Neutrophils Promote Metabolic Dysfunction‐Associated Steatotic Liver Disease Through Extracellular Vesicle‐mediated Lipid Transfer Authors: Sanjeeb Shrestha, Soo‐Jung Jung, Seong‐Min Lim, Yu‐Bin Lee, Hyeung‐Wook Shin, Hoyul Lee, Hee‐Kyung Jin, Jae‐Sung Bae, Jae‐Han Jeon, Lai Guan Ng, Chang‐Won Hong Published: 2026 ABSTRACT ABSTRACTMetabolic dysfunction‐associated steatotic liver disease (MASLD) and metabolic dysfunction‐associated steatohepatitis are leading causes of chronic liver disease, yet the contribution of neutrophils to early hepatocellular lipid accumulation remains poorly understood. Here, we investigated the role of neutrophils in hepatic lipid deposition during early MASLD development. Neutrophils acquired extracellular fatty acids (FAs) via FATP2 and CD36 and stored them as triglycerides (TGs). These lipid‐laden neutrophils (LNs) did not utilize FAs for energy production or lipid mediator synthesis but instead transferred lipid cargo to hepatocytes through extracellular vesicles (EVs). Neutrophil‐derived EVs were enriched with TGs and lipid metabolism‐regulating microRNAs, augmenting TG accumulation in hepatocytes. Peripheral neutrophils isolated from high‐fat diet‐fed mice exhibited a lipid‐laden phenotype, and adoptive transfer of LNs and EVs derived from LNs increased hepatic fat accumulation in recipient mice. In patients with MASLD, circulating neutrophils showed lipid droplet accumulation with increased TGs and enrichment of lipid‐associated miRNAs while plasma EVs were also enriched with TGs and lipid‐associated miRNAs. Single‐cell RNA sequencing of peripheral immune cells identified a neutrophil subpopulation characterized by enhanced lipid‐handling and EV‐related gene expression. These findings identify neutrophil‐mediated lipid transfer via EVs as a mechanistic link between innate immune activation and hepatic lipid accumulation during early MASLD.Keywords: extracellular vesicles, lipid‐laden neutrophils, metabolic dysfunction‐associated steatotic liver disease, neutrophils Abbreviations abhydrolase domain containing 5 acetyl‐CoA acyltransferase acetyl‐CoA carboxylase alpha ATP citrate lyase 1‐acylglycerol‐3‐phosphate O‐acyltransferase acyl‐CoA synthetase long‐chain family member cluster of differentiation 36 (a fatty acid translocase) carnitine palmitoyltransferase 1 extracellular acidification rate extracellular vesicles fatty acid‐binding protein fatty acids fatty acid synthase fatty acid transport protein 2 G0/G1 switch gene 2 gene ontology hydroxyacyl‐CoA dehydrogenase/3‐ketoacyl‐CoA thiolase/enoyl‐CoA hydratase alpha subunit hypoxia‐inducible lipid droplet‐associated protein in vivo imaging system Kyoto encyclopaedia of genes and genomes lipid droplets lipase E neutrophil‐derived extracellular vesicles from lipid‐laden neutrophils metabolic dysfunction associated steatohepatitis metabolic dysfunction associated steatotic liver disease mannosyl‐glycoprotein beta‐12‐N‐acetylglucosaminyltransferase monoglyceride lipase microRNAs neutrophil extracellular traps neutrophil‐derived extracellular vesicles from normal neutrophils oxygen consumption rate patatin‐like phospholipase domain containing 2 RNA sequencing single‐cell RNA sequencing triglycerides uniform manifold approximation and projection Introduction Metabolic dysfunction‐associated steatotic liver disease (MASLD) is characterized by the abnormal accumulation of hepatic lipids, primarily in the form of triglycerides (TGs) (Reddy and Rao 2006, Friedman et al. 2018, Rinella et al. 2023). Metabolic disorders such as obesity, hyperlipidemia, and type II diabetes mellitus (T2D) are major risk factors for MASLD and are strongly associated with hepatic steatosis (Diehl and Cause 2017). Excessive dietary lipid intake and increased adipose tissue lipolysis contribute to hepatic fat accumulation, while upregulation of de novo lipogenesis (DNL) further exacerbates steatosis in MASLD (Hwang et al. 2021). This lipid overload creates a lipotoxic hepatic environment that promotes recruitment and activation of innate immune cells, thereby contributing to hepatic inflammation and fibrotic remodeling (Diehl and Cause 2017, Hwang et al. 2021). Neutrophils are among the earliest immune cells recruited to the liver in response to lipotoxic injury (Friedman et al. 2018). Previous studies have primarily focused on their inflammatory functions, including the production of reactive oxygen species (ROS), release of granule contents, and formation of neutrophil extracellular traps (NETs), which aggravate hepatic inflammation (van der Windt et al. 2018, Ou et al. 2017, Wang and Liu 2021, Tang et al. 2021). Consistently, depletion of neutrophils or inhibition neutrophil‐derived inflammatory mediators has been shown to delay progression from simple steatosis into more advanced stages such as metabolic dysfunction associated steatohepatitis (MASH) and cirrhosis (Ou et al. 2017, Hwang et al. 2020). These studies suggest an important role for neutrophils in the inflammatory progression of fatty liver disease, however, whether neutrophils directly contribute to hepatocellular lipid accumulation remains unclear. Beyond their established inflammatory roles, neutrophil have been shown to participate in lipid handling in specific pathological conditions. A recent study demonstrated that neutrophils infiltrating premetastatic niches acquire lipids from resident mesenchymal cells and deliver them to premetastatic cancer cells, thereby supporting tumor progression (Li et al. 2020). While this study revealed a lipid‐delivering function of neutrophils, the potential contribution of such mechanisms to hepatic lipid accumulation in MASLD has not been explored. In the present study, we investigated whether neutrophils contribute to hepatic fat accumulation during MASLD development. We show that neutrophils acquire extracellular fatty acids (FAs) released from adipocytes, store them as TGs, and polarize into a lipid‐laden phenotype. These lipid‐laden neutrophils (LNs) transfer stored lipids to hepatocytes via extracellular vesicles (EVs), influencing hepatocellular lipid metabolism, and suggest a neutrophil‐mediated lipid transfer pathway that may contribute to early lipid accumulation and hepatic steatosis along with other established contributors (de novo lipogenesis, impaired fatty acid oxidation) during MASLD. Results Neutrophils Develop a Lipid‐Laden state Through Fatty Acid Uptake To determine whether neutrophils uptake free FAs, human neutrophils were exposed to FAs and intracellular lipid droplet (LD) accumulation was assessed using BODIPY staining (Figure 1A). Exposure to palmitic acid (PA), oleic acid (OA), and linoleic acid (LA) significantly increased LD accumulation in human neutrophils (Figure 1B and C). Neutrophils are known to express CD36 (Castleman et al. 2015) and FA transporter 2 (FATP2) (Al‐Khami et al. 2017), both of which mediate long‐chain FA uptake. Accordingly, treatment with sulfo‐N‐succinimidyl esters (SSO), a CD36 inhibitor, or lipofermata (Lipo), a FATP2 inhibitor, markedly reduced LD accumulation in FA‐exposed neutrophils (Figures 1C and S1A). The FA concentrations used did not affect neutrophil viability, NET formation, or ROS generation (Figure S1B–D). To examine whether neutrophils can uptake lipids from adipocytes, mouse neutrophils were co‐cultured with differentiated 3T3‐L1 (d3T3‐L1) adipocytes under inflammatory conditions induced by tumor necrosis factor‐α (TNF‐α) (Jack et al. 2022, Chen et al. 2010), and LD accumulation in neutrophils were assessed (Figure 1D). Neutrophils co‐cultured with d3T3‐L1 adipocytes exhibited significantly increased LD accumulation compared with those cultured with undifferentiated 3T3‐L1 cells (Figure 1E and F). Inhibition of CD36 or FATP2 attenuated LD accumulation (Figure 1F). Neutrophils develop a lipid‐laden state through free fatty acid uptake. (A–C) Neutrophils uptake free fatty acids (FAs). Human neutrophils were exposed to free fatty acids (PA, palmitic acid, 50 µM; OA, oleic acid, 50 µM; LA, linoleic acid, 50 µM). Lipid droplets (LDs) were visualized by BODIPY staining and immunofluorescence microscopy. SSO, sulfo‐N‐succinimidyl esters (CD36 inhibitor, 100 µM); Lipo, lipofermata (FATP2 inhibitor, 5 µM). (A) Schematic illustration of free FAs uptake by neutrophils. Representative immunofluorescence images (B) and quantification of LDs (C) in neutrophils exposed to free FAs. (D–F) Neutrophils acquire lipids from adipocytes. Mouse neutrophils were co‐cultured with differentiated 3T3‐L1 (d3T3‐L1) in the presence of TNF‐α, and LDs within the neutrophils were examined using immunofluorescence microscopy. (D) Schematic illustration of adipocyte‐to‐neutrophil lipid transfer. (E) Representative immunofluorescence images of neutrophils co‐cultured with adipocytes. Adipocytes are delineated by dashed lines, and neutrophils surrounding the adipocytes are indicated by arrowheads. (F) Quantification of LDs in neutrophils co‐cultured with adipocytes. Neutrophils co‐cultured with adipocytes were harvested, and intracellular LDs were measured using BODIPY staining. (G and H) Metabolic pathways in lipid‐laden neutrophils. (G) qPCR analysis of lipid metabolism‐related genes in vehicle‐ and OA‐treated neutrophils. (H) Heatmap showing the relative mRNA expression levels in OA‐treated neutrophils compared with the vehicle‐treated neutrophils. (I) Intracellular concentrations of FAs and TGs in OA‐treated neutrophils. (J–M) Lipidomics analysis in neutrophils exposed to OA. (J) Principal component analysis of glycerolipids in neutrophils. (K) Differential metabolite analysis in OA‐treated neutrophils. Scatter plot showing the log2 fold change of metabolites between OA‐ and vehicle‐treated neutrophils, with upregulated DG and TG species marked. (L and M) Heatmap showing the relative expression levels of selected DG and TG species enriched in OA‐treated neutrophils. Each heatmap annotates lipid species and their fatty acid composition. ABHD5, Abhydrolase domain containing 5; ACAA2, Acetyl‐CoA acyltransferase; ACACA, Acetyl‐CoA carboxylase alpha; ACAT, Acetyl‐CoA: cholesterol acyltransferase; ACLY, ATP citrate lyase; AGPAT, 1‐acylglycerol‐3‐phosphate O‐acyltransferase; ACSL, Acyl‐CoA synthetase long‐chain family member; CD36, Cluster of differentiation 36; CPT, Carnitine palmitoyltransferase; DGAT, (Diacylglycerol O‐acyltransferase); FABP, Fatty acid‐binding protein; FASN, Fatty acid synthase; G0S2, G0/G1 switch gene 2; HADHA, Hydroxyacyl‐CoA dehydrogenase / 3‐ketoacyl‐CoA thiolase / enoyl‐CoA hydratase alpha subunit; HILPDA, Hypoxia‐inducible lipid droplet‐associated protein; LIPE, Lipase E; MGAT, Mannosyl‐glycoprotein beta‐1,2‐N‐acetylglucosaminyltransferase; MGLL, Monoglyceride lipase; PLIN, Perilipin; PNPLA, Patatin‐like phospholipase domain containing; SLC27A4, Solute carrier family 27 member 4. All results are expressed as mean ± SEM. *p < .05; **p < .01; ***p < .001. We next examined the expression levels of genes associated with lipid metabolism in FA‐exposed neutrophils. Among the fatty acids tested, OA was selected as the principal stimulus for subsequent experiments, as it exhibits lower lipotoxicity potential and more efficiently promotes LD biogenesis without inducing substantial cytotoxicity (Dai et al. 2025, Zeng et al. 2020). OA‐exposed neutrophils exhibited increased expression of genes associated with lipid transport, FA synthesis, lipolysis and inhibition of lipolysis, including G0S2 and HILPDA, which are critical for LD accumulation (Figure 1G and H). Although carnitine palmitoyl transferase 1A (CPT1A) and CPT1C showed divergent expression changes, analysis of the extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) revealed no significant changes in either human or mouse neutrophils exposed to FAs (Figure S1E), suggesting that acquired FAs are not substantially utilized for cellular energetics. Consistently, supernatants from FA‐exposed neutrophils did not induce directional chemotaxis of naive neutrophils but instead increased random migratory behavior, indicating limited production of lipid chemoattractants such as leukotriene B4 (LTB4) (Figure S1F). We next examined whether neutrophils store the acquired FAs intracellularly. Quantification of intracellular lipids species showed that neutrophils store taken‐up FAs predominantly in the form of TGs (Figure 1I). Consistently, lipidomics analysis revealed significant increases in TGs and diacylglycerols (DGs) in OA‐exposed neutrophils. Lipid profiling further demonstrated increased abundance across multiple lipid classes, including glycerophospholipids (phosphatidylserine, PS; phosphatidylcholine, PC; phosphatidylethanolamine, PE; phosphatidylglycerol, PG; phosphatidylinositol, PI), sphingolipids (sphingomyelin, SM; ceramide, Cer), and glycerolipids (DGs and TGs) (Figure S2A–C). Principal component analysis (PCA) for glycerolipids species revealed distinct differences between vehicle‐ and OA‐treated neutrophils (Figure 1J). Differential metabolite analysis identified significant upregulation of four DG species and twenty‐four TG species in OA‐treated neutrophils (Figure 1K). Notably, DG (36:2) and TG (72:4) species associated with OA (18:1) were markedly increased (Figures 1L–M and S2D–E), consistent with incorporation of OA into intracellular DG and TG pools. Based on these results, we refer these cells as lipid‐laden neutrophils (LNs). Lipid‐laden Neutrophils Transfer Lipids to Hepatocytes We next examined whether LNs transfer stored lipids to hepatocytes. LNs were generated by co‐culturing neutrophils with d3T3‐L1 adipocytes, and AML12 hepatocytes were subsequently exposed to LNs (Figure 2A). AML12 cells cultured with LNs exhibited significantly increased LD accumulation compared with cells exposed to normal neutrophils (NNs) or d3T3‐L1 adipocytes (Figure 2B). To directly trace FA movement between cells, d3T3‐L1 adipocytes were loaded with BODIPY‐PA, a fluorescent PA analog (Figure S3A), and co‐cultured with neutrophils. Neutrophils acquired BODIPY‐PA from adipocytes and incorporated into LDs (Figure S3B), demonstrating direct FA transfer from adipocytes to neutrophils. When these BODIPY‐PA‐laden neutrophils were subsequently co‐cultured with AML‐12 hepatocytes, BODIPY‐PA was transferred to AML‐12 cells and incorporated into their LDs (Figure S3C), suggesting onward FA transfer from neutrophils to hepatocytes. These findings were further validated using human neutrophils and hepatocytes. Exposure of HepG2 cells to human LNs resulted in increased LD accumulation (Figure 2C) and confirmed onward FA transfer from PA‐treated neutrophils to hepatocytes (Figure S3D and E). Lipid‐laden neutrophils transfer lipids to the hepatocytes. (A–C) Lipid transfer from lipid‐laden neutrophils (LNs) to hepatocytes. Mouse neutrophils were co‐cultured with either undifferentiated 3T3‐L1 or differentiated 3T3‐L1 (d3T3‐L1) and subsequently co‐cultured with AML12 cells for 2 h. Neutrophils were removed and intracellular fat levels in AML12 cells were quantified using BODIPY staining. NNs, normal neutrophils; LNs, lipid‐laden neutrophils; AML12+OA, AML12 cells treated with oleic acid (OA, 50 µM); AML12+LA, AML12 cells treated with linoleic acid (LA, 50 µM); AML12+d3T3L1, AML12 cells co‐cultured with d3T3‐L1 cells at the ratio of 1:1; AML12+NNs, AML12 cells co‐cultured with NNs; AML12+LNs, AML12 cells co‐cultured with LNs. (A) Schematic representation of the experiment. (B) Left, representative immunofluorescence images. Right, quantification of BODIPY fluorescence in AML12 cells. (C) qPCR analysis of the metabolic genes in AML12 cells. (D–F) Lipid transfer from neutrophils exposed to free FAs to hepatocytes. Human neutrophils exposed to OA were subsequently co‐cultured with human HepG2 cells for 2 h. Neutrophils were removed and intracellular fat levels were quantified using BODIPY staining. HepG2+Veh; HepG2 cells treated with vehicle, HepG2+OA; HepG2 cells treated with OA (50 µM), HepG2+NNs, HepG2 cells co‐cultured with NNs; HepG2+LNs, HepG2 cells co‐cultured with LNs. (D) Schematic representation of the experiment. (E) Quantification of BODIPY fluorescence in HepG2 cells (F) qPCR analysis of metabolic pathway genes in HepG2 cells. Dgat1, Diacylglycerol O‐Acyltransferase 1; LPIN, Lipin family member; Scd, Stearoyl‐CoA Desaturase. All results are expressed as mean ± SEM. *p < .05; **p < .01; ***p < .001. Consistently, qPCR analysis revealed increased expression of genes associated with lipid metabolism in LN‐exposed AML12 cells and HepG2 cells (Figures 2E–F and S4A–B), including genes associated with lipid uptake, transport, mitochondrial fatty acid oxidation, de novo lipogenesis, TG synthesis, and inhibition of lipolysis without significant effects on cell viability (Figure S4C). HepG2 cells exposed to LNs exhibited increased OCR, suggesting enhanced utilization of lipid substrates (Figure S4D). These results suggest that LNs transfer lipid cargo to hepatocytes, leading to hepatocellular lipid accumulation and metabolic remodeling. Neither AML‐12 nor HepG2 cells exposed to NNs or LNs showed detectable release of TNF‐α or IL‐1β, and intracellular cytokine staining confirmed no significant accumulation of these cytokines, indicating that LN‐induced lipid accumulation in hepatocytes is not accompanied by a detectable inflammatory response. Extracellular Vesicles Contribute Lipid Transfer From Neutrophils to Hepatocytes To investigate the mechanism by which LNs transfer lipids to hepatocytes, we examined whether direct cell‐cell contact is required. HepG2 cells cultured with LNs in a transwell chamber, which prevents direct contact, exhibited significantly increased LD accumulation compared with cells exposed to NNs (Figure 3A). Additionally, HepG2 cells treated with conditioned media from LNs showed increased LD accumulation (Figure 3B), suggesting that soluble factors released from LNs contribute to lipid transfer. The comparable magnitude of LD accumulation observed across transwell, conditioned medium, and direct co‐culture conditions indicates that the principal lipid‐inducing effects of LNs are mediated by secreted soluble and vesicular factors rather than direct cell‐cell contact, although a contribution from contact‐dependent mechanisms cannot be entirely excluded. Lipid‐laden neutrophils transfer lipids to hepatocytes via extracellular vesicles. (A, B) LNs transfer lipids to HepG2 cells in the absence of direct cell to cell contact. (A) HepG2 cells were co‐cultured with either NNs or LNs in a transwell chamber, and intracellular BODIPY fluorescence was quantified. (B) Conditioned media from LNs mediate lipid transfer to HepG2 cells. HepG2 cells were treated with supernatants derived from either NNs or LNs, and intracellular BODIPY fluorescence was quantified. (C) Extracellular vesicles (EVs) mediate lipid transfer from LNs to hepatocytes. HepG2 cells were treated with EVs isolated from either NNs (NNEVs) or LNs (LNEVs), and intracellular BODIPY fluorescence was quantified. Veh, HepG2 cells treated with vehicle; +NNEVs, HepG2 cells treated with NNEVs; +LNEVs, HepG2 cells treated with LNEVs. (D) LNEVs exhibit increased lipid content. Left, representative immunofluorescence images. Right, quantification of BODIPY fluorescence in neutrophil‐derived EVs. (E) Quantification of lipid contents in neutrophil‐derived EVs. (F) In vivo distribution of fluorescence‐labeled neutrophil‐derived EVs. Mice were intravenously injected with fluorescence‐tagged neutrophil‐derived EVs, and bioluminescence was assessed 24 h after administration using IVIS spectrum in vivo imaging system. All results are expressed as mean ± SEM. *p < .05; **p < .01; ***p < .001. Neutrophils are known to release EVs under both resting and activated conditions, facilitating intercellular communication through delivery of diverse molecular cargoes (Youn et al. 2021, Hong 2018). We therefore examined whether EVs derived from LNs (LNEVs) contribute to transfer of lipids to hepatocytes. Exposure of HepG2 cells to LNEVs resulted in significantly increased LD accumulation compared with cells treated with EVs derived from NNs (NNEVs) (Figures 3C and S5A). Characterization of LNEVs and NNEVs using nanoparticle tracking analysis (NTA) revealed no significant differences in vesicle size. However, LNs released a greater amount of EVs than NNs, both under basal conditions and following activation (Figure S5B). Consistent with their lipid‐laden state, LNEVs exhibited increased BODIPY fluorescence and higher TG levels compared with NNEVs (Figure 3D and E). To further characterize the isolated EV preparations in accordance with MISEV (Minimal information for studies of extracellular vesicles) guidelines (Welsh et al. 2024), Western blot analysis confirmed the presence of EV‐associated marker proteins CD9, Flotillin‐1, and Caveolin‐1 in both NNEV and LNEV fractions, while the lipoprotein marker ApoB was detected exclusively in the neutrophil lysate, supporting the preparation purity and excluding substantial lipoprotein contamination (Figure S5C). Flow cytometric analysis further confirmed the vesicular nature of the isolated particles, as particle counts were markedly reduced following lysis with 1% Triton X‐100 (Figure S5D). Treatment of HepG2 cells with EVs isolated from BODIPY‐PA‐loaded neutrophils also induced BODIPY‐PA accumulation in LDs, supporting the role of EVs as vehicles for lipid cargo delivery to hepatocytes (Figure S6A). To assess the in vivo distribution of neutrophils and neutrophil‐derived EVs, fluorescence‐labeled neutrophils were intravenously injected and tracked using an in vivo imaging system (IVIS). Labeled neutrophils were detectable in the liver within 4 h after injection (Figure S6B). We further tracked fluorescence‐labeled neutrophil‐derived EVs, and fluorescent signals were predominantly observed in the liver (Figure 3F), indicating preferential hepatic accumulation of neutrophil‐derived EVs. Neutrophil‐derived Extracellular Vesicles Modulate Lipid Metabolism‐related Gene Expression in Hepatocytes To investigate how neutrophil‐derived EVs affect hepatocytes, we performed bulk RNA sequencing (RNA‐seq) of HepG2 cells exposed to NNEVs and LNEVs (+NNEVs, +LNEVs) (Figure 4A). HepG2 cells treated with either EV population displayed gene expression distinct from vehicle‐treated control cells (HepG2) (Figure S7A–C). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed enrichment of genes associated with MASLD and alcoholic liver disease in both +NNEVs and +LNEVs compared with HepG2 (Figures 4B and S7D). KEGG analysis further identified enrichment of metabolic pathways and endocytosis‐related genes in EV‐treated cells (Figures 4C and S7E). Consistently, Gene Ontology (GO) enrichment analysis revealed upregulation of genes involved in vesicle‐mediated transport, lipid metabolic processes, lipid biosynthesis, and fatty acid metabolism following EV exposure (Figures 4D and S7F). Differentially expressed gene (DEG) analysis showed increased expression of genes related to FA transport, FA metabolism, lipogenesis, and LD regulation in EV‐treated HepG2 cells (Figures 4E and S7G). While gene expression changes induced by NNEVs and LNEVs were largely overlapping, subtle differences between +NNEVs and +LNEVs were observed (Figures S7H–K). Intriguingly, EV‐treated HepG2 cells exhibited increased expression of several genes implicated in lipid metabolism and MASLD pathophysiology (Dai et al. 2025, Sun et al. 2025), including fatty acid desaturase 1/2 (FADS1/2), β‐galactosidase 1 (GLB1), malic enzyme 1 (ME1), 5'‐AMP‐activated protein kinase catalytic subunit alpha‐2 (PRKAA2), and patatin‐like phospholipase domain‐containing protein 3 (PNPLA3) compared with HepG2. These results suggest that neutrophil‐derived EVs per se induce metabolic changes in HepG2 cells. Transcriptomic profiling of HepG2 cells exposed to extracellular vesicles derived from lipid‐laden neutrophils. HepG2 cells were treated with extracellular vesicles (EVs) derived from NNs (+NNEVs) or LNs (+LNEVs) for 2 h and subsequently subjected to bulk RNA sequencing. (A) Schematic representation of the experimental procedure. (B) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis highlighting enrichment of human disease‐associated pathways in EV‐treated HepG2 cells. (C) KEGG pathway analysis of metabolic processes altered by neutrophil‐derived EVs. (D) Gene Ontology (GO) Biological Process enrichment analysis of differentially expressed genes. (E) Metabolic pathways modulated by LNEVs in HepG2 cells. Upper panel, a heatmap depicting differentially expressed genes (DEGs) in +LNEVs relative to vehicle‐treated HepG2 cells. Lower panel, a schematic summary illustrating alterations in lipid metabolic genes in +LNEVs based on KEGG metabolic pathway mapping. EVs are enriched in non‐coding RNA species, including miRNAs, which can post‐transcriptionally modulate gene expression in recipient cells (Valadi et al. 2007). Neutrophil‐derived EVs have been reported to contain various miRNAs targeting macrophage and hepatocyte function (Youn et al. 2021, He et al. 2021). Notably, HepG2 cells exposed to neutrophil‐derived EVs exhibited increased intracellular miRNAs levels compared with vehicle‐treated cells (Figure S8A). To identify candidate miRNAs potentially involved in hepatocellular lipid regulation, we integrated miRNAs profiles from EV‐treated HepG2 cells with published datasets of neutrophil‐associated, neutrophil EV‐associated, and MASLD‐related miRNAs. This analysis identified thirty‐four overlapping miRNAs shared across these datasets (Figures S8B and C). Profiling of neutrophil‐derived EVs revealed enrichment of several miRNAs, including hsa‐miR‐21‐3p, hsa‐miR‐34‐5p, hsa‐miR‐150‐5p, hsa‐miR‐192‐5p, hsa‐miR‐324‐3p, hsa‐miR‐483‐3p, hsa‐miR‐574‐3p, hsa‐miR‐760, and hsa‐miR‐3188, with LNEVs exhibiting higher levels of most miRNAs compared with NNEVs, except for hsa‐miR‐150‐5p (Figure 5A). Neutrophil‐derived extracellular vesicles modulate hepatocyte lipid metabolism. (A) qPCR validation of miRNAs enriched in neutrophil‐derived EVs. The expression level of each miRNA was normalized to UniSp6 as an external control. (B–D) Effects of miRNA mimic transfection on lipid metabolism in HepG2 cells. (B) Schematic illustration of experimental procedure. (C) Left panel, the expression levels of G0S2 and HILPDA in HepG2 cells transfected with their respective miRNA mimics. Right panel, a heatmap depicting normalized expression levels of G0S2 and HILPDA relative to β‐actin (ACTB). (D) Left panel, BODIPY fluorescence normalized to DAPI fluorescence, and the mean lipid droplet (LD) counts per cell in HepG2 cells transfected with miRNA mimics. Right panel, a heatmap summarizing relative BODIPY fluorescence and LD counts compared with mock‐transfected cells. (E) GO Cellular Component enrichment analysis of HepG2 cells treated with LNEVs. (F) Effects of pharmacological inhibitors of endocytic pathways in EV‐mediated gene regulation in hepatocytes. Expression of G0S2 and HILPDA was assessed in HepG2 cells treated with neutrophil‐derived EVs. (G) Representative BODIPY fluorescence images of and quantification of BODIPY intensity in HepG2 cells co‐cultured with LNs in the presence of the EV biogenesis inhibitor Y‐27632 (ROCK inhibitor, 50 µM) or the EV uptake inhibitor cytochalasin D (50 µg/mL). CytoD, Cytochalasin D; Wort, Wortmannin; CQ, Chloroquine; Pit, Pitstop; Gen, Genistein; Fil, Filipin. All results are expressed as mean ± SEM. *p < .05; **p < .01; ***p < .001. To assess whether these miRNAs modulate hepatocellular lipid handling, HepG2 cells were transfected with individual miRNAs mimics, and expression of G0S2 and HILPDA, LD accumulation, and BODIPY fluorescence were evaluated (Figure 5B). Mimics of hsa‐miR‐34‐5p, hsa‐miR‐192‐5p, hsa‐miR‐574‐3p, and hsa‐miR‐760 significantly increased expression of both G0S2 and HILPDA (Figure 5C). Mimics of hsa‐miR‐21‐3p, hsa‐miR‐34‐5p, hsa‐miR‐192‐5p, hsa‐miR‐324‐3p, hsa‐miR‐483‐3p, hsa‐miR‐574‐3p, hsa‐miR‐760, and hsa‐miR‐3188 enhanced lipid uptake in hepatocytes (Figures 5D and S8D). These results indicate that miRNAs enriched in neutrophil‐derived EVs are capable of promoting lipid accumulation‐associated transcriptional changes in hepatocytes. Next, we explored the mechanism underlying EV uptake by hepatocytes. RNA‐seq revealed enrichment of genes associated with endocytosis, phagocytic vesicles, and clathrin‐mediated endocytosis in EV‐treated HepG2 cells (Figures 5E and S8E–F). To assess the contribution of endocytic pathways, we examined the effects of pharmacological inhibitors on expression of the G0S2 and HILPDA in HepG2 cells exposed to LNEVs. Inhibition of actin polymerization with cytochalasin D, micropinocytosis and endosomal trafficking with chloroquine, or clathrin‐mediated endocytosis with genistein significantly attenuated the induction of G0S2 and HILPDA in LNEV‐treated cells (Figure 5F). Furthermore, inhibition of EV biogenesis by the ROCK inhibitor Y‐27632 or EV uptake by cytochalasin D significantly reduced the LD accumulation in LN‐exposed HepG2 cells (Figure 5G). These findings indicate that uptake of neutrophil‐derived EVs by hepatocytes involves actin‐dependent endocytic processes, including clathrin‐mediated pathways. Impact of Lipid‐laden Neutrophils on Hepatic Fat Accumulation in a Murine Model of MASLD To elucidate the contribution of LNs to hepatic fat accumulation in vivo, we examined the effects of neutrophil depletion in a high‐fat diet (HFD)‐induced murine model of MASLD. C57BL/6J mice were fed an HFD for eight weeks and neutrophils were depleted using the Ly6G‐specific antibody 1A8, as previously described (Ou et al. 2017) (Figures 6A and S9A–B). Peripheral neutrophils isolated from HFD‐fed mice exhibited increased LD accumulation, as indicated by elevated BODIPY fluorescence (Figure 6B). Consistently, immunofluorescence analysis of liver sections revealed the presence of Ly6G+ BODIPY+ neutrophils within the hepatic parenchyma of HFD‐fed mice (Figure 6C), supporting the accumulation of LNs in the liver during diet‐induced steatosis. HFD‐fed mice developed marked hepatic lipid accumulation, as visualized by Oil Red O staining and BODIPY fluorescence (Figures 6D–E and S9C–D). Intriguingly, depletion of neutrophils significantly reduced hepatic lipid accumulation (Figures 6D–E and S9C–D). Lipid‐laden neutrophils contribute to hepatic fat accumulation in vivo. (A–E) Effect of neutrophil depletion on hepatic fat accumulation in a murine model of diet‐induced obesity. (A) Schematic representation of the experimental design. C57BL/6J mice were fed either a high‐fat diet (HFD) or normal chow (NC). Neutrophils were depleted via intraperitoneal administration of an anti‐Ly6G antibody. (B) Representative immunofluorescence images showing BODIPY staining of peripheral neutrophils isolated from mice. (C) Immunofluorescence analysis of liver sections, showing the presence of BODIPY+Ly6G+ lipid‐laden neutrophils in livers of HFD‐fed mice. (D and E) Visualization of hepatic lipid accumulation using Oil Red O staining (D) and BODIPY immunofluorescence staining (E). (F–J) Effects of adoptive transfer of LNs on hepatic fat accumulation. (F) Schematic illustration of the experimental design. C57BL/6J mice were intravenously injected with either LNs or NNs at two‐day intervals for 10 days. (G and H) Representative images and quantification of hepatic lipid accumulation assessed by Oil Red O staining. (I and J) Representative images and quantification of hepatic lipid accumulation assessed by BODIPY fluorescence. All results are expressed as mean ± SEM. *p < .05; **p < .01; ***p < .001. To further evaluate whether LNs are sufficient to promote hepatic fat accumulation, mice were intravenously injected with either LNs or NNs, and hepatic lipid accumulation was subsequently assessed (Figure 6F). Adoptive transfer of LNs resulted in a significant increase in hepatic lipid accumulation compared with NNs, as demonstrated by enhanced Oil Red O staining and BODIPY fluorescence (Figures 6G–I and S9E–F). This increase in hepatic lipid content occurred in the absence of overt liver injury, as assessed by histological analysis (Figure S9G). Spatial analysis further revealed that regions of increased hepatic lipid accumulation closely overlapped with neutrophil infiltration (Figure 6I and S9F), suggesting a localized contribution of LNs to hepatocellular lipid deposition. To directly assess the contributions of neutrophil‐derived EVs to hepatic fat accumulation in vivo, NNEVs or LNEVs were intravenously administered to neutrophil‐depleted HFD‐fed mice twice weekly from week 2 to week 4. Immunofluorescence analysis and Oil Red O staining of liver sections revealed restoration of hepatic lipid accumulation in LNEV‐treated animals (Figures S10A and B). These findings provide direct in vivo evidence that LNEVs are sufficient to recapitulate hepatic fat accumulation in the absence of intact neutrophils, supporting EV‐mediated lipid delivery as a functionally relevant mechanism. Neutrophils From Patients With MASLD Exhibit a Lipid‐laden Transcriptional State To characterize immune cell states associated with MASLD, we performed single‐cell RNA sequencing (scRNA‐seq) of peripheral blood immune cells from healthy volunteers (n = 3) and patients with MASLD (n = 3) using the BD Rhapsody platform (Figure 7A). A total of 59,218 cells were captured and transcriptomically profiled. Unsupervised clustering and Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction identified ten major immune cell populations, including neutrophils, monocytes, dendritic cells, CD4+ T cells, CD8+ T cells, B cells, NK cells, eosinophils, basophils, and platelets (Figure S11A). Cell identities were annotated based on canonical marker expression, including CD14 and LYZ for monocytes, CD3E and CD4 for CD4+ T cells, CD3E and CD8A for CD8+ T cells, and S100A8, S100A9, CXCR2, CXCR4, FCGR3B, and CSF3R for neutrophils (Figure S11B–D). Neutrophils constituted the most abundant population across all donors (Figure S11E). GO analysis revealed enrichment of genes associated with lipid storage, lipid droplet, and EV biogenesis in neutrophils compared with other immune cell types (Figure S11F). Notably, neutrophils exhibited higher expression of G0S2 relative to other immune subsets (Figure S11G). Single‐cell transcriptomic profiling identifies a lipid‐associated neutrophil subpopulation in MASLD patients. (A) Schematic overview of the experimental workflow. Peripheral blood immune cells were collected from MASLD patients (n = 3) and healthy volunteers (n = 3) and subjected to single‐cell RNA sequencing using BD Rhapsody platform. (B) Left panel, uniform manifold approximation projection (UMAP) plot showing unsupervised clustering of neutrophil subpopulations (N1–N4) across all subjects. Right panels, relative distribution of neutrophil subclusters in MASLD patients and healthy volunteers. (C) Proportional distribution of neutrophil subclusters within total neutrophil compartment for individual samples. (D) UMAP feature plots and dot plots showing representative marker genes for each neutrophil subcluster. (E) KEGG and GO pathway enrichment analyses of neutrophil subclusters. (F) Violin plots displaying module scores for G0S2 expression, lipid droplet, lipid metabolic process, glycolysis, and apoptosis across neutrophil subclusters. (G) UMAP plots and violin plots illustrating G0S2 expression levels across neutrophil subpopulations in MASLD patients and healthy volunteers. (H) Schematic summary of transcriptional features characterizing each neutrophil subcluster. All results are expressed as mean ± SEM. *n < .05; **n < .01; ***n < .001. Unbiased clustering of neutrophils identified four transcriptionally distinct subpopulations (N1–N4) (Figure 7B). All four clusters were present in both healthy individuals and patients with MASLD, although the relative proportion of the N3 population was increased in MASLD patients (Figure 7C). Marker gene analysis revealed that N1 expressed high levels of canonical neutrophil markers, including FCGR3B, CSF3R, PTGS2, CXCR2, and CXCR4, whereas N2 displayed modestly elevated expression of VCAN, CD83, CD84, FTX, and MAML2, suggesting subtle transcriptional differences potentially related to regulatory or senescence‐associated features (Figures 7D and S12A–B). The N3 population was characterized by increased expression of inflammatory and NET‐associated genes, including S100A8, S100A9, S100A12, PADI4, HMGB2, and MMP9, while N4 showed enrichment of interferon‐stimulated genes such as IFIT1, IFIT2, IFIT3, and ISG15 (Figures 7D and S12A–B). KEGG and GO enrichment analyses revealed that the N3 population was enriched for pathways associated with MASLD, NET formation, lipid droplets, exosome biogenesis, and EV biogenesis (Figure 7E). Consistently, violin plot analyses demonstrated higher scores for G0S2, lipid droplet‐associated genes, and lipid metabolic processes in N3 neutrophils (Figure 7F). UMAP visualization further showed increased G0S2 expression in neutrophils from patients with MASLD across neutrophil populations (Figure 7G). Neutrophils from patients with MASLD also exhibited higher lipid droplets, lipid metabolic process, and lipid biosynthesis scores compared with those from healthy volunteers (Figure S12C). Correlation analysis indicated that G0S2 expression within the N3 population was associated with both cell frequency and lipid droplet scores (Figure S12D). These transcriptional features are consistent with a lipid‐associated neutrophil state that is enriched in patients with MASLD (Figure 7H). We further validated these transcriptomic findings at the functional level in neutrophils isolated from patients with MASLD. Compared with neutrophils from healthy subjects, neutrophils from patients with MASLD displayed increased BODIPY fluorescence and LD accumulation (Figure 8A–C). Indeed, neutrophils isolated from MASLD patients also showed an increased accumulation of TGs (Figure 8D) and exhibited significant upregulation of genes associated with lipid transport, FA synthesis, lipolysis, and lipolysis inhibition (Figure 8E). Furthermore, EVs isolated from the plasma of patients with MASLD exhibited increased TG content and enrichment of several lipid‐associated miRNAs, including hsa‐miR‐21‐5p, hsa‐miR‐150‐5p, hsa‐miR‐192‐5p, hsa‐miR‐324‐3p, hsa‐miR‐483‐3p, hsa‐miR‐760, and hsa‐miR‐3188 (Figure 8F and G). Consistently, neutrophils isolated from patients with MASLD exhibited increased levels of these miRNAs, and correlation analysis demonstrated a positive correlation between miRNA levels in neutrophils and those in circulating plasma EVs from patients with MASLD (Figure 8H and I), suggesting the contribution of neutrophils to the circulating EV‐associated miRNAs signature observed in MASLD. Neutrophils from patients with MASLD exhibit a lipid‐laden phenotype. (A) Schematic overview of the experimental workflow for the isolation and analysis of neutrophils and EVs from patients with MASLD and healthy volunteers. (B–E) Phenotypic and metabolic characterization of neutrophils isolated from patients with MASLD. Neutrophils were isolated from healthy volunteers and patients with MASLD and subjected to BODIPY staining, intracellular TG quantification, and qPCR analysis of the lipid metabolic genes. (B) Representative immunofluorescence images of neutrophils isolated from healthy volunteers and patients with MASLD stained with BODIPY to visualize LDs, with LDs indicated by arrows. (C) Quantification of LD counts per neutrophils from healthy volunteers and patients with MASLD. (D) Intracellular TG levels in neutrophils isolated from healthy volunteers and patients with MASLD. (E) qPCR analysis of lipid metabolism‐related genes in neutrophils, comparing expression levels between patients with MASLD and healthy volunteers. (F and G) Characterization of circulating EVs isolated from patients with MASLD. EVs were isolated from serum of healthy volunteers (HV EVs) and patients with MASLD (MASLD EVs) and analyzed for miRNAs expression and TG content. (F) qPCR analysis of selected miRNAs in HV EVs and MASLD EVs. (G) Quantification of TG concentrations in HV EVs and MASLD EVs. (H) qPCR analysis of selected miRNAs in neutrophils isolated from HV (HV Neu) and MASLD patients (MASLD Neu). (I) Spearman correlation analysis between miRNA levels in circulating EVs and neutrophils from patients with MASLD. All results are expressed as mean ± SEM. *p < .05; **p < .01; ***p < .001. Discussion In this study, we identified an additional role of neutrophils in hepatic lipid accumulation during the early stages of MASLD. Neutrophil have traditionally been recognized as first responders to hepatic injury, with extensive evidence demonstrating their involvement in inflammation, hepatocyte damage, and fibrogenesis during MASLD (Hwang et al. 2021, Ou et al. 2017). Clinical and experimental studies consistently show that neutrophil infiltration and elevated neutrophil‐to‐lymphocyte ratios correlate with disease severity, including steatosis, hepatocyte ballooning, fibrosis, and overall liver injury (Yilmaz et al. 2015, Alkhouri et al. 2012, Gadd et al. 2014). These observations strongly establish neutrophils as contributors to inflammatory amplification and disease progression rather than as initiators of steatosis. In the present study, our findings suggest that under conditions of metabolic stress, neutrophils can acquire a lipid‐associated state that enables them to participate in hepatocellular lipid accumulation prior to overt inflammatory liver injury. By taking up extracellular FAs, adipocyte‐derived lipids and storing them as TGs, neutrophils function as transient lipid carriers capable of delivering lipid cargo to hepatocytes. This process appears to operate alongside canonical pathways of hepatic lipid accumulation, such as adipocyte lipolysis, de novo lipogenesis, and dietary lipid intake. The identification of this neutrophil‐associated lipid‐handling mechanism broaden the functional repertoire of neutrophils in MASLD and highlights their potential contribution to shaping the hepatic metabolic microenvironment during early disease development. Hepatic fat accumulation is driven by three major pathways: increased delivery of FAs from adipose tissue, enhanced de novo lipogenesis within hepatocytes, and excess dietary lipid intake (Friedman et al. 2018, Donnelly et al. 2005). Our data suggest that neutrophils may participate in this process by delivering FAs from adipose tissue to the liver. Neutrophil infiltration into adipose tissue is a well‐described feature of obesity and metabolic dysfunction and is closely linked to chronic low‐grade inflammation (Bijnen et al. 2018, Uribe‐Querol and Rosales 2022). Such inflammatory conditions promote adipocyte lipolysis, increasing the availability of extracellular FAs that can be taken up by neutrophils. The presence of LNs may therefore reflect sustained metabolic stress within adipose tissue and serve as an indicator of increased lipid flux toward the liver. Notably, neutrophils do not utilize these acquired lipids for energy production or lipid mediator synthesis. Instead, they store FAs as neutral lipids, positioning them as reservoirs capable of delivering lipid cargo through the circulation via EVs which reach the liver via portal and hepatic blood flow. The mechanism by which lipids are transported from adipose tissue to the liver remain incompletely understood beyond the well‐established role of albumin‐bound FAs and lipoprotein‐associated lipid transport. Our findings suggest that neutrophils and neutrophil‐derived EVs may represent an additional route for lipid delivery. Neutrophils are known to release EVs under both resting and activated conditions in response to inflammatory stimuli, microbial products, cytokines, and immune complexes (Youn et al. 2021, Hong 2018). These EVs have been shown to modulate immune responses by influencing macrophage polarization, cytokine production, and lymphocyte trafficking (Youn et al. 2021, Lim et al. 2015). A previous study has also reported the anti‐inflammatory and antifibrotic effects of neutrophil‐derived EVs in experimental MASH (He et al. 2021), suggesting the functional diversity of these vesicles. In the present study, EVs released from LNs were enriched with TGs and efficiently taken up by hepatocytes through actin‐dependent endocytic pathways, including clathrin‐mediated mechanisms. Consistent with this, higher lipid contents in peripheral neutrophils and circulating EVs from patients with MASLD support that this EV‐mediated pathway may operate in human diseases, particularly during early hepatic lipid accumulation and steatosis alongside canonical pathways of hepatic lipid accumulation and hepatic steatosis. Beyond serving as lipid carriers, neutrophil‐derived EVs also influenced the expression of genes associated with lipid metabolism in hepatocytes. Transcriptomic analysis of hepatocytes exposed to neutrophil‐derived EVs revealed coordinated upregulation of genes involved in FA uptake, TG synthesis, LD organization, and metabolic remodeling. These transcriptional changes resemble metabolic adaptation observed in MASLD livers, suggesting that EV‐mediated signaling can modulate hepatocyte lipid handling in addition to delivering lipid cargo. The enrichment of miRNAs within neutrophil‐derived EVs provides a plausible mechanism for such effects. Several miRNAs identified in this study, including hsa‐miR‐21‐3p, hsa‐miR‐192‐5p, and hsa‐miR‐760, have been implicated in lipid metabolism, cellular stress responses, and fibrogenic signaling. Functional experiments using miRNA mimics indicate that these miRNAs are capable of promoting lipid accumulation‐associated gene expression and phenotypes in hepatocytes. While these findings support a contributory role of EV‐associated miRNAs, they do not exclude additional effects mediated by other EV cargoes, such as proteins or lipids, which may also participate in hepatocellular metabolic remodeling. Notably, NNEVs and LNEVs induced largely overlapping transcriptional changes in hepatocytes, suggesting that miRNAs cargo and other molecular constituents, rather than differential lipid content alone, are the primary drivers of transcriptional remodeling. The principal distinction between NNEVs and LNEVs may therefore in their differential capacity for direct lipid delivery, as supported by higher TG content and BODIPY fluorescence of LNEVs. Transcriptional reprogramming mediated by EV‐associated miRNAs and direct lipid transfer via vesicular cargo are therefore likely complementary contributors to hepatocellular lipid accumulation. The preferential hepatic accumulation of LNEVs observed in biodistribution analysis may additionally reflect differences in hepatocyte EV uptake efficiency attributable to their distinct lipid composition, though this interpretation requires dedicated mechanistic validation. The influence of neutrophil‐derived EVs may not be limited to hepatocytes. Metabolic reprogramming is increasingly recognized as a key determinant of immune cell differentiation and function (O'Neill et al. 2016). Anti‐inflammatory macrophages and regulatory T cells rely on FA oxidation and oxidative metabolism to support their phenotypic programs, and recent evidence suggests that neutrophil‐derived structures such as NETs can influence T cell differentiation through metabolic reprogramming (Wang et al. 2021). Given that neutrophil‐derived EVs are enriched in miRNAs and lipids capable of modulating cellular metabolism, it is plausible that these vesicles also affect the metabolic state and functional polarization of neighboring immune cells within metabolic tissues. Further studies will be required to determine whether such effects contribute to immune remodeling in MASLD. Our scRNA‐seq analysis of peripheral blood immune cells from patients with MASLD identified a neutrophil population enriched for transcriptional programs related to lipid metabolism, lipid storage, and EV biogenesis, suggesting a disease‐associated neutrophil state that reflects systemic metabolic stress. Functional validation confirmed that neutrophils from patients with MASLD exhibit increased LD accumulation, elevated TG content, and release EVs enriched in lipid‐associated miRNAs. These observations raise the possibility that LNs or their EVs could serve as accessible circulating indicators of metabolic and immunological alterations in MASLD. Compared with liver biopsy, which remains the diagnostic gold standard but is invasive and limited by sampling variability (Diehl and Cause 2017), blood‐based biomarkers derived from neutrophils may offer complementary information regarding disease activity and progression. Limitations of the Study This study has several limitations. Although our data support a role of neutrophils in lipid transport during the early stages of MASLD, the precise regulatory mechanisms by which EV‐associated miRNAs influence hepatocellular lipid metabolism remain incompletely defined. Furthermore, the specific lipid cargo carried and transferred by EVs are not fully characterized, and a direct link between our in vitro findings on NDEVs and circulating EVs population in patients with MASLD remains unresolved. In addition, the clinical relevance of LNs and neutrophil‐derived EVs as biomarkers requires validation in larger and longitudinal patient cohorts. The 8‐week HFD model represents an early‐to‐intermediate disease stages and does not fully recapitulate advanced fibrotic MASH, and future studies integrating temporal disease staging and functional perturbation approaches will be essential to delineate the full contribution of this pathway to MASLD progression. Methods All Reagents and Resources Used in this Study Are Summarized in Supplementary Table 1 Human Blood Samples This study involved nine patients diagnosed with Metabolic Dysfunction Associated Steatotic Liver Disease (MASLD) at Kyungpook National University Hospital between January 2022 and May 2026. All human subject research was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of Kyungpook National University (KNU‐2025‐0545). Venous blood was collected from healthy volunteers aged 18 years or older. All participants provided written informed consent, and information regarding gender and age was collected. Participation was limited to male subjects for healthy volunteers. Blood collection was performed using vials containing anti‐coagulant, and neutrophils were isolated using either density gradient or immunomagnetic negative selection (EasySep Human Neutrophil Isolation Kit, Stemcell Technologies, #17957). Whole blood was layered over an equal volume of Histopaque‐1077 (Sigma‐Aldrich, #10771) and centrifuged at 1400 × g for 30 min at room temperature. The granulocyte‐enriched supernatant was collected and allowed to sediment with 5% (w/v) Dextran (Pharmacosmos, #40016) for 45 min at 4°C. The neutrophil‐enriched supernatants were collected, and residual red blood cells were removed via hypotonic lysis. Neutrophils were then stabilized for 1 h in RPMI medium (Gibco, #11875‐093) supplemented with 5% fetal bovine serum (Hyclone, #SH30919.03) at a density of 1 × 107 cells/mL. Neutrophil purity, assessed via Wright‐Giemsa staining, consistently exceeded 95 %. Clinical variables and demographic characteristics of MASLD patients are summarized in Supplementary Table 2. Cell Lines The mouse preadipocyte cell line (3T3‐L1, The Korean Cell Bank, #10092.1), mouse hepatocyte cell lines (AML‐12, ATCC, #CRL‐2254), and human hepatocyte cell line (HepG2, The Korean Cell Bank, #88065) were purchased from the respective repositories. The 3T3‐L1 and HepG2 cell lines were cultured in DMEM (Gibco, #11965092) supplemented with 10% FBS and 1% antibiotics (Gibco #15240‐062). The AML12 cell line was maintained in DMEM supplemented with 1 × insulin‐transferrin‐selenium (ITS‐G, Gibco, #41400045), 10% FBS, and 1 % antibiotics. To induce differentiation of 3T3‐L1, the medium was replaced with MDI medium comprising DMEM, insulin (10 µM, Sigma‐Aldrich, #I2643), dexamethasone (1 µM, Sigma‐Aldrich, #D2915), IBMX (0.5 mM, Sigma‐Aldrich, #I5879), and indomethacin (200 µM, Sigma‐Aldrich, #I7378). Following 2‐day incubation, the medium was replaced with DMEM supplemented with insulin every other day for 10 days. Mouse Study The experimental procedures involving animals in this study were ethically approved by the Institutional Animal Care and Use Committee (IACUC) of Kyungpook National University (KNU‐2022‐0167). Male C57BL/6J mice, aged 8–9 weeks, were purchased from Central Lab Animal Inc. (Seoul, Republic of Korea). Mouse neutrophils were isolated from peripheral blood and bone marrow by positive selection of Ly6G cells using MACS isolation kit (Miltenyi Biotec, #130‐120‐337). Assessment of Lipid Accumulation in Neutrophils Human neutrophils (1 × 107 cells) were exposed to free fatty acids, oleic acid, (OA, Sigma‐Aldrich, #O1008), palmitic acid, (PA, Sigma‐Aldrich, #P5585), or linoleic acid, (LA, Cayman Chemical Company, #90150) for 2 h. Adipocytes and neutrophils were co‐cultured using a direct contact (cell‐contact‐dependent) method, as described by Monk et al. (Monk et al. 2020), with minor modifications. Briefly, 3T3‐L1 preadipocytes were chemically differentiated for 9 days prior to immune cell treatment. Differentiated adipocytes were then stimulated with TNF‐α (10 ng/mL, R&D Systems, #MAB610) for 20 h to induce a pro‐inflammatory state, as reported by Chen et al. (2010) and Jack et al. (2022) and the TNF‐α‐containing medium was replaced with fresh medium. TNF‐α concentration in the washed conditioned medium were confirmed to be below the detection limit by ELISA prior to neutrophil co‐culture. Freshly isolated bone marrow neutrophils were subsequently added directly onto the adipocyte monolayer at a 10:1 neutrophil‐to‐adipocyte ratio in 6‐well plates and co‐incubated for 2 h at 37°C with 5% CO2. Following incubation, neutrophils were collected by gently aspirating the basal medium and washing with 1 × PBS without disturbing the adipocyte layer, resulting in an adipocyte contamination of less than 3 % in the collected neutrophil fraction. Following transfer onto AML12 hepatocytes and a 2‐h co‐incubation period, hepatocytes were washed thoroughly to remove non‐adherent neutrophils prior to BODIPY staining, resulting in residual neutrophil and adipocyte contamination of approximately 2% and 0.5%, respectively. To inhibit lipid transport, neutrophils were treated with inhibitors of long‐chain fatty acid transporters, sulfo‐N‐succinimidyl esters (SSO, Sigma‐Aldrich, #135661‐44‐8, 100 µM) or Lipofermata (Lipo, Cayman Chemical Company, #297180‐15‐5, 5 µM). Cells were washed with 1×PBS and were incubated with staining solution 4,4‐Difluoro 1,3,5,7,8‐Pentamethyl‐4‐Bora‐3a,4a‐Diaza‐s‐Indacene (BODIPY 493/503, Invitrogen, USA, California #3922) for 15 min at 37°C. Cells were washed and fixed with phosflow buffer I (BD Biosciences, #557870) for 30 min and mounted using 4′,6‐diamidino‐2‐phenylindole (DAPI, Invitrogen #P36962) on an 8‐well chamber slide (Nunc Lab‐Tek II Chamber Slide, ThermoFisher, #154453). The BODIPY fluorescence was visualized by either immunofluorescence microscopy (Olympus Optical Co Ltd, IX83) or confocal microscopy (Carl Zeiss Meditec AG, LMS800). Quantitative Real‐time PCR for Lipid Metabolic Genes Total RNA was extracted using TRIzol reagent (Ambion, #15596018) and was then reverse‐transcribed using a RT2 first strand cDNA synthesis kit (Qiagen, #205411). Real‐time PCR was performed using SYBR Green reagent (Qiagen, #204143) on a StepOnePlus system (Applied Biosystems). Gene‐specific amplification was confirmed by comparative cycle threshold (Ct) analysis using StepOne software. For quantification, the average Ct values of the genes of interest was normalized to the average Ct values of GAPDH which was run in a parallel reaction to the genes of interest. The relative ΔΔCt method was used to calculate fold‐change values between the groups. The primers used in this study are as follows: human primers, ABHD5 (Qiagen, #PPH08221A), ACAA2 (Qiagen, #PPH13528A), ACACA (Qiagen, #PPH02316A), ACLY (Qiagen, #PPH00021A), ACSL1 (Qiagen, #PPH72508A), ACSL3 (Qiagen, #PPH15368F), ACSL5 (Qiagen, #PPH06335A), ACSL6 (Qiagen, #PPH08013A), ACTB (Qiagen, #PPH00073G), AGPAT1 (Qiagen, #PPH19453A), AGPAT2 (Qiagen, #PPH00845D), CD36(Qiagen, #PPH01356A), CPT1C (Qiagen, #PPH11204A), CPT1A (Qiagen, #PPH15298E), FABP4 (Qiagen, #PPH023282F), FABP6 (Qiagen, #PPH02466A), FASN (Qiagen, #PPH01012B), LPIN2 (Qiagen, #PPH20709B), G0S2 (Qiagen, #PPH20079C), GAPDH (Qiagen, #PPH00150F), HADHA (Qiagen, #PPH10000B), HILPDA (Qiagen, #PPH09691A), LIPE (Qiagen, #PPH02383A), MGAT1 (Qiagen, #PPH11508C), DGAT1 (Qiagen, #PPH23420F), DGAT2 (Qiagen, #PPH17129A), MGLL (Qiagen, #PPH19478A), PLIN2(Qiagen, #PPH02583A), PNPLA2 (Qiagen, #PPH11403B), SLC27A4 (Qiagen, #PPH00471A); mouse primers, ABHD5 (Qiagen, #PPM37242A), ACLY (Qiagen, #PPM05509B), ACSL1 (Qiagen, #PPM33300G), ACSL3 (Qiagen, #PPM63570C), ACSL5 (Qiagen, #PPM58825B), AGPAT2 (Qiagen, #PPM03322A), CD36 (Qiagen, #PPM03796D), CPT1A (Qiagen, #PPM25930C), DGAT1 (Qiagen, #PPM26329A), FABP4 (Qiagen, #PPM04517A), FASN (Qiagen, #PPM03816F), G0S2 (Qiagen, #PPM24951A), GAPDH (Qiagen, #PPM09206E), GPAT4 (Qiagen, #PPM32982A), HILPDA (Qiagen, #PPM29151A), LIPE (Qiagen, #PPM03313A), PLIN2 (Qiagen, #PPM05362F), PNPLA2 (Qiagen, #PPM27696A), SCD1 (Qiagen, #PPM05664F), SLC27A4 (Qiagen, #PPM03287A). Assay for Accumulation of Triglyceride and Fatty Acid in Cells Triglyceride (TG) and fatty acids (FAs) were isolated from cells (5×107 cells) or EVs (1×109 particles). EVs were suspended in chloroform (Junsei, #28560‐0350) and methanol (ThermoFisher Scientific, #A452‐4) at 3:1 ratio, mixed vigorously, and sonicated for 30 min on ice. The suspension was centrifuged at 10,000 × g for 15 min at 4°C, and the supernatant was collected. The debris was resuspended in chloroform and methanol solution, and the step was repeated until there was minimal pellet. Deionized distilled water was added to the suspension from chloroform and methanol solution, mixed vigorously, and centrifuged for 30 min. The aqueous phase was collected and dried out. The final precipitate was dissolved into the respective assay buffer. The accumulation of TG (Abcam, #ab65336) and FAs (Abcam, #ab65341) were measured. Lipidomics Analysis Human neutrophils (2 × 107 cells) were exposed to free fatty acid, oleic acid (OA), for 2 h. Cells were frozen in liquid nitrogen and shipped to BGI Genomics Co., Ltd, China, for lipid metabolite extraction and identification. The complete process, including metabolite extraction, LC‐MS/MS detection, peak extraction, identification, preprocessing, and quality control, was conducted at BGI Genomics Co. Ltd. Lipid metabolites were identified using the high throughput target quantification for metabolites kit (HML1600 Metabolome). The normalized data were further analyzed using R. The absolute quantities of total lipid molecules, expressed in micromoles per gram or liter, were calculated by summing the concentrations of metabolite species within each class and presented as mean ± SEM. Principal component analysis (PCA) to examine patterns and trends associated with glycerolipids was performed using the factoextra package. Differential glycerolipids were quantified as log2 fold change between glycerolipids in OA‐treated and vehicle‐treated neutrophils. Glycerolipids exhibiting an absolute fold change of ≧1.2 (log2FC ≧0.263) and p‐value < 0.05 were considered significant. Visualization of the results included a volcano plot prepared using the EnhancedVolcano package and a heatmap was generated using the heatmap3 package and Graphpad Prism 10. Assessment of Effector Functions of Neutrophils Extracellular NET formation was measured using SYTOX green (Invitrogen, #S7020), a cell‐impermeable DNA staining dye. Neutrophils (2×105 cells) were exposed to indicated free fatty acids for 2 h and stained with SYTOX Green (5 µM) for 20 min. Cells were then washed and the SYTOX green fluorescence was measured using a fluorescence microplate reader (Spectramax M2/e, Molecular Devices) at an excitation wavelength of 504 nm and an emission wavelength of 523 nm. Intracellular reactive oxygen species (ROS) levels were measured using a fluorescent probe, 2′7′‐ dichlorodihydrofluorescein diacetate (DCF‐DA, #C6827) assay. Neutrophils (2×105 cells) were exposed to indicated free fatty acids for 2 h and stained with DCF‐DA (5 µM) for 20 min. Cells were then washed and the DCF‐DA fluorescence was measured using a fluorescence microplate reader (Spectramax M2/e, Molecular Devices) at an excitation wavelength of 488 nm and an emission wavelength of 535 nm. The migration of neutrophils was determined using a μ‐slide chamber (Ibidi, #80326). Migration of neutrophils against fMLP (1 µM), PA (50 µM), OA (50 µM), LA (50 µM), supernatants of vehicle‐treated neutrophils, and supernatants of PA‐treated neutrophils was visualized by Olympus microscopy and analyzed by ImageJ and Chemotaxis and migration tool (Ibidi). Apoptosis Assay Apoptotic rates were measured using flow cytometry. Neutrophils (1 × 106 cells), HepG2 cells (1 × 106 cells), or HepG2 cells (1 × 106 cells) cocultured with either vehicle‐ or FFA‐exposed neutrophils (1 × 106 cells) for 2 h were stained using Annexin V‐FITC apoptosis detection kit (Abcam, #ab14085). Flow cytometric analysis for apoptotic rates was examined using BD FACSCalibur flow cytometer (BD Biosciences) in the FL1 (Annexin–FITC) and FL2 (Propidium iodide–PE) channels, and acquired data were analyzed using FlowJo software (Treestar). Seahorse XF Analysis The mitochondrial respiration and glycolytic function were measured using an XF‐96e Extracellular Flux Analyzer (Seahorse Bioscience). Neutrophils (1 × 105 cells) or HepG2 cells (2 × 104 cells) were seeded in the XF assay medium and plated on XF‐96e tissue culture plates precoated with Cell‐Tak (Corning, #354241). The extracellular acidification rate (ECAR) was measured using an XF base medium (Agilent Technologies, #103334‐100) supplemented with 1 mM Seahorse XF glutamine solution (Agilent Technologies #103579‐100), followed by sequential injection of glucose (10 mM, Sigma‐Aldrich #G7528), oligomycin (2 µM, Sigma‐Aldrich #75351), and 2‐DG (50 mM, Sigma‐Aldrich #D6134). The oxygen consumption rate (OCR) of neutrophils was measured using an XF medium containing RPMI medium supplemented with 10 mM Seahorse XF glucose solution (Agilent Technologies #103577‐100), 1 mM Seahorse XF pyruvate solution (Agilent Technologies #103578‐100), and 1 mM Seahorse XF glutamine solution (Agilent Technologies), followed by the sequential injection of oligomycin (1 µM), carbonyl cyanide‐4‐(trifluoromethoxy) phenylhydrazone (FCCP, 2 µM, Sigma‐Aldrich #C2920), and rotenone (1 µM, Sigma‐Aldrich, #R8875), and antimycin A (1 µM, Sigma‐Aldrich, #A8647). XF analysis was performed at intervals of 6 min over a 90 min cycle at 37°C. Hepatocytes Uptake Lipid From Neutrophils HepG2 cells (1 × 106 cells) and AML‐12 cells (1 × 106 cells) were co‐cultured with respective lipid‐laden neutrophils (1 × 107 cells) for 2 h. Lipid‐laden neutrophils were generated with either exposure to OA (50 µM) or d3T3‐L1 for 2 h. To assess the indirect delivery of lipids by neutrophils, HepG2 cells were co‐cultured with human lipid‐laden neutrophils in a transwell chamber with a 3 µm pore size Transparent PET membrane (Corning, #3452) or treated with supernatant (conditioned media) obtained from neutrophils exposed to OA. Extracellular Vesicle Isolation and Treatment Extracellular vesicles (EVs) were isolated from neutrophils as previously described (Youn et al. 2021). Neutrophils (5 × 107 cells) were incubated with OA (50 µM) in RPMI only for 2 h and supernatants were harvested. Cell debris was further eliminated by centrifugation at 1400×g and the supernatants were filtered through 1.2 µm syringe filter, followed by ultracentrifugation at 100,000 × g for 90 min. The pellets were harvested, washed with filtered 1 × PBS and resuspended in RPMI only. EVs were suspended to RPMI at a ratio of 100 µL per 5 × 106 parent cell equivalent. Isolated EVs were subjected to Nanoparticle Tracking Analysis (NTA) using NanoSight NS300 (Malvern Panalytical). EVs were injected into the sample chamber and measured in duplicate with a camera setting of 14, acquisition time of 25 s, and detection threshold setting of 5. At least 1400 frames were analyzed per video and analyzed by NTA analytical software version 3.4. For in‐vitro treatment 1 × 107 EVs particles (equivalent to EVs isolated from 5 × 106 neutrophils) were used. HepG2 cells (5×105 cells) were exposed to isolated EVs for 2 h. EVs were isolated from 200 µL of serum of MASLD patients or healthy donors using a EasySep Human Pan‐Extracellular Vesicle Positive Selection Kit following manufacturer's instruction (Stemcell Technologies, #17891). Pharmacological Inhibition of EV Biogenesis and Uptake To inhibit EV uptake by hepatocytes, HepG2 cells (1×106 cells) were exposed to EVs for 2 h in presence of wortmannin (100 nM, TOCRIS, #SL‐2052), chloroquine (2 µM, Abcam, #142116), genistein (100 µM, Sigma‐Aldrich, #G6649), pitstop (15 µM, Sigma‐Aldrich, #SML1169), filipin (20 µg/mL, Sigma‐Aldrich, #F9765) or cytochalasin D (50 µg/mL, TOCRIS, #1233). To examine the functional contribution of EV biogenesis and uptake to neutrophil‐mediated lipid transfer HepG2 cells (1 × 106 cells) were co‐cultured with lipid‐laden neutrophils (1 × 107 cells) for 2 h in the presence of ROCK inhibitor Y‐27632 (50 µM, MedChemExpress, #HY‐10071) to suppress EV formation, or cytochalasin D (50 µg/mL) to inhibit EV uptake. In Vivo Imaging Neutrophil‐derived EVs were stained with CellTracker Red CMTPX (Invitrogen, #C34552) and were administered into mice through intravenous injection via tail vein (3 × 108 particles per mouse). Mice were sacrificed 24 h after administration and the organs were harvested for analysis. The biodistribution of CMTPX‐labeled EVs was monitored using an In Vivo Imaging Spectrum (IVIS, Perkin Elmer). The harvested organs were subjected to imaging for 5 sec with an excitation wavelength of 580 nm and emission wavelength of 620 nm and acquired data were analyzed using the IVIS live imaging software (Perkin Elmer). Mouse lipid‐laden neutrophils (1 × 107 cells) were stained with CellTracker Red CMTPX and intravenously injected into mice. Mice were sacrificed 4 h after administration and cells in the liver were harvested. The harvested liver was minced and digested using Collagenase A (5 mg/mL, Sigma‐Aldrich, #10103578001), DNase I (20 U, Ambion, #AM2222), and Dispase (1 mg/mL, Gibco, #17105‐041) by incubation for 30 min at 37°C with agitation at 100 rpm. Subsequently, the digested sample was strained through a 100 µm strainer followed by red blood cell (RBC) lysis. The cells were then fixed followed by anti‐Ly6G (Abcam, #ab25377) staining and analyzed using flow cytometry (BD FACSCalibur). The acquired data were analyzed using FlowJo software. Bulk RNA Sequencing Total RNA was extracted from HepG2 cells (5 × 107 cells) using TRIzol reagent (Ambion). DNA contamination was removed using DNase. For the construction of an RNA library, samples were prepared using the TruSeq Stranded Total RNA Sample Preparation Kit (Illumina, # 20020596). Paired‐end cDNA libraries were prepared for each sample, and sequencing was performed using a NovaSeq 6000 system (Illumina). Both library construction and RNA sequencing were conducted at Rokit Genomics Korea. The bulk RNA‐seq data were analyzed utilizing available packages in R. The short reads were aligned with the human genome with the align function (Rsubread v2.8.1) using a reference index built from GRCh38. The abundance or expression level of transcripts as read counts was finally estimated using the featureCounts function (Rsubread v2.8.1, GRCh38.gtf annotation with default parameters). Differential expression analysis was performed using edgeR (3.36.0) using log transformation and the trimmed mean of M‐values (TMM) method for normalization. The transcripts were annotated using a species‐specific genome‐wide annotation database for humans, the org.Hs.eg.db database (v3.14.0). Pathway enrichment analysis of significant genes was conducted with enrichGO and enrichKEGG functions from the clusterProfiler package (v.4.2.2.) with Biological Process (BP) and Cellular Component (CC) as Gene Ontology (GO) annotation term as pathways for GO enrichment analysis. MiRNA Quantification miRNA was isolated from EVs or neutrophils using a miRNA isolation kit (miRNeasy Serum / Plasma kit, Qiagen, #04053228006244), and cDNA was synthesized using the miRCURY LNA miRNA starter kit (Qiagen, #339320). For cDNA synthesis, 50 ng of total RNA or 16 μ of RNA extracted from serum or EVs were used. Real‐time PCR was performed using SYBR Green reagent, the miRCURY LNA miRNA starter kit on a StepOnePlus system. The amplification of miRNAs was confirmed by Ct analysis using StepOne software. The average Ct values of the miRNAs of interest were normalized to the average Ct values of miR‐103‐a‐3p, an internal control which was run in a parallel reaction to the gene of interest, and the relative ΔΔCt was used to calculate fold‐change values between the groups. These fold changes were further normalized to Unisp6, serving as an external control. The miR primers used in this study are as follows: miR‐21‐5p, miR‐27b‐5p, miR‐34a‐5p, miR‐103‐a‐3p, miR‐150‐5p, miR‐192‐5p, miR‐233‐3p, miR‐324‐3p, miR‐483‐3p, miR‐574‐3p, miR‐760, miR‐1307‐3p, miR‐3188, UniSp6. MiRNA Transfection HepG2 cells (1×106 cells) were transfected with 50 pM of miRNA mimics or agomir (Applied Biological Materials) using a Nepa21 electroporator (NEPAGENE) at 225 V. Cells were incubated for 2 h in DMEM for 2 h and further incubated for 24 h with supplementation of FBS at a concentration a concentration of 10%. miRNA mimics and agomir used in this study are as follows: hsa‐miR‐21‐5p (Applied Biological Materials, #MCH01562), hsa‐miR‐27b‐3p (Applied Biological Materials, #MCH01639), hsa‐miR‐34a‐5p (Applied Biological Materials, #MCH01909), hsa‐miR‐150‐5p (Applied Biological Materials, #MCH0147), hsa‐miR‐192‐5p (Applied Biological Materials, #MCH01454), hsa‐miR‐223‐3p (Applied Biological Materials, #MCH01592), hsa‐miR‐324‐3p (Applied Biological Materials, #MCH01868), hsa‐miR‐483‐3p (Applied Biological Materials, #MCH02736), hsa‐miR‐574‐3p (Applied Biological Materials, #MCH03153), hsa‐miR‐760 (Applied Biological Materials, #MAH03810), hsa‐miR‐1307‐3p (Applied Biological Materials, #MCH01263), hsa‐miR‐3188 (Applied Biological Materials, #MAH01831). In Vivo Experiment We employed a murine diet‐induced obesity model to investigate the effects of neutrophil depletion on hepatic fat accumulation. C57BL/6J (6‐weeks‐old, male) were segregated into four groups: a normal diet control group (NC, n = 4), a high‐fat diet group (HFD, n = 4), a normal diet group treated with an anti‐Ly6G antibody (NC + anti‐Ly6G, n = 4), and a high‐fat diet group treated with anti‐Ly6G antibody (HFD + anti‐Ly6G, n = 4). Mice were fed a high fat diet (60 kcal% Fat, Research Diets, #12492) for 8 weeks. For depletion of neutrophils, mice were weekly intraperitoneally injected with anti‐Ly6G antibody (Bio‐Connect Life Science, #BP0075‐1) at a dose of 200 µg per mouse. To determine the ability of LNs to deliver lipids to the liver in vivo, mice were intravenously injected with 2 × 106 cells per injection of either NNs or LNs (n = 5 per group). The injections were administered five times at two‐day intervals. NNs and LNs were prepared by co‐culturing mouse neutrophils with either undifferentiated 3T3‐L1 cells or d3T3‐L1 cells, respectively, for 2 h to induce lipid‐laden state in neutrophils. Mice were sacrificed 11 days post‐injection, and liver sections were collected for lipid analysis. The sections were subjected to Oil Red O staining to visualize lipid accumulation and to immunofluorescence staining to further characterize the tissue distribution of lipids and cellular markers. Histology Liver tissues were embedded in FSC22 Clear Frozen Section Compound (Leica) and sectioned into 10 µm sections. These sections were placed onto Micro Slide Glass (Matsunami, #HMA‐S1214) and fixed by incubating ice‐cold acetone (Sigma‐Aldrich, #270725) and methanol (1:1) solution for 10 min within a slide staining jar at ‐20°C. The slides were then washed twice with PBS and blocked in 1% bovine serum albumin (BSA, Roche Diagnostics, #10735086001) for 1 h at RT. For immunostaining, slides were incubated with the primary antibody (Rat anti‐mouse Ly6G, Abcam, #ab25377) in 1% BSA in PBS for 2 h at RT. Slides were washed with PBS and incubated with secondary antibody (Goat anti‐Rat IgG Alexa Fluor 647, Abcam, # ab150167) and BODIPY (4 µM) for 1 h at RT. Slides were again washed, followed by counterstaining with DAPI. The prepared sections were examined using an immunofluorescence microscope (Olympus Optical, IX83). Oil Red O Staining For lipid visualization, liver tissues were cryosectioned into 6 µm‐thick slides and stained using Oil Red O (Sigma‐Aldrich, #O0625). Sections were hydrated with distilled water and 60 % isopropanol (Sigma‐Aldrich #1.09634.1011) and stained with Oil Red O for 15 min, followed by rinsing with 60% isopropanol and distilled water. Hematoxylin (Sigma‐Aldrich, #03971) was used for counterstaining. The stained sections were then visualized using an immunofluorescence microscope (Olympus Optical, IX83). Single Cell RNA Sequencing and Data Analysis Peripheral blood immune cells were isolated from MASLD patients and healthy volunteers following blood lysis using Red Blood Cell Lysis Solution (Miltenyi Biotech # 130‐094‐183). Isolated immune cells were processed with the BD Rhapsody system for microwell‐based single‐cell partitioning and library preparation and tagged using antibody‐conjugated polyadenylated DNA barcode for human cells (BD Hu Single cell sample multiplexing kit, BD Biosciences, #633781). Cell viability was assessed using DRAQ7 (BD Pharmingen, # 564904) and Calcein AM (Thermo Fischer Scientific, #C1430). Optimal cell input concentration (60,000 cells/sample) for multiplexing was determined using the BD Rhapsody Scanner. Cells were pooled and loaded with BD Rhapsody 8‐lane cartridge using the Enhanced Cartridge Reagent Kit (BD Biosciences, #644887) on the BD Rhapsody HT express system, following the manufacturer's instruction. Subsequent lysis, cDNA synthesis, and exonuclease I treatment were conducted using the BD Rhapsody cDNA kit (BD Bioscience, #633773). Whole‐transcriptome libraries were prepared using the BD WTA amplification kit (BD Biosciences, #6338011) and sequenced on an Illumina NovaSeq X Plus platform. Raw sequence data were processed with the BD Seven Bridges WTA Analysis Pipeline (v2.3) using the RhapRef_Human_WAT_2023‐02 reference genome. This pipeline generated MolsPerCell (containing information about genes and molecules for each cell) and gene expression matrices (UMI counts per gene per barcode) for downstream analysis. Gene expression matrices were analyzed using Seurat (v5.3.0). Data were imported using the Read10X function and merged into a single Seurat object using Rstudio (2024.04.2.764). Quality control filtering excluded cells with: (i) <100 unique features, (ii) >5000 features, or (iii) >20% mitochondrial gene content. Putative doublets were identified and removed using the DoubletFinder package (v2.0.4). After filtering, a total of 59,218 cells were retained for downstream analysis. Data normalization was performed using the LogNormalize method. Highly variable genes (n = 2,000) were identified with FindVariableFeatures, followed by PCA. Harmony integration was performed using the IntegrateLayers function with 15 principal components, followed by non‐linear dimensionality reduction via UMAP. The filtered datasets containing singlets were subjected to a standard workflow for visualization and clustering, including ScaleData to minimize cell‐cell variation in gene expression driven by batch effects, followed by RunPCA, IntegrateLayers, FindNeighbors (15 dimensions, reduction = “harmony”), FindClusters, and RunUMAP (15 dimensions, resolution = 0.5). Cell identities were assigned based on canonical markers: neutrophils (CXCR2, FCGR3B, CSF3R), monocytes (CD14, LYZ, VCAN, FCN1), dendritic cells (FCER1A, CD1C, NRP1, CLEC4C), T cells (CD3E, CD4, CD8A, IL7R, CCR7), B cells (MS4A1, IGHM, CD79A), NK cells (GNLY, NKG7, KLRD1, NCAM1), platelets (PPBP), and basophil/eosinophils (IL5RA, SIGLEC8, ENPP3, IL3RA), as previously described (Hautz et al. 2023). Ten major immune cell types were annotated accordingly. Neutrophils (n = 40,691, four subpopulations) were extracted from the dataset using the subset function in Seurat to generate a separate object for subcluster analysis. Differentially expressed genes (DEGs) among neutrophil subclusters and between healthy and MASLD samples were identified using FindAllMarkers or FindMarkers functions with Student's t‐test and filtered by |log2FC| > 0.1 and adjusted p < 0.05. Module scores for biological pathways were computed using the AddModuleScores function with gene sets curated from the GSEA‐Molecular Signature Database (gsea‐msigdb.org). Violin plots were generated to visualize module scores. Comparison between groups was made using unpaired Student's t‐tests. Functional enrichment analysis for GO (biological process, GOBP; cellular component, GOCC; and molecular function, GOMF) and KEGG pathways was performed using compareCluster from clusterProfiler package. Selected pathways were visualized as dot plots using the dotplot function in enrichplot. Correlations between G0S2 expression, neutrophil subcluster proportions, BMI, and lipid droplet score were assessed using simple linear regression and logistic models to evaluate the association between gene expression and phenotypic traits. Statistical Analysis All statistical data were analyzed using GraphPad Prism (GraphPad software Ltd.). Data are presented as the mean ± SEM and comparison was performed using either Student's t‐test (unpaired t‐test). P values < 0.05 were considered statistically significant. Patients and Public Involvement Patients or members of the public were not involved in the design, conduct, reporting, or dissemination of this research. Human blood samples from patients with metabolic dysfunction–associated steatotic liver disease (MASLD) and healthy volunteers were collected with informed consent under protocols approved by the Institutional Review Board of Kyungpook National University. The key findings of the study were shared with participants through representative visual summaries to promote transparency and feedback. All patient data were anonymized, and the study complied with the principles of the Declaration of Helsinki. Author Contributions Sanjeeb Shrestha: conceptualization, methodology, funding acquisition, writing – original draft, data curation, formal analysis, investigation. Soo‐Jung Jung: data curation, formal analysis, investigation, writing – original draft, methodology. Seong‐Min Lim: conceptualization, investigation, formal analysis, methodology. Yu‐Bin Lee: conceptualization, methodology, investigation, formal analysis. Hyeung‐Wook Shin: methodology, investigation, formal analysis. Hoyul Lee: methodology, investigation, formal analysis. Hee‐Kyung Jin: resources. Jae‐Sung Bae: resources. Jae‐Han Jeon: supervision. Lai Guan Ng: supervision, data curation, formal analysis, writing – review and editing. Chang‐Won Hong: conceptualization, investigation, validation, formal analysis, visualization, funding acquisition, writing – original draft, writing – review and editing, resources, project administration, supervision. Funding This work was supported by National Research Foundation of Korea Grant 2022R1A2C2006898 (CWH), 2020R1A4A2002691 (HKJ, JSB, and CWH), RS‐2023‐00222188 (SS), and Korea Health Industry Development Institute Grant RS‐2024‐00438983 (CWH). Conflicts of Interest The authors declare no competing interests. Supporting information Supporting Information: jev270350‐sup‐0001‐FigureS1‐S12.docx Supporting Information: jev270350‐sup‐0002‐TableS1.docx Supporting Information: jev270350‐sup‐0003‐TableS2.docx