Adipose extracellular vesicles carrying miR-210-3p drive macrophage inflammation and nicotine-induced atherosclerosis

Summary

Visceral adipose tissue (VAT)-derived extracellular vesicles (EVs) have emerged as key mediators of inter-organ communication, yet their role in nicotine-induced atherosclerosis remains poorly defined. Here, we demonstrate that nicotine markedly enhances secretion of VAT-EVs and that these EVs are preferentially internalized by macrophages within atherosclerotic plaques, thereby accelerating lesion progression. Functionally, nicotine-induced VAT-EVs promote macrophage inflammation, oxidative stress, and foam cell formation. High-throughput profiling identified miR-210-3p as a dominant pro-atherogenic cargo within VAT-EVs, and its inhibition significantly attenuated nicotine-induced atherosclerosis in vivo. Mechanistically, miR-210-3p directly targets Krüppel-like factor 7 (KLF7), amplifying macrophage inflammatory responses and promoting plaque progression. Collectively, these findings uncover a previously unrecognized role of adipose-derived EVs in smoking-related vascular injury and highlight EV-derived miR-210-3p as a promising therapeutic target in nicotine-associated atherosclerosis.

Graphical abstract

Highlights

• •Nicotine increases VAT-EVs release into blood and uptake by plaque macrophages
• •Nicotine-induced adipose EVs drive inflammation and foam cell formation
• •MiR-210-3p is the key pro-atherogenic cargo in adipose-derived EVs
• •MiR-210-3p targets KLF7 to modulate nicotine-induced atherosclerosis

Cardiovascular medicine; molecular medicine; molecular biology experimental approach

Introduction

Atherosclerotic cardiovascular disease (ASCVD) remains a leading cause of morbidity and mortality worldwide, fundamentally driven by atherosclerosis.¹ Cigarette smoking is a prominent modifiable risk factor for ASCVD, wherein nicotine acts as a key bioactive component directly linked to excess cardiovascular mortality.^2,3 Notably, global nicotine consumption has surged in recent years, fueled by the widespread adoption of electronic cigarettes (e-cigarettes).^4,5 Despite these clinical observations, the precise molecular mechanisms by which nicotine accelerates atherosclerosis remain incompletely understood, particularly mechanisms involving systemic inter-organ communication.

Adipose tissue is recognized as a dynamic endocrine organ, with visceral adipose tissue (VAT) accumulation being intimately linked to dyslipidemia and metabolic complications, including atherosclerosis.^6,7 While adipose tissue comprises a heterogeneous stromal vascular fraction, adipocytes constitute the primary metabolic unit and a dominant source of circulating extracellular vesicles (EVs).⁸ Beyond their canonical role in lipid storage, adipocytes utilize these EVs to encapsulate diverse bioactive cargos—including microRNAs (miRNAs), long non-coding RNAs, and proteins—to orchestrate long-distance intercellular communication and modulate the phenotype of remote organs.^8,9,10 In the context of atherosclerosis, plaques are densely populated by immune cells, among which macrophages act as the predominant drivers of inflammation and plaque progression. Given this landscape, we hypothesized that VAT—specifically via adipocyte-derived EVs—may remotely regulate plaque stability. This concept is supported by evidence in obesity models, where adipocyte-derived EVs carrying miR-27b-3p trigger endothelial inflammation to promote atherosclerosis.¹¹ However, whether nicotine-induced adipocyte-EVs specifically target macrophages—the central architects of plaque inflammation and foam cell formation—remains largely unexplored.

Emerging evidence indicates that nicotine exerts profound detrimental effects on adipose tissue, compromising both its structural integrity and functional homeostasis.¹² While nicotine use is frequently associated with transient weight loss, chronic exposure paradoxically promotes VAT accumulation and adipose insulin resistance, which in turn fuels systemic lipid dysregulation.^13,14 These observations are corroborated by animal models, which further underscore the long-term—and potentially intergenerational—metabolic perturbations induced by nicotine.^15,16,17

Given that nicotine compromises adipose tissue homeostasis and that stressed adipocytes are known to alter their EV secretory profiles, we hypothesized that nicotine-stimulated VAT releases pro-atherogenic EVs that specifically reprogram plaque macrophages to accelerate atherosclerosis.

Here, we investigated the mechanism underlying this inter-organ communication. We demonstrate that nicotine potently stimulates the secretion of VAT-EVs and reprograms their cargo, leading to a selective enrichment of miR-210-3p. Mechanistically, we show that EV-transferred miR-210-3p exacerbates atherogenesis by directly targeting the transcription factor KLF7, thereby driving macrophage inflammation. Collectively, our work establishes the nicotine/VAT-EVs/miR-210-3p/KLF7 axis as a critical driver of smoking-related cardiovascular pathology and identifies adipose-derived EV-miR-210-3p as a promising therapeutic target.

Results

Nicotine triggers visceral adipose inflammation and augments adipose-derived extracellular vesicle secretion

To evaluate the impact of nicotine on atherosclerosis, ApoE^−/− mice were fed either a normal chow diet (NCD) or a high-fat diet (HFD) and treated with subcutaneous nicotine injection (Figures 1A and as previously described¹⁸). H&E and oil red O staining revealed that nicotine treatment significantly exacerbated atherosclerosis compared to HFD alone, resulting in larger aortic lipid plaques (Figures 1B–1D). Additionally, serum lipid profiling indicated dyslipidemia (Figure S1A), particularly under combined HFD and nicotine exposure, which may contribute to the observed exacerbation of atherosclerosis.

Figure 1.

Nicotine induces inflammatory remodeling of visceral adipose tissue and promotes extracellular vesicle release

(A) Schematic of the experimental protocol for establishing the nicotine-enhanced NCD or HFD-induced atherosclerosis model in ApoE^−/− mice.

(B) Representative images of aortic arch plaques (scale bars, 1 mm), H&E-stained (aortic root) and oil red O-stained (aortic root) atherosclerotic plaques (scale bars, 200 μm), and H&E-stained visceral adipose tissue (VAT) (scale bars, 50 μm). Scale bars apply to all panels within their respective columns.

(C) Quantification of atherosclerotic plaque area in the aortic root based on H&E staining (n = 8).

(D) Quantification of lipid-positive area within atherosclerotic plaques (aortic root) based on oil red O staining (n = 4).

(E and F) VAT weight (E) and VAT weight-to-body weight ratio (F). Epididymal white adipose tissue (eWAT) was used as a representative model of VAT in mice (n = 7).

(G) Quantification of adipocyte diameter in VAT and the number of crown-like structures (CLS) in F4/80-stained VAT sections (n = 8–10).

(H) Relative mRNA expression of pro-inflammatory mediators (IL-1β, MMP3, MMP9, TNF-α, MCP-1, and IL-6) and PPARγ in VAT from HFD (red) and HFD+Ni (blue) groups. (n = 3).

(I) Schematic of the protocol for isolating VAT-derived extracellular vesicles (EVs) via differential centrifugation.

(J) Transmission electron microscopy (TEM) images of VAT-derived EVs (bilayer membrane, 100–200 nm; scale bars, 200 nm). Left: EVs (black arrows) from HFD and HFD+Ni VAT; Right: Quantification of EV counts from VAT (n = 5).

(K) Western blot analysis of EV-specific markers (CD9, CD63, and TSG101) and the absence of calnexin (an endoplasmic reticulum marker) in VAT-derived EVs.

(L) Nanoparticle tracking analysis (NTA) of size distribution for eWAT-derived EVs, showing a peak diameter of 100–200 nm.

(M) Concentration of VAT derived EVs determined by bicinchoninic acid (BCA) assay (based on total protein quantification, n = 3).

(N) Schematic of EV isolation from primary adipocytes and stromal vascular fraction (SVF) of VAT (left), and analysis of EV concentration by BCA assay in these fractions (right) (n = 6).

Values are shown as mean ± SEM. Two-group comparisons were performed using unpaired two-tailed Student’s t tests, and multi-group comparisons were analyzed by one-way ANOVA followed by Tukey’s multiple-comparisons post-hoc test. Sample sizes (n) indicate biological replicates per group, determined based on prior studies and pilot experiments to ensure reproducibility and adequate statistical sensitivity. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Given that visceral adipose tissue (VAT) accumulation is closely associated with atherosclerotic cardiovascular risk, we further explored the effects of nicotine on VAT. Surprisingly, despite nicotine reducing body weight, VAT mass was still increased under HFD conditions, independent of body weight changes (Figures S1B, 1E, and 1F). We next examined whether this nicotine-induced increase in VAT mass was associated with pathological tissue remodeling. H&E staining together with immunofluorescence analyses revealed a reduction in the mean adipocyte diameter and an increased frequency of crown-like structures, both hallmark features of adipose tissue inflammation and remodeling (Figures 1G and S1C). Moreover, nicotine significantly upregulated the expression of pro-inflammatory mediators (IL-1β, MMP3, MMP9) in VAT under HFD feeding (Figure 1H). These findings demonstrate that nicotine drives VAT into a pro-inflammatory and remodeled state. While nicotine provoked a slight increase in inflammatory markers in the NCD group, these alterations were substantially less pronounced than those induced by the combination of nicotine and HFD (Figure S1D). Consequently, we focused our downstream mechanistic investigations on the HFD model.

Remodeled and inflamed adipose tissue alters its secretion profile, and VAT, being a major source of extracellular vesicles (EVs), contributes to inter-organ communication and the progression of vascular diseases.¹⁹ Based on this, we hypothesized that nicotine-induced VAT promotes atherosclerosis through the release of pathogenic EVs. To test this hypothesis, we isolated EVs from the VAT of control and nicotine-treated mice using ultracentrifugation (Figure 1I). EVs were characterized by TEM, which revealed ovoid vesicles with bilayer membranes (Figure 1J). Western blot confirmed the presence of EV markers (CD63, CD9, and TSG101) and the absence of the negative control protein Calnexin, confirming the purity of the samples (Figure 1K). Nanoparticle tracking analysis (NTA) showed that the predominant vesicle diameter was between 100 and 200 nm (Figure 1L).

Quantitatively, nicotine stimulated a 1.3-fold increase in VAT-EVs release compared to HFD controls (Figure 1M). To determine the cellular origin of these EVs, we fractionated VAT into primary adipocytes and the stromal vascular fraction (SVF). This analysis revealed that nicotine specifically augmented EVs output from adipocytes, while EVs release from the SVF remained unchanged (Figure 1N). Collectively, these data demonstrate that nicotine exacerbates atherosclerosis, remodels VAT into a pro-inflammatory and hypotrophic state, and selectively enhances EV secretion from adipocytes.

Nicotine-stimulated adipose-derived EVs promote atherosclerotic plaque progression by targeting plaque-resident macrophages

To determine the functional contribution of VAT-EVs from nicotine-exposed mice to atherosclerosis, we isolated EVs from the VAT of donor mice fed a high-fat diet (HFD-EVs) or an HFD plus nicotine (HFD+Ni-EVs) and intravenously injected them into ApoE^−/− mice recipients (Figure 2A).

Figure 2.

Nicotine-stimulated visceral adipose-derived EVs promote atherosclerotic plaque progression and preferentially target plaque-resident macrophages

(A) Schematic illustration of the experimental design evaluating the effect of visceral adipose–derived EVs on atherosclerosis. ApoE^−/− recipient mice were fed an HFD for 8 weeks, followed by 4 weeks of tail vein injection with EVs isolated from the VAT of HFD-fed or HFD+nicotine (HFD+Ni)-treated donor mice.

(B) Representative images of aortic sinuses: gross morphology, H&E-stained sections, and oil red O-stained sections (scale bars, 1 mm for gross images; 200 μm for stained sections).

(C) Quantification of atherosclerotic plaque parameters in aortic sinuses based on H&E staining (n = 8) and lipid accumulation based on oil red O staining (n = 4).

(D) Immunohistochemical staining of aortic sinuses for pro-inflammatory cytokines (IL-6, IL-1β, and TNF-α) and the antioxidant enzyme SOD2 (scale bars, 100 μm).

(E) Quantification of expression levels of pro-inflammatory cytokines and antioxidant markers in aortic sinuses (n = 6).

(F and G) Confocal fluorescence images showing co-localization of PKH67-labeled EVs (green) with CD68⁺ macrophages (red, F) and α-SMA⁺ vascular smooth muscle cells (red, G) in atherosclerotic plaques; DAPI (blue) stains cell nuclei (scale bars, 50 μm in F and 100 μm in G).

(H) Quantification of PKH67-labeled EVs co-localized with CD68⁺ macrophages, demonstrating significantly greater uptake of HFD+Ni EVs by plaque-resident macrophages compared with HFD EVs (n = 6).

Values are shown as mean ± SEM. Two-group comparisons were performed using unpaired two-tailed Student’s t tests. Sample sizes (n) indicate biological replicates per group. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Morphological and biochemical analysis revealed enhanced atherogenesis following HFD+Ni-EVs treatment. Compared to HFD-EVs, mice treated with HFD+Ni-EVs exhibited significantly increased aortic root plaque area and lipid accumulation, as demonstrated by H&E and oil red O staining (Figures 2B and 2C). This structural progression was accompanied by pronounced inflammatory changes within the plaques. Specifically, plaques from HFD+Ni-EVs-treated mice exhibited increased positive areas of IL-1β, IL-6, and TNF-α, while superoxide dismutase 2 (SOD2) positive area was diminished (Figures 2D and 2E), indicating an inflammatory microenvironment with compromised antioxidant defenses. These local plaque effects coincided with partial lipid alterations. HFD+Ni-EVs administration increased serum TC and TG levels without affecting body weight, LDL-C, or HDL-C concentrations (Figures S2A and S2B).

To elucidate the cellular mechanisms underlying these pro-atherogenic effects, we performed in vivo tracking of PKH67-labeled EVs. Given the centrality of macrophages and smooth muscle cells in atherosclerotic plaque, we specifically examined the association of EVs with these cell types. Confocal imaging revealed significantly greater colocalization of PKH67-labeled HFD+Ni-EVs with CD68⁺ macrophages compared to HFD-EVs, while their association with α-smooth muscle actin (α-SMA)⁺ smooth muscle cells did not differ significantly between the two groups (Figures 2F–2H). Further supporting these observations, immunostaining for the EV marker TSG101 in aortic sections showed that TSG101⁺ signals were predominantly colocalized with macrophages rather than smooth muscle cells (Figure S2C). Together, these data indicate that nicotine-stimulated adipose EVs accelerate atherosclerotic progression by preferentially associating with plaque-resident macrophages and being linked to an amplified inflammatory milieu within the lesions.

Nicotine-stimulated adipose-derived EVs impair macrophage function in vitro

Given that nicotine-stimulated adipose EVs were observed to associate with macrophages in vivo, we next investigated their direct effects on macrophages in vitro. EVs were isolated from VAT explants and from mature adipocytes (Figure 3A). In vitro, mature adipocytes were differentiated from 3T3-L1 preadipocytes, with successful adipogenic differentiation verified (Figure S3A). Nicotine exposure dose-dependently increased EVs secretion from both adipocytes and VAT. A significant elevation was already observed at 10 μM nicotine, which was selected for subsequent functional assays. It should be noted that this concentration is supraphysiological compared to human plasma levels, representing a limitation of our in vitro study design (Figures S3B and S3C). TEM, NTA, and western blot confirmed the presence and characteristic size/marker profile of adipocyte-derived EVs (Figures S3D–S3F).

Figure 3.

Nicotine-stimulated visceral adipose-derived EVs impair macrophage function by enhancing inflammation, oxidative stress, and foam cell formation

(A) Schematic of the experimental design: EVs isolated from mouse VAT or differentiated 3T3-L1 adipocytes (control or nicotine-treated) were co-cultured with RAW264.7 macrophages to assess the effects of EVs on macrophage function.

(B) Representative fluorescence images showing the uptake of PKH67-labeled EVs (green) by macrophages. Macrophages were incubated with PBS (Con), control EVs, or Ni-EVs derived from VAT or 3T3-L1 adipocytes. Nuclei are stained with DAPI (blue). Scale bars, 50 μm.

(C–E) RT-qPCR analysis of IL-1β, IL-6, and iNOS mRNA expression in macrophages treated with (C) VAT-derived control EVs (AT-EVs) or nicotine-EVs (AT-Ni-EVs), (D) 3T3-L1-derived control EVs (3T3-L1-EVs) or nicotine-EVs (3T3-L1-Ni-EVs), and (E) EV-depleted supernatants from control or nicotine-exposed adipose tissue (sAT-EV-depleted; EVs removed by ultracentrifugation). n = 3–4.

(F and G) Representative fluorescence images (F) and quantification (G) of DCFH-DA fluorescence intensity (intracellular reactive oxygen species, ROS) in macrophages treated with control or nicotine EVs (scale bars, 50 μm; n = 8).

(H) Representative fluorescence images and quantification of intracellular BODIPY fluorescence intensity (lipid-droplet accumulation) in macrophages treated with control or nicotine EVs derived from VAT (scale bars, 50 μm; n = 6).

(I and J) Representative oil red O staining (I) and quantification (J) of lipid accumulation in macrophages. Cells were treated with ox-LDL alone or in combination with AT-EVs or AT-Ni-EVs for 48 h (scale bars, 100 μm; n = 4).

Values are shown as mean ± SEM. Multi-group comparisons were analyzed by one-way ANOVA followed by Tukey’s multiple-comparisons post-hoc test. Sample sizes (n) indicate biological replicates per group; each dot represents one biological replicate. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To establish EV–macrophage interactions, PKH67-labeled EVs were incubated with macrophages and visualized by fluorescence microscopy, demonstrating efficient EV uptake by macrophages (Figure 3B). Following uptake, nicotine-exposed EVs (Ni-EVs) from both VAT and 3T3-L1 adipocytes induced significant pro-inflammatory transcriptional changes in macrophages, including elevated iNOS, IL-1β, and IL-6 mRNA levels compared with control EVs (Figures 3C and 3D). To exclude contributions from soluble adipose-derived factors, EVs were depleted from adipose tissue supernatants (sAT) by ultracentrifugation, validated by the absence of EV markers (TSG101, CD63) in the supernatant (Figure S3G). Under these conditions, nicotine-exposed and control sAT did not differentially affect IL-1β, iNOS, or IL-6 expression (Figure 3E), supporting that the pro-inflammatory response was attributable to EVs themselves rather than to soluble factors.

In addition to provoking inflammation, Ni-EVs also increased intracellular ROS levels in macrophages (Figures 3F and 3G) and promoted ox-LDL-induced macrophage transformation into foam cells, a hallmark of atherosclerotic progression. Oil Red O and BODIPY staining revealed greater neutral lipid accumulation and larger foam cell areas after Ni-EV treatment, whereas control VAT-EVs exerted minimal effects; similarly, Ni-EVs from 3T3-L1 adipocytes elicited greater foam cell formation than their respective controls (Figures 3H–3J). Together, these results indicated that nicotine-stimulated adipose EVs impaired macrophage homeostasis by promoting inflammatory activation, oxidative stress, and foam-cell formation.

Inhibition of EV production alleviates nicotine-induced macrophage dysfunction

To determine whether adipose-derived EVs mediate nicotine-induced macrophage activation, we treated adipose tissue explants and adipocytes with GW4869, a neutral sphingomyelinase inhibitor that blocks EV release. Adipose tissue explants were cultured in serum-free medium, and the resulting conditioned supernatant from adipose tissue (sAT) was applied to macrophages for 24 h (Figure 4A). Supernatants from nicotine-treated adipose tissue significantly upregulated inflammatory gene expression compared with vehicle controls, while GW4869 co-treatment attenuated this upregulation (Figure 4B).

Figure 4.

Inhibiting EV generation attenuates the impact of nicotine-stimulated adipose EVs on macrophages

(A) Schematic of the experimental workflow for the sAT-RAW model. Adipose tissue was treated with PBS, nicotine (Ni), or Ni + GW4869. The conditioned supernatant from adipose tissue (sAT) was then collected and applied to RAW264.7 macrophages.

(B) qRT-PCR analysis of inflammatory cytokine mRNA expression (IL-1β, IL-6, and MCP-1) in macrophages after stimulation with sAT from vehicle-, Ni-, or Ni + GW4869-treated visceral adipose tissue (VAT; n = 3).

(C) Schematic of the transwell co-culture system (3T3-L1-RAW): 3T3-L1 adipocytes in the upper transwell insert were treated with vehicle, Ni, or Ni + GW4869, while RAW264.7 macrophages were cultured in the lower chamber—enabling paracrine interactions via soluble factors (including EVs) that diffuse through the insert pores.

(D) RT-qPCR analysis of inflammatory cytokine mRNA expression (IL-1β, IL-6, MCP-1, and TNF-α) in macrophages after transwell co-culture with vehicle-, Ni-, or Ni + GW4869–treated 3T3-L1 adipocytes (IL-1β, n = 6; IL-6, n = 3; MCP-1, n = 4; TNF-α, n = 6).

(E) Representative fluorescence images of intracellular ROS in macrophages from the sAT-RAW (scale bars, 50 μm) and 3T3-L1-RAW (scale bar, 100 μm) models; ROS was detected via DCFH-DA staining (sAT-RAW, n = 4; 3T3-L1-RAW, n = 3).

(F and G) Quantification of mean DCFH-DA fluorescence intensity in macrophages from the sAT-RAW (F) and 3T3-L1–RAW (G) models (n = 4 and n = 3, respectively).

(H and I) Representative fluorescence images (H) and quantification (I) of intracellular BODIPY fluorescence intensity in macrophages from the sAT-RAW model (scale bars, 50 μm; n = 6).

(J–L) Representative Oil Red O staining (J) and semi-quantitative analysis of oil red O-positive lipid accumulation in macrophages from the sAT-RAW (K) and 3T3-L1-RAW (L) models (scale bars, 100 μm for sAT-RAW and 50 μm for 3T3-L1-RAW; n = 3–4).

Values are shown as mean ± SEM. Multi-group comparisons were analyzed by one-way ANOVA followed by Tukey’s multiple-comparisons post-hoc test. Sample sizes (n) indicate biological replicates per group; each dot represents one biological replicate. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To validate these findings, we established a transwell co-culture system using differentiated 3T3-L1 adipocytes and RAW264.7 macrophages (Figure 4C). Nicotine treatment enhanced the expression of inflammatory cytokines including IL-1β, IL-6, MCP-1, and TNF-α mRNA compared to vehicle controls (Figure 4D). GW4869 co-treatment significantly attenuated these responses, supporting the involvement of EVs in nicotine-induced inflammatory activation. ROS production was next examined by DCFH-DA fluorescence imaging. Nicotine exposure markedly increased ROS generation in both the sAT-RAW (adipose supernatant-RAW) and 3T3-L1-RAW co-culture systems (Figures 4E–4G).

GW4869 treatment substantially reduced ROS production in both experimental models, suggesting that EV release plays an important role in nicotine-induced oxidative stress. BODIPY staining revealed increased neutral lipid accumulation in macrophages following nicotine treatment, which was markedly reduced by GW4869 (Figures 4H and 4I). Oil red O staining similarly showed enhanced lipid droplet formation in both oxidized LDL-primed and nicotine-treated macrophages, with GW4869 significantly reducing lipid accumulation (Figures 4J–4L). These results indicate that pharmacological inhibition of EV production attenuates nicotine-induced macrophage inflammation, oxidative stress, and foam cell formation.

Nicotine-induced reprogramming of adipose EV cargo by miR-210-3p drives macrophage dysfunction

To investigate how nicotine affects the miRNA composition of adipose-derived EVs, we performed high-throughput sequencing on VAT-derived EVs from ApoE^−/− mice fed HFD or HFD+Ni. Differential expression analysis (fold change ≥1.5, p ≤ 0.05) identified 26 significantly altered miRNAs, including 23 upregulated and 3 downregulated species in the HFD+Ni group (Figures 5A and 5B).

Figure 5.

Nicotine reprograms the miRNA cargo of adipose-derived EVs, with miR-210-3p as a key mediator of macrophage dysfunction

(A) Experimental workflow for miRNA sequencing of adipose-derived EVs: EVs were isolated from visceral adipose tissue (VAT) of mice in high-fat diet (HFD) and HFD + nicotine (HFD + Ni) groups.

(B) Volcano plot of differentially expressed miRNAs (fold change ≥1.5, p ≤ 0.05) in adipose-derived EVs from HFD +Ni versus HFD mice.

(C) Heatmap of the top 20 differentially expressed miRNAs between HFD EVs (n = 4) and HFD+Ni EVs (n = 3) isolated from VAT of ApoE^−/− mice.

(D) RT-qPCR validation of representative differentially expressed miRNAs (e.g., miR-210-3p, miR-192-3p) in adipose-derived EVs from nicotine-treated vs. control ApoE^−/− mice (n = 3). miR-210-3p is highlighted in red.

(E) Gene ontology (GO) biological-process enrichment analysis of targets of differentially expressed miRNAs, highlighting processes related to atherogenesis, including macrophage activation, inflammatory cytokine secretion, and foam-cell formation.

(F) Expression of miR-210-3p in EVs and cells from nicotine-stimulated 3T3-L1 adipocytes and VAT-derived EVs from HFD-fed mice treated with nicotine (3T3-L1, n = 8; VAT, n = 6).

(G) Expression of miR-210-3p in nicotine-stimulated 3T3-L1 adipocytes and VAT from HFD-fed mice treated with nicotine (3T3-L1, n = 8; VAT, n = 6).

(H) qRT-PCR analysis of miR-210-3p expression in serum-derived EVs from HFD and HFD+Ni mice (n = 6).

(I) Transfection efficiency of miR-210-3p mimic in RAW264.7 macrophages (n = 3).

(J and K) qRT-PCR analysis of mRNA expression of inflammatory mediators in macrophages transfected with the miR-210-3p mimic (J: IL-6, IL-1β, iNOS, TNF-α, NF-κB, NLRP3, n = 3) or miR-210-3p inhibitor (K: IL-1β, n = 6; IL-6, n = 3).

(L and M) Representative fluorescence images and quantification of intracellular ROS levels after miR-210-3p mimic transfection (scale bars, 50 μm; n = 6).

(N and O) Representative oil red O staining and quantification of lipid accumulation after miR-210-3p mimic transfection (scale bars, 100 μm; n = 3).

(P and Q) BODIPY-cholesterol staining and quantification of lipid accumulation in macrophages transfected with the miR-210-3p mimic (scale bars, 50 μm; n = 6).

Values are shown as mean ± SEM. Two-group comparisons were performed using unpaired, two-tailed Student’s t tests, and multi-group comparisons were analyzed by one-way ANOVA followed by Tukey’s multiple-comparisons post-hoc test. Sample sizes (n) indicate biological replicates per group. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

A heatmap of the top 20 differentially expressed miRNAs visually illustrates this reprogramming (Figure 5C). Gene ontology analysis of predicted mRNA targets revealed significant enrichment for terms related to macrophage activation, cytokine secretion, and foam cell formation (Figure 5E). To validate these predictions in vivo, we performed immunohistochemistry on aortic sections from nicotine-treated mice (NCD+Ni and HFD+Ni) compared with their respective controls, which revealed increased expression of pro-inflammatory markers and decreased SOD2, reflecting enhanced vascular inflammation and oxidative stress (Figures S4A–S4H).

We selected eight candidate miRNAs from the sequencing data for further validation by RT-qPCR. This analysis confirmed that four miRNAs—miR-210-3p, miR-192-3p, miR-205-5p, and miR-449a-5p—were significantly elevated in nicotine-treated VAT-EVs, with miR-210-3p exhibiting the most pronounced upregulation (Figure 5D). Consistently, in HFD-fed mice, nicotine robustly increased miR-210-3p abundance across multiple EV sources, including EVs released from 3T3-L1 adipocytes, VAT, and serum (Figures 5F and 5H). In parallel, miR-210-3p levels were also elevated within nicotine-exposed adipose tissue and cultured 3T3-L1 adipocytes themselves (Figure 5G). Notably, nicotine similarly increased EV-associated miR-210-3p in VAT-derived EVs and circulating serum EVs in NCD-fed mice (Figures S5A and S5B), despite no detectable changes in lesion area or systemic metabolic parameters under NCD. In paired VAT fractionation analyses, nicotine increased EV-miR-210-3p in adipocyte-derived EVs but not SVF-derived EVs (Figures S5C and S5D). Together, these data are consistent with nicotine promoting both miR-210-3p induction and its EV association. To extend these findings to humans, we examined two independent public datasets (Database: GSE59421, GSE105449). miR-210-3p expression was significantly higher in CAD patients compared with controls (p = 0.019), and showed an increasing trend in smokers versus non-smokers (p = 0.076) (Figure S5E). Together, these analyses reinforce the link between smoking-related risk factors and elevated miR-210-3p expression.

Given its prominent enrichment in nicotine-reprogrammed EVs, we next examined the functional role of miR-210-3p in macrophages. Transfection with miR-210-3p mimics markedly increased the mRNA expression of pro-inflammatory mediators, including IL-6, IL-1β, iNOS, TNF-α, NF-κB, and NLRP3 (Figures 5I and 5J), whereas inhibition of miR-210-3p attenuated IL-6 and IL-1β mRNA expression (Figure 5K). MiR-210-3p overexpression also enhanced oxidative stress, as shown by increased ROS levels (Figures 5L, 5M, S5F, and S5G), and profoundly altered macrophage lipid metabolism. Specifically, miR-210-3p promoted lipid-droplet accumulation and impaired cholesterol efflux, leading to accelerated foam cell formation. This phenotype was consistently demonstrated by oil red O staining and BODIPY fluorescence imaging (Figures 5N–5Q, S5H, and S5I). Collectively, these results indicate that miR-210-3p plays a central role in nicotine-induced remodeling of adipose EV cargo and contributes to macrophage dysfunction.

miR-210-3p contributes to macrophage dysfunction by suppressing KLF7

To identify the mechanistic target of miR-210-3p, we utilized three prediction databases (TarBase, mirDIP, and TargetScan) to generate a list of ten candidate genes (Figure 6A). Among these, KLF7 emerged as the strongest candidate: RT-qPCR screening identified KLF7 as the top candidate (Figure S5J). To test whether Klf7 is a direct target of miR-210-3p, we performed a dual-luciferase reporter assay using wild-type (Klf7-WT) and mutant (Klf7-MT) constructs containing the predicted miR-210-3p binding site (Figure 6A). miR-210-3p mimics significantly reduced luciferase activity of the Klf7-WT reporter, whereas no suppression was observed with the Klf7-MT construct (Figure 6B), confirming that miR-210-3p directly binds the Klf7 3′UTR. Consistent with the reporter assay, miR-210-3p markedly decreased endogenous KLF7 expression at both the protein and mRNA levels, while inhibition of miR-210-3p produced the opposite effect (Figures 6C–6E). Immunofluorescence staining further verified this reciprocal regulation, showing robust KLF7 loss after miR-210-3p overexpression and clear restoration following miR-210-3p inhibition (Figures 6F and 6G). Collectively, these data establish KLF7—a reported atheroprotective regulator of macrophage metabolism—as a bona fide target of miR-210-3p.

Figure 6.

miR-210-3p promotes macrophage inflammatory activation and foam cell formation through KLF7 suppression, which is reversed by KLF7 overexpression

(A) Venn diagram showing 10 common downstream target genes of miR-210-3p identified from three databases (TarBase, mirDIP, and TargetScan); KLF7 was selected as a key target for further investigation. Potential binding sites of miR-210-3p on the KLF7 mRNA 3′ untranslated region (3′UTR) (3′UTR) are also shown.

(B) Dual-luciferase reporter assay in HEK293T cells co-transfected with miR-210-3p mimics (or NC mimics) and luciferase vectors containing wild-type (Klf7-WT) or Mutant (Klf7-MT) 3′UTR (n = 9).

(C–E) Validation of KLF7 suppression in macrophages. Western blot images (C), quantification of KLF7/GAPDH protein ratio (D, n = 6), and qRT-PCR analysis of Klf7 mRNA (E, n = 3) after transfection with miR-210-3p mimics or inhibitors.

(F and G) Representative immunofluorescence images (F) and quantification (G) of KLF7 expression after transfection with miR-210-3p mimics or inhibitors (KLF7, red; nuclei, blue [DAPI]; n = 6; scale bars, 20 μm).

(H and I) RT-qPCR (H, n = 6) and western blot (I, n = 3) analyses confirming transduction efficiency and KLF7 overexpression in macrophages following lentiviral transduction with Lv-KLF7 or empty lentivirus (control).

(J and K) qRT-PCR analysis of IL-1β and IL-6 mRNA expression in macrophages post-transfection with miR-210-3p mimics and/or transduction with Lv-KLF7 (n = 6).

(L and M) Representative immunofluorescence images (L) and quantification (M) of KLF7 expression in RAW264.7 macrophages transfected with miR-210-3p mimics and transduced with Lv-KLF7 (KLF7, green; nuclei, blue [DAPI]; n = 4; scale bars, 50 μm).

(N and O) Representative DCFH-DA fluorescence images (N) and quantification (O) of intracellular ROS levels (green). Red fluorescence (mCherry) indicates macrophages successfully transduced with the lentiviral vectors. Note that KLF7 overexpression attenuates miR-210-3p-induced oxidative stress. (scale bars, 50 μm; n = 6).

(P and Q) Representative BODIPY-cholesterol fluorescence images (P) and quantification of intracellular fluorescence intensity (Q) in macrophages transfected with miR-210-3p mimics and transduced with Lv-KLF7 (n = 3; scale bars, 50 μm).

(R and S) Representative oil red O staining (R) and semi-quantitative analysis of oil red O-positive lipid area (S) in macrophages transfected with miR-210-3p mimics and transduced with Lv-KLF7 (n = 5; scale bars, 100 μm).

Values are shown as mean ± SEM. Two-group comparisons were performed using unpaired, two-tailed Student’s t tests, and multi-group comparisons were analyzed by one-way ANOVA followed by Tukey’s multiple-comparisons post-hoc test. Sample sizes (n) indicate biological replicates per group. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To functionally validate this interaction, we transduced macrophages with lentiviral vectors encoding KLF7 (Lv-KLF7) or a control reporter (Lv-NC). Transduction efficiency was confirmed by RT-qPCR and western blotting, which verified robust KLF7 overexpression (Figures 6H and 6I). Immunofluorescence imaging confirmed that Lv-KLF7 transduction successfully restored nuclear KLF7 levels even in the presence of miR-210-3p mimics (Figures 6L and 6M). Importantly, restoring KLF7 expression substantially reversed the pathological effects induced by miR-210-3p mimics. Specifically, KLF7 overexpression markedly reduced miR-210-3p-driven IL-1β and IL-6 expression (Figures 6J and 6K), significantly attenuated the elevated intracellular ROS levels as shown by confocal microscopy (Figures 6N and 6O), and substantially decreased lipid accumulation and foam cell formation (Figures 6P–6S). Together, these data demonstrate that re-establishing KLF7 expression can counteract the pro-atherogenic effects of miR-210-3p, establishing the miR-210-3p-KLF7 axis as a critical regulator of macrophage dysfunction in nicotine-associated atherosclerosis.

Inhibition of miR-210-3p mitigates nicotine-induced atherosclerosis

To assess the therapeutic potential of miR-210-3p inhibition in nicotine-induced atherosclerosis, we used an AAV9 vector—chosen for its high affinity for cardiovascular tissues—to deliver an miR-210-3p inhibitor in ApoE^−/− mice (Figure 7A). En face staining of the aortic arch showed a marked reduction in plaque area following miR-210-3p inhibition (Figure 7A). Consistently, cross-sectional H&E and oil red O staining demonstrated decreased lesion size and lipid deposition (Figures 7B–7E). Inhibition of miR-210-3p also alleviated vascular inflammation and oxidative stress. Expression of SOD2 was significantly increased (Figures 7F and 7G), whereas pro-inflammatory cytokines—including IL-6, TNF-α, and IL-1β—were substantially reduced (Figures 7H–7M). In addition to reducing inflammatory cytokines, miR-210-3p inhibition markedly restored vascular KLF7 expression. Immunohistochemistry showed a significant increase in KLF7-positive staining within atherosclerotic plaques following anti-miR-210-3p treatment (Figures 7N and 7O). Confocal immunofluorescence further revealed that KLF7 predominantly colocalized with CD68⁺ macrophages and was strongly restored upon miR-210-3p inhibition (Figures 7P and 7Q), supporting KLF7 as a functional downstream effector in vivo.

Figure 7.

Inhibition of miR-210-3p in atherosclerotic plaques alleviates nicotine-induced atherosclerosis

(A) Left: Representative en face images of aortic arch plaques from negative control (NC) and AAV9-anti-miR-210-3p-treated (miR-210-3p inhibition) mice. Right: schematic illustration of the experimental design. ApoE^−/− mice were injected with either AAV9-NC inhibitors or AAV9–miR-210-3p inhibitors, followed by HFD+Ni treatment for 12 weeks before analysis (scale bars, 1mm).

(B and C) H&E-stained sections of aortic sinuses from nicotine-treated ApoE^−/− mice and corresponding quantification of lesion area (n = 8; scale bars, 200 μm).

(D and E) Oil red O-stained sections of aortic sinuses and quantification of oil red O-positive area (n = 4; scale bars, 200 μm).

(F–M) Immunohistochemical staining and corresponding quantification of SOD2 (F and G), IL-6 (H and I), TNF-α (J and K), and IL-1β (L and M) in aortic sinus sections from nicotine-treated ApoE^−/− mice (n = 8; scale bars, 200 μm).

(N and O) Immunohistochemical staining and quantification of KLF7 expression in the aortic sinus (n = 8; scale bars, 200 μm).

(P and Q) Representative immunofluorescence images (P) of aortic-root atherosclerotic lesions stained for KLF7 (red), CD68(green; macrophage marker), and DAPI (blue; nuclei), with corresponding quantification of KLF7 fluorescence intensity (Q; n = 8; scale bars, 100 μm).

(R) Alterations in body weight of ApoE^−/− mice over a 12-week period in NC and miR-210-3p inhibition groups.

(S–V) Serum lipid profiles, including TC, TG, LDL-C, and HDL-C, in NC and miR-210-3p inhibition groups (n = 9).

Values are shown as mean ± SEM. Statistical significance was determined using unpaired, two-tailed Student’s t tests. Sample sizes (n) indicate biological replicates per group; each dot represents one biological replicate (individual mouse). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To assess systemic metabolic improvements, serum lipid profiling demonstrated significant reductions in TC, TG, and LDL-C levels, accompanied by an increase in HDL-C (Figures 7S–7V). Body weight remained comparable between groups throughout the intervention, excluding weight change as a confounding variable (Figure 7R). Collectively, these findings indicate that systemic miR-210-3p inhibition mitigates nicotine-induced atherosclerosis by restoring KLF7 expression, suppressing inflammation, and improving lipid metabolism, ultimately reducing plaque burden.

Discussion

In this study, we elucidate a novel mechanism through which nicotine exacerbates atherosclerosis. We demonstrate that nicotine acts on VAT to reprogram the miRNA cargo of its secreted EVs, leading to a marked enrichment of miR-210-3p. These adipose-derived EVs are then transported to atherosclerotic plaques, where they are internalized by macrophages. Subsequently, EVs-delivered miR-210-3p directly suppresses the expression of the transcription factor KLF7, thereby driving a pro-inflammatory, pro-oxidant, and pro-foam cell phenotypic shift in macrophages that ultimately accelerates plaque progression (Figure 8).

Figure 8.

Schematic mechanism of nicotine-exacerbated atherosclerosis via the adipose-macrophage axis

Nicotine stimulates visceral adipose tissue to release EVs enriched with miR-210-3p, which traffic to atherosclerotic plaques and are internalized by macrophages. EV-delivered miR-210-3p then suppresses KLF7, triggering pro-inflammatory signaling, oxidative stress, and foam cell formation, which accelerates plaque progression.

Nicotine interacts with adipocytes through nicotinic acetylcholine receptors, directly influencing the release of cytokines and free fatty acids from these cells.²⁰ While smoking and obesity are established co-risk factors for coronary heart disease, the specific metabolic impact of nicotine remains controversial, with conflicting reports of both impaired and enhanced insulin sensitivity in clinical versus animal models.^21,22

To bridge the gap between complex clinical comorbidities and isolated animal models, we utilized an in vitro adipose tissue culture model derived from HFD-fed ApoE^−/− mice.²³ This system mimics the inherent metabolic disturbances of obese smokers, allowing us to dissect how nicotine amplifies adipose inflammation and altering its secretion profile in atherosclerosis.

While elevated levels of circulating EVs have been observed in smokers with coronary heart disease, the cellular origin, biodistribution, and functional mechanisms of these EVs remain poorly defined.²⁴ Our study demonstrates that nicotine stimulates EVs release from visceral adipose tissue—a major source of circulating EVs²⁵—and promotes their preferential accumulation within macrophages inside atherosclerotic plaques. Given the regulatory potential of EV cargo, we sought to characterize the functional miRNAs mediating this cross-talk.²⁶ Notably, adipose tissue significantly contributes to the pool of circulating EVs-derived miRNAs, as evidenced by a 4-fold reduction in circulating miRNAs following adipose tissue-specific Dicer knockout in mice.²⁷ In vivo and in vitro analyses demonstrated that nicotine-primed adipose EVs are potently atherogenic: they accelerate plaque growth, are readily internalized by macrophages, and induce a comprehensive pro-atherogenic phenotype. Although nicotine can directly affect macrophages as well as other vascular and immune cells (e.g., endothelial and smooth muscle cells), our EV-depletion and GW4869 experiments indicate that VAT-derived EVs are key mediators of macrophage dysfunction in this model. Nevertheless, direct nicotine effects and other EV-independent pathways cannot be fully excluded. Future studies using genetic disruption of EV biogenesis (e.g., Rab27a or nSMase2 loss-of-function) will be important to further validate these findings.

Our sequencing analysis reveals that nicotine specifically reshapes the miRNA landscape of adipose-derived EVs, culminating in a marked enrichment of miR-210-3p. Notably, obesity-associated miR-210-3p has been traced to adipose tissue macrophages (Patra et al.²⁸). Given that adipocytes are a major source of adipose tissue EVs,^29,30 we fractionated VAT to identify the cellular origin of EV-miR-210-3p. Our analysis revealed that adipocytes are the primary contributors to the nicotine-induced elevation of EV-miR-210-3p, whereas no significant induction was observed in SVF-derived EVs (Figures 1N and S5D). This distinction implies that nicotine does not merely amplify the basal obesity-associated inflammatory signal but rather mobilizes a distinct adipocyte-vascular communication axis. Thus, consistent with a “second-hit” model, nicotine exploits the metabolically primed adipose tissue to unleash a specific, potent pro-atherogenic EV cargo. Clinically, analysis of GEO datasets reveals elevated serum miR-210 levels in smokers and coronary artery disease (CAD) patients. This translational relevance is further supported by Karakas et al., who identified circulating miR-210 as an independent predictor of cardiovascular mortality.³¹ Although miR-210-3p is milieu-dependent, with prior studies demonstrating its cardioprotective effects under acute hypoxia—such as reducing ischemia-reperfusion injury and promoting angiogenesis—its role in our model emerges in the setting of chronic nicotine-associated metabolic inflammation in atherosclerosis.^32,33 Consistent with Virga et al.³⁴ and Karshovska et al.,³⁵ who showed that miR-210 drives macrophage inflammation and necroptosis, we demonstrate that EV-delivered miR-210-3p acts as a potent driver of macrophage dysfunction, enhancing pro-inflammatory cytokine secretion, ROS production, and foam cell formation. These findings underscore the cell-type specificity of miR-210-3p: in contrast to its stabilizing effects on vascular smooth muscle cells (VSMCs), its pro-inflammatory effects on macrophages dominate plaque progression in early atherosclerosis.^32,35 Importantly, we provide the first in vivo evidence that AAV-mediated inhibition of miR-210-3p alleviates nicotine-induced atherosclerosis, identifying it as a critical therapeutic target.

To investigate the downstream mediators through which miR-210-3p regulates macrophage function, we integrated target predictions from TarBase, mirDIP, and TargetScan, identifying KLF7 as a key transcription factor target, which we validated using a luciferase reporter assay. Notably, KLF7 is a zinc-finger transcriptional regulator³⁶ that has been linked to metabolic and cardiovascular disease and is considered a central factor in CAD-associated networks.^37,38 In vascular and myeloid contexts, KLF7 exerts a protective role. For example, it attenuates ox-LDL–induced endothelial ferroptosis and dysfunction and limits macrophage glycolytic reprogramming, oxidative stress, and plaque formation via mechanisms involving transcriptional regulation of ferroptosis (ALKBH5/ACSL4) and metabolism (HDAC4/NCOR1).^39,40

Consistent with this protective profile, in our study, we identified a miR-210-3p-KLF7 regulatory axis that modulates macrophage inflammatory activation in nicotine-accelerated atherosclerosis. Although the precise downstream mediators remain to be fully elucidated, our findings, together with prior literature, support a model in which KLF7 restrains macrophage glucose metabolism and inflammatory programs via multiple pathways. For instance, prior work suggests that KLF7 can restrain macrophage glucose metabolism via the HDAC4-miR-148b-3p-NCOR1 pathway,⁴⁰ which likely contributes to the protective metabolic effects observed in our study. Beyond metabolic control, KLF7 has been reported to induce cyclin-dependent kinase inhibitor p21. Given that p21 inhibits NF-κB signaling by favoring the formation of inhibitory p50-p50 homodimers,^41,42 we speculate that miR-210-3p-mediated suppression of KLF7 compromises this protective KLF7-p21-NF-κB module, thereby unleashing sustained pro-inflammatory signaling in macrophages. Although KLF7 has been reported to modulate inflammatory signaling in adipocytes under obesity-related metabolic stress,⁴³ its role in vascular and myeloid cells supports a context where it primarily limits inflammatory activation. Furthermore, given that KLF7 acts upstream of PPARγ in adipocytes⁴⁴—a master regulator of lipid metabolism and anti-inflammatory gene expression—we speculate that the miR-210-3p-KLF7 axis may similarly impinge on PPARγ-dependent transcriptional programs in macrophages. This would provide an additional layer of immunometabolic regulation, although direct validation is warranted. Finally, while our study focused on nicotine, both miR-210 and KLF7 respond to broader metabolic and hypoxic stressors. Thus, KLF7 emerges not merely as a target, but also as a critical node integrating metabolic reprogramming and inflammatory resolution, suggesting a broader mechanism that could operate beyond smoking-related injury.

In conclusion, our study elucidates a novel mechanism of inter-organ communication in nicotine-induced atherosclerosis, whereby nicotine reprograms the miRNA cargo of adipose-derived EVs to enrich miR-210-3p. Upon trafficking to plaques, these EVs suppress macrophage KLF7, thereby driving inflammatory activation and foam cell formation. These findings not only bridge the gap between metabolic disruption and vascular inflammation but also position the miR-210-3p-KLF7 axis as a promising diagnostic biomarker and therapeutic target for smoking-associated cardiovascular disease.

Limitations of the study

The principal constraints of this study are as follows: First, our analyses focused on EVs derived from visceral adipose tissue, and it remains uncertain whether other fat depots—particularly perivascular adipose tissue—respond similarly to nicotine. Second, although adipose-derived EVs were predominant in our model, EVs from other tissues may also contribute to vascular effects. Third, while KLF7 was validated as a downstream target of miR-210-3p, additional targets of this miRNA may likewise influence macrophage behavior and plaque phenotypes. Fourth, the use of young male ApoE^−/− mice with subcutaneous nicotine administration may not fully recapitulate human smoking, which involves inhalation exposure, aging, and sex-specific factors.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Xuedong Wang (doctor_wang@vip.126.com).

Materials availability

This study did not generate new unique reagents. All materials generated in this study are listed in the key resources table.

Data and code availability

• •Public datasets used in this study were obtained from GEO Database: GSE59421 (Kok et al.) and Database: GSE105449 (de Ronde et al.), which are openly available.
• •The extracellular vesicle (EV) miRNA-seq dataset generated in this study has been deposited in the GEO Database: GSE314694 and will be made publicly available upon publication of the manuscript.
• •This study does not report any original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon reasonable request.

Acknowledgments

We would like to express our gratitude to all those who contributed to this study. We would also like to thank the reviewers and editors for their invaluable input, which has enhanced the quality of this study. This work was supported by the National Natural Science Foundation of China (81900309, 82000330, and U23A20480) and the Open Research Fund of Key Laboratory of Myocardial Ischemia, Ministry of Education (KF202224 and KF202413).

Author contributions

Conceptualization, Y.L., X.W., and J.H.; methodology, M.Y., Q.L., and X.H.; investigation, Y.L., H.L., and M.Y.; data curation, M.Z., Z.Y., Y.L., and Y.S.; writing—original draft, Y.L.; writing—review & editing, all authors, with lead contribution from Q.L., X.W., and J.H.; funding acquisition, X.W., Q.L., and J.H.; resources, J.H. and X.W.; supervision, X.W., Q.L., and J.H.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

No generative AI or AI-assisted tools were used for data analysis or content generation in this work, except for grammar and spelling correction.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
TSG101	Abcam	Cat# ab125011; RRID: AB_10974262
CD63	Abcam	Cat# ab134045; RRID: AB_2800495
CD9	Abcam	Cat# ab263019; RRID: AB_3076464
Calnexin	Abcam	Cat# ab133615; RRID: AB_2864299
F4/80	Abcam	Cat# ab16911; RRID: AB_443548
CD68	Abcam	Cat# ab283654; RRID: AB_2922954
α-SMA	Cell Signaling Technology	Cat# 19245; RRID: AB_2734735
Perilipin-1	Cell Signaling Technology	Cat# 9349; RRID: AB_10829911
GAPDH	Abcam	Cat# ab9485; RRID: AB_307275
KLF7	Absin	Cat# abs140715
KLF7	UpingBio	Cat# YP-Ab-01839
IL-6	Affinity Biosciences	Cat# DF6087; RRID: AB_2838055
IL-1β	Cell Signaling Technology	Cat# 12242; RRID: AB_2715503
TNF-α	Abcam	Cat# ab8348; RRID: AB_306503
SOD2	Novus Biologicals	Cat# NB100-1969; RRID: AB_2191816
Goat Anti-Rabbit IgG	ZsgbBio	Cat# ZB-2301; RRID: AB_2747412
Goat Anti-Mouse IgG	ZsgbBio	Cat# ZB-2305; RRID: AB_2747415
Goat Anti-Mouse IgG (Fc) Polymer (IHC specific)	ZsgbBio	Cat# PV-6002; RRID: AB_2864334
Goat Anti-Rabbit IgG Polymer (IHC specific)	ZsgbBio	Cat# PV-6001; RRID: AB_2864333
Goat Anti-Rabbit IgG/Alexa Fluor 594	ZsgbBio	Cat# ZF-0516; RRID: AB_2936330
Goat Anti-Rat IgG/FITC	ZsgbBio	Cat# ZF-0315; RRID: AB_3662146
Goat Anti-Mouse IgG/FITC	ZsgbBio	Cat# ZF-0312; RRID: AB_2716306
Goat anti-rat IgG	ZsgbBio	Cat# PV-9004; RRID: AB_2868453
Bacterial and virus strains
AAV9-mmu-miR-210-3p inhibition virus	GeneChem	Serotype: AAV9; Vector: GV481; U6-MCS-CAG-mCherry; Titer: 1 × 10ˆ12 vg/mL
AAV9-negative control virus	GeneChem	Serotype: AAV9; Vector: GV481; U6-MCS-CAG-mCherry; Titer: 1 × 10ˆ12 vg/mL
Lentiviral KLF7 overexpression virus	GeneChem	Vector: GV737; CMV-MCS-EF1a-mCherry-T2A-puromycin
Lentiviral negative control virus	GeneChem	Vector: GV737; CMV-MCS-EF1a-mCherry-T2A-puromycin
Chemicals, peptides, and recombinant proteins
Fetal bovine serum (FBS)	ExCell Bio	Cat# FSS500
Calf serum (CS)	Gibco	Cat# 16010159
IBMX	MedChemExpress	Cat# HY-12318
Dexamethasone	MedChemExpress	Cat# HY-14648
Rosiglitazone	MedChemExpress	Cat# HY-17386
Insulin	HTBT	Cat# S20180003
Ox-LDL	Yiyuan Biotechnology	Cat# YB-0010
Nicotine	MICXY reagent	Cat# 54-11-5
GW4869	Sigma-Aldrich	Cat# D1692
PKH67 dye	Invitrogen	Cat# MIDI6-1 KT
DNAfectin™ Plus Transfection Reagent	Applied Biological Materials (abm)	Cat# G2500
DMEM	Gibco	Cat# 11965092
Critical commercial assays
DCFH-DA ROS probe	Beyotime	Cat# S0033S
Oil Red O staining kit	Solarbio	Cat# G1261
BODIPY-cholesterol	MedChemExpress	Cat# HY-125746
Total cholesterol assay kit	Nanjing Jiancheng Bioengineering	Cat# A111-1-1
Triglyceride assay kit	Nanjing Jiancheng Bioengineering	Cat# A110-1-1
LDL-C assay kit	Nanjing Jiancheng Bioengineering	Cat# A113-1-1
HDL-C assay kit	Nanjing Jiancheng Bioengineering	Cat# A112-1-1
RNA Isolation Kit for EVs	Rengenbio	Cat# EXORNA20B-1
BCA Protein Assay Kit	Beyotime	Cat# P0012
ReverTra Ace™ qPCR RT Kit	Toyobo	Cat# FSQ-101
2X M5 HiPer SYBR Premix EsTaq Plus	Mei5bio	Cat# MF787-01
Dual-Luciferase Reporter Assay Kit	Beyotime	Cat# RG027
Deposited data
EV miRNA sequencing data (this study)	Gene Expression Omnibus	Database: GSE314694
public data: mRNA sequence	Gene Expression Omnibus	Database: GSE59421
public data: mRNA sequence	Gene Expression Omnibus	Database: GSE105449
Experimental models: Cell lines
3T3-L1 (mouse preadipocyte)	National Biomedical Experimental Cell Repository	RRID:CVCL_0123; STR-authenticated; Mycoplasma-free
RAW264.7 (mouse macrophage)	National Biomedical Experimental Cell Repository	RRID:CVCL_0493; STR-authenticated; Mycoplasma-free
Experimental models: Organisms/strains
ApoE−/− mice	HFK Bioscience Company	Cat#14012A
Oligonucleotides
mmu-miR-210-3p mimics	RiboBio	Cat#miR10000658-1-5
Mimic negative control (NC)	RiboBio	Cat#miR1N0000001-1-5
Inhibitor negative control (NC)	RiboBio	Cat#miR2N0000001-1-5
mmu-miR-210-3p inhibitor	RiboBio	Cat#miR20000658-1-5
Recombinant DNA
pmirGLO-Klf7-3′UTR-WT	This paper	Custom synthesis
pmirGLO-Klf7-3′UTR-MT	This paper	Custom synthesis
Software and algorithms
GraphPad Prism 10.0	GraphPad	https://www.graphpad.com/features
ImageJ	ImageJ	https://imagej.nih.gov/ij/index.html
TargetScan 8.0	TargetScan	https://www.targetscan.org/
TarBase v.8	DIANA Tools	https://dianalab.e-ce.uth.gr/
mirDIP	mirDIP	https://ophid.utoronto.ca/mirDIP/
Bowtie	Bowtie	https://bowtie-bio.sourceforge.net/

Experimental model and study participant details

Mice

All animal procedures were approved by the Animal Ethics Committee of the Second Hospital of Harbin Medical University (Approval No. YJSDW2022-152) and were performed in accordance with institutional guidelines for animal care and use. Male ApoE^−/− mice (8 weeks old, C57BL/6J background, specific pathogen-free [SPF] grade) were purchased from HFK Bioscience (hfkbio; Cat# 14012A) and housed under SPF conditions (22 ± 2°C, 50–60% humidity, 12-h light/dark cycle) with ad libitum access to food and water; all efforts were made to minimize animal suffering. Sex as a biological variable was not assessed, as only male ApoE^−/− mice were used to reduce variability and maintain consistency with prior atherosclerosis studies; this is a limitation.

Nicotine-accelerated atherosclerosis model

To establish a nicotine-induced atherosclerosis model, mice were randomly assigned to four groups.

• (1)NCD, normal chow diet;
• (2)NCD + nicotine (Ni), receiving nicotine (2 mg/kg/day, MICXY reagent) via subcutaneous injection;
• (3)HFD, high-fat diet (HFKbio, H10141);
• (4)HFD + nicotine (HFD+Ni).

Control groups received equal volumes of saline. After 12 weeks of treatment, mice were euthanized with isoflurane, and tissues were harvested for further analysis. Histological, imaging, and molecular analyses were performed by investigators blinded to group allocation. No animals were excluded except those that died accidentally or developed unrelated illnesses prior to endpoint assessment.

Extracellular vesicles (EVs) transfer experiment

To evaluate whether EVs derived from nicotine-stimulated adipose tissue promote atherosclerosis, visceral adipose tissue (VAT) was obtained from donor ApoE^−/− mice fed either HFD (HFD-EVs) or HFD plus nicotine (HFD+Ni-EVs). Equal tissue masses were processed in parallel to ensure comparable EV yields. Recipient ApoE^−/− mice were maintained on an HFD and received EV injections via the tail vein, administered three times per week for four consecutive weeks. EV dosage was normalized based on EV protein concentration, with each injection containing 0.3 μg/μL EV protein in 100 μL total volume.

AAV-mediated miR-210-3p inhibition in vivo

To determine the contribution of miR-210-3p to nicotine-driven atherosclerosis, the male ApoE^−/− mice (8 weeks old) were randomly divided into two groups and treated with either AAV9-mmu-miR-210-3p inhibition virus or a negative control AAV9 (GeneChem). Mice simultaneously received nicotine and HFD as described above. AAVs were administered via tail vein injection at a dose of 2.5 × 10¹¹ vg in 200 μL per mouse. After 12 weeks of treatment, mice were euthanized for tissue collection and downstream assays.

Cell lines

Mouse 3T3-L1 preadipocytes and RAW264.7 macrophages were obtained from the National Biomedical Experimental Cell Repository (Beijing). Both cell lines were authenticated by short tandem repeat (STR) profiling and were tested negative for mycoplasma contamination prior to use. All cells were cultured at 37°C in a humidified 5% CO₂ incubator. 3T3-L1 cells were induced into adipocytes and then subjected to further experiments. RAW264.7 cells were cultured in DMEM supplemented with 10% FBS (ExCell Bio). To induce foam cell formation, macrophages were treated with oxidized LDL (ox-LDL; Yiyuan Biotechnology) at a final concentration of 80 μg/mL for 24–48 h depending on the experimental design.

Method details

Adipose tissue, adipocyte culture and drug treatment in vitro

Visceral adipose tissue (VAT) was isolated from 8-week-old male ApoE^−/− mice fed a high-fat diet for four weeks. For ex vivo culture, 100–200 mg of VAT from each mouse was washed and incubated in serum-free DMEM (Gibco) supplemented with or without nicotine (10 μM) for 24 h at 37°C. After incubation, the conditioned medium was collected for subsequent EV isolation or macrophage stimulation.

To investigate the effects of adipose-derived EVs on macrophages, two experimental systems were established.

• 1.Tissue-level model
  Macrophages were treated with:

  • a.
    EVs derived from control VAT (AT-EVs);
  • b.
    EVs derived from nicotine-treated VAT (AT-Ni EVs);
  • c.
    untreated medium (control).

• 2.Cell-level model
  Differentiated 3T3-L1 adipocytes were cultured with or without nicotine (10 μM) for 24 h, and EVs were isolated to generate:

  • a.
    EVs from control adipocytes (3T3-L1 EVs);
  • b.
    EVs from nicotine-treated adipocytes (3T3-L1-Ni EVs);
  • c.
    no-EV control medium.

To determine the role of EV release in adipocyte–macrophage communication, GW4869 (20 μM) was used to inhibit EV biogenesis in cultured adipocytes or VAT explants.

Two models were constructed.

• 1.Adipocyte–macrophage co-culture model (3T3-Raw)
  Co-culture was performed for 24 h and divided into four groups.

  • a.
    Raw (vehicle control);
  • b.
    3T3-Raw (vehicle co-culture control);
  • c.
    3T3-Raw-Ni (nicotine-treated);
  • d.
    3T3-Raw-Ni-GW4869 (nicotine + EV inhibitor).

• 2.VAT-derived supernatant stimulation model (sAT-Raw)
  Raw264.7 macrophages were treated with VAT supernatant (sAT) under three conditions.

  • a.
    sAT-Raw (vehicle control);
  • b.
    sAT-Raw-Ni (nicotine-treated);
  • c.
    sAT-Raw-Ni-GW4869 (nicotine + EV inhibitor).

Lipogenic induction of 3T3-L1 cells

The medium configuration was as follows: (1) basal medium: DMEM +10% calf serum (CS); (2) differentiation induction medium: DMEM +10% FBS supplemented with IBMX (0.5 mmol/L; MedChemExpress), dexamethasone (1 μmol/L; MedChemExpress), insulin (5 μg/mL; HTBT), and rosiglitazone (2 μmol/L; MedChemExpress); and (3) maintenance medium: DMEM +10% FBS + insulin (5 μg/mL).3T3-L1 preadipocytes were cultured in basal medium until they reached full confluence and were then maintained for 48 h to achieve growth arrest. Adipogenic differentiation was initiated by replacing the basal medium with differentiation induction medium for 48 h. Thereafter, cells were cultured in maintenance medium, which was renewed every 2 days, until more than 90% of cells exhibited uniform intracellular lipid droplets.⁴⁵ Fully differentiated adipocytes were then used for subsequent experiments.

Extraction, identification and staining of adipose derived EVs

Conditioned medium from adipose tissue explants or mature 3T3-L1 adipocytes was collected and centrifuged at 300 × g for 10 min to remove cells, and the pellet was discarded. The supernatant was then centrifuged at 3,000 × g for 10 min and at 10,000 × g for 30 min at 4°C to remove cell debris, large vesicles, and apoptotic bodies. The resulting supernatant was ultracentrifuged at 100,000 × g for 70 min at 4°C. The pellet was resuspended in PBS and ultracentrifuged again at 100,000 × g for 70 min at 4°C to further purify extracellular vesicles (EVs). The final EV pellet was resuspended in PBS for subsequent experiments.The morphology of EVs was examined by transmission electron microscopy (TEM; H-7500, Hitachi, Japan), and EV protein concentration was determined using a BCA Protein Assay Kit (Beyotime). Nanoparticle tracking analysis (NTA) was performed to assess particle size distribution and concentration. For in vitro uptake assays, EVs were labeled with PKH67 (Invitrogen) according to the manufacturer’s instructions. After labeling, EVs were washed with PBS and collected by ultracentrifugation at 100,000 × g to remove excess dye, and the pellet was resuspended in PBS. PKH67-labeled EVs were added to macrophages cultured in 2 4-well plates, and EV uptake was observed under a fluorescence microscope at 2 h and 6 h after incubation.For in vivo tracking experiments, PKH67-labeled EVs derived from VAT were injected into ApoE^−/− mice at a dose of 0.3 μg/μL (100 μL per mouse).

Isolation of RNA from EVs

Total RNA was isolated from 300 μL of purified EVs using the RNA Isolation Kit for EVs (Rengenbio, Cat.EXORNA20B-1), following the manufacturer’s instructions. An exogenous spike-in control (C. elegans miR-39-3p) was added to the samples for normalization. The extraction process involved sequential treatment with Lysate A, Lysate B, anhydrous ethanol, Wash Solution A, and Wash Solution B using adsorbent columns. The final RNA was eluted in Eluent A and stored at −80°C for downstream experiments.

Cell transfection

RAW264.7 macrophages were seeded to reach 30–50% confluence at the time of transfection. miRNA mimics or inhibitors were diluted in RNase-free water and transfected into cells using riboFECT CP Reagent (RiboBio) according to the manufacturer’s instructions in antibiotic-free medium, and cells were incubated for 48 h before subsequent experiments. The following oligonucleotides were used (all from RiboBio): mmu-miR-210-3p mimic, micrON mimic negative control, micrOFF inhibitor negative control, and micrOFF mmu-miR-210-3p inhibitor. For lentiviral transduction, KLF7-overexpressing lentivirus (GeneChem; vector GV737; CMV-MCS-EF1α-mCherry-T2A-puromycin) and the corresponding negative control lentivirus (GeneChem) were transduced into RAW264.7 cells using HiTransG P (GeneChem) following the supplier’s protocol. Transduced cells were selected with puromycin (2 μg/mL; MedChemExpress).

Immunohistochemical staining

Briefly, after paraffin embedding, the sections were sequentially sectioned, dried, dewaxed, and hydrated. Antigen retrieval was performed, followed by sequential incubation of primary and secondary antibodies. Finally, chromogenic detection was performed using DAB, cell nuclei were stained, and sections were mounted. The results were observed under a microscope and recorded. The antibodies used were: anti-IL-6 (1:100), anti-IL-1β (1:100), anti-TNFα (1:100), anti-SOD2 (1:100), anti-F4/80 (1:100), anti-TSG101 (1:100), and anti-KLF7 (1:100). Secondary antibodies included HRP-conjugated goat anti-mouse IgG (Fc), goat anti-rabbit IgG polymer, and goat anti-rat IgG polymer.

Immunofluorescence

For immunofluorescence staining of cells on coverslips or frozen sections, a cell suspension was prepared and seeded onto coverslips, and cells were treated with the indicated drugs for 48 h. Next, paraformaldehyde fixation was performed, followed by 0.3% Triton X-100 permeabilization and blocking with goat serum. Primary antibodies were incubated at 4°C overnight. The following day, samples were incubated with fluorescent secondary antibodies at room temperature for 1 h in the dark and then observed using laser confocal microscopy. The antibodies used were: anti-TSG101 (1:200), anti-F4/80 (1:200), anti-PLIN1 (1:200), and anti-KLF7 (1:100). Secondary antibodies included goat anti-rabbit IgG/Alexa Fluor 594 (1:200), FITC-labeled goat anti-rat IgG (1:200), and FITC-labeled goat anti-mouse IgG (1:200).

Cell and tissue RNA extraction, reverse transcription and quantitative Real-time PCR (qRT-PCR) reaction

Tissue was ground into powder and added to TRIzol Reagent (Invitrogen, USA), while cell samples were added to TRIzol Reagent after discarding the supernatant. The procedure of RNA extraction was as follows: chloroform (350 μL per 1 mL TRIzol; Xilong Scientific) was added, and the mixture was then centrifuged for 15 min at 12000 rpm at room temperature. The supernatant was mixed with 500 μL of isopropanol (Xilong Scientific) and left to stand for 10 min at room temperature, then centrifuged for 10 min at 12000 rpm at 4°C. After discarding the supernatant, 1 mL of 75% anhydrous ethanol (Xilong Scientific) was added and centrifuged for 5 min at 7500 rpm at 4°C. The supernatant was discarded, and the white solid at the bottom was preserved; when the precipitate was sufficiently dried, enzyme-free water was added, and the concentration of RNA was measured and the RNA was stored at −80°C for later use. Subsequently, cDNA was obtained by reverse transcription reaction using Toyobo reverse transcription kit, followed by qRT-PCR reaction using 2X M5 HiPer SYBR Premix EsTaq plus (with Tli RNaseH) (Mei5bio).

Relative gene expression was calculated using the 2ˆ-ΔΔCt method. Data were normalized to GAPDH for mRNA targets, to U6 for cellular miRNAs, and to cel-miR-39-3p for EV-derived small RNAs. Candidate miRNAs for RT-qPCR validation were selected from the RNA-seq dataset based on a combination of high fold-change (log2 fold-change ≥1.5), robust expression levels (high raw read counts), and reported relevance to atherosclerosis or macrophage inflammation. Low-abundance transcripts or those with unknown functions in vascular biology were excluded from the validation panel. The analyzed mRNA targets included MCP-1, IL-1β, IL-6, TNF-α, iNOS, KLF7, NF-κB, NLRP3, MMP3, MMP9, PPARγ, GPDL1, B4GALT5, NDUFA4, PPTC7, KMT2D, CPEB2, TET2, HIF3A, KCMF1, and GAPDH. The analyzed miRNA targets included miR-210-3p, miR-210-5p, miR-192-3p, miR-205-5p, miR-129-2-3p, miR-1a-3p, miR-137a-3p, miR-449a-5p, U6, and cel-miR-39-3p. Sequences of primers used in qRT-PCR are shown in Table S1.

EVs-MicroRNA sequencing and analysis

In total, 100 mg adipose tissue from ApoE^−/− mice was used for extracting EVs. Total RNA was isolated from extracted adipose EVs (HFD group vs. HFD+Ni group). The raw image data files obtained from sequencing on the Illumina HiSeq2500 platform were converted into raw sequence reads by base calling, and then the raw data were subjected to quality control. Using Bowtie software, clean reads were aligned with the Silva database, GtRNAdb database, Rfam database and Repbase database to filter ribosomal RNA (rRNA), transfer RNA (tRNA), small nuclear RNA (snRNA), small nucleolus RNA (snoRNA) and other ncRNAs and repeat sequences. The ncRNAs and repetitive sequences were removed to obtain unannotated reads containing miRNAs, which were used for subsequent analysis. The mouse reference genome was downloaded from ftp://ftp.ensembl.org/pub/release-101/fasta/mus_musculus/dna. Unannotated reads were aligned with the reference genome using Bowtie software to obtain the position information on the reference genome, i.e., mapped reads. miRDeep2 was used for the identification of known and novel miRNAs and for further analysis of miRNA differential expression. Small RNA sequencing and analysis were performed by Echo Biotech Co., Ltd. (Beijing, China). Eight samples were sent for testing, one of which was rejected due to quality control failures, leaving seven samples for further analysis. Finally, we completed small RNA sequencing of 7 samples, detecting 1,479 miRNAs, of which 920 were known miRNAs and 559 were newly predicted miRNAs. The miRNA expression abundance of each sample was quantified, and differentially expressed miRNA screening results were obtained (a total of 26 differentially expressed miRNAs, of which 23 were differentially up-regulated and the other 3 were differentially down-regulated). A total of 21,776 miRNA target genes were predicted, and the differentially expressed miRNAs were subjected to GO classification of target genes, GO enrichment analysis, KEGG functional annotation of target genes and KEGG pathway enrichment analysis of target genes. Detailed sequencing information is provided in the Table S2. The EV miRNA sequencing data generated in this study have been deposited in the GEO Database: GSE314694.

Western blot protein assay

Protein was extracted from cell or tissue samples using RIPA lysate, followed by sonication, centrifugation, denaturation, and storage at −20°C. SDS-PAGE gels (Epizyme) were prepared, and equal volumes of samples were loaded. Electrophoresis was performed at a constant voltage of 20 V for 20 min, followed by 80 V for 30 min, and finally 120 V for separation. The membrane was transferred while wet using a constant current of 300 mA for 30 min. Following this, the membrane was blocked with 5% skimmed milk at room temperature for 1 h. Next, the membrane was incubated with primary antibodies and washed. Finally, the membrane was incubated with secondary antibodies, washed, developed, and photographed. The antibodies used were: anti-TSG101 (1:1000; ∼45 kDa), anti-CD63 (1:1000; ∼26–65 kDa), anti-CD9 (1:1000; ∼25 kDa), anti-Calnexin (1:1000; ∼75 kDa), anti-KLF7 (1:1000; ∼41 kDa), anti-GAPDH (1:1000; ∼36 kDa). Optical density was used to quantify band intensities from three or more independent experiments using ImageJ software. Relative expression levels of target proteins were normalized to GAPDH. All uncropped gels are listed in the Data S1.

Detection of intracellular reactive oxygen species (ROS) levels

Cells were treated with the corresponding drugs, miRNA mimics, or miRNA inhibitors for the indicated durations. The DCFH-DA probe (Beyotime, S0033S) was diluted 1:1000 in serum-free medium and added to the cells. The culture flask was gently shaken during this period. After incubation for 20–30 min at 37°C in the dark, cells were washed three times with PBS to remove excess probe. Fluorescence was observed and photographed using a standard fluorescence microscope. Mean fluorescence intensity was quantified using ImageJ software.

Oil red O staining

Intracellular lipid accumulation was assessed using a commercially available Oil Red O staining kit (Solarbio). The stain was freshly prepared and applied for 15 min. The sample was then permeabilised with drops of 60% isopropanol for 40 s, followed by drops of filtered haematoxylin stain for 40 s. The staining process was terminated by drops of tap water, and the sample was photographed using a light microscope.

BODIPY-cholesterol staining

Sterol uptake was assessed using BODIPY–cholesterol (MedChemExpress), a fluorescent cholesterol analog. BODIPY-cholesterol can be used to monitor sterol uptake and inter-organelle sterol flux in cells. The analogue is reconstituted in DMSO to a stock concentration of 2 mM. Cells diluted with PBS are washed (300×g, 10 min) to remove excess diluent and the loose pellet resuspended with vehicle media to a proper concentration. Cell suspensions are labeled with BODIPY-cholesterol (final concentration of 1 μM), mixed thoroughly and incubated for 10–20 min at 37°C. The cells are diluted with each of the capacitation media conditions as outlined in Incubation media to a final concentration and incubated for 2 h prior to flow cytometric assessment.

Measurement of serum lipid parameters

Serum lipid profiles, including total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C), were assessed using commercial assay kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer’s instructions. Briefly, serum samples were thawed on ice and diluted appropriately with phosphate-buffered saline. Enzymatic colorimetric methods were employed for all measurements: TC and TG levels were determined via cholesterol oxidase-peroxidase and glycerol phosphate oxidase-peroxidase methods, respectively, while HDL-C and LDL-C were measured using direct methods with polyethylene glycol-modified enzymes. Absorbance was read at appropriate wavelengths using a microplate spectrophotometer, and concentrations were calculated based on standard curves generated with provided calibrators. All samples were assayed in duplicate.

Dual-luciferase reporter assay

293T cells were cultured in DMEM supplemented with 10% FBS at 37°C in 5% CO₂. The wild-type mouse Klf7 3′UTR fragment containing the predicted miRNA binding site (5'- … GCATGAATGCACGCACACGCCAGGGAT … -3') and the corresponding mutant fragment (5'- … GCATGAATGCTGCGTGTCGCCAGGGAT … -3') were synthesized and cloned into the pmirGLO vector (Generalbiol). The full sequences of the constructed plasmids are provided in Table S3. Cells were co-transfected with these reporter plasmids and miR-210-3p mimics or negative controls using DNAfectin Plus Transfection Reagent (abm). After 48 h, luciferase activity was measured with the Dual-Luciferase Reporter Assay Kit (Beyotime, China) according to the manufacturer’s instructions. Firefly luciferase activity was normalized to Renilla luciferase activity, and results were expressed as the Firefly/Renilla ratio.

Quantification and statistical analysis

All statistical analyses were performed using GraphPad Prism 10.0 software. Data are presented as mean ± SEM. Unless otherwise stated, “n” represents the number of independent biological replicates per group (e.g., individual mice or independent cell culture experiments), and exact sample sizes are indicated in the corresponding figure legends.

Sample Size Determination: Sample sizes were determined based on established standards in the field for ApoE^−/− atherosclerosis models and our pilot experiments indicating robust effect sizes for the primary endpoints. Although a formal a priori power analysis was not performed, post hoc analysis confirmed that the selected sample sizes provided sufficient statistical power to detect biologically meaningful differences.

Public datasets: Publicly available miRNA expression datasets were obtained from GEO (Database: GSE59421, GSE105449) and analyzed as described in the corresponding supplementary figure.

Statistical Comparisons: Data normality was assessed using the Shapiro-Wilk test or Kolmogorov-Smirnov test. Homogeneity of variance was verified using the F-test (for two groups) or Brown-Forsythe test (for multiple groups). For normally distributed data with equal variance: Comparisons between two groups were performed using unpaired two-tailed Student’s t-tests. Comparisons among three or more groups were performed using one-way ANOVA followed by Tukey’s multiple comparisons test. P-value <0.05 was considered statistically significant (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

Published: February 26, 2026