A Bioinspired in vitro Lung Model to Study Particokinetics of Nano-/Microparticles Under Cyclic Stretch and Air-Liquid Interface Conditions

Abstract

Evolution has endowed the lung with exceptional design providing a large surface area for gas exchange area (ca. 100 m²) in a relatively small tissue volume (ca. 6 L). This is possible due to a complex tissue architecture that has resulted in one of the most challenging organs to be recreated in the lab. The need for realistic and robust in vitro lung models becomes even more evident as causal therapies, especially for chronic respiratory diseases, are lacking. Here, we describe the Cyclic I n VI tro Cell-stretch (CIVIC) "breathing" lung bioreactor for pulmonary epithelial cells at the air-liquid interface (ALI) experiencing cyclic stretch while monitoring stretch-related parameters (amplitude, frequency, and membrane elastic modulus) under real-time conditions. The previously described biomimetic copolymeric BETA membrane (5 μm thick, bioactive, porous, and elastic) was attempted to be improved for even more biomimetic permeability, elasticity (elastic modulus and stretchability), and bioactivity by changing its chemical composition. This biphasic membrane supports both the initial formation of a tight monolayer of pulmonary epithelial cells (A549 and 16HBE14o^-) under submerged conditions and the subsequent cell-stretch experiments at the ALI without preconditioning of the membrane. The newly manufactured versions of the BETA membrane did not improve the characteristics of the previously determined optimum BETA membrane (9.35% PCL and 6.34% gelatin [w/v solvent]). Hence, the optimum BETA membrane was used to investigate quantitatively the role of physiologic cyclic mechanical stretch (10% linear stretch; 0.33 Hz: light exercise conditions) on size-dependent cellular uptake and transepithelial transport of nanoparticles (100 nm) and microparticles (1,000 nm) for alveolar epithelial cells (A549) under ALI conditions. Our results show that physiologic stretch enhances cellular uptake of 100 nm nanoparticles across the epithelial cell barrier, but the barrier becomes permeable for both nano- and micron-sized particles (100 and 1,000 nm). This suggests that currently used static in vitro assays may underestimate cellular uptake and transbarrier transport of nanoparticles in the lung.

Keywords: ALI culture; bioinspired membrane; cyclic stretch; lung cell model; particle study.

Figure 1

The CIVIC system used in this study. (A) Schematic depiction of the CIVIC system under (top) static (unstretched) and (bottom) dynamic (stretched and perfusion) conditions with the pressure-based strain/elasticity monitoring system. A thin, permeable, and stretchable membrane (BETA) placed in the (PDMS-free) main chamber of the bioreactor separates the apical (air) and basal (medium) compartments. Lung cells are grown on the membrane at ALI and perfused with culture medium by circulating the medium in the basal compartment with a perfusion pump to mimic blood flow. Cyclic mechanical stretch is applied to the cells on the membrane by applying cyclic (positive) pressure (P₁) to the apical compartment. The cell/membrane stretch profile can be monitored via a pressure sensor in the air volume of the medium reservoir (P₂), which is connected to the main chamber. The apical compartment of the bioreactor can be connected to a nebulizer to deliver aerosolized particles/drugs to the cells. (B) Snapshot of the main chamber of the CIVIC bioreactor system. (C) Photograph of the BETA membrane in the PC holder of the CIVIC system, which is transparent, thus favorable for direct cell imaging applications.

Figure 2

Characterization of the biphasic (BETA) membrane used in this study. (A) Schematic depiction of the biphasic membrane concept. (Top) During phase I, gelatin forms a hydrogel due to contact with water (cell culture medium), which serves as adhesion point for epithelial cells of the lung and facilitates subsequent cell proliferation until a confluent epithelial cell layer is formed. After 4 days (phase II), the gelatin has been dissolved in water leaving behind a network of interconnected pores in the PCL membrane, which provides space for further cell spreading and at the same time enhances both membrane permeability and elastic modulus. (Bottom) Different membranes with various combinations of mixing ratio of PCL and gelatin—in the PCL/gelatin solution used for membrane manufacturing—expected to obtain a wide range of physicomechanical properties. (B) Top view of the ultrastructure of the “optimum” membrane (9.35% PCL and 6.34% gelatin [w/v of TFE], i.e., P/G = 9.35/6.34), which is also used for the analysis presented in (C,D). The scale bar is 100 μm. (C) Surface topography of an 80 × 80 μm² section of the membrane analyzed by Atomic Force Microscopy (AFM; left) and its corresponding z-amplitude profile (right) showing an average roughness height of 1.31 μm. (D) Cross-sectional analysis of the membrane using Focused Ion Beam-Scanning Electron Microscopy (FIB-SEM, left panel) and Energy Dispersive X-ray Spectroscopy FIB-SEM (EDS-FIB-SEM, right panel), indicating that gelatin is distributed throughout the PCL membrane during the late stage of phase I. The thickness of the membrane thickness ca. 5 μm. The scale bar is 2 μm. The newly manufactured membranes (P15G6, P15G8, and P15G10) are analyzed (and compared to previously reported results (Doryab et al., 2020) with respect to (E) Water Contact Angle (WCA), (F) Young's modulus (uniaxial, phase I, under dry conditions), (G) porosity obtained empirically by the liquid displacement method—and theoretical (maximum) porosity (gelatin volume fraction), (H) cell viability analyzed by WST1 assay on days 2, 4, and 6 of culture relative to PET Transwell® cell culture insert (H), and (I) cytotoxicity (LDH assay at day 6) of the three newly investigated membranes with different mixing ratios of PCL and gelatin. The LDH release for each membrane was normalized by the maximum possible LDH level (LDH contained in all cells). Typically, LDH < 10% is considered non-cytotoxic. There is no significant difference between the LDH release of Transwell® inserts and the different mixing ratios of BETA membranes. Optimum values for WCA, Young's modulus, empirical porosity, and theoretical (maximum) porosity were 69 ± 5 [°], 9.0 ± 1.9 [MPa], 15.32 [%] and 37.6 [%], respectively. Data are reported as the mean ± SD, n = 3.

Figure 3

Bioactivity of the BETA membrane (optimum: 9.35% PCL and 6.34% gelatin [w/v of TFE]). (A) Z-stack Confocal Laser Scanning Microscopy (CLSM) of human alveolar epithelial cells (A549) on the membrane (under static submerged culture conditions for 6 days) demonstrating the formation of a confluent uniform cell layer. The cell nuclei (DAPI, blue), expression of the cell-cell adhesion protein E-cadherin (red), and formation of F-actin filaments (green). The scale bar is 10 μm. (B) SEM image of A549 cells after proliferation on a (left) biphasic membrane and (right) a commercial Transwell^® insert (6 days of submerged culture). The scale bar is 10 μm. (C) Effect of leaching from BETA and PDMS membrane on cell viability (WST1 assay; A549 cells). BETA and PDMS membranes were incubated for 2 days in cell culture medium and this medium was used to grow A549 cells for 1 and 4 days in a standard 12-well tissue culture well plate. The reduced cell viability for PDMS-conditioned medium after 4 days indicates that some substances (e.g., uncured oligomers) leaching from the PDMS have a cytotoxic effect. This effect is not seen for the BETA membrane. The viability data were normalized that of a standard 12- well cell culture plate with fresh medium. Data are reported as mean ± SD. n = 8; ^****P < 0.00001 by one-way ANOVA with Dunnett test. (D) Z-stack and orthogonal CLSM view of (XY) with side views of YZ (right) and XZ (bottom) optical projection of the human bronchial epithelial cells (16HBE14o⁻) on the membrane visualizing a confluent cell monolayer and the formation of tight junctions (culture conditions: 6 days submerged and 24 h ALI culture); cell nucleus (DAPI, blue), F-actin filaments (green); ZO-1 tight junction (red). The scale bars are 100 μm (for the 20× projection view) and 20 μm (for the 63× projection view).

Figure 4

Cellular uptake and membrane-association of aerosolized nano- and microparticles for alveolar epithelial (A549) cells and translocation across the epithelial barrier (2 h incubation). (A) Schematic of the CIVIC bioreactor system for particle study. A549 cells were seeded on the BETA membrane (cell density: 2 × 10⁵ cells cm⁻², 4 days LLC, and 1-day ALI culture). Amine-modified polystyrene (PS-NH₂) nano- and microparticles (100 and 1,000 nm diameter, respectively) are then nebulized onto the cells with the nebulizer of the bioreactor. After 2 h, the cells were fixed and prepared for CLSM analysis. (B) 3D reconstruction z-stack of CLSM images presented as orthogonal (XY) and side views (YZ, right) of monolayered, confluent cells on the membrane after nebulization of 100 and 1,000 nm fluorescently labeled, amine-modified polystyrene (PS-NH₂) particles under non-stretched and physiologically stretched (10% linear, 0.33 Hz for 2 h) under ALI conditions. Cell nucleus (DAPI, blue), particles (red) and F-actin filaments of the cytoskeleton (green). Arrows, arrowheads, and asterisk indicate internalized particles, cell-membrane associated (extracellular) particles (on the apical cell surface) and particles located between cells, respectively. (Scale bar: 20 μm). (C) Quantitative cellular uptake of particles measured by fluorescence intensity of z-stacks, showing that physiologic cyclic mechanical stretch enhances uptake of 100 nm NPs as compared to static conditions, while there is no effect on 1,000 nm microparticles [Representative images (z-stacks) were recorded at 5 independent fields of view for each sample (n = 4); region of interest: 134.95 × 134.95 μm]. Y-axis is presented fluorescence intensity data in a log scale. Data are reported as the mean ± SD; ^*P < 0.01 by two-way ANOVA and data were corrected by Sidac for multiple comparison tests. (D) Translocation of 100 and 1,000 nm particles across the cell layer grown on unstretched PET Transwell^® inserts and on the BETA membrane (under unstretched and stretched conditions) (n = 3). ^** Show the comparison between stretched with the corresponding experiment under unstretched conditions. Data are reported as the mean ± SD; ^**P < 0.001 by two-way ANOVA and data were corrected by Tukey for multiple comparison tests.

Figure 5

Qualitative comparison between key properties of conventional porous membranes (PET and PDMS) used for static and dynamic (cell-stretch) in vitro lung models at the air-liquid interface and the BETA membrane presented here. The scores are based on biomimetic relevance as compared to the alveolar basement membrane. The permeability was measured as apparent permeability (Papp) for FITC-dextran (4 kDa). Stretchability and through pores refer to the elastic modulus and pores connecting the apical and basal sides of the membrane, respectively.