Respirable Particulate Matter Induces Natural Killer Cell Dysfunction via Oxidative Stress-Mediated Lysosomal and Mitochondrial Injury: Implications for Mineral-Rich Mining Dust Exposure

Abstract

Epidemiological and occupational studies increasingly link respirable particulate matter (PM), including mineral dust, exposure with elevated lung cancer risk, yet the immunological mechanisms underlying these associations remain poorly defined. In this study, we investigated how compositionally distinct respirable PMs affect natural killer (NK) cells, a key component of immune surveillance. Using in vitro exposure and K562 target-cell challenge models, we show that respirable mineral-containing PMs induce dose- and composition-dependent NK-cell injury, including increased apoptosis/necrosis, altered cytokine profiles, changes in perforin/granzyme B-associated functions, and impaired effector responses under tumor-cell challenge conditions. Silica- and silicate-rich particles caused the most pronounced dysfunction, whereas iron-bearing and carbonaceous dusts produced distinct patterns of oxidative injury and cell loss. Mechanistic analyses identified mitochondrial depolarization, reduced mitochondrial mass, and lysosomal deacidification as convergent organelle stress pathways associated with altered NK-cell function. Ascorbic acid-associated improvements in organelle function and IFN-γ production further support a contribution of ROS-associated organelle dysfunction to impaired NK-cell responses. Collectively, these findings provide the first integrated adverse outcome pathways linking respirable PM exposure to NK-cell dysfunction, highlighting oxidative stress-driven lysosomal and mitochondrial injury as central pathways that may compromise NK-cell-mediated immune surveillance under tumor-cell challenge conditions.

1. Introduction

Air pollution, particularly exposure to respirable particulate matter (PM), is a major global environmental health concern. A growing body of evidence links both urban and industrial PMs to a range of adverse respiratory outcomes, including lung cancer. The Global Burden of Disease 2017 Risk Factor Collaborators estimated that ambient PM pollution, especially PM_2.5, and occupational exposure, particularly to crystalline silica (CS), were the second and third leading contributors to global lung cancer mortality, following tobacco smoking. The carcinogenic effects of respirable PM have traditionally been attributed to its genotoxic potential. Organic components such as polycyclic aromatic hydrocarbons (PAHs) form DNA adducts, while mineral components, including crystalline silica and asbestos, promote mutagenesis by driving persistent inflammation and oxidative DNA damage. , Additionally, recent findings highlight a complementary, nongenotoxic route, in which PM exposure disrupts immune balance and possibly contributes to emerging tumors.

Natural killer (NK) cells, a subset of innate lymphoid cells accounting for 10–15% of resident and circulating pulmonary lymphocytes, , play a critical role in surveillance and elimination of virally infected and transformed cells in the lung. Dysfunction of NK cells has been increasingly linked to various diseases, including cancer progression, pneumonia, chronic obstructive pulmonary disease (COPD), and viral infection. For example, NK cells from smokers exhibit reduced cytotoxicity, lower perforin expression, and diminished secretion of IFN-γ and TNF-α, while in COPD, lung NK cells show excessive cytotoxicity against autologous epithelial cells that correlates with disease severity. − These findings support the conclusion that the imbalance of NK cells through either suppression or dysregulated activation drives pulmonary pathology and, possibly, provides a reasonable explanation for the prevalence of smoking-associated lung cancer. As such, we hypothesize that environmental exposure to respirable PM may similarly disrupt lung NK cell function, thereby altering immune surveillance and potentially contributing to elevated lung carcinogenesis risk.

The source and composition, especially the mineral composition, of PM critically influence its toxicity. − CS-containing dust exposure shows a higher likelihood of developing lung cancer as a complication of silicosis, whereas such cancer is less frequently observed in classic coal mining dust-associated lung dysfunction. , Asbestos exposure further illustrates the importance of mineral composition, as it is a well-established cause of mesothelioma. , Iron mineral-bearing dusts, such as urban magnetic dust containing both magnetite and elemental iron (Fe⁰) or coal dust containing siderite and pyrite, are highly reactive, promoting ROS formation and potentially contributing to neurodegeneration and pulmonary injury. , Kaolin, a silicate mineral, exhibits weaker but measurable toxicity at higher concentrations or after prolonged exposure to lung macrophages. , Such compositional differences may be essential in determining how PM alter the NK cell function.

In this study, we hypothesize that inhalation of respirable mineral PM will lead to NK cell dysfunction, thereby contributing to the increased prevalence of lung cancer. As such, we comprehensively investigated how PM with distinct major mineral components impairs NK cell function and weakens immune surveillance. Using the NK-92MI cell line, we evaluated the effects of mineral-bearing PM exposure on cellular states, cytokine production, oxidative stress profiling, expression of key cytotoxic proteins, and, ultimately, the impact on target cells, with particular emphasis on the integrity and function of lysosomes and mitochondria. Our findings provide the first direct evidence that respirable mineral-bearing PM exposure compromises NK-mediated immune surveillance. Additionally, we established an adverse outcome pathway (AOP) framework linking PM exposure to immune dysfunction, thereby bridging additional mechanistic insights with epidemiological evidence of PM-associated lung cancer.

2. Materials and Methods

2.1. Chemicals and Reagents

Details of chemicals and reagents used can be found in Text S1 and Table S1 in the Supporting Information.

2.2. Respirable Particulate Matter Preparation

Three environmental bulk samples were sourced from coal mining sites located in the eastern states of Australia. These bulk samples represent the typical environmental component of crystalline silica, silicates, iron-bearing minerals, and carbonaceous matter that workers and surrounding communities are commonly exposed to during mining activities. Both SiO₂ and kaolin standards were also employed in this work to better elucidate the role of distinct mineral composition in disease progression, given their widespread occurrence in sandstorms and respirable ambient dust in mining and construction sites. − All bulk environmental samples were size-fractionated to enrich the respirable range (<10 μm) as reported previously (Text S2). Dust characterizations were conducted to measure key properties associated with in vivo PM toxicity using a variety of established techniques (Text S3), including X-ray fluorescence (XRF), X-ray diffraction (XRD), particle size, surface charge, and surface area.

2.3. Cell Culture, Cell Treatments, IFN-γ Measurement, and Total Intracellular GSH Level

NK-92MI cells and K562 cells were both obtained from the American Type Culture Collection (ATCC). The NK-92MI cells were cultured in α-MEM medium without nucleosides supplemented with 0.1 mM 2-mercaptoethanol, 0.2 mM myo-inositol, 0.02 mM folic acid, 12.5% fetal bovine serum (FBS), 12.5% horse serum, and 1% penicillin-streptomycin in a 5% CO₂ humidified incubator at 37 °C as suggested by ATCC. K562 cells were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin at 37 °C in a humidified incubator containing 5% CO₂, and were washed and resuspended in the NK-cell exposure medium before co-culture experiments. All cells were serum-starved in phenol red-free DMEM/F12 medium for 12 h before exposure to mineral-bearing respirable PM concentrations of 0, 50, 200, and 400 mg/L, equivalent to cell-normalized doses of 0, 100, 400, and 800 pg/cell, respectively. Details of dose and experimental duration justifications and sample sterilization are provided in Texts S4 and S5. After 7 and 24 h of PM exposure, the production of IFN-γ from NK-92MI was measured using a human IFN-γ ELISA kit according to the manufacturer’s instructions and analyzed using a microplate reader (FLUOstar Omega; BMG Labtech). The intracellular GSH level of NK cells after 4 h of dust exposure was measured using a commercial kit (ab65322, Abcam) according to the manufacturer’s instructions and analyzed using a microplate reader.

2.4. Western Blotting

NK-92MI intracellular proteins were homogenized in RIPA buffer containing 1% protease inhibitors on ice after 4 h of 200 and 400 pg/cell respirable PM treatment to quantify the protein levels of perforin, granzyme B, and total LAMP1. Further detailed information is provided in Text S6.

2.5. Flow Cytometry Analysis of Apoptosis, Oxidative Stress, Organelle Function, and NK-Cell Effector Responses

Apoptosis and necrosis were assessed in NK cells exposed to PM at 100, 400, or 800 pg/cell for 4 or 24 h. Intracellular ROS in NK-92MI cells exposed to different PM concentrations were quantified at 1, 4, or 24 h using the DCFH-DA assay following established protocols. , CD107a degranulation was evaluated after 4 h exposure to PM at 400 or 800 pg/cell using published methods with minor modifications. , Mitochondrial membrane potential and total mitochondrial mass were measured at 4 h using TMRM and MitoTracker Green FM, respectively, and quantified by flow cytometry according to the manufacturer’s instructions. The lysosome-like organelles’ functional status after 4 h of dust exposure was measured by LysoTracker Deep Red according to the manufacturer’s instructions. Target cell killing was assessed by co-culturing PM-pre-exposed NK-92MI cells with K562 target cells, followed by flow-cytometric quantification of carboxyfluorescein diacetate succinimidyl ester (CFSE)-positive and PI-positive K562 cells. Further detailed information is provided in Texts S7–S12.

2.6. Immunofluorescence Staining

Live NK-92MI cells were stained with MitoTracker Green FM and LysoTracker Deep Red according to the manufacturers’ instructions to visualize mitochondria and lysosomes, respectively, with NK-92MI nuclei labeled with DAPI (1 mg/L, shown in blue). Cells were not fixed or permeabilized before staining. Under this condition, DAPI-positive cells suggest compromised membrane integrity. Confocal images were acquired on a Zeiss Cell Discoverer 7 with LSM 900 (Ramaciotti) and analyzed by FIJI ImageJ.

2.7. Statistical Analysis

One-way ANOVA, two-way ANOVA analysis, or t-tests were performed using GraphPad as appropriate for comparisons among individual treatment groups. The normality tests were conducted using the Q-Q plot method (shown in Figure S1). All data are expressed as mean ± standard deviation. The confidence level used in statistical analysis was 95% (N ≥ 3), with significance defined as * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, or otherwise not significant (n.s.).

3. Results and Discussion

3.1. Key Properties and Elemental Compositions of Dust Samples

XRF and XRD analyses (Figures A–B) identified three distinct lithologies among the environmental samples, namely, coal, iron-rich rock, and silica/silicate-rich rock, which were accordingly classified as high-grade coal, sideritic waste band, and siliceous roof rock based on their elemental compositions. The coal sample was dominated by carbon (78%) with trace silica (4.6%) and alumina (3.3%), consistent with its predominantly amorphous phase shown in Figure S2 (99.8% amorphous and 0.2% kaolinite). The Rock (Fe) contained moderate carbon (23%) and high Fe₂O₃ (46%) with minor SiO₂ and Al₂O₃ detected, aligned with the trace quartz (0.1%) in XRD refinement (Figure S2). The substantial Fe₂O₃ content in Rock (Fe) corresponded to the dominance of siderite (99.9% of crystalline phases and 65.5% of bulk composition shown in Figures B and S2). Rock (Si) was characterized by high SiO₂ (77.8%) and Al₂O₃ (9.7%), with negligible Fe₂O₃ (1.9%) and no detectable carbon (Figure A). The high proportions of silica and alumina were confirmed by XRD as quartz and kaolinite, consistent with the classification of siliceous roof rock. Analytical-grade SiO₂–S and kaolin standards were purchased, and their compositions were verified by XRD and XRF analyses, which confirmed high purity with only 1.2% Illite present in the kaolin standard.

1.

Key physicochemical properties of selected PMs that are associated with in vivo toxicity. (A) Elemental composition of respirable environmental PM (Coal, Rock (Fe), and Rock (Si)) and purchased SiO₂ and kaolin standard (SiO₂_S and Kaolin) determined by WD-XRF and elemental analysis. Results are reported as corresponding oxide forms rather than individual elements. (B) Mineral phase composition measured by XRD analysis. (C) Surface charge of the PMs determined by the zeta potential measurement. (D) Surface area of PMs measured by Brunauer–Emmett–Teller (BET) analysis. Note that the XRD refinement results, including both the amorphous and mineral phases of the PMs, are shown in Figure S2. Details of the R Bragg value and fitting goodness of XRD analysis can be found in Table S3 and Table S4.

All PMs exhibited negative surface charges (−58.30 ± 2.03 to −9.76 ± 0.21 mV, Figure C), consistent with previous studies. , The strongest negative charges occurred in silica and silicate-rich samples (SiO₂_S, Rock (Si), Kaolin), likely due to the presence of deprotonated silanol (SiO^–) and aluminol (AlO^–) groups. BET surface areas ranged from 7.69 ± 0.12 to 20.06 ± 2.83 m²/g (Figure D), with higher values linked to kaolinite content in the PM (19.3% in Rock (Si) and 98.8% in Kaolin, Figure S2). The elevated surface area of kaolinite-rich dust reflects both its finer particle size (Table S2) and platy morphology (Figure S3). Although kaolinite was the dominant crystalline phase in the Coal sample used in this work (Figure B), its overall abundance in this PM was only 0.2% (Figure S2), explaining its relatively low surface area.

3.2. Respirable PMs Induce Apoptosis/Necrosis of NK Cells

NK cells are essential innate cytotoxic lymphocytes that eliminate transformed cells through multiple mechanisms, including perforin/granzyme B-mediated cytolysis, FasL/Fas-mediated apoptosis, and cytokine-associated effector functions. Reduced functional NK cell populations and/or decreased production of these cytolytic proteins and cytokines may contribute to cancer progression via impaired immune surveillance. Quantitative analysis revealed that all selected respirable PMs induced NK cell apoptosis and necrosis in a dose- and time-dependent manner, with PMs associated with waste band and roof rock containing greater amounts of minerals, such as iron minerals, quartz, and kaolin (Rock (Fe), Rock (Si), Kaolin), causing greater damage (Figure ). The crystalline or mineral-rich particles tended to result in greater necrosis in NK cells than that of the amorphous phase. For instance, while quartz is a well-documented carcinogenic material, the standard material (SiO₂–S) containing only 10% of crystalline phase led to only slightly higher necrosis of NK cells than Coal, with a statistical difference not achieved (Figure B, 24 h, p > 0.05). Rather than uniformly reducing viable cells at early time points, PM exposure appeared to play a more critical role in accelerating the transition of early apoptotic NK cells into late apoptosis and necrosis phase, thereby increasing the overall proportion of necrotic cells, particularly at high doses and later time points. After 24 h exposure, apoptosis and necrosis in the control group were 26.4 ± 13.1% and 22.9 ± 2.9%, respectively, whereas exposure to Coal at 400 pg/cell reduced apoptosis to 12.5 ± 9.9% (p < 0.01) and increased late apoptosis and necrosis to 42.0 ± 10.3% (p < 0.0001). This transition was greater for kaolin (3.7 ± 1.7% for apoptosis vs. 56.9 ± 2.2% for late apoptosis and necrosis, p < 0.0001). The greater cytotoxicity of Kaolin may partly reflect its platy morphology, which could result in greater membrane disturbance during internalization, consistent with previous reports of stronger cytotoxicity from plate-like than spherical or rounded particles. Collectively, these results indicate that PM exposure substantially decreases viable NK cell populations, potentially compromising their immune surveillance capacity.

2.

Apoptosis and necrosis of NK cells after 4 h (A) and 24 h (B) of respirable PM exposure were analyzed by flow cytometry (N = 3). The mean and standard deviation (SD) values for all cellular state measurements are summarized in Table S5.

3.3. Respirable PM Alters NK Cell Cytotoxic Functions and Impairs NK-Cell Responses under Target-Cell Challenge Conditions

IFN-γ, an essential cytokine in NK cell-mediated anti-tumor immunity, was measured together with IL-10, an immunoregulatory cytokine associated with anti-inflammatory and effector immune modulation, to evaluate changes in the basal functional state of NK cells following PM exposure. − ELISA analysis showed a clear dose- and time-dependent decline in IFN-γ production (Figure A–B), closely linked to respirable PM type (Figure A). Rock (Si) and Kaolin caused the most pronounced reductions, with IFN-γ levels after 24 h exposure to PM at 800 pg/cell decreasing to 9.3 ± 1.0 ng/L and 3.6 ± 2.0 ng/L, respectively, compared with 45.6 ± 15.3 ng/L in the control group (both p < 0.0001, Figure B). By contrast, Coal showed only a modest reduction at 7 h (p < 0.01) (Figure A) and no significant decrease at 24 h even at 800 pg/cell PM exposure (Figure B). Interestingly, although SiO₂_S-treated NK cells maintained relatively high viability, with viable cell numbers at 800 pg/cell remaining at 89–91% of the control level (Figure ), IFN-γ production at both 7 and 24 h was markedly lower than that in cells with comparable viability, particularly Coal-treated cells (p < 0.001). ELISA analysis showed particle- and dose-dependent changes in IL-10 production following PM exposure over time, with SiO₂_S and Rock (Fe) inducing the strongest increase (p < 0.0001), particularly at 400 pg/cell post 24 h exposure, whereas Coal showed comparatively weaker responses (Figure C). Silica/silicate-rich particles, including Rock (Si) and Kaolin, produced more variable responses across doses. Collectively, these cytokine profiles indicate that PM exposure alters the basal immune state of NK cells in a composition-dependent manner, with potential relevance to NK cell-mediated immune surveillance. The functional consequences of these basal changes under activating conditions are examined in the following section.

3.

Effects of different respirable PMs on NK-cell basal immune state and effector responses. (A) IFN-γ production in the presence of 800 pg/cell PM at 7 h (N = 5) and (B) 100, 400, and 800 pg/cell at 24 h measured by ELISA (N = 5). (C) IL-10 production in the presence of 100 and 400 pg/cell PM at 7 and 24 h (N > 3). (D-F) Expression of perforin and granzyme B in NK cells after 4 h of 200 and 400 pg/cell PM exposure was analyzed by Western blotting (N = 3), with GAPDH and tubulin used as the housekeeping proteins for perforin and granzyme B, respectively. LMW Granzyme B denotes the lower-molecular-weight (∼26 kDa) Granzyme B species. (G) Degranulation of NK cells as indicated by the surface expression of CD107a post 4 h of 400 pg/cell and 800 pg/cell PM exposure measured by flow cytometry (N > 3). (H) CD107a-positive control was obtained by co-culturing NK cells with A549 as target cells at an effector-to-target (E:T) ratio of 1:1 for 4 h. (I) IFN-γ production in the presence of K562 cells at an NK:K562 ratio of 10:1 measured by ELISA (N > 3). (J) Schematic of the experimental design for K562 target-cell challenge assays, including short-term PM pre-exposure of NK cells followed by co-culture with K562 cells and (K) Flow cytometry quantification of K562 cell death by using CFSE and PI staining, including particle-only controls (green), total K562 death (gray), and NK-mediated killing after background correction (pink) following PM pre-exposure. The fold changes shown in panels D, E, and G were calculated as [value]_PM/[value]_Control. Error bars represent mean ± standard deviation. Statistical significance is indicated as follows: * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, or otherwise not significant (n.s.).

The synthesis and release of perforin and granzyme B are essential to NK cell-mediated cytotoxicity against virally infected and transformed cells, which are crucial for suppressing lung cancer metastasis. Briefly, perforin forms pores in the membrane of the target cell, allowing granzyme B to enter the cytoplasm, where it subsequently triggers apoptosis. Western blot analysis revealed significant particle-dependent alterations in both perforin and granzyme B expression following PM exposure (Figures D–F). Perforin was generally elevated at 200 pg/cell, whereas responses at 400 pg/cell varied among PM types (p < 0.01), except for Kaolin, which showed no clear increase. The lower-molecular-weight granzyme B species (∼26 kDa) was significantly increased in SiO₂_S-, Coal-, and Rock (Fe)-treated cells at both doses (p < 0.05), whereas Rock (Si) and Kaolin showed non-significant increases or reduced responses at the higher dose. In parallel, changes in the higher-molecular-weight granzyme B species (∼32 kDa) were also observed (Figure S4). These findings indicate particle-dependent alterations in the relative distribution of granzyme B molecular forms following PM exposure. These changes occurred despite comparable viable NK-cell proportions across treatments (82–86%, p > 0.05, Figure ), suggesting that PM composition, rather than differences in cell viability alone, contributed to the altered cytotoxic granule-associated phenotype. The dose-dependent reduction in detectable perforin and granzyme B in Rock (Si)- and Kaolin-treated cells may be associated with intracellular organelle dysfunction, as discussed below.

NK-cell degranulation involves the fusion of LAMP1-positive perforin- and granzyme-containing granules with the plasma membrane to deliver cytotoxic molecules to target cells. Upon fusion with the plasma membrane, LAMP1 is exposed externally and can be detected as surface CD107a, , serving as an indicator of the exocytosis capability of NK cells toward target cells. To accurately quantify total CD107a expression and prevent its internalization after degranulation, monensin was added according to established protocols, , with procedural details illustrated in Figure S5. Flow cytometry analysis revealed that PM exposure significantly reduced the translocation of perforin- and granzyme-containing lytic granules to the plasma membrane (p < 0.0001) (Figure G), indicating disruption of granule trafficking or exocytotic processes, which may impair NK cell-mediated immune surveillance and limit the clearance of damaged or transformed target cells. The unchanged total LAMP1 expression across doses and PM types (p > 0.05) (Figure S6) suggests that the reduction of surface CD107a (Figure G) reflects impaired granule exocytosis.

To determine whether these basal changes translated into impaired responses under target-cell challenge conditions, we assessed IFN-γ production in the presence of K562 cells, a classical NK target-cell line, and evaluated NK-mediated K562 killing after short-term PM pre-exposure. Consistent with the dust-only results, co-incubation of NK cells with selected PMs and K562 cells at an NK:K562 ratio of 10:1 substantially reduced IFN-γ production, particularly in the presence of silica/silicate-rich particles (Figure I, p < 0.001). Short-term PM exposure caused limited direct K562 cell death (green bars in Figures J–K, p > 0.05). However, NK cells pre-exposed to silica/silicate-rich PMs showed markedly reduced K562 killing after background correction (pink bars in Figures J–K, p < 0.0001). These findings indicate that PM exposure reshapes the basal immune state of NK cells and, for selected particles, can further impair NK-cell responses under tumor-cell challenge conditions. These composition-specific alterations in cytotoxic proteins may explain the stronger association between silica/silicate-rich dust exposure and lung cancer risk compared with high-grade coal dust. −

3.4. Respirable PM Causes Oxidative Stress and Damage to Both Mitochondria and Lysosomes

Intracellular reactive oxygen species (ROS) play essential roles in diverse biological processes, including signal transduction and mitochondrial dynamics. − Flow cytometry analysis of NK cells exposed to various PMs revealed a clear dose- and time-dependent response (Figure A). In contrast to the IFN-γ and IL-10 responses shown in Figure , low-dose PM exposure (100 pg/cell) resulted in only minor, non-significant fluctuations in ROS levels across all PM types (p > 0.05). A marked reduction in intracellular ROS was observed when PM concentrations were greater than 400 pg/cell, either via a consistent decline (SiO₂_S, Coal, Rock (Fe), and Kaolin, p < 0.01) or a biphasic pattern (Rock (Si), p < 0.0001), with an initial rise followed by a significant decrease. Among all PM types, Rock (Fe) induced the most pronounced early increase in ROS (178.0 ± 15.6% and 197.5 ± 34.6% of control at 400 and 800 pg/cell, respectively, p < 0.0001), likely reflecting labile iron-mediated ROS generation and a rapid consumption of intracellular antioxidants. ,− The sharp decrease in intracellular ROS in Figure A may be attributed to either enhanced antioxidant buffering, reduced mitochondrial mass or metabolic activity, and/or clearance of damaged mitochondria following mitochondrial depolarization. ,−

4.

PM-induced oxidative stress, organelle dysfunction, and altered effector responses in NK cells. (A) Intracellular ROS levels in NK cells measured by flow cytometry using the DCFH-DA probe (N = 3). Mean ± SD values are summarized in Table S6. The positive control for ROS generation was obtained by incubating NK cells with 100 μM H₂O₂ for 4 h (Figure S10). (B) Representative confocal images of live NK cells stained with MitoTracker Green (mitochondria), LysoTracker Deep Red (lysosomes), and DAPI (nuclei/membrane-compromised cells) following 4 h exposure to 400 pg/cell Kaolin. Scale bar = 10 μm. (C) Mitochondrial membrane potential (ΔΨm) in NK cells after 4 h exposure to 400 pg/cell PMs in the absence or presence of 0.5 mM Asc, measured by flow cytometry (N = 3). The positive control for mitochondrial depolarization was obtained by incubating NK cells with 600 μM H₂O₂ for 4 h (Figure S10). (D) Total mitochondrial mass in NK cells after 4 h exposure to 400 pg/cell PMs in the absence or presence of 0.5 mM Asc, measured by flow cytometry (N = 3). (E) LysoTracker fluorescence intensity in NK cells after 4 h exposure to 100 pg/cell PMs, measured by flow cytometry (N = 5). (F) Dose-dependent changes in LysoTracker fluorescence intensity across different PM types after 4 h exposure. (G) Comparison of LysoTracker fluorescence intensity in the absence and presence of 0.5 mM Asc after 4 h exposure to 400 pg/cell PMs (N = 3). The fold changes shown in panel G were calculated as [value]_PM/[value]_Control. (H) IFN-γ production under K562 target-cell challenge conditions in the absence or presence of Asc supplementation following PM exposure. Data in panel H were analyzed using ordinary two-way ANOVA with PM treatment and Asc supplementation as factors, including their interaction. Within-treatment comparisons between the no-Asc and Asc groups were performed using Šídák’s multiple-comparisons test. Statistical significance is indicated as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, or otherwise not significant (n.s.).

To further investigate these possibilities, we first quantified total intracellular glutathione (GSH) levels. Fluorometric analysis showed no statistically significant differences among PM-treated groups after 4 h of 400 pg/cell PM exposure (Figure S7), suggesting that the total GSH pool was maintained despite early ROS fluctuations, likely through dynamic GSH consumption. , The slightly lower GSH level in cells treated with Rock (Fe) PM (p > 0.05) aligns with the pronounced ROS fluctuation (from 178% to 88%) observed during this period (Figure A) and the well-established iron redox cycling that accelerates localized GSH oxidation. The preserved total GSH pool suggests that early NK-cell injury in Figure A is unlikely to result from substantial depletion of the total cellular glutathione reserve or cytosolic oxidative overload but may instead involve mitochondrial stress, as mitochondria can both generate ROS during acute stress and undergo depolarization, fragmentation, and lysosome-mediated clearance under sustained damage. ,

To evaluate the role of mitochondrial dysfunction, we applied both confocal imaging and flow cytometry to visualize and quantify the mitochondrial membrane potential and mass. Confocal imaging with MitoTracker Green and LysoTracker Deep Red was conducted on Kaolin for its elevated apoptosis and necrosis proportion (Figure , p < 0.001) and considerable variation in cytotoxic proteins, IFN-γ, IL-10, and intracellular ROS (Figures A–F and Figure A) at early stages. PM-exposed NK cells displayed increased DAPI positivity and numerous small, diffuse MitoTracker Green puncta (Figure B), suggesting compromised plasma-membrane integrity and altered mitochondrial morphology. The increased DAPI positivity is consistent with the increased PI-positive late apoptotic/necrotic populations (p < 0.001) observed in Figure . Increased mitochondrial–lysosomal colocalization, shown by yellow merged signals, further suggests enhanced interaction between damaged mitochondria and lysosomes in Kaolin-treated NK cells. , Flow cytometry analysis of live cells showed a broad dose-dependent decrease in ΔΨm across PM-treated groups (p < 0.0001, Figure C and Figure S8), indicating mitochondrial depolarization and impaired bioenergetic function. This was accompanied by reduced total mitochondrial mass (at least p < 0.05, Figure D and Figure S9), suggesting possible enhanced clearance of damaged mitochondria, consistent with the increased apoptosis/necrosis (Figure ) and altered ROS profile (Figure A). These provide a plausible explanation for the decreased CD107a surface expression observed in Figure G, as impaired mitochondrial function limits intracellular ATP availability required for granule polarization and exocytosis. , Among the different PM types, Rock (Fe), Rock (Si), and Kaolin exhibited both lower ΔΨm and reduced mitochondrial mass, likely reflecting oxidative burst-induced mitochondrial damage (Figure A).

Lysosomes, including the secretory lysosomes that store perforin and granzyme B, are LAMP1⁺ acidified organelles (pH ∼ 4.5–5.5) whose acidity is maintained by the ATP-driven H⁺ pump (V-ATPase). Such an acidic milieu is crucial for protease function and membrane trafficking, , ensuring their proper maturation and safe storage before regulated exocytosis. Flow cytometry analysis of lysosomal acidity using LysoTracker Deep Red revealed a dose- and PM type-dependent decrease in LysoTracker intensity (Figures E and F), indicating a loss of acidity and/or compartment integrity. For example, only 4 h of 100 pg/cell PM exposure resulted in at least a 40% reduction in the acidity compared to the control group (p < 0.0001), though no statistical difference was observed among all PM-treated groups (Figure E). Upon increasing the PM dose to 800 pg/cell, LysoTracker fluorescence fell to <50% in SiO₂_S (p < 0.0001) and to ∼10% of the control in Rock (Si) (p < 0.0001) within 4 h (Figure F). The loss of lysosomal acidity may be associated with mitochondrial dysfunction (Figures C–D), which can impair ATP-dependent V-ATPase activity and reduce LysoTracker intensity.

To further examine whether and how ROS contributed to organelle dysfunction, 0.5 mM ascorbic acid (Asc) was added to NK cells during exposure to PM at 400 pg/cell. Asc supplementation partially restored mitochondrial mass in Rock (Si) and Kaolin-treated groups (Figure D, p < 0.05) and significantly restored lysosomal acidity in all PM-treated groups (at least p < 0.05), except for Rock (Fe) (p > 0.05) (Figure G). The lack of lysosome acidity recovery in Rock (Fe) may reflect iron redox cycling-mediated Asc consumption, which could limit its antioxidant effect against iron-rich PM-induced oxidative stress. Together with the unchanged LAMP1 levels (Figure S6) and the faint or diffuse lysosomal puncta observed by confocal imaging (Figure B), these results suggest that PM exposure primarily disrupts lysosomal acidification and/or compartment integrity via oxidative burst rather than reducing lysosomal abundance. Such lysosomal dysfunction may impair the intracellular processing, storage, and trafficking of cytotoxic granule proteins. − Consistent with this possibility, silica/silicate-rich PMs were associated with changes in the lower-molecular-weight (∼26 kDa) Granzyme B species, the 26 kDa/32 kDa Granzyme B band-intensity ratio, and perforin expression (Figures E–F and Figure S4). Functionally, two-way ANOVA showed significant main effects of PM treatment (F(5,24) = 62.87, p < 0.0001) and Asc supplementation (F(1,24) = 116.09, p < 0.0001) on K562-stimulated IFN-γ production, with a significant PM treatment × Asc interaction observed (F(5, 24) = 2.64, p = 0.0485) (Figure H), indicating that the magnitude of the Asc-associated increase differed among PM treatments. Collectively, the Asc-associated improvements in mitochondrial mass, lysosomal acidity, and K562-stimulated IFN-γ production support a contribution of ROS-associated organelle dysfunction to PM-associated impairment of NK-cell effector responses. However, the incomplete and particle-dependent responses suggest that ROS is unlikely to be the sole driver and that additional PM-specific mechanisms may also contribute.

3.5. Mechanistic Links between PM Elemental Composition and NK Cell Dysfunction

Respirable PMs induced NK-cell dysfunction through convergent yet composition-dependent adverse outcome pathways (AOPs). Following endocytosis, the negatively charged PMs investigated in this study are likely trafficked into the endo-lysosomal system, where early local ROS generation may destabilize lysosomal membranes and disrupt lysosomal acidification (Figure G). The preserved total intracellular GSH pool together with the partial rescue of mitochondrial function by Asc argues against global oxidative overload and, instead, supports compartmentalized ROS-associated organelle stress as an early initiating event. Lysosomal dysfunction, including reduced acidity and compromised granule integrity, may directly contribute to and/or be amplified by the observed reduction in mitochondrial mass and pronounced mitochondrial depolarization after PM exposure. The resulting mitochondrial dysfunction can limit ATP availability required for NK-cell degranulation and cytotoxic responses, while also contributing to apoptosis and necrosis through mitochondrial stress pathways. In turn, reduced ATP supply may further impair V-ATPase-dependent lysosomal acidification, reinforcing a bidirectional lysosome–mitochondria stress loop. This organelle stress axis may contribute to the particle-dependent alterations in perforin and the lower-molecular-weight (∼26 kDa) Granzyme B species observed following exposure to silica/silicate-rich PMs, particularly at higher doses.

Among the three PM subcategories, silica/silicate-rich particles produced the strongest suppression of NK-cell effector function, which was closely associated with pronounced mitochondrial depolarization and lysosomal deacidification, particularly at higher PM doses (Figures C–G). This enhanced immunotoxicity may be linked to the highly reactive silica/silicate surface chemistry, including stronger negative surface charge from deprotonated silanol/aluminol groups and defect-rich silanol structures that can destabilize biological membranes and promote persistent endo-lysosomal stress following particle internalization. ,, Such surface reactivity has been widely implicated in silica-induced membranolysis, lysosomal dysfunction, and inflammasome activation in phagocytic cells. − Consistently, the silica/silicate-rich PMs in this study produced the greatest disruption in lysosomal acidity and mitochondrial integrity, together with the strongest suppression of NK-cell effector responses. Interestingly, despite being silica-based, SiO₂_S showed relatively preserved mitochondrial mass and LysoTracker intensity compared with other silica/silicate PMs, particularly at elevated concentrations. This weaker effect may reflect its higher amorphous content and lower ROS-generating capacity, as amorphous silica can still be toxic but may lack the persistent crystalline or defect-rich surface reactivity that drives stronger lysosomal destabilization and organelle dysfunction. ,, The particularly strong cytotoxicity of Kaolin may further reflect its platy morphology, which could increase particle–cell contact area and cause greater membrane disturbance during internalization compared with more rounded particles, consistent with previous reports of morphology-dependent particle cytotoxicity.

By contrast, the iron-bearing PM Rock (Fe), although highly effective at inducing ROS generation and necrotic cell loss, caused comparatively weaker suppression of the NK cell effector function. This suggests that iron-rich PMs may preferentially induce acute oxidative and membrane-compromising injury through iron redox cycling, whereas silica/silicate-rich particles more strongly disrupt organelle systems directly linked to NK effector function, particularly lysosome-like cytotoxic granules and mitochondria. A similar pattern may also explain the weaker functional impairment observed for coal. Collectively, the extent of cell death alone did not fully predict NK functional impairment, instead, the dominant mode of injury appeared to depend on the PM composition.

While these findings support an important role for elemental composition in shaping NK-cell responses, causality should be interpreted with caution. Endotoxin or other non-viable bioactive contaminants cannot be fully excluded, despite exposures being conducted under UV treatment and antibiotic-supplemented conditions. In addition, differences in particle morphology, surface area, and aggregation/agglomeration in suspension may have influenced effective particle–cell interactions. The present study was also limited to the tested dose range and exposure duration. Future studies using lower concentrations and longer or repeated exposure scenarios would be beneficial for better defining the dose- and time-dependent progression of PM-induced NK-cell dysfunction. As such, the present findings should be interpreted as mechanistic and comparative evidence rather than as a direct quantitative assessment of real-world exposure risk. Another limitation is that the study used NK-92MI cells as a controlled and reproducible NK-cell model. Although this model is useful for mechanistic comparison across particle types, it does not fully recapitulate the heterogeneity of primary human NK cells or tissue-resident NK cells in vivo. Further validation using primary human NK cells and more lung-relevant co-culture systems, such as NK cells co-cultured with lung epithelial cells and/or macrophage-containing models, would strengthen the physiological and translational relevance of the findings. Such models would also help clarify how interactions among inhaled particles, macrophages, epithelial cells, and NK cells contribute to immune dysfunction in the pulmonary microenvironment. Collectively, these considerations indicate that elemental composition is an important, but not exclusive, determinant of the observed NK-cell responses.

4. Environmental Implications

Although increasing epidemiological and occupational evidence links respirable particulate matter (PM) exposure, especially crystalline silica, coal, and iron-bearing dusts, to elevated risks of pneumoconiosis and lung cancer, the mechanistic pathways connecting PM inhalation to impaired immune surveillance remain poorly defined. Our findings suggest that exposure to practical in vitro concentrations of respirable PM is associated with compromised cytotoxicity and cytokine production in NK cells via the ROS-lysosome-like organelles-mitochondrial axis. These changes may impair immune surveillance and provide a potential mechanistic link between particulate exposure and early lung carcinogenesis. By linking PM exposure to NK cell dysfunction, this work proposes the first integrated adverse outcome pathway framework connecting respirable PM exposure with innate immune suppression and increased lung cancer susceptibility. From a policy perspective, the findings support the incorporation of particle composition and immune-relevant biological effects into future dust hazard assessment and control frameworks rather than relying solely on mass-based exposure metrics.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.6c07515.

• Chemicals and reagents; respirable PM sample preparation procedure; analysis procedures of key PM properties; justification of PM exposure concentrations and experimental durations; cell treatments; Western blot analyses; apoptosis and necrosis measurement; intracellular reactive oxygen species (ROS) analysis; CD107a degranulation assay; measurement of mitochondrial depolarization and total mitochondrial mass; measurement of lysosome-like organelle function status; flow cytometric measurement of NK-mediated killing of K562 cells under continuous PM exposure; particle size distribution; R Bragg values of individual phases; agreement indices of XRD analyses; details of the mean ± SD value for all the cellular state data; details of the mean ± SD value for the ROS data; normality test and Q-Q plots for all the data used in this work; XRD Rietveld refinement results of PM samples; SEM image of the kaolinite sample; expression of granzyme B molecular forms and the 26 kDa/32 kDa band-intensity ratio; schematic diagram of the measurement procedure of CD107a; effects of PM exposure on the expression of total LAMP1 in NK cells; intracellular GSH levels after 4 h of PM exposure; dose-dependent reduction in mitochondrial membrane potential across all PM types; dose-dependent reduction in the mass of mitochondria across all PM types; positive control for intracellular ROS and mitochondrial depolarization (PDF)

The authors declare no competing financial interest.