Engineering a dynamic modelof the alveolar interface for thestudy of aerosol depositionRoberta Nossa,1,2 Joana Costa,1Ludovica Cacopardo,1 Arti Ahluwalia1,21Research Center “E. Piaggio”,University of Pisa, Pisa; 2Department ofInformation Engineering, University ofPisa, Pisa, Italy

AbstractNano and micro particles are widely

used in industrial, household and medicinalapplications. To understand the interactionbetween particles and epithelial cells, wedeveloped a dynamic model of the alveolarinterface. This system, named DALI(Dynamic Model for the ALveolarInterface), is a modular bioreactor com-posed of two chambers divided by a porousmembrane where epithelial lung cells areseeded. The membrane is the support of thealveolar barrier that separates the two com-partments of the alveolus: the air and bloodside. The system integrates the followingelements: i) Air/Liquid interface, thanks tothe two chambers divided by the mem-brane: ii) Cell culture media flow, thanks tothe presence of a peristaltic pump; iii) Lungbreathing motion, thanks to an airflow thatallows the stretching of the membrane; iv)Aerosol deposition system, to study theeffects of drug efficacy or particle toxicityon the epithelial layer; v) Quartz CrystalMicrobalance, to quantify the amount ofaerosolized particles.

IntroductionIn order to improve the physiological

relevance of the lung model and to investi-gate the deposition of aerosolised nanopar-ticles on the alveolar barrier, a bioreactorable to mimic breathing movements wasdesigned. The system, named DALI(Dynamic model for ALveolar Interface),consists of a commercial aerosol generator,two bioreactors with a moving membraneplaced between an air-liquid interface(Figure 1a), and a Quartz CrystalMicrobalance (QCM) to measure the effec-tive nanoparticle dose on the membrane(Figure 1c).

Materials and MethodsThe bioreactor is composed of two

polycarbonate cylindrical chambers (A andB in Figure 1b): the upper one for the airflow (height: 24 mm, diameter: 24 mm) andthe bottom one for the medium flow(height: 20 mm, diameter: 24 mm).Between them there is the porous stretch-able membrane fixed in a holder that con-sists of two annular magnets covered byPDMS (C in Figure 1b). The upper chamberis connected to an aerosol system fornanoparticles deposition (D in Figure 1b),and to a pressure regulator that ensures thecyclic stretching of the membrane. Theelectronics and the pressure regulators areplaced in a control box. Potentiometers onthe control box allow regulating the stretch-ing level of the membrane for mimickingdifferent stretching conditions.

The membrane was fabricated by elec-trospinning using a 1:1 (w/w)Bionate®:Gelatin solution. The Bionate® II80A, a commercial poly(carbonate)ure-thane copolymer, was used to replicate thebasement of the alveoli, as it guaranteesmembrane flexibility necessary to mimicthe cyclic motion during the breathing.Additionally, in order to increase cell adhe-sion, gelatin was used in combination withBionate® to obtain the final formulation forthe membrane. Cell adhesion and biocom-patibility were assessed by the Alamar Blueassay and cell staining with DAPI and rho-damine conjugated phalloidin.

A FEM model was used to simulatemembrane displacement due to the pressureimbalance on the two sides of the air/liquidinterface. The model was based on the FluidStructure Interaction module of ComsolMultiphysics 4.3a software and consists of acylindrical chamber connected to the exter-nal system with an inlet and an outlet tube(5 mm in diameter). The 80-μm thick mem-brane was modelled as a disk on the top ofthe bioreactor and undergoes a constantpressure from the top. The inlet velocitywas fixed at 100 μL/min, and the fluiddynamics solved in the Laminar Flowregime. Bioreactor walls were set as wallswith the no slip condition, and water chosenas a reference fluid. The FEM model wassolved for different pressures (1 to 15 kPawith a step of 1 kPa), in order to establishthe pressure at which the membrane dis-placement in z-direction is 7mm, corre-sponding to a linear distention of ≈ 17%,and so mimicking pathological levels ofstretching.1,2

ResultsIn the DALI, the membrane is placed

between annular magnets covered byPDMS. During static experiments, the hold-er with the membrane can be placed both ina 6-well multiwell, or inserted between thetop and bottom chambers. The bioreactor isclosed tightening wing nuts and the tight-ness of the bioreactor is ensured by the pres-ence of the membrane holder enclosed inPDMS, which is self-adhesive anddeformable. The bioreactor can be sterilizedby ethanol solution, gas plasma, or ultravio-let light.

During dynamic experiments, the baso-lateral chamber is connected to a commer-cial peristaltic pump with an inlet tube, andto the reservoir with an outlet tube (Figure1). The apical chamber is connected to theaerosol device and to a compressed air sys-tem, with an interposed pressure regulatorput inside a control box. Potentiometers onthe control box allow regulating the stretch-ing level of the membrane: 17% for mimicking differentstretching conditions, both physiologicaland pathological.

A FEM model was used to simulate thevelocity field of the fluid flowing throughthe bottom chamber, and to simulate themembrane displacement due to the pressureimbalance on the two sides of the air/liquidinterface. When the pressure in the upperchamber increases, the membrane stretches,moving into the basolateral chamber, until itreaches an equilibrium with the hydrody-

Correspondence: Roberta Nossa, ResearchCenter “E. Piaggio”, University of Pisa, Pisa,Italy.E-mail: roberta.nossa@ing.unipi.it

Key words: In vitro model; alveolar interface;aerosol; dynamic model; bioreactor.

Conference presentation: this paper was pre-sented at the Second Centro 3R AnnualMeeting - 3Rs in Italian Universities, 2019,June 20-21, University of Genoa, Italy.

Received for publication: 28 October 2019.Accepted for publication: 11 November 2019.

This work is licensed under a CreativeCommons Attribution NonCommercial 4.0License (CC BY-NC 4.0).

©Copyright: the Author(s), 2019Licensee PAGEPress, ItalyBiomedical Science and Engineering 2019; 3(s3):115doi:10.4081/bse.2019.115

[page 24] [Biomedical Science and Engineering 2019; 3(s3):115]

Biomedical Science and Engineering 2019; volume 3(s3):115

Non-c

omme

rcial

use o

nly

namic pressure of the flowing liquid. Withthe FEM model, it was possible to predictthe pressure ranges that the external systemmust apply in order to achieve the desiredstretching field on the membrane. Theapplied pressure corresponding to a z-dis-placement of almost 7 mm is 14 kPa. Thisdisplacement corresponds to a linear disten-

tion of ≈ 17%, and so mimicking patholog-ical levels of stretching.1,2

ConclusionsTo conclude, we present a bioreactor

that is able to replicate the cyclic motionduring the breathing. The flexible movingmembrane causes the rhythmic stimulationof epithelial cells, leading to the study of theinteraction between them and the particles,in a system that replicates in vivo condi-tions. The electrospun 1:1 (w/w)Bionate®:gelatin membrane has beenselected as suitable membrane for our appli-cation, as it is biocompatible and highlyflexible, allowing physiological deforma-tion levels. This study, based on the 3R’sStatement, paves the way towards thedevelopment of an actuation device forphysiologically relevant studies of aerosoland drug delivery and toxicology.

References 1. Patel HJ. Characterization and applica-

tion of dynamic in vitro models ofhuman airway. PhD thesis, Departmentof Biomedical Engineering, Utah StateUniversity, Logan, Utah.

2. Cavanaugh KJ, Margulies SS.Measurement of stretch-induced loss ofalveolar epithelial barrier integrity witha novel in vitro method. Am J PhysiolCell Physiol 2002;283:C1801-8.

[Biomedical Science and Engineering 2019; 3(s3):115] [page 25]

Article

Figure 1. a) Schematic representation of the DALI system (on the left) and its picture (onthe right); b) pictures of the component parts of the bioreactor; c) picture of the QCMdevice.

Non-c

omme

rcial

use o

nly