Physicochemical Properties and Bioreactivity of Sub-10 μm Geogenic Particles: Comparison of Volcanic Ash and Desert Dust

Abstract

Exposure to ambient particulate matter (PM) with an aerodynamic diameter of <10 μm (PM₁₀) is a well-established health hazard. There is increasing evidence that geogenic (Earth-derived) particles can induce adverse biological effects upon inhalation, though there is high variability in particle bioreactivity that is associated with particle source and physicochemical properties. In this study, we investigated physicochemical properties and biological reactivity of volcanic ash from the April 2021 eruption of La Soufrière volcano, St. Vincent, and two desert dust samples: a standardized test dust from Arizona and an aeolian Gobi Desert dust sampled in China. We determined particle size, morphology, mineralogy, surface texture and chemistry in sub-10 μm material to investigate associations between particle physicochemical properties and observed bioreactivity. We assessed cellular responses (cytotoxic and pro-inflammatory effects) to acute particle exposures (24 hr) in monocultures at the air-liquid interface using two types of cells of the human airways: BEAS-2B bronchial epithelial cells and A549 alveolar type II epithelial cells. In acellular assays, we also assessed particle oxidative potential and the presence of microorganisms. The results showed that volcanic ash and desert dust exhibit intrinsically different particle morphology, surface textures and chemistry, and variable mineralogical content. We found that Gobi Desert dust is more bioreactive than freshly erupted volcanic ash and Arizona test dust, which is possibly linked to the presence of microorganisms (bacteria) and/or nanoscale elongated silicate minerals (potentially clay such as illite or vermiculite) on particle surfaces.

Keywords: bioreactivity; desert dust; particulate matter; physicochemical properties; respiratory hazard; volcanic ash.

Figure 1

Particle size and morphology distributions of the Arizona test dust (ATD), Gobi Desert dust (GDD) and St. Vincent volcanic ash (SVA) respirable samples. (a) Cumulative particle size distributions in volume percent (Vol.%) and number percent (No.%) of the three samples acquired by laser diffraction. (b) Particle solidity and aspect ratio distributions in the three samples obtained by image analyses of particles with equivalent circular diameters >0.5 μm in scanning electron microscope images (n is the number of particles measured). Distributions are presented as Tukey box and whisker plots, where the thick middle line is the median, the lower and upper hinges mark the 25th and 75th percentiles, the lower and upper whiskers extend to 1.5 × the interquartile range (i.e., distance between the first and third quartiles), and outliers are plotted individually.

Figure 2

Phase quantification in the three respirable samples Arizona test dust (ATD), Gobi Desert dust (GDD) and St. Vincent volcanic ash (SVA), performed by image analyses of the EDS maps acquired with the FlatQuad detector using the ImageJ macro “EDS‐pie.” Example multi‐elemental EDS maps are shown in (a), (c) and (e). Color keys are displayed on the maps. The corresponding phase map constructions are shown in (b), (d) and (f). Color key is in (h). The phase salts is inclusive of all identified sulfates/sulphides and chlorides. Phase proportions (in area %) for the three samples (calculated from 3 to 4 maps for each, all provided in Supplementary file 2 in Supporting Information S1) are shown in (g) with the corresponding color key in (h). The values for the “total particle area quantified” correspond to the mean of the two users (due to differences in image cropping). In (f), the brown phase corresponds to Fe‐Ti oxides, that are counted together with olivine and pyroxene in (g).

Figure 3

Surface texture of respirable particles from the Arizona test dust (ATD), Gobi Desert dust (GDD) and St. Vincent volcanic ash (SVA) samples imaged by scanning electron microscopy in secondary electron mode. (a–e) ATD particles imaged at 10 kV, 0.1 nA and a working distance of 5 mm. (f–k) GDD particles imaged at 2 kV, 50 pA and a working distance of 4 mm. (l–q) SVA particles imaged at 2 kV, 50 pA and a working distance of 4 mm.

Figure 4

Chemical composition of adhering grains present at the surface of Gobi Desert dust (GDD) particles. (a) Scanning Transmission Electron Microscopy (STEM) image in Dark Field (DF) mode of the electron transparent thin section extracted from a calcite particle. (b) EDS map of the thin section in (a) showing the distribution of silicon (Si; blue), iron (Fe; yellow) and calcium (Ca; red). The yellowish outer rim, as indicated by the white arrows in (a) and (b), represents the platinum‐based protective layers deposited on the calcite grain before extraction of the thin section. (c) EDS spectra of the core of the calcite particle (red spectrum) and the silica‐rich rim (blue spectrum) shown in blue in (b). The unlabeled peaks at 0.9 and 2.0 keV in (c) are the L band of copper (material constituting the thin section stand) and the M band of platinum, respectively.

Figure 5

Oxidative potential (OP) of the three respirable geogenic samples: Arizona test dust (ATD), Gobi Desert dust (GDD) and St. Vincent volcanic ash (SVA). Data are expressed as oxidative activity (rate of reagent consumption) normalized to sample mass in nmol/min/μg, as determined by measurement of depletion of (a) dithiothreitol (OP_DTT) and (b) ascorbate (OP_AA). Data are presented as Tukey boxplots with the mean represented as a + and the median as a horizontal line. Data are from measurement replicates n = 16 for OP_DTT and n = 12 for OP_AA. Statistical significance between OP of different samples is indicated as * (p ≤ 0.05).

Figure 6

Cytotoxicity toward (a) A549 and (b) BEAS‐2B cells of the respirable Arizona test dust (ATD), Gobi Desert dust (GDD) and St. Vincent volcanic ash (SVA) at two doses (0.1 and 10 μg/cm²) after 24 hr exposure at an ALI, measured by lactate dehydrogenase (LDH) activity. LDH release is expressed as a fold change relative to untreated cells (serum‐free cell culture medium only) measured from three (A549 N = 3) or four (BEAS‐2B N = 4) independent experiments. Triton X‐100 (TX) at 0.2% in phosphate buffered saline acted as the positive assay control. The data are presented as the median with range. The points are the average value of individual measurements of the replicates within an experiment. Statistical significance between treated and untreated cells is indicated as * (p < 0.05).

Figure 7

Pro‐inflammatory response in alveolar epithelial cells (A549) assessed by cytokine production following acute exposure to the respirable Arizona test dust (ATD), Gobi Desert dust (GDD) and St. Vincent volcanic ash (SVA) at two doses (0.1 and 10 μg/cm²) measured after 24 hr exposure at an ALI in culture supernatants on duplicates or triplicates of three independent experiments (N = 3). Protein production is presented as the cytokine concentration normalized to the total protein concentration and expressed as a fold change relative to untreated cells for (a) IL‐6 and (b) IL‐8. Lipopolysaccharide (LPS, from E. coli, at 10 μg/mL) was used as a pro‐inflammatory stimulant. Data are presented as the median with range. The points are the average value of individual measurements of the replicates within an experiment. Statistical significance between treated and untreated cells is indicated as * (p < 0.05).

Figure 8

Pro‐inflammatory response in bronchial epithelial cells (BEAS‐2B) assessed by cytokine production following acute exposure to the respirable Arizona test dust (ATD), Gobi Desert dust (GDD) and St. Vincent volcanic ash (SVA) at two doses (0.1 and 10 μg/cm²) measured after 24 hr exposure in culture supernatants on duplicates of four independent experiments (N = 4). Protein production is presented as the cytokine concentration normalized to the total protein concentration and expressed as a fold change relative to untreated cells for (a) IL‐6, (b) IL‐8, (c) IL‐1β, and (d) TNF‐α. The background level of TNF‐α (i.e., in untreated cells) was unquantifiable in one experiment, and therefore, a fold change could not be calculated for this experiment and the data are presented for N = 3. Lipopolysaccharide (LPS, from E. coli, at 1 μg/mL) was used as a pro‐inflammatory stimulant. Data are presented as the median with range. The points are the average value of individual measurements of the replicates within an experiment. Statistical significance between treated and untreated cells is indicated as * (p < 0.05).