Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 5.
Published in final edited form as: J Hazard Mater. 2020 Apr 13;395:122687. doi: 10.1016/j.jhazmat.2020.122687

Mimicking the human respiratory system: online in vitro cell exposure for toxicity assessment of welding fume aerosol

Ryan X Ward 1,†, Trevor B Tilly 1, Syeda Irsa Mazhar 1,2, Sarah E Robinson 3, Arantzazu Eiguren-Fernandez 4, Jun Wang 5, Tara Sabo-Attwood 3, Chang-Yu Wu 1,*
PMCID: PMC7276288  NIHMSID: NIHMS1586729  PMID: 32330784

Abstract

In assessing the biological impact of airborne particles in vitro, air-liquid interface (ALI) exposure chambers are increasingly preferred over classical submerged exposure techniques, albeit historically limited by their inability to deliver sufficient aerosolized dose. A novel ALI system, the Dosimetric Aerosol in Vitro Inhalation Device (DAVID), bioinspired by the human respiratory system, uses water-based condensation for highly efficient aerosol deposition to ALI cell culture. Here, welding fumes (well-studied and inherently toxic ultrafine particles) were used to assess the ability of DAVID to generate toxicological responses between differing welding conditions. After fume exposure, ALI-cultured cells showed reductions in viability that were both distinct between welding conditions and linearly dose-dependent with respect to exposure time; comparatively, submerged cell cultures ran in parallel did not show these trends across exposure levels. DAVID delivers a substantial dose in minutes (> 100 μg/cm2), making it preferable over previous ALI systems, which require hours of exposure to deliver sufficient dose, and over submerged techniques, which lack comparable physiological relevance. DAVID has the potential to provide the most accurate assessment of in vitro toxicity yet from the perspectives of physiological relevance to the human respiratory system and efficiency in collecting ultrafine aerosol common to hazardous exposure conditions.

Keywords: Air-liquid interface, Ultrafine particles, Condensation, Aerosol deposition

Graphical abstract

graphic file with name nihms-1586729-f0001.jpg

1. Introduction

Since the Industrial Revolution, human inhalation exposure to ultrafine particles (UFP, diameter < 100 nm) has significantly increased, and with the advent of and rapid developments in the field of nanotechnology, this exposure route will only continue to expand (Lioy et al., 2010; Oberdörster et al., 2005). When inhaled, airborne particles encounter numerous natural defense mechanisms along the respiratory tract; for example, mucociliary action can entrap and eliminate particles from the lung airways (Murgia et al., 2016). UFP are of critical concern compared to larger size modes because of their enhanced ability to evade these defense mechanisms, penetrate deep within the respiratory system, and ultimately deposit into the alveolar region in high number concentration, increasing their potential translocation through the cardiovascular system and to various parts of the body, such as the brain and liver (Kreyling et al., 2006; Oberdörster et al., 2004; Xing et al., 2016). Commonly, characteristic properties of UFP, such as their high surface area to mass ratio, high mobility within the body, and affinity to bond with and carry toxic gaseous compounds, are cited for adverse health risks associated with UFP (Frampton and Rich, 2016; Sturm, 2016).

While years of epidemiological evidence has clearly demonstrated the adverse health effects associated with the inhalation of airborne particles (Kreyling et al., 2004), current focus is geared towards the toxicological analysis of these particles, or the identification of the species and physicochemical properties of these species that generate the adverse health effects (Adami et al., 2011). The use of in vitro studies has greatly furthered our understanding of the interaction between UFP and living systems, but these studies are limited in their ability to depict accurately a physiologically relevant human exposure setting (Grass et al., 2010). Often times, in vitro analyses are conducted offline, which generally involves the collection of aerosols and dispersion of these particles into submerged cell culture (Lichtveld et al., 2012). In reality, aerosols interact dynamically with the human respiratory system, and their properties relating to number concentration, particle size distribution, agglomeration, and morphology are significantly altered by the change in phase from airborne to liquid-borne (Lewinski et al., 2017). For this reason, air-liquid interface (ALI) cell culture exposures are preferred and often considered a more physiologically relevant mechanism for exposure and toxicity analysis of UFP (Stone et al., 2017). These ALI methods allow for an online, real-time exposure of aerosols to cell culture compared to the previously standard offline methods mentioned; however, these systems are much more complicated and not yet well established.

While many technologies exist currently for ALI exposure, such as gravitational settling and electrostatic precipitation (Fujitani et al., 2015; Zavala et al., 2014), a trade-off is often observed between deposition efficiency to ALI cell culture and physiological accuracy. The ideal type of ALI exposure system would be one that most accurately reflects the conditions present in the human lungs, namely a relative humidity (RH) greater than 95% and a temperature of approximately 37 °C which induces the water-based condensational growth that particles experience within the respiratory system (Ferron et al., 1985; Longest et al., 2008; Zavala et al., 2017). In 2016, a laminar-flow water-based condensation device was shown to efficiently collect fine and ultrafine virus aerosol from ambient air (Pan et al., 2016), and has since been adapted for ALI exposure and termed DAVID, the Dosimetric Aerosol in Vitro Inhalation Device (Tilly et al., 2019). Employing the laminar flow condensation mechanism proposed by Hering and Stolzenberg (2005), DAVID could precisely control temperature and RH to activate aerosol for condensational growth, enhancing inertial deposition of aerosol particles to ALI cell culture. DAVID maintained its exposure chamber at state conditions akin to the human body, which reduced exposure-chamber-related stresses and allowed for a more accurate representation of respiratory system exposures. Additionally, under operating exposure conditions, DAVID was shown to maintain cell viability at greater than 97% for airflow exposure periods up to 20 minutes, which demonstrated its potential to maintain viable cell culture and separate the toxicological effects of specific aerosols from inherent viability reduction due to the nature of the exposure system.

The present study aims to provide the first application of DAVID in an ALI toxicity assessment through analysis of welding fume aerosol. Commonly studied due to its nano-scale size, high number concentration, and known respiratory and carcinogenic effects, welding fume is a major occupational exposure risk (Keane et al., 2016; Kim et al., 2005; Li et al., 2015; Zimmer and Biswas, 2001). Previously, the toxicities of both welding fumes and of welding fumes mitigated for toxicity using an amorphous silica precursor (tetramethylsilane - TMS) were demonstrated through offline toxicological analysis of E. coli biotoxicity (Yu et al., 2011). The present study aims to confirm this result through a proof-of-concept analysis via DAVID, in turn validating the ability of DAVID both to generate distinct toxicological responses from different aerosol exposures and to provide a more physiologically relevant ALI exposure.

2. Materials and Methodology

2.1. Experimental Design

The full exposure and analysis is composed of multiple simultaneously operated pieces, and an overview of the experimental set-up is shown in Figure 1. This set-up can be broken down into four primary components: welding fume generation, particle sizing, DAVID exposure, and filter collection. Each of these processes is herein described. In short, welding fume aerosol was generated and drawn into three parallel flows for online sizing, online exposure to cell culture through DAVID, and select collection on filters for gravimetric analysis. After the exposure, cell cultures were analyzed for particle toxicity through a viability assay. For comparison, further toxicity analysis was performed on liquid submerged cell cultures, following previously established standard protocol for in vitro toxicity analysis.

Figure 1.

Figure 1.

Open in a new tab

Experimental schematic diagram for exposure of welding fumes. Tetramethylsilane (TMS) was infused at the welding torch through a modified torch tip, allowing for the amorphous silica encapsulation. Excess fumes were ventilated through a high-volume pump as sample fumes were drawn from the same location through separate lines for particle sizing (SMPS), filter collection, and online ALI exposure (DAVID).

2.2. Welding Fume Generation

Generation and sampling of welding fumes were performed based on the American Welding Society (AWS) methodology for the controlled measurement and assessment of welding emissions (AWS, 2006). A conical chamber was used for the controlled fume generation environment, shown in Figure 1. At the base, the chamber measured 91.4 cm in diameter, and at the top the chamber measured 20.3 cm in diameter. The chamber itself rested on top of five symmetrically distributed bricks in contact with the floor, and a high-volume pump (General Metal Works GL-2000H, Cleves, OH) evacuated fumes at 40 liters per minute (LPM) from the top of the chamber during all welding events and for at least 10 seconds (s) once welding finished.

While the welding machine (Lincoln Power MIG 140C, Cleveland, OH) was exterior to the chamber (see Fig. 1), welding and subsequent fume generation both occurred interior to the chamber. The metal inert gas (MIG) welding process fed wire continuously to a mild steel base plate at a set voltage of 12.5 volts (V) and a wire feed rate of 2 meters per minute. The MIG process used ER 308L stainless steel welding wire (Harris Corporation, Melbourne, FL), which has the chemical composition: 19.5 – 22.0% Chromium, 9.0 – 11.0% Nickel, 1.0 – 2.5% Manganese, 0.30 – 0.65% Silica, and 0.03% Carbon by weight. The base metal plate was fixed to a rotating turntable (MK Products Aircrafter T-25, Irvine, CA), which doubled as the ground for the MIG process. A 25% CO2 and 75% Argon (Ar) mixture served as the primary shielding gas.

The welding gun was fixed to a stand outside of the chamber and extended through a fitted cut in the chamber wall. The gun was extended over the base plate and the tip-to-plate distance was fixed at 1.3 cm. The trigger of the welding gun was locked open, which allowed for continuous welding at a fixed height without the need for manual operation of the MIG process. Due to a plate-warping effect, noted after extended weld-time initial trials, continuously operated welding was limited to two minutes, wherein the welding machine would be shut off, the chamber allowed to ventilate, and the welding machine re-started.

Two variants of welding were employed, differing in the use of a TMS additive to the shielding gas. For the case without the TMS additive, welding proceeded as described. For the case with the TMS additive, the upper branch of Figure 1 was additionally engaged in the welding protocol. A 100% Ar carrier gas was passed through a Teflon impinger (Apex Instruments T507G, Fuquay-Varina, NC) containing liquid TMS. This impinger was submerged into an ice bath to lower the vapor pressure, as TMS rapidly volatilizes at room temperature. A mass flow controller (Omega FMA5500, Stamford, CT) was used to regulate the TMS-carrier gas mixture flow rate. A modified torch tip was used to mix the shielding gas and TMS-carrier gas (Wang et al., 2014).

2.3. Particle Sizing

High spatial and temporal variation exists when sampling welding fumes for sizing purposes, and protocols in sampling are further complicated due to high number concentrations and high temperatures (Biswas and Thimsen, 2011; Zimmer and Biswas, 2001), necessitating sample dilution. A size distinction between the welding conditions with and without TMS (due to coating of and agglomerating between particles) has been shown through transient sampling of welding fumes (Topham et al., 2011); however, a real-time steady-state particle size distribution, useful for aerosol statistics relating to the human respiratory system deposition, is employed here.

Particles were sized by a scanning mobility particle sizer (SMPS), which employed an aerosol classifier (Model 3080, TSI Inc., Shoreview, MN), a differential mobility analyzer (DMA, Model 3081, TSI Inc., Shoreview, MN), and a condensation particle counter (CPC, Model 3010, TSI Inc., Shoreview, MN). The sheath flow was set to 4 LPM and the size bins ranged from 14.1 nm to 478.3 nm. Number concentration size distributions were recorded and differences in geometric mean particle size for the TMS and non-TMS cases were assessed by one-way ANOVA. Particles were diluted before particle sizing to measure particle size distribution concurrently with fume generation. A high temperature dilution probe developed by Biswas and Thimsen (2011) was used to achieve dilution. This method used compressed air to dilute welding fumes, and the dilution ratio (set to 20:1) was controlled by the regulator on the compressed air cylinder.

2.4. Filter Collection and Gravimetric Analysis

During welding, fumes were pulled at 4 LPM across a 47-mm diameter glass microfiber filter (GF/A 47 mm 1820–057, Whatman, Maidstone, Kent, UK), matching the flow rate and exposure length experienced by welding fume aerosol within DAVID. Filters were conditioned in an anhydrous calcium sulfate desiccator and weighed (MC 210 S, Sartorius AG, Goettingen, Germany) pre- and post-collections to determine fume mass deposition onto the filter.

2.5. Exposure System & Online Biological Protocol

Details about the original DAVID, including calibration, optimization, and schematic drawings, have been described by Tilly et al. (2019); as such, only a brief overview is provided here. DAVID consists of eight parallel growth tubes which pass particles through three temperature-controlled stages: the conditioner (6 °C), where the aerosol is cooled, the initiator (45 °C), where condensation of water vapor onto the aerosol surface occurs, and the exposure chamber (37 °C), where aerosols are impacted into cell culture (shown in Figure 2). The condensational growth mechanism allows particles as low as 5 nm to grow to a sufficient size for efficient inertial impaction to cell culture at the base of the growth tube (Lednicky et al., 2016; Tilly et al., 2019) and conditions aerosols to temperature and humidity conditions found in the lungs and for minimization of extracellular stresses unto the cell cultures. Custom, 3D printed cell culture platforms were fabricated in house to allow for simultaneous exposure of four cell cultures, seen in Figure 2. Polylactic acid (PLA) filament was employed by the 3D printer to create the exposure platforms and individual cell culture wells, polyethylene terephthalate (PET) membrane was treated by O2 plasma and used at the base of the wells for cell culture, and silicon-based glue was used to join membranes and 3D printed materials. Cell cultures were maintained at the base of DAVID in the cell exposure chamber, and this section of DAVID was maintained at 37 °C to mimic the temperature in vivo. Noteworthy in Figure 2 is the presence of aerosol deposition sites (circled in the figure) outside of the wells containing cell culture; i.e., the current design only uses four of the eight growth tube and exposure channel combinations to impact samples to cell cultures, due to constraints in internal spacing. Future designs should modify this configuration to utilize all eight channels.

Figure 2.

Figure 2.

Open in a new tab

(a) Simplified schematic view of DAVID’s condensational growth stages for one of the four exposure channels (seen in Figure 2(c) labeled ‘Well’). Polydisperse aerosols enter DAVID and are grown (in high supersaturation) to a larger and more monodisperse size distribution for impaction to cell culture. (b) Particle size shift due to condensational growth. (c) Top view of the experimental exposure platform containing four removable wells. The wells are situated atop media and separated from the media via a membrane, creating the ALI environment upon which cell cultures rest. Each well is situated under three jets for aerosol deposition, creating highly localized particle deposition sites, three examples of which are circled. Note that not all depositions are directly in the cell cultures, as there are eight total growth tubes but only four exposure channels in the current design.

In this study, a modified and enhanced DAVID compared to the original described by Tilly et al. (2019) was used. The enhanced DAVID had a moderator (30 °C) after the initiator to minimize water accumulation during exposure by removing excess water vapor from the air. Furthermore, the wetting of growth tube wicks was changed from capillary force wetting to direct-injection pumps to overcome the drying issues at high flow rates and/or long exposure times and allow for greater control of the growth tube environment.

Thirteen different exposure times (between two and eight minutes) were used for both the TMS and non-TMS exposures. Welding fume aerosols were drawn into DAVID through a line at 4 LPM, a flow rate which was shown to demonstrate negligible intrinsic toxicity to cell culture (Tilly et al., 2019). Four cell cultures were exposed simultaneously at the ALI using the 3D printed platforms. After exposure, cells were returned to the incubator and viability was assessed 24 hours after exposure through lactate dehydrogenase (LDH; Thermo Scientific, Waltham, MA) and deep blue (DB; Biolegend, San Diego, CA) assays. Between the TMS and non-TMS exposures, linear regression analysis was used to compare differences in viability for each exposure time length.

The A549 alveolar-derived epithelial-like cell line (ATCC, Manassas, VA), a commonly studied cell line (Paur et al., 2011), was cultured and used for all exposures. Cells were maintained with glutamate-free RPMI 1640 base media supplemented with 10% fetal bovine serum (Heat Inactivated FBS, Gibco, Carlsbad, CA), 100-fold penicillin-streptomycin (Pen: 100 U/mL, Strep: 100 μg/mL, Gibco, Carlsbad, CA), and 100-fold dilution of GlutaMAX Supplement (35050061, Gibco, Carlsbad, CA). Once the cells were cultured to at least 70% confluence, they were treated with trypsin-EDTA (0.25% Trypsin-EDTA, with phenol red, Gibco, Carlsbad, CA) to detach from the flask, counted by hemocytometer, and seeded at a concentration of 165,000 cells/cm2. The medium was replaced after three days, at which point apical side medium was removed and prepared for exposure in DAVID at the ALI. Once exposed, cells were incubated for 24-hours for the LDH assay.

2.6. Offline Biological Protocol

Beyond the online exposure and toxicity assessment through DAVID, an offline analysis was performed through classical, submerged cell culture protocol. Welding fume particles were sampled by 35 mm polycarbonate Cyclopore track-etched membrane filters with a 0.2 μm pore diameter (Whatman, Maidstone, Kent, UK), which were mounted in an open-faced cassette (Millipore #XX50 057 10, MilliporeSigma, Burlington, MA); sampling flow rate of 4 LPM was controlled by a rotameter (FLDA3216ST and FT-054–01-ST-VN, Omega Engineering Inc., Norwalk, CT) and a vacuum pump. After particle collection, the mass of particles collected was assessed gravimetrically on a scale (Sartorius MC210S, Goettingen, Germany; readability 10 μg). The particles were then extracted from each filter in 5 mL of ultrapure water contained in 50 mL conical tubes by five ultrasonication cycles of 10 s in a bath sonicator (M8800, Branson Ultrasonics Corp., Danbury, CT). An equivalent volume of ultrapure water was added to each conical tube to create a final stock concentration of 2 mg/mL. This methodology for particle extraction from filters is well established, having previously employed various immersion fluids (diluted water, phosphate-buffered saline, or RPMI cell media) and various particle extraction methods (brushing or eluting filters) (Antonini et al., 2005, 1999; Mcneilly et al., 2004).

For submerged cell culture exposure, confluent cells in 96-well plates were exposed in groups of six to various dosing solutions. The dosing solutions were prepared by diluting the original 2 mg/mL stock solution in the supplemented cell growth medium to concentrations ranging between 1 and 500 μg/mL. The stock and dosing solutions were treated by a vortex mixer (Digital Vortex Mixture, Fisher Scientific, Waltham, MA) and were freshly prepared on the day of the exposure to prevent time-dependent dissolution that could alter the toxicity outcome. 100 μL of dosing solution was administered to each well through a multichannel pipette, and cell cultures were incubated for 24 h. After incubation, cell viability of the cells exposed to welding fume and that of the untreated controls was assessed via LDH and DB assays.

2.7. Calculation of Mass Dose

While exposure length can be linked to decreases in cell viability, a useful metric for comparison of ALI studies is the mass deposition onto the cell culture. Deposition of UFP to the cells is a function of their effective density and diameter (Cohen et al., 2013; Deloid et al., 2014), which can be used for estimation of the delivered dose. This mass deposition can be approximated indirectly through statistics of aerosol measurements as:

Mass Dose=πdp3ρpCpQefft6Awell (1)

where dp is the diameter corresponding to the average volume of the particle (171 nm and 187 nm for the cases without and with TMS, respectively), ρp is the density of the particles, Cp is the time-weighted average number concentration of the particles, Qeff is the effective flow rate onto a single cell culture (or the flow rate through one of DAVID’s growth tubes), t is the exposure duration, and Awell is the surface area of a single cell culture well (0.36 cm2; see Figure 2). From SMPS measurements, the number concentration (adjusted for the dilution ratio) and diameter values were recorded. The effective flow rate through a single growth tube was 0.5 LPM, calculated as the total flow rate through DAVID (4 LPM) divided by the number of growth tubes (8 total). In accordance with previously reported values, the density of welding fume was 5.9 g/cm3 for the welding without TMS and 3.31 g/cm3 for the case with TMS (a mass weighted average employing density = 2.3 g/cm3 for amorphous silica fume and 1:2.33 ratio welding fume to silica) (Hewett, 1995; Muller et al., 2015). Consequently, between the two welding conditions, only number concentration and particle diameter are unique parameters, and when held constant within each condition, calculated mass dose becomes a linear function dependent only on time. DAVID has previously been shown to effectively capture > 98% of influent particles, negating any need to correct for an over-assumption (Pan et al., 2016).

As a means of providing a physical approximation for this deposition, filters collected welding fumes for select exposures for the same length at the same flow rate (Section 2.4). By normalizing the mass of the filter to the size of the area of particle collection on each filter (roughly 77% of the total filter area), the resulting mass deposition over time was calculated as a surrogate for the amount of mass passing through DAVID and ultimately deposited per unit area.

3. Results and Discussion

3.1. Aerosol Size Distribution

Presented in Figure 3 are the dilution-factor-adjusted particle size distributions for both welding conditions. For the welding without TMS-infusion, the geometric mean diameter was 83.0 nm and the geometric standard deviation (GSD) was 3.00; for the case with the addition of TMS, the geometric mean diameter was 112 nm and the GSD was 1.79. The design of the diluter employed in this study maintains a short (< 1 s) residence time in the dilution chamber, effectively maintaining particle size distribution to accurately depict the in situ condition (Biswas and Thimsen, 2011).

Figure 3.

Figure 3.

Open in a new tab

Welding fume particle size distributions for the two welding conditions. The number concentration is shown against the particle diameter given by SMPS readings after a 20:1 dilution of the initial sample. The geometric mean diameter for the welding without TMS was 83.0 nm while the geometric mean with TMS was 112.4 nm. An upward shift in the distribution is demonstrated with the addition of the TMS to the shielding gas.

Previously, infusion of TMS into the shielding gas of the welding process has demonstrated an increase in particle size of over an order of magnitude, from 18 nm to median diameters > 100 nm (Topham et al., 2011). Particle sizing in that study involved a transient, non-steady-state method for sampling, wherein welding fumes were generated for 10 s, the welder was stopped, and sampling via SMPS was started. For calculation of the mass deposition (described later), however, it is ideal to have a steady-state measure of particle size, which is an integral component of the mass deposition calculation equation. This allows for the most accurate characterization of the welding fume aerosol in both DAVID and in the human respiratory system. Capturing such a particle size distribution, though, can be difficult due to the high spatial-temporal variability associated with the evolution of welding fumes in time and with dilution (Cena et al., 2016; Zimmer and Biswas, 2001). Despite the less drastic difference in particle size due to the coating and agglomerating effects of the TMS observed in this study compared to previous analyses (Wang et al., 2014), one-way ANOVA indicated that there is significant difference at the 95% CI (p < 0.05) between measurements of median particle diameter for the TMS and non-TMS cases. This difference illustrates the effectiveness of the amorphous silica in shifting the particle size distribution to a larger median particle size. This can be seen in the particle size distribution presented in Figure 3, which demonstrates a shift to larger particle sizes on a particle number concentration scale (dN/dlogdp) for the welding with TMS-infusion compared to the welding without TMS.

3.2. Toxicological Analysis in ALI Cell Culture

The results of the LDH viability analyses of the 13 different time-length ALI exposures are shown in Figure 4. The mean cell viability of the four cell cultures in each exposure is plotted against the length of each exposure, and error bars represent their standard error. After eight minutes, the maximum length of exposure, a vastly different result was observed between the cell cultures exposed to welding fumes alone and those exposed to welding fumes with the TMS additive. The cell culture exposed to welding fumes without TMS exhibited a 51% decrease in viability on average, and a roughly linear decrease in cellular viability was observed (R2 = 0.85) for measurements spanning 1.5 – 8 minutes. Previously, through in vitro submerged cell culture mechanisms, welding fumes were shown to exhibit dose-dependent, cytotoxic responses (Mcneilly et al., 2004). Additionally, DAVID was shown to efficiently deposit UFP to cell cultures with a linear increase as a function of time (Tilly et al., 2019). It can be inferred that welding fumes generate a toxic response that is roughly linear with respect to time through ALI techniques. Indeed, while there is some evidence for linear dose-responses by ALI cultures (Svensson et al., 2016), our result is one of the most robust assessments of the time-dependent dose, as many prior studies have explored the total mass dose for the exposure (Lichtveld et al., 2012; Paur et al., 2011) rather than temporal trends, which are potentially important to ALI studies as compared to submerged cell culture studies. Nonetheless, further studies must be performed to validate this pattern for longer exposure periods and for different biologic endpoints, as ALI studies of welding fumes have not yet been widely performed.

Figure 4.

Figure 4.

Open in a new tab

Viability of A549 Air-Liquid Interface cell culture, measured through a lactate dehydrogenase (LDH) assay, as a function of exposure length in DAVID. Error bars represent the standard error of four simultaneously exposed cell culture.

By comparison, cell culture exposed to welding fume with TMS exhibited only a 10% decrease in viability for the same maximum length of exposure (8 minutes). The toxic response was lower for all exposure lengths (R2 = 0.28), supporting previous evidence for the effectiveness of the silica coating in suppressing toxic responses (Yu et al., 2011); the low value of R2 further suggests almost no linear decline in viability over all exposure lengths tested in this study. The results of the linear regression, which suggest two differing toxic responses over time, demonstrate the ability of DAVID to effectively generate distinct toxicological responses to particles with different physicochemical properties in a dose-dependent manner. This expands the DAVID optimization results (Tilly et al., 2019), demonstrating its capabilities as an effective tool for ALI cell exposure to UFP for toxicity analysis. While our previous study focused on the design and optimization of the procedure for cell culture exposure, the toxicity results observed here are the first illustrating that DAVID is able to deliver sufficient mass in very short sampling periods and effectively generate a toxic response to aerosol (> 50% reduction in viability) in a short period of time (8 minutes).

The DB assay for the ALI exposures did not have the same toxicity results, however, as the viability of the control and exposed cells were similar post-exposure. This may be attributed to the difference in mechanisms by which the DB assay endpoints respond to cell viability markers compared to the LDH assay. The localized sites of aerosol deposition onto the densely-seeded ALI membranes allows for the proliferation of many cells, whose mitochondrial activity likely caused the DB assay to saturate very quickly (measured by fluorescence) compared to the submerged cultures (discussed in Section 3.3).

Moving forward, two design parameters must be re-considered when analyzing ALI toxicity through DAVID. Firstly, the delivered particles to each cell culture well were localized to three spots per well due to the nozzle design in each of DAVID’s growth tubes (see Figure 2), which resulted in a nonhomogeneous distribution of deposited particles throughout the cell culture. Although analogous with deposition in the human respiratory tract, nonuniform deposition may influence the cellular response compared to particles homogenously distributed across the entire cell culture, as not all cells were exposed. At the same time, the ability to study the effects of this nonuniform distribution on intercellular responses (e.g. cell signaling) at the ALI has been made possible through DAVID. So, the heterogeneity of the deposition must be further considered. Secondly, the responses incurred by (1) this localized stress and (2) this exposure mechanism may not best be captured through viability assays. While only the LDH assay was shown to demonstrate a toxic response for ALI analysis, other assays which consider toxicity through differing pathways, such as oxidative stress, may yield results that are divergent from the present data. In vivo toxicity is systemic in that the immune system allows for responses throughout the body, and consequently a viability assay does not necessarily capture a systemic toxicological response.

3.3. Toxicological Analysis in Submerged Cell Culture

The DB analysis results of the submerged cell culture, shown in Figure 5, illustrate that there was no significant difference in viability between the welding fumes with and without TMS infusion for dosing concentrations less than 100 μg/mL. At exposure doses of 100 μg/mL and 250 μg/mL, the viability of cell cultures exposed to TMS-coated welding fumes was almost 30% higher than cultures exposed to the welding fumes without TMS coating. At ultra-high exposure concentrations (e.g. 500 μg/mL), this difference becomes indiscernible.

Figure 5.

Figure 5.

Open in a new tab

Viability of A549 submerged cell culture, measured through a Deep Blue (DB) assay, as a function of exposure concentration. Measured viability above an administered dose concentration of 100 μg/mL shows significant difference between the TMS-coasted and non-TMS-coated exposures, but this difference becomes indiscernible at ultra-high exposure concentrations (i.e. 500 μg/mL).

It is important to note that as the dosing concentration increases, the physiological relevance of the exposure substance tends to decrease; at high exposure concentrations, many substances will incur toxic responses in vitro regardless of the chemical species used for exposure. For instance, 500 μg/mL (the highest dose presented here) is already well beyond a relevant human exposure condition. At these high concentrations, the effects of the TMS-coating were outweighed by the high dosing concentrations. At these observed low viability rates (< 50% viability), it is difficult to attribute the toxic effects to the chemical nature of the exposure substance without considering the effects due to high particle concentration as a major contributor to the reduction in viability. Conversely, at low concentrations there was little to no evidence of the difference in viability generated by these two welding fume scenarios. In both extreme scenarios, the signal-to-noise ratio obscured the toxicological conclusions, differing from the results of the ALI exposure and serving to highlight the effectiveness of the DAVID system.

In the present study, the ALI cell cultures were notably more responsive to the dose delivered by DAVID, due in part to the high efficiency in delivering dose to cell culture. This sensitivity to delivered dose in ALI cell culture has been previously reported. For instance, Lichtveld et al. (2012) suggested that offline toxicological techniques modify the toxicity of airborne particles and underestimate their toxicity; furthermore, Lenz et al. (2012) demonstrated that for distinct ALI exposure techniques and exposure compounds, ALI cell cultures were more sensitive to delivered doses. Across exposures ranging one to eight minutes, DAVID was able to generate two unique responses from the cell culture at the ALI under differing welding conditions while maintaining the viability of cell culture above 50%, all over a short exposure period.

Although both the LDH and DB assays were performed, only the results of the DB analysis are presented, because no reduction in viability was observed through the LDH assays, contrasting the observable results of the ALI analysis. This is in line with prior studies using the LDH assay for the submerged cell culture (Han et al., 2011; Holder et al., 2012) that showed false outcomes of toxicity due to interference due to particle binding with the LDH.

3.4. Mass Deposition in DAVID

The SMPS-based measured mass doses are presented graphically as linear models in Figure 6. When considering 2 min and 8 min as the minimum and maximum exposure lengths (as was the experimental condition), deposition values ranged from 46 – 183 μg/cm2 for the case without TMS and 85 – 341 μg/cm2 for the fumes with TMS. Between the SMPS-based calculated mass deposition and the gravimetrically assessed deposition, parity exists in the relationships between the scatter points and linear functions for both the cases with and without TMS, demonstrating an approximate knowledge of the deposition to cell culture.

Figure 6.

Figure 6.

Open in a new tab

Comparison of mass deposition between SMPS-based aerosol measurements and gravimetric assessment. The SMPS-based estimates are illustrated in dashed lines, and the filter-based estimates are displayed as scatter points. There is a general agreement between both approaches, as the scatter points situate roughly around their corresponding linear mass deposition model.

Since Equation 1 is third order with respect to particle diameter, the determination of a standard particle diameter that is most physiological comparable to the human respiratory system is critical. Because of the high temporal and spatial variability of welding fume aerosols with respect to time, dilution, and sampling conditions (Cena et al., 2016; Zimmer et al., 2002), it is challenging to determine which condition is most reflective of the in situ condition. This problem is not only limited to welding fumes, as ultrafine particles have been shown to coagulate and change in particle size, affecting the size distribution significantly (Lee and Wu, 2005). Additionally, while the mechanisms of particle amplification and efficiency of particle deposition in DAVID have been explored (Tilly et al., 2019), the relationship between particle evolution in DAVID and the human respiratory system should be further assessed to understand the spatial-temporal condition at which the particle size distribution should be sampled. It should also be noted in the presented linear model that while time was the only model parameter changed, other parameters, such as number concentration and diameter, can be resolved with continuous SMPS measurements to create a model dependent on multi-variable inputs. Only time-weighted average values are presented in this work.

Discrepancies between SMPS data and filter measurements may result from the size range limitations of the SMPS during experimental operating conditions. There is clear truncation in both size distributions, and these upper diameters likely contribute more substantially to mass dose (as larger diameters often contribute to a majority of the mass distribution in an aerosol statistics). Future studies of deposited dose through aerosol statistics should consider a measured resolution which better represents the size range of the aerosol in question.

Regardless, the effectiveness of DAVID in capturing particles of the ultrafine size regime and creating a significant mass dose to ALI cell culture was confirmed (Lednicky et al., 2016; Tilly et al., 2019). In comparison to other systems, DAVID provides one of the greatest dose deliveries UFP as a function of exposure time. In part, this can be attributed to the fact that DAVID was adapted from a highly-efficient aerosol sampler. Previous ALI systems (Lenz et al., 2012; Zavala et al., 2014) have generally demonstrated significantly less aerosol deposition (two orders of magnitude with respect to deposited mass per unit area of culturable cells) over much larger time scales (hours vs minutes) as compared to DAVID. While there is precedent for the use of water-based condensation to facilitate particle concentration, these systems are bulky and limited in their ability to sample beyond a laboratory setting and resolve particle sizes below 30 nm (Gupta et al., 2004; Saarikoski et al., 2019; Stone et al., 2017) and not yet well-adapted for application to ALI cell deposition. By facilitating high particle deposition efficiency through water-based condensational growth and preconditioning the exposure environment to maintain viable cell culture, DAVID is applicable across a range of environmental exposure conditions and to an array of UFP, such as previously unrealized low ambient particle regimes. Furthermore, current ALI deposition techniques (e.g. electrostatic precipitation) are unrepresentative of human physiology, which hinders the relevance of the ALI exposure as compared to a liquid submerged exposure. Employing similar principles as the condensational growth mechanisms of the human respiratory system, DAVID demonstrates a conceptually superior, online method of ALI exposure and toxicity analysis compared to previous ALI devices that are less physiologically relevant, offline analyses (i.e. liquid submerged cell culture).

4. Conclusions

The novel ALI exposure chamber, DAVID, was demonstrated to deliver highly concentrated doses (> 100 μg/cm2) of UFP in short periods of time (minutes). Using water-based condensation to promote efficient delivery of particles to ALI cell culture, DAVID was shown to provide an online exposure for particles that is both comparable to the physiology of particle deposition in the respiratory system and highly efficient for laboratory studies of particle toxicology. By employing two types of welding, a baseline condition and a reduced toxicity condition through the administration of amorphous silica precursor to the welding fume’s shield gas flow, DAVID’s ability to provide unique, dose-dependent toxicological exposures was shown. Additionally, the application of the TMS precursor was shown to significantly mitigate the welding fume’s toxicity over a range of exposure lengths, consistent with previous investigations on this material application. In comparison to currently available ALI exposure chambers, DAVID provides a conceptually superior exposure platform for future studies in both the toxicity of nano-sized materials and the response of cell culture to these exposures.

Highlights.

  • Novel system for analyzing airborne particle toxicity, akin to respiratory system.

  • Demonstrate high deposition and temporal efficiency compared to current systems.

  • Evidence for welding fume toxicity mitigation using amorphous silica precursor.

Acknowledgments

The authors would like to thank Katherine Rankine, Lauren Judah, Mayuko Mizutani, Jiva Luthra, Ricardo Aleman, and Meg Simms for assistance in operating welding fume exposures and Dr. Pratim Biswas for the loan of the diluter used in this study.

Funding

This work was supported by the University of Florida’s University Scholars Program and the National Institutes of Health Grant numbers: R21AI123933 and 1R43ES030649-01.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Adami H, Berry SCL, Breckenridge CB, Smith LL, Swenberg JA, 2011. Toxicology and Epidemiology: Improving the Science with a Framework for Combining Toxicological and Epidemiological Evidence to Establish Causal Inference 122, 223–234. 10.1093/toxsci/kfr113 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Antonini JM, Lawryk NJ, Krishna Murthy GG, Brain JD, 1999. Effect of welding fume solubility on lung macrophage viability and function in vitro. J. Toxicol. Environ. Heal. - Part A 58, 343–363. 10.1080/009841099157205 [DOI] [PubMed] [Google Scholar]
  3. Antonini JM, Leonard SS, Roberts JR, Solano-Lopez C, Young SH, Shi X, Taylor MD, 2005. Effect of stainless steel manual metal arc welding fume on free radical production, DNA damage, and apoptosis induction. Mol. Cell. Biochem 279, 17–23. 10.1007/s11010-005-8211-6 [DOI] [PubMed] [Google Scholar]
  4. AWS, 2006. F1.2:2006 Laboratory method for measuring fume generation rates and total fume emission of welding and allied processes. Am. Weld. Soc [Google Scholar]
  5. Biswas P, Thimsen E, 2011. High Temperature Aerosols: Measurement and Deposition of Nanoparticle Films, in: Aerosol Measurement: Principles, Techniques, and Applications. Wiley, pp. 723–738. [Google Scholar]
  6. Cena LG, Chen BT, Keane MJ, 2016. Evolution of Welding-Fume Aerosols with Time and Distance from the Source. Weld. J 95, 280–285. [PMC free article] [PubMed] [Google Scholar]
  7. Cohen J, DeLoid G, Pyrgiotakis G, Demokritou P, 2013. Interactions of engineered nanomaterials in physiological media and implications for in vitro dosimetry. Nanotoxicology 417–431. 10.3109/17435390.2012.666576 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Deloid G, Cohen JM, Darrah T, Derk R, Rojanasakul L, Pyrgiotakis G, Wohlleben W, Demokritou P, 2014. Estimating the effective density of engineered nanomaterials for in vitro dosimetry. Nat. Commun 5, 1–10. 10.1038/ncomms4514 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ferron GA, Haider B, Kreyling WG, 1985. A method for the approximation of the relative humidity in the upper human airways. Bull. Math. Biol 47, 565–589. [DOI] [PubMed] [Google Scholar]
  10. Frampton MW, Rich DQ, 2016. Does Particle Size Matter? Ultrafine Particles and Hospital Visits in Eastern Europe. Am. J. Respir. Crit. Care Med 194, 1180–1182. 10.1164/rccm.201606-1164ED [DOI] [PubMed] [Google Scholar]
  11. Fujitani Y, Sugaya Y, Hashiguchi M, Furuyama A, 2015. Particle deposition efficiency at air-liquid interface of a cell exposure chamber. J. Aerosol Sci 81, 90–99. 10.1016/j.jaerosci.2014.10.012 [DOI] [Google Scholar]
  12. Grass RN, Limbach LK, Athanassiou EK, Stark WJ, 2010. Exposure of aerosols and nanoparticle dispersions to in vitro cell cultures: A review on the dose relevance of size, mass, surface and concentration. J. Aerosol Sci 41, 1123–1142. 10.1016/j.jaerosci.2010.10.001 [DOI] [Google Scholar]
  13. Gupta T, Demokritou P, Koutrakis P, 2004. Development and Performance Evaluation of a High-Volume Ultrafine Particle Concentrator for Inhalation Toxicological. Inhal. Toxicol 16, 851–862. 10.1080/08958370490506664 [DOI] [PubMed] [Google Scholar]
  14. Han X, Gelein R, Corson N, Wade-Mercer P, Jiang J, Biswas P, Finkelstein JN, Elder A, Oberdörster G, 2011. Validation of an LDH assay for assessing nanoparticle toxicity. Toxicology 287, 99–104. 10.1016/J.TOX.2011.06.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hering SV, Stolzenburg MR, 2005. A Method for Particle Size Amplification by Water Condensation in a Laminar, Thermally Diffusive Flow A Method for Particle Size Amplification by Water Condensation in a Laminar, Thermally Diffusive Flow. Aerosol Sci. Technol 39, 428–436. 10.1080/027868290953416 [DOI] [Google Scholar]
  16. Hewett P, 1995. The Particle Size Distribution, Density, and Specific Surface Area of Welding Fumes from Smaw and GMAW Mild and Stainless Steel Consumables. Am. Ind. Hyg. Assoc. J 56, 128–135. 10.1080/15428119591017150 [DOI] [PubMed] [Google Scholar]
  17. Holder AL, Goth-Goldstein R, Lucas D, Koshland CP, 2012. Particle-induced artifacts in the MTT and LDH viability assays. Chem. Res. Toxicol 25, 1885–1892. 10.1021/tx3001708 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Keane M, Siert A, Stone S, Chen BT, 2016. Profiling stainless steel welding processes to reduce fume emissions, hexavalent chromium emissions and operating costs in the workplace. J. Occup. Environ. Hyg 13, 1–8. 10.1080/15459624.2015.1072634 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kim JY, Chen J, Boyce PD, Christiani DC, 2005. Exposure to welding fumes is associated with acute systemic inflammatory responses. Occup. Environ. Med 62, 157–163. 10.1136/oem.2004.014795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kreyling WG, Semmler-Behnke M, Möller W, 2006. Ultrafine Particle - Lung Interactions: Does Size Matter? J. Aerosol Med 19, 74–83. 10.1089/jam.2006.19.74 [DOI] [PubMed] [Google Scholar]
  21. Kreyling WG, Semmler M, Möller W, 2004. Dosimetry and Toxicology of Ultrafine Particles. J. Aerosol Med 17, 140–152. 10.1089/0894268041457147 [DOI] [PubMed] [Google Scholar]
  22. Lednicky J, Pan M, Loeb J, Hsieh H, Eiguren- A, Hering S, Fan ZH, Wu C, 2016. Highly efficient collection of infectious pandemic influenza H1N1 virus (2009) through laminar-flow water based condensation. Aerosol Sci. Technol 50, 1–4. 10.1080/02786826.2016.1179254 [DOI] [Google Scholar]
  23. Lee SR, Wu CY, 2005. Size distribution evolution of fine aerosols due to intercoagulation with coarse aerosols. Aerosol Sci. Technol 39, 358–370. 10.1080/027868290931753 [DOI] [Google Scholar]
  24. Lenz A, Karg E, Brendel E, Hinze-heyn H, Maier KL, Eickelberg O, Stoeger T, Schmid O, 2012. Inflammatory and Oxidative Stress Responses of an Alveolar Epithelial Cell Line to Airborne Zinc Oxide Nanoparticles at the Air-Liquid Interface: A Comparison with Conventional, Submerged Cell-Culture Conditions. Biomed Res. Int 2013, 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lewinski NA, Liu NJ, Asimakopoulou A, Papaioannou E, Konstandopoulos A, Riediker M, 2017. Air-Liquid Interface Cell Exposures to Nanoparticle Aerosols, in: Biomedical Nanotechnology: Methods and Protocols. pp. 301–313. 10.1007/978-1-4939-6840-4_21 [DOI] [PubMed] [Google Scholar]
  26. Li H, Hedmer M, Monica K, Björk J, Stockfelt L, Tinnerberg H, Albin M, Broberg K, 2015. A Cross-Sectional Study of the Cardiovascular Effects of Welding Fumes. PLoS One 10, 1–15. 10.1371/journal.pone.0131648 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lichtveld KM, Ebersviller SM, Sexton KG, Vizuete W, Jaspers I, Je HE, 2012. In Vitro Exposures in Diesel Exhaust Atmospheres: Resuspension of PM from Filters versus Direct Deposition of PM from Air. Environ. Sci. Technol 46, 9062–9070. 10.1021/es301431s [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lioy PJ, Nazarenko Y, Han T, Lioy MJ, Mainelis G, 2010. Nanotechnology and Exposure Science: What is Needed to Fill the Research and Data Gaps for Consumer Products? Int. J. Occup. Environ. Health 16, 378–387. 10.1179/107735210799160057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Longest PW, Xi J, Longest PW, Xi J, 2008. Condensational Growth May Contribute to the Enhanced Deposition of Cigarette Smoke Particles in the Upper Respiratory Tract. Aerosol Sci. Technol 42, 579–602. 10.1080/02786820802232964 [DOI] [Google Scholar]
  30. Mcneilly JD, Heal MR, Beverland IJ, Howe A, Gibson MD, Hibbs LR, Macnee W, Donaldson K, 2004. Soluble transition metals cause the pro-inflammatory effects of welding fumes in vitro. Toxicol. Appl. Pharmacol 196, 95–107. 10.1016/j.taap.2003.11.021 [DOI] [PubMed] [Google Scholar]
  31. Muller ACA, Scrivener KL, Skibsted J, Gajewicz AM, Mcdonald PJ, 2015. Influence of silica fume on the microstructure of cement pastes: New insights from 1H NMR relaxometry. Cem. Concr. Res 74, 116–125. 10.1016/j.cemconres.2015.04.005 [DOI] [Google Scholar]
  32. Murgia X, Pawelzyk P, Schaefer UF, Wagner C, Willenbacher N, 2016. Size-Limited Penetration of Nanoparticles into Porcine Respiratory Mucus after Aerosol Deposition. Biomacromolecules 17, 1536–1542. 10.1021/acs.biomac.6b00164 [DOI] [PubMed] [Google Scholar]
  33. Oberdörster G, Oberdörster E, Oberdörster J, 2005. Nanotoxicology: An Emerging Discipline Evolving from Studies of Ultrafine Particles. Environ. Health Perspect 113, 823–839. 10.1289/ehp.7339 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Oberdörster G, Sharp Z, Atudorei V, Elder A, Van Gelder R, Kreyling WG, Cox C, 2004. Translocation of inhaled ultrafine particles to the brain. Ina. Toxicol 16, 437–445. 10.1080/08958370490439597 [DOI] [PubMed] [Google Scholar]
  35. Pan M, Hsieh H, Hering SV, Lednicky J, Hugh Z, 2016. Efficient collection of viable virus aerosol through laminar-flow, water-based condensational particle growth. J. Appl. Microbiol 805–815. 10.1111/jam.13051 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Paur H, Cassee FR, Teeguarden J, Fissan H, Diabate S, Aufderheide M, Kreyling WG, Otto H, Kasper G, Riediker M, Rothen-rutishauser B, Schmid O, 2011. In-vitro cell exposure studies for the assessment of nanoparticle toxicity in the lung—A dialog between aerosol science and biology 42, 668–692. 10.1016/j.jaerosci.2011.06.005 [DOI] [Google Scholar]
  37. Saarikoski S, Williams LR, Spielman SR, Lewis GS, Eiguren-Fernandez A, Worsnop DR, Timonen H, 2019. Laboratory and field evaluation of the Aerosol Dynamics Inc. concentrator (ADIc) for aerosol mass spectrometry. Atmos. Meas. Tech 12, 3907–3920. 10.5194/amt-12-3907-2019 [DOI] [Google Scholar]
  38. Stone V, Miller MR, Clift MJD, Elder A, Mills NL, Møller P, Schins RPF, Vogel U, Kreyling WG, Jensen KA, Kuhlbusch TAJ, Schwarze PE, Hoet P, 2017. Nanomaterials Versus Ambient Ultrafine Particles: An Opportunity to Exchange Toxicology Knowledge. Environ. Health Perspect 1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sturm R, 2016. Local lung deposition of ultrafine particles in healthy adults: experimental results and theoretical predictions. Ann. Transl. Med 4, 1–9. 10.21037/atm.2016.11.13 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Svensson CR, Ameer SS, Ludvigsson L, Ali N, Alhamdow A, Messing ME, Pagels J, Gudmundsson A, Bohgard M, Sanfins E, Kåredal M, Broberg K, Rissler J, 2016. Validation of an air–liquid interface toxicological set-up using Cu, Pd, and Ag wellcharacterized nanostructured aggregates and spheres. J. Nanoparticle Res 18, 86 10.1007/s11051-016-3389-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Tilly TB, Ward RX, Luthra JK, Robinson S, Eiguren-Fernandez A, Lewis GS, Salisbury RL, Lednicky JA, Sabo-Attwood TL, Hussain SM, Wu C-Y, 2019. Condensational particle growth device for realiable cell exposure at the air-liquid interface to nanoparticles. Aerosol Sci. Technol 53, 1415–1428. 10.1080/02786826.2019.1659938 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Topham N, Wang J, Kalivoda M, Huang J, 2011. Control of Cr6+ Emissions from Gas Metal Arc Welding Using a Silica Precursor as a Shielding Gas Additive. Ann. Occup. Hyg 56, 233–241. 10.1093/annhyg/mer103 [DOI] [PubMed] [Google Scholar]
  43. Wang J, Wu CY, Franke G, 2014. Effectiveness of amorphous silica encapsulation technology on welding fume particles and its impact on mechanical properties of welds. Mater. Des 54, 79–86. 10.1016/j.matdes.2013.08.002 [DOI] [Google Scholar]
  44. Xing Y, Xu Y, Shi M, Lian Y, 2016. The impact of PM2.5 on the human respiratory system. J. Thorac. Dis 8, 69–74. 10.3978/j.issn.2072-1439.2016.01.19 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Yu KM, Topham N, Wang J, Kalivoda M, Tseng Y, Wu CY, Lee WJ, Cho K, 2011. Decreasing biotoxicity of fume particles produced in welding process. J. Hazard. Mater 185, 1587–1591. 10.1016/j.jhazmat.2010.09.083 [DOI] [PubMed] [Google Scholar]
  46. Zavala J, Greenan R, Krantz QT, Demarini DM, Higuchi M, Ian M, White PA, 2017. Regulating temperature and relative humidity in air–liquid interface in vitro systems eliminates cytotoxicity resulting from control air exposures. Toxicol. Res. (Camb) 6, 448–459. 10.1039/c7tx00109f [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Zavala J, Lichtveld K, Ebersviller S, Carson JL, Walters GW, Jaspers I, Jeffries HE, Sexton KG, Vizuete W, 2014. The Gillings Sampler - An electrostatic air sampler as an alternative method for aerosol in vitro exposure studies. Chem. Biol. Interact 220, 158–168. 10.1016/j.cbi.2014.06.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zimmer AT, Baron PA, Biswas P, 2002. The infuence of operating parameters on numberweighted aerosol size distribution generated from a gas metal arc welding process. J. Aerosol Sci 33, 519–531. 10.1016/S0021-8502(01)00189-6 [DOI] [Google Scholar]
  49. Zimmer AT, Biswas P, 2001. Characterization of the aerosols resulting from arc welding processes. J. Aerosol Sci 32, 993–1008. 10.1016/S0021-8502(01)00035-0 [DOI] [Google Scholar]

RESOURCES