For all biological studies, dilutions and formulations were performed using sterile Dulbecco’s Modified Eagle Medium (DMEM) (Corning) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin-streptomycin or sterile molecular grade water (Cytiva). All other reagents were used as provided by the manufacturer and stored following manufacturer’s instructions. Reagents were obtained from Fisher Scientific unless otherwise noted within subsections below.

Poly(ethylene)-glycol diacrylate (PEGDA)-based hydrogel nanoparticles (NPs) were synthesized via a reverse emulsion photopolymerization technique at 50 wt% solids as described previously (Jarai and Fromen, 2022). Briefly, NP pre-polymer solution was composed of ∼89 wt% PEGDA₇₀₀ (MW = 700), 10 wt% charge-establishing functional monomer, 1 wt% photoinitiator diphenyl(2,4,6-trimethylbenzoyl)phosphine oxide (PI, Sigma Aldrich), and .05 or .1 wt% fluorescent dye (Cy5 Maleimide, Fluoroprobe) in methanol. The cationic functional monomer was 2-aminoethyl methacrylate (AEM, Sigma Aldrich) and the anionic was 2-carboxyethyl acrylate (CEA, Sigma Aldrich), creating two PEGDA NP formulations which will be referred to as (+) NP and (−) NPs, respectively. Pre-polymer solution compositions for cationic and anionic particles are shown in Figure 2. A volume of 100 μl of the polar monomer mix was added to 1 ml of a continuous, non-polar phase of silicone oil AP1000 (Sigma Aldrich). The polar phase was vortex mixed with the non-polar phase for 1 min until homogenous then tip sonicated. The NPs were polymerized via UV light (APM LED UV Cube with a wavelength of 365 nm at a distance of ∼28 cm from the light source, ∼5–10 mW/cm²) for 30 s. Following polymerization, 800 μl of hexanes were added to the emulsion to break up the remaining oil in solution. The NPs were centrifuged at 18,200 RCF for 5 min and sequentially washed twice with ethanol.

Stock concentration of NPs was measured via Thermogravimetric Analysis (TGA) using TA Instruments TGA 550. Two technical replicates of the NP stock at 50 μl were added to platinum pans on the TGA. The mass of the sample was determined by ethanol evaporation via a temperature ramp up to 90°C followed by 10 min isothermal incubation.

Dynamic Light Scattering (DLS) was performed using Malvern Zetasizer Nano S instrument. NP samples were prepared by diluting stock solutions to ∼0.1 mg/ml in either ethanol or water. Hydrodynamic diameters (D_h), polydispersity indices (PDIs), and zeta potential at room temperature were averaged from at least two independent replicates, 3 technical replicates each.

Pre and post nebulizer samples for both (+) NPs and (−) NPs were prepared at 1 mg/ml in water. Aliquots of 2 μl were flash frozen using the High-Pressure Freezer Leica EM ICE. The samples were then sputter-coated with platinum for 60 s and imaged under high vacuum using an Apreo VolumeScope Scanning Electron Microscope at 2 kV and a variety of magnifications, specified under figure captions. NP geometric diameter was quantified using the measuring tool on ImageJ and scale bars were autogenerated from SEM images (n ≥ 4).

An optical particle sizer spectrometer 3330 (OPS, TSI Inc.) was utilized to calculate the aerodynamic diameter of aerosol droplets carrying (+) or (−) NPs generated from the Aeroneb^® Lab Nebulizer, standard VMD 4.0–6.0 μm (Kent Scientific). NPs were prepared at 50 μg/ml and then 100 μl samples were added to the nebulizer chamber. NP solutions were nebulized for 1 min into a holding chamber directly above the OPS device, which sampled continuously from the holding chamber at a flow rate of 1 LPM for 5 min until aerosols were no longer being detected to ensure the quantification of the entire nebulized sample. The mass median aerodynamic diameters (MMAD) and the geometric standard deviation (GSD) were calculated for each sample according to Eqs 1, 2.

A deposition adapter was designed with Autodesk^® Fusion 360 and 3D printed on a Carbon M1 3D printing platform (Carbon, Inc.) following manufacturer’s instructions. The adapter was printed using the proprietary material PR 25 White manufactured by Carbon following the dimensions displayed in our results. The top was designed to fit on the Aeroneb^® Lab Nebulizer and the bottom has a smaller opening to fit the 24-well transwell plate. The design was inspired by prior work (Horstmann et al., 2021) with one major modification; there is an internal circular lip to lessen the abrupt pressure drop caused by the internal diameter change to the 24-well plate and to collect excess liquid droplets from condensation. The adapter allows a portion of aerosolized volume to deposit in the transwell plate, and the rest of the volume is collected in the internal lip to keep liquified large droplets from forming on the adapter and subsequently dripping into the plate.

Following initial TGA measurements of NP stocks, subsequent NP concentrations were determined using fluorescence measurements using a plate reader. Samples were read at Ex. 590 nm, Em. 640 nm using the Biotek Cytation 5 Multimode Imager. The gain for each assay was adjusted as indicated in figure captions. Median fluorescence intensity (MFI) was directly compared for each NP type by running 200 μl of 10 times diluted NP stock through Novocyte Flow Cytometer (n = 2) and comparing intensity on the APC channel (Ex. 640 nm/Em. 660 nm, Gain:800), shown in Supplementary Figure S1.

PEGDA NP solutions were diluted to 1 mg/ml in 1 ml of 0.5 M sodium phosphate buffers adjusted to varied pHs. The prepared buffers were NaH₂PO₄ (pH 5), Na₂HPO₄ (pH 10), or a 50/50 mol mixture of the solutions (pH 7). These are referred to as pH 5, 10, and 7 respectively.

NP samples were spun down at 18,200 RCF then resuspended in 1 ml of water. Prepared samples were nebulized using the Aeroneb^® Lab Nebulizer (VMD: 4.0–6.0 μm). Recovery through the nebulizer without the 3D adapter was measured by collecting aerosols and resulting droplets as the device was held over a 24-well plate. The recovery through the 3D adapter was measured by collecting aerosols onto a transwell (Falcon 24-Well Hanging Transwell, PET membrane, 0.4 μm pore) and then resuspending that deposited sample in 100 μl of water. The percentage of NPs recovered post-nebulization was determined via fluorescence reading on BioTek Cytation 5 MultiMode Imager using Eq. 3.

The murine macrophage-like cell line, RAW 264.7 (ATCC), was cultured in complete DMEM media (Corning) with 10% FBS (certified Gibco heat inactivated, United States of America origin) and 1% peninicillin-streptomycin. All experiments were performed with cell lines not exceeding a passage number of 10. Air-Liquid Interface (ALI) studies were performed in Falcon 24-Well Hanging Transwells (PET membrane, 0.4 μm pore), while Liquid-Liquid Interface (LLI) studies were performed in 96-well plates. Cells were seeded with 100 μl volume at a density of 8 × 10⁴ cells/well in the growth area with an additional 300 μl of media added to the basal chamber for ALI cultures. ALI was established via removal of apical chamber media 4 hours-post seeding and incubation at 37°C, 5% CO₂. All cell plates were incubated overnight at 37°C, 5% CO₂ before NP treatment.

ALI and LLI cultures were treated with the following sample groups (n = 3 biological replicates): untreated (UT), 25 or 50 μg/ml (+) NPs, 25 or 50 μg/ml (−) NPs. For ALI samples, NPs were re-suspended in sterile water at appropriate concentration based on NP recovery calculated using Eq. 3. NPs were dosed at the indicated concentrations of NP treatment via inserting the Aeroneb^® Lab Nebulizer into the 3D printed device and holding directly over individual transwells throughout dosing. Volumes of 100 μl were delivered for roughly 30 s, and then the set-up was discharged with 100 μl of sterile water for 30 s in between every sample. The device was wiped down in between deliveries to remove any leftover condensation. For LLI samples, NPs were re-suspended in complete DMEM media at concentrations equivalent to the mass of each NP formulation deposited post-nebulization and then dosed via pipette.

Samples were left to incubate for either 6- or 24-h time points at 37°C, 5% CO₂. After, 24 h, adherent LLI and ALI imaging samples were prepared by fixation in a 4% PFA solution at room temperature for 20 min in the dark. Following fixation, cells were washed once in PBS and then sequentially stained with Fluorescent Dye 488-I Phalloidin (Fisher Scientific) for 30 min in the dark and 4′, 6-diamidino-2′-phenylindole, dihydrochloride (DAPI) (Fisher Scientific) for 5 min in the dark with two PBS washes after removal each stain. Samples were then resuspended in a FACS solution (2% FBS in PBS) and stored at 4°C. Images of samples were collected using the Biotek Cytation 5 MultiMode Imager at 40x.

NP internalization and viability after 6 and 24 h was quantified via flow cytometry using NovoCyte Flow Cytometer. At the appropriate time point, cells were washed twice with PBS, detached using 0.25% Trypsin-EDTA (Corning), spun down for 5 min at 500 RCF, and then washed twice with PBS. Cells were stained with Zombie Yellow™ Live/Dead Stain (BioLegend) for 20 min at room temperature in the dark and then washed once with PBS. Uptake and viability was measured on the APC channel (Ex. 640 nm/Em. 660 nm, Gain:800) and Pacific Orange Channel (Ex. 405 nm/Em. 572 nm), respectively.

Metabolic activity after 24 h was determined via a CellTiter-Glo 2.0 Cell Viability Assay (Promega) following manufacturer’s instructions. Luminescence was measured on Biotek Cytation 5 Multimode Imager and relative luminescence was calculated by dividing sample luminescence by a 50/50 media and CellTiter-Glo control well.

GraphPad Prism 9 (GraphPad Software Inc.) was used to perform statistical analyses. Numerical data are represented as mean ± standard deviation (SD) or geometric standard deviation (GSD) as reported in the figure captions. Appropriate post hoc statistical tests were performed as reported in figure captions. Tukey’s multiple comparisons test as part of one-way or two-way ANOVAs was used unless stated otherwise. Results shown are representative of at least two independent experiments, with NP or biological replicates reported in the figure captions.

Our overall goal was to compare macrophage uptake between traditional liquid delivery at the liquid-liquid interface (LLI) and aerosol delivery at the air-liquid interface (ALI) for both positively and negatively charged PEGDA-based NPs (Figure 1). Two PEGDA hydrogel NP formulations were successfully synthesized with cationic AEM or anionic CEA functional monomers following the compositions shown in Figures 2A, B, referred to as (+) NP and (−) NPs, respectively. During the formulation phase, (+) NPs were sufficiently fluorescent at 0.05 wt% Cy5 Maleimide incorporation, while (−) NPs required an increased addition of Cy5 that was incorporated 0.1 wt%. However, it was still found that the (+) NP showed statistically higher fluorescence intensity compared to the (−) NP stock (Supplementary Figure S1).

Following synthesis, the two PEGDA NP formulations were characterized using DLS and cryo-SEM (Table 1; Figure 3, respectively). Cryo-SEM images of both formulations revealed an abundance of spherical NPs ranging in geometric diameters from ∼200–250 nm (Figures 3A, B), which was confirmed via DLS. DLS measurements in both water and ethanol determined differential stability for the two formulations; (−) NPs showed aggregation in ethanol that was mitigated in water, while (+) NPs exhibited the reverse phenomenon (Supplementary Table S1). Given this preference, DLS measurements were reported in the preferred solvent for each formulation, with (−) NPs measured in water, while (+) NPs were measured in ethanol (Table 1). DLS measurements for the as-synthesized, pre-nebulized (+) NP and (−) NP formulations resulted in hydrodynamic diameters (Dh) between 200 and 250 nm in their preferred solvent with PDI below 0.15, indicating reasonable homogeneity and minimal aggregation of the synthesized NPs. As expected, zeta potential for the (+) NP was found to be +29 ± 1 mV, while the (−) NP was found to be −32 ± 4 mV.

Nebulized samples were produced using the Aeroneb^® Lab Nebulizer (VMD 4–6 µm) and collected in 24-well plates. All samples were nebulized in water and the post-nebulizer sample was resuspended in the preferred solvent (i.e., ethanol for (+) NP and water for (−) NP) for consistent DLS characterization. Shown in Table 1, (−) NPs demonstrated minimal change in size and zeta potential following nebulization from the pre-nebulized samples. In contrast, (+) NPs demonstrated aggregation, resulting in an increased Dh to 350 ± 25 nm and a 2x increase in PDI, along with a decrease in zeta potential after passing through nebulizer mesh. Given the overall decreased stability for (+) NPs in water, nebulizing in an aqueous solution is likely responsible for these changes. Indeed, cryo-SEM images confirm this trend between pre- and post-nebulization (Figure 3). All NPs retained their spherical shape post nebulization, as seen by visual inspection. Additionally, an ImageJ sizing analysis demonstrated non-significant changes post nebulization, where (−) NPs average size pre- and post-nebulizing were 250 ± 90 and 170 ± 40 nm, respectively, and (+) NPs were 150 ± 80 and 230 ± 40 nm, respectively. However, a few cases of particle aggregation were found in the (+) NPs, which were not present in (−) NP samples (Supplementary Figure S2). Importantly from both DLS and cryo-SEM results, both NP formulations were largely unaltered by the nebulization process, demonstrating minimal change to critical physiochemical features of the as-synthesized formulation.

Previous studies (Horstmann et al., 2021) provided the framework to create custom 3D-printed adapters for the Aeroneb^® Lab Nebulizer for consistent and efficient aerosol delivery of cargo to transwell plates. As shown in Figure 4, we developed a new custom design with a well bottom to collect condensation and enhance aerosol delivery specifically to 24-well transwells. As shown in Figures 4A, B, the design of the adapter consisted of a top inlet portion designed to fit snugly on the Aeroneb^® Lab Nebulizer, with an outlet of decreased diameter to fit on top of the 24-well transwell plate. The addition of an internal circular lip served to lessen the abrupt pressure drop caused by the internal diameter change to the 24-well plate and was able to collect excess liquid droplets. The adapter allowed a portion of aerosolized volume to deposit in the transwell plate and the rest of the volume was collected in the internal lip to keep large, coalescing droplets from forming on the adapter and subsequently dripping into the plate. The as-printed part is shown in Figure 4C and was able to able to successfully interface with both the Aeroneb^® Lab Nebulizer and 24-well plate opening.

It is known that cargo recovery through mesh nebulizers is relatively poor, with prior work reporting efficiencies of only ∼10% aerosolized from the Aeroneb^® Lab Nebulizer (Horstmann et al., 2021). Losses can stem from both inefficiencies in the nebulizer and deposition within the nebulizer itself, which can be exacerbated with the introduction of any additional surface area, such as an adapter. Thus, it was important to confirm post-nebulizer NP efficiency of the NP used here with the Aeroneb^® Lab Nebulizer and the adapter interface. Percent NP recovery was determined by finding the ratio of the fluorescence between the nebulizer deposited NPs into a 24-well plate or transwell and the total loaded NP amount to the nebulizer (Eq. 1). It was found that with or without the adapter, the (+) NP recovery through the nebulizer was around 12%. Interestingly, (−) NP recovery decreased considerably from 30% without the adapter to 7% using the 3D-printed adapter device. Thus, the use of the adapter did not significantly reduce the aerosol output of either formulation and the introduction of the lipped-adapter ensured a comparable ∼10% efficiency through the adapter for both (−) and (+) NPs.

Both (+) and (−) NPs were loaded in water at 50 μg/ml in 100 μl inlet volume and aerosolized through the Aeroneb^® Lab Nebulizer. The resulting aerosol size with the loaded NPs was confirmed using an Optical Particle Sizer (OPS) to determine MMAD and GSD (Eqs 2, 3). It was found that without the 3D-printed adapter, MMAD values did not significantly deviate across any samples, including the NP-free control (water), or the (+) and (−) NP groups and that they were within the expected range given the selected mesh size of the Aeroneb^® Lab Nebulizer (Figure 5A). Aerosols delivered through the 3D-printed device exhibited a slight increase in size that was not statistically different from those without the adapter, falling within the expected 4–6 µm range based on the mesh properties. The slight increase in aerosol size with the adapter may stem from increased coalescence within the adapter; however, most importantly, the addition of 50 μg/ml of either NP formulation did not contribute to altering the NP stability within the aerosol droplet. Additionally, aerosol counts without the adapter were not statistically significantly different from those with the adapter (Figure 5B), indicating that there was no notable decrease in aerosol formation observed with the introduction of the lipped-adapter.

Building on the desire to better evaluate PEGDA NP formulations following aerosolization, we sought to evaluate NP uptake following deposition at the ALI. Here, monocultures of macrophage RAW 264.7 cells were cultured on both traditional liquid-liquid interface (LLI) and compared to ALI cultures to determine differences between surface charge dependent NP internalization. Macrophages are a critical cell type present in the lung microenvironment whose role is to engulf foreign bodies and activate innate immunity, making them likely targets for NP internalization (Hussell and Bell, 2014).

Here, LLI cultures were grown using traditional methods on cell culture plastic in 96-well plates. Contrastingly, ALI cultures were seeded in transwells of similar surface area to LLI but had media from the top compartment removed 4 h post seeding to establish the appropriate interface. All samples were incubated overnight at 37°C, 5% CO₂ prior to NP dose at either ALI or LLI. LLI cultures were dosed directly with liquid NP treatment via pipette at low and high NP concentrations, 25 μg/ml and 50 μg/ml, respectively. ALI wells were individually treated with NP aerosols using the Aeroneb^® Lab Nebulizer fitted to the 3D printed adapter above the well plate. NP concentration for ALI was increased according to the percent NP recovered calculated previously, such that equivalent NP masses to the LLI conditions would be delivered. To visualize NP uptake and morphology of cells, fixed samples dosed via traditional methods at LLI and with the nebulization method at ALI were stained to demonstrate actin and nuclei using phalloidin and DAPI, respectively. Morphology between non-NP internalizing cells and NP internalizing cells appeared similar, inferring that there were no obvious cytotoxicity concerns as due to NP uptake (Figure 6). At 24-h post-delivery, it appeared that (+) NPs demonstrated dose-dependent uptake that was higher than that of the (−) NP post-delivery in the LLI condition (Figures 6A–D). In contrast, after the same period the ALI culture demonstrated minimal to no uptake of either particle formulation for both the high and low dose (Figure 6E–H).

To quantify particle uptake overtime, a flow cytometry analysis was used to measure NP internalization. Uptake of NPs was measured at 6- and 24-h post-dosage via fluorescence shifts from flow cytometry using two parameters. Median fluorescence intensity (MFI) was used to indicate overall population shifts towards NP uptake, while the percent of samples with a fluorescence intensity greater than 1% of the appropriate untreated samples, known as %NP⁺, was used to quantify NP uptake. Representative gating is shown in Supplementary Figure S4.

MFI values were compared between treated groups at different dosing conditions (Figure 7A) and fold changes were determined as the ratio of the treated group divided by that of the untreated group at the same delivery interface. After 6 h, (−) NPs showed a ∼1.5-fold increase in MFI for LLI cultures only at 50 μg/ml and a ∼1.6 and ∼1.8-fold increase in MFI for 25 and 50 μg/ml dosed ALI samples, respectively (Figure 7B). The MFI of (+) NPs dosed LLI samples was significant and dose dependent at 6-h post-delivery, where low and high NP concentrations increased by ∼3.3 and ∼7.2-fold, respectively (Figure 7B). After 24 h, all LLI dosed cultures showed significant fold-changes that had increased from 6-h samples (Figure 7D). At this time point, (−) NP-dosed LLI cultures MFI increased for low and high NP concentrations by ∼2.2- and ∼3.1-fold, respectively, while (+) NP-dosed LLI cultures increased by ∼5.2- and ∼16-fold, respectively (Figure 7D). However, all ALI cultures after 24 h showed decreased MFI compared to untreated samples (Figure 7D).

The percent of NP uptake (%NP⁺) generally followed MFI trends at 6 and 24 h for both (+) NP and (−) NP samples (Figures 7C, E). At 6-h post-delivery, LLI populations demonstrated 33% and 53% NP-internalized populations for low and high %NP+ population, respectively (Figure 7C), and at 24 h, the NP-positive populations increased to 48% and 76%, respectively (Figure 7E). Interestingly, while MFI data did not show significant increase for ALI delivery at 6- or 24-h post-delivery, the high concentration dose for (+) NP did show significant increase in the percent %NP+ population at both time points (Figure 7C, E). ALI uptake for the high dose of (+) NP yielded 10% and 7% %NP⁺ populations at 6 and 24 h, respectively. However, while this positive population was significant at both 6- and 24-h post-delivery, the percentage of NP-internalizing did not increase after 24 h (Figures 7C, E). In fact, unlike at the LLI, every treatment group demonstrated lower %NP + at 24 h post-delivery compared to 6 h post-delivery (Figures 7C, E).

To determine if the designed NPs demonstrated any pH sensitivity that may impact their fluorescence signal when internalized by cells, NPs were resuspended in three different 0.5 M sodium phosphate buffer conditions (Supplementary Figure S3) at a concentration of 1.0 mg/ml. It was observed that both NP formulations had the highest fluorescent signal in the basic conditions. However, minimal changes to fluorescence intensity were observed for either NP between pH 7.0 and pH 5.0. Thus, fluorescent intensity of the NPs is expected to be constant between extracellular and internalized environments, ensuring consistent detection efficiency for in vitro uptake studies that could not be attributed to the low ALI results.

Previous studies (Goodman et al., 2004; Fröhlich, 2012) have reported surface charge dependent cytotoxicity effects after NP treatment. Here, post-delivery cellular viability was determined via two methods using a live/dead stain for flow cytometry and luminescence-based metabolic assay of plated cells. It was found that at 6- and 24- hours post-delivery for LLI and ALI conditions, both NP types showed a majority of live cells via flow cytometry that did not statistically deviate from untreated samples (Figure 8A). The same trend was observed at 24 h (Figure 8B), indicating that cells were still alive at both LLI and ALI. Morphometric assessments of cells in Figure 6 further supports the presence of viable, live cells in both LLI and ALI. In contrast, metabolic activity measurements of the LLI and ALI cultures via CellTiter Glow assay indicated differences in metabolic activity between the cultures; as shown in Figure 8C, higher metabolic activity was observed for all of the LLI conditions when compared to the ALI conditions. While no trends were observed between NP dosed samples, this difference in metabolic activity likely contributed to the decreased NP uptake compared to ALI.