All key reagents used in this study and their corresponding catalog numbers are outlined in Supplementary Table 1 .

Experimental protocols were approved by the Institutional Animal Care and Use Committee at Michigan State University (MSU) (Animal Use Form [AUF] #PROTO201800113) in accordance with guidelines established by the National Institutes of Health. Six- to 8-wk-old C57BL/6 mice (cat. #000664) were procured from the Jackson Laboratory (Bar Harbor, ME) and were given freely accessible food and water. Animal facilities were maintained with a 12 h light/dark cycle at consistent temperature (21-24°C) and humidity (40-55%). Mice were bred for FLAM isolation as previously described (43, 44). At 14-18 gestational days, pregnant dams (8-10 wk of age) were euthanized by CO₂ asphyxiation for 10 min to ensure death to the neonatal mice (14-18 gestational days), which are resistant to anoxia. As a secondary method of euthanasia, cervical dislocation was performed on the dam and blood supply was cut off to the fetuses. Upon removing the fetuses from the dam, fetal livers were carefully removed and dissociated in ice-cold DPBS^-/- (DPBS without calcium and magnesium) by gentle pipetting. Liver cells were pelleted by centrifugation at 220 x g for 5 min at 4°C, resuspended in complete FLAM medium (RPMI 1640 medium, 10% fetal bovine serum, 1% penicillin-streptomycin, 30 ng/ml recombinant mouse GM-CSF, and 20 ng/ml recombinant human TGF-β1), seeded in 100 mm tissue-culture treated dishes, and incubated at 37°C and 5% CO₂. Medium was changed every 1-2 d and non-adherent dead cells discarded. Adherent fetal liver monocytes were lifted with DPBS^-/- containing 10 mM EDTA and gentle scraping upon reaching 70-90% confluency. After 1-2 wk of culture, cells developed a round morphology resemblant of AMs and were frozen for future use. Mature, differentiated FLAMs were routinely confirmed to be CD14⁺ low, SiglecF⁺ high, and CD11c⁺ high by flow cytometry as previously described (43). Cells between passage 5 and 10 were used for this study.

FLAMs were seeded in 6-well plates at a density of 4.50×10⁵ cells/well in complete FLAM medium. Cells were incubated overnight to achieve 70-90% confluency before beginning treatments. The next day, cells were washed once with sterile DPBS^-/-, and were then incubated for 24 h in fresh complete FLAM medium containing either 1) ethanolic DHA (25 µM) or ethanol (EtOH) vehicle (VEH), or 2) 25 µM DHA complexed to BSA at a 3:1 ratio as previously described (27, 28) or 8.33 µM BSA VEH. Following treatments, FLAMs were pelleted and stored in 100% methanol at -80°C before total fatty acids were measured at OmegaQuant Inc. using gas chromatography (GC) with flame ionization detection as previously described (45).

A standard mixture of fatty acids (GLC OQ-A; NuChek Prep, Elysian, MN) was used to generate calibration curves for individual fatty acids and identify fatty acids present in each sample. The following 24 fatty acids were identified (by class): i) saturated (14:0, 16:0, 18:0, 20:0, 22:0, 24:0), ii) cis monounsaturated (16:1, 18:1, 20:1, 24:1), iii) cis ω-6 polyunsaturated (18:2ω6, 18:3ω6, 20:2ω6, 20:3ω6, 20:4ω6, 22:4ω6, 22:5ω6), and iv) cis ω-3 polyunsaturated (18:3ω3, 20:5ω3, 22:5ω3, 22:6ω3). Proportions of individual fatty acids in each sample were expressed as a percentage of total identified fatty acids.

Two biomarkers of ω-3 tissue incorporation, percent ω-3 PUFAs and highly unsaturated fatty acid (HUFA) score, were calculated as previously described (23) using the following equations, with total fatty acids (FA) equaling the sum of all analyzed FA and total HUFA equaling the sum of dihomo-γ-linolenic acid (DGLA; 20:3ω6), ARA (20:4ω6), eicosatetraenoic acid (EPA; 20:5ω3), docosatetraenoic acid (DTA; 22:4ω6), ω-6 docosapentaenoic acid (DPA; 22:5ω6), ω-3 DPA (22:5ω3), and DHA (22:6ω3):

For lipidomics analyses, FLAMs were seeded in 6-well plates at a density of 4.50×10⁵ cells/well in complete FLAM medium. Cells were incubated overnight to achieve 70-90% confluency before beginning treatments. The next day, cells washed once with sterile DPBS^-/-, fresh complete FLAM medium containing either 25 µM ethanolic DHA or EtOH VEH, then cells incubated for 24 h. Following DHA or VEH treatment, cells were washed once with sterile DPBS^-/-, treated with either 20 ng/ml LPS in DPBS^-/- or DPBS^-/- VEH in DPBS^+/+ (DPBS containing calcium and magnesium) for 2 h, then exposed to 12.5 µg/cm² cSiO₂ or DPBS^-/- VEH for 1.5 or 4 h. These timepoints were selected for cSiO₂ exposure to assess acute toxic effects of the particle on FLAMs instead of prolonged effects. LPS and cSiO₂ exposures were done in DPBS^+/+ to minimize interference from fatty acids present in cell culture medium during oxylipin analyses. Each treatment condition was tested using three biological replicates and one technical replicate.

Cell culture supernatants and cells were collected at 0, 1.5, and 4 h after cSiO ₂ treatment and pooled together prior to lipidomic analysis. Following treatments, ice-cold methanol was added to each well, resulting in a final sample volume of 3 ml (1 ml cell culture supernatant + 2 ml methanol). To each sample, 60 µl of antioxidant cocktail (0.2 mg/ml butylated hydroxytoluene, 0.2 mg/ml triphenylphosphine, 0.6 mg/ml EDTA) was added to achieve a total cocktail concentration of 5% (v/v) (46). After subjecting the cells and supernatant to centrifugation at 0°C and 10,000 x g to separate the cellular debris, the resulting supernatant was stored in tubes at -80°C under nitrogen gas. Cells and supernatants within each well were pooled together, then samples were frozen at -80°C until liquid chromatography-mass spectrometry (LC-MS) analysis. Targeted LC-MS lipidomics for 156 lipid metabolites was conducted at the Lipidomics Core Facility at Wayne State University as previously described (47–49). Briefly, 100 µl aliquots of cellular samples were thawed and spiked with a cocktail of deuterated internal standards (5 ng each of PGE₁-d₄, RvD2-d₅, LTB₄-d₄, and 15[S]-HETE-d₈; Cayman Chemical, Ann Arbor, MI) for quantification of oxylipins and recovery. Then, oxylipins were extracted by using C18 extraction columns that were washed with 15% (v/v) methanol and subsequently hexane, dried in a vacuum, eluted with methanol containing 0.1% (v/v) formic acid, dried under nitrogen gas, and dissolved in a 1:1 mixture of methanol:25 mM aqueous ammonium acetate. Extracted oxylipins were subjected to high-performance liquid chromatography (HPLC) using a Luna C18 (3 µm, 2.1×150 mm) column connected to a Prominence XR system (Shimadzu, Somerset, NJ) then analyzed with a QTrap5500 mass spectrometer (AB Sciex, Singapore) set to negative ion mode. Analyst 1.6 software (AB Sciex) and MultiQuant software (AB Sciex) were used to collect and quantify the data in units of ng, respectively. Lipid metabolite classifications are provided in Supplementary Table 2 , and mass spectra of representative metabolites are given in reference (50).

FLAMs were seeded in 24-well plates at a density of 1.50×10⁵ cells/well in complete FLAM medium and then cultured with DHA or VEH, LPS or VEH, and cSiO₂ or VEH under the conditions described for the lipidomics study. LPS and cSiO₂ exposures were done in DPBS^+/+ to maintain consistent experimental conditions between oxylipin analyses and other analyses conducted in this study. Cell culture supernatants were collected by centrifugation at 200 x g at 0, 1.5, and 4 h post cSiO₂ exposure and analyzed for proinflammatory cytokines, cathepsin B activity, and lactate dehydrogenase (LDH) release.

Cathepsin B activity was determined using a fluorescent assay as previously described (51). Briefly, 50 µl of cell culture supernatant and 2 µg Z-LR-AMC were combined in 96-well plates and adjusted to a final volume of 150 µl/well, then all samples were incubated for 1 h at 37°C. Sample fluorescence was then measured using a FilterMax F3 Multimode plate reader (Molecular Devices, San Jose, CA) set to 380 nm excitation and 460 nm emission. Cathepsin B activity in each well was calculated in units of relative fluorescence units (RFU) by the following equation:

LDH activity was measured as previously described (27, 28). Separate wells of untreated FLAMs were included and designated as max-kill (MK) samples and incubated with 0.2% Triton-X (Millipore Sigma) for 5 min. After all treatments, supernatants were collected from all wells and 50 µl was transferred to non-treated, flat-bottom 96-well plates in duplicate. DPBS^+/+ was used as a sample blank, and DPBS^+/+ containing 0.2% (v/v) Triton-X was used as a MK blank. 100 µl of LDH assay reagent (15 µM 1-methoxyphenazine methosulfate [PMS], 2 mM iodonitrotetrazolium [INT], 3.2 mM β-nicotinamide adenine dinucleotide [NAD] sodium salt, and 160 mM lithium lactate in 0.2 M Tris-HCl, pH 8.2) was added to each well, and assay plates were incubated at room temperature for 15 min in the dark. Sample absorbance was then measured using a FilterMax F3 Multimode plate reader (Molecular Devices, San Jose, CA) set to a wavelength of 492 nm. Percent cell death in each well was calculated by the following equation:

Concentrations of proinflammatory cytokines (i.e., IL-1α, IL-1β, TNF-α) were quantified by enzyme-linked immunosorbent assay (ELISA) using corresponding mouse R&D Systems DuoSet kits according to the manufacturer’s instructions.

FLAMs were seeded in 48-well plates 0.625×10⁵ cells/well in complete FLAM medium to achieve 50% confluency after overnight incubation. The next day, cells were washed once with sterile DPBS^-/- then treated with either 25 µM ethanolic DHA or EtOH VEH as a control in complete FLAM medium. After 24 h, cells were washed once with sterile DPBS^-/- and primed with 20 ng/ml LPS or PBS VEH for 1.5 h. Following LPS priming, cells were washed with sterile DPBS^+/+ then subsequently incubated for 30 min at 37°C in the dark with 50 nM LysoTracker Red DND-99 in DPBS^+/+ to label lysosomes, 25 nM MitoTracker Red CMXRos to label mitochondria, or 200 nM SYTOX Green to detect cell death. All fluorescent dyes were diluted in DPBS^+/+ prior to addition to the cells, and 20 ng/ml LPS was simultaneously added to appropriate wells to allow 2 h of total LPS priming prior to cSiO₂ treatment. After 30 min of cell staining, a freshly prepared cSiO₂ stock suspension was added dropwise to a final concentration of 0 or 12.5 µg/cm². LPS and cSiO₂ exposures were done in DPBS^+/+ to maintain consistent experimental conditions between oxylipin analyses and other analyses conducted in this study.

Live-cell fluorescence microscopy began immediately after adding cSiO₂ to the cells. Images of live cells were taken at 0, 1.5, and 4 h post cSiO₂ treatment using an EVOS FL Auto 2 Cell Imaging System (ThermoFisher Scientific) with an onstage, temperature-controlled incubator. Each experimental condition was tested using 3 biological replicates. For each well, 2-4 images were acquired using fields of view defined before the beginning of the experiment. LysoTracker Red and MitoTracker Red were detected using the Texas Red light cube (Ex: 585/29 nm, Em: 628/32 nm), and SYTOX Green was detected using the GFP light cube (Ex: 482/25 nm, Em: 524/24 nm).

Acquired images were analyzed for lysosomal integrity, mitochondrial integrity, and cell death using CellProfiler 4.2.1 as previously described (43, 52). Briefly, lysosomal integrity, mitochondrial integrity, and cell death were assessed by quantifying the number of LysoTracker Red⁺, MitoTracker Red⁺, and SYTOX Green⁺ cells, respectively. LysoTracker Red⁺ and MitoTracker Red⁺ puncta were omitted if fluorescent intensity fell below preset minimum thresholds, which were chosen to omit false positives quantified from background fluorescence. Raw counts of LysoTracker Red⁺, MitoTracker Red⁺, and SYTOX Green⁺ cells were further analyzed using RStudio 2022.07.1 + 554 (Posit, Boston, MA).

For oxylipin data, MetaboAnalyst Version 5.0 (Xia Lab, Quebec, Canada, www.metaboanalyst.ca/) (53) was used to conduct statistical analyses. First, raw ng values were converted to corresponding pmol values in Microsoft Excel (50). Then, in MetaboAnalyst, the one factor statistical analysis module was chosen, and data were uploaded as a comma separated values (.csv) file with samples in unpaired columns and features (i.e., metabolites) in rows. Features with >70% missing data were removed from the dataset, and remaining missing values were estimated by replacing with the corresponding limits of detection (LODs; 1/5 of the minimum positive value of each variable). After the data was cleaned, the data was normalized by auto scaling only, then the data editor option was used to select experimental groups of interest to compare. For comparisons between experimental groups, one-way analysis of variance (ANOVA) (FDR = 0.05) followed by Tukey’s honestly significant difference (HSD) post-hoc test was used, with FDR q<0.05 considered statistically significant. Asterisks (*) indicate statistically significant differences (FDR q<0.05) for cSiO₂-treated groups and their corresponding non-cSiO₂-treated controls. Hashes (#) indicate statistically significant differences (FDR q<0.05) for DHA-treated groups and their corresponding non-DHA-treated controls. Crosses (†) indicate statistically significant differences (FDR q<0.05) for LPS-treated groups and their corresponding non-LPS-treated controls.

For all other endpoints, GraphPad Prism Version 9 (GraphPad Software, San Diego, CA, www.graphpad.com) was used to conduct statistical analyses. The ROUT outlier test (Q = 1%) and the Shapiro-Wilk test (p<0.01) were used to identify outliers and assess normality in the data, respectively. For comparisons between two experimental groups, non-normal and semiquantitative data were analyzed by the Mann-Whitney nonparametric test. The F test (p<0.05) was used to test the assumption of equal variances across both groups. Normal data with unequal variances were analyzed using an unpaired t test with Welch’s correction. Normal data that met the assumption of equal variance were analyzed using an unpaired t test. For comparisons between more than two experimental groups, non-normal and semi-quantitative data were analyzed by the Kruskal-Wallis nonparametric test followed by Dunn’s post-hoc test. The Brown-Forsythe test (p<0.01) was used to test the assumption of equal variances across treatment groups. Normal data with unequal variances were analyzed using the Brown-Forsythe/Welch analysis of variance (ANOVA) followed by Dunnett’s T3 post-hoc test. Normal data that met the assumption of equal variance were analyzed by standard one-way ANOVA followed by Tukey’s post-hoc test. Data are presented as mean ± standard error of the mean (SEM), with a p-value < 0.05 considered statistically significant.

Two major methods for introducing ω-3 PUFAs to cell cultures, ethanolic suspensions and BSA complexes (54), were evaluated for their suitability for incorporating DHA into FLAMs as shown in Figure 1A . Ethanolic DHA-treated FLAMs had significantly greater DHA content (19.3% total fatty acids) compared to EtOH VEH-treated FLAMs (4.4% total fatty acids) ( Figure 1B ). Corresponding with these findings, the ω-9 monounsaturated fatty acid (MUFA) oleic acid (OA), the main fatty acid found in fetal bovine serum (55), and the ω-6 PUFA arachidonic acid (ARA), the major precursor PUFA for eicosanoid biosynthesis, were significantly decreased in DHA-treated FLAMs (16.5% and 7.2% total fatty acids, respectively) compared to VEH-treated FLAMs (26.5% and 10.6% total fatty acids, respectively). No notable changes were found for other saturated and unsaturated fatty acids that were analyzed. FLAMs incubated with DHA-BSA complexes or BSA VEH displayed similar DHA membrane incorporation at the expense of OA and ARA ( Figure 1C ) to that seen for ethanolic DHA ( Figure 1B ).

Total fatty acid findings were related to two biomarkers: i) percent ω-3 fatty acids, which is the sum of eicosapentaenoic acid (EPA) and DHA as a percentage of all measured FA, and ii) ω-3 HUFA score, which equals the sum of ω-3 HUFAs (i.e., EPA, ω-3 DPA, and DHA) as a percentage of all measured ω-3/6 HUFAs (i.e., 20:3ω6, 20:4ω6, 20:5ω3, 22:4ω6, 22:5ω6, 22:5ω3, and 22:6ω3) (23). In FLAMs treated with ethanolic DHA, percent ω-3 fatty acids was 22%, while in VEH-treated FLAMs it was 6% ( Figure 1D ). FLAMs treated with ethanolic DHA also demonstrated a comparatively higher ω-3 HUFA score (69%) compared to VEH-treated FLAMs (38%). Similarly, FLAMs treated with DHA-BSA complexes exhibited significant increases in percent ω-3 fatty acids (18%) and ω-3 HUFA score (70%) compared to VEH-treated cells (5% and 39%, respectively) ( Figure 1E ). Based on its relative simplicity, ethanolic DHA delivery was used for all subsequent studies.

The effects of cSiO₂ on the combined intracellular and extracellular lipidome in VEH- and DHA-treated FLAMs was compared in unprimed and LPS-primed FLAMs ( Figure 2A ) and are illustrated in Figure 2B and Supplementary Figure 1 . In addition, summarized oxylipin quantities ( Supplementary Tables 3 - 5 ), and individual oxylipin quantities (50) were compared between experimental groups at each designated timepoint. Oxylipin profile shifts were highly pronounced following cSiO₂ exposure ( Figure 2B ); therefore, we focused our analysis on the 1.5 h and 4 h post cSiO₂ timepoints. cSiO₂ elevated total levels of ARA-derived oxylipins and DHA-/EPA-derived oxylipins produced from VEH- and DHA-treated FLAMs, respectively ( Figure 3A , Supplementary Tables 4, 5 ). In DHA-treated FLAMs, cSiO₂-induced levels of ARA-derived oxylipins were significantly decreased, and, correspondingly, cSiO₂-induced levels of DHA-/EPA-derived oxylipins were significantly elevated. cSiO₂-induced levels of EPA-derived oxylipins also increased modestly in VEH-treated FLAMs. LPS priming elicited a marked increase in cSiO₂-triggered ARA-derived oxylipins production in VEH-treated FLAMs and a marked decrease in cSiO₂-triggered EPA-derived oxylipins levels in DHA-treated FLAMs. Findings at 4 h post cSiO₂ reflected those found at 1.5 h with higher overall quantities of ARA-, EPA-, and DHA-derived metabolites.

At 1.5 h and 4 h, cSiO₂ induced robust increases in total prostaglandins ( Figure 3B ), leukotrienes ( Figure 3C ) and thromboxanes ( Figure 3D ) compared to VEH-treated FLAMs ( Supplementary Tables 4 , 5 ). Levels of representative oxylipins from each metabolite class including PGE2, PGJ2, PGD2, PGF2α, and PGA2 ( Figure 4A ), LTB4 ( Figure 4B ), and TXB2 ( Figure 4C ) were increased in like manner in the presence of cSiO₂ (50).

LPS priming augmented cSiO₂-triggered production of total and individual prostaglandins, leukotrienes, and thromboxanes ( Figures 3B-D , 4 ). DHA significantly reduced cSiO₂-induced prostaglandin, leukotriene, and thromboxane production, yet induction of these metabolites was still significant compared to baseline levels in DHA-treated FLAMs. Interestingly, LPS priming did not significantly impact cSiO₂-induced levels of prostaglandins, leukotrienes, and thromboxanes in DHA-treated FLAMs.

Total hydroxy fatty acid (HFA) metabolites were significantly increased by cSiO₂ exposure in both VEH-treated and DHA-treated FLAMs at 1.5 h and 4 h ( Figure 5A , Supplementary Tables 4 , 5 ). DHA supplementation significantly reduced ω-6 HFAs and increased ω-3 HFAs at both timepoints ( Figures 5B, C ). The suppressive effect of DHA on ω-6 HFA levels was more marked at 1.5 h than at 4 h, while ω-3 HFA levels steadily increased in DHA-treated FLAMs over time. LPS priming further increased levels of cSiO₂-induced ω-6 HFAs at 1.5 h and decreased levels of cSiO₂-induced ω-3 HFAs at 4 h.

Observed changes in ARA-derived HFAs reflected those in total ω-6 HFAs ( Figures 5B, D , Supplementary Tables 4 , 5 ). cSiO₂ triggered significant increases in ARA-derived HFA levels in both VEH-treated and DHA-treated FLAMs ( Figure 5D ). DHA significantly reduced metabolite levels, and LPS priming further potentiated cSiO₂-induced metabolite production at 1.5 h and 4 h. Likewise, DHA- and EPA-derived HFA levels reflected total levels of ω-3 HFAs, as these significantly increased in DHA-treated cells exposed to cSiO₂ starting at 1.5 h and continuing at 4 h ( Figures 5C, E , Supplementary Tables 4 , 5 ). In both VEH-treated FLAMs and DHA-treated FLAMs, quantities of ARA-, EPA-, and DHA-derived HFAs were ranked as follows for both timepoints: ARA > DHA > EPA ( Figures 5D-F ). Furthermore, ARA-derived HFAs accounted for the majority of total measured ω-6 HFAs, whereas EPA and DHA both accounted for the majority of total measured ω-3 HFAs.

cSiO₂-induced ω-6 HFAs in VEH-treated FLAMs at 1.5 h and 4 h consisted primarily of ARA-derived 5-HETE, 8-HETE, 9-HETE, 11-HETE, 12-HETE, and 15-HETE ( Figure 6 ) (50). Relative abundance of selected HETEs was: 5-HETE > 11-HETE > 15-HETE > 8-HETE > 12-HETE > 9-HETE. HETE levels also increased in DHA-treated FLAMs exposed to cSiO₂ but were found to be significant only for 5-HETE, 8-HETE, and 9-HETE at 4 h and 15-HETE at both timepoints. In line with total ARA-derived HFA levels ( Figure 5D ), DHA significantly suppressed 8-HETE at 4 h and 9-HETE at both timepoints ( Figure 6 ). Levels of other cSiO₂-induced HETEs (e.g., 5-HETE, 11-HETE, 12-HETE, 15-HETE) also were reduced in DHA-treated FLAMs, but the findings were not statistically significant. LPS priming elicited a significant increase in cSiO₂-induced 8-HETE and a non-significant trend toward increased levels of other HETEs in VEH-treated FLAMs at both timepoints.

In DHA-treated FLAMs, cSiO₂ induced significant increases in several DHA-derived HDoHEs (i.e., 4-HDoHE, 7-HDoHE, 8-HDoHE, 10-HDoHE, 11-HDoHE, 14-HDoHE, 16-HDoHE, 17-HDoHE, 20-HDoHE) ( Figure 7 ) and in several EPA-derived HEPEs (i.e., 5-HEPE, 8-HEPE, 9-HEPE, 11-HEPE, 12-HEPE, 15[S]-HEPE) at both timepoints ( Supplementary Figure 2 ) (50). Relative abundance of selected HDoHEs was: 4-HDoHE ≈ 20-HDoHE > 16-HDoHE > 7-HDoHE ≈ 8-HDoHE ≈ 10-HDoHE ≈ 11-HDoHE > 14-HDoHE > 17-HDoHE and relative abundance of selected HEPEs was: 5-HEPE > 8-HEPE ≈ 11-HEPE > 15(S)-HEPE > 12-HEPE > 9-HEPE. While cSiO₂ exposure led to significant increases in EPA-derived 5-HEPE, 11-HEPE, and 12-HEPE in VEH-treated cells during the time-course, DHA-derived HDoHEs did not undergo significant increases in VEH-treated cells exposed to cSiO₂. The effects of LPS priming on cSiO₂-induced HEPEs and HDoHEs were minimal at 1.5 h and 4 h.

Specialized pro-resolving mediators (SPMs) are a class of oxylipins comprised of resolvins, maresins, protectins, and lipoxins derived from ARA, EPA, ω-3 DPA, and DHA that limit proinflammatory cytokine release and promote dead cell clearance by macrophages (56–58). Most SPMs assessed in our LC-MS oxylipin panel were not detected at any timepoint (50). DHA supplementation did, however, cause a modest increase in RvD6 (4,17-DiHDoPE) and MaR1_{ω-3 DPA} at 1.5 h and 4 h ( Figure 8 ) (50). In DHA-treated FLAMs, cSiO₂ exposure did not significantly influence production of RvD6 at either timepoint ( Figure 8A ) but significantly increased MaR1_{ω-3 DPA} at 1.5 h and decreased MaR1_{ω-3 DPA} at 4 h ( Figure 8B ). LPS priming significantly suppressed MaR1_{ω-3 DPA} production in DHA-treated FLAMs at 4 h, modestly inhibited MaR1_{ω-3 DPA} at 1.5 h, and modestly decreased RvD6 production at both 1.5 h and 4 h.

Total epoxy fatty acids (EpFAs) and CYP450-derived dihydroxy fatty acids (DiHFAs) were quantified from VEH-treated and DHA-treated FLAMs ( Figure 9A , Supplementary Tables 4 , 5 ). In VEH-treated FLAMs, cSiO₂ modestly induced production of EpFA metabolites at 1.5 h and 4 h and did not significantly impact production of DiHFA metabolites. Conversely, cSiO₂ triggered significant increases in total EpFAs and DiHFAs in DHA-treated FLAMs at both timepoints. Interestingly, LPS priming significantly reduced total DiHFA metabolite levels in the absence of cSiO₂ in DHA-treated FLAMs. Separate and simultaneous LPS priming and cSiO₂ exposure elicited modest increases in EpFA : DiHFA ratios in both VEH-treated and DHA-treated FLAMs during the time-course.

Effects of cSiO₂ and DHA were also analyzed for selected CYP450 oxylipin products of ARA (i.e., 14,15-EpETrE, 14,15-DiHETrE) and DHA (i.e., 19,20-EpDPE, 19,20-DiHDoPE) ( Figure 9B ) (50). cSiO₂ evoked production of 14,15-EpETrE and 14,15-DiHETrE starting at 1.5 h and continuing through 4 h in VEH-treated FLAMs and, to a lesser degree, in DHA-treated FLAMs. LPS priming also modestly increased cSiO₂-triggered production of 14,15-EpETrE in VEH-treated FLAMs. Changes in 14,15-EpETrE levels were not significant, and cSiO₂-induced production of 14,15-DiHETrE was significant only at 1.5 h. In contrast, DHA treatment promoted robust production of 19,20-EpDPE and 19,20-DiHDoPE at both timepoints. Exposure to cSiO₂ resulted in a subtle, yet non-significant, increase in 19,20-EpDPE and corresponding decrease in 19,20-DiHDoPE at both timepoints. Intriguingly, LPS priming alone significantly decreased levels of 19,20-EpDPE and 19,20-DiHDoPE during the experiment. Overall, levels of 19,20-EpDPE and 19,20-DiHDoPE were found to be higher than levels of 14,15-EpETrE and 14,15-DiHETrE.

The concurrent impacts of DHA on proinflammatory cytokine release from unprimed and LPS-primed FLAMs were evaluated at 1.5 h and 4 h post cSiO₂-treatment ( Figure 10A ). At both timepoints, unprimed FLAMs treated with either VEH or cSiO₂ alone secreted negligible amounts of IL-1α, IL-1β, and TNF-α ( Figures 10B–D ). In contrast, LPS-primed FLAMs released robust amounts of IL-1α, IL-1β, and TNF-α at both timepoints following cSiO₂ exposure, with much higher cytokine levels observed at 4 h compared to 1.5 h post cSiO₂. DHA pretreatment in LPS-primed FLAMs markedly reduced cSiO₂-induced release of IL-1α, IL-1β, and TNF-α release at both timepoints. No notable DHA effects on proinflammatory cytokine release were evident in FLAMs treated with VEH or cSiO₂ alone.

In addition to analyzing the impacts of DHA, LPS, and cSiO₂ on the cellular lipidome, we also measured lysosomal cathepsin and LDH activities in collected supernatants ( Figure 10A ). Cathepsin activity in VEH-treated cells exposed to cSiO₂ alone was higher at 1.5 h (1.1×10⁷ MFI) than at 4 h (5.5×10⁶ MFI) post cSiO₂ treatment ( Figure 10E ). Interestingly, DHA caused a moderate increase in cathepsin activity in FLAMs exposed to cSiO₂ alone or to both LPS and cSiO₂ at 1.5 h but not at 4 h. LDH release in VEH-treated FLAMs treated with cSiO₂ alone was higher at 4 h (9.9%) than at 1.5 h (0.4%) ( Figure 10F ). LPS priming of FLAMs increased cSiO₂ -induced extracellular lysosomal cathepsin and LDH response at 1.5 h post cSiO₂ in both VEH- and DHA-treated FLAMs but not at 4 h post cSiO₂ ( Figures 10E, F ). In line with cathepsin activity analyses, DHA caused a slight increase in LDH release in FLAMs exposed to cSiO₂ alone at 1.5 h and did not significantly impact LDH release at 4 h.

Live-cell imaging using LysoTracker Red (LTR), MitoTracker Red (MTR), and SYTOX Green (SG) was employed to further assess the impacts of DHA on cSiO₂-induced lysosomal membrane permeabilization (LMP), mitochondrial depolarization, and cell death, respectively ( Figure 11A , Supplementary Figure 3A ). In a preliminary experiment, we found that LPS priming slightly expedited cSiO₂-induced loss of LTR⁺ cells ( Supplementary Figure 3B ). LPS did not significantly impact cSiO₂-induced development of SG⁺ cells ( Supplementary Figure 3D ). Intriguingly, LPS priming perpetuated cSiO₂-triggered loss of MTR⁺ cells ( Supplementary Figure 3C ). In unprimed FLAMs, the proportions of LTR⁺ cells ( Figure 10B ), MTR⁺ cells ( Figure 11C ), and SG⁺ cells ( Figure 11D ) at time of cSiO₂ addition were nearly 100%, 100%, and 0%, respectively, relative to total cells. Loss of lysosomal integrity occurred at very similar rates up to 4 h post cSiO₂ in VEH- and DHA-treated unprimed FLAMs exposed to cSiO₂ ( Figure 11B ). Mitochondrial depolarization progressed at approximately the same rate in VEH- and DHA-treated unprimed FLAMs from 0 to 1.5 h, and DHA slightly protected FLAMs from further mitochondrial depolarization from 1.5 to 4 h ( Figure 11C ). Minimal cell death was observed from 0 to 1.5 h for both VEH- and DHA-treated cells, and DHA slightly, albeit insignificantly, suppressed cell death from 1.5 to 4 h ( Figure 11D ). Interestingly, DHA pretreatment modestly enhanced cSiO₂-induced loss of LTR⁺ ( Figure 11B ) and MTR⁺ cells ( Figure 10C ) while increasing SG⁺ cells ( Figure 11D ). Altogether, DHA had negligible effects on cSiO₂-induced lysosomal membrane permeabilization, mitochondrial toxicity, and death in FLAMs.