Establishing An Epithelial-Endothelial Co-Culture Lung-on-a-chip For Investigation In Healthy and COPD Conditions

Summary

This protocol establishes a dynamic lung-on-a-chip co-culture system combining human primary airway epithelial and pulmonary microvascular endothelial cells for two models: healthy and COPD model. The system reproduces airway-vascular interactions under air-liquid interface and flow, and provides barrier assessment, cell phenotyping and RNA analysis.

Abstract

Two dynamic lung-on-a-chip co-culture models were developed to recreate the airway–vascular interface by incorporating primary airway epithelial and pulmonary microvascular endothelial cells. One model uses commercially available cells from healthy donors, whereas the other employs cells derived from patients with chronic obstructive pulmonary disease (COPD), enabling direct comparison of healthy and diseased conditions under identical culture conditions. Both models support epithelial differentiation under continuous perfusion and air–liquid interface (ALI) conditions, resulting in a physiologically relevant epithelial barrier supported by an underlying endothelial layer. The workflow includes extracellular matrix coating, cell seeding, ALI establishment, barrier function assessment, cell phenotyping, and cell type-specific RNA extraction. Permeability assays demonstrated robust barrier integrity in both systems, while immunofluorescence confirmed continuous localization of Zonula Occludens-1 (ZO-1) and vascular endothelial cadherin (VE-cadherin) at cell junctions. Epithelial differentiation was confirmed by β-tubulin IV staining and visualization of ciliary beating. High-quality RNA was obtained separately from each compartment, supporting cell type-specific molecular analyses. The COPD model recapitulated key features of the disease, notably enhanced mucus production, without compromising barrier function or cell viability. The use of patient-derived epithelial and endothelial cells preserves individual pathological characteristics, enabling investigation of intercellular mechanisms in COPD. This approach provides a physiologically relevant alternative to static in vitro and animal models and supports preclinical studies of inhaled drugs, pathogen exposure, and epithelial–endothelial interactions. Importantly, the healthy and COPD lung-on-a-chip models can be used either in combination for direct comparative studies or independently for mechanistic investigations, therapeutic testing, and drug discovery. The system can also be adapted to other respiratory conditions involving altered epithelial or endothelial compartments, offering a versatile and translational platform for respiratory research.

Introduction

The interaction between airway epithelial cells and pulmonary microvascular endothelial cells is essential for maintaining lung homeostasis and coordinating responses to injury, infection, and inflammation. Together, these cell types form a functional barrier that regulates gas exchange, immune surveillance, and vascular permeability. Their dynamic cross-talk involves paracrine signaling, direct cell–cell interactions, and responses to mechanical stimuli, all of which are critical for lung function and structural integrity¹.

In chronic obstructive pulmonary disease (COPD), a progressive lung disorder characterized by chronic inflammation, airway remodeling, and destruction of the alveolar–capillary interface, the barrier function of both the airway epithelium and pulmonary vascular endothelium is impaired and accompanied by aberrant cytokine release and dysregulated tissue repair mechanisms^2,3. However, the interplay between airway epithelial and pulmonary microvascular endothelial cells remains poorly understood. This limitation is partly due to the shortcomings of existing in vitro and in vivo models. Although animal models have provided important insights into respiratory physiology and pathology, their relevance is limited by interspecies differences in immune function, lung architecture, cellular signaling, and drug metabolism, which can hinder translation to human biology.

Conventional in vitro systems, such as two-dimensional (2D) cultures of immortalized airway epithelial cell lines, fail to recapitulate key features of the airway epithelium, including proper differentiation and biomechanical responsiveness. Although three-dimensional (3D) air–liquid interface (ALI) monocultures of primary airway epithelial cells improve physiological relevance, they remain static and lack interactions with other cell types and mechanical stimuli⁴. The lung is a dynamic organ continuously exposed to airflow, shear stress, extracellular matrix stiffness, and cyclic stretch generated by breathing. These factors strongly influence cellular behavior and tissue remodeling⁵. Therefore, static models are insufficient to replicate processes such as leukocyte adhesion, immune cell trafficking, and real-time responses to drug exposure.

To address these limitations, organ-on-a-chip platforms have emerged as advanced systems for modeling complex cellular interactions. Lung-on-a-chip models integrate microfluidic technology with human-derived cells to replicate key structural and mechanical features of the lung microenvironment. These platforms enable co-culture of airway epithelial and pulmonary vascular endothelial cells within a dynamic, three-dimensional environment that more closely mimics in vivo conditions⁶. The incorporation of physiological breathing motions, fluid flow, and real-time monitoring enhances the study of epithelial–endothelial communication under both normal and diseased conditions^7–9.

Despite these advances, most existing lung-on-a-chip systems rely on cells derived from healthy donors¹⁰, limiting their ability to model disease-specific interactions. In addition, many platforms lack the flexibility to incorporate patient-derived cells or to recreate the complex COPD microenvironment. Furthermore, detailed protocols for co-culture of primary airway epithelial and pulmonary microvascular endothelial cells—whether from commercial sources or from COPD patients—using cell-type-specific extracellular matrix (ECM) coatings remain limited.

In this study, a commercially available two-channel microfluidic lung-on-a-chip platform was used to establish two co-culture models comprising primary airway epithelial and pulmonary microvascular endothelial cells. A healthy model was generated using commercially available cells derived from different donors, whereas a COPD model was established using airway epithelial and endothelial cells obtained from the same patient. Both models employ a cell-type-specific ECM coating to better mimic the in vivo lung microenvironment. These systems enable investigation of epithelial–endothelial interactions under physiological and pathological conditions. The healthy and COPD models can be used either in combination for direct comparative studies or independently for mechanistic investigations and therapeutic testing.

This co-culture system reflects the anatomical organization of the airway epithelium and pulmonary vasculature, which function as an integrated epithelial–endothelial unit. The airway epithelium is a pseudostratified layer composed of basal, secretory, and ciliated cells, located above a vascularized connective tissue layer (lamina propria) and separated from the endothelium by a basement membrane. In this model, the ECM coating (collagen type I, fibronectin, and laminin) approximates the composition of epithelial and endothelial basement membranes. The porous membrane separating the two channels mimics the subepithelial connective tissue, although it lacks vascular complexity. Multiple mechanisms contribute to epithelial–endothelial communication, including paracrine signaling, coordinated regulation of permeability, immune cell recruitment, and responses to mechanical stimuli such as breathing-induced stretch, which influence vascular resistance and mechanotransduction pathways.

This methodological study outlines the steps required to establish a fully differentiated airway epithelium cultured under air–liquid interface (ALI) conditions with an underlying pulmonary microvascular endothelium. The system enables qualitative and functional assessment of barrier integrity, achieving reproducible apparent permeability (Papp) values ≤ 1 × 10⁻⁶ cm/s. In addition, protocols are provided for cell phenotyping and for the extraction and quantification of RNA from each compartment, consistently yielding high-quality RNA (RNA integrity number > 7). This method is not intended for quantitative mRNA analysis or detailed assessment of mucus production, nor does it investigate the effects of mechanical stretching, although such applications are feasible with this system.

Protocol

All use of primary human cells was approved by the Local Ethics Committee Research UZ Leuven (S51577/S57114/S70453). Commercially available primary human bronchial epithelial cells (HBEC) and pulmonary microvascular endothelial cells (HPMEC) were obtained from two non-smoking donors (Supplementary Tables 1, 2). Primary cells for the COPD model were isolated from explanted lung tissue of three COPD patients with a history of smoking (Supplementary Table 3), according to a well-established protocol (ethical approval S51577) and stored in the BREATHE biobank. Both HBEC and HPMEC were obtained from the same COPD patient.

To establish the healthy lung-on-a-chip model, commercial HBEC and HPMEC were used. For the COPD model, primary bronchial epithelial and pulmonary microvascular endothelial cells were isolated from explanted lung tissue obtained from the same patient following lung transplantation. The lung-on-a-chip system used in this study is a microfluidic two-channel platform comprising parallel airway (top) and vascular (bottom) microchannels separated by a porous, extracellular matrix-coated membrane (Figure 1). The same protocol was used to build both healthy and COPD models. The chemicals, reagents, and kits used in the protocol are listed in the Table of Materials.

1. Establishment of the lung-on-a-chip co-culture models under laminar flow

1. Chip preparation
   1. Place the chip cradle into a 120-mm square cell culture dish, with the corners facing up, to ensure correct orientation, then place the chips into the cradle. Ensure that all chips are placed properly prior to the activation step (Figure 2).
   2. Prepare the activation solution by dissolving the activation reagent powder in its corresponding buffer, both supplied by the manufacturer, to obtain a final concentration of 0.5 mg/mL.
   3. Activate the inner surfaces of the chip channels to promote extracellular matrix (ECM) attachment by carefully introducing approximately 35 µL of the activation solution through the bottom channel inlet. Pipette until the solution begins to exit from the bottom channel outlet. Repeat the procedure by introducing approximately 50 µL of the solution into the top channel inlet.
      NOTE: The activation reagent powder is light-sensitive. Prepare the activation solution immediately before use and discard any remaining solution 1 h after reconstitution.
   4. Remove all excess activation solution from the surface of the chip by gentle aspiration. Be sure only to remove the activation solution from the chip surface — do not aspirate it from the channels.
      NOTE: Prior to ultraviolet (UV) activation, inspect the channels under a microscope to detect bubbles. If any are present, dislodge them by gently washing with activation solution until all bubbles are removed.
   5. Expose the chips to a UV lamp for 10 min, then wash each channel twice with 200 µL of the corresponding buffer.
   6. Pipette fresh activation solution into both channels as described above and expose the chips to UV for an additional 5 min.
   7. Fully aspirate the solution from both channels, wash each channel twice with 200 µL of corresponding buffer, then wash each channel once with 200 µL of sterile cold phosphate-buffered saline without Ca²⁺ and Mg²⁺ (DPBS -/-).
   8. Prepare an ECM coating solution in DPBS (-/-) containing human collagen type I (100 µg/mL), human fibronectin (50 µg/mL), and human laminin (50 µg/mL). Use 100 µL of solution as reference volume per chip.
   9. Add 50 µL to the top channel and 35 µL to the bottom channel, while simultaneously aspirating the outflow through the outlet ports.
      NOTE: Inspect the channels under a microscope to detect any bubbles. If present, gently wash with the ECM solution to remove them.
   10. Leave small droplets of ECM solution on both channel inlet and outlet ports to prevent evaporation during overnight (ON) incubation.
   11. Incubate the chips ON at 37 °C with 5% CO₂. The next day, aspirate the ECM solution and wash both channels with 200 µL of pre-warmed DPBS (-/-) at 37 °C.
2. Cell seeding and submerged culture
   Expand HBEC in T25 cm² flasks coated with human collagen type IV (150 µg/mL in DPBS (-/-)) using airway epithelial cell expansion medium until 60%–70% confluency is reached.
   In parallel, expand HPMEC in T75 cm² flasks coated with gelatin (2 mg/mL in DPBS (-/-)) using endothelial cell (EC) growth medium until 90%–100% confluency is reached.
   Refer to Supplementary Table 4 for the composition of key reagents and culture media.
3. Seeding HPMEC
   1. Aspirate the medium. Wash cells with 10 mL DPBS (-/-).
   2. Add 3 mL of 0.25% trypsin-EDTA (1X). Incubate for 30 s at room temperature (RT).
   3. Add 12 mL EC growth medium (4× the volume of trypsin used) to inactivate trypsin. Add 5 mL EC growth medium to collect any residual cells (final volume = 20 mL).
   4. Confirm complete cell detachment under the microscope.
   5. Take ~20 µL to count live cells using trypan blue.
   6. Centrifuge the remaning cell suspension at 900 × g for 5 min at RT in 50 mL conical tube.
   7. Resuspend the pellet in EC growth medium. Adjust to 7 × 10⁶ cells/mL.
   8. Seed HPMEC (passage 5) into the bottom channel (~20 µL; 1.4 × 10⁵ cells/channel).
   9. Immediately invert the chips and incubate for 2 h at 37 °C with 5% CO₂ (Figure 3).
   10. Return chips to original orientation.
   11. Wash with 200 µL warm (37°C) EC growth medium.
   12. Verify attachment under the microscope. Extend incubation if needed.
4. Seeding HBEC
   1. Aspirate the medium. Wash with 5 mL warm DPBS (-/-).
   2. Add 1 mL trypsin. Incubate for 5 min at 37 °C with 5% CO₂.
   3. Monitor detachment under the microscope.
   4. Add 1 mL defined trypsin inhibitor (DTI) at a 1:1 ratio to inactivate trypsin.
   5. Add 8 mL airway epithelial cell expansion medium (final volume = 10 mL).
   6. Transfer to a 50 mL tube.
   7. Take ~20 µL for cell counting using trypan blue.
   8. Centrifuge the remaining cell suspension at 200 × g for 5 min at RT.
   9. Discard the supernatant.
   10. Resuspend the pellet in airway epithelial cell expansion medium. Adjust to 3.5 × 10⁶ cells/mL.
   11. Supplement the airway epithelial cell expansion medium with 50 nM synthetic light-stable retinoic acid and 10 ng/mL recombinant human epidermal growth factor (rhEGF).
   12. Seed HBEC (passage 2) into the top channel (~35 µL; 1.2 × 10⁵ cells/channel).
   13. Incubate for 4 h at 37 °C with 5% CO₂ (Figure 4).
   14. Replace medium in both channels.
   15. Incubate overnight (ON) under static conditions at 37 °C with 5% CO₂.
       NOTE: From this step onwards, use only degassed medium for all steps. Degassing must be performed with a Steriflip unit connected to a vacuum system to minimize the risk of bubble formation in the chip.
5. Connection to flow
   1. Add warm (37°C) airway epithelial expansion medium supplemented with synthetic light-stable retinoic acid and rhEGF to the top inlet reservoir.
   2. Add warm (37°C) EC growth medium to the bottom inlet reservoir.
   3. Fill each inlet reservoir with 3 mL medium and each outlet reservoir with 300 µL.
   4. Place the portable modules into the culture module (Figure 5).
   5. Press the dial to select the “Prime” cycle using the rotary dial.
   6. Press the dial to initiate the cycle.
   7. Allow the cycle to run (~1 min).
   8. Inspect the bottom of each module for droplets at all four fluidic ports.
   9. Repeat the “Prime” cycle if droplets are not visible.
   10. Wash both chip channels once with the appropriate medium.
   11. Leave a droplet on each inlet and outlet port to prevent bubble formation.
   12. Hold the chip in the dominant hand and the portable module in the other hand.
   13. Slide the chip into the tracks until fully seated.
   14. Place the assembled unit into the culture module.
   15. Set the flow rate to 30 µL/h for both channels using the rotary dial.
   16. Maintain liquid–liquid interface (LLI) conditions.
   17. Keep chips to the module under liquid-liquid interface (LLI) to allow uniform epithelial and endothelial layer formation
   18. Replace medium every 3 days. Aspirate medium from outlet reservoirs. Add 3 mL fresh medium to each inlet reservoir.
       NOTE: Monitor for bubbles under the microscope. If present, disconnect the chip, flush channels with medium, and reconnect.
6. Epithelial cell differentiation and air liquid interface culture
   1. Monitor cells daily under the microscope.
   2. Confirm formation of a homogeneous epithelial cell layer without gaps (typically 3–5 days).
   3. Establish the air–liquid interface (ALI).
   4. Aspirate medium from the top inlet and outlet reservoirs.
   5. Flush out residual medium from top channel by increasing flow rate to 600µL/h using the rotary dial
   6. Rapidly aspirate medium from the top outlet reservoirs.
   7. Aspirate medium from the bottom inlet reservoirs.
   8. Add 3 mL epithelial differentiation medium supplemented with 10 ng/mL vascular endothelial growth factor (VEGF) and 1 µg/mL vitamin C.
   9. Set the top channel to “air” using the rotary dial.
   10. Set the bottom channel flow rate to 30 µL/h.
   11. Maintain ALI culture for 2 weeks.
   12. Observe the development of epithelial differentiation under the microscope.
   13. Confirm formation of a pseudostratified epithelium. Monitor cilia beating and mucus production.
   14. Monitor endothelial cell morphology at inlet and outlet regions.

Figure 1: Microfluidic lung-on-a-chip platform. Cross-section (left) and top view (right) of the two-channel microfluidic lung-on-a-chip device. The schematic illustrations are provided for illustrative purposes only. Please click here to view a larger version of this figure.

Figure 2: Positioning of the chip in the chip cradle. Representative image showing the correct orientation of the chip cradle and placement of the chip. Please click here to view a larger version of this figure.

Figure 3: Endothelial cell seeding in the microfluidic chip. Representative brightfield images of pulmonary microvascular endothelial cells immediately after seeding (A) and after 2 h of incubation (B). Scale bar: 100 µm. Please click here to view a larger version of this figure.

Figure 4: Airway epithelial cell seeding in the microfluidic chip.
Representative brightfield images of airway epithelial cells immediately after seeding (A) and after 4 h of incubation (B). Scale bar: 100 µm. Please click here to view a larger version of this figure.

Figure 5: Configuration of the microfluidic lung-on-a-chip system.
The upper part of the figure shows the portable module with four reservoirs (two inlets and two outlets) used for medium supply and effluent collection. The microfluidic device is connected to a culture module that regulates flow rates and to a hub module that provides gas, air, and power. The schematic representation is provided for illustrative purposes only. Please click here to view a larger version of this figure.

2. Assessment of barrier integrity

Barrier integrity was assessed in two ways: firstly, functionally, by evaluating permeability to fluorescently labeled dextran over time (before ALI and after 10 days of ALI), and secondly, qualitatively, by immunodetection of Zonula Occludens-1 (ZO-1) in the bronchial epithelium and Vascular Endothelial-cadherin (VE-cadherin) in the pulmonary vascular endothelium.

1. Barrier function
   1. Under the laminar flow, add 1 mL of epithelial differentiation medium supplemented with VEGF and Vitamin C and containing 100 µg/mL fluorescently labeled dextran to the top inlet reservoir of the airway channel. Simultaneously, add 1 mL of tracer-free medium to the bottom inlet reservoir of the vascular channel.
   2. Introduce the medium containing the tracer into the top channel. Using the rotary dial, set the flow rate in both channels to 600 µL/h for 3 min, then aspirate medium from both top and bottom outlet reservoirs.
   3. Place the chip, connected to the portable module, back into the culture module. Using the rotary dial, set both channels to a flow rate of 45 µL/h and maintain this flow rate for 3 h.
   4. Collect 100 µL of medium from top and bottom outlet reservoirs and measure fluorescence intensity using a plate reader.
   5. Calculate apparent permeability (Papp) using the permeability equation:
      
      where SA represents the surface area of the co-culture channel (0.17 cm²), Q_R and Q_D denote the fluid flow rates in the dosing and receiving channels respectively (cm³/s) and C_R,0 & C_D,0 correspond to the recovered concentrations in the dosing and receiving channels, respectively.
      NOTE: A Papp value below 1 x 10^-6 cm/s indicates strong barrier integrity, reflecting low permeability of the epithelial layer to the tracer. This threshold is set by the system manufacturer as an internal quality control measure.
2. Immunodetection of tight and adherens junctions
   1. After 2 weeks of ALI, carefully remove the chip from the portable module.
   2. Fix the cells by adding 50 µL of 4% paraformaldehyde (PFA) to both channels for 20 min at RT.
      CAUTION: PFA is toxic; perform all steps in a certified chemical fume hood.
   3. Wash both channels twice with 200 µL of DPBS (-/-) and store the chips at 4 °C until staining.
   4. Permeabilize the cells by adding 50 µL of 0.1% Triton X-100 in DPBS (-/-) for 5 min at RT. Then block with 50 µL 0.1% Triton X-100 in DPBS (-/-) containing 10% goat serum, per channel for 1 h at RT.
   5. Incubate the top (bronchial epithelial) channel with 50 µL of mouse monoclonal anti-human ZO-1 primary antibody (1:100), and the bottom (pulmonary vascular endothelial) channel with 50 µL of rabbit monoclonal anti-human VE-cadherin primary antibody (1:600).
   6. Dilute both antibodies in antibody solution (0.1% Triton X-100 in PBS (-/-) with 1% goat serum) and incubate ON at 4°C under static conditions.
   7. Wash the chips three times with 200 µL of DPBS (-/-). Then incubate each channel for 1 h at RT with 50 µL of appropriate fluorophore-conjugated secondary antibodies diluted in antibody solution (1:500): goat anti-mouse secondary antibody in the top channel and goat anti-rabbit secondary antibody in the bottom channel.
   8. Wash both channels three times with 200 µL DPBS (-/-). Perform nuclear staining with 50 µL for each channel of 4′,6-diamidino-2-phenylindole (DAPI, 1 µg/mL in DPBS (-/-)) for 10 min at RT.
   9. Perform a final wash with 200 µL of DPBS (-/-) and store the chip in DPBS (-/-) at 4°C in the dark until imaging.
   10. Image the chips on coverslips (20x60 mm with 0.16-0.19 mm thickness) using an inverted fluorescence microscope equipped with a spinning-disk module, a high-sensitivity camera, and a 20X objective (NA 0.8). Use appropriate image acquisition and analysis software for post-processing.

3. Phenotyping

1. Preparation of chip sections
   1. Fix the chips with 4% PFA as described in section 2.2.2 .
   2. Cut away the surrounding polydimethylsiloxane (PDMS) from the chip using a razor blade.
   3. Fill both channels with 50 µL of 2% low-melting agarose in DPBS (-/-), and allow agarose to solidify at RT.
   4. Cut the chip transversely into two halves using a razor blade, embed each half in an optimal cutting temperature (OCT) embedding medium and store at −80 °C until sectioning.
2.  Cryosection
   1. Prepare a cryostat and cut 200 µm cryosections from chip halves.
   2. Collect chip sections with tweezers into wells of a 48-well plate filled with DPBS (-/-) and store them at 4 °C until immunostaining.
3. Immunofluorescent staining
   NOTE: Perform each step in the 48-well plate, using 400 µL as the reference volume.
   1. Permeabilize the cells and block nonspecific binding as previously described in section 2.2.4.
   2. Incubate each section ON at 4°C with primary antibodies specific for epithelial ciliated cells (β-tubulin IV, 1:2000) or for endothelial cells (CD31, 1:25), diluted in antibody solution (0.1% Triton X-100 in DPBS with 1% goat serum).
   3. Wash sections three times with DPBS (-/-).
   4. Incubate sections for 1 h at RT with appropriate fluorophore-conjugated goat anti-mouse secondary antibodies (1:500) for β-tubulin IV and CD31, diluted in antibody solution, protected from light.
   5. Wash the sections three times with DPBS (-/-), and perform nuclear staining with DAPI (1 µg/mL in DPBS (-/-)) for 10 min at RT.
   6. Wash sections once with DPBS (-/-) before mounting.
   7. Using tweezers, place one spacer on each microscope slide, then place a section in the middle. Add 1–2 drops of mounting medium and place a circular coverslip.
   8. Let them dry ON in the dark at RT, then seal the coverslip edges with transparent nail polish.
   9. Store the slides at 4°C in the dark until imaging.
   10. Image the chips sections using an inverted fluorescence microscope, equipped with a spinning-disk module, a high-sensitivity camera, and a 40X objective (NA 1.25). Process images using appropriate microscopy software.

4. RNA extraction and quantification

1. Collection of cells from chips under laminar flow
   1. After 2 weeks of ALI culture, disconnect the chips from the portable modules, and wash both channels twice with 200 µL of DPBS (-/-) prewarmed at 37 °C.
   2. Add 35 µL of 0.25% Trypsin-EDTA (1X) to the endothelial cells in the bottom channel and incubate for approximately 30 sec at RT.
   3. Inspect the cells under the microscope to confirm complete detachment. Then collect the cell suspension into a tube and centrifuge at 900 × g for 5 min at RT.
   4. Resuspend the cell pellet in 50 µL of phenol-based RNA isolation reagent and store at -80 °C until RNA extraction.
   5. Add 50 µL of phenol-based RNA isolation reagent directly to the top channel to lyse epithelial cells, seal ports with tips, and incubate on ice for 3-4 min.
   6. Gently pipette to ensure complete lysis, collect lysate into a tube, and store it at -80 °C until RNA extraction.
   7. Inspect both channels under a microscope to confirm complete cell collection.
2. RNA extraction
   1. Add chloroform (0.2 volume per 1 volume of phenol-based RNA isolation reagent used for lysis; 10 µL per sample), mix tubes thoroughly by shaking and incubate on ice for 2-3 min.
   2. Centrifuge at 12,000 × g for 15 min at 4 °C to separate phases and carefully transfer the upper aqueous phase into a new tube without disturbing the interphase. Then, add 5 µg of RNA-free glycogen to the aqueous phase as a carrier.
   3. Add isopropanol (0.5 volume per 1 volume of lysis reagent; 25 µL per sample), incubate samples at 4 °C for 10 min, and centrifuge at 12,000 × g for 10 min at 4 °C.
   4. Discard the supernatant while retaining the RNA pellet.
3. Washing and resuspension of RNA
   1. Wash the RNA pellet with 75% ethanol (1 volume per 1 volume of lysis reagent; 50 µL per sample), vortex briefly and centrifuge at 7,500 × g for 5 min at 4 °C.
   2. Discard the supernatant and allow the pellet to air-dry for 5–10 min.
   3. Resuspend RNA in 30 µL of RNase-free water by gentle pipetting and store RNA at -80 °C until further use.
   4. Assess RNA integrity and quantification using a microfluidic chip-based nucleic acid separation system.

Results

Establishment of the lung-on-a-chip co-culture models

Brightfield imaging revealed cell adhesion and growth over time in the lung-on-a-chip co-culture system. Bronchial airway epithelial cells proliferated in submerged culture at the liquid–liquid interface (LLI), forming a uniform layer along the top channel (Figure 6A), while pulmonary vascular endothelial cells expanded to cover the entire bottom channel, forming a confluent monolayer (Figure 6B).

Once the co-culture was established, endothelial cells were monitored at the inlet and outlet regions of the bottom channel, which are the only areas accessible for microscopic inspection. As shown in Figure 7, endothelial cells remained viable and exhibited an elongated morphology aligned with the direction of flow. After 14 days of ALI culture, the epithelial layer in both healthy and COPD models displayed visible ciliary beating (Supplementary Video). As expected, the COPD model exhibited markedly increased mucus production compared to the healthy model (Figure 8), recapitulating key features of COPD pathology.

Figure 6: Co-culture of airway epithelial and endothelial cells in the microfluidic device.
Representative images of primary bronchial epithelial cells on the upper surface (A) and primary pulmonary microvascular endothelial cells on the lower surface (B) of the porous membrane under liquid–liquid interface (LLI) conditions. Scale bar: 100 µm. Please click here to view a larger version of this figure.

Figure 7: Endothelial cells in the bottom microfluidic channel during air–liquid interface.
Representative brightfield images showing pulmonary microvascular endothelial cells at the inlet (upper) and outlet (lower) of the bottom channel during air–liquid interface (ALI). Scale bar: 100 µm. Please click here to view a larger version of this figure.

Figure 8: Mucus accumulation in the airway channel.
Representative images showing mucus production after 14 days of air–liquid interface (ALI) in healthy (A) and COPD (B) co-culture models. The brown layer in the airway compartment represents accumulated mucus. Scale bar: 100 µm. Please click here to view a larger version of this figure.

Quantitative assessment of barrier integrity

Functional assessment of barrier integrity showed lower apparent permeability (Papp) in both healthy and COPD models compared to the positive control (ECM-coated chip without cells) (Figure 9).

In the absence of cells, high Papp values indicated minimal barrier formation. In contrast, in the presence of epithelial and endothelial layers, the fluorescent dextran tracer did not cross the ECM, resulting in markedly decreased permeability and indicating a strong barrier. This effect was observed both prior to ALI and after 10 days of ALI, demonstrating that barrier function is established early during differentiation.

As shown in Figure 10, in the healthy model, mean Papp values before ALI were already below the threshold indicative of robust barrier integrity (1 × 10⁻⁶ cm/s) and decreased further during ALI culture. In the COPD model, mean Papp values were also within the strong barrier range before ALI and, although slightly increased after 10 days of ALI, barrier integrity remained robust.

Figure 9: Barrier function assessment in the lung-on-a-chip model.
Apparent permeability (Papp) measured in ECM-coated chips without cells (positive control for leakage; black), and in healthy (pink) and COPD (green) co-culture models before and after 10 days of air–liquid interface (ALI). Each dot represents an individual chip. The healthy model includes cells from separate donors, whereas the COPD model uses cells derived from a single patient. Please click here to view a larger version of this figure.

Figure 10: Comparison of barrier function in healthy and COPD models.
Apparent permeability (Papp) measured before and after 10 days of air–liquid interface (ALI) in healthy (left) and COPD (right) co-culture models. Each dot represents an individual chip. Values are mean ± SD. The COPD model includes cells from three patients, whereas the healthy model combines cells from different donors. Please click here to view a larger version of this figure.

Immunodetection of tight and adherens junctions

Immunofluorescence analysis confirmed the structural organization of junctional proteins. ZO-1 showed continuous localization along bronchial epithelial cell borders, indicating an intact epithelial barrier (Figure 11A). Similarly, VE-cadherin staining revealed uniform distribution at endothelial junctions, consistent with a confluent endothelial monolayer. Endothelial cells exhibited an elongated morphology aligned with the direction of flow (Figure 11B).

No differences were observed between the healthy and COPD models in the localization of tight and adherens junction proteins. These findings are consistent with the functional barrier data, supporting the presence of a physiologically relevant epithelial–endothelial interface with preserved junctional integrity in both models.

Figure 11: Junctional protein expression in the lung-on-a-chip co-culture model.
Top-view immunofluorescence images obtained after 14 days of air–liquid interface (ALI) showing ZO-1 localization in the airway epithelial compartment (A) and VE-cadherin localization in the endothelial compartment (B). Nuclei are stained with DAPI. The arrow indicates flow direction. Scale bar: 100 µm. Please click here to view a larger version of this figure.

Phenotyping

Immunofluorescent staining of chip cryosections revealed β-tubulin IV-positive cells on the apical side of the membrane, indicating differentiation of airway epithelial cells into ciliated cells, a defining feature of mature bronchial epithelium (Figure 12A). CD31 staining confirmed the presence of a confluent endothelial monolayer along the basal side of the membrane (Figure 12B). These results demonstrate that the co-culture system supports the establishment of physiologically relevant epithelial and endothelial phenotypes in both models.

Figure 12: Immunofluorescence images of lung-on-a-chip cross-sections. Representative immunofluorescent staining of chip cross-sections after 14 days of air–liquid interface (ALI) showing β-tubulin IV expression in the airway epithelial compartment (A) and CD31 expression in the endothelial compartment (B). Nuclei are stained with DAPI. Scale bar: 50 µm. Please click here to view a larger version of this figure.

RNA extraction and quantification

RNA isolated from airway epithelial cells yielded concentrations ranging from 59 to 135 ng/µL, with RNA Integrity Numbers (RIN)¹¹ of 7.4–9.6, whereas RNA from microvascular endothelial cells ranged from 12 to 46 ng/µL, with RIN values of 9.6–10, in both models. Data indicated that both models achieved sufficient quantities of high-quality RNA, suitable for downstream molecular analyses.

Supplementary Table 1: Donor information for the bronchial airway epithelial cells. The table represents bronchial airway epithelial cells used in the healthy model.Please click here to download this file.

Supplementary Table 2: Donor information for the pulmonary microvascular endothelial cells. Donor information for pulmonary microvascular endothelial cells used in the healthy model.Please click here to download this file.

Supplementary Table 3: Patient information. The table represents information for airway epithelial and pulmonary microvascular endothelial cells used in the COPD model.Please click here to download this file.

Supplementary Table 4: Overview of key reagents and culture media used in the protocol.Please click here to download this file.

Supplementary Video 1: Ciliary beating in the airway channel (healthy model).
Representative video showing ciliary beating after 14 days of air–liquid interface (ALI).Please click here to download this file.

Supplementary Video 2: Ciliary beating in the airway channel (COPD model).
Representative video showing ciliary beating after 14 days of air–liquid interface (ALI).Please click here to download this file.

Discussion

This study presents a detailed protocol for establishing a lung-on-a-chip co-culture system to generate both a healthy model, using commercially available bronchial epithelial and pulmonary microvascular endothelial cells, and a COPD model using patient-derived cells from explanted lungs. In both models, the two cell types are separated by a porous, stretchable membrane coated with cell type-specific extracellular matrix (ECM), enabling a microenvironment that more closely resembles the in vivo lung. The protocol also includes methods for assessing barrier integrity, cell phenotyping, and RNA extraction from each compartment. Both models were successfully maintained under dynamic perfusion and air–liquid interface (ALI) conditions, resulting in a physiologically relevant epithelial–endothelial barrier with a fully differentiated pseudostratified airway epithelium. Endothelial cells aligned with the direction of flow, as observed in vivo, and high-quality RNA was obtained from both compartments.

Several critical steps are essential for establishing this co-culture system successfully. Precise epithelial cell seeding is required to achieve uniform coverage, as non-uniform distribution can result in incomplete surface coverage and compromised barrier formation. In such cases, the apical compartment should be gently aspirated, and epithelial cells reseeded. The establishment of ALI should only be performed once both epithelial and endothelial layers form continuous monolayers, as premature ALI exposure may lead to leakage of vascular medium into the airway compartment and impaired differentiation. If confluency is not achieved, ALI establishment should be delayed.

Incomplete removal of apical medium or inconsistent ALI exposure may lead to heterogeneous epithelial maturation. Residual medium in the top channel can be removed by increasing the flow rate to 600 µL/h for 3–4 min. Bubble formation, which disrupts perfusion and affects cell viability, may occur due to insufficient equilibration. Bubbles can be removed by temporarily increasing the flow to 600µL/h for 3-4 min or, if persistent, by disconnecting the chip and manually dislodging them using a pipette. In the COPD model, excessive mucus production may obstruct the airway channel and impair perfusion; in such cases, gentle washing with DPBS followed by re-establishment of ALI is recommended.

Most existing co-culture systems combining airway epithelial and pulmonary endothelial cells rely on commercially available primary cells and generic ECM coatings such as collagen or fibronectin^10,12. While these support adhesion, they only partially reproduce the biochemical and mechanical properties of the airway–vascular interface. In contrast, the present system employs cell type-specific ECM coatings, providing a microenvironment that more closely mimics human lung tissue¹³.

Barrier function is commonly assessed using transepithelial/transendothelial electrical resistance (TEER), although this method can be influenced by epithelial multilayer organization¹⁴. In this study, dextran permeability was used to assess functional integrity. The observed Papp values were below the threshold for strong barrier integrity (1 × 10⁻⁶ cm/s), indicating a robust epithelial–endothelial interface. The airway epithelium exhibited differentiation features including ciliary beating and mucus production, along with continuous localization of ZO-1 and VE-cadherin at cell junctions. In addition, high-quality RNA was consistently obtained from both compartments, enabling cell-type-specific downstream analyses.

A key advance of this work is the development of a COPD lung-on-a-chip model in which both epithelial and endothelial cells are derived from the same patient. To recent knowledge, such a patient-matched co-culture has not been previously reported. Earlier studies have incorporated COPD-derived epithelial cells^7,8 but relied on commercially sourced endothelial cells, limiting investigation of disease-specific epithelial–endothelial interactions. The present model overcomes this limitation by preserving patient-specific cellular interactions.

The COPD model recapitulated key disease features, including increased mucus production³, while maintaining endothelial alignment and junctional integrity under flow. Importantly, the co-culture of patient-derived epithelial and endothelial cells did not compromise barrier function or viability, despite increased mucus production. The system remained stable under ALI conditions and continuous perfusion, supporting its use for long-term studies. In addition, the platform allows mechanical stretch, further enhancing its physiological relevance.

An important strength of this workflow is the ability to evaluate healthy and COPD conditions within the same platform under comparable experimental parameters. This enables direct assessment of disease-associated differences in epithelial differentiation, endothelial function, and barrier integrity. The models can also be used independently, depending on the research question. The healthy model is suitable for mechanistic studies and pathway modulation, whereas the COPD model provides a platform for therapeutic evaluation and drug testing. The ability to extract high-quality RNA from each compartment separately further enables cell-type-specific molecular profiling. Using patient-matched cells in the COPD model preserves disease-relevant interactions that would otherwise be lost when combining unrelated cell sources.

Despite these advantages, several limitations should be considered. The use of lung-on-a-chip systems requires specialized equipment and technical expertise, which may limit accessibility. In addition, the model does not currently include other relevant lung cell types, such as immune or stromal cells, which may influence responses to pathogens and drugs. However, the platform can be adapted to incorporate circulating immune cells, enabling investigation of their interactions with epithelial and endothelial compartments. Access to patient-derived COPD material may also be limited, and inter-patient variability may affect reproducibility. Furthermore, excessive mucus production may interfere with perfusion.

The impact of cyclic mechanical stretching was not investigated in this study but represents an important future direction. The stretchable membrane allows simulation of breathing-associated mechanical forces, which are known to influence cellular function, repair, and homeostasis^15,16. Previous studies have shown that mechanical strain can enhance nanoparticle uptake and transport across the epithelial barrier¹⁰.

In conclusion, the healthy and COPD lung-on-a-chip models presented here provide a physiologically relevant platform for studying epithelial–endothelial interactions in health and disease. The system's dynamic nature enables precise control of flow conditions and supports applications such as drug testing and pathogen exposure studies. This protocol offers a robust alternative to traditional in vitro and animal models and represents a valuable tool for advancing translational respiratory research.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
0.25% trypsin-EDTA (1X)  (for endothelial cells)	Gibco	25200-056	Used for enzymatic detachment of pulmonary microvascular endothelial cells.
120 mm x 120 mm square cell culture dish	VWR	688161	Used to hold chip cradle during chip preparation and handling.
2100 Bioanalyzer	Agilent	G2939A	Used to assess RNA quality and integrity.
2-propanol	Sigma-Aldrich	I9516	Used to precipitate RNA during extraction.
4′,6-diamidino-2-phenylindole (DAPI)	Invitrogen	D1306	Used to stain cell nuclei during immunofluorescence.
50 mL conical tubes	Cellstar	227261	Used for cell collection, centrifugation, and preparation of cell suspensions.
Agilent RNA 6000 Nano Kit	Agilent	5067-1511	Used for RNA quality analysis with the Bioanalyzer.
Amphotericin B	Gibco	15290-026	Used to prevent fungal contamination during cell culture.
Automatic cell counter	NanoEntek	EVE PLUS	Used to determine cell concentration and viability prior to seeding.
Cell culture multiwell plates, 48 well, sterile	Falcon	353078	Used to hold cross-sections during staining procedures.
Centrifuge	Hettich	320R	Used to pellet cells prior to seeding.
Chip Cradle	Emulate	support@emulatebio.com	Used to support and hold chips during surface activation, coating, and cell seeding.
Chloroform	Sigma-Aldrich	366927	Used for phase separation during RNA extraction.
Cicero spinning disk module	CrestOptics S.p.A.		Used to perform high-speed confocal fluorescence imaging.
Cover slips	Epredia	BB02400600AC13MNZ0	Used to support chips during confocal imaging. Desciption: 24 x 60 mm, #1.5
Cryostat	Thermo Fisher Scientific	NX70	Used to section OCT-embedded samples.
Culture module (Zoë-CM)	Emulate	support@emulatebio.com	Used to control environmental conditions for cell culture within the microfluidic system.
Define trypsin inhibitor (DTI)	Thermo Fisher	R007100	Used to neutralize trypsin activity after epithelial cell detachment.
Dextran Cascade Blue 3KDa	Thermo Fisher	D7132	Used as a fluorescent tracer to assess barrier permeability.
Endothelial cells Growth Medium MV2 kit	PromoCell	C-22022	Used for maintenance and expansion of pulmonary microvascular endothelial cells.
ER-1  surface activation reagent (light-sensitive activation reagent powder)	Emulate	support@emulatebio.com	Used to activate chip surfaces for subsequent coating.
ER-2  surface activation reagent (buffer)	Emulate	support@emulatebio.com	Used as a buffer during surface activation of the chip.
Ethanol absolute	Thermo Fisher Scientific	437433T	Used to wash RNA pellets during extraction.
Fluorescent plate reader	Agilent	BioTek Sinergy H1 microplate reader	Used to measure fluorescence intensity for permeability assays.
Gelatin from bovine skin	Sigma-Aldrich	G9391	Used to coat culture surfaces to promote endothelial cell adhesion.
Goat Anti-Mouse Alexa Fluor 488 (for ZO-1 and β-tubulin IV)	Thermo Fisher	A11001	Used to detect mouse primary antibodies with green fluorescence.
Goat Anti-Mouse Alexa Fluor 594 (for CD31)	Thermo Fisher	A11005	Used to detect mouse primary antibodies with red fluorescence.
Goat Anti-Rabbit Alexa Fluor 555 (for VE-cadherin)	Thermo Fisher	A21429	Used to detect rabbit primary antibodies with orange-red fluorescence.
Hub module (Orb-HM)	Emulate	support@emulatebio.com	Used to supply gas and power to the culture module.
Human Bronchial Epithelial Cells (HBEC)	Lonza	CC-2540S	Used as airway epithelial cells to establish the healthy co-culture model.
Human Collagen  IV	Sigma	C5533	Used to coat T25 flasks to promote epithelial cell attachment.
Human Collagen type I	Advanced Biomatrix	5007	Used as a component of the extracellular matrix solution for chip coating.
Human Fibronectin	Sigma-Aldrich	F0895	Used as a component of the extracellular matrix solution for chip coating.
Human Laminin	Sigma -Aldrich	L4544	Used as a component of the extracellular matrix solution for chip coating.
Human Pulmonary Microvascular Endothelial Cells (HPMEC)	PromoCell	C-12281	Used as pulmonary microvascular endothelial cells to establish the healthy co-culture model.
human recombinant epidermal growth factor (hrEGF)	PromoCell	C-60182	Used to promote epithelial cell proliferation.
Hydrocortisone stock solution	StemCell Technologies	7925	Used as a supplement in culture media to support airway epithelial cell differentiation.
ImageJ	Fiji software	version ImageJ 1.54p	Used to process and analyze acquired images.
Inverted microscope for cell culture	Leica Microsystems	DMi1	Used to monitor cell morphology and confluency during culture.
iSpacer	SUNJin Lab	IS307	Used to create a defined chamber for sample mounting during imaging. Thickness: 0.3 mm.
Kinetix camera	Teledyne Photometrics		Used to capture high-resolution fluorescence images.
Microscope Camera	Leica Microsystems	FLEXA CAM C1	Used to capture images of cells during culture and experiments.
Microscope cover glass circular	Marienfeld	0111620	Used for covering chip sections on glass slides. Description: 22 mm
Monoclonal Mouse Anti Human CD31 - clone JC70A	Agilent	M082329-2	Used to detect endothelial cell marker CD31.
Monoclonal Mouse Anti ZO-1 - clone ZO1-1A12	Invitrogen	33-9100 (100µg)	Used to detect tight junction protein ZO-1.
Monoclonal Mouse Anti β-tubulin IV antibody - clone ONS.1A6	Sigma-Aldrich	T7941  (0.2 mL)	Used to detect ciliated epithelial cells.
Monoclonal Rabbit Anti VE-Cadherin	Cell Signaling Tech.	D87F2	Used to detect endothelial junction protein VE-cadherin.
Mounting medium	Ibidi	50001	Used to mount samples and preserve fluorescence signal during imaging.
Nikon Ti2 inverted microscope	Nikon		Used to image chip samples under fluorescence conditions.
NIS-Elements	Nikon Instruments Europe B.V.	version 6.10.01 / 6.20.01	Used to control image acquisition during microscopy experiments.
Normal goat serum	Abcam	ab7481	Used to block non-specific antibody binding during immunostaining.
Optimal Cutting Temperature (OCT)	VWR	361603E	Used to embed samples for cryosectioning.
Paraformaldehyde (PFA)	Sigma-Aldrich	1.04003	Used to fix cells prior to staining.
Penicillin/Streptomycin	Gibco	15140-122	Used to prevent bacterial contamination during cell culture.
Perfusion module (Pod)	Emulate	support@emulatebio.com	Used to drive controlled fluid flow through microfluidic chips.
Phosphate Buffered Saline (DPBS -/-)	Gibco	14190-094	Used for washing cells, preparing extracellular matrix solutions, and performing washing steps during staining.
Pic 21 Microcentrifuge	Thermo Fisher Scientific	75002415	Used to pellet RNA and separate phases during extraction.
PneumaCult ALI medium (epithelial differentiation medium) kit	StemCell Technologies	5002	Used to support differentiation of airway epithelial cells at air-liquid interface.
PneumaCult ExPlus medium (airway epithelial expansion medium) kit	StemCell Technologies	5041	Used for expansion of airway epithelial cells.
Razor blade	Thermo Fisher Scientific	11904325	Used to cut  the chip.
RNA-free Glycogen	Thermo Fisher Scientific	R0551	Used as a carrier to enhance RNA precipitation.
RNAse ZAP	Invitrogen	AM9780	Used to decontaminate surfaces from RNases.
RNase-free water	Thermo Fisher Scientific	R0551	Used to resuspend purified RNA.
Steriflip-HV, 0.45 µm, PVDF, sterilized	EMD Millipore	SE1M003M00	Used to filter and degas medium prior to use.
Stretchble chip (Chip-S1)	Emulate	support@emulatebio.com	Used to culture cells under dynamic mechanical stimulation within a microfluidic environment.
Superfrost TMPlus Adhesion Microscope Slides	Epredia	J1800AMNZ	Used to mount and retain chip sections during staining.
Synthetic light-stable retinoic acid (EC-23)	Tocris	4011	Used to promote differentiation of airway epithelial cells.
T25 flask	Falcon	353108	Used to culture and expand airway epithelial cells.
T75 flask	Falcon	353136	Used to culture and expand pulmonary microvascular endothelial cells.
Triton X-100	Sigma-Aldrich	T8787	Used to permeabilize cells during immunostaining.
TRIzol (phenol-based RNA isolation reagent )	Invitrogen	15596026	Used to lyse cells and isolate RNA.
Trypan Blue	Invitrogen	T10282	Used to assess cell viability.
TrypLE Express (for epithelial cells)	Thermo Fisher	12604013	Used for enzymatic detachment of airway epithelial cells.
UV lamp	nailstar professional	NS-01-EU	Used to activate chip surfaces.
Vacuum set-up	/	/	Used to degas culture media prior to use.
Vascular endothelial growth factor (VEGF)	Gibco	PHC9394	Used to support endothelial cell growth and maintenance.
Vitamin C	Lonza	CC-4398	Used to support endothelial cell growth and maintenance.
Vortex	VWR	444-0996	Used to mix samples during RNA extraction.