Ex vivo airway safety assessment.

To evaluate tissue compatibility under the tested exposure regimen, we next assessed the effects of combined MNB-TCEP and US mucus disruption on tissue integrity and viability.

3.5.1. LDH Release.

To screen for epithelial membrane damage (Fig. 6A), lactate dehydrogenase (LDH) release was quantified from ex vivo porcine tracheal biopsies prepared 10 min and 24 h after exposure to (i) PBS or free TCEP (10–50 mM; Fig. S4A), (ii) micro/nanoscale bubble formulations alone (PBS, MNB-Cation, MNB-NAC, MNB-TCEP; Fig. 6B), (iii) ultrasound-only control conditions spanning the tested intensity/exposure settings (Fig. S4B), and (iv) combined micro/nanoscale bubble and ultrasound vs. ultrasound alone (Fig. 6C, Fig. S4C). Across experiments, LDH release was not significantly different across conditions at either time point.

Figure 6. Ex vivo airway membrane viability and gene delivery after exposure to MNB formulations ± ultrasound.

(A) Porcine tracheal biopsies cultured at air-liquid interface were exposed apically to PBS or MNB formulations ± therapeutic ultrasound, with the addition of AAV1 carrying NLS-tdTomato for gene delivery studies. Basolateral supernatants were collected at 10 min and 24 h for LDH quantification (created with BioRender.com). (B) LDH release (n = 6/group) at 10 min and 24 h. (C) LDH release for PBS or MNB-TCEP + ultrasound (n = 6/group) at 10 min and 24 h. (D) Percent live cells on fluorescence imaging at 24 h post exposure to PBS, TCEP, or MNB-TCEP ± ultrasound. (E) Representative images of nuclear localization signal (NLS)-tdTomato signal overlapping with Hoechst-stained nuclei in porcine tracheal explants treated with MNB-TCEP, US only, both, or untreated control. (F) Quantification of imaging data associated with (E). Data are shown as mean ± SD with individual values overlaid in (B, C, F). Percent live cells were calculated from Hoechst-positive nuclei (C1), propidium iodide-positive nuclei (C2), and overlapping nuclei (C3) as 1 - C2/(C1 + C2 - C3). The Live/Dead quantification workflow is illustrated in Fig. S5. For Live/Dead analysis in (D), four 20X ROIs were captured from each of two tissue biopsies per condition, yielding n = 8 ROI-level measurements per condition.

3.5.2. Live/Dead Staining.

We also examined the safety of combinations of TCEP, micro/nanoscale bubbles, and ultrasound treatment on ex vivo tissue by direct nuclear/dead staining (Fig. 6D). Viability remained high (>96%) across all treatment groups, with no significant differences.

3.6. Exploratory AAV1 reporter gene-delivery readout after bubble-ultrasound conditioning.

We next performed an exploratory AAV1-NLS-tdTomato reporter experiment to assess whether the treatment regimen was compatible with viral exposure and could potentially modulate epithelial gene-transfer readouts. Localized confocal image-field quantification showed the highest percentage of NLS-tdTomato-positive nuclei in the combined MNB-TCEP + ultrasound group, with lower values in the MNB-TCEP-only and ultrasound-only groups (Fig. 6E–F). We therefore interpret these data as exploratory evidence that bubble-ultrasound conditioning is compatible with AAV exposure and may increase localized reporter-positive nuclei in this ex vivo setting, warranting future powered studies.

4. Discussion

We show that mucolytic-associated micro/nanoscale bubble treatment improves nanoparticle transport through CF-mimetic airway mucus, a key upstream barrier to drug and gene delivery into diseased lungs. Given that progressive pulmonary disease leading to respiratory failure remains the principal cause of morbidity and mortality in CF [34,35], approaches that improve delivery across hyperconcentrated, crosslinked airway mucus address a central translational challenge [9]. Here, the strongest incremental benefit of micro/nanoscale bubble formulation was observed for a stringent ~200 nm anionic nanoparticle stress-test, selected to challenge mucus-barrier remodeling rather than to serve as a direct surrogate for AAV.

MNB-TCEP produced a pronounced transport advantage for 200 nm anionic nanoparticles across a CF-mimetic mucus barrier compared with free TCEP at the same reducing-equivalent dose (Fig. 3E). Under a more restrictive thick-mucus configuration, therapeutic ultrasound further increased penetration only in the presence of MNB-TCEP (Fig. 4E). Thus, we demonstrate chemical-mechanical modulation of CF-mimetic mucus using mucolytic-associated MNB formulations.

The formulation-dependent gain observed with MNB-TCEP suggests that improved transport is not explained by mucolytic concentration alone but also depends on the method of reducing-chemistry presentation. The anionic 200 nm case provides a stringent benchmark for barrier-conditioning strategies in CF-mimetic mucus, where steric and electrostatic constraints are expected to be substantial [6,7,36]. One possible explanation for the formulation-dependent transport response is that micro/nanoscale bubble formulation alters how reducing capacity is presented at mucus-interacting interfaces. For example, bubble-associated reducing equivalents may be retained or concentrated near mucus components as nanoparticles move through the remodeled network. Although ζ-potential measurements showed no statistically significant differences across MNB formulations under the final 10 mM NaCl measurement condition (Fig. 2G), differences in mucolytic presentation, surface composition, and particle–mucus interactions may still contribute to formulation-dependent transport enhancement.

Beyond formulation dependence, particle class also influenced transport after mucolytic treatment. Free TCEP increased penetration for cationic and nominally neutral nanoparticles but produced minimal improvement for anionic nanoparticles (Fig. 4A–C). The enhanced transport of cationic nanoparticles following TCEP treatment is notable because electrostatic considerations alone predict restricted transport. In PBS-treated control mucus, cationic nanoparticle transport was minimal, whereas TCEP-mediated reduction of disulfide crosslinks may loosen the mucus matrix and reduce steric/network obstruction, allowing increased particle transport despite expected adhesive particle–mucus interactions. This particle-class-dependent response suggests that transport gains after mucolytic remodeling may vary with particle properties and mucus composition [36].

Because the commercial polystyrene probes differ in surface functionalization and the experiments used a controlled CF-mimetic mucin hydrogel, future studies using particle libraries with controlled surface chemistry and patient-derived CF sputum will be needed to further define charge- and surface-dependent transport behavior. These mechanistic ideas remain hypotheses for future study.

Having established that MNB-associated mucolytic activity enhanced nanoparticle transport, we next examined whether ultrasound could further augment this effect under a more restrictive mucus barrier. Under the thick-mucus configuration, therapeutic ultrasound increased anionic 200 nm nanoparticle penetration only in the MNB-TCEP condition (Fig. 4E), indicating that the additional transport benefit depended on the mucolytic-associated MNB formulation rather than ultrasound exposure alone. Ultrasound may promote bubble-mediated mechanical activity (e.g., microstreaming and/or collapse) that locally perturbs the mucus network when gas bodies are present, consistent with literature establishing gas bodies as cavitation nuclei that concentrate mechanical stresses under ultrasound [20,37,38]. The microscale mechanisms of mucus disruption were not directly assessed in this study.

The limited effect of cationic bubbles lacking reducing capacity indicates that the bubble formulation alone was insufficient to substantially overcome the mucus barrier under these conditions. In contrast, NAC- and TCEP-associated MNB formulations enhanced nanoparticle penetration, and ultrasound further increased transport in the MNB-TCEP group under the more restrictive mucus condition. These findings are consistent with a coupled chemical-mechanical mechanism in which mucolytic reduction loosens the disulfide-crosslinked mucin network, increasing the susceptibility of the mucus matrix to bubble-associated acoustic/mechanical disruption during ultrasound exposure. Overall, these data support a coupled chemical-mechanical interpretation based on endpoint nanoparticle transport, although the present study does not directly resolve the relative contributions of mucus network loosening and ultrasound-associated mechanical disruption.

In addition to barrier conditioning, MNB-TCEP retained ultrasound detectability in gel phantoms and produced airway-associated echogenicity ex vivo (Fig. 5). Because the present study used the complete unfractionated MNB formulation, the observed functional and imaging effects should be interpreted as effects of a polydisperse preparation containing nanoscale/submicron components as well as larger micron-scale bubble-like structures, rather than a purified nanoscale fraction. The constraints of lung ultrasound also need to be taken into account, as contrast visualization is most feasible in fluid-accessible, non-aerated regions where acoustic coupling is maintained, in contrast to fully aerated parenchyma [39–41]. Within this context, ultrasound-triggered bubble activity is expected to be localized to bubble-containing, mucus-filled regions, limiting effects on aerated lung and instead concentrating disruption to mucus-obstructed areas. Future work using fractionated preparations and orthogonal sizing methods will be needed to define which bubble-size populations contribute most strongly to mucus remodeling and ultrasound responsiveness.

We also sought to explore whether the platform utility could extend to gene-delivery applications by using an AAV1 reporter in a tissue-level test after MNB-TCEP and ultrasound conditioning. AAV1-NLS-tdTomato was selected as a widely used airway gene-transfer vector that provides an interpretable reporter-expression endpoint in ex vivo airway tissue. Although AAV is substantially smaller than the 200 nm anionic polystyrene nanoparticles that have been used in the mucus transport assays, smaller viral vectors are not necessarily exempt from mucus-mediated restriction, as the motion of viral vectors can be hindered in CF sputum due at least in part to adhesive vector-sputum interactions, with steric obstruction also contributing for AAV [11,42]. The AAV experiments thus enable evaluation of the airway tissue. The 200 nm anionic polystyrene nanoparticles should still be interpreted as restrictive model probes for mucus-barrier remodeling rather than direct surrogates for AAV, which has distinct size, capsid composition, and surface biology [43,44]. These results suggest that the platform may extend to gene-transfer applications, but they remain exploratory and require future studies to determine the magnitude, reproducibility, and mechanism of any gene-transfer enhancement. Future studies should evaluate AAV, lipid nanoparticles, and other non-viral carriers with clinically relevant size and surface-chemistry in mucus-containing airway models using prospectively powered designs.

Notably, none of the treatment conditions detectably reduced acute viability of ex vivo tracheal biopsies (Fig. 6), supporting compatibility of the testing regimen with ex vivo airway tissues. While our primary focus in this work was mucus barrier disruption, the small-scale viral gene delivery study included here indicates that even in the absence of thick mucus, ultrasound-micro/nanoscale bubble systems may also be suitable for disrupting other epithelial barriers to gene delivery. Recent work has emphasized the need for airway pre-conditioning strategies to achieve meaningful delivery rates of gene therapeutics [45–47]. In that context, the increase in localized AAV reporter-positive nuclei observed in this study suggests that bubble-based chemical–mechanical conditioning may improve access across multiple airway barriers relevant to luminal gene transfer. If rigorously confirmed, this would extend the present transport findings beyond nanoparticle penetration. Whether the increase in gene delivery that we observed is due to disruption of the native mucus layer, epithelial cell tight junction disruption, or additional mechanism(s) is an interesting avenue for future investigation.

This study also has several limitations. Nanoparticle transport was evaluated in a mucin hydrogel model and static Transwell configuration, which improved experimental control but simplified airway secretion complexity [32,48]. The disulfide-crosslinked model was selected to represent a chemically reducible mucus-barrier feature targeted by this chemical-mechanical conditioning strategy, since some mucus-mimetic systems are designed primarily around rheological tuning rather than deliberate disulfide crosslinking [49], and to enable standardized head-to-head comparisons that are difficult to conduct in freshly expectorated sputum due to its heterogeneity.

Furthermore, barrier restrictions were modeled by increasing hydrogel thickness/volume at fixed mucin and PEG-4SH concentrations. This approach isolated the effect of barrier thickness/path length but did not capture the full biophysical consequences of increased mucin solids content, which can alter crosslink density, mesh structure, viscoelasticity, and mucolytic responsiveness. Future studies should independently vary hydrogel thickness and mucin/PEG-4SH concentration to distinguish the contributions of mucus accumulation versus mucus hyperconcentration to nanoparticle transport and treatment responsiveness.

In addition to mucus-model limitations, the particle probes used for screening also introduce constraints. Because non-PEGylated carboxylated polystyrene nanoparticles were used as model anionic probes, penetration may reflect surface-mediated particle–mucus interactions as well as particle diameter. Mucus-inert probes, such as PEGylated particles, could help better isolate diameter-dependent transport in future studies. Relatedly, size-dependent screening required different Transwell pore sizes (3 μm versus 8 μm) to avoid membrane-limited transport and preserve a mucus-limited readout, such that absolute transport values were not directly comparable across particle sizes. The Transwell assay also provided an endpoint readout rather than intralayer gradients or particle trajectories [50]. Although this enabled efficient screening across formulations and dose conditions, future studies could integrate high-speed microscopy with multiple particle tracking (MPT) to better resolve transport heterogeneity [51,52].

Although MNB-NAC and MNB-TCEP enhanced nanoparticle penetration even without ultrasound, this effect is most consistent with accessible mucolytic reducing activity weakening the disulfide-crosslinked mucus network. DTNB-based reducing-equivalent quantification was used operationally to normalize dosing across formulations, but Ellman readouts report total accessible reducing capacity and do not distinguish free, surface-associated, bubble-associated, or potentially entrapped mucolytic pools. As a result, the present data do not define the relative contribution of each mucolytic pool or whether ultrasound alters reducing-equivalent accessibility. Future time-resolved studies could determine whether ultrasound changes reducing-equivalent accessibility or redistribution among mucolytic-associated bubble compartments.

Furthermore, we did not evaluate the full range of higher soluble mucolytic concentrations that may be achievable with inhalational delivery. Although MNB-TCEP enhanced nanoparticle transport relative to free TCEP under the same reducing-equivalent dosing, higher concentrations of free NAC or TCEP could potentially produce mucus-barrier disruption through bulk reducing-agent exposure. The intended role of the mucolytic-associated MNB approach is not to replace soluble mucolytic treatment, but to provide an adjunct strategy in which accessible reducing activity can be paired with localized ultrasound-responsive acoustic/mechanical disruption. This may be useful when regional enhancement of nanoparticle transport is desired in selected mucus-obstructed areas. The observed echogenicity of MNB-TCEP in phantom and fluid-accessible ex vivo lung regions suggests the possibility of image-guided or regionally activated use, although this was not tested as an in vivo targeting strategy. Future studies should compare high-concentration free mucolytics with mucolytic-associated MNB formulations to define their relative effects on mucus-barrier disruption and nanoparticle transport.

Because the therapeutic strategy we studied depends on ultrasound activation, the exposure window was selected to permit bubble-associated ultrasound activation and measurable nanoparticle-transport enhancement without detectable acute tissue damage under the tested conditions. However, this bounded ex vivo window was not designed to reproduce in vivo airway residence time or to establish the optimal exposure regimen. Future studies should evaluate efficacy across ultrasound exposure windows and clearance conditions that account for mucociliary clearance, disease state, mucus burden, formulation retention, and delivery route.

Finally, the safety of therapeutic treatment was assessed using LDH release and live/dead staining as first-line indicators of acute membrane compromise, but these do not capture inflammatory signaling, barrier function, and other sublethal endpoints [53,54]. Within this boundary, LDH release was not increased by MNB-TCEP exposure, and ultrasound did not significantly elevate LDH relative to no-ultrasound conditions (Fig. 6); live/dead imaging likewise showed high viability. Future work should assess barrier integrity and mucociliary clearance, both ex vivo and in vivo. While our pilot gene delivery study was limited by small sample size, the observed localized increase in AAV reporter-positive nuclei supports further investigation of ultrasound/MNB-mediated enhancement and its underlying mechanism. Sample sizes varied across experimental modules, with Transwell nanoparticle transport assays serving as the primary quantitative evidence and lower-throughput ex vivo lung imaging and viral reporter studies interpreted as feasibility/pilot endpoints requiring future prospective validation.

5. Conclusion

Overall, the collected data position mucolytic-associated MNB formulations as an integrated mucus-conditioning strategy for CF-mimetic pathological mucus, with MNB-TCEP producing a robust transport advantage over free TCEP in a challenging 200 nm anionic nanoparticle test case. Therapeutic ultrasound provided an additional transport advantage only in the MNB-TCEP condition under the most restrictive barrier configuration, consistent with bubble-mediated localization of acoustic mechanical activity. These efficacy gains were achieved without detectable acute viability loss or increased LDH release in ex vivo large-animal airway tissue under the tested exposure regimen, while retaining ultrasound detectability in phantoms and airway-associated echogenicity in ex vivo lungs. Ultrasound-responsive mucolytic-associated MNB formulations therefore represent a promising airway-conditioning platform for improving therapeutic access through pathological mucus in CF and potentially other muco-obstructive lung diseases.

Data availability

All collected data are included within the article and Supplementary Information. The additional imaging, ultrasound, and plate-reader files are available from the corresponding author upon request.