Comparison of two in vitro systems to assess cellular effects of nanoparticles-containing aerosols

Highlights

► A new VITROCELL – Pariboy system was evaluated for testing of aerosolized NPs. ► Deposition rates differed between marker compounds and NPs. ► The manual aerosolizer MicroSprayer was suitable for cytotoxicity testing of NPs. ► Polystyrene nanoparticles acted more cytotoxic as aerosols than as suspensions.

Abstract

Inhalation treatment with nanoparticle containing aerosols appears a promising new therapeutic option but new formulations have to be assessed for efficacy and toxicity. We evaluated the utility of a VITROCELL®6 PT-CF + PARI LC SPRINT® Baby Nebulizer (PARI BOY) system compared with a conventional MicroSprayer. A549 cells were cultured in the air–liquid interface, exposed to nanoparticle aerosols and characterized by measurement of transepithelial electrical resistance and staining for tight junction proteins. Deposition and distribution rates of polystyrene particles and of carbon nanotubes on the cells were assessed. In addition, cytotoxicity of aerosols containing polystyrene particles was compared with cytotoxicity of polystyrene particles in suspension tested in submersed cultures. Exposure by itself in both exposure systems did not damage the cells. Deposition rates of aerosolized polystyrene particles were about 700 times and that of carbon nanotubes about 4 times higher in the MicroSprayer than in the VITROCELL®6 PT-CF system. Cytotoxicity of amine-functionalized polystyrene nanoparticles was significantly higher when applied as an aerosol on cell cultured in air–liquid interface culture compared with nanoparticle suspensions tested in submersed culture. The higher cytotoxicity of aerosolized nanoparticles underscores the importance of relevant exposure systems.

1. Introduction

Inhalation is considered a suitable route for both topical and systemic pharmaceutical applications. Asthma, chronic obstructive pulmonary disease and pulmonary infections are targets for topic inhalation treatment. In addition, inhalation may also be appropriate to treat systemic diseases. Absorption by the lung is high since the alveolar surface is quite large (80–140 m²; (Weibel, 1963)) and the air–blood barrier (0.1–0.2 μm thick) is more permeable than other epithelial barriers. No other non-invasive application route provides the same systemic bioavailability and speed of action as inhalation. For therapeutic gene delivery via inhalation a lower risk of immunogenicity and toxicity has been reported in cystic fibrosis and alpha-1-trypsin deficiency compared to conventional viral vectors (Roy and Vij, 2010). Macromolecules for systemic inhalation treatment also include hormones, especially insulin, growth factors, different interleukins and heparin (Siekmeier and Scheuch, 2008). Using nanoparticle-based medication, a more efficient treatment of inflammation and mucus hypersecretion in asthma, chronic obstructive pulmonary disease and cystic fibrosis is expected. Nanoparticle-based medications also offer the possibility of increased mucus layer penetration since they can be designed with positive charge, better mucoadhesive properties, enhancers for drug absorption, mucolytic agents and compounds that open epithelial tight junctions. Using these tools an increased delivery of drugs in nanoparticle-based aerosol formulations is expected (Mansour et al., 2009). Physiological relevant testing of aerosols is needed to assess these nanoparticle formulations but established in vitro systems are rare and complicated to operate. In vivo systems face problems with interspecies differences in the morphology and physiology of the respiratory tract, with the ease of application and low deposition rates.

The relevant biological evaluation of nanoparticle-based medication requires a physiological exposure system, and deposition rates should be high enough to also enable cytotoxicity testing required for safety reasons. To assess cytotoxicity, in vitro studies are most commonly done on cell lines rather than on primary cells because cell lines yield more reproducible results. A549 cells are still the most commonly used cell line for cytotoxicity testing of nanoparticles (e.g., Akhtar et al., 2012; Lankoff et al., 2012; Stoehr et al., 2011), although tightness of intercellular junctions is lower than that of other cell lines derived from the respiratory system, such as H358, H596, H322 cells. The later cell lines, however, are used less often in pharmacological and toxicological testing because they are less well characterized.

To test aerosol exposure, respiratory cells are often exposed in submersed culture, although this does not reflect their normal physiological situation. More advanced in vitro exposure models use culture in the air–liquid interface (ALI) where cells are cultured on semi permeable membranes of a transwell insert. The insert is placed into a culture well, medium is supplied from the basal site only and cells are exposed to an aerosol at the apical part. Transwell cultures were first used for permeability studies of gastrointestinal cells, like Caco-2 cells, and later adapted to other cell types (Hidalgo et al., 1989). Several systems are available to expose transwell cultures to aerosols: the Voisin chamber (Voisin et al., 1977; Voisin and Wallaert, 1992), the Minucell system (Bitterle et al., 2006; Tippe et al., 2002), the Cultex system (Aufderheide and Mohr, 2000; Ritter et al., 2003) and the modified Cultex system, the VITROCELL system (Aufderheide and Mohr, 2004). These systems have been used for volatile organic compounds and carbon or cerium oxide nanoparticles in the atmosphere (Bakand et al., 2006; Bitterle et al., 2006; Gasser et al., 2009; Paur et al., 2008; Rothen-Rutishauser et al., 2009). For nanoparticle-containing aerosols the ALICE (air liquid interface exposure) system (Brandenberger et al., 2010a,b; Lenz et al., 2009) and the MicroSprayer has been used (Blank et al., 2006).

In this study, we evaluated a new test system based on the VITROCELL system by assessing the deposition rate of nanoparticle-containing aerosols in respiratory cells compared to a macromolecular reference substance. We were particularly interested in the suitability of this new system when using a nebulizer type also frequently used by patients. This VITROCELL based system was compared to a manual aerolizer, the MicroSprayer, which allows the direct application of aerosols to cells. Cellular effects observed by direct application of the aerosol to cells cultured in ALI were compared to those obtained by testing of nanoparticle suspension on cells cultured in submersed culture. These data can help to decide whether larger work and material efforts of aerosol exposure testing are justified.

For the evaluation of the system two particle types were used. Polystyrene particles, which can be obtained in reproducible form in different sizes and with different surface functionalization, were used as models for spherical particles. These particles have the additional advantage that they also allow the determination of the deposition rate in their core-fluorescently labeled form. Multi-walled carbon nanotubes with different diameters were used to identify the influence of shape and the aggregation behavior of the nanoparticles since carbon nanotubes show high aggregation and polystyrene particles low aggregation (Wiesner and Colvin, 2005). Carbon nanotubes are also candidates for numerous medical applications (Zhang et al., 2010).

2. Methods

2.1. Nanoparticles and reference substances

20, 40, 100, and 200 nm red (580/605) fluorescent labeled carboxyl-functionalized polystyrene particles (FluoSpheres) were purchased from Invitrogen (Vienna). Carboxylated short multi-walled carbon nanotubes (0.5–2 μm long, purity >95%) with outer diameters <8, 20–30 and >50 nm were obtained from CheapTubes Inc. (Brattleboro, Vermont). To identify a potential difference in cytotoxicity between exposure in submersed culture and exposure as aerosol, 20 nm amine-functionalized polystyrene particles were used (Estapor Microspheres, Merck Chimie S.A.S., Fontenay-sous-Bois). For exposure all nanoparticles were diluted and the suspensions were put into an Elmasonic S40 water bath (ultrasonic frequency: 37 kHz, Elma, Singen) for 20 min prior to all experiments. For the VITROCELL system and the MicroSprayer cells were cultured for 24 h after the exposures. Fluorescein sodium salt (Sigma Aldrich, Steinheim) was used as a macromolecular reference substance.

2.2. Physico-chemical characterization by photon correlation spectroscopy and transmission electron microscopy

Nanoparticle-sizes were determined routinely by dynamic light scattering using a Zetasizer Nano ZS (Malvern Instruments, Malvern) equipped with a 532 nm HeNe laser, taking into account viscosity as well as refraction index. Polystyrene particles were diluted with the same solvent used for the exposures (distilled water for VITROCELL PT/PARI BOY LC Sprint system and DMEM + 10% FBS for MicroSprayer) to a concentration of 200 μg/ml and sonicated for 20 min before measurements. To study the stability of the nanoparticles in the VITROCELL PT/PARI BOY LC Sprint system samples of the condensate from the vial at the end of the glass tube (Fig. 1a) were also tested. After equilibration of the sample solution to 25 °C, scattered light was detected at a 173° angle with laser attenuation and the dynamic fluctuations of light scattering intensity caused by Brownian motion of the particles was evaluated. Polydispersity Indices <0.2 are interpreted as indication for homogenous samples (Stancampiano et al., 2008).

Fig. 1.

Exposure systems. (a): The main compounds of the VITROCELL®6 PT/PARI BOY exposure system are indicated: PARI BOY® SX compressor connected by tubes to the PARI BOY LC SPRINT® Baby nebulizer produces the aerosol. The flow of the aerosol from the nebulizer via a glass tube to the three individual compartments of the exposure module until the collection of aerosol not delivered to the compartments in a plastic vial is indicated. Airflow is regulated by a vacuum pump (not seen) that aspirates air from each of the three compartments of the exposure module through plastic tubes. For temperature maintenance at 37 °C a water bath connected to the exposure unit by plastic tubing, is used. (b): The tube of the MicroSprayer IA-1C aerosolizer is attached with a clamp to a tripod and the position of the tip is fixated at a distance of 11 cm from the rim of the exposure well. The long flexible tube allows the activation of the syringe to generate the aerosol without displacing the aerosolizer tip.

The viscosities of the different solutions were 0.88 cP (water) and 0.94 cP (DMEM + 10% FBS). Additionally, the refraction indices of all investigated media were found to be around 1.33 (1.33 for water, 1.345 DMEM + 10% FBS. These values were very similar and showed no impact on the measured particle sizes.

CNTs were suspended in DMEM + 10% FBS at 1 mg/ml. Prior to sizing, the solution was put through another sonication cycle for dispersion and diluted to 1 μg/ml for measurement in the ZetaSizer. NNLS software was used for sample analysis.

The zeta potential was measured by Laser Doppler Velocimetry (LDV) coupled with Photon Correlation Spectroscopy using a Zetasizer Nano ZS (Malvern Instruments, Malvern). The experiments were conducted at 25 °C and a scattering angle of 17°. The Zetapotential was calculated out of the electrophoretic mobility by applying the Henry equation.

Although Photon Correlation Spectroscopy has its limitations for the assessment of fibrous particles it is an accepted technique to describe physicochemical parameters of CNTs in solvents (Ito et al., 2004; Lee et al., 2005). Hence, this method has also been used by several other groups for the characterization of CNTs for biological experiments (e.g., (Bhirde et al., 2010; Wang et al., 2011; Yang et al., 2012)). To verify this data by another independent method, CNTs were also characterized by transmission electron microscopy. The CNTs were dispersed in DMEM + 10% FBS at 1 mg/ml and treated with ultrasound for 20 min. Five Microlitre of this solution were placed on a carbon coated copper grid that had previously been treated with a Pelco EasyGlow glow discharge device (Ted Pella, Inc., Redding, CA). After 1 min incubation, the solution was withdrawn using non hardened microscopic filter paper (Whatman, VWR International). Images were taken using a FEI Tecnai G² 20 transmission electron microscope (FEI Eindhoven) with a Gatan ultrascan 1000 ccd camera. Acceleration voltage was 80 kV. Sizes of CNTs were measured from the TEM images.

2.3. Cell culture

A549 human lung adenocarcinoma cells (ATCC) were cultured in DMEM + 10% fetal bovine serum in 6-well multiwell plates with polycarbonate membrane transwells (ThinCerts, Greiner bio-one, Frickenhausen). Cells were seeded with 500,000 cells/well. Cells in transwells were cultured in both liquid, submersed culture (LCC, cell culture medium in apical and basal compartment) and air–liquid interface (ALI) (apical compartment air and basal compartment cell culture medium) at 37° C in a 95% air/5% CO₂ atmosphere. For the exposures in the VITROCELL/PARI BOY and in the MicroSprayer, cells were seeded, medium was removed after 24 h and cells were cultured for an additional 7–8 days prior to the exposures.

2.4. Immunocytochemistry

2.4.1. Staining of tight junction proteins

The expression of tight junction proteins in cells was studied by the immunocytochemical localization of zona occludens protein-1 (ZO-1) and claudin-1. E-cadherin was chosen as a representative protein that is present in adherent junctions.

Cells were fixed by incubation in 100% ethanol for 20 min, in 100% methanol for 2.5 min and in 1:1 ethanol/acetone 10 min at −20 °C. Thereafter, first antibodies and negative controls were added for 30 min at RT, followed by incubation with the secondary antibodies for 30 min at RT and counterstained with Hoechst 33342 for 15 min. Between all incubations, cells were rinsed three times for 5 min in PBS. The following antibodies and dilutions were used: ZO-1 (rabbit polyclonal antibody, 1:100, Invitrogen), Claudin-1 (mouse monoclonal antibody, 1:40, Invitrogen), E-Cadherin (mouse monoclonal antibody, 1:100, Biozym, Vienna), AlexaFluor546 labeled anti-mouse IgG (goat polyclonal antibody, 1:200 Invitrogen), AlexaFluor488 labeled anti-rabbit (goat polyclonal antibody, 1:400 Invitrogen), negative controls (rabbit IgG and mouse IgG, Linaris, Bellingen). Cells were mounted in fluorescence mounting medium and viewed at a LSM 510 Meta Laser Scan microscope (Zeiss, Vienna) with the following settings: 488 nm excitation wavelength using a BP 505–530 nm band-pass detection filter for AlexaFluor488 and 543 nm excitation wavelength in conjunction with a LP 560 nm long pass filter for the red channel (AlexaFluor546).

2.4.2. Cytoskeleton staining (Actin)

After exposure, cells were rinsed in PBS, fixed in 3.7% paraformaldehyde for 10 min at RT and washed (3 × 5 min) in PBS. Cells were permeabilized by incubation in acetone for 3 min at −20 °C and rinsed again. Cells were stained with 165 nM phalloidin AlexaFluor 488 (Invitrogen, 1:40 dilution of stock solution in methanol) for 20 min at RT in the dark, rinsed in PBS, counterstained by immersion in 1 μg/ml Hoechst 33342 (Invitrogen) in PBS for 10 min, rinsed again in PBS and mounted in fluorescence medium. Pictures were taken using a LSM 510 Meta with 488 nm excitation wavelength using a BP 505–530 nm band-pass detection filter.

2.5. Transepithelial electrical resistance (TEER)

The formation of tight junctions indicating healthy cell monolayers was studied by measuring the transepithelial electrical resistance. To follow the development of TEER cells were cultured for up to 18 days. 2 ml DMEM were added to the apical and 3 ml DMEM were added to the basal compartment for TEER measurement with a EVOM STX-2-electrode (World Precision Instruments, Berlin). Calculation of TEER:

$$\text{TEER}(\mathrm{\Omega }\ast {\text{cm}}^2)=\text{Sample}-\text{bank resistance}(\text{Transwell without cells})\ast \text{Membrane area}$$

2.6. Quantification of deposition and distribution rate

For deposition and distribution studies, solutions of 2 mg/ml and 200 μg/ml FluoSpheres (VITROCELL/PARI BOY) and 1 mg/ml (MicroSprayer) were aerosolized. A549 cells in transwells were exposed to these solutions for 1 h in the VITROCELL/PARI BOY or up to three doses in the MicroSprayer and cultured for additional 24 h.

To quantify deposition and distribution rates, cells were lysed by adding 10 μl of lysis solution (one part 70% ethanol + one part Triton X100 to 500 μl distilled water) for 10 min at 37 °C. Fluorescence was read at a FLUOstar optima (BMG) at 485/520 nm for fluorescein and at 584/612 nm for red FluoSpheres. Calculation of deposition:

$$\text{Deposition}(\%)=\frac{\text{Signal sample}\times \text{dilution}}{\text{Signal}(\text{nebulized solution})\times \text{dilution}\times \text{volume nebulized}}\times 100$$

To take into account a potential influence of the cell lysate, 10 μl cell lysate of non-exposed cells was also added to the stem solution sample used for aerosolization for the measurement. For the deposition of CNTs absorbance of the lysates was read at 360 nm using a SPECTRA MAX plus 384 photometer (Molecular Devices).

To test for differences between the three compartments in the VITROCELL/PARI BOY exposure unit the distribution rate was calculated according to the equation:

$$\text{Distribution}(\%)=\frac{\text{Signal}(\text{signal compartment})}{\text{Signal}(\text{all compartments})}\times 100$$

2.7. Cytotoxicity by formazan bioreduction (MTS assay)

As an indication for healthy monolayers of A549 cells, viability according to formazan bioreduction (MTS) was used. Formazan bioreduction by cellular dehydrogenases was assessed by CellTiter 96® Aqueous Non-Radioactive Cell Proliferation Assay (Promega, Mannheim) using a water-soluble tetrazolium salt according to the manufacturer’s instructions. In short, after the 24 h following the exposure of the cells to polystyrene particles, medium was removed for the submersed cultures. To all wells the combined MTS/PMS solution (200 μl + 1 ml medium) was added. Plates were incubated for 2 h at 37 °C in the cell incubator. Absorbance was read at 490 nm on a plate reader (SPECTRA MAX plus 384, Molecular Devices). To correct for absorbance by the polystyrene particles alone, the signal of MTS/PMS + particles (in the absence of cells) was subtracted. All values are referred to solvent-exposed cells as 100%.

For the evaluation of CNTs the MTS assay was performed in a slightly different protocol because pilot experiments showed that the absorbance of CNTs interfered with the MTS signal. Therefore, to ensure that the signal of formazan bioreduction was not influenced by the absorbance of CNTs, cells were washed three times with PBS at the end of the incubation with the CNTs. Subsequently the combined MTS/PMS solution (200 μl + 1 ml medium) was added to the wells and after formation of the formazan product (2 h at 37 °C) the supernatant was transferred to a new plate for the measurement.

2.8. VITROCELL® PT/PARI BOY LC Sprint system (Fig. 1a)

For the exposures the following parts of a commercial VITROCELL System (VITROCELL Systems GmbH, Waldkirch) were used: VITROCELL®6 PT-CF stainless steel exposure unit with three compartments for transwell inserts of a 6-well plate. The thermostat HAAKE C10 P5 (Thermoscientific, München) regulated the temperature in the exposure block and the vacuum pump N840 FT.18 (Neuberger GmbH, Freiburg) controlled the air flow through the exposure unit. This unit was connected to a PARI BOY® SX compressor (Pari GmbH, Starnberg) in combination with Pari LC Sprint Nebulizer Baby¹ for generation of the aerosol. This nebulizer has an output rate of 150 mg/min and a mass median diameter of 2.5 μm and a mass percentage below 5 μm of 82%.

Tubings were connected according to the pre-established protocol provided by VITROCELL. In pilot experiments, specific parameters (nebulizer type, tube types, temperature, velocity, solvent) were varied to optimize the deposition rate. The delivery of substances to cells was higher for Pari LC SPRINT baby nebulizer than for Pari LC SPRINT junior. The Pari LC SPRINT junior produced more aerosol, but a high fraction of this aerosol condensed on the glass tubes. Best deposition rates were obtained using the glass tube and not the steel tube. Steel tubes are recommended for dry materials and, apparently, are less suited for liquids due to a higher condensation rate. The system worked optimally at temperature between 21 and 25 °C, without external cooling or heating of the glass tube. All experiments were performed under a hood in an air-conditioned room (variations between 21 and 25 °C). Mass flow was varied between 1 ml/min and 10 ml/min with best deposition rates at 5 ml/min. Deposition rates of fluorescein at 1 ml/min and at 10 ml/min were 0.19–0.36% (3rd compartment – 1st compartment) and 0.38–0.42% (3rd compartment – 1st compartment) of the deposition at 5 ml/min, respectively. Aerosolization in a variety of solvents (distilled water, PBS, 0.9% saline, DMEM, DMEM + 2% FBS) did not cause morphological damage to the exposed cells. As nebulization in distilled water produced the highest deposition rates, this solvent was used for the exposures of polystyrene particles.

2.8.1. Aerosol exposure of cells

The established system used in all experiments worked with PariLC SPRINT baby, glass tube as inlet, at room temperature, with a flow rate of 5 ml/min and distilled water was used as solvent. For FluoSpheres an optimal deposition rate was seen at 200 μg/ml whereas, 50 and 500 μg/ml showed lower deposition rates. CNTs were assessed at 50 μg/ml. Cells were exposed for 1 h and a volume of 10 ml for FluoSpheres and 8 ml for CNTs was nebulized.

2.9. MicroSprayer® IA-1C exposure (Fig. 1b)

The MicroSprayer® IA-1C aerosolizer (PennCentury Inc., Wyndmoor, PA) consists of a thin, flexible, stainless steel tube measuring 0.64 mm in diameter and 50.8 cm in length attached to the light, hand-operated, high-pressure syringe FMJ-250. A unique patented atomizer at the very tip of the tube generates the aerosol with a mass median diameter of 16–22 μm (http://www.penncentury.com/products/IA_1C.php). The MicroSprayer was fixated at a distance of 11 cm between tip of the MicroSprayer and the rim of the 6-well plate. This distance was determined as optimal for a reproducible delivery of the aerosol. To deliver the aerosol in a reproducible way the syringe was actuated in one fast push. For safety reasons all exposures were performed in a HERAsafe® KS 9 clean bench (Thermo Scientific, Vienna) equipped with UPLA filters of both filter grades U15 and H14. Aerosols with the MicroSprayer were generated with the same solvent as the VITROCELL/PARI BOY system (distilled water, PBS, 0.9% saline, DMEM, DMEM + 2% FBS) but in addition allowed aerosolization of substances in DMEM + 10% FBS. The maximum concentration of particles, which could be aerosolized without clogging of the aerosolizer tip and the maximum number of spray doses, which did not result in a continuous liquid layer on top of the cells, were determined.

2.9.1. Aerosol exposure of cells

Polystyrene nanoparticles (1000 μg/ml suspended in DMEM + 10% FBS) and CNTs (500 μg/ml suspended in DMEM + 10% FBS) were applied in three spray doses (600 μl aerosol). For the exposures, transwells were transferred to another plate, the exposure plate, and subsequently replaced and cultured for additional 24 h.

2.10. Statistical analysis

Data from three to five independent experiments (number of experiments indicated in the figures) were subjected to statistical analysis and are represented as means ± SD data were analyzed with one-way analysis of variance (ANOVA) followed by Tukey-HSD post hoc test for multiple comparisons (SPSS 19 software). Results with p-values of less than 0.05 were considered to be statistically significant.

3. Results

3.1. Physicochemical characterization of nanoparticles

The size of all polystyrene particles was increased in DMEM + 10% FBS compared with distilled water (Table 1). The size increase of the amine-functionalized particles was larger than that of the carboxyl-functionalized particles and the size of smaller particles increased more than that of the larger particles. Sample heterogeneity for carboxyl-functionalized polystyrene particles, measured with the polydispersity index, was higher in DMEM + 10% FBS than in water, indicating a greater tendency for aggregate formation in protein-containing medium. The opposite trends were seen for CNTs, in distilled water aggregates predominated and the polydispersity index was high, whereas in DMEM + 10% FBS sizes were much smaller and the polydispersity index lower. Zeta-potential values of carboxyl- functionalized polystyrene particles were negative when suspended in distilled water and positive for amine-functionalized ones. When suspended in DMEM + 10% FBS zeta-potential values of both polystyrene particle types were close to neutral. Zeta-potential values of CNTs in distilled water and in DMEM + 10% FBS were in the slightly negative range. Transmission electron microscopical analysis showed that all CNTs were shorter than indicated by the producer with maximum length of 450 nm. CNT8, CNT20 and CNT50 had diameters of 4.7 ± 0.48, 18.9 ± 0.9 and 62.8 ± 5.7 nm, respectively. To assess the influence of nebulization on the particles, 20 and 200 nm carboxyl-functionalized polystyrene particles were also characterized in aerosols collected at the end of the glass tube. In addition to agglomerates predominant peaks at 46 nm for the 20 nm polystyrene particles and 234 nm for the 200 nm polystyrene particles were recorded, suggesting that the particles are stable in the aerosol.

Table 1.

Solvent-dependent changes in the physico-chemical parameters of carboxyl (CPS) and amine (AMI) functionalized polystyrene particles and carbon nanotubes (CNTs). Mean from two experiments of dynamic light scattering (size) and electrophoretic light scattering (surface charge) is given.

Particles	Size (nm)		Polydispersity index		ζ-Potential (mV)
	Aqua dest.	DMEM, 10% FBS	Aqua dest.	DMEM, 10% FBS	Aqua dest.	DMEM, 10% FBS
CPS 20	42	80	0.09	0.157	−44.9	−8.3
CPS 40	86	127	0.138	0.20	−34.6	−11.6
CPS 100	159	173	0.04	0.21	−41.2	−9.6
CPS 200	245	320	0.145	0.57	−46.6	−11.9
AMI 20	21	54	0.495	0.384	40.1	−9.2
CNT8	4581	26	1.00	0.545	−15.6	−0.64
CNT20	1140	38	0.597	0.34	−10.7	−8.47
CNT50	885	50	0.449	0.442	−16.9	−11.0

3.2. Physiology and morphology of A549 cells in the exposure systems

Cells cultured in ALI had a slightly lower viability (85 ± 8%) than those cultured in submersed culture, which may be due to a lower hydration of cells in ALI culture. The viability of ALI cultured cells exposed to solvent without particles from the VITROCELL PT/PARI BOY system was 110 ± 10% of the non-exposed cells in ALI culture and similar to cells cultured in submersed culture. Viability of cells exposed to aerosols without nanoparticles generated by MicroSprayer was 112 ± 7% of the non-exposed cells in ALI culture.

TEER values were determined over two weeks to determine the stability of the ALI culture. Values increased during the first 13 days up to 230 ± 17.33  cm² and subsequently decreased from day 16 on (Fig. 2a); cells were routinely used after 7–8 days of culture. All tested proteins related to the formation of cellular junctions were detected in the A549 cells (Fig. 2b). Co-localization of ZO-1 and E-cadherin was seen at all-time points. Co-localization of ZO-1 and claudin-1 was absent after 24 h and weak at later time points.

Fig. 2.

Physiological and morphological characterization of not exposed A549 cells cultured in air–liquid interface (ALI) culture. (a): Development of the transepithelial electrical resistance (TEER) in A549 cells cultured in transwells. Means ± SD of n = 5 experiments are given. (b–e): Immunocytochemical detection of tight junction proteins after 7 days in ALI culture: Claudin-1 (red, b), ZO-1 (green, c) and E-cadherin (red, d) is seen between A549 cells cultured on transwells. Co-labeling of ZO-1 with E-cadherin (e) shows that the proteins are localized in different membrane areas. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.3. Deposition and distribution rate of nanoparticles

Both, VITROCELL PT/PARI BOY and MicroSprayer system showed a similar homogeneous deposition of nanoparticles on the A549 monolayer (Fig. 3, VITROCELL PT/PARI BOY system shown) but the deposition rate showed pronounced differences between the two systems (Table 2). Nanoparticle deposition rate with the VITROCELL/PARI BOY system was 0.038 ± 0.0068% for the 20 nm FluoSpheres and 0.029 ± 0.0073% for larger (40–200 nm) FluoSpheres (Fig. 4a). The total deposition rate in the VITROCELL/PARI BOY system was 6.3 times higher for the fluorescein reference compared with the 20 nm FluoSpheres and 7.9 times higher compared with the larger FluoSpheres (p < 0.001, Fig. 4a). The differences in deposition rate between small and large nanoparticles were not significant. CNTs show significantly higher deposition rates between 8.81 ± 0.92% for CNT8 and 1.25 ± 0.07% for CNT50 (Fig. 4b).

Fig. 3.

Distribution of fluorescently labeled (red) carboxyl-functionalized polystyrene particles (FluoSpheres, FS) on A549 cells 24 h after exposure in the VITROCELL®6 PT-CF/PARI BOY system. Cells are stained by phalloidin (green) and counterstained with Hoechst 33342 (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 2.

Overview on deposition rate maximum applied concentration and cytotoxic effect at that concentration in air–liquid interface culture (ALI) and in submersed culture (LCC) of carboxyl (CPS) and amine (AMI) functionalized polystyrene particles and carbon nanotubes (CNTs).

Exposure	VitroCell			MicroSprayer
	Deposition (%, n = 3) 1st–3rd compartment	Maximum concentration (μg/cm2)	Viability at maximum concentration (%, mean ± SD, n = 3), ALIa	Deposition (%, mean ± SD, n = 4)	Maximum concentration (μg/cm2)	Viability at maximum concentration (%, mean ± SD, n = 4), ALI	Viability at maximum concentration (%, mean ± SD, n = 4), LCC
Air	n.a.	n.a.	85 ± 8	n.a.	n.a.	85 ± 8	n.a.
Solvent	n.a.	n.a.	110 ± 10	n.a.	n.a.	112 ± 7	100
CPS 20	0.011–0.038	0.0007	98 ± 7	28 ± 1.96	148.67	98 ± 8	100 ± 7
CPS 40	0.012–0.036	0.0006	n.d.	28 ± 1.97	148.67	n.d	108 ± 1
CPS 100	0.009–0.032	0.0006	n.d.	28 ± 1.98	148.67	n.d	113 ± 4
CPS 200	0.003–0.021	0.0004	n.d.	28 ± 1.99	148.67	n.d	104 ± 1
AMI 20	n.a.b	0.0007	95 ± 8	n.a.b	89.07	2 ± 2	78.5 ± 3
CNT8	6.4–6.81	0.0241	99 ± 12	25 ± 2.5	66.37	93 ± 6	103 ± 2
CNT20	4.04–4.66	0.0165	n.d.	25 ± 2.5	66.37	n.d	97 ± 5
CNT50	0.81–1.66	0.0059	n.d.	25 ± 2.5	66.37	n.d	97 ± 12

Abbreviations: n.a., not applicable; n.d., not determined.

^aNo LCC data are given because at all concentrations no cytotoxic effect is expected.

^bDetermination not possible because the particles are not colored and not fluorescent.

Fig. 4.

Deposition rate (% of nanoparticles in the respective compartment referred to total aerosolized nanoparticles) and distribution rate (% of nanoparticles in the respective compartment normalized to the total amount deposited) of fluorescein, fluorescently-labeled 20 nm and larger (40–200 nm) carboxyl-functionalized polystyrene particles (FluoSpheres, FS, a) and carbon nanotubes of <8 nm (CNT8), 20–30 nm (CNT20) and >50 nm (CNT50, b) in the VITROCELL®6 PT-CF/PARI BOY system. Data are presented as means ± SD of n = 3 experiments. The three chambers of the VITROCELL®6 PT-CF exposure unit are termed as 1st, 2nd, and 3rd compartment. Significant differences in the deposition and distribution of the respective compartment to the 1st compartment are indicated by an asterisk and significant differences to the respective values in the 2nd compartment by a hash (p < 0.05).

The deposition rate with the MicroSprayer showed minimal variation between the reference substance and FluoSpheres of all sizes: fluorescein was deposited with an efficacy of 27 ± 3%, FluoSpheres with an efficacy of 28 ± 1.96% and CNTs with an efficacy of 25 ± 2.5%.

Significant differences in the distribution rate among the three VITROCELL/PARI BOY compartments were noted for both nanoparticles and fluorescein but the differences were larger for the polystyrene particle-containing aerosols (55%, 31%, 14%) compared with the reference substance (39%, 32%, 29%). No significant difference in the distribution among the compartments was seen for the tested CNTs.

3.4. Cytotoxicity testing with the MicroSprayer

Cytotoxicity testing was performed with the MicroSprayer, where higher concentrations could be applied. Cytotoxicity of aerosolized amine-functionalized polystyrene in ALI cultured cells at all doses was significantly higher than that of nanoparticle suspensions in submersed cells (Fig. 5, p < 0.05). Significant cytotoxicity was detected as low as a 31 μg/cm² of aerosolized nanoparticles in ALI cultures. In submersed (LCC) culture exposed to a nanoparticle suspension a significant cytotoxic effect was detected only at a concentration of 62 μg/cm² (Fig. 5, p < 0.05). At 62 μg/cm² TEER values were significantly decreased and staining against ZO-1 showed a diffuse cytoplasmic pattern. For CPS 20 and CNT8 no significant reduction in viability at the maximum concentration, which could be applied, was seen. The other CPS and CNTs were not evaluated for cytotoxicity because they were less toxic in submersed culture than CPS 20 and CNT8.

Fig. 5.

Cytotoxicity of amine-functionalized polystyrene particles in A549 cells assessed by the MTS assay. The particles were tested either in suspension on submersed culture (liquid culture, LCC) or as aerosol on ALI cultures with the MicroSprayer. Concentrations are indicated as μg/cm² and are based on a deposition rate of 29% as determined in previous experiments. Viability of solvent exposed cells is set as 100%. Data are presented as means ± SD of n = 4 experiments and significant differences (p < 0.05) are indicated by an asterisk.

4. Discussion

In this study, A549 cells cultured in ALI were exposed to aerosols for a more physiological testing of therapeutic nanoparticle containing aerosols. In this culture system the use of a new combination of test systems (VITROCELL + PARI BOY) was compared to a manual aerosol generating system, the MicroSprayer. PARI LC SPRINT nebulizers are commonly used for inhalation treatment by patients and the aerosol composition used in our study is therefore similar to that inhaled by patients. With both aerosol-generating systems the amount of nanoparticles, which could be applied to cells was limited and concentration, where cytotoxicity was expected based on conventional testing in suspensions, were only reached for amine-functionalized polystyrene particles. With these particles a significantly higher cytotoxicity was seen upon aerosol exposure than when applying nanoparticles in suspension.

The VITROCELL/PARI BOY system presented in this study allowed testing of nanoparticle based aerosols in a physiological exposure and without causing cell damage by the exposure system itself. The deposition rates of 0.175% for the reference substance and a maximum of 0.037% for aerosolized polystyrene particles in the VITROCELL/PARI BOY system are, however, lower than those of other existing systems. For instance, using the ALICE system (Lenz et al., 2009) for in vitro exposure 7.2% of the dose were delivered to an area of two 6-well plates (215.9 cm²). Cells cultured in an insert, therefore, would receive 0.157% of the total nebulized nanoparticle dose. Using a nose-only inhalation in mice only 0.008% of the nebulized dose reached the lung (Nadithe et al., 2003). Even upon instillation into the lung at the bifurcation of the trachea only 5% of the aerosol reaches the lung periphery where absorption can take place. In rabbits with tracheostoma, a model for the neonatal lung, deposition by nebulizers has been reported between 0.05% and 1.96% (Cameron et al., 1991; Flavin et al., 1986). Regarding the deposition rate of polystyrene particles, the MicroSprayer is much more efficient because the delivery rate is more than 700 times higher than the VITROCELL/PARI BOY system. For the assessment of conventional substances and polystyrene particles the VITROCELL/PARI BOY system also has the disadvantage that the deposition rate is not the same for all compartments of the system. The observed decrease in the deposition rate from the 1st to the 3rd compartment appears to be inherent to the system but affects aerosolized conventional substances and nanoparticles to different degrees. Taking the low absolute deposition rates of the polystyrene particles in this system and the sensitivity of the fluorescence plate reader into account, the significance and the relevance of the observed differences could, however, be questioned. For the evaluation of CNTs the differences between VITROCELL/PARI BOY system and MicroSprayer were less pronounced; the distribution between the compartments of the VITROCELL/PARI BOY was more homogeneous and the Microsprayer delivered only about 4 times more CNTs to the wells. CNTs have a much higher tendency to form aggregates (Lee et al., 2007) and, therefore, may form aggregates of appropriate size for sedimentation on the cells. The MicroSprayer delivers more aerosolized nanoparticles to the cells than the VITROCELL/PARI BOY system, which is important for cytotoxicity testing. On the other hand application with the MicroSprayer might damage cells by generation of shear stress because high flow rates are needed for effective particle deposition. Decreases in cell viability due to impaction of aerosols have been shown by Mühlhopt et al. (Mülhopt et al., 2007). Although adverse effects on cells cannot be excluded this study do not provide any indication for cell damage by using the MicroSprayer.

Both aerosol generating systems were assessed with respect to cytotoxicity testing. This assessment is an important first step in the toxicological assessment of compounds. Routine cytotoxicity testing, the exposure by addition of the test compounds to the medium above cells seeded in plastic wells (submersed culture), is not physiological for respiratory cells. It may lead to a sub-estimation of their cytotoxicity because a direct contact of the nanoparticles with the plasma membrane is not likely. Therefore, cells cultured in ALI and exposed to aerosols are recommended for physiologically relevant in vitro testing. This recommendation is supported by data showing the higher induction of the anti-oxidative enzyme HO-1 in A549 upon exposure to ZnO nanoparticles in ALI than in submersed culture (Lenz et al., 2009). The higher cytotoxicity of aerosolized polystyrene nanoparticles reported in this study also suggests a stronger effect upon aerosol application. It may be suspected that for nanoparticles with a greater tendency for aggregation, like CNTs, the exposure condition (aerosol or suspension) has a much smaller influence on the cytotoxicity.

For cytotoxicity testing, where high concentrations have to be tested to determine safety margins, the use of the MicroSprayer appears indicated because much higher doses than with the VITROCELL/PARI BOY system can be applied and the application itself did not cause adverse effects on cells. These data together with data from other groups (Fiegel et al., 2003; Knebel et al., 2001; Savi et al., 2008) show that higher aerosol delivery rates can only be obtained by a less physiological application mode. To assess the efficacy of aerosolized nanoparticles at therapeutic doses the VITROCELL/PARI BOY system appears better because it mimics better the low flow velocities in the alveoli. Providing every compartment with one nebulizer could decrease the differences in the deposition rates between the compartments.